- Email: [email protected]
Adsorption of Dyes on Nanosize Modified Silica Particles
JOURNAL OF COLLOID AND INTERFACE SCIENCE ARTICLE NO.
195, 222–228 (1997)
CS975156
Adsorption of Dyes on Nanosize Modified Silica Particles 1 Guangwei Wu, Athanasia Koliadima, Yie-Shein Her, and Egon Matijevic´ 2 Center for Advanced Materials Processing, Clarkson University, Box 5814, Potsdam, New York 13699-5814 Received June 17, 1997; accepted August 28, 1997
The adsorption of several anionic dyes on nanosize aluminamodified silica particles of different compositions and modal sizes has been studied. These silica cores have the same surface properties as alumina dispersed in aqueous solutions. The negatively charged dyes are electrostatically attracted to positively charged cores and chemisorbed by forming a surface Al lake. The application of so obtained pigments in the preparation of color films and their optical characteristics are described. q 1997 Academic Press Key Words: adsorption of dyes; color films; nanosize pigments; silica/dye pigments.
INTRODUCTION
The interactions between dyes and solid surfaces may involve covalent bond formation or physical forces ( electrostatic and / or van der Waals ) . For example, reactive dyes were grafted to the surface of derivatized silicas to prepare water-dispersible pigments for ink-jets ( 1 ) . A number of studies ( 2 – 7 ) dealt with the adsorption of dyes on charged solids. Thus, McKay et al. ( 2 ) investigated the interactions of dyestuffs with chitin. Because of the porosity of the latter, it was difficult to establish the adsorption mechanism; over a limited dye concentration range, both Langmuir and Freundlich isotherms could fit the data. When alumina was used as adsorbent ( 3 – 5 ) , the uptake of dyes was strongly dependent on the particle morphology, method of preparation, pretreatment of the solids, and the equilibrium pH conditions. This work investigates the interactions of several anionic dyes with alumina-modified silica particles of different sizes and compositions. The advantage of this adsorbent is in its small size ( õ20 nm) and the ability of the dyes to form chemical bonds with surface GAlOH groups of the core particles. The ultimate objective of the study is to produce well-defined nanosize pigments for special applications, such as for ink-jets and color filters for flat panel displays. Previous studies (8–11) carried out in this laboratory 1 2
Supported by the U.S. Display Consortium. To whom correspondence should be addressed.
JCIS 5156
/
6g35$$$281
EXPERIMENTAL
Materials The dyes used in this study are listed in Table 1. The D& C Red #6 dye (Sun Chemical Co.) with purity of ú95% was used as received, while Naphthol Yellow S (Aldrich, Acid Yellow #1, 75% pure) and Acid Blue #25 (Aldrich, 45% pure) were recrystallized twice in water. The Guinea Green B (Aldrich, 50% pure) was used as received. To prepare saturated solutions, the dye powders were dissolved in hot water ( Ç857C), immediately filtered through 0.8 mm pore-size membranes, and then cooled to room temperature. After about 1 day, the solutions were passed through 0.2 mm filters to remove any undissolved dye. Ludox CL particles, supplied by DuPont, were obtained by first producing nanosized silica and then coating it with alumina (12). The characteristics of different samples used in this study are listed in Table 2. The specific surface areas were determined by the titration method (13) before alumina was applied. The number following CL is the equivalent silica particle diameter without the alumina shell, calculated from the specific surface area, assuming smooth spheres, and the numbers in the parentheses give the Si/Al molar ratio. Adsorption Isotherms A fixed amount (0.5 cm3 ) of a Ludox CL sample was added into 9.5 cm3 dye solutions of different concentrations. The samples were equilibrated for several hours in an ultrasonic bath and then separated by centrifugation at 30,000 rpm for 15 min. The dye concentration in the supernatant solution was determined spectrophotometrically. For some dyes the visible absorption spectrum was affected by the pH of the solution, as shown in Fig. 1 for the D&C Red #6. Since little change in absorption is observed at pH ú7, the concentration of the dye in aqueous solutions was deter-
222
0021-9797/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
showed that adsorbing dyes on solid surfaces or incorporating them into inorganic particles can yield reproducible pigments of superior optical and stability properties.
11-04-97 05:52:10
coidas
DYE ABSORPTION ON Al-MODIFIED Si PARTICLES
TABLE 1 Organic Dyes Used in This Work Name
Formula COONa
HO D & C Red 6
CH‹©
©N®N© SO‹Na ONa
Acid Yellow 1
NaO‹S
NO¤
NO¤ O
NH¤
Acid Blue 25
223
LSA 440 (Coulter Electronics) instrument. The data for Ludox CL-5 (3.5) in aqueous solution containing 1.2 wt% solids, as a function of the pH, yield the isoelectric points (iep) of Ç9.2 (Fig. 4). Since the iep of silica and alumina are very different, i.e. at pH Ç2.0 and 9.3, respectively (15, 16), it can be assumed that the surface of these particles consists essentially of hydroxylated alumina. The mobility of other Ludox CL samples at pH Ç4.0 is the same within the experimental error ( {10%) (Table 2), indicating the identical nature of particle surfaces of all samples. Thermogravimetric analysis (TGA) was carried out over the temperature range of 30–10007C at a scan rate of 257C min 01 in air. Based on the obtained curves for Ludox samples and pigments, the weight loss below 3507C was attributed to the evaporation of the solvent used in the preparation of the pigment.
