The color of foods dramatically influences consumers'

Transcription

1 452 MELÉNDEZ-MARTÍNEZ ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 89, NO. 2, 2006 FOOD COMPOSITION AND ADDITIVES Influence of White Reference Measurement and Background on the Color Specification of Orange Juices by Means of Diffuse Reflectance Spectrophotometry ANTONIO J. MELÉNDEZ-MARTÍNEZ,ISABEL M. VICARIO,andFRANCISCO J. HEREDIA 1 University of Seville, Laboratory of Food Color and Quality, Department of Nutrition and Food Science, Faculty of Pharmacy, Seville, Spain The color of citrus beverages, in general, is related to the consumers' perception of flavor, sweetness, and other characteristics related to the quality of these products, so it is important to develop suitable methods for accurately assessing this attribute. In this study, the color of 2 different kinds of orange juices (ultrafrozen orange juices and orange juices from concentrate) were measured using different white references in order to demonstrate the influence of white reference measurement on the color parameters obtained. The results of this work indicated that the election of a wrong white reference measurement can affect dramatically the accuracy of the objective specification of orange juice color. Thus, the application of analysis of variance ( = 0.01) revealed that, in most cases, the parameters obtained were significantly different. In terms of color differences, the change of the background for the same white reference affected the measurements less than the changes of white references for the same background. On the other hand, important chromatic differences between the 2 types of orange juices studied were found. The color of foods dramatically influences consumers' preferences. In this sense, some studies revealed that the color of citrus beverages, in general, are related to the consumers' perception of flavor, sweetness, and other characteristics in relation to the quality of these products (1, 2). Due to this fact, many studies have been carried out in order to accurately determine the color of these products. Color measurement of orange juice has developed at the same time as the techniques of colorimetry. Before the development of specific apparatus, this kind of analysis was made by means of comparison with colored plastic standards (3). However, in order to eliminate the subjectivity linked to that kind of measurement, different instruments were designed for the Received June 7, Accepted by SG September 6, Author to whom correspondence should be addressed; heredia@us.es objective evaluation of the color of the product (4). Huggart et al. (5) evaluated the relationship between the color of orange juice, the visual assessment of its quality, and the readings obtained with the Hunter colorimeter. As a result of this and other experiments, many suitable techniques and instruments were developed, of which the Citrus Colorimeter (6) stood out. As a consequence of several studies (7 9), the validity of many other instruments for the objective measurement of orange juice color, some of which had a diffuse reflectance sphere attached, was demonstrated. Recently, the objective assessment of color has also been applied for determining the carotenoid content of orange juice for quality control purposes (10). Independent of the instrument used, it is necessary to make a white reference measurement prior to the color measurements, which is different depending on the technique used. Nevertheless, within the literature there are no specific papers in which the influence of the white reference measurement on the color specification of orange juice is evaluated. In this study, the colors of 2 different kinds of orange juices were measured with a diffuse reflectance sphere attached to a spectrophotometer using different white references. The 2 main aims were, firstly, to study how the different white references tested affect the color parameters of the uniform color space CIELAB (11). On the other hand, this work was also aimed at evaluating the color differences between the 2 types of orange juices studied. Juice samples were measured against a white background and a black background, because some studies have shown that sometimes reflectance measurements against a black background (12 15) are better correlated with visual assessment of color and other parameters. Experimental Samples Twenty-three orange juices were analyzed, 15 of which were orange juices from concentrate (OJFCs), and 8 ultrafrozen orange juices (UFOJs). OJFCs, which are subjected to high temperatures industrially, were purchased locally, whereas UFOJs were kindly provided by Zumos Vitafresh (Almonte, Spain). In the

