glass working ocr

8/14/2019 Glass Working OCR

1/192

C O N T E N T S* #

+ *

FOREWORDACKNOWLEDGEM ENTSPREFACE1. INTRODUCTION . .2. PHYSICAL AND CHEMICAL PROPERTIES

# *

+ +

3.

* * * 4 * * 4 4

fr

# *

The structure of glassGeneral physical properties of glassGeneral chemical properties of glassDevitrificationThermal strain in glassAnnealing of glassSome types of glass for general useSome types of glass for sealing to metals . .Some types of glass for special purposes . .The ease of working different types of glassDistinguishing between different types of glassGLASS-WORKING EQUIPMENTThe glass-blower's tableBlowpipesHand torchesCylinder heads and valvesBellowsAir blowers . .Glass-working toolsWax for toolsCarbonsGauges for measuringRubber capsRubber stoppersBlowing tubesThe uses of asbes tos..Glass-blowing spectaclesGlass knives and diam ondsTube-cutting device V l l

-I


2/192

C O N T E N T SClampsAbrasivesGlass holdersLarge tube suppo rtsAdjustable rollersGlass-blower's swivelPolarized light strain-viewerGlass-cutting wheelsLapping wheelCarborundum grinding wheelAnnealing ovenTreatment of slight burnsThe amount of equipment required for different purposes

4. BASIC GLASS-WORKING OPERATIONSThe preliminary p reparation of materialsThe cleaning of glass tubes . .Breaking glass tubesKnocking off sm all bits of tubingHolding and rotating tubes in the bench flameBends and spiralsPutting a handle on the end of a tubeDrawing tubes off to spindlesRo und-bottom ed closures of tubesFab rication of thin glass rodBlowing holes in glass tubesJoining tubes of similar sizesJoining tubes of different sizesT-jointsY-joints4- and 5-way junctionsW orking capillary tubesBlowing bulbs in the middle of a tubeBlowing bulbs at the end of a tubeMultiple perforations in bulbsFlanging, flaring or bordering tub ingInternal sealsThin glass windowsSpinning out feetTapering glass tubesSealing-in sintered glass discsMending cracksDetection and removal of leaks and holes

* 4

* *

PAGS393940414242434444454546464848484951525355565758596064656868687071727273777879798081

+ * *vin


3/192

C O N T E N T S

ayers

Grinding glassReleasing frozen glass surfacesPolishing glassDrilling holes in glassCementing glassSilvering of glassDepositing copper on glassMetallizing preparations for the firing on of metalon glass

3. THE MANIPULATION OF LARGE TUBINGGeneral observationsBending big tubing . .Draw ing off large tubesClosing large flasks . .Joining big tubesBig T-jointsBig interna l sealsMending cracks in large tubingMending cracks near complex an d large seals

6. SOME OPERATIONS W ITH A GLASS-WORKINGMACHINE . .7. METAL-TO-GLASS SEALSMatched and m ismatched sealsSealing tungsten into PyrexSealing tungsten into special glassesMultiple wire seals . .Prepared coppe r-tungsten-nicke l wiresSealing platinum into soda or lead glass

Sealing platinum into PyrexSpecial alloys for sealing to glass . .Co pper-to-glass seals8. THE CONSTRUC TION OF SOME TYPICAL SINGLEPIECES OF EQUIPMENTAmpoulesSealed tubes for reactionsBreak-tip sealsDistillation flasks ..

Distillation splash headsDew ar sealsDewar vesselsColdfingerrefrigerant traps

PAGE82828383848587879090909192939495969697

102102103105105107108109110113117117118119120121121123125

IX


4/192

C O N T E N T S

* # # * J *

* *

* #

* *

* *

V *

Filter pumpsPhosphorus pentoxide trapsCondensersFrac tional distillation condensersFractionating columnsA fractional distillation receiverDouble surface condensersMercury cut-offsMcLeod gaugesTraps to catch mercury displaced in apparatusValvesSoxhlet extraction apparatusMercury vapour pum psElectrolytic gas generator .Therm ostat regulatorsGas flow metersSpectrum tubesDischarge tube lampsApparatus con taining many electrodesLeaks for m olecularflowof gases . .Bourdon gauges and glass spiral gaugesCirculating pumpsApparatus for semi-micro qualitative analysisGas analysis appara tus

9, THE ASSEMBLY OF COMPLEX APPARATUSGeneral observationsJoining closed systemsMaking more than one joint at onceRemoval of strain and clamping of apparatusAnnealing by flameUse of a bent blowpipe and double tipping deviceDanger of flame cracks

10. THE MA NIPULATION OF SILICA . .The properties of fused silica and general remarksGeneral technique for working fused silicaSilica torsion fibres ,.AUTHOR INDEXSUBJECT INDEX 4 *+ # # +* t

* #

+ ** # # *

P A G E126127128131132132133135136140140142144149150152153154156158159160162163166166167168169170170171173173.174175179181

x


5/192

A C K N O W L E D G E M E N T SWe thank the following Com panies and ind ividua ls:Aero Research Ltd., Duxford, Cam bridge, for information on cementsfor glassTlie British Heat-Resisting Glass Company Ltd., Phoenix Works,Bilston, Staffs., for information about Phoenix glassTile British Thomson-Houston Company Ltd., Rugby, for infor

mation about glasses manufactured by themEdwards High Vacuum Ltd., Manor Royal, Crawley, Sussex, forinformation about their glass-working machine and for permissionto reproduce FIGURES 34 and 35A. Gallenkamp & Com pany Ltd., 17-29 Sun Street, Finsbury Square,London, E.C.2., for information about the Davies double surfacecondenser, and for permission to reproduce FIGURE 4The General Electric Com pany Ltd., Osram Glass W orks, East Lane,Wembley, Middlesex, for information about glasses manufacturedby themlam es A. Jobling & C ompany Ltd., Wear Glass Works, Sunderland,for information abou t Pyrex glassJohnson, Matthey & Company Ltd., 73-83 Hatto n Garden, London,E.C.I., for information about preparations manufactured by themfor the production of fired-on metallized layers on glassJohn Moncrieff Ltd., North British Glass Works, Perth, Scotland,for information about Monax glassPlowden & Thom pson Ltd., Dial Glass W orks, Stourbridge, Worcs.,for information abou t glasses manufactured by themStone-Chance Ltd., 28 St. James's Square, London, S.W.I., for information about glass-working burners m anufactured by themThe Thermal Syndicate Ltd., Wallsend, Northumberland, for information about VitreosilWood Brothers Glass Company Ltd., Borough Flint Glass Works,Barnsley, Yorkshire, for information about Firmasil glassD. W. Bassett and J. A. Stone of King's College, London , for readingthe original manuscript and making many suggestionsV. J. Clancey for information on the method used by him for makingfused silica fibres.

X I


6/192


7/192

P R E F A C E J-A MBO for working glass is frequently encountered in scientificUIMrch and teaching, particularly in the field of physical chemistry.h many laboratories this problem is solved by the employment ofprofessional glass-blowers, and the research worker requires little or1 0 skill in the manipulation of glass. Or it may be that a laboratoryhiS no glass-blower and the scientist has to rely on his own resources,Ombined perhaps with the services of some outside specialist. InOther laboratories the scientist may find the available services to beilfltient in various respects, or overloaded, so that more satisfactoryprogress is made when he himself becomes an am ateur glass-worker.Tllli possibility was emphasized by W. A. Shenstone in 1889, whenitt wrote that the amateur with practice can make almost all the^apparatus he needs for lecture or other experiments with a consider-rlole saving in expense 'and, which very often is more important,out the delay that occurs when one depends upon the pro-

onal glass-worker.' Th is latter advantage is, he writes, forIf a very weighty one.W e became interested in working glass ourselves when engaged inous researches in the Chemical Laboratories at King's College,don, and over the last few years we have become increasinglyvinced that the scientific glass-worker can use methods which'er from those of the professional. The scientist is primarilyted in apparatus which gives him the results he seeks for, andthis apparatus lacks elegance in appearance and is made byods which are looked upon unfavourably by the professionalblower, it by no means follows that the apparatus is defectivea scientific point of view.W e have therefore compiled this work in the hope that it will beto scientists faced with problem s of glass-working. By avoid-the more difficult manipulations involved in professional glass-king, it seems possible for a scientist to assemble quite complexatus, including, for exam ple, his own m ercury diffusion pum ps,spectrometer tubes, molecular beam generators and silica; the preliminary practice required, which depends upon in-c ability, may take some hours a week for a period of two orm onths. We have also endeavoured to bring together data

t * in

xin


8/192

P R E F A C Escattered in the literature, and to give an accountmore difficult methods of manipulating glass.

The researches which led to our interest in glass-working weresupported by grants from the Research Fund of the University ofLondon, from the Institute of Petroleum and from the Departmentof Scientific and Industrial Research (maintenance grants to A. J. Cand J. D.). A. J. B. R.December, 1956 D . J. F.

J . D .

xiv


9/192

Chapter 1I N T R O D U C T I O N

GLASS has been known to man kind for some thousands of years. Themanufacture and working of glass became a stable industry in Egypti t the beginning of the 18th dynasty. W. M. Flinders Petrie (1909)hfti described how, about 1370 B.C. in the time of Amenhotep IV, inDie works at Tell-el-Amarna, a lump of glass was patted into aCylinder and then rolled into a rod which was drawn into cane about| In, thick. This cane was wound on a mandrel to produce hollowLater the art of pressing glass into open moulds was dis-OOvered. Th e blowpipean iro n tube 4-5 feet long with a mouth-pteoe at one endwas probably discovered about the beginning ofChristian era, and, according to G. W. MOREY (1938), caused anIndustrial revolution. The use of tongs for manipu lating glass wasknown to the Romans in A.D. 300. By this time, therefore, severalOf the basic methods which are now useful in constructing complexislentific apparatus had been developed. The glass in those earlywas similar in many respects to a modern soda-lime-silicag)tM. Scientific method does no t seem to have influenced glass pro -igttCtion very much before the present century , although of importan tiarly investigations we may mention those of K. W. Scheele andA. L. Lavoisier on the durability of glass exposed to water and weak d d s , and those of W . V. Harcourt and M. Farad ay on the productionAlld properties of glass.

