Anal. Chem. 1997, 69, 2559-2566
Simultaneous Determination of Ionization Constants and Isoelectric Points of 12 Hydroxy-s-Triazines by Capillary Zone Electrophoresis and Capillary Isoelectric Focusing Ph. Schmitt,*,† T. Poiger,† R. Simon,‡ D. Freitag,‡ A. Kettrup,‡ and A. W. Garrison§
National Research Council and National Exposure Research Laboratory, U.S. Environmental Protection Agency, 960 College Station Road, Athens, Georgia 30605-2720, and GSF-Forschungszentrum fu¨ r Umwelt und Gesundheit GmbH, ¨ kologische Chemie, Schulstrasse 10, D-85356 Freising-Attaching, Germany Institut fu¨ r O
Capillary zone electrophoresis (CZE) was used to separate and determine simultaneously the pK1, pK2, and pI values of 12 environmentally relevant hydroxytriazines (hydroxymetabolites of atrazine, terbuthylazine, simazine, and propazine and four (arylamino)-s-triazines) and observe the effects of the alkylamino and arylamino substituents on the measured values. Capillary isoelectric focusing (CIEF) methods were developed to measure the pI of these compounds and compare those values with the CZE-measured pI’s. CZE and CIEF can provide accurate pK and pI values reasonably fast, and pI values measured by the two techniques agree. Knowledge of the pK and pI values of hydroxytriazines is important for an understanding of the binding mechanisms of these molecules in environmental matrices. Because the hydroxytriazines may exist as a myriad of speciessneutral, charged, zwitterionic, and keto-enol tautomerics depending on structure and pH, we briefly addressed the existence of these species relative to their electrophoretic analysis. s-Triazines are widely used as herbicides in many parts of the world. Atrazine, for example, was the most commonly used herbicide in the United States in 1993.1 Because of environmental concerns, the commercial use of atrazine has been forbidden in Germany since 1991,2 when it was replaced by terbuthylazine, which is more strongly adsorbed by soils, less mobile, and less likely to reach groundwater.3 After field application, the triazine herbicides are subjected to various degradation processes (photolysis, oxidation, hydrolysis, biodegradation, etc.), leading primarily to dealkylation of the amine groups in positions 4 and 6 (see structure in Table 1) and/or hydrolysis of the substituent in position 2.4 The latter process gives the corresponding hydroxytriazines, which have been found as contaminants in stream and reservoir water.5 †
National Research Council. Institut fu ¨r O ¨ kologische Chemie. § National Exposure Research Laboratory. (1) Lerch, R. N.; Donald, W. W. J. Agric. Food Chem. 1994, 42, 922-927. (2) Pflanzenschutzanwendungsverordnung vom 22.03, Bundestaggesetzblatt 1 1991; 796-798. (3) Dousset, S.; Louvet, C.; Schiavon, M. Chemosphere 1994, 28, 467-476. (4) Schmitt, Ph.; Freitag, D.; Sanlaville, Y.; Lintelman, J.; Kettrup, A J. Chromatogr., A 1995, 709, 215-225. (5) Adams, C. D.; Randtke, S. J. Environ. Sci. Technol. 1992, 26, 2218-2227. ‡
S0003-2700(96)00945-6 CCC: $14.00
© 1997 American Chemical Society
s-Triazine herbicides can be protonated in acidic solution6 and partition to clays in soils by cation exchange at low pH or by hydrophobic adsorption at neutral pH. Hydroxylated triazines can also be present as anions because of the ionization of the hydroxy groups in more basic solutions. Only scattered data can be found in the literature on values of ionization constants (pK) or isoelectric points (pI) of hydroxytriazines, yet these are essential parameters for understanding and predicting triazine environmental behavior. Capillary zone electrophoresis (CZE) has been found to be a good alternative to classical titration and UV spectroscopy