Erosion Behaviour Of Nano Filled Silicone Elastomers

Proceedings of the XIVth International Symposium on High Voltage Engineering, Tsinghua University, Beijing, China, August 25-29, 2005

C-09

Erosion Behaviour of Nano Filled Silicone Elastomers Stephanie Rätzke*, Josef Kindersberger Lehrstuhl für Hochspannungs- und Anlagentechnik der Technischen Universität München, Deutschland; *E-mail: [email protected] Abstract: The effect of nano sized particles compared to micro sized particles is explained by the high surface to volume ratio. The influence of different types and concentrations of nano and micro sized fillers on the erosion resistance, the thermal conductivity and the dielectric properties of an unfilled silicone elastomer is investigated. The dispersion of the filler particles in the matrix material seems to play a decisive role. Key Words: Nanodielectrics, nanometric fillers, advanced insulation materials, erosion behaviour

Fraction of interfacial layers To illustrate the highly increased fraction of interfacial layers, one has to dwell on the consequences of the smaller particle size. On the one hand a spherical particle with a diameter of 1 µm has approximately the same volume as one billon particles with a diameter of 1 nm (Fig. 1).

INTRODUCTION

1 nm

In polymeric materials for outdoor insulation fillers are used for various reasons, e.g. to improve the mechanical stiffness, the resistance to arcing, resistance to tracking and erosion, etc. The contact area between filler and matrix material is called interfacial layer. The size of typical conventional filler particles is 1 - 150 µm. These filler materials shall further be called µ-filler. The potentials of new nano structured materials, in which the properties of the interface layers dominate the volume effect, which today is called interface effect, were already indicated in 1992 [1]. It was illustrated in 2001 that the interface effect can be used for tailored materials [2], for instance by using nanometric fillers. Nanometric fillers are defined as particles with a size between 1 – 100 nm. In this paper they are called n-fillers or nano particles and the materials consisting of n-fillers in a dielectric medium are called nanodielectrics. Several publications point out in past years the potential of n-fillers [3, 4, 5]. It is remarkable, that the published improvements already occur at a filler loading of 1 – 10 wt. %. By using conventional fillers usually a large amount of filler material (55 – 65 wt. %) is necessary to achieve the required properties [9]. The present work investigates the influence on the resistance to HV-arcing (according to IEC 61621) by using n-fillers and µ-fillers in a silicone elastomer. Furthermore the dielectric properties were measured.

1 µm Fig. 1 Illustration of dielectrics with µ- and n-fillers One can see from this example that as a result of the smaller particle size the absolute number of particles is highly increased at the same volume fraction. Furthermore, with decreasing particle diameter the fraction of surface atoms is increased. In particles with a diameter of 20 nm already 10 % of the atoms in the particle are surface atoms (Table 1). At a particle size of 1 nm nearly all atoms are located at the particle surface. Therefore, nano particles have a highly increased fraction of surface atoms. Particle diameter [nm] 20 10 5 2 1

Number of atoms in a particle 250 000 30 000 4 000 250 30

Fraction of surface atoms in a particle [%] 10 20 40 80 99

Table 1 Relation of particle diameters to contained atoms as well as to the fraction of surface atoms

FUNDAMENTALS OF THE NANO EFFECT

Thus, for a constant filler fraction in the material it is found that with decreasing particle size - the absolute number of atoms in the particle is highly increasing and - a larger fraction of the atoms in the filler particle belong to the interfacial layer. As a result the fraction of the interfacial layers is increased enormously.

In materials with µ-fillers the volume effect is dominating, i.e. due to the high volume fraction the filler material contributes to the properties of the whole material. Contrary to that the properties in nanodielectrics are affected by the interfacial layer between the filler and the matrix material. The reason for that is the highly increased fraction of interfacial layers in the material.

1

Proceedings of the XIVth International Symposium on High Voltage Engineering, Tsinghua University, Beijing, China, August 25-29, 2005

C-09

interfacial layers is large enough, the properties of the whole material are neither the properties of the matrix material A nor the properties of the particle B (Fig. 3).

