V O L U M E 25, N O . 1, J A N U A R Y 1 9 5 3 Table I.
177
L~~
torr.
1 3 ~ 14R 13 16 3-90
Observed Absorbancy, mp 285 275 300 0.337 0.312 0.229 0.295 0.267 0.194 0 334 0.309 0.222 0 337 0.306 0.229 0.261 0.228 0.163
0313 1204 1213
0 481 0,530 0..558
0,506
0.423 0.495
0.301 0.355 0.361
3-Si1ccinoxy-7.12-diketo-10 cholctnic acid
0.074
0.618
0.510
0.458 0.478 0.473 0.424
7-Keto-3.12-dihydroxvcholanic acid (impure)
0 174
0.126 0.329
0.103 0.228
0.169 0.354
Compound Dehydrocholic acid
SO.
Sodiuiiidehydrocholate
3-106 6530
0.371
.4b. 0.250 0.245
0.257 0.265 0.253
O , l o o r
l
275
280
of commercial samples of dehydrocholic acid and concentrated solutions of the sodium Wt. Wt. M x . Found, salt of dehydrocholic acid. Taken. f hIg./hIl. Mg./MI. I n the study of saturated 3.71 0.94 1.00 polyketosteroids the number of 3.77 0.93 1.00 3.77 0.97 1.00 keto groups as determined by 3.77 1.00 1.00 a typical hydrazone assay may 3.77 0.95 1.00 be compared u-ith the quality 3.98 1.82 2.00 3.98 1.91 2.00 and intensity of the spectrum 3.98 1.88 2.00 of the parent compound. Such 7.06 3.00 3.00 a comparison would present 11.34 1.92 2.12 added information as to the 11.34 4.01 5.00 r e l a t i v e p r o x i m i t y of the groups in question. This procedure, of course, should be cairicd out on pure substances tinti should not require baekground absorption correction. Typical data obtained from various commercial preparations of saturated ketocholanic acids appear in Table I. Molecular extinction coefficients used in all calculations are:
Assaj- of Typical Saturated Ketocholanic Acids [Me./nil. = corrected absorbancv X (M.W./c)]
l
,
l
285
,
290
l
,
295
l
Saturated monoketocholanic acid = 35.6 Saturated diketocholanic acid = i1.2 Saturated triketocholanic acid = 106.8
,
300
ACKNOW L E D G V E? T
W A V E LENGTH,m@
Typical Irrelevant ihsorption from Cholic Acid Oxidation Jlixture
Figure 4.
X = 300 mp. (Use of a Beckman 1Iodel D C spectrophotometri equipped u-ith a n ultraviolet accessory unit or an equivalent spectrophotometer is recommended.) Calculate the concentration in milligrams pc’r milliliters of the i-keto acid intermrdiate in the original reaction mixture as follom: Corrected absorbancy = 4.29 X (observed absorbancy 285 mp) - 2.574 X (observed a h s o r h n c y 275 mp) - 1.716 X (observed absorbancy 300 niH)
hIilligrams of keto acid milliliter
Thanks are extended to Hugh B. Corhitt for his hrlpful criticism and suggestions in the preparation of this paper, and to George Jacobs for his assistance in the preptrdtion of the graphs. Samples of 3,6-diketorholanic acid and 3-succinoxy-7,12diketocholanic acid were generouslv furnished by J. S. Buck and C. 11. Suter of the Sterling-Jl‘iiithrop Iteqearch Institute. LITERATURE C I T E D
(1) Henze and Soeer. J . A m . Chem. ,Sot. 64.522 (1942). ( 2 ) Klein, Keiner. and Gordon = ~ S I L CHEV., 20, 174
absorbanm ) (101.5) (25) - (:is 0(corrected ) (volumc of sample talipn 111 milliliters)
01
-
1Iillierams of keto acid - (corrected absorbancyi 1284.1) volume of sample taken in milliliters milliliter Further application of this procedure is found in the evaluation
(1948). (3) &ladigan, Zeniio, and Pheasant, Ibzd., 23, 1691 (1951). (4) Morton and Stubbs, A n a l y s t . 71, 348 (1946). (5) Shi.iiier and Fuson, “Identification of Organic Compounds,” 3rd ed., p. 202, New York, John JViley 6E Sons, 1948. (6) Troaaolo and Lieber. .-~N.LL. CHEM..22. 764 (1950). (7) Werthessen and Baker, Endocrinoloyi, 36, 351 (1945). (8) Wolfe, Hershberg, and Fieser, J Btol Chem., 136, 653 (1940). RECEIVED f o r review AIurcIi 12, 1952. .iccepted Beptemher 9 , 1 0 3 ~ .
