ý,-AUG 24 1960 AD-A286 620

SFused-Quartz Fibers

A Survey of Properties, Applications, and Production Methods

! - : •

M~IR

~~3TIC ELECTE

A~ IUG 161994 DLSG

UNITED STATES DEPARTMENT OF COMMERCE

NATIONAL BUREAU OF STANDARDS

94-25826 D-rC . "A.

94 8 P 024

BestAvai~lable

Copy

* PERIODICALS OF THE NATIONAL BUREAU OF STANDARDS

(Published monthly)

The National Bureau of Standtards is engaged in fundamental and applied research in physics, chemistry,mathematics, and engineering. Projects are conducted in fifteen fields: electricit y and electronics, opticsand metrology, heat and power, atomic and radiation physics, chemistry, mechanics, organic and fibrousmaterials, metallurgy, mineral products, building technology, applied mathematics, data processing sys-tems, cryogenic engineering, radio propagation, and radio standards. The Bureau has custody of thenational standards of measurement and conducts research leading to the improvement of scientific andengineering standards and of techniques and methods of measurement. Testing methods and instrumentsare developed; physical constants and properties of materials are determined; and technical processes areinvestigated.

Journal of Research

The Journal presents research papers by aldliorities in the specialized fields of physics, mathematics,chemistry, and engineering. Complete details of the work are presented, including laboratory data,experimental procedures, and theoretical and mathematical analyses. Annual subscription: domestic,$4.00; $1.25 additional for foreign mailing.

Technical News Bulletin

Summaries of current research at the National Bureau of Standards are published each month in theTechnical News Bulletin. The articles are brief, with emphasis on the results of research, chosen on tilebasis of their scientific or technologic importance. Lists of all Bureau publications during the precedingmonth are given, including Research Papers, Handbooks, Applied Mathematics Series, Building Mate-rials and Structures Reports, Miscellaneous Publications, and Circulars. Each issue contains 12 or moretwo-column pages; illustrated. Annual subscription: domestic, $1.00; 35 cents additional for foreignmailing.

Basic Radio Propagation Predictions

The Predictions provide the information necessary for calculating the best frequencies for communica-tion between any two points in the world at any time during the given month. The data are important toall users of long-range radio communications and navigation, including broadcasting, airline, steamship,and wireless services, as well as to investigators of radio propagation and ionosphere. Each issue, cover-ing a period of one month, is released three months in advance and contains 16 large pages, includingpertinent charts, drawings, and tables. Annual subscription: domestic, $1.00; 25 cents additional forforeign mailing.

CATALOGS OF NBS PUBLICATIONS

National Bureau of Standards Circular 460 and its Supplement lists all Bureau publications from 1901through June 1952, including Applied Mathematics Series, Building Materials and Structures Reports,Circulars, Handbooks, Research Papers, and Miscellaneous Publications. Brief abstracts for thepublications issued after January 1, 1942, are also included.

National Bureau of Standards Circular 460, 375 pages, $1.25. Supplement to Circular 460, 223 pages,75 cents. (A free mimeographed list of publications issued since June 1952 is available on request tothe National Bureau of Standards.)

()ralýu atnfpf tS it " ot 't

K]

UNITED STATES DEPARTMENT OF COMMERCE • Sinclair Weeks, Secretary

NATIONAL BUREAU OF STANDARDS * A. V. Astin, Director

Fused-Quartz Fibers

A Survey of Properties, Applications, and Production Methods

Nancy J. Tighe

Accesion For

NTIS CRA&IDTIC TABU. :arýnourncedJustificationJustf:,c tion ......... .................... .

B y ............. .....................Di:t ib ,tio -I

"Availability Codes

OF Availado

National Bureau of Standards Circular 569Issued January 25, 1956

Contents

1. Introduction ......................... i 5. Some properties of fused silica-Con, Page

2. Applications of fused-silica fibers ........ 2 5.4. Sorrtion ............................. 153. Production and fabrication methods ......... 4 5.5. Hardness ............................. 15

3.1. Fused silica .......................... 4 5.6.- Density .............................. 15

3.2. Fused-silica fibers ................... 4 j.7. Prvitrificaticn ....................... 16

3.3. Silica-fiber apparatus ................ 6 5.8. Thermal expansin a..................... 16

4. Mechanical properties of fused-silica 5.9. Viscosity ............................ 17fibers .................................... 9 6. Bibliography ............................... 18

4.1. Strength .............................. 9 6.1. Books and ýeviews ..................... 184.2. Elastic properties .................... 12 6.2. Applications of silica fibers ......... 18

5. Some properties of fused silica ............ 14 6.3. Production an'ý fabrication methods.... 205.1. Structure ............................. 14 6.4. Properties of fused silica and glass.. 205.2. Chemical durability ................... 14 6.5. Miscellaneous referesc-2. .............. 25

5.3. Permeability .......................... 15 7. Addendum ................................... 26

(II)

V

FUSED-QUARTZ FIBERS

A Survey of Properties, Applications, and Production MethodsNancy J. Tighe

Fused-silica fibers have an important function in many measuring instruments usedin scientific research. Much of the information on the production and fabricationmethods and on the properties of the fibers is widely scattered throughout the techni-cal literature. This Circular is a survey of this literature and a summary of thefindings. A bibliography of pertinent references on the subject is included to pro-vide the sources of more complete and detailed information necessary for specificapplications.

1. INTRODUCTIONThis Circular is the result of a litera- of the terms has been discussed frequently

ture survey covering the properties and the [7, 112, 301, 303 to305]1 but little uniformityuses of fused-silica fibers. Silica fibers has resulted. The terms used are meant toor, as they are frequently called, quartz indicate the degree of transparency or thefibers, have had wide application in preci- source of the raw material. For instance, insion measuring instruments despite the fact the glass industry the term fused silica maythat the factors which affect their behavior refer to the transparent material formed byare not fully known. The survey of the lit- the fusion of quartz sand or other forms oferature revealed that while some investiga- silica, while the term fused quartz may refer

tions are made in order to obtain specific to the transparent material formed by thevalues of particular properties, other fusion of pure quartz rock crystal. However,studies are made primarily to develop more this designation of terms is not used con-

completely the theoretical treatment of sistently, for the material is frequentlyglasses and their relation to liquids and described as clear or translucent fused

solids. Although this Circular covers prima- quartz or as clear or translucent fusedrily those properties related to the actual silica. In this Circular the term fuseduse of the fibers, it mentions those theories silica will be used in general for that glassand properties needed for better understand- formed by the fusion of silica. Where neces-

ing of the variations observed in the testing sary, this term is modified by the adjectivesand use of silica fibers. It is hoped that clear or translucent. For instance, wherethis summary will aid in the production of the condition of transparency is evident fromsilica fibers having more reproducible and the use, such as in fibers which are drawn

predictable properties. from clear fused silica, the fibers are re-Throughout the literature such terms as ferred to simply as silica fibers.

fused silica, fused quartz, vitreous silica,silica glass, quartz glass are used to refer 1 Figures in brackets refer to the references listed into types of fused silica. The use or misuse the bibliography.

1

2. APPLICATIONS OF FUSED-SILICA FIBERSAlthough the first fused-silica fibers made here of the major developments in fiber

were made by Gaudin in 1839 [64] and some microbalances. Reviews of silica fibertubes and spirals were made and exhibited by microbalance developments [12 to 18, 503 andGautier in 1869 [67], these articles were papers on specific balances should be con-more curiosities than useful tools. It was suited for complete information about theo-not until about 1887 when Boys [1] drew fi- ries and principles of operation and design.bers and used them in his gravimetric work The earliest microbalance using silica[65] that their usefulness was exploited, fibers was that designed by Salvoni [19].

Today silica fibers in the form of paper, This type of balance [20 to 24] relies on thematting, cr wool are used in sizable quanti- rigidity of a cantilever fiber to indicateties in heating and insulating devices [77, displacement due to load. Shortly afterward80, 313]. These devices are well suited for Nernst [25] introduced a microbalance con-high temperature work because of the high- sisting of a horizontal silica fiber attachedsoftening range, low-heat conductivity, and to and supporting a fine glass rod with aresistance to thermal shock exhibited by the counterpoised pointer to indicate deflectionfibers. however it is in the usage of single due to load. This type of balance, whichfibers that the unique properties such as uses a silica fiber as a knife edge, is stillhigh strength, elasticity, thermal, electri- in use although in forms slightly modifiedcal, and chemical resistance are most fully from the original [26 to 36]. Steele andutilized. It is with these individual fibers Grant [37, 38] were the first to construct anand their adaptability to precision measuring arm balance entirely of fused silica. Theinstruments that this review is chiefly con- balance, made from fine fused silica rodscerned. with fiber load suspensions and a fuse.-sil-

Individual fibers, with or without metal- ica knife edge, utilizes a method of weighinglic coatings, have been used as suspensions, by pressure. Several modifications in theas sensing elements and as unit assemblies in design or in the operation of the Steele-many precision instruments. Silica fibers Grant balance have been made for adaptationhave been used as galvanometer suspensions to particular uses [39 to 42, 46, 48]. Pet-[1, 65], in gravimetric balances [65, 66], in tersson [43 to 45] introduced suspension ofelectroscopes and electrometers [9, 70, 74], the arm balance by two fine vertical fibersin low pressure manometers [78], in radia- as a replacement for the conventional knifemeters [75], in ionization chambers [79], and edge; and developed further the theory ofin magnetometers [73]. They may possibly be such suspension. Neher [9] designed a micro-of some use as electrodes in electron tubes. balance which consisted of a silica-fiber

