v* ^ 1:^ *ff- [ ' J ♦J^ x^ -?* m The person charging this material is re- sponsible for its return to the library from which it was withdrawn on or before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University. UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN mm FFB 9137$ JAN SEP 18 MAYS L161 — O-1096 JL*1 V&H a r> 1**L rV. £ f J* v$ ,y UNIVERSITY OF ILLINOIS BULLETIN Vol. 4. FEBRUARY 16, 1907 No. 12 [Entered at Urbana, Illinois, as second-class matter] STUDIES FROM THE SCHOOL OF CERAMICS NUMBER TWO. STUDIES IN GLAZES. PART I. FRITTED GLAZES. BY R. C. PURDY AND H. B. FOX. PUBLISHED FORTNIGHTLY HY THE UNIVERSITY [Reprinted from the Transactions of the American- Ceramic Society, Vol. IX. Paper read at St. Louis meeting, February, 1D07.] FRITTED GLAZES; A STUDY OF VARIATIONS OF THE OXYGEN RATIO AND THE SILICA-BORAdC ACID MOLECULAR RATIO. BY ]{<>ss C. Pubdy AND HARBY B. Fox, Champaign, Ills. Two systematic studies of raw lead glazes, published in the Transactions of the American Ceramic Society, have done much to indicate the limits of variation in their composition and heat treatment. Accounts of similar studies of fritted glazes could not be found by the writers, after a search in the literature of ceramics. This want of definite information is certainly not because fritted glazes are unimportant or little used, for, on the contrary, they form the basis of decoration for the most costly wares, and are to the white-ware manufac- turer the glaze par excellence in the making of non-crazing china. This dearth of information regarding fritted glazes is no doubt due largely to their complexity in composition and the consequent difficulties in making the necessary calculations. When there are but two variables as in the case of raw lead glazes, only the simplest calculations are required to formulate series in which one or both numbers vary in some predetermined ratio, but when three variable factors are to be considered, as in the case of fritted glazes, the planning of a like series presents serious difficulties. Mr. Ashley 1 , in his able discussion of the two papers on the composition of biscuit bodies that appeared in Vol- ume VII of the Transactions, demonstrated clearly the in- completeness that is apt to follow an attempt to determine the limits of variation of three factors when taken in pairs Trans. Am. Or. gOC., Vol. VII. p. 90. 3 FRITTED GLAZES. in six or more wholly independent series. As a further illustration, several series of fritted glaze studies were formulated, in each of which either the Al 2 O s , Si0 2 or jB 2 3 was varied in an arithmetical ratio. Several good iglazes were developed in each series, and consequently it i was thought a very wide range in the composition of fritted glazes at certain temperatures had been determined. Such, however, did not prove to be the case, for, when the glazes had been compared as to their oxygen ratio, Si0 2 — B 2 3 ratio, and A1 2 3 content, it was found that but a very nar- row range in composition had been used. As a result, the large amount of experimenting had practically been for naught. Confident that the difficulties in the case of fritted glazes could be overcome, the senior writer gave consider- able thought to methods by which the range in the varia- tions of the three factors, A1 2 3 , Si0 2 , and B 2 3 could be determined. The one given in the following report seemed to be the most feasible for the purpose. Many improve- ments in the details of the original plan were made by the junior writer, and it is felt that the method as here pre- sented is simple in its detail, and permits a very broad study of the limits of variation in fritted glaze composition. There are important details of the fritted glaze prob- lem which cannot be determined by any consideration of the oxygen ratio, the molecular ratio, nor in some cases, even the ultimate chemical formulae of the glazes. As an illustration, oxygen ratio does not seem to be a factor in determining the kind and quantity of the several ingre- dients that should be incorporated into the fritt, and what should be added 'raw', in order to obtain a given fusibility, coefficient of expansion and contraction, fluidity, solvent effect on body, production of color tints, etc., etc., conse- quently such questions are not considered in the experi- ments here reported, although as stated before, their im- portance is recognized. Owing to the scarcity of reliable scientific data on the fritted glazes actually used in the different ceramic indus- FRITTED GLAZES. O tries, but very little idea of the limits of variation in the composition and heat treatment of such glazes could be obtained. The following are a few of the formulae studied: White Ware Glaze, E. Mayer, Trans. Am. Cer. Soc, Vol. I, p. 57. 0.064 K 2 ^ ro ai ft . n 0.192 Na 2 C>! Ain J J ' 81 Sl ° 2 SSSSJ i«» c — 4 Earthen-ware Glaze, Dr. H. Hecht, Thonindustrie-Zeitung, 1897, quoted by Ashley, Trans. Am. Cer. Soc, Vol. VII, p. 92. 0.10 KoO 1 f2.1 SiOo 0.30 CaO } 0.2 AI2O3 4 0.60 PbO j 10.4 B 2 ;! Cone 010 White Ware Glaze, K. Langenbeck, p. 122, Chemistry of Pottery. 1 1. 25 KNaCn f3.0 Si0 2 0.50 CaO } 0.3 A1 2 3 < 0.25 PbO J 10.5 BsOj Cone 04 WaU Tile Glaze, R. C. Puxdy, private notes. 0$?N?ol fl.835Si0 2 S CaO W2 ^ Wall Tile Glaze, R. C Purdy, Trans. Am. Cer. Soc, Vol. VII, p. 81. on CaO °- 19 A1 * ^ o'lgpbo l0.57B 2 O., Conel GENERAL PLAN OF INVESTIGATION. A study of these successful commercial fritted glazes revealed the fact that there were four factors to be con- sidered in the make-up of the chemical formulae. 1st. The character of R. O. i. e. the kind and equiva- lent amounts of each present. 2nd. The oxygen ratio, i. e. the ratio of the total oxy- gen in the acids to the total oxygen in the bases. 3rd. The molecular ratio of the silica to boracic acid. 4th. The equivalent content of alumina. It is quite obvious that since there are oxides of seven or eight different basic elements used for different purposes in fritted glazes, there are a great many possible BO com- binations that could be and should be tried if the studv of 6 FRITTED GLAZES. the entire field of fritted glazes is to be attempted. Since for each 110 there is a large number of possible variations in the other factors, it is evident that it would be imprac- tical to include all possible combinations in one investiga- tion. It was decided, therefore, to limit this study to one arbitrarily chosen RO. The RO. The following was taken because it repre- sents the average character of the RO used in white-ware and wall-tile glazes. 0.126 Na 2 Q 0.124 K z O 0.500 CaO 0.250 PbO The Oxygen Ratio. In order to not only cover the range of variation in the oxygen ratio of the commercial fritted glazes studied in the preliminary survey of the sub- ject, but also to determine if possible the practical limits of such variation, it was thought best to go beyond Avhat was taken to be the extreme minimum and maximum. Since the most workable oxygen ratio in a raw lead glaze has been found to be t'^2, and since one of the fundamental objects in the use of fritted glazes is to secure increased acidity and consequent reduction of the coefficient of ex- pansion and contraction, this ratio (1 : 2 ) was adopted as the minimum. At the time the study here reported was made, it was thought that the oxygen ratio of 1. : 4 was the maximum in use, and therefore a ratio that would repre- sent a step beyond this limit was chosen as the maximum. During the progress of the investigation it was found that the maximum limit possible under all conditions had not been chosen, for, not only, were there good glazes de- veloped with this oxygen ratio, but it has since been found that Seger had used leadless barium fritted glazes that ranged as high as 1 : (> with reported success. FRITTED OI.AZK9. The following are the oxygen ratios adopted Oxygen in Oxygen In Total Bases Total Acids 2.00 2 60 3.00 3.50 3 75 4.00 4.50 The Silica-Boracic Acid Molecular Ratio. It was in- ferred from a study of commercial fritted glazes that the silica-boracic acid molecular ratios varied from 1 : to 1 : 0.20, and therefore 1 : 0.25 was chosen as the maximum. Here also it has since been learned that the possible maxi mum ratio was not adopted, for Edwards and Wilson 1 report the successful use of a much higher ratio. The silica-boracic acid ratios adopted in this study were : 0.25 0.20 1 0.11 1 0.13 1 09 1 05 0.01 and The Range in Alumina Content. It is quite obvious to those who have made a study of glazes in general, that the permissible maximum equivalent of A1 2 : . is dependent upon the temperature at which the glaze is designed to mature, and since cone 10 had been arbitrarily chosen as the maximum temperature at which to burn the glazes of this study, it was decided that 0.45 equivalent would be a larger equivalent than would be used commercially in the vast majority of instances. It is common experience that a small equivalent of Al 2 O a is quite accessary to the devel- opment of an insoluble or stable glaze, and thai as a rule white ware glazes contain from 0.25 to 0.40 equivalents. It 'Trans. Eng. C. S. 1904-5, p. 24. FRITTED GLAZES. was thought, therefore, that 0.1 equivalents could, with justice to the study, be considered as the minimum amount feasible. The equivalent molecular variations of A1 2 3 chosen were 0.1; 0.15; 0.20; 0.25; 0.30; 0.35; 0.40 and 0.45. Shown in tabular form these three variable factors can be represented as follows : Division into Groups by Oxygen Ratio Division of Groups into Series by the Molecular Ratio of SIO, : B 2 3 Division of Series into Members Differentiated by their Al a 3 content Group No. O. R. Series No. Ratio Member Designation A1,0 3 Equiv. I n ni IV V VI VII 2.00 2.50 3.00 3.50 3.75 4-00 4.50 1 2 3 4 5 6 7 8 1 1 1 1 1 1 1 1 0.25 0.20 0.17 0.13 0.09 0.05 0.01 0.00 : c d e f g h 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 CALCULATION OF THE GLAZES. The calculation of the proportions in which the var- ious extremes in these series are to be blended to produce any given glaze of the 448 provided for in the above scheme is somewhat perplexing, and as before stated, was made the subject of a good deal of thought. The following method of calculation was finally found to be the most satisfactory. Considerable more space is given therefore to the detailed explanation of the method than would perhaps be justifi- able in ordinary cases. First Method : Formula of Glazes. The oxygen ratio of a fritted glaze can be expressed by the formula 2y+3z < A > °- K - = 3iTr in which O. R. stands for oxygen ratio; y represents the molecular equivalent of Si0 2 present; z the molecular equivalent of boracic acid; x the molecular equivalent of alumina; and the numeral 1, the oxygen in the RO when the total equivalent of RO is reduced to unity. In the FRITTED GLAZES. 