Self-cleaning glass and method of making thereof

SELF-CLEANING GLASS AND METHOD OF MAKING THEREOF

0002Field of Invention

0003This invention relates to the photocatalytic oxidative stripping of organic contaminants from the surface of glass and the process of making such photocatalytic glass.

0004Background ofthe Invention

0005It has been known for some time that photocatalysts, particularly titanium dioxide in the anatase phase, accelerate the air-oxidation of organic compounds upon exposure to light, usually ultraviolet, absorbed by the photocatalyst. See, for example, Photocatalytic Purification of Water and Air, D.F. Ollis and H. Al-Ekabi, eds., Proceedings ofthe First International Conference on TiO₂ Photocatalytic Purification and Treatment of Water and Air, London, Ontario, Canada, 1993, Elsevier, Amsterdam. The major sections ofthe book describe the theory and fundamentals of titanium dioxide

0006(TiO₂) photocatalysis, photocatalyzed water and air treatment, reactor design and photocatalytic oxidation process economics. In all of these applications the photocatalyst is bound to a ceramic substrate to which it adheres. For example, on page 123, R. W. Mathews describes TiO₂ coated glass mesh and TiO₂ coated glass tube-based photoreactors. Another reactor with a TiO₂ coated glass tube is described by T. Ibusuki et al. on page 376. The photocatalytic films described were all light scattering, as they were made of photocatalysts such as Degussa P25, e.g., with an abundance of titanium dioxide particles approximately 0.1-0.3 microns in diameter. This particle size, even in the thinnest films, produces a milky appearance. Such light scattering films are not efficient or useful in applications such as clear, self-cleaning glass surfaces for windows and mirrors.

0007It is also known that clear and adherent, non-light scattering TiO₂ films can be made. Such known films are applied to optical lenses of optical instruments to provide scratch resistance and are also applied in anti-reflective optical coatings, usually by reactive evaporation or by reactive sputtering of titanium in an oxygen-containing atmosphere. Such coatings can also be made by applying a solution containing a precursor of a photocatalyst, e.g., TiO₂, to a glass surface, forming a precursor film and heating to a high temperature where organic matter in the precursor film is oxidized and TiO₂ is crystallized. However, on substrates consisting of glasses comprising more than about 10% by weight of combined alkali metal oxides, particularly sodium oxide (Na₂O) and potassium oxide (K₂O), these films are poor photocatalysts, i.e., when exposed to sunlight in air they do not oxidize organic contaminants at a rate adequate to maintain a clean glass surface. For example, when coated with a film of stearic acid, they oxidize it under 2.4 m Wcm^"2 365 nm irradiance at a rate of 4nm per hour or less, i.e., reduce the thickness ofthe stearic acid film by less than 4nm per hour. Under the same conditions, a good photocatalyst strips a film of stearic acid at a rate of about 20nm per hour or more, i.e., reduces the thickness ofthe stearic acid film by about 20 nm per hour (or more).

0008It has now been found that migration of sodium into the photocatalytic film, (e.g., leaching out of common soda lime glass), particularly during the formation of the film from precursors, results in severely reduced activity ofthe photocatalytic films and reduced photocatalytic efficiency of self-cleaning glass. It would be highly advantageous to provide a means for reducing the deleterious effects of migrating alkali metal oxides and/or sodium on the photoactivity of self-cleaning, photocatalytic film- coated glasses.

0009Summary ofthe Invention We have discovered a barrier to the migration of alkali metal oxides from a glass substrate into the photocatalytic coating that is thin, yet effective. This barrier layer slows or blocks the migration of alkali metal oxides into the photocatalyst layer during its formation from a precursor and also after its formation from the precursor.

0010The preferred barrier is formed by first introducing into the near surface region ofthe glass to be coated, protons, i.e. hydrogen ions, by exchanging alkali metal ions with protons of an acid and/or by hydrolytic cleavage of silicon-oxygen-silicon bonds with an acid, in a process called "acid etching" or simply "etching". The hydrogen or proton-containing glass layer is then reacted with a precursor of an oxide ofa four-valent element, preferably a precursor of crystalline titania or zirconia comprising inorganic oxide. In this process, a thin, sodium-migration blocking layer comprising titanium, silicon and oxygen and/or zirconium, silicon and oxygen is formed. A sodium migration blocking layer is also formed when the proton or hydrogen-containing glass (acid glass), is reacted with a precursor of silica. A preferred process of forming the acid glass layer includes etching with acid, most preferably boiling in 9 M sulfuric acid. The sodium migration reducing barrier layer is preferably formed upon heating the acid glass with a precursor of titania, and/or zirconia to a temperature in excess of about 300°C and less than about 500°C and preferably about 400°C.

0011The photocatalytic activity ofa formed, optically clear TiO₂ layer on glass is also enhanced by treating the TiO₂ coated glass with a material that reacts with sodium oxide, particularly with a dilute acid that does not dissolve TiO₂ in the form in which it is included in the coating. In general, the anion of an acid that does not dissolve TiO₂ does not form a strong complex with four valent titanium. These types of acids include protic acid, Lewis acid and Bronsted acid, and can be in liquid or gas form. For example, nitric, perchloric, and tetrylfluorboric acids are useable in the invention, as they are not known to form complexes with four valent titanium. Chlorides, fluorides and sulfates are generally not useable in the invention because they are known to strongly complex with titanium (TV).

0012The imposition ofa sodium migration blocking layer on the surface ofa sodium containing glass, e.g., common soda lime glass, assures formation ofthe desired photocatalytic anatase phase by prohibiting migration of sodium from the glass and into the precursor ofthe photocatalytic film. Tms barrier also reduces sodium contamination ofthe anatase phase that lowers the photocatalytic activity. A particularly useful sodium migration blocking layer is formed by applying a film of an organotitanate that decomposes upon heating in air, a nascent, yet non-crystalline precursor of anatase TiO₂, and reacting the precursor with the sodium-depleted acid glass prior to final calcining of the coated glass and at an elevated temperature, preferably about 450°C. After final calcination, a transparent, non-scattering, adherent nanocrystalline photocatalytic oxide film is produced, with a distinct, sodium migration-blocking interface between the glass and the photocatalytic film.

0013The photocatalyst coated glass ofthe invention, when exposed to ultraviolet light, efficiently cleans itself of organic contaminants. This photocatalyst coated glass ofthe invention is particularly useful in applications such as photocatalytically self-cleaning windows, windshields and mirrors. Brief Description ofthe Figures

0014Figure 1 is a graph showing the UV absorption spectra of TiO₂ films on fused silica.

0015Figure 2 is FTIR spectra of stearic acid coated on clear TiO₂ film on fused silica prior to (dotted line) and after (solid line) exposure to UVA light for 7.5 minutes.

0016Figure 3 is a graph showing the UV absorption spectra of TiO₂ films on etched (dashed line) and on non-etched soda lime glass (solid line).

0017Figure 4 is a graph showing the effect ofthe calcination temperature on UVA, i.e., near UV, photoactivity of clear films of titanium dioxide:(A) Two layers of TiO₂ on fused silica, (B) Two layers of TiO₂ on etched glass, (C) One layer of TiO₂ on etched glass, (D) One layer of TiO₂ on one layer of ZrO₂ on etched glass.

0018Figure 5 is a graph showing the effect of etching duration on the photoefficiency of clear TiO₂ films on soda lime glass.

0019Detailed Description ofthe Preferred Embodiment The invention is directed toward self-cleaning glass and methods of making the same. Photocatalytic films can form the basis for self-cleaning or photooxidatively cleaning glass, useful, for example, as self-cleaning windows, mirrors, optical components, eyeglass lenses, and automotive windshields. When using photocatalyst films, e.g. TiO₂, in these applications, the following should be optimized: the absence of scattering of visible light; abrasion resistance to an extent that the film is typically not damaged when cleaned or when impacted by dust particles; and an adequate photooxidation rate or efficiency in order to maintain a relatively clean glass surface. The coated self-cleaning glasses ofthe invention provide an abrasion resistant, photoefficient, optically clear, self-cleaning glass.

0020Deposition of organic contaminants on glass usually reduces visibility. Furthermore, light particularly from headlights of oncoming cars and from the sun when the sun is low on the horizon, interferes with driving when the windshield is contaminated. Films of organic contaminants smeared on the windshield by wipers operating in rain also add to the hazard of driving. Light scattered from contaminants on outside or inside rearview mirrors of cars impairs visibility in these mirrors. Dirt or fingerprints on lenses or eyeglasses impair vision. The inventive films maintain surfaces clean of organic contaminants. Inorganic, non-oxidizable contaminants are readily removed by being blown or washed off, once the organic matter that makes them stick to the glass surface is oxidized. The most preferred method for preparing photocatalytic glass is by acid etching the glass followed by application ofthe photocatalyst composition and then calcination.