SO‹Na
Pigment and Thin Film Preparation O
NH© SO‹2
CH¤CH‹ 1
Guinea Green B
N©CH¤©
C
CH¤CH‹
SO‹Na
NCH¤©
mined at pH Ç7. The amount of the dye adsorbed was then calculated from the difference of the original concentration and that found in the solution after equilibration with silica.
The well-defined nanosized pigments were prepared by dropwise addition of modified silica into a concentrated dye solution ( Ç0.02 mol dm03 ) under stirring at room temperature, resulting in an immediate interaction between the cores and the solutes. The pigment thus obtained was separated by centrifugation and washed twice with distilled water to remove physically adsorbed and/or trapped dye. To have a common solvent (propylene glycol methyl ether acetate, PMA) with the dispersing medium (polymer), the pigment slurry was washed several times with ethanol and finally with PMA. The pigment slurry was then dispersed into an acrylic copolymer (CFP-18, supplied by Shipley Company) solution under stirring and/or ultrasonication. The thin film was prepared by spin-coating the pigment–polymer dispersion on a glass substrate at controlled speed, followed by baking at 807C for 5 min to remove the solvent.
Characterization The particle size was evaluated by scanning electron microscopy and/or dynamic light scattering. The scanning electron micrographs (Fig. 2) of Ludox CL-12 before and after adsorption of the D&C Red #6 dye show that the pigment remains in the nanosize range. To determine the hydrodynamic radii of Ludox CL samples by dynamic light scattering, all samples were diluted with water to Ç1.2 wt% concentration to minimize any interparticle interaction, and the pH was adjusted to Ç4.0. Figure 3 shows a typical size distribution curve for Ludox CL-5 (3.5) in terms of relative weight percentages, assuming that the particles behave as hard spheres, as analyzed using the CONTIN program (14). The obtained hydrodynamic radii, Rh , are listed in Table 2. The electrophoretic mobilities were measured with a DE-
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
FIG. 1. The spectra of the D&C #6 Red dye solutions of different pH values. The dye concentration in all samples was 5.44 1 10 05 mol dm03 .
coidas
224
WU ET AL.
TABLE 2 Specifications of Ludox Samples
A B C D E
Ludox
Specific surface areaa (m2/g)
CL-5 (3.3) CL-5 (3.5) CL-5 (4.4) CL-7 CL-12
530 530 530 380 230
pH
Si/Al molar ratio
Solid content (wt%)
Rhb (nm)
Mobility at pH 4.0 (mm/s/V/cm)
Max. adsorption of D&C Red #6 [g(dye)/g(Ludox)]
3.0 3.0 3.0 4.0 4.0
3.3 3.5 4.4 — 5.1c
Ç12.5 Ç12.5 Ç12.5 Ç27 Ç33
6.5 7.4 8.7 8.4 10.0
2.7 2.6 2.4 2.9 2.8
0.30 0.29 0.18 0.19 0.12
a
Determined by titration before coating with alumina. Hydrodynamic radius, determined by light scattering. c Calculated from the composition provided with the sample. b
RESULTS AND DISCUSSION
Effect of the Adsorbent The effect of the adsorbent on the uptake of dyes was investigated by using Ludox CL samples of different sizes and compositions and the D&C Red #6. Figure 5 shows that all isotherms are of the Langmuir type. Because the electrokinetic behavior of these adsorbents indicated the same surface composition, the difference should be ascribed to their particle size. The saturation adsorption per unit weight (Table 2) decreases with increasing particle size of different adsorbents, due to the smaller specific surface areas. The equilibrium concentrations of the dye (Ceq ) is close to zero at their low initial concentrations (C0 ) and then increases linearly above a critical value C *0 (Fig. 6), which is characteristic of solutes that are chemisorbed. Indeed, the supernatant solutions are colorless below C *0 , indicating that essentially all of the dye is adsorbed by Ludox CL particles. As expected, the critical value decreases with increasing particle size when the same amount of different adsorbents is used. Effect of Adsorbates The adsorption isotherms for different dyes on Ludox CL5 (3.5) show the same trend, except the plateau values vary
(Fig. 7). The maximum adsorbed amounts obtained from these isotherms are listed in Table 3. Once again, in all cases the plot of Ceq versus C0 is characteristic of chemisorption (Fig. 8), which is not surprising because all of these dyes contain sulfonic groups. To verify these data, thermogravimetric analysis was used to determine the amount of the adsorbed dye, and a typical result with the red pigment is illustrated in Fig. 9. Since an organic stabilizer was present in Ludox CL, its weight content had to be determined in order to calculate the actual amount of the dye. For this purpose, two approaches were employed to prepare Ludox CL samples for TGA experiments. In the first, the Ludox CL sample was dried under vacuum to retain the stabilizer in the solid, resulting in a total weight loss of 10% on heating from 350 to 9007C. In the second approach, particles were aggregated and settled on addition of K2SO4 and then washed three times with distilled water. In the latter case, the weight loss was Ç3% because part of the stabilizer was leached out on washing. Thus, the amount of the stabilizer in these pigments should be between the two extreme values. The weight loss on heating the red and yellow pigments, evaluated by the TGA, was found to be Ç6% larger than calculated for the adsorbed amount of dye. The difference was most likely due to the stabilizer in Ludox, although some other minor additives could have been present. It should pointed out that
TABLE 3 Data for Adsorption of Dyes on Ludox CL-5 at 25 { 17C Dye
D&C Red #6
Acid Yellow #1
Acid Blue
Guinea Green B
Molecular weight (M) Dye/Ludox [Gmax, g/g (adsorption)] Dye/Ludox [Gmax , g/g (TGA)] Gm* ax (1004 mol/g*)a Area (nm2/molecule)b
430 0.30 0.36 7.0 1.1
358 0.19 0.24 5.0 1.5
416 0.70 0.63 13.7 0.6
691 0.60 0.61 8.0 1.0
a The Gm* ax used here is calculated from the TGA results after correcting for the contribution from the organic stabilizer in the Ludox CL sample (Å Gmax(TGA) 0 0.06). b The specific surface area used is derived from the titration.
