antioxidants

Article
Microwave Pretreatment for the Extraction of Anthocyanins
from Saffron Flowers: Assessment of Product Quality
Ana Álvarez, Sara Terreros, María J. Cocero and Rafael B. Mato *

BioEcoUva, Research Institute on Bioeconomy, High Pressure Processes Group, Department of Chemical
Engineering and Environmental Technology, University of Valladolid, 47011 Valladolid, Spain;
[email protected] (A.Á.); [email protected] (S.T.); [email protected] (M.J.C.)
* Correspondence: [email protected]; Tel.: +34-983-423169

Abstract: The potential of saffron flowers as a source of polyphenols, and in particular anthocyanins,
for the extraction of bioactive compounds and the production of a cyanic colorant was analyzed. A
microwave pretreatment, prior to the conventional solid–liquid extraction process, was proposed as
a feasible intensification step. The effectiveness of microwave pretreatment was assessed in terms of
increased yield and improved quality of the final product. The operational variables studied were
the pretreatment temperature (60–120 ◦ C) and the solid–liquid ratio (0.30–0.50 g/mL). It was found
that the addition of the microwave pretreatment to the conventional process allowed one to reduce
extraction time by up to 12 times and to greatly improve the characteristics of the final product,
using microwave energy densities as low as 0.16–0.54 kJ/mL. The extract quality was evaluated in
terms of polyphenol richness (25% increase), product composition (80% of the anthocyanins was
delphinidin), antioxidant capacity (boosted by the pretreatment) and color (variations in red and



blue hue depending on conditions). To conclude, a microwave pretreatment in which the material is

 heated to a temperature of 65 ◦ C with a solvent ratio of 0.30 g/mL was selected as the optimum to
Citation: Álvarez, A.; Terreros, S.; maximize process efficiency and product quality.
Cocero, M.J.; Mato, R.B. Microwave
Pretreatment for the Extraction of Keywords: saffron flowers; microwave pretreatment; solid–liquid ratio; kinetic analysis; color
Anthocyanins from Saffron Flowers:
Assessment of Product Quality.
Antioxidants 2021, 10, 1054. https://
doi.org/10.3390/antiox10071054 1. Introduction
Extracts from natural products have drawn the attention of food industry as an
Academic Editor: Francisco J. Barba
alternative to synthetic complements due to the antioxidant properties and health benefits
of their main components: polyphenols [1,2]. Saffron biofloral residues have recently been
Received: 30 April 2021
Accepted: 25 June 2021
nominated as an outstanding source of polyphenols, especially anthocyanins, due to their
Published: 29 June 2021
great potential and availability [3,4]. The production of the spice entails a large generation
of residues. In particular, 68 L of flowers is needed to produce 1 kg of spice [5]. The
Publisher’s Note: MDPI stays neutral
petals and stigmas constitute this byproduct, which has no real use but can be valorized to
with regard to jurisdictional claims in
improve the sustainability of the process.
published maps and institutional affil- A rich polyphenolic extract stemming from saffron flowers would be highly valued in
iations. the food industry as a natural colorant. Actual market trends show a predilection for such
tints instead of synthetic ones because of their health benefits [6–8]. In the case of saffron
flowers, it is not only their high content of antioxidant compounds that makes them a
convenient food ingredient [9], but also their composition in minerals, dietary fiber, sugars,
Copyright: © 2021 by the authors.
anions and organic acids [10,11]. In addition, anthocyanins are the phytochemicals that
Licensee MDPI, Basel, Switzerland.
play the most important role as food additives due to their ability to scavenge superoxide
This article is an open access article
radicals [11] and the cyanic hues they exhibit [12]. Blue colorants are rare in nature [7].
distributed under the terms and Few examples have been described in literature, such as the blue pigments isolated from
conditions of the Creative Commons Port red wine [13]. Most natural blue colorants are now under research [14]. Among the
Attribution (CC BY) license (https:// anthocyanins, delphinidin and petunidin derivatives have been reported to provide the
creativecommons.org/licenses/by/ cyanic colors (purple, lilac, mauve and blue) in saffron flowers. Malvidin is also present,
4.0/). but in a lower percentage [15–17].

Antioxidants 2021, 10, 1054. https://doi.org/10.3390/antiox10071054 https://www.mdpi.com/journal/antioxidants

Antioxidants 2021, 10, 1054 2 of 19

According to previous work, there is a correlation between the concentration of active

compounds and color [18], and thus color must be an important attribute to take into
account. To characterize colors for industrial applications, the use of the CIELAB space
is recognized as an international standard [18]. In this standard, a sphere is defined with
axes that correspond to lightness (L*), green vs. redness (a*) and yellow vs. bluish (b*) [4].
Table 1 specifies the range of values that these variables can have and their meaning.

Table 1. CIELAB parameters meaning.

Lightness Hue
L=0 black a* > 0 red b* > 0 yellow
L = 100 white a* < 0 green b* < 0 blue

Most of the published works on the valorization of saffron flower residues focus on
their in vivo properties, thus demonstrating the potential of the extracts as chemopreven-
tive agents [8,19–23]. As for the extraction process, conventional liquid extraction is not
very convenient, as it requires low solid–liquid ratios and long extraction times. Low
solid–liquid ratios imply high solvent consumption, which is contrary to the principles of
green chemistry [24]. Long extraction times (prolonged for more than one hour) seem quite
excessive for the extraction of polyphenols from petals, since their structure is not that
complex to represent a large mass transfer resistance. These two unfavorable characteristics
favored the application of intensification techniques: high-pressure extraction (subcritical
water and supercritical carbon dioxide) [3,25,26], enzymatic extraction [27], ultrasound-
assisted extraction (UAE) [28] and microwave-assisted extraction (MAE) [28,29]. The latter
is particularly interesting because of the good results published in the literature, but the
extraction conditions have not been analyzed as extensively, and the few available results
present a notable disparity. Table 2 presents a review of optimized extraction conditions
for conventional solid–liquid extraction processes.

Table 2. Yields and optimum conditions of conventional solid–liquid extraction reported in the literature.

S:L Ratio
Raw Material Solvent T (◦ C) t (min) Yield Reference
(g/mL)
Freeze-dried and Water:HCl 13.57 ± 1.16
1/30 25 60 [17]
milled saffron flowers 100:1 (v/v) mgGAE /gDry SF
Ethanol:Water 11.34 ± 3.00
Saffron flowers 1/3 66 104 [30]
59% (v/v) mgGAE /gDry SF
Dried and milled Ethanol:acidified water
1/15 25 1440 1712 mgCGE /L [31]
saffron flowers 50% (v/v)
Dried and milled Ethanol:Water
1/20 25 1440 1609 mgCGE /L [32]
saffron flowers 25% (v/v)

The different authors who have optimized the conventional extraction process have
reported very different extraction conditions. While in some works the maximum yield is
obtained with water as solvent and at ambient temperature [17], others find that a mixture
of organic and water solvents at high temperature provides the best efficiency [11,30].
However, when the same temperature and ethanol percentage suggested as optimal for
conventional solid–liquid extraction by Ahmadian-Kouchaksaraie et al. [30] were used by
Da Porto and Natolino [28], the obtained yields using the same conventional extraction
process were four-fold higher. On the other hand, all authors agree that an acidified solvent
improves the extraction and stability of the active compounds, especially anthocyanins [33].
To clarify the effect of the microwaves on the extraction of bioactive compounds from
saffron flowers, a new study is required where all the parameters of interest (polyphenol
richness, product composition, antioxidant capacity and color) are analyzed and where not
Antioxidants 2021, 10, 1054 3 of 19

only the external operational variables are characterized but also the efficiency of the mi-
crowave heating. For this purpose, the absorbed energy density (Qabsorbed , kJ/mL) [34] is
reported in the present paper for all experiments. The product of the power and irradiation
time selected for each experiment in the microwave oven determines the total amount of
energy irradiated, but of this, only a fraction is absorbed by the sample. This fraction (ab-
sorption efficiency) usually ranges between 40% and 95%, depending on the experimental
conditions (characteristics and geometry of the oven, amount and location of the sample in
the oven, solvent composition, etc.). The calculation of the microwave absorbed energy,
as described in Equations (1)–(4), helps elucidate whether the differences in the results
reported by different authors may come from different qualities of raw materials or from
different microwave absorption efficiencies, as the latter depends on many variables.
This work focused on the optimization of microwave extraction. However, in contrast
to usual MAE processes, in this study, radiation has been used as a previous step to a
conventional solid–liquid extraction. Bench-scale MAE experiments typically employ
temperature control involving an initial power spike, to rapidly reach set temperature
conditions, and an intermittent short power radiation to maintain those conditions [35].
This, together with the difficulties in adequately scaling a large cavity, has led to a preference
for the use of microwaves as a brief pretreatment. By providing the extraction sample with
a peak of microwave energy, a high temperature in a short time is achieved. Thus, the
possible disruption of the matrix and the rapid extraction of the compounds take place [36].
In addition, due to their short duration, the degradation of thermosensitive compounds is
very low if pretreatment is followed by rapid cooling [37]. Regarding the scale-up for the
industrial application, the use of a low residence time in the microwave allows the design
of a compact oven that can be easily implemented at a larger scale. In addition, previous
work has shown that a microwave pretreatment improves not only the process yield, but
also the polyphenol richness of the final product [38].
Thus, this work is aimed to develop an efficient and feasible scale-up process to obtain
a functional antioxidant colorant.

2. Materials and Methods

2.1. Raw Material
Saffron flowers were donated by a local farmer. Saffron was collected before sunlight
in October 2017 and manually processed. Floral byproducts were immediately frozen
and stored at −20 ◦ C. Before being used, flowers were thawed for one hour at 4 ◦ C.
Table 3 gathers the characterization of the raw material employed in this work. Moisture
was determined gravimetrically by drying the sample until constant weight at 105 ◦ C. Fat
and extractives were quantified by a Soxhlet extraction with n-hexane, and ethanol and
water, respectively. The protein content was computed by the Kjeldahl method, using a
conversion factor of 6.25 [39]. Ashes were measured by means of the char formed at 550 ◦ C.
The fiber content is assumed to be 173.61 mg/gSF , which is the unanalyzed remainder, and
is in accordance with literature data [40].

