Adsorption Study of Cationic Dyes Having a Trimethylammonium Anchor Group on Hectorite Using Electrooptic and Spectroscopic Methods

Journal of Colloid and Interface Science 245, 16–23 (2002) doi:10.1006/jcis.2001.7978, available online at http://www.idealibrary.com on

Adsorption Study of Cationic Dyes Having a Trimethylammonium Anchor Group on Hectorite Using Electrooptic and Spectroscopic Methods S. Holzheu1 and H. Hoffmann2 Department of Physical Chemistry I, University of Bayreuth, D-95440 Bayreuth, Germany Received November 28, 2000; accepted September 14, 2001

tion behavior of different dyes on various clay minerals (7–9). Dye adsorption was also applied to characterize clay mineral surfaces (10) and Rytwo et al. used methylene blue and crystal violet for determination of exchangeable cations in montmorillonite (11). Recently Yamaoka and Sasai (12) presented pulsed electric linear dichroism (ELD) as a useful method to obtain additional information about the orientation of the adsorbed dyes. In a detailed theoretical analysis they calculated roll, tilt, and inclination of the dye molecule at the clay mineral surface. Most of the dyes involved in the studies were cationic dyes with the charge delocalized over the whole chromophore. In our present study we have chosen cationic dyes with a trimethylammonium cationic group. Structurally these dyes are more similar to surfactants having a ionic head group and the chromophore as hydrophobic tail. Ionic surfactants are known to form micellelike aggregates on oppositely charged mineral surfaces. These aggregates are usually referred to as hemi- or admicelles (13–15). The aggregation on the charged surfaces takes place at surfactant concentration notedly below the critical micelle concentration. Therefore we expect that cationic dyes with a trimethylammonium cationic group will form hemimicellelike aggregates on the clay mineral surface. The aim of this paper is to clarify the influence of dye loading and molecular structure on the adsorption behavior of dyes on hectorite. We will characterize the adsorption by means of UV-VIS spectroscopy, transient electric birefringence, and ELD.

Clay minerals are natural or synthetic material of colloidal dimensions. Due to the sheetlike structure clay minerals offer a huge specific surface area and hence optimal properties for modification through adsorption. The current work studies the adsorption of five cationic dyes on the synthetic clay mineral hectorite. All dyes have a trimethylammonium anchoring group in common. The adsorbed dye molecules are characterized by means of pulsed electric linear dichroism and UV-VIS spectroscopy. With increasing dye loading a continuous shift in the absorption spectra is observed. But there is no occurrence of a new absorption band. Therefore we conclude that the dyes preferentially adsorb as amorphous aggregates on the clay surface. At low dye loadings the dye molecules lie flat on the clay mineral surface. Increasing dye concentration leads to a continuous increase in average tilt angle. However the orientation of the dye molecules is very sensitive to functional groups. The introduction of a nitro group to a particular dye increases significantly the tendency to lie flat on the surface whereas the introduction of a methoxy group at the same position has the opposite effect. °C 2002 Elsevier Science Key Words: cationic dyes; hectorite; adsorption; transient electric dichroism; spectroscopy.

1. INTRODUCTION

In the view of a colloidal chemist clay minerals are fascinating materials. The diameter of the primary particles is in a range of hundred nanometers whereas the thickness is only one nanometer. This aspect ratio offers an extremely high surface area and hence a lot of interesting possibilities for modifying clay particles through adsorption. A second advantage of clay minerals is their easy availability. There are many natural and synthetic clay minerals offered commercially. Consequently one can find many applications using clay minerals as adsorption medium: e.g., wastewater treatment (1–3) or image fixation (4–6) just to mention the most important. From the scientific point of view special interest was devoted to the adsorption of dyes on clay minerals. From detailed UVVIS spectrum studies it was possible to clarify the aggrega-

2. EXPERIMENTAL

2.1. Materials The synthetic clay mineral hectorite (SKS 21) was obtained as a gift from Clariant (Germany). It has a cec of 0.88 meq/g and was already used in former studies (16). The cationic dyes were provided from Wella (Germany). The chemical structures of the dyes are presented in Fig. 1. All dyes have a trimethylammonium cationic group in common. Direct yellow is a mixture of two isomers. All chemicals were used as received without further purification.

