trap into cooled, 1 :1 ethanol-hydrochloric solution. Additional 1% potassium permanganate was added to the trap and the trap solution was boiled for 15 minutes. Solid thiourea was dissolved in the resultmt receiving solutions to give a concentration of 4%. The mixture was heated to 85" C., allowed to cool and th: absorbance mas measured spectrophotlsmetrically at 480 mp . The results for salted samples are recorded in Table V . The first four samples in Table V deal with recoveries from samples containing only 80 grams of oxidized concentrate salted with 161 pg. of osmium but without a subsequent hydrogenation. Each of the last 23 samples contain the basic 80 grams of unsalted oxidized concentrate in addition to the amounts of salted conrentrate recorded in Table V. Sources of Osmiiim Loss. Slags were analyzed according t o t h e proce-
dure b y Plummer and Beamish (7) and pot walls according to the procedure by Kavanagh and Beamish ( 4 ) . N o appreciable amounts of osmium were detected. The residue from the button parting was treated to determine its osmium content. This predominantly sulfurous residue was collected and burned in oxygen in a Vycor tube at 800" C. The gas was passed through a cooled mixture of 1 :1 ethanol-HC1 to which thiourea was added and the solution subsequently heated. X o indications of osmium were obtained from any sample. ACKNOWLEDGMENT
The authors express their appreciation to the h'ational Research Council for Studentships given to J. C. Van Loon and to the Falconbridge Kickel Mines, Ltd., for the provision of ores and concentrates.
LITERATURE CITED
( 1 ) .411an, W. J., Reamish. F. E., ANAL. CHEM.24.1569 (1952). ( 2 ) Griffith,' L., Trans. 'Can. Inst. ;Mining M e t . 43, 153 (1940). (3) Hoffman, I., Schweitzer, J. E., ANAL. CHEM.25,1091 (1953). (41 Kavanaueh. J. M.. Beamish. F. E..
' Ibid., 32,45011960). ' ( 5 ) Lovely, W.13. C., Chena. Eng. Mining Rev.33, 199 (1941 ). ( 6 ) O'X,eill, J. J., Gunning, H. C., "Platinum and Allied Metal Deposits of Canada," Canada Department of Mines, Ottawa Economic Geology Series X o . 13, 1934. ( 7 ) Plummer. IT. E. \-.. Beamish. F. E.. ANAL.C H E31, ~ 1141'(3959). ' ( 8 ) Rueden, H., Henderson, J., J . Chem. M e t . Mininq SOC. S. Africa 28, 181 (1928). (9) Westland, A. D., Beamish, F. E., AKAL.CHEM.26,739 (1954). (10) Westland, A. D., Beamish, F. E., Am. Mineralogist 43, 503 (1958). (11) Zachariasen, H., Reamish, F. E., Talantn 4 , 4 4 (1960). RECEIVEDfor review October 21, 1963. Accepted Januarv 7 , 1964. \
,
Identificatiion of the Pyrolyzates of Substituted Barbituric Acids by Gas Chromatography DONALD F. NELSON' and PAUL L. KIRK School of Criminology, University of California, Berkeley, Calif.
b
The nitriles are shown to be important in both abundance and variety among the pyrolysis products of the substituted barbituric acids. One route of degradation, yielding a nitrile which incorporates the C-5 substituents intact, is shown to predominate in the pyrolytic decc'mposition of the sodium salts of six barbituric acids. A seventh barbiturate e.lthibited a different pattern.
D
among the barbiturates of criminalistic interest by means of Flyrolysis and gas chromatography has been described (6). The identification of the pyrolysis products of a selected group of these drugs has been undertaken to assess the validity of this empirical technique. The combination of pyrolysis with gas chromatography, first suggested about a decade ago, hes been applied t o the analysis of many classes of large organic molecules, both natural and synthetic. Its use in the study of drugs and poisons is relatively recent. Most significant was t'ie work of Janak (4, 5 ) mho reported the study of 15 barbiturates, along with a variety of 'other compounds. The findings, under IFFERENTIATION
* Present address, Dominion Laboratory, Auckland, New Zealand.
