390
ANALYTICAL CHEMISTRY
only a fraction of a milligram may be accurately analyzed by this technique. By observing carbonization in the ampoule during combustion, and the background in the mass spectrometer in the mass range 27 to 52, it was found that 1 hour at 800" C. was a satisfactory time for the combustion procedure. A large peak a t m / e 27 for some nitrogen compounds, as well as above average peaks in the mass range 47 to 52, was present when much shorter combustion times or lower temperatures were used.
LITERATURE CITED
Consolidated Engineering Corp., "Operation and Maintenance Manual," CEC.2012a, p. 6:08, 1948. (2) Crosse, A. V., Hindin, S.G., Kirschenbaum, A. D., A 4 CHEM. ~ ~ ~ . (1)
21, 386-90 (1949). (3) Kirsten, W., Ibid., 26, 1097 (1954). (4) Siederl, J. B., Trauts, 0. R., IND.ENG.CHEM..ASAL. En. 3, 151-2 (1931).
RECEIVED for review
May 2 5 , 1955. iccepted Xovernber 22, 1955.
Isotope Dilution-Static Combination Method for
Organic Carbon in Submilligram Specimens R. N. BOOS, S. L. JONES, and N. R. TRENNER Research Laboratories, Merck & Co., Inc., Rahway, N. J
The method of elemental anall-sis described was undertaken to fill the need for a rapid, convenient, and precise means for quantitative determination of elemental carbon, especially where the available sample is very small, as in paper chromatography. The method enables one to determine elemental carbon in organic compounds with a precision comparable to that achieved in the standard micro-Pregl determination but with sample sizes of well under 1 mg. It is rapid, convenient, and easily adaptable to routine operation.
THE
quantitative determination of the carbon content of submilligram specimens has become of ever-increasing importance because of the current availability and application of micromethods of isolation of new naturally occurring products. Unterzaucher (5)described a method for the determination of carbon and hydrogen in submilligram quantities of material based on the determination of oxygen in the water and carbon dioxide resulting from the combustion of the sample. Kirsten ( 4 ) has stated, however, that Unterzaucher's procedure adds the troubles of the oxygen determination to those of the carbon and hydrogen determination. Grosse, Hindin, and Kirshenbaum ( 2 ) introduced an isotope dilution procedure for the determination of carbon, oxygen, and nitrogen on 20- to 60-mg. quantities of organic compounds. The method provides for the static combustion of the sample after addition of a known quantity of C13O2, O2**,or X21b volumetrically, equilibration of the combustion gases, and subsequent measurement of the isotope ratio in the gas mixture. Rather than measure the tracer volumetrically, it was felt that i t would be more convenient to use the same method of measurement as that required for the sample. The method, herein presented, is based on the static combustion of an accurately weighed mixture of the unknown material and a known carbon-13-labeled tracer compound. ThP ratio of carbon dioxide of mass 45 to carbon dioxide of mass 44 is compared with the ratio for the gases obtained when the tracer alone is burned. The combustion is performed in evacuated, sealed Vycor ampoules in a furnace a t 775" to 800" C. The ampoule, after completion of the combustion, is broken within the evacuated inlet system of the mass spectrometer and the C1302/C1202ratio ascertained. Recently Wilzbach and Sykes (6) published a somewhat similar combustion technique for the ultimate determination of carbon14.
