Hydrogen bonding interactions and miscibility between phenolic resin and octa(acetoxystyryl) polyhedral oligomeric silsesquioxane (AS-POSS) nanocomposites

Phenolic Resin and Octa(acetoxystyryl) Polyhedral

Oligomeric Silsesquioxane (AS-POSS) Nanocomposites

Department of Environmental Science and Engineering, National Ping-Tung University of Science and Technology, Ping-Tun, Taiwan, Republic of China

Received 15 June 2005; accepted 20 November 2005 DOI: 10.1002/polb.20731

Published online in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: We have synthesized a polyhedral oligomeric silisesquioxane (POSS) derivative containing eight acetoxystyryl functional groups [octa(acetoxystyryl)octasil-sesquioxane (AS-POSS)] and then blended it with phenolic resin to form nanocompo-sites stabilized through hydrogen bonding interactions between the phenolic resin’s hydroxyl group and the AS-POSS derivative’s carbonyl and siloxane groups. One-and two-dimensional infrared spectroscopy analyses provided positive evidence for these types of hydrogen bonding interactions. In addition, we calculated the interas-sociation equilibrium constant, based on the Painter–Coleman asinteras-sociation model (PCAM), between phenolic resin and POSS indirectly from the fraction of hydrogen-bonded carbonyl groups; quantitative analyses indicate that the hydroxyl–siloxane interassociation from the PCAM is entirely consistent with the classical Coggesthall and Saier (C and S) methodology. From a thermal analysis, we observed that the mis-cibility between phenolic and AS-POSS occurs at a relatively low AS-POSS content, which characterizes this mixture as a polymer nanocomposite system. VVC2006 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 44: 673–686, 2006

Keywords: 2D-FTIR; hydrogen bonding; nanocomposite; POSS

INTRODUCTION

Incorporating nanoparticles into polymer matri-ces to enhance their properties has attracted enormous interest in recent years because of the potential to create candidate materials that bridge the gap between polymers and nanopar-ticles. In particular, the use of polyhedral oligo-meric silsesquioxanes (POSS) derivatives is one suitable approach because they embody an inor-ganic–organic hybrid architecture comprising a

well-defined inorganic framework composed of silicon and oxygen (SiO1.5)xand organic substitu-ents containing nonreactive or reactive function-alities. By designing the functionality of the organic substituents, it is possible to create octa-functional or monoocta-functional macromonomers to fit a desired application. Therefore, these nano-structured compounds can be incorporated readily through copolymerization into such polymers as polysiloxane,1poly(methyl methylacrylate),2,3poly (styrene),4 epoxy,5 polyurethane,6 polyimide,7and polynorborbornene.8

The physical properties of polymer/POSS nano-composites are strongly influenced by the misci-bility between the host polymer and the POSS moiety. Random copolymerization of the organic Correspondence to: S.-W. Kuo (E-mail: [email protected].

edu.tw)

Journal of Polymer Science: Part B: Polymer Physics, Vol. 44, 673–686 (2006)

V

VC2006 Wiley Periodicals, Inc.

functional groups of the POSS derivative is one approach to improving the miscibility. Therefore, in a previous study,9 we synthesized a series of poly(vinylphenol-co-vinylpyrrolidone-co-POSS) (PVPh-co-PVP-co-POSS) copolymers that exhibit a significant glass-transition temperature in-crease relative to the corresponding nonPOSS PVPh-co-PVP copolymers because of the strong hydrogen bonds that exist between the PVPh and POSS units. The synthesis of a random copolymer is generally more complicated and time-consum-ing than is prepartime-consum-ing a physical blend; thus, poly-mer blending is seen as a more convenient method of preparing polymer/POSS nanocomposites. Be-cause the combined entropy contribution to the free energy of mixing two polymers is negligi-bly small, specific intermolecular interactions are generally required to enhance the miscibility of polymer blends. To improve the properties and miscibility of hybrid materials, it is necessary to ensure that favorable, specific interactions exist between these components, such as hydrogen bonding,10 dipole–dipole, and acid–base interac-tions.

