Food Bioprocess Technol

DOI 10.1007/s11947-014-1464-x

ORIGINAL PAPER

Anthocyanins During Baking: Their Degradation Kinetics

and Impacts on Color and Antioxidant Capacity of Bread
Xiaonan Sui & Pei Yi Yap & Weibiao Zhou

Received: 1 May 2014 / Accepted: 18 December 2014

# Springer Science+Business Media New York 2015

Abstract Anthocyanin-rich black rice powder was incorpo- Keywords Black rice powder . Anthocyanins . Bread .
rated into bread, and the stability of two specific anthocyanins, Non-isothermal modeling . Antioxidant capacity . Total
cyanidin-3-glucoside and cyanidin-3-rutinoside, during bak- phenolic content . Baking
ing was investigated. Various baking conditions included
different baking temperatures (200, 220, and 240 °C) and
baking durations (0, 2, 4, 6, 8, 10, and 12 min). Non-
isothermal kinetic models were successfully established for
the two anthocyanins in both bread crumb and crust. The
derived degradation rate (kref) of cyanidin-3-glucoside and Introduction
cyanidin-3-rutinoside in bread crumb at the reference temper-
ature (Tref) of 65 °C were 2.49×10−5 and 1.93×10−5 s−1, Anthocyanins, a subgroup of water-soluble flavonoids which
respectively. The kref values of cyanidin-3-glucoside and occur naturally in the plant kingdom, are responsible for many
cyanidin-3-rutinoside in bread crust (Tref =125 °C) were of the colors observed in nature. Studies have demonstrated
5.37×10−4 and 5.72×10−4 s−1, respectively. The color devel- that anthocyanins can help in the prevention of cardiovascular
opment of bread crust and crumb was measured and expressed and neurological diseases, cancer, and inflammation
as L*C*H° values. While the color of bread crust was signif- (Konczak and Zhang 2004). They have also been shown to
icantly subjected to variations in oven operating parameters, play a role in obesity and diabetes control and in improving
the color of bread crumb was relatively less affected by baking visual functions (Tsuda 2012). Owing to their antioxidant
conditions. The antioxidant capacity and total phenolic con- capacity and the associated health benefits, anthocyanins have
tent of bread samples were measured using DPPH and Folin- gained increased attention in recent years.
Ciocalteu assays, respectively. Results showed that in bread Currently, a number of foods have been identified to be rich
crumb, both the antioxidant capacity and the total phenolic sources of anthocyanins. These include fruits such as berries,
content decreased; however, an increase in both was observed grapes, plums, and cherries (Fernandes et al. 2013) as well as
in bread crust. grains and vegetables such as black rice, black carrots, purple
sweet potatoes, and red cabbages (Wiczkowski et al. 2014).
According to Chun et al. (2007), the daily anthocyanin con-
sumption was found to be ranging from 3 to 15 mg/person in
the United States. This value is subjected to great variation
X. Sui : P. Y. Yap : W. Zhou due to differences in dietary choice. To increase the amount of
Food Science and Technology Programme, c/o Department of
anthocyanins consumed, many food products have been cre-
Chemistry, National University of Singapore, 3 Science Drive 3,
Singapore 117546, Singapore ated from anthocyanin-rich food materials or have been forti-
fied with anthocyanins such as cereals, tortillas, baby foods,
W. Zhou (*) soft drinks, and dairy products (Bueno et al. 2012; Shipp and
National University of Singapore (Suzhou) Research Institute, 377
Abdel-Aal 2010; Vera de Rosso and Mercadante 2007;
Linquan Street, Suzhou Industrial Park, Jiangsu 215123, People’s
Republic of China Hernández-Herrero and Frutos 2014a). As bakery products
e-mail: chmzwb@nus.edu.sg are among the most commonly consumed food products in the
 Food Bioprocess Technol

world, it is of considerable interest to incorporate anthocya- ABRP, cyanidin-3-glucoside and cyanidin-3-rutinoside

nins in bakery products to improve their value as a health- accounted for about 89.6 and 1.7 %, respectively (Sui and
promoting food. A recent study conducted by Rodriguez- Zhou 2014). Cyanidin-3-glucoside and cyanidin-3-rutinoside
Mateos et al. (2013) showed that anthocyanin content in standards were purchased from Polyphenols Laboratories
blueberry-enriched bread decreased significantly during the (Sandnes, Norway). DPPH (2,2-diphenyl-1-picrylhydrazyl),
baking process. However, the authors did not proceed to an in- formic acid, and Folin-Ciocalteu reagent were purchased from
depth study to develop mathematical models that enable the Sigma–Aldrich (Sigma–Aldrich, St Louis, MO, USA).
prediction of anthocyanin content under various baking con-
ditions. Moreover, the effect of baking on the antioxidant Bread Sample Preparation
properties of anthocyanin-rich bread was not studied. As the
antioxidant capacity of anthocyanin-rich foods is among the Bread samples were prepared based on a no-time bread-making
major properties to make them attractive, it is important to process described by Wang and Zhou (2004) with some mod-
study the effect of baking on their antioxidant properties ifications. In brief, anthocyanin-fortified flour was prepared by
which measure their ability to scavenge free radicals. In addi- incorporating ABRP into flour at a concentration of 2 g/kg
tion, the color development of the blueberry-enriched bakery flour. Bread dough was prepared by mixing 1 kg anthocyanin-
product was not addressed, which could be a critical factor in fortified flour, 590 g water, 40 g sugar, 30 g shortening, 12 g
determining consumers’ acceptance of the food product. salt, and 10 g instant dry active yeast at a slow speed for 1 min
Black rice (Oryza sativa L. indica) has been regarded as a followed by an intense mixing for 5 min in a mixer (WAG-
health-promoting food due to the large amount of anthocyanins RN20, Varimixer, Globe, US). After a resting period of 15 min
contained in its aleurone layer (Lee 2010). Cyanidin-3- at room temperature, the dough was divided and molded to 60 g
glucoside has been found to be the most abundant anthocyanin each using an electronic molder (DR Robot, Daub Bakery
in black rice, and it is also one of the most predominant Machinery B.V., Holland), followed by 70 min of proofing at
anthocyanins present in nature. Besides, cyanidin-3-rutinoside, 40 °C/85 % relative humidity in a proofer (Climatic chamber-
which also was found in black rice, has been found to be the KBF, Binder, Germany) before the dough proceeded to baking.
most thermally stable anthocyanin (Rubinskiene et al. 2005). The designated temperatures of the oven (Eurofours, France)
Hence, these two anthocyanins were selected as the focus of were 200, 220, and 240 °C, and the baking durations were 0, 2,
this study. By incorporating anthocyanin-rich powder extracted 4, 6, 8, 10, and 12 min at each temperature. Type T thermo-
from black rice to a standard formulation of bread, this project couples were placed in the oven, at dough surface (around
aimed to study the kinetic profile of anthocyanins during bread 1 mm below the dough surface), and at dough center to record
baking process. Based on the kinetic profile obtained, a math- the temperature profiles at those positions during baking. A
ematical model was developed which could allow the predic- diagram depicting the positions of type T thermocouples in
tion of the amount of anthocyanins in the fortified bread under bread and sampling area is shown in Fig. 1.
various baking conditions including different baking tempera-
tures and durations. Moreover, the antioxidant capacity and Moisture Content Analysis
total phenolic content of the bread before and after baking as
well as the color development of the bread crust and crumb Moisture content was analyzed based on AOAC Official
during baking were also evaluated. Method 945.15 (AOAC 2000) with minor modifications.
Around 2 g of sample were placed in a pre-dried and cooled