9 statement of conditions for any one glaze, the oxygen ratio and alumina equivalent are given, i. e. known. Each for- mula thus contains only y and z as unknowns. Since the ratio of y to z is also given in the siliea-boracic acid ratio, we have the equation ( B ) y : z : : 1 : b or z=by. With these two simultaneous equations y and z are readily determined. Since the A1 2 3 equivalent varies regularly in each series, it is quite obvious that the Si0 2 and B 2 3 content of the several glazes in each series differs by a constant factor that can be obtained by subtracting the Si0 2 and B 2 3 in the first member (a) from the Si0 2 , and B 2 O x equivalent in the second (b), so that only the first two members of the series need be calculated by equations (A) and (B). Second Method. It was developed in the actual carry- ing out of the calculations by the first method that only the first two terms of the series in the end groups need be calculated by the above equations. After obtaining the Si0 2 and B 2 3 for the members of each series in Groups I and VII as given in the first method, the corresponding serial members of the intermediate groups, it was found, could be obtained by a blending calculation on the basis of the difference in the oxygen ratio and the rule of extreme and mean differences. The blending factors for the groups are Parts of Group I Parts of Group VII Group I 1.0 0.0 n 0.8 0.2 m 0.6 0.4 IV 0.4 0.6 v 0.3 0.7 VI 0.2 0.8 u vn 0.0 1.0 Equiv. Si0 2 for S 1, Q 1 "a" (.9464) times 0.8 = 0.75632 " '« " SI, GVII" a" (2 1 "3) times 0.2 = 0.42547 " Sl,GII"a" then 0.75632 + 0.42546 or 1.1818 10 FRITTED GLAZES- By this method the Si0 2 and B 2 3 content of the glazes in all the groups can be very readily obtained by the use of simple factors as above shown. Blending of the Glazes. There are many ways in which the glazes indicated in the above table could be blended, but it was decided that this could be done most readily and with the greatest ac- curacy when the number of "original" or weighed glaze batches was the smallest, so that the fritts might be melted in large quantities at one time, and thus errors in weighing fractional quantities made as small as possible. Each individual glaze can then be made by weighing portions of these few made-up glazes as explained later. The minimum number of "weighed" glazes that could be used in this blending scheme was found to be eight, as follows : 1. Group I. Series 1 member a 2. Group I. *' 1 h 3. Group I. it 8 " a 4. Group I. " 8 h 5. Group VII. •• 1 a 6. Group VII. " 1 h 7. Group VH. " 8 a 8. Group VII. c< 8 h Of these eight glazes the first four are from Group I and the last four are from Group VII. In the first group, therefore, the maximum amount of each of the first set of four glazes will be used and in Group VII the maximum amount of each of the last set of four glazes. The amounts of the glazes in each set that are used in the intervening groups are in the same proportion as the oxygen ratio of those several groups. By again noting the extreme and consecutive variations in oxygen ratios in the tabulated statement of the groups and series, it will be seen that these proportional amounts for each group as noted above are as follows: FRITTED GLAZES. Group I 1.0 0.0 Group II 0.8 0.2 Group III 0.6 0.4 Group IV 0.4 0.6 Group V 0.3 0.7 Group VI 0.2 0.8 Group VII 0.0 1.0 Total, 3.3 37 11 That is, in the first group, 10-33 of the total amount of the first four weighed glazes will be used ; in the second group 8-33 of the first four and 2-37 of the second four, etc. Having thus calculated the distribution of the eight weighed glazes in the several groups, the next step is ob- viously the determination of the proportional amount of each of the "weighed" glazes in each series. Since the "weighed" glazes represent series 1 and 8 of each group, by making blending calculations on the basis of the silica content, the intermediate series of each group are made up of proportional parts of the extreme :*Tiefl as follows : Serie. Weighed Glazes of Weighed Glazes of 1st Series 8th Series 1 1.0000 0.0000 2 0.8461 0.1539 3 0.7451 0.2549 4 5984 0.4016 '» 0.4363 0.5637 6 2558 0.7442 T 00542 0.9458 8 0000 1.0000 Total 3-9359 4.0(541 To illustrate how this proportional distribution of the weighed glazes is made in the separate groups take Group II, Series 4. It was determined that for Group II there should be used 8-33rds of the total amount of the weighed glazes belonging to Group I and 2-37ths of the 12 FRITTED GLAZES. weighed glazes belonging to Group VII. Group II, therefore, there would be used : /0J5984 _8\ f weighed glaze j and 2 . V3.9359 33/ (Hi x ^)° ,wei8hedgIaK!6a,,d6 - 1 of weighed glaze 3 and 4. 1 of weighed glaze 7 and 8. In Series 4, of 3.9359 0.4016 8 4.0641 X 33 0.4016 _2_ 4.0641 X 37 Similar calculations were made for each series in each of the groups. The sum or mixture of the proportional parts of the total amount of "weighed" glazes 1, 3, 5, and 7, as shown above, would constitute the first member or "a" of Series 4, Group II and the mixture of the proportional parts of the total amount of "weighed" glazes 2, 4, 6, and 8, as above shown, would constitute the last member or "h" of Serie* 4, Group II. Having thus the proportional part of the total amount of the "weighed" glazes in the first and last members of each series in each of the groups, the further distribution of the "weighed" glazes in the intermediate members of each series must be made on the blending proportions as obtained on the basis of the difference between the alumina content of each member. These proportional factors are found to be as follows : Proportion of "a" Proportion of " b" Members of Series in each Member in each Member a 1.0000 0.0000 b 0.8571 0.1429 c 0.7142 0.2858 d 0.5714 0.4286 e 0.4286 0.5714 d 0.2858 0.7142 f 0.1429 0.8571 g 0.0000 1.0000 Total 4.0000 4.0000 FRITTED GI>AZES. 13 Series 4 Group II then may be made up by blending the eight "weighed" glazes in the proportional amount of the total of each of the "weighed" glazes as in the table on page 106-107. To demonstrate that the factors there developed give true results, the chemical formula of glaze d, Series 4, Group II, is calculated as follows : 0.00526 Eqv. glaze No. 1 contains 0.00394 0.00341 0.00266 0.00117 0.00088 0.00076 0.00057 No. 2 No. 3 No. 4 No. 6 No. 6 No. 7 No. 8 Total 0.0052 0.0039 0.0034 0.0026 0.0011 0.0008 0.0007 0.0005 Al„O a 0.0181 0.0005 0.0018 0.0003 0.0012 0.0001 0.0003 0.0001 0.0002 0.0046 SiO, 0.0049 0.0067 0.0044 0.0060 0.0026 0.0034 0.0022 0.0030 B.O. 0.0012 0.0016 0.0006 0.0008 0.0331 0.0042 Multiplying these totals through by an amount which will bring the RO to unity, or 55.24, we have the following result : (1.8284 Si0 2 . 9998 RO, . 2486 A1 S 8 ■? (0.2820 B 2 0, of which the oxygen ratio is 1:2.49 + Since the sum of the elements brought in by the re- spective amounts of the various weighed glazes employed makes a glaze whose formula satisfies all the conditions for glaze 'd', Series 4, Group II, the above method of calcula- tion must be correct in principle. In the same manner, the equivalent amounts of the eight "weighed" glazes required for a mixture or blend, having the chemical composition of each of the 448 glazes required for the whole investigation can be calculated and tabulated. These equivalents or factors are parts by molecules but not by weight. In the calculation of the required parts by weight of the eight "weighed" glazes, the total amount of each glaze required, and the difference in the combining 14 PKITTKD GLAZES. Calculation for Compounding 0.5984 3.9359 4016 4.0641 5984 8 9369 4016 4 0641 X X X- x- 8 33 8 33 2 37 2 37 "a" X "h" X 0.8571 4.000 0.1429 4.000 "a" X "h" X 0.7142 4.0000 0.2858 4.000 "a" X "h»" X 0.6714 4.000 0.4286 4.000 "a" X "h"X 0.4286 4.000 0.5714 4 000 "a" X "h"X 0.2858 4.000 0.7142 4.000 "a" X- h" X 0.1429 4 . 000 0.8571 4.0D0 5984 3.9359 0.4016 4.0641 0.4016 4.0641 0.5984 3.9369 X X- X- 8 33 2 37 8 33 2 37 Proportion of Weighed Glaze Number 1 0368 0079 00657 00526 00394 0.00263 001315 Proportion of Weighed Glaze Number 2 0.001375 0.00263 0.00394 0.00526 0.00657 0.007885 0368 Proportion of Weighed Glaze Number 3 0.0239 00512 0.004267 0.003414 00256 00181 0.000854 KRJTTKD GLAZES. 16 Scries 4, Group II. Proportion of Weighed Glaze Number 4 Proportion of Weighed Glaze Number 5 Proportion of Weighed Glaze Number 8 Proportion of Weighed Glaze Number 7 Proportion of Weighed Glaze Number 8 00822 0.00536 0.00176 001146 0.000864 000294 0.000191 0.00147 0.00096 0.00181 000587 000382 0.001174 0.00076 0.00256 0.00088 0.000573 00088 000573 003414 0.001174 0.00076 000587 0.000382 0.004267 0.00147 0.00(96 0.00612 0.000294 0.00176 0.000191 i 0.001146 0289 00822 1 i 0.00685 1 18 FRITTED GLAZES. or batch weights of each of the "weighed" glazes must be considered. This latter consideration has not until lately been noted in the blending of series of glazes or bodies and, so far as the writers can learn, it was first suggested and made use of by the senior writer, in class exercises in ce- ramics at the Ohio State University in 1903, and was first mentioned in published papers by E. Ogden, who was a student at Ohio at that time. 1 Mr. Ogden has amply set forth the necessity of noting the difference in the batch weights of the extremes, so that further discussion of this point at this time is superfluous. The importance of tak- ing cognizance of the differences in combining weights of the several glazes to be blended is illustrated in the fol- lowing : Calculation of required total amount of "weighed" glazes for entire experiment. By trial it was found that about 60 grams of fritted glaze was necessary to make a coating one-sixteenth to one-eighth inch thick on five 3 inch by 1 inch by 1^2 inch wall tiles. It is obvious that the maximum quantity of any one of the "weighed" glazes in a given blend would be required in the case where the "weighed" glaze is used alone, or unblended with any of the others. Considering this as a safe criterion by which the total amount of each the "weighed" glazes required in the entire series of blends can be estimated, it is quite ob- vious that since, in the case of "weighed" glaze No. 1, for illustration, as shown in the development of the blending factors, 10-33 of the total amount is required for the first 1 group; and of the quantity required for the first 3.9359 group is used in the first series ; and 14 of that used in the first series is required in the first member or 'a', it follows 10 1 1 10 that — X X — or of the total amount of 33 3.9359 4 519.54 'Trans. A. C. S., Vol. VII, p. 378. FRITTKJ) GLAZES. 17 "weighed" glaze No. 1 used in the entire system of blends 10 must be equal to 60 grams. Therefore 60 -4- or 519.54 3117.24 grams is the total amount of "weighed" glaze No. 1 required. Having calculated the batch weight of the eight '•weighed" glazes, as shown later, it was found that the combining weight of ''weighed" glaze No. 1 was 213.78. Since 3117.24 was so nearly 15 times the combining weight of "weighed" glaze No. 1, 15 was adopted as a fac- tor by which the combining weights of each of the eight "weighed glazes should be multiplied to ascertain the total amount of each required for blending the entire number of glazes in this experiment, as shown iu the following table : Weighed Glazes Combining Weight Factor Amount Required 1 213.78 15 3207 2 296.78 15 4452 3 206.52 15 3097 4 305.34 15 4580 5 293.00 15 4395 6 462.32 15 6935 7 303.97 15 4560 8 481.47 15 7222 Means of Minimizing the Number of Weighings in Above Blending Scheme. By simple calculations it was found that if the above scheme of blending was carried out in detail as given, 2976 separate weighings would be re- quired. By first weighing up the "weighed" glazes pro- portionally into groups, then thoroughly mixing the blended glazes by passing them through a 60 mesh sieve six or seven times; taking these blended glazes, now 16 in num- ber, and blending them proportionally into series, and then finally into the separate members, in other words, by mak- ing the blends in three stages, there would be required only 656 or less than Vi of the number of weighings that would be required if the amount of each of the "weighed" glazes •/ FRITTED GLAZES. requisite for the proper blending of each member, was weighed out direct. This was accordingly done. THE FRITTS. Opinion and custom differs widely as to what parts of the glaze should be incorporated in the fritted portion. As a rule, there is no cognizance taken of the solubility of the resultant fritt, nor any attempt to harmonize its chemical composition with that of the glaze. Since there is neither law nor general custom in this matter, the following rules were arbitrarily chosen : 1st. The fritt should constitute at least 50% by weight of the glaze. 