1

E-mail: [email protected]. To whom correspondence should be addressed. E-mail: heinz.hoffmann@ uni-bayreuth.de. 2

0021-9797/02 $35.00

° C 2002 Elsevier Science

All rights reserved.

16

17

ADSORPTION OF CATIONIC DYES

cell. Before each measurement the photocurrent of the reference and the detector are balanced. The change in amplifier output is directly proportional to the relative change of light intensity at the detector. There are two alternative high voltage sources: (a) a highvoltage pulse generator (Cober 606) delivering rectangular pulses up to 2.5 kV and a duration of 3 ms and (b) a capacitor producing exponentially decreasing fields with a onset voltage of maximum 20 kV. To calculate the reduced dichroism 1A/A the ELD data were normalized by the adsorption obtained from UV-VIS spectroscopy. Transient electric birefringence was measured through a selfbuilt instrument. The principle operation of the apparatus was already described earlier (17). The light source of the instrument is a He-Ne laser (λ = 632.8 nm). High-voltage rectangular pulses were produced by the same high-voltage pulse generator (Cober 606) used in the dichroism apparatus. 3. THEORY

FIG. 1. Chemical structure of the dyes studied.

2.2. Sample Preparation Stock solutions of dyes and hectorite were prepared in doubledistilled water. The solutions were diluted and mixed to achieve the final concentrations. The hectorite concentration was kept constant at 0.1 g/L and different amounts of dyes were added to obtain the different mixing ratios. In all samples the dye concentration was below the cation exchange capacity of the hectorite. The samples were ultrasonicated for 1 h and afterward kept at rest for at least 1 day. The pH of the samples was between 9 and 10.

For a better understanding of the experimental results of ELD, we will give a brief introduction into the theoretical background of this measurement technique with a special focus on clay dye systems. A dye molecule can only interact with an incoming light beam when a component of the transition moment of the dye molecule points in the same direction as the electric vector of the plane polarized light. In a three-dimensional space on average only one-third of the dye molecules will have the required orientation. By orienting the dyes in a crystal or an external field one can significantly increase or decrease the percentage pointing in a certain direction. In the case of dye clay systems the orientation force on the clay particle in an electric field is orders of magnitudes larger than the orientation force on the dye molecules. Therefore ELD probes the relative orientation of the dye to the clay particle. Figure 3 shows the two borderline cases: (a) transition moment

2.3. Measurements UV-VIS measurements were performed using a Perkin-Elmer Lambda 19 spectrophotometer and 1-cm quartz cells. During the measurement the cells were thermostated at 25◦ C. ELD measurements were carried out through a modified Tjump apparatus. The schematic representation of the apparatus is shown in Fig. 2. The light source of the apparatus is a high-pressure mercury lamp. The white light passes a monochromator. After the monochromator the beam is split and a fraction routed to a reference photomultiplier. The rest of the light is linearly polarized in front of the sample cell. During measurements the sample cell was kept at constant temperature (25◦ C). A photomultiplier after the sample cell detects the light transmitted by the sample

FIG. 2. Schematic representation of the pulsed electric linear dichroism apparatus.

18

HOLZHEU AND HOFFMANN

particles the orientation function is defined as follows: ¿

À 1 2 8 = (3 cos θ − 1) . 2

[2]

The brackets denote an average over all particles and θ is the angle between the normal vector and the electric field. The definition of the orientation function of the dye relative to the clay particles is equivalent. The average is taken over all molecules and θ is the angle between the normal vector of the clay particle and the transition moment. 1A/A represents the product of the two orientation functions: 1A = 3 · 8dye · 8clay (E). A

FIG. 3. Principles of electric dichroism in clay dye systems: (a) transition dipole moment (represented by bold double arrows) perpendicular to clay mineral surface; (b) transition dipole moment parallel to clay mineral surface.

