tween various sources of crude oils, far more complex pyrograms must be developed. This can be accomplished by performing pyrolysis at lower temperatures and increasing GC separation of produced derivative composition. The practical use of this method for oil spill surveillance and source identification is further compounded by the fact that the composition of spilled oil is unstable in aquatic environment. This instability is characterized by the loss of volatile and water soluble fractions, leading to eventual formation of so-called “tar balls.” The residual, environmentally stable compositions of oil spills such as found in “tar balls,” could provide a key
for oil source identification. If a sample and a standard are pretreated to remove volatile and water soluble fractions, the comparison between the patterns produced by the residual matter in a sample and a standard should be possible.
RECEIVED for review May 30, 1972. Accepted July 27, 1972. The research upon which this paper is based was performed pursuant to Contract No. 14-12-802 with the Environmental Protection Agency, Water Quality Office. H. P. Nicholson served as Project Officer.
Influence of Solvent upon the Phosphorescence Characteristics of Several Sulfonamides at 77 O K D. R. Venning,’ J. J. Mousa, R. J. Lukasiewicz,2 and J. D. Winefordner3 Unicersitp of Florida, Department of’Chemistry, Gainesville, Fla. 32601
PHOSPHORIMETRY has been shown to be a n extremely sensitive analytical method of considerable use for the trace analysis of molecules, especially when coupled with separation methods. However, until the recent work of Lukasiewicz, Rozynes, Sanders, and Winefordner ( I ) and of Lukasiewicz, Mousa, and Winefordner (2, 3) phosphorimetry was limited to nonaqueous solvents and solvent mixtures which would form clear, rigid glasses a t liquid nitrogen temperature, 77°K. However, the latter group of authors (2, 3) introduced the quartz capillary cell for low temperature studies; the quartz capillary cell allowed measurement of phosphorescence of molecules in aqueous or predominantly aqueous solvents with no danger of cracking the cell. In addition, capillary cells are much simpler to clean, and so contamination is not anywhere near the problem as with previous commercial closed cells. Furthermore, the cell is much simpler to fill and empty, and so the time of analysis is considerably reduced. Because phosphorimetry has been limited to the measure5 ment of sample solutions in rather small diameter ( ~ mm cm) quartz cells, considerable positioning i d . ) , long ( ~ 2 0 errors result with standard commercial holders. However, Hollifield and Winefordner (4, Zweidinger and Winefordner (5) and later Lukasiewicz, Mousa, and Winefordner (3) have developed a means of reducing random positioning errors and of allowing tneasurement of snowed and highly cracked sample matrices, as well as clear, rigid glasses by simply rotating the sample cell. The precision of measurement with the rotating phosphorimetric cell is of the same order as in fluorimetry of samples a t room temperature. Present address, 1101 Coral Way, Coral Gables, Fla. 33134. Present address, P.O. BOY76. Brea, Calif. 92621. Author to whom all correspondence should be sent. (1) R. J. Lukasiewicz, P. Rozynes. L. B. Sanders, and J. D. Winefordner, ANAL.CHEW.. 44,237 (1972). ( 2 ) R. J. Lukasiewicz, J. J. Mousa, and J. D. Winefordner, ibid.. p 1339. ( 3 ) Ibid., p 963. (4) H. C . Hollifield and J. D. Winefordner, ibid., 40, 1759 (1968). ( 5 ) R. A. Zweidinger and J . D. Winefordner, ihid., 42, 639 (1970).
