METHOD FOR HEATING GLASS
Background of the Invention This invention pertains to processes during which glass workpieces are stacked or placed adjacent one another and heated. This invention also pertains to methods for flattening, annealing and/or densifying glass. There are certain manufacturing processes during which glass is heated. These processes include annealing, densifying and flattening. When a glass sheet is initially manufactured, regions within the glass can be under tensile or compressive stress due to manufacturing conditions such as thermal gradients that exist during cooling. To eliminate these stress regions, it is known to anneal glass by heating the glass and then slowly cooling it. If one heats glass that is formed by certain manufacturing techniques (e.g. the float method), the density of that glass increases. Glass used to form certain types of devices (e.g. LCD or TFT displays) is often heated during the device manufacturing process. It is undesirable to have that glass change size (e.g. shrink) during the manufacturing process. (Thermal shrinkage of glass is discussed by K. Mori et al., "Study on the Thermal Shrinkage of Annealed Coming Code 1737 Glass for Low- Temperature Poly-Si Application", presented at SID 1999 in San Jose CA and Anma et al, "Substrate for Poly-Si TFT Applications", presented at IDW 1996 in Kobe,
Japan, each incorporated herein by reference.) Thus, it is known to heat the glass workpiece prior to forming the LCD or TFT displays in order to "densify" the glass. During this process, the glass becomes more dense. Densified glass does not shrink as much upon subsequent heating as non-densified glass. Thus, it is often desirable to use densified glass when making LCD and TFT displays. Another manufacturing process during which glass is heated is "flattening." Glass formed by various methods, including down drawing, sometimes lacks the flatness needed for certain applications. Accordingly, it is known to subject such glass to a flattening process. Fig. 1 illustrates one type of prior art flattening process during which a glass sheet 1 is sandwiched between two optically flat plates 2, 3. Sheet 1 is heated above its softening temperature. The combination of high temperature and pressure from plates 2 and 3 flattens glass sheet 1. It would be desirable to enhance the efficiency of this process by flattening several sheets at once. One might be tempted to try to flatten sheets 4, 5, 6, 7 and 8 between plates 2 and 3 as shown in Fig. 2. (One might also be tempted to stack glass sheets one on top of another, during annealing or densifying.) Unfortunately, if sheets 4 to 8 are heated above their softening temperature and pressed together, they tend to bond or fuse to one another — an undesirable result. It is known to try to separate such sheets from one another, and thereby prevent such bonding during flattening, by placing separator material, e.g. paper, between the sheets. Unfortunately, due to the fact that this process occurs at a high temperature, the paper tends to carbonize. In addition, surviving carbonized fibers from the paper can introduce non-uniform irregularities in the glass sheet surface.
U.S. Patent 5,916,656, issued to Kitayama et al, discusses a method whereby an aqueous NaOH, KOH, Ba(OH)2, Ca(OH) , or Na2SO4 solution is applied to glass workpieces. (Kitayama lists a few other water-soluble materials, e.g. at col. 8, lines
48-55.) Kitayama then removes his workpieces from the aqueous solution, dries the workpieces, stacks them, and flattens them. A residual film left over from the above-
mentioned solution is described in Kitayama as preventing the glass workpieces from
bonding. Kitayama' s draining and drying steps are critical. The drying process may
result in a film that is not uniform in thickness, e.g. because of capillary effects that
are difficult to eliminate.
Summary
A method in accordance with the present invention comprises providing a set of workpieces adjacent to or stacked on one another and heating the workpieces. In
one embodiment, this is done during the course of flattening the workpieces, and the workpieces are sandwiched between a set of plates and heated. In another embodiment, the workpieces are annealed or densified, (typically without the
presence of the above-mentioned plates). The workpieces are typically glass.
