Journal of Food Engineering 318 (2022) 110869 Contents lists available at ScienceDirect Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng Prediction of riboflavin and ascorbic acid concentrations in skimmed heat-treated milk using front-face fluorescence spectroscopy Ulises Alvarado a, b, Anna Zamora a, Oscar Arango c, Jordi Saldo a, Manuel Castillo a, * a Centre D’Innovació, Recerca i Transferència en Tecnologia Dels Aliments (CIRTTA), Department de Ciència Animal I Dels Aliments, Facultat de Veterinària, Universitat Autònoma de Barcelona, 08193, Bellaterra, Spain b Escuela Professional de Ingeniería Agroindustrial, Facultad de Ciencias Agrarias, Universidad Nacional Del Altiplano, Av. Floral 1153, Puno, 21001, Peru c Facultad de Ingeniería Agroindustrial, Universidad de Nariño, Ciudad Universitaria Torobajo, Pasto, Nariño, Colombia A R T I C L E I N F O A B S T R A C T Keywords: The objective of this work was to obtain degradation kinetic models of riboflavin and ascorbic acid, and to Front-face fluorescence spectroscopy evaluate the potential of front-face fluorescence (FFF) spectroscopy to predict their concentration in milk after Milk thermal processing. A factorial design with three temperatures (70, 80 and 90 ◦ C) and nine times (from 0 to 60 Heat treatment min) of heat treatment was used. Degradation kinetics of both vitamins were of first order. Predictive kinetic Riboflavin Ascorbic acid models using the Arrhenius equation had a variability coefficient of 2.26 and 3.64% for riboflavin and ascorbic Prediction acid, respectively. With fluorescent predictors such as tryptophan, Maillard compounds and riboflavin, predic­ tion models with coefficients of variation smaller than 3.57% for riboflavin and ranging from 4.24 to 14.25% for ascorbic acid were obtained. In conclusion, FFF is a viable alternative to predict riboflavin and ascorbic acid content of milk which could allow in-line monitoring and control of thermal damage. 1. Introduction oxidative stress (Rajendran et al., 2014). It is ingested mainly through a diet rich in fresh fruits and vegetables; however, milk is not an important Bovine milk is an excellent source of vitamin B2, which can occur as source of the recommended daily intake (Morrissey and Hill, 2011). free riboflavin or in the form of one of the two coenzymes, flavin Being the most unstable and labile vitamin, ascorbic acid is very easily mononucleotide and flavin adenine dinucleotide (Koop et al., 2014). altered not only by temperature, but also by the action of light, changes Much of riboflavin in milk is present as free riboflavin in up to 1 mg/L in pH, and metal ions (Morrissey and Hill, 2011). Thus, in the food in­ and so milk is one of the main contributors to the dietary intake of dustry, ascorbic acid is considered as an index of nutrient retention and riboflavin (Saedisomeolia and Ashoori, 2018), fulfilling recommended an indicator of thermal damage (Sun et al., 2012). allowance of 1.1 and 1.3 mg for adult females and males, respectively There are several techniques to quantify these two vitamins in milk, (Nohr et al., 2011). This vitamin is characterized by being the most heat being the chromatographic methods the most widely used and in resistant of the water-soluble family, but it is very labile to the action of particular high-performance liquid chromatography (HPLC) technique visible and ultraviolet light (Lešková et al., 2006). Conventional steril­ (Ardö et al., 2011; Romeu-Nadal et al., 2006; Bueno-Solano et al., 2009; ization causes losses of up to 18%, in skimmed milk (Nohr et al., 2011), Poulsen et al., 2015). Most of these methods are accurate, but require whilst in whole milk, the destruction percentage does not exceed 10% sophisticated laboratory implementation, trained personnel, large (Veisseyre, 1988). This difference could probably be due to the fact that sample amount, complex sample preparation, appreciable amounts of riboflavin would be more protected in whole milk than in skimmed milk, reagents and long analysis times (Huang et al., 2021). Currently, the because fat globules diminish heat transfer (Fox et al., 2015). food industry needs to develop fast and inexpensive alternative Like riboflavin, vitamin C or ascorbic acid is an essential vitamin, methods. In this sense, electrochemical sensors are being developed to which fulfils vital processes in the human body. This compound has an quantify vitamins in foods (Huang et al., 2021). However, owing to the antioxidant function that delays or inhibits oxidation, reduces the con­ fact that other alternative methods providing real-time information centration of free radicals and provides protection against diseases could guarantee process control to ensure food safety and security, op­ (arthritis, atherosclerosis, cancer, diabetes, ischemia, etc.) that involves tical methods, such as NIR and fluorescence spectroscopies, are gaining * Corresponding author. E-mail address: manuel.castillo@uab.es (M. Castillo). https://doi.org/10.1016/j.jfoodeng.2021.110869 Received 2 June 2021; Received in revised form 13 October 2021; Accepted 25 October 2021 Available online 5 November 2021 0260-8774/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). U. Alvarado et al. Journal of Food Engineering 318 (2022) 110869 more interest (Shaikh & O’Donnell, 2017; Pu et al., 2020). In particular, This device was equipped with a 15 W xenon pulsed lamp and an front-face fluorescence spectroscopy (FFF), which measures the fluo­ accessory with “front-face” geometry adjusted to an angle of incidence rescence on the surface of the analyzed sample, can be directly applied of 35◦ , which minimizes both the phenomena of specular reflection by to cloudy samples such as milk, showing great application potential the surface of the cuvette and of internal absorption by the sample. (Karoui and Blecker, 2011). So far, several markers of thermal damage, Excitation and emission wavelengths were selected based on the such as furosine, lactulose and retinol have been estimated using FFF fluorophore to be determined, using a 5 nm slit in all cases. Tryptophan detecting fluorophores such as tryptophan, dityrosine, Maillard’s com­ fluorescence spectrum (FTRP) was scanned at 290 nm excitation and the pounds and riboflavin