Table 3. Saffron flowers proximate composition.

Moisture Fat Ash Protein Extractives

(gH2O /gDry SF ) (mg/gDry SF ) (mg/gDry SF ) (mg/gDry SF ) (mg/gDry SF )
5.096 ± 0.001 17.69 ± 0.70 36.28 ± 3.46 45.66 ± 0.30 726.76 ± 10.18

Saffron heat capacity was measured in a Setaram Micro DSC II microcalorimeter [41]
from 20 to 85 ◦ C and was found to be 3.89 ± 0.02 J/g K.

2.2. Chemical Reagents

APPH (2,2-azobis 2-amidopropane dihydrochloride), gallic acid, sodium fluorescein
and trolox were purchased from Sigma. Acetonitrile, ammonium phosphate monobasic,
chlorohydric acid, ethanol, Folin-Ciocalteu reagent, phosphate salts (NaH2 PO4 ·2H2 O
Antioxidants 2021, 10, 1054 4 of 19

and Na2 HPO4 ·12H2 O), phosphoric acid, potassium chloride, sodium acetate and sodium
carbonate were obtained from Panreac. HPLC standards catechin, delphinidin, epicatechin,
gallic acid, malvidin and quercetin were bought from Extrasynthese, except for kaempferol
which was purchased from TCI. A Millipore unit was used to purify the water used as
solvent.

2.3. Extraction Procedure

A conventional solid–liquid extraction process was used as a control reference to
assess the influence of the microwave pretreatment. In all experiments, an amount of
30 g of saffron flowers was weighed and mixed with the corresponding volume of solvent
to achieve the desired solid–liquid ratio. An ethanol–water mixture was used as solvent.
Ethanol was chosen as the organic solvent because of its GRAS (generally recognized as
safe by the FDA) certification [5]. To improve the yield and stability of the anthocyanins,
water acidified to pH = 2 with hydrochloric acid was used [11].
Once the saffron flowers were mixed with the solvent, they were homogenized for
5 min at room temperature and stirred gently. The extraction itself was set to begin when
the flask was place in a thermostatic bath at the set temperature and with vigorous stirring.
For experiments in which a microwave pretreatment was added to the process, it was
integrated after the homogenization and before the thermostatic and stirring bath. The
microwave pretreatment was carried out in a CEM Discover One unit (CEM Corp). It
was devised as an intense and short pretreatment, ease to scale up for a future industrial
application. Thus, the maximum power, 300 W, was irradiated for shorter times than
traditional MAE, between 40 and 145 s, to reach temperatures of 60, 80, 100 and 120 ◦ C.
These temperatures were measured with an optical fiber (FOTEMP4, OPTcon GmbH),
weekly calibrated at 0 ◦ C with a bath of ice and distilled water. The pretreatments at 60
and 80 ◦ C were conducted in an open vessel round bottom flask, whereas for the 100 and
120 ◦ C, a pressure cell was needed to avoid sample evaporation by boiling. A 100 mL
QianCap glass pressure vessel (QLabtech) was used.
Unlike most published work on MAE, microwave pretreatment has been quantified
in terms of absorbed energy density rather than using the radiation conditions selected in
the oven. The absorbed energy density (Qabsorbed ) was calculated by means of an energy
balance, based on the two main contributions: sensible heat and latent heats, as expressed
in Equation (1). Losses to the environment have been dismissed in accordance to Sólyom’s
conclusions [34].
Qabsorbed = Qsensible + Qlatent (1)
The sensible heat (Qsensible ) was evaluated using Equation (2), which calculates the
energy required to increase the temperature of the extraction media (solvent and saffron
flower) to the final value specified in the pretreatment.

Qsensible = ∑ m Cp ∆T (2)

where m is the mass of every component in the sample, Cp its specific heat, and ∆T the
temperature rise experienced during the pretreatment.
The latent heat (Qlatent ) is calculated according to Equation (3).

Qlatent = nevaporated λ (3)

where λ is the solvent vaporization heat.

The number of moles of solvent evaporated (nevaporated ) was calculated using two
different procedures, depending on whether the pretreatment was carried out in an open
vessel or in a pressure cell. For the open vessel experiments (pretreatments up to 60
and 80 ◦ C), it was calculated by weight loss. However, when the specified pretreatment
temperature exceeded the normal boiling point of the solvent (pretreatments up to 100 and
120 ◦ C), pretreatment had to be carried out in a closed vessel. In this case, the evaporated
solvent accumulates in the closed gas space, increasing the pressure inside the vessel. The
Antioxidants 2021, 10, 1054 5 of 19

number of evaporated moles in this case was calculated from the registered maximum final
pressure (P) reached during the pretreatment, assuming ideal gas behavior:

PV0 − RTn0
nevaporated = (4)
RT − ρ P
mol

where V0 is the empty cell volume, R the gas constant, n0 the number of air moles in the
initial gas space, T the maximum final temperature, V0 the volume of the gas space (cell
minus liquid volume), and ρmol the molar density of the liquid solvent. Liquid composition
changes due to evaporation were neglected since only a very low fraction of the solvent is
evaporated (between the 0.18% and 0.35%).
Once the pretreatment was finished, the media was rapidly cooled down using an
ice-water bath. In those cases where the solid–liquid ratio used in the pretreatment was
different from the one used in the subsequent extraction stage, the extra cold solvent
was added just after the pretreatment to accelerate the cooling stage. Previous work has
shown that the use of different solid–liquid ratios can improve extraction efficiency [38].
The medium was then placed in the thermostatic bath, and the process continued as in a
conventional solid–liquid extraction for 180 min. Liquid samples were taken throughout the
process to determine the extraction kinetics and richness. The calculated concentration data
(ccal ) were adjusted to the experimental values (cexp ) using the first order kinetic equation
described in Equation (5), minimizing the average relative deviation (ARD) defined in
Equation (6).
ccal = co + cf [1 − exp(−kt)] (5)
1 cexp − ccal
ARD =
n ∑ cexp
(6)

Concentrations are expressed in mg/gdry saffron flowers . The regressed parameters in

Equation (5) were the initial concentration c0 , the pre-exponential factor cf , and the rate
constant k.
The use of this kinetic model makes it easy to assess the extraction efficiency of a
particular extraction process by means of the derived parameters initial extraction rate
(u0 ) and yield (c∞ ). The initial extraction rate, calculated by Equation (7), represents the
extraction rate at t = 0 (start of the extraction period after the pretreatment). The higher
this parameter is, the shorter the time required to reach maximum yield.

∂ccal
uo = = cf k (7)
∂t t=0

The yield (c∞ , Equation (8)) determines the maximum concentration of polyphenols
that can be achieved after endless extraction.

c∞ = Limccal = c0 + cf (8)
t→ ∞

The polyphenol and anthocyanin richness were also calculated to characterize the
extract. These parameters indicate the fraction of polyphenols or anthocyanins in the total
amount of solid extract product, since they are also mixed with sugars, fibers, etc. Since
extracts are usually commercialized as solid products, their specific richness is crucial
as it determines the quality and price of the final product. Richness was expressed in
mg/gDry Extract , and it was calculated from the polyphenols or anthocyanins concentration
(mg/gExtract ) and the solid extract residue (gDry Extract /gExtract ).

2.4. pH Measurement
pH throughout the extraction was recorded by a Jenway 3505 pH meter.
Antioxidants 2021, 10, 1054 6 of 19

2.5. Extract Characterization

2.5.1. Total Polyphenol Content
The Folin–Ciocalteu method was used to determine the total polyphenol content
(TPC). In brief, a volume of 40 µL of sample was mixed with 3000 µL of distillate water and
200 µL of Folin–Ciocalteu reagent. After 5 min, 600 µL of 20% sodium carbonate solution
was added. The mixtures were left for 30 min at 40 ◦ C before recording their absorbance
at 765 nm (Shimadzu UV/VIS spectrophotometer). Results were expressed in gallic acid
equivalents (mgGAE /g).

2.5.2. Anthocyanin Content

The pH differential method was used to quantify the anthocyanin content (AC).
Samples were diluted in two buffers: one at pH = 1 of potassium chloride 0.025 M and
another at pH = 4.5 of sodium acetate 0.4 M. The increase in absorbance between 520 and
700 nm was used together with an extinction coefficient of 26,900 L/(mol cm) to calculate
the anthocyanin content, expressed in cyaniding-glucoside equivalents (mgCGE /g).

2.5.3. Solid Residue

The solid residue was determined gravimetrically by drying the extracts at 105 ◦ C
overnight.

2.5.4. Antioxidant Capacity

Antioxidant activity (AAO) was computed as the ability of the extracts to quench
oxygen radicals from the decomposition of AAPH (α,α0 -azodiisobutyramidine dihydrochlo-
ride, 240 mM). Sodium fluorescein (100 nM) was used as probe and trolox as the standard
(200 µM). Phosphate buffer (10 mM, pH = 7.4) was used as solvent in the assays.
In a 96-well plate, 150 µL of fluorescein and 25 µL of sample (standard or diluted
extract) were poured. After 30 min incubation at 37 ◦ C, 25 µL of APPH was added.
Fluorescence was then recorded in a Fluorstar Optima (BMG Labtech) at an emission
wavelength of 530 ± 25 nm and excitation wavelength 485 ± 20 nm, until the signal
variation was null. Trolox calibration was repeated for each assay, and samples were
measured at least six times to minimize experimental error. Results were expressed as
trolox equivalents (µmolTE /g).