In a previous study,11 we demonstrated that simple blending of POSS derivatives containing nonreactive or inert diluent functional groups (in that case, octaisobutyl-POSS) with phenolic resin provide unsatisfactory results because of poor miscibility. The interassociation equilibrium con-stant between the phenolic hydroxyl group and the octaisobutyl-POSS siloxane group (38.6) is lower than the self-association equilibrium con-stant of pure phenolic (52.3), based on the Painter– Coleman association model10 (PCAM). This result indicates that this POSS derivative tends to be only partially miscible with phenolic in the phe-nolic/POSS hybrid because of the poor degree of interaction between the phenolic resin and the octaisobutyl-POSS. Functionalization of POSS such that it displays pendent hydrogen-bond-acceptor groups is expected to improve the misci-bility with phenolic resin. Functionalization of Q8M8H [HSiMe2OSiO1.5]8 can be achieved by hydrosilylation of its Si

H groups onto acetoxys-tyrene in the presence of a platinum catalyst to form AS-POSS. Previously, we used infrared and solid-state NMR spectroscopy to thoroughly inves-tigate the hydrogen bonding interactions between the carbonyl groups of poly(acetoxystyrene) (PAS) and the hydroxyl groups of phenolic resin.12 We found that the interassociation equilibrium con-stant for the phenolic/PAS blend (64.6) is higher than the self-association equilibrium constant of

pure phenolic (52.3), which implies that the ten-dency for hydrogen bonding between the phenolic resin and PAS dominates over the self-association (intramolecular hydrogen bonding) of the phenolic resin in the mixture.

Observing the carbonyl, hydroxyl, and siloxane vibrations by infrared (FTIR) spectroscopy is an excellent tool for detecting intermolecular poly-meric interactions.13This tool can be used to study the mechanism—both qualitatively and quantita-tively—of interpolymer miscibility through the formation of different types of hydrogen bonds. In addition, the generalized two-dimensional (2D) IR correlation spectroscopy 14–19 has been applied widely in polymer science in recent years. In 2D IR, a spectrum is obtained as a function of two independent wavenumber axes, and peaks located on the spectral plane. This novel method monitors spectral fluctuations as a function of time, temper-ature, pressure, and composition and allows the specific interactions that exist between polymer chains to be identified. 2D IR correlation spectro-scopy can identify different intra- and intermolecu-lar interacting sites through the monitoring of selected bands from the one-dimensional vibration spectrum. In this study, we used generalized 2D IR correlation spectroscopy to explore the hydrogen bonding interactions present in blends of AS-POSS and phenolic resin.

The interassociation equilibrium constant be-tween the phenolic hydroxyl groups and the POSS siloxane groups cannot be quantified directly through IR spectroscopic analysis in this binary blend because no carbonyl groups are available to measure the fraction of groups that are hydrogen bonded. The siloxane stretching mode near 1100–1200 cm–1, which presents sig-nals for both the free and hydrogen-bonded silox-ane absorptions, is a highly coupled mode that is conformationally sensitive but cannot be decom-posed readily into two peaks. In a previous study,11we calculated the interassociation equili-brium constant (KA) between the hydroxyl groups of phenolic resin and the siloxane group of octai-sobutyl-POSS by using the classical Coggesthall and Saier (C and S)20methodology. The interasso-ciation equilibrium constant obtained from low molecular weight compounds, however, is not exactly the same as that calculated for a true poly-mer blend because intramolecular screening and functional group accessibility affect the miscibility of a polymer blend.21 Fortunately, in our previous study,11 octaisobutyl-POSS was the low-molecular-weight compound and, thus, this calculation can be

considered to provide the true interassociation equi-librium constant for the interactions between the phenolic hydroxyl groups and the POSS siloxane groups. To recheck the interassociation equilibrium constant between the phenolic hydroxyl groups and the POSS siloxane groups in this present study, we determined the value of KAindirectly from a least-squares fitting procedure of the experimental frac-tion of hydrogen-bonded carbonyl groups of AS-POSS in this binary blend. We have found a good correlation between these two methods for deter-mining the interassociation equilibrium constant of the hydroxyl–siloxane interactions.

EXPERIMENTAL

Materials

The phenolic used in this study was synthesized through a condensation reaction with sulfuric acid to give average weights of Mn¼ 500 and Mw ¼ 1200. Q8M8Hwas purchased from Hybrid

Plas-tics Co. Platinum divinyltetramethyldisiloxane complex, Pt(dvs), and acetoxystyrene were ob-tained from Aldrich Chemical Co. Inc. Octa(ace-toxystyryl)octasilsesquioxane (AS-POSS) was syn-thesized according to the following method. Synthesis of Octa(acetoxystyryl)octasilsesquioxane (AS-POSS)

