A Modified Method for the Determination of Fluorine with Special

Volumetric Method for Determination of Fluorine. H. H. Willard and O B. Winter. Industrial & Engineering Chemistry Analytical Edition 1933 5 (1), 7-10...
0 downloads 0 Views 1MB Size
Download PDF
1116

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 TECHNIC OF THE METHOD

The detail of t h e procedure is as follows: Carefully precipitate t h e lime as oxalate, reprecipitating, as is usually done, where t h e magnesium content is appreciable. Ignite in a small platinum dish (or platinum or porcelain crucible) over a Bunsen burner (or in a muffle) until t h e filter is completely incinerated. For each approximate 0 . 2 g. of C a C 0 8 add enough of a n equal part a n d jFnely ground and dried mixture of ammonium sulfate a n d ammonium chloride t o insure a n excess of approximately 0.3 g. of sulfate. Effect a thorough mixture of t h e fusion salts with t h e lime residue in t h e crucible by means of a small glass rod, enlarged and flattened a t one end. Quantitative comparisons indicate t h a t i t is decidedly preferable t h a t t h e mixing be done in t h e crucible, rather t h a n by transfer t o a mortar. T h e volatilization of t h e excess of salts may be efficiently carried out as follows: insert t h e crucible in a circular opening cut in a piece of asbestos board, placed horizontally. The upper half of t h e crucible should extend above t h e upper surface of t h e asbestos. Direct a nearly horizontal flame from a small Bunsen burner across t h e surface of t h e crucible in such a manner as t o have t h e side of the crucible nearest t h e flame intensely heated. The conducted heat will effect volatilization without spattering. Should duplicates fail t o agree within a few tenths of a milligram, t h e analyst may verify t h e results by moistening t h e ignited Cas04 with a few drops of I : I O H2SO4, evaporating excess of water and again igniting. The above procedure has been tested thoroughly against reagents of known purity and as t o ease in close duplication of results in a large number of unknowns, a n d has given complete satisfaction. We have found t h e modification t o be especially well adapted t o high occurrences of lime and t o sets containing widely varying percentages. I

SUMMARY

I-Though not in general use, t h e determination of lime as C a S 0 4 is authoritatively stated in accepted texts t o be accurate. 11-The procedure generally given calls for t h e use of H2SO4 as the converting agency. 111-In addition t o loss of weight as a result of t h e dissociation of sulfate because of high tempwature, there occurs a loss due t o the reduction of sulfate t o sulfide, as induced b y t h e small amount of organic matter contained in “C. P.” HzS04. IV-The use of recrystallized ammonium sulfate eliminates this objectionable feature. V-It was found t h a t a mixture of ammonium chloride a n d ammonium sulfate afforded a very satisfactory conversion mixture which gives very accurate results. VT-The detail of t h e procedure, as perfected, is given. AGRICULTURAL EXPERIMENT STATTON UNIVERSITY os T B N m c a s s s . KNOXVILL~

Vol. g , No.

12

A MODIFIED METHOD FOR THEDETERMINATION OF FLUORINE WITH SPECIAL APPLICATION TO THE ANALYSIS OF PHOSPHATES’ By CARY R. WAGNERAND WILLIAMH. Rosa Received September 13. 1917

I n t h e preparation of phosphoric acid from phosphate rock by t h e sulfuric acid method there may be found as impurities in the acid any of t h e constituents of t h e raw materials used. I n t h e volatilization method, however, which consists in smelting t h e rock with silica and coke in a n electric or other furnace, t h e impurities may be limited t o those constituents of t h e charge which are volatile a t t h e temperature used in t h e process. The most important of these volatile constituents are t h e alkalies, sulfur, chlorine a n d fluorine. The last mentioned is usually t h e most abundant and is found associated with lime phosphate in all mineral deposits a n d in t h e bones of animals. I n t h e scrubbing tower method of recovering t h e phosphorus oxide fumes evolved in t h e heat treatment of phosphate rock n o , separation of hydrofluoric from t h e phosphoric acid takes place Since both are absorbed in the process. I n experiments made in this laboratory it has been shown2 t h a t by use of t h e Cottrell process of electrical precipitation in place of t h e scrubbing tower method much more efficient and economic recovery of t h e acid is possible, and in addition a separation of t h e hydrofluoric acid from t h e phosphoric acid may be brought about a t t h e same time. I n carrying out this process t h e phosphorus pentoxide, as it is evolved from t h e furnace, quickly combines with t h e moisture driven off from t h e charge or with t h a t present in t h e current of air which is passed through t h e furnace for t h e oxidation of t h e evolved phosphorus, and is thus precipitated in t h e form of a solution of phosphoric acid. If t h e temperature of precipitation is below rooo t h e strength of the solution will depend on t h e amount of moisture associated with the acid. When the precipitation is made a t a temperature above IOO’, however, t h e strength of t h e acid will be independent of t h e moisture with which it is associated. Under these conditions any excess moisture will remain in t h e gaseous state while passing through t h e precipitator pipes and will consequently escape without being precipitated. By use of a proper temperature for precipitation, acid may thus be recovered directly of such a concentration t h a t it will crystallize t o a solid mass on cooling. T h e same conditions which bring about a recovery of the phosphoric acid in concentrated form also contribute toward its purification from such volatile constituents as hydrofluoric acid with which i t is initially associated as it escapes from t h e furnace. At a temperature above 100’ these constituents will remain in the gaseous state in passing through the treater pipes and will therefore undergo no precipitation excepting t h a t portion represented by the solubility of t h e respective constituents in t h e phosphoric acid under the conditions of the precipitation. As referred 1

Published by permission of the Secretary of Agriculture. Ross, Carothers and M e n , T H I S JOURNAL. 9 (1917), 26.

