Critical Reviews in Food Science and Nutrition

ISSN: 1040-8398 (Print) 1549-7852 (Online) Journal homepage: http://www.tandfonline.com/loi/bfsn20

Computer simulation for improving radio

frequency (RF) heating uniformity of food
products: A review

Zhi Huang, Francesco Marra, Jeyamkondan Subbiah & Shaojin Wang

To cite this article: Zhi Huang, Francesco Marra, Jeyamkondan Subbiah & Shaojin Wang (2017):
Computer simulation for improving radio frequency (RF) heating uniformity of food products: A
review, Critical Reviews in Food Science and Nutrition, DOI: 10.1080/10408398.2016.1253000

To link to this article: https://doi.org/10.1080/10408398.2016.1253000

Accepted author version posted online: 28

Nov 2016.
Published online: 02 Jun 2017.

Submit your article to this journal

Article views: 316

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=bfsn20
CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION
https://doi.org/10.1080/10408398.2016.1253000

Computer simulation for improving radio frequency (RF) heating uniformity of food
products: A review
Zhi Huanga, Francesco Marrab, Jeyamkondan Subbiahc, and Shaojin Wanga,d
a
College of Mechanical and Electronic Engineering, Northwest A&F University, Yangling, Shaanxi, China; bDipartimento di Ingegneria Industriale,
Universit
a degli studi di Salerno, Fisciano, SA, Italy; cDepartments of Biological Systems Engineering and Food Science & Technology, University of
Nebraska-Lincoln, Lincoln, Nebraska, USA; dDepartment of Biological Systems Engineering, Washington State University, Pullman, Washington, USA

ABSTRACT KEYWORDS
Radio frequency (RF) heating has great potential for achieving rapid and volumetric heating in foods, Computer simulation; food
providing safe and high-quality food products due to deep penetration depth, moisture self-balance products; heating uniformity;
effects, and leaving no chemical residues. However, the nonuniform heating problem (usually resulting in model; radio frequency
hot and cold spots in the heated product) needs to be resolved. The inhomogeneous temperature
distribution not only affects the quality of the food but also raises the issue of food safety when the
microorganisms or insects may not be controlled in the cold spots. The mathematical modeling for RF
heating processes has been extensively studied in a wide variety of agricultural products recently. This
paper presents a comprehensive review of recent progresses in computer simulation for RF heating
uniformity improvement and the offered solutions to reduce the heating nonuniformity. It provides a brief
introduction on the basic principle of RF heating technology, analyzes the applications of numerical
simulation, and discusses the factors influencing the RF heating uniformity and the possible methods to
improve heating uniformity. Mathematical modeling improves the understanding of RF heating of food
and is essential to optimize the RF treatment protocol for pasteurization and disinfestation applications.
Recommendations for future research have been proposed to further improve the accuracy of numerical
models, by covering both heat and mass transfers in the model, validating these models with sample
movement and mixing, and identifying the important model parameters by sensitivity analysis.

1. Introduction
electrodes. During RF treatments, heat is generated within the
Radio frequency (RF) heating (also known as capacitive dielec- product due to molecular friction resulting from oscillating
tric heating) is a recognized rapid electro-heating technology, dipole molecules and migrating ions caused by the applied
which forms a part of innovative techniques based on electro- alternating electric field (Piyasena et al., 2003). Theoretically,
magnetic heating and other nonthermal methods. The available therefore, RF heating is expected to deliver more uniform heat
systems for producing and transferring RF power to dielectric at a faster rate than conventional heating. Conventional heating
heating can be divided into two distinct groups: the more wide- relies on conduction and convection to transport heat from the
spread conventional RF heating equipment and the more heating sources to the product, which requires relatively longer
recent 50-V RF heating equipment (Marchand and Meunier, period of time, whereas RF heating has the potential to deliver
1990). Whether conventional or 50-V dielectric heating sys- heat instantly throughout the product (Zhao et al., 2000). RF
tems are used, the RF applicator has to be designed for the par- heating offers the advantages of providing more uniform heat-
ticular product being heated. Although the size and shape of ing over sample geometry than microwave heating does due to
the applicator can vary enormously, they mostly fall into one of deeper wave penetration into the sample and is more economi-
three main types: the through field applicator, the fringe field cal than microwave at higher power levels (Luechapattanaporn
applicator, and the staggered through field applicator (Jones et al., 2004). These features of RF heating have generated great
and Rowley, 1996). Conceptually, a through field RF applicator interests for applications in the food-processing industry.
(with a pair of parallel plates) is the simplest and the most com- The potential use of RF technology for food processing was
mon design of RF applicators in which the lossy material is recognized since the 1940s (McCormick, 1988). These early
placed between the two electrodes to form a parallel plate efforts employed RF energy for applications, such as the cook-
capacitor. This type of applicator is mainly used for heating ing of processed meat products, heating of bread, dehydration,
bulk materials or related large and thick objects. High voltage is and blanching of vegetables at 150 MHz (Moyer and Stotz,
applied onto one of the electrodes while the other is grounded, 1947). The next generation of commercial applications for RF
thus causing propagation of electromagnetic waves between the energy in the food industry was postbake drying of cookies and

CONTACT Shaojin Wang, Ph.D, Professor [email protected] College of Mechanical and Electronic Engineering, Northwest A&F University, Yangling,
Shaanxi 712100, China.
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/bfsn.
© 2017 Taylor & Francis Group, LLC
2 Z. HUANG ET AL.

