228
T H E JOURiVAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
comparatively a much smaller stem and finer graduations. They should be calibrated with mercury before using, in which case a long drawn capillary tube is useful in filling and emptying. It is also useful in washing out precipitates after a potash determination. Linings for the centrifuge tube shields can be made from large corks t o prevent breakage of the potash tubes. NOTES ON METHOD
The solution of t h e sample should contain approximately 1 g. of the K 2 0 per 100 cc. Before diluting to the mark, i t should be rendered alkaline with sodium hydroxide and acidified with glacial acetic acid, using phenolphthalein as indicator. If i t contains insoluble matter, filter through a dry paper, and centrifuge 5 cc. of the clear filtrate. Samples containing ammonium salts should be weighed and ignited before bringing into solution. The stem of the tube should be full of water before adding the nitrite solution. The temperature at which the determination is made is 22' C. or above. The precipitate reading of t h e sample should not be over five small divisions above or ten below t h a t of the standard, which should be about 10.5. A 4-tube head centrifuge allows three samples t o be run with the one standard. A P P L I C A T I O N S O F METHOD
The ease with which samples are prepared for analysis, the rapidity of obtaining reliable results, and the minimum costs of chemicals per sample make t h e method ideal. Samples that used to take the most time are now determined quickest. For instance, the K 2 0 in molasses is determined as follows: Transfer 26 g. more or less t o a 100-cc. volumetric flask by aid of hot water, render alkaline with sodium hydroxide, acidify with acetic acid, cool, fill to the mark with water, and mix. Centrifuge 5 cc. and calculate the per cent K20. This requires 12 min. in all. Distillery and sugar factory wastes can be determined in from 4 t o 8 min. Altogether over 1700 determinations have been made in this laboratory by the method, about 60 of them being on 50-ton car shipments of crude potash. The results obtained on twelve car samples by the centrifugal method, in comparison with those of a wellknown public analyst and in the case of disputed samples those of an umpire chemist in New York, are given below: Lab. No. 8.5 88 89 97 101 102 103 104 111 I12 P 1 P 2
Centrifugal Method Per cent KzO Moisture 32.80 35.62 35.54 0:&3 28.75 30.87 1.45 1.30 30.57 29.55 1.59 30.02 1.36 36.73 1.6s 0.73 40.78 41.51 0.83 0.82 41.39
..
..
Puhlic Analyst Analysis Umpire Analysis Per cent Per cent K1O Moisture KaO 32.68 5.63 3.95 34.12 4.11 3316.5 33.74 26.36 1.11 ... 1.47 30.92 1 ..58 29.76 2.20 28.10 1.53 ?9:38 28.38 1.32 37.28 ... 1.01 39.44 ... 41.30 40.80
...
...
The results given by the centrifugal method were those obtained the first time run, not averages of two or more determinations, and not over 25 min. were required for any result. In each case they were just one of several samples run during the day.
V O ~ 13, . NO. 3
RAPID IODOMETRlEUC METHOD FOR DETERMINATION OF CHROMIUM IN CHROMITE By Ernest Little and Joseph Costa RUTGERSCOLLEGE, N E W BRUNSWICK, N J Received October 2, 1920
The determination of chromium after oxidation t o sodium chromate immediately suggests the rapid accurate method of iodometry. Because of t h e analogous action of the dichromate and ferric ions on the iodide ion, however, the analysis of chromite by this method presents a problem, and our purpose here is t o present a method whereby a n iodometric determination of chromium may be effected with its usual rapidity and accuracy without time-consuming, intermediate procedures for the elimination of iron. Practically all the methods in use for the analysis of chromite prescribe t h a t the Ore be fused with sodium peroxide, or sodium peroxide and sodium carbonate, in a spun iron crucible. The methods differ after t h e extraction of the melt and are of two classes: first, those in which the iron is removed as ferric hydroxide by filtration, and the filtrate of sodium chromate is analyzed by any of the usual methods, including a n iodometric method;' and second, those in which the chromium is determined in the presence of iron in a n acid solution. The objection t o the methods of the first class is that the filtration of a solution containing in suspension a voluminous precipitate of ferric hydroxide is a tedious operation and quite likely not t o give quantitative results on the first filtration; three filtrations and subsequent reprecipitation are very often necessary.2 I n methods of the second class, the extract is acidified with either hydrochloric or sulfuric acid, a weighed excess of Mohr's salt or ferrous sulfate is added, and the excess titrated with standard permanganate or dichromate. The seortcomings or inconveniences of these methods are well known. I n the case of the potassium permanganate, when the titration is carried out in a moderately small volume, the end-point is obscured by the bright blue-green color of the chromic ion; when larger volumes are used in order t o overcome the above-mentioned difficulty, blank tests on the water are necessary. Furthermore, permanganate is rather unstable in solution, and frequent restandardizations are necessary. When dichromate is used, an outside indicator with its inconveniences is imperative. The use of a n outside indicator is cspecially difficult in this analysis, owing t o t h e high concentration of chromic ion, the color of which makes the end-point more difficult t o determine. An iodometric method has not been considered possible here on account of the presence of the ferric ion. ,4 method has been outlined in which the interference of the ferric ion is claimed t o have been removed by the addition of a solution of phosphoric acid, in the presence of which iron forms a very slightly ionized ferric acid p h ~ s p h a t e . ~This method, however, has not been tried in the presence of large excesses of iron, 1 2
3
Brunn, Z . anal. Chem., 52, 401. Schorlemmer, Collegium, 1918, 145. 0. I,. Barnebey, J . A m . Chem S O L ,39 (19171, 604.
