The invention relates to a thermal insulation material comprising an outer radiation reflective surface, comprising a protective coating comprising a transparent fluoropolymer, to the use of such a polymeric composition comprising such a transparent fluoropolymer for coating a radiation reflective surface of a thermal insulation material and to a method for coating an outer radiation reflective surface with such a fluoropolymer.

Thermal insulation materials comprising an outer radiation reflective surface are used as insulation material that prevent heat transfer by thermal radiation. Such materials reflect radiation heat and prevent transfer from the outer side, where the thermal radiation hits the material, to the opposite inner side. Such materials are radiant barriers, in particular against infrared red (IR), providing reflective insulation. Examples of such materials are insulation materials, such as foam panels, glass wool, stone wool and rubbers, such as poly(styrene-butadiene-styrene) (SBS) and ethylene propylene diene monomer (EPDM) rubbers, where the insulation material can be provided with a radiation reflective surface, such as a metal layer.

The outer reflective surface of such thermal insulation material preferably reflects not only IR radiation, but also visible light and UV radiation. In the building industry, such thermal insulation material is often used indoors, e.g. below roofing surfaces, as it is avoided to expose the radiation reflective surfaces known in the art to environmental conditions such as variable temperatures, humidity, corrosion, (acid) rain etc. It is therefore desired to coat the radiation protective surface with a protective layer to protect the radiation reflective surface from weathering influences and to render such surfaces suitable for outdoors usage.

Fluoropolymer coatings are strong and can form an oxygen barrier to prevent corrosion of the coated material. In the art, many fluoropolymers are known, such as polytetrafluorethylene (PTFE), polyvinylidene fluoride (PVDF), ethylene tetrafluoroethylene (EFTE), copolymers such as tetrafluoroethylene perfluoromethylvinylester (MFA) and terpolymers such as tetrafluoroethylene-hexafluoropropylene-vinylidene fluoride terpolymer (THV). Because of the strong covalent C-F bond, fluoropolymers have a high chemical inertness, UV stability and provide a strong oxygen barrier. However, most of the fluoropolymers are not transparent and are therewith not suitable as a transparent coating.

Protective coating materials based on fluoropolymers are known, e.g. polyvinylidene fluoride (PVDF) resin, a copolymer of vinylidene fluoride and tetrafluoroethylene, marketed under the name Kynar® of (Arkema, France). Indeed, PVDF provides an improved lasting performance in coating applications for the building industry. However, it was found that PVDF containing coatings have a limited durability, caused e.g. by the absorption profile of PVDF in the IR and UV range. By exposure to thermal radiation, PVDF absorbs IR and UV resulting in heat generation, leading to deterioration of the polymer cured coating in time. Emission rates, i.e. the radiation that is passed through the coating are undesirably low and become even lower in time.

From US4,710,426 a thermal insulation material of the above kind is known, wherein a transparent surface comprises an infrared reflection layer, which layer is coated with a fluoropolymer. However, the fluoropolymer is void of crosslinkable hydroxyl groups.

US2013/0040148 relates to a coating composition comprising a fluorinated copolymer.

US2012/301728 describes a coating composition for back coating of a solar heat collecting reflector, wherein the coating composition comprises a crosslinkable fluoropolymer. The fluoropolymer is a choice out of a broad set of proposed monomers. The fluor containing monomer is at least one of tetrafluoroethylene, chlorofluoroethylene, hexafluoropropylene, vinylidene fluoride and vinyl fluoride, whereas only examples of chlorotrifluoroethylene are given. The coating of US2012/301728 is prepared by polymerising chlorotrifluoroethylene with hydroxybutyl vinyl ether and a vinyl ether. The presence of chlorine groups renders the coating susceptible to weathering by radical formation.

US2018/0222168 describes a similar fluorpolymer as described in US2012/301728, now in combination with a polyaspartic acid ester.

In the art, roofing materials from polyvinylchloride (PVC) and thermoplastic elastomer olefin (TPO) are also used, but these materials are not transparent and are not or to a very small extent durable reflective for UV radiation.

Therefore, an improved coating material is desired with improved weathering resistance, emissivity, and reflection profile, while still having the protective strength and oxygen barrier characteristics of the known fluoropolymers.

