TS 700 
 
 ■■ .t3 
 
 
 Copy 2 
 
 
 
■*<fe 
 
 
 
 
 * ^ 
 * ** ^ -J8 K* -^ 
 
 
 v o^ 
 
 
 "* • • • >A& ^* 
 
 
 
 
 
 J\ 
 
 w 
 
 # 
 

 %<** 
 
 p 
 ** 
 
 
 
 
 ^ % 
 
 
 
ENAMELS 
 
 Robert D^ Landrum 
 
 rflS 
 
 I feV\V 
 
 

 mix 
 
 Publtaker 
 API 5 1S18 
 
 //■ 
 
 /O 
 
FOREWORD 
 
 This little book is a collection of the various 
 articles on Enamel published by the writer in the 
 various technical journals to which credit is given 
 with each article. 
 
 There are also included some tables of interest 
 to the Enameling industry and space is provided 
 at the end of the volume for additional data of 
 this kind. 
 
 The book is published for the benefit of the 
 Enameling industry by THE HARSHAW FULLER 
 & GOODWIN COMPANY, Cleveland, Ohio, and is 
 presented with its compliments. 
 
 Robert D. Landrum 
 January 1st, 1918. 
 
CONTENTS 
 
 Page 
 I. Enamels for Sheet Steel 7 
 
 II. The Function of the Various Raw Materials in a 
 
 Sheet Steel Enamel 15 
 
 III. Resistance of Sheet Steel Enamels to Solution 
 
 by Acetic Acids of Various Strengths 26 
 
 IV. A Comparison of Ten White Enamels for Sheet 
 
 Steel i 34 
 
 V. The Necessity of Cobalt in Ground Coat Enamels 
 
 for Sheet Steel 60 
 
 VI. Methods of Analysis for Enamel and Enamel Raw 
 
 Materials 70 
 
 VII. Atomic and Molecular Weights and Factors used 
 
 in Ceramic Calculations 101 
 
 VIII. Cubical Coefficient of Expansion! 106 
 
ENAMELS FOR SHEET STEEL *i 
 
 Enamels for sheet steel are boro-silicates of sodium, 
 potassium, calcium and aluminum and are, in every sense 
 of the word, glasses. Such enamels are so compounded 
 that they form a homogeneous, glossy coating on the sur- 
 face of the sheet steel utensil, which will not be corroded 
 by the acids or alkalies used in cooking and which will 
 resist punishment both by impact and by rapid changes 
 of temperature. 
 
 Although an enamel is a glass, the fact that it must 
 adhere to steel and resist the abuse common to cooking 
 utensils taakes necessary the addition of other ingredi- 
 ents besides those used in manufacturing ordinary glass. 
 In enamels, ground quartz, flint or sand supply the silica, 
 and feldspar and clay, the alumina. Fluorspar or cal- 
 cite is added to supply the lime and cryolite to render 
 the enamel translucent. Soda ash and pearl ash are 
 fluxes adding sodium oxide or potassium oxide to the 
 product, and borax furnishes the boric anhydride, which 
 adds many desirable qualities, such as greater ductility 
 and elasticity. Sodium or potassium nitrate is used in 
 white enamels and manganese dioxide in dark colored 
 enamels as an oxidizing agent. Oxide of cobalt is used 
 in enamels which come directly in contact with the steel 
 and adds adhesiveness to this coating. 
 
 For producing white enamels, oxide of tin is used; 
 for blue, cobalt; for violet and brown, manganese; for 
 gray, nickel ; for green, copper or chromium ; for yellow, 
 uranium or titanium ; and for red, iron, selenium or gold. 
 
 ♦Reprinted from the Journal of Industrial and Engineering Chemistry, Vol. 4. 
 No. 8, August, 1912. 
 
 1 Delivered before the Chemists' Club of Rochester at the University of Rochester, 
 Rochester, New York, April 1, 1912. 
 
Enameling is still held as a secret art, and the for- 
 mulas are carefully guarded. Most companies allow 
 very few visitors to go through their plants and some 
 keep their employees in ignorance by various schemes. 
 In one American works, each of the enamel raw-materials 
 is given a number. They are ordered, shipped, kept ac- 
 count of, and stored under their respective numbers, and 
 only those in authority even know what materials are 
 used. In this same factory, employees of one department 
 are not allowed in another and after being employed in 
 one department, a man is barred from employment in any 
 other. Some works have the formula for each enamel 
 divided into two parts, one of which is mixed by one man, 
 the other by a second, and certain proportions of each are 
 then mixed together by a third man. In practically all 
 enameling works, the materials are weighed on a scale, 
 the beam of which is hidden from the laborers, who are 
 also generally of foreign birth and are changed fre- 
 quently. 
 
 The "Black Shape." — The sheet steel which is used 
 for enameled ware is as nearly as is possible free from 
 carbon, silicon, sulphur and phosphorus, and its manga- 
 nese content is generally about 0.2 per cent. These sheets 
 come in squares and oblongs from 27 to 20 gauge and are 
 circled, stamped and spun with as little heat treatment 
 as possible and with the use of a lubricant that can easily 
 be cleaned off. The ears, handles and other trimmings 
 are, as far as is practical, welded on, as riveted joints are 
 difficult to enamel. 
 
 Pickling Process. — The surfaces of the completed 
 steel vessels are thoroughly freed from carbonaceous 
 matter by annealing at a low red-heat and are then 
 pickled in hot dilute acid, thoroughly rinsed in water, 
 and then in weak alkali solution. After a quick drying 
 they are ready to be enameled. 
 
 The Enamel. — In the making of an enamel, the vari- 
 ous raw-materials are loaded from their respective bins 
 
 8 
 
into small cars called "dollies." These are filled to a 
 line which approximates the correct weight, then they 
 are pulled on a scale, the beam of which is hidden from 
 the workman, and the enamel-master indicates whether 
 the load is light or heavy, and the workmen correct this 
 by shoveling on more or taking some off. When each of 
 the "dollies" is corrected so that the required amount of 
 material for a mix is in it, all are dumped on a large, hard 
 maple floor, the coarser material on the bottom and the 
 finer on the top. This pile is thoroughly mixed by shovel- 
 ing, and is loaded into an electric elevator, which hoists 
 it to its bin. There is a bin for each different kind of 
 enamel, and a traveling bucket which holds a melt (about 
 1200 pounds) carries the mix to the tank furnaces where 
 it is melted into a liquid glass. 
 
 These tank furnaces are regenerative, reverberatory 
 furnaces like those used in the manufacture of glass, and 
 natural gas or crude oil is an ideal fuel for them. How- 
 ever, in the older enameling works, coal is used directly, 
 and in the later ones producer-gas is used as a fuel. The 
 temperature required for smelting the different enamels 
 varies from 1000° C. for a glaze to 1300° C. for a ground 
 coat, and, in most enameling works, pyrometers are in- 
 stalled to assist in controlling these temperatures. Each 
 furnace will give seven or eight melts in twenty-four 
 hours. 
 
 After the enamel is melted into a liquid glass, a fire- 
 clay plug in the front of the furnace is pulled out and 
 the glowing liquid stream plunges out and is caught in 
 a tank of cold running water. The reaction is terrific 
 and the glass mass is torn and shredded, cracking into 
 small pieces like popcorn, each of which is a myriad of 
 microscopic seams and fissures. This "quenching," as 
 the process is called, toughens the enamel and facilitates 
 the process of grinding which comes next. 
 
 The water is drained from the tanks, leaving the 
 "enamel frit." This is shoveled into pans (a certain 
 weight to a pan) and is ready for grinding. 
 
 9 
 
In the mill room, the enamel frit is ground in large 
 ball mills for about thirty hours. These mills are cylin- 
 drical, about five feet long and six feet in diameter, and 
 are lined with porcelain bricks. The frit is put into them 
 with fifty per cent, of water and several per cent, of white 
 ball-clay. For the white cover-coat enamels, tin oxide 
 is also added. The mill revolves and the constant impact 
 of the flint stones against the glass particles grinds them 
 to an impalpable powder, which mixes with the water 
 and the clay, forming a mass which has the consistency 
 of rich cream. This is loaded into tanks, where it is 
 allowed to age a week or so. 
 
 Application of the Enamel. — From the mill room the 
 enamel is taken to the dipping room, where it is put into 
 tanks that are like large dish-pans. These are sunk into 
 tables, and at each tank a slusher works. The slusher 
 takes the stamped-out steel vessel, which has been 
 thoroughly cleaned, and plunges it into the enamel. 
 When taken out, the wet enamel forms a thin film over 
 the entire surface. By a gentle swinging motion, the 
 excess of enamel is thrown off, and the vessel is placed 
 bottom down on three metal points projecting from a 
 board. Three or four vessels are put on a board ; these 
 are placed on racks and when the vessels are thoroughly 
 dry they are carried to the furnace room. 
 
 The furnace room contains a long bank of muffle- 
 furnaces and in these the ware is put after drying. The 
 temperature in these furnaces is about 1000° C. and here 
 the little powdered particles of enamel are fused together 
 in a solid glass coating over the vessel, the process re- 
 quiring from three to five minutes. 
 
 Each coat is burned separately. For instance, we 
 have a pudding pan that is to be a three-coat white in- 
 side, turquoise-blue mottle outside. It is first dipped in 
 the ground coat enamel, the excess is shaken off and the 
 vessel put on a three-pointed rack and dried. After dry- 
 ing, the enamel stands in little grains all over the surface 
 
 10 
 
of the ware, adhering to the metal on account of the raw 
 clay ground with it. At this stage every care must be 
 taken, for a scraping, even of the finger nail, would take 
 off some of the powdered particles of the enamel. This 
 pan is then put into the muffle of the furnace, and the 
 heat fuses all the little particles together, leaving a tight- 
 holding vitreous coating all over the surface of the vessel. 
 This fundamental coating is nearly black, due to the 
 oxides of cobalt and nickel which it contains, and shines 
 with a glass-like luster. 
 
 After the vessel has cooled at the ordinary tempera- 
 ture of the room, it is again brought to the slushing room, 
 and here is covered with an enamel — this time a white. 
 It goes through the same process as before, except that 
 a black bead is brushed around the rim. On account of 
 the dark color of the first coat showing through, this 
 second coat, after it is burned, has a gray appearance, 
 and is called the "gray coat" or "first white." The vessel 
 is again sent to the slushing room, and is dipped into a 
 white enamel, the excess shaken off, and before drying 
 the blue-green enamel is sprayed on the outside. 
 
 This spraying process was at one time done by dip- 
 ping a wire brush into the wet blue-green enamel and 
 the slusher shaking it over the surface of the vessel, caus- 
 ing the blue enamel to fall in little speckles all over the 
 white enamel. In most factories, however, spraying ma- 
 chines, which work on the principle of an atomizer, have 
 been installed. A tank full of the colored enamel stands 
 over the table and the enamel is forced out through a 
 nozzle in a spray by compressed air. The flowing of the 
 enamel is controlled by the foot of the slusher as he holds 
 the vessel in the spray. The vessel is then dried and the 
 coating is fused in the muffle-furnace, the result being 
 turquoise-blue spots on a white background. 
 
 The finished ware is assorted into three lots: firsts, 
 seconds, and job lots. Some of the seconds and job lots 
 
 11 
 
are fit for redipping. They may have some little spots 
 where the original vessel was not properly cleaned and 
 where, on account of the rust or dirt, the enamel did not 
 adhere. These spots are filed or are held under a sand- 
 blast until the exposed surface is perfectly clean, and 
 then the vessel is covered with another coat of enamel. 
 
 There are schemes for saving money in all manu- 
 facturing plants, and in the enameling business a large 
 part of the profit comes from the residues. For instance, 
 every bit of enamel is scraped from the tanks and tables, 
 all sweepings from floors are saved, and all the waste 
 water from the various departments is first carried into 
 catch basins, and every few days these are cleaned and 
 the residue, which has settled to the bottom, is taken out. 
 The residues from all these sources are again melted with 
 the proper amount of fluxing material and coloring mat- 
 ter, and this dark-colored enamel is used for coating the 
 cheaper wares. 
 
 A German White Enamel. — In order to give an idea 
 of the composition of a white cover-coat frit, such as is 
 used on cooking utensils, and to show the method used 
 by ceramists to calculate its so-called molecular formula, 
 the following enamel, the formula of which is taken from 
 the 1911 1 edition of the "Taschenbuch fur Keramiker," 
 is used: 
 
 Feldspar 38.6 per cent., quartz 19.0 per cent., borax 
 15.4 per cent., cryolite 11.7 per cent., saltpeter 6.5 per 
 cent., calcite 6.5 per cent., fluorspar 1.3 per cent, and 
 magnesium carbonate 1.0 per cent. 
 
 Enamel Materials. — All the materials used were 
 practically pure except the feldspar, which was a peg- 
 matite of the following composition : 
 
 1 Page 18. Published by Keramische Rundschau, Berlin, N. W., 21. 
 
 12 
 
Per cent. 
 
 Silica (SiO») 70.66 
 
 Alumina AhO 16.85 
 
 Potassium oxide K2O 5.93 
 
 Sodium oxide NaaO 4.61 
 
 Lime CaO 0.52 
 
 Carbon dioxide CO 0.41 
 
 Moisture H2O 1.02 
 
 This figures to a "molecular" formula of 
 
 0.45Na2(n f7.11SiOi 
 
 0.39 K2O V AUOs 4 0.34 H2O 
 0.06 CaO J 10.06 CO2 
 
 the molecular weight of which would be 602. 
 
 The other materials used were : 
 
 Equivalent 
 Material Formula weight 
 
 Quartz SiO* 60 
 
 Borax NasO 2B2O3 10ILO 382 
 
 Cryolite 2NasAlFe, giving 3NasOAl20s-6F3 . . . 420 
 
 Saltpeter 2KX) NsO* 202 
 
 Calcite CaO CO2 100 
 
 Fluorspar CaFs, giving CaOFa 78 
 
 Magnesium carbonate MgO CO2 84 
 
 Feldspar (Given above) 602 
 
 The above total corresponds to the following molecu- 
 lar formula of enamel : 
 
 0.497 Na20 1 r2.513Si02 
 
 0.186 K2O I J 0.262 B*Ot 
 
 0.278 CaO 1 0.299 AbO* 1 0.599 F2 
 0.039 MgO J I 
 
 Research Laboratory 
 
 Lisk Manufacturing Co., Ltd., 
 
 Canandaigua, N. Y. 
 
 13 
 
tf : 
 
 o 
 
 to 
 
 o 
 
 00 
 
 © 
 © 
 
 •. 00 t- 
 
 O 10 © 
 
 ■<# eo 
 © © 
 
 CD 
 © 
 
 o" 
 
 o «> 
 
 
 U 
 
 o 
 
 
 
 o 
 
 
 
 
 r-l 
 
 
 o 
 
 W 
 
 o 
 
 10 
 
 3 
 
 w 
 
 o 
 
 £ 
 
 
 d 
 
 w 
 
 
 
 w 
 
 o 
 
 00 
 
 on 
 
 
 n 
 
 <M 
 
 fc 
 
 O 
 
 o t> 
 
 \a © 
 
 CD r-l 
 
 O © 
 
 d d 
 
 eo c- 
 o eo 
 rtf oo • • • • 
 
 ©©.... 
 
 d d 
 
 •S S Ht-MOiNO|>ffl 
 g "3 -*tOOt-NW(OH 
 
 (CHTfNMtOHH 
 
 © eo © © © © © © 
 dddddddd 
 
 X > 
 
 5 °* 
 
 <N O CM O <N O 
 
 © «D 00 <M © © 
 
 «D M ■* <N H 
 
 00 ""* 
 t> 00 
 
 5^3 tOOi|t-10IOMO 
 °m 00O)IOH«(SriH 
 >- 3 CO l-H r-l iH 
 
 +> -P CI 
 
 J) S O fc (j «^ 
 
 fe CPQ O OT O fe 
 
 00 
 CO 
 
 \o 
 
 r-l 
 
 f 
 
 o 
 
 as 
 
 
 * 
 
 
 
 a) 
 
 g 
 
 
 o 
 
 s 
 
 ft 
 
 5 
 
 <0 
 
 
 rt 
 
 T3 
 
 3 
 
 s 
 
 g 
 
 V 
 
 C 
 
 
 <u 
 
 a 
 
 rt 
 
 
 > 
 
 
 
 1 
 
 ■9 
 
 3 
 
 
 
 
 rt 
 
 U 
 
 ■a 
 
 a 
 
 8 
 
 14 
 
THE FUNCTION OF THE VARIOUS RAW MATERIALS 
 IN A SHEET STEEL ENAMEL* 
 
 An enamel, such as is used on cooking utensils, is a 
 glass of such nature that it will adhere to steel and on 
 account of this, its composition is more complicated than 
 that of ordinary glass. The materials used to make ordi- 
 nary glass are sand, limestone and soda ash, with lead 
 oxide added to certain types. These glasses, however, 
 will not adhere to iron and have a co-efficient of expan- 
 sion entirely different from that of sheet steel, and also 
 the temperature at which they must be melted would 
 soften the steel shape. An enamel, then, although a glass 
 in every sense of the word, and containing the elements 
 introduced by these glass materials, also contains others 
 added to modify the physical properties and to suit it for 
 the purpose intended. 
 
 Quartz and Flint 
 
 Sand is the material which furnishes silica (Si0 2 ) to 
 glass and it is sometimes used in enamels. However, 
 enamels must melt at a much lower temperature than 
 glass and thus require the silica-furnishing material to 
 be very finely powdered in order that it may combine 
 with the other materials at this lower temperature. As 
 it is more expensive to pulverize sand than it is to pul- 
 verize quartz or flint, one of these minerals — each hav- 
 ing the same chemical composition as sand — is generally 
 used. 
 
 It may be taken as a general rule that other things 
 remaining constant, the higher the per cent of silica the 
 higher will be the melting point of the enamel and the 
 
 •Reprinted from original communications, Eighth International Congress of 
 Applied Chemistry. Vol. XXV— Page 317. 
 
 15 
 
greater its acid resistance. Silica also has a low co-effi- 
 cient of expansion and increasing it in an enamel lowers 
 the co-efficient of expansion of that enamel. Therefore, 
 one method of regulating an enamel coating is to increase 
 the silica when the enamel is inclined to split off when 
 cooling after muffle burning. This remedy is indicated 
 when the curve of the enamel chips, showing that the 
 enamel contracts less rapidly on cooling than the iron. 
 
 Soda Ash 
 
 In glass, all of the sodium oxide is introduced as soda 
 ash, while in enamel, this material is only used when it is 
 desirable to add the sodium oxide without the introduc- 
 tion of any other ingredient. 
 
 Up to a certain point the addition of silica to 
 an enamel formula may lower the melting point. An 
 exaggerated example is that Calcium Oxide (lime) alone 
 is practically infusible, but on adding silica, a fairly easily 
 fusible glass is formed. 
 
 Soda Ash is commercially pure anhydrous sodium 
 carbonate (Na 2 C0 3 ) and, therefore, besides adding the 
 sodium oxide to the finished enamel, gives off carbon-di- 
 oxide gas during smelting and the melted mass is then 
 quite thoroughly stirred by the evolution of this gas and 
 many impurities are carried off. 
 
 The function of the sodium oxide is to combine with 
 the other materials (especially the silica) and form a 
 vitreous product. The larger the proportion of sodium 
 oxide (in comparison with the amount of silica present) 
 the lower will be the melting point of the product and the 
 less resistant to acids it will be. At the same time, how- 
 ever, an increase of the sodium oxide tends to make the 
 product more flexible and less brittle. 
 
 Many enamels do not include soda ash in their batch 
 mix formulas as sufficient sodium oxide is furnished by 
 the feldspar, cryolite and borax. 
 
 16 
 
Fluorspar and Calcite 
 
 Fluorspar and Calcite are the minerals which are 
 used to supply the lime (CaO) to enamels. In glass, 
 limestone furnishes this important ingredient, but on ac- 
 count of the impurities, always present in this mineral, 
 it cannot be used in the more carefully compounded enam- 
 els. Both Fluorspar and Calcite have their advantages 
 and disadvantages and, as their cost is about the same, 
 a careful consideration of these is necessary before one 
 can decide which is the better to use in an enamel. Ameri- 
 can practice is inclined to favor the use of Fluorspar. 
 
 Fluorspar 
 
 The calcium present in Fluorspar seems to combine 
 more easily with the other ingredients of the enamel 
 batch than does the calcium oxide of calcite. This is 
 evident from the fact that Fluorspar enamels melt at a 
 lower temperature and become homogeneous in a shorter 
 time than calcite enamels. In enamels, which derive some 
 of their opacity from cryolite, this is of particular advan- 
 tage, for it is a well known fact that the longer a cryolite 
 enamel is smelted, the less opaque it becomes. 
 
 The disadvantages pertinent to the use of fluorspar 
 instead of calcite in an enamel formula are due to two 
 things; this mineral contains fluorine and is a powerful 
 reducing agent at the temperature attained in the smelter. 
 The fluorine is set free during the smelting and, although 
 the virtues of fluorspar are very likely due to the ener- 
 getic action of this element, it is also active in corroding 
 the furnace linings and the life of the smelter is short- 
 ened. 
 
 The reducing action of this mineral makes it very 
 necessary to carefully regulate the smelter so that the 
 atmosphere is always oxidizing and where a large amount 
 is used the percentage of nitrate in the batch mix must 
 be increased. 
 
 17 
 
Calcite 
 
 Calcite in enamels does not act as a reducing agent 
 nor does the gas given off by it (C0 2 ) attack the furnace 
 linings. Therefore, the life of the smelter is longer and 
 there is no necessity for adding more nitrate nor of so 
 carefully controlling the atmosphere of the smelter. 
 
 The disadvantages common to a calcite enamel are 
 due to its requiring a higher temperature and a longer 
 time to combine with the other ingredients of the batch. 
 With cryolite enamels, this longer smelting is certain to 
 cause them to lose some of their opacity. 
 
 Defects Caused by Fluorspar and Calcite 
 
 Unless sufficiently smelted and thoroughly oxidized, 
 fluorspar enamels are not stable and, on standing, lose 
 their gloss, while calcite enamels, unless smelted entirely 
 homogeneous and free from carbonate, are inclined to 
 chip and scale and to form bubbles and blowholes where 
 the enamel is applied thickest. 
 
