US3524737A - Method for thermochemical strengthening of glass articles - Google Patents

Description

United: States Patent ABSTRACT OF THE DISCLOSURE This invention relates to the strengthening of glass articles by creating a surface layer of compressive stress therein through an ion exchange process. More particularly, this invention relates to the strengthening of sodium aluminosilicate glasses wherein potassium ions are exchanged for sodium ions and contemplates heating such glasses in air prior to the ion exchange process.

A recent development in the field of glass technology has been the discovery that alkali silicate glass articles can be strengthened several fold by means of an ion exchange process. In this process, a glass article containing lithium or sodium ions is contacted with a source of alkali metal ions having a larger ionic radius than lithium or sodium at an elevated temperature but below the strain point of the glass for a period of time such that i the larger ions will migrate into the surface of the .glass article and replace the smaller alkali ions therein. Since this exchange is carried out at a temperature below the strain point of the glass, there can be no viscous flow thereof to relieve the compressive stresses set up by the crowding of the larger ions into the glass structure in place of the smaller ions. Thus, a surface compression layer is set up in the glass article which greatly enhances the mechanical strength thereof.

The means for carrying out this large-ion-for-small-ion type of ion exchange has generally utilized two approaches. The first, represented by US. Pat. No. 3,218,220, involves an electrochemical technique while the second, represented by British Pat. No. 917,338, is dependent solely upon thermal diffusion of the ions. This second mode of operation has founded a new dimension in commercial glass manufacturing since it has permitted the production of glass articles exhibiting mechanical strengths heretofore not obtainable. And it is with this second method that the present invention is an improvement thereon.

British Pat. No. 966,733 discusses the unusual effect the addition of substantial amounts, i.e., greater than 5% by weight, of A1 0 to alkali silicate glasses has upon the strength than can be imparted thereto through ion exchange of the large-ion-for-small-ion type. Specific examples are recited therein of the exchange of potassium ions for sodium ions in sodium aluminocilicate glasses and the strengths obtained thereby compared with other sodium silicate glasses wherein the A1 0 content is less than 5% by weight. That patent also observed that a true realistic appraisal of the strength of glass involves the measurement thereof after the surface of the article has been subjected to at least moderate abrasion. Hence, the pristine strength of newly-formed glass articles is very high but usually of very short duration, since normal handling of the articles causes surface damage which drastically reduces the initial strength. Thus, the mechanical strength of a glass article is generally of little significance unless measured after the surface has been subjected to abrasion to simulate conditions like that which the glass might see in service. Therefore, two techniques were devised for simulating the abrasion underice gone by glass articles in service. These techniques were described in that patent and are set out below since these tests were utilized in examining the mechanical strengths of the glasses developed by the instant invention.

In the first type of abrasion, a test piece, generally a 4 inch by inch diameter glass cane, is mechanically mounted and rapidly rotated for about 30 seconds in contact with 150 grit silicon carbide paper under a small constant pressure to maintain um'form contact.

The second type of abrasion, referred to as tumble abrasion, comprises placing ten similar-sized glass rods ma number 0 ballmill jar, adding 200 cc. of 30 grit silicon ca'rlgide particles, and then rotating the jar at 100 r.p. rn. for 15 minutes.

Sprface scratches resulting from the first type of abriasion simulate abrasion encountered in service as a result of rubbing against hard materials as, for example, glass articles rubbing against each other. Surface flaws produced in the tumbling action of the second type of ab fgsion simulate those resulting from a combination of such rubbing abrasion and actual impact. We believe the turiible abrasion more closely simulates actual service abrasion and this type of abrasion practice was utilized in .inost of our test work.

Ip determining the mechanical strength of the abraded glass, modulus of rupture measurements were conducted in the conventional manner on the samples. Such measurements yielded the abraded strength of the glass and reflect the useful or practical strength of the glass.

The invention comprising British Pat. No. 966,733 lay in'the discovery that the presence of substantial amounts of A1 0 i.e., at least 5% by weight, resulted in glasses exhibiting high abraded or practical strengths. Thus, after subjecting the alkali aluminosilicate glasses of that invention to either of the above-described abrasion tests, the glasses would still exhibit etrengths several fold that of annealed glasses of the same composition. Comparisons of this property were made between the aluminosilicate glasses and glasses containing little or no A1 0 and, specifically, the common soda lime glasses of commerce. These comparisons demonstrated that while alkali silicate glasses containing little or no A1 0 could be initially strengthened by ion exchange to high mechanical strengths, the surface abrasion growing out of the abovedescribed tests severely reduced the added strength such that the advantage over annealed glass was nominal. Thus, that improvement in the basic thermal diffusion technique of the large-ion-for-small-ion exchange process comprising British Pat. No. 966,733 has enabled the commercial exploitation of the enhanced strengths obtainable by that technique.

