Waste and Biomass Valorization (2019) 10:2851–2861

https://doi.org/10.1007/s12649-018-0317-7

ORIGINAL PAPER

Optimization of Lycopene Extraction from Tomato Processing Waste

Using an Eco-Friendly Ethyl Lactate–Ethyl Acetate Solvent: A Green
Valorization Approach
Yasmini P. A. Silva1   · Tânia A. P. C. Ferreira1 · Giovana B. Celli2 · Marianne S. Brooks3

Received: 15 January 2018 / Accepted: 2 May 2018 / Published online: 10 May 2018
© Springer Science+Business Media B.V., part of Springer Nature 2018

Abstract
Lycopene is a highly-prized antioxidant with associated health benefits and is abundant in natural sources. A green valoriza-
tion approach was used to extract lycopene from tomato processing waste. Ultrasound-assisted extraction was applied to the
tomato waste using an eco-friendly solvent mixture containing ethyl lactate and ethyl acetate for the extraction of lycopene.
Extraction parameters were: X1 = extraction temperature (°C), X2 = proportion of ethyl acetate in solvent mixture (% v/v),
X3 = solvent:sample ratio (mL/g), and X4 = extraction time (min). A Box–Behnken design was used to define experimental
conditions, and response surface methodology was then conducted to determine the optimized conditions: X1 = 63.4 °C,
X2 = 30% (v/v), X3 = 100 mL/g, and X4 = 20 min. The experimental optimized extraction yield of lycopene was 1334.8 µg/g
(d.w.), in agreement with the predicted yield. At the same conditions without ultrasound, a yield of 1209.5 µg/g (d.w.) was
obtained (9.4% lower). Ultrasound increases extraction yield, and tomato processing by-products are a viable alternative
source of extractable lycopene. This represents a greener strategy for the extraction of lycopene in comparison to conventional
methods using organic solvents, and shows a promising alternative use for a food processing waste.

Keywords  By-product · Green extraction · Carotenoids · Box–Behnken design · Response surface methodology · Biomass
utilization · Extraction/separation

Introduction

Electronic supplementary material  The online version of this The use of agricultural by-products as source material for the
article (https​://doi.org/10.1007/s1264​9-018-0317-7) contains extraction of high value-added compounds has been heavily
supplementary material, which is available to authorized users.
investigated in recent years [1, 2]. Due to their organic com-
* Marianne S. Brooks position, these wastes represent alternative raw materials
su‑ling.brooks@dal.ca that could be exploited for the recovery of bioactive com-
Yasmini P. A. Silva pounds that have positive effects on human health and also
yasminiportes@gmail.com widespread technological applications, as coloring agents
Tânia A. P. C. Ferreira and antioxidants. Lycopene is an example of a high-value
taniaferreira@ufg.br biomolecule that could be potentially recovered from food
Giovana B. Celli processing wastes, as a valorization approach to minimize
gb477@cornell.edu the environmental impact of waste disposal and increase the
1
sustainability of the overall process. Lycopene is a carote-
Faculty of Nutrition, Federal University of Goiás, Rua 227, noid and a valuable dietary antioxidant, with activity against
qd. 68, Setor Leste Universitário, Goiânia, GO 74605‑080,
Brazil diseases such as cancer and cardiovascular diseases [3–5]. It
2 is widely used by various industries, as a dietary supplement
Department of Food Science, Cornell University, Ithaca,
NY 14853, USA for human nutrition, and as a natural colorant and antioxi-
3 dant in products with high lipid content [6].
Department of Process Engineering and Applied Science,
Dalhousie University, PO Box 15000, Halifax, NS B3H 4R2, Lycopene is usually extracted from natural sources,
Canada especially red-coloured vegetables such as tomatoes [6],

13
Vol.:(0123456789)

2852 Waste and Biomass Valorization (2019) 10:2851–2861

red pepper [7] and watermelon [8]. Synthetic routes for be commercially obtained from lactic acid and ethanol, both
lycopene synthesis have been developed [9], but given the of which can be obtained from fermentation processes [23].
abundance of the compound in natural materials and the Recently, EL has shown good potential for the extraction of
market preference for natural products, the extraction of the carotenoids from tomato and other plant materials [24–27].
carotenoid from these sources is preferred. The major source Therefore, both EA and EL represent promising green alter-
of lycopene is tomato, due to its high lycopene content [3] natives to hexane for carotenoid extraction.
and low cost compared to other lycopene-rich fruits and veg- The use of assisting technologies, such as the application
etables [6]. Many companies extract lycopene from tomatoes of ultrasound, to improve the efficiency of solvent extrac-
worldwide, and a high-lycopene tomato variety, containing tion from biological materials is another strategy for reduc-
around 150–250 mg/kg (fresh weight, f.w.), has been devel- ing the environmental impact of extraction, and has been
oped specifically to be used as source for lycopene extraction widely studied in recent years [28, 29]. The increase of
[10]. Commercial extraction of lycopene involves crushing extraction yield observed in ultrasound-assisted processes
the raw tomatoes to separate fibrous materials (such as skin has been associated with the acoustic cavitation phenom-
and seeds) from the juice, then extracting the compound enon, in which ultrasound waves cause a violent collapse of
from the juice using an organic solvent. However, most of gas bubbles present in the solvent [30]. Ultrasonication has
the lycopene in tomatoes (72–92%) is found in the skin frac- been applied as assisting technology for the extraction of
tion [11]. Therefore, tomato pomace, the by-product from several bioactive compounds from food processing wastes,
industrial processing, can be considered a potential source such as phenolic compounds from sunflower seed cake [31]
of lycopene as it is mostly composed of skin and seeds. and from apple peels [32], flavonoids from citrus peels [33],
The extraction of lycopene from vegetables requires the among many others. Also, it has been previously studied for
use of organic solvents, such as hexane, due to the hydro- the extraction of lycopene from tomato [21, 34] and tomato
phobic nature of the compound. Hexane has advantages processing by-products [35–37]. However, to the best of our
of low cost, good capacity for solubilizing lipophilic com- knowledge, no study has yet been published evaluating the
pounds, and low boiling point [12], and is considered as a ultrasound-assisted extraction of lycopene from tomato pom-
food grade solvent by the FDA in residual concentrations of ace using EL, nor its use combined with EA.
up to 25 ppm [13]. However, this solvent has a high envi- In this study, a three-fold strategy was applied to develop
ronmental impact as it is obtained from petroleum [14], thus a greener alternative for lycopene extraction using (1) food
requiring safe handling and storage procedures. Alternative by-products (i.e. tomato processing waste) as source mate-
strategies can be developed using the principles of green rial, (2) green solvents to replace conventional organic
chemistry to overcome the disadvantages associated with the solvents, and (3) ultrasound to improve extraction yield.
use of organic solvents for extraction, where green chemistry Although EL has shown good potential for the extraction
promotes the reduction or elimination of hazardous materi- of carotenoids from plant materials, this is the first report
als and processes. Green solvents are alternatives to conven- that combines these three approaches for extraction of
tional organic solvents due to their low toxicity, high biodeg- lycopene from tomato-based material. Thus, the objec-
radability, and sustainable production from non-petroleum tive of this study was to establish optimal conditions for
sources. One example is ethyl acetate (ethyl ethanoate) (EA), the ultrasound-assisted extraction (UAE) of lycopene from
an environmentally friendly organic solvent commonly used tomato pomace with a green solvent mixture of ethyl lactate
for extraction of lipophilic compounds [15]. Although over- and ethyl acetate. Extraction parameters investigated were
exposure to EA may cause health hazards such as skin irri- temperature, proportion of ethyl acetate in solvent mixture,
tation and unconsciousness, it is considered a low-toxicity solvent:sample ratio and time. A Box–Behnken experi-
solvent due to its rapid hydrolysis to ethanol and acetic acid mental design and response surface methodology (RSM)
during metabolism [16]. Its presence in food products is were used to optimize extraction conditions, which were
accepted in concentrations of up to 25 mg/kg, and it is even experimentally verified, and compared to extraction with-
used as a flavouring agent due to its fruity aroma [17]. EA out ultrasound.
is considered a green alternative to hexane because it can be
produced from bio-based ethanol, and is biodegradable in air
and water [18]. The solvent has been used for extraction of Materials and Methods
lycopene from tomato, resulting in yields that are compara-
ble to hexane extraction [19–21]. Another promising green Chemicals
solvent is ethyl lactate (ethyl 2-hydroxypropanoate) (EL), a
non-corrosive, non-toxic (non-carcinogenic and non-terato- Ethyl lactate (CAS 687-47-8) and ethyl acetate (CAS
genic), non-ozone depleting, biodegradable, considered as 141-78-6) were purchased from Sigma-Aldrich (Oakville,
generally recognized as safe (GRAS) solvent [22]. EL can ON, Canada), and were of analytical grade. The lycopene