 Calcium Oxide (CaO) 
 
 The calcium oxide introduced by either of these ma- 
 terials will replace sodium oxide in an enamel formula 
 and, while not affecting the melting point to any great 
 extent, makes the enamel much harder, more acid resist- 
 ant, more glossy and more opaque. Calcium oxide in- 
 creases the brittleness of an enamel, however, but at the 
 same time it allows the addition of more boric anhydrid 
 and this ingredient counteracts this effect. 
 
 Feldspar 
 
 Feldspar is generally included in the batch mix of 
 any enamel. The new element introduced by this ma- 
 terial is aluminum in the form of alumina (A1 2 3 ). It 
 also adds sodium oxide (Na 2 0) or potassium oxide (K 2 0) 
 or both and at a much cheaper price per pound than 
 they can be bought in any other form. 
 
 18 
 
The other ingredient introduced by feldspar is silica 
 (SiO z ) and it is likely that its introduction thus has the 
 advantage over its introduction as quartz or flint, in that, 
 in feldspar, nature has already combined the alumina and 
 the silica. An enamel, therefore, getting its silica from 
 feldspar requires less heat in smelting than one getting 
 the silica otherwise. 
 
 The alumina added by the feldspar makes the glass 
 softer and also less resistant to acids, but when other 
 suitable ingredients are present, causes the enamel to be- 
 come translucent. It also lessens the tendency of the 
 enamel to chip but makes it more liable to craze. It com- 
 bines directly with the other ingredients forming a homo- 
 geneous product and, therefore, does not aid in improving 
 the stretching qualities of the product as does the alumina 
 added as clay, which material will be taken up later. 
 
 Borax and Boric Acid 
 
 One of the most characteristic ingredients of an en- 
 amel is boric anhydrid (B 2 3 ) and this is furnished to 
 the batch mix by borax or boric acid. 
 
 Borax is generally used for this purpose, as it is 
 much cheaper and the only enamel in which its use is 
 prohibited are those whose full quota of sodium oxide is 
 furnished by the feldspar and cryolite. Borax contains 
 16% of sodium oxide, so in such enamels, boric acid 
 (B 2 3 .3H 2 0) would be used. 
 
 Boric anhydrid (B 2 3 ) makes the enamel more elas- 
 tic, less brittle and changes the co-efficient of expansion in 
 such a way that the glass produced may be used on steel. 
 It is like the silica in many ways, increasing the acid re- 
 sistance and allowing more alkalies and metal and earth 
 oxides to be used in the mix and like silica causes the 
 enamel to chip when present in excess. 
 
 Unlike silica, however, it increases the viscosity of the 
 smelted enamel and lengthens the period between the 
 temperature at which the enamel will melt to form a 
 
 19 
 
homogeneous vitreous coating and the temperature at 
 which it will "burn off," thus making the enamel less 
 liable to be spoiled by poor shop practice. 
 
 This property of viscosity, which the boric anhydrid 
 increases, causes the enamel to remain thick and gummy 
 on the black shape, while being heated in the muffle, in- 
 stead of getting thin and running off as would ordinary 
 
 glass. 
 
 Clay 
 
 Clay is always used in the mill mix of an enamel that 
 is to be slushed on wet and is sometimes included in the 
 ingredients that go into the smelter. In the latter case, 
 it is used to introduce alumina and silica, as does feldspar, 
 but without introducing any alkalies. Clay is quite in- 
 fusible and very finely divided, thus adding opacity to the 
 enamel. When added at the mill, it gives besides the 
 qualities already mentioned plasticity to the wet enamel 
 and holds up the glass particles during slushing. It also 
 causes the powdered glass particles to adhere to the steel 
 shape during the drying before muffle burning. 
 
 Whether added at the mill or in the smelter, clay 
 adds to the stretching qualities of the finished enamel 
 coating. This is accounted for by the infusibility of the 
 clay, with property keeps it from entering into complete 
 combination with the other materials. Instead, it holds 
 the glass masses (with which each particle of clay is sur- 
 rounded) apart during the contraction of the steel and 
 allows them to pull away from it without chipping dur- 
 ing expansion. Under the microscope, enamels with a 
 high clay content are very porous — the higher the clay 
 the more porous — and, therefore, clay in an enamel al- 
 ways lessens the gloss. This is the prime factor in limit- 
 ing the amount of clay that can be used in an enamel. 
 
 Stellmittle 
 
 The plasticity of clays in water can be greatly in- 
 creased by adding to the water very small quantities of 
 acids, bases or salts which dissociate in the water. These 
 
 20 
 
cause the clay to assume a colloidal form quite jelly-like 
 in nature and thus make the enamel batch in which they 
 are present much thicker without removing any of the 
 water. 
 
 These are called "Stellmittle" or fixation materials 
 by the German enamelers. Any acid will answer, but, 
 as the effect on the finished product is deleterious, these 
 are seldom used. Borax (a saturated solution in hot 
 water) or a mixture of this with a saturated solution of 
 sodium carbonate is generally used in ground coat enam- 
 els and may be used in small quantities in white enamels. 
 Magnesium Sulfate has much favor with most enamelers 
 and its principal virtue is that less of it is required than 
 of any of the others. Magnesium sulfate cannot be used 
 in the ground coat, as it will cause the iron to rust. 
 
 Magnesium Oxide and Carbonate are used in the 
 mill mix and are good for this purpose, as also is ammo- 
 nium chloride, ammonium carbonate and ammonium ace- 
 tate. Some enamelers also use the sulfate of sodium but 
 any of the sulfates will impair the gloss of the product. 
 
 Oxide of Tin 
 
 Tin oxide is one of the most important and, at the 
 same time, most expensive of the enameling materials. 
 Indeed a French writer defines an enamel as "an opaque 
 glaze containing tin oxide." 
 
 Added to the smelting batch, up to about 3% of the 
 stannic oxide — the commercial oxide of tin — is reduced 
 to stanous oxide, which forms a transparent compound 
 with the silica. All over 3%, and practically all added 
 at the mill, remains in suspension in the enamel, and, 
 keeping its intense white color, makes the enamel opaque. 
 
 The writer has used as high as 30% of this material 
 in an excellent, though costly, enamel, but the amount 
 used in white enamels for cooking utensils seldom runs 
 below 5% or above 15% in the finished product. 
 
 21 
 
All efforts to entirely replace this material with a 
 cheaper one have led to the conclusion that some oxide 
 of tin is absolutely necessary in a white enamel. Other 
 materials can replace part of it, but the question as to 
 whether there is any actual saving in so doing is a matter 
 of dispute. 
 
 The higher the percentage of tin oxide in an enamel 
 the thinner it may be applied and attain a given standard 
 of whiteness. This thin application reduces the produc- 
 tion cost in two ways, viz., less enamel is required and the 
 .number of seconds, caused by the scaling off of the too 
 thick coatings, will be greatly reduced. But of still more 
 importance is the fact that the thin application of the 
 enamel adds greatly to the durability of the ware under 
 punishment both by impact and by sudden changes of 
 temperature, thus adding to the reputation of the manu- 
 facturer. Then too, the opacity produced by this mate- 
 rial is practically "fool proof." The tin oxide substitutes 
 must be handled with the greatest of care during every 
 operation and the slightest variation of method of proce- 
 dure is likely to spoil the enamel. 
 
 Cryolite 
 
 The only material which produces opacity besides 
 the tin oxide and which has stood the test of time is cryo- 
 lite. This material must be added in the smelter and al- 
 though it adds no new elements those added are so geo- 
 logically combined that at a certain temperature they 
 make the enamel frit quite translucent and even opaque 
 when thickly applied. The amount of this material which 
 can be used in a given enamel is limited by the large 
 amount of sodium oxide which it introduces. Cryolite 
 is a double fluoride of aluminum and sodium and in the 
 enamel about 20% of its mass escapes as fluorine gas and 
 is replaced by oxygen. Thus it is very necessary to keep 
 the atmosphere of the smelter oxidizing, and this is best 
 done by having sufficient nitrate present in the enamel 
 batch. The opacity produced by cryolite is quite elusive 
 
 22 
 
and the greatest care must be taken to have the tempera- 
 ture, both of the smelter and the muffle furnace, right 
 and the length of time for smelting and burning correct. 
 
 Oxide of Antimony 
 
 Antimony Oxide is another substitute for tin oxide 
 and like cryolite this is added in the smelter. Antimony- 
 containing enamels are quite opaque, if applied in thick 
 layers; thin coats of such enamels, however, are quite 
 transparent and it is doubtful whether or not antimony 
 has any great coloring effect on a properly applied 
 enamel. Antimony oxide is in itself quite poisonous and 
 when used in quantities large enough to give the desired 
 opacity, there is a danger that some of it may be made 
 soluble in the cooking acids and thus be detrimental to 
 the health of the consumer. In practice most antimony 
 enamels contain less than 5% of this material. 
 
 Both antimony and tin oxide and, in fact, all metallic 
 oxides add lustre to the enamel coating, for they increase 
 the density and it has been shown that the gloss of a glass 
 or enamel increases directly with the density. 
 
 Zinc Oxide 
 
 The oxide of zinc is the only metallic oxide, except 
 that of tin, which can be safely added to a white enamel 
 for the purpose of increasing its luster, and even it will 
 lower the resistance to corrosion by acids of such an en- 
 amel very markedly. 
 
 Used in large quantities and in very soft enamels it 
 produces some opacity. Its main use, however, is as a 
 substitute for the objectionable lead oxide in formulas 
 for colored enamels. Like lead, it has the property of 
 making such colors more brilliant. 
 
 Saltpeter and Chili Saltpeter 
 
 Potassium nitrate (saltpeter) or sodium nitrate 
 (Chili Saltpeter) is used in enamels, which require ox- 
 idizing agents in the smelter. 
 
 23 
 
Oxidizing agents are necessary in enamel formulas 
 which contain fluorspar or cryolite to replace the fluorine 
 given off in the smelter and also in white enamels con- 
 taining iron as an impurity. In the latter case, they 
 change all of the iron to the higher oxide which has a less 
 intense coloring action on the enamel. 
 
 Sodium nitrate is much cheaper than potassium ni- 
 trate but must be stored in air tight containers as it is 
 very deliquescent. The impracticability of storing the 
 material has forced manufacturers to use the more ex- 
 pensive nitrate or to employ a chemist to make daily de- 
 terminations of its moisture content. Then too, experi- 
 ment has shown that the presence of potash in an enamel 
 already containing soda tends to increase the gloss and 
 to lower the melting point without affecting the other 
 properties. This would be a reason for using nitrate of 
 potash instead of nitrate of soda in an enamel, in which 
 potash is not supplied by some other material, for potas- 
 sium oxide remains from the saltpeter, while sodium oxide 
 remains from the Chili saltpeter. 
 
 Manganese Di-Oxide 
 
 Manganese di-oxide (Mn0 2 ) is a strong oxidizing 
 agent and its use in enamels is primarily due to this fact. 
 Used in an enamel batch, it disintegrates during smelting 
 into manganous oxide (MnO) and oxygen gas and the 
 latter, besides stirring the molten enamel, changes some 
 of the ingredients to their highest possible oxides. 
 
 This action is especially desirable in ground coats 
 which must stand a hot fire in the muffle, as it renders 
 over-burning during this operation less probable. 
 
 Manganese compounds cannot be used in white en- 
 amels in more than minute quantities as it colors the glaze 
 an amethyst purple. Its second use in enamels is due 
 to this coloring action and it is used in many colored 
 enamels. 
 
 24 
 
Oxide of Cobalt 
 
 Oxide of Cobalt (C0 3 4 ) is added to enamels either 
 to give them a blue color or to make them adhere directly 
 to the steel. For the second reason, all successful ground 
 coats contain oxide of cobalt. An enamel may be so com- 
 pounded that its co-efficient of expansion will be exactly 
 that of the sheet steel upon which it is to be used, and, 
 yet without the addition of oxide of cobalt, according 
 to the writer's experience, it cannot be made to adhere 
 to the steel as well as with this addition. There are 
 many theories as to the exact function of cobalt in a 
 ground coat enamel and the popular one at present, is that 
 silicate of cobalt in the enamel frit is reduced during 
 muffle burning to a lower silicate and perhaps to metallic 
 cobalt. The oxygen, which is given off in either case then 
 unites with the iron of the black shape and is taken into 
 the enamel as ferrous silicate and, if metallic cobalt is 
 left, this unites with the iron of the black shape, forming 
 a widely distributed porous alloy. At any rate, there is 
 an interaction between the cobalt of the enamel and the 
 iron of the black shape which binds the enamel to the 
 steel. Whether this is the correct explanation of the ac- 
 tion, we do not know, but it is certain from practical ex- 
 perience that the cobalt-containing ground coat enamels 
 are more easily burned correctly by the men in charge 
 of the muffle furnaces. This is explained by the fact that 
 there is a definite, though delicate, color change from 
 blue to green in cobalt-containing ground coat enamel? 
 at just the point at which the enamel so fired will adhere 
 most firmly to the steel coat. Such an enamel, when cor- 
 rectly burned, will have a very characteristic greenish 
 tinge; when under-burned a blue color, and, when over- 
 burned, a brownish black color. 
 
 25 
 
RESISTANCE OF SHEET STEEL ENAMELS TO 
 SOLUTION BY ACETIC ACIDS OF 
 VARIOUS STRENGTHS* 
 
 Certain enameled wares have been advertised as 
 capable of withstanding 80 per cent, or 90 per cent, acetic 
 acid solutions, and although this was found to be a true 
 statement still it was misleading, for these very wares 
 were unable to resist the action of ordinary vinegar which 
 contained but 5 per cent, acetic acid. This phenomenon 
 is explained by the chemist as being due to the fact that 
 a 5 per cent, solution of acetic acid is very much more dis- 
 sociated than one of 80 or 90 per cent, and therefore its 
 dissolving action (which is directly proportional to its dis- 
 sociation) is much the greater. 
 
 To show the action of various mixtures of this com- 
 mon cooking acid and water, the two series of tests were 
 made upon enamels typical of some of the cheap wares 
 on the market. 
 
 The First Series of experiments was made upon a 
 gray enamel, mottled with dark brown. As the dark 
 enamel was made up from residues, and as the propor- 
 tions of the two enamels on the dishes is uncertain, the 
 exact molecular formula of the finished enamel coating 
 cannot be given, but the formulas as calculated from the 
 batch mix of the gray enamel and the analysis of the dark 
 enamel frit (before milling) is given below. 
 
 0.667 N'asO 
 0.108 K2O 
 0.083 CaO 
 0.057 MgO 
 0.085 ZnO 
 
 THE ENAMELS 
 
 Soft Gray Enamel 1 
 
 "1.065 Si02 
 
 0.272 AI2O3 
 
 .402 B2O3 
 
 0.532 F2 
 Milled with 7% clay 
 
 *Reprinted from Transactions of American Ceramic Society. Vol. XIII, page 494. 
 
 1 So called, as it is translucent instead of opaque and when milled without tin 
 oxide — as was the case above — the dark ground coat shows through giving a gray 
 effect. 
 
 26 
 
0.430 Na=0 
 0.083 K2O 
 0.200 CaO 
 0.045 MgO 
 0.056 CuO 
 0.010 CoO 
 0.176 MnO 
 
 Soft Dark Enamel for Spray 
 
 rl.700 SiO* 
 
 0.114 AM)s 
 
 0.430 BzOs 
 
 .0.178 Fa 
 
 Milled with 7% clay 
 
 The Test 
 
 Nineteen miniature wash-basins about 8.5 centi- 
 meters in diameter and 2 centimeters in depth were 
 slushed and burned in a dark colored ground coat, and 
 then dipped into the gray enamel slush and when the 
 excess was shaken off a light spray of the dark-colored en- 
 amel was flipped in with a brush. After drying they were 
 burned at about Seger cone 09 in a muffle furnace. They 
 were cooled in a desiccator and weighed accurately to 
 one-tenth of a milligram (0.0001 gram). Into one of them 
 was accurately measured (from two burettes graduated 
 to 1/10 of a cubic centimeter) 0.25 cubic centimeter acetic 
 acid 1 and 24.75 cubic centimeters distilled water, making 
 a 1 per cent, solution by volume. Into a second dish was 
 measured (as before) sufficient acid and water to make 
 a 2 per cent, solution. Into a third a 3 per cent, and so on 
 as given in the table following. These basins were then 
 placed upon a gas hot-plate and evaporated to dryness 
 without allowing them to boil vigorously. 
 
 When baked dry (fifteen minutes after apparent dry- 
 ness) the dissolved residue was washed out at the tap, the 
 dishes were scrubbed with a finger covered with a rubber 
 finger-stall, rinsed thoroughly with distilled water, placed 
 upon the hot-plate again until dry, cooled in a desiccator 
 and again weighed. The loss in weight, which is equal 
 to the amount of enamel dissolved in each case, is given 
 below in milligrams. 
 
 1 This acetic acid was the ordinary 
 tained 98.99% acid by weight. 
 
 c. p. 99% (1.06 sp. gr.) and by analysis con- 
 
 27 
 
The Results 
 
 Enamel Enamel 
 
 Per cent dissolved Per cent dissolved 
 
 acid Mg acid Mg 
 
 1 4.8 25 17.9 
 
 2 6.9 30 18.9 
 
 3 10.1 40 13.0 
 
 4 12.1 50 8.9 
 
 5 14.9 60 5.0 
 
 7 16.9 70 3.1 
 
 9 18.5 80 1.6 
 
 15 21.3 90 0.3 
 
 17 22.0 100 0.0 
 
 20 22.4 
 
 The Second Series was undertaken upon an enamel, 
 the definite molecular formula of which can not be given. 
 
 The Enamel used was "soft gray enamel," the 
 molecular formula of which is given above. 
 
 The Test used was the same as with the first series 
 except that the miniature wash-basins were slushed and 
 burned in three coats, viz., a dark ground, a good opaque 
 white, and the "soft gray enamel" given above which 
 was the top coat. These dishes are on exhibition and 
 the results are given in the following table. 
 
 
 The 
 
 Results 
 
 
 
 Enamel 
 
 
 Enamel 
 
 Per cent 
 
 dissolved 
 
 Per cent 
 
 dissolved 
 
 acetic acid 
 
 Mg 
 
 acetic acid 
 
 Mg 
 
 1 
 
 3.4 
 
 21 
 
 14.0 
 
 2 
 
 5.7 
 
 23 
 
 12.0 
 
 3 
 
 7.0 
 
 25 
 
 9.9 
 
 4 
 
 9.3 
 
 30 
 
 11.3 
 
 5 
 
 11.5 
 
 40 
 
 10.3 
 
 7 
 
 15.1 
 
 50 , 
 
 7.9 
 
 9 
 
 15.9 
 
 60 , 
 
 4.6 
 
 11 
 
 16.3 
 
 70 , 
 
 2.8 
 
 13 
 
 16.6 
 
 80 
 
 1.4 
 
 15 
 
 16.7 
 
 90 
 
 0.2 
 
 17 
 
 16.3 
 
 100 
 
 0.0 
 
 19 
 
 14.7 
 
 
 
 28 
 
The Enamel Solubility Curves 
 
 The accompanying sketch shows graphically the re- 
 sults of these two series of experiments. The various 
 percentages of acetic acid solutions are laid off horizon- 
 tally and the lengths of the vertical lines are proportional 
 to the amount of enamel dissolved by the corresponding 
 acid, one centimeter length of vertical line being equal 
 to one milligram of dissolved enamel. The results of the 
 first series, i. e., the one using the mottled enamel, are 
 marked "X," while those of the second series, i. e., of 
 the solid-colored enamel are marked "o." 
 
 N. B. The dip of the two curves from 21 per cent, to 
 30 per cent, acid is unexplained. Several independent 
 trials at those points tended towards proving that this dip 
 is not due to experimental error. 
 
 Discussion 
 
 MR. STALEY: This paper is interesting and in- 
 structive. As a practical method of testing the relative 
 solubility of enamels in acid solutions, the method de- 
 scribed has the commendable feature of being easily and 
 rapidly performed. In point of accuracy, it is capable of 
 being materially improved. 
 
 The shape of the solubility curve derived is very in- 
 teresting. That the solubility should decrease as the acid 
 becomes very concentrated is in accord with common ex- 
 perience. 1 But why should the solubility be greatest at 
 15 to 20 per cent, acid? Dissociation can hardly be at a 
 maximum at this high concentration. Nor are we willing 
 to grant that the solvent action of acetic acid is directly 
 proportional to its dissociation. If we leave out of con- 
 sideration the possibility that the acid solution may act 
 toward the enamel simply as a solvent, dissolving it as 
 water dissolves sugar, and treat the phenomena as a case 
 of chemical attack by an acid, we must keep in mind the 
 following considerations : 
 
 1 Foerster, "Action of Acids on Glass," Zeitschrf. Instraum., XIII, 457. 
 
 29 
 
y. 0°' 
 
 
 to 
 
 
 t 
 
 
 Q 
 
 
 to 
 
 
 ^ 
 
 In, 
 
 .£ 
 
 kj 
 
 ^ fe 
 
 \ 
 
 V § 
 
 Q- 
 
 u ? 
 
 ^ 
 
 £ . fr 
 
 •o 
 
 
 X 
 
 <b J? s 
 
 k 
 
 
 ;vj 
 
 JilO^ 
 
 L M 
 
 * o ^ 
 
 C5 
 
 x ^ . 
 
 ^>l 
 
 > * t 
 
 8 
 
 > t! ^ 
 
 it Q <0 
 
 
 8*$ 
 
 r: 
 
 
 V 
 
 ■< 
 
 ^ 
 
 T 
 
 Ql 
 
 ^ 
 
 
 kj 
 
 
 pasi/ossip /awoug /o swBj£e//ip\/ «j 
 
 30 
 
1. In dilute acid solutions, the acid is more disso- 
 ciated than in concentrated solutions. This of itself means 
 simply that we will have more action in a given time per 
 unit of acid and does not mean that the more highly ion- 
 ized acid is capable of dissolving more enamel if the reac- 
 tions are allowed to come to equilibrium. 
 