2.5.5. HPLC
Catechin, delphinidin, epicatechin, gallic acid, kaempferol, malvidin and quercetin
were quantified. A Waters e2695 separation module and a Waters 2998 photodiode array
detector (DAD) were employed. A previous method was used as reference [42]. A volume
of 20 µL of sample was injected into the 1 mL/min eluent flow and passed through a
Teknokrima C18 (250 × 4.6 mm, 5 µm) column dotted with an OptiGuard precolumn at
35 ◦ C. The eluents used were (A) ammonium phosphate monobasic 50 mM and pH = 2.6,
(B) a mixture of 80% acetonitrile and 20% of eluent A, and (C) acidified water with phos-
phoric acid to pH = 1.5. The eluent gradient is detailed in Table 4. DAD signals were
recorded at 360 and 520 nm as well as the UV/VIS spectra. As they were very complex
samples, the whole spectrum was recorded to compare it with the standard. The gallic acid
was measured at 280 nm. Compound identification was performed by comparing retention
time and UV/VIS spectra with the standards.
Antioxidants 2021, 10, 1054 7 of 19

Table 4. Eluent gradient.

Time (min) Eluent A 1 (%) Eluent B 2 (%) Eluent C 3 (%)

0 100 0 0
2 100 0 0
5 92 8 0
17 0 14 86
22 0 18 82
29.5 0 21 79
55 0 33 67
70 0 50 50
75 0 50 50
78 20 80 0
81 20 80 0
86 100 0 0
1 NH4 H2 PO4 50 mM, pH = 2.6; 2 80% acetonitrile + 20% eluent A; 3 H3 PO4 200 mM, pH = 1.5.

2.6. Color Measurement

A spectrophotometric method [43] was used to determine extract colors. This method
allowed to compute color coordinates in the XYZ space by applying Equations (9)–(11)
with the measured transmittances (T) at the selected wavelengths (see Appendix A). These
measurements correspond to an illuminant C and a 2◦ standard observer.

X = 0.03269 ∑ TX (9)

X = 0.03333 ∑ TY (10)
X = 0.03938 ∑ TZ (11)
To avoid errors due to different dilutions of the extracts, the samples were dried and
diluted again to have a similar concentration. Liquid samples were concentrated in a
rotary evaporator (Heidolph) to remove the ethanol and freeze-dried at −50 ◦ C and 0.1 bar
(Telstar LyoQuest). The dried extracts were then rediluted to a concentration of about
500 ppm. This final solution was placed in a ShimadzuUV/vis spectrophotometer where a
transmittance scan from 700 to 400 nm was performed. Since CIELAB is the most extended
color space in food industry, XYZ coordinates were transformed into the CIELAB ones (L*,
a*, b*) by means of Equations (12)–(14). In these expressions, Xn , Yn and Zn represent the
illuminant C tristimulus, which for the conditions used here have values of Xn = 98.0681,
Yn = 100 and Zn = 118.2313.
L∗ = 116(Y/Yn )1/3 − 16 (12)
h i
a∗ = 500 (X/Xn )1/3 − (Y/Yn )1/3 (13)
h i
b∗ = 200 (Y/Yn )1/3 − (Z/Zn )1/3 (14)

It must be noticed that, since no direct measurement of the final product was made
(but a dilution), these measurements only allow for comparison of color changes. For ease
of comparison, the total color difference (∆E in Equation (15)) was used. In this equation,
the subindex i denotates the analyzed sample, while 0 stands for the control. To assess the
influence of microwave pretreatment on product color, conventional solid–liquid extraction
(without microwaves) was used as the control.
q

2 2
∆E = (Li∗ − L0∗ )2 + ai∗ − a0∗ + (bi∗ − b0∗ ) (15)

For an easy CIELAB parameters interpretation, the software JMP has been employed
to estimate ellipses with a 90% coverage of the experimental dispersion.
Antioxidants 2021, 10, 1054 8 of 19

2.7. Experimental Design and Statistical Analysis

The pretreatment variables studied were the solid–liquid ratio (S:L) and the final
temperature at the end of the microwave pretreatment (TMW ). Experimental conditions
are gathered in Table 5. A factorial experimental design was employed. ANOVA tables
with a confidence of 95% (p-value of 0.05) were developed to determine significant effects.
To compare the influence of variables in a similar range, coded variables have been used
(−1/0/+1). They are identified in parentheses after the variable values. Design Expert
software has been used for this analysis.

Table 5. Experimental pretreatment operating variables and extraction kinetics results.

Pretreatment Conditions Total Polyphenols Anthocyanins Extraction

Run Energy C∞ u0 C∞ u0
S:L (g/mL) TMW (◦ C)
(kJ/mL) (mgGAE /gDry SF ) (mgGAE /gDry SF /min) (mgGCE /gDry SF ) (mgGCE /gDry SF /min)
1 0.50 (0) 80 (0) 0.39 26.62 2.85 6.97 0.58
2 0.50 (0) 80 (0) 0.40 29.00 2.56 8.00 0.58
3 0.50 (0) 100 (+1) 0.51 24.99 2.45 6.97 0.58
4 0.50 (0) 80 (0) 0.37 27.68 3.00 8.43 0.72
5 0.70 (+1) 60 (−1) 0.31 21.05 1.96 5.00 0.56
6 0.30 (−1) 80 (0) 0.35 28.12 10.45 7.94 2.51
7 0.50 (0) 80 (0) 0.36 26.62 1.77 6.70 0.49
8 0.30 (−1) 100 (+1) 0.51 29.53 9.25 8.05 2.38
9 0.70 (+1) 100 (+1) 0.51 24.17 2.57 7.17 0.68
10 0.50 (0) 60 (−1) 0.29 23.01 3.52 9.34 0.97
11 0.30 (−1) 60 (−1) 0.23 23.82 1.62 7.62 0.42
12 0.70 (+1) 80 (0) 0.36 27.64 1.14 6.93 0.34
13 0.70 (+1) 80 (0) 0.40 26.01 7.11 6.52 1.52
14 0.30 (−1) 120 (+2) 0.58 28.62 10.98 5.51 1.83
15 0.70 (+1) 120 (+2) 0.58 25.64 1.03 4.55 0.15
16 0.50 (0) 120 (+2) 0.64 26.68 0.98 5.06 0.18
Control – – 26.05 1.23 6.80 0.20

3. Results and Discussion

3.1. Conventional Solid–Liquid Extraction
The experimental conditions (temperature, solvent composition and solid–liquid ratio)
used to carry out the conventional solid–liquid extractions, both in the control experiments
and to complete the extraction in the experiments with microwave pre-treatment, were
selected based on the analysis of three inputs: (1) the optimal conditions reported in the
literature and listed in Table 2 for conventional extraction processes; (2) industrial and scale-
up aspects, such as solvent consumption, extract concentration or the use of saffron flowers
without any conditioning (grinding or drying); and (3) some preliminary experimental
work, presented below in Figure 1, to evaluate the influence of the solid–liquid ratio on
extraction efficiency.
Regarding extraction temperature, previous authors (Table 2) suggested the conve-
nience of employing extraction temperatures of 30 and 60 ◦ C. Similar polyphenol yields and
richness were achieved in both cases. However, an anthocyanin yield decrease was found
at 60 ◦ C due to the thermal degradation of these compounds after 40 min of extraction. For
these reasons, 30 ◦ C was selected as the extraction temperature.
As for the solvent, water was found to greatly enhance the extraction of anthocyanins
but not polyphenols to the same extent. Therefore, the use of only water as a solvent
was discarded in favor of a 50% ethanol:water solution (water acidified at pH = 2), since
this mixture allowed one to achieve a convenient compromise between polyphenol and
anthocyanin extraction yields.
 iments and to complete the extraction in the experiments with microwave pre-treatment,
were selected based on the analysis of three inputs: (1) the optimal conditions reported in
the literature and listed in Table 2 for conventional extraction processes; (2) industrial and
scale-up aspects, such as solvent consumption, extract concentration or the use of saffron
Antioxidants 2021, flowers
10, 1054 without any conditioning (grinding or drying); and (3) some preliminary experi- 9 of 19
mental work, presented below in Figure 1, to evaluate the influence of the solid–liquid
ratio on extraction efficiency.

