Substrate with photocatalytic coating

The invention relates to generally transparent or semi-transparent substrates, in particular glass, plastic, glass-ceramic, which are provided with a coating having photocatalytic properties to give them an anti-fouling or, more precisely, self-cleaning function. .

An important application of these substrates relates to glazing, which can be of very diverse applications, from utility glazing to glazing used in household appliances, glazing for vehicles to glazing for buildings.

It also applies to reflective glazing of the mirror type (home mirror or vehicle rear view mirror) and opaque glazing of the light type.

The invention also applies, similarly, to non-transparent substrates, such as ceramic substrates or any other substrate which can in particular be used as an architectural material (metal, tiles, etc.). It preferably applies, whatever the nature of the substrate, to substantially planar or slightly curved substrates.

Photocatalytic coatings have already been studied, in particular those based on titanium oxide crystallized in anatase form. Their ability to degrade soiling of organic origin or microorganisms under the effect of UV radiation is very interesting. They also often have a hydrophilic character, which allows the evacuation of mineral soiling by spraying water or, for exterior glazing, by rain.

This type of coating with anti-fouling, bactericidal and algicidal properties has already been described, in particular in patent WO97 / 10186, which describes several methods of obtaining it.

The invention therefore aims to improve the deposition techniques for this type of coating, in particular with a view to simplifying them. At the same time, it also aims to improve the appearance of the coating, more particularly to improve the optical properties of the substrate which is provided with it.

The invention firstly relates to a process for the sputtering deposition of a coating with photocatalytic properties comprising titanium oxide at least partially crystallized in anatase form on a transparent or semi-transparent carrier substrate. The characteristic of the invention consists in carrying out the spraying on the substrate, under a deposition pressure of at least 2 Pascals. It is preferably at most 6.67 Pa and in particular at least 2.67 Pa (that is to say at least 15 millitors, in particular between 20 and 50 millitors).

In fact, as is known from the aforementioned patent WO97 / 10186, this type of coating can be deposited by sputtering. It is a vacuum technique which allows, in particular, very fine adjustment of the thicknesses and the stoichiometry of the deposited layers. It is generally assisted by magnetic field for more efficiency. It can be reactive: we start from an essentially metallic target, here based on titanium (possibly alloyed with another metal or silicon), and the spraying is done in an oxidizing atmosphere, generally an Ar / O mixture.₂. It can also be non-reactive, starting from a so-called ceramic target which is already in the oxidized form of titanium (possibly alloyed).

However, the layers obtained by this type of technique are generally amorphous, while the functionality of the coating according to the invention is directly linked to the fact that it must be significantly crystallized. This is the reason why, as recommended in the aforementioned patent, there is a need to crystallize (or increase the crystallization rate) of the coating by subjecting it to a heat treatment, for example of the order of 30 minutes to several hours at least 400 ° C.

It has been shown according to the invention that such a high pressure promotes a particular crystallization of the layer, a level of density / roughness which had a significant impact on the level of the photocatalytic properties of the coating. In some cases, annealing can become optional. To give an idea, the deposition pressures generally used for metallic oxides are usually in the range of 2 to 8 millitors (i.e. 0.27 to 1.07 Pa): the invention therefore chooses deposition pressures entirely - quite unusual in this area.

It has also been shown in the context of the present invention that the post-deposition treatment step could possibly be omitted, or at least make it optional (and / or limit in duration or in temperature), by spraying the layer on the hot substrate, and not at room temperature, in particular at least 100 ° C. This heating during deposition is alternative or cumulative with the use of high pressures mentioned above.

This heating has at least five advantages:<ul id="ul0001" list-style="dash" compact="compact"><li>energy savings during manufacturing,</li><li>the possibility of using substrates which could not withstand heat treatments at temperatures of 400 or 500 ° C at least without degradation,</li><li>in the case where the annealing required to interpose between substrate and photocatalytic coating a barrier layer to the diffusion of elements of the substrate (of the alkaline type when it is a question of glass), the possibility of using a thinner barrier layer, or even to completely remove the barrier layer, since the heat treatment according to the invention is much less aggressive than annealing,</li><li>a much shorter manufacturing cycle (since the heat treatment of the substrate is significantly shorter and at a significantly lower temperature),</li><li>the elimination of storage of "semi-finished" products to be annealed.</li></ul>

However, photocatalytic activity levels are obtained for the coatings quite similar to those of the coatings deposited and then annealed.

However, this was not a gamble won in advance, since it could be expected that prolonged annealing would be essential to gradually grow the germs crystallized within the amorphous oxide matrix. This was not the case: hot deposition promotes the deposition of the directly crystallized layer at least in part.

It was also not obvious that the coating thus deposited "hot" would preferentially crystallize in anatase form rather than in rutile form (the anatase form is much more photocatalytic than the rutile or broockite form of titanium oxide).

To carry out the invention, there are different variants, in particular depending on the type of spraying device available. It is thus possible to heat the substrate prior to the actual deposition, outside of the vacuum enclosure. The substrate can also be heated during deposition, when the deposition chamber is equipped with ad hoc heating means. The substrate can therefore be heated before and / or during spraying of the coating. It can also be progressive during deposition, or affect only part of the thickness of the deposited layer (for example the upper part).