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
coidas
DYE ABSORPTION ON Al-MODIFIED Si PARTICLES
225
FIG. 2. Scanning electron micrographs (SEM) of Ludox CL-12 particles (A) and after adsorbing D&C Red #6 (B).
such experiments with blue and green pigments are less reliable, because of the uncertainty with respect to the purity of the used dyes. To make results of different pigments comparable, data are listed in moles of dye per unit weight of adsorbent (Table 3). The chemisorption must be due to the interaction of sulfonic ( –SO 3– ) and/or carboxylic ( –COO – ) groups of
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
dyes with GAlOH sites on the surface of modified silica particles. The chemisorption mechanism is also supported by the fact that there is no observable leaching in water, ethanol, or the organic solvent (PMA) over a long experimentation times. The adsorption behavior of alumina-coated silica used in this work differs from that of acid-pretreated alumina (4, 5); the
coidas
226
WU ET AL.
FIG. 3. A typical size distribution of the Ludox CL-5 (3.5) (Table 2) sample, determined by the dynamic light scattering and analyzing the correlation function with the CONTIN program.
latter shows a reversible uptake of the dye, in strong dependence on the pH over the range of 1–9 and on the acid used to pretreat the adsorbent. Coadsorption of Yellow and Green Dyes To modify the color properties of the pigment, coadsorption of Acid Yellow #1 and Guinea Green B was studied by determining adsorbed amounts on Ludox CL from solutions containing these dyes in different molar ratios. For this purpose, the content of dyes in the pigments was calculated from the TGA data, which was possible because the weight loss for yellow and green pigments is very different, i.e., 0.24 and 0.61 g / g, respectively. The total dye concentration used ( Ç0.02 mol dm03 ) in these experiments exceeded the saturation level. Assuming that each dye in a mixture follows the same adsorption pattern on silica as when present alone in the
FIG. 4. The electrophoretic mobility as a function of the pH of Ludox CL-5 (3.5) sample at 25.07C.
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
FIG. 5. Adsorption isotherms of the D&C Red #6 dye for five different Ludox CL samples (Table 2) at 25 { 17C.
solution, the molar ratio in the solid should equal that in the solution. Consequently, the weight loss of pigments prepared with different mixtures of dyes can be calculated from the linear combination of the results observed with individual pigments. Figure 10 gives such data for the green and yellow dyes, which implies that there is no competition between the two for the Ludox adsorbent. Conformation of Dye Molecules on the Surface of Modified Silica Particles The conformation of the adsorbed dyes can be estimated from the surface area occupied by each molecule. D&C Red #6 is used as an example because of its purity. One difficulty in such calculation is the evaluation of the correct specific surface area of the adsorbent. The conformation of dye mole-
FIG. 6. The equilibrium concentration of D&C Red #6 dye ( Ceq ) as a function of its initial concentration (C0 ) for adsorbents listed in Table 2.
coidas
DYE ABSORPTION ON Al-MODIFIED Si PARTICLES
FIG. 7. Adsorption isotherms of D&C Red #6, Acid Yellow #1, Acid Blue 25, and Guinea Green B on Ludox CL-5 (3.5) at 25 { 17C.
cules was usually studied at the liquid/air interface to avoid the complexity of this problem (17). The area occupied by dye molecules calculated from adsorption on porous silica is much larger than that calculated from a close packed monolayer (18). However, the porosity effect should be negligible with nanosize modified silica particles used in this work. Two cases are considered here. In the first, it was assumed that the entire surface area of Ç210 m2 / g, as determined by titrating Ludox CL-12, was accessible to dye molecules, which yielded an area per dye molecule, A , of Ç1.3 nm2 . In the second case, the hydrodynamic radius from dynamic light scattering was used to calculate the specific area by assuming particles to be smooth. When the density of the solids is taken as Ç2.26 g / cm3 , the area occupied by each dye molecule is estimated to be 0.8 nm2 .
FIG. 8. The equilibrium concentration of dyes in the supernatant solution as a function of the initial concentration for the systems shown in Fig. 7.
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
227
FIG. 9. Typical thermogravimetric data for Ludox CL-5 (3.5) before and after adsorbing D&C Red #6 at the scanning rate of 257C min 01 . The weight loss below 3507C (arrow) is attributed to the removal of solvents. ‘‘Ludox/dry’’ refers to the sample dried in vacuum, and ‘‘Ludox/wash’’ was prepared by coagulating Ludox CL-5 (3.5) with potassium sulfate and then washing with water.