Q8M8H (2.00 g, 1.96 mmol) was dissolved in tol-uene (20 mL) in a 100-mL Schlenk flask, equipped with a reflux condenser and a magnetic stirrer, and then 4-acetoxystyrene (3.2 g, 19.6 mmol) was added. Pt(dvs) (2 mM solution, 0.2 mL) was added through a syringe. The reaction mixture was heated to 808C under nitrogen. The reaction was complete within 4 h. The toluene was evaporated under reduced pressure, and the residue was dried in a vacuum oven at 808C for 24 h to give AS-POSS (3.17 g, 93%). AS-POSS is a colorless, viscous liquid that is soluble in THF, CHCl3, and acetone. The chemical structure and scheme for the synthesis of AS-POSS are shown here:

Blend Preparation

Blends of phenolic/AS-POSS of various composi-tions were prepared by solution blending. A THF solution containing 5 wt % of the mixture was stirred for 6–8 h and then the solvent was evapo-rated slowly at room temperature over 1 day. To ensure total elimination of solvent, the powder of the blend obtained was dried in a vacuum oven at 608C for 2 days.

Characterization

Nuclear Magnetic Resonance (1H NMR) Spectroscopy

1

H NMR spectra were recorded on a Varian Unity Inova 500 FT NMR spectrometer operat-ing at 500 MHz, usoperat-ing CDCl3 as the solvent; chemical shifts are reported in parts per million (ppm).

Fourier Transform Infrared (FTIR) Spectroscopy Infrared spectroscopic measurements were re-corded on a Nicolet Avatar 320 FTIR spectropho-tometer; 32 scans were collected with a spectral res-olution of 1 cm–1. Infrared spectra of polymer blend films were obtained through the conventional NaCl disk method. All sample preparations were per-formed under a continuous flow of nitrogen to mini-mize sample oxidation or degradation. Samples were prepared by casting a THF solution directly onto a NaCl disk, which was dried under conditions similar to those used in the bulk preparation.

2D IR correlation analysis was conducted using Vector 3D software (Bruker Instrument Co.). All of the spectra were normalized before being subjected to 2D correlation analyses. All of the spectra subjected to the 2D correlation analy-ses were normalized and classified into two sets: A and B. The spectra in set A are, in order, pure phenolic, phenolic/AS-POSS¼ 5/95,

phenolic/AS-POSS ¼ 10/90, and phenoic/AS-POSS ¼ 20/80.

Those in set B are, in order, phenolic/AS-POSS ¼ 30/70, POSS ¼ 40/60, phenolic/AS-POSS¼ 60/40, and pure AS-POSS. Shaded areas indicate regions of negative intensity of auto-peaks or crossauto-peaks in the 2D correlation spec-trum; unshaded areas indicate positive-intensity regions. Synchronous 2D spectra were used to study the specific interactions between phenolic and AS-POSS in the blends.

Differential Scanning Calorimetry (DSC)

Thermal analysis was performed using a DuPont DSC-9000 differential scanning calorimeter at a scan rate of 208C/min over a temperature range from
50 to 150 8C. Temperature and energy cali-brations were undertaken using indium. Approxi-mately 5–10 mg of each blend was weighed and sealed in an aluminum pan. This sample was quickly cooled to
50 8C from the melt of the first scan and then it was scanned between
50 and 1508C at 20 8C/min. The glass-transition temper-ature was obtained at the midpoint of the specific heat increment.

RESULTS AND DISCUSSION

AS-POSS Analyses

Figure 1(a) presents the FTIR spectra of Q8M8 H

, acetoxystyrene, and the final hydrosilylation prod-uct, AS-POSS. The strong absorption peak at 1100

cm–1for both Q8M8Hand AS-POSS represents the vibrations of the siloxane Si

O

Si groups and is a general feature of POSS derivatives. The

charac-teristic stretching vibrations of the vinyl

(ArCH¼¼CH2) and Si

H groups appear as peaks at 1650 and 2200 cm–1, respectively. In AS-POSS, these peaks have disappeared completely, indicat-ing that complete reaction was achieved. Further-more, the peak for the carbonyl groups of the ace-toxystyrene (1765 cm–1) remained in AS-POSS, which provides evidence for the successful attach-ment of the acetoxystyrene units to the POSS core. Figure 1(b) displays the corresponding 1H NMR spectra of AS-POSS. Clearly, the peaks for the vinyl (ca. 5.8 ppm) and Si

H protons (4.7 ppm) have dis-appeared in the spectrum of AC-POSS, which sup-ports the complete reaction. The spectrum in Fig-ure 1(b) indicates that the vinyl groups of acetoxys-tyrene underwent hydrosilylation of the Si