his conditions, concerned chiefly both alkyl and aryl hydrocarbons as products of pyrolytic decomposition, in contrast to the findings in this study, in which the chief and most significant products were nitriles. EXPERIMENTAL
Apparatus. Two gas chromatographs were employed: a Pye argon instrument with a strontium-90 detector a n d a 10-mv. MinneapolisHoneywell-Brown recorder, a n d an ilerograph A-90 arranged for preparative work. Three 4-foot glass columns of 5-mm. i.d. were used with t h e P y e chromatograph; t h e stationary phases were polypropylene glycerol, 10%; Carbowax 600, 6%; and Carbowax 20N, 6%. I n the two preparative columns polypropylene glycol, 10% and Carbowax 20M, 6% were used; the copper tubes were 5 feel long and l/r-inch 0.d. Firebrick 100- to 120mesh was the supporting medium in all the columns. The column packing was prepared by dissolving the liquid phase in sufficient volatile solvent to form a thick slurry with the supporting medium; t h e slurry then was dried on a steam b a t h under partial vacuum with occasional gentle turning with a spatula. The pyrolysis unit described previously (6) was used. This was a modification of t h a t described b y Strassburger et al. (8). A platinum foil strip bridged the gap between two
platinum contact wires which, for convenient manipulation, had been sealed through a ground-glass stopper of standard taper. It was heated t o a dull red by a current of about 6 volts and 6 amperes in approximately 1 second. This unit was adapted for use with the preparative chromatograph by enlarging its foil receptacle to accommodate samples u p to about 10 mg. Neasurements of absorption in the ultraviolet and in the infrared were made with a Beckman D B recording spectrophotometer and a Perkin-Elmer Infracord, respectively. Melting points were determined with a Kofler hotstage. Reagents. T h e sodium salts of seven substituted barbituric acids studied are as listed under Results I n six of them the ring is unmodified except a t the C-5 atom, one of t h e substituents a t t h a t location being ethyl, and t h e other a straight or branched alkyl chain. T h e seventh derivative has a n additional methyl group on t h e 5-1 atom. T h e acids were obtained free of excipients and were recrystallized fmm aqueous ethanol to constant melting point and were desiccated under vacuum. The salts were prepared by reaction in ethanol solution of equivalent weights of sodium hydroxide and the barbituric acids and were also desiccated. Procedure. A preliminary investigation was made of t h e pyrolysis VOL. 36, NO. 4, APRIL 1964
875
Tablel.
Matching of Retention Times of Reference Standards and Major Pyrolyzates of Barbiturates
Compounds compared Major peak, Na barbital 2-Ethylbutanenitrile Major peak, Na butethal 2-Ethylhexanenitrile A ma:or peak, Na metharbital Methyl isocyanate Formaldehyde
Retention times (minutes) Polypropylene glycol Carbowax 600 column column 40' C. 75" C. 125' C. 40" C. 75'C. 79.0 14.1 2.95 37.0 8.0b 13.8 2.9 37.0 8.2 c 7.85 25.5 c 55.0 7.75 25.5 3.15 3.05
1.4
3.1
1.35
1.3
0.75
2.25 2.25
1.1
Not seen
1.75
0.95
0.8b
c
Carbowax C OM column
125" C. 1.9 1.9 3.7 3.75
0.71 0.8verv asymmetrical Notseen
Flow rate throughout: 50 ml./min. Chromatogram not run long enough for emergence. Peak somewhat smaller than is typical of major products. e Not chromatographed at this temperature. ~
Table II.
Probable Identifications of Pyrolyzates of Barbiturates
Sodium barbiturate acetone (large) ethanenitrile (medium) lJ2-epoxyethane hydrogen cyanide ethanol water Barbituric acid acetone (medium-variable) ethanenitrile (medium-variable) water Sodium barbital 2-ethylbutanenitrile (very large) 1,Qepoxyethane hydrogen cyanide butanenitrile pentanenitrile hexanenitrile propenenitrile 2-methylpropenenitrile acetone benzene dimethyl carbonate is0butyraldehyde water Barbital 2-ethylbutanenitrile butanenitrile propenenitrile 2-methylpropenenitrile 1,2-epoxyethane methyl isocyanate is0butyraldehyde dimethyl carbonate benzene Sodium butethal and butethal 2-ethylhexanenitrile (very large) hexanenitrile butanenitrile water hydrogen cyanide acetone acrolein (The products from butethal and sodium butethal have not yet been chromatographed at 40" C.)
