lable at ScienceDirect

Journal of Food Engineering 172 (2016) 9e18

Contents lists avai

Journal of Food Engineering

journal homepage: www.elsevier .com/locate/ j foodeng

Drying characteristics and quality of grape under physicalpretreatment

G. Adiletta a, P. Russo b, *, W. Senadeera c, M. Di Matteo a

a Department of Industrial Engineering, University of Salerno, Via Giovanni Paolo II 132, 84084 Fisciano, SA, Italyb Department of Chemical Engineering Materials Environment, University of Rome “La Sapienza”, Via Eudossiana 18, 00184 Roma, Italyc School of Chemistry Physics and Mechanical Engineering, Faculty of Science and Engineering, Queensland University of Technology, 2 George Street,Brisbane, QLD 4001, Australia

a r t i c l e i n f o

Article history:Received 7 October 2014Received in revised form29 May 2015Accepted 23 June 2015Available online 27 June 2015

Keywords:GrapeDryingAbrasive pretreatmentAntioxidant activityShrinkageEmpirical models

* Corresponding author.E-mail address: paola.russo@uniroma1.it (P. Russo

http://dx.doi.org/10.1016/j.jfoodeng.2015.06.0310260-8774/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

Grape drying is a slow and energy intensive process because the waxy peel has low permeability tomoisture. Therefore, peel chemical and physical pretreatments are considered before drying in order tofacilitate water diffusion. However, they cause heterogeneity in the waxes removal and problems duringshelf-life.

In this paper an alternative abrasive pretreatment of grape peel, for enhancing the drying rate andpreserving the samples, was applied to Red Globe grapes. Convective drying experiments were carriedout at 40e70 �C and at 2.3 ms�1 air velocity. The effect of wax abrasive pretreatment on the drying ki-netics and quality parameters of raisins was investigated. The results were compared with those ofsamples pretreated by dipping in alkaline ethyl oleate solution and untreated grapes. All the driedsamples are darker than fresh one and shrunked. The samples pretreated by peel abrasion and dried at50 �C showed the lowest color changes, less shrinkage and the best rehydration capacity. The dryingkinetics and shrinkage curves were also analyzed using some commonly available empirical models.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Grape is a non-climacteric fruit that grows on the perennial anddeciduous woody vines of the genus Vitis. It contains large amountsof phytochemicals including phenolics, flavonoids, anthocyaninsand resveratrol, which offer health benefits. Antioxidant com-pounds include vitamins, phenols, carotenoids, and flavonoids.Among the last group, flavones, isoflavones, flavanones, flavonols,anthocyanins and catechins are the most important, and exhibitsubstantial antioxidant activity (Wang et al., 1997). The high con-tent of grapes in phenolics, flavonoids, and anthocyanins have beensuggested to be responsible for their health benefits (Yang et al.,2009).

Grapes are one of the major dietary sources of anthocyanins,which are responsible for the coloring of black, red and purplegrapes. Anthocyanins are reported to have antioxidant activity,anti-inflammatory activity, anticancer activity, apoptotic inductioneffect, a-glucosidase inhibition activity, vision benefits and effects

).

on collagen, blood platelet aggregation and capillary permeabilityand fragility (Hou, 2003).

Grapes have relatively high sugar content and moisture content,are very sensitive to microbial spoilage during storage. Therefore,after harvest they must be consumed or processed into variousproducts in a fewweeks in order to reduce economic losses. Dryinggrapes into raisins is a major processing method in almost allgrape-growing countries. Traditionally raisins are obtained by sundrying of the fruit for 8e10 days, which substantially reduces watercontent. This drying method is cheap, but there is a risk of damagedue to dust and insect infection (Pangavhane and Sawhney, 2002).An alternative to it is hot air drying. In general, the dehydrationcauses damages in texture, color, taste and nutritional value of fooddue to the high temperatures (60e75 �C) and long drying timesrequired in the process (2e3 days) (Singh et al., 2012). According toCarranza-Concha et al. (2012), the dehydration of grapes affectstheir content of polyphenols, ascorbic acid and antioxidant activity.That is why efforts should be made to reduce drying times but alsoto decrease the temperatures used in the drying processes in orderto obtain better quality products.

Drying of grape is difficult because grapes contain a wax outerpeel layer which acts as a barrier to moisture movement across the

G. Adiletta et al. / Journal of Food Engineering 172 (2016) 9e1810

membrane. Therefore, various drying methods and pretreatmentswere investigated as reviewed by Esmaiili et al. (2007a).

Currently, chemical pretreatment methods are used frequentlyto dissolve the wax layer and accelerate drying rate (Bingol et al.,2012; Doymaz, 2006; Doymaz and Pala, 2002; Esmaiili et al.,2007b). Chemical pretreatment consists of dipping grapes into analkaline solution, i.e. NaOH (Berna et al., 1991; Femenia et al., 1998;Carranza-Concha et al., 2012), or in oil emulsion: ethyl oleate andK2CO3 (Cinquanta et al., 2002; Doymaz, 2004) for several minutes.By this way, the wax is dissolved, thus reducing the resistance towater diffusion through the peel.

Doymaz and Pala (2002) studied the influence of dipping solu-tions such as ethyl oleate and potassium carbonate on drying timeand color quality of dried untreated and pretreated grapes. Theyfound that grapes dipped into alkaline emulsion of ethyl oleateshowed shorter drying times than those untreated, or pretreatedwith potassium carbonate solution. About color analysis, the bestresults were obtained for grapes which were pretreated with ethyloleate and dried at 60 �C.

Esmaiili et al. (2007b) studied the effect of different pre-treatments (i.e. dipping in alkaline emulsion of ethyl oleate or in hotwater) on moisture diffusivity of seedless grapes. They showed thatboth pretreatments activated the diffusion process by reducing theskin resistance to water transfer at the beginning of the process. Theethyl oleate pretreatment was found much more efficient by main-taining higher internal diffusivities at the end of the drying process.Moreover, the increase in mass transfer coefficients at differenttemperatures for the ethyl oletate pretreated sampleswas two timesthat for the hot water pretreated samples during the drying.