perpendicular to the clay mineral surface and (b) transition dipole moment parallel to clay mineral surface. In the isotropic case (no field) the clay particle can have three principal orientations. Having the transition moment perpendicular on the surface there are just three clay dye configurations (Fig. 3a, left). The absorption of the sample in every direction would be equal and is typically referred to as the normal absorption A of the sample. In the presence of a strong electric field the clay orients in the direction of the electric field (normal vector perpendicular to the electric field) (18–20) and only two configuration survive (Fig. 3a, right). The absorption parallel to the electric field (Ak ) is now zero as there is no transition moment pointing into this direction. The absorption perpendicular (A⊥ ) is 1.5A (one of two instead of one of three transition moments in the observation direction). The reduced dichroism defined as Ak − A⊥ 1A = A A

[1]

becomes −1.5. When the transition moment is parallel to the surface each clay orientation splits into two clay dye configurations representing the two directions of the surface. In the absence of the field there are two configurations for every direction (Fig. 3b, left). By orienting the clay particles in the electric field the two configurations with the transition moment perpendicular to the electric field disappear (Fig. 3b, right). Ak becomes 1.5A and A⊥ 0.75A. The reduced dichroism is 1A/A is 0.75. Of course this only highlights the extreme cases. In reality neither the orientation of the clay nor that of the dye molecule relative to the clay particle is exactly parallel or perpendicular. Therefore one has to deal with orientation functions. For clay

[3]

8clay is first of all a function of the field strength whereas 8dye gives information about the relative orientation of the transition moment of the dye to the clay mineral. Most dyes do not have only one transition moment but several. In this case the reduced dichroism will be a function of the wavelength λ. The reduced dichroism at a wavelength λ can be calculated as a superimposement of the dichroism of the partial absorption bands i weighted by their absorption Ai (λ), 1A (λ) = 3 · A

P

Ai (λ)8dye,i P · 8clay (E). i Ai (λ)

i

[4]

Typically each partial absorption band is modeled as a Gaussian function in the wavenumber space. Experimentally one has first to deconvolute the measured isotropic absorption into partial absorption bands and second to adjust the relative orientation functions 8dye,i in a way that they reproduce the measured (1A/A) (λ). For more details we refer to Refs. (12, 21, 22). 4. RESULTS AND DISCUSSION

4.1. Signal Form and Field Strength Dependency Typical ELD signals using the capacitor as high-voltage source are presented in Fig. 4. At low field strength the signal increases very fast and shows a continuous decrease with a positive curvature. At high field strength there is a significant part of the decay having a negative curvature indicating the saturation of the signal.

FIG. 4. Typical ELD signals with an exponentially decreasing field.

ADSORPTION OF CATIONIC DYES

FIG. 5. Dependency of 1A of the applied field strength. The inset shows the extrapolation to infinitive field strength using the method proposed by (18).

As discussed under Experimental the measured dichroism is a superimposement of the orientation function (8clay (E)) of the clay particle and the orientation function 8dye of the transition moment of the dye relative to the clay particle. To extract the average angle θ of the dye transition moment relative to the clay normal axis (Eqs. [2] and [3]) it is necessary to know the orientation function of the clay particle. In general 8clay (E) approximates −0.5 for infinitively high field strength. As shown in Fig. 5 1A E→∞ can be extrapolated from 1A(E) plotted against E −2 . The analysis of the data for different dyes and different wavelength revealed that 8clay (E) is independent of dye and wavelength. The average 8clay (E) at a voltage of 18 kV corresponding to a field strength of 13 kV/cm was −0.45 ± 0.01. In order to reduce the number of measurements ELD spectra were determined at a constant field strength of 13 kV/cm.

