October 15, 1932
INDUSTRIAL AND ENGINEERING CHEMISTRY CALCD.
solve soiuble gold salts in dilute sulfuric acid, add 3 to 5 cc. of hydrogen peroxide, and warm. The precipitated mass of gold is washed with water, then with alcohol, dried a t 105" C., and finally ignited. The results are as follows: CALCD.Au Gold sodium thiosulfate
FOUND
%
%
37.0
37.0
GERMANIUM The germanium compound is oxidized by sulfuric acid and hydrogen peroxide in a Kjeldahl flask, the sulfide precipitated from strongly acid solution, filtered as soon as coagulation is complete, and dissolved in strong ammonium hydroxide. This ammonia solution is filtered directly into a large crucible, the sulfide decomposed by Superoxol, evaporated to dryness, and ignited to germanium dioxide as in the method of Johnson and Dennis (4). The results are as follows:
403 %
Tetrabenryl germanium
16.62
FOUND % 16.45
LITERATURE CITED (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Fenimore and Wagner, J . Am. Chem. Soc., 53, 2468 (1931). Graham, J. Assoc. O$ciaZ Agr. Chem., 13, 156 (1930). Jamieaon, J. IND. ENG.CHEX.,11, 296 (1919). Johnson and Dennis, J . A m . Chem. Soc., 47, 790 (1925). Kleeman, 2. angew Chem., 34, 625 (1921). Koch and McMeekin, J . Am. Chem. Soc., 46, 2066 (1924). Myers, J. Lab. Clin. Med., 16, 272 (1931). Newberry, J. Chem. Soc., 127, 1751 (1925). Oakdale and Powers, J . Am. Pharm. Assoc., 20, 881 (1931) Schulek and Villecz, 2. anal. Chem., 76, 81 (1929). Willard and Thompson, J . Am. Chem. Soc., 52, 1893 (1930). Youngberg and Farber, J . Lab. Clin. Med., 17, 363 (1932).
RECEIVEDJune 9, 1932. Presented before the Division of Medicinal Chemistry at the 83rd Meeting of the American Chemical Society, New Orleans, La., March 28 to April 1, 1932.
Microdetermination of Carbon Improvements in Nicloux Method PAULL. KIRK AND PEARL A. WILLIAMS,University of California Medical School, Berkeley, Calif. HE method for microdetermination of carbon developed by Nicloux (3) would appear to have many possibilities, particularly in the field of biochemistry, since it, is a comparatively rapid and moderately accurate method which can be applied both to organic compounds and to mixtures containing carbonaceous compounds either in the solid state or in solution. In its original form, the Nicloux method had many shortcomings, part of which were chemical and part technical. Boivin (1) has so modified the method as to overcome most of the difficulties which arose from faulty combustion and absorption of the resulting carbon dioxide. Various other modifications have been developed, as by Osuka (4), and Schadendorff and Zacherl (6), both of whom have improved the method of handling carbon-containing solutions, such as urine, etc. From the technical standpoint, the method remained clumsy and difficult to carry out. I n this paper, further modifications are reported, aiming a t the reduction of these technical difficulties and the simplification of the procedure. These modifications also make possible an increase in the accuracy and decrease the time necessary to carry out an analysis.
T
APPARATUS UBED The apparatus is shown in Figure 1. It is essentially the apparatus used by Boivin, except that instead of having a bulb blown directly in the upper chamber to hold the absorbing caustic solution, a detachable absorption chamber is introduced. The rubber tube at G in the Nicloux and Boivin apparatus is replaced by a stopcock, as well as the plunger type of stopcock in funnel D. Originally Nicloux used an ordinary stopcock in funnel D and later replaced it with the plunger type to prevent entrance of carbon from the stopcock grease. Such a replacement may be desirable but does not seem to be necessary from the authors' experience, I n addition to the changes in design, it has been found advantageous to use the microfilters to be obtained from the Central Scientific Co., Chicago, Ill., previously described by Kirk and Schmidt (5') for separating the barium
carbonate precipitate, instead of centrifuging as in the original method.