0021Photocatalytic Film The photocatalytic film coated glasses ofthe invention, e.g. containing photocatalytic particles and adherent to glass, are optically clear. They may have a tint or color, but do not absorb or scatter visible light so as to impair visibility through the glass or cause severe glare. The photocatalytic films also adhere to glass and resist abrasion. An inventive film adhering to glass cannot be removed by pressing adhesive tape against it, as discussed below in Example 1 , and rapidly pulling on it (Tape Test). An abrasion resistant film ofthe invention is not damaged when cleaned with wet or dry paper or cloth, and it typically is not scribed by a pencil of H2 hardness or softer. In general, the photocatalytic film is formed of photocatalytic particles, e.g., TiO₂, cast onto acid glass and calcined for specific adherence to the glass. The photocatalyst-containing films ofthe invention contain a material that, upon exposure to light, particularly UV light, accelerates the oxidation in air of organic compounds absorbed or deposited on the film. One example of such a film is a film containing crystalline, preferably anatase, titanium dioxide. The film is well bound to glass in the present invention, generally through an intermediate barrier layer that prevents migration of alkali metal oxides, e.g. sodium, yet is transparent to visible light.

0022Glass

0023The glasses useful in the present invention have varied compositions. The most commercially important and most common glasses comprise sodium and calcium ions and have a network formed of bonds of silicon and oxygen atoms.

0024The self-cleaning glasses of this invention are usually photocatalytic films cast on common glass, which is formed of silicon dioxide, alkali metal oxides (oxides of Column I metals ofthe periodic table), particularly sodium and potassium, and oxides of alkaline earth metals (oxides of Column 2 metals ofthe periodic table), particularly calcium. In general, the ratio ofthe number of oxygen atoms to the number of silicon atoms is between 2.2 and 2.7 in the glasses ofthe invention. In the preferred glasses, the ratio of oxygen atoms to silicon atoms is between 2.2 and 2.5. The glasses may contain other oxides, such as oxides of trivalent or tetravalent rare earths or oxides of aluminum, boron, antimony, germanium, lead and tin. The glasses are transparent, meaning that they can be seen through and are useful as windows, mirrors, windshields, and the like. An example of glass commonly used in such applications and useful in the invention is soda lime glass.

0025The method ofthe invention is intended to prevent migration of alkali metal oxides from glass, particularly of sodium and/or potassium, into the precursor ofthe photocatalytic film and/or into the photocatalyst film by creating a barrier to the migration. When excessive sodium migration is prevented, the photocatalytic film formed upon calcining of its precursor, can efficiently clean its surface of oxidizable contaminants such as carbon-rich organic films.

0026Glass that is commercially used in windows, mirrors, and optical lenses typically contains significant amounts of alkali metal oxides, such as sodium oxide and potassium oxide (usually at least about 10% by weight). Thus, the methods and coated glasses ofthe present invention provide for the first time an economically feasible self- cleaning clear glass.

0027Photocatalyst

0028Photocatalysts useful in this invention are generally photoconductors or semiconductors having band gaps greater than 2.5eV and smaller than 4.5eV. The photocatalytic films are generally less than one micron thick, preferably about 40-80 nm thick, and consist of sufficiently small particles to avoid scattering of visible light. Alternatively, the crystallites ofthe photocatalytic particles are densely packed and oriented so that they do not scatter visible light. The preferred photocatalysts are crystalline oxides, particularly crystalline oxides comprising titanium, tin, tungsten or molybdenum. A particularly useful photocatalyst is titanium dioxide in the anatase phase. Other photocatalysts include TiO₂ with co-catalysts such as Pt, Pd, Au, Ag, Cu, W, Mo, or their sulfides and oxides; compound oxides such as (SrTiO₃) or CaTiO₃, and TiO₂ in the rutile phase or in the mixed anatase and rutile phases. While a preferred photocatalyst, titanium dioxide, is exemplified herein, it is understood that other photocatalysts, e.g., those described above, forming clear films on glass may also be used. Titanium dioxide in the anatase or the rutile phases has an index of refraction of visible (yellow) light greater than about 2.4. Coating of photocatalyst films on glass having a low refractive index, e.g., soda lime glass which has a refractive index below 1.6, causes an increase in the refractive index. In the specific case of automotive windshields such an increase can be undesirable, because as the angle between the dashboard and the windshield is reduced or the index of refraction ofthe windshield is raised, the reflected image ofthe dashboard becomes visible to the driver looking through the windshield. Such reflection is reduced by forming the photocatalytic film ofa combination ofthe photocatalyst and a material having a lower refractive index than that ofthe photocatalyst. An example of such a film is one comprising non-crystalline silicon dioxide (SiO₂) and anatase or rutile TiO₂. By way of example, if the film contains up to 90 weight% SiO₂ then the index of refraction, in the visible, is only about 1.46. (As compared with 2.4-2.7 in the absence of SiO₂.)

0029Photocatalyst Precursor In the present invention, a photocatalyst precursor is generally a film formed of a non-crystalline, three, four, or five- valent element, preferably of an oxide of such an element, which film forms an active photocatalytic film, eg., on calcining in air. The oxide is non-volatile at about 600°C, and the preferred three, four, or five-valent elements are titanium, tin, tungsten, or molybdenum. Most preferred is Ti ⁺. The photocatalyst-precursor films can be formed by their deposition from a liquid phase, or from a vapor phase. Useful photocatalytic precursor compositions include alkoxides, halides and oxyhalides of titanium, tin, tungsten, or molybdenum, e.g., a titanium tetralkoxide. A most preferred photocatalyst precursor is a film formed upon partial hydrolysis of titanium tetraalkoxide, followed by polymerization by condensation ofthe hydrolysate.

0030When a photocatalyst-comprising film having a lower refractive index than that ofthe photocatalyst itself is desired, for example for use in automotive windshields as described above, the sol of which the photocatalysts precursor film is cast, in addition to the photocatalyst precursor, also contains a precursor ofa lower refractive index film. For example, a preferred second component or precursor sol is a precursor of silicon dioxide, e.g., formed by co-hydrolyzing a silicon alkoxide, such as a silicon te raalkoxide silicon, alkyltrialkoxide or dialkyldialkoxide, co-dissolved with titanium alkoxide acetylacetonate. The ratio ofthe amounts ofthe silicon and titanium oxide precursors in the sol are adjusted as needed to obtain the desired refractive index. The photocatalytic film compositions range from pure titanium dioxide to compositions having a 1 : 10 titanium dioxide:silicon dioxide molar ratio. Preferred low-refractive index films comprise vitreous silicon dioxide and crystalline titanium dioxide phases.

0031Barrier Layer

0032In the present invention, a barrier layer is defined as a barrier that slows or stops the diffusion or migration of alkali metal ions (e.g., sodium ions) and or alkali metal oxides (e.g., sodium oxide) into the photocatalyst precursor film or into the photocatalyst film. The barrier layer operates at the termperatures that the inventive photocatalyst precursor and photocatalyst films experience, and for the duration ofthe films.

0033In general, the barrier ofthe invention is the product ofthe reaction between hydrogen glass and a photocatalyst precursor as defined above. In the barrier, the most preferred three, four, or five-valent elements are Ti⁴⁺, Zr⁴⁺, Ge⁴⁺, Sn ⁺, and Si⁴⁺.

0034The most preferred barrier film is the reaction product between acid glass and the precursor of TiO₂, and includes the elements silicon, titanium, and oxygen.

0035Forming Photocatalyst Coating on Glass A photocatalyst containing layer can be formed on the surface ofthe glass from a vapor phase or from a precursor dissolved or dispersed in a liquid. The preferred liquids containing the precursor are long lived sols. An example of these is described in Example 1. Stable sols can be formed, for example, of titanium tetraalkoxides, by reacting these first with acetylacetone, then with water. The sols contain polymers ofthe precursor species or crystallites ofthe photocatalyst that do not have a longest dimension greater than about 30nm and most preferably not greater than about 20 nm, and are preferably smaller than about 5nm in their larger dimension. The preferred liquid phase in which the sol is dissolved or dispersed comprises alcohol in excess. Alcohols such as n-propanol, methanol, and butanol are useable.

0036A film of this liquid is applied to the glass surface, preferably to acid or etched glass, formed by acid etching (boiling in 9M sulfuric acid). The volume applied and the concentration ofthe precursor are selected so that the final thickness ofthe photocatalyst-containing layer will not be less than lOnm nor more than 500nm. The preferred final thickness is 20-200nm. The film can be formed by known methods, such as spraying microdroplets while the glass is cold or hot; dipping the glass in the liquid then removing it; pouring the liquid onto the glass and leveling the liquid layer mechanically or by spinning.

