uu
T H E JOL-Ri\-AL
Aug., 1913
OF I i Y D U S T R I A L A-1-D E S G I N E E R I N G CHE-IJISTRY
Per cent Laboratory of boras number Sone 79188 Borated skin soap.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 79189 Antiseptic toilet soap, borated. . . . . . . . . . . . . . . . . . 1 96 . . . . . . . . . . . None 79190 Best borax soap.. . . . . . . . . . . . . . . . . . . . . . . 0.1: 79191 Borax soap.. . . . . . . . . . . . . . . . 0 10 79192 Finest quality borax soap.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 85 79193 10.8s 79194 Real boras soap . . . . . . . . . . . . . . . . . . . . . . . . . 6 44 79195 XVhite boras soap.. . . . . . . . . . . . . . . . . . . . . . . . . . . 1.35 79196 Refined borax soap.. . . . . . . . . . . . . . . . . . . . . . . . . . . 0.03 79518 Best borax soap.. . . . . . . . . . . . . . . . . Sone 79519 Best borax soap.. . . . . . . . . . . . . . . . . 3.42 i9520 Borax soap.. . . . . . . . . . . . . . . . . . . . . . 4one 79521 Borax soap.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . None 79522 Borax soap powder.. , , , , >-one 82569 Borax soap.. . . . . . . . . . Tone 82570 Cocoa boras soap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sone 82571 Borax nashing compoun
tain borax vary from I , 3 j to I O ,88 per cent. I n conclusion, I wish to acknowledge my indebtedness to my assistants, Messrs. J . A . Flanagan. JI. J. Vinick and E. C. h i l l e r , for their assistance in the analytical work. 1.EDERLE
LABORATORIES
39 WEST 3
8 STREET ~ ~ YEW YORKCITT
__
DETERMINATION O F PHOSPHORUS IN STEELS CONTAINING VANADIUM B y J. R. CAIS
AND
F. H. TUCKER
Received May 19, 1913
I t is known that vanadic acid. even when present in small quantity, interferes with the precipitation of phosphorus as phosphomolybdate. where the quantity of phosphorus is small relatively to the vanadium, as in steel, the vanadium not only contaminates the precipitate which is obtained, but also retards the rate at which it is formed and may prevent complete precipitation. While developing a method in this laboratory for the determination of vanadium by co-precipitation iTith phosphomolybdate,‘ it was observed that the vanadium is precipitated in this manner only when present in the quinquivalent condition ; quadrivalent vanadium was not precipitated, neither was pervanadic acid. In some experiments made by adding enough hydrogen peroxide to a solution containing vanadium to convert all of i t t o the pervanadate, not the slightest co-precipitation was observed with large or small amounts of added phosphate. The precipitates of phosphomolybdate obtained had the normal canaryyellow color, in contrast to the orange-red tint of predipitates containing vanadium. Other tests also showed no vanadium in such precipitates. These observations suggested the possibility of determining phosphorus in the presence of vanadium by converting the latter to pervanadic acid before adding the molybdate reagent. To test this question, determinations of phosphorus in the B. S. Vanadium Standard ( N o . 24). in the R. S . Chrome-Vanadium Standard (No. 30), and in some synthetic solutions containing varying proportions of iron, chromium and vanadium, were carried out by the alkalimetric method, except that the vandium was converted, be’
l Cain and Hostetter, B . ‘5. Tech. Paper S o . 8 : (1912).
THISJou~s.41,.
