Journal of Colloid and Interface Science 336 (2009) 872878

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

An investigation into the improvement of adhesive strength of polyimides by incorporation of elastomeric nanoparticles
Sepideh Khoee a,*, Ali Reza Mahdavian b, Warahram Bairamy a, Mohsen Ashjari b
a b

Polymer Laboratory, Chemistry Department, School of Science, University of Tehran, P.O. Box 14155-6455, Tehran, Iran Polymer Science Department, Iran Polymer & Petrochemical Institute, P.O. Box 14965/115, Tehran, Iran

a r t i c l e

i n f o

a b s t r a c t
Copoly(styrenebutyl acrylateethyleneglycoldimethacrylate) (StBAEGDMA) nanoparticles were prepared using miniemulsion polymerization technique. Then the dispersed nanoparticles in DMAc were added to in situ condensation polymerization media of pyromellitic dianhydride (PMDA) and oxydianiline (ODA) and consequently, homogenous polyamic acid solution containing the nanoparticles was obtained. Novel polymerpolymer nanocomposites were prepared by casting of PMDAODA polyamic acid solution with various content of the above elastomeric nanoparticle (ENP) on a glass plate followed by thermal imidization. All samples were characterized after preparation by FT-IR spectroscopy, transition electron microscopy (TEM), thermal gravimetry analysis (TGA) and differential scanning calorimetery (DSC). To investigate the adhesion strength of polyimides lled with (StBAEGDMA) nanoparticles, lap shear strength (LSS) test was examined on different metallic surfaces. Effect of nanoparticles content on the adhesion properties of this polymer was considerable for aluminum surface. Lapshear strength and adhesive energy of the bonded samples were found to initially increase with the increase in ENP wt%, but decrease after a critical value. It was shown that by increasing the nanoparticles amount up to 25 wt%, the adhesion strength of polyimides increased due to the good wetability of surfaces. After that and by increasing the nanoparticles amount, the adhesion strength decreased according to the diminished strength between polyimide chains. Scanning Electron Microscopy (SEM) micrographs of the fractured surfaces were taken to determine the failure mode. They showed that by increasing the nanoparticle content in the polyimide matrix, failure mode was converted from adhesion failure to cohesion one. 2009 Published by Elsevier Inc.

Article history: Received 23 November 2008 Accepted 22 April 2009 Available online 3 May 2009 Keywords: Polymerpolymer nanocomposites Polyimide adhesive Elastomeric nanoparticles Ultrasound

1. Introduction Polyimides (PIs) are a family of heterocyclic polymers that have received extensive evaluation as adhesives, bers, lms, molding, composite matrices, coatings and membranes. As a class of polymers, polyimides posses a favorable combination of properties, that make them attractive for many applications. The high thermal, thermo-oxidative and chemical stability coupled with good mechanical properties, makes these polymers particularly attractive for use in harsh environments [1]. Several polyimides have displayed excellent properties as adhesives. High-performance adhesives are needed for joining metals, composites, ceramics and plastics to themselves and to each other. These adhesives nd use in a variety of applications in the aerospace, automotive, electronic, domestic household, oil and other industries. High performance adhesive polyimides should have high molecular weight. These polyimides impart good mechanical
* Corresponding author. Fax: +98 21 6649 5291. E-mail address: Khoee@Khayam.ut.ac.ir (S. Khoee). 0021-9797/$ - see front matter 2009 Published by Elsevier Inc. doi:10.1016/j.jcis.2009.04.067

and thermal properties but on the other hand, limit their applications as they are insoluble in most organic solvents and have high melting temperature in the range of 300400 C. The processing of such PIs involves use of temperature higher than 400 C, which often leads to degradation of the polymer [2]. High melting temperature (Tm) leads to increase the ow temperature of polyimide adhesives and results in poor wetting ability of these adhesives, which limit their applications. The subject of many contemporary researches is to overcome this limitation. Preparation of polyimide containing oxyalkylene units [3], partially uorinated polyimide [4], polyimide sulphone [5] and polyimides containing aliphatic segments [6] are typical methods for changing in the chemical structure of primary polyimides. Another method is using polyimide with low molecular weight or oligoimides that have low glass transition temperature (Tg) [7]. All of these methods improve wettability of the adhesives but has improper effect on their mechanical properties. Saeed and Zhan used two different types of nanoparticles namely multi-walled carbon nano-tubes and aluminum nitride (AlN) nanopowder to reinforce PI adhesive matrices which had lowest effect on their melt ow [8,9].

