Analysis of the Solvatochromic Behavior of the Disubstituted

The coincident spectral components of the main visible absorption band of the triphenylmethane dye Brilliant Green can be resolved using a nonlinear ...
0 downloads 0 Views 64KB Size
Anal. Chem. 1998, 70, 4212-4217

Technical Notes

Analysis of the Solvatochromic Behavior of the Disubstituted Triphenylmethane Dye Brilliant Green Kerry K. Karukstis* and Aaron V. Gulledge

Department of Chemistry, Harvey Mudd College, Claremont, California 91711

The coincident spectral components of the main visible absorption band of the triphenylmethane dye Brilliant Green can be resolved using a nonlinear least-squares fitting routine. This analysis reveals two Gaussian components, with centers near 600 and 630 nm, that possess additional spectral characteristics with a pronounced sensitivity to the dye’s microenvironment. In particular, both the ratio of the heights of these Gaussian peaks and the ratio of their areas exhibit an inverted dependence on solvent dielectric constant, i.e., a reversal of solvatochromic behavior in media of low polarity. Such an absorption spectrum analysis facilitates the use of triphenylmethane dyes that lack symmetric trisubstitution of the phenyl rings as environment-sensitive spectroscopic probes. We demonstrate further the utility of this approach through the spectrophotometric determination of the critical micelle concentration of the anionic surfactant sodium dodecyl sulfate using Brilliant Green as an optical probe. The para-substituted triphenylmethane dyes represent a class of synthetic dyes of commercial and analytical importance.1 Numerous applications capitalize on the intensity, range, and lightfastness of color exhibited by these dyes. For example, triphenylmethane dyes function as colorants in the textile, food, cosmetic, and ink industries, as saturable absorbers in laser mode locking, as reagents in protein assays, as histological stains, and as indicators in spectrophotometric determinations of surfactants, metal ions, and pesticides.1 The spectral features of triphenylmethane dyes in liquid media have been investigated extensively.1-4 The main visible absorption band of this class of dye molecules primarily consists of two overlapping spectral components, designated as R and β, with the more intense R band red-shifted relative to the β band. At high dye concentrations, a shorter-wavelength band (designated γ or * To whom correspondence should be addressed: (e-mail) Kerry_Karukstis@ hmc.edu; (fax) (909) 607-7577; (phone) (909) 607-3225. (1) Duxbury, D. F. Chem. Rev. 1993, 93, 381-433. (2) Lueck, H. B.; McHale, J. L.; Edwards, W. D. J. Am. Chem. Soc. 1992, 114, 2342-2348. (3) Sheppard, S. E.; Geddes, A. L. J. Am. Chem. Soc. 1944, 66, 1995-2002. (4) Lewis, G. N.; Magel, T. T.; Lipkin, D. J. Am. Chem. Soc. 1942, 64, 17741785.

