WO2008129032A1

WO2008129032A1 - An impregnation medium

Info

Publication number: WO2008129032A1
Application number: PCT/EP2008/054800
Authority: WO
Inventors: Tommaso Auletta; Henrik Hillborg; Su Zhao; Linda Schadler
Original assignee: Abb Research Ltd
Priority date: 2007-04-20
Filing date: 2008-04-21
Publication date: 2008-10-30
Also published as: EP2137740A1

Abstract

An impregnation medium for impregnation of a porous fibrous matrix to achieve an electric insulation material, where the impregnation medium comprises filler particles consisting of Al2O3 dispersed in a thermosetting resin, where the filler particles constitute between 0.1-10 vol-% of the impregnation medium.

Description

An impregnation medium

FIELD OF THE INVENTION AND PRIOR ART

The invention relates to an impregnation medium for impregnation of a porous fibrous matrix to achieve an electric insulation material, where the impregnation medium comprises filler particles consisting Of AI₂O₃ dispersed in a thermosetting resin.

In this description and the subsequent claims the term "porous fibrous matrix" refers to a fibrous matrix that has such permeability that the impregnation medium is capable of penetrating into the cavities between the fibres of the matrix.

Traditionally, machine insulation systems consist of mica tape impregnated with epoxy or polyester resins. These systems are robust and reliable, but have however drawbacks, such as generation of electrical discharges in voids between the mica and the matrix. These voids can be generated by delamination as a result of vibration, thermal cycling etc. Mica tape usually comprises two layers, one is called mica paper, which is mica flakes enriched with binder resin, such as epoxy, and the other layer is a support layer. The support layer is usually made of a porous structure, such as glass fibres or polymeric fibres.

Epoxy is a thermosetting epoxide polymer that cures (polymerize and crosslinks) when mixed with a curing agent and a catalyst. Epoxies find significant use in many applications including: paints and coatings, adhesives, industrial tooling and compos- ites, electrical systems and electronics because of their good thermal, mechanical and electrical properties, low cost, ease of processing and good chemical resistance. However, like other thermosetting materials, most epoxy resins are brittle. Therefore many kinds of micro-sized fillers have been added into epoxy resin to form composites with better combination of mechanical, thermal and electrical properties. Toughening epoxy by introduc- tion of soft particles, such as rubber, has proven to be an effective way and widely used. However, it reduces the stiffness of the epoxy resin. Rigid particles have also been used to improve the stiffness of epoxy resin. The limitation of such filler particles is that they cause a decrease in ductility and opacity. Recently, a large window of opportunity has open to overcome these limitations through the use of filler particles with sizes in the nanometre range. When compared to conventional composites based on micrometer-sized filler particles, the interface between the filler particles and the matrix in a polymer nanocomposite constitutes a much greater area within the bulk material, and hence influences the composite's properties to a much greater extent, even at rather low filler loading.

WO20061 18536 discloses an electric insulation material formed by a porous fibrous matrix that is impregnated with an impregnation medium. One of the aims of the invention described in the above mentioned document is to provide an electric insulation material with high thermal conductivity, which is demanded for the specific use described in the document, which is as insulation material in electrical bushings. The above mentioned document describes the use of amounts of particle fillers in the impregnation medium, up to 25 vol-%. Nanocomposites with such high concentration of filler particles are usually brittle, although the thermal conductivity usually is very high. WO20061 18536 also indicates that a plurality of materials for the filler particles can be used, of which most is not suitable for the purpose of the present invention.

Therefore, there is a need for an impregnation medium for impregnation of a porous fibrous matrix to achieve an electric insulation material, being improved with respect to impregnation mediums already known, for instance by providing an insulation material with improved ductility and increased stiffness without losing strength, thermal conductivity and electrical resistivity, e.g. for the purpose of electrical machine insulation. SUMMARY OF THE INVENTION

The object of the present invention is to provide an impregnation medium for impregnation of a porous fibrous matrix to achieve an electric insulation material being improved with respect to such impregnation media already known by at least partially addressing said need.

This object is according to the invention obtained by providing an impregnation medium and an electric insulation material of the type defined in the introduction, in which the impregnation medium comprises filler particles consisting of AI₂O₃ dispersed in a thermosetting resin, and that the filler particles constitute between 0.1 -10 vol-% of the impregnation medium.