2nd. Its oxygen ratio should be the same as that of the whole glaze. 3d. The fritt should contain (1) all of the alkaline salts including the feldspar; (2) all of the boracic acid and borax ; (3) all but 0.05 equivalent of the required clay; (4) all but 0.10 equivalent of the total calcium oxide; (5) all of the free aluminum oxide; (6) only sufficient silica to maintain the required oxygen ratio. This left to be added raw to each of the "weighed" glazes the following : . 25 Eqv. white lead 0.10 Eqv. whiting 0.05 Eqv. china clay X Eqv. flint The fritts were made in a crucible fritt furnace 1 fired by gas. The fritts when fused dropped from the crucible into cold water for granulation. It was found that all but two of the fritts were reasonably insoluble in water, but the comparatively ready solubility of the fritts belonging to weighed glazes No. 3 and 7 developed a belief that the fritts should not be run into water but rather out 'Manufactured and donated to the Ceramic Department, Univer- sity of Illinois, by W. D. Gates, American Terra Cotta and Ceramio Company, Chicago, Illinois. FRITTED GI.AZKS. 19 onto a told slab, refritted and then ground dry. Mr. J. F. Krehbiel suggested this plan, and has carried it out with good effect in the crystaline glaze experiments. With the exception of the solubility of the two fritts cited, all eight fritts seemed to be normal. Their fluidity was sufficient to permit free flowage through the orifice of the crucible, and in no case was excessive heat or time re- quired to affect the complete melting of the fritt. Materials Used. The following materials were used in this series of experiments : O. P. Sodium Carbonate NajCO.-? C. P. Potassium Nitrate KNO3 Whiting CaCOs White Lead Pb(OH) 2 ZPbCOs Flint SiOs Borax NajB^O? 10H 2 O Boracic Acid (Flaky) B»0 8 3H 2 Calcined Aluminum Oxide Al 2O3 Soda Feldspar. The analysis was as foUows : Per cents. SiOs 69.36 AljOs 17.00 Fes0 3 0.63 CaO 0.62 MgO 0.88 K s O 6.31 NaiO 4.79 The molecular formula of the above is : of which the combining weight is 669 . 6. Potash Feldspar, from Brandywine Summit, Pennsylvania. The analysis was as follows : Per cents. SiOa. 68.60 AI2O3 18.40 FeaOs 0.54 CaO 0.30 MgO 0.14 KzO 10.52 NaaO 2.00 Moist 0.17 l\ A I -. 20 FRITTED GLAZES. The molecular formula of which is : 0.7314 K 2 ) n ™ iln , 0.2109 Na*0 1 11<9 M2 °* \ - .- «,.<. 0.0349 CaO f n n . 9 ~ n f 747 blL>? 0.0228 MgOJ - 042re2 ° 3j of which the combining weight is 663.79. Georgia Kaolin. The analysis was as follows : Per cents. Si0 2 44.02 AI2O3 39.51 Fe 2 3 1.09 CaO 0.36 MgO 0-12 KNaO 0.23 H 2 14.60 The molecular formula of the above is : KNaO 0.0075^| I.OOAI2O3 ^ 1.89 SiO a OaO 0.0165 } V MgO 0.0077 J 0.0175 Fe 2 3 J 2.09 H 2 of which the combining weight is 257.78. PREPARATION OF THE GLAZES. The fritts were wet-ground to pass freely a 100 mesh sieve, dried in newly-made plaster evaporating molds, and finally dried thoroughly by subjection to artificial heat. As no water was poured off and thrown away, the only portion of the soluble materials in the fritts that was lost, was the trifle that entered the pores of the plaster evaporating molds. The uncertainty as to the absolute constitution of the fritts when finally ready for use was the only known irregularity in the whole experiment, and the writers believe that this is not very serious. The eight "weighed" glazes were then weighed, wet- ground for a half hour, passed through a 100 mesh sieve, and dried in plaster evaporating molds. o to tO JS o tO 35 X cc lO IN © tO ■ r^ cc tt 05 : cr -r CC © W C: O © C ec © ec CC cs c .a lO tc K •- t^ r~ rt „ to ec IC t» l~ S3 s ■<* ec T» cc r~ O N ao © X iC oc to eg CC CC t^: CC C) £ © ©" o a CM CN © .- CM tO CN cc -~ to CN t^ e t^ © oo X i- X X s _' CC Ci CN ■^ -r CO pa CO I- a c 'J' to co o •* lO •* cc CN X X K c: e co ■* a" © ©' © cc cc CN i-i © © •E OB CC r- to -S" •^ to X CN « T*< ~ aa n _' « X cc cc OS o «b CB pa cc -r CN CC to co © o BE tO ■ - CC 't to (M C CM •^ ~*f to tO X CM © © = © © a © Q X r— jC to c c to tO t» « ,_i 1> co 00 X 03 co tO r- CN ■* o CI t- to tO tC CC © cr: tC c CN H co CN © c © CM G C3 IO OS o - X cn 1 — — s CC -? © ?> cc CC © ~ © © o © tO CB X © A tO 1 — X X ' t~ CC -r I- X CO cc ec tO CM CO co T*< X CN *""* 5S © -h — o © t- tO • r CC X CN iO • — 3 •3 S 00 g, -o - o o © u B © s s eo o s "3 > o c B "3 6 ± a i p 43 "as • •-1 3 '3 1 i i 7 ft CO 3 X X 1 C a, CD a o .o o GO C 5 c8 J < 6 3 < p c - o 43 "C Z, H, E aw um !".[ qo >«a £ s — W ~~* 21 dj Xi 1 a O X a PQ -o a CO i o _• i-J IO CM OOO O c a o 3 S o O 00 <* © oo CM id u M PQ lO CM CM CO co CM 00 © © © CM © iO © © > a" W lO CO d © CO © ■O © IO CM © © I— 1 © t^ o o IO CM as CM IO CM 00 © © © CM © IO © © © f— 1 > w IO CO © 3 CO o "O 1 IO © CM © 1 © © to IO •** o m oo CD OS d o PQ © 1—1 CO co CM 00 i-H © © © CM © IO © © © > a- W iO © © © CO o IO © d CM © © d IO o © CN CM CM co •o d o << pa co © CM 00 © © © CM © lO © © © > w iO © ©' 3 CO © IO © © lO CM © © »— 1 © ■** »o d iO CO CM u « PQ !M © CO 2 CM to CN © © CM 1-1 © IO © © © V W IO © © iO © © lO © © IO CM © © © CO o d CO d o PQ © ( W lO © © as © © lO © © iO CM © © © CI lO d as © CO CM o j» u 09 CO © 00 CM CO CM © I © © IO CM tH — i | © © © r-t a 1 W IO © © © © IO © © IO CM d © © ~ d «o as © co CM d X 03 © © CM © CN 1 © CO © -# CM 01 r-l © lO © © © > W iO © d 1 © o © to CM © © © CO q C c O a .5 : 3 : co • 60 "" ££ 41 CS O eg a O "3 CO CD i-2 CD 43 a -*- X D 22 KRITTKD GLAZKS. 23 From the foregoing tables it is readily seen that the difference in composition and batch formulae lies wholly in the constitution of the fritt, except that each group con- tains progressively more and more free flint. For each group, however, there is a definite equivalent of free silica. The equal equivalent of fritt used in each case insures uni- formity in the make-up of the several glazes, so that the difference in behavior of the glazes, one from another, can be said to lie wholly in their chemical constitution. The actual formulae of the glazes, giving the A1 2 :5 , Si0 2 and B 2 3 content of each, has been prepared in tabu- lar form, but in order to facilitate close comparison be- tween the composition of the glazes and their results on firing, the tables, marked Group I, Group II, etc., appear in connection with the results, instead of at this place. BODY USED. The body used in these experiments was furnished by the U. S. Encaustic Tile Co., of Indianapolis, Ind., through the courtesy of Mr. E. M. Ogle. It was delivered to the laboratories of the Ceramic Department of the University of Illinois in the shape of normally burned biscuit wall tile of apparently uniform density. The composition of the body as given by Mr. Ogle is as follows: Ball Clay 40 China Clay 40 Cornwall Stone ..20 Flint 22.5 122.6 PREPARATION OF THE TRIAL PIECES. The question of how thick a layer of the glazes should be placed on the tile was considered seriously, for it was doubted if the difference between the behavior of the sev- eral glazes would be sufficiently exaggerated or emphasized if applied as thin as is the practice of the china and white ware potters, or even the wall tile manufacturers. Between 24 FRITTED GLAZES. one-sixteenth and one-eighth of an inch was finally adopted as the thickness. The glazes that stand as "good"' at this thickness will surely stand well when applied thinner, and those glazes which would have a tendency to craze or shiver would have that tendency increased by increased thickness, and thereby display their peculiarities almost at once after drawing from the kiln. Five tile, marked in pencil with the proper group, ser- ies and member symbols, were thoroughly saturated with distilled water, placed side by side and the glaze paste ap- plied over all five tile at once, with a spatula. After dry- ing, the tiles were separated, fettled, and re-marked with a cobalt stain. PLACING OF THE TRIAL PIECES. The tiles were placed in the tile setters of such capacity that each held one whole series. The members of each ser- ies were placed in a setter in regular order, so that in case any of the tile should be stuck to the bottom of the setter, as a consequence of the running off of a portion of the ex- cessively thick glaze, as happened in a few cases, each mem- ber of the series could be readily identified by its position in the setter. The identification of one specimen in each setter was, therefore, sufficient for the identification of the remaining members of the series. For nearly all burns above cone 010, the tile were placed on small wads with sufficient space between the tile and the setter bottom to permit of considerable running of the glaze without seriously cementing the tile to the setter. BURNING OF THE GLAZES. The glazes were burned in a side down-draft kiln de- signed by the senior writer, and built for the Ceramic Lab- oratory of the University of Illinois, as shown in Plate I. K KITTED GLAZES. 26 • 1 I > t 26 FRITTED GLAZES. This kiln was fired with coke, on the following time sche- dule: Cone 010 in 12 hours, 1 hour soaking, rapid cooling. Cone 05 in 14 hours, 1 hour soaking, rapid cooling. Cone 1 in 16 hours, 1 hour soaking, rapid cooling. Cone 5 in 18 hours, 1 hour soaking, rapid cooling. Cone 10 in 18 to 24 hours, 1 hour soaking, rapid cooling. These heats were easily attained in the time allotted, with a very thin (3 inches) bed of live coals and an average of 20 to 30 minute firing periods. In fact, in all but the cone 10 burns, the raising of the heat was intentionally checked, so as to insure as close approximation to the above temperature schedule as possible. In one of the cone 010 burns and one of the 05 burns, the kiln was "smoked" in the early part of the burn, which caused a deposition of carbon in the glazes that colored them black. This, however, did not injure the character or reduce the value of the results, for fortunately the smoked glazes were not in the series that matured at these tem- peratures. Twenty-eight series or one-half of the entire 448 glazes were burned at a time, thus necessitating ten burns in all. The setters were placed in four bungs. In the first few burns, cones were placed near the top and bottom of each bung. In as much as the cones burned down equally in all eight positions, a fact that checked similar experience in this kiln in several previous burns on other forms of ware, sufficient confidence in the absolutely equal distribution of heat in all parts of the firing chamber was established, so that in the later burns only two and sometimes one set of cones was used in a burn. RESULTS OF THE VARIOUS BURNS. Group I. A table of the compositions of the 64 glazes composing this group follows on page 27. 3ooo £MOpu o ^ o O CM CN C IC Q E CI I i-l o = 8° "I C0« c o ~v «• COM 55 ^ £0! 00 coP, c o z - co* O CN eo CN (M CN ~cn co cT~ _ rt ,-< ,-H rl CM CN 000000c 0000000 00 CN I- M W « eo co eo lO CN 05 OS O .-I O «-> CN CO IQ CO lO I— CO O »-t CO l» OJ ih * «0 M I» O t* t>- O CO CM «* IO CO CO 05 IN 41 r- _ rt< 1^ O co CO r~ -+ O co CO OS iO xH iO t~ 05 CN CO iO 1-1 l-H •""' CN CN CN EN t- CO CO •<* o CO i-H CO CO OS lO o iO .