4212 Analytical Chemistry, Vol. 70, No. 19, October 1, 1998

µ) appears. For example, the wavelengths of the R, β, and µ bands for the triphenylmethane dye crystal violet are reported at 590, 550, and 510 nm, respectively.3 Considerable controversy exists as to the origin of the R and β bands, with studies attributing the phenomenon to dye aggregation, dye tautomerism, conformational isomerism, and dye-counterion interactions.1 The spectral sensitivity of the R and β bands to solvent varies with the substitution pattern of the three phenyl rings.2 Symmetrically para-trisubstituted triphenylmethane dyes exhibit a pronounced shoulder for the β band in aqueous solution and clearly resolved R and β bands in nonpolar solvents.2 Triply substituted triphenylmethane dyes of lower symmetry also show well-resolved R and β bands in nonpolar solvents.2 For example, the R and β bands of symmetrically trisubstituted ethyl violet and asymmetric triply substituted Victoria pure blue are separated by more than 100 nm in benzene.2 The relative intensities of the R and β bands of triply substituted dyes also exhibit a notable sensitivity to solvent dielectric constant2,5 and concentration.1 For example, as the solvent is changed from water to methanol to benzene, the relative absorbance values at the maximal wavelengths designated for the β and R bands (β:R) change from approximately 0.67 to 0.77 to 0.65 for crystal violet and from 0.7 to 1.0 to 0.85 for Victoria pure blue.2 Furthermore, for aqueous solutions of these dyes, increasing dye concentration initially diminishes the R band while enhancing the β component; at even higher dye concentrations, both bands are diminished as the short-wavelength γ or µ band appears. In contrast, a much reduced sensitivity of spectral shape to solvent2,3 and concentration6,7 is observed for asymmetric doubly substituted triphenylmethane dyes. Red-shifts of less than 10 nm have been reported for the absorbance maximum of a dimethylamino-substituted dye as the solvent is changed from methanol to benzene.2 For 100-fold changes in dye concentration, only small enhancements of the absorption at the wavelength of the R band with no change in the absorbance at the β band wavelength have been observed.6 As a consequence of the relatively negligible change in spectral parameters with microenvironment, asymmetric (5) Korppi-Tommola, J.; Kolehmainen, E.; Salo, E.; Yip, R. W., Chem. Phys. Lett. 1984 104, 373-377. (6) Oshima, M.; Motomizu, S.; Doi, H. Analyst 1992 117, 1643-1646. (7) Michaelis, L.; Granick, S. J. Am. Chem. Soc. 1945, 67, 1212-1219. S0003-2700(98)00318-7 CCC: $15.00

© 1998 American Chemical Society Published on Web 08/29/1998

doubly substituted triphenylmethane dyes are generally not selected for spectrophotometric determination of analytes.1,6 The asymmetric disubstituted triphenylmethane dye known as Brilliant Green is a cationic dye with p-ethylamino substitution on two of the phenyl rings. In aqueous solution at pH values above 2.6, the basic form of Brilliant Green predominates with an absorption maximum at 624 nm. We have examined the spectral dependence of Brilliant Green in solvents of varying dielectric constant. As anticipated, the experimentally observed absorption spectra exhibit small and irregular changes in response to the microenvironment. However, using data analysis techniques capable of resolving an individual absorbance spectrum into a sum of overlapping Gaussian absorption bands, we have revealed spectral parameters that exhibit an unusual and pronounced sensitivity to solvent dielectric constant. Furthermore, we illustrate the use of this spectral analysis technique to determine the critical micelle concentration (cmc) of the common anionic surfactant sodium dodecyl sulfate (SDS) in the presence of Brilliant Green. EXPERIMENTAL SECTION Materials. Brilliant Green (Aldrich, CA registry number 63303-4, N-[4-[[4-(diethylamino)phenyl]phenylmethylene]-2,5-cyclohexadien-1-ylidene]-N-ethylethanaminium sulfate (1:1), also known as Basic Green 1, Diamond Green G, Fast Green J, Ethyl Green, and Emerald Green) was obtained as the hydrogen sulfate salt. The surfactant sodium dodecyl sulfate (Sigma, 99%) was used without further purification. Solvent studies were performed using the following solvents (supplier, Aldrich, unless otherwise specified): methanol (99.9+%), ethanol (Quantum Chemical Corp., 200 proof, dehydrated), 1-propanol (99.5+%), 1-butanol (99.4+%), 1-pentanol (99%), 1-hexanol (98%), 1-heptanol (98%), 1-octanol (99%), 1-decanol (99%), 2-propanol (Spectrum, 99.5%), 1,2ethanediol (Spectrum, 99.0%), 1,2-propanediol (Spectrum, 99.5%), 1,3-butanediol (99+%), cyclopentanol (99%), 1,4-dioxane (99+%), deuterium oxide (D2O, Norrell, 99.9 atom % D), dimethyl sulfoxide (DMSO, Spectrum, 99.9%), N,N-dimethylformamide (DMF, Eastman, 99.9%), 1,1-dimethoxyethane (DMOE, Eastman, 99%), acetone (99.5%), and tetrahydrofuran (THF, Spectrum, 99.0%). Absorbance Measurements and Analyses. Absorption spectra over the range of 400-700 nm were recorded at 25 °C on a Beckman DU-650 UV/VIS spectrophotometer at a scan rate of 1200 nm/min. A nonlinear least-squares fitting routine (PeakFit, Jandel Scientific) was used to deconvolute individual absorption spectra into a sum of overlapping Gaussian functions with frequency as the independent variable. All spectra were fit using an iterative Marquardt-Levenberg fitting algorithm to obtain the minimum number of absorbing components that yielded an r2 of at least 0.999 with a random scattering of residuals. The center, amplitude (or height), width, and area of each Gaussian function were characterized. RESULTS Absorption Spectra of Brilliant Green in Various Solvents. For all solvents examined, the absorption spectrum of Brilliant Green shows a main absorption maximum (the R band) near 630 nm with a detectable shoulder (the β band) near 590 nm. A small absorption band also occurs near 430 nm. Figure 1 shows three spectra for 5 µM Brilliant Green in (a) deuterium oxide (λmax )

