W Y S O R ON THE DETERMINATION OF PHOSPHORUS. desired. It seems that the amount of insoluble bromides in oils of the same kind but from different sources varies to a considerable extent.' I t is on account of heating and other oxidizing processes to which the oil are subjected before they reach the market that no real quantitative method will ever be worked out on the lines laid out above. The fact, however, that the precipitate formed can be weighed, and is indeed in direct proportion to the amount of fish oil present in a given sample, elevates this reaction above the level of a n ordinary qualitative test, and therein lies a very important adirantage over the methods now in use. The short time required for the performance of the test and its applicability to both raw and boiled linseed oil are other factors of importance. CONCLUSIONS.
The following test for fish oils of any kind in admixtures with vegetable oils or similar products is suggested. IOO drops of oil are dissolved in 3 cc. of chloroform and 3 cc. of glacial acetic acid. Bromine is added slowly, and after about I O minutes the test tubes are placed in boiling water. All vegetable oils clear up, while fish oils remain cloudy. By filtering and weighing the precipitate formed, it is possible to obtain objective results. I n case of boiled oils it is necessary to remove the metallic salts before adding the bromine. Fish oils that have been heated to 260' C. or more for some time will not respond to this test. The authors desire to express their gratitude to Miss Gertrude Stock who performed a large proportion of the analytical work carried out in connection with the above investigations. LABORATORY OF AM LINSEED Co., So CHICAGO,ILL.
THE DETERMINATION OF PHOSPHORUS RETAINED WITH FERRIC CHLORIDE IN THE ETHER SEPARATION? BY R. J. WYSOR. Received December 23, 1909.
The use of ether as a medium for effecting the separation of iron from other metals in oxidized 1 I n corroboration of our results we quote from a private communication of L. M . Tolman. of July 7, 1909. "I must say, however, t h a t m y experience with these fish oils is such as would make i t seem almost impossible to make any standard on their composition, the various determinations showing such a tremendous variation on oils of absolutely known origin. This wide variance you will note in connection with the iodine absorption of the cod liver oils themselves, and also the amount of bromine precipitate formed. " I n connection with this matter, we must state t h a t we found i t impossible to obtain samples of absolutely known origin through the Bureau of Fisheries a t Washington. I n reply t o our request for fish oil samples of known origin, H. hf. Smith, Acting Commissioner of the Bureau of Fisheries a t Washington. stated t h a t he had no samples of oil on hand, and n-as not in a position to procure them for us. Referring to OUT remarks t h a t commercial fish oils vary considerably in their constants, and t h a t market samples probably had been oxidized, MI. Smith wrote t h a t "menhaden oil is subjected to heat when the fish are cooked. and also after the oil is pressed o u t , and in the process of refining." 2 Read a t the December meeting of the Pittsburg Section. -4merican Chemical Society.
45
hydrochloric acid solutions has long had practical application in analytical chemistry. J. W. Rothe, in his original paper on this subject,' discusses in detail his experiments and the method by which the ether separation is accomplished. Carnot, in a later publicat i o q 2 pursues an investigation along the same general lines as Rothe, and adds one or two metals to the list of those which may be separated from the bulk of iron by the ether method. Recently, Andrew A. Blair3 has shown that molybdenum, in common with iron, exhibits an affinity for ether, and that a complete separation of this element may thus be effected from the other metals. The writer has discovered no literature which treats of the behavior of phosphorus, or other non-metallic elements, in company with iron in the ether separation. I n reviewing Rothe's article, the statement was found that no attempt had been made to determine whether ferric phosphate, or ferric sulphate could be separated from acid solutions. It is the purpose of this paper to consider the behavior of relatively small amounts of phosphorus in the ether separation, under somewhat empirical conditions, in order that a certain critical value may be attached to the discussion. Conditions of experiment commonly occurring in practice were chosen, thus affording a commentary upon an existing application of the ether method, to which reference will subsequently be made. I n a number of preliminary experiments, a steel of very low phosphorus content was employed, together with a standard phosphorus solution, prepared by dissolving diammonium hydrogen phosphate in hydrochloric acid (1.13 sp. gr.). d sufficient volume of this solution to represent the amount of phosphorus desired was added to I, z or 3 grams of the steel. A further addition of enough hydrochloric acid (1.13 sp. gr.) to make a total volume of 40 cc. was then made, and after solution had been effected, the ether separation was made in the same manner as later described. Although interesting results were obtained, comparable with those secured in the ensuing work, their practical value was not evident, and the following materials were selected as the basis of experimen t . Two iron ores were chosen of the same soluble iron content, 5 7 . j per cent., and of widely variant soluble phosphorus content, 0.060 per cent. and 0 . j I o per cent. These ores were mixed in such proportions as to represent approximately even gram weights of metallic iron, and definite weights of phosphorus as subsequently noted. An iron ore of 68.6 per cent. soluble iron content Mztthezlunpen aus den Ronaolzch Tech. Versuchsanstalien zu Berlin, 1892, P a r t 111. 2 "Methodes d'Analyse des Fontes, des Fers, et des Aczers," 1895, p 1 2 3 , et seq Journal American Chemzcal Soczety, 1908, p. 1229.