0037The film can also be formed by other methods, including a dry process, such as sputtering or evaporating a metal and then oxidizing it; or by reacting a low molecular weight molecular or metallic precursor in the gas phase prior to its deposition on the glass. For example, a titanium tetraalkoxide can be evaporated and decomposed either en route to or on the surface ofthe glass. TiCl₄ can be reacted with water to form

0038TiO₂ en route to the surface. In addition, metallic Ti can be reactively sputtered in an O₂ containing atmosphere to form a TiO₂ film.

0039In contrast to prior attempts at coating common glass, relatively active photocatalytically self-cleaning films are formed on common window glass (soda lime glass) when the glass surface has been treated with a reactant that produces a hydrogen- containing acid glass by exchanging sodium ions with protons and/or by hydrolysis of Si-O-Si bonds, both of which may occur upon acid etching. When materials such as sodium oxide in the glass diffuse into part ofthe photocatalytic film layer near the glass, phases other than the desired crystalline photocatalyst phase form, and the desired crystalline phase can become excessively sodium-contaminated. As a result, some ofthe photoactivity ofthe film is lost. Reactants that are useful in exchanging sodium ions or other alkali metal ions ofthe glass at or near its surface (by protons) and thus increase the photocatalytic activity ofthe coated glass, are generally acids. When the acid glass reacts with a precursor ofthe crystalline TiO₂, ZrO₂ or SiO₂ film, a sodium migration reducing barrier layer is formed. Even in the presence of this barrier layer, some photoactivity of the crystalline titanium dioxide-comprising films on glass is still lost when sodium oxide from the glass diffuses into part ofthe titanium dioxide layer near the glass. Photoactivity can be partially restored by subsequent acid treatment.

0040Acid Etching Glass-Pretreatment Etching is the process whereby the reactive acid glass is formed.

0041Chemically, etching may involve one or both ofthe following processes: Exchange of alkali metal (e.g., sodium) ions ofthe glass with protons; and scission, through hydrolysis, of Si-O-Si bonds. In both processes, a glass having SiOH junctions is produced. An enxample of an etchant is boiling 9M aqueous sulfuric acid. This is the preferred etchant. In a preferred process of making the photoactive film on glass, the glass is first exposed to acid so that sodium ions are extracted from its surface, being exchanged by protons and thereby forming an acid glass with silicon-bound OH-groups. The treatment with acid can be at ambient or, preferably, at higher temperatures, e.g., at the boiling temperatures ofthe acid. The glass surface can be rinsed following exposure to the acid with water, preferably deionized water, so as to remove any water soluble sodium salt. It has been noticed that treating ofthe glass with an acid prior to applying the photocatalytic coating to the glass leads to a higher photoefficiency. If the rinsing step is performed, the rinsing solution can be deionized water, or it can also contain a volatile base or ion such as ammonium hydroxide or an ammonium salt. Preferred acids for etching glass include those which form a hydrogen glass upon reaction with the glass. An example of such an acid is 9M (50%) sulfuric acid. Most preferably, the glass is reacted with boiling 9M H₂SO₄.

0042The photocatalyst precursor or photocatalyst containing film is then deposited on the acid-treated surface. The acid treated glass should not be calcined prior to application ofthe catalytic film, as such calcining lowers the photocatalytic activity. It is suggested that reaction between the acid-treated glass and the photocatalyst crystalline precursor establishes a sodium-migration barrier at the interface ofthe glass and the film.

0043Photocatalyst Coating The acid-treated and optionally rinsed glass is coated with the photocatalyst precursor. When coating is with a liquid, it is preferred that the precursor be in the form ofa sol having an ambient temperature shelf life longer than a day. The sol can be formed, for example, from a titanium tetraalkoxide, such as titanium tetraisopropoxide. The tetraalkoxide is first reacted in an alcohol solution, preferably in an excess alcohol, and preferably in an alcohol solution where the alcohol differs from the one that is evolved from the tetraalkoxide upon its hydrolysis. Although the sol can be formed by adding water to the titanium tetraalkoxide solution, in the preferred process the titanium tetraalkoxide is first reacted with a bifunctional Ti⁴⁺ complexing agent, such as acetylacetonate, to form a complex where the titanium to acetylacetonate ratio is 1 : 1. This complex is then hydrolyzed, preferably at room temperature, by adding water (preferably dissolved in alcohol), preferably at a molar ratio of 10 moles of water per 1 mole of titanium. The resulting precursor sol is generally stable, meaning that when stored for at least a day at a temperature between 5°C and 35°C the solution remains substantially clear.

0044A uniform film ofthe photocatalyst-containing compounds is cast on the glass through a process such as spinning, dipping, painting, spraying or applying an excess of solution then spreading it with a blade. The cast film is then allowed to dry.

0045Calcination

0046The dried-film coated-glass is next calcined, e.g., heated at a temperature to form photocatalytic crystallites and cause the photocatalyst layer to adhere to the glass. Calcination is preferably by heating in air at a rate resulting in a temperature increase of

004750°C per minute, then holding at the desired calcining temperature, preferably for about 15 to 30 minutes. The coated glass is then cooled. At this point, this glass is photocatalytically self-cleaning within the scope ofthe invention.

0048The calcination temperature is generally in excess of 275°C and less than 650°C, and preferably is in the range of 400°C - 650°C. For calcining films on glass, the most preferred temperature is in the range of 400-550°C; for films on silica, the most preferred range is 550-600°C.

0049Acid Wash - Post Treatment The photocatalytic activity (i.e. the rate at which the glass cleans itself of an organic contaminant) ofthe already photocatalyst-coated glass can be further increased by a second treatment, termed "post-treatment", with acid. Post-treatment, even in the absence of initial acid etching to form the acid glass, increases the low photocatalytic activity, though only to a lesser level than that observed when the glass was acid etched prior to deposition ofthe photocatalyst precursor. Applicants have found that multiple acid treatments can increase the self-cleaning glass's photoefficiency. In this second, acid wash, post treatment step, the photocatalytic-film- coated glass is again exposed to an acid that reacts with or neutralizes sodium oxide or calcium oxide or a product of these. The preferred acids for this process step are strong mineral acids, the ions of which do not form strong complexes with four valent ions such as titanium ions (Ti⁴⁺). Dilute aqueous nitric acid, and particularly nitric acid of O.IM to 3M concentration, with a preferred concentration of 0.2M is useful for this process step. Other useful acids include tetra fluoboric acid and dilute perchloric acid. Examples of acids that are not useful are 6M aqueous hydrochloric acid and 6M aqueous sulfuric acid, both of which dissolve or damage the photocatalytic titanium dioxide film on the glass. In the final process stage the glass may be rinsed, preferably with water, then dried.

0050Properties of Self-Cleaning Glass

0051The inventive self cleaning glasses include a substrate glass, barrier and a photocatalyst, as described above.

0052Self-cleaning glasses ofthe invention prepared as described above have coatings that strongly adhere to the glass and are abrasion resistant. These self-cleaning glasses have a photoefficiency, as defined in Example 1, when a film stearic acid is photoreacted, of at least 3.5 X lO^" . In more general terms, the glasses ofthe invention have a photooxidation rate sufficient to oxidize daily, in direct sunlight, organic contaminant films at a rate of 50 nm per day or more. At this rate, impaired vision due to a dirty windshield or lens of an eyeglass is minimized. 50 nm thick spots of contaminants interfere with vision, for example, by scattering light and causing glare. The glasses ofthe invention, as discussed more fully in Example 1, can also withstand successive applications and removals of scotch tape on their coated surfaces (tape test). Glasses of the invention can withstand being scribed with pencils of hardness H2 or softer as described in Example 1. EXAMPLES

0053The invention may be further understood by reference to the following examples, which are not intended to limit the scope ofthe invention in any way.

0054Example 1

0055Preparation of clear photoactive films of titanium dioxide on fused silica slides by spin coating

0056Clear, approximately 60 nm thick films of titania were produced on fused silica by a sol-gel process. The sol was made of a precursor solution prepared by mixing 4.5 mL Ti(OCH(CH₃)₂)₄, (97% in propanol) with 10.0 mL n-propanol and 1.6 mL acetylacetone (acac) to provide a stock solution having Ti:propanol:acac molar ratio of about 1 :9: 1.05. After aging for a week at room temperature (about 20°C - 25°C), a casting solution was prepared by mixing 1.0 ml ofthe precursor solution with a 1.8 ml water/n- propanol solution (1:9 v/v), the resulting water to titanium ratio being about 11:1. A clear, yellow casting solution, stable for at least two months and having a viscosity of 2.3 cp was obtained. Viscosity was measured using a falling ball type viscometer.

0057Prior to casting, the fused silica slides were rinsed in a cleaning solution (usually methanol), washed thoroughly with de-ionized water, and dried in a stream of air. The casting solution was then spread on the substrate (0.03 mL per 2.5x2.5 cm slide) which was spun, after the application, for 2 minutes at 4000 rpm to dryness. In the next stage, the coated silica slides were heated in air to 500°C at a rate of 50°C min^" and were calcined at this temperature for 30 minutes. The calcination transformed the product of the hydrolytic reaction into a microcrystalline oxide, stripped all organic residues, and bound the TiO₂ film to the substrate. Multiple TiO₂ layers were produced by applying a layer, drying in an oven (90°C, 10 minutes), applying another layer and finally calcining.