4, 2 5 0
647
fore precipitation of the phosphorus, to pervanadic acid by means of hydrogen peroxide. While this method yielded satisfactory results with the majority of the large number of samples tested, there were occasional irregularities which we were not able to eliminate. In cases where the method failed, the precipitates \\-ere red, the results were low, and no further precipitate formed after a long period of time. I n a series of determinations with a given sample faulty results occurred only occasionally. A great deal of work v a s done in trying to find the causes of these apparentlv accidental failures by varying the concentration, temperature, acidity, etc., of the solutions. but xv’ithout success. -4pparently the trouble is due t,o a decomposition of the hydrogen peroxide and of the peroxidized vanadium before complete precipitation of the phosphorus takes place and this decomposition is probably caused by the presence of one or more of the numerous catalyzers which decompose hydrogen peroxide. On-ing to this fact and to the further objection that hydrogen peroxide also peroxidizes the molybdenum of the precipitant, thereby causing possible abnormalities in the action of the latter, it was decided to abandon this method for general work. I t is quite possible, however, that correct results on a given sample may be obtained by making several determinations and eliminating from the average those which are evidently n-rong when judged by the criteria above indicated. Experiments which resulted in the discovery of a satisfactory method based on reduction of the vanadium to the quadrivalent condition were then begun. The fact that phosphorus may be quantitatively precipitated by the molybdic reagent when present with vanadium in this condition1 has long been known. We have found that the difficulty in applying this principle t o the analysis of steel arises from the fact that whereas original solution of the metal must be made in nitric acid, if the direct molybdate precipitation is to succeed, the presence of this acid complicates the reduction of the vanadium as the operation is ordinarily carried out. I t has often been attempted t o reduce vanadium to the quadrivalent state by means’ of ferrous iron, but the statements in textbooks referring to this method indicate, directly or indirectly, that success has not been attainid. Thus, Johnson2 states, and gives results indicating, that such reduction, as carried out in his experiments, did not give correct results for phosphorus. Brearley and Ibbottson3 give determinations on two samples of ferrovanadium containing large amounts of phosphorus, where the vanadium had been reduced under conditions specified by them. Whether or not the phosphoruscontent which Brearley and Ibbottson reported xi-as correct cannot be deduced from their statements, for this element was not determined by any other method known to give a correct result. It is equally impossible to decide from their experi1
Treadwell, “Kurz. Lehrbuch der -4naly. Chern.,” pp. 227 and 228,
4th ed., Val. 11, 1907. 2
3
“Chemical Analysis of Special Steels, et “The Analysis of Steel \\-arks Materials
648
T H E JOURiVAL OF I l Y D U S T R I A L A N D E N G I N E E R I N G CHE.IIISTRY.
ments whe.ther the vanadium was completely reduced a t the time of precipitation. If i t was not completely reduced (causing low results for phosphorus), there would have been produced a corresponding amount of the orange-colored precipitate which always forms in the presence of vanadic acid, but this, when associated with the large amount of normal phosphomolybdate which they obtained (their samples contained 0 . 6 5 per cent and I . 3 5 per cent phosphorus, respectively), might easily have escaped their notice. From other statements made by Brearley and Ibbottson on this subject* it would appear that they believe the formation of an orange-colored precipitate has little significance in affecting phosphorus results, the deficiency in phosphorus precipitated in such cases being compensated by the positive error caused by the vanadium associated with the precipitate. I n this connection it may be stated that our results show that formation of a red or orange-colored phosphomolybdate always indicates incomplete precipitation of the phosphorus. Having confirmed, by a series of results, the fact that phosphorus is quantitatively precipitated as phosphomolybdate in the presence of quadrivalent vanadium alone, the disturbing factors operating when vanadic acid is reduced by ferrous iron in nitric acid solutions of steels as a preliminary to the determination of phosphorus were investigated. Duplicate nitric acid solutions of steel containing added vanadic acid were reduced with excess of ferrous sulfate. TO one of the pair there was added the usual proportion of ammonia used in a regular phosphorus determination, and, after cooling, the color of the solution was compared with the companion solution to which no ammonia had been added. Not only had the blue color of the first solution due to reduced vanadium disappeared, being replaced by the yellow tint of vanadic acid, but the treated solutions usually gave but a slight test for ferrous iron. It thus appears that quadrivalent vanadium is oxidized