Figure 1. Solid–liquid
Figure 1. Solid–liquid ratio influence
ratio influence on conventional
on conventional solid–liquid
solid–liquid extraction
extraction for (A) polyphenol
for (A) polyphenol and (B) anthocyanin
and (B) anthocyanin extraction. ● S:L = 0.10 g/mL, ■ S:L = 0.20 g/mL, ▲ S:L = 0.30 g/mL, ▼
extraction. • S:L = 0.10 g/mL,  S:L = 0.20 g/mL, N S:L = 0.30 g/mL, H S:L = 0.50 g/mL,  S:L = 0.75 g/mL. S:L =
0.50 g/mL, ♦ S:L = 0.75 g/mL.
The solid–liquid ratio was the most influential factor; therefore, a thorough analysis
Regardingwas extraction
performed.temperature, previous
Ratios ranging fromauthors
0.10 to(Table 2) suggested
0.75 g/mL the conven-
were studied. Results can be found
ience of employing extraction
in Figure 1. temperatures of 30 and 60 °C. Similar polyphenol yields
and richness were achieved
Differentinfinal
both cases.were
yields However,
obtained anfor
anthocyanin yield decrease
total polyphenols, waswith increasing
decreasing
found at 60 °C S:Ldueratio.
to the thermal degradation of these compounds after
The influence of solvent composition on mass transfer is considered 40 minutes of as the reason
extraction. For these reasons,
for this behavior.30 °C
Forwaslowselected
S:L valuesas the extraction
(large volumetemperature.
of solvent), depletion of the extractives
As PEER
Antioxidants 2021, 10, x FOR for the solvent,
could bewater
REVIEW was found
achieved because to of
greatly enhance
the large the extraction
concentration of anthocyanins
gradient between the solid 10 and
of the
21
but not polyphenols
dilutedto the same However,
solvent. extent. Therefore, the use of
these extracted only water were
polyphenols as a solvent was diluted in the
also highly
discarded in favor of a 50%
solvent; ethanol:water
therefore, solution (water
a low concentrated acidified
liquid product atwas
pH =finally
2), since this
obtained. By contrast,
mixture allowed one
whenthe to achieve
a low a convenient
solvent andvolume compromise
was used (high between
S:L ratio), polyphenol
a smaller and an-
concentration gradient
drove extraction, consequently, a smaller fraction of polyphenols was extracted.
thocyanin extraction
droveyields.
the extraction, and consequently, a smaller fraction of polyphenols was extracted. As
As the polyphenols accumulated in the low-volume liquid phase, a more concentrated
The solid–liquid ratio was the
the polyphenols most influential
accumulated factor; therefore,
in the low-volume liquidaphase,
thorough analysis
a more concentrated extract
extract could be obtained.
was performed.couldRatiosberanging from 0.10 to 0.75 g/mL were studied. Results can be found
obtained.
Another aspect considered was the likely influence of the acidified solvent. Since
in Figure 1. Another aspect considered was the likely influence of the acidified solvent. Since
acidified water was employed as a solvent constituent, the presence of these ions could
acidified
Different final yieldswater was employed
were obtained for totalaspolyphenols,
a solvent constituent,
decreasing the presence
with of these ions could
increasing
interfere with the structure of saffron, reducing the binding of polyphenols and enhancing
S:L ratio. The influence of solvent composition on mass transfer is considered as the reason and enhancing
interfere with the structure of saffron, reducing the binding of polyphenols
their release. Figure 2 represents the pH evolution during the extraction at different S:L
for this behavior.their
For release. Figure(large
low S:L values 2 represents
volume of the pH evolution
solvent), depletion during
of thethe extraction at different
extractives
ratios. Lower ratios entailed a larger solvent consumption, and therefore, the pH main-
could be achieved S:L because
ratios. Lower ratiosconcentration
of the large entailed a larger solvent
gradient consumption,
between the solid andand the
therefore, the pH
tained during the extraction closer to that of the fresh solvent (pH = 2). The greater con-
diluted solvent.maintained
However, these duringextracted
the extraction closer towere
polyphenols that also
of the fresh diluted
highly solvent (pH
in the = 2). The greater
centration of the hydronium ion present in the extraction media improved the extraction
concentration
solvent; therefore, of the hydronium
a low concentrated ion present
liquid product wasinfinally
the extraction
obtained.media improved the extraction
By contrast,
of compounds by acid hydrolysis.
when a low solventof compounds
volume was by acid
usedhydrolysis.
(high S:L ratio), a smaller concentration gradient
pH

Figure
Figure2.2.
pHpH evolution during
evolution the extraction
during for a solid
the extraction for aliquid
solidratio of ●ratio
liquid S:L =of0.10 g/mL,
• S:L ■ S:Lg/mL,
= 0.10 =
0.20 g/mL,
 S:L = 0.20 ▲ S:Land
andg/mL, = 0.30 g/mL.
N S:L = 0.30 g/mL.

Unlike total polyphenols, anthocyanin extraction achieved almost the same final
yield regardless of the solvent (except in the case of 0.75 g/mL, which was proved to be
limited by solubility). This indicated that anthocyanin was easily released from saffron
Antioxidants 2021, 10, 1054 10 of 19

Unlike total polyphenols, anthocyanin extraction achieved almost the same final yield
regardless of the solvent (except in the case of 0.75 g/mL, which was proved to be limited
by solubility). This indicated that anthocyanin was easily released from saffron flowers.
Finally, a S:L ratio of 0.30 g/mL was selected as the optimum. Although this ratio
did not lead to the highest yield of polyphenols (23% less was achieved), it made it pos-
sible to obtain a much more concentrated extract. Specifically, 230% more polyphenols
was obtained per liter of solvent (9.6 mgGAE/mL compared to 4.1 mgGAE/mL when
0.10 g/mL was used) and 276% more in the case of anthocyanins (2.2 mgGGE/mL com-
pared to 0.8 mgGGE/mL when 0.10 g/mL was used). This increase is interesting for
favoring the profitability and sustainability of the process, since much less energy is re-
quired to evaporate the solvent to obtain the final solid extract.

3.2. Microwave Pretreatment Extraction

This section analyzes the effect on the kinetics of adding microwave pretreatment to
conventional solid–liquid extraction, the main results of which are shown in Tables 5 and 6.
The table allows one to identify the effects of each variable (linear and quadratic), as
well as possible interactions.

Table 6. Extraction results: richness, antioxidant activity (AAO) and color measurement.

TPC Richness AC Richness Delphinidin Richness AAO

Run L* a* b* ∆E
(mgGAE /gDry Extract ) (mgCGE /gDry Extract ) (mg/gDry extract ) (mmolTE /gDry Extract )
1 45.51 11.71 7.43 1.09 ± 0.06 87.54 29.18 −11.23 3.82
2 43.69 11.76 7.71 1.96 ± 0.22 86.97 29.04 −9.46 4.35
3 44.40 9.88 6.04 2.30 ± 0.08 88.49 24.77 −8.78 8.75
4 47.50 13.36 7.58 1.42 ± 0.08 85.78 31.08 −11.53 1.47
5 41.52 10.58 5.99 2.04 ± 0.23 88.80 24.15 −8.83 9.38
6 42.08 11.74 6.37 1.61 ± 0.15 82.07 37.81 −13.43 6.67
7 41.95 10.23 5.49 1.56 ± 0.11 87.96 25.14 −8.98 8.20
8 45.94 12.52 7.34 1.43 ± 0.10 86.44 29.13 −10.78 3.61
9 44.02 12.37 7.89 1.11 ± 0.08 86.98 29.61 −10.72 3.33
10 42.80 14.18 6.12 1.32 ± 0.08 82.79 39.24 −14.82 7.95
11 44.78 12.01 5.03 0.74 ± 0.04 84.82 34.06 −12.58 1.98
12 42.27 10.89 6.28 1.04 ± 0.10 88.94 22.46 −4.46 12.85
13 45.18 11.51 5.38 1.32 ± 0.08 87.89 28.04 −11.36 4.97
14 52.21 10.06 6.64 1.70 ± 0.11 88.63 26.67 −10.43 6.64
15 45.35 8.28 6.49 0.77 ± 0.09 87.81 28.31 −11.34 4.69
16 53.34 9.88 5.98 1.27 ± 0.07 88.41 26.75 −9.94 6.61
Control 41.82 12.00 6.51 1.04 ± 0.04 85.80 32.53 −11.80 0.00

The operating variables of the pretreatment, S:L and TMW , determined the total
microwave energy absorbed by the extraction media, shown in Table 5. This is the value
that must be considered when designing an industrial scale oven. The absorbed energy
density in these experiments ranged between 0.23 and 0.64 kJ/mL.
The conventional solid–liquid extraction (without a microwave pretreatment) was
taken as the control reference to assess the efficiency of the pretreatment. The corresponding
values are also included in Tables 5 and 6.

3.2.1. Initial Extraction Rate

The most relevant effect of the of microwave pretreatment application was the sig-
nificant acceleration of the initial extraction rate, both in polyphenols and anthocyanins.
The acceleration caused by pretreatment ranged from 2 to 12 times the conventional value.
How the S:L ratio and the pretreatment temperature influenced this acceleration can be
observed in Figures 3 and 4, respectively.
Antioxidants
Antioxidants 2021, 10, x FOR
FOR PEER
PEER REVIEW 12 of
of of21
Antioxidants2021,
2021,10,
10,x 1054 REVIEW 12 11 2119

0,MW/u
uu0,MW 0,CON
/u0,CON

Figure
Figure3.
Figure 3.3.Solid–liquid
Solid–liquidratio
Solid–liquid ratioinfluence
ratio influence on
influence on initial
on extraction
initial extraction
initial rate
extractionrate respect
respecttoto
raterespect the
tothe control
thecontrol forfor
control for a
a pretreat-
a pretreatment
pretreat-
ment temperature
temperature of of
80 ◦80
C °C
( (■ polyphenol
polyphenol extraction,
extraction, ∆Δ anthocyanin extraction, ○ and ●
ment temperature of 80 °C (■ polyphenol extraction, Δ anthocyanin extraction, ○ and ● experimental
anthocyanin extraction, # and •experimental
experimental
points
pointsfor
points forpolyphenols
for polyphenolsand
polyphenols andandanthocyanins,
anthocyanins,respectively).
anthocyanins, respectively).
respectively).

