- Email: [email protected]
Theoretical and Experimental Study of Spectral Reflectance and Color of Concentrated Oil-in-Water Emulsions
Journal of Colloid and Interface Science 218, 324 –330 (1999) Article ID jcis.1999.6428, available online at http://www.idealibrary.com on
Theoretical and Experimental Study of Spectral Reflectance and Color of Concentrated Oil-in-Water Emulsions Withida Chantrapornchai, Fergus Clydesdale, and D. Julian McClements 1 Biopolymers and Colloids Laboratory, Department of Food Science, University of Massachusetts, Amherst, Massachusetts 01003 Received April 8, 1999; accepted July 12, 1999
The influence of droplet and dye concentration on the optical properties of oil-in-water emulsions has been investigated. The spectral reflectance and tristimulus color coordinates (L, a, b) of a series of n-hexadecane oil-in-water emulsions with the same median droplet diameters (;0.3 mm), but different droplet concentrations (0.25 to 38.3 wt%) and red dye concentrations (0 to 0.2 wt%), were measured. Spectral reflectances decreased with increasing dye concentration and decreasing droplet concentration and had troughs at the same wavelengths as the peaks in the absorption spectra of the dyes. Emulsion L-values (“lightness”) decreased and a- and b-values (“chromacity measures”) increased as the dye concentration increased or the droplet concentration decreased. There was good qualitative agreement between the measured spectral reflectance of the emulsions and that predicted by the Kubelka–Munk theory of diffuse reflectance. Excellent agreement between theory and experiment could be obtained using an empirically determined correction-factor that accounts for cuvette effects. © 1999 Academic Press Key Words: emulsion; color; spectral reflectance; light scattering; cuvette effects.
INTRODUCTION
Optical properties of emulsions, such as spectral reflectance, spectral transmittance, angular light scattering, and color, are determined by the way that they interact with radiation in the visible region of the electromagnetic spectrum (1–5). When a beam of white light is incident upon the outer surface of a concentrated emulsion most of it is (initially) transmitted into the emulsion, while the remainder is reflected because of the difference in refractive index of the emulsion and its surroundings. For example, about 96% of a light beam that encounters an air– emulsion boundary is initially transmitted into the emulsion. Once the transmitted part enters the emulsion it is partly absorbed by absorbing species and partly scattered by particles before emerging (3, 6 –9). Absorption is largely responsible for the color of emulsions, whereas scattering is responsible for their turbidity or lightness (6). The light wave that is transmitted into the emulsion propa1
To whom correspondence should be addressed. Fax: 1 413 545 1019. E-mail: [email protected]. 0021-9797/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
gates through both the aqueous phase and the droplets, and will be absorbed by any chromophoric molecules present in either phase. The extent of absorption depends on the concentration and (wavelength-dependent) absorptivity of the chromophores, the wavelength of light, and the distance that each wavelength travels through the emulsion. Some wavelengths are absorbed more strongly than others are, so the color of the light emerging is no longer white. When a light beam encounters a droplet within an emulsion, it is scattered (1, 2). The characteristic scattering pattern of a single spherical droplet depends on its radius and relative refractive index, and can be accurately calculated (10). As the droplet concentration is increased, the fraction of light that is scattered increases, which leads to an increase in turbidity and reflectance (3). Once the droplet concentration exceeds a certain level it is not possible to transmit light through an emulsion because of the high degree of attenuation caused by scattering. The appearance of a concentrated emulsion is therefore determined mainly by that portion of light that is scattered by the droplets near its surface. In general, it is difficult to develop theoretical models to predict the optical properties of nondilute emulsions because of multiple scattering effects (2, 3, 11). Nevertheless, theoretical models have been developed for highly concentrated systems, where the light propagates through the emulsion primarily through diffusion (12). One of the most widely used of these diffusion theories is the Kubelka–Munk theory (3). This theory can be used to predict the spectral reflectance of light waves from concentrated emulsions, and thus their color (5). The influence of droplet size and concentration on the optical characteristics of concentrated oil-in-water emulsions containing a blue dye have recently been investigated (13). Emulsion “lightness” (L-value) increased with increasing droplet concentration and decreasing droplet size, while the opposite occurred for measures of the emulsion “chromacity” (a- and b-values). Concentrated emulsions containing red and green dyes exhibited similar behavior (14). As would be expected, peaks in the absorption spectra of the dyes corresponded to troughs in the reflectance spectra. Recently, we developed a theory to relate the optical properties of concentrated emulsions to the characteristics of the droplets and chromophores
324
OPTICAL PROPERTIES OF OIL-IN-WATER EMULSIONS
they contain (15). This theory is based on the Kubelka–Munk theory and predicts that the spectral reflectance (R ` ) of an emulsion is determined by the wavelength dependence of the ratio of the absorption coefficient (K) to the scattering coefficient (S), F KM 5
K ~1 2 R ` ! 2 5 , S 2 R`
[1]
where, F KM is the Kubelka–Munk function (3). The values of the scattering and absorption coefficients can be calculated from diffuse scattering theory (12), K 5 2S a S5
3 1 S @1 2 g# 2 S a, 4 s 4
[2] [3]
where S a and S s are the absorption and scattering cross sections of the droplets and g is the asymmetry factor. The development of the Kubelka–Munk theory is based on a number of assumptions that can limit its range of application. First, it is assumed that the continuous phase of the suspension is the same as the medium from which the radiation impinges, so that no additional reflection losses occur at the boundary because of the refractive index difference (3). Mathematical and empirical methods have been developed to overcome this limitation (3, 14). Second, it is assumed that the droplets within the emulsion are distributed evenly throughout the volume, which is the case in nonflocculated and noncreamed systems. Third, it is assumed that the light propagation through the emulsion is entirely diffuse, whereas in practice there may also be some nondiffuse propagation, especially in relatively dilute emulsions (15). Finally, it is assumed that the droplets are far enough apart (.3 3 radius) that the calculation of the scattering coefficients are not influenced by droplet– droplet interactions, which means the equations should not be applicable to highly concentrated emulsions. The emulsions used in the present study had relatively small droplet sizes (r 5 0.15 m m), so the droplet characteristics could be calculated using the Rayleigh–Gans–Debye theory (15). The absorption coefficient depends on the type and concentration of chromophores present, while the scattering coefficient depends on the size, refractive index, and concentration of the droplets present. Normally it is assumed that the scattering coefficient is proportional to the droplet concentration, while the absorption coefficient is proportional to the chromophore concentration. In the present study we examined the influence of droplet and dye concentration on the optical properties of concentrated oil-in-water emulsions and interpreted the data using the above theory. We also develop an empirical method of correcting theoretical predictions to take into account losses associated with the cuvette.