snack foods, which began in the late 1970s (Holland, 1974). and have discussed mathematical modeling and the computer
Later in 1990s, the area of RF pasteurization was studied with simulation of RF systems. Neophytou and Metaxas (1996,
attempts made to improve energy efficiency and solve technical 1997, 1998, 1999) attempted to model the electrical field for
problems (Zhao et al., 2000). This in turn has led to recent industrial-scale RF heating systems and compared solutions
investigations on RF heating applications in food processing, as from both electrostatic and wave equations. Later on, Marra
well as dielectric properties of the product and system specific et al. (2007) proposed and experimentally validated a mathe-
factors that influence the RF heating (Piyasena et al., 2003). matical model for RF heating of solid-like food (meat batter),
Subsequently, Marra et al. (2009) and Jojo and Mahendran based on so-called quasi-static approach, where the Gauss law
(2013) reviewed several additional applications for RF heating was coupled with classical transient heat transfer equation, tak-
in food processing and their impacts on general quality aspects ing into account, beside the conductive term, a volumetric
of foods. The most recent focus of this novel heating technol- power generation. After that, further steps towards improving
ogy was investigated for baking and roasting (Awuah et al., the understanding of RF heating were taken for various food
2014a), cooking (Kirmaci and Singh, 2012; Laycock et al., 2003; products, such as dry foods (Huang et al., 2016a; Jiao et al.,
Marra et al., 2007; Rincon et al., 2015), controlling insects (Hou 2015a; Tiwari et al., 2011a, b), fruits (Birla et al., 2008a, b),
et al., 2016; Jiao et al., 2012; Wang et al., 2007a, b), defrosting meat lasagna (Wang et al., 2012), peanut butter (Jiao et al.,
(Farag et al., 2009; Llave et al., 2014, 2015), drying (Jumah, 2014a, b), and raisins (Alfaifi et al., 2014). Overall, the consen-
2005; Wang et al., 2011, 2014), pasteurization (Awuah et al., sus is that RF heating is only able to heat certain food products
2005; Gao et al., 2011; Geveke et al., 2002; Ha et al., 2013; Kim uniformly, and sometimes processing aids (Jiao et al., 2015b) or
et al., 2012; Liu et al., 2011; Nagaraj et al., 2016), sterilization modifications to the RF applicator (Tiwari et al., 2011a) are
(Guan et al., 2004; Wang et al., 2003b), thawing (Farag et al., needed to achieve good RF heating uniformity. These include
2008a, 2011; Uyar et al., 2015), and many more. hot water-assisted treatment (Tiwari et al., 2008), combining
Although there are many research evidences suggesting pos- with moving or rotating method (Birla et al., 2008a; Chen et al.,
sible advantages in the food industry, RF energy has not yet 2016), intermittent mixing (Chen et al., 2015b), electrode type
been exploited to its fullest potential in the industrial applica- modification (Tiwari et al., 2011a), and similar dielectric mate-
tions. One of the major drawbacks of RF heating is the exis- rial surrounding methods (Huang et al., 2015c; Jiao et al.,
tence of hot spots in several zones depending on product 2015b).
geometry (Alfaifi et al., 2014; Fu, 2004; Tiwari et al., 2011b; The general purpose of this paper was to review the current
Wang et al., 2006b; Wang et al., 2008a). Sample size, shape, and state of numerical modeling of overheating that usually occurs
nonhomogeneous dielectric properties (DPs) could result in at corners and edges of the RF-treated products in rectangular
nonuniform electric field distribution, which in turn causes containers for better understanding the causes of nonuniform
nonuniform temperature distribution (Birla et al., 2008a; Marra temperature distributions in foods and provide an optimal
et al., 2007; Tiwari et al., 2011a). The nonuniform temperature solution for improving the RF heating uniformity in these
distribution has been studied in different agricultural commod- foods. The prospects of further research were examined and
ities, such as dry nuts (Wang et al., 2001, 2002, 2005; Wang various applications of numerical models were discussed. The
et al., 2014), dried fruits (Alfaifi et al., 2014), eggs (Lau et al., specific objectives of this review were (1) to introduce the basic
2016), fresh fruits (Birla et al., 2004, 2005; Tiwari et al., 2008; mechanism of RF heating and present the fundamental of
Wang et al., 2006a), legumes (Huang et al., 2015c; Jiao et al., mathematical modeling of RF heating, (2) to review the litera-
2011; Wang et al., 2008c, 2010) and meat (Llave et al., 2015; tures of computer simulation on RF heating applications in
Marra et al., 2007; Romano and Marra, 2008; Uyar et al., 2014, food processing, (3) to explore the nonuniform RF heating
2016). Major problems in RF heating, such as poor end quality, behavior with the developed simulation model, (4) to analyze
pathogens/insects survivals, microbial safety concerns, and the main factors that affect the RF heating uniformity and dis-
overheating, are related to nonuniform heating during RF treat- cuss the simulated methods to improve RF heating uniformity
ment (Birla et al., 2004; Brunton et al., 2005; Jiao et al., 2012; in food processing, and (5) to propose recommendations for
Kim et al., 2012; Kirmaci and Singh, 2012). Several researchers developing computer aided engineering of RF processes on an
have studied the temperature distribution during RF heating in industrial-scale.
food materials and suggested measures to reduce the nonuni-
form temperature distribution (Chen et al., 2015b; Huang et al.,
2016b; Ikediala et al., 2002; Jiao et al., 2014a; Tiwari et al., 2. Overview of RF heating modeling
2011a). The trial and error procedures are time consuming,
2.1. RF heating system
costly, and often provide limited information. Due to rapid
progress in computing capability, computational modeling has RF technology uses electromagnetic energy within a frequency
increasingly been used to predict RF power and temperature range of 3 kHz to 300 MHz to heat target foods. Federal Com-
distribution in food products without the necessity of extensive munication Commission (FCC) assigned 13.56 § 0.00678,
experiments. 27.12 § 0.16272, and 40.68 § 0.02034 MHz for RF treatments
Numerical techniques, such as the finite element method in industrial, scientific and medical (ISM) applications to pre-
(FEM), have been extensively used in simulations of RF heat- vent disturbance in telecommunications (Jones and Rowley,
ing. Its capability was first investigated by Baginski et al. (1989) 1996). RF heating is based on the transformation of alternating
and again later by Choi and Konrad (1991). Starting from the electromagnetic field energy into thermal energy by affecting
mid 1990s, a number of papers have appeared in the literature the polar molecules and charged ions of a material. Dielectric
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 3