Physical Fundamentals of the Nano Effect The binding structures in interfacial layers are very different in comparison with the filler or matrix material. Beside the classic physical effects quantum physical effects are also possible. The interfacial layers present a range between macroscopic and quantum mechanical properties. This is evident if one compares the size of a nano particle (1 – 100 nm) with the radius of an aluminium atom (0.25 nm) or the single bond length between two carbon atoms (0,154 nm) [2]. In a homogenous material, like the matrix material or interior of a particle, the atomic constitution is also homogenous. That means every atom is surrounded only by atoms of the same type and therefore the interactions are equal for all directions. But, atoms of the interfacial layer interact with atoms of the matrix material and the filler material likewise. This means that the forces acting to the atoms of the interfacial layer are different from those in a homogeneous material. The interfacial layer can have a thickness of only 1 nm if only strong interaction is acting or can be more than 10 nm thick if charge carriers are situated in the interfacial layer [7, 15]. The effects in the interfacial layer may be explained by a model [7]. The interfacial layer between two different materials defines also a transition of an intensity of a special property (for example concentration or energy) in material A (IA) to another in material B (IB). There must be a continuous transition between the two intensities at the border (Fig. 2). This transition can have different slopes (a) or even a maximum (b) is possible, which is higher than the intensities of both materials IA and IB [6, 7].

I d is reduced

IB

IA

material A

particle B material A

Fig. 3 Intensity I depending on the particle diameter d (according to [7]) These changes are not limited only to the macroscopic properties: The higher intensity in the area of the particle B can for instance be a barrier for an electron (in the tunnel effect), for a photon or a phonon. If always just the energy quantity of an electron, photon or phonon can be transported across the energy barrier, the electrical conductivity, the light or the thermal conductivity becomes quantised. An example for this effect is the fluorescence behaviour of CdTe nano particles. Dilutions with nano particles in a range from 2 – 5 nm change their fluorescence with increasing particle diameter from green to red [12]. Another example is the quantisation of the thermal conductivity, which was detected by Schwab 2000 [13]. EROSION BEHAVIOR OF POLYMERIC MATERIALS

I

(b)

Degradations of insulations for outdoor application are basically caused by high energy stresses. The most important stresses are: - Pollution initiated discharges, - Arcs and - UV-radiation. Usually the stresses of the pollution initiated discharges and arcs are dominating for polymeric materials for outdoor insulation. They can lead to tracking and erosion. Tracking can be presumed as a special form of erosion. It is characterized by a continuous conductive path, which is caused by local erosion on the surface of the material [14]. The resistance to erosion and arcing of materials for high voltage outdoor insulation is evaluated by two standardized tests, the IEC 60587 and IEC 61621 [14, 21]. Within a material family a high resistance to tracking and erosion usually corresponds to a long duration of the arcing test [11]. Therefore the considerably faster arcing test according to IEC 61621 suffices for exploratory analysis in terms of a screening.

IB (a) IA material A

material B

Fig. 2 Transition of an intensity I of a property in material A to material B with a continuous slope (a) or a maximum (b) An example for (a) is the electron-hole concentration in a p-n-junction of a diode. For instance the characteristic of (b) represents the potential barrier at the tip of the electrode of a scanning electron microscope. It will be tunnelled by electrons, which are needed to determine the distance to the sample. It is assumed, that a particle of material B is dispersed in the matrix material A. If the particle diameter is small enough no constant value IB is reached. The intensity of the property is a value on the curve. If the fraction of the

2

Proceedings of the XIVth International Symposium on High Voltage Engineering, Tsinghua University, Beijing, China, August 25-29, 2005

C-09

There are different studies on the impact of the particle size of silica in RTV silicone elastomers. Particle sizes of 1.5 – 12 µm were investigated at filler loadings of 10 -50 % wt. The studies show a better resistance to erosion and a higher thermal conductivity by using smaller particles at the same filler fraction [8, 19]. It has to be studied if this behaviour continues at nanometric scale. EXPERIMENTAL RESULTS Investigated materials Fig. 4 TEM picture of the matrix material with 4 wt. % n-SiO2, 10 000-times magnified

Unfilled HTV silicone rubber with a filler content of (2), 4, 10, and 40 wt. % was investigated. Table 2 shows medium primary particle diameter of the used filler materials. Furthermore, samples without filler material were investigated. Primary particles are mostly spherical clusters of amorphous filler material. The primary particles can not be split up (sheared) at the dispersion process. Aggregates are formed by chemical bonds between the primary particles. Several aggregates can form agglomerates by weaker surface forces. They are sensible to shear at the dispersion process. Filler