Methods

Materials and Chemicals

Wheat flour (11.7 % protein, 14.0 % moisture) was obtained

from Prima Ltd., Singapore. Pure cane sugar (fine grain,
NTUC Fairprice Cooperative Ltd., Singapore), fine salt
(NTUC Fairprice Cooperative Ltd., Singapore), vegetable
shortening (Bake King, Gim Hin Lee Ltd., Singapore), and
instant dry active yeast (Saccharomyces cerevisiae, S.I.
Lesaffre, France) were purchased from a local supermarket.
Commercial anthocyanin-rich black rice powder (ABRP) with
a total anthocyanin content of 20 % was acquired from Shaan- Fig. 1 A scheme showing the position of type T thermocouples and
xi Taiji Huaqing Technology Co., Ltd., China. Within the sampling area on bread
Food Bioprocess Technol

dish and heated in an oven (Memmert, Germany) at 100 °C disturbed, which renders them unsuitable for the subsequent
until a constant weight was achieved. At the end of drying, the sampling. Triplicate batches of bread baking were conducted.
sample was immediately transferred to a desiccator and Sample was freeze-dried, ground, and well mixed; after which,
weighed after it had reached room temperature. The moisture 5 g sample was placed in a centrifuge tube and macerated with
content was calculated based on the percentage change in the 10 mL acidified methanol (0.01 %v/v HCl in methanol) for
weight of the sample. 30 min. During maceration, the mixture was constantly agitated
using the orbital shaker at 200 rpm. Subsequently, the slurry was
pH Measurement filtered through Whatman No. 1 filter paper via vacuum suction
using a Buchner funnel to separate the liquid fraction, and the
The pH measurement was performed according to AOAC solid fraction remained on the filter paper was re-extracted for
Official Method 943.02 (AOAC 2000). Briefly, 5 g of sample several rounds (the number of extraction rounds was determined
were placed in a reagent bottle followed by the addition of 50 as described below). The liquid fraction was condensed by
mL deionized water. The mixture was agitated for 30 min evaporating methanol at 40 °C under vacuum. Afterwards, the
using an orbital shaker (IKA VXR basic Vibrax, Staufen, condensed liquid extract was made up to 5 mL with acidified
Germany) at 200 rpm and were left to stand for 10 min before deionized water (5 %v/v formic acid) and stored in the dark at
measuring the pH of the supernatant using a pH meter −20 °C until its anthocyanins, antioxidant capacity, and total
(Metrohm 744 pH meter, Switzerland). phenolics were measured. The three analytical methods were
performed on samples from the same bread liquid extract.
Color Measurement To determine the number of extraction rounds required to
extract all anthocyanins from the sample, 3, 4, 5, and 6 rounds
Color evaluation of sample was conducted using a spectro- of extraction were performed as described previously. It was
photometer (CM-5 spectrophotometer, Konica Minolta, found that 4 rounds of extraction were sufficient as it allowed
Japan) with D65 as the light source, and readings were the maximum extraction of anthocyanins at approximately
expressed as L* (lightness), C* (chroma or saturation), and 65 %, while 5 and 6 rounds of extraction did not show a
H° (hue) values. The L*C*H° system was adopted instead of significant increase in the amount of anthocyanins extracted
the commonly used L*a*b* system as the former allows a (data not shown). Hence, 4 rounds of sample extraction were
better interpretation and comparison of color shades (Homann conducted in this study. To determine the recovery rate of the
2009). In contrast, L*a*b* values do not express chroma and analysis method, freeze-dried plain dough powder was spiked
hue directly and are hard to interpret independently (Reyes with ABRP. Sample extraction and analysis were conducted as
and Cisneros-Zevallos 2007). The coordinates of CIE L*C*- described previously. The recovery rate was obtained based on
H° were computed from L*, a*, and b* values based on the the ratio of the measured amount of anthocyanin content to the
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
  original spiked amount of anthocyanins. The recovery rate of
following equations: C * ¼ a*2 þ b*2 and H°=tan− the anthocyanins was found to be above 97 % (data not shown).
1
b*/a*, where L* represents the lightness measuring bright-
ness with 100 and 0 corresponding to absolute white and Quantification of Anthocyanins Using HPLC/DAD
absolute black, respectively. Chroma (C*) measures intensity
or saturation, and H° indicates hue angle which is expressed Analysis of anthocyanins was carried out using an analytical
on a 360° grid, with 0° and 180° corresponding to +a* axis C18 Sunfire column (250×4.6 mm/5 μm; Waters, Wexford,
(red) and −a* (green), respectively, 90° and 270° for the +b* Ireland) on a Shimadzu HPLC system (Shimadzu, Japan)
axis (yellow) and −b* (blue), respectively. For both crust and equipped with a diode array detector (DAD). Elute A was
crumb of bread, triplicate measurements at different spots composed of 5 % formic acid (v/v) in deionized water and
were performed for each sample to account for non- elute B was 100 % acetonitrile. A gradient elution process was
homogeneous color distribution in the crust and the crumb. adopted as follows: 0 % B for 5 min, followed by a linear
increase to 10 % B in the next 15 min, 13 % at 40 min, 20 % at
Extraction of Anthocyanins 44 min, 25 % at 50 min, and 100 % at 55 min. Prior to HPLC
analysis, sample extract (1 mL) was filtered through a
Sample extraction was conducted according to the method by 0.45-μm Nylon filter (Whatman, NJ, USA). The sample in-
Rodriguez-Saona and Wrolstad (2001). As described in jection volume was 50 μL, while the flow rate and column
Section 2.2, bread was sampled at the baking durations of 0, 2, temperature were set at 1 mL/min and 25 °C, respectively.
4, 6, 8, 10, and 12 min for each of the three baking temperatures, Detection of anthocyanins was performed at 520 nm, and
i.e., 200, 220, and 240 °C. For every sampling time, a new batch quantification of anthocyanins was achieved through external
of bread baking was necessary as, otherwise, the temperature calibration. The standard curves were produced using a series
profile of those non-sampled ones would have been severely of standard solutions (prepared by dissolving anthocyanin
 Food Bioprocess Technol

