ISDUSTRIAL AND ETGINEERISG CHEMISTRY
J u l y , 1933
755
glasses in the temperature zones corresponding to the change points in the thermal expansion curve. LITERATURE CITED (1) Andrade, E. N. da C., Nature, 125, 308 (1930). (2) Day, A. L., and Allen, E. T., Am. J . Sci., 19, 93 (1905). (3) English, S., Howes, H. W., Turner, W. E. S., and Winks, F., J . SOC.Glass Tech., 12, 31 (1928). (4) Fulda, M., SprechsaaZ, 60, 810 (1927). (5) Hanlein, W., Glastech. Ber., 10, 126 (1932). (6) Klemm, A., and Berger, E., Ibid., 5, 405 (1927). (7) Lillie, H. R., J. Am. Ceram. SOC.,12, 505, 516 (1929). (6) Ibid., 14, 502 (1931). (9) Lillie, H. R., J. Rheol., 3, 121 (1932).
, Jw’
T€M< “C. ~
5w’
4W‘
FIGURE 18. SORPTION
I # 630“
HEAT ABCURVES(16)
40
80
/zo
(10) Lillie, H. R , paper presented before the meeting of the Am. Ceram. SOC.,Feb. 14, 1933. (11) Littleton, J. T., J . SOC.Glass Tech., 15, 262 (1931). RELATIONSHIP (12) Peters, C. G., and Cragoe, C. H., J . Optical SOC.Am., 4, 105 *mperaturp,T ,o zm 2.0
0
FIGURE 19. BETWEEY DIELECTRIC CONSTANT
AND
TEhfPER.4TURE
(1920).
(13) Rasoh, E., and Hinrichsen, L. W., 2. Elektrochem., 14, 41
(l5)
1. Chilled lass
2.
a
Annealei glass
any dielectric properties are completely masked. Such observations should serve as a test for the existence of the low-temperature change points indicated by the thermal expansion data of Turner and Winks (19). Observations by Strutt (15) on three different glasses are shown in Figure 19. The different frequencies a t which the measurements are made are indicated on the curves, and it is apparent that there are no indications of any constitutional changes in the
(1906).
Schonborn, H., 2. Physik, 22, 305 (1924). Strutt, M. J. O., Arch. Elektrotech., 25, 715 (1931). Tool, A. Q., and Eichlin, C. G., J . Optical SOC.Am., 8,419 (1924). Tool, A. Q., Lloyd, D. B., and Merritt, G. E., J . Am. Ccram. SOC.,13, 632 (1930). (16) Tool, A. Q., and Valasek, J., Bur. Standards, Sci. Paper 358
(14) (15) (16) (17)
(1920). (19) Turner, W. E. S., and Kinks, F., J . SOC.Glass Tech., 14, 84 (1930). (20) Waterton, S. C., Ibid., 16, 244 (1932).
RECEIVED March
30, 1933
Chemical Composition of Commercial Glasses DONALD E. SHARP,Bailey & Sharp Company, Hamburg, N. Y.
T
T h e ramges of composition of a wide cariety glass w i t h t h a t of a n c i e n t . of modern c.ommerciai glasses are giren and Analyses of museum specimens of tion of glass acceptable e q u a l l y to t’he physical ancient glass, a few of which are of the reasons lchy the compositions have been chemist, the organic chemist, the given in Table I,have shown the g r a d u a b dtered from those in use glass technologist, and the engiyears c o m p o s i t i o n to be much the ago are pointed out. Analyses of m a n y of the neer. Manysubstancesdiffering same as that of modern glass. glasses are included, and colored glass, technical Such differences as e x i s t a r e w i d e l y in composition are freglass, and optical glass compositionsare treated principally in the alkali concluently designated as glass because they bear a resemblance in tent, t h e a n c i e n t glasses, as briejly. some r e s p e c t to the material a rule, b e i n g m u c h higher in c o m m o n l y known a s glass. s o d a a n d potash t h a n t h e Synthetic resins, for example, are frequently considered to be modern glass. Other differences can be accounted for by glasses because they are transparent and glossy. However, be- impurities that must have been present in the materials cause of the absence of silica in such materials the glass tech- available to the early glassmakers. nologist is inclined to consider them more as glass substitutes TABLEI. .4SALYSES O F h C I E N T GLASSES(16, 17) than as glass. A consideration of commercial glass composi( I n per cent) tions will be sufficiently comprehensive if it is confined only to 10 2 3 4 5 6 those compositions that contain a substantial amount of silica. i:2le: 63.72 63.86 65.95 66.26 68.48 67.44 3.26 0.70 2.98 Glasses containing silica are produced by the fusion, a t $ : ? $ 1.04 0.65 2.49 0.54 0.67 0.28 0.78 0.91 0.51 relatively high temperatures, of three general types of maCaO 9.13 7.86 6.89 7.09 5.71 4.80 5.20 4.18 1.37 1.48 5.28 5.64 terials-(1) silica, ( 2 ) alkaline fluxes, and (3) That may be 20.63 22.66 20.30 19.33 14.95 13.94 ?$, 0.41 0.80 0.96 0.37 2.83 1.93 termed “stabilizing ingredients.” Silica is a primary in... ... 0.97 0.61 ... 0.70 gredient in all the glasses; the alkaline flux may contain soda, cuo ... ... ... 0.95 ... ... ... ... 1.08 ... 0.54 0.84 potash, or both; and the stabilizing ingredient, may be an &$: ... ... ... ... 0.95 1.01 oxide Of Calcium, magnesium, barium, zinc, alumirLum,lead, or a Samples 1 and 2 , colorless glass from Tel el Amarna, E g y p t , 1400 E . c.; boron, Or a combination of two or more of these. ’$hen silica 3, colorless glass from Elephantine, Egypt, 200-100 E . c.: 4, dark blue glass from Elephantine, Egypt, 200-100 B. c.; 5, window glass, about 900 A. D.; and soda only are fused t’ogether, the product is a glass, but a 6, white blown glass about goo A. D. soluble one. The addition of stabilizing ingredients, however, brings about a tremendous reduction in solubility, so that, In general these old glasses are shown to contain roughly to all intents and purposes, t,he resultant glass is completely 65 to 70 per cent silica, 6 t o 10 per cent calcium and magpermanent. nesium oxides, and 16 t o 23 per cent alkali principally as soda. It is interesting to (compare the composition of present day Such glasses are designated by the glass technologist as sodaHERE is no simple defini-
g:g
i56
IKDUSTRIAL AND ENGINEERING CHEMISTRY
-
Vol. 23, No. 7
c h e m i c a l composition of a glass. The results show considerable variation for most U-IIW of the properties, although they are reasonably accurate in the case of the specific heat. Since the time of Winkelman and Schott many others have atteTpted to develop m o r e a c c u r a t e values for these constants. There are a half-dozen sets of factors for computing the density of glass, some of w h i c h g i v e surprisingly good r e s u 1t s so long as the glass on which they are employed is not of extreme composition. Morey and M e r w i n (16) have recently shown why a simple additive relation cann o t b e used to express accurately the density or refractive index of such simple, three-component g l a s s e s as the soda-calcium-silica group. They determined the lines of 524 constant density and refractive index over the range of FIGURE1. RELATION BETWEEN DENSITY AND COMPOSITION OF GLASSES IN TERNARY SYSTEM c o m p o s i t i o n s included in NapO.SiO2-CaO.SiO~-SiO~(15) the t r i a n g l e 40 to 100 per cent silica, 0 to 60 per cent lime glasses, the calcium and magnesium oxides being con- calcium oxide, and 0 to 60 per cent sodium oxide. They sidered as “lime.” The comparatively large amounts of showed these iso-lines in general to be definitely curved, alumina and iron present in many of the ancient glasses whereas, if the properties were simple additive functions doubtless had their origin in impurities in the batch materials of composition, the iso-lines would necessarily be straight or in corrosion of the clay containers in which the glass was and parallel. Figure 1 shows the triaxial diagram of Morey fused. The great bulk of modern glassware also consists and Merwin with the lines of constant density. The straight primarily of silica, lime, and alkali, but the silica content line WS represents the compositions for a density of 2.5 generally is above 70 per cent while the alkali content seldom calculated from the factors of Winkelman and Schott. Their exceeds 17 per cent. density factors are: silica, 2.3; soda, 2.6; lime, 3.3. These The reasons for the similarity in the composition of ancient are substituted in the equation, and modern glass are due primarily to the fundamental nature 100 AI A P As = and character of glass itself which require that a glass (1) z, ... S must be fusible a t commercially obtainable temperatures, (2) must be capable of being fashioned into usable forms without in order to calculate the density. In this equation AI, A*,Aa, crystallization taking place, and (3) must be reasonably per- etc., are the percentages by weight in the glass of the oxides for manent in the use t o which it will be put. It is not surprising, which the corresponding factors