AI₂O₃ particles have a surface which to a very high extent is terminated by OH groups. These OH groups can be chemically modified introducing functionalizing groups which bind to a matrix in which the particles are dispersed. By providing a strong interface between the filler particles and the thermosetting resin in an impregnation medium, significant mechanical advantages can be achieved for the material. The concentration of filler particles in the impregnation medium affects the mechanical, electrical and thermal properties of the material. If the concentration of filler particles lies within the range described above the impregnation medium will show a significant toughening compared to an impregnation medium without filler particles. The toughening is due to higher ductility in the impregnation medium because the filler particles provide for derealization of plastic deformation. Also, the filler particles cause crack deflection in the impregnation medium, which results in difficulties for cracks to travel through the material and cause failure. It has surprisingly been seen that an optimization of the concentration of filler particles within the range described above contributes to give an electric insulation material impregnated with the impregnation medium of the invention a better thermal conductivity as well as mechanical properties well- suited for the use of the electric insulation material in e.g. electrical machines, while the good electrical insulation from the thermosetting resin is maintained. It has also been seen that a too high concentration (over 10 vol-%) of filler particles in the impregnation medium, especially when the filler particles are coated with a surface modifier, results in a decrease in ductility of the impregnation medium. The DC-resistivity is also dependent upon the filler particle concentration, where the DC- resistivity is substantially inversely proportional to the concentration of the filler particles. The electric insulation material described in WO20061 18536 is well-suited for applications where the material is not subjected to wear and impacts, such as in electrical bushings where a high thermal conductivity is more important than toughness. However the materials de- scribed in that document are not optimized with regard to both thermal conductivity and mechanical toughness, while maintaining the resistivity of the thermosetting resin.

According to an embodiment of the present invention the filler particles constitutes between 2-8 vol-%, preferably 3-6 vol-% of the impregnation medium.

It has turned out that the mechanical properties as well as the thermal conductivity and the electrical resistivity has an optimum for the desired applications when the filler particle concentration is between 2-8 vol-%, preferably 3-6 vol-%

According to another embodiment of the invention the thermosetting resin comprises an epoxy resin. Epoxy resin has the high electrical resistance demanded by the invention. It also has a viscosity, which is low enough for enabling homogenous impregnation of the porous fibrous matrix. The epoxy can, after impregnation of the porous fibrous matrix, be the subject of curing to form the electric insulation material of the invention.

According to another embodiment of the invention the average size of the filler particles are in the size range of 1 -100 nm, preferably 20-60 nm, most preferred 30-50 nm.

In order to make the impregnation medium to penetrate the porous fibrous matrix completely during the impregnation the filler particles have to be smaller than 100 nm, i.e. substantially smaller than the pores or cavities in the porous fibrous matrix. In comparison with micro particles, i.e. particles with sizes in the micrometer region, nanoparticles have a tendency to better remain in dispersion within the impregnation medium than the micro particles, without causing sedimentation and wear of the matrix. Particles in the size range of 20-60 nm, preferably 30-50 nm, are especially well-suited to fulfil the requirements during the impregnation of the porous fibrous matrix.

According to another embodiment of the invention the filler parti- cles have a narrow size distribution.

In this description and the subsequent claims the term "narrow size distribution" refers to a size distribution in which more than 90 % of the particles have a particle size in the range of 0.2-2 times the mean particle size.

A narrow size distribution of the filler particles allows for a homogenous impregnation of the porous fibrous matrix with the impregnation medium comprising the filler particles. In a case where the sizes of the particles differs a lot, i.e. have a broad size distribution; larger particles can become concentrated in the surface region of the bulk of the porous fibrous matrix, whereas the smaller particles are the only ones reaching all the bulk. This can cause anisotropic behaviour of the mechanical, thermal and electrical properties.

According to another embodiment of the invention the filler particles are substantially uniformly dispersed in the impregnation medium.

A uniform dispersion of the filler particles in the impregnation medium results in isotropic properties like thermal conductivity, mechanical strength, ductility and electrical resistance. This is desired in most electric insulation material, e.g. for use as electric insulation of electrical machines.