Advantageously, the substrate is, during the spraying of the layer, at a temperature between 150 and 350 ° C, preferably at least 200 ° C and in particular between 210 and 280 ° C. Surprisingly, it was therefore possible to obtain sufficiently crystallized layers without having to heat the substrate to the temperatures generally used for annealing, from at least 400 ° C. to 500 ° C.

Generally, when the coating is essentially based on titanium oxide (TiO₂), and when it is deposited by sputtering ("hot" or at room temperature), it has a fairly high refractive index, greater than 2 or 2.1 or 2.15 or 2.2. It is generally between 2.15 and 2.35 or between 2.35 and 2.50 (it may be slightly sub-stoichiometric), in particular between 2.40 and 2.45. This is a fairly specific characteristic of this type of deposit, since similar coatings deposited by other techniques, for example by sol-gel, tend to be much more porous and to have significantly lower refractive indices ( below 2 and even below 1.8 or 1.7). The invention makes it possible to obtain layers by sputtering which have a porosity and / or roughness (in particular an RMS roughness of between 2.5 and 10 nm amplifying its photocatalytic properties. Therefore, they may have refractive indices of the order of 2.15 or 2.35, lower than those usually obtained by sputtering, indirect proof of their porosity. This is an advantage from an optical point of view, since they have, with a lower refractive index, a less reflective appearance at a given thickness.

It has been observed that the crystallographic structure of the coatings is influenced by the fact that they are deposited cold and then annealed or deposited hot. Thus, quite unexpectedly, the coatings deposited "hot" and / or at high pressure, in accordance with the invention generally have an average size of TiO crystallites₂ generally less than or equal to 50 or 40 or 30 nm, in particular between 15 and 30 nm or between 20 and 40 nm. The coatings deposited in a standard manner, in particular "cold" then annealed, tend to comprise crystallites of larger size, at least 30 nm or 40, generally between 40 and 50 nm when standard deposition pressures are used.

On the other hand, if, according to a variant of the invention, the coating is deposited at room temperature but at high pressure, and an annealing operation is then carried out, the size of the crystallites is of smaller size (20-40 nm), and comparable to that of crystallites from hot-deposited coatings, whether at high or low pressure.

The photocatalytic activity of coatings deposited at room temperature at high pressure and then annealed is much better than that of coatings deposited at room temperature at low pressure and then annealed: all other things being equal, it is clear that the deposition pressure influences the performance of the coating, particularly in the case of "cold" deposition, and this in a striking manner.

Simultaneous heating with the growth of the layer leads to the formation of a microstructure conducive to roughness and / or porosity favorable to a photocatalytic property. This is somewhat the same case when using a high deposition pressure (with a "cold" deposition followed by annealing for example).

Thanks to the method according to the invention (by hot deposition and / or at high pressure), it is possible to obtain coatings having an RMS (Root Mean Square) roughness measured by atomic force microscopy, by making measurements on the same surface with steps of 2 micrometers:<ul id="ul0002" list-style="dash" compact="compact"><li>at least 2 nm, in particular at least 2.5 nm, preferably between 2.8 nm and 4.6 nm in the case of deposition at room temperature at high pressure within the meaning of the invention (2 to 5 Pa), followed by annealing,</li><li>at least 4 nm, in particular at least 5 nm, preferably between 5.5 and 6.0 nm in the case of hot deposition (around 250 ° C) without annealing, whether at high or low pressure .</li></ul>

By way of comparison, the roughness of coatings deposited at room temperature at standard pressure (in particular 2.10-3 millibars, ie 0.2 Pa) and then annealed is only 2 nm at best: this proves that the use of high pressures achieves surprisingly high roughnesses for layers deposited by sputtering, which results in an improvement of the photocatalytic properties of the coating.

Advantageously, the coating has a geometric thickness of less than 150 nm, in particular between 80 and 120 nm or between 10 and 25 nm. It turned out that, even very thin, the coating could have sufficient photocatalytic properties (at least for certain applications), with the additional optical advantage of being not very reflective.

As seen above, sputtering of the coating can be reactive or non-reactive. In either case, the target to be sprayed can be doped, in particular with at least one metal. It can be one or more metals chosen from the following list: Nb, Ta, Fe, Bi, Co, Ni, Cu, Ru, Ce, Mo, Al.

The deposition method according to the invention can be preceded and / or followed by one or more steps of deposition of other thin layer (s), in particular with an optical, anti-static, anti-color function. , anti-reflective, hydrophilic, protective, or to amplify the roughness of the coating with photocatalytic properties. It has thus been observed that it may be advantageous to deposit a (at least) layer so that it is particularly rough, for example by pyrolysis or sot-get, then the photocatatytic coating; the coating then tends to "follow" the roughness of the underlying layer and to actually present, too, a significant roughness, while the layers deposited by sputtering tend to be rather rough. It is thus possible to make stacks with a sublayer (of RMS roughness of for example at least 5 or 10 nm), of SiO type₂, SiOC or SiON deposited by gas phase pyrolysis (CVD), then the photocatalytic layer by sputtering.

The invention therefore includes any combination between the deposition of one or more layers by sputtering (at least the photocatalytic coating) and the deposition of the other layer or layers of the stack by a technique involving thermal decomposition, in particular a pyrolysis (in the liquid, gas or powder phase) or a sol-gel technique.