Microdetermination of Carboxy Groups In .4romatic and Aromatic-Type Heterocyclic Acids AIORTON BER0Z.i Bureau of Entomology and Plun t Quarantine, U . S. D e p a r t m e n t of Agriculture, Beltsrille, M d . S STUDIES on the struetuie of four insecticidal alkaloids
I isolated from Trtpterygzrcrri todjordzz ( I ) , a small quantityOnof
Cl1H,aOJ and Cl,H,,O,S, was isolated. the basis of experimental data. the d i b a w acids nere believed to have the following structuial formulas
h o dibasic acids,
fiCOO€€
,,-,CH,CHCH~COOH I kH3
I
1 1
!
!
I
y
~
CH3 I1
Each of these compounds hai one carlios)- group attached to the ring, and it should be possible t o t1ecarho~:~~l:ite them and
obtain 1 mole of carbon dioxide from a c h mole of dibasic acid. The hydroxy-dibasic acid (11) should yield an additional mole of carbon dioside offing to the a-hydroxy carboxylic acid structure of the side chain. Hubacher (2’) has described both a gravimetric and a volumetric method suitable for the quantitative estimation of carboxy groups attached to an aromatic ring, but these methods require a large amount, of sample, generally 1 gram or more. T h e ~ ~ l ~ ~ ~ ~ ~ following micromethod, which requires only 5 t.0 10 mg. of sample, is a modification of Hubacher’s gravimetric method. Analyses on seven known acids were in good agreement with those obtained by Hubacher on a macro scale. The method has also been found useful for decarboxylation of aromatic-type heterocyclic acids, such as pyridine, pyrazine, and furan acids. With the present
~
~
178
ANALYTICAL CHEMISTRY
method the pyridine dicarboxylic acid I gave approximately 1 mole of carbon dioxide and acid I1 gave 2 moles of carbon dioxide per mole of acid, as anticipated. REAGEKTS
Quinoline, synthetic, redistilled before use. Coal tar quinoline (Iteilly) also gives satisfactory results if purified as follo\vs: 2.5 grams of basic cupric carbonate is added to 260 grams of dry quinoline and the mixture is refluxed for 24 hours. The quinoline is then distilled directly from the catalyst. The fraction distilling a t 23G0 to 237" C. and at,mospheric pressure is collected. Basic cupric carbonate, Ascarite (a sodium hydrate asbestos adsorbent), Anhydrone (magnesium perchlorate), concentrated sulfuric acid, and diethvlene glycol. .411 chemicals were c.P., and the acids analyzed were of high purity as determined by melting point. Glass wool and Kronigs glass cement. APPAR.ATUS
The apparatus is shoir n in Figure 1. A small reaction flask, A , is attached with springs to a section of glass tubing, B , fitted with ground joints. B connects t o a bubbler, C, containing about 2 ml. of concentrated sulfuric acid, which in turn is sealed to tube D containing Anhydrone. During the analysis a weighing tube, F , such as is used in carbon-hydrogen microanalyses, is attached to tube D by means of a paraffin-impregnated rubber tube, E, that is lubricated ivith a minimum of glycerol. The first three fourths of tube F is filled with Ascarite and the last one fourth with Anhydrone, the two layers being separated with glass wool. Joints c, d, and e are sealed with Kronigs glass cement. The reaction flask, A , is constructed v-ith a heating jacket,, G, so that by boiling a liquid in a 100-ml. flask, H , a constant temperature, corresponding to the boiling point of the liquid, is maintained around the reaction flask. A voltage regulator (Variac, not shovm) controls the heat input to t,he hrating mantle, I , so that theliquidwill reflux halfm-ayup tube J . The heat'ing jacket, G, and the tube connecting it to flask H are covered with glass wool (not shown) to prevent heal loss, and to maintain a uniform temperature in the jacket. This heating arrangement requires no attention after the Variac is set.