 2. Dilute acid solutions contain fewer units of acid. 
 
 3. Very highly concentrated acid solutions have 
 little action. 
 
 4. The concentration of the acid solutions varied 
 continuously as they were boiled, becoming more and 
 more concentrated as the boiling progressed. Therefore, 
 the more dilute the acid the longer the time in which ac- 
 tive concentrations would be operating. It also follows 
 from this that the slower the rate at which the acid is 
 concentrated, the greater will be its solvent action. 
 
 In accordance with these conflicting tendencies, we 
 find the acid solutions of maximum solvent action are 
 those of medium concentrations. 
 
 Acetic acid is one of the few organic acids that does 
 not form a mixture of constant boiling point with water. 
 The pure acid boils at 118° C. and in water solutions the 
 water will start to come off at 100°C, and will come off 
 the more rapidly the higher the temperature. So, if we 
 should start with a given volume of what would be in this 
 method approximately a 5 per cent, solution by volume 
 of acetic acid and place it on a hot gas plate, we would 
 soon have a smaller volume of 10 per cent., then 20 per 
 cent, and so on up to a very small volume of 100 per cent, 
 acid. The resulting solvent action would probably vary 
 materially from what would be obtained by a treatment 
 for a given time with an acetic acid solution of 5 per cent, 
 strength. The latter results which would truly corre- 
 spond to the title of the paper under discussion could be 
 obtained by the use of a return condenser. 
 
 31 
 
In order to determine the effect of the rate of evapo- 
 ration on the solvent action of an acid solution of given 
 strength, the following tests were made by the writer. 
 Four pans coated with the same enamel were treated 
 according to this method, the only variation in their treat- 
 ment being that two pans were placed on a hot portion 
 and two on a cooler portion of the same gas hot-plate. 
 Violent boiling did not occur in either case. 
 
 The results are tabulated below: 
 
 Time for Enamel 
 
 Per cent of evaporation dissolved 
 
 No. of sample acetic acid in minutes Mg 
 
 1 15 65 20.9 
 
 2 15 65 19.1 
 
 3 15 155 37.8 
 
 4 15 155 39.0 
 
 It would seem that in a test of this kind a constant 
 temperature bath should be employed. 
 
 MR. LANDRUM : I agree with Mr. Staley that my 
 title is rather misleading and might infer that this paper 
 is intended as a research in pure chemistry instead of 
 being merely a statement of the results of a series of 
 practical tests made to demonstrate in a quick and con- 
 vincing way the fact that an enameled ware may with- 
 stand the action of boiling 90 per cent, acid and still be 
 attacked by acid solutions even as dilute as those used 
 in cooking. 
 
 In these tests the conditions were very carefully kept 
 as uniform as possible, and I might add that acetic acid 
 and the method of boiling to dryness were used simply 
 because I was trying to duplicate the method used on the 
 ware advertised as "90 per cent, acid proof." I also 
 would like to state that in these series of tests all the 
 dishes in each series were put on the hot plate at the same 
 time and that this plate was of the type given an even 
 temperature to all parts of the plate (see E. H. Sargent's 
 catalog for cut of plate No. 2406). While, as stated, the 
 
 32 
 
solutions were not allowed to boil vigorously, they were 
 allowed to boil down as rapidly as possible without spat- 
 tering. From eighteen to twenty minutes were required 
 to boil to dryness. 
 
 I certainly do not advocate this as a method for test- 
 ing the acid-resistance of enameled wares and agree with 
 Mr. Staley that for a research as that seeminly indi- 
 cated by my title, a constant-temperature bath and a re- 
 flux condenser should be used. However, for purposes 
 of duplicating the treatment received by an enameled 
 dish in actual use this method of showing the action of 
 various acetic acid solutions might have some points in 
 its favor over the more accurate one suggested. 
 
 33 
 
A COMPARISON OF TEN WHITE ENAMELS FOR 
 SHEET STEEL* 
 
 This paper is the record of the manufacturing and 
 description of the physical properties of ten white enam- 
 els. It is given not with the idea of presenting to ceramic 
 literature a set of commercial formulas, but to illustrate 
 a method for testing, arranging the data, and arriving 
 at the comparative values of any enamels upon which it 
 might be desirable to experiment. 
 
 The ten enamels are those given in the "Taschenbuch 
 fur Keramiker, 1911," 1 pages eighteen and nineteen. It 
 should be stated, however, that some changes have been 
 made in the milling where it was deemed necessary, and 
 also that a feldspar high in silica has been used where 
 the formula calls for pure feldspar. 
 
 All the materials, except the borax and the saltpeter, 
 were finely ground. Crystals of these were used. The 
 enamel batches were weighed, a kilogram at a time, on 
 a balance sensitive to one hundredth of a gram. They 
 were then very thoroughly mixed and were smelted, 200 
 grams at a time, in a gas-fired crucible furnace, 2 at tem- 
 peratures varying from 1050° to 1200° C. This smelting 
 required from twelve to twenty-five minutes and, as is the 
 custom in practice, the molten enamel was poured into 
 cold water to facilitate subsequent grinding. 
 
 The resulting frits were milled — after drying— with 
 the required amount of tin oxide, clay, magnesia and 
 
 * Reprinted from the Transactions of the American Ceramic Society. Vol. XIV. 
 (Paper read at Chicago, 111., Meeting, March, 1912.) 
 
 1 Published by the Keramische Rundschau, Berlin, N. W., 21, Germany. 
 
 2 See E. H. Sargent's Catalogue No. 2098 for illustration and description of this 
 furnace. 
 
 34 
 
water, about four hundred grams at a time, in a small 
 porcelain ball mill. (This mill is manufactured by the 
 Abbe Engineering Co., N. Y., and is illustrated on page 
 eleven of their catalogue.) The time required for mill- 
 ing varied from 3% to Q% hours. 
 
 The wet enamel from the mill was slushed upon minia- 
 ture wash basins which had been previously coated with 
 a good cobalt groundcoat. After drying and burning, a 
 second coating of the same white enamel was applied. 
 Both white cover-coats were applied as thin as possible 
 and were burned in the regular muffle furnaces. 
 
 These dishes were then tested as to their resistance 
 against corrosion by acetic acid; their behavior during 
 rapid expansion and contraction; and their brittleness, 
 elasticity and adhesiveness under punishment by impact; 
 and were examined as to their opacity, gloss, etc., as a 
 finished ware. 
 
 METHODS FOR TESTING THE WARES 
 Test as to Corrosion by Acetic Acid 
 
 Each dish was carefully dried and weighed correctly 
 to 0.0001 gram; and 15 cc. of 20 per cent, acetic acid 3 
 (20 per cent, by volume of 99.5 per cent, acid) were 
 measured into it. It was then placed on a gas-fired hot 
 plate and boiled to dryness, the plate being so regulated 
 that about thirty minutes were required to bring the ves- 
 sel to dryness. The enameled dish was then washed out 
 thoroughly with distilled water, rinsed, dried on the hot 
 plate, cooled in a desiccator and again weighed. The 
 difference in weight is the amount of enamel dissolved 
 by the acid, and is recorded as "Acid Loss." The ten 
 enamels were then arranged in a list according to their 
 relative resistance to corrosion, the dish losing the least 
 being first, and so on. The position of each enamel in 
 this list is also given under "Acid Loss." 
 
 s This has been shown to be about the strongest mixture of acetic acid and water, 
 as measured by its action on an enamel. See page — . 
 
 35 
 
Tests of Adhesion Under Rapid Expansion and 
 Contraction 
 
 Test 1. — Twenty-five cubic centimeters of water 
 were heated to boiling in the dish, on a wire gauze over 
 a Bunsen flame, and the dish was then plunged into cold 
 water. The effect of this treatment on the enamel was 
 recorded. 
 
 Test 2. — The dish from Test 1 was dried, heated on 
 the wire gauze over the Bunsen flame for one minute, and 
 then plunged into cold water and any results noted. 
 
 Test 2,\/%. — In the dish from Test 2 a few cubic centi- 
 meters of water were boiled away — over the Bunsen flame 
 as in the other two tests — and then the dish was allowed 
 to remain, dry, over the flame for one minute and was 
 again plunged into cold water. (This test may seem a 
 duplication of the one above. It is not ; many commercial 
 wares fail with this test as it is especially severe.) 
 
 Test 3. — The dried dish from Test 2^ was very 
 gradually heated in the blast flame until the bottom be- 
 came red-hot. The results of this rapid expansion were 
 noted. 
 
 Test 4. — While the dish was still red-hot from Test 3 
 it was plunged into cold water and the effect of this rapid 
 contraction upon the enamel coating was recorded. 
 
 A description of the behavior of each enamel under 
 these tests is given under "Expansion and Contraction," 
 and it is to be noted that when a test is not mentioned 
 the ware was unaffected by it. Again the enamels have 
 been listed, this time according to their adhesiveness 
 under rapid changes of temperature, and their position 
 in this list is also given under "Expansion and Contrac- 
 tion." 
 
 36 
 
Test as to Adhesion Under Punishment by Impact 
 
 A testing machine, by means of which a five-pound 
 hammer with a three-quarter inch rounded head can be 
 dropped twenty and one-quarter inches onto the middle 
 of the bottom of the inverted basin, was used. The sam- 
 ple dishes were weighed correctly to 0.01 gram before 
 and after the hammer was dropped upon them, and the 
 
 TRANS. AM.CER.3CC.M1..X/Y LA/VO/fl/M 
 R/6. /. 
 
 A 
 
 A 
 
 £? 
 
 !:3L 
 
 M 
 
 "W 
 
 S~7-\ 
 
 Impact Testing Machine. 
 
 grams loss and a brief description of the effect upon the 
 enamel coating is recorded under "Loss under Hammer." 
 As before, the enamels have been arranged in a list which 
 shows their relative adhesion under punishment by im- 
 pact. 
 
 Examination as to Opacity 
 
 The finished dishes were arranged in a series accord- 
 ing to their opacity, and their position in this series as 
 well as other details as to their appearance are given 
 under "Appearance of the Ware." 
 
 37 
 
Arrangement of the Data 
 
 It is the custom of the Lisk Manufacturing Com- 
 pany's laboratory to make a complete record of each 
 enamel on a single sheet of special form, and although 
 this cannot be followed exactly in publishing this article, 
 the general arrangement and form of report will be re- 
 tained. 
 
 Immediately under the heading of the enamel the 
 batch mix in percentages and the calculated graphic 
 formula is given, and under the latter the oxygen ratio 
 and the ratio of the silica to the boric anhydride 
 
 Material Formula Equiva- Loss Cost 
 
 lent on per 
 
 Feldspar 4 0.45 Na^O, 0.39 K2O, weight ignition pound 
 
 0.06 CaO, AUO, 7.11 SiCh. .602 1.43% $0.0035 
 
 Borax NasO, 2B2O, IOH2O 382 47.0 0.0375 
 
 Quartz SiO 60 0.0025 
 
 Cryolite 2Na 3 AlFe 420 20.0 0.0600 
 
 Soda Na20, CO* 106 41.5 0.0085 
 
 Fluorspar CaFa 78 28.2 0.0045 
 
 Calcite CaO, CO2 100 43.9 0.0060 
 
 Saltpeter K2O, N2O5 202 53.4 0.0525 
 
 Carb. magnesia. .MgO, CO2 84 52.1 0.0800 
 
 Magnesia MgO 40 0.1000 
 
 Clay AbOs, 2.8 SiO*, 1.6 H*0 300 10.0 0.0100 
 
 Tin oxide SnOa 151 0.4600 
 
 4 Chemical analysis of feldspar : 
 
 Per cent. 
 
 Silica 70.66 
 
 Alumina 16.85 
 
 Potash (K 2 0) 6.93 
 
 Soda (Na 2 O f by diff.) 4.61 
 
 Lime (CaO) 0.B2 
 
 Carbon dioxide 0.41 
 
 Water 1.02 
 
 100.00 
 
 (Si0 2 /B 2 O s ). The oxygen ratio is given considering 
 A1 2 3 both as a base (ORb) and as an acid (ORa). Fol- 
 lowing these ratios is the calculated loss on smelting. 
 
 38 
 

 . '-;: 
 
 
 INSIDES 
 
 
A chemical analysis of each of the materials used 
 was made, and from them the following formulas were 
 derived. These and the other constants given in the table 
 above were used in making the calculations. 
 
 A brief description of the action of each enamel dur- 
 ing smelting and of the resulting frit is given. The mill- 
 ing follows in percentages of the weight of frit charged. 
 Thus "12 per cent, tin oxide" means "12 grams of tin 
 oxide to each hundred grams of frit." Water is added 
 to each milling equal to 50 per cent, of the weight of 
 the frit. 
 
 The latter part of the record of each enamel needs 
 no further explanation. 
 
 The Charts. — Figs. 2 and 3 show a method of ar- 
 ranging all the data in regard to an enamel in such a 
 form that it may very easily be compared with any other 
 enamel. The data of the enamels may be separated by 
 cutting along the horizontal lines of the chart and these 
 slips can be arranged in any order desired for making 
 comparisons. This is especially desirable in a series 
 where but one factor is changed at a time . 
 
 The Photographs of the Dishes. — The two half-tones 
 show the effect of the "Hammer Test" and the "Expan- 
 sion and Contraction Tests" upon both the outside and 
 inside enamel coatings of the ten wares. 
 
 Remarks. — The types of these enamels are so differ- 
 ent that the writer will not try to draw any conclusions. 
 In figuring the graphic formulas the customary method 
 has been used. In figuring the oxygen ratio, the tin oxide 
 and the fluorine have been ignored. This is incorrect in 
 almost every case, for it has been proved by experiment 
 that only about 20 per cent, of the fluorine is driven off, 
 and it is also a fact that some of the tin oxide combines 
 to form stannous silicate. This is especially shown in 
 Enamel X, for although this melted enamel contains 11.7 
 
 43 
 
per cent, of tin oxide, it is a transparent glaze. Enamel 
 II also illustrates this, for although its oxygen ration is 
 4.5 when figured in the ordinary way, it is one of the most 
 easily smelted of the enamels under discussion. 
 
 Enamel I 
 
 Feldspar 38.6% 0.497 Na20 ~| r 2 gl3 giQ2 
 
 Quartz 19.0 0.186 K2O I Q 2Q9 ., Q J „" „ B Q * 
 
 Borax 15.4 0.278 CaO °'^ 8 Ab °' 1 n Sq £ 
 
 Cryolite 11.7 0.039 MgO J I 
 
 Saltpeter 6.5 
 
 Calcite 6.5 ORb = 3.1 ORa = 6.7 SiO/B*Oa = 9.6 
 
 Fluorspar 1.3 
 
 Mg carbonate .... 1.0 
 
 Loss on Smelting. — 17.34 per cent. 
 
 Milling. — Four hours with 12 per cent, tin oxide, 7 
 per cent. Vallendar clay and ^4 P er cent, magnesia. 
 
 Smelting. — Smelted at about 1200° C, hundred gram 
 batches required about 15 minutes. The melted enamel 
 is quite viscous and is inclined to be lumpy. 
 
 The Frit. — The frit is fairly opaque but is translu- 
 cent in spots. 
 
 Acid Loss. — 0.0101 gram (fifth in list). 
 
 Expansion and Contraction. — This ware was unaf- 
 fected by heating to redness in blast flame (Test 3) and 
 came off only over a medium sized surface on the outside 
 and a comparatively small surface on the inside when 
 plunged, red-hot, into cold water. This enamel is very 
 adhesive, according to this test, and is second only to 
 Enamel II. 
 
 Loss Under Hammer. — This was 1.07 grams, and the 
 manner in which the enamel comes off shows it to be of 
 average brittleness and elasticity. It is placed fifth in 
 the list according to this test, but it is to be noticed that 
 
 44 
 
:S 
 
 z9 
 
 S w 
 "z 
 
 IS §3 
 
 8££ 
 
 eg 
 
 o 
 
 Cc 
 O 
 
 O 
 
 <JD 
 
 i> 
 
 s 
 
 I 1 
 
 
 CM 
 
 <f> 
 
 
 
 &2 
 
 
 
 $ 
 
 
 M3 
 
 <T> 
 
 ^ 
 
 a n % 
 
 
 ^ ^ ?vj ^ 
 
 
 *> 
 
 "So 
 
 -51 I 
 
 <5> Sx. 
 
 I 1 
 
 °o 
 
 
 
 
 ^ - 
 
 ^ 
 ^ 
 
 "ft 
 
 \=\ 
 
 
 T 5 
 
 O) 5"*> 
 
 ") 
 
 1> ~ 
 
 B. 
 
 M 
 
 ^ 
 O? 
 ■^ 
 
 
 «. 
 
 F 5 
 
 
 ^ 
 ^ 
 
 & 
 
 ^ 
 ^ 
 
 Hi 
 
 
 
 a 4 
 
 
 -*s 
 
 
 £5 
 
 5 s 
 
 W&1 
 
 0> 
 
 SO 
 
 
 T5 
 
 .**! 
 
 § (j & 
 
 1=1 
 
 §1 
 
 
 
 
 
 tx 
 
 I 
 $ 
 
 
 i n 
 
 & 
 
 
 <*3 
 
 "55 
 
 ^ ^ 
 ^ $ 
 
 <*S 
 
 «v? « * 
 
 w 
 
 # 
 
 # 
 
 & 
 
 ^ 
 
 <* 
 
 
 5> v 
 
 « 
 
 # 
 
 * 
 
 # 
 
 IIP 
 
 111* 
 
 45 
 
z £ 
 iq_IL 
 
 8 
 
 CUD 
 
 ggvQ 
 
 Bgg 
 
 eg 
 o 
 
 o 
 
 3 
 
 
 
 
 HL 
 
 
 
 
 
 
 
 
 r* J\j *vj 
 
 § s & S 
 
 
 « 
 
 
 fc 
 ^ 
 
 
 & 
 
 <* 
 
 
 
 & — ^ 
 
 I 
 
 Sf» 
 
 ¥1 
 
 K 
 
 "5 
 
 KS §1 
 
 
 
 fill 
 
 
 34 
 
 19 
 
 
 1_S 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1* £ «*5 
 
 W 
 
 
 ^. 
 
 fi 
 
 fcOi 
 
 29 
 
 fi 
 
 ^ Ift 
 
 I 
 
 
 s 
 
 
 
 0r> 
 
 OQ 
 
 
 w^ <sm ^ >cj 
 
 Mr 
 
 tx| 
 
 # 
 
 * 
 
 
 $> 
 
 
 mm 
 
 46 
 
all the enamels, excepting VIII, which is much the best, 
 and IX and III, which are much the worst, stand very 
 close together. 
 
 Appearance of the Ware. — This ware sets a very 
 high standard in its appearance and is much more opaque 
 than many with higher tin oxide content. In fact, it 
 would be marketable with a much smaller proportion of 
 this oxide added at the mill, and with this change would 
 be a good commercial enamel. As it stands, it is fourth 
 in the list, according to opacity. 
 
 Cost. — The cost of the materials for the finished 
 enamel is $6.70 per hundred pounds; and this could be 
 reduced by using less potassium nitrate, and, as said be- 
 fore, by using less tin at the mill. 
 
 Remarks. — If the cost is not considered, but consid- 
 ering everything else, this is the second best enamel of 
 the ten. If cost is considered, it is the very best. 
 
 Enamel II 
 
 Quartz 29.7% 0.563 NaaO ) ( 1.636 SiO« 
 
 Tin oxide 24.0 0.131 K2O f ] 0.396 B2O3 
 
 Borax 22.9 0.306 MgO ) ( 0.525 SnO* 
 
 Sodium carbonate 11.7 
 
 Saltpeter 8.0 OR = 4.5 SiCh/BsOs = 4.1 
 
 Magnesia 3.7 
 
 Loss on Smelting. — 19.89 per cent. 
 
 Milling. — Six and three-fourths hours with 4.3 per 
 cent, quartz, 2.1 per cent, tin oxide and 8 per cent. Val- 
 lendar clay. 5 
 
 Smelting. — Smelted at about 1200° C, hundred 
 gram batches required about 15 minutes. To get the 
 maximum opacity it was necessary to mechanically stir 
 just before pouring. The viscosity was medium and melt 
 was free from lumps. 
 
 The Frit. — The frit was extremely opaque and quite 
 hard and tough. 
 
 B Necessary for correct slushing but not given in original directions. 
 
 47 
 
Acid Test. — The acid test showed this enamel to be 
 absolutely unaffected by 20 per cent, acetic acid solutions, 
 and thus it is first in the list as to resistance to corrosion 
 by acid. 
 
 Expansion and Contraction. — When plunged, red- 
 hot, into cold water (Test 4) this enamel came off over 
 very small surfaces, both inside and outside, and in such 
 a manner that it is easily singled out as the most adher- 
 ent enamel of the ten. Few enamels have ever been 
 tested by the writer that have withstood sudden changes 
 of temperature as well as this one. 
 
 Loss Under Hammer. — This was 1.02 grams. This 
 enamel is of average brittleness and elasticity and is 
 placed third in the list as to adhesion under punishment 
 by impact, but an examination of the sample used for this 
 test shows that it is not as well suited to the ground coat 
 as is Enamel I. 
 
 Appearance of Ware. — This ware is very opaque, 
 and it certainly should be, as the finished enamel contains 
 something like 32 per cent, of tin oxide. It is second in 
 opacity only to Enamel VI, which contains a little over 
 23 per cent, of tin oxide when on the ware. This enamel 
 is so opaque that it may be put on in very thin coatings. 
 This is an advantage, as the thinner the enamel is applied 
 the more durable the product. 
 
 Cost. — The cost of this enamel is $15.07 per hundred 
 pounds, and this would make it entirely impractical to 
 use on a commercial ware. It is a freak enamel in every 
 sense, and is of interest on this account. 
 
 Enamel III 
 
 Borax 30.0% 0.894 Na20 ") f 2.217 SiOa 
 
 Feldspar 22.0 0.097 K2O i-0.147 AhO»^ 0.632 B*0« 
 
 Quartz 17.5 0.009 CaO J [ 0.399 SnCh 
 
 Tin oxide 15.0 
 
 Sodium carbonate . 13.5 ORb = 4.4 ORa = 6.8 SiO/BsOs = 3.5 
 Saltpeter 2.0 
 
 48 
 
Loss on Smelting. — 21.08 per cent. 
 
 Smelting. — Smelted at about 1200° C. for 15 minutes 
 or less; the melt pours with medium viscosity. 
 