— _ •«* r~ _ ^h r^ cc CO CO l~ I— t^ 00 K 1 - 05 CO iO r- — 1 CM CN CN CN CN cc co ko r- Oi 1— 1 co o 1— ci eo 10 co CO CO CN CN o —. — o o O CO CO OS CO O CN "* CO OS CN CN CN CN CN s a 1— 1 co ■** co OS CO I- co CN r~ t— C5 O IO CO GO 2 8 fi co co OS S r~ OS CO CO CO CD a> CN OS 1^ CO CO OS Tf< r~ OS CN OS n CM CN CN CO CO eo CO •»* z« O O cr O OS CN «*c •o CO 1^ CO OS OS ^4 CN iO T* CN OS os OS O V <* -r iO co 1^. on OS OS co OS 53 05 O f-H CN eo -3< 10 l~ O . O O O O 10 O IO F1 S L ~ „ O 1—1 CM O CO CO O c a per eqv. 2 3 3 u •% .2 • 5 3 5? ■ 3 d 53 =3 ^2 O •d © =w u * OQ 27 28 F KITTED GLAZES. Fired at Cone 010. Series 1. All fused to glasses; none attacked the body. Members a, b, c, devi trifled and badly crazed in both thin and thick places. Member d slightly devitrified, crazed, good gloss on most of surface. Members e, g, f, h, good glazes where thin, but have small pin holes on surface where thick. Pin-holing increases as B 2 3 decreases. Crazed whether thick or thin. Series 2, 3, 4. All devitrified and all effloresced. Devitri- fication decreases with increase of A1 2 3 . Series 5, 6, 7. Are like 2, 3 and 4 except that with low A1 2 3 crackling or separation of the glaze begins in series 5 and increases progressively until it is the most manifest in series 8. Series 8. Is very badly crackled, the glaze patches exhib- iting a vitreous sheen. Summary of Group I at Cone 010. (1) The Eqv. content of A1 2 3 with which the best glazes are developed at this temperature ranges from 0.30 to 0.40, with the ratio of Si0 2 to B 2 3 1 :0.25. (2) Devitrification decreases with increase of A1 2 3 and the decrease of B 2 O s . (3) Efflorescence was general in whole series. Fired at Cone 05. Series 1. Devitrification is less at this temperature than at cone 010. Crazing is finer meshed in the glazes which have lower equivalent of A1 2 3 . Crazing more pronounced in all glazes at this temperature than at 010. Series 2, 3. All glazes of these two series are dimmed less at this temperature than they were at 010. The most fusible member of both series is "e," having an A1 2 3 content of 0.30 Eqv. In both series, it has but a trace of dimness. FRITTED GLAZES. 29 The members f, g, h, in both series show graded in- crease in refractoriness. Series 4, 5, t>. These three series exhibit a most peculiar appearance in that they have passed from what ap- peared to be devitrification in the cone 010 burn, to a blistered dull surface due to the boiling that precedes quiet fusion. Members e and f seem to be the most fusible of these series. Series 7. Members a, b, c, present same crackled appear- ance as at cone 010. Members d, e, f, are devitrified, matt-like in appear- ance; e being a beautiful matt, but badly crazed. Series 8. Carbonized badly, but are apparently more fused than at cone 010. Summary of Group I at Cone 05. (1 (2 (3 (4 (5 (6 (7 (8 Good glazes having an A1 2 3 equivalent of 0.3 to 0.4 occur in Series 1, 2 and 3. Devitrification decreases with increase of A1 2 3 and decrease of B 2 3 . None effloresced. Fair matt surface occurs in Series 7 with 0.30 Eqv. A1 2 3 . Crazing is more pronounced in cone 05 burn than at 010. Glaze e, series 1, has the longest range so far devel- oped, being good at 010 and at 05. Glazes f and g are very nearly matured at 010 and fully so at 05. (razing decreases with increase of A1 2 3 and changes from fine mesh to long hair lines. Fired at Cone 1. Series 1 . Members a, b, c, d have good gloss ; are perfectly matured ; crazed in fine meshes and have eaten into the body, glaze "a" being the worst in this respect. 30 FRITTED GLAZES. Members e, f, g are good, well matured glazes; "e" is considerably crazed but the craze lines are long and some distance apart ; "f" is less crazed and "g" has but one craze line. Member g is not quite matured. What is stated in the discussion of the 05 burn in re- ference to the decrease in devitrification phenomena of ser- ies 1, seems to find confirmation in the cone 1 burn, for the glazes that were badly devitrified at 010 were less so at 05, and not at all at cone 1, but at cone 1 the edges of the tile are eaten away by the glaze, and this eating away of the edges decreases as the glaze increases in A1 2 3 . This lends support to the doctrine (1) that A1 2 3 counteracts devitri- fication; (2) that glazes having the acidity of this group must contain at least 0.30 Eqv. of A1 2 3 to satisfy the acid content. This series brings out another very significant fact that has been noted in many other isolated examples, to- wit: that when the glaze is compelled to feed upon the body to gain the A1 2 3 required to make a perfect glaze, fine mesh crazing is sure to follow. This increase in craz- ing with increase of A1 2 3 does not accord with Seger's law in which the introduction of bases of high molecular weight is credited with power to decrease crazing. But, as will be noted later, Seger's law in regard to decrease of crazing with increase of A1 2 3 does hold true when the original content of A1 2 3 of the glaze is considered. In- crease in A1 2 3 by eating into the body increases rather than decreases crazing. No scumming or devitrification was apparent in any member of this series. Series 2, 3 and 4. The members of these series resemble series 1 in every respect save that of maturity. In series 1, all members but h are fully matured, while in series 2, 3 and 4 g is likewise unmatured. The fine-mesh crazing in a, b, c, d, and to some extent in e, is shown in these series as in series 1, but the KBITTED GLAZES. 31 members f, g, h, are freer from crazing as the 1^0. ; decreases, i. e. progressively from series 1 to scries 4 inclusive. This would seem to be rather significant in view of the fact that all the succeeding series have dimmed surfaces which resemble devitrification more than immaturity. Series 5 and 6. Not a single member of these series is matured. Members a, b, c, d, and e show progressive decrease in dimness (probably devitrification) with increase of A1 2 3 . None have eaten into the body. Series 6. Members a and b are smooth, devitritied and fine-mesh crazed. Other members are in the "boiling" stage. Member d with 0.25 A1 2 3 exhibits this more than the others. Series 7. All members of this series cover the tile perfectly showing that increase of heat treatment from cone 05 to 1 has been sufficient to cause these glazes to flow enough to pass from a erackeled condition to a perfect coating of glass. Series 8. All members are dim with slight evidence of devi- trification. All are crazed in fairly fine meshes. Summary of Group I at Cone 1. (1) Good Glasses developed at this heat are as follows: Series 1 a to h inclusive. Series 2 a to g inclusive. Series 3 a to g inclusive. Series 4 a to g inclusive. (2) Good glazes, having only hair line crazes, or none ;it all and being free from pinholes. Series 1 e, f, g, h. Series 2 e, f, g. Series 3 f, g. Series 4 f, g. 32 FRITTED GLAZES. (3) Glazes d, e, f, series 7, that exhibited good matt text- ure at cone 05 were "boiling'' at cone 1, showing that at cone 05 their matt surface was due entirely to im- maturity. (4) At this temperature, eating into the body decreases with decrease in B 2 ;! and increase in A1 2 ;; . (5) Crazing passes from fine mesh to total absence with increase to A1 2 3 . With low A1 2 3 (0.1 to 0.25 and 0.8 inclusive) decrease in B 2 3 does not seem to affect the character or amount of crazing, but with higher A1 2 3 there is shown a progressive decrease in crazing with decrease of B 2 3 . This would indicate at least four important facts. (a) Increase in A1 2 3 decreases crazing. (b) Increase of Si0 2 , retaining constant oxygen ratio, likewise decreases crazing, but (c) A1 2 3 is a more powerful factor than Si0 2 in checking crazing at this heat treatment and oxygen ratio. (d) The facts noted in a and c only hold true when the original Al 2 O a content is consid- ered. When the glaze is compelled to borrow A1 2 3 from the body, a strain is established that increases crazing in proportion to the ex- tent to which the glaze has attacked the body. On looking down into the thicker portions of the glazes which have eaten into the body, and which exhibit this fine mesh crazing to the greatest degree, there appears to be a separation or splitting between portions of the glazes next to the body and those nearer the surface. It is quite evi- dent from this, that the portion of glaze in contact with the body is of a different composition from that near the sur- face. There cannot be ready diffusion of materials in glazes of this oxygen ratio, even when relatively high in B 2 3 . From the fact that the glazes lowest in A1 2 3 have eaten into the body most, exhibit fine-mesh crazing to FBITTBD GLAZES. 33 the greatest degree, and show a separation be! ween the por- tion of glaze contiguous to the body and that above, it is concluded that this fine-mesh crazing is due more largely to extraction of A1 2 3 than to the extraction of Si0 2 from the body, which owing to its viscosity, diffuses very re- luctantly, and further, that this fine-mesh crazing increases with the increase of A1 2 3 so obtained by the glaze. (6) Devitrification extends only from series 5 to 8 in- clusive, or over a proportional range of Si0 2 B 2 3 from 1 : 0.9 to 1 : 0. (7) It is indeed a most surprising fact that with the oxygen ratio of 2, which is best suited to the development of good glossy raw-lead glazes free from boracic acid, there 6hould not be developed a good glass, when a portion of the glaze is fritted, until the ratio of silica to boracic acid has been raised to at least 1 : 0.13 and then only within a very narrow range of variation in A1 2 3 . Indeed, as will be seen in the study of this same group at cone 5, fritted glazes having an oxygen ratio of 2 have a heat range that is limited to but a slight variation from cone 1 until the ratio of Si0 2 to B 2 3 has reached 1 : 0.2. Even at the Si0 2 — B 2 3 ratio of 1 : 2.0 at least cone 1 is required to ma- ture the glaze, and at cone 5 it has withstood its maximum heat treatment. On the other hand, when the boracic acid has been in- creased until the ratio of Si0 2 to B 2 3 stands at 1 : 0.25 there seems to be established a degree of fusibility and a restraint against devitrification that permits of heat treat- ment ranging from cone 010 to at least cone 5 inclusive, provided that A1 2 3 content originally incorporated in the glaze is at least equal to 0.3 or 0.4 equivalents. Fired at ('one 5. Scries 1, 2, 3 and 4. All glazes are crazed, fine-mesh craz- ing the most pronounced with low ALO.. and increas- ing BoO ; . Glazes that were not crazed at Cone 1 are crazed at Cone 5 in long hair lines. 34 FRITTED GLAZES. ' Except for crazing, the following are good glazes: Series 1 d, e, f, g, h. Series 2 c, d, e, f, g, h. Series 3 none. Series 4 f, g, h. In series 1, fine-niesh crazing is about as it was in same series at Cone 1, except in case of member d, which has flowed and run off the tile, leaving only a comparatively thin coating of glaze. While the body shows evidence of having been attacked to some extent, the glaze is coarser crazed and freer from horizontal crazing than member e, which is thicker. This fact suggests three things. (1) Fine-mesh crazing can be decreased by de- crease in thickness of glaze, thus permitting equal diffusion of the A1 2 3 obtained from the body through- out the whole mass. (2) The body will be attacked less the thinner the glaze. (3) That fine-mesh crazing is due almost en- tirely to the unequal coefficient of expansion and con- traction of the upper and lower portion of the glaze layer. In all of the series of this group, the fine-mesh crazing is exhibited in glazes which at cone 05 were either free from crazing or were crazed only in hair lines. Members g and h in all series are still free from this fine-mesh crazing, but it is evident that as the heat increases in in- tensity, even though not in length of time, the glazes originally higher in A1 2 ;! are beginning to attack the body, causing tension between the upper and lower portions of the glaze, that causes either actual or incipient horizontal as well as vertical crazing and as a consequence, fine- mesh phenomena ; and that in the glazes originally low in Alo0 3 the alumina incorporated from the body is much more thoroughly diffused, causing, as a r- on sequence, a de- FRITTED GLAZES. 