found in milk (Kulmyrzaev and Dufour, 2002; emission range of 300–450 nm. For Maillard compounds (FMC) the Ayala et al., 2017; Liu et al., 2018). excitation wavelength was 330 nm, and between 350 and 500 nm for the As was previously mentioned, both riboflavin and ascorbic acid emission. Dityrosine (FDT) was read at the excitation wavelength of 315 could play a major role in evaluating the thermal damage of milk. nm with an emission range from 350 to 500 nm. Finally, riboflavin was However, to the best of our knowledge, no kinetic studies about ribo­ excited at 267, 370 and 450 nm (FRb267, FRb370 and FRb450 respectively) flavin and ascorbic acid degradation due to heat treatment of milk have with a wavelength emission range of 470–570 nm. been published. However, two studies have shown that it is feasible to The maximum fluorescence intensity value of each fluorophore was quantify both vitamins using FFF. Diez et al. (2008) built mathematical used to construct the mathematical prediction models using fluorescent models that allowed estimating ascorbic acid with a relative error of markers. 12% during heat treatment at seven temperatures (72–115 ◦ C) and six different times (2–9.5 min) in infant milk formula serum-based. 2.3. Chemical quantification of vitamins Recently, Alvarado et al. (2019) also developed and validated a math­ ematical model for the prediction of riboflavin concentration in milk. 2.3.1. Riboflavin The validation of the model yielded a standard error of prediction of Riboflavin concentration was analyzed by the method of Albala-­ 0.13 mg/L and a coefficient of variation of 4.14%. Finally, the riboflavin Hurtado et al. (1997). Thawed milk samples (5 g) were accurately concentration prediction model based on FFF data developed by weighed in a centrifuge tube to which 0.5 g of trichloroacetic acid was Alvarado et al. (2019) was compared with the concentration measured added. The mixture was stirred for 10 s and subsequently centrifuged for using two other HPLC conventional techniques. Results obtained with 15 min at 2,500 rpm (Sigma laboratory centrifuge, 4K-15, SN. 93250, commercial milk showed no significant differences between the vitamin Osterode am Harz, Germany). Supernatants were stored and 0.5 mL of concentration estimated using FFF and HPLC. Thus, both investigations 4% trichloroacetic acid was added to the residual pellets for a second suggest the great potential of FFF as it is a simple and fast method extraction by mixing for 1 min before being centrifuged at 2,500 rpm for compared to conventional measurements. 10 min. The two acid extracts were combined in a 10 mL tube and were Due to the lack of studies of ascorbic acid and riboflavin degradation kept protected from light. Just before injection into HPLC, extracts were kinetics in heated milk and the need to lay the foundations for the filtered through 0.22-μm filters. development of an optical sensor that allows monitoring and evaluating A stock solution of 100 mg L− 1 of riboflavin in the form of flavin thermal damage in milk, the present study proposes to develop kinetic mononucleotide (R9504, Sigma Aldrich, Saint Louis, Missouri) prepared prediction models of the loss of riboflavin and ascorbic acid in milk in 2.4% (v/v) of acetic acid was used to obtain calibration solutions (0.5, during heat treatment and its evaluation with fluorescent markers 1.0, 2.0, 4.0 mg L− 1), diluting the stock solution in 2.4% aqueous acetic applying FFF. acid. All standard solutions were filtered through a 0.22-μm filter just before injection into HPLC. 2. Materials and methods Chromatography measurements were performed in triplicate using a Dionex P680 HPLC UV–Vis detector (Dionex, Germering, Germany). The 2.1. Sample preparation analytical column used was Waters Spherisorb ODS 2 C18, with a par­ ticle filling of 5 μm in diameter, 4.6 × 150 mm (Tecknokroma, Barce­ The study was carried out with standard skimmed milk powder dried lona, Spain). Chromatographic separation was carried out under by atomization, of high functional and microbiological quality (extra- isocratic conditions in a mobile phase of 5 mM octane sulfonic acid, grade, “low heat”), supplied by Chr. Hansen S.L. (Barcelona, Spain). The 0.5% triethylamine, 24% glacial acetic acid and 15% methanol. The milk was reconstituted at 12% (w/w) with distilled water tempered at eluent flow rate was 1.0 mL min− 1 and the column temperature was 40 ◦ C, magnetically stirring for 15 min at 150 rpm (Agimatic-E, JP. 25 ◦ C. Time analysis of each sample was 20 min. Selecta, Barcelona, Spain). It was then allowed to stand for 30 min at room temperature in the dark to facilitate proper hydration of the milk 2.3.2. Ascorbic acid constituents. The same equipment with an analytical column Tracer Extraxil ODS In order to study the kinetic evolution of riboflavin and AA, heat 2 C18; 0.46 mm × 250 nm and particles of diameter 5 × 5 mm I.D. treatments were applied to 5 mL samples of milk, using Pyrex test tubes (Tecknokroma, Barcelona, Spain) was used for the quantification of of 30 mL capacity, tightly sealed and immersed in a thermostatic bath ascorbic acid following the method of Romeu-Nadal et al. (2006) with (OvanTherm model TC00E C, Suministros Grupo Esper SL, Badalona, some modifications. Thawed milk samples (1.50 mL) were centrifuged Spain). The following conditions were evaluated: three levels for tem­ at 12,000 rpm for 15 min (Hettich Universal, Mikro 12–24, Germany). perature (70, 80 and 90 ◦ C) with nine levels for time (0, 5, 10, 15, 20, 25, 0.4 mL of the supernatants were carefully extracted, mixed with 1.00 mL 30, 45 and 60 min). At the different treatment time intervals, the sam­ of 0.55% (w/v) metaphosphoric acid solution and manually stirred for ples were removed from the bath and placed in cold water until reaching 30 s. Supernatants obtained by a second centrifugation at 12,000 rpm room temperature to measure fluorescence, while for the kinetic study of for 10 min were filtered through a 0.22-μm filter just before injection the vitamins they were frozen at − 30 ◦ C until their subsequent chemical into HPLC. analysis. Chromatographic separation was performed in triplicate by isocratic elution using a mobile phase of Milli-Q water with acetic acid (0.1% v/v) 