13
Waste and Biomass Valorization (2019) 10:2851–2861 2853

standard (CAS 502-65-8) was purchased from Fisher Sci- Table 1  Input variables used in Box–Behnken design (BBD), their
entific (Ottawa, ON, Canada). coded (ki) and natural (Xi) values, and the equation for conversion
from ki to Xi
Input variable Natural unit Levels Equation
Plant Material –1 0 + 1

Tomato pomace was collected from a tomato processing X1 = Extraction temperature °C 40 55 70 k = X1 −55
1 15
plant located in the state of Goiás (Brazil), immediately X2 = Proportion of EA in %, v/v 30 65 100 k =
2
X2 −65

after being produced. The moisture content of the fresh solvent mixture 35

pomace was ~ 62.8%, and it was composed of 61.5% skin X3 = Solvent:sample ratio mL/g 50 75 100 k = X3 −75
3 25
and 38.5% seeds. The sample was freeze-dried in a bench- X4 = Extraction time min 20 35 50 k =
4
X4 −35
15
top freeze dryer (Liotop L108, Liotop, São Carlos, Bra-
zil) for ~ 16 h to a final moisture content of ~ 5.6%. The
moisture content of samples was determined in vacuum Ultrasound‑Assisted Extraction (UAE) of Lycopene
oven (Lindberg Blue M, Thermo Scientific, Asheville, NC,
USA) at 72 °C and 14 psi until constant weight (AOAC Preliminary analyses indicated that the proposed solvent
method 934.06) [38]. Freeze-dried samples were vacuum mixture EL–EA could obtain 73% higher lycopene yield
packed, transported to Halifax (NS, Canada), and kept at than a mixture of hexane and acetone, both at 1:1 (v/v)
− 20 °C until further analysis. Immediately before extrac- [42]; therefore, its use as solvent for the UAE was further
tion, samples were ground using a manual grinder (Smart- evaluated. The UAE of lycopene from tomato pomace was
grind, Black and Decker, Mississauga, ON, Canada) and conducted in an ultrasound water bath (Branson 2510R-
sieved through a 0.5-mm (32 mesh) sieve. DTH, Branson Ultrasonics Corp., Danbury, CT, USA), with
fixed frequency (40 kHz) and power (100 W), with inter-
nal dimensions of 24 × 14 × 10 cm (L × W × D) and 2.8 L
Box–Behnken Experimental Design capacity, where a constant water level was used as well as
the placement of samples in the bath. The solvent mixture
Optimization of UAE was performed using a Box–Behnken was prepared immediately before use, and placed in water
design (BBD) to determine the experimental points. The bath to achieve extraction temperature before being added
extraction parameters used in the experimental design to the sample. According to the experimental condition
were chosen based on results from other lycopene (Table 2), the appropriate amount of sample was added to
extraction studies [15, 21, 34, 39]. The parameters were 10 mL screw-capped glass tubes. Then, 5 mL of the solvent
X1 = extraction temperature (°C), X2 = proportion of ethyl mixture were added, the tubes were vortexed for 10 s, and
acetate in the ethyl lactate–ethyl acetate solvent mixture placed in the ultrasound bath. The extraction temperature
(%, v/v), X3 = solvent:sample ratio (mL/g) and X4 = extrac- was controlled within ± 1 °C of the set temperature with
tion time (min). The experimental design was established a calibrated thermometer and adjusted using hot and cold
according to a BBD for four factors [40]. In this design, water as needed. At the end of the extraction time, the
each variable (Xi) assumes three levels (ki), coded as − 1, tubes were removed from the bath and centrifuged (Sorvall
0, and + 1. Extreme high (k i = + 1) and low (k i = − 1) RT1, Thermo Scientific, Madison, WI, USA) at 2916×g for
natural values of each parameter were defined based on 10 min at 20 °C for separation of the pellet from the lyco-
preliminary studies, and the centre level (ki = 0) was then pene extract. The supernatant (extract) was then transferred
calculated as the median of the extremes. The natural (Xi) to another tube for spectrophotometric analysis. Extraction
and coded (ki) levels used in the design for each variable was performed under reduced light to reduce carotenoid
are shown in Table 1. degradation.
The BBD matrix is presented in Table 2, with six centre
points. To reduce possible effects of external variations, Determination of the UAE Lycopene Yield
the experiments were conducted in randomized order as
shown in Table 2. The six centre point runs (runs 25–30) Lycopene Calibration Curves
were not randomized, but evenly interspersed among the
other experimental points, since they provide a measure of A lycopene standard curve was determined for each solvent
process stability and allow estimating inherent variability mixture with different EL:EA ratio, where spectrophotomet-
of the process [41] and therefore must be regularly verified ric analyses were performed under dimmed light to reduce
throughout the experiment. carotenoid degradation. A 500 µg/mL lycopene stock solu-
tion was prepared by dissolving pure lycopene standard in