One of the more extensive uses of silica crossarm attached to a horizontal silica tor-fibers is in the microbalance field. The sion fiber, the twist of which is propor-entire weighing system can be made of fibers, tional to the load. This balance can bethus avoiding any errors due to differing brought to an equilibrium position by rotat-densities and coefficients of expansion of ing the torsion fiber with a wheel calibratedthe members. Silica fibers are particularly to indicate the amount of twist in massdesirable for these instruments because of units.their high strength in tension and in tor- From this point microbalances using sili-sion, nearly perfect elasticity, negligible ca fibers were usually designed by combininghysteresis. Further, the resistance to chem- certain principles developed in the earlierical attack, low rate of sorption, low per- fiber balances. The balance of Kirk andmeability, and ease of cleaning allow such Craig [23, 47, 491 and modifications bybalances to be used under many difficult Carmichael [53], EI-Badry and Wilson [50],weighing conditions. Only brief mention is and Garner [54, 310] consist of a nominally

2

equal-arm beam with fiber pan suspension anrd McBain-Bakr balance [551. The silica springuse a torsion fiber to determine weight dii- balance can be used for sorption measurements,ferences. These balances differ slightly in density determinations, measurements of heatdesign and operation, the greatest design loss and evaporation. Discussions of designdifference being in that of Garner which has characteristics [1071 and application to par-a single vertical suspension fiber in addi- ticular measurements can be found in the ref-tion to the torsion fiber. erences cited. The springs resist corrosion

The simplest and apparently most widely and can be easily cleaned but have very littleused fiber balance is the helical spring type damping due to internal friction.[55 to 63], frequently referred to as a

3

3. PRODUCTION AND FABRICATION METHODS

3.1 Fused Silica ite are about the same below about 1,500*C.Above this temperature the formation of glass

One of the greatest drawbacks in the corn- is able to keep ahead of the crystallization

mercial use of fused-silica fibers has been so that as the temperature is raised rapidly

the difficulty and expense of preparing the to above about 1,700*C there is no tendency

bulk product from which the fibers are drawn, to crystallize [7,109]. The fused product at

Prior to World War II no fused silica of good this point is a soft plastic mass becoming

optical quality was produced commercially in fluid only after the temperature is raised to

the United States [87]. However since that the range 2,0000 to 2,500"C rapidly enough to

time many of the technologic difficulties prevent too much loss due to volatilization

have been overcome and fused-silica rods are [109]. The development and refinement as

available in many grades and sizes, well as the difficulties of fusion techniques

The two forms of fused silica, transpar- for fused silica are described in several re-ent and nontransparent, are products of view articles [4, 7, 10, 85 to 88, 109] and

quartz rock crystal and of quartz sand, re- in the original horks of Shenstone [3], Paget

spectively. Sand contains occluded gases and [6], and Hutton [82].

impurities which are difficult to remove en- Silica can be fused in arc furnaces, re-

tirely in the cleaning process or in the fu- sistor furnaces, graphite molds, and by di-

sion process with the result that the fused rected flame. Each of the methods has cer-

product contains volatile and nonvolatile im- tain difficulties including those caused by

purities of approximately 0.06 and 0.1 per- losses from volatilization and by reduction

cent, respectively [8]. These impurities are in the presence of carbon and hydrogen. With

revealed in the masses of tiny bubbles which carbon or graphite electrodes or molds, car-

make the product nontransparent in varying bon monoxide is formed and silicon carbide

degrees. On the other hand quartz rock crys- and elemental carbon vapor or silicon metal

tal can be selected with few if any internal may be formed [881. A purer product is ob-

impurities and can be more effectively tained if the fusion is done in an oxidizing

cleaned so that the occluded gases and other atmosphere which negates the effects of the

impurities on the surface are removed [87]. reducing tendencies of carbon. There is no

The resulting product is relatively free of tendency for quartz or fused silica to adhere

bubbles and is highly transparent. Quartz- to the carbon if both are pure, for the ox-

ites, natural deposits of quartz and vein ides of carbon prevent close contact between

quartz, which contain more impurities than the molten silica and the carbon electrodes

rock crystal, are poor substitutes for rock or molds [82,88]. However, particles from

crystal; they do however make a better prod- the electrodes, molds or from the walls of

uct than quartz sand for the nontransparent the furnace deposited in the melt cause in-

form of fused silica [8]. A more recent de- fusible hard zones which impair the qualityvelopment is the production of fused silica of the final product. Further refinements of

from "noncrystalline" materials [293,314]. the fusion techniques may eliminate this type

The production and forming methods for of impurity.

fused silica in no way resemble those for

commercial glasses [87]. This is because of

the high temperatures needed for complete 3.2 Fused-Silica Fibersfusion of the highly viscous melt, compli-

cated by volatilization. Although the melt- With the developments in the fusion and

ing point of quartz is probably below 1,470 0 C, forming techniques for fused silica, some of

the rate of fusion of quartz and the rate of the difficulties of fiber production have

change of the resulting glass into cristobal- been overcome. Early workers with fused-sil-

4

ica fibers [1 to 3, 92] were hampered by the the two ends drawn rapidly apart to produce a

lack of commercially available stocks of rather heavy fiber; or by the gravity methodtransparent rods from which fibers could be [94] in which the stock, weighted at on- end,drawn. By fusing small pieces of fused sil- is heated at a point until melted sufficiently

ica into the form of a rod and working this so that the falling weight draws out a fiber.rod until it was fairly smooth, these men ob- Boys [1] and Threlfall [21 each developed atained , rod from which fine sticks and then simple mechanical method for drawing fibers.

fibers .juld be drawn. This was a tedious Boys used a crossbow with a straw arrow whichand time-consuming process and limited the when released drew from the fiber stock aquantity and quality of fibers which could be fine fiber as long as 60 feet. Threlfall

drawn in a reasonable time. Threlfall [2] used a modified catapulL the slider of whichmentioned spending 14 days to get two fibers when released drew short thick fibers.

about 2.5 microns in diameter by about 13 Machines for continuous drawing of silicainches long. fibers generally have three elements: a

The production of silica fibers may be chuck to hold the silica stock stationary, toclassified by two major methods. One method rotate it about its axis, or to move it

includes those processes used for limited lengthwise into the flame; a holder to main-production of single fibers. Namely, blowing tain the torch at the proper angle and eitherof fibers with a flame and the drawing of fi- to hold it steady or to move it along the fi-bers by hand, and by simple mechanical means ber stock; a rotating reel to draw the fiberwhereby the length is limited or the amount from the molten stock at a given rate [11,

drawn is small. The other method includes 102 to 104, 1073. These elements can bethose processes in which automatic or semi- geared to be regulated independently or as a

automatic machines are used for continuous unit in order to draw a desired size of fiberproduction of single fibers, under readily reproduced conditions. A con-

The process of blowing fibers in a flame venient base for the elements is a lathe or a

depends on the action of friction of the drill press used either horizontally or ver-gases used. Two slightly different methods tically. This type of machine has been usedare used, one [9, 1951 producing a long fine to draw fibers from 600 to 0.7 microns from

fiber, the other [2, 92] a mass of short fi- stock of 12 to 0.15 millimeters in diameter.hers. In the first process the stock, a fi- Once the flame is adjusted to give sufficientber about 25 microns in diameter, is held heat for thorough melting of a particularvertically in a long flame until a finer fi- size of stock then che reel speed and the rateber is blown out. The size of the fiber thus of feed of the stock into the flame can beblown depends on the size of the original related to the approximate diameter of thestock, the temperature and size of the tiame fiber [11, 103]. The ratio of fiber pullingand the time interval between starting the to fiber melting can be regulated within lim-fiber and its removal from the flame. The its above which the fiber breaks off and be-

fibers .re smooth in appearance but have a low which the stock is not sufficientlyslight curl which decreases the strength melted for uniform drawing [11]. Drive slip-

[195]. In the second method the fiber stock page and speed recovery accompanied by alter-is first drawn apart in the flame and the two nate drawing of cold and hot silica into thepieces stroked back and forth in the flame, fiber results in variations in strength andwhich is placed in a horizontal postion. The in the diameter of the fiber [103].

fibers are blown out horizontally and caught Many gas mixtures have been used in all

on a cloth or board placed a few feet from of the above methods: oxy-hydrogen, oxy-coalthe flame. These fibers are deposited in a gas, oxy-natural gas, oxy-acetylene. It has

tangled mass and must be eased apart. They not been mentioned whether the gases used af-are fairly short and may be damaged by super- fect the fibers, although the heat of theficial scratches from the contacts. Fibers flame must be regulated so that it is neither

can also be drawn by a hand method [1, 5, 9] high enough to volatilize an excessive amountin which the stock is melted at a point and of the fused silica nor low enough to prevent

5

sufficient melting. With the thicker stock, bers [50, 3M5, 316]. Finger marks on largecurly fibers result from uneven heating [1, 2, pieces of fused silica appear as devitrifica-

94, 1951, so curly in some cases as to resem- tion marks on the material after it has been

ble watch springs [2]. These fibers can be worked in a flame [300]. Opinions regarding

straightened by "soft flaming" but are weak- methods to counteract this contamination andened as a result [1951. A slight bowing due subsequent weakening differ among workers.

to the effect of gravity on larger fibers has Soae workers prefer to prevent contamination

been noticed when they are drawn horizontally; by working under scrupulously clean conditions,however, the small fibers can be drawn either using only those fibers which are freshly

horizontally or vertically with no noticable drawn, free of dust and other contamination

effects [316]. and by never touching that part of the fiber

With flame-blown silica fibers about 21 which is used in a piece of apparatus [9,

microns in diameter and 15 cm in length, Hor- 3151. Others prefer to remove any contamina-

ton [185] observed variations in the diameter tion on the fibers by cleaning with a solution,

of about I to 4 percent. The fibers had the such as chromic acid, before the joints are

appearance of a sequence of irregular conic fused or the fibers worked [104, 1061. The

sections within an over-all tapered fiber, finished piece is also cleaned to remove any

