Color of orange juices in relation to their carotenoid contents as assessed from different spectroscopic data

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This is an Accepted Manuscript of an article published by Elsevier in Journal of Food Composition and Analysis on September 2011, available at: https://doi.org/10.1016/j.jfca.2011.05.001 .” 1 1 COLOR OF ORANGE JUICES IN RELATION TO THEIR 2 CAROTENOID CONTENTS AS ASSESSED FROM DIFFERENT 3 SPECTROSCOPIC DATA 4 Running title: Orange juice color as assessed from different spectra 5 A.J. Meléndez-Martínez1, L. Gómez-Robledo2, M. Melgosa2, I.M. Vicario1 and F.J. 6 Heredia1 * 7 1 Food Colour & Quality Lab., Dept. Nutrition & Food Science. Universidad de 8 Sevilla. Facultad de Farmacia, 41012 Sevilla, Spain ([email protected]). 9 2 Department of Optics, Faculty of Sciences (Mecenas Building), University of 10 Granada, 18071 Granada, Spain 11 12 *Author to whom correspondence should be addressed 13 Francisco J. Heredia Mira 14 Food Colour & Quality Lab., Dept. Nutrition & Food Science. Universidad de 15 Sevilla. Facultad de Farmacia, 41012 Sevilla, Spain 16 Telephone: ++34 9545 56761; Fax: ++ 34 9545 57017 17 e-mail: [email protected] 18 19 2 20 ABSTRACT 21 22 The color of orange juices influences , hence the 23 assessment of this attribute has been paid much importance for decades. The 24 instrumental measurement of orange juices color can be harnessed for the rapid, 25 quality-control oriented estimation of its carotenoids, which are health-promoting 26 compounds. In this work we compared and analyzed different spectroscopic data that 27 can be used for these purposes, like the reflectance spectra of orange juices and the 28 absorption spectra of their carotenoid extracts. Additionally, we have revisited the 29 Kubelka-Munk theory and have concluded that its parameters are suitable to assess 30 the carotenoid content of the samples, but not as much as the CIELAB color 31 coordinates. In this regard, we have observed that K/S is the Kubelka-Munk 32 parameter better correlated with the carotenoid content of the orange juices surveyed 33 (r = -0.84), although better correlations were observed when the CIELAB coordinate 34 a* was considered (r = 0.86 and 0.88 for measurements with white and black 35 background, respectively). However, in our opinion this fact should not lead to 36 dismiss the usefulness of the Kubelka-Munk theory to estimate carotenoid contents, 37 since its application can lead to obtain valuable information about the absorption and 38 scattering properties of the samples. 39 40 41 KEYWORDS: Absorption spectra; carotenoids; color; Kubelka-Munk; orange juice; 42 reflectance spectra. 43 44 45 3 46 INTRODUCTION 47 The color of orange juices (OJ) 48 deal (Tepper, 1993). The hues of most OJ range from yellow to orange and are 49 mainly due to their content in carotenoids. However, other oranges also accumulate 50 anthocyanin pigments such that their juices exhibit a characteristic reddish color 51 (Arena, Fallico, Maccarone, 2000, Kirca Cemeroglu, 2003, Meléndez-Martínez, 52 Vicario, Heredia, 2005). Recently, two red-fleshed orange mutants that owe their 53 reddish color mainly to the carotenoid lycopene and not to anthocyanins have also 54 been described and studied (Lee, 2001, Liu et al., 2007, Alquezar, Rodrigo, Zacarías, 55 2008). Anyhow, the genotype is not the only factor determining the carotenoid 56 pattern of oranges and therefore their color, since the climate of the area of 57 production and the industrial processing, among others, are also related to the 58 pigment content (Mouly, Gaydou, Lapierre, Corsetti, 1999, Meléndez-Martínez, 59 Britton, Vicario, Heredia, 2008). 