13

2854 Waste and Biomass Valorization (2019) 10:2851–2861

Table 2  Box–Behnken design matrix with experimental factors in coded (ki) and natural (Xi) values, experimental and predicted results for lyco-
pene extraction yield (µg/g d.w.)
Assay Run order Factora Lycopene yield (µg/g Prediction
d.w.) error (%)
Extraction temperature Proportion of EA Solvent:sample ratio Sonication time Experimental Predicted
in solvent mixture
k1 (X1) (°C) k2 (X2) (%, v/v) k3 (X3) (mL/g) k4 (X4) (min)

1 20 − 1 (40) − 1 (30) 0 (75) 0 (35) 1147.1 1125.3 − 1.9

2 22 + 1 (70) − 1 (30) 0 (75) 0 (35) 1004.1 1047.1 4
3 10 − 1 (40) + 1 (100) 0 (75) 0 (35) 849.5 878.7 3.4
4 11 + 1 (70) + 1 (100) 0 (75) 0 (35) 857.5 800.5 − 6.6
5 17 0 (55) 0 (65) − 1 (50) − 1 (20) 1070.3 1016.7 − 5.0
6 5 0 (55) 0 (65) + 1 (100) − 1 (20) 1249.4 1256.2 0.5
7 14 0 (55) 0 (65) − 1 (50) + 1 (50) 1108.7 1086.7 − 2.0
8 6 0 (55) 0 (65) + 1 (100) + 1 (50) 931.3 969.6 4.1
9 29 − 1 (40) 0 (65) 0 (75) − 1 (20) 1065.9 1000.7 − 6.1
10 3 + 1 (70) 0 (65) 0 (75) − 1 (20) 1187.8 1151.6 − 3.0
11 9 − 1 (40) 0 (65) 0 (75) + 1 (50) 1145.0 1121.5 − 2.1
12 16 + 1 (70) 0 (65) 0 (75) + 1 (50) 809.0 814.3 0.6
13 4 0 (55) − 1 (30) − 1 (50) 0 (35) 1067.3 1115.9 4.6
14 21 0 (55) + 1 (100) − 1 (50) 0 (35) 760.9 869.4 14.3
15 23 0 (55) − 1 (30) + 1 (100) 0 (35) 1264.5 1177.1 − 6.9
16 15 0 (55) + 1 (100) + 1 (100) 0 (35) 1022.4 930.5 − 9.0
17 18 − 1 (40) 0 (65) − 1 (50) 0 (35) 1009.8 1030.5 2.0
18 27 + 1 (70) 0 (65) − 1 (50) 0 (35) 1017.8 952.3 − 6.4
19 12 − 1 (40) 0 (65) + 1 (100) 0 (35) 1031.0 1091.7 5.9
20 2 + 1 (70) 0 (65) + 1 (100) 0 (35) 903.1 1013.5 12.2
21 26 0 (55) − 1 (30) 0 (75) − 1 (20) 1193.5 1200.6 0.6
22 28 0 (55) + 1 (100) 0 (75) − 1 (20) 873.7 954.1 9.2
23 24 0 (55) − 1 (30) 0 (75) + 1 (50) 1081.9 1092.4 1.0
24 8 0 (55) + 1 (100) 0 (75) + 1 (50) 915.0 845.8 − 7.6
25 1 0 (55) 0 (65) 0 (75) 0 (35) 1137.7 1082.3 − 4.9
26 7 0 (55) 0 (65) 0 (75) 0 (35) 991.5 1082.3 9.2
27 13 0 (55) 0 (65) 0 (75) 0 (35) 1034.9 1082.3 4.6
28 19 0 (55) 0 (65) 0 (75) 0 (35) 1189.0 1082.3 − 9.0
29 25 0 (55) 0 (65) 0 (75) 0 (35) 1104.4 1082.3 − 2.0
30 30 0 (55) 0 (65) 0 (75) 0 (35) 1012.4 1082.3 6.9
a
 Coded level (natural value)

EA, which was kept at − 20 °C for up to 1 week for determi-

nation of the standard curve. The stock solution was diluted
with the appropriate solvent mixture as needed to give
concentrations varying from 5 to 30 µg/mL. Absorbance Table 3  Wavelength for peak absorbance (λmax) and lycopene stand-
readings were performed immediately after dilution using a ard equation for each solvent (L is lycopene concentration in µg/mL
UV–Vis spectrophotometer (Genesys 10S UV–Vis, Thermo and Amax is the absorbance at the λmax)
Scientific, Madison, WI, USA) in duplicate for each dilution. Solvent (% EA, λmax (nm) Regression equation R2
Wavelength scans from 350 to 550 nm were used to con- v/v)
firm the lycopene three-peak absorbance spectrum, which
30 477 L = 9.65 × Amax 0.9999
consisted of peaks at approximately 450, 470 (maximum),
65 474 L = 7.51 × Amax 0.9985
and 505 nm, with the specific wavelength varying according
100 473 L = 6.98 × Amax 0.9995
to the solvent used [4, 43]. The maximum absorbances (at

13
Waste and Biomass Valorization (2019) 10:2851–2861 2855

approximately 470 nm, Table 3, Online Resource 1) for the (Ypred). Three-dimensional surface plots of lycopene yield
lycopene solutions were used to construct standard curves versus each combination of 2 factors (with remaining two
for each solvent mixture, and linear regression equations factors held at centre level, ki = 0) were used to illustrate the
were subsequently obtained by plotting absorbance against influence of the variables in the response.
lycopene concentration (µg/mL), going through the origin.
4 4 3 4
The regression equations obtained for each calibration curve ∑ ∑ ∑ ∑
are shown in Table 3, where L is lycopene concentration in
Y = 𝛽0 + 𝛽i Xi + 𝛽ii Xi2 + 𝛽ij Xi Xj (2)
i=1 i=1 i=1 j=i+1
µg/mL and Apeak is the absorbance value at wavelength of the
absorption peak (λpeak). The standard equations (­ R2 > 0.99)
were used to calculate lycopene concentration for each Optimization, Model Validation and Comparison
extract obtained from the BBD experimental conditions. Without Ultrasound

