Methods for measuring the asymmetry of a glass-sheet manufacturing process

Abstract

Methods are provided for measuring the asymmetry of glass-sheet manufacturing processes. The methods include subjecting glass sheets or test samples taken from glass sheets to an ion-exchange process and measuring warp values. Metrics for the asymmetry of the glass-sheet manufacturing process are then obtained from the warp values. In one embodiment, the metric is independent of the geometry of the glass sheets or the test samples (the BM.sub.1 metric); in another embodiment, the metric is independent of the geometry of the glass sheets or the test samples and substantially independent of the ion-exchange process used in the testing (the ASYM metric).

Claims

1. A method for producing low warp, ion-exchanged, glass sheets comprising: (a) producing a population of glass sheets using a glass-sheet manufacturing process, each glass sheet in the population having an R value that is less than or equal to 2.5×10.sup.−4, where the R value is the ratio of the thickness of the glass sheet to the length of the glass sheet's diagonal; and (b) ion exchange strengthening the population of glass sheets produced in step (a); wherein the glass-sheet manufacturing process has an ASYM value that is less than or equal to 1,500 ppm and a BM.sub.1 value that is less than or equal to 0.002 microns, where the ASYM value is determined by: (IA) producing a glass sheet using the glass-sheet manufacturing process; (IIA) measuring the mass M.sub.0 of the glass sheet or a test sample from the glass sheet; (IIIA) subjecting the glass sheet or the test sample to an ion-exchange process; (IVA) measuring the change in mass ΔM of the glass sheet or the test sample as a result of step (IIIA); (VA) measuring a warp value W for the glass sheet or the test sample after step (IIIA); and (VIA) computing an ASYM value for the glass sheet or the test sample using an equation of the form: $\mathrm{ASYM}=\frac{2R^2}{3}\left(\frac{2}{L_x}+\frac{2}{L_y}+\frac{2}{L_z}\right)\frac{W}{\Delta M}\frac{2M_0\left(W_{in}-W_{\mathrm{out}}\right)}{W_{\mathrm{Glass}}}$ where W.sub.Glass is the molecular weight in grams/mole of the glass making up the glass sheet or the test sample prior to step (IIIA), W.sub.in and W.sub.out are, respectively, the molecular weights in grams/mole of the ions being exchanged into and out of the glass in step (IIIA), taken singly and not as oxides, R is the glass sheet's or the test sample's thickness-to-diagonal ratio, and L.sub.x, L.sub.y, and L.sub.z are, respectively, the length, width, and thickness of the glass sheet or the test sample in the same units as the warp value W; wherein the ASYM value computed in step (VIA) is an average value based on measurements made on multiple glass sheets or multiple test samples; and where the BM.sub.1 value is determined by: (IB) producing a glass sheet using the glass-sheet manufacturing process; (IIB) subjecting the glass sheet or a test sample from the glass sheet to an ion-exchange process; (IIIB) measuring a warp value W for the glass sheet or the test sample after step (IIB); and (IVB) computing a BM.sub.1 value for the glass sheet or the test sample using an equation of the form:
BM.sub.1=K.Math.W.Math.R.sup.2 where K is a constant and R is the glass sheet's or the test sample's thickness-to-diagonal ratio; wherein the BM.sub.1 value computed in step (IVB) is an average value based on measurements made on multiple glass sheets or multiple test samples and K=⅔.

2. The method of claim 1 wherein: (a) the R value of the glass sheets is less than or equal 2.0×10.sup.−4, the ASYM value of the glass-sheet manufacturing process is less than or equal to 1,000 ppm, and the BM.sub.1 value of the glass-sheet manufacturing process is less than or equal to 0.001 microns; or (b) the R value of the glass sheets is less than or equal 1.5×10.sup.−4, the ASYM value of the glass-sheet manufacturing process is less than or equal to 500 ppm, and the BM.sub.1 value of the glass-sheet manufacturing process is less than or equal to 0.0005 microns.

3. An ion-exchanged glass sheet having an R value that is less than or equal to 2.5×10.sup.−4 and a W value that is less than or equal to 48 millimeters, where the R value is the ratio of the thickness of the glass sheet to the length of the glass sheet's diagonal and the W value is the glass sheet's gravity-free warp value, wherein the glass sheet is produced by a process that comprises: (i) producing the glass sheet with a glass-sheet manufacturing process that has an ASYM value that is less than or equal to 1,500 ppm and (ii) ion exchange strengthening the glass sheet.

4. An ion-exchanged glass sheet having an R value that is less than or equal to 2.5×10.sup.−4 and a W value that is less than or equal to 48 millimeters, where the R value is the ratio of the thickness of the glass sheet to the length of the glass sheet's diagonal and the W value is the glass sheet's gravity-free warp value, wherein the glass sheet is produced by a process that comprises: (i) producing the glass sheet with a glass-sheet manufacturing process that has a BM.sub.1 value that is less than or equal to 0.002 microns and (ii) ion exchange strengthening the glass sheet.

5. The ion-exchanged glass sheet of claim 3 wherein the R value is less than or equal to 2.0×10.sup.−4, the W value is less than or equal to 37.5 millimeters, and the ASYM value is less than or equal to 1,000 ppm.

6. The ion-exchanged glass sheet of claim 3 wherein the R value is less than or equal to 1.5×10.sup.−4, the W value is less than or equal to 33 millimeters, and the ASYM value is less than or equal to 500 ppm.

7. The ion-exchanged glass sheet of claim 3 wherein the glass-sheet manufacturing process has a BM.sub.1 value that is less than or equal to 0.002 microns.

8. The ion-exchanged glass sheet of claim 7 wherein the R value is less than or equal to 2.0×10.sup.−4, the W value is less than or equal to 37.5 millimeters, the ASYM value is less than or equal to 1,000 ppm, and the BM.sub.1 value is less than or equal to 0.001 microns.

9. The ion-exchanged glass sheet of claim 7 wherein the R value is less than or equal to 1.5×10.sup.−4, the W value is less than or equal to 33 millimeters, the ASYM value is less than or equal to 500 ppm, and the BM.sub.1 value is less than or equal to 0.0005 microns.

