CA1144762A - Thermal treatment of glass - Google Patents

Abstract

Abstract Glass is thermally treated by contacting the glass, when it is hot, with a gas-fluidised mixture of particulate materials, at least one of which has gas-generating properties when heated by the hot glass. The materials are mixed in selected proportions which impart to the mixture a thermal capacity and flowability which are such that a required thermal treatment is achieved. The method is suitable for the thermal toughening of glass sheets for vehicle windscreens.

Description

6;~

, ~ R~I~L ~ L~S
Thls inven-tLon rela-tes -to -the thermal ~trea~tmen-t o~f glass ancl more especially -t.o -the tll,ermal toughenin~ of flcr~t glass or ben-t glass sheets, -,For example g].ass shee-ts for use sing'l.y as a mQtor vehicle windscreen7 or as par-t t.~f a laminated motor vc~hicle windscreen, or a side light or rear -I,ight for a rnotor veh.icle, or for use in -the con s-truc'r,.ion o~ wil~dscreeln assem~lies for aircraft ar.cl .r way locomo'.ives.
In our Canadian Pa-t,ent Application No. 259724 there is descri.bed a method :~or thermally -treatirl~ g:L~ss aL ~ticle,g 7 ~r ].1.eating ea~h g].ass ar-ticle to a t:e~npera~tur~
above i-ts strain poi.nt7 a~d. quench~ g the g],,.~ s ar~t1.-,lt.~s :i.n turn in a ga,s~flvidised bed of particulate ma-terjal ~.~h.i.ch ls placed ln a c~uiescen-t urliforrnly expanded stc~~te ~ par iculate ~ idisa-tion by c~nt~ol o~ the di,stribu-tion of ~luidlsi.ng ~as in the particu.l~te rnateri.al at, a g~s flo~
velocl-t~ -through t'he par-ticu'la-te materi--ll be-t~Jeen that veJ.-oci.l.~ corr~sporlding to i.n~ipi.ent flui.disa.tion and that ve~ot,i.~;y corres~ond.lng to rnaxi~l~ e:~pansio~n o:~f -the part;-icu],atr.-~rr!a-i::erial.
This state o ,F .~lui di.sation of -t,he '~ed is SUCh that agi+,ation o.f -the .L`lui..dised partlcula-~ve rnateria I .i.s engen-dered on. t~e l.lot immersed surfaces OI the glass as -the g].ass cools i.n ~the i`..ui.dl.sed bed, but, any transierlt tensile ~t,resses i.f~duc~.?d !.n -"he sl.ll:face o~-t;'ne hot &'.l.ass as ts l.ta~di.~lg ed~e fi.t~s t c~ -taccs the~ flu dised be~d are nc~t so SeVf~?7'e 2S to e~n('.~3~ er -the ,,lass, The~t?fo?.e tl~e proc~ss has a ~.? g'l~ yi.eld~ , :
.

.

'762 ; -2-Th~ dQ~r~ee of toughenlng of a glass sheet which :is i~nersed in such a fluidised bed depends on -the ra-te of hea-t transfer be-tween the fluidi.sed par-ticulate ma-terial and the hot shee-t i~mersed in i.t~ and on -the rapid trans~
fer of hot particles away from the vicinity of t,he glass sheet wi-th a. concurrent rapid supply of cooler particles from the body of the flu.idised bed into the v.icinit-~ o~
the glass sheet.
The movement of'par-ticles in the vicinity of the glass surfaces is more rapid -than the movemen-t of the particles in -the b~lk of the bed~ because of rapid agi.-tation o~ the ~luidised ~ar-ti.cul~t,e material which is engendered on -the hot immersed surfaces of -the glass due to.heating of the particulate ma-terial by -the glass wh:i.ch co:~tinues as -the glass cools in the flui.dised.`bed. .
Agita-tiGn o~ -the particula-te ma-terial at -the ~lass .surfaces is consid.erab].y e~lanced when using a selec-ted par-ticulate rna-terial which ha.s late~l-t gas-evolntion pro--per-ti.e- such tha-t there i.s a rapicl e~olut.iorl o~` gas :f'rGm the parti.culate material when heated in proxi.mity -to the glass ,surfaces.
It has now been ~ound that t-here are three factors which dc~m.inate in cont~olling the tllermal toughenirlg o gla~ss in a gas-f:!.uLdised particulâte ma-terial.~ and ln particul.ir wllich contro] the degree o.~ -toughenlng of a l,.ot glass sheet when cont,acted wit,ll a gas~:flui.di.sed 1-ar:r,.~cul.cl~te materi~l.

7~Z

.
- These f`actors are as follows:-1. The gas-generating properties of the ~ar-ticulate material.

2. The thermal capaci-ty per -unit volume of -the par-ticu~
late ma-terial a-t minimwm fluidisa-tion. This is de-rived from the specific heat of the ma-terial measured of the bed . at 50C and the density of -the ma-terial/measu-red at minimum fluidisa-tion of -the material.
. The "flowabili-ty" of -the par-ticulate ma-terial 7 as defined below, ~hich is the sum of :~our poi.nt scores which. are awarded -to the ma-terial by assessmen-t of four characteristics of -the flowable particulate material. The term "flowabi.li-ty" when used nerein has tha-t meaning.
~hese four characteristics oi' a flowable par-ticulate ma-terial a~ld -the manner of awardin~ poin-t scores are des-cribed in -the ar-ticle "Evaluating Flow Propert;ies O:r Solids" by Ralph I.. Carr Jr., ~hemical Engineeri.ng Volume 72, Number 2, Jamlary 18, 1965 and are as ~ollows:~
1. Compressi.bility = 100 ~ -A~ %
where P = packed bulk densi.-ty and A = aerated bulk density 2. .,~ le of Repose : this is the angle irl degrees between the horizontal and the slope of a heap o~ -the particulate materiai. dropped from a polrlt above the horiz.onlal until a constan-t angle is measured.

3. ~ngle of Spat1lla ~. a spatula is .i.nserted '''' , .

.

~4_ horizontally im o -the bottom of a mass of the dry par-ticulate ma-terial and is lifted straight up and out of the ma-terial. An average value of the angle in degrees to -the horizontal of the side of the heap of material on -the spa1~i.1a is the Angle of Spatula.

4. Par-ticle Size Distr.ibution (called Uniformity C'o-eff-icient in -the above Ir~entioned article):
this is described in the abov~ men-tione~ arti-cl~ as the nu.merical value ar.lived a-t hy divi-ding the width of sieve opening ~i.e particle size) which will pass 60% of the particlllate rnaterial by the width of sieve openlng which will just pass 10% of -the particulate matcrial~
15 All the values of ~article size distribu-tiorl re.ferrec, to herein were measured in known manner by a method ~sing a Coulter counter -to determil1e t~le particle d.:ia~
meters a~propriate -to retained cumulative weight per--centa~es of 40% and 90% corresponding to ~ridths o~
20 ~.leve openings which will pclSS 60% and will just paCs 10~o Of the particula+e material.
The nwmerica1 values of Compressibility, An~le o.f P~epose, an~ ~ngle of Spa-tulca were measured using a Powder Tester manufactured by the Hosakawa 25 i~iicrome-rrics Laboratory, of The Hos~1cawa Iron l~orks, Osaka, Japan, frhic~. Po~c~er Te~ter is speci~iGally ce signed for use in the determin.3.tion o~ the "flowabi1..ty"
of po~.~ders as defi.ned above.

, ~ .

'7~Z .

. The flowability of a particula-te ma-terl.al is basi-cally rela-l.ed to fac-tors such as the rnean particle size~
.
. the particle size distribution, and the shape of the particles which is sometimes referred -to as -the angulari.ty

5 of the particles, -tha-t is whether.they have a rounded or angular shape The'value of. flowabili-ty increases ~li-th increase of -the mean particle size, with decrease of -the . particle size distribution, a~d with decrease in the an-. gulari-ty of the par-ticles.
10 The -thermal capacity per Imit volume a-t minimum fluidisat~.on is dependent on the s~ecific heat of the .' material and on the density of -the fluidised bed at rnini.~ :
mlm fluiid.isa-tion9 which densi-ty increases wil;h decrease o~ the par-ticle size distribu-tion.
A hi.gh value of toughening stress is produced in glass wh~n i.-t is quenched in a fluidised bed ha~ing an.
' .optimwn flowability. Some materials'which produce re-- ' quired toughenillg s-tresses may be comrnercially available.
Other commercially available materials may 'be mod:ified to 20 produce the required toughening stresses by sieving -the rnateriaI to change its mean par-ticle size and particle size distribulion.
A ~roblem exists however in that -there is a limit t~ he ~x-tent 1.o which -the degree o~ toughening st:ress 2 induced i.n gl.ass can be controlled by variation o.~ -the fl.owabili.ty of comrnercially avail.a'ble ma-ter.ials Mater-ials ha~ing t~.e req~ired flo~A,ablli.ty may not ~ie coramer--cially availab'le. The produc-tion'of a large ql.lanti.ty of a r~a1;eri.a~ ha~ing the r~quired flowabi'l.ity may .involve tthe siie~i-n~ o~ a large quan-ti.ty of part;iculate .