techniques for the rapid measurement of the pK7-11 and pI12 values of ionizable organics; this is partly because of the higher sensititivity of CE relative to the classical methods. Capillary isoelectric focusing (CIEF) has been proven to be a rapid and powerful analytical technique for separating and measuring the pI’s of proteins and DNA fragments using low amounts of sample,13 with possible application to small molecules.14 CE is being increasingly applied to the analysis of organic and inorganic environmental contaminants. For example, the analysis of striazines by CZE as a function of pH has recently been reported.15-17 The objectives of this study were (1) to use CZE to determine simultaneously the pK1, pK2, and pI values of 12 environmentally relevant hydroxytriazines and observe the effects of the alkylamino and arylamino substituents on the measured values and (2) to apply CIEF to measure the pI of these compounds and compare those values with the CZE-measured pI’s. Because the hydroxytriazines may exist as a myriad of speciessneutral, charged, zwitterionic, and keto-enol tautomericsdepending on structure and pH, we briefly address the existence of these species relative to their electrophoretic analysis. (6) Russel, J. D.; Cruz, M.; White, J. L.; Bailey, G. W.; Payne, W. R., Jr.; Pope, J. D., Jr.: Teasley, J. I. Science 1968, 160, 1341-1342. (7) Beckers, J. L.; Everaerts, F. M.; Ackermans, M. T. J. Chromatogr., A 1991, 537, 407-428. (8) Gluck, S. J.; Cleveland, J. A., Jr. J. Chromatogr., A 1994, 680, 43-48. (9) Gluck, S. J.; Cleveland, J. A., Jr. J. Chromatogr., A 1994, 680, 49-56. (10) Cleveland, J. A.; Benko, M. H.; Gluck, S. J.; Walbroehl, Y. M. J. Chromatogr., A 1993 652, 301-308. (11) Cao, J.; Cross, R. F. J. Chromatogr., A 1995, 695, 297-308. (12) Yao, Y. J.; Khoo, K. S.; Chung, M. C. M.; Li, S. F. I. J. Chromatogr., A 1994, 680, 431-435. (13) Chen, S. M.; Wiktorowicz, J. I. Anal. Biochem. 1995, 206, 84-90. (14) Slais, K; Friedl, Z. J. Chromatogr., A 1994, 661, 249-256. (15) Foret, F.; Sustacek, V.; Bocek, P. Electrophoresis 1990, 11, 95-97. (16) Desiderio, C.; Fanali, S. Electrophoresis 1992, 13, 698-700. (17) Schmitt, Ph.; Garrison, A. W.; Freitag, D.; Kettrup, A. J. Chromatogr., A 1996, 723, 169-177.
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997 2559
Table 1. Structures of the 12 Studied Hydroxy-s-triazines OH N R1 N
N N
N R2
hydroxy metabolites
commercial names
R1
R2
short namea
2-hydroxy-4-(ethylamino)-6-(tert-butylamino)-1,3,5-triazine 2-hydroxy-4-(isopropylamino)-6-(isopropylamino)-1,3,5-triazine 2-hydroxy-4-(ethylamino)-6-(isopropylamino)-1,3,5-triazine 2-hydroxy-4-(ethylamino)-6-(ethylamino)-1,3,5-triazine 2-hydroxy-4-amino-6-(tert-butylamino)-1,3,5-triazine 2-hydroxy-4-amino-6-(isopropylamino)-1,3,5-triazine 2-hydroxy-4-amino-6-(ethylamino)-1,3,5-triazine 2-hydroxy-4-amino-6-amino-1,3,5-triazine 2-hydroxy-4-isopropylamino-6-anilino-1,3,5-triazine 2-hydroxy-4-amino-6-anilino-1,3,5-triazine 2-hydroxy-4-(isopropylamino)-6-(3′,4′-dimethoxyanilino)1,3,5-triazine 2-hydroxy-4-amino-6-(3′,4′-dimethoxyanilino)-1,3,5-triazine
hydroxyterbuthylazine hydroxypropazine hydroxyatrazine hydroxysimazine hydroxydesethylterbuthylazine hydroxydesethylatrazine hydroxydesisopropylatrazine hydroxydiaminotriazine (ameline) R234 R243 R240
CH2CH3 CH(CH3)2 CH2CH3 CH2CH3 H H H H CH(CH3)2 H CH(CH3)2
C(CH3)3 CH(CH3)2 CH(CH3)2 CH2CH3 C(CH3)3 CH(CH3)2 CH2CH3 H C6H5 C6H5 C6H3(OCH3)2
[Et, tBu]-ameline [iPr, iPr]-ameline [Et, iPr]-ameline [Et, Et]-ameline [H, tBu]-ameline [H, iPr]-ameline [H, Et]-ameline ameline [iPr, Ar]-ameline [H, Ar]-ameline [iPr, mAr]-ameline
R237
H
C6H3(OCH3)2 [H, mAr]-ameline
a
Short name adopted as a function of different substitution of ameline.