It has now been surprisingly found that a crosslinked saturated aliphatic unbranched terpolymer composed of tetrafluoroethylene monomers, ethylene monomers, and hydroxy ethylene monomers, wherein the crosslinks are between hydroxyl groups of the terpolymers provide for such an improved coating. Such terpolymers have been found to have such improved emissivity and reflection profiles. Such polymers are fully saturated, i.e. being free of double C=C bonds, free of chlorine, and unbranched, meaning that the polymer chains, before being crosslinked, have a single linear carbon backbone with only hydrogen, fluorine and hydroxy moieties.

The said terpolymers are fully transparent for infra-red and visible light while being extreme UV stable. This means that such coatings are highly resistant against weathering, caused by UV radiation. It was observed that UV is passed through the coating or converted therein into IR. This conversion is without heat generation because of the full transparency of the terpolymer for IR. The radiation that passes through the coating and the IR generated therein are reflected by the radiation reflective surface of the thermal insulation material without heat generation. The radiation arrives at the surface of the reflective material where it bounces away from the reflective material without penetration and concomitant heat generation. Indeed, PVDF is not a fully transparent coating.

Therefore, the invention provides a thermal insulation material comprising a protective coating comprising a transparent fluoropolymer, wherein the fluoropolymer is a crosslinked saturated aliphatic unbranched terpolymer composed of tetrafluoroethylene monomers, ethylene monomers, and hydroxy ethylene, wherein the crosslinks are between hydroxyl groups of the terpolymer. It was observed that such thermal insulation materials are not only suitable to be used below roofing surfaces, but can be used as roofing surface itself, as these materials are extremely resistant against the above- mentioned environmental conditions and are weathering and waterproof.

The terpolymer is preferably defined by the atomic content of its constituents C, F and O. The atomic content can conveniently be determined by known methods in the art, in particular by energy-dispersive X-ray spectroscopy using scanning electron microscopy, also known as SEM-EDX (Corbari et al., 2008, Biogeosciences 5(5):1295-1310; Goldstein, 2003, Scanning Electron Microscopy and X-Ray Microanalysis, Springer, ISBN 978-0-306- 47292-3. It is to be noted that the atomic hydrogen content cannot be determined using the SEM-EDX technique.

Accordingly, the uncrosslinked terpolymer preferably has an atomic fluorine content of 5 - 40%, more preferably of 10 - 30% and most preferably of 15 - 25%. An atomic fluorine content of 5% means that 5 out of 100 atoms of the terpolymer are fluorine atoms. The atomic oxygen content of the terpolymer is preferably 10 - 60%, more preferably 20 - 40% and most preferably 25 - 35%. The atomic carbon content of the terpolymer is preferably 20 - 80%, more preferably 35 - 65% and most preferably 45 - 55%.

The molecular weight of the terpolymer can be measured with methods, known in the art, such as with gel permeation chromatography (GPC). This provides number a.o. average molecular weight (M _n ), weight average molecular weight (M _w ), and the polydispersity index (PDI). Based on the percentages of functional groups, the molecular weight the mole percentage thereof can be quantified.

Accordingly, expressed in mole percentage, in order to provide for the envisaged radiation and weathering resistance, the uncrosslinked terpolymer has a fluoridisation degree of 10 - 70 mol%, more preferably of 15 - 50 mol%, and most preferably of 20 - 40 mol%. Other attractive ranges are 15 - 70 mol%, in particular of 25 - 60 mol%, more in particular 30 - 45 mol%. In addition to the above-discussed SEM-EDX analysis, the fluoridisation degree can be determined by numerous techniques, known in the art, such as by scanning electron microscopy Nicolae and Amariei, OHDM 2011 , vol. 10, nr. 4, p 199 - 208) or near infrared spectroscopy (Tamburini et al., Sensors 2016, vol. 16, 1216).

Accordingly, the uncrosslinked terpolymer preferably comprises 30 - 85 mol% ethylene and hydroxy ethylene monomers, more preferably 40 - 75 mol% and most preferably 55 - 70 mol%. The skilled person is aware of how to determine the molar percentage of ethylene and hydroxy ethylene monomers in the terpolymer, and may use a method, such as titration, Fourier-transform infrared (FTIR) spectroscopy as defined by ASTM E168 and ASTM E1252 standards, X-ray photoelectron spectroscopy and mass spectroscopy (MS) see for suitable method also the textbook “Characterisation of polymer blends, Grohens and Jyotishkumar, in particular chapter 20 therein.