325
EXPERIMENTAL METHODS
Materials The n-hexadecane (.99% pure) and sodium dodecyl sulfate (SDS) were obtained from Sigma Chemical Company (St. Louis, MO). Red food dye (FD&C red 3 and red 40) was obtained from Durkee, a division of Burns Philp Food Inc. (San Francisco, CA). Stock solution of this dye was prepared by dilution with distilled water. Doubly distilled and deionized water was used to prepare all solution and emulsions. Emulsion Preparation A stock emulsion with relatively small droplets (r 32 ' 0.15 m m) was prepared by homogenizing 40 wt% n-hexadecane and 60 wt% surfactant solution (20 mM SDS in water) using a biohomoginizer (M133/1281-0, Biospec Products Inc., Bartlesville, OK) for 1 min (30 s at low speed and 30 s at high speed) and then using a sonicator for about 20 min at a power level of 280 W with a repeating duty cycle of 0.5 s (Braun Biotech, Allentown, PA). Emulsions containing droplet concentrations ranging from 0.25 to 38.3 wt% were prepared by diluting the stock emulsion with different masses of distilled water. For each droplet concentration a series of emulsions with different dye concentrations (0 – 0.2 wt%) was prepared by adding an equal mass (but different ratios) of stock dye solution and distilled water. Droplet Size Characterization A static light scattering instrument (Horiba LA-900, Horiba Instruments Incorporated, Irving, CA) was used to measure the droplet size of the emulsions. A relative refractive index of 1.08 (5 refractive index of oil/refractive index of aqueous phase) was used by the instrument to calculate the droplet size distribution. To avoid multiple scattering effects emulsions were diluted with distilled water prior to analysis so that the final droplet concentration was ;0.005 wt%. Droplet size measurements are reported as the volume–surface mean radius: r 32 5 Sn i r i3 /Sn i r i2 , where n i is the number of droplets of radius r i . The droplet size of the emulsions remained constant throughout the experiments, which indicated that no coalescence or Ostwald ripening occurred. All light scattering measurements were carried out before the dye was added to the emulsion to avoid complications in droplet size analysis associated with dye absorbance. Absorbance and Spectral Reflectance Measurements Absorbance of dye solutions and spectral reflectance of emulsions were measured using a UV-visible spectrophotometer (UV-2101PC, Shimadzu Scientific Instruments, Columbia, MD). During the measurements emulsions were contained in quartz cuvettes with a 1-cm path length. Spectra were obtained over the wavelength range 380 –780 nm using a scanning speed
326
CHANTRAPORNCHAI, CLYDESDALE, AND McCLEMENTS
absorbance increased less steeply (Fig. 2). The nonlinearity of the A 520nm versus concentration curve at higher concentrations is probably because of dye– dye interactions. Influence of Droplet Concentration on the Spectral Reflectance of Emulsions
FIG. 1. Influence of dye concentration on the absorption spectra of aqueous red dye solutions.
of 700 nm min 21. Absorbance measurements were made using a standard double-beam arrangement, with the absorption of the dye solution being measured relative to that of a reference cell containing distilled water. Spectral reflectance measurements were made using an integrating sphere arrangement (ISR-260, Shimadzu Scientific Instruments, Columbia, MD). The spectral reflectance of the emulsions was measured relative to a barium sulfate (BaSO 4) standard white plate. Colorimetric Measurements The color of the emulsions was measured using an instrumental colorimeter (Labscan II, Hunter Associates Laboratory, Reston, VA), which was calibrated using a white color standard tile with tristimulus values X 5 78.54, Y 5 83.18, and Z 5 85.80 (Standard No. LS-13556, Hunter Associates Laboratory, Reston, VA). Daylight (D 65) was used as a standardized light source. A fixed amount of emulsion sample was poured into the measurement cup, which was then surrounded with a black paper strip, and covered with a white tile before the measurement was carried out. The instrument provides the color of the samples in terms of the L, a, b color space system (16). In this color space, L represents the lightness and a and b are color coordinates, where 1a is the red direction, 2a is the green direction, 1b is the yellow direction, and 2b is the blue direction (16).
Figure 3 shows predicted and measured spectral reflectance of a series of n-hexadecane oil-in-water emulsions with different droplet concentrations (0.25–38.3 wt%), but the same median droplet diameter (0.3 mm) and the same dye concentration (0.1 wt%). There was a large increase in the measured spectral reflectance when the droplet concentration was increased from 0.25 to 5.0 wt%, followed by a less steep increase at higher droplet concentrations (Fig. 3b), as has been observed in previous studies (13, 14). The emulsions had a single broad spectral reflectance trough at 380 – 600 nm, with a minimum around 520 nm, which corresponded to the peak in the absorption spectra of the dye solutions (Fig. 1). The trough occurred because energy is selectively absorbed by the red dye at these wavelengths. The emulsion therefore appears “red” because violet-to-yellow light (380 –590 nm) is absorbed, leaving predominantly orange-to-red light (590 –760 nm) in the reflected beam. The increase in spectral reflectance with increasing droplet concentration occurred because the droplets scattered light more effectively, and therefore a greater portion of the light beam was reflected back from the emulsion. Theoretical predictions of the spectral reflectance made using Eqs. [1] to [3] were in good qualitative agreement with experimental data: the spectral reflectance had a trough between 380 and 600 nm and increased with increasing droplet concentration (Fig. 3a). Nevertheless, there were some significant differences between the predictions and measurements. The predicted spectral reflectance was close to 100% at wavelengths where the red dye did not absorb (600 to 780 nm), but the measured spectral reflectance was only 80% in the most concentrated emulsion at these wavelengths and decreased appreciably with decreasing droplet concentration.
RESULTS AND DISCUSSION
Absorption Spectra of Dye Solutions Absorption spectra of red dye solutions (0 to 0.2 wt%) are shown in Fig. 1. The red dye had a single broad absorption peak at 380 – 600 nm, with a maximum at 520 nm. As expected, the height of the absorption maximum increased as the dye concentration increased. There was a linear relationship between the height of the absorption maximum ( A 520nm) and dye concentration up to about 0.1 wt% dye, after which the
FIG. 2. Concentration dependence of the absorption of red dye solutions at 520 nm.