Figure 1. Absorption spectra for Brilliant Green in (a) deuterium oxide (λmax ) 623 ( 1 nm), (b) 1-pentanol (λmax ) 630 ( 1 nm), and (c) dimethyl sulfoxide (λmax ) 637 ( 1 nm).

623 ( 1 nm), (b) 1-pentanol (λmax ) 630 ( 1 nm), and (c) dimethyl sulfoxide (λmax ) 637 ( 1 nm). Table 1 summarizes the maximal absorption wavelengths recorded for 5 µM Brilliant Green in each solvent. No simple dependence on solvent dielectric constant is observed. However, for solvents of similar chemical structure such as the primary straight-chain alcohols, the maximal absorption wavelength shifts to the red as the length of the hydrocarbon chain increases, from a minimum of 625 ( 1 nm for methanol to a maximum value of 631 ( 1 nm for primary alcohols of seven carbons or more. Gaussian Analysis of Brilliant Green Absorption Spectra in Various Solvents. The absorption spectra of Brilliant Green in the various solvents were deconvoluted into a sum of Gaussian peaks. The main absorption band observed near 630 nm was resolved into three peaks centered on average near 632 ( 2, 604 ( 2, and 528 ( 3 nm, with the latter Gaussian contributing only about 1% of the area of the overall spectrum. The shortwavelength absorption band at 430 nm was resolved into two Gaussians centered on average at 431 ( 2 and 405 ( 3 nm. With the proximity of the center of the 405-nm peak to the spectrum boundary, no further analysis on the low-wavelength absorption band was conducted. Figure 2 illustrates the deconvolution of the Brilliant Green absorption spectrum in water. Table 1 reports the center of each Gaussian resolved for the 630-nm absorption band in all solvents studied. As the solvent varied, the two dominant Gaussians of the main absorption band maintained a generally constant separation of centers of 28 ( 1 nm. The centers of the 604- and 632-nm peaks exhibited no regular dependence on solvent dielectric constant. However, Figure 3 presents the dependence on solvent dielectric constant of (a) the ratio of heights of the two Gaussian components near 600 and 630 nm comprising the long-wavelength absorption band (height at 630 nm/height at 600 nm) and (b) the ratio of the percent area contributions of the 600-nm and 630-nm components (percent area 630 nm/percent area 600 nm). Pronounced increases in both the height ratio and the area ratio occur as the solvent  decreases to near 10-15, followed by a decrease in both ratios for solvents of lower dielectric strength. The dotted lines in Figure 3 are included to highlight these trends. Table 1 also summarizes the height and area ratios of the main Gaussian bands near 604 and 632 nm for each solvent. Absorption Spectra of Brilliant Green as a Function of SDS Concentration. The absorption spectrum of an aqueous Analytical Chemistry, Vol. 70, No. 19, October 1, 1998