In line with the above, as determined by SEM-EDX analysis, the uncrosslinked terpolymer preferably comprises 5 - 50 mol%, more preferably 10 - 40 mol% hydroxy ethylene monomers. Other attractive ranges are 3 - 40 mol%, in particular 10 - 30 mol%. Expressed differently, the terpolymer preferably has an OH number, or hydroxyl value, of 10 - 150 mg KOH /g terpolymer, in particular of 20 - 120 mg KOH/g terpolymer, more in particular of 30 - 100 mg KOH/g terpolymer, even more in particular of 30 - 60 mg KOH/g terpolymer, and still even more in particular of 50 - 55 mg KOH/g terpolymer. The hydroxyl value is defined as the number of milligrams of potassium hydroxide required to neutralize the acetic acid taken up on acetylation of one gram of the terpolymer and can e.g. be determined according to the ASTM D1957 and ASTM E222-10 standards. The hydroxyl value is a measure of the content of free hydroxyl groups in the terpolymer, expressed in units of the mass of potassium hydroxide (KOH) in milligrams (mg) equivalent to the hydroxyl content of one gram of the terpolymer. The analytical method used to determine the hydroxyl value traditionally involves acetylation of the free hydroxyl groups of the terpolymer with acetic anhydride in pyridine solvent. After completion of the reaction, water is added, and the remaining unreacted acetic anhydride is converted to acetic acid and measured by titration with potassium hydroxide.

As the monomers of the terpolymer are tetrafluoroethylene, ethylene, and hydroxy ethylene, the uncrosslinked terpolymer is substantially free from nitrogen, silicon and chlorine atoms. Such atoms may be determined by SEM-EDX as contaminants only, with an atomic content for each of the contaminants of preferably not above 2%, more preferably not above 1%. Preferably, the atomic content of each of Si and Cl is not more than 1%, preferably, the atomic content of both Si and Cl is not above 1%. The above hydroxyl value, or content of hydroxy ethylene monomers in the terpolymer renders the terpolymer soluble in a polar organic solvent, and allows for crosslinking of the polymer with a stable crosslinker, known in the art.

Suitable organic solvents are e.g. ketones and acetals, such as 2-butanon, ethylbenzene, cyclohexanone, butyl acetate, ethyl acetate; a saturated hydrocarbon solvent such as xylene or toluene.

In particular the above atomic content, but also the numbers for fluoridisation degree, ethylene and hydroxy ethylene monomer content and hydroxyl value, clearly define the terpolymer by way of relative content of the different monomers.

The transparent terpolymer is preferably solvent-based, to allow application of the polymer onto the radiation reflective material by spraying or other known techniques. The majority of fluoropolymers are however hardly soluble in organic solvents. Perfluoroalkoxy (PFA) has a high degree of transparency, but has a low solubility, rendering it impossible for application of thin layers onto a substrate. The hydroxyl groups of the terpolymer of the invention makes the terpolymer more readily soluble in organic solvents such as xylene, ethylbenzene and 2-butanon, also known as methylethylketone (MEK). The more hydroxyl groups, the more soluble the polymer is. The terpolymers as described herein are very suitable in view of solubility, transparency, UV resistance and inertness for IR and visible light. Suitable terpolymers are e.g. Lamoral 100 series or 400 series (Lamoral Coatings, Netherlands).

The solubility makes the polymers particularly suitable to be coated on surfaces, e.g. by roll-to-roll converting/coating or any other suitable technique, known to the skilled person.

In another attractive embodiment, the terpolymer of the coating is crosslinked by a polyisocyanate. A polyisocyanate hardens, or cures, the terpolymer by reacting with the hydroxyl groups to produce polyurea and urethane linkages. By crosslinking the hydroxy groups of the terpolymer, a coating of the envisaged qualities is obtained. Said crosslinking preferably occurs while a composition of the terpolymer, such as a solution, and the crosslinking agent is applied and drying off solvents on the substrate surface, i.e. the thermal insulation material, whereafter the curing on the said substrate surface takes place resulting in the envisaged coating.

The skilled person is aware of suitable polyisocyanates for use in crosslinking the terpolymer of the invention. Preferably, the polyisocyanate comprises a diisocyanate, more preferably an aliphatic diisocyanate. Polyisocyanates comprising double C=C bonds or cyclic structures are less preferred, as these are less stable and may lead to less weather durability of the cured terpolymer coating. Hexamethylene diisocyanate is one of the preferred aliphatic diisocyanates. Hexamethylene diisocyanate is e.g. available as Lamoral C20 or C100 (Lamoral, The Netherlands). Lamoral C20 has 20% solids in MEK whereas Lamoral 100 is solvent free (i.e. 100% solids). Minimisation of the solvent is preferred.

