722
Anal. Chem. 1988, 60.722-724
considerable promise in model and diagnostic studies of redox reactions of biological compounds. Factors controlling thermospray ionization as well as the screening potential of this method are being further explored. Registry No. Uric acid, 69-93-2; 6-thioxanthine, 2002-59-7; imine alcohol, 6960-30-1; allantoin, 97-59-6; alloxan, 50-71-5; 5-hydroxyhydantoin-5-carboxamide, 36597-25-8;bicyclic imidazolone, 112459-94-6;5-hydroxyhydantoin,29410-13-7;xanthine, 69-89-6; 2-hydroxypurine, 2308-57-8.
/
Diol
LITERATURE CITED
Bicyclic lmidazolone M W 140
i:
0
“0:
0“
Alloxan MW 142
H 5-Hydroxyhydantoin5-carboxamide M W 159
OH
JJ, H A l l o x a n Monohydrate M W 160
H
5-Hydroxyhydantoin MW 116
Allantoin MW 158
(1) Zimm, S.; Johnson, G. E.; Chabner, B. A.; Poplack, D. G. Cancer Res. 1985, 45, 4156. (2) Langen, P. Antimetabolites of Nucleic Acid Metabolism; Gordon and Breach: New York, 1975. (3) Astwocd. D.; Lippincott, T.; Deysher, M.; D’Amico, C.; Szurley, E.; Brajter-Toth, A. J . Elecfroanal. Chem. 1983, 259, 159. (4) Kraske, P. J.; Brajter-Toth, A. J . Electroanal. Chem. 1986, 207, 101. (5) McKenna, K.; Brajter-Toth, A. J . Electroanal. Chem. 1987, 233, 49. (6) Miner, D. J.; Rice, J. R.; Riggin, R . M.; Kissinger, P. T. Anal. Chem. 1981, 5 3 , 2258. (7) Rice, J. R.; Kissinger. P. T. Biochem. Siophys. Res. Commun. 1982, 104, 1312. (8)Andrews, P. A.; Pan, Su-shu; Bachur, N. R. J . Am. Chem. SOC. 1988, 108, 4158. (9) Dryhurst, G.; Kadish, K. M.; Scheller, F.; Renneberg, R. Siological Electrochemistry; Academic: New York, 1982. (10) Brajter-Toth. A.; Dryhurst, G. J . Electroanal. Chem. 1981, f22, 205, and ref 5(d) therein. (11) Hambtzer, G.; Heitbaum, J. Anal. Chem. 1986, 58. 1067. (12) Rudewicz, P.; Straub, K. M. Anal. Chem. 1986, 5 8 , 2928. (13) Perchalski, R . J.; Yost, R. A,; Wilder, E. J. Anal. Chem. 1982, 5 4 , 1466. (14) Doerr, I . L.; Wempen, I.; Clarke, D. A.; Fox, J. J. J . Org. Chem. 1971, 2 6 , 3401.
Figure 1. Electrochemical oxidation of uric acid. Molecular weights are shown for intermediate and products observed by on-line EC/
Kevin J. Volk Mike S. Lee Richard A. Yost* Anna Brajter-Toth*
TSPIMSIMS.
at these potentials (5,141. It has been previously inferred from the identified final products (5) that the oxidation of 6thioxanthine at potentials more positive than +0.7 V vs SCE must lead to the oxidation of the purine ring. This is now confirmed (Table I) by the identification of an imine alcohol intermediate (MW 184). The same intermediate forms in the oxidation of uric acid, clearly indicating oxidation of the purine ring (Figure 1). As expected, exhaustive oxidation of 6thioxanthine a t potentials >+0.7 V is accompanied by the disappearance of the ions characteristic of 6-thioxanthine. The structure of the compound responsible for the negative ion at mlz 142 is a t present uncertain. As demonstrated here, the on-line combination of electrochemistry with thermospray/tandem mass spectrometry shows
Department of Chemistry University of Florida Gainesville, Florida 32611
RECEIVED for review July 17,1987. Accepted November 24, 1987. We acknowledge ESA, Inc., for the loan of the electrochemical cell. This work was supported, in part, by grants from the U.S. Army Chemical Research, Development, and Engineering Center (R.A.Y., No. DAAA15-85-C-0034),Research Corporation (A.B.T.),NIH (A.B.T., through Grant GM 35341-01A2),and the Division of Sponsored Research at the University of Florida (A.B.T.).