As we saw above, TiO-based coatings₂ photocatalytic have a high refractive index. This means that they are reflective and that they give their carrier substrate a reflective appearance which is often considered unattractive. In addition, the color in reflection, apart from this brilliant character, can be undesirable. It is not easy to improve this aspect in reflection, because the photocatalytic functionality presents constraints: The coating must generally be in contact with the external atmosphere to receive UV and degrade external dirt. It cannot therefore be overcome with a low index layer (unless it is very thin and / or porous). It must also have a given minimum thickness to be sufficiently effective.

Another aspect of the present invention therefore consisted in improving the appearance in reflection of the substrate, without disturbing the photocatalytic activity of the coating, in particular by lowering its light reflection as well as possible and / or by giving it a color in reflection which is the as neutral as possible.

The invention therefore also relates to the transparent or semi-transparent substrate defined above and which is provided on at least part of at least one of its faces with a photocatalytic coating comprising at least partially crystallized titanium oxide anatase, this coating having a high refractive index of at least 2 or 2.1 or 2.2. According to the invention, this coating is considered to be part of a stack of thin anti-reflective layers, the coating being the last layer (that is to say the layer furthest from the carrier substrate). The antireflection stack is composed of an alternation of high and low index layers, and therefore ends in the present case with the high photocatalytic index layer. This term "anti-reflective" is used for convenience: generally, it is used when one seeks to obtain a light reflection lower than that which one would have with the substrate alone. In the context of the invention, it is more a question of limiting the increase in light reflection (and / or of modifying or attenuating its color in reflection) caused by the use of a coating containing titanium oxide. .

For the purposes of the invention, the term "layer" means a single layer or a superposition of layers. If it is a superposition of layers, it is considered that its overall thickness is the sum of the thicknesses of each of the layers and that its overall index is the average of all the refractive indices of said layers. This also applies to the photocatalytic coating. It can also be combined with another high index layer.

Within the meaning of the invention and as recalled above, the term “antireflection” is understood to mean the function which makes it possible to lower the light reflection value of the coated substrate, and / or to attenuate its color in reflection, in particular to make it more paler and more neutral, more aesthetic possible (this is also called "anti-color" effect).

This is a fairly free and unexpected adaptation of conventional anti-reflective stacks. In fact, in a known manner, these stacks alternate layers with high and low indices and end with layers with low index (as close as possible to the refractive index, equal to 1, of air) and which are usually SiO-based layers₂, from MgF₂... But here, the stack ends with a high index layer, which is quite paradoxical. However, by appropriately selecting the characteristics of the different layers, this particular anti-reflective stack manages to significantly attenuate the reflective character intrinsic to TiO₂ high index, and to give the substrate an acceptable reflection color (neutral, in pale shades avoiding reds and other warm colors, considered unattractive, in favor of gray, blue, or green in particular).

Advantageously, the photocatalytic coating has a refractive index greater than or equal to 2.30, in particular between 2.35 and 2.50, or between 2.40 and 2.45 (as we have seen previously, it is also possible place it so that it has an index of only 2.10 to 2.30). It is preferably deposited by sputtering. Advantageously, its optical thickness is selected, together with those of the other layers of the stack, in order to lower the light reflection from the substrate. It has been shown that an optimal optical thickness is preferably around λ / 2 with λ around 580 nm. This corresponds to an optical thickness between 250 and 350 nm, in particular between 270 and 310 nm; and at a geometric thickness between 80 and 120 nm, in particular between 90 and 110 nm. This range of geometric thickness has proven to be sufficient to obtain, in parallel, a photocatalytic activity considered to be sufficient (the photocatalytic activity in fact depends on many parameters, including the thickness but also the surface roughness, the crystalline morphology of the layer, its porosity, ...). It is also possible to use significantly thinner layers, in particular between 10 and 25 nm in geometric thickness.

Depending on whether the coating is deposited by "hot" spraying or at cold and annealed room temperature, it contains crystallites of variable size as seen above (generally less than 30 nm "hot", and range 30 to 50 nm or more at room temperature at standard pressure , as seen above).

The antireflection stack of the invention, in its simplest embodiment, comprises three layers, including, successively, a high index layer, a low index layer, then the high index photocatalytic coating.

The layer or layers with a high index of the stack, apart from the photocatalytic coating, generally have an index of at least 1.9, in particular between 1.9 and 2.3 or between 1.9 and 2.2. It can be zinc oxide, tin, zirconium, aluminum nitride or silicon nitride. It can also be a mixture of at least two of these compounds.

The optical thickness of these high index layers is selected. Their optimal optical thickness is preferably around λ / 10 with λ around 580 nm. This corresponds to an optical thickness between 48 and 68 nm, especially between 53 and 63 nm, and a geometric thickness between 20 and 40 nm, especially between 25 and 35 nm. It is also possible to choose a lesser thickness, in particular between 20 and 48 nm.

The low index layer (s) generally has an index of between 1.4 and 1.75, especially between 1.45 and 1.65. They can for example be based on silicon oxide, aluminum oxide or a mixture of the two. The optical thickness of these low index layers is selected: their optimal optical thickness is preferably around λ / 20 with λ around 580 nm. This corresponds to an optical thickness between 20 and 79 nm, in particular between 19 and 39 nm, in particular between 25 and 35 nm, and in a geometric thickness between 12 and 50 nm, in particular between 15 and 30 nm, for example between 20 and 28 nm.