S. Khoee et al. / Journal of Colloid and Interface Science 336 (2009) 872878

873

A miniemulsion is an aqueous dispersion of relatively stable oil droplets with a size range of 50500 nm, prepared by shearing a system containing oil, water, surfactant and hydrophobe (cosurfactant). Because of the small droplet size and large surface area, the monomer droplets can effectively capture the radicals and become the dominant sites for particle nucleation in a miniemulsion polymerization process [10]. In miniemulsion polymerization, both particle nucleation and subsequent propagation reaction occur primarily in submicrometer monomer droplets. Each of those droplets can be regarded as an individual nanophase reactor [11]. On the other hand, ultrasonic irradiation is a useful technique for preparing novel materials with unusual properties. Ultrasound has been widely used in chemical industries, such as dispersing, emulsifying, crushing, organic synthesis and initiating polymerization of the monomers [12]. Ultrasonically initiated miniemulsion polymerization has many advantages, such as chemical-free initiation and low reaction temperature. In this study, StBAEGDMA nanoparticles were prepared using miniemulsion polymerization technique and the obtained nano-size elastomeric particles were applied to improve wettability of polyimide adhesives. Here, the polyimide is based on pyromellitic dianhydride (PMDA) and oxydianiline (ODA). Vespel is the trademark of this durable high-performance polyimide manufactured by DuPont [13]. This rather expensive polymer is mostly used in aerospace, semiconductor and transportation technologies. It combines heat resistance, lubricity, dimensional stability, chemical and creep resistance of those used in hostile and extreme environmental conditions. This polymer has a Tg of approximately 289 C. After preparing polymeric matrix, the elastomeric StBAEGDMA nanoparticles were added to the above matrix. The rst step in the adhesion study was selection of an appropriate Tg for the binding polyimide. Polyimides inherently show a very high melt viscosity due to their rigid aromatic backbone. This has been one of their major drawbacks and efforts have continued to lower their viscosity to improve their processability. Wetting of the substrate and good interfacial contact are critical points in obtaining a strong and durable interface. It was found that lowering the Tg to 177 C, decreases the viscosity substantially, while still produces breakable adhesives. 2. Materials and methods 2.1. Materials Styrene (St) from Merck Chemical Co. (analytical grade) was washed with 5% sodium hydroxide aqueous solution (wt%) to remove the inhibitor, dried over calcium chloride and stored at 0 C. PMDA was recrystalized in dry acetic anhydride and ODA was puried by sublimation before use. N,N-dimethylacetamide (DMAc) was dried over BaO and then distilled under reduced pressure. Butyl acrylate (BA) and ethylene glycol dimethacrylate (EGDMA) from Fluka, sodium dodecyl sulfate (SDS) from Aldrich, hexadecane (HD) and benzoyl peroxide (BPO) from Merck Chemical Co. were used without further purication. 2.2. Apparatus The equipment employed in this research was a 20 kHz 500 Hz ultrasonic generator, SONOPULS Ultrasonic homogenizer, Model HF-GM 2200 (BANDELIN electronic GmbH & Co. KG) and the probe type was a titanium microtip MS-73 with 3 mm in diameter. FT-IR spectra were recorded on a FT-IR BRUKER- IFS 48 spectrophotometer (Germany) using KBr pellets. Mean particle size was measured by SEMATECH light scattering (France) with 633 nm wavelength. Thermogravimetric analysis (TGA) were recorded on a Pyris1 Perkin-Elmer under N2 atmosphere with heating rate of