are Z1, Z2, Z3, etc. The therefore, that ancient glass, a t least that which has survived, factors given by Baillie for silica, soda, and lime are 2.24, 3.20, is similar in composition to modern glass. We may likewise and 4.30, respectively. The corresponding factors of English conclude that ancient glass that was too high in lime could and Turner are 2.2, 3.47, and 5.00. The use of these factors not be worked into ware without crystallization, while any assumes the validity of the additive relationships which that was too high in soda may have disintegrated completely Morey and Merwin have disproved. Finn and Thomson (9) and is not now available for study. and Finn and Faick employ an expression: If it were possible to conceive of glass as a pure mixture of D di d, d3 its constituent oxides, its properties could then be computed in advance from the properties and proportions of its con- where di, d2, etc. = partial densities determined from an equation, d = a ( p c ) b stituents; that is, the properties would be additive functions where a, b, c = constants which dider for each oxide of the composition. Many attempts have been made to p = percentage oxide in glass determine for each oxidic component a suitable factor to represent its effect. It has been found that the physical Such a three-constant equation naturally could fit the curves of properties are somewhat additive but only in a rough sense. Morey and Merwin much more closely than could any Winkelman and Schott (10) determined factors for the straight line. The line BB of Figure 1 was calculated from mechanical and thermal properties of a number of the oxides the factors of Baillig (3) for the same density. The line ET entering into glass composition, By means of these, density, was calculated from the factors of English and Turner ( 7 ) . tensile strength, compression strength, elasticity, scratch These lines representing data from three different sets of hardness, specific heat, thermal expansion, and thermal factors show how far the additive relations depart from a true conductivity may be computed, within limitations, from the representation of the facts. hterpofatron Sca/es
z + + z,
+ +
+
July, 1933
INDUSTRIAL AND ENGINEERING CHEMISTRY
Earlier work of Norey and Bowen ( l 2 , 1 4 )gives equilibrium diagrams of the ternary system sodium metasilicate-calcium metasilicate-silica and some data also on the effect of additions of magnesia, boron oxide, and alumina. These diagrams show the crystallization relationships and the substances which crystallize out on cooling the material. The glasses of commerce lie in a rather small area in which the crystallization temperature is lower than, or a t least not much higher than, the normal working range of temperature. Circle W’ in Figure 2 shows the approximate region of the composition of the old handmade window glass; circle B indicates roughly the composition range of modern bottle glass. The ability to obtain a glass of a composition expressed by any point on this diagram is largely a matter of how rapidly the glass can be fashioned into useful form, from a temperature above the liquidus, without crystallization taking place. while the glass is being fashioned and is cooling, twoopposing forces come into play-namely, the tendency of the primary phase to crystallize and the resistance to crystallization from the decreased mobility of the molecules as the fluidity of the glass decreases ~ t the h drop in temperature. Thus the rate of change of viscosity and the crystallization tendencies of any glass determine, in a major degree, its suitability for commercial purposes. Prior to the twentieth century there Rere, in the main, but
two types Of glass: lime and lead* glass working was done by hand processes, and the glassworker, according to his skill, fashioned articles from almost any sort of glass that was available. A difference in viscosity between two glasses afforded him Only the opportunity to make from a glass what h e might be more accustomed to make from a “softer” glass. As a consequence window glass and bottles were generally made from the same material. It had not become necessary to suit the glass composition to a particular process, for the glassworker suited the process to the material. We find, therefore, that both window glass and bottle glass of the nineteenth century, had, in general, a composition approximately as follows: APPROX RAWQE MOST SlO? A1203
CaO NazO
% 69-72 12 5-13 1-4 5 13-15
REPREsErrlTrVE
70
71 5 131 0 5 14 0
The flint glass of this period varied in lead content according to the requirements of the ware. Usually potash was the alkali rather than soda. The composition was, in general, as follows:
751
process. When glass was worked by hand, the glassworker by his skill compensated for variations in temperature or viscosity, but the automatic machine cannot without human aid adjust itself to such changes. When automatic processes came to be employed, uniformity of conditions became more and more important as the productive capacity of the process was increased. Moreover, increase in productive capacity itself demanded the use of a glass of such composition as would give the best conditions for rapid fashioning of the ware, taking into consideration also the three points mentioned earlier-namely, easy fusibility, freedom from crystallization, and the permanence of the product. WIKDOWGLASS The window or sheet glass industry has progressed rapidly in recent years from hand-blown and -flattened glass, through the stages of semi-automatic manufacture by the machine cylinder process of Lubbers, to the modern methods of drawing a continuous flat sheet Or ribbon from a huge Pool of glass. The glass used for the semi-automatic cylinder process differed little in composition from that of its handmade predecessor as may be seen from the analyses given in Table 11. In the case of the fully automatic processes the situation There isis here quiteandifferent. example of what has happened in several branches of the industry. The first step toward machine manufacture of window glass was an imitation of the hand process in which part of the work (the blowing of a large cylinder) was done by a machine. The subsequent cutting and flattening into sheets was done by hand. The old hand glass composition was satisfactory, In the modern processes, as carried out by Libbey-Owens, by the Pittsburgh Plate Glass Company, and by the Fourcault method a t several plants, the sheet is drawn directly as a large flat ribbon which may be cut into sheets of any desired size. It is obvious that uniformity of temperature and composition are prime considerations in the creation of such a sheet of glass of uniform thickness and width. M o r e o v e r , in the development of the above mentioned processes it was found that glass, containing approximately equal parts of lime and soda,
MOBT
APPROX
5102
KzO
PbO
+ (NazO)
RANGE ‘To 50-65 8-14 20-40
REPREUENTA-
TIYE
With the advent of modern i n d us t r i a l requirements, a n d w i t h the engineering developments in the glass industry in the last t h i r t y y e a r s , Sio, new emphasis took place on adaptation Weight Per cent S L O , of compositions t o FIGURE2. BOUNDARY CURVESAND TIELINESIN TERNARY SYSTEM Na20+X02-Ca0 SiO&iiO, ( 1 2 )
INDUSTRIAL AND ENGINEERING CHEMISTRY
758
had a setting rate that was too abrupt, and that its compos'ition was unfavorable for avoiding crystallization a t temperatures found necessary for drawing. We know now from the work of Morey ( l b , I 4 )of the favorable influence of a higher soda content and of small amounts of alumina and magnesia in lowering the temperature above which crystallization will not take place. However, these important alterations in composition of the commercial glasses were worked out a t the time quite empirically.
FOURCAULT ETROPEAAGLASS FOURCAULT AMERICAX GLASS SiO? FetOt CaO
{
1.42 2.59 1.90 1.82 1.06) 0.78 13.34 13.40 12.40 13.31 12.76 13.46 12.58 14.16 12.68 0 0 26 0.26 Trace 1.07 . . . 0.22 14.01 14:55 li'.bO 15.i2 13.25 12.46 13.37 11.96 12.69 SO8 0.97 0.29 Not detd. a Sample 7, American hand-cylinder ( 8 ) ; 8, Belgian hand-cylinder (41; 9, English hand-cylinder ( 8 ) . 10, French hand-cylinder ( 8 ) . 11, G e r m a n hand-cylinder (8). 12, Britikh handmade (go); 13,. Lubbe& (8,p. 781): 14, European cylihder (1930) (2); 14.4, American cylinder (1920)(2).
%&
..................
