An unusual, completely miscible, ternary hydrogen-bonded polymer blend of phenoxy, phenolic, and PCL

(1)

An Unusual, Completely Miscible, Ternary Hydrogen-Bonded Polymer

Blend of Phenoxy, Phenolic, and PCL

Shiao-Wei Kuo,* Shih-Chi Chan, Hew-Der Wu, and Feng-Chih Chang Institute of Applied Chemistry, National Chiao Tung University, Hsinchu, Taiwan Received December 21, 2004; Revised Manuscript Received March 17, 2005

ABSTRACT: We have investigated the miscibility and hydrogen-bonding behavior of ternary blends of phenoxy, phenolic, and poly(-caprolactone) (PCL) by using differential scanning calorimetry (DSC) and Fourier transform infrared spectroscopy. On the basis of DSC analyses, we observed a rare totally miscible ternary hydrogen-bonded polymer blend in the amorphous phase: all compositions of this ternary blend display a single glass transition temperature. In addition, these single glass transition temperatures can be predicted well by extending the Kwei equation from the binary polymer blend to this ternary polymer blend. The infrared spectra indicate that the intermolecular hydrogen bonding of each pair of binary components still exists in the ternary polymer blend. We used the ternary totally miscible blend to determine the interassociation equilibrium constant between the hydroxyl groups of phenolic and the hydroxyl groups of phenoxy indirectly from the fraction of hydrogen-bonded carbonyl groups of PCL. Quantitative analyses suggest that interassociation between the hydroxyl groups of phenolic and the hydroxyl groups of phenoxy is more favorable than the hydroxyl-carbonyl interassociations of either phenolic/PCL or phenoxy/PCL and the hydroxyl-hydroxyl self-association of the pure phenolic and phenoxy homopolymers at room temperature.

Introduction

Polymer blending is a convenient and attractive route for obtaining new polymeric materials, and there is considerable interest in the phenomenon of miscibility in binary polymer blends.1-3_{In contrast, ternary}

poly-mer blends have received relatively less attention because of the complexity of calculating their phase diagrams and the problems associated with experimen-tal accuracy. Although increasing the number of poly-mer components does indeed lead to complications, there are several good reasons for studying the phase behavior of ternary polymer blends, especially because of their significant industrial importance. For example, Scott4

and Tompa5_{reported ternary polymer blends in which}

a polymer B, miscible with each of polymers A and C, compatibilizes the immiscible binary pair A-C. Clas-sical examples of this system include the ternary blends of poly(vinylidene fluoride) (PVDF)/poly(methyl meth-acrylate) (PMMA)/poly(ethyl methmeth-acrylate) (PEMA),6

PVPh/PMMA/PEMA,7_{and SAN/PMMA/PEMA.}8_When

all three binary pairs (B-A, B-C, and A-C) are individually miscible, a completely homogeneous or a closed immiscibility loop phase diagram may exist.9_The

phase separation is caused by differences in the solubil-ity parameters and interassociation equilibrium con-stants of the binary systems, the so-called “∆χ”10_and

“∆K”9_{effects observed in ternary polymer blends such}

as the phenoxy/PMMA/poly(ethylene oxide) (PEO),11

PVPh/poly(vinyl acetate) (PVAc)/PEO,9

poly(styrene-co-acrylic acid)/PMMA/PEO,12 _{and phenolic/PEO/PCL}13

blend systems. Totally miscible ternary polymer bends may offer unique opportunities to develop new polymer materials from a flexible combination of the three components, but only a very few ternary polymer blends have been reported to be homogeneous over their entire range of compositions. These totally miscible ternary blends include poly(epichlorohydrin) (PECH)/PMMA/

PEO,14 _{PVDF/PVAc/PMMA,}15 _{PECH/PVAc/PMMA,}16

poly(3-hydroxybutyrate)/PEO/PECH,17 _poly(ether

di-phenyl ether ketone)/poly(ether ether ketone)/poly(ether imide) (PEI),18_{PEI/poly(ethylene terephthalate) (PET)/}

poly(butylenes terephthalate) (PBT),19 _and

PCL/poly-(phenyl methacrylate)/poly(benzyl methacrylate),20_which

all possess low ∆χ effects and hydrogen-bonding interac-tions between their polymer segments. As a result, they exist as totally miscible ternary polymer blend systems because the ∆K effect may be neglected. Coleman and Painter have noted9_{that only in very rare cases, such}

as the PVPh/PVAc/poly(methyl acrylate) (PMA) ternary blend,9_{can completely miscible ternary polymer blends}

exist because the ∆χ and ∆K interactions must be so finely balanced. The chemical structures of the PVAc and PMA repeat units are isomorphous, and thus, this ternary polymer blend displays a completely homoge-neous amorphous phase. We became curious about the following question: Other than isomers of two polymers, is it possible to obtain a totally miscible ternary polymer blends in which hydrogen bonding exists between the respective polymer segments?

