Surface Modification of Poly(ether imide) by Low-Energy Ion-Beam

Yoonkwan Kim , Seongeun Park , and Yongsok Seo ... Sehyun Kim , Jiseok Lee , Hoyun Kim , Youngwook P. Seo , Soon Man Hong , Atsushi Takahara , Hyoung ...
0 downloads 0 Views 334KB Size
Download PDF
Langmuir 2002, 18, 6185-6192

6185

Surface Modification of Poly(ether imide) by Low-Energy Ion-Beam Irradiation and Its Effect on the Polymer Blend Interface Sehyun Kim,†,‡ Ki-Jun Lee,‡ and Yongsok Seo*,† Supercomputational Modeling and Simulation Laboratory, Future Technology Research Division, Korea Institute of Science and Technology (KIST), P.O. Box 131, Cheongryang, Seoul 130-650, Korea, and School of Chemical Engineering, Seoul National University, Shinlimdong 56-1, Kwanakku, Seoul 151-742, Korea Received January 10, 2002. In Final Form: April 23, 2002 A low-energy Ar+ ion beam was used to modify the surface of a poly(ether imide) (PEI) powder. The modification was promoted by the oxygen gas injected during the irradiation. The surface functional groups of the modified PEI were identified with X-ray photoelectron spectroscopy (XPS). The XPS results show that the bonds originally contained in the imide ring structure are destroyed as a result of random collisions with incident ions and energy transfer to the PEI atoms. Ion-beam irradiated PEI (IBPEI) blends with a thermotropic liquid crystalline polymer (TLCP) show more homogeneous rheological behavior than pristine PEI blends, which is attributed to an interaction between the IBPEI and the TLCP phase. Ionbeam irradiation has a strong effect on the blend morphology, the size of the dispersed phase being remarkably reduced (from 5.3 to 3 µm) and the TLCP phase being finely dispersed. Also, the adhesion between the TLCP phase and the IBPEI is definitely improved. The theoretically calculated interfacial tension of the TLCP/IBPEI blend is smaller than that of the TLCP/PEI blend, which indicates a greater interaction between the IBPEI and the TLCP phase. TLCP/IBPEI blends show more orientation (fibrillation) of the TLCP phase and better adhesion at the interface than TLCP/PEI blends do. The mechanical properties of the TLCP/IBPEI blends are improved due to the structures being more deformed and to the strong adhesion at the interface, which are correlated to the interaction between the IBPEI and the TLCP phase.

1. Introduction Compatibilization, the modification of normally immiscible blends to give alloys with improved end-use performance, is an important factor in almost all commercial blends and has been the subject of numerous experimental investigations, for which many of the results remain proprietary.1 Thermodynamically, a miscible blend means a system that has a negative interaction to reduce the system’s free energy. Thus, compatibilization is an operation to impose some kind of interaction between two immiscible phases. Many different methods exist for compatibilization. One of them is the attachment of functional groups, which can interact with the functional groups of other polymers, to a polymer backbone.2 This can be achieved by modifying the polymers. Various modification methods have been employed, including wet chemical methods and vacuum techniques.3-6 Since vacuum techniques require no solvent and are quite efficient, plasma or the gas-phase approaches have been applied to achieve the desired surface func* To whom correspondence should be addressed. Email: [email protected]. † Korea Institute of Science and Technology. ‡ Seoul National University. (1) Utracki, L. A. Commercial Polymer Blends; Chapman & Hall: London, 1998. (2) Al-Malaika, S. Reactive Modifiers for Polymers; Chapman & Hall: London, 1997. (3) Bertoti, I.; Menyhard, M.; Toth, A. In Handbook of Surface and Interface Analysis; Riviere, J. C., Mihra, S., Eds.; Marcel Dekker: New York, 1998; Chapter 8. (4) Tan, B. J.; Fessehaie, M.; Suib, S. L. Langmuir 1993, 9, 740. (5) Valenza, A.; Spadaro, G.; Calderaro, E., Acierno, D. Polym. Eng. Sci. 1993, 33, 845. Spadaro, G.; Acierno, D.; Dispenze, C.; Calderaro, E.; Valenza, A. Radiat. Phys. Chem. 1996, 48, 207. (6) Golub, M. A.; Lopata, E. S.; Finney, L. S. Langmuir 1994, 10, 3629. Golub, M. A. Langmuir 1996, 12, 2360.