14.00
14.00

12.00
12.00

10.00
10.00
0,CON
/u0,CON

8.00
8.00
0,MW/u
uu0,MW

6.00
6.00

4.00
4.00

2.00
2.00

60
60 80
80 100
100 120
120
T
TMWpretreatment (ºC)
MWpretreatment (ºC)
Pretreatment temperature
Figure 4.Pretreatment
Figure influence on initial extraction rate respect to the control for
Figure 4.
4. Pretreatment temperature
temperature influence
influence on
on initial
initial extraction
extraction rate
rate respect
respect to
to the
the control
control for
for aa
a solid–liquid ratio of 0.30 g/mL ( polyphenol extraction, ∆ anthocyanin extraction, # and •
solid–liquid
solid–liquid ratio
ratio of
of 0.30
0.30 g/mL
g/mL (■
(■ polyphenol
polyphenol extraction,
extraction, Δ
Δ anthocyanin extraction, ○
anthocyanin extraction, ○ and
and ●
● experi-
experi-
experimental
mental points points
for for polyphenols
polyphenols and and anthocyanins,
anthocyanins, respectively).
respectively).
mental points for polyphenols and anthocyanins, respectively).
Lower S:Lratios
Lower ratios ledtotofaster
faster extractions.The The samebehavior,
behavior, althoughless less pro-
Lower S:LS:L ratios led
led to faster extractions.
extractions. The same same behavior, although
although less pro-
pro-
nounced, wasalso
nounced, also observedininthethe conventionalsolid–liquid
solid–liquid extraction(Figure
(Figure 5).This
This
nounced, was was also observed
observed in the conventional
conventional solid–liquid extractionextraction (Figure 5).
5). This
effect
effect was explained in terms of the concentration gradients. Lower S:L ratios implied
effect was
was explained
explained in in terms
terms of
of the
the concentration
concentration gradients.
gradients. Lower
Lower S:L
S:L ratios
ratios implied
implied
largersolvent
larger solventvolumes
volumesand andmore
more diluted
diluted extracts,
extracts, leading
leading toto aa higher
higher mass
mass transfer
transfer driving
driv-
larger
force. solvent volumes
Nevertheless, itand
was more
showndiluted
that extracts,
the leading
application to
of a
the higher mass
pretreatment transfer driv-
significantly
ing
ing force.
force. Nevertheless,
Nevertheless, it
it was
was shown
shown that
that the
the application
application of
of the
the pretreatment
pretreatment significantly
significantly
increasedthe
increased the initialextraction
extraction rate(u(u 0 ), demonstrating the enhancementofofmass mass transfer
increased
due to the initial extraction rate
rate (u00),
initial pretreatment.
microwave ), demonstrating
demonstrating the the enhancement
enhancement of mass transfer
transfer
due
due toto microwave
microwave pretreatment.
pretreatment.
Antioxidants 2021, 10, x FOR PEER REVIEW 13 of 21
Antioxidants 2021, 10, 1054 12 of 19

Figure
Figure 5. 5. Coupled
Coupled effect
effect of S:L
of S:L ratioratio
and and pretreatment
pretreatment temperature
temperature onpolyphenols
on the the polyphenols
initialinitial
ex-
extraction rate at 60 ◦ C () and 120 ◦ C ().
traction rate at 60 °C (■) and 120 °C (□).

As for the pretreatment temperature, the initial extraction rate accelerated with tem-
As for the pretreatment temperature, the initial extraction rate accelerated with tem-
perature up to 100 ◦ C. At higher temperatures (120 ◦ C), the polyphenol content showed
perature up to 100 °C. At higher temperatures (120 °C), the polyphenol content showed
no further improvement, while the anthocyanin content appeared to show a decrease,
no further improvement, while the anthocyanin content appeared to show a decrease, de-
despite the experimental error. Considering the anthocyanin richness (results shown in
spite the experimental error. Considering the anthocyanin richness (results shown in Ta-
Table 6), a 17% decrease in anthocyanin richness was observed at 120 ◦ C, which corrobo-
ble 6), a 17% decrease in anthocyanin richness was observed at 120 °C, which corroborated
rated the hypothesis of thermal degradation. Thus, it can be stated that as the pretreatment
the hypothesis of thermal degradation. Thus, it can be stated that as the pretreatment tem-
temperature increased up to 100 ◦ C, the extraction of polyphenols and anthocyanins
perature increased up to 100 °C, the extraction of polyphenols and anthocyanins acceler-
accelerated rapidly, compensating their thermal degradation rate. However, at higher
ated rapidly, compensating
pretreatment temperatures, their
the thermal
influence degradation
of degradation rate. However, at higher
exerted a greater pretreat-on
influence
ment temperatures, the influence of degradation exerted a greater
the balance than the increase in extraction yield. This agrees with [10], who reported influence on the balance
that
than the ◦ increase in extraction yield. This agrees with [10], who
70–90 C was the best temperature range to maintain the antioxidant properties of saffron reported that 70–90 °C
was the best
flowers temperature
during a drying range
compared to maintain
to higher thetemperatures
antioxidant properties
(110 and 125 of ◦saffron
C). On flowers
the other
during
hand, it contrasts with the results of [37] on the stability of grape pomaceother
a drying compared to higher temperatures (110 and 125 °C). On the hand, it
anthocyanins,
contrasts
where 10% withdegradation
the results of was [37] on the after
reported stability
3 min of at
grape
120 ◦pomace
C. In thisanthocyanins, where
work, the extraction
10% degradation was reported after 3 min at 120 °C. In this
media was exposed to that temperature for 2.4 min. In the present work, almost twicework, the extraction media
was
theexposed
thermal to that temperature
degradation for 2.4 for
was observed minutes.
saffronInanthocyanins
the present work, almosttotwice
with respect the
the value
thermal degradation was observed for saffron anthocyanins with
observed by Solyom for grape anthocyanins, which may be due to differences in the raw respect to the value ob-
served
material by matrix.
Solyom for grape anthocyanins, which may be due to differences in the raw
material It ismatrix.
worth considering the coupled effect of the interaction between the S:L ratio and
theItpretreatment
is worth considering
temperaturethe coupled effect of
on the initial the interaction
extraction rate. Nobetween the S:L
difference was ratio andin
found
the pretreatment
the temperature
initial extraction rate at lowon the initial extraction
pretreatment rate. No difference
temperatures (Figure 5),was found in theof
independently
initial extraction rate
the solid–liquid at low
ratio. pretreatment
However, at higher temperatures
temperatures, (Figure 5), independently
a substantial acceleration of the
was
solid–liquid ratio. However, at higher temperatures, a substantial acceleration
observed for low S:L ratios. In this case, both variables contribute to improve the extraction was ob-
served
process: for the
lowhigh
S:L ratios. In this case,
temperature both variables
accelerated the rate,contribute
and the low to S:L
improve the extraction
ratio ensured a large
process:
drivingthe high
force thattemperature
enhanced mass accelerated
transfer.the rate,
This and the low
interaction wasS:L ratio ensured
significant a large
for polyphenol
driving
extractionforcebutthatwas
enhanced
not as mass transfer.
crucial This interaction
for anthocyanins was significant
(a significance for polyphenol
of 10%), although it
extraction
showed the butsame
was tendency.
not as crucial for anthocyanins (a significance of 10%), although it
showed The theapplication
same tendency.of microwave pretreatment significantly improved the initial extrac-
tionTherate, allowing aofconsiderable
application reduction of the
microwave pretreatment extractionimproved
significantly time. the initial extrac-
tion rate, allowing a considerable reduction of the extraction time.
3.2.2. Extraction Yield
Extraction yield (polyphenols extracted from the saffron flowers) was not affected by
the addition of the microwave pretreatment. The same yield than in the conventional solid–
 3.2.2. Extraction Yield
Extraction yield (polyphenols extracted from the saffron flowers) was not affected by
Antioxidants 2021, 10, 1054 the addition of the microwave pretreatment. The same yield than in the conventional 13 of 19

solid–liquid extraction was obtained for both polyphenols (around 26.19 ± 2.26
mgGAE/gDry SF) and anthocyanins (around 6.92 ± 1.30 mgCGE/gDry SF). Thus, it can be
concluded that saffron
liquid extraction flowers were
was obtained completely
for both exhausted
polyphenols by the
(around extraction.
26.19 ± 2.26 mgGAE/gDry
SF) and anthocyanins (around 6.92 ± 1.30 mgCGE/gDry SF). Thus, it can be concluded
3.2.3. Productflowers
that saffron Richness were completely exhausted by the extraction.
Polyphenol and anthocyanin richness (the concentration of active compounds ob-
3.2.3. Product
tained in the dry Richness
product after solvent evaporation of the extract) remained constant
throughout the extraction.
Polyphenol and anthocyanin This meant that the(the
richness extraction of active
concentration ofcompounds and unde-
active compounds ob-
sired
tained compounds
in the dry(sugars,
productfibers, etc.) took
after solvent place always
evaporation of in
thethe same kinetics
extract) remained ratio. Pre-
constant
treatment
throughout resulted in the same
the extraction. Thisacceleration
meant thatin theboth desiredofand
extraction undesired
active compoundscompounds,
and un-
unlike
desired in compounds
the case of grape (sugars,pomace [38].etc.)
fibers, In this
tooklatter case,
place polyphenol
always in the extraction
same kineticswas sub-
ratio.
Pretreatment
stantially resulted
enhanced by in the same acceleration
pretreatment but no other in both desired extraction,
compounds and undesired compounds,
which decayed
unlike in the
polyphenol case of grape
richness. This allowedpomaceone [38].
to In this
find anlatter case,
optimal polyphenol
time at which extraction
polyphenols was sub-
were
stantially enhanced by pretreatment but no other compounds
extracted but other compounds that decay polyphenol richness were not. However, in the extraction, which decayed
polyphenol
case of saffron richness.
flowers,This it was allowed one to find
not possible an an
to find optimal
optimal time at which
time polyphenols
to improve were
polyphenol
extractedTo
richness. but other compounds
facilitate the analysisthat decay
of the polyphenol
influence of therichness were not.
experimental However,
conditions, inav-
the the
case of
erage saffronofflowers,
richness the product it wasthroughout
not possiblethe to extraction
find an optimal time Pretreatment
was used. to improve polyphenol
tempera-
richness.
ture To facilitate
was found the analysis
to be significant of the influence
according of theof
to the analysis experimental conditions,
variance. However, the
an ex-
average richness of the product throughout the extraction was
haustive analysis of this variable cannot be performed due to the large experimental error, used. Pretreatment tem-
perature was
especially found to be significant
for anthocyanins, as it can be according to the 6.
seen in Figure analysis of variance.
Nonetheless, However,
from these an
results,
exhaustive analysis of this variable cannot be performed due
it can be concluded that polyphenol richness improved with higher microwave pretreat- to the large experimental error,
especially
ment for anthocyanins,
temperatures. Considering as it that
can betheseen in Figure 6.
conventional Nonetheless,
extraction fromathese
provides results,
richness of
it canmg
41.82 beGAE
concluded
/gDry extractthat polyphenol
, polyphenol richness
richness wasimproved
especiallywith higher when
enhanced microwave pretreat-
temperatures
ment temperatures.
above the solvent boiling Considering that the conventional
point (combination extraction
of microwaves andprovides
pressure)a richness
were em- of
41.82 mg /g , polyphenol richness was especially
ployed in the pretreatment. A 9.8% and a 24.8% rich improvement stood out at 100 °C and
GAE Dry extract enhanced when temperatures
above
120 °C,the solvent boiling
respectively. On the point (combination
other of microwavesin
hand, no improvement and pressure) were
anthocyanin employed
richness was
in the pretreatment. A 9.8% and a 24.8% rich improvement stood out at 100 ◦ C and 120 ◦ C,
observed due to microwave pretreatment but a decrease of 8.8% caused by thermal deg-
respectively.
radation due to On the othertohand,
exposure highernotemperatures.
improvement in anthocyanin richness was observed
due to microwave pretreatment but a decrease of 8.8% caused by thermal degradation due
to exposure to higher temperatures.