In this paper, we report another completely miscible ternary hydrogen-bonded polymer blend: that between phenoxy, phenolic, and PCL. Phenolic and phenoxy are well-known hydrogen bond donor polymers that interact favorably with polyacrylate, polyester, polyether, and polyvinylpyridine. To the best of our knowledge, how-ever, only a few reports exist that describe binary polymer blends incorporating these self-association polymers. We reported that the phenolic/phenoxy21-23

and PVPh/phenoxy24_{blends are totally miscible in the}

amorphous phase because of the hydrogen bonds that exist between their polymer segments. In those previous studies,21-24_{we calculated the interassociation}

equilib-rium constants between each binary blend from analy-ses of suitable model compounds. The interassociation equilibrium constants obtained from model compounds, however, are not exactly the same as those of the true polymer blends because of intramolecular screening and * To whom correspondence should be addressed: e-mail

[email protected]; Tel 886-3-5131512; Fax 886-3-5719507.

(2)

functional group accessibility effects25-30 _{in miscible}

polymer blends. In our present study, we determined indirectly, for the first time, the interassociation equi-librium constant between the hydroxyl groups of phe-nolic and the hydroxyl groups of phenoxy from a least-squares fitting procedure based on the experimental fraction of hydrogen-bonded carbonyl groups in the phenolic/phenoxy/PCL ternary blend.

Experimental Section

Materials. The polymers used in this study were phenolic,

poly(hydroxy ether of bisphenol A) (phenoxy), and poly(-carprolactone) (PCL). The phenolic was synthesized through a condensation reaction using sulfuric acid; its average mo-lecular weights were Mn) 500 and Mw) 1200. The phenolic resin does not contain any reactive methylol groups that are capable of causing cross-linking upon heating. The phenoxy was obtained from Union Carbide (Mn) 23 000; Mw) 48 000). The PCL used in this study was TONE Polymer P-787 (Mn) 80 000) purchased from Union Carbide.

Preparation of Blend Samples. Ternary polymer blends

of phenoxy/phenolic/PCL having various compositions were prepared by solution blending. A tetrahydrofuran (THF) solution containing 5 wt % polymer mixture was stirred for 6-8 h and then left to evaporate slowly at room temperature for 1 day. The film of the blend was then dried at 50 °C for 2 days to ensure total removal of the residual solvent.

Differential Scanning Calorimetry (DSC). The glass

transition temperatures (Tg) of the polymer blends were determined by differential scanning calorimetry (Du-Pont, DSC model 2900). A sample (5-10 mg) was placed on the DSC cell and then heated at a scan rate of 20 °C/min within the range from 0 to 150 °C to avoid possible ester interchange known to occur above 170 °C, and then the specimen was quickly cooled to -100 °C after the first scan. The value of Tg was obtained as the midpoint of the transition point of the heat capacity (Cp) change at a scan rate of 20 °C/min over a temperature range from 0 to 150 °C.

Infrared Spectra. Infrared spectra were recorded using a

Nicolet Avatar 320 FT-IR spectrometer. In all cases, at least 32 scans with an accuracy of 1 cm-1were signal-averaged. Infrared spectra of polymer blend films were determined using the conventional NaCl disk method. A THF solution containing the blend (5% w/v) was cast onto a NaCl disk and dried under conditions similar to those used in the bulk preparation. The films used in this study were sufficiently thin to obey the Beer-Lambert law.

Results and Discussion

Binary Blend System. We used differential scan-ning calorimetry to assess the miscibility of the polymer blend by measuring the glass transition temperature of the blend composition. Figure 1 displays the values of Tgobtained using various compositions of each binary

blend of phenolic/phenoxy,23_{phenolic/PCL,}31_and

phe-noxy/PCL.32 _{All these binary compositions exhibit a}

single Tg, which strongly suggests that all of these

compositions are miscible and possess a homogeneous phase. The dependence of Tgon the composition of these

blends is presented in Figure 1; these plots fit well to the Kwei equation33

where W1 and W2 are the weight fractions of the

components, Tg1and Tg2represent the components’ glass

transition temperatures, and k and q are fitting con-stants. The Kwei equation can apply to polymers that possess specific interactions, such as hydrogen bonds,