H bonds of Q8M8H in both a and b configurations, which is a mixture of these two orientation exists. From the integrated areas of the two types of methyl groups attached to the Si atoms, we esti-mate that a-carbon atom attachment was three times as prevalent asb-carbon atom attachment. 1D IR Spectral Analyses of Phenolic/AS-POSS Nanocomposites

In a previous study,11 we used 1D and 2D FTIR spectra to discuss in detail the hydrogen bonding interactions that exist between the phenolic hydroxyl groups and the POSS siloxane groups. In addition, we have also studied22the weak spe-cific hydrogen bonding interactions between the carbonyl groups of PAS and the methylene units of PEO. In this study, we synthesized an acetox-ystyrene (AS)-grafted POSS (AS-POSS) to inves-tigate the specific interactions between phenolic and AS-POSS. Figure 2 displays infrared spectra of phenolic/AS-POSS blends in different composi-tions and Table 1 lists detailed peak assignments for phenolic and AS-POSS.

Figure 3(a) presents scaled infrared spectra (2700–4000 cm–1) recorded at room temperature from pure phenolic and various phenolic/AS-POSS nanocomposites. The pure phenolic poly-mer exhibits two bands in the hydroxyl stretching region of the infrared spectrum. We attributed the very broadband centered at 3350 cm–1_{to the} wide distribution of hydrogen-bonded hydroxyl groups, while a narrower shoulder band at 3525 cm–1 represents the free hydroxyl groups. Figure 3(a) indicates clearly that the intensity of the free

hydroxyl absorption (3525 cm–1) decreases gradu-ally as the AS-POSS content of the blend increases from 5 to 90 wt %. The band for the hydrogen-bonded hydroxyl units in the phenolic tends to shift to higher frequency (toward 3465 cm–1) upon increasing the AS-POSS content. This

change is due to the switch from hydroxyl–hydro-xyl interactions to the formation of hydrohydroxyl–hydro-xyl–car- hydroxyl–car-bonyl and/or hydroxyl–siloxane hydrogen bonds.

Figure 3(b) displays the infrared spectra (1680–1820 cm–1) measured at room temperature for various phenolic/AS-POSS blend

composi-Figure 1. (a) IR spectra recorded at room temperature for Q8M8 H

, acetoxystyrene, and AS-POSS and (b) the1H NMR spectrum of AS-POSS.

Figure 2. Infrared spectra recorded at room temperature for phenolic/AS-POSS blends of various compositions.

Table 1. Frequency and Peak Assignments of the FTIR Spectral Bands of Phenolic and AS-POSS

Phenolic (cm
1) AS-POSS (cm
1) Assignments 3525 Free OH-stretching 3350 Hydrogen-bonded OH-stretching

3037 Benzene ring CH stretching

3016 Benzene ring CH stretching

2957 CH2stretching

2916 CH2stretching

2841 CH2stretching

1763 Free C¼¼O stretching

1611,1595 C

C stretching of ring in plane

1603 C

C stretching of ring in plane

1510 Phenyl

OH stretching

1505 Phenyl

O

streching

1478,1439 Asymmetric CH2stretching

1455,1421 Asymmetric CH2stretching 1368 Symmetric carboxylate stretching

1357 COH bending

1223 phenyl

OH stretching

1230 Si

CH2stretching

1215 Acetate stretching

1193 Acetate asymmetric stretching

1102 C

O stretching

1087 Si

O

Si Stretching

1015 C

O stretching

911 CH2out of line bending 846 CH2out of line bending

tions. The carbonyl stretching frequency is split into two bands at 1763 and 1735 cm–1, which cor-responds to the free and hydrogen-bonded car-bonyl groups, respectively. The band can be read-ily decomposed into two Gaussian peaks that cor-responds to the areas of the hydrogen-bonded carbonyl (1735 cm–1) and free carbonyl (1763 cm–1) peaks, as indicated in Figure 4. To obtain the fraction of the hydrogen-bonded carbonyl units, we required the absorptivity ratio for the contri-butions of the hydrogen-bonded and free carbonyl units. We have employed a value foraHB/aFof 1.5, as calculated previously by Coleman et al.10Table 2 summarizes the fractions of hydrogen-bonded carbonyl groups obtained through curve fitting. As expected, this value rises upon increasing the phenolic content.

2D IR Spectral Analyses of Phenolic/AS-POSS Nanocomposites

In the studies of the 1D-IR spectra earlier, we found that the signal of the hydroxyl groups of phenolic shifts dramatically higher (or lower) upon adding AS-POSS, and we propose that this change is due to the switch from hydroxyl– hydroxyl interactions to hydroxyl–carbonyl and/ or hydroxyl–siloxane hydrogen bonds. In this

sec-tion, we describe how we used 2D-IR spectra to investigate these types of hydrogen bonds.