materials are composed of atoms or molecules irrespective of its magnetrons or klystrons, and can be transmitted to an applica-
phase (solid, liquid, or gas) and atoms or molecules are com- tor or antenna through a waveguide or coaxial transmission
posed of negatively charged electrons and positively charged line. The output of such tubes tends to be in a range from 0.5
nucleus. These constituents may be locked into regular struc- to 100 kW and requires a special power supply (Piyasena et al.,
tures of crystals or free to wander through the structure. There- 2003). RF systems are generally simpler to construct than MW
fore, several types of electric displacements of these systems, which usually contain an oscillator and applicator cir-
constituents, such as electronic, atomic, molecular, and ionic, cuit. The RF heating also offers simple uniform field patterns as
are possible (Hartshorn, 1949). The molecular displacement opposed to the complex nonuniform standing wave patterns in
involves molecules that are initially asymmetrical in structure a MW oven (Piyasena et al., 2003). Compared with MW
and have a definite electric moment in addition to their own energy, RF energy has the advantage of heating bulk food
electric field when placed in an external electric field (Piyasena because of its relatively longer wavelength (7.4–22.1 m in vac-
et al., 2003). When an external electric field is applied, the bipo- uum). As a result, a better heating pattern for RF processing
lar molecules tend to behave like microscopic magnets and would be obtained compared to MW processing.
attempt to align themselves with the field. When the electrical A typical RF system consists of two main components: gen-
field is changing millions of times per second (e.g. 27.12 MHz), erator and applicator. The generator part is mainly used to gen-
these molecular magnets are unable to withstand the forces act- erate the RF energy, and the applicator is a metal structure,
ing to slow them. This resistance to the rapid movement of the which directs the RF field to the product or load to be heated.
bipolar molecules creates friction and results in heat dissipation Two main kinds of RF systems are used in commercial applica-
in the material exposed to the RF radiation. This interaction tions: the open circuit (also called free running oscillator) and
results in translation motions of free or bound charges and the 50 V technology based on their wave generation mecha-
rotation of dipoles, and causes losses that, in turn, produce the nisms, components, and properties (Marra et al., 2009). The
volumetric heating (Piyasena et al., 2003). The ability of RF open circuit RF system is used widely in many industrial appli-
radiation to penetrate and couple with materials provides an cations, but the 50-V system is a relatively new technology to
attractive method for obtaining controlled and precise heating. provide a fixed frequency compared to self-oscillatory circuit,
In RF treatment, heat is generated throughout the material and also to precisely control power and feedback (Jones and
by molecular friction in high-frequency alternating electric Rowley, 1996). Thus the heating rate can be tightly controlled
fields, leading to faster heating rates, which was similar to the and precise final temperatures can be achieved easily in a 50-V
microwave (MW) heating (Tang et al., 2005). Both RF and system. Because power is fed through standard 50-V cables, the
MW heating methods are nonionizing radiation since the fre- generator can be remotely located and can be quickly con-
quencies in RF and MW ranges cannot produce sufficient nected. It has a more stable frequency output since the match-
energy to ionize molecules (Awuah et al., 2014b). Therefore, RF ing system in the RF heater is automatically adjusted to
and MW heating is also called dielectric heating or dielectric maintain the load impedance at 50 V. The 50-V systems are
loss heating. RF and MW systems have been recognized to be more expensive than the open circuit ones, and have not been
50–70% heating efficient in comparison to 10% efficiency with popularly used by industries. The open circuit RF system gener-
conventional heating ovens (Mermelstein, 1997). RF heating ates heat by means of an RF generator that produces oscillating
also differs from the higher frequency MW heating (915 MHz fields of electromagnetic energy and comprises a powersupply
and 2450 MHz). In RF heating, the electrical field is commonly and control circuitry, a parallel plate and a system for support-
generated in a directional manner between a pair of electrode ing processed material as described in Fig. 1. The target mate-
plates, but the electrical field could approach the product from rial placed between top and bottom electrodes is moved on a
all directions in MW heating depending on different mode conveyor belt to simulate continuous processes, and acts as a
designs. RF and MW system configurations are totally different, capacitor to store electrical energy and a resistor to transfer
since MW energy is generated by special oscillator tubes, electric energy to thermal energy. The top electrode plate with

Figure 1. Schematic diagram of the radio frequency (6 kW, 27.12 MHz) heating system.
4 Z. HUANG ET AL.

0 00
adjustable height is inductively coupled to the tank oscillator permittivity (e r ), and the relative loss factor (er ) are related as
circuit via feed strips. The power supplied to the sandwiched follows:
object is typically controlled by varying the electrode gap in
most industrial systems. er D e 6 e0 D e 0 6 e0 ¡ je00 6 e0 D e 0r ¡ je00r (2)

2.2. Mechanisms of RF heating where e0 is the permittivity of electromagnetic wave in free

space (8.854 £ 10¡12 F/m).
DPs influence reflection of electromagnetic waves at interfaces An important thermo-physical property that affects RF
and the attenuation of the wave energy within materials (Birla heating of foods is electrical conductivity (s, S/m), which is
et al., 2008b). That in turn determines the amount of energy defined as the ability of a food to conduct electric current. Elec-
absorbed and converted into heat. DPs related to RF heating trical conductivity is related to the ionic depolarization that
are permeability, permittivity (capacitivity), and electrical con- occurs within the food during RF heating. s of a material can
ductivity of the heated material. It is generally accepted that, be expressed by the following equation (Piyasena et al., 2003):
for most dielectric materials, permeability has no or very small
contribution to dielectric heating, and so it is usually not con- s D 2pf e00 D 2pf e0 e00r (3)
sidered (Zhang and Datta, 2001). The permittivity that deter-
mines the dielectric constant (e0 ), the dielectric loss factor (e00 ), where f is the electromagnetic frequency (Hz).
and the loss angle, influences the RF heating. The e0 and e00 , The tangent of the dielectric loss angle (tan d) is often called
which are the real and imaginary parts of the complex permit- the loss tangent or the dissipation (power) factor of the mate-
tivity, respectively, of e, are given by: rial. For a given material, this is equivalent to the ratio of e
00

