Classifications

• • C—CHEMISTRY; METALLURGY
  • C03—GLASS; MINERAL OR SLAG WOOL
  • C03C—CHEMICAL COMPOSITION OF GLASSES, GLAZES OR VITREOUS ENAMELS; SURFACE TREATMENT OF GLASS; SURFACE TREATMENT OF FIBRES OR FILAMENTS MADE FROM GLASS, MINERALS OR SLAGS; JOINING GLASS TO GLASS OR OTHER MATERIALS
  • C03C23/00—Other surface treatment of glass not in the form of fibres or filaments
  • C03C23/007—Other surface treatment of glass not in the form of fibres or filaments by thermal treatment
• • C—CHEMISTRY; METALLURGY
  • C03—GLASS; MINERAL OR SLAG WOOL
  • C03C—CHEMICAL COMPOSITION OF GLASSES, GLAZES OR VITREOUS ENAMELS; SURFACE TREATMENT OF GLASS; SURFACE TREATMENT OF FIBRES OR FILAMENTS MADE FROM GLASS, MINERALS OR SLAGS; JOINING GLASS TO GLASS OR OTHER MATERIALS
  • C03C2218/00—Methods for coating glass
  • C03C2218/30—Aspects of methods for coating glass not covered above
  • C03C2218/35—Exuding
• • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
  • Y10—TECHNICAL SUBJECTS COVERED BY FORMER USPC
  • Y10T—TECHNICAL SUBJECTS COVERED BY FORMER US CLASSIFICATION
  • Y10T428/00—Stock material or miscellaneous articles
  • Y10T428/13—Hollow or container type article [e.g., tube, vase, etc.]
  • Y10T428/131—Glass, ceramic, or sintered, fused, fired, or calcined metal oxide or metal carbide containing [e.g., porcelain, brick, cement, etc.]