 The Frit. — The frit was creamy-white and very 
 opaque. Some of the tin oxide remained in suspension, 
 but, except for this, the frit was homogeneous. 
 
 Milling. 6 — Five hours with 10 per cent. Vallendar 
 clay, and 14 per cent, magnesium oxide. 
 
 Acid Loss. — The acid loss was but 0.0016 gram. This 
 is remarkably low and places this enamel second in the 
 list as to acid resistance. 
 
 Expansion and Contraction. — This too is an excellent 
 enamel, according to the manner in which it withstands 
 punishment by rapid changes of temperature. It was 
 unaffected by any of the tests up to Test 4 and when 
 plunged, red-hot, into water during this test came off in 
 flakes, both from the inside and outside of the dish. Very 
 little steel was laid bare, but the surface of the ground 
 coat enamel which was exposed is larger than on either 
 Enamel I or Enamel II. This enamel is placed third on 
 the list. 
 
 Loss Under Hammer. — This was 1.31 grams. This 
 shows up the chief fault in this enamel — brittleness and 
 lack of elasticity — and places it next to last in the list, 
 arranged in accordance with their relative ability to with- 
 stand punishment by impact. 
 
 Appearance of Ware. — Considering the fact that this 
 enamel contains about 19 per cent, of tin oxide when on 
 the ware, it is very poor in opacity indeed. The writer 
 has made many white enamels with no other opacifier 
 than cryolite that were more opaque than this one. It is 
 fourth from the poorest among these ten. 
 
 6 Original directions gave no milling. 
 
 49 
 
Cost. — The cost of this enamel is $9.81 per hundred 
 pounds. If the cost is not considered, this enamel stands 
 fourth, considering all its properties. 
 
 Enamel IV 
 
 Quartz 31.40% 0.834 Na*0 ) f 9 571 qin 
 
 SSSST :::::::S5S S:SS?SS k 2 * 9A1 * ' fe-r 
 
 Borax 16.20 0.061 MgO J 1"'' 
 
 Sodium carbonate. 9.30 
 
 Saltpeter 3.10 ORb = 3.4 ORa = 6.8 SiO/B^Os = 9.5 
 
 Magnesia 0.80 
 
 Loss on Smelting. — 16.61 per cent. 
 
 Smelting. — Smelted at about 1100° C, four hundred 
 
 gram batches required 15 minutes. The viscosity of the 
 melt was medium. 
 
 The Frit. — The frit had little, if any, opacity and 
 with a smelting of 18 minutes became a clear glass. 
 
 Milling. — Four hours with 6.67 per cent, tin oxide, 
 4.44 per cent. Vallendar clay, and 14 per cent, magnesia. 
 
 Acid Loss. — 0.0033 gram, placing this enamel third 
 on the list. 
 
 Expansion and Contraction. — While being heated to 
 dryness (Test 21^) innumerable "nail chips" flew off, 
 both from the inside and outside surface, and when 
 heated to redness in the blast flame (Test 3) a few more 
 came off. When plunged red hot into cold water (Test 
 4) the coatings peeled off over medium sized surfaces, 
 leaving the ground coat almost bare. Notwithstanding 
 the apparent adhesiveness under Tests 3 and 4, the utter 
 failure of this enamel under Test 2 1 /£ causes it to be 
 deemed the poorest of the ten, according to its resistance 
 to punishment by rapid changes of temperature. 
 
 Loss Under Hammer. — This was 1.07 grams; and 
 although this is not far different from several others, the 
 
 50 
 
manner in which the remaining enamel adhered places 
 this ware sixth in the list. 
 
 Appearance of the Ware. — Although this enamel 
 contains but eight per cent, of tin oxide in its finished 
 coating, it is opaque enough to be placed fifth in the list. 
 The gloss is quite poor. 
 
 Cost. — The cost of this enamel is $5.03 per hundred 
 pounds. Its final rating places it fifth in the list. 
 
 Enamel V 
 
 Feldspar 35.30% 0.504 NaaO ] ("2.447 SiO* 
 
 Quartz 20.50 0.176 K2O V 0.281 A1»0»^ 0.284 B2O 
 
 Borax 16.80 0.320 CaO J Lo.636 Fa 
 
 Cryolite 12.00 
 
 Calcite 7.00 ORb = 3.1 ORa = 6.6 SiOVB^Os = 8.6 
 
 Saltpeter 6.40 
 
 Fluorspar 2.00 
 
 Loss on Smelting. — 17.86 per cent. 
 
 Smelting. — Smelted at about 1100° C, in from 18 to 
 20 minutes. The melt, when poured, was very thick and 
 sticky. This mixture is inclined to melt unevenly and 
 should be mechanically stirred for best results. 
 
 The Frit. — The frit has a translucent white color, 
 fairly good for the amount of cryolite it contains. It is 
 not noticeably hard and tough. 
 
 Milling. — Five hours with 11.76 per cent, of tin 
 oxide, 5.88 per cent, of Vallendar clay, and 14 per cent, 
 magnesia. 
 
 Acid Loss. — 0.0084 gram. This is of about the same 
 acid resistance as the best wares on the market and is 
 excelled only by Enamels I, II and IV in this list. 
 
 Expansion and Contraction. — When boiled to dry- 
 ness on asbestos gauze, this enamel chipped slightly 
 (Test 2i/£) but was not further affected when plunged 
 
 51 
 
into cold water. When heated to redness (Test 3) more 
 chips flew off the outside and the inside enamel blistered 
 somewhat. When plunged, red hot, into cold water ac- 
 cording to Test 6, the enamel came off very badly from 
 both surfaces of the dish, and much steel was laid bare. 
 This enamel stands sixth according to this test, and is 
 typical of the action of many of the wares on the market. 
 
 Loss Under Hammer. — This was 1.00 gram, showing 
 this enamel to be quite elastic. It stands next to the best, 
 according to this test. 
 
 Appearance of the Ware. — This ware is quite 
 opaque, standing third among the ten, but when one con- 
 siders that 12 per cent, of cryolite was used in the smelt 
 and that the finished enamel contains 12 per cent, of tin 
 oxide, he comes to the conclusion that much of the color 
 must have been lost. 
 
 Cost. — The cost of this enamel is $6.67 per hundred 
 pounds and its final rating places it third best, when all 
 its physical properties are considered. 
 
 Enamel VI 
 
 Feldspar 32.00% ._ „ ft 1 r 1.558 SiO* 
 
 Borax 26.00 "^» £ a fU I J .415 B2O3 
 
 Tin oxide 11.00 J'JJJ J?° [ °- 198 AK)2 1 0.452F* 
 
 Sodium carbonate. 9.00 UZ4d ^ aU J L 0.222 SnO« 
 
 Quartz 8.00 
 
 Fluorspar 6.00 ORb = 2.7 ORa = 5.0 SiOVB^Os = 3.8 
 
 Cryolite 5.00 
 
 Saltpeter 3.00 
 
 Loss on Smelting. — 20.71 per cent. 
 
 Smelting. — Smelted at about 1150° C. in from 20 to 
 22 minutes. The viscosity of this enamel was rather low. 
 
 The Frit. — The frit was creamy-white and quite 
 opaque but inclined to be non-homogeneous. Threads of 
 this enamel were brittle. 
 
 52 
 
Milling. — Five hours with 9.3 per cent, tin oxide, 7 
 per cent. Vallendar clay, and 14 P er cent, magnesia. 
 
 Acid Loss. — 0.0190 gram, placing this enamel eighth 
 on that list. 
 
 Expansion and Contraction. — During Test 3 a few 
 large chips came off while the dish was being heated to 
 redness in the blast flame. When plunged red-hot into 
 cold water, a large surface of the ground coat was laid 
 bare on the outside, but the inside was affected very much 
 less. This enamel is seventh best, according to this test. 
 
 Loss Under Hammer. — This is 1.13 grams, which is 
 the average, but places this enamel third from the last 
 when listed according to its relative resistance to punish- 
 ment by impact. 
 
 Appearance of the Ware. — This is the best appear- 
 ing enamel of the ten, and its great opacity is not to be 
 wondered at when we consider that as a finished enamel 
 it contains about 23 per cent, of tin oxide. The gloss is 
 splendid. 
 
 Cost. — The cost of this enamel is $11.10 per hundred 
 pounds; but as much more oxide of tin has been used 
 than is necessary for even a very high grade ware, this 
 could be greatly reduced. Considering everything, this 
 is the sixth best enamel of the ten. 
 
 Enamel VII 
 
 Feldspar 39.00% 0.500 NaaO -} f 2 525 gi02 
 
 Q uartz ^°° lillfl 0.303 AM)J 0.255 *0. 
 
 Borax 15.00 0.282 CaO n rqo fr„ 
 
 Cryolite 12.00 0.039 MgO J I U ' 
 
 Calcite 7.00 
 
 Saltpeter 6.00 ORb = 3.0 ORa = 6.7 SiOa/B^Os = 9.9 
 
 Fluorspar 1.00 
 
 Mg carbonate . . . 1.00 
 
 Loss on Smelting. — 17.09 per cent. 
 
 Smelting. — Smelted at about 1200° C. The melt 
 poured thin and somewhat lumpy. 
 
 53 
 
The Frit. — The frit was quite translucent and non- 
 homogeneous, a hard glassy frit. 
 
 Milling. — Three and one-half hours with 11.1 per 
 cent. Vallendar clay and ^ per cent, magnesia. 
 
 Acid Loss. — 0.0204 gram, or next to the worst in the 
 ten. 
 
 Expansion and Contraction. — This enamel chipped 
 somewhat when boiled to dryness in Test 2i/£>, but no 
 more chipping was observed while heating to redness in 
 blast flame (Test 3). When plunged, red-hot, into cold 
 water (Test 4) the enamel adhered fairly well, especially 
 on the inside, but the failure in Test 21/2 places this 
 enamel next to the poorest, according to this test. 
 
 Loss Under Hammer. — This was 1.06 grams and this 
 ware comes fourth as to its adhesiveness under punish- 
 ment by impact. 
 
 Appearance of Ware. — Considering the fact that 
 there is no oxide of tin in this enamel, it is quite opaque 
 indeed and deserves further trial with tin oxide added at 
 the mill. Even as it stands, it is more opaque than Enamel 
 X, which contains almost 12 per cent, of tin oxide. It 
 stands next to the last in the list, arranged according to 
 the relative opacity of the ware. 
 
 Cost. — On account of the absence of tin oxide in the 
 make-up of this enamel, its cost is but $2.17 per hundred 
 pounds. With the proper addition of this oxide at the 
 mill, this enamel would cost about $5.00 per hundred 
 pounds. Everything except the cost considered, this 
 enamel is rated as ninth best. 
 
 Enamel VIII 
 
 Feldspar 38.40% Q 661 N Q ] r 1.509 SiCh 
 
 S° raX .;, ?™° 111 S) 0.203 AM)J 0-463 BK). 
 
 Tin oxide 13.90 n 99S r n 1 0.215 F2 
 
 Sodium carbonate 11.30 ° ^ au J 1 0.293 SnO 
 
 Fluorspar 5.30 
 
 Saltpeter 2.00 ORb = 2.7 ORa = 5.0 SiO»/B»Os = 3.3 
 
 Quartz 1.30 
 
 54 
 
Loss on Smelting. — 20.87 per cent. 
 
 Smelting. — Smelted at about 1200° C. for from 15 
 to 18 minutes. The viscosity was rather low. Although 
 the melting point of this enamel is low, it was difficult to 
 drive off all the C0 2 . 
 
 The Frit. — The frit was creamy and very opaque, 
 but rather brittle. 
 
 Milling. — Four hours with 7.87 per cent, tin oxide, 
 4.5 per cen. Vallendar clay, and Y& per cent, magnesia. 
 
 Acid Loss. — 0.0353 gram. This is entirely too high 
 for a cooking utensil and places this ware last in the list 
 as to its resistance to corrosion by acetic acid. 
 
 Expansion and Contraction. — While heating this dish 
 to redness, a few chips flew off and when plunged, red- 
 hot, into cold water (Test 4) the enamel flaked off over 
 quite a large surface, but only to the ground coat and 
 even this was fairly well covered by the adhering cover 
 coat. This ware stands fourth according to this test. 
 
 Loss Under Hammer. — This was 0.56 gram, so that 
 according to this test the enamel is very excellent indeed. 
 It stands punishment by impact better than any of the 
 rest. 
 
 Appearance of Ware. — The opacity of this ware is 
 very low, considering the fact that it contains 25 V^ per 
 cent, of tin oxide. It stands sixth when listed according 
 to opacity. 
 
 Cost. — The cost of this ware is $21.08 per hundred 
 pounds. This is the most expensive of the wares, except 
 one, and yet it stands sixth according to opacity and 
 eighth when everything is considered. 
 
 55 
 
Enamel IX 
 
 g uartz 35.30% 0.670 Na*0 . |fl ffi F f2 . 406 Si02 
 
 Borax 20.50 0.062 K*0 I Q 22g AhQ I Q 323 BiQi 
 
 Cryolite 19.40 0.005 CaO ] n ooa -p 
 
 Feldspar 17.70 0.263 MgO J i d 
 
 Magnesia 3.50 
 
 Sodium carbonate 1.80 ORb = 3.4 ORa = 6.5 SiOa^Oa = 7.4 
 
 Saltpeter 1.80 
 
 Loss on Smelting. — 15.48 per cent. 
 
 Smelting. — Smelted at about 1200° C. for from 15 to 
 18 minutes. This enamel was difficult to pour from the 
 crucible. The enamel was quite viscous and lumpy. 
 
 The Frit. — The frit was very hard and, although 
 quite opaque, was translucent in spots. 
 
 Milling. 7 — Three and a half hours with 10 per cent. 
 Vallendar clay and *4 per cent, magnesia. 
 
 Acid Loss. — 0.0159 gram, or seventh in the list of 
 ten enamels. 
 
 Expansion and Contraction. — When plunged, red- 
 hot, into cold water (Test 4) this enamel came off over a 
 large surface on the outside and a small surface on the 
 inside of the dish. The steel was laid bare in many places. 
 In the list of wares, arranged according to their resistance 
 to punishment by change of temperature, this enamel 
 stands eighth, as it scaled slightly during Test 2i/ 2 when 
 water was boiled to dryness in it. 
 
 Loss Under Hammer. — This was 1.42 grams, which 
 is more than any of the other dishes lost, and places this 
 enamel last in the list. 
 
 Appearance of Ware. — This enamel is remarkable in 
 its opacity, all of which comes from the cryolite and clay. 
 It contains no tin oxide and stands eighth in the list ac- 
 cording to opacity. 
 
 7 Milling not given in original directions. 
 
 56 
 
Cost. — The cost is but $2.87 per hundred pounds. 
 With sufficient oxide added at the mill to bring this up to 
 a remarkable standard, this enamel would cost about 
 $5.50 per hundred pounds. Considering everything, this 
 is the poorest enamel of the ten. 
 
 Enamel X 
 
 Feldspar 45.70% 0.841 Na^O 1 f2.013 SiO 
 
 Borax 32.00 0.142 K^O f 0.283 AK>»^ 0.625 B2O3 
 
 Sodium carbonate 11.40 0.017 CaO J 1 0.227 SnO 
 
 Tin oxide 9.20 
 
 Saltpeter 1.70 ORb = 3.2 ORa = 6.8 SiO/B^Ot = 3.2 
 
 Loss on Smelting. — 21.33 per cent. 
 
 Smelting. — Smelted at about 1050° C. for 20 minutes; 
 this enamel became transparent, although it contains 9.2 
 per cent, tin oxide. Its viscosity was medium. 
 
 The Frit. — The frit was a colorless glass, compara- 
 tively hard and quite tough. 
 
 Milling. — Four and one-fourth hours with 6.4 per 
 cent. Vallendar clay and 14 per cent, magnesia. 
 
 Acid Loss. — 0.0119 gram, placing this enamel sixth 
 in the list. 
 
 Expansion and Contraction. — While heating to red- 
 ness in the blast flame, the enamel bubbled slightly, and 
 when plunged, red-hot, into cold water, came off to the 
 steel over a small surface on the inside and a larger sur- 
 face on the outside. This enamel stands fifth, according 
 to this test. 
 
 Loss Under Hammer. — This was 1.09 grams, which 
 places this cover coat seventh in the list. 
 
 Appearance of Ware. — Although this enamel con- 
 tains 11.7 per cent, tin oxide, it stands last in opacity, be- 
 ing nothing more nor less than a clear glass. 
 
 Cost. — The cost of this enamel is $7.01 per hundred 
 pounds. 
 
 57 
 
Discussion 
 
 MR. RANKIN: Mr. Landrum, I would like to ask 
 if you have ever made any experiments in the line of sub- 
 stituting sodium nitrate for potassium nitrate? 
 
 MR. LANDRUM : I have made experiments on a 
 small scale and a large scale. The substitution of sodium 
 nitrate for potassium nitrate is successful only where you 
 can make routine analyses of the sodium nitrate. Sodium 
 nitrate takes up water from the air, and thus varies in 
 strength. In the sheet steel enameling industry, the va- 
 riation of water content will spoil the enamel unless this 
 is taken into account. 
 
 I worked in one place where they determine this very 
 nicely by weighing about ten or fifteen pounds and then 
 drying and reweighing it, thus getting results which were 
 as accurate as taking a smaller sample and testing it in 
 the laboratory. 
 
 MR. STALEY: I wish to say that I consider this a 
 very able paper. 
 
 The Germans do not use the empirical formula in 
 their enamel industries to any large extent. They publish 
 their results in percentage of the various oxides. When 
 it comes to fluorides, they publish their results in percent- 
 age of the various fluorides. I think that is a fairly good 
 way. In my own papers, I prefer to consider the melted 
 weights of the various minerals; consider the feldspar, 
 for instance, as feldspar rather than as split up into the 
 oxides. The method I use is very similar to that of the 
 Germans. Mr. Landrum advocates the use of empirical 
 formulas in connection with batch weights. This diverg- 
 ence in methods of calculation shows very nicely that, 
 provided a man works in a systematic manner and has 
 enough experience, he can get results from a variety of 
 methods of calculation. 
 
 Enamels III and X are two cases where there are 
 high sodium and potassium oxides and practically no 
 
 58 
 
other basic oxide. I believe that early in the history of 
 this Society Mr. Burt gave the results of some experiments 
 in which he showed that you can melt tin oxide and 
 sodium carbonate, or tin oxide and potassium carbonate, 
 together and get a clear glass. It is known that there 
 are sodium stannates and potassium stannates in existence 
 which are translucent. I think it is more plausible to say 
 that in a material as basic as enamel the loss in opacity 
 is due to the fact that the potash and soda are high rather 
 than to attribute the poor opacity to any hypothetical 
 effect that the alkaline earths have in preventing the tin 
 from forming stannous silicate. I have always found that, 
 with a given amount of tin oxide, when sodium and po- 
 tassium run up the opacity decreases. To verify Mr. 
 Landrum, I find that increasing calcium or barium oxide 
 increases the opacity. To my mind this means simply 
 that opacity was increased by decreasing the percentage 
 of sodium and potassium. 
 
 MR. PURD Y : I would like to ask Mr. Landrum if 
 it makes any difference in the power of tin oxide to pro- 
 duce opacity, whether you frit it or use it raw? 
 
 MR. LANDRUM : Yes, there is a very marked dif- 
 ference. This has already been brought out in the Trans- 
 actions. It never pays to put tin oxide in the smelt for 
 opacity. 
 
 I have found, however, that it is good practice to put 
 a small amount of tin oxide in the smelt for gloss. In this 
 case you get no opacity from the tin oxide. For maximum 
 color, the tin oxide should be added at the mill. 
 
 59 
 
THE NECESSITY OF COBALT OXIDE IN GROUND- 
 COAT ENAMELS FOR SHEET STEEL* 
 
 There has been a great deal of discussion recently 
 in German ceramic literature over the function of the 
 ground coat and especially as to the necessity of cobalt 
 being present in such an enamel. As the durability of an 
 enameled ware is primarily dependent upon the physical 
 properties of this fundamental coating, it might be well 
 to review what has been said upon the subject. 
 
 In the Transactions of the American Ceramic Society, 
 XI, p; 115, J. B. Shaw says that cobalt-containing grounds 
 have the advantage that they change color during burn- 
 ing, the iron taken up from the steel destroying the blue 
 cobalt color. "The ground coat is considered well burnt 
 when the blue color is no longer visible. It seems to be 
 quite generally believed that CoO has a great affinity for 
 iron and that a good ground coat cannot be made without 
 using CoO." Shaw goes on to say that from his experi- 
 ments along this line he feels perfectly safe in stating 
 that this belief is ungrounded. 
 
 About the only other time that this subject is men- 
 tioned in the Transactions is by J. H. Coe, Volume XIII, 
 p. 549, and he states that the value of cobalt oxide in the 
 ground coat for cast iron is doubtful. 
 
 Dr. Grunwald in his book "Enameling on Iron and 
 Steel," p. 22, says: "Cobalt oxide possesses valuable 
 physical characteristics which make it suitable for the 
 preparation of ground (coat) enamels, for these derive 
 the property that their coefficients of expansion are as 
 near as possible the same as sheet iron." This is disputed 
 by M. Mayer and B. Havas 1 who find that ground coat 
 
 ♦Reprinted from the Transactions of the American Ceramic Society. Vol. XIV. 
 (Paper read at Chicago 111., Meeting, March, 1912.) 
 1 Chem. Ztg. Vol. XXXIII, p. 1314. 
 
 60 
 
enamels have a much lower coefficient of expansion than 
 that of sheet steel. 
 
