Aug., 1913
T H E J O V R S A L OF I,\-DUSTRIAL
especially determined in the kernels. The fat content appears to vary but a few per cent in the kernels of nuts a pear old or fresh. The fresh nuts naturally contain more moisture. 1 year-old Fresh nuts nuts Per cent Per cent Moisture.. . . . . . . . . . . . . . . . . . . . . 3.55 7.14 F a t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 . O O 66.25 rlsh . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.56 3.05 Protein.. . . . . . . . . . . . . . . . . . . . . . . . 18.75 19.88 Fiber.. . . . . . . . . . . . . . . . . . . . . . . . . 2.14 1.39 I\-itrogen-free extract b y difference.. 7 .OO 2.29
__
__
100,oo
100.G O
Immature nuts Per cent 13.39 56.7
...
... ...
...
The fiber and the ash are very low, the principal constituents being fat and protein. I t is of interest to note in this oil-seed the small percentage of hydrolyzable carbohydrates (only I . 4 0 per cent by analysis). A qyalitative test showed the absence of starch in the kernels. The quantity of fat, if calculated to the waterfree basis, is quite constant in the three samples. Analyses made of the kernels of this same nut, Alcztrites wzolucca?zn, by various chemists are given in the table belox-: (1) Per cent
IYater. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.00 Oil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.175 Protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.653 Sitrogen-free extract. . . . . . . . . . . . . . . . . . . . 6.827 -4sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.345 Fiber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100.00
0) Per cent s.25 59.93 8.04 17.62 3.56 2.62 100.00
~ 1H ) . Semler, Die Trofiische Agrikullur, 2, 5 1 5 .
‘ 2 ) N. S . Wales, Agri. Gazelle, 17, 859 (1906).
The mineral constituents of the kernel were also determined in this laboratory, as shown in the following table : Per cent 0.03 ....... . . . . . . . 0.17 ....................... 0.60 P20j. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.59 .K*O. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.75 ~ l n 3 0.r. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
On extraction of the oil these constituents’ increase in proportion t o the dry matter. The residue contains, therefore, large amounts of phosphoric acid, potash and nitrogen, all valuable as fertilizing ingredients. Analysis of the residue of ground up kernels extracted with ether for two days is as follows : Moisture.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K20
....................................
...
Protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Per cent 2.42 2.79 2.i7 53.75
Press cake obtained commercially by expressing the oil from kukui nuts would compare favorably with the residue, as most of the oil can be expressed. According to an article in the Agricztltziral Gazette,2 55 per cent oil can be obtained from kernels containing 60 per cent. Analyses of press cake made elsewhere shows i t to be valuable as a fertilizer. 1
Bull. I m p . I n s t . , loc. czt.
.4pr. G Q Z .:Y.S.W’.,17, 859 (1906).
AND E.VGI-VEERI.YG C H E N I S T R Y
645
ANALYSES OF P R E S S C:AKE
(1)
P e r cent Oil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.8 . . . . . . . . . . . . . 10.00 Ash . . . . . . . . . . . . . . . . . . . . . . . Protein. . . . . . . . . . . . . . . . . . . . Fiber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.47 PZOj . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.39 PZO . . . . . . . . . . . . . . . . . . . . . . . 1.95 M g and C a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
(2) P e r cent 5.5 10.25
...
47.81
... 3.68 1.53 7.19
J . SOC.Chem. Ind., 20, 509 (1901). (2) H.Semler, Die Tropische Agrik., 2 , 51.5. (1’1 Lewkowitsch,
Press cake cannot be used as a fodder in spite of its apparent food value, as i t has a poisonous effect upon stock.1 The proportions of fat to kernel and whole nut were determined in several samples. In a sample one year old in which the fleshly husk that covers the shell when the nut is green, had decayed away, the kernel weighed 3 1 . 5 per cent of the hard nut. As the fat content of the kernel was 65 per cent, the fat content of the nut minus the husk, averaged about 2 0 per cent. In a sample of fresh nuts, the kernel was 2 9 . 3 per cent of the nut minus the husk and 1 2 . I per cent of the nut with the husk on. The fat content of the kernel was about 66 per cent. I t was, therefore, about 8 per cent of the nut with the husk, on and about 1 9 . 4 per cent of the hard nut minus the husk. I n another sample of fresh nuts, the kernel was I Z per cent of the nut with the husk on and 3 0 . 4 per cent of the hard nut minus the husk. The kernel contained about 5 7 per cent fat. The fat content of the whole, therefore, was about 7 per cent and of the nut minus the husk about 17 per cent. HAWAIIEXPERIMENT STATION HONOLULU
T H E DETERMINATION O F SODIUM BORATE IN SOAPZ B y PAULPOETSCHKE
The customary method employed for the determination of borax in soap has been to extract the dry soap with 9 j per cent alcohol, or, the soap without drying is extracted with absolute alcohol. The residue, insoluble in alcohol, is then treated with water and the borate determined by titration. I t has been assumed that borax is insoluble in alcohol, undoubtedly, from the fact that this statement appears quite generally in the text-books. About a year ago the author had occasion to conduct quantitative analyses of quite a number of commercial borax soaps and i t was soon established that extraction with alcohol was unsatisfactory, owing to the solubility of borax in this liquid, even if the soap was thoroughly dried and extracted with absolute alcohol. When chemically pure sodium borate was boiled with absolute alcohol and carefully filtered until the filtrate was absolutely clear, i t was found that the filtrate upon evaporation left a residue which readily dissolved in cold water, gave a n alkaline reaction to methyl orange and a positive test for boron. The same Agric. G o z . ,Y.S. W’.,17, 859 (1906). :: Read before X, Y.Section, ilmerican Chemical Society. May 9. 1913
(r I
*
T H E JOURlYAL OF I l V D U S T R I A L A N D E i V G I X E E R I S G CHE*\fISTRE'
646
sodium borate which had been thus boiled with alcohol was treated twice in the same manner and the individual filtrates evaporated to dryness. Both of these filtrates gave a residue which was readily soluble in cold water, the later solution giving an alkaline reaction to methyl orange and a positive test for boron. An alcoholic solution of boric acid left no residue on evaporation. This demonstrates conclusively that the material dissolved by the alcohol consisted of borax and that i t could not have been due t o free sodium hydroxide or free boric acid. Two samples of commercial borax soaps were thoroughly dried a t 100' C. and extracted with absolute alcohol in a Soxhlet extractor. The insoluble residue and the alcoholic extract were separately analyzed for the quantity of borax contained in each. Table I gives the results of these experiments, together with the total borax content of the soaps as determined by the method described in this paper. TABLEI Boras in alcohol extract Per cent
Sample A , ,, , . .. ,
B.. . . . . , .
