392
ANALYTICAL EDITION
for water cooling, which gives additional protection to the de Khotinsky seal. OPERATION OF FURNACE
-
The furnace operates for all except the highest vacuums without using the mercury seal. No difficulty was encountered in pumping out to a pressure of 0.002 mm. of mercury a t room temperature with a good oil pump. I n the neighborhood of 1000° C., which is about as high a temperature as may be reached safely with the copper block, the apparatus pumped down to 0.1 mm. of mercury after small amounts of impurities in the copper had been removed by distillation and the original rapid degassing of metal and alundum cement had diminished. The heating unit and monel metal shields proved to be entirely satisfactory, 4.5 and 13.8 amperes through the 8-ohm coil giving, respectively, temperatures of 412" and 906" C. The experience with this furnace has definitely proved the great value of radiation shields. Monel metal was used because it retains a good reflecting surface after being heated in vacuo to temperatures around 1000" C., and has a relatively low thermal conductivity. On dismantling the apparatus, the outer shield had the appearance of never having
Vol. 4, No. 4
been hot, which is substantiated by the fact that the temperature of the cooling water was only slightly higher on leaving the furnace than on entering. I n reducing the heat losses, the radiation shields also protect the de Khotinsky seals, which were a great source of trouble in a previously used furnace which was not equipped with radiation shields completely surrounding the heater unit. The method of bringing out lead wires also proved very satisfactory and is considered a marked improvement over methods formerly employed. Fitting T was constructed to give good thermal contact between copper cylinder V and copper tube U,which is surrounded by the cooling water. This construction, with the cooling unit 2, makes practically certain the permanence of the deKhotinsky seal in well Y . ACKNOWLEDGMENT The success of this furnace is largely due to designs contributed by C. C. Maier, supervising engiqeer of the Pacific Experiment Station, and to the excellent workmanship of C. M. Bell, mechanician a t the Pacific Experiment Station. RH~CEIVED April 30, 1932. Published by permission of the Director, U. 8. Bureau of Mines. (Not subject to oopyright.)
Determination of Fluorine and Boron in Organic Compounds DANIELJ. PFLAUM AND HERMAN H. WENZKE,University of Notre Dame, Notre Dame, Ind.
I
N SOME of the researches being carried on in this department on organic boron trifluoride compounds, the problem of finding a dependable method of analysis for both fluorine and boron arose. A number of more or less unsuccessful attempts to do this have been made in the past. Kraus and Brown (4) prepared various amino derivatives of boron trifluoride, and analyzed them for fluorine, boron, and nitrogen, in establishing their structures. Their analyses for fluorine and boron indicate the need of a more reliable method. Bowlus and Nieuwland ( 1 ) were able to analyze for boron with an average error of about 6 per cent, by decomposition of the sample through continued heating with fuming nitric acid in a sealed tube, and titrating the boric acid formed with base and mannitol, according to the standard procedure (6). These authors did not report fluorine on the analyses. Vaughn and Nieuwland (10) presented a new method of analysis for organic fluorides, using sodium in liquid ammonia, but boron was not present in their compounds. There are a number of difficulties peculiar to the analysis of compounds containing boron and fluorine. Fusions of the samples with alkali carbonates (6) proved to be unsuccessful, as there was difficulty in decomposing the compounds, and there was evidently some loss through volatilization. It was finally decided to destroy the organic matter by combustion in a Parr sulfur bomb with an oxidizing mixture of sodium peroxide, potassium chlorate, and sugar. Most of the organic boron trifluoride compounds are deliquescent fuming substances, liquids or solids, which interact instantly with the fusion mixture. Recourse was had to weighing them in gelatin capsules and placing them in the bomb, thus inclosed. The capsules were completely destroyed by the ignition and did not interfere with quantitative combustion of the compound.