According to another embodiment of the invention the surfaces of the filler particles are coated with a surface modifier configured to inhibit agglomeration of the filler particles, e.g. by making the surfaces of the filler particles more hydrophobic and/or configured to form covalent bonds with the matrix during cross linking of the thermosetting resin..

Surface modifiers will inhibit agglomeration of the filler particles in the impregnation medium. Agglomeration can cause formation of aggregates several micrometers in size. This can decrease the mechanical strength of the impregnation medium and cause crack formation, since the aggregates can work as crack initiation sites. Also, by introducing surface modifiers containing functional groups which can form covalent bonds to the thermosetting resin during crosslinking the bonds between the filler particles and the thermosetting resin in the impregnation medium can be strengthened and can also promote interfacial debonding which would also contribute to the toughening of the material. According to another embodiment of the invention the surface modifier is an organic compound, a silane or a compound which can be described as a combination of an organic compound and a silane, such as an aminosilane.

The compounds mentioned above are especially useful as surface modifiers and fulfils the demands mentioned above of the present invention.

The invention also relates to an electric insulation material comprising a porous fibrous matrix impregnated with an impregnation medium, where the impregnation medium is the impregnation medium according to the invention.

According to another embodiment of the invention the porous fibrous matrix comprises mica or mica and at least one in the group consisting of: cellulose fibres, glass fibres and polymeric fibres.

According to another embodiment of the invention the porous fibrous matrix is in the form of paper, pressboard, laminate, tape, weave or sheets.

The invention also relates to the use of an electric insulation material according to the invention in a machine insulating system.

Other advantages and advantageous features of the invention will appear from the dependent claims and the subsequent description.

BRIEF DESCRIPTION OF THE DRAWINGS

With reference to the appended drawings, below follows a spe- cific description of embodiments of the invention cited as examples. In the drawing:

Fig 1 shows representative stress-strain graphs for neat polymer, 3.1 vol-% NT-AI₂O₃/epoxy nanocomposite and 3.1 vol-% APTES-AI₂O₃/epoxy nanocomposite,

Fig 2 shows graphs for strain-to-break versus particle concentration for neat polymer, NT-AI₂O₃/epoxy nanocomposites and APTES-AI₂O₃/epoxy nanocompo- sites,

Fig 3 shows the fatigue crack growth rate versus applied stress intensity range curve for neat polymer, 3.1 vol- % NT-AI₂O₃/epoxy nanocomposite and 3.1 vol-% APTES-AI₂O₃/epoxy nanocomposite,

Fig 4 shows graphs over the loss tangent data at 50 Hz for neat polymer, NT-AI₂O₃/epoxy nanocomposites and APTES-AI₂O₃/epoxy nanocomposites, at 25°C, 70⁰C, 120⁰C and 155°C,

Fig 5 shows graphs over the relative change in electric breakdown strength (%) for the NT-AI₂O₃/epoxy nanocomposites and APTES-AI₂O₃/epoxy nanocomposites compared to the neat polymer as a function of nanoparticle concentration (vol-%), and

Fig 6 shows graphs over the DC-resistivity at different electric fields (1 kV/mm and 3kV/mm) and temperatures (25°C and 70⁰C) for (a) NT-AI₂O₃ nanocomposites, and (b) APTES-AI₂O₃ nanocomposites.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS OF THE INVENTION Explained herein are preferred embodiments of the invention, describing optimization of fatigue and fracture qualities of the nanoparticle-filled nanocomposites. The invention may, however, be embodied in many different forms and should not be construed as being limited to the exemplary embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the invention to those skilled in the art.

In this disclosure three types of impregnation media are compared:

^■ well-dispersed non-treated AI₂O₃ nanoparticles in epoxy resin, denoted as NT-AI₂O₃/epoxy nanocomposite,

^■ 3-aminopropyltriethoxysilane treated AI₂O₃ nanoparticles in epoxy resin, denoted as APTES- AI₂O₃/epoxy nanocomposite, and

^■ epoxy resin which not was added with filler particles, denoted as neat polymer. The neat polymer is not a part of the invention but is included in this description for comparison.