In particular, the weight ratio between the fluoropolymer and the polyisocyanate, based on dry weight, is 1 - 30 : 1 , preferably 3 - 25 : 1. Other attractive ranges are 5 - 25:1 , in particular 10 - 20:1.

In an attractive embodiment, the crosslinked terpolymer of the thermal insulation material has a density of 0.9 - 1.4 g/m ² . The glass transition temperature of the cured terpolymer is preferably 180 - 210°C.

Particularly, the coating of the article described herein has a thickness of 1 - 10 pm, in particular of 2 - 5 pm. The thickness can be measured as known in the art, e.g. according to the iso3882 standard or the iso19840 standard.

The coating preferably transmits more than 95% of the radiation having a wavelength of between 300 and 2500 nm, preferably more than 98%, even more preferably at least 99% thereof.

The radiation reflective surface preferably reflects at least 80% of the radiation (i.e. UV, visible light, and IR). Such materials are suitable as reflective insulation materials. However, the radiation reflective surface preferably has a reflectivity for UV, visible light and I R of at least 90%, preferably at least 95%, 96%, 97%, 98%, 99% or even 100%. Reflection can be measured as is known to the skilled person. For example, a UV-Vis spectrophotometer can be used, such as the Shimadzu UV-3600 with a LISR-3100 150 mm diameter integrating sphere.

The coating is impermeable for oxygen, vapour, and water, and is therefore optimally suitable to be applied to a radiation reflective surface of an article according to the invention, wherein the radiation reflective surface is corrosion sensitive, such as a metal surface. The coating renders the substrate, in particular the metal weather and durable waterproof.

The radiation reflective surface preferably comprises a metal.

In an attractive embodiment, the article of the invention is a building element, in particular chosen from the group, consisting of roofing tiles, roofing panels, wall elements, thermal insulation panels, thermal insulation wool, such as stone wool or glass wool, in particular provided with a radiation reflective surface of e.g. metal. The thermal insulation material is preferably waterproof, in particular as a result of the terpolymer coating applied thereon. In another aspect, the invention relates to the use of a polymeric composition comprising the saturated aliphatic unbranched terpolymer as defined above, and a crosslinking agent, for coating an outer radiation reflective surface of a thermal insulation material, where the crosslinking agent and the thermal insulation material are preferably as described above. It is preferred that the composition comprises both the terpolymer and the crosslinking agent. The composition, and therefore also the protective coating comprising the crosslinked terpolymer, are preferably free of polyaspartic acid ester.

The invention further relates to a method for coating an outer radiation reflective surface with a terpolymer as defined herein, comprising the steps of a. preparing a 15 - 35 w/w% solution of a mixture of 1 : 10 - 50 polyisocyanate : terpolymer on weight basis in an organic solvent; b. applying the mixture on the reflective surface; and c. allowing the mixture to cure on the reflective surface.

The isocyanate is preferably as described above.

A 15 - 30 w/w% solution is made by mixing, on weight basis, 1 weight part polyisocyanate with 10 to 50 weight parts of the terpolymer in a suitable solvent such as 2- butanon. Another attractive solution is a mixture of 1 : 10 - 40 polyisocyanate : terpolymer. To this end, both the polyisocyanate and the terpolymer may be provided as separate mixtures and combined together. The solutions can be in a high concentration and be diluted to the envisaged values. For example, the isocyanate can be provided as a 20 w/w% in 2-butanon, and the terpolymer can be provided in a concentration of 40 w/w% in 2-butanon or ethyl benzene or xylene, and both can be mixed in a 1 : 10 ratio, resulting in a 34 w/w% solution that can be further diluted to 20 - 25 w/w% in e.g. acetone. Such solution can optimally be used for reel to reel application on envisaged surfaces.

The invention will now be further explained by way of the following non limiting examples and figures, wherein

Figure 1 is a SEM image of a mesh coated with the terpolymer;

Figure 2 is a graph, showing transmission data of a PE foil coated with the terpolymer;

Figure 3 is a graph, showing reflection data of an aluminium foil coated with the terpolymer;

Figure 4 is a graph, showing reflection data of an aluminium foil coated with the terpolymer and a TPO foil of the art.

