Journal of Food Engineering 89 (2008) 460465

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier.com/locate/jfoodeng

Drying kinetics of grape seeds

John S. Roberts a,*, David R. Kidd b, Olga Padilla-Zakour b
a
Western Regional Research Center, Agricultural Research Service, United States Department of Agriculture, 800 Buchanan Street, Albany, CA 94710, USA
b
Department of Food Science and Technology, Cornell University, New York State Agricultural Experiment Station, Geneva, NY 14456, USA

a r t i c l e i n f o a b s t r a c t

Article history: Drying of grape seeds representing waste products from white wine processing (Riesling), red wine
Received 31 October 2007 processing (Cab Franc), and juice processing (Concord) was studied at 40, 50, and 60 C and constant
Received in revised form 21 May 2008 air velocity of 1.5 m/s. Equilibrium moisture content had a signicant effect on the normalized drying
Accepted 23 May 2008
curve and was determined for each grape seed at each drying temperature. Effective moisture diffusivity
Available online 6 June 2008
ranged between 1.57 and 3.96  1010 m2/s for Riesling seeds, 2.935.91  1010 m2/s for Concord seeds,
and 3.898.03  1010 m2/s for Cab Franc seeds. The temperature dependence of the effective diffusivity
Keywords:
followed an Arrhenius relationship, and the activation energies were 40.14 kJ/mol for Riesling seeds,
Diffusion
Drying
30.45 kJ/mol for Concord seeds, and 31.47 kJ/mol for Cab Franc seeds. Three thin-layer models were used
Equilibrium moisture content to predict the drying curves: Page model, Lewis model, and the HendersonPabis model. All three models
Grape seeds were found to produce accurate predictions compared to the mass average moisture loss for each grape
Thin-layer models seed variety (percent error less than 10%), and the Lewis model was shown to be an excellent model for
predicting all three grape seed varieties (percent error less than 5%).
Published by Elsevier Ltd.

1. Introduction
In 2005, the US processed over 4 million metric tons (over 80
million metric tons worldwide) of grapes for wine production,
and over 0.5 million metric tons of grapes for juice production
(USDA, 2006). Grape pomace is the solid waste product leftover
from wine and juice processing and generally consists of pulp,
skins, and seeds. Depending on the condition of the grape at har-
vest and the type of press used, 13.514.5% of the grapes crushed
becomes grape pomace waste with extreme waste production as
high as 20% (Russ and Meyer-Pittroff, 2004). Thus in the US alone,
between 56 and 80 thousand metric tons (11.2 million metric tons
worldwide) of pomace was generated as waste in 2005. This high
volume waste product contains valuable components, such as tan-
nins, phytochemicals and oil. Depending on the variety, grape
seeds constitute approximately 26% of the pomace.
One product produced from grape pomace waste is grape seed
oil, which has the following nutritional properties: cholesterol free,
low in saturated fats, contains linoleic acid and high-density lipo-
protein, and rich in Vitamin E and antioxidants (Arvanitoyannis
et al., 2006). Grape seed oil is a high-quality culinary oil having a
high smoke point, 252 C, making it a good choice for frying and
other high temperature food applications. Due to its stability and
uidity properties, grape seed oil spreads and mixes better with
* Corresponding author. Tel.: +1 510 559 5656.
E-mail address: jsrob25@hotmail.com (J.S. Roberts).
food, requiring 50% less than other oils (Arvanitoyannis et al.,
2006).
Fresh grape seeds are highly perishable, and dehydration is a
useful means to increase the shelf-life of seeds for further use. Cold
mechanical pressing of dried seeds is an environmentally friendly
method to extract oil where no chemicals are used. Dry seeds with
moisture content below 0.10 g/g dry solids are required for a cold
press operation, thus dehydration is a critical processing step.
There are a limited number of reported studies on drying kinetics
of seeds: tomato seeds (Sogi et al., 2003), pumpkin seeds (Sacilik,
2007), and rapeseeds (Pathak et al., 1991). To these authors knowl-
edge, drying kinetic studies on grape seeds have not been reported.
The objective of this research is to quantify the drying kinetics
of grape seeds to facilitate production of valuable grape seed oil
from high volume waste pomace.
2. Materials and methods
2.1. Sample preparation
Grape seeds used in drying experiments were provided from
grapes Vitis vinifera of the variety Riesling and Cabernet Franc, rep-
resenting waste materials from white wine and red wine process-
ing, respectively, and grape seeds from Vitis labrusca of the variety
Concord, representing waste material from grape juice processing.
Pomace was obtained from New York wineries and grape juice
manufacturers, and the seeds were separated using a very efcient
separator (turbo nisher/extractor, Bertocchi Inc., ITALY). Riesling