However, although abraded strengths several times greater than those exhibited by annealed glasses and, frequently, three-to-four times than those demonstrated by thermally tempered glasses were obtained in ion exchanged alkali alumino-silicate glasses, means were sought for improving the strengths thereof still further. We have discovered that improvements in strength of up to about 50% can be imparted to a specific range of sodium aluminosilicate glasses by subjecting such glasses to a carefully defined heat treatment prior to the ion exchange process. Hence, our invention comprises heating glass articles of certain compositions in air or other inert atmosphere at temperatures from about 25 C. below the transformation range to temperatures within the transformation range of the glass and then carrying out a potassium-for-sodium-ion exchange at temperatures below the strain point of the glass.

In practicing our invention, glass-forming batches consisting essentially, in weight percent on the oxide basis, of about 50-65% SiO 10-25% A1 0 and 10-20% Na O were melted in open crucibles at temperatures between about l5001600 C. for a period of about-16 hours. In most instances, the melts were mechanically stirred to insure homogeneity. Cane about A" in diameter was drawn from each crucible and, frequently,'-%" thick sheet was rolled.

The glass articles were then placed in an electricallyheated kiln and the temperature raised therein to slightly below or within the transformation range of the particular glass. The transformation range of a glass is that 'temperature area within which a liquid melt is deemed 'to I have become an amorphous solid. This temperature range has been considered as lying between the strain pointcand the annealing point of a glass. The strain points of the glasses of this invention range between about 550-600 C. and the annealing points between about 600 6507 C. The glass articles were maintained within the kiln operating at temperatures slightly below or within the transformation range of the individual glasses for at-least four hours. Much longer times can be utilized, often-with advantage, as far as strength improvement is concerned, but dwell periods much longer than about 24 hours do not significantly enhance the strength with respect to shorter maintenance times and, hence, are economically unfavorable. Therefore, a holding period of about 24 hours is deemed to be a practical maximum although much longer dwell times can be safely employed. For greatest strengths, we have learned that this heat treatment should be conducted at temperatures ranging between about -35 C. above the strain point of the glass. Theheat treatment was conducted in the atmosphere of air present in the kiln it will be appreciated that any inert atmosphere can be employed. 1

After the air heating step, the glass articles were im mersed in a bath of molten KNO operating at an elevated temperature but below the strain point of the g'lass involved. We have found that the highest strengthslare obtained where the ion exchange is carried out at temperatures between about 50150 C. below the strain point of the glasses. The glass articles were then removed from the salt bath, the adhering salt removed with water, and thereafter abraded according to one of the abovedescribed procedures preparatory to testing for modulus of rupture.

Laboratory and field testing has demonstrated that the compression layer resulting from the replacement or sodium ions with potassium ions should 'be at least five microns in depth in order to provide the desired practical or abraded strength. Furthermore, abraded strength commonly increases a maximum value with increase in time or temperature of the ion exchange treatment. This factor illustrates a complex relationship involving ion-exchanged layer thickness, depth of weakening flaws in the 'glass surface, and mechanical strength. In general, we have determined that the ion exchange treatment should be conducted for at least two hours with much longer periods being useful and, often, advantageous. It must be borne in mind, nevertheless, that the ion exchange should not be carried out for such a length of time that the replacement of sodium ions with potassium ions take place throughout the entire thickness of'the article.

Table I records various glass compositions, expressed in weight percent, on the oxide basisascalcuIated from the batch, which were treated in accordance with this invention. The batches may be composed of any materials, either oxides or other compounds, which, on being melted together, are converted to the desired oxide compositions in the proper proportions. Although the melts of Table I are of fairly low viscosity, a conventional fining agent such as AS203 may be included in the batch. Commonly, about 0.5-1.0% by weight is added but, since the quantity remaining in the glass after the batch has been melted is too small to have any significant effect on the fundamental properties of the glass, this constituent, if employed, was not included in the table.

TABLE I 62. 45 53. 6 58. 9 58. 6 60. l 62. 7 18. 0 25. 0 15. 4 17. 5 22. 0 16. 5 l2. 5 15. 0 l8. 8 11. 8 12. 0 ll. 8 4. 05 5. 0 3. 3 3. 6 2. 5 4. 0 2.65 3.4 0.8 2.6 0. 35 0. 4 0. 2 3. 8 0. 4 0. 4 2.0 1.0 3.9 1.0 2.0 B20 2. 0

We have determined that the proportions of SiO;, A1 0 and Na O are critical in assuring the desired superior mechanical strength and that the addition of various compatible metal oxides should be severely limited, not exceeding about 20 weight percent total and, preferably, less than 10 weight percent total. Such compatible metal oxides include K 0, the alkaline earth oxides, ZrO PbO, P 0 TiO and B 0 Such additions should not be made in individual amounts exceeding about 10% by weight, and, preferably, are held to less than 5% by weight. The presence of Li O has a deleterious effect upon the strength of the glass and should not be present in an amount greater than about 1% by weight.