Mar., 1921
T H E J O U R N A L O F I N D U S T R I A L AiVD E N G I N E E R I N G C H E M I S T R Y Iodometric Method-
No. 1
2 3 4
5 6 7 8 9 10 11
12 13
Wt. Sample
0.9600 0.9600 0.9600 0.9600 0.9600 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100 0.4100
-NazSzO-
cc. Used
Normality
114.46 114.11 115.40 92.96 92.63 52.45 52.31 45.17 45.40 52.50 52.44 75.12 75.00 68.31 84.42 84.44 83.53 91.20 91.37 96.80 96.47 98.05 97.81
0.1235 0.1235 0.1235 0.1235 0.1235 0.1235 0.1235 0.1235 0.1235 0.1235 0.1235 0.1030 0.1030 0.1030 0.1030 0.1030 0.1030 0.1030 0.1030 0.1023 0.1023 0.1023 0.1023
Grams CrzOa
0.3581 0.3569 0.3610 0.2909 0.2898 0.1641 0.1637 0.1413 0.1409 0.1643 0.1641 0.1960 0.1957 0.1782 0.2203 0.2206 0.2180 0.2380 0.2384 0.2509 0.2500 0.2541 0.2535
37.30 37.19 37.61 30.30 30.19 40.02 39.91 34.47 34.37 40.05 40.01 47.79 47,72 43.47 53.72 53.80 53.16 58.04 58.16 61.09 60.98 61.95 61.82
and the data given are hardly sufficient t o warrant its acceptance a t this time. F E R R I C - F L U O R I D E COMPLEX
It is known t h a t when the fluoride ion is added t o a solution containing the ferric ion, a very slightly ionized, but fairly soluble complex is formed, probably FeFG=. The very low ionization of this complex is well demonstrated by the fact t h a t such a slightly soluble substance a s ferric hydroxide will dissolve quite readily in t h e presence of the fluoride ion. Also if potassium iodide and starch paste are added t o a solution containing the ferric fluoride, no blue color is produced. T h e theoretical considerations in the reaction of the Fe+++, F-, and I- ions will not be entered into here, as they have already been fully presented.' The ferric-fluoride complex is, however, broken up by large excesses of either acid or alkali, but is stable in acid concentrations such a s are suitable for the chromite analysis. I n the case of the analyses here outlined, twice the ,amount of acid prescribed had t o be added before trivalent iron from the complex reacted with the iodide. If less acid is used, the oxidizing potential of t h e dichromate is too low, and the titration is greatly retarded. A false end-point may appear, the blue color returning after a few seconds. This leads t o no inaccuracy, however; the solution may be allowed t o stand a few minutes longer, or 2 t o 3 cc. more acid added, and the titration completed. The ferric-fluoride complex is, of course, least ionized in the presence of a n excess of the fluoride ion, but a large excess is not necessary. From 1 t o 4 g. excess ammonium fluoride were used, and the results were identical in each case. While spot tests with potassium ferrocyanide were used t o insure the absence of the ferric ion, this is hardly necessary. Ammonium fluoride was used in most cases, but the results with potassium fluoride, as would be expected, were found t o be identical. Hydrofluoric acid was used in a few instances, and found satisfactory but inconvenient. When very large amounts of iron are present, t h e ferric fluoride is precipitated, giving a n opaque whii e 1
Little and Hults, THISJ O L R S A I , ,
12 (19201, 270.
