octavinyl polyhedral oligomeric silsesquioxane composite dielectrics

European Polymer Journal 45 (2009) 2172–2183

Contents lists available at ScienceDirect

European Polymer Journal journal homepage: www.elsevier.com/locate/europolj

Macromolecular Nanotechnology

Morphology studies and ac electrical property of low density polyethylene/octavinyl polyhedral oligomeric silsesquioxane composite dielectrics Xingyi Huang a, Liyuan Xie a, Pingkai Jiang a,*, Genlin Wang a, Yi Yin b a

MACROMOLECULAR NANOTECHNOLOGY

Department of Polymer Science and Engineering, Shanghai Key Lab of Electrical Insulation and Thermal Aging, Shanghai Jiao Tong University, Shanghai 200240, China b Department of Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China

a r t i c l e

i n f o

Article history: Received 23 February 2009 Received in revised form 28 April 2009 Accepted 16 May 2009 Available online 22 May 2009

Keywords: Dielectric properties LDPE Morphology Polyhedral oligomeric silsesquioxane (POSS) Rheology Supermolecular structure

a b s t r a c t This paper presents the results of morphological and ac electrical investigations on low density polyethylene (LDPE) composites with octavinyl polyhedral oligomeric silsesquioxane (POSS). It has been shown that at low loadings, the frequency dependence of dielectric constant and dielectric loss for the LDPE/POSS composites showed unusual behaviors when compared with conventional (micro-sized particulates) composites. The ac breakdown strength was measured and statistical analysis was applied to the results to determine the effects of POSS loadings on the dielectric strength of LDPE. The morphological characterization showed that the presence of POSS additives apparently altered the supermolecular structure of LDPE and resulted in more homogeneous morphology when compared with the neat LDPE. The structure–property relationship was discussed and it was concluded that the final dielectric properties of the composites were determined not only by the incorporation of POSS additives but also by the supermolecular structure of LDPE. Rheological analyses of LDPE/POSS composite were also performed and the results showed that the octavinyl-POSS had good compatibility with LDPE. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction In the past decades, the mixing of organic polymers with nano-fillers to form so-called nanocomposites with improved properties attracted considerable attentions [1–6]. One interest of the researchers is to seek the design rules that would allow the composites to be engineering materials with the desirable electrical properties. The purpose of this paper is to investigate the ac electrical properties of low density polyethylene (LDPE)/polyhedral oligomeric silsesquioxane (POSS) composites and to evaluate the possibilities of using LDPE/POSS composites as potential dielectrics or insulating materials in high voltage application. * Corresponding author. E-mail addresses: [email protected] (X. Huang), pkjiang@sjtu. edu.cn (P. Jiang). 0014-3057/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.eurpolymj.2009.05.019

POSS is a class of cage-shaped molecules represented by the formula (RSiO1.5)n with an silica-like core (Si6O9, Si8O12 or Si10O15) surrounded by 6, 8 or 10 organic corner groups, which are becoming one class of versatile inorganic fillers for preparing polymer composites with desirable properties [7]. POSS can be dispersed in polymer matrix by copolymerization, chemical grafting or blending and consequently lead to some dramatically improved properties, such as, increase of thermal stability [8] and mechanical modulus [9], as well as reduction in dielectric constant [10], crystallization time [11] and flammability [12]. Although, POSS have been proved to be efficient to decrease the dielectric constant of some polymers, little is known about the dielectric properties of polymer/POSS composites. This paper constitutes a contribution to the understanding of the effects of POSS on the dielectric properties of LDPE. For materials used as dielectric applications, the dielectric strength is one of the properties that must be

taken into account in order to check the ability to withstand high electric field [13–15]. Therefore, we paid much attention to the investigations on dielectric breakdown strength of the composites. Dielectric strength is defined as a relationship between the breakdown voltage and the sample thickness, representing the maximum field what the materials can withstand. There exist many factors influencing the ac dielectric strength of polymers [16,17]: environment conditions (temperature, moisture), applied voltage frequency, electrode conditions and the polymers themselves involving morphology, additives, defects and so on. In the following a brief review is given of early work in which the dielectric strength of solid polymers has been related to various factors. (i) Temperature: the temperature dependence of breakdown mechanisms for most thermoplastic polymers like LDPE has been extensively investigated and could be divided into three categories [18]: low temperature region, middle high temperature region and high temperature region. In low temperature region, the polymers show a glass-like state and the electron avalanche breakdown is supposed to be dominant, on the basis of the results of breakdown time lag, the thickness dependence of breakdown and the effects of impurity etc. The main mechanisms in middle high temperature region where the polymers show a rubber-like state, are considered to be thermal and free volume breakdown. In higher temperature region, the polymers are in melting or softening states and the breakdown characteristics were explained by thermal and electromechanical breakdown. (ii) Applied voltage frequency [16,17]: charge carriers in practical polymers are trapped in localized electronic states and a higher frequency voltage can yield higher electrical conductivity resulting in lower breakdown strength when thermal breakdown takes place. (iii) Structure and morphology: over the last three decades, many speculations have been put forward to explain the relationship between structure and morphology of semicrystalline polymers and their dielectric strength [6,19–25]. The majority of authors suppose that the damage of semicrystalline polymers in a high electric field is related to macroscopic inhomogeneity of structure [26]. However, clearly there exist some controversies in understanding the effects of the supermolecular structure on the breakdown strength: (i) it has been shown by Ceres and Schultz that the electrical lifetime of polypropylene decreased with increasing spherulite size [25]; (ii) results of Ishida and Okamoto from cross-linked polyethylene (XLPE) specimens indicated that the improvement in dielectric strength was probably due to the increase in spherulite size [21]; (iii) Vaughan et al. mentioned that no clear trends emerged to indicate that sperulite size variations were reflected in the short-term electrical failure processes of a series of LDPE systems [6]. In the case of the effects of lamellar thickness and crystallinity on the dielectric breakdown strength, however, it seemed that different authors have the similar conclusions: (i) a semicrystalline morphology consisting of extensive thick lamellae produces the highest dielectric strength results; (ii) high degree of crystallinity can result in improvement in dielectric strength of polyolefin.

