Table 1.
Metal A1 Be
v Ti
Ba
Analytical Data
Wavelength, A.
Sensitivity (p.p.m./l%)
3962 2349 3184 4379 3653 3643 3999 3635 3342 3371
6.0 0.2 7.0 100.0 12.0
12.0 16.0 16.0 20.0 53.0 3.5
5535
greatly expanded list of elements susceptible t o analysis by atomic absorption spectrophotometry.
a chloride. All of our experiments were performed with solutions containing an organic solvent. This increased the sensitivity, as has been reported by others ( 1 ) . Using a different experimental arrangement, Fassel and his co-workers (5, 6) have observed the absorption of some of these metals as well as a number of the rare earths. While this report must be considered preliminary, the analytical data in
Salt used Chloride Chloride ...
Organic used Ethanol Ethanol Ethanol
Chloride Chloride Chloride Chloride Chloride Chloride Acetate
Isopropanol Isopropanol Isopropanol Isopropanol Isopropanol Isopropanol Isopropanol
EXPERIMENTAL
Reagents. Tri-n-butylphosphine was prepared b y reaction of phosphorus trichloride with a n excess of n-butyl magnesium chloride in ether under nitrogen (2, 4). The ethereal solution of tri-n-butylphosphine was treated with aqueous ammonium chloride to hydrolyze the excess Grignard reagent, then the organic layer was separated and distilled. The portion boiling a t 149 =t= 1' C. and 50 mm. Hg was taken as the butvlwhosDhine and " _ redistilled (6). Tri-n-butylphosphine sulfide was prepared by direct addition of sulfur t o the butylphosphine under a nitrogen I
254
ANALYTICAL CHEMISTRY
(1) Allan, J. E., Spectrochini. A c t a 17, 467 (1961). (2) illlan, J. E., Zbid., 18, 259 (1962). (3) Box, G. F., Walsh, -I., Ibid., 16, 2a.3 (1960). (4) Dean, J: d.,A n a l y s t 85, 621 (1960). (5) Fassel, V. A,, Proceedings of the
International Symposium on Molecular Structure and Spectroscopy, Tokyo, September 1962. (6) Fassel, V. h.,Curry, R . H., Myers, R. B., Kniseley, R . S . ,Xth Colloquium, College Park, Md., June 1962. (7) Fassel, V. A,, Myers, 11. B., Kniseley, R. S . ,Spectyochitn .-lcta, in press. (8) Gatehouse, B. AI., Willis, J. B., Spectrochiiri. A c t a 17, 710 ( 1961). (9) Gilbert, P. T., 12th .innual Symposium on Spectroscopy, May 1961. (10) Leen, M . W., .itwood, J. G., Pittsburgh Conference, March 1961. (11) Robinson, J. m., ASAL. CHEar. 33,
Table I may help guide others. The sensitivity reported represents the concentration of the metal in parts per million that will produce 1% absorption. We are continuing this work to determine the sensitivities that may be obtained with other metals previously unavailable by atomic absorption spectrophotometry. Because emission spectra have been observed from all but a few of the metallic and semimetallic elements in flames, we may expect a
Tri-n-butyl phosphine Sulfide SIR: This paper deals with a brief investigation into the behavior of trin-butylphosphine sulfide as an extractant for metal ions. This class of organic extractant, not previously investigated, provides the sulfur analog to the trialkylphosphine oxides (which includes well known TOPO). A trialkylphosphine sulfide seemed preferable to an aryl derivative because it was expected that the former compound would be prepared more easily and, furthermore, the corresponding trialkylphosphine oxides have received more study than the triarylphosphine oxides.
LITERATURE CITED
Table
I.
1067 (1961).
WALTERSLAVIX D. C. MAS~SIXG
The Perkin-Elmer Corp. Norwalk, Conn. R E C X I ~ Zfor D review Xovember 8, 1962. accepted December 19, 1962. Presented in part at the Xinth Ottawa Symposium Ottawa, Canada, September 1962.
as an Organic Extractant
Survey of Extraction of the Elements by Tri-n-Butylphosphine Sulfide)
Degree of extractiono Element
5111 HCl
Aluminum* Antimony( 111) Bismuth Boronbpc Cadmium Cerium( 111) Chromium(111) Cobalt Copper(11) Iridium Iron( 111) Lead( 11) Manganese( 11) Mercury( 11) Ruthenium Selenium(IV) Silver Strontium Thallium( 111) Thorium Tin(I1) Vanadium(V)b Zinc Zirconium-niobium
N
a
C
0.1M HCl
Water
Aqueous 3"
N
N N
iv N
-2' P
N N N
av 1V
N
N
N
P
S-P
N = no extraction (99% extraction). Extraction followed by flame spectrometric methods. Reagent concentration was 0.05M.