Indirect Fluorometric Detection of Anions in Thin-Layer Chromatography Sir: Thin-layer chromatography (TLC) is a broadly applicable separation technique. With the recent growth of interest and products for high-performance liquid chromatography (HPLC), TLC has not received the attention it deserves. A major factor is that the approach is still not highly “instrumental” compared to HPLC, both in the separation step and the detection step, even though much progress has been made ( 1 ) . TLC provides many advantages as an analytical technique. It is easily adapted for two-dimensional separation (Z),for “whole-column” detection ( 3 ) , and for handling multiple samples. As a complementary technique to HPLC, one can use TLC as an initial screening step to
optimize conditions. This is because of the ease of equilibration with a new mobile phase, the speed of development, the freedom from contamination and carry over, the applicability to highly retained solutes, and the low cost of stationary and mobile phases used. Detection in TLC lags behind the corresponding technology in HPLC. The two-dimensional geometry of the spots limits the path length available for absorption photometry, although nanogram levels are readily detectable. The sensitivity problem can be partially solved by using photothermal spectroscopy ( 4 ) or photoacoustic spectroscopy ( 5 ) . Nonlinearity still exists because of the scattering nature of the
0003-2700/88/0360-0722$01.50/00 1988 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 60, NO. 7. APRIL 1. 1988
substrate (6),and the Kubelka-Munk function is needed for correction. If the analyte fluoresces, laser fluorometry has been shown to allow detection of the order of 5 x lo-" mol per spot (7).AU of these require derivatization for the majority of analytes because of the lack of absorption, either before or after development. Besides being an undesirable complication, unreliable chemistry and irreproducible operation can drastically affect quantitation. Recently, several indirect detection methods for LC have been demonstrated (8-10). Briefly, the detector responds to some physical property of the chromatographic eluent. So, there is a constant background signal generated at the detector when no analytes are present. When the analyte elutes, it displaces an equal amount of the eluent a t the detector. Even though the detector does not respond to the analyte, the lower eluent concentration at the detector causes a decrease in signal. The analyte can then be monitored as a negative signal, i.e., indirectly. So, it should he possible to devise a detection method for TLC based on indirect fluorometry. It may then be poesible to extend the advantages for fluorescencedetection to species that do not themselves absorb or fluoresce. There is a technique for TLC quantitation known as fluorescence quenching (11). In fact, commercial TLC plates are available with a fluorophor already incorporated for this purpose. I t is really a misnomer because rarely does the analyte actually change the fluorescence quantum yield of the fluorophor. In reality that is only a modified absorption technique. The analyte absorbs the excitation light (or the emitted light), and depletes the amount of fluorescence. The advantage is the elimination of UV optica in the observation of emission, so that direct visualization is possible for UV chromophors. Sensitivity is a t best comparable to U V transmimion or UV reflectance spearometry. In this article, we describe the detection of nonfluorescing and nonabsorbing anions. Displacement of a fluorescing (background) eluting ion by the analyte ion due to electroneutrality gives rise to the response.
EXPERIMENTAL SECTION The setup for TLC is conventional, with development in the vertical direction. Samples are applied by a microsyringe calibrated in 0.1-pL units. Cetyltrimethylammonium bromide (cetrimide) was obtained from Aldrich (Milwaukee,WI). Samples and various solutions are prepared from reagent grade chemicals (Fisher, Fair Lawn,NJ),the former being the sodium salts of the respective anions. TLC plates used are either polyester-backed [(diethylamino)ethyl]-tert-aminecellulase (Anspec, Warrenville, IL, Whatman 100-pmCEL3OO DEAE) or glass-hacked silica gel (Alltech. Deerfield. IL, Adsorbosil-Plus l ) ,cut to 5-em widths. Pretreatment of TLC plates is accomplished by immersing the plates for predetermined time intervals in solutions in a beaker, rinsing with deionized water, and drying with a heat gun. Visualivltion is aided by a W lamp (Cole Palmer, Chicago, IL,9815, 312 nm). Photographs are unretouched and are taken by a 35mm camera using Plus-X film (Kcdak,Rochester, NY)and a macro lens with a 1:4 or lower magnification.so that the plate fills moat of the frame. The same UV lamp is used for excitation.