Because of the recent work by Lukasiewicz, Mousa, and ‘Winefordner ( 2 ) o n the influence of mixed solvents (alcoholwater and alkali halide-water) on the phosphorescence signals of several organic molecules, the present authors decided to investigate the phosphorescence characteristics (phosphorescence excitation and emission spectra, phosphorescence lifetimes, phosphorescence analytical curves, and phosphoIescence limits of detection) of sulfanilamide, sulfathiazole, sulfamethazine, sulfadiazine, sulfaguanidine, and sulfacetamide in four different solvents: 10% sodium iodide in water; 10% sodium chloride in water; 1 0 % methanol-water; and 10% methanol with 1 % sodium iodide in water. Previous work had shown that these solvents should result in maximal phosphorescence signals--i.e., plateau of plots of phosphorescence signal cs. solvent composition. The sodium iodide was of particular interest because of the combined matrix and heavy atom effects (3). The results of this study are compared with the results obtained by Hollifield and Winefordner (6) using a clear, rigid solvent of 100 % ethanol. Sulfonamides were chosen for this study because they have been used as chemotherapeutic agents and are still being used as bacteriostatic agents. EXPERIMENTAL
Apparatus. All phosphorimetric measurements were made with a n Aminco-Bowman spectrophotofluorometer a i t h a n Aminco-Keirs phosphoroscope attachment, a 150-W xenon arc lamp, a potted R C A IP28 photomultiplier tube, and a n X-Y recorder (American Instrument Company, Silver Spring, Md.). The basic instrument was arranged as previously described (2, 6 ) with the following exceptions. A Keithley Model 244 high voltage supply (Keithley Instruments, Cleveland, Ohio) was used to provide power to the photomultiplier tube. A rotating sample cell apparatus ( 4 ) consisting of a Varian A60-A High Resolution Nuclear Magnetic Resonance Spectrometer Spinner Assembly (Varian Associates, Palo Alto, Calif.) was modified for use with a (6) H. C. Hollifield and J. D. Winefordner, A i d . Cl7im. Actn, 36,
352 (1966).
ANALYTICAL (CHEMISTRY, VOL. 44, NO. 14, DECEMBER 1972
2387
Table I. PhosphorescenceParameters of Quartz Tube and Solvent Backgroundsa Excitation Emission Phosphorescence Standard deviation wavelength, nmb wavelength, nmb background signal, in A in background signal, in A 291 504 8 . 0 X IO-" 0.18 X 1O-I' 293 473 2.9 x 0.70 x 292 473 1 . 2 x 10-10 0.43 x 10-'0 290 473 1.3 X 0.07 x 10-1"
Background Quartz tube 10% Sodium iodide 10% Sodium chloride 10% Methanol 10% Methanol1 Sodium iodide 295 473 1 . 8 X 10-lo 0.13 X 10-l0 Solvent backgrounds included quartz capillary tube. The above wavelengths are uncorrected for instrumental characteristics and are at the mean peak intensities. The actual spectra are quite broad compared to the sulfonamide spectra and thus make it hard to pinpoint the actual maximum wavelength. Reproducibility of setting the wavelengths 2 2 nm. Q
Table 11. Phosphorescence Characteristicsof Several Sulfonamides in Several Solvents Emission Excitation Emission Limit of Limit of detection, wavelength, wavelength, lifetime. detection Sulfonamide Solvent mg/mlC nm4 nmn secb ngc Sulfanilamide 10% NaI 0.6 0.06 275 410 3.0 X lo0 10% NaCl 0.06 1.5 274 411 3.0 X 10" 276 409 2.0 1 . 0 x 100 10% MeOH 0.02 10% MeOH/l% NaI 410 2.0 x 100 0.04 278 1 .o 100% EtOHd 270 405 1.3 0.5 1.0 x 1 0 0 Sulfathiazole 10% NaI 1 . 0 x 102 426 2.0 278 1.3 10% NaCl 4 . 0 X 101 426 277 1.5 0.8 10% MeOH 2.1 7 . 0 x 10' 2.0 416 280 10% M e O H / l z NaI 5 . 0 X 10' 1.6 1 .o 412 277 1007, EtOHd 0.9 420 1 . 0 x 10: 5.0 310 102 N a I Sulfamethazine 6 . 0 x 10-1 0.01 410 1.1 278 10% NdCl 0.01 6 . 0 x 10-1 410 1.3 275 6.0 x IO-' 10% MeOH 410 0.01 2.1 286 6.0 x 10-1 0.01 10% MeOH/1 NaI 285 411 1.5 6 . 0 x 10-1 0.3 0.8 410 100% EtOHd 280 Sulfadiazine 0.06 10% NaI 0.5 3 . 0 X IOo 271 410 1.5 6 . 0 x 100 10% NaCl 0.12 27 3 410 0.04 2.0 x IO0 10% MeOH 1.6 410 277 10% MeOH/l NaI 1 . 