.
KMnOs
Mohr's Salt Reduction Method E uiv 01N Cc. %.I, 2 Ox. Reagt. Mohrs Used Grsms Salt Sqlt cc. Crz03 6.0072 153.19 13.09 0.3549 3.31 0.3595 5.7863 145.25 5.8816 145.99 7.32 0.3614 4.7770 121.82 7.62 0.2893 4.9339 125.82 11.94 0.2985 2.6000 66.30 1.79 0.1634
KMn04
2.3913
60.98
5.02
0.1417
34.56
KMnOl
2.7300
69.62 86.27 81.85 77.92 93.42 91.76 94.42 100.00 100.00
5.10 8.77 4.66 7.58 6.59 5.89 8.68 6.13 6.21 1.60 1.46 13.88
0.1634 0.1963 0.1955 0.1782 0.2199 0.2175 0.2172 0.2378 0.2376 0.2492 0.2496 0,2546
39.86 47 .E8 47.70 43.46 53.65 53.05 52.97 58.00 57.95 60.80 60.86 62.09
r-
Per cent CrzOa
229
Reagent Used KMnOa
KMnOa KMnOa KzCrzOz
KMn0.r KMnOr KzCrnOz KzCrlOz
KzCr20z KzCrzOn
KzCrvOz
Weight Mohr's
"3 "2;: 3.0554 3.6632
:;g:i 3.9214 3.9217 3.9214 3.9214 4.4850
100.00
100.00 114.37
Per cent
CraOa 36.97
37.45 37.65 30.13 30.05 39.86
character t o the solution. The formation of this precipitate interferes in no way, but rather aids t h e endpoint, which goes sharply from the usual deep blue t o a n opaque white, instead of t o the transparent, bluish green chromic solution. S U G G E S T E D METHOD
Four-tenths of a gram of chromite were mixed thoroughly in a 25-cc. iron crucible with 3 g. of sodium peroxide, and covered with 2 g. more. This mixture was heated a t a low heat for about 5 min., and then fused for 15 min. a t a higher temperature. The crucible was allowed t o cool completely, and placed in a beaker containing 150 cc. water. After ebullition had ceased, the crucible was thoroughly washed, and removed. A half gram of peroxide was then added t o the solution t o insure complete oxidation, and the excess peroxide driven off by boiling. The solution was cooled and hydrochloric acid added until the ferric hydroxide dissolved, when 5 cc. excess concentrated acid were added for each 100 cc. of solution. Ammonium fluoride was now added until the solution no longer reacted for ferric ion with potassium ferrocyanide on a spot plate, and 1 g. in excess used. Three grams of potassium iodide were then added, the solution allowed t o stand 3 min. and titrated against a standard thiosulfate solution, using starch solution a s a n indicator. The crucibles used in the fusion weighed about 25 g. and lost in each fusion a n average of 0.5 g. in weight. It is evident, therefore, t h a t the determinations were carried out in the presence of large excesses of iron. The above table shows the results obtained by this method in the analysis of thirteen typical chromite ores, compared with the results of analyses made by reducing the chromate with Mohr's salt and titrating the excess reducing agent with either standard permanganate or potassium dichromate. CONCLUSIONS
1--An iodomctric titration for dichromic acid may be easily carried out in the presence of large excesses of iron by means of the formation of the ferric-fluoride complex.