The Fourcault glass composition developed in two directions, as may be seen from the analyses in Table 111. In Europe freedom from crystallization difficulties (always the bone of contention in the Fourcault process) was obtained by
0.5
8.0 3.0 15.5
Rolled sheet and plate glass are made by methods differing from the drawing processes in that they permit the glass to be kept a t quite a high temperature up to the moment when it is formed into a sheet. During the forming of the sheet the glass is cooled with reasonable speed from this high temperature, so that it passes rapidly through the temperature range that is most favorable to crystallization. Consequently, there is ordinarily no acute crystallization problem, and a composition higher in lime and lower in soda than Fourcault glass may be used. The analyses in Table IV show rather clearly that the compositions do not differ much from glasses used in hand processes.
bf.4CHISE CYLISDER
::A t:!: ::A
% 73.0
ROLLEDSHEETAND PLATE
-NINETEENTH CENTCRY GLASSGLASS Sample: 7'3 8 9 10 11 12 13 14 14.4 Sios 72.2669.48 71.40 69.65 72.68 72.10 71.60 71.74 72.88
{
% 72.0 1.6 10.4 16 :o
NalO
(In per cent)
CaO
+ A1201
hlg0
TABLE 11. ANALYSESOF WIXDOWGLASS
$::
Vol. 25, No. 7
COXTAISERS When bottles and jars were made by hand methods, the old high lime-low soda glass was considered entirely satisfactory. However, with the advent of semi-automatic machines the glass was changed slightly to a content of about 10 per cent
TABLE111. ANALYSESOF FLAT-DRAWN WINDOWGLASS FOCRCaCLT PROCESS
Sample: Si09 Alios FerOs CaO
E$:
71.82 1.40 0.07 8.90 3.31 14.27
16 71.00 0.54 0.09 11.26 1.56 15.13
17 72.14 0.84 0.08 9.06 1.29 16.23
18 72.43 0.80 0.20 10.16
19 70.51 1.42 0.16 10.67 16:?4
0.31
0.44
...
15:88
so0
0.45
0.76
Trace
...
0.93
Trace ..,
MnO
ABZOS BaO
Kr0
15a
... ... ...
...
...
...
... ...
...
... ...
...
...
...
...
( I n per cent) SHEET
20 70.64 0.77 0.11 10.58 0.09 17.02 Trace
0.78
... ... ...
21 22 72.27 72.72 0.30 075 0.10 10.06 9.54 2.01 2.15 14.31 Not d e t d . 0.19 ... 0.71
{
.
I
.
... ...
...
1.30
...
-
23 72.9 0.7 7.9 2.8 15.0
-LIBBEY-O~ENS
]
..
24 72.14
25 70.87 0.52 1.06 0.15 0.29 11.24 10.75 2.62 2.42 14.06 12.60 0.05 0.01 Not detd. 0.85
...
.... ..
017
... ... ...
PROCESJ-
26 72.55 0.64 0.14 12.54 0.19 13.16
27 72.15 0.56 0.12 10.64 2.66 13.45
28 71.76 0.82 0.10 11.55 2.41 12.65
... ...
0.47
Trace
...
... ... ...
... 0.19
0.93
... ... ... ...
a Samples 15 to 17 made in U.S . 1929 ( 5 ) ; 18 made in Belgium 1924 (6); 19 made in Belgium 1925 ( 5 ); 20 made in Europe (20): 21 made in Bohemia (6, p. 781); 22 and 23 made in E. S.(2); 24, Libbey-Owens ( 2 0 ) ; 25 t o 27, Libbey-Oaens made in Europe 125 a n d 26 (8,p. 781), 27 (6,p. 779)); 28, Libbey-Owens, 1930 ( 2 ) .
lowering the liquidus temperature by means of a minor addition of alumina, and by substituting about 2 per cent of soda for an equal amount of calcium. A good share of the European Fourcault glasses therefore contain about 10 per cent calcium oxide, 1 to 2 per cent alumina, and 15 to 16.5 per cent sodium oxide. In this country a similar result was obtained by the same substitution of sodium for calcium, and, because dolomitic lime is abundant here, by the substitution of magnesia instead of alumina for some of the calcium. American Fourcault glasses generally contain, therefore, less than 1 per cent alumina, from 2 to 4 per cent magnesia, 6 to 8 per cent calcium oxide, 15 to 16 per cent sodium oxide, and about 73 per cent silica. Of recent years in this country there has been a tendency away from the magnesia glass and towards a glass containing alumina, and, more recently still, to the use of small quantities of barium. In either case-that is, the substitution of magnesia or alumina for part of the calcium-the results were satisfactory. The glasses had similar viscosity-temperature curves, each avoided crystallization, and each gave a sufficiently durable glass. Any glass having a composition somewhere between these two average compositions is, as a rule, quite suitable for the Fourcault process. The glassmaker may therefore choose a composition suited to the materials that are readily available. The range is roughly as follows:
calcium and 15 per cent soda. Such a composition worked reasonably well on fully automatic, suction-feed machines, because the glass was kept quite hot up to the very moment when the glass was drawn by suction from its receptacle to be fashioned into a bottle. OF ROLLED ASD FIGURED SHEETWIRE TABLEIV. AKALYSES A N D PLATE GLASS
(In per cent) Sample:
29'" 71.78
::$ {
13.12
30 71.54 0.72 13.50 0.39 13.13
31 70.82 0.42 f 12.65 1.71 13.67
32 70.50 1.67 0.17 14.04
...
...
...
33 34 35 71.07 72.68 70.86 1.27 0.50 0.47 0.13 0.07 0.06 13.35 12.95 11.13
12:30 i3:87 ii:i7 ij:07 13:82 0.34 0.44 0.23 0.76 i i o t detd. Not detd. 0.45 0.18 Not detd. Not detd. Not detd. . . .
...
...
0:is ......
Sample 29, figured glass, European (a); 30, figured glass, American, 1916 ( 8 ) ' 31 figured glass American 1931 (2). 32, rolled plate European 1931 (6): 33,'rolled plate, European, i932 (6);54, polished plat;, American: 1929 (6); 35, wire plate, American, 1927 (6). a
The development of the gob feeder for feeding slugs of molten glass into the forming molds brought about problems similar to those encountered in the Fourcault process. Moreover, limitations of the machines which pressed and blew the material into shape made it necessary to employ a glass which remained plastic over a greater temperature interval.
INDUSTRIAL AND EKGINEERING CHEMISTRY
July, 1933
759
As a consequence, the composition used for containers does that could be melted in large, continuous furnaces, and that not depart much from the typical analysis given in Table V. could be reworked repeatedly in a flame without surface TABLEV. ANALYSIS OF REPRESESTATIVE CONTAINER GLASS crystallization. Here, too, recourse was had to the use of some magnesia in a high-soda glass, and today the majority of % % electric light bulbs are made from a glass containing approxi3 . .5 SiOr 74.0 17.0 Fez03 + A1203 0.5 mately 70 per cent silica, 20 per cent soda, and roughly equal CaO 5.0 quantities of calcium and magnesium oxides. This glass is I n this country dolomitic limestone is used very extensively melted in huge tanks containing many tons of the molten as the source of lime in container glass. Consequently, al- material and is turned into bulb enclosures by a marvelous though the total lime-magnesia is much lower in this type of type of machine at the rate of over three hundred bulbs glass than in most of the other commercial glasses, the pro- per minute. Machines such as are used for making bulbs might be portion of magnesia is higher. This combination of calcium and magnesium, together with a high alkali content, yields adapted to the manufacture of other thin glass articles such as a glass with a long working range, and with nearly the op- vacuum bottles, lamp chimneys, drinking glasses, and laboratimum composition for avoiding crystallization. Although tory ware. The bulb glass composition, however, is not glass-blowing machines may be accelerated if IL quicker setting necessarily suited to any of them. Lamp chimneys, for
E2
TABLE1-1. Sample: SiOn A1203 FezOa CaO .MgO NazO K20 MnO BaO
so3
3W 69.1
37 70.9
11,s 1.e 10.6 1.585
::;: ... ...
14.02 0.49 10.66
39 74.43 1.16 4.80 3.53 15.60
0.30
...
...
...
...
40 41 74.16 73.52 1.25 1.15 4.64 5.60 3.41 3.76 15.08 S o t d e t d .
...
... ...
... ...
...
