considerable background activity, mainly due to the presence of Na24. The activity a t the point mhere 0.002 ininole of sodium nitrate had been spotted was about 20 times that due to background (at 1.4 m.e.v.), so the sodium n-as easily detected in the presence of the background. On the sheet on which an equivalent amount of potassium nitrate was spotted, no activity due to K42 could be found. This mas expected because of the low rrlative abundance of K4’, the isotope which forms radioactive K42. &%ftera total of 144 hours of decay, the background due to NaZ4 was negligible, although some background activity due to unknown impurities remained. The activity at the place in the paper where 0.002 mmole of sodium had been spotted was no higher than background. A number of rubidium and cesium spots were scanned with the y-ray spectrometer. The total counts under the energy maxima for the two ions were determined (rubidium at 1.08 m.e.17.; cesium at 0.60 and 0.80 m.e.17.). Plots of this activity us. mmole of salt taken were linear in the range of 0,0001 to 0.001 mmole for each ion. For equal molar quantities, the cesium spots were about 100 times more active than the rubidium spots. This reflects the much greater cross section of cesium compared with rubidium. These preliminary observations
warrant some significant conclusions regarding the combination of electrochromatography and neutron activation for the estimation of the alkali metal cations. U‘ith irradiation of the ions separated in the paper, determination of sodium and potassium is interfered with by the high background of the paper, and the determination of cesium and rubidium is limited by the instability of the paper which restricts the degree to which these ions may be activated. Activation of the mixture before separation of the group of alkali metal cations from all the other cations eliminates the paper background. With short irradiation periods plus y-ray spectrometry, it provides an effective method for sodium and potassium. With longer irradiation periods plus “cooling” and y-ray spectrometry, it provides an extremely sensitive method for rubidium and cesium. Activation after electrochromatographic separation from all other cations and subsequent elution from the paper has the same features as activation before electrical migration and reduces the total activity due to the cations removed by migration. For the determination of sodium and potassium, neutron activation offers few advantages over flame photometry. For the determination of rubidium and cesium, electrochromatography preceded or followed by neutron activation, folloffed in turn by “cooling” and y-ray
spectrometry, offers enormous sensitivity. ACKNOWLEDGMENT
The authors are indebted to the staff of the Argonne CP-5 reactor for irradiating various samples; to the Radiation Safety Division of the laboratory for guidance in handling radioactive material; and to R. W. Bane and J. J. Hines of the Analytical Division of the laboratory for helpful discussions and for performing various analyses. LITERATURE CITED
(1) Barreto, H. S. R., Barreto, R. C. R.,
J . Chromatog. 4, 153 (1960). (2) Cook, C. S., Am. Scientist 45, 245 (1957). (3) Erlenmeyer, H., Hahn, H. v., Sorkin, E., Helv. Chim. Acta 34, 1419 (1951). (4) Evans, G. H., Strain, H. H., -4NAL. CHEM.28, 1560 (1956). ( 5 ) hleinke, W. IT., Science 121, 177 (1955). (6) Miller, C. C., Magee, R. J., J . Chem. SOC.1951, 3183. ( 7 ) Strain, H. H., Binder, J. F., Evans, G. H., Frame, H. D., Jr., Hines, J. J., AKAL.CHEM.33, 527 (1961). (8) Tuckerman, bl. M., Strain, H. H., Ibid., 32, 695 (19601. (9) Wood, S. E., Strain, H. H., Zbid., 26, 1869 (1954). RECEIVED for review September 22, 1961. Accepted October 26, 1961. Based on work performed under the auspices of the 11.S. Atomic Energy Commission.
Paper Chromatography of Organic Mercury Compounds JOSEPH
N. BARTLETT
and GEORGE
W. CURTIS’
Department of Chemistry, St. Joseph’s College, Philadelphia 3 7 , Pa.
b A procedure has been developed for the paper chromatography of organic mercury compounds. Several developing systems may be used, if aqueous ammonia is one of the components. The most suitable combinaethanol-28% tion is 1 -butanol-95% ammonia (8: 1 :3), A chloroform sohtion of dithizone or an aqueous sodium stannite solution may b e used as spraying agent; the former i s more sensitive but less selective than the latter. Both the ascending and the descending techniques were used and the R, values for several mercurials are reported.