PROCEDURE A sample of dry material containing 2 mg. of carbon is weighed. Unless the percentage content of carbon is very low, this must be carried out on a microbalance, or by dilution of a weighed quantity of the unknown material with a larger weighed quantity of noncarbonaceous solid material and subsequent weighing of a sample of the mixture on an analytical balance. Dry sodium sulfate may serve for such a diluting material if allowance is made for it in adding the reagent later. The dry material is transferred to sample tube A . To this is added 0.6 gram of silver chromate and 0.1 gram of anhydrous sodium sulfate. The sample tube is attached to the upper part of the apparatus and 3 ml. of concentrated carbon-free sulfuric acid placed in funnel D. The sulfuric acid is rendered free from carbon by heating it under a good vacuum on a boiling water bath with a little chromate, This heating should take place behind a protective glass screen because of a certain risk from explosions. In one case out of a considerable number of such heatings, the reagent exploded with considerable violence. No reference to this hazard has been found in the literature. If a stopcock is used, as shown here, it should be greased scantily with vaseline. The sulfuric acid apparently does not dissolve or attack the trace of vaseline with which it comes in contact. Another alternative is to use a phosphoric acid mixture as a lubricant, as described by Stevens (6).
The absorption cup, C, which is made from the bottom of a test tube, is now charged with 0.6 ml. of 2 N sodium hydroxide which must be carbonate-free. This is insured by the previous addition of a little barium hydroxide to the sodium hydroxide solution. The cup, inserted in the stopper containing a hole bored partly through, is inserted into chamber B. The system is now thoroughly evacuated and closed. I n case any leakage develops, a little clear lacquer has been found to be very effective in sealing i t without introduc-
ANALYTICAL EDITION
404
ing difficulties. This, of course, must never come in contact with the oxidizing reagent, but may be safely applied around the stopper holding the absorption cup. Another possible point of leakage is the stopper through which filament F is introduced. De Khotinsky cement will seal this point and is practical for use, since this stopper need never be removed. It is likewise possible, if desired, to seal platinum or tungsten leads directly through the glass, rather than to use a stopper. Such seals are subject to considerable breakage due to heat from the filament. After evacuation, 2 ml. of the carbon-free sulfuric acid are run into A f r o m the f u n n e l . The tube is i m m e r s e d in a b o i l i n g water b a t h f o r a b o u t 20 m i n u t e s . A microburner is then applied directly to the tube until the contents turn definitely green, which marks the end of the react i o n . T h e t u b e i s now cooled in a beaker of water and dilute sulfuric acid run in from t h e f u n n e l until t h e l i q u i d level is about 0.5 cm. f r o m t h e e n d of tube E. A current of short d u r a t i o n is now p a s s e d t h r o u g h f i l a m e n t F by means of a t a p p i n g key. The current should be adj u s t e d so as to h e a t the filament rapidly to a bright red color. A considerable n u m b e r , fifty or sixty, of these intermittent currents are a p p l i e d . The consequent heating and cooling F~~~~~ 1. D~~~~~~OF of t h e filament catalyzes the oxidation of the carbon APPARATUS monoxide present to carbon dioxide, and thoroughly circulates the gases over the caustic in cup C, producing complete absorption of carbon dioxide. The procedure thus far has been that recommended by Boivin with the exception of the use of a different absorption apparatus. From this point on, it is decidedly preferable to abandon his method. The removable cup is simply taken out without disturbing the remainder of the apparatus, and the carbonate is precipitated in the cup with a solution of barium chloride. This is stirred thoroughly and immediately poured into