by the nitric acid when the solution is heated by adding the necessary amount of ammonia. The vanadic acid thus formed would have prevented complete precipitation of phosphorus; moreover, most of the excess of ferrous iron, which would have tended t a keep the vanadium reduced, had disappeared. Solutions of steel containing vanadium and phosphorus were then prepared for precipitation of the latter by reducing with ferrous sulfate under the proper conditions to insure an excess of the reducing agent being present a t the time of adding the molybdic reagent, When precipitation was made a t 3 5 0 t o 4 5 0 c., the solutions usually gave no test for ferrous iron after completion of the five-minute period of shaking generally recommended. When the reducing agent had thus disappeared, the oxidation of the hypovanadic acid was indicated by change in color of the solution, by abnormal color of the precipitate, and by low results. However, if precipitation was made a t 1 5 0 c., excess of ferrous iron was almost always present. LOC.cat., p . 165
Vol. j, NO. 8
. We sought to determine whether phosphorus may be quantitatively precipitated as phosphomolybdate a t low temperatures. With prolonged and thorough shaking or agitation of the solution, quantitative precipitation can be made a t temperatures of 1 5 O to 20' C., and our successful determinations, given in Table I , have been made a t these temperatures. A I O minute shaking suffices ; the precipitate thus formed settles rapidly and filters as readily as that formed a t 35' to 45' C., as in the ordinary procedure for determining phosphorus in steels. . Even with an excessof ferrous iron, as indicated by the ferricyanide drop test, red or orange-colored precipitates and low results were often obtained. The solutions in such cases had a deep red color. There was a strong odor of oxides of nitrogen and sometimes even evolution of these gases. It was evident, then, that under certain conditions vanadic acid in nitric acid solution would not be completely reduced by ferrous iron, even with a considerable excess of the latter present. The trouble was found to be caused by the presence of oxides of nitrogen, due to interaction of ferrous iron and nitric acid; for when the gases were expelled from the solution, for instance by passing carbon dioxide in the cold for a sufficient time, the deep red color disappeared, the phosphorus precipitates were normal in color, and the results were quantitatively correct. To further test this point, a vanadate solution was reduced by sulfurous acid, the excess of reducer removed, and the vanadyl solution thus formed added to the nitric acid solution of a vanadium-free steel. Oxides of nitrogen, formed during the solution of zinc in nitric acid, were passed into the solution thus prepared. A gradual disappearance of the blue color, due to reduced vanadium, was observed, and the phosphomolybdate precipitate obtained from the solution of the steel was orange-colored, both results indicating that the oxides of nitrogen had caused oxidation of the vanadium. Some experiments were also made by precipitating phosphorus from steels containing vanadium, by reducing the latter with a n excess of ferrous sulfate, the flask in which precipitation was made being filled with carbon dioxide; under these conditions there was complete precipitation and the phosphomolybdate had the normal color. I t thus appears probable that the nitric oxide (NO) formed by action of the ferrous salt in nitric acid serves as a carrier of atmospheric oxygen to the vanadyl compound. The method of preparing solutions of steels for phosphorus determination in the experiments described above was the usual one of dissolving in nitric acid of I. 135 specific gravity, oxidizing with excess of permanganate solution, and destroying the excess of oxidizer (with simultaneous partial or complete reduction of the chromic and vanadic acids in solution) by ferrous sulfate, sulfurous acid, or a sulfite. After the observations above noted had been made, it was finally found that, observing the precautions shown to be necessary by the preceding experiments, samples in which the preliminary reductions of excess of permanganate, manganese dioxide,
Aug., 1913
T H E ]Oli‘R.\’AL
OF Ii\;DLTSTRIAL
a ~ v DE-VGISEERISG
CHE.UISTRY
649
etc., were made with ferrous sulfate almost always gave trouble through incomplete precipitation, red or orange-colored precipitates, etc., whereas, on the other hand, samples in which this reduction was made b y sulfurous acid or a sulfite nearly always gave good results. I n the latter class of samples, also, after addition of ferrous sulfate t o complete the reduction ofgthe vanadium before adding the molybdate reagent, no odor of nitrous fumes was noticeable. Experiments were then made with the idea of reducing the vanadium by sulfurous acid alone, without the use of ferrous salt a t all. This can be done; the essential conditions are to keep the temperature of the solution low and to allow time enough. Phosphomolybdate precipitated from such solutions has the normal color, and precipitation is quantitative. On account of the long time necessary, however, reduction bj7 sulfurous acid alone is not t o be recommended for routine work.