through the 10/16 ground joiiit by means of a medicine dropper pipet. The orifice is then closed with stopper a, and screw clamp K is cautiously opened until bubbler C shows a nitrogen flow rate of about 3 bubbles per second. The heat is turned on, and the Variac is adjusted so that the liquid refluxes half way up tube J . I n all the experiments described in this paper diethylene glycol (boiling point 245" C.) was used to maintain the temperature of the heating jacket. However, any other liquid that boils a t the same point may be used in its place. or to maintain a lover temperature, a loivcr-boiling liquid may he used. After about 30 minutes all the carbon dioxide in the quinoline and in the system is flushed out and tube F is attached. RIeanwhile 5 to 10 mg. of the acid to be analyzed (enough to give 1.0 to 3.0 mg. of carbon dioxide) and about 1.0 to 1.5 mg. of catalyst are weighed into the platinum boat, xhich is then placed in the cup of the spoon. The spoon is held aside and tube F is removed with the flannels for weighing. The hole is wiped out, and tube F is wiped with the flannels and chamois in the usual manner (5),placed directly on the balance, and weighed after 10 minutes. This tube is again connected to rubber tube E and the flow of nitrogen is speeded up t o about 4 bubbles per second. Stopper a is removed and the spoon containing the sample and catalyst is introduced in its place. The flow of nitrogen is then cut d o m to about 2 bubbles per second and the spoon is inverted to drop the boat. Hubacher gave the time necessary for the reaction t o go to completion with the seven known acids, -4bout 30 t o 45 minutes bevond these times is allon-ed before tube F is removed for n-eighing t o be sure the carbon dioxide that developed is flushed into it. To remove tube F the following procedure is followed. Tube F is pulled almost out, and the flow of nitrogen through the system is stopped, It is then removed and weighed as above. Xeanwhile, the open end of rubber tube E is stoppered R-ith a glass rod, Tube F is then reconnected, and nitrogen is passed through the system a t a rate of 2 bubbles per second. If another weighing, made after an hour, checks the previous one, the analysis is complete. The amount of carbon dioxide evolved is equal t o the difference between the initial and final weighings. With unknown samples a weighing is made 2 hours after addition of the sample and again everv hour or two until a constant xveight is obtained. K i t h the pyridine dicarboxylic acids ( I and 11) the reaction is complete after 2 hours. I t does no harm t o run the analysis for long periods of time. Blank analyses, in which only the catalyst was added, gave after 6 hours, only the amount of carbon dioxide due t o the catalyst within the error of n-eighing. When the determination is terminated, the products of the reaction may be removed n i t h a dropping pipet that will reach the bottom of the reaction flask. The flask is then rinsed with a suitable solvent. Some catalyst adheres to the sides of the flask, but most of it may be readily removed with fine steel wool held on a wooden stick a i t h fine wire. The loosened particles are rinsed out with acetone and the flask is dried. Before the next analysis is started, the vertical portion of glass tubing B is wiped out with acetone on a cotton swab, and then with a dry cotton swab t o remove any quinoline. The platinum boat is cleaned in hot nitric acid, rinsed with water, acetone, and dried. The percentage of aromatic -COOH in the sample is calculated according t o the formula
Per cent -COOH
K
= 102.3
w1
- w2
~
nihere
H
4
0 5 IO Centimeters
Figure 1. Apparatus
Nitrogen is conducted through pressure rubber tubing fitted with a screw clamp, K , t o a U-tube, L, the first half of which is filled with Ascarite and the second half with Snhydrone. After passing the U-tube the nitrogen is dry and free of carbon dioxide. It is led by means of a rubber tube t o the inlet tube, M , of the reaction flask, A . When the stopper, a, is inserted, the flow of nitrogen through the system may be observed by watching the bubbler, C. A small platinum boat which fits into the cup of spoon B is used to introduce the sample and the catalyst. il semimicrobalance sensitive to 0.02 mg., flannels and chamois ( S ) , and a medicine dropper pipet are also needed. PROCEDURE
