Application of a vidicon tube as a multiwavelength ... - ACS Publications

shows the regular behavior in the LFE function of the aprotic solvents. Because of the high correlation coefficients between the E values forall of th...
0 downloads 0 Views 335KB Size
donor, CHC13. The fundamental work on Phenol blue by Figueras (5),as well as ours (6), has demonstrated that the solvent polarity contribution to the free energy transition (ET) of the dye can be derived from the two-parameter McRae equation:, and the influence of a given hydrogen bond donor (m-cresol or CHC13) conforms to a simple additivity as a perturbation energy contribution. However, in equilibrium 3, chloroform serves as an acceptor (HBA) toward phenol and shows the regular behavior in the LFE function of the aprotic solvents. Because of the high correlation coefficients between the ET values for all of the structurally related dyes in Table I, it appears that this whole series of bathochromic dyes may share a common phenomenological response to the Lewis basicity of the solvent. Finally, it should be clear that the equilibria in Equations 1 and 3 are nonequivalent chemical processes and should have dissimilar ET vs. AG, plots. The solvent displacement reaction with DMSO assigns a measurable Lewis base strength to CCld with reference to phenols as hydrogen bond donors. The available thermodynamic data for the two equilibria are not overlapping so that direct comparisons of solvent basicity between the two processes cannot be made at the present time. When a weaker Lewis base (i.e., pyridine) is substituted for DMSO in equilibrium 3, the two hydrogen bond donor sol-

vents, chloroform and 1,2-dichloroethane, compete more effectively with phenol to interact with the HBA; and the LFE relationship fails (7).Although the LFE function for reaction 3 includes only the weaker aprotic bases, similar correlations have not been observed between thermodynamic parameters related to solvent basicity and the blue shift scales proposed in the past as measures of solvent polarity in aprotic media. LITERATURE CITED (1) 0. W. Kolling, Anal. Chem., 48, 884 (1976). (2) M. Kamlet and R. W. Tafl, J. Am. Chem. SOC.,98,377 (1976). (3) J. Spencer, R. Harner, and C. Penturelli, J. fhys. Chem., 79, 2488 (1975). (4) E. Arnett, E. Mitchell, and T. Murty, J. Am. Chem. SOC.,96, 3875 (1974). (5) J. Figueras, J. Am. Chem. SOC.,93, 3255 (1971). (6) 0.Kolling and J. Goodnight, Anal. Chem., 45, 160 (1973). (7) J. N. Spencer et al., J. fhys. Chem., 80, 81 1 (1976).

Orland W. Kolling Chemistry Department Southwestern College Winfield, Kansas 67156

RECEIVEDfor review May 17, 1976. Accepted July 14, 1976.

Application of a Vidicon Tube as a Multiwavelength Detector for Liquid Chromatography Sir: In this report, we describe preliminary data for the application of a silicon target vidicon as a multiwavelength detector for liquid chromatography (LC). Most photometers currently used as uv detectors in LC are limited to one or two wavelengths. For these systems, the selectivity of any analysis is limited to that which can be obtained via the chromatographic process, and the analytical wavelength is seldom optimized fior all components to be detected. The vidicon detector, when used with appropriate optics, is capable of monitoring many wavelength resolution elements simultaneously ( I - 3 ) , and as such should serve as a very versatile detector for liquid chromatography ( 4 , 5 ) . The vidicon spectrometer system used in this work has been described in detail previously (1,Z). The liquid chromatograph was assembled from commercially available components and is similar to that described earlier (6) except that the flow cell from a stopped-flow mixing system (Aminco-Morrow Model B30-68109; American Instrument Co., Silver Spring, Md. 20910) modified by replacing the Teflon input chamber by a Teflon exhaust chamber, was used as the observation cell. No attempt was made to optimize the performance of the chromatographic system. Spectral data were collected, processed, and dispLayed by an on-line computer. An aqueous mixture of uric acid, theophylline, and phenobarbital was selected as an illustrative example (7). Uric acid is an important biological compound and phenobarbital is often included in theophylline preparations. These compounds were eluted from an anion exchange column with an ammonia buffer at pH 10. The performance of the column had been degraded by previous operation. Spectra in the range of 225 to 450 nm were recorded every 10 s. Figure 1represents the spectra after 220,330, and 350 s for uric acid, theophylline, and phenobarbital, respectively, added to and eluted from the column separately. Figure 2 represents the absorbance measured at 10-s intervals at the absorbance maximum for each of the respective components. The spectra show that there is