2173

(iv) Fillers: the introduction of fillers can significantly modify the dielectric strength of polymers depending on the electrical characteristic of fillers [15,27,28]. They appear in polymer matrix as a separate phase, and in this case the dielectric strength of the polymer is determined by the dielectric constant or electrical conductivity of the fillers because the local field distortion and enhancement around the fillers are caused by the difference in dielectric constant or electrical conductivity between the fillers and the base polymer. (v) Defects like voids and pores: experiment [28,29] and simulation [30] results have confirmed that the breakdown strength of solid samples is related to the void introduced in the sample preparation process. The existence of voids in insulators produces local electrical failures, resulting in partial discharges and thereby reducing the electrical breakdown strength of the insulating material and the lifetime of the insulation apparatuses [29]. In a quasi-homogeneous field configuration, the dielectric breakdown of solid polymers is usually controlled by the presence of major flaws such as fillers and voids found within the polymer matrix. The molecular structure of an octavinyl-POSS is shown in Fig. 1. Octavinyl-POSS are selected because that they contain C@C bonds, which can offer an opportunity to tailor the morphology and properties of LDPE/POSS composites by physical blending or reactive blending. In this work, we focused on the morphology and ac electrical properties of LDPE/octavinyl-POSS composites prepared with physical blending. It was well understood from previous work that, for polymer/particle systems, the particle–particle interaction, particle–polymer interaction and the compatibility between polymer matrix and particles can be sensitively reflected by the rheological characterization [31]. On the other hand, polyethylene is a semicrystalline polymer which consists of crystallites and disordered regions between crystallites [32,33]. The electrical properties of the polyethylene could be significantly influenced by its semicrystalline nature and rheological characteristics. In addition, the incorporation of fillers can result in changes in degree of crystallinity, supermolecular structure and crystalline phases [33]. Therefore, the morphological, rheological and thermal characteristics were also investigated in order to understand the structure–dielectric property relationships for the composite systems.

Fig. 1. Chemical structure of an octavinyl-POSS molecule.

MACROMOLECULAR NANOTECHNOLOGY

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

2174

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

2. Experimental 2.1. Materials The low density polyethylene with Ziegler-Natta catalysts was purchased from ExxonMobil in Saudi Arabia, with a Melt Flow Index (MFI) of 2.0 g/10 min by ASTM D1238 and a density of 0.9225 g/cm3 by ASTM D4703/ D1505. The octavinyl POSS, Grade OL1160 was purchased from Hybrid Plastic, USA.

MACROMOLECULAR NANOTECHNOLOGY

2.2. Preparation of LDPE/POSS composites It has been shown from our previous work that, solution compounding is an effective method to prepare nano-filler/ polymer composites [34,35]. Therefore, the composite masterbatch containing 10 wt% POSS was firstly prepared by solution compounding. Briefly, the POSS were dissolved in trichloromethane (CHCl3). At the same time, LDPE was dissolved in chemically pure para-xylene at 368 K. Then, the solution of POSS/CHCl3 was added to the solution of LDPE/para-xylene, and the resulting mixture was stirred with a electromagnetic stirrer at 368 K for 2 h. Afterward, the composites were prepared by precipitation of the mixed solution into methanol, and then the precipitate composites were dried at 403 K in vacuum for 10 h. Films with a thickness of around 250 ± 10 lm were obtained using compression molding at 140 °C under a pressure of about 10 MPa. Then, the film in the mould was transferred to another compression molding machine and the temperature of the film in the mould was slowly decreased to room temperature under the condition of the pressure of 10 MPa. Next, the film was removed from the mould. Poly(ethylene terephthalate) films were used as backing. For the sake of convenience, the composites were denoted using the following notation: PE–POSS weight, thus PE-1.0 indicates the composite with 1.0 wt% POSS. 2.3. Characterization 2.3.1. SEM and TEM The dispersion state of POSS in the surface of masterbatch sample was observed by a Hitachi X-650 scanning electron microscopy (SEM). For morphological investigation of LDPE, cylindrical samples were broken in liquid nitrogen and then the cross sections were etched at room temperature for 4 h in a 1% w/v solution of potassium permanganate in 5 parts concentrated sulphuric acid to 2 parts orthophosphoric acid to 1 part water. The etched cross sections of the cryo-fractured samples were observed using a JEOL JEM-7401 field emission SEM. A JEOL JEM-2100F high-resolution transmission electron microscopy (TEM) was used to observe the dispersion of POSS in the composites. The specimens for TEM observation were trimmed using a microtome machine, and the thickness of the specimens was about 50–100 nm. 2.3.2. XRD The wide-angle X-ray diffraction (XRD) patterns were recorded on a Rigaku D/MAX-2200/PC automatic diffrac-

tometer. All measurements were performed at the atmospheric pressure and room temperature with nickelfiltered Cu target Ka radiation at 40 kV and 20 mA with a scintillation counter system. Data were recorded in the range of 2h = 10–35° at the scanning rate and the step size of 6.0°/min and 0.02°, respectively. 2.3.3. Differential scanning calorimetry (DSC) The crystallization and melting behaviors were analyzed using a Perkin-Elmer Pyris-1 DSC in a dry N2 atmosphere, and all the samples were of the same shape and size and accurately weighted (5–8 mg). The samples were heated from 20 to 150 °C at a heating rate of 10 °C/min. The instrument was calibrated with an Indium standard. Based on the DSC melting curves, the degree of crystallinity X c was determined with the following equation:

X DSC ¼ C

DH f DHm ð1  /Þ

ð1Þ

where DHf is the heat of fusion obtained from integrating the area under the normalized melting curves and DHm is the enthalpy corresponding to the melting of a 100% crystalline sample and DHm ¼ 293 J/g for LDPE. / is the weight fraction of POSS in the base polymer. 2.3.4. Rheological measurements A controlled strain rotational rheometer, Gemini 200HR (Bohlin instruments, UK) with a parallel-plate geometry (diameter of 25 mm), was used for rheological measurements. Tests were carried out at 140 °C with a gap of 0.85–0.95 mm. Strain sweep tests were firstly performed from the initial strain value of 0.1% to a final strain value of 100% with the frequency of 1 rad/s to determine the linear viscoelastic region of the samples. For frequency sweep tests, a small amplitude oscillatory shear was performed in the frequency range 0.1–100 rad/s. A strain of 10% was used, which was in the linear viscoelastic regime for all samples. Storage melt viscosity (g0 , real part of complex viscosity) and loss melt viscosity (g00 , imaginary part of complex viscosity) were measured and used for Cole–Cole plots. Dynamic storage modulus (G0 ) versus dynamic loss modulus (G00 ) were also measured and used for Han–Chuang plots. 2.3.5. Dielectric measurements Dielectric constant and dielectric loss measurements in the frequency range of 500–1 MHz were performed using an impedance analyzer (Aglient 4294A) with 16451B Dielectric Test Fixture at room temperature, applying 1.0 V ac voltage across two opposite sides of the plane samples. The data of impedance and phase angle measured were converted into the relative dielectric constant and dielectric loss, considering the appropriate geometric coefficient. Thin films with thickness of around 250 ± 10 lm were used for dielectric measurements. Gold electrodes were evaporated on the front and rear surfaces of the samples. The sample is considered as a plane capacitor and described by a parallel resistor–capacitor circuit system. Dielectric constant and dielectric loss measurements at 50 Hz was carried out on a high-voltage Schering Bridge

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

er ¼ C0 ¼

CX C0

e0 A t

ð2Þ

¼

e0 p 4

ðd1 þ gÞ2

ð3Þ

where d1 is the diameter of front electrode of the specimen (50 mm in this case). C 0 is the capacity with a free space between the electrodes and g the guard gap (1 mm). For 50 Hz dielectric measurements, gold electrodes were evaporated on the front and rear surfaces of the samples (a front electrode, 50 cm in diameter, surrounded by a guard ring, and a back electrode deposited on the whole rear surface). Dielectric breakdown strength was measured using an AHDZ-10/100 ac dielectric strength tester (Shanghai Lanpotronics Corp., China) according to ASTM D 149-2004. The specimens were placed between two 10-mm-diameter copper ball electrodes and the electrode system containing the measured sample was immersed in the pure silicon oil in order to prevent the surface flashover. The test setup and electrode configuration are presented in Fig. 2. A 50Hz ramp voltage was applied across two ball-typed electrodes and was increased with a rate of 2 kV/s until the sample was punctured. Forty breakdown tests were repeatedly performed on each specimen. Under a quasi-homogeneous field (sphere–sphere electrode configuration), the field to breakdown mainly depends on the maximum voltage and the electrode gap (specimen thickness).

E¼a

V d

ð4Þ

where a is the non-uniformity coefficient of the electric field. It is a function of geometry of the electrodes and could be defined as 1 in this case. Thus the dielectric breakdown strength measured in this study is just the quotient of the breakdown voltage and effective thickness of the specimens [36]. Owing to both structural and technological differences, breakdown strength of apparently identical insulating specimens varies from one to another in a random way. Thus, it is necessary to use statistic methods to treat and obtain the characteristic values of dielectric strength. We treated the measurement data using two-parameter Weibull statistical distribution method. The Weibull statistical distribution in the case of ramp voltage test can be written as

"   # b Ei E0

P ¼ 1  exp 

ð5Þ

where Ei is an experimental breakdown strength, P is the cumulative probability of electrical failure, b is the shape parameter which is related to the scatter of the data (inverse function of the variation of the data), E0 is the characteristic dielectric breakdown that represents the breakdown strength at the cumulative failure probability of 63.2%, the scale parameter, which is often used to compare the dielectric breakdown of various samples with one another. Commonly, the parameter is used to compare differences in dielectric strength among the specimens. Rearranging and taking logarithms twice the above equation can be represented in the following form

log½ lnð1  PÞ ¼ b log Ei  b log E0

ð6Þ

According to the recommendation of the IEEE 930-2004 standards, a good, simple, approximation for the most likely probability of failure is presented in the equation

Pi ¼

i  0:44  100% n þ 0:25

ð7Þ

where i is the i-th result when the values of the E are sorted in ascending order and n is the number of specimens; for this study, n = 40. 3. Results and discussion 3.1. Dispersion

Fig. 2. Electrode configurations used to measure dielectric breakdown.

Dispersion of POSS in the surface of the masterbatch sample is shown in Fig. 3(a). It is observed that POSS is well distributed and exists in the form of crystal particles with the size of 0–500 nm. On the other hand, a small number of POSS particles of 100–500 nm in size could be seen in the composites (Fig. 3b), indicating that crystallization of POSS molecules in LDPE matrix does occur. These results can be understood if one takes into account the fact that POSS is a highly crystalline material. The nature of the crystal like particles as POSS was verified with energy dispersive Xray spectroscopy (EDX) analyses, as shown in the spectrum of middle panel of Fig. 3(c), where the signals from the C matrix, the Si from the POSS and Au from the metallization are marked.

MACROMOLECULAR NANOTECHNOLOGY

(TETTEX AG Instrument, Switzerland) according to ASTM D 150-2004, applying 1000 V/mm ac field across two opposite sides of the plane samples. Disk electrodes with guard-ring were used and dielectric constant er can be determined from measured capacitance data C X by employing the following relations:

2175

2176

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

MACROMOLECULAR NANOTECHNOLOGY

and the diffraction peak positions of the neat LDPE agree well with the values of unit cell (orthorhombic) parameters to be a = 4.14 Å, b = 7.47 Å, c = 2.48 Å, reported in other literatures [37]. This indicates that the POSS particles do not alter or distort the crystal structure of LDPE. DSC measurement results are shown in Fig. 5. It is observed that the values of initial and final melting temperatures of the LDPE/POSS composites are almost the same, suggesting similar size and size distribution of the LDPE crystals according to Gibbs-Thomson equation [38]. Crystallinity values of LDPE obtained from DSC curves were 41.6 ± 0.3%, 39.6 ± 1%, 40.4 ± 0.5%, 40.6 ± 0.4%, 40.5 ± 0.6% and 40.2 ± 1% for the neat LDPE, PE-0.5, PE-1.0, PE-2.0, PE-4.0 and PE-6.0, respectively, which revealed that there was no significant influence of POSS on the crystallinity of LDPE. The supermolecular structure of LDPE with and without POSS was investigated using an etching method. The supermolecular structure is related to the arrangements of the individual lamellar crystallites into a larger scale of organization [38]. Fig. 6 shows the etched external surface of the neat LDPE and PE-6.0. It can be seen that the morphology of the superstructure of the LDPE was changed with the addition of POSS. The neat LDPE exhibited a well – defined spherulite texture (Fig. 6a) and the diameter of the spherulite from FE-SEM is about 5 lm. The growing front and curved individual lamellae or lamellae bundles can be observed. The composites, however, displayed the severely disrupted spherulites features (Fig. 6b), and the lamella bundles were arranged in such a disordered way that it was impossible to resolve individual spherulites. Although, external surface morphology observation enables direct visualization of the polymer superstructure, it only yields surface texture information. Therefore, FESEM micrographs of an etched cross section of a cryofractured sample, which provide bulk information, were presented to supply the structure investigation. Fig. 7 shows the FE-SEM micrographs of the etched cross sections of cryo-fractured samples for the neat LDPE and PE-6.0. As shown in Fig. 7, spherulites in the range of 3–5 lm could be seen for the neat LDPE though some of