Prior to heating, a thin layer of inorganic material, e.g. a layer of particles, is
applied to one or both surfaces of the workpieces. In one embodiment, the thin layer
of material is a "monolayer", i.e. a single layer of particles, on the workpiece surface. Preferably, the particles are uniform in size, and are sub-micron in size. In one
embodiment, the particles have a size greater than or equal to 10 nm in diameter, and
preferably greater than or equal to 0.1 μm. Also in one embodiment, the particles
have a size less than or equal to 10 μm in diameter, and preferably less than 5μm in
diameter. In one embodiment, the particles are 0.5μm in diameter. The monolayer
can be a metal oxide, e.g. SiO2, AI2O3, TiO2, SnO , CeO2, ZnO2, Sb2O5 and Y2O3. Because the particles are in the form of a monolayer, the thickness of the particles across the workpiece is uniform. This can be accomplished independently of any draining or drying steps. In one embodiment, the particles are provided in an aqueous dispersion. The particles have an electrostatic charge that is opposite to the charge residing on the workpiece. After heating, the particles can be removed. In accordance with another embodiment of the invention, instead of applying the above-mentioned monolayer, the workpieces are immersed in a solution such as an HC1 solution. It as been discovered that immersing the workpieces in such a solution can prevent the glass workpieces from bonding together. Of importance, glass typically contains a number of oxide materials such as Na2O, K2O, and BaO, Li2O or other materials. Unreacted oxide materials within the glass can sometimes reside at or near the surface of the glass workpiece. It is believed that these materials are responsible for causing the glass workpieces to bond together when the glass is heated above its softening temperature. It is further believed that these unreacted oxides are dissolved by the HC1 solution and removed. Therefore, glass workpieces treated in this manner do not bond when heated above their softening temperature. In lieu of HC1, other materials can be used to dissolve these unreacted oxides, e.g. acids such as H2SO4, HNO3, CH3COOH or mineral acids. In one embodiment, the workpieces are rinsed after immersion in acid.
In accordance with another embodiment of the invention, one or more workpieces are heated, e.g. during annealing, flattening or densification. The workpieces are typically a silica-based material. Prior to heating, the one or more workpieces are coated with the above-mentioned layer of particles. The particles prevent or reduce sticking between the one or more workpieces and a structure that the one or more workpieces touch (e.g. other workpieces or a platform on which the workpiece rests) during heating. In accordance with another embodiment of the invention, a platform or other structure on which one or more workpieces rest or touch during heating (e.g. a flattening plate) is covered with the above-mentioned layer of particles. In one embodiment, this structure is pure silica or a non-pure silica-based material. For example, this structure can be quartz. The particles prevent or reduce sticking between the one or more workpieces and the above-mentioned structure. In accordance with another embodiment of the invention, one or more workpieces are immersed in the above-mentioned solution and then heated. The immersion in the solution prevents or reduces sticking between the one or more workpieces and a structure that the one or more workpieces touch during heating. During heating, two or more of annealing, densifying and flattening can be accomplished. (For example, both annealing and densifying can be accomplished, or both annealing and flattening can be accomplished, etc.)
Brief Description of the Drawings Fig. 1 illustrates a glass workpiece being flattened using a method in accordance with the prior art.
Fig. 2 illustrates a set of glass workpieces being flattened by being sandwiched between a pair of optically flat plates. Fig. 3 schematically illustrates a glass workpiece covered by a monolayer of metal oxide particles. Fig. 4 illustrates a set of glass workpieces covered by a monolayer of metal oxide particles being subjected to a flattening process in accordance with the present invention. Fig. 5 illustrates the relation between surface charge on a glass workpiece, AI2O3 particles, and CeO2 particles versus the pH of a solution in which the workpiece or particles are immersed. Fig. 6 illustrates the temperature applied to glass workpieces vs. time during a flattening, annealing and/or densifying process in accordance with one embodiment of the present invention.
Detailed Description A method in accordance with the invention begins by providing a workpiece 10 (Fig. 3). In one embodiment, the workpiece is a glass rectangle (optionally a square) between 0.5 and 2.8 mm thick, but this shape and thickness are merely exemplary. The glass rectangle is typically cut from a glass sheet. The glass sheet can be formed by drawing, the float method or pressing, although the workpiece can be formed by other methods as well. A layer 11 of material is then formed on the glass workpiece 10. Typically, layer 11 is a monolayer of particles formed on the surface of workpiece 10, i.e. layer 11 is one particle thick. In one embodiment, these particles are sub-micron in size.