OPTICAL PROPERTIES OF OIL-IN-WATER EMULSIONS
327
sions, but again measured values were consistently lower than predicted values. Empirical Correction of Measured Spectral Reflectance The most likely reason for the observed discrepancies between theory and experiments is that the Kubelka–Munk theory assumes that the reflection occurs at an interface between a pure continuous phase and an infinitely thick and infinitely wide emulsion (15). In reality the sample is contained within a quartz cuvette that is of finite thickness and width, and the reflection occurs at an air– quartz–sample interface (15, 18, 19). As a result some of the light that is incident upon the emulsions reverberates within the cuvette walls, which leads to a reduction in the reflectance because light is absorbed each time the wave encounters the emulsion. In addition, it is possible that some of the light propagates through the emulsions and is transmitted through the side or end of the cuvettes, which would also lower their reflectance. In a previous study, we developed a theoretical method of improving the agreement between the predicted and measured spectral reflectance by analyzing the interaction of a light wave with an air– cuvette–
FIG. 3. Influence of droplet concentration on the spectral reflectance of n-hexadecane oil-in-water emulsions containing 0.1 wt% red food dye: (a) theoretical predictions (Eqs. [1] to [3]), (b) experimental measurements (points) compared to corrected theoretical predictions (curves).
Influence of Dye Concentration on the Spectral Reflectance of Emulsions Predicted and measured reflectance spectra of a series of n-hexadecane oil-in-water emulsions with the same median droplet diameter (0.3 mm) and concentration (10 wt%) but different dye concentrations (0 to 0.20 wt%) are shown in Fig. 4. In the absence of dye, there was a small broad peak in the measured reflectance spectra at wavelengths between 400 and 450 nm, which was probably due to absorption in the quartz cuvette (6), followed by a slight decrease in reflectance with increasing wavelength at higher wavelengths (Fig. 4b), which was probably due to absorption by the water (17). These emulsions appeared “whitish” because the spectral reflectance did not change appreciably over the entire wavelength range. In the presence of dye, there were troughs in the reflectance spectra that corresponded to the peaks in the absorption spectra of the red dye (Fig. 1). These troughs became deeper as the dye concentration increased because more light was selectively absorbed by the dye molecules at these wavelengths. There was a good qualitative agreement between the predicted (Fig. 4a) and measured (Fig. 4b) spectral reflectance of the emul-
FIG. 4. Influence of dye concentration on the spectral reflectance of n-hexadecane oil-in-water emulsions containing 10 wt% droplets: (a) theoretical predictions (Eqs. [1] to [3]), (b) experimental measurements (points) compared to corrected theoretical predictions (curves).
328
CHANTRAPORNCHAI, CLYDESDALE, AND McCLEMENTS
FIG. 5. Dependence of empirical calibration-factor, C(K, S) [5R M/R P(K, S)], on the absorption coefficient (K) and scattering coefficient (S).
emulsion system (15). Nevertheless, the quantitative agreement was poor at wavelengths where there was low absorption or low scattering. In the present study we took a different approach, which involved establishing an empirically determined correction-factor to relate the measured reflectance (R M) to the predicted reflectance (R P): C~K, S! 5
RM . R P~K, S!
[4]
The absorption coefficient (K), scattering coefficient (S), and spectral reflectance (R P) were calculated as a function of wavelength (l 5 380 to 770 nm, in 10-nm increments) using Eqs. [1] to [3] for emulsions containing 0, 0.1, and 0.2 wt% red dye and droplet concentrations between 0.25 and 30 wt%. The correction factor was determined at each wavelength by dividing the measured spectral reflectance by the predicted one. An interpolation technique was then used to calculate C(K, S) for any pair of K and S values (MathCad Version 7.0). The dependence of C(K, S) on the predicted scattering and absorption coefficients is shown in Fig. 5. As the scattering coefficient decreased (i.e., the emulsion became more dilute), the measured reflectance became appreciably smaller than the predicted value, so C(K, S) decreased. The correction factor was stored in the computer program used to calculate the spectral reflectance of the emulsions, and was applied to the predicted reflectance values in order to take into account cuvette effects. There was excellent agreement between the measured spectral reflectance and the predicted values corrected for cuvette effects using Eq. [4] (Figs. 3b and 4b), which highlights the usefulness of this approach. For all calculations we assumed that the emulsion droplets were monodisperse, whereas in fact they were polydisperse. Nevertheless, numerical calculations carried out in our laboratory using Kubelka–Munk theory
established that droplet polydispersity did not have a significant influence on the predicted spectral reflectance for the emulsions studied in this work. We wanted to show that the empirical calibration-factor, C(K, S), determined above could be used to correct the spectral reflectance of other emulsions measured using the same experimental arrangement. In a previous study we measured the spectral reflectance of 10 wt% n-hexadecane oil-in-water emulsions (r 5 0.15 m m) containing red, green, and blue dyes (14). In Fig. 6 we compare these measurements with values predicted using Eqs. [1] to [3] and corrected for cuvette effects using the empirical correction-factor (Eq. [4]). There was excellent agreement between the measurements and the corrected predictions, which again highlights the usefulness of the empirical approach for correcting the predicted spectra. This suggests that once the calibration-factor has been established for a particular experimental arrangement it can be used in the analysis of all future samples. Use of the Kubelka–Munk Function to Characterize Colored Emulsions It is useful to have some measurable parameter which indicates the way that the reflectance of a colored emulsion changes with dye or droplet concentration. In dilute emulsions it is possible to relate the dye concentration to the absorbance and the droplet concentration to the turbidity. In concentrated emulsions it is necessary to use a different approach because it is not possible to transmit light through the system. To a first approximation the absorption coefficient of a concentrated emulsion is proportional to the concentration of dye (K 5 K9c), while the scattering coefficient is proportional to the concentration of droplets (S 5 S9 f ). Consequently, the Kubelka–Munk function should be proportional to the dye concentration, but the reciprocal of the Kubelka–Munk function should be proportional to the droplet concentration (Eq. [1]). Kubelka–Munk functions of emulsions containing differ-
FIG. 6. Influence of dye type on the spectral reflectance of n-hexadecane oil-in-water emulsions containing 10 wt% droplets, experimental measurements (points) compared to corrected theoretical predictions (curves).
OPTICAL PROPERTIES OF OIL-IN-WATER EMULSIONS
329
FIG. 7. Dependence of the Kubelka–Munk function of red dye solutions at 520 nm on dye concentration.
ent dye and droplet concentrations were calculated from measurements of the spectral reflectance at 520 nm, i.e., at the absorption peak of the red dye (Figs. 7 and 8). For all droplet concentrations, the Kubelka–Munk function was approximately linearly related to the dye concentration up to about 0.1 wt% (Fig. 6), and then increased less steeply at higher concentrations, which was in close agreement with the variation of absorbance with dye concentration (Fig. 2). For all dye concentrations, the reciprocal of the Kubelka–Munk function was linearly related to the droplet concentration up to about 3 wt%, but became increasingly nonlinear at higher droplet concentrations (Fig. 8). The appreciable decrease in 1/F KM that occurs at high concentrations was probably due to multiple-scattering and particle–particle interactions that are not accounted for in the theory used to calculate the scattering cross section and asymmetry factor of the droplets (20). Colorimetric Measurements The tristimulus coordinates (L, a, and b values) of n-hexadecane oil-in-water emulsions with the same median
FIG. 8. Dependence of the reciprocal of the Kubelka–Munk function of red dye solutions at 520 nm on droplet concentration.