2.2. Fluorescence determination and methanol in a ratio of 95:5 (v/v). The eluent flow rate was 0.7 mL min− 1, the column temperature was 25 ◦ C and the analysis time of each Spectroscopic fluorescence measurements were performed in tripli­ sample was 20 min. Ascorbic acid was identified by comparing the cates in Suprasil quartz cuvettes (Agilent Technologies, Madrid, Spain) retention time of the sample peak at 254 nm. Quantification was carried with a spectrofluorometer model Cary Eclipse (Agilent Technologies). out using a calibration curve of ascorbic acid standard solutions (1, 5, 2 U. Alvarado et al. Journal of Food Engineering 318 (2022) 110869 10, 20 and 30 ppm) prepared by dissolving ascorbic acid (Sigma- utilized to obtain the best one, two- and three-parameter models for Aldrich, Spain) in 0.55% (w/v) metaphosphoric acid dissolved with predicting riboflavin concentration with FTrp, FDt, FMC, FRb267, FRb370 and Milli-Q water. FRb450. The procedure to obtain the prediction models of the two vita­ mins was correlated with the maximum fluorescence intensities ob­ tained from the emission spectra for each fluorescent marker. 2.4. Kinetic data modeling 3. Results and discussion Kinetic equations of zero, first and second order were used to fit the data of riboflavin and ascorbic acid degradation during milk heat 3.1. Effect of heat treatment on riboflavin and ascorbic acid degradation treatment and obtain corresponding kinetic data through regression analysis using MS Excel software 2013 (Microsoft, USA). Determination Fig. 1a shows the effects of heat treatment on riboflavin content of coefficients (R2) were used to select the order of the kinetic model that reconstituted skimmed milk. An initial concentration of 1.72 mg L− 1 was better explains the degradation kinetics, according to the following found in non-heat-treated milk, equal to that found by Poulsen et al. equations: (2015). Variations of riboflavin in fresh cow milk are mainly due to Zero order as breed and/or biotypes, diet, and climate effects (Fox et al., 2015). At the [C] = [C]0 + kT t highest temperature evaluated (90 ◦ C), riboflavin loss was 2.99 and 22.55% after 5 min and 60 min, respectively, showing that this vitamin First order as is not very sensitive to heat. Meha (1994) found decreases of 2 and 4% of ln[C] = ln[C]0 + kT t riboflavin during heat treatments for 30 min at 80 and 90 ◦ C, respec­ tively, in whole milk; values that are lower than the results of the present Second order as study. This difference could be due to the fact that riboflavin shows 1 / [C] = 1/[C]0 + kT t, higher degradation in high-fat milk than in skim milk (Fox et al., 2015). Saad (1980), cited by Meha (1994), found 12% loss of riboflavin as a where, [C]0 and [C] = initial and final concentration of vitamin result of low pasteurization (63 ◦ C, 30 min). Such loss was only observed respectively (expressed in mg⋅L− 1); kT = reaction rate as a function of at 90 ◦ C for 30 min in the present work. Asadullah et al. (2010) found a temperature (expressed in different units depending on the reaction decrease of 16.5 and 27% in riboflavin concentration of whole cow milk order); t = time (expressed in min). after heating at 100 ◦ C for 10 and 15 min, respectively. The effect of temperature on the reaction rate was modeled using the In many cases, ascorbic acid is considered as an indicator molecule in Arrhenius equation (Van Boekel, 2008). Furthermore, the following food processing because it is highly reactive and very sensitive to the thermal parameters described by Ordóñez et al. (2013) were physicochemical conditions in the environment (Al Fata et al., 2017). determined: Therefore, although milk is not the main source of this vitamin, it could Decimal reduction time (DT) as be used as a potential marker of thermal damage to milk during pro­ cessing. Fig. 1b shows the effects of time and temperature of heat Ln(10) DT = treatment on the concentration of ascorbic acid in reconstituted skim­ kT med milk. The amount found in non-heat-treated milk was 19.50 mg with kT being the reaction rate. L− 1, which is in the range of concentrations reported by other authors Half-life time (t0.5) as (Esteire et al., 2014; Fox et al., 2015; Nalame et al., 2009). Its concen­ tration depends on the biological conditions (breed, food, etc.) and, t0.5 = Ln(0.5) . possibly, the season of the year, being apparently higher in winter kT (Tamime, 2009). The loss of ascorbic acid significantly increased (p < Thermal coefficient (Q10) as 0.05) with increasing heat treatment temperature and time. As mentioned, ascorbic acid is quite thermolabile and is quickly [ ]T 10− T k2 2 1 destroyed when the milk is subjected to heat treatment. The results Q10 = k1 obtained in the present study confirm its thermolability. Meha (1994) observed degradation values of 26 and 36% for treatments of 30 min at with k1 and k2 corresponding to the rate constants at T1 = 70 ◦ C and T2 80 and 90 ◦ C, respectively. Damjanovic and Birlouez-Aragon (2011) = 90 ◦ C, respectively. found losses of up to ~23% due to treatments at 80 ◦ C for 9.5 min in Thermal resistance constant or z value infant food, being therefore slightly greater than our results. In other investigations, the reduction was even greater. For example, Bendicho 10Ln(10) z= . et al. (2002) found losses of up to ~50% after 30 min treatment at 63 ◦ C LnQ10 in skim milk, while Haddad and Loewenstein (1983) found ~21% loss Arrhenius law after treating milk at 80 ◦ C for 16 s. It has been shown that temperature − Ea/ and matrix are important factors that affect the stability of ascorbic acid k T = A0 e RT (Stešková et al., 2006). where A0 is the constant or pre-exponential factor, Ea the activation 3.1.1. Kinetic models for riboflavin degradation energy of the process (expressed in J⋅mol− 1), R is the general gas con­ Different kinetic equations for the degradation of riboflavin in milk stant (8.3141 J mol− 1K− 1) and, finally, T is the absolute temperature during heat treatment were obtained (Table 1). The first-order kinetic is expressed in K (◦ C + 273). the model that had the best fit in all treatments, showing in general the highest R2. It was not possible to find information about riboflavin 2.5. Prediction models using fluorescent