With the apparatus assembled but tube F not attached, about 0.8 t o 1.0 ml. of quinoline is introduced into the reaction flask
w1 = weight of carbon dioxide developed in the analysis wz = weight of carbon dioxide due to catalyst
s = weight of sample all in milligrams. I n all the analyses basic cupric carbonate was the catalyst used. Hubacher found that it was the best allaround catalyst and that it developed the theoretical amount of carbon dioxide (19.901, of its &-eight). DISCUSSION
Analyses on the seven known acids in Table I show that the results Ivith the present method are in good agreement with those reported by Hubacher. Hubacher obtained better results with his gravimetric method than with his volumetric one. The gravimetric procedure was used on a micro scale, since it was believed that the increased vapor pressure of the decarboxylated acid might seriously interfere with a volumetric procedure. Furthermore, it is not difficult t o weigh with reasonable accuracy the 1.0 to 3.0 mg. of carbon dioxide obtained with the gravimetric method, whereas the measurement of its correqponding volume of 0.5 to 1.5 ml. with
V O L U M E 2 5 , NO. 1, J A N U A R Y 1 9 5 3 Table I.
179
Determination of Aromatic and Aromatic-Type Carboxyl Acid
Theory 7
Found
-
Literature (3)
%
36.9 3 5 . 1 , 37.0 32.6 31.9,32.4 31.9, 3 2 . 0 32.6 32.8 33.2,32.9 32.8 32.1,32.4 21.2 21.3,21.6 19 9 2 0 . 4 , 2 0 .2 20.2 20.2, 19.7 Acid I 36.3,36.3 37.7 Acid I1 36.6 37.5,36.3 Nicotinic acid 36.6 37.8,37.8 Isonicotinic acid 53.6 52.7,53.2 Pyraaine 2,3-dicarboxylic acid 53.9 52.6,52 2 Quinolinic acid 40.2 42.7,40 9 2-Furoic acidQ a Solution of 1 plus 1 sulfuric acid used in bubbler.
34 8 29 8 31 8 31 9 31 0 21.3 19 6
Benzoic Saiicviic
..
..
_
_
the same accuracy would require temperature control and therefore cumbersome apparatus. I t was found that the water-cooled condenser used by Hubacher could be eliminated, since only a small quantity of quinoliiie got beyond the glass tube, B, and this amount was retained by the sulfuric acid in the bubbler. Tube D was included to eliminate any possibility of sulfuric acid being mechanically carried over. It also ensured that the nitrogen would be dry upon entering tube F . Anhydrone was placed in the last quarter of tube F to prevent moisture loss by the Sscarite. In the early work on this method the catalyst and acid were weighed out directly on the spoon without the use of the platinum boat. However, the acid did not drop completely when the 5poon was inverted. The quinoline vapors were supposed to wash down into the reaction mixture any material adhering to the spoon. In two determinations, the time of decarbosylation was much beyond the time expected and tube F did not reach
constant weight by the end of the day. These determinations which had to be discarded, indicated that the sample did not wash down completely or that the catalyst was deactivated (see next paragraph) by the time the sample finally reached the reaction mixture. The use of the platinum boat not only eliminated this difficulty, but in experiments with benzoic acid, gave yield8 of carbon dioxide that were closer to the theoretical value. Deactivation of the catalyst apparently takes place during a run. Anthranilic acid was decarboxylated in the usual manner. iinother sample of acid was added to the reaction mixture after the determination was complete, but practically no carbon dioxide was evolved. X h e n the evperiment was repeated-except that the acid sample plus catalyst v-as added to the reaction mixture-the proper amount of carbon dioxide was evolved. No~alteration in procedure was necessary to determine carboxy groups attached to pyridine and pyrazine rings. However, to deteimine carbovy groups attached to furan acids, a modification in the procedure was necessary since the results were too high. An aqueous solution of 1 1 sulfuric acid (by volume) was used instead of the concentrated acid in the bubbler. With thi.; change, the results were close to theory even though the acid solution turned brown slowly and the Anhydrone in tube D became wet, so that the acid and Anhydrone had to be replaced after a few runs. For further information on aromatic decarboxylations the reader is referred to Hubacher’s paper ( 2 ) .