OWES3 &lACHISE
GOB-FEDGLASS
38 72.43 0.50 4.80 3.60 17.40
i::: {
GOB FEEDER, AND
ANALYSES 0 1 7 COSTAISER,
CONTAINER GLASS
.
...
.
...
...
42 72.16 1.36 6.40 0.51 16.43 0.52
...
0.65
...
...
43 74.74 0.82 5.47 3.57 15.23
44 45 74.43 63.9 1.31b) 8.12 13.9 0.61 1.2 15.53
2":; 2":;
...
... ... ...
...
...
I
...
. . I
GLASSES
-OWEM
...
BOTTLEGLASS-
46 73.0 8.59 1.49 15,64
47 70.54 0.17 2.08 r.77 1.53 19,02
48 71.98 o ,85 0.06 9.07
0.48
0.45
0.08
g::
{
, . .
...
0.09
5.5 4.1 16,93) 15.0
...
. . I
...
49 74.50
I
0.8
.. ..
... ...
...
... 0.76
...
~~
.
.
,..
... ...
a Samples 36 a n d 37 hand-fed t o O'Xeill machine (PO); 38 crystal bottle American 1915 ( 2 ) . 39 crystal bottle. American. 1928 ( 2 ) : 40. rrvqtal .............................. ...........h n t t l e . American. 1931 ( 2 ) ; 41: milk bottle, American, 1928 ( 2 ' ) ; 42, 'crystal perfumk bottle. -F;.-nrh (9):k.?. krnher hott .- . ~~ -...le American ( 2 ) ; 44, emerald green bottle American'(: 3 ) ; 45, av. of 32 different German glasses, 1930 (6, p. 1094); 46 and 47, a mber, German (6, p . 1110);48,'crystal, German (6,p. 1110); 49,crystal: American ( z i . b Includes CrnO:.
.
..............
T ~ B LVI[. E ASALYSES OF ELECTRIC LIGHT BULB GLASS (In per cent) Sample: Si01 Alios FetOa PbO BaO CaO MgO Na10
51a 53.7
52 56.5
53 62.3
54 68.1
35:O
29:s
2i:4
6:s
..
..
..
..
..
"
..
..
7.1
55 63.0
{
0.61
21.0 0:3 0.2
g;:
56 72.07
::: ... ...
57 72.4
"0:. .
58 70.4
{
3.02 5:3 3.86 3.7 { 2 0 , ~ ~ ]17.4
O.') 2:2
5.6 3.4 16.5 1.3
59 58-59 0.2-0.4
60 71.5-73.5 1-2
28-30
..... .....
0.04-0.06
..... ..... .....
0.06-0.1
5-6 3.5-4.5 14-17 0-1.5
61 72.1 2.0
62 73.6
... ...
... ... ...
2.77 1.95 21.1
5.37 3.67 17.23
...
...
2:0 5:4 7:0 16:4 7-8 9.3 8.6 8.3 7.6 .. 4-5 ... ... AszOa .. .. .. .. 0.05 0.37 .. ..... ... ... M nO .. .. .. .. 0.2 0.06 .. ..... ..... ... ... a Samples 51 t o 53 lead glasses German 1904-14 (6 P. 1323)' 54 barium lead glass. German (6,p. 1325). 55 lead glass American (E),; ;6 modern, English (6, p. 1328);i7, American (6, P. 1326);58, G e r m i n (6, P . 1829)';59, range of composition, modern lead bu'lb &ass; 60, ra&e of composltioi , modern lime bulb glass; 61, lime bulb glass ( I @ ; 6 2 , lime bulb glass, n e s t l a k e machine, calcd. from batrh (go).
Kz0
glass is used (that is, if lime is increased arid soda lowered slightly) the actual yield of good ware in such an event is generally less, because of the likelihood of small cracks or checks developing in the edges of the ware. A composition that is plastic over a longer temperature range avoids this checking. With such a glass, speeds of thirty-five averagesize bottles per minute from one machine are not uncommon. Analyses of a number of container glasses are given in Table VI.
THIS-BLOWK WARE Enclosures for electric light bulbs were made entirely by hand up t o about twenty years ago. A soda-potash-lead silica glass was generally employed, although barium was sometimes substituted for part of the lead, and soda for the potash, particularly during the World War. These compositions (Table VII) yielded glass which had all the properties most desirable for making electric bulbs. It was easily fusible, readily Mown to shape, resistant to atmospheric attack and, what was very important, could be softened without trouble in a small gas flame for the purpose of attaching it to the glass tube bearing the filaments. Properly chosen wires could be sealed, gas-tight, into such a glass n-ithout breakage of the glass. For years, bulb glass and the tubing used n-ith it were melted in covered pots or crucibles. The advent of modern automatic machine processes called for a cheaper product
example, must be somewhat resistant to temperature changes; laboratory beakers must be decidedly resistant to corrosion by water or chemical reagents, and a highly alkaline glass is incompatible with such resistance. Analyses of some glasses used for chimneys, vacuum flasks, and so forth are given in Table VIII. TABLE b-111.
ANALYSES OF F'ACLX7M ETC.
BOTTLES,L I X P CHIMNEYS,
(In per cent) SamDle: Si02 A h o r f Fe202 CaO MgO Nag0 Ii?0 PbO AS?O3
__
63" 64 65 66 fi7 fiR tiQ .. 71.28 73.52 54.92 76.78 73.88 75.27 68.0 1.92 3.68 1.28 0.72 2.24 1.91 3.87 6 . 6 2 Trace 6.52 0.63 , .2 3.19 0.10 0 . 2 0 0.24 . . . 0.07 12.13 14.50 2.08 11.14 6.67 4.96 2 i : o 1.01 . . . 4.54 4.74 Trace . . . . . . . . . . . 34.93 . . . . . . . . . .. 0.13 . . . 0.99 . . . 0.73 ... .. 1.21 . . . 0.82 . . . 4.70 1.20 . . . . . . i6:48 ii:k . . . . . . . . . . . . . . . 1.98
...
"I
2,'
ZnO B?OI Sb20a a Sample 68. American vacuum bottle ( 2 ) ; 6 4 German vacuum bottle ( 2 ) ' 65 miner's lamp, French analysis by W. H. &they (18)' 66, lamp Aus; trian, analysis by Withe; (18)' 67, Jena, incandescent gks chimne; ( 1 8 ) ; 68. Jena, heat-resisting globe f h high-measure gas (18) ; 69, Christmas-tree ornament glass ( 2 ) .
TABLETTARE hlost tableware of the cheaper variety is made by automatic or semi-automatic process and has practically the same composition as the container glass in Table T'. Hand-pressed
INDUSTRIAL AND ENGINEERING CHEMISTRY
760
tableware generally contains slightly less soda and more lime because the more rapid rate of solidification permits faster production. For better quality ware, small amounts of lead oxide are employed because of the added brilliancy of the product. The best quality of crystal tableware is generally a straight potash-lead-silica glass. Analyses of a number of these glasses are given in Table IX. TABLEIX. ANALYSESOF AMERICASTABLEWARE (2) Sample: SiOz AliOa Fez08 PbO
69A0 72.39 1.01 0.08
(In per cent) 70 74.2
..... .
...
BaO
0.15 8.10 0.15 18.12
CaO MgO Naz0
4:3 3.2 17.7
71 73.1
72 67.2
i: 1
1418
i:0
0:9
..
9:5 7.1 0:2 0.5 ... 0.4 .. .. 0 Samples 69A and 70, pot, lime crystal; 71, part lead crystal; 72, light lead crystal.
... ...
Kz0 B ~-.0..