compounds thermal degradation in milk, but for other products, disappearance of the riboflavin responds also to a first-order kinetics, like in soy milk Data were processed and analyzed using “Statistical Analysis Sys­ (Kwok et al., 1998), whole green soybeans (Nisha et al., 2005a) or tem” (SAS, version 9.2, 2009, SAS Institute Inc, Cary, NC, USA). Pear­ spinach (Nisha et al., 2005b). son’s correlation coefficients (r) were determined by the correlation Table 2 shows the estimated kinetic parameters assuming a first- procedure (CORR). The maximum R2 procedure (REG, MAXR) was order for degradation of riboflavin in milk. As expected, the values of 3 U. Alvarado et al. Journal of Food Engineering 318 (2022) 110869 Fig. 1. Thermal degradation of a) riboflavin and b) ascorbic acid in skimmed milk (first order kinetics). Table 1 Comparison of different kinetic equations of riboflavin and ascorbic acid degradation during heat treatment in skim milk. Equations 70 ◦ C 80 ◦ C 90 ◦ C kT units 2 2 R k70 R k80 R2 k90 Zero-order Rb 0.969 − 0.0015 0.971 − 0.0042 0.973 − 0.0062 mg⋅L− 1⋅min− 1 AA 0.989 − 0.187 0.992 − 0.187 0.964 − 0.194 First-order Rb 0.965 − 0.0009 0.983 − 0.0027 0.983 − 0.0041 min− 1 AA 0.990 − 0.011 0.986 − 0.014 0.980 − 0.016 Second-order Rb 0.965 0.0005 0.959 0.0017 0.956 0.0027 mg− 1⋅L⋅min− 1 AA 0.962 0.001 0.951 0.001 0.947 0.001 N = 27; number of replicates, Nr = 3; Rb, riboflavin; AA, ascorbic acid; kT, reaction rate; t, time (min); R2, coefficient of determination. Table 2 First order kinetic parameters of riboflavin and ascorbic acid degradation in milk. Vitamin T ◦C kT (min− 1) DT (min) t0.5 (min) Q10 z (◦ C) Ea (kJ⋅mol− 1) Rb 2.13 ± 0.33 30.41 ± 4.08 75.84 ± 10.59 70 − 0.0009 ± 0.00007 2560.79 ± 203.2 770.87 ± 61.2 80 − 0.0026 ± 0.00023 864.84 ± 85.2 260.34 ± 25.6 90 − 0.0039 ± 0.00062 563.11 ± 89.3 169.51 ± 26.9 AA 1.28 ± 0.004 96.96 ± 7.96 19.07 ± 2,85 70 − 0.0110 ± 0.0002 221.78 ± 8.64 63.08 ± 1.10 80 − 0.0141 ± 0.0002 160.59 ± 3.28 49.11 ± 0.76 90 − 0.0159 ± 0.0010 137.93 ± 6.41 43.66 ± 2,90 N = 27; number of replicates, Nr = 3; Rb, riboflavin; AA, ascorbic acid; kT, kinetic constant; DT, decimal reduction time; t0.5, half-life; Q10, thermal coefficient; Ea, activation energy; z, z-value. the reaction constant (kT) increased with temperature. Fig. 2a shows the 20–45 ◦ C up to 12 months. relation between Ln kT and the inverse of absolute temperature. In green Few authors have approached the kinetic study of the degradation of soybeans, decimal reduction time (DT) and half-life (t0.5) of riboflavin at riboflavin in milk during heat treatment. However, a considerable 90 ◦ C were 3268 min and 122 min, respectively (Nisha et al., 2005a). variation in the kinetic parameters of this compound has been docu­ The z value found in the present study was slightly lower than that found mented in other foods. For example, Ea was reported to be 21.7 kJ mol− 1 for thermal degradation of riboflavin in soy milk (36 ◦ C) (Kwok et al., in spinach (Nisha et al., 2005b), 29.8 kJ mol− 1 in green soybean (Nisha 1998). Using the Arrhenius equation based on a first-order kinetic model et al., 2005a) and 37.2 kJ mol− 1 in rosehip nectar (Kadakal et al., 2018). with R2 ~0.95, the obtained activation energy (Ea) was 75.8 kJ mol− 1. Thermal degradation of riboflavin at 50–70 ◦ C produced β-keto acid and This value was higher than that found by Singh et al. (1975) for a dioxo compound as the isoalloxazine ring cleavage products at pH light-induced riboflavin loss in whole milk (~27.5 kJ mol− 1), and close 9–13 (Sheraz et al., 2014). Little information is available regarding to 83.3 kJ mol− 1 reported for thermal degradation of riboflavin in soy riboflavin thermal degradation in aqueous solution. Riboflavin milk (Kwok et al., 1998). Likewise, Galdi et al. (1989) found a wide destruction rate increased when oxygen was present during storage and range of Ea values from 34 to 44 kJ mol− 1 in infant milk, stored at with increasing water activity. Other components in food may also affect 4 U. Alvarado et al. Journal of Food Engineering 318 (2022) 110869 Fig. 2. Effect of temperature (70, 80 and 90 ◦ C) on the degradation rate of a) riboflavin and b) ascorbic acid in skim milk based on Arrhenius model (k values used corresponded to first order kinetic models). the riboflavin degradation rates (Choe et al., 2005). The differences − 9121 found in the results with respect to those of other studies may be a [Rb]t = [Rb]0 × e(3.47×10 ) 8 ×t×e T consequence of temperature ranges evaluated, the chemical composi­ tion of food, pH and the presence of light during the production process where [Rb]t is the concentration of riboflavin at time t (mg⋅L− 1); [Rb]0, (Saidi and Warthesen, 1995). the initial concentration (mg⋅L− 1); t, time (min); T, temperature (K); the The kinetic equation to predict the concentration of riboflavin as a value 3.47 × 108 min− 1, is the pre-exponential factor (A0), determined function of time and temperature is given by the following mathematical using Arrhenius equation. With this equation it was possible to predict model: riboflavin thermal degradation with a standard error of prediction (SEP) and a variation coefficient (CV) of 0.03 mg L− 1 and 2.26%, respectively Fig. 3. Concentration of experimental values with respect to those predicted for a) riboflavin and b) ascorbic acid are estimated according to the kinetic model of order one in skim milk. 5 U. Alvarado et al. Journal of Food Engineering 318 (2022) 110869 (Fig. 3a). 