Dec., 1917

T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y

t o in experiments which are t o be described more in detail in a later publication complete removal of the hydrofluoric acid or other impurities m a y subsequently b e effected in one of several ways, as by bubbling h o t air through t h e acid, b y recrystallization or by chemical means. I n carrying o u t t h e work thus outlined o n the preparation of crystallized phosphoric acid free from hydrofluoric acid, it was found necessary t o make numerous analyses for fluorine when present in phosphoric acid in widely varying amounts. No official method is available for the quantitative determination of fluorine and the methods given i n t h e literature were found t o b e of limited application. Accordingly an a t t e m p t was made t o devise a suitable method for the determination of fluorine in t h e presence of phosphates or phosphoric acid. T h e object of t h e present paper is t o give an account of the results obtained. P R O P O S E D M E T H O D S F O R THE ANALYSIS O F BLUORINE

T h e various methods which have been proposed for the determination of fluorine m a y b e conveniently divided into four classes as follows: 1-Gravimetric methods, as the Berzelius-Rose method in which the fluorine is weighed as calcium fluoride. 2-Etching methods, in which comparative tests are made of the action of hydrofluoric acid o n glass or quartz. 3-Colorimetric methods. 4- Volatilizatiolt methods, as the Offermann method, i n which t h e fluorine is distilled as silicon fluoride and collected in water. I . GRAVIMETRIC METHODS-In the Berzeliusl method as modified by Rose2 and by Treadwell and Koch,3 the sample is fused with sodium and potassium carbonates, the silica precipitated with ammonium carbonate and ammoniacal zinc oxide, phosphates removed from the neutrai solution-by silver nitrate, the excess of silver taken out by sodium chloride, and CaF2 CaC03 precipitated with a large excess of calcium chloride. Calcium carbonate is dissolved after igniting the mixed precipitate in I .5 N acetic acid and the calcium fluoride weighed. The method gives low results on account of the solubility of calcium fluoride in water and acetic acid. If phosphates are present the solution becomes acid when silver nitrate is added, as indicated in the equation 2NaNOs Ag3P04 "Os. Na2HP04 3AgNOs This acidity prevents complete precipitation of silver phosphate and, furthermore, will cause volatilization of fluorine when the solution is heated, unless the acidity is neutralized. This is a point that seems to have escaped notice in texts on the subject and could easily be overlooked by an analyst working on an unknown sample. Ditte' and Deussen and Kesslel-6 have devised methods depending upon the formation of calcium fluoride in a fusion, the remainder of the melt being removed later by solution. Dinwiddies proposed precipitation of calcium fluoride with powdered calcium suifate and treatment of the combined precipitate with sulfuric acid. Such a mixture filters much more easily than a calcium fluoride-carbonate mixture. Starck and Thorin' suggested precipitation of calcium fluoride and oxalate.

+

+

+

Pogg. Ann., 1, 69; Sckweigg. Jour., 16 (1816). 426. Liebig's Ann., I2 (1849), 343; Ibid.. 1 9 , 115. anal. Ckem.. 43 (1904). 469. 4 Comfit. rend., 80 (1875). 561. 8 Monatsk. Ckem.. 26 (1907). 163. 8 Am. J. Sci., [4] 42 (1916). 464. 7 2. anal. Chem., 61 (1912). 14. 1 3

* Z.

+

1117

Heintz' precipitated calcium fluoride and phosphate, d e t a i n e d the weight of Caa(PO&, and got CaFI by difference. Kneeland' used Berzelius' method, but determined lime in the precipitate, JannaschS distilled hydrofluoric acid from the sample in a platinum retort and then used Berzelius' method on the neutralized distillate. Pisani' claimed excellent results by precipitating ThFr.qH20 and igniting to thorium dioxide. Starcks precipitated PbFCl with saturated lead chloride solution. Of the various methods referred to in this class that of Berzelius-Rose is the only one commonly used when phosphates are present. The method, however, is applicable only when the percentage of fluorine with respect to the phosphate present in the sample is not below a certain limiting amount. For the determination of small amounts of fluorine in phosphoric acid, the method was found to be entirely impractical and therefore could not be used in our work. 2 . ETCHING METHODS-The methods of this class are dependent upon the corrosive action of hydrofluoric acid on glass or quartz. Carles,' Ost,' and Woodman and Talbot* compared the extent to which a glass surface is attacked with that resulting from the use of a known amount of fluorine. Westerbergs and KobelllO determined the loss in weight of a watch glass exposed to the action of the hydrofluoric acid evolved from a sample. Many other minor modifications of apparatus and procedure have been proposed but none seem to take into account the effect of silica in the sample." When a sample containing silica is treated with acid a greater or less proportion of the fluorine, depending on the amount of silica present, will pass off as silicon fluoride. Therefore unless due precautions are taken for the initial removal of the silica a negative test may frequently be obtained even when considerable fluorine is present. Phosphoric acid prepared by the volatilization method always contains silica and to remove this constituent completely without danger of losing some of the fluorine would be very difficult if not impossible. The use of etching methods for the determination of fluorine in phosphoric acid was therefore considered impractical. 3. COLORIMETRIC METHODS-The Steiger12-Merwin13method depends upon the fact that fluorides bleach oxidized titanium solutions. Another method proposed by Gautier and ClausmannI4 depends upon the formation of lead fluoride and its transformation into lead sulfide. The turbidity caused by distilling hydrofluoric acid into a solution of a calcium salt has been recommended by PetersenI5 as a means of determining fluorine. The method of Steiger and Merwin is the best known of this class and gives good resultsx6for the determination of fluorine in the absence of such constituents as phosphoric acid, which have a bleaching effect on the titanium solution. All colorimetric methods, moreover, are applicable only when working with small amounts of fluorine and are therefore not of convenient application to the analysis of phosphate materials containing varying amounts of fluorine. 4. VOLATILIZATION METHODS-In this Class are to be found a Pogg. Ann.. 11, 267. Eng. and M i n . Jour.. SO (1905). 1212. 8 Z . anorg. Chem.. 9 (1895). 267. 4 Compt. rend., 162 (1916), 791. 8 Z. anorg. Chem.. IO (1911). 173. 6 Compt. rend., 144 (1907), 37. 7 Bn.. 26 (1893). 151. * J . Am. Ckem. Soc.. 26 (1906). 1437. 0 Chem. Ztg.. 26 (1902). 967. 10 J . prakf. Chem., [ l ] 92 (1864), 385. 11 The same oversight is apparent in many texts on the subject. as may be noted, for example, in the etching tests given in Scott's "Standard Methods of Chemical Analysis." 18 J . Am. Chem. Soc.. SO (1908). 219. Fenton, I. J . Chem. See also € SOL. 93 (1900). 1064. 18 ~ m J . sci.. [41 a8 (1909). 119. 14 Comgt. rend., 164 (1912). 1469, 1670, 1753. '6 Z. anal. Ckem.. 36 (18961, 585. 18 Adolph, J . Am. Chem. Soc.. 81 (1915), 2500. 1