0
over e :
e D e 0 ¡ je00 (1)
tand D e00 6 e 0 D e00r 6 e 0r (4)

where e is permittivity (F/m), e0 is often called the dielectric The DPs of food products are affected by many factors,
constant or “capacitivity”, describing the degree of a material’s including frequency, temperature, and moisture content. The
interaction with an alternative electrical field, and quantifying mechanisms that contribute to the dielectric loss in heteroge-
its ability for reflecting, storing, and transmitting electromag- neous mixtures include dipole, electronic, atomic, and
netic energy. e00 is commonly called dielectric loss factor, which Maxwell–Wagner responses (Fig. 2). For moist dielectric
is a measure of the ability of a material to dissipate electrical materials, ionic conductivity plays a major role at lower
energy into heat, and j D . ¡ 1/0:5 . A material with a low e00 frequencies (e.g.,<200 MHz), whereas both ionic conductiv-
may absorb less energy and could be expected to heat poorly in ity and dipole rotation of free water play a combined role at
an electrical field due to its greater transparency to electromag- MW frequencies. The Maxwell–Wagner polarization effect
netic energy (Marra et al., 2009). peaks at about 0.1 MHz, but in general, its contribution is
In practice, the DPs of materials relative to that of vacuum small compared to that of ionic conductivity (Tang et al.,
are used. The relative permittivity (er ), relative complex 2005). For foods with low moisture content, dipole

Figure 2. Contribution of various mechanisms to the loss factor of food materials as functions of frequency and temperature (Tang et al., 2002).
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 5

dispersion of free water molecules is negligible, so the bound 2.3. RF power density
water plays a major role in dielectric heating in the fre-
An open circuit RF system consists of a high voltage trans-
quency range from 20 to 30,000 MHz. Bound water is a
former, a rectifier, an oscillator tube, a tuned circuit, an
form of water that has its mobility between ice and free
impedance coupling and matching circuit, and an applicator
water. Bound water molecules have a lower relaxation fre-
(Fig. 3). The line power from the wall is transformed to a
quency than free water at around 100 MHz (20 C), and the
high voltage, and converted into DC power by a rectifier.
magnitude is also much smaller than free water. It could be
The oscillatortube excites high frequency alternating electro-
found that both the bound water and Maxwell–Wagner’s
magnetic waves and transports them to the tank circuit to be
effects contribute to the DPs at a very low range comparing
tuned to a specific working frequency and to match the load.
with free water and ionic effects. This explains the low values
The applicator receives the high-frequency electromagnetic
of DPs in low moisture foods.
wave, and the load in it converts the electromagnetic wave
From Fig. 2, we can see that the influence of ionic conduc-
into heat.
tion is always positive when temperature increases. At the RF
In principle, if there is a dielectric material between two elec-
and MW frequency of practical importance and current appli-
trodes, the material would become a resistance, and the current
cations in food processing (RF: 1 to 50 MHz and MW: 915 and
flowing through the resistance is in phase with the applied volt-
2450 MHz), ionic conduction and dipole rotation are dominant
age. For the most commonly used flat capacitor, an AC dis-
loss mechanisms (Ryyn€anen, 1995):
placement current passes throughout the food, the capacitance
can be expressed as (Zhao et al., 2000):
s
e00 D e00d C e00s 0 D e00d C (5)
2pf e0
e0 e 0r A
CD (7)
0 d
where subscripts d and s stand for contributions due to dipole
rotation and ionic conduction, respectively. where C is the capacitance (F), A is the area of an electrode
Penetration depth (dp, m) of RF power is defined as the (m2), and d is the distance between the two electrodes (m).
depth where the power is reduced to 1/e (e D 2.718), about Figure 3 shows the circuit diagram of a dielectric heating
37%, of its value at the surface of the material. The dp value in a system and the current directions. The current through the
lossy material can be calculated as follows (Marra et al., 2009): capacitor can be expressed by:

c A
dp D  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 12 (6) IC D vVC D vVe0 e 0r (8)
pffiffiffi e00 2 d
2 2pf e 0 1 C e0 ¡ 1
where IC is the electric current (A),v is angular frequency (v D
2pf, rad/s), and V is the electric potential between the two elec-
where c is the speed of light in free space (3 £ 10 m/s). Due to 8
trodes (volt) and is related to the electrical field strength (E D
the much longer wavelength (11 m at 27 MHz), RF energy pen-
V/d, volt/m). The current through the load (IR) can be
etrates further into food materials than MW (0.12 m at
expressed by:
2450 MHz). This factor is important in the selection of the
appropriate thickness of a material bed to ensure a uniform
A !
heating during the RF process. IR D IC tand D vVe0 e 0r tand D 2pf j E j e0 e00r Atand (9)
d

Figure 3. Scheme of a typical open circuit RF system (Zhao et al., 2000).