876
ANALYTICAL CHEMISTRY
Sodium hexethal 2-ethyloctanenitrile (very large) octaneni trile butanenitrile propenenitrile hydrogen cyanide acetone acrolein pyrrole (Not chromatographed at 40" C.) Sodium probarbital 3-methyl-2-ethylbutanenitrile (very large) 2-ethylbutanenitrile (medium) butanenitrile propenitrile hydrogen cyanide benzene acetone water acrolein (Not chromatographed a t 40" C.) Sodium amobarbital 5-methyl-2-ethylhexanenitrile (very large) 2-ethylhexanenitrile hexanenitrile 2-ethylbutanenitrile butanenitrile propenenitrile acetone benzene (Not chromatographed at 40" C.) Sodium pentobarbital 3-methyl-2-ethylhexanenitrile (very large) 2-ethylhexanenitrile hexaneni trile butanenitrile 2-ethylbutanenitrile propenenitrile hydrogen cyanide acetone benzene (Not chromatographed at 40' C.) Sodium metharbital The largest peak has not yet been identified methyl isocyanate (very large) 2-ethylbutanenitrile but anenitrile propenenitrile 2-methylpropenenitrile benzene 1,2-epoxyethane
1'5
lb
5
MINUTES
Figure 1 . Pyrolysis products of sodium barbital chromatographed at 75" C. on polypropylene glycol, 10% on firebrick 100/120. The large peak was produced by 2-ethylbutanenitrile
products of barbituric acid and of sodium barbiturate. As in the earlier study ( 6 ) , there was no attempt at quantitation. Sample size was judged by eye on the basis of experience, being a few tenths of a milligram in the regular runs and about 10 mg. in the preparative runs. T o increase the yield of certain desired functions, quantities of barbiturate approximating 3 grams were heated vigorously in small test tubes, and the products were recovered with a watercooled condenser. Such a sample yielded about 1 ml. of crude liquid product, which was injected into the preparative gas chromatograph, the major fraction being collected in a Utube cooled with acetone and dry ice. The s e w n substituted barbituric acids were pyrolyzed both as free acids and as their sodium salts. The retention times of the pyrolysis products were compared with those of standard reference compounds determined under the same gas chromatographic conditions. Table I illustrates the matching of retention times applied to the major pyrolyzates of sodium barbital and sodium butethal, and to one of the two major products of sodium metharbital. The retention times of formaldehyde have been included t o illustrate the necessity of employing more than one liquid phase and of operating a t several temperatures. Certain unobtainable standards were synthesized: nitriles by the method of Friedman and Schechter ( 2 ) or from the acid via the acid chloride and the acid amide ( I ) , and isonitriles from silver cyanide and the appropriate alkyl iodides (3). The matching, within the limits of experimental error, of retention times obtained on either the polypropylene glycol or a Carbowax column was regarded as indicating a "possible" identity, the possibility being strengthened by comparable matching on both
Table 111. Barbiturate pyrolyzed Na barbital Na butethal Na hexethal Na probarbital
5,5-Substituents ethyl, ethyl ethyl, n-butyl ethyl, n-hexyl ethyl, isopropyl
Na amobarbital
ethyl, 3-methylbutyl
Barbiturates Studied and Their Major Pyrolysis Products
Ka pentobarbital ethyl, 1-methylbutyl Na metharbital
Figure 2.