According to another variant, in the three-layer stack mentioned above, the high-index layer / low-index layer sequence can be replaced by a layer with an "intermediate" refractive index, that is to say, of preferably greater than 1.65 and less than 1.9. The preferred range of indices is between 1.75 and 1.85. It can be based on silicon and / or aluminum oxynitride. It can also be based on a mixture of a low index oxide such as SiO₂ and at least one oxide with a higher index such as SnO₂, ZnO, ZrO₂, TiO₂. (the relative proportion between the oxides makes it possible to adjust the index).

This intermediate layer can also be used to replace the first high index layer / low index layer sequence of a stack containing not three but five or seven layers for example.

The optical thickness of this intermediate index layer is selected. The optimal optical thickness is around λ / 4 with λ around 580 nm. This corresponds to an optical thickness between 120 and 150 nm, in particular between 125 and 135 nm, and to a geometric thickness between 65 and 80 nm, in particular between 68 and 76 nm.

As mentioned above, these different selections of optical thicknesses take into account the whole of the aspect in reflection of the substrate: an effort is not only made to lower the value of light reflection R_L but also to give it a shade deemed aesthetic nowadays (that is to say rather in cold colors than towards yellow or red) and the least intense possible. It is therefore necessary to find the best compromise so that, as a whole, the aspect in reflection of the substrate is better. Depending on the applications, the lowering of the value of R may be preferred._L or rather the selection of a particular colorimetry in reflection (for example quantified by the values ​​of a * and b * of the colorimetry system L, a *, b * or by the value of dominant wavelength associated with the purity of color).

Advantageously, all of the layers of the antireflection stack can be deposited by sputtering, one after the other, on the same production line.

According to an optional variant of the invention, it is possible to insert between the substrate and the antireflection stack a barrier layer to the species liable to diffuse from the substrate. These include alkalis when the substrate is glass. It is, for example, based on silicon oxide (or oxycarbide): SiO₂ can be deposited by sputtering and SiOC, in known manner, by gas phase pyrolysis (CVD). It preferably has a thickness of at least 50 nm, for example between 80 and 200 nm. Chosen in this type of material, with a relatively low index (around 1.45 to 1.55), it is in fact, generally, largely "neutral" optically. The silicon oxide can contain minority elements, in particular chosen from Al, C, N.

A subject of the invention is also glazing, in particular simple glazing (a rigid substrate), laminated glazing, multiple glazing of the double glazing type and which comprises at least one substrate coated in the manner described above.

Said glazing preferably has, thanks to the anti-reflection effect of the invention, a light reflection R_L (layer side) which remains at most 20%, in particular at most 18%. Preferably, this light reflection has a pleasant tint in blues or greens, with values ​​of a * and b * in the colorimetry system (L, a *, b *) negative and in particular less than 3 or 2.5 in absolute values. The shade is thus a color that is both pleasing to the eye and pale, not very intense.

The glazing may also include one or more other functional coatings (deposited by sputtering or pyrolysis or sol-gel), either on the same face of the substrate provided with the photocatalytic coating, or on the opposite face of this substrate, or on one side of the '' another substrate associated with the first in a glazing (double glazing or laminated glazing). It is also possible to have double glazing of the glass / gas slide / glass type with on the outside face (s) of the glasses the photocatalytic coating and on the internal faces (facing the gas slide) a stack of one or two layers of money. The same type of configuration applies to laminates.

The other functional coating (s) may be, in particular, an anti-fouling, anti-solar, low-emissive, heating, hydrophobic, hydrophilic, anti-reflective, anti-static coating, another photocatalytic coating, etc. in particular, anti-solar or low-emissive stacks with one or more layers of silver, or of nickel / chromium or of titanium nitride or of zirconium. In the case of layers based on metallic nitride, a CVD technique can be used.

The invention will be described below in more detail, with nonlimiting exemplary embodiments.

Comparative examples 1 and 1 relate to the hot deposition of TiO layers₂ photocatalytic by sputtering.

<u style="single">EXAMPLE 1</u>

A first layer of SiOC by CVD, of 80 nm, then a second layer of TiO was deposited on a clear silica-soda-lime glass, 4 mm thick.₂ 90 nm photocatalytic, (we can also replace the SiOC layer with a SiO layer₂ : Al obtained by reactive sputtering from an Al doped Si target).

The TiO layer₂ was deposited by sputtering assisted by magnetic field. It is a reactive spray, in the presence of oxygen from a titanium target. The glass is preheated to a temperature of about 220 ° C to 250 ° C. This temperature is kept constant to within 5 ° during the spraying of the layer, using a heating device placed opposite the target.

The TiO layer₂ obtained at a refractive index of 2.44. It is crystallized in anatase form (it can also include amorphous zones), with an average size of crystallites less than 25 nm.