The particles are typically inorganic, and can be a metal oxide such as silica, ceria or alumina. The layer can be formed on workpiece 10 by immersing workpiece 10 in an aqueous dispersion of such solid particles. The immersion time can be between 1 second and 1 minute, and the dispersion can be at room temperature, but again, these parameters are merely exemplary. It is believed that the particles uniformly adhere to workpiece 10 due to electrostatic attraction. While not wishing to be bound by theory, it is believed that the reason is as follows. The surface of a glass workpiece, when immersed in water, generally has a negative surface charge assuming the water has a neutral pH. Ceria and alumina, when dispersed in water, have a positive electrical surface charge. Thus, ceria or alumina particles adhere to the glass workpiece due to electrostatic attraction. Further, ceria or alumina particles tend to adhere to the glass workpiece in the form of a monolayer. This is because after a first, positively charged monolayer is formed on workpiece 10, there is nothing to further attract additional positively charged ceria or alumina particles to the monolayer. Thus, to the extent that there are any extra ceria or alumina particles on workpiece 10, these particles only adhere weakly to the glass. After dipping the workpiece in a dispersion of ceria or alumina, any particles more than a monolayer thick are removed from the workpiece by rinsing using DI water. (If course, due to the nature of a single layer, there may be some spots on the workpiece where particles are missing, and other spots where the layer of particles on top the workpiece is a couple of particles thick. However, in general, the average thickness of the monolayer is one particle thick.) Workpiece 10 is then withdrawn from the aqueous dispersion of such particles and dried. Thereafter, workpiece 10 is subjected to a heating process. In one
embodiment, the heating process can be a flattening process as shown in Fig. 4. Referring to Fig. 4, workpieces 10a, 10b, 10c, lOd and lOe are pressed between plates 2 and 3, and heated above the softening temperature of the workpieces. Of importance, monolayers 11a, 1 lb, 1 lc, 1 Id and lie surround workpieces 10a, 10b, 10c, lOd and lOe, respectively. These monolayers prevent workpieces 10 from bonding to one another. After the above-mentioned heating process, workpieces 10 are cooled and separated from one another. As mentioned above, ceria and alumina are positively charged when placed in an aqueous pH neutral dispersion, whereas glass is negatively charged when placed in such an aqueous dispersion. Fig. 5 illustrates the charge on the surfaces of glass, alumina and ceria (curves 30, 32 and 34) as a function of the pH of the aqueous solution. Preferably the alumina or ceria particles have a charge that is the opposite of the glass charge so that these particles adhere to the glass. Accordingly, as can be seen from Fig. 5, this means that the aqueous dispersion pH should be between 2 and 7.5, and preferably between 4 and 6. Normally, silica particles in an aqueous dispersion are negatively charged. However, one can place silica particles from an aqueous solution onto a glass workpiece by applying to the glass workpiece an agent such as positively charged electrolyte, e.g. polyethyleneimine (PEI), to render the glass surface positively charged. In one embodiment, a 1 % PEI aqueous solution, e.g. as available from Sigma- Aldrich Co. of Milwaukee, Wisconsin, can be applied to glass workpiece 10 prior to immersing workpiece 10 in an aqueous silica dispersion. PEI is a polyelectrolyte. A polyelectrolyte is generally a high polymer substance, either natural (e.g. a protein or gum arabic) or synthetic (PEI or a polyacrylic acid salt)
containing ionic constituents which may be either cationic or anionic. Typical cationic polyelectrolytes, besides PEI, include Purifloc (available from Dow Chemical), Cat-Floe (Calgon Corp.) or Cato (available from Starch and Chemical Corp.). In one embodiment, the workpiece is dipped at a room temperature 1% PEI solution and rinsed well, thereby leaving a monolayer of PEI on the workpiece. The workpiece is then dried, and then dipped in the silica dispersion. The effectiveness of a method in accordance with the present invention was demonstrated by performing an experiment in which prior to flattening, five sets of glass workpieces were provided as follows:
Sample 1 Workpieces in sample 1 were untreated and uncoated glass.
Sample 2 Workpieces in sample 2 were coated with sub-micron size silica particles by dip-coating the workpieces in an aqueous silica dispersion. The silica dispersion was a 5% wt./vol. dispersion, i.e. product no. MP-4540, manufactured by Nissan Chemical
Co. of Japan. The average particle size of the dispersion was 0.45μm. The
workpieces in sample 2 were then dried. (If one dips the workpiece in a silica dispersion and then dries the workpiece, one does not obtain a monolayer of silica particles. Rather, one will obtain some regions on the workpiece covered by a layer varying numbers of silica particles thick, e.g. comprising regions covered by clumps of silica particles. This will ultimately negatively impact waviness of the workpiece
after the flattening process. Although a monolayer of particles may be preferable in some applications, use of silica particles without a polyelectrolyte, and non- monolayer silica particles, come within an embodiment of the present invention.)