36 crease in the fine-mesh crazing over that shown with less intense heat. These facts are very clearly shown at cone 10, where the glazes having the finest mesh crazing were originally highest in AUO3 and at lower heats, in some cases, are entirely free from crazing. Series 5, 6, 7 and 8. Devitrification is now shown only in members a, b and c of series 6 and 7, and in all members of series 8. None are free from crazing. Members having highest original content of A1 2 3 ex- hibit pin-holing. No glazes of promise shown in any portion of these series. Summary of Group I at Cone 5 will be included with the summary of the facts deduced from the Cone 10 burn. Fired at Cone 10. All members of every series of Group I at this heat treatment are crazed in fine meshes. In series 1, member a is not so very finely crazed, but the fine craze meshes increase regularly from a up to f, which had originally 0.35 Eqv. ALO.j and then decrease slightly from member f to'h. In series 2, member f again marks the point of maximum fiue-mesh crazing. From series 2 to 8, the area of mazimum fine-mesh crazing increases until in series 8, every member is crazed in exceedingly fine meshes. Every glaze ate into the body considerably, those showing maximum fine-mesh crazing being no worse in this respect than those showing this feature in a less degree. Summary of Group I at Cones 5 and 10. (1) It is quite evident that for fritted glazes of this oxygen ratio and RO, on this body, cone 5 is beyond the p, A P.— 8. 36 FBITTED GLAZES. maximum limit of temperature, for all the glazes are more or less crazed in fine meshes. On a body that would not give up any of its constituent parts to the glaze, or if the glazes were dipped as thin as is the practice in the white ware industry, many of these glazes may have a heat range that would include at least cone 5 if not cone 10. The ex- treme thickness at which these glazes were applied, per- mitted the formation of two strata, the one next to the body containing without a doubt additional A1 2 3 , while the upper strata was not altered materially in composition, except by the normal volatilization of B 2 3 and alkalies. Conclusion mi Group I. 1. The good glazes developed at the several tem- peratures are noted in the following table. The glazes which have a question mark beside them are good glasses, which might have been good glazes if they had been ap- plied thin enough to prevent the formation of two strata in the glaze layer, thus causing fine-mesh crazing. Normal crazing is not taken into account in designating a glaze as good. Series Cone 010 Cone 05 Cone 1 Cone 5 Cone 10 1 efgh d?e?fgh b?c?d?efgh b? c?d?e?fgh (atoh)? 2 e?f b?c?d?efg b? c?d?e?f?g? (a toh)? 3 ef b?c?d?efg (a tog)? (atoh)? 4 c? d? e f g (a to g)? (b to h)? 5 f?g? (b to c)? 6 f?g? none 7 e? f ? g? none 8 (btof)? 2. Series 8 of this group demonstrates the fact which was stated by the senior writer, in 1904 ; viz., a fritted glaze must have a higher oxygen ratio than 1 : 2 or that normally used in raw lead glazes. True, good glazes with a fair temperature range were developed in this group, but they required the maximum content of B 2 3 . FRITTKD GLAZES. 37 3. Fine-mesh crazing appears to be due to unequal dissemination of constituents taken from the body, and is more pronounced at the low temperatures in those glazes that are lowest in Al 2 O a irrespective of their B 2 3 content, and as the intensity of the heat increases, fine-mesh craz- ing decreases in the glazes low in A1 2 3 , in consequence of the compounds extracted from the body, and progresses steadily with increase in intensity of heat until the glazes having an A1 2 3 content of 0.3 to 0.4 Eqv. that were per- fect at the lower heat treatment because crazed in fine meshes at the higher heat treatment. 4. Devitrification decreases as the A1 2 3 increases, either as originally added or taken from the body. Group II. A table of the compositions of the 64 glazes composing this group follows on page 38. ?ooo — • 1-HU5 CM © o © © o w pi o o CO o C CO K 8.2 Q CM IO ci oo o o o qo o to t- ©3 CM © o o IO © —* r-l CM CI o CO C O pq © T-H © © I-H o © CM © CM cm © * 00 CM © -i lO >o © CO 00 CM t~- — ' CM CO to © 00 © i— I to © co © CM © © CO © CM CO CO CO ■* 10 a 00 CM IO © H M T* W © CM © — — ,-f — — i — < CM CM S3 £l O «H Jio -A .as 0© =g CO 43 co ,• « r>3 38 •Sod '43 05 =1-1 FRITTED GLAZES. 89 Fired at Cone 010. Series 1-8 1. All were heavily impregnated with carbon, so that but little can be said of the behavior of the glazes of this group at 010. 2. Members d, e and f, of series 1, 2, 3 and 4, seem to be the most fusible. Fired at Cone 05. (1) Members d and e of series 1, were the only two good glazes developed with this heat treatment and thickness of glaze. The entire group had evidently been subjected, in burning, to the influence of carbon, shortly after mem- bers d and e had been fused into perfect glasses, for the less fusible glazes show either an undulating or a pimply surface, where they are thick, on account of the expulsion of CO or C0 2 generated by the com- bustion of the carbon, but were smooth, well developed glazes where thin. (2) The glazes retaining the carbon to the end of the burn are to the left of a line drawn diagonally from member a in series 1 to member h in series 8. This according to Seger 1 demonstrated that the glazes to the right of this diagonal line were nearly formed into glasses at the time that the kiln was ''smoked.'* Further, those glazes which were fairly well devel- oped at cone 010 and are good at cone 1, show least of this undulating and pimply surface. From these facts at least three conclusions can be drawn. (a) The most fusible mixture in series 1, 2 and 3 is that with 0.30 Eqv. of A1 2 3 . (b) In series 4, the most fusible mixture contains 0.35 Eqv. of A1 2 3 ; in series 5 and 6, 0.45 Eqv. of A1 2 3 . 'Collected Writings of Herman A. Seger. Amer. Cer. Soc. trans. Vol. II, p. 592. 40 FRITTED GLAZES. (3) Judging from the appearance of the glaze where thin, and the relative degree of maturity at cone 010 and cone 1, the following glazes would have been well developed had they not been "smoked." Series 1 d, e, f, g, h ? Series 2 c,?d,?e, f, g? Series 3 c?d?e, f, g? Series 4 e? f ? Series 5 0. Dimness of surface is equally pronounced in this group when either Al 2 O s is low and B 2 3 high, or A1 2 3 is high and B 2 3 low. Fired at Cone 1. (1) The more refractory glazes of the first four series exhibited surface pinholing or pitting, especially where the glaze is thick. The surface pinholing could readily be taken as indicating an over-burned condi- tion, but such could not be the case, for they show no indication of being over-burned at cone 5. (2) Members of series 4, 5 and 6 are likewise pinholed, but owing to their being less fusible than the members of the first three series, most of the trial pieces appear to have a narrow border of normally fused glaze sur- rounding a more boiled and pitted patch in the center. (3) Series 7 and 8 are slightly blackened by carbon, showing that the whole group had been smoked. These three facts suggest that : (a) Pinholing, which appears at times when every condition seems to be normal, may be traced largely to the carbon which was entrapped when the glaze was almost matured, the combustion of which produced gas that devel- oped blisters, or blibs, which finally bursted, forming small pits or pin holes. These pinholes differ somewhat from the blisters due to over-burning, in that the latter extend much deeper into the glaze layer, and are frequently much larger in diameter. KKITTED GLAZKS. 41 (b) It cannot as yet be determined whether the car- bon reduced the lead, thus altering its chemical activity in respect to the boro-silicate formation that is taking place, or whether the presence of carbon and the consequent car- bonic gases make the glaze more viscous. There is evi- dence in this group that might be taken to substantiate either claim. The essential fact is that if one portion of a glaze has been smoked, and another portion not smoked, the former will have every appearance of being over-fired save that of the nature of the pinholes, while the portion not smoked may be a normally developed glaze. (4) The glazes of group II, which either are good, or would have developed into good glazes at cone 1 were it not for their having been smoked, are as follows : Series 1 c, d, e, f, g> h. Series 2 c, d, e, t g, h. Series 3 b, c, d, e, f, g, h Series 4 b?c?d, e, f? < « « ei w w n O pa" CM © O o o 0C m CM O oo CM O CO o co © iO CM CO o i— CO o O o o o o o o © p" 55 CM r. © CM •*» CM cm CM CO oo iO CM © I- O oo CM © CM o CO CM CM co © CM ■** CO 6 pa" t- o OS o CM O CO 1-H 1-H CM CM CO CM co CO co iO O co © o o O o C c o q o 00 CO CO CI o CD CM CO CM os CM iO CD »o o CD CO 00 © CD © OS CI pH CM CI — CM CM CO co coiococMooor-ic «*ICMOOOCOCO'-IOS «3t-OSOC-lr* •** t- 00 CM pq -: i CM co co co T* Tf ^ •* ^^ O © o © © © © © © i— i © © nr i co © © CI IO r~ co CO CM © © © co O iO CO © t~ IO CM © © Ax iO t~ © © CM 1" © 00 1-H .— 1 _ CI CM cm CM CM I 3 © CM Of) **> _ 1— ^> s -* © CO 00 CO r- CM m 1 3 © o co 1- Q "* © do pa , co co *# •<♦" f IO >o © 5.2 © © © © © © © © co 1— c CO © © r _ C i CM CO •*r iC »o 8 co © © CI iO CJ() — c t^ © © 2 © i/.^ m © oo © i— ' co t- _ _ _ 01 01 Cl d CM o r~ CM ,_ i © © •Y iO © t» 00 © © © o iO os co r- a © -r 6o pa CO CO ^< ^r >o © © © 3? 5.2 © © © © © © © © CO © © oo s — © CM CO **> 00 © CM © © s o 1-H oo ■>* © r- CO © © © •>.ai ^t< © t- © © C» so iO _M ,_! _, art 01 Tl Ol CM .3 oO fl >P C 5 © to © ID © lO © iO •^ © « »o* © 1-H © CM © co © co © © © «°3 O M^ I? .2 S > 'J3 i «w "3 S.O ■ V- * — o n -n— LO CM o a. o o -5 o oo 2- _ o Q LO © lO o o Q LO o I— a CM lO t~ CO CM o t- CO o CO CM on ■o CM 10 j- o co o oo i-H * S.2 c/3 35 Z^ io« o o r. — Z- kK CM CM CM CO CO CO CO "»*<©©CO©iOCN00 •— • © co r- io^co— « -♦■©LQ^r^coWLO 04lOt«OCNlOL-*0 CNo CO _ © CM CM co T LO © O 00 © * r- © CM •o 00 © co ■* LO r- CO © © CM co o © eo © © CM © © CM X * © CM ■* © © •"< co © ■M CN CM o ~f" CO >o t» © CO o o lO iO •o iO "0 © © CO Pi CO t- © *1 CO LO r- © *1 — < t~» CM ci CM CN CM co SO <-H © © © LO LO © © © © *¥ lO © •f CO 20 00 © — © © oo © <— — cm oi ri CM CM CM <— — CM CM CO © LO © © co X> 55 >&0 ay? -§$ a - J? 54 FIUTTED GLAZES. Group IV is so very much like Group III in every way, that it need not be described in detail. The well matured glazes of this group are as follows : Series Cone 010 Cone 05 Cone 1 Cone 6 Cone 10 1 (btoe)? abcde abcdefg abcdefgh? abcdefg 2 abed abcdefg abcdefgh? abcdefg 3 abc abedef abcdefgh? bede 4 abedef abedef h? 5 bed abcdefgh 6 abcdefgh 7 c d e 8 Group V. The table on page 55 shows the compositions of the 64 glazes entering into this group. fcMOPn C 8 g eo I 30 o o ■ oo a 2-2 c" CO | JO cm cm © U3 X 1 eo •** oo io cm to -J3 N co eo eo eo ** t»< 2^ o o oo 8.2 tXaS o i- iO CO ,_, on s ■** •*J1 © 35 CM ia r^ eo o CM cm CM CO eo eo "f ■*»< SQ o © © © © © © © © © © © © © © © iQ so T o $ i— 00 © ~t> »o CM <£> eo © o © t^ lO eo on § rt< ■"!