within blend systems. The parameter q corresponds to the strength of hydrogen bonding; it reflects the balance between breaking the intramolecular hydrogen bonds and forming intermolecular hydrogen bonds. Figure 1 indicates that q ) -10 and k ) 1 for the phenolic/PCL blend, q ) -100 and k ) 1 for the phenoxy/PCL blend, and q ) -15 and k ) 1 for the phenolic/phenoxy blend. The deviation in the values of Tgin Figure 1a can be

interpreted as reflecting the fact that, upon blending, the self-association of phenolic is broken, with its hydroxyl groups becoming diluted within the blend. The phenoxy molecule, with its long repeating unit, provides a smaller number of potential hydrogen-bonding sites that are available to form interactions with the other blend components. The extent of forming interassocia-tion hydrogen bonds is too small to overcome the increasing entropy due to the reduction of the number of self-associating hydrogen bonds of each polymer’s hydroxyl groups.23 _{The deviation of the experimental}

value of Tgfrom the Kwei equation at high PCL content

in Figure 1b,c is due to the crystallization of PCL in the blends during quenching. This phenomenon indi-cates not only that crystallization of PCL in the blends changes the amorphous phase but also that the crystal of PCL acts as a physical cross-linking point that hinders the molecular mobility of the amorphous phase.31

Ternary Blend System. Thermal Analyses. Figure 2 displays the DSC thermograms of several phenolic/ T_g)W1Tg1+ kW2Tg2

W₁+ kW₂ + qW1W2 (1)

Figure 1. Plots of Tgvs composition for the individual binary

blends: (a) phenolic/phenoxy, (b) phenoxy, and (c) phenolic/ PCL.

Figure 2. DSC thermograms of phenoxy/phenolic/PCL (w/w/ w) blends having different compositions.

(3)

phenoxy/PCL ternary blends having various composi-tions; it reveals that each ternary blend has only a single glass transition temperature. A single value of Tg

strongly suggests that the ternary polymer blend is fully miscible over its total range of compositions. On the basis of this evidence, we suggest that any phenolic/ phenoxy/PCL blend composition is miscible at temper-atures within the range from -80 to +120 °C. Figure 3 displays the phase diagram of this ternary polymer blend with respect to the value of Tgof each composition.

Clearly, the value of Tgof this ternary blend increases

upon increasing the phenoxy content at a constant phenolic/PCL ratio and upon increasing the phenolic content at a constant phenoxy/PCL ratio.

The well-known Fox equation34_{has been proposed to}

predict variations of glass transition temperatures of copolymers and blends as a function of composition:

where Tgis the glass transition temperature of the blend

and W is the weight fraction, whose subscripts “1” and “2” indicate polymer 1 and 2, respectively. This equation is generally applied to binary blend systems that are compatible and not too strongly polar. For three-component mixing, this equation can be extended to the Tg-composition relationship

In addition, the weight-average values calculated from the linearity prediction has been determined in totally ternary miscible blends, such as PEO/PHB/PECH,17_as

follows:

Table 1 summarizes the calculated values of Tg and

those measured from the DSC analyses of each compo-sition. It is obvious that the values of Tgcalculated from

the Fox equation and the linearity prediction do not fit well to the values of Tg obtained by thermal analysis;

for many compositions, we observe a large negative deviation. This negative deviation between the experi-mental data and values calculated from the Fox equa-tion and the linearity predicequa-tion probably arises from the ternary blend system containing some strong

inter-molecular interactions. Therefore, we extended the Kwei equation for a binary blend to describe a ternary blend such that we could predict the nature of this ternary hydrogen-bonded polymer blend system:

where qijis the interassociation strength of each binary

blend that has been calculated previously. Table 1 also summarizes the values of Tgcalculated from the Kwei

equation and those measured from the DSC analyses at each composition. Figure 4 indicates that the agree-ment between the experiagree-mental and calculated values is quite satisfactory.

FT-IR Spectroscopic Analyses. Fourier transform infrared spectroscopy has been used widely in the study of polymer blends. This method is useful for verifying the presence of intermolecular interactions between various hydrogen bond donor and acceptor groups because of its sensitivity to hydrogen bond formation. Figure 5 displays scale-expanded infrared spectra (in the region 4000-2700 cm-1) of the hydroxyl group stretching absorptions of pure phenolic, pure phenoxy, and various phenoxy/phenolic/PCL blends having their phenoxy content fixed at 50 wt %. Pure phenolic and phenoxy present two distinct bands in the hydroxyl stretching region of the infrared spectra. We attribute the very broad bands centered at 3350 and 3400 cm-1 to the wide distribution of the hydrogen-bonded hy-droxyl groups and the sharp bands at 3525 and 3570 cm-1to the free hydroxyl groups of pure phenolic and pure phenoxy, respectively. The intensity of the signal of the free hydroxyl groups decreased upon increasing the PCL content in the phenoxy/PCL blend. Meanwhile, the signal for phenoxy’s broad hydrogen-bonded

hy-Figure 3. Ternary phase diagram of the phenoxy/phenolic/ PCL system with respect to the individual values of Tg

displayed in each cycle.

