Journal of Food Engineering 80 (2007) 96–110

www.elsevier.com/locate/jfoodeng

Porous media approaches to studying simultaneous heat

and mass transfer in food processes. II: Property data
and representative results
A.K. Datta *

Department of Biological and Environmental Engineering, Cornell University, 208 Riley-Robb Hall, Ithaca, NY 14853-5701, USA

Received 15 December 2005; received in revised form 27 April 2006; accepted 2 May 2006
Available online 24 July 2006

Abstract

A more general multiphase porous media model is shown to effectively describe a number of heat and mass transfer processes in
foods, particularly processes involving internal evaporation. Results from application of such a model to convective heating, baking
(with and without volume change), frying and microwave heating are included. Limitations of such models in applying to foods include
lack of availability of input parameters to such models and the lack of implementation in commercially available software.
 2006 Elsevier Ltd. All rights reserved.

Keywords: Drying; Frying; Baking; Microwave; Pressure

1. Introduction and overview in the range of parameters such as heating rates and phys-
ical properties. The simplified versions of this general for-
This article follows its companion (Datta, 2006a). It first mulation, presented in the companion paper are more
provides a very short overview of some of the solution readily solved and are not discussed here. Cited equation
options available for the models presented in the compan- numbers, when not from this article, are from the compan-
ion paper. Next, it discusses the types of property data (and ion paper (Datta, 2006a).
some of their literature sources) that would be needed in
the models described. This is followed by specific results 2. Solution techniques
(e.g., temperature and moisture profiles) from a few of
the models as applied to convective and microwave heat- Detailed discussion of solution techniques is not possi-
ing, drying, baking and frying. Discussion here is limited ble due to space limitations. Complexities of heat and mass
to the application of the more general formulation for transfer in food are due to several factors including a
small pores (given by Eqs. (31)–(34) in the companion change in the dominant transport mechanisms during heat-
paper by Datta, 2006a) or its equivalent, without and with ing, properties changing significantly during the heating
deformation. This restriction to one particular model is process and strong coupling between heat and mass trans-
intended to serve two purposes—to conform to the space fer through the evaporation process. Rigorous analytical
limitations of this article and, more importantly, to demon- solutions are only available for very limited cases such as
strate the performance of a general model in simulating dif- the use of the simple diffusion equation (e.g., see Eqs.
ferent food processes that involve variations in physics and (48) or (50) in Datta, 2006a) with constant coefficient and
in regular geometries (Carslaw & Jaeger, 1959; Crank,
*
Tel.: +1 607 255 2482; fax: +1 607 255 4080. 1975). Even when using a simple diffusion equation, when
E-mail address: [email protected]. diffusivity is a function of moisture (which is almost always

0260-8774/$ - see front matter  2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2006.05.012
 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110 97

true when a food is heated, due to evaporation), analytical included in these codes or very few applications have been
solution is mostly unavailable (Crank, 1975). Although reported that use the evaporation and resulting multiphase
there are recent works focusing on either the availability capabilities of those codes. In one commercial software,
of solution to more complex problems or the approxima- PORFLOW (www.acri-in.fr/software/porflow/), for exam-
tion techniques to obtain the final form of the solution ple, although evaporation can be included, solution to
(Banaszak & Kowalski, 1999), numerical solutions are problem formulation (Eq. (31)–(34) in Datta, 2006a) or its
most prevalent. equivalent has not been reported. A public-domain code,
As in other heat and mass transfer studies, common dis- TOUGH2, developed at the Lawrence Berkeley laboratory
cretization methods are used in heat and mass transfer in of the US (www-esd.lbl.gov/TOUGH2/tough2v2.html) and
food, e.g., finite element method (Chen & Pei, 1989; Iru- intended also for geophysical applications, seems to address
dayaraj & Wu, 1994; Jackson, Zhang, & Ungan, 2000; the type of physics relevant to food, i.e., heating, strong
Lamberg & Hallstrom, 1986), finite difference method evaporation and multiphase transport. It appears to be
(Kuang, Thibault, & Grandjean, 1994; Plumb, Spolek, & the most promising, but it has not been applied to food or
Olmstead, 1985) and finite volume method (Boukadida & other similar processing problems so far.
Ben Nasrallah, 1995; Chang, Ruan, & Chen, 1998). When
considering strain or shrinkage, finite element method is 2.3. Commercial CFD software
used more often (Irudayaraj & Wu, 1993; Itaya, Kobay-
ashi, & Hayakawa, 1995). We can group the available solu- Use of commercial CFD software may seem like a good
tion possibilities into three somewhat overlapping choice since the conservation based governing equations
categories—(1) individual researcher codes developed for can be implemented in most of these software. They also
specific applications; (2) more general commercial and pub- generally have some porous media capabilities. Some of
lic-domain codes for solving porous media problems with the complications that need to be addressed for food pro-
evaporation; and (3) general CFD solvers that may be cus- cessing include multiphase capability, coupled liquid and
tomizable for solving porous media problems with vapor phases (i.e., liquid to vapor and vice versa), and
evaporation. unsaturated (capillary) flow. Adapting a problem having
these features to the available formulations in a typical
2.1. Individual researcher codes commercial CFD software (e.g., Fluent) is non-trivial.
Much detailed understanding of both the physics and the
Examples of codes that can solve Eqs. (31)–(34) in Datta CFD solver is required. For example, the commercial
(2006a) include Ni, Datta, and Torrance (1999), Yamsa- CFD software typically require an explicit function for
engsung and Moreira (2002) and Zhang et al. (2005). In the evaporation rate (see I_ in Eqs. (31)–(34) in Datta,
Zhang, Datta, and Mukherjee (2005), these equations are 2006a) to be included, which is possible with a non-equilib-
coupled with deformation of the matrix and mechanical rium formulation, but not an equilibrium formulation
stress analysis. For a glimpse into some of the numerical (these formulations have been discussed in Datta, 2006a).
difficulties associated with similar formulations of govern- It is not clear if the multiphase porous media capabilities,
ing equations, the reader can start, for example, from as needed for food processing problems, will be available
Turner and Mujumdar (1996). Numerical issues in relation in commercial CFD software in the near future. Continued
to Eqs. (31)–(34) in Datta (2006a), for which the results are effort is underway at the author’s research group to imple-
presented in paper, are discussed in Ni et al. (1999). For ment the types of problems discussed in this paper in avail-
research purposes, code development from scratch is the able commercial CFD software. Also, when a code has
preferred route, but it is typically a multi-year project. significant flexibility in incorporating additional equations
Commercial and public-domain software (perhaps Bixler, beyond those that are built-in (as in COMSOL), such an
1985) can potentially reduce or eliminate the code develop- arrangement cannot be expected to automatically handle
ment time, however, such software are not readily available the numerical challenges inherent in the additional equa-
as discussed below. Also, based on the discussion below, tions. Thus, complexities resulting from rapid evaporation
commercial or public-domain software that include defor- in porous media (a physics that is not typically built-in) will
mation with the multiphase transport and evaporation continue to be a challenge even in such apparently flexible
would be even more difficult to locate. codes.