4213

Table 1. Absorption Parameters of Brilliant Green in Various Solvents Gaussian analysisc solvent

a

observed λmax/nmb

λ1/nm

λ2/nm

λ3/nm

ht1/ht2d

%A1/%A2e

water deuterium oxide dimethyl sulfoxide ethylene glycol N,N-dimethylformamide methanol propylene glycol 1,3-butanediol ethanol acetone 1-propanol 2-propanol cyclopentanol 1-butanol 1-pentanol 1-hexanol 1-heptanol 1-octanol 1-decanol tetrahydrofuran dimethoxyethane p-dioxane

78.54 78.25 45.0 37.0 36.7 32.63 32.0 28.8 24.3 20.7 20.1 18.3 18.0 17.1 13.9 13.3 12.1 10.3 8.1 7.6 7.1 2.21

624 623 637 632 632 625 631 631 627 627 629 627 631 629 630 630 631 631 631 631 629 633

627 625 639 634 635 627 632 632 629 629 630 629 634 631 632 632 632 633 633 633 631 636

600 600 612 606 605 599 604 606 601 600 602 601 605 602 602 604 604 604 604 603 605 607

530 530 533 526 529 524 525 525 524 527 528 523 529 528 533 527 529 527 532 527 525 534

0.73 0.69 0.76 0.82 0.91 0.87 0.85 0.89 0.93 0.93 1.00 0.92 0.94 0.97 1.00 0.97 0.98 0.91 0.85 0.91 0.91 0.77

1.59 1.50 1.79 1.95 1.98 1.98 2.08 2.12 2.10 2.03 2.01 2.10 2.18 2.20 2.24 2.14 2.14 2.17 2.08 2.00 1.93 1.50

a Dielectric constants taken from ref 20. b λmax value of observed absorption band reported to nearest nanometer. c λmax values of resolved Gaussian curves reported to nearest nanometer. d Ratio of height of Gaussian peak at λ1 to height of Gaussian peak at λ2. e Ratio of percent area of Gaussian peak at λ1 to percent area of Gaussian peak at λ2.

Figure 2. Deconvolution of the Brilliant Green absorption spectrum in water (25 µM Brilliant Green). Five overlapping Gaussian curves with centers at 627, 600, 530, 427, and 400 nm contribute to the overall spectrum (indicated by dotted lines). The experimental absorption spectrum and the theoretical fit to the experimental curve from the sum of the contributing Gaussian components are superimposed.

solution of 25 µM Brilliant Green changes dramatically as the surfactant SDS increases in concentration. Figure 4 presents representative spectra obtained for an SDS concentration range from 0 to 22 mM. Extensive variation in spectral shape is observed at low SDS concentrations, particularly from 250 µM to 2.0 mM, with no change in spectral appearance at SDS concentrations above 8.0 mM. No change in absorbance intensity is observed above 12 mM SDS. The wavelength of maximum absorbance shifts steadily from 623 nm in the absence of SDS to 631 nm at or above 8 mM SDS, as illustrated in Figure 5. Gaussian Analysis of Brilliant Green Absorption Spectra as a Function of SDS Concentration. Figures 5 and 6 summarize the Gaussian analysis of the Brilliant Green absorption spectra recorded for 25 µM Brilliant Green as a function of SDS concentration. The main absorption band centered near 630 nm 4214 Analytical Chemistry, Vol. 70, No. 19, October 1, 1998

Figure 3. Gaussian analysis of the absorption spectra of Brilliant Green in a variety of solvents. The dependence on solvent dielectric constant of (a) the ratio of heights of the two Gaussian components near 630 and 600 nm comprising the long-wavelength absorption band (height at 630 nm/height at 600 nm) and (b) the ratio of the percent area contributions of the 630-nm and 600-nm components (percent area 630 nm/percent area 600 nm). The straight lines are included to underscore the trend in the ratios with varying solvent dielectric constant. The straight lines are linear regression lines for dielectric constants in the ranges of 2-11 and 12-80 for height ratios and in the ranges of 2-14 and 17-45 for area ratios.

is resolved into two principal Gaussian components whose centers and ratios of heights and peak areas exhibit dramatic variations as SDS concentration increases. The longer-wavelength component at 627 nm in the absence of SDS steadily shifts to a center at 640 nm at 1.5 mM SDS and then reverses direction to 634 nm