Similar irregularities in shape of about six deposits which result from handling during

percent in a quarter of an inch were observed the working of the fibers. The effects of

in some machine drawn fibers [3161. The long cleaning solutions on the strength of fibersterm variations could be associated with var- and on the finished piece are not known or

iations in the stock and with the pulling have not been detected. The effects are com-phase. For example, rungs or tapes used on plicated by the fact that while cleaning,

sectional reels cause an irregular rate of rubbing with the fingers, and adsorption of

pulling which results in a slight taper in moisture reduce the -.ensile strength, these

the fiber between the markers 1103], a varia- processes tend to increase the resistance to

tion which is especially noted in fine fibers. scratcaing or to any further surface injury

The short term variations in the diameter [1131.

could arise from insufficient melting of thestock, oscillations of the flame, vibrations 3.3 Silica-Fiber Apparatusin the drawing equipment, or even a natural

oscillation between the rate of drawing and The fineness of the silica fibers usedthe rate of melting. Although these diameter necessitates the use of holders or jigs tovariations were not mentioned by many workers, prevent the fibers from blowing away while

it is possible that they were not sought or they are being fabricated into the variousthat they were so small as to be hardly per- devices. A number of different types of

ceptible with the measuring equipment used. holders and jigs are used. These aids, used

Although fused silica fibers seem to lose mostly in the fabrication of microbalances

strength with age [188 to 191, 199, 2001 the and electrometers, are described more thor-real effects of actual storage conditions on oughly by the individual experimenters.

this loss of strength are not known. It The holders used by Neher [91 and bywould appear that a method of storing fibers Carmichael [315] corksist of forks with two

might be determined which would help to pre- prongs whose spacing is eitner adjustable or

serve the strength and to prevent damage by is fixed for a particular length of fiber.dust, moisture, certain atmospheres, or acci- These forks hold individual fibers only atdental scratching. the ends of the fiber and ailow the fiber to

Contamination and scratching of the ii- be held in any position while joints are

bers from dust, salts in the air, and from fused. The holders can be held in a micro-

handling with the fingers tends to weaken the manipulator, or the proper sized holder can

fibers. When the fibers are fused or heated be placed in position in a jig designed tothe silicates formed by impurities on the insure reproducible construction of a partic-

surface can cause brittle joints and weak fi- ular device. The used portion of a fiber is

6

never touched so that any damage which may any disadvantage and thus determine the

result from contact with a holder or with the method to be followed in individual applica-

fingers is avoided. With this type of holder tions.

the fibers can be removed directly from a In the construction of spiral or helical

storage reel, thus eliminating extra handling springs, long even silica fibers are needed.

steps. Although shorter fibers have been fused to-

Kirk [104,311] used two types of holders; gether to make a fiber long enough for a

one is a holder with multiple prongs on which spring [106] this practice may affect the

"a fiber is supported and shaped; the other is performance of the spring."a Y-shaped holder on which a fiber is sup- Silica springs are wound on a suitableported. Fibers are clipped or fastened to mandrel by feeding the fiber in place and

these holders at intervals along the usuable heating it to conform to the shape of the

length and can then be held in any desired mandrel. The mandrel may be tapered for ease

position. A predesigned assembly of these of removal and to suit the design character-

holders allows for repeated construction of istics of the spring [107]. Several mate-

fiber apparatus. However, where the holders rials have been used for mandrels including

are made of heavy silica fibers or other hard carbon and graphite [55, 95 to 98, 100, 101,

materials there is some danger of scratching 106], Pyrex [l05]1 Vycor and fused-silica

resulting from the contact between holder and [99, 107] rods or tubes. Of these, fused sil-

fiber. ica and Vycor mandrels have been considered

In some applications where fairly heavy most suitable because they have the slight

fibers are used for a beam or supporting mem- adherence necessary to prevent stretching and

ber, metal, carbon, or asbestos blocks appro- a coefficient of expansion the same as or

priately grooved are used as templates. The similar to that of the fibers wound [107].

fibers are laid in the grooves, the joints Carbon and graphite mandrels may damage the

fused or cemented and the fine fibers either spring by overheating and by contamination.

drawn out from the assembly or fused on in Springs wound on carbon or graphite mandrels

the appropriate places [28, 32, 35 to 37, 39, may have squared rather than smooth coils.

40]. Joints made in this way are frequently Over the years many methods of windingbrittle and weak because of overheating and silica springs have been developed. In the

may be contaminated by the block itself [50]. earliest successful method the mandrel was

Two methods using templates as patterns turned about its tilted axis by hand and, as

or guides in the construction process have the fiber coiled about the mandrel, a flame

been used. The templates used by O'Donnell bent it to the mandrel [55, 95, 106]. This

[102] consist of thin glass plates equipped method has been greatly improved and mechan-

with pegs to locate the fibers and with sil- ized to produce more regular springs [97, 98,

ica fiber clips to hold the fibers in place 100, 105]. In the latest and most completely

for fusing. The plates are held in proper mechanized method, as described by Ernsberger

positions in jigs so that all the joints can and Drew [107], the mandrel is turned in a

be fused through the appropriate holes drilled lathe which has controls for the fiber and

in the plates. EI-Badry and Wilson [50] used flame guides to allow for uniformity in the

a series of patterns drawn on matt paper. spacing and in the coils, and for proper

Fibers attached to microscope slides are heating of the fiber.

placed in position over the pattern and the As with all procedures for fabricating

joints cemented in successive steps until the apparatus from silica fibers, the precautions

assembly is completed. All the fibers can be against contamination and damage resulting

attached in this manner, from handling and from the actual fabrication

While there are certain disadvantages to process should be observed.

the use of some of the methods such as con- For many applications a conducting sur-

tamination, overheating of the joints, face of metal or of graphite [79] on silica

scratching, the ease of manipulation and the fibers has proved very satisfactory. There

character of the final assembly may outweigh are several methods for putting this coating

368o21 0 - 56 -2 7

on the fibers including sputtering, evaporat- heat radiat ion, the fumes ai'd ,ilsagdi' lideing, plating, or baking a painted coat 12, 5, odors cause fatigue and nausea among most

79, 91, 94]. A coating or plated surface on workers. For these reasons and because ofthe ends of fibers has also been used to fa- the possibility of silicosis most workers it,

cilitate holding fibers in certain apparatus. this country work only 2 or I hours with

Additional equipment for fiber work in- quartz or fused silica. Speciaal glasses such

cludes a binocular microscope, nacromanipula- as those using Noviweld lenses should be used

tors and holders, torch holders, small to protect the eyes from the intense infrared

torches made from fine fused-silica tubing or radiations. A suitable exhaust system should

from hypodermic needles. Detailed descrip- be used to carrý awa' the heat anid the fumes.

tions of the techniques of construction of Phi le these disconforts have beei, re-

fiber devices and descriptions of the equip- ferred to in large scale operations there is

merit used in their fabrication are given by little reference to such effects when workii,,

Neher [91, O'Donnell [1021, Haring [103], and with silica libers. This could be because ofKirk [104]. Individual construction details the smaller masses involved and the consequent

and specialized design problems are discussed smaller amounts of heat and fumes. II is

most completely' by the workers who developed well to note, however, that there are some

the apparatus (see applications). Frecautions to be taken in quartz fusing op-

In regular fusing operations involving erations.quartz or fused silica [299, 3001 the intense

8

4. MECHANICAL PROPERTIES OF FUSED-SILICA FIBERS4.1 Strength pear to be of some help. However an analysis

S tr t shows dissimilarities in individual testing

Strength is defined as the resistance of procedures, in the specimens, handling, ther-a material to fracture or quantitatively as mal treatment, temperature and humidity con-the critical value of stress at which frac- trols, in the test equipment and in the pres-ture occurs r22 3 1. The mechanical strength of entation of the results; all of which tend tofused silica can be considered in three cate- affect the values determined for strength andgories: compressive strength, bending thus make any strict comparison misleading.strength, and tensile strength. The compres- The results of the various investigators showsive strength of fused silica is 2 or 3 times trends in the strength such as an increasethat of ordinary glasses and is about one in strength with decreasing diameter and

fifth less than that of quartz (7]. The length, some variations due to the atmos-bending strength, modulus of rupture, of sil- phere, to thermal treatment, to age, to draw-ica fibers is generally higher than the ten- ing methods, to handling methods. The pos-sile strength [7, 188 to 190, 241] and is af- sible errors that may result from takingfected by the same factors which cause so values at random from experimental curves canmuch scatter in the values of the tensile be illustrated by a plot of all the previous-

strength. The tensile strength of silica ly determined values of tensile strength with

fibers is smaller than the crushing strength. respect to the diameter, on a log-log scale.However, the individual values vary much more Here where a value of the strength for athan the uncertainty of measuremlent whi-h 100-micron fiber is about 35 to 40 kg/m 2 ,makes comparison of values difficult. the tensile strength given for a 10-micron

lhe tensile strength of silica fibers ex- fiber may range from 2 to 12 times larger

ceeds that of most metals and other materi- than this value, while those for a 3-micronals. This high strength coupled with an al- fiber may range from 5 to 100 times larger

most perfect elasticity makes silica fibers than the values given for the 100-micron fi-desirable for many applications. However the ber. The trend on these curves for fibersgreat range of individual values for strength over 100 aicrons is a slow approach to aand the slight loss of strength due to cer- constant value, while that for fibers less

tain conditions of use prevent or limit those than 100 microns is a rapid increase inapplications needing more than the minimum strength in a linear manner.values of strength. For this reason there is In the experimental curves showing the

a need to inves igate the strength of silica strength of silica fibers as determined by an

fibers with special atten'ion given to the investigator different methods of showing the

methods of production and to Lhe conditions results are used. Such curves may indicateunder which they are used. Most measurements maximum values [188 to 1911, values adjustedof tensile and bending strength on any quan- for the presence of flaws [1951 , averagw,tity of silica fibers [188 to 195] were made values [192, 1931, values without the condi-with fibers produced by flame, hand, bow and tions specified [11], and in some cases allarrow or other simple methods. The only the values determined [1, 2, 188 to 190].available measurements on machine drawn fi- Without some standardization of the testingbers [11) as published are n't very complete. procedure, comparisons of such values are

The scattering in the determinations of frequently meaningless. For example the di-tensile and bending strength of fibers is so ameter measurements can greatly affect thelarge that the assignment of a definite results [307, 309] especially with small di-strength is difficult. A graph showing ex- ameter fibers. An interesting discussion ofperimental values obtained for tensile the problem of correlation of strength meas-

strength and for bending strength might ap- urements on glass is presented by Bailey [2071.