tionality. Electron beam or ion-beam irradiation methods are other alternatives. They have their own merits; i.e., chemical or physical changes occur on the polymer surface without affecting the bulk properties.4,5 We recently used low-energy ion-beam irradiation under a reactive gas environment for polymer surface modification.7 The lowenergy (less than 5 keV) ion-beam irradiation process reduces chain degradation or the cross-linking of irradiated polymers, and added gas molecules, such as oxygen or ammonia, are chemically adsorbed onto the surface to form functional groups. These functional groups can then react or interact with other functional groups on other polymers. Using this technique, we investigated the compatibilization between two immiscible homopolymers (high-density polyethylene, HDPE, and Nylon 66).7 The surface functionalization resulted in good adhesion at the interface, which enabled the stress to be transmitted to the dispersed HDPE phase and to deform it. In this study, we further extended this idea to in situ composite preparation. In situ composites are blends of thermotropic liquid crystalline polymers (TLCPs) and other thermoplastics, in which the TLCPs are used as reinforcing fillers that are not present as a solid phase during processing of the blends; instead, the fibril-shape solid phase is formed when the material is cooled to the solid state (“in situ” shaping during processing).8,9 In situ composites have attracted a great deal of interest because they can solve some problems that arise during the processing of conventional fiber-reinforced composites. However, in situ composites have some drawbacks. One (7) Kim, H. J.; Lee, K.; Seo, Y.; Kwak, S.; Koh, S. Macromolecules 2001, 34, 2546. (8) Seo. Y. In Handbook of Engineering Polymeric Materials; Cheremisinoff, N. P., Ed.; Marcel Dekker: New York, 1997; Chapter 38. (9) Lee, S. M.; Seo, Y.; Hong, S. M.; Hwang, S. S.; Park, T. S.; Kim, K. U., Lee, J. W. Polymer 1994, 35, 519.

10.1021/la020038k CCC: $22.00 © 2002 American Chemical Society Published on Web 07/09/2002

6186

Langmuir, Vol. 18, No. 16, 2002

Kim et al.

Chart 1

of the drawbacks is that most thermoplastics are incompatible with TLCPs. This incompatibility between the matrix polymers and the reinforcing TLCPs leads to poor interfacial adhesion, which brings about a reinforcing effect less than that expected from the law of mixtures.9 If any compatibility between the TCLP phase and the matrix polymer is provided, the interfacial tension will decrease, which means less resistance against deformation of the dispersed TLCP phase and good adhesion at the interface. Our goal in this paper is to gain an understanding of the structure development of the TLCP phase due to surface functionalization of the matrix polymer. We investigate the effect of the poly(ether imide) (PEI) surface functionalization on the physical properties as well as the interfacial properties of TLCP/PEI in situ composites. We chose PEI as the matrix polymer because it is resistant to heat-induced degeneration up to a temperature on the order of 450 °C; thus possible beaminduced heating effects can be prevented. Since we have already investigated in situ composites of PEI/TLCP,9,10 we can also easily study the effects of surface functionalization on the blend properties. 2. Experimental Section 2.1. Materials. The chosen TLCP was an all aromatic liquid crystalline poly(ester amide), Vectra B950 (VB) (a copolymer based on 6-hydroxy-2-naphthoic acid (60%), terephthalic acid (20%), and aminophenol (20%)) produced by Hoechst Celanese Co. Poly(ether imide) (PEI), commercially known as Ultem 1000 (an amorphous polymer made by G.E.), was used as the matrix. Chart 1 shows the chemical structures of these polymers. VB, 25 wt %, was blended with PEI. 2.2. Ion-Beam Irradiation and Blending. Since the details of the ion-beam irradiation equipment were presented in our previous paper,7 we briefly introduce the basic tools. The ionbeam-assisted reaction system was composed of a conventional ion-beam system, a reactive gas feeding system, and the polymer sample (PEI pellets or powder) mixing bowl. Figure 1 shows a schematic diagram of the ion-beam irradiation reactor. The working pressure in the reaction chamber was kept under 10-4 Torr. The Ar+ ion beam was generated from a 5-cm, cold hollowcathode ion source, and its potential energy was maintained at less than 1 keV. The flow rate of Ar gas, which was ionized to Ar+ by the ion source, was fixed at 2 cm3/s. The mixing bowl was equipped with a rotor blade for uniform mixing during ion-beam irradiation. Reactive O2 gas was constantly injected from the bottom of the chamber. The flow rate of the O2 gas was 3 cm3/s and was controlled by using a mass flow controller (MassFlo 9121). PEI and VB pellets were dried in a vacuum oven at 100 °C for 24 h. The dried PEI pellets (100 g) were ground into powder and then put into the mixing bowl in the ion-beam chamber. The mixing bowl was covered with a circular Faraday cup to protect the powder from the ion beam before irradiation. Ar gas was injected into the ion gun, and the appropriate current density for (10) Seo, Y.; Hong, S. M.; Hwang, S. S.; Park, T. S.; Kim, K. U.; Lee, S. M.; Lee, J. W. Polymer 1995, 36, 515 and 525.

