Feb., 1919
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
low rather t h a n t o o high. This is what one would naturally expect, since there are a number of precipitations and filtrations, with opportunity in each operation f o r a slight loss of cadmium, b y solution as well as b y mechanical loss. Brasses can be run with a t least t h e same degree of accuracy, since t h e salt mixtures used contained a greater proportion of possible interfering substances t h a t is likely t o be met with in a n ordinary brass. Table I11 shows t h a t excellent agreement can b e obtained. As a final check, and t o guard against t h e possibility of t h e results obtained being due t o a balancing of errors, t h e cadmium sulfate residues from t h e salt mixtures, as well as from t h e brasses, were carefully examined. They were pure white a n d completely soluble in water. T h e combined solutions failed t o give a n y test for interfering substances which could conceivably have been present. T h e procedure for determination of cadmium in brass as finally adopted is as follows: Dissolve IO g. in 75 cc. of nitric acid (I : I). Boil the solution ‘/z hr. and filter off metastannic acid. Dilute to 350 cc. which should bring the concentration of nitric acid to about 3 per cent. Add I O cc. of sulfuric acid (I:I) and electrolyze with a current of 4 amp., using rotating anode, until the copper is nearly all removed, which should take about zl/z hrs. Siphon off the electrolyte, add another IO cc. of sulfuric acid, and evaporate over night nearly to dryness. Dissolve in IOO cc. of water, neutralize with ammonia, add 5 cc. of I : I sulfuric acid, pass hydrogen sulfide into the cold solution 30 min. until the sulfides settle. Add ammonia until zinc begins to come down and pass in hydrogen sulfide 30 min. more. Filter, using water containing ammonium chloride and hydrogen sulfide to wash the precipitate out of the beaker. Boil the precipitate 15 min. with 20 cc. of I :5 sulfuric acid, maintaining the volume by adding water occasionally. Filter and wash, keeping the volume down to 50 cc. Add 5 cc. of strong ammonium hydroxide, pass in hydrogen sulfide 112 hr., adding ammonia until the cadmium begins to come down. Filter and treat the precipitate with a little cold hydrochloric acid (I :I). Neutralize the solution with ammonia and add 5 cc. of I :5 sulfuric acid, keeping the volume down to about 30 cc. Pass in hydrogen sulfide as before and add ammonia to start the precipitation of the cadmium, Filter, dissolve the cadmium sulfide on the filter in a little I :I hydrochloric acid. Collect the solution in a weighed platinum or porcelain dish, add I cc. of sulfuric acid, evaporate to fumes, and add a little concentrated nitric acid. Evaporate carefully to dryness and finally weigh as sulfate. If there is any possibility of the presence of tin, dissolve the cadmium sulfate in 2 5 cc. of water and 5 cc. of I :5 sulfuric acid and reprecipitate with hydrogen sulfide, adding a little ammonium chloride to the solution. Let the sulfide settle, decant through a filter, add a little ammonium polysulfide and ammonium chloride to the precipitate in the beaker, warm for several minutes, and pour through the filter. If there is any turbidity, refilter until the solution comes through clear. Finally dissolve the cadmium sulfide in hydrochloric acid and proceed as before with the evaporation and weighing. In conclusion, it will be well t o call attention t o cert a i n possible sources of error a n d important precautions to be observed, Because of t h e mass of salts present, t h e first precipitation with hydrogen sulfide must be made i n a somewhat larger volume t h a n t h e later precipitations, i. e., under less favorable conditions. Two means are t a k e n t o counteract this. A small p a r t of t h e copper is allowed t o remain in solu-