Definitions

• the present invention relates to a silicate glass article, such as a glass container, with a modified surface region.
• the modified surface has, among other advantageous properties, an improved chemical durability, an increased hardness, and/or an increased thermal stability, such as thermal shock resistance.
• the present invention relates to a process for modifying a surface region of a silicate glass article.
• the surface of a Fe 2+ -bearing silicate glass article can be modified by using redox reactions.
• the new surface can be obtained by heat-treating the said glass article in atmospheric air at temperatures near the glass transition temperature (T 9 ) for suitable durations.
• the heat-treatment leads to oxidation of ferrous iron (Fe 2+ ) to ferric iron (Fe 3+ ), which causes a diffusion of divalent cations (primarily Mg 2+ ) from the interior of the glass towards the surface (called outward diffusion).
• the oxidation process does not cause oxygen to diffuse into said glass article to a significant degree as Fe 3+ is formed due to reaction between electronic species (electron holes) and Fe 2+ .
• a crystalline layer forms on the surface of the glass article as the divalent cations react with oxygen at the surface.
• This surface layer exhibits excellent thermal performance, i.e., this finding has potential to be industrially applied.
• the effect of the surface layer on the physical and chemical properties (mechanical properties, chemical durability, inertness, optical properties, etc.) of glassy materials is still unknown.
• thermochemical reduction of alkali-lead-silicate glass does not lead to any significant incorporation of hydrogen in the surface, but greatly sensitizes the surface to the chemical and physical adsorption of water.
• the treatment creates a thin (25 nm) compositionally modified layer, perhaps microporous, of silica-rich glass at the surface due to depletion of Pb where the hydrogen concentration due to adsorption is irreversible.
• the time and temperature of thermochemical treatment influence the initial kinetics of the adsorption.
• soda-lime-silica glass can be improved by thermal treatment of the glass in N 2 and NH 3 gases. They argue that the observed strength improvement might be due to the reaction of water and ammonia with the glass. This causes an exchange process of network-modifying cations with protons.
• the soda-lime-silica glass is turned into a glass of a high silica content by a process that is rate-limited by the diffusion of water.
• the US patent 3,460,927 describes a thermal treatment method that improves the flexural strength (the ability to resist deformation under load) of a polyvalent element-containing glass by reducing it in a hydrogen atmosphere. The treatment is carried out well below the glass transition temperature.
• an improved method to create silicate glass products with a relatively thick surface of high silica content would be advantageous as an economically more favourable option than using the expensive bulk silica in the production, or as a favourable option to the use of coatings of metal oxides or polymers, ion exchange between glass and salt melt, fire polishing and so on.
• an object of the present invention relates to improving the properties of silicate glass articles.
• one aspect of the invention relates to a silicate glass article comprising a bulk part and a surface region, said silicate glass article comprises network- modifying cations (NMC); wherein the concentration of the network-modifying cations in the surface region is lower than in the bulk part, wherein the composition in the surface region of the network-modifying cations is a consequence of an inward diffusion.
• NMC network- modifying cations
• the invention is particularly, but not exclusively, advantageous for obtaining an improved silicate glass article having improved chemical durability, an increased hardness, and/or an increased thermal stability.
• the network modifying cations occupy interstitial positions within the network and thereby create non-bridging oxygens.
• NMC network modifying cations
• the mechanical properties e.g. the hardness
• the mechanical properties are augmented due to the increased connectivity of the surface layer resulting in an increased effective silica concentration in the said surface layer.
• the glass type comprises a relatively large weight percentage of silica such as 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, or 90%.
• the silicate glass article according to the invention has a weight percentage of silica of at least 10-35%, preferably at least 30-49%, and even more preferably at least 50%.
• Other components than silicate may be comprised in the silicate glass article, such as alkali oxides, alkaline earth oxides and polyvalent metal oxides.
• the silicate glass article according to the invention has a weight percentage of alkali oxides of at least 0- 90%, such as 0.5-85%, preferably at least 1-80%, such as 3-75%, preferably at least 5-50%, such as 7-30%, preferably at least 10-20%.
• the silicate glass article according to the invention has a weight percentage of alkaline earth oxides of at least 0-90%, such as 0.5-85%, preferably at least 1-80%, such as 3-75%, preferably at least 5-50%, such as 7- 30%, preferably at least 10-20%.
• the silicate glass article according to the invention has a weight percentage of polyvalent metal oxides of at least 0.001-90%, such as 0.5-85%, preferably at least 1-80%, such as 3-75%, preferably at least 5- 50%, such as 7-30%, preferably at least 10-20%.
• the surface layer exerts a strong impact on the surface properties as silica increases the connectivity of the glass. In particular, it considerably enhances the chemical durability (in both acid and alkali solutions) and the hardness of the glass.
• the silicate glass article according to the invention has a silicate bridging oxygen content that is substantially higher in the surface region than in the bulk region, i.e. the network connectivity of the surface region is higher than that of the bulk region.
• the silicate glass article according to the invention has a decrease in the number of non-bridging oxygen atoms per tetrahedron, NBO/T, in the surface region of at least 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 100%.
• the silicate glass article according to the invention has a concentration of SiO 2 in the surface region that is substantially higher than in the bulk part.
• Oxidation of an iron-bearing glass by thermal treatment in atmospheric air causes the Mg 2+ , Ca 2+ , and Fe 2+ ions to diffuse from the interior towards the surface (called outward diffusion). This observation is consistent with the results of previous studies based on basaltic glass systems.
• the diffusion of Mg 2+ is predominant in the overall diffusion process, and at the surface, Mg 2+ ions react with external oxygen to form periclase (MgO) crystals.
• MgO periclase
• the Fe 2+ ions that diffuse to the surface are oxidized to Fe 3+ at the surface.
• the surface region or layer enhances the hardness of the glass and protects it from attack of an acid solution, but makes it more vulnerable against an alkali solution.
• the reduction of Fe 3+ to Fe 2+ or V 5+ to V 4+ in H2/N2 (10/90) operates by permeation of H 2 into the glass.
• hydroxyl groups form and are incorporated into the glass structure.
• the incorporation of hydroxyl groups increases the rate of the cationic diffusion even though the reduction of Fe 3+ does not cause the diffusion in H2/N2 (10/90).
• the created OH groups reduce the stability of the glass against crystallization and the mechanical properties of the glass.
• the inward diffusion is caused by reduction by a reducing gas and/or a reducing liquid.
• the thickness of the silica-rich layer can be controlled by the content of the polyvalent element. Therefore, in another embodiment according to the invention, the depth of the surface region is a function of the inward diffusion process.
• the composition in the surface region of the network-modifying cations is a consequence of inward diffusion, wherein the inward diffusion is caused by reduction of a polyvalent element.
• the reduced element has lower mobility than the earth alkaline ions in the glass network.
• the thickness of the silica-rich layer can also be controlled by tuning the temperature and duration of the heat. Therefore, in still another embodiment according to the invention, the depth of the surface region is a function of time, temperature, field strength of diffusing ions, partial pressure of the reducing gas, concentration and redox ratio of a polyvalent element, and/or glass type. Hence, the layer thickness can be tuned according to specific requirements.
• the diffusion is characterized by chemical diffusion.
• the diffusion is rate-limited by the reduction kinetics in a manner where the diffusion is parabolic with time.
• the polyvalent element should in certain embodiments of the invention have a redox state that is relatively easy to reduce in a weak reducing atmosphere, e.g. in about 0.001, 0.01, 0.02, 0.03, 0.07, or 0.09 bar H 2 .
• the element and the redox state may determine the color of the glass article depending on the glass application, e.g. transparency of glass, specific color for a specific application, art glass, specific UV-absorption to protect, e.g., medicals, beer, wine, and other liquids or chemicals from degradation.
• the present invention relates to a silicate glass article, said silicate glass article being: a glass container for storage of chemicals, a glass fiber, art glass, a glass container for storage of beer, wine, and other liquids.
• the present invention is advantageous for storage of harsh or aggressive chemicals or for use in mechanically harmful environments.
• the silicate glass is transparent in the optical range of 10-1200 nm, preferably in the visible range of 380-750 nm.
• the silicate glass is capable of absorbing UV-light in the range of between 400-10 nm, 400-315 nm, 315-280 nm, or 280-100 nm, preferably in the range of between 400-100 nm.
• the Vickers hardness (H v ) test has been developed as a method to measure the hardness of materials.
• the Vickers hardness measurements reveal that the heat-treated glasses are harder than the original glass.
• the hardness increases with duration and temperature of the heat- treatment, i.e., the hardness increases when the thickness of the modified layer increases.
• the silicate glass article has a hardness of the silicate glass that is substantially higher in said surface region than in the corresponding surface region of untreated glass, e.g. at least +10%, +20%, +30%, +40%, +50%, +100%, +200%, +300%, +1000% higher H v .
• the said nanolayer exerts a strong impact on the surface properties as silica increases the connectivity of the glass. In particular, it may considerably enhance the chemical durability in both acid and alkali solutions. In acid solutions, leaching of alkali ions from the glass is the dominant dissolution mechanism. In one example of the present invention (figure 6B), the sodium leached from the glass article into a HCI solution is decreased more than five times as compared to the untreated glass.
• the silicate glass article has a chemical durability in the said surface region that is substantially higher than in the corresponding surface region of untreated glass, e.g. at least 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 3, 5, 10, 30, 50, 100, 1000 times better than in the corresponding surface region of untreated glass.
• the modified surface has, among other advantageous properties, an increased thermal stability, such as thermal shock resistance.
• the silicate glass article has a thermal shock resistance that is substantially higher than the thermal shock resistance of the corresponding untreated glass, e.g. at least 1.5, 2, 3, 5, 10, 30, 50, 100, 1000 times better than the thermal shock resistance of the corresponding untreated glass.
• the thickness of the silica-rich layer may be controlled by the content and reduction of the polyvalent element.
• the silicate glass article comprises transition metallic cations.
• the present invention relates to a silicate glass article wherein at least some of the transition metallic cations are network- modifying cations (NMC).
• NMC network- modifying cations