Fig. 3. SEM micrograph and EDX spectrum showing dispersion of POSS in the surface of the masterbatch sample (a and c) and TEM micrograph showing dispersion of POSS inside of PE-6.0 (b). Arrows indicate POSS crystals.

3.2. Crystal structure and morphology The wide-angle X-ray diffraction (WAXD) measurements between 5° and 35° were carried out to examine the unit cell dimension and crystal size of LDPE. As shown in Fig. 4, the observed crystalline peak positions for all composite samples are close to those of the neat LDPE

Fig. 4. WAXD curves of POSS and LDPE/POSS composites.

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

2177

Fig. 7. Micrographs showing the cryo-fractured surface morphologies of neat LDPE (a) and PE-6.0 (b) observed after etching. Circles indicate spherulites.

Fig. 6. Micrographs showing the external surface morphologies of neat LDPE (a) and PE-6.0 (b) observed after etching.

them exhibit coarser texture. However, the composite sample shows more homogeneous texture and it was hard to resolve individual spherulites or lamellar-bundles in the observed areas, indicating the presence of POSS crystals

remarkably deteriorated the spherulitic organization of the LDPE. These results are consistent with that obtained from the external surface morphology observations. Such a study also demonstrated that the external surface morphology observations can be used to obtain valuable information on the internal organization of the spherulites for LDPE. This finding is significant because that a complicated sample preparation process is usually needed for observing the internal organization of polymers using SEM or other apparatus. Taking the aforementioned results into account, it could be concluded that the presence of POSS did not alter the unit cell dimensions and the lamellar thickness of LDPE. In addition, the crystallinity did not change in the composites. However, the polymer morphology, evidenced by FE-SEM, did change significantly in the composites. These results are consistent with previous researches. Ma reported that the presence of the TiO2 nanoparticles, with the various surface conditions, did not alter the degree of LDPE crystallinity, the unit cell dimensions or the average lamellar thickness. However, the nanoparticles did affect the internal arrangement of intraspherulitic crystalline aggregates [37].

MACROMOLECULAR NANOTECHNOLOGY

Fig. 5. DSC melting curves of LDPE and LDPE/POSS composites.

2178

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

MACROMOLECULAR NANOTECHNOLOGY

3.3. Rheological and compatibility analysis The rheological properties of polymer/particle composites have both practical importance related to composite processing and scientific importance as a probe of the composite microstructure. The complex viscosity (g ) of LDPE/ POSS composites at 140 °C is measured and Fig. 8 presents the influence of POSS loadings on complex viscosity of the composites. The g reaches minimum at 2.0 wt% POSS loading level at both low frequency (x = 0.1 rad/s) and high frequency (x = 10 rad/s). It is proposed that there could be two competitive factors to affect the rheological behavior of the LDPE/POSS composites: POSS–POSS interaction and POSS–polymer interaction. At low POSS loading levels (0.5–2.0 wt%), the POSS–POSS interaction (attractive van der Waals force) is relatively weak because of the long POSS-to-POSS distance, which may lead to a fine dispersion of POSS particles in the LDPE matrix. When the POSS loading level is very low (below about 0.5%), the effective interface area between the POSS and LDPE is so small that interaction between LDPE chains and POSS particles can be neglected and hence no significant change of g is found. The further increase of POSS loadings causes the increase of effective interface area between POSS and LDPE. Therefore, interaction between LDPE chains and POSS particles tends to further and further increase as the loading level becomes higher. This interaction probably causes chain disentanglements and more free volume in the melt, as the POSS acts as a lubricant of LDPE, which is consistent with the rheological properties of octamethyl POSS/HDPE [39] and octavinyl POSS/PP [40] composites. The lubrication effect makes the values of g for the composites tend to decrease with the increasing POSS loadings. At relatively higher loading levels, e.g., over about 2.0%, the POSS–POSS distance decreases exponentially from a purely geometric consideration [41] and the more attractive forces begin to cause the significant agglomeration of POSS units with further increasing of POSS, as shown in Fig. 3b. Such small POSS–crystal agglomerates can undoubtedly induce a steric barrier or act as physical cross-linking, increasing the resistance to polymer chain motion. Additionally, the sig-

Fig. 8. Influence of POSS loading levels on complex viscosity of LDPE/ POSS composites at 140 °C.

nificant agglomeration of POSS molecules can also result in a reduction of effective interface area between POSS and LDPE, reducing the free volume in the composites. That is, the variation of viscosity for the composites is the net result of both of the effects: free volume and steric barrier. Starting at 2% of POSS concentration, the steric barrier effect of POSS may play a dominant role in determining the viscosity and thus raise the complex viscosity of the composites. The compatibility was further investigated by Cole–Cole plots [42] and Han–Chuang plots [43,44]. Cole–Cole plots represent storage melt viscosity (g0 ) versus loss melt viscosity (g00 ) of the blends as well as the pure resins on a linear scale. If it forms a semicircular relationship, then the blend is considered to be compatible. The Cole–Cole plots of the composites are shown in Fig. 9a. It is observed that the curves for each composition are close to semicircular shapes, indicating all the composites are compatible. Han–Chuang plots represent the log dynamic storage modulus (G0 ) versus the log dynamic loss modulus (G00 ). In this method, if a blend is compatible, the same slope is observed between the blend compositions and the pure components; otherwise, it is considered to be an incompatible or phase separated blend. As can seen from Fig. 9b any concentration dependency of G0 versus G00 for LDPE/POSS composites could not be found. According to the previously