a filter previously prepared as described by Kirk and Schmidt ( 2 ) - The solution is sucked through and the precipitate thoroughly washed with water saturated with barium carbonate. Otherwise] either alkali is retained by the asbestos, or carbonate is dissolved. The precipitate is transferred to a test tube by pushing out the asbestos pad from the bottom of the filter with a small stirring rod. The cup and the sides of the filter are rinsed down by means of 0.05 N hydrochloric acid solution delivered from a microburet or accurate pipet, and followed by a little water. About 10 ml. of the acid are necessary to insure complete solution of the barium carbonate. A drop of methyl red is now added to the solution in the test tube, and the tube is heated by immersion in a boiling water bath until gas evolution ceases. The excess acid is back-titrated while still hot, using 0.05 N sodium hydroxide, and the methyl red as indicator. The acid required to react with the carbonate is
Vol. 4, No. 4
found by subtracting the base added from the total acid added, and the number of milligrams of carbon in the sample is obtained by multiplying the milliliters of 0.05 N acid by 0.3, since one ml. of 0.05 N acid is equivalent to 0.3 mg. of carbon. A blank must be run on the reagents and, if the sulfuric acid is first heated with silver chromate as recommended by Boivin, this blank should not exceed 0.3 ml. of 0.05 N hydrochloric acid. The blank must he subtracted from the amount of acid used in the determination. For analysis of physiological fluids, this method has been recommended by Boivin ( I ) , Osuka (4), and Schadendorff and Zacherl (5). Urine analysis in particular is difficult on account of its content of volatile carbon compounds which are released in the evacuated chamber without undergoing oxidation. The method of Boivin for urine analysis has been tried and found to give unsatisfactory results. The last work mentioned undoubtedly avoids many of the difficulties of the Boivin technic. The modifications proposed here are equally applicable to any of these methods and will result in an improvement in the method. RESULTS
I
i
Since the general method used is the same as that of Boivin it remains only to show that the modified method gives reliable results and is an actual improvement over the older type of technic. In Table I are listed a series of typical analyses of pure compounds. TABLE I. TYPICAL ANALYSES OF PURE COMPOUNDS SUBSTANCE
WT. OF SAMPLE
0.0502 N HC1
Mg.
MZ.
Sucrose
8.45 7.67 5.10
11.798 10.743 7.115
p-Amino.benzoio acid
3.20 3.48
Tyrosine Glutamic acid
THBORETICAL AMT.or C C OBTAINED M Q. Mg . 3.56 3.23 2.14
3,539 3.22 2,134
6.61 7.02
1.98 2.13
1.983 2.106
3.5s 3.36
6.996 6.994
2.135 2.06
2,098 2.097
5.31 5.13
7.264 7.10
2.17 2.09
2.179 2.13
From the results given in Table I it is seen that the method as modified compares very favorably with that of Boivin as regards accuracy when applied to pure compounds. In fact, it is theoretically possible to attain a higher degree of accuracy, since the exposure to the carbon dioxide of the air is reduced, the complete removal of the carbonate from the apparatus is more readily accomplished, and the filtration and washing give more nearly quantitative manipulation of the precipitate than is possible when the centrifuge method is used. I n addition t o the question of accuracy, the saving of time is considerable, which is particularly important in handling the absorption liquid t o keep it free from carbon dioxide of the air. The ease of manipulation is considerably increased, thus making it possible to develop the necessary technic with a minimum of effort. LITERATURE CITED (1) Boivin, A., Bull. soc. chim. b i d , 11, 1269 (1929). (2) Kirk, P. L., and Schmidt, C. L. A., J. Bid. Chem., 83, 311 (1929). (3) Nicloux, M., Compt. rend., 184, 890 (1927): Bull. soc. chim. bid., 9, 639 (1927). (4) Osuka, T., Biochem. Z., 244, 284 (1932). (5) . , Schadendorff. E.. and Zacherl, M. K., Mikrochemie [N. S.1, 99 (1931). (6) Stevens, H. N., J. Am. Chem. Soc., 52, 635 (1930). RECEIYED July 2, 1932.
