Εικόνες σελίδας
PDF
Ηλεκτρ. έκδοση

vapor. When the same or similar fibers were dried in CaCl2 the strength increased about one and a half to two times the strength in water vapor [190, 192, 193]. When water or alcohol vapor was admitted into the evacuated apparatus the strength immediately decreased to its original values. For fibers of a given diameter the strength in oil, and in alcohol is greater than that in air. In addition, the strength of those fibers which were in the air for a long time was the same as that of fibers broken in water, and the strength of fibers stored in a room saturated with oil or alcohol vapor was the same as that of fibers broken in the liquids.

Further experiments [188 to 193] on the effects of etching in hydrofluoric acid revealed a 3 to 5 fold increase in strength of silica fibers after the etching. There was, however, a considerable amount of scattering of the values which was partially explained [193] as a result of the etching process which alternately smoothes and then exposes flaws; the strength thus depending on the condition of the surface at the time of removal from the acid. The strength of fibers broken in the acid did not differ much from that of fibers broken in the air. Furthermore, the strength gradually increased with etching until a certain maximum thickness had been removed, after which no further increase was noticeable [190]. These etched fibers were very susceptible to damage from the air or from scratching although a protective coating of shellac helped them retain their high strength [193].

The strength of silica fibers is decreased by scratches from atmosphere dust [199, 200] and from other fibers. There is an immediate loss of strength resulting in fracture when a fiber under strain is touched by another silica fiber. A further decrease of strength results from contact with the fingers [213, 315], although those parts rubbed with the fingers are protected from further injury [113]. It has been suggested occasionally that silica fibers be coated to prevent damage from the various atmospheric conditions or to increase the strength. There is at present no published data available on the effects of coatings on silica fiHowever, investigations [195, 227] of

the effects of coatings on textile glass fibers revealed that while the coating itself did not increase the strength it did lesse the effects of atmospheric attack in some i stances. The chief use of a resin or wax coating is to prevent seizing of glass fiber when they are woven or pressed. Silica fibers are much more resistant to damage than commercial glass fibers and in present app cations careful handling will avoid much harmful damage. A protective coating may prove useful however.

The tensile strength of silica fibers it creases slightly with increasing temperature up to the flow region, showing minimum strength at room temperature [108, 116]. This can be explained by the fact that the healing of the cracks by surface diffusion a the temperature is raised outweighs the influence of thermal motion which tends to increase tension and to spread the cracks [213]. In addition, the absorbed moisture layer which is thick and active at room tem perature is driven off at higher temperatures and is inactive at low temperatures [213]. Experiments [10] showed an increase in strength of silica fibers after they were heated to 1, 188°C for 4 hours and then allowed to cool. However, heating silica fibers in a flame makes them weaken considerably [2]. Although the strength may change with temperature changes, the strength at relatively low temperatures does not depend on the temperature [193].

The values of strength thus vary depending on the conditions of test and of use. Any improvement in these conditions should increase the apparent strength. A strength test gives not the true value of strength but rather the residual value after the fiber has been damaged in some way [114].

4.2 Elastic Properties

Silica fibers exhibit almost perfect elasticity up to the breaking point; any de viations are so slight as to be barely meas urable.

Although the elastic constants have been determined since silica fibers were first used [1], there is not complete agreement be tween early values and those determined

ther recently. Again as with strength, sting methods and conditions of the test d of the specimens considerably affect the nal results with the errors either magnied or diminished. Absolute determinations elastic constants require accurate values size, density, and mass [244]. Temperare effects must also be considered if invidual measurements made under different nditions are to be related or compared. elastic effects which cause the strain to ig behind the stress, must be accounted for nce the elastic constants are related harniously only if the rate of stress appliation is the same [224].

is.