The optimal extraction conditions for maximizing lycopene

UAE Lycopene Yield yield with UAE were determined from the final model using
the response optimizer function in Minitab® (v. 17.3.1).
The lycopene extraction yield obtained from each treat- Experiments were then performed in triplicate at the opti-
ment was determined according to Strati and Oreopoulou mal extraction conditions to validate the model. In addi-
[25]. Since lycopene is the major carotenoid in tomatoes tion, results were compared with solvent extraction without
and tomato products [44, 45], lycopene standard curves ultrasound (SE) at the same optimal extraction conditions, to
have been used by other researchers for estimating the total evaluate the effect of ultrasound on extraction yield.
carotenoid content in tomato products [39, 46]. The concen-
tration of lycopene (µg/mL) in each extract was therefore
determined using the peak absorbance in the UV–Vis spec- Results and Discussion
tra in relation to the appropriate standard curve (Table 3).
Then, the lycopene extraction yield was calculated accord- Lycopene Standard Curves in Different Solvent
ing to Eq. 1, where L is the concentration of lycopene in the Mixtures
extract calculated using the regression equations (µg/mL), V
is the final volume of extract obtained (mL), w is the sample The absorbance spectrum of a carotenoid is highly depend-
weight used for extraction (g), and m is the moisture of the ent on the solvent in which the compound is dissolved [43].
sample. Results are expressed as µg of lycopene per g dry To allow the correct estimation of the carotenoid content of
weight (d.w.). the extracts, standard curves of pure lycopene were obtained
L ×V for each solvent mixture used in the extraction (Fig. 1). The
Lycopene yield (μg∕g) = (1) increase in the amount of EA in the solvent mixture from 30
w × (1 − m)
to 100% caused a hyperchromic effect in all concentrations
Response Surface Analysis used [48].

RSM was used to optimize lycopene yield under the stud-

ied conditions, using Minitab® version 17.3.1 software
(Minitab Inc., PA, USA). Responses were fitted to a sec-
ond-degree polynomial equation, as shown in Eq. 2, where
Y is the response variable (lycopene yield, µg/g), β0 is the
constant (or model intercept); βi is the linear coefficient of
each variable Xi; βii is the quadratic coefficient; βij is the
cross-product coefficient of interaction between variables Xi
and Xj; and Xi and Xj are the independent variables. Models
were obtained for coded and natural factors, and analysis
of variance (ANOVA) was performed on the full model
to determine the significant variables affecting extraction
yield (p < 0.05). The initial full quadratic model was refined
by stepwise backward elimination of non-significant terms
[47], using a significance level of 5%. ANOVA was used Fig. 1  Lycopene standard curves in different solvent mixtures:
to evaluate the significance and goodness-of-fit of the final ( ) 100% ethyl acetate, ( ) 65% ethyl acetate–35%
reduced model, which was used to predict lycopene yield ethyl lactate, and ( ) 30% ethyl acetate–70% ethyl lactate

13

2856 Waste and Biomass Valorization (2019) 10:2851–2861

The position of the absorption peak (λmax) of carotenoids the quadratic model fitted the observed experimental results
depends on several factors, such as the molecular environ- well. Also, the lack-of-fit was not significant (p = 0.652),
ment in which the carotenoid molecule is inserted [4, 48]. further suggesting that model goodness-of-fit is satisfactory
For this reason, the same compound shows different absorp- [50].
tion peaks for different solvents. Considering the three sol- The reduced models containing only significant terms are
vent mixtures evaluated, the increase in the amount of ethyl shown in Eqs. 3 and 4, using coded (ki, Eq. 3) and natural fac-
acetate in the solvent mixture caused a hypsochromic shift, tors (Xi, Eq. 4). The regression model was also used to calcu-
that is, a displacement in the λmax to a shorter wavelength late the predicted yields for each treatment, thus allowing the
[48]. The λmax observed for lycopene in 100% EA (473 nm, comparison of the experimental yield with the corresponding
as shown in Table 3) was the same [25] or within a ± 1 nm predicted values (Table 2), which shows that the prediction
variation [49] from the values reported in other studies. error was below 10% for most of the treatments, indicating
good adequacy of the predicted model. Equation 3 shows that
Lycopene Extraction Yield and Modeling the strongest effect on lycopene yield (largest coefficient) is
due to the negative first-order effect of factor X2. Here, the
The lycopene yield at each experimental treatment is shown proportion of EA in the solvent mixture is the factor influ-
in Table 2. Experimental lycopene yield varied between encing the extraction yield the most, and an increase in the
760.9 and 1264.5 µg/g. The centre point replicates ranged proportion of EA causes a decrease in lycopene yield.
from 991.5 to 1189.0 µg/g. Besides providing a measure of Lycopene (μg∕g) = 1082.3 − 39.1k1 − 123.3k2 + 30.6k3 − 54.1k4
process stability and variability, the centre points also pro-
− 60.3k12 − 59.1k22 − 114.5k1 k4 − 89.1k3 k4
vide an estimation of quadratic terms [41]. In this study, the
centre points average was 1078.3 ± 77.7 µg/g, representing (3)
a coefficient of variation of 7.2%. Lycopene (μg∕g) = −1129 + 44.7X1 + 2.75X2 + 9.54X3 + 42.2X4
ANOVA testing was performed on the full quadratic poly- − 0.27X12 − 0.05X22 − 0.51X1 X4 − 0.24X3 X4
nomial model to determine the variables with significant (4)
effect on extraction yield (p < 0.05). The ANOVA results Factors X1 (temperature) and X2 (proportion of EA in sol-
for the final reduced quadratic model are shown in Table 4. vent mixture) showed negative quadratic effects on lycopene
Although the linear terms X1 (p = 0.075) and X3 (p = 0.158) yield. For temperature, while an initial increase will increase
did not have a significant effect on extraction, both were mass transfer rates and extraction yield, a further increase in
included in the model for hierarchical purposes. The final temperature decreases lycopene yield. This is likely due to
reduced model was significant (p < 0.0001), indicating that thermal degradation of lycopene at high temperatures [51,
52]. High extraction yields can be observed at higher tem-
peratures due to several effects, including the decrease in
Table 4  ANOVA results of significant factors in quadratic model for surface tension and solvent viscosity, leading to improved
UAE of lycopene
sample wetting and matrix penetration; the increase in
Source DF SS MS F-value p value molecular motion of the solvent, leading to higher solubil-
Model 8 376,968 47,121 9.00 < 0.0001
ity; and the destruction of the plant matrix, increasing the
 Linear 4 247,118 61,780 11.80 < 0.0001
availability of the target compound to the solvent [53, 54].
  X1 = Temperature 1 18,342 18,342 3.50 0.075 However, an upper temperature limit must be defined to
  X2 = %Acetate 1 182,380 182,380 34.83 < 0.0001 avoid thermal degradation. In the present study, the higher
  X3 = Solvent:sample 1 11,221 11,221 2.14 0.158 temperature likely caused carotenoid degradation, since the
  X4 = Time 1 35,175 35,175 6.72 0.017 optimal temperature was below the maximum of 70 °C.
 Square 2 45,623 22,812 4.36 0.026 The negative quadratic effect of EA proportion in the sol-
  X12 1 25,860 25,860 4.94 0.037 vent mixture (X2) indicates that increasing the amount of EA
  X22 1 24,831 24,831 4.74 0.041 beyond an optimal value causes a decrease in the lycopene
 2-Way interaction 2 84,226 42,113 8.04 0.003 yield, suggesting that EL is more efficient for extraction than
  X1 × X4 1 52,441 52,441 10.02 0.005 EA. The extraction of carotenoids from tomatoes has been
  X3 × X4 1 31,786 31,786 6.07 0.022 previously investigated using EA [19–21] or EL [24–27]
Error 21 109,947 5236 individually, with yields comparable to or higher than
 Lack-of-fit 16 79,740 4984 0.82 0.652 hexane. Our results show that the mixture of EL with EA
 Pure error 5 30,207 6041 was effective for obtaining high lycopene extraction yield
Total 29 486,914 (between 760.9 and 1264.5 µg/g), and the highest yield was
achieved with the lowest proportion of EA, indicating that
DF degrees of freedom, SS sum of squares, MS mean square