D
purity and stability studies of some organic mercury compounds, a method for detecting very small amounts of mercurials in the pres80
WRING
ANALYTICAL CHEMISTRY
ence of relatively large amounts of other mercurials was necessary. Kone of the methods reported in the literature could be satisfactorily adapted, so this led to the development of a paper chromatographic method for mercurials in general. APPARATUS
The development tank for the ascending technique is a KO.4 American Xedical Museum jar. The tank used for the descending technique is the conventional 12 x 24 inch glass cylindrical jar with a plate glass lid. The tray for holding the solvent and the paper was made of stainless steel after the design of Porter ( 5 ) . While Whatman No. 1 filter paper was used to obtain the data reported in this paper, Eaton-Dikeman papers 048 and 248 gave good results but
different R/ values than the Rhatman paper. REAGENTS
The following reagent grade chemicals were used without further purification: acetone, 28% aqueous ammonia, 1butanol, chloroform, dioxane, dithizone (Eastman Kodak Co. White Label), ethanol, ethyl acetate, and stannous chloride. The mercurials were obtained from their manufacturers or synthesized in this laboratory. Stock solutions of organic mercury compounds were made by dissolving sufficient mercurial in dioxane to contain 5 pg. of mercury per microliter of solution. 1 Present address, Socony Mobil Oil Co., Inc., Paulsboro, N. J.
EXPERIMENTAL
Before attempting to find a solvent system for developing chromatograms of colorless substances it is necessary to find a suitable spraying agent to locate the spots. Miller, Polley, and Gould (4) had used dithizone in the spectrophotometric determination of mercurials and a chloroform solution of this compound 11 as tested as a spraying agent. Yellow t o pink spots appeared upon spraying filter paper which had been spotted with mercurials. Another spraying agent, not so sensitive as dithizone but much more selective, is a 0.3X solution of freshly prepared sodium stannite, n hich gives gray to black spots of free mercury. Because of the greater sensitirity, dithizone nas used almost exclusively in this investigation, except when there was some doubt as to whether or not a given spot was actually caused by a mercurial, whereupon its presence or absence v a s verified with sodium stannite. Using the test tube technique of Rockland and Dunn ( 7 ) , a large numIber of solvent systems were tested, invluding butanol saturated with water, I\ hich had been used by Kanazawa et al. ( 1 ) . Little success for the purpose a t hand was obtained until aqueous ammonia was introduced into the system. Kanazawa and Sat8 (2) subsequently used this same component in their work nith some of the organic mercury compounds. The most satisfactory system was 1-butanol-95yo ethanol-ZS~o aqueous ammonia (B-E-A) (8:1 :3). Ethyl acetate-95yo ethanol-Z8% aqueous ammonia (EA-E-A) (8 : 3 :3) worked well and moved about twice as fast as the butanol system but did not give such well defined and reproducible spots. The qnme was true of the acetone-ammonia .ystem(A-A) (1 :1). All of these mixtures are single-phase systems, which greatly facilitates their preparation and use. Procedure. The small scale ascending technique of paper chromatography developed by Rockland, Blatt, and Dunn (6) was first used and proved successful. The filter paper, spotted with dioxane solutions of mercurials, 2 cin. apart and each spot containing 3 to 5 pg. of mercury, was clamped between halves of wooden dowels 10 mm. in diameter and suspended in the museum jar. The dimensions of the paper were approximately 12 by 14 t:m. No equilibration time was required and the temperature was maintained a t 23" i- 1" C. The time required for the solvent to rise 10 em. ranged from one-half hour for the EAE--4 system t o 1 hour for the B-E-A system. A 0.005% solution of dithizone in chloroform was used to spray the chromatograms. The Rf values are shown in Table I.