standards in DI water) between 0.0001 and 0.1 mg/mL for the (8.314 J mol−1 ·K−1); T is absolute temperature (K); and Tref is
two anthocyanins. the reference temperature (K). In this study, Tref was chosen to
be 125 and 65 °C for crust and crumb samples, respectively,
Non-Isothermal Kinetic Modeling which were the average values of the temperature range they
experienced during baking.
Based on the previous study conducted by Sui and Zhou During baking, the temperature of bread increased gradu-
(2014), it was assumed that the thermal degradation of the ally; thus, the baking process was a non-isothermal process.
two specific anthocyanins in this experiment followed a first- By rearranging Eq. 1, change in the anthocyanin concentration
order reaction, which was to be verified by examining the over a small time interval Δti =ti+1 −ti during which tempera-
modeling results. The first order kinetic model could be de- ture was nearly constant (Ti) could be expressed as in Eq. 3.
scribed using Eq. 1.
C iþ1 ¼ C i exp ð−k i Δt i Þ ð3Þ
Ct
ln ¼ −kt ð1Þ
C0

where Ct is the concentration of anthocyanins in mg/L at time where Ci+1 is the anthocyanin concentration at time ti+1; Ci
t; C0 is the initial concentration of anthocyanins in mg/L; and k is the anthocyanin concentration at time ti; and ki is rate
is the rate constant in s−1. Temperature dependence of the constant at each time interval (Δti) with temperature Ti. Based
degradation rate constant of the anthocyanins was represented on Eq. 3, the final anthocyanin concentration (Cend) at the end
by the Arrhenius model (Eq. 2) as previously suggested by of a non-isothermal process could be calculated using Eq. (4).
others (Sólyom et al. 2014).
   C end ¼ C 0 ½exp ð−k 0 Δt0 Þ  exp ð−k 1 Δt 1 Þ…  exp ð−k end Δt end Þ
Ea 1 1 ¼ C 0 exp ½−ðk 0 Δt 0 þ k 1 Δt1 þ ⋯ þ k end Δtend Þ
k t ¼ k re f exp − − ð2Þ
R T t T re f ð4Þ

where kref is the rate constant (s−1) at reference temperature

(Tref); Ea is activation energy (kJ/mol); R is ideal gas constant Substituting Eq. 2 into Eq. 4:

       
Ea 1 1 Ea 1 1
C end ¼ C 0 exp − k re f exp − − Δt 0 þ k re f exp − − Δt 1
  R T 0 Tref  R T 1 T re f
ð5Þ
Ea 1 1
þ⋯ þ k re f exp − − Δt end
R T end T re f

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
The activation energy (Ea) of anthocyanin degradation in a u1 X m
2
RMSE ¼ t C iMod −C iExp ð6Þ
bread system was assumed to be independent of the medium m i¼0
environment and equivalent to that in an aqueous system
(Wang et al. 2008). To find the optimum value of kref, the Ea
values of cyanidin-3-glucoside and cyanidin-3-rutinoside
were taken to be at 92.5 and 91.5 kJ/mol, respectively, ac- where CMod is modeled value; CExp is experimental value; m
cording to our previous work evaluating the stability of is the number of experimental value.
cyanidin-3-glucoside and cyanidin-3-rutinoside in an aqueous
system at pH 5 (Sui et al. 2014). Matlab software (version Antioxidant Capacity Analysis
8.1.0.604 R2013a, the MathWorks Inc., Natick, Massachu-
setts, USA) was used to conduct nonlinear regression to find DPPH assay was performed according to the method reported
the optimum value of kref through Eq. 5 by minimizing the by Brand-Williams et al. (1995). Briefly, 0.1 mL bread liquid
root mean square error (RMSE) between the experimental and extract, which had been obtained using the method described in
modeled values using Eq. 6. The generated RMSE value was Section 2.6 or Trolox was mixed with 3.9 mL DPPH stock
further used to examine the quality of the established mathe- solution (60 μM). The mixture was allowed to stand in the dark
matical model. at room temperature for 2 h. The absorbance of the mixture was
Food Bioprocess Technol

then measured at 515 nm using the spectrophotometer. A way analysis of variance (ANOVA) followed by Tukey’s test
Trolox calibration curve was constructed, and the antioxidant (P<0.05) was performed using SPSS software (version 22.0,
capacity of the sample was expressed as Trolox equivalent. SPSS Inc., Chicago, USA) to evaluate the differences among
groups.
Total Phenolic Analysis