Within the experimental error, the elastic >duli can be considered to remain fairly conant over the range of sizes used. The expernents of some investigators [1, 2, 5, 9, 188 > 191] show a dependence on the fiber diameer of Young's modulus and of the shear moduHowever, later investigations showed hat these moduli were independent of the dineter and of the length, and pointed to light errors in former measurements [11, 193]. A summary and discussion of much of the xperimental work on the elastic constants of used-silica fibers can be found in Sosman's reatise on silica [7]. Some of the experients [1, 2, 182, 185], and later work [184, 87, 188 to 191, 244] show that Young's moduus and the shear modulus do not vary appreiably over the range of temperatures used in he tests. There is a slight linear increase In both of these constants with increasing emperature up to the region of viscous flow. The moduli continue to decrease from values measured at room temperature to those measired down to -200°C [245]. The change in shear modulus with change in temperature affects the readings of very sensitive measiring instruments [62, 63, 107]. For example, Sheft and Fried [62] found an increase in rigidity, in a spring balance, of about 0.02 percent per degree centigrade which was reflected in a change in elongation of 0.02 percent of the total load per degree centigrade. The change seemed to indicate strengthening with increasing temperature. The elastic moduli vary with the surrounding atmosphere, exhibiting higher values

ture.

in a vacuum and in alcohol vapor than in water or moist air [192], which increase could not be due to the weight of the moisThis effect was also observed in silica springs where an increase in elongation occurred in water or alcohol vapor but not in inorganic vapors when these springs were exposed to the vapors after being in a vacuum [196]. This was, however, erroneously attributed to an expansion of the silica because of absorption. Silica fibers etched in hydrofluoric acid had values of elastic moduli in close agreement with unetched fibers with some uncertainty regarding an actual increase [192].

Slight deviations from "perfect elasticity" of silica fibers can be attributed to a slight delayed elastic effect noted when fibers are twisted through large angles. The classification of the types of deformation due to applied load generally used are: an instantaneous strain which is completely recoverable, a delayed elastic strain which is recovered slowly, and a viscous flow which is not recoverable but which appears to be nonexistent in fused silica at temperatures below about 800°C [7, 117]. The delayed elastic effect is similar in effect to a progressive increase in viscosity and frequently is confused with a viscous flow or creep [117]. Although delayed elastic effects are of some concern in commercial glass fibers, the effects in fused-silica fibers are rather negligible, approximately 100 times less than in other glass fibers [217, 218]. A delayed elastic effect has been noticed in instruments using a torsion fiber. However, a constant correction could be applied, as after a short time the rate became constant. In addition the effect decreased with decreasing fiber diameter so that in the finest fibers the effect is barely perceptible [1, 66]. A part of the apparent delayed elastic effect could also be attributed to the mountings used for the fibers [24, 53]. With increasing temperature the delayed elastic effect, expressed as the ratio of the delayed strain to the instantaneous strain, increases. Fused silica has the lowest value of the ratio of all the glasses [117].

5. SOME PROPERTIES OF FUSED SILICA

The properties of fused silica observed at room temperature combine to make it such a useful material. While some properties may change with time, temperature, or surrounding atmosphere the magnitude of the changes is usually small compared with similar changes in other materials. The properties discussed in this section usually are determined for fused silica in bulk form, but are of some importance in the use of silica fibers.

Since fused silica is a glass, the most simple glass, it would seem appropriate to place glass in reference to other states of matter. The ASTM describes glass [301] simply as "an inorganic product of fusion which has cooled to a rigid condition without crystallizing." The vitreous or glassy condition is rather difficult to define as is evidenced by the numerous papers on the subject some of which attempt to describe the condition of glass as distinct from or similar to the solid and liquid states [124 to 164].

5.1 Structure

The present picture of the atomic arrangement of fused silica developed from the laws of crystal chemistry, analysis of X-ray diffraction patterns, and studies of certain physical properties. It was from studies on crystal structure that Goldschmidt [126], Zachariasen [126], Sosman [7], and others [124, 151, 155] developed theories on the atomic arrangement and formation of glasses. The early methods and the apparent similarity between crystal and glass led some observers [124, 151] to believe that fused silica and other glasses were made up of particles of crystalline material called "crystallites." However, even without the analytical tools later used by Warren and his co-workers [127 to 134], Zachariasen postulated the existence of a random three-dimensional network with energy comparable to that of the corresponding crystalline network in fused silica and glasses in general and laid down certain conditions as necessary for the formation of oxide glasses. The network picture thus

postulated by Zachariasen was later verified by Warren and his co-workers. These experimenters using a Fourier analysis of the X-ray diffraction pattern, first of fused silica and later of other glasses, were able to show not only the arrangement but also the average interionic distances and the mean bond angles. They also showed that any particles in fused silica showing similarity to crystals are too small to be described as crystalline

matter.