0260-8774/$ - see front matter Published by Elsevier Ltd.

doi:10.1016/j.jfoodeng.2008.05.030
 J.S. Roberts et al. / Journal of Food Engineering 89 (2008) 460465
Nomenclature
A intercept constant in the HendersonPabis model
Deff effective moisture diffusivity, m2/s
D0 pre-exponential factor of the Arrhenius equation, min1
K drying rate constant, min1
Ea activation energy, kJ/mol
M moisture content (dry basis), kg H2O/kg dry solid
Mi initial moisture content (dry basis), kg H2O/kg dry solid
seeds were the smallest seeds of the three varieties, and a 6 mm
screen was used to separate the seeds from the pomace. An
8 mm screen was used to separate the seeds from the Cabernet
Franc and Concord pomace. After bulk separation, the seeds were
washed to remove the remaining pomace. Initial moisture contents
of all samples were determined by the oven drying method at 70 C
in a vacuum oven (Model 3608, Lab-Line Instrument, Inc.) for 24 h
(AOAC, 1990).
2.2. Convective hot air drying
The schematic diagram of the apparatus used for convective
hot air drying experiments is shown in Fig. 1. The basic design
of the convective drying apparatus is described by Roberts et
al. (2002). Samples were suspended beneath an analytical bal-
ance (AG204, Mettler Toledo, Inc., Hightstown, NJ) into the oven
and within a polycarbonate air-ow cylinder using a fabricated
mesh basket. The balance was interfaced to a PC computer by a
RS-232 cable, and the weight loss of the samples was recorded
on-line every 30 s throughout drying using software for the bal-
ance (Balance Link, Mettler Toledo, Inc., Hightstown, NJ). High
drying temperatures used in drying seeds prior to oil extraction
may have deleterious effects on the nal oil quality (Gomes
and Caponio, 2001); however, little is known about the extent
of thermal damage caused by the drying process prior to oil
extraction (Ggs and Maskan, 2006). Thus, the temperatures
used in this study, 40, 50, and 60 C, were in the range of the
RS-232 Cable
Air-Flow:
1.5 m/s
Fig. 1. Convective hot-air drying apparatus.
461
Me equilibrium moisture content (dry basis), kg H2O/kg dry
solid
N constant in the Page model
Rc universal gas constant, = 8.314  103 kJ/mol K
T absolute temperature, K
t time, s
low end of typical drying temperature. External resistances to
mass transfer become important for a Biot number for mass
transfer, Bim = kL/Dqs less than 10 (Vaccarezza et al., 1974). In
addition, experimental studies of the effect of air velocity on
the moisture content of the material during drying can be used
to determine whether external resistances are important to mass
transfer (Vaccarezza et al., 1974). When convective hot air is
equal to or greater than 1 m/s, external resistances become neg-
ligible (Jason, 1958). Srikiatden and Roberts (2006) showed that
the Biot number for mass transfer for 0.7 cm samples was in
the range of 1.44  1049.70  104 using an air velocity above
1.5 and air temperatures between 50 and 70 C. This Biot number
for mass transfer is far above 10 below which external resistance
to mass transfer occurs. Thus, the air velocity used in this study
was 1.5 m/s with the assumption that the Biot number for mass
transfer for the seed samples will also be above 10 insuring neg-
ligible external resistance to mass transfer. Moisture contents at
each time interval were calculated from both weight loss data
and dry solid weight of the sample, which was determined at
the end of drying by the vacuum oven-drying method at 70 C
for 24 h (AOAC, 1990).
2.3. Drying analysis
Simplied drying models have been used to quantify drying
kinetics of various grains and some seeds (Sogi et al., 2003; Parti,
1993; Bruce, 1985). Three common models are the Page model:
Analytical
Balance
Support Table
Oven
Air Flow
Cylinder:
0.088 m O.D.,
0.082 m I.D.,
0.60 m height
Mesh Sample Holder:
462 J.S. Roberts et al. / Journal of Food Engineering 89 (2008) 460465