60 The citrus industry has been aware of the importance of the color of OJ since 61 long ago, especially in the United States, where this attribute has been long used as a 62 quality parameter for the commercial classification of the product. In relation to this 63 it is important to mention that much research has been conducted in this country to 64 standardize its assessment to the extent that many instruments have been evaluated, 65 and that even a color number scale was developed for the rapid classification of the 66 juices (Huggart Wenzel, 1954, Huggart, Barron, Wenzel, 1966, Hunter, 1967, 67 Eagerman, 1978, Buslig Wagner Jr., 1984, Buslig Wagner Jr., 1985, Meléndez68 Martínez et al., 2005). Anyway, in our opinion, internationally recommended color69 specification systems, like CIELAB (CIE, 2004), must be always employed with 4 - - 70 independence we consider OJ, virgin olive oils (Moyano, Melgosa, Alba, Hita, 71 Heredia, 1999) or any other product. 72 Apart from its evident relationship with the perceived quality and the 73 es, with the derived economic repercussions, the study of the color 74 of OJ is especially challenging for several reasons. On one hand, the carotenoid 75 pattern of OJ is very complex and includes, among others, pigments with provitamin 76 A activity -cryptoxanthin) and other biological 77 functions or actions (besides the former, lutein, zeaxanthin and, in some genotypes, 78 lycopene). This is important because the instrumental measurement of the color of 79 OJ, apart from serving to assess this important quality attribute, could also be 80 harnessed for the rapid, quality-control oriented estimation of these phytochemicals. 81 In this sense, the application of tristimulus colorimetry in conjunction to multivariate 82 statistical methods has proved a powerful tool that allows the determination of 83 individual carotenoids (Meléndez-Martínez, Vicario, Heredia, 2003), and even 84 hypothetical vitamin A activity (Meléndez-Martínez, Vicario, Heredia, 2007b) in OJ. 85 In other studies carried out in our laboratory, multivariate statistical methods have 86 been applied to gain insight into how the diverse carotenoids occurring in OJ 87 produce a final color (Melendez-Martinez et al., 2010). 88 The instrumental measurement of the color of these products is also 89 especially challenging because they are neither transparent nor opaque, but 90 translucent. Whilst the color of transparent products can be readily ascertained from 91 transmission measurements and that of opaque ones from reflection measurements, 92 the behavior of light when reaches a translucent sample makes the instrumental 93 measurement more complicate (Meléndez-Martínez et al., 2005). Over and above 5 94 this fact, the characteristic turbidity of OJ also plays an important role in its 95 appearance, which is also an important factor to be considered when measuring its 96 color (Rummens, 1970, Arena et al., 2000). The importance of the pulp particles in 97 this regard is double, as they also contain the carotenoid pigments (Meléndez98 Martínez, Vicario, Heredia, 2009). 99 Although the color definition of OJ continues attracting the interest of both 100 scientists and the citrus industry in the first years of the 21st century (Arena et al., 101 2000, Lee Castle, 2001, Choi, Kim, Lee, 2002, Lee Coates, 2003, Meléndez102 Martínez, Vicario, Heredia, 2004, Meléndez-Martínez, Britton, Vicario, Heredia, 103 2005, Pérez-López, Beltran, Serrano-Megías, Saura López, Carbonell-Barrachina, 104 2006, Cortes, Esteve, Frigola, 2008, Tiwari, Muthukumarappan, O' Donnell, Cullen, 105 2008), little attention has been paid to methodological aspects. In this paper we 106 analyze different spectroscopic approaches that can be used to define the color of the 107 juices with a view to determine their pigment content as the correlation between the 108 color of several foodstuffs and their carotenoid content is raising much interest 109 (Meléndez-Martínez et al., 2003, Humphries, Graham, Mares, 2004, Ruiz, Reich, 110 Bureau, Renard, Audergon, 2008). In this regard, we discuss the absorbance spectra 111 of OJ carotenoid extracts (which has been used for very long to quantitate them), the 112 reflectance spectra of OJ (which have proved very useful to estimate carotenoid 113 levels without having to extract them) and spectra derived from the application of the 114 Kubelka-Munk theory, which has been used for translucent samples and whose 115 application to the estimation of OJ carotenoid levels has not been assessed in detail. 