6 Z. HUANG ET AL.

The power dissipated in the load can be expressed by: made, Eqs. (12)–(15) become:
! !
Q D IR V D 2pf j E j 2 e0 e00r dA (10) !
r £E D ¡ jmvH
!
(18)
! ! ! !
where j E j is electric field strength (V/m) in the food load and r £H D je0 er vE (19)
dA is the volume of load. So the power dissipated per unit vol-
ume (P, W/m3) can be rewritten as follows: Under time harmonic condition, the effect of the magnetic
field is negligible, Eq. (18) can be ignored and Eq. (19) can be
! !
P D 2pf j E j 2 e0 e00r D 5:56£10 ¡ 11 f e00r j E j 2 (11) changed to:
 
The electromagnetic field energy is converted into thermal ! ! !
r £H D .s C jve/E D jv e C
s ! ! !
E D jvecE D J (20)
energy through interactions with a dielectric material. jv

In existing RF applicators, the wavelength in the RF

2.4. Electromagnetic field and Maxwell equation solutions
range is often much larger than the maximum size of the
Modeling RF treatments in food processing applications object to be heated, quasi-static approximation can be
involves the simulation of the electric and magnetic fields applied for the solution of Maxwell’s electromagnetic field
that pass throughout the food, and its interaction with food equations. Since the
!
scalar
!
voltage potential (V) is related
! !
to
molecules and the simulation of heat transfer within the electric field by E D ¡ r V, and current density r J D 0,
product. A brief theoretical explanation of what RF energy Eq. (20) can also be modified to Laplace equation (Metaxas,
is and how it interacts with food matteris needed to under- 1996):
! !
stand their general behaviors. The general approach for ¡ r ..s C j2pf e0 er / r V/ D 0 (21)
solving electromagnetic field in 2-and 3-D requires the
application of Maxwell equations in differential form (Neo- The electric field at any point inside the electrodes is
phytou and Metaxas, 1998): governed by Eq. (21). In this case, as boundary conditions,
! !
the top electrode is set as the electromagnetic source since
! @B! @H it introduces the high frequency electromagnetic energy and
r £E D D ¡m (12)
@t @t the bottom electrode is set as ground conditions (V D 0).
! !
! ! ! @D ! @E RF applicator walls are electrically insulated, so last bound-
r £H D J C D s c E C e0 er (13) ary conditions are:
@t @t ! !
! ! ! ! r E D 0 (22)
r  D D r  eE D re (14)
! ! ! !
The analytical solution of Laplace equation for the electric
r  B D r  mH D 0 (15)
potential between RF electrodes and heat dissipation in a slab
! ! sandwiched between electrodes can be coupled together, and
where E is electric! field intensity (V/m), B is! magnetic flux
an expression for electric potential on upper electrode can be
density (Wb/m2), H is magnetic field (A/m2), D is electric flux
obtained as (Birla et al., 2008a):
density (C/m2) of the dielectric material, m is magnetic perme-
!
ability (H/m), re is free volume charge density (C/m3), J is the  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!
current density (A/m2), and s c is electrical conductivity (S/m) 2 2 rcp dT
V D dair ðe 0 Þ C ðe00 Þ C dmat (23)
of the material. pf e0 e00 dt
In problems of wave propagation, the behavior of an electro-
magnetic wave is concerned in a source-free region where re where dair is the air gap between top electrode and food sample
and J are zero. Equations (12)–(15) can be combined to give a (m), and dmat is the height of the food material (m). The elec-
second-order homogeneous vector wave equation in E and H trode voltage would change as a function of spatial location
alone. between the two electrodes.
!
! @2 E
r E ¡ me 2 D 0
2
(16)
@t
Field vectors that vary with space coordinates and are sinu- 2.5. Heat transfer modeling
soidal functions of time can similarly be represented by vector The common objective in modeling RF heating is to obtain
phasors that depend on space coordinates but not on time. The the temperature profile inside the food. The equations rep-
RF field can be seen as a time harmonic field, and the electric resenting the electromagnetic and the thermal phenomena
field strength can be expressed below: must be solved inside the food itself. Once electromagnetic
! !
model has been solved, heat generation can be derived from
E.x; y; z; t/ D Re½E.x; y; z/ejvt  (17) electromagnetic fields and material properties. Therefore, a
link has to be created between the electromagnetic equa-
where Re[E(x,y,z)ejvt] is the real part of [E(x,y,z)ejvt], and t is tions and heat transfer equations. The amount of power (P,
the time period (s). When the time harmonic assumption is W/m3) converted from electromagnetic energy to thermal
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 7

energy is related to the DPs!

of treated materials. For a given which that equation is considered. In this way, it is possible to
electric field intensity ( j E j , V/m), P was already introduced model the selective RF heating, since different substrates may
above (Eq. (11)), as described below (Choi and Konrad, 1991): heat faster or slower depending on their different loss factors.
The same concept applies when the goal is to study the effect of
! !
P D 2pf e0 e00 j E j 2 D v e0 e00 j E j 2 (24) RF heating on insects in host products (such as fruits). When
the applied frequency and electric fields are the same both for
! the pests and the host products, the heat produced in the insects
where j E j is the modulus of the E-field, that—in case of RF
and the host products might be different, resulting in the differ-
heating—can be evaluated in accordance with the Laplace
ent final temperatures of the insects and the host products after
equation (Eq. (21)), stated for electroquasi-static conduction
the same period of RF heating (Huang et al., 2015a).
in inhomogeneous materials. Here, volumetric energy genera-
For the heat transfer equation to determine the temperature
tion term depends on the electric fields experienced within
distribution inside the food undergoing RF heating, initial and
the food, the frequency of electromagnetic field, and DPs of
boundary conditions are needed. A uniform temperature T0
the material to be heated.
can be assumed within the food sample as initial condition. On
The mathematical description of heat transfer within the
boundary conditions, convective heat transfer from the external
food product placed between the electrodes, is given by unsteady
surfaces, in accordance with the Newton law formulated as fol-
heat-conduction equation (assuming that a solid-like foodstuff is
lows:
processed in the RF applicator) with a generation term, coupled ! !
with the quasi-static electro-magnetic field equations: ¡ k r T  n D h .T ¡ Tair /: (26)
@T ! !
where h is the overall external convective heat transfer coeffi-
rcp D r k r T C P (25)
@t cient (W/m2K), Tair is the air temperature inside the RF cavity
!
(K) and n is the normal vector of the surface.
where r is density (kg/m3), cp is specific heat (J/kgK), k is thermal Moisture transfer can occur during many heating operations
conductivity (W/m K), T is the temperature (K), and t is the time when a certain amount of the energy being transferred to the
(s). While the solution of heat transfer is needed just within the food sample. The amount of thermal energy required to change
sample, the Gauss law must be evaluated for the space between the state of water (that is contained in the food) from liquid to
the two electrodes, which includes the sample and the air around vapor at a given saturation temperature and pressure is called
it (Marra et al., 2007). the latent heat of evaporation. Under such condition, the latent
It must be emphasized that in the whole RF systems (i.e., the heat of evaporation must be supplied. The internal vapor pres-
RF applicator plus the food sample undergoing RF heating) the sure caused by evaporation of the liquid water is generated
displacement of the electromagnetic field is needed (as quickly, and this pressure is possibly the main driving force
described by Eq. (21)), while in the food sample the coupling behind moisture transfer into vapor and/or liquid forms. The
with the heat transfer equation is strictly needed (as described vapor pressure of a liquid increases with increasing tempera-
by Eqs. (24) and (25)). This concept is summarized in the fol- ture. The molecules in the liquid are more energetic at higher
lowing Fig. 4, where—as an example—the case of RF heating of temperatures, and more molecules can escape from the liquid
a food product constituted by two components is sketched, phase into the gas phase. A pressure of 20 kPa and a tempera-
together with the reference to the equations needed to mathe- ture of 60 C correspond to the “liquid-vapor” separate region.
matically describe the RF heating of the considered food prod- Therefore, the driving force of the moisture transfer from the
uct. It is clear from this sketch that, while solving the set of sample to the surrounding air starts to be important when pro-
equations, one has to take into account the physical (density, cesses characterized by temperatures above 60 C (it means that
specific heat, and thermal conductivity) and the dielectric the vapour pressure is higher than 20 kPa) for long time. In
(dielectric constant and loss factor) properties of the material in operations conducted below 60 C and for short processing
times, in a first mathematical model formulation of RF heating,
the moisture transfer and its effect on the heat transfer can be
ignored.
When the moisture transfer from the sample to the sur-
rounding air cannot be ignored, the thermal effects of evapora-
tion must be taken into account at least at the boundaries
between the sample and the surrounding air. On these surfaces,
the internal conduction is balanced by the external convective
heat transfer and by the heat loss due to the evaporation of
moisture, as in the following expression (Marra et al., 2007):