Vazquez et al. (1997), pretreating the grape peel by potassiumcarbonate solution, found the presence of micro-fissures in thegrape peel.

However, the chemical additive residue in the raisins may causefood safety problems and how to deal with larger quantities ofcorrosive chemicals is a serious problem.

Beside chemical pretreatment, some physical pre-treatmentshave also been developed for drying grapes. An alternative phys-ical pre-treatment, consisting of abrasion of the peel of grapes, wasproposed by Di Matteo et al. (2000). They found that the pattern ofthe drying curves was quite similar for both chemical (alkalineethyl oleate solution) and physical treated samples. Therefore,abrasion was found to be as effective as the traditional chemicalmethod. The calculated mass transport coefficient for physicaltreated samples was about 4 times greater than that determined foruntreated samples.

The same pretreatment was performed on plums, before dryingat 60 �C, without significantly altering the qualitative characteris-tics of the plums, reducing drying time and causing a smaller loss ofsugars (Cinquanta et al., 2002; Di Matteo et al., 2002).

The main purpose of this study was to investigate the effects ofabrasive pre-treatment and of air drying conditions (i.e. tempera-ture) on the quality properties of raisins, with the aim of produceminimally processed high quality raisins. We then compared theresults with those relevant to raisins obtained by grape pretreatedwith alkaline ethyl oleate solution and untreated raisins.

Moreover, in order to represent the observed change in mois-ture during drying experiments, data of moisture ratio were fittedusing empirical models available in the literature.

Since shrinkage influences drying and consequently rehydra-tion, it has to be taken into account in mathematical modeling ofsuch processes. Several types of models predicting volume changesare also available in the literature (Brasiello et al., 2013; Khalloufiet al., 2009; Mayor and Sereno, 2004). Some empirical modelswere here tested to verify their ability to represent experimentaldata of shrinkage volume during drying.

2. Materials and methods

2.1. Sample preparation

Red grapes (Vitis Vinifera) cv. Red Globe with an initial moisturecontent of 6.43 ± 0.02 kg water/kg db and average diameter of24.4 ± 1.95 mm were used for the experiments. The fresh grapeberries with uniform size and color and without surface damage ordiseases were selected for the experiments and stored at 4 ± 0.1 �Cbefore handling (up to 12 h).

Three kinds of samples were compared in this study: untreatedgrape (UTR), abraded grape (TR-Abr) and dipped grape in chemicalsolution (TR-EtOl).

Before drying, TR-Abr samples were submitted to a physicalabrasive pretreatment. The abrasion of the grape peel was carriedout in a motorized rotating drum (D ¼ 240 mm, L ¼ 250 mm) madeof plexiglass, lined inside with sandpaper (US CAMI grit 400). Therotation speed of drum was 10 rpm, the pretreatment time was15 min and the mass of grapes was 4 kg.

TR-EtOl samples were dipped in an aqueous solution of 2% (v/v)ethyl oleate and 2.5% (v/v) Na2CO3 at 40 �C for 3 min.

2.2. Drying experiments and empirical models

Drying of grapes was carried out in a convective dryer (ZanussiFCV/E6L3) at constant temperature at 40, 50, 60 and 70 �C and at airvelocity of 2.3 m/s until the water content plateau was reached(about 0.30 kg water/kg db). During drying at regular intervals thesamples were weighted by means of a digital balance (mod.Gibertini E42, Italia).

The results were reported in terms of Mt/M0 vs time (min),where Mt was the moisture content (kg water/kg db) at a givendrying time and M0 was its initial value. They are reported asaverage of three sets of experiments.

All the raisins obtained by different pretreatment and atdifferent drying temperatures were then characterized.

Empirical models that are commonly applied for vegetable foodmaterials were here adopted (Table 1). The empirical constants forthe drying models were determined experimentally from normal-ized drying curves (moisture ratio MR ¼Mt/M0 vs time) at eachdrying temperature. Non-linear least square regression analysiswas used to evaluate the parameters of the selected model with theLevenbergeMarquardt procedure. The goodness of fit for eachmodel was evaluated based on the statistical parameters R2, and c2,RMSE calculated as follows:

c2 ¼PN

i¼1 MRexp;i�MRpre;i

� �N � z

2

(1)

RMSE ¼ 1N

XNi¼1

MRpre;i�MRexp;i

� �2" #1=2(2)

The reduced c-square is the mean square of the deviations be-tween the experimental and calculated values for the models. Thelower the values of the reduced c-square, the better the goodnessof the fit. The RMSE, root mean square error, gives the deviationbetween the predicted and experimental values and it is required toreach zero.

2.3. Color

For all samples, color was obtained through a colorimeterMinolta Chroma Meter II Reflectance CR-300 (triple flash mode

Table 1Mathematical models applied to drying curves.

Model name Equation References

Henderson and Pabis MR ¼ a exp (�kt) Henderson and Pabis (1961) and Park et al. (2002)

Page MR ¼ exp (�ktn) Doymaz (2005) and Kashaninejad and Tabil (2004)

Logarithmic MR ¼ a exp (�kt) þ c Yagcioglu et al. (1999)

Two term MR ¼ a1 exp (�k1t) þ a2 exp (�k2t) Henderson (1974)

G. Adiletta et al. / Journal of Food Engineering 172 (2016) 9e18 11

aperture 10 mm). It was recorded using CIE L*a*b* uniform colorspace (CIE-Lab). The lightness value (L*) indicates the darkness/lightness of the sample, a* is a measurement of the greenness/redness of the sample and b* is the extent of blueness/yellowness.The Hue angle (H), which indicates how an object’s color isperceived by human eye: red, orange, green or blue, was calculatedas follows:

H ¼ tan�1b*�a* (3)

2.4. Total phenolic content and antioxidant activity

Total phenolics were extracted from fresh and dried sampleswith 80% acetone solution following the method described by Yanget al. (2009). All extractions were performed in triplicate.