116 Moreover, some observations regarding the use of black or white 117 backgrounds for the color measurement of the juices are made. 6 118 MATERIAL AND METHODS 119 Samples 120 Seventy commercially available OJ were surveyed. The samples were 121 divided in two main groups according to the industrial treatment undergone: 122 ultrafrozen orange juices (UFOJ, juices not subjected to heat treatment that are 123 stored and marketed at T < 18°C; , n = 26) and thermally-treated orange juices 124 (TTOJ, n = 44). The UFOJ were supplied at different intervals by Zumos Vitafresh 125 (Almonte, Huelva, Spain), whilst the TTOJ were purchased from several retailers in 126 Seville at different times. All the samples were stored as recommended (the UFOJ at 127 T < -18°C and the TTOJ at room temperature or 4°C) until their analysis. UFOJ 128 were thawed at room temperature. 129 Instrumental color measurement 130 For the color readings, the samples were placed in a plastic cuvette (47.5 35 131 10 mm) for reflection measurements. The measurements were carried out with 132 subdued illumination to avoid possible light interferences. Moreover, the cuvettes 133 were placed inside a cabin with grey walls. The reflectance visible spectra (380-770 134 nm, = 1 nm) were recorded by means of a CAS 140 B spectroradiometer 135 (Instrument Systems, Munich, Germany) fitted with a Top 100 telescope optical 136 probe (Instrument Systems, Munich, Germany) and a Tamron zoom mod. SP 23A 137 (Tamron USA, Inc., Commack, NY, USA). The zoom, to which the probe was 138 attached, was held at a fixed distance of 50 cm in a straight line from the sample. For 139 all the measurements the source of light was an external incandescent lamp providing 140 a 45o incident illumination with respect to the perpendicular to the cuvette. The 141 apparatus was set to take three consecutive readings, so the color parameters 7 142 obtained were averages of three measurements. The instrument blank measurements 143 were made with the cuvette filled with distilled water against a reference white 144 pressed plate (SRS-99-010, Labsphere Inc. North Sutton, NH, USA). The OJ 145 samples were measured against a white background (WB, the pressed white plate) 146 and a black background (BB, a round plastic piece with homogeneous black color). 147 The illuminant D65 and the 10° Standard Observer were considered as 148 references. The illuminant D65 is a standard illuminant defined by International 149 Commission on Illumination and the 10° Standard Observer was considered to 150 represent best average spectroscopic response of human observers 151 The color coordinates corresponding to the approximately uniform color 152 space CIELAB (CIE, 1978) were obtained directly from the apparatus. Within this 153 color space, two color coordinates, a*10 and b*10, and a psychometric index of 154 lightness, L*10, are defined. a*10 takes positive values for reddish colors and negative 155 values for the greenish ones, whereas b*10 takes positive values for yellowish colors 156 and negative values for the bluish ones. L*10 is an estimation of the relative 157 luminosity, and according to this parameter any given color can be regarded as 158 equivalent to a member of a grey scale, between black (L*10 = 0) and white (L*10 = 159 100). From a*10 and b*10, the psychometric parameters chroma (C*ab,10) and hue160 angle (hab,10) are defined: 161 C*ab,10 = [(a*10)2 + (b*10)2]1/2 (Eq. 1) 162 hab,10 = tan-1 (b*10/a*10) (Eq. 2) 163 C*ab,10 (related to the quantitative attribute of colorfulness) allows to 164 determine for each hue its degree of difference in comparison to a grey color with the 165 same lightness, whilst hue-angle (hab,10) is the attribute according to which colors 166 have been traditionally defined as reddish, greenish, etc. 8 167 168 Application of the Kubelka-Munk theory to the spectroscopic data 169 The theory can be briefly summarized into several main points (Calvo, 1993): 170 - The layer of sample can be divided into elementary layers with parallel 171 sides to the total layer and identical optical properties. 172 - The elementary layer or sheet is defined as a sheet with parallel and infinite 173 sides, so that the effect of the edges is eliminated. However its thickness is finite. 174 - The thickness of the elementary sheet is small to the total thickness of the 175 sample, but large compared to the size of the particles. 176 - At any wavelength, the optical properties of the sample can be described by 177 one scattering (S) and one absorption (K) coefficient. Such coefficients describe the 178 amount of light scattered and absorbed when passing through the sample. 179 - There are an ascending flow and a descending flow of diffuse light. Each 180 flow loses on its way through the sample an amount of light proportional to its 181 energy and to K, due to absorption and another amount of light proportional to its 182 energy and to S, owed to scattering. Likewise, each flow gains an amount of light 183 proportional to the energy of the other one and to S, due to scattering. 184 Considering that i is the descending flow and j the ascending one. The 185 following equations can be deduced: 186 (Eq. 3) 187 dx + Sj dx (Eq. 4) 188 These equations, in turn, lead to the following formulas (Judd, Wyszecki, 189 1975, Wyszecki, Stiles, 1982, Hutchings, 1994): 190 a = 1/2[R + (R0 + Rg /R0 Rg )] = (S + K )/S (Eq. 5) 191 b = (a 1/2 (Eq. 6) 15 339 extract them, which offers many advantages (more efficient throughput of samples, 340 little risk of pigment degradation, etc.). 341 The usefulness of the CIELAB color coordinates to estimate the individual or 342 total carotenoid content of OJ has been demonstrated in some of our previous works 343 (Meléndez-Martínez, Vicario, Heredia, 2003, Meléndez-Martínez, Vicario, Heredia, 344 2007). The average CIELAB color coordinates of the samples analyzed are displayed 345 in Table 1 with their corresponding standard deviations, whilst the correlation 346 coefficients between them and the total carotenoid content of the samples are shown 347 in Table 2. All the correlations were significant at p<0.05, the highest values of linear 348 correlation coefficient r (over 0.8) corresponding to a*10 regardless of the 349 background used for the measurements. The differences in the carotenoid levels and 350 color coordinates within and between types of orange juices can be easily explained 351 considering that it is well-known that the accumulation of secondary metabolites in 352 general and of carotenoids in particular in plants depend on the genotype, climate 353 and agronomic factors and type of processing and storage conditions, among others 354 factors. Considering the averaged color data, it can be claimed that, overall, the 355 TTOJ appear lighter and less vivid than the UFOJ. 356 357 Due to the interesting similarities found between the reflectance spectra of 358 carotenoid-containing samples and some of the spectra obtained considering the 359 Kubelka-Munk theory commented before, it appeared interesting to assess the 360 usefulness of the Kubelka-Munk parameters to assess the carotenoid content of the 361 juices analyzed in this study. The results of this assessment are shown in Figure 4, 362 where