When the products of the fusion were dissolved in water, a solution containing a high concentration of alkali hydroxide and carbonate was obtained, and these interfered with the precipitation of calcium fluoride, When an attempt was made to remove them by treatment with hydrochloric acid, low results for fluorine were invariably obtained. According to Mellor (7),it is possible that the stable fluoborate ion was formed and not completely broken up by the weakly alkaline solutions permissible for the precipitation of calcium fluoride. Whatever the explanation, any appreciable concentration of hydrogen ion resulted in low values for fluorine. Although acetic or other weak acids are suitable for destroying hydroxide and carbonate without causing too high hydrogen-ion concentration, they interfere with the subsequent titration for boron. The method finally chosen was to destroy the carbonate and alkali by boiling the solution with ammonium chloride. Some trouble was had in the proper coagulation of the precipitate of calcium fluoride, but this was overcome by the use of Fisher "filter accelerators." It was found that calcium fluoride was filterable when these were used, even without the addition of ammonium hydroxide, acrording to the method of CarriBre and Rouanet ( 2 ) . Coprecipitation of fluoride and carbonate (9) was unnecessary, and has been attacked recently on theoretical grounds by Mougnaud and others (8). Clarke and Bradshaw ( 3 ) have shown that the calcium fluoride method can be made to give accurate results by substituting these paper-pulp filter-aids for the coprecipitation and subsequent washing with acetic acid. The next difficulty occurred through the presence of the excess ammonium chloride in the filtrate containing boron. Ammonium ion causes the end point to appear too late, through buffer action. It was removed by the addition of
October 15, 1932
INDUSTRIAL
AND E N G I N E E R I N G CHEMISTRY
strong sodium hydroxide solution and boiling until the vapors ceased to affect litmus paper. The calcium ion in excess from the calcium fluoride precipitation, which partially precipitated as hydroxide, was removed by filtration, the solution transferred to a volumetric flask, and aliquot portions titrated by the standard method, using mannitol ( 5 ) . Good results were also obtained for boron. The experimental procedure as finally decided upon was as follows: Approximately half-gram samples were accurately weighed in No. 00 gelatin capsules and placed in the cup of a Parr sulfur bomb of the electrical ignition type, together with a fusion mixture of 10 grams of sodium peroxide, one gram of potassium chlorate, and 0.5 gram of sugar. The bomb was tightly closed, well shaken, and shorted with direct or alternating current of about 20 volts and 8 amperes. For the fusion wire, standard iron wire “for analysis” was used in order not to introduce additional impurities. The bomb was permitted to become quite hot, and then immersed in cool water until it could be handled. The cup was then removed, and the contents dissolved out by placing it on its side in a 400-cc. beaker, covering it with water, and warming gentli. After boiling for a few minutes to coagulate the heavy metallic hydroxides from the cup and the fusion wire, the solution was filtered. Fifteen grams of ammonium chloride were roughly weighed and introduced into the solution, which was boiled until the odor of ammonia no longer persisted (one to two hours). If a new precipitate formed during the boiling, it was filtered out before proceeding. To the hot solution, 10 cc. of a 2 N solution of calcium nitrate were added drop by drop, with stirring. This provided a large excess of calcium ion to depress the solubility of the precipitate. A filter accelerator was dropped in and macerated well, and the boiling was continued for a few minutes. Occasionally it was necessary to add one cc. of approximately 3 N ammonium hydroxide to assist in coagulatc ing the precipitate. The solution was cooled in running water until cold, and filtered, using strong suction and a platinum filter cone. The filtrate was reserved for the boron analysis. The volume of wash-water should not exceed 50 cc. The paper containing the precipitate was dried a t 110” C. and ignited to constant weight in a platinum crucible over a Fisher burner. To the filtrate was added sufficient 4 N sodium hydroxide to produce a precipitate of calcium hydroxide and carbonate. The solution was then boiled until the vapors
393
no longer affected red litmus paper, and filtered into a 250-cc. volumetric flask. After being diluted to the mark and mixed thoroughly by shaking, a 50-cc. portion was withdrawn. This was titrated to neutrality with methyl orange, using 0.1 N hydrochloric acid. The remainder of the solution was neutralized with an equivalent amount of hydrochloric acid, but no methyl orange was added. The necessary amount of mannitol was added, together with a few drops of phenolphthalein, and the solution titrated to the first faint tinge of pink with 0.1 N sodium hydroxide solution (carbonatefree). The buret reading was calculated to boron. Using this procedure, the following results were obtained: ALIQUOT
QAafPm
Gram
NaOH SOLUTION Amt. B B used FOUND CALCD.