MATERIAL AND METHODS Separations were performed with a Beckman P/ACE 5000 Series HPCE with Beckman System Gold Chromatography Software version 8.1. Capillary Zone Electrophoresis. For CZE separations, the fused-silica CE column (75 µm i.d.; 375 µm o.d.; 50 cm length to detector and total length of 57 cm) was obtained from Beckman Instruments, Inc. The separation runs were done at constant temperature (30 °C) and voltage (20 kV) with UV/visible filter detection at 230 nm. Hydrodynamic sample injection for 5 s was the sample introduction mode for all experiments. Two buffer systems were used, covering a pH range of 2.0512.45. The citrate/phosphate buffers were prepared by diluting 1/10 in distilled water the mixed stock solutions of disodium hydrogen phosphate (0.2 M) and citric acid (0.1 M) in ratios of 1/50 (pH 2.2) to 50/1 (pH 7.8). The carbonate buffers were prepared by diluting 1/10 in distilled water the mixed stock solutions of sodium carbonate (0.1 M) and sodium dihydrogen carbonate (0.1 M) in ratios of 1/20 (pH 9.0) to 20/1 (pH 11.0). The buffers at pH under 2.5 and over 12.0 were prepared by adjusting the citrate/phosphate and the carbonate buffer with 0.1 N HCl and 0.1 N NaOH, respectively. All the buffer pH were checked before each run (errors in pH measurements were of (0.01 pH unit at room temperature). Disodium hydrogen phosphate, disodium citrate, citric acid, sodium carbonate, and sodium dihydrogen carbonate, all p.a. grade, were obtained from Fisher Scientific Co. A 2 min washing cycle with 0.1 M NaOH was followed by a 2 min conditioning of the capillary with the run buffer before the sample injection; each measurement was ended with a 2 min 0.1 M NaOH washing cycle; this washing cycle between the measurements assured a good conditioning of the capillary wall surface, thus avoiding hysteresis effects with changes in the pH of the running buffer. Mesityl oxide was used as internal EOF marker and was purchased from Micro-Solve CE, Scientific Resources Inc., Eatontown, NJ. Herbicide stock solutions were prepared by dissolving 2.0 mg of each herbicide in 12.5 mL of pesticide grade 2560
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
methanol slightly acidified with 1 N HCl; 20 µL of this solution was mixed with 380 µL of distilled water to reach a final concentration of the mixture of the 12 triazines of 8 ppm each. All buffers and stock solutions were kept under refrigeration (4 °C). Capillary Isoelectric Focusing. (a) pH 3-10. The isofocusing separation was done following the instructions in the Beckman CIEF kit, which contained all required supplies. Briefly, 4 µL of ampholyte mixture (pI 3-10) was added to 200 µL of CIEF gel (pI 3-10) provided in the kit. A small amount of this gel/ampholyte mixture (40 µL) was inserted in a 50 µL Beckman CE vial and mixed with 3 µL of triazine solution (40 ppm) and 1 µL of the protein marker solution; final triazine concentration was 2.7 ppm. Four protein markers were provided with the kit: ribonuclease A (pI 9.45), carbonic anhydrase II (pI 5.9), β-lactoglobulin A (pI 5.1), and CCK flanking peptide (pI 2.75). The neutral eCAP capillary (50 µm i.d., 20 cm to detector) was first conditioned with distilled water (10 min), then with 10 mM phosphoric acid (5 min), and then filled hydrodynamically with the gel/ampholyte/pesticide/ marker mixture (2 min). The focusing step (2 min, 500 V/cm) was followed by the mobilization step, which consisted of the combined application of low pressure (0.5 psi) and voltage (13.5 kV, 500 V/cm) from the anolyte (91 mM phosphoric acid in the CIEF gel) to the catholyte (20 mM NaOH in water). The initial current of a few microamperes decreased during the first minutes to reach 1 µA during the mobilization; the filling of the capillary with the anolyte caused an increase of the current at the end of the mobilization. After about 40 min, the CCK flanking peptide passed the detector. The capillary was rinsed with 10 mM phosphoric acid (1 min) and distilled water (2 min) after each separation. (b) pH (5-8). Commercially available ampholytes, Pharmalyte pH 5-8 and Ampholyte pH 5-8, were purchased from Sigma (St. Louis, MO). For calibration of the separations with these ampholytes, IEF standards with pI range 4.45-9.6 from Bio-Rad Laboratories (Hercules, CA) were used. Procedures and other chemicals were the same as for the pH 3-10 ampholytes from Beckman.