0058The process yielded uniform, clear and non-scattering films, as measured with a HP 8452A spectrometer (Figure 1). Comparison with the UV absorption spectrum of "milky" film coatings made of larger particles (Degussa P-25) revealed that the absorption edge ofthe clear films was shifted by 30-40 nm to the blue, as expected for nanoparticles in which electrons are confined.

0059The thickness ofthe films, 60±15 nm for a one layer film, was determined with an Alpha step 200 profilometer (Tencor Instruments). Transmission electron diffraction patterns and images ofthe thin films were obtained using JEOL- 1200EX and JEOL-2010 microscopes. For these measurements, the films were detached from their supports by pressing 200 mesh copper grids against the films while boiling in potassium hydroxide solution (6 M), following by thorough washing ofthe grids with water to remove any potassium residues. The electron diffraction ring patterns obtained by this method were indexed as that of T1O2 in the anatase phase and the dark field imaging suggested that the crystallites were segmented, with a typical segment being approximately 3 nm in diameter.

0060The resultant films were not damaged when wiped aggressively with any of several types of paper, including "Kimwipe", office copying machine paper, and newsprint, whether dry or wet. Furthermore, the films could not be removed by 20 successive applications and removals of Scotch® adhesive tape (3M-810), nor damaged when scribed with pencils of hardness H2 or softer.

0061The photoactivity ofthe various coated slides was tested by casting thin films of stearic acid (CH₃(CH₂)₁₆COOH) on the TiO₂-coated substrates and measuring at a defined irradiance the rate of decrease in the integrated absorbance ofthe ensemble ofthe C-H stretching vibrations between 2700 cm^"1 and 3000 cm^"1.

0062The measurements were performed on batches of 8-12 slides. The organic films were cast by applying 3x10"2 ml of 8.8x10"-^ M stearic acid in methanol per slide and spinning at 1000 rpm for 2 minutes to dryness. The integrated IR absorbance ofthe stearic acid films was measured by a Nicolet Magna IR-750 FTIR. The actual number of stearic acid molecules on the surface was calculated based on the integrated absorbance of densely packed monolayers of homologs having a known area per molecule (such as octadecyl trichlorosilane, arachidic acid and behenic acid). A typical stearic acid film had, prior to illumination, an integrated absorbance of 0.6 cm" 1 corresponding to ~1.9xl0^ molecules cm"2 .

0063The UV light source for the photoefficiency measurements was either a UVA wide band lamp, the peak emission being at 365 nm (Hideaway 6000 Solarium, Helitron Ltd.) or a 254 nm line-emitting mercury lamp (Sylvania G30T3). The irradiance, measured at the slide surface, was 2.4 ± 0.4 mW cm"2 for the UVA source and 0.8 ± 0.15 mW cm^~2 for the 254 nm source, corresponding to respective fluxes of 4.4x1015 photons sec" cm"2 and 1.0x10 IS photons sec"-- cm"2. These fluxes resulted in photocarrier generation rates differing only by a factor of 1.3, because ofthe large fraction of 254 nm UV photons absorbed in the 60 nm thick films in comparison to the lower fraction absorbed at the longer wavelength. The values given for the irradiance are the average between the readout of a JB A 100 power meter and a second measurement by a well known photochemical method, potassium ferrioxalate actinometry.

0064Within the context of this application, the photoefficiency is defined as the number of carbon - hydrogen bonds of stearic acid stripped from the photocatalyst surface per incident photon, dAxKxN/f, where dA is the change in the integrated IR absorbance ofthe C-H vibrations ofthe stearic acid per minute; K is the number of C-H bonds in a stearic acid molecule (35); N is the number of stearic acid molecules per cm^ per integrated absorbance unit, having a value of 3.17x101-5; and f is the irradiance

0065(2.6x 10*7 photons cm"2 minute"! f_{or me} UVA source and 6x10^ photons cm"2 minute" 1 for the mercury lamp).

0066The integrated absorbance ofthe organic contaminant's infrared C-H stretching vibrations was measured as a function ofthe exposure time. Figure 2 presents the FTIR spectra ofa Tiθ2 film on silica contaminated with stearic acid prior to and after exposure to the UVA light for 7.5 minutes. In general, the rate at which the stearic acid film was stripped remained constant during the exposure, as long as a continuous stearic acid film remained on the surface ofthe photocatalyst. Thus, the efficiency ofthe photoreaction was generally independent of time, remaining constant throughout the reaction period. Table 1 presents the integrated absorbance of stearic acid on silica coated with the photoactive Tiθ2 films following illumination with the UVA light. The efficiency in the table was calculated based on the change following 7.5 minutes of illumination.

0067The photoefficiency of a fused silica substrate coated with a single layer of Tiθ2 film was between 14xl0"3 and 21xl0"3 upon illumination with the UVA light and between 45x10^~3 and 81xl0"3 upon illumination with the 254 nm light, the ratio between efficiencies being scaled with the ratio in the number of photons absorbed per cm-2 per second at each wavelength. Values, averaged over more than 25 slides are presented in table 2.

0068Table 1

0069Slide No. of Integrated Integrated Integrated Integrated efficiency No. Ti02 absorbance absorbance absorbance absorbance layers t=0 t=3.75 min. t=7.5 min. t^*=15 min.

007065 1 0.88 0.55 0.31 18.8xl0"³

0071205 1 0.63 0.51 0.35 0.2 15.9xl0'³

0072206 2 0.78 0.60 0.45 0.2 18.8X10'³

0073247 2 0.89 0.70 0.58 0.24 17.6x10-³

0074249 2 0.85 0.70 0.57 0.25 15.9xl0-³

0075Table 1: Changes in the integrated absorbance ofthe FTIR C-H stretch band of stearic acid on silica slides coated with clear and photoactive Tiθ2 upon illumination with UVA light. The efficiency given in the table was calculated based on the change after 7.5 minutes of exposure.

0076Table 2

0077Glass substrate Silica substrate

0078365 nm 254 nm 365 run 254 nm

0079Ti0₂

0080(clear) (0.7±0.35)xl0"³ (17.5±3.5)xl0"³ (17.5±3.5)xl0-³ (63±17.5)xl0"³ on non-etched substrate τio₂

0081(clear) (9.1±2.8)xl0"³ (58±18)xl0-³ (21±3.5)xl0-³ (73.5±28)xl0-³ on etched substrate

0082Table 2: Photoefficiency (number of C-H bonds consumed per impinging photon) of Tiθ2 clear films on glass and on fused silica substrates, showing the effect of H2SO4 etching ofthe substrate upon the efficiency. The values are averages for 20-30 samples. Example 2

0083Preparation of clear photoactive films of titanium dioxide on soda lime glass slides by spin coating

0084Clear, photoactive films of titanium dioxide were produced on soda lime glass (Corning 2947, composed of O (60 atom%), Si(24.5 atom%), Na(10 atom%), Ca(2.5 atom%), Mg(2 atom%), and Al(l atom%) by the sol-gel method described for Example 1. In the first step of their preparation, the substrates were cleaned by an organic solvent (usually methanol or chloroform), then etched by boiling them for at least 30 minutes in fuming 50% (9M) sulfuric acid at 240°C. After being cooled to ambient temperature, the slides were washed with de-ionized water and dried in a stream of air. Then, the same organotitanate coating solution described in Example 1 was applied onto the surface (0.04 ml per 3.75x2.5x1 mm slide) which was spun, as described for Example 1. The coated glass substrates, were calcined usually at a temperature of 400 °C, thus producing clear and homogeneous films, denoted as GE films. For comparison, Tiθ2 films on non-etched glass slides, denoted as GN, and Tiθ2 films on etched and on non-etched silica (denoted SE, SN respectively) were produced in the same manner. Thicker films were produced either by applying a first layer ofthe titanate precursor solution, oven drying (90 °C, 10 minutes), applying a second layer and calcination or by repeating the single layer preparation process. The former class is denoted as (2), whereas the latter is denoted as (1+1).