1 T_g) W1 T_g1+ W2 T_g2 (2) 1 T_g) W₁ T_g1+ W₂ T_g2+ W₃ T_g3 T_g) W₁T_g1+ W₂T_g2+ W₃T_g3 (3)

Table 1. Characteristics of Ternary Polymer Blends (°C)

prediction composition

phenoxy/phenolic/PCL exptl data Kwei Fox linear

10/0/90 -49 -53 -54 -45 10/18/72 -29 -31 -36 -22 10/36/54 -1 -7 -16 1 10/54/36 16 16 7 24 10/72/18 35 42 35 46 10/90/0 51 58 69 70 30/0/70 -34 -34 -30 -14 30/14/56 -11 -14 -15 4 30/28/42 7 7 3 22 30/42/28 29 29 24 40 30/56/14 49 49 47 57 30/70/0 71 72 75 76 50/0/50 -6 -6 -4 18 50/10/40 9 9 10 31 50/20/30 25 25 25 43 50/30/20 44 43 42 56 50/40/10 62 61 60 69 50/50/0 79 79 82 82 65/0/35 21 21 20 42 65/7/28 34 33 32 51 65/14/21 46 45 44 60 65/21/14 59 58 57 69 65/28/7 74 72 71 78 65/35/0 82 82 87 87 80/0/20 50 50 49 33 80/4/16 59 59 57 71 80/8/12 65 65 65 76 80/12/8 73 73 73 81 80/16/4 82 82 82 87 80/20/0 89 89 91 92 0/20/80 -38 -38 -44 -36 0/40/60 -14 -14 -24 -10 0/60/40 12 12 0 14 0/80/20 31 39 29 40 T_g) W₁T_g1+ W₂T_g2+ W₃T_g3+ q₁₂W₁W₂+ q₁₃W₁W₃+ q₂₃W₂W₃ (4)

(4)

droxyl band shifted to higher frequency upon increasing the PCL content (50 wt %) at 3430 cm-1. Taking into account the interassociation and self-association equi-librium constants of the phenoxy/PCL blend, the inter-association equilibrium constant between the hydroxyl group of phenoxy and the carbonyl group of PCL (KA)

7) is smaller than the self-association equilibrium constant of pure phenoxy35_(K

B) 25.6). This observed

change in the hydroxyl stretching region arises from switching from strong intramolecular hydroxyl-droxyl hydrogen bonds to weaker intermolecular hy-droxyl-carbonyl hydrogen bonds. In addition, the broad

hydrogen-bonded hydroxyl band for phenolic and phe-noxy shifts to a slightly lower frequency (3335 cm-1) upon increasing the phenoxy content (50 wt %), which suggests that the interassociation equilibrium constant between the hydroxyl groups of phenolic and phenoxy is greater than the self-association equilibrium con-stants of the hydroxyl groups of the pure phenolic and pure phenoxy homopolymers. We provide further evi-dence in the next section.

In addition, Figure 5 indicates that, for the ternary blend, the band at 3430 cm-1shifted to 3335 cm-1, a lower wavenumber, upon increasing the phenolic/PCL ratio. This change arose from a switch from lecular hydroxyl-carbonyl hydrogen bonds to intermo-lecular hydroxyl-hydroxyl hydrogen bonds between the phenolic and phenoxy segments, which indicates that there are hydrogen-bonding interactions between the hydroxyl groups of phenoxy and phenolic resin. It also reveals that the hydroxyl-hydroxyl intermolecular interactions between the phenolic and phenoxy seg-ments dominates in these ternary blends, and thus, it is reasonable to assign the band at 3335 cm-1 to the hydroxyl groups of phenolic and phenoxy segments that are hydrogen bonded to other hydroxyl groups. Coleman et al.36_{employed the frequency difference (∆ν) between}

the absorptions of the hydrogen-bonded and free hy-droxyl groups to determine the relative strength of different intermolecular interactions. We obtained an increase in the ∆ν value upon increasing the phenolic/ PCL ratio; this result is consistent with the result of the DSC analyses. The value of Tgof this ternary blend

increased upon increasing the phenolic content because the average strength of hydrogen bonding increased.