Sample 3 Workpieces in sample 3 were coated with a 5% wt./vol. dispersion of alumina particles (0.5 micron average size). The dispersion was product number WA20000, manufactured by Fujimi Chemical Co. of Japan. Coating was performed by dipping the workpieces in an aqueous dispersion of alumina particles followed by a thorough rinse with DI water. Prior to coating, the acidity of the dispersion was adjusted to a pH of 6 in order to maximize surface adhesion of the alumina to the glass.
Sample 4 The workpieces of sample 4 were rendered positively charged by immersion in a 1% solution of PEI. Following immersion, the glass was rinsed with water and dried. After drying, a monolayer of silica particles was deposited on the surface by immersing the workpieces in an aqueous dispersion (5 % wt./vol.) of silica particles. In particular, product no. MP-4540, manufactured by Nissan Chemical Co. of Japan was the silica dispersion used. An excess of silica particles was rinsed off, leaving a
strongly adhering monolayer of silica particles. The particle size was about 0.45 μm in diameter.
Sample 5 The workpieces were soaked in a diluted solution of hydrochloric acid (3%) for about 3 hours. The solution was at an elevated temperature (i.e. 75 ° C). (In lieu of
a temperature of 75 ° C, other temperatures, e.g. a higher temperature, can be used.) The workpieces were then dried. After treating the various workpieces, they were pressed between optically flat plates and heated as described above. Fig. 6 illustrates the temperature applied to the workpieces vs. time during this process. A stack of five glass squares with a topical load of 20 kg was treated at a time. These experiments were repeated using optically flat flattening plates (i.e. plates 2 and 3) made of stainless steel, aluminum, and quartz. The results are set forth in Table I. As can be seen in Table I, there was no bonding or welding between the glass workpieces for those workpieces coated with silica or alumina particles. Also, bonding or welding between glass workpieces did not occur for those workpieces soaked in an HC1 solution. As mentioned above, it is believed that the HC1 solution treatment aids in removal of acid-soluble alkaline oxides that might be responsible for bonding. For example, the HC1 reacts with additives such as Na2O as follows: Na2O + 2HC1 - 2NaCl + H2O HC1 reacts with the other oxide components within glass, e.g. K2O, Li2O, and/or BaO. As seen in Table I, those glass workpieces that were left untreated and uncoated with metal oxide particles exhibited the bonding or welding problem solved by the present invention.
Table I lists average values for "flatness" and "waviness" for the various samples. As known in the art, flatness is a measure of the maximum peak to valley distance when looking along a line on the surface of the workpiece about 1 cm long. Waviness is the measure of the maximum peak to valley distance when looking along a line on the surface of a workpiece about 2 or 3 mm long. As can be seen, flatness was improved using a process in accordance with the present invention. In some instances, e.g. using silica particles without PEI, waviness was degraded during flattening. In some applications such degradation in waviness is undesirable. In other applications, such degradation of waviness is not a major problem. Accordingly, in those applications, silica particles without PEI can be applied to the workpieces prior to flattening. In one embodiment the glass workpieces are cut into different shapes or dimensions. Flattening can occur after the workpieces are cut, but preferably, flattening occurs before such cutting. (This is because prior to flattening there can be internal stresses within the workpieces. If the workpieces are heated above the softening temperature after being cut, these stresses could distort the shape of the workpieces — an undesirable result.) Another experiment was performed to assess the effectiveness of ceria particles for preventing the bonding phenomenon. During this experiment, two sets of samples were subjected to the flattening process. Sample 6 comprised workpieces without application of a dispersion of particles . Sample 7 of Table II comprised flat glass workpieces treated with a dispersion of micron-size ceria particles (a dispersion marketed under the name "Big C", manufactured by Universal Photonics, located in Hicksville, NY). The ceria was
applied by rubbing with a soft sponge, followed by rinsing with DI water. The
workpieces in the second sample were then allowed to dry in a room temperature
environment. The same treatment was given to the quartz flattening plate.