The projected area of the D&C Red #6 molecule is Ç1.3 nm2 , implying that the adsorbed dyes lie entirely flat or are slightly tilted. Optical Properties of Nanosize Pigments By adsorbing anionic dyes on alumina-modified silica particles well-defined pigments can be prepared, which is a more effective approach than grafting the dye to the surface of derivatized silica (1, 9). In addition, the described procedure yields uniform and reproducible pigments in contrast to the traditional method, which is based on reducing the
FIG. 10. The content of coadsorbed Guinea Green B and Acid Yellow #1 on Ludox CL-5 (3.3) as a function of the molar ratios of these dyes in the solution. The line represents the linear combination of green and yellow pigments of the same molar ratios.
coidas
228
WU ET AL.
particle size by grinding. Since the pigments retain essentially the same size and shape as the carrier particles (Fig. 2), their optical properties can be controlled by selecting the core materials. If the latter is in the nanosize range, as used in this work, it is expected that color films prepared with such pigments would have high transmittance at the desired wavelengths. Figure 11 shows the transmittances of red, blue, and green thin films containing 60 wt% solid pigments. The red and blue pigments were prepared by adsorbing D&C Red #6 and Acid Blue #25 on Ludox CL-5, respectively. The green pigment was obtained by using solutions containing both Acid Yellow #1 and Guinea Green B dyes at a weight ratio of 2:1 with the same silica. The transmittance is excellent, especially for the red thin film which reaches Ç100% even with a film thickness of 8 mm. The chromaticities of thin color films, determined by evaluating the transmitted spectra by means of commercial ( PECOL ) software, based on the CIE ( International Commission on Illumination ) system, are displayed in Fig. 12. The thin color films ( a few micrometers thick ) , prepared by using pigments described in this work, show characteristics to be very close to the NTSC ( National Television Standards Committee ) values ( red, x Å 0.64, y Å 0.33; green, x Å 0.29, y Å 0.60; blue, x Å 0.15, y Å 0.06 ) . The differences become smaller with increasing film thickness or with higher pigment contents in the film of the same thickness. However, the mechanical proper-
FIG. 11. The transmittance spectra of thin color films prepared by dispersing the red (D&C Red #6), blue (Acid Blue 25), and green (Acid Yellow #1 and Guinea Green B mixture in a weight ratio of 2:1) pigments in an acrylic copolymer at a solid loading of 60%. The film thickness is Ç2, Ç4, and Ç2 mm for red, blue, and green, respectively.
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
FIG. 12. Chromaticity diagram based on the CIE (International Commission on Illumination) system. The solid squares represent the reference chromaticities for red, green, blue, and the CIE light source, respectively. The open triangles are the chromaticities of thin color films displayed in Fig. 11.
ties of such films deteriorate after the solid loading exceeds an upper limit of Ç70 wt%. REFERENCES 1. Winnik, F. M., Keoshkerian, B., Fuller, J. R., and Hofstra, P. G., Dyes Pigm. 14, 101 (1990). 2. McKay, G., Blair, H. S., and Gardner, J. R., J. Appl. Polym. Sci. 27, 3043 (1982). 3. Fleming, H. L., in ‘‘Fundamentals of Adsorption’’ (A. I. Liapis, Ed.), pp. 221. Engineering Foundation, New York, 1987. 4. Jain, J. K., Mundhara, G. L., and Tiwari, J. S., Colloid Surf. 29, 373 (1988). 5. Mishra, R. K., Mundhara, G. L., and Tiwari, J. S., J. Colloid Interface Sci. 129, 41 (1984). 6. Charreyre, M.-T., Zhang, P., Winnik, M. A., Pichot, C., and Graillat, C., J. Colloid Interface Sci. 170, 374 (1995). 7. Jan, D. E., and Raghavan, S., Colloid Surf. 92, 1 (1994). 8. Giesche, H., and Matijevic´, E., Dyes Pigm. 17, 323 (1991). 9. Hsu, W. P., Yu, R., and Matijevic´, E., Dyes Pigm. 19, 179 (1992). 10. Carotenuto, G., Her, Y.-S., and Matijevic´, E., Ind. Eng. Chem. Res. 35, 2929 (1996). 11. Tentorio, A., Matijevic´, E., and Kratohvil, J. P., J. Colloid Interface Sci. 77, 418 (1980). 12. DuPont Specialty Chemicals, ‘‘Ludoxt Colloidal Silica: Properties, Uses, Storage, and Handling.’’ Wilmington, DE. 13. Iler, R., ‘‘The Chemistry of Silica, Solubility, Polymerization, Colloid and Surface Properties, and Biochemistry.’’ Wiley, New York, 1979. 14. Provench, S. W., Comput. Phys. Commun. 27, 215 (1982). 15. Parks, G. A., and Bruyn, P. L., J. Phys. Chem. 66, 967 (1962). 16. Brace, R., and Matijevic´, E., J. Inorg. Nucl. Chem. 35, 3691 (1973). 17. Schmitt, F.-J., Meller, P., Ringsdorf, H., and Knoll, W., Prog. Colloid Polym. Sci. 83, 136 (1990). 18. Parida, S. K., and Mishra, B. K., J. Colloid Interface Sci. 182, 473 (1996).