10. The ion-exchanged glass sheet of claim 4 wherein the R value is less than or equal to 2.0×10.sup.−4, the W value is less than or equal to 37.5 millimeters, and the BM.sub.1 value is less than or equal to 0.001 microns.

11. The ion-exchanged glass sheet of claim 4 wherein the R value is less than or equal to 1.5×10.sup.−4, the W value is less than or equal to 33 millimeters, and the BM.sub.1 value is less than or equal to 0.0005 microns.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a graph illustrating the type of warp observed after ion exchange of a glass sheet (glass plate) produced by a manufacturing process, e.g., a float process, having a high level of Z-axis asymmetry.

(2) FIG. 2 is a graph showing concentration versus distance-from-surface after ion exchange for a glass sheet produced by a float process. The upper curve is for the air side of the sheet and the lower curve is for the tin side.

(3) FIG. 3 is a plot of warp in millimeters predicted by Eq. (5) for a representative BM.sub.1 value of 0.003 microns and a sheet thickness of 0.5 mm.

(4) FIG. 4 is a plot of warp in millimeters predicted by Eq. (5) for a representative BM.sub.1 value of 0.003 microns and a sheet size of 2000 mm×1500 mm.

DETAILED DESCRIPTION

(5) As discussed above, the present disclosure relates to quantifying intrinsic asymmetry (Z-axis asymmetry) in the manufacture of ion-exchangeable glass which results in warping or curvature of a glass sheet as a consequence of asymmetric ion distribution through the thickness of the sheet after ion exchange. The quantification provided herein serves as a guide to the manufacture of low-warp, ion-exchangeable glass through the creation and/or maintenance of low asymmetry in the manufacturing process.

(6) The warp resulting from performing ion exchange on a glass sheet produced by a manufacturing process having a high level of Z-axis asymmetry is more severe when the glass sheets are larger in length and width and/or when thinner. As the marketplace evolves toward larger displays that use cover glass (such as pad sizes, notebook sizes, laptop, and TV cover glass) and as the marketplace evolves toward thinner cover glass thicknesses, the problem of warping becomes aggravated.

(7) Glass sheets that develop warp through ion exchange are difficult to handle on removal from a salt bath and difficult to transport within a manufacturing plant, as warp can cause glass sheets to touch each other or to touch parts of cassettes or transport devices with consequent damage to the glass. Thus, even if the warp in a final product might be small due to small product dimensions, when large glass sheets are being processed prior to cutting into small individual products, warp can still be a problem because it can interfere with the overall process for making the product. Processing of larger sheets is a commercial trend and thus the need to control (reduce) warp is an ever pressing problem in the art.

(8) FIGS. 1 and 2 provide background for the analysis presented below. FIG. 1 shows an example of the warp observed for a typical glass plate produced by a manufacturing process, e.g., a float process, having a high level of Z-axis asymmetry and subsequently ion exchanged. For a float process, the sense of warp is concave towards the side of the sheet that was exposed to the molten metal bath (molten tin bath) and persists even when the metal (tin) is removed by grinding and polishing or by etching. The bowl shape seen in FIG. 1 is the shape predicted by the gravity-free analysis used in the equations presented below.

(9) FIG. 2 shows an example of the asymmetry in potassium concentration that can be observed after Na/K ion exchange performed on a glass sheet produced by a manufacturing process having a high level of Z-axis asymmetry, specifically, in this figure, a float process. As can be seen, the difference in ion penetration is rather small, i.e., about 1 micron, with the air side of the glass sheet exhibiting a slightly greater diffusivity. The non-zero asymmetry seen in FIG. 2 creates a non-zero value for the intrinsic asymmetry integral of Eq. (1) discussed below.

(10) It should be noted that the asymmetry shown in FIG. 2 is not always detectable even when warp in a full size sheet is clearly present. This is so because the magnitude of the asymmetry in concentration can be so small that it is nearly within the noise of the measurement. A benefit of the quantification (measurement) techniques of the present disclosure is that they do not depend on the direct measurement of such small concentration differences, but rather employ warp and overall weight change measurements that can be readily and accurately performed using commonly-available laboratory equipment. These easily-performed measurements are in direct comparison to the complex measurement techniques used in, for example, the above-referenced International Publication Number WO 2013/005588.

(11) With the foregoing as background, we now turn to the problem of identifying and quantifying asymmetries of physical properties that give rise to warp after ion exchange. To simplify the analysis, we focus on nominally flat glass sheets of rectangular shape and uniform thickness, it being understood that the processing of other shapes can also have trouble with systematic warp from ion exchange.

(12) The asymmetry of interest is intrinsic (native) asymmetry of physical properties through the thickness dimension of the glass sheet as a result of Z-axis asymmetries in the glass-sheet manufacturing process. We identify two main sources of warp from asymmetric physical properties: (1) asymmetry of the diffusivity (such as can arise from asymmetry of the thermal history of the glass) and (2) asymmetry of the surface chemistry of the glass. The former influences how far and how many ions enter each surface during ion exchange, while the latter affects both how many ions enter and the magnitude of exchanged ion concentration at each surface. Warp results when either or both of these quantities are asymmetric. Other sources of asymmetry of physical properties, such as asymmetry of lattice dilation coefficient (discussed below), asymmetry of Young's modulus, or asymmetry of the Poisson ratio, are also included as these also induce warp during ion exchange.

(13) We denote the concentration of the exchanged ion (e.g., the concentration of K.sub.2O that replaces Na.sub.2O in a K/Na ion exchange) by C(z) where z is the dimension through the thickness. This can be expressed in concentration units such as mole %. In accordance with an embodiment, we represent the intrinsic asymmetry by a concentration integral of the following form, which can be thought of as a “bending moment per unit length” concentration integral in view of the multiplication of the concentration difference by z:

(14) $\begin{array}{cc}{M}_1=\frac{1}{L_z}{\int }_{-L_z/2}^{L_z/2}\left(C\left(z\right)-C_{\mathrm{base}}\right)zdz& \left(1\right)\end{array}$

(15) In this expression L.sub.z is the thickness of the glass sheet, C(z) is the concentration of the larger ion after ion exchange, and C.sub.base is the concentration of the larger ion in the base glass prior to ion exchange. The bottom surface is at z=−L.sub.z/2 and the top surface is at z=L.sub.z/2. The M.sub.1 integral has dimensions of mole % times distance. Thus, the M.sub.1 intrinsic asymmetry value can be thought of as a measure of the asymmetry of the concentration through the thickness times the asymmetry of the ion penetration depth through the thickness, giving rise to concentration times thickness. This integral would be zero if the concentration profile were perfectly symmetric because the value of z would take on equal but opposite values between the center and each respective surface while the concentrations above baseline would be the same.