2 MELÉNDEZ-MARTÍNEZ ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 89, NO. 2, industry, the latter juices are cooled and immediately frozen using liquid nitrogen in an industrial freezing tunnel; that is, they are never subjected to high temperatures. The commercial samples are kept in freezing chambers between 18 and 21 C until distribution. The samples in this study were kept at 21 C until their analysis. Thawing was carried out by means of a Samsung (Ridgefield Park, NJ) microwave oven at 800 watts for 11 min. Color Measurement Color measurements were made by means of diffuse reflectance spectrophotometry. For this purpose, an ultraviolet-visible (UV-Vis) diode-array spectrophotometer Model HP-8453 (Hewlett-Packard, Palo Alto, CA), with a Labsphere diffuse reflectance sphere Model RSA-HP-53 (Labsphere Inc., North Sutton, NH) attached, was used. Samples were placed into plastic cuvettes ( mm) for reflection measurements. The whole visible spectrum ( nm) was registered ( =1nm), and Illuminant D65 and 10 Observer were considered as references. The color parameters corresponding to the uniform color space CIELAB were obtained by means of the software PCROM (16). Within the uniform color space CIELAB, 2 color coordinates, a* and b*, as well as a psychometric index of lightness, L*, are defined. Coordinate a* takes positive values for reddish colors and negative values for the greenish ones, whereas b* takes positive values for yellowish colors and negative values for the bluish ones. L* is an approximate measurement of relative luminosity, which is the property according to which each color can be considered as equivalent to a member of the grey scale, between black and white, taking values within the range From L*, a*, and b*, new parameters are defined, such as chroma (C* ab ) and hue (h ab ). Chroma (C* ab ) is the attribute that allows each hue to be determined by its degree of difference in comparison to a grey color with the same lightness, so it is considered the quantitative attribute of colorfullness. Hue angle (h ab ) is the attribute according to which colors have been traditionally defined as reddish, greenish, etc. It is the attribute that allows a color to be distinguished with respect to a grey color with the same lightness. This attribute is related to the differences in absorbance at different wavelengths and is considered the qualitative attribute of color: ab 2 2 C* a * b* (1) h arctan b* ab a* The angles are given in degrees using the following conventions: 0 <h ab <90, if a* > 0, b* > 0 90 <h ab < 180, if a* < 0, b* > <h ab < 270, if a* < 0, b* < <h ab < 360, if a* > 0, b* < 0 (2) Table 1. Subscripts referring to the background and white reference used for color measurement Color differences are calculated as the Euclidean distance between 2 points in the 3-dimensional space defined by L*, a*, and b*: E* L* a* b* ab In this study, a pressed plate of BaSO 4 (Labsphere USRS ) has been considered as the standard white reference because it is the one recommended for the instrument. For white reference measurement, the pressed plate was placed in the plane of the aperture. However, due to the fact that orange juice is a translucent aqueous product and 2 backgrounds (white and black pieces placed on the back of the cuvette) were used for the color measurements, plastic cuvettes filled with water and in contact to the 2 backgrounds were also considered as white references. Immediately after the white reference measurements, samples were measured against the white and the black backgrounds, using for this purpose homogeneous materials. As white background, a square cork piece whose spectrum was almost identical to that corresponding to the BaSO 4 pressed plate was used, in order to avoid the deterioration of the pressed plate. As black background, a round plastic piece was used. Six subscripts have been used to refer to the color parameters as a function of the white reference and the background used for the color measurements, as shown in Table 1. Results and Discussion White references Background BaSO 4 background H 2O+Black H 2O+White background Black White Average values and standard deviations (SDs) of the color variables as a function of white reference measurement and background are shown in Table 2; results corresponding to the Situations 3 and 4 (Table 1) have been omitted because incoherent results were obtained. Thus, when the white reference measurement was made with water, average values of the color variables for OJFCs were L* = , a* = 2.84, b* = 73.18, C* ab = 73.30, and h ab = with black background, and L* = , a* = 9.13, b* = 93.28, C* ab = 93.84, and h ab = with white background. For UFOJs, average values obtained were L* = , a* = 12.13, b* = , C* ab = , and h ab = with black background, and L* = , a* = 24.38, b* = , C* ab = , and h ab = with white background. It can be observed that anomalous values, especially for L*, were obtained. These incoherent results are related to the white (3)