3

T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y

111s

large number of methods and modifications dependent on the volatilization of fluorine as silicon fluoride. Oettel' measured the volume of silicon fluoride formed. Freseniusz observed the gain in weight of a tube filled with moist pumice when silicon fluoride was passed through it. Wohlera recommended volatilizing the silicon fluorideand noting the loss in weight of the flask and contents. Daniel's4 method is a combination of the two. Penfields distilled silicon Iluoride into water, added potassium chloride and alcohol, and titrated with ammonia the hydrochloric acid set free. Offermanns titrated the aqueous solution with potassium hydroxide using phenolphthalein. Off ermann's method has been modified somewhat in the manner of procedure by Drawe,' Hileman* and Ad01ph.~ Browning'O suggested decomposing silicon fluoride on moist black paper and noting the amount of silica formed. Korovaeff," Stadelerl*and Weinlandla observed the loss in weight due t o volatilizationof silicon fluoride. CarnotI4 passed silicon fluoride into a solution of potassium fluoride and weighed the precipitate of potassium silico-fluoride. Carnot's method was modified in some respects by Prost and BalthasarI6 and by Burk.I6 Somewhat similar methods had been proposed previously by Stolba,I7Liversidgel* and Tammann.19 Reinz0 collected and weighed the silica formed when silicon fluoride is decomposed by water. LasneZ1and SchneideF distilled silicon fluoride into water and after removing silica precipitated CaFz CaC03 and weighed the CaFz. Hileman23 made use of an entirely different action of HZSiFs when he titrated the iodine liberated in a solution of KI KIOs by fluosilicic acid.24

+

+

EXPERIMENTAL

T h e volatilization methods, unlike t h e methods of t h e preceding classes, are all applicable t o t h e analysis of phosphates as well as other materials. The methods of this class, however, which are based on a determination of loss in weight due t o t h e volatilization of silicon fluoride, are not suited t o t h e analysis of material containing a low percentage of fluorine. T h e principle of all t h e remaining methods consists in passing t h e volatilized silicon fluoride into water and then determining, either volumetrically or gravimetrically, t h e amount of hydrofluosilicic acid formed. This procedure should be applicable t o t h e determination of fluorine i n amounts varying between rather wide 1Z.

anal. Chem.. 26 (1886). 505. Ibid.. 6 (1866). 190. 8 Pogg. Ann., 48 (1839). 87. 4 Z . anorg. Chem.. 38 (1904), 257. 8 A m . Chem. J.. l ( 1 8 7 9 ) . 27; Z. anal. Chem.. 21 (1882). 120. 6 Z . angew. Chem., 3 (1890). 615. 7 Zbid.. 26 (1912), 1371. 8 Z . anorg. Chem., 6 1 (1906). 157; A m . J . Sci., [41 22 (1906). 329. 9 LOG. cit. 10 A m . J . Sci., [4] 32 (1911). 249. 11 J . prakt. Chem., 111 86 (1862). 442. 12 I b i d . . [ l ] 99 (1866). 66. 18 Z . anorg. Chem.. 21 (1899), 45. 14 Bull. SOL. chim., [3] 9, 71; Z . anal. Chem.. 36 (18961, 580; Comfit. rend., 114 (1892). 750. 10 Bull. assoc. belg. chim.. 13 (1899). 453. 16 J. A m . Chem. Soc.. 23 (1901). 825. 17 J . prakt. Chem., [ l ] 89 (1863). 129. 18 Chem. News, 24 (1871). 226. 19 Z . anal. Chem.. 24 (1885). 328; Z . physiol. Chem., 12 (1888). 322. 20 Z . anal. Chem.. 26 (1887). 733. $1 A n n . chim. anal.. 2 (1897). 182. 1%Oesterr. Z . Berg.-Hilttenw., 61, 365; C. A , . 7 (1913), 3582. A m . J . Sci., [4] 22 (1906). 383. 1 4 A volumetric method, belonging t o none of these four classes, developed by Guyot [Comfit. rend.. 71 (1870). 274; 7s (1871). 2731 and Greef [Ber.. 46 (1913). 25111 depends on the precipitation of NaaFeFe in neutral solution by standard FeCla. As modified by Bellucci [ A n n . chim. applicata, 1 (1914), 4411. it gives good results for solublefluorides. but is not applicable to the analysis of phosphates. since FeP04 mould be precipitated along with the fluorine compound. 2

Vol. 9, No.

12

limits. A comparative s t u d y was accordingly undertaken t o ascertain t h e most satisfactory details t o b e followed in t h e determination of fluorine according t o this general scheme. I n carrying out these experiments use was made of a special double-trapped generating flask, designed by t h e Victor Chemical Co., for t h e volatilization of silicon fluoride. This flask (C), shown in Fig. I , was called t o our attention by Dr. H . E. P a t t e n of t h e Bureau of Chemistry. The remainder of t h e equipment and t h e procedure followed in t h e preliminary experiments were in principle t h e same as t h a t used by Offermann,' Adolph,l a n d others. The sample t o be analyzed was mixed with a small proportion of ground quartz a n d then treated in t h e generating flask with 98.5 per cent sulfuric acid. A current of carbon dioxide washed with sulfuric acid served t o carry t h e silicon fluoride evolved in t h e generating flask through a U-tube filled with glass beads, t o remove accompanying fumes of sulfur trioxide. a n d t h e n through water in a large test-tube, where decomposition of t h e silicon fluoride took place as represented in t h e equation 3SiFd zHzO = zH2SiFs SiO?. The solution thus obtained was boiled t o expel dissolved carbon dioxide a n d sulfur dioxide, a n d t h e hydrofluosilicic acid determined both volumetrically b y titrating with standard sodium hydroxide, using phenolphthalein as indicator, a n d gravimetrically by precipitating with saturated lead chloride solution. Satisfactory results were sometimes obtained b y this procedure, b u t , as a rule, a n d particularly when working with small amounts of fluorine, t h e values obtained were very greatly in error. After numerous tests had been made, t h e principal sources of error were ascertained t o be as follows: ( I ) Glass beads were found t o be only moderately efficient in removing t h e sulfur trioxide fumes carried over from t h e generating flask. The result is t h a t a portion of t h e fumes which escape passes into solution in t h e absorbing tube and thus gives high results whether t h e hydrofluosilicic acid is determined volumetrically or gravimetrically. ( 2 ) Sulfur dioxide, which also escapes from t h e generating flask, is partly taken u p b y t h e solution in t h e absorbing t u b e a n d a portion there undergoes oxidation t o sulfuric acid. This gives high results i n t h e same way as with sulfur trioxide. (3) When t h e solution of hydrofluosilicic acid was boiled t o expel carbon dioxide a n d sulfur dioxide preparatory t o t h e determination of t h e fluorine either gravimetrically or volumetrically, loss of fluorine was found t o result when using t h e volume of solution ordinarily specified. (4) Any chlorides or nitrates in t h e sample analyzed will result in t h e evolution of t h e corresponding acid and its absorption in t h e solution of hydrofluosilicic acid. This will give high results when t h e fluorine is determined volumetrically.