+
LITERATURE CITED
(1) Beroza, M,, J . Am. Chem. Soc., in press. ( 2 ) Hubacher, & H., I.ANAL.CHEM.,21, 948 (1949).
(3) Pregl, F., “Quantitative Organic Microanalysis,” 3rd English ed., p. 43, Philadelphia, Blakiston Co., 1937. RECEIVED for review July 26, 1952. Accepted September 1 2 , 1962.
Titration of Phenols in Nonaqueous Solvents JAMES S. FRITZ
AND ROBERT T. KEEN’ Iowa S t a t e College, A m e s , Iowa, a n d K e y s t o n e Oil Refining Go., Detroit, Mich.
ROUABLY the most satisfactorv and widel? used methods ‘for determining phenols involve bromination (6) or acetylation ( 7 ) . The fact that aromatic amines interfere with bromination methods (6) and that both amines and alcohols interfere with acetylation methods ( 7 ) makes other general procedures for phenol determination desirable. Coupling methods (8) are quite general for phenols but are lacking in convenience and accuracy. Titration of phenols as acids has long been inviting but difficult owing to the weakly acidic nature of phenols. I n 1948, however, Moss, Elliott, and Hall ( 5 ) showed t h a t phenols behave as inodrrately strong acids in anhydrous ethylenediamine. They xere unable to find a satisfactory visual indicator but were able to carry out potentiometric titrations successfully using sodium aminoethoxide in ethlenediamine-ethanolamine as the titrant. 1iat.z and Glenn (4)used essentially the same procedure but added a recording device to improve the accuracy ~ i t hTvhich the end point could be detected. Fritz and Lisicki (3) !yere able to titrate phenols potentiometrically in butylamine with sodium methoxide but found no successful visual indicator. Azo violet has been shon n to be a satisfactory indicator for titration of salts ( 1 ) and sulfonamides (2) in ethylenediamine, dimethylformamide, and butylamine, but it gives a gradual, premature end point in the titration of phenol. This paper shows that most phenols Kith negative substiturnts are sufficiently acidic to permit titration in dimethrlformxmide 1
Present address, 4tomic Energ3 Cornmisalon, Oah Ridge Tenn
using azo violet indicator. Phenol and alkyl-substituted phenols are some\\ hat weaker acids but can be titrated in ethylenediamine using o-nitroaniline as the indicator. These titrations give reasonably good accuracy and are convenient to carry out since only the simplest equipment is needed, and commercially available chemicals which may be used without further purification ale employed. REAGENTS AKD SOLUTlONS
Acetone, reagent grade. Acetonitrile, Eastman white label. Benzene, ACS grade from unopened bottle. Benzoic acid, primary standard grade. Dimethylformamide, technical grade (Du Pont j. Ethylenediamine, 95 to 100% (Eastman or Hach Chemical CO.). Methanol, ACS grade. Phenol samples, commercial samples (98 to 1007, purity) analyzed as received. -420 violet, a saturated solution of azo violet (p-nitrobenzeneazoresorcinol) in benzene. p-Hydroxyazobenzene, 0.2 gram dissolved in 100 ml. of benzene. o-Nitroaniline, 0.15 gram dissolved in 100 ml. of benzene. Potassium methoxide, 0.1 ?;. -4dd about 4 grams of freshly cut potassium metal to a mixture of 20 ml. of methanol and 50 ml. of benzene in a loosely covered flask or bottle. When the reaction is complete, add methanol u i t h stirring until the solution becomes homogeneous. Add benzene until the solution remains cloud^ n i t h stirring, then add more methanol until the solution clears. Repeat this dilution procedure until 1 liter