AszOr
l2:7 4.6 0.5
TECHNCAL GLASS There are many uses for glass in which special physical properties are important or absolutely necessary. Such glasses, for want of a better name, may be classified as technical glasses. This class includes laboratory ware, thermometer glass, special tubing, optical glass, spectacle glass, and so forth. LABORATORY WARE A glass to be satisfactory for beakers and flasks should be highly resistaht to corrosion by reagents, to rapid changes in temperature, and to mechanical shock. In addition, it should have as many as possible of the ordinary characteristics of glass, such as being transparent, easily melted and fashioned, and reasonably cheap to produce. Fused quartz would be ideal for most purposes in the laboratory, but it lacks the two latter qualities. The early laboratory glasses were lime glasses, high in silica and lime and somewhat low in alkali, because such a composition was reasonably resistant to chemical attack. Resistance to temperature changes was obtained by making the articles very thin. The first real progress in laboratory ware was made by Schott who introduced the old Jena glassware, made from a glass in which part of the soda was replaced by boric oxide in order to lower the alkali content without making the glass too difficult to fuse. Boric oxide, like soda and potash, is a flux. It can be used in place of part of the soda although its fluxing power is much less than that of either soda or potash. However, it has an additional advantagenamely, it produces a much lower coefficient of thermal expansion than the alkali it replaces. The old Jena glass was followed after a few years by the new Jena glass, which was a further improvement upon anything that had gone before. Many other glasses of good chemical resistance were produced, and at one time there were on the
Vol. 25, No. 7
market as many as a dozen glasses that were more or less suitable for laboratory use. Analyses of some of the best of these are given in Table X. In general, these laboratory glasses contained about 10 per cent of soda. Besides relying upon a relatively low alkali content for increasing the chemical resistance of the glass, the makers employed fairly large amounts of alumina, zinc oxide, magnesia, or other stabilizing oxides. Alumina in large amounts is noted for the “stiffness” it imparts to molten glass. Consequently, high-alumina, low-alkali glasses are difficult to melt and refine, and they solidify a t such a high temperature that working them is not easy. To overcome these difficulties more alkali had to be used than was advantageous from other standpoints. The laboratory glass most nearly approaching the ideal was introduced about fifteen years ago in Pyrex. The makers of this glass faced squarely the manufacturing difficulties and developed a product which could not be melted by ordinary means but had all the other desirable qualities-low alkali content, high resistance to corrosion, low expansion, and high resistance to temperature changes and to mechanical shock. The original Pyrex had approximately the following composition: ,O 6 2.0 11.9 4.4 1.1
The makers of Pyrex also manufacture other heat-resisting glasses of compositions suited to particular purposes. In general, these glasses are similar to Pyrex in composition with slight modifications either to obtain the special property which is most important or t o obtain facility of manufacture.
THERMOMETER GLASS Another contribution of the Jena laboratory was the glass used for thermometers. If a thermometer is made of ordinary glass and exposed only to ordinary temperatures, the thin bulb which contains the mercury contracts slightly as time goes on. The contraction gradually makes the thermometer give higher and higher readings, although the rate of change diminishes with time. By heating the thermometer to about 100” C. and permitting it to cool, the situation is reversed, and it is found that the volume of the bulb is increased. Accordingly, if a thermometer has been lying on a shelf for some time, it will, if heated to the boiling point of water and cooled, display the phenomenon known as depression of the zero. Obviously, such a useful instrument as a thermometer should not be made from material exhibiting this afterworking phenomenon. The studies a t Jena showed that the afterworking characteristic was greatest in glasses which contained about equal quantities of soda and potash. The smallest depression of the zero was obtained when only one alkali was. present, although it made little difference whether it was
GLASS TABLE X. ANALYSES OF LABORATORY (In per cent) Sample: Si02
Bios
Alios
LO
73” 77.0
.. ..
74 79.1
.. .. .. ..
75 66.4 4.0 2.4
76 75.9
77 84.
78
79
0:2
... 7.1
...
...
80 67.3 6.2 2.5 0.23 7.8 0.79 3.4
... 10.9
81 80.62 11.90 2.00 0.16
82 80.75 12.00
0.22 0.29
0:30
... ...
...
...
83 80.5 11.8
b:i9 0.06
...
4.10 4.4 3.83 10.8 0.10 0.20 0.30 0.61 0.3 ... ... . . . . . . 0 . 0 2 S0a 0.40 Tiace 0.66 0.7 0.14 0.02 AszOn .. ... 0 . 6 2 ... .. . . . 0 . 6 0 ... Sbz0; .. .. a Sample 73 Stas 1888 (6 p. 1301). 74 Kavslier, 1879 (e, p. 1301); 78, Jena, prior t o 1910 (6,p. 1301). 76 Kavalier ( 2 2 ) ’ 77 Jena, l 9 l l , ( 6 , p. 1301). 78, Jena (21);’79, Maobeth-Evans ( $ 1 ) ; 80, Nonsol ($1); $1, Pyrex, 8ndy818 by Dept. Glass Tech , She5eld’(4): 82, Pyrex’7Od EJ,snalyei. by Cornin; Glass Works ( 1 9 ) ;83 ( 2 1 ) . &a0
5:0
..
..
0:4 6.7
.. ..
1516
.... ..
7.9 0.20
...
?:8 0.3
.. ..
7.5 0.37
...
...
INDUSTRIAL AND ENGINEERING CHEMISTRY
July, 1933
potash or soda alone. The result was the development of the Jena normal thermometer glass 16"' which for many years was the standard for thermometers. This glass, which contains only soda as alkali (Table XI) has a depression constant about one-tenth as large as the glass formerly in use. A later glass, 59"', contains less alkali and more boric anhydride. It may well be considered the earliest example of low-expansion borosilicate glass. TABLE XI. ANALYSESOF THERMOMETER GLASSES Sample: SiOr
BnOs
All08 Fezor ZnO CaO
81" 67.3 2.0 2.5
7 :0 7.0
(In per cent) 85 86 66.58 72.0 12.0 0.91 3.84 5.0
...
6.24 7.18 0.17
..
87 72.86 10.43 6.24 Trace
.. ..
0.35 0.20
..
...
88 60.77 1.27 3.02 0.17
89 57.45
...
1.12 Trace
...
0.10 22.76 10.21 1.05 0.52 0.14 0.11
...
3.00 0.09
...
MgO 21.99 ... PbO 10.44 9.82 14:o 14:so ii:o NazO 0.17 Trace .. 0.10 Kz0 3.82 . . . .. AstOa ... MnO 0:2 0.28 .. Tiace 1.90 ... .. ... PlOS a Sam le 84, Jena normal 16111 (fa): 85, Jena normd, analysis by W. H Withey 58). 86 Jena borosilicate 59111 (IS)' 87, Jena borosilicate 59111 analysis by Withdy (18); 88. lead glass ( 8 ) : 86, opal backing in lead glas6 ,.
..
(2).
The white backing in thermometers is an opal glass generally made by the addition of calcium phosphate and arsenic to a base batch which without them would have resulted in a clear glass. This bone ash-arsenic opal glass is characterized by an intense whiteness even in small thicknesses. Thermometer glass must have a composition that will permit repeated heating of the tube in the glass blower's flame without discoloration or crystallization. Many years ago it was discovered that glass, made from a certain sand obtained near the Thuringian Forest in Germany withstood repeated heating in a flame without change. Upon thorough investigation this sand was found to contain about 4 per cent of alumina, which impurity later was demonstrated to be responsible for the excellent properties of the glass. Even today in glass tubing for thermometers and other articles, crystallization is avoided by the incorporation of sufficient alumina, which now is added purposely either as feldspar, china clay, or alumina hydrate. Glass containing lead may not be heated in a reducing flame, or a black metallic luster may be developed from reduction of the lead silicate to metallic lead. Arsenic and antimony, too, should be avoided if possible in glasses which are to undergo repeated heating, as they also may cause discoloration.