9

The mechanical strength of glasses has severe quench from the rapid rate of coolingbeen termed a "structure sensitive" property provides a fire polished surface which re-[213]. The random structure with the result- sists chemical reactions and decreases or de-ing variations in bond strength and in the lays the effects of cracks or flaws. Recent-interatomic distances gives glasses and espe- ly the importance of the rate of cooling oncially fused silica an initially high value commercial glass fibers was brought out byfor strength. However, the theoretical Slater [227] who found little variation instrength is somewhat modified by certain dis- strength between fibers 0.002 in. in diametercontinuities in the structure and by the ef- and those 0.0002 in. in diameter when bothfects of various external factors on them were drawn at the same rate of cooling. The[116, 213]. The differences between the the- rate of cooling can be regulated in the fiberoretical and the experimental strength of drawing machines by adjusting the drawingglasses led Griffith [199, 200] to postulate temperature and the drawing speed within cer-the existence of certain concentrations of tain limits [11]. The thermal history of fi-energy in the form of submicroscopic cracks bers as determined by drawing conditions isor flaws throughout the material. These one of the important single factors in theflaws could occur either during the manufac- strength of silica fibers.turing process or in subsequent treatment of From further considerations of the draw-a glass. The effect of flaws or discontinui- ing process two theories were developed inties is to produce local stresses which ex- order to explain the difference in strengthceed the average stress and as a result lead between fibers and the bulk product fromto failure. Calculations of the effects show which they are drawn. One theory was ad-that they could conceivably account for the vanced by Murgatroyd [172] who suggested thatdifferences between theoretical and actual in order to allow the continuity of the fiberstrength [200, 208, 223]. Proof of the ex- as drawn the strongest bonds were selected

istence of flaws was demonstrated in experi- from the melt and alined parallel to the di-ments [163, 164] which showed fine cracks in rection of drawing while the weakest bondsplaces on annealed fused silica rods and were alined perpendicular to the direction oftubes where no accidental scratching could drawing. In the other theory Bickerman andoccur. The Griffith flaws appeared to be Passmore [220] believed that the flaws in aabout 10 microns long by about 0.01 microns fiber must be oriented favorably to enabledeep. the fiber to be drawn. These theories em-

phasize the importance of the drawing processa. Production rather than the reduction of size which ac-

The drawing process for silica fibers has companies it. Although there are many whosome effect on the final strength. In regard support or have supported these ideas [143,to the method of production fibers drawn by 149, 173, 213, 215, 222], recent experimentsbow and arrow appear to have higher tensile [148, 161] on annealed glass fibers indicate

bow nd rrowappar o hae hghertenile that there is neither an orientation of bondsstrength and show less scattering of values

nor one of flaws in fibers which could be re-than those blown in a flame which in turn ap-pear to have higher tensile strength and show sponsible for the high strength. Additionalless scatter than those fibers drawn by hand evidenc ed ta an s tructure[195]. Comparative tests with machine drawn wat idened by X-a sudie [0athat if an orientation at the surface ex-silica fibers are not available in the cur- isted, its influence on the strength could

rent literature. It is entirely probable not ed, tet ed o193h.that such fibers could show greater strengthand less scatter because of the controlled b. Sizeconditions under which they can be drawn.

The drawing process can be regulated with Experiments on the breaking strength ofrespect to drawing temperature, rate of silica fibers show an increase in strengthdrawing, and rate of cooling and thus deter- with decreasing fiber diameter [1, 11, 188 tomine to some extent the strength [103]. The 195]. Similar results occur with other glass

10

fibers [195, 2351 and in glass slabs where a noticeable effect of the surface and its

small area is tested [236]. An additional treatment on the strength of silica fibers.

increase in strength is found when the speci- The chemical and physical nature of the sur-

men length is decreased [190,220]. Assuming face of glass has been described as different

that flaws exist in fibers, then the apparent from that of the body of a glass [160].

dependence of strength on fiber size can be Structurally, this influence of the surface

explained in a statistical manner by the de- can be explained by a broadening of the bond-

creasing probability of effective flaws oc- strength distribution nea- the surface which

curring as the volume or the surface of a fi- tapers off toward the interior and leaves

ber is decreased [193, 213, 214, 220]. Al- weak bonds at the surface. Although flaws

though there is some basis for assuming a re- are considered to be distributed throughout

lation between the probability of flaws oc- the material, the strength is determined by

curring in relation to the size of a specimen, the most dangerous flaws on the surface [193,

an exact expression for the distribution 199, 200], for a surface flaw can exert twice

function of the cracks or flaws is difficult the stress of the same sized inner flaw [221].

to formulate [223]. The stress variations in individual fibers

Orowan [223) considered that the neces- resulting from the existence of flaws,

sarily arbitrary assumptions as to the size scratches on the surface, and the effects of

of cracks and to the number of cracks in a surrounding conditions on them, may cause

given volume or over a given area without ac- fracture before the theoretical elastic limit

counting for the individual history of the is reached. Because of the effects oncracks led to results more interesting math- strength of these conditions it is difficultematically than correct physically." How- to bring values of individual fibers within aever, the thinnest silica fibers which have small range of error. It may be appropriate

been investigated have diameters less than to mention here the often quoted idea ofthe supposed size of an ordinary Griffith Little that one measures not the strength of

the uppsedsiz ofan odinry rifith a material but the weakness of its surface.crack and the strength of these fibers ap-

pears to approach the theoretical values de- A discussion of the chemical and physical as-

termined for the material. These fibers pects of the surface condition of glasses is

could not contain flaws as large as those in presented in a series of works by Weyl [116,

larger specimens and unless the flaws were 175, 213].

large enough to weaken the fiber by reducing d. Environmentits cross section, the strength, Orowan con- Glass fibers are tested and used undercluded, should be higher than that of a thickrod. Thi cosidraton o th nuberandadverse conditions because of the characterr o d . T h (! c on s i d e ra t io n o f t h e n u m b e r an dof t e m e r a . F r h e c u l n r al o -

size of flaws in a given length particularly of the material. For the actual normal con-affets masuemens o beningstregthditions of atmosphere are corrosive to com-affets masuemens o beningstregthmercial glasses and to some extent to fused

where the probability of flaws occurring at a merca glasses and to so e nt tosed

certain point of maximum stress is smaller silicarrost a etain al ateresare corrosive to metals and other materialsthan that of flaws occurring over a givenlength, with the result that values for bend- 112131. Experiments on strength of fused sil-

lengh, iththereslt tat alus fr bnd- ica fibers in certain controlled atmospheresing strength are generally higher than those icafe i ertin cotrled mostereaindicate a relatiorship between moisture ab-for tensile strength. In addition, the non-

uniformity in strength over the entire length sor bed and the strengt [ 1of a fiber [241] due to the existence of After silica fibers have been exposed toofas fiser a[p4ssi e caushe fo sthenreat r e the air, water vapor is absorbed which de-flaws is a possible cause for the great range cr a e th s r ng h f o t at b ai dof values determined for any one size of fi- create streth fr ta obained

ber. A well-designed statistical test proce-

dure might help to remedy the situation. this layer of moisture is removed the tensile

strength increases. The tensile strength of

c. Surface fibers, which have been heated in a vacuum toThe large surface area in comparison with remove all absorbed moisture, increases up to

the volume of fibers may account for the 3 or 4 times the strength in water or water11

vapor. When the same or similar fibers were the effects of coatings on textile glass fi-dried in CaCi2 the strength increased about bers revealed that while the coating itselfone and a half to two times the strength in did not increase the strength it did lessenwater vapor [190, 192, 193]. When water or the effects of atmospheric attack in some in-

alcohol vapor was admitted into the evacuated stances. The chief use of a resin or waxapparatus the strength immediately decreased coating is to prevent seizing of glass fibersto its original values. For fibers of a when they are woven or pressed. Silica fi-given diameter the strength in oil, and in bers are much more resistant to damage thanalcohol is greater than that in air. In ad- commercial glass fibers and in present appli-dition, the strength of those fibers which cations careful handling will avoid muchwere in the air for a long time was the same harmful damage. A protective coating mayas that of fibers broken in water, and the prove useful however.strength of fibers stored in a room satu- The tensile strength of silica fibers in-rated with oil or alcohol vapur was the same creases slightly with increasing temperatureas that of fibers broken in the liquids, up to the flow region, showing minimum

Further experiments [188 to 193] on the strength at room temperature [108, 116].effects of etching in hydrofluoric acid re- This can be explained by the fact that thevealed a 3 to 5 fold increase in strength of healing of the cracks by surface diffusion assilica fibers after the etching. There was, the temperature is raised outweighs the in-however, a considerable amount of scattering fluence of thermal motion which tends to in-of the values which was partially explained crease tension and to spread the cracks[193] as a result of the etching process [2131. In addition, the absorbed moisturewhich alternately smoothes and then exposes layer which is thick and active at room tem-

flaws; the strength thus depending on the perature is driven off at higher temperaturescondition of the surface at the time of re- and is inactive at low temperatures [2131.moval from the acid. The strength of fibers Experiments [10] showed an increase inbroken in the acid did not differ much from strength of silica fibers after they werethat of fibers broken in the air. Further- heated to 1,188°C for 4 hours and then al-more, the strength gradually increased with lowed to cool. However, heating silica fi-etching until a certain maximum thickness had bers in a flame makes them weaken consider-been removed, after which no further increase ably [2]. Although the strength may changewas noticeable [1901. These etched fibers with temperature changes, the strength atwere very susceptible to damage from the air relatively low temperatures does not dependor from scratching although a protective on the temperature [193].coating of shellac helped them retain their The values of strength thus vary depend-high strength [193]. ing on the conditions of test and of use.