(16) We denote the top and bottom surface concentration values above baseline by:
ΔC.sub.t=C.sub.surf,t−C.sub.base  (2)
and
ΔC.sub.b=C.sub.surf,b−C.sub.base  (3)

(17) These two values may or may not be identical; when they differ they constitute another source of intrinsic asymmetry of the glass sheet. This intrinsic asymmetry is only revealed by performing ion exchange but it is intrinsic to the glass because the surface concentration is the result of a dynamic equilibrium between the salt bath and the glass, so when the two surfaces reach a different equilibrium it reflects a fundamental difference of the glasses near the top and bottom surfaces.

(18) We denote the surface compressive stress, as measured by, for example, a FSM instrument such as that manufactured by Orihara Industrial Company, by CS. Since this will differ on the top and bottom surfaces in the presence of intrinsic asymmetry, we denote the top and bottom surface values by CS.sub.t and CS.sub.b respectively. As shown below, a surprisingly small difference in CS values can give rise to substantial warp. Because small differences are difficult to determine accurately, in accordance with the present disclosure, intrinsic asymmetry is quantified using a warp measurement rather than by trying to measure a difference in CS.

(19) In accordance with the present disclosure, the difference in CS from the top to bottom surfaces, where each is taken as a positive number representing the magnitude of compressive stress, is related to the M.sub.1 intrinsic asymmetry value by the following equation:

(20) $\begin{array}{cc}{\mathrm{CS}}_t-{\mathrm{CS}}_b=\frac{\mathrm{BE}}{1-v}\left[\Delta C_t-\Delta C_b-\frac{12M_1}{L_z}\right]& \left(4\right)\end{array}$

(21) In Eq. (4), B represents the “lattice dilation coefficient”, E is Young's modulus, v is the Poisson ratio of the glass, and L.sub.z is again the glass thickness. As used herein, the lattice dilation coefficient is defined as a coefficient that converts concentration above baseline into strain, i.e., BΔC.sub.t gives the ion exchange-induced free strain on the top surface and BΔC.sub.b gives the ion exchange-induced free strain on the bottom surface. As shown in Eq. (4), when these differ, their difference contributes directly to a difference in CS between top and bottom surfaces.

(22) The other term in Eq. (4) involves the bending moment integral of Eq. (1), i.e., the M.sub.1 asymmetry value. When more ions enter say the top surface than the bottom surface, then M.sub.1>0. This contributes to CS.sub.b being larger than CS.sub.t from Eq. (4). Physically, added strain in the top half of a glass plate bends the plate in a concave sense toward the bottom half; this introduces increased compressive stress (higher CS.sub.b) on the bottom surface and reduced compressive stress (lower CS.sub.t) on the top surface. Thus warp and differences in CS are correlated.

(23) Using the above definition of warp as a number, the overall warp W in terms of the bending moment integral M.sub.1 can be written as:

(24) $\begin{array}{cc}W=-\frac{6}{L_z^2}\left({\left(\frac{L_x}{2}\right)}^2+{\left(\frac{L_y}{2}\right)}^2\right){\mathrm{BM}}_1& \left(5\right)\end{array}$
where the sheet length and width are L.sub.x and L.sub.y, the thickness is still L.sub.z, and B is again the lattice dilation coefficient.

(25) Because M.sub.1 has units of concentration times distance, BM.sub.1 is dimensionless strain times distance, so overall W has the dimension of distance as it should. Like M.sub.1, BM.sub.1 constitutes an intrinsic asymmetry value in accordance with the present disclosure.

(26) As can be seen from Eq. (5), for a given intrinsic asymmetry expressed as a M.sub.1 value or a BM.sub.1 value, the warp is worse for a larger sheet size (larger L.sub.x and/or L.sub.y value) and for a thinner sheet (smaller L.sub.z value). FIGS. 3 and 4 illustrate the rapid rise in gravity-free warp values with sheet size and sheet thickness as predicted by Eq. (5) for a BM.sub.1 value of 0.003 microns, a typical value for a float process. In FIG. 3, sheet thickness is held constant at 0.5 mm, while in FIG. 4, sheet size is held fixed at 2000 mm×1500 mm. Table 1 tabulates typical warp values for a few representative sheets sizes and thicknesses. The connection between (1) sheet size and thickness and (2) resulting warp that is illustrated in these figures and table is critical for controlling warp for larger and thinner parts.

(27) To simplify the analysis, the warp of Eq. (5) is gravity-free warp. Gravity will tend to flatten a warped glass sheet that is laid out horizontally but the underlying warp of the sheet is still problematic due to the possibility of edge curl or other warp-induced effects such as ripple in the sheet. As discussed below, gravity-free warp values can be obtained from warp values measured under gravity conditions by multiplying the measured value by a correction factor, e.g., by 1.13 if the warp is measured with the sheet in a dome orientation (i.e., concave down) and 1.07 if the warp is measured with the sheet in a bowl orientation (i.e., concave up), which makes the latter configuration the preferred orientation for the warp measurement.