3.2. Prediction of riboflavin and ascorbic acid using fluorescent The results confirm that riboflavin is quite thermostable since it only compounds degraded in small percentages (e.g., ~22% at 90 ◦ C for 60 min). In the dairy industry, riboflavin has been studied normally during products 3.2.1. Riboflavin prediction storage to evaluate its photodegradation. However, results indicate that Riboflavin is a fluorescent compound with 3 excitation maxima, it can also be a potential marker to monitor thermal damage during high reason why it might be complicate to use its fluorescence to directly heat treatments. assess its concentration due to the interference of other self-fluorescent compounds naturally present in milk. Among those, we can highlight 3.1.2. Kinetic models for ascorbic acid degradation some of the most relevant aromatic amino acids as tryptophan, tyrosine Comparison of different kinetic equations for ascorbic acid degra­ and phenylalanine, having this last one a low quantum yield and dation at different temperatures, according to zero, one and two-order contributing little to protein fluorescence. Di-tyrosine is a product of models are shown in Table 1. The data confirm that ascorbic acid proteins oxidative stress, which is also fluorescent and can provide in­ disappearance follows first-order kinetics (Fig. 2b), in accordance with formation on changes undergone by proteins. Some other self- that obtained, in the same biological matrix, by Bendicho et al. (2002) fluorescent compounds, as NADH or even chlorophylls, are present in and in other foods (Karhan et al., 2004; Stešková et al., 2006). Table 2 low amounts and have a limited stability. shows the first-order kinetic parameters of ascorbic acid degradation in One-, two- and three-parameter models were obtained from fluo­ milk. The obtained k90 ◦ C was 0.0159 min− 1, much lower than 0.08 rescent markers for prediction of riboflavin concentration at 70, 80 and min− 1 reported by Bendicho et al. (2002) but similar to 0.014 min− 1 90 ◦ C using the maximum R2 procedure of SAS (Table 3). At 70 ◦ C, FMC observed by (Kadakal et al., 2018) in rosehip nectar. The obtained t0.5 at and FDt were the main predictors, both being inversely proportional to 70 ◦ C was higher than the 49 min found by Ordóñez-Santos and the concentration of riboflavin and with Pearson’s correlation co­ Yoshioka-Tamayo (2012) in mango pulp treated at the same tempera­ efficients of − 0.89 (p < 0.0001). However, at 80 ◦ C, the main predictors ture. For DT, the obtained values were lower than those found by were FTrp and λFTrp with r values of 0.73 and − 0.79 (p < 0.0001), Dhuique-Mayer et al. (2007) at the same temperatures in citrus juice; respectively. Similarly, at 90 ◦ C, both FTrp and λFTrp showed the highest r but, higher than those found in guava fruits, at 75 and 95 ◦ C (2.88 and values with respect to riboflavin concentration (0.91 and − 0.81 with p 1.5 min) (Ordóñez et al., 2013). The z value was higher than those < 0.0001, respectively). Except in model IV, R2 values for all prediction observed in guava (~77 ◦ C) and in mango pulps (66 ◦ C) models, even for those with only one-parameter, were between 0.80 and (Ordóñez-Santos and Yoshioka-Tamayo, 2012; Ordóñez et al., 2013). 0.86 with SEP below 0.06 mg L− 1. These results show that some FFF Furthermore, the Q10 value found was similar to that of some fruits e.g. parameters are good predictors of riboflavin concentration in milk heat- 1.35 in guava (Ordóñez et al., 2013), 1.4 in mango pulp (Ordóñez-Santos treated at different temperatures. and Yoshioka-Tamayo, 2012) and 1.2 in rosehip pulp (Karhan et al., Table 3 also shows the general mathematical models for the esti­ 2004). The obtained Ea was lower than that found by Bendicho et al. mation of riboflavin concentration obtained with data from the three (2002) and Damjanovic and Birlouez-Aragon (2011), which reported temperatures studied. The three-parameter model (XII) included data at values of 63.66 kJ mol− 1 for skim milk and 67.94 kJ mol− 1 for infant all temperatures and is the one with a better fit. This model explained formula respectively. A wide range of Ea values (7.54–125.6 kJ mol− 1) 81% of the variability of riboflavin concentration with a SEP of 0.043 were found in liquid foods (Lee and Labuza, 1975). Ascorbic acid can be mg L− 1 (Fig. 4a). Both FTrp and λFTrp had a high relationship with ribo­ reversibly oxidized to dehydroascorbic acid (DHAA), without losing flavin concentration, with r ≥ 0.80, while the relationship with the third vitamin C functionality. Further degradation of DHAA (e.g. to 2,3-dike­ parameter, FMC, was low (r = 0.26). Predictor λFTrp is attributed to the togulonic acid) results in the loss of biological activity. These conversion redshift of FTrp due to the denaturation of whey protein as a consequence reactions are influenced by pH, moisture content, temperature, presence of exposure of tryptophan residues from a non-polar environment to the of oxygen and metal ions (Belitz et al., 2004; Mendoza-Corvis et al., aqueous phase of milk (Alvarado, 2017; Ayala et al., 2020). At first, 2015). Oxygen is probably the most determining factor in vitamin C Becker et al. (2003) studied the potential use of FFF with the statistical degradation, temperature is frequently the accelerating factor, whereas method partial least squares (PLS) for the estimation of riboflavin in anaerobic degradation occurs mainly when oxygen is depleted (Oey yogurt during storage (5 weeks at 4 ◦ C), constructed a two-component et al., 2008; Hiatt et al., 2010). In thermal experiments, each of the two multiple regression model that was attributed to FRb (450 nm excita­ degradation reactions was found to follow first-order kinetics (Lima tion and 530 nm emission) and FTrp (290 nm excitation and 350 nm et al., 2010; Verbeyst et al., 2013). emission). Multivariate regression showed a correlation of 0.99 and a Substituting the Arrhenius equation into the first-order kinetic prediction error of 0.092 μg of riboflavin per g of yoghurt. Christensen equation, we obtained the following mathematical model to predict the et al. (2005) also took advantage of the FFF with the application of concentration of ascorbic acid at a given time and at a given heat parallel factor analysis (PARAFAC) to predict the same vitamin with the treatment temperature. same biological matrix and under the same environmental conditions as − 2300.7 the previous study. The PARAFAC analysis with the multiple linear [AA]t = [AA]0 × e(9.15×t×e ) regression method showed an R2 of 0.97, and a prediction error of 0.094 T ppm, with FRb and lumichrom (photochemical degradation of riboflavin) Where [AA]t, final concentration of ascorbic acid (mg⋅L− 1); [AA]0, its being the main predictors. Recently, Alvarado et al. (2019) established a initial concentration (mg⋅L− 1); t, time (min); T, temperature (K); the prediction model based on FRb with point mediated at 370 nm of exci­ value 9.15, is the pre-exponential factor (A0), which is given from the tation and 530 nm of emission to estimate the concentration of ribo­ intercept of linear regression. Prediction of ascorbic acid concentration flavin in commercial milk which was compared with HPLC methods. The using this equation had SEP = 0.03 mg L− 1 and CV = 2.26% (Fig. 3b). validation of the model was R2 = 0.99 and SEP = 0.13 mg L− 1. It should From the results obtained, it is confirmed that ascorbic acid is very be noted that in these investigations there was no prior manipulation of sensitive to processing by heat treatment. Therefore, it can be used as an the sample for analysis. indicator to assess thermal nutritional damage in the dairy industry (Sun Therefore, all these results suggest the convenience of using FFF et al., 2012). spectroscopy as a simple and non-destructive method for online and offline determinations of riboflavin in dairy products. 