46
T H E JOURNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY.
and 0.050 per cent. soluble phosphorus content was used in conjunction with a standard phosphorus solution. This solution was prepared by dissolving diammonium hydrogen phosphate in dilute hydrochloric acid, and was carefully standardized so that each cubic centimeter contained 0.0005 gram of phosphorus. Several other iron ores were subjected to sufficient experiment to indicate that approximately the same results were obtainable as those described in this paper. Method of Procedwre.-The requisite weight of the single ore, or of the two ores to represent I , 2 or 3 grams of metallic iron, was employed. I n the latter case, the respective ores were properly proportioned to yield 0.001, 0 . 0 0 2 , 0.004, 0.008 or 0.016 gram of phosphorus as desired in the mixture. In the treatment of the single ore, the proper volume of the standard phosphorus solution was added, which togethFr with the phosphorus content of the ore would afford the desired weight of this element. All calculations were based on the soluble iron and phosphorus content of the ores as previously mentioned. The numerical results obtained by these analogous procedures appear in the appended tables in parallel columns. The calculated weight of the single or mixed ores was digested in strong hydrochloric acid, with the aid of a gentle heat, for two hours, or until no further solvent action was apparent. The solution was diluted and filtered, the residue, after being washed free of iron discoloration, being discarded. If the standard phosphorus solution was to be added, the addition was made a t this point, and the filtrate evaporated to dryness. The residue was dissolved in 40 cc. of hydrochloric acid (1.13 sp. gr.), gentle heat being applied to aid solution. The solution was cooled in water, and transferred to a single-chambered, cylindricalshaped, separatory funnel, provided with a ground glass stopper, and having a capacity of about 175 cc. Anhydrous ether (distilled over sodium) was used for all the separations, 40 cc. being employed for I gram of iron, 7 5 cc. for 2 grams and IOO cc. for 3 grams. P a r t of the measured amount of ether was used to rinse the dish, from which the solution had been transferred, the rinsings with the remainder of the ether being transferred to the separatory funnel. After thorough agitation under cold, running water, the solutions were allowed to stand several minutes to permit as complete a separation as possible. The lower stratum was then carefully drawn off and discarded. The ethereal solution, containing the bulk of ferric chloride, if not subjected to further treatment, was washed with a sufficient quantity of water to remove all the iron and concomitant phosphorus. A second series of determinations was made in the Same manner as just described, except that after the
Feb., 1910
lower, acid solution had been drawn off, the ether solution was rewashed with I O cc. of ether-saturated hydrochloric acid (1.13 sp. gr.). The solutions were again allowed to stand several minutes, the lower stratum was drawn off and discarded, and the ether solution washed with water as previously described. The final solution, containing nearly all the iron, with the residual amount of phosphorus to be determined, was warmed to volatilize the remaining ether, then the excess hydrochloric acid expelled by evaporation. The determination of the phosphorus was completed by the well-known gravimetric method. The interesting behavior of phosphorus under the varying conditions of experiment may readily be interpreted from the tabulated results, and the accompanying diagrammatic presentation. TABLEI -PERCENTAGEOF PHOSPHORUS RETAINEDI N
THE
ETHEREAL
SOLUTIONS OF IRON.
.
Total phosphorus present. Gram. 0.00104 0.002 0.004 0.008
0.016
..... 0.00208 0.004 0.008
0.016
..... 0.004 0,008
0.016
Mixed ores. Percentage of total phosphorus retained with 1 gram iron. 23.2 19.0 18.3 17.5 16.9
Ore plus standard phosphorus solution. Total Percentage of phosphorus total phosphorus present. retained with Gram. 1 gram iron. 0.001 24.0 0.002 18.7 0,004 17 . O 0,008 15.9 0.016 15.8
2 grams iron.
..
0,004 0.008 0.016
2 grams iron. 32.3 31.3 30.9 29.5 29.0
0.00219 0.004 0.008 0.016
3 grams iron. 37.7 3.5.1 35.5 35.9
0.00146 0.002
33.0 32.3 30.6 29.7 3 grams iron.
..
35.2 34.8 35.4
TABLEII.-PERCENTAGEOF PHOSPHORUS RETAINED IN THE ETHEREAL SOLUTIONS O F I R O N , REWASHED W I T H 10 C C . O F ETHEREALIZED HC1 (1.13
Total phosphorus present. Gram. 0.00104 0.002 0.004 0,008 0.016
..... 0,00208 0.004 0.008 0.016
'
SP.