 Dr. Vondracek, in the Sprechsaal, 1909, No. 14, 
 seems to have first propounded the theory of the function 
 of cobalt in a ground-coat enamel which is most popular 
 at present. He considers that the iron, at the melting 
 temperature of the ground coat, is oxidized at the ex- 
 pense of the cobalt oxide and that the latter, or rather 
 the cobalt silicate, is changed to a compound of lower 
 oxygen content, or is even reduced to the metal. As a 
 result of this, the clean surface of the iron is attacked so 
 that the enamel joins very intimately with the metal and 
 the danger of the chipping off of the enamel coats is 
 lessened. In another place, Dr. Vondracek, although he 
 repeats that cobalt improves the adhesiveness of enamels, 
 says : "I have, notwithstanding, often obtained a very ad- 
 hesive ground-coat enamel without using cobalt oxide." 2 
 
 Philip Eyer in his book "Die Eisenemailierung," pp. 
 10-12, says: "The idea of adding cobalt oxide to the glaze 
 so that it will go into combination with the iron is faulty, 
 for a good, adherent ground-glaze can be prepared with- 
 out the use of that oxide. However, the application of 
 such a ground coat is impossible in practice." Also in 
 writing of cast-iron enamels in the "Glashutte," Vol. 
 XLI, pp. 737-8, 764-5, he says that cobalt and nickel are 
 necessary as they form a weak alloy with the iron. 
 
 C. Tostman, in the Keramische Rundschau, XIX, pp. 
 5, 65 and 107, discusses this subject and emphasizes 
 especially, what is, in the writer's opinion, the best argu- 
 ment as to the necessity of cobalt in a ground coat. He 
 agrees with Shaw that cobalt acts as an indicator for cor- 
 rectly burning the enamel. He says in part: 
 
 "Only in cobalt oxide grounds does a blue color appear on 
 smelting. If one should discontinue the heating just at this point, 
 it would not adhere firmly enough to the steel, even though it had 
 already become molten and glass-like. One can also burn it so long 
 
 2 See Chem. Ztg., XXX, 575-7. 
 
 61 
 
that the color becomes black. Now how is this (definite) color 
 change, which is so markedly essential for a proper adhesion, to be 
 explained? The only explanation I find for it is that the enamel has 
 taken up the iron from the surface (of the black shape) in some 
 form of oxidation." 
 
 He goes on to state that while this oxygen might 
 come from the air in the muffle, it is more probable that 
 it is given up by the oxide of cobalt, which is in turn re- 
 duced to metal. "These small amounts of very finely di- 
 vided metallic cobalt could then perhaps form a very 
 porous alloy with the iron on the surface of the shape. 
 To this, the enamel would be able to adhere firmly, while 
 the silicate flux would take the place of the cobalt which 
 alloyed with the iron." He gives as an argument that this 
 oxygen is furnished to the iron by the cobalt and not 
 from any other source, the fact that in ground coats, 
 which are not colored by cobalt, "an exceedingly smaller 
 change of color takes place during the burning." He 
 also mentions the fact that the addition of borax to an 
 enamel causes it to chip and the further addition of co- 
 balt oxide seems to correct this. 
 
 Dr. Bela Havas 3 replies to Tostman that he agrees 
 with him that the cobalt silicate in the ground coat is re- 
 duced to a lower stage of oxidation and, as previously 
 published by him in cooperation with M. Mayer, 4 this is 
 indicated by the change of color of the enamel coating 
 from blue to green. However, he states that it is im- 
 probable that the reduction of the cobalt silicate at the 
 temperatures involved would go far enough to produce 
 metallic cobalt. 
 
 The writer has no new theories to offer, but he is very 
 strongly of the opinion that cobalt is a necessary ingredi- 
 ent in a successful ground coat for two reasons, which 
 have been given above : 
 
 First, it is an excellent indicator which will inform 
 the "burner" exactly the point at which his charge of 
 ware is correctly burned. 
 
 "Sprechsaal XLIV 72-3. 
 *Ibid., XLIII, 727-9. 
 
 62 
 
Fig. 1 
 
Second, whatever may be the reason, the fact re- 
 mains that cobalt grounds adhere more firmly to the steel 
 than those not containing this metal, and it is the opin- 
 ion of the writer that any non-cobalt ground coat can be 
 improved in its adhesive properties by the correct addi- 
 tion of cobalt to its formula. 
 
 To illustrate this a ground coat was prepared which 
 is practically a transparent glaze. When it is coated on 
 a piece of ware the bright steel surface shows through 
 the glaze and makes this appear a very light colored 
 
 coating (see Fig. I, 1). The composition is practically 
 the same as that of the Mayer and Havas ground No. 1 
 as given in the Sprechsaal, 1909, No. 34. 
 
 M-H. Ground Coat Number One 5 
 
 Batch mix Graphic formula 
 
 Feldspar 36.34% 0.642 Na^O -) r 
 
 Borax 35.64 O.O8O2K2O 2.256 SiO 
 
 Quartz 14.38 0.243CaO V 0.204 AI2O -I 0.231 F2 
 
 Soda 7.42 0.025 MnO 0.629 B2O3 
 
 Fluorspar 5.34 0.010 CoO J I 
 
 Manganese oxide. 0.65 
 
 Cobalt oxide 0.23 ORb 4.0, ORb 7.0, SiO/B^Oa 3.6 
 
 Milled with 6 per cent, clay and 2^ per cent, dis- 
 solved borax. 
 
 A cobalt ground coat has been prepared which has 
 the same chemical composition except that 0.03 equiva- 
 lent part of CaO has been replaced by 0.03 equivalent 
 part of CoO (see Fig. I, 2). This enamel has been made 
 as follows : 
 
 Cobalt Ground Coat 5 
 
 Batch mix Graphic formula 
 
 Feldspar 36.42% 0.642 Na 2 "I ,2.260 SiO* 
 
 Borax 35.54 0.080 K2O 
 
 Quartz 14.38 0.213 CaO fOT203 A1*0»J 0.628 BaOs 
 
 Soda 7.42 0.040 CoO 
 
 Fluorspar 4.64 0.025 MnO 1 0.201 F2 
 
 Manganese oxide. 0.65 
 
 Cobalt oxide 0.95 ORb 4.0, ORa 7.0, S^/B^Os 3.6 
 
 5 The same materials and method of manufacturing and a similar arrangement of 
 data have been used as in "Comparison of Ten White Enamels for Sheet Steel," 
 this volume. 
 
 65 
 
Milled with 6 per cent, clay and 21/2 per cent, dis- 
 solved borax. 
 
 Each of these enamels has been coated with two 
 white cover coats and then tested under the hammer (see 
 article on white enamels), and it is to be noticed that the 
 non-cobalt ground coat leaves the steel entirely bright 
 and bare (see Fig. I, 3), while the one containing cobalt 
 still adheres in concentric ridges (see Fig. I, 4). It is 
 very evident that the cobalt must cause this extra ad- 
 hesiveness. 
 
 Discussion 
 
 MR. LANDRUM : As an example of the reduction 
 of a metal oxide (or silicate) in an enamel coating by the 
 steel of the shape it may be interesting to examine a 
 sample dish coated with the following enamel: 
 
 0.451 Na*0 1 r 0.886 SiO» 
 
 0.019 K2O 
 
 0.123 CaO f 0.111 AM)* -{ 0.282 BsO* 
 
 0.285 ZnO 
 
 0.122 CuO L 0.123 F2 
 
 Milled with 6 14 per cent. clay. 
 
 This enamel when used as a cover coat (separated 
 from the steel by a ground-coat enamel) is green but 
 when applied directly to the steel becomes red through 
 the reduction of the copper compound. It may be ob- 
 served on the sample which has been dented in the test- 
 ing machine that some of the copper compound has been 
 reduced to the metal and that this is plated on the steel 
 surface. 
 
 MR. PURDY : You have tried oxides other than co- 
 balt? Have you tried nickel, for instance? 
 
 MR. LANDRUM : I have tried oxide of nickel and 
 it works fairly well, but does not promote adhesiveness 
 to as great a degree as does the oxide of cobalt. It is 
 best used in combination with that oxide. 
 
 66 
 
MR. PURDY: There is no actual oxygen for it to 
 give up. 
 
 MR. LANDRUM: We have nickelic and nickelous 
 compounds, and the change may be from one to the other 
 and not necessarily a change of the oxides ; then, too, the 
 conclusion might be drawn that the reduction is to metal- 
 lic nickel. In the case of the enamels under discussion, 
 it is easy to see that the one containing cobalt (see vessel 
 4 in illustration) does adhere to the steel after being 
 dented while the one which does not contain cobalt (ves- 
 sel 3) does not adhere but leaves the steel surface bright. 
 
 PROF. STALEY : You will have to look at it with 
 a microscope. 
 
 MR. LANDRUM: I have examined this and other 
 cobalt grounds under the microscope and have not been 
 able to see any evidence of an alloy being present. The 
 dark particles remaining after the dish is dented have a 
 gloss that would lead one to think that they are particles 
 of the glaze rather than of metal. Microscopic examina- 
 tion is difficult as it is practically impossible to get a good 
 cross section of an enameled steel sheet. 
 
 PROF. STALEY : In other words, this alloy theory 
 is used just because it fits in — because it is within reason? 
 
 MR. LANDRUM: Yes, and even then there is a 
 doubt that it is within reason at enameling temperatures. 
 
 MR. PURDY: In our transactions I stated as my 
 opinion that cobalt in the ground coat merely furnished 
 a mechanical means of holding the enamel onto the iron. 
 It is easy to enamel most metals directly, but very difficult 
 to enamel iron. The cobalt is merely suspended in the 
 ground coat furnishing the required easily enameled "go 
 between." 
 
 MR. LANDRUM : This may be true and would be 
 a good explanation. There is an interaction between the 
 iron and the enamel when you have cobalt present ; when 
 
 67 
 
you do not have cobalt present there is no interaction. 
 Sample 1 (see illustration), which is enameled with the 
 non-cobalt ground coat, is nicely covered; but under im- 
 pact the enamel comes off, leaving the bright steel (see 
 dish 3). Sample 2 is covered with the same enamel with 
 1 per cent, of oxide of cobalt added, and when subjected 
 to the same blow, adheres firmly in ridges, barely expos- 
 ing the steel between them (see dish 4). Samples Nos. 
 5 and 6 also show how the adhesiveness, under punish- 
 ment by rapid changes of temperature, is promoted by 
 the addition of oxide of cobalt. Both have been heated 
 red hot and plunged into cold water. Number 5, whose 
 fundamental coating contains no cobalt, peels off clean, 
 to the steel, while 6, whose fundamental coating is the 
 cobalt ground, exposes no steel at all. 
 
 PROF. STALEY: I would like to have Mr. Landrum 
 prove that there is an interaction between the ground 
 coat and the iron when cobalt is present and no interaction 
 when cobalt is absent. I would also like to see him de- 
 monstrate the presence of a porous alloy between the 
 enamel and iron. In regard to what Prof. Purdy has said, 
 I would like to know on what evidence he bases his state- 
 ment that cobalt is merely suspended in the ground coat 
 and, furthermore, to explain the mechanics of just how 
 such a suspension, if it should exist, would make a more 
 tenacious ground coat. We have one fact, namely, that 
 the addition of cobalt to a ground coat for sheet steel 
 enamels makes it a better ground coat. This is admir- 
 ably shown by Mr. Landrum's paper. Beyond this one 
 fact, we can merely speculate until some one produces 
 some evidence. 
 
 MR. LANDRUM : We are merely speculating as to 
 the "mechanics" of the action of the cobalt, but I believe 
 that these samples show to the naked eye that there has 
 been an interaction between the steel and the cobalt- 
 containing enamel, and that there has not been such an 
 interaction in the case of the non-cobalt enamel. It might 
 
 68 
 
throw light on the subject to consider the fact that even 
 a cobalt-containing ground coat will not adhere well to 
 a steel surface unless that surface has been made rough, 
 or porous or crystallized by pickling, sand-blasting, or 
 annealing. It may be that the enamel simply sinks down 
 into the pores of the steel, but the question is: Why 
 should an enamel have to contain cobalt to do this? Now 
 if some one has the facilities to make a cross-section of a 
 piece of enameled steel, I would like to see it. Such a 
 section might explain this matter. 
 
 * 
 
 69 
 
METHODS OF ANALYSIS FOR ENAMEL AND 
 ENAMEL RAW MATERIALS* 
 
 Introduction. 1 The fact that practically nothing has 
 been published on the above subject, and the remem- 
 brance of the many long hours spent in digging out these 
 methods and adapting them to enamels and enamel raw 
 materials, has led the author to put them in this form for 
 others who might use them. While he claims little origi- 
 nality in the methods themselves, he does claim originality 
 in the adaptations here given. Each and every one of 
 these methods has been thoroughly tried out, either in 
 the laboratory of the Columbian Enameling and Stamp- 
 ing Company, at Terre Haute, Ind., or in the chemical 
 laboratories of the University of Kansas. 
 
 PART I. 
 
 The Analysis of an Enamel 
 
 The analysis of an enamel presents one of the most 
 difficult and complicated problems with which the an- 
 alyst comes in contact. An enamel is generally an in- 
 soluble silicate containing besides silica, iron, alumina, 
 calcium, magnesium and the alkalies, generally boron, 
 fluorine, manganese, cobalt, antimony and tin, and some- 
 times phosphorus and lead. Before attempting the quan- 
 titative analysis of any enamel a thorough qualitative 
 analysis should be run, and this will enable one to choose 
 
 * Reprinted from Vol. XII, page 144, Transactions of American Ceramic Society. (Read 
 at Pittsburgh Meeting, February, 1910. 
 
 1 This paper was prepared as a thesis for the master's degree at Rose Polytechnic 
 Institute. The author desires to render thanks to Dr. W. A. Noyes and Dr. John 
 White, his former instructors, for advice freely given, and to Dr. E. H. S. Bailey 
 and Dr. H. P. Cady for suggestions offered. Methods, especially from the following 
 sources, have been freely used, and adapted to the specific uses herein described ; 
 Treadwell and Hall's "Analytical Chemistry" ; Classen's "Ausgewahlte Methoden der 
 Analytischen Chemie" ; Sutton's "Volumetric Analysis" ; Lunge and Keane's "Techni- 
 cal Methods of Chemical Analysis" ; "Methods of Agricultural Analysis" (Bui. 107, 
 U. S. Dep't of Agric.) ; Hillebrand's "Analysis of Silicate Rocks" (U. S. Geol. Survey 
 Bui. 306) ; and the files of the Journals of the various Chemical Societies. 
 
 70 
 
a quantitative separation. One of the most important 
 aids to a correct analysis is a thorough grinding. The 
 sample should be ground to an almost impalpable powder, 
 and every conceivable precaution for accuracy taken. 
 
 The analysis of a sample of enamel to be taken from 
 a piece of ware involves an extra difficulty. The coating 
 of enamel almost always consists of two or more layers — 
 the lower a large ground coat, and the upper ones white 
 or colored enamels. For an illuminating analysis these 
 must be separated. The author has found the following 
 method of V. de Luyeres 1 good for doing this : The sur- 
 face is scratched lightly with a piece of emery cloth or a 
 file, and a coating of gum acacia or glue is applied. The 
 vessel is placed in an air-bath and heated. The glue on 
 hardening generally carries with it some of the outer 
 coat. The glue or gum is then broken off, dissolved in 
 water and the enamel pieces collected on a filter paper. 
 Some obstinate enamels require painstaking methods, 
 such as chipping off with a chisel and separating the dif- 
 ferent coats — which always vary somewhat in color — by 
 picking out and sorting, using a pair of forceps. A large 
 reading glass will be useful in making these separations. 
 Any iron from the vessel which may adhere to the enamel 
 may be removed by means of a magnet after the sample 
 is ground. 
 
 Analysis of an Enamel Containing Fluorine 
 
 In an enamel containing fluorine the usual methods 
 for silicates cannot be used, as silicon-tetra-fluoride would 
 be volatilized in the evaporation with hydrochloric acid 
 for the separation of the silica. 
 
 Fluorine. One gram sample is very finely ground, 
 slowly fused with two grams each of potassium carbonate 
 and sodium carbonate. The melt should be kept in quiet 
 fusion over as low a flame as possible for one hour. The 
 melt is transferred, (after cooling quickly by giving the 
 
 1 Compte Rendus 8, p. 480. 
 
 71 
 
crucible a gyratory motion while held in the tongs, caus- 
 ing the melt to cling to the sides instead of forming a 
 solid cake in the bottom), to a platinum dish where it is 
 covered with a watch glass and boiled vigorously with 
 one hundred cc. of water. The residue is filtered off and 
 is saved for the determination of the metallic oxides and 
 the silica. 
 
 The covered solution is digested on a steam bath for 
 an hour with several grams of ammonium carbonate, and 
 on cooling more carbonate is added and the solution is 
 allowed to stand for twelve hours. The precipitate of 
 silica, alumina, etc., is filtered off, washed with ammo- 
 nium carbonate water and is saved for further determi- 
 nations. 
 
 The solution containing all the fluorine and traces of 
 silica, phosphate, etc., is evaporated until gummy, then 
 diluted with water and neutralized as follows: Phe- 
 nolphthalein is added, and nitric acid (double normal) 
 drop by drop until solution is colorless. 
 
 The solution is boiled and the red color which reap- 
 pears is again discharged with nitric acid, boiled again 
 and neutralized again until one cc. of acid will discharge 
 the color. 
 
 The last traces of silica, etc., are now removed, as 
 recommended by F. Seemann (Zeit. Anal. Chem. 44, p. 
 343), by the addition of 20 cc. of Schaffgotsch solution. 
 This solution is made as follows: 250 grams of ammonium 
 carbonate are dissolved in 180 cc. of ammonia (0.92 sp. 
 gr.) and the solution is made up to one liter. To the cold 
 solution 20 grams of freshly precipitated mercuric oxide 
 are added and the solution is vigorously shaken until the 
 mercuric oxide is dissolved. 
 
 The precipitate caused by the Schaffgotsch solution 
 is filtered off and saved, and the solution is evaporated 
 to dryness and the residue taken up with water. 
 
 72 
 
Any phosphorus from the bone ash used in some 
 enamels, and chromium which may be present, are re- 
 moved from this alkaline solution by adding silver nitrate 
 in excess. Phosphate, chromate and carbonate of silver 
 are here thrown down and may be determined if desired. 
 
 The excess of silver is removed from the solution by 
 sodium chloride, and one cc. double normal sodium car- 
 bonate solution is added to the nitrate, and the fluorine 
 is precipitated by boiling with a large excess of calcium- 
 chloride solution. 
 
 The precipitate, consisting of a mixture of calcium 
 carbonate and fluoride, is collected on a blue ribbon filter 
 paper and is washed, dried, ignited at low red heat, sep- 
 arated from the filter paper, and the residue with the ash 
 of the paper is treated with dilute acetic acid until carbon 
 dioxide is no longer given off on heating. The liquid is 
 then evaporated to dryness, the residue taken up with hot 
 water (slightly acidified with acetic acid) filtered, dried 
 and gently ignited and weighed as CaF 2 . This may be 
 checked by heating with sulfuric acid, driving off all the 
 excess of acid and reweighing as CaS0 4 . This method 
 gives results for the amount of fluorine checking within 
 0.2%, but which are generally from 2% to 4% low. 
 
 Silica. For the estimation of silica and the metallic 
 oxides, first the precipitate from the Schaffgotsch mer- 
 curic oxide solution is ignited to drive off the mercuric 
 oxide, and the silica left is weighed. The residue from 
 the original melt, together with the precipitate obtained 
 by ammonium carbonate (after the drying and removal 
 from the filter paper whose ash is added) are then dis- 
 solved in hydrochloric acid. The solution is evaporated 
 to dryness and moistened with hydrochloric acid. It is 
 diluted with water and the silica is filtered off, weighed, 
 and this with that previously obtained is the total silica. 
 
 Iron, Alumina and Manganese. The solution from 
 the silica is raised to boiling and the iron and aluminum 
 
 73 
 
are precipitated as hydroxides. Then 5 cc. of bromine 
 water is added and the boiling continued for five minutes. 
 The precipitate is dried on filter-paper and ignited separ- 
 ately from it in a weighed platinum crucible, to which 
 the ash of the filter-paper is afterwards added. The pre- 
 cipitate consists of A1 2 3 , Fe 2 3 , and Mn 2 3 , and is 
 weighed as such. It is then fused with fifteen times its 
 weight of potassium pyrosulfate over a low flame for 
 three hours with the crucible covered. The crucible, con- 
 tents and cover are placed in a beaker and dilute sulfuric 
 acid (10:1) is added. By warming and continued shak- 
 ing of liquid complete solution may be obtained. It is 
 then drawn through a Jones Reductor to change all the 
 iron to ferrous and titrated with N/10 potassium perman- 
 ganate solution. The iron is calculated to Fe,0 3 and the 
 alumina determined by difference. 
 
 If manganese is present it is determined in a separate 
 sample in a method given later and is subtracted from the 
 iron in the above. In white enamels containing only a 
 trace of iron the manganese may be determined in the 
 solution from the pyrosulfate fusion. A freshly prepared 
 solution of potassium ferrocyanide is added to oxidize 
 the manganese, then the solution is made alkaline with 
 sodium hydroxide solution and the manganese-dioxide 
 thus formed is filtered off. The solution is then made 
 acid and the ferrocyanide is titrated with N/10 potassium 
 permanganate solution. (1 cc. KMn0 4 = 0.00435 gram 
 MnO.) 
 
 Calcium Oxide. The filtrate from the iron and 
 alumina is raised to boiling, treated with boiling ammon- 
 ium oxalate solution and digested on water bath until 
 precipitate readily and quickly settles after being stirred. 
 The calcium oxalate is now filtered off and ignited wet in 
 platinum to constant weight over a strong blast. 
 
 Magnesium Oxide. The solution is evaporated to 
 dryness and the residue ignited to remove ammonium 
 salts. The residue is treated with a few drops of hydro- 
 
 74 
 
chloric acid and taken up with boiling water and filtered 
 from the carbonaceous residue. To the boiling solution 
 is added drop by drop a solution of sodium ammonium 
 phosphate and is allowed to cool. Half as much concen- 
 trated ammonium hydroxide is added as there is solution 
 and it is allowed to stand over night. The precipitate is 
 collected on a filter, washed with 3% ammonia water, 
 dried in oven and ignited separate from the filter. The 
 heat is applied gently at first and finally with the highest 
 heat of a good Bunsen burner. It is then weighed as 
 Mg 2 P 2 7 . 
 
 1 gram Mg2PsO = .3625 grams MgO 
 
 The alkalies are determined by the method of J. 
 Lawrence Smith from a gram sample finely powdered. 
 This method is standard and need not be given here. 
 