0085All films, whether GE, GN, SE, or SN withstood the abrasion tests described in Example 1. A ratio of BET (Brunauer-Emmett-Teller) to geometrical surface area of approximately 20 was determined for the Tiθ2 coated glass slides by N2 BET adsorption isotherms, using a Micromeritics AccuSorb system. Characterization ofthe films, i.e. UV, TEM and profilometry measurements were performed as described in Example 1. Figure 2 shows the UV absorption spectra of clear Tiθ2 films on etched glass (dashed line) and on non-etched glass (solid line), the latter being blue shifted by 4-5 nm with respect to the former. As seen below, the difference between the spectra of films on etched glass (GN type) may have resulted ofthe GN type films not having the anatase phase or of smaller crystalline domains in the GN type films. "Plan view" TEM images ofthe Tiθ2 films, detached from their substrates in the same manner as described in Example 1, revealed a distinct difference between type GE and type GN films. Type GE films were found to be identical to the SN films described in Example 1 (i.e. consisted of segmented nanocrystallites having the anatase phase) whereas the type GN films were typical for materials having no long range order, with a diffused ring in their selected area diffraction pattern corresponding to an interplanar distance of 2.6 -3.6 A.

0086The compositions ofthe supported films, as well as ofthe glass substrates, were measured by x-ray photoelectron spectroseopy (VG-ESCALAB). To obtain depth profiles, the samples were sputtered with an argon gun (Varian 981-2043, 3 kV, 25 mA) and re-measured. The estimated sputtering rate was 0.15 nm min." 1. For these measurements, the samples were cut into 8 mm x 8 mm slides, that were attached to their pedestals by a conducting adhesive tape to reduce charging during the sputtering process.

0087Table 3 presents the percentage of sodium atoms in various Tiθ2 clear films on soda lime glass, and in glass slides, some pre-etched, some non-etched and some calcined after being etched, as deduced from the XPS measurements. These values are based on the area ofthe sodium and oxygen IS peaks taking sensitivity factors (sf) of 2.51 and 0.63, respectively, and on the area ofthe 2P doublet (2P3/2 + 2Pl/2) pea of titanium, taking a sensitivity factor of 1.59. Whenever silicon was found, for example in the glass samples or after prolonged sputtering ofthe Tiθ2 samples, its atomic fraction was calculated from its 2P peak (sf=0.17). No constituents apart from titanium, oxygen, and sodium were found in the Tiθ2 films.

0088Table 3

0089glass Ti0₂ on glass Ti0₂ on glass (one layer) (two layers)

0090Pre- Sputtering Not HND₃ HN0₃ treatment time calcined calcined untreated treated untreated treated (min.)

0091Non-etched 0 7.2 35.0 21.5 2.0 14.0 2.5

009230 60 13.6

009390 9.1 12.8 2.5

0094180 3.4 8.7 14.5

0095300 5.3 14.1

0096420 17.1 18.0

0097Etched 0 1.8 7.5 8.4 3.2 7.7 0.0

009830 0.5 1.1

009960 7.0

010090 3.4

0101120 4.4 0.0

0102150 1.6

0103300 3.0

0104Atomic percentage (%) of sodium measured by XPS, in soda lime glass and in clear films of Ti0₂ on soda lime glass.

0105A high atomic fraction of sodium was measured not only at the interface between the Tiθ2 and the glass but also at the air interface. The concentration of sodium at the air interface was actually higher than in the bulk ofthe films. Evidently, the sodium diffusion length exceeded the film thickness and sodium segregated at the surface ofthe titanium dioxide film. The effect of etching the glass prior to coating with the Tiθ2 precursor on the sodium concentration at the Tiθ2 film surface is evident, the atomic percentage of sodium in the GE films being 2-3 times lower than that in the GN films. In the GN films the atomic percentage of sodium exceeded that in the bulk ofthe soda lime glass (10 atom%), showing that the film was not only contaminated with sodium, but actually extracted sodium from the glass. The amount of sodium in the thicker photocatalyst films made by applying two layers ofthe organo-titanate precursor and then calcining, was only slightly smaller than in films made of one layer. A high atomic percentage (7.5%) of sodium was found on the glass surface after it was etched and calcined. Nevertheless, the fraction of sodium at the surface ofthe non-etched but calcined glass was far higher (35%). Thus, the pre-etching ofthe glass prior to application and calcining the photocatalyst precursor film caused effective blocking of sodium migration into the photocatalyst film. The photoefficiency ofthe photocatalytic stearic acid stripping process, as defined in Example 1, was between 5xl0^~3 and 12xl0"3 for the GE slides illuminated with the UVA light whereas for the Tiθ2 films made on non-etched glass (the GN films) it was at least 7 times less. The photoefficiency results obtained for a batch containing 12 slides, half of which were GE type while the other half were GN type are given as an example in Table 4. In Table 2 the average photoefficiencies obtained for Tiθ2 clear films on fused silica, on glass etched by H2SO4 prior to coating with the titanium precursor, and on non-treated glass are presented. Each value listed in the table represents an average for more than 25 samples. For soda lime glass substrates, a large difference in the photoactivity was found between GE type and GN type samples, the latter being several times less efficient than the former. This difference was greater for excitation upon 365 nm photons than by 254 nm photons. In contrast to the observed difference in efficiency between films on etched and on non-etched glass, etching did not increase or reduce the photoactivity when the films were coated on fused silica.

TABLE 4

0107Slide No. Pre-treatment Photoefficiency (%)

0108398 50% H₂SO₄ (boiled), 30 min. 4.9 X 10^"3

0109399 50% H₂SO₄ (boiled), 30 min. 5.25 X 10^"3

0110400 50% H₂SO₄ (boiled), 30 min. 8.4 X lO^"3

0111401 50% H₂SO₄ (boiled), 30 min. 8.05 X 10^"3

0112402 50% H₂SO₄ (boiled), 30 min. 7.7 X lO^"3

0113403 50% H₂SO₄ (boiled), 30 min. 8.05 X 10^"3

0114404 without pre-treatment .35 X 10^"3

0115405 without pre-treatment .7 X lO^"3

0116406 without pre-treatment .7 X 10^"3

0117407 without pre-treatment .35 X 10^"3

0118408 without pre-treatment .35 X 10^"3

0119409 without pre-treatment 0.0

0120410 without pre-treatment .7 X 10^"3 The dependence ofthe efficiency of clear films of titanium dioxide on the calcination temperature is seen in Figure 4. In etched glass type (GE) films on soda lime glass, the efficiency decreased when the calcination temperature exceeded 450°C (B,C), whereas on fused silica no decrease in efficiency was observed at higher calcination temperatures (A). The absence of loss of efficiency in silica-supported films indicated that the cause of loss in the photoactivity was not sintering, leading to lower surface area. The results obtained with 254 nm light were similar, but the decrease in photoactivity of Tiθ2 on soda lime glass upon calcination at elevated temperatures was less pronounced.

0121The effect ofthe duration ofthe etching in sulfuric acid at 240°C on the 365 and 254 nm photoefficiencies is seen in Figure 5. While etching for 10 minutes sufficed to produce a film which was highly photoactive under 254 nm light (filled circles), more than 45 minutes of etching were needed for best activity under 365 nm light (open circles).

0122Table 5

0123Batch Film Type As is Double NaN0₃ NaN0₃ NaOH NaOH -number and wave¬ calcination treatment treatment + soaking soaking + and (# of length calcination calcination samples)

01241 (2) P-25, 365 55.3 x 10^"3 40.6 x lO^"3 40.95 x lO^"3 9.45 x 10^"3 4.2 l0^{' b} 6.3 x 10^'3b nm Degussa

01252 (2) P-25, 254 92.75xl0^"3 - 136.85x10 ^'3a 33.6 x l0^"3a 75.6 x 10^*3b 10.15xl0^"3b nm Degussa

0126Etch glass,

01273 (2) 4 layers, 14 x l0^"3 14.35xl0^'3 6.3 x lO^*3** 0.35 x 10^*3a 0" 0^b 365 nm.

0128Etch glass,

01294 (2) 4 layers, 12.25x10^3c 13.65xl0^{" c} 9.1 x 10^"3a*c 2.8 x l0 1^3a-^c 10.85 x 10^'3₀b.c 365 nm.

0130Etch glass,

01315 (1) 4 layers, 16.45 x l0^"3 12.6 x l0^"3 17.6 x l0^"3d 07.35 x l0^{" d} - - 365 nm. 12.95 x l0^{'3 e}

0132Etch glass,

01336,7 (4) 7 layers, 18.2xl0^"3 - - 11.2 x l0-^3f - 11.2xl0^"3g 365 nm.

0134Table 5: Photoefficiency of Ti0₂ films, expressed as the number of C-H bonds in a steric acid film consumed per impinging photon. a-Dipping in NaN0₃ (1 M, 10 minutes, 20°C). b-Dipping in NaOH (1 M, 10 minutes, 20°C). c-Batch 4 obtained by washing batch 3 slides with methanol to remove stearic acid residues, immersion in water and drying. d-Dipping in NaN0₃ (0.2 M, 10 minutes, 20°C). e- Washing after calcination. f-3 layers of Ti0₂, calcination, NaN0₃ layer by spin coating (0.2 M, 1250 rpm, 1 minute), calcination, washing and coating with 4 additional layers of Ti0₂. g-3 layers of Ti0₂, calcination, NaN0₃ layer by spin coating (0.2 M, 1250 φ , 1 minute), calcination, washing and coating with 4 additional layers of Ti0₂.