The strength of silica fibers is de- Any improvement in these conditions shouldcreased by scratches from atmosphere dust increase the apparent strength. A strength[199, 200] and from other fibers. There is test gives not the true value of strength butan immediate loss of strength resulting in rather the residual value after the fiber has

fracture when a fiber under strain is touched been damaged in some way [114].by another silica fiber. A further decreaseof strength results from contact with the 4.2 Elastic Propertiesfingers [213, 3151, although those partsrubbed with the fingers are protected from Silica fibers exhibit almost perfectfurther injury [113]. It has been suggested elasticity up to the breaking point; any de-occasionally that silica fibers be coated to viations are so slight as to be barely meas-prevent damage from the various atmospheric urable.conditions or to increase the strength. Although the elastic constants have beenThere is at present no published data avail- determined since silica fibers were firstable on the effects of coatings on silica fi- used [1], there is not complete agreement be-bers. However, investigations [195, 227] of tween early values and those determined

12

rather recently. Again as with strength, in a vacuum and in alcohol vapor than in

testing methods and conditions of the test water or moist air [192], which increase

and of the specimens considerably affect the could not be due to the weight of the mois-

final results with the errors either magni- ture. This effect was also observed in sil-

fied or diminished. Absolute determinations ica springs where an increase in elongation

of elastic constants require accurate values occurred in water or alcohol vapor but not in

of size, density, and mass [244]. Tempera- inorganic vapors when these springs were ex-

ture effects must also be considered if in- posed to the vapors after being in a vacuum

dividual measurements made under different [196]. This was, however, erroneously at-

conditions are to be related or compared. tributed to an expansion of the silica be-Anelastic effects which cause the strain to cause of absorption. Silica fibers etched in

lag behind the stress, must be accounted for hydrofluoric acid had values of elastic modu-

since the elastic constants are related har- li in close agreement with unetched fibers

moniously only if the rate of stress appli- with some uncertainty regarding an actual in-

cation is the same [224]. crease [192].Within the experimental error, the elastic Slight deviations from "perfect elastic

moduli can be considered to remain fairly con- ity" of silica fibers can be attributed

stant over the range of sizes used. The exper- slight delayed elastic effect noted when

iments of some investigators [1,2, 5, 9, 188 bers are twisted through large angles. Ti.

to 191] show a dependence on the fiber diame- classification of the types of deformation

ter of Young's modulus and of the shear modu- due to applied load generally used are: an

lus. However, later investigations showed instantaneous strain which is completely re-

that these moduli were independent of the di- coverable, a delayed elastic strain which isameter and of the length, and pointed to recovered slowly, and a viscous flow which is

slight errors in former measurements [11, 193] not recoverable but which appears to be non-A summary and discussion of much of the existent in fused silica at temperatures be-

experimental work on the elastic constants of low about 800°C [7, 117]. The delayed elas-

fused-silica fibers can be found in Sosman's tic effect is similar in effect to a progres-

treatise on silica [7]. Some of the experi- sive increase in viscosity and frequently is

ments [1, 2, 182, 185], and later work [184, cLnfused with a viscous flow or creep [117].187, 188 to 191, 244] show that Young's modu- Although delayed elastic effects are of some

lus and the shear modulus do not vary appre- concern in commercial glass fibers, the ef-

ciably over the range of temperatures used in fects in fused-silica fibers are rather neg-the tests. There is a slight linear increase ligible, approximately 100 times less than in

in both of these constants with increasing other glass fibers [217, 218]. A delayed

temperature up to the region of viscous flow. elastic effect has been noticed in instru-

The moduli continue to decrease from values mcnts using a torsion fiber. However, a con-

measured at room temperature to those meas- stant correction could be applied, as after a

ured down to -200 0C [245]. The change in short time the rate became constant. In ad-

shear modulus with change in temperature af- dition the effect decreased with decreasing

fects the readings of very sensitive meas- fiber diameter so that in the finest fibersuring instruments [62, 63, 107]. For ex- the effect is barely perceptible [1, 66]. A

ample, Sheft and Frjpd [62]found an increase part of the apparent delayed elastic effect

in rigidity, in a spring' balance, of about could also be attributed to the mountings

0.02 percent per degree centigrade which was used for the fibers [24, 53]. With increas-

reflected in a change in elongation of 0.02 ing temperature the delayed elastic effect,percent of the total load per degree centi- expressed as the ratio ot the delayed strain

grade. The change seemed to indicate to the instantaneous strain, increases.

strengthening with increasing temperature. Fused silica has the lowest value of the

The elastic moduli vary with the sur- ratio of all the glasses [i17].

rounding atmosphere, exhibiting higher values

13

5. SOME PROPERTIES OF FUSED SILICAThe properties of fused silica observed postulated by Zachariasen was later verified

at room temperature combine to make it such a by Warren and his co-workers. These experi-useful material. While some properties may menters using a Fourier analysis of the X-raychange with time, temperature, or surrounding diffraction pattern, first of fused silicaatmosphere the magnitude of the changes is and later of other glasses, were able to showusually small compared with similar changes not only the arrangement but also the averagein other materials. The properties discussed interionic distances and the mean bond an-in this section usually are determined for gles. They also showed that any particles infused silica in bulk form, but are of some fused silica showing similarity to crystalsimportance in the use of silica fibers. are too small to be described as crystalline

Since fused silica is a glass, the most matter.simple glass, it would seem appropriate to The structural arrangement of fused sil-place glass in reference to other states of ica as confirmed by Warren's analysis con-

matter. The ASTM describes glass [301] sim- sists of a short order structure which isply as "an inorganic product of fusion which tetrahedral in form with 1 silicon ion bondedhas cooled to a rigid condition without crys- to 4 oxygen ions, with each oxygen ion bondedtallizing." The vitreous or glassy condition in turn to 2 silicon ions. The tetrahedrais rather difficult to define as is evidenced share the oxygen corners to form a long rangeby the numerous papers on the subject some of three-dimensional random network. A two-which attempt to describe the condition of dimensional picture of this structure showsglass as distinct from or similar to the sol- the network made up of a series of irregularid and liquid states [124 to 164]. rings, where the average number [154] of

tetrahedra per ring is six and the number of5.1 Structure tetrahedra in individual rings varies from 3

to 10 or more. Where the bond angle betweenThe present picture of the atomic ar- tetrahedra is nearly 1800 and varies slightly

rangement of fused silica developed from the for successive tetrahedra, a buildup of alaws of crystal chemistry, analysis of X-ray random or "disordered" [143] network is al-diffraction patterns, and studies of certain lowed. It is this flexibility in bond anglesphysical properties. It was from studies on which gives fused silica and other glassescrystal structure that Goldschmidt [126], the long-range disorder of a liquid; whileZachariasen [126], Sosman [7], and others the ordered distances and angles within a[124, 151, 155] developed theories on the tetrahedron gives them the short-range orderatomic arrangement and formation of glasses. of a crystal.The early methods and the apparent similaritybetween crystal and glass led some observers 5.2 Chemical Durability[124, 151] to believe that fused silica andother glasses were made up of particles of At room temperature, fused silica is at-crystalline material called "crystallites." tacked by hydrofluoric acid. Near 300 toHowever, even without the analytical tools 400-C, phosphoric acid starts to attack sil-later used by Warren and his co-workers [127 ica. At high temperatures and up to 1,000°C,to 134], Zachariasen postulated the existence weak and moderately concentrated solutions ofof a random three-dimensional network with basic salts and of metallic oxides and basicenergy comparable to that of the correspond- salts react with fused silica. Fused silicaing crystalline network in fused silica and is reduced to silicon by carbon at tempera-glasses in general and laid down certain con- tures over 1,600 0 C [110]. It is also reducedditions as necessary for the formation of at high temperatures by hydrogen which formsoxide glasses. The network picture thus silicon hydride. When this material is

14

cooled the part of the silicon hydride not behavior of measuring instruments [62, 106,oxidized decomposes into silicon and hydro- 1961.gen and forms a deposit of silica and sil-icon on the surface of the material [281]. 5.5 HardnessThis deposit, observed on fibers fused in aflame containing hydrogen, has frequently The "hardness" of fused silica is notbeen confused with devitrification of the fiber, only a function of the strength and the elas-

tic properties but must be greatly influenced5.3 Permeability by the chemical resistivity of the glass and

by the environment [116]. The hardness ofFused silica is permeable to neon, hydro- fused silica has been determined on the var-

gen, and helium at elevated temperatures and ious hardness scales although an exact defi-under pressure [7, 10, 109, 112, 296]. De- nition has not been formulated. The Nbhs'vitrified and nontransparent types of fused hardness numbers given for fused silica rangesilica are more permeable than the trans- from 5 to 7 [9, 10, 3121. The Knoop indenta-parent form. The rate of diffusion may de- tion hardness number given for fused silicapend to some extent on the size of the open- is approximately 475 [284] where the valueings in the random network [116, 296] and on for quartz is approximately 710 to 790 [302].the size of the gas atoms. Thus fused silica This Knoop indentation scale shows fused sil-with larger random openings than other ica to be harder than some other glasses andglasses would allow larger gas molecules to many metals including brass, gold. silver,pass through and the smaller gas molecules tantalum, and some stainless stec'Lh.could pass through more quickly than in someother glasses [2963. Reviews of work on per- 5.6 Densitymeability indicate considerable differencesin observations on the amounts of diffusion The density of transparent fused silicathrough fused silica [4, 109, 112]. is approximately 2.21 g/cm3 while that of

nontransparent fused silica is about 2.07g/cm3 . The density of the specimens of fused