! !
¡ k r T n j boundary ¡ _ w λ j boundary
D ½h.T ¡ Tair / C m C (27)

where m_ w is the amount of water evaporated per unit of timeand

Figure 4. Sketch of a RF heating systems, with a food product constituted by two
substrates (red and orange color regions) surrounded by air (blue region), and ref- surface area (kg/m2 s) and λ is the latent heat of vaporisation per
erence to modeling equations applying in the different regions. unit mass of evaporated water (J/kg).
8 Z. HUANG ET AL.

Of course, the amount of water evaporated per unit of time- evaluated by comparing numerical results with measured
and surface area (m_ w ) could be computed by adding another temperature profiles. Multiphysics phenomena during RF
equation that describes the mass transfer at the boundary heating of a foodstuff, shaped as cube, cylinder, or sphere,
between the food sample and the surrounding air: have been theoretically studied by Romano and Marra
(2008). Tiwari et al. (2011a) employed a finite element
_ w D km ðcws ¡ cw air Þ
m (28) model to explore the effects of sample size, shape, position,
and dielectric properties of wheat flour and the surrounding
where cws and cwair are the mass concentration of water (kg/m3) medium, electrode gap, and top electrode configuration on
at the food external surface and in the air far from the food the RF power uniformity in wheat flour placed into a
sample, respectively and km (m/s) represents a coefficient of 12 kW, 27.12 MHz parallel plate RF system and showed
convective mass transfer. that these factors significantly affected the RF power distri-
A finer modeling approach would also require coupling the bution in RF-treated flour. Then, a sensitivity study indi-
mass transfer (Dhall and Datta, 2011; Pace et al., 2011; Zhu cated that the heating uniformity of the raisins was most
et al., 2015), thus adding another partial differential equation in affected by the density of the samples followed by the top
a porous-media (Datta et al., 2007; Ni et al., 1999) or conjugate electrode voltage, the DPs, the thermal conductivity and the
approach (Marra et al., 2010). heat transfer coefficient (Alfaifi et al., 2014). Simulation
results also demonstrated that the variation in sample den-
sity and specific heat, especially thermal conductivity, had a
2.6. Computer simulation of RF heating in food processing
relatively slight effect on RF heating rate (Huang et al.,
Activities of computer simulation fall into two general cate- 2015b). The RF heating rate was significantly influenced by
gories, the first being the simulation of heat transfer within electrode gap, top electrode voltage, and the DPs and mois-
the product (load) between the electrodes and focusing ture content of the sample (Jiao et al., 2014b).
mainly on the description of transport phenomena inside the Huang et al. (2015c) developed a computer simulation
food. The second area has been modeling RF heating in model using a finite element-based commercial software and
terms its electric and magnetic fields. Table 1 gives a sum- validated it using experimental dry soybeans packed in a rect-
mary of recent development of computer simulation for RF angular container (300 £ 220 £ 60 mm3) and heated in a
heating in food processing and postharvest treatments. It can 6 kW, 27.12 MHz RF system. The experimental results were in
be seen from literatures that the computer simulation could good agreement with the simulation ones, and both showed
be used to understand the mechanisms, test new strategies, higher temperature values in the middle and bottom layers
optimize parameters, and design appropriate RF treatment compared with those of the top layers. Corners and edges were
conditions for specific food products. heated more than center areas in all layers. Therefore, rounding
the edges and corners in the containers or bending the top and
2.6.1. Computer simulation of RF heating and drying bottom electrodes was key methods to control the electrical
The first attempt to model RF systems was reported in the 1990s field inside the RF system (Alfaifi et al., 2016; Huang et al.,
(Baginski et al., 1989, 1990). These efforts attempted to model 2016b). More recently, Uyar et al. (2016) developed a numerical
the electromagnetic and thermodynamic phenomena for indus- model to analyze power absorption, temperature distribution,
trial-scale RF drying chambers. Subsequently, Neophytou and heating rates and heating uniformity in processed food material
Metaxas (1996, 1997, 1998, 1999) recommended that solution of during RF heating when different projection areas and different
Laplace equation is adequate for small-sized applicators, whereas distances between electrodes were considered. Results showed
wave equations should be used for large size electrodes. This that geometrical factors, such as the projection of sample
opened opportunity for analyzing the electromagnetic field dis- exposed surface on electrode surface and the distance between
tribution, heating rate, and heating pattern in foods with FEM electrodes, have a significant influence on heating rate, heating
during RF heating. The application of a three-dimensional (3D) uniformity and power absorption in block-shaped foods under-
finite-element computer program package, TLM-FOODHEAT- going RF heating.
ING, on the simulation of capacitive RF dielectric heating of rad-
ish and alfalfa seeds was investigated by Yang et al. (2003). They 2.6.2. Computer simulation of RF pasteurization and
successfully solved the EM field by the transmission line method sterilization
and the heat diffusion by the standard explicit finite difference In the specific application of RF energy to food sterilization
time domain method. Chan et al. (2004) developed an effective processes, both simulation and experiment results confirmed
model to simulate an actual RF heating cavity using the wave that the hot spots were located at the places near the corner of
equation applied in 3D instead of the conventional electrostatic the sample (mashed potato), and the cold spots were at the cen-
method. Similarly, Jumah (2005) solved the partial differential ter of the sample (Wang et al., 2008a). The temperature differ-
equations for modeling the RF-assisted fluidized bed drying of ences between the simulation and experiment at cold spots
grains. were within 5% (i.e., 6 C at 120 C), which limited the steriliza-
The temperature profiles and the nonuniformity of tem- tion value (F0) error to 3 min. During the food sterilization pro-
perature distribution occurring during RF heating of cylin- cess, due to the exponential nature of the sterilizati on
drical meat batters, were analyzed by mathematical calculation, a small difference at temperatures above 121 C can
modeling of both electromagnetic and thermal phenomena introduce a large variation in sterilization value. For example,
(Marra et al., 2007). The goodness of model fit was an 11% temperature difference at 121 C represents a 13 C
Table 1. Summary of computer simulation for RF heating in food processing and post-harvest treatments.