T H E J O U R N A L OF INDUSTRIAL A N D ENGINEERING C H E M I S T R Y
230
2-This method has been found t o be rapid, accurate, and highly satisfactory with chrome iron ores, a n d should adapt itself for use in control work, in t h e analysis of such ores. A RAPID VOLUMETRIC METHOD FOR DETERMINING
ALCOHOL By Arthur Lachman I43 FOURTEENTH AVENUE,SANFRANCISCO, CALIFORNIA Received October 25, 1920
The accurate estimation of alcohol by means of t h e density of water-alcohol mixtures requires great care, especially in regard t o temperature control. The tables of the Bureau of Standards are carried out t o five figures, with alcohol values in terms of hundredths of per cents; but such accuracy requires a temperature Atmospheric changes may adjustment of about 0.01 introduce fluctuations of more t h a n 0.15 per cent, involving reduction of weights t o vacuum, The tables of t h e Bureau have been compiled with all possible care, as have those of the German NormalAichulzgs-Amt; yet these two tables differ in parts by as much as 0.10 per cent, or more t h a n ten times t h e limit of accuracy postulated in the tables themselves. The method herein briefly described gives a high degree of accuracy, and is exceedingly rapid. It is based on the determination of the critical point of a n equilibrium of the third order. A fixed weight of aniline (25.00 9.) is pipetted into a definite volume (50.00 cc.) of the alcohol-water mixture whose strength is t o be determined. If t h e aniline does not dissolve completely, some convenient fixed volume, such as 25.00 cc. of strong alcohol of known strength, is added until solution occurs. Water is run into the clear solution from a buret until a permanent turbidity occurs. The end-point is exceedingly sharp; a single drop of water converts the perfectly clear, or slightly opalescent, liquid into a milky suspension t h a t cannot possibly be mistaken. If t h e end-point is overshot, the vessel is slightly warmed in the hand, and a drop or two of water added again. When the end-point is reached, the temperature of the mixture is noted t o 0.1' C. The operation is then complete, requiring merely 2 or 3 min. From the known volume of sample, of added alcohol, and of added water, the percentage of alcohol in the sample can be calculated. The following tabulation shows the character of the results obtained :
'.
-Determined byDensity Titration
20.10 22.94
20.04 20.02 20 04 22.91 22.95 22.91
VOLUMEPER CENT
-Determined by.Density Titration
32.54
32.57 32.58
50.63
50.60 50.59 50.62 50.66 96.04 96.07 96.08 96.07
96.03
25.15
25.12 25.12
99.84
99.86 99.83
The calculation depends upon the experimentally established fact t h a t the total volume of solvent (alcohol plus water) is a nearly strictly linear function
Vol. 13, No. 3
of the volume of contained alcohol. If a number of points on the curve are determined, the intermediate values may be obtained by graphic interpolation without serious error. I n the following condensed table are given the total solvent volume and the corresponding alcohol volumes. By deducting the known volume of added alcohol, we find the volume of alcohol in the sample: RELATIONBETWEEN TOTAL VOLUME OF SOLVENT A N D VOLUMEOF CONTAINED ALCOHOL (For 25.00 G. Aniline at 15 '6 C.) Total Total Solvent Alcohol Solvent Alcohol
50 60
70 80 90
..
22.28 25.38 28.40 31.43 34.42
...
100 110 120 130 140 146
37.41 40.30 43.05 45.80 48.50 50.00
Several corrections must be made before the final result is obtained. Tables for these have been calculated, but owing t o lack of space they cannot be given here, and a brief enumeration must suffice. The total solvent volume given above holds only for the normal alcohol temperature of 15.6'. T h e temperature coefficient happens t o be almost exactly 1 per cent of the total solvent volume per degree, for a range of 2' or 3' in both directions. The temperature during titration may be kept close t o the normal by immersing the flask occasionally in cold water. T h e temperature of t h e sample and of t h e added alcohol may be kept between 14' and 17" without appreciable effect upon t h e results; larger deviations require correction. T h e volume of water added from the buret may require correction if the room temperature differs by more t h a n 5' from normal. T h e most troublesome correction is caused by t h e contraction of volume which has previously taken place in the sample, It may be ascertained by making a n approximation value, then computing the contraction, and recalculating. Tables have been worked out for this correction, but cannot be given here. The above method has been used in commercial control work over a period of nearly 10 yrs. Where routine work is done over a comparatively limited range of strength, i t is possible t o condense all calculations into one set of tables, and t o obtain percentages of alcohol directly from the buret readings.
The anti-trust suit of the Federal Government against the Eastman Kodak Company was settled February 1, 1921, with the filing of a decree in the U. S. District Court in Buffalo, requiring the company to dispose of approximately $4,000,000 of its assets, which total $90,000,000. Among other things the decree orders the sale of the Premo factory and the CenturyFolmer and Schwing factory in Rochester and the Aristo plant plants which were acquired from competiin Jamestown, N. Y., tors, and not developed as part of the industry built up by George Eastman. It is stated that the decree will result in no substantial disruption of the organization, since a radical move for dissolution has practically been stopped, and the company will carry on its activities with renewed confidence. Notice of appeal was withdrawn after a conference of the company's representa\ tive with the Attorney General in Washington.