Figure 5 displays synchronous 2D correlation maps (900 to 1300 cm–1) of blends and indicates that strong auto and cross peaks exist at

wave-numbers between 1100 and 1250 cm–1. The

absorption bands of the POSS derivative that appear in the spectral range from 1250 to 1000 cm–1_{are those at 1100 and 1230 cm}–1_{, which}

cor-respond to siloxane Si

O

Si and Si

CH2

stretching vibrations, respectively; for phenolic, only one appears (at 1223 cm–1_{; it is due to} phe-nyl

OH stretching vibrations) based on Table 1. The two positive cross peaks in Figure 2 indicate that the hydrogen bonding interactions occur between the siloxane groups of AS-POSS (1100 cm–1) and the phenyl

OH groups of phenolic (1223 cm–1), which are similar to those observed for phenolic/isobutyl-POSS blends.11

Figure 6 also displays the synchronous 2D cor-relation maps (1500–1800 cm–1) of blends. The three strong auto peaks and two pairs of cross peaks near the 1510, 1750, and 1600 cm–1regions correspond to vibrations of the phenol

OH, car-bonyl, and aromatic groups, respectively. In the 1750 cm–1 region, we observe two auto peaks clearly at 1735 and 1763 cm–1; they correspond to the free carbonyl groups of AS-POSS and those

Figure 3. Infrared spectra recorded at room temperature for phenolic/AS-POSS blends and displaying the (a) hydroxyl stretching and (b) carbonyl stretching regions.

that are hydrogen bonded to the phenolic hy-droxyl groups, respectively. In the 1600 cm–1 region, we observe two auto-peaks at 1590 and 1610 cm–1 that correspond to the phenyl groups in both phenolic and AS-POSS. Figure 6 indicates clearly that hydrogen bonding interactions do indeed exist between the carbonyl groups of AS-POSS (1763 cm–1) and the phenyl

OH groups of the phenolic resin (1510 cm–1).

Hydroxyl–Siloxane Inter-Association Equilibrium Constant Determined through PCAM Analysis According to the PCAM, the interassociation equilibrium constant between a noncarbonyl group component and a hydrogen bond-donating component can be calculated using the classical Coggesthall and Saier method.20 To recheck the interassociation equilibrium constant between the phenolic hydroxyl groups and the POSS

silox-Figure 4. Deconstructed models of the carbonyl stretching bands [in Fig. 3(b)] with respect to the weight percentages of the phenolic/AS-POSS blends at various compositions.

ane groups in this present study, we determined the value of KAindirectly from a least-squares fit-ting procedure of the experimental fraction of hydrogen-bonded carbonyl groups of AS-POSS in this binary blend. Figure 7 displays plots of the experimental data and theoretically predicted

curves as a function of the composition at room

temperature. The results demonstrate that

PCAM has the ability to predict the degree of hydrogen bonding on the carbonyl group. Figure 7 indicates that the experimental values are gen-erally lower than the predicted values when

Figure 5. The synchronous 2D correlation map of set A in the 900–1300 cm–1 region. (The y-axis is wavenumber (cm
1)). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Table 2. Curve Fitting of the Area Fractions of the Carbonyl Stretching Bands in the FTIR Spectra of Phenolic/AS-Poss Blends Recorded at Room Temperature

Phenolic/ AS-POSS (Wt Ratio)

Free C¼¼O H-Bonding C¼¼O

fb a t (cm
1_{) W} 1/2(cm
1) Af(%) t (cm
1) W1/2(cm
1) Ab(%) 0/100 1763 23 100 – – – – 10/90 1765 18 52.6 1739 30 47.4 0.375 20/80 1765 17 41.5 1737 28 58.5 0.484 30/70 1764 17 36.1 1736 28 63.9 0.541 40/60 1764 16 32.6 1734 26 67.4 0.580 50/50 1763 17 29.9 1734 26 70.1 0.610 70/30 1763 17 25.6 1732 26 74.4 0.660 80/20 1762 18 23.3 1731 25 76.7 0.688 90/10 1762 18 22.1 1731 25 77.9 0.701 a

fb, fraction of hydrogen bonding interaction.