in solutions containing as high as I . 5 per cent of this element. This point was verified by testing phosphomolybdate precipitates (from a steel containing 0 . I 1 2 per cent phosphorus and I . j per cent vanadium) €or vanadium by dissolving the precipitates in concentrated sulfuric acid, adding a drop or two of nitric acid and heating till fumes were given off strongly; on cooling, no yellow color developed, showing that no vanadium had been coprecipitated.1
Our conclusion, then, is that the presence of small amounts of sulfurous acid in the solutions where that reagent was used as a reducing agent for destroying permanganate, etc., counteracted the effect of, or prevented the formation of, oxides of nitrogen upon adding ferrous salt. The formation of these oxides being thus avoided, their effect in preventing complete reduction of vanadic acid by ferrous salt was eliminated, and hence correct results could be obtained for phosphorus.
Cr
TABLE
Phosphorus found with-
_---
B . S. steel standards
.. ..
steel
~~,
> 0.035
{ ~:~~~~~ . . .. 3o 0 . 2 1 ( a ) 1.35(0) 0 . 0 4 3 .( 0 , 21( a ) 1.35(a) . . . .
V steel No. 29..
0.2 B. 0. H. 0.2 C 1 0 . 4
0.5 1 .o
2 .o 3.0 4.0 0.5 1 .o
2 .o 3 .O 4.0 0.5 1.0
As a result of these investigations, then, the following four conditions should be observed in precipitating phosphomolybdate in nitric acid solutions of steels containing vanadium by the method of reduction of the vanadium t o the quadrivalent state: ( I ) The temperature of precipitation should be held a t a point ( I j O t o 20 ”) where the nitric acid does not oxidize the excess of ferrous salt or the reduced vanadium before complete precipitation of phosphorus takes place ; ( 2 ) the partial neutralization with ammonia, frequently used when phosphorus is precipitated as phosphomolybdate, must be made before reduction of the vanadic acid, otherwise the heat of neutralization causes reoxidation of most of the ferrous iron and reduced vanadium b y the nitric acid; (3) care must be taken to prevent the action of oxides of nitrogen, formed by interaction of ferrous salt and nitric acid, on the red‘uced vanadium; since these substances seem t o catalyze the oxidation of the vanadyl salt and may in some cases completely prevent precipitation of the phosphorus, owing to the large amount of vanadic acid so produced; (4) efficient means for shaking or agitation of the solutions in which precipitation is t o take place must be provided. A fifth condition, not likely, however,to give trouble with ordinary care, is the avoidance of too great a n excess of ferrous salt. Under certain conditions, not fully investigated by us, a large excess of ferrous iron in the presence of vanadic and molybdic acids causes reduction of the latter, which, of course, is to be avoided for the present purpose. When all these conditions are observed there is no co-precipitation of vanadium
1
0.2 B . 0 . H. 0.2 C Renewal 9 a . .
.
,E,:
0.8 I .O
>
0.037 0.038
... ...
.. ,
0.046 0,048
..
,
Differences (in percentages on steel) with -
> +0.002
..
fO.003
, , ,
. .. ...
.
+ O ,003 0.005
+
0.0082 0 . 0 0 7 0.008 . . . . 0 . 0 0 8 0.007 . . . . 0 . 0 0 8 0.006 . . . . 0 . 0 0 8 0.010 . . . . 0.007 0.008
-0.001 -0.000 -0.000 -0.000
-0.001
-0.001 -0,002 f0.002 -0,000
0.2460 0.024 . . . . 0.025 . . . . 0.024 . . . . 0.023 . . . . 0.023
-0.001
-0,002
f0.000
-0,003 -0,001 -0,001 -0,003
0.045
.... .... ....