The total phenolic content was determined using theFolineCiocalteu colorimetric method (Singleton et al., 1999), whichwas modified by Yang et al. (2004). The grape extracts wereoxidized by the FolineCiocalteu reagent in sodium carbonate

Fig. 1. Experimental (symbols) and predicted (lines) drying curves of untreated (UTR) and

solution. The absorbance was measured at 760 nm after 90 min atroom temperature by a spectrophotometer (Lambda Bio 40; PerkinElmer, Waltham, MA, USA). All values were expressed as the mean(mg GAE/100 g db) ± SD of three replications.

The DPPH radical scavenging activity of fresh and dried grapeswas determined using the method proposed by Katalinic et al.(2006). Aliquot (50 mL) of the undiluted extract was placed in acuvette, and 2 mL of 6 $ 10�5 M methanol solution of DPPH radicalwas added. Absorbance measurements commenced immediately.The decrease in absorbance at 517 nmwas determined after 16 minfor all samples. Methanol was used to zero spectrophotometer. Theabsorbance of the DPPH radical without antioxidant, i.e. the control,was measured daily.

The % inhibition of DPPH radical caused by the samples wasdetermined according to the following equation:

% inhibition of DPPH ¼ ACð0Þ � ASðtÞ� �.

ACð0Þh i

*100 (4)

where AC(0) is the absorbance of the control at t ¼ 0 min and AS(t) is

pretreated (TR-ABR, TR- EtOl) samples at (a) 40 �C, (b) 50 �C, (c) 60 �C and (d) 70 �C.

Table 2Correlation coefficients (R2, RMSE, c2) of the drying models.

Model name Temperature (�C) Samples

UTR TR-Abr TR-EtOl

Henderson and Pabis 40 R2 0.9808 0.9951 0.9951RMSE 0.0408 0.0214 0.0218c2 0.001666 0.000698 0.000544

50 R2 0.988 0.9857 0.9992RMSE 0.0306 0.0342 0.0084c2 0.002113 0.001385 0.000070

60 R2 0.9836 0.9948 0.9988RMSE 0.0414 0.0233 0.0108c2 0.002804 0.000555 0.000120

70 R2 0.9882 0.9931 0.9972RMSE 0.0373 0.0236 0.0153c2 0.001483 0.000559 0.000242

Page 40 R2 0.9855 0.9951 0.996RMSE 0.0358 0.0216 0.0199c2 0.001264 0.001078 0.000392

50 R2 0.9964 0.9898 0.999RMSE 0.0167 0.0292 0.0091c2 0.000272 0.000862 0.000087

60 R2 0.9887 0.9951 0.9996RMSE 0.0348 0.0230 0.0067c2 0.006769 0.000509 0.000059

70 R2 0.9972 0.9981 0.9991RMSE 0.0183 0.0126 0.0086c2 0.000741 0.000362 0.000143

Logarithmic 40 R2 0.9975 0.9954 0.9951RMSE 0.0148 0.0210 0.0220c2 0.000218 0.000452 0.000504

50 R2 0.999 0.9989 0.9992RMSE 0.0089 0.0096 0.0076c2 0.010529 0.000368 0.000359

60 R2 0.9973 0.9992 0.9991RMSE 0.0169 0.0095 0.0095c2 0.005950 0.000105 0.000090

70 R2 0.9961 0.9995 0.9985RMSE 0.0217 0.0063 0.0111c2 0.000565 0.000070 0.000166

Two term 40 R2 0.9808 0.9951 0.998RMSE 0.0415 0.0218 0.0142c2 0.001720 0.000724 0.000760

50 R2 0.988 0.9989 0.9992RMSE 0.0309 0.0097 0.0085c2 0.002139 0.000372 0.000072

60 R2 0.9836 0.9992 0.9997RMSE 0.0422 0.0096 0.0053c2 0.002919 0.000109 0.000091

70 R2 0.9882 0.9931 0.9972RMSE 0.0382 0.0243 0.0157c2 0.001559 0.000592 0.000256

G. Adiletta et al. / Journal of Food Engineering 172 (2016) 9e1812

the absorbance of sample at t ¼ 16 min. Experiments were carriedout in triplicate and results obtainedwere reported asDPPH%/mg db.

2.5. Shrinkage and empirical models

The initial grape volume (V0) was calculated by measuring eachsample (about 10 berries) in all three orthogonal axes (length,width and depth) by means of a digital Vernier caliper (0.01 mmaccuracy) and considering spheroidal shape of grape, as proposedby Correa et al. (2011).

At given time during drying experiments for the same berries,we measured the dimensions along three orthogonal axes and wecalculated the volume (Vt).

For the evaluation of shrinkage during drying, the volumeshrinkage (Vt/V0) was reported as a function of the relativemoisturecontent (Mt/M0).

The simplest way to model shrinkage during dehydration is toobtain an empirical correlation between shrinkage and moisture

content. Several examples of empirical models were presented inthe literature and were commonly applied for vegetable food ma-terials (Adiletta et al., 2014; Mayor and Sereno, 2004; Ratti, 1994;Senadeera et al., 2014; Simal et al., 1996).

The mathematical models taken for identifying the most suit-able model to describe the shrinkage behavior were reported inTable 5. Non-linear least square regression analysis was used toevaluate the parameters of the selected model with the Lev-enbergeMarquardt procedure.

2.6. Microstructure analysis

Images of peel surface of fresh and dehydrated samples, withand without pretreatments, were captured using a scanning elec-tron microscope (SIGMA ZEISS, featuring GEMINI® technology).Samples prior the analysis were coated with a thin layer of silver ina sputter coater (AGAR Auto Sputter Coater, mod.108 A, England) for5 min.