M  Me N or due to uctuating relative humidity during drying, the equilib-

MR expkt
Mi  Me
the Lewis model:
M  Me
MR expkt
Mi  Me
and the HendersonPabis model:
M  Me
MR a expkt
Mi  Me
where MR represents the unaccomplished moisture content or
moisture ratio, M is the moisture at any time t during drying, Mi
is the initial moisture content, Me is the equilibrium moisture con-
tent, k is the drying rate constant, and N and a are constants.
The empirical constants for the thin-layer drying models were
determined experimentally from normalized drying curves at dif-
ferent temperatures, which were evaluated based on coefcient
of determination (R2). The form of the normalized Page equation is
ln lnMR lnk N lnt 4
where the drying constants, k and N, are determined from the inter-
cept and slope of the ln(ln (MR)) vs ln (t) curve, respectively. The
form of the normalized Lewis equation is
InMR kt 1
and of the HendersonPabis equation is
InMR kt a
where the drying constants k and a are determined from the slope
and intercept, respectively, of the ln (MR) vs time curve. For the Le-
wis equation, the intercept is set equal to 1. The equilibrium mois-
ture content (Me) was obtained by extending the drying time until
no measurable weight loss was observed. The air temperature and
humidity were measured by a temperature/humidity logger (Digi-
sense 91090-00, Cole-Parmer Instrument Co.).
The goodness of t for each model was evaluated based on root
mean square error (RMSE), chi square (v2), and relative percent er-
ror (PE). The predicted moisture ratio was compared to the exper-
imental moisture ratio using root mean square error and chi square
as shown in the following equations (McMinn, 2006):
" #1
1X n
  2
RMSE MRexp;i  MRpredict;i
n i1
1
rium moisture content can be assumed to be 0 g/g dry solid (Saci-
lik, 2007; McMinn, 2006; Doymaz and Pala, 2003; Midilli et al.,
2002; Yaldiz and Ertekin, 2001; Diamente and Munro, 1993). This
assumption is valid only at the beginning of drying, but as the sam-
2
ple dries the moisture content approaches the equilibrium mois-
ture content. This can signicantly affect the slope and linearity
of the normalized drying curve, as shown in Fig. 2. Also, if the dry-
ing curve is not allowed to continue until it reaches equilibrium
3
and thus the equilibrium moisture content is measured to be too
high, the slope is shown to be greater and departs from linearity.
As the unaccomplished moisture content reduces by more than
90% (or one log-cycle), the slope of the semi-log plot is particularly
sensitive to changes or deviations during the second log reduction.
Thus, if the equilibrium moisture content is measured accurately,
the slope of the semi-log unaccomplished moisture content graph
will be linear throughout drying. Table 1 shows the equilibrium
moisture content of each grape seed variety at each temperature
and equilibrium relative humidity condition. Initial moisture con-
tent for each grape seed variety is also provided in Table 1 and
ranges between 0.324 and 0.387 g/g dry-solid (d.s.), which is with-
in the low-end range reported for grape seeds by Rice (1976).
3.2. Effective moisture diffusivity and activation energy
5
Drying of most food materials occurs in the falling rate period
(Wang and Brennan, 1992), and moisture transfer during drying
6
is controlled by internal diffusion (Saravacos and Charm, 1962).
Ficks second law of diffusion, as shown in Eq. (1), has been widely
used to describe the drying process during the falling rate period
for most biological materials (Saravacos and Charm, 1962; Sablani
et al., 2000; Saravacos and Maroulis, 2001) as follows:
1
Minf = 0
7
MR

0.1

1 Xn
v2 MRexp;i  MRpredict;i 2 8
Nn i1

Minf = 0.1
As RMSE and v2 approach zero, the closer the prediction is to Minf = 0.068
experimental data. Where RMSE and v2 compare the differences
between the predicted moisture ratios to the experimental mois-
ture ratios, relative percent error compares the absolute differ- 0.01
ences between the predicted moisture contents with the 0 2000 4000 6000 8000 10000 12000 14000 16000
experimental moisture contents throughout drying (McLaughlin Time (sec.)
and Magee, 1998; zdemir and Devres, 1999):
Fig. 2. The effect of equilibrium moisture content on the normalize plot of
100 Xn jM exp;i  M predict;i j unaccomplished moisture content.
PE % 9
n i1 Mexp;i
The relative percent errors below 10% indicate good t (McLaughlin
and Magee, 1998). Table 1
Equilibrium moisture content