• © © CM lO t» CO cm cm CM eo eo eo -f ■**• —i cm eo io © © © « © IH CM s ffl iO - a CM CM eo eo eo eo r» o z ~ SQ 8.2 Jr a o ioi in eo©©c © © r~ © _ eo T* © eo t— 00 © O •»»< © 3 © ** © -r © oo « •»»< ■«*< © © © t~ 00 z« © o © © © © © © t- eo Of) ■** © •o s © CM t— © s lO ■«i< i <• o r~ t^- 00 OO © © © v.X t~ © '- , co lO t- © CM — fH CM Ol CM CM co co s »cr © iO © CM 1? CM eo lO CO © © CO © o © CM (M M « © ■<»< *> T* © © CO cc t^ t~ © © © -«f © IM © o ■^1 © o O IM © ■* © t- © o do 03 ■"" ' rH I-H •"J •"• '-' M 55 t^. © u O © © © © © © © ©' © © © © l» t- © © © © M4 © r- © © ^ Tl« © (M CM O ^ rt o © t— © « © t^ 55* 55 t* © © IM m CO © CO © © CM • © CM © © © CO © © O © »H ^J< © »H Tt« h- ■a CO do tS CM « © © rf* ** • © © g CO t~ I-H re © M © © ^i © © wi © CO © t-- © * © o © CO © © IM © © © o' mtX 55 © S3 ■f © © •^ © © oi CM CM * © O N - V z as © o T3 =t-i tc |S6 > 0> r,8 FRITTED GLAZES. 61* Fired at Cone 010. Group VI at this cone is wholly immature. No efflor- escence. Members of Series 8 are perfectly flat and are por- celain-like in character. Fired at Cone 05. Series 1, 2, 3. The following series were well matured and completely in solution. Series 1 a, b. c, d. Series 2 a, b, c, d. Series 3 a, b, c, d. While the above members had reached good maturity, members g and h of the same series were just sub- siding from the first boiling stage. Series 4, 5. The members of these series were all imma- ture glasses more or less full of undissolved materials. Members a and b are almost clear. Series 6, 7, 8, were porcelanic in character, decreasing in glossiness with decrease in proportion of B 2 3 to Si0 2 . Fired at Cone 1. Series 1 and 2. Members a to g inclusive are good glazes. h has some undissolved material, a and b are the onh- members of this series that are crazed. Series 3 and 4. Members a to e of this series are well developed, g and h have not settled down to a flat surface. Series 5 and 6. Members a, b, c, d of this series are prom- ising. The remaining members have not reached quiet fusion as yet. Series 7. Member a is fairly good glaze, while the re- maining series are porcelanic in character. Series 8 are porcelanic in appearance, resembling an aver- age white Bristol glaze. 60 FRITTED GLAZES. Fired at Cone 5. Series 1 to 5 inclusive. Every member of these series ex- cept members a of series 4 and 5, are good glazes, a of series 4 and 5 are crazed in extremely fine meshes. Series 6 and 7. This is the only group which seems to have been smoked at Cone 5, and as a consequence there is in these series the phenomena, before noted, of patches of pin-holes in an otherwise smooth, well-ma- tured matrix. Members c, d, e, and f, present such an appearance in series 6 and 7. Series 8. Member d of Series 8, Group VI, is the first instance of any of the glazes of this series maturing, except in case of d Series 8, Group 1. Fired at Cone 10. The only good glazes in this group at Cone 10 were Series 5 and 6, members g and h. All other glazes were over-burned. Summary of Group VI. The good glazes are as follows: Cone 10 gb gh (2) Opalescence phenomena are shown in member "a" of Series 1, 2, 3, 4, at Cone 5 and 10. Group VII. The table on page 61 gives the composition of the 64 glazes composing this group. ries Cone 010 Cone 05 Cone 1 Cone 6 1 abed atog a to h 2 abed a tog a toh 3 abed a b cd e a to h 4 a b abode btoh 5 a b abed btoh 6 abed c d e 7 a c d e 8 d S?ooo O O © c 00 o o c" © n t^ z « © © o kO s iC O © © lO 35 CM r^ LO CM % r^ © O cm eo O co I-- © C0« «j o> CM © © CM © © CM CM CO CO co •»** -r -»• © „ CO 8 © © SB CM kO 3 CM © CM o CM CO CO S "* -9< LO o d sa o CO © © © © © © Z - o o © © © © © © CM © © 8 so eo CO co 00 © © © o oo 1 CO OS 2 CO i © © !/3« CO oo CM CO CM © © CM ■M CO co CO •<# -9" i« © ^ t-- Tj< ,— , 88 lO CM © co t~- CO ■"*! © © o CO © © 00 © « T o'o oa rH I— 1 f* "■"' CM CM CM © © o o © © s © © © o» © © © © ca en CO I s - t^ © © o ^« 00 © © © 00 © © O CM eo ^» © r- © ■S.X i~« o eo eo © CM © © © © CM CO cc eo eo •<*» ■<»< «* © 8R © iO CM CM £ © © CM © © o CO © CO eo CO © 1-H oo 83;; S.2 « (M O CM o CM © eo © CO © © « © © eo © © © CM lO oo c* © © © © •*»< H © © © oo CM © W35 o © ao f-t CO CD 8 © © © © © CM CM eo eo eo ■**« ■»* Tfl CM co en © s © lO © © © CM © o lO © CM © © eo t- Oo pa" eo co CO «■* ■*»« © © © « —4 55^ 5 — o © © © © © © © 1 © © © © O CO r- O-l © r- o CM LO f- © CM ■*»• 94 © o 35 Tt< CO © e- © ■"f 34 «* t- © ■^ t~ © ■<* © CM CM CM © TO eo 'J" ■<*< CM © © ^r __ © © CM t~ eo © •*»« © © o OS •+ oo 3 t~ CM »— 00 CO T" ^< © © © 5« ca © © © © © © © O a* © © CM © © 5 © © © CO tO © CO kO CM CM CO O © © eo © t~ ■*t> © eo © © T»< © © CM cm CM CM CO © CO « Tt< o © on co L^- sc © ■^ eo >o t- © ^H « **S o © © lO o i0 © © T* © © © © t- I s - © ■M CM z - u ° oa o © © © © © © © © CM © © o 8 © © CM © oo oo $ oo 1^ CO I— 13 O § s © CM © 00 CM © © © $ © © CM CM CM co eo © © ■# 00 CM CO © CO t- o ■"f co LO t- 00 co © © c- © © ca © © © © © © © © © © © © © © CM © I-H 5 CO CM cm 00 © © LO © 8 CC «3 insi g t/5 cm eo © © CO © lO eo © © © CM CM CM CM w » eo CO C "> o © © © CM © © © eo 5 © a © •3© a © js? .HoO 4 5 U o © © © © © © © Ol,°0 o ih O a r> a S x§ J vi O (0 ol Of _ K v t > -J V. ! 1 i t_... .,_ ... - €> © i 1 / © \ V > ^""**^ 5' ft/ ^ \ > / &\ / i ® y--'" • ** V® / \ f \ \"***« o\ © © ! 1 d « ! lr \' 8 ^ CM 4 S-OIXVy N19AXO TO FRITTED GLAZES. Glazes plotted in area IX have a heat range from Cone 1 to Cone 10, being fully matured at Cone 1. Glazes plotted in area V have a heat range from Cone 05 to Cone 10, being fully matured at Cone 05. The maximum heat range then of glazes containing 0.2 equivalent of A1 2 3 , is to be found in Series 1, 2 and 3 of Groups III to V, inclusive. Figure 6. Glazes plotted in area I are over burned at Cone 5. Glazes plotted in area II are immature at Cone 5, but over-burned at Cone 10. Glazes plotted in area III have a heat range from Cone 1 to Cone 5. Those of the lower portion are harder than those of the upper portion of this area. Glazes plotted in area IV have a variable heat range, those having an oxygen ratio of 4.5 showing a range from Cone 1 to Cone 5; those having an oxygen ratio of 3.5 to 4, inclusive, showing a range from Cone 1 to Cone 10, when B 2 3 is high, while those having an oxygen ratio of 2.5 to 3.5 show a range from Cone 05 to Cone 5. No explanation of this variableness in heat range of the glazes of the area can be offered. Glazes plotted in area V have a heat range of Cone 1 to 10. Glazes plotted in area VI have a heat range from Cone 05 to Cone 10. Glazes plotted in area VII have a heat range from Cone 010 to Cone 10. Glazes plotted in area VIII have a heat range from Cone 010 to Cone 5. Glazes plotted in area IX have a heat range that aver- ages from Cone 05 to Cone 5. Glazes plotted in area X have a heat range that ex- tends onlv from 1 to 10. FRITTED GLAZES. 71 •» « . dOoa» > UJ a. O E K \ 1 1 1 1 ©1® i i ^ 1 / / ® / u. Q Z N \ ©\ ©V J ! j / / > O tr 0- \ \ / ® ! / / / / / \ *> — * **^* . 1*^ \ g) 1 j ® . 1 ! / 1 v • 1 ^* 1 ^ • \ * i V \ 1 / \ i c . z J M 1 © / 1 1 J j o o si's o k \ © 1 1 1 < 1 (0 2 I cr 5 i \ 8 § 1 a 6 CD (ft O soixvy 72 FRITTED GLAZES. Glazes plotted in areas XI and XII have a heat range from 05 to 10, those in area XI being immature and those in area XII being fully matured at Cone 05. The glazes having the longest heat range are those plotted in areas V, VI, VII, VIII, IX and XII, or those having an oxygen ratio of from 2.5 to 4.0, inclusive, and a Si0 2 -B 2 O s ratio of 0.17 to 0.25. Figure 7. The description of Figure 6 applies to Figure 7, ex- cept for minor details. FRITTED GLAZES. 73 cu3flwnN aiyotiO Cpl-LVU KJ30AX0 74 FRITTED GLAZES. Figure 8. Glazes of all areas except I, VII and VIII have heat ranges, the maximum points of which are Cone 5 in the case of II and III, and Cone 10 in the case of IV, V and VI, and the minimum points of which lie in the curve bounding them on the right, except in case of III. The longest heat range shown in the case of glazes having 0.35 Eqv. content of A1 2 3 , is found at the oxygen ratios of 2.0 to 3.50, inclusive, and a Si0 2 -B 2 3 ratio of from 0.25 to 0.17. KBITTBD OLAZK8. 75 Sa39wnN dnouO .JOLLVii N3 9AXO 76 FBITTBD GLAZES. Figure 9. Glazes plotted in area I were immature even at Cone 10, except when the oxygen ratio is low. Glazes plotted in area II have a heat range from Cone 1 to Cone 5. Glazes plotted in area III have a heat range from Cone 1 to Cone 10. Glazes plotted in area IV have a heat range from Cone 05 to 5 or more, being somewhat immature at Cone 05. Glazes plotted in area V have a heat range from Cone 05 to nearly Cone 10, being fully matured at Cone 05. Glazes plotted in area VI have a heat range that ex- tends from a little above Cone 05 to Cone 10, while those of area VII have a heat range that extends from Cone 05 to Cone 10. The glazes plotted in area VIII have a heat range from Cone 010 to 10, inclusive. The glazes having an A1 2 3 content of 0.40 equivalents and showing the longest heat range, are found within the oxygen ratios of 2.0 to 3.5, inclusive and Si0 2 -B 2 3 ratio from 0.25 to 0.20, inclusive. FRITTED GLAZES. 77 V -- * VO 4 • sa38kMON ■> 1 henomena observed in raw lead glazes, there seems to be no doubt that the statements in regard to the cause of pin-holing given above are correct. In conclusion the writers wish to express their appre- ciation of the hearty support and substantial encourage- ment they received from the authorities of the University of Illinois in general, and from Professor C. W. Rolfe, Di- rector of the Ceramic Department, in particular. Had they not granted every facility within their power, the exe- 'Trans. Am. Cer. Soc. Vol . VII, p. 356. K KITTED GLAZES. 81* cution of this study In its details would have been impos- sible. We wish also to express our further appreciation of the privilege of offering the results of this research to the American Ceramic Society previous to its issuance as a University Bulletin. DISCUSSION. The Chair: By the presentation of this very able paper, I am impressed at the outset with the growing im- portance of making provision, as soon as our financial con- dition shall warrant it, for the publication of such papers prior to our meetings, so that we may study them, and be better able to discuss them when we come to the meeting. We have realized this necessity during past sessions, but the need is growing more apparent all the time. Still this paper contains so many interesting points that we ought to have a general discussion on it. Mr. Par melee: I feel, of course, wholly unable to discuss the detail of this work. It certainly is a big sub- ject, and I think brings pleasure to all members to see such a substantial contribution on it. Did I understand you to say, Mr. Fox, that coke was used as a fuel ? Mr. Fox: Yes, sir. Mr. Parmelee: Then how do you attribute the pin- holing to carbon deposited on the glaze? In what form is it deposited on the glaze? Mr. Purdy: Coke certainly contains carbon, which on combustion is driven off. Not all the carbon, however, is oxydised to either carbon monoxide or carbon dioxide; some of it may be carried by the draft as particles of free carbon. Mr. Parmelee: Then these are particles of fuel, car- ried mechanically and deposited mechanically? Mr. Purdy: We know that black smoke contains about two percent, of carbon, and when the coke is freshly thrown on the grate bars there will be an evolution of smoke or black gases, and that is sufficient to cause a de- 90 FRITTED GLAZES. posit of carbon on the glaze, notwithstanding the fact that the saggers are thoroughly sealed. The Chair : You assume that in all of the kiln firings you formed this deposit, but in some were able to free the ware from it more completely than in others? Mr. Purdy: No. When deposited prior to fusion, then the carbon would be incorporated and cause pin- holing ; but when deposited after the quiet stages of fusion has been reached, the glaze would not then be affected. The Chair : But all appear to have the deposit in the earlier stage, but some are able to free themselves? Mr. Purdy : No. Carbon is effective in producing pin- holes only when it becomes incorporated into the glaze. If carbon is deposited on the glaze during the boiling period, it will become incorporated. A glaze which has its initial boiling period early, or which has a very short or mild boiling period, will not be affected as much as the glaze that boils later and consequently harder. The Chair: I will ask Mr. Purdy what method he used to determine over-firing? Mr. Purdy. We determined that by its appearance, i. e. when it began to pass into the second boiling stage. I do not know why it should boil the second time, but it does. Mr. Par melee: I have never had any experience in coke firing, but am surprised to learn that in burning coke you have black smoke. Mr. Purdy : You will have some ; and if you put one speck of carbon into the glaze at the period of troubled fusion, a pinhole will be the consequence, after the glaze is matured. Mr Plusch : Do these particles of carbon burn through and produce pinholes? Mr. Purdy : Not in all cases. Mr. Plusch : I find, in my experience, that carbon settling on the green ware before it is glazed, or settling on the glaze before the glaze is burned, does not produce pin- holing in the firing. Very often particles fall on, after the pieces are sprayed, and we take no account of it. FRITTED GLAZES. 