2. Materials and methods

2.1. Materials

Tomato processing waste devoid of seeds was obtained from Pezziol SPA, Parma, Italy. The average particle size and water con tent of the sample were 0.5 mm and 35%, respectively. HPLC grade acetone, n-hexane, methanol and acetonitrile were pur chased from Sigma-Aldrich, USA Lycopene standard of HPLC grade ( 95%) and solid sodium chloride (NaCl) ( 99.5%) were obtained from Sigma-Aldrich, USA. Milli-Qultrapure water with a resistivity of 18.3 MO em was used in all the steps (Millipore, Germany).

2.2. Lycopene extraction and analysis

Lycopene extraction was carried out in a 35 ml screw-top vial placed in a thermostated water bath (0.1 C). 0.5g of sample

was added to the vial and agitated with 20 ml of acetone/n-hexane mixture under appropriate conditions (Tables 1-3). The vial was connected to a peristaltic pump that continuously draws the extract and passes it through a quartz flow cell (Hellma, Germany), mounted on a UV-VIS spectrophotometer (HP 8452A Diode Array Spectrophotometer, Agilent, USA).The extract was passed through a 0.45 J.U11 hydrophobic PTFE membrane filter (Whatman) to elim inate suspended solids before entering the quartz flow cell. Absor bance of the extract was measured continuously every 60 s for

45 min at 472 and 503 nm against a blank of pure solvent mixture.

To minimise the interference from other carotenoids, the concen tration of lycopene was calculated at 503 nm using the molar extinction coefficient 17.2 x 1Q4M-1 cm-1 (Fish, Perkins-Veazie,

8z Collins, 2002). Lycopene content was expressed as mg/100 g fresh weight.

In a second set of experiments, we sought to determine how washing the extract with an aqueous NaCl solution affected the results. At set times (5, 10, 15, 20 and 40 min), 1.00 ml of extract was pipetted manually and washed with 10.0 ml of 1M NaCI solu tion for 15 min and allowed for phase separation. Absorbance of the hexane layer (dehydrated with the addition of anhydrous

CaC03) was measured against a hexane blank and compared with absorbance of non washed extract.

2.3.Totallycopene content detennination

Total lycopene content was determined according to Fish et al. (2002) with minor modifications. Briefly, 1.0 g of sample was extracted with 30 ml of acetone/n-hexane (1:3, vJv) for 15 min and the extraction procedure was repeated until absorbance of the extract at 503 nm was lower than the instrumental noise

3.3.1.1.Effect of mnperature and solvent composition on the rom of extraction and equilibrium concentration. Based on the results obtained by model fitting, the equilibrium concentration of lycopene decreased as the temperature increased (Table 1). According to rate constants, no linear effect of temperature on extraction was evident Overall, the lowest experimental temperature of 30 "C afforded the bighest yield, whJch is unusual, given that nonnally an increase of temperature ravours extraction by enhandng the solubility of the compounds to be extracted. These unexpected results were also observed by Kaur, Wani, Oberoi, and Sogi (2008), when they used response surface methodology to study how conditions affected lycopene extractions l'rom the skin discarded in tomato processing. In their study, the response surface plots of temperature vs. extractioo time revealed that lycopene recovery was highest at the lowest temperature studied ( 20 "C). the lower yields of lycopene at higher temperatures could be due to oxidatidegradation and isomerisation (E-Z) reactions favoured by the hilber extraction temperatures. Isomerisation of lycopene causes a decrease in the visible-band absorption in lN VlS spectrophotometric analysis (Sadler et al., 1990; Tan & Soderstrom, 1989), thereby, indicating a lower extraction yield during analysis. In addition, the cis isomers showed an additional absorptioo band around 320-360 nm (Berg et al., 2000), which did not account for the lycopene quantities obtained in our study.

the acetona-hexano mezcla (50 C) .Estos resultados indican que el uso de disolvente n-hexano no polar, la transferencia de masa mejorada y en consecuencia ofrece una produccin de licopeno superior. No hubo relacin lineal observable entre la composicin solvente y constante de velocidad

3.3.2. Modelo de transferencia de masa

Desde el modelo de estado estacionario (Spiro &] hace, 1982). lo observado constante de velocidad (. Koa) puede expresarse de la siguiente manera:

(15)

)

where the solvent diffusion coeffident in the mattix isD (m1 min-1

The Calvo, Dado, and Santa-Marfa (2006) group found that the

yield of lycopene isomers (E and Z), obtained from the ethanol extraction of lycopene from tomato skin powder at 60oc, was lower compared to theat SO oc,indicating extensive isomerisation and predominantly oxidative degradation at high tempera ture in ethanol. However, they observed that lycopene ( isomers) concentration Increased with the increase in the extrac

tioo temperature from 25 oc up to SO ac. They also noted that the

yield of another compound of interest, p-carotene, deaeased with

the increase in the extractioo temperature.The highest P-carotene concentration was obtained at 25 C and the lowest values were

found at 60 c.