Figure6.6.Polyphenol
Figure () and
Polyphenol (■) and anthocyanin
anthocyanin (∆)
(∆)richness
richnessininthe extraction
the with
extraction microwave
with pretreatment
microwave pretreat-
at 0.30
ment at g/mL S:L ratio
0.30 g/mL (# polyphenol
S:L ratio (○ polyphenol • anthocyanin
andand ● anthocyaninexperimental points).
experimental points).

3.3. End Product Characterization

In addition to the richness of the product, there are other characteristics that determine
the quality of the end product.
Antioxidants 2021, 10, 1054 14 of 19

3.3.1. Product Composition

As for the identification of individual polyphenols, the presence of gallic acid, catechin,
epicatechin, quercetin, kaempferol and maldivin has been reported in similar extracts [44,45],
but they have not been detected in this study. On the other hand, delphinidin was found in
a very large concentration. It represented 80% of the total anthocyanins detected by HPLC
at 520 nm wavelength [11] and 15% of the total polyphenols. Delphinidin concentrations
obtained in microwave pretreatment varied from 5.03 to 7.89 mg/gDry Extract . These con-
centrations represent an average improvement of 28% over the conventional extract. The
extraction conditions used to maximize delphinidin richness agreed with the previous de-
scribed tendencies to optimize anthocyanin extraction. In essence, it can be said that, although
the richness of the extract can be considered low (an average of 5% for polyphenols and 1%
for anthocyanins), the fact that a high-value compound such as delphinidin was present in
such high concentrations greatly enhances the quality of the final product. The addition of a
purification step would make it possible not only to obtain an extract rich in polyphenols but
also a product specifically rich in delphinidin.
The characterization of the extract was concluded with an analysis of its antioxidant
capacity. Most of the antioxidant activities of saffron flower extracts reported in other
works are expressed as the extract ability to inhibit DPPH [21,30]. However, this kind
of radical is scarce in biological systems [46]; therefore, the capacity to quench alkoxyl
radicals was assessed instead. Again, improvements could be observed due to the addition
of microwave pretreatment. An average improvement of 36% was found, although the
antioxidant activity increased to more than 100% over conventional extraction without
pretreatment when mild pretreatment conditions (low pretreatment temperature and low
solid–liquid ratio) were used.

3.3.2. Color
Because one of the likely applications of this extract is as food colorant, variations
in color have been analyzed. The three CIELAB parameters considered for this purpose
(L*, a* and b*) are gathered in Table 5. It must be noticed that color was measured in
a solution prepared with the dried extract; therefore, these values can only be used to
identify differences, not to characterize final extract color. However, since the dry extract
is expected to be the preferable form of the commercial product, this may be considered
the analysis of interest. From the results, it can be concluded that the application of
microwave pretreatment only slightly influenced the color of the extract. An average total
color difference of 5.96 was found when microwaves were employed compared to the
conventional solid–liquid extraction. The main difference was found in chroma, while hue
(L*) was almost constant. However, it is worth noting that each of the CIELAB parameters,
L*, a* and b*, showed a relationship with anthocyanin content, as already observed by
other authors [47]. Luminosity decreased with the concentration of anthocyanins (darker
extracts were obtained), and red hue was improved (a* tended to higher values), as
well as blue (b* tended to negative). According to the optimal anthocyanin extraction
conditions, experiments performed with a low S:L ratio and low pretreatment temperature
(0.30 g/mL and 60 ◦ C, which entailed an absorbed energy density of 0.23 kJ/mL) gave a
high anthocyanin richness and, consequently, the highest color increase in terms of red
and blue. An intense violet extract was obtained. If more pronounced conditions were
employed (80–100 ◦ C and 0.30–0.50 g/mL, with absorbed energy density between 0.39
and 0.58 kJ/mL), a blue hue similar to that of the conventional extraction was found, but
a lower a* parameter was obtained, indicating a tendency to green. These results are
shown in Figure 7. The most severe conditions (120 ◦ C and 0.50–0.70 g/mL, with absorbed
energy densities between 0.58 and 0.64 kJ/mL) caused a deviation from the desired color.
A brighter and more yellow-green colored extract was obtained compared to the one
obtained in the conventional extraction. A tendency to yellow in saffron flower extracts
was also observed by other authors when the extract was encapsulated [4] or when saffron
flowers were dried [10]. The latter attributed the color change to thermal degradation of
Antioxidants 2021, 10, x FOR PEER REVIEW 16 of 21

severe conditions (120 °C and 0.50–0.70 g/mL, with absorbed energy densities between
Antioxidants 2021, 10, 1054 0.58 and 0.64 kJ/mL) caused a deviation from the desired color. A brighter and more 15 yel-
of 19
low-green colored extract was obtained compared to the one obtained in the conventional
extraction. A tendency to yellow in saffron flower extracts was also observed by other
authors when the extract was encapsulated [4] or when saffron flowers were dried [10].
Theanthocyanins, which
latter attributed theseems
colortochange
be a reasonable
to thermaljustification.
degradationTherefore, if a bluerwhich
of anthocyanins, extract
is sought, a mild pretreatment (according to optimal anthocyanin extraction conditions)
seems to be a reasonable justification. Therefore, if a bluer extract is sought, a mild pre-
should be
treatment used. to optimal anthocyanin extraction conditions) should be used.
(according

Figure
Figure 7. CIELAB
7. CIELAB parametersa*a*and
parameters andb*
b* as a function
functionofofthe
thepretreatment temperature:
pretreatment × conventional
temperature: × conven-
solid–liquid
tional extraction,
solid–liquid  TMW
extraction, = 60=◦60
□ TMW # TMW
C, °C, ○ TMW= 80= ◦80
C, °C, ▲ TMW
N TMW = 100= ◦100
C, and  TMW
°C, and = 120 =◦ C.
♦ TMW
120 °C.
3.4. Extraction Optimization
Optimization
3.4. Extraction was performed using response surface methodology with the statistical
Optimization
analysis software DesignExpert
Optimization was performed using v. 11 response
(StatEase,surface
Minneapolis, MN, USA).
methodology The
with the operating
statistical
variables S:L ratio and pretreatment temperature were optimized to maximize
analysis software DesignExpert v. 11 (StatEase, Minneapolis, MN, USA). The operating extraction
kineticsS:L
variables and to improve
ratio the characteristics
and pretreatment temperature of the final
were product to
optimized (richness andextraction
maximize color). The
optimization method allows for a multiple goal target. To do this, the different variables
kinetics and to improve the characteristics of the final product (richness and color). The
of interest to be optimized must be ranked according to the specific ‘Objective’ (max-
optimization method allows for a multiple goal target. To do this, the different variables
imize/minimize/none) and the weight of each variable in the final objective function
of interest to be optimized must be ranked according to the specific ‘Objective’ (maxim-
(desirability function) must be quantified in terms of ‘Relative importance’ (0–5). Greater
ize/minimize/none) and the weight of each variable in the final objective function (desir-
relative importance (value = 5) was attributed to richness and color (blue and red), since
ability function) must be quantified in terms of ‘Relative importance’ (0–5). Greater rela-
these factors are closely related to the final application of the product. Kinetic improve-
tive importance (value = 5) was attributed to richness and color (blue and red), since these
ments (extraction rate rather than yield) were considered to be of medium importance
factors are closely related to the final application of the product. Kinetic improvements
(value = 4), and a low-medium importance (value = 3) was chosen for yields and antioxidant
(extraction rate rather than yield) were considered to be of medium importance (value =
activities. Results are presented in Table 7 in the ‘Theoretical optimum’ column:
4), and a low-medium importance (value = 3) was chosen for yields and antioxidant activ-
ities. Results are presented in Table 7 inresults.
Table 7. Optimization the ‘Theoretical optimum’ column:

Relative Theoretical Experimental

Variable Objective Range
Importance Optimum Optimum
S:L ratio (g/mL) 0.30–0.70 0.30 0.30 ± 0.00
Pretreatment temperature
60–120 65 66 ± 0
TMW (◦ C)
Absorbed Energy Density
0.16–0.54 0.24 0.24 ± 0.01
(kJ/mL)
Polyphenol yield
maximize 21.05–29.53 3 29.52 31.33 ± 3.50
(mgGAE /gDry SF )
Antioxidants 2021, 10, 1054 16 of 19

Table 7. Cont.

