USRE24795E

USRE24795E - Temperature

Info

Publication number: USRE24795E
Authority: US
Publication date: 1960-03-15

Description

March l5, 1960 F. A. HUMMEL THERMAL SHOCK RESISTANT CERAMIC BODY 2 Sheets-,Sheet l Original Filed June 23, 1950 `wo immun soasonfnow TEMPERATUREC |00 D500-005mm 700 M am TEMPERATURE C ZE no ,Lilo r A1105: Tsaoz (SYNTHETIC) Lzomhof asm, {snmmlc} Lizomrzoy asox (SYNTHETIC) gg JzVm/Toi BY ATTORNEYS.

D00 ZDDBNWMM'NOM |000 TEMPERATURE'C TEE- F. A. HUMMEL Re. 24,795

THERMAL SHOCK RESISTANT CERAMIC BODY March 15, 1960 Sheets-Sheet 2 Original Filed June 23, 1950 g g g o l l 1 l l l l t l ^^^NNMW^MM lo zo so 60 4o 5o 7o ao 90 A,

` Lazozmzo3 *o* Lilo :Algo554'si0g FUSED SILICA couwcnou, MM PER noo MH TEMPERATURE' c I l l l l I l 1 o l 2W 5m 400 5N m 7m 000 o ATToRN'zvsk.

United States Patent O THERMAL SHQCK RESISTANT CERAMIC BODY Floyd A. Hummel, University Park, Pa., assignor, byV

mesne assignments, to The Carborundum Company, Niagara Falls, N.Y., a corporation of Delaware 8 Claims. (Cl. 106--65) -VMatter enclosed in heavy brackets appears in the original patent but forms no part of this reissue speelfication; matter printed in italics indicates the additionsV made by reissue.

This invention relates to a new ceramic body having improved and desirable thermal shock resistant properties, and, more particularly,` to a ceramic body formed of a monotropic lithium aluminosilicate having a crystalline structure.

This application is a continuaton-in-part application of my pending application Serial No. 72,802, tiled January 26, 1949, now abandoned.

The thermal shock characteristics of most conventional ceramic bodies are such that the bodies are unsatisfactory [which] when they are subjected in use to rapid variations in temperature over a wide range. The inability of the bodies to withstand thermal shock is particularly evident when they are subjected to temperature variations of the order [to] of several hundred degrees C. in time periods ranging from a few seconds to several minutes. The vfailure of ceramic bodies when subjected to thermal shocks of this nature is evident from the appearance of cracks in and breakage of the ceramic bodies. In those cases, such as insulators on furnaces, cooking vessels, crucibles, etc., where the body is subjected to rapid temperature changes of considerable magnitude, the conventional practice is to use the materials alumina, zircon, or cordierite (commercial form) in the formulation of the ceramic body, ceramic bodies formed from these materials exhibiting better thermal shockv characteristics than other conventional ceramic bodies.

One ofthe principal objects of this invention is to provide a novel ceramic body which is capable of withstandingk repeated thermal shocks of considerable magnitude. To this end, the ceramic body is formed from monotropic lithium aluminosilicate crystals in a mannel: .to be more fully described below. I have found that a monotropic lithium aluminosilicate crystal ceramic body has a coeicient of expansion which compares favorably with alumina, zircon, or cordierite bodies, is capable of withstanding repeated thermal shocks, and that the coecient of expansion of the body will vary with the molecular ratio of the silica with respect to the alumina and lithia in the crystal structure of the body.

Lithium aluminosilicates are found in the natural state in the materials known as spodurnene and petalite. Typical chemical analyses of spodumene and petalite are as follows:

spodumene Percent by Petalite Percent by weight weight y 118.24395 Reissued Mar. l5, 1960 Spodumene has been considered to have a theoretical molecular ratio of lithia-alumina-silica of 121:4. Petalite has been generally considered to have a similar molecular ratio of 1:1:8.