(28) It should be noted that when the lattice dilation coefficient is not known, Eq. (4) can be used to estimate B from an equation of the form

(29) $\begin{array}{cc}B=\frac{1-v}{E}\frac{\left({\mathrm{CS}}_t-{\mathrm{CS}}_b\right)}{\left[\Delta C_t-\Delta C_b-\frac{12M_1}{L_z}\right]}& \left(6\right)\end{array}$

(30) This equation allows us to relate warp directly to the difference in CS values by substituting Eq. (6) in Eq. (5) to get

(31) $\begin{array}{cc}W=-\frac{6}{L_z^2}\left({\left(\frac{L_x}{2}\right)}^2+{\left(\frac{L_y}{2}\right)}^2\right)M_1\frac{1-v}{E}\frac{\left({\mathrm{CS}}_t-{\mathrm{CS}}_b\right)}{\left[\Delta C_t-\Delta C_b-\frac{12M_1}{L_z}\right]}& \left(7\right)\end{array}$
When the surface concentrations are symmetric, this expression simplifies to

(32) $\begin{array}{cc}W=\frac{1}{2L_z}\left({\left(\frac{L_x}{2}\right)}^2+{\left(\frac{L_y}{2}\right)}^2\right)\frac{1-v}{E}\left({\mathrm{CS}}_t-{\mathrm{CS}}_b\right)& \left(8\right)\end{array}$

(33) This last equation is convenient because M.sub.1 divides out and we have a simple way to relate CS differences to warp. To provide a specific example, we assume that the glass sheet is produced by a float glass process for which a typical observed warp after ion-exchange is approximately −0.63 mm for a part with L.sub.x=182 mm, L.sub.y=237 mm, and L.sub.z=0.8 mm. The negative sign indicates that the warp is concave “down” or concave toward the tin side of the glass. The Young's modulus for such a glass is roughly 70 GPa and the Poisson ratio roughly 0.2.

(34) Using these values, Eq. (8) gives an estimate for CS.sub.t−CS.sub.b of 3.95 MPa. This shows that a surprisingly small difference in surface compressive stress, generally too low to measure accurately, results in a substantial degree of warp that is readily observed. From Eq. (5), we can estimate BM.sub.1 for such a glass sheet to be approximately 3.0×10.sup.−6 mm. Taking the lattice dilation coefficient to be approximately 0.001/mole %, we can then estimate M.sub.1 to be approximately 0.0030 mm mole %.

(35) In terms of average diffusivities D.sub.t and D.sub.b of the top and bottom halves of the plate, M.sub.1 can be expressed by integrating an erfc-shaped concentration profile to get

(36) $\begin{array}{cc}{M}_1\approx {\mathrm{\Delta C}}_t\left(\sqrt{\frac{D_{t}t}{\pi }}\right)-\Delta C_b\left(\sqrt{\frac{D_{b}t}{\pi }}\right)& \left(9\right)\end{array}$
where t is the time of diffusion. When characterizing the depth of penetration of ions using for example a FSM instrument manufactured by Orihara, this expression, when written in terms of FSM DOL values for the sheet's top and bottom surfaces, takes the form

(37) $\begin{array}{cc}{M}_1\approx \frac{1}{2.8\sqrt{\pi }}\left[\Delta C_t\left({\mathrm{DOL}}_t\right)-\Delta C_b\left({\mathrm{DOL}}_b\right)\right]& \left(10\right)\end{array}$
where DOL.sub.t is the DOL measured for the top half and DOL.sub.b is the DOL measured for the bottom half.

(38) If the surface concentrations above baseline are the same for top and bottom, and are equal to about 12 mole %, we can estimate the difference in penetration depths DOL for the above M.sub.1 value of approximately 0.0030 mm mole % to be about 1.2 μm. This is again a barely measurable difference in diffusivity or ion penetration depth, but it gives rise to readily observable warp.

(39) The above analysis illustrates 1) the substantial amount of warp that can be observed on larger and thinner parts and 2) the difficulties associated with directly measuring the asymmetry of the physical properties that gives rise to warp after ion exchange. In view of these considerations, in accordance with the present disclosure, direct measurement of warp is used as the experimental starting point for characterizing intrinsic asymmetry of the glass-sheet manufacturing process that produced the warped glass sheet. In some cases, when intrinsic asymmetry is large enough, the integral of Eq. (1) can be evaluated using directly measured concentration profiles from, for example, an electron probe micro-analyzer (EPMA) measurement. However, even in these cases, direct experimental observation of warp is generally more sensitive.

(40) By choosing a piece size (L.sub.x,L.sub.y) that is large enough and a thickness L.sub.z that is thin enough, a consistently large warp value can be observed after ion exchange when a glass manufacturing process produces glass with intrinsic asymmetry. Equation (5) is then the primary equation for transforming such measured values of warp into an asymmetry value (asymmetry metric) for the glass-sheet manufacturing process. In particular, the glass' intrinsic asymmetry (native asymmetry) as revealed by a given ion exchange process can be characterized by the BM.sub.1 value of Eq. (5), independent of the length and width of a given glass plate.

(41) Once a BM.sub.1 value is known for a given glass produced by a given glass-sheet manufacturing process and a given ion exchange process (e.g., time, temperature, and salt quality), the expected gravity-free warp for any glass geometry produced by that manufacturing process can be estimated by Eq. (5) for the same or substantially the same ion exchange process. BM.sub.1 is thus a valuable measure of the intrinsic (native) asymmetry of a glass-sheet manufacturing process and when coupled with a defined ion exchange process, provides a useful method for comparing sheet manufacturing processes and for tracking the effects on warp of changes to such processes.

(42) From the above, it can be seen that the BM.sub.1 metric is of the form:
BM.sub.1=K.Math.W.Math.R.sup.2
where R is the ratio of the glass sheet's thickness to its diagonal given by:

(43) 0$\begin{array}{cc}R=\frac{L_z}{\sqrt{L_x^2+L_y^2}}& \left(11\right)\end{array}$
In Eq. (5) above, K equals ⅔.

(44) Although a single test sample can be used, the final result is normally obtained by averaging BM.sub.1 values for a plurality of test samples, e.g., four or more samples. As a general guide, a BM.sub.1 value (e.g., an average BM.sub.1 value) less than or equal to 0.002 microns, (for example, BM.sub.1≦0.001 microns or BM.sub.1≦0.0005 microns), obtained using a commercially viable ion-exchange process and calculated for K equal to ⅔, is indicative of a glass-sheet manufacturing process that has intrinsic (native) symmetry, and thus glass sheets produced by the process can be expected not to encounter warp problems. In particular, these values are applicable to glass sheets having R values that are less than or equal to 2.5×10.sup.−4 (for example, R≦2.0×10.sup.−4 or R≦1.5×10.sup.−4).

(45) As indicated above BM.sub.1 values depend on the ion exchange conditions used in the measurement process. Specifically, ion exchange conditions can make the warp more or less severe depending on the resulting surface boundary concentrations above baseline, i.e., ΔC.sub.t and ΔC.sub.b, and the depth of penetration of the concentration profile that occurs in Eq. (1).