Sample preparation and testing

In these studies, a Huntsman Araldite® epoxy system was chosen as the thermosetting matrix polymer, including (i) Araldite F - bisphenol A liquid epoxy resin ; (ii) HY905 - modified dicarbox- ylic anhydride hardener; (iii) DY062 - amine catalyst. The mixing ratio of epoxy resin to hardener was 1 : 1 by weight. NanoTek® Aluminium Oxide (AI₂O₃) was purchased from Nanophase Technologies Corporation, with an average particle size of 45 nm. 3-aminopropyltriethoxysilane (APTES) from Gelest Inc. was chosen as surface modifier for the aluminium oxide nanoparticles. The aluminium oxide nanoparticle surface modification procedure was: (i) addition of 1 g AI₂O₃ nanoparticles to 50 ml of 95% ethanol; (ii) sonication of the mixture for 5 minutes using a wand; (iii) addition of 1 .5 g APTES and sonication of the mixture for another 10 minutes; (iv) refluxing the mixture for 3 days at 80 ⁰C in an oil bath; (v) centrifuging and washing the nanoparticles with ethanol and hexane to remove the by-products and extra silane, followed by drying of the nanoparticles in a vacuum oven overnight at room temperature.

NT-AI₂O₃ and APTES-AI₂O₃/epoxy nanocomposites with 1 .6, 3.1 , 4.6 and 6.1 vol-% nanoparticle concentrations were prepared. Before use, the NT-AI₂O₃ nanoparticles were dried in vacuum at 190⁰C overnight. Three steps were used to prepare the nanocomposites. First, the nanoparticles were dispersed in the liquid epoxy resin to prepare a "masterbatch". This was the key step for the whole procedure. A well-dispersed masterbatch was achieved by shear mixing using a Hauschild SpeedMixer®. This machine provides high shear stress gradients to disrupt particle aggregates. Because the surface modified AI₂O₃ nanoparticles aggregate easily after the centrifuging and drying processes, for the APTES-AI₂O₃/epoxy nanocomposites, the dispersion was improved by adding 1/8" aluminium oxide balls during mixing. The balls were removed before curing. Second, the hardener and catalyst were added as per composition requirements into the masterbatch; Third, the mixture was mixed, degassed at room temperature, cured at 80⁰C for 6 h, and then post-cured at 135°C for 10 h. Neat polymer samples were also made to compare with the nanocomposites.

Scanning electron microscopy (SEM) on a FESEM JEOL 6335, was performed for evaluation of the filler particle dispersion in the epoxy resin. The images used for evaluation are not included in this disclosure.

Tensile tests were performed on an lnstron 4204 tensile test machine with a constant loading speed of 0.1 mm/min. The Young's modulus, strain-to-break and stress-strain curves of the materials were obtained. At least 8 samples were tested for each condition and the results are presented in Table 1 . The stress- strain and stress-to-break curves are presented in Fig 1 and Fig 2, respectively.

Table 1. Properties for neat polymer, NT-AI₂O₃/epoxy and APTES- Al₂θ₃/epoxy nanocomposites.

Fracture toughness of the materials was measured on an lnstron 4204 machine based on ASTM Standard E1820 and D5045 with compact tension fracture specimen at a loading speed of 1 mm/min. Then the fracture energy was calculated based on

Tapered-double-cantilever-beam (TDCB) fracture specimens were used to measure the fatigue crack propagation rates of the materials. The tests were performed on an lnstron 8562 with a 5 KN load cell. The samples were pin loaded. A triangle frequency of 1 Hz was applied with a load ratio (R = K_min/K_maχ) of 0.1 . Crack lengths were measured optically. The TDCB geometry provided a crack-length-independent relationship between ap- plied stress intensity factor Kl and load P. A constant range of Mode-I stress intensity factor ΔKI was achieved by applying a constant range of load ΔP independent of crack length. Samples were pre-cracked with a razor blade. The log-log plot of fa- tigue-crack-growth rate da/dN versus the applied range of stress intensity factors ΔKI was determined and fitted to the Paris law and is presented in Fig 3, where n is the Paris exponent.

dN ° ⁷

The results from the test are presented in Table 2.

Table 2. . Fracture toughness and fracture energy of neat epoxy polymer, NT-AI₂O₃/epoxy and APTES-AI₂O₃/epoxy nanocompo- sites.