absorption was determined as a 20Oj, solution in CS2. .A sample of 2-ethylbutanenitrile, prepared from 2-ethylbutyric acid and purified in the preparative chromatograph, was similarly scanned. The two IR spectra may be compared in Figure 2. (The broad absorption band a t about 1150 cm.-' resulted from staining of the microcells. It was present in the spectra of CS2 blanks made at this time, though it had not been present initially.) From the unknown major pyrolyzate of sodium barbital, by reaction with sulfuric acid and subsequent chloroform extraction and sublimation, very fine acicular needles of 2-ethylbutyramide were obtained which melted sharply at 112°-112.50 C. d sample of P-ethylbutyramide prepared from 2-ethylbutyric acid and purified by recrystallization and sublimation also melted at 112°-112.50 C. The mixed melting point of the two was unchanged. By similar methods the major pyrolyzates of sodium butethal and sodium hexethal were identified as 2-ethylhexanenitrile and 2-ethyloctanenitrile, respectively. I n the latter case the retention times were compared by means of extrapolation, as shown in Figure 3. I n the absence of suitable standards for comparison, the identifica-
columns. Standard chemical and physical methods of identification were then followed. The chemical identity of a given pyrolyzate obtained from the pyrolysis unit and by the test-tube method was tested. RESULTS AND DISCUSSION
The pyrolysis products identified are listed in Table 11. h t e n t i o n is called to the wide variety of compounds listed and particularly to the importance and variety of the nitriles among them. Each of the chromatograms obtained from the sodium s d t s of six of the barbituric acids showed a single peak which was conspicuoL sly large compared with the others. Figure 1 shows a representative chromatogram obtained from sodium barbital. Purified samples of 2-ethylbutanenitrile, synthesized by two different routes, were found to have the same retention times as the pyrolyzate of sodiun barbital which produced this large peak. The fraction corr1:sponding to this large peak was collected from the preparative instrument and its infrared CeHi3CH(C2H$CN
F
;
;
;
;
[c
Identification firm firm
virtually certain reasonably certain probable probable
probable
Infrared comparison
Upper curve: major pyrolyzate obtained from sodium barbital Lower curve: 2-ethylbuttinenitrile
F
l-methyl-5,5-diethyl
Major pyrolyzate 2-ethylbutanenitrile Sethylhexanenitrile 2-ethyloctanenitrile 3-methyl-2-ethylbutanenitrile 5-methyl-2-ethylhexanenitrile 3-methyl-2-ethylhexanenitrile (1) methyl isocyanate (2) still unknown
tion of the major products of the sodium salts of probarbital, amobarbital, and pentobarbital was made, in part, by analogy and deduction. Results obtained from all of the six closely related sodium salts indicated t h a t among several routes of degradation one route predominated. The formation of the major pyrolyzates was summarized thus :
Particulars are given in Table 111. Sodium metharbital, the seventh drug studied, produced two very large peaks, one emerging early and one late. The early peak was identified as that of methyl isocyanate; a possible mode of its production is shown in Figure 4. The late peak has not been identified. For purposes of comparison butethal was subjected to parallel runs as the free acid and the sodium salt on the Carbowax 600 column at i 5 O C. Although the retention times of the major pyrolysis products were in exact agreement, the peak obtained from the free acid was noticeably smaller than that derived from the salt. Similarly, when unmodified barbituric acid and sodium barbiturate were compared, the acid failed to give any products which were conspicuously abundant, whereas the salt produced acetone and ethanenitrile in considerable amounts. Much re-
1
ond unknown
9
C$l,CHlC,H$CN
mrtharbitol
C2HsCHlCzHs)CN
minutes
io
'100
O =c\
Columns and remperotures A : Corbowor: 2 0 M ; 125OC
64 Polypropilene glycol; 125OC. C : Corbowo) 6 0 0 ; 7 Y C .
a:observed volues minus 0 5 minute 0:voIues determined by extrapolation
-/N-c\C/c-c
&/
y-c
I\
c-c
CH3
mbssrved volues for unknown minus 0 5 minute
Figure 3. Determination of retention times of 2-ethyloctanenitrile by extrapolation, and comparison with the observed retention times of the major pyrolyzate of sodium hexethal
CHjNCO
Figure 4. A possible source of methyl isocyanate, one of two major products of sodium metharbital VOL. 36,
NO. 4,
APRIL 1964
877
mains to be understood about the interplay of chemical activity among the various parts of these structures, particularly the substituents on the ring. There are indications that some of the nitriles originally formed may react with free radicals to increase their chainlength- for example, the apparent presence of pentanenitrile, hexanenitrile, and 2-methylpropenenitrile among the products of sodium barbital. If this is so, i t is likely that some of the unidentified peaks may also be nitriles. The loss of a side chain, or more than one, may be expected to result in the formation of unsaturated products; hence further studies should give consideration to unsaturated nitri!es, in addition to the propenenitrile which is reported here. Hydrocarbons, the group of products reported by Janak (4, 5 ) , were not explored in this study as they did not appear to be major products. The probable presence of benzene among the products of several barbiturates was noted, and in fact the data reported here are not inconsistent with the presence of a considerable quantity of hydrocarbons. It appears probable that the conditions of chromatography utilized by Janak may have led to retention of nitriles in the column.