The manner in which lithia imparts desirable characteristics to glasses and glazes, particularly from the standpoint of improved iluxing properties, is well-known in the ceramic art. Because of its lithia content, spodumene and petalite have been added to ceramic compositions in varying amounts ranging up to about 15% by weight of such compositions with desirable results, and these materials have been subjected to considerable research investigations as a possible source of lithium. This research has shown that the crystal structure of both spodumene and petalite undergoes a change by the api plication of heat thereto. When heated to a temperature of about 1000 C., the natural crystals of spodumene undergo a physical change and invert to a form conventionally referred to as beta spodumene. The change which takes place results in a noticeable increase in volume of the material and a marked decrease in specific gravity. Natural spodumenes have a specific gravity between 3.13 and 3.20, and natural petalites have a specic gravity between 2.39 and 2.46. The change is an irreversible one, and the crystals will not revert to their original state upon further application of heat thereto. It is with this changed form of spodumene crystal, that is, beta spodumene, that this invention is primarily concerned. Since the change which takes place upon the inversion of the natural spodumene crystal to the beta form is an irreversible one due to the application of heat, the changed form for the purposes of this invention is designated as a monotropic lithium aluminosilicate to distinguish the material with which this invention is concerned from lithium aluminosilicates in ther natural state.

Petalite undergoes a change similar to spodurnene when subjected to heat. Heating of petalite converts the crystal stmcture from its natural state to beta spodumene with the additional molecules of silica in solid solution in the crystal. The conversion of petalite by heating is accompanied with irreversible volume and specific gravity changes as in the case of spodumene. Accordingly, it will be seen that the beta spodurnene resulting from heating of petalite may similarly be defined as a monotropic lithium aluminosilicate.

In a manner to be more specically referred to, the monotropic lithium aluminosilicate crystal may be produced synthetically. I have also discovered that the coefficient of thermal expansion of the resultant material varies with the number of molecules of silica in solid solution in the beta spodumene crystal, and that a control of the coeflicient of expansion may be had by varying the number of molecules of [silicia] silica in solid solution in the crystal.

Various specific examples illustrative of the practice of the invention will now be referred to under headingsl corresponding to the materials from which the low expansivity lithium aluminosilicate crystals are derived.

SPODUMENE Spodumene in its natural state crushed to a size which will pass a 200 mesh screen is preferably subjected to an acetic-acid grind and slip cast by a plaster of Paris molding process to the shape of the articles desired in accordance with conventional practice. The foregoing is merely representative of one method of forming the material into an article of desired shape. Alterna- 0 tively, the material, with or vwithout common organic binders, can be pressed, cast, or extruded into the shape [and] of article desired. Additions of clay in small ask Zircon` for controlling the' porosity"l of the body;r maybe added in accordance with conventional practice. After shaping, the article is dried and in the dry state has astrength adaptable*tomachining'of` the article, and whichcomparestfavorably withconventional driediceramic products.

The dried article is then heated to a temperature; and for a time depending upon the porosity desired in the nished'body. The'heat treatment requires a minimum temperature of' 1000 C; for a period of about one hour iny order that the spodumene crystals will invert from their natural stateto-the monotropic form, the temperature of 1000 C. being a critical temperature at which complete inversion of the spodumene to its low expansivity form takestplace` as will be more fully explained in connection with the accompanyingl graphs. Subjecting the body to heat treatment for a longer period ofl time will haveno effect other than to decrease the porosity ofthe body'. The body may be tired at higher temperatures, butfsuch'temperatures mustbe kept under` 1400 C.-1f4-25' C. atwhich the' crystal structure will' be destroyed byk meltingAY and the formation of a gia'ss.`

Fromv apractical standpoint; novr beneficial results will be obtained from' heatingn above" 1300gr as a non-porous body is formed'at this temperature: atthelower temperature beyondVv thaty necessary to contr'ol the porosityrn'ay also resuit in theA formation' of' glass. Generally" speaking, the body should` not be tired for periods longer than 5 to 6 hours at' the inversion temperature of about 1000` C., and should'y be tired for lesser periods at the higher temperatures, and in no case should the temperature be' takenv above the melting orliquidus temperature of 1400 C.1425 C.