Folin-Ciocalteu assay was performed based on the method

reported by Waterhouse (2001). Twenty-microliter bread liq- Results and Discussion
uid extract, which had been obtained using the method de-
scribed in Section 2.6, 1.58 mL deionized water, and 100 Temperature-time Profile
μL Folin-Ciocalteu reagent were mixed in a cuvette, and the
mixture was incubated for 5 min in the dark. Three hundred- The temperature profiles during the baking processes under
microliter sodium carbonate solution (0.7 M) was then added the three stipulated temperatures, 200, 220, and 240 °C are
into the mixture and was kept in the dark for 2 h at room shown in Fig. 2. It was observed that the oven temperature
temperature. The absorbance of the mixture was measured at oscillated considerably during baking due to the nature of the
765 nm using the spectrophotometer. A gallic acid calibration oven’s temperature control system. According to Ryckaert
curve was constructed, and the total phenolic content of the et al. (1999), the temperature of an operating oven could
sample was expressed as gallic acid equivalent. deviate greatly from the set point and large fluctuations in
temperature could also occur. Despite the large temperature
Statistical Analysis deviation observed in this study, the extent of variation was
observed to be similar for the three temperature set points.
All analyses were performed in at least triplicate, and the The temperature increase in the crumb showed a sigmoidal
results were presented as mean ± standard deviation. One- pattern with an inflection point at around 50 °C, after which

Fig. 2 Temperature profiles of oven, crumb, and crust during baking

 Food Bioprocess Technol

the temperature increased gradually before becoming constant content of the crust sample decreased significantly with longer
at 100 °C. Owing to the presence of water, the crumb temper- baking duration. During baking, the bread surface was ex-
ature maintained largely at 100 °C (Eliasson 2003). In con- posed to the high oven temperature and hot air; hence, rapid
trast, the temperature increase in the crust was divided into water evaporation could occur. Additionally, the partial water
two stages. During the first 2 min, most of the heat energy vapor pressure of the surrounding air was far from saturation,
supplied was used to increase the crust temperature. This allowing diffusion of water vapor into the air. This caused the
explains the rapid increase in crust temperature to 100– surface to dry out, resulting in crust formation (Therdthai and
120 °C during the initial period of the baking process. After Zhou 2003).
that, the evaporation period commenced where a more gradual
temperature increase was observed as the heat energy was pH Variation
used to evaporate water from the sample (Feyissa et al.
2011). As heat transfer occurred more rapidly in the crust, Previous studies have shown that pH is one of the factors that
the crust temperature was observed to be much higher than the affect anthocyanin stability (Cavalcanti et al. 2011;
crumb temperature. During baking, volume expansion oc- Hernández-Herrero and Frutos 2014b). In this study, the pH
curred as the embedded bubbles inflated, resulting in oven of the bread samples ranged narrowly from pH 4.9–5.2 (data
spring. At the same time, rapid moisture loss occurred on the not shown). Thus, it could be deduced that the difference in
dough surface due to the high temperature, resulting in crust anthocyanin degradation rate observed in this study was nei-
formation when moisture was exhausted (Zhou and Therdthai ther due to pH change of the bread, nor resulting in a signif-
2007). icant change in the bread’s pH.

Moisture Content Analysis Color Development

The moisture contents of the crumb samples under various A general decreasing trend was observed for both L* and C*
baking conditions were statistically the same (P>0.05) for all values of the crumb sample (Fig. 3, dash lines). This suggests
the three baking temperatures (Table 1). This result was con- that the crumb became darker and a less vivid color was
sistent with the study by Purlis and Salvadori (2010), in which observed with increased baking duration. The reduced L*
the moisture content of crumb was shown to be stable and value could be attributed to the absorption of light by the
independent of baking duration. This could be attributed to the crumb due to the nature of its structure. With an increased
presence of two generalized forces in the crumb, which were a baking duration, a more developed gluten network was
result of the moisture concentration gradient and temperature formed in the crumb, which was more likely to absorb light
gradient (Wang et al. 2008). During baking, moisture traveled due to the presence of air cells with uneven sizes and deep
from the wetter inner region to the drier outer surface due to cavities. Subsequently, light reflection was impeded and
the concentration gradient. Conversely, the temperature gra- crumb appeared to be comparatively darker (NPCS Board of
dient served as an opposing force and drove water vapor Consultants and Engineers 2011). Similarly, the decreased C*
toward the inner region of the bread. Owing to the relatively value could be explained by the development of gluten net-
lower temperature of the inner region, water vapor condensed. work which led to a more expanded structure, resulting in a
Consequently, the moisture content of crumb remained fairly reduced color intensity. In general, the crumb sample appeared
constant during the baking process. However, the moisture to transform from a more bluish shade to a more yellowish

Table 1 Moisture content (%) in crumb and crust samples under various baking conditions*

Control 2 min 4 min 6 min 8 min 10 min 12 min

Crumb
200 °C 44.4±0.5a 42.3±0.3b 42.4±0.1b 42.5±0.1b 42.5±0.2b 42.6±0.3b 42.3±0.1b
220 °C 44.4±0.5a 42.3±0.1b 42.3±0.3b 42.3±0.5b 42.6±0.4b 42.2±0.5b 42.6±0.2b
240 °C 44.4±0.5a 42.4±0.4b 42.8±0.1b 42.6±0.3b 42.6±0.1b 42.4±0.3b 42.4±0.1b
Crust
200 °C 44.4±0.5a 26.1±0.9b 21.3±0.5c 20.5±0.3c 19.8±0.9c 17.8±0.1d 16.4±0.6d
220 °C 44.4±0.5a 25.1±0.3b 21.1±1.0c 19.7±0.4cd 18.0±0.6de 16.5±0.9ef 14.6±1.5f
240 °C 44.4±0.5a 23.4±1.2b 20.5±0.8c 18.3±1.0cd 16.4±0.8de 14.7±0.8ef 13.3±0.9f

*Values are reported as mean ± standard deviation (SD) of triplicate samples. Values in the same row followed by different superscript letters are
significantly different from each other (P<0.05). Control refers to proved dough without baking
Food Bioprocess Technol