2.2. Specialized porous media software 3. Various properties needed for the models

There are several well developed numerical codes related The formulations described in the companion paper
to transport in porous media, most of which are intended to (Datta, 2006a) require a large number of input parameters.
study the flow of fluids and transport of dissolved materials Although these parameters are measurable, some of the
beneath the earth’s surface (e.g., oil and gas reservoirs, con- measurements are quite difficult and data for food systems
taminated aquifers). Strong evaporation, of critical impor- are often hard to find. Following is a discussion of how these
tance to problems in food processing, is either not properties can be obtained and the extent of available data.
98 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110

3.1. Density pore size distribution in some food materials such as a

bread can be very significant (Datta, Sahin, Sumnu, &
Density includes apparent density and true density with Keskin, 2006), but the distribution effect cannot easily be
reference to bulk volume and true volume, respectively. included in modeling. Theoretical methods to predict
True volume is that occupied by the solid phase only. porosity are generally unavailable.
One way to obtain the volume of the solid phase of a sam- Apparent porosity for granular materials with clear
ple is by mechanically compacting the sample and measur- boundary, e.g., a cereal pile, can be derived solely from
ing its volume (Sumnu, Datta, Sahin, Keskin, & Rakesh, geometry (Rahman, 1995) or it can be obtained directly
2006). True density is mass divided by the true volume. based on the definition
Apparent density in multiphase porous media can be writ- qapp
ten as (Ni, 1997) /¼1 ð2Þ
qs
qapp ¼ qs ð1  /Þ þ /S w qw ð1Þ where / is the porosity, provided the true density, qs, is
known. Equations for porosity based on solely moisture
where qapp and qs are the apparent density and the true are also reported in Lozano, Rotstein, and Urbicain (1983).
density, respectively. In this equation, the mass of gas has In the results presented in the following section, that
been ignored. A series of empirical correlations of apparent uses Eqs. (31)–(34) from the companion paper (Datta,
density in terms of moisture content are given in Rahman 2006a), the quantity apparent porosity is used for / that
(1995). One possibility to calculate the true density is to is defined analogous to porosity in a non-hygroscopic
use the nutritional composition data (e.g., in Watt & Mer- material where no water is bound to the matrix. This
rill, 1975). Density data of major food components for the apparent porosity is related to moisture content as (Ni,
temperature range of 40 C to 150 C are listed in Choi 1997)
(1985).
qapp ð1  M w Þ
/¼1 ð3Þ
3.2. Porosity qs
where Mw is the moisture content on wet basis and qapp is
Porosity refers to volume fraction of pores or void given by Eq. (1). Using this equation, apparent porosity is
space. This void space is filled with one or more fluids. calculated from the initial moisture content and stays con-
Pores in a porous media can be divided into three groups stant during the entire heating process while the saturation
as illustrated in Fig. 1. These are closed pores that are of water, whose initial value is calculated from
closed from all sides, blind pores that have one end closed
and flow-through pores where the flow of fluid typically M w ð1  /Þqs
Sw ¼ ð4Þ
takes place. Many different definitions of porosity are pos- ð1  M w Þ/qs
sible (Rahman & Potluri, 1990). For non-hygroscopic changes due to transport, both spatially and with time. The
materials, porosity does not change with change in mois- common knowledge of increasing porosity (not apparent
ture content. Porosity of hygroscopic materials normally porosity) during a process such as drying (e.g., McDonald
increases as moisture content decreases (Marousis & Sarav- & Sun, 2001; Sumnu et al., 2006) is captured here using
acos, 1990). Available data on porosity change with heat- change in saturation since lower saturation is equivalent
ing of hygroscopic materials are very limited. Also, the to higher porosity.

3.3. Thermal conductivity and specific heat

True solid
volume, Vts
Flow through pore Thermal conductivity of food depends on its composi-
volume, Vfp Blind pore
volume, Vbp tion (particularly the moisture content) and temperature.
Measurement of thermal conductivity as a function of
composition and temperature (Rahman, 1995; Sweat,
1995) can be difficult and models of thermal conductivity
are often used. Most commonly used models are the
weighted (by volume fraction) addition of the conductivity
of each component, as shown in Eqs. (35) and (36) of the
Total pore volume companion paper (Datta, 2006a), that was also used in
= Vbp +Vcp +Vfp Whitaker (1977). It is important to note that thermal con-
Closed pore ductivity is particularly sensitive to the structure of the por-
volume, Vcp ous (food) material. For example, in a meat, thermal
conductivity along the direction of muscle fibers can be
Fig. 1. Schematic showing three different types of pores in a porous considerably different from that in the direction perpendic-
medium. ular to the fibers. Thus, prediction of thermal conductivity
 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110 99

using such equations can have significant error when needs to be considered in the total effective diffusivity.
applied to a particular material. More generalized form Related information and measurement techniques are
of equations for multi-component food have been given available in Karlsson (1985). Gas diffusivities in liquid
in Rahman (1995). Similar empirical models, e.g., statisti- and solid are much smaller than in gases (Saravacos,
cal correlations, were developed by treating food as a mix- 1995). Thus, molecular diffusion of gases through liquid
ture of water, protein, carbohydrate, lipid and ash (Choi & and solid matrix would often be insignificant compared
Okos, 1984; Sweat, 1995). Thermal conductivity of major to that in the gas phase.
food components at the temperature range of 40 C to
150 C are listed in Choi (1985). Other thermal conductiv- 3.5. Capillary diffusivity
ity models are summarized in Rahman (1995) and Mura-
kami and Okos (1988). In Eq. (16), capillary diffusivity has two components,
Similar to thermal conductivity and density, specific Dc, due to moisture gradient, and DT, due to temperature
heat of composite food can be a weighted addition of each gradient. Capillary diffusivity data is generally unavailable
component (Choi, 1985). Prediction of specific heat this (see Fig. 2 for soil) and data separating the two effects Dc
way is more reliable than similar prediction of thermal con- and DT do not exist with perhaps the only exception being
ductivity described above. Specific heat of a single compo- the work by Zhou, Puri, and Anantheswaran (1994), where
nent is generally a function of temperature. Thermal DT is shown to be insignificant in the context of microwave
conductivity, density and specific heat of a wide variety heating. Studies generally ignore DT. Effective diffusivity
of foods at different moisture content can also be found data (discussed later) available in the literature is close to
from Rahman (1995) and Sweat (1995). Some sources of capillary diffusivity when the material is very wet because
literature of thermal conductivity for several porous foods the vapor diffusion is insignificant. Data for Deff in potato,
are listed in Murakami and Okos (1988). Databases (e.g.,
Wolf et al., 2004) containing food physical properties data D c (m2/sec)*0.86x105
are also becoming available. -h m(m) K (cm/day)
10000 1000
3.4. Molecular diffusivity