Figure 5 is a photograph showing the results of a droplet test. EXAMPLES

Characterisation of the terpolymer

Lamoral 100 series and 400 series (Lamoral Coatings, The Netherlands) are both 40% methylethylketone solutions of a terpolymer of tetrafluoroethylene monomers, ethylene monomers and hydroxy ethylene monomers having a density of 0.9-1.0 kg/I (at 20°C). The atomic content and mole percentage of the atoms present in the terpolymer (apart from hydrogen) can be determined as follows, using the SEM-EDX technique.

Sample preparation

A 130 g/m ² fabric of 86% polyethylene terephthalate (PET) and 14% spandex (Miti, Italy) was coated with Lamoral 400 without the presence of crosslinking agent by dipping the fabric in the terpolymer solution, padding the coated material with a 10 kg roller and drying in ambient air. The weight of the dry coated sample was 112 - 114% of the sample before coating.

The coated fabric was cut in pieces of 1 x 1 cm and rendered electrically conductive by sputter coating the sample with gold, using an Mbraun MB20G vacuum plasma sputter coater with a run time of 2 minutes and a vacuum of 10 ^_1 Pa, and following the instructions of the manufacturer.

Percentages moieties and Mole percent analysis

The samples were analyzed with a Phenom Pharos Desktop SEM (Thermo Fisher Scientific, US) with element identification software package + fully integrated energy dispersive spectrometer (EDS), using a standard backscattered electron detector with a medium vacuum mode, and following the instructions manual of the supplier. The zooming settings were set on analysis 15kV - point. The quality of the image was set high with a high frame rate for live footage as set out in table 1 :

Table 1 : Imaging Specifications

For the analysis of atomic content, Phenom’s integrated EDS software was used. Analyzing the surface composition was done at the same time as when SEM pictures were made (see figure 1). Every time when a picture is taken a full screen EDS surface composition was made. With the EDS composition analysis, the type of element and atomic concentration / weight concentration of the surface were determined at ambient conditions. The specifications of the EDS were as set out in table 2.

Table 2: EDS specifications The atomic content as measured by SEM-EDS is given in table 3. where the values given is an average of multiple measurements. The weight content and the mole percentage were calculated accordingly. An atomic content of 29% for O means that 29% of the atoms of the terpolymer are oxygen atoms. A weight content of 42 % for C means that 42% of the weight of the terpolymer is by carbon atoms. The hydrogen atom content in the terpolymer is not determined. Similar results were obtained for similar samples made with Lamoral 100 series.

Table 3: Atomic and weight content of coating Preparation of crosslinked samples

Lamoral series 100 as described above is mixed with hexamethylene diisocyanate Lamoral C20 (Lamoral Coatings, The Netherlands) a hexamethylene diisocyanate solution in methylethylketone, having a solids content of 20% and a density of 0.8 - 0.9 kg/I (at 20°C), in a weight ratio of 10:1. This mixture was diluted with acetone to 15 - 30% solids.

Several substrates were used, varying from plain aluminium foil (50 pm thickness) and plain PET film.

Sheets having a size varying from DIN A4 (210 x 297 mm) to DIN A6 (105 x 148 mm) were cut from the above-mentioned samples and were placed on the rubber surface of a hand coater set (K101 Hand Coater, RK Print Coat Instruments Ltd., UK). A set of K-bars (RK Print Coat Instruments Ltd., UK) with a closed wound from the firm RK Print Coat instruments were used to apply desired coating weights in g/m ² .

When using K-bar no 2 with a closed wound wire diameter of 0.15 mm, the wet film deposit is approximately 12 pm leading to a dry film deposit of approximately 3 pm when using a solids content of 25%.

To remove the solvent, an air dryer with an outlet temperature of 80°C was used for one minute and the samples were afterwards placed in an oven at 80°C for 5 minutes.

In view of variation in applied coating thickness due to pressure, speed, solids etc. in the manual coating application, the actual coating thickness on several samples of a coated PET film was measured according to ISO 19840, wherein the coating thickness was measured with an Elcometer 456 Ferro layer thickness meter with a Ferro probe (iso 19840). The Elcometer measures the coating thicknesses on metal substrates with an accuracy of 1%. The thickness of coating and substrate was measured and compared with the thickness of the substrate to calculate the coating thickness, see table 4.