Varios otros autores tambin han experimentado resultados similares durante la extraccin de otros fitoqumicos. Karacabey y Mazza (2008) estudiaron la cintica de extraccin de cido ferlico de fresadas bastones de uva, en el que las temperaturas ms bajas produjeron mayores cantidades de cido ferlico. Cacace y Mazza (2003) tambin observaron resultados similares para la extraccin de antocianinas de las bayas molidas. En el estudio actual, la presencia de una cantidad substancial de agua r1 (aproximadamente 35X) en la materia prima utilizada para la intervencin podra ser otra razn para rendimientos ms bajos de licopeno a temperaturas ms altas. El aumento de temperatura aumenta la solubilidad en agua en el sistema disolvente y la presencia de agua disminuye la solubilidad del licopeno en la fase lquida, lo que resulta en rendimientos de extraccin ms bajas.

Se observ una relacin lineal entre la composicin del disolvente y la concentracin de equilibrio. Las concentraciones de equilibrio calculados para el licopeno disminuyeron al aumentar la concentracin de acetona. Se observ que el rendimiento ms alto de licopeno en 1: 3 en acetona-hexano (30 "C), mientras que la ms baja se observ en 3: 1

and the particle size of the tomato paste is d (m1 By assuming that the paste is homogenous and all the samples had the same particle size distribution. then d oc n, where n is a pure number.As a direct consequence of this assumption:

).

K""' oc D(16)

The goodness of fit was approoml by ANOVA table, and the results of application of the diffusion model on the experimental data set are shown in Table 2.

3.3.2.1. Elfect of temperature and solvent composition on K'..6. or diffusion co .J1fdent Increase in the temperature caused a negati effect on K,x,.. indicating lower diffusivity at these temperatures (Table 2). Overall, a higher rate of diffusivity was observed at lower temperatures.11lf! extraction yield also decreased with increase In the temperature. This is rather unexpected, as temperature rise should increase the solubility and diffusivities (Goula, 2013). On the basisof this behaviour, one could assume that inthe extraction of lycopene from tomato sing waste, the rate determining step, espedally in our case, may be the penetration of solvent into the mattix or the bond creation between the analyte and the sol vent (the first step of solid-liquid extraction) rather than diffusion (the second step). Since the matrix CDiltaiDS about 3S% water, the penetration of solvent mixture into the bulk of the matrix may be hindered. as the solvent mixture CDiltains a non polar hexane phase. In additioo, if the rate determining step was diffusioo, the diffusion coefficient should vary with the temperature in accor dance with the .AIThenius type of relationship Eq. (17) (cacace & Mazza, 2003) (where Dis the diffusivity, E,is the energy of activa-

fils. z. CompM!Jon ot aoodness of fit for dlJferent madlemadQI modeb (A) r..,. and RMSI!. (B) MBI! and 7!'.

M.M. Poojary,l'. PaJsamtmlj/flood OlenUmy 173(2015) 943-950

tion for diffusion, R is the universal gas constant, and Tis the abso lute temperature), and the plot of In D vs 1/T should be a straight line Uaganyi 8t Price. 1999). But our data sets did not match with this behaviour and the correlation between D and 1/Twas not lin ear (data not shown).

KoJJs ex: D =Aexp-&fr (17)

Thus we can assert that diffusivity was dependent on the nature of the solvent mbcture.lnaease In the acetone content had a neg ative effect on D in a linear manner. Enhanced diffusivity was observed with inaease inn-hexane content in the solvent mixture. The solubility of different natural products varies with the nature

of solvents. This is especially the case for polarity:polar solute is

soluble in polar solvents whereas nonpolar solutes dissolve in

non-polar solvents.Since lycopene is fat soluble, diffusion is pro

moted by inaeases in the non-polar content of the solvent.