Figure 1. Schematic diagram of the ion-beam irradiation reactor. treatment could be set by adjusting the discharge voltage and the ion-beam potential. After the ion gun had reached a stable condition and the current density had reached a steady value, ion-beam irradiation was started while stirring the powder with the rotor blade. O2 gas was injected from the bottom of the mixing bowl. After a predetermined period, Ar and O2 gas injection was terminated. The irradiated PEI powder was then premixed in a container with dried VB pellets at a predetermined weight ratio. Finally, the mix was blended in a twin-screw extruder (PRISM) at 330 °C. The twin-screw extruder was equipped with a drawing unit to impart different drawing ratios. The drawing ratio is defined as the ratio of die exit diameter to the final fiber diameter. 2.3. Scanning Electron Microscopy. Scanning electron microscopy (SEM) observations of the composites were performed on a Hitachi S-2500 C model. The fractured surfaces of the blends were prepared by using cryogenic fracturing in liquid nitrogen followed by a coating with gold in an SPI sputter coater. The morphology was determined using an accelerating voltage of 15 keV. 2.4. X-ray Photoelectron Spectroscopy (XPS). The chemical components on the surface of the ion-beam-treated PEI (IBPEI) were analyzed by using X-ray photoelectron spectroscopy (XPS), and the spectra were recorded with a Surface Science 2803-S spectrometer (hν ) 1.5 keV). A basic pressure of 2 × 10-10 Torr was maintained during the analysis, and the energy resolution was 0.48 eV. The XPS spectra were obtained using a monochromatic Al KR X-ray source, and the peaks were referenced to the main component of the C 1s peak of PEI at a binding energy of 284.6 eV.4,6 The irradiation treatment generally resulted in a small shift of all of the peaks (up to ca. 0.6 eV) toward higher binding energies, implying an increased conductivity for the modified surfaces. Overlapping peaks were resolved by using a peak synthesis method based on Gaussian peaks. 2.5. Rheometry. The rheological properties were measured using a UDS200 (Physica, Germany) rheometer on which a 25 mm diameter cone and plate were mounted. The frequency range was set at 0.1-600 rad/s, and the applied strain was 5%. Before the measurement, the samples were prepared using a compression molder at 330 °C. The measurements were done under a nitrogen atmosphere. 2.6. Mechanical Properties. Testing of the mechanical properties of the blends was undertaken using an Instron Universal Testing Machine (model 4204) at a constant temperature. A crosshead speed of 10 mm/min was used. All the reported results are averages of at least 10 measurements. 2.7. FT-IR Spectrum. FT-IR spectra were obtained using a Bruker 200 spectrometer (IF 66) with an average of 200 scans at a resolution of 4 cm-1. Attenuated total-reflection (ATR) adsorption spectra were recorded using an ATR accessory at a reflection angle of 30°. 2.8. Correlation between Rheology and Morphology. In polymer blends, most of the common polymers are incompatible, so they form multiphase systems. In the case of two components, the minor phase is usually dispersed in the form of spherical inclusions of different sizes into the major phase (matrix). Since