19

but there is a continuous change in dye molecule environment through increasing adsorption density. In Fig. 6b the dichroism is plotted as a function of the wavelength. This is the primary experimental output of the ELD apparatus. This picture will be omitted for the other dyes. The results show a pattern similar to that of the absorption spectra. By normalizing the results of Fig. 6b with the absorption data of Fig. 6a and taking into account that the orientation function of clay mineral is −0.45 one can calculate the reduced dichroism extrapolated to infinitely high fields (Fig. 6c). A striking feature of the reduced dichroism spectra is the steady decrease with increasing dye loadings. According to Eqs. [2] and [3] one can immediately deduce that the average angle between the transition moments and the normal vector of clay particle becomes smaller when the dye adsorption density rises. A quantitative analysis will be given below. Figures 7 to 10 show UV-VIS and ELD spectra of the dyes direct blue, direct brown, direct yellow, and direct red. The ELD spectra are extrapolated to infinitively high fields. As in the case of straw yellow the UV-VIS spectra of the adsorbed dyes are

4.2. UV-VIS and ELD Spectra When discussing UV-VIS or ELD spectra one must know the exact dye loadings. We did not explicitly carry out adsorption isotheme measurements. However for all dyes studied a totally colorless supernatant was obtained when exceeding the cation exchange capacity (cec: 88 µmol/L for hectorite concentration of 0.1 g/L) of the system. We expect therefore that the dyes adsorb quantitatively in the concentration range investigated. By increasing the mixing ratio between dye and clay we automatically increase the dye loading. Figure 6a shows the absorption spectra of straw yellow at different dye loadings. The spectra of the adsorbed dyes are almost identical to the spectrum of straw yellow in water. With increasing dye concentration the maximum is shifted to smaller wavelengths and the intensity decreases. There is no indication of a distinct isosbestic point nor the appearance of new adsorption bands. We conclude from these observations that straw yellow does not form well-defined aggregates like dimers or trimers

FIG. 6. (a) UV-VIS spectra of straw yellow in the presence of 0.1 g/L hectorite. For comparison the spectrum of the dye in water is also plotted. (b) Corresponding ELD results of the same samples as a function of the wavelength. (c) Normalized ELD spectra.

20

HOLZHEU AND HOFFMANN

FIG. 7. (a) UV-VIS spectra of direct blue in the presence of 0.1 g/L hectorite. For comparison the spectrum of the dye in water is also plotted. (b) Corresponding normalized ELD results extrapolated to infinitively high fields of the same samples as a function of the wavelength.

FIG. 9. (a) UV-VIS spectra of direct yellow in the presence of 0.1 g/L hectorite. For comparison the spectrum of the dye in water is also plotted. (b) Corresponding normalized ELD results extrapolated to infinitively high fields of the same samples as a function of the wavelength.

similar to those of the free dyes. Again neither the appearance of additional adsorption bands nor distinct isosbestic points can be detected. However one observes a steady regression of the main adsorption band in the visible wavelength region. We attribute

this to a continuous change in the environment of the adsorbed dyes. Also the ELD spectra show a decrease with increasing dye loading. The spectra are not constant over the wavelength region

FIG. 8. (a) UV-VIS spectra of direct brown in the presence of 0.1 g/L hectorite. For comparison the spectrum of the dye in water is also plotted. (b) Corresponding normalized ELD results extrapolated to infinitively high fields of the same samples as a function of the wavelength.

FIG. 10. (a) UV-VIS spectra of direct red in the presence of 0.1 g/L hectorite. For comparison the spectrum of the dye in water is also plotted. (b) Corresponding normalized ELD results extrapolated to infinitively high fields of the same samples as a function of the wavelength.

21

ADSORPTION OF CATIONIC DYES

TABLE 1 Average Angle between the Transition Moments of the Partial Absorption Bands and the Normal Vector of the Clay Surface Calculated from Dichroism Measurements

Straw yellow Direct blue Direct brown Direct yellow Direct red Area per a molecule [A2 ]

λ-region

10 µM

20 µM

50 µM

80 µM

300–500 400–680 400–630 400–630 430–580

81(72/90) 72(62/80) 67(60/71) 79(67/89) 57(47/63) 1300

69(67/72) 69(64/72) 61(56/64) 69(64/74) 55(46/60) 660

67(63/71) 65(63/67) 56(52/58) 62(58/65) 51(45/56) 270

64(64/65) 62(61/64) 55(52/56) 59(56/61) 51(46/54) 170

Note. Average is taken over partial adsorption bands with the maximum in the specified λ-region and a extinction coefficient of more than 2000 1 · cm−1 · mol−1 ; values in parentheses are the minimum and maximum value, respectively; area per molecule were calculated through Eq. [5].