20 C/min. Differential scanning calorimetry thermograms of the copolymers were obtained using a computer-interfaced calorimeter (Perkin-Elmer Pyres DSC) under a nitrogen atmosphere and a heating rate of 10 C/min from ambient to 350 C. The glass transition temperature (Tg) of the copolymers was measured according to their thermograms. TEM micrographs were taken by a ZEISS 902A (Germany) at an acceleration voltage of 80 keV. The PI samples were suspended in DMAc and diluted up to 20 times, stained with OsO4 and dropped on a copper grid for TEM analysis. The surface morphology was studied by Cambridge S-360 SEM scanning electron microscopy (SEM). Lap-shear strength (LSS) of bonded joints was measured in accordance with ASTM D1002. SANS 100 kN tensile testing machine equipped with a PC was used to acquire data. Due to the guide lines provided by ASTM D1002, head speed was 1.3 mm/ min and testing temperature was maintained at 25 2 C. Determination of LSS at high temperature was carried out at 250 C. During the tests, samples were exposed to test temperature in a furnace (mounted on tensile testing machine) for 1 h and then subjected to load for measuring LSS. 2.3. Miniemulsion polymerization StBAEGDMA monomers with molar ratios of 0.97:1:0.03 were polymerized isothermally through miniemulsion copolymerization in a 500 ml four necked round bottomed glass reactor, which was equipped with a reux condenser, a mechanical stirrer and a nitrogen gas inlet. The monomers were premixed through ultrasonication for 5 min at room temperature in order to reach the optimum dispersion of the monomeric phase. HD as a hydrophobic agent and BPO as an initiator were dissolved in the monomeric phase and the obtained mixture was added into the water phase containing SDS and a denite amount of double distilled water. Because of miniemulsication, the nal mixture was ultrasonicated for 10 min until a white milky miniemulsion obtained. The above miniemulsion was moved into the reaction vessel. The initial charge was stirred at 300 rpm and heated up to 75 C by using of a thermostatic bath and the polymerization reaction was progressed under N2 atmosphere for 10 h. The system was cooled to room temperature and the nal latex was obtained without any post-preparative treatments. After removing water, nanoparticles with 98% conversion relative to the initially added non-volatile components were obtained. 2.4. Preparation of PAA containing nanoparticles A predetermined amount of StBAEGDMA nanoparticles were mixed with DMAc and homogenized by pulsy ultrasonic and intensive agitation, to form a stable transparent suspension. The nanoelastomeric lled poly(amic acid)s (PAA)s were prepared by addition of 1.01 gr (5.04 103 mol) ODA and 1.09 gr (5.00 103 mol) PMDA to the homogeneous dispersions of StBAEGDMA nanoparticles at 5 C and then stirring for 6 h at 20 C in a three-necked round bottomed ask equipped with a magnetic stirrer, nitrogen inlet and a drying tube to afford a homogenous poly (amic acid)/ ENP solution. Finally, creamy solutions of PAA containing nanoprticles were obtained. 2.5. Preparation of polyimidic nanocomposites The nanoelastomeric lled poly(amic acid)s (PAA)s in DMAc was casted onto glass plates. Thermal imidization was achieved by holding the samples at 100 C for 24 h, raising the temperature to 150 and 200 C and remaining at each of these temperatures for 1 h and nally to 250 C for 2 h. After completion of the cycle, the

874

S. Khoee et al. / Journal of Colloid and Interface Science 336 (2009) 872878

plates were allowed to cool down to room temperature before removal from the oven. The lms were carefully stripped off the glass plates and dried for another 24 h at 100 C under vacuum. 2.6. Surface treatment In order to make a strong interface, different surface treatments were evaluated. The aluminum, copper and steel coupons used in this study were free of any rust or other surface contaminations. However, the surface was polished with a grade 1000 emery paper for appearance of the new chemically active surface after removal of the protective oxide layer [14]. The coupons utilizing the

standard blasting were washed in water, dried in an oven and subsequently wiped with acetone. 2.7. Preparation of samples for lap-shear strength (LSS) analysis Samples for LSS analysis were prepared in accordance with ASTM standard D 1002 [15]. Size of the sample was 108 25.4 1.62 mm and the bonded area of the joint was 25.4 12.7 mm (317.8 mm2). PI/ENP nanocomposite powders were poured in DMAc to prepare a 50 wt% solution. The solution was mixed till a consistent paste-like mixture was obtained. The paste was then applied on the adherents surface on the required

Scheme 1. Schematic representation of preparation process of elastomeric copolymer, polyimide matrices and PI-ENP nanocomposite lm.