The inversion ofv spodumene or the reaction between synthetic oxidesy of A1203-, SiOz, andv Li2O is defined in the claims as aY reaction pro'motediby` sintering. ThisV inversion involves: thel heating of the A1203, SiO and LizO at a' temperature (l`000`E tovv 1300 C.) to promote; a reaction between the three ingredients to makev a crystalline solid' solution without fusing or melting to form glass-like products.

- In the above described'example of the invention, the inversion of the'natural spodumeneV to its beta low expansivity crystal has been described as taking placeV during the tiring of thel article. As anv alternative'v method of practicing the invention, the [sopdumeue] spodumena maybev tiredI to form the beta orv low expansivity crystal before being shaped intov an article, in which case, the tired material is reground, shaped, and then sintered to form the desired ceramic [bady] body Low' expansivity' lithium aluminos'ilicate PETALITE The mineral petalite may be used to form a low expansivity lithium aluminosilicate ceramic body by following the same procedure as described above for the mineral spodumene. The only' dill'erence' in the use of petalite is the heating temperature to which the mineral is subjected. TheV critical minimum temperature at which'vthe petalite convertsl into beta s'podumene with the additional mois. ofL silica in solid solution. is about? 1050'v C. To obtain a non-porous' body, a slightly" The' liquidus temperalture-y vvhieh` must not. be' exceeded:

crystal* ceramic bodies derived from theomineral spodumene, andv fabricated as described above,1have been found capable` The time of heating 4' in any event is also higher, being ofthe order of about 14-5016.

The performance of bodies derived from the mineral petalite show the same desirable thermal shock characteristics as the bodies derived from spodumene. As will bel presently apparent, the thermal expansiony of the body derived from petalite is less per given temperature change than that of the body-l derived fromispodufmene, and will. accordingly be more desirable for-i some purposes;

SYNTHETIC bonate, any'other material which will yield lithia' (.LiZ'O) may be employed'. The three materials/are proportioned. in predetermined amounts in the mixture to yiel'dftliet-y desired ratio of lithia-alumina-silica such as the ratio 1:1:4 typical of spodumene, the ratiol 1:1'28 typical?, of petalite, and other ratiosl with different molecular/cou tents ofsilica to as high a ratio'[of],as 1: 1120,.` Heating@ the intimate-.mixture at temperatures between 1000" C..4 to l400 C. will produce large quantities of a. purelfsyn;v thetic-.betatspodumene-solidsolutionA phase. The heating may take place after the intimate mixture of raw; materials have been formedv to the shapeoftheeramic: body desired, or the intimatemixture of raw materials mayk be'v calcined at the temperature indicated abovetm bring about the formation of the desired phase after: which the particle size may be reduced', the article'-` shapcd, and linally sintered to` finish. TheL heat treat*` mentrequiresa minimum temperature' of'1000 C.' andi a temperature of 1400*" C.-1450 C. to producefay nom. porous body. Continuing the heat treatment for too-y long afl period at any of the temperatures may result i112v the production of a glassphase in the'bodywhich s'fto` be avoided;

In all of the above examples of the invention, the*v articles maybe formed entirely from the raw materials.- before tiring, or entirely from the beta spodurnene crys'-I tals broughty about fromv tiring the raw materials, the latter' case requiring subsequentv steps of grinding,` sha'p;y ing, and4 sintering. In accordance with conventional? ceramic practice, additive' materials. such. as clay andi Zircon maybe employed', but inlno'rca'se should theadd'ed materialsy bel over 2030% by weight of the finished ceramic body. The additive materials will generally foundk to` increase the thermal coethcient of expansion and decrease the ability of the body to withstand thermal?- shock. For example, the addition of clay as a: bonding'ff materialv to aid in the shaping of the article will formt'L mullite` on firing which has. a relatively` highy thermali coecient: of expansion. the best thermal shock properties aretobe. obtained; and? since 5`% clay by weight of. the vbcidy is' suflicient"topfo-Y vide good binding characteristics, it will beobvious thatLA the use off greater quantitiesV of clay will result in.l no*` additional advantages.v