FIG. 11. (a) Deconvolution of UV-VIS spectrum of basic yellow (50 µM) in partial adsorption. The numeral in parentheses for each partial band denotes the angle θ between the transition moment and the normal vector of the clay particle. (b) ELD spectra; solid line is the best fit according to Eq. [4].

studied but exhibit a characteristic pattern for each dye. With increasing concentration the patterns stay unchanged but the spectra are shifted to lower values. To carry out a quantitative analysis the UV-VIS spectra (Figs. 6a–10a) are deconvoluted into partial adsorption band. In a second step relative orientation functions 8dye,i (Eq. [4]) are assigned to reproduce the measured ELD spectra. As an example the deconvolution of basic yellow is shown in Fig. 11a. The numbers in parentheses represent the angle θi between the transition moment of the partial adsorption band and the normal vector of the clay particle. The solid line in Fig. 11b is the calculated ELD spectrum according to Eq. [4]. There is an excellent agreement between the measured values and the calculation. We want to call attention to the point that the calculated angles θi must be regarded as average angle in terms of Eq. [2]. The primary result of the calculation is the value of the orientation function 8dye,i . A random oriented transition moment and a transition moment oriented at 55◦ would yield the same value zero of the orientation function. In order to be able to determine the orientation of the dye molecule one has to assign the partial adsorption bands to known transitions derived from molecular orbital calculations. With this knowledge it is possible to calculate exactly roll, tilt, and inclination angle as shown by Yamaoka et al. (12, 21). Unfortunately quantum mechanical calculations are complex and one has to have a large computational facility and an advanced knowledge to interprete the results. For this reason we will carry out a simplified analysis. In dyes having a conjugated π electron system transitions with a high oscillator strength are typically π → π ∗ transitions. Transition

moments of this character are located in the plane of the π electron system. Out of plane transitions of the n → π ∗ character usually exhibit lower oscillator strength (23). Therefore we will confine the interpretation to the most intense partial absorption bands in the visible wavelength region. Assuming that these bands represent π → π ∗ transitions the calculated angles θi give information about the orientation of the dyes’ π electron system relative to the clay particle. Table 1 summarize the results of the five dyes at different concentrations. For simplicity only the average, the minimum, and the maximum values are given. The average is the simple arithmetic mean of the angles θi of the selected partial adsorption bands. As a general trend we observe a decrease of all values (average, minimum, and maximum) with increasing dye loading. Hence at low dye coverage the dye molecules prefer to lie flat on the clay mineral surface whereas with increasing adsorption density more dyes are pointing away from the clay surface. The total angle change in Table 1 seems to be rather small. However one should keep in mind that the average angle calculated according to Eq. [2] of random oriented molecules would be 55◦ . The angles at the highest dye loadings are always close to these 55◦ . Therefore it is probably reasonable to interpret the angle changes of Table 1 more in a sense of a transition to more random aggregates as pictured in Fig. 12 (left) than in a continuous increase in tilt angle as schematically shown on the right side.

FIG. 12. Schematic representation of two possible adsorption structures: (a) random aggregates, (b) highly ordered aggregates with an exactly defined tilt angle.

22

HOLZHEU AND HOFFMANN

The formation of random aggregates is supported by the UVVIS spectra. Highly ordered aggregates should show new absorption bands originating from π-π interactions. This was not observed for the dyes studied. To get a better understanding of the conditions on the clay mineral surface it is useful to have a look at the surface area a per dye molecule. Using the known particle thickness (10 A) and density (2.5 g/cm3 ) we can calculate the crystalograpic surface area with

hydrophobic effect. Additionally one could think about specific interactions of the surface OH groups and the NO2 group. Similar effects were observed studying the cosorption behavior of nitrobenzene and benzene on surfactant-covered silica (24). Nitrobenzene showed a particular affinity to the silica surface. This additional interaction of the dye with the minerals surface could explain the increased tendency of direct yellow to lie flat on the clay surface. 4.3. Electric Birefringence

m V =2· A =2· h ρh

[5]