13
Waste and Biomass Valorization (2019) 10:2851–2861 2857

EL is the major component responsible for the extraction

efficiency.
The positive linear effect of solvent:sample ratio (X3)
shows that an increase of this ratio (more solvent for a fixed
amount of sample) could further increase the extraction
yield. However, the maximum ratio would be limited by cost
and environmental considerations for the amount of solvent
used. Thus, the maximum amount of solvent recommended
was fixed at 100 mL/g of sample, which is a reasonable level
close to the maximum amount tested in other studies [34,
39].
Extraction time (X4) showed a negative linear effect,
where higher lycopene yields were observed at lower extrac-
tion times. The decreased yield observed at higher extraction
time could have been caused by the longer exposure of the Fig. 3  Response surface plot for effect of interaction X3 × X4
(Solvent:sample ratio  × Time) over lycopene yield (X1 = 55  °C,
bioactive compound to agents that cause carotenoid degrada- X2 = 65% acetate in solvent mixture)
tion, such as light and oxygen [49, 52]. Although the expo-
sure to these agents was minimized during the experimental
procedure, their presence could not have been completely interactions (interaction X1 × X4 and X3 × X4, respectively)
excluded, similarly to the conditions that would be observed over lycopene yield. For each set of two parameters shown in
in a large-scale process. Therefore, it is possible that some the plots, the other two were held at the centre value (ki = 0).
level of thermal- and photo-degradation could have taken The interaction between time and temperature (X1 × X4)
place when the extraction procedure was carried out for a (Fig. 2) shows that high lycopene yields can be achieved
longer period of time. with a combination of either high temperature for short time
or low temperature for longer time. This could be related
RSM Analysis and Optimization of UAE to the ability of the solvent to penetrate the plant cell to
extract the target compound, as discussed previously. At
RSM is a powerful tool for achieving high performance higher temperatures, mass transfer processes are enhanced,
from processes at reduced costs. In this study, RSM was thus less time is required to achieve a high yield, while at
used to optimize the ultrasound-assisted extraction (UAE) lower temperatures more time is required for the solvent
of lycopene from tomato pomace. From the quadratic mod- to reach the cell interior. The sharp decrease in lycopene
els obtained (Eqs. 3, 4), three-dimensional surface plots yield at low temperature and short extraction time observed
were constructed to illustrate the influence of the interac- in Fig. 2 indicates this insufficient access of the solvent to
tion of two factors at a time over the lycopene extraction the plant matrix. Interestingly, a sharp decrease was also
yield. Figures 2 and 3 show the effect of the two significant observed at the combination of high temperature and high
extraction time levels. This could indicate degradation of
lycopene due to prolonged exposure of the sample to the
high temperatures, since exposure to heat is one of the main
factors that can cause lycopene degradation [52].
Figure  3 illustrates the influence of the interaction
between solvent:sample ratio and time (X3 × X4) on lycopene
yield, with the highest yield obtained at high solvent:sample
ratio and short extraction time. The relationship between
solvent:sample ratio and time can be explained by the satura-
tion of the solvent with the extracted compound. At higher
ratios, more solvent is available to dilute the target com-
pound, thus the mass transfer rates are maintained at high
levels for longer [53]. As the highest yield is achieved with
short extraction times, this suggests that the solvent pen-
etrates the sample easily, quickly reaching the cell matrix
and solubilizing the target compound.
Fig. 2  Response surface plot for effect of interaction X1 × X4 (Temper-
ature × Time) over lycopene yield (X2 = 65% acetate in solvent mix- The optimal extraction conditions obtained from
ture, X3 = 75:1 mL/g solvent:sample ratio) RSM analysis were X 1 = 63.4  °C, X 2 = 30% (v/v) EA,

13

2858 Waste and Biomass Valorization (2019) 10:2851–2861

X3 = 100 mL/g solvent:sample, and X4 = 20 min (Table 5), vary considerably among different studies, depending on
with a predicted lycopene yield of 1343.9 µg/g. Validation the raw material and the extraction conditions applied, as
studies were performed in triplicate under the optimal con- shown by a recent review [54]. When using UAE, extrac-
ditions to verify the predicted yield. The average experi- tion yields also depends on the ultrasound equipment char-
mental yield at the optimal extraction conditions was then acteristics, such as power, frequency, shape, and size [28,
1334.8 µg/g, showing good agreement with predicted values 36]. For example, Kumcuoglu et al. [37] optimized UAE of
(only 0.7% error). These results further validate the model, dried tomato processing waste and obtained an extraction
indicating good prediction capacity. yield of 90.1 mg/kg (0.09 mg/g), much lower than the yield
RSM has been used to optimize the UAE of lycopene from the present work (1.3 mg/g), using a solvent mixture
from tomatoes and its by-products using different extrac- of hexane/acetone/ethanol (2:1:1) containing 0.05% (w/v)
tion conditions (Table 6). Lycopene extraction yield can BHT, at a lower solvent:sample ratio. In comparison, Eh and
Teoh [34] obtained a lycopene yield of 5.11 mg/g from dried
tomato pulp by UAE using hexane/acetane/ethanol (2:1:1),
Table 5  Optimized levels of experimental factors in coded (ki) and at a lower temperature and amount of solvent, for a longer
natural (Xi) values
extraction time (Table 6).
Input variable Optimized level The optimized lycopene yield of 1.33 g/kg (d.w.) achieved
Coded (ki) Natural (Xi) in this study shows the effectiveness of a greener extraction
process for obtaining a high value-added product from a
X1 = Extraction temperature 0.6162 63.4 °C widely available agro-industrial waste. The use of tomato
X2 = Proportion of ethyl acetate in –1 30%, v/v pomace as raw material for lycopene extraction represents an
solvent mixture
efficient and environment-friendly use of a food waste, help-
X3 = Solvent:sample ratio 1 100 mL/g
ing to reduce the environmental impact of food production
X4 = Extraction time –1 20 min
[55]. In addition, the use of green solvents and ultrasound