S. Khoee et al. / Journal of Colloid and Interface Science 336 (2009) 872878

875

area using a spatula. The adherents were then placed in an oven and heated till 200 C for 2 h to facilitate the evaporation of volatile materials such as DMAc. After drying process, the other half was placed in a way that an overlap of 12.7 mm was obtained. Five samples were prepared for each test condition and average of LSS for ve samples was reported as the nal result. 2.8. Cure cycle for bonded joints After matching, coupons were placed under hot press for curing cycle. For samples prepared with PI/ENP nanocomposites, the temperature was raised up to 250 C under 1.8 MPa pressure, which is above the recorded Tg and held for 2 h. During this stage, good wetting of the adherent surface was provided. 3. Results and discussion 3.1. Miniemulsion polymerization As shown in Scheme 1, step 1, a mixture of three monomers (St, BA and EGDMA) was prepared as an oil phase at rst. In the next step, the oil phase was dispersed in the water phase including SDS and HD in order to form minidroplets. During miniemulsication process, monomer droplets are formed with a range of 50 500 nm depending on the SDS/HD ratio [16], meanwhile this diameter may be 1020 lm in the conventional emulsions. When an ultrasonicator is used to form a miniemulsion, the ultrasonic waves can break up the monomer droplets. It is also notable that the droplet size may have an inverse proportionality with ultrasonication time and latex particles will follow this trend subsequently [17]. Due to the combined and synergistic effect of surfactant and hydrophobe, monomer droplets will become stable enough to prevent diffusion of monomers to the continuous phase. This will result in disappearing of micelles and consequently, impossibility of nucleation in the continuous phase. The detailed amounts of ingredients have been listed in Table 1. Here, crosslinked nanoparticles with copoly(styrenebutyl acrylateethyleneglycol dimethacrylate) composition were prepared. FT-IR spectra were recorded to conrm the structure of copolymer (Fig. 1). The characteristic peak of butyl acrylate and ethylene glycol dimethacrylate appears at 1730 cm1, which corresponds to the stretching vibration of BA carbonyl group. Those characteristic peaks for styrene monomers are at 700 and 760 cm1 relating to phenyl groups. The bonds at 2854 and 2924 cm1 correspond to aliphatic CAH stretching modes. Absorption bands at 1450 and 1377 cm1 proves the presence of CH2 and CH3 groups of BA and EGDMA segments too. This shows the presence of termonomers in the prepared copolymer. These data are in good accordance with our previously reported spectrum [18]. Particles size of the obtained latex and their distribution were determined by dynamic light scanning (DLS) technique. The obtained mean particle size diameter was about 114 nm (Fig. 2). 3.2. Polyimide synthesis Uniform nanoparticles of ENP, in the polymer matrix without any aggregation, were obtained via dispersing of the ENP lm in the reaction solvent (DMAc) by sonication prior to further
Table 1 The recipe for miniemulsion polymerization.a Components Amounts
a b

Fig. 1. The FT-IR spectrum of copoly(styrenebutyl acrylateethyleneglycol dimethacrylate).