no wavelength within this range at which phenobarbital is free of interference from the other two components and, similarly, that there is no wavelength at which theophylline is free of interference from uric acid. The elution peaks show that uric acid is reasonably well separated from the other two components, but that the two drugs are poorly separated. These observations will be useful in interpreting data presented below. Figures 3A and 3B represent spectra at selected times during the elution of a mixture (0.83 pg each) of the compo-

b(nm> Figure 1.

Absorption spectra of uric acid, theophylline, and pheno-

barbital. All compounds were added to and eluted from the cation exchange column separately. Eluting reagent was 0.1 M ammonia buffer at pH 10. (a) Uric acid (5fig) at 220 s, (b)Theophylline (2 fig) at 330 s,and (c) Phenobarbital (2 fig) at 350 s

ANALYTICAL CHEMISTRY, VOL. 48, NO. 12, OCTOBER 1976

1815

A

t i m e (S) Figure 2. Absorbance vs. time for elution of pure samples of uric acid, theophylline, and phenobarbital ( a )Uric acid monitored at 290 nm, ( b )Theophylline monitored at 275 nm, and

(c)Phenobarbital monitored at 245 nm

nents. The spectra at 220 and 300 s in Figure 3A correspond closely to the spectra of uric acid and theophylline in Figure 1,suggesting that reasonably pure components are eluting a t these times (see Figure 2). Figure 3B represents spectra at times when theophylline and phenobarbital are eluting together. The peak near 275 nm results almost exclusively from theophylline (see Figure 1)while the absorbance at shorter wavelengths result from both drugs. Figures 4A and 4B represent absorbance vs. elution time at selected wavelengths. Figure 4A represents the absorbance at 290 nm (curve a ) and at 275 nm (curve b ) . Curve a is recorded a t the absorption maximum for uric acid and curve b is recorded at the absorption maximum of theophylline where phenobarbital does not interfere. Thus, uric acid could be determined with maximum sensitivity from the first peak in curve a and theophylline could be determined from the second peak in curve b without interference from phenobarbital even though the two drugs are not well separated. If either 275 or 290 nm were selected for both uric acid and theophylline, then small fluctuations in wavelength settings would have greater

effects on one of the components than is the case when absorption maxima are used. Figure 4B represents the elution curves recorded a t 245 nm where all components absorb. Uric acid could be quantified from the first peak, and some functions of the sum of the two drugs could be quantified from the second peak. Actually, since theophylline could be determined at each point from the peak at 275 nm in Figure 4A, the second peak in Figure 4B could be “corrected” for absorbance due to this component, leaving at least an estimate of the absorbance due to phenobarbital. These data were not treated in this fashion because the concept is relatively straightforward, but more importantly because we do not feel it would represent the best use of the data which are available from this type of experiment. It should be noted that the data displayed in Figures 3 and 4 were selected from a large volume available in the computer. In other words, during a 400-s run, we would have collected 40 complete spectra and these spectra would be more effective in differentiating quantitatively between the two drugs than would the simpler two-wavelengthapproach mentioned above. For example, we have used multiwavelength data and matrix algebra to resolve two-, three- and four-component mixtures of compounds whose spectra overlap significantly (8). We suggest that the most effective use of the data available from the experiments described above, would be to use such a multiwavelength approach to resolve the effluent quantitatively at different points in time and then to use numerical integration methods to quantify each component in the original sample. In fact, if one were attempting to resolve these components in an otherwise clean matrix, then there would be little advantage in attempting to separate them on the ion-exchange column used in this work. On the other hand, if they were in a complex matrix, then the separation step could be useful in isolating the drugs into a two-component mixture and the multiwavelength detector could alleviate the need for a complete separation. We are presently working on software which will permit us to combine the mutliwavelength approach with numerical integration methods for chromatographic applications. We have already suggested that one potential advantage of the rapid scanning detector for components which are well separated from others is the ability to monitor these components a t or near their absorption maxima. We believe there are other useful features of the multiwavelength data. Kissinger et al. (9) have suggested that liquid chromatography