RESULTS AND DISCUSSION DEAE cellulose plates are natural anion exchangers a t neutral pH. They can be used without modification in this scheme. Analogous to equilibration of the ion-exchange column in nonsuppressed ion chromatography (IC) (12),the commercially available plates are first immersed in a solution of 5 x M sodium salicylate. This is to avoid depletion of the concentration of salicylate in the developing solution. The p r d u r e also creates a uniform fluorescence background on the plate. After drying, the plate can be spotted with the sample solutions and developed as usual. The developing solution used contains 1.5 X M sodium salicylate and 25% ethanol. The latter was included to avoid any hydrophobic
723
Fl@m1. Indirect flwromeMc detection of anbns separated on a DEAE d i u l o ~ e plate. ~magesize is 4.5 x 5.5 cm. s he four samples are spotted about 0.7 cm apart. A bight R w e s m c e spot corresponding to the system peak (salicylate) is visible in each lam, at the sobent front. The sample atthe mntalns 15 n d of NO;, 15 n m l of I-, and 7.5 n m l of S2-, showing up as dark spots from top to bonom. The other three samples contain the same ions at concentrations ' I 2 . 'I,. and ' I , $of that at Me right.
nc
interactions between the plate and the larger anion, salicylate. After developing for 8 min, the fluorescence image can be viewed under a UV lamp after drying. The result is shown in Figure 1. Clearly. the spots correspondingto each ion are visible even a t the lowest concentration, despite an inhomogeneous background cauwd by the rough surface of the plate. It must be emphasized that the photographs shown here have essentially the same contrast as the images observed visually in a normally lit laboratory, i.e., the contrast has not been enhanced photographically. These anions do not absorb significantly a t either the absorption or the emission wavelengths of salicylate. The nature of the response is one of replacement. Local charge neutrality (uniform cation distribution) and competition for available anion exchange sites cause a depletion of salicylate where the analyte ions reside, producing a lower fluorescence signal. A t the solvent front in Figure 1, one can observe bright spots corresponding to the 'system peak". The associated cations of the analyte are not retained. and they sweep along an equivalent amount of salicylate ions to where they reside. The detection of the order of 1C-100 ng of an anion in TLC is not particularly impressive, even though Figure 1 demonstrates the principle of indirect fluorescence detection. The optimization of detectability can he achieved following the corresponding scheme in IC (13.14). In the indirect detection method, the dynamic reserve (which is defined as the ratio between the background signal and ita noise level), the concentration of the visualization reagent, and the displacement ratio (which is defined in this work as the number of visualization ions which are transferred by one analyte ion) all play important roles in the sensitivity that can be achieved. The concentration detectability (Ch)a t the detector is given by these parameters as follows: C,
= CJRD
(1)
where C. is the concentration of the visualization reagent, R is the displacement ratio, and D is the dynamic reserve. The daplacement ratio for monovalent ions is expected to be unity in ion-exchange chromatography (independent of the mobile phase concentration). So, one should decrease the concentration of the eluting ion, C,. Ion-exchange stationary phases
724
ANALYTICAL CHEMISTRY. VOL. 60. NO. 7. APRIL 1. 1988
npure 2. lndkecl fluoromefic detenion of a n h s separated on a dynamicaity (cetrimlde) mod^ silica gel n c piate. ~magesize is 4.5 X 5.5 an. The f a r samples are sponed abart 0.7 cm apart. Tht sampb at the right wntains 1 0 nnwi of IO;. 5 nnwi of NO.; and 5 nmol of I'. showing up as dark spots from top lo bottom. The 0th thee samplescontah the arm b s at cowenbatbns of '1,. 'Ila. and '1% of that a1 Vm right. having very low capacities are required to demonstrate the separation of ions by use of a mobile phase with a low concentration of visualization reagent. The cellulose plates cannot be modified to have a lower capacity, so eluent concentrations below lo-* M salicylate cannot be used for separating these ions. We therefore dynamically modify silica gel plates with cetrimide to provide anion exchange sites, similar to a procedure reported for IC (14). From those experiments i t may be concluded that cetrimide forms a double layer on the silica surface. The first layer of the reagent is formed by electrostatic interaction between nitrogen with a positive charge and a silanol group. while the secondary layer is formed by the hydrophobic interaction between the cetyl groups of the first layer and those of the reagent in the treating solution. The electrostatic interaction between the silica surface and the first layer is much stronger than the hydrophobic interaction between the first and