coidas
195, 222–228 (1997)
CS975156
Adsorption of Dyes on Nanosize Modified Silica Particles 1 Guangwei Wu, Athanasia Koliadima, Yie-Shein Her, and Egon Matijevic´ 2 Center for Advanced Materials Processing, Clarkson University, Box 5814, Potsdam, New York 13699-5814 Received June 17, 1997; accepted August 28, 1997
The adsorption of several anionic dyes on nanosize aluminamodified silica particles of different compositions and modal sizes has been studied. These silica cores have the same surface properties as alumina dispersed in aqueous solutions. The negatively charged dyes are electrostatically attracted to positively charged cores and chemisorbed by forming a surface Al lake. The application of so obtained pigments in the preparation of color films and their optical characteristics are described. q 1997 Academic Press Key Words: adsorption of dyes; color films; nanosize pigments; silica/dye pigments.
INTRODUCTION
The interactions between dyes and solid surfaces may involve covalent bond formation or physical forces ( electrostatic and / or van der Waals ) . For example, reactive dyes were grafted to the surface of derivatized silicas to prepare water-dispersible pigments for ink-jets ( 1 ) . A number of studies ( 2 – 7 ) dealt with the adsorption of dyes on charged solids. Thus, McKay et al. ( 2 ) investigated the interactions of dyestuffs with chitin. Because of the porosity of the latter, it was difficult to establish the adsorption mechanism; over a limited dye concentration range, both Langmuir and Freundlich isotherms could fit the data. When alumina was used as adsorbent ( 3 – 5 ) , the uptake of dyes was strongly dependent on the particle morphology, method of preparation, pretreatment of the solids, and the equilibrium pH conditions. This work investigates the interactions of several anionic dyes with alumina-modified silica particles of different sizes and compositions. The advantage of this adsorbent is in its small size ( õ20 nm) and the ability of the dyes to form chemical bonds with surface GAlOH groups of the core particles. The ultimate objective of the study is to produce well-defined nanosize pigments for special applications, such as for ink-jets and color filters for flat panel displays. Previous studies (8–11) carried out in this laboratory 1 2
Supported by the U.S. Display Consortium. To whom correspondence should be addressed.
JCIS 5156
/
6g35$$$281
EXPERIMENTAL
Materials The dyes used in this study are listed in Table 1. The D& C Red #6 dye (Sun Chemical Co.) with purity of ú95% was used as received, while Naphthol Yellow S (Aldrich, Acid Yellow #1, 75% pure) and Acid Blue #25 (Aldrich, 45% pure) were recrystallized twice in water. The Guinea Green B (Aldrich, 50% pure) was used as received. To prepare saturated solutions, the dye powders were dissolved in hot water ( Ç857C), immediately filtered through 0.8 mm pore-size membranes, and then cooled to room temperature. After about 1 day, the solutions were passed through 0.2 mm filters to remove any undissolved dye. Ludox CL particles, supplied by DuPont, were obtained by first producing nanosized silica and then coating it with alumina (12). The characteristics of different samples used in this study are listed in Table 2. The specific surface areas were determined by the titration method (13) before alumina was applied. The number following CL is the equivalent silica particle diameter without the alumina shell, calculated from the specific surface area, assuming smooth spheres, and the numbers in the parentheses give the Si/Al molar ratio. Adsorption Isotherms A fixed amount (0.5 cm3 ) of a Ludox CL sample was added into 9.5 cm3 dye solutions of different concentrations. The samples were equilibrated for several hours in an ultrasonic bath and then separated by centrifugation at 30,000 rpm for 15 min. The dye concentration in the supernatant solution was determined spectrophotometrically. For some dyes the visible absorption spectrum was affected by the pH of the solution, as shown in Fig. 1 for the D&C Red #6. Since little change in absorption is observed at pH ú7, the concentration of the dye in aqueous solutions was deter-
222
0021-9797/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
showed that adsorbing dyes on solid surfaces or incorporating them into inorganic particles can yield reproducible pigments of superior optical and stability properties.
11-04-97 05:52:10
coidas
DYE ABSORPTION ON Al-MODIFIED Si PARTICLES
TABLE 1 Organic Dyes Used in This Work Name
Formula COONa
HO D & C Red 6
CH‹©
©N®N© SO‹Na ONa
Acid Yellow 1
NaO‹S
NO¤
NO¤ O
NH¤
Acid Blue 25
223
LSA 440 (Coulter Electronics) instrument. The data for Ludox CL-5 (3.5) in aqueous solution containing 1.2 wt% solids, as a function of the pH, yield the isoelectric points (iep) of Ç9.2 (Fig. 4). Since the iep of silica and alumina are very different, i.e. at pH Ç2.0 and 9.3, respectively (15, 16), it can be assumed that the surface of these particles consists essentially of hydroxylated alumina. The mobility of other Ludox CL samples at pH Ç4.0 is the same within the experimental error ( {10%) (Table 2), indicating the identical nature of particle surfaces of all samples. Thermogravimetric analysis (TGA) was carried out over the temperature range of 30–10007C at a scan rate of 257C min 01 in air. Based on the obtained curves for Ludox samples and pigments, the weight loss below 3507C was attributed to the evaporation of the solvent used in the preparation of the pigment.