Figure 4. Representative absorption spectra for 25 µM Brilliant Green in aqueous solutions in the presence of (a) 0, (b) 1.0, (c) 2.0, (d) 4.0, (e) 8.0, and (f) 12.0 mM SDS. Figure 6. Analysis of the dominant Gaussian components of the 630-nm absorption band for 25 µM Brilliant Green as a function of SDS concentration. The long-wavelength component at 627-640 nm is denoted as peak 1; the shorter-wavelength component at 600611 nm is denoted as peak 2. Both (a) the ratio of peak heights (peak 1/peak 2) and (b) the ratio of percent area contributions (peak 1/peak 2) of the Gaussians display a minimum at 1.0 mM SDS and a relatively constant value at SDS concentrations above the known cmc. As described in the text, the cmc of SDS is quantitatively determined as the intersection of the two linearly extrapolated lines through the rising portion of ratios (estimated as 4-7 mM SDS) and the essentially constant portion of ratios (estimated as 10-22 mM SDS).12, 13 The intersection occurs at 8.4 ( 0.2 mM in (a) and at 8.1 mM ( 0.2 in (b).

Figure 5. Wavelength of maximum absorbance of 25 µM Brilliant Green in aqueous solutions of varying SDS concentration, given by the open squares (0). The λmax value shifts steadily from 623 nm in the absence of SDS to 631 nm at or above 8 mM SDS. Gaussian analysis of the Brilliant Green absorption spectra recorded for 25 µM Brilliant Green as a function of SDS concentration results in the deconvolution of the main 630-nm absorption band into two principal Gaussian components represented by b and 2. The longerwavelength component at 627 nm in the absence of SDS steadily shifts to a center at 640 nm at 1.5 mM SDS and then reverses direction to 634 nm at 8 mM SDS and above. The shorter-wavelength Gaussian at 600 nm in aqueous solution also red-shifts its center to 611 nm at 1.0 mM SDS and then reverses direction to 604 nm at [SDS] g 7 mM.

at 6 mM SDS and above (Figure 5). The shorter-wavelength Gaussian at 600 nm in aqueous solution also red-shifts its center to 611 nm at 1.0 mM SDS and then reverses direction to 604 nm at [SDS] g 7 mM (Figure 5). Both the ratio of peak heights (Figure 6a) and the ratio of percent area contributions of the Gaussians (Figure 6b) also display dramatic variations as SDS concentration is altered, with a minimum in both parameters observed at 1.0 mM SDS and a relatively constant value at SDS concentrations at or above 10 mM. For all SDS concentrations, a third minor Gaussian exhibits a rather fixed center at 531 ( 1 nm with a constant area contribution to the overall spectrum of 2.2 ( 0.3%. For SDS concentrations of 250, 500, and 750 µM, a

fourth Gaussian is resolved at 670 ( 2 nm with an area contribution of 10, 15, and 26%, respectively. DISCUSSION The optical and solubility properties of molecular probes are of fundamental importance in the study of surfactants and micellar systems. In particular, the sensitivity of the absorption wavelength or intensity to the polarity of the probe’s environment is essential for the characterization of aqueous micellar structure. Furthermore, dyes possessing a high absorptivity in aqueous solution as well as solubility in a wide range of solvents are most desirable. The ability to form 1:1 complexes with ionic surfactants also permits the detection of low concentrations of such amphiphilic molecules for analytical determinations. The disubstituted triphenylmethane dye Brilliant Green satisfies several of the requisite criteria for molecular probes. Although a cationic dye, Brilliant Green exhibits solubility in a wide range of media. A high molar absorptivity is also displayed in both polar and nonpolar solvents. However, by the conventional spectral parameters of maximal absorption wavelength or overall absorption intensity, Brilliant Green appears to be a less sensitive optical probe of microenvironment. Using a new approach involving data analysis techniques capable of resolving individual absorbance spectra into a sum of overlapping Gaussian functions, we have revealed notable new findings. In particular, the height ratios and percent area ratios of the two principal Gaussian peaks Analytical Chemistry, Vol. 70, No. 19, October 1, 1998