FIG. 9. Dependence of L, a, b values on droplet and dye concentration for n-hexadecane oil-in-water emulsions containing a red food dye.
droplet diameter (0.3 mm), but different droplet concentrations (0.25 to 38.3 wt%) and red dye concentrations (0 to 0.20 wt%), were measured using a colorimeter (Fig. 9). In the absence of dye, the L-value increased steeply with increasing droplet concentration from 0 to 3 wt%, and then leveled off at approximately L 5 100% at higher concentrations. A similar type of behavior was observed for the emulsions containing dye, except that the initial increase in L-value was less steep as the dye concentration increased, and the final L-value was lower than for the dye-free emulsions. These results indicate that the emulsions became lighter as the droplet concentration increased and the dye concentration decreased. This has important consequences for the formulation of many types of commercial emulsion-based products. For example, in the food industry, there is great emphasis on the development of reduced fat emulsions that have the same desirable attributes as the original product (21, 22). Our results indicate that both the fat and the dye content of a food emulsion must be carefully controlled, since the increase of dye concentration and the reduction of fat content below about 3 wt% leads to an appreciable decrease of the lightness, which would adversely effect its consumer acceptability.
330
CHANTRAPORNCHAI, CLYDESDALE, AND McCLEMENTS
The a- and b-values of the emulsions were also strongly influenced by the droplet and dye concentration (Fig. 9). In the absence of dye, the a-value of all emulsions and the b-values of the concentrated emulsions (f . 5 wt%) were close to zero. At lower droplet concentrations, the b-value of the dye-free emulsions became increasingly negative as the droplet concentration was reduced, which indicated that they become bluer in appearance. The a-values (5 to 50) and b-values (0 to 18) of the emulsions containing dye indicated that they were in the redyellow region of the color space (16). Emulsion color was strongly influenced by droplet and dye concentration. Emulsions became more intensely colored (i.e., a- and b-values moved away from zero) as the droplet concentration decreased (down to about 0.5 wt%) and the dye concentration increased. CONCLUSIONS
The optical properties of concentrated oil-in-water emulsions depend strongly on the concentration of dye and droplets present. The spectral reflectance of emulsions could be described qualitatively using the Kubelka–Munk theory of diffuse reflectance, but the quantitative agreement was fairly poor because of light losses associated with the measurement cell. Nevertheless, we developed an empirical method of correcting the predicted spectra that gave excellent agreement between theory and experimental data. The Kubelka–Munk function increased approximately linearly with dye concentration and was closely related to the absorbance of the dye solutions. The reciprocal of the Kubelka–Munk function increased linearly with droplet concentration at relatively low concentrations (f , 3 wt%), but decreased at higher values because of multiple scattering and droplet– droplet interactions. The information presented in this study would be useful for the formulation of emulsions with specific colors and lightness, e.g., foods, paints, cosmetics, and pharmaceuticals. The dependence of the optical properties of emulsions on particle size suggests that droplet aggregation would have a significant impact on their appearance. For this reason we are currently examining the influence of floc size and structure on the optical properties of concentrated emulsions. It should be noted that diffusing wave spectroscopy (DWS) techniques are being applied more frequently for the characterization of the structure, interactions, and dynamics of concentrated colloidal systems. These techniques also rely on the diffusion of light waves through concentrated colloidal dispersions. Nevertheless, they utilize temporal fluctuations in the intensity of mono-
chromatic scattered light caused by particle diffusion to provide information about the system, whereas the optical reflectance spectroscopy technique used in this study utilizes measurements of the wavelength dependence of the timeaveraged back-scattered light (23). ACKNOWLEDGMENTS The authors thank the Massachusetts Agricultural Experiment Station (Project MAS00745) for supporting this work.
REFERENCES 1. van de Hulst, H. C., “Light Scattering by Small Particles.” Wiley, New York, 1957. 2. Kerker, M., “The Scattering of Light.” Academic Press, New York, 1969. 3. Kortum, G., “Reflectance Spectroscopy: Principles, Methods, Applications.” Springer-Verlag, New York, 1969. 4. Heimenz, P. C., and Rajagopalan, R., “Principles of Colloid and Surface Science.” Dekker, New York, 1997. 5. McClements, D. J., “Food Emulsions: Principles, Practice and Techniques.” CRC Press, Boca Raton, FL, 1998. 6. Wyszecki, G., and Stiles, W. S., “Color Science: Concepts and Methods, Quantitative Data and Formulas.” Wiley, New York, 1967. 7. Wendlandt, W. W., “Modern Aspects of Reflectance Spectroscopy.” Plenum, New York, 1968. 8. Frei, R. W., and MacNeil, J. D., “Diffuse Reflectance Spectrocopy in Environmental Problem-Solving.” CRC Press, Cleveland, OH, 1973. 9. Frei, R. W., Frodyma, M. M., and Lieu, V. T., in “Comprehensive Analytical Chemistry” (G. Svehla, Ed.), Chap. 3. Elsevier, Amsterdam, 1975. 10. Wickramasinghe, N. C., “Light Scattering Functions for Small Particles.” Wiley, New York, 1973. 11. Bailey, A. E., and Cannell, D. S., Phys. Rev. E. 50, 4853 (1994). 12. Mudget, P. S., and Richards, L. W., Appl. Optics 10, 1485 (1971). 13. Chantrapornchai, W., Clydesdale, F. M., and McClements, D. J., J. Agric. Food Chem. 46(8), 2914 (1998). 14. Chantrapornchai, W., Clydesdale, F. M., and McClements, D. J., Colloids Surf. A. 155, 373 (1998). 15. McClements, D. J., Chantrapornchai, W., and Clydesdale, F. M., J. Food Sci. 63(6), 935 (1998). 16. Hutchings, J. B., “Food Colour and Appearence.” Blackie, London, 1994. 17. Smith, R. C., and Baker, K. S., Appl. Optics 20, 177 (1981). 18. Seyfried, M., and Fukshansky, L., Appl. Optics 22, 1402 (1983). 19. Seyfried, M., Fukshansky, L., and Schafer, Appl. Optics 22, 492 (1983). 20. Ma, Y., Varadan, V. K., and Varadan, V. V., J. Acoust. Soc. Am. 87, 2779 (1990). 21. Akoh, C. C., Food Technol. 52(3), 47 (1998). 22. McClements, D. J., and Demetriades, K., Crit. Rev. Food Sci. Nutr. 38, 511 (1998). 23. Weitz, D. A., Zhu, J. X., Durian, D. J., Gauy, H., and Pine, D. J., Phys. Scr. T49, 610 (1993).