G. Adiletta et al. / Journal of Food Engineering 172 (2016) 9e18 13

2.7. Rehydration experiments

Dried samples were rehydrated in distilled water at 25 �C for 5 huntil constant water content was reached. The approximate ratio ofraisins and water volume was kept as 1:30 (Russo et al., 2013).

The rehydration capacity, described as percentage water gain,was calculated from the sample weight difference before and afterthe rehydration as follows:

Weight gain ð%Þ ¼ ðweight of rehydrated samples�weight of dried samplesÞðweight of dried samplesÞ � 100 (5)

The rehydration experiments were repeated in three setsindependently, each set having 3 replicates and the averages werereported with standard deviation

Table 3Parameters of the drying models.

Model name Temperature (�C) Parameters

Henderson and Pabis 40 ak

50 ak

60 ak

70 ak

Page 40 kn

50 kn

60 kn

70 kn

Logarithmic 40 akc

50 akc

60 akc

70 akc

Two term 40 a1k1a2k2

50 a1k1a2k2

60 a1k1a2k2

70 a1k1a2k2

2.8. Statistical analysis

The means and standard deviations of experimental resultswere calculated from three replicates. One-way ANOVA (analysis ofvariance) at the level of significance p < 0.05 using Tukey test wasperformed for comparison of means.

3. Results and discussion

3.1. Drying kinetics: experiments and empirical models

In order to analyze the effect of different drying temperatures

Samples

UTR TR-Abr TR-EtOl

1.0002 0.9676 0.93730.000093 0.000200 0.0003001.0334 0.9832 0.98880.000200 0.001200 0.0005001.0109 1.0099 0.96970.000400 0.002600 0.0012001.0609 1.0407 1.02690.001300 0.001900 0.002400

0.000027 0.000400 0.0011001.1359 0.9201 0.84840.000042 0.002600 0.0006001.1953 0.8871 0.97910.000100 0.003400 0.0020001.1569 0.9484 0.92960.000200 0.000700 0.0013001.2469 1.1594 1.0981

1.5206 0.9895 0.93350.000045 0.000200 0.000300�0.5552 �0.0279 0.00531.3579 0.9691 1.00840.000100 0.001500 0.000500�0.3604 0.0624 �0.02411.3594 0.9902 0.95900.000200 0.000290 0.001300�0.3841 0.0376 0.02151.1922 1.1617 1.05540.000900 0.001400 0.002100�0.1721 �0.1540 �0.0432

0.4985 0.4587 0.12650.000093 0.000200 0.0036000.5018 0.5089 0.87920.000093 0.000200 0.0003000.5171 0.9691 0.48120.000200 0.001500 0.0005000.5163 0.0624 0.50760.000200 2.10 , 10�12 0.0005000.5064 0.9902 0.07550.000400 0.002900 0.0088000.5046 0.0376 0.93210.000400 3.03 , 10�13 0.0012000.5474 0.5263 0.51800.001300 0.001900 0.0024000.5135 0.5144 0.50890.001300 0.001900 0.002400

Table 4Color parameters for fresh, pretreated and untreated dried samples. Values with no letters in common are significantly different (p < 0.05).

Sample L* a* b* H

Fresh 40.62a ± 0.43 20.07a ± 0.20 14.08a ± 0.92 35.08a ± 1.25UTR 40 �C 21.63c ± 1.70 10.42bc ± 0.84 �1.17c ± 1.39 353.58e ± 5.69TR-Abr 40 �C 25.68bc ± 0.38 11.84b ± 1.41 2.84bc ± 0.35 13.49c ± 0.54TR-EtOl 40 �C 22.01c ± 1.70 11.14bc ± 0.61 �0.27c ± 0.42 358.60e ± 2.07UTR 50 �C 27.43b ± 0.08 8.76c ± 0.12 0.02c ± 1.16 0.15e ± 1.81TR-Abr 50 �C 26.14 bc ± 1.07 10.21bc ± 0.38 4.33b ± 1.56 22.99b ± 0.53TR-EtOl 50 �C 24.70bc ± 0.36 10.09bc ± 1.38 2.22bc ± 1.39 12.41c ± 0.83UTR 60 �C 23.02bc ± 3.40 10.10bc ± 0.05 �0.10c ± 2.84 359.46e ± 0.20TR-Abr 60 �C 22.30bc ± 3.67 10.59bc ± 1.08 3.17bc ± 1.83 16.66c ± 2.97TR-EtOl 60 �C 22.85bc ± 0.13 10.43bc ± 0.92 �0.63c ± 1.04 356.54e ± 0.10UTR 70 �C 25.41bc ± 2.68 10.14bc ± 0.94 �0.85c ± 3.10 355.20e ± 1.44TR-Abr 70 �C 24.37bc ± 1.97 11.70b ± 1.45 1.14c ± 1.15 5.57d ± 2.82TR-EtOl 70 �C 24.11bc ± 1.36 10.26bc ± 0.190 �0.79c ± 0.98 355.60e ± 1.12

G. Adiletta et al. / Journal of Food Engineering 172 (2016) 9e1814

(40e70 �C) and pretreatments on the drying kinetics of Red Globegrapes, the curves of moisture ratio Mt/M0 vs drying time (min)were shown in Fig. 1. A constant-rate period was not observed inany of the experiments, hence, the entire drying process fortreated/untreated grapes occurred in the range of the falling-rateperiod. Therefore, diffusion was the dominant physical mecha-nism governing moisture transfer in the samples. Similar resultswere obtained by different authors on drying of various fruits(Ramos et al., 2004; Senadeera et al., 2014). As expected, for all theuntreated and treated samples the moisture content decreasedwith higher rate, increasing the drying temperature. At each tem-perature, the drying rate was higher for both pretreated samples(TR-ABR, TR-EtOl) with respect to the untreated ones. Moreover, at50 and 60 �C the highest rate was found for the abraded grapes. Inparticular, at these temperatures the drying time for raisins withabraded peel was about 1/3 the time required to dry the untreatedgrapes.