949

3.3.3. Peleg's model

The applicability of Peleg's equation on food material has been demonstrated extensively (Odriozola-Serrano, Soliva-Fortuny, Gimeno-Afl6, 8t Martin-Belloso, 2008: Park, Bin. Reis Brod, 8t Brandini Park, 2002:PlaniniC. VeJiC. Tomas, BiliC. 8t BuciC. 2005: Turban, Sayar, 8t Gunasekaran. 2002). In the present study, Peleg's model was adapted to experimental conditions (Eq.(12)1and used for experimental data approximation.The calculated parameters of

Peleg's constants, namely Kt and Kz, and values of Bo, ('_,. rAI!I

RMSE. MBE and X: are shown in Table 3.It should be noted that

a lower K1 value in Eq. (13) implies a faster rate of the process,

whilst the lower Kz value in Eq.(14) indicates maximum yield

(Ghafoor, Misra, Mahadevan. 8t Tiwari, 2014). Peleg's model

showed a better fit to the experimental data compared to the other models investigated in this study.The model fitted well with the

experimental data, with good accuracy (average r.A4 =0.986.aver

age RMSE= 0.06213, average MBE= 0.00543 and average

i'" 0.005157) and revealed that there was a good concordance

between experimental and predicted values of yield.

(C)

4.0

..,.

3.904

3.850

J./00

3.300

----

4.0&:1

4,0241

3.987

-3.911

l.,

3.3.3.1.Effect of temperature and solvmt composition on the K1 and

. Solvent extraction significantly affected the Kt parameter, but

3.93,854

noo

3.8 3.742

did bave such marked effect on one. PeJeg's constant Kz

,..3..665

cn s:r

inaeased with an increase in the temperature, indicating that

3.629

there is a lower extraction yield of lycopene at higher tempera

tures; the parameter K1also increased with increase in tempera ture, indicating a higher rate of reaction at lower temperatures. The highest yield was observed at the lowest experimental tem perabJre (30 OC), which is in agreement with the results obtained

with the first order kinetic model and the mass transfer model.

Peleg's rate constant (K1)was not linearly correlated to the sol vent composition. However, solvent composition affected Kz and Ceq linearly.Aninaease in the acetone concentration showed posi tive effects on theKl parameter, indicating lower extraction yields.

3.4.Comparison of ldnetic models

Generally, the values of adjusted correlation coeffident ( ). the root mean square error (RMSE), the mean bias error (MBE) and the reduced chi-square (i'") are used to evaluate the goodness of fit and to select the best model for a given experimental data set

(Hayaloglu et al., 2007).The higher the r-44 value and the lower the

RMSE. i'" and MBE. the better the model fits the experimental data.

Fig.2 compares variations of rAI!I RMSE, MBE and i'" respectively

for each tested model.All tested models exhibited a good fitting

performance for the experimental data set. Amongst the models studied, the Peleg model fitted best with the experimental data and exhibited good accuracy by showing the highest average value

J"j'3.572

j,516

.._13-D smflre plots sbowlogtbe efJeds of extt.adion l2mpel3l:uJe mel solvent CDIIIpOSition 011 lycopene yield. (A) Pirst order kioetic model, (B) mass tr.msfer moclel, ami {C) Peleg'5 model (3-D smface plots wen! obtlined by griddiog sperlmeotal data Into equally spaced 20 K 20 matrix).

of r..14i and the lowest average values of RMSE, MBE and y_2 (Fig.2). The data fitting and accuracy was in the order: Peleg's model > mass transfer model> first order kinetic model. Fig. 1 illustrates the concordance between extraction yields obtained from the experimental data set and from the mathematical models. It also compares the fitting accuracy of each model for the given experi mental data. The 3-D surface plots shown in Fig. 3 indicate the effects of extraction temperature and solvent composition on lycopene yield (mgf100gA maximum yield of lycopene was

observed at the lowest experimental temperature (30oq and at the solvent composition containing the highest quantity of n-hexane (acetone/n-bex.ane 1:3, v/v).

950 M.M.Poajary, P. Passamanti/Food Chemistry 173 (2015) 943-950

4. Conclusions

Kinetic modelling was used to optimise the extraction of lyco pene from tomato processing waste.The results demonstrated that the process variables, namely temperature and solvent composi tion have a strong effect on the extraction efficiency. Based on the models investigated, the best operational conditions for the extraction of lycopene were the lowest temperature (30 C) and a solvent mixture containing the highest quantity of n-hexane (ace tone/n-hexane 1:3, v/v). Peleg's model appears to offer the best fit for the experimental data, signifying the highest AI!Iand the low

est RMSE, MBE and X: values, compared to other tested models.

Aclmowledgemenb

The authors of this manuscript are thankful to Prof. Giorgio Mancini, Department of Physics, University of Camerino, Italy for assisting in mathematical modelling.

References

Arab, L, & Steck, S. (2000). Lycopene and cardiovascular disease. The American journal of Clinical NuiTition, 71,1691s-1695s.