(46) In accordance with a further embodiment, the effects of ion exchange conditions on the measurement of Z-axis asymmetry for a glass-sheet manufacturing process are minimized through the evaluation of the process' ASYM value. That is, while both BM.sub.1 and ASYM values measure the intrinsic (native) asymmetry of a glass-sheet manufacturing process, ASYM values have the advantage of being substantially independent of ion exchange conditions, provided the ion exchange conditions are sufficient to produce ion exchange strengthening of glass sheets, i.e., although not wedded to any particular ion exchange conditions, an ASYM measurement does assume the use of ion exchange conditions that are reasonable, e.g., ion exchange conditions suitable for use commercially for the glass sheets of interest.

(47) In general terms, an ASYM value is a normalization of a BM.sub.1 value by the number of atoms exchanged. The normalization is achieved by a measurement of fractional mass difference before and after ion exchange, with the fractional mass difference being itself normalized using molecular weights and sample size. While BM.sub.1 values have the units of length, e.g., microns or nanometers, ASYM values are dimensionless, e.g., of the order of a few hundred to a few thousand ppm.

(48) In particular, in accordance with an embodiment, ASYM values are determined using an equation of the form:

(49) $\begin{array}{cc}\mathrm{ASYM}\equiv {\mathrm{BM}}_1\left(\frac{2}{L_x}+\frac{2}{L_y}+\frac{2}{L_z}\right)\frac{M_0}{\Delta M}\frac{2\left(W_{\mathrm{in}}-W_{\mathrm{out}}\right)}{W_{\mathrm{Glass}}}& \left(12\right)\end{array}$
or upon using Eq. (5) for BM.sub.1 and Eq. (11) for R:

(50) $\begin{array}{cc}\mathrm{ASYM}=\frac{2R^2}{3}\left(\frac{2}{L_x}+\frac{2}{L_y}+\frac{2}{L_z}\right)\frac{W}{\Delta M}\frac{2M_0\left(W_{\mathrm{in}}-W_{\mathrm{out}}\right)}{W_{\mathrm{Glass}}}& \left(13\right)\end{array}$
where W, BM.sub.1, L.sub.x, L.sub.y, and L.sub.z are as defined above and are all expressed in the same units, e.g., microns, M.sub.0 is the mass of the sample (i.e., the sample whose warp was measured) just before ion exchange, ΔM is the change in sample mass as a result of the ion exchange and is obtained by using the mass just after ion exchange and cleaning minus M.sub.0, W.sub.Glass is the molecular weight of the base glass prior to ion exchange in grams/mole, W.sub.in represents the molecular weight of the ion (e.g., K.sup.+, not K.sub.2O) being exchanged into the glass in grams/mole, and W.sub.out represents the molecular weight of the ion (e.g., Na.sup.+, not Na.sub.2O) being exchanged out of the glass in grams/mole. The molecular weight of the base glass is obtained by taking the molecular weights of the components of the glass (e.g., SiO.sub.2, Al.sub.2O.sub.3, etc.), multiplying each molecular weight by the component's mole fraction in the glass, and then summing the results.

(51) Although a single test sample can be used, the final result is normally obtained by averaging ASYM values for a plurality of test samples, e.g., four or more samples. As a general guide, an ASYM value (e.g., an average ASYM value) less than or equal to 1,500 ppm, (for example, ASYM≦1,000 ppm or ASYM≦500 ppm), obtained using a commercially viable ion-exchange process and evaluated for K equal to ⅔ in the BM.sub.1 calculation if Eq. (12) is used, is indicative of a glass-sheet manufacturing process that has intrinsic (native) symmetry, and thus glass sheets produced by the process can be expected not to encounter warp problems. This is especially so for glass sheets having R values that are less than or equal to 2.5×10.sup.−4 (for example, R≦2.0×10.sup.−4 or R≦1.5×10.sup.−4).

(52) Eqs. (12) and (13) are based on the observation that the measured native asymmetry grows in proportion to the number of ions exchanged. Accordingly, normalization can be performed using the number of ions exchanged, i.e., by the fractional mass increase produced by an ion exchange process, a readily measurable value. Fractional mass increase is proportional to the average of the top and bottom depths of layer (DOLs), while the native asymmetry is proportional to the difference in the depths of layer. Since both grow with sqrt(time), there is no need to explicitly divide by sqrt(t).

(53) Accordingly, in general terms, based on the above observation, the ASYM value can be expressed as:

(54) $\begin{array}{cc}\mathrm{ASYM}\equiv B\frac{\Delta C_t\left(\sqrt{D_{t}t}\right)-\Delta C_b\left(\sqrt{D_{b}t}\right)}{\Delta C_t\left(\sqrt{D_{t}t}\right)+\Delta C_b\left(\sqrt{D_{b}t}\right)}& \left(14\right)\end{array}$
where D.sub.t and D.sub.b are the diffusivities for the top and bottom of the glass sheet. This equation shows how ASYM is a measure of the difference in number of ions that have crossed the two surfaces normalized by the total number of ions that have crossed the two surfaces. This equation includes the glass' lattice dilation coefficient B since larger B gives larger warp in proportion and thus plays a role in the asymmetry of strain and thus warp, but, as will now be shown, its effects are automatically included in the warp measurement and thus a value for B is not needed to evaluate ASYM.

(55) We begin by assuming a single erfc concentration profile on each side of the glass so that warp can be expressed as

(56) $\begin{array}{cc}W=-\frac{3B\sqrt{t}}{2\sqrt{\pi }}\left[\Delta C_t\left(\sqrt{D_t}\right)-\Delta C_b\left(\sqrt{D_b}\right)\right]\frac{\left(L_x^2+L_y^2\right)}{L_z^2}& \left(15\right)\end{array}$
We also have from Eq. (5) above

(57) $\begin{array}{cc}W=-\frac{3}{2}\frac{\left(L_x^2+L_y^2\right)}{L_z^2}{\mathrm{BM}}_1& \left(16\right)\end{array}$
which implies

(58) $\begin{array}{cc}{\mathrm{BM}}_1=-\frac{B}{\sqrt{\pi }}\left[\Delta C_t\left(\sqrt{D_{t}t}\right)-\Delta C_b\left(\sqrt{D_{b}t}\right)\right]& \left(17\right)\end{array}$
which corresponds to Eq. (9) above.