Medium diameter of primary particle n-SiO2 7 nm µ-SiO2 212 nm n-Al2O3 13 nm µ-Al2O3 4100 nm Table 2 medium diameter of primary particles of the used filler materials

Fig. 5 TEM picture of the matrix material with 4 wt. % µ-SiO2, 10 000-times magnified

Fig. 4 to Fig. 7 show the TEM (Transmission Electron Microscope) pictures of the investigated materials. All samples were manufactured in the same way. To avoid agglomeration strong shear forces were applied at the dispersing process. As it can be seen from the TEM pictures the agglomeration could not be avoided completely, anyway. The best dispersion was achieved with n-SiO2 (Fig. 4). The filler particles are equally dispersed and partly form aggregates or agglomerates. Nevertheless the particles mostly have a size of 10 – 50 nm, which are nano particles by definition. It can be seen in Fig. 6 that n-Al2O3 tend to form large agglomerates, which are not nano particles. The particle dispersion is well in some parts but not satisfying in general. Due to their not sufficing dispersion the particles and agglomerates are not interspersing the matrix material contrary to the n-SiO2. Because of this it is to be expected that the n-Al2O3 behave generally as µparticles. Both types of µ-particles are obviously larger and do not form aggregates or agglomerates (Fig. 5 and Fig. 7). They have a medium diameter of approximately 250 nm and 4 µm, respectively.

Fig. 6 TEM picture of the matrix material with 4 wt. % n-Al2O3, 10 000-times magnified

Fig. 7 TEM pictures of the matrix material with 4 wt. % µ-Al2O3 8000-times magnified

3

Proceedings of the XIVth International Symposium on High Voltage Engineering, Tsinghua University, Beijing, China, August 25-29, 2005

In general it is more difficult to achieve a good dispersion of n-fillers than of µ-fillers because of their high specific surface and their resulting affinity to form aggregates or agglomerates. Furthermore it is important to consider the combination of filler and matrix material. If the binding structure of the filler material is similar to the one of the material, the tendency to form aggregates is less pronounced. It is assumed, that the dispersion behaviour of the filler types is similar at higher filler concentrations. Unfortunately the contrast of the TEM pictures at higher filler concentrations is insufficient.

C-09

thermal conductivity [W/mK]

0,23

n-SiO2 µ-SiO2 n-Al2O3 µ-Al2O3

0,22 0,21 0,20 0,19 0,18 0,17 0,16 0

5

10

15

20

25

30

35

40

filler concentration [wt. %]

Resistance to HV-Arcing

Fig. 9 Thermal conductivity of investigated materials

The resistance to arcing was tested according to IEC 61621. The size of the samples was 15x30x6 mm. For each material a test series was made consisting of 12 measurements. Fig. 8 shows the mean values of the test duration (arcing time) and the associated minimum and maximum values in dependence on filler type and filler concentration. The results show increasing test durations with increasing filler concentrations. The graph for both µ-fillers and for the n-Al2O3 exhibit no significant differences. But the test duration of the n-SiO2 is significantly higher at 40 % wt.

Dielectric Properties The dielectric loss factor and the relative permittivity of the investigated materials were measured at 3 kV and plotted in Fig. 10 and Fig. 11. For both SiO2-fillers and the n-Al2O3 filler there is no significant difference for the loss factor at 10 and 40 wt %. The value of 10 wt % µ-Al2O3 is smaller than the value of 40 wt %. The loss factors of the SiO2-fillers are smaller than for the Al2O3fillers. There is a remarkably peak for three of the filler types at a filler concentration of 4 wt %. Only the value of nAl2O3 is smaller for 4 wt % than for 10 wt %.

300

n-SiO2 µ-SiO2 n-Al2O3 µ-Al2O3

200

14

loss factor [10^-3] at 3 kV

test duration [s]

250

150 100 50 0 0

5

10

15

20

25

30

35

40

filler concentration [wt. %]

Fig. 8 Duration of the arcing according to IEC 61621 for different filler materials and filler concentrations

n-SiO2 µ-SiO2 n-Al2O3 µ-Al2O3

12 10 8 6 4 2 0 0

5

10

15

20

25

30

35

40

filler concentration [wt. %]

Thermal conductivity

Fig. 10 loss factor of the investigated materials To investigate the influence of the thermal conductivity on the resistance to arcing, measurements of the thermal conductivity were undertaken. Fig. 9 shows the thermal conductivities of the investigated materials. The values for all filler materials rise approximately linear to the filler concentration. There is no significant difference between the four filler materials.