+

+

M E T H O D S F O R T H E A B S O R P T I O N O F SULFUR T R I O X I D E

ANALYSIS-when t h e Sulfuric acid is boiled in t h e generating flask sulfur trioxide is carried I N FLUORINE 1 LOG. cit

Dec., 1917

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

over as a white fume which is difficult t o condense. T h e proportion retained b y t h e glass beads in t h e U-tube will depend on t h e r a t e at which t h e carbon dioxide is passed through t h e apparatus, b u t even when t h e flow of gas is as slow as it is expedient t o use, a white fume may still be seen t o issue from t h e end of t h e t u b e containing t h e glass beads. It was found, however, t h a t complete absorption of t h e fumes could be effected b y joining in the apparatus, in addition t o t h e t u b e containing glass beads, a straight piece of glass tubing ( F ) , filled with glass wool as shown in Fig. I. A perceptible coloration of t h e glass wool on heating t o 100-130’ will indicate the depth t o which t h e fumes have permeated. Care should be taken t o replace t h e t u b e before t h e glass wool becomes completely saturated. NETHODS FOR THE ABSORPTION O F SULFUR DIOXIDE

F L U O R I N E ANALYSIS--When t h e fluorine treated in t h e generating flask is considerable, t h e greater p a r t of i t passes over before t h e acid in t h e flask begins t o boil a n d before a n appreciable dissociation of the sulfuric acid takes place. It was also noted t h a t IN

1119

hydrochloric acid were used. Table I shows some results obtained when using t h e volumetric method of determining t h e hydrofluosilicic acid. It was observed, however, t h a t when organic matter t o t h e extent of 0 . I gram or more is contained in t h e sample analyzed, sulfuric acid then appears in t h e solution after boiling out dissolved gases, even though hydrochloric acid h a d previously been added. So much sulfur dioxide is evolved under these conditions t h a t t h e hydrochloric acid present cannot prevent t h e oxidation of a part of it. It is well known t h a t solutions of sulfur dioxide are rapidly oxidized by exposure t o air and some work has been done on preventive measures. Thus Saillard’ found t h a t a solution of sucrose exerted an inhibitive effect on this oxidation. Accordingly 3 0 per cent and 6 0 per cent sucrose solutions were made up a n d used in place of water in the absorption tube. T h e effect was very much less t h a n t h a t produced by hydrochloric acid. An oxidizing solutionZ was then prepared consisting of chromium trioxide in 98.5 per cent sulfuric acid. This was placed with a n excess of t h e trioxide in t h e

Bowen potash bulb ( E , Fig. I ) . When n o organic matter was used in t h e generating flask t h e quantity of sulfur dioxide t h a t escaped was found t o undergo such complete oxidation in t h e trioxide solution t h a t t h e amount carried over was too small t o give rise t o a n y appreciable quantity of sulfuric acid in t h e absorption tube. When t h e sample analyzed contains a large proportion of organic matter, however, so much TABLE I-RETARDATIONOF THE OXIDATION OF SULFUR DIOXIDEIN ACID SOLUTIONS sulfur dioxide is then evolved t h a t quick reduction Grams Cc. ”10 HCl -Cc. N/10 NaOHHzSO4 FORMED of t h e trioxide takes place. I n this case t h e most conNaF Added t o Equivalent of ReExpressed as Cc. Taken Absorption Tube N a F + HC1 quired of N/10 Solution venient procedure is first t o remove t h e greater part None None None 0.75 0.75 of t h e organic matter b y ignition a t a temperature 1.00 None 1.00 1.65 0.65 None 3.00 3.00 3.30 0.30 below t h e melting point of t h e material. T h e comNone 5.00 5.00 5.15 0.15 10.00 None 10.00 10.01 0.01 plete combustion of certain phosphate materials is, 10.00 10.00 None 10.06 0.06 0.0346 10.00 18.24 18.31 0.07 however, sometimes difficult and if in such cases some 0.0194 10.00 14.62 14.40 None carbonaceous material remains unburned in t h e t h e amount of SO:” present, for sulfurous acid is very sample, t h e sulfur dioxide evolved may be sufficiently SO8 . T h e larger great t o escape complete absorption in t h e chromium slightly dissociated into 2H+ p a r t of i t is in t h e form of H+ HSO3’ and t h e pres- solution. I n t h e analysis of such materials standard ence of a n acid would therefore force back t h e disso- hydrochloric acid solution should be used in t h e abHS0’3 H+ SOa.” Several de- sorption tube as a n added precaution. The effectiveciation of terminations were run using varying quantities of N/IO ness of t h e chromium trioxide solution in t h e absorphydrochloric acid in t h e absorption tube. It was tion of sulfur dioxide is shown by t h e results given in found t h a t t h e amount of sulfuric acid formed de- Table 11. creased steadily with increasing amounts of acid until 1 keo. gen. chim., 16 (1913). 412. practically none was formed when I O cc. of N / I O G . Tammann, Z. anal. Chem.. 24 (1885). 328. when using pure samples of sodium fluoride in t h e generating flask, little or no sulfuric acid was formed in t h e absorbing t u b e when t h e fluorine taken amounted t o 0.015 gram or more. From these observations it was thought possible t h a t t h e hydrogen-ion concentration affected t h e oxidation of t h e sulfur dioxide. This would be expected if t h e oxidation depends on

-

+ + +

T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y

I120

TABLE 11-ABSORPTION

OF SULFUR DIOXIDE M SULRURIC ACID OF CHROMIUM TRIOXIDE -

SOLUTION

Cc. N/10 HCl -Cc. N/10 NaOHHnSO4 FORMED SAMPLE Added to E uiv. of REQUIRED (Cc. N/10 Solution) TAKEN Abs. Tube Nag HCI +CrOa No CrO; +CrOa No CrOl None None None None 0.75 None 0.75 0.0214~ None 16.82 16.80 None NanSiFs

+

1

0.0050 g. NaF

Starch : g z t e : ,

...