3.2.2. Ascorbic acid prediction Using maximum R2 procedure of SAS, highly significant models with 6 U. Alvarado et al. Journal of Food Engineering 318 (2022) 110869 Table 3 Prediction models of riboflavin concentration in heated milk by means of front-face fluorescence. Mathematical models β0 β1 β2 β3 R2 SEP CV T I [Rb]*** = β0 + β1FMC 2.40*** − 0.001*** – – 0.80 0.014 0.84 4 II [Rb]*** = β0 + β1FMC + β2FDt 2.55*** − 5.5 × 10− − 8.4 × 10− 4 – 0.83 0.013 0.80 70a 4 III [Rb]*** = β0 + β1FMC + β2FDt + β3FRb450 2.43*** − 5.8 × 10− − 9.8 × 10− 4* − 8.4 × 10− 4 0.84 0.013 0.78 IV [Rb]*** = β0 + β1λFTrp 11.1*** − 0.027*** – – 0.62 0.051 3.14 V [Rb]*** = β0 + β1FTrp + β2FMC 1.39*** 0.013*** − 0.002*** – 0.83 0.054 2.81 80a VI [Rb]*** = β0 + β1FTrp + β2FMC + β3FRb267 1.31*** 0.015*** − 0.002*** 2.1 × 10− 4* 0.86 0.032 1.99 VII [Rb]*** = β0 + β1FTrp 0.98* 0.006* – – 0.83 0.049 3.15 VIII [Rb]*** = β0 + β1FTrp + β2FMC 1.51* 0.007* − 0.001 – 0.85 0.047 3.00 90a IX [Rb]*** = β0 + β1FTrp + β2FMC + β3FRb450 1.15* 0.007* − 0.001* 0.002 0.86 0.046 2.96 X [Rb]*** = β0 + β1λFTrp 6.81*** − 0.015*** – – 0.64 0.06 3.57 XI [Rb]*** = β0 + β1FTrp + β2FMC 1.87*** 0.008*** − 0.001*** – 0.72 0.05 3.06 Allb XII [Rb]*** = β0 + β1FTrp + β2FMC + β3λFTrp 4.07*** 0.006*** − 0.001*** − 0.006* 0.81 0.04 2.65 *p < 0.05, **p < 0.001, ***p < 0.0001. a N = 27. b N = 81; R, determination coefficient; SEP (mg⋅L− 1), standard error of prediction; CV (%), coefficient of variation; T (◦ C), temperature; β0, β1, β2, β3, regression coefficients; [Rb], riboflavin concentration; λFTrp, maximum intensity wavelength of tryptophan fluorescence; FTrp, FMC, FDt, FRb267, FRb450, maximum fluorescence intensity of tryptophan, Maillard compounds, dityrosine and riboflavin (excited at 267 and 450 nm). Fig. 4. Relationship between the experimental values of a) riboflavin and b) ascorbic acid versus the predicted values of model XII from Tables 3 and 4, respectively. high R2 and small SEP values were obtained for prediction of ascorbic However, at 90 ◦ C the best model was VIII, which included two pre­ acid concentration at 70 and 90 ◦ C but using different fluorescent dictors to estimate the concentration of ascorbic acid; FTrp and FMC had a markers for each temperature and having as main predictors FMC and positive correlation with ascorbic acid (r = 0.95, p > 0.0001 and 0.59, p FTrp, respectively (Table 4). At 80 ◦ C, the fit of the models was lower > 0.001, respectively). Although it is true that at high temperatures and because the main predictor (λFTrp) had a moderate negative correlation prolonged times Maillard reactions take place, causing changes in the coefficient (r = − 0.78, p > 0.0001). This predictor is related to the fact color, taste and odor of milk, in the present study both at 80 and 90 ◦ C, that there is a shift in the maximum wavelength of fluorescence emission the FMC shows an increase up to the minute 30 and 5, respectively, and throughout the heat treatment, both at 80 and 90 ◦ C, while at 70 ◦ C no subsequently a decrease which could be related to the disappearance of change in the wavelength of the maximum fluorescence emission peak is intermediate compounds and production of advanced stages of the observed. Similar results were found by Ayala et al. (2020) in the same Maillard reaction or, in a lesser extent, to the internal filter effect caused biological matrix. Model I was the best one-parameter model found to when the fluorophores are in very high concentrations (Geddes and predict the concentration of ascorbic acid at 70 ◦ C. It included an Lakowicz, 2006). intercept (β0) and the predictor FMC, which had an inverse correlation The general models to estimate the concentration of ascorbic acid for with ascorbic acid (r = − 0.97, p > 0.0001). The intersection and the the three temperatures studied are shown, also in Table 4. All the predictor were significant, and this simpler model explained the vari­ mathematical models were highly significant (p < 0.0001). It was found ability of ascorbic acid concentration in 95%, with a SEP of 0.69 mg L− 1. that FRb450 and FRb370 were inversely proportional to ascorbic acid Little improvement was achieved by the two or three-parameter models. concentration with r values of − 0.13 and − 0.72, respectively. As with 7 U. Alvarado et al. Journal of Food Engineering 318 (2022) 110869 Table 4 Prediction models of ascorbic acid degradation in heated milk by means of front-face fluorescence. Mathematical models β0 β1 β2 β3 R2 SEP CV T I [AA]*** = β0 + β1FMC 91.22*** − 0.11*** – – 0.95 0.69 4.53 II [AA]*** = β0 + β1FMC + β2FRb450 82.92*** − 0.12*** 0.05 – 0.95 0.65 4.29 70a III [AA]*** = β0 + β1FMC + β2FRb450 + β3FTrp 76.97*** − 0.12*** 0.05 0.06 0.95 0.65 4.24 IV [AA]*** = β0 + β1λFTrp 423.3*** − 1.19*** – – 0.62 2.12 14.25 V [AA]*** = β0 + β1FTrp + β2FMC 4.56 0.57*** − 0.07*** – 0.70 1.39 9.33 80a VI [AA]*** = β0 + β1FTrp + β2FMC + β3FRb370 19.96 0.53*** − 0.05* − 0.08 0.85 1.33 8,97 VII [AA]*** = β0 + β1FTrp − 5.315** 0.22*** – – 0.91 1.07 7.78 VIII [AA]*** = β0 + β1FTrp + β2FDt 16.75* 0.28*** − 0.05* – 0.94 0.84 6.08 90a IX [AA]*** = β0 + β1FTrp + β2FDt + β3λTrp 36.51 0.27*** − 0.05* − 0.05 0.95 0.84 6.14 X [AA]*** = β0 + β1FRb370 105.7*** − 0.30*** – – 0.51 1.39 10.13 XI [AA]*** = β0 + β1FRb370 + β2FRb450 76.28*** − 0.39*** 0.21*** – 0.64 1.06 7.74 Allb XII [AA]*** = β0 + β1FTrp + β2FRb450 + β3FMC 38.84*** − 0.29*** 0.04 − 0.09*** 0.75 0.97 7.05 *p < 0.05, **p < 0.001, ***p < 0.0001. a N = 27. b N = 81; R, determination coefficient; SEP (mg L− 1), standard error of prediction; CV (%), coefficient of variation; T, temperature (◦ C); β0, β1, β2, β3, regression coefficients; [AA], ascorbic acid concentration; λFTrp, maximum intensity wavelength of tryptophan fluorescence; FTrp, FMC, FDt, FRb370, FRb450, maximum fluorescence intensity of tryptophan, Maillard compounds, dityrosine and riboflavin (excited at 370 and 450 nm), respectively. riboflavin, it was necessary to include three parameters in model XII, Belitz, H.