using the value of KA of 64.6 obtained from phe-nolic/PAS blends. This result also indicates that the hydroxyl groups of phenolic not only interact with the carbonyl groups of the acetoxystyrene units but also with the siloxane groups of the POSS core, which is consistent with our results from a previous study.11 In other words, the AS carbonyl groups compete with the siloxane groups of the POSS core in forming hydrogen bonds with the hydroxyl groups of the phenolic resin. We employed a numerical method to deter-mine the value of KA of the phenolic/AS-POSS blend, according to the PCAM, based on the frac-tion of hydrogen-bonded carbonyl groups. The approximate equations10_{are as follows:}

B¼ B1 2 1þKAA1 rA
 ð1Þ A¼ A1½1þ KAB1 1 ð2Þ where 1¼ 8 > > :1
K2 KB 9 > > ; þK2 KB 8 > > :_ð1
K1 BB1Þ 9 > > ; ð3Þ 2¼ 8 > > :1
K2 KB 9 > > ; þK2 KB 8 > > : 1 ð1
KBB1Þ2 9 > > ; ð4Þ

and/Aand/Bdenote the volume fractions of the nonself-associated species A (AS-POSS) and the self-associating species B (phenolic), respectively; /A1 and /B1 are the corresponding volume frac-tions of the isolated AS-POSS and phenolic seg-ments, respectively; r is the ratio of molar vol-ume, VA/VB. The self-association equilibrium con-stants, KB and K2, describe the formation of multimers and dimers, respectively. Finally, KAis the equilibrium constant describing the associa-tion of A with B. The values of KBand K2of pure phenolic at 258C are 23.3 and 52.3, respectively.11 To calculate the values of the interassociation constants KA, we used a least-squares method that we had described previously.12 Table 3 lists all of the parameters required by the Painter–

Figure 6. The synchronous 2D correlation map of set A in the 1500–1800 cm–1 region. (The y-axis is wavenumber (cm
1)). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Coleman association model to estimate the ther-modynamic properties for this phenolic/AS-POSS blend. We obtained an interassociation equili-brium constant of 26.0 for the phenolic/AS-POSS blend. It is difficult to exactly separate the seg-ment of AS and POSS because of the chemical bond between these two segments, and so we assume that the molar volume of AS and POSS segment is the same (ca. 1000 mL/mol). The value of KA, however, for the phenolic/PAS blend is 64.6, which implies that the value of KAbetween the hydroxyl group of phenolic and the siloxane group of POSS is equal to 38.6 (i.e. 64.6
26.0 ¼ 38.6), which is entirely consistent with the value reported previously based on classical Cog-geshall and Saier (C and S) methodology. There-fore, there is a good correlation between these two different methods when determining the val-ues of the interassociation equilibrium constants for hydroxyl–siloxane interactions.

Thermal Analyses

In general, DSC analysis is one of the most con-venient methods for determining the miscibility of blend systems. DSC can determine whether one or two values exist for Tg: a single value of Tg is the most conventionally used criterion for

establishing the miscibility of polymer blends; an immiscible polymer blend exhibits more than one value of Tg. A single compositionally dependent glass transition indicates full miscibility with dimensions of the order of 20–40 nm. The misci-bilities of most polymer/nanoparticle blend sys-tems have not been studied using DSC analyses, however, because most of these nanoparticles

Figure 7. Fraction of hydrogen-bonded carbonyl groups plotted with respect to the composition of the blend: (n) FT-IR spectroscopic data, (–) theoretical values from phenolic/PAS blends (KA ¼ 64.6), and (
) theoretical values from phenolic/AS-POSS blends (KA¼ 26.0) calculated at 25 8C.

Table 3. Self- and Inter-Association Equilibrium Constants and Other Thermodynamic Parameters of Phenolic/AS-Poss Blends at 258C Polymer V Mw Equilibrium Constant K2 KB KA Phenolica _{84 105 23.3 52.3} PASb 128.6 162.2 64.6 8-Isobutyl POSSc 778.6 872.2 38.6 AS-POSS 2058.6 2314.6 26.0 a_reference11_. b_reference12_. c reference11.