3 454 MELÉNDEZ-MARTÍNEZ ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 89, NO. 2, 2006 Table 2. Average values SD of the color variables Backgrounds blanks L* a* b* C* ab h ab OJFCs UFOJs reference itself, due to the fact that, as a result of the white reference measurement, a black material behind a transparent liquid is assigned a value of L* = 100 by the apparatus, although in fact there is no reflection and L* should be equal to 0. Consequently, the samples of orange juice, which were indeed lighter than the white reference, were assigned higher values, which were in fact out of the range defined for L* (0 100). As a consequence, measurements made with this white reference were erroneous and have been omitted. It can be observed in Table 2 that the parameter most affected by the type of white reference is L*. When the white reference measurement was made with water and white background, L*, a* (in absolute terms), b*, and C* ab values were higher than those obtained with the standard white reference measurement. These differences can be explained by taking into consideration that the use of the cuvette filled with a transparent liquid for white reference measurement changes the geometry of the system, the path length being 10 mm longer. In relation to this, it must be taken into account that, if the orange juices analyzed were almost completely opaque, light would be reflected by the external layers of the samples, which is what takes place when the white reference measurement is made with the BaSO 4 reference plate placed in the plane of the aperture. On the other hand, if the juices were almost transparent, reflection would take place mainly on the white background, and white reference measurements should be made with the cuvette filled with water + white background. In the particular case of orange juices, h ab, which is considered the qualitative index of color, and a* (red-green axis) are useful parameters to have preliminary information of their color, although, as is well-known, L*, a*, and b* or L*, h ab, and C* ab must be considered together to specify completely and objectively any color, due to the tridimensional nature of color. Coordinate a* vs hue diagrams are shown in Figures 1 and 2. To test whether the use of the 2 different white references led to significant differences between the color parameters, analysis of variance (ANOVA) for repeated measurements ( = 0.01) were applied. Results are summarized in Table 3. Figure 1. a* versus h ab plane for OJFCs.! White Figure 2. a* versus h ab plane for UFOJs.! White

4 MELÉNDEZ-MARTÍNEZ ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 89, NO. 2, Table 3. Results of the ANOVA test for repeated measurements a Backgroundsblanks L* a* b* C* ab h ab OJFCs UFOJs a P values ( = 0.01); numbers in italics indicate the existence of significant differences. In the case of OJFCs, results showed that there were significant differences between all the parameters, except for h ab, when the white background was used for the measurements (p = 0.686). For UFOJs, ANOVA results indicated that there were significant differences between all the parameters as a function of the white reference made, with the exception of a* and h ab, when measurements were made with black background (p = and 0.011, respectively). These findings may show that the qualitative attribute of color, h ab, is less affected by the different white references tested. Location of the samples within the (a*b*) diagram is shown in Figures 3 and 4. For OJFCs it can be clearly observed that, when the white background was used, all of the samples were located within the first quadrant (positive values of a* and b*). When the measurements were made against the black background, the samples were located in the second quadrant (negative values of a* and positive values of b*). As for b* values, it can be clearly seen that, independent of the background used for the measurements, when the white reference was water + white background, they increased slightly with respect to those obtained for the white reference BaSO 4. Regarding UFOJs, all of the samples were located in the first quadrant, both for measurements made against the white and the black backgrounds. The evolution of b* values as a function of the white reference measurements can be explained in the same terms as for OJFCs. Regarding a* values, it was observed that they increase slightly from the measurements made with the white reference BaSO 4 to those made with the white reference water + white background. Independent of the white reference and the background used, a* and b* values were higher for UFOJs, indicating that these samples were redder and more yellow than OJFCs. Diagrams of C* ab versus h ab are shown in Figures 5 and 6. These diagrams are quite different, depending on the type of orange juice considered. For OJFCs, the differentiation between the measurements made against white and black backgrounds is clear. When the black one is used, samples were located in the interval of hue values , almost linearly. On the other hand, when the measurements were made against the white background, the interval of hue values was wider, and they were always under 90. Regarding UFOJs, the separation among the samples within the diagram as a function of the white references and the backgrounds was not so clear. Chroma values were higher in UFOJs independent of the white Figure 3. Location of the OJFCs within the (a*b*) plane.! White reference BaSO 4, black background; O white reference BaSO 4, white background; # white reference H 2 O + white background, black background; G white reference H 2 O + white background, white background. Figure 4. Location of the UFOJs within the (a*b*) plane.! White reference BaSO 4, black background; O white reference BaSO 4, white background; # white reference H 2 O + white background, black background; G white reference H 2 O + white background, white background.