I I3
tion after electrolyzing and t h e copper sulfide precipitated tends t o a c t as a carrier and bring down t h e cadmium. I n addition t h e acidity is reduced so a s t o permit t h e precipitation of p a r t of t h e zinc, under which condition t h e solubility of cadmium sulfide is greatly lowered. I n later preFipitations t h e volume of solution is so small t h a t t h e cadmium sulfide forms readily, and t h e acidity can be so regulated as t o effect a separation from zinc without a n y great proportion of t h e cadmium remaining in solution. I t may sometimes happen, however, if t h e concentration of hydrogen ion is too small, t h a t some zinc will come down with t h e third precipitate of cadmium sulfide. This can usually be detected b y t h e color of t h e precipitate or by t h e appearance of t h e cadmium sulfate residue, a n d when it occurs a fourth precipitation must be made. Finally, it must be remembered t h a t cadmium sulfide is insoluble in a solution containing z or 3 per cent sulfuric acid only when t h e solution is saturated with hydrogen sulfide. Such solutions, however, are unstable and if allowed t o s t a n d i n a n open beaker, gradually lose their hydrogen sulfide. As a result of t h e removal of t h e sulfide ion t h e solubility product of cadmium sulfide is no longer exceeded, t h e reaction is reversed, and finally all of t h e cadmium goes back into solution. On t h e other hand, if insufficient time is allowed, t h e cadmium sulfide remains in a colloidal condition and runs through t h e filter. As pointed o u t above, t h e presence of ammonium salts t e n d s t o coagulate t h e sulfide, which can be successfully filtered under t h e conditions described. If a n y runs through, a second filtration will suffice t o retain it. I n case i t is desired t o allow t h e precipitate t o stand, t h e solution must be placed in a tightly stoppered vessel or i n communication with a hydrogen sulfide reservoir so t h a t equilibrium m a y be established between t h e solution a n d a definite pressure of t h e gas. It is doubtless through neglect t o consider one or several of t h e conditions dwelt on i n this paper t h a t low results f o r cadmium have been sometimes reported. RESEARCH DEPARTMENT AMERICAN ZINC,LEADAND SMELTING COMPANY a b ST. LOUIS,MISSOURI
THE DETERMINATION OF PHOSPHORUS IN VANADIUM STEELS, FERROVANADIUM, NON-VANADIUM STEELS, AND PIG IRON B y CHAS. MORRISJOHNSON Received July 12, 1918 I-METHOD
FOR
STEEL CONTAINING
2.6
VANADIUU UP
TO
PERCENT
Ibbotson a n d Brearley, i n 1902,~ recommended t h a t phosphorus be precipitated from t h e reduced solution of vanadium, using ferrous sulfate as a reducing agent, claiming thereby t o aid in t h e precipitation of phosphorus when vanadium is present. Cain a n d Tucker, i n 1g13,2 recommended t h a t t h e vanadium present be reduced with ferrous sulfate and sulfurous acid for t h e same purpose. 1 2
“Analysis of Steel Works Materials.” THISJOURNAL, 5 (1913), 647.
T H E J O U R N A L OF I N D U S T R I A L AND ENGINEERING C H E M I S T R Y
,