0135Table 5 demonstrates the deleterious effect of sodium contamination on the efficiency of titanium dioxide films on glass. Dipping ofthe clear Tiθ2 films in NaOH (IM, 20°C) reduced their efficiency under 365 nm light to nil (Table 5, batch 3), and subsequent washing with de-ionized water led to recovery ofthe efficiency (Table 5, batch 4). However, when the NaOH treated slides were calcined, the loss of their efficiency could not be reversed (Table 5, batches 3,4). Double calcination of GE films (Table 5, batches 3-5) did not increase or decrease their efficiency. Films made of Degussa P-25 that were immersed in NaOH were 10 fold less efficient than untreated films, under 365 nm light (Table 5, batch 1). Immersion of P-25 samples in NaOH had, however, a less severe effect on their photoactivity at 254 nm (Table 5, batch 2). Untreated P-25 films showed a moderate decrease in efficiency upon a second calcination (Table 5, batchl), possibly because their surface area was reduced. Etched glass type (GE) films did not show a significant loss in photoefficiency at 365 nm after immersion in a 0.2M NaNO3 solution (Table 5, batches 3-5). Dipping in >lM NaNO3 reduced, however, the 365 nm efficiency by ~ 50%. The efficiency of films made of P-25 was reduced by dipping in IM NaNO3 by 25% at 365 nm (Table 5, batch 1). No loss in efficiency was observed for 254 nm light (Table 5, batch 2). Films calcined following their immersion in NaNO3 had only 10-40% ofthe efficiency of untreated films under 365 nm light (Table 5, batches 3-5). The efficiency ofthe P-25 films subjected to the same treatment was only 20% ofthe efficiency of untreated films under UVA light (Table 5, batch 1), and 40% under 254 nm light (Table 5, batch 2). Washing with water after the calcination of NaNO3 treated GE films increased the efficiency ofthe calcined samples, probably by removing some ofthe sodium ions on the surface (Table 5, batch 4). In a sandwiched structure, where a NaNO3 layer, decomposing upon calcination to Na2θ, was introduced on a calcined Tiθ2 film made of 3 precursor layers, then covered with 4 layers of Tiθ2 precursor and calcined, the efficiency was only 60% ofthe efficiency of clear films made of 7 precursor layers without NaNO3 (Table 5, batch 5).

0136To probe the effect ofa second calcination, two batches of slides was prepared. In one the first precursor layer was applied to the etched glass, oven dried (90 oC, 10 min.), coated with a second layer, then calcined. These slides were termed GE(2). The second batch was made by repeated coating and calcination of an etched glass slide. These slides were termed GE(1+1). The absorption spectra ofthe two batches were identical, showing that the sizes ofthe crystallites, their phases and the thickness ofthe films did not differ. The GE(1+1) films, despite being calcinated twice, did not show lesser photoactivity than the GE(2) films. When the glass was etched, then calcined (GC type samples) and then overcoated with the Tiθ2 precursor and re- calcined the efficiency was only 20-40% of that ofthe standard 1 layer GE film.

0137The XPS results show that sodium migrates from the soda lime glass substrate into the titanium dioxide layer during the calcination step. When the sodium was extracted from the surface of the soda lime glass, and when the extracted glass was overcoated with the organo-titanate precursor and calcined, then the sodium content in the titania films was significantly lower. Such lowering affected the titanium containing phase that formed and thereby the photocatalytic behavior. It is known that at temperatures below 550°C, the extraction of sodium from glass is governed by an ion exchange mechanism and is limited by the rate ofthe diffusion of sodium ions rather than by the rate of diffusion of protons. If the process of etching would not introduce a change in the glass structure and if charge neutrality would be maintained only through protonation ofthe originally sodium occupied sites, i.e. if a "hydrogen glass" would have formed, than calcination at 400°C should have resulted in complete recovery from the extraction, the Na⁺ being redistributed simply because ofthe higher Na⁺ diffusion coefficient at the calcination temperature of 400°C relative to that at the extraction temperature of 240°C (2x10"9 cm-2 sec'l versus 2x10"! 1 cm-2 sec- respectively, for a soda lime glass of similar composition). However, if the extraction of sodium and the formation ofa "hydrogen glass" were followed by loss of water, then a layer of denser glass, containing fewer ion exchanging Si-OH functions and less permeable to Na⁺ ions would be formed. Formation of such a layer was proposed for dealkalization of soda lime glass in atmospheres containing water vapor and SO2- Such a sodium diffusion limiting layer is formed in glass by etching and calcination. Despite limiting the sodium diffusion, this layer does not completely prevent the migration of sodium, as is evident from the difference between the atomic percentage of sodium in etched glass prior to (1.8%) and after (7.5%) calcination. The results show that the atomic percentage of sodium in the surface layer of soda lime glass that was etched and then calcined was much smaller than in glass calcined without being etched (7.5% vs. 35% at the surface, and 3.4% vs. 12.8% after 90 minutes of argon sputter etching).

0138Depth profile measurements show a constant concentration of sodium across the Tiθ2 film in both GE type and GN type films, with an increase in the sodium concentration at the air interface. Apparently, the nanocrystalline Tiθ2 film has a greater affinity for sodium than the silicate network of soda lime glass, particularly when the surface ofthe crystallites is hydrated and the protons are sodium exchangeable, H2Ti_nO2n+l being a stronger acid than H2Sinθ2n+l • In silicate glasses, but not necessarily in Tiθ2 films on glass, sodium also accumulated at the surface upon hydration ofthe silicate network, sodium segregation at the surface being coupled with depletion of sodium from the layer below. The sodium depleted .layer observed in silicate glasses was not observed in the GE films, showing that Na⁺ accumulated throughout the film, not only at its air interface. It was particularly noteworthy that the concentration of sodium in the GN films was higher than in the soda lime glass itself. This higher concentration can be rationalized by Ti-OH being a stronger acid than Si-OH, because ofthe more electropositive nature of Ti⁺4 relative to Si⁺4 , combined with fast sodium diffusion across the film. Because the sodium ions were uniformly redistributed by the 400°C calcination step, the ions diffusing rapidly through the nanocrystalline Tiθ2 layer, similar sodium concentrations in the "type GE" and the "type GN" films would have been observed, were it not for the existence of a sodium transport limiting layer at the Tiθ2 - extracted glass interface. The massive difference between the GE type and GN type films is explained by creation of a zone in the GE-type films through which sodium permeation was slow. When the precursor temperature is raised nucleation, crystal growth, and oxidation of organic residues take place. The sodium concentration in the titanium dioxide film was not related to its final concentration at the glass-titania interface, but to its concentration at this interface at the start ofthe calcination, when the unique sodium diffusion limiting layer was formed. Sodium extraction followed by a first calcination, then by coating with the Tiθ2 precursor and by a final calcination did not provide films as photoactive as those obtained when in the first calcination the Tiθ2 precursor was already present. Therefore, dehydration and compacting ofthe glass accounted only for part ofthe sodium transport characteristics ofthe layer. The unique sodium transport limiting characteristics are explained by reaction ofthe decomposing anatase Tiθ2 precursor and the dehydrating hydrogen glass, and/or by fast nucleation and growth at the interface between the glass and the Tiθ2 precursor layer.

0139In the type GN films, in contrast with the type GE films, there was no long range order. Earlier work has shown that titanium dioxide films on non etched soda lime glass were of brookite phase when formed by rapid heating of a film ofa hydrolyzed titanium alkoxide precursor(>25°C/min), and that a Na2O-xTiO2 phase was formed when the temperature was raised sufficiently slowly to allow diffusion of the sodium ions from the glass into the TiO₂ layer. The absorption edge of Na2θ-xTiθ2 was found to be shifted by 10-12 nm to shorter wavelengths relative to the edge of anatase, whereas the absorption edge for the brookite structure was identical with that of anatase. the data did not reveal the existence of well defined phases in the GN films, possibly because a slowly polymerizing chelated precursor was used. Hence, sodium could diffuse into the nascent Tiθ2 film in amounts sufficient to prevent its crystallization.

0140The photoefficiency invariably decreased when the Na⁺ concentration increased in the Tiθ2 films. Furthermore, the photoactive GE films lost efficiency when soaked in NaNO3 and calcined at temperatures where the NaNO3 decomposed to sodium oxide. The lesser efficiency of films on soda lime glass relative to films on fused silica presented in Example 1 could have resulted either ofthe presence of non- anatase phases or the presence of sodium oxide at the air interface, where it would have raised the local pH. In films on non etched glass the sodium fraction exceeded 10 atom% and, as a result, formation of an ordered anatase, brookite or sodium titanate phase was inhibited. This reduced the efficiency practically to nil for 365 nm illumination. In contrast, on etched glass there were only crystalline anatase domains, even though their sodium atom percentage was as high as 8%.