second layer. Once treated, the cetrimide stays on the plate even though the subsequent solutions used do not contain cetrimide. By adjustment of the organic solvent fraction, the pH, and the time of interaction, the amount of cetrimide (thus the capacity) on the TLC plate can he controlled. Figure 2 shows the separation of anions on a dynamically modified TLC plate visualized by indirect fluorescenee. The commercial TLC plate was treated with 10 mM cetrimide in water (pH 7.38 with 0.03 M phosphate) for 2 h. After being rinsed with deionized water and dried, the plate was treated with a solution of 1 X lo-' M sodium salicylate for 30 min. After being rinsed and dried, the plate was spotted with samples and developed with 1 X 10" M salicylate for 1.5 min. Because of the low capacity of the plate, NO; (middle spot) begins to show tailing at the highest concentration tested. The results show that even 0.1 nmol(4.6 ng of NO;) of ions can be visually detected. This is an order of magnitude lower in concentration than the results obtained by using the cellulose plate. The developing solution is roughly an order of magnitude lower in salicylate concentration. The detectability found agrees well with the predictions of eq 1. The fluorescence intensity created by lo-' M salicylate is marginal for visual observation, because the blue emission of salicylate is
in a low sensitivity region of our eyes. Thus we were unable to experiment with still lower salicylateconcentrations to take advantage of eq 1. The results here are by no means fully optimized. Several improvements in detectability should be possible with further work. Visual observation is obviously poor compared to photoelectric recording (especially with signal averaging) with respect to D in eq 1. Fluorescence with high-frequency modulation should allow D to improve (increase) by 2 to 3 orders of magnitude (10). The high sensitivityof laser-excited fluorescence should make it compatible with much lower salicylate concentrations, provided the natural fluorescence of the TU: substrate can be discriminated against. Multiple development (15) can effectively control the spread of the spots, so that the local concentration can be higher for the same quantity of analyte. One can also consider other schemes for dynamic modification to approach even lower capacities, analogous to adsorbing cetrimide on reversed-phase materials for IC (16). In summary, we have presented a novel detection method for TLC based on indirect fluorescence. The demonstration system is chosen to elucidate the nature of the response and to establish the criteria for optimization. and not to specifically compete with IC for the same analytes. For example, other fluorescinganions can be used instead of salicylate to enhance separation and/or detection. Since detection is by charge displacement, any analyte anion should give a similar response. The method should also be applicable to the detection of cations (17) and even nonelectrolytes (13) on TLC plates. Finally, the pretreatment of the TLC plates here is necessitated by the lack of suitable commercial products. Those steps could very easily have been incorporated into the manufacturing process to allow essentially real-time monitoring of the separation process. Registry No. NO;, 18851-77-9;I-, MdB1-54-5; S-,1849&-256; IO3-, 15454-31-6;cetrimide, 65-452. LITERATURE CITED (1)
Fried. 0.; shama.J. nh&p ChwnWop~My.2nd ed.; Dekka: Nsw VOn. 1986.
(2) oiddhgs. J. C. Anal. M*nn. 1084. 56, 1258A-1270A. (3) Geldedoo9. 0. 0.;Rovkn. K. L.: M s . J. W.: AT. J. P.:Enks. C. G. AMI. Unn. 191111. 58. 900-903. (4) Fotiou. F. K.: Monk. M. 0. A&. Spscarrc. 19W. 40, 700-704. (5) McClellend. J. F. Anal. Warn. 19113. 55. 89A-105A. (8) Touchslone. J. C.: Dobbins. M. F. Ractlce 01 Thin layer ChromalqFaphy. 2nd ed.: Wiley: New Ymk. 1983; pp 235-273. (7) Belenkii. B. G.; Ganklna. E. S.: AdamovW. T. 0.; LObBzov. A. R.; Necbev. S. V.; Soh-3rko. M. 0. J . chromelogr. 191111. 385. 31537n
(8) S i i l l . H; M k , 1.E. Anal. Unn. 19112. 54. 462.469. (9) EobbMl. D. R.: V a w g . E. S. A M I . &m. 19114. 56. 1577-1581 1101 Mho. S. 1.: Y-. E. S. Anal. Warn. 19115. 57. 2253-2256. ( l l i P&k. V: J . &?q. 1977. ~133.49-57. ' (12) Fritz. J. S.; *de. 0. T.: pohleandt. C. Ion Chromalogaphy; w i l t 6 r g . 1982. (131 Takeurn. 1.;Ysunp. E. S. J . chrometogr. 19116. 368. 145-152. (14) Takeuchi. T.; Yeung. E. S. J . chometogr. 19811. 370.83-92. (15) Jupille. 1. H.: Pew. J. A. J . Chromatog. 1974. 99, 231-239. (16) Pleller. W. D.: T a k d . T.: Ysung. E. S. UVomatqlapMa. in pm. (17) Sherman. J. H.: Oanklsm. N. D. AMI. Chsm. 1987. 59. 1483-1485.
m:
Ames Laboratory-USDOE Chemistry Iowa State University Ames. Iowa 50011
Yinfa M a Edward S. Yeung' and Department of
R ~ C E for ~ Dreview June 25.1987. Accepted December 14, 1987. The Ames Laboratory is operated for the US. Department of Energy by Iowa State University under Contract No. W-7405-Eng-82. This work was supported by the Office of Basic Energy Sciences.