Barba, A. I. 0Hurtado, M. C., Mata, M. C. S., Ruiz, V.F., & De Tejada, M. L S.{2006).

Application of a UV-vis detection-HPLC method for a rapid detennination of

lycopene and 1!-carotene in vegetables. Food Chemistry, 95,328-336.

Berg. H. Van Den, Faulks, R., Granado, H. F., Hirschberg.)., Olmedilla, 8., Sandmann, G., et al. (2000). The potential for the improvement of carotenoid levels in foods and the likely systemic effects. journal of the Science of Food and Agriculture, 80

880-912.

Cacace, J.E.,8t Mazza, G.{2003). Mass transfer process during extraction of phenolic compounds from milled berries. journal of Food Engineering, 59, 379-389.

Calvo, M. M., Dado, D., &. Santa-Marfa, G. {2006). Influence of extraction with ethanol or ethyl acetate on the yield of lycopene, fl-carotene, phytoene and phytofluene from tomato peel powder. European Food Research and Technology,

224, 567-571.

Chalabi, N., Delort, L, LeCorre, L, Satib, S., Bignon, J8t Bernard-Gallon, D.{2006).

Gene signature of breast cancer cell lines treated with lycopene.

Pharmacogenrmrics, 7, 663-672.

Cheung, Y...C., Siu, K.-C., Ill Wu, J.-Y.{2012). Kinetic models for ultrasound-assisted extraction of water-soluble components and polysaccharides from medicinal fungi. Food and Bioprocess Technology, 6, 2659-2665.

Ointon, s. K. (1998). Lycopene: Chemistry, biology, and implications for human

health and disease. Nulrition Reviews, 56(2 Pt 1), 35-51.

Crank, J. (1975). The mathematics of diffusion (2nd ed.). Oxford: Oarendon press. Eh, A. L-S., 8t Teoh, S.-G.{2012Novel modified ultrasonication technique for the

extraction oflycopene from tomatoes. Ultrasonics Sonochemistry, 19,151-159. Fish, W. W., Perkins-Veazie, P& Collins, j. K. (2002). A quantitative assay for

lycopene that utilizes reduced volumes of organic solvents. journal of Food

Composition and Analysis, 15, 309-317.

Ghafoor, M., Misra, N.N., Mahadevan. K.,IllTiwari,B.K.{2014). Ultrasound assisted hydration of navy beans (Phaseolus vulgaris). Ultrasonics Sanochemistry 21

409-414. '

Goula, A. M. (2013). Ultrasound-assisted extraction of pomegranate seed oil - Kinetic modeling.journal of Food Engirleering. 117, 492-498.

Hatami, T., Cavalcanti, R. N., Takeuchi, T. M., Ill Meireles, M. A. A. {2012).

Supercritical fluid extraction of bioactive compounds from Macela (AchyrocHne satureiofdes) flowers: Kinetic, experiments and modeling. The journal of SUpercrttical Rukls, 65, 71-77.

Hayaloglu, A. A., Karabulut, 1., Alpaslan, M., 8t Kelbaliyev, G. {2007Mathematical

modeling of drying characteristics of strained yoghurt in a convective type tray

dryer.joumal of Food E1J8ineering. 78(1), 109-117.

Holzapfel, N. PHolzapfel, B. M., Champ, S., Feldthusen, J., Cements, J., 8t Hutmacher, D. W. (2013). The potential role of lycopene for the prevention and therapy of prostate cancer: From molecular mechanisms to dinical evidence. International journal of Molecular Sciences. 14, 14620-14646.

Jaganyi, D., & Price, R. D. {1999). Kinetics of tea infusion: The effect of the manufacturing process on the rate of extraction of caffeine.Foad Chemistry 64

27-31.

Karacabey, E.. & Mazza, G. {2008). Optimization of solid-liquid extraction of resveratrol and other phenolic compounds from milled grape canes (Vilis vtnifera). journal of Agrlculturul and Food Chemistry, 56, 6318-6325.

Kaur, DWani, A. A., Oberoi, D. P. SIll Sogi, D. S.{2008). Effect of extraction conditions on lycopene extractions from tomato processing waste skin using response surface methodology. Foad Chemistry, 108, 711-718.