Figure 6 displays the carbonyl stretching region (1680-1780 cm-1) of these ternary blends from infrared spectra recorded at room temperature. The absorption

Figure 4. Relationship between the values of Tgobserved by

DSC (9) and those calculated using the Kwei equation (0).

Figure 5. Infrared spectra recorded in the region 2700-3700 cm-1for a series of compositions of phenoxy/phenolic/PCL (wt/ wt/wt %) blends: (a) pure phenoxy, (b) 50/10/40, (c) 50/20/30, (d) 50/30/20, (e) 50/20/30, (f) 50/40/0, (g) 50/50/0, and (h) pure phenolic.

Figure 6. Infrared spectra recorded at room temperature in the range 1680-1780 cm-1for phenoxyl/phenolic/PCL blends: (a) 50/0/50, (b) 50/10/40, (c) 50/20/30, (d) 50/30/20, and (e) 50/ 40/10.

(5)

at 1734 cm-1represents the free carbonyl group, while the signals of the hydrogen-bonded carbonyl groups appear at 1705 and 1715 cm-1, corresponding to bonding to phenolic and phenoxy, respectively. Figure 6 indicates that the carbonyl stretching frequencies split into the three main bands fit well to a Gaussian function. The fraction of hydrogen-bonded carbonyl groups can be calculated by using an appropriate absorptivity ratio (aR

) aHB/aF) 1.5), which has been discussed previously.37

Table 2 summarizes the results from curve fitting; Figure 7 displays the fraction of hydrogen-bonded carbonyl groups in this ternary polymer blend. Clearly, these results indicate that the fraction of hydrogen-bonded carbonyl groups increases upon increasing the relative ratio of phenolic to PCL. In addition, the fraction of PCL that is hydrogen bonded at a high phenolic content in this ternary blend system is lower than that present in the binary phenolic/PCL polymer blend, which indicates that hydrogen bonding exists between the phenoxy and phenolic segments. In other words, the PCL carbonyl units compete with the hy-droxyl groups of phenoxy in forming hydrogen bonds to

the hydroxyl groups of the phenolic resin. Furthermore, the fraction of PCL that is hydrogen bonded at a low phenolic content in this ternary blend system is higher than that observed in the binary polymer blend of phenolic/PCL because hydrogen bonding also exists between the carbonyl groups of the PCL and the hydroxyl groups of phenoxy. Therefore, the complicated intermolecular hydrogen-bonding interactions that exist in this ternary polymer blend system behaves like a network and the phenolic resin is oligomeric, so they do get some help from the entropy of mixing. Even though different hydrogen-bonding strengths exist in each binary blend, the ternary blend remains completely miscible. Again, we conclude here that if intermolecular hydrogen bonding exists between each pair of binary components in a ternary blend, a completely miscible ternary blend may be obtained, as in this phenolic/ phenoxy/PCL system.

Interassociation Equilibrium Constant (KA) be-tween Phenolic and Phenoxy Segments. The origi-nal Painter-Coleman association model37_{is incapable}

of estimating the thermodynamic properties of a blend containing two self-associating polymers. Therefore, in this study we extended the original PCAM to estimate the thermodynamic properties of such a blend. The equilibria can be described as follows:

Table 2. Results of Curve Fitting of the Infrared Spectroscopy Data Recorded at Room Temperature for Phenoxy/ Phenolic/PCL Ternary Blends