4215

comprising the main absorption band of the Brilliant Green spectrum show a pronounced dependence on the polarity of the microenvironment. The ratios of the amplitudes and percent area contributions of the two principal Gaussian components reveal an inverted dependence on solvent dielectric constant, i.e., inverted solvatochromism or a reversal of solvatochromic shifts in media of low polarity. This unusual solvatochromic behavior may have been anticipated from the observation noted above for the solvent dependence of the triply substituted triphenylmethane dyes crystal violet and Victoria pure blue (albeit a study limited to three solvents): the ratio of absorbance values at the maximal wavelengths designated for the β and R bands (β/R) increases and then decreases as solvent dielectric constant decreases.2 Inverted solvatochromism has also been previously observed for several merocyanine dyes8-10 and a class of bischromophoric cyanine dyes11 using the maximal absorption wavelength as the solventdependent spectral parameter. The sensitivity of the Brilliant Green Gaussian peak parameters to microenvironment may be utilized in various analytical applications. For example, the spectrophotometric determination of the cmc of the anionic surfactant SDS is readily accomplished using either the ratio of peak heights or the ratio of percent area contributions for the 600- and 630-nm Gaussian components of the Brilliant Green absorption spectrum. At SDS surfactant concentrations far below the cmc, the sharp changes in these Gaussian peak parameters (and the appearance of a Gaussian centered at 670 nm) are attributed to the electrostatic interaction of Brilliant Green and SDS to form dye-surfactant ion pairs.1 As SDS concentration is increased further, the ratios of Brilliant Green peak heights and peak areas increase and then become almost constant. It is well understood that, over these SDS concentrations, surfactant monomers first dominate the solution until micelle formation occurs over a very narrow concentration range. Abrupt changes in a variety of solution properties signal micellization, and the cmc is defined as the concentration corresponding to the maximum change in a gradient in the solution property versus concentration curve (i.e., d3φ/dC3 ) 0 at C ) cmc for solution property φ and total surfactant concentration C).12 Thus, the cmc of SDS is determined as the intersection of the two linearly extrapolated lines through the rising portion of ratios (estimated as 4-7 mM SDS) and the essentially constant portion of ratios (estimated as 10-22 mM SDS).12,13 This procedure is illustrated by the dashed lines in Figure 6. For peak height ratios, the intersection of extrapolated lines occurs at 8.4 ( 0.2 mM SDS; for peak area ratios, the intersection suggests a cmc value of 8.1 ( 0.2 mM SDS. These numbers are in excellent agreement with reported SDS cmc values at 25 °C that include 8.1,12 8.2,14 and 8.39 mM.15 (8) Jacques, P. J. Phys. Chem. 1986, 90, 5535-5539. (9) Niedbalska, M.; Gruda, I. Can. J. Chem. 1990, 68, 691-695. (10) Aliaga, C.; Galdames, J. S.; Rezende, M. C.; J. Chem. Soc., Perkin Trans. 2 1997, 1055-1058. (11) Mishra, B. K.; Kuanar, M.; Mishra, A.; Behera, G. B. Bull. Chem. Soc. Jpn. 1996, 69, 2581-2584. (12) Phillips, J. N. Trans. Faraday Soc. 1955 51, 561-569. (13) Evans, D. F.; Wennerstrom, H. The Colloidal Domain: Where Physics, Chemistry, Biology, and Technology Meet; VCH Publishers: New York, 1994; Chapter 1. (14) Aniansson, E. A. G.; Wall, S. N.; Almgren, M.; Hoffmann, H.; Kielmann, I.; Ulbricht, W.; Zana, R.; Lang, J.; Tondre, C. J. Phys. Chem. 1976, 80, 905922.