• the present invention relates to a silicate glass article, wherein at least some of the network-modifying cations (NMC) are from Group Ha in the Periodic Table, e.g. Be 2+ , Mg 2+ , Ca 2+ , Sr 2+ , Ba 2+ , and Ra 2+ .
• NMC network-modifying cations
• the present invention relates to a silicate glass article, wherein the polyvalent element is selected from a group consisting of: Au, Ir, Pt, Pd, Ni, Rh, Co, Mn, Ag, Se, Ce, Cr, Sb, Cu, U, Fe, As, Te, V, Bi, Eu, Ti, Sn, Zn, and Cd.
• the polyvalent element is selected from a group consisting of: Au, Ir, Pt, Pd, Ni, Rh, Co, Mn, Ag, Se, Ce, Cr, Sb, Cu, U, Fe, As, Te, V, Bi, Eu, Ti, Sn, Zn, and Cd.
• the present invention relates to a silicate glass article, wherein the transition metallic cations are selected from a group consisting of: Ti 4+ , Ti 3+ , V 5+ , V 4+ , V 3+ , Cr 6+ , Cr 5+ , Cr 3+ , Mn 7+ , Mn 6+ , Mn 5+ , Mn 4+ , Mn 3+ , Fe 5+ , Fe 4+ , Fe 3+ , Co 4+ , Co 3+ and Ni 3+ .
• the present invention relates to a silicate glass article, wherein the transition metallic cations are selected from a group consisting of:, Ti 2+ , V 2+ , Cr 2+ , Mn 2+ , Fe 2+ , Co 2+ , Ni 2+ , Cu 2+ , Zn 2+ , , Zr 2+ , Nb 2+ , Mo 2+ , Ru 2+ , Rh 2+ , Pd 2+ , Ag 2+ , Cd 2+ , , Ta 2+ , W 2+ , Re 2+ , Os 2+ , Ir 2+ , Pt 2+ , Hg 2+ and Ra 2+ .
• the transition metallic cations are selected from a group consisting of:, Ti 2+ , V 2+ , Cr 2+ , Mn 2+ , Fe 2+ , Co 2+ , Ni 2+ , Cu 2+ , Zn 2+ , , Zr 2+ , Nb 2+ , Mo 2+
• the process of the invention leads to a silicate glass article with a surface of high silica content, thereby avoiding the need to produce glass articles of the bulk silica glass.
• the latter requires very high temperature (up to 2400 0 C) for melting and forming. Therefore, the present invention is economically more favorable than bulk silica glass production.
• the invention creates an improved silicate glass article, having improved chemical durability, an increased hardness, and/or an increased thermal stability, without using the extrinsic coating technology that requires additionally expensive raw materials.
• another aspect of the invention relates to a process for modifying a surface region of a silicate glass article, said process comprises the step of heat-treating the silicate glass article in an atmosphere comprising a reducing gas, said process resulting in an inward diffusion of the network-modifying cations (NMC) into deeper regions of the silicate glass article, whereby the concentration of the network-modifying cations in the surface region is lowered.
• NMC network-modifying cations
• Still another aspect of the invention relates to said process, wherein the reducing gas is a mixture of one or more reducing gasses. Still another aspect of the invention relates to said process, wherein the reducing gas is further mixed with one or more inert gasses.
• a preferred aspect of the invention relates to said process, wherein the atmosphere comprises a mixture of nitrogen gas and hydrogen gas.
• Another preferred aspect of the invention relates to said process, wherein the atmosphere comprises a mixture of carbon monoxide gas and carbon dioxide gas.
• the atmosphere comprises a mixture of gasses selected from a group consisting of: SbH 3 , AsH 3 , B 2 H 6 , CH 4 , PH 3 , SeH 2 , SiH 4 , SH 2 , SnH 4 , Cl 2 , NO, N 2 O, CO, H 2 , N 2 O 4 , SO 2 , C 2 H 4 , and NH 3 .
• a preferred aspect of the invention relates to said process, wherein the reducing gas is substantially impermeable in the untreated silicate glass.
• a preferred aspect of the invention relates to said process, wherein the heat-treatment is performed so as to obtain a thickness of said surface region of at least 100 nm, 200 nm, 400 nm, 500 nm, 600 nm, 700 nm, 1000 nm, 1500 nm, or 3000 nm.
• the thickness of the silica-rich layer can be controlled by tuning the temperature and duration of the heat-treatment.
• a preferred aspect of the invention relates to said process, wherein the heat-treatment is performed at e.g. 0.1-3.0, 0.5-3.0, 0.6-3.0, 0.7-3.0, 0.8-2.0, or 0.9-2.0 times the glass transition temperature (T 9 ) of the silicate glass.
• Another aspect of the invention relates to said process, wherein the heat- treatment is performed in the interval of 0.001-36, 0.01-36, 0.1-36, 0.1-30, 0.1- 24, 0.2-36, 0.2-34, 0.2-20, 0.3-36, 0.3-25, 0.3-18, 0.4-36, 0.4-27, 0.4-12, 0.5- 36, 0.5-15, 1-5, 1-4, or 1-3 hours. Even shorter or longer times are within the teaching of the invention.
• Regulating the pressure of the surrounding atmosphere in said process has an important impact on the temperature and/or duration of heat-treatment.
• Yet another aspect of the invention relates to said process, wherein the pressure of the said atmosphere in the interval of 0.001-20 atm., 0.001-10 atm., 0.01-10 atm., 0.01-5 atm., 0.1-5 atm., or 1-10 atm. Any of the lower limits in the said intervals may also be minimum values.
• FIGS. IA-B show schematic representations of different proposed mechanisms of surface modification, IA shows the formation of an MgO/CaO layer, IB shows the formation of a silica-rich layer,
• Figure 2A-D show schematic representations of SNMS depth profiles of the untreated 6wtFe glass and of the 6wtFe glass heated in H2/N2 (1/99) at different conditions
• 2A shows a profile of the original 6wtFe glass
• 2B shows a profile of the 6wtFe glass heated at T 9 for 2 hours
• 2C shows a profile of the 6wtFe glass heated at T 9 for 16 hours
• 2D shows a profile of the 6wtFe glass heated at 1.05 T 9 for 2 hours
• Figure 3A-D show schematic representations of FT-IR reflectance spectra of the untreated and heat-treated 6wtFe glass at different conditions
• 3A shows a spectra of the 6wtFe glass heated at T 9 for different durations in H2/N2 (10/90)
• 3B shows a spectra of the 6wtFe glass heated for 2 hours at different temperatures in H2/N2 (10/90)
• 3C shows a spectra of the
• Figure 4A shows a plot of squared thickness of the divalent-cation-depleted region ( ⁇ ) versus heat-treatment duration (t a ) for the 6wtFe glass samples in H2/N2 (1/99) at T 9
• figure 4B shows a table describing the dependence of the difference in the Fe 2+ concentration before and after treatment (Ac(Fe 2+ )) on the initial iron- content of the glass and the heat-treatment condition
• 4C shows CEMS spectra of 6wtFe glasses (untreated, heated in air at T 9 for 16 h, and heated in H2/N2 (1/99) at Tg for 16 h,
• Figure 5A shows a UV-VIS-NIR spectra of 0.20 mm thick 6wtFe glass samples heated in H2/N2 (1/99) at T 9 for different durations
• 5B shows the corresponding Fe 2+ concentrations expressed as the dependence of the difference in the Fe 2+ concentration before and after treatment (Ac(Fe 2+ )) on the heat-treatment duration (t a ),
• Figure 6A shows a table with data of Vickers hardness and water contact angle of the untreated and thermally treated 6wtFe glasses at different temperatures and durations
• 6B shows a table with data of chemical durability of the untreated and thermally treated 6wtFe glasses at different temperatures and durations
• FIG. 7 shows a schematic overview of the experimental strategy and the employed analytical techniques
• FIG 11 shows the onset crystallization temperature T c as a function of ⁇ n re ⁇ ,dyn, ⁇ n re ⁇ ,dyn is the normalized number of moles of the polyvalent element that were oxidized during dynamic heating in air at 10 °C/min to 975 0 C, T c was determined from a DSC experiment performed at 10 °C/min in air, and
• polyvalent element can be found in numerous articles in the field of glass science and technology. In the present context this term will refer to an element which may exist in different redox states. The best direct definition one may find in Pye et al. (2005) p. 28: "In this chapter, all those elements will be considered to be polyvalent, which may occur in a glass melt in at least two different oxidation states, even if extreme oxidizing or reducing conditions are necessary.”
• the glass transition temperature (T 9 ) is defined as the onset of change of heat capacity due to the glass transition when heating a glass as defined in Shelby (2005).
• SRO short-range order
• LRO long-range order
• SRO exists when the local atomic bonding units (nearest neighbour configuration of atoms) are uniform in the entire solid.
• LRO exists when the arrangement of atoms in space is periodic [Gersten & Smith, 2001].
• SRO existing in a given glass is ideally identical to that found in the corresponding crystals.
• Crystals are defined to be solids with perfect LRO which implies perfect periodicity of the atomic arrangement.
• At least three corners in each oxygen polyhedra must be shared in order to form a three-dimensional network.
• the chemical components in an oxide glass can be divided into different categories according to their role in the structural arrangement of the glass.
• Stanworth (1971) has classified oxides into three groups based on the electronegativity of the cation, i.e., the oxides are classified according to the fractional ionic character of the cation-anion bond as the anion is oxygen in every case. If the cation forms bonds with oxygen with a fractional ionic character near or below 50 %, the cation will act as a network former [Shelby, 2005]. All glasses contain at least one network former as it is the primary source of the structure.
• silicon acts as the network former and it exists as silicon- oxygen tetrahedral that are linked by bridging oxygen (BO) atoms.
• BO bridging oxygen
• the tetrahedral themselves are very ordered.
• the required lack of LRO is introduced by variability in the Si-O-Si angle, rotation of adjacent tetrahedral around the point occupied by the oxygen atom linking the tetrahedral, and rotation of the tetrahedral around the line connecting the linking oxygen with one of the silicon atoms [Shelby, 2005].
• Cations which form highly ionic bonds with oxygen are termed network modifiers as they only serve to modify/interfere with the network structure without becoming part of the primary network [Shelby, 2005].
• Network modifiers provide non-bridging oxygen (NBO) atoms with a negative charge as they are introduced as oxides and have coordination number > 6.
• NBO non-bridging oxygen
• Both alkali (e.g., Na + and K + ) and alkaline earth (e.g., Ca 2+ and Mg 2+ ) ions can act as modifiers. Every alkali ion has one neighbouring NBO, while every alkaline earth ion has two neighbouring NBOs.
• the strength of the network is dependent on the amount of network formers and modifiers. An increase in the amount of network modifiers results in an increase in the amount of NBOs which decreases the connectivity (or the degree of polymerization) of the. This lowers the melting temperature and several other properties of the glass [Shelby, 2005].
• Figures IA-B show schematic representations of two proposed mechanisms of surface modification, for explaining the present invention.
• Figure IA shows a known mechanism for the formation of a crystalline MgO/CaO layer 2 on a silicate glass sample or article 1 comprising Mg 2+ , Ca 2+ and Fe 3+ .
• the glass sample 1 is illustrated as having a surface 6, a surface region 3, a bulk part 4, and a so-called redox front 5.
• the heat-treatment leads to oxidation of ferrous iron (Fe 2+ ) to ferric iron (Fe 3+ ), which causes an outward diffusion of divalent cations (primarily Mg 2+ ) from the interior of the glass towards the surface.
• a crystalline layer 2 forms on the surface 6 as the divalent cations react with ionic oxygen at the surface. This surface layer 2 exhibits excellent thermal performance.
• Figure IB shows a mechanism according to the present invention for the formation of a silica-rich layer in the surface region 3.
• the schematic representation is a still shoot of a dynamic process.
• the Fe 3+ ions are converted to Fe 2+ ions and electron holes (h * ).
• the extremely low partial oxygen pressure in the atmosphere provides a large driving force for the removal of oxygen from the glass article 1.
• oxygen anions surrender two electrons to fill the h * and are subsequently released from the free surface 6 via reaction with H 2 to form H 2 O.
• the diffusion of h * towards the surface is charge- balanced by an inward migration of the divalent cations (including Fe 2+ ) as a mirror-image of the oxidation mechanism.
• the inward diffusion is driven by reduction of the high valence to the low valence state of the polyvalent element.