Theoretical and Experimental Study of Spectral Reflectance and Color of Concentrated Oil-in-Water Emulsions Withida Chantrapornchai, Fergus Clydesdale, and D. Julian McClements 1 Biopolymers and Colloids Laboratory, Department of Food Science, University of Massachusetts, Amherst, Massachusetts 01003 Received April 8, 1999; accepted July 12, 1999
The influence of droplet and dye concentration on the optical properties of oil-in-water emulsions has been investigated. The spectral reflectance and tristimulus color coordinates (L, a, b) of a series of n-hexadecane oil-in-water emulsions with the same median droplet diameters (;0.3 mm), but different droplet concentrations (0.25 to 38.3 wt%) and red dye concentrations (0 to 0.2 wt%), were measured. Spectral reflectances decreased with increasing dye concentration and decreasing droplet concentration and had troughs at the same wavelengths as the peaks in the absorption spectra of the dyes. Emulsion L-values (“lightness”) decreased and a- and b-values (“chromacity measures”) increased as the dye concentration increased or the droplet concentration decreased. There was good qualitative agreement between the measured spectral reflectance of the emulsions and that predicted by the Kubelka–Munk theory of diffuse reflectance. Excellent agreement between theory and experiment could be obtained using an empirically determined correction-factor that accounts for cuvette effects. © 1999 Academic Press Key Words: emulsion; color; spectral reflectance; light scattering; cuvette effects.
INTRODUCTION
Optical properties of emulsions, such as spectral reflectance, spectral transmittance, angular light scattering, and color, are determined by the way that they interact with radiation in the visible region of the electromagnetic spectrum (1–5). When a beam of white light is incident upon the outer surface of a concentrated emulsion most of it is (initially) transmitted into the emulsion, while the remainder is reflected because of the difference in refractive index of the emulsion and its surroundings. For example, about 96% of a light beam that encounters an air– emulsion boundary is initially transmitted into the emulsion. Once the transmitted part enters the emulsion it is partly absorbed by absorbing species and partly scattered by particles before emerging (3, 6 –9). Absorption is largely responsible for the color of emulsions, whereas scattering is responsible for their turbidity or lightness (6). The light wave that is transmitted into the emulsion propa1
To whom correspondence should be addressed. Fax: 1 413 545 1019. E-mail: [email protected]. 0021-9797/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
gates through both the aqueous phase and the droplets, and will be absorbed by any chromophoric molecules present in either phase. The extent of absorption depends on the concentration and (wavelength-dependent) absorptivity of the chromophores, the wavelength of light, and the distance that each wavelength travels through the emulsion. Some wavelengths are absorbed more strongly than others are, so the color of the light emerging is no longer white. When a light beam encounters a droplet within an emulsion, it is scattered (1, 2). The characteristic scattering pattern of a single spherical droplet depends on its radius and relative refractive index, and can be accurately calculated (10). As the droplet concentration is increased, the fraction of light that is scattered increases, which leads to an increase in turbidity and reflectance (3). Once the droplet concentration exceeds a certain level it is not possible to transmit light through an emulsion because of the high degree of attenuation caused by scattering. The appearance of a concentrated emulsion is therefore determined mainly by that portion of light that is scattered by the droplets near its surface. In general, it is difficult to develop theoretical models to predict the optical properties of nondilute emulsions because of multiple scattering effects (2, 3, 11). Nevertheless, theoretical models have been developed for highly concentrated systems, where the light propagates through the emulsion primarily through diffusion (12). One of the most widely used of these diffusion theories is the Kubelka–Munk theory (3). This theory can be used to predict the spectral reflectance of light waves from concentrated emulsions, and thus their color (5). The influence of droplet size and concentration on the optical characteristics of concentrated oil-in-water emulsions containing a blue dye have recently been investigated (13). Emulsion “lightness” (L-value) increased with increasing droplet concentration and decreasing droplet size, while the opposite occurred for measures of the emulsion “chromacity” (a- and b-values). Concentrated emulsions containing red and green dyes exhibited similar behavior (14). As would be expected, peaks in the absorption spectra of the dyes corresponded to troughs in the reflectance spectra. Recently, we developed a theory to relate the optical properties of concentrated emulsions to the characteristics of the droplets and chromophores
324
OPTICAL PROPERTIES OF OIL-IN-WATER EMULSIONS
they contain (15). This theory is based on the Kubelka–Munk theory and predicts that the spectral reflectance (R ` ) of an emulsion is determined by the wavelength dependence of the ratio of the absorption coefficient (K) to the scattering coefficient (S), F KM 5
K ~1 2 R ` ! 2 5 , S 2 R`
[1]
where, F KM is the Kubelka–Munk function (3). The values of the scattering and absorption coefficients can be calculated from diffuse scattering theory (12), K 5 2S a S5
3 1 S @1 2 g# 2 S a, 4 s 4
[2] [3]
where S a and S s are the absorption and scattering cross sections of the droplets and g is the asymmetry factor. The development of the Kubelka–Munk theory is based on a number of assumptions that can limit its range of application. First, it is assumed that the continuous phase of the suspension is the same as the medium from which the radiation impinges, so that no additional reflection losses occur at the boundary because of the refractive index difference (3). Mathematical and empirical methods have been developed to overcome this limitation (3, 14). Second, it is assumed that the droplets within the emulsion are distributed evenly throughout the volume, which is the case in nonflocculated and noncreamed systems. Third, it is assumed that the light propagation through the emulsion is entirely diffuse, whereas in practice there may also be some nondiffuse propagation, especially in relatively dilute emulsions (15). Finally, it is assumed that the droplets are far enough apart (.3 3 radius) that the calculation of the scattering coefficients are not influenced by droplet– droplet interactions, which means the equations should not be applicable to highly concentrated emulsions. The emulsions used in the present study had relatively small droplet sizes (r 5 0.15 m m), so the droplet characteristics could be calculated using the Rayleigh–Gans–Debye theory (15). The absorption coefficient depends on the type and concentration of chromophores present, while the scattering coefficient depends on the size, refractive index, and concentration of the droplets present. Normally it is assumed that the scattering coefficient is proportional to the droplet concentration, while the absorption coefficient is proportional to the chromophore concentration. In the present study we examined the influence of droplet and dye concentration on the optical properties of concentrated oil-in-water emulsions and interpreted the data using the above theory. We also develop an empirical method of correcting theoretical predictions to take into account losses associated with the cuvette.