Table 4: correlation between theoretical and actual applied coating thickness The coated samples were stored at ambient temperature for 7 days to reach final cure.

UV, Visible and IR transmittance

UV, visible and IR transmittance was measured by coating a polyester film as described above and measuring the transmittance with UV-VIS spectrophotometer LISR- 3100 (Shimadzu, Japan) with the settings shown in table 5 below.

The Polyester film was coated with several coating thicknesses of Lamoral 100 series + C20 as described above. Coating thickness measured was done according to Iso 19840 and Transmission curves obtained as in figure 2.

It can be concluded that the coating is UV stable and protects the film from UV impact, while the difference in IR transmission between a coated film versus an uncoated film is negligible. This means that the coating is inert for radiation from 400 nm and above (visible and IR). Similar results were obtained with samples wherein the terpolymer was Lamoral 400 series, as well as in samples wherein Lamoral C100 was used as a crosslinking agent.

Reflection - aluminium foil

Reflection was measured by coating plain Aluminium foil (50 pm) with Lamoral 100 + C20 series with a thickness of 3 pm as described above and measuring the transmittance with UV-VIS spectrophotometer LISR-3100 (Shimadzu, Japan) with a 150 mm diameter integrated sphere and a 50W deuterium lamp with the settings shown in the table 5. The reflection was measured and compared with a control uncoated aluminium foil sample.

The reflection data, shown in figure 3, show that there is no significant difference in reflection over the full range from UV (300 - 400), visible (400 - 700) and Infrared (>700 nm) between coated and uncoated aluminium foil. This means that the coating does not disturb or interfere with solar radiation - especially the IR radiation - that is reflected from the aluminium foil. This means that the coating is an ideal radiant barrier coating, e.g. for keeping roofs cool by an optimized radiant reflection.

A similar reflection test was done with the above coated aluminium foil over the visible and infrared spectrum and compared with TPO (Versify™, Dow Chemical Company, US). The measurements were performed in conformity with the ASTM E903-96 standard on a Perkin Elmer Lambda 1050 in combination with the 150mm InGaAs Integrating Sphere at TNO, Netherlands. The measurements were performed in the 300 - 2500 nm range in steps of 5nm, see figure 4. The detector switch from InGaAs (NIR) to PMT (UV-Vis) occurred at 860nm. The lamp switch took place at 320nm. The baseline was set against a calibrated white Spectralon standard (barium sulphate). The reflection was measured by placing the samples facing the lamp.

It can be clearly seen from figure 4 that the reflectivity of TPO is low in the infrared spectrum, whereas the Lamoral coated aluminium as a reflectance of 95% or even more in this spectrum. Similar results were obtained with samples wherein the terpolymer was Lamoral 400 series, as well as in samples wherein Lamoral C100 was used as a crosslinking agent.

Table 5: settings UV-VIS spectrophotometer Acid corrosion test

Several solutions of H ₂ SO ₄ were prepared with a pH of 2, 3, 4, 5 by diluting a concentrated H ₂ SO ₄ solution (98%) with demineralised water. Coated aluminium foil as described above was contacted with vapour of the above solutions or brought in direct contact therewith as follows.

The acid solutions were kept in a glass container of 10 ml. The opening was covered with the aluminium foil. On top of the foil, a cap was placed to close the container tightly to avoid water or other molecules to evaporate during the test. For the vapour test, the containers were kept with the foil and lid up, allowing contact of the vapour of the acid solution with the foil. For the direct contact test, the containers were put upside down, bringing the foil in direct contact with the acid solution. This was done for a period of 4 days in an oven at 40°C. The results are shown in table 6. Similar results were obtained with samples wherein the terpolymer was Lamoral 400 series, as well as in samples wherein Lamoral C100 was used as a crosslinking agent.

Table 6: Acid corrosion test

Droplet test

Drops of a H2SO4 solution having a pH of 1 ,64 were contacted with both uncoated aluminium foil, and coated as described above, having a coating thickness of 3 pm. Already after 4 hours at ambient temperature, the uncoated foil was fully corroded and etched away while the coated sample was not corroded or etched even after 24 hours, see figure 5. Similar results were obtained with samples wherein the terpolymer was Lamoral 400 series, as well as in samples wherein Lamoral C100 was used as a crosslinking agent.