Surface-Functionalized PEI/TLCP Blend Interface

Langmuir, Vol. 18, No. 16, 2002 6187

the ultimate properties of the blends depend on the size distribution of the minor phase and the interfacial tension between components, good insight into the relationship between the blend morphology and its rheological behavior is essential to optimize the final physical properties of the blends. The relationship is reciprocal in the sense that the applied flow history can change the blend morphology while the two-phase structure affects, in turn, the rheology.7 Recent advances in the field of polymer-blend rheology, which were based on the use of theoretical models, have allowed us to obtain quantitative relationships between linear rheological properties on one hand and interfacial properties on the other hand.12 Thus, in principle, rheological data can be used to evaluate the average particle size and/or the interfacial tension; i.e., if a rheological model and information about the size of the dispersed phase are provided, the interfacial tension between the blend components can be determined. In 1990, Palierne published a model which has become the most widely used rheological model for polymer blends with a matrix and spherical inclusions.13 The Palierne model describes the complex modulus of molten blends, GB*, as a function of the complex modulus of each phase, GM* for the matrix and GD* for the dispersed phase. The viscoelasticity of both phases, the hydrodynamics interactions, the droplet sizes, the droplet-size distribution, and the interfacial tension are included in this formulation, but steric interactions and anisotropic effects which often occur in concentrated systems are not included. This limits the use of the Palierne model to moderate concentrations.14,15 The interfacial tension, R, is the sole parameter describing the interfacial properties between components.12 If the effects of gravity and inertia are neglected, GB* can be expressed as a function of the volume fractions, φi, of droplets of radius Ri by

GB*(ω) )

1 + 3ΣiφiHi*(ω) 1 - 2ΣiφiHi*(ω)

GM*(ω)

(1)

where Hi* is given by

Hi*(ω) ) {4(R/Ri)[2GM*(ω) + 5GD*(ω)] + [GD*(ω) GM*(ω)][16GM*(ω) + 19GI*(ω)]}/{40(R/Ri)[GM*(ω) + GD*(ω)] + [2GD*(ω) + 3GM*(ω)][16GM*(ω) +19GI*(ω)]} (2) and R being the interfacial tension between the two polymer blend components and G*(ω) the complex shear modulus, G*(ω) ) G′(ω) + iG′′(ω). G′ is the dynamic storage modulus and G′′ is the dynamic loss modulus at a given frequency ω. Since a smallamplitude, oscillatory flow does not affect the morphology, it can be used to study the effect of the morphology on the rheology in a nondestructive manner. Equation 1 can only be used in the linear viscoelasticity region, that is, for cases of small-amplitude, oscillatory shear flows. Therefore, this model cannot predict the morphological changes during the flow. However, we can use the model since we are interested in the interfacial tension, not in the morphology changes during blending. The summation is carried out over the distribution of particle sizes (φi being the volume fraction of particles of radius Ri). For distributions which are not too broad (the polydispersity in size, dv/dn, where dv is the volume-averaged diameter and dn is the number-averaged diameter, does not exceed 2.3), the sum over (φi,Ri) in eq 1 can be replaced by a single term H(φ,Rv), where φ is the total volume fraction of inclusions and Rv is the volume-averaged particle radius.14,15

3. Results and Discussion 3.1. Surface Functionalization. The details of the chemical reactions induced by particle bombardment at (11) Frydman, E.; Cohen, H.; Maoz, R.; Sagiv, J. Langmuir 1997, 13, 5089. (12) Carreau, P. J.; DeKee, D. C. R.; Chbara, R. P. Rheology of Polymeric Systems; Hanser: Munich, 1997; Chapter 8. (13) Palierene, J. F. Rheol. Acta 1990, 29, 204. (14) Graebling, D.; Mulle, R.; Palierene, J. F. Macromolecules 1993, 26, 320. (15) Bousmina, M.; Bataille, P.; Supieha, S.; Schreiber, H. P. J. Rheol. 1995, 39, 499.