SO‹Na
Pigment and Thin Film Preparation O
NH© SO‹2
CH¤CH‹ 1
Guinea Green B
N©CH¤©
C
CH¤CH‹
SO‹Na
NCH¤©
mined at pH Ç7. The amount of the dye adsorbed was then calculated from the difference of the original concentration and that found in the solution after equilibration with silica.
The well-defined nanosized pigments were prepared by dropwise addition of modified silica into a concentrated dye solution ( Ç0.02 mol dm03 ) under stirring at room temperature, resulting in an immediate interaction between the cores and the solutes. The pigment thus obtained was separated by centrifugation and washed twice with distilled water to remove physically adsorbed and/or trapped dye. To have a common solvent (propylene glycol methyl ether acetate, PMA) with the dispersing medium (polymer), the pigment slurry was washed several times with ethanol and finally with PMA. The pigment slurry was then dispersed into an acrylic copolymer (CFP-18, supplied by Shipley Company) solution under stirring and/or ultrasonication. The thin film was prepared by spin-coating the pigment–polymer dispersion on a glass substrate at controlled speed, followed by baking at 807C for 5 min to remove the solvent.
Characterization The particle size was evaluated by scanning electron microscopy and/or dynamic light scattering. The scanning electron micrographs (Fig. 2) of Ludox CL-12 before and after adsorption of the D&C Red #6 dye show that the pigment remains in the nanosize range. To determine the hydrodynamic radii of Ludox CL samples by dynamic light scattering, all samples were diluted with water to Ç1.2 wt% concentration to minimize any interparticle interaction, and the pH was adjusted to Ç4.0. Figure 3 shows a typical size distribution curve for Ludox CL-5 (3.5) in terms of relative weight percentages, assuming that the particles behave as hard spheres, as analyzed using the CONTIN program (14). The obtained hydrodynamic radii, Rh , are listed in Table 2. The electrophoretic mobilities were measured with a DE-
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
FIG. 1. The spectra of the D&C #6 Red dye solutions of different pH values. The dye concentration in all samples was 5.44 1 10 05 mol dm03 .
coidas
224
WU ET AL.
TABLE 2 Specifications of Ludox Samples
A B C D E
Ludox
Specific surface areaa (m2/g)
CL-5 (3.3) CL-5 (3.5) CL-5 (4.4) CL-7 CL-12
530 530 530 380 230
pH
Si/Al molar ratio
Solid content (wt%)
Rhb (nm)
Mobility at pH 4.0 (mm/s/V/cm)
Max. adsorption of D&C Red #6 [g(dye)/g(Ludox)]
3.0 3.0 3.0 4.0 4.0
3.3 3.5 4.4 — 5.1c
Ç12.5 Ç12.5 Ç12.5 Ç27 Ç33
6.5 7.4 8.7 8.4 10.0
2.7 2.6 2.4 2.9 2.8
0.30 0.29 0.18 0.19 0.12
a
Determined by titration before coating with alumina. Hydrodynamic radius, determined by light scattering. c Calculated from the composition provided with the sample. b
RESULTS AND DISCUSSION
Effect of the Adsorbent The effect of the adsorbent on the uptake of dyes was investigated by using Ludox CL samples of different sizes and compositions and the D&C Red #6. Figure 5 shows that all isotherms are of the Langmuir type. Because the electrokinetic behavior of these adsorbents indicated the same surface composition, the difference should be ascribed to their particle size. The saturation adsorption per unit weight (Table 2) decreases with increasing particle size of different adsorbents, due to the smaller specific surface areas. The equilibrium concentrations of the dye (Ceq ) is close to zero at their low initial concentrations (C0 ) and then increases linearly above a critical value C *0 (Fig. 6), which is characteristic of solutes that are chemisorbed. Indeed, the supernatant solutions are colorless below C *0 , indicating that essentially all of the dye is adsorbed by Ludox CL particles. As expected, the critical value decreases with increasing particle size when the same amount of different adsorbents is used. Effect of Adsorbates The adsorption isotherms for different dyes on Ludox CL5 (3.5) show the same trend, except the plateau values vary
(Fig. 7). The maximum adsorbed amounts obtained from these isotherms are listed in Table 3. Once again, in all cases the plot of Ceq versus C0 is characteristic of chemisorption (Fig. 8), which is not surprising because all of these dyes contain sulfonic groups. To verify these data, thermogravimetric analysis was used to determine the amount of the adsorbed dye, and a typical result with the red pigment is illustrated in Fig. 9. Since an organic stabilizer was present in Ludox CL, its weight content had to be determined in order to calculate the actual amount of the dye. For this purpose, two approaches were employed to prepare Ludox CL samples for TGA experiments. In the first, the Ludox CL sample was dried under vacuum to retain the stabilizer in the solid, resulting in a total weight loss of 10% on heating from 350 to 9007C. In the second approach, particles were aggregated and settled on addition of K2SO4 and then washed three times with distilled water. In the latter case, the weight loss was Ç3% because part of the stabilizer was leached out on washing. Thus, the amount of the stabilizer in these pigments should be between the two extreme values. The weight loss on heating the red and yellow pigments, evaluated by the TGA, was found to be Ç6% larger than calculated for the adsorbed amount of dye. The difference was most likely due to the stabilizer in Ludox, although some other minor additives could have been present. It should pointed out that
TABLE 3 Data for Adsorption of Dyes on Ludox CL-5 at 25 { 17C Dye
D&C Red #6
Acid Yellow #1
Acid Blue
Guinea Green B
Molecular weight (M) Dye/Ludox [Gmax, g/g (adsorption)] Dye/Ludox [Gmax , g/g (TGA)] Gm* ax (1004 mol/g*)a Area (nm2/molecule)b
430 0.30 0.36 7.0 1.1
358 0.19 0.24 5.0 1.5
416 0.70 0.63 13.7 0.6
691 0.60 0.61 8.0 1.0
a The Gm* ax used here is calculated from the TGA results after correcting for the contribution from the organic stabilizer in the Ludox CL sample (Å Gmax(TGA) 0 0.06). b The specific surface area used is derived from the titration.