Ethanol was the only alcohol indicated, and acetone, the only ketone. The absence of other ketones would suggest that the acetone is formed from the barbituric acid ring itself. Water, which was found among the products of several barbiturates, would probably have been found in all of them if they had all been chromatographed a t 40' C. Gases such as CO, CO,, Sz, 02, and NzO were not investigated in this study. One or more of them might reasonably be expected among the pyrolysis products. A quantitative study of their formation might usefully complement the data presented here. Two interesting sidelights appeared in this study. (a) The matching of retention times revealed the likely identity of an early negative peak as being that of the long-postulated hydrogen isocyanide, H N C ( 7 ) . (b) Contrary to the usual laboratory methods for preparing methane and acetone by the pyrolysis of sodium and calcium acetates, respectively, acetone was found to be the main pyrolytic product of both of these salts under the conditions of this investigation. This may indicate that the hydrocarbons produced by the pyrolysis of the sodium salts of the fatty acids result from secondary reactions among the primary products.
ACKNOWLEDGMENT
We acknowledge the technical assistance of Charles R. Fontan in the preparation of nitriles and isonitriles used as reference standards, and the advice of James Cason of the College of Chemistry. Chemical standards were supplied by the Chemistry Department. LITERATURE CITED
1,LaboratoryJames, Rapoport, Henry, Text in Organic Chemis-
(1 Cason,
try," pp. 72, 84, 89, Prentice-Hall, K.J. ( 2 ) Friedman, L., Schechter, H., J . Org. Chem. 25, 877 (1960). (31 Jackson, H. L., McKiusick, B. C., Org. Syn. 35, 62 (19;5). (4) Janak, J., C'ollectzon of Czech. Chem. Commun. 25, 1780 (1960). ( 5 ) Janak, J., Suture 185, 684 (1960). (6) Selson, D. F., Kirk, P. L., AXAL. CHEW34,899 (1962). ( 7 ) Selson, D. F., Kirk, P. L., J. Chromatog. 12, 167 (1963). (8) Strassburger, J., Brauer, G. M., Tryon, M., Forziati, A. E., Ax.4~. CHEM.32,454 (1960). RECEIVED for revie-- September 5 , 1963. Accepted January 22, 1954. Fall Meeting, 1962, The California Association of Criminalists, Concord, Calif. This work was supported by grants from the Kational Institutes of Health, U. S. Public Health Service (RG-4372 and RG-5802) and from the Research Committee, University of California.
Correlation between Apparent pH and Acid or Base Concentration in ASTM Medium OREST POPOVYCH' Analytical Research Division, Esso Research & Engineering Co., linden, N. 1.
b
Equations were derived which correlate the apparent pH of acid and base solutions in amphiprotic nonaqueous solvents with their concentrations. Within moderate concentration ranges, these relationships are closely approximated by straight-line equations. The slopes of these lines are a function of the nature of the ionic dissociation, while the intercepts depend on the magnitude of the ionic dissociation constant, on changes in the liquid-junction potentials and on the primary medium effect. The above relationships were verified for solutions of perchloric, hydrochloric, SUIfuric, nitric, benzoic, acetic, a mixture of benzoic and acetic acids, potassium hydroxide, piperidine, and lutidine in the ASTM medium. The latter consists of 50.0% toluene, 49.5% isopropyl alcohol, 0.5% water, by volume, and i s used in the ASTM titrations of acids and bases ( I ) . Elec-
878
ANALYTICAL CHEMISTRY
trolytic conductance of HC104, HzS04, HN03, and HCI in the ASTM medium was also studied. A combination of apparent pH and conductance data showed that the sum of liquid-junction and primary medium effects was roughly a constant characteristic of the medium only.
T
definition of pH (8)makes it possible to convert to apparent p H any e.m.f., E , developed between a conventional glasscalomel electrode pair previously standardized against aqueous standard p H buffers : pH
HE OPERATIONAL
=
+
pH, (E
- E8)/(2.3026 RT/F) (1)
where the subscript s designates ap aqueous standard buffer solution. The above procedure coupled with the availability of meters which read pH
directly has extended the measurement of apparent p H to partially and totally nonaqueous solvents. The potential misinterpretation of such p H readings has prompted extensive reviews of the concept and the limitations of p H (2-5, 10, 17, 25). Although it is generally recognized that pH readings do not represent hydrogen ion activity outside of dilute aqueous solutions, the apparent p H in nonaqueous solveiits is by now well established in industrial use as an index of relative acidity. .Is a result, need does exist for obtaining a better insight into the meaning of apparent p H in nonaqueous solvents. I n the petroleum industry, those acidbase measurements which involve ASTM specifications are made in a medium consisting of 50.0yo toluene, 49.5% 1 Present addrese, Department of Chemistry, Brooklyn College of the City University of New York, Brooklyn 10, N. Y.