Numerous bodies have been madefup of each ofthe materials spodumene, petalite, and synthetically com pounded lithia-alumina-silica, all tired to produce thel low expansivity lithia aluminosilicate crystall ceramicbodiesi All of such bodiesy have shown extremely-goodN thermal shock characteristics when subjected to repeated therrnall shocks resulting from rapid: reduction of temperatre' from temperatures of as highas 1200'ov CL to' rjoom temperature. The ability of these bodies toy witiigl-f stand thermal shock, is much improvedl as` compared? to' conventional thermalI shock resistant bodies such as alu?. mina, zircon andl Cordier-itey bodies.

In: the accompanying drawings, there is; shown a? ofi graphs illustrating. the thermal expansionY curves. ci typical ceramicl bodies formed in accordance with the Sincer this isto be-a'voided teachings of this invention, and in which all are compared 4with that resulting from the expansion of a body formed of fused silica glass, together with a triaxial diagram illustrating the system in which this invention falls. In this showing:

Fig. 1 is a graph illustrating the terminal expansion of bodies derived from spodumene and tired at temperatures of, respectively, 800 C.; 900 C.; and 1000 C. for one hour;

Fig. 2 is a graph similar to Fig. 1 illustrating the thermal expansion of bodies derived from spodumene and fired respectively at 1100 C. for one hour and at l200 C. for one hour.

Fig. 3 is a similar graph for bodies derived from petalite tired at temperatures respectively of 1000 C. for l5 minutes and of 1100 C. for one hour;

Fig. 4 is a graph similar to Fig. 3 for a body derived from petalite and tired at 1250' C. for one hour, this graph being on an enlarged scale to give a better cornparison with fused silica glass;

Fig. 5 is a graph showing and comparing typical expansion curves of alumina, Zircon, cordierite (commercial), silica glass, and synthetic low expansivity lithia aluminosilicate crystalline bodies containing the different mols. of silica indicated;

Fig. 6-is a triaxial diagram representing the system LizO; A1203; Si02 and showing compositions in this system demarcating the field of this invention; and

Fig. 7 is a graph similar to Figs. 1 through 5 illustrating the thermal expansion of 4bodies in a negative range. To obtain the data from which the graphs shown in Figs. l and 2 were drawn, natural spodumene was pressed into bars having the dimensions 1 cm. x 1 cm. x 11 cm. and tired to 800 C., 900 C., 1000 C., 1100 C., and 1200 C., respectively, and holding at each temperature for one hour. From these curves, it will be noted that the low or natural form of spodumene has a moderate expansion from room temperature to 900 C.,

giving a coefficient of approximately 10X10-'7 cm./ cm./ C. in this range. The beginning of the alpha-beta transformation is seen in the curve for the 900 C. lire, since a rapid increase in rate of expansion is noted around 875 C. Apparently, only a partial conversion of the low form to the high, (beta) form took place during the original heat treatment of 900 C. The expansion curve for the sample which was red at 1000 C. is radically different from the first two curves due tothe fact that the spodumene is now rather completely converted to the beta form. The coefficient of expansion in the range room temperature to 1000 C. has been lowered to about half its original value and is now about 19x10-7 cm./ cm./ C. Commercial cordierite bodies usually show a` coeicient of 25 10'I or more in the limited range from room temperature to 600 C., and hence beta spodumene provides a good basis for shock resisting bodies which show distinct advantages over the cordierite type. The investigation of the expansion behavior after firing at 1100 C. and l200 C. (Fig. V2) shows that the coeiiicient increases slightly to about 24Xl07 cm./cm./ C.