 Separation and Determination of Antimony, Tin, 
 Manganese and Cobalt in Enamel 
 
 Decomposition. Two grams finely powdered sample 
 are transferred to a platinum dish, and after moistening 
 with a little water, pure hydrofluoric acid is added and 
 the whole is mixed with a platinum spatula. The dish is 
 digested on steam bath for five hours covered with plati- 
 num cover (a larger platinum dish may be used for cover 
 if no other is at hand). After the decomposition is com- 
 plete the solution is evaporated to dryness on steam bath. 
 The residue is moistened with enough dilute sulfuric acid 
 (1:1) to make a thin paste, and evaporated as far as 
 possible on a steam bath and then on a hot plate, all the 
 time being covered to prevent spirting. As soon as fumes 
 of sulfuric anhydride cease to be evolved the cover is 
 strongly heated until fumes cease to be driven off, when 
 it is removed. The contents are heated by bringing the 
 dish to dull redness directly over a Bunsen burner. The 
 sulfates thus formed are moistened with strong hydro- 
 chloric acid, a little hot water is added and the solution 
 boiled with repeated additions of acid and water until 
 
 75 
 
completely in solution. In some enamels — especially 
 those with high melting points — the stannic oxide remains 
 undissolved, and a fusion of the residue with sulfur and 
 sodium carbonate as given later under "The Analysis of 
 Oxide of Tin" may be necessary. 
 
 Treatment with' H 2 S. The solution containing at 
 least 30 cc. double normal hydrochloric acid is transferred 
 to a 500 cc. Ehrlenmeyer flask fitted with a double bored 
 stopper. Through one of the holes a right-angled piece 
 of glass tubing is introduced that just reaches to the lower 
 edge of the stopper, while through another hole another 
 right-angled glass tube is fixed so that it almost reaches 
 the bottom of the flask. 
 
 A Kipp H 2 S generator is connected to the longer tube 
 and H 2 S is passed through for half an hour and the solu- 
 tion is let stand for another half an hour, after which the 
 sulfides of antimony and tin are transferred to a filter 
 paper and the solution is kept for the determination of 
 manganese and cobalt. 
 
 Antimony and Tin. The precipitated sulfides are 
 dissolved in a solution of potassium polysulfide — if any 
 lead or copper is present it will remain undissolved and 
 may be determined separately — by pouring this succes- 
 sively through the filter into a 300 cc. Jena beaker, and 
 finally washing with water containing a small amount of 
 potassium polysulfide. 
 
 Antimony. The antimony and tin in this solution 
 are separated by F. W. Clark's method as modified by 
 F. Henz 1 , as follows: 
 
 To the solution in the Jena beaker 6 grams caustic 
 potash and 3 grams tartaric acid are added. To this 
 mixture twice as much 30 per cent, hydrogen peroxide is 
 added as is necessary to completely decolorize the solu- 
 tion, and the latter is now heated to boiling and kept there 
 
 1 Treadwell, Vol. II, p. 188. 
 
 76 
 
until the evolution of oxygen is over, in order to oxidize 
 the thiosulphate formed. All of the excess of peroxide 
 cannot be removed successfully at this point. The solution 
 is cooled somewhat, the beaker covered with a watch- 
 glass, and a hot solution of 15 grams pure recrystallized 
 oxalic acid is cautiously added (5 gms. for 0.1 gm. of the 
 mixed metals) . This causes the evolution of considerable 
 carbon dioxide. Now, in order to completely remove the 
 excess of hydrogen peroxide the solution is boiled vigor- 
 ously for ten minutes. The volume of the liquid should 
 amount to from 80 to 100 cc. After this a rapid stream 
 of hydrogen sulfide is conducted into the boiling solution, 
 and for some time there will be no precipitation, but only 
 a white turbidity formed. At the end of five or ten min- 
 utes the solution becomes orange colored and the anti- 
 mony begins to precipitate, and from this point the time is 
 taken. At the end of fifteen minutes the solution is di- 
 luted with hot water to a volume of 250 cc, at the end of 
 another fifteen minutes the flame is removed, and ten 
 minutes later the current of hydrogen sulfide is stopped. 
 The precipitated antimony pentasulfide is filtered off 
 through a Gooch crucible which, before weighing and 
 after drying, has been heated in a stream of carbon di- 
 oxide at 300° C. for at least one hour. The precipitate 
 is washed twice by decantation with 1 per cent, oxalic 
 acid and twice with very dilute acetic acid before bring- 
 ing it in the crucible. Both of these wash liquids should 
 be boiling hot and saturated with hydrogen sulfide. 
 
 The crucible is heated in a current of carbon dioxide 
 (free from air) to constant weight, and its contents 
 weighed as Sb 2 S 3 . 
 
 The filtrate is evaporated to a volume of about 225 
 cc, transferred to a weighed unpolished platinum dish, 
 and electrolyzed at 60° to 80° C. with a current of 0.2 to 
 0.3 ampere (corresponding to 2 to 3 volts). For very 
 small amounts of tin, a current of not over 0.2 ampere 
 should be used. At the end of six hours 8 cc of sulfuric 
 
 77 
 
acid (1:1) are added, and at the end of twenty-four 
 hours the solution is transferred to another dish. The 
 deposited tin has a beautiful appearance, similar to silver. 
 
 Tin. The plated tin is washed thoroughly with water 
 and the dish is dried in an air oven at 110° and weighed. 
 
 The solution containing the cobalt and manganese is 
 boiled until free from H 2 S. The iron is oxidized back to 
 the ferric state by the addition of bromine water and 
 boiling until the excess of the latter is expelled. Ten cc. 
 double normal ammonium chloride is added and the iron 
 and alumina are precipitated by the addition of ammonia 
 and are filtered off. (The iron alumina may be deter- 
 mined from this precipitate if desired.) 
 
 The solution still containing the manganese and co- 
 balt is transferred to an Ehrlenmeyer flask fitted for pass- 
 ing in H 2 S, as before described, and 3 cc. strong ammonia 
 is added. H 2 S is passed through for some time, and after 
 precipitation ceases 3 cc. more of ammonia are added and 
 the flask is filled to the neck (300 cc. flask), is corked and 
 set aside for twelve hours at least. The precipitate is 
 collected and washed on a small filter with water contain- 
 ing ammonium chloride and sulfide. 
 
 Manganese. The manganese is extracted from the 
 precipitate on the filter by pouring through it strong H 2 S 
 water acidified with 1-5 its volume hydrochloric acid (sp. 
 gr. 1.11) . This solution from the extraction is evaporated 
 to dryness, ammonium salts are destroyed by evaporation 
 with a few drops of sodium carbonate solution, hydro- 
 chloric acid and a drop of sulfurous acid are added to 
 decompose excess of carbonate and to dissolve the pre- 
 cipitated manganese, and the latter is reprecipitated at 
 boiling heat by sodium carbonate after evaporating off 
 the hydrochloric acid. The manganese is weighed as 
 Mn 3 4 and calculated to Mn0 2 , in which form it is prob- 
 ably present in the enamel. 
 
 The residue of cobalt sulfide left after extracting the 
 manganese is burned in a porcelain crucible, dissolved in 
 
 78 
 
aqua regia, and evaporated with hydrochloric acid; the 
 platinum — and copper if any is present — are thrown 
 down by heating and passing in hydrogen sulfide. The 
 filtrate from the platinum and copper is made.ammonia- 
 cal, and cobalt is thrown down by hydrogen sulfide. This 
 is filtered off and washed with water containing ammo- 
 nium sulfide. This is either ignited and weighed as oxide 
 or more accurately determined by dissolving in an ammo- 
 niacal solution of ammonium sulfate, containing 10 grams 
 of ammonium sulfate and 40 cc. of concentrated ammonia 
 for each 0.3 grams of cobalt, and electrolyzing in a 
 weighed platinum dish at room temperature with a cur- 
 rent of 0.5 to 1.5 ampere, and an electromotive force of 
 2.8 to 3.3 volts. The electrolysis is finished in three hours. 
 The circuit is broken and the liquid poured off, and the 
 platinum dish is washed with water, then with absolute 
 alcohol (distilled one hour) and finally with ether, al- 
 lowed to dry in oven at 95° for one minute and then 
 weighed. The metallic cobalt is calculated as CoO, in 
 which form it is present in the enamel. 
 
 The Determination of Boric Anhydride in Enamel 
 
 The boron is determined in a separate sample of 
 about 0.3 grams. This finely pulverized sample is fused 
 with three grams sodium carbonate for fifteen minutes, 
 is taken up with thirty cc. dilute hydrochloric acid and a 
 few drops of nitric acid. The melt is heated in a 250 cc. 
 round-bottomed flask almost to boiling, and dry precipi- 
 tated calcium carbonate is added in moderate excess. The 
 solution is boiled in the flask after it has been connected 
 with a six-inch worm reflux condenser. The precipitate 
 is filtered on an 8 cm. Buchner 1 funnel, and is washed 
 several times with hot water, taking care that the total 
 volume of the liquid does not exceed 100 cc. 
 
 The filtrate is returned to the flask, a pinch of cal- 
 cium carbonate is added and the solution is heated to 
 
 1 See Method of Wherry and Chapin, Jr., Am. Chem. Soe. 30, p. 1688, for Deter- 
 mination of Boron in Silicates. 
 
 79 
 
boiling to remove the free carbon dioxide. This is best 
 done by connecting the flask to a suction pump, and the 
 suction is applied during boiling. The solution is cooled 
 to ordinary temperature, filtered if the precipitate has a 
 red color, and four or five drops of phenolphthalein is 
 added and N/10 sodium hydroxide solution is run in 
 slowly until liquid has a strongly pink color. A gram of 
 mannite (or 150 cc. of neutral glycerol) is added, where- 
 upon the pink color will disappear. Continue to run in 
 N/10 sodium hydroxide until end point is reached. Add 
 more mannite or glycerol and if necessary more alkali, 
 until a permanent pink color is obtained. 
 
 1 cc. N/10 Sodium Hydroxide = .0035 g. BaO« 
 
 Lead. The enamel for cooking utensils should never 
 contain lead. To determine whether a cooking utensil 
 contains lead, E. Adam gives the following simple quali- 
 tative method : A small piece of filter paper moistened 
 with hydrofluoric acid is placed upon the enamel and al- 
 lowed to remain for some minutes; the paper, together 
 with any pasty mass adhering to the enamel, is then 
 washed off into a small platinum basin, diluted with 
 water, and tested for lead by passing H 2 S through the 
 solution. 
 
 J. Grunwald (Oesterr. Chem. Ztg. 8, p. 46) gives an- 
 other quick test for lead : Wet small portion of surface 
 with HN0 3 (cone.) and heat until acid is evaporated. 
 Add several drops of water and a few drops 10% potas- 
 sium iodide solution, and if even a trace of lead is present 
 yellow lead-iodide will be produced. 
 
 Determination of Phosphoric Anhydride in Enamel 
 
 Enamels containing bone ash to give opaqueness are 
 analyzed for P 2 3 as follows: 
 
 To a gram sample of very finely pulverized enamel 
 in a platinum crucible one cc. of sulfuric acid is added 
 and the crucible is filled half full (about ten cc. are re- 
 
 80 
 
quired) with hydrofluoric acid. The crucible is heated 
 on the water bath until most of the solution is evaporated 
 and then gently on a hot plate to remove all the fluorine 
 as silicon-tetra-fluoride and as hydrofluoric acid, but no 
 sulfuric acid fumes should evolve, as P 2 5 is volatile. The 
 residue is dissolved in nitric acid and taken to dryness, 
 moistened with nitric acid, diluted with water, filtered 
 and washed with a very little water. 
 
 Add aqueous ammonia to the solution from above 
 until the precipitate of calcium phosphate first produced 
 just fails to redissolve, and then dissolve this by adding a 
 few drops of nitric acid. Warm the solution to about 
 70° C. and add 50 cc. ammonium molybdate solution (70g. 
 Mo0 3 per liter). Allow the mixture to digest at 50° for 
 twelve hours. Filter off precipitate washing by decanta- 
 tion with a solution of ammonium nitrate made acid with 
 nitric acid. 
 
 The precipitate on the filter is dissolved by pouring 
 through it dilute ammonia solution (one volume of 0.96 
 sp. gr. ammonia to three volumes of water). 
 
 The solution is received in the beaker containing the 
 bulk of the precipitate, all of which is dissolved in the 
 ammonia solution. 
 
 An excess of magnesium ammonium chloride ("mag- 
 nesia mixture") solution is added very slowly and with 
 constant stirring. Let solution stand over night. Decant 
 clear solution through a filter and wash by decantation 
 with ammonia water (1: 3). Dissolve the precipitate by 
 pouring a little hydrochloric acid (sp. gr. 1.12) through 
 the filter, allowing the acid solution to run into the beaker 
 containing most of the precipitate. When all the precipi- 
 tate on the filter and in the beaker is dissolved wash the 
 filter paper with a little hot water. To the solution add 
 2 cc. magnesia mixture and then strong ammonia, drop by 
 drop, with constant stirring until distinctly ammoniacal. 
 Stir several minutes, then add strong ammonia equal to 
 one-third of the liquid, let stand two hours and filter off 
 
 81 
 
the precipitate of magnesium ammonium phosphate. 
 Wash with dilute ammonia water, dry the precipitate, 
 ignite separately from the filter, first at low temperature 
 and gradually raise to full blast. Weigh precipitate as 
 Mg 2 P 2 7 and calculate as P 2 5 in sample. 
 
 PART II. 
 
 THE ANALYSIS OF ENAMEL RAW MATERIALS 
 The Analysis of Borax 
 
 Sampling. A handful is taken from the middle of 
 every tenth bag as it is unloaded. The sample from the 
 entire car-load is then quartered down to two pounds. 
 This is crushed so that it will pass through a forty mesh 
 sieve. This is futher quartered to about thirty grams. 
 Sample is then accurately weighed and thoroughly dis- 
 solved in about 600 cc. hot — not boiling — water in a liter 
 volumetric flask, and when cool is diluted to the mark. 
 One hundred cc. of this, representing one-tenth of the 
 sample, is then taken for analysis. 
 
 Determination of Sodium Oxide and Boric Acid. Ti- 
 trate, with normal sulfuric or hydrochloric acid solution, 
 using methyl orange as indicator. 
 
 Number cubic centimeters Normal Acid X .031 = g, 
 Na 2 0. The solution is now boiled, covered with a watch 
 glass to expel C0 2 , and on cooling may turn pink. Add 
 normal KOH solution (a drop will do) to bring back yel- 
 low color. At this stage all the boric acid exists in a free 
 state. 
 
 (2Na + +B407~ -) +HsO+ (2IP+2C1-) = (2Na + +2Cl-) +4 (H + +4BOsr) 
 
 Add as much neutral glycerol as there is solution 
 (about 150 cc.) and titrate with normal potassium hy- 
 droxide, using phenolphthalein as indicator. If end is 
 not distinct add more glycerol and more indicator. The 
 addition of glycerol causes the boric acid to become more 
 
 82 
 
dissociated, probably due to the formation of boroglyceric 
 acid, and the end-point is quite distinct. The following 
 equation represents essentially what takes place: 
 
 (4IT+4B02-) + (4K++40H-) = (4K + +4BO) +4H=0 
 1 cc. normal KOH solution — .035 g. BaO« 
 
 If the analysis gives more Na 2 than is required to 
 calculate all the B 2 3 to Na 2 B 4 7 , the remainder comes 
 from sodium carbonate with which it has been adulte- 
 rated. 
 
 Calculation of Results. The analysis of the borax is 
 very important, as many times samples are adulterated, 
 and even when not adulterated seldom contain exactly 
 enough water to give the formula Na 2 B 4 7 - 10H 2 O. It is 
 necessary to know the strength of the borax not only to 
 buy intelligently, but also so that each and every mix of 
 enamel will contain the same amount of borax. 
 
 It is customary to calculate from the percent of B 2 3 
 in sample the percent strength of the sample as Na 2 
 B 4 O 7 10H 2 O. 
 
 %B20s X 2.7307 = %Na2B*0 • lOIfcO 
 
 When the sample has dehydrated of course this will 
 run over 100%, and thus the correspondingly fewer 
 pounds of borax may be used in the mix of enamel. 
 
 Moisture. On account of the large amount of water 
 of crystallization in borax it is difficult to determine the 
 moisture directly, therefore it is calculated by subtracting 
 the % Na 2 B 4 7 10H 2 O and the % Na 2 C0 3 (if any is 
 present) from 100%. 
 
 The Analysis of Ground Sand, Flint and Quartz 
 
 Fineness. These, as are most of the raw materials 
 used in the enamel, are tested for fineness. One kilogram 
 is weighed on balance sensitive to 1/10 gram and is 
 shaken on a 100 mesh sieve. The material remaining on 
 
 83 
 
the sieve is weighed. This is then shaken on an 80 mesh 
 sieve and the residue weighed. From this is calculated 
 percent through 100 mesh and percent through 80 mesh. 
 The finer the material the better it is for use in making 
 enamel. 
 
 An analysis for Si0 2 , Fe,0 3 and MgO is run when a 
 new material is being tried, but generally only the Si0 2 
 and Fe 2 3 are determined. In this case the acid solution 
 from the silica is reduced by passing through a Jones 
 Reductor and is titrated with N/10 potassium bichromate. 
 
 Preparation for Analysis. The material is carefully 
 sampled by quartering down to several grams. This is 
 ground in an agate mortar to pass completely through a 
 hundred mesh sieve. This grinding is generally done by 
 hand but an enameling works laboratory should be 
 equipped with a McKenna Grinder, (manufactured by 
 McKenna Bros. Brass Company, Ltd., of Pittsburgh), in 
 which the material can be ground in an agate mortar by 
 power. 
 
 The method followed for the analysis of flint and 
 other forms of silica as well as clays and feldspars, is in 
 all essentials, a well known method given by Hillebrand 
 in analysis of silicate rocks, U. S. Geological Survey Bull. 
 305, and for reasons of space this method will not be 
 given here. 
 
 The Determination of Titanium in Enamels, Clays and 
 Silicate Minerals 
 
 Titanium is determined after the determination of 
 the iron by titrating with permanganate. This solution 
 (after titrating) is diluted to 1000 cc. and is treated with 
 hydrogen peroxide and the titanium determined by A. 
 Weller's Colorimetric Method, 1 from one-half the solution. 
 
 This determination depends upon the fact that acid 
 solutions of titanium sulphate are colored intensely yel- 
 
 1 Berichte 15. p. 25-93. 
 
 84 
 
low when treated with hydrogen peroxide; the yellow 
 color increases with the amount of titanium present and 
 is not altered by an excess of hydrogen peroxide. On the 
 other hand, inaccurate results are obtained in the pres- 
 ence of hydro-fluoric acid (Hillebrand) ; consequently it 
 is not permissible to use hydrogen peroxide for this de- 
 termination which has been prepared from barium per- 
 oxide by means of hydrofluosilicic acid. Furthermore, 
 chromic, vanadic, and molybdic acids must not be present, 
 since they also give colorations with hydrogen peroxide. 
 The presence of small amounts of iron do not affect the 
 reaction, but large amounts of iron cause trouble on ac- 
 count of the color of the iron solution. If, however, phos- 
 phoric acid is added to the colored ferric solution it be- 
 comes decolorized, and from such a solution the determi- 
 nation of titanium offers no difficulty. The solution in 
 which the titanium is to be determined must contain at 
 least 5 per cent, of sulfuric acid; an excess does not in- 
 fluence the reaction. The reaction is so delicate that 
 0.00005 gm. of Ti0 2 present as sulphate in 50 cc. of solu- 
 tion give a distinctly visible yellow coloration. 
 
 For this determination a standard solution of titan- 
 ium sulfate is required. This can be prepared by taking 
 0.6000 gm. of potassium titanic fluoride which has been 
 several times recrystallized and gently ignited (corres- 
 ponding to 0.2 gm. of Ti0 2 ) . This is treated in a platinum 
 crucible several times with a little water and concentrated 
 sulfuric acid, expelling the excess of acid by gentle igni- 
 tion, finally dissolving in a little concentrated sulfuric 
 acid and diluting with 5 per cent, sulfuric acid to 100 cc. 
 One cubic centimeter of this solution corresponds to 0.002 
 gm. Ti0 2 . 
 
 The determination proper is carried out in the same 
 way as the colorimetric determination of ammonium in 
 the sanitary analysis of water. 
 
 50 cc. of the solution which has been brought to a 
 definite and accurately measured volume is placed in a 
 
 85 
 
Nessler tube beside a series of other tubes, each contain- 
 ing a known amount of the standard titanium solution, 
 filled up to the mark with water and each treated with 
 
 2 cc. of 3 per cent, hydrogen peroxide 1 (free from hydro- 
 fluoric acid). The color of the solution in question is 
 compared with the standards. This method is only suit- 
 able for the estimation of small amounts of titanium, as 
 the shades of strongly colored solutions cannot be com- 
 pared accurately. 
 
 The Analysis of Oxide of Tin 
 
 Stannic Oxide. As this is one of the most important 
 and most expensive of the raw materials used in enamel- 
 ing, an analysis is very necessary. The oxide is bought 
 to contain not less than 99.5% Sn0 2 , and in this the im- 
 purities will consist of minute traces only of other ma- 
 terials. For an oxide of this kind from .2 to .3 of a gram 
 of the sample is placed in a porcelain casserole, about 
 10 cc. of C. P. nitric acid of a sp. gr. 1.2 is added and the 
 solution is slowly evaporated to a volume of about 2 or 
 
 3 cc, diluted to about 30 or 40 cc. of water, kept warm 
 for about a half hour, filtered on a small blue-ribbon filter 
 paper, and washed with warm water, slightly acidulated 
 with nitric acid, being careful to avoid letting the preci- 
 pitate creep up. 
 
 The precipitate is dried on filter paper in the funnel 
 by placing in a hot air bath. The dried tin oxide is then 
 removed as completely as possible from the filter paper 
 and the paper is ignited in a porcelain crucible, being 
 sure that there is an excess of air so that there will be no 
 metallic tin reduced. 
 
 The balance of oxide of tin is now added to the cru- 
 cible and the whole is moistened With a drop of nitric 
 acid, the temperature under the crucible is gradually 
 raised until it comes to a bright red heat over the blast 
 flame. 
 
 1 The hydrogen peroxide solution is prepared shortly before using: by dissolving 
 commercial potassium percarbonate in dilute sulfuric acid. 
 