Fig. 9. Cole–Cole plots of LDPE/POSS composites (a) and Han–Chuang plots of PE-6.0 (b).

proposed criteria by Han and Chuang, this observation means that the LDPE/POSS systems are compatible. 3.4. Dielectric parameters The frequency dependence of dielectric constant and dielectric loss for the LDPE/POSS composites at different loading levels was shown in Fig. 10. It can be seen from Fig. 10 that the effective dielectric constant of the neat LDPE and the LDPE/POSS composites is very weakly dependent on frequency, which is the typical characteristics of non-polar polymers. Although, dielectric constant is a frequency dependent parameter in polymer systems, the neat LDPE contains no dipolar units and thus do not show apparent frequency characteristics in the range of 500– 1  106 Hz. For polymer composites with fillers, electrical properties of the fillers and the polymer matrix usually are known to be considerably different from each other, especially conductivity, which can lead to the formation of electric charge layers at the interfaces, that is, the interfacial polarization when voltage is applied to the materials. The interfacial polarization can cause an increase of dielectric constant and dielectric loss in polymer composites when compared with the neat polymers. Therefore, the

Fig. 10. Frequency dependence of dielectric constant and dielectric loss for LDPE/POSS composites with different POSS loading levels.

2179

above observations suggest that interfacial polarization does not occur in the LDPE/POSS composites. This is possible due to two factors: one is that the dielectric constant measurements are all performed at such an low applied voltage of 1 V (4 V/mm) for durations <3 min that interfacial polarization could not occur at conductivity barriers in composites bulk for the present range of measurement frequencies; Another might be ascribed to the low electrical conductivity of the POSS bulk which is comparable to that of the neat LDPE. It is observed from Fig. 10 that at POSS loading levels lower than 2.0%, the dielectric constant values of the composites at all frequencies are lower than those of the neat LDPE, but at loading levels higher than 4.0%, opposite results are observed. The occurrences of lower dielectric constant in polymer nanocomposites have been reported by several research groups [45–47]. Dielectric constant usually increases when polymers are filled with inorganic fillers, and the reasons have been attributed to the higher dielectric constant of fillers by nature than the polymer matrix and interfacial polarization. However, lower values of dielectric constant of composites than that of the neat polymer matrix were really observed in some polymer nanocomposites such as epoxy/TiO2 and epoxy/ZnO, in which not only the fillers have much higher dielectric constant than the base polymer, but also the interfacial polarization does occur [45]. Therefore, plausible mechanisms should be raised to explain the lower dielectric constant in the composites. Firstly, it is proposed that free volume increase may cause the reduction of dielectric constant. Our rheological analyses have given us a hint that at POSS loading levels lower than 2.0%, the POSS could introduce more free volume. Increases in free volume can be attributed to the greater steric volume of POSS molecules, which may interfere with efficient chain packing. A correlation of high free volume with low dielectric constant has previously been found in several polymers [48]. It is also found from Fig. 10 that only those with low loadings show lower dielectric constant, which is consistent with the concentration dependence of complex viscosity in the composites. The restriction of macromolecular chain movement by POSS particles may be another reason for the reduction of dielectric constant. For polar polymers such as epoxy resin, the polymer chain immobility could be thought one of the important reasons for the reduction of dielectric constant because that the mobility of dipolar groups in epoxy resin contributes to the dielectric constant of the composites [45]. For non-polar polymer such as polyethylene, however, the effect of polymer chain immobility on dielectric constant may be secondary. The dielectric loss of the neat LDPE and the LDPE/POSS show the increasing tendency with the increase of the frequency because a higher frequency voltage can yield higher electrical conductivity (Fig. 10). Unlike the loading dependence of dielectric constant, a clear correlation between the POSS loadings and dielectric loss can be stated, that is, the dielectric loss of the composites increases with POSS loading. And also, the dielectric loss of the composites is less than that of the neat LDPE except PE-6.0. It is proposed that there could be two competitive factors to affect the dielectric loss of LDPE/POSS composites depending

MACROMOLECULAR NANOTECHNOLOGY

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

MACROMOLECULAR NANOTECHNOLOGY

2180

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

on the combination of the hindrance in charge transport and the incorporation of charge. At low POSS loading levels, the incorporation of POSS can result in a large volume fraction of interfaces and polymer chain entanglement because of the good dispersion of POSS, which in turn cause immobility of charge carriers or reduction in electrical conductivity, and thus causing a reduction of dielectric loss. With the loadings increasing, there is an enhancement of charge carriers contributing the dielectric loss and in consequence the introduction of POSS starts to cause an increase of dielectric loss in the composites. At the highest loading level of 6.0%, as mentioned earlier, the agglomeration of POSS molecules can also result in an apparent reduction of effective interface area between POSS and LDPE, and thus the effect of immobility of charge carriers caused by POSS on reduction in electrical conductivity is far less important than the influence of charge carriers introduced by POSS loading, therefore, composite PE-6.0 shows higher dielectric loss than that of the neat LDPE in the whole frequency range. For dielectrics used in high-voltage electric power equipments, the aforementioned dielectric data of LDPE/ POSS composites could not present valuable information for evaluating the effects of POSS on dielectric loss of the LDPE composites. Not only is the applying voltage for dielectric loss measurements so low when compared with the operating voltage of the electric power equipments, but also the dielectric loss information under power frequency used for breakdown strength measurements was not involved. Therefore, it is necessary to provide the power frequency data obtained from relatively high-field measurements. Fig. 11 shows the power-frequency dielectric data versus POSS loading. It can be seen that the dielectric loss and ac conductivity increase gradually with POSS loading level up to 1.0%. Starting at 1.0%, any variation could not be found. These results are different from the observations in Fig. 10 and can be understood according to the difference between the voltages used for frequency dependence dielectric parameter and power-frequency dielectric measurements. Power-frequency dielectric data were obtained under the higher electric field, in this case there is an enhancement of charge carriers contributing the electrical

Fig. 11. Dielectric measurement results using high voltage bridge (50 Hz, applied voltage 1000 V/mm).