* Present address, General Electric Co.. Schenectady,
N. T.
APPARATUS AND REAGEYTS
The samples were weighed into platinum boats after it was ascertained that rather high carbon dioxide blanks were obtained when boats fashioned from electrolytic copper foil were employed. The Vycor ampoules, ampoule breaker, and absorption tower were similar to those employed by Jones and Trenner (5),except that the absorption tower contained only Dehydrite. A Consolidated-Nier mass spectrometer, Model 21-201, was used for the isotope ratio measurements after the gas-handling system (Figure 1) had been changed to include a Toepler pump, so that smaller volumes of gas could be manipulated and compressed to give a satisfactory inlet pressure (see Figure 1) and to allow for the removal of excess gas when necessary. The cupric oxide was prepared by passing dry electrolytic oxygen over pure electrolytic copper powder a t 800" C. in a silica tube. Succinic acid, containing 30 atom % carbon-13, was prepared by refluxing overnight on the steam bath 3.5 ml. of ethylene bromide and 6.25 grams (20% excess) potassium cyanide (60 atom % CI3) in 40 ml. of 50% ethyl alcohol. Following distillation of the solvent and drying of the residue, the succinonitrile was extracted with four 50-ml. portions of absolute ethyl alcohol and filtered, and the solution was evaporated to dryness. The residue was treated with 50 ml. of concentrated hydrochloric acid and evaporated to dryness on the steam bath. The resulting succinic acid was extracted with four 50-ml. portions of methyl ethyl ketone. The solution was treated with activated Xorite to decolorize it, filtered, and concentrated to crystallize the succinic acid. The acid was recrystallized from ethyl acetate, sublimed a t 117' C. under vacuum, and recrystallized again from ethyl acetate. The filtered crystals were dried for 4 hours under vacuum (melting point = 189-192' C.). The yield was 57% (2.7 grams) based on the ethylene bromide. ANALYTICAL PROCEDURE
The sample and a quantity of succinic acid tracer, containing approximately the same weight of carbon as that expected in the sample, are weighed accurately into a platinum boat. The boat and approximately 50 mg. of cupric oxide are put into a 6-inch Vycor ampoule marked with a diamond pencil for identification. After the lower half of the ampoule has been cooled in a dry iceacetone bath, the center portion is heated in the oxygen-hydrogen flame to constrict it to capillary size for subsequent seal-off. The ampoule is connected to the manifold of a mercury diffusion pump and evacuated to a few microns before the final sealing at the constriction. Once sealed, the ampoule is rotated to effect an intimate mixture of the contents, and put into a furnace preheated to 775' to 800" C. for 1 hour. After the completion of the combustion, the ampoule is removed from the hot furnace, cooled to room temperature, and transferred to the ampoule breaker, which is attached as shown in Figure 1 to the gas-handling train of the mass spectrometer. After the entire inlet system has been evacuated for 5 minutes, with a mercury diffusion pump, all the stopcocks are closed, the ampoule is broken, the combustion gases are dried for approximately 1 minute over the Dehydrite in the absorption tower, and in the meantime the amplifiers of the instrument are balanced.
391
V O L U M E 28, NO. 3, M A R C H 1 9 5 6 The dried combustion gases are then so compressed by means
and for carbon-12 is:
of the Toepler pump against the mass spectrometer that the ion voltage on collector 1 is equal to 20 volts. After the ClSOz of mass 45 is peaked on collector 2, a t an ion-accelerating voltage of 1192 volts, the C130z/C120z or mass 45/44 ratio is read directly
on the instrument panel. Several ratio values are obtained and the average value is used in the calculations. After the average ratio value has been multiplied by theffactor of the mass spectrometer ( 1 ) and corrected for the natural abundance of oxygen-17 and oxygen-18, it is converted to atom per cent excess carbon-13 in the usual manner.
where
TVT = weight (in mg.) of tracer in a given combustion -TIT = molecular weight of the succinic acid tracer
nz TT
= number of carbon atoms per molecule of tracer = true C13/C12 ratio of the tracer a t large specimen
weights, where the background is negligible
1 11 X 10-2 and 0 989 = known natural abundances of
The mass spectrometer measures:
C13 and C12, respectively
W C = weight (in mg.) of the carbon background C'3 Let T$ = - be the mass spectroscopically observed ratio for C'Z a given combustion of this tracer only. Then by Equations 4 and 5
Iff is the fraction of atoms of a given kind, then
12.01 WT '6 ( T T
IVc = .Z~T(TT
- T$)
+ 1) (0.989 - 1.11 X T$
(6)
and
(3)
A study of the carbon-13 content of the Euccinic acid tracer, carried out using varied specimen weights, revealed that there was a small but significant carbon contamination (background) involved in the apparatus and technique. This, of course, became of greater importance as the tracer specimen weight used in a given combustion was decreased. This background effect was evaluated in the follotying manner: Let T refer to the tracer, whence the number of mole atoms of carbon-13 is given by:
For the succinic acid tracer of C4Hs04
whence M T = 119.3 Equation 6 then reduces to: 0,2761 W~(0.4576- T $ ) - 1.11 x 10-2
IVC = 0.989 T $
(7)
Table I shows the application of Equation 7 to the experimental observations carried out on this tracer. The results clearly
Figure 1. Modified gas-handling system for Consolidated-Nier mass spectrometer
ANALYTICAL CHEMISTRY
392 illustrate the statistical constancy and magnitude of the background effect encountered in this work. I n view of the good constancy of WCit is appropriate t o derive the follonring corrected equation for use in computing carbon analyses:
~
1
r z x nE x rx 1 + Ms ws x ng X 0.989 +
= 1 W
wc
__ XO.989 (9) 12.01
C‘S
e = rs
where Ws, M s , and n8 are the same quantities as used above but refer to the sample being analyzed. r,s is the observed mass spectroscopic ratio in t,he combustion of a sample plus tracer mixture. If X C is the quantity sought-Le., the weight fraction of carbon in the sample-then
____-.___
Table 11. D e t e r m i n a t i o n of Carbon in Known C o m p o u n d s Compound Sample Carbon, Wt., % Mg. Succinic acid 40.7 1.307 1.228 0.910 0.878 0.719 0.571 0.497 0.447 0.255 0.156
Tracer Wt., Mg.