Table 6  Ultrasound-assisted extraction of lycopene from tomato or tomato by-products

Raw material Extraction solvent Optimized UAE conditions Optimal lycopene yield Reference

Tomato pomace, freeze- EL/EA (7:3) • 100 mL/g S:S • 1.33 mg/g (d.w.) Present work
dried • 63.4 °C • UAE increased yield by
• 20 min 10% compared to extrac-
tion under same condi-
tions, without ultrasound
Tomato processing waste, Hexane/ acetone/ ethanol • 35:1 mL/g S:S • 90.1 mg/kg (0.09 mg/g) Kumcuoglu et al. [37]
dried (2:1:1) with BHT 0.05% • 90 W • UAE required less solvent,
(w/v) • 30 min lower temperature, and
less time to achieve similar
yield to the optimized
organic solvent extrac-
tion without ultrasound
(93.9 mg/kg yield achieved
at 50:1 mL/g S:S, 60 °C,
40 min)
Tomato, fresh EA • 8.0:1 mL/g S:S • 89.4% of maximum Lianfu and Zelong [21]
• 86.4 °C lycopene yield achieved at
• 29.1 min optimized UAE
• UAE required shorter
period of time and lower
solvent volume to achieve
similar yield to solvent
extraction with 2% dichlo-
romethane in petroleum
ether, without ultrasound
Tomato pulp, freeze-dried Hexane/ acetane/ ethanol • 74.4:1 mL/g S:S • 5.11 mg/g (d.w.) Eh and Teoh [34]
(2:1:1) • 47.6 °C • Same extraction condi-
• 45.6 min tions, without ultrasound:
1.23 mg/g (d.w.)

S:S solvent:sample ratio

13
Waste and Biomass Valorization (2019) 10:2851–2861 2859

to improve the extraction yield also reduces environmental without ultrasound at optimum conditions was 1209.5 µg/g
impact of extraction processes, which generally involve haz- which was 9.4% lower than with UAE, showing that the
ardous organic solvents and large energy inputs to achieve applied ultrasound treatment promoted an increase in extrac-
desirable yields [28]. Furthermore, with the use of GRAS tion yield. This is the first study evaluating the ultrasound-
solvents there is potential for creating further value from assisted extraction of lycopene from tomato pomace using
the solids remaining after lycopene extraction, and further EL in combination with EA. These results indicate the effec-
work should be conducted to investigate the extraction of tiveness of the novel green solvent mixture proposed for
macronutrients (dietary fiber, proteins, or oils), which would lycopene extraction, and represent a more environmentally
provide greater use of the waste. sustainable approach for lycopene extraction from tomato
pomace and the valorization of industrial tomato process-
Comparison of UAE with Solvent Extraction Without ing waste.
Ultrasound
Acknowledgements  The authors would like to thank the processing
company who provided the tomato pomace for this study and are grate-
To evaluate the effect of ultrasound on lycopene yield of the ful to Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
optimized process, solvent extraction (SE) was performed (Coordination for the Improvement of Higher Education Personnel)
under the same optimal extraction conditions, without ultra- (CAPES—Brazil), the Department of Foreign Affairs, Trade and
sound. A yield of 1209.5 µg/g (d.w.) was obtained, which Development (DFATD—Canada), and the Natural Sciences and Engi-
neering Research Council (NSERC—Canada) for financial support.
was 9.4% lower than with UAE, showing that the applied
ultrasound treatment promoted an increase in extraction
yield, when the other extraction parameters were kept con-
Compliance with Ethical Standards 
stant, contributing towards a greener method for lycopene Conflict of interest  The authors declare that they have no conflict of
extraction. Ultrasonication has been reported to promote an interest.
increase in extraction yield due to acoustic cavitation occur-
ring in the solvent [30]. The corresponding collapse of the
gas bubbles in the liquid results in several physical effects on
the plant matrix, such as fragmentation, erosion, capillarity, References
detexturation, and sonoporation [56], facilitating the release
1. Galanakis, C.M.: Recovery of high added-value components from
of the target compound from the plant matrix to the solvent food wastes: conventional, emerging technologies and commer-
media. This increase in the mass transfer rates contributes cialized applications. Trends Food Sci. Technol. 26, 68–87 (2012)
towards a greener process by the reduction of the amount of 2. Mirabella, N., Castellani, V., Sala, S.: Current options for the val-
orization of food manufacturing waste: a review. J. Clean Prod.
solvent and energy required to achieve a desired extraction
65, 28–41 (2014)
yield. In comparison to this study, Eh and Teoh [34] reported 3. Rao, A.V., Agarwal, S.: Role of lycopene as antioxidant carotenoid
a higher increase of 27% on lycopene extraction yield with in the prevention of chronic diseases: a review. Nutr. Res. 19(2),
the use of ultrasound compared to SE without ultrasound. 305–323 (1999)
4. Rodriguez-Amaya, D.B., Kimura, M.: Harvestplus Handbook for
The difference in the degree of improved extraction observed
Carotenoid Analysis. HarvestPlus Technical Monograph 2. Inter-
in the present study could be due to ultrasound equipment national Food Policy Research Institute (IFPRI) and International
characteristics and could also indicate a higher efficiency Center for Tropical Agriculture (CIAT), Washington, DC (2004)
of the green solvent mixture for penetrating plant cells and 5. Singh, P., Goyal, G.K.: Dietary lycopene: its properties and anti-
carcinogenic effects. Compr. Rev. Food Sci. Food Saf. 7, 255–270
reaching the target lycopene compound.
(2008)
6. Ciriminna, R., Fidalgo, A., Meneguzzo, F., Ilharco, L.M., Pagli-
aro, M.: Lycopene: emerging production methods and applications
Conclusions of a valued carotenoid. ACS Sustain. Chem. Eng. 4(3), 643–650
(2016)
7. Arimboor, R., Natarajan, R.B., Menon, K.R., Chandrasekhar, L.P.,
A green valorization approach was used for lycopene extrac- Moorkoth, V.: Red pepper (Capsicum annuum) carotenoids as a
tion from tomato processing waste, and optimized using source of natural food colors: analysis and stability—a review. J.
RSM. Tomato pomace was used as the raw material, a novel Food Sci. Technol. 52(3), 1258–1271 (2015)
8. Celli, G.B., Teixeira, A.G., Duke, T.G., Brooks, M.S.: Encapsula-
and efficient green solvent mixture (EL and EA) provided
tion of lycopene from watermelon in calcium-alginate microparti-
an alternative to conventional organic solvents, and ultra- cles using an optimized inverse-gelation method. Int. J. Food Sci.
sound was applied to improve extraction yield. The opti- Technol. 51, 1523–1529 (2016)
mized extraction conditions (63.4 °C, 30% (v/v) EA in sol- 9. Shen, R., Jiang, X., Ye, W., Song, X., Liu, L., Lao, X., Wu, C.:
A novel and practical synthetic route for the total synthesis of
vent mixture, 100 mL/g solvent:sample ratio, and 20 min),
lycopene. Tetrahedron 67, 5610–5614 (2011)
achieved a high lycopene yield of 1334.8 ± 83.9 µg/g in a 10. Rath, S., Olempska-Beer, Z., Kuznesof, P.M.: Lycopene extract
relatively short time with a mild temperature. The yield from tomato. Chemical and Technical Assessment (CTA).