The value of glass as a labora tory material is very great. Aoderately skilled worker can fabricate complex glass equipmentMing simple tools; and perhaps of even greater value in research isthe ease with which complex glass apparatus can be modified andidded to with little or no dismantling . Glass is sufficiently chemicallyinert for most purposes, and vitreous silica may be used when extra-Ordinary inertness is required . The transparency of glass is oftenValuable. Glass is a good electrical insulator, and m etal electrodes inft glass envelope can be raised to incandescence by eddy-currentsinduced by a coil, outside the envelope, carrying high-frequencyCurrent. In vacuum researches glass is valuable because of the easeWith which leaks are found with a Tesla coil, and on account of its

1


10/192

I N T R O D U C T I O Nready outgassing on bakin g. For nearly all practical purposes glassis vacuum-tight. G lass-to -m etal seals of various kinds may be made,and different types of glass can be sealed together with appropriateintermediate glasses.The manipulation of glass is a craft and has been handed downover the centuries largely by personal example and tra d iti oa Venicewas the most important centre of glass-working for the four centuries following its rapid development in the eleventh century to adominating position; in 1279 a system of apprenticeship needingeight years was introduced there, and the closed and somewhat secretive nature of the craft was reinforced by the employm ent of assassinsto remove glass-blowers who seemed likely to give away valuablesecrets to other centres of the industry. Perhaps the aura of mysterywhich still to some extent surrounds scientific glass-working can betraced back to these times.The basic techniques developed for scientific glass-working involvethe manipulation of glass in the form of tubing or rod , using a flameas a source of hea t. The article to be fashioned is held by the op eratorin both hands, and the part to be worked becomes a semi-moltenmass in the flam e and is shaped principally by blowing. Generally itis necessary to rotate the article in the flame, and this often requiresa very high degree of muscular co-ordination, which can be developedby professional workers to a remarkab le extent. This rotation operation is not easy; M . FARADAY (1842) stated the outstanding difficultyin the following words: 'But when the heat has brought the glassinto a soft state, it is by no means easy so exactly to turn the tube atbo th ends alike, and so lightly yet equally to hold them, that the softpart shall retain its cylindrical shape; being neither twisted, nor bent,nor elongated, nor thrust up .' R. E. THRELFALL (1946) considers thatone third of the a rt of glass-blowing consists essentially in being ableto move both hands about, rotating a tube with each finger andthumb, and keeping constant both the distance between the handsand the speed of rotation . Considerable practice is necessary to gaingreat mastery of this technique, which we can call the bench-flamemethod of work, since the flame is in a fixed position on the bench.The problem facing the scientist who has to engage in his own glass-working is that of simplifying or modifying those operations whichdepend on extremely good m uscular co-ordination. We consider thisproblem to be soluble when elegance can be sacrificed to u tility.The requirements of a scientist differ from those of a professionalglass-worker. Th e starting point of a new research is new ideas, andwhen these require subsequent experimental investigation the apparatus used need be only good enough to give results which are

2


11/192

M-N -

I N T R O D U C T I O Nsatisfactory from a scientific po int of view. A glass appara tus must, infact, hold together and work, but it may have an unsightly appearance. However, we must also note that experimental skill in itself isof great value to the scientist in pursuing his ideas, for as Faraday(1842) wrote: The person who could devise only, without knowinghow to perform, would not be able to extend his knowledge far, ormake it useful; and where the doubts or questions that arise in themind are best answered by the results of an experiment, whateverenables the philosopher to perform the experiment in the simplest,quickest, and most correct man ner cannot but be esteemed by him asof the utm ost v alue.' In addition to these considerations, experimental skill is of particular value in developing new techniqu es; and

may lead to many unexpected developments. H . E. ARMSTRONG(1924) tells how the introduction in 1865 of the Sprengel pum;fairly simple piece of glass-workingrevolutionized the Englishwater supply. Again we may note the im po rtan t practical consequences following on the early work of H. L. CALLENDAR (1892) onVitreous silica tubes, and that of C. V. Boys and R. Threlfall onvitreous silica fibres.In view of the difference in objective between the scientist andprofessional glass-worker, we consider that glass-working for theicientist should develop as its own craft, and indeed this has happened to a certain extent. An early systematic account is that of J. J.BERZELIUS (1833), which describes some of the basic operations.Faraday's Chemical Manipulation is a masterly treatm ent, referred tofrequently in R. THRELFALL'S notable work On Laboratory Arts(1898), which devotes 107 pages to glass-blowing and manipulationof g lass. Threlfall seems to have been one of the first to describewhat is now often called 1n-place' glass-blowing, in which the glasspieces are kept stationary by clamps, and the flame is moved. Jointsare made with thin glass rod, now often called welding rod, by amethod very similar to that used in 1370 B.C. in the Tell-el~Amarnafactory. The difficulty of holding and rotating the glass is avoided,and thus, as Threlfali says, the method is most useful to the experimenter who wants to get on to otter things before sufficient skill isacquired for the rotation method. The tradition of the craft of glass-working for scientists is continued, we think, by the works of W. A .SHBNSTONB (1889), T. BOLAS (1898), B. D. BOLAS (1921), F. C.FlUltY, C. S, TAYLOR and J. D. EDWARDS (1928), R. H. WR I GHT(1943), and J. D. HELDMAN (1946). Amongst eminent scientists whohave carried on the glass-working craft with their own hands we maymention R . W. Bunsen, whose skill at the oil-fed blowpipe and whoseplary patience when one of his pupils rapidly and several times

3


12/192

I N T R O D U C T I O Nin succession broke the apparatus glass-blown by the master himselfare described by H. E. ROSCOE (1901); J. Dewar, whose vacuum-jacketed glass vessels m arked a new era in low temperature research ;M. Bodenstein and W. Ramsay, both of whom are described by E.K . RIDEAL (1951) as uncan nily skilled in the art of glass-blow ing; andF . W . Aston, whose glass bulb discharge tubes and cooled c ha rco al-in-glass pumping systems first gave those exact mass relations ofisotopic chemistry which contributed decisively to the opening of theatomic age. In modern times the old tradition that the masters ofscience should also be themselves masters of the practical craft ofscience has failed to persist, and the more eminent scientists are nowusually no t to be found at the wo rking bench in the research labo ratory. There is, we think, a consequent weakening of the craft basisof practical scientific work, and some retreat from the view expressedby I. Langmuir th at wo rk in the laboratory can be fun.The increasing use in scientific research of borosilicate glasses overthe last thirty years has not, on the whole, been accompanied bymuch departure from the traditional methods of glass-working.Generally it is much easier to rotate the flame about a stationaryclamped article than to rotate the article in a stationary flame.Quite complex app aratus can be constructed by working with a handtorchgiving a movable flame held in the handon clamped apparatus. This method of work is specially suitable for the borosilicateglasses of low thermal expansion, which can be worked into quiteknobby apparatus without there being much danger of cracking oncoolingin contrast to sod a-lime-silica glass. The joining of twoclamped tubes with a movable flame has been described by SHEN-STONE (1889), THRELFALL (1898) and TRAVERS (1901). A X . REIMANN(1952) has described some further uses of this general method, butthe great number of complex operations which can be carried outwith facility using a hand torch in place of a bench torch does notseem on the whole to have been realized; indeed, HELDMAN writes:'End-seals with both tubes of approximately the same diameter andT-seals are, with practically no exceptions, the whole repertory ofin-place glass blowing.5 We are by no means in agreement with thisstatement. M any operations can be carried out with a hand torchon completely fixed glass, or on a fixed piece of glass to which someother part of glass can be joined by holding it in one hand whilst theother hand m anipulates the hand torch. The results are usually notas elegant as those obtained by a skilled worker using rotation in abench torch , but less skill and practice are required for the hand torchmethod . It has the further great advantage th at the complexity of theapparatus being constructed can be steadily increased without

4


13/192

I N T R O D U C T I O NfTMtly increasing the glass-working difficulty. Fu rtherm ore, thelOtatist will probably only work glass from time to time, dependingtpon the requirements of the research or other work, and the handtorch method is far more suitable than the bench torch method forthe operator who does not want to spend time on preliminarypractice; it is also far more suitable for the scientist working iniwkward positions on complex apparatus. We find that very com-pttft apparatus may be assembled with a hand torch.A clear accoun t of professional m ethods of glass-working has beengfvtn by W. E. BARR and V. J. ANHORN (1949). Valuable informationme glass-working problems is also found in the works of J.

FfftONG (1940), M. C. NOKES (1948), A. J. ANSUEY (1950), A. ELLIOTTMd J . HOME DICKSON (1951), and H. J. J. BRADDICK (1954).ounts of the simpler operations have been given by W. E.AUK-WINDER (1947) and E. H . MORGAN (1953).