-D., Grosch, W., Schieberle, P., 2004. Food Chemistry. Springer-Verlag, Berlin. Becker, E., Christensen, J., Frederiksen, C.S., Haugaard, V.K., 2003. Front-face which contained data at the three temperatures, to obtain an acceptable fluorescence spectroscopy and chemometrics in analysis of yogurt: rapid analysis of goodness of fit (R2 = 0.75, SEP = 0.97 mg L− 1) as shown in Fig. 4b. Diez riboflavin. J. Dairy Sci. 86 (8), 2508–2515. et al. (2008) demonstrated that FFF spectroscopy associated with PLS Bendicho, S., Espachs, A., Arántegui, J., Martín, O., 2002. Effect of high intensity pulsed calibration is a quick and simple method to evaluate the nutritional electric fields and heat treatments on vitamins of milk. J. Dairy Res. 69 (1), 113–123. Bueno-Solano, C., Campas-baypoli, O.N., Díaz-garcía, A.S., Izaguirre-flores, E.I., impact of heat treatment at seven temperatures (72, 80, 87, 95, 100, 110 Verdugo-zamorano, W., Estrada-alvarado, M.I., 2009. Quantification of riboflavin and 115 ◦ C) for six different times (from 2 min to 9.5 min) in (vitamin B2) in dairy products by HPLC. Rev. Chil. Nutr. 36 (2), 136–142. serum-based infant food samples. The model allowed to obtain an esti­ Choe, E., Huang, R., Min, D., 2005. Chemical reactions and stability of riboflavin in foods. J. Food Sci. 70 (1), 28–36. mate of ascorbic acid concentration with a relative error of 12% using Christensen, J., Becker, E., Frederiksen, C.S., 2005. Fluorescence spectroscopy and FTrp and FMC. PARAFAC in the analysis of yogurt. Chemometr. Intell. Lab. Syst. 75, 201–208. Damjanovic, S., Birlouez-Aragon, I., 2011. The FAST index – a highly sensitive indicator of the heat impact on infant formula model. Food Chem. 124 (3), 1043–1049. 4. Conclusions Diez, R., Ortiz, M.C., Sarabia, L., Birlouez-Aragon, I., 2008. Potential of front face fluorescence associated to PLS regression to predict nutritional parameters in heat Using FFF spectroscopy markers, it was possible to obtain prediction treated infant formula models. Anal. Chim. Acta. 606, 151–158. Dhuique-Mayer, C., Tbatou, M., Carail, M., Caris-Veyrat, C., Dornier, M., Amiot, M.J., models for predicting of riboflavin and ascorbic acid degradation during 2007. Thermal degradation of antioxidant micronutrients in Citrus juice: kinetics milk heat treatment, being tryptophan fluorescence the most important and newly formed compounds. J. Agric. Food Chem. 55 (10), 4209–4216. predictor. Esteire, L., Cenzano, E., Madrid, A., 2014. In: Madrid Vicente, A. (Ed.), Queserías. Nuevo Manual Técnico, 1o Edición. The proposed technique to predict and/or quantify riboflavin and Fox, P.F., Uniacke-Lowe, T., McSweeney, P.L.H., O’Mahony, J.A., 2015. Dairy chemistry ascorbic acid is a viable alternative to control thermal damage but it and biochemistry. In: Dairy Chemistry and Biochemistry, second ed. Springer. could also be used to evaluate milk powder. FFF spectroscopy has the Galdi, M., Carbone, N., Valencia, M.E., 1989. Comparison of ferric glycinate to ferrous sulfate in model infant formulas - kinetics of vitamin losses. J. Food Sci. 54 (6), advantage of revealing information quickly, directly and non- 1530–1533. destructively. In addition, it is friendly to the environment as does not Geddes, C.D., Lakowicz, J.R., 2006. Reviews in fluorescence 2006. In: Geddes, C.D., require reagents. Therefore, the implementation is encouraging enough Lakowicz, J.R. (Eds.), Reviews in Fluorescence, vol. 2006. Haddad, G.S., Loewenstein, M., 1983. Effect of several heat treatments and frozen to promote future applications for in-line process monitoring and storage on thiamine, riboflavin, and ascorbic acid content of milk. J. Dairy Sci. 66 control. (8), 1601–1606. Hiatt, A., Tylor, L., Mauer, L., 2010. Influence of simultaneous variations in temperature and relative humidity on chemical stability of two vitamin C forms and implications References for shelf life models. J. Agric. Food Chem. 58, 3532–3540. Huang, L., Tian, S., Zhao, W., Liu, K., Guo, J., 2021. Electrochemical vitamin sensors: a Al Fata, N., Georgé, S., André, S., Renard, C.M.G.C., 2017. Determination of reaction critical review. Talanta 222 (1–15), 121645. orders for ascorbic acid degradation during sterilization using a new experimental Karhan, M., Aksu, M., Tetik, N., Turhan, I., 2004. Kinetic modeling of anaerobic thermal device: the thermoresistometer Mastia. LWT - Food Sci. Technol. (Lebensmittel- degradation of ascorbic acid in rose hip (Rosa canina L) pulp. J. Food Qual. 27 (5), Wissenschaft -Technol.) 85, 487–492. 311–319. Albala-Hurtado, S., Veciana-Nogues, M.T., Izquierdo-Pulido, M., Marine-Font, A., 1997. Kadakal, C., Duman, T., Ekinci, R., 2018. Thermal degradation kinetics of ascorbic acid, Determination of water-soluble vitamins in infant milk by high- performance liquid thiamine and riboflavin in rosehip (Rosa canina L) nectar. Food Sci. Technol. 38 (4), chromatography. J. Chromatogr. A 778 (1–2), 247–253. 667–673. Alvarado, U., 2017. Aplicación de indicadores nativos de fluorescencia para la Karoui, R., Blecker, C., 2011. Fluorescence spectroscopy measurement for quality evaluación rápida de daño térmico en el procesado de leche. Universitat Aùtonoma assessment of food systems-a review. Food Bioprocess Technol. 