0141Because a sodium transport blocking layer forms ofthe hydrogen glass and the nascent, yet non-crystalline Tiθ2 layer, sodium diffusion takes place mainly during the beginning ofthe calcination process, and a second calcination has only a minor effect on the efficiency ofthe films (Table 5). That the critical step for prevention of sodium ion migration is the reaction ofthe hydrogen glass with the nascent Tiθ2 layer was also suggested by comparison ofthe efficiency of films prepared by two methods involving (a) applying a first layer, calcination, then applying a second layer and calcining again and (b) applying two successive layers and calcining only once. There was little or no difference in the photoefficiencies ofthe two films. The fact that the Tiθ2 film made on etched and calcined (GC type) glass had only 20- 40% ofthe efficiency of a Tiθ2 film made by coating the precursor on etched but not calcined glass suggests that the sodium transport limiting layer, formed by interdiffusion of Tiθ2 and hydrogen glass, was a superior sodium transport blocker to the film formed on the dehydrated hydrogen glass.

0142Transport of sodium from the substrate into the photocatalyst film precursor and into the photocatalyst films is detrimental to their photocatalytic activity. The Tiθ2 films have a higher affinity for sodium ions than the soda lime glass itself and extract the sodium ions from the glass during the calcination step. Sodium transport to the Tiθ2 layer can be retarded by forming a blocking layer. Such a layer forms upon calcining the Tiθ2 precursor film on the hydrogen glass, formed upon extraction ofthe sodium with hot acids such as sulfuric acid.

0143Example 3

0144Preparation of clear photoactive films of titanium dioxide with enhanced photoactivity on soda lime glass

0145Clear films of titanium dioxide were made on soda lime glass either with or without acid pre-treatment (types GE and GN respectively) as described in Example 2. Following their calcination, the samples were post-treated with dilute nitric acid, usually by soaking them in 0.2 M HNO3 for 15 minutes at 22 ± 3°C, then washed with water to remove acid residues.

0146In Table 3 the effect of post-treatment on the atomic percentage of sodium is presented. The atomic percentage of sodium in all films (GE(1), GN(1), GE(2), GN(2) as defmed in Example 2) was reduced significantly by the post treatment. For example, from 8.4% to 3.2% at the surface of a titanium dioxide film on etched glass and from 21.5% to 2% at a the surface of a Tiθ2 film on non-etched glass, manifesting the replacement of sodium ions in the films by protons from the acid.

0147Table 6 presents the efficiency (as defined in Example 1) measured for a batch of slides containing GN and GE slides, with and without post -treatment, upon illumination with the UVA source. In Table 6, pretreatment means boiling the uncoated glass in 9M sulfuric acid for 30 minutes, "without water wash" means that the slide was dried with a residue of nitric acid. "With water wash" means that any residual nitric acid was removed by washing with de-ionized water. The average efficiencies of slides with a post-treatment versus the average efficiencies without such a treatment are presented in Table 7. In Table 7, "treated with HNO3 " means soaked in HNO3 0.2 M for 15 minutes, followed by a water wash.

0148Table 6

0149Slide No. Pre-treatment Post-treatment Photoefficiency (%) Photoefficiency (%) without water wash with water wash

0150656 yes yes 14.7 x lO^"3 10.15 x lO^"3

0151657 yes yes 14.0 x lO^"3 11.9 x lO^'3

0152658 yes no 9.1 x lO^"3 7.35 x 10^"3

0153659 yes no 7.0 x lO^"3 7.35 x 10^"3

0154662 no no 1.05 x lO^"3 1.05 x lO^"3

0155663 no no 0.35 x 10^"3 0.35 x 10^'3

0156666 no yes 3.85 x lO^"3 2.1 x lO^"3

0157667 no yes 3.15 x lO^"3 1.75 x lO^"3 Table 7

0158<img file="WO9707069A1_D0001.tif" />

0159HNO3 post-treatment significantly increased the photoefficiency of non- etched glass (GN type) Tiθ2 films although their photoefficiency remained much smaller than that ofthe standard etched glass (GE) Tiθ2 films. In GE type films the improvement was slighter. Evidently, post-treatment was not an effective substitute for pre-etching ofthe glass. The gain efficiency upon HNO3 post treatment was partially lost with time, as observed in repeated measurements. No increase in efficiency was observed upon HNO3 treatment of films prepared according to Example 1 on fused silica (Table 8), manifesting that the observed increase in efficiency in the Tiθ2 coated glass was not due to a lowered surface pH. The cause ofthe improvement was reaction or neutralization of sodium oxide or its products that migrated to the titanium dioxide layer's surface during the calcination process.

0160Table 8

0161Slide No. Slide Type Post-Treatment Photoefficiency (%) a TiCyquartz untreated 14.7 x lO^"3 b Ti0₂/quartz untreated 11.9 x lO^"3 c Ti0₂/quartz soaking in HN0₃ 0.2 M 12.6 x lO^"3 (22°C, 15 min.) d Ti0₂/quartz soaking in HN0₃0.2 M 10.85 x 10° (22°C, 15 min.)

0162For treating titanium dioxide films on soda lime glass, the acid chosen was nitric acid. The anion of this acid, in contrast to with the anions of sulfuric and hydrochloric acid does not complex the four valent titanium ion. The rate of dissolution ofthe photoactive film in nitric acid was consequently much slower than its rate of dissolution in sulfuric or hydrochloric acid at the same normalities. Within this context, it is noteworthy that GE films were more resistant to soaking in concentrated acids than the GN films.

0163Soaking ofthe Tiθ2 films in dilute acids (like HNO3) reduced the concentration of sodium near the surface and improved the efficiency by exchanging sodium ions with protons without degrading the mechanical properties ofthe Tiθ2 films. The photoefficiency increased particularly in type GN films. Although such soaking was not an adequate substitute for etching the glass prior to the application of the titania precursor, it was an appropriate finishing step for the process of forming a photocatalytic film on soda lime glass.

0164Example 4 Preparation of clear photoactive films of titanium dioxide on soda lime glass from coatmg solutions having controlled viscosities

0165The coating solution prepared in the manner described in Example 1, (denoted as "Ti" solution in this example) was mixed with various non-interacting organic solvents such as hexane, methylene chloride, chloroform, thus obtaining coating solutions, their viscosities depending upon the volumetric ratio between the coating solution and the non-interacting solvent. For comparison, films made from the "Ti" solution, as well as from a mixture ofthe "Ti" solution with methanol which may interact, like other alcohols, with the titanium dioxide precursor, were prepared. The mixed solutions were spread on acid - etched glass substrates (40 ml per 3.75x2.5 cm) which were spun, after the application, in N2 environment, for 2 minutes at 1000, 2500 or 4000 rpm to dryness. The slides were then calcined at 400°C for 35 minutes, in air. Clear, homogenous, well adhered films, were obtained.

0166The photoefficiency ofthe various types of slides was measured simultaneously. The results are presented in Table 9. For spin coating at 1000 rpm, there was practically no adverse effect on the efficiency following coating by solutions containing a "Ti" / non-interacting solvents mixtures, despite the lower amount of Tiθ2 in these films, due to the reduced viscosity ofthe coating solution. For spin coating at 2500 rpm, films made of Ti / non-interacting solvents solutions had 30% - 45% ofthe efficiency of films made ofthe "Ti" solutions, when illuminated with 365 nm light, and app. 70% - 100% when illuminated with 254 nm light. In contrast, the efficiency, measured with 365 nm light, of Tiθ2 films made ofthe Ti / methanol solution was only 6% of that ofthe regular GE film when coated at 2500 rpm and 62% when coated at 1000 rpm.

0167Table 9

0168<img file="WO9707069A1_D0002.tif" /> Table 9: Efficiency values of slides prepared by spin coating of mixtures ofthe Tiθ2 precursor coating solution with various non-interacting solvents. In the table, "Ti" represents the regular coating solution, its preparation being described in example 1. It can be concluded that photoactive clear and homogenous films can be formed on glass by means of diluting the coating solution with solvents that do not interact with the titanium dioxide precursor and that this method is especially viable for cases in which relatively low rates of spinning are required.

0169Example 5

0170Preparation of clear photoactive films of titanium dioxide on soda lime glass by dip-coating

0171Coating solutions of controlled viscosities, prepared in the manner described in example 4, were used to produce clear, photoactive and homogeneous films of Tiθ2 on glass. For that, corning 2947 soda lime glass slides, 7.5x2.5x0.1 cm in size, were etched as described in example 2, connected to a speed controlled elevator and dip coated with an up-stroke speeds ranging between 0.5 cm min'l and 4.3 cm min^~l . The

0172Tiθ2 precursor coated films were than calcined at 400 - 450 °C for 35 minutes. A clear, homogeneous, well adhered film of Tiθ2, was obtained. Best results, in terms of clarity and homogeneity, were obtained by performing the process under N2 atmosphere, enriched with the solvents' vapor. The photoefficiencies of several dip-coated films, as measured by the stearic acid test described in example 1, are presented in table 10.