R E F E R E N C E SttUY, A. J., 1950, An Introduction to Laboratory Technique, 2ndE dn ; London, M acmillan.IMTltONG( H. E., 1924, Chemistry in the Twentieth Century, edited

by E. F. Armstrong; London, Benn.p i , W, E . and ANHORN, V. J., 1949, Scientific and Industrial Glass* Blowing and Laboratory Techniques', Pittsburgh, InstrumentsPublishing Co .i , J, J. 1833, Traite de C himie, (Trans. Esslinger) Vol. 8;Paris, Firmin D idot Freres.B. D. 1921, A Handbook of Laboratory Glass-Blowing;London, Routledge.T. 1898, Glass Blowing and Working; London, Dawbarn andWard.CK, H. J. J., 1954, The Physics of Experimental Method;London, Chapman & Hall.>AR, H . L., 1892, / . Iron St. Inst., 1,164.; A. and HOME DICKSON, J., 1951, Laboratory Instruments;London, Chapman & Hall.V, M.t 1842, Chemical Manipulation, 3rd Edn; London,Murray.PBTRIE, W. M., 1909, The Arts and Crafts of Ancient Egypt;Edinburgh and London, Foulis.

fcAIY, F. C , TAYLOR, C. S. and EDWARDS, J. D., 1928, Laboratory1 Class Blowing, 2nd Ed n; New York, M cGraw-Hill.hLDMAN, J. D., 1946, Techniques of Glass Manipulation in ScientificI Btstarch ; New York, Prentice-Hall.5


14/192

I N T R O D U C T I O NMOREY, G. W., 1938, The Properties of Glass; New York, Reinhold.M O R G A N , E. H., 1953, Newnes Practiced Mechanics; issues of October, November and December.NOKES, M. C , 1948, Modem Glass Working and Laboratory Technique, 3rd Edn ; London, Heinemann.PARK-WINDER, W. E., 1947, Simple Glass-blowing for Laboratoriesand Schools; London, Crosby Lockwood.REIMANN, A, L., 1952, Vacuum Technique; London, Chapman &Hall.RIDEAL, E. K 1951, / . Chem. Soc, 1640.ROSCOE, H. E., 1901, Chemical Society Memorial Lectures 1893-

1900; London, Gurney & Jackson.SHENSTONE, W. A., 1889, Hie Methods of Glass Blowing, 2nd Edn;London, Rivingtons.STRONG, J., 1940, Modern Ph ysical Laboratory Practice; London andGlasgow, Blackie.THRELFALL, R., 1898, On Laboratory A rts; London, Macmillan.THRELFALL, R. E., 1946, Glass Tubing; London, British Associationof Chemists.TRAVERS, M. W., 1901, The Experimental Study of Gases; London,Macmillan.W R I G H T , R. H., 1943, Manual of Laboratory Glass-Blowing; Brooklyn, N.Y ., Chemical Publishing C o.

6


15/192

Chapter 2P H Y S I C A L A N D C H E M I C A LP R O P E R T I E S O F G L A S S

The Structure of GlassA OLAS5 is a product of fusion which has cooled to a rigid conditionWithout crystallizing. This definition includes a large number ofOrganic glasses, and does not restrict the term 'glass* to inorganicgubltances, which is a frequent practice in the U.S.A. This restrictionn t n u somewhat arbitrary, particularly when we consider how G.immann established the general principles of the glass-like state bylarch on organic glasses, and how the devitrification of technicahue s is paralleled by that of organic glasses. Tam mann concluded|tt a glass could be regarded as a supercooled liquid in which theNational movements of the molecules had been frozen (see W. E.,, 1952). In fact, as R. Boyle described it about 1660, 'theof the glass agitated by the heat, were surpriz'd by the coldthey could make an end of those motions which were requisite| their disposing themselves into the most durable texture.' Inpdarn terminology, a glass is thermodynamically unstable withto the corresponding crystal.lalline silica (quartz, tridymite or cristobalite) in its variouslions is built up of Si0 4 tetrahedra linked together in amanner so tha t every oxygen is between two silicons. Thera therefore share corners. The arrangem ent in space of the'a is different in the various crystalline forms, but is alwaysregular. A silica glass, in contrast, again contains S i0 4 tetra-with every corner sh ared; b ut by slight distortions of the valencycompared with the crystal, a continuous and irregular three-ional network is built up. The orientation about the Si-O -Siof one S i0 4 tetrahedron with respect to another can be practi-random. Thus a two-dimensional picture of a silica glass would

a series of irregular rings, with an average number of about sixra in each ring, but with the number of tetrahedra in indi-rings varying from three to ten or more. The silica glassUltflw the condition for glass formation proposed by W. H.7


16/192

P H Y S I C A L A N D C H E M I C A L P R O P E R T I E S O F G L A S SZACHARIASEN (1932), namely that the substance can form extendedthree-dimensional networks lacking periodicity, with an energycontent comparable with that of the corresponding crystal network.A glass does not, therefore, prod uce a regular diffraction pa tternwith x-rays; but a monochromatic x-ray beam incident on a glass isscattered, and a radial distribution curve may be constructed. Thespace average of the distribution of atoms round a given one can bededuced (see J. T . RANDALL, 1938), M uch work of this kind has beencarried out by B. E. W arren and his colleagues.A soda-silica glass results from the fusion of N a 3 0 w ith SiO a. Thenumber of oxygens is more than twice the number of silicons, andsome of the oxygens are bonde d to only one silicon, A silicon bond edto one of these oxygens is at the centre of a tetrahe dro n w hich sharesonly three corners w ith other tetrahedra . W ith each singly bond edoxygen there is associated one negative charge. The sodium ions arefound in the holes in the three-dimensional silicon-oxygen network.On the average, each sodium is surroun ded by abo ut six oxygens, andeach silicon by four oxygens. In a so da -bor ic oxide glass of low sodacontent the extra oxygen is bonded between two borons, and thereare no singly bonded oxygens. This can happen because in a boricoxide glass the co-o rdina tion of bo ro n by oxygen is triang ular, and inthe mixed glass some of the boron atoms become tetrahedrally coordin ated by oxygen. W hen there is more than ab out 13-16 per centof Na 2 0 in the glass, the boron atoms cease to change their co-ordination, and some singly bonded oxygens exist (B. E. WARREN,1942).Soda-silica glasses are not formed when the soda content exceedsthat given by the formula Na 2 S i0 3 . For this form ula, if every siliconatom is surrounded tetrahedrally by four oxygen atoms, then on theaverage two oxygens round every silicon are singly bonded, and acon tinuou s n etwork is just possible. W ith still mo re oxygen it is notpossible. In a soda-s ilica glass with much less soda , there are manyS i0 4 tetrahedra sharing every corner, and a number sharing onlythree corners. The way in which these different tetrahedra are distribu ted is not yet quite clear. Th ere may be small regions where allthe tetrahedra share four corners, and such regions are compo sed ofpure silica; they may alternate with regions of, for example,N a 2 0 2Si0 2 . The composition may vary through the glass whensufficiently small regions are considered .

The general picture of a glass as a negatively charged irregularframework containing holes with positive ions in them enables adistinction to be made between the network-forming ions, whichcomprise the framework, and the network-modifying ions which go8


17/192

G E N E R A L P H Y S I C A L P R O P E R T I E S O F G L A S Sn the holes. Silicon, bo ron an d phosphorus are im po rtan t network-brming ions. Sodium and potassium are important network-nodifying ions. Other ions can act in both capacities. This isprobably true of aluminium , beryllium, zinc, iron and titanium . In aad- si lie a glass it seems that lead atoms can take pa rt in the networkind link S i0 4 tetrahed ra together. Co balt ions in network-modifyingpositions tend on heating to move into the network, and this canMUM a colour change from pink to blue.Most commercial glasses are based on silicates or borosilicates. Atypical hard borosilicate glass for chemical work may contain 80 perWit Si0 2 f 12 per cent B 2O s and 4 per cent N a aO . A soft soda-lime -lllica glass (usually referred to as soda glass) may contain 70 per3tnt St 0 8 ,1 7 per cent Na aO and 5-4 per cent CaO. Lead glasses, usedfor lamp and valve stems, may contain 30 per cent PbO, 57 per centWOt, 5 per cent Na sO and 7 per cent K 2 6 . These glasses have highritctrical resistance. Glasses with exceptionally high softening temperatures contain 20-25 per cent of A1203. Borate glasses, sub-Htntially free from silica (8 per cent Si0 2) are used for sodiumUncharge lam ps.General accounts of the structure of glass have been given by J. E.ITANWORTH (1950), B. E. WARREN (1940) and C. J. PHILLIPS (1948).

General Physical Properties of Glass[ht physical properties of a given specimen of glass may dependipon the previous history of the specimen. This is particularly thefor the mechanical strength under tension, when the surface pre-iM tment of the specimen is of decisive importance. The therm alnsion and viscosity of glass also depend to some extent on theiiy of the specimen. The importance of this factor has beenphasized by A. E. DALE and J, E. STANWORTH (1945). R, W.LAS (1945) has given a valuable review of the physical pro of glass.

f0Chanical Strengthimportant property for the practical worker is the strength of| m under tension. The surface of glass very probably containslUMrous extremely small cracks extending into the glass, and whentensile stress is applied there is a concentration of stress at the ends)f time cracks, which causes them to grow further into the glass,Bttil at some crack breakage occurs and is propagated through theto iin en . Glass usually breaks in a direction at right angles to the~"~n of maximumfrom

9


18/192

i'

P H Y S I C A L A N D C H E M I C A L P R O P E R T I E S O F G L A SSfibre; but touching a new fibre,even w ith thefingers,greatly weakensit. An old fibre is actually strengthened by removing the surfacelayer with hydrofluoric acid, even though the cross-section is reduced.The strength of a glass under tension varies from one specimen toano ther. A further complication is the variation of the tensilestrength with the time for which the stress is applied. A tensile stresswhich does not cause fracture after a short time of application maydo so after a lone time. There is in fact a delaved fracture of glass.

j>

.1

0-007

Breaking stress kg/nun:FIGURE 1. Nature of therelation between time of loadingand breaking stress for glass. This property is of astatistical nature, and the particular curve shown can

only be taken as representativeThe nature of the relation between time of loading and breakingstress for a borosilicate or soda-lime-silica glass is shown in FIGURE1. The curve given must be taken as representative only. A typicalfigu re for the safe tensile strength for prolonged loading times is0-7 kg/mm2 (1000 lb/in .2). Similar results for the relation betweenbending stress and time to fracture are found when a tensile stress isproduced by bending a glass rod into an arc of a circle. FromFIGURE 1 we note that an increase of stress by a factor of 4 reducesthe time required for fracture by over 106 times. If a certain load issupported for one hour by a certain piece of glass, one quarter of theload should be supported for a million hours. This can be made use