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
coidas
DYE ABSORPTION ON Al-MODIFIED Si PARTICLES
225
FIG. 2. Scanning electron micrographs (SEM) of Ludox CL-12 particles (A) and after adsorbing D&C Red #6 (B).
such experiments with blue and green pigments are less reliable, because of the uncertainty with respect to the purity of the used dyes. To make results of different pigments comparable, data are listed in moles of dye per unit weight of adsorbent (Table 3). The chemisorption must be due to the interaction of sulfonic ( –SO 3– ) and/or carboxylic ( –COO – ) groups of
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
dyes with GAlOH sites on the surface of modified silica particles. The chemisorption mechanism is also supported by the fact that there is no observable leaching in water, ethanol, or the organic solvent (PMA) over a long experimentation times. The adsorption behavior of alumina-coated silica used in this work differs from that of acid-pretreated alumina (4, 5); the
coidas
226
WU ET AL.
FIG. 3. A typical size distribution of the Ludox CL-5 (3.5) (Table 2) sample, determined by the dynamic light scattering and analyzing the correlation function with the CONTIN program.
latter shows a reversible uptake of the dye, in strong dependence on the pH over the range of 1–9 and on the acid used to pretreat the adsorbent. Coadsorption of Yellow and Green Dyes To modify the color properties of the pigment, coadsorption of Acid Yellow #1 and Guinea Green B was studied by determining adsorbed amounts on Ludox CL from solutions containing these dyes in different molar ratios. For this purpose, the content of dyes in the pigments was calculated from the TGA data, which was possible because the weight loss for yellow and green pigments is very different, i.e., 0.24 and 0.61 g / g, respectively. The total dye concentration used ( Ç0.02 mol dm03 ) in these experiments exceeded the saturation level. Assuming that each dye in a mixture follows the same adsorption pattern on silica as when present alone in the
FIG. 4. The electrophoretic mobility as a function of the pH of Ludox CL-5 (3.5) sample at 25.07C.
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
FIG. 5. Adsorption isotherms of the D&C Red #6 dye for five different Ludox CL samples (Table 2) at 25 { 17C.
solution, the molar ratio in the solid should equal that in the solution. Consequently, the weight loss of pigments prepared with different mixtures of dyes can be calculated from the linear combination of the results observed with individual pigments. Figure 10 gives such data for the green and yellow dyes, which implies that there is no competition between the two for the Ludox adsorbent. Conformation of Dye Molecules on the Surface of Modified Silica Particles The conformation of the adsorbed dyes can be estimated from the surface area occupied by each molecule. D&C Red #6 is used as an example because of its purity. One difficulty in such calculation is the evaluation of the correct specific surface area of the adsorbent. The conformation of dye mole-
FIG. 6. The equilibrium concentration of D&C Red #6 dye ( Ceq ) as a function of its initial concentration (C0 ) for adsorbents listed in Table 2.
coidas
DYE ABSORPTION ON Al-MODIFIED Si PARTICLES
FIG. 7. Adsorption isotherms of D&C Red #6, Acid Yellow #1, Acid Blue 25, and Guinea Green B on Ludox CL-5 (3.5) at 25 { 17C.
cules was usually studied at the liquid/air interface to avoid the complexity of this problem (17). The area occupied by dye molecules calculated from adsorption on porous silica is much larger than that calculated from a close packed monolayer (18). However, the porosity effect should be negligible with nanosize modified silica particles used in this work. Two cases are considered here. In the first, it was assumed that the entire surface area of Ç210 m2 / g, as determined by titrating Ludox CL-12, was accessible to dye molecules, which yielded an area per dye molecule, A , of Ç1.3 nm2 . In the second case, the hydrodynamic radius from dynamic light scattering was used to calculate the specific area by assuming particles to be smooth. When the density of the solids is taken as Ç2.26 g / cm3 , the area occupied by each dye molecule is estimated to be 0.8 nm2 .
FIG. 8. The equilibrium concentration of dyes in the supernatant solution as a function of the initial concentration for the systems shown in Fig. 7.
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
227
FIG. 9. Typical thermogravimetric data for Ludox CL-5 (3.5) before and after adsorbing D&C Red #6 at the scanning rate of 257C min 01 . The weight loss below 3507C (arrow) is attributed to the removal of solvents. ‘‘Ludox/dry’’ refers to the sample dried in vacuum, and ‘‘Ludox/wash’’ was prepared by coagulating Ludox CL-5 (3.5) with potassium sulfate and then washing with water.