325
EXPERIMENTAL METHODS
Materials The n-hexadecane (.99% pure) and sodium dodecyl sulfate (SDS) were obtained from Sigma Chemical Company (St. Louis, MO). Red food dye (FD&C red 3 and red 40) was obtained from Durkee, a division of Burns Philp Food Inc. (San Francisco, CA). Stock solution of this dye was prepared by dilution with distilled water. Doubly distilled and deionized water was used to prepare all solution and emulsions. Emulsion Preparation A stock emulsion with relatively small droplets (r 32 ' 0.15 m m) was prepared by homogenizing 40 wt% n-hexadecane and 60 wt% surfactant solution (20 mM SDS in water) using a biohomoginizer (M133/1281-0, Biospec Products Inc., Bartlesville, OK) for 1 min (30 s at low speed and 30 s at high speed) and then using a sonicator for about 20 min at a power level of 280 W with a repeating duty cycle of 0.5 s (Braun Biotech, Allentown, PA). Emulsions containing droplet concentrations ranging from 0.25 to 38.3 wt% were prepared by diluting the stock emulsion with different masses of distilled water. For each droplet concentration a series of emulsions with different dye concentrations (0 – 0.2 wt%) was prepared by adding an equal mass (but different ratios) of stock dye solution and distilled water. Droplet Size Characterization A static light scattering instrument (Horiba LA-900, Horiba Instruments Incorporated, Irving, CA) was used to measure the droplet size of the emulsions. A relative refractive index of 1.08 (5 refractive index of oil/refractive index of aqueous phase) was used by the instrument to calculate the droplet size distribution. To avoid multiple scattering effects emulsions were diluted with distilled water prior to analysis so that the final droplet concentration was ;0.005 wt%. Droplet size measurements are reported as the volume–surface mean radius: r 32 5 Sn i r i3 /Sn i r i2 , where n i is the number of droplets of radius r i . The droplet size of the emulsions remained constant throughout the experiments, which indicated that no coalescence or Ostwald ripening occurred. All light scattering measurements were carried out before the dye was added to the emulsion to avoid complications in droplet size analysis associated with dye absorbance. Absorbance and Spectral Reflectance Measurements Absorbance of dye solutions and spectral reflectance of emulsions were measured using a UV-visible spectrophotometer (UV-2101PC, Shimadzu Scientific Instruments, Columbia, MD). During the measurements emulsions were contained in quartz cuvettes with a 1-cm path length. Spectra were obtained over the wavelength range 380 –780 nm using a scanning speed
326
CHANTRAPORNCHAI, CLYDESDALE, AND McCLEMENTS
absorbance increased less steeply (Fig. 2). The nonlinearity of the A 520nm versus concentration curve at higher concentrations is probably because of dye– dye interactions. Influence of Droplet Concentration on the Spectral Reflectance of Emulsions
FIG. 1. Influence of dye concentration on the absorption spectra of aqueous red dye solutions.
of 700 nm min 21. Absorbance measurements were made using a standard double-beam arrangement, with the absorption of the dye solution being measured relative to that of a reference cell containing distilled water. Spectral reflectance measurements were made using an integrating sphere arrangement (ISR-260, Shimadzu Scientific Instruments, Columbia, MD). The spectral reflectance of the emulsions was measured relative to a barium sulfate (BaSO 4) standard white plate. Colorimetric Measurements The color of the emulsions was measured using an instrumental colorimeter (Labscan II, Hunter Associates Laboratory, Reston, VA), which was calibrated using a white color standard tile with tristimulus values X 5 78.54, Y 5 83.18, and Z 5 85.80 (Standard No. LS-13556, Hunter Associates Laboratory, Reston, VA). Daylight (D 65) was used as a standardized light source. A fixed amount of emulsion sample was poured into the measurement cup, which was then surrounded with a black paper strip, and covered with a white tile before the measurement was carried out. The instrument provides the color of the samples in terms of the L, a, b color space system (16). In this color space, L represents the lightness and a and b are color coordinates, where 1a is the red direction, 2a is the green direction, 1b is the yellow direction, and 2b is the blue direction (16).
Figure 3 shows predicted and measured spectral reflectance of a series of n-hexadecane oil-in-water emulsions with different droplet concentrations (0.25–38.3 wt%), but the same median droplet diameter (0.3 mm) and the same dye concentration (0.1 wt%). There was a large increase in the measured spectral reflectance when the droplet concentration was increased from 0.25 to 5.0 wt%, followed by a less steep increase at higher droplet concentrations (Fig. 3b), as has been observed in previous studies (13, 14). The emulsions had a single broad spectral reflectance trough at 380 – 600 nm, with a minimum around 520 nm, which corresponded to the peak in the absorption spectra of the dye solutions (Fig. 1). The trough occurred because energy is selectively absorbed by the red dye at these wavelengths. The emulsion therefore appears “red” because violet-to-yellow light (380 –590 nm) is absorbed, leaving predominantly orange-to-red light (590 –760 nm) in the reflected beam. The increase in spectral reflectance with increasing droplet concentration occurred because the droplets scattered light more effectively, and therefore a greater portion of the light beam was reflected back from the emulsion. Theoretical predictions of the spectral reflectance made using Eqs. [1] to [3] were in good qualitative agreement with experimental data: the spectral reflectance had a trough between 380 and 600 nm and increased with increasing droplet concentration (Fig. 3a). Nevertheless, there were some significant differences between the predictions and measurements. The predicted spectral reflectance was close to 100% at wavelengths where the red dye did not absorb (600 to 780 nm), but the measured spectral reflectance was only 80% in the most concentrated emulsion at these wavelengths and decreased appreciably with decreasing droplet concentration.
RESULTS AND DISCUSSION
Absorption Spectra of Dye Solutions Absorption spectra of red dye solutions (0 to 0.2 wt%) are shown in Fig. 1. The red dye had a single broad absorption peak at 380 – 600 nm, with a maximum at 520 nm. As expected, the height of the absorption maximum increased as the dye concentration increased. There was a linear relationship between the height of the absorption maximum ( A 520nm) and dye concentration up to about 0.1 wt% dye, after which the
FIG. 2. Concentration dependence of the absorption of red dye solutions at 520 nm.
OPTICAL PROPERTIES OF OIL-IN-WATER EMULSIONS
327
sions, but again measured values were consistently lower than predicted values. Empirical Correction of Measured Spectral Reflectance The most likely reason for the observed discrepancies between theory and experiments is that the Kubelka–Munk theory assumes that the reflection occurs at an interface between a pure continuous phase and an infinitely thick and infinitely wide emulsion (15). In reality the sample is contained within a quartz cuvette that is of finite thickness and width, and the reflection occurs at an air– quartz–sample interface (15, 18, 19). As a result some of the light that is incident upon the emulsions reverberates within the cuvette walls, which leads to a reduction in the reflectance because light is absorbed each time the wave encounters the emulsion. In addition, it is possible that some of the light propagates through the emulsions and is transmitted through the side or end of the cuvettes, which would also lower their reflectance. In a previous study, we developed a theoretical method of improving the agreement between the predicted and measured spectral reflectance by analyzing the interaction of a light wave with an air– cuvette–
FIG. 3. Influence of droplet concentration on the spectral reflectance of n-hexadecane oil-in-water emulsions containing 0.1 wt% red food dye: (a) theoretical predictions (Eqs. [1] to [3]), (b) experimental measurements (points) compared to corrected theoretical predictions (curves).