F F PART NorCaFz FOUND CALCD.TITRATED mality Gram % ’ % METHYL ACETATE B O R O N TRIFLUORIDE
0.5398 0.4446 40.12 40.19 0.5057 0,3716 o:iioo 40:i4 40:i9 0.5861
....
...
...
iji
o:iiis o:iiis
476
Cc.
%
%
(CBSCOOCBx:BB8) 2i:is 2i:io
i:io i:ii
i:ii7
+:ii7
ETHYL ACETATE BORON TRIFLUORIDE (CHSCOOCZHS:RFS)
MONOAMMONO B O R O NTRIFLUORIDE
0.2540 0.2976
0.3510 67.32 67.18 0,4106 67.20 67.18
:$:
(NHSBF~
0.1258 18.90 12.66 12.75 0.1268 22.30 12.75 12.75
These compounds were freshly made and purified by crystallization or distillation immediately before the analysis. The three are all solids, easy to obtain in a pure condition, and were chosen for that reason. LITERATURE CITED (1) Bowlus and Nieuwland, J. Am. Chem. SOC.,53, 3835 (1931). (2) Carrikre and Rouanet, Compt. rend., 189, 1281 (1929). (3) Clarke and Bradshav, AnuZUst, 57, 138 (1932). (4) Kraus and Brown, J. Am. Chem. SOC.,51, 2690 (1929). (5) Mahin, “Quantitative Analysis,” p. 205, McGraw-Hill, 1924. (6) Marignac, Chem. Cent., 39, 490 (1863). (7) Mellor, “Inorganic and Theoretical Chemistry,” p. 125, Longmans, 1924. (8) Mougnaud, Compt. rend., 194, 1507 (1932). (9) Rose, Ann., 72, 343 (1849). (10) Vaughn and Nieuwland, IND. EXQ.CHEM.,Anal. Ed., 3, 274 (1931).
RECEIVED June 3, 1932. From a thesis submitted by Daniel J. Pflaum to the Graduate School of the University of Notre Dame in partial fulfilment of-the requirements for the degree of master of science.
A Gas Buret for Catalase Apparatus DEANA. PACK,The Birdseye Laboratories, Gloucester, Mass.
T
HE buret described is believed to be of some general service as well as being an improvement for the catalase apparatus devised by Appleman (1). This buret was constructed so as to be placed in the water bath for maintaining n constant temperature for the gas to be measured. The apparatus consists of a short length of buret, a parallel reservoir tube, a siphon which allows the leveling bottle to be operated from without the bath, and connections for the reaction bottle and the outlet. The lower ends of the buret, B, and reservoir tube, R, are drawn out into tube, S, which serves as the apparatus support and forms one arm of the siphon connecting with the leveling bottle. The upper ends of B and R are drawn out into tube A , which connects with the reaction bottle. Tube A is also provided with an outlet tube, C. The buret with its reservoir tube is calibrated in
A’
centimeters for the readings on the buret. This apparatus permits all the chambers containing gas to be submerged in water. It must be kept in mind that as the buret is shortened and the volume of the r e s e r v o i r is increased, the errors of reading are increased in proportion. This arrangement has been in service for one year with good results. LITERATURE CITED (1) Appleman, C. O., Bot. Guz., 50, 182-92 (1910).
R E C Z I V ~August D 2, 1932.