free CdO H-bond CdO with phenoxy H-bond CdO with phenolic

phenoxy/phenolic/PCL (wt %) ν, cm-1 W1/2, cm-1 Af, % ν, cm-1 W1/2, cm-1 Af, % ν, cm-1 W1/2, cm-1 Af, % fba(%) 10/18/72 1734 19 62.8 1716 12 2.8 1704 22 34.4 28.3 10/36/54 1734 18 55.0 1715 12 6.6 1703 20 38.4 35.2 10/54/36 1733 18 50.4 1715 13 9.3 1703 20 39.3 39.1 10/72/18 1733 18 34.0 1715 13 10.0 1703 20 56.0 56.4 30/14/56 1733 18 65.6 1715 11 1.7 1708 22 32.7 25.9 30/28/42 1732 19 50.0 1715 11 5.6 1704 20 44.4 40.0 30/42/28 1732 18 32.4 1715 12 20.9 1702 19 47.7 58.5 30/56/14 1732 18 22.8 1715 13 25.0 1702 19 52.2 69.3 50/10/40 1733 18 60.2 1715 13 11.1 1705 20 28.7 30.6 50/20/30 1733 18 53.0 1716 12 12.4 1705 20 34.6 37.2 50/30/20 1732 18 40.9 1716 12 13.5 1704 20 45.6 49.1 50/40/10 1732 17 27.6 1716 13 13.6 1704 20 58.8 63.6 65/7/28 1732 17 63.0 1715 12 7.0 1707 21 30.0 28.1 65/14/21 1732 18 55.8 1715 12 7.3 1706 20 36.9 34.6 65/21/14 1732 17 50.0 1715 12 7.4 1706 20 42.6 40.0 65/28/7 1732 17 48.8 1716 12 8.0 1707 20 43.2 41.1 80/4/16 1732 17 69.5 1716 12 6.2 1707 20 24.3 22.6 80/8/12 1732 17 64.2 1716 12 7.8 1707 20 28.0 27.1 80/12/8 1732 17 61.6 1715 12 8.1 1707 20 30.3 29.3 80/16/4 1732 19 60.2 1716 12 8.6 1707 19 31.2 30.6 10/0/90 1734 21 96.2 1715 28 3.8 2.5 30/0/70 1734 20 86.8 1715 28 13.2 9.2 50/0/50 1734 20 77.6 1715 28 22.4 16.1 65/0/35 1734 20 69.8 1715 29 30.2 22.3 80/0/20 1734 20 62.3 1715 29 37.7 28.7 0/20/80 1734 18 62.1 1708 28 37.9 28.9 0/40/60 1734 20 39.1 1707 28 60.9 50.9 0/60/40 1733 16 23.9 1705 27 76.1 67.9 0/80/20 1733 16 15.8 1704 27 84.2 78.0 a_f

b) fraction of hydrogen-bonded carbonyl groups.

Figure 7. Fraction of hydrogen-bonded carbonyl groups in the ternary phenoxy/phenolic/PCL blend system.

B₁+ B₁798 K2 B₂ (5) Bh+ B1798_K B Bh+1(h g 2) (6) Bh+ Ai798_K A Bh + A_i (7) A1+ A1798_K C2 A2 (8) Ai+ A1798_K C Ai+2(i g 2) (9)

(6)

The terms K2and KBare the self-association constants

for dimer and multimer formation, respectively, of B; KA is the interassociation constant for the interaction

between Aiand Bh; KC2and KCare the self-association

constants for dimer and multimer formation, respec-tively, of A; r ) VA/VBis the ratio of the segmental molar

volumes; ΦBh is the volume fraction of the chains of

length h, and ΦAiis the volume fraction of the chains of

length i at any instant in time.

The stoichiometric relationships are obtained readily from material balance considerations. The total volume fractions of all of the A and B units present in the mixture are given by

Therefore, the total volume fractions of these two self-associating polymers (A and B) can be extended as follows:

These results for ΦAand ΦΒare similar to those for one

self-associating polymer and one nonself-associating polymer, which can be expressed as follows:

In our ternary hydrogen-bonded blend system, the hydrogen bonding between the phenolic hydroxyl group and the PCL carbonyl group and between the phenoxy hydroxyl group and the PCL carbonyl group is repre-sented by

These six equilibrium constants can be expressed as follows in terms of volume fractions:

where

ΦΒ, ΦA, and ΦCare the volume fractions of the repeat

units in the blend, ΦΒ1, ΦΑ1, and ΦC1 are the volume

fractions of isolated units in the blend, and rA) VA/VB,

rB) VB/VA, rC) VC/VB, and rD) VC/VAare the ratios of

the segmental molar volumes. The values of the self-association constants of phenolic (K2 ) 23.3 and KB)

53.3)23_{and the self-association constant of phenoxy (K} C2

) 14.4 and KC ) 25.6)35 and the interassociation

constants for phenolic and PCL (KD ) 116.8)31 and

phenoxy and PCL (KE ) 7)35 have been determined

previously. Equations 23-25 are too complicated to determine the interassociation equilibrium constant (KA) between the hydroxyl groups of phenolic and

phenoxy. For convenience, we take into account the fact that, for the interassociation and self-association in phenolic/PCL and phenoxy/PCL blends, the value of KE

of the phenoxy/PCL blend is considerably smaller than the value of KDof the phenolic/PCL blend and that the

value of KCfor pure phenoxy is also smaller than that

of KBfor pure phenolic. Therefore, we may ignore the

(16) ΦB) ΦB₁

[

(

(

1 (1 - K_CΦA₁) 2

)

][

1 + KAΦB₁ r_B

]

(18) ΦB) ΦB₁

[

(

(

1 (1 - K_BΦB₁)

(29) ΦB) ΦB1Γ2

[

1 + K_AΦA1 rA +KDΦC1 rC

]

(30) ΦA) ΦA1[1 + KAΦB1Γ1] (31) ΦC) ΦC1[1 + KDΦB1Γ1] (32)

(7)