conduction and in consequence the introduction of POSS causes increase of dielectric loss and electrical conductivity in the composites. 3.5. Dielectric breakdown strength The breakdown strength data of all samples are shown in Fig. 12 and the Weibull parameters are listed in Table 1. As shown in Fig. 12 and Table 1, the values of characteristic breakdown strength of the composites decrease gradually with an increase in POSS loading. However, all the composites show higher b values than that of the neat LDPE. And also, it was observed from Fig. 12 that the breakdown strength of the neat LDPE at lower cumulative failure probability shows lower values than those of the composites and the neat LDPE shows the lowest dielectric strength in the minimum cumulative failure probability. This observation is of vital importance for engineering application because the dielectric rupture always occurs at the weakest points, in other words, the real dielectric strength of products such as wire and cables is determined by the weakest part of their insulation. In practice, the dielectric breakdown probably does not occur at the cumulative failure probability of 63.2% but at lower values. Firstly, we discussed the effects of structure and morphology on the dielectric strength of LDPE. As mentioned earlier, the lamellar thickness, lamellar thickness distribution and crystallinity of LDPE are not modified by POSS and are almost independent of the POSS loading levels. Therefore, only the effects of changes of supermolecular structure on the dielectric strength are considered here. Previous researches indicated that [18], breakdown of polyethylene occurs more often at the spherulite boundaries, especially at the triple point of boundaries, than inside the spherulites. Thus one of the effective methods for the decrease of scatter for breakdown data is to reinforce weak amorphous parts of crystalline boundaries by modification of polyethylene, in other words, to homogenize the morphology of polyethylene. Taking these considerations and the aforementioned morphological observations into account, the higher values of b and

Fig. 12. Weibull plots of the dielectric strength for LDPE and LDPE/POSS composites.

Table 1 Weibull parameters for LDPE and LDPE/POSS composites. Sample

E0 (kV/mm)

b (100% > P > 0)

b (100% > P > 10%)

LDPE PE-0.5 PE-1.0 PE-2.0 PE-4.0 PE-6.0

172.0 ± 1.3 162.8 ± 0.6 159.4 ± 0.8 158.0 ± 0.6 139.8 ± 0.8 136.5 ± 0.7

5.3 ± 0.2 9.8 ± 0.2 7.7 ± 0.3 10.4 ± 0.3 8.3 ± 0.3 10.5 ± 0.4

5.8 ± 0.2 9.8 ± 0.3 7.9 ± 0.2 10.0 ± 0.2 7.4 ± 0.2 9.3 ± 0.3

breakdown strength of LDPE/POSS composites at low accumulated failure probability should be attributed to the more homogeneous morphology of the composites when compared with the neat LDPE, as shown in Fig. 7. Our previous work indicated that gamma irradiation could result in more homogeneous morphology and in consequence cause a significant increase in the slope of the Weibull statistical plots of breakdown strength–breakdown probability for the XLPE cable insulation after irradiation [49]. POSS molecules can be thought as the smallest particles of silica with relatively low dielectric constant. Therefore, in our case, the POSS themselves might not act as local electric field enhancing defect centers and the modification of dielectric strength, e.g., the decrease of characteristic dielectric strength of LDPE/POSS composites, could not be explained by the electric field distortion in the presence of POSS. In fact, there exist several polyethylene/silica composites with dielectric strength much higher than that of the polymer matrix [46]. In the temperature region (285 ± 2 K) where our measurements were taken, the breakdown mechanism of LDPE should be partly related to free volume breakdown. In the free volume breakdown theory set up by Artbauer, breakdown is initiated by free electrons accelerated by electric field in the largest of the holes present in the amorphous phase of all polymers [50]. Some additives, e.g., plasticizer, however, tend to cause a decrease of dielectric strength as the plasticizer increases the free volume of polymers, which can lead to an increase of electron mean free path and cause early breakdown [51]. As mentioned earlier, the plasticization effect of POSS for LDPE was clearly observed. The increase of the POSS concentration effectively increase the distance between adjacent polymer chains and lead to a further increase of the mean size of local free volume as the POSS loading becomes higher, which may be one of the reasons resulting in a gradual decrease in dielectric strength of LDPE with POSS loading. It should be noted that, as reported in previous work, some plasticizers could significantly increase the dielectric strength of their composites though they show apparent lubrication effect [52]. This phenomenon could be related to the following mechanism: charge scattering or trapping by the functional groups belonging to the plasticizers and excitation effects related to the plasticizers. For POSS fillers used in this study, it is considered that they do not have these effects, therefore, the increase of characteristic dielectric strength for the composites was not observed when compared to the neat LDPE, which gave us a hint that choosing POSS with special functional groups might be an effective approach to prepare POSS/polymer composites with im-

2181

proved electrical properties, e.g., characteristics dielectric strength. The dimensions of free volume are of molecular order, while voids, even the smallest, are orders of magnitude larger. In a quasi-homogeneous field configuration, the dielectric breakdown is usually controlled by the presence of major flaws such as fillers and voids found within the polymer matrix. The influence of POSS themselves on the breakdown of the LDPE has been discussed earlier, and it is considered that the POSS units themselves only show a very small effect, indicating that the dielectric breakdown of the LDPE may be controlled by the presence of the voids besides the aforementioned free volume. The incorporation of POSS can introduce voids and their volume fraction may increase with the POSS loading because of the POSS particles tend to form agglomerates (Fig. 3b). Therefore, the increase of voids may be another reason causing the decrease of dielectric strength of LDPE with POSS loading. Starting at 4.0%, further increase of POSS concentration can cause the increase of both the number and size of POSS agglomerates, which in return can introduce larger voids and pores because of the strong re-agglomeration tendency of POSS small crystals. Therefore, we observed an apparent decrease of characteristic breakdown strength of PE-4.0 when compared with PE-2.0. The ac breakdown strength is also associated with thermal heating due to dielectric loss and electrical conduction [16,17]. According to the observations in Fig. 10, the dielectric strength should have the same level in the POSS loading range of 1–6%, since the composites with POSS loading higher than 1.0% almost have the same dielectric loss and ac conductivity. The dielectric strength, however, gradually decreases with POSS loadings. Therefore, this results shown in Fig. 10 exclude the possibility of the ordinary thermal breakdown by dielectric loss heating and electrical conductivity heating. The factors involving applied voltage frequency, environment and electrode conditions would not be considered here since they were not changed during the dielectric strength measurements. 4. Conclusions In this study, LDPE/POSS composites are prepared with masterbatch using melt mixing. We have investigated the morphology and property performances of the composites by several characterization techniques. Rheological analyses of the composite indicated that octavinyl-POSS had a good compatibility with LDPE and could play a lubricant role in LDPE matrix. The frequency dependence of dielectric parameters of the LDPE/POSS composites showed unusual behaviors when compared with conventional polymer/filler systems at a very low voltage (4 V/mm), the dielectric constant and dielectric loss in the composites with low POSS concentration are found to be lower than those of the neat LDPE. When the applied voltage was high (1000 V/mm), however, the measured dielectric constant and loss of all the composites by a Schering Bridge show marginally higher values than those of the neat LDPE. The ac breakdown strength of the neat LDPE and LDPE/POSS composites was