Differential Scanning Calorimetry (DSC) model Q100 from TA® Instruments was used to check the glass transition temperature (Tg) of the nanocomposites and neat polymer. The samples were heated from 25°C to 160⁰C with a constant heating rate of 10°C/min. Data obtained from the second heating run were used, and at least 3 samples were tested for each material. The results are presented in Table 1 . The dielectric response, permittivity and loss tangent of neat polymer, NT-AI₂O₃/epoxy and APTES-AI₂O₃/epoxy nanocompo- sites were performed on a Heraeus Vόtsch machine with an IDA data collection system. A full frequency sweep from 10^~3 to 10³ Hz of each material was measured at 25°C, 70⁰C, 120⁰C and 155°C (T_g~1 10°C). The results are summarized in the graphs over the loss tangent data in Fig 4.

Short-term electric strength measurements were measured under AC conditions (50 Hz) with a ramp rate of 500 V/s in oil. In order to avoid the problem with the sample thickness, the same thickness range (between 0.1 -0.2 mm) for the reference neat polymer sample and each nanocomposite sample was chosen and analyzed by ReliaSoft's Weibull ++6.0 software. At least 7-9 data points were used for each material. The results are shown in Table 3. And the relative change (%) of breakdown strength for the nanocomposites compared to the neat polymer as a function of filler loading (vol-%) was plotted in Fig 5.

Table 3. Electric strength results for the reference neat polymer and the nanocomposites.

DC-resistivity measurements of the neat polymer, as well as the NT-AI₂O₃/epoxy and APTES-AI₂O₃/epoxy nanocomposites were performed on an in-house-made machine. In order to check if the filler introduces some field dependence or temperature dependence, each material was measured at two electrical stresses of 1 kV/mm and 3 kV/mm, and two temperatures of 25°C and 70⁰C. After the field and temperature cycling, a measurement at 1 kV/mm and 25°C was performed again for each material to check for any aging. The charging time for each sample was 4 h and discharging time was 48 h. The results are summarized in Fig 6a and Fig 6b.

Good dispersion is critical for achieving higher ductility of the nanocomposites of the invention. From SEM micrographs (not included here) it was concluded that many of the nanoparticles are individually dispersed, but small aggregates were observed for both NT-AI₂O₃/epoxy and APTES-AI₂O₃/epoxy nanocompo- sites. A comparison showed that the nanoparticle/epoxy adhesion was stronger for the APTES-AI₂O₃/epoxy nanocomposites.

One indication of proper curing in epoxies is the glass transition temperature (T₉). For these samples, there was no change in the glass transition temperature for 1 .6, 3.1 , 4.6 vol-% and even for 6.1 vol-% NT-AI₂O₃/epoxy nanocomposites, which indicates that the crosslink density of the epoxy resin did not change significantly upon nanoparticle addition (Table 1 ). For the APTES-AI₂O₃/epoxy nanocomposites, the glass transition temperature decreased at high filler concentration (1 1⁰C de- crease at 6.1 vol-%), which indicates a change in curing reaction due to the surface modification of AI₂O₃ nanoparticles. This is likely due to a reaction of the APTES with the epoxy system.

A major part of this effort is to produce nanocomposites with improved ductility and increased stiffness without losing strength. Figure 1 shows the representative stress-strain curves for the 3.1 vol-% NT-AI₂O₃/epoxy nanocomposites and the 3.1 vol-% APTES-AI₂O₃/epoxy nanocomposite obtained by tensile tests using ASTM D638-03. The mean strain-to-break values are listed in Table 1 . It is clear that the strain-to-break, which indicates the ductility of the material, for the 3.1 vol-% APTES- AI₂O₃/epoxy nanocomposite increased significantly (39%) compared to the neat polymer. The strain-to-break of the 3.1 vol-% NT-AI₂O₃/epoxy nanocomposite did not change compared to the neat polymer. In addition, the Young's modulus increased for all the nanocomposites compared to the neat polymer.

The strain-to-break as a function of nanoparticle concentration for both of the NT-AI₂O₃/epoxy and APTES-AI₂O₃/epoxy nanocomposites is presented in Figure 2, and the corresponding data are listed in Table 1 .