Molecular diffusion (in the gas phase) becomes more

important when the water saturation decreases and more
1000 100
of the pore space is occupied by the gas phase. Note that
if Eq. (50) in the companion paper is used to express the Hydraulic
moisture movement in the form of Fick’s law, using an conductivity,
effective diffusivity, the driving force is the total moisture K
content, including that in the liquid phase. This effective 100 10
diffusivity in materials is different from molecular diffusiv-
ity that refers to the gas phase only, particularly when a Capillary
head, hm
material has large amounts of liquid water.
Molecular diffusivity of vapor depends on temperature, 10 1
moisture, porosity and tortuosity of food. As an example,
measured vapor–air diffusivity in apple tissues at different Capillary
water contents can be seen in Forbito, Urbicain, and Rot- diffusivity,
Dc
stein (1981). Binary gas diffusivity (needed in Eqs. (28) and
1 0.1
(29) in the companion paper) data are available (Reid &
Sherwood, 1958; Welty, Wicks, & Wilson, 1984). Theoret-
ically, binary gas diffusivity increases with temperature as
the 32 th power (Hirschfelder, Bird, & Spotz, 1949). Effect
of structure (porosity and tortuosity) change during heat- 0.1 0.01
ing is difficult to obtain and the following expression devel-
oped for soils was used by Ni (1997) to model food systems
4
Dg ¼ Dva ðS g /Þ3 ð5Þ 0.01 0.001
0 0.1 0.2 0.3 0.4
where Dg is the molecular diffusivity in the gas phase of the Moisture content (m3 of water/m3 of solid)
porous media, Sg is the saturation of the gas phase and Dva
Fig. 2. Typical relationship of capillary diffusivity to the moisture content
is the binary diffusivity of vapor in the air. in a porous media. Shown is soil moisture diffusivity Dc as calculated from
When the mean path of molecular movement is close in soil water characteristics (matric potential) hm and hydraulic conductivity
size to the capillary tubes or pores, Knudsen diffusivity K.
100 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110

10
-8 non-existent. Water permeability data on potato is shown
Diffusivity (m2/s)

in Fig. 4, more data can be seen in Datta (in press). Air per-
-9 meability of pre-gelatinized dough under different moisture
10
contents have also been measured (Goedeken & Tong,
1993). Food materials can have many physical and bio-
-10
10 chemical changes during heating (gelatinization, cell col-
lapse, etc.) that can change its permeability. For
0 1 2 3
Moisture content (dry basis)
multiphase flow in a porous media, permeability, k, is fur-
ther broken down into intrinsic permeability, ki, and rela-
Fig. 3. Example of capillary diffusivity vs. moisture content (dry basis) tive permeability, kr, of the liquid or the gas phase, given
data for potato tissue (Kiranoudis et al., 1994). by the following expression (Bear, 1972):

green pepper, carrot and onion are reported for a large k ¼ kikr ð7Þ
moisture range (Kiranoudis, Maroulis, Marinos-Kouris,
& Saravacos, 1994), an example of which is shown in The intrinsic permeability, ki, represents the permeability
Fig. 3. In Ni (1997), the following equation for capillary of a liquid or gas in fully saturated state. Thus, the mea-
diffusivity of water, Dw (presumed due to concentration sured data in Fig. 4, that are for the pores being fully sat-
gradient only) in potato: urated with water, are that of intrinsic permeability. The
relative permeability for a phase varies between 0 (none
Dw ¼ 1:0  108 expð2:8 þ 2:0MÞ ð6Þ of the phase) and 1 (saturated with that phase). Expres-
was assumed to represent the entire region from low to high sions exist that relate the relative permeability to the water
moisture content. Here M = cw/(1  /)qs is the moisture or gas content in between 0 and 1 in a material, which is
content on dry basis and Dw is the capillary diffusivity in borrowed for use in food materials, e.g., see work on
m2/s. Assumption of the above relationship for capillary dif- microwave heating (Ni et al., 1999) where the following
fusivity for potato was partly based on available data (Kira- equations are used:
noudis et al., 1994; Ruan, Schmidt, Schmidt, & Litchfield,
1991) on effective (not capillary) diffusivity data. Eq. (6) k gr ¼ 1  1:1S w S w < 1=1:1
translates approximately to a value for Dw > 1.0 · k gr ¼ 0 S w > 1=1:1 ð8Þ
107 m2/s in the constant rate region and Dw < 1.0 ·  3
S w  S ir
109 m2/s in the falling rate region. k wr ¼ S w > S ir
1  S ir
3.6. Effective diffusivity k wr ¼ 0 S w < S ir ð9Þ

The simplified and semi-empirical formulations referred Further discussion on permeability data for food applica-
to in Section 8 of the companion paper make use of the tions can be seen in Datta (in press) and Ni (1997).
property termed effective diffusivity, Deff. This effective
moisture diffusivity is related to the total moisture trans-
port and therefore it lumps capillary flow of liquid and
-17
molecular diffusion of vapor together. As mentioned ear- 10

lier, effective diffusivity is very close to capillary diffusivity

when the material is significantly wet. However, it can be Potato tissue
quite different from capillary diffusivity as the material
Intrinsic permeability[m ]

dries because the vapor diffusion increases while the liquid

2

Sample 1 (2.62 mm slice)

capillary flow becomes much smaller. For a very dry mate-
rial, effective diffusivity refers to essentially molecular diffu-
sivity. Thus, effective diffusivity is also a strong function of 10-18
water content. Effective moisture diffusivity is the most
common type of data reported in food literature (Sarava- Sample 2 (2.40 mm slice)

cos, 1995). Even so, effective diffusivity over a large range

of moisture content, is generally unavailable. More data
of moisture diffusivity in foods can be found in Saravacos
(2001) and Okos et al. (1992).

3.7. Vapor and liquid permeability 10

-19

0 50000 100000 150000 200000

Differential pressure [Pa]
Although multiphase porous media models often
require permeability data, such data on foods are almost Fig. 4. Water permeability of raw potato (from Datta, 2004).
 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110 101