'The thermal expansion curves for similarly shaped bar s formed from petalite and heated respectively to 1000 C. flor l5 minutes; to 1100 C. for one hour; and to 1250 C. for one hour are shown in Figs. 3 and 4. Petalite begins to convert into beta spodumene at 1000 C. and the curve showingthe expansion of a bar subjected to this temperature for l5 minutes indicates that only partial conversion has taken place with most of the movement comingV in the initial stages of heating and giving a coecient of only 19X 10-7 cm./cm./ C. Firing to 1100 C. for an hour brings about a very unusual changev in' that the expansion jis` a straight line and, about' the same as fused silica.- As shown in the graph in Fig.

4 with an enlarged scale, "a`hetltreatmentat 1250. C.rv

for `one hour shows that the expansion lower than that of fused silica.

In Fig. 5, there are shown the expansion curves for three bodies produced synthetically as described above. In the top curve, the raw materials were mixed in proportions to give a ratio of lithia-alumina-silica in the resultant monotropic crystal of 121:4. Similarly, the bodies, from which the lower two expansion curves were obtained, were produced from raw materials mixed to produce resultant low expansivity crystals in the ratios respectively of 1:1:6 and l:1:8. From the three curves for the synthetic materials shown in Fig. 5, it will be noted that the thermal expansion of the resultant low expansivity crystal bodies decreases as the number of mols of silica in solid solution in the beta1 spodumene crystal structure is increased. Numerous other bodies were produced synthetically with varying silica molecular contents, and the thermal expansion characteristics of such bodies were noted. As a result of the tests on these bodies, it was learned that the thermal expansion continuously decreased as the ratio of the lithia-aluminasilica was increased from 1:1:4 to an intermediate ratio containing over 8, but less than l0 parts silica, at which intermediate ratio, the thermal expansion of the resultant body [start] starts rising. A body having a lithiaalumina silica ratio of 1:1210 has an expansion curve slightly: above that of fused silica. Starting with a ratio of about 1:1:l2, the expansion rises to an impractical value due to the presence of the minerals quartz or [cristabolite] cristobalite in the body, both of which have objectionable crystallographic inversions with attendant volume changes which disrupt the ceramic body. In tests on bodies having varying lithia-alumina-silica ratios varying from 1:1:14 to l:l:20, it was found that the ratio of 1:1:12 was the upper limit at which a body was obtained which has practical thermal shock resistant properties. Attention is particularly directed to the fact that the data obtained from the bodies produced synthetically proves that the thermal expansion of the resultant bodies may be controlled by varying the silica content in the resultant low expansivity lithia-alumina-silica crystal.

In Fig. 5, typical expansion curves for commercial alumina, Zircon, and cordierite ceramic bodes have been drawn for comparison purposes. From these curves it will be noted that the thermal expansion of ceramic bodies formed in accordance with this invention are much lower than those for the conventional ceramic bodies customarily used where thermal shock resistance is a desirable consideration. Particular attention is directed to the fact that the expansion characteristics of the is appreciably ceramic bodies of this invention compare favorably with and may be lower than that for fused silica glass. The importance of the invention in this respect will be particularly evident in view of the fact that it has been heretofore considered impossible to obtain a ceramic body having expansion characteristics as low as that of fused silica glass.

From the foregoing, it will be apparent that this invention provides a ceramic body having greatly improved thermal shock resistance, and a much lower thermal expansion than heretofore considered possible in the ceramic art. This is accomplished by the provision of a ceramic body comprising essentially a low expansivity lithium aluminosilicate crystal structure. As hereinbefore explained, I use the term low expansivity lithium alumino` silicate as generic of forms of that material which has herein sometimes been designated beta spodumene with' The above disclosure describes monotropic lithium aluminoslicate bodies in which the silica content is varied' In this range,it has: been shown that the thermal expansion decreases as the. silica content is increased upv to a mol. content of about* over the range of 1:1:4 to 1:l:20.