The method here described gives a successful and convenient method for complete precipitation of phosphorus in t h e presence of highly oxidized vanadium: Place 1.63 g. sample in Ijo CC. beaker. Dissolve in 45 CC. HhTo3 (1.13sp. gr.) over a low flame. When solution is clear and brown fumes have been driven off, add 3 cc. KMn04 sohtion; boil 3 min.; then add 3 cc. FeSOd solution to dissolve precipitate due to the excess permanganate; boil until brown fumes are gone. Avoid large excess of ferrous sulfate. Add 40 to 50 cc. conc. " 0 3 (1.42 sp. gr.); bring to a boil; rinse cover and sides of beaker with least amount of distilled water; then add jo cc. of the water solution of ammonium molybdate. Stir vigorously for a minute or two and let stand over night. Decant through a 7 paper, placing a little paper Pulp in apex of filter, keeping the main precipitate in the beaker. Wash iron out of filter by washing 15 times with the nitric wash, then proceed t o pour the main precipitate onto filter. Transfer all of precipitate as far as possible, with the' aid of the nitric wash. Add about z cc. of wash to the beaker and remove adhering particles with a rubber-tipped glass rod. Use separate beaker for catching washings. Wash 15 times with the dilute nitric wash; then 25 times with K N 0 3 wash, or until the outside of the filter has no sour taste, especially along the double thickness. Place filter containing washed precipitate in 150 CC. beaker; add enough standard NaOH solution to cause the yellow color of the precipitate to disappear on macerating the paper to a pulp with a rubber-tipped stirring rod. Dilute t o about 30 cc. with distilled water; add a drop of phenolphthalein solution, which should cause a deep red coloration, otherwise more standard NaOH solution is needed; then titrate back carefully with standard H N 0 3 solution until the pink just disappears. Subtract total cc. of acid used from total cc. of alkali used. The difference in cc. x 0.01= percentage P. EFFECTO F INCREASING AMOUNTSOF NITRIC ACID (1.42 SP. GR.) ON PHOSPHORUS RECOVERY' 50 cc. "Os 30 cc. HNOa 40 cc. "03 Added Added Added Added P found P found P found P found
TABLEI-SHOWING Vanadium Added Per cent
15 cc. HNOa
0.121 None 0:iii 0.121 0.121 0.119 0.20 0.118 0.121 0.114 0.40 0.117 0.120 0.111 0.60 0.116 0.121 0,110 0.80 o:iis 0.120 0.115 0.110 1.00 0.122 0.121 0.112 0.110 1.40 0.121 0.112 0.119 0.105 1.80 0.120 0.119 0.111 0,099 2.20 0.117 0.117 0.108 2.60 0.093 1 This work was done with 1.63 g . samples of U. S. Bureau of Standards, 0.120 P standard, l o b .
...
REMARKS-The tests with t h e higher vanadium content were more t a r d y in precipitating t h a n t h e lower ones, though all settle'd well after 40 min. standing. Those t o which ~j cc. conc. H N 0 3 were added gave deep orange precipitates, darkening in color about in proportion t o t h e vanadium content, clinging tightly t o t h e stirring rods and beakers. The three highest vanadiums are lower in yield of phosphorus t h a n t h e preceding ones because of t h e impossibility of detaching t h e adhering precipitate f r o m t h e stirring rods. Those with 30 cc. conc. H N 0 3 were a n improvement in speed of precipitation, color of precipitate, and cleaning from beakers and rods. Those with 40 cc. and 5 0 cc. showed t h e best colored precipitates, i. e., nearest t h e normal yellow phosphomolybdate color and gave water-white washings, from which no recoveries of phosphorus were made. T h e 5 0 cc. ones were a little prompter t h a n t h e 40 cc. though both sets were
.
Vol.
11,
No.
2