 86 
 
This method gives results which check within one- 
 tenth of a per cent. 
 
 Some brands of oxide of tin on the market contain a 
 number of impurities in considerable quantities. Lead, 
 iron, silica, Sodium chloride, sodium sulfate and water 
 are the most common of these. These are determined as 
 follows: 
 
 Direct Method. Methods for the direct determina- 
 tion of the tin have proven quite unsatisfactory but the 
 following, with very careful manipulation, yields results 
 checking within 0.2%,: 
 
 Five-tenths grams of oxide is mixed in a porcelain 
 crucible with 3 grams each of powdered sulfur and dry 
 carbonate of soda, which both of course must be C. P., 
 especially free of metals and earths. The covered cruci- 
 ble is heated for about an hour at low heat first, and later 
 at the heat of a regular Bunsen burner ; then let cool with- 
 out lifting the cover. The cold mass is dissolved in water, 
 filtered and washed with water to which was added a 
 little sulfide of ammonia ; the residue is brought back in 
 the crucible and the melting process repeated, of which 
 the solution is filtered to the first melting. The sulfide tin 
 solution then is acidulated with hydrochloric acid and the 
 precipitated sulfide of tin is allowed to settle clearly, after 
 which it is filtered and washed with sulfide of hydrogen 
 water. 
 
 The wet precipitate of sulfide of tin is transferred 
 to an Ehrlenmeyer flask and treated with dilute hydro- 
 chloric acid and bromine until completely dissolved, at a 
 low heat. The filter left after the solution is filtered off 
 is washed and the SnCl 2 solution is precipitated with am- 
 monia and a little nitrate of ammonia, allowed to settle, 
 filtered and washed. After drying, the precipitate is 
 ignited at white heat and is weighed as SnO,. 
 
 Reduction Method. When the qualitative analysis 
 shows no metal other than tin present, a very satisfactory 
 
 87 
 
method is to reduce a weighed quantity of the sample in 
 a Rose crucible by heating to redness in a stream of hy- 
 drogen. The silica, if any is present, may be determined 
 by dissolving out the tin with hydrochloric acid and 
 weighing the residue. 
 
 Combined Water. In oxides which are prepared by 
 certain precipitation methods, the combined water runs 
 as high as ten per cent. To determine this, a two gram 
 sample is heated in a porcelain crucible at a white heat 
 to constant weight. The loss is combined water. 
 
 Lead. The lead is determined from the nitric acid 
 solution and washings from the tin oxide determination 
 by precipitation as the sulfate. 
 
 Iron. Digest about one gram with twenty-five cc. 
 of concentrated hydrochloric acid. As much water is 
 added and the solution is boiled for five minutes. The 
 residue is filtered off and about a cubic centimeter of con- 
 centrated sulfuric acid is added and the solution is evap- 
 orated until the sulfuric fumes come off. The solution is 
 diluted, passed through a Jones Reductor and titrated 
 with N/10 potassium permanganate solution. 
 
 Soluble Salts. About two grams of the sample is 
 boiled with water for thirty minutes. The residue is fil- 
 tered on a blue-ribbon paper and is dried in an air bath. 
 It is then separated as completely from the paper as is 
 possible. The paper is burned in a platinum crucible. A 
 drop of nitric acid is added and the crucible is raised to 
 bright red. The whole of the residue is now added and 
 heated to white heat for some time. (If there was com- 
 bined water present in the sample of course it will be 
 driven off, and this must be taken into calculation). 
 
 The loss in weight (minus the above correction) is 
 the soluble salts — usually sodium chloride and sulfate. 
 
 If desired these may be determined definitely by 
 usual methods. (Titration of an aliquot part with N/10 
 silver nitrate solution for the chloride and precipitation of 
 
 88 
 
the sulfate as barium sulfate in another aliquot part 
 slightly acidifies with nitric acid.) 
 
 Silica. To the residue in the platinum crucible from 
 the above determination several drops of sulfuric acid are 
 added, and the crucible is filled within a quarter of an 
 inch of the rim with pure hydrofluoric acid. This is 
 volatilized, carrying with it any of the silica as hydrofluo- 
 silicic acid. Loss of weight = Si0 2 . 
 
 The Analysis of Pyrolusite 
 
 Pyrolusite has two uses in enamel, first as an oxidiz- 
 ing agent, and second to give an amethyst color to the 
 enamel frit. Its grading, however, is generally made on 
 its oxidizing value. This is found as follows : 
 
 Manganese Dioxide. A sample is carefully taken 
 from each barrel of the shipment, and after quartering 
 down to about ten grams is ground so as to pass through 
 a 200 mesh sieve. (It is better to test this by seeing if 
 any grit can be detected when the powder is placed be- 
 tween the teeth.) The sample is dried, spread out on a 
 watch glass, at 110° for one hour, transferred to a stop- 
 pered weighing tube, and after weighing, about one-half 
 gram is transferred into a 250 cc. Ehrlenmeyer flask. For 
 each gram of sample weighed out add at least 0.9 grams 
 pure, tested oxalic acid (H 2 C 2 4 2H 2 0) weighing the acid 
 accurately and recording the same. Add about 30 cc. of 
 water and 30 cc. 5 normal sulfuric acid and drive off car- 
 bon dioxide by heating gently. 
 
 It is seldom necessary to filter after some practice, so 
 the solution is titrated hot for the excess of oxalic acid 
 with N/10 potassium permanganate solution. Calculate 
 amount of oxalic acid oxidized by the pyrolusite. The 
 reaction is 
 
 MnOa + H1C2O* • 2HaO + H2SO = MnSO* + 2COs + 4H a O 
 
 Each gram oxalic acid oxidized therefore corre- 
 sponds to .6902 g. Mn0 2 . 
 
 89 
 
As pyrolusite is added to some enamels only to give 
 color it is sometimes necessary to know its coloring power, 
 and this is dependent upon the total manganese. 
 
 Total Manganese. One-half gram sample is boiled 
 with strong hydrochloric acid until chlorine ceases to be 
 evolved. The solution is neutralized with calcium carbon- 
 ate and an excess of a strong filtered solution of bleaching 
 powder is added. The solution is boiled until deep red, 
 then alcohol is added until the red color disappears. The 
 whole of the manganese now exists as Mn0 2 and may be 
 reduced with oxalic acid and titrated for its oxidizing 
 power as before with N/10 permanganate of potassium. 
 Each gram of oxalic acid oxidized corresponds to .4361 g. 
 Mn. 
 
 The Analysis of Soda Ash and Pearl Ash 
 
 Generally it is only necessary to determine the total 
 alkali in a sample of either soda ash or pearl ash, and to 
 calculate from this the percentage of Na 2 or K 2 0. A 
 more complete analysis includes the determination of in- 
 soluble matter, iron, chloride, sulfate and moisture, as 
 well as the total alkali. 
 
 Insoluble Matter. 50 g. weighed on rough balance 
 (sensitive to 0.1 g.) and sufficient water added to dissolve 
 the ash, shaking until dissolved. After an hour's digestion 
 the solution is filtered through a weighed Gooch crucible 
 with a circle of filter paper covering the bottom. This is 
 dried at 105° and the increase in weight is insoluble 
 matter. 
 
 Iron. The iron in the above insoluble matter is dis- 
 solved by pouring hot dilute hydrochloric acid through 
 the precipitate in the Gooch crucible. The iron is precipi- 
 tated from this by ammonium hydroxide and filtered on a 
 white ribbon filter paper. The still moist precipitate is 
 dissolved in sulfuric acid, reduced by means of a Jones 
 Reductor and titrated with N/10 permanganate. 
 
 90 
 
Chloride. Three gram samples are dissolved in 
 water and nitric acid added until the solution is neutral 
 (test with litmus paper). It is then titrated with N/10 
 silver nitrate solution. 
 
 Sulfate. Five or ten grams are dissolved in hydro- 
 chloric acid and the sulfate precipitated from the almost 
 boiling solution by the addition of hot barium chloride 
 solution. 
 
 Total Alkali. Twenty-five grams are dissolved in 
 water in a 500 cc. volumetric flask and 50 cc. are titrated 
 with N. hydrochloric acid, using methyl orange as indi- 
 cator. 
 
 Hydroxide. To 50 cc. from above, precipitate all the 
 carbonate with barium chloride. Without filtering, add 
 phenolphthalein and titrate until colorless with normal 
 hydrochloric acid. 
 
 Moisture. Ten gram samples are dried at 120° for 
 two hours. 
 
 The Analysis of Saltpeter and Chili Saltpeter 
 
 Moisture. Ten gram samples are heated to constant 
 weight in an air-bath at 130°. 
 
 Insoluble Matter. Twenty grams are dissolved in 
 boiling water and filtered through a weighed Gooch cru- 
 cible with a circle of filter paper on the bottom. After 
 drying at 110° in air bath to constant weight, the increase 
 in weight is the insoluble matter. 
 
 Chlorine. The solution from above — this should be 
 about 500 cc. — is placed in a 1000 cc. volumetric flask and 
 25 cc. (representing 0.5 g. sample) is titrated with N/10 
 silver nitrate, using potassium chromate as indicator. The 
 result is calculated to sodium chloride. 
 
 Sulfate. Twenty cc. are heated to boiling and precip- 
 itated by adding hot barium chloride solution, a drop at 
 
 91 
 
a time and with constant stirring. After two hours diges- 
 tion ( or until precipitate settles quickly after agitating), 
 filter through a Gooch crucible with ignited asbestos filter, 
 ignite and weigh as barium sulfate. This is calculated to 
 calcium sulfate. 
 
 Calcium and Magnesium. From five hundred ccT"of 
 the above solution (equal to 10 grams sample) at boiling 
 temperature precipitate the calcium as oxalate by the ad- 
 dition of ammonium oxalate, being careful not to add 
 much excess, as magnesium is to be determined in the 
 same sample. Filter on a white ribbon filter paper, after 
 an hour's digestion on the steam bath, ignite wet paper in 
 platinum crucible, gradually increase to full blast and 
 heat to white heat to constant weight. Weight as calcium 
 oxide. 
 
 Determine the magnesium in filtrate from the calcium 
 by addition of a solution of microcosmic salt and after- 
 ward one-third the volume of concentrated ammonium 
 hydroxide, added drop by drop. The precipitate, ignited 
 separate from the filter paper, is heated at first gently 
 and at last with the full heat of a Bunsen burner, and 
 weighed as magnesium pyrophosphate (Mg 2 P 2 7 ). 
 
 Perchlorate. Ten grams of the sample of which the 
 chloride content has already been determined, is mixed 
 with an equal quantity of chemically pure sodium carbon- 
 ate, and is heated in a large, covered platinum crucible to 
 quiet fusion. Ten or fifteen minutes are required. The 
 product is then dissolved in nitric acid and the chloride 
 estimated as usual. 
 
 Nitrogen. This is determined by the Kjeldahl 
 method after reducing the nitrate to ammonia. Twenty 
 grams of the sample are ground coarsely and dissolved in 
 water in a liter flask, and solution is diluted to the mark. 
 Twenty-five cc. (equal to 0.5 g. sample) of this solution is 
 mixed in a 800 cc. Kjeldahl flask with 15 cc. concentrated 
 
 92 
 
sulfuric acid to which 2 grams salicylic acid have been 
 added, then add gradually 2 grams zinc dust and shake 
 flask to mix contents. Digest over low flame with neck of 
 flask slightly inclined until danger of frothing has passed. 
 Increase flame until the acid boils briskly and until white 
 fumes cease to come off. This usually takes about ten 
 minutes. 
 
 Add .7 gram mercuric oxide and continue boiling, 
 adding acid if necessary to keep solution from solidifying. 
 Solution should be clear in a short time. Complete oxida- 
 tion by adding a little powdered potassium permanganate 
 and allow the contents to cool. Add about 200 cc. am- 
 monia-free water and 25 cc. potassium sulfide solution 
 (40 g. commercial salt to the liter) and shake thoroughly. 
 Add several pieces of granulated zinc and then pour care- 
 fully down the side of the neck 100 cc. sodium hydroxide 
 solution (500 g. per liter), avoiding shaking and thereby 
 mixing the acid and alkali. After washing the neck with 
 ammonia-free water connect the flask immediately with a 
 previously set up block tin condenser, which has been 
 thoroughly washed and the tips of whose delivery are im- 
 mersed in 30 cc. standard acid solution (half normal), 
 colored with methyl orange contained in a 150 cc. Phillip's 
 flask. Mix contents of digestion flask by shaking thor- 
 oughly, then heat carefully, then slowly (taking about an 
 hour) distill over 200 cc. of the liquid. Titrate excess of 
 acid with standard half normal alkali solution, and from 
 this calculate percentage of nitrogen in sample. 
 
 Lung Nitrometer Method. Where frequent analyses 
 are made the Lung 1 Nitrometer method is better. A nitro- 
 meter modified especially for the use of the determination 
 of nitrate in saltpeter is here illustrated. The Nitrometer 
 "A" and the leveling tube "B" are filled with mercury. 
 From a twenty gram sample which has been dried at 110° 
 to constant weight as nearly as is possible 0.35 grams is 
 put into a weighing tube. This is then accurately weighed 
 
 iBerichte 1885, 18, 139L 
 
 93 
 
and the contents shaken into the entry tube "C." The 
 weighing tube is again weighed and the difference in 
 weight is the grams sample employed. This should be 
 close to 0.35 grams so that the gas evolved will be more 
 than 100 cc. and less than 130 cc. at ordinary tempera- 
 ture and pressure. 
 
 About .5 cc. water is then poured in and the solution 
 and crystals (after a minute's standing) are drawn into 
 the measuring tube by opening the three-way cock into 
 the entry tube "C" and lowering the leveling bulb cau- 
 tiously. The cup is washed, using less than 1 cc. of water, 
 and about 15 cc. of strong sulfuric acid is admitted 
 through the entry tube into the measuring tube. (More 
 than IV2 cc. H 2 renders the acid too dilute and the 
 mercury is attacked) . After the cock is closed the leveling 
 tube is placed in a clamp, the measuring tube is thor- 
 oughly shaken and the following reaction takes place : 
 
 MnOH-H*OOi : 2H20+H2SO*=MnSO*+2C0 2 +4H 2 
 
 The measuring tube is now placed in clamp on a level 
 with the levelling tube and solution is allowed to cool 
 for an hour. 
 
 The tube is then accurately leveled, allowing one 
 division of mercury for each six and one-half divisions of 
 acid, and the gas volume read off. The temperature and 
 barometric pressure are read and the gas corrected to 
 standard conditions. Each cc. NO gas corresponds to 
 .0037986 g. NaN0 3 or .003845 g. KN0 3 . 
 
 The Analysis of Cryolite 
 
 Cryolite is a mineral occurring in large quantities in 
 Greenland, and is the sodium salt of hydrofluo-aluminic 
 acid, Na 3 AlF 6 . It is used in enamels and is fused, finely 
 ground with the frit giving it a milky opaqueness which 
 enamellers call "body." It is a very expensive material 
 and is most always far from pure, either being deliber- 
 ately adulterated or merely naturally impure. 
 
 94 
 
Methods used by most chemists for its analysis are 
 at the best crude. The direct determination of the fluorine 
 is the only satisfactory means of properly grading it. The 
 method used for cryolite is exactly the same as that em- 
 ployed in the analysis of fluorine-bearing enamel as given 
 in the beginning of this article, except that one gram of 
 the cryolite is finely ground with about three grams (ac- 
 curately weighed) of pure silica, and this mixture is fused 
 with about eight grams of equal parts of sodium carbon- 
 ate and potassium carbonate. In determining the silica 
 in the cryolite, this silica which has been added must be 
 deducted from that found. 
 
 The alkalies are determined by the method of J. Law- 
 rence Smith 1 from a gram sample finely powdered. 
 
 Combination of Results. All soda is combined with 
 sufficient fluorine to form sodium fluoride (NaF 2 ). The 
 remainder of the fluorine is combined with aluminum as 
 aluminum fluoride (A1F 3 ) . The remainder of the alumi- 
 num is calculated as alumina (A1 2 3 ). 
 
 The Analysis of Fluorspar 
 
 Fluorspar is analyzed especially for the fluorine con- 
 tent by the same method as that given under Cryolite. It 
 is a material seldom adulterated and a mere fusion with 
 six times the weight of sodium carbonate, the taking to 
 dryness with hydrochloric acid as in ordinary silicate 
 analysis, the removal of iron and alumina as hydroxide 
 with ammonia, and the precipitation with ammonium 
 oxalate of the calcium and its final weighing as calcium 
 oxide, is sufficient in most cases. All the calcium may be 
 calculated as CaF 2 . The determination of the fluorine, 
 however, is of course the only exact method of accurately 
 grading this material. 
 
 The approximate method for determining the fluorine 
 is as follows: 
 
 Approximate Method for Fluorine. About one gram 
 of sample finely ground and accurately weighed is in- 
 
 1 Am. Jour. Science (2) 50, p. 269. Treadwell-Hall Anal. Chem., Vol. II, p. 394. 
 
 95 
 
timately mixed in agate mortar with about the same quan- 
 tity of pure silica. The whole is transferred to a 250 cc. 
 Ehrlenmeyer flask — rinsing the mortar with more silica 
 The flask is weighed and a weighed quantity of concen- 
 trated sulfuric acid is added. The record should now 
 show the weight of flask, silica and acid. The flask is 
 gently heated and the loss of weight is calculated as sili- 
 con fluoride. 
 
 Iron. For use in light colored enamels the iron con- 
 tent of the fluorspar is important. Five grams, finely 
 ground, are heated in a platinum dish with an excess of 
 sulfuric acid as long as hydrofluoric acid is given off. 
 
 After cooling it is diluted with 100 cc. of water, and 
 after reducing by drawing through a Jones Reductor the 
 solution is titrated with N/10 potassium permanganate 
 solution. 
 
 Accurate Method for Fluorine. The fluorine may be 
 accurately determined by the following method : 
 
 One gram sample (ground to pass through 200 mesh 
 sieve) is mixed with three grams silica and three grams 
 each sodium carbonate and potassium carbonate in a 
 platinum crucible. Heat gradually until it is in quiet 
 fusion. The thin liquid fusion soon changes to a thick 
 paste or only sinters somewhat. The reaction is complete 
 when there is no further evolution of carbon dioxide. 
 
 After fusion the melt is treated with water and after 
 cooling the insoluble residue is filtered off and thoroughly 
 washed. The solution contains all fluorine and consider- 
 able silica. Remove the silica by adding four grams solid 
 ammonium carbonate. Heat liquid at 40 °C. for some 
 time and let stand over night. Filter in morning and wash 
 with ammonium carbonate water. 
 
 Evaporate on water bath almost to dryness in plati- 
 num dish (keep covered, as liquid foams). Dilute with a 
 little water. Add a few drops of phenolphthalein. Add 
 
 96 
 
dilute HC1 until colorless. Heat on steam bath and color 
 will return. Cool and repeat operation until 1.5 cc. dou- 
 ble normal HC1 is sufficient to make colorless. Remove 
 last traces silica by treating the solution with a solution 
 of moist zinc oxide in ammonia water. Boil until am- 
 monia is completely expelled. Filter off silica and zinc 
 oxide and wash with water. 
 
 Precipitate fluorine as calcium fluoride and calcium 
 carbonate by adding an excess of calcium chloride. Filter, 
 using blue ribbon paper, and wash thoroughly with hot 
 water. Dry precipitate on funnel. Transfer as much as 
 possible to a platinum crucible. Burn filter and add ash. 
 Ignite contents of crucible. 
 
 After cooling the mass is covered with a slight excess 
 of dilute acetic acid (this changes the calcium oxide to 
 soluble acetate). Evaporate to dryness on steam bath. 
 Take up with water. Filter, wash and dry. Transfer 
 most of precipitate to weighed platinum crucible. Burn 
 filter paper. Add ash. Ignite and weigh as calcium 
 fluoride CaF 2 . To confirm the results add cautiously little 
 concentrated sulfuric acid. Evaporate off excess sulfuric 
 acid, ignite and weigh as calcium sulfate. 
 
 The Analysis of Oxides of Antimony 
 
 Arsenic. One gram of oxide of antimony is dis- 
 solved in 10 cc. of strong hydrochloric acid — at as low a 
 temperature as possible. The solution is then cooled and 
 packed in ice and the arsenic, which is almost invariably 
 present, is removed by passing through H 2 S for several 
 hours. The As 2 S 3 is filtered off in a weighed Gooch cru- 
 cible, washed first with CS, and alcohol then with concen- 
 trated hydrochloric acid and dried at 100°, and weighed 
 
 Antimony. The filtrate from above is put into 250 cc. 
 volumetric flask, rinsing the beaker well with concen- 
 trated hydrochloric acid and an equal part of water. All 
 
 97 
 
the H 2 S is removed by passing through a current of air. 
 Five grams of tartaric acid are added and the liquid 
 diluted to the mark. 
 
 Twenty-five cc. of the solution are measured out with 
 a pipette and are neutralized with dry sodium bi-car- 
 bonate — keeping covered to avoid loss — finally a pinch of 
 sodium bi-carbonate and a cubic centimeter of clear 
 starch solution is added and the mixture is titrated with 
 N/10 iodine solution. 
 
 1 cc. N/10 Iodine = 0.0060 grams Sb 
 
 The Analysis of Oxide of Cobalt 
 
 Arsenic. One gram finely pulverized sample is fused 
 at low heat with ten grams bisulfate of potassium for 
 three hours. The melt is extracted with water acidified 
 with sulfuric acid and the arsenic is precipitated from the 
 warm acid solution with H 2 S, collected in a weighed 
 Gooch crucible, washed with water containing H 2 S and 
 dried at 100° for one hour and weighed as As 2 S 3 . 
 
 Cobalt. The filtrate from above is boiled, and at the 
 same time air is drawn through to remove the H 2 S, and it 
 is then treated by Fisher's Potassium Nitrite method 1 to 
 separate the cobalt and the nickel. 
 