13

2860 Waste and Biomass Valorization (2019) 10:2851–2861

http://www.fao.org/filea​dmin/templ​ates/agns/pdf/jecfa​/cta/71/ 29. Vilkhu, K., Mawson, R., Simons, L., Bates, D.: Applications
lycop​ene_extra​ct_from_tomat​o.pdf (2009). Accessed 28 Aug and opportunities for ultrasound assisted extraction in the food
2017 industry—a review. Innov. Food Sci. Emerg. 9, 161–169 (2008)
11. Sharma, S.K., Le Maguer, M.: Lycopene in tomatoes and tomato 30. Suslick, K.S.: Sonochemistry. Science 247(4949), 1439–1445
pulp fractions. Ital. J. Food Sci. 8(2), 107–113 (1996) (1990)
12. Wang, L., Weller, C.L.: Recent advances in extraction of nutra- 31. Zardo, I., Sobczyk, A.E., Marczak, L.D.F., Sarkis, J.: Optimi-
ceuticals from plants. Trends Food Sci. Technol. 17, 300–312 zation of ultrasound assisted extraction of phenolic compounds
(2006) from sunflower seed cake using response surface methodology.
13. FDA (U.S. Food and Drug Administration): Food Additive Sta- Waste Biomass Valoriz. (2017). https​://doi.org/10.1007/s1264​
tus List. U.S. Department of Health and Human Service, Silver 9-017-0038-3
Spring, MD. https​://www.fda.gov/Food/Ingre​dient​sPack ​aging​ 32. Deng, G., Xu, D., Li, S., Li, H.: Optimization of ultrasound-
Label ​ i ng/FoodA​ d diti ​ vesIn ​ g redi ​ e nts/ucm09 ​ 1 048.htm#ftnH assisted extraction of natural antioxidants from sugar apple
(2018). Accessed 07 Apr 2018 (Annona squamosa L.) peel using response surface methodology.
14. ATSDR (Agency for Toxic Substances and Disease Registry): Molecules 20, 448–459 (2015)
Toxicological profile for n-Hexane. U.S. Department of Health 33. Londoño-Londoño, J., Lima, V.R., Lara, O., Gil, A., Pasa, T.B.C.,
and Human Services, Public Health Service, Atlanta, GA. https​ Arango, G.J., Pineda, J.R.R.: Clean recovery of antioxidant flavo-
://www.atsdr​.cdc.gov/toxpr​ofile​s/tp113​.pdf (1999). Accessed 21 noids from citrus peel: optimizing an aqueous ultrasound-assisted
June 2017 extraction method. Food Chem. 119, 81–87 (2010)
15. Strati, I.F., Oreopoulou, V.: Process optimisation for recovery of 34. Eh, A.L., Teoh, S.: Novel modified ultrasonication technique for
carotenoids from tomato waste. Food Chem. 129, 747–752 (2011) the extraction of lycopene from tomatoes. Ultrason. Sonochem.
16. Hartwig, A.: Ethyl acetate. In: The MAK-Collection for Occu- 19, 151–159 (2012)
pational Health and Safety. Part I: MAK Value Documentations. 35. Ahmadi, M., Heidari, O., Nafchi, A.R.M.: Optimization of lyco-
pp. 167–176. Wiley, Hoboken (2012) pene extraction from tomato waste with the integration of ultra-
17. WHO (World Health Organization): Evaluations of the Joint FAO/ sonic—enzymatic processes by response surface methodology. J.
WHO Expert Committee on Food Additives (JECFA). http://apps. Ind. Eng. Res. 1(2), 29–34 (2015)
who.int/food-addit​ives-conta​minan​ts-jecfa​-datab​ase/chemi​cal. 36. Konwarh, R., Pramanik, S., Kalita, D., Mahanta, C.L., Karak,
aspx?chemI​D=2357 (1996). Accessed 07 Apr 2018 N.: Ultrasonication—a complementary ‘green chemistry’ tool
18. Cascant, M.M., Breil, C., Garrigues, S., de la Guardia, M., to biocatalysis: a laboratory-scale study of lycopene extraction.
Fabiano-Tixier, A.S., Chemat, F.: A green analytical chemistry Ultrason. Sonochem. 19, 292–299 (2012)
approach for lipid extraction: computation methods in the selec- 37. Kumcuoglu, S., Yilmaz, T., Tavman, S.: Ultrasound assisted
tion of green solvents as alternative to hexane. Anal. Bioanal. extraction of lycopene from tomato processing wastes. J. Food
Chem. 409, 3527–3539 (2017) Sci. Technol. 51(12), 4102–4107 (2014)
19. Calvo, M.M., Dado, D., Santa-Maria, G.: Influence of extraction 38. AOAC (Association of Official Analytical Chemists): Method
with ethanol or ethyl acetate on the yield of lycopene, β-carotene, 934.06: Moisture in Dried Fruits. Official Methods of Analysis
phytoene and phytofluene from tomato peel powder. Eur. Food of the AOAC International, 18 edn. AOAC International, Gaith-
Res. Technol. 224, 567–571 (2007) ersburg (2005)
20. Lavecchia, R., Zuorro, A.: Improved lycopene extraction from 39. Kaur, D., Wani, A.A., Oberoi, D.P.S., Sogi, D.S.: Effect of extrac-
tomato peels using cell-wall degrading enzymes. Eur. Food Res. tion conditions on lycopene extractions from tomato processing
Technol. 228, 153–158 (2008) waste skin using response surface methodology. Food Chem. 108,
21. Lianfu, Z., Zelong, L.: Optimization and comparison of ultra- 711–718 (2008)
sound/microwave assisted extraction (UMAE) and ultrasonic 40. Box, G.E., Hunter, J.S., Hunter, W.G.: Statistics for Experiment-
assisted extraction (UAE) of lycopene from tomatoes. Ultrason. ers: Design, Innovation, and Discovery, 2 edn. Wiley, Hoboken
Sonochem. 15, 731–737 (2008) (2005)
22. Villanueva-Bermejo, D., Reglero, G., Fornari, T.: Recent advances 41. NIST/SEMATECH: NIST/SEMATECH e-Handbook of Statistical
in the processing of green tea biomolecules using ethyl lactate. A Methods. 5. Process Improvement. http://www.itl.nist.gov/div89​
review. Trends Food Sci. Technol. 62, 1–12 (2017) 8/handb​ook/pri/pri.htm (2012). Accessed 05 Apr 2017
23. Pereira, C.S.M., Silva, V.M.T.M., Rodrigues, A.E.: Ethyl lactate 42. Silva, Y.P.A., Pereira, V.A., Brooks, M.S., Ferreira, T.A.P.C.:
as a solvent: properties, applications, and production processes—a Effect of solvent on lycopene extraction yield from tomato
review. Green Chem. 13, 2658–2671 (2011) pomace. Paper presented at the CSBE/SCGAB Annual General
24. Ishida, B.K., Chapman, M.H.: Carotenoid extraction from plants Meeting and Technical Conference joint with CIGR VI Technical
using a novel, environmentally friendly solvent. J. Agric. Food Symposium, Winnipeg, Canada. http://www.csbe-scgab​.ca/publi​
Chem. 57, 1051–1059 (2009) catio​ns/meeti​ng-paper​s/csbe-techn​ical-confe​rence​s/8072 (2017).
25. Strati, I.F., Oreopoulou, V.: Effect of extraction parameters on the Accessed 25 Aug 2017
carotenoid recovery from tomato waste. Int. J. Food Sci. Technol. 43. Zang, L., Sommerburg, O., Van Kuijk, F.J.G.M.: Absorbance
46, 23–29 (2011) changes of carotenoids in different solvents. Free Radic. Biol.
26. Strati, I.F., Gogou, E., Oreopoulou, V.: Enzyme and high pres- Med. 23(7), 1086–1089 (1997)
sure assisted extraction of carotenoids from tomato waste. Food 44. Tonucci, L.H., Holden, J.M., Beecher, G.R., Khachik, F., Davis,
Bioprod. Process. 94, 668–674 (2015) C.S., Mulokozi, G.: Carotenoid content of thermally processed
27. Strati, I.F., Oreopoulou, V.: Recovery and isomerization of carot- tomato-based food products. J. Agric. Food Chem. 43(3), 579–586
enoids from tomato processing by-products. Waste Biomass Val- (1995)
oriz. 7(4), 843–850 (2016) 45. Gama, J.T., Tadiotti, A.C., Sylos, C.M.: Comparison of carotenoid
28. Chemat, F., Rombaut, N., Sicaire, A., Meullemiestre, A., Fabiano- content in tomato, tomato pulp and ketchup by liquid chromatog-
Tixier, A., Abert-Vian, M.: Ultrasound assisted extraction of food raphy. Aliment. Nutr. 17(4), 353–358 (2006)
and natural products. Mechanisms, techniques, combinations, 46. Periago, M.J.S., Rincón, F., Agüera, M.D., Ros, G.: Mixture
protocols and applications. A review. Ultrason. Sonochem. 34, approach for optimizing lycopene extraction from tomato and
540–560 (2017) tomato products. J. Agric. Food Chem. 52, 5796–5802 (2004)