In order to estimate the moisture content as a function of dryingtime, the empirical equations, reported in Table 1, were fitted andstatistical parameters were shown in Table 2. The correlation co-efficient (R2) was one of the primary criteria to select the bestmodel to account for variation in the drying curves of dried sam-ples. In addition to R2, the statistical parameters, such as: thereduced c-square (c2) and the rootmean square error (RMSE), wereused to determine the quality of the fit. A good fitting between theexperimental data and the correlations was obtained when therewas a combination of the high R2 value, and the values of c2 andRMSE, which should be as low as possible (Menges and Ertekin,2006;Vega-G�alvez et al., 2008). The values of parameters for eachmodel were reported in Table 3.

From the analysis of the results given in Table 2, the logarithmicmodel was the best model for describing the behavior of abradedgrape at all the temperatures investigated and of untreated grapeexcept at 70 �C, where the Page model gave the highest correlationfactor. On the contrary the two term model was the best fitting forethyl oleate dipped grape excluding 70 �C, where also in this casethe Page model was more satisfactory.

With reference to the logarithmic model, from Table 3 resultedthat the calculated parameters (a, k, c) decrease as the temperatureincreases for untreated grapes. On the contrary, for both treatedsamples the k values showed an opposite trend increasing at highertemperatures.

Comparison between experimental data and best-fitting modelresults were reported in Fig. 1.

Fig. 2. Total phenolics of fresh and dried samples (40, 50, 60 and 70 �C). Bars with noletters in common are significantly different (p < 0.05).

3.2. Color evaluation

Color is one of the most important quality indices of foods and

agricultural products. Unsuitable changes in color of agriculturalproducts would make the food having low quality and marketingvalue. The color of raisins is mostly the result of pigments formedby the effect of the enzymatic and non-enzymatic reactions, takingplace during grape dehydration (Karadeniz et al., 2000). Mean colorvalues of fresh and dried samples are shown in Table 4.

Results showed that drying process has a significant effect onthe color parameters of raisins. All Hunter color parameters (L*, a*,b*) changed significantly after drying. The decrease of L* value in-dicates that the dried samples, which are not significantly different(p < 0.05), were turned darker than fresh grape. Also the parame-ters a* and b* decreased after drying. No significant differences(p < 0.05) in a* values were found between dried samples (exceptfor TR-Abr 40 �C, TR-Abr 70 �C and UTR 50 �C samples). On thecontrary, the b* values were significantly different among driedsamples; in particular, negative values of b* for untreated and ethyloleate treated grapes (except at 50 �C) indicate the bluer color ofthese samples, while positive b* values for abraded samples evi-dence the yellower color. Moreover, the Hue angle values of un-treated and abraded dried samples were significantly differentfrom the fresh sample and from each other; but no significantdifferences (p < 0.05) were observed between untreated andtreated by ethyl oleate dried samples.

The smallest differences of b* and Hue angle values with respectto fresh grape were observed for the sample dried at 50 �C afterabrasive pretreatment.

It can be concluded that, generally, the color of the dried sam-ples is darker than fresh grapes, more for untreated and chemically

Fig. 3. Antioxidant activity of fresh and dried samples (40, 50, 60 and 70 �C). Bars withno letters in common are significantly different (p < 0.05).

G. Adiletta et al. / Journal of Food Engineering 172 (2016) 9e18 15

treated samples than abraded grape berries. This browning in grapeberries, occurring by both enzymatic and non-enzymatic browningreactions, seems to be of lesser extent for abraded samples.

3.3. Total phenolic content and antioxidant activity evaluation

During the drying process the grapes were exposed to hightemperature for a long time which contributed to the loss of

Fig. 4. Experimental data (symbols) and prediction (curves) of volume shrink

antioxidants (Kalt et al., 2000).The results of the effect of each treatment and temperaturewere

summarized in Figs. 2 and 3 for total phenolic content and anti-oxidant activity, respectively. Fresh grape had the highest totalphenolics. The drying of grape, for different times and differenttemperature conditions, induced a decrease in the total phenoliccontent up to 83% for TR-EtOl sample dried at 70 �C. There was nostatistical difference (p < 0.05) in total phenolics between allsamples dried at 40 �C and 50 �C and the sample TR-Abr dried at60 �C. Differences were observed for the samples UTR and TR-EtOldried at 60 �C and all the samples dried at 70 �C. This behaviorseems due to the combined effect of exposure time and tempera-ture: higher temperatures and longer exposure time (for both UTRand TR-EtOl dried grapes) produced more degradation of phenoliccompounds. Based on these results, the best drying temperaturewas 50 �C for both untreated and pretreated samples.

The results of antioxidant activity for all samples were shown inFig. 3. Fresh sample had the highest antioxidant activity equal to1.71 DPPH%/mg db. The scavenging effect of extracts on the DPPHradical decreased as increasing the drying temperature. No statis-tical differences (p < 0.05) were found between the samples TR-Abrat 40 �C, TR-Abr 50 �C and TR-EtOl at 50 �C, which showed thehighest values of DPPH activity.

There was not observed simple quantitative correlation of theantioxidant activity values with the total amount of polyphenols.The lack of relationship in this study is in agreement with otherauthors (Ghasemi et al., 2009; Ruberto et al., 2007; Sengul et al.,2009). In particular Sengul et al. (2009) reported no correlation

age of grape during drying at (a) 40 �C, (b) 50 �C, (c) 60 �C and (d) 70 �C.

Table 5Mathematical models applied to shrinkage curves.

Model name Equation References

Linear Vt/V0 ¼ a1 þ a2 (Mt/M0) Simal et al. (1996)

Quadratic Vt/V0 ¼ a1 þ a2 (Mt/M0) þ a3 (Mt/M0)2 Mayor and Sereno (2004)

Exponential Vt/V0 ¼ a1 exp (k Mt/M0) Mayor and Sereno (2004)

G. Adiletta et al. / Journal of Food Engineering 172 (2016) 9e1816

between total phenolic content and antioxidant capacities of anumber of medicinal plant extracts. No correlation between totalphenolic content and antioxidant capacity can be explained byconsidering that the antioxidant capacity was not solely due to thephenolic contents, but could possibly be due to the presence ofsome other phytochemicals which also contribute to the totalantioxidant capacity.