GLASSTUBING Glass tubing is unique in that it is the first step in the fabrication of many small articles that are blown by hand or by machine with the aid of blast lamps. Besides its ability to withstand repeated heating without crystallization or discoloration, it often must have other special properties. For example, electric light bulbs and radio tubes are made of two main glass parts--an enclosure and a filament support. The enclosure has already been discussed. The filament support is a piece of glass tubing into which wires have been sealed; the wires bear the filament and, in the radio tube, the electrodes also. In order that the two glass parts may seal together perfectly without cracking, they must have substantially the same coefficient of thermal expansion. Furthermore, the filament support should have the same expansion as the leadin wires that are fused into it. The filament support becomes quite hot when the tube is in use. Hot glass conducts electricity much more freely than does cold. Accordingly, if the current flows always in the same direction, as it may in radio tub-, ions migrate through the glass between the
i6 1
wires, and presently the glass cracks because of the strain set up by this action. It is consequently qecessary that the glass have a low eleotrical conductivity a t elevated temperatures. This may be obtained by keeping the alkali content as low as practicable and by using a combination of soda and potash for the alkali. From the foregoing it is evident that the quantity of alkali present in glass is very important, determining in many cases the suitability of a glass for a particular purpose. Pyrex, for example, because of its low alkali content, is particularly suited to uses as an electrical insulator. Spun glass may not be made from a material that contains much alkali, since the great surface exposed per unit of weight renders the glass sensitive to the corrosive influence of even such moisture as is present in air. OPTIC.4L GLASS Until the time of Schott and Abbe there were but two types of optical glass-crown and flint. The former was a lime glass, the latter contained considerable lead oxide. The refractive index, or light-bending power, of the lead glass was slightly greater than that of the crown, but its dispersion, or
4 ~ * " ~ 1 ~ " i k o 1 ' " ' ~ c
IS5
RL-FRACTIVC
INDEX
FIGURE3. RELATION BETWEEN REFRACTIVE INDEX ( n D ) AND MEAN (nF - n c ) I N OPTICAL DISPERSIOX GLASSES
ability to spread the rays out into a spectrum, was far greater than that of crown glass. Consequently, by a suitable combination of a crown with a flint lens, the spreading of light into colors could be practically eliminated without impairing the ability of the combination to bring the rays to a focus. There were other shortcomings in lenses which such a simple combination of glasses could not overcome, and Schott and Abbe set out to conquer these. They developed and put into practical use transparent glasses containing fluorides, boric anhydride, zinc oxide, and barium oxide. Optical glasses are classified by the physical constants, refractive index, %D, and mean dispersion An. The old crown and flint glasses were limited by a definite relation between %D and An, and so, except for the benefit to be obtained from extremes of refractive index, nothing could be gained by selection from the old types of glasses. Figure 3 illustrates the dependence of dispersion on refractive index in these old glasses. Schott and Abbe found that boric oxide in a lime glass increased its transparency and lowered its mean dispersion. They found, moreover, that barium increased the refractive index of a crown glass materially without increasing its dispersion. The circles in Figure 3 illustrate the wide divergence of optical properties in the newer glasses as compared to the old. From this work of Schott and Abbe the anastigmatic lens and the apochromatic objective came into being, as well as other lenses and prisms which have been such effective tools for the chemist. The wide variation in composition of optical glasses is apparent from Table XI1 which contains only a few of the hundreds of varieties.
IN DU STR IA L A N D E N GI N EE R IN G CH E M ISTR Y
762
Vol. 25, No. 7
TABLEXII. ANALYSESOF OPTICALGLASS ( I n per cent) Sample: nD:
V:
SiOn
BzOa
Ah01 Fez01 As201
z&82_06 nu CaO MgO BaO PbO
Kz0
Sa20 MnzOa
H20
CROWX GLASS
90" 91 92 93 94 1.5179 1.5250 1.5100 1.5164 1.5409 60.2 59.1 64.0 64.0 59.6 70.6 71.34 72.15 69.58 59.13 .. 5.88 9.91 3.04 .. o,09] 0.11 0.02 0.'3 0.90 0.08 0.09 0.34 .. ... 0.20 0.22 5.00 ... i i : o ii:ii 2.04 0.07 0.13 .. 0.07 0.07 ... ... ... 2.54 9.25 . . . . . Tia'ce ... ... 16 0.10 13.85 8.37 9.70 2.0 13.42 5.16 8.44 3.16 0.1 ... ... ... ... 0.08 0.06 ... 0.12 0.08 . .. ... 0.06 0.06 . .. ..
{ ...
:E; .
.
I
...I
..
.. ..
.. .. ..
95 96 1.5726 1.6095 57.4 56.8 47.73 40.17 3.90 5.96 0.65 2.79 0.01 0.02 0 38 0.03 0.49 8.61 8.17 0.15 0.03 0.02 29.88 42:i5 ... ... 7.16 0.03 1.14 0.13
.:.
...
0.14 0.04 0.03
... ... .. .. .. ...
97 98 1.5159 1.5221 53.6 52.2 53.5 67.40 ... 20.0 1.72 .. 0.02 .. 0.14
..
.. ..
..
.. .. 615
..
.. .. ..
... 3.85 0.39
...
1O:ii 0.14 15.13 0.04 0.15
...
99 1.5711 43.0 54.75 0.45 0.04 0.02 0.06 0.14 0.96 0.05
...
1.64 29.30 7.99 4.31
...
0.20
FLINT GLASS 100 101 102 1.5802 1.5832 1.6163 41.0 46.3 36.6 52.53 49.80 46.12 ... 0.30 ... 0.01 0.05 0.06 0.02 0.01 0.02 0.13 0.01 0.04 ... 0.51 0.25 0.03 8.03 ... 0.04 0.07
...
34:42 8.21 3.78
...
...
0.04 0.02 0.22
.. .. ..
0.13
...
l3:36 18.74 8.20 1.24
45:i3 6.78 1.72
5i:i3 6.93 0.61
0.08
0.03 0.10
0.08
...
...
. . . . . .
0.09 0.10 0.03
...
103 104 1.6485 1.7510 33.8 27.6 40.99 29.3
...
...
.. ..
0:2
....
.. ..
67:5 3.0
.. .. .. ..
801 ... ... ... C1 .. ... ... 0.06 .. SbzOs ... ... ... ... ... ... 20 Sample 90 hard, calcd. from batch, Zsch S o . 80 ( 2 9 ) . 91, plate analysis by E. T. Allen (2.2).92 borosilicate Schott 0.144,analysis b y Allen ( 1 ) ; 93 borosilicate, S6hott 0.3832,analysis b y Allen, ( I ) ; 94, lighi barium, kchott 0.227,analyais by Zies' ( 1 ) : 95, medium barium, Schott 0.211,analysis b y Alled (P2); 96, dense barium, Schott 0.1209,analysis b y Allen ( f ) ; 97,telescope, Schott 0.608 calcd. from batch, Zsch No. 131 (Pa); 98,extra light, Schott 0.381 analysis b y E.Poaniak ( 2 2 ) .99 light, Schott 0.154,,analyeis by Allen ( 1 ) ; 100, light, Schott 0.340,analysis b y Allen ( 2 2 ) ; 101, barium, Schott 0.578,analysi; b y Posnjak ( 1 ) ; 102, m e d i i m , kchott 0.167, analysis by Zies ( 1 ) ; 103, dense, Schott 0.102,analysis b y PosnJak ( 8 2 ) ; 104, extra dense, Schott 0.165, calcd. from batch, Zsch No. 59 ( 2 9 ) .
...
S P E C T a C L E GL.4Ss
Spectacle glass is a material partaking of the characteristics of both optical and plate glass. I n general, it is made by the older methods used for plate glass-that is, by melting in an open pot or crucible and pouring the molten glass on a large iron table where it is rolled into a sheet. However, unlike plate glass its refractive index must be practically the same from melt to melt. It must have as high a transparency as possible and should not discolor under the action of sunlight. It must be extraordinarily free from bubbles and inhomogeneity. The ordinary crown glass used for most spectacles is primarily a soda-lime glass similar in composition to plate. The crown and flint glasses used for fused bifocal lenses must have the same total expansion from softening point to room temperature, and the glasses must not cloud or fog in the fusion process. I n general, spectacle crown contains 69 to 72 per cent silica, 10 to 13 per cent calcium oxide, and 14 to 16 per cent alkali, which latter is generally soda, although some potash may be employed. In addition to these major ingredients, oxides of arsenic, antimony, zinc, and boron are sometimes employed. The flint glass generally used for bifocals in this country has an index of refraction of 1.616 and therefore has a lead oxide content approximating 45 per cent, with alkali (almost invariably a mixture of soda and potash) constituting 6 to 9 per cent. Montgomery (11) has given a number of compositions of commercial and experimental spectacle glasses, a few of which are reproduced in Table XIII. Transparency is obtained in optical glass by keeping it free from impurities, such as iron, which color the glass. Conversely, if a colored glass is desired, it is obtained by adding appropriate metallic oxides to a foundation which without them would have produced a clear glass. Accordingly, for blue glass, cupric oxide or cobalt oxide is used. A green color is obtained from chromium compounds or from iron. Manganese produces violet, while uranium causes a canary yellow tint. Carbonaceous materials in a lime glass produce a fine amber, or selenium may be employed in a lead glass for the same purpose. Red is obtained by the use of selenium, cuprous oxide, or gold. Obtaining a definite color is not as simple as it sounds. For example, iron may produce any color from yellow, through green, to bright blue, according to the proportions of ferric and ferrous iron present. A glass containing one per cent of ferrous iron is almost identical in color with a glass colored by cupric oxide. If the same glass containing iron is melted under oxidizing conditions, it will be yellow, and the
...