1~47~;~

-6 .
. material. In addition ~hen u.s.in~ a single mater.ial,the only way of modifying the therrnal capaci-ty of the flu.i dised bed is by narro~ring the partic].e size distribution9 . so that there is no way of modi.~yi.ng -the therrnal capacity in~ependently of the variation of ~lowabili-ty which is produc~ed by narrowi.ng t.he particle s.ze dis-tributi.on~
It has no~ been :~ound that a particula-te rnaterial can be produced having optimum gas~genera-ti~g proper-ties, :~ thermal capacity and'flowability for the production of required toughening s-tresses in a glass ar-ticle by use of a mixture of particulate materials eaGh o:E whi.ch CO~
, tributes to op-timum pI operties O.L the mix-ture. By ~' selection o~ particula-te materials and the proportior~s in which they are mixed the g~as-fluidised particulate .
ma-terial can be tailored to provide any requ.ired tough~
' ening s-tresses wi-thin a wide,range.
~ ' According to the invention -there is provlded a ; . method of' therlllally treating glass in which glass ls ~.ea-ted to a vrede-termined temperature and is con-tacte~d with a gas~ .idised particulate material~ characteri-sed by employing a par-ticulate mater.al whic~l comprises a mixture o~ a number o~` selected pa:rticulate ma-te.rials~
selec-.;i.na at least one of said particulat;e materials -to have gas~generatlllg prroperties when hea-ted b~r said l~o-t g1ass 9 and rnixln~ -the rla-terials in selec-ted. predeter-mi.-ied prc.~portions which impart to t,he gas-.~'luidised mi.~ture o:~ ,ar-ticulate~ materials a therm,ll capacit~ axld ~:Lowab.lity suoh that a requirecl thermal -treatment o~
the ~lass is achieve'dO

. .
.,, ,.,7, , .. , ~-.. ~. .. . ' '' 1 , '~ ' ':
, ;, . . 7 .
The ~etho~.may be characterised by mlxing said materictls in selected predetermined proportions such that required -toughening stresses are induced in the glass as it cools in said gas--fluidised par-ticulate 5 material from a temperature above i-ts s-train point.
The invention further provides for selecting as the gas~generating particulate material a ma-teri.~.l which is capable of evolving from 4% to 37% of its o~.~m weight of gas when heated to a cons-tant weight at 80()QC~
and mixing said particulate ma-terials in predeterrn:i.red proportions which impart to -the mix-ture a -thermal capa-~i.ty per unit volume at minimum ~:I.uidisation i.l~. the range 1~02 to 1.73 MJ/m3K and a ~lowabili-ty in the .
range 60 to 86.
~hen -thermally toughening a shee-t o~ soda-lime~
silica glass o~ th:ickness in t,he range 2 mm to 2.5 mr~.
by t;h* glass sheet being heated to a temperature i.n. the glass sheet being heated to a -temperature in the range 61~C -to 680C, and the mi.xture being malnGa.ined io. a quiesc~n-t ~miforml~J ex.panded state of particulate fluid.isation, the me-thod is characterised by ~onsti-tu~
ting the par-ticulate ma-terial to engender ln the glass sheet a cen-tral tensile stress in the ra.n~;e 35 to 57 ~IPa.
For ~roducing a windscreen ~lass, the r.~e-t~lod may oe characterised by mi.xing said part.iculate ma-terials in - prede-termlned proportions ~hlch impart to the mixt~ re a ~lo~ab.l.lity in the ra~lge 71 to 8~ 5 and a t~lerm,3.:L capa-city per unit voll~ne at mirllrnum fluidls,~'Gi.on in t~
range 1~0~) MJ~m3K to 103~ MJ/rn3K.

' '76Z

` -8-The gas~generati.ng Material may be ~-alumina~
The ~-aluMina rnay be mixed with ~ a.lv.mina. The mi~.ture ma~T comprise from 7% to 86% by weight o~ ~6~alumina.
Another way of carrying out -the invention is charac-teri.sed by chilling a ho-t glass sheet in a gas-fluidi.sed mixture of a ~as~generating particulate material and at l.east one particulate metal oxide whose thermal capacity per unit volume a-t minimum fluidisa-tion is in the range ~rom 1.76 MJ~m3K to 2~01 MJ/m3K7 and mixing the par^t.iculate materi.als in prede-ter!nined proportions which i.mpart -to the mixt.ure a therma:L capa-city ~er ~mi-t volume at ~i.nimum fluidisa-tio~ the range 1.27 MJ~m3K to 1.76 MJ/m3K ancl a flowability in the range 71 to 82*
The particula-te me-tal oxide may be spheroidal. iron oxide ~a~Fe~03)~ The mix-ture may comprise from 30% -to 70% by weight o~ spheroidal iron oxide. Further~ -the mixture may comprise :Erom 70/0 to 30% by weight o~
alumina as gas-generating ma-terial.~ ~n an~-ther way o:~
operatirlg this aspect of the method 7 the mlxture may cornprise frorn 28% to ~5% by weight of spheroidal iron o~i.de~ and from 45% to 56% by we:ight o~ a~alumina? the re.rnainder being X alumlna as gas-generating mater;.al, The ~articulate me-tal oxide may be ~ircon (ZrO~, SiO2). l'he mixture may comprise frorn 10,~ to 70% by weight of aluminlurn monohydr~-te (Al~0301H20) as gas~
genera'iirlgJ mater.ial and fron ~0~0 '~o 30~ ~y '.leight~ 0:~
æircon.
T~. alfJother e~bodiment -the gas-generat.ing particu~

.~ .. . . .
, ~L4~7~Z
_g late materirll jAS an aluminosilica-te. The aluminos:ilicate may ke zeolite, aind from ~% -to 10% by weigh-~ o~ the zeo lite is mixed with from 90% to 92% by weight of a-alumina to constitute the mixture~
The~ gas-generating particula-te rnaterial may be aluminium monohydrate (Al203.1H20).
Further according to the invention the mixture may comprise si].icon carbide (SiC) mixed with a gas-generating particulate material. In one embodiment the gas-generat~
ing particulate material is aluminl1m1 monohydrate (Al203.1H20), and the mixture comprises 17% by weight of ~ alumini~ rQonohyd.rate mixed wi-th ~3% by weight of si:Licon carbide.
In ano-ther way of Garrying out -the method the gas-generating particula-te material is alurniniwn trihydrate ~A~.203. :3H20), The mixture may comprise two gas~generating par~
culat,e ma-terials, aluminium trihydra-t.e (A~20~.3H20) and '~-alumina in equal proportions.
I~ yet another way o~ carrylng out the~ inven-tion the gas genera-ting particulate ma-terial is so;llum bicar-bonate (NaHC03)~ The mixture m.ay be eonsituted by 1OSb by weight of sodiun1 bic~rbonate mixed with 90% by weigh-t.
of a-alumina.
The inverl-tion also comprehend.s apparltus .~`or thermally -treating g'as.r~ comprlslng a container f'or Oa~-:~iul~ised par-t,iculate ma,;eria:l., gas su-,oply means connec~
te~ '~o the contalner for main-~ainl.ng -t.he state of flu.l.di.~-~L4~i62 --10-- .
sation o~' -the particulate rnaterial~ and means for posi tioning hot glass in the container, characterised in that the container holds a gas-fluidised mixture of selec-ted par-t.iculate materials at leas-t one o~ which ha.s gas-generating properti.es when heated by the hot glass7 which particulate materlals are mlxed in selec-ted pre-determined ploportions which impa~t to the gas~-,.luidised mixture a thermal capaci-ty and ~lowabili-ty such that a required thermal treatment of the glass is achievecl.
10 The invention further comprehends -therma1ly trea-ted glass, for example a thermally -toughened Llat glass shee-t, produced by the method of the invention .Some embodiments will now be described, by way of example, with reference -to the accompanying drawlngs in which -Figure 1 illustra-tes diagra~nmatically a ver-tical sec-tion -through apparatus for -th.ermally toughening gla.ss sheets by the method of the invention, . Figure 2 is a graph of central -tensile s-tress plot-ted agains-t the proportions o~ the mixture of particulate materials cons-ti-tuting the gas-fluidised bed and illus-trating varia~
tion o~ stress wi-th ~ari.ation o~ -those 2~ proportions, u.re 3 is a graph si.mil..ar to Figure ~ lustrat ing varia-tion i.n een-tlal tensile stress ~i-th varia-tions of the proportions in .: ' ':
. . ' . :

~.~L44~'~62 another mix-ture of par-ticulate rna-terials~
Figure 4 is a graph similar to Figure 3 illustra-t-i~g the variatlon of surface cornpressive s-tress in 2.3 mm glass wi-th variation in . 5 the composi-tion of the gas-fluidised bed., - . Figure 5 is a graph similar to Figure 3 lllustrat-ing central tensiIe stress as induced in glass 6 mm -thick, Figure 6 is a graph similar to Figure 4 illus-trat-ing variation of surface compressive stress induced in glass 6 mm -thick, . Figure 7 is a graph similar to Figure ~ ~or 12 mm glass, - ~igure 8 is a graph similar to Figure 4 for 12 mm glass, and Figures 9~ 10 and 11 illustrate the variation of central tensile stress with composi-tion of the mixture of particulate materials in three o-ther ways of carrying ou-t -the method of the invention.
Referring to Figure 1 of the drawings, a vertical toughening ov~n indica-ted generally at 1 has side walls 2 and a roof 3. The side walls 2 and the roof ~ are made of the usual refractory material and the bottom of -the oven has an open mouth defined by an elonga.ted aper-ture 4 in a ~ase plate 5 on which -the oven 1 is supported~
A slidable shu-tter~ not shown3 is provided to close th~

.

.