Sources of Hydroxytriazines. The hydroxy-(alkylamino)-striazines (Table 1) were purchased in greater than 99% purity grade from Dr. Ehrenstorfer GmbH, Augsburg, Germany, or from Riedel-de Hae¨n (Pestanal grade), Munich, Germany. The hydroxy-(arylamino)-s-triazines (R234, R237, R240, and R243 Table 1) were synthesized from the corresponding 2-chloro-4-(arylamino)-6-(alkylamino)-s-triazines by acidic hydrolysis in 1:1 acetonitrile/water at 80 °C. The 2-chloro-4-(arylamino)-6-(alkylamino)s-triazines were synthesized from cyanuric chloride as reasonable model compounds to study the fate of bound atrazine residues in soil and water systems;18 the hydroxy analogues have been found as their major abiotic degradation products. There are no particular safety precautions necessary for research with hydroxy-s-triazines beyond those taken for any pesticide; e.g., the neat standards should only be handled in a hood. Dilute solutions such as used in this research should be handled like any pesticide standard solution. The toxicities of hydroxy-s-triazines are generally unknown but are not expected to be more than that of the parent triazines, which are not acutely toxic. The LD50 of atrazine, for example, is 1750 mg/kg. Fitting of the experimental data was performed with Kaleida Graph version 2.1 for MacIntosh. THEORETICAL BACKGROUND Hydroxy-s-triazines are weak bases; a ring nitrogen is protonated at acidic pH. At higher pH values the ionization of the hydroxy group leads to the formation of an anionic species. Electrophoretic separations of hydroxy-s-triazines were achieved with CZE in previous studies both at low pH (cationic species) and high pH (anionic species) as a function of their charge-tomass ratio.17 The degree of ionization determines the electrophoretic mobility of these substances and allows their separation in mixtures. The ionization of the hydroxy-s-triazine (HZ) is governed by both the protonation of the overall neutral HZ to give the overall cationic species H2Z+ and its dissociation to give the overall anionic species Z-: K1
pK1T ) pK1 + log γH2Z+ K2T )
γZ- {H+}[Z-] γZK ) γHZ [HZ] γHZ 2
pK2T ) pK2 - log γZ-
(1’) (1′′) (2’) (2’’)
assuming that γHZ is equal to 1. The extended Debye-Huckel theory defines the activity coefficient of an anion A- in dilute solution (ionic strength 0.999) was found for these measurements; the relative standard deviation of retention times of the markers was 3.5% (n ) 14). The reproducibility in the calculation of the pI of [Et, iPr]-ameline from the linear regression was found to be 0.05 pH unit with a maximum difference of 0.1 pH unit (n ) 5); thus, 0.1 was taken as the error value. The measured pI values of all 12 hydroxytriazines with the pI 3-10 ampholytes are given in Table 2. The goal was then to try to establish a CIEF method with a narrow pH gradient to allow more accurate pI determination of the zwitterionic triazines. Several ampholytes in the range pI 5-8 are commercially available; Amplolyte 5-8 and Pharmalyte 5-8 were tried in this study. Ampholyte 5-8 showed linearity of the Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
2563
Figure 5. CIEF electropherograms of the separation of the 12 hydroxy-s-triazines in the Pharmalyte 5-8 pH gradient at 230 and 254 nm (two runs, pesticide concentration of 11.3 ppm): (A) [Et, tBu]ameline; (B) [iPr, iPr]-ameline, [Et, Et]-ameline and [H, Et]-ameline; (C) [Et, iPr]-ameline; (D) [H, iPr]-ameline and [H, tBu]-ameline; (E) ameline; (F) [iPr, Ar]-ameline and [iPr, mAr]-ameline; (G) [H, Ar]ameline and [H, mAr]-ameline.