• the network modifying cations in this example Mg 2+ , Ca 2+ and Fe 3+
• a silica-rich surface layer 3 is formed.
• oxygen anions surrender the electrons to h * at the surface, H 2 molecules in the surrounding atmosphere are ultimately the source of the electrons.
• one aspect of the invention relates to a silicate glass article 1 comprising a bulk part 4 and a surface region 3, said silicate glass article comprises network- modifying cations (NMC), e.g. Mg 2+ , Ca 2+ , and Fe 2+ as indicated in the Figure IB.
• NMC network- modifying cations
• concentration of the network-modifying cations (NMC) in the surface region 3 is lower than in the bulk part 4, and generally speaking the composition in the surface region of the network-modifying cations is a consequence of above- mentioned inward diffusion as it will be explained in more detail below.
• Figure 2A-D show schematic representations of secondary neutral mass 5 spectroscopy (SNMS) depth profiles of the original 6wtFe glass and of the 6wtFe glass heated in H2/N2 (1/99) at different conditions.
• the depth profile of each element is normalised to the bulk concentration.
• the H 2 partial pressure is lowered to 0.01 bars in order to create the silica-rich surface by decreasing the rate of the gaseous permeation.
• SNMS depth profiles show that this effort was 10 successful.
• Table 1 The chemical composition of the glasses and various other relevant data is given in Table 1 below:
• NBO/T non- bridging oxygen atoms per tetrahedron
• NBO/T for the untreated 6wtFe glass is 0.81 as stated in Table 1.
• the inventors have calculated the NBO/T ratio in a 200 nm surface layer of the glass treated in U 2 /N 2 (1/99 v/v) at T 9 for 16 h. This surface layer has NBO/T ⁇ 0.45. This is the only sample for which the inventors at present have calculated NBO/T.
• Figure 2A shows a profile of the original 6wtFe glass.
• the depth profile of the untreated glass reveals that the ion concentrations do not vary with depth.
• Figure 2B shows a profile of the 6wtFe glass heated at T 9 for 2 hours.
• Heat- treatment of the 6wtFe glass under a U 2 /N 2 (1/99) gas at T 9 , for 2 h leads to the inward migration of the divalent cations and a remarkable increase in the silica concentration near the surface.
• FIG. 2D shows a profile of the 6wtFe glass heated at 1.05 T 9 for 2 hours.
• Figure 2B it is seen that by increasing the temperature of the heat-treatment, the result is an increase in the thickness of the modified surface layer, i.e. the depth resulting from the combined heating and reduction according to the present invention is larger.
• Figure 3A-D show schematic representations of FT-IR reflectance spectra of the untreated and heat-treated 6wtFe glass at different conditions.
• IR reflectance spectra of the heat-treated samples only data in the range 900- 1200 cm “1 will be shown as no changes occur at lower wavenumbers. This is consistent with previous studies as the position and intensity of a peak at 480 cm " 1 vary little with glass composition.
• Figure 3A shows spectra of the 6wtFe glass heated at T 9 for different durations in H2/N2 (10/90)
• figure 3B shows spectra of the 6wtFe glass heated for 2 hours at different temperatures in H2/N2 (10/90).
• the FT- IR reflectance spectrum displays major peaks near 480 cm "1 and 1100 cm "1 that are assigned to Si-O-Si bond rocking and Si-O-Si antisymmetric stretching vibration, respectively.
• Fig. 3A heat-treatment duration
• Fig. 3B temperature
• Figure 3C shows spectra of the 6wtFe glass heated at T 9 for different durations in H2/N2 (1/99), and figure 3D shows spectra of the 6wtFe glass heated for 2 hours at different temperatures in H2/N2 (1/99).
• the peaks assigned to the vibration of Si-OH and Si-N bonds are also present in these IR spectra, but the intensities of the peaks are lower than those observed for glasses heated in H2/N2 (10/90).
• the IR absorption measurements show that less OH groups are formed with decreasing hydrogen pressure. This explains the lower intensity of the Si-OH peaks.
• the Si-O-Si antisymmetric stretching wavenumber and peak intensity increase with increasing t a and T 3 .
• Figure 4A shows a kinetic analysis by a plot of squared thickness of the of the modified surface region ( ⁇ ) versus heat-treatment duration (t a ) for the 6wtFe glass samples in H2/N2 (1/99) at T 9 .
• ⁇ modified surface region
• t a heat-treatment duration
• Figure 4B shows a table describing the dependence of the difference in the Fe 2+ concentration, before and after treatment ( ⁇ c(Fe 2+ )), on the initial iron-content of the glass and the heat-treatment condition.
• the untreated glasses contain more Fe 3+ ions with increasing total iron content.
• Figure 4B reveals that ⁇ c(Fe 2+ ) increases with increasing total iron content of the glass.
• FIG. 4C shows conversion electron M ⁇ ssbauer spectroscopy (CEMS) spectra of 6wtFe glasses (untreated, heated in air at T 9 for 16 h, and heated in H2/N2 (1/99) at Tg for 16 h), the fitted doublets of Fe 3+ and Fe 2+ are shown.
• CEMS can be used to study the iron redox state in the surface region ( ⁇ 200 nm) of a sample, i.e., it is different from conventional M ⁇ ssbauer spectroscopy that determines the redox state in the bulk.
• conventional M ⁇ ssbauer spectroscopy the absorption peaks of the resonantly absorbed gamma rays are recorded.
• CEMS the energy released from the excited (metastable) iron nuclei in the sample is studied.
• the excited iron nuclei in the sample return to their ground state by three processes.
• Approximately 90 % of the absorbed energy is released by so-called internal conversion and approximately 10 % is released as gamma rays.
• the internal conversion includes transfer of the energy via X-rays or to so-called conversion electrons.
• the conversion electrons are emitted because the excited nucleus can transfer its energy to an electron that has a certain probability of being in the nucleus.
• the conversion electrons emitted from the excited nuclei are recorded. These electrons are strongly attenuated when they pass through the sample, i.e., the signals only come from the uppermost layer (approximately 200 nm) of the sample.
• the isomer shifts of Fe 3+ and Fe 2+ are determined to 0.27 ⁇ 0.06 and 1.07 ⁇ 0.09 mm/s, respectively.
• the quadrupole splittings are 1.13 ⁇ 0.09 and 1.7 ⁇ 0.2 mm/s for Fe 3+ and Fe 2+ , respectively. These values are in good agreement with literature data.
• the Fe 3+ /Fe to t ratio is estimated for each sample by measuring the relative areas of the two doublets and assuming that no metallic iron is present in the glasses. The calculated ratios are stated in Figure 4C.
• the relatively high errors of the ratios are due to i) the use of a weak source and ii) the small surface areas of the samples.
• Fe 3+ /Fe to t equals 68 ⁇ 7 % for the untreated 6wtFe glass. This is consistent with the result found by conventional M ⁇ ssbauer spectroscopy that determined the redox ratio of a powdered sample. As expected, heat- treatment of the glass in air results in an increased amount of Fe 3+ ions compared to the amount of Fe 2+ ions near the surface, whereas the opposite is valid for treatment in U 2 ZN 2 (1/99).
• FIG. 5A shows UV-VIS-NIR spectra of 0.20 mm thick 6wtFe glass samples heated in H2/N2 (1/99) at T 9 for different durations.
• the iron redox state is investigated as a function of the different heat-treatment conditions.
• UV-VIS-NIR spec-troscopy is the main method used for that purpose. To use this method quantitatively, the molar absorption coefficient of Fe 2+ is preliminarily determined.
• Figure 5B shows the corresponding Fe 2+ concentrations expressed as the dependence of the difference in the Fe 2+ concentration before and after treatment (Ac(Fe 2+ )) on the heat-treatment duration (t a ).
• the increase in the intensity of the Fe 2+ peak with increasing t a is less than the observed increase for glasses heated in H2/N2 (10/90)(not shown).
• the iron redox ratio is shifted to the more reduced state with increasing hydrogen partial pressure in the treatment atmosphere. This is explained by the increased solubility of H 2 (S H 2) in the glass at higher pressures.
• Figure 6A shows a table with data of Vickers hardness (H v ) and water contact angle of the untreated and thermally treated 6wtFe glasses at different temperatures and durations.
• the Vickers hardness is measured by microindentation. 25 indentations were performed for each sample at widely separately locations with a load of 0.25 N and a hold time at the maximum load of 5 s.
• the Vickers hardness measurements reveal that the heat-treated glasses are harder than the original glass.
• the hardness increases with duration and temperature of the heat-treatment, i.e., the hardness increases when the thickness of the modified layer increases.
• the contact angle measurements show that the surface becomes more hydrophobic as a result of the thermal treatments. Hardness measurements were done with accuracies better than ⁇ 0.3 GPa.
• Figure 6B shows a table with data of chemical durability of the untreated and thermally treated (in H2/N2 (1/99)) 6wtFe glasses at different temperatures and durations.
• the chemical resistance of the samples was examined in 0.25 M HCI and 0.25 M KOH solutions. After immersing a sample in plastic container with the test solution (20 cm 3 for 1 cm 2 of the glass surface area), the container was mounted on a thermostatic shaking assembly at 90 0 C (agitated at 100 rpm). After 12 h, the sample was removed from the solution. The concentrations of leached Na + and Mg 2+ ions were measured in the test solution using atomic absorption spectroscopy (AAnalyst 100, Perkin Elmer).
• the dissolution of the glasses was tested in both acid and alkali solutions.
• acidic solutions primarily the monovalent alkali ions leave the glass and are replaced by H + and/or H 3 O + .
• alkali solutions the liquid directly attacks the network bonds as hydroxyl ions can break the Si-O bonds leading to the formation of silanolgroups and hence, a continuous dissolution of the glass.
• the thermally treated glasses possess a higher resistance towards both acid and alkali solutions than the untreated glass (cf. Fig. 6B).
• the increase in alkali resistance is caused by the high network connectivity of the treated glasses.
• the network modifying cations NMC occupy interstitial positions within the network creating nonbridging oxygens (NBO).
• the connected network on the surface of the treated glasses makes it difficult for ions to diffuse through the glass, impeding ions such as OH " and H + to penetrate the network and react with the glass species.
• OH " diffusion is difficult in the thermally treated glasses increasing their alkali resistance.
• the increase in acid resistance is caused by the impeded diffusion of H + and to a minor extent the depletion of sodium near the surface in the treated samples.
• the hardness and chemical durability of the glasses are increased as a result of the surface modification resulting from the combined heating and reduction according to the present invention.
• NMC network modifying cations
• a change of T a or t a can preferably be used to change the extent of the surface modification.
• the heat-treatment atmosphere determines primarily how the surface is modified in terms of composition, morphology, and/or redox state. The effects of the heat-treatment atmosphere on the investigated glass properties are summarized in Table 2 below:
• Table 2 may be used to select the appropriate surface modification method in order to achieve some desired properties. For most applications of glasses, the effects of treatment in H2/N2 (1/99) are the most favourable.
• Figure 7 shows a schematic overview of the experimental strategy and the employed analytical techniques.
• a r A is stated for a coordination number of 6. T 9 and m have been determined by DSC and viscosity measurements, respectively.
• All iron is reported as Fe 2 O 3 .
• c n.d. not determined. Cylindrical glass samples (diameter ⁇ 8-10 mm; thickness 3 mm) were prepared. The samples for the diffusion experiments were ground flat on one surface to a thickness of ⁇ 2 mm by a six-step procedure with SiC paper under ethanol. The surfaces were carefully polished afterwards with 3 ⁇ m diamond paste and finally cleaned with acetone. To study the reduction reactions, ultraviolet-visible-near- infrared (UV-VIS-NIR) spectroscopy measurements were performed. The samples for these experiments were ground coplanar to achieve uniform thickness, and then they were polished to a thickness of 0.2 mm using the above-mentioned procedure.