Table 1. XPS Spectrum Results of Ion-Beam-Treated PEI Powder carbon (%)

oxygen (%)

irradiation time

CH

CN

C-O

CdO

COO

CdO

C-O

(theoretical) 0 min (pristine) 5 min 10 min 20 min 40 min

73 78.7 80.8 74 72.5 79.3

5.4 5.1 0.9 0.8 0.7 0.1

10.8 10.6 10.3 14.8 15.5 11.5

10.8 5.6 5.0 5.7 5.8 5.3

3.1 4.7 5.5 3.8

66.7 63.2 62.4 61.0 58.9 61.1

33.3 36.8 37.6 39.0 41.1 38.9

low energy in poly(N,N′-(p,p′-oxydiphenylene)pyromellitimide) have been studied by Marletta et al.16 using XPS and reflection energy-loss spectroscopy. They identified complicated reaction channels such as the destruction of imidic rings through a random mechanism involving the production of recoiling oxygen atoms, which form additional ether or hydroxyl groups with the backbone. Since we added reactive oxygen gas to the irradiated PEI, its reaction channels were more complicated. In the present work, however, we were more concerned with the surface functionalization after ion-beam irradiation than with the true chemical reaction mechanism. Thus, we used XPS to investigate as a function of irradiation time both the formation of the functional groups and their changes. The modified surface of the thin layer was suitably characterized by using XPS. Figure 2 shows the characteristic modification trend for the C 1s peaks for untreated and 20-min ion-beam-irradiated PEI (IBPEI). The peak synthesis and the assignment of components were performed using the data in the literature.3,6 The C 1s peak of untreated PEI showed two separate peaks, one at 288.5 eV (CdO peak) and the other at 286 eV. The latter was an asymmetric peak due to the overlapping of three peaks (the C-H peak at 284.8 eV, the C-N peak at 285.8 eV, and the C-O peak at 286.5 eV). In the case of the IBPEI powder, the binding energy of the carbon peak shifted slightly to higher energy, and a new peak appeared at 288.8 eV corresponding to the CO-O group. The O 1s peaks in Figure 2b show the CdO peak at 532 eV and the C-O peak at 533.6 eV. With the addition of the oxygen gas, the C-O and the CO-O peaks became larger with irradiation. The functional group ratios of PEI before and after irradiation were calculated from the integral intensities of the peaks of each functional group and are shown in Table 1. The amount of nitrogen element was significantly reduced while that of oxygen was increased. In the oxygencontaining groups, the amount of CdO was decreased with irradiation time while that of C-O bond was increased. This is consistent with another observation that oxygen and nitrogen atoms contained in polyimide films are selectively sputtered by the incident particles.17 The most remarkable difference between the samples before and after irradiation was the emergence of the CO-O group. A good correlation was observed between the component ratio of the virgin PEI powder and theoretical ratio, except for the CdO double bond. The measured CdO ratio was 5.6% while the theoretical value was 10.8%. This deficiency of CdO bonds has also been reported by a number of researchers18,19 and is ascribed to hydrogen bonding and/ or cross-linking. This deficiency of CdO bonds also appears in the O 1s component, which has a measured ratio of (16) Marletta, G.; Iacoa, F.; Toth, A. Macromolecules 1992, 25, 3190. (17) Ektessabi, A. M.; Hakamota, S. Thin Solid Films 2000, 377, 621. (18) Flitch, R.; Shih, D. Y. J. Adhes. Sci. Technol. 1996, 24, 3325. (19) Silverman, B. D.; Bartha, J. W.; Clabes, J. G.; Ho, P. S.; Rossi, A. R. J. Polym. Sci., Part A: Polym. Chem. 1986, 24, 3325.

6188

Langmuir, Vol. 18, No. 16, 2002

Kim et al.

Figure 3. FTIR spectra of (a) pristine PEI, (b) 20-minirradiated PEI, and (c) the difference.

Figure 4. Dynamic viscosity (|η*|) for untreated PEI (9) and IBPEI (0). Storage modulus G′ of untreated PEI (b) and IBPEI (O). Loss modulus G′′ of untreated PEI (1) and IBPEI (3).

Figure 2. (A) XPS C 1s spectra of untreated PEI (a) and of PEI ion beam irradiated for 20 min (b). (B) XPS O 1s spectrum of PEI ion beam irradiated for 20 min.