10


19/192

G E N E R A L P H Y S I C A L P R O P E R T I E S O F G L A S SOf in testing the extent to which stress can be applied to a glassipp aratu s. Four times the stress the glass must support can beapplied for a short time.The delayed fracture of glass, shown in FIGURE 1, must be borne inmind in assembling app aratu s. If, for example, an appara tus isdamped so that bending stress is introduced, the apparatus maybreak after a long interval. Chemical reactions at the surface of theglass may be partially responsible for delayed fracture: C GURNEYami S. PEARSON (1952) found a soda-lime-silica glass to be strongerin vacuum, and to be weakened by carbon dioxide and water in thesurrounding atmosphere.Thermal ExpansionThe coefficient of linear thermal expansion is almost constant, formost types of glass, for temperatures up to 400~-600C. The actualvalue depends on the chemical constitution of the glass. It thenIncreases rapidly above a certain temperature, often called the

to o 00 300Temperature m SCO VFIQVEM 2. A typical expansion curve for a hard borosilicate glass(Phoenix)

transformation point'. This is not, however, a characteristic temperature, since it depends on the thermal history of the specimen andthe rate of heating. At a higher temperature the glass softens andMases to expand. This is sometimes called the 'softening temperature* or the 'M g point'. Confusion may result from another definition of softening temperature, depending upon the rate of extensiona fibre by viscous flow. This latter softening temperature, whichs to a viscosity of 107*8 poises, is very much higher thanMg point. A typical linear expansion curve for a borosilicateglass (Phoenix) is shown in FIGURE 2. A is the transformation point

d B the Mg point. The temperature corresponding to A is oftenOalled the iower annealing temperature', and corresponds to aViscosity of about 1014 poises; tha t corresponding to B is often calledthe *upper annealing temperature', and corresponds to a viscosity of11

i i .


20/192

P H Y S I C A L A N D C H E M I C A L P R O P E R T I E S O F G L A S Sabout 1G12 poises. An other definition in comm on use, especially inthe U.S.A., is to call the 'annealing temperature' that at which theviscosity is 1013*4 poises. This is then between A a,n&B on the thermalexpansion curve. It is useful to bear in mind the confused state ofterminology and definitions when using tabulated data on thethermal properties of glass. The viscosities mentioned above are,perhaps, not established with certainty. At the lower annealingtem pera ture, annealing is actually extremely slow. This temp eratureis not used for the practical annealing of laboratory app aratus.Thermal EnduranceThis measures the ability of the glass to stand sudden changes oftemperatu re without fracture. W hen a specimen of glass is suddenlyheated uniformly over all its surface, the heat penetrates slowly intothe interior. The outside layers are heated first, an d being unable toexpand fully they become subject to a compressive stress, whilst theinner layers become subject to a tensile stress. When the specimen ata uniform temperature is suddenly cooled over all its surface, thesurface layers are subject to tensile stress. Since the mechanism offracture usually involves surface cracks, glass is more likely to breakon sudden cooling than on sudden heating. The m agnitude of thestress produced on sudden cooling depends on the modulus ofelasticity and the coefficient of linear thermal expansion, and, in away not important in practice, on Poisson's ratio. Thermal endurance is measured by somewhat empirical methods, and is again astatistical quan tity. A heat-resisting glass is one having a highthermal endurance; a hard glass has a high softening temperature.A 1-mm thick beaker of a hard borosilicate glass, such as Pyrex,Phoenix or Firmasil, will require a thermal shock, by sudden cooling,of abou t 325C to give appreciable probability of fracture. Fo r asoda-lime-silica beaker the corresponding figure is about 120C.Beakers of Monax glass stand a much greater thermal shock thanthe soda-lime-silica beaker; the beakers of standard thickness canusually survive a thermal shock of 24O~250C. Thick glass fractureswith less thermal shock th an th in glass.The glass-worker subjects tubing to thermal shock by suddenlyplacing it in a ho t flam e. T he inner surface of the glass tube is then notheated directly, and is very quickly subjected to tensile stress. The hardborosilicate glasses as tubes can usually be placed immediately in anoxy-coal gas flame without fracture, but complex apparatus, especially when internal seals are present, requires more gentle heating.Soda-4ime-silica glass tubes need gentle warming a t first, particularlywhen the end of a tub e which has not been fire-polished is pu t in the

12


21/192

G E N E R A L P H Y S I C A L P R O P E R T I E S O F G L A S SBamc. The end of a tube is fire-polished by fusing it in a flame, andthis process closes up surface cracks. Vitreous silica has very greatthermal endurance: small red-hot articles can be quenched in waterwithout cracking.Thermal ConductivityFor the hard boros ilicate glasses and the soda-lim e-silica glasses thisft aroun d 0-0025 cal C"1 c m - 1 sec1. For vitreous silica ( Vitreosil) inthe transparent form it is 0-0025 up to 500C, and 0-0035 from 500to 1000C; for the translucen t form it is 0-0033.Viscosity and Softening TemperaturesThese properties have already been mentioned in connection withthermal expansion. The viscosity decreases rapidly with increasingtemperature. A linear relation is found between the logarithm of theviscosity an d the reciprocal of the absolute tem perature. This isConvenient for extrapo lation. W hen the viscosity has th e value 107"6poises the glass is mobile enough to be drawn into threads, and thetemperature is sometimes called the softening tem perature (see p . 11).At temperatures between the lower and upper annealing temperatures (A and B in FIGURE 2) the viscosity can change with t i m e -When the glass is suddenly cooled the viscosity slowly increases to anquilibrium value and when the glass is heated the viscosity slowlydecreases to an equilibrium valuein fact time is required for thequilibrium viscosity values to be attained . Glass is often wo rkedWhen its viscosity is about 104 poises; for a hard borosilicate glassthis corresponds to a temp erature of about 1200C.Elastic Propertiesfyrex Chemical Resistance Glass has a Youn g's mo dulus of 6-1 x 1 0 niynes/cm 2, a modulus of rigidity of 2-5 x 1011 dynes/cm 2 and aPoisson's ratio of 0-22. Similar values ar e found for oth er glasses.The extension of an amorphous material under a tensile force canbe resolved into three parts; first, an immediate elastic extension,Which is immediately recoverable on removing the tensile force;lecondly, a delayed elastic extension which is recoverable slowly ; an dthirdly, a plastic extension, viscous flow, or creep, which cannot berecovered. With glass at ordinary temperatures, this plastic exten-tlon is practically absent. A very slow delayed elastic extensionoccurs. This effect can be troublesome in work with torsion fibres.The delayed elastic effect in vitreous silica fibres is 100 times less thanto Other glass fibres, and visa) us flow of silica is negligible below100*C (N. J. T IG H E , 1956). For exact work vitreous silica torsionfbres are therefore used.

13


22/192

P H Y S I C A L A N D C H E M I C A L P R O P E R T I E S O F G L A S STherma l Capacity therm al250C is given by C (calgm- 10C) =0-174 + 0-00036t where t is thetemperature in C.Electrical ResistanceTh e resistance of vitreous silica (Vitreosil) in the translucent form atroom temp erature exceeds 2 x 1014 ohm cm. Glasses containingmetal ions in network-modifying positions are ionic con du ctors. Ina sod a-lime-s ilica glass, for example, the curren t is carried by so diumions and the resistance at 150C may be around 108 ohm cm.Lemington W.L, a hard borosilicate glass, has a resistance of about1010 ohm cm at 200C. A typical lead glass, Wembley L.I., has a verymu ch greater resistance both at room temperature and norm al lamp-operating temperatures tha n a soda~4ime~silica glass, and is thereforevaluable for lamp and valve pinches. The resistance of LJ. at 150Cis 1012 oh m cm. Generally the volume resistance due to ionic conduction decreases rapidly with tem perature. The logarithm of theconductivity is a linear function of the reciprocal of the absolutetem pera ture. Th e surface of mo st glasses is very hydrophilic, an dthere is a surface conductivity which depends upon the relativehumidity. For Phoenix glass, for exam ple, the volume resistance of acentimetre cube at room temperature is about 3 x 1014 ohm, but thesurface resistance at 60 per cen t relative hum idity is 7 x 1011 ohm, andat 81 per cent relative humidity it is 5-4 x 109 ohm. In very humidatmospheres it is possible to have an electrical shock by touch ing thesurface of a soda glass apparatus containing electrodes at highpo ten tial. Th e water layer on the glass becomes slightly alkalineafter a time by reaction with sodium from the glass; the apparatusshould be wiped from time to time with a cotton cloth. It is best inthese cases to use a borosilicate glass. The surface conductivity ofglass was discovered by M . FARADAY (1830).DensityFor the soda-lime-silica glasses this is about 2-5 gm/em 3; for theboro silicate glasses it is very nearly 2*25 gm/cm 3 and hardly changeswith slight variations in composition. Wem bley LJ. lead glass has adensity of 3-08. A very dense lead glass has a density of 5-2.HardnessGenerally glasses with a high silica content are more resistant toabrasion than low silica content glasses. The hardness thereforeincreases with increase of softening temp eratu re. Lead glasses canbe scratched quite easily.