A similar but slighter peak is found for the relative permittivity (Fig. 11) of n-SiO2 and µ-Al2O3. In general the relative permittivity is raising slightly with higher filler concentrations. All values lies between 2.8 and 3.2.

4

Proceedings of the XIVth International Symposium on High Voltage Engineering, Tsinghua University, Beijing, China, August 25-29, 2005

suggested that the nano particles restrict chain movement [3]. The behaviour can not be confirmed for the Al2O3 fillers. The reason is probably the unsatisfactory dispersion of the n-Al2O3. Here the agglomerates have a similar size as conventional µ-particles. An explanation for the peak of the SiO2-fillers at 4 wt % can not be given at the moment.

relative permittivity at 3 kV

3,6

n-SiO2 µ-SiO2 n-Al2O3 µ-Al2O3

3,4

C-09

3,2

3,0

CONCLUSION

2,8

- The interface effect is based on the high fraction of interfacial layers in the dielectric. The fraction of interfacial layers becomes dominant by using nfillers. - Theoretically the properties of the interfacial layer do neither match the properties of the matrix material nor the properties of the filler. If the fraction of interfacial layers is high enough, the properties of the whole dielectric are changing. - A good dispersion of the filler in the matrix material is decisive for the fillers to make an impact, in particular for n-fillers. Only if the whole matrix material is interspersed with filler particles, one can expect a change of the properties of the n-filled dielectric contrary to the µ-filled dielectric. - The necessary dispersion was only achieved with the n-SiO2 filler, which improves the resistance to arcing significantly at higher filler contents contrary to the corresponding µ-filler. Also the loss factor and the relative permittivity are decreasing slightly. - n-Al2O3 acts as a µ-filler because of its large aggregates. That is the reason why there is no remarkable effect on the resistance to arcing or the dielectric properties by using n-Al2O3 instead µAl2O3.

2,6 0

5

10

15

20

25

30

35

40

filler concentration [w t. %]

Fig. 11 relative permittivity of the investigated materials DISCUSSION The erosion behaviour of filled silicone elastomers was investigated with the HV arcing test according to IEC 61621. The test durations increase approximately linear with increasing filler content. This is in good correlation with the thermal conductivity, which is supposed to improve the resistance to erosion. An influence of the particle size at lower filler concentrations becomes not apparent. The graphs for both Al2O3 fillers and the µ-SiO2 show also no significant difference at higher filler concentrations. Only the test duration of the n-SiO2 is remarkably higher at 40 wt %. This result is plausible considering the filler dispersion. Only the particles of the n-SiO2 exist as nano particles in the matrix material. The n-Al2O3 act as a µ-filler because of its large aggregates. Therefore its behaviour is similar to both µ-fillers.

ACKNOWLEDGEMENT

Another fact to discuss is that the enhanced resistance to arcing by using n-SiO2 is achieved only at high filler concentrations. The published improvements of material properties mostly deal with mechanical properties like surface degradation caused by partial discharges and they already occur at low filler concentrations (1 - 10 % wt). In this case the strong interfacial bonding and small inter-filler spaces of the nanodielectrics are supposed to constrain the material degradation. This is relevant, if the upper layer of the matrix material is degraded and the particles are now situated at the surface [22]. The erosion of a material is caused by thermal decomposition. It is supposed, that the stronger interfacial bondings can mitigate the pyrolysis of the polymer chains. This is relevant if the dielectric consists mostly of interfacial layers. To achieve this a higher filler concentration is necessary.

The authors thank to Dr. H.-J. Winter for his cooperation in manufacturing samples and for valuable discussions. REFERENCES [1] Gleiter, H.: „Materials with Ultrafine Microstructures: Retrospectives and Perspectives”. Nanostruc. Materials, Vol.1, pp. 1-19, 1992 [2] Fréchette, M. F.; Trudeau, M.; Alamdari, H. D.; Boily, S.: “Introductory remarks on NanoDielectrics”. CEIDP, pp. 92-99, 2001 [3] Nelson, J. K.; Fothergill, J. C.: “Internal charge behaviour of nanocomposites”. Nanotechnology, Vol. 15, pp. 586-595, 2004 [4] Kozako, M.; Fuse, N.; Ohki, Y.; Okamoto, T.; Tanaka, T.: “Surface Degradation of Polyamide Nanocomposites Caused by Partial Discharges Using IEC (b) Electrodes”. IEEE Trans. Dielectrics and Electrical Insulation, Vol. 11; No. 5; pp. 833839, Oct. 2004;