7~Z
~. .
~ - -12-mouth 4 in known manner. A sheet of glass to be curvedand subsequently thermally toughened is suspended in the oven 1 by tongs 7 which grip the upper margin of the glass sheet 6. The tongs 7 are suspended from a tong bar 8 which is suspended from a conventional hoist, not shown, and which runs on vertical guides 9 which extend downwardly from the o~en to guide the lowering and raising of the tong bar.
A pair o~ bending dies 10 and 11 are located immediately below the mouth 4 of the oven in a heated `chamber 12 whi.ch i.s maintained at a tempera-ture such that the dies are at the same temperature as the hot glass which they bend. The chamber 12 is hea-ted by hot gases fed through ports 12a. When the dies are open -they are located on either side of the path of -the glass sheet 6 The die 10 is a solid male die mounted on a rarn 13 and has a curved front face which de:Eines the curvature to be imposed on the hot glass sheet. The die 11 is a ring frame female die carried by struts 14 mounted on a back ing plate 15 which is mounted on a ram 16, The curvature of the die frarne 11 matches the curvature of the face o~
the male die 10.
The guide rails 9 extend downwardly -to elther side of the bendi.ng dies towards a container for a fluidised bed 17 of particulate refractory material in which the hot cu-~ved glass sheet is~-to be chil-led by lowering the sheet do~nwardly into the bedO
The container f:or the fluidised bed comprises an . .

7~ Z

~ 13-open-topped rectangular tank 18 which is moun-ted on a scissors-lift pla-tform 19. When the pla-tform 19 is in its raised position -the -top edge of the tank 18 is just . below the bending dies 10 and 11.
. 5 A high pressure drop micro-porous'membrane 20 extends across the base of the tank 180 The edges of the,membrane 20 are fixed between a flange 21 on the tank and a flange 22 on a plenum chamber ;. 23 which forms the base of the tank~ The flanges and the edges of the membrane 20 are bolted -together as indicated at 24. A gas inlet duct 25 i.s co~ected to the plemlm chamber and fluidising air is supplied to the duct at a regulated high-pressure. There is a high pressure drop , of at least 60% of the plenum pressure across the mem~
brane 20 which resul-ts in a uniform distri'bu-tion of fluid,ising air in the parti,culate material at a gas flow velocity through the particulate material between that velocity corresponding -to minimum fluidisation with -the particles ~ust suspended in the upwardly flowing air, and that ~elocity corresponding to maximum expansion of the particulate material i-n which dense-phase fluidi-sation is maintained. The expanded bed is in a sub-stantially bubble-free quiescent state of particulate fluidisation with a horizontal quiescent surface through ~Jhich the glass sheet enters the bed.
- The membrane,20 may comprise a steel plate which has a re~ular'dlstribution of hole3 and a number of layer.s of , . . .

4~7~;2 strong micro-porous paper laid on the plate. For example fifteen sheets of paper may be used. The membrane is com-pleted with a woven wire mesh laid on top of the layers of paper, for example stainless steel mesh.
. A basket for catching culle-t may be loca-ted near the 5 membrane 20 and is designed so as not to inter:Eere wi.th ., the uni~orm flow of fluidising gas upwardly from the mem-brane~
The guide rails 9 extend downwardly to a position belo~ the bending dies and terminate in the region of the 10 upper edge of the tankc A ixed. fra~ie 27 is mounted in the tank 1~ and has upturned ,feet 28 at its base to rec,eive the lower edge of a glass sheet immer~sed in -the fluidised bed when the tong bar is lowered beyond the bending dies by the hoist~ ' .
In order to load a glass sheet into the apparatus the scissors-lif-t table 19 is lowered and with -the tong bar in i-ts lowest position at the bottom.of the guide rails the glass sheet to be curved and toughened is loaded onto the tongs 7.
The hoist then raises the suspended glass sheet into `the oven 1 which is maintained at a tempera-ture for example 850C so that the glass shee-t is rapidly heated to a tem-perature near its strain point for example in the range 610C to 6~0C~ Wken the glass skee-t has reached -the 2~ required tempe:ratur,e uniformly, the shu-tter closing the mou-th 4 is opened and a hot glass sheet is lo~ered'by -the hoist into position be ~esn -the open bending clles 10 and 110 .

4 ~7 -15- .

' The rams 13 and 16 are operated and -the dies close to bend -the sheet to a desired curva-ture. When the required curvature has been impar-ted to the sheet, for example to enable the sheet to be used as a component of a laminated windscreen for a motor vehicle, the dies open and -the hot curved glass -sheet is rapidly lowered into the fluidised bed in the tank 18 which has been raised to quenching position by raising -the scissors-:lift table 19 while -the glass sheet was being hea-ted in the oven 1. The fluidised bed is maintained a-t a -tel~erature of be-tween 30C and 150C by a water cobling jacket fixed to -the flat longer walls of the tank 18.
The flui.dised bed 17 is constitu-ted by gas-fluidised particulate ma-terial which is a mix-ture in predetermined propor-tions of a number of part.iculate materials one at least of whlch has gas-generating proper-ties and is capable of evolving gas when the fluidised material contacts -the hot glass.

. ..... ~ , ~
, 7~Z

A suitable gas-generating particulate material is capable of evolving from 4 0% to 37% of i-ts own weight of gas when hea-ted to constant weight at 800C, suitable materials are ~-alumina (~Al203) which is porous and 5 contains. water adsorbed in its pores; aluminosilicates whi.ch are porous and contain water adsorbed in their pores; alumina hydrates such as aluminium trihydrate : (A120~.~H20) containing combined water o~ crystalliza-tion, and aluminium monohydrate (Al203.1H20) containing water o~ crystallization and which is porous wi~h water also adsorbed in the pores; and materials which generate gases o-ther than water, ~or example sodium bicarbonate (NaHCO~).
To produce required toughening stresses in the glass the components of the mixture of par-ticulate ma-terials mixed in predetermined proportions which impart to the mixture a flowability in the range 60 to 86 and. a thermal capacity per unit volume at minimum fluidlsa-tion in the range 1002 to 1~75 ~J/m~K.
Other components of the mixture which are mixed with the gas-genera-ting particulate material are particulate materials which are inert in the sense that substantially no gas is evolved f:rom -the rna-terial when heated. .~xamples are ~-alumina ~x~-Al203); zircon (ZrO2.SiO2); silicon carbide; and spheroidal iron oxlde (a-Fcz03).

: .

.:

' . , '', ' ' ,4t7~Z

~17-These particulate materials are o~ dense non-porous form and are selected to have a flowability and thermal capacity different from that of -the gas~generating particulate material so tha-t, depending on the proportion of the dense non-porous materia]. employed, they are effec~
tive to modify the flowability and thermal capacity of the mixture of particulate materials ~GO an extent such that a required degree of toughening stress is produced in the glass.
It ls thought that when a hot glass shee-t is quenched in a gas-fluidised bed of such a mixture of particulate materials, rapid evolution and expansion of the gas evolved from -the gas-generating particulate mate-rial due to heating of the particulate material in the vicinity of the glass surfaces enhance the localised agitation of the mixture of particulate ma-terials on the glass surfaoes in a manner akin to the boiling of' a llquid~
with the result that there are agitated layers of gas and ': ' ' '' // ' , ~_ '.' , ''.

' ~ ~

3L44~62 ` -18-particulate materi.al streaming over the glass surfaces as the glass is chilled in the fluidised bed.
By the blending of the components of the mixture in predetermined proportions, there is optimum hea-t transfer away from the glass surfaces into the bulk of -the bed which induces the stresses desired to be introduced in-to the glass, and there is continual dissipation -to.the remoter parts of the bed of the heat which is extrac-ted from the glass by the agitation of the fluidised particu-late ma-terial immediately surrounding the glass sheet.
The water cooled jacket 29 keeps the remoter parts of the bed cool so that they act in effect as a h.~at sink.
Severe agitation of -the particulate material on the glass surfaces cont.inues until well after the glass has cooled below its strain point7 which ensures that the centre-to-surface temperature gradients initially induced i.n the glass as the glass is in the fluidised bed, are maintained as the glass cools through its strain poin-t, and the re-quired toughening s-tresses are developed thereafter during the continuous cooling of the glass while it is still immersed in the bed.
The lower edge of the hot glass sheet is uniformly chilled as the lower edge enters the horizon-tal quiescent surface of the expanded fluidised bed. Su.bstantially the same tenslle stresses are generated in different areas of the surface of the edge of the glass sneet so -that there is a very low .incidence of fracture. During the descen-t of the lower edge of the glass into the bed, every . .... : ., .. , - - :
:' ` ~ , : ~ .
:: , ~ 44~7~iZ

"
part of the lower edge always contacts fluidised material which is in a quiescent uniformly expanded state of particulate fluidisation, and this uniform treatmerL-t o..
the lower edge, regardless of s-treaming flow of particulate material which may be generated on the hot glass surfaces by gas evolution rom the gas-generating .
component of the blend, largely ob~iates fracture and the consequent problems of dealing wlth glass fragmen-ts in the bed~ ThisJtogether with the avoidance of losses of glass sheets due to change o~ shape of the glass sheets and/or damage to the surface quality, ensures a commer cially ~riable yield of toughened glass sheets.
Some examples of operation of the invention with selected blends of par-ticulate material are gi.ven below.
In each o~ these examples -the numerical value of the product of the particle density~ ln g/cm~, and the mean particle size in ~lm of each o~ the components o~ the mixture is less than 220. This is a criterion which has been useful for assessing whether an individual particu-late material is suitable ~or fluidisation in a quiescentuni:formly expanded state of particulate fluidisation, when operating with air at ambient conditi.ons of normal temperature and pressure. A mixture of the individual Particulate materials is then capable of being fluidised in a quiescen-t uniformly expanded state of par-ticu].ate fluidisatlcn~ .

~, ` ~ ' ' .