3. Results and discussion Grape seed Mi (g/g d.s.) Me (g/g d.s.)

variety
40 C, 12% 50 C, 7.5% 60 C, 4.5%
3.1. Equilibrium moisture content ERH ERH ERH
Riesling 0.344 0.024 0.102 0.005 0.078 0.006 0.061 0.006
Several papers have suggested that since the equilibrium mois- Cab Franc 0.387 0.022 0.074 0.008 0.054 0.004 0.040 0.001
Concord 0.324 0.024 0.074 0.001 0.057 0.001 0.051 0.006
ture content is signicantly less than the initial moisture content
 J.S. Roberts et al. / Journal of Food Engineering 89 (2008) 460465 463

Table 2 where NFi is the Fick number ((Defft)/r2). For long drying time when
Effective diffusivities and corresponding activation energies for the three grape seed the Fick number, Deff t/r2, is greater than 0.1 or MR is less than 0.6,
varieties
the rst term of the series solution dominates (Crank, 1975; Rizvi,
Grape seed variety Temp (C) Deff (m2/s)  1010 Ea (kJ/mol) R2 1986):
Riesling 40 1.57 40.14 0.9999  
5:783Deff
50 2.51 InMR 0:692  t 13
60 3.96 R2
Cab Franc 40 3.89 31.47 0.9916 The effective moisture diffusivity can be determined from the
50 5.33 slope of the normalized plot of the unaccomplished moisture ratio,
60 8.03
ln (MR) vs time, using the following equation:
Concord 40 2.93 30.45 0.9999
50 4.21 sloper 2
60 5.91 Deff 14
5:783
The effective moisture diffusivities for each seed variety are
within a range to those reported for hull-less seed pumpkin (Saci-
oM lik, 2007).
rDeff rM 10
ot The effective moisture diffusivity for each seed variety in-
Deff is the effective moisture diffusivity representing the con- creased with increasing temperature. Temperature dependence of
ductive term of all moisture transfer mechanisms. This parameter the effective diffusivity has been shown to follow an Arrhenius
is usually determined from experimental drying curves (Saravacos relationship (Jason, 1958; Saravacos and Charm, 1962; Vaccarezza
and Maroulis, 2001). Based on the assumptions of uniform initial et al., 1974; Suarez et al., 1980):
 
moisture distribution, negligible external resistance, negligible Ea
temperature gradients, negligible shrinkage during drying, and Deff D0 exp  15
RC T
constant diffusion coefcient, the analytical solution of the diffu-
sion equations for innite cylinder is given as the following equa- where Do is the pre-exponential factor of the Arrhenius equation
tion (Crank, 1975): (m2/s), Ea is the activation energy (kJ/mol), Rc is the universal gas
constant (kJ/mol K), and T is the absolute air temperature (K). The
M  M e X1 4 activation energy is determined from the slope of the Arrhenius
MR i1 r 2 a2
expDeff a2n t; J o ran 0 11
Mi  Me n plot, ln (Deff) vs. 1/T. Table 2 shows the activation energy for each
where Jo(x) = 0 is the Bessel function of the rst kind and zero order, seed variety, which ranges between 30.45 and 40.14 kJ/mol and cor-
and ans are the roots of this function. Expansion of Eq. (11) of the responds to reported activation energies for food materials (Sarav-
rst three terms is as follows: acos and Maroulis, 2001) and similar to the activation energy
reported for hull-less seed pumpkin, which was 33.15 kJ/mol (Saci-
MR 0:692e5:78NFi 0:131e30:5NFi 0:0534e74:9NFi 12 lik, 2007). Thus, the temperature dependence of the three grape

Table 3
Empirical constants of the Page, Lewis, HendersonPabis equations

Grape seed variety Temp (C) Page equation Lewis equation HendersonPabis equation
k (min1) N R2 k (min1) R2 k (min1) a R2
4 4 4
Riesling 40 7.42  10 0.8328 0.9999 1.42  10 0.9901 1.45  10 1.0271 0.9909
50 9.14  104 0.8628 0.9794 2.83  104 0.9938 2.79  104 0.9726 0.9942
60 1.13  103 0.8900 0.9830 4.24  104 0.9971 4.24  104 1.0504 0.9971
Cab Franc 40 2.80  103 0.7062 0.9603 2.23  104 0.9857 2.19  104 0.9588 0.9864
50 3.42  103 0.7149 0.9829 3.06  104 0.9848 3.10  104 1.0364 0.9874
60 4.11  103 0.7260 0.9946 4.27  104 0.9984 4.26  104 0.9959 0.9990
Concord 40 1.28  103 0.7996 0.9916 2.13  104 0.9898 2.01  104 0.9015 0.9960
50 2.38  103 0.7412 0.9918 2.47  104 0.9955 2.39  104 0.9411 0.9975
60 3.90  103 0.7101 0.9876 3.30  104 0.9846 3.21  104 0.9377 0.9860