3.8. Moisture isotherms and solid–vapor equilibrium Examples of empirical relations to compute water potential
curves include the one by van Genuchten (1980)
Eqs. (31)–(34) in the companion paper (Datta, 2006a) cw  cwr 1
do not form a closure as there is one more unknown than ¼ h<0
cws  cwr ½1 þ ðajhjÞb c
equation. The additional equation is that for vapor pres-
sure, pv, as function of moisture and temperature, for a ¼1 h¼0 ð12Þ
hygroscopic material. The relationships between vapor Here cwr and cws are the concentrations of water at residual
pressure and moisture content at equilibrium at a certain moisture level and under saturation, respectively. In the
temperature are called moisture isotherms. The isotherms above equation, absolute value of h is used that stands
are strongly dependent on a specific material and the bind- for the magnitude of negative head under unsaturated con-
ing ability of the bound water to the solid. Many empirical dition. The quantities a, b, c are empirical constants that
equations have been developed to express the relationships are obtained from fitting measured data, mostly available
(Chung & Pfost, 1967; Lewicki, 1998; Rahman, 1995; in applications involving soil. Measurements of water po-
Schuchmann, Roy, & Peleg, 1990; Sun, 1998). For tential curves are certainly needed for food materials.
instance, the sorption relationship of potato has been used
as (Ratti, Crapiste, & Rotstein, 1987) 4. Application of the distributed evaporation model to
p multiple processes
ln v ¼ 0:0267M d1:656 þ 0:0107e1:287M d M 1:513
d ln ps ðT Þ
ps ðT Þ
The versatility of using a multiphase porous media
ð10Þ
model in describing a number of processes is illustrated here
using the distributed evaporation model (Eqs. (31)–(34)) in
where pv and pvs are vapor pressure and saturated vapor
the companion paper (Datta, 2006a). Computed results are
pressure, respectively, in Pa and Md is moisture content
presented for five different processes of convective heating
(dry basis). The ratio pv/pvs(T) is known as the water activ-
(Ni, 1997), baking (Ni & Datta, 1999), frying (Ni & Datta,
ity. When selecting the isotherms, attention should be given
1999) and microwave heating (Ni et al., 1999), that cover a
to the materials under study and the specific range of tem-
wide range of heating modes and processing situations. In
perature and moisture. Different models for various mate-
addition, short discussion is also added for microwave com-
rials and experimental methods are available (Karel, 1989;
bined with infrared heating (Datta & Ni, 1999), using the
Pezzutti & Crapiste, 1997; Rahman, 1995).
same model. Fig. 5 shows the schematic of the 1D geometry
that is applicable to all of the simulations. The input data
3.9. Water potential curves
file for the simulations are presented in Table 1. Lack of
space here makes it impossible to provide more details on
When a porous media formulation is desired in terms of
each of these simulations and the reader is referred to the
capillary pressure (head) as the driving force, as discussed
respective publications just mentioned in this paragraph.
in the companion paper, perhaps the biggest stumbling
block is the unavailability of the relationship between cap-
illary head, h, and the amount of water, cw, in a food mate-
Air convects away
rial (e.g., see Lee, Farid, & Nguang, 2006). This energy and moisture
relationship is highly dependent on the specific food mate-
rial and its physical conditions, and is very difficult to Line of
Symmetry
obtain. Since almost no direct measurement of h is avail-
Microwaves
able today, researchers have tried to obtain h indirectly (when present)
from other available material properties data. One such
example is the Kelvin’s equation that relates capillary pres-
sure head, h, to water activity Water
(capillarity, pressure)

RT Vapor
h¼ ln Aw ð11Þ (diffusion, pressure)
qw gV m Infrared with
zero penetration
(for combination
where Aw is the water activity of the material and Vm is the heating only) Oil
molar volume of water. Significant data exists for water (for frying)
activity of a food material as a function of its water content
and such data are also relatively easy to measure. Using x=0 x=L
such data, the term oh/ocw (needed in Eq. (47)) is obtained. Surface Centerline
This equation has been shown to work for soils (Campbell, (open boundary) (closed boundary)
Gee, Campbell, & Campbell, 1992) and has been used for Fig. 5. Schematic of the 1D system that corresponds to the results in Figs.
soaking of tea leaves (Lian, Weerts, & Martin, 2003). 6–9.
102 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110

Table 1
Input parameters used in the simulations
Parameter Convective low mois. Convective high mois. Baking Frying MW heating
cps (J/kg K) 1566 1566 1566 1566 1566
cpw (J/kg K) 4180 4180 4180 4180 4180
Dw (m2/s) = 1.0 · 108exp(2.8 + 2.0M) for all studies
Deff,g (m2/s) 2.6 · 105 2.6 · 105 2.6 · 105 2.6 · 105 2.6 · 105
Fmicro,0 (W/m2) – – – – 30,000
h (W/m2 K) 20 20 15 250 20
hmv (m/s) 0.01 0.01 0.007 0.015 0.01
ks (W/m K) 0.21 0.21 0.21 0.21 0.21
kskin (W/m K) – – 0.2 – –
kw (W/m K) 0.64 0.65 0.65 0.65 0.64
kg (W/m K) 0.026 0.026 0.026 0.026 0.026
k0 (W/m K) – – – 0.17 –
kg (m2) 10 · 1014 10 · 1014 2 · 1015 10 · 1014 10 · 1014
kw (m2) 5 · 1014 5 · 1014 5 · 1016 5 · 1014 5 · 1014
k0 (m2) – – – 5 · 1014 –
kskin (m2) – – 5 · 1016 – –
L (cm) 1 1 0.635 1.27 1
Swi 0.5 0.8 0.8 0.5 0.8
S0,surface – – – 0.35 –
Sir 0.09 0.09 0.08 0.08 0.09
Tamb (C) 177 177 192 180 20
Ti (C) 20 20 20 20 20
dmicro (cm) – – – – 1 (variable)
/ 0.75 0.88 0.88 0.88 0.88
k (J/kg) 2.435 · 106 2.435 · 106 2.435 · 106 2.435 · 106 2.435 · 106
l0 (Pa s) – – – 0.0317 –
lg (Pa s) 1.8 · 105 1.8 · 105 1.8 · 105 1.8 · 105 1.8 · 105
lw (Pa s) 5.468 · 104 5.468 · 104 5.468 · 104 5.468 · 104 5.468 · 104
q0 (kg/m3) – – – 920 –
qs (kg/m3) 1419 1419 1419 1419 1419
qv,amb (kg/m3) 0 0 0 0 0
qw (kg/m3) 1000 1000 1000 1000 1000

For microwave and infrared heating, the energy flux is metry (closed boundary), no heat or mass exchange takes
assumed to decay exponentially, which leads to the source place, thus the flux values for water (liquid and vapor
terms being formulated as (Datta & Ni, 1999) together), air, and heat are set equal to zero.
 Z x  At the surface (open boundary), there is exchange of
F micro;0 dx energy, water vapor, liquid water and air. Two types of for-
q_ micro ðxÞ ¼ exp  ð13Þ
dmicro ðxÞ 0 dmicro ðxÞ mulations are considered depending on whether the liquid
 