14'7~2 . ~20-~ .
.
The fluidised bed 17 was constituted by a mixture of ~-alumina as gas-genera-ti.ng particula-te rna-terial and a-alumina.
The ~alumi.na used was a microporous material having pores of diame-ters in the range 2.7 to 4.9 rim and having from 20% to 40% of free pore space. The pores contain adsorbed water which is released as gas when the ma-t-erial is heated.
The ~-alumina used had the follow~ng characteristics mean particle size a 119 ~m particle size distribution - 2. 3~F
flowabili.ty = 90.25 water content (weight loss = 4. 3%
at 800C) thermal capacity per unit = 1.09 MJ/m3K
volume at minimum fluidisation The ~-alumina used was a dense non-porous form having the following properties:
mean particle size - 30 ~m particle size distribu-tion = 1.22 flowability = 70 thermal Gapacity per unit = 1.3 MJ/m3 volume a-t minimum fluidisa-tion ExPerimen-ts were conducted wi-th mixtures of the 25 ~-alumina and a-alumina which were blended together in predetermi~ed proportions. Shee-ts of glass of soda- -lime-silica composit~on 2.3 mm thick were cut and the , : -- , .

7~ Z

; , - .
edges of the cut sheets were flnished by being rounded using a fine diamond grit wheel. Each sheet was heated `
to 660C ln the furnace 1 before being bent and quenched in the fluidlsed mixture which was in a quiescent uniformly expanded state of particulate fluidisation.
Table I sets out the characteristics o~ different . ......................... ..
mixtures of these materials in the range 30% to 90% by ;~ ~ weight of a-alumina and 70% to 10% by weight of X -alumina, and also the central -tensile stress induced in glass sheets when quenched. For comparison~ the cen-tral tensile stress produced when using the a-alumina and the ~-alumina alone are also included in Table I.
T~BLE I
____ _.
Percentage by weight in mixture - __ ..__ ~ _.
a-alumina 0/0 3o% 5o% 7o% 90% 100%
. _ , _ _ _ ~ ~
~-alumina100% 7o% 5o% 3o% 10% 0%
, _ ___ _ __ _ _ ._ _ __ flowability of mixture 90.25 81.5 75 7L~ 72.25 7o . _ _ . _ _ ~
thermal capaci-ty of mixture per unit vol-ume at minimu~ flui~
isation (MJ/m~K)~ 1.09 1.16 1.20 1.24 1.28 1.3 - ~ _ _ Central Tensile Stress ~
~MPa) 41 43 49 _49 47 32 --~--. ' .

..
Figure 2 lllustrates the variation o~ cen-tral ten-sile stress with composltion of -the mix-ture.

:, .

.

, .

.

4~76;~

The ~-alumina alone has a flowability which is too high for the production of maximum -toughening stress in the glass sheets due particularly to its large rnean particle size and the fact that the particles are of rela-tively smooth non-angular shape. The addition of a proportion of a-alumina, which has a lower flowabi-lity than the ~-alumina'because of the smaller mean particle size of the a-alumlna and the angularity of i-ts individual particles lowers the flowability of the mix-ture. The flowability of the mixture decreases as theamount of a-al~nina in the mixtur~ is increased 7 and there is a commensurate increase in -the central -tensile s-tress produced. A maximum central tensile stress of 49 MPa is achieved when the flowability has been ad-justed to -the optimum value of 74, and the mixture con-tains about 70% by weight of a alumina and 30% by weight of ~-alumina.
The a-alumina has a higher therma'L capacity than the '~-alumina, and as the propor-tion of a-~alumina in -the mixture i,s increased there is a progressive increase in the thermal capacity of the mixture which con-tri-butes to the increase in stress which is achieved~
Further addition of a-alumina in a proportion above 70% by weight increases the -thermal capacity slightly more and maintains a reasonable flowability, hut de-creases the central tensile stress induced because the proportion of the gas-generating consti-tuent 9 ~ alumina, has been reduced to a low level~ ' ,, , ~x~m~e~
The fluidised bed,was consti-tuted by a mix-ture of .
' , 4~'76~
. . . -23- .
. . .
~-alumina as gas-generating partioulate material, and a~
alumina. .
The ~-alumina used was a microporous material having pores of size in the range 2~7 to 409 nm and having ~rom 20% to 40% of free pore space. The pores contai.n adsor-bed water which is released as gas when the material is heated. - . .
The ~-alumlna used had the following characteri-stics:-10 . mean particle size = 64 ~m particle size distribu-tion = 1.88 flowability - . = ~4 water content (weight loss at 800C) - 4~0 the.rmal capacity per unit volume at minimum fluidi.sa-- tion = 1.06 MJ/m3K
The a-alumina used was the same as in E~ample 1.

..
- .
Experiments were conducted with mixtures of the ~-alumina and a-alumina which were blended together in predetermined proportions from 100% ~-alumina and 0% a-alumina to 0% ~-alumina and 100% a-a].umina.
Table ~sets ou-t the thermal ca.pacity per unit vclume . .

, ' .

,, ' ;' ' ::~.l,'~L~t;i~Z
~ , .
, ,24 ,- . .

at minimum fluidisation and -the flowability Gf the mixtures used:~ .
TABLE II

_ __ _ ~ ~ ~
Mlxture Thermal Capacity V_i~bt ~ ~ r~owc~ t~

~-Alumlna ~ Alumina ~ .~ ~_ ~
100 0 1 . 05 ~34 B6 14 1 .09 82 .75 61 39 1~15 ~ ;79 10 40: 60 1~20 76 22. 78 1.25 . '73.25 7 93 1 ~ 29 71 0100 1 . 30 . 70 Sheets of glass of soda-lime-silica composltion, 2.3 mm thick, were cut, and the edges of the cut sheets were fi~lished b~J being rounded using a fine diamond grit wheelO Each sheet was suspended by the tongs 7 and was heated in -the furnace 1 before being bent and quenched.
The rffs~ts are illustrated in Fig~lres 3 and 4~ The absc.is~a oi each curve ~epresents the composition oi~ -t.he blend in ~rei~ t %. In each of Fi~.lres 3 and 4 there are four cil:c~es corresponding to the ~entral tensj.3e s-tvress (Fi~lrc- ~ ajld to the s~lr;,'ace compressive stress (Figure L~) i.nduced in the 2.3 mm thick glass sheets ~hich have 25 heated to a temperature of 610C7 63GC~ 650C or 670C9 , .

,~ . .

1144 7~
.

and then quenched in the fluidissd bed 1'7 which was maintained in a quiescen-t uniformly expanded state of particula-te fluidisati~n and in the temp~rature rarge 60C to 80C.
The curves show that it was preferable to use from about 7y to about 86% ~y weight o~ a-alumina mixed ~ith ~-alumina. As the proportions of ~-alumina in the mix ture is increased, the central -tensile stress and the surface compressive stress induced in the glass in the thermal toughening process increases up to a ma~im~lm which is achieved when the-amoun~ o.~ a-alun~ina is about 70% to 80% by weight o~ the mixtureO Generally the highest stresses were induced when the amount o~ a-alumi.na present is ~rom 55% to 85~ by weig~llt of +,he m.ix-ture, A
higher proportion of a-alumina in the mixture produces a falling o~ in the induced stresses.
By suitable selection of` the proportions o~ ~'alumi.n~
and a-alumina, the mixture had gas~genera-ting properties, a therma:L capacity per unit volume at mlnim~ fluid~.sation at 50C, and ~30wability, ~hich produced consistently high values of central tensile stres.s and surface compressive ,stress in the shee-t,s o~
glass 2~3 mn thick, For example5 when the glass is heated to 670C and -then ~JLuenched, a required cen-tral tensile stress~ in -the range 4s~ Pa to l~g I~Pa, and corresporlcliLlg s~lr~ace corn-pressi.ve stress, in the range 83 MPa to 103 MPa, can be induced in 1,he glass by se~lQcting the pre~eterminecl ,. ':

1~47~iZ

.
propcrti.ons of ~-al~ina and ~-alumina in the mixture ln -the range lrom 7/0 to 86% by weigh-t o~'~-alumina and frorn 9~% to 14% by weight of ~-alumina.
Exam~le 3 Sheet,s o~ soda-lime-silica glass 6 mm thick were cu-t and edge finished and were,-then heated and quenched in a fluidised ~ed in a quiescent uniformly expanded state of particulate fluidisation cons-tituted by a blend of' the ,.~ same particulate ~-alumi.na and ~-a3umina materials as described in Example 2. Figures 5 and 6 are graphs slmi-lar to Figures 3 and 4, which illustrate -the results obtained ~or glass sheets hea-ted to terlperatures of - 610C, 630C, 650C and 670C and then quenched.
The resu]-ts show that requlred -toughening s-tresses ' can be induced in th.e glass which are a ~unc-tion of -the proportions of ~-alumina and ~-alumina in the mix-ture.
Maximurn stresses were achi.eved when the mix-ture oomprlsed abou-t 65% to 95% by weigh-t of ~'-alllrnirla. For example ' wh~n -the ~lass was heated to 670C and then quenched in a fluidised mi.x-ture of 22~o by weight of ~-al~nina and 78% by weight of ~-alumina, the central tensile stress induced in the glass was 9~ MPa and -the surface comp-ressive stress was 216 ~Pa.
This hi~h s-trength 6 mm thick glass is used in the manufact,u~e of window assemblies ~or aircraft and railway -locomo.,ives.

.