Influence of Dye Concentration on the Spectral Reflectance of Emulsions Predicted and measured reflectance spectra of a series of n-hexadecane oil-in-water emulsions with the same median droplet diameter (0.3 mm) and concentration (10 wt%) but different dye concentrations (0 to 0.20 wt%) are shown in Fig. 4. In the absence of dye, there was a small broad peak in the measured reflectance spectra at wavelengths between 400 and 450 nm, which was probably due to absorption in the quartz cuvette (6), followed by a slight decrease in reflectance with increasing wavelength at higher wavelengths (Fig. 4b), which was probably due to absorption by the water (17). These emulsions appeared “whitish” because the spectral reflectance did not change appreciably over the entire wavelength range. In the presence of dye, there were troughs in the reflectance spectra that corresponded to the peaks in the absorption spectra of the red dye (Fig. 1). These troughs became deeper as the dye concentration increased because more light was selectively absorbed by the dye molecules at these wavelengths. There was a good qualitative agreement between the predicted (Fig. 4a) and measured (Fig. 4b) spectral reflectance of the emul-
FIG. 4. Influence of dye concentration on the spectral reflectance of n-hexadecane oil-in-water emulsions containing 10 wt% droplets: (a) theoretical predictions (Eqs. [1] to [3]), (b) experimental measurements (points) compared to corrected theoretical predictions (curves).
328
CHANTRAPORNCHAI, CLYDESDALE, AND McCLEMENTS
FIG. 5. Dependence of empirical calibration-factor, C(K, S) [5R M/R P(K, S)], on the absorption coefficient (K) and scattering coefficient (S).
emulsion system (15). Nevertheless, the quantitative agreement was poor at wavelengths where there was low absorption or low scattering. In the present study we took a different approach, which involved establishing an empirically determined correction-factor to relate the measured reflectance (R M) to the predicted reflectance (R P): C~K, S! 5
RM . R P~K, S!
[4]
The absorption coefficient (K), scattering coefficient (S), and spectral reflectance (R P) were calculated as a function of wavelength (l 5 380 to 770 nm, in 10-nm increments) using Eqs. [1] to [3] for emulsions containing 0, 0.1, and 0.2 wt% red dye and droplet concentrations between 0.25 and 30 wt%. The correction factor was determined at each wavelength by dividing the measured spectral reflectance by the predicted one. An interpolation technique was then used to calculate C(K, S) for any pair of K and S values (MathCad Version 7.0). The dependence of C(K, S) on the predicted scattering and absorption coefficients is shown in Fig. 5. As the scattering coefficient decreased (i.e., the emulsion became more dilute), the measured reflectance became appreciably smaller than the predicted value, so C(K, S) decreased. The correction factor was stored in the computer program used to calculate the spectral reflectance of the emulsions, and was applied to the predicted reflectance values in order to take into account cuvette effects. There was excellent agreement between the measured spectral reflectance and the predicted values corrected for cuvette effects using Eq. [4] (Figs. 3b and 4b), which highlights the usefulness of this approach. For all calculations we assumed that the emulsion droplets were monodisperse, whereas in fact they were polydisperse. Nevertheless, numerical calculations carried out in our laboratory using Kubelka–Munk theory
established that droplet polydispersity did not have a significant influence on the predicted spectral reflectance for the emulsions studied in this work. We wanted to show that the empirical calibration-factor, C(K, S), determined above could be used to correct the spectral reflectance of other emulsions measured using the same experimental arrangement. In a previous study we measured the spectral reflectance of 10 wt% n-hexadecane oil-in-water emulsions (r 5 0.15 m m) containing red, green, and blue dyes (14). In Fig. 6 we compare these measurements with values predicted using Eqs. [1] to [3] and corrected for cuvette effects using the empirical correction-factor (Eq. [4]). There was excellent agreement between the measurements and the corrected predictions, which again highlights the usefulness of the empirical approach for correcting the predicted spectra. This suggests that once the calibration-factor has been established for a particular experimental arrangement it can be used in the analysis of all future samples. Use of the Kubelka–Munk Function to Characterize Colored Emulsions It is useful to have some measurable parameter which indicates the way that the reflectance of a colored emulsion changes with dye or droplet concentration. In dilute emulsions it is possible to relate the dye concentration to the absorbance and the droplet concentration to the turbidity. In concentrated emulsions it is necessary to use a different approach because it is not possible to transmit light through the system. To a first approximation the absorption coefficient of a concentrated emulsion is proportional to the concentration of dye (K 5 K9c), while the scattering coefficient is proportional to the concentration of droplets (S 5 S9 f ). Consequently, the Kubelka–Munk function should be proportional to the dye concentration, but the reciprocal of the Kubelka–Munk function should be proportional to the droplet concentration (Eq. [1]). Kubelka–Munk functions of emulsions containing differ-
FIG. 6. Influence of dye type on the spectral reflectance of n-hexadecane oil-in-water emulsions containing 10 wt% droplets, experimental measurements (points) compared to corrected theoretical predictions (curves).
OPTICAL PROPERTIES OF OIL-IN-WATER EMULSIONS
329
FIG. 7. Dependence of the Kubelka–Munk function of red dye solutions at 520 nm on dye concentration.
ent dye and droplet concentrations were calculated from measurements of the spectral reflectance at 520 nm, i.e., at the absorption peak of the red dye (Figs. 7 and 8). For all droplet concentrations, the Kubelka–Munk function was approximately linearly related to the dye concentration up to about 0.1 wt% (Fig. 6), and then increased less steeply at higher concentrations, which was in close agreement with the variation of absorbance with dye concentration (Fig. 2). For all dye concentrations, the reciprocal of the Kubelka–Munk function was linearly related to the droplet concentration up to about 3 wt%, but became increasingly nonlinear at higher droplet concentrations (Fig. 8). The appreciable decrease in 1/F KM that occurs at high concentrations was probably due to multiple-scattering and particle–particle interactions that are not accounted for in the theory used to calculate the scattering cross section and asymmetry factor of the droplets (20). Colorimetric Measurements The tristimulus coordinates (L, a, and b values) of n-hexadecane oil-in-water emulsions with the same median
FIG. 8. Dependence of the reciprocal of the Kubelka–Munk function of red dye solutions at 520 nm on droplet concentration.