MACROMOLECULAR NANOTECHNOLOGY

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

2182

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

analyzed according to Weibull statistic distribution method. Although, the characteristics dielectric breakdown of LDPE/POSS composites decreased gradually with POSS concentration, the breakdown strength of the LDPE/POSS composites at lower cumulative failure probability showed higher values than those of the neat LDPE which shows the lowest dielectric strength in the minimum cumulative failure probability. Our study concluded that LDPE/POSS composites have potential engineering application because that the real dielectric strength of products such as wire and cables is determined by the weakest part of their insulation. It was found from the morphological observations that incorporation of the POSS particles could not result in significant alterations in lamellar thickness and thermal characteristics of LDPE, whereas the supermolecular structure of LDPE was distorted by POSS particles.

MACROMOLECULAR NANOTECHNOLOGY

Acknowledgement The authors are grateful for the financial supports of the National Science Foundation of China (Project No. 5067 7037). References [1] Tiemblo P, Hoyos M, Gomez-Elvira JM, Guzman J, Garcia N, Dardano A, et al. The development of electrical treeing in LDPE and its nanocomposites with spherical silica and fibrous and laminar silicates. J Phys D Appl Phys 2008;41:125208. [2] Takala M, Karttunen M, Salovaara P, Kortet S, Kannus K, Kalliohaka T. Dielectric properties of nanostructured polypropylene-polyhedral oligomeric silsesquioxane compounds. IEEE Trans Dielectr Electr Insul 2008;15:40–51. [3] Lisunova MO, Mamunya YP, Lebovka NI, Melezhyk AV. Percolation behaviour of ultrahigh molecular weight polyethylene/multi-walled carbon nanotubes composites. Euro Polym J 2007;43:949–58. [4] Smith RC, Liang C, Landry M, Nelson JK, Schadler LS. The mechanisms leading to the useful electrical properties of polymer nanodielectrics. IEEE Trans Dielectr Electr Insul 2008;15:187–96. [5] Hoyos M, Garcia N, Navarro R, Dardano A, Ratto A, Guastavino F, et al. Electrical strength in ramp voltage AC tests of LDPE and its nanocomposites with silica and fibrous and laminar silicates. J Polym Sci B Polym Phys 2008;46:1301–11. [6] Green CD, Vaughan AS, Mitchell GR, Liu T. Structure property relationships in polyethylene/montmorillonite nanodielectrics. IEEE Trans Dielectr Electr Insul 2008;15:134–43. [7] Li GZ, Pittman Jr CU. Polyhedral oligomeric silsesquioxane (POSS) polymers, copolymers, resin nanocomposites. In: Abd-El-Aziz AS, Carraher Jr CE, Pittman Jr CU, Zeldin M, editors. Macromolecules containing metal, metal-like elements, vol. 4. Wiley; 2005. p. 79–131. [8] Liu YL, Chang GP, Hsu KY, Chang FC. Epoxy/polyhedral oligomeric silsesquioxane nanocomposites from octakis(glycidyidimethylsiloxy)octasilsesquioxane and small-molecule curing agents. J Polym Sci A Polym Chem 2006;44:3825–35. [9] Misra R, Fu BX, Morgan SE. Surface energetics, dispersion, and nanotribomechanical behavior of POSS/PP hybrid nanocomposites. J Polym Sci B Polym Phys 2007;45:2441–55. [10] Wang YZ, Chen WY, Yang CC, Lin CL, Chang FC. Novel epoxy nanocomposite of low D-k introduced fluorine-containing POSS structure. J Polym Sci B Polym Phys 2007;45:502–10. [11] Zhou Z, Cui LM, Zhang Y, Zhang YX. Yin NAW: preparation and properties of DOSS grafted polypropylene by reactive blending. Euro Polym J 2008;44:3057–66. [12] Bourbigot S, Duquesne S, Jama C. Polymer nanocomposites: how to reach low flammability? Macromol Symp 2006;233:180–90. [13] Nelson JK. Breakdown strength of solids chapter 5. In: Bartnikas R, Eichhorn RM, editors. Engineering dielectrics, vol. 2A. Philadelphia PA: ASTM Press; 1983. [14] Dissado LA, Fothergill JC. Electrical degradation, breakdown in polymers. London: Peter Peregrinus; 1992.