F infra;0 x from the interior crosses this open boundary with or with-
q_ infra ðxÞ ¼ exp  ð14Þ out a complete phase change. The last situation occurs
dinfra dinfra
when the surface is saturated (so that capillary attraction
Here x is the distance from the surface and F,0 are the is zero) and there is positive pressure gradient toward the
respective fluxes at the surface. As can be seen, the formu- surface, i.e., when the following conditions are satisfied
lations for the source terms for the two heating modes are (pw is the pressure in the liquid water):
exactly the same, except that for infrared, constant penetra- S w jsurface ¼ 1 ð15Þ
tion depth is used in Datta and Ni (1999), whereas for 
opw 
microwave, penetration depth is allowed to vary with the >0 ð16Þ
local moisture content. For the infrared results presented ox surface

here, a zero penetration depth is used, which is imple- liquid can flow across the boundary without evaporation,
mented as a heat flux boundary condition instead of Eq. in a ‘pumping’ manner. Thus, water flux at the surface is
(14) (no heat generation term is present). considered to be partly from liquid water that leaves with-
out evaporation and the remaining that does evaporate at
4.1. Boundary conditions the surface. Partly to avoid the complexities associated
with this formulation, this ‘pumping’ effect is considered
The general formulation of the boundary conditions for only in the cases of convective heating and microwave heat-
the simulations are somewhat similar. At the line of sym- ing from the list in Table 1, i.e., not considered in baking
 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110 103

and frying. When the conditions given by Eqs. (15) and from latent heat relating to evaporation at the surface. This
(16) are not satisfied (and in case of baking and frying), energy at the surface is also considered to be convected
any water flux reaching the surface from the interior is con- away and is treated using a heat transfer coefficient. This
sidered to fully evaporate at the surface (i.e., not leave the heat transfer boundary condition is also modified when
surface as liquid water). Any vapor amount at the surface ‘pumping’ condition, as discussed above, is present, for
(from evaporation or transported from the interior) is con- which further details can be seen in Ni et al. (1999).
vected away to the surroundings and is computed using a
convective mass transfer coefficient. In the case of baking 4.2. Temperature profiles
of potato, where experiment showed the formation of a
skin during baking, additional resistance (both thermal Fig. 6 shows temperature profiles in convective heating,
and mass transport) was added for this layer that was con- baking, frying and microwave heating. Temperature pro-
sidered 0.5 mm in thickness. Details of this can be seen in files for convective heating of a high moisture potato are
Ni and Datta (1999). Pressure at the surface (open bound- shown in Fig. 6a. In the high moisture material, the interior
ary) is set at ambient pressure (one atmosphere). temperature tends to be constant at its wet bulb tempera-
Boundary condition for energy transfer at the surface ture, irrespective of boundary temperature. This is due to
(open boundary) is written in terms of flux from the surface the dynamic equilibrium between the surface evaporation
(which can be contributed by the convection of liquid and and the capillary diffusion of liquid water at a very large
vapor in addition to conduction), with any modifications capillary diffusivity corresponding to large water content.

Temperature T (°C) 120

50
Temperature T (°C)

40, 60 100
20 min (a)
60
40 80
40
60
30 20 min
40
Time=0 Time=0
20 20
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(a) x (cm) (b) x (cm)

160
140
Temperature (˚C)

100
Temperature (°C)

120
40 min 1 min 9 min
90 30 min 100

80 20 min 80
60
70
0 min
40
60 10 min
20
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2
(c) Distance from the surface (cm) (d) Location x (cm)

3, 4, 5
100
Temperature T (°C)

6
80
2

60

40 1 min
Time=0 (a)

20
0.0 0.2 0.4 0.6 0.8 1.0
(e) x (cm )

Fig. 6. Temperature profiles in a 1D slab, due to: (a) convective heating (high moisture material); (b) convective heating (low moisture material); (c)
convective baking; (d) frying and (e) microwave heating. Here x is distance from surface, as shown in Fig. 5. Input data used for the computations are in
Table 1.
104 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110

As the moisture level drops, capillary diffusivity is reduced the relatively dry region near the surface. Temperature
and the profiles begin to be diffusion-like. Fig. 6b shows profiles during frying (Fig. 6d) also show two distinctive
temperatures when starting with a low moisture material. regions (as in baking and drying), although with a much
Here the surface temperature increases very rapidly but sharper distinction between the wet and dry regions.
the temperature inside remains at a low value of 70 C. For microwave heating of a high moisture material,
Such profiles are characteristics of capillarity driven drying however, temperature profiles (Fig. 6e) are qualitatively
processes. The low thermal conductivity of the dry surface different from the four processes mentioned earlier. This
region, together with evaporation, leads to the significant is particularly true during the initial periods of heating.
drop in temperature from the surface to the interior. The Surface temperature stays colder, a characteristic of micro-
almost linear temperature profiles are indicative of a wave heating. Temperature rises quickly for high rates of
pseudo-steady state heat conduction process. Temperature heating in a high moisture material and reaches boiling
is somewhat uniform in the core region, primarily due to temperatures. After about 3 min, the temperatures start
the presence of the evaporation zone that acts as a sink to drop because the moisture inside becomes lower,
for the incoming energy from the surface. decreasing the microwave absorption, and therefore, the
In baking (Fig. 6c), temperature profiles are quite simi- evaporation rate and pressure development. Addition of
lar to convective drying process as mentioned above. In infrared heating (with zero penetration depth) to micro-
fact, baking and drying are implemented almost the same waves in combination heating increases surface tempera-
process when defined in the context of multiphase porous ture (Fig. 9a) and this effect is higher at increased
media transport. Most of the temperature drop occurs over infrared power level.

0.8
Time=0 0.5
Water saturation Sw

Water saturation Sw

Time=0
20 min 0.4
0.6

40 0.3 20 min
(b)
0.4 0.2 40
60 (b) 60
0.1
0.2
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(a) x (cm) (b) x (cm)
0.5
0.8 10 min
Water saturation Sw

0.4
Water saturation, Sw

0.6
0.3 0 min
1 min
0.4 20 0.2 9 min
30
0.2 0.1
40 min

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2
(c) Distance from the surface (cm) (d) Location x (cm)

1.0
Time=0
Water saturation Sw

0.8

0.6 3
2 min
0.4
4
(b) 6
0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0
(e) x (cm)

Fig. 7. Moisture profiles in a 1D slab due to: (a) convective heating (high moisture material); (b) convective heating (low moisture material); (c) convective
baking; (d) frying and (e) microwave heating. Here x is distance from surface. Input data used for the computations are in Table 1.
 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110 105

4.3. Moisture profiles tion appears to occur over a region rather than at a
sharp interface.
The water saturation profiles in a high moisture material In baking (Fig. 7c), as expected, the moisture profiles are
are shown in Fig. 7a. The moisture levels drop quite uni- qualitatively similar to the heating process mentioned in
formly due to the large capillary diffusivities at high initial previous paragraph. The dry region does not penetrate as
moisture contents. This is equivalent to a lumped parame- much as in Fig. 7b since the material starts with a higher
ter situation in heat transfer. The non-uniformity in the moisture content (more water to evaporate). Under inten-
moisture profile starts to develop once the moisture level sive frying conditions (Fig. 7d), water saturation near the
near the surface reduces to a value such that the capillary surface decreases very rapidly. Initially the dry layer region
diffusivity decreases significantly. The water saturation increases rapidly. However, it slows down as the crust
profiles (Fig. 7b) when starting with a material (potato) becomes thicker and its low thermal conductivity decreases
of low moisture show that the surface moisture drops very the rate of heat transfer. As expected, frying is a quite rapid
fast due to the quick rise in surface temperature. However, process as compared to simple heating or baking.
the inside moisture drops slowly because of a much Microwave heating of a high moisture material is shown
reduced capillary diffusivity in these regions. This leads to in Fig. 7e. At lower temperatures, there is insignificant
the characteristic diffusion profiles in Fig. 7b. As the sur- evaporation and pressure driven flow. As temperatures
face dries, the evaporation region moves inside. Evapora- reach near boiling, evaporation increases and the internal