Temperature dependent
Purpose Food Items Software Technique properties Dimension Affiliation References

Insect control Almonds Visual Basic FEM ¡ ¡ Northwest A&F University, China Wang et al. (2013)
Apple FEMLAB 3.2 FEM x 3D Washington State University, USA Birla et al. (2008a)
Avocado FEMLAB 3.4 FEM x 3D Washington State University, USA Birla et al. (2008b)
Dry soybeans COMSOL 4.3a FEM x 3D Northwest A&F University, China Huang et al. (2015a, c)
Dried nuts Visual Basic FEM ¡ ¡ Washington State University, USA Wang et al. (2003a)
Grapefruit FEMLAB 3.4 FEM x 3D Washington State University, USA Birla et al. (2008b)
In-shell walnuts Visual Basic ¡ ¡ ¡ Washington State University, USA Wang et al. (2005)
Mung beans COMSOL 4.3a FEM x 3D Northwest A&F University, China Huang et al. (2015b)
Orange FEMLAB 3.4 FEM x 3D Washington State University, USA Birla et al. (2008b)
Raisins COMSOL 4.2a FEM x 3D Washington State University, USA Alfaifi et al. (2014)
Stored-grain MATLAB R2011a ¡ ¡ ¡ University of Saskatchewan, Canada Shrestha and Baik (2013)
Wheat COMSOL 4.3a FEM x 3D Northwest A&F University, China Chen et al. (2016)
Wheat flour, FEMLAB 3.4 FEM x 3D Washington State University, USA Tiwari et al. (2011b)
Wheat kernels COMSOL 4.2 FEM x 3D Shanghai Jiao Tong University, China Jiao et al. (2015a)
Pasteurization In-shell eggs COMSOL 3.4 FEM ¡ 3D McGill University, Canada Dev et al. (2012)
or sterilization Peanut butter COMSOL 4.2a FEM ¡ 3D Washington State University, USA Jiao et al. (2014a, 2015b)
Shell eggs COMSOL 4.3a FEM x 3D University of Nebraska – Lincoln, USA Lau et al.(2016)
Soybean flour COMSOL 4.3a FEM x 3D Northwest A&F University, China Huang et al. (2016a)
Wheat COMSOL 4.3a FEM x 3D Northwest A&F University, China Chen et al. (2015b)
Temperature and Alfalfa and radish seeds TLM-FOOD HEATING Finite different time x 2D-axi University of Connecticut, USA Yang et al. (2003)
electrical field analyses domain method
1% of CMC powder HFSS FEM x 3D Washington State University, USA Chan et al. (2004)
Meat batters FEMLAB 3.1 FEM x 3D Universit
a degli studi di Salerno, Italy Marra et al. (2007)
Meat batter COMSOL 3.5 FEM x 3D Universit
a degli studi di Salerno, Italy Uyar et al. (2014)
Meat batter COMSOL 4.3b FEM x 3D Universit
a degli studi di Salerno, Italy Uyar et al. (2016)
Raw potato FAM FEM ¡ 2D-axi University of Cambridge, UK Neophytou and Metaxas
(1996, 1997, 1998, 1999)
Salt solution CCC ¡ x ¡ Washington State University, USA Jiao et al. (2014b)
Thawing Lean beef COMSOL 4.3b FEM x 3D Universit
a degli studi di Salerno, Italy Uyar et al. (2015)
Mashed potato COMSOL 3.2a FEM x 3D Washington State University, USA Wang et al. (2008a)
Tuna FEMAP, PHOTO-Wave-jv FEM x 3D Tokyo University of Marine Science and Llave et al. (2015)
and Photo-Thermo Technology, Japan
software
CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION
9
10 Z. HUANG ET AL.