(59) Again assuming a single erfc concentration profile, an analysis of mass difference during ion exchange gives the following relationship:

(60) $\begin{array}{cc}\frac{\Delta M\left(t\right)}{M_0}=\left(\frac{2}{\sqrt{\pi }}\right)\left(\frac{1}{L_x}+\frac{1}{L_y}+\frac{1}{L_z}\right)\left[\Delta C_t\left(\sqrt{D_{t}t}\right)+\Delta C_b\left(\sqrt{D_{b}t}\right)\right]\left[\frac{2\left(W_{\mathrm{in}}-W_{\mathrm{out}}\right)}{W_{\mathrm{Glass}}}\right]& \left(18\right)\end{array}$

(61) Substituting Eqs. (17) and (18) into Eq. (14) then gives Eq. (13), with the dependence on B removed and the resulting asymmetry value (asymmetry metric) normalized by the molecular weight of the glass, the masses of exchanging atoms, and the sample geometry.

(62) To summarize, in outline, ASYM values are obtained by:

(63) (A) Measuring fractional mass difference on a plurality of samples before and after ion exchange and calculating for each sample

(64) $\begin{array}{cc}\frac{\frac{\Delta M\left(t\right)}{M_0}}{\left(\frac{2}{L_x}+\frac{2}{L_y}+\frac{2}{L_z}\right)\left[\frac{\left(W_{K_{2}O}-W_{{\mathrm{Na}}_{2}O}\right)}{W_{\mathrm{Glass}}}\right]}=\left(\frac{1}{\sqrt{\pi }}\right)\left[\Delta C_t\left(\sqrt{D_{t}t}\right)+\Delta C_b\left(\sqrt{D_{b}t}\right)\right]& \left(19\right)\end{array}$
(Note that the samples' edges may need to be polished prior to testing to avoid errors due to edge chipping.)

(65) (B) On the same samples, measuring warp magnitudes W, including correction for gravity, and calculate for each sample

(66) $\begin{array}{cc}{\mathrm{BM}}_1=\frac{2}{3}\frac{L_z^2}{\left(L_x^2+L_y^2\right)}W=\frac{B}{\sqrt{\pi }}\left[\Delta C_t\left(\sqrt{D_{t}t}\right)-\Delta C_b\left(\sqrt{D_{b}t}\right)\right]& \left(20\right)\end{array}$

(67) (C) Forming the ratio

(68) 0$\begin{array}{cc}\frac{B\left[\Delta C_t\left(\sqrt{D_{t}t}\right)-\Delta C_b\left(\sqrt{D_{b}t}\right)\right]}{{\mathrm{\Delta C}}_t\left(\sqrt{D_{t}t}\right)+\Delta C_b\left(\sqrt{D_{b}t}\right)}=\frac{\frac{2}{3}\frac{L_z^2}{\left(L_x^2+L_y^2\right)}W}{\frac{\frac{\Delta M\left(t\right)}{M_0}}{\left(\frac{2}{L_x}+\frac{2}{L_y}+\frac{2}{L_z}\right)\left[\frac{2\left(W_{\mathrm{in}}-W_{\mathrm{out}}\right)}{W_{\mathrm{Glass}}}\right]}}=\frac{2}{3}\frac{L_z^2}{\left(L_x^2+L_y^2\right)}\left(\frac{2}{L_x}+\frac{2}{L_y}+\frac{2}{L_z}\right)\frac{W}{\Delta M}\frac{2M_0\left(W_{\mathrm{in}}-W_{\mathrm{out}}\right)}{W_{\mathrm{Glass}}}& \left(21\right)\end{array}$
the last expression in Eq. (21) being the ASYM value of Eq. (13).

(69) As a more specific, non-limiting, example, BM.sub.1 and ASYM values are determined by the following procedure:

(70) (1) From one or more glass sheets manufactured using the glass-sheet manufacturing process whose asymmetry is to be measured, cut or extract at least 4 square test samples (test pieces) of size 100 mm on each edge. The thickness is set by the target thickness of the glass sheet to be tested.

(71) (2) Accurately measure the initial mass M.sub.0 of each test sample just before ion exchange in preparation for tracking the change in mass that occurs as a result of ion exchange. Also measure the sample dimensions L.sub.x, L.sub.y, and thickness L.sub.z for later use in calculations.

(72) (3) Subject the test samples to an ion exchange process that is identical to or at least similar to that which is to be used on parts manufactured using the glass-sheet manufacturing process whose asymmetry is being measured.

(73) As used in this step, a “similar” ion exchange process means subjecting the test samples to similar conditions of time in the salt bath, temperature of the salt bath, and purity or composition of the salt bath, with the same heating up and cooling down before and after immersion in the salt bath. If any polishing or etching is done before or after ion exchange in the manufactured part, the same should be done to the test sample. The resulting surface compressive stress CS, which can be measured by the optical measurement instrument FSM-6000 by Orihara, or measured by any other means, is preferably within 10% of the highest CS observed on manufactured parts. This implies that the salt bath should be “fresh”, i.e. relatively pure, and not contaminated by the exchanged ion. High CS is a sensitive measure of the purity of the salt bath. In order to partly remove the influence of asymmetry of the salt bath itself, half the test samples are preferably placed in one orientation of the salt bath relative to the glass manufacturing orientation and half placed in the opposite orientation (see discussion of extrinsic asymmetry below).

(74) Note that if only ASYM values are of interest, the ion-exchange conditions need not be based on those used or to be used on parts manufactured using the glass-sheet manufacturing process of interest, but can be based on other considerations, e.g., easy of use in a testing setting.

(75) (4) Accurately measure the mass just after ion exchange and after cleaning the sample so that only glass remains in the mass measurement.

(76) (5) After all steps of the ion exchange process are completed, including any cleaning and polishing that is intended for the commercial object of interest, measure the maximum warp of the test sample using any of several instruments capable of measuring surface profiles. An example of one such instrument is the KLA Tencor P16 Profilometer, which uses a stylus. Another example is the Tropel Flatmaster MSP-150, which is a non-contact interferometer. Measure the maximum vertical difference from center to edge or corner. Perform the measurement with the part held horizontally.