1 None

1

10.00 10.00

1.19

...

10.84 10.84

10.87

...

2.00

ii:io

..

..

0.81

0.03

..

O:i6

LOSS O F F L U O R I N E I N BOILING SOLUTIONS O F HYDRO-

ACID-when proper precautions were taken t o prevent t h e formation of sulfuric acid in t h e absorption tube, low results for fluorine were frequently obtained, as Adolph' observed in his experiments with t h e Offermann method. This was finally traced t o a loss of fluorine on boiling t h e hydrofluosilicic acid solution t o expel dissolved carbon dioxide and sulfur dioxide. As t h e volume of t h e solution was increased, t h e loss of fluorine decreased, and beyond a N / ~ o o o concentration of t h e hydrofluosilicic acid no loss of fluorine took place, as shown by t h e results given in Table 111: FLUOSILICIC

Vol. 9 , No.

IO

nitric acid will be evolved a t first, b u t as t h e temperature is raised, this will decompose t o give nitrogen dioxide. I n either case t h e product evolved will be taken up by t h e solutions in t h e Schmitz and Bowen tubes a n d consequently no nitric acid will be formed in the absorption tube as shown by t h e results given in Table IV. DETAILS O F P R O C E D U R E I N T H E M O D I F I E D M E T H O D O F DETERMINING FLUORINE

APPARATUS-The equipment finally adopted in t h e analysis of fluorine is represented in Fig. I. A is a cylinder of compressed carbon dioxide or nitrogen fitted with a reducing valve or other safety device for regulating the flow of gas. B B are wash bottles containing concentrated sulfuric acid for washing t h e gas. C is a generating flask of 2 5 0 cc. capacity provided with o traps containing 9 8 . 5 per cent sulfuric acid, t h a t next the flask being half full while the acid in t h e second is just sufficient t o make a seal. D is a Schmitz tube having a I O per cent solution of silver sulfate in 9 8 . 5 per cent sulfuric acid in t h e bulbed arm of t h e TABLE111-Loss OF FLUORINE ON BOILINGA SOLUTION OF HYDROFLUO-tube and glass beads in t h e other. E is a Bowen bulb snIcIc ACID containing a saturated solution of chromium trioxide Cc. Acid -Cc. N/10 NaOHSolution Boiling Period Equivalent to Rein 9 8 . 5 per cent sulfuric acid. F is a straight piece Taken Min. HZSiFs Taken auired 100 10 11.14 10.95 of glass tubing filled with glass wool. G is a n absorp100 15 11.14 10.85 tion tube consisting of a large test-tube containing 100 20 11.14 10.75 150 20 11.20 10.96 about jo cc. of water. 175 20 11.20 11.13 20 11.20 11.18 250 On account of t h e pressure t h a t must be generated DETERMINATION O F FLUORINE I N PRESENCE OF in the apparatus t o cause a flow of gas against the head C H L O R I D E S A N D NITRATES-Ill most phosphatic ma- of sulfuric acid, great care must be taken in assembling terials a limited amount of chlorides is usually pres- t h e apparatus t o make all joints tight. Ends of glass ent. On treating such material in the generating tubing must be brought together and the rubber t u b flask hydrochloric acid would be evolved along with ing used should be a good grade of heavy-walled gum t h e silicon fluoride and would be taken up by the solu- tubing. As a precaution against diffusion of gas tion in t h e absorption tube. This would interfere through the joints t h e rubber tubing may be covered with the volumetric determination of the hydrofluo- with a good coating of shellac. Glass stopcocks silicic acid. With a view t o preventing any hydro- should be paraffined also. An asbestos screen may be chloric acid reaching t h e absorption tube, t h e gases used under t h e flask and t o minimize t h e danger of from t h e generating flask were made t o pass through breaking t h e flask a ring of asbestos paper should be a Schmitz tube (D,Fig. I ) containing a solution of placed around it for protection against cold air cursilver sulfate in 98.5 per cent sulfuric acid. Glass beads rents. A mantle of asbestos over t h e bulb of t h e flask were placed in t h e small arm of t h e tube and as thus must also be used t o prevent burning t h e rubber conequipped it took t h e place of the U-tube used in t h e nection a t t h e t o p of t h e flask. The exclusion of preliminary experiments. T h a t a solution of silver every trace of water from all parts of t h e apparatus sulfate in sulfuric acid is able t o absorb effectively is essential. The following additional reagents are a n y hydrochloric acid given off from t h e generating required: 0ask without interfering in any way with t h e de(I) Anhydrous copper sulfate. termination of fluorine is shown by t h e results given (2) Ground quartz. in Table IV. I n carrying out these experiments t h e (3) 9 8 . 5 per cent sulfuric acid prepared by boiling acidity of t h e solution in t h e absorption tube was de- C. P. concentrated sulfuric acid for about 20 minutes termined volumetrically after t h e precautions already in a n open vessel. noted were taken t o prevent the formation of sulfuric (4) A I O per cent solution of silver sulfate in 9 8 . 5 acid or loss of fluorine through boiling. per cent sulfuric acid. The silver sulfate should be When nitrates are present in t h e sample analyzed, ignited with a n excess of sulfuric acid before using TABLEIV-DETERMINATION OF FLUORIDE I N PRESENCE OR CHLORIDES t o drive off any volatile acids present. AND NITRATES ( 5 ) A saturated solution of chromium trioxide in -Cc. N/10 NaOHGRAMSREAQENT TAKEN Equivalent to Re9 8 . 5 per cent sulfuric acid. NaF NaNOa NaF Taken quired NaCl (6) N / I O sodium hydroxide, preferably standard0.2 None None 0.05 None 0.0093 0.05 None 2.21 2.22 ized against chemically pure sodium fluosilicate. None None 0.2 None 0.0 ( 7 ) N / I O hydrochloric acid, carefully standardized 0.0050 0.5 1.19 1.20 None against t h e N / I O sodium hydroxide. 1 LOG. cit.