14


23/192

E N E R A L P H Y S I C A L P R O P E R T I E S O F G L A S STransmission of LightA 1-mm thick sheet of Phoenix glass will transmit 90 per cent or m oreOf the light incident on it, for wavelengths of 350 millimicrons toAlmost 2 microns. In the infra-red region a strong absorption occursat 3 microns and little transmission beyond 4 microns. In the ultraviolet region increasing absorption occurs as the wavelength fallsbelow 350 millimicrons and very little transmission occurs below 270millimicrons. Th e transp aren t variety of vitreous silica (fused quartz)has very superior optical properties, and is widely used in photochemical and op tical researches. In the ultra-violet region it trans mits at high efficiency down to 1850 Angstrom units (185 millimicrons). 'Quality O.H. VitreosiV of The Thermal Syndicate Ltd isA special optical quality in which the absorption ban d at 2400Ang strdrns has been eliminated. A special quality of fused qua rtz isalso available which transmits infra-red up to 3*5 microns approximately ('I.R. quality VitreosiV of Th e The rma l Syndicate L td). In thisVitreosil the absorption band at 2*7 microns has been much reduced.The Stress-Optical CoefficientI t IS no t usu al to take quan titative m easuremen ts of the strain in glassapparatus made for research; when a strain-viewer is used (p. 43)qua litative observations are normally made. Qu antitative measure-Kits can be made when the stress-op tical coefficient is know n. Theinvolved requires a knowledge of the optical behaviour ofdoubly refracting materials and depends on the fact that a ray ofp!anepolarized light ente ring strained glass is brok en in to two raysthe 'ordinary ray' and the 'extraordinary ray'vibrating at rightllgles to each othe r. Fo r glass subject to simple axial tension o rCompression, the extraordinary ray vibrates in the plane whichincludes the axis of the stress. The birefringence of strained glass isproportional to the strain, and thus to the stress. The stress-opticalOOefficient is the maximum double refraction or birefringence ob-gerved in polarized sodium light for 1 cm path length when there is aUniform stress of 1 kg/cm2. It is expressed either in wavelengths ofOdium light or in millimicrons. This coefficient varies from oneglass to another; it is around 3*5 millimicrons, or 0*006 wavelengthsof sodium light A. JOHANNSEN (1918) has given an account ofmethods for determining double refraction, and very valuable datafor practical work are given by J. H. PARTRIDGE (1949).

General Chemical Properties of GlassResistance to Chemical ActionsVittoous silica is the most chemically inert glass for most purposes.

15

M -- ij ^ ' t


24/192

P H Y S I C A L A N D C H E M I C A L P R O P E R T I E S O F G L A S SIt is not affected by halogens or acids, except for phosphoric andhydrofluoric acids. Phosphoric acid attacks fused silica at temperatures of 300-400C, and hydrofluoric acid attacks it at room temperature, forming silicon tetrafluoride an d wa ter. A t high tem pera turessilica reacts with caustic alkalis, certain metallic oxides, and somebasic salts, and cann ot be used for incinerating these m aterials. Ov er1600C, fused silica is reduced to silicon by ca rbo n. It can also bereduced at high tem pera ture by hydrog en. It is unaffected by wa terunder normal conditions but is attacked by strong solutions ofalkalis.The hard borosilicate glasses are highly resistant to attack bywater; but just as the sodium ions in the glass are slightly mobileunder the influence of an electric field (p. 14), so also they can bemobile by thermal agitation and escape from the glass into water incontact with it and be replaced by hydrogen ions. This effect isslight: for example, a Firmasil beaker in an autoclave containingwater at 150C loses about 0-00015 gm of sodium per dm 2 in fourhours. A soda-lime-silica glass loses sodium to water at a muchgreater ra te. The resistance of borosilicate glass to mo st acids is verygood, but strong aqueo us alkalis produce visible attack. The network of triangles and tetrahedra is attacked, so the glass tends todissolve as a whole. So da- lim e-silica glass usually has less chem icalresistance than a borosilicate glass. Alkaline attack, however, becom es mu ch greater o n glasses with high silica con tent. Alkalis canalso leach ou t bo ric oxide from a borosilicate glass. Hy drofluoricacid dissolves glass, and glacial phosphoric acid attacks most kindsof glass.The Weathering of GlassA reaction between sodium from the glass and atmosp heric water andcarb on dioxide can lead to the formation of sodium ca rbon ate, w hichcrystallizes in fine needles. A pota sh glass forms potas sium carbon ate, w hich is too deliquescent to crystallize out . A lead glass ca nreact with hydrogen sulphide, and to a smaller extent with carbondioxide, sulphur dioxide, and acid vapours.Phenomena Arising from the Heating of GlassA rapid evolution of adsorbed water first occurs on heating glass;this is followed by a persistent evolution, due to gas (mostly water)diffusing from the interior. Ab ove 300C the two processes are fairlyclearly separated. The adsorb ed water is rapidly and completelyremoved, and the quantity of gas evolved by the persistent evolution


25/192

G E N E R A L C H E M I C A L P R O P E R T I E S O F G L A S Sis proportional to the square root of time. The process has anactivation energy. For a soda~lime~silica glass over 98 per cent ofthe evolved gas is wa ter. B. J. T O D D (1955) has studied these effectsin detail. The ad sorbed w ater on glass can be troublesom e in gaseousmanipulation , as R. W . Bunsen first appreciated.A t high temp eratures glass loses its more volatile com po nen ts. Th eloss of silica, lime, magnesia and alumina is negligible, but boricoxide, lead oxide, sodium oxide and po tassium oxide can also be lost.W hen the glass is hea ted in a flame, reaction may occu r with some ofthe flame gases; sulphur dioxide can react with soda glass and leadglass to form sodium sulphate and lead sulphate respectively, and ofthese only the former can be washed off. A n acco unt of these effectsis given by W. E. S. TURNER (1945). The loss of weight of vitreoussilica on ignition is negligible; crucibles can be heated to 1050C, andprecipitates can be ignited at 100GC in crucibles with a porous baseof v itreous silica.Diffusion through GlassThe mobility of the sodium ions in a soda-lime-silica glass at elevated tem perature s is fairly hig h; if an evacuated bulb of such a glassis dipped into m olten sodium nitrate and electrolysis is bro ug ht abo utby bom bard ing the inside of the bu lb with electrons, the circuit beingcompleted with an electrode in the sodium nitrate, then metallicsodium appears in the bulb . By immersing the bulb in other m oltensalts the sodium ions can be replaced by ions of silver, copper,thallium and van adiu m. These ions also diffuse into glass from theirmolten salts in the absence of an electric field. W hen potassium isdistilled in a borosilicate glass vessel it becomes slightly co ntam ina tedwith sodium which diffuses from the glass and is replaced by potassium (D. K. C. M A CD O N A L D and J. E. STANWORTH, 1950). Vitreoussilica allows helium, hydrogen, neon, nitrogen, oxygen and argon todiffuse throug h it, with the permeab ility decreasing in the o rder given.Th e permeability of silica becom es greater if the glass devitrifies. Thepermeability to helium of soda-lime-silica glass is 105 (or more)times less than that of vitreous silica. Fo r practical vacuum pu rposessoda and borosilicate glasses can be regarded as impermeable togases at ordinary temperatures, except in work at extremely lowpressures when the diffusion of atmo spheric helium through the glassmay become significant.

The permeability of glass at high temperatures seems to have beendiscovered by R. Boyle. In his collected works published in 1744there is a paper in Volume III 'A discovery of the perviousness ofglass to ponderable parts of flame' in which he writes \ . . it is plain17


26/192

P H Y S I C A L A N D C H E M I C A L P R O P E R T I E S O F G L A SSthat igneous particles were trajected through the glass, which agreeswith the Epic ure an s.. . 'Some Surface PropertiesTh e chemical properties of glass are largely determined by its surface,because the metal ions present in most types of glass are in fact reactive enough, bu t only those at the surface can react. The very slowrates of diffusion in glass at ordinary temperatures limit its reactivity, so that, as M. FARADAY (1830) wrote 'Glass may be consideredrather as a solution of different substances one in another, than as astrong chemical compound; and it owes its power of resisting agentsgenerally to its perfectly compact state, and the existence of aninsoluble and unchangeable film of silica or highly silicated matterupon its surface.* Th e surface com position of glass may be verydifferent from the bulk com position, for volatilization occurs duringthe forming process, and weathering occurs subsequently; bothprocesses produce a surface resembling vitreous silica. It is possiblethat the Si0 4 tetrahedra on the surface terminate in OH groups towhich adsorbed water is normally bou nd. The glass surface can bemade hydrophob ic instead of hydrophilic by allowing adsorbed waterand surface hydroxyls to react with various monoalkyldichloro-silanes (RHSiCl2), when hydrogen chloride is formed by elim inatioaThe surface properties of glass are of great importance in manyreaction kinetic studies, particularly those involving the terminationof reaction chains on the walls of the vessel. W hen a glass reactionvessel is used in such cases, it is usually found that it must bematured, by carrying out a number of reactions in it, before reproducible results can be obtained . In many cases, reproducible resultsare only obtained when the vessel is kept continuously at the reactiontemperature, and exposed only to the reaction mixtures; if theapparatus is cooled, and air let in to make an alteration or repair,different results may be obta ined subsequently. This is particularlythe case in oxidation reactions.

DevitrificationThis is the process of the crystallization of one or more of the constituents of glass. Generally a glass is thermodynam ically unstablewith respect to these crystals, but at ordinary temperatures thecrystallization ra te is quite negligible. Crystallization may occurwhen the glass is worked at high tem perature. The crystals whichappear in a supercooled melt are not necessarily those of the stablesolid phase a t the temperature concerned: for example, cristobalite canappear at temperatures for which tridymite is the stable crystalline

18


27/192

I

D E V I T R I F I C A T I O Nof silica. The crystals which are most likely to separatesoda-lime-silica glass are those of calcium silicate, tridymite

id cristobalite. Calcium silicate occurs usually as the m onoclinicinn wollastonite, bu t sometimes in a hexagonal form. The m ono-ic form tends to appear as very long, thin crystals bunched to-lgBther to give a brush-like appearance. The hexagonal form is notpcicular. Tridym ite is hexagonal, and forms hexagonal stars , likemow, by twinning at 60. The cristobalite forms twins at 90. Devitrification on the surface of glass seems to depend upon loss of alkal-oxides, and may be assisted by dust particles.Transparent vitreous silica (transparent Vitreosil) is liable to de-vitrify if potassium o r sodium compounds are present. The surfaceof the material must be thoroughly cleaned, and the part to beworked should then not be touched w ith the fingers.When devitrification is observed in the soda and borosilicate typesof glass, the crystals may sometimes be removed by fusion of theglass in the flame; alternatively the semi-molten devitrified portionmay be removed with tongs, and replaced by fresh glass added as thinrod. If there is a large extent of devitrification the portio n of glassshould be completely cut out and replaced. Old soda glass apparatusis very liable to devitrify when repaired: F. C. FRARY, C. S. TAYLORand J. D . EDWARDS (1928) advise, in such cases, a preliminary washing of the glass with dilute hydrofluoric acid to remove the surfacelayer.