(77) If there is a detectable curvature to the part, measure it in the “bowl” orientation (see FIG. 1), i.e. with the center lower than the edges, in order to minimize the influence of gravity. If curvature is not readily detectable then measure both bowl and “dome” (center higher than edges) orientations. If those measurements reveal which orientation corresponds with the bowl shape then use the maximum vertical difference from that measurement, otherwise use the smaller of the two maximum vertical differences for the two orientations. Express the result in, for example, microns.

(78) (6) Apply a correction for the influence of gravity by taking each warp value from the previous step and multiplying by a correction factor of 1.07. This correction factor approximately accounts for the flattening influence of gravity on a bowl-shaped piece of glass of the chosen size.

(79) The correction factor of 1.07 (a 7% increase in warp) has been found to be appropriate over the full range of thicknesses from 1 mm to 0.2 mm for square plates of 100 mm on an edge. Density and Young's modulus also affect the gravity correction but these influences are approximately at the 1% level so they can be ignored. In the dome orientation the correction factor is about 1.13 (a 13% increase in warp). The bowl orientation has less impact from gravity and thus is preferred, although the dome orientation and its correction factor can be used if desired.

(80) (7) For each test sample, calculate the value of BM.sub.1 by using a ratio R of thickness to diagonal given by

(81) $R=\frac{L_z}{\sqrt{L_x^2+L_y^2}}$
where L.sub.z is the thickness in the expression
BM.sub.1=⅔W.Math.R.sup.2
where W (the warp in microns) is the warp corrected for gravity of step 6. (Note that if only ASYM values are desired, Eq. (13) can be used and a separate calculation of BM.sub.1 values need not be performed.)

(82) (8) For each test sample, calculate a native (intrinsic) asymmetry ASYM value using Eq. (12) or Eq. (13) above. The final result is obtained by averaging the ASYM values (and BM.sub.1 values, if determined) for the various test samples.

(83) With regard to step (6) of the above procedure, i.e., correction for gravity, it should be noted that there is a trade-off between the sheet size (where larger sheets have more warp) and the impact of gravity on sheet shape (where larger sheets have their shape more influenced by gravity). By keeping the size of the sheet smaller, one also has recourse to various commercial instruments that are capable of highly accurate measurements of the sheet shape. These include the use of interferometry such as the Tropel Flatmaster referred to above. A part size of up to about 200 mm on a side should have an easily detectable warp if there is intrinsic asymmetry in the manufacturing process used to produce the part, and should have little influence of gravity on its shape. Smaller sizes can also be used if the measurement equipment is sensitive enough.

(84) Instead of measuring warp, it may be more convenient to measure an average radius of curvature near the center of the sheet. This radius of curvature C is related to the warp W discussed above by the relation

(85) $\begin{array}{cc}W=\frac{1}{2C}\left({\left(\frac{L_x}{2}\right)}^2+{\left(\frac{L_y}{2}\right)}^2\right)& \left(22\right)\end{array}$

(86) If it is desired to measure the warp or radius of curvature of a larger sheet and obtain the gravity-free result, then an apparatus such as is described in U.S. Pat. No. 7,225,665, entitled “Process and apparatus for measuring the shape of an article”, which describes a measurement method that incorporates immersing the sheet in a fluid to approximate neutral buoyancy, may be used.

(87) It should be noted that even a glass sheet manufactured to have no intrinsic asymmetry of its physical properties through the thickness of the sheet will generally have some nonzero warp after being ion exchanged. This can occur through either slight variations in the manufacturing process or slight asymmetry of the ion exchange process itself. Fortunately, intrinsic (native) asymmetry can be separated from extrinsic (random) asymmetry by performing repeated measurements and averaging.

(88) In particular, extrinsic asymmetry produced by the ion exchange process itself can be eliminated by varying the orientation of the glass sheets within the salt bath relative to their orientation during manufacturing. Using “top” and “bottom” to denote the two sides of the manufactured sheet, and “left” and “right” to denote the two possible orientations in a salt bath, ion exchanges can be performed both with the top side facing left and also with the top side facing right. When these are compared, extrinsic asymmetry from the salt bath will tend to make the warp have the same orientation relative to the salt bath (left and right) but it will have a different orientation relative to the manufacturing process (top and bottom). Intrinsic asymmetry from the manufacturing process will always have the same warp orientation relative to the manufacturing process (top and bottom), but as a result of flipping, it will have different orientations relative to the salt bath (left and right). Random warp will have values (typically small) that are random relative to the manufacturing process orientation. Accordingly, when multiple sheets are ion exchanged with both orientations of top and bottom relative to left and right of the salt bath, the sign of the observed warp will flip when the top and bottom sides are exchanged, but not when left and right are exchanged. Averaging will then tie any remaining nonzero value to systematic warp that is intrinsic to the glass prior to ion exchange. That is, averaging will help remove extrinsic warp from the ion exchange process itself and will average over random warp that could arise from the warp measurement gauge or from random (small) variations in the manufacturing process. The choice of how many sheets to average depends on how sensitively the warp can be measured.

(89) Using an eight-step measurement procedure of the type described above with flipping and a typical, commercially-acceptable ion exchange process, BM.sub.1 and ASYM values were determined for glass sheets produced by a fusion process and glass sheets produced by a float process. The results are shown in Table 2, where samples 1-9 are fusion glass samples and samples 10-22 are float glass samples. W.sub.Glass for the fusion and float samples were 63.88 grams/mole and 63.18 grams/mole, respectively; W.sub.in (K.sup.+) was 39.09831 grams/mole and W.sub.out (Na.sup.+) was 22.98977 grams/mole. The column labeled “Corrected Warp” in Table 2 refers to measured warp values corrected for the effects of gravity as described above.

(90) As can be seen in Table 2, the fusion glass had substantially smaller BM.sub.1 and ASYM values than the float glass, the average over all of the samples of the BM.sub.1 values being ˜0.0005 microns for the fusion process and ˜0.003 microns for the float process, and the average over all of the samples of the ASYM values being ˜240 ppm for the fusion process and ˜2080 ppm for the float process. Consequently, the float glass can be expected to exhibit warp problems when used in various applications for chemically-strengthened glass sheets.