Fig. 3 Color profile of bread crumb (dash lines) and crust (solid lines) under different baking conditions: 200 °C, 220 °C, 240 °C,
200 °C, 220 °C, and 240 °C

shade as indicated by its increased H° value as the baking regarded as the lag phase where the surface condition was
duration increased. This could be associated with anthocyanin insufficient for browning reactions to occur (Purlis and
degradation during the baking process as increased H° value Salvadori 2009; Tan and Zhou 2009). During the later stage
usually indicates the degradation of anthocyanins to yellow of baking, Maillard reactions and caramelization commenced,
chalcone species (Reyes and Cisneros-Zevallos 2007; Yang resulting in the darkening of the crust layer. The formation of
et al. 2008). Maillard reaction products (MRPs) such as melanoidin and
The color change in crust (Fig. 3, solid lines) was more hydroxymethylfurfural (HMF) conferred the crust layer its
evident than that in crumb. Contrary to the crumb which was characteristic brown color (Purlis and Salvadori 2007).
relatively insulated and less affected by baking condition, the An initial rise in C* value was observed during the first
crust layer was more subjected to variations in the oven 4 min of baking and subsequent baking resulted in a near
operating parameters. Consistent with other studies, L* value constant or a decline in the value. This could be associated
of the crust sample showed an initial rise followed by a with crust browning as MRPs were formed, causing the
gradual decrease as the baking duration increased. Purlis and C* value to increase which corresponded to the increased
Salvadori (2007) attributed this observation to physical chang- intensity of brown shade of the crust layer. Moreover, the
es of the dough surface during the early stage of baking. After decreased C* value of the crust under baking tempera-
the proofing process, dough surface appeared creased. Upon tures of 220 and 240 °C indicated the burning of the
baking, the surface texture became smoother due to volume crust layer as the baking duration increased. The effects
increase of the dough. This change in surface texture could be of Maillard reactions and caramelization also explained
the explanation for the initial increase in lightness of the the increased H° value of the crust as browning of the
sample as a greater amount of light could be reflected from a crust layer occurred. For visualizing the color change of
smooth surface as compared to a creased surface. Surface bread during baking, the top, side, and cross-section
drying could also contribute to the lightening of the sample views of the bread samples with various baking durations
(Shibukawa et al. 1989). This early stage of baking could be at 220 °C are presented in Fig. 4.
 Food Bioprocess Technol

Fig. 4 Top, side, and cross-

section view of breads baked at
220 °C. From left to right: dough,
2, 4, 6, 8, 10, and 12 min of
baking

Non-isothermal Kinetic Modeling reported by Kırca et al. (2007) when the solid content of a
black carrot juice was increased.
Kinetic models are commonly employed for an objective and The higher anthocyanin stability in bread could be associ-
efficient prediction of the influence of thermal processing on ated with the lower availability of oxygen in the system.
critical quality parameters. Knowledge of kinetic parameters During bread preparation, oxygen was assimilated by yeast
such as the degradation rate constant, activation energy, and in the production of carbon dioxide which helped in the rising
frequency factor is imperative to allow the prediction of of dough (Sharma and Zhou 2011). This minimized anthocy-
quality or nutrient loss in foods during thermal processing anin oxidation. According to Cavalcanti et al. (2011), antho-
(Patras et al. 2010). cyanin degradation could be greatly enhanced in the presence
The optimized kinetic parameters for the two anthocyanins of oxygen, and elimination of oxygen had been shown to
in the crumb and crust are shown in Table 2. The kref values of reduce thermal degradation of anthocyanins in berry juices.
the two anthocyanins, in bread system at Tref =65 and 125 °C, When comparing the kref values of anthocyanins in the crumb
obtained in this study were found to be much lower than those and crust, as expected, the degradation rates of anthocyanins
in aqueous system; for example, in a model blackberry juice, were found to be lower in the crumb. As anthocyanins are
kref was 0.9 to 3.5×10−3 s−1 for total monomeric anthocyanins thermally unstable, anthocyanins in the crumb were less ther-
at Tref =120 °C and water activity from 0.99 to 0.34 (Jiménez mally degraded due to the much lower temperature profile of
et al. 2012), and in a model blackcurrant juice system, it was the crumb than the crust.
2.8×10−3 s−1 at 140 °C for total monomeric anthocyanins In addition, cyanidin-3-rutinoside was observed to be more
(Harbourne et al. 2008). This suggests that anthocyanins thermally stable in the crumb as compared to cyanidin-3-glu-
exhibit lower degradation rates when present in a bread sys- coside. As demonstrated by Rubinskiene et al. (2005), retention
tem. Although the higher solid content in bread was thought to of cyanidin-3-rutinoside was found to be higher (65 %) than
be a major factor for the decreased degradation rates observed, cyanidin-3-glucoside (47 %) when both anthocyanin extracts
many studies reported otherwise. A study conducted by Wang were heated at 95 °C. The difference in thermal stability could
and Xu (2007) showed that increased solid content in black- be attributed to their chemical structures as anthocyanin disac-
berry concentrate resulted in an increase in the degradation charide derivatives have been reported to be more stable than
rate. Decreased thermal stability of anthocyanins was also monoglycoside derivatives (Fleschhut et al. 2006). However,
cyanidin-3-rutinoside exhibited a slightly higher degradation
rate than cyanidin-3-glucoside at the reference temperature
Table 2 Kinetic parameters of cyanidin-3-glucoside and cyanidin-3- (5.72×10−4 s−1 vs. 5.37×10−4 s−1). This could be due to the
rutinoside in crumb and crust
significantly higher temperature profile of the crust which the
Ea (kJ/mol) kref (s−1) RMSE two anthocyanins experienced, resulting in their remaining
amounts to be very low and, therefore, their degradation rates
Crumb Cyanidin-3- 92.5 2.49×10−5 0.0013
to be both high with diminished difference.
(Tref =65 °C) glucoside
Cyanidin-3- 91.5 1.93×10−5 7.4×10−5
rutinoside Model Validation
Crust Cyanidin-3- 92.5 5.37×10−4 0.0023
(Tref =125 °C) glucoside
As shown in Table 2, the low values of RMSE obtained
Cyanidin-3- 91.5 5.72×10−4 1.4×10−4
rutinoside indicate that the differences between the modeled data and
the experimental values were very small, suggesting that the
Food Bioprocess Technol