0 x 100 0.02 277 410 0.9 100% EtOHd 1.0 x 10" 0.5 275 410 0.7 Sulfacetamide 8 . 0 x 10-1 0.02 272 41 3 10% NaI 1.1 7.0 x lo-' 0.01 269 10% NaCl 1.6 41 1 2 . 0 x 10-1 0,004 10% MeOH 2.0 289 411 2 . 0 x lo-' 0.004 1.6 10% MeOH/l% NaI 287 411 0.05 1 . 0 x 10-1 410 1.3 100% EtOHd 280 Sulfaguanidine 4 . 0 X loo 0.08 270 410 0.6 10% NaI 4.0 X IO" 0.08 1.7 10% NaCl 267 408 0.08 2.0 4 . 0 x 100 273 410 10% MeOH 4 . 0 X 100 0.08 274 1.1 411 10% MeOH/l% NaI 5.0 0.7 1.0 x IO' 405 305 100% EtOHd a All wavelengths are uncorrected for instrumental characteristics. Reproducibility of setting wavelengths and of grating drives is about 1 5 nm. Lifetime taken as time for phosphorescence signal to drop to lie of initial value after termination of exciting light. Limit of detection was taken as concentration resulting is a signal of twice the standard deviation of the background readings. The absolute limit of detection (in ng) was taken as the concentrational value (ng X ml) multiplied by volume of sample (cell used for all aqueous samples held 20 p1 of sample and cell for EtOH held 500 PI.) 100% EtOH solutions were previously measured by Hollifield and Winefordner (6).
quart7 capillary sample tube. The sample tubes were made from T21 Suprasil capillary tubing (Amersil Inc., Hillside, N.J.) having 5-nim outer diameter and 0.90-mm internal diameter. Reagents. Standard stock solutions of the following sulfonamides were prepared with 5-6 mg of the compound per 50 ml of deionized water: sulfanilamide, sulfathiazole, sulfamethazine, sulfadiazine, sulfaguanidine, and sulfacetamide (all from Nutritional Biochemical Corporation, Cleveland, Ohio). Solvents were prepared using water, sodium chloride, sodium iodide (Fisher Scientific Co., Fair Lawn, N.J.), and spectroquality methanol (Matheson, Coleman and Bell, East Rutherford, N.J.). Reagents were used without further purification. Deionized water was obtained directly from a commercial ion exchange column. 2388
Procedure. The phosphorescence spectra for the six molecules were obtained in four aqueous solvents using t h e quartz capillary sample tube spinning and suhpended in liquid nitrogen in a quartz window Dewar flask. The slit arrangement used for all studies was 3, 4, 3, 3. 4.3 (in mm, corresponding to 17 nm and 22 nm spectral half-band pass). The spectra were not corrected for instrumental response. The stock solutions of approximately 10-i.44 of each sulfonamide in deionized water were diluted ten-fold in solvents of 10% sodium iodide, 10% sodium chloride, 10% methanol, and 10% methanol plus 1 % sodium iodide added, where all spectra were measured. From the phosphorescence spectra, the peak excitation and emission wavelengths were recorded. The mean emission lifetime, which i b the length of time required for the phosphorescence intensity to decrease to 1 : ~
ANALYTICAL CHEMISTRY, VOL. 44, NO. 14, DECEMBER 1972
of its initial value, was also recorded a n d determined for each of the above solutions. Serial dilutions of (1 :lo) of the above solutions with respective solvents were made until the phosphorescence signal of the sulfonamide solution was the same as the background signal. The phosphorescence background is taken t o be the solvent signal with n o drug added. (All results were taken as the average of three o r more dilrerent measurements.) Sample handling procedures were followed as previously described (7). RESULTS AND DISCUSSION