3.4. Shrinkage and empirical models

Shrinkage of food materials has negative consequences on thequality of the dehydrated product and increasedwith the volume ofremoved water (Adiletta et al., 2014). Macroscopic shrinkage iscommonly referred to the ratio of the volume of the product at agiven drying time to the initial volume. In Fig. 4 the variation of Vt/V0 as a function of the relative moisture content (Mt/M0) was rep-resented. It was generally observed a decrease of grape volumeproportional to the water content decrease during drying. An

Table 6Correlation coefficients (R2, RMSE, c2) of the shrinkage models.

Model name Temperature (�C) Samples

UTR TR-Abr TR-EtOl

Linear 40 R2 0.9397 0.9831 0.9666RMSE 0.0656 0.0370 0.0466c2 0.004298 0.001366 0.002170

50 R2 0.9535 0.8951 0.9838RMSE 0.0542 0.0504 0.0319c2 0.002942 0.002536 0.001018

60 R2 0.9756 0.9574 0.9733RMSE 0.0414 0.0586 0.0381c2 0.001710 0.003437 0.001449

70 R2 0.9868 0.9888 0.994RMSE 0.0381 0.0296 0.0205c2 0.001449 0.000874 0.000421

Quadratic 40 R2 0.978 0.9908 0.9755RMSE 0.0401 0.0277 0.0404c2 0.001605 0.000766 0.001632

50 R2 0.9616 0.9269 0.9877RMSE 0.0496 0.0573 0.0282c2 0.002457 0.003275 0.000798

60 R2 0.9805 0.9844 0.9821RMSE 0.0378 0.0368 0.0318c2 0.001427 0.001351 0.001012

70 R2 0.9927 0.9889 0.996RMSE 0.0295 0.0304 0.0175c2 0.000870 0.000926 0.000307

Exponential 40 R2 0.9739 0.9719 0.966RMSE 0.0431 0.0478 0.0470c2 0.001860 0.002280 0.002211

50 R2 0.9544 0.9202 0.9792RMSE 0.0537 0.0416 0.0361c2 0.002885 0.001732 0.001306

60 R2 0.966 0.9744 0.9771RMSE 0.0485 0.0454 0.0353c2 0.002354 0.002063 0.001243

70 R2 0.9874 0.9546 0.9713RMSE 0.0371 0.0596 0.0450c2 0.001375 0.003553 0.002024

exceptional behavior was observed at 50 �C where the abrasivepretreatment produced less shrunk raisins (Vt/V0 equal to 0.38)with respect to both untreated (Vt/V0 equal to 0.20) and driedgrapes by ethyl oleate dipping (Vt/V0 equal to 0.15).

For the empirical models which correlate the shrinkage andmoisture content, the statistical parameters (R2, RMSE and c2) werereported in Table 6. The values of parameters for different modelswere reported in Table 7.

Quadratic model showed an acceptable fit to experimental datafor all the samples and temperatures investigated, being theexponential model the one leading to larger deviation betweenexperimental and predicted values, in agreement with Mayor andSereno (2004). The comparison between experimental data andquadratic model results was reported in Fig. 4.

3.5. Microstructure evaluation

SEM analysis was used to examine the peel structure of un-treated, dipped in ethyl oleate solution and abraded grapes beforedrying (Fig. 5aec).

Fig. 5a showed the typical surface wax of untreated fresh grape.On the peel of the abraded grape the waxy layer was almostcompletely removed in a quite uniform way (Fig. 5b). As conse-quence of waxy solubilization by ethyl oleate, a non-uniform dis-tribution of the waxy components on the grape surface resulted

Table 7Parameters of the shrinkage models.

Model name Temperature (�C) Parameters Samples

UTR TR-Abr TR-EtOl

Linear 40 a1 �0.0370 0.0289 0.0970a2 0.9555 0.9338 0.8082

50 a1 0.1143 0.3160 0.0598a2 0.8135 0.4731 0.8277

60 a1 0.0492 0.1280 0.1236a2 0.8311 0.8146 0.7271

70 a1 0.1462 0.1740 0.1113a2 0.8156 0.8646 0.8396

Quadratic 40 a1 0.1376 0.0900 0.1395a2 0.0137 0.0588 0.0512a3 0.7268 0.3303 0.3067

50 a1 0.1825 0.4232 0.1079a2 0.0446 �0.0377 0.0554a3 0.3415 0.9537 0.2546

60 a1 0.0970 0.1767 0.1591a2 0.0572 0.0340 0.0418a3 0.2407 0.5231 0.3348

70 a1 0.1831 0.1777 0.1327a2 0.0424 0.0839 0.0696a3 0.4064 0.0268 0.1466

Exponential 40 a1 0.0149 0.0176 0.0189k 1.9334 1.8036 1.6842

50 a1 0.0225 0.0337 0.0172k 1.495 0.9077 1.7231

60 a1 0.0178 0.0196 0.0184k 1.682 1.6994 1.6609

70 a1 0.0206 0.0271 0.0212k 1.6329 1.4209 1.5942

Fig. 5. SEM images of peel for: (a) fresh untreated sample, (b) fresh abraded sample, (c) fresh sample dipped in ethyl oleate solution and for dried sample at 50 �C: (d) pretreated byethyl oleate (e) untreated and (f) pretreated by abrasion.

G. Adiletta et al. / Journal of Food Engineering 172 (2016) 9e18 17

(Fig. 5c). After drying the presence of cracks on the peel surface ofethyl oleate pretreated raisins appeared (Fig. 5d) which are notevident for those untreated (Fig. 5e) and pretreated by abrasion(Fig. 5f). These results are in agreement with previous works byVazquez et al. (1997), Riva et al. (1986), Saravacos et al. (1988).These researchers analyzing the effect of chemical pretreatmentsreported thatmicro-fissures in the grape peel were formed by usingcarbonate solution and ethyl oleate solution. Moreover, theyconcluded that the increased drying rate by chemical pre-treatments was due to the formation of micropores on the surfaceof peel.