Fig. 2. Mean particle sizes of dispersed nanoparticles in the latex (- - -) and DMAc () obtained from dynamic light scattering analysis.

processing. The mean particle size diameter of the above dispersed nanoparticles in DMAc was about 140 nm. The polyimide samples were synthesized by adding the equimolar amounts of solid PMDA and ODA in DMAc to a sufcient amount of ENP in dry DMAc at 5 C to obtain a 10, 20, 25, 30, 40 wt% for elastomer concentration as shown in Table 2. Then the reaction was allowed to stir for 6 h at 20 to afford a homogenous poly (amic acid)/ENP solution. In the second step, a stepwise thermal imidization procedure was utilized. The prepared nanocomposites were thermally cured under vacuum at 100 C for 24 h, 150 C for 60 min, 200 C for 60 min and 250 C for 2 h (Scheme 1, Step 2). The obtained PI nanocomposites were removed from the glassy plates and dried under vacuum at 100 C for another 24 h. 3.3. FT-IR analysis of PI/ENP nanocomposites FT-IR spectra of nanocomposite specimens containing different percent of ENP have been shown in Fig. 3. The conversion of polyamic acid to polyimide could be observed by increasing the two
Table 2 Amount of reactants in preparation of PI/ENP. Sample 0 10 20 25 30 40
a b

ENP (g) 0 0.21 0.42 0.525 0.63 0.84

PMDA (g) 1.09 1.09 1.09 1.09 1.09 1.09

ODA (g) 1.01 1.01 1.01 1.01 1.01 1.01

DMAc (ml)a 4.31 9.78 12.94 16.57 25.87

DMAc (ml)b 18.9 16.5 13.5 11.7 9.7 4.5

Yield (%) 98.5 99.0 98.6 96.3 97.4 98.8

St 0.97 mol

BA 1 mol

EGDMA 0.03 mol

HD 7.5 wt%
b

SDS 3.5 wt%

b

BPO 1.5 wt%b

The amount of ingredients were set to have a latex with 20 wt% solid content. Based on the total weight of monomers.

The required amount for preparation of ENP dispersions. The required amount for preparation of a polyimide precursor solution containing ENP dispersions.

876

S. Khoee et al. / Journal of Colloid and Interface Science 336 (2009) 872878

Fig. 4. TEM micrograph of PI nanocomposites with 30 wt% of ENP.

Fig. 3. FT-IR spectra of the PI nanocomposites with (a) 0, (b) 10, (c) 20, (d) 30 and (e) 40 wt% of ENP.

overlapped imidic carbonyl peaks at 1720 (imide C@O asymmetric stretching) and 1780 cm1 (imide C@O symmetric stretching). All of the spectra exhibited a strong absorbance at 1377 (CAN stretching) and 723 cm1, which shows the presence of the imidic heterocycle in these polymers. Appearance of a new peak at 2854 and 2924 cm1 in FT-IR spectrum of PI/ENP nanocomposites and its accretion in Fig. 3b to 3e proves the presence of nanoparticles in polymeric matrices. The unreacted polyamic acid, despite thermal imidization, could be found by appearance of a broad acidic hydroxyl peak in all samples at 33003500 cm1. According to the specic application of these PIs, the reactants were heated 2 h instead of 4 h in the nal step of thermal imidization and consequently, imidization step was not completed and some amic acid remained unchanged during thermal cyclization process. These amic acid groups could react with adherent surfaces during the curing cycle. 3.4. Determination of particle size and morphology of PI/ENP nanocomposites Fig. 4 shows the morphology of modied PI by TEM. It could be seen that the spherical elastomeric nanoparticles with particle size of about 4060 nm are well dispersed. The primary nanoparticles and their dispersion in DMAc are about 114 and 140 nm (Fig. 2). It seems that the nanoparticles media could affect on their aggregation degree. The mechanism of dispersion of nanoparticles may be inuenced by the type of chemical or physical interactions between nanoparticlenanoparticle and nanoparticlemedia, such as the Van der Waals or polar interactions and etc. Due to the organic characteristics of the applied nanoparticles, they were enwrapped well in the PI matrix and consequently, the obtained nanocomposite shows smaller particles with good dispersion. 3.5. Thermal analysis Thermogravimetric analysis is a convenient method for evaluation of the thermal stability of polyimides. Thermal degradation of