• UV-VIS-NIR ultraviolet-visible-near- infrared
• T 9 glass transition temperature
• DSC differential scanning calorimetry
• C p isobaric heat capacity
• T 9 was determined by a second upscan at 10 K/min in order to ensure a uniform thermal history of the glasses.
• T 9 was defined as the cross point between the extrapolated straight line of the glass C p curve before the transition zone and the tangent at the inflection point of the sharp rise curve of C p in the transition zone.
• Viscosity was measured by beam-bending (T > T 9 ) and concentric cylinder (T > Thq Ul dus) experiments.
• T > T 9 beam-bending
• T > Thq Ul dus concentric cylinder
• a 300 g weight was used to explore the viscosity range from approximately 10 12 to 10 9lS Pa-s at a constant heating rate of 10 K/min.
• the viscosity was calculated according to DIN ISO 7884-4.
• the low viscosities ( ⁇ 10 3 Pa-s) were measured using a concentric cylinder viscometer.
• the viscometer consisted of furnace, viscometer head, spindle, and sample crucible.
• the viscometer head (Physica Rheolab MCl, Paar Physica) was mounted on top of a high temperature furnace (HT 7, Scandiaovnen A/S). Spindle and crucible were made of Pt 8 oRh 2 o-
• the viscometer was calibrated using the National Bureau of Standards (NBS) 710A standard glass.
• the polished glasses were heat-treated at 1 atm in an electric furnace under a flow of H2/N2 (1/99 v/v) gas.
• the glass samples were inserted into the cold furnace and the gas-flow was turned on.
• the furnace was then heated at 10 K/min to the pre-determined heat-treatment temperature and kept at this temperature for a given duration. Afterwards, the furnace was cooled down to room temperature at 10 K/min.
• the ternary Si-Ca-Fe and Si-Na-Fe glasses were treated at their respective T 9 for 2 h.
• compositional analysis of the surfaces was carried out using electron-gas secondary neutral mass spectroscopy (SNMS).
• SNMS electron-gas secondary neutral mass spectroscopy
• the measurements were performed on an INA3 (Leybold AG) instrument equipped with a Balzers QMH511 quadrupole mass spectrometer and a Photonics SEM XP1600/14 amplifier.
• the analyzed area had a diameter of 5 mm and was sputtered using Kr plasma with an energy of ⁇ 500 eV.
• the time dependence of the sputter profiles was converted into depth dependence by measuring the depth of the sputtered crater at 12 different directions on the same sample with a Tencor Pl profilometer.
• A Na, K, Rb, Cs.
• An increase in viscosity with increasing ionic radius r A of the alkali ion at a given temperature is observed for both the low and high temperature data.
• MYEGA Mauro-Yue-Ellison-Gupta-Allan
• the fragility index is defined as the slope of the log ⁇ versus T g /T curve at T 9 :
• the viscosity at T 9 is set equal to 10 12 Pa-s since this has been shown to be equivalent to the calorimetrically measured T 9 values for oxide glasses.
• the MYEGA equation offers improved accuracy in performing low temperature extrapolations compared to the Vogel-Fulcher-Tammann (VFT) and Avramov-Milchev (AM) equations.
• VFT Vogel-Fulcher-Tammann
• AM Avramov-Milchev
• the fragility decreases with increasing size of the alkali ion.
• the glass transition temperature increases with r A .
• a maximum absorption peak is found at approximately 1075 nm, which is due to the presence of Fe 2+ ions.
• the maximum absorbance and position of the Fe 2+ peak is the same (within ⁇ 5 %) for all glasses. This indicates that the initial [Fe 3+ ]/[Fe to t] ratio is approximately the same in all the glasses.
• M ⁇ ssbauer spectroscopy we have found that the untreated Si-Ca-Fe-Na glass contains 77 ⁇ 2 % of its Fe ions as Fe 3+ .
• the Si-Ca-Fe-Na glass has been heat-treated at its T 9 of 892 K for 2, 8, 16, and 60 h.
• the absorbance of the Fe 2+ band increases because Fe 3+ is reduced to Fe 2+ .
• the change in Fe 2+ concentration increases approximately linearly with the square root of the treatment duration, implying that diffusion-controlled kinetics occurs.
• the change in absorbance of the Fe 2+ peak increases with increasing radius of the alkali ion.
• the glasses are treated at their respective T 9 , the trend is qualitatively the same but the differences between the glasses are more pronounced. This is because the Si-Ca-Fe-Cs glass has the highest T 9 and therefore it is treated at the highest temperature.
• a depletion of calcium, potassium, and iron is observed near the surface.
• the extent of the calcium depletion is larger than that of potassium and iron.
• Qualitatively all six glasses display the same type of surface depletion of network-modifying cations as a result of heat-treatment for 2 h at temperatures around their respective T 9 .
• the surface depletion is caused by an inward diffusion of these ions induced by the reduction of Fe 3+ to Fe 2+ .
• An important consequence of the inward diffusion is the creation of a silica-rich surface layer. Before heat-treatment in H2/N2 (1/99), the glasses do not show any variation in composition as a function of depth.
• the Si-Ca-Fe-Na glass has been heat-treated at its T 9 for 1 hour and 0.2 hours.
• the glass treated for 0.2 h displays an about 50 nm layer depleted in calcium and iron, whereas the concentrations of silicon and sodium are higher in this layer than in the bulk.
• the duration of the treatment is increased to 1 h, the thickness of the layer depleted in calcium and iron increases and inward diffusion of sodium occurs.
• an enrichment of sodium is observed in the depth interval from approximately 50 to 100 nm.
• the ternary Si-Ca-Fe and Si-Na-Fe glasses have been heat-treated at their respective T 9 for 2 h.
• k' is proportional to the product of the diffusion coefficient of the rate-limiting species (divalent cation) and a thermodynamic driving force (gradient in oxygen activity).
• the activation energy of calcium diffusion (f d ) can be obtained by plotting the data in Arrhenius coordinates ( Figure 9).
• the diffusion data for each glass reveal an Arrhenius dependence on temperature in the studied temperature range.
• E ⁇ is calculated from the slope of each line and is plotted as a function of the ionic radius of the alkali ion. E ⁇ decreases with increasing size of the alkali ion.
• the diffusion data show that the isothermal inward diffusion of Ca 2+ is fastest in the Si-Ca-Fe-Na glass, i.e., in the glass with the lowest molar volume.
• the former process dominates the reduction reaction at 0.01 bar of H 2 . Therefore, the thickness of modified surface layer (as measured by SNMS) cannot be directly correlated with the degree of reduction (as measured by UV-VIS-NIR spectroscopy). In other words, a large extent of Fe 3+ reduction does not necessarily result in a thick SiO 2 -rich surface layer because two processes contribute to the reduction of Fe 3+ .
• the Si-Ca-Fe-Cs glass has the highest T 9 of the glasses and it has been shown that this type of redox-induced diffusion begins at temperatures around 0.87 g (in K). Apparently, the process requires some degree of viscous softening even though the motion of the alkaline earth ions is decoupled from that of the network. Therefore, the glass with lowest Tg will have the fastest Ca 2+ diffusion when the glasses are heat-treated at the same temperature.
• alkali ions decreases the diffusivity of the alkaline earth ions because the extent of Ca 2+ diffusion is smaller in the Si-Ca-Fe-Cs glass than in the ternary Si-Ca-Fe glass. These two glasses also have approximately the same T 9 .
• the presence of relatively slow alkali ions may therefore block the diffusion of the faster alkaline earth ions in the interconnected channels, i.e., the slow alkali ions occupy interstices and hereby increase the packing density. These interstices can no longer be used for alkaline earth migration.
• An alkaline earth ion neutralizes two electron holes, whereas an alkali ion neutralizes only one electron hole.
• the alkaline earth ions are more mobile than trivalent modifier ions (e.g., Al 3+ ) because the latter ones are more strongly bound to the oxygen anions.
• the diffusion coefficient of the alkali ions is smaller in the inward diffusion process compared to what is found in the literature because the latter results have predominantly been obtained by the use of a radioactive tracer.
• Seven soda-lime silicate glasses each of which contains one of the following polyvalent metals: Fe, Mn, Cu, Ce, Ti, V, and Cr, are oxidized in air and reduced in H2/N2 (1/99) at their respective glass transition temperatures for some period.
• a crystalline oxide surface layer is created on the glasses (except the vanadium- bearing glass) under the oxidizing condition, since the metallic ions are oxidized from lower to higher valence state, and thereby calcium ions diffuse outward and react with oxygen ions.
• a silica-rich surface layer is created on the glasses under the reducing condition, since sodium and calcium ions diffuse inward.
• Seven glasses were prepared from three main analytical reagent-grade chemicals (SiO 2 , Na 2 CO 3 , and CaCO 3 ) and one minor analytical reagent-grade polyvalent metal oxide (Cr 2 O 3 , MnO 2 , CeO 2 , V 2 O 5 , CuO, Fe 2 O 3 , or TiO 2 ).
• the batch materials were melted in a Pt 90 Rhi 0 crucible in an electric furnace (SF6/17, Entech) at 1500 0 C for 3 h.
• the melt was then cast onto a brass plate and pressed to obtain cylindrical glasses of 7-10 cm diameter and ⁇ 5 mm height.
• the prepared glasses were immediately annealed at 640 0 C for 10 min and then cooled naturally down to room temperature in the closed furnace.
• the chemical compositions of the glasses were analyzed by x-ray fluorescence (S4-Pioneer, Bruker-AXS) and are listed in Table 4.
• the main impurity was AI 2 O 3 ( ⁇
• T 9 The glass transition temperature [T 9 ) was measured using differential scanning calorimetry (DSC). The DSC measurements were performed on a simultaneous thermal analyser (STA) (449C Jupiter, Netzsch). All the glasses were subjected to two runs of up- and downscans at 10 °C/min. The onset temperature of the endothermic C p (isobaric heat capacity) jump during the second upscan was assigned as T 9 .
• the STA instrument was also used for recording both DSC and thermogravimetric (TG) signals during both iso-thermal (i.e, constant temperature) and dynamic (i.e., increasing temperature at constant heating rate) heating, from which the oxidation degree of the polyvalent ions were determined.
• the measurements were performed on powdered samples by crushing and sieving the glass samples. The 45-63 ⁇ m size fraction was collected for each glass.
• a platinum crucible containing the glass sample and an empty platinum crucible were placed on the sample carrier of the STA at room temperature. To evaporate water from the samples, the crucibles were initially heated at a rate of 10 °C/min to 300 0 C and held for 15 min before cooling down to room temperature.
• both crucibles were then held 5 min at an initial temperature of 60 0 C and heated at a rate of 10 °C/min to the respective T 9 of the glasses and held at this temperature for 6 or 12 h.
• the crucibles were also held 5 min at an initial temperature of 60 0 C but were then heated at a rate of 10 °C/min to 975 0 C.
• the crucibles were cooled down to 250 0 C at a rate of 10 °C/min, and finally down to room temperature at a natural rate. Atmospheric air dried by a molecular sieve was used as purge gas.
• UV-VIS-NIR ultraviolet-visible-near-infrared Specord 200 spectrophotometer
• UV-VIS-NIR spectra were recorded over the wavelength range of 300-1100 nm. The measurements were performed on 2.0 mm thick samples ground by a six-step procedure with SiC paper, followed by polishing with 3 ⁇ m diamond suspension. The UV-VIS-NIR spectra were used to qualitatively determine the redox states of the polyvalent elements present in the untreated glasses.
• the bulk glasses were cut in cylinders of 10 mm diameter and 2-3 mm height. One surface of each sample was then ground by a six-step procedure with SiC paper, followed by polishing with 3 ⁇ m diamond suspension.