Mixed ores. Percentage of total phosphorus retained with 1 gram iron. 22.3 15.8 13.4 12.3 11.6 2 grams iron.
..
24.7 24.2 23.5 23.3 3 grams iron.
.....
..
0.004 0,008 0.016
28 .O 26.0 26.5
GR.). Ore plus standard phosphorus solution Total Percentage of phosphorus total phosphorus present. retained with Gram. 1 gram iron. 0.001 22.0 0.002 15.4 0.004 12.7 0.008 12.2 0.016 12.2 2 grams iron. 0.00146 26.2 0,002 24.6 0.004 23.4 0.008 22.8 0.016 22.4 3 grams iron. 0.00219 26.8 0.004 26.6 0,008 25.2 0.016 26.1
..
From the preceding results it is evident that under fixed conditions, a fairly constant percentage of the phosphorus present remains with the iron, in the ether separation. This phenomenon a t once suggested a 1 Described in the methods of the United States Steel Corporation for t h e Commercial Sampling and Analysis of Iron Ores.
LOWENSTEIN A N D D U N N E ON E T H E R E X T R A C T OF SPICES. practical adaptation of the results obtained. I n methods for the estimation of alumina, involving a n ether separation of the iron, followed by a n ammonia precipitation, i t has been assumed that all the phosphorus present remained with the alumina and the other metals. I n the ether method for the determination of alumina, as described in the methods of the United States Steel Corporation for the Commercial Sampling and Xnalysis of Iron Ores, a correction is DL9GRAM OF THE PERCENTAGE OFPHO5PHORU5 RETA/NED /N THf .LTHEPSO?TUN3OFFfRR/CG" OR/E
47
average of the corresponding results presented in the preceding tables. The figures here tabulated are additive values, to be applied as direct increments to the weight of the mixed oxides of aluminum, iron and phosphorus, previously noted. It will be seen that there is a n increasing amount of phosphorus lost in the ether solutions with an increase of both the phosphorus and iron content of the ores, though the percentage error in the present ether alumina method is reduced by the use of large samples. Although corrective estimates are not given for intermediate variations of phosphorus and iron content, i t seems rational to assume that they may be approximated by interpolation. TABLE111.
Weight soluble phosphorus pentoxide in ore. Gram. 0.0023 0.0046 0.0092 0.0183 0.0366 0.0023 0.0046 0.0092 0.0183 0.0366
(Calculated from Table I.) Weight phosphorus pentoxide t o be added to weight of alumina residue found.
-
A -
1 gram iron. Gram. 0.0005 0.0009 0.0016 0.0031 0.0060
2 grams iron.
Gram.
....
....
0.0015 0.0029 0.0055 0.0108
0.0017 0.0032 0.0064 0.0130
(CalculatedIfrom Table 11.) 0.0005 0.0007 0.0012 0.0022 0.0044
3 grams iron. Gram.
....
....
0.0011 0.0022 0,0042 0,0084
0.0012 0.0024 0.0047 0.0097
In order to determine the behavior of increased amounts of phosphorus in the ether separation, several experiments were performed according to the conditions of the second series of determinations, with weights of phosphorus ranging upwards to I gram, a strong standard solution being prepared for this purpose. The results obtained indicate that about the same percentage of phosphorus remains in the ether solutions with the presence of relatively greater amounts, as with the largest weight (0.016 gram) of phosphorus previously employed. Expressed diagrammatically, the lines representing the percentage of phosphorus remaining with the ferric chloride in the ethereal solutions, if prolonged, would suffer no further, appreciable depression. applied to the weight of the mixed oxides of iron, aluminum and phosphorus by deducting the combined weights of iron (determined by titration) as ferric oxide, and the total phosphorus content of the ore as pentoxide. The conditions of experiment represented in Table I of this paper are practically the same as those existing in the above-mentioned method. The approximate error incident to this method may, therefore, be calculated and the proper correction applied, 'For convenience in reference, Table I1 I, appearing below, has been prepared for weights of phosphorus pe'ntoxide and metallic iron; these conditions, of course, are to be observed in effecting the proper corfections. This table has been constructed from the
.
LABORATORY OF D U Q U E S N E STEEL WORKS.
THE INFLUENCE O F THE METHOD O F HEATING ON THE NON-VOLATILE ETHER EXTRACT O F SPICES. By
ARTHUR
LOWENSTEIN A S D \YILLIAM P. DUNNE. Received January 6 , 1910.
In a paper recently reported by us on "Spanish Paprika" we called attention to the fact that in heating the non-volatile ether extract of this spice to constant weight, the method of heating was an important factor. The official method of the A. 0. A. C. calls for "heating to constant weight at 110' C.," without specifying how this is to be accomplished, and we take i t that i t is usually interpreted to mean