temperature difference between experiment and simulation, higher after 4 min of RF heating at 27 MHz than that of the
which would cause a significant error in predicting the F0 by walnut kernels. The heating rate for the insect slurry is 1.4–
simulation results. To guarantee the safety of commercially 1.7 times faster than for walnut kernels, confirming that the
sterilized food, it is necessary to assure that the F0 of the cold insects are indeed preferentially heated in walnuts. Therefore,
spot reaches the designated value. when the applied frequency and intensity of the electric fields
According to the recommendations of USDA-FSIS for the are the same both for the pests and the host products, the heat
pasteurization of eggs, egg white must be heated up to 57.5 C, produced in the insects and the host products might be differ-
and the egg yolk has to be heated up to 61.1 C for 2 min (Dev ent due to different loss factor, resulting in different final tem-
et al., 2012). For the individual whole eggs, which heated with RF peratures of the insects and the host products after the same
heating in a vertical alignment in air would form a coagulation period of RF heating (Huang et al., 2015a). To examine the rel-
ring at the interface between the albumen, air cell, and shell ative electric field intensities in the insects (Ei) and the host
before the yolk was sufficiently pasteurized. This phenomenon products (Ep), a theoretical model was developed on the basis
was caused by redirection and concentration of the electric field of interaction of electromagnetic waves with multi-phase mate-
in the shell-air cell-albumen interface due to starkly different rials as following (Shrestha and Baik, 2013):
dielectric properties (Lau et al., 2016). Maxwell’s equations and
Fourier’s equation were solved for obtaining the temperature dis-  
3ep
tribution in eggs and the locations of hot and cold spots were E i D Ep (29)
2ep C ei
determined (Dev et al., 2012). Their results showed that egg white
gets heated up faster than the egg yolk in the parallel plate RF
applicator, which was not desirable to sustain the functional where ei is the relative complex permittivity of the insect (F/m),
properties of the egg. When the individual whole eggs were ep is the relative complex permittivity of host products (F/m),
rotated between the electrodes, the egg yolk was heated more Ei is the electric field within insect bodies (V/m), and Ep is the
than the egg white as preferred. For the egg products in packages, electric field within host products (V/m).
it was possible to achieve more heating of the yolk than the white To analyze the electric field distribution of insects under dif-
when the eggs were kept static and heated in the coaxial cavity ferent conditions, Eq. (29) can be shown as:
RF applicator (Dev et al., 2012). This set up would be suitable for
applications at an industrial scale. Therefore, two types of RF  
Ei 1
applicators namely parallel plate RF applicator and coaxial cavity Eip D D3 (30)
Ep 2 C ei 6 ep
design were simulated and it was found that both the processes
would be suitable for in-shell pasteurization of eggs.
Commercial peanut butter in a cylindrical jar was used as a where Eip is the ratio of E-fields per unit volume in the insect
model of low moisture food subjected to RF heating when the relative to that in host medium.
cold spot location reached the target pasteurizing temperature To investigate the feasibility of RF selective heating of insect
(70 C) in a 6 kW, 27.12 MHz RF system. Based on both experi- larvae in soybeans, Huang et al. (2015a) solved the coupled
ment and simulation results, the cold spot location was found to electromagnetic and heat transfer equations for developing a
be at the center of top and bottom surfaces for samples with or simulation model. Simulated and experimental results both
without PEI sheets. Their results indicated that the use of the showed that cold spots were located at the center part of each
PEI addition method has improved heating uniformity by reduc- layer. After 6 min RF heating, the mean temperature differences
ing the maximum temperature difference from 28 to 18 C in the between insects and soybeans at the top, middle, and bottom
peanut butter and therefore enhanced the potential for pasteuri- layers were 5.9, 6.6, and 6.2 C, respectively. Their results
zation of low moisture foods heated in RF systems. After that, revealed that the heating rate of insects was 1.4 times greater
Jiao et al. (2015b) showed that a pair of PEI blocks with a than that of soybeans. The simulated selective heating of insects
diameter of 8 cm among all five diameters (2, 4, 6, 8, and 10 cm) in soybeans may provide potential benefits in developing prac-
added to the cold spots of a given peanut butter sample could tical RF treatments to ensure reliable control of insect pests
make the sample reach the best heating uniformity. Further- without adverse effects on product quality.
more, the best height of PEI blocks with a diameter of 8 cm that
allows the sample to be heated most uniformly was found to be 2.6.4. Computer simulation of RF thawing
1.3 cm after sweeping from 0.1 to 2.3 cm with a step of 0.1 cm. The coupled electro-thermal problem of modeling an RF process
becomes more complicated for simulation of thawing since the
2.6.3. Computer simulation of RF disinfestations phase change process requires dealing with evolving large latent
Differential heating or selective heating is a main advantage of heat over a small range of temperature. Llave et al. (2015) con-
RF heating as compared to the conventional and MW heating structed a 3D geometric model using FEMAP, the boundary and
for disinfestations. To provide a theoretical basis and experi- loading conditions, electric intensity, and other parameters used
mental evidence to support the hypothesis that insect larvae in the model were set in Photo-Wave-jv for analysis of the elec-
can be preferentially heated in food products by RF heating for tromagnetic field, and set in Photo-Thermo for the heat transfer
pest control. Wang et al. (2003a) developed a mathematical analysis to estimate the temperature distribution. To validate the
model based on heat transfer and dielectric heating to predict simulated results, the tuna samples were thawed from ¡60 to
preferential heating of insect pests in dry nuts. The model ¡3 C (center temperature) using a 13.56 MHz parallel plate RF
results showed that the temperature of insects is 14.3 § 1.1 C system and by thawing in air. The validated model was then
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 11

successfully applied to studying the effects of electrode size on serious concerns in meat products because of the survival of the
temperature uniformity in the RF thawing of frozen tuna muscle. bacteria resulting in health hazards (Rincon et al., 2015).
Uyar et al. (2015) also developed a computational model in