Consider first the NT-AI₂O₃/epoxy nanocomposites. For the 1 .6, 4.6 and 6.1 vol-% nanocomposites, the mean strain-to-break decreased compared to the neat polymer. However, the 3.1 vol-% nanocomposite stands out. When comparing the fracture morphology from SEM micrographs (not shown here) of the 3.1 vol-% NT-AI₂O₃ nanocomposite with that of the neat polymer, it was clear that the former had a rougher surface than the latter. 3D hackle markings appeared in the 3.1 vol-% NT-AI₂O₃ nanocomposite, in which small scale-like cracks are deflected around the nanoparticles. Careful observation of lower and higher magnification images showed plastic deformation of the matrix, crack deflection, and particle pull-out for the 3.1 vol-% NT-AI₂O₃/epoxy nanocomposite. These are indications of poten- tial toughening mechanisms. However, although a relatively good dispersion were obtained for all of the nanocomposites, there were occasional aggregates with a diameter of 3-6 μm in each material. The size of these aggregates was on the order of the critical crack length for epoxy. Therefore, it is assumed that the nanoparticles helped increase ductility because the deformation was not localized, but initiated at the nanoparticles as plastic deformation and particle pullout throughout the composite which increases the volume of material deforming. In addition, crack deflection also occurred. These benefits, however, were balanced by the stress concentrations caused by the aggregates.

The APTES-AI₂O₃/epoxy nanocomposites achieved better properties. All of the strain-to-break values for the APTES- AI₂O₃/epoxy nanocomposites are higher than both neat polymer and NT-AI₂O₃/epoxy nanocomposites at the same particle loading, except for 6.1 vol-%. An 1 1 °C decrease in glass transition temperature, which indicates incomplete curing, might explain the decreased ductility at 6.1 vol-%. t-test (a statistical hypothe- sis test) results show that the mean strain-to-break of the 3.1 vol-% APTES-AI₂O₃/epoxy nanocomposites is significantly higher than the neat polymer at a 95 % confidence level; and for 4.6 vol-% APTES-AI₂O₃/epoxy nanocomposites, it is significantly higher than the neat polymer at 90 % confidence level. The frac- ture morphology of the 3.1 vol-% APTES-AI₂O₃/epoxy nanocomposite was also investigated. At low magnification it looks similar to the 3.1 vol-% NT-AI₂O₃/epoxy nanocomposite. This indicates that the toughening mechanisms of plastic deformation of matrix, crack deflection and particle pull-out also exist in the APTES-AI₂O₃/epoxy nanocomposites. Comparing higher magnification SEM micrographs of NT-AI₂O₃/epoxy nanocomposites and APTES-AI₂O₃/epoxy nanocomposites, crack pinning, which was indicated by a tail structure left behind each particle, can be clearly seen in the APTES-AI₂O₃/epoxy nanocomposite. In addition, the stronger interface between the APTES treated AI₂O₃ particles and the epoxy matrix could cause significant interfacial debonding which would also contribute to the toughening of the APTES-AI₂O₃/epoxy nanocomposites.

The fracture toughness and fracture energy of both NT-AI₂O₃ and APTES-AI₂O₃ filled epoxy nanocomposites improved compared with the neat polymer, and were optimized at nanoparticle loading of 3.1 -4.6 vol-%, Table 2.

The fatigue-crack propagation in neat polymer, 3.1 vol-% NT- AI₂O₃/epoxy and 3.1 vol-% APTES-AI₂O₃/epoxy nanocomposites were measured. The values of the Paris exponent n dropped from 10 for neat polymer to 7 for 3.1 vol-% NT-AI₂O₃/epoxy nanocomposites, and further to 5 for 3.1 vol-% APTES- AI₂O₃/epoxy nanocomposites (Fig 3). The results indicate the retardation of fatigue crack growth rate by adding the nanoparti- cles into the polymer system, especially for the surface modified nanoparticles.