Dec., 1917

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 CHEMISTRY

I121

DETERMINATIO?i-The sample t o be analyzed must it will be found unnecessary t o add standard hydrobe carefully dried. A quantity containing between chloric acid t o the water in the absorption tube, pro0.OOI and 0.I g. of fluorine is weighed into the generatviding the fluorine in t h e sample is equivalent t o IO ing flask with 0.1-1.0 g. of silica and 5 g. of anhydrous CC. or more of N/IO acid. If t h e fluorine taken is copper sulfate: IOO cc. of 9 8 . 5 per cent sulfuric acid are less t h a n this, or if there is any possibility of organic then added, allowing t h e first portions t o run into t h e matter being present, then IO cc. of N/IO hydrochloric traps, and t h e flask quickly connected t o t h e source acid should be added as a precaution against oxidaof carbon dioxide. The valve of t h e carbon dioxide tion of sulfur dioxide t o sulfuric acid. T o insure t h a t cylinder is adjusted t o give a flow of gas through t h e no error has been introduced from this source, t h e soluabsorption bulbs a t a rate of about 2 or 3 bubbles per tion, after titrating, may be concentrated t o about 5 0 sec., and this rate should be maintained t h e same cc., acidified with hydrochloric acid and a small throughout t h e course of the experiment. The con- amount of barium chloride solution added. I n case tents of t h e flask are shaken until well mixed and then a precipitate of barium sulfate occurs, this may be filheated gradually t o boiling. At this point a white tered off, ignited and weighed, and t h e proper deducscum indicating fluorine will appear on t h e inside of tion made from the volume of sodium hydroxide ret h e flask. The flame under the flask should then be quired. adjusted so t h a t t h e condensing sulfuric acid will When undertaking a second determination, a portion wash this scum freely and completely into t h e first of t h e acid in t h e t r a p equivalent t o t h a t which distrap. Care should be taken, however, t o avoid heat- tilled over in t h e preceding determination must be reing so strongly t h a t white fumes will be evolved in turned t o the flask. This may be done by removing noticeable quantity, nor should the acid in t h e first t h e stopper in t h e Schmitz tube a n d simply lowering t r a p be made t o boil. When the second t r a p is half t h e flask into the proper position. Whether or not full of acid t h e heating is discontinued. On turning t h e contents of the flask should be changed with each off the flame, particular pains must be taken t o regulate determination will depend on the character of the samt h e Aow of gas so t h a t t h e relatively cool acid from ple taken for analysis. Using samples containing litt h e traps does not flow back into t h e flask and thus tle carbon and weighing z g. or less, t h e same acid may cause it t o break. The valve in the carbon dioxide be used in making several determinations. I n this cylinder must therefore be sensitive enough t o admit case the sample t o be analyzed, mixed with the proper of quick adjustment. After a minute or two of ad- amount of silica, is placed in a thin-walled glass capjustment t h e valve may be set t o give as before a uni- sule, dropped into t h e flask and t h e latter then quickly form flow of carbon dioxide, which is continued from connected with t h e source of carbon dioxide a s before. 2 5 t o 30 minutes, so as t o wash all silicon fluoride into I n this way little time is lost between successive det h e absorption tube. T h e latter is then removed, its terminations. The copper sulfate may be renewed contents transferred t o a 3 5 0 cc. Erlenmeyer flask, with each change of acid. I t s purpose is t o serve t h e solution made up t o 2 0 0 - 2 5 0 cc. and boiled gently as a dehydrating agent and t o prevent bumping. for IO t o 15 minutes t o expel dissolved gases. This The accuracy of the method described a n d its range operation should consume as little time as possible, of application is shown by the results given in Table because any prolonged exposure t o t h e air before boil- V. For t h e sake of comparison some results obtained ing leads t o oxidation of t h e sulfur dioxide in solution. with t h e Berzelius-Rose method are also given. The solution is allowed t o cool somewhat and then TABLE V-RSSULTS OBTAINED IN THB ANALYSIS OF FLUORINE COMPOUNDS titrated with N / I O sodium hydroxide, using phenolPBR CBNT FLVORINSIN SAMPLB Cc. N/IO HCI Used in Modified Berzeliusas phthalein as indicator. T h e end-point must be apSample Grams Absorption Volatilization Rose CalcuUsed Taken Tube Method Method lated proached slowly but is fairly sharp when the fluorine 45.63 ... 45.24 45.24 0.0346 10.00 present amounts t o less t h a n 0 . 0 2 g. The reaction N a F ........... . . 0.0202 44.49 ... 10.00 45.06 45.24 0.2000 None occurring is represented by the equation 0.2000 ... ... 4j:il 45.24 ... 42.95 45.24 0.2000 ... HzSiFs 6NaOH + 6 N a F 2 H z 0 Si(OH)4. 4 . 4 8 ~. 4.52 0.4614 ... 4.52 4.52 ... 0.4220 It will require about a n hour t o make one determina4 .52 4.54 ... 0 5000 4 .52 4 . 4 3 ... 0.5000 None tion. 4 . 52 2 . 9 2 1 .0000 ... ... 4.52 2.62 1.0000 If a n appreciable amount of organic matter is pres0 . 4 52 0 438 1o:oo 1. IO35 0.452 0.434 e n t in the sample t o be analyzed, it must first be reI . I610 10.00 0 . 4 5 2 0: i o ... 1.0000 ... moved by burning. This is done by adding sufficient 0.452 0.18 1 . 0000 0 . 0 4 5 o:t)37 1o:oo ... 5.0000 sodium carbonate t o make the material alkaline and 0.045 0,037 10.00 ... 5 moo 0,009 0 . 0 1 5 10.00 10.0000 then igniting in a muffle furnace a t a dull red heat NarSiFs + 0.2341 10.00 6 . 0 14 ... 6.05 6.05 until a white, or nearly white, ash remains. Care must Car(P0d)s) 0.2647 10.00 be taken t h a t t h e temperature does not get high enough a t any time t o melt t h e material or t o volatilize I n Table VI are given Some results obtained seParately by each of us in the analysis of some commercial sodium fluoride. t h e determination of fluorine in phosphoric acid materials which are ordinarily supposed t o Offer Special t h e latter must first be prepared for analysis by making difficulty in t h e determination of fluorine. neutral t o phenolphthalein with sodium hydroxide An examination of t h e results given in Tables a n d V I will show values obtained for the fluorine in a n d then drying. I n t h e analysis of samples free from organic matter, different materials ranging from 0.01 to 4 5 . 2 per cent.