After applying ceria to the workpieces in Sample 7, they were subjected to a
flattening process using quartz flattening plates. (The quartz plates were also coated with ceria particles using the above-described method.) The flatness and waviness data for Samples 6 and 7 are set forth in Table II. After flattening, the workpieces in Sample 7 were polished using a ceria
polishing slurry to ensure a desired surface finish (roughness). (During this polishing
step, lOμm of material was removed from each side of the sample.) Table II also lists
the flatness and waviness for these workpieces (see the third line of data in Table II).
As seen in Table II, waviness and flatness numbers for the workpieces in
Sample 6 (lacking the ceria monolayer) were rougher than for sample 7 (workpieces coated with ceria prior to polishing). For some disks, waviness and flatness values of "NR" are provided. NR means that the waviness and flatness values were not
readable, e.g. because their values were too great. The waviness and flatness values
for ceria coated workpieces were invariably better than for workpieces not coated with ceria. Ceria has several advantages when used as a monolayer material. First, often,
after flattening, the workpieces are subjected to polishing using ceria polishing
particles. Accordingly, the residual presence of ceria on the workpieces after
flattening does not represent a problem.
Second, plates 2, 3 often move relative to the workpieces, e.g. because of
thermal expansion and contraction of the plates relative to the workpieces. (This can
be either because the plates have a different coefficient of thermal expansion from the workpieces, or because of temperature differences between the plates and the
workpieces.) Ceria particles can serve as a "lubricant" to permit the workpieces or plates to expand and/or move relative to one another without applying a shear force to
the workpieces that would otherwise distort the workpieces. (Ceria is believed to be
better than silica because no polyelectrolyte is needed. Ceria is believed to be better to alumina because it is less likely to leave abrasive scratches on the workpiece.)
As mentioned above, in the above-described experiment, micron size ceria particles were used from Universal Photonics. In other embodiments, glass workpieces are coated with other ceria slurries. The concentration of ceria in the
dispersion can be 25% wt./vol. The ceria dispersion can be product no. Mirek L-50, manufactured by Yochiyo Microscience Inc. The dispersion can be applied using a soft sponge. The glass workpieces are typically rinsed and dried prior to heat
treatment.
To summarize, monolayers of preferably monosized inorganic particles
provided superior waviness. Preferably the particles have a consistent size, e.g. between plus or minus 20% of a nominal value, and preferably 10% of a nominal value. (This is useful for AFM Ra purposes.) Significant improvement in flatness and waviness of the glass workpieces can be realized by coating the glass surface with monolayers of inorganic particles such as silica, alumina and ceria. No thermal
bonding or fusing was observed when the particulate films composed of the above
oxides were applied to the workpieces. Silica dispersions of particles can be used in a simple mode by dipping a glass
workpiece in an aqueous dispersion of silica, draining and drying. This approach
results in a dried film consisting of many layers of dried particles, and can have non- uniform streaks and generally non-conformal "skin". To obtain a uniform "skin", one
can deposit a monolayer consisting of mono-dispersed silica particles which are conformal to the surface. Monodisperse silica particles, of any size, are readily
commercially available and are not too expensive. However, obtaining a monolayer
of silica particles usually involves charge reversal on the glass surface by treatment,
e.g. with PEI. Alumina, by contrast, does not require PEI to stick since the glass surface and
the surface of the alumina particles have opposite charges (the glass being negative and the alumina being positive) and forming a monolayer is readily and speedily
accomplished. Ceria particles are electrostatically comparable to alumina and they also stick to glass without the aid of PEI. The disadvantage of both alumina and ceria dispersions is their poorer monodisparity (i.e. non-uniform particle size) which results in high "spots" (due to over-sized particles) which degrades surface flatness and
waviness.
POST-HEATING CLEANING PROCESS
In one embodiment, after flattening (or after annealing, densifying or other
heating process), the particles deposited on the workpieces are removed by polishing. In another embodiment, they are removed with a fluoride solution. One example of such a solution comprises 5% ammonium bifluoride and 5% sulfuric acid. The
workpieces can be immersed in the solution (typically room temperature) for a time
period between 10 seconds and 5 minutes (e.g. 1 minute), followed by rinsing in DI water and drying. During immersion in the fluoride solution, the solution is typically
ultrasonically agitated at a frequency greater than or equal to about 28 KHz, e.g. 68
KHz. In another embodiment, the workpieces are cleaned in a NaOH solution. This solution can comprise between 5 to 50% NaOH (typically 10% NaOH).