' . :, ' ~ ' .... .

z ~
. . -27-~Similar results were ob-tained when toughening sheets of soda-lime silica glass 10 mm -.thlck. Such glass sheets are-used in the manufacture of wi.ndow assemblies .for air craft which may ~or examp]e comprise two shsets o~
toughened glas,s 10 ~n -thick and an outer sheet 3 mm thick.
- The sheets are lamina-ted together with plastic interlayers of known kind~ .
E~-L~
.
.~ Sheets of soda-lime~silica glass 12 mm thick were GUt and edge finished~ and then were h.eated and quenched in a : . fluidised bed con.stituted by a mix~.ure nf ~ alumina and ~-alumlna in prede-termined proportions in the same ~ray as descr~bed i.n Example 2.
.
Results were obtained for glass shee-ts heated to.
610C, 630C? 650C and 670C with a range of proporti.ons o:E ~a~.umina and ~ alumina, and thb results are il1.ust-rated by the curves of F:igures 7 and 8~
Maximl1m values o~ s'~ress wer~ measured when the ~luid.lsed rnixture comprises about 65% I;o 85% by we.-ight 20. of ~--alumina. When a sheet was hea-ted -to 670C and . quenched in a fluidised. bed o~ a mixture of 22% by welght OL a^--alumlna and 78% by weigh-t of ~-~aluminal tll~ central tensile stress in the gas was 124 MPa and the sur~ace corl~p~essive stress was 261 MPa.
E`~.gutes 7 a~d 8 il1.ustra~e how a wide ranee of values o.ï i,ou~g~hen.~ng stre.sses can b~ induced in the glass, as reql.lire~ y selection of the proportlons of the const~
tuents of the mixture o~ par-t,icula-te ma-t,erials as appro-.
' ' ' ' 4~7~;%
. -28-priate to the te.rnperature to which the glass is heated before cluenchin~.
The results illustrated in F.i.gures 3 to 8 have in common that higher toughening stresses are achieve(l as . 5 the proportion in the mixture of the constituent of higher thermal capaci-ty (a-alumina) is increased up to a point where further increase in the proportion reduces the proportion of gas-generating constituent (~-alumi.n.l) to an inadequate level.
10The range OI proportions of the gas-~generating materlal and the other const.ituent or constituen-ts of the mix-ture ensure a f'lowability of -the mix-ture in the range 60 to 86 which is-such that the na-ture of the agi-ta-t,ion is favourable for cooli.ng the glass at a rate 15 which achieves the required values of stress in the glass.
Cooling of the glass occurs due to the rapld agi~
t;ati.on of the particulate ma-terial in the vicinit~ of' the glass sv.rface, which agita-tion is substan-tial:ly due to the e~olut;ion o~ water vapour .~rom ~--alumina ~onstlt-uerlt o.i' ~,he mixture.
A higher proportion of a-alumina enhances the rate o~ heat ex-traction from.the glass as well as modifying the I'lowability of the mixture.

' 76;~

~29 - ~xa~
The :El.uidised bed was cons-tituted by a mixture of ~-alumina as the gas generating component with a pro-portion of spheroidal lron oxi.de (a~Fe203) and one or two kinds o:~ a-alumina~
The ~-alumi.na ha~ the following ch,arac-teristics~--mean par-ticle size = 84 ~m particle size distribution = 1094 flowabili-ty = 87~25 10 water content (weight loss at 800C) - 6yo thermal capacity per uni-t = 1.053 MJ/m3K
volume at minimum fluidisation The spheroidal iron oxide had the following charac~
~eri.c~.ics:~
mean particle size , - 41 ~m particle size distribution -- 1.69 flowability = 76.5 ~herma1. c,apacity per un:lt ~olume a-t minimum fluiclisati.on - 2.~1 MJ!m3K
llhe first ~-alumina was that as used in Example 1.
The seco~d ~-alun~ina hacl the ~ollowing characteristic,s:~
mean par-ticle size = 24 ~m particle size dlstribution = 1.'25 flowab:ility - 66 thermal capacity per ~nit .~
volume at mlnimum fluidisa-tion = 1.1~2 MJ/~-K
Glass sheets of soda-lime-sili.ca composition 2.3 mm thick ~lere heatec~..to 650C and. quenched in fluidised.

., :
. , .

- : . .

~4~'7~

mixtures o~ th~ above materials which wore in a quiescent uniformly expandéd state of par-ticulate fluidisation.
The characteri.stics of the mixtures and -the result-ing central tensile stresses produced in the glass sheets were as given in Table III.
TABLE III
.~ .. , _ Percentage by weight in mixture ~ (1) ~ (2) ~ (L~) _ 10 ~-alumina 70% 5o% 3oo/o 20% 16%
. . , , ~ _~ _ _~ _ Spheroidal iron oxide 3o% 5o% 7oo/o 35% 28%
. ~ . ~ ~ . ._ ~
a-alwnlna ( 1 ) 45~/D36D/o a~alumina ~2) _ _ _ _ 20%
~_ ~ ~ _ __ __ __ . flowability of mixture 82 79 78 7L~ 73.5 - , _ _ _ __ _ _ . __ _~
15 thermal capaci-ty of mixture per unit vol-ume a-t minimum ~luid-isa-ti.on MJ/m3K 1. 347 1 . 54 1.726 1.502 1. 4L~
_ _. ____ ____ ____ Central Tensile Str~ss _ 45 49 50 57 53.0 Figure 9 illustrates the variation of central ten-sile stress with composition of the mixtures (1), (2), and (3) of ~-alumina and ~-Fe203 in Table III. The cen-tral tensile stresses resulting from use of the ~-25 al~ina alone and the a-FE203 alone were 41 MPa and 32 ~Pa respectively.
As in Example 2 -the ~-alumina used in -this example has a f1owaoility which is too higll lor the production - of maximw!l toughening s-tress in glass sheets. The .
. : ~ . .. .
.

~.
.

4'~ ~
31- . . . .
spheroidal iron oxide has a lower flowability than the ~-alurnina par~ticularly because of its smaller mean par-ti-cle size. The addition of increasing amount,s of' the spheroidal iron oxide to the ~6-alumina in the mixtures (1), (2) and (3~ o~ Table III has a progressive effect of lowering the flowability of the mixture by progressively lowering the mean particle size o'~ the mixture'as the amount o~ spheroidal iron oxide in the mix-ture is in-creased. As the flowability of.the mixture decreases 10 there is a progressive increase in the central tensile stress produced in the glass shee-ts. A maximu~ central tensile stress of 50 ~a is achieved when the mixture contains about 70% spheroidal iron oxide and 30~o~-alu~ina.
The flowa~ility of the spheroidal iron oxide is not as low as -that of the a-~alumina used in Example 1 be-cause.it is Qf larger mean particle ~ize and the parti~
cles are smoo-th].y rounded as compared wi-th the angular particles o~ the a-alumina. Therefore the ~.pheroidal 20 iron oxide is no-t so effective in lowering the :Elowa-bility o~ the mixture as is the a-alu~lina of Example 1.
Mi,x-tu-re (3) of the present example con-taining 70%
by weight spheroidal iron oxide and 30%,by weight~S-alumi.na, which produces -the maximum central tensile 25 stress cf 50 MPap has a.flowability of 78 which i.s higher th.an the optimum flowability of 74 of the mixture o:f 70% b-~ weight of a-alumina and. 30% by weight of ~
al~nina which produces high central tensile stress; ln Example 1~

, . .

' 4~2 The maximum central tensile stress produced by mixture (3) in the present example is howe~er about the same as -the maximum central -tensile stress produced by the mixture of Example 1. This is because, al.-though the flo~rability of mixture (3) is a little higher than -the optimum flowability at ~hich maximurn stress will be pro~
duced, the spheroidal iron oxide used in mixture (3) has a signifi.cantly higher thermal capacity than that o~ the ~-alumina used in Example 1.
Because the flowability of mixture (3) was though-c to be slightly too high~ mix-ture (4) was then made up, containing a propor-tion of the ~-alumina as used in Example 1~ This reduced flowability of the mi.xture to an optim~ ~alue of 74, and the mixture produced a further increase in central tensile s-tress to 57 MPa, despite -the louering of the thermal capacity.
Mixture (4~ has the same optimum value of flowa-bility, 74, as the mixture of Example 1 consisting o~
30% ~-alumina and 70% a-alumina, which produces a maxiJnurn central tensile stress of L~g MPa. The fact that rni.xtu.re (4) produces a higher central tens:ile stress of 57 MPa is due to the higher thermal capaclty : o~ mixture (4), ~hat is.1.502 MJ/m~K as compared with 1.24 MJ/m~K of the mixture of Example 1.
The further reduc-cion in the flowabi.li-ty resul-ting ~rom the inc.usion of a proportion of a second ~-alv~ina in mixture (5) has produced a reduction of the thermal ,, , ~ , a'7~;2 capacity of -the mixture as compared with rnix-ture (4) with an acco~panying small reduction in central -tensile stress.

, /

.
. . ~, , :
, ;~"

.:

6~
-~4-' ~5 ' . , ~ he fluidised bed was constituted b~J a mixtu~e of the gas^generatin~ particulate material aluminium monohydra-te (A1203 1~20) and zircon (ZrO2.SiO2)0 .The aluminium ~onohydrate was in the form of Boehmite~ ~hich is a porous ~aterial containing 15,' by weight of combined water of crystallization and 13~
by ~eight of ~ater in its pores~ Adsor~ed ~ater released during ~uenchin~ of the glass is mainly effective to provide the gas generation ~.~hich gives rise to the enhanced agitation o~ the particulate material in the vîcini-t~ of the ~lass surfaces~
r~he aluminium monoh~drate used had the follo~in~
chara~teristics~
mean particle size = 51 pm particle size distrlbution = 1~70 flo~7ability = 78 water content (~7ei htol)oss = 28~4%

th~rmal capacit~ per unit = 1.18 MJ/~3K
~olume al. minimum fluidisatlon ~` ~'he zircon ~hich is an inert non-porous zirconia-o~thosilicate of hi~,her thermal capacity than ~-alumina had ~he follo~-Ti~ charactexistics:-mean partlcle size - = 34 paxl,icle size distribution = 1.73 i`~o~!~bili-J~ - . - 67 ttler~al ~paci~y per u~i~ - 1.76 MJ/m3K
vol~lme P ~ ni..~
~luidisati.on .