1.00000
Total pressure P (0.1MPa)

Time=0 Total pressure P (0.1MPa)

0.99999 1.004 60

0.99998 60
1.003 40
40
0.99997
20 min 1.002
0.99996 (c) 20 min
1.001 (c)
0.99995 Time=0
1.000
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(a) x (cm) (b) x (cm)

1.010 9 min
40 min
1.10
Total pressure (0.1MPa)

1.008
Pressure (0.1MPa)

1.08 30 min
1.006
1.06
1.004
1 min
20 min
1.04 1.002
10min 1.000
1.02

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2
(c) Distance from the surface (cm) (d) Location x (cm)

3
Total pressure P (0.1MPa)

1.30 (c)
1.25 4

1.20

1.15 6
1.10

1.05 1, 2 min
Time=0
1.00
0.0 0.2 0.4 0.6 0.8 1.0
(e) x (cm)

Fig. 8. Pressure profiles in a 1D slab due to: (a) convective heating (high moisture material); (b) convective heating (low moisture material); (c) convective
baking; (d) frying and (e) microwave heating. Here x is distance from surface. Input data used for the computations are in Table 1.
106 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110

2
140 6000W/m

Surface water saturation

Surface temperature (°C)
0.6 no
infrared
120 2
4000W/m 0.5
2
100 2000W/m
2
0.4 2000W/m
80
0.3
60 no 2
infrared 4000W/m
0.2
40
0.1 6000W/m
2

20
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Heating time (min) Heating time (min)

Fig. 9. Effect of adding infrared to microwave heating on surface temperature and moisture in combined microwave-infrared heating for the schematic
shown in Fig. 5. The flux values in W/m2 are those at the surface from the infrared source, with a zero penetration depth of the infrared (all absorbed at the
surface). From Datta and Ni (1999).

1 minute 4 minutes

Moisture
Moisture

Temperature
Temperature

Vapor Vapor
pressure pressure

Gas
pressure

Gas
pressure

0 0.5 1 1.5 2 0 1 2 3

Distance from the center (cm) Distance from the center (cm)
Fig. 10. Profiles of temperature, moisture, vapor pressure and total pressure computed using a multiphase porous media transport model fully coupled
with a solid mechanics model incorporating viscoelasticity and large deformation of the dough (Zhang et al., 2005). Although the profiles are drawn to
scale, the scales are omitted here for clarity and only overall trends are of relevance here.

pressure starts to build. Even small amounts of pressure mulation near the surface. At high rates of evaporation,
can cause enough moisture to reach the surface, exceeding significant overpressure is reached. The surface starts to
its moisture removal capacity, and causing moisture accu- accumulate moisture until it becomes almost fully satu-
 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110 107

rated. The generated pressure continues to push interior overpressure probably makes the pressure driven flow
liquid water to the surface, until the surface cannot hold insignificant (Ni, 1997) for a simple heating process.
any more. The capillary force tends to zero when the mate- In baking (Fig. 8c), pressure variation is quite similar to
rial is fully saturated. After this, liquid is pumped across heating process, as expected. The pressures are small,
the open boundary without undergoing a phase change. although higher than the simple heating process. Pressure
The large drop in internal moisture after about three min- variation is confined in the sorption region where pressure
utes (Fig. 7e) is due to this surface pumping phenomena. driven flow occurs.
In combination microwave-infrared heating (Fig. 9b), In frying (Fig. 8d), pressure reaches the maximum value
addition of infrared having zero penetration increases sur- near the evaporation front. Near the surface, the pressure
face evaporation and reduces surface moisture. With builds up quickly. However, after the pressure reaches
increased infrared power level, the surface moisture can the maximum, it starts to decrease as the surface dries
be reduced to a very low value. out. While the pressure inside keeps increasing (within
the calculated period) due to high moisture. Although the
4.4. Pressure profiles pressure build-up is only 1 kPa, its effect on the moisture
transport is still important. In the crust region, vapor is dri-
The pressure profiles for the processes under discussion ven by the pressure gradient from the evaporation front to
are shown in Fig. 8. In heating of a high moisture food the surface. The vapor convective flux magnitudes are com-
(Fig. 8a), the pressures stay close to atmospheric, although parable to those for diffusional vapor flux, making the con-
it can be slightly below atmospheric, as reported by Wei, vective term important in the crust layer. In the core
Davis, Davis, and Gordon (1985). The pressure profiles region, water convective flux due to pressure is toward
for the low moisture potato are shown in Fig. 8b. The the center. The magnitude of the convective flux is small
boundary is forced to being at atmospheric pressure. The but comparable with that from capillary diffusional flux,
linear pressure profile is also representative of a pseudo making both transport mechanisms for water flux impor-
steady-state process. Although the pressure gradient can tant in the core region.
be constant, pressure driven flow may not be so, since the In rapid microwave heating, pressures can reach high
permeability varies with position. Pressure increases with value (Fig. 8e) compared to heating, baking, or frying pro-
time due to both the temperature increase and receding cesses. Rapid development of this pressure profile funda-
zone of maximum evaporation that leads to increased mentally changes the moisture transport process, as
vapor resistance to Darcy flow (more distance). The small discussed in the earlier section under moisture profiles.

6 6

4 minute
1 minute 5
5

4 4
Height (cm)

Height (cm)

3 3 80 o C
60 o C
60 oC
2 80 o C 2 100 o C

1 1

0 1 2 3 4 5 0 1 2 3 4 5
Radius (cm) Radius (cm)
Fig. 11. Changes in temperature and shape during a bread baking process computed using a multiphase porous media transport model fully coupled with
a solid mechanics model incorporating viscoelasticity and large deformation of the dough (Zhang et al., 2005).
108 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110