V, molar volume (ml/mol); Mw, molecular weight (g/mol);

K2, dimmer self-association equilibrium constant; KB,

multi-mer self-association equilibrium constant; KA,

have been inorganic materials that do not display a glass transition at relatively lower tempera-tures (<400 8C). Fortunately, we obtained conven-tional second-run DSC thermograms of the phe-nolic/AS-POSS blends at various compositions, as presented in Figure 8. The glass-transition tem-peratures of the pure components used in this study, phenolic and AS-POSS, are 53 and
15 8C, respectively. Interestingly, even though the molar masses of pure phenolic and AS-POSS are only 1200 and 2305 g/mol, they display single-Tg behavior at 52.5 and
15 8C, respectively, during the second heating scan. We must emphasize that phenolic resin contains a high density of hydroxyl groups and that hydrogen bonding serves as a physical crosslink to increase its glass-transition temperature. In addition, phenolic resin pos-sesses a higher value of Tgthan do other materi-als that have similar molecular weights because of its high density of hydrogen bonds. Pure AS-POSS can be considered as an oligomer of the siloxane and acetoxystyrene. We also rechecked the thermal analysis of Q8M8

H

by DSC. We observed no glass transition, which indicates that the glass-transition temperature of AS-POSS is a result of the presence of the acetoxystyrene seg-ments.

We observed clearly that a single value of Tg existed for those blends having relatively low AS-POSS contents (<40 wt %) and that the value of Tg decreased upon increasing the AS content. In addition, the Tg breadth also increased upon

increasing the AS content (from 12 8C for pure phenolic to 348C for phenolic/AS-POSS ¼ 60/40), which implies that the homogeneity decreases at the molecular scale of the blend system upon an increase in the POSS content. At a higher AS-POSS content (60 wt %), however, we found two values of Tg for those phenolic/AS-POSS blends, indicating their immiscibility. According to the infrared spectral analyses, the fraction of hydro-gen-bonded carbonyl groups decreased upon increasing the AS-POSS content; thus, phase sep-aration may occur at higher AS-POSS content because the lower fraction of hydrogen bonding interactions does not overcome the strong ability of AS-POSS nanoparticles to aggregate. This result is entirely likely because we found in this study that the interassociation equilibrium con-stant (KA¼ 64.6) for the interaction between the carbonyl groups of AS-POSS and the hydroxyl groups of phenolic is not significantly higher than the self-association equilibrium constant of pure phenolic (KB¼ 52.3). Therefore, at a higher AS-POSS content, the ability of AS-AS-POSS nanopar-ticles to aggregate immiscibly dominates over the benefits of hydrogen bonding. In summary, we have demonstrated that the miscibility behavior in this phenolic/AS-POSS blend is dependent on the extent of the hydrogen bonding interactions, while the immiscible aggregation ability of AS-POSS nanoparticles, which is similar to that of many polymer nanocomposite systems, decreases upon increasing the nanoparticle’s content.23

If we take into account, only the results obtained at relatively low AS-POSS contents (<40 wt %), these compositions display miscibility and variable glass-transition temperatures as a function of the composition of the blend, as indi-cated in Figure 9. A number of equations have been designed to predict the variations in the glass-transition temperatures of miscible blends in relation to their composition. The most widely used equation is the Kwei equation,24which pre-dicts the glass-transition temperature of a misci-ble misci-blend featuring hydrogen bonding interac-tions as a function of its composition:

Tg¼W1Tg1þ kW2Tg2

W1þ kW2 þ qW1W2 ð5Þ

where W1 and W2 denote the weight fractions of the compositions, Tg1and Tg2represent the corre-sponding glass-transition temperatures of the blend’s components, and k and q are fitting con-stants. Figure 9 displays plots of the values of Tg of the blends versus their compositions; clearly, the linear and Fox equations do not fit the

exper-imental data well. The Kwei equation, however, correlates well with the experimental data. On the basis of the nonlinear least-squares best fit of this data, we obtained k ¼ 1 and q ¼ 25; q is a parameter that corresponds to the strength of the hydrogen bonds in the blend and reflects the balance between the breaking of any self-associa-tion and the formaself-associa-tion of the interassociaself-associa-tion hydrogen bonds. Compared with the phenolic/ PAS blend system (k ¼ 1, q ¼
245),12 the phe-nolic/AS-POSS blend system (k ¼ 1, q ¼ 25) seems to have the stronger average hydrogen bonding interactions. This result may arise from two phenomena. One is that the star-shaped ace-toxystyrene-POSS presents a larger fraction of hydrogen-bonded carbonyl groups than does the linear PAS, which is similar to the findings we made in a previous study of the phenolic/poly(-methyl methacrylate) blend system.25 The other reason is that the siloxane groups of the POSS core also take part in hydrogen bonding interac-tions with the hydroxyl groups of the phenolic to result in organic/inorganic polymer nanocompo-sites.26

CONCLUSIONS

We have synthesized a new nanomaterial based on AS-POSS and investigated its hydrogen bond-ing with phenolic by usbond-ing 1D and 2D FTIR spec-troscopic analyses. Hydrogen bonds exist between the phenolic hydroxyl groups and both the car-bonyl and siloxane groups of AS-POSS. We deter-mined the interassociation equilibrium constant between the hydroxyl groups of the phenolic and the siloxane groups of the POSS indirectly from a least-squares fitting procedure based on the experimental fraction of hydrogen-bonded car-bonyl groups in this blend system. The value of KA (38.6) we obtained for the hydroxyl–siloxane interactions is equal to the value determined using the classical Coggeshall and Saier (C and S) methodology.