In Table 1, the phosphorescence excitation and emission mean peak wavelengths and relative phosphorescence signals are listed for the quartz capillary tube cell and the various solvent backgrounds (at 77 ”K). In Table 11, the phosphorescence characteristics (phosphorescence excitation and emission wavelengths uncorrected for instrumental response, phosphorescence emission lifetimes, and concentrational (ng’ml) and a bsolute (ng) phosphorimetric limits of detection) a r e listed tor sulfanilamide, sulfathiazole, sulfamethazine, sulfadiazine. sulfaguanidine, and sulfacetamide in the four frozen (at 77 ‘K) aqueous solutions mentioned previously (10% sodium iodide, 1 0 % sodium chloride, 10% methanol, and 10z methanol-1 % sodium iodide, water being the primary solvent in all cases) as well as results previously obtained by Hollifield and Winefordner (6) for the same sulfonamides in 100 ethanolic solvent. The sulfonamides investigated showed good phosphorescence intensities except for sulfathiazole. Although there is n o correlation a t present between phosphorescence intensity and molecular structure of the compounds studied, it should be noted that sulfathiazole is the only one of the six sulfonamides studied which has a sulfur atom located within a ring structure. The concentrational limits of detection obtained in the four new solvents showed comparable results with those obtained in 100 ethanol. Exceptions were sulfathiazole which showed a decreased (better) concentrational detection limit of ten to twenty times depending on the solvent used a n d sulfamethazine which showed a n increased (poorer) concentrational detection limit of five to fifty times as compared to the concentrational detection limits for the same species (7) J. D. Winefordner, P. A. St. John, and W. J. McCarthy, Chapter on “Phosphorirnetry” in “Fluorescence Assay in Biology and Medicine,” S. Udenfriend, Ed., Academic Press, New York,
N.Y., 1Y70.
in 100% ethanolic solvents. However, it should be kept in mind that the sample volume of sulfonamides in 100% ethanol was about 500-pl sample cell, whereas the sample volume of sulfonamides in the aqueous solvents was about 20 ,d. Although there has not been any significant gain in the concentrational limits of detection of any of the sulfonamides (except for sulfathiazole), the absolute limits of detection (in ng) are generally much less because ten to one hundred times less sample can be used to obtain the same phosphorescence signal. All of the sulfonamides in the various aqueous plus ethanolic solvents had phosphorescence lifetimes in the narrow lifetime range from 0.5 t o 2.1 sec. The sulfonamides in 10% methanol-water generally exhibited the longest phosphorescence lifetimes, a n d the sulfonamides in 10 iodide-water exhibited the shortest lifetimes, as expected based upon the influence of the heavy atom solvent. The phosphorescence lifetimes of the sulfonamides in 10 sodium chloride-water and 10% methanol--1 % sodium iodide-water lifetimes fell somewhere between the lifetimes of sulfonamides in the other two solvents. These observations were consistent with the heavy atom perturbation theory, but none of the sulfonamides showed a n increase in phosphorescence intensity due to the heavy atom erect from the salt solution, i.e., only the matrix effect discussed by Lukasiewicz, Mousa, and Winefordner (3) was present. It is thought that there is possibly a weak complex formed between the sulfonamides and the iodide or chloride which accounts for the same or even a decreased signal (3). The four aqueous solvent combinations used in this study have been shown to produce comparable results to the nonaqueous solvent (e.g., ethanolic solvent) used in the past. Of the four new aqueous solvents studied, the 10 methanol in water seemed to be the best all around solvent as it was the easiest to prepare, usually produced the highest phosphorescence signal, and has one of the lowest background signals. However, the 10% sodium iodide and 10% sodium chloride, although not showing the heavy atom increase in signal, would yield comparable results if an aqueous solvent resembling that of human blood, for example, was desired. RECEIVED for review June 7 , 1972. Accepted August 9, 1972. Research carried o u t as a part of a study o n the phosphorescence characteristics of drugs in blood and urine, supported by U.S. Public Health Service G r a n t (GIM-11373-09).
ANALYTICAL CHEMISTRY, VOL. 44, NO. 14, DECEMBER 1972
2389