4 (3), 364–386. de Barcelona. Koop, J., Monschein, S., Macheroux, E.P., Knaus, T., Macheroux, P., 2014. Determination Alvarado, U., Zamora, A., Liu, J., Saldo, J., Castillo, M., 2019. Rapid quantification of of free and bound riboflavin in cow’s milk using a novel flavin-binding protein. Food riboflavin in milk by front-face fluorescence spectroscopy: a preliminary study. Chem. 146, 94–97. Foods 9 (1), 1–6. Kulmyrzaev, A., Dufour, É., 2002. Determination of lactulose and furosine in milk using Ardö, Y., Chatterton, D.E.W., Varming, C., 2011. Vitamin C. In: Fuquay, J.W. (Ed.), front-face fluorescence spectroscopy. Lait 82 (6), 725–735. Encyclopedia of Dairy Sciences, second ed. Academic Press, London, pp. 55–62. Kwok, K.C., Shiu, Y.W., Yeung, C.H., Niranjan, K., 1998. Effect of thermal processing on Asadullah, Khair-un-nisa, Tarar, O.M., Ali, S.A., Jamil, K., Begum, A., 2010. Study to available lysine, thiamine and riboflavin content in soymilk. J. Sci. Food Agric. 77 evaluate the impact of heat treatment on water soluble vitamins in milk. Jpma. J Pak (4), 473–478. Med Assoc 60 (11), 909–912. Lee, S.H., Labuza, T.P., 1975. Destruction of acorbic acid as a function of water activity. Ayala, N., Zamora, A., González, C., Saldo, J., Castillo, M., 2017. Predicting lactulose J. Food Sci. 40 (2), 370–373. concentration in heat-treated reconstituted skim milk powder using front-face Lešková, E., Kubíková, J., Kováčiková, E., Košická, M., Porubská, J., Holčíková, K., 2006. fluorescence. Food Control 73, 110–116. Vitamin losses: retention during heat treatment and continual changes expressed by Ayala, N., Zamora, A., Rinnan, Å., Saldo, J., Castillo, M., 2020. The effect of heat mathematical models. J. Food Compos. Anal. 19 (4), 252–276. treatment on the front-face fluorescence spectrum of tryptophan in skim milk. J. Food Compos. Anal. 92 (1–8), 103569. 8 U. Alvarado et al. Journal of Food Engineering 318 (2022) 110869 Lima, J.R., Elizondo, N.J., Bohuon, P., 2010. Kinetics of ascorbic acid degradation and Pu, Y.-Y., O’Donnell, C., Tobin, J.T., O’Shea, N., 2020. Review of near-infrared colour change in ground cashew apples treated at high temperatures (100-180 ◦ C). spectroscopy as a process analytical technology for real-time product monitoring in Int. J. Food Sci. Technol. 45 (8), 1724–1731. dairy processing. Int. Dairy J. 103 (1–11), 104623. Liu, J., Zamora, A., Castillo, M., Saldo, J., 2018. Using front-face fl uorescence Rajendran, P., Nandakumar, N., Rengarajan, T., Palaniswami, R., Gnanadhas, E.N., spectroscopy for prediction of retinol loss in milk during thermal processing. LWT - Lakshminarasaiah, U., 2014. Antioxidants and human diseases. Clin. Chim. Acta Food Sci. Technol. (Lebensmittel-Wissenschaft -Technol.) 87, 151–157. 436, 332–347. Meha, M.A., 1994. Vitamin C and riboflavin contento in Camels milk: effects of heat Romeu-Nadal, M., Morera-Pons, S., Castellote, A.I., López-Sabater, M.C., 2006. Rapid treatments. Food Chem. 50, 153–155. high-performance liquid chromatographic method for Vitamin C determination in Mendoza-Corvis, F.A., Hernández, E.J., Ruiz, L.E., 2015. Efecto del Escaldado sobre el human milk versus an enzymatic method. J. Chromatogr. B, Analyt. Technol. Color y Cinética de Degradación Térmica de la Vitamina C de la Pulpa de Mango de Biomed. Life Sci. 830 (1), 41–46. Hilacha (Mangífera indica var Magdalena river). Inf. Tecnol. 26 (3), 9–16. Saedisomeolia, A., Ashoori, M., 2018. Riboflavin in human health: a review of current Morrissey, P.A., Hill, T.R., 2011. Vitamin C. In: Fuquay, J.W. (Ed.), Encyclopedia of Dairy evidences. Adv. Food Nutr. Res. 83, 57–81. Sciences, second ed. Academic Press, London, pp. 2683–2690. Saidi, B., Warthesen, J.J., 1995. Effect of heat and homogenization on riboflavin Nalame, N., Chaisri, W., Suriyasathaporn, W., 2009. Loss of L-ascorbic acid in photolysis in milk. Int. Dairy J. 5 (7), 635–645. commercial drinking milk caused by milk processing and storage times. Southeast Shaikh, S., O’Donnell, C., 2017. Applications of fluorescence spectroscopy in dairy Asian J. Trop. Med. Publ. Health 40 (4), 848–851. processing: a review. Curr. Opin. Food Sc 17, 16–24. Nisha, P., Singhal, R.S., Pandit, A.B., 2005a. A study on degradation kinetics of riboflavin Sheraz, M.A., Kazi, S.H., Ahmed, S., Anwar, Z., Ahmad, I., 2014. Photo, thermal and in green gram whole (Vigna radiata L.). Food Chem. 89 (4), 577–582. chemical degradation of riboflavin. Beilstein J. Org. Chem. 10, 1999–2012. Nisha, P., Singhal, R.S., Pandit, A.B., 2005b. A study on degradation kinetics of riboflavin Singh, R.P., Heldman, D.R., Kirk, J.R., 1975. Kinetic-analysis of light-induced riboflavin in spinach (Spinacea oleracea L.). J. Food Eng. 67 (4), 407–412. loss in whole milk. J. Food Sci. 40 (1), 164–167. Nohr, D., Biesalski, H.K., Back, E.I., 2011. Vitamins riboflavin. In: Fox, Patrick F., Stešková, A., Morochovičová, M., Lešková, E., 2006. Vitamin C degradation during McSweeney, P.L.H. (Eds.), Encyclopedia of Dairy Sciences, second ed. Elsevier, storage of fortified foods. J. Food Nutr. Res. 45 (2), 55–61. pp. 704–706. Sun, D.-W., Manzi, P., Pizzoferrato, L., 2012. Ultrahigh temperature thermal processing Ordóñez-Santos, L., Yoshioka-Tamayo, L., 2012. Cinética de degradación térmica de of milk. In: Sun, D.-W. (Ed.), Thermal Food Processing. New Technologies and vitamina C en pulpa de mango (Manguifera indica L). Vitae 19 (1), 81–83. Quality Issues, second ed., p. 686. Ordóñez, L.E., Portilla, M.A.O., Rodríguez, D.X.R., 2013. Cinética de degradación Saad, M.H., 1980. Vitamin C and Riboflavin in Milk and Some of its Products. MSc thesis. térmica de vitamina C en frutos de guayaba (Psidium guajava L.). Rev. Lasallista In: Alexandria University, Alexandria, Egypt. Invest. 10 (2), 44–51. Tamime, A.Y., 2009. In: Tamime, A.Y. (Ed.), Milk Processing and Quality Management, Oey, I., Van der Plancken, I., Van Loey, A., Hendrickx, M., 2008. Does high pressure first ed. processing influence nutritional aspects of plant based food systems? Trends Food Van Boekel, M.A.J.S., 2008. Kinetic modeling of food quality: a critical review. Compr. Sci. Technol. 19 (6), 300–308. Rev. Food Sci. Food Saf. 7 (1), 144–158. Poulsen, N., Rybicka, I., Poulsen, H.D., Larsen, L.B., Andersen, K.K., Larsen, M.K., 2015. Veisseyre, R., 1988. In: Acribia (Ed.), Lactología Técnica, second ed. Seasonal variation in content of riboflavin and major minerals in bulk milk from Verbeyst, L., Bogaerts, R., Van der Plancken, I., Hendrickx, M., Van Loey, A., 2013. three Danish dairies. Int. Dairy J. 42, 6–11. Modelling of vitamin C degradation during thermal and high-pressure treatments of red fruit. Food Bioprocess Technol. 6, 1015–1023. 9