 The concentrated solution containing salts of both 
 metals is treated with pure potassium hydroxide to alka- 
 line reaction, made slightly acid with acetic acid, and to 
 this a concentrated solution of pure potassium nitrite 
 that has been made slightly acid with acetic acid is added. 
 After vigorous stirring, the mixture is allowed to stand 
 twenty-four hours in a warm place. Before filtering, a 
 little of the clear solution is pipetted off and treated with 
 more potassium nitrite to see if the precipitation of the 
 cobalt has been complete. If a precipitate is formed, the 
 whole solution is treated with more potassium nitrite and 
 again allowed to stand until complete precipitation is ef- 
 
 1 Treadwell, Vol. II, p. 130. 
 
 98 
 
fected. The precipitate is then filtered and washed with 
 a barely acid 5 per cent solution of potassium nitrite until 
 1 cc. of the filtrate, after being boiled with hydrochloric 
 acid and treated with caustic potash and bromine water, 
 no longer gives a black precipitate of nickelic hydroxide. 
 The cobalt precipitate is then transferred to a porcelain 
 dish, covered, and hydrochloric acid is gradually added 
 until there is no further evolution of nitric oxide, and 
 after filtering, the cobalt is precipitated by means of caus- 
 tic potash and bromine water. 
 
 The precipitate is filtered off, using blue ribbon filter 
 paper, dried, and ignited. After cooling it is treated with 
 water in order to remove the small amount of alkali which 
 is always present, after which the residue is ignited in a 
 stream of hydrogen and weighed as metal. After weigh- 
 ing, the metal is dissolved in hydrochloric acid, evapor- 
 ated to dryness, the dry mass moistened with hydrochlor- 
 ic acid, then treated with water, and the small residue of 
 silicic acid is filtered off. This residue is ignited and its 
 weight subtracted from that obtained after the ignition 
 in hydrogen. 
 
 Nickel. The filtrate containing the nickel is treated 
 with hydrochloric acid until the nitrite is completely de- 
 composed, and the nickel is precipitated with potassium 
 hydroxide and bromine water as brownish-black nickelic 
 hydroxide [Ni (OH) 3 .] 
 
 The precipitate — which seldom contains more than 
 ten milligrams of nickel — is washed with hot water, col- 
 lected on a filter and is dried, ignited separately from 
 the filter, and weighed as NiO, in which form it was prob- 
 ably present in the oxide. 
 
 Steel Plate 
 
 The steel best adapted for enameled ware is of very 
 low carbon value and extremely low in the other impuri- 
 ties, in fact, the nearer pure iron the better. Of the steel 
 
plate used by the Columbian Enameling and Stamping 
 Company, the best satisfaction was obtained from those 
 giving the following analysis : 
 
 Sulfur from .040% to .050%; phosphorous from 
 .030% to .090%; silica less than .01%; manganese from 
 .060% to .040%, and carbon less than 0.10%. The sheets 
 must be of an even gauge for seamless drawn work and of 
 a dark soft quality, which allows them to be drawn with- 
 out tearing. When the vessel is made without drawing and 
 sheets are used flat, this evenness of gauge is not so much 
 an object. The grain in all cases must be as open as pos- 
 sible. The sheet must be low in carbon and sulfur, as 
 these develop gases at temperatures of the muffle, which 
 would cause the enamel to peel off. 
 
 Samples of the steel plate are obtained from drillings 
 taken from eight or ten sheets stacked in a pile, and 
 drilled holes are run every two inches on the diagonal of 
 the plate. Drillings are sampled down to twenty-five 
 grams, which are kept in stoppered bottles. The method 
 of analysis is that commonly employed by steel-works 
 chemists, and can easily be found in print elsewhere, and 
 for that reason will not be given here. 
 
 100 
 
ATOMIC AND MOLECULAR WEIGHTS AND 
 
 FACTORS USED IN CERAMIC 
 
 CALCULATIONS! 
 
 Aluminum, 
 
 
 
 
 
 oxide, 
 hydrate, 
 
 A1 2 3 
 A1 2 3 
 
 *102.2 
 .3H 2 156.3 
 
 tR 2 3 
 R 2 3 
 
 — 102 
 
 — 156 
 
 
 1 Al — oxide 
 
 = 1.529 Al— hydrate 
 
 = 6.536 Al— sulfate 
 
 = 8.893 Ammonia Alum 
 
 = 2.534 China Clay 
 
 = 2.548 Cryolite 
 
 = 9.305 Potash Alum 
 
 = 5.154 Soda Feldspar 
 
 = 3.069 Am. Enamel Clay 
 
 
 Antimony, 
 
 
 
 
 
 oxide, 
 tetroxide, 
 pentoxide, 
 sulfide, 
 
 Sb 2 3 
 Sb 2 4 
 Sb 2 5 
 
 Sb 2 S 3 
 
 1 Sb — trioxide 
 
 288.4 
 304.4 
 320.4 
 336.6 
 
 = 1.167 Sb— trisulfide 
 
 R 2 3 
 R 2 3 
 R,0 3 
 R 2 3 
 
 — 288 
 
 — 304 
 
 — 320 
 
 — 337 
 
 = 1.111 Sb — pentoxide 
 
 Arsenic, trioxide, As 2 (X 
 
 197.9 R,0c 
 
 198 
 
 Barium, 
 
 
 
 
 carbonate, 
 
 BaC0 3 
 
 197.4 RO 
 
 — 197 
 
 sulfate, 
 
 BaS0 4 
 
 233.4 RO 
 
 — 233 
 
 
 1 Ba — carb. = 
 1 Ba — oxide = 
 1 Ba — sulfate = 
 
 0.777 Ba — oxide 
 1.287 Ba— carb. 
 0.657 Ba — oxide 
 
 
 Boric Acid, 
 
 B 2 3 .3H 2 
 
 124.0 B 2 3 
 
 — 124 
 
 (Fused), 
 
 B 2 3 
 
 70.0 B 2 3 
 
 — 70 
 
 Borax, 
 
 Na 2 B 4 O 7 .10H 2 O 382.2 R0.2B 2 3 
 
 — 382 
 
 (Fused), 
 
 Na 2 B 4 7 
 
 202.0 R0.2B 2 3 
 
 — 202 
 
 
 1 Boric Acid = 
 1 B 2 3 = 
 1 Borax = 
 1 Borax = 
 1 Borax = 
 1 Borax (Fused) = 
 1 Borax " = 
 1 Borax " = 
 
 0.548 B 2 3 
 1.771 Boric Acid 
 0.529 Borax Fused 
 0.162 Na — oxide 
 0.366 B 2 O a 
 0.307 Na — oxide 
 0.693 B 2 3 
 1.892 Borax 
 
 
 1 Compiled by Robert D. Landrum, mainly from Report of Committee of Equiva- 
 lent Weights, Transactions of American Ceramic Society, Vol. II, pages 196 to 278. 
 The reader is referred to this report for full explanation and method of using these 
 equivalent weights. 
 
 ♦Molecular Weight. fEquivalent Weight. 
 
 101 
 
Cadmium sulfide, 
 Calcium, 
 
 CdS 
 
 *144.5 fCdS —145 
 
 carbonate, 
 
 fluoride, 
 
 oxide, 
 
 phosphate, 
 
 sulphate, 
 
 China Clay, 
 
 CaC0 3 
 
 CaF 2 
 
 CaO 
 
 Ca 3 (P0 4 ) 2 
 
 CaS0 4 .2H 2 
 
 1 Ca — fluoride 
 
 100.1 
 
 78.1 
 
 56.1 
 
 310.3 
 
 172.2 
 
 RO 
 RO 
 RO 
 RO 
 RO 
 
 100 
 
 78 
 
 56 
 
 103 
 
 172 
 
 1 Ca — oxide 
 
 1 Ca — phosphate 
 
 = 0.718 Ca — oxide 
 
 = 0.487 F. 
 
 = 1.786 Ca — carbonate 
 
 = 1.321 Ca— hydrate 
 
 = 3.073 Ca— sulfate (Gypsum) 
 
 = 0.542 Ca— oxide 
 
 Al 2 3 2Si0 2 
 2H 
 
 Cerium 
 
 dioxide, 
 sesquioxide, 
 
 Chromate of Barium, BaCr0 4 
 
 Ce0 2 
 Ce 9 0< 
 
 Chromate of Lead, PbCr0 4 
 Chromium Oxide, Cr 2 3 
 
 258.0 
 
 172.3 
 328.3 
 
 253.5 
 
 323.1 
 
 152.0 
 
 R 2 3 7 
 2Si0 2 | — 258 
 
 Ce0 2 
 R 2 3 
 
 2RO 
 R 2 3 
 
 2RO 
 R 2 3 
 
 2RO 
 R 2 3 
 
 172 
 
 328 
 
 !- 
 
 !- 
 I- 
 
 506 
 646 
 152 
 
 1 Cr — oxide = 1.658 Am — bichromate 
 
 " s= 2.001 Am — chromate 
 
 " = 1.828 Cr— hydrate 
 
 " = 5.261 Cr — nitrate 
 
 " = 4.710 Cr— sulfate 
 
 " = 6.288 Cr — ammonia alum 
 
 Cobalt 
 
 carbonate, 
 
 chloride, 
 
 nitrate, 
 
 oxide, 
 
 oxide, 
 
 oxide (blk.), 
 
 sulphate, 
 
 CoC0 3 119.0 
 
 CoCl 2 .6H 2 238.0 
 Co(N0 3 ) 2 .6H 2 291.1 
 
 Co 2 3 " 166.0 
 
 CoO 75.0 
 
 Co 3 4 241.0 
 
 CoS0 4 .7H 2 281.2 
 
 RO 
 RO 
 RO 
 RO 
 RO 
 RO 
 RO 
 
 119 
 
 238 
 
 291 
 
 83 
 
 75 
 
 80 
 
 281 
 
 1 Co — carbonate 
 1 Co — chloride 
 1 Co — nitrate 
 1 Co — ic oxide 
 1 Co — ous oxide 
 
 0.630 
 0.315 
 0.258 
 0.904 
 1.587 
 3.173 
 1.173 
 3.749 
 
 Co — ous oxide 
 Co — ous oxide 
 Co — ous oxide 
 Co — ous oxide 
 Co — carbonate 
 Co — chloride 
 Co — ic oxide 
 Co — sulfate 
 
 ♦Molecular Weight. fBquivalent Weight. 
 
 102 
 
Copper 
 
 oxide (red) 
 oxide (black) 
 sulphate, 
 
 1 Cu- 
 
 1 Cu- 
 1 Cu- 
 1 Cu- 
 
 Cu 2 
 
 CuO 
 
 CuS0 4 .5H 2 
 
 -oxide (Blk.) = 0.899 Cu- 
 -oxide (Blk.) = 3.137 Cu- 
 -oxide (Red) = 1.112 Cu- 
 -sulfate = 0.319 Cu- 
 
 *143.1 |RO — 
 
 79.6 RO — 
 
 249.7 RO — 
 
 ^oxide (Red) 
 -sulfate 
 -oxide (Blk.) 
 -oxide (Blk.) 
 
 72 
 
 80 
 
 250 
 
 Cornwall Stone, 
 
 1R0.2.5 A1 2 3 
 
 20.SiO 2 1550.0 1RO] 
 
 2.5R 2 O 3 fl550 
 20.0 SiOJ 
 
 
 Cryolite 
 
 Na 7 AlF 6 
 
 210.0 
 
 3RO \ 
 
 Al 2 3 j— 
 
 420 
 
 Feldspar (Soda) 
 
 Na 2 . A1 2 3 
 6Si0 2 
 
 524.0 
 
 1RO 1 
 1R 2 3 U- 
 
 6Si0 2 J 
 
 524 
 
 Feldspar (Lime-Soda)CaO) A1 9 3 
 Na 2 0) 5Si0 2 
 
 520.0 
 
 1RO ] 
 1R 2 3 ^ — 
 5Si0 2 J 
 
 520 
 
 Feldspar (Potash) 
 
 K 2 0.A1 2 3 
 6Si0 2 
 
 557.0 
 
 1RO 1 
 1R 2 3 U_ 
 
 6Si0 2 J 
 
 557 
 
 Ferric Oxide 
 
 Fe 2 3 
 
 159.7 
 
 R2O3 — 
 
 160 
 
 Ferrous Sulphide 
 
 FeS 
 
 87.9 
 
 R 2 3 — 
 
 176 
 
 Ferric Oxide 
 
 
 
 
 
 (Magnetic) 
 
 Fe 3 4 
 
 231.5 
 
 R 2 3 — 
 
 155 
 
 Ferrous Sulphate 
 
 FeS0 4 .7H 2 
 
 278.0 
 
 R2O3 — 
 
 556 
 
 Lead Acetate Pb (C 2 H 3 2 ) 2 
 3H 2 
 
 carbonate (basic) 2(PbC0 3 ) 
 
 Pb(OH) 2 
 oxide (red) Pb 3 4 
 
 379.2 
 
 775.3 
 685.3 
 
 RO — 
 
 RO — 
 RO — 
 
 379 
 
 258 
 228 
 
 Litharge 
 
 PbO 
 
 223.1 
 
 RO — 
 
 222 
 
 Magnesium 
 
 carbonate 
 oxide 
 
 1 Mg- 
 1 Mg- 
 
 MgC0 3 
 MgO 
 
 -carbonate = 0.479 Mg- 
 -oxide = 2.089 Mg- 
 
 Iquivalent Weight. 
 103 
 
 84.3 
 40.3 
 
 —oxide 
 —carbonate 
 
 RO — 
 RO — 
 
 84 
 40 
 
 *Molecular Weight, tl 
 
 
Manganese 
 
 
 
 
 
 
 carbonate 
 
 
 MnC0 3 
 
 *114.9 
 
 fRO — 
 
 115 
 
 di-oxide 
 
 
 Mn0 2 
 
 86.9 
 
 RO — 
 
 87 
 
 Nickelic Oxide 
 
 
 NiO 
 
 74.7 
 
 RO — 
 
 75 
 
 Nickelous Oxide 
 
 Ni 2 3 
 
 165.4 
 
 RO — 
 
 83 
 
 Nickel Sulfate 
 
 
 NiS0 4 .7H 2 
 
 280.9 
 
 RO — 
 
 281 
 
 
 1 Ni 2 3 = 0.903 NiO 
 
 
 
 
 
 1 NiO 
 
 = 1.107 Ni 2 3 
 
 
 
 
 1 NiO 
 
 = 3.760 Ni- 
 
 -sulfate 
 
 
 
 Potash Alum 
 
 
 K 2 S0 4 A1 2 (S0 4 ; 
 
 U 
 
 RO 
 
 
 
 
 24H 2 
 
 949.1 
 
 A1 2 3 — 
 
 949 
 
 Potassium 
 
 
 
 
 
 
 antimonate 
 
 
 
 
 RO 
 
 
 
 
 KSb0 3 
 
 207.3 
 
 Sb 2 5 — 
 
 •414 
 
 bichromate 
 
 
 K 2 Cr 2 7 
 
 294.2 
 
 RO I— 
 2R 2 3 j— 
 
 294 
 
 carbonate 
 
 
 K 2 C0 3 .2H 2 
 
 174.2 
 
 RO — 
 
 ■174 
 
 carbonate 
 
 
 
 
 
 
 (calcined) 
 
 
 K 2 C0 3 
 
 138.2 
 
 RO — 
 
 ■138 
 
 nitrate 
 
 
 KN0 3 
 
 101.1 
 
 RO — 
 
 202 
 
 ] 
 
 L K — nitrate = 0.466 K — oxide 
 
 
 
 
 " 
 
 s= 0.683 K- 
 
 -carbonate 
 
 
 
 
 " 
 
 = 0.737 K— chloride 
 
 
 
 
 " 
 
 s= 0.863 K- 
 
 -sulfate 
 
 
 
 ] 
 
 L K— oxide = 1.467 K- 
 
 -carbonate 
 
 (anhydrous) 
 
 
 
 it 
 
 = 1.848 K — carbonate 
 
 (crystal) 
 
 
 
 " 
 
 = 1.582 K- 
 
 -chloride 
 
 
 
 
 it 
 
 = 1.192 K- 
 = 2.145 K- 
 
 -hydrate 
 -nitrate 
 
 
 
 
 it 
 
 = 1.849 K— sulfate 
 
 
 
 
 « 
 
 = 5.927 K- 
 
 -feldspar 
 
 
 
 
 (i 
 
 = 10.066 K- 
 
 -alum 
 
 
 
 Selenium 
 
 
 Se 
 
 79.2 
 
 Se0 2 — 
 
 ■ 79 
 
 Oxide 
 
 
 Se0 2 
 
 111.2 
 
 Se0 2 — 
 
 ■111 
 
 Silica 
 
 
 Si0 2 
 
 60.3 
 
 Si0 2 — 
 
 ■ 60 
 
 ♦Molecular Weight. ^Equivalent Weight. 
 
 104 
 
Sodium 
 
 antimonate NaSb0 3 
 
 carb. (cryst.) Na 2 C0 3 
 
 carb. (soda ash) Na 2 C0 3 
 
 chloride NaCl 
 
 hydrate NaOH 
 
 nitrate NaN0 3 
 
 silico fluoride Na 9 SiF R 
 
 * 191.2 fR0Sb 2 E 
 10H,O 286.2 
 
 382 
 286 
 
 106.0 RO —106 
 
 58.5 RO —117 
 
 40.0 RO — 80 
 
 85.0 RO —170 
 
 188.3 R0SiF 6 —189 
 
 1 Na — carbonate 
 (crystals) 
 
 1 Na — carbonate 
 (crystals) 
 
 1 Na carbonate 
 
 (anhydrous) 
 
 1 Na carbonate 
 
 (anhydrous) 
 1 Na — chloride 
 1 Na — hydrate 
 1 Na — nitrate 
 1 Na 2 
 
 0.371 Na — carbonate (anhydrous) 
 
 = 0.216 
 2.698 
 
 = 0.585 
 0.531 
 
 — 0.774 
 = 0.365 
 = 6.161 
 = 3.258 
 = 4.613 
 = 1.710 
 = 1.887 
 s= 1.290 
 = 2.742 
 = 5.193 
 
 — 2.290 
 = 8.500 
 
 Na 2 
 
 Na — carbonate (crystal) 
 
 Na 2 
 
 Na 2 
 
 Na 2 
 
 Na 2 
 
 Borax (crystal) 
 
 Borax (anhydrous) 
 
 Na — carbonate ( crystal ) 
 
 Na — carbonate (anhydrous) 
 
 Na — chloride 
 
 Na — hydrate 
 
 Na — nitrate 
 
 Na — sulfate (crystal) 
 
 Na — sulfate (anhydrous) 
 
 Na — feldspar 
 
 Strontium 
 
 
 
 
 
 
 
 carbonate 
 sulfate 
 
 SrC0 3 
 SrS0 4 
 
 
 
 147.6 
 183.7 
 
 RO — 
 RO — 
 
 ■148 
 
 •184 
 
 Tin 
 
 
 
 
 
 
 
 chloride 
 oxide 
 
 SnCl 2 
 Sn0 2 
 
 2H 2 
 
 226.0 
 151.0 
 
 RO — 
 RO — 
 
 -226 
 ■151 
 
 Titanium Oxide 
 
 Ti0 2 
 
 
 
 80.0 
 
 Ti0 2 - 
 
 ■ 80 
 
 Uranium 
 
 v 
 
 
 - 
 
 
 
 
 oxide 
 
 oxide (com'l) 
 
 uo 2 
 
 Na 2 
 6H 2 
 
 U 2 
 
 o 6 
 
 270.5 R 2 — 
 
 746.0 RO I 
 
 R 2 3 j- 
 
 ■543 
 -746 
 
 Zirconium Oxide 
 
 Zr0 2 
 
 juivalent Weight. 
 105 
 
 
 122.6 
 
 Zr0 2 — 
 
 ■123 
 
 * Molecular Weight. tE< 
 
 
CUBICAL COEFFICIENT OF EXPANSION 
 
 in Millimeters per degree Centigrade, as determined 
 by Winkelmann and Schott and Mayer and Havas 
 
 (See Sprechsaal 1911, No. 13) 
 
 A1F 3 = 4.4 X 10- 7 MgO = 0.1 X 10- 7 
 
 A1 2 3 = 5.0 X 10~ T MnO = 0.1 X 10~ 7 
 
 As 2 5 = 2.0 X 10- 7 Na 3 AlF 6 = 7.4 X 10" 7 
 
 BaO = 3.0 X 10- 7 NaF = 7.4 X 10~ 7 
 
 BeO = 4.7 X 10- 7 Na 2 =10.0 X 10~ 7 
 
 B 2 3 = 0.1 X 10- 7 NiO = 4.0 X 10~ 7 
 
 CaF 2 = 2.5 X 10- 7 PbO = 4.2 X 10~ 7 
 
 CaO = 5.0 X 10- 7 P 2 5 = 2.0 X 10~ 7 
 
 Ce0 2 = 4.2 X 10- 7 Sb 2 5 = 3.6 X lO" 7 
 
 CoO = 4.4 X 10- 7 Si0 2 = 0.8 X 10~ 7 
 
 Cr 2 3 = 5.1 X 10- 7 Sn0 2 = 2.0 X 10~ 7 
 
 CuO = 2.2 X 10- 7 Th0 2 = 6.3 X 10~ 7 
 
 Fe 2 3 = 4.0 X 10- 7 Ti0 2 = 4.1 X 10~ 7 
 
 K 2 = 8.5 X 10- 7 Zr0 2 = 2.1 X 10~ 7 
 
 Li0 2 = 2.0 X 10- 7 ZnO = 1.8 X 10~ 7 
 
 106 
 
 237 90 
 

 ,••• 
 
 
 
 
 W 
 
 <> 
 
 
 '•••' 
 
 
 &°+ 
 
 4? • 
 
 ,••• 
 
 • *0 
 
 • •o 
 
 
 iV«^ 
 
 L$^ 
 
 > v 
 
 
 ^ 
 
 «*°* 
 
 .••• 
 
 
 • •» 
 
 
 
 W 
 
 
 4 
 
 
 
 
 
 
 
 
 
 • ••/ 
 
 
 
 • #< 
 
 j\^^aT» •? 
 
r* 
 
 
 
 ^ 
 
 .••• 
 
 V* 
 
 H«3* 
 
 
 - V ^ % 
 
 
 :♦ y% ; -SK : /\ 
 
 * 
 
 
 
 
 
 HECKMAN 
 
 BINDERY INC. 
 
 ^ MAY 90 
 
 N. MANCHESTER, 
 ^^ INDIANA 46962