A N A L Y T I C A L CHEMISTRY
130 Table I. Analysis of Refined Selenium Powder Se,
%
Te,
Sios,
Cu,
Pb,
Fe,
Hg,
Sb.
Sample
%
%
1
99.65 99.64 99.65 99.49
0.30
0.077 0.029 0.014 0.024
0.0014 0.0010 0.0017 0.0039
0.012
0.021 0.017
0.010 0,0009
0.0046 0.0033 0,0033 0,0039
Xi1 Nil h-il Nil
2
3 4
%
0.26
0.23 0.23
70
%
%
0.017 0.013 0.017
%
0.024
0.018
%
0.007
0.0005
and 4 gave the values 98.73, 99.15, 98.81, and 98.73% selenium, respectively. Sodium Selenite. The permanganate procedure applied to sodium selenite was found to give selenium values in good agreement with the theoretical value, 45.65y0. Typical samples analyzed 45.45 and 45.67y0 selenium. Sodium Selenate. As no convenient means is available for reducing sexivalent selenium to the quadrivalent state, the most satisfactory procedure of analyzing sodium selenate by the permanganate method was found to be reduction of the selenate to elemental selenium by hydroxylamine hydrochloride, followed by oxidation to the quadrivalent state by nitric acid. Typical analyses gave 41.84, 41.77, and 41.84%,,values agreeing well with the theoretical value, 41.79%. Selenide, The selenium in iron may be determined readily by the permanganate met,hod following dissolution
xron
S.
of the sample in hot mixed nitric-sulfuric acid. In order to check the selenium value indirectly, the iron content of a number of samples was det'ermined by hydrolytic precipitation from a separate aliquot of the sample solution following the removal of selenium and tellurium. A typical analysis gave selenium 58.49%, iron M.49%, and tellurium 0.47%. The theoretical selenium figure is 58.57%. ACKNOWLEDGMENT
The authors are indebted to Herbert >farshall for the determination of the impurities in the refined selenium samples. LITERATURE CITED
Deshmukh, G. s., and Sant, B. R., Analust 77, 272 (1952). (2) E ~B. s,, ~Ibid.,~67, S46 ~ (1g42). , (3) Gooch, F. A . , and Clemons, C. F., Am. J . Sei. 50, 5 1 (1895). (4) Marshall, H., Canadian Copper Refiners, Ltd., unpublished work. ( 5 ) Schrenk, W. T., and Browning, B. L., J . Am. Chem. SOC. 48, (1)
139 (1926).
(6) Ibid.,p. 2550. (7) Schulek, E., and Koros, E., 2. anal. Chem. 139, 20 (1953). (8) Stammy H.3 and Goehring, LI.,I b d . , 120, 230 (1940).
RECEIVED for
review Aplil 4, 1
9
Accepted ~ ~ June 28, 1955.
Coulometric Determination of Organic Bases in Acetonitrile CARL A. STREULI Research Division, American Cyanamid Co., Stamford, Conn.
The coulometric method of analysis is extended to the determination of amines in essentially nonaqueous solution. Nonaqueous solvents have been shown to be most useful in the determination of these weaker bases. The solvent employed is acetonitrile containing 0.O.W lithium perchlorate trihydrate. The water content of this solution is approximately 0.3%. Hydrogen ion is produced by anodic oxidation of the water and the ion detected by the conventional glass-calomel electrode combination. Pyridine, diphenylguanidine, triethylamine, and benzylamine have been titrated in milligram quantities; the usual error is less than 2%. Aromatic amines tested by this method cannot be titrated. They apparently undergo partial or complete oxidation by the oxygen in solution which is produced concurrently with hydrogen ion at the anode.