emes

.101m 12. flnadditfion, tests have beenrun on bodiescontaining less than 4 mols. silica, and on bodies containing varyingamounts ofsilica overa range offrom 4to less vthan 2 mois. of silica. These Ltests have developedfunexpected results in that the thermal expansion of the bodies has been found to decrease as the mol. `content of silica is `decreased below 4, and that it is possible to obtain bodies having a zero coefficient of expansion, and to obtain bodies havinga negative coefficient of thermal expansion, by decreasingthe mol. silica content of the body within limits.

l tGenerally stated, it has been found that low expansivity lithium aluminosilicate bodies having a molecular composition of 1:1:2, thatis, eucryptite, have a negative coeliicient of thermal expansion. As the silica content is reduced below 4 mols., a reduction in the coefficient of thermal expansion is had, and when such mol. content is reduced to 3 mols., a body having a negative coeicient of expansion is obtained. Since it has been determined that a body containing 3.5 mols. of silica has apositive coeicient of thermal expansion. and one containing 3.0 mp1s. `of silica has negative coefficient of expansion, an indication is given that a body having a practically zero coeicient of thermal expansion exists somewhere between the limits of 1:1:3 and 1:1:3.5. The exact composition of a body having a zero coeicient of expansion has not been determinedlnor would it be of commercial importance and probably not capable of absolute determination. Experirnental data has definitely indicated that the coetiicient of expansion of bodies may be'controlled by varying the silica content.

In Fig. 7., the thermal expansion of various bodiesbetween spodumene 1:1:4 and eucryptite 1:1:2 have been plotted and comparedA with the thermal expansionl vcurve vfor fused silica. From these curves, it will be apparent that the coeicient of expansion may be controlled Iby v varying the silica content of the bodies.

A large number of bodies have been produced syn-` theticallyas described above and tests have been frun on such bodies to determine their expansion characteristics. From these bodies, 40 have been selected as `examples ofthe invention. The bodies thus selected were formed and iired as described above. bodies with the constituents thereof given in percent by weight is Vgiven in the following table:

Percentage composition of mixtures composition Number 11140 A1101 sro 10 `1o` so 2o 10l 7o 3o 1o 6o x '40 10 `50 5 5 15 so s 15 15 '7o 55 25 151 eo V35 15 5o 5 2o 75 15 -25 25 25` 5o 5 so e5 15 a5 5o 25 a5 4o 5 4o 55 5 '50Ty v4s m 15 5oV a5 25 5o '25 Kif-1..........1.1...V 10A 6 3e 2 53.2

...r---..,.,........-.-........ s. 7 f2s. s "51:5 2 @Mi-.vvv'xm-.uqu'rm-sv-r -9-6 3217. 557.1'

The composition of these Percentage composition of mxturesa-.Centnued 4tomposition'lsjirnbar 11130 l A110; 4 SiO,

4.1 i 1a. 9 s2. o a. 5 12. o 84.5 2. 9 9; 9 87. 2

11. s 4o. 5 47.7- s. o 24. 7 54. e 4. 9 15. 5 7s. 5

`5 v55'y 4o As explained above, vM2603 is employed in amounts which will yield the above percentages of LizO upon firing.

The mol. compositions of the above mixtures, with adjustments madeto vindicate the Li2O equal to one mol., are as follows:

Mol composition of mixtures o'mpositlon Number f [LzO] E120 A120; S101A 1 .48s 11. 90 1 .293 0.119 `14 1. 75 s. 5av

1 685 l"1.565 1 A 412 0. 79 1 2. a 5. 4s

1 `1 Il 5 1 1 V3. 0 1 1 BJ 0 nuerypm- 1 1 '2.o Spodumene. 1y 1 4.0 Petallte (beta spodumene solid solution) 1 1 '8; 0,

A comparison of the thermal coefficient of expansion` from Aroom temperature (R.T.) to 1000" C. is `given in thefollowingtable:

Thermal :expansion coefficient in the 'range room temperature to 1000I C.