... ...
..
color depth will be much less noticeable. The proportions of ferrous and ferric iron in commercial glasses depend on the temperature of melting, the composition of the glass, and the presence or absence of various substances which may in themselves upset the equilibrium. Therefore glass containing any iron a t all generally contains both ferrous and ferric iron, one coloring it blue and the other yellow, so that the result is green. Colored glasses for signals on railroads, highways, lighthouses, and so forth, must be quite precise in their transmission of a particular portion of the spectrum. The red glass is generally a soda-zinc-borosilicate glass colored by cadmium, sulfur, and selenium. Analyses show such glasses to contain about 70 per cent silica, 12 per cent zinc oxide, and 1 to 1.5 per cent cadmium determined as oxide; the balance i s alkali and boric oxide in proportions depending upon the process employed in making the glass and the use for which it is intended. Much more selenium must be added to the batch than can be found in an analysis. Moreover, temperature of melting, type of crucible, purity of batch ingredients, and many other factors are just as important to obtaining consistently good results as is the batch composition itself. Green signal glass is colored by cupric oxide. The casual observer in daylight would probably say that it is blue rather than green. Actually, it has a light transmission that will give the maximum impression of green to the observer when a kerosene flame is used as a light source. Such a glass, if used with a high-intensity electric lamp instead of a kerosene flame, appears much bluer than it should. Consequently, selection of the color of such a green glass takes into consideration the source of light that will be used. These green glasses are fundamentally soda-lime glasses that contain almost exactly one per cent of copper oxide as the main colorant, sometimes with other materials such as iron, boric oxide, titanium, chromium, and so forth, to modify the character of the color. The amber glasses are generally soda-lime glasses colored with carbon, although lead glasses colored with selenium have been employed, as well as combinations of iron and manganese, uranium and manganese, cadmium sulfide and selenium, etc. The type of clear glass used as a foundation for the color makes considerable difference in the color obtained. For example, in a cadmium sulfide yellow the sharpest cut-off is obtained in a barium glass. A borosilicate gives a broader region over which absorption commences and proceeds to completion. It is almost as difficult to obtain constant results in making amber glasses as it is in making red. Analyses of a number of these glasses are given in Table XIII.
I N D U S T R I A I.,
July, 1933
A N D E N G I N E E R I K G C H E hI I S T R I’
TABLEXIII. AXALYSESOF
763
SPECTACLE AND SIGXAL G L A ~ S E S
(In per cent) Sample: SiOz
Bios
105“ 69.04 0.25
Ah08 Fez03
... ...
A6103
I
ZnO CaO
MgO BaO PbO
Kz0
N320 SbzOs
MnOz
CUO
CdO Ti02
.
.
1i: 07
...
... ii:i5 5.95 0.90
... ... ...
...
106 70.63
...
0.65 0.05
...
SPECTACLE GLASS 107 108 70.36 44.45
... ...
109 40.00
0.31
0.13
...
0.20
0.15
...
... ...
2.36
42120 8.40
5i.’67 2.94 1.98
0.18
14:21
0.90 12.77
...
.. .. ..
... ...
12:44
... ... ... ...
...
li :92 Not detd. 0.97 ... ... ...
...
...
...
1.01
...
...
... ... ...
r
110 66.59 13.30 1.09
... ...
{
...
11.20
...
... ... ... ...
...
...
... ... ... ...
0.58 5.32
...
...
...
1.52
...
111 67.5 10.3 1.28
SIGKAL GL.ASS 112 113 64.53 65.00
...
...
3.10
2.9
...
...
... ... ...
, . . , . .
l7:46
...
1.12 13.50
5.80 9.5
ii:k ...
0.32 6.44
... ... ...
1.26
...
...
... ... ...
...
...
,..
10.4 7.8
...
... ... ... ... ...
114 72.9
... ... ...
4.72
... ... . .
0.22 16.47
...
i.05 ...
115 57.22 rot’ h’etd. 1.50
... ... ... ... ...
21.75
ii.44
... ...
3.86
...
2.72 ... 4merican ( I f ) ’ 106 American “strengthened” glaRs for protection goggles ( 2 , ’ 107, amethyst colored ( 2 ) : 108, flint, n~ 1.616 ( I f ) ; 109, 0 Sample 105 flint, n~ 1.66 (fij; 110 and 111,’red Lolored by selenium ( 8 ) ; 112 a n d 113, yellow, colored by caibon and sulfur ( 2 ) : 114, green, colored by copper ( 2 ) ; 115, old lead glass formerly used for green lanterns ( 2 ) .
About the most important phase of the coloration of glass removal of coloring materials. Consequently, optical glass actually is that of making it colorless. If impurities could be which must have a high transparency, is made from the purest kept out, there would be no difficulty in making colorless material available, and the only additional decolorizing is the glass. The worst offender among the impurities is iron, oxidation of the iron to the fullest practicable degree by the which always is present to some extent even in the purest of use of relatively large quantities of white oxide of arsenic and commercial sands. Less than 0.01 per cent of iron oxide, sodium nitrate. The foregoing illustrations of the use of decolorizers in glass if present completely in the ferrous form, gives a noticeable blue cast to a one-inch cube of soda-lime glass. The same are characteristic examples of effects that may be obtained quantity of iron in the ferric state gives no noticeable color. through the addition of minor ingredients. A detailed Commercial bottle glasses generally contain 0.05 to 0.07 discussion of these minor ingredients is far beyond the scope per cent of iron. Obviously such an amount causes a decided of this paper, but their importance should not be overlooked. I n this connection, fluorides and alumina added to a clear color in any ordinary sized bottle if present in the ferrous state. Even if the iron were completely oxidized to ferric glass in relatively small quantities produce opalescent glasses. oxide, the glass might have a yellowish cast. In practice it is The sulfate and chloride or other halogen ion will change an seldom possible to put the iron into either the ferrous or the opalescent to an alabaster glass. Water, added to a glass ferric state and keep it there. This results from the fact that batch in quantities of less than 5 per cent, will, in some cases, ferric oxide dissociates into ferrous oxide and oxygen a t the hasten melting. Small quantities of iron injure the transtemperatures encountered in making glass. The equilibrium mission of an ultra-violet transmitting glass. The color varies with temperature so that the higher the temperature obtained from manganese is increased many-fold if small amounts of potassium dichromate are added. In glasses the greater the proportion of ferrous oxide. As a consequence, ordinary bottle glass would have, if steps containing large amounts of boric oxide, the color obtained were not taken to avoid it, a greenish color formed from the from nickel may be altered from a yellow to a violet by the blending of the blue from ferrous iron and the yellow from addition of a halogen compound. -4similar glass containing ferric iron. The depth of color would depend primarily on cobalt may be changed from a reddish color to a green. the percentage of ferrous iron, which in turn would depend Such effects from small quantities of materials have only commenced to be appreciated in recent years, and are deservupon the temperature of melting. Fortunately there are substances which upset the ferrous- ing of much more thorough study. It is to be regretted that space will not permit detailed ferric equilibrium. Arsenious oxide is oxidized to arsenic pentoxide a t relatively low temperatures. Probably the mention of colored glasses, heat-absorbing glasses, ultrapentoxide dissociates into arsenious oxide at higher tempera- violet transmitting glasses, and opal and alabaster glass for tures, and the equilibrium of this reaction also depends on the illuminating purposes. The absurdity of listing “glass” in temperature, more oxygen being liberated as the temperature tables of physical properties without mentioning the comincreases. In any event, in the presence of arsenic the position or type of glass should be evident to the reader of this ferrous-ferric equilibrium is upset, and the proportion of brief treatment of glass compositions. If this paper brings ferrous iron present does not vary greatly with temperature. anyone to a realization of the incompleteness of such data, The use of arsenic in glass thus permits a constant color, or its purpose will have been fulfilled. rather an appearance of freedom from color to be obtained by LITERATCRE CITED the addition of some coloring matter such as selenium or manganese that will impart to the glass the pink color which (1) -411en and Zies, J . Am. Ceram. SOC.,1, 784 (1918:. (2) Bailey &. Sharp Co., Hamburg, ?;. Y., private communication, is complementary to the iron green. The effect of such a and data from records. decolorizer is analogous to the whitening of lstundered clothes (3) Baillie, J . SOC.Chem. I d . , 40, 141 (1921). by the addition of bluing. (4) Cauwood and Turner, J . SOC.Glass Tech., 2, 219 (1918). Selenium is almost invariably used as the decolorizer in (5) Chance Bros. & Co., Birmingham, Eng., private communication. (6) Dralle-Keppeler, “Die Glasfabrikation,” 1-01,2 11931). bottle glass because of the relatively constant results to be (7) English and Turner, J . SOC.Glass Tech., 4-7 (192&3). obtained. Manganese is still employed for high-grade pot ( 8 ) Fettke, “GlasJ Manufacture and Glass Sand Industry,” TODOglass. For more precise decolorization of container glass, graphic and Geol. Sun-ey of Pa , 1918. when conditions demand it, the process is carried still farther. (9) Finn and Thomson, J . Am. Ceram. SOC.,8 , 505 (1925). (10) Hovestadt, “Jena Glass,” tr. by Everett, Macmillan, 1902 The selenium pink is not exactly complementary to the iron green but leaves a faint yellow tinge in thcl glass. Cobalt (11) hfontgomery, J . Am. Ceram. SOC.,12, 274 (1929). (12) Morey, Glass I d . , 12, G9 (1931). oxide is added t o overcome this residual yellow. (13) hIorey, International Critical Tables, 1-01. 11, p. S9, hIcGrawThe fact remains that the best decolorization is complete Hill. 1927.