+

+

+

~~

v

T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY

I122

I n general, the method may be expected t o give results accurate t o within o . 0005 g. of fluorine, and, if organic matter is carefully excluded, even greater accuracy may be secured. The lowest limit for t h e determination of fluorine t o which t h e method is applicable may be taken as 0 . 0 1 per cent. The concentration of fluorine which admits of most convenient analysis ranges from about 0 . 5 t o I O per cent and it is between these limits t h a t the most accurate results are usually obtained. I n t h e analysis of high-grade material like sodium fluoride a slight error in weighing may produce a considerable variation in t h e results. The source of error may be minimized by diluting t h e material with a known amount of a n inert substance like calcium phosphate, or by taking a sample containing a larger amount of fluorine t h a n t h a t ordinarily used. I n TABLEVI-MISCELLANEOUSFLUORINE ANALYSES Cc. N/10 HC1 PER CENT Added to FLUORINE FOUND

SAMPLE Phosphate R o c k . .

Absorption Tube

................. None 10.00 Phosphoric A c i d . . ................. 1 0 . 0 0 None Monocalcium Phosphate (commercial) 1 0 . 0 0 10.00 10.00 10.00

Baking Powder..

.................

Sample A 1.58 3.74 0.40 0.67 0.63 0.031 0.014 0.043

Sample B 1.47 3.85 0.47 0.70 0.73 0.032 0.023 0.040

the latter case care must be taken t o use a delivery cube sufficiently flared a t t h e point where it meets the surface of t h e water in t h e absorption tube t o prevent its being stopped up by the silica t h a t is deposited. T h e sodium fluoride used in testing t h e accuracy of t h e method as represented in Table V was prepared by neutralizing C. P. sodium bicarbonate with an excess of hydrofluoric acid, igniting t o drive off the excess of acid and then recrystallizing several times. The purity of the product finally obtained was verified by changing it into t h e sulfate and noting t h e increase in weight. The sodium fluosilicate used was prepared from “Baker’s analyzed” product by heating t o 400’ for 5 minutes, filtering off t h e insoluble residue, recrystallizing three times in a large platinum dish and finally drying a t 130’. The purity of t h e product thus prepared was established by titrating with a standard solution of sodium hydroxide. DETERMINATION

OF

HYDROFLUOSILICIC ACID GRAVI-

METRICALLY-The results so far given in t h e analysis of fluorine are all based on t h e volumetric determination of t h e hydrofluosilicic acid in the absorption flask. If desired, t h e results thus obtained may be checked by means of a gravimetric determination of t h e acid. I n carrying out this procedure t h e solution obtained after titration with standard sodium hydroxide is concentrated t o about 50 cc. and t h e fluorine precipit a t e d by adding 2 0 0 cc. of a saturated solution of lead chloride, as directed in Starck’s method. After standing over night t h e precipitate is filtered through a Gooch crucible, washed with saturated lead chloride solution or, better, with saturated lead fluochloride, and finally with a few cc. of water, dried a t 1 2 0 ’ and weighed as PbFCl. I n Table VI1 t h e results obtained by this method are compared with those found volumetrically.

Vol. 9 , No. 1 2

The values given in Table VI1 show t h a t when t h e weight of fluorine taken is below 0.01 g. t h e gravimetric method gives lower results t h a n t h e volumetric TABLEVII-COMPARISON OF RESULTSOBTAINEDIN

THE VOLUMETRIC GRAVIMETRIC DETERMINATION OF HYDROFLUOSILICIC ACID GRAMSFLUORINE GRAMSFLUORINE GRAMSFLUORINE No. Vol. Grav. No. Vol. Grav. No. Vol. Grav. 1 0.02395 0,02410 5 0,00818 0,00798 9 0,00467 0,00462 2 0.02270 0.02295 6 0.00762 0.00758 10 0.00388 0.00382 7 0.00589 0,00498 1 1 0.00306 0.00295 3 0.02185 0,02235 12 0,00185 0.00170 4 0.00988 0,00982 8 0,00560 0.00457 AND

and where t h e weight of fluorine is over 0 . 0 2 g. the gravimetric results are usually higher. Although not definitely proved, it seems probable t h a t this result is due t o two factors, t h e solubility of lead fluochloride and the adsorption of t h e silicic acid set free by t h e titration of the hydrofluosilicic acid. For small amounts of fluorine the solubility of t h e lead fluochloride more t h a n offsets t h e adsorption of silicic acid, b u t for larger amounts t h e reverse is true. This appears reasonable from t h e fact t h a t t h e weight of lead fluochloride remaining in solution will be practically t h e same whether little or much fluorilie is present, but t h e weight of silicic acid carried down will increase with increasing concentration of fluorine. For a n intermediate amount varying between 0.0075 g. and 0.01jo g. of fluorine, the two errors practically balance each other and within these limits results by t h e two methods closely agree. The procedure of precipitating t h e fluorine with saturated lead chloride solution may thus be used as a check on t h e results obtained by t h e volumetric method, and it serves as a visible means of identifying fluorine in solution. I n t h e case of routine analyses, however, it is apparent from t h e results given t h a t there is nothing t o be gained by lengthening t h e method beyond t h e volumetric determination of t h e hydrofluosilicic acid. ACKKOWLEDGMENT

The possibilities of adapting a volatilization method of determining fluorine in t h e presence of phosphates was called t o our attention by Dr. H. E. Patten, Referee for baking powders of the Association of Official Agricultural Chemists, and we wish fully t o acknowledge t h e assistance which he and also t h e Victor Chemical Co. have cheerfully given in t h e course of this investigation. STJMMARY

An experimental study has been made of t h e various processes t h a t have been proposed for t h e determination of fluorine in t h e presence of phosphates. From t h e results obtained it would appear t h a t t h e procedure which offers most promise for the determination of fluorine over a wide percentage range consists in volatilizing t h e fluorine as silicon fluoride a n d then collecting t h e latter in water t o form hydrofluosilicic acid. When sulfuric acid is used t o bring about t h e volatilization of t h e silicon fluoride there is evolved a t t h e same time sulfur trioxide, sulfur dioxide a n d also such other products which may be present as hydrochloric and nitric acids. It was found, however, t h a t by use of various reagents a selective removal of these products from t h e silicon fluoride may be effected, yielding a solution of hydrofluosilicic acid en-

Dec., 1917

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

tirely free from other acid constituents. By titrating this solution with standard sodium hydroxide, using phenolphthalein as indicator, closely agreeing results for t h e fluorine in t h e samples analyzed may be obtained. I n t h e case of samples freed from water a n d from organic matter (by burning) a complete analysis b y this procedure may be made in t h e course of a n hour. T h e method is applicable t o t h e analysis of material having a fluorine content as low as 0.01 per cent. BUREAUOF SOILS

u. s. DEPARTMENT OF AGRICGLTURE D. C. WASRINGTON.