• the polished glasses were heat-treated at 1 atm in an electric furnace under a flow of H2/N2 (1/99 v/v) gas. The glass samples were inserted into the cold furnace and the gas-flow was turned on. The furnace was then heated at 10 °C/min to the respective T g of the glasses and kept at this temperature for 6 h. The diffusion process was ended by cooling the furnace to room temperature at 10 °C/min.
• the polished glasses were heat-treated by applying an identical heating procedure under atmospheric conditions.
• the diffusion profiles were determined by electron-gas secondary neutral mass spectroscopy (SNMS).
• SNMS is used to determine the elemental concentrations as a function of the depth within the glass.
• the measurements were performed by using an INA3 (Leybold AG) instrument equipped with a Balzers QMH511 quadrupole mass spectrometer and a Photonics SEM XP1600/14 amplifier.
• the analyzed area had a diameter of 5 mm and was sputtered using Kr plasma with an energy of ⁇ 500 eV.
• the time dependence of the sputter profiles was converted into depth dependence by measuring the depth of the crater at 12 different locations on the same sample with a Tencor Pl profilometer.
• UV-VIS-NIR spectra was made (not shown) of the untreated Cr- and Ti-containing glasses.
• Cr can exist as Cr 2+ , Cr 3+ , and Cr 6+ in silicate glasses. No absorption bands due to Cr 2+ are observed, whereas both Cr 3+ (at 445, 640, 660, and 690 nm) and Cr 6+ (at 360 nm) are observed.
• Ti can exist as Ti 4+ and Ti 3+ .
• Ti 4+ has d 0 electron configuration, in which only charge transfer transitions occur in the L)V range. Hence, Ti 4+ is colorless and no absorption bands can be observed in the spectrum.
• Fe and Mn can exist as Fe 2+ and Fe 3+ and Mn 2+ and Mn 3+ , respectively.
• the di- and trivalent states are detected in both glasses.
• V can exist as V 3+ , V 4+ , and V 5+ , but only V 4+ is detected.
• V 5+ is expected to be present in silicate glasses melted in air, but its strong charge transfer bands are found in the UV-range, which causes a sharp UV-absorption edge.
• Cu can exist as Cu 0 , Cu + , and Cu 2+ , and Cu 2+ is observed in the spectrum of the untreated glass, whereas Cu + is colorless.
• Ce can exist as Ce 3+ and Ce 4+ . Both redox states cause absorption peaks in the UV-range, and hence, they cannot be observed due to a sharp UV-absorption edge.
• Table 4 exhibits the chemical compositions and the T 9 values of the seven glasses studied in this work. As expected, the glasses with the higher field strength polyvalent metal ions (Ti, Ce, Cr, V) have higher values of T 9 than the rest. For example, Ti 4+ is known to act as a network former and V 5+ may serve to increase the polymerization degree of a silicate network.
• both DSC and TG measurements were carried out in air at a heating rate of 10 °C/min on powdered samples.
• the energy response of the samples to the dynamic heating and the mass change of the glasses during the heating are measured using DSC and TG, respectively.
• An increase in mass of 0.20 % at temperatures above 550 0 C was observed.
• the increase in mass is caused by oxidation of Cr 2+ to Cr 3+ and/or Cr 3+ to Cr 6+ .
• the stoichiometry of the second reaction is also valid for oxidation Of Ti 3+ to Ti 4+ , V 4+ to V 5+ , Fe 2+ to Fe 3+ , Cu + to Cu 2+ , and Ce 3+ to Ce 4+ . It is assumed that the mass increase of the samples during heating is solely due to incorporation of oxygen. M ⁇ ssbauer spectroscopy experiments have confirmed this assumption for the oxidation of Fe 2+ to Fe 3+ during heating of iron-bearing aluminosilicate glass fibers in air. The normalized number of moles that has been oxidized ( ⁇ n re ⁇ ) can then be calculated by the following equation: . x- Am
• m 0 is the initial mass of the sample
• ⁇ m is the maximum increase in mass
• M 0 is the molar mass of oxygen
• x is the ratio between the number of moles of the polyvalent element being oxidized and the number of moles of O 2 being consumed in the oxidation process, i.e., x is 4/3 for oxidation of Cr 3+ to Cr 6+ and x is 4 for the other oxidation reactions.
• the values calculated using the equation for the dynamic heating procedure ( ⁇ n re ⁇ ,dyn) are listed in Table 5.
• Cu + is oxidized to the largest extent, whereas very limited amounts Of Ti 3+ and V 4+ are oxidized.
• ⁇ n re ⁇ , ⁇ S o and ⁇ n re ⁇ ,dyn are the normalized number of moles of the polyvalent element that were oxidized during iso-thermal heating in air at T 9 for 6 h and during dynamic heating in air at 10 °C/min to 975 0 C, respectively.
• the characteristic temperatures were determined from DSC measurements performed at an upscanning rate of 10 °C/min in air.
• T c onset temperature of crystallization
• T 9 peak temperature of crystallization
• the crystallization begins at a lower temperature when a larger mass increase occurs, i.e., a higher degree of oxidation during the dynamic heating (Fig. 11).
• XRD x-ray diffraction
• AFM atomic force microscopy
• SNMS secondary neutral mass spectroscopy
• SNMS is used to determine the concentration depth profiles of the seven glasses that prior to the measurements have been oxidized in air or reduced in H2/N2 (1/99) at their respective T 9 for 6 h. It should be noticed that the untreated glasses show no changes in composition as a function of depth.
• the normalized concentration depth profiles of Si, O, Ca, and Na of the Cu-containing glass oxidized in air are not shown.
• a high surface concentration of calcium is found, which is due to outward diffusion of Ca 2+ .
• the low surface concentration of sodium and silicon is due to the enrichment of calcium and oxygen near the surface.
• concentration profiles of the Cu-containing glass as a result of heat- treatment in H 2 /N 2 (1/99) are not shown.
• the calculated values are listed in Table 6. A positive correlative between A and D is found.
• a Ca , O ⁇ is plotted as a function of ⁇ re ⁇ , ⁇ so (Fig. 12). Both A Ca , O ⁇ and ⁇ n re ⁇ , ⁇ so were obtained from the iso-thermal heating procedure, but it should be noticed that y4 Ca ,ox was determined by using a bulk sample, whereas was determined by using a powdered sample. If the mass changes were obtained by using bulk samples, the mass increase would have been below the detection limit of the apparatus due to the small surface area.
• Figure 12 shows that the degree of Ca 2+ diffusion increases approximately linearly with increasing degree of oxidation of the polyvalent element.
• the untreated glasses do not have significantly different chemical durability. Some of the glasses that have been oxidized in air display a lower resistance towards the basic solution than the corresponding untreated glasses. This is explained by the low surface concentration of silicon because few Si-O bonds need to be broken in order to dissolve a relatively large amount of Ca 2+ ions.
• the increase in basic resistance of the glasses reduced in U 2 /N 2 (1/99) is caused by the high network connectivity of the treated glasses due to the inward diffusion of Ca 2+ . It seems that for both samples heated in air and U 2 /N 2 (1/99), the chemical durability depends on the thickness of the modified surface layer.
• Oxidation and reduction of seven polyvalent elements in silicate glasses result in diffusion processes near the surface.
• the oxidation process leads to formation of a crystalline oxide surface layer, whereas the reduction process leads to formation of a silica-rich layer.
• the crystalline surface layer lowers the onset temperature of the primary crystallization process, whereas the silica-rich surface layer enhances the chemical resistance of the glass in a hot basic solution.
• the diffusion mechanisms of modifying ions appear to be universal for all polyvalent element-containing glasses at temperatures around T 9 .
• the polyvalent element must be present in the glass as a mixture of oxidized and reduced ions.
• Cu is the optimal ingredient for formation of the thickest surface layers under the same redox treatment condition.
• both a kinetic fragility index m i.e., steepness of the log viscosity vs. T 9 ZT curve at T 9
• a thermodynamic index C p ⁇ /C pg i.e., the ratio of the liquid to the glassy isobaric heat capacity at T 9
• the ratio of the activation energy of the inward diffusion E d near T 9 to the activation energy for viscous flow E n at T 9 increases with increasing fragility of the liquid.
• the inward cationic diffusion can be enhanced by lowering the fragility of glass systems via varying the chemical composition.
• Sample preparation Four iron-bearing alkali-alkaline-earth silicate glasses (see Table 8) were prepared from analytical reagent-grade SiO 2 , Na 2 COs, MgO, CaC ⁇ 3, SrC ⁇ 3, BaC ⁇ 3, and Fe 2 ⁇ 3 powders.
• the mixed batch materials were melted in an electrical furnace (SF6/17, Entech) at 1500 0 C in a Pt 90 Rhi 0 crucible for 3 h. The melt was then cast onto a brass plate and pressed to obtain cylindrical glasses of 7-10 cm diameter and ⁇ 5 mm height.
• the prepared glasses were annealed 10 K above their respective glass transition temperatures for 10 min and then cooled down to room temperature within 20 h.
• the glass transition temperature (T 9 ) was measured using a differential scanning calorimetry (DSC) instrument (STA 449C Jupiter, Netzsch).
• DSC differential scanning calorimetry
• the C p curve for each measurement was calculated relative to the C p curve of a sapphire reference material after subtraction of a correction run with empty crucibles. Measurements were carried out in a purged Ar atmosphere. The following heating procedure was carried out to determine T 9 . First, the sample was heated at 10 K/min to a temperature 1.11 times the respective T 9 (in K) of each sample. Subsequently, the sample was cooled to room temperature at 10 K/min. Then, T 9 was determined by a second upscan at 10 K/min in order to ensure a uniform thermal history of the four glasses.
• the ratio C p ⁇ /C pg was also determined from this scan.
• the viscosity was measured by beam-bending (T > Tg) and concentric cylinder (T > Ti ⁇ qu ⁇ dus ) experiments.
• T > Tg beam-bending
• T > Ti ⁇ qu ⁇ dus concentric cylinder
• a 300 g weight was used to explore the viscosity range from approximately 10 12 to 10 10 Pa-s at a constant heating rate of 10 K/min.
• the viscosity was calculated according to DIN ISO 7884-4.
• the low viscosities ( ⁇ 10 2 Pa-s) were measured using a concentric cylinder viscometer.
• the viscometer consisted of four parts: furnace, viscometer head, spindle, and sample crucible.
• the viscometer head (Physica Rheolab MCl, Paar Physica) was mounted on top of a high temperature furnace (HT 7, Scandiaovnen A/S). Spindle and crucible were made of Pt 8 ORh 2 O-
• the viscometer was calibrated using the National Bureau of Standards (NBS) 710A standard glass.
• the polished glasses were heat-treated at 1 atm in an electrical furnace under a flow of HVN 2 (1/99 v/v) gas.
• HVN 2 (1/99 v/v) gas.
• the presence of oxygen in the furnace is not completely avoidable.
• Fe 2 Os and Fe3 ⁇ 4 powders were mixed in the molar ratio 3:2 and placed inside the furnace together with the samples.
• the glass samples and redox buffer were inserted into the cold furnace and the gas-flow was turned on.
• the furnace was then heated at 10 K/min to the pre-determined heat-treatment temperature T a and kept at this temperature for the duration t a .
• the diffusion process was ended by cooling the furnace to room temperature at 10 K/min.
• the glasses were treated at 0.95, 1.00, 1.025, and 1.05 times their respective T 9 (in K) for 2 h and at their T 9 for 16 h.
• the Mg-containing glass was treated at its T 9 for 0.5 and 8 h.
• the diffusion profiles were determined by electron-gas secondary neutral mass spectroscopy (SNMS). SNMS is used to determine the elemental concentrations as a function of the depth within the glass. The measurements were performed by using an INA3 (Leybold AG) instrument equipped with a Balzers QMH511 quadrupole mass spectrometer and a Photonics SEM XP1600/14 amplifier.
• the analyzed area had a diameter of 5 mm and was sputtered using Kr plasma with an energy of ⁇ 500 eV.
• the time dependence of the sputter profiles was converted into depth dependence by measuring the depth of the crater at 12 different locations on the same sample with a Tencor Pl profilometer.
• the relative spectral areas of Fe 3+ (doublet and sextet) and Fe 2+ (doublet) are used to calculate the Fe 3+ /Fe to t ratio for each of the untreated glasses (Table 8).