0.40 2.53
Boras in extracted residue Per cent
Total boras in extract and residue .Per cent
1.47 8.18
1.87 10.71
Total boras content of soap Per cent 1.87 10.88
I t is seen from the data in Table I that the error resulting from the solubility of the borax in alcohol amounts t o 21.4per cent of the total borax content in A , and 2 3 . 2 per cent in B, and that the borax found in the residue and extract agrees well with the total borax content of the soap. An attempt was made to determine the borax by ignition of the soap with alkali carbonates. The ash was dissolved in water, slightly acidified with hydrochloric acid and boiled under a reflux condenser in order t o expel carbon dioxide. After neutralizing the excess of acid with sodium hydroxide, using methyl orange as a n indicator, the boric acid was titrated with N / I O alkali in the presence of glycerine, using phenolphthalein as indicator. Owing to the presence of silica in the soaps, this method failed to yield satisfactory results. Wherry1 has described a volumetric method for the determination of boric acid in insoluble silicates and a n adaptation of this method was found t o yield satisfactory results. It was found t h a t silica wTas not always present in the soaps in sufficient quantity t o give a satisfactory precipitation, so t h a t it became necessary to add a fixed quantity of silica t o the alkali carbonate used in the fusion mixture. The method as finally developed is as follows: Weigh I O grams of the soap (or j grams if more than 5 per cent of borax is present) into a platinum dish and add 2 . 1 5 grams of fusion mixture (consisting of 2 0 0 grams sodium carbonate, r 5 grams silica in fine powder). To this mixture add 1 5 cc. of alcohol, mix with the aid of a glass rod and after washing the rod with a little alcohol, evaporate the mass t o dryness on the water bath. Ignite until the combustible material is destroyed, cover the dish with a piece 1
"Determination of Boric Acid in Insoluble Silicates." E. T. Wherry
and W. H. Chapin. J . A m . Chem. SOC.,30. 1687.
Vol. j, NO. 8
of platinum foil and fuse. Completely disintegrate the fusion by boiling with water and transfer the solution t o a 2 5 0 cc. round-bottomed flask. Acidify with 2 0 cc. of dilute hydrochloric acid ( I : I ) , heat nearly to boiling and add a moderate excess of dry precipitated calcium carbonate. Connect with a reflux condenser and boil vigorously for I O minutes. Filter out the precipitate through a folded filter, washing several times with hot water, but keeping the total volume of liquid below I O O cc. Return the filtrate to the flask; add a pinch of calcium carbonate and again boil under a reflux condenser. Remove the flame and connect the top of the condenser with a water pump. Apply the suction until the boiling has nearly ceased. Cool t o ordinary temperature, add j o cc. of neutral glycerine and titrate the solution with 0 . I normal sodium hydroxide, free from carbonate, using phenolphthalein as indicator. After the end point is reached add I O cc. more of glycerine and again titrate. Repeat this process'until the addition of glycerine causes no further action on the end point. The number of cubic centimeters required multiplied by o.oog5j will give the equivalent of borax (Na,B,O,.IoH,O) present in the solution. A number of analyses were made in which a definite amount of chemically pure borax was added t o castile soap which was free from borax. The results of these determinations are given in Table 11: TABLEI1 Per cent sodium borate found Per cent sodium borate added
A
B
10
10 00
9.99 6.96 4.96 2.96 1.96 0.98
6.96 5 3 7
1
4.91
2.95 1.97 0 95
It is evident from the table that the method gives results which are in close agreement with the actual borax content of the soap. I t is always advisable to test a soap qualitatively for borax before undertaking the quantitative determination. This is accomplished by placing one gram of the soap in a test tube, together with I O cubic centimeters of dilute hydrochloric acid. The mixture is then heated to boiling which causes the fatty acids to rise to the surface. After cooling the tube under a tap of running water, the acid liquid is filtered through a wetted filter. A strip of turmeric paper is immersed in the filtered liquid and dried. when it will acquire a deep red color, which changes t o green or blue on the addition of ammonia, if borax is present. A number of tests conducted with known amounts of borax show that the test performed in this manner is sensitive t o 0.05 per cent of borax in the soap. There is a considerable variation in the borax content of commercial borax soaps. Analyses of fourteen different brands are given in Table 111. It is thus apparent that a considerable number of the commercial borax soaps contain absolutely no borax, or practically none, whereas those which con-
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