a vast improvement over t h e ~j cc. and 30 cc., and both cleaned very easily from beakers and rods and gave good results. The precipitate from t h e 15 cc. cone. " 0 3 in particular seemed t o be too finely divided and passed through t h e pores of t h e filter, necessitating reprecipitation of the washings in all those of higher vanadium content. The vanadium was added as nitrate from a n acid made as follows: Fuse 4.5 g. T'Zoj (56.14 Per cent IT) with 1 0 g. NazC03 in a Platinum crucible for I O t o 1 5 min. Let cool; place in a 400 cc. beaker; add I O cc. water; then H N O s ( 1 . 2 0 sp. gr.) until effervescence ceases and 2 5 cc. in excess. Rinse off crucible thoroughly with distilled water; cool solution to temperature; transfer to 5oo cc. flask; dilute t o t h e mark with water; stopper, and shake t o mix thoroughly. I cc. = o . o o j o j g. V. I per cent V in r.63 g. = 3.2j The vanadiunl was added after the tests were weighed and before t h e 4 5 CC. "03 (1.13 Sp. gr.) were added. The vanadium was, in each instance, reduced from yellow t o green due t o t h e action of t h e reducing gases liberated b y t h e reaction between t h e acid and t h e steel. ~h~ 4o to 5o cc, cone, nitric added just before removing from t h e fire caused t h e yellow vanadic color t o return. SoLUT1oNS REQUIRED-NitriC Acid f o r Titratirtg35.4 cc. (I*'k2 sp* gr*) to 4Ooo cc* with distilled water. Stock Solution of Sodium Hydroxide-Dissolve 150 g. NaOH (c. P. sticks) and I g. Ba(OH)* in 1000 cc. water. Stir well and allow t o s t a n d 24 hrs. Filter off t h e clear solution (or decant if preferable) and dilute with a n equal volume of distilled water. P u t into a 2-liter bottle and close with a rubber stopper. Sodium Hydroxide Solution f o r Titrating-Dilute 566 cc. stock solution t o 8000 cc. with distilled water. Ferrous Sulfate Solutiort-Dissolve j5 g. steel (low in phosphorus and sulfur) in 1000 cc. beaker in 7 2 0 cc. HeS04 ( I : 3). Add a little water occasionally t o prevent salting out, and heat gently, b u t do not boil. Filter, cool, and dilute t o 1000 cc. Concentrated Potassium Permanganate Solution-Dissolve 5 0 g. KMn04 in 1000 cc. water. 8 g. KNOB in 8000 cc. Nitrate Wash-Dissolve water. Acid Wash-Dilute 102.4 cc. " 0 3 (1.42 sp. gr.) t o 8000 cc. with water. Faintly Ammoniacal Water Solutior, o j Ammonium Molybdale-Into each of four goo cc. beakers weigh 55 g. ammonium molybdate and 50 g. ammonium nitrate, and add 4 0 cc. ammonium hydroxide ( 0 . g j sp. gr.). Dilute each t o 7 0 0 cc. with water. Heat for about 30 min., stirring once in a while until'all salts are in solution. Combine contents of t h e four beakers by pouring into a large bottle; then dilute t o 4000 cc. with water. Let s t a n d over night. Filter t h e insoluble material through double 1 5 cm. papers. Do not wash. The clear solution t h u s obtained should remain clear indefinitely. Do not filter out t h e in-
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
Feb., 1919
Add about 3 cc. of the KMn04 solution, boiling for 3 min., then just enough FeS04 solution (about 3 cc.) to dissolve the oxides of manganese. Boil off brown fumes. Add 15 cc. HNOa (1.42 sp. gr.). Rinse cover and sides of beaker with least amount of water, add 50 cc. ammonium molybdate solution, and stir briskly until phosphorus precipitate is completely Let stand 1/2 hr. or less. formed (about 2 min. stirring). Filter through 7 cm. paper; wash 15 times with dilute HNOs wash; then with nitrate wash until the outside foldof the filter has no sour taste. In case of high phosphorus samples this may mean as much as 35 or 40 times. (The washings in such a case were kept separate from the main filtrate, concentrated to IO cc., acidified with 15 cc. conc. “ 0 3 , and 50 cc. of the ammonium molybdate solution added. It was then allowed to stand over night. There were no recoveries, showing the insolubility of the precipitate in the wash.)
soluble material until the total mixture from the four beakers has stood over night. 21-METHOD STET:LS
ADDED
FOR AND
28
PHOSPHORUS
PIG
IRONS
IN
TO
S.
U.