63.2% but a theoretical ratio of 66.7%. As the irradiation time was increased, the CH and CN peaks decreased significantly whereas the C-O and the CO-O peaks, which were not evident in the pristine PEI spectrum, increased rapidly. This result implies that the imide ring of PEI (especially the CN bond) degrades rapidly with

ion-beam irradiation.16,17 This can be corroborated by the FTIR spectrum. Figure 3 shows the FTIR spectra of the pristine PEI, the 20-min-irradiated PEI, and their difference. The peaks corresponding to the CdO, the CdC, and the CN bonds decreased, while the peak corresponding to the C-O bond increased. The carbonylic component was substantially depleted by the ion-beam irradiation whereas the ether-like band showed apparent stability, which was attributed to the high radiation resistance of the ether linkage which was stabilized by two contiguous benzene rings.16 After a long irradiation time, the etherlike band slightly decreased. The late increase in C-H is due to the formation of a C-C-C structure at 285 eV. Thus, it may be concluded that polyimide films show complicated behaviors when irradiated by an ion beam.16,17 3.2. Rheological Properties. The rheological data for PEI and ion-beam-irradiated PEI are displayed in Figure 4. Both polymer melts show an almost constant dynamic viscosity, but the IBPEI melt shows a slightly higher zeroshear viscosity, possibly due to the high molecular weight portion produced during the irradiation. The frequency dependence of the dynamic moduli are also presented in Figure 4. IBPEI shows larger values of the modulus

Surface-Functionalized PEI/TLCP Blend Interface

Figure 5. Dynamic viscosity (|η*|) for the untreated VB/PEI blend (9) and the VB/IBPEI blend (0). Storage modulus G′ of the untreated VB/PEI blend (b) and the VB/IBPEI blend (O). Loss modulus G′′ of the untreated PEI /VB blend (1) and the IBPEI/ VB blend (3).

Langmuir, Vol. 18, No. 16, 2002 6189

Figure 7. log G′ versus log G′′ for the untreated VB/PEI blend (filled symbols) and the VB/IBPEI blend (open symbols).

Figure 8. SEM micrographs of the fractured surfaces: (a) untreated VB/PEI blend; (b) VB/IBPEI blend.

Figure 6. log G′ versus log G′′ for untreated PEI (filled symbols) and IBPEI (open symbols).

(especially the dynamic storage modulus) at low frequency. The complex viscosity and dynamic moduli for VB/PEI blends are shown in Figure 5. A comparison of the dynamic viscosity values between Figure 4 and Figure 5 reveals that the blends have lower dynamic viscosity values than the matrix polymers. This indicates the role of the thermotropic liquid crystalline polymer as a processing aid.20 However, for the same reason, the IBPEI blend still shows higher viscosity values than the pure PEI blend does. Figure 6 presents plots of log G′ vs log G′′ for both polymers. The untreated PEI melt shows a homogeneous phase behavior, that is, the plot is independent of the temperature and has a slope of 2 in the low-frequency region. To the contrary, the ion-beam-treated PEI melt does not show the same behavior. Although its log G′ versus log G′′ plot is independent of the temperature, the slope in the terminal region is less than the value of 2 required (20) Seo, Y.; Kim, H.; Kim, B.; Hong, S.M.; Hwang, S. S.; Kim, K. U. Korea Polym. J. 2001, 9, 238.

for a truly homogeneous phase. Thus, the ion-beamirradiated PEI melt cannot be regarded as being truly homogeneous,21 which means that although the morphological state does not change with temperature, microheterogeneity still exists in the melt.22 This homogeneity difference disappears in the blends. The dynamic modulus values for both blends show nonlinear behaviors in the low-frequency range by mixing with VB, which has a long relaxation time (Figure 7). Both blends show similar behaviors at 330 °C. Because of the low modulus values for VB,9 the difference between the dynamic modulus values for the two blends decreased with temperature, but the VB/IBPEI blend showed a more linear behavior than the VB/PEI blend, which implies a more uniform blend. It is worthy of note that the VB/IBPEI blend showed less heterogeneous behavior than the VB/PEI blend. This is attributed to the interaction between IBPEI and VB. More details are discussed below and are correlated with the morphological observation. 3.3. Morphology and Interfacial Tension of the Blends. As mentioned in the Introduction, one of the most popular methods for making a compatible polymer blend is to provide some interaction between two immiscible polymers, which brings about a reduction in the total free energy.1,7 By using ion-beam-assisted reactions, we can simply provide some functional groups on the backbone (21) Kim, J.; Lee, H. H.; Son, H. W.; Han, C. D. Macromolecules 1998, 31, 8566. (22) Eaton, P. J.; Graham, P.; Smith, J. R.; Smart, J. D.; Nevell, T. G.; Tsibouklis, J. Langmuir 2000, 16, 7887.

