Anal. Chem. 1981, 53, 562-563
magnetic field,lie-shape analysis, spectra simulation. The signal processing system consisting of "andi" and the cbm-computer is applicable to all sorts of data processing and handling in laboratory application.
RESULTS Figure 2 shows the ESR spectrum of the naphthalene anion. The corresponding autocorrelation is represented in Figure 3. The absolute maxima of 0.183 and 0.490 mT correspond exactly to the coupling constants for the and cy protons. The relative maxima a t 0.307 and 0.674 mT (indicated by an asterisk) result from the sum and the difference of aHDl and aw. Because of its simultaneous function as an optimum fiiter (6), the autocorrelation procedure is also applicable to spectra with a low signal-to-noise ratio. Figure 4 shows the central part of the ESR spectrum of the N,N,I?',I?'-tetramethyl-p-phenylenediamine radical cation in absolute ethanol a t 22 "C (Wurster's blue perchlorate). The
values of the coupling constants (7) are indicated by arrows in the autocorrelation function (Figure 5).
ACKNOWLEDGMENT We are indebted to W. Jaenicke for a number of helpful discussions. We also thank F. Dickert for useful comments. LITERATURE CITED Allen, L. C. Nature (London) 1982, 198, 663-664. Swalen, J. D.; Gladney, H. M. IBMJ. Res. Dev. 1984, 7, 515-526. Zieglec, E.; Hoffmann, E. (3. 2. Anal. Chem. 1988, 240, 145-156. Dohrmann, J.; Rakowsky, T. Ber. Bunsenges. Phys. Chem. 1979, 83, 495-500. (5) Wyard, S. J. J . Scl. Instrum. 1985, 42, 769-770. (6) Plato, M.; Wlus, K. Messtechnlk (Braunschwelg)1972, 8, 224-234. (7) Wertz, J. E., Bolton, J. R. "Electron Spin Resonance"; Mc8aw-HIII: New York, 1972; Appendix D. (1) (2) (3) (4)
RECEIVED for review July 24, 1980. Accepted November 19, 1980.
Determination of Acetohydroxamic Acid in ,8-Mercaptoethanol H. P. S. Makkar, 0. P. Sharma, I?. K. Dawra, and S. S. Negi' Biochemistry Laboratory, Regional Research StaNon, Indian Veterhary Research Institute, Palampur, H.P. 176 06 1, India
The current work originated with the observation in our laboratory during studies on chemical interaction of acetohydroxamic acid (AHA) and urease (I) that ,%mercaptcethanol at high concentrations interferes in the Angeli-Rimini reaction (2) for AHA (AHA Fe3+ colored complex, A,, 520 nm) and that the estimation of AHA is not possible by the procedure of Fishbein (3). We report in this communication the removal of P-mercaptoethanol interference in AHA estimation by vacuum drying. The method is simple and does not require sophisticated instruments.
EXPERIMENTAL SECTION Chemicals. Acetohydroxamic acid was prepared by the method of Fishbein et al. (4). After three crystallizations from ethyl acetate solution, the white compound had a melting point of 87-89 O C , which is in agreement with the reported value (3, 4 ) . 0Mercaptoethanol was purchased from Sigma Chemical Co. All other reagents used were of analytical grade. Procedure. Aliquots (1.0 mL each) of AHA solution (135 pg/mL) with and without @-mercaptoethanolwere dried in test tubes under vacuum on a water bath kept at 75 2 "C. One milliliter of distilled water was added to each test tube, the test tube was shaken, and AHA was estimated colorimetrically (3) by the addition of 1 mL of 2% FeC13 (anhydrous) in 0.1 N HCl in a total volume of 3 mL. The absorbance was measured at 520 nm with a spectrophotometer, Type CS866B/C (Electronics Corporation of India Limited). RESULTS AND DISCUSSION Effect of Different Concentrations of 8-Mercaptoethanol on AHA Estimation. AHA solution containing different concentrations of P-mercaptoethanol was prepared, and the color was developed by the above mentioned method for the estimation of AHA. @-Mercaptoethanola t a concentration of 0.25 and 0.5% produced 7 and 17% interference in the color complex formation. Whereas 1% or more of P-mercaptoethanol produced 100% interference. Effect of Vacuum Drying the AHA Solution in Boiling Water Bath. One milliliter of AHA solution without @mercaptoethanol was vacuum dried in a boiling water bath and then AHA was estimated by the standard procedure. It 0003-2700/81/0353-0562$0 1 .OO/O
was observed that the absorbance of vacuum-dried AHA was less than that of the one which was not dried. I t was found that the absorbance decreased as the samples were kept for a longer time in the boiling water bath after the samples had been dried. A 22, 34, and 44% decrease in absorbance was observed when the samples were kept for 2, 4, and 6 min, respectively, after the solution was just dried. The decreased absorbance could be due to the decomposition of AHA when it was dried on a boiling water bath under vacuum. Effect of Vacuum Drying the AHA Solution at 75 f 2 "C. When AHA was vacuum dried at 75 f 2 "C, no decrease in absorbance was found even after 6 min of ita being dried, suggesting that vacuum drying a t 75 f 2 "C is quite safe. Effect of Vacuum Drying the AHA Solution Containing @-Mercaptoethanolat 75 f 2 OC. The interference in samples containing 0.25 and 0.5% 0-mercaptoethanol can be eliminated when these are dried under vacuum a t 75 f 2 "C to 0.5 and 0.25 mL, respectively. At high concentrations of P-mercaptoethanol(1,2, and 5%) vacuum drying to even 0.25 mL is not sufficient for the complete removal of interference. However, complete drying under vacuum removed the interference caused by P-mercaptoethanol not only at low concentrations but also a t high concentrations. The time taken for completely drying 1 mL of the solution was 8-13 min. However, the time can be further reduced by taking smaller volumes (0.25 mL or 0.5 mL) if the AHA content of the sample is high. The precaution which needs to be taken is that the sample containing AHA should be completely dry before following the standard procedure for its estimation. Incomplete drying of the samples can give rise to large error especially a t high concentrations of 0-mercaptoethanol. The present results show that P-mercaptoethanol interferes in AHA estimation. The vacuum drying of the sample containing P-mercaptoethanol on a boiling water bath is not recommended for the removal of interference. However, complete drying under vacuum a t 75 f 2 "C eliminates the interference caused by 8-mercaptoethanol in the estimation of AHA. 0 1981 American Chemical Society
Anal. Chem. 1981, 53, 563-584
9ev. 1943, 33, 209-256. nter, T. S.;Davidson, J. D. J . Bioi. Chem: 1965,
LITERATURE CITED (1) Makkar, H. P. S.;Sharma, 0.P.; Dawra, R. K.; Negi, Sci., in press.