Relative Theoretical Experimental

Variable Objective Range
Importance Optimum Optimum
Polyphenol initial extraction
maximize 0.98–10.98 4 10.98 15.24 ± 6.70
rate (mgGAE /gDry SF /min)
Anthocyanin yield
maximize 4.55–9.34 3 6.92 10.19 ± 1.26
(mgCGE /gDry SF )
Anthocyanin initial extraction
maximize 0.15–2.51 4 2.38 3.25 ± 1.43
rate (mgCGE /gDry SF /min)
Polyphenol richness
maximize 41.52–53.34 5 47.18 54.25 ± 3.06
(mgGAE /gDry Extract )
Anthocyanin richness
maximize 8.28–14.18 5 11.14 12.32 ± 0.87
(mgCGE /gDry Extract )
Antioxidant activity
maximize 1.26–2.53 3 1.96 2.40 ± 0.16
(mmolTE /gDry Extract )
L* minimize 82.07–88.94 5 87.38 90.56 ± 0.93
a* maximize 22.46–39.24 5 29.63 27.60 ± 1.49
b* minimize −14.82–(−4.46) 5 −11.18 −16.22 ± 0.45
∆E none 1.47–12.85 0 - 8.15

4. Conclusions
The use of a microwave pretreatment, consisting of a short but intense heating step
followed by rapid cooling before conventional solid–liquid extraction, has been shown to
accelerate the extraction of polyphenols while minimizing their degradation. The variables
analyzed were the maximum temperature reached and the solid–liquid ratio used, both
referring to pretreatment. The effect of pretreatment was compared with a conventional
extraction (without microwave pretreatment), the conditions of which were also optimized
in this work. In the optimization study of the conventional solid–liquid extraction, the ratio
of saffron flowers to solvent (S:L ratio) proved to be of vital importance. It determined
whether a concentrated liquid product was obtained (high S:L ratios) or whether the
polyphenol content in the saffron flowers was completely depleted (low ratios). The acidity
of the solvent used was also considered significant, as it could have contributed to the
hydrolysis of the petal structure, thus facilitating the release of the active compounds.
Finally, a ratio of 0.30 g/mL was selected for the conventional solid–liquid extraction, as it
implied a balance between solvent consumption, polyphenol extraction and liquid-extract
concentration. It also ensured maximum yield of anthocyanins, which were the compounds
of greatest interest. This extraction medium, at a temperature of 30 ◦ C, was also used for
all extractions after microwave pretreatment.
The main advantages obtained with the application of microwave pretreatment were
the reduction of the extraction time and the improvement of the quality of the final product.
The initial extraction rate was the most affected parameter. It was greatly accelerated when
moderately high temperatures and low solid–liquid ratios were used. This improvement
made it possible to obtain a high yield in 5 min, instead of almost an hour, as required
by the conventional process. Moreover, there was an increase in product quality in terms
of richness, composition, antioxidant activity and color. It was found that the richness in
polyphenols and anthocyanins (proportion of active compounds in the final dry product)
showed opposite trends. The former increased with higher pretreatment temperatures, up
to 25% at 100 ◦ C, while anthocyanin richness decreased by 9% at the same temperature
increase, due to thermal degradation. Thus, optimal anthocyanin extraction conditions
were established at a low pretreatment temperature and at low S:L ratio. These conditions
also maximized the antioxidant activity of the extract (130%) and the color of the final
Antioxidants 2021, 10, 1054 17 of 19

product. With respect to the latter, mild conditions enhanced the red and blue tones, while
more severe conditions caused the extract to move away from the desired tannic hue to
a greener, yellowish product, probably due to thermal degradation. Finally, as for the
composition of the extract, delphinidin was the main compound detected. It represented
80% of total anthocyanins and 15% of total polyphenols. Such specificity raised the value
of the extract, as it can be easily converted into a delphinidin extract by the addition of a
purification step.
In summary, it was concluded that a gentle microwave pretreatment (with a ratio of
0.30 g/mL and irradiating the medium up to a temperature of 65 ◦ C) was very convenient
as it allowed for the obtaining of a high-quality product in an efficient extraction process.

Author Contributions: Conceptualization, A.Á. and R.B.M.; methodology, A.Á., S.T. and R.B.M.;
software, A.Á. and S.T.; validation, A.Á., S.T. and R.B.M.; formal analysis, A.Á., S.T. and R.B.M.;
investigation, A.Á. and S.T.; resources, A.Á., S.T., M.J.C. and R.B.M.; writing—original draft prepara-
tion, A.Á., S.T. and R.B.M.; writing—review and editing, A.Á., S.T., M.J.C. and R.B.M.; supervision,
M.J.C. and R.B.M.; project administration, M.J.C. and R.B.M.; funding acquisition, M.J.C. All authors
have read and agreed to the published version of the manuscript.
Funding: This research was funded by Spanish regional government Junta de Castilla y León
(VA040U16).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Acknowledgments: A. Sahagun is kindly thanked for providing the saffron flowers, F. Ronda for the
color assay guidance, and R. Páramo for the measurements of the specific heat of saffron flowers. A.
Álvarez is grateful for her Ph.D. grant FPU13/04678.
Conflicts of Interest: The authors declare no conflict of interest.

Appendix A
Wavelengths for color determination:
Wavelength (nm)
# X Y Z # X Y Z
1 424.4 465.9 414.1 16 585 558.5 454
2 435.5 489.5 422.2 17 588.7 561.9 455.9
3 443.9 500.4 426.3 18 592.4 565.3 457.9
4 452.1 508.7 429.4 19 596 568.9 459.9
5 431.2 515.2 432 20 599.6 572.5 462
6 474 520.6 434.3 21 603.3 576.4 646.1
7 531.2 525.4 436.5 22 607 580.4 466.3
8 544.3 529.8 438.6 23 610.9 584.8 468.7
9 552.4 533.9 440.6 24 615 589.6 471.4
10 558.7 537.7 442.5 25 619.4 594.8 474.3
11 564.1 541.4 444.4 26 624.2 600.8 477.7
12 568.9 544.9 446.3 27 629.8 607.7 481.8
13 573.2 548.4 448.2 28 636.6 616.1 487.2
14 577.4 551.8 450.1 29 645.9 627.3 495.2
15 581.3 555.1 452.1 30 663 647.4 511.2

References
1. Shukla, S.; Mehta, A.; John, J.; Singh, S.; Mehta, P.; Vyas, S.P. Antioxidant activity and total phenolic content of ethanolic extract of
Caesalpinia bonducella seeds. Food Chem. Toxicol. 2009, 47, 1848–1851. [CrossRef]
2. Castro-López, C.; Ventura-Sobrevilla, J.M.; González-Hernández, M.D.; Rojas, R.; Ascacio-Valdés, J.A.; Aguilar, C.N.; Martínez-
Ávila, G.C.G. Impact of extraction techniques on antioxidant capacities and phytochemical composition of polyphenol-rich
extracts. Food Chem. 2017, 237, 1139–1148. [CrossRef] [PubMed]
Antioxidants 2021, 10, 1054 18 of 19