pure PI and PI/ENP samples has been displayed in Fig. 5. For pure PI, the degradation stage took place at 579.6 C according to the degradation of imide segment [19]. Unlike the pure PI, all the nanocomposite degradation graphs show a two-step transition, which conrms the presence of a new phase. For PI/ENP nanocomposite lms, thermal degradation occurred in two stages. The rst degradation stage appeared at 348.21 C because of the degradation of aliphatic segments of elastomeric repeating unit i.e. butyl acrylateethyleneglycol dimethacrylate. The second stage took place at 551.85 C according to the degradation of imide segments. Generally, it was observed that after introducing nanoelastomeric particles in the PI matrix, the ENP particles must degrade before imide segments and promote the direct thermal decomposition of polymer matrix. Consequently, the remained char yield decreased by increasing the elastomer weight percent. According to Table 3, it can be seen that the thermal stability of pure PI lm is more than PI/ENP hybrid lms. It is attributed to the fact that addition of the elastomeric moieties and disordered structure of PI/ENP hybrid lms could reduce thermal stability of the resulting PI/ENP lms. However, wetting ability characteristics must be investigated if the polyimides are going to be chosen as a good candidate for adhesive materials. This phenomenon may occur above glass transition temperature of the polymer without any signicant weight loss. These parameters can adversely affect recrystallization of the polyimide and indeed this seems to be a useful case for most of the semicrystalline thermoplastic polyimides. Flow viscosity is also an important property for applying thermoplastic adhesives.

Fig. 5. TGA thermograph of nanocomposite lled with (a) 0, (b) 10, (c) 20, (d) 25, (e) 30 and (f) 40 wt% ENPs.

S. Khoee et al. / Journal of Colloid and Interface Science 336 (2009) 872878 Table 3 Thermal properties of some polyimidesilica hybrid lms obtained from TGA thermograms. Sample 0 10 20 25 30 40 IDT (C)a 348.97 335.71 327.41 321.38 314.20 327.21 T5 (C)b 432.14 378.57 385.71 378.63 370.96 375.01 T10 (C)c 560.34 421.43 410.71 401.06 390.32 391.67 Y700 (%)d 55.45 45.65 38.69 37.27 36.09 32.08

877

a IDT (initial decomposition temperature) is the temperature at which initial mass loss was observed. b Temperature at which 5% weight loss was recorded by TGA with a heating rate of 20 C/min under N2 atmosphere. c Temperature at which 10% weight loss was recorded by TGA with a heating rate of 20 C/min under N2 atmosphere. d Weight percent of the non-decomposed material by TGA at 700 C with a heating rate of 20 C/min under N2 atmosphere.

For enhancing the spread-ability and for the ease of applicability of the adhesive, lower viscosity is required [20]. Often, this requirement is obtainable in the presence of lower polymer chain interaction or by using suitable diluents [20]. However, it is well known that the degree of interactions also affects mechanical properties of the polymer. At high nanoparticles percentage, entanglements of the polymer chains are reduced and mechanical properties of the adhesives are diminished. But at the low nanoparticles percentage, the viscosity increases with an increase in the above mentioned entanglements and consequently, a substantial improvement is observed in the properties. Therefore, most hotmelt adhesives can be used above their Tg and afford low viscosities. However, for rigid chain polymers like aromatic polyimides, which often show very high ow viscosities, this problem is especially important. Fig. 6 represents the glass transition temperature (Tg) of pure PI and PI/ENP hybrid lms measured by DSC. Pure PI exhibits the highest Tg value of 289 C. Due to the presence of StBAEGDMA nanoparticles with low Tg, a trend in the decrease of Tg is observed by increasing the elastomeric nanoparticles content in the PI lms. 3.6. LSS analysis It is well known that for any hot-melt adhesive application, melt temperature, adhesion time and pressure are three important parameters that would affect the degree of spreading. In order to attain good adhesion, these parameters need to be properly controlled [20]. Any way, using high temperature could degrade the polymeric adhesive and the interface may lead to failure at the working temperature as a consequence. In this study, the results are presented for specimens that were held for 2 h under 1.8 MPa pressure at 250 C, which is well above the recorded Tg. It was expected that during this stage, a good wetting of adherent surfaces was provided. Measurement of LSS by subjecting single