= b
N
1111HCl
atmosphere until the liquid became faintly yellow. As the addition of sulfur progressed, it became necessary t o heat the reaction mixture sufficiently to melt t h e sulfur. As prepared, the tri-n-butylphosphine sulfide contained small amounts of unreacted sulfur and tri-vi-butylphosphine. Tri-n-butylphosphine sulfide is now available commercially. Procedure. T h e distribution ratio, D , was measured using gamma-emitting radioisotopes. T h e activity of each phase was determined b y counting in a gamma-scintillation counter. Carrier solutions contained approximately 10-251 of the metal. A 5% (v./v.) solution of tri-n-butylphosphine sulfide in CCI, n-as used. Equal volumes, 10 ml., of both phases were contacted for a period of 30 minutes. Phase separation was assisted by centrifugation a t 2000 r.p.m. for 10 minutes. Elements for n-hich suitable radiotracers mere unavailable were checked using flame spectrometric methods. For these, 0.1Msolutions of the reagent in cyclohexane were used. RESULTS AND DISCUSSION
The results of the survey of 21 elements are listed in Table T. Similar
studies have been reported for three trialkylphosphine oxides b y Ross and White (6, 7 ) . Tri-n-butylphosphine sulfide is much more selective than the trialkylphosphine oxides. Only tIvo metal ions among those tested extracted well, silver(1) ( D = 625) and mercury(I1) (D = 282), both class b acceptors according to Ahrland, Chatt, and Davies ( 1 ) . Elements which are partially extracted arc border region acceptors and included bismuth ( D = 0.56), copper(I1) ( D = 0.026 from 0.1M HCl), lead ( D = 0.27), and zinc ( D = 0.27). The pronounced selectivity of the reagent for silver and mercury is in accord with that reported for neutral phosphate esters containing a thiophosphoryl group, such as tri-nbutyl- or triisooctyl thiophosphate (S), but contrasts sharply with the behavior observed for trialkylphosphine o.cides. The latter compounds react largely with class a acceptors. Triphenyl phosphine sulfide in CHC1, also extracted silver iD = 520) well from 6M HKOa. It proved very difficult to strip niercury(I1) from the organic phase n-ith common reagents such as nitric acid, aqua regia, or sodium peroxide.
LITERATURE CITED
(1) Bhrland, S., Chatt, J., Davies, N.
R.,
Quart. Rev. (London) 12,265 (1958). ( 2 ) Baldwin, W. H., Oak Ridge National Laboratory, Oak Ridge, Tenn., private communication. 1960. (3) Handley, T. 'H., Dean, J. A , , A 4 ~ ~ ~ . CHEM.32, 1878 (1960). (4)Kosolapoff, G. ill., "Organophoephorus Compounds," p. 16, Wiley, Sew Tork, 1950. (5) Ibid., p. 32. 16) Ross. W. J.. U. S. Atomic Enerw Con& Rept. ORNL-CF-56-9-18 (1956). (7) Ross, W. J., Khite, J. C., Ibid., ORNL-CF-57-1-5 (1957). \
,
ROBERTB. HITCHCOCK' JOHN -4.DEAN Department of Chemistry University of Tennessee Knoxville 16, Tenn. THOMAS H. HANDLEY
Analytical Chemistry Division Oak Ridge Sational Laboratory Oak Ridge, Tenn. RECEIVED for review November 16, 1962. Accepted December 5, 1962. Taken in part from the Ph.D. thesis of R.B. Hitchcock, submitted t o the Graduate School of the University of Tennessee, August 1962. 1 Present address, Chemistry Department, Southwestern Louisiana University, Lafayette, La.
Determination of Hydrogen Cyanide and Cyanogen by Gas Chromatography SIR: I n the course of a study of the reactions of hydrogen cyanide, we have developed a chromatographic method for the determination of hydrogen cyanide and cyanogen. M7001mington ( 2 ) separated hydrogen cyanide, water, and nitrogen with a column of Celite or Chromosorb impregnated with 20% polyethylene glycol 1500. On the basis of his results and the solubilities of hydrogen cyanide and cyanogen, we examined columns of Chromosorb coated with several moderately polar, lorn molecular weight liquids for the separation of hydrogen cyanide, cganogen, and carbon dioxide. Glycerol, diethylene glycol, and triethanolamine were not suitable, but triacetin (glyceryl triacetate, KO. 256, Eastman Chemical Products, Inc.) or tributyrin (glyceryl tributyrate, S o . 726, Eastman) effected the desired separation. Triacetin seemed t o give better results. -4 triacetin column can be used for short periods at 90" C. and for several months
a t 60" to i 5 " C. Triacetin separates chlorine and cyanogen chloride also, but the effectof continued exposure of tile column and the instrument to these gases is not known. Type 5A or 13X molecular sieve (Wilkins Instrument and Research,
Component COI
CI? (CY 12 0 9
CXC1 ?r'?
HCN
Inc.) was used as received to separate oxygen, nitrogen, methane, and carbon monoxide. Each material also resolved nitric oxide in the absence of oxygen and methane. Two columns, each prepared from '/(-inch (5 mm. i.d.) copper tubing, were used in series. The first column
Table 1. Retention Time and Volumetric Factors Retention time, minutes Volumetric factor,a K', Molecular sieve with molecular sieve Triacetin 5A 13X' 5A 13X 0.54 mb m 1.38 1.25 __ 0.80 m m 1.02 m m 1.66 1.60 0.4sc 1 .'io 1.46 1.00 1.oo 2.03
0.48
4.35
m
3.35 m
m
2.06 m
-_
1.10 1. O O d
5,33 2.87 1.03 CH, 5.52 3.25 0.99 co 0.48 15.3 4.00 1.00 a Determined by method of Johns ( l ) ,based on peak area. * Irreversibly adsorbed. Composite of permanent gases was eluted in 0.48 minute. Determined with mixtures containing 35y0 HCN or less, K' = 1.00. KO
0.18 0.48
-
1.09
1.OOd
1.06 1 00
1,18
VOL. 35, NO. 2, FEBRUARY 1963
255