not oxidized to yield hydrogen ion anodically-the reaction water so readily undergoes. The small amount of water required in the solvent as the source of hydrogen ion was introduced in the form of a hydrate of lithium perchlorate. A 0.05,V solution of lithium perchlorate trihydrate in acetonitrile served as the solvent for the determinations. The water content of this solution is 0.3%. A potentiometric titration curve for pyridine dissolved in ace-
THE
value of coulometry as a method for the determination
of small amounts of materials has been amply demonstrated in recent years ( 2 ) . DeFord ( 4 ) has applied this method for the
titration of acids and bases in aqueous systems. Carson and KO ( 1 ) have titrated acids coulometrically in a 70/30 mixture of isopropyl alcohol and water. Extensive work (6, 7 )has shown, however, that weak acids and bases are most easily titrated in nonaqueou3 solutions. An extension of coulometry to essentially nonaqueous systems should then be useful in the determination of small amounts of these compounds. Coulometry has a further advantage in t h a t it is readily adapted t o automatic procedures. The present paper deals with a coulometric method for determining weak bases in acetonitrile. The usual source of hydrogen ion in the aqueous coulometric titrations for base is the solvent itself. Acetonitrile, however, is
ml. 0.IN HCQ in C Y N
Figure 1. Titration of pyridine in acetonitrile solution
V O L U M E 28, NO. 1, J A N U A R Y 1 9 5 6 Table I. Compound Pyridine
1,3-Diphenylguanidine
Triethylamine
Benaylamine
Results on Four Amines
Sample, Current, Mg. Ma. 0.581 2.044 2.041 2.039 2.039 1.984 1.986 1.992 1.998 1.400
1.092
0.638
131
2.018 1.976 1.998 1.994 2.044 2.056 2.064 2.070 1.897 1,894 1.894 1.999 2.005 2.012
Time,
Result, Mg. 346.8 0.581 353.4 0.591 351.0 0.587 349.8 0,585 364.2 0,592 361.8 0.589 350.4 0.572 352.8 0.678 Av. 0 . 5 8 4 1.400 316.8 1.396 324.0 1.401 319.2 1.398 320.4 Av. 1.399 1.097 512.4 1.097 509.4 1.089 503.4 1.087 501.0 Av. 1.092 Sec.
304.2 301.8 289.8 296.4 286.2 283.8 Av.
0.640 0.634 0.609 0.657 0,637 0.634 0,635
Error, Mg.
0.000 0.010 0.006
0.004 0.011 0.008 -0.009 -0,003 10.006 0.000 -0.004 0.001 -0,002 10.002 0.005
0.005 -0.003 -0.005
10.005 0.002
-0,004
-0.029 0.019 -0.001 -0.004 10.010
%
Error 0.0 1.7 1.0 0.7 1.9 1.4
chlorate salt is manufactured by the G. Frederick Smith Chemical Co. The amines were reagent grade chemicals and were tested for purity by conventional nonaqueous titration with perchloric acid. Samples were prepared by dissolving weighed quantities of the amines in acetonitrile and serially diluting to the desired value. The supporting solution was made by dissolving 8 grams of !ithiurn perchlorate trihydrate in 1 liter of acetonitrile and filtering.
-1.5
-1.0 11.0 0.0
-0.3
0.1
-0.1 10.1 0.4 0.4
-0.3 -0.4 10.4 0.3 -0.6 -4.5 3.0 -0.2 -0.0 rl.5
tonitrile containing 2% of water is shown in Figure 1, together with the titration curve for pyridine in the pure solvent. Although the end-point potential is lowered for the solution containing water, a clearly defined titration break is apparent, and titration values are identical. Coulometric titrations in a solvent containing only 0.3% of water should not then offer endpoint problems.
PROCEDURE
The procedure is essentially the same as that described by Cooke, Reilley, and Furman ( 3 ) . One hundred milliliters of 0.05N base solution is placed in the 150-ml. beaker, the solution is stirred, and the recording is begun. Current is allowed to flow until the indicator potential shows a value between 300 and 350 mv. (near the inflection point of the blank curve). Without stopping the generation current a 2.0-nil. sample of amine solution is pipetted into the beaker. The indicator potential falls immediately to a lower value, and then rises again to the end-point potential as the generation proceeds. Generation current is measured during this interval. When the indicator potential has again risen to its previous value. another sample is added to the same solution and the procedure is continued. Five to six samples may be added to the original solution in this manner during one run. The time elapsed to completion of the titration is determined by measuring the distance between points of equal potential on the upward sloping curve of the recording, and multiplying by the appropriate factor. Final results are obtained by multiplying current, time, and the appropriate factor in milligrams per microcoulomb for the amine under consideration. As Cooke, Reilley, and Furman (3) have pointed out, measurements need not be made at the exact end-point value of a titration, as long as they are made a t point of equal indicator potential for both blank and sample, if this value is reasonably near the true end point. Data collected during this investigation showed that time measurements made of the same sample 100 mv. apart were virtually identical.