Considering the dielectric properties (see Fig 4), below and around T₉ (at 25°C, 70⁰C and 120⁰C), there was no significant change of dielectric properties for both NT-AI₂O₃/epoxy and APTES-AI₂O₃/epoxy nanocomposites compared with the neat polymer, which indicates that the dielectric properties of the ep- oxy were maintained after introducing AI₂O₃ nanoparticles. At 155°C, compared with the neat polymer, an increase in the dielectric properties occurred. The loss tangent at 50 Hz increased from 0.18 for neat polymer to 0.39 for 6.1 vol-% APTES-AI₂O₃/epoxy nanocomposite and 0.57 for 6.1 vol-% NT- AI₂O₃/epoxy nanocomposite. All of the APTES-AI₂O₃/epoxy nanocomposites had lower loss tangent values compare to the NT-AI₂O₃/epoxy nanocomposites at the same filler particle concentration. And the relative permittivity increased slightly from 7.01 for neat polymer to 8.20 for 6.1 vol-% APTES- AI₂O₃/epoxy nanocomposite and 8.38 for 6.1 vol-% NT- AI₂O₃/epoxy nanocomposites.

A clear improvement in the electric strength for all the nanocomposites compared with the neat polymer was observed (see Table 3). From the Weibull statistic analysis (see Table 3 and Fig 5), it is clear that there is a statistically significant improvement in breakdown strength for both NT-AI₂O₃/epoxy and APTES-AI₂O₃/epoxy nanocomposites compared to the neat polymer. The APTES-AI₂O₃/epoxy nanocomposites had higher electric strength than NT-AI₂O₃/epoxy nanocomposites at the same filler loading. However, there is no clear nanoparticle concentration dependence on the observed improvement in the breakdown strength.

For the DC-resistivity the results show that there was no significant change in DC-resistivity of the epoxy introduced by the nanoparticles (see Fig 6a and Fig 6b). The DC-resistivity decreased with nanoparticle concentration for both NT-AI₂O₃/epoxy and APTES-AI₂O₃/epoxy nanocomposites. At 25°C, there was no electric field dependence for all the materials. At 70⁰C, a higher electrical field gave higher DC-resistivity values for all of the materials. And with the increasing temperature, the DC-resistiv- ity decreased for all the materials. However, all of the nanocomposites were still highly insulating. And there was no aging for any of the composites.

The invention is of course not in any way limited to the embodiments described above. On the contrary, several possibilities to modifications, such as other types of substrate materials and additional layers in the coating, thereof should be apparent to a person skilled in the art without departing from the basic idea of the invention as defined in the appended claims.

Claims

1 . An impregnation medium for impregnation of a porous fibrous matrix to achieve an electric insulation material, where the impregnation medium comprises filler particles consisting of AI₂O₃ dispersed in a thermosetting resin, characterized in that the filler particles constitute between 0.1 -10 vol-% of the impregnation medium.

2. An impregnation medium according to claim 1 , characterized in that the filler particles constitutes between 2-8 vol- %, preferably 3-6 vol-% of the impregnation medium.

3. An impregnation medium according to claim 1 or 2, characterized in that the thermosetting resin comprises an epoxy resin.

4. An impregnation medium according to any of the preceding claims, characterized in that the average size of the filler particles are in the size range of 1 -100 nm, preferably 20-

60 nm, more preferred 30-50 nm.

5. An impregnation medium according to any of the preceding claims, characterized in that the filler particles have a narrow size distribution.

6. An impregnation medium according to any of the preceding claims, characterized in that the filler particles are substantially uniformly dispersed in the impregnation me- dium.

7. An impregnation medium according to any of the preceding claims, characterized in that the surfaces of the filler particles are coated with a surface modifier configured to inhibit agglomeration of the filler particles, e.g. by making the surfaces of the filler particles more hydrophobic and/or configured to form covalent bonds with the matrix during crosslinking of the thermosetting resin.

8. An impregnation medium according to claim 7, character- ized in that the surface modifier is an organic compound, a silane or a compound which can be described as a combination of an organic compound and a silane, such as an aminosilane.

9. An electric insulation material comprising a porous fibrous matrix impregnated with an impregnation medium, characterized in that the impregnation medium is the impregnation medium according to any of claims 1 -8.

10. An electric insulation material according to claim 9, characterized in that the porous fibrous matrix comprises mica or mica and at least one in the group consisting of: cellulose fibres, glass fibres and polymeric fibres.

1 1 . An electric insulation material according to claim 9 or 10, characterized in that the porous fibrous matrix is in the form of paper, pressboard, laminate, tape, weave or sheets.

12. Use of an electric insulation material according to any of claims 9-1 1 in a machine insulating system.