(91) Using BM.sub.1 and/or ASYM values as a guide, the intrinsic (native) asymmetry of the float process can be reduced so as to minimize its warp problems. Likewise, one or both of these values can be used as a guide to maintain the high level of symmetry of the fusion process. In connection with both processes, the techniques disclosed herein can be used to monitor Z-axis asymmetry during on-going production and/or to track changes in Z-axis asymmetry associated with changes to existing processes. In these ways, thin and large glass sheets with low warp after ion exchange can be produced by achieving and/or maintaining the intrinsic symmetry of the physical properties of the glass through the thickness direction, including its thermal history.

(92) In particular, based on the ASYM and the BM.sub.1 metrics disclosed herein, as well as the analysis of the physical processes underlying warp discussed above, novel, low-warp, glass sheets are produced which, in an embodiment, have R values that are less than or equal to 2.5×10.sup.−4 and have ASYM values that are less than or equal to 1,500 ppm (BM.sub.1 values (K=⅔) that are less than or equal to 0.002 microns). In other embodiments, the glass sheets have R values that are less than or equal to 2.0×10.sup.−4 and ASYM values that are less than or equal to 1,000 ppm (BM.sub.1 values (K=⅔) that are less than or equal to 0.001 microns), while in still further embodiments, the glass sheets have R values that are less than or equal to 1.5×10.sup.−4 and ASYM values that are less than or equal to 500 ppm (BM.sub.1 values (K=⅔) that are less than or equal to 0.0005 microns).

(93) For these novel glass sheets, the ASYM and the BM.sub.1 values are determined for the sheets in question or test samples taken from the sheets in question, while the R values are for the full sheets. For example, for a glass sheet having a thickness of 0.5 mm, an R value of 2.5×10.sup.−4 corresponds to a sheet diagonal of 2000 mm, an R value of 2.0×10.sup.−4 corresponds to a sheet diagonal of 2500 mm, and an R value of 1.5×10.sup.−4 corresponds to a sheet diagonal of 3333 mm. For a sheet thickness of 0.7 mm, the diagonals are larger, being 2800 mm, 3500 mm, and 4667 mm for R values of 2.5×10.sup.−4, 2.0×10.sup.−4, and 1.5×10.sup.−4, respectively.

(94) In view of their low ASYM (BM.sub.1) values, the full sheets exhibit gravity-free warp values (W values) that are less than or equal to 48 mm for the R=2.5×10.sup.−4, ASYM=1,500 ppm (or BM.sub.1=0.002 microns) combination, full sheet W values that are less than or equal to 37.5 mm for the R=2.0×10.sup.−4, ASYM=1,000 ppm (or BM.sub.1=0.001 microns) combination, and full sheet W values that are less than or equal to 33 mm for the R=1.5×10.sup.−4, ASYM=500 ppm (or BM.sub.1=0.0005 microns) combination. Full sheets having such low warp values facilitate handling of the sheets during manufacture and use. The ultimate product also benefits from such reduced warp, especially where the chemically-strengthened glass incorporated in the ultimate product has a large surface area.

(95) A variety of modifications that do not depart from the scope and spirit of the invention will be evident to persons of ordinary skill in the art from the foregoing disclosure. The following claims are intended to cover the specific embodiments set forth herein as well as modifications, variations, and equivalents of those embodiments.

(96) TABLE-US-00001 TABLE 1 Representative Calculated Gravity- Free Warp Values for Float Glass (Lx{circumflex over ( )}2 + Lx(mm) Ly(mm) Lz(mm) Ly{circumflex over ( )}2)/Lz{circumflex over ( )}2 R × 10.sup.3 Warp(mm) 60 44 0.5 22144 6.71 0.10 1910 1575 1.1 5065062 0.44 22.79 136 170 0.8 74056 3.68 0.33 182 237 0.8 139520 2.67 0.63 500 500 0.5 2000000 0.71 9.00 2000 1500 3 694444 1.20 3.13 2000 1500 1 6250000 0.40 28.13

(97) TABLE-US-00002 TABLE 2 Representative ASYM and BM.sub.1 Values For Glasses With and Without Intrinsic (Native) Asymmetry Corrected Sample Thickness Warp Warp BM.sub.1 ASYM Number (mm) ΔM/M0 (mm) (microns) R (microns) (ppm) 1 0.55 0.00393 0.0485 51.895 0.003889 0.000523 247 2 0.55 0.00308 0.048 51.36 0.003889 0.000518 312 3 0.55 0.00461 0.059 63.13 0.003889 0.000637 256 4 0.55 0.00381 0.033 35.31 0.003889 0.000356 173 Average 247 5 1 0.00202 0.014 14.98 0.007071 0.000499 254 6 1 0.00195 0.013 13.91 0.007071 0.000464 245 7 1 0.00197 0.0125 13.375 0.007071 0.000446 233 8 1 0.00197 0.014 14.98 0.007071 0.000499 261 9 1 0.00201 0.0115 12.305 0.007071 0.00041 210 Average 241 10 0.55 0.00264 0.2245 240.215 0.003889 0.002422 1722 11 0.55 0.00279 0.2855 305.485 0.003889 0.00308 2070 12 0.55 0.00287 0.2705 289.435 0.003889 0.002918 1908 13 0.55 0.00292 0.242 258.94 0.003889 0.002611 1676 14 0.55 0.00288 0.304 325.28 0.003889 0.00328 2137 15 0.55 0.00308 0.2715 290.505 0.003889 0.002929 1784 16 0.55 0.00278 0.2515 269.105 0.003889 0.002713 1828 Average 1875 17 1.1 0.001194 0.072 77.04 0.007778 0.003107 2467 18 1.1 0.001383 0.0845 90.415 0.007778 0.003647 2499 19 1.1 0.001414 0.081 86.67 0.007778 0.003496 2343 20 1.1 0.001459 0.0785 83.995 0.007778 0.003388 2200 21 1.1 0.001305 0.0765 81.855 0.007778 0.003301 2397 22 1.1 0.001386 0.067 71.69 0.007778 0.002891 1976 Average 2314