Its photocatalytic activity has been quantified using a test using palmitic acid: This involves depositing a given thickness of palmitic acid on a photocatalytic coating, to expose it to UV radiation centered on 365 nm with a surface power of approximately 50 W / m2 for the entire duration of the test, then measuring the rate of disappearance of palmitic acid according to the following relationship: <maths id="math0001"><math display="block"><mrow><mi mathvariant="normal">V</mi><mrow><mo>(</mo><mi mathvariant="normal">nm</mi><mn>.</mn><msup><mrow><mi mathvariant="normal">h</mi></mrow><mrow><mo>-</mo><mn>1</mn></mrow></msup><mo>)</mo></mrow><mo>=</mo><mrow><mo>[</mo><mi mathvariant="normal">palmitic acid thickness</mi><mrow><mo>(</mo><mi mathvariant="normal">nm</mi><mo>)</mo></mrow><mo>]</mo></mrow><mo>/</mo><mrow><mo>[</mo><mn>2</mn><mo>×</mo><msub><mrow><mi mathvariant="normal">t</mi></mrow><mrow><mn>1</mn><mo>/</mo><mn>2</mn></mrow></msub><mi mathvariant="normal"> disappearance</mi><mrow><mo>(</mo><mi mathvariant="normal">h</mi><mo>)</mo></mrow><mo>]</mo></mrow></mrow></math><img file="EP1679389A2_D0001.tif" /></maths>

With the layer according to the invention, a photocatalytic activity of at least 10 nm.h is obtained by this calculation.^-1, especially at least 20 nm.h^-1, especially between 20 and 100 nm.h^-1, depending on the choice of deposition parameters of the pressure, temperature type.

The glass thus coated with the two layers has, according to Illuminant D₆₅, a light reflection R_L 23%, with values ​​of a * and b * in reflection according to the colorimetry system (L, a *, b *) of the order of 17 and 28 respectively.

The photocatalytic activity of the layer is therefore interesting, but its optical appearance is still clearly reflecting, with too intense a color.

Note that it is possible to increase the photocatalytic activity of the layer by subjecting it, after deposition, to conventional annealing (of one or more hours at at least 400 ° C.).

<u style="single">COMPARATIVE EXAMPLE 1</u>

Example 1 is repeated, but this time the TiO layer₂ is deposited on an unheated substrate, then treated for four hours at approximately 500 to 550 ° C. In addition, the SiO underlay₂ is thickened up to 100 nm. The morphology of the layer is a little different, with an average size of crystallites rather greater than 30 nm.

Its photocatalytic activity is similar to that of the layer of Example 1 without annealing, but it is lower if a smaller thickness of SiO sublayer is chosen.₂.

This therefore confirms that the "hot" deposition according to the invention, making it possible to "save" an often long annealing operation, is not obtained at the expense of the performance of the layer. This also confirms a subsidiary advantage of the invention: by hot depositing, by avoiding annealing, it is possible to use, with identical photocatalytic performances, a thinner barrier sublayer (hence, here again, a time and a cost reduced manufacturing).

Examples 2 and following relate to the incorporation of a photocatalytic layer into TiO₂ with high index, in particular deposited by cathode sputtering, in antireflection stacks to improve their optical properties.

<u style="single">EXAMPLE 2 - (REALIZED)</u>

The following stack of layers is deposited on a 4 mm thick silica-soda-lime float glass:<maths id="math0002"><math display="block"><mrow><mtable><mtr><mtd><mi mathvariant="normal">Glass</mi><mo>/</mo></mtd><mtd><msub><mrow><mi mathvariant="normal">Yes</mi></mrow><mrow><mn>3</mn></mrow></msub><msup><mrow><msub><mrow><mi mathvariant="normal">NOT</mi></mrow><mrow><mn>4</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><msup><mrow><mtable><mtr><mtd><msub><mrow><mi mathvariant="normal">SiO</mi></mrow><mrow><mn>2</mn></mrow></msub></mtd></mtr></mtable></mrow><mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msup><mrow><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></msup></mtd><mtd><mi mathvariant="normal"> </mi></mtd></mtr><mtr><mtd><mi mathvariant="normal"> </mi></mtd><mtd><mn>30</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>22</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>104</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mrow><mo>(</mo><mtable><mtr><mtd><mi mathvariant="normal">geometric thicknesses</mi></mtd></mtr></mtable><mo>)</mo></mrow></mtd></mtr></mtable></mrow></math><img file="EP1679389A2_D0002.tif" /></maths>

The Si layer (1)₃NOT₄ is deposited by reactive sputtering in the presence of nitrogen from an Al doped Si target.

The layer (2) of SiO₂ is deposited by reactive sputtering in the presence of oxygen from an Al doped Si target.

TiO layer (3)₂ is photocatalytic and was hot deposited as described in Example 1.

Optionally, an additional layer can be inserted between the glass and the Si layer₃NOT₄, a layer of SiO₂ about 100 nm obtained as the other layer of SiO₂ (2) described above. It has almost no influence on the optical properties of the substrate and can serve as a barrier layer against alkalis with respect to glass. It is optional, especially since the layers of the antireflection coating under the photocatalytic layer, namely the layers (1) and (2) themselves constitute barrier layers which are entirely satisfactory, in addition to their optical properties: these two layers already form a 100 nm barrier to species capable of diffusing out of the glass.

The photocatalytic activity of layer 3 is 80 nm.h^-1.