APPARATUS
RESULTS
The generation apparatus and end-point detection system are of conventional design, similar to that described by Cooke ( 2 ) . Essentially constant current is obtained from two 45-volt dry cells connected in series through a 33,000-ohm resistor. Current is measured by determining the voltage drop through a precision resistor of 24.00 ohms with a standard potentiometer. hleasurement of current during the course of a number of determinations indicated that the current variation was of the order of 3 or 4 parts per thousand. A large platinum gauze electrode immersed in the body of the solution served as the generator anode. The cathode was a platinum wire spiral isolated in a separate compartment containing an aqueous 1% solution of the lithium salt. The two compartments were connected by a salt bridge of 1% lithium perchlorate trihydrate, 5% agar. The resistance of the bridge decreases the current passed through the solution, but this is necessary, as when the cathode was immersed in the acetonitrile solution constant generating currents could not be obtained. These bridges are slowly dehydrated by the acetonitrile solution and must be replaced from time to time. The indicating system consisted of glass and calomel electrode-, a Leeds & Northrup line-operated pH meter, and n Speedomas recorder. The recorder's speed can be varied t o 1 or 2 inches of strip chart er minute and gives a full scale deflection of 350 or 700 mv. T f e recorder was used in place of a clock, since its speed is precise. Time can be measured t o within 0.3 second, which iq 1part per thousand on a 300-second run. Studies were also made using a Beckman Model G pH meter and a time clock reading to 0.1 second. Comparable results were obtained. All values reported, however, were obtained from the recorded data. Titrations were carried out in a 150-ml. tall-form beaker. Solutions were stirred with a magnetic stirrer. All electrodes were contained in a large rubber stopper fitted to the beaker. Recorder, pH meter, stirrer, and coulometer control box were all grounded to eliminate stray voltages, and the potentiometer wa8 shielded with aluminum foil. Generator and potentiometer leads were of shielded wire. Shielding and grounding appear t o be essential for these titrations, as the detection system is extremely sensitive to voltage transients.
A recording of the indicating electrode potential with time, obtained during hydrogen ion generation in the acetonitrile solution, is illustrated in Figure 2. Generation current was 2.073 i 0.005 ma. The inflection point of the curve is between 300 and 350 mv. Results obtained on the four amines determined are listed in Table I together with error in milligrams and per cent erior.
REAGENTS
time
The acetonitrile used in this work is commercial grade obtained from the Carbide and Carbon Chemicals Co. The per-
current 2.073 ma.
420
--!
'5
140
01
Figure 2.
Coulometric generation of hydrogen ion in acetonitrile solution
132
ANALYTICAL CHEMISTRY
The first value in any group of runs is not included, as this result vas invariably lower than any of the following values. Average crrors run less than 2% for any compound and individual errors are also less than t,hisvalue, except for two of the values for benzylamine. I n most cases an error of 1 second in determining the elapsed time of titration from reading the recordings would be equal t o 2 or 3 y of amine if the current were about 2 ma. Attempts were also made to determine several aromatic amines hy the same procedure. Sone of the compounds tested (diphenylamine, p-toluidine and phenylenediamine) gave qnantitative results. When samples of diphenylamine m r e added to the pretitrated qolvent, t,he entire solution turned violet; this color eventually faded to yellox brown. S o break in the indicator current was noted when the sample was added. This behavior is believed to be due to oxidation of the amine by oxygen generated in solution at, the anode concurrently with hydrogen ion. Fieser and Fieser (6) state that diphenylamine is readily oxidized by chemical means to tetraphenylhydrazine. This may be the reaction undergone in this solution. If the amine and perchloric acid are mixed in acetonitrile, no rolor is exhibited. The color reaction occurs in pretitrated solution, however, even after the generating current has been turned o f f . The reaction cannot then be direct oxidation of the amine :it the anode. Phenylenediamine and p-toluidine both give breaks in the indicator current, h u t rewlts are very low, indicating that partial
oxidation has occurred. The phenylenediamine solution turns bright yellow during the course of the reaction. Caution must be exerted in making sure that chlorides are not present in solution when hydrogen ion i- generated. Chloride is evidently oxidized to chlorine in acetonitrile and high results will be obtained. The solution also turns yellow with prolonged generation if chlorides are present. ACKNOW LEDG3lENT
The author gratefully acknon-ledges the assistance of Everett
II-.Hobart in constructing the generation system. LITERATURE CITED
(1)
Carson, W.S . , and KO, R.,
~ ~ N AC L H. E U .