s.s. J . ~ a / w
RECEIVED for review July
29, 1980. Accepted November 14,
Prediction of Optimum Solvent Composition for Thin-Layer and Liquid Chromatography David Nurok” and Michael J. Richard Department of Chemistry, Indiana University-Purdue University at Indlanapolls, Indianapolis, Indiana 46205
In recent years there has been a renaissance in thin-layer chromatography (TLC) with the introduction of high-performance plates, sophisticated spectrophotometric TLC scanners, and novel spotting techniques. An important problem that remains to be solved both in high-performance liquid chromatography and in TLC is the development of a logical method for optimizing the solvent composition of the liquid phase to give a predictable separation between compounds. It is well-known that a binary mixture of two solvents with different polarity (E) values is often satisfactory for separation where the pure solvents are unsuitable. Snyder and Kirkland (1) have described methods for modifying solvent composition to improve separation, but some trial and error is necessary in their method. Soczewinski and co-workers (2-5) have shown that a plot of RM vs. log X , results in a family of skraight lines for TLC on silica gel with a given binary mixture of a polar and a nonpolar solvent. RM is the natural log of the capacity factor ( k )and X,is the mole fraction of the polar component of the binary mixture. Perry (6)has reported BL similar linearity when RM is plotted against the values of binary mixtures of carbon tetrachloride and chloroform. We have used the results of Perry and Soczewinski together with a simple equation that we have developed to investigate a method for determining optimum solvent composition for a given separation. The method is described below. DISCUSSION The relationship between retardation factor Rf and capacity factor k is 1 R, (1)
l + k
Consider two solutes 1 and 2 in a TLC system such that Rnl > Rf,2. Then ARf = Rf,, - R,2 1 - 1 1 + k , 1 + k2
The actual separation in millimeters between the spot centers of two solutes on a TLC plate is S D = ARfL (3) where SD = center-to-center distance between two spots measured in millimeters and L = path length in millimeters between the spot origin and the solvent front. Clearly eq 2 and 3 can be used to calculate spot separation a t a given solvent composition provided values of kl and kz are known. These values have been calculated for a few of the many systems for which results have been published by Soczewinski and co-workers (2-5) that show a linear relationship between
In k and In X,.These results on silica gel were used to calculate regression lines for three compounds. Values of k were then used to calculate ARf values for pairs of compounds at different mole fractions of an acetone/cyclohexane mixture. These results are shown in Figure 1 for three pairs of compounds together with the experimentally obtained separations. There is good agreement between the experimental results and the theoretical curve. It is clear from the diagram that optimum solvent composition is at an acetone mole fraction of 0.40 with (ARJmaXvarying from 0.08 to 0.30 for the three pairs of compounds. If a 10-cm path length is used on a silica gel plate with this solvent composition, the separation of the closest pair of spots (o-cresol/phenol) is calculated from eq 3 as 8 mm. The mole fraction range for a given ARf can also be obtained from the curve. When the mole fraction is extrapolated to hypothetical values of greater than 1 (i.e., more than 100% acetone) a second maximum is found for two of the pairs, implying that a (ARf)max with the inverse spot order may occur a t solvent strength greater than that of pure acetone. In this particular case the (ARf), for the hypothetical mole fraction is less than that for the true mole fraction. The beauty of this approach is that optimum solvent strength can in principle be estimated for any mixture of solutes by performing TLC at two mole fractions with a binary solvent mixture provided that there is a linear relationship between In k and In X,.This is the most common case (7). Care should however be taken as there are situations when the relationship is nonlinear. In any event it will always be possible to choose an optimum solvent strength for a given separation provided there is a predictable relationship between k and X,. In Figure 1all three solvents pairs have an optimum separation at the same solvent strength. It is expected that there will be many instances where maximum ARf for each pair occurs at a different solvent strength. In such a case the optimum solvent strength would be chosen for the pair of compounds with lowest (ARf)max. For this pair of compounds the required path length can be estimated by rearranging eq 3.
L = S,/ARf
In Figure 1 the lowest (ARf),, is 0.08. About 5 mm is an average value of SDfor complete spot separation on highperformance TLC plates. Thus
ARf can also be optimized for the solvent polarity parameter ( e ) . Perry (6) has shown that for certain dyes there is a linear relationship between RM and the e value of a binary solvent system. Figure 2 was constructed for two pairs of dyes using
0003-.2700/81/0353-0563$01.00/0 0 1981 American Chemical Society