Alternatively, a layer of TiO can be used₂ cold deposited then annealed as described in Comparative Example 1.

In reflection on the layer side, the result for such a stack is as follows:<dl id="dl0001" compact="compact"><dt>R_L (according to illuminant D₆₅):</dt><dd>17,3%</dd><dt>a * (R_L)=</dt><dd>-2</dd><dt>b * (R_L)=</dt><dd>-2,8</dd><dt>λ_d(nm) =</dt><dd>494 nm (dominant wavelength of light reflection)</dd><dt>ρe (%) =</dt><dd>2.5% (purity of color in reflection).</dd></dl>

We see, compared to Example 1, a significant drop in the value of R_L, we get here a color in blue-green, rather pale. Overall, we therefore have an aspect in reflection aesthetically and significantly improved.

<u style="single">EXAMPLE 3</u>

It is very close to Example 2, only the thickness of the TiO layer changes a little₂.

Here we have:<maths id="math0003"><math display="block"><mrow><mtable><mtr><mtd><mi mathvariant="normal">Glass</mi><mo>/</mo></mtd><mtd><msub><mrow><mi mathvariant="normal">Yes</mi></mrow><mrow><mn>3</mn></mrow></msub><msup><mrow><msub><mrow><mi mathvariant="normal">NOT</mi></mrow><mrow><mn>4</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><msup><mrow><mtable><mtr><mtd><msub><mrow><mi mathvariant="normal">SiO</mi></mrow><mrow><mn>2</mn></mrow></msub></mtd></mtr></mtable></mrow><mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msup><mrow><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></msup></mtd><mtd><mi mathvariant="normal"> </mi></mtd></mtr><mtr><mtd><mi mathvariant="normal"> </mi></mtd><mtd><mn>30</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>22</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>99</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mrow><mo>(</mo><mtable><mtr><mtd><mi mathvariant="normal">geometric thicknesses</mi></mtd></mtr></mtable><mo>)</mo></mrow></mtd></mtr></mtable></mrow></math><img file="EP1679389A2_D0003.tif" /></maths>

The result in light reflection is as follows (with the same conventions as for example 2):<dl id="dl0002" compact="compact"><dt>R_L =</dt><dd>17,9 %</dd><dt>a * =</dt><dd>-0,8</dd><dt>b * =</dt><dd>-0,7</dd><dt>λ_d(nm) =</dt><dd>494 nm</dd><dt>ρe (%) =</dt><dd>0,8 %</dd></dl>

So here we have a slightly different compromise, with a value of R_L slightly higher but values ​​of a * and b * lower in absolute values.

<u style="single">EXAMPLE 4 - (MODELING)</u>

It is very close to Example 2, only changes the thickness of the first layer in Si₃NOT₄ :<maths id="math0004"><math display="block"><mrow><mtable><mtr><mtd><mi mathvariant="normal">Glass</mi><mo>/</mo></mtd><mtd><msub><mrow><mi mathvariant="normal">Yes</mi></mrow><mrow><mn>3</mn></mrow></msub><msup><mrow><msub><mrow><mi mathvariant="normal">NOT</mi></mrow><mrow><mn>4</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><msup><mrow><mtable><mtr><mtd><msub><mrow><mi mathvariant="normal">SiO</mi></mrow><mrow><mn>2</mn></mrow></msub></mtd></mtr></mtable></mrow><mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msup><mrow><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></msup></mtd><mtd><mi mathvariant="normal"> </mi></mtd></mtr><mtr><mtd><mi mathvariant="normal"> </mi></mtd><mtd><mn>25</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>22</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>104</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mrow><mo>(</mo><mtable><mtr><mtd><mi mathvariant="normal">geometric thicknesses</mi></mtd></mtr></mtable><mo>)</mo></mrow></mtd></mtr></mtable></mrow></math><img file="EP1679389A2_D0004.tif" /></maths>

The result in light reflection is as follows (always with the same conventions):<dl id="dl0003" compact="compact"><dt>R_L =</dt><dd>15,8 %</dd><dt>a * =</dt><dd>0</dd><dt>b * =</dt><dd>-9</dd><dt>λ_d(nm) =</dt><dd>475 nm</dd><dt>ρe (%) =</dt><dd>4,9 %</dd></dl>

Here we have greatly lowered the value of R_L, but the color in reflection has changed color.

<u style="single">EXAMPLE 5 - (MODELING / COMPARATIVE)</u>

Here, compared to Example 2, all the thicknesses change. We have :<maths id="math0005"><math display="block"><mrow><mtable><mtr><mtd><mi mathvariant="normal">Glass</mi><mo>/</mo></mtd><mtd><msub><mrow><mi mathvariant="normal">Yes</mi></mrow><mrow><mn>3</mn></mrow></msub><msup><mrow><msub><mrow><mi mathvariant="normal">NOT</mi></mrow><mrow><mn>4</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><msup><mrow><mtable><mtr><mtd><msub><mrow><mi mathvariant="normal">SiO</mi></mrow><mrow><mn>2</mn></mrow></msub></mtd></mtr></mtable></mrow><mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msup><mrow><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></msup></mtd><mtd><mi mathvariant="normal"> </mi></mtd></mtr><mtr><mtd><mi mathvariant="normal"> </mi></mtd><mtd><mn>28</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>30</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>75</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mrow><mo>(</mo><mtable><mtr><mtd><mi mathvariant="normal">geometric thicknesses</mi></mtd></mtr></mtable><mo>)</mo></mrow></mtd></mtr></mtable></mrow></math><img file="EP1679389A2_D0005.tif" /></maths>