The projected area of the D&C Red #6 molecule is Ç1.3 nm2 , implying that the adsorbed dyes lie entirely flat or are slightly tilted. Optical Properties of Nanosize Pigments By adsorbing anionic dyes on alumina-modified silica particles well-defined pigments can be prepared, which is a more effective approach than grafting the dye to the surface of derivatized silica (1, 9). In addition, the described procedure yields uniform and reproducible pigments in contrast to the traditional method, which is based on reducing the
FIG. 10. The content of coadsorbed Guinea Green B and Acid Yellow #1 on Ludox CL-5 (3.3) as a function of the molar ratios of these dyes in the solution. The line represents the linear combination of green and yellow pigments of the same molar ratios.
coidas
228
WU ET AL.
particle size by grinding. Since the pigments retain essentially the same size and shape as the carrier particles (Fig. 2), their optical properties can be controlled by selecting the core materials. If the latter is in the nanosize range, as used in this work, it is expected that color films prepared with such pigments would have high transmittance at the desired wavelengths. Figure 11 shows the transmittances of red, blue, and green thin films containing 60 wt% solid pigments. The red and blue pigments were prepared by adsorbing D&C Red #6 and Acid Blue #25 on Ludox CL-5, respectively. The green pigment was obtained by using solutions containing both Acid Yellow #1 and Guinea Green B dyes at a weight ratio of 2:1 with the same silica. The transmittance is excellent, especially for the red thin film which reaches Ç100% even with a film thickness of 8 mm. The chromaticities of thin color films, determined by evaluating the transmitted spectra by means of commercial ( PECOL ) software, based on the CIE ( International Commission on Illumination ) system, are displayed in Fig. 12. The thin color films ( a few micrometers thick ) , prepared by using pigments described in this work, show characteristics to be very close to the NTSC ( National Television Standards Committee ) values ( red, x Å 0.64, y Å 0.33; green, x Å 0.29, y Å 0.60; blue, x Å 0.15, y Å 0.06 ) . The differences become smaller with increasing film thickness or with higher pigment contents in the film of the same thickness. However, the mechanical proper-
FIG. 11. The transmittance spectra of thin color films prepared by dispersing the red (D&C Red #6), blue (Acid Blue 25), and green (Acid Yellow #1 and Guinea Green B mixture in a weight ratio of 2:1) pigments in an acrylic copolymer at a solid loading of 60%. The film thickness is Ç2, Ç4, and Ç2 mm for red, blue, and green, respectively.
AID
JCIS 5156
/
6g35$$$282
11-04-97 05:52:10
FIG. 12. Chromaticity diagram based on the CIE (International Commission on Illumination) system. The solid squares represent the reference chromaticities for red, green, blue, and the CIE light source, respectively. The open triangles are the chromaticities of thin color films displayed in Fig. 11.
ties of such films deteriorate after the solid loading exceeds an upper limit of Ç70 wt%. REFERENCES 1. Winnik, F. M., Keoshkerian, B., Fuller, J. R., and Hofstra, P. G., Dyes Pigm. 14, 101 (1990). 2. McKay, G., Blair, H. S., and Gardner, J. R., J. Appl. Polym. Sci. 27, 3043 (1982). 3. Fleming, H. L., in ‘‘Fundamentals of Adsorption’’ (A. I. Liapis, Ed.), pp. 221. Engineering Foundation, New York, 1987. 4. Jain, J. K., Mundhara, G. L., and Tiwari, J. S., Colloid Surf. 29, 373 (1988). 5. Mishra, R. K., Mundhara, G. L., and Tiwari, J. S., J. Colloid Interface Sci. 129, 41 (1984). 6. Charreyre, M.-T., Zhang, P., Winnik, M. A., Pichot, C., and Graillat, C., J. Colloid Interface Sci. 170, 374 (1995). 7. Jan, D. E., and Raghavan, S., Colloid Surf. 92, 1 (1994). 8. Giesche, H., and Matijevic´, E., Dyes Pigm. 17, 323 (1991). 9. Hsu, W. P., Yu, R., and Matijevic´, E., Dyes Pigm. 19, 179 (1992). 10. Carotenuto, G., Her, Y.-S., and Matijevic´, E., Ind. Eng. Chem. Res. 35, 2929 (1996). 11. Tentorio, A., Matijevic´, E., and Kratohvil, J. P., J. Colloid Interface Sci. 77, 418 (1980). 12. DuPont Specialty Chemicals, ‘‘Ludoxt Colloidal Silica: Properties, Uses, Storage, and Handling.’’ Wilmington, DE. 13. Iler, R., ‘‘The Chemistry of Silica, Solubility, Polymerization, Colloid and Surface Properties, and Biochemistry.’’ Wiley, New York, 1979. 14. Provench, S. W., Comput. Phys. Commun. 27, 215 (1982). 15. Parks, G. A., and Bruyn, P. L., J. Phys. Chem. 66, 967 (1962). 16. Brace, R., and Matijevic´, E., J. Inorg. Nucl. Chem. 35, 3691 (1973). 17. Schmitt, F.-J., Meller, P., Ringsdorf, H., and Knoll, W., Prog. Colloid Polym. Sci. 83, 136 (1990). 18. Parida, S. K., and Mishra, B. K., J. Colloid Interface Sci. 182, 473 (1996).
coidas