the linear correlation coefficient r between the Kubelka-Munk parameters (at 363 500nm, wavelength where K /S is approximately 1) and the carotenoid content of 16 364 each OJ is shown. When comparing the values of r with the ones displayed in Table 365 2, it can be readily concluded that they are similar but not higher than the highest 366 found with CIELAB color parameters. This seems to be indicative that the Kubelka367 Munk parameters (above all K/S, which was found to be better correlated to the 368 pigment content than K and S considered individually) can be used to estimate the 369 carotenoid content of OJ, although they are not as suitable as the CIELAB 370 colorimetric coordinates. Apart from these observations it has to be considered that 371 the application of the Kubelka-Munk theory requires two measurements (with WB 372 and BB), a more complex mathematical treatment and involves more error sources 373 than those linked to a CIELAB colorimetric measurement. It is interesting to note 374 that there are large differences in the correlations between the K/S index and the 375 carotenoid levels as a function of the juices considered (r = 0.44 and 0.75 for UFOJ 376 and TTOJ, respectively). This could reflect marked differences in the absorption and 377 scattering of light between UFOJ and TTOJ, which could in turn be related to the 378 characteristics of their pulp particles, these being interesting aspects to be addressed 379 in future research. 380 Despite the Kubelka-Munk theory has been widely applied to foods, paints, 381 biological material, etc. (Huang, Francis, Clydesdale, 1970, Hetherington Mac 382 Dougall, 1992, Calvo, 1993, Yang, Celmer, Koutcher, Alfano, 2000, Berns 383 Mohammadi, 2007) for several purposes, in the case of orange-based drinks and 384 other foodstuffs sometimes it does not lead to higher correlations with other 385 parameterss as compared to other methodologies (Rummens, 1970, Gullett, Francis, 386 Clydesdale, 1972). In this regard, as we reviewed some years ago (Meléndez387 Martínez, Vicario, Heredia, 2005), some authors reported that the use of the color or 388 spectroscopic information obtained with only one background could be more 17 389 valuable for some purposes (Little, 1964, Durán, Rodrigo, Alcedo, 1976, Lafuente, 390 Gasque, Nieto, Izquierdo, 1979). Considering the spectra represented in Figure 1 the 391 main differences, in spectroscopic terms, derived from the use of the WB and the BB 392 can be readily inferred. Thus, it is evident that the spectra of the juices are largely 393 very similar within the interval of wavelengths from 380 nm to ca. 550 nm, the 394 region of the electromagnetic spectrum where the OJ carotenoids absorb maximally. 395 From ca. 550 onwards, the reflectances of the samples measured with the BB are 396 much lower. This is logical since the light that reaches the BB is not reflected, which 397 can occur when the WB is used. From our previews studies on the matter some other 398 conclusions derived from the use of either background can be drawn. For instance, 399 we have reported that the use of a black background led to a better discernment of OJ 400 dilutions both visually and instrumentally (Meléndez-Martínez, Vicario, Heredia, 401 2004). On the other hand, the use of either background did not affect much the 402 correlations between individual carotenoids and the CIELAB color coordinates 403 (Meléndez-Martínez, Vicario, Heredia, 2003). In the case of the correlation between 404 those color coordinates and the retinol activity equivalents (parameter used to assess 405 the potential vitamin A activity of a product) of OJ, it was concluded that the use of 406 the BB led to slightly better values of r. At this point, it is pertinent to stress that the 407 actual vitamin A activity of the juices depend on many factors (human genotype, 408 food matrix, carotenoid isomeric form, etc.) and that therefore, we are very far to use 409 color data to accurately estimate it. However, since the objective measurement of 410 color has proved useful to determine the hypothetical vitamin A activity, its 411 applicability in food labelling appears very promising (Meléndez-Martínez, Vicario, 412 Heredia, 2007). 