... . . .. . . . .. ...
,~ ' , 7~2 .
~35-Glass sheets 2.3 rnm thick were heated to 660C' and quenched in mixtures of -the aluminium rnonohydra-te and zircon as set out in Table IV? which sets out the characteristics of the mixtures and the central tensile 5 . stress produced in the giass sheets.
TABLE IV .
~ _~_____ . .
. Percentage by weigh-t . in mixture _ aluminium monohydrate 100~o 7o% 5o% 20~ 10% 0%
~.. . ~ _ _ _ _~ _~
zircon % 3o% 5o% ~0% 90% 100%
. . . ~ . . _ ~ ~
flowability of mixture 7~ 5.5 7L~ 73 71 67 ~__ . _ _ ___ , ,, ___ _~_ thermal capacity of mixture per unit volume a-t minimum .
fluidisation MJ/m3K 1.005 1.277 1.41 1.62 1.70 1.76 _ ~., _" . , .
Central l'ensile Stress .
. . . 37 42 44 46-5 - 23 _ .__ .
Figure 10 illustrates the variation o~ central tensile stress with composi-tion of the mixture.
Aluminiurn monohydrate has good gas generating pro-perties and a lower value of flowability -than tha-t of the ~--aluminas referred to in Examples 1 and 5. However the flowability of the aluminium monohydrate is hi.gher than the optimum flowability which will.produce maximum central tensile stress and the thermal capacity is re-latively low.. Zircon has a lower flowab.ility and hi~her ther.mal capacity than the aluminium monohydrate, and as the proportion of zircon in the mixture is increased there is a progressive increase in the centra].

.. .
. ~, .
' ' , .
.. , . . .~ .. . . . . . . ......

, tensile stress produced in the glass due both to the progressive lowering o the ~lowability and increase in the thermal capacity of the mixture.
~ le æircon has a high thermal capacity which con-tributes significantly -to theincrease ln central tensile stress produced in the glass shee-ts inthe same way as the spheroidal iron oxide of Example 5. Becavse the zircon has a lower flowability than the spheroidal iron oxide of ~xample 5, it is more effective in reducing the flowability of the mixture, and therefore makes a greater contribution to the increase in central tensile stress brought about by the reduction in the value of flowability of the mixture.
The maximum central tensile stress of 46.5 MPa is achieved when the mixture contains about 20% aluminium mono~lydrate and 80% zircon, which mixture has a flowa-bility at an op-timum value of 73.
Further addition of zircon abou-t 80% by weight raise.s the thermaJ capacity of the mi~ture, but results in a decrease in central tensile stress due to signi-ficant reduction of the flowability below the optimum value, and reduction o the proportion of the gas-generating constituent, aluminium monohydra-te, to a less effec-tive level.

The fluidised bed was constituted by a miY~ture of`
a-alwmina witll equal proportions o each of four ~
aluminas designated A? B, C and D in Table V which se-ts out the characteristics of -the ~-al~inas.
, ... ~ - -- '-'~

'76'~
~37-.
` TABLE

. .
~- aluminas A B C D

______~ _ __ . __ mean particle size (~m) 70 61 57 72 ~ __ _ _ ~ .. _ .
partlcle size distribution 1.47 1.67 1.66 1O65 ~ . _ ~ ~ ~
flowability 88.5 88 85 86 . ... ~ __ __ ____ water con-tent (% weight . loss at 800C) 7 7 7 7 ,. ____. __~ _~ ~ ~

thermal capacity per unit 10 volum7 at mlnirnum fluidisation : (MJ/m~K) 1.16 1~16 1.12 1.12 The characteristics of the ~-alumina were as follows mean particle size = 22 ~m particle size distribution - 1.69 flowability = 63 thermal capacity per unit volume at rnini.murn fluidisation = 1.24 MJ/m3~.

G~as,s sheets of soda-lime silica co1npos.ition 2.3 mrn thick were heated to 660C and quenched in gas-fluidised 20 mi~tures of the above materials which we~e i.n a quiescent uniforrrlly expanded state of particulate fluidisation~

The charac-teristics of :the mixtures and the resu.lt-~ing cen-trcal -tensile stresses produced in the glass sheet werc. as gi~Ten in Table VI.

' ` .

.. . .

...

~ .

~4~'7~%

- TABLE VI

Percentage by weight ; in mix-ture ___._______ _~ _ _ ~
Mixture o f 4 ~-al~ninas 100~' L~0% 20% 10% 0~
________ _ _ _. _ . . _ c~-alumina 0% 60% 80% 90% 100%
~ . .__ __ . . ~_ .
.~lowabili-ty of mixture 87 7o 67 65 63 ~ ~ v~ ~ ~ _~ _~
thermal capacity oi' mi,xture per ~mit volume at minimum ~luiclisa-tion (MJ/m3K) 1~14 1.20 1.22 1.23 1.24 .. ~ __.. ___ _ _ " 10 Cen-tral Tensile Stress .
(l~Pa) 39 L~0 35 31 ~ 25 ' Figure 11 illustrates -the variation of cen-tral ten-~i1e streSs With composition o~ the mix-ture.
The previous examples have shown how higher stresses 15 can be produced by mix-tures of a gas~generating par-tic-ula-te ma-terial with an inert material, than can be pro-duced by use of -the gas-genera-ting particulate mater-,al ' alone. However i.t may be desi,red to produce lower stress values than can. be achieved by use of the gas-20 generating particulate ma-terial alone.
In this example this is achieved by use o~ an a-alumlna ~aving a small mean particle size and a rela--tively wide particle size d:i.stribu-tion, whi.ch resul-ts in a signif:ican-tly lower flowabili.ty than that. of the a-a].u~,nirl.~s'usecl in the previous examples.
~ 'he maxi.mum central tensile 'stress produced ~.ras L10 MPa u.--,.ln~; a ~l~ixture o~ 40% ~-~alwnina~and 60% a--alumina hav.ing ~ f'lo~abili ty 0~ '1'0. This i.s only .

~ ,4t7~
- 39_ .
.
marginally higher than the central tensi.le s-tress of 39 MPa procluced when using the ~ alumi~.a alone~
Progressive addition o~ further ~-alu.mina in the mixtures rapidly reduces the fl.owability o~ the mix-. tures to such low valu.es such tha-t -the central tensile stress produced in the glass sheets is less th.an that produced by use of the ~-alumi~as alone~
e_ m e fluidised bed was constltuted. by a mixture of 9% by weight of zeolite, which i.s a porous, crystal~
line aluminosilicate having water adsorbed i.n its pores, and~9l% by weight of a-alumina.
The zeolite had the fol].owing charac-terist-ics:
mean particle size = 2Li ~m particle size distribution = 4 Plowabili.ty - 5 water content (weight loss at 800~C) - 20%
thermal capacity per unit vo].ume at minimum ~luidi~
sation = 0.8 ~J/m~K
The ~-alumina had -the following characteristics:--mean particle size = 37 ~lm particle size distribu-tion - 1.682 flowability = 70 thermal capacity per unit volume a-t minimum fluidi sa-tlon - 1.4 MJ/rn-3K
.-he thermal capaci-ty per uni.t volume at minimum . fluicli.sat;ion of the mi~-~u-f~ was 1.34 MJ/m3~ and the - *lowclbl.l.ity of the mixt;ur~-~ was 60.

. ..

- .- , ilL~gL47~
- ~ . L~o A. gl.~s,s shee-t 2.3 mm thick was hea-ted -to 660C and ~,,uenched in the flu,.dlsed mixture,.and produced a cen-tral tenSlle StJ:~e~;S of 41' MPa i.n the sheet. By varying the selectecl propor-tions of the constituent~ o:~ 1,he rnlxture a central tensile s-tress within -the rarlge 25 MPa to 41 MPa - could be induced,in the sheet.
Exa~e 9 ' The mixture o:f par-ti.culate materials for fluidisa-tion comprised 20% by weight o~ ~-alumir..~.a, and 40,h b~ weight of ~each of -two a-alurninas which were readi]y avai3abl.e and which were used in p:Lace of a single scarcer a~alumina.
The charac-teristics oE -the ~-alumi.na were as follows ineall par-ticle size ' = 57 - pa~-ti.c:le ~ize distribution - 1.66 Elol.~ability . = 85 wa-ter con-tent (weigh-t los,s . a-t ~300C) = 7%

thermal capaci-ty per unit volume a-t minimum flu:idi.sa-tion = 1.18 MJ/m3~
The ch.arac-teri.c;-l;ic.s o:E the two ~-alllminas, A and B, .were a.s in Table VI]:~
TABLE VII

. ' a-alumina A B

mean particle si.ze (~m) 38 24.
par tic3 e siz~: distrlbu-tion 1.1~ 1.25 L'lo~.~abi.'!.l1;y . ' 75 66 c~lerrna''. c.-lpa~;ity per uni-t voll~le .
a t 113:ir~ 'i.r~ L l~lidl.s~lt;ioll (MJ/m3K) 1~ 14 1.19 . .