A WOOL FAT (LANOLIN) SUBSTITUTE AND THE PREPARATION OF CETYLIC ALCOHOL B y SOL. AXELRAD Received October 17, 1917

T h e question of substitutes for various materials has been one of prime importance during these war times, especially those substances a n d chemicals used in medicine a n d pharmacy. A year a n d a half ago t h e supply of wool fat was very limited a n d t h e price asked was four times more t h a n t h a t under normal conditions. A substitute called “Eucerin” imported from Germany was also scarce a n d t h e agency for this product had only four ounces left. I t was claimed t h a t “Eucerin” was made from t h e washings obtained in t h e manufacture of wool fat. The uses of lanolin are many, especially in pharmacy; as a vehicle for ointments, in t h e preparation of bougies, suppositories, cold creams a n d plasters of various kinds, etc. This investigation on a wool f a t substitute was undertaken with t h e idea of making a f a t t y composition which would have all of t h e desirable properties of lanolin, such as body, tenacity, power of absorbing water readily, taking up solutions of various chemicals used in pharmacy, dry powders, etc., etc. Liebreich claimed that t h e absorbing power of wool fat was due t o t h e cholesterin ethers i t contained. Lifschuetzl isolated t h e cholesterin ethers of Liebreich a n d proved t h a t they h a d very little power of absorbing water. He concluded from his experiments t h a t t h e absorbing power of wool fat was due t o t h e f a t t y alcohols of is0 a n d oxycholesterols. He separated from t h e alkaline washings of partially saponified wool fat two saturated alcohols and one unsaturated alcohol. His experiments further proved t h a t t h e more purified lanolin was, t h e lower was its power t o absorb water, owing t o t h e fact t h a t during its purification t h e is0 and oxycholesterols were partly removed. Unna,* in his paper “Ointment Bases,” states t h a t “Eucerin” is a mixture of alcohols of t h e is0 and oxycholesterin group with petrolatum. He does not, however, give a commercial method for t h e preparation of these alcohols. T h e absorbing power of “Eucerin” is due t o these alcohols, b u t Roemer (see below) claims t h a t with t h e employment of cholesterols for ointment bases, hydrocarbons, such as mineral oil, benzol, etc., are essential, for i t is due t o t h e m in combination t h a t t h e absorbing property is imparted. Roemer,3 in his Ber.. 29, 2890. J . Am. Pharm. Assoc., 1 , 673. a Ibid., 2, 971.

1123

paper on “ T h e Pharmacy of t h e Oxycholesterine Ointment Bases,” confirms t o a certain extent t h e work of Unna, b u t he also fails t o give a method for t h e preparation of t h e alcohols. I;nnal states t h a t “Eucerin” has been used in skin preparations in Germany. especially in pure form, for ichthyosis. According t o t h e U. S. Dispensatory, 19th Edition, page 97, experiments have been carried on in reference t o the absorption of wool fat by t h e skin. Patschkowski and Kaspar claim t h a t t h e skin readily absorbs lanolin, b u t Ritter a n d Pfeiffer in a long series of experiments were unable t o verify these results. Grimm% recommends, in his paper “Ceber die Verwendung von Aethal in der Hautpflege,” cetyl alcohol, C16Hs10H, for skin preparations on account of its absorption by t h e skin and he further states t h a t he has found i t useful in t h e treatment of prurigo, weeping eczema and other skin infections. I n view of t h e fact t h a t Unna and Roemer have writt e n about t h e cholesterols, which are aliphatic higher alcohols, and Grimm has given a favorable report on t h e use of cetyl alcohol. t h e writer came t o the conclusion t h a t t h e use of this alcohol was advantageous in a substitute for wool fat. , A review of t h e literature failed t o show any commercial method for t h e preparation of cetyl alcohol. There were many references as t o its preparation from spermaceti b y saponification with caustic potash a n d shaking t h e aqueous soap solution with petroleum ether, this being analogous t o t h e extraction of unsaponifiable matter. Spermaceti is essentially t h e cetylic ester of palmitic acid. Chevreul in 1818 isolated t h e alcohol by t h e above method. Krafft3 prepared this substance by t h e reduction of palmitic acid t o t h e aldehyde and heating it with barium formate. He also made t h e alcohol4 a n d by heating t h e palmitic aldehyde with zinc dust and acetic acid and hydrolyzing t h e acetate formed. Levene5 made this alcohol by t h e reduction of ethyl palmitate with sodium and absolute alcohol. Schorlemmer6 distilled a dry mixture of barium oxide with sebacic acid. A method given in several text books for t h e preparation of cetyl alcohol was t h e saponification of spermaceti with alcoholic potash, evaporating t h e alcohol, taking u p t h e residue with water, adding calcium chloride solution t o form calcium soap and extracting with suitable solvents. The above methods are useful for preparing small quantities of cetyl alcohol, b u t for commercial quantities these methods fall down for obvious reasons. T h e method of Schorlemmer was found impracticable on account of sebacic acid not being a commercial substance. The extraction process causes a considerable loss of t h e solvent employed a n d formation of troublesome emulsions. It has been found t h a t cetyl alcohol distills a t about 340 t o 350’ C. without decomposition and t h e writer’s 1

1

4

2

5

6

Proc R o y SOL.M e d , L o n d o n , 1911-1912, p. 220. Dnmalologische Z., 6, 158. Be7 , 13, 1416. Ibzd., 16, 1721; 17, 1627. J. Biol Chem., 20, 121. Ber., 3, 616.