of the jar serves as separator, if two slides are placed in the chamber at the same time. Alternately, a test tube of about 30 mm. i.d. and 90 to 100 mm. in length can be used. For a larger number of slides, rectangular glass jars, with guide rails such as are used for the preparation of slides in botany, may be employed. The drying and developing of the spots follow conventional procedures. Several experiments demonstrated the practicability of the method. All were executed on silica-gel G (Merck) layers of 2Wmicron thickness. Essential oils of peppermint and sassafras diluted with alcohol to a concentration of 10% were run with benzene. The front was reached in about 5 min. The spots were developed with a saturated solution of antimony trichloride in chloroform, with subsequent heating to about 120' C. Morphine and rutine dissolved in alcohol in a concentration of 1% were run with butanol containing 10% of acetic acid by volume and saturated
I
A
2
4
B Figure 2. Schematic side ( A ) and bottom ( E ) view of applicator 1. Cylinder. 2. Recess In base. 3. Guide rail. 4. Curved extremity of guide rail to even o d small irregularities in slide width
with water. The front was reached in about 30 min. The spots were developed with Wasicky's reagent (p-dimethyl-
aminobenealdehyde in sulfuric acid) with subsequent heating to about 70" C. Vsnic acid dissolved in chloroform in a concentration of 1% waa run with benzene. The spot was examined by ultraviolet light. Aminopyrine dissolved in alcohol in a concentration of 1% was run with methanol. The front was reached in about 5 minutes. The spot was deseloped with an aqueous solution of ferric chloride a t about 1%. In all cases the spots were well defined. The essential oils separated readily into their components, making possible an easy verification of adulterants. The R , values of the principal substances tested, as described, a t 22' C., were: menthol, 0.43; safrol, 0.94; morphine, 0.26; rutin, 0.6; usnic acid, 0.43; aminopyrine, 0.63.
ROBERTOWASICKY Departmenta de Farmacognoaia da Faculdade de Farmlcia e Odontologis da Univeraidade de SBo Paulo Rua Trea Rim, 363 SHo Paulo, Brad
Microdetermination of Fluoride Using Null-Point Potentiometry SIR:In the null-point poteiitiotuetric determination of fluoride previously described [O'Donnell, T. A.. Stewart, D. F., ANAL. CHEM.33, 337 (196111, fluoride solution of unknown concentration was added to one of two identical half-cells containing a cerium (1V)-cerium(II1) solution. Fluoride complexed the cerium(1V) strongly and the redox potential of that half-cell was decreased. Standard fluoride solution was then titrated into the other half-cell until there was no potential difference hetween the two half-cells. The lower limits of applicability of this method were 5 X 10-3M fluoride with an accuracy of 0.5% or 10F3M fluoride with an accuracy of 1%. In the present work the basic method has been developed as a micromethod, an accuracy of 0.5% being obtained in the determination of 5 X 10-4M (10 p.p.m.) fluoride. In addition, the experimental procedure previously reported has been simplified significantly. The improvement in accuracy a t low concentrations has been obtained by several modifications of the original method. First, improved electrical connection between the two half-cells has resulted from the replacement of the agar-KC1 salt bridge by an asbestos fiber. One of the difficulties in the original method arose from the necessity for adding water from a second buret during titration with sodium fluoride solution in order to keep the COII-
centrations of cerium(1V)-ceriuni(Il1) solutions comparable in each half-cell. This has been eliminated by adding concentrated fluoride solution from a microburet, so that the increase in volume during a titration is negligible. A third modification is one of general application for this type of potentiometric titration. By using a volume
of cerium(IV)-cerium(II1) solution in the titrant half-cell which is lsrger by a certain factor, say ten, than that in the half-cell containing the fluoride for analysis, it is possible to titrate a small quantity of unknown fluoride with ten times the amount of standard fluoride, with a consequent increase in accuracy. EXPERIMENTAL
A Figure 1. A. 6. C. D.
c
Apparatus used
Asbestos fiber connection sealed in glass Platinum foil electrodes Stirrer Microburet
Apparatus. The apparatus used is shown in Figure 1. The null-point detection was by means of a Pye portable potentiometer (Catalog No. 7569 P) with a n external Pye Scalamp galvanometer (Catalog No. 7902/S). Platinum foil electrodes were used. I n a typical analysis the internal half-cell consisted of a glass tube (6 x 6 / * inch) with a n asbestos fiber sealed into the bottom. Several different tubes employing asbestos fiber, cracked glass, and sintered glass electrical connections were tested before selecting the asbestos fiber connection as being the most satisfactory-Le., having a low resistance and negligible transfer of liquid through the connection. The external half-cell was a 250ml. beaker. The microburet used had a delivery volume of 0.2 ml. and could be read to an accuracy of =tO.OOOl ml. All reagents used were of analytical grade. Procedure. For analysis of a solution 5 x lO-4M in fluoride, dissolve 0.64 gram of ceric ammonium sulfate and 0.28 gram of cerous sulfate in 200 ml. of water and 14 ml. of 18M sulfuric acid, and dilute to 1 liter. VOL 34, NO. 10, SEPTEMBER 1962
0
1347
Table 1. Results of Determinations of Fluoride Added as Sodium Fluoride Series No.