Table 4
Model prediction evaluation

Grape seed variety Temp (C) Page equation Lewis equation HendersonPabis equation
RMSE X2 (102) PE RMSE X2 (102) PE RMSE X2 (102) PE
Riesling 40 0.0730 0.295 7.16 0.0236 0.058 2.35 0.0256 0.069 2.10
50 0.0393 0.196 5.86 0.0355 0.167 4.28 0.0385 0.188 4.44
60 0.0417 0.188 5.43 0.0217 0.055 2.97 0.0222 0.057 4.64
Cab Franc 40 0.0528 0.298 7.32 0.0355 0.134 2.88 0.0364 0.140 3.13
50 0.0664 0.539 10.49 0.0293 0.106 4.28 0.0308 0.116 5.78
60 0.0387 0.213 6.56 0.0205 0.047 4.37 0.0212 0.050 4.70
Concord 40 0.0194 0.041 2.79 0.0452 0.219 4.09 0.0550 0.324 2.65
50 0.0338 0.148 3.41 0.0224 0.055 3.14 0.0292 0.098 2.56
60 0.0346 0.188 3.00 0.0395 0.195 4.98 0.0449 0.240 4.48
464 J.S. Roberts et al. / Journal of Food Engineering 89 (2008) 460465

seed varieties can be represented by the following Concord:

equations:Riesling:  
3662:16
  Deff 3:519  105 exp  18
4828:51 T
Deff 7:79  105 exp  16
T
Cab France: 3.3. Thin-layer drying models
 
3785:11
Deff 6:77  105 exp  17 The empirical drying constants for each grape seed variety at
T
each temperature are given in Table 3. The coefcients of determi-
nation (R2) for the drying rate constants of all three thin-layer dry-
ing models for three grape varieties were above 0.96 for all values,
with all but two being above 0.98. These high coefcients of deter-
a 0.4 mination are due to the highly linear plots of the unaccomplished
0.35
40 C moisture content, which are perhaps due to accurate equilibrium
Riesling 50 C moisture contents. The drying rate constants for all three thin-layer
Moisture Content (g/g d.s.)

0.3 60 C models and for all three grape seed varieties were similar to those of
hull-less seed pumpkin (Sacilik, 2007), and the drying rate con-
Lewis Equation
0.25 stants for all three grape seed varieties in Pages model were similar
to those of tomato seeds (Sogi et al., 2003). The drying constants for
0.2
the Lewis and HendersonPabis models were similar, which can be
0.15 explained by the intercept constants for the HendersonPabis mod-
el being close to 1. An interesting physical characteristic compari-
0.1 son of all three seed varieties was the obvious softness of the Cab
Franc seeds, perhaps as a result of being in a fermentation environ-
0.05 ment. Also, the drying rate constant and the effective diffusivity val-
ues for Cab Franc seeds were the highest. Extraction of
0
0 50 100 150 200 250 300 phytochemicals from seeds occurs during fermentation of red
Time (min.) wines (Kovac et al., 1992), and after 57 days of fermentation, seed
coat cuticle is dissolved by ethanol produced and extraction of tan-
b 0.4 nins predominates (Zoecklein, 2006). This removal of the cuticle
during fermentation may result in softening of the seeds and allow
40 C
0.35 moisture to be removed at a greater rate during drying. In contrast,
Cab Franc
Moisture Content (g/g d.s.)