Equations 30-32 are capable of describing both the PVPh/PMA/PEO and phenolic/PEO/PCL blend systems. The value of the interassociation constant KA may be

determined indirectly from a least-squares fitting pro-cedure of the fraction of hydrogen-bonded carbonyl groups obtained experimentally for the present ternary polymer blend. If the equilibrium constants (K2, KB, KC),

segment molar volumes, and fractions of hydrogen-bonded carbonyl groups are known, the value of KAcan

be calculated from eqs 26, 27, and 30-32 by using a least-squares fit based on the fraction of hydrogen-bonded carbonyl group obtained experimentally. To minimize errors in this calculation, we considered the blend containing a relatively low phenoxy content (10 wt %) so that we could ignore the effect of hydrogen bonding between the phenoxy hydroxyl groups and the PCL carbonyl groups. Figure 7 indicates that only 3% of the carbonyl groups are hydrogen bonded in the phenoxy/PCL (10/90) blend. We obtained a value for KA

of 230.0 for the phenolic/phenoxy blend at room tem-perature. The value of KAcalculated from the ternary

blend is higher than that obtained for the model compound. The same trend also has been found in previous studies of PVPh/PVAc/PEO and phenolic/PEO/ PCL blend systems. Table 3 lists all of the parameters required by the Painter-Coleman association model to estimate the thermodynamic properties of this ternary phenolic/phenoxy/PCL blend; the data imply that the interassociation equilibrium constant for hydroxyl-hydroxyl interactions of phenolic/phenoxy is indeed greater than the interassociation equilibrium constants for the hydroxyl-carbonyl interactions of phenolic/PCL and phenoxy/PCL blends and the self-association con-stants of pure phenolic and pure phenoxy at room temperature. Here, we need to emphasize that Painter et al. proposed two self-associating polymers,37_which

also can be handled using approach first developed by Boris Veytsman38 _{recently. We will extend this}

treat-ment for ternary blend systems in the future. Conclusions

We have investigated the phase behavior and hydro-gen bonding present within a ternary blend of phenolic, phenoxy, and PCL by using DSC and FTIR spectroscopic analyses. On the bais of DSC analysis, we observed that this unusual, completely miscible, ternary hydrogen-bonded blend possesses a single glass transition tem-perature over its entire range of compositions. Infrared spectra indicated that intermolecular hydrogen bonding exists within this ternary polymer blend. In addition, we calculated the interassociation equilibrium constant for the interactions between the hydroxyl groups of

phenolic and phenoxy. Even though different intermo-lecular hydrogen-bonding strengths exist in each binary blend, the ternary blend remains completely miscible because the intermolecular hydrogen bonds that exist within the individual binary blends create a networklike structure.

Acknowledgment. This research was financially supported by the National Science Council, Taiwan, Republic of China, under Contract NSC-93-2216-E-009-018. We also thank Professor Paul Painter for the constructive comments, suggestions, and discussions in this paper.

References and Notes

(1) Utracki, L. A. Polymer Alloys and Blends; Hanser Publish-ers: Munich, 1989.

(2) Bonner, J. G.; Hope, P. S. In Polymer Blends and Alloys; Folkes, M. J., Hope, P. S., Eds.; Blackie: Glasgow, UK, 1993; p 46.

(3) Paul, D. R. In Polymers Blends; Paul, D. R., Newman, S., Eds.; Academic Press: New York, 1978; Vol. II, p 35. (4) Scott, R. L. J. Chem. Phys. 1949, 17, 279.

(5) Tompa, H. Trans. Faraday Soc. 1949, 45, 1140.

(6) Kwei, T. K.; Frisch, H. L. Radigan, W.; Vogel, S.

Macromol-ecules 1977, 10, 157.

(7) Pomposo, J. A.; Cortazar, M.; Calahorra, E. Macromolecules

1994, 27, 252.

(8) Fowler, M. E.; Barlow, J. W.; Paul, D. R. Polymer 1987, 28, 1177.

(9) Zhang, H.; Bhagwagar, D. E.; Graf, J. F.; Painter, P. C.; Coleman, M. M. Polymer 1994, 35, 5379.

(10) Patterson, D. Polym. Eng. Sci. 1982, 22, 64.

(11) Hong, B. K.; Kim, J. K. Jo, W. H.; Lee, S. C. Polymer 1997,

38, 4373.

(12) Jo, W. H.; Kwon, Y. K.; Kwon, I. H. Macromolecules 1991,

24, 4708.

(13) Kuo, S. W.; Lin, C. L.; Chang, F. C. Macromolecules 2002,

35, 278.