413 CONCLUDING REMARKS 18 414 Despite the color of OJ continues attracting much attention from scientists 415 and the industry, few studies have dealt with methodological aspects in the last years. 416 In this paper we have commented the features of different spectra that can be used to 417 evaluate the color of these products and of the pigments accounting for it. 418 Furthermore we have applied the Kubelka-Munk theory to the spectroscopic data to 419 ascertain whether K , S or K /S can lead to better correlations with the carotenoid 420 content of the samples as compared to those obtained with the color coordinates of 421 the CIELAB space. In this sense, we have concluded that K/S is the Kubelka-Munk 422 parameter better correlated with the carotenoid content of orange juices subjected to 423 different processing conditions, although it does not involve a better estimation of the 424 carotenoid content. In conclusion, it can be stated that, apart from the traditional 425 quantification of total carotenoid contents by means of the absorption spectra of their 426 extracts, it is also possible to use the color coordinates (like a*) computed from the 427 reflectance spectra of the product for that purpose. This has the advantage that the 428 carotenoids do not have to be extracted, so the assessment is more rapid and risks of 429 degradation and/or formation of artifacts are minimized. Additionally, good 430 correlations (r > 0.8 in absolute values) were observed between the K/S index and the 431 total carotenoid levels, which indicate that this Kubelka-Munk parameter also seems 432 useful for the quantification of OJ carotenoids. Although the correlation of this 433 parameter with the total carotenoid level was slightly lower as compared to a* (r > 434 0.85), the applicability of this theory for that purpose should not be dismissed since it 435 offers the possibility to obtain valuable data about the absorption and scattering 436 properties of the samples. 437 438 ACKNOWLEDGMENTS 19 439 Research project FIS2007-64266, Ministerio de Educación y Ciencia (España), with 440 European Regional Development Fund (ERDF) support. 441 442 REFERENCES 443 Alquezar, B., Rodrigo, M. J., Zacarias, L. (2008). Regulation of carotenoid 444 biosynthesis during fruit maturation in the red-fleshed orange mutant Cara 445 Cara. Phytochemistry, 69(10), 1997-2007. 446 Arena, E., Fallico, B., Maccarone, E. (2000). Influence of carotenoids and pulps on 447 the color modification of blood orange juice. Journal of Food Science, 65(3), 448 458-460. 449 Berns, R. S.,& Mohammadi, M. (2007). Single-constant simplification of Kubelka450 Munk turbid-media theory for paint systems - A review. Color Research and 451 Application, 32(3), 201-207. 452 Britton, G. (1995a). Structure and properties of carotenoids in relation to function. 453 The FASEB Journal, 9, 1551-1558. 454 Britton, G. (1995b). UV/Visible Spectroscopy. In G. Britton, S. Liaaen-Jensen, & H. 455 Pfander (Eds.), Carotenoids. Volume 1B: Spectroscopy, (pp. 13-62) Basel, 456 Switzerland: Birkhäuser. 457 Buslig, B. S.,& Wagner Jr., C. J. (1984). General purpose tristimulus colorimeters for 458 color grading orange juice. Proc.Fla.State Hort.Soc., 97, 74-76. 459 Buslig, B. S.,& Wagner Jr., C. J. (1985). Instrumental measurement of orange juice 460 color. Food Technology, 39(9), 95-97. 20 461 Calvo, C. (1993). La medida del color en alimentos translúcidos: teoria de Kubelka462 Munk. Rev.Esp.Cienc.Tecnol.Aliment., 33 (6), 597-605. 