The result in light reflection is as follows:<dl id="dl0004" compact="compact"><dt>R_L =</dt><dd>25,8 %</dd><dt>a * =</dt><dd>-0,3</dd><dt>b * =</dt><dd>-0,7</dd><dt>λ_d(nm) =</dt><dd>492 nm</dd><dt>ρe (%) =</dt><dd>0,5 %</dd></dl>

If the substrate has a satisfactory reflection color, on the other hand it has a value of R_L well above 20% which is too high: the thicknesses chosen are not optimal.

<u style="single">EXAMPLE 6 - (MODELING / COMPARATIVE)</u>

Here we move further away from the layer thicknesses recommended by the invention, with the following stacking:<maths id="math0006"><math display="block"><mrow><mtable><mtr><mtd><mi mathvariant="normal">Glass</mi><mo>/</mo></mtd><mtd><msub><mrow><mi mathvariant="normal">Yes</mi></mrow><mrow><mn>3</mn></mrow></msub><msup><mrow><msub><mrow><mi mathvariant="normal">NOT</mi></mrow><mrow><mn>4</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><msup><mrow><mtable><mtr><mtd><msub><mrow><mi mathvariant="normal">SiO</mi></mrow><mrow><mn>2</mn></mrow></msub></mtd></mtr></mtable></mrow><mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msup><mrow><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></msup></mtd><mtd><mi mathvariant="normal"> </mi></mtd></mtr><mtr><mtd><mi mathvariant="normal"> </mi></mtd><mtd><mn>20</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>20</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>60</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mrow><mo>(</mo><mtable><mtr><mtd><mi mathvariant="normal">geometric thicknesses</mi></mtd></mtr></mtable><mo>)</mo></mrow></mtd></mtr></mtable></mrow></math><img file="EP1679389A2_D0006.tif" /></maths>

The result in light reflection is as follows:<dl id="dl0005" compact="compact"><dt>R_L =</dt><dd>30 %</dd><dt>a * =</dt><dd>2,3</dd><dt>b * =</dt><dd>7,2</dd><dt>λ_d(nm) =</dt><dd>587 nm</dd><dt>ρe (%) =</dt><dd>14 %</dd></dl>

It has both a value of R_L very high, a color in reflection little sought after and more intense. Its reflection aspect is therefore not satisfactory.

<u style="single">EXAMPLE 7 - (REALIZED)</u>

The stack is this time:<maths id="math0007"><math display="block"><mrow><mtable><mtr><mtd><mi mathvariant="normal">Glass</mi><mo>/</mo></mtd><mtd><mi mathvariant="normal">Sn</mi><msup><mrow><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><msup><mrow><mtable><mtr><mtd><msub><mrow><mi mathvariant="normal">SiO</mi></mrow><mrow><mn>2</mn></mrow></msub></mtd></mtr></mtable></mrow><mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msup><mrow><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></msup></mtd><mtd><mi mathvariant="normal"> </mi></mtd></mtr><mtr><mtd><mi mathvariant="normal"> </mi></mtd><mtd><mn>30</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>27</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>105</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mrow><mo>(</mo><mtable><mtr><mtd><mi mathvariant="normal">geometric thicknesses</mi></mtd></mtr></mtable><mo>)</mo></mrow></mtd></mtr></mtable></mrow></math><img file="EP1679389A2_D0007.tif" /></maths>

So we replaced the Si₃NOT₄ SnO₂, deposited by reactive sputtering in the presence of oxygen from a tin target.

The result in light reflection is as follows:<dl id="dl0006" compact="compact"><dt>R_L =</dt><dd>17,4 %</dd><dt>a * =</dt><dd>-2,8</dd><dt>b * =</dt><dd>-2,7</dd><dt>λ_d(nm) =</dt><dd>496 nm</dd><dt>ρe (%) =</dt><dd>2,8 %</dd></dl>

The aspect in reflection is close to that obtained in Example 2.

<u style="single">EXAMPLE 8 - (MODELED)</u>

Here, the first two layers are replaced by a single layer of index 1.84 in silicon oxynitride SiON.

So we have the stack:<maths id="math0008"><math display="block"><mrow><mtable><mtr><mtd><mi mathvariant="normal">Glass</mi><mo>/</mo></mtd><mtd><mi mathvariant="normal">If we</mi><mo>/</mo><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mtd><mtd><mi mathvariant="normal"> </mi></mtd></mtr><mtr><mtd><mi mathvariant="normal"> </mi></mtd><mtd><mn>72</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>101</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mrow><mo>(</mo><mtable><mtr><mtd><mi mathvariant="normal">geometric thicknesses</mi></mtd></mtr></mtable><mo>)</mo></mrow></mtd></mtr></mtable></mrow></math><img file="EP1679389A2_D0008.tif" /></maths>

The result in light reflection is as follows:<dl id="dl0007" compact="compact"><dt>R_L =</dt><dd>17,4 %</dd><dt>a * =</dt><dd>0</dd><dt>b * =</dt><dd>-1,08</dd><dt>λ_d(nm) =</dt><dd>480 nm</dd><dt>ρe (%) =</dt><dd>14 %</dd></dl>

The aspect in reflection is therefore satisfactory.