Thermal Strain in GlassWhen a block of glass is suddenly heated on all its faces, the outerlayers are under compression and the inner layers are under tension,as explained on p. 12. This strain is temporary, however, since itvanishes as soon as the temperature gradient vanishes. Tem porarystrain is similarly produced when the surface of a block of rigid hotglass is cooled. Above the upper annealing tem perature (p. 11) astress can only exist in glass for a short time, because the glass flowsto relieve the stress. Stress is relieved only very slowly at the lowerannealing temperature. Thus over a certain temperature range glasschanges from a viscous to a rigid body. Consider a block of glass tobe rapidly cooled through this temperature range, so that there isalways a temperature grad ient. In the viscous region the glass isstrain-free, and thus when it first becomes rigid it is also strain-free.Thus there is a rigid block of glass containing a temperature gradientbu t free from strain. When this tem perature gradient is removed, theinner layers of the glass are in tension and the outer layers in compression. There is then a permanent strain in the glass. Clearly, when

c 19


28/192

P H Y S I C A L A N D C H E M I C A L P R O P E R T I E S O F G L A S Sa com plex glass object is cooled rapidly there willfinallybe a complexdistribution of permanent strain, and this can be great enough tocause fracture of the glass.The drops of glass produced by Prince Rupert of Bavaria bydropping molten glass into oil become rigid when there is a greattemperature gradient and the surface is consequently in strong compression. This makes the drops very strong, bu t they disintegrateviolently as soon as the tail of the drop is cut off, when the internalstresses are no longer balanced.

Annealing of GlassThe object of this process is to prevent permanen t strain arising fromthe cooling of glass. The glass must be cooled slowly throu gh thecritical temperature range in which it becomes rigid and ceases torelieve internal stresses by viscous flow. The rate at which thesestresses are relieved in the annealing range of temp eratures (A to B inFIGURE 2) depends on temperature; when this is such that theviscosity is 1013*4 poises , the glass will becom e practically stress-freein 15 minutes (A. E. DALE and J. E. STANWORTH, 1945). Below thelower annealing tem perature the glass can be cooled quickly withoutintroduction of permanent strain, but the temporary strain couldbecome great enough to fracture the article.Annealing is carried out most satisfactorily in an oven (p. 45).Complex articles of Pyrex glass can be annealed at 560C for 30minutes, followed by slow cooling with the oven do or sh ut. Articlesof Firmasil glass should be annealed at 575C, but even at 475Cstrain is very slowly removed. For Phoenix, the upper annealingtemperature is 600C and the lower annealing tem perature is 520C.Annealing at 560-5 80C is therefore satisfactory for th is glass. Theannealing temperatures of these borosilicate glasses are not at allcritical. The article must not, of course, be made too ho t, or it willdeform. Annealing is of great importan ce for articles made of asoda-lime-silica glass. Wembley X.8. soda glass should be annealedin the range 520-400C, and the General Electric Company, whichmakes this glass, recommends annealing at a high temperature of520C for 5-10 minutes, followed by cooling to an intermediatetemperature of 460C at a rate dependent upon the glass tubingthickness. These rates are:3C per minute for J mm wall thickness

2C per m inute for 1 mm wall thickness1C per m inute for 3 mm wall thickness.The glass should be cooled from the intermediate temperature of460C tc a low temperature of 400C at double the above rates. The20


29/192

A N N E A L I N G O F G L A S SArticle can then be cooled to room temperature at any rate possibleWithout cracking it by temporary thermal strain. The correspondinghl|h t intermediate and low temperatures for Wembley L.L lead glassIff 430C t 390C, and 340C. The same annealing schedule can bemid. For Wembley M.6. 'white neu tral' glass the temperature range| | $I0"45OOC. Again the same schedule should be used.Complex apparatus assembled on the bench must be annealed bylame, and this method must also be used when no oven is available(Mt p. 170). In our experience, very complex apparatus of Pyrex glasstail be flame-annealedsatisfactorily. Usually with Pyrex the apparatus either cracks in a day or two after making, or else not at all.Complex apparatus of soda-Kme-silica glass can be annealed byflame, but we do not find this satisfactory. Fo r research apparatus itIft beet to avoid this kind of glass. With a complex vacuum apparatusOf boroiilicate glass a fracture can often lead to unfortunate conse-quancet, especially when there are many mercury cut-offs present;iDd in such cases it is well, before evacuating, to wait for a few daysaltar a repair or alteration has been made in a position where flamea&nealing is difficult. J - BIIS Some Types of Glass for General UseMany different kinds of glass are made. In this Section and theMowing Sections we mention only a few of these which are useful|S the laboratory.Wmbley X.8.fllis is a soda-Hme-silica glass, containing magnesia and boric oxidethan 1 per cent), made by the General Electric Co. It is oftenbed as GEC X.8. or simply as X.8. The linear coefficient ofal expansion between 20 and 350C is 9-65 0-10 x 10~6. Thisis available as tubing and rod in a wide range of sizes.&). 1-5-cm diameter tubing, for the water circulation, isdrawn ou t to a taper at its end, and closed. 8-mm diameter tubing isprepared for the side arm A and 5 mm tubing for the water inlet andoutlet tubes B and C. The 3 cm tubing is rounded off and a holeblown jus t large enough to let the 1-5 cm tube slide into it. A norm alinternal seal is next made, but in this case the joint must be blownfrom both ends. The side arm A is then added, followed by the sidearm C. The larger tubing to which C is joined is then rounded off,blowing through C, and the central tube B is inserted in the usualway, but the internal seal is somewhat more difficult to make sincethe side arm C is already present. The whole apparatus requirescareful annealing.

Alternatively, the whole water-circulating part can be m ade separately and internally sealed into the wide tube, using a hand torch, and131


140/192

S OM E T Y P I C A L S I N G L E P I E C E S O F E Q U I P M E N Tblowing from both ends. An other design employs a Dew ar seal as inFIGURE 52, II , and the water-carrying tubes are inserted with a rubber

Fractionating ColumnsThese are devices for establishing contact between ascending vapourand descending liquid in a distillation. A variety of columns can bemade. Perhaps the simplest (FIGURE 60, II) consists of a supp ort in atube on which a packing rests, made up of short glass tubes, glassrings or pieces of metal gauze. The support for this packing is made

E

I m ir vFIGURE 60. Two types of fractionating column

by heating the tube with a small hot flame and pushing the wall inwith a spike almost to the centre of the tu b e; this is done a number oftimes to give the support shown in FIGURE 60, /. A thorough annealing should be given.Another fractionating colum n consists of a number of pear-shapedbulbs (FIGURE 60, F ) . A series of bulbs is blown (FIGURE 60, III), andtheir ends are heated and blown and draw n so that pear-shaped bulbsare formed (FIGURE 60, IV). The ends of these bulbs are heated (at Aetc.) and they are gently pushed in. The column shou ld then bethoroug hly annealed. The whole sequence of operations can becarried o ut very easily with a glass-working machine.

A Fractional Distillation ReceiverA receiver for low pressure distillations designed by G. A. R. KON(1930) is quite easy to make: the joints D and C (FIGURE 61bo th standard B.19 joints ; the two taps a t A andiJare standard designsof three-way taps. For the part G a tube of about 3 cm diameter

132


141/192

A F R A C T I O N A L D I S T I L L A T I O N R E C E IV E Ris joined at bo th ends to tube of abou t 8 mm diameter. One of thesmall tubes is joined to the tap B, To prevent distortion of thebar rel of this tap, the join t should be made 1-2 cm away from theAn internal seal is made into the cone D and a side armattached. The top of the internally sealed tube is then joined to thetap.B. The two tubes E and F ar e joined to the tap A. The socket C isthen joined to the larger tube .

6

F

FIGURE 61. A fractional distillation receiver

An advantage of this receiver is that receiving vessels connected atD can be taken off and p ut on w ithout altering the low pressure in themain apparatus. When tap B is shut during the operation of changinga receiver the distillate collects in G.Double Surface CondensersA simple form of double surface condenser th at can be m ade ^too grea t difficulty by an unskilled worker is shown in FIGUREThe parts seen in FIGURE 62, 7, must first be prepared.

The ring seal at A is made in the early stages and is pannealed . This pa rt of the assembly is then inserted in to the pitube B which is held clamped, and, when seated well, is itself c(FIGURE133


142/192

S OM E T Y P I C A L S I N G L E P IE C E S O F E Q U I P M E N Tup the whole area slowly and uniformly the ring seal at D is nowmade by tooling, a little at a time, the glass flare of tube B on to the

\ / /

c

(

D

8

WaferI E

IF WaferFIGURE 62. A simple form of double surface condenser

bulge of the inner tube. Care m ust be taken to keep the whole of b othring seals ho t while this is being d on e; a large flame should thereforebe used and only reduced to a sharp hot flame for the final blowingout. The water outlet tube E is added while the whole area is hot.134


143/192

D O U B L E S U R F A C E C O N D E N S E R SThis tube can be held by Jiand and a pin-point flame used to seal it inposition. Every few seconds theflamesize should be increased and thefar side of the assembly re-warmed before continuing the sealing ofE inposition . Finally the w hole section must be very carefully annealed.The assembly may now be clamped ready for the final large ringseal as in FIGURE 62 , III. Fo r this operation , and to a lesser extent forthat described in the previous paragraph, it is advisable to have asecond worker, also with a hand torch, who can keep the far side ofthe apparatus h ot while one side is being worked. Ex tra carefulwarming up is required at this stage before the large final ring seal canbe made. The warming should be commenced with hot air for ab out10 minutes, then with the tips of large luminous flames and finallywith slowly increasing oxygen content. It will be found with such aprocedure that the initial double ring seal will survive without cracking. The large ring seal is then worked in the same manner as the sealA and, with the entire section kept hot, the side arms F and G areadded. This region with th reeringseals needs extremely careful annealing. We have found, however, thatflame-annealing s adequate.