FIG. 9. Dependence of L, a, b values on droplet and dye concentration for n-hexadecane oil-in-water emulsions containing a red food dye.
droplet diameter (0.3 mm), but different droplet concentrations (0.25 to 38.3 wt%) and red dye concentrations (0 to 0.20 wt%), were measured using a colorimeter (Fig. 9). In the absence of dye, the L-value increased steeply with increasing droplet concentration from 0 to 3 wt%, and then leveled off at approximately L 5 100% at higher concentrations. A similar type of behavior was observed for the emulsions containing dye, except that the initial increase in L-value was less steep as the dye concentration increased, and the final L-value was lower than for the dye-free emulsions. These results indicate that the emulsions became lighter as the droplet concentration increased and the dye concentration decreased. This has important consequences for the formulation of many types of commercial emulsion-based products. For example, in the food industry, there is great emphasis on the development of reduced fat emulsions that have the same desirable attributes as the original product (21, 22). Our results indicate that both the fat and the dye content of a food emulsion must be carefully controlled, since the increase of dye concentration and the reduction of fat content below about 3 wt% leads to an appreciable decrease of the lightness, which would adversely effect its consumer acceptability.
330
CHANTRAPORNCHAI, CLYDESDALE, AND McCLEMENTS
The a- and b-values of the emulsions were also strongly influenced by the droplet and dye concentration (Fig. 9). In the absence of dye, the a-value of all emulsions and the b-values of the concentrated emulsions (f . 5 wt%) were close to zero. At lower droplet concentrations, the b-value of the dye-free emulsions became increasingly negative as the droplet concentration was reduced, which indicated that they become bluer in appearance. The a-values (5 to 50) and b-values (0 to 18) of the emulsions containing dye indicated that they were in the redyellow region of the color space (16). Emulsion color was strongly influenced by droplet and dye concentration. Emulsions became more intensely colored (i.e., a- and b-values moved away from zero) as the droplet concentration decreased (down to about 0.5 wt%) and the dye concentration increased. CONCLUSIONS
The optical properties of concentrated oil-in-water emulsions depend strongly on the concentration of dye and droplets present. The spectral reflectance of emulsions could be described qualitatively using the Kubelka–Munk theory of diffuse reflectance, but the quantitative agreement was fairly poor because of light losses associated with the measurement cell. Nevertheless, we developed an empirical method of correcting the predicted spectra that gave excellent agreement between theory and experimental data. The Kubelka–Munk function increased approximately linearly with dye concentration and was closely related to the absorbance of the dye solutions. The reciprocal of the Kubelka–Munk function increased linearly with droplet concentration at relatively low concentrations (f , 3 wt%), but decreased at higher values because of multiple scattering and droplet– droplet interactions. The information presented in this study would be useful for the formulation of emulsions with specific colors and lightness, e.g., foods, paints, cosmetics, and pharmaceuticals. The dependence of the optical properties of emulsions on particle size suggests that droplet aggregation would have a significant impact on their appearance. For this reason we are currently examining the influence of floc size and structure on the optical properties of concentrated emulsions. It should be noted that diffusing wave spectroscopy (DWS) techniques are being applied more frequently for the characterization of the structure, interactions, and dynamics of concentrated colloidal systems. These techniques also rely on the diffusion of light waves through concentrated colloidal dispersions. Nevertheless, they utilize temporal fluctuations in the intensity of mono-
chromatic scattered light caused by particle diffusion to provide information about the system, whereas the optical reflectance spectroscopy technique used in this study utilizes measurements of the wavelength dependence of the timeaveraged back-scattered light (23). ACKNOWLEDGMENTS The authors thank the Massachusetts Agricultural Experiment Station (Project MAS00745) for supporting this work.
REFERENCES 1. van de Hulst, H. C., “Light Scattering by Small Particles.” Wiley, New York, 1957. 2. Kerker, M., “The Scattering of Light.” Academic Press, New York, 1969. 3. Kortum, G., “Reflectance Spectroscopy: Principles, Methods, Applications.” Springer-Verlag, New York, 1969. 4. Heimenz, P. C., and Rajagopalan, R., “Principles of Colloid and Surface Science.” Dekker, New York, 1997. 5. McClements, D. J., “Food Emulsions: Principles, Practice and Techniques.” CRC Press, Boca Raton, FL, 1998. 6. Wyszecki, G., and Stiles, W. S., “Color Science: Concepts and Methods, Quantitative Data and Formulas.” Wiley, New York, 1967. 7. Wendlandt, W. W., “Modern Aspects of Reflectance Spectroscopy.” Plenum, New York, 1968. 8. Frei, R. W., and MacNeil, J. D., “Diffuse Reflectance Spectrocopy in Environmental Problem-Solving.” CRC Press, Cleveland, OH, 1973. 9. Frei, R. W., Frodyma, M. M., and Lieu, V. T., in “Comprehensive Analytical Chemistry” (G. Svehla, Ed.), Chap. 3. Elsevier, Amsterdam, 1975. 10. Wickramasinghe, N. C., “Light Scattering Functions for Small Particles.” Wiley, New York, 1973. 11. Bailey, A. E., and Cannell, D. S., Phys. Rev. E. 50, 4853 (1994). 12. Mudget, P. S., and Richards, L. W., Appl. Optics 10, 1485 (1971). 13. Chantrapornchai, W., Clydesdale, F. M., and McClements, D. J., J. Agric. Food Chem. 46(8), 2914 (1998). 14. Chantrapornchai, W., Clydesdale, F. M., and McClements, D. J., Colloids Surf. A. 155, 373 (1998). 15. McClements, D. J., Chantrapornchai, W., and Clydesdale, F. M., J. Food Sci. 63(6), 935 (1998). 16. Hutchings, J. B., “Food Colour and Appearence.” Blackie, London, 1994. 17. Smith, R. C., and Baker, K. S., Appl. Optics 20, 177 (1981). 18. Seyfried, M., and Fukshansky, L., Appl. Optics 22, 1402 (1983). 19. Seyfried, M., Fukshansky, L., and Schafer, Appl. Optics 22, 492 (1983). 20. Ma, Y., Varadan, V. K., and Varadan, V. V., J. Acoust. Soc. Am. 87, 2779 (1990). 21. Akoh, C. C., Food Technol. 52(3), 47 (1998). 22. McClements, D. J., and Demetriades, K., Crit. Rev. Food Sci. Nutr. 38, 511 (1998). 23. Weitz, D. A., Zhu, J. X., Durian, D. J., Gauy, H., and Pine, D. J., Phys. Scr. T49, 610 (1993).