5 456 MELÉNDEZ-MARTÍNEZ ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 89, NO. 2, 2006 Table 4. Average color differences ± SD OJFCs UFOJs E* ab E* ab E* ab E* ab Figure 5. C* ab versus h ab plane for OJFCs.! White reference and background used, indicating that the color of these juices was deeper than the color of OJFCs. Taking into account Figures 3 6, it can be concluded that although there were differences between the measurements made against the black and the white backgrounds, the background affected the color measurements of OJFCs, to a greater degree, at least in qualitative terms. To evaluate quantitatively the effect of the backgrounds and the white reference measurements on the color specification of the orange juices studied, the corresponding CIELAB color differences ( E* ab ) were calculated. It is important to bear in mind that E* ab can be only computed for 2 color samples viewed under the same conditions (that is, with only 1 white reference). Thus, E* ab1 5 or E* ab2 6 are pseudo color differences measured in 2 different color spaces. Figure 6. C* ab versus h ab plane for UFOJ.! White Table 4 shows the average color differences between the different white references with black background ( E* ab1 5 ), the different white references and white background ( E* ab2 6 ), and the different backgrounds considering the same white reference ( E* ab1 2, E* ab5 6 ). It was noticeable that E* ab values for the same pairs were, in general, similar for both types of orange juices, despite the large color differences among them. Taking into account E* ab for the same white reference and different backgrounds ( E* ab1 2, E* ab5 6 ), it was seen that the white reference that led to the lowest color differences was BaSO 4. In the case of OJFCs, the SD values associated to these pairs (4.94 and 7.80, respectively) were larger than those corresponding to UFOJs (0.96 and 4.06, respectively), possibly due to the fact that the samples were more heterogeneous. It can be claimed that the changes of backgrounds for the same white reference had less effect, in terms of color differences, than the changes of white references for the same background. Furthermore, it was also seen that E* ab1 5 E* ab2 6, which indicated that, in the comparisons between the different white references, the background used for the measurements did not affect E* ab values significantly. Conclusions In conclusion, the results of this work indicate that the selection of a wrong white reference measurement can affect dramatically the accuracy of the objective specification of orange juice color. Agood illustration of this is the fact that the change of the background for the same white reference affected E* ab less than the changes of white references for the same background, independent of the type of orange juice considered. It was also particularly noticeable that incoherent values of L* were obtained when the measurements with white reference water black background was made. Due to the opacity of the samples analyzed, the standard white reference seemed to be the most suitable for the measurements. Thus, the correct white reference is the measurement of the BaSO 4 plate at the aperture of the integrating sphere, because the color sample to be measured is the set orange juice + background. The choice of H 2 O + white background as the white reference is incorrect because, in this case, the background is already a part of the color sample. The choice of H 2 O + black background as the white reference is also incorrect as it has been explicitly indicated.

6 MELÉNDEZ-MARTÍNEZ ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 89, NO. 2, In relation to the 2 types of orange juice studied, important color differences were found. Considering the (a*b*) diagrams, it can be clearly seen that, in the case of OJFCs, samples were located in the first or the second quadrant depending on the background used for the measurements, because the use of the black background led to negative a* values. Likewise, plots of the samples as a function of the psychometric parameters, C* ab and h ab, are different depending on the type of juice considered. In this sense, it can be observed that the differences between the h ab values obtained with black and white backgrounds are higher for OJFCs. Acknowledgments We gratefully acknowledge Zumos Vitafresh (Almonte, Spain) for its collaboration in this study. References (1) Huggart, R.L., Fellers, P.J., De Jager, G., & Brady, J. (1979) Proc. Fla. State Hort. Soc. 92, (2) Tepper, B.J. (1993) J. Sensory Stud. 8, (3) Francis, F.J. (1995) Food Qual. Prefer. 6, (4) Huggart, R.L., & Wenzel, F.W. (1955) Food Tech. 9, (5) Huggart, R.L., Barron, R.W., & Wenzel, F.W. (1966) Food Tech. 20, (6) Meléndez-Martínez, A.J., Vicario, I.M., & Heredia, F.J. (2005) J. Sci. Food Agric. 85, (7) Eagerman, B.A. (1978) J. Food Sci. 43, (8) Buslig, B.S., & Wagner, Jr, C.J. (1984) Proc. Fla. State Hort. Soc. 97, (9) Buslig, B.S. (1991) Proc. Fla. State Hort. Soc. 104, (10) Meléndez-Martínez, A.J., Vicario, I.M., & Heredia, F.J. (2003) J. Agric. Food Chem. 51, (11) Commission Internationale d'eclairage (CIE) (1986) Colorimetry, 2nd Ed., Publication CIE No. 15.2, CIE Central Bureau, Vienna, Austria (12) Meléndez-Martínez, A.J., Vicario, I.M., & Heredia, F.J. (2005) Food Qual. Prefer. 16, (13) Huang, I.-L., Francis, F.J., & Clydesdale, F.M. (1970) J. Food Sci. 35, (14) Huang, I.-L., Francis, F.J., & Clydesdale, F.M. (1970) J. Food Sci. 35, (15) Terrab, A., Díez, M.J., & Heredia, F.J. (2003) Food Sci. Tech. Int. 8, (16) Álvarez, C., & Heredia, F.J. (1994) PCROM Cálculos Colorimétricos, v. 2.0., Spanish patent SE-2374