:

The ~ rll~i.?ity o~ -the mixture was '73.5 and its thermal capac.ity per unit volurne a-t minimurn :Eluidisat.ion was 1.25 MJ/m~K.
A glass sheet 2.3 ~ thick was hea-ted to 660C and 5. quenched 1.n the fluidised r~lix-ture and the central -tensile stress induced in the sheet was 48 MPa~ By va:rying the selected relative proportions o~ the consti-tuen-ts of the mi.xture a selected cen-tral tensile s-tress in -the range ~4 MPa to 48 MPa could be induced in. the shee-t.
10 E~ le 10 ~ ~. ._ . ,.
The versa-tility of the me-thod o~ -the inven-tion is further lllustrated by tailoring a mix-ture by blending several gas-gencrating consti-tuen-ts and several inert cons-ti-tuen-ts, all o~ ~hich are avallable and. relatlvely 15 cheap mat*rial.s~ in order to produce a rnixture having gas-generating proper~ties, and an optimum flowability and thermai. capacity which produces required s-tresses :in t~le glass ~uenched in that mix-ture ~hen ln a quiesc.ent ~ ormly ex~)anded sta-te of particulate ~luidisation.
In this exclmple tlle mi.xture included 5% by weight 0~ eaC~l O.L four ~-al~wminas, A, B7 ~ and D whose charac-teristi.cs are set out .n Table VIII~

.~

` ~ ~ 4 ~'7~ Z

_ _~
. ~-alumina .~ ~_~ ~_ . A B C D
____ _ ___ _ _. ___ ___ mean particle size (~m) 70 61 57 72 .. _ _~ ~, ~_ ~.
particle size distribution ~,47 1.67 1.66 1.6~
_~ __ ~ _ flowabili-ty . 88.5 88 8586 __ _ __ ~ . ~_~ __ water con-tent (% welght loss .
at 800C) 7 7 7 7 , ___ _ ~ _ .
ther.mal capacity per unit volume at minimum fluidi- :
sation (MJ/m3K) 1.16 1.16 1012 1.l2 _ . . ............ _ ~ _ , _ This to-tal of 20% by weight of ~-alumina, wa,s mixed with 26~67% by weight of each of three a-alumin~s E, F
and G whose characteristics were as set out in Table IX.
TABLE IX
, a-al~mli.n.a __ ____ __,__.
. E F G
~ .. __ __ I ~' si Z G ( ~--) 3~ 30 24 paJ.~-ticle size distribùtion 1O19 1.22 1.25 _.~ _.~ _~. ~
. . ~lowability 75 70 66 ~ ~ _~_~ ,~_ thermal capacity per unit . .
volume a-t mi.ni,ml~ fluidisa-tion (~i,J/rl3K) . 1~3~ 1.3 101 ~_~.~.. ~_ ~w~ ~ ~

.

'7~ Z

The flowabili-ty of the mix-ture was 7l~ and i-ts thermal capacity per unit volume a-t minimurn :flui.d.isa-tjor ~as 1.26 MJ/m3K.
A glass sheet 2.3 rnm thick was heated -to 660C and quenched in the fluidised mix-ture and the cen-tIa~ -tensile stre~s induced in -the sheet was 49 MPa. B-y varying the selected proportions of thé ~ alurninas cons-ti-tuting 20%
by weight of the mixture~ or by varying the proportions of the a-alumina constituting 80% by weight of the mix-ture, or by varying the relati~e proport.ions o~ to-ta] ~-alurnina to -total a-alumina in the mix-ture a selected cen-tral tens'le stress in the range ~2 to 49 MPa could be induced.
E~ e 11 The fluidised particula-te material consisted of 17%
by weigh-t of the aluminium monohydrate of Example 6~
mi.xed wi-th 83% by weight of sllicon carbide ha~ing 1-,he following characteristics:-mean par~ticle size = I.~o ~m particle size distribution = 1.32 f`lowability = 72.75 thermal ca~acity per unit volume at m:i.nimum ~luidisa-tion - = 1.21 MJ/m3K
The flowability of the mix~ture was 75, and i-ts thermal CapaCit~J per unit ~olurne a~t minimum ~luidisation - was 1.02 MJ/rl!~K, ~ ~;lass sheet 2.7~ rQm thic~ was heated -to r.~60C and que.rlched in the fl.uidised mlx-t,ure, and the c~ntral -te~lslle s-~ress inducecl i.n the.shee-t ~as 51 MPa. These materials ; , .

~44'7~Z

- provlded the facili-ty f'or inducing a selected cen-tral tensi.le s~ress in a broad range9 32 MPa to 51 MP~, by selection of the predetermined proportions of the con-stituents of the mix-ture to tailor -the ~luidised mixture to the production of -the required stress ln the glassO
~ ' .
In a two~componen-t mixture, bo-t,h particulate mat-erials may ha~e gas~generating properties. A mixture was -made of equal proportions by weight of ~-alumina and alu-minium trihydrate (Al20~.3H20). A proportion of the wa'Ger of crys-tallisa-tion of the aluminium -trihydra-te ls released on heating~ whlch adds to the effec-t o~ the wa-t~r released from -the pores o~ the y-al~nina.
The characteristics o~ the ~-alumina were as follows mean par-ticle size = 60 y~
particle .size distribu-tion = 1.9 ; ' ~lowability = 84 water content (weight loss at 800C) = ~ :
thermal capacity per uni-t volume at minim~n ~lui~isation = 1~05 MJ/m~K
The characterist,ics o~ the alumini.um -trihydra-te were as follows mean partlcle size = 86,~m particle size d.istri.bution = 1.42 ~lowabilit~ = 86 water conJcent (weight loss at 800C) = 3L~%
thermal. capacity per UnlG ~o~..urne at m:inimurn flui.disation = 1.56 MJ/m3K

z The ~'lowabi,l.ity o~ -the mix-ture was ~5.25 and its thermal capacl-ty per unit volurne at minimum flui~isation was 1.31 MJ/rn3K.
' A glass shee-t 2.3 mm thick was hea-ted -to 660C and - 5 quenched in -the fluidised mixture~ and the central tenslle stress induced i.n the sheet was 47 MPa. A selected cen~
tral tensile s-tress in the range 42 MPa to 47 MPa could be - induced in the glass sheet by suitable selection of the relative proportions of the -two gas-generating ~aterials, Gas-generating particulate ma-terials which will 'evclve gas~s other than wa-ter vapour on hea-ting can be used, for example sodium bicarbonate (NaHC03) which re- , leases carbon dioxide as well as water. A mixture of 10%
15 by weight of sodium bicarbona-te containing 0.6% by weigh-t ' of colloidal silica, to improve i-ts flowability~ with 90%' by weigh-t of the a-alurnina A of ~xample 9~ was used.
The characteris-tics of the sodium bicarbonate/
colloidal silica mix-ture were as follows 20 , msan part.icL~ size = 70 ~m particle si~,e distribu-tion = 1~98 flowabili-ty = 75 H~O~C02 content ~weigh-t loss - ~ at 800C) - = 37%
therma~ capacit~ per ~lit volume 3 a-t m'nirllu,rn fluidisation = 1.41 M~/m K
The flowa~ility of the f~vidised rnixture was 75 9 and it.s thermal cap~city per ~it ~ol~ne at minimum ~luidisa-tion .~Jas 1.3~. MJ/m3K.

... .. .

.
.

-` ~144'762 A glas.s shee-t 2 3 mrn thick was hea-tecl to ~60C a~d -then quenched in the fluidised mixture, and -the central tensile s-tress induced in the sheet was 53.5 MPa. By suitable selection of the relative propor-tions o~ -the constituen-ts of the mixture a selec-ted cen-tral -tensile stress in the range 34 MPa to about 55 MPa could be induced in the glass sheet.
In many of -the examples an indi.cation o:E the stresses induced in the glass when quenched in the Plui-dised mixture of particulate materials, is given in -terms of the stresses induced in a 2.3 mm -thi.ck sheet of soda-llme-silica glass heated to 660C and then quenched. I.n the same way as described ln Examples 27 3 and 4 dif~erent s-tresses can be achieved by varying the temperature to which -the g].ass is hea-ted, and proportionate:ly higher s-tresses are incluced in thicker glass.
'~he examples aIl illustrate how the selection o.E a proportion of a gas~generating particulate rnat~rial which is capable o~ evolving ~rom 4% to 47% of i-ts own wei.ght of gas when heated to a cons-tant weight at 800C, and -then mixing -that gas--generating material in predetermined pro-por-tions with other gas-generating particula-te l~aterials or with inert materials, the mixture can be ~tailored to produce a desired ~lowabili~ty in the range 60 -to 86 and a -thermal capacity per uni-t volurne at rn.inimum ~luidisation in the range 1.02 to 1.75 MJ/m3K~ which ensures tha-t the glass sheet quenched in the mixture is thermally toughened to the requi.red degree, indicated in the examples by the cen-tral tensil2 s-tress. As is usual in -thermQlly .oug e d :,.

11~4'762 - - ~7-glass, -the ratio OL sur~ace compres5ive stress to central tensile stress is of the order o~ 2:1 and the surface compressive stress induced is usually about twice the quoted central tensile stress~
5 . . By employing the method of -the invention toughening conditions can be readily reproduced and it is possible to make use of a wide range of particulate materials, as available, and to employ blends of cheaper and more readily available particulate materials in place of 10 scarcer ancl more expensive single components o~ the mix ture, SG that operating costs are reduced.
Further, by appropriate selection of the particulate ma-terials and of the proportions in whic.h they are mixed7 it is possible to produce in glass selected higher 15 toughenlng stresses than the stresses which could be achieved by any of the constituents o~ the mixture used alone.
Some o.f the particulate ma-terials described above were co~nercia:Lly available with an appropriate mean 20 par-ticle siæe, particle size distribution, flowabili-ty, and thermal capacit~J.
I~hen these characteristics o:E the required material, e.g. ~ -alumina, were not present in commerciall~f a.va.ilable materials sieving is employed to produce refined pa.ticu--25 latc- ~aterials having required characteris~i.cs ~or mixing with o-ther constltuen^ts to produce a mi~-ture which~ when fluidi.sed, woul~l induce required -tougheni.ng stresses i.n the glass.