as 800 m2 per g. With this value and the known dye concentration one obtains the average area per dye molecule. The calculated values are listed in Table 1. At all loadings the surface area per molecule is considerably larger than the typical area of a flat a dye molecule (about 130 A2 ). In principle it would be possible for the dye molecules to lie flat on the clay mineral surface even at the highest concentration. Consequently the decrease in average angle θ is not induced through steric hindrance but through spontaneous aggregation on the clay surface driven by attractive interactions. It is interesting to pay additional attention to the differences between the three dyes direct brown, direct yellow and direct red. In relation to direct brown the only structural difference is an additional NO2 -group in direct yellow and an additional MeO group in direct red, respectively. The absorption spectra are quite similar (Figs. 8a–10a). The NO2 group induces a blue shift of the spectrum and the MeO group a red shift. The patterns of the dichroism spectra are comparable and show the same shifts as the UV-VIS spectra (Figs. 8b –10b). However the absolute values decrease in the order direct yellow, direct brown, direct red. Obviously the introduction of a NO2 group increases significantly the tendency of the molecule to lie flat on the surface whereas the MeO group has the opposite effect. Talking about ionic surfactant adsorption on oppositely charged mineral surfaces, there are two main driving forces governing the adsorption: (a) electrostatic attraction between the ionic head group of the surfactant and the mineral surface and (b) the hydrophobic effect of the alkyl chains. In the case of dyes we have to take at least one additional effect into account: specific interactions of functional groups of the dye with the clay mineral surface like H-bonding or Lewis acid-basic interactions. The electrostatic attraction stays unchanged with the introduction of a NO2 group or MeO group, respectively. Therefore we have to attribute the observed differences to changes in hydrophobic effect or specific interactions. The MeO group is rather hydrophobic. This could lead to an increase in hydrophobic effect and consequently to an increase in aggregation tendency. From steric considerations effective aggregation is only possible when the dyes do not lie flat on the surface. The NO2 group is more hydrophilic. This probably weakens the

Another possibility to get information about the orientation of the dye molecules is electric birefringence measurements with a laser wavelength outside the absorption bands of the dyes. Our electric birefringence apparatus is equipped with a He-Ne laser (632.8 nm). This allows reasonable measurements for direct red and straw yellow. Figure 13a shows the stationary birefringence of hectorite with increasing concentrations of these dyes. The field strength is 5 kV/cm. This is close to saturation. Therefore the changes in 1n stat can be directly attributed to changes in the optical anisotropy factor 1g of the particle. Low dye loadings lead to an initial increase of 1g. With increasing dye loadings 1g passes through a maximum and decreases again. The measured anisotropy factor can be regarded as a superimposement of the intrinsic 1g of the clay particle and the 1g of the dye layer on the hectorite surface. By subtraction of the stationary birefringence of the pure hectorite and normalization

FIG. 13. (a) Stationary birefringence of dye covered hectorite (0.1 g/L) as a function of the dye concentration; E = 5 kV/cm. (b) Normalized stationary birefringence representing the specific optical anisotropy of the dye layer.

23

ADSORPTION OF CATIONIC DYES

with the dye concentration (Fig. 13b) one obtains a quantity representing the specific optical anisotropy of the dye layer. It is reasonable to assume that the optical polarizability of a dye molecule is larger in plane of the π electron system than perpendicular to the π electron system. Therefore a positive specific optical anisotropy of Fig. 13b represents a flat dye orientation whereas a negative can be attributed to dye molecules pointing away from the clay mineral surface. Hence the results of the birefringence measurements coincide well with ELD measurements and support the interpretation given above. Direct red showing a sign inversion of the dichroism also exhibits a sign inversion in specific optical anisotropy of the dye layer whereas straw yellow having a positive dichroism in the whole concentration range only shows a decrease. 5. CONCLUSION