<u style="single">EXAMPLE 9 - (MODELED)</u>

He replicates Example 8, but with an index of 1.86 for the SiON layer.

The aspect in reflection is slightly modified:<dl id="dl0008" compact="compact"><dt>R_L =</dt><dd>17,8 %</dd><dt>a * =</dt><dd>-1,1</dd><dt>b * =</dt><dd>-1,5</dd><dt>λ_d(nm) =</dt><dd>494 nm</dd><dt>ρe (%) =</dt><dd>1,3 %</dd></dl>

<u style="single">EXAMPLE 10 (REALIZED)</u>

We have the stacking:<maths id="math0009"><math display="block"><mrow><mtable><mtr><mtd><mi mathvariant="normal">Glass</mi><mo>/</mo></mtd><mtd><msub><mrow><mi mathvariant="normal">Yes</mi></mrow><mrow><mn>3</mn></mrow></msub><msub><mrow><mi mathvariant="normal">NOT</mi></mrow><mrow><mn>4</mn></mrow></msub><mo>/</mo></mtd><mtd><mtable><mtr><mtd><msub><mrow><mi mathvariant="normal">SiO</mi></mrow><mrow><mn>2</mn></mrow></msub></mtd></mtr></mtable><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msup><mrow><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></msup></mtd></mtr><mtr><mtd><mi mathvariant="normal"> </mi></mtd><mtd><mn>24</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>17</mn><mo>,</mo><mn>5</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>24</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>92</mn><mo>,</mo><mn>5</mn><mi mathvariant="normal"> nm</mi></mtd></mtr></mtable></mrow></math><img file="EP1679389A2_D0009.tif" /></maths>

The last high index "layer" is therefore the superposition of a layer of Si₃NOT₄ and TiO₂. Light reflection on the R layer side_L is between 16.5 and 17.5%. Photocatalytic activity is around 80 nm.h-1.

<u style="single">EXAMPLE 11 (REALIZED)</u>

It uses the type of stacking from Example 3, with different thicknesses. It is :<maths id="math0010"><math display="block"><mrow><mtable><mtr><mtd><mi mathvariant="normal">Glass</mi><mo>/</mo></mtd><mtd><msub><mrow><mi mathvariant="normal">Yes</mi></mrow><mrow><mn>3</mn></mrow></msub><msup><mrow><msub><mrow><mi mathvariant="normal">NOT</mi></mrow><mrow><mn>4</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><msup><mrow><mtable><mtr><mtd><msub><mrow><mi mathvariant="normal">SiO</mi></mrow><mrow><mn>2</mn></mrow></msub></mtd></mtr></mtable></mrow><mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></mrow></msup><mo>/</mo></mtd><mtd><mi mathvariant="normal">Ti</mi><msup><mrow><msub><mrow><mi mathvariant="normal">O</mi></mrow><mrow><mn>2</mn></mrow></msub></mrow><mrow><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow></msup></mtd></mtr><mtr><mtd><mi mathvariant="normal"> </mi></mtd><mtd><mn>14</mn><mo>,</mo><mn>5</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>43</mn><mi mathvariant="normal"> nm</mi></mtd><mtd><mn>14</mn><mo>,</mo><mn>5</mn><mi mathvariant="normal"> nm</mi></mtd></mtr></mtable></mrow></math><img file="EP1679389A2_D0010.tif" /></maths>

the light reflection on the layer side is between 13 and 16%. The optical variations of the substrate thus coated, if the layers of the stack are varied by 3%, are the following:<dl id="dl0009" compact="compact"><dt>ΔR_L:</dt><dd>0,8%</dd><dt>Δa * (R_L):</dt><dd>0,3</dd><dt>Δb * (R_L):</dt><dd>1,3</dd></dl>

This example has a photocatalytic activity of approximately 15 to 20 nm.h^-1.

This example is interesting for several reasons: it is very little sensitive to variations in thickness, so it will be easy to produce industrially. It remains sufficiently photocatalytic, although the titanium oxide layer is very thin. It is satisfactory from a colorimetric point of view.

In conclusion, the invention has developed a new mode of vacuum deposition of layers comprising TiO₂ photocatalytic. It has also developed a new type of anti-reflective / anti-color stack ending in a high-index layer, a stack that is simple to produce industrially and which considerably reduces the reflective appearance of TiO₂ without degrading the photocatalytic properties. It makes it possible to obtain glazing in blues or in pale greens in reflection, while retaining substantial thicknesses of photocatalytic layer, of the order of a hundred nanometers. The choice of a significantly thinner photocatalytic layer, 12-30 nm, is also possible.

The invention in its two aspects (product and process) can be applied in the same way to photocatalytic coatings which contain not only TiO₂.

The invention therefore provides "hot" deposits of these coatings, and, alternatively, deposits at room temperature followed by appropriate heat treatments, preferably with particular control of the deposition pressure, in order to obtain layers deposited under vacuum having completely unusual characteristics, resulting in remarkable anti-fouling properties.