(14) Min, K. E.; Chiou, J. S.; Barlow, J. W.; Paul, D. R. Polymer

1987, 28, 172.

(15) Guo, Q. P. Eur. Polym. J. 1996, 32, 1409. (16) Guo, Q. P. Eur. Polym. J. 1990, 26, 1329. (17) Goh, S. H.; Ni, X. Polymer 1999, 40, 5733.

(18) Yau, S. N.; Woo, E. M. Macromol. Rapid Commun. 1996, 17, 615.

(19) Woo, E. M.; Tseng, Y. C. Macromol. Chem. Phys. 2000, 201, 1877.

(20) Lee, S. C.; Woo, E. M. J. Polym. Sci., Polym. Phys. 2002, 40, 747.

(21) Wu, H. D.; Ma, C. C. M.; Chu, P. P. Polymer 1997, 38, 5419. (22) Wu, H. D.; Ma, C. C. M.; Chang, F. C. Macromol. Chem. Phys.

2000, 201, 1121.

(23) Wu, H. D.; Chu, P. P.; Ma, C. C. M.; Chang, F. C.

(24) Kuo, S. W.; Wu, H. D.; Lin, C. L.; Chang, F. C. J. Polym.

Res. 2003, 10, 87.

(25) Painter, P. C.; Veytsman, B.; Kumar, S.; Shenoy, S.; Graf, J. F.; Xu, Y.; Coleman, M. M. Macromolecules 1997, 30, 932.

Table 3. Self-Association and Interassociation Equilibrium Constants and Thermodynamic Parameters for Phenolic/ PCL, Phenoxy/PCL, and Phenolic/Phenoxy Blends at 25°Ca

equilib constant enthalpy (kcal/mol)

polymer V Mw δ DP K2 KB h2 hB phenolicb ₈₄ ₁₀₅ _12.05 ₆ _23.3 _52.3 _-4.1 _-6.1 phenoxyc ₂₁₆ ₂₈₄ _10.22 ₈₀ _14.4 _25.6 _-2.5 _-3.4 PCLd ₁₀₇ ₁₁₄ _9.21 ₇₁₄ KA hA phenolic/PCLe _116.8 _-4.6 phenoxy/PCLc _7.0 _-2.0 phenolic/phenoxy 230.0 -5.1b

a_{V ) molar volume (mL/mol), M}

w) molecular weight (g/mol), δ ) solubility parameter (cal/mL)1/2, DP ) degree of polymerization, K2 ) dimer self-association equilibrium constant, KB) multimer self-association equilibrium constant, KA) interassociation equilibrium constant, h2) enthalpy of dimer self-association formation, hB) enthalpy of multimer self-association formation, and hA) enthalpy of interassociation formation.b_{Reference 23.}c_{Reference 35.}d_{Reference 37.}e_{Reference 31.}

(8)

(26) Coleman, M. M.; Pehlert, G. J.; Painter, P. C. Macromolecules

1996, 29, 6820.

(27) Pehlert, G. J.; Painter, P. C.; Veytsman, B.; Coleman, M. M.

Macromolecules 1997, 30, 3671.

(28) Pehlert, G. J.; Painter, P. C.; Coleman, M. M. Macromolecules

1998, 31, 8423.

(29) Coleman, M. M.; Guigley, K. S.; Painter, P. C. Macromol.

Chem. Phys. 1999, 200, 1167.

(30) Painter, P. C.; Coleman, M. M. Polymer Blends; Paul, D. R., Ed.; John Wiley & Sons: New York, 2000; Vol. 1.

(31) Kuo, S. W.; Chang, F. C. Macromol. Chem. Phys. 2001, 202, 3112.

(32) Kuo, S. W.; Huang, C. F.; Chang, F. C., J. Polym. Sci., Polym.

Phys. Ed. 2001, 39, 1348.

(33) Kwei, T. J. Polym. Sci., Polym. Lett. Ed. 1984, 22, 307. (34) Fox, T. G. J. Appl. Bull. Am. Phys. Soc. 1956, 1, 123. (35) Iriarte, M.; Alberdi, M.; Shenoy, S. L.; Iruin, J. J.

(36) Moskala, E. J.; Varnell, D. F.; Coleman, M. M. Polymer 1985,

26, 228.

(37) Coleman, M. M.; Graf, J. F.; Painter, P. C. Specific

Interac-tions and the Miscibility of Polymer Blends; Technomic

Publishing: Lancaster, PA, 1991.

(38) Park, Y.; Veytsman, B.; Coleman, M.; Painter, P. C.

Macro-molecules, in press.

(39) Veytsman, B. J. Phys. Chem. 1993, 97, 7144. MA047371A