pH gradient in the column but had too high a background absorbance at 254 nm. Pharmalyte 5-8 gave a low background at 230 and 254 nm and was chosen for this pI determination. The runs were stopped after 40 min, giving a linear pH gradient in the measured range of 8.2 to 6.5 using five protein markers: lentil lectin (two peaks of pI 8.2 and 8.0), human hemoglobin C (pI 7.5), human hemoglobin A (pI 7.1), equine myoglobin (two peaks of pI 6.8 and 7.0), and human carbonic anhydrase (pI 6.5). A good linearity was found between the migration times and the pI for these six markers (r2 > 0.991) with a relative standard deviation of 1.8-4.6% in the retention times (n ) 5) in consecutive measurements. Triazines were focused with the markers, and the pI was determined from the linear regression of the internal standards (as in the case with the pI 3-10 ampholytes). [iPr, iPr]-ameline (hydroxypropazine) was tested for the reproducibility of the calculated pI and showed a standard deviation of 0.005 pH unit (n ) 5). The calculated pI for all 12 hydroxytriazines with the Pharmalyte 5-8 are given in Table 2. The separation of a mixture of the 12 hydroxytriazines with the Pharmalyte 5-8 is illustrated in Figure 5. Due to the close pI values of some of the hydroxytriazines, only 7 of the 12 triazines could be separated. Their elution order is the same as their pI values measured in separate runs; furthermore, by calculating the pI of the hydroxytriazines with the equation for linear regression between the the pI and the elution times of the two extreme triazines ([Et, tBut]-ameline, [H, Ar]-ameline), one finds pI values in the range of the measured values with errors of 0.03 pH unit. Note that the 25 min run in Figure 5 shows a separation window with a linear pH gradient from 7 to 8 of only 1 pH unit; the corresponding 40 min run with the 3-10 ampholyte provided by the Beckman kit had a much higher background (factor of 5) and showed no separation of B and C. DISCUSSION A good parallel was found between the pK1 and the pK2 values measured by CZE, as shown in Figure 3A. The calulated pKa values can be related to the structure of the hydroxy-s-triazines (Figure 3A): the more the triazines are alkyl-substituted, the higher are the pK values; this is in accordance with previous results found with methoxy-s-triazines.20 Terbutyl substitution has the highest influence on the pK values; [Et, tBu]-ameline and [H, tBu]-ameline have higher measured pK values than the (diiso2564 Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
propyl-substituted) [iPr, iPr]-ameline and [H, iPr]-ameline, respectively. Methyl groups are less electronegative than H and are thus better able to repel electrons on the attached nitrogen atom; thus tertiary groups (tBu) increase the electron density of the basic ring nitrogens more than secondary (iPr) and primary (Et) groups.24 Furthermore, all dealkylated (proton-substituted) molecules have a lower pK1 value than their corresponding alkylated species. The substitution of an aryl group for alkyl groups of the hydroxy-s-triazines showed even lower pK1 and pK2 values than ameline. Aryl groups (Ar, mAr) withdraw electrons more than the alkyl groups, so they decrease the basicity of the triazines.23 No significant differences could be observed upon substitution of two methoxy groups on the aromatic substituent ([H, mAr]) but N-alkylation of the aryl-substituted triazines increased the pKa values. The pI's vary in the same direction as the pK’s as a function of the alkyl/aryl substitution for the same reasons (Figure 3B). The best correlation between the CZE calculated pI and the CIEF measured pI values was found with the Pharmalyte 5-8. The ampholyte 3-10 measurements underestimated the calculated pI relative to the CZE values, while the 5-8 ampholytes overestimated most of them (Figure 3B). The differences in the measured pI values with the two ampholyte systems may be due to specific interactions of the triazines with the ampholyte matrix. Pharmalytes are branched polyamino polycarboxylic acids, while Ampholines, which include the ampholyte 3-10 used here, are linear polyamino polycarboxylic acids, and Servolytes are composed of a mixture of polyamino polysulfonic acids and polyamino polycarboxylic acids.27 Both ionized and neutral hydroxytriazines have potential binding possibilities with the zwitterionic ampholytes (through ionic, polar, or electrostatic interactions); hydrogen bonding was, for example, found to be a major interaction mechanism between hydroxyatrazine and carboxyl groups as a function of pH.28,29 In summary, Pharmalyte 5-8 is the best ampholyte system for determination of the isoelectric points of the hydroxytriazines with CIEF. As shown here, the hydroxytriazines can be measured as either cations or anions by CZE, depending on the pH of the buffer system, and can exist as neutral species so that they migrate with the electroosmotic flow. In fact, these compounds may exist in a variety of formssneutral (uncharged), zwitterionic (with balanced charges), charged (overall positive or negative), and keto-enol tautomers, depending on their structure and the pH of the medium. The pKa values in Table 2 were obtained by a fit assuming that only three speciessZ-, HZ, and H2Z+sexist. The good fit for all compounds supports this assumption (Figure 2). However, the residuals of the mobilities (Figure 2B) show clear trends for all compounds investigated, indicating that the situation is more complex; i.e., the observed macroconstants (pK1, pK2, and pI in Table 2) are probably combined values for several very closely related structures, each of which has a slightly different microconstant. To illustrate, some possible subspecies of the three major species are shown in Figure 6. For example, two sets of closely related structures are depicted for a generic cationic hydroxytri(26) Weber, J. B. Soil Sci. Am. Proc. 1970, 34, 401-404. (27) Righetti, P. G.; Gianazza, E.; Bianchi Bosisio, A. In Recent Developments in Chromatography and Electrophoresis; Frizero, Renoz, Eds.; Elsevier: Amsterdam, 1979, pp 1-37. (28) Welhouse, G. J.; Bleam, W. F. Environ. Sci. Technol. 1993, 27, 500-505. (29) Welhouse, G. J.; Bleam, W. F Environ. Sci. Technol. 1992, 26, 959-964.