We thank the National Science Council, Taiwan, Republic of China, for supporting this research finan-cially under Contract No. NSC-93-2216-E-009-021.

REFERENCES AND NOTES

1. Schwab, J. J.; Lichtenhan, J. D. Appl Organomet Chem 1998, 12, 707.

Figure 9. Plots of Tg versus composition based on the experimental data and the linear, Fox, and Kwei equations.

2. Lichtenhan, J. D.; Otonari, Y.; Carri, M. G. Mac-romolecules 1995, 28, 8435.

3. Shochey, E. G.; Bolf, A. G.; Jones, P. F.; Schwab, J. J.; Chaffee, K. P.; Haddad, T. S.; Lichtenhan, J. D. Appl Organomet Chem 1999, 13, 311. 4. Haddad, T. S.; Lichtenhan, J. D. Macromolecules

1996, 29, 7302.

5. Abad, M. J.; Barral, L.; Fasce, D. P.; Williams, R. J. J. Macromolecules 2003, 36, 3128.

6. Fu, B. X.; Zhang, W.; Hsiao, B. S.; Johansson, G.; Sauer, B. B.; Phillips, S.; Balnski, R.; Rafailovich, M.; Sokolov, J. Polym Prepr 2000, 41, 587.

7. Leu, C. M.; Chang, Y. T.; Wei, K. H. Macromole-cules 2003, 36, 9122.

8. Mather, P. T.; Jeon, H. G.; Romo-Uribe, A.; Had-dad, T. S.; Lichtenhan, J. D. Macromolecules 1999, 32, 1194.

9. Xu, H.; Kuo, S. W.; Lee, J. S.; Chang, F. C. Poly-mer 2002, 43, 5117.

10. Coleman, M. M.; Graf, J. F.; Painter, P. C. Specific Interactions and the Miscibility of Polymer Blends; Technomic: Lancaster, PA, 1991.

11. Lee, Y. J.; Kuo, S. W.; Huang, W. J.; Lee, H. Y.; Chang, F. C. J Polym Sci Part B: Polym Phys 2004, 42, 1127.

12. Kuo, S. W.; Chang, F. C. Macromol Chem Phys 2002, 203, 868.

13. Kuo, S. W.; Chang, F. C. Macromolecules 2001, 34, 4089.

14. Noda, I. J Am Chem Soc 1989, 111, 8116.

15. Ren, Y.; Murakami, T.; Nishioka, T.; Nakashima, K.; Noda, I.; Ozaki, Y. Macromolecules 1999, 32, 6307.

16. Makashima, K.; Ren, Y.; Nishioka, T.; Tsubahara, N.; Noda, I.; Ozaki, Y. J Phys Chem B 1999, 103, 6704.

17. Haung, H.; Malkov, S.; Coleman, M.; Painter, P. Macromolecules 2003, 36, 8148.

18. Haung, H.; Malkov, S.; Coleman, M.; Painter, P. Macromolecules 2003, 36, 8156.

19. Shen, Y.; Wu, P. J Phys Chem B 2003, 107, 4224. 20. Coggesthall, N. D.; Saier, E. L. J Am Chem Soc

1951, 71, 5414.

21. Coleman, M. M.; Painter, P. C. Prog Polym Sci 1995, 20, 1.

22. Kuo, S. W.; Huang, W. J.; Chan, S. C.; Huang, C. F.; Chang, F. C. Macromolecules 2004, 37, 4164. 23. Ginzburg, V. V. Macromolecules 2005, 38, 2362. 24. Kwei, T. K. J Polym Sci Polym Lett Ed 1984, 22, 307. 25. Haung, C. F.; Kuo, S. W.; Lin, H. C.; Chen, J. K.;

Chen, Y. K.; Xu, H.; Chang, F. C. Polymer 2004, 45, 5913.

26. Xu, H.; Kuo, S. W.; Lee, J. S.; Chang, F. C. Macro-molecules 2002, 35, 8788.