Fig. 6. Rehydration kinetics of untreated and treated samples dried at 50 �C.

3.6. Rehydration kinetics

Rehydration is a complex phenomenon affected by numerousfactors (e.g., pretreatments, drying method and process, physicalstructure, chemical composition). Therefore, it is expected that theporous microstructure will play important role in the rehydrationmechanism (Russo et al., 2013).

Samples previously dried over longer periods showed lesserrehydration, indicating the presence of modified structures.

Typical rehydration kinetics for raisin were reported in Fig. 6.

During the absorption of water, the increase of weight with timewas almost linear (R2 ¼ 0.9831) for untreated samples, while alogarithmic rise of weight to a maximum value was observed fordried samples obtained by abrasion (R2 ¼ 0.9952) and by dipping in

G. Adiletta et al. / Journal of Food Engineering 172 (2016) 9e1818

ethyl oleate solution (R2 ¼ 0.9926).Moreover, untreated samples showed lesser rehydration (33%)

after longer drying periods, indicating the presence of shrunkenand closed structures that obstacle the absorption of water inagreement with shrinkage results (Fig. 5b). On the contrary theraisins by ethyl oleate dipping showed the highest weight gain(334%) due to the presence of crack on the peel surface, as observedby SEM images (Fig. 5d).

4. Conclusions

The effect of abrasive pretreatment on drying rate and quality ofraisins was studied and compared with chemical pretreatment inalkaline ethyl oleate solution. Grapes obtained by abrasion methodand with ethyl oleate solution showed shorter drying times (up to1/3) than untreated samples.

Furthermore, peel abrasion combined with optimal dryingtemperature, which was 50 �C, was not only effective in reducingthe drying time, but also in improving the rehydration capability,preserving the antioxidant activity and reducing the shrinkage ofgrapes. Differences in quality properties for raisins obtained byethyl oleate pretreatment seem due to the formation of cracks onthe grape peel during drying.

Among the common used empirical models, the logarithmicmodel was the best fitting model for describing the drying kineticsof abraded grape at all the temperatures investigated except at70 �C, where the Page model gave the highest correlation factor.While for the shrinkage, quadratic model showed an acceptable fitto experimental data for all the samples and temperaturesinvestigated.

References

Adiletta, G., Iannone, G., Russo, P., Patimo, G., De Pasquale, S., Di Matteo, M., 2014.Moisture migration by magnetic resonance imaging during eggplant drying: apreliminary study. Int. J. Food Sci. Technol. 49, 2602e2609. http://dx.doi.org/10.1111/ijfs.12591.

Berna, A., Rossell�o, C., Mulet, A., 1991. Drying kinetics of Majorcan seedless grapevariety. Technol. Today 3, 134e137.

Bingol, G., Roberts, J.S., Balaban, M.O., Devres, Y.O., 2012. Effect of dipping tem-perature and dipping time on drying rate and color change of grapes. DryingTechnol. 30, 597e606.

Brasiello, A., Adiletta, G., Russo, P., Crescitelli, S., Albanese, D., Di Matteo, M., 2013.Mathematical modeling of eggplant drying: shrinkage effect. J. Food Eng. 114,99e105.

Carranza-Concha, J., Benlloch, M., Camacho, M.M., Martínez-Navarrete, N., 2012.Effects of drying and pretreatment on the nutritional and functional quality ofraisins. Food Bioprod. Process. 90, 243e248.

Cinquanta, L., Di Matteo, M., Estia, M., 2002. Physical pre-treatment of plums(Prunus domestica). Part 2. Effect on the quality characteristics of differentprune cultivars. Food Chem. 79, 233e238.

Correa, P.C., Resende, O., Garin, S.A., Jaren, C., De Oliveira, G.H.H., 2011. Mathematicalmodels to describe the volumetric shrinkage rate of red beans during drying.Eng. Agríc. Jaboticabal 31 (4), 716e726.

Di Matteo, M., Cinquanta, L., Galiero, G., Crescitelli, S., 2000. Effect of a novelphysical treatment process on the drying kinetics of seedless grapes. J. FoodEng. 46, 83e89.

Di Matteo, M., Cinquanta, L., Galiero, G., Crescitelli, S., 2002. Physical pre-treatmentof plums (Prunus domestica). Part 1. Modelling the kinetics of drying. FoodChem. 79, 227e232.

Doymaz, I., 2004. Effect of dipping treatment on air drying of plums. J. Food Eng. 64(4), 465e470.

Doymaz, I., 2005. Drying characteristics and kinetics of okra. J. Food Eng. 69,275e279.

Doymaz, I., 2006. Drying kinetics of black grapes treated with different solutions.J. Food Eng. 76, 212e217.

Doymaz, I., Pala, M., 2002. The effects of dipping pretreatments on air-drying ratesof the seedless grapes. J. Food Eng. 52, 413e417.

Esmaiili, M., Sotudeh-Gharebagh, R., Mousavi, M.A.E., Rezazadeh, G., 2007. Grapedrying: a review. Food Rev. Int. 23, 257e280.

Esmaiili, M., Sotudeh-Gharebagh, R., Mousavi, M.A.E., Rezazadeh, G., 2007. Influence

of dipping on thin-layer drying characteristics of seedless grapes. Biosyst. Eng.98, 411e421.

Femenia, A., S�anchez, E.S., Simal, S., Rossell�o, C., 1998. Effects of drying pre-treatments on the cell wall composition of grape tissues. J. Agric. Food Chem.46, 271e276.

Ghasemi, K., Ghasemi, Y., Ebrahimzadeh, A.M., 2009. Antioxidant activity, phenoland flavonoid contents of 13 citrus species peels and tissues. Pak. J. Pharm. Sci.22 (3), 277e281.