Composition Number Ooemcientot Expansion 1 Too low melting. 2 Do. :L no. i Do. l

Thermal expansion coecient'in the range room tempen ature to 1000 C.-Continued Composition Number Coeiciexl Ixpansion o 3F (R.T. to 800 0.). 7 45 (RT. to 800 0.). 8 Too low melting. 9 5. 10 Not determined. 11 37 5. 12 4, 13 12. 14 7. 15 8.5. 16 s. 17 -L 1B 10. 19 -44. 20 -25 (RT. to 600 C.) 21 17. 22 40. 23 105. 24 18. 25 -13. 25 140. 27 12.5. 28 6.5. 20 10.0. 30 4.0. 31 5.0. 32 Unstable. 33 vDo.

Eucryptlte -70. Spodumene 10. Petallte (beta spodumene solid solution). 2.

To facilitate consideration of the characteristics of the above bodies, they have been spotted by number on the triaxial diagram of Fig. 6, in which P designates the synthetic composition corresponding to petalite, S the synthetic composition corresponding to spodumene, and E the synthetic composition corresponding to eucryptite.

From the triaxial diagram of Fig. 6 in conjunction with Table No. 3, it will be apparent that useful com.- positions will be found within the area bounded by solid lines and ranging from a few percent to 25% LiZO, from 30% to 82% SiO3 and 13% to 70% A1203. The limitation of a few percent Li2O is meant to be less than 5%, since compositions containing a smaller percentage will be mainly silica and alumina and will contain fairly high percentages of mullite along with spodumene or eucryptite crystals. Some mullite in the bodies will not be found harmful since it increases the refractory properties and does not increase the expansion characteristics too much. The limitation of few percent Li2O is intended to eliminate bodies containing 0% LizO or bodies' consisting essentially of mullite with which this invention is not concerned.

From the accompanying diagram and Table No. 3, it will be noted that the boundaries are established by compositions l through 4 and 8, which bodies melted at too low a temperature and displayed eutectic properties, and compositions 32 and 33 which were unstable. The lower limit of silica content was determined by noting the increasing coeiiicient of thermal expansion. For example, the lowest negative coeflicient of expansion was given by -the composition E, eucryptite. As the mol. silica content of the compositions was reduced below E, that is below 1:1:2, as in the case of compositions 34-37, the negative coellicient of expansion started to fall off. This was due to the production of a composition consisting of eucryptite crystals having a binary compound (lithia alumina), a material having a high coefiicient of expansion, suspended therein. This feature excludes compositions having a smaller percentage of silica than 30% since such compositions would contain too much LizO, A1203 and would result in a body of two crystals respectively having high negative and positive coeflcients of expansion. Such bodies would have poor thermal shock properties, and

would for this 'reason be undesirable. In the range of compositions containing more than 2 mols. of silica, only a single crystal is had which is one having varying mols. of silica in solid solution therewith, that is, essentially either a spodumene or eucryptite crystal. Of course, near the outer limits of the area demarcated on the diagram, small amounts of binary compounds or crystals will be found suspended in the ternary composition, but the composition will be found to be essentially a ternary composition, that is, a monotropic lithium, aluminosilicate composition.

From the foregoing, it will be apparent that the ternary compound, low expansivity lithium aluminosilicate, of this invention is of use over the entire range of from less 1:l:2 to about 1:1:12 as described above. The useful compositions in this system for the purpose of this invention will be found in an area adjacent that and on both sides of the line in which the lithia and alumina are present in the mol. ratio of 1:1. As the mol. silica content is varied in this area, compositions are obtained having different but not necessarily proportional coeflicients of thermal `expansion which vary over a reproducible range from a negative value to a positive value. Since the useful area of the system is that encompassing the line in which the mol. content of alumina and lithia has a ratio of 1:1, it will be understood that any denition calling for substantially 1 mol. lithia, 1 mol. alumina, and a variable number of mols. of silica, is meant to in'- clude those useful compositions in the area about the line representing a ratio of 1:'1 and in which the ratio of lithia to alumina may not be precisely l to 1.