Fig. 5 Modeled values against

experimental values for a
cyanidin-3-glucoside and b
cyanidin-3-rutinoside content in
bread crumb; c cyanidin-3-
glucoside and d cyanidin-3-
rutinoside content in bread crust

model quality is good. To illustrate the quality of the models, regression, samples processed at 200 °C and 8 min, 220 °C
the modeled values against experimental values were plotted and 10 min, and 240 °C and 6 min were selected for validating
in Fig. 5. The even distribution of these data points around the the models. The validation results (Fig. 6) clearly show that
45° lines demonstrates a good agreement between the exper- the modeled results for the stability of anthocyanins in bread
imental and modeled results. To exclude the possibility that under the three selected validating conditions were in good
the developed mathematical models might have been over- agreement with the experimental results, indicating that the
fitted to the experimental results caused by the nonlinear developed models are valid and of high quality.

Fig. 6 Model validation plots. a

cyanidin-3-glucoside and b
cyanidin-3-rutinoside content in
bread crumb; c cyanidin-3-
glucoside and d cyanidin-3-
rutinoside content in bread crust.
As shown in plot (a), symbols in
the four plots, from left to right
represent anthocyanin content at
the baking condition of 200 °C
and 8 min, 220 °C and 10 min,
and 240 °C and 6 min,
respectively
 Food Bioprocess Technol

Table 3 ANOVA analysis of the antioxidant capacity (mg Trolox/100 g Sample)*

Control 2 min 4 min 6 min 8 min 10 min 12 min

Crumb
200 °C 20.2±0.7a 18.8±0.2ab 17.4±1.2bc 17.4±0.5bc 17.7±0.3bc 17.5±0.6bc 16.8±0.6c
220 °C 20.2±0.7a 18.8±0.2ab 19.7±0.4a 18.8±0.9ab 17.2±0.8b 17.5±0.8b 17.1±0.3b
240 °C 20.2±0.7a 19.4±0.1ab 17.0±1.1bc 16.8±1.3bc 18.3±0.8abc 16.0±1.9c 16.8±0.8bc
Crust
200 °C 20.2±0.7a 16.1±2.6c 16.9±0.7abc 16.8±1.0bc 17.6±0.3abc 19.6±1.0ab 20.0±0.7ab
220 °C 20.2±0.7abc 17.3±2.6c 19.5±1.0bc 21.3±1.2ab 22.2±1.3ab 23.0±0.5ab 23.2±0.4a
240 °C 20.2±0.7c 16.6±0.3d 21.6±0.1b 22.1±0.3b 23.3±0.3a 23.3±0.5a 23.4±0.2a

*Significant differences of values within the same row are indicated by different superscript letters (P<0.05)

Antioxidant Capacity and Total Phenolic Content large molecules which could be less extractable (Peng et al.
2010). In addition, furfural derivatives generated during
In general, thermal degradation of anthocyanins begins with caramelization and breakdown of sugar could undergo con-
the hydrolysis of sugar moieties, such as glucoside and densation with phenolic compounds, which may lead to the
rutinoside. This results in deglycosylation of anthocyanins to loss of total phenolics in the crumb, and subsequently, a
form anthocyanidins, which are further degraded to chalcones. decrease in the antioxidant capacity (Holtekjølen et al. 2008).
Subsequent breakdown of chalcones results in the formation Similarly, decreased antioxidant capacity and total phe-
of phenolic acids and carboxyaldehydes (Hendry and nolic content were observed in the crust samples during
Houghton 1996). It has been suggested that the degradation the first 2 min of baking. The decrease was more signif-
products of anthocyanins retain antioxidant properties; and icant in the crust as compared to that in the crumb.
hence, thermal degradation may not have a significant impact However, further baking resulted in an increase in the
on their antioxidant capacity (Slavin et al. 2013). antioxidant capacity and total phenolic content in the crust.
It was observed that the antioxidant capacity and total This observation might be attributed to the generation of
phenolic content of the crumb samples generally decreased MRPs in the crust layer during baking, which also
after baking (Tables 3 and 4). This observation was in contrast corresponded to the browning of the crust layer after
with what reported by Sui and Zhou (2014), which showed 2 min of baking at all the three baking temperatures as
negligible losses in the antioxidant capacity of aqueous solu- indicated by the decreased L* value. It has been well
tions of anthocyanins after thermal treatment at various tem- documented that MRPs possess antioxidant properties in-
peratures (100–165 °C). Generation of other phenolic com- cluding radical chain breaking activity, decomposition of
pounds from anthocyanin degradation was suggested to be the hydrogen peroxide as well as scavenging of oxygen rad-
reason for the nearly constant antioxidant capacity observed in icals (Vhangani and Van Wyk 2013). Hence, it could be
the aqueous solutions. However, in a complex system such as deduced that the increased antioxidant capacity and total
bread, the phenolic compounds produced could complex with phenolic content observed in the crust were a result of the
proteins and carbohydrates present in the system, forming MRPs formed during baking.

Table 4 ANOVA analysis of the total phenolic content (mg gallic acid/100 g sample)*

Control 2 min 4 min 6 min 8 min 10 min 12 min

Crumb
200 °C 42.8±1.1a 43.3±2.2a 33.9±3.9b 36.5±1.1b 38.5±1.2b 36.9±1.6b 33.1±2.2b
220 °C 42.8±1.1a 40.5±1.2ab 36.4±2.8b 37.3±1.7ab 36.8±2.5b 35.4±2.5b 33.3±1.8b
240 °C 42.8±1.1a 36.3±4.3ab 34.2±3.0b 33.5±5.3b 34.5±1.1ab 32.9±2.0b 33.7±1.5b
Crust
200 °C 42.8±1.1bc 32.7±6.3c 43.4±2.3b 45.3±3.5b 48.9±1.4b 63.6±2.9a 62.0±5.6a
220 °C 42.8±1.1c 32.3±2.4c 62.3±6.7b 70.3±2.2ab 72.7±3.8ab 77.5±5.7a 77.0±5.6a
240 °C 42.8±1.1cd 38.5±4.1d 64.9±5.5bc 73.0±4.4b 84.0±13.4ab 88.0±11.9ab 102.4±13.4a

*Significant differences of values within the same row are indicated by different superscript letters (P<0.05)
Food Bioprocess Technol

Conclusion Feyissa, A. H., Gernaey, K. V., Ashokkumar, S., & Adler-Nissen, J.