4216 Analytical Chemistry, Vol. 70, No. 19, October 1, 1998

Further analysis of the data presented in Figure 6 suggests that the variations in peak height and peak area ratios may be interpreted in terms of the expected concentration of free amphiphile in solution as total surfactant concentration is varied. As detailed by Tanford,16 as surfactant concentration is initially increased at levels below the critical micelle concentration, the monomeric surfactant concentration increases linearly (i.e., free monomeric surfactant concentration equals total surfactant concentration). The transition from free amphiphile to the micellar state occurs over a narrow concentration range, but in this transition zone, only a small fraction of surfactant molecules are aggregated as micelles. As the surfactant concentration is increased above the cmc, increasing amounts of amphiphile are present in micellar form with a fairly constant concentration of monomeric amphiphile. Thus, we postulate that the rising and constant segments of the data in Figure 6 reflect the concentration of free surfactant monomer. In particular, the minimum height or area ratio in Figure 6 occurs at the SDS concentration where the maximum number of surfactant-dye ion pairs exist and thus free Brilliant Green and free surfactant are at a minimum. (The steadily decreasing peak height and area ratios at lower SDS concentrations reflect the transition of Brilliant Green molecules from free dye to dye molecules incorporated in surfactant-dye ion pairs.) As increasing amounts of surfactant are added, the Brilliant Green absorption spectrum reflects an increasing number of free surfactant molecules. The nature of the Brilliant Green molecules (i.e., free dye or ion pairs) in this SDS concentration range is not directly ascertainable. However, given that the cmc determination matches other independent measures of SDS cmc in the absence of Brilliant Green, the addition of surfactant to reach the cmc presumably leads to a dissociation of surfactantdye ions pairs to regenerate free Brilliant Green molecules. At and above the cmc, the spectral parameters suggest that the Brilliant Green dye molecules exist in an environment with a relatively constant number of surfactant monomers. The slight differences in the spectral parameters of free Brilliant Green molecules in aqueous solution (i.e., at 0 mM SDS) and in SDS micellar-rich solutions (e.g., at 12-22 mM SDS) likely reflect the different microenvironments for the free dye molecules. Both physical properties (e.g., electrical conductivity and surface tension) and optical properties (e.g., light scattering and spectral changes) of solutions may signal micellization. In some instances, widely disparate cmc values can be obtained by independent methods.17 Furthermore, some variation in cmc value can result from determinations using the solvatochromic behavior of the absorbance wavelength and intensity of a single optical probe.18 While the use of an extrinsic optical probe in analytical determinations should always be critically evaluated for perturbation effects resulting from surfactant-probe interactions,19 (15) Ottewill, R. H. In Surfactants; Tadros, Th. F., Ed.; Academic Press: New York, 1984; pp 1-18. (16) Tanford, C. The Hydrophobic Effect: Formation of Micelles and Biological Membranes, 2nd ed.; John Wiley & Sons: New York, 1980; Chapter 7. (17) Diaz Garcia, M. E.; Sanz-Medel, M. E. Talanta 1986 33, 255-264. (18) Kessler, M. A.; Wolfbeis, O. S. Chem. Phys. Lipids 1989 50, 51-56. (19) Johnson, I. In Molecular Probes: Handbook of Fluorescent Probes and Research Chemicals; Haugland, R. P., Ed.; Molecular Probes, Inc.: Eugene, OR, 1992; pp 1-4. (20) Lide, D. R., Ed. Handbook of Chemistry and Physics, 74th ed.; CRC Press: Boca Raton, FL, 1993; pp 6-148-6-155.

the spectral parameters obtained from the deconvolution of the absorption spectrum of Brilliant Green provide a promising environment-sensitive response for numerous applications. ACKNOWLEDGMENT This research was supported by an award from Research Corp. This research was also supported in part by a grant from the National Science Foundation Research Experiences for Undergraduates Program (CHE-9322804). Acknowledgment is also made

to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for the partial support of this research. K.K.K. acknowledges the Henry Dreyfus TeacherScholar Awards Program for support of this research.

Received for review March 18, 1998. Accepted July 24, 1998. AC980318Y

Analytical Chemistry, Vol. 70, No. 19, October 1, 1998

4217