2.0 3 .O 4.0
....
0.5
0.084
1 .o 2 .o 3 .O 4.0
.... .... ....
0.5
0.112
1 .o 2 .o 3 .O 4.0
....
.... ....
.... ... .
0.023 0.022 0.024 0.024 0.022
-0.001 -0.002
-0.002
0.044 0.046 -0.001 0 . 0 4 6 0.047 f O . 0 0 1 0 . 0 4 4 0.047 -0.001 0.046 0.050 f 0 . 0 0 1 0.049 0.044 + 0 . 0 0 4
-0.000
+0.001 +0.002 f0.002 +0.005
-0.001
0.087 0.085 + 0 . 0 0 3 + o . o 0 1 0.084 0.083 0.000 -0.001 0.084 0.083 0.000 -0,001 0.086 0.084 f 0 . 0 0 2 o.ooo 0.084 0.083 0.000 -0.001 0.106 0.111 0.111 0.110 0.106
0.109 0.117 0.109 0.116 0.106
-0.006 -0.001 -0.001 -0,002 -0.006
-0,003
+0.005 -0.003 f0.004 -0.006
The method, which is described below, is adapted to any steel containing vanadium, no other element which ordinarily complicates the determination of phosphorus being present. If tungsten, titanium, arsenic, tin, etc., are present, their disturbing influence is eliminated by the usual methods for steels containing no vanadium. Nickel, copper, chromium, molybdenum, or aluminium, when present as alloying elements along with the vanadium, do not interfere. The details of the method are as follows: The solution of the steel, up to the point where precipitation of the phosphorus is t o be made, is prepared as for a n ordinarye phosphorus determination by the alkalimetric method, viz., solution of I. to z grams in roo cc. of nitric acid (specific gravity I . 135), oxidation with slight excess of permanganate solution while boiling, destruction of the excess of permanganate, etc., by a slight excess of sulfur dioxide or a sulfite, cooling of the solution and addition of 40 cc. of ammonia (specific gravity 0.96). Any steel which does not 1
Cain and Hostetter. loc. czt.
T H E J O U R Y A L OF IiVD U S T R I A L A S D E-VGINEERING C H E M I S T R Y
650
give all its phosphorus to the solution by this method must be treated in a n appropriate manner, so that the acidity, concentration of ammonium nitrate, and total volume obtained are always the same as by the method of solution directed. Slight deviations from these conditions are, however, probably without significance. Assuming solution of the steel and partial neutralization t o have been made as directed, the solution is cooled to I j ot o 20' C., 5 cc. of a saturated solution of ferrous sulfate and two to three drops of concentrated sulfurous acid are added. After addition of 40 cc. of molybdate reagent the solution is shaken in a n efficient manner for I O minutes. The precipitate, after settling (which is quite rapid), is then filtered off, washed in the usual manner, and t i trated by the alkalimetric method. The following table gives results by this method on synthetic solutions made with B. S. standard acid open-hearth, basic open-hearth, and Bessemer steels, of varying phosphorus and carbon content, to which vanadium was added as sodium vanadate and chromium as chromium nitrate. The acid and alkali solutions used for titrations were standardized against B:S. standard steels No. 19n (A.0.H. 0 . 2 ) and No. ga (Bes. 0 . 2 renewal). The individual determinations in each series are single determinations only; the good agreement in all cases with the certificate values for phosphorus of the individual steels shows the reliability and accuracy of the method. ' The time required is practically that for determining phosphorus in an ordinary steel. BUREAUOF STANDARDS ~VASHISGTOS -__
~
_
_
A RAPID MODIFIED CHLORPLATINATE METHOD FOR THE ESTIMATION OF POTASSIUM' By
\V. B HICKS
Received June 27, 1913
H I S T 0R I CA L