• the isobaric heat capacity (C p ) curves recorded during DSC upscans for the four glass compositions are not shown.
• T 9 is determined at the cross point between the extrapolated straight line of the glass C p curve before the transition zone and the tangent at the inflection point of the sharp rise curve of C p in the transition zone.
• the Tg values within an accuracy of ⁇ 3 K are given in Table 9.
• T 9 is plotted as a function of the ionic radius of the alkaline earth ion (r) (not shown).
• the radii of the alkaline earth ions are listed in Table 8 for a coordination number of 6. T 9 is found to decrease with increasing r.
• the equation may be simplified as the following expression: In this equation, there are only 3 parameters. The data are fitted with this modified AM equation by using the Levenberg-Marquardt algorithm. It is found that this equation fits the data better than both the Vogel-Fulcher-Tammann (VFT) equation and the Adam-Gibbs (AG) equation.
• VFT Vogel-Fulcher-Tammann
• AG Adam-Gibbs
• the F values for the four compositions are given in Table 8. Fragility can also be described by the index m, which is defined as the slope of the log ⁇ versus T 9 /T curve at T 9 :
• C P9 a linear function is fitted to the C p values at temperatures below T 9 .
• the value of this function at T 9 is reported as C P9 .
• the results in Table 9 show that C p ⁇ /C P9 increases with increasing r.
• the step change in the heat capacity (C pr C P9 ) at the glass transition is also calculated and listed in Table 9. Similar tendencies for C p ⁇ /C P9 and C pr C P9 are observed.
• the kinetic fragility (quantified by F or m) shows a positive correlation with the thermodynamic fragility (quantified by C p ⁇ /C pg or C p rC pg ).
• thermodynamic fragility quantified by C p ⁇ /C pg or C p rC pg .
• the fragility is found to increase with increasing r of the alkaline earth ions in the glass series studied in this work. This may be explained as follows. For a more fragile liquid, there is a larger change in the structure of the liquid with temperature than for a less fragile liquid.
• the high field strength of Mg 2+ causes a high degree of short range order, which prevents the structure from a rapid break-down with increasing temperature.
• the Ca 2+ ions should then be the fastest and have the lowest E 0 as the radius of Na + (1.02 A) is very similar to that Of Ca 2+ (1.00 A).
• the iron reduction causes the alkali ions to diffuse (role of Na + in diffusion process is discussed later). These factors limit the ability of Na + to jump into the empty alkaline earth ion sites, which might explain why Ca 2+ is not found to be the fastest ion in our glasses.
• the network modifying oxides form interconnected channels (i.e., a percolative network) at sufficiently high concentration.
• the threshold for percolation occurs at 16 vol% of modifying oxides, which is exceeded by the glass compositions studied in this work.
• the alkaline earth ions should diffuse fastest through the channels when their size is smallest which explains our diffusion results at isokom temperatures.
• the activation energy of diffusion is the sum of an electrostatic term due to the Coulomb interaction between the cation and the NBO plus an elastic part to open up doorways into neighboring sites. In our glasses, the latter term governs the activation energy as the smallest alkaline earth ion has the lowest E ⁇ since it most easily moves through the channels.
• the channels are constituted by [SiO 4 ] tetrahedra, i.e., the required displacement of oxygen is relatively small for a small alkaline earth ion.
• E d is plotted against E n in Fig. 15 and a clear linear correlation is observed, but the EJE n ratio is smaller than 1. According to the Stoke- ⁇ instein equation, the activation energy of diffusion increases with increasing viscosity.
• the equation cannot be used to predict ion mobilities, because the ions use the transportation route with the lowest activation energy, i.e., they flow faster than the cooperative rearrangements of the structural units. In other words, the diffusion of alkali and alkaline earth ions is decoupled from the network change.
• the inward diffusion process is driven by reduction of Fe 3+ to Fe 2+ , but Fe 2+ is capable of diffusing itself.
• the ionic radius of Fe 2+ in 6-fold coordination is 0.78 A for the high spin state.
• the diffusion coefficients of the network-modifying divalent cations are calculated and they are different for the glasses treated in the CO and H 2 gases.
• the H 2 -bearing gas creates the silica-rich layer more effectively than the CO-bearing gas.
• the layer increases the hardness and chemical durability of the glass due to the silica network structure in the surface layer.
• SiO 2 was introduced as quartz, Fe 2 O 3 as Fe 2 O 3 , and MgO as Mg(OH) 2 -(MgCOs) 4 -(H 2 O) 5 .
• Conventional transmission 57 Fe M ⁇ ssbauer spectroscopy measurements on powdered samples were used to determine the iron redox state of the untreated iron-containing glasses.
• the T 9 values of 6wtFe and 3wtFe were measured using differential scanning calorimetry (DSC), and found to be 926 K and 921 K, respectively.
• the obtained glasses were cut in cylinders and then ground by a six-step procedure with SiC paper under ethanol, followed by polishing with 1 ⁇ m diamond suspension.
• Heat-treatments in the H2/N2 (1/99) atmosphere were conducted at 1 atm in an electrical furnace.
• the glass samples were inserted into the cold furnace and the gas-flow was turned on. Heating and cooling of the furnace were conducted at 10 K/min. Treatments in CO/CO 2 (98/2) were conducted similarly, but the heating and cooling rate was 5 K/min.
• the partial pressure of oxygen was kept at a known value in the H2/N2 (1/99) atmosphere by using a Fe3 ⁇ 4 /Fe2 ⁇ 3 redox buffer. Fe 2 Os and Fe3 ⁇ 4 powders were mixed in the molar ratio 3:2 and placed inside the furnace together with the samples. In the CO/CO2 (98/2) atmosphere, the oxygen partial pressure was controlled by the CO-CO2-O2 equilibrium.
• FT-IR Fourier transform infrared
• UV-VIS- NIR ultraviolet-visible-near infrared
• the Fe 2+ ion has a maximum absorption peak near 1050 nm but the position and intensity of this peak varies with glass composition.
• A absorbance
• c concentration
• ⁇ absorption coefficient
• t the sample thickness
• compositional analysis of the glass surfaces was carried out using electron-gas secondary neutral mass spectroscopy (SNMS) with an INA 3 (Leybold AG) instrument.
• the analyzed area had a diameter of 5 mm and was sputtered using Kr plasma with an energy of ⁇ 500 eV.
• the time dependence of the sputter profiles was converted into depth dependence by measuring the depth of the crater at 10 different locations on the same sample with a Tencor Pl profilometer. Two properties of the heat-treated glasses were tested. Vickers hardness was measured 25 times for each sample using a Struers Duramin 5 microindentor at a load of 0.25 N and a hold time at the maximum load of 5 seconds.
• the lengths of the indentation diagonals were measured using an optical microscope (reflection method). Chemical durability was tested by measuring leached amounts of Na + and Mg 2+ ions after dissolution in 0.25 M HCI and KOH solutions. The samples were immersed in plastic containers with 20 cm 3 test solution for each 1 cm 2 of the glass surface area. The containers were mounted on a thermostatic shaking assembly at 90 0 C (agitated at 100 ppm) and after 12 h, the samples were removed from the solutions. Atomic absorption spectroscopy (AAnalyst 100, Perkin Elmer) was employed to measure the concentrations of Na + and Mg 2+ in the test solutions.
• Figure 16 shows UV-VIS-NIR spectra of glasses heat- treated at Tg for 16 h in U 2 /N 2 (1/99) or CO/CO 2 (98/2), respectively. A maximum absorption peak is seen near 1050 nm, which is attributed to the existence of the Fe 2+ ions.
• the intensity of the Fe 2+ band increases, indicating that Fe 3+ is reduced to Fe 2+ .
• NaFe 3+ O 2 represents a Fe 3+ , which is tetrahedrally coordinated with oxygen and charge-balanced by Na + .
• SiOSi corresponds to a bridging oxygen connecting two silica tetrahedra.
• SiOH is a silica tetrahedron containing a hydroxyl group.
• SiCO 3 Ca 0 .5 is a carbonate species connected to a NBO and Ca 2+ .
• SiO(Fe 2+ )o.5, SiONa, and SiOCa 0 . 5 represent that Fe 2+ (octahedral coordination), Na + , and Ca 2+ are connected to a NBO, respectively.
• the SNMS depth profile of the 6wtFe glass heat-treated in CO/CO 2 (98/2) at its T 9 for 16 h is not shown.
• a pronounced decrease of the concentration of Mg 2+ , Ca 2+ , and Fe 2+ towards the surface is observed (thickness: 300 ⁇ 350 nm).
• Na + also diffuses towards the interior.
• alkali ions are normally found to be faster than earth alkaline ions in glasses due to their lower charge, the diffusion depth of Na + is smaller than that of Mg 2+ , Ca 2+ , and Fe 2+ , which is in agreement with the above studies.
• the inward diffusion occurs to charge-balance the outward flux of electron holes, and the charge might be most effectively transferred by the divalent cations.
• the SNMS profile of the glass treated in CO/CO 2 indicates that the mechanism of Fe 3+ reduction in CO/CO 2 (98/2) is the same as that in U 2 ZN 2 (1/99).
• the internal reduction of Fe 3+ generates electron holes (h * ).
• An outward flux of h' occurs, which is driven by the gradient in oxygen activity across the reaction zone, h' are filled by electrons released by ionic oxygen at the surface since oxygen is released into the reducing atmosphere as CO 2 .
• the outward flux of h' is accompanied by inward flux of network-modifying cations to maintain the charge-balance.
• the inward cationic diffusion is driven by reduction of the high valence to the low valence state of the polyvalent cation.
• diffusion coefficients for the divalent cations should be calculated and compared to known values of diffusion coefficients for divalent cations in similarly polymerized glasses.
• the diffusion coefficient for a divalent cation ( D M2+ ) can be calculated by using the following equation:
• ⁇ is the thickness of the modifier layer
• J ⁇ 2+ is the cation mole fraction of the divalent cation M 2+
• ⁇ t is the reaction time
• p' o is the partial pressure (i.e., activity) of oxygen at the free surface
• p" 0 is partial pressure of oxygen at the internal reaction front.
• H 2 and CO molecules must first penetrate into the uppermost surface layer, subsequently be dissolved in the structure, and simultaneously contact and reduce the ferric ions in the glass structure. The penetration, and hence, reduction process is easier when the molecule is small.
• the hardness and chemical resistance of the untreated and heat-treated samples are displayed in Table 10. From the structural point of view, the earth alkaline and alkali cations disrupt the continuous Si-O random network, and so introduce NBOs to the glasses. Their removal from the surface clearly increases the hardness and chemical resistance of the glasses. The increase is most pronounced as a result of the H 2 -treatment as treatment in this atmosphere creates the thickest silica-rich layer.
• Table 10 Effect of the atmosphere of the heat-treatment on the Vickers hardness (AVv) and chemical durability of the 6wtFe glasses.
• Chemical durability of the glasses is expressed by the leached amount of Na + after 12 h in 0.25 M HCI solution (C(Na + )ac,d) and Mg 2+ after 12 h in 0.25 M KOH solution (C(Mg 2+ ) a
• a silica-rich surface layer can be created by heat-treating an iron- bearing glass at its T 9 in both CO- and H 2 -containing atmospheres.
• the layer is created due to the inward diffusion of network-modifying cations.
• the glass surface becomes structurally more polymerized due to the removal of network-modifying cations from the surface. Consequently, the hardness and chemical durability of the glasses are enhanced.
• the extent of the inward diffusion is larger as a result of the H 2 - treatment than of the CO-treatment. This is attributed to the fact that H 2 has a smaller size than CO, so that the former more readily reduces the ferric ions in the surface structure than the latter.
• H 2 will be more or less soluble. Hence, different cone, of H 2 will result in inward diffusion (probably: when H 2 is more soluble, a lower H 2 cone, is needed to obtain inward diffusion).