(2011). Modelling of coupled heat and mass transfer during a
contact baking process. Journal of Food Engineering, 106, 228–
Non-isothermal kinetic models were successfully developed 235.
for cyanidin-3-glucoside and cyanidin-3-rutinoside degrada- Fleschhut, J., Kratzer, F., Rechkemmer, G., & Kulling, S. E. (2006).
tions during bread baking process. The anthocyanins exhibit- Stability and biotransformation of various dietary anthocyanins
in vitro. European Journal of Nutrition, 45, 7–18.
ed lower degradation rates when present in a bread system as
Harbourne, N., Jacquier, J. C., Morgan, D. J., & Lyng, J. G. (2008).
compared to aqueous systems. The two anthocyanins showed Determination of the degradation kinetics of anthocyanins in a
different degradation rates and kinetic parameters; cyanidin-3- model juice system using isothermal and non-isothermal methods.
rutinoside was found to be more thermally stable. Bread Food Chemistry, 111, 204–208.
crumb was shown to display a shielding effect against thermal Hendry, G. A. F., & Houghton, J. D. (1996). Anthocyanins and betalains.
In G. A. F. Hendry & J. D. Houghton (Eds.), Natural food colorants
degradation of anthocyanins. Significant changes in moisture (2nd ed., p. 261). Glasgow: Blackie Academic and Professional.
content and color were observed in the bread crust after Hernández-Herrero, J. A., & Frutos, M. J. (2014a). Effect of concentrated
baking. The antioxidant capacity and total phenolic content plum juice on physicochemical and sensory properties of yoghurt
of the anthocyanin-fortified bread crumb decreased after bak- made at bench-top scale. International Journal o f Diary
Technology, 67, 123–128.
ing while an increase was observed in the crust. While the
Hernández-Herrero, J. A., & Frutos, M. J. (2014b). Color and antioxidant
decrease might be due to the low extractability and loss of capacity stability in grape, strawberry and plum peel model juices at
phenolic compounds, the increase was likely a result of the different pHs and temperatures. Food Chemistry, 154, 199–204.
MRPs generated. Further research needs to be carried out to Holtekjølen, A. K., Bævre, A. B., Rødbotten, M., Berg, H., & Knutsen, S.
identify the major compounds that are generated during the H. (2008). Antioxidant properties and sensory profiles of breads
containing barley flour. Food Chemistry, 110, 414–421.
baking of bread fortified with anthocyanins. Besides, it will Homann, J.-P. (2009). Color theory with realistic colors. In J.-P. Homann
also be necessary to conduct sensory tests to evaluate the (Ed.), Digital color management: principles and strategies for the
bread fortified with anthocyanin-rich black rice powder to standardized print production (p. 44). Berlin: Springer.
determine its acceptability to consumers. Jiménez, N., Bohuon, P., Dornier, M., Bonazzi, C., Pérez, A. M., &
Vaillant, F. (2012). Effect of water activity on anthocyanin degrada-
tion and browning kinetics at high temperatures (100–140 °C). Food
Acknowledgments Support from A*STAR through research grant
Research International, 47, 106–115.
SERC 112 177 0038 and the National University of Singapore
Kırca, A., Özkan, M., & Cemeroğlu, B. (2007). Effects of temperature,
(Suzhou) Research Institute under the grant number NUSRI2011-007 is
solid content and pH on the stability of black carrot anthocyanins.
gratefully acknowledged. The first author also likes to thank China
Food Chemistry, 101, 212–218.
Scholarship Council (CSC) and the National University of Singapore
Konczak, I., & Zhang, W. (2004). Anthocyanins—more than nature’s
(NUS) for financial support.
colors. Journal of Biomedicine and Biotechnology, 5, 239–240.
Lee, J. H. (2010). Identification and quantification of anthocyanins from
the grains of black rice (Oryza sativa L.) varieties. Food Science and
References Biotechnology, 19, 391–397.
NPCS Board of Consultants and Engineers. (2011). Handbook on
fermented foods and chemicals. Delhi: Asia Pacific Business Press.
AOAC. (2000). Official methods of analysis of AOAC international (17th Patras, A., Brunton, N. P., O’Donnell, C., & Tiwari, B. K. (2010). Effect
ed.). Gaithersburg: AOAC International. of thermal processing on anthocyanin stability in foods; mechanisms
Brand-Williams, W., Cuvelier, M. E., & Berset, C. (1995). Use of a free and kinetics of degradation. Trends in Food Science and
radical method to evaluate antioxidant activity. LWT - Food Science Technology, 21, 3–11.
and Technology, 28, 25–30. Peng, X., Ma, J., Cheng, K.-W., Jiang, Y., Chen, F., & Wang, M. (2010).
Bueno, J. M., Saez-Plaza, P., Ramos-Escudero, F., Jimenez, A. M., Fett, The effects of grape seed extract fortification on the antioxidant
R., & Asuero, A. G. (2012). Analysis and antioxidant capacity of activity and quality attributes of bread. Food Chemistry, 119, 49–53.
anthocyanin pigments. Part II: chemical structure, color, and intake Purlis, E., & Salvadori, V. O. (2007). Bread browning kinetics during
of anthocyanins. Critical Reviews in Analytical Chemistry, 42, 126– baking. Journal of Food Engineering, 80, 1107–1115.
151. Purlis, E., & Salvadori, V. O. (2009). Modelling the browning of bread
Cavalcanti, R. N., Santos, D. T., & Meireles, M. A. A. (2011). Non- during baking. Food Research International, 42, 865–870.
thermal stabilization mechanisms of anthocyanins in model and Purlis, E., & Salvadori, V. O. (2010). A moving boundary problem in a
food systems—an overview. Food Research International, 44, food material undergoing volume change—simulation of bread
499–509. baking. Food Research International, 43, 949–958.
Chun, O. K., Chung, S. J., & Song, W. O. (2007). Estimated dietary Reyes, L. F., & Cisneros-Zevallos, L. (2007). Degradation kinetics and
flavonoid intakes and major food sources of U.S. adults. Journal of color of anthocyanins in aqueous extracts of purple- and red-flesh
Nutrition, 137, 1244–1252. potatoes (Solanum tuberosum L.). Food Chemistry, 100, 885–894.
Eliasson, A.-C. (2003). Utilization of thermal properties for understand- Rodriguez-Mateos, A., Cifuentes-Gomez, T., George, T. W., & Spencer,
ing baking and staling processes. In G. Kaletunc & K. J. Breslauer J. P. E. (2013). Impact of cooking, proving, and baking on the
(Eds.), Characterization of cereals and flours: properties, analysis (poly)phenol content of wild blueberry. Journal of Agricultural
and applications (pp. 65–66). New York: Marcel Dekker. and Food Chemistry. doi:10.1021/jf403366q.
Fernandes, I., Marques, F., de Freitas, V., & Mateus, N. (2013). Rodriguez-Saona, L. E., & Wrolstad, R. E. (2001). Extraction, isolation,
Antioxidant and antiproliferative properties of methylated metabo- and purification of anthocyanins. In R. E. Wrolstad (Ed.), Current
lites of anthocyanins. Food Chemistry, 141, 2923–2933. protocols in food analytical chemistry. New York: Wiley.
 Food Bioprocess Technol