0173A comparison between the photoefficiency ofthe dip - coated films (table 10) and films of similar thickness made by spin coating (Table 2) reveals that the efficiency ofthe dip-coated films was not inferior to that ofthe spin-coated films, thus enabling large scale implementation.

0174Table 10

0175Slide Coating Solution Solution Up-stroke Film Efficiency No. (v/v) Viscosity Speed Calcination Absorbance at at 365 ran

0176(cp) cm min"* 300 nm

01771636 Ti/hexane (l :l) 0.88 4.3 400°C, 0.65 12.3xl0"³ 35 min.

01781667 Ti/hexane (l:l) 0.88 2.68 450°C, 0.57 l l.OxlO"³ 35 min.

01791664 Ti /hexane (l:l) 0.9 3.22 450°C, 0.55 9.1xl0-³ 35 min.

01801666 Ti/hexane (2:3) 1.02 1.34 450°C, 0.61 8.8xl0"³ 35 min.

01811650 Ti/hexane (2:3) 1.02 0.54 450°C, 1.13 13.7xl0"³ 35 min.

01821623 Ti hexane (3:7) 0.66 4.3 400°C, 0.28 5.6xl0"³ 35 min.

01831655 Ti^#/hexane (3:7) 0.65 0.54 450°C, 1.00 14xl0"³ (2 layers) 35 min.

0184Table 10: Efficiency values of slides prepared by dip-coating in mixtures hexane and the Tiθ2 precursor coating solution. In the table, "Ti" represents the regular coating solution, the water to titanium ratio being 11:1, while Ti# represents a coating solution where the water to titanium ratio was app. 4: 1

0185Example 6

0186Preparation of clear photoactive films of titanium dioxide with sodium diffusion limiting layer

0187Silica films were produced on soda lime glass by the same manner described in Examples 1 and 2, using silicon tetrapropoxide instead of titanium tetra-i- propoxide. Here, the silicate precursor solution consisted of 10 ml n-propanol, 1.6 ml acetylacetonate (acac) and 4.9 ml ofa Si(OPr)₄ solution (Aldrich Cat. No. 23,574, 95%) by weight). The coating solution was made from 1.0 ml ofthe silicate precursor solution and 1.8 ml of 1:10 (v:v) water in n-propanol. 3.75 cm x 2.5 cm x 1 mm soda lime glass slides were then etched in boiling sulfuric acid for 30 minutes as described in Example 2, and coated by spin coating as described therein. Following drying at 80°C for 20 minutes, a second layer, made ofthe same coating solution described in Example 1 (denoted as "overlayer"), was applied by spin coating as described therein. The slides were then calcined at 450°C for 30 minutes to produce clear films comprised ofa silica layer (denoted as "underlayer") between the glass substrate and the photoefficient titanium dioxide overlayer. Glass slides coated with a film containing a first layer of zirconia and a second, photoactive layer of titania were obtained in the same manner. Here, the zirconia coating solution was made of zirconium tetra n-propoxide (Aldrich 33,397 -2, 70% in 1-propanol) with the same molar ratios between zirconium, acac, water and propanol as described for the titania precursor coating solution in Example 1. No post-treatment with nitric acid was used.

0188The efficiency ofthe slides was measured in the manner described in Example 1. For comparison, films containing one and two layers of TiO₂ on glass, were made from the same batch of etched glass slides and with the same TiO precursor coating solution. The efficiency of these slides was measured simultaneously with that ofthe glass-silica-titania films. As shown in Table 11, the efficiency ofthe two layered structure, containing a silica underlayer and TiO₂ overlayer, was higher than that ofa single layer of titanium dioxide by a factor of between 1.4 and 2.0, but less than that ofa film containing two layers of TiO₂.

0189Table 11

0190Slide No. Structure Photoefficiency

0191701 1 layer of TiO₂ on soda lime glass 3.5 x l0^"J

0192706 1 layer of TiO₂ on soda lime glass 5.6 x lO^"J

0193709 1 layer of TiO₂ on a silica 8.75 x 10^"J underlayer applied on soda lime glass

0194710 1 layer of TiO₂ on a silica 10.85 x 10° underlayer applied on soda lime glass

0195699 2 layers of TiO₂ on soda lime glass 14.35 x 10^"J

0196712 2 layers of TiO₂ on soda lime glass 12.95 x 10°

0197713 2 layers of TiO₂ on soda lime glass 12.6 x 10°

0198A two-layered structure, consisting of a zirconia underlayer and one TiO₂ overlayer coated on glass, had higher efficiency than that of a glass coated with a single film of TiO₂, when made at calcination temperatures higher than 500°C, where the diffusion of sodium into the forming TiO₂ film is fast enough to reduce its efficiency even in acid etched glass substrates (Figure 4). This makes such a sodium diffusion blocking structure appropriate for applications where the glass has to withstand high temperatures, for example in products made by tempering ofthe glass.

0199Example 7

0200Preparation of clear photoactive films of titanium dioxide on a layer made of a mixture of a TiO₂ precursor and a SiO₂ precursor

0201A layer composed of a mixture of TiO₂ and SiO₂ (1:1) molar ratio was produced on soda lime glass and on fused silica by preparing the appropriate coating solution and applying it onto the substrate by spin coating as described in examples 1 and 2. A precursor solution was made of 10 ml n-propanol, 2.25 ml titanium-tetra-i- propoxide (98% in propanol, density: 1.033 gr. cm^"3), 1.6 ml acetylacetonate (acac) and 2.43 ml of silicon tetrapropoxide (Aldrich 23,574-1, 95% in propanol, density:0.916 gr. cm^" ). After aging the precursor solution for a day, the coating solution was made by mixing 1.0 ml ofthe mixed precursor solution with 1.8 ml of a water/ n-propanol solution (1:10 v/v). The coating solution was then spread by spin coating at 4000 rpm on H₂SO₄ etched glass slides (40 μl per 3.75 x 2.5 cm slide) and on non-etched fused silica slides (30 μl per 2.5 x 2.5 cm slide). The slides were then dried in an oven at 100°C for 10 minutes. On part ofthe slides, a second layer, consisting ofthe TiO₂ precursor coating solution, mentioned in Example 1 , was applied in the same manner described in Examples 1,2. For reference, several other etched glass and fused silica slides were coated with two layers ofthe same TiO₂ precursor coating solution without being coated with the mixed underlayer. All slides were then calcined at 450°C for 30 minutes to produce clear, well adhered films. No post-treatment with nitric acid was used.

0202The efficiency ofthe slides was measured in the manner described in Example 1. The films made ofthe TiO₂/SiO₂ mixture were totally inactive. However, films consisting of a TiO overlayer on an inactive underlayer consisting of TiO₂/SiO were found to e between 50% and 100% more efficient than films made of one layer of TiO₂, regardless ofthe substrate (Table 12). Table 12

0203Slide No. Substrate Structure Efficiency

0204A Fused Silica Mixed SiO₂/TiO₂ (1:1) layer 0.3 x 10°

0205B Fused Silica Mixed SiO₂/TiO₂ (1:1) layer 1.7 x 10°

0206C Fused Silica One layer of TiO₂ on a mixed 13.7 x 10° SiO₂/TiO₂ (1:1) layer

0207D Fused Silica One layer of TiO₂ on a mixed 16.1 x 10° SiO₂/TiO₂ (1:1) layer

0208E Fused Silica One layer of TiO₂ 4.9 x 10°

0209F Fused Silica One layer of TiO₂ 9.0 x 10°

0210G Fused Silica Two layers of TiO₂ 13.1 x 10°

0211H Fused Silica Two layers of TiO₂ 17.2 x 10°

0212684 Etched Glass Mixed SiO₂/TiO₂ (1:1) layer 0.0 x 10°

0213685 Etched Glass Mixed SiO₂/TiO₂ (1:1) layer 0.0 x 10°

02141711 Etched Glass One layer of TiO₂ on a mixed 16.6 x 10° SiO₂/TiO₂ (1:1) layer-

02151712 Etched Glass One layer of TiO₂ on a mixed 16.7 x 10° SiO₂/TiO₂ (1:1) layer

02161701 Etched Glass One layer of TiO₂ 8.6 x 10°

02171702 Etched Glass One layer of TiO₂ 8.6 x 10°

02181707 Etched Glass Two layers of TiO₂ 18.7 x 10°

02191708 Etched Glass Two layers of TiO₂ 17.4 x 10°

0220Table 12: The photoefficiencies of films composed of one TiO₂ layer on a silica/ titania underlayer in comparison to films made of one TiO₂ layer, two TiO₂ layers, and a mixture of TiO₂/SiO₂.