To obtain better resolution of mixtures of mercurials, the so-called "large scale" descending technique was also used. The hole in the lid of the large cylindrical tank was fitted with a Bunsen valve to relieve the ammonia pressure which occurred when the developing solvent was added to the tray in the jar. Strips of filter paper 17 cm. wide and 54 em. long were spotted with mercurials 2 em. apart and suspended from the stainless steel tray. The paper was equilibrated for 4 hours, using the same solvent as was subsequently used for development of the chromatograms. It required 13 hours for the B-E-A system to travel 40 cm., whereas it took approximately half that time for the EA-E--4 system to travel the same distance. Bfter development] the papers were thoroughly dried and sprayed with dithizone solution as in the small scale method. To compensate for any heterogeneity in the paper structure or any unevenness of the solvent front, the R, (ratio to a standard) of several mercurials were determined by adding a reference standard, which in this case was phenylmercuric chloride, to each spot of mercurial. The ratio of the distance traveled by a given mercurial to that traveled by the phenylmercuric chloride, using the descending technique, is recorded in Table 11. Because some of the simple inorganic mercury salts are often used in the synthesis of mercurials, a number of these salts were subjected to the same chromatographic procedure as used for the mercurials. Therefore the chloride, acetate, sulfate, and nitrate salts were individually chromatographed for 13 to 14 hours; in each case the spot did not migrate enough from the origin to obtain a measurable Rf value. DISCUSSION
While the Rf values for the EA-E--4 system are not listed, they were similar to those found in Table I, but the spots were not as sharply defined. The various phenylmercuric salts such as the nitrate, chloride, and acetate gave, as expected, virtually the same Rfvalues and hence cannot be separated by any of the solvent systems tested. Because of the large difference in R, values between the 0- and p-isomers of chloromercuriphenol, it was possible to detect the presence of a small amount of p-isomer in a sample of the o-isomer which had been subjected t o repeated recrystallizations and gave indications of high purity by the conventional methods. Di-p-tolylmercury gave no visible color reaction with dithizone solution until the paper had been exposed to HC1 fumes for 15 to 20 seconds after drying. This apparently resulted in a cleavage of
Table 1.
Rj Values of Some Organic Mercury Compounds
(B-E-A developing system) Small Large Scale Scale (As(Decend- scendCompound ing) ing) 0.39 Phenylmercuric chloride 0.38 0.49 Tolylmercuric chloride 0.48 0.22 o-Chloromercuriphenol 0.28 0.07 p-Chloromercuriphenol 0.07 ... 0.94 Di-p-lylmercury 0.22 Mercaptomerin sodium. 0.16 0.17 0.15 Mercuhydrinb 0.43 0.41 Mercuritalc 0.70 0.75 Merthiolated 0.40 0.39 SalyrganO 4 N-( r-carboxymethylmercaptomercurip-methoxy)propylcaniphoramic acid disodium salt. b N - ( p - hydroxymercuri - y - niethoxy)propyl-N '-succinylurea. Sodium salt of 0-[(3-hydroxymercuri2-methoxypropyl) carbamyl] phenoxyacetic acid. d Sodium ethylmercurithiosalicylate. 0,e
Table II.
R, Values of Some Organic Mercury Compounds
(Phenylmercuric chloride used as standard) Compound R. Value o-Chloromercuriphenol 0.61 p-Chloromercuriphenol 0.19 Tolylmercuric chloride 1.22 Mercaptomerin sodium 0.38 Mercuhydrin 0.28 Merthiolate 2.02
the molecule suggested by IUlarasch and Flenner (3) as follows:
Mercurochrome (disodium salt of dibromohydroxymercurifluorescein) was chromatographed only with difficulty because of excessive tailing when the period of development was the same as for other mercurials. It was necessary to extend the development time to 33 or 34 hours to observe fluorescent spots under the ultraviolet. Eight or nine spots were found, but no Rfvalue could be obtained because the solyent ran off the paper. Several other commercially available mercurials used as diuretics or antiseptics yielded one and sometimes two organic mercury spots in addition to the main constituent, indicating impurities. The fact that inorganic mercury salts do not give measurable R, values is an obvious advantage of the method, because it provides a means of differentiating between organic and inorganic mercury impurities in a given sample. The choice of techniques for a separation depends largely upon the resolution demanded, which, in turn, depends upon the difference in Rf values of the VOL. 34, NO. 1, JANUARY 1962
e
81
components in the mixture. The small scale ascending technique has the advantage of speed, since the total time required to develop, dry, and spray the chromatogram is only about 2 hours, compared to about 20 hours for the larger scale descending technique. study of the concentration limits showed that the least amount of mercurial which could be detected after chromatographic development contained from 1 t o 1.5 pg. of mercury. The mavimuni amount n hich could be
chromatographed without tailing contained 20 pg. of mercury. From Table I it appears that the method works regardless of whether the mercury atom is attached to one or more benzene rings or whether it is in a straight-chain compound. Hence it is believed to be general in application. LITERATURE CITED
(1) Kanazawa, Jun, Koiama, Iiiyoshi, Aya, Masahiro, Satb, Rokurb, .Yzppon I\'ogei-Kagaku Kaashz 31, 872 (1957).