50 C Concord seeds were from grape juice processing and do not under-
0.3 60 C go fermentation, and Riesling seeds are from white wine processing
Lewis Equation where the skins and seeds are removed prior to fermentation (Boul-
0.25
ton et al., 1999). Therefore, Concord and Riesling seeds were not ex-
0.2 posed to ethanol, so their cuticles remained intact resulting in their
rate of moisture loss being less than that of the Cab France seeds.
0.15 Table 4 shows the results of the RMSE, v2, and PE for the three
thin-layer drying models for each seed variety. Overall, the three
0.1
models provided good predictions, where the percent error ranged
0.05
between 4.98% and 2.35% using the Lewis model, between 5.78%
and 2.10% using the HendersonPabis, and between 10.49% and
0 2.79% using the Page model. Fig. 3 shows the drying curves of each
0 50 100 150 200 250 300 grape seed variety at each temperature and the corresponding Le-
Time (min.) wis Model prediction. For drying predictions of Riesling and Cab
Franc seeds, the Lewis model was a better t based on moisture ra-
c 0.4 tio goodness of t analysis (lower RMSE and v2) and moisture con-
tent t analysis (lower PE). For Concord seeds, the Page model was
0.35 40 C
Concord a better t model, though all three thin-layer models had an error
Moisture Content (g/g d.s.)

0.3 50 C below 5%.