The examples of the invention given in the above tables were all formed by mixing chemically pure lithium carbonate in an amount which would yield the desired amount of ilthia upon firing, chemically pure alumina, and silica in the form known as potters flint. These materials in powder form were mixed intimately and then water was added to facilitate molding. Thereafter the mixture was dried in'the mold at 110 C., and when dry, was heated to 1300" C. and held at this temperature for 24 hours. The molded material was then cooled and crushed in a steel mortar. The crushed material was then mixed with less than 1% of an organic binder, methyl-cellulose and carbo-wax, and pressed in a metal mold to form a bar having the dimensions ll cm. x l cm. x l cm. The bar was then sintered at l300 C. for a period of 30 to 60 minutes to remove the binder by oxidizing. The coellicient of thermal expansion was then taken for each bar over a range from room temperature to 1000 C. In those cases where the range was not taken to 1000 C., the maximum temperature to which the bar was subjected has been noted in the above table.

While I have illustrated and described several embodiments of my invention it will be understood that this is merely by way of illustration, and that various changes and modications may be made therein within the contemplation of my invention and under the scope of the following claims.

I claim:

1. A method of making a thermal shock resistant ceramic body which comprises forming and shaping to predetermined size and contour linely divided particles of LizCOs, A1203 and SiO: which yield upon firing a crystalline structure having essentially the composition LigO, A1203 and Si02 in the approximate range of 1:1:2 to 1:l:l0 of said oxides in the order named, and sintering the body so formed at a temperature between about l000 C. and the liquidus temperature of the mass between about 1300" C. and l450 C.

2. The method of making a thermal shock resistant body which comprises forming and shaping to predetermined size and contour nely divided particles of the beta form tl.. temperature .between :about 1000' larn'1mvth-e -=liqnidus temperature of the ymass between about 130.05 C. fand 14.50 C. until the body has become sintend.

3. The method defined in claim 2 in which thenel-y divided particles are united by a. temporarybond.

v `[4. The method denedrin `claim 2 -in which `the LigO Vis produced in situ yfrom LizCO, .initially lineorptmated into the body and reduced upon heating] 5. The method defined in claim 2 `in which .thefmel'y divided particles are `united'by a` temporarylbond'in which the temporary bond comprises not more than about 5.% of clay. v

6. The method of making a thermal Vshock 'resistant body which comprises formingfandshapingto predetermined size and contour finely divided particles oft-a mineral of the `class consisting of spOdumene, :petalite `and eucryptite and firing the body so formed ata :temperature between about 1000 C. and ythelquidastemperature.of the mass between about 1300" AC.and 11450 QC. fumilythe body has become sintered.

1....4 method according toclaim 6 in which 'not ymore than V5% `of clay is incorporatedin ythe initially formed shape.

'8. .The method of making a thermal shock' resistant body which comprises forming and shaping to predeterminedsze and contour nely ldivided'particles` of a mineral of the classconsistingtof spodumene, `petaltei-and eucryptte, firing the body so formed vat a temperature of about 10001 C. Vto transform the mineral particlesdo `the 4beta form of .ternary crystal structure, andfurther 12 A rngsad body at a temperature.betweentabaut1000C. andthe liquidus temperature lof :the mass between about 1300 C. and 14.50 C. until the body has become sintered.

9. A method of making a thermal .shock resistant ceramic body which comprises forming and shaping to predetermined size and contour Afinely divided particles of LZCOS, Alz03 and Si02 which yield upon firing .a crystalline structure having essentially the composition LgO, A1203 and SiOz in the approximate range of 1:1:2 to 1 :1:10 of said oxides in the order named, firing the body so formed at a temperature of about 100.0 C. to convert the finely divided particles to the beta 'form of a ternary crystalline structure, and Vfurther firing .to sinter the body so formed ata temperature between about 1000 C. and the lquidas temperature .of Vthe mass be tween about 13007 C. and 1450 C.

References Cited in the file of this patent* or theorigmal -patent