91 Mr. Purdy : Carbon deposited either in or on the glaze before tiring, will as a rule be burnt out before the glaze has reached its first boiling period. Hence you ought not to expect it to produce pin-holing under those condi- tions. Mr. Gray: During the Boston meeting I brought up the question as to the appearance of pinholes in different parts of the kilns and at regular periods. I never found a probable solution until I made some experiments last fall, and I now believe the trouble is wholly in the carbon. Mr. PI it sch : In my experience I found that pin-holing occurred in the bottom of the kilns, especially when we fired quickly, or in other words, water-smoked too rapidly. I have entirely overcome this by water-smoking more slow- ly. Pin-holing on glazed terra cotta is also produced by the burning out of finely divided carbonaceous materials acci- dentally introduced into the body, or by the premature fusion of low-heat glaze materials introduced with the saggars used for grog. In both of these cases cavities are produced on the surface of the ware, under the glaze, before it lias matured, and show up on the burned ware as pinholes. Mr. Purdy : In other words, you have overcome the pin-holing by permitting the carbon that is in the glaze to burn out, and by not permitting any more to deposit on the glaze after fusion has begun. Mr. Hope : We fire with gas and there is no possible chance of having free carbon formed, yet we have pin- holing. Mr. Purdy : You use a fritted glaze? Mr. Hope : Yes. I put it down to the boiling of the glaze. Mr. Goodwin : Did you say, Mr. Hope, that there were no fumes from the gas? Mr. Hope: I meant that there was no uncombined carbon ; that is, no free carbon in the kiln. Mr. Goodwin : My experience goes to prove the con- trary. I have seen smoke from gas, almost as from coal. 92 FRITTED GLAZES. Mr. Hope: I accept the correction. I have seen the same thing myself, but failed to think of it in this connec- tion, before. Mr. Goodwin: I will ask Mr. Purdy if he has had experience in glazes high in A1 2 3 , as to the probability of their spitting out? Mr. Purdy : That is a matter with which I have had no experience. Mr. Goodwin : I have found glazes of that nature in- clined to spit out. Mr. Millar: I make enameled bricks and burn in muffle kilns, and use a raw glaze and am not troubled much with pin-holing. But I find occasionally here and there throughout the kiln, bricks which are covered with pin- holes, while all around them, other bricks will be perfect. I have been trying to gather from Mr. Purdy's paper and remarks, what could be the cause of that, but I have not been able to do so. Mr. Purdy : When I was working in the stoneware business, I found if the biscuit ware was not thoroughly brushed, we would have pinholes. The dust is sometimes common inorganic dirt and oftentimes it is soot or carbon. If that dirt was not brushed off, there would be pinholes as a consequence. But when the ware was thoroughly brushed, it would be free from pin-holing, provided we were careful not to smoke the kiln in the early stages of firing. The Chair: In regard to this question of pin-holing, I had the question put to me some time since by a member, how to prevent it. Not working on Mr. Purdy's theory, but rather on one of my own which I presented to the so- ciety some two or three years ago, I said the way to prevent it was to be careful not to cool the kiln too rapidly. A sud- den chill produces pin-holing, and we have been practically able to eliminate it from our ware, merely by a slow cool- ing. At the Cleveland summer meeting, while visiting a stoneware works in Akron where they were just firing off a kiln, the old burner was covering up every opening in sight. I asked him what he was doing? He said, "Why, if I don't FRITTED GLAZES. 93 get this kiln closed tight as quickly as possible, the ware will be all piiiholed." I attribute it, as I did in that paper, to the liberation of carbonic acid gas. Mr. Purdy: In parts of this paper which were not read, we discuss how the probable reduction of lead in a glaze makes the glaze more viscous, and that the glaze does uot fuse as readily as when free from the carbon, but that by continued soaking in the finishing heat, the normal condition would be re-produced, the carbon burned out, the lead oxydized, and we would have the glaze at its greatest fluidity, and consequently would not have pin-holing. In other words, carbon indirectly stiffens the glaze and pre- vents the healing over of these pinhole defects. Besides this, pin-holing is probably not all due to one cause. There are probably other causes. Mr. Binns: I am very glad we have drawn that confession from Mr. Purdy. I was afraid that he was com- mitting himself to the opinion that this was the one and only cause, whereas everyone, I am sure, knows there are many causes. It is a new thought to me, that particles of carbon deposited on the surface of the glaze can cause pin- holing, and I was about to make other suggestions which Mr. Purdy's last remark renders unnecessary. I think it must be true that when carbon is deposited on the glaze in the early stages of the burn it is not deposited as particles but as a film, and I hardly sec how a film of carbon should break into minute particles and cause pin-holing. The particles could not be carried in through the saggars, and the smoke deposited would be in the nature of a film. We must be cautious in claiming as facts what only appear as phenomena. It is a pity we could not have seen the samples for this piece of work; apart from that, we have here a mass of material which it will take a long time to digest. I will ask what means were used to determine when a glaze was immature or when overtired? What is meant by an "over- fired" glaze? Is it a glaze which has proven too fluid, and has escaped control, or a glaze which has partly volatilized 94 FRITTED GLAZE. and lost its beauty? This must be largely a matter of opin- ion. I do not know of any test by which a man can say a glaze is immature or overtired, and I think we ought to have Mr. Purdy's point of view on that point. Mr. Purdy : In the case of our glazes, as the heat in- creased, the glaze passed gradually from a porous to a vitrified coating, then to that stage of bubbly fusion, then into quiet fusion. As soon as the glossiness of the glaze became dim by overheating, or as soon as the glaze began to show the second boiling, we called it overtired. We did not, however, have any dimness due to overfiring in our cases. Mr. Goodwin : Do you mean that at the second boil- ing you got pin-holing? Mr. Purdy : No ; at that stage it looked like "curdled cream." Mr. Goodwin : My experience has been that you can get pin-holing at the last stage as well as during the earlier stage, but it is of a different type. Mr. Purdy : We explained that in a part of the paper which was not read, viz., that the pinholes due to over- firing seemed to go clear through and were larger, while the others, due to the carbon, were smaller, and apparently only surface phenomena. FRITTED GLAZES 87 as to produce a cloudy or milky appearance. When over- tired, the effect is similar to the curdiness produced in a glaze by addition of raw tin oxide. When fired to the best maturity of the glaze, the opalescent effects produced by supersaturation with Si0 2 are distinctly crystalline. (5) While the writers have advanced the opinion that this opalescent effect is produced by supersaturation and separating out of Si0 2 , they can offer no evidence con- tradictory to the hypothesis that the material thus separ- ated out is a borate. High content of B 2 3 is required, and the opalescence here described not only does not appear unless BoO ; > is proportionately high, but it increases in in- tensity and is produced in wider and wider variation as the B^Oo increases in proportion to Si0 2 . The writers have based their opinion that this phenomenon is due to the supersaturation of the glaze by Si0 2 , on the chemical com- position of milky opals as they occur in nature. ATTACK ON THE BODY. The supposition has been advanced that A1 2 3 is the principal constituent taken from the body by the glaze. The facts that support this assumption are as follows: (1) The glazes lowest in A1 2 3 (0.10—0.20 Eqv.) show this phenomenon to the greatest extent. (2) The glazes lowest in A1 2 3 show the fine-mesh crazing which results from incorporating constituents from the body into the glaze layer that is contiguous to the body, irrespective of oxygen ratio. It is a fact, however, that fine-mesh crazing decreases as the oxygen ratio in- creases, but this is attributed to the approximate satura- tion of the matrix with acid, as is evidenced in the appear- ance of opalescence as early as Group VI. (3) Glazes having a low oxygen ratio, and low A1 2 0. 5 content, decrease in fine-mesh crazing as the intensity of heat increases. (4) Glazes that craze in fine meshes when thick are crazed only in hair lines when thin. P. A F.— 7. 88 FRITTED GLAZES. (5) Increase of originally added AL0 3 decreases fine-mesh crazing under less intense temperatures, but in- creases it as the temperature becomes more intense. (6) Fine-mesh crazing is certainly due to lack of homogeniety in the composition of the upper and lower sur- face of the glaze layer, and this lack of homogeniety is more pronounced, the more viscous the glaze is rendered by the incorporated constituent. (7) Pine-mesh crazing developed at temperatures below Cone 5, can be counteracted by increasing the A1 2 3 content of the glaze. If it is developed at higher heats, it can be counteracted by decreasing the A1 2 3 content, es- pecially when the oxygen ratio is low. PIN-HOLING. Pin-holing and blistering has been proven in these studies to be primarily due to the combustion of carbon that had .been entrapped during the fritting stage of glaze formation. The glazes that come to a quiet fusion earliest are less likely to exhibit this defect. Raw feldspar glazes show this defect more than the raw Cornwall stone glazes, but as was shown by Mr. Coulter 1 the stone glazes are more fusible, and have a longer heat range, and therefore are subjected for a shorter time to the influence of carbon in the fire gases before quiet fusion sets in. As the evidence in this study in the case of fritted glazes agrees with simi- lar phenomena observed in raw lead glazes, there seems to be no doubt that the statements in regard to the cause of pin-holing given above are correct. In conclusion the writers wish to express their appre- ciation of the hearty support and substantial encourage- ment they received from the authorities of the University of Illinois in general, and from Professor C. W. Eolfe, Di- rector of the Ceramic Department, in particular. Had they not granted every facility within their power, the exe- 'Trans. Am. Cer. Soc. Vol. VII, p. 356. n:i ITKI) QOjAZBB. 