5. Application of the distributed evaporation model to The author’s research group has succeeded in implement-
deformable systems: bread baking ing multiphase porous media model for some important
food processes, such as the one discussed here, in commer-
Transport formulation for which results were just pre- cial solvers. We hope to make these available to others in
sented were fully coupled with deformation in Zhang the future. The other difficulty is with properties. Here
et al. (2005), i.e., effect of transport on deformation as well things may not look so bad with the current situation. If
as the effect of deformation on transport are both included. we accept the fact there is significant variability in the food
The dough matrix is treated as viscoelastic whose property properties due to biological and other sources of variability
changes with baking. The driving force for the deformation in the food properties, it is often little use to obtain results
is the vapor generated from internal evaporation and the for any given data. Rather, what would be of interest is the
carbon dioxide generated by the yeast in the dough. The sensitivity of the model to data. Thus, starting from some
transport equations follow the formulation given by the reasonable data (which can be based on multiple sources
Eqs. (31)–(34) and provide the total pressure values at a of available information), we can bracket, i.e., we can
location and at any time. These pressure values are the obtain desired output for a range of properties data. This
driving force for deformation in a solid mechanics model will show important trends and understanding of the sys-
that includes viscoelasticity of the dough and large defor- tem even when very precise data is not available. Such sen-
mation capability. An example of results obtained from sitivity analysis actually can direct us into areas where
this model can be seen in Figs. 10 and 11. Fig. 10 shows more precise data collection is necessary. Notice that pro-
how temperature, moisture, vapor pressure and total gas cedures exist today to measure any of the data discussed
pressure profiles exist in the bread from the center at two in these fundamental-based models. It will simply take time
different times during heating. Fig. 11 illustrates simulta- to make more measurements.
neous changes in temperature and deformation during
the bread baking process. Although not shown in this fig- Acknowledgements
ure, transient moisture distribution is also calculated and
is a critical part of the model. Much of the results presented here on non-shrinkable
porous media is the work of Haitao Ni. Capillary pressure
6. Summary and recommendations head formulation is from the work of Srikanth Geedipalli.
Results on swellable porous media are from the work of
A multiphase porous media model with distributed Jifeng Zhang. Their work and interactions with them made
evaporation is shown to be capable of describing a number this manuscript possible. Careful proofreading of the man-
of simultaneous heat and mass transfer processes in foods uscript by Ashish Dhall, Amit Halder and Vineet Rakesh is
that involve strong internal evaporation. Variations of also appreciated.
the model were used to simulate microwave and convective
heating, baking (without volume change), frying and References
microwave heating. A similar model with the incorporation
of large deformation was able to simulate bread baking Banaszak, J., & Kowalski, S. J. (1999). Stresses in viscoelastic plate dried
process with volume change. Being fundamental based, convectively. Drying Technology, 17(1–2), 97–117.
these models have many input parameters for which more Bear, J. (1972). Dynamics of fluids in porous media. New York: American
Elsevier Publishing Company, Inc.
data for specific foods need to be available in the future.
Bixler, N. E. (1985). NORIA – a finite element computer program for
Also, availability of commercial solvers that can readily analyzing water, vapor, air, and energy transport in porous media.
solve such models would make modeling a more useful tool Report No. SAND84-2057, UC-70, Sandia National Laboratories,
for the food process and product designer. Albuquerque, New Mexico.
This is an evolving field and it is hard to make recom- Boukadida, N., & Ben Nasrallah, S. (1995). Two dimensional heat and
mass transfer during convective drying of porous media. Drying
mendations. Although the more detailed models provide
Technology, 13(3), 661–694.
more physical insight and can be used across a number Campbell, G. S., Gee, G. W., Campbell, M. D., & Campbell, J. H. (1992).
of processes, they do require a large number of parameters, Rapid measurement of low soil water potentials using a water activity
many of which are not available for food materials. Thus, meter. Soil Science Society of America Journal, 56, 1068–1070.
the choice between more empirical models and fundamen- Carslaw, H. S., & Jaeger, J. C. (1959). Conduction of heat in solids. Oxford,
UK: Clarendon Press.
tal-based model is not as obvious. It is here that the author
Chang, K., Ruan, R. R., & Chen, P. L. (1998). Simultaneous heat and
takes a stance. In the development of science, we generally moisture transfer in cheddar cheese during cooling. I. Numerical
strive to be more fundamental, not less, even if this means simulation. Drying Technology, 16(7), 1447–1458.
all the answers are not available at any point in time. Thus, Chen, P., & Pei, D. C. T. (1989). A mathematical model of drying
the author feels that unless the application is strictly in processes. International Journal of Heat and Mass Transfer, 32(2),
297–310.
industrial context where the time and competence in the
Choi, Y. (1985). Food thermal property prediction as affected by
detailed subject matter can be severely limited, one should temperature and composition. PhD thesis, Purdue University.
go for the more fundamental-based models. There are two Choi, Y., & Okos, M. R. (1984). Effect of temperature and composition on
major sets of difficulties with these – properties and solvers. the thermal properties of foods. In M. LeMaguer & P. Jelen (Eds.).
 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110 109