60 C
0.25
Lewis Equation 4. Conclusion
0.2
This study showed that the drying of grape seeds can be accu-
0.15
rately predicted using the thin-layer models of either the Page, Le-
0.1 wis, or HendersonPabis models. The Lewis model proved to be an
overall better prediction model for drying of each grape seed vari-
0.05 ety. The moisture transfer can be described by diffusion, where the
effective diffusivities were within ranges of other seeds, and the
0
temperature dependence of the effective moisture diffusivities
0 50 100 150 200 250 300
was shown to follow an Arrhenius relationship. This study also
Time (min.) showed the importance of the equilibrium moisture content on
Fig. 3. Drying curves at 40, 50, and 60 C for: (a) Riesling seeds, (b) Cab Franc seeds, the slope and linearity of the normalized unaccomplished moisture
and (c) Concord seeds. content curve in determining the drying rate constants.
 J.S. Roberts et al. / Journal of Food Engineering 89 (2008) 460465
Drying grape seeds is an essential in both stabilizing the seeds
for preventing microbial and chemical degradation and producing
the seeds in a form suitable for oil extraction. Seeds from grape and
wine processing plants are contained within waste pomace. From a
practical standpoint, drying the pomace and the seed together may
not be a promising practice due to the extensive energy require-
ment to remove the added signicant moisture load from the pom-
ace. Also, from preliminary experience, it is much more difcult to
separate dry pomace from dry seeds particularly since the ratio of
pomace to seeds is very high and re-wetting was required. By sep-
arating the pomace rst, drying of the seeds is preferred where
separating a small amount of dry pomace, or chaffe, is much easier.
This study provides information to dry high volume and highly
valuable seed waste from wineries and grape juice manufactures
to produce grape seed oil.
References
AOAC, 1990. Ofcial Methods of Analysis, 15th ed., vol. 2. Association of Ofcial
Analytical Chemists, Inc., Arlington.
Arvanitoyannis, I.S., Ladas, D., Mavromatis, A., 2006. Potential uses and applications
of treated wine waste: a review. International Journal of Food Science and
Technology 41, 475487.
Boulton, R.B., Singleton, V.L., Bisson, L.F., Kunkee, R.E., 1999. Principles and Practices
of Winemaking. Kluwer Academic/Plenum Publishers, New York.
Bruce, D.M., 1985. Exposed-layer barley drying, three models tted to new data up
to 150 C. Journal of Agricultural Engineering Research 32, 337347.
Crank, J., 1975. The Mathematics of Diffusion, Second ed. Oxford University Press,
London, UK.
Diamente, L.M., Munro, P.A., 1993. Mathematical modeling of the thin layer solar
drying of sweet potato slices. Solar Energy 5 (4), 271276.
Doymaz, I., Pala, M., 2003. The thin-layer drying characteristics of corn. Journal of
Food Engineering 60, 125130.
Ggs, F., Maskan, M., 2006. Air drying characteristics of solid waste (pomace) of
olive oil processing. Journal of Food Engineering 72, 378382.
Gomes, T., Caponio, F., 2001. Possibility of improving the quality characteristics of
olive-pomace oil and enhancing its differentiation from rened olive-pomace
oil. Journal of the Science of Food and Agriculture 1, 6267.
Jason, A.C., 1958. A study of evaporation and diffusion processes in the drying of sh
muscle. In: Fundamental Aspects of Dehydration of Foodstuffs. Soc. Chem. Ind.,
London and MacMillan Co., New York, pp. 103135.
Kovac, V., Alonso, E., Bourzeix, M., Revilla, E., 1992. Effect of several enological
practices on the content of catechins and proanthocyanidines of red wines.
Journal of Agricultural and Food Chemistry 40, 19531957.
McLaughlin, C.P., Magee, T.R.A., 1998. The determination of sorption isotherm and
the isosteric heats of sorption for potatoes. Journal of Food Engineering 35, 267
280.
465
McMinn, W.A.M., 2006. Thin-layer modeling of the convective, microwave,
microwave-convective and microwave-vacuum drying of lactose powder.
Journal of Food Engineering 72, 113123.
Midilli, A., Kucuk, H., Yapar, Z., 2002. A new model for single-layer drying. Drying
Technology 20 (7), 15031513.
zdemir, M., Devres, Y.O., 1999. The thin layer drying characteristics of hazelnuts
during roasting. Journal of Food Engineering 42, 225233.
Parti, M., 1993. Selection of mathematical models for grain drying in thin-layers.
Journal of Agricultural Engineering Research 54, 339352.
Pathak, P.K., Agrawal, Y.C., Singh, B.P.N., 1991. Thin-layer drying model for rapeseed.
Transactions of American Society of Agricultural Engineers 34, 25052508.
Rice, A.C., 1976. Solid waste generation and by-product recovery potential from
winery residues. American Journal of Viticulture and Enology 27, 2126.
Rizvi, S.S.H., 1986. Thermodynamic properties of foods in dehydration. In: Rao, M.A.,
Rizvi, S.S.H. (Eds.), Engineering Properties of Foods. Marcel Dekker Inc, New
York, pp. 133214.
Roberts, J.S., Tong, C.H., Lund, D.B., 2002. Drying kinetics and timetemperature
distribution of pregelatinized bread. Journal of Food Science 67 (3), 10801087.
Russ, W., Meyer-Pittroff, R., 2004. Utilizing waste products from the food
production and processing industries. Critical Reviews in Food Science and
Nutrition 44, 5762.
Sablani, S., Rahman, S., Al-Habsi, N., 2000. Moisture diffusivity in foods an
overview. In: Mujumdar, A.S. (Ed.), Drying Technology in Agriculture and Food
Sciences. Science Publishers, Inc., Eneld, NH.
Sacilik, K., 2007. Effect of drying methods on thin-layer drying characteristics of
hull-less seed pumpkin (Cucurbita pepo L.). Journal of Food Engineering 79, 23
30.
Saravacos, G.D., Charm, S.E., 1962. Effect of surface-active agents on the dehydration
of fruits and vegetables. Food Technology 16 (1), 9193.
Saravacos, G.D., Maroulis, Z.B., 2001. Transport Properties of Foods. Marcel Dekker
Inc., New York.
Srikiatden, J., Roberts, J.S., 2006. Measuring moisture diffusivity of potato and carrot
(core and cortex) during convective hot air and isothermal drying. Journal of
Food Engineering 74 (1), 143152.
Sogi, D.S., Shivhare, U.S., Garg, S.K., Bawa, A.S., 2003. Water sorption isotherm and
drying characteristics of tomato seeds. Biosystems Engineering 84 (3), 297301.
Suarez, C., Viollaz, P., Chirife, J., 1980. Diffusional analysis of air drying of grain
sorghum. Journal of Food Technology 15, 523531.
USDA Foreign Agricultural Service, 2006. http://fas.usda.gov/agx/ISMG/
GlobalWineReportFinalAug2006.pdf.
Vaccarezza, L.M., Lombardi, J.L., Chirife, J., 1974. Heat transfer effects on drying rate
of food dehydration. Canadian Journal of Chemical Engineering 52, 576579.
Wang, N., Brennan, J.G., 1992. Effect of water binding on the drying behavior of
potato. In: Mujumdar, A.S. (Ed.), Drying, vol. 92. Elsevier Science Publishers B.V,
London, pp. 13501359.
Yaldiz, O., Ertekin, C., 2001. Thin layer solar drying of some vegetables. Drying
Technology 19 (34), 583597.
Zoecklein, B., 2006. Tannins, color and the 2006 season, continued: phenolic
compounds and red wine processing. Enology Notes #117, August 30, 2006.
http://www.vtwines.info.