Table II reports various heat treatments and salt bath treatments applied to Example 1 along with modulus of rupture (MOR) measurements made after the tumble abrasion test. The strain point of this glass is 581 C. and the annealing point is 631 C. A bath of molten KNO was utilized in each instance although other potassium salts which are molten at the required temperature range can obviously be employed, as can mixtures of potassium salts be utilized. Table II also supplies a comparison in mechanical strength obtained employing the preliminary heat treatment of the present invention with that attained where the preliminary heat treatment is omitted. Each MOR figure represens an average of measurements made on six samples and is believed to be within an error of :3000 p.s.i.

TABLE II MOR (p.s.i.)

Ion Exchange Tumble Ex. No. Heat Treatment Treatment Abrasion 1.- 565 for24hours..-. 500 C. ior24hours--. 87,000 1.. 570 C. (1 90,000 1.. 570C 88,000 1.. for 1 hour .do 73,000 1-- .for2hours.-.. ...do.... 73,000 1. 85, 000 l. 84, 000 1. 88, 000 1. 74, 000 1 74, 000 1 84, 000 1 89, 000 1 90, 000 1 68, 000 1 75, 000 1 75, 000 1 85, 000 1 88, 000 1 89, 000 1 for 24 hours -do 93, 000 1 600 C.lor 48 hours... 500 C. for 48 hours... 95,000 1 None 525 C. for 16 hours...- 64, 000 1 600 C for 24 hours 37, 000 1 None 550 C. for 6 hour 61,000 1 600 O.for24hou o 80,000 1 None 550 C. for 16 hours. 64, 000 I 600 0.10124 hours 0 86,000 1... None 450 C.for48hours. 64,000 1. for 24 hours ..do 85, 000 1 for 1 hour- 500 C. for 24 hours. 75, 000 1 for 2 hours do 76, 000 1 for 4 hours 87, 000 1 for 8 hours. 88, 000 l... for24hou1s ..do 90,000

Table II clearly demonstrates the effectiveness of the preliminary heat treatment in improving the mechanical strength of Example 1, the preferred glass for this invention, based upon such practical considerations as meltability and formability as well as the capability of exhibiting very high mechanical strengths. The preferred strengthening technique involves heating the glass article in air for about 24 hours. Such procedure generally assures a tumble abraded strength of about 90,000 p.s.i. (r3000 p.s.i.). Where the glass is abraded with silicon carbide paper rather than being tumble abraded, this preferred strengthening technique has yielded modulus of rupture measurements of about 110,000 p.s.i. as compared with about 72,000 p.s.i. wher no preliminary heat treatment in air is employed.

Table III records the effect of a preliminary heat treatment in enhancing the mechanical strengths of the other examples reported in Table I. In each instance, the glasses were immersed in a bath of molten KNO for 24 hours at 500 C. to accomplish the exchange of potassium for sodium ions.

TABLE III Annealing point Strain MOR point (p.s.i.) tumble abrasion Ex. No. Heat treatment for 24 hours for 24 hours None 600 C. for 24 hours The modulus of rupture data reported in Table III (each figure an average of six measurements) amply illustrate the remarkable improvement in mechanical strength which can be imparted to glasses of this invention by employing a preliminary heat treatment prior to the ion exchange step. The depth of the ion exchanged layer generally varied between 24 mm. Example 7 demonstrates that this heat treatment should be carried out at a temperature from about 25 C. below the transformation range of the glass to within the transformation range in order to attain the greates improvement in strength although, as is manifested there, some enhancement in strength is obtained where the heat treatment is undertaken at about C. below the transformation range.

We claim:

1. In a method for strengthening a sodium aluminosilicate glass article by replacing the sodium ions in a surface of the glass article with potassium ions by contacting a surface of the glass article with a source of potassium ions at an elevated temperature but below the strain point of the glass to place a compression layer in the surface of the article, the improvement which comprises forming the glass article to be strengthened from a sodium aluminosilicate glass consisting essentially, by weight on the oxide basis, of about 1020% Na O' 10- 25% A1 0 and 50-65% SiO heating said glass article to a temperature from about 25 C. below the transformation range to within the transformation range of said glass for at least four hours, and then contacting a surface of said glass article with a source of potassium ions at a temperature between about 50-150 C. below the strain point of said glass for a period of time sufficient to place a compression layer in the surface of the glass to a depth of at least five microns.

2. A method according to claim 1 wherein said glass article is exposed to a temperature between 5 -35 C. above the strain point of the glass for at least four hours.

3. A method according to claim 1 wherein said glass article is contacted with a source of potassium ions for at least two hours.

References Cited UNITED STATES PATENTS 3,287,200 11/1966 Hess et al. -30 XR 3,396,075 8/1968 Morris 65-30 XR 3,445,316 5/1969 Megles 6530 S. LEON BASHORE, Primary Examiner J. H. HARMAN, Assistant Examiner US. Cl. X.R.

65-ll6; l0639, 52