23, 1019 (1951).
( 2 ) Cooke, W. D., “Coulometric llethods, Organic Analysis,”
1.01.
11. pp. 169-93, New York, Interscience, 1954. (3) Cooke, W.D., Reilley, C. S . .and Furman. S . H., ANAL.CHEM. 23, 1662 (1952). (4) DeFord. D. D., Pitts. J. S . , and Johns. C. J., Ibid., 23, 938
(1951). (5) Fieser, L. F., and Fieser, AI., “Organic Chemistry,” pp. 610-11, Heath, Boston, 1944. (6) Fritz, J. S.,“.kcid-Base Titration in Sonaqueous Media,” G. Frederick Smith Chemical Co.. Columbus, Ohio, 1952. ( i )Pifer, C. W., Wollish, E. G., and Schrnall, 31.. J . Am. Pharm. -4ssoc. 42,509 (1953). R E C E I V Efor D review September 21, 1055. .iccepted October 19, 1955.
Separation of 2,4=Dinitrophenylhydrazones of Aldehydes and Ketones by Paper Chromatography WILLIAM S. LYNN, JR., LOIS A. STEELE, and EZRA STAPLE Department o f Biochemistry, School o f Medicine, University
4 convenient and highly sensitive paper chronlatographic method for separating homologs of low molecnlar weight 2,4-dinitrophenylhydrazones is described. Ksing phenoxyethanol-impregnated paper as the stationary phase and heptane as the mobile phase it has been found possible to obtain good separation of as much as 230 -/ of 2,4-dinitrophenylhydrazones in a single spot.
.1RIOUS methods have been described for the separation of 2,4-dinitrophenylhydrazones of aldehydes and ketones by paper chromatogiaphy ( 1 , 2, 4 ) . However, many of these systems powess serious limitations in ability t o separate homologs and have other inconvenient features. Some require eqpecially impregnated paper which has t o be prepared by difficultly repi oclucible technique. I n others, the quantity of 2,4-dinitrophenylh j drazone is limited to such small amounts that detection of the spots becomes uncertain. Attempts t o increase the quantity of 2,4-dinitrophenylhydrazonein such systems leads t o strenlmg and poor separation. Neher and Wettstein ( 3 ) have described a two-phase paper chromatographic system for separating steroid compound3 This system, using phenoxyethanol-impregnated paper as the stationary phase and heptane as the mobile phase, has been found t o be well adapted to the separation of homologs of 10x5 molecular weight 2,4dinitrophenylhydrazones. As much as 250 -1 of a 2,4-dinitrophenylhydrazone in B spot 1 cni 111 diameter ran be chromatographed with no appreciable streaking. The spots obtained with such amounts of material are readily visiblp arithout spraying or use of ultraviolet light. .41so, a s all
o f Pennsylvania,
Philadelphia
4, Pa.
solvents used are easily evaporated, the separated 2,4-dinitrophenylhydrazones can be readily eluted and recovered in pure state from the paper. EXPERIM EKT 4 L
Descending development of the papei chromatograms was used. Strips of Whatman S o . 7 filter paper were dipped in a solution of 10% phenoyyethanol in acetone, blotted free of eycess solution, and then dried in air for a fen minutes. The 2,4-dinitrophenylhydrazones, dissolved in methanol, were applied to the paper to give spots 0.5 to 1 em. in diameter. The paper strips
Table I.
Mobilities of 2,4.-Dinitrophenylhydrazones
2,4-Dinitrophenylhydrazone
Formaldehyde Acetaldehyde Propionaldehyde Butyleldehyde Valeraldehyde Heptaldehyde Acetone P-Butanone 2-Pentanone %Pentanone 4-Hydroxy-+methyl2-pentanone Benzaldehyde Anisaldehyde Veratraldehyde Solvents.
hIelting Point. ’ C. 107
169
1,56
122 108 108 122 117 144 1.56
202 239 256
265
Nobility in 20 Hours. Cm. f r o m Origin
Keight,
.5 . 5 9.0 12.0 1d , 3
200
i
20.5 28.5
200 200 200 200 200 200 200 200 200
21.5
200
Streaked Streaked
50 50 50
22.0
29.0 13.5 20.5
2.0
Stationary phase phenoxxethanol. Moving phase, hkptane saturated with phenoxyethanol. Paper Whstman S o . 7 Derelbpment, descending.