Alternatively, KOH can be substituted for the NaOH. The solution temperature can
be between room temperature and 95°C (e.g. 85°C), and immersion in this solution can be between 1 and 15 minutes, followed by rinsing in DI water and drying. In one
embodiment, the solution is subjected to ultrasonic agitation, e.g. at a frequency
greater than or equal to about 28 KHz, e.g. 68 KHz.
As mentioned above, in one embodiment, in lieu of or in addition to the above-mentioned cleaning process, the workpieces are polished. Any adhering layer of particles is polished off the substrates very quickly. Typically, polishing removes
several tens of microns in thickness from each surface of the glass. Since any adhering particles typically range in thickness from about 12 nm (e.g. for a monolayer of particles such as alumina-coated silica) to a micron or so (e.g. for ceria particles), any residual particles are quickly removed.
For processes including immersion in PEI, there is typically no PEI left on the
workpiece after flattening. As an organic material, PEI is burnt completely
(producing water and CO2).
EXAMPLES OF OTHER EMBODIMENTS For an embodiment in which the workpieces are immersed in an acid such as HCl prior to flattening, the acid solution can be at different appropriate temperatures,
e.g. between 50 and 100 ° C. The acid concentration can be from 1% to fully
concentrated acid. (For HCl, fully concentrated is about 38%. For H2SO , fully concentrated can be as high as 98%, but H2SO4 at this concentration may be unpleasant or problematic to use, so a concentration of 1 to 10% may be more practical.) In one embodiment, acid immersion is followed by application of particles, e.g. a monolayer of particles. In one example of such a process, the workpieces are immersed in a hot 5% HCl solution for at least 3 hours, followed by rinsing and then dipping the workpieces in a dispersion of colloidal silica that has its surface modified by alumina. Silica which has its surface modified by alumina is a form of silica particles having an alumina outer layer. During manufacturing, the silica particles are exposed to an aqueous solution of alumina ions, such as Al , e.g. as can be found in an Al(NO3)3 solution. This allows alumina ions to be incorporated onto the silica surface. In fact, the particles are coated with a thin alumina layer, e.g. a monolayer of AI2O3 a few angstroms thick, which makes the silica with its surface modified by alumina positively charged. These particles behave like alumina particles and form a monolayer on the glass workpiece. Such a colloidal silica dispersion is available under the product name Ludox CL, and is manufactured by DuPont. These parties are monodisperse (i.e. of a relatively uniform size) with an average size of 12 nm. After dipping in the Ludox dispersion, the workpieces are rinsed, dried, and ready for the flattening heat treatment. (In one embodiment, the alumina-modified silica particles can also be used if the workpieces are not soaked in acid, e.g. if the above-mentioned oxides within the glass are of a suitably low concentration.)
In other embodiments, other materials (e.g. ceria) are applied to the silica particles so that the silica particles exhibit a surface charge that facilitates electrostatic attraction to glass or formation of a monolayer on glass.
EXAMPLES OF HEATING CYCLES FOR USE DURING FLATTENING Flattening can be performed at different appropriate temperatures. The glass is typically heated above the glass transition temperature, and then cooled slowly to about 400 ° C, followed by relatively fast cooling to room temperature. (Below
400 ° C, the shape and flatness of glass parts are not affected by the cooling rates.) Generally, during flattening, the temperature is raised quickly to a value between 500 and 600 ° C (preferably between 545 and 565 ° C), and held at that temperature between 4 and 10 hours (e.g. about 4 hours). The workpieces are then cooled very
slowly to 400°C. Ramping down to 400 ° C can take place over a time span of4 to 10 hours, e.g. 7 hours. For example, in one embodiment, ramping down to 400°C can be done at a rate of 25°C/hour. After ramping the workpiece temperature down to 400°C, it can then be brought rapidly to room temperature.