Landscapes

• Chemical & Material Sciences (AREA)
• Physics & Mathematics (AREA)
• Thermal Sciences (AREA)
• Life Sciences & Earth Sciences (AREA)
• Engineering & Computer Science (AREA)
• Chemical Kinetics & Catalysis (AREA)
• General Chemical & Material Sciences (AREA)
• Geochemistry & Mineralogy (AREA)
• Materials Engineering (AREA)
• Organic Chemistry (AREA)
• Glass Compositions (AREA)
• Surface Treatment Of Glass (AREA)

Priority Applications (4)

Applications Claiming Priority (2)

Publications (1)

Family

ID=40510556

Family Applications (1)

Country Status (5)

Cited By (4)

Families Citing this family (13)

Citations (1)

Family Cites Families (3)

• 2009
  • 2009-09-03 JP JP2011525407A patent/JP2012501940A/ja active Pending
  • 2009-09-03 CN CN2009801344545A patent/CN102143923A/zh active Pending
  • 2009-09-03 EP EP09811067A patent/EP2346793A1/en not_active Withdrawn
  • 2009-09-03 US US13/061,398 patent/US20110159219A1/en not_active Abandoned
  • 2009-09-03 WO PCT/DK2009/050224 patent/WO2010025735A1/en not_active Ceased

Patent Citations (1)

Non-Patent Citations (2)

Also Published As

Similar Documents

Legal Events