Flgure 3. Absorption spectra during elution of a mixture of uric acid, theophylline, and phenobarbital(0.83 pg each) ( A ) Spectra at 220 ( a )and 300 ( b ) s and (6) Spectra at 330 (a),350 (b), and 380 ( c )s

1816

ANALYTICAL CHEMISTRY, VOL. 48, NO. 12, OCTOBER 1976

time(s) Figure 4. Eilution peaks

A

for a mixture of uric acid, theophylline, and phenobarbital

( A )Peaks monitored at 290 (a)and 275 ( b )nm and (6)Peaks monitored at 245 nm

with electrochemical detection (LCEC) may represent a good choice as a reference method for uric acid. We believe multiwavelength spectral detection could represent a complementary approach because the spectral data could be used to evaluate the purity of the aliquots which were being monitored in the effliuent. For example, if one selected some wavelength in the uric acid spectrum (e.g., 290 nm) as a reference, then the ratio of the absorbance a t any other wavelength t o that a t the reference wavelength (or the difference in log A ) should be a concentration independent constant for pure uric acid. Such ratios (or differences) evaluated a t multiple wavelengths should prove very useful in evaluating the efficiency of a separation process. Such procedures could be useful both in semiroutine applications and in more basic chromatographic studies. We do anticipate some limitations to this approach. The detection system, including the requirement of computing equipment to take full advantage of it, will certainly be more expensive than a single wavelength detector monitored by a strip-chart recorder. More important, there is little chance that a rapid scanning multiwavelength detector can be operated with the same reliability as the well regulated single wavelength detectors (IO,11). Accordingly, detection limits will most certainly be less impressive for multiwavelength detectors than with highly regulated single wavelength detectors. Accordingly, we view these as complementary approaches which will fill different needs for different workers.

LITERATURE CITED (1) M. J. Milano, H. L. Pardue, T. E. Cook, R. E. Santini, D. W. Margerum, and J. M. T. Raycheba, Anal. Chem., 46,374 (1974). (2) M. J. Miiano and H. L. Pardue, Anal. Chem., 47, 25 (1975). (3) H. L. Pardue, A. E. McDowell, D. M. Fast, and M. J. Milano, Clin. Chem. (Winston-Salem,N.C.), 21, 1192 (1975). (4) M. S. Denton, T. P. DeAngelis, A. M. Yacynych, W. R. Heineman, and T. W. Gilbert, Anal. Chem., 48, 20 (1976). (5) R. E. Dessy, W. G. Nunn, C. A. Titus, and W. R. Reynolds, J. Chromatogr. Sci., 14, 195 (1976). (6) P. T. Kisslnger, L. J. Felice, R. M. Riggin, L. A. Pachla, and D. C. Wenke, Clin. Chem. ( Winston-Salem, N.C.), 20, 992 (1974). (7) M. Weinberger and C. Chidsey, Clin. Chem. ( Winston-Salem, N.C.), 21, 834 (1975). (8) A. McDoweil and H. L. Pardue, Clin. Chem., ( Winston-Salem, N.C.), submitted for publication. (9) W. D. Slaunwhite, L. A. Pachia, D. C. Wenke. and P. T. Kissinger, Clin. Chem., ( Winston-Salem, N.C.), 21, 1427 (1975). (10) J. J. Kirkland, Anal. Chem., 40, 391 (1968). (11) T. E. Hewitt and H. L. Pardue, Clin. Chem. ( Winston-Salem, N.C.), 21, 249 (1975).

Alan McDowell Harry L. Pardue* Department of Chemistry Purdue University West Lafayette, Ind. 47907 RECEIVEDfor review May 17, 1976. Accepted July 2, 1976. This work was supported by USPHS Research Grant No. GM-13326-10 from the National Institutes of Health.

ANALYTICAL CHEMISTRY, )/OL. 48, NO. 12, OCTOBER 1976

1817