WHICH
STAKDARD
HAVE
BEEN
PER CENT VANADIUM A N D I N FERRO-
VANADIUM
CONTAINING
56.7
PER
CENT
VANADIUX
PROCISDURE-weigh Out 0.5 g. Steel and 0.5 g. VzOs (56.14 per cent V). Transfer to 250 cc. porcelain dish. Digest with a mixture of 30 cc. HCl (1.20 sp. gr.) and 30 cc. ” 0 3 (1.42 sp. gr.) for about I hr.; then rinse off cover, add IOO cc. H N 0 3 (1.42 sp. gr.), and take to dryness. Bake 5 min. a t 750’ C. in an electric muffle furnace. Dissolve the oxides in 35 cc. conc. HCl, evaporate to IO cc.; add 50 cc. HNOa (1.42sp. gr.), take to I O cc. volume, then a d d IO cc. more strong nitric and heat awhile with cover glass .on. Filter through a platinum Gooch crucible with a thin pad of acid-washed asbestos, using suction. Wash 15 times, using the following wash: 200 cc. HNOa (1.42 sp. gr.), IOO cc. water, and 2 0 g. Fe(N03)a. (Ferric nitrate made by dissolving 5 g. low phosphorus, melting bar steel in 50 cc. HC1 (I:I) and (1.42 sp. gr.) taking to sprupiness twice with 50 cc. “ 0 8 each time). Concentrate the filtrate from the separated V205 to I O cc. in 1 5 0 cc. beaker and filter out the second crop of vanadium “rust” as before. A third concentration t o IO cc. should show no “rust” (V205). To the third concentration add 40 cc. “ 0 3 (1.42 sp. gr.), and bring to a boil; rinse cover and sides with water and precipitate with 50 cc. of the faintly ammoniacal ammonium molybdate solution. Stir vigorously for about 2 min. Let stand one hour; filter and wash as described for steels containing up to 2.6 per cent V. Titrate with alkali and acid in the usual manner.
TABLEIV-COMPARING RESULTS OBTAINEDWHEN USING FAINTLY AMMONIACAL WATERSOLUTION WITH THOSE OBTAINEDBY USE OF A NITRIC ACID SOLUTION
AMMONIUM MOLYBDATE* Slightly Ammoniacal Nitric Acid Water Solution of Ammonium Solution of Molybdate Molybdate Per cent SAMPLE Per cent P found No. P found 0,054 0.056 9 ............................... 34 ............................. 0.023 0.023 0.019 0.018 7132 0.056 0.054 3622. 0.013 0.014 1 . . ............................ 0.058 0.058 0.042 0.044 0.052 0.052 5. . . . . . . . . . . 0.052 0.055 70.055.. .... 0.042 0.044 80.042 ......................... 0.730 0.736 0.006 0.004 3641 .......................... 0,050 0.049 0.025 0.025 34 ............................. 0.056 0.058 39 ............................. 0.096 Pig Iron Std. A . . 0.097
........................... ..........
Standard Iron D, 6a VzOs . . . . . . . . . . . . . Iron D, 6b VzOr 10b VzOs.. . . . . . . . . . . . 10b VzOa No. 20 Std. VzOi No. 20 Std. VeOs . . . . . . . . . . . .
Added Ma. 500 500 500 600 500 600
P Found Per cent 0.522 0.531 0.124 0.115 0.033 0.033
............. .., . . . ................... ............
I n t h e above procedure an average of
................
Given b y U. S. Govt. 0.526 0.531 0.120 0.120 0.031 0.031
0.007
...........
TABLE V-SHOWING RESULTSOF TESTSON U. S. GOVT.STANDARDS U. S. Bureau P Found Per cent SAMPLE Pig Iron, D, third s e t . . . . . . . . . . . . 0.610 Pig Iron, B 6 a . . . . . . . . . . . . . . . . . . 0.104 Pig Iron, D 6 a . . 0.540 Steel, 10b 0.119
TABLE1I’-RESULTS OBTAINEDUSING THISMETHODON U. S. STANDARDS VZOI Per cent P No. 1 2 3 4 5 6
OF
per cent
P was found in the Vz05 added. TABL& 111-PHOSPHORUS IN FERROVANADIUM, 56.7 PER CENT VANADIUM
................ .......................
P Value Per cent 0,602 0.105 0.545 0.120
This method is t h e same as the one published by t h e author in Iron A g e , April 5 , 1917, except t h a t i t was afterwards found t h a t more rapid precipitation of t h e yellow precipitate was secured by t h e addition of I ; cc. of concentrated nitric acid (or 30 cc. I : I acid is safer t o handle) t o t h e boiling solution of t h e sample just before removing from t h e fire t o a d d t h e molybdate solution; a n d 1.13 sp. gr. HNOa is used instead of 1.20 as t h e steels in general dissolve more rapidly in t h e 1.13 acid. ADVANTAGES O F T H E FOREGOING
M E T H O D S F O R ORDI-
N A R Y R O U T I N E W O R K ON P L A I N A N D O T H E R S T E E L S Per cent P 0.414 0.418 0.397 0.401
... III--RII3THOD
Per cent P 0.240 0.232 0.235 0.228 0.234 0.260 0.220
Per cent P 0.053 0.108
... ... ... .