INDUSTRIAL AND ENGINEERING CHEMISTRY
764 (14) Morey and Bowen,
J. SOC.Glass Tech., 9, 226 (1925); Morey,
J . Am. Ceram. SOC.,13, 683, 714, 718 (1930). Morey and Merwin, J. Optical SOC.Am., 22; 632 (1932). Keumann, 2. angew. Chem., 38, 776-857 ($925).
Ibid., 40, 963 (1927). Rosenhain, “Glass Manufacture,” p. 242, Van Iiostrand, 1919. (19) Sullivan and Taylor, U. S. Patent 1,408,145 (Feb. 28, 1922).
(20) (21) (22) (23)
Vol. 25, No. 7
Turner, J . SOC.Glass Tech., 10, 81-9 (1926). Walker and Smither, J. IND.EXQ.CHEM.,9, 1090 (1917). Wright, J . Am. Ceram. Soc., 3, 783-832 (1920). Zschimmer, in Doelter’s “Handbuch der Mineralchemie,” Vol. I, pp. 869-89, T. Steinkopff, Dresden, 1914.
RECETVED April 7, 1933.
(SYMPOSIUM TO BE COKTINUED IN ACQUBTISSUE.)
The Whaling Industry LEWISRADCLIFFE, Bureau of Fisheries, Washington, D. C.
F
ISHIXG for whales dates back a thousand years. The Basques developed an important whaling industry as early as the twelfth century and by the end of the sixteenth century had extended their operations to Iceland, Newfoundlancl, and Labrador. Whaling vessels of Dutch, English, German, Danish, and Norwegian registry engaged in the fishing a t Spitzbergen and then continued westward to the coast of Greenland, Davis Strait, and Baffin Bay; this continued until the end of the eighteenth century. At the beginning of the eighteenth century the Iiew England colonists had begun fishing in the open seas for the sperm whale, and a t the time of the Revolutionary War there were more than 300 American fishing vessels engaged in this occupation. By 1846 United States vessels were reported to number 735, representing more than 230,000 tons, and there were employed in all branches of the industry about 40,000 persons. The investment was placed a t $40,000,000 and the crude value of the annual catch was about $8,000,000. New Bedford, Mass., was the center of the American industry, parts of the fleet making their headquarters a t Kantucket, Sag Harbor, Gay Head, and other Atlantic ports. During this early history the right whale and the sperm whale were the species chiefly sought, as they were relatively slow of motion and the thick layer of blubber usually sufficed to keep them afloat after they were killed, essentials to the mode of fishing then employed of harpooning the whales from rowboats. The early fishery was prosecuted chiefly for whale and sperm oil. In 1854 the United States fleet alone produced more than 12,000,000 gallons of oil; production declined rapidly thereafter. This intensive fishing led to a reduction in the supply of slow-moving whales such as the right or bowhead, sperm, and other species commonly caught by the older method. As the supply of these diminished, this type of whaling grew unprofitable and attention then turned to more abundant species such as the blue, finback, sei, and humpback. These could not be taken satisfactorily by the former method of fishing with the hand harpoon and rowboat, as they would swim away a t terrific speed when struck with the harpoon, and did not float after being killed. The perfection of the harpoon gun in 1868 by a Norwegian and of the air pump for filling the abdomen of the dead whale with air until the carcass would float, initiated a new phase in the whaling industry. Vessels propelled by steam replaced the frail rowboats and revolutionized the fishery. As these vessels were capable of towing the catch to shore, the establishment of land whaling stations was made possible. In northern seas the new whaling era reached its peak in 1905 when twenty-four companies with sixty-three vessels caught 2671 whales yielding 83,000 barrels of oil, besides by-products. The success of the Norwegian method attracted somuch attention that stations were erected in every part of the world
where conditions were favorable. Increasing expenses soon reached the point where the fishery again became unprofitable, and it was necessary to seek new sources of supply. MODERNWHALINGINDUSTRY In 1904 a Norwegian company experimented with a floating factory ship in the Spitzbergen grounds. I n 1905 this ship and two whale “catchers” visited the Falkland Islands ground and the neighborhood of Admiralty and the South Shetlands, returning with a full catch. This represents the beginning of the most recent great era in the whale fishery. Before the war about twenty-four steamers had been converted into whaling factories working chiefly in harbors. After the war about the same number wm outfitted to work in the open seas. In the beginning these floating factories were steamers of 2300 to 6000 gross tonnage and used only the blubber of the whale. These ships were gradually increased in size and provision was made for more complete utilization of the carcass. One of these ships used in 1929 is reported to have been of 32,000 tons displacement, 570 feet long, 77 feet broad, and 50.5 feet deep. This ship was equipped with the latest machinery for the extraction of oil from a maximum catch of whales. On the upper deck there was a large slipway, aft, up which the carcasses of the whales were drawn by powerful winches and derricks, making possible the cutting up of whales on deck without the flensing difficulties of former days, when the whales were cut up alongside the ship, with all the attendant handicaps. Accompanying each of the floating factories were the whale “catchers”emall steamers of 150 to 180 tons with speeds of 12 to 15 knots, or fast motor boats. These catchers cruised the adjacent seas in search of whales and towed the catch of five to nine whales back to the mother ship. Following the successful venture of these large factory ships through the ice pack into the Ross Sea, in 1923-24 whaling centered in the Antarctic. Under the conditions which obtain in those waters, it is necessary for the factory ship to negotiate the difficult passage through the ice a t the earliest practicable moment, to fill her holds with oil in about 10 weeks, and hasten out before passage becomes impossible. The large factory ships are capable of bringing back from 3.5 to 6 million gallons of oil. So successful are these ventures that oil carriers now accompany the factory ships, adding greatly to the volume of oil which may be brought out of the frozen Antarctic. Thus it has been possible for the larger factory ships to bring back the oil of eight hundred or more whales killed in a single season lasting but a few weeks. On the modern factory ship two or more whales may be cut up a t a time; great sheets of blubber 40 feet long, 20 feet Kide, and 12 to 20 inches thick are cut. Fifteen or more