(2) Kanazawa, Jun, Satb, Rokur6, Bunseki Kagaku 8,322-3 (1959). (3) Kharasch, 11. S., Flenner, -4. L., J . Am. Chem. Soc. 54, 675 (1932). (4) Miller, Y.L., Polley, D., Gould, C. J., ANAL.CHEM.23, 1286 (1951). (5) Porter, W.L., Zbid., 26, 439 (1954). (6) Rockland, L. B., Blatt, J. L., Dunn, 11.S., Ibid., 23, 1142 (1951). (7) Rockland, L. B., Dunn, 11. S., Science 109, 539 (1954).
RECEIVEDfor review June 30, 1961. Accepted October 30, 1961. Division of -4nalytical Chemistry, 134th Meeting, ACS, Chicago, Ill., September 1958.
Determination of Normal Paraffins in C,, to C,, Paraffin Waxes by Molecular Sieve Adsorption Molecular Weight Distribution by Gas-liquid Chromatography JOHN G. O'CONNOR, FRANK H. BUROW, and MATTHEW S. NORRIS Gulf Research & Development Co., Pittsburgh, Pa.
b A method is described for the determination of normal paraffins in CZOto C32 paraffin waxes. Adsorption of the normal paraffins is accomplished b y refluxing an iso-octane solution of the wax in the presence of pelleted Molecular Sieves, Type 5-A. The sieves are filtered from solution and the solvent is evaporated to leave a residue consisting of branched and cyclic paraffins. The normal paraffin content of the wax i s represented by the weight difference between the unadsorbed residue and the original sample. By employing this method, the normal paraffin content can be determined with a standof the amount urd deviation of 1.1 present. Also described are procedures for recovering the adsorbed normal paraffins and determining the molecular weight distribution of the w a x by gas-liquid chromatography.
column of Molecular Sieves. Extension of this procedure to waxes was not possible, however, because the column did not permit sufficient contact time for complete adsorption. To overcome this difficulty, a batchwise adsorption was investigated wherein the wax was ad-
flIA
%
A
knowledge of the hydrocarbon types in paraffin waxes is of interest to the petroleum industry and serves in evaluating the physical characteristics of these products. For instance, a high normal paraffin content tends to make a wax brittle, whereas the presence of branched and cyclic paraffins constitutes a more flexible type wax. Recently, the authors reported (3) that normal paraffins can be determined in the gasoline to light gas oil range (c6 to Cz0) by adsorption on a 82
PRECISE
ANALYTICAL CHEMISTRY
Figure 1. apparatus A.
B. C.
Adsorption
24/40 T outer, fitted with water condenser Erlenmeyer flask, 250-ml. capacity Fritted-glass disk, medium
sorbed from a boiling solution. This technique showed considerable promise and was developed into a method which permits an accurate determination of the normal paraffin content of waxes of chain length from C20 to C32. The method also provides for the recovery of adsorbed normal paraffins and determination of their carbon number distribution by gas-liquid chromatography. The standard deviation calculated for the accuracy is 1.1%. MATERIALS AND APPARATUS
Solvents.
Iso-octane and isopentane (pure grade, Phillips Petroleum Co., Bartlesville, Okla.). These solvents were purified by percolating 500 ml. through a column (2.5 cm. X 1 meter) packed with powdered Molecular Sieves. At a flow rate of 100 ml. of eluent per hour all traces of normal hydrocarbon impurity were removed. n-Pentane (pure grade, Phillips Petroleum Co., Bartlesville, Okla.). This solvent does not require a pretreatment step. Adsorbents. Molecular Sieves, Type 5A, '/&nch pellets and Type 5A. powder (Linde Air Products Co., New York, K.Y.). Both sieves were dried for 6 hours a t 250' C. a t 1 to 5 mm. of mercury pressure in a vacuum oven. Folloning the drying procedure the oven was cooled under vacuum and vented to dry air. Llolecular Sieves rapidly adsorb water vapor from the air and must be stored in sealed containers.