5.4 Sorption silica may vary about 0.05 percent or even asFused silica exhibits a combination of a much as 0.1 percent, a variation of about 10

reversible adsorption plus a slow permanent times that of specimens of crystallinesorption of water [285]. This reversible ad- quartz [7]. Sosman thought that it might besorption was reduced about 35 percent by acid possible to correlate these variations withtreatment, and an additional 30 percent by the source, with the thermal treatment orheat treatment. The rate of permanent sorp- with the state of strain in the fused silica.tion as found by Barrett [285] in finely Since that time these factors have been re-powdered transparent fused silica, fell off lated to density changes in a number of ex-rapidly in about a half hour after exposure periments [170, 289, 292].to moisture and then became constant at about Density measurements at temperatures up4x10"'' g/crr.2/hr. This process seemed to be an to 1,7000C showed that there is an equilib-actual diffusion of water into the silica. rium density which increases with temperatureFused silica apparently does not adsorb in- [289]. The specimens used changed from theorganic vapors in sufficient amounts to be initial density to the equilibrium value formeasurable [196]. It does however adsorb al- the temperature at which the specimen wascobol vapor [192, 193] and the vapor of par- heated. The rate of change increased as theaffin oil [175] though in smaller amounts temperature increased so that at 1,3000C theythan the water vapor. This adsorption of had a density corresponding to the high-tern-moisture affects the strength and the elas- perature density. These changes in densitytic constants such as Young's modulus and with heat treatment were found to occur inthe shear modulus which in turn affect the addition to the normal thermal expansion, in

15

effect a slow change of volume in addition to 108]. Specimens with a great deal of surface

an instantaneous change in volume, area such as powdered specimens, or those

Under uniaxial pressure an increase in with a large number of bubbles such as trans-

density was noted [289, 292]. Bridgman and lucent types devitrify more readily than

Simon [292] found a threshold pressure near large pieces and transparent fused silica

100 atm above which the structure of fused [7, 108, 183, 249].

silica seemed to collapse. Rapid increases A possible form of devitrification oc-

in density of 7.5 percent and occasionally at curred when fused-silica articles were ex-

high as 17.5 percent, to 2.61 g/cm3 ,were posed to radium salts at the temperature of

found. After X-ray examination showed the boiling water but did not occur at lower tern-

specimens to be amorphous, the increase was peratures [2461.

attributed to a folding up of the network; a A dependence of the rate or of the amount

bending of the Si-O bonds rather than a of devitrification on previous thermal treat-

shortening of them. This folded-up structure ment is evidenced by the appearance of less

was mechanically stable at ordinary tempera- devitrification on specimens made from high

tures, but with heat treatment the density temperature melts than on those made from

decreased to the original value. Both Doug- lower temperature melts [71, as well as a

las [289] and Bridgman and Simon concluded greater tendency toward devitrification ex-

that there appears to be an equilibrium con- hibited by annealed specimens.

figuration which may be approached from ei- The presence of fluxes such as calcium

ther direction. Their experiments support carbonate, calcium oxide, liquid silicates,

Sosman's theory that heat treatment and and certain metallic oxides hastens devitri-

strain could cause variations in individual fication [253]. Although titanium oxide and

specimens. zirconium oxide have been added to fused sil-ica to decrease any tendency toward devitri-

5.7 Devitrification fication, the resulting glasses devitrifymore rapidly than pure fused silica [7].

The rate of crystallization or the de-

vitrification of fused silica depends on the

temperature to which it is heated, the pericd 5.8 Thermal Expansionof heating and on the degree of subdivision. Investigators have long been interested

It is also affected by atmospheric dust [251], in the thermal expansion of fused silica,

by the previous thermal treatment [7], and which is smaller than that of almost all oth-may be related to the viscosity [248]. er materials, because of its usefulness in

While fused silica is thermodynamically measuring instruments. Critical examinations

unstable at all temperatures below 1,710°C, of the experimental work on expansion of

its molecular sluggishness (viscosity) at fused silica presented by Sosman [7] and by

room temperature is so great that no change Souder and Hidnert [260] show that differ-

toward crystallization has been observed at ences in testing equipment and in methods of

such temperatures [7]. Fused silica de- testing are responsible for mady variations

vitrifies only after prolonged heating at observed in the results of many investiga-

high temperatures. At 1,000°C [7] or even at tors. With transparent and nontransparent

1,2000 C [10] devitrification is hardly per- fused silica tested over a range of -125°to

ceptible. Above these temperatures fused i, 000°C, Souder and Hidnert determined a

silica slowly devitrifies into cristobalite. critical temperature of about -80°C where

When samples were powdered, almost complete specimens had a minimum length. Viscous flow

devitrification occurred after heating at which started around 800°C hindered their

1,100°C for 6 days or at 1,6000C for I hour measurements of expansion above that tempera-

[83]. ture.

Although devitrification could not be The expansion of fused silica changes

called a purely surface phenomenon [112], it slightly with heat treatment. Heat treatment

does start at a surface and work inward [7, over the range 200 to 750°C increased the ex-

16

pansion by about 20 percent of the total ex- been studied. Much of the work done on vis-

pansion [170] which results in a slightly cosity is with specimens of commercial glasslarger coefficient for annealed specimens types, for studies on fused silica are lim-than for unannealed ones [2601. In addition ited by insolubility and by the high tempera-to the change in expansion, a permanent in- tures required to decrease the viscosity

crease in length occurs after heat treatment. enough for present experimental conditionsAn average variation after cooling from [306].

1,000°C was 0.001 percent with a maximum of At room temperature and below several0.003 percent [108,260]. hundred degrees centigrade the viscosity is

The coefficient of expansion of trans- so high that the material is considered asparent fused silica is slightly larger than solid; for instance, the viscosity at room

that of the nontransparent [108, 255, 260], temperature has been estimated at about 1060a difference amounting to about 25 ppm on the or 1070poises which is rather incomprehensi-heating cycle and to about 30 ppm on the ble for practical meaning [267]. Prestoncooling cycle [108]. thus considered that viscosity has an upper

Thermal expansion or the instantaneous limit of definition as well as a lower limit,change of volume with change of temperature with a viscosity over 1014 approaching infi-

[289] can be explained by structural consid- nity as a quantitative measure. Even in theerations and defined in terms of thermal his- melt the viscosity is so high that fused sil-tory and the temperature of measurement. The ica never really becomes fluid [10]. Thesmall coefficient for fused silica as corn- viscosity of fused silica at room temperaturepared to that of quartz could be a direct re- is so high that movements toward an equilib-sult of the random structure of the glass rium configuration (stabilization) or a con-[133, 166]. Since a small thermal expansion figuration appropriate to the temperature at

can take place by a change in average inter- which the glass is used are prevented [117].atomic distance or by a configuration change The high temperature configuration of the

such as a change in bond angles, an increase rapidly cooled fibers is retained permanentlyof disorder or an increase in structural for all general purposes. The fibers havebinding would decrease the expansion coef- what is referred to as a low viscosity com-ficient [289], where bond angle changes in pared with slowly cooled fused silica, a dif-the random structure can compensate for other ference which may account for the differences

effects which tend to increase the coeffi- in properties between the two forms [117].cient [153]. The negative coefficient is due Studies of the viscosity of fused silicato an increase of structural binding with de- and other glasses are incomplete. The proc-crease in temperature 1166, 170]. ess of viscosity changes due to thermal

treatment and the effects of this on the

5.9 Viscosity properties of glasses are not completelyknown. A survey and discussion of the avail-

It is principally with reference to the able work and its implications was made bydevelopment and improvement of methods of Jones [117]. This and other articles [174,melting and working glasses that the rela- 267, 268, 270, 276 to 280] should be con-tionship between viscosity and temperature, sulted for details of the developments and

time, composition, and heat treatment have problems in studies of viscosity.

17

6. BIBLIOGRAPHY

6.1 Books and Review Articles on [211 A. Friedrich, Uber ein verbissertes Nidell dervereinfachten Sal voni -FederaageoNkrocheaie 15,Fused Silica and Fused Silica Fibers 35 (1934).

£22] H. K. Alber, Improved apparatus for microprepara-(11 C. V. Boys, Production, properties and suggested tive work, Ind. Eng. Clem. (Anal. Ed.) 13, 656

uses of the finest threads, Phil. Mag. 23, 489 (1941).(1887). [231 G. T. Seaborg, The transuranium elements, Science

£21 R. Threlfall, Laboratory Arts,(Macmillan & Co., 104, 382 (1946).Ltd., London 1898). [241 B. B. Cunningham, L. B. Werner, The first isola-

(31 W. A. Shenstone, Vitrified quartz, Nature 64, 65 tion of Plutonium, J. Am. Chem. Soc. 712, 1524(1901). (1949).

(41 P. Gunther, Quarzglas, (Julius Springer, Berlin1911). c. Nernst Type Balance

[5] C. V. Boys, Quartz fibers, Dict. Appl. Phys. III,695 (1923). [251 W. Nernst, E. H. Riesenfeld, Uber quantitative

£61 B. Paget, Hardness, density and uses of fused Gewichtsanalyse mit sehr kleinen Substanzmengen,silica, J. Boy. Soc. Arts 72, 323 (1924). Ber. deut. chem. Ges. 362, 2086 (1903).

[71 R. B. Sosman, The properties of silica, (Chemical (26] 0. Brill, Uber einige Erfahrerngen beim GebrauchCatalog Co., New York 1927). der Mikrowage fi•r Analysen, Ber. deut. chem. 38,

£8] J. Mbore, Fused silica in industry, J. Soc. (Cem. 140 (1905).Ind. 50, 671 (1931). (27] E. H. Reisenfeld, H. F. M61ler, Eine neue Mikro-

[91 H. V. Neher, in J. Strong, Procedures in experi- wage, Z. Elektrochem. 21, 131 (1915).mental physics, (Prentice-Hall, Inc., New York [281 T. S. Taylor, A detennination of the density of

1939) (lap. V, VI. helium by means of a quartz microbalance, Phys.[10] Anon., Fused vitreous silica-its properties and Rev. 10, 653 (1917).

uses, Ceramics 2, 181 (1950). [291 A. Stock, G. Bitter, Gasdichtebestimmungen mit£111 E. Eberhardt, H. Kern, H. Klumb, Untersuchungen an der Schwebewage, Z. rhysik chim. 119, 333 (1926).