The structural arrangement of fused silica as confirmed by Warren's analysis consists of a short order structure which is tetrahedral in form with 1 silicon ion bonded to 4 oxygen ions, with each oxygen ion bonded in turn to 2 silicon ions. The tetrahedra share the oxygen corners to form a long range three-dimensional random network. A twodimensional picture of this structure shows the network made up of a series of irregular rings, where the average number [154] of tetrahedra per ring is six and the number of tetrahedra in individual rings varies from 3 to 10 or more. Where the bond angle between tetrahedra is nearly 180° and varies slightly for successive tetrahedra, a buildup of a random or "disordered" [143] network is allowed. It is this flexibility in bond angles which gives fused silica and other glasses the long-range disorder of a liquid; while the ordered distances and angles within a tetrahedron gives them the short-range order of a crystal.

5.2 Chemical Durability

At room temperature, fused silica is attacked by hydrofluoric acid. Near 300 to 400°C, phosphoric acid starts to attack silica. At high temperatures and up to 1,000°C, weak and moderately concentrated solutions of basic salts and of metallic oxides and basic salts react with fused silica. Fused silica is reduced to silicon by carbon at temperatures over 1,600°C [110]. It is also reduced at high temperatures by hydrogen which forms. silicon hydride. When this material is

cooled the part of the silicon hydride not oxidized decomposes into silicon and hydrogen and forms a deposit of silica and silicon on the surface of the material [281]. * This deposit, observed on fibers fused in a flame containing hydrogen, has frequently been confused with devitrification of the fiber.

[ocr errors]

5.3 Permeability

Fused silica is permeable to neon, hydrogen, and helium at elevated temperatures and under pressure [7, 10, 109, 112, 296]. Devitrified and nontransparent types of fused silica are more permeable than the transparent form. The rate of diffusion may depend to some extent on the size of the openings in the random network [116, 296] and on the size of the gas atoms. Thus fused silica with larger random openings than other glasses would allow larger gas molecules to pass through and the smaller gas molecules could pass through more quickly than in some other glasses [296]. Reviews of work on permeability indicate considerable differences in observations on the amounts of diffusion through fused silica [4, 109, 112].

5.4 Sorption

Fused silica exhibits a combination of a reversible adsorption plus a slow permanent 3 sorption of water [285]. This reversible adsorption was reduced about 35 percent by acid treatment, and an additional 30 percent by heat treatment. The rate of permanent sorption as found by Barrett [285] in finely powdered transparent fused silica, fell off rapidly in about a half hour after exposure to moisture and then became constant at about 4x10-10 g/cm2/hr. This process seemed to be an actual diffusion of water into the silica. Fused silica apparently does not adsorb inorganic vapors in sufficient amounts to be measurable [196]. It does however adsorb al{ cohol vapor [192, 193] and the vapor of paraffin oil [175] though in smaller amounts than the water vapor. This adsorption of moisture affects the strength and the elastic constants such as Young's modulus and the shear modulus which in turn affect the

behavior of measuring instruments [62, 106, 196].

5.5 Hardness

The "hardness" of fused silica is not only a function of the strength and the elastic properties but must be greatly influenced by the chemical resistivity of the glass and by the environment [116]. The hardness of fused silica has been determined on the various hardness scales although an exact definition has not been formulated. The Mohs' hardness numbers given for fused silica range from 5 to 7 [9, 10, 312]. The Knoop indentation hardness number given for fused silica is approximately 475 [284] where the value for quartz is approximately 710 to 790 [302]. This Knoop indentation scale shows fused silica to be harder than some other glasses and many metals including brass, gold, silver, tantalum, and some stainless steels.