6190

Langmuir, Vol. 18, No. 16, 2002

Kim et al.

Figure 9. Comparison of model predictions with experimental data: (A) the untreated VB/PEI blend at 330 °C and (B) the VB/IBPEI blend at 330 °C. Dashed lines are model predictions. Symbols are experimental data (G′(b,O) and G′′(1,3)).

Figure 10. (A) Tensile strengths of binary blends with different draw ratios: (b) untreated VB/PEI blends and (9) VB/IBPEI blends. (B) Tensile moduli of binary blends with different draw ratios: (b) untreated VB/PEI blends and (9) VB/IBPEI blends. The lines are guides for the eyes.

of PEI, and we expect some of the attached functional groups to interact with VB. To verify this interaction between PEI and VB, we investigated the morphology of the blends. SEM photographs of the fractured surfaces of binary blends are shown in Figure 8. The untreated PEI blend shows large discrete domains of VB (volumeaveraged diameter dv ) 5.3 µm). In the VB/PEI blends, poor interfacial adhesion due to poor interaction and immiscibility is evident from the many round holes around the dispersed VB phase. Ion-beam irradiation has a strong effect on the blend morphology. The size of the dispersed phase is remarkably reduced (from 5.3 to 3 µm), the VB phase is finely dispersed, and adhesion between VB and IBPEI is definitely improved. The better dispersion and better adhesion clearly demonstrate the existence of some interaction between IBPEI and VB. The differences in the morphology are related to different interfacial tensions. A theoretical model allows us to obtain quantitative relationships between linear rheological properties on one hand and morphological interfacial properties on the other hand.12-15 Hence, the rheological data can be used to evaluate the average particle size and/or the interfacial tension.11,12 If the particle size is provided, the interfacial tension can be determined by fitting the theoretical predictions to the experimental data. By applying the theoretical model with only one adjustable parameter, namely the ratio (R/R), we produced the results in Figure 9. Since the dynamic loss modulus G′′ is not sensitive to droplet deformation, we discuss the storage modulus G′ first. Because of the long relaxation time of the VB phase,23 no shoulder

Figure 11. Calculated aspect ratio: (b) untreated VB/PEI blends and (9) VB/IBPEI blends.

(23) Marucci, G, In Thermotropic Liquid Crystal Polymer Blends; La Mantia, F. P., Ed.; Technomic: Lancaster, 1993; Chapter 2.

corresponding to dispersed-phase relaxation at low frequency appears in G′. In the model fitting, R/R was the optimized value, and the optimized values of R for the Palierne model were 2 mN/m for the VB/PEI blend and 0.7 mN/m for the VB/IBPEI blend. A decrease in the interfacial tension is direct evidence for a strong interaction between IBPEI and VB. Though the content of VB was large (25 wt %), the results from the Palierne model showed a good correlation with the experimental data over the whole frequency range. Failures in Palierne model predictions have been reported in concentrated systems which can be affected by steric interactions or anisotropic effects.15 The good agreement between the predictions of the Palierne model and our experimental data may be possibly due to the viscosity of the dispersed phase (VB) being lower than that of the matrix (PEI or IBPEI). Since the loss modulus of polymer melts is much larger than

Surface-Functionalized PEI/TLCP Blend Interface

Langmuir, Vol. 18, No. 16, 2002 6191

Figure 12. SEM micrographs of the fractured surfaces at different draw ratios: (A) untreated VB/PEI blends and (B) VB/IBPEI blends.