Quarzfaden, Z. angew. Phys. 3, 209 (1951). D3OI W. Cawood, H. S. Patterson, Atomic weights of C,N and F by microbalance method, Phil. Trans.Roy. Soc. (A] 236, 77 (1936).

6.2 Applications of Silica Fibers [311 E. A. Gulbransen, Vacuum microbalance for studyof chemical reactions on metals, Rev. Sci. Inst.

a. Review Articles 15, 201 (1944)..32] H. H. Podgurski, A microbalance for oxidation

(121 F. Ezuich, Methoden der Mikrochemie, Zweiten Teil: rate studies of Aluminum, Phys. Sci. Develop.Quantitative Method, Handbuch der Biologischen Shops, U. of Chicago (1945).Arbeitsmethoden, Abderholden, Urban und [331 E. A. Guibransen, New developments in the studySchwarzenberg, Berlin (1923) p. 183. of surface chemistry, Metal Progress, 49, 553

£13] F. A. Gould, Balances, Dict. Appl. Phys. 3, 107 (1946).(1923). [34] J. T. Stock, M. A. Fill, A direct reading micro-

[141 G. Gorbach, Somnelieferat: die Mikrowaage, Mi- balance for preparative work, Metallurgia, 37-8,krochemie 20, 254 (1936). 108 (1947-8)

[151 B. B. Cunningham, Micro-chemical methods used in [351 F. C. Edwards, R. R. Baldwin, Magnetically con-nuclear resetrch, Nucleonics 5, 62 (1949). trolled quartz fiber microbalance, Anal. Chem.

£161 G. Ingram, The suboicro balance and its applica- 23, 357 (1951).tions, Metallurgia 40, 231, 284 (1949). [36] R. S. Bradley, A silica micro-balance; its con-

[17] I. M. Korenman, Y. N. Fertel'meister, Ultramicro- struction and manipulation, and the theory ofbalance, Zavodskaya Lab. 15, 785 (1949). its action, J. Sci. Inst. 30, 84 (1953).

[181 E. Singer, Etat actuel de la technique de la mi-crobalance, J. phys. Radium 12, 534 (1951). d. Beam-Knife Edge and Torsion-Restoration

h. Salvoni Type Balance [371 B. D. Steele, K. Grant, Sensitive microbalancesand a new method of weighing minute quantities,

[191 E. Salvoni, Misura di masse comprese fra g 10-1 Proc. Roy. Soc.(London) [A] 82, 580 (1909).e g 10-6, Messina (1901). £381 B. D. Steele, Attempt to determine changes in

[203 C. B. Bazzoni, Loss of weight of musk in a current weight accompanying disintegration of radium,of dry air, J. Franklin Inst. 180, 463 (1915). Nature 84, 428 (1910).

18

[391 R. W. Grey, W. Ramsay,Density of Niton and Dis- [60) A. H. Weber, Sister Gonzaga Plantenberg, Rapidentegration theory, Proc. Roy. Soc. 84, 536 and direct measurement of vapor pressure of liq-(1910); 86, 276 (1912). uid metals, Phys. Rev. 69, 649 (1946).

(401 W. Ramsay, Les mesures de quantitie infinitesimales (611 R. C. ninn, H. H. Pomeroy, A vapor-pha3e sorptionde matieres, Comp. rend. 151, 126 (1910); J. study of Iodine and active MgD utilizing thephys. 1, 429 (1911). McBain-Bakr spring balance, J. Phys. Coll. Chem.

[411 W.BRamsayR. W. Gray, Molecular weight of radium 51, 9%1 (1947).emanation, Engineering 90, 423 (1910). [621 1. Sheft., S. Fried, Quartz spring Jolly balance,

[421 F. W. Aston, A simple form of micro-balance for Rev. Sci. Inst. 19, 723 (1948) AECD-2082.determining the densities of small quantities of [631 W. 0. Milligan, W. C. Simpson, G. L. Bushey,gases, Proc. Roy. Soc. 89, 442 (1913). H. If. Rachford, R. L. Draper, Precision multiple

[431 H. Pettersson, A new micro-balance and its use, sorption-desorption apparatus, Anal. Chem. 23,Gateborgs, Kungl. Veteaskaps och Vitterhets- 739 (1951).Sanballes, Handlingar 16th series (1914) Thesis,U. of Stockholm (1914). f. Miscellaneous Applications

[441 H. Pettersson, R. Strcmberg, A new kind of micro-balance, Instrument Fabriks A. B. Lych, [641 H. Gaudin, Sur les proprietes du cristal de rodeMaImtargagatan 6, Stockholm, Sweden (1918). fondu, Compt. rend. 8, 678 (1839), from

[451 H. Pettersson, Experiments with a new micro-bal- Shenstone 1901.ance, Proc. Phys. Soc. London 32, 209 (1920). [651 C. V. Boys, On the Newtonian constant of gravity,

[461 E. J. Hartung, Observations on the construction Phil. Trans. Roy. Soc. (London) 186, 65 (1895).and use of the Steele-Grant microbalance, Phil. [661 R. Threlfall, J. A. Pollock, On a quartz threadMag. 43, 1056 (1922). gravity balance, Phil. Trans. Roy. Soc. (London)

[471 R. A. Staniforth, The Kirk-Craig quartz fiber mi- [A] 193, 215 (1899).crobalance (Model A), AEC report N-2112 (Sep- [671 A. Gautier, Sur les aptiareils en quartz fondu,tember 17, 1945). Compt. rend. 130, 816 (1900).

[481 S. J. Gregg, M. F. Wintle, An automatically re- [68] B. E. Brauer, The construction of a balance, Inc.cording electrical sorption balance, J. Sci. Soc. of Inspectors of Weights and Measures,Inst. 23, 259 (1946). Edinburgh (1909) p. 189.

[49] P. L. Kirk, R. Craig, J. E. Gillberg, R. Q. [69W 0. W. Richardson, Note on gravitation, Phil. kag.Boyer, Quartz microgram balance, Ind. Eng. Chem. 43, 138 (1922).(Anal. Ed.) 19, 427 (1947). [701 F. A. Lindemann, A. F. Lindemamn, T. C. Kerley, A

[50] H. M. El-Badry, C. L. Wilson, The construction new form of electrometer, Phil. Mag. 47, 577and use of a quartz microgram balance. Roy. (1924).Inst. Chem., Repts. Sympos. Microbalances, No. [711 F. Auerbach, W. Hort, Handbuch der Physikalischen4, 23 (1950). and Technischen Mechanik, Johann A. Barth,

[51] I. Eyraud, Automatic adsorption balance, J. Chem. Leipzig 6, 113 (1928).Phys. 47, 104 (1950). [721 E. Wiesinberger, Die Anwendung der elektromag-

[521 P. L. Kirk, Quantitative Ultramicroanalysis (John netischen Mikrowaage bei der ausf~hrung vorWiley & Sons, Inc., New York, 1950) chap. 4. Ruckstandsbestimmingen u Flektolysn, Mikrochemie

[53] H. Carmichael, Design and performance of a fused 10, 10 (1931).silica microbalance, Can. J. Rhys. 30, 524 [731 E. A. Johnson, W. F. Steiner, An astatic magne-(1952). tometer for measuring susceptibility, Rev. Sci.

(541 J. A. Kuck, P. L. Altieri, A. K. Towne, The Inst. 8, 236 (1937).Garner heavy duty quartz fiber micro balance, [741 C. C. Lauritsen, T. Lauritsen, A simple quartzMikrochim. Acta 3, 254 (1953). fiber electrometer, Rev. Sci. Inst. 8, 438

(1937).e. Springs [751 H. Klumb, H. Schwarz, Uber ein absolutes Manome-

ter zur Messung niedrigster Gasdrucke, Z. Phys.[551 J. W. McBain, A. M. Bakr, A new sorption balance, 122, 418 (1944).

J. Am. Chem. Soc. 48, 690 (1926). [761 A. Base, A new crystat of the gas flow type and[56] P. T. Newsome, McBain-Bakr balance for sorption quartz fiber microbalance adopted for magne-

of vapors by fibrous and film materials, Ind. crystalline measurements, Ind. J. Phys. 21, 274Eng. (lem. 20, 827 (1928). (1947).

[571 J. W. McBain, H. G. Tanner, A robust microbalance [771 R. H. Barnard, Glass fibers research development,of high sensitivity, for weighing sorbed films, Glass Ind. 33, 144 (1952).Proc. Roy. Soc., (London) [A] 125, 579 (1929). [781 K-D. Mielenz, E. Sch&nheit, Zur Theorie des

[581 A. J. Stamt, S. A. Woodruff, A convenient six- Quarzfadenmanometers, Z. angew. Phys. 5, 90tube vapor sorption apparatus, Ind. Eng. Chem. (1953).(Anal. Ed.) 13, 836 (1941). [791 H. V. Neher, An automatic ionization chamber,

[591 G. H. Wagner, G. C. Bailey, #. G. Eversole, De- Rev. Sci. Inst. 24, 99 (1953).termining liquid and vapor densities in closed [801 L. Parker, B. J. Zack, Silica fibers, U. S.systems, Ind. Eng. Chea. (Anal. Ed.) 14, 129 (1942). 2,624,658, Jan. 6, 1953 L. Parker, Method of

19

forming batts of silica fibers, U. S. 2,635,390, [101 A. King, C. G. Lawson, J. S. Tapp, G. H. W~atson,April 21, 1953. Manufacture of helical silica springs, J. Sci.