The n-SiO2 filler has also the lowest loss factor and the lowest relative permittivity at 40 wt %. The values of the corresponding µ-type are higher. This is in good correlation with already published results where it is

5

Proceedings of the XIVth International Symposium on High Voltage Engineering, Tsinghua University, Beijing, China, August 25-29, 2005

C-09

[14] IEC 60587: „ Test methods for evaluating resistance to tracking and erosion of electrical insulating materials used under severe ambient conditions “; 1999 [15] Fröhlich, H.: “Theory of Dielectrics”; Oxford Clarendon Press, 2nd Edition; 1958 [16] Gächter, R.; Müller, H.: „Kunststoff- Additive“; Hanser Verlag München; 2. Ausgabe; 1983 [17] Kind, D.; Kärner, H.: „Hochspannungsisoliertechnik“; Vieweg, 1982 [18] Kumagai, S.; Yoshimura, N.: „Tracking and Erosion of HTV Silicone Rubber and Suppressing Mechanism of ATH”; IEEE Trans. Dielectrics and Electrical Insulation, Vol. 8; No. 2; pp. 203-211; 2001 [19] Omranipour, R.; Meyer, L. H.; Jayaram, S. H.; Cherney, E. A.: “Tracking and Erosion Resistance of RTV Silicone Rubber: Effect of Filler Particle Size and Loading”; Annual Report Conference on Electrical Insulation and Dielectric Phenomena; pp. 371-374, 2002 [20] Rätzke, S.; Kindersberger, J.: “Werkstoffe mit Nanofüllstoffen für Freiluftisolierungen”. ETGFachbericht 99, VDE-Verlag GmbH Berlin Offenbach, 2004 [21] IEC 61621: “Dry, solid insulating materials Resistance test to high-voltage, low-current arc discharges”, 1998 [22] Kozako, M.; Kido, R.; Fuse, N.; Ohki, Y.; Okamoto, T.; Tanaka, T.: „Difference in Surface Degradation due to Partial Discharges between Polyamide Nanocomposite and Microcomposite”. Annual Report Conference on Electrical Insulation and Dielectric Phenomena; pp. 398-401; 2004

[5] Henk, O. P.; Kortsen, T. W.; Kvarts, T.; Saeidi, A.: “Increasing the PD-endurance of epoxy and XLPE insulation by nanoparticle silica dispersion in the polymer”. Nordic Insulation Symposium Stockholm, 2001 [6] Lewis, T. J.: “Interfaces are the Dominant Feature of Dielectrics at the Nanometric Level”. IEEE Trans. Dielectrics and Electrical Insulation, Vol. 11; No. 5; pp. 739-753, 2004 [7] Lewis, T. J.: “Interfaces and Nanodielectrics are Synonymous”. International Conference on Solid Dielectrics, pp. 792-795, 2004 [8] Meyer, L. H.; Cherney, E. A.; Jayaram, S. H.: „The Role of Inorganic Fillers in Silicone Rubber for Outdoor Insulation – Aluminia Tri-Hydrate or Silica”. IEEE Electrical Insulation Magazine, Vol. 20, No. 4; pp. 13-21; 2004 [9] Krämer, A.: „Über das Erosionsverhalten und die Wasseraufnahme von Silikonelastomeren und unterschiedlichen cycloaliphatischen EpoxidharzFormstoffen“. TU Braunschweig; Dissertation 1987 [10] Oberbach, K.; Meyer, B.: „Saechtling Kunststoff Taschenbuch“. Carl Hanser Verlag München Wien, 26. Ausgabe, 1995 [11] Jahn, H.: „Zur Bewertung stofflicher und herstellungsbedingter Einflussgrößen auf das Hydrophobie- und Erosionsverhalten von Silikonelastomeroberflächen“; Dissertation TU Dresden; 2003 [12] Luther, W.; Bachmann, G.: „Eine Ortsbestimmnung“; Technik in Bayern; Nr. 6, 7. Jg. Ausgabe Süd, 2003 [13] Roukes, M.L.: „Unten gibt’s noch viel Platz“; Spektrum der Wissenschaft spezial: Nanotechnologie; 2001

6