The completion of the outer water jacket, with ring seal at H andwater inlet tube J, is relatively straightforward. The tubes G and Cmay be joined w ith a short piece of condenser tubing, though a moreelegant finish to the condenser is achieved if they are bent and sealedto give a closed circuit as in FIGURE 62, IV . If this is done the upperring seals must be very carefully protected with asbestos paper, butotherwise the procedure \$ again straightforward.The design of this double surface condenser can be changed asdesired; for example, FIGURE 62, V, shows a double surface reflux condenser which may be made by modifying the above general procedure.The type of double surface condenser most often found in laboratories is the 'Davies improved double surface condenser' whichappeared in 1905, and was designed by J. Davies of A. Gallenkamp &Co. as a direct development from Thresh's modification of the Bidetcondenser. All rubber bung s and tubing were eliminated in theDavies condenser, which has retained its popularity for 50 years. Tomake a condenser of this kind the inner water jacket is connected tothe outer one by two internal seals which are made by a slightmodification of Method 2 described on p. 75. The rest of theassembly is similar to t ha t for an ordinary Liebig condenser.

Mercury Cut-offsMercury cut-offs are used in vacuum apparatus instead of taps whentap grease is undesirable. A frequently used type is shown in FIGURE63. A two-way tap is attached at the top of a reservoir for mercury.


144/192

S OM E T Y P I C A L S I N G L E P IE C E S O F E Q U I P M E N TThe tdown into the reservoir, but this necessitates very careful annealing.It is easier tofirst oin the capillary to no rmal tubing of equal externaldiameter, and then to seal this normal tube into the larger tube. Thetop pa rt of the cut-off consists of an internal seal at B and a side arm.This cart is connected to the reservoir bv about 70 cm of catrillarv

VacuumAtmosphere

FIGURE 63. A mercury cut-offtube, which gives strength and reduces the volume of mercury required. To use the cut-off the m ercury level is raised and loweredbetween C and D.W. E. BARR and V. J. ANHORN (1949) describe a mercury cut-offwith mercury return lines which prevent mercury surging into avacuum system.

McLeod GaugesThe McLeod gauge is one of the oldest instruments for the measurement of vacua and it has remained virtually unchanged since itsintroduction in 1874 (H. M C L E OD, 1874).A simple bench type McLeod gauge is shown in FIGURE 64 , III.Before starting to make such a gauge an estimate should be made of

136


145/192

M C L E O D G A U G E Sthe volume required to give pressure readings in the required range:a gauge with a volume of 300 ml and using 0*75-mm bore capillarymay be used to read pressures of 10"

5mm of mercury with reasonableaccuracy, bu t if relatively high pressures are to be measured it will bemore satisfactory to use wider bore capillary, say 2 mm . This avoidsdifficulties encountered with mercury sticking in the capillary. Thebore of the capillary should be determined accurately by directmeasurement along its length to ensure uniform boreunless

H I

vacuum

To atmospWe

[VFIGURE 64. A McLeod gauge for bench use

precision-bore tubing is used. This precision-bore tubing may becontaminated by the lubricant used to prevent it sticking to the metalmandrel used in its manufacture, and consequently great care shouldbe taken in cleaning it. Stripping the surface with hydrofluoric acidgives a uniform clean surface if normal cleaning methods fail.The simplest way of making a McLeod gauge is to make it in twobasic sections which are finally joined together. Since the capillaryon bo th parts of the gauge should be identical in bore and in surfaceproperties, it is test to make use of a single length of tubing so thatbo th capillaries are subjected to identical cleaning procedures. Thecapillary should be carefully inspected for particles of dirt which may137

JL


146/192

S OM E T Y P I C A L S I N G L E P I E C E S O F E Q U I P M E N Tnot have been removed by the cleaning. FIGURE 64 shows a gaugefitted to the mercury reservoir by a standard ground joint; this isusually the easiest method of assembly.The pa rt in FIGURE 64, // , can be made either by blowing a bulb ofsuitable volume or by using a round-bottomed flask. If the latter isused a slight bulge should be blown centrally in the bo ttom so that asharp shoulder is avoided at the capillary seal. A 15-20 cm length ofcap illary is sealed to this bulge and its upper end left open. Thiscapillary can then be used as a han dle an d blowing tube for the subsequen t work on the neck of the flask. The neck of theflaskshould thenbe drawn down, a smaller tube joined to it and cut off short to formthe inner tube defining the cut-off volum e. Alternatively, the cut-offvolume can be defined by sealing a flanged tube in to the neck of theflask, which then does no t have to be drawn down. The standardcone is then joined to the mercury supply tube b y a ring seal to theflared out end of the tube. This part is then joined t o the end of theflask as shown, allowing sufficient room between the two seals for theattachment of the side arm to the vacuum system. The two side armsshould be inserted while the glass is still ho t from w orking the internalseals and the whole section annealed thoroughly.

The next step is the calibration of the cut-off volume: this can bedone by filling with water and weighing the volume contained in thecapillary and bulb u p to the cut-off po in t; for small volumes mercuryis a better medium but becomes unmanageable for larger volumes.Once a satisfactory volume calibration has been obtained the end ofthe capillary is closed to a square end. This can be easily achieved bydrawing down a piece of glass rod until it just fits into the bore of thecapillary and cutting a short length of this rod to plug the end of thecapillary before fusing it. If this is done carefully the capillary will beclosed with a square end and the walls of the tube will not be distorted at the end of the gas column, which introduces uncertainties inreading this position. It is im portan t to avoid lens effects here ifaccu rate readings are to be obtained .The part of the gauge in FIGURE 64, 1, is simply constructed bylengths of normal tubing to either end of the length ofcapillary, bend ing these to a suitable angle and blowing o ut the ends.One end of this composite tube is then blocked with a stopper, andthe other end joined to the main tube at an angle by a norma l T-jointA hole is then blown in the main tube for the second T-joint and theshape of the bends in the tubing is adjusted by gentle heating andbending until the end is brou ght in to close contact with the hole. Thisjoin t is then completed. The parts / a n d / / ( F I G U R E 64) are then joinedso that the two capillaries are parallel and a few millimetres apart,138


147/192

M C L E O D G A U G E SA useful safety precaution is to draw the end of the tube dippinginto the mercury reservoir down to a capillary to retard the rate of

flow of mercury into the gauge if excessive air pressure is aaccidentally. The drawing down of the air inlet on the two-way stopcock makes the gauge much easier to operate. Various other methodshave been used for raising and lowering the mercury in the gauge butthe m ost convenient and most widely used is tha t shown the use ofatmospheric pressure and a low vacuum which can be supplied by anormal water pump if necessary. If this method is used it is advisableto have th e high vacuum connection a t more than atmospheric heightabove the mercury reservoir.

FIGURE 65. A small rotatingtype of McLeod gaugeThe range of a McLeod gauge may be extended by using varioussized tubes in conjunction with a small diameter capillary and similar

sized compensating tubes (L. DUNOYER, 1926). In this way comparatively high pressures may be determined on the same gauge as maylow pressures. The calibration technique is more complex as thevolume is not a simple fraction of the length of the compressed gascolumn (see FIGURE 64, IV).Small McLeod gauges have been developed for rapid pressuremeasurements: they avoid the tedious raising and lowering of a largevolume of mercury by the gauge rotating about a ground glass jointin the high vacuum connection. A gauge of this type is shown inFIGURE 65. The volumes used are relatively small and can be madeconveniently from short lengths of tubing; the reservoir bulb shouldbe large enough to permit free passage of gas into the gauge when allthe mercury is in the reservoir. The volume of mercury required isdetermined by the volume of the bulb and the volume of the tubing

139


148/192

S OM E T Y P I C A L S I N G L E P IE C E S O F E Q U I P M E N Tthe mercury level in the compensation capillary should be levelwith the top of the closed capillary when the gauge is in a verticalposition and under vacuum . These gauges are best calibrated againsta no rm al type of gauge.

Traps to catch Mercury displaced in ApparatusWhen parts of an apparatus contain mercury, as in cut-offs andM cLeod gauges, it very often happ ens th at air rushing in or out of thesystem will displace mercury in an undesirable manner. W ith cut-offsacross a differential pressure, bubbling of gas through mercury isunavoidable, and in tubing of internal diameter less than 3^4 mmbubbling will no t occur b ut the gas will raise a column of mercury infront of it. It is therefore des irable to insert anti-splash traps wher-

ii * S r/ mFIGURE 66. Trap for catching displaced mercury

ever bubbling is likely to occur. A simple form of trap is shown inFIGURE 66,777.In the construction a bulb of the required size is first blown andthen a bulge is formed at A (FIGURE 66, I): the tip of the bulge isblown out resulting in a hole which will just take the size of tubingbeing used. A short piece of tubing is bulged and bent to shape(FIGURE 66, II) and a ring seal is formed at A. Care m ust be taken tosee tha t the lower end of the inside tub e finally slopes downw ards sotha t d rops of mercury will no t lodge in it.