89 cution of this study in its details would have been impos- sible. We wish also to express our further appreciation of the privilege of offering the results of this research to the American Ceramic Society previous to its issuance as a University Bulletin. DISCUSSION. The Chair: By the presentation of this very able paper, I am impressed at the outset with the growing im- portance of making provision, as soon as our financial con- dition shall warrant it, for the publication of such papers prior to our meetings, so that we may study them, and be better able to discuss them when we come to the meeting. We have realized this necessity during past sessions, but the need is growing more apparent all the time. Still this paper contains so many interesting points that we ought to have a general discussion on it. Mr. Parmelee: I feel, of course, wholly unable to discuss the detail of this work. It certainly is a big sub- ject, and I think brings pleasure to all members to see such a substantial contribution on it. Did I understand you to say, Mr. Fox, that coke was used as a fuel ? Mr. Fox: Yes, sir. Mr. Parmelee: Then how do you attribute the pin- holing to carbon deposited on the glaze? In what form is it deposited on the glaze? Mr. Purdy. Coke certainly contains carbon, which on combustion is driven off. Not all the carbon, however, is oxydised to either carbon monoxide or carbon dioxide; some of it may be carried by the draft as particles of free carbon. Mr. Parmelee: Then these are particles of fuel, car- ried mechanically and deposited mechanically? Mr. Purdy: We know that black smoke contains about two percent, of carbon, and when the coke is freshly thrown on the grate bars there will be an evolution of smoke or black gases, and that is sufficient to cause a de- 90 FRITTED GLAZES. posit of carbon on the glaze, notwithstanding the fact that the saggers are thoroughly sealed. The Chair : You assume that in all of the kiln firings you formed this deposit, but in some were able to free the ware from it more completely than in others? Mr. Purdy : No. When deposited prior to fusion, then the carbon would be incorporated and cause pin- holing ; but when deposited after the quiet stages of fusion has been reached, the glaze would not then be affected. The Chair : But all appear to have the deposit in the earlier stage, but some are able to free themselves? Mr. Purdy : No. Carbon is effective in producing pin- holes only when it becomes incorporated into the glaze. If carbon is deposited on the glaze during the boiling period, it will become incorporated. A glaze which has its initial boiling period early, or which has a very short or mild boiling period, will not be affected as much as the glaze that boils later and consequently harder. The Chair: I will ask Mr. Purdy what method he used to determine over-firing? Mr. Purdy: We determined that by its appearance, i. e. when it began to pass into the second boiling stage. I do not know why it should boil the second time, but it does. Mr. Par melee: I have never had any experience in coke firing, but am surprised to learn that in burning coke you have black smoke. Mr. Purdy: You will have some; and if you put one speck of carbon into the glaze at the period of troubled fusion, a pinhole will be the consequence, after the glaze is matured. Mr Plusch : Do these particles of carbon burn through and produce pinholes? Mr. Purdy : Not in all cases. Mr. Plusch : I find, in my experience, that carbon settling on the green ware before it is glazed, or settling on the glaze before the glaze is burned, does not produce pin- holing in the firing. Very often particles fall on, after the pieces are sprayed, and we take no account of it. FRITTED GLAZES. 91 Mr. Purdy : Carbon deposited either in or on the glaze before firing, will as a rule be burnt out before the glaze has reached its first boiling period. Hence you ought not to expect it to produce pin-holing under those condi- tions. Mr. Gray : During the Boston meeting I brought up the question as to the appearance of pinholes in different parts of the kilns and at regular periods. I never found a probable solution until I made some experiments last fall, and I now believe the trouble is wholly in the carbon. Mr. Plusch : In my experience I found that pin-holing occurred in the bottom of the kilns, especially when we fired quickly, or in other words, water-smoked too rapidly. I have entirely overcome this by water-smoking more slow- ly. Pin-holing on glazed terra cotta is also produced by the burning out of finely divided carbonaceous materials acci- dentally introduced into the body, or by the premature fusion of low-heat glaze materials introduced with the saggars used for grog. In both of these cases cavities are produced on the surface of the ware, under the glaze, before it has matured, and show up on the burned ware as pinholes. Mr. Purdy: In other words, you have overcome the pin-holing by permitting the carbon that is in the glaze to burn out, and by not permitting any more to deposit on the glaze after fusion has begun. Mr. Hope : We fire with gas and there is no possible chance of having free carbon formed, yet we have pin- holing. Mr. Purdy : You use a fritted glaze? Mr. Hope : Yes. I put it down to the boiling of the glaze. Mr. Goodwin: Did you say, Mr. Hope, that there were no fumes from the gas? Mr. Hope: I meant that there was no uncombined carbon ; that is, no free carbon in the kiln. Mr. Goodicin : My experience goes to prove the con- trary. I have seen smoke from gas, almost as from coal. 92 FRITTED GLAZES. Mr. Eope: I accept the correction. I have seen the same thing myself, but failed to think of it in this connec- tion, before. Mr. Goodicin: I will ask Mr. Purdy if he has had experience in glazes high in A1 2 3 , as to the probability of their spitting out? Mr. Purdy : That is a matter with which I have had no experience. Mr. Goodwin : I have found glazes of that nature in- clined to spit out. Mr. Millar: I make enameled bricks and burn in muffle kilns, and use a raw glaze and am not troubled much with pin-holing. But I find occasionally here and there throughout the kiln, bricks which are covered with pin- holes, while all around them, other bricks will be perfect. I have been trying to gather from Mr. Purdy's paper and remarks, what could be the cause of that, but I have not been able to do so. Mr. Purdy: When I was working in the stoneware business, I found if the biscuit ware was not thoroughly brushed, we would have pinholes. The dust is sometimes common inorganic dirt and oftentimes it is soot or carbon. If that dirt was not brushed off, there would be pinholes as a consequence. But when the ware was thoroughly brushed, it would be free from pin-holing, provided we were careful not to smoke the kiln in the early stages of firing. The Chair: In regard to this question of pin-holing, I had the question put to me some time since by a member, how to prevent it. Not working on Mr. Purdy's theory, but rather on one of my own which I presented to the so- ciety some two or three years ago, I said the way to prevent it was to be careful not to cool the kiln too rapidly. A sud- den chill produces pin-holing, and we have been practically able to eliminate it from our ware, merely by a slow cool- ing. At the Cleveland summer meeting, while visiting a stoneware works in Akron where they were just firing off a kiln, the old burner was covering up every opening in sight. I asked him what he was doing? He said, "Why, if I don't FRITTED GLAZES. OS get this kiln closed tight as quickly as possible, the ware will be all piiiholed." I attribute it, as I did in that paper, to the liberation of carbonic acid gas. Mr. Purdy: In parts of this paper which were not read, we discuss how the probable reduction of lead in a glaze makes the glaze more viscous, and that the glaze does not fuse as readily as when free from the carbon, but that by continued soaking in the finishing heat, the normal condition would be re-produced, the carbon burned out, the lead oxydized, and we would have the glaze at its greatest fluidity, and consequently would not have pin-holing. In other words, carbon indirectly stiffens the glaze and pre- vents the healing over of these pinhole defects. Besides this, pin-holing is probably not all due to one cause. There are probably other causes. Mr. Binns: I am very glad we have drawn thai confession from Mr. Purdy. I was afraid that he was com- mitting himself to the opinion that this was the one and only cause, whereas everyone, I ani sure, knows there are many causes. It is a new thought to me, that particles of carbon deposited on the surface of the glaze can cause pin- holing, and I was about to make other suggestions which Mr. Purdy's last remark renders unnecessary. I think it must be true that when carbon is deposited on the glaze in the early stages of the burn it is not deposited as particles but as a film, and I hardly see how a film of carbon should break into minute particles and cause pin-holing. The particles could not be carried in through the saggars, and the smoke deposited would be in the nature of a film. We must be cautious in claiming as facts what only appear as phenomena. It is a pity we could not have seen the samples for this piece of work; apart from that, we have here a mass of material which it will take a long time to digest. T will ask what means were used to determine when a glaze was immature or when overtired? What is meant by an "over- fired" glaze? Is it a glaze which has proven too fluid, and has escaped control, or a glaze which has partly volatilized 94 FRITTED GLAZE. and lost its beauty? This must be largely a matter of opin- ion. I do not know of any test by which a man can say a glaze is immature or overtired, and I think we ought to have Mr. Purdy's point of view on that point. Mr. Purdy : In the case of our glazes, as the heat in- creased, the glaze passed gradually from a porous to a vitrified coating, then to that stage of bubbly fusion, then into quiet fusion. As soon as the glossiness of the glaze became dim by overheating, or as soon as the glaze began to show the second boiling, we called it overtired. We did not, however, have any dimness due to overfiring in our cases. Mr. Goodwin : Do you mean that at the second boil- ing you got pin-holing? Mr. Purdy : No ; at that stage it looked like "curdled cream." Mr. Goodwin : My experience has been that you can get pin-holing at the last stage as well as during the earlier stage, but it is of a different type. Mr. Purdy : We explained that in a part of the paper which was not read, viz., that the pinholes due to over- firing seemed to go clear through and were larger, while the others, due to the carbon, were smaller, and apparently only surface phenomena. ml r ■-^ ■&>.. **&*--pr k««5? 'y ,«. ^v^i, -X: ?'•#*&££ *w- ^Jc* *» j. >-^. *■»**•; * ^-j v i> -: <>£ UNIVERSITY OF ILLINOIS-URBANA L_3 0112 052567101 M*0 * a / 1 A % . V \ £ -A&rm Skm^ > .«- y «»a