Figure 6. Some of the possible subspecies of the three major species (cationic, neutral, anionic) of a generic hydroxytriazine.
azine at low pH, at which either one of the basic ring nitrogens is protonated or the oxygen deprotonated. One set of resonance hybrids is shown enclosed in brackets at the top of the figure; these would constitute one structure with one microconstant. Similar resonance sets could be depicted for initial protonation at each of the other two ring nitrogens. These three sets of structures would have three slightly different microconstants because the molecule is assymetric (R1 * R2) and the “added” protons are in different environments. In addition, there could
exist a fourth, fifth, and sixth set of resonance structures (and microconstants) caused by protonation of the oxygen. One of these sets is shown in the brackets at the bottom of Figure 6. These six sets of structures are in a state of equilibrium which shifts with pH, ionic strength, and other properties of the medium, changing the observed macroconstant. Several structural variants can be depicted at neutral pH, also (Figure 6). There is only indirect evidence for most of these structures. For example, zwitterions (and the keto form), as shown in the Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
2565
top bracket at neutral pH in Figure 6, are speculated to be much more predominant than the oxygen protonated (enol) form; a charged species would be more stable in water than an uncharged one (the enol) because of dipole interactions. In support of this speculation, the decreased water solubility of hydroxyatrazine at neutral pH relative to atrazine may be because of the high crystal energy imparted to solid hydroxyatrazine by charge interactions between the zwitterions, which cannot occur with atrazine. Further discussion of these structural variants is beyond the scope of this paper and, indeed, generally beyond the state of the art. However, recent work to expand the computer models for prediction of chemical reactivity constants used with the SPARC (SPARC Performs Automated Reasoning in Chemistry) program indicates that models may soon be available for predicting pK and pI values of such species as depicted here, although the ketoenol tautomers offer a particular challenge.30 CONCLUSION Knowledge of the pK and pI values of hydroxytriazines is important for an understanding of the binding mechanisms of these molecules in environmental matrixes. Binding mechanisms, for example, electrostatic binding of hydroxytriazines with humic substances, invariably depend to some extent on the charge(s) on the molecule, which is related by pH to its pKa, and to spatial distribution of that charge. CZE and CIEF can provide accurate pK and pI values reasonably fast, and pI values measured by the two techniques
agree. Use of the CIEF technique for the analysis of zwitterionic pesticides is limited by the following: (1) the type of ampholyte, which determines the pH window, the linearity of the gradient, the base line of the electropherogram, and the optimum detection wavelength; (2) the purity of the pI markers, because CE is much more sensitive to impurities than classical IEF separation techniques; and (3) the gel medium, which determines the viscosity and UV transparency of the background. Interactions of the ampholyte system with the analyte can occur, and this factor should not be underestimated when measuring pI and pKa. Finally the values are macroconstantsscomposite values of the microconstants corresponding to the many pH-dependant structural variants of the hydroxytriazines. ACKNOWLEDGMENT This work was performed while Ph.S. held a National Research Council Research Associateship at the Ecosystem Research Division, Natural Exposure Research Laboratory, U.S. EPA, Athens, GA. The authors thank Dr. R. Bae from Beckman Instruments, Inc., for providing the Beckman CIEF method development kit. Special appreciation is expressed to Sam Karickhoff, Lionel Carreira, and Joe D’Angelo for profitable discussions on hydroxytriazine structure and chemistry. Received for review September 17, 1996. Accepted March 31, 1997.X AC960945S
(30) Hilal, S., Carreira, L. A.; Karickhoff, S. W., submitted to Quant. Struct. Act. Relat.
2566
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
X
Abstract published in Advance ACS Abstracts, June 1, 1997.