lap joints to tensile loads is a method for determination of adhesion strength of the bonded joints. This method allows analyzing the failure modes in the bonded joints in addition to the measurement of LSS as shown in Fig. 7. Here, three kinds of coupons: aluminum, copper and stainless steel were selected as adherents. The highest lap-shear strength was obtained for aluminum specimen, whereas the adhesive layer between copper or stainless steel coupons were very fragile after setting. According to Table 4 and due to the differences in thermal expansion coefcient of adhesive and adherent, some thermal stresses may build up and lead to failure of the adhesive layer. Low difference between expansion coefcient of the adhesive and aluminum would be the reason for the presence of stronger adhesion. For aluminum coupons jointed by PI/ENP nanocomposite, LSS amounts were found to increase initially with the increase in nanoparticles weight percent. When the nanoparticle content exceeded more than 25 wt%, a reduction in LSS was observed (Fig. 8). Despite the increase in wetability of the adhesive with the increase in ENP wt%, high nanoparticle content resulted in the poor polymer chain entanglements. This poor interaction hinders the proper and strong adhesion to the adherent surfaces and results in a reduced LSS. 3.7. Investigation of the failure surface The failure of adhered joints can occur inside the adhesive layer (cohesion failure with adhesive residues on both surfaces) or at the interface between the adhesive layer and the adherent surface (adhesive failure), as illustrated in Fig. 9 [22]. The fractured surface of the broken lap-shear samples was analyzed using SEM and micrographs have been shown in Fig. 10. It is

Fig. 7. Jointing of two coupons for LLS analysis.

Table 4 Thermal expansion of adhesive and different adherents. Material LTECa

a

Unlled PI 45 [13]

Al 23 [21]

Cu 17 [21]

Stainless steel 1113 [21]

Linear thermal expansion coefcient (106/K).

Fig. 8. LSS measurement of PI/ENP nanocomposites.

Fig. 6. Variation in glass transition temperature of PI nanocomposites with respect to the ENP wt%.

Fig. 9. Types of failure in adhesive bonds.

878

S. Khoee et al. / Journal of Colloid and Interface Science 336 (2009) 872878

Fig. 10. SEM micrograph of fractured surfaces (a) 0, (b)10, (c) 20, (d) 25, (e) 30 and (f) 45 wt% samples.

evident that with 10 wt% ENP, the nanocomposite provides a very rough and deformed fracture surface. Similar features were observed on both sides of the fractured lap-shear specimen. It can be noted that failure mode was totally adhesive in nature as the adhesive matrix had a very poor melt ow behavior. This kind of rough fractured surface is not common for most high-performance adhesives. Fig. 10b shows the SEM micrograph of the fractured bonded surfaces prepared with PI nanocomposite containing 25 wt% of ENP. A mixed failure mode i.e. cohesive/adhesive was observed for this matrix. In fact, the change in viscoelastic behavior of PIs by ENP causes an increase in wetting of the adherent surfaces and shows good metaladhesive adhesion. Consequently, the failure mode moved toward adhesive with decreasing ENP wt% and toward cohesive with increasing ENP wt%. Thus, an increase in ENP up to 25 wt%, increases the LSS strength of the composite adhesive. Above this amount, due to the decrease in polymer chain entanglements, the lap-shear strengths decreased too. 4. Summary In order to investigate the effect of elastomers on the adhesive properties of thermoplastic polyimides (PI) at nanoscale, several PIs were successfully synthesized through an in situ polymer polymer nanocomposite preparation from pyromellitic dianhydride4,40 -oxydianiline (PMDAODA) and copoly(styrenebutyl acrylateethyleneglycol dimethacrylate) (StBAEGDMA) nanoparticles. The PIs were prepared via a two-step method involving poly(amic acid) (PAA) step followed by thermal imidization to obtain polyimide in powder form. The (StBAEGDMA) nanoparticles solutions with 60 nm diameter as elastomeric nanoparticles (ENP) were mixed with high content (1040 wt%) during synthesis at PAA stage and through mechanical agitation. Adhesive strengths for pure PI and its hybrids were measured on single-lap joints test method. The fracture energy of the PI hybrids was rstly increased and then decreased with the increase in the elastomeric nanoprticle content. The fracture energy of the hybrid with 25 wt% ENP was the highest one among the other PI nanocomposites. These