Food engineering and process applications (Vol. 1, pp. 93–101). New Marousis, S. N., & Saravacos, G. D. (1990). Density and porosity in
York: Elsevier. drying starch materials. Journal of Food Science, 55(5), 1367–1372.
Chung, D. S., & Pfost, H. B. (1967). Adsorption and desorption of water McDonald, K., & Sun, D. (2001). Pore size distribution and structure of a
vapor by cereal grains and their products. Part I. Heat and free energy cooked beef product as affected by vacuum cooling. Journal of Food
changes of adsorption and desorption. Transactions of ASAE, 10, 549. Process Engineering, 24, 381–403.
Crank, J. (1975). The mathematics of diffusion. Oxford, UK: Clarendon Murakami, E. G., & Okos, M. R. (1988). Measurement and prediction of
Press. thermal properties of foods. In R. P. Singh & A. G. Medina (Eds.),
Datta, A. K. (2006a). Porous media approaches to studying simultaneous Food properties and computer-aided engineering of food processing
heat and mass transfer in food processes. I: Problem formulations. systems, Part 1. NATO ASI series (pp. 3–48). The Netherlands: Kluwer
Journal of Food Engineering, in press, doi:10.1016/j.foodeng.2006. Academic Publishers.
05.013. Ni, H. (1997). Multiphase moisture transport in porous media under
Datta, A. K. (in press). Hydraulic permeability of food tissues. Interna- internal heating of microwaves. PhD thesis, Cornell University, Ithaca,
tional Journal of Food Properties. New York.
Datta, A. K., & Ni, H. (1999). Infrared and hot air additions to microwave Ni, H., & Datta, A. K. (1999). Heat and moisture transfer in baking of
heating of foods for control of surface moisture. Journal of Food potato slabs. Drying Technology, 17(10), 2069–2092.
Engineering, 51, 355–364. Ni, H., Datta, A. K., & Torrance, K. E. (1999). Moisture transport in
Datta, A. K., Sahin, S., Sumnu, G., Keskin, S. O. (2006). Porous media intensive microwave heating of biomaterials: A multiphase porous
characterization of breads baked using novel heating modes. Journal of media model. International Journal of Heat and Mass Transfer, 42(8),
Food Engineering, in press, doi:10.1016/j.foodeng.2006.01.046. 1501–1512.
Forbito, P., Urbicain, M. J., & Rotstein, E. (1981). Water vapor–air effective Okos, M. R., Narsimham, G., Singh, R. K., & Weitnauer, A. C. (1992).
diffusivity in apple tissues. Journal of Food Science, 46, 1960–1961. Food dehydration. In D. R. Heldman & D. B. Lund (Eds.), Handbook
Goedeken, D. L., & Tong, C. H. (1993). Permeability measurements of of food engineering (pp. 437–562). New York: Marcel Dekker.
porous food materials. Journal of Food Science, 58(6), 1329–1331. Pezzutti, A., & Crapiste, G. H. (1997). Sorption equilibrium and drying
Hirschfelder, J. O., Bird, R. B., & Spotz, E. L. (1949). Chemical Reviews, characteristics of garlic. Journal of Food Engineering, 31, 113–123.
44, 205–231. Plumb, O. A., Spolek, G. A., & Olmstead, B. A. (1985). Heat and mass
Irudayaraj, J. & Wu, Y. (1993). Heat, mass, and pressure transfer during transfer in wood during drying. International Journal of Heat and Mass
drying. In American Society of Agricultural Engineers, editor. Paper Transfer, 28(9), 1669–1678.
for 1993 International Winter Meeting of the American Society of Rahman, M. S. (1995). Food properties handbook. Boca Raton, Florida:
Agricultural Engineers, Vol. 936584 (pp. 1–16), St. Joseph, Michigan, CRC Press, Inc.
1993. American Society of Agricultural Engineers. Rahman, M. S., & Potluri, P. L. (1990). Shrinkage and density of squid
Irudayaraj, J., & Wu, Y. (1994). Finite element analysis of coupled heat, flesh during air drying. Journal of Food Engineering, 12, 133–143.
mass, and pressure transfer in porous biomaterials. Numerical Heat Ratti, C., Crapiste, G. H., & Rotstein, E. (1987). A new water sorption
Transfer, Part A, 26, 337–350. equilibrium expression for solid foods based on thermodynamic
Itaya, Y., Kobayashi, T., & Hayakawa, K. (1995). Three-dimensional heat considerations. Journal of Food Science, 54(3), 738–747.
and mass transfer with viscoelastic strain–stress formation in compos- Reid, R. C., & Sherwood, T. K. (1958). The properties of gases and liquids.
ite food during drying. International Journal of Heat and Mass New York: McGraw-Hill Book Company.
Transfer, 38(7), 1173–1185. Ruan, R., Schmidt, S. J., Schmidt, A. R., & Litchfield, J. B. (1991).
Jackson, T. H., Zhang, Q., & Ungan, A. (2000). Numerical modeling of Nondestructive measurement of transient moisture profiles and the
microwave induced natural convection. International Journal of Heat moisture diffusion coefficient in a potato during drying and absorption
and Mass Transfer, 43(12), 2141–2154. by NMR imaging. Journal of Food Process Engineering, 14, 297–313.
Karel, M. (1989). Role of water activity. In R. P. Singh & A. G. Medina Saravacos, G. D. (1995). Mass transfer properties of foods. In M. A. Rao
(Eds.), Food properties and computer-aided engineering of food process- & S. S. H. Rizvi (Eds.), Engineering properties of foods (2nd ed.,
ing systems, Part 2. NATO ASI series (pp. 135–155). The Netherlands: pp. 169–221). New York: Marcel Dekker, Inc.
Kluwer Academic Publishers. Saravacos, G. D. (2001). Mass transfer properties of foods. New York:
Karlsson, E. I. (1985). Gaseous diffusion in solid food systems and the Marcel Dekker, Inc.
dependence on structure and temperature. PhD thesis, University of Schuchmann, H., Roy, I., & Peleg, M. (1990). Empirical models for
California, Davis. moisture sorption isotherms at very high water activities. Journal of
Kiranoudis, C. T., Maroulis, Z. B., Marinos-Kouris, D., & Saravacos, G. Food Engineering, 55(3), 759–762.
D. (1994). Estimation of the effective moisture diffusivity from drying Sumnu, G., Datta, A. K., Sahin, S., Keskin, S. O., Rakesh, V. (2006).
data. Application to some vegetables. In T. Yano, R. Matsuno, & K. Transport and related properties of breads baked using various heating
Nakamura (Eds.). Developments in food engineering: Proceedings of modes. Journal of Food Engineering, in press, doi:10.1016/j.foodeng.
the 6th international congress on engineering and food (Vol. 1, 2006.01.010.
pp. 340–342). New York: Blackie Academic & Professional. Sun, D. (1998). Selection of EMC/ERH isotherm equation for shelled
Kuang, H., Thibault, J., & Grandjean, P. A. (1994). Study of heat and mass corn based on fitting to available data. Drying Technology, 16(3-5),
transfer during IR drying of paper. Drying Technology, 12(3), 545–575. 779–797.
Lamberg, I., & Hallstrom, B. (1986). Thermal properties of potatoes and a Sweat, V. E. (1995). Thermal properties of foods. In M. A. Rao & S. S. H.
computer simulation model of a blanching process. Journal of Food Rizvi (Eds.), Engineering properties of foods (2nd ed., pp. 99–138). New
Technology, 21, 577–585. York: Marcel Dekker, Inc. (Chapter 3).
Lee, K. T., Farid, M., & Nguang, S. K. (2006). The mathematical Turner, I., & Mujumdar, A. S. (1996). Mathematical modeling and numerical
modelling of the rehydration characteristics of fruits. Journal of Food techniques in drying technology. New York: Marcel Dekker, Inc.
Engineering, 72, 16–23. van Genuchten, M. T. (1980). A closed-form equation for predicting the
Lewicki, P. P. (1998). A three-parameter equation for food moisture hydraulic conductivity of unsaturated soils. Soil Science Society of
sorption isotherm. Journal of Food Processing Engineering, 21, 127–144. America Journal, 44, 892–898.
Lian, G., Weerts, A. H., & Martin, D. R. (2003). Modeling the hydration Watt, B. K., & Merrill, A. L. (1975). Composition of foods. In Agriculture
of foodstuffs: temperature effects. AIChE Journal, 49(5), 1334–1339. handbook No. 8. US Department of Agriculture, Washington, DC.
Lozano, J. E., Rotstein, E., & Urbicain, M. J. (1983). Shrinkage, porosity Wei, C. K., Davis, H. T., Davis, E. A., & Gordon, J. (1985). Heat and
and bulk density of foodstuffs at changing moisture contents. Journal mass transfer in water-laden sandstone: Convective heating. AIChE
of Food Science, 48, 1497–1553. Journal, 31(8), 1338–1348.
110 A.K. Datta / Journal of Food Engineering 80 (2007) 96–110

Welty, J. R., Wicks, C. E., & Wilson, R. E. (1984). Fundamentals of Yamsaengsung, R., & Moreira, R. G. (2002). Modeling the transport
momentum, heat and mass transfer (3rd ed.). New York: John Wiley & phenomena and structural changes during deep fat frying – part 1:
Sons. Model development. Journal of Food Engineering, 53(1), 1–10.
Whitaker, S. (1977). Simultaneous heat, mass and momentum transfer in Zhang, J., Datta, A. K., & Mukherjee, S. (2005). Transport processes and
porous media: A theory of drying. Advance in Heat Transfer, 12, volume expansion during baking of bread. AIChE Journal, 51(9),
119–203. 2569–2580.
Wolf, W., Gekas, V., Jarvis, D., Sadd, P. A., Johns, A. I., Nesvadba, P., Zhou, L., Puri, V. M., & Anantheswaran, R. C. (1994). Measurement of
et al. (2004). Database of physical properties of agro-food materials. coefficients for simultaneous heat and mass transfer in food products.
Journal of Food Engineering, 61(4), 497–503. Drying Technology, 12(3), 607–627.