.
I
...
Per cent P 0.185 0.183
... ... ... ... ...
F O R D E T E R M I N I N G P H O S P H O R U S I N NON-
VANADIUM S T E E L S AND P I G I R O N USING A PAINTLY AISMONIACAL W A T E R S O L U T I O N OB A M M O N I U M & 0! LY I B D AT E
PROCICDURE-weigh I .63 g. sample into 150 cc. beaker. (1.13 sp. gr.) over low flame. (In Dissolvc: in 45 cc. “ 0 s case of pig iron and certain chrome steels, when all metal is in solution filter off carbon residue through 7 cm. paper; wash 15 times with dilute “ 0 3 , wash, catching filtrate in 1 5 0 cc. beaker. Concentrate filtrate to original volume.)
‘
I-The simplicity of t h e preparation of t h e faintly ammoniacal water solution of ammonium molybdate as compared with the elaborate method of preparing the old acid molybdate solution in nitric acid. 2-The handling of t h e acid molybdate is disagreeable, hard on t h e clothing and fingers of t h e operator and on table tops. The slightly ammoniacal water solution is harmless in these particulars. 3-As stated, t h e slightly zmmoniacal water solu tion keeps as clear as distilled water after it has been properly filtered; and its precipitating power is cons t a n t , whereas t h e old acid solution soon becomes turbid and therefore its precipitating power is continuously lessened. 1
See Iron A g e , April 5 , 1917.
T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
I 16
4-The interference of vanadium u p t o at least per cent V in t h e determination of phosphorus is prevented in a simple way and t h e t r u e phosphorus is obtained from ferrovanadium containing t h e very high vanadium content of 56.7 per cent. ACKNOWLEDGMENT
Credit is due Mr. F. D. Hawkins for his careful analytical work when making t h e trial analyses required b y t h e author t o prove t h e foregoing. RESEARCH DEPARTMENT AMMERICA PITTSBURGH, PA.
CRUCIBLE STEEL COXPANY O F
SOME MIX-CRYSTALS OF CALCIUM FERRITE AND ALUMINATE By EDWARD D. CAMPBELL Received June 25, 1918
1 2
8
4
THISJOURNAL,
5 (1913), 627. I b i d . , 6 (1914), 706. I b i d . , 7 (1915), 835 A m . J . Sci., 28 (1909), 293.
II?
So.
2
showed t h e existence of a definite calcium aluminate of t h e empirical formula 5Ca0.3A1203 and gave the melting point in a pure state as 1386’ C., which was lowered b y the addition of a small amount of CaO t o 1382’. I n a later publication by Rankin and Wright1 t h e melting point of pure 5Ca0.3Al2O3 is given as 1455’ C., which is first lowered b y t h e addition of lime t o 13g5’, and then gradually raised as t h e concentration of t h e solution increases until, when the concentration reaches t h a t of t h e empirical formula 3CaO.AlzO3, i t has reached 1535’. I n formulating a hypothesis of t h e formation of clinker b y recrystallization of t h e basic silicate, assumption was made t h a t t h e aluminate which acted as a solvent was t h e one having t h e empirical formula gCa0.3A1203 and also, although there was no experimental evidence t o support i t , t h a t since i t had been shown t h a t tricalcic silicate could be crystallized from calcium ferrite as easily as it could from calcium aluminate, probably a calcium ferrite of the empirical formula j C a 0 . 3Fe203 might exist and function exactly like t h a t of t h e jCa0.3A1203. I n a “Preliminary Report on the System LimeFerric Oxide,”2 Sosman and Merwin seem t o have demonstrated pretty clearly t h e probable existence of only two calcium ferrites having t h e formulas, t h e dicalcic ferrite, zCaO.Fe203, and t h e monocalcic ferrite, CaO.FezO3. I n t h e present investigation, t h e experimental work of which was carried out by W. C. Kwong, B.S.E., there were two points on which i t was hoped t o secure some additional information: t h e first was t o redetermine t h e existence or non-existence of a calcium ferrite of t h e empirical formula gCa0.3Fe203; and t h e second, to obtain some experimental evidence as t o what is formed when a mixture of calcium oxide, ferric oxide, and alumina is melted and then cooled at a rate slow enough t o permit t h e material t o be nearly in equilibrium during t h e entire cooling period.