I
I1
Added 0.950
0.891 0.885 0.467 0.612 0.730 0.940
Fluoride, Mg. Found 0.951, 0.946, 0.959. 0.946,0.940,0.961. 0.940,0.947,0.956, 0.955, 0.942, 0.961 0.895 0.886 0.464 0.609 0.727 0.949
Pipet 5 ml. of this solution into the internal half-cell and 50 ml. into the external half-cell, and check that the e.m.f. of the cell is zero. Pipet a suitable aliquot-e.g., 10 ml. of unknown fluoride-into the internal half-cell and add ten times this volume of distilled water to the external half-cell. With continuous stirring, add standard 0.551
sodium fluoride from the micrcburet until the null-point is obtained. In this example, the quantity of fluoride in the titrant will be approximately ten times the quantity of fluoride in the unknown fluoride sample, the exact value of this factor depending on precise pipet delivery volumes. The factor is usually best determined experimentally by titration against standard 5 x 10-4.11 fluoride. The volumes quoted above are for a typical analysis. Obviously, the size and concentration of the sample for analysis xi11 govern the choice of dimensions of the internal and external half-cells and of the volumes of unknoivn fluoride solution, water, and cerium (1V)-cerium (111) solutions that will be used. The size of the internal half-cell and of the unknown solution used could be reduced considerably. RESULTS
The results of eighteen determinations by this method are shown in Table I. In Series I, the volume of unknown fluoride taken for analysis was 10.00 ml. but in Series I and 11, the volume of
Determination of Carbon by W e t Combustion. Explosives and Initiators SIR: Recently we described a method for determining carbon in alkyldecaboranes based on oxidation in an evacuated apparatus by combustion fluid containing chromic, iodic, phosphoric, and sulfuric acids [ANAL. CHEM.33, 1598 (196l)l. Decomposition was
Table I.
smooth and complete and a n analysis took 1 hour. Dry combustion of these compounds often resulted in violent decomposition unless samples were heated slowly and carefully in oxygen diluted with nitrogen. We have used this met combustion
Determination of Carbon in Explosives and Initiators
Behavior Compound Pentaerythritol tetranitrate Glycerol trinitrate Nitroguanidine 1,ZDinitraminoethane
1,3,5Trinitro-1,3,5-triazacyclohexanc 1,3,5,7-Tetranitro-l,3,5,7-tetraazac~clooctane
Calcd. 19.00 15.87 11.53 16 01 16.22 16.22
Carbon, 7' Found 19.06 15.94 11.48 16.00 16.27
on (3xida-
Diff.
+O.OG $0.07 -0.05 -0.01 +0.05
tion h a h 1
1)
h
2,4,6-Trinitrotoluene 2,4,6-Trinitro-m-xylene 2,4,6-TrinitrophenoI 2,4-Dinitroreaorcinol 4,6-Dinitroresorcinol Lead salt of 2,4-dinitroresorcinol Basic lead salt of 4,6-dinitroresorcinol
37.02 39.84 31.46 36 01 36 01 li.73
16.37 17.68 20.82 36.90 39.83 31.42 36 02 36.20 17.81
[C6H*(NOn)*(OnPb)!alph(OH),l~
12.73
12.74
$0.01
a
18.12
18.06
-0.06
b
1,3-Dinitro-5-nitroso-1,3,5-triazacyclohcxnnr 17.48 1,3,5Trinitroso-l,3,5triaz~cyclohexnnc 20.70
Banum salt of 2,4,6trinitroresorcinol monohydrate
$0.15 +0.20 +o. 12 -0.12 -0.01 -0.04 +0.01
+o. 19 $0.03
b a a a
tl
5 C C
unknown fluoride was varied. For Series 1 and 11, the mean percentage deviations ure 0.7 and 0.45%, respectively. The results quoted here are for solutions 5 X 10-4M in fluoride but results of similar accuracy were obtained for solutions 5 X 10-a~M and 5 x 10-*M in fluoride using these techniques with suitable adjustments of reagent concentrations. Interferences and optimum experimental conditions are the same as for the original method of O'Donnell and Stewart, except that to obtain stable e.ni.f.'S it is better not to reduce the c~eriuni(III)-cerium( llr) concentration below 10-3Alf and, hence, the concentrations quoted for this reagent in this paper are ten times greater than those one would expect by a direct scaling down of the original conditions. T. A. O'DONNELL D. F. STEWART Department of Chemistry University of Melbourne Parkville, N. 2 Victoria, Australia
Application to
method to analyze typical explosive and initiatory substances including nitric esters, nitramines, nitrosamines, nitroaromatic compounds, and lead and barium salts of nitroresorcinols. Even the heavy metal salts of nitroaromatic compounds, for example, barium styphnste, were oxidized safely and rapidly. Of the compounds examined, only tetrazene reacted too violently with combustion fluid to be analyzed by this method. EXPERIMENTAL
The experimental procedure was identical with that described previously apart from the use of combustion fluid containing 5y0 (w./w.) of water for thc analysis of nitramines and nitrosamines. Sample weights of about 0.1 gram were used; only in the analysis of glycerol trinitrate was sand added to the weighed sample. Preliminary experiments were performed to estimate the hazard involved in treating initiators with combustion fluid. Twenty milliliters of the fluid were added to 0.1 gram of the initiator mixed with 1 gram of potassium iodate in a stainless steel beaker placed in a safety screen. RESULTS AND DISCUSSION
a Reacts readily a t room temperature. b Oxidizes completely below 100' C. c Requires heating to 100" to 200' C. for complete oxidation.
1348
ANALYTICAL CHEMISTRY
The results presented in Table I show that the analyses are usually accurate to better than 0.2% absolute.