Many attempts have been made in recent years to develop a short practical modification of the chlorplatinate method for the estimation of potassium. These have proceeded along three essentially different lines, all of which have sought to avoid the necessity of estimating other constituents than potassium and sodium chlorides before undertaking the precipitation of the potassium. The first two modifications, the shortened method of Fresenius and the LindoGladding method, are so u-ell kno7v-n that they need not be discussed here. I n the third modification the potassium solution is evaporated directly with chlorplatinic acid, the residue washed u-ith alcohol, the potassium chlorplatinate reduced and either the chlorine titrated or the platinum weighed. Although such a procedure has been the 5ubject of numerous investigations in Europe, it seems never t o have found its way into American literature Because of this fact, and the rapidity, reliability and practicability of this procedure. the author feels justified a t this time in directing atten1
Published by permission of the Director of the LT. S. Geological
S u n ey
Vol. 5 , No. 8
tion to this excellent method, and in presenting the results of experiments along this line. Finkener,I in 1866, first used a reduction method for the estimation of potassium. He evaporated the potassium solution with chlorplatinic acid, washed the residue first with alcohol and then with ammonium chloride solution, reduced the K,PtCl, by igniting in hydrogen, extract.ed the residue with water and weighed the KCl, obtaining excellent results in the presence of various other salts. In 1873 Mohr* proposed to precipitate, wash and dry potassium chlorplatinate in the usual way, then t o mix this precipitate x i t h sodium oxalate, ignite, leach out with water, acidify with acetic acid a n d titrate the chlorine, using .K,CrO, as indicator, De Konincks suggested, in 1881, to dissolve the potassium chlorplatinate in hot water, reduce with magnesium ribbon and titrate the chloride with silver nitrate solution. To avoid the formation of oxychlorides in the reduction with magnesium, Fabrei suggested the addition of a drop of sulfuric acid which should be neutralized by means of CaCO, previous t o the titration with AgNO,. Diamants preferred to reduce the cblorplatinate solution with zinc dust rather than magnesium, because under the conditions .no oxychlorides of zinc were formed. I n a second investigation on the subject, De Koninck6 effected the reduction of the chlorplatinate by warming the solution with calcium formate. Stolba7 proposed, in 1884, to reduce the K,PtCl, with amorphous silver. This reduction, according to Atterberg,8 proceeds too far and is not to be recommended. Many reduction methods in which platinum is weighed have been proposed. Thus Hilgardg reduced the K,PtCl, by heating, while MeisslIo and Vogel and HaefckeII effected the reduction by heating the precipitate contained in a porous Gooch crucible in a current of hydrogen. NeubauerI' substituted coal gas for hydrogen because it is more practical and gives equally good results. Sonstadt13 accomplished the reduction by heating the precipitate with mercury. In all cases after complete reduction the soluble salts were washed out and the platinum weighed. I n the wet way the oldest reduction method is that of Corenwinder and Contamine.14 According t o their procedure, the K,PtCl, dissolved in hot water is added to a boiling solution of sodium formate. This method is often slow and leaves adherent deposits on the walls of the vessel which are difficult t o remove according to Jean and Trillat,~swho proposed to reduce P o g g . Ann., 9, 637 (1866). Z . anal. Chem.. 12, 1 3 7 (1873). Ree. Uniu. des Mines, 9, S o . 2 (1SS1). 4 Chem. Z., 20, 5 0 2 (1896). 6 Ibid., 22, 99 (1898). 6 Ibid., 19, 901 (1895). 7 I b i d . , 8. 456 ( 1 8 8 4 ) . 8 Ibid., 22, 538 (1898). 9 Hilgard, Versuchsstatio%en. 42, 174 (1893). 10 .Meissl. Ibid., 42, 173 (1893). 11 Landw'. V e r s . - S f a .47, , 109 (1896). 12 Z . anal. Chem.. 39, 481 (1900). 13 J . Chem. S o c . , 67, 984 (1895). 14 Bull. de la SOC. Indusfrielle du S o r d , 1879. 15 Bull. de la S O C Chim. . de Paris. [3] 7 , 2 2 8 (1892). 1
2
3