463 Choi, M. H., Kim, G. H., Lee, H. S. (2002). Effects of ascorbic acid retention on 464 juice color and pigment stability in blood orange (Citrus sinensis) juice 465 during refrigerated storage. Food Research International, 35, 753-759. 466 CIE (1978). Recommendations on Uniform Color Spaces, Color-Difference 467 Equations, Psychometric Color Terms, CIE Publication No. 15 (E-1.3.1) 468 1971, Supplement 2. Vienna: CIE Central Bureau. 469 CIE (2004). CIE Publication 15:2004. Colorimetry. 3rd Edition. Vienna: CIE 470 Central Bureau. 471 Cortes, C., Esteve, M. J., Frigola, A. (2008). Color of orange juice treated by High 472 Intensity Pulsed Electric Fields during refrigerated storage and comparison 473 with pasteurized juice. Food Control, 19(2), 151-158. 474 Durán, L., Rodrigo, M., Alcedo, M. J. (1976). Medida del color y de la turbiedad en 475 zumo de naranja. Rev.Agroquim.Tecnol.Aliment., 16(1), 98-105. 476 Eagerman, B. A. (1978). Orange juice color measurement using general purpose 477 tristimulus colorimeters. Journal of Food Science, 43, 428-430. 478 Fish, W. W. (2006). Interaction of Sodium Dodecyl Sulfate with Watermelon 479 Chromoplasts and Examination of the Organization of Lycopene within the 480 Chromoplasts. Journal of Agricultural and Food Chemistry, 54(21), 8294481 8300. 21 482 Francis, F. J. (1987). Food colorimetry: Measurement and interpretation. In R. 483 Jowitt, F. Escher, M. Kent, B. McKenna, & M. Roques (Eds.), Physical 484 properties of foods-2., (pp. -) London: Elsevier Appl. Sci. 485 Gullett, E. A., Francis, F. J., Clydesdale, F. M. (1972). Colorimetry of foods: orange 486 juice. Journal of Food Science, 37, 389-393. 487 Hetherington, M. J.,& Mac Dougall, D. B. (1992). Optical properties and appearance 488 characteristics of tomato fruit (Lycopersicon esculetum). Journal of the 489 Science of Food and Agriculture, 59 (4), 537-543. 490 Huang, I.-L., Francis, F. J., Clydesdale, F. M. (1970). Colorymetry of foods. 2. Color 491 measurement of squash puree using the Kubelka-Munk concept. Journal of 492 Food Science, 35, 315-317. 493 Huggart, R. L., Barron, R. W., Wenzel, F. W. (1966). Evolution of the Hunter Citrus 494 Colorimeter for measuring the color of orange juice. Food Technology, 20, 495 109-111. 496 Huggart, R. L.,& Wenzel, F. W. (1954). Measurement and control of color of orange 497 concentrate. Proc.Fla.State Hort.Soc., 67(210), 216-. 498 Humphries, J. M., Graham, R. D., Mares, D. J. (2004). Application of reflectance 499 color measurement to the estimation of carotene and lutein content in wheat 500 and triticale. Journal of Cereal Science, 40, 151-159. 501 Hunter, R. S. (1967). Development of the Citrus Colorimeter. Food Technology, 21, 502 906-911 22 503 Hutchings, J. B. (1994). Food color and appearance. Glasgow: Blackie Academic 504 Professional. 505 Judd, D. B., Wyszecki, G. (1975). Color in business, science and industry, 3rd ed. 506 New York: John Wiley Sons. 507 Kirca, A., Cemeroglu, B. (2003). Degradation kinetics of anthocyanins in blood 508 orange juice and concentrate. Food Chemistry, 81, 583-587. 509 Kön, S., Kolbe, H., Korger, M., Köpsel, C., Mayer, B., Auweter, H., Lüddecke, E., 510 Bettermann, H., Martin, H.-D. (2008). Aggregation and interface behaviour 511 of carotenoids. In G. Britton, S. Liaaen-Jensen, & H. Pfander (Eds.), 512 Carotenoids. Volume 4: Natural functions, (pp. 53-98) Basel, Switzerland: 513 Birkhäuser. 514 Lafuente, B., Gasque, F., Nieto, P., Izquierdo, L. (1979). Nota. Métodos para obtener 515 mezclas de zumos de naranja de color preestablecido. 516 Rev.Agroquim.Tecnol.Aliment., 19(4), 549-553. 517 Lee, H. S.,& Coates, G. A. (2003). Effect of thermal pasteurization on Valencia 518 orange juice color and pigments. Lebensmittel-Wissenschaft und 519 Technnologie-Food Science and Technology, 36, 153-156. 520 Lee, H. S. (2001). Characterization of carotenoids in juice of red navel orange (Cara 521 Cara). Journal of Agricultural and Food Chemistry, 49, 2563-2568. 