the storage modulus in the low-frequency range, viscoelastic behavior at low frequency is predominantly governed by the zero-shear viscosity. The loss modulus is also in good agreement with the predictions of the Palierne model because it is less sensitive to deformation. 3.4. Mechanical Properties and Interfacial Interaction. The tensile strength and the tensile modulus of the blends are shown in Figure 10 as functions of the draw ratio, which is defined as the ratio of the die exit diameter to the fiber diameter at far down stream where the fiber has a constant diameter. The tensile strength increases with the draw ratio due to the reinforcing effect of the dispersed VB phase, which has a fibril shape. The tensile strength of ion-beam-treated blend (VB/IBPEI) is higher than that of the pristine PEI blend. The tensile modulus of the ion-beam-treated blend also shows higher values than those of the VB/PEI blend. However, the differences are not so remarkable as they are for the compatibilized blend case.10 This indicates that there are some interactions between the ion-beam-irradiated PEI and the VB. However, those interactions are not as strong as the compatibilizer effect, so they cannot significantly change the mechanical properties. In other words, the surface functional groups have less of an effect than expected. VB is known to react with maleic anhydride groups,24 but the surface functional groups of IBPEI do not react or interact so much with the VB molecules to form a kind of block or graft copolymers that can compatibilize the PEI and VB. This implies that surfacefunctionalized PEI does not provide as strong an interaction as the reactive compatibilizer does and/or that the reaction did not occur sufficiently to provide enough (24) Seo, Y. J. Appl. Polym. Sci. 1997, 64, 359.

compatibilizer. The high viscosity of PEI, and thus the low chance of encounters between the reactive functional groups, is thought to be the reason. Figure 11 shows the theoretical aspect ratio for the VB phase calculated by using the Tsai-Halpin equation.9 The aspect ratio of the VB/IBPEI blend is higher than that of the VB/PEI blend, which is due to lower interfacial tension and better adhesion at the interface. Figure 12 shows the fractured surfaces of strands with different draw ratios. In the VB/IBPEI blends, more orientation (fibrillation) of the VB phase and better adhesion at the interface are clearly observable. Also, the VB/IBPEI blends show smaller domains of the dispersed VB phase. For the VB/IBPEI blends having improved adhesion at the interface, the stress is transmitted to the dispersed VB phase, and the break proceeded through the dispersed phase as well as the matrix. Pull-out of the deformed dispersed phase requires extra energy to overcome the frictional force due to improved adhesion.10 As a result, the tensile strength of the system is increased. Also, the elongation at break was improved for the ionbeam-irradiated VB/PEI blends (3.5% and 4.6% for VB/ PEI blends and 4.7% and 5.8% for VB/IBPEI blends at a draw ratio of 1 and 8, respectively) because the dispersed fibrils maintained contact with the sheath of the matrix surrounding the fibrils before breaking. In untreated PEI blend systems, the propagating stress passes around the VB phase because the adhesion at the interface is quite poor. The VB phase does not maintain contact with the sheath of the matrix and is easily pulled out; thus, elongation does not increase. The simultaneous increases in tensile strength and elongation are decisive evidence of improved adhesion due to interactions at the interface.

6192

Langmuir, Vol. 18, No. 16, 2002

4. Conclusions In this study, we experimentally demonstrated that the ion-beam-assisted gas reaction technique induced structural modification of the matrix polymer and significantly changed the physical properties of immiscible VB/PEI blends by causing some interactions between the two phases. The interactions are attributable to surface functional groups (mostly carboxyl groups and carbonyl groups) added on the surfaces of PEI particles. Though the modification proceeds at a relatively shallow depth below the surface, the physical properties of the ion-beamirradiated PEI (IBPEI) were changed by the structural variation (chain scission and recombination, and imide ring degeneration). The interfacial properties of the VB/ PEI blends are also clearly changed due to the improved adhesion at the interface and the reduced interfacial tension. The interfacial tension was evaluated using an emulsion model (Palierne model). The value of the interfacial tension calculated for the VB/IBPEI blend was less than that calculated for the VB/PEI blend, which shows the effect of functionalization, which was caused by the ion-beam-assisted gas reaction, on the PEI surface.

Kim et al.

Good adhesion at the interface enabled the stress to be transmitted to the dispersed VB phase and to deform the droplets. Hence, extra energy was consumed by plastic deformation of VB. Also, extra energy was consumed in the tensile extension process to overcome the larger frictional force due to better adhesion. As a result, the tensile strength of the system increased. However, the improvements in the mechanical properties were not as significant as in compatibilizer added systems.10 This implies that surface-functionalized PEI did not provide such a strong interaction as a reactive compatibilizer and /or the reaction did not occur sufficiently to provide enough compatibilizer. The high viscosity of PEI, and thus the low chance of an encounter between reactive functional groups, is thought to be the reason. Acknowledgment. The financial support by KIST (No. 2E17421) and MOCIE (No. 2M11320) is greatly appreciated. LA020038K