, , ~ " : : .

7~;~

. 48 . The fluidised mixtures of Exarnples 1 to 4 and 13 have been fou~d -to be par-tlcularly suitable for the thermal -toughening of glass sheets for lamination in the manufacture of automobile windscreens. The flowability of such mixtures is in the range 71 to 83, their gas content in terms of weight.loss when heated to constant weight a-t 800C is ln the range 4% to.37% and their thermal capacity per unit volume at minimu~ flllidisation is in the range 1.09 -to 1.38 MJ/m3K.
Selection of the proportions o~ the constituen-ts of mixture is possible -to produce lower stresses in the . glass than those produced by the gas-generating consti-tuent alone. This is illustrated by ~xample 7.

,.

Claims (28)

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:-

1. A method of thermally treating glass in which glass is heated to a predetermined temperature and is contacted with a gas-fluidised particulate material, characterised by employing a particulate material which comprises a mixture of a number of selected particulate materials, selecting at least one of said particulate materials to have gas-generating properties when heated by the hot glass, and mixing the materials in selected predetermined proportions which impart to the gas-fluidised mixture of particulate materials a thermal capacity and flowability such that a required thermal treatment of the glass is achieved.

2. A method according to claim 1, characterised by mixing said materials in selected predetermined proportions such that required toughening stresses are induced in the glass sheet as it cools in the gas-fluidised particulate material from a temperature above its strain point.

3. A method according to claim 2, characterised by selecting as gas-generating particulate material a material which is capable of evolving from 4% to 37%
of its own weight of gas when heated to a constant weight at 800°C, and mixing the particulate materials in predetermined proportions which impart to the mix ture a thermal capacity per unit volume at minimum fluidisation in the range from 1.02 to 1.73 MJ/m3K and a flowability in the range 60 to 86.

4. A method according to claim 3, of thermally toughening a sheet of soda-lime-silica glass of thickness in the range 2mm to 2.5mm, in which the glass sheet is heated to a temperature in the range 610°C to 680°C, and the mixture is maintained in a quiescent uniformly expanded state of particulate fluidisation, characterised by constituting the particulate material to engender in the glass sheet a central tensile stress in the range 35 MPa to 57 MPa.

5. A method according to claim 3 or claim 4, characterised by mixing said particulate materials in predetermined proportions which impart to the mixture a flowability in the range 71 to 83, and a thermal capacity per unit volume at minimum fluidisation in the range 1.09 MJ/m3K to 1.38 MJ/m3K.

6. A method of thermally treating glass in which glass is heated to a predetermined temperature and is contacted with a gas-fluidised particulate material which comprises a number of selected particulate materials one of which is .gamma.-alumina having gas-generating properties when heated by the hot glass, and mixing the materials in selected predetermined proportions which impart to the gas-fluidised mixture of particulate materials a thermal capacity and flowability such that a required thermal treatment of the glass is achieved.

7. A method according to claim 6, characterised in that the .gamma.-alumina is mixed with .alpha.-alumina.

8. A method according to claim 6, characterised in that the mixture comprises from 7% to 86% by weight of .gamma.-alumina and from 93% to 14% by weight of .alpha.-alumina.

9. A method according to claim 3, characterised by chilling a hot glass sheet in a gas-fluidised mixture of a gas-generating particulate material and at least one particulate metal oxide whose thermal capacity per unit volume at minimum fluidisation is in the range from 1.76 MJ/m3K to 2.01 MJ/m3K, and mixing the particulate materials in predetermined proportions which impart to the mixture a thermal capacity per unit volume at minimum fluidisation in the range 1.27 MJ/m3K to 1.76 MJ/m3K and a flowability in the range 71 to 82.

10. A method according to claim 9, characterised in that the particulate metal oxide is spheroidal iron oxide (.alpha.-Fe2O3).

11. A method according to claim 10, characterised in that the mixture comprises from 30% to 70% by weight of spheroidal iron oxide.

12. A method according to claim 11, characterised in that the mixture comprises from 70% to 30% by weight of .gamma.-alumina as gas-generating material.

13. A method according to claim 10, characterised in that the mixture comprises from 28% to 35% by weight of spheroidal iron oxide, and from 45% to 56% by weight of .alpha.-alumina, the remainder being .gamma.-alumina as gas-generating material.

14. A method according to claim 9, characterised in that the particulate metal oxide is zircon (ZrO2.SiO2).

15. A method according to claim 14, characterised in that the mixture comprises from 10% to 70% by weight of aluminium monohydrate (Al2O3.1H2O) as gas-generating material and from 90%
to 30% by weight of zircon.

16. A method of thermally treating glass comprising heating the glass to a predetermined temperature above its strain point, contacting the hot glass with a gas-fluidised particulate material which comprises a particulate alumino-silicate which has gas-generating properties when heated by hot glass and at least one inert material in selected predetermined proportions which are tailored to impart to the mixture a thermal capacity and a flowability such that required toughening stresses are induced in the glass as it cools in the gas-fluidised mixture.

17. A method according to claim 16, characterised in that the aluminosilicate is zeolite, and from 8% to 10% by weight of the zeolite is mixed with from 90% to 92% by weight of .alpha.-alumina to constitute the mixture.

18. A method according to claims 1 to 3, characterised in that the gas-generating particulate material is aluminium monohydrate (Al2O3.1H2O).

19. A method of thermally treating glass comprising heating the glass to a predetermined temperature above its strain point, contacting the hot glass with a gas-fluidised particulate material which comprises aluminium monohydrate (Al2O3.1H2O) and at least one inert particulate material, and mixing the aluminium monohydrate and the inert material in selected predetermined proportions which are tailored to impart to the mixture a thermal capacity and a flowability such that required toughening stresses are induced in the glass as it cools in the gas-fluidised mixture.

20. A method according to claim 19, characterised in that the mixture comprises 17% by weight of aluminium monohydrate mixed with 83% by weight of silicon carbide.

21. A method according to claims 1 to 3, characterised in that the gas-generating particulate material is aluminium trihydrate (Al2O3.3H2O).

22. A method according to claims 1 to 3, characterised in that the mixture comprises two gas-generating particulate materials, aluminium trihydrate (Al2O3.3H2O) and .gamma.-alumina in equal proportions.

23. A method of thermally treating glass comprising heating the glass to a predetermined temperature above its strain point, contacting the hot glass with a gas-fluidised particulate material which comprises a proportion of sodium bicarbonate (NaHCO3), and tailoring the proportions of sodium bicarbonate and at least one other particulate material in said mixture to impart to the mixture a thermal capacity and a flowability such that required toughening stresses are induced in the glass as it cools in the gas-fluidised mixture.

24. A method according to claim 23, characterised in that 10% by weight of sodium bicarbonate is mixed 90% by weight of .alpha.-alumina to constitute the mixture.

25. Apparatus for thermally treating glass by a method according to claim 1, comprising a container for gas-fluidised particulate material, gas supply means connected to the container for maintaining the state of fluidisation of the particulate material, and means for positioning hot glass in the container, characterised in that the container holds a gas-fluidised mixture of selected particulate materials at least one of which has gas-generating properties when heated by the hot glass, which particulate materials are mixed in selected predetermined proportions which impart to the gas-fluidised mixture a thermal capacity and flowability such that a required thermal treatment of the glass is achieved.

26 A method of thermally toughening glass comprising:
heating the glass to a temperature above its strain point;
chilling the hot glass by contact with a gas-fluidised particulate material which comprises a mixture in predetermined proportions of a number of selected particulate materials, at least one of which has gas-generating properties;
placing the gas-fluidised particulate material in a quiescent uniformly expanded state of particulate fluidisation by control of the distribution of fluidising gas in the particulate material at a gas flow velocity through the particulate material between that velocity corresponding to minimum fluidisation and that velocity corresponding to maximum expansion of the particulate material;
selecting the gas-generating particulate material from the group consisting of y-alumina, aluminium trihydrate, aluminium monohydrate, aluminosilicate and sodium bicarbonate; and mixing said selected particulate materials in predetermined proportions which are tailored to impart to the mixture a thermal capacity per unit volume at minimum fluidisation in the range 1.02 to 1.73 MJ/m3K and a flowability in the range 60 to 80.

27. The method of Claim 26, wherein the mixture includes a particulate material selected from the group consisting of .alpha.-alumina, zircon, silicon carbide, spheroidal iron oxide and mixtures thereof.
28. A method of thermally treating glass comprising:
heating the glass to a predetermined temperature;
contacting the hot glass with a gas-fluidised particulate material which comprises a mixture of a number of selected particulate materials, at least one of said particulate materials being capable of evolving from 4% to 37% of its own weight of gas when heated to a constant weight of 800°C;
placing said gas-fluidised particulate material in a quiescent uniformly expanded state of particulate fluidisation by control of the distribution of fluidising gas in the particulate material at a gas flow velocity through the particulate material between that velocity corresponding to minimum fluidisation

Claim 28 - cont'd ,..
and that velocity corresponding to maximum expansion of the particulate material; and mixing said selected materials in selected predetermined proportions which are tailored to impart to the mixture of particulate materials a thermal capacity, determined as the thermal capacity per unit volume at minimum fluidisation in the range 1.02 to 1.73 MJ/m3K, and a flowability in the range 60 to 86 such that a required thermal treatment of the glass by the gas-fluidised mixture is achieved whereby the tailored mixture imparts higher stresses than the stresses which could be achieved by any of the constituents of the tailored mixture used alone.