13
Waste and Biomass Valorization (2019) 10:2851–2861 2861

47. Lundstedt, T., Seifert, E., Abramo, L., Thelin, B., Nyström, Å, 53. Palma, M., Barbero, G.F., Piñero, Z., Liazid, A., Barroso, C.G.,
Pettersen, J., Bergman, R.: Experimental design and optimization. Rostagno, M.A., Prado, J.M., Meireles, M.A.A.: Extraction of nat-
Chemom. Intell. Lab. 42(1), 3–40 (1998) ural products: principles and fundamental aspects. In: Rostagno,
48. Brittton, G.: UV/visible spectroscopy. In: Britton, G., Liaaen- M. A., Prado, J. M. (eds.) Natural Product Extraction: Principles
Jensen, S., Pfander, H. (eds.) Carotenoids. Volume 1B: Spectros- and Applications, pp. 58–88. The Royal Society of Chemistry,
copy, pp. 13–62. Birhäuser Verlag, Basel (1995) Cambridge (2013)
49. Takehara, M., Nishimura, M., Kuwa, T., Inoue, Y., Kitamura, C., 54. Strati, I.F., Oreopoulou, V.I.: Recovery of carotenoids from tomato
Kumagai, T., Honda, M.: Characterization and thermal isomeriza- processing by-products—a review. Food Res. Int. 65, 311–321
tion of (all-E)lycopene. J. Agric. Food Chem. 62, 264–269 (2014) (2014)
50. Berger, P.D., Maurer, R.E., Celli, G.B.: Experimental Design with 55. Sala, S., Anton, A., McLaren, S.J., Notarnicola, B., Saouter, E., Son-
Applications in Management, Engineering, and the Sciences, esson, U.: In quest of reducing the environmental impacts of food
2 edn. Springer, Cham (2018) production and consumption. J. Clean Prod. 140, 387–398 (2017)
51. Henry, L.K., Catignani, G.L., Schwartz, S.J.: Oxidative degrada- 56. Toma, M., Vinatoru, M., Paniwnyk, L., Mason, T.J.: Investiga-
tion kinetics of lycopene, lutein, and 9-cis and all-trans β-carotene. tion of the effects of ultrasound on vegetal tissues during solvent
J. Am. Oil Chem. Soc. 75(7), 823–829 (1998) extraction. Ultrason. Sonochem. 8(2), 137–142 (2001)
52. Lee, M.T., Chen, B.H.: Stability of lycopene during heating and
illumination in a model system. Food Chem. 78, 425–432 (2002)

13