USE OF THE INVENTION DURING DENSIFYING AND/OR ANNEALING The present invention can be used in applications other than flattening glass, e.g. during the manufacture of LCD and TFT displays. For example, glass workpieces can be stacked during annealing or densifying. During such embodiments a stack of glass workpieces (typically between 1 and 25 workpieces) can be placed in a furnace. One prevents the glass from workpieces from sticking together using the methods described above. The workpieces are typically heated to a temperature at
which they would otherwise be caused to stick together (e.g. 565 °). For example, the workpieces are heated to their softening temperatures. Thereafter, the workpieces are cooled. (While the workpieces are typically stacked one on top of another, they can be placed side by side and against one another.) In one embodiment, during annealing or densifying, the glass is heated to a temperature between 400 and 600°C for a time period between 1 and 12 hours and then slowly cooled to the glass strain point. (This cooling can be at a rate between 5 and 30°C/hour.) After reaching the glass strain point, the cooling rate is typically increased. The specific times and temperatures can be selected as a function of glass
) composition and temperature-related characteristics of the glass (e.g. the annealing point and the strain point). The densifying temperature can also be a function of the temperature to which the glass will be exposed during subsequent manufacturing processes. In another embodiment, the temperatures and times applied to the glass during densifying and/or annealing can be as shown in Fig. 6.
EMBODIMENTS INCLUDING OTHER PARTICLES THAT CAN BE USED As mentioned above, in lieu of silica, ceria or alumina, other oxide particles can be used during a method in accordance with the present invention. For example, particles such as TiO2, SnO2, ZnO2, Sb2O5 and Y2O3 can be used. In one embodiment, adsorption and adhesion to the glass surface is accomplished by selecting particles that carry an electrostatic charge that is the opposite of the surface
charge of the workpiece. Most of the above-mentioned oxides have a positive charge in the pH regime between 3 and 5 where glass has a negative surface charge. The spacing between the workpieces can be controlled by selecting particles
with a suitable size, e.g. between 0.1 to 10 μm.
In one embodiment, The oxide particles can contain other components in addition to the metal oxides. In lieu of oxides, carbides or nitrides such as SiN, SiC, refractory carbides or nitrides can be used. In addition, refractory or other metal oxides can be used. Also, other ceramic materials can be used. Each particle can comprise components of different materials. These components can be in the form of overcoats (as in the case of alumina-coated silica discussed above) or in the form of solid solutions. Thus, for example, silica particles can be pure silica, substantially silica, or predominantly (mostly) silica. Ceria particles can be used that are pure ceria, substantially ceria or predominantly ceria. Alumina particles can be used that are pure alumina, substantially alumina, or predominantly alumina.) While the invention has been described with respect to specific embodiments, those skilled in the art will appreciate that changes can be made in form and detail without departing from the spirit and scope of the invention. For example, workpieces can be immersed in dispersions and/or solutions of different concentrations and temperatures for different lengths of time. The dispersions can be either aqueous or non-aqueous. (For example, the particles can be dispersed in alcohol.) During heating, the workpieces can be stacked over one another or side by side. The workpieces can be heated to different temperatures (e.g. greater than 300°C but typically less than 800°C). In one embodiment, the workpieces are heated at or above their annealing point. The workpieces can comprise silica glasses having
different compositions and additives. For example, the workpieces can be aluminosilicate or borosilicate glasses. In other embodiments, the workpieces can be a soda lime glass or lead glass. The workpieces can also comprise glass ceramic. In one embodiment, one or more structures that the workpieces rest against or touch during heating (e.g. the flattening plates) are coated with the above-mentioned particles to prevent the workpieces from sticking to these structures. (This is useful if the one or more structures comprise silica, e.g. quartz.) Different aspects of the invention can be practiced either independently or in conjunction with other aspects of the invention. Accordingly, all such changes come within the present invention.
TABLE I
Type of Glass Sample Initial average Average flatness Initial Waviness after Plates Treatment Number flatness (μm) after flattening (μm) Waviness (nm) flattening (nm) Bonding
Stainless steel None 1 41 33 Yes
Soaked in HCl 5 36 20 23 31 No Coated with silica 2 29 18 16 121 No
Alumina None 1 35 22 Yes
Soaked in HCl 5 39 23 39 40 No
Coated with silica 2 44 19 24 96 No
Quartz None 1 42 22 Yes
Soaked in HCl 5 34 19 33 10 No
Coated with silica 2 33 11 26 109 No
Coated with alumina 3 44 12 19 16 No
Coated with silica 4 31 9 36 23 No
TABLE II
Workpiece Type Sample Number Flatness (μm) Waviness (nm) Without particles 6 31.6 44.4 29.1 NR 10.5 NR NR NR Annealed 7 11.2 9.6 8.5 7.6 8.2 7.7 7.3 5.9 Polished 7 10.0 8.4 6.6 3.6 1.5 2.1 2.0 1.5