Henderson, S.M., 1974. Progress in developing the thin layer drying equation. Trans.ASAE 17, 1167e1168.

Henderson, S.M., Pabis, S., 1961. Grain drying theory I: temperature effect on dryingcoefficient. J. Agric. Eng. Res. 6, 169e174.

Hou, D.X., 2003. Potential mechanisms of cancer chemoprevention by anthocya-nins. Curr. Mol. Med. 3 (2), 149e159.

Kalt, W., McDonald, J.E., Donner, H., 2000. Anthocyanins, phenolics, and antioxidantcapacity of processed lowbush blueberry product. J. Food Sci. 65, 390e393.

Karadeniz, F., Durst, R.W., Wrolstad, R.E., 2000. Polyphenolic composition of raisin.J. Agric. Food Chem. 48, 5343e5350.

Kashaninejad, M., Tabil, L.G., 2004. Drying characteristics of purslane (Portulacaoleraceae L.). Drying Technol. 22, 2183e2200.

Katalinic, V., Milos, M., Kulisic, T., Jukic, M., 2006. Screening of 70 medicinal plantextracts for antioxidant capacity and total phenols. Food Chem. 94, 550e557.

Khalloufi, S., Almeida-Rivera, C., Bongers, P., 2009. A theoretical model and itsexperimental validation to predict the porosity as a function of shrinkage andcollapse phenomena during drying. Food Res. Int. 42 (8), 1122e1130.

Mayor, L., Sereno, A.M., 2004. Modelling shrinkage during convective drying of foodmaterials: a review. J. Food Eng. 61, 373e386.

Menges, H.O., Ertekin, C., 2006. Modelling of air drying of Hacihaliloglu-typeapricots. J. Sci. Food Agric. 86, 279e291.

Pangavhane, D.R., Sawhney, R.L., 2002. Review of research and development workon solar dryers for grape drying. Energy Convers. Manage. 43, 45e61.

Park, K.J., Vohnikova, Z., Brod, F.P.R., 2002. Evaluation of drying parameters anddesorption isotherms of garden mint leaves (Mentha crispa L.). J. Food Eng. 51,193e199.

Ramos, N., Silva, C.L.M.C.L.M., Sereno, A.M., Aguilera, J.M., 2004. Quantification ofmicrostructural changes during first stage air drying of grape tissue. J. Food Eng.62, 159e164.

Ratti, C., 1994. Shrinkage during drying of foodstuffs. J. Food Eng. 23, 91e105.Riva, M., Peri, C., Lovino, R., 1986. Effects of pretreatments on kinetics of grapes

drying. In: Le Maguer, M., Jelen, P. (Eds.), Food Engineering and Process Ap-plications, Transport Phenomena, vol. 1. Elsevier Applied Science, London,pp. 461e472.

Ruberto, G., Renda, A., Daquino, C., Amico, V., Spatafora, C., Tringali, C., DeTommasi, N., 2007. Polyphenol constituents and antioxidant activity of grapepomace extracts from five Sicilian red grape cultivars. Food Chem. 100,203e210.

Russo, P., Adiletta, G., Di Matteo, M., 2013. The influence of drying air temperatureon the physical properties of dried and rehydrated eggplant. Food Bioprod.Process. 91, 249e256.

Saravacos, G.D., Marousis, S.N., Raouzeos, G.S., 1988. Effect of ethyl oleate on the rateof airdrying of foods. J. Food Eng. 7, 263e270.

Senadeera, W., Adiletta, G., Di Matteo, M., Russo, P., 2014. Drying kinetics, qualitychanges and shrinkage of two grape varieties of Italy. Appl. Mech. Mater. 553,362e366.

Sengul, M., Yildiz, H., Gungo, N., Cetin, B., Eser, Z., et al., 2009. Total phenolic con-tent, antioxidant and antimicrobial activities of some medicinal plants. Pak. J.Pharm. Sci. 22 (1), 102e106.

Simal, S., Mulet, A., Catal_a, P.J., Canellas, J., Rossell�o, C., 1996. Moving boundarymodel for simulating moisture movement in grapes. J. Food Sci. 61 (1), 157e160.

Singh, S.P., Jairaj, K.S., Srikant, K., 2012. Universal drying rate constant of seedlessgrapes: a review. Renew. Sust. Energy Rev. 16, 6295e6302.

Singleton, V.L., Orthofer, R., Lamuela-Raventos, R.M., 1999. Analysis of total phenolsand other oxidation substrates and antioxidants by means of FolineCiocalteureagent. Methods Enzymol. 299, 152e178.

Vazquez, G., Chenlo, F., Moreira, R., Cruz, E., 1997. Grape drying in a pilot plant witha heat pump. Drying Technol. 15 (3e4), 899e920.

Vega-G�alvez, A., Miranda, M., Bilbao-S�ainz, C., Uribe, E., Lemus-Mondaca, R., 2008.Empirical modeling of drying process for apple (Cv. Granny Smith) slices atdifferent air temperatures. J. Food Process. Pres. 32, 972e986.

Wang, H., Cao, G., Prior, R.L., 1997. Oxygen radical absorbing capacity of anthocya-nins. J. Agric. Food Chem. 45, 304e309.

Yagcioglu, A., Degirmencioglu, A., Cagatay, F., 1999. Drying characteristics of laurelleaves under different drying conditions. In: Proceedings of the Seventh In-ternational Congress on Agricultural Mechanisation and Energy, May 26e27,Adana, Turkey. pp. 565e569.

Yang, J., Meyers, K.J., Heide, J., Liu, R.H., 2004. Varietal differences in phenoliccontent and antioxidant and antiproliferative activities of onions. J. Agric. FoodChem. 52, 6787e6793.

Yang, J., Martinson, T.E., Liu, R.H., 2009. Phytochemical profiles and antioxidantactivities of wine grapes. Food Chem. 116, 332e339.