522 Lee, H. S.,& Castle, W. S. (2001). Seasonal changes of carotenoid pigments and 523 color in Hamlin, Earlygold, and Budd Blood orange juices. Journal of 524 Agricultural and Food Chemistry, 49, 877-882. 23 525 Little, A. C. (1964). Color Measurement of Translucent Food Samples. Journal of 526 Food Science, 29, 782-789. 527 Liu, Q., Xu, J., Liu, Y., Zhao, X., Deng, X., Guo, L., Gu, J. (2007). A novel bud 528 mutation that confers abnormal patterns of lycopene accumulation in sweet 529 orange fruit (Citrus sinensis L. Osbeck). Journal of Experimental Botany, 530 58(15-16), 4161-4171. 531 Meléndez-Martínez, A. J., Vicario, I. M., Heredia, F. J. (2003). Application of 532 Tristimulus Colorimetry to estimate the carotenoids content in ultrafrozen 533 orange juices. Journal of Agricultural and Food Chemistry, 51(25), 7266534 7270. 535 Meléndez-Martínez, A. J., Vicario, I. M., Heredia, F. J. (2004). Correlation between 536 visual and instrumental color measurements of orange juice dilutions. Effect 537 of the background. Food Quality and Preference, 16, 471-478. 538 Meléndez-Martínez, A. J., Vicario, I. M., Heredia, F. J. (2005). Instrumental 539 measurement of orange juice color: A review. Journal of the Science of Food 540 and Agriculture, 85, 894-901. 541 Meléndez-Martínez, A. J., Britton, G., Vicario, I. M., Heredia, F. J. (2005). Color 542 and carotenoid profile of Spanish Valencia late ultrafrozen orange juices. 543 Food Research International, 38, 931-936. 544 Meléndez-Martínez, A. J., Britton, G., Vicario, I. M., Heredia, F. J. (2007). 545 Relationship between the color and the chemical structure of carotenoid 546 pigments. Food Chemistry, 101, 1145-1150. 24 547 Meléndez-Martínez, A. J., Vicario, I. M., Heredia, F. J. (2007a). Carotenoids, color 548 and ascorbic acid content of a novel frozen-marketed orange juice. Journal of 549 Agricultural and Food Chemistry, 55(4), 1347-1355. 550 Meléndez-Martínez, A. J., Vicario, I. M., Heredia, F. J. (2007b). Rapid assessment of 551 vitamin A activity through objective color measurements for the quality 552 control of orange juices with diverse carotenoid profiles. Journal of 553 Agricultural and Food Chemistry, 55(8), 2808-2815. 554 Melendez-Martinez, A. J., Britton, G., Vicario, I. M., Heredia, F. J. (2008). The 555 complex carotenoid pattern of orange juices from concentrate. Food 556 Chemistry, 109(3), 546-553. 557 Meléndez-Martínez, A. J., Vicario, I. M., Heredia, F. J. (2009). Effect of ascorbic 558 acid on deterioration of carotenoids and color in ultrafrozen orange juice. 559 Journal of Food Composition and Analysis, 22(4), 295-302. 560 Melendez-Martinez, A.J., Escudero-Gilete, M.L., Vicario-Romero, I.M., Heredia, 561 F.J., 2010. Study of the influence of carotenoid structure and individual 562 carotenoids in the qualitative and quantitative attributes of orange juice 563 colour. Food Research International, 43 ,1289 -1296. 564 Mouly, P. P., Gaydou, E. M., Lapierre, L., Corsetti, J. (1999). Differentiation of 565 several geographical origins in single-strength Valencia orange juices using 566 quantitative comparison of carotenoid profiles. Journal of Agricultural and 567 Food Chemistry, 47, 4038-4045. 1 1.2 0.9 0.6 0.3 0.0 1 2 400 450 500 550 600 650 700 750 3 UFOJ 4 TTOJ1 5 TTOJ2 6 7 Figure 2. Absorption visible spectra of the carotenoid fraction of an ultrafrozen 8 orange juice (UFOJ) and two thermally treated orange juices (TTOJ) in hexane 1 1 2 3 2 4 5 Figure 3. Spectroscopic representation of the Kubelka-Munk parameters of an 6 ultrafrozen orange juice (UFOJ) and two thermally treated orange juices (TTOJ) 1 1 2 3 4 5 2 6 7 Figure 4. Representation of the levels of carotenoids of the samples (UFOJ n= 26, 8 TTOJ n = 44) vs the Kubelka-Munk parameters, linear fits and simple regression 9 coefficients 10