behaviors were attributed to the effect of Tg and chemical interactions of ENP nanoparticles with PI matrix. The fractured bonded surfaces were examined by scanning electron microscopy to determine failure patterns. The results showed that an appropriate amount of ENP in PI matrix could effectively reduce melt ow temperature of the PI adhesive and restrain the formation of crystalline PIs. On the other hand, the viscoelastic behavior of PIENP composites changed from solid-like to liquid-like by increasing elastomeric content. References
[1] D. Wilson, H.D. Stenzenberger, P.M. Hergenrother, Polyimides, Blackie & Son Ltd., New York, 1990. [2] S. Tamai, T. Kuroki, A. Shibuya, A. Yamaguchi, Polymer 42 (2001) 2373. [3] G. Hougham, P.E. Cassidy, K. Johns, T. Davidson (Eds.), Fluoropolymers 2, Properties, Kluwer Academic Publishers, New York, 2002. [4] M. Ding, Prog. Polym. Sci 32 (2007) 623. [5] T.L. St. Clair, D.A. Yamaki, in: K.L. Mittal (Ed.), Polyimides, vol. 1, Plenum Press, New York, 1984, p. 99. [6] D. Yao, Z.Y. Wang, P.R. Sundararajan, Polymer 46 (2005) 4390. [7] V. Ratta, A. Ayambem, J.E. McGrath, G.L. Wilkes, Polymer 42 (2001) 6173. [8] M.B. Saeed, M.S. Zhan, Int. J. Adhes. Adhes. 27 (2007) 306. [9] M.B. Saeed, M.S. Zhan, Int. J. Adhes. Adhes. 27 (2007) 319. [10] J.M. Asua, Prog. Polym. Sci. 27 (2002) 1283. [11] K. Landfester, N. Bochthhold, S. Foster, M. Antinietti, Macromol. Rapid Commun. 20 (1999) 81. [12] N. Feng, H.M. Li, Sonochemistry and its Applications, Anhui Science & Technology Publishing House, Hefei, 1992. [13] Original DuPont Datasheets. Available from: <http://www2.dupont.com/ Vespel/en_US/Literature/sp1.pdf>. [14] K. Ikegami, T. Fujii, H. Kawagoe, H. Kyogoku, K. Motoie, K. Nohon, T. Sugibayashi, F. Yoshida, Int. J. Adhes. Adhes. 16 (1996) 219. [15] Standard test method for apparent shear strength of single-lap joint adhesively bonded metal specimens by tension loading (Metal-to-metal), ASTM D1002, ASTM International, PA, USA, 2001. [16] M. Antonietti, K. Landfester, Prog. Polym. Sci. 27 (2002) 689. [17] Z.Z. Xu, C.C. Wang, W.L. Yang, Y.H. Deng, S.K. Fu, J. Magn. Magn. Mater. 277 (2004) 136. [18] A.R. Mahdavian, M. Ashjari, H. Salehi-Mobarakeh, J. Appl. Polym. Sci. 110 (2008) 1242. [19] M. Foroutan, S. Khoee, J. Appl. Polym. Sci. 104 (2007) 3228. [20] M.M. Gauthier, Engineered Materials Handbook, Adhesives and Sealants, vol. 3, ASM International, Materials Park, OH, 1995, p. 80. [21] D.R. Lide, Handbook of Chemistry and Physics, 85th ed., CRC Press, 20042005. [22] A. Sheppard, D. Kelly, L. Tong, Int. J. Adhes. Adhes. 18 (1998) 385.