3. Goleroudbary, M.G.; Ghoreishi, S.M. Response surface optimization of Safranal and Crocin extraction from Crocus sativus L. via
supercritical fluid technology. J. Supercrit. Fluids 2016, 108, 136–144. [CrossRef]
4. Ahmad, M.; Ashraf, B.; Gani, A.; Gani, A. Microencapsulation of saffron anthocyanins using β glucan and β cyclodextrin:
Microcapsule characterization, release behaviour & antioxidant potential during in-vitro digestion. Int. J. Biol. Macromol. 2018,
109, 435–442. [CrossRef] [PubMed]
5. Putnik, P.; Barba, F.J.; Španić, I.; Zorić, Z.; Dragović-Uzelac, V.; Bursać Kovačević, D. Green extraction approach for the recovery
of polyphenols from Croatian olive leaves (Olea europea). Food Bioprod. Process. 2017, 106, 19–28. [CrossRef]
6. Mahdavi, S.A.; Jafari, S.; Ghorbani, M.; Assadpoor, E. Spray-Drying Microencapsulation of Anthocyanins by Natural Biopolymers:
A Review. Dry. Technol. 2014, 32, 509–518. [CrossRef]
7. Araújo, P.; Basílio, N.; Azevedo, J.; Fernandes, A.; Mateus, N.; Pina, F.; de Freitas, V.; Oliveira, J. Colour modulation of blue
anthocyanin-derivatives. Lignosulfonates as a tool to improve the water solubility of natural blue dyes. Dye. Pigment. 2018, 153,
150–159. [CrossRef]
8. Hosseini, A.; Razavi, B.M.; Hosseinzadeh, H. Saffron (Crocus sativus) petal as a new pharmacological target: A review. Iran. J.
Basic Med. Sci. 2018, 21, 1091–1099. [CrossRef] [PubMed]
9. Serrano-Díaz, J.; Sánchez, A.M.; Maggi, L.; Martínez-Tomé, M.; García-Diz, L.; Murcia, M.A.; Alonso, G.L. Increasing the
applications of Crocus sativus flowers as natural antioxidants. J. Food Sci. 2012, 77, C1162–C1168. [CrossRef]
10. Serrano-Díaz, J.; Sánchez, A.M.; Alvarruiz, A.; Alonso, G.L. Preservation of saffron floral bio-residues by hot air convection. Food
Chem. 2013, 141, 1536–1543. [CrossRef]
11. Ongkowijoyo, P.; Luna-Vital, D.A.; Gonzalez de Mejia, E. Extraction techniques and analysis of anthocyanins from food sources
by mass spectrometry: An update. Food Chem. 2018, 250, 113–126. [CrossRef] [PubMed]
12. Xiong, S.; Melton, L.D.; Easteal, A.J.; Siew, D. Stability and antioxidant activity of black currant anthocyanins in solution and
encapsulated in glucan gel. J. Agric. Food Chem. 2006, 54, 6201–6208. [CrossRef]
13. Mateus, N.; Silva, A.M.S.; Rivas-Gonzalo, J.C.; Santos-Buelga, C.; de Freitas, V. A new class of blue anthocyanin-derived pigments
isolated from red wines. J. Agric. Food Chem. 2003, 51, 1919–1923. [CrossRef]
14. Martins, N.; Roriz, C.L.; Morales, P.; Barros, L.; Ferreira, I.C.F.R. Food colorants: Challenges, opportunities and current desires of
agro-industries to ensure consumer expectations and regulatory practices. Trends Food Sci. Technol. 2016, 52, 1–15. [CrossRef]
15. Nørbæk, R.; Brandt, K.; Nielsen, J.K.; Ørgaard, M.; Jacobsen, N. Flower pigment composition of Crocus species and cultivars
used for a chemotaxonomic investigation. Biochem. Syst. Ecol. 2002, 30, 763–791. [CrossRef]
16. Goupy, P.; Vian, M.A.; Chemat, F.; Caris-Veyrat, C. Identification and quantification of flavonols, anthocyanins and lutein diesters
in tepals of Crocus sativus by ultra performance liquid chromatography coupled to diode array and ion trap mass spectrometry
detections. Ind. Crop. Prod. 2013, 44, 496–510. [CrossRef]
17. Serrano-Díaz, J.; Sánchez, A.M.; Martínez-Tomé, M.; Winterhalter, P.; Alonso, G.L. Flavonoid Determination in the Quality
Control of Floral Bioresidues from Crocus sativus L. J. Agric. Food Chem. 2014, 62, 3125–3133. [CrossRef]
18. Sant’Anna, V.; Gurak, P.D.; Ferreira Marczak, L.D.; Tessaro, I.C. Tracking bioactive compounds with colour changes in foods—A
review. Dye. Pigment. 2013, 98, 601–608. [CrossRef]
19. Zheng, C.-J.; Li, L.; Ma, W.-H.; Han, T.; Qin, L.-P. Chemical constituents and bioactivities of the liposoluble fraction from different
medicinal parts of Crocus sativus. Pharm. Biol. 2011, 49, 756–763. [CrossRef]
20. Sánchez-Vioque, R.; Santana-Méridas, O.; Polissiou, M.; Vioque, J.; Astraka, K.; Alaiz, M.; Herraiz-Peñalver, D.; Tarantilis, P.A.;
Girón-Calle, J. Polyphenol composition and in vitro antiproliferative effect of corm, tepal and leaf from Crocus sativus L. on
human colon adenocarcinoma cells (Caco-2). J. Funct. Foods 2016, 24, 18–25. [CrossRef]
21. Tuberoso, C.I.G.; Rosa, A.; Montoro, P.; Fenu, M.A.; Pizza, C. Antioxidant activity, cytotoxic activity and metabolic profiling of
juices obtained from saffron (Crocus sativus L.) floral by-products. Food Chem. 2016, 199, 18–27. [CrossRef]
22. Favot, L.; Martin, S.; Keravis, T.; Andriantsitohaina, R.; Lugnier, C. Involvement of cyclin-dependent pathway in the inhibitory
effect of delphinidin on angiogenesis. Cardiovasc. Res. 2003, 59, 479–487. [CrossRef]
23. Meiers, S.; Kemény, M.; Weyand, U.; Gastpar, R.; von Angerer, E.; Marko, D. The anthocyanidins cyanidin and delphinidin are
potent inhibitors of the epidermal growth-factor receptor. J. Agric. Food Chem. 2001, 49, 958–962. [CrossRef]
24. Anastas, P.; Eghbali, N. Green Chemistry: Principles and Practice. Chem. Soc. Rev. 2010, 39, 301–312. [CrossRef] [PubMed]
25. Ahmadian-Kouchaksaraie, Z.; Niazmand, R.; Najafi, M.N. Optimization of the subcritical water extraction of phenolic antioxidants
from Crocus sativus petals of saffron industry residues: Box-Behnken design and principal component analysis. Innov. Food Sci.
Emerg. Technol. 2016, 36, 234–244. [CrossRef]
26. Ahmadian-Kouchaksaraie, Z.; Niazmand, R. Supercritical carbon dioxide extraction of antioxidants from Crocus sativus petals of
saffron industry residues: Optimization using response surface methodology. J. Supercrit. Fluids 2017, 121, 19–31. [CrossRef]
27. Lotfi, L.; Kalbasi-Ashtari, A.; Hamedi, M.; Ghorbani, F. Effects of enzymatic extraction on anthocyanins yield of saffron tepals
(Crocos sativus) along with its color properties and structural stability. J. Food Drug Anal. 2015, 23, 210–218. [CrossRef]
28. Da Porto, C.; Natolino, A. Extraction kinetic modelling of total polyphenols and total anthocyanins from saffron floral bio-residues:
Comparison of extraction methods. Food Chem. 2018, 258, 137–143. [CrossRef] [PubMed]
29. Sarfarazi, M.; Jafari, S.M.; Rajabzadeh, G.; Galanakis, C.M. Evaluation of microwave-assisted extraction technology for separation
of bioactive components of saffron (Crocus sativus L.). Ind. Crop. Prod. 2020, 145, 111978. [CrossRef]
Antioxidants 2021, 10, 1054 19 of 19

30. Ahmadian-Kouchaksaraie, Z.; Niazmand, R.; Maskooki, A. Optimization of extraction conditions of bioactive components from
saffron petal using response surface method (RSM). Res. Innov. Food Sci. Technol. 2016, 5, 39–54.
31. Jafari, S.-M.; Mahdavi-Khazaei, K.; Hemmati-Kakhki, A. Microencapsulation of saffron petal anthocyanins with cress seed gum
compared with Arabic gum through freeze drying. Carbohydr. Polym. 2016, 140, 20–25. [CrossRef] [PubMed]
32. Khazaei, K.M.; Jafari, S.M.; Ghorbani, M.; Kakhki, A.H.; Sarfarazi, M. Optimization of Anthocyanin Extraction from Saffron
Petals with Response Surface Methodology. Food Anal. Methods 2016, 9, 1993–2001. [CrossRef]
33. Mateus, N.; Freitas, V. Anthocyanins as Food Colorants. In Anthocyanins; Springer: New York, NY, USA, 2008; pp. 284–304.
34. Sólyom, K.; Mato, R.B.; Pérez-Elvira, S.I.; Cocero, M.J. Influence of absorbed energy of microwave treatment on biogas production
from wastewater sludge. In Proceedings of the Microwave and RF Power Applications: Proceeding of the 13th International
Conference on Microwave and Radio Frequency Heating, AMPERE 2011, Toulouse, France, 5–8 September 2011; Ampere Europe,
2011.
35. Galan, A.-M.; Calinescu, I.; Trifan, A.; Winkworth-Smith, C.; Calvo-Carrascal, M.; Dodds, C.; Binner, E. New insights into the role
of selective and volumetric heating during microwave extraction: Investigation of the extraction of polyphenolic compounds
from sea buckthorn leaves using microwave-assisted extraction and conventional solvent extraction. Chem. Eng. Process. Process
Intensif. 2017, 116, 29–39. [CrossRef]
36. Yedhu Krishnan, R.; Rajan, K.S. Microwave assisted extraction of flavonoids from Terminalia bellerica: Study of kinetics and
thermodynamics. Sep. Purif. Technol. 2016, 157, 169–178. [CrossRef]
37. Sólyom, K.; Solá, R.; Cocero, M.J.; Mato, R.B. Thermal degradation of grape marc polyphenols. Food Chem. 2014, 159, 361–366.
[CrossRef] [PubMed]
38. Álvarez, A.; Poejo, J.; Matias, A.A.; Duarte, C.M.M.; Cocero, M.J.; Mato, R.B. Microwave pretreatment to improve extraction
efficiency and polyphenol extract richness from grape pomace. Effect on antioxidant bioactivity. Food Bioprod. Process. 2017, 106,
162–170. [CrossRef]
39. Kjeldahl, J. Neue Methode zur Bestimmung des Stickstoffs in organischen Körpern. Z. Für Anal. Chem. 1883, 22, 366–382.
[CrossRef]
40. Serrano-Díaz, J.; Sánchez, A.M.; Martínez-Tomé, M.; Winterhalter, P.; Alonso, G.L. A contribution to nutritional studies on Crocus
sativus flowers and their value as food. J. Food Compos. Anal. 2013, 31, 101–108. [CrossRef]
41. Páramo, R.; Zouine, M.; Casanova, C. New Batch Cells Adapted To Measure Saturated Heat Capacities of Liquids. J. Chem. Eng.
Data 2002, 47, 441–448. [CrossRef]
42. Gómez-Alonso, S.; García-Romero, E.; Hermosín-Gutiérrez, I. HPLC analysis of diverse grape and wine phenolics using direct
injection and multidetection by DAD and fluorescence. J. Food Compos. Anal. 2007, 20, 618–626. [CrossRef]
43. Eaton, A.D.; Clesceri, L.S.; Greenberg, A.E.; American Water Works Association; Water Environment Federation. Standard Methods
for the Examination of Water and Wastewater, 19th ed.; American Public Health Association: Washington, DC, USA, 1995.
44. Menghini, L.; Leporini, L.; Vecchiotti, G.; Locatelli, M.; Carradori, S.; Ferrante, C.; Zengin, G.; Recinella, L.; Chiavaroli, A.; Leone,
S.; et al. Crocus sativus L. stigmas and byproducts: Qualitative fingerprint, antioxidant potentials and enzyme inhibitory activities.
Food Res. Int. 2018, 109, 91–98. [CrossRef] [PubMed]
45. Serrano-Díaz, J.; Estevan, C.; Sogorb, M.Á.; Carmona, M.; Alonso, G.L.; Vilanova, E. Cytotoxic effect against 3T3 fibroblasts cells
of saffron floral bio-residues extracts. Food Chem. 2014, 147, 55–59. [CrossRef]
46. Kameya, H.; Watanabe, J.; Takano-Ishikawa, Y.; Todoriki, S. Comparison of scavenging capacities of vegetables by ORAC and
EPR. Food Chem. 2014, 145, 866–873. [CrossRef] [PubMed]
47. Jiménez-Aguilar, D.M.; Ortega-Regules, A.E.; Lozada-Ramírez, J.D.; Pérez-Pérez, M.C.I.; Vernon-Carter, E.J.; Welti-Chanes, J.
Color and chemical stability of spray-dried blueberry extract using mesquite gum as wall material. J. Food Compos. Anal. 2011, 24,
889–894. [CrossRef]