Rubinskiene, M., Jasutiene, I., Venskutonis, P. R., & Viskelis, P. (2005). Tsuda, T. (2012). Dietary anthocyanin-rich plants: biochemical basis and
HPLC determination of the composition and stability of recent progress in health benefits studies. Molecular Nutrition and
blackcurrant anthocyanins. Journal of Chromatographic Science, Food Research, 56, 159–170.
43, 478–482. Vera de Rosso, V., & Mercadante, A. Z. (2007). Evaluation of color and
Ryckaert, V. G., Claes, J. E., & Van Impe, J. F. (1999). Model-based stability of anthocyanins from tropical fruits in an isotonic soft drink
temperature control in ovens. Journal of Food Engineering, 39, 47– system. Innovative Food Science and Emerging Technologies, 8,
58. 347–352.
Sharma, A., & Zhou, W. (2011). A stability study of green tea Vhangani, L. N., & Van Wyk, J. (2013). Antioxidant activity of Maillard
catechins during the biscuit making process. Food Chemistry, reaction products (MRPs) derived from fructose-lysine and ribose-
126, 568–573. lysine model systems. Food Chemistry, 137, 92–98.
Shibukawa, S., Sugiyama, K., & Yano, T. (1989). Effects of heat transfer Wang, W.-D., & Xu, S.-Y. (2007). Degradation kinetics of anthocyanins
by radiation and convection on browning of cookies at baking. in blackberry juice and concentrate. Journal of Food Engineering,
Journal of Food Science, 54, 621–624. 82, 271–275.
Shipp, J., & Abdel-Aal, E.-S. M. (2010). Food applications and physio- Wang, R., & Zhou, W. (2004). Stability of tea catechins in the
logical effects of anthocyanins as functional food ingredients. Open breadmaking process. Journal of Agricultural and Food
Food Science Journal, 4, 7–22. Chemistry, 52, 8224–8229.
Slavin, M., Lu, Y., Kaplan, N., & Yu, L. (2013). Effects of baking on
Wang, R., Zhou, W., & Jiang, X. (2008). Mathematical modeling of the
cyanidin-3-glucoside content and antioxidant properties of black
stability of green tea catechin epigallocatechin gallate (EGCG)
and yellow soybean crackers. Food Chemistry, 141, 1166–1174.
during bread baking. Journal of Food Engineering, 87, 505–513.
Sólyom, K., Solá, R., Cocero, M. J., & Mato, R. B. (2014). Thermal
degradation of grape marc polyphenols. Food Chemistry, 159, 361– Waterhouse, A. L. (2001). Determination of total phenolics. In R. E.
366. Wrolstad (Ed.), Current protocols in food analytical chemistry.
Sui, X., & Zhou, W. (2014). Monte Carlo modelling of non-isothermal New York: Wiley.
degradation of two cyanidin-based anthocyanins in aqueous system Wiczkowski, W., Topolska, J., & Honke, J. (2014). Anthocyanins profile
at high temperatures and its impact on antioxidant capacities. Food and antioxidant capacity of red cabbages are influenced by genotype
Chemistry, 148, 342–350. and vegetation period. Journal of Functional Foods, 7, 201–211.
Sui, X., Dong, X., & Zhou, W. (2014). Combined effect of pH and high Yang, Z., Han, Y., Gu, Z., Fan, G., & Chen, Z. (2008). Thermal degra-
temperature on the stability and antioxidant capacity of two antho- dation kinetics of aqueous anthocyanins and visual color of purple
cyanins in aqueous solution. Food Chemistry, 163, 163–170. corn (Zea mays L.) cob. Innovative Food Science and Emerging
Tan, M. Y., & Zhou, W. (2009). Modelling of bread colour development Technologies, 9, 341–347.
during baking. Food Manufacturing Efficiency, 2(2), 9–15. Zhou, W., & Therdthai, N. (2007). Three-dimensional CFD modeling of a
Therdthai, N., & Zhou, W. (2003). Recent advances in the studies of continuous industrial baking processes. In D.-W. Sun (Ed.),
bread baking process and their impacts on the bread baking tech- Computational fluid dynamics in food processing (pp. 288–289).
nology. Food Science and Technology Research, 9, 219–226. Boca Raton: CRC Press.