rs
Carbon Found, %
Mean
0.940 1,047 0.986 0.932 0.684 0.540 0.902 1.521 0.260 0.181
0,1572 0.1752 0,1995 0.1990 0.1880 0,1861 0.2587 0,2101 0.1938 0,2059
41.2 40.9 40.4 40.6 40.3 40.7 40.4 40.1 40.4 40.6
40.610.3
Benzamide
69.4
0.987 0.913
1.356 1.390
0.1682 0,1797
70.4 69.9 70.1-1-0 2
P-Nitrobenzoic acid
50.3
1.177 0.881 0.804 0.167
1.958 1.151 1.453 0.144
0.2351 0,1965 0.2335 0,1555
50.3 51.1 50.7 49.2
50.3 1 0 . 3
Nicotinic acid
58.5
1.186 0.781 0.661
1.232 0,990 1.323
0.1602 0.1786 0.2290
57.8 58.8 58.4
58.3 =to 2
Cystine
30.0
1.254 1.059 0.481 0.395 0.385 0.365
1.192 1.694 0.462 0.565 0,345 0.204
0.2192 0.2788 0.2218 0.2657 0.2135 0.1594
30.3 29.7 29.7 29.6 29.7 30.8
30.010 2
Cortisone
69.9
0.542 0.509
0.862 0.818
0,1849 0.1852
69.4 69.9 69.6 z t O . 2
9wFluorohydro- 65.4 cortisone acetate
0.731 0.560
0.588
1.042
0.1783 0.1483
66.3
65,s
66.0 2 ~ 0 . 2
~~
SUMMARY
Combining Equations 8, 9, 10, and 11 and solving for X C one gets:
The proposed method has made possible the quantitative determination of carbon in organic compounds on samples as small as 200 y with an accuracy within lY0 using an Ainsworth microanalytical balance. Recently a hlodel E ultramicrobalance has been obtained which is sensitive to 0.05 y. This quartz fiber torsion balance is supplied by Microtech Services Co, Box 121, Berkeley, Calif., and is excellent for this type of analysis. LITERATURE CITED
Table I. WT
0.187 0 216 0.233 0.363 0,380 0.423
C a r b o n Background Effect T+
(Obsd.)
U’C
0 4444 0 4434 0 4456 0.4473 0 4507 0 4503
x
103
1 64 2 04 1.86 2.49 1 7fi 2 07 1 98 =t 0 22
Mean
( 1 ) Consolidated Engineering Corp., “Operation and Maintenance Nanual,” CEC-ZOlZa, p . 6:08, 1948. ( 2 ) Grosse, A. T’., Hindin, S.G . , Kirshenbaum, =1. D., 4 N . 4 L . CHEM.
21, 386-90 (1940). (3) Jones, S. L., Trenner, S . R., Ibid.,28, 387 (1956). (4) Kirsten, W., Ibid.,25, 74-86 (1953). (5) Unterzaucher, J., Chem.-lng.-Tech. 22, 39 (1950). (6) Wilebach, K. E., Sykes, W.Y., Science 120, 494-6 (1954) RECEIVEXI for review l l a y 2 5 . 19.53. .%cc:pteri Sovernber 22, 1955.
As in this work rT = 0.4576, ng = 4, M T 1.98 X 10-3, Equation 12 becomes:
=
119.3, and Wc =
Displacement of the Nitro Group during Determination of Nitrophenols and Nitroanalines by the Koppeschaar Method-Correction the “corrected” equation used for computing the analytical results given in Table 11, which shows this method applied to compounds of known carbon content and varied structural types. In carrying out isotope dilution assays of this kind maximum mass spectroscopic precision is achieved ( 3 ) \%-hen rs E 1/2 rT
kept a t about 1.7.
2x0
+
0 2 +
2x02
The equation on page 1498 should read:
Thus for an average organic compound where, generally, X C = 0.50, Equation 13 dictates that the quantity
I n the article on “Displacement of the Nitro Group during Determination of Nitrophenols and Nitroanilines by the Koppeschaar Method [ A N A L .C H E x 27, 1494 (1955)], the eighth equation on page 1494 should read:
should be
HgO
+ 2Br2 + HQO
+ 2HOBr
+ HgBrz L. D. JOHNSON W. hl. M C ~ A B B E. C. WAGNER