2.6
About five years ago in an article on “The Constitution of Portland Cement Clinker,”l some experimental evidence was given t o show t h a t t h e formation of Portland cement clinker comprises essentially a series of solutions of certain silicates and lime in a magma which is fluid a t the clinkering temperature and serves as a solution in which t h e recrystallization of t h e basic silicates characteristic of Portland cement takes place. As early as 1897 Tornebohm had applied t h e t e r m celite t o the magma in which t h e crystalline portions of t h e clinker were imbedded, and states t h a t this fusible magma promoted t h e crystallization of t h e alite, as t h e basic silicates were termed. Prior t o 1913 no satisfactory method for separating celite from alite had been described, b u t in t h e article above referred t o a method was given for separating t h e celite present in clinker a t different temperatures, and i t was further shown t h a t t h e celite of ordinary Portland cement clinker consists essentially of a mixture of calcium aluminate and ferrite in which both ortho-calcium silicate and lime are soluble, t h e solubility of these being a function of t h e temperature of the fluid solution. I n an article on “Synthetic Celite and Large Crystals of Tricalcic Silicate,”2 i t was shown t h a t if pure silica, alumina, and calcium oxide were mixed in t h e same molecular proportions as those in which they had been shown t o exist in t h e celite previously obtained from Portland cement clinker and t h e liquid mixture was very slowly cooled, t h e tricalcic silicate crystallized out in large thin plates, thus demonstrating t h a t it was t h e aluminate which constitutes t h e solvent in which t h e crystals had formed. I n a third paper on “The Function of Ferric Oxide in t h e Formation of Portland Cement Clinker,”3 i t was shown t h a t if a mixture of pure silica, ferric oxide, and calcium oxide of t h e same molecular proportions as those found in t h e celite drawn off from t h e Portland cement clinker is melted and slowly cooled, large crystals of tricalcic silicate will separate, thus demonstrating t h a t calcium ferrite may act as a solvent for both calcium silicate and lime in a manner almost exactly similar t o t h a t of t h e calcium aluminate. As early as 1909, Shepherd, Rankin, a n d Wright4
Vol.
MATERIALS U S E D
-
Kahlbaum’s purest analyzed calcium carbonate and ferric oxide were used as t h e source of CaO and Fe203, t h e &Os being t h e purest available ignited oxide, giving b y analysis 99.60 per cent AI&. These materials were thoroughly dried at about 105 ’ before being accurately weighed in molecular proportions prior t o mixing and melting. The mixture t o be melted was placed in a flatbottom cylindrical platinum crucible 48 mm. in diameter and 5 5 mm. deep, having a capacity of about IOO cc. Enough material was used for each experiment t o give from 7 j t o 92 g. of t h e final melt. It was usually necessary t o heat t h e mixture in two lots over night t o a temperature of about 1100’ in order t o shrink it sufficiently t o enable the entire charge t o be p u t in the crucible before melting. During t h e melting and cooling t h e crucible was covered, first with the platinum cover on which was placed enough ignited magnesium oxide t o fill t h e concavity of the cover, and then with a pure magnesium oxide crucible cover. This arrange1 2
Am. J . Sci., 39 (1915), 1. J . Wash. A c a d . Scz., 6 (1916), N o . 15.