[15] Cruz SA, Zanin M. Assessment of dielectric behavior of recycled/ virgin high density polyethylene blends. IEEE Trans Dielectr Electr Insul 2004;11:855–60. [16] Miyairi K. Frequency dependent dielectric breakdown of thin polyimide films prepared by vapor deposition polymerization. Jpn J Appl Phys 2001;40:1297–9. [17] Miyairi K. Frequency-dependent dielectric breakdown in thin polyvinylcarbazole films. Jpn J Appl Phys 2003;42:5153–7. [18] Ieda M, Nagao M, Hikita M. High-field conduction and breakdown in insulating polymers. Present situation and future prospects. IEEE Trans Dielectr Electr Insul 1994;1:934–45. [19] Vaughan AS, Zhao Y, Barre LL, Sutton SJ, Swingler SG. On additives, morphological evolution and dielectric breakdown in low density polyethylene. Euro Polym J 2003;39:355–65. [20] Vaughan AS, Hosier IL, Dodd SJ, Sutton SJ. On the structure and chemistry of electrical trees in polyethylene. J Phys D Appl Phys 2006;39:962–78. [21] Ishida M, Okamoto T. Effect of surfactants on superstructure and dielectric breakdown strength of crosslinked polyethylene insulation. Electr Eng Jpn 1992;112:21–32. [22] Hosier IL, Vaughan AS, Swingler SG. On the effects of morphology and molecular composition on the electrical strength of polyethylene blends. J Polym Sci B Polym Phys 2000;38:2309–22. [23] Hosier IL, Vaughan AS, Swingler SG. Structure–property relationships in polyethylene blends: the effect of morphology on electrical breakdown strength. J Mater Sci 1997;32:4523–31. [24] Hosier IL, Vaughan AS, Patel D, Sutton SJ, Swingler SG. Morphology and electrical conductivity in polyaniline/polyolefin blends. IEEE Trans Dielectr Electr Insul 2001;8:698–704. [25] Ceres BV, Schultz JM. Dependence of electrical breakdown on spherulite size in isotactic polypropylene. J Appl Polym Sci 1984;29: 4183–97. [26] Gefle OS, Lebedev SM. Effect of morphology, proton irradiation and pre-stressing on the breakdown of LDPE. IEEE Trans Dielectr Electr Insul 2006;13:788–94. [27] Cruz SA, Zanin M. Dielectric strength of the blends of virgin and recycled HDPE. J Appl Polym Sci 2004;91:1730–5. [28] Chen G, Davies AE. The influence of defects on the short-term breakdown characteristics and long-term dc performance of LDPE insulation. IEEE Trans Dielectr Electr Insul 2000;7:401–7. [29] Anderson RA, Lagasse RR, Russick EM, Schroeder JL. Effects of void size and gas content on electrical breakdown in lightweight, mechanically compliant, void-filled dielectrics. J Appl Phys 2002;91:3205–12. [30] Wu K, Okamoto T, Suzuoki Y. Simulation study on the correlation between morphology and electrical breakdown in polyethylene. J Appl Phys 2005;98:114102. [31] Huang XY, Kim CU, Ma ZS, Jiang PK, Yin Y, Li Z. Correlation between rheological, electrical, and microstructure characteristics in polyethylene/aluminum nanocomposites. J Polym Sci B Polym Phys 2008;46:2143–54. [32] Huang XY, Ke QQ, Kim CN, Zhong HF, Wei P, Wang GL, et al. Nonisothermal crystallization behavior and nucleation of LDPE/Al nano- and microcomposites. Polym Eng Sci 2007;47:1052–61. [33] Huang XY, Jiang PK, Kim CN, Duan JK, Wang GL. Atomic force microscopy analysis of morphology of low density polyethylene influenced by Al nano- and microparticles. J Appl Polym Sci 2008;107:2494–9. [34] Huang XY, Jiang PK, Kim CU. Electrical properties of polyethylene/ aluminum nanocomposites. J Appl Phys 2007;102:124103. [35] Huang XY, Kim CU, Jiang PK, Yin Y, Li Z. Influence of aluminum nanoparticle surface treatment on the electrical properties of polyethylene composites. J Appl Phys 2009;105:014105. [36] Vouyovitch L, Alberola ND, Flandin L, Beroual A, Bessede JL. Dielectric breakdown of epoxy-based composites: relative influence of physical and chemical aging. IEEE Trans Dielectr Electr Insul 2006;13:282–92. [37] Ma DL, Akpalu YA, Li Y, Siegel RW, Schadler LS. Effect of titania nanoparticles on the morphology of low density polyethylene. J Polym Sci B Polym Phys 2005;43:488–97. [38] Mark JEM, Ngai K, Graessley W, Mandelkern L, Samulski E, Koenig J, et al., editors. Physical properties of polymers. Cambridge, England: Cambridge University Press; 2004. [39] Joshi M, Butola BS, Simon G, Kukaleva N. Rheological and viscoelastic behavior of HDPE/octamethyl-POSS nanocomposites. Macromolecules 2006;39:1839–49. [40] Zhou ZY, Zhang Y, Zhang YX, Yin NW. Rheological behavior of polypropylene/octavinyl polyhedral oligomeric silsesquioxane composites. J Polym Sci B Polym Phys 2008;46:526–33.

X. Huang et al. / European Polymer Journal 45 (2009) 2172–2183

[47] Roy M, Nelson JK, MacCrone RK, Schadler LS. Candidate mechanisms controlling the electrical characteristics of silica/XLPE nanodielectrics. J Mater Sci 2007;42:3789–99. [48] Tanaka T. Dielectric nanocomposites with insulating properties. IEEE Trans Dielectr Electr Insul 2005;12:914–28. [49] Huang XY, Kim CU, Jiang PK, Yin Y. Effects of high dose gamma irradiation on water treeing deterioration of XLPE cable insulation. IEEE Trans Dielectr Electr Insul, submitted for publication. [50] Artbauer J. Electric strength of polymers. J Phys D Appl Phys 1996;29:446–56. [51] Sabuni MH, Nelson JK. The effects of plasticizer on the electric strength of polystyrene. J Mater Sci 1979;14:2791–6. [52] Katsuyoshi S, Shigetaka F. The effects of plasticizer on the mechanical, electrical characteristics of PLA. In: international symposium on electrical insulating materials. Japan: Yokkaichi City; 2008. pp. 267–70.

MACROMOLECULAR NANOTECHNOLOGY

[41] From a purely geometric consideration, assuming a spherical shape for the POSS particles and a uniform size distribution and dispersion, the interparticle distance, l, is expressed as l ¼ r½ð4p=3/Þ1=3  2 where r is the radius of the POSS particles and / is the volume fraction. [42] Utracki LA. Two-phase polymer systems. Munich, Germany: Hanser Publishers; 1992. [43] Han CD, Chuang HK. Criteria for rheological compatibility of polymer blends. J Appl Polym Sci 1985;30:4431–54. [44] Chuang HK, Han CD. Rheological behavior of polymer blends. J Appl Polym Sci 1984;29:2205–29. [45] Singha S, Thomas MJ. Dielectric properties of epoxy nanocomposites. IEEE Trans Dielectr Electr Insul 2008;15:12–23. [46] Roy M, Nelson JK, Reed CW, MacCrone RK, Keefe RJ, Zenger W, et al. Polymer nanocomposite dielectrics – the role of the interface. IEEE Trans Dielectr Electr Insul 2005;12:629–42.

2183