ValvesValves are of two kinds: those meant to close a system as far aspossible, and those designed to operate as check valves which reduceth eflow hrough a system.A simple and effective check valve which impedes the flow of m ercury is seen in FIGURE 67, L It consists of a small plunger floating onmercury, which is pushed into a seating when the mercury level rises.Th e plunger P is made by blowing an elongated bulb at the end of a

140


149/192

V A L V E Stube and then drawing out a neck. The bu lb is about half filled withmercury and the neck sealed off. Th e tube C is joined to the tube Aand the edge of the join at B is heated in places and pushed in to forma few spikes to hold the plunger (FIGURE 67, / / ) . The plunger is theninserted in C ; the end of C is rounded off an d D is joined. This mustbe done with the tubing held vertically, or nearly so, to prevent theplunger coming in contact with the hot g lass; the joint can be m adevery easily with a han d torch.Valves that are to close a system must have two surfaces whichm ake intim ate c on tac t; this is readily effected by grinding the plunger

A

B

C

M JV nFIGURE 67. Some kinds of valves

into its seating. A simple design is illustrated in FIGURE 67, / / / : theplunger is made by blowing a bulb o n the end of a small tube, heatingthe end of the bulb , touching it with a piece of hot glass, and pullingou t into a conical shape. The plunger must be able to pass down thetubing to be used as C. Th e tubing A is joined to B by a taper w ith anangle equal to that of the head of the plunger, and the plunger isground into this taper with fine carborundum until a continuousgrou nd ba nd 1-2 mm wide is formed on the glass. The cone is thenwashed and dried, and B is joined to C. The end of the plunger istapered as shown, mercury is added, and the plunger is sealed off andput into B through C. The junction of C and B is then push ed in at afew places so that the glass spikes just allow enough freedom ofmovement of the plunger.Instead of a glass plunger a ball bearing may be used, as in FIGURE67, IV and V. In bo th of these examples a larger tube is joined toa smaller tube and the joint is ground to fit the ball bearing. The141


150/192

S OM E T Y P I C A L S I N G L E P I E C E S O F E Q U I P M E N Tgrinding can be carried out with the ball bearing or with a glass grinding tool of diameter equal to that of the ball bearing. This tool is aglass rod with a sphere mad e on the end by rotating the end in a ho tflame. Th e glass sphere is ground into the join t, an d since bot h surfacesare ground away a slightly tapered seating is produced which is bettertha n tha t produced by direct grinding with the ball bearing. The ballbearing can float on the mercury as in FIGURE 67, F, or be kept inplace by a small indentation in the glass (FIGURE 67, IV ) made bypush ing it in with a spike. Th e former valve is closed by the mercuryrising, and the latter may be closed by mercury also, or by a greatenou gh flow of gas or liquid pushing the ball bearing into place. Inthis case the bulb shown in FIGURE 67, IV 9 may be replaced by a tubejus t bigger tha n the ball bearing.A no the r form of valve which we find useful is shown in FIGURE 67,VI, To make the plunger a sphere is made on the end of a glass ro d,and to this sphere a thin glass rod is attached, using a very smallflame so tha t the sphere is not distorted. The sphere is then removedfrom the larger glass rod with a small and very hot flame, and it isrotated with the thin rod until it is spherical. The glass sphere is nextground into the end of the tube to be used as B; this can be doneeasily by rotating the tubing with a lathe chuck and keeping thesphere stationery. The tube B is then internally sealed into C, theplunger is put in and finally C is rounded off and D is added. Th etail of the plunger should be long enough to prevent the plungercoming o ut from the tube B. A valve of this kind is useful for liquidsand ga ses ; it is opened by pressure of the fluid in B, but flow canno toccur from D to B.The apparently complex Soxhlet apparatus (FIGURE 68, III) is madeby th ree simple ope ration s: first the large tube A is joined to the lowermain tube H 9 then the vapour bypass is fitted, and finally the liquidsipho n is pu t on. The main tube is usually made from 3-cm diametertubing abou t 15 cm long, and this is joined to 1-5-cm diameter tubingH. Th ere is not a continuous tub e when these are joined, and hencethe jo in t m ust be blown from b oth ends. The joint is made by touching the small tube against the rou nded end of the large tube, and thendirecting the flame against the smaller tube, which is shrunk andblow n a few times, and then straightened ou t. Directing the flameagainst the smaller tube prevents the larger tube becoming distorted.A h an d torch can be used for this joint.

To m ak e the vap our bypass, 1-cm diameter tubing is bent with tworight angles, and the ends are cut off so tha t the tube jus t fits at C and142


151/192

SOX HLET EX TRACTI ON APPARATUSD and is parallel to the larger tube. Holes are then blown at C andD. One end of the vapour tube is closed w ith a stopper and the jointD is made; then, after removing the stopper, the upper right anglebend of the vapour tube is heated and the tube pushed until the lowerend meets the hole at C. The end of the vapour tube is then heatedand pushed with a spike on to H until there are only a few smallholes; the joint is finished in the usual way. The top angle is thenmade smooth and the whole annealed (FIGURE 68, /). It is advantageous to have a bunsen burner handy with which the top joint iskept hot while the lower joint is made.

/ /y*

X 9lb*

EA

SM

. A

6H M

For the siphon tube some 2-3-mm diameter tubing has a circularbend made a t one end through about 180, and then about 8 cm fromthis bend the tube is ben t sharply again through 180 (FIGURE 68, II).The straight end of the tube is bent about 3 cm below the end of thecircular bend, into a right angle in line with the circle. The end is cutjr tubes as shown. Anotherpiece of the small tube is bent in to a right angle to make (J. This pieceis internally sealed into H, and afineho t flame is directed against thecentre of the join , which is blown out to a bubble. G is m ade central,and the siphon tube FIE is joined to G by the appropriate internalseal procedu re: E is closed with a stopper. This is removed, a holeis blown in the large tube A (for joining on E) and, w ith F heated, E ismanoeuvred into position. The join t is then m ade with the help of aspike as usual. F is then m ade smooth and the whole annealed.

143


152/192

S OM E T Y P I C A L S I N G L E P I E C E S OF E Q U I P M E N TIn making the second joint E it is helpful if the ends of the tube tobe pushed to a hole with a spike are thickened and slightly flangedbefore the first joint is made . This is especially the case for sm all

diameter tubing. The holes blown for bo th the bypass and the siphonshould be so placed that both side tubes are parallel with the maintube.Mercury Vapour PumpsMercury diffusion pumps are standard pieces of laboratory equipment and, for most cases, quite adequate ones can be constructedfrom glass. In general, Pyrex glass (or similar glass) is necessarybecause hot mercury vapour circulates and differential strain is

S \\ \ G High vacuum

FIGURE 69. A simple diffusion pumprapidly set up at various, sometimes com plicated, join ts. Below,however, is described a very simple glass diffusion pump which isquite satisfactory for many purposes, and which can easily be constructed from soda or other glasses.Very Simple Glass Mercury Diffusion PumpThis pum p (FIGURE 69) is due to H . P. WARAN (1923) and is designedsuch tha t it may be constructed by the most average of experimentalists. Soda glass is satisfactory because all joints are subject to aminimum, if not a zero, temperature gradient. Pyrex glass would, ofcourse, be an advantage but is not a necessity.The boiler A can be made from an ordinary laboratory conicalflask: a side arm is added and then the top of the flask is drawn offand roundedif a labo ratory filter suction flask is available, and hassufficient thermal endurance for a boiler, the side arm already existsand need only be lengthened.

144


153/192

M E R C U R Y V A P O U R P U M P SThe side arm is next bent slightly downwards to form the nozzle Bwhich can be tapered as shown if desired. The mouth of the nozzleshould be of the order of 5-10 mm diameter, though the size is not

critical provided a good rotary , oil, backing pum p is to be used. Thelarger the nozzle the faster the pumping tha t can be obtained, whilethe higher the backing pressure employed the finer must be the nozzleand the higher the boiler pressure of mercury vapour.The ring seal at C is made not too far from the flask so that the ho tmercury vapour has to travel as short a path as possible. The top ofthe boiler and the exposed portion of the side arm may be laggedwith asbestos rop e or paper to reduce loss of heat. The remainder ofthe pum p is easily constructed. The condenser E can be any reasonable length, for the length does not affect the pump performance. Dis a 3-5 mm bore tube for returning mercury to the boiler, and mustlie below the level of the mercury in the boiler. The condenser jacketis mad e separately an d isfixed o the inner tube with rubber stoppersor co rks which are waxed to give a watertight seal. The jacket mustbe positioned before the T-junction at F is made. If Pyrex glass isemployed then the condenser jacket may, if desired, be joined to theinner tu be with ring seals at either end in the no rmal m anner.The boiler can be heated w ith an ordinary bunsen burner but its baseshould be coated with a ^ in. layer of asbestos pastethis is absolutely necessary with soda glass. It m ay be desirable to employ anasbestos board G for deflecting radiant heat from the boiler, therebyprotecting the ring seal and the waxed cork of the water jacke t.Waran reports that an average pump of this design, workingagainst a backing pressure of less than 1 mm of mercury, will pum p a1 litre volume at atmospheric pressure down to 10~3 mm of mercuryin 2 -3 m inutes. W hen a rotary backing pump is not available, adiffusion pum p w ith a 5 mm nozzle can be opera ted, with less efficiency, against a backing pressure of 7 mm which can be obtainedwith an ordinary water vapour pum p. Sufficient heat must be supplied to the boiler, in the latter case, to produce a mercury vapourpressure of 1-2 cm.Glass Mercury Diffusion Pum p suitable for High Vacuum WorkA useful glass diffusion pump, due to J. Po llard , is shown in FIGURE70, VIII. This pum p is quite satisfactory for high vacuum w ork, a ndour m easurements on pum ps made by han d torch methods show thata pum ping speed for air of about 10 litres/second at 10 -5 mm ofmercury can be obtained at the mouth of a pump of this design.FIGURE 71 shows some results for the speed of this pump communicated to us by J. Pollard.

145


154/192

S OM E T Y P I C A L S I N G L E P I E C E S O F E Q U I P M E N T

Thisfangfhnotcritical1/c m

#m m

A

Tx&rind to I3 mm sharp edge -*

/

3$mm , /5mm

B ?mm

E

C

Iff IF

> vir

FIGURE 70 . A mercury diffusion pum p due to J. Pollard146

8/14/2019 Glass Working

glass working ocr

Documents