Lavecchia. R., Ill Zuorro, A. (2008).Impd lycopene extraction from tomato peels using cell-wall degrading enzymes. European Food Research and Technology,228,

153-158.

liZ., Ill Zelo, L (2008): Optimization and comparison of ultrasound( rrucrowave asststed extraction {UMAE) and ultrasonic assisted extraction (UAE) of lycopene from tomatoes. Ultrasonics Sonochernistry, 15, 731-737.

lin, C. H., & Chen, B. H.(2003). Determination of carotenoids in tomato juice by

liquid chromatography. journal a/Chromatography A. 1012, 103-109.

Naviglio, D., Caruso, T., Iannece, P., AragOn, A., 8t Santini, A.{2008). Characterization

of high purity lycopene from tomato wastes using a new pressurized extraction

approach. journal of Agrlculturul and Food Chemistry, 56,6227-6231.

Odrlozola-Serrano, l., Sollva-Fortuny, R., Gimeno-Afl6, V., 8t Martfn-Belloso, o.

(2008). Modeling changes in health-related compounds of tomato juice treated

by high-intensity pulsed electtic fields. journal of Food Engineering. 89, 210-216.

0sterlie, M., Ill Lerfall, J. {2005). Lycopene from tomato products added minced

meat: Effect on storage quality and colour. Food Research Inremationa38,

925-929.

Papaioannou, E. H., Ill Karabelas, A. j. {2012). Lycopene recovery from tomato peel under mild conditions assisted by enzymatic pre-treatment and non-ionic surfactants.Acta Biochimica Polonim, 59, 71-74.

Park, K. j., Bin, A., Reis Brod, F. P., & Brandini Park, T. H. K. {2002). Osmotic

dehydration kinetics of pear D'anjou {communis L). journal of Food

E118ineering, 52, 293-298.

Peleg, M. (1988). An empirical model for the description of moisture sorption

curves.Joumal of Food Sdence,53,1216-1219.

Perretti, G., Troilo, A., Bravi, ll., Marconi, o., Galgano, F., 8t Fantozzi, P. (2013).

Production of a lycopene-enriched fraction from tomato pomace using supercritical carbon dioxide. The journal of Supercritical Fluids, 82, 177-182.

PlaniniC, M., VeliC, D., Tomas, s., Bilic, M., a. Bucic, A. (2005). Modelling of drying and

rehydration of carrots using Peleg's modeL European Food Research and

Technology, 221, 446-451.

Rali, M. M., Yadav, P. N., 8t Reyes, M. (2007). Lycopene inhibits LPS-induced proinflammatory mediator inducible nitric oxide synthase in mouse macrophage cells. journal of Food Sdenre, 72, 569-574.

Rao, A. V., 8t Agarwal, S. (1999). Role of lycopene as antioxidant carotenoid in the prevention of chronic diseases: A review. NuiTilion Research, 19, 305-323.

Sadler, G., Davis, j., a. Dezrnan, D. (1990). Rapid extraction of lycopene and !>

Carotene from reconstituted tomato paste and pink grapefruit homogenates.

journal a/Food Science, 55,1460-1461.

Spiro, M., a.Jago, D. s. (1982).Kinetics and equilibria of tea infusion. part 3.Rotation

disc experiments interpreted by a steady state model. journal of Chemical Society

Transactions, !(78), 298-305.

Tan, B., & Soderstrom, D.N.(1989QJialitative aspects ofUV-vis spectrophotometry

of fl-carotene and lycopene. journal of Chemlcal Education, 66, 258. Taungbodhitham, A. K., jones, G. P., Wahlqvist, M. L, & Briggs, D. R. (1998).

Evaluation of extraction method for the analysis of carotenoids in fruits and vegetables. Food Chemistry, 63, 577-584.

Turban, M., Sayar, S.,8t Gunasekaran, S.(2002). Application of Peleg model to study

water absorption in chickpea during soaking. journal of Food Engineering, 53,

153-159.

Xu, Y., 8t Pan, S. (2013). Effects of various factors of ultrasonic treatment on the extraction yield of all-trans-lycopene from red grapefruit (atrus parndise Mad.). UltrasonU:s Sonochemistry,20(4), 1026-1032.

Zuknik, M.H., Nile Norulaini, N.A., & Mohd Omar, A. K.{2012).Supercritical carbon

dioxide extraction of lycopene: A review. journal of Food Engineering. 112

253-262.

Zuorro, A., Fidaleo, M., 8t Lavecchia, R. {2011). Enzyme-assisted extraction of lycopene from tomato processing waste. Enzyme and Microbial Technology 49

567-573.

Web references

Global Agricultural Information Network; http;/(gain.fas.usda.guv( Recent%20GAlN%20PUblications(Tomatoes%20Annual%20Report.RomeJtaly_

6-4-2009.pdf {last accessed 28/03/14).