The adsorption behavior of five different cationic dyes on hectorite has been studied by means of UV-VIS spectroscopy, electric linear dichroism, and electric birefringence. The common property of the dyes is a trimethylammonium cationic group. ELD measurements give interesting information about the orientation of the dyes on the clay mineral surface which are difficult to obtain by UV-VIS spectrum measurements alone. The results revealed that the dye orientation strongly depends on the dye loading. At low concentrations the dyes tend to lie flat on the surface whereas with increasing concentration the dichroism decreases. This was attributed to the formation of more random aggregates. The dye aggregation behavior is in principle comparable to surfactant aggregation. However there is a strong influence of functional groups. The introduction of a MeO group seems to strengthen the hydrophobic effect and the tendency to form aggregates. A NO2 group has the opposite effect. Probably there are additional specific interactions between the functional groups of the dye and the functional groups of the mineral surface. ACKNOWLEDGMENT We thank the Deutsche Forschungsgesellschaft (DFG) for the financial support of this work.

REFERENCES 1. Ramakrishna, K. R., and Viraraghavan, T., Water Sci. Technol. 36, 189 (1997). 2. Lambert, S. D., Graham, N. J. D., Sollars, C. J., and Fowler, G. D., Water Sci. Technol. 36, 173 (1997). 3. Juang, R. S., Wu, F. C., and Tseng, R. L., Environ. Technol. 18, 525 (1997). 4. Ito, K., Kuwabara, M., Fukunishi, K., and Fujiwara, Y., Dyes Pigm. 34, 297 (1997). 5. Dominguez, J. M., and Eje Central Cardenas, L., Proc. SPIE-Int. Soc. Opt. Eng. 3227, 88 (1997). 6. Ito, K., Kuwabara, M., Fukunishi, K., and Fujiwara, Y., J. Imaging Sci. Technol. 40, 275–280 (1996). 7. Garfinkel-Shweky, D., and Yariv, S., Clay Miner. 34, 459 (1999). 8. Lopez Arbeloa, F., Herran Martinez, J. M., Lopez Arbeloa, T., and Lopez Arbeloa, I., Langmuir 14, 4566 (1998). 9. Cione, Ana P. P., Miguel Neumann, G., and Gessner, F., J. Colloid Interface Sci. 198, 106 (1998). 10. Lopez Arbeloa, F., Lopez Arbeloa, T., and Lopez Arbeloa, I., Trends Chem. Phys. 4, 191 (1996). 11. Rytwo, G., Serban, C., Nir, S., and Margulies, L., Clays Clay Miner. 39, 551 (1991). 12. Yamaoka, K., and Sasai, R., J. Colloid Interface Sci. 225, 82 (2000). 13. Fuerstenau, D. W., J. Phys. Chem. 60, 981 (1956). 14. Harwell, J. H., Hoskins, J. C., Schlechter, R. S., and Wade, W. H., Langmuir 1, 251 (1985). 15. Koopal, L. K., in “Coagulation and Flocculation” (Dobias, Ed.), p. 102, Surfactant science Series, Vol. 47. Dekker, New York, 1993. 16. Holzheu, S., and Hoffmann, H., Progr. Colloid Polym. Sci. 115, 265 (2000). 17. Schorr, W., and Hoffmann, H., J. Phys. Chem. 85, 3160 (1981). 18. Sasai, R., and Yamaoka, K., J. Phys. Chem. 99, 17754 (1995). 19. Yamaoka, K., and Sasai, R., J. Colloid Interface Sci. 209, 408 (1999). 20. Peikov, V., Sasai, R., Stoylov, S. P., and Yamaoka, K., J. Colloid Interface Sci. 197, 78 (1998). 21. Yamaoka, K., Sasai, R., and Takata, N., Colloids Surf. A 175, 23 (2000). 22. Sasai, R., Shin’ya, N., Shichi, T., Takagi, K., and Gekko, K., Langmuir 1999 15, 413 (1999). 23. Holmen, A., Broo, A., Albinsson, B., and Norden, B., J. Am. Chem. Soc. 119, 12240 (1997). 24. Holzheu, S., Behrends, T., and Herrmann, R., in “Surtactant-Based Separations, Science and Technology” (J. F. Scamehorn and J. H. Harwell, Eds.), p. 314, ACS Symposium Series 740. Am. Chem. Soc., Washington, DC, 1999.