[81] E. W. Gelewitz, H. C. Thomas, Adsorption studies Inst. 12, 249 (1935).on clay minerals III A Torsion pendulum adsorp- [1023 T. J. O'Donnell, Drawing and working quartz fi-tion balance, Rev. Sci. Inst. 25, 55 (1954). bers, Argonne National Laboratory, Chicago,

Ill. March 1946, reissued-Central developmentshop, University of Chicago.

6.3 Production and Fabrication [103] M. M. Haring, Notes on quartz fiber drawing sup-Methods plementing Argonne manual on "Drawing and work-

ing quartz fibers ", Info. rep. 35, MonsantoChemical Co., Central Re.search Dept., Dayton,

a. Fused Silica Ohio (July 30, 1947).[104] P. L. Kirk, B. Craig, Reproducible construction[82] B. S. Hutton, (On the fusion of quartz in the o urzfbrdvcs e.Si nt 9

electric furnace, Trans. Am. Electrochem. Soc. 77 (1948).

2, 105 (1902).(1948).[83] A. L. Day, E. S. Shepherd, Quartz glass, Science [1051 P. L. Kirk, F. L. Schaffer, Construction and

23, 670 (1906). special uses of quartz helix balances, Rev.

[84] F. Bottomley, Fused silica*, J. Soc. Chem. Ind. Sci. Inst. 19, 785 (1948).

36, 577 (1917). [106] L. Singh, A note on the construction of silica

[85] H. George, Survey of methods of fusing raw springs, J. Sci. Ind. Research [A] 11, 63

quartz and shaping resulting products, Bull. (1952).

soc. encour. ind. natl. 127, 373 (1928). [107] F. M. Ernsberger, C. M. Drew, Improvements in

[86] J. Fortey, Fused quartz manufacture in Germany, design and construction of quartz helix bal-

B.I.O.S. report No. 202 (1947). ances, Rev. Sci. Inst. 24, 117 (1953).

[87] R. B. Ladoo, Clear fused quartz, Mining Met. 28,501 (1947). 6.4 Properties of Fused Silica

[88] B. A. Rogers, W. J. Kroll, H. P. Holmes, Produc-tion of fused silica, Electrochem. Soc. 92, 115 and Glass(1947).

[89] C. Martinez, Influence of small quantities of a. Review Articlesimpurities on the fusion of quartz, Bull. inst.verre 6, 14 (1947). [108] A. Blackie, On the behavior of fused silica at

[901 F. A. Kurlyankin, The influence of pressure on high temperature, Trans. Faraday Soc. 7, 158the removal of bubbles in quartz glass during (1911); Collected Researches NPL 8, 139 (1911);melting, Steklo i Keram. 8 (1951). Chem. News 104, 77, 86 (1911).

[109] H. L. Watson, Some properties of fused quartz

b. Fused Silica Fibers and Apparatus and other forms of silicon dioxide, Bull. Am.Ceram. Soc. 9, 511 (1926).

[91] C. V. Boys, The attachment of quartz fibers, [110] B. Moore, Influence of characteristics andPhil. Mag. 37, 463 (1894). treatment of the raw material on the properties

[92] W. W. Coblenz, Investigations of infra-red spec- of fused silica products and the effect oftra: Note on blowing fine quartz fibers, Car- treatment of the fused product on their prod-negie Inst. Publ. 65, 126 (1906). ucts, Trans. Brit. Ceram. Soc. 32, 45 (1932-3).

[93] H. J. Denham, The construction of simple micro- [111] G. Slater, Strength and physical properties ofbalances, J. Textile Inst. 15, 10 (1924). fine glass fibers and yarns, J. Am. Ceram. Soc.

[94] T. C. Keeley, The preparation and silvering of 19, 335 (1936).quartz fibers, J. Sci. Inst. 1, 369 (1924). [112] G. W. Morey, The properties of glass, ACS Mono-

[95] K. Sliupas, Spiral springs of quartz, Nature graph,(Reinhold Publ. Corp., New York, N. Y.115, 943 (1925). 1938).

[96] C. V. Boys, Note on spiral springs of quartz by [113] C. J. Phillips, Glass: The miracle maker,(Pit-Sliupas, Nature 115, 944 (1925). man Pubi. Corp., New York, Chicago 1941).

[97] H. D. H. Drane, Quartz helical springs, Phil. [114] F. W. Preston, Mechanical properties of glass,Mag. 5, 559 (1928). J. Appl. Phys. 13, 623 (1942).

[98] J. S. Tapp, A convenient mechanical means of [1151 W. C. Hynd, Physics and the glass industry, Sci.winding quartz spirals, Can. J. Research 6, 584 J. Roy. Coll. Sci. 17, 80 (1947).(1932). [116] W. A. Weyl, Chemical aspects of some mechanical

[99] H. W. Weinhart, A machine for winding springs properties of glass,Research 1, 50 (1947).of vitreous silica, Rev. Sci. Inst. 4, 350 [117] G. 0. Jones, Viscosity and related properties in(1933). glass, Repts. Progr. Phys. 12, 133 (1948-9).

[100] L. M. Pidgeon, Improved method for construction [118] J. M. Stevels, Progress in the theory of theof quartz spirals, Can. J. Research 10, 252 physical properties of glass (Elsevier Publish-(1934). ing Co., Amsterdam 1948).

20

£119] R. N. Haward, The strength of plastics and glass [1431 G. 0. Jones, What experiments are needed in(Cleaver-Humie Press, Ltd. London 1949). glass science?, J. Soc. Giass Technol. 32, 382

£120] G. 0. Jones, F. E. Simon, What is a glass?, En- (1948).deavor 8, 175 (1949). [1441 H. O'Daniel,Zur Struktur und Nbntgenogxaphie der

£121] J. E. Stanworth, Physical properties of glass Silikatgl~serGlastech. Ber. 22, 11 (1948).(COxford University Press 1950). [145] J. E. Stanworth, On the structure of glass, J.

(122] J. Gialan,4lla, Fused quartz, Chem. Eng. 60, 302 Soc. Glass Technol. 32, 154 (1948).(1953). [1461 J. E. Stanworth, The ionic structure of g]ass,

[123] F. E. Wright, Fused quartz, a versatile indus- J. Soc. Glass Technol. 32, 366 (1948).trial material, Materials & Methods 37, 98 [147] G. Pincus, Influence of structural chemistry on(1953). glass composition,Ceram. Age. 53-54, 299

(1949).

b. Theory of Glass-Structure, Constitution [1481 W. H. Otto, F. W. Preston, Evidence against anoriented structure in glass fibers, J. Soc.

[124] R. W. Wyckoff, G. W. Morey, X-ray photographs of Glass Technol. 34, 63 (1950).glass, J. Soc. Glass Technol. 9, 265 (1925). (1491 1. Peych9s, The nature of the vitreous state.

[125] E. Berger, Contributions to the theory of glass Glass Ind. 32, 17, 77, 123, 142 (1951>'formation and the glassy state, J. Am. Ceram. Silicates Ind. 15, 77, 96 (1950).Soc. 15, 647 (1932). [150o F. Singer, Nature of glasses, glazes and hard-

£126] W. H. Zachariasen, The atomic arrangement in ness, Glass Id. 31, 241, 305, 366 (1950).glass, J. Am. Chem. Soc. 543, 384 (1932); J. [1511 R. Suzoki, A study of silicate melts, Tech. Rep.Chem. Phys. 3, 162 (1935). Tohoku U. Sendai 15, 11 (1950).

[127] B. E. Warren, X-ray determination of the struc- [1521 P. Bremond, What is glass?, Glaces et Verres 24,ture of glass, J. Am. Ceram. Soc. 17, 241) (1034). 24 (1951).

£1281 B. E. Warren, H. Krutter, 0. Morningstar, Fou- [153] W. E. Houth Jr., Crystal chemistry in ceramics,rier analysis of X-ray patterns of vitreous VIII, The crystal chemistry of glass, Bull. Am.Si0 2 and B203, J. Am. Ceram. Soc. 19, 202 Ceram. Soc. 30, 203 (1951).(1936). £1541 H. Moore, M. Carey, Limiting composition of bi-

£129] B. E. Warren, X-ray determination of the struc- nary glasses, J. Soc. Glass Technol. 35, 44ture of liquids and glasses, J. Appl. Phys. 8, (1951).645 (1937). [1551 T. Moriya, The constitution of glass, Japan Sci.

[130] B. E. Warren, J. Biscoe, The structure of silica Rev. 1, 61 (1951).glass by X-ray diffraction studies, J. Am. f156] A. G. Smekal, On the structure of glass, J. Soc.Ceram. Soc. 21, 49 (1938). Glass Technol. 35, 411 (1951).

£1311 B. F. Warren, Geometrical considerations in £157] A. E. Prebus, J. W. Michener, Structure in sili-glass, J. Soc. Glass Technol. 24, 159 (1940). cate glasses, Bull. Am. Phys. Soc. (May 1,

£132] B. E. Warren, Summary of work on the atomic ar- 1952).rangement in glass, J. Am. Ceram. Soc. 24, 256 [158] R. Yokota, Color centers in fused quartz, J.(1941). Phys. Soc. Japan, 7, 316 (1952).

£133] B. E. Warren, The basic principles involved in £159] A. Winter-Klein, Notions generales sur la forma-the glassy state, J. Appl. Phys. 13, 602 tion du verre, Verres et refractaires 6, 271(1942). (1952).

£134] B. E. Warren, X-ray diffraction study of glass, [160] B. K. Banerjee, X-ray study of glass fibers, J.Chem. Rev. 26, 237 (1948). Am. Ceram. Soc. 36, 294 (1953).

£135] N. A. Shishakov, Crystals of quartz glass, [161] R. T. Brannen, Further evidence against the ori-Compt. rend. acad. sci. URSS, 23, 788 (1939). entation of structure in glass fibers, J