5.6 Density

The density of transparent fused silica is approximately 2. 21 g/cm3 while that of nontransparent fused silica is about 2.07 g/cm3. The density of the specimens of fused silica may vary about 0.05 percent or even as much as 0.1 percent, a variation of about 10 times that of specimens of crystalline quartz [7]. Sosman thought that it might be possible to correlate these variations with the source, with the thermal treatment or with the state of strain in the fused silica. Since that time these factors have been related to density changes in a number of experiments [170, 289, 292].

Density measurements at temperatures up to 1,700°C showed that there is an equilibrium density which increases with temperature [289]. The specimens used changed from the initial density to the equilibrium value for the temperature at which the specimen was heated. The rate of change increased as the temperature increased so that at 1,300°C they had a density corresponding to the high-temperature density. These changes in density with heat treatment were found to occur in addition to the normal thermal expansion, in

effect a slow change of volume in addition to an instantaneous change in volume.

Under uniaxial pressure an increase in density was noted [289, 292]. Bridgman and Simon [292] found a threshold pressure near 100 atm above which the structure of fused silica seemed to collapse. Rapid increases in density of 7.5 percent and occasionally at high as 17.5 percent, to 2.61 g/cm3, were found. After X-ray examination showed the specimens to be amorphous, the increase was attributed to a folding up of the network; a bending of the Si-O bonds rather than a shortening of them. This folded-up structure was mechanically stable at ordinary temperatures, but with heat treatment the density decreased to the original value. Both Douglas [289] and Bridgman and Simon concluded that there appears to be an equilibrium configuration which may be approached from either direction. Their experiments support Sosman's theory that heat treatment and strain could cause variations in individual specimens.

5.7 Devitrification

The rate of crystallization or the devitrification of fused silica depends on the temperature to which it is heated, the period of heating and on the degree of subdivision. It is also affected by atmospheric dust [251], by the previous thermal treatment [7], and may be related to the viscosity [248].

While fused silica is thermodynamically unstable at all temperatures below 1, 710°C, its molecular sluggishness (viscosity) at room temperature is so great that no change toward crystallization has been observed at such temperatures [7]. Fused silica devitrifies only after prolonged heating at high temperatures. At 1,000°C [7] or even at 1,200°C [10] devitrification is hardly perceptible. Above these temperatures fused silica slowly devitrifies into cristobalite. When samples were powdered, almost complete devitrification occurred after heating at 1,100°C for 6 days or at 1,600°C for 1 hour [83].

Although devitrification could not be called a purely surface phenomenon [112], it does start at a surface and work inward [7,

108]. Specimens with a great deal of surface area such as powdered specimens, or those with a large number of bubbles such as translucent types devitrify more readily than large pieces and transparent fused silica [7, 108, 183, 249].

A possible form of devitrification occurred when fused-silica articles were exposed to radium salts at the temperature of boiling water but did not occur at lower temperatures [246].

A dependence of the rate or of the amount of devitrification on previous thermal treatment is evidenced by the appearance of less devitrification on specimens made from high temperature melts than on those made from lower temperature melts [7], as well as greater tendency toward devitrification exhibited by annealed specimens.

a

The presence of fluxes such as calcium carbonate, calcium oxide, liquid silicates, and certain metallic oxides hastens devitrification [253]. Although titanium oxide and zirconium oxide have been added to fused silica to decrease any tendency toward devi trification, the resulting glasses devitrify more rapidly than pure fused silica [7].

5.8 Thermal Expansion

Investigators have long been interested in the thermal expansion of fused silica, which is smaller than that of almost all other materials, because of its usefulness in measuring instruments. Critical examinations of the experimental work on expansion of fused silica presented by Sosman [7] and by Souder and Hidnert [260] show that differences in testing equipment and in methods of testing are responsible for many variations observed in the results of many investigators. With transparent and nontransparent fused silica tested over a range of -125° to 1,000°C, Souder and Hidnert determined a critical temperature of about -80°C where specimens had a minimum length. Viscous flow which started around 800°C hindered their measurements of expansion above that tempera

ture.

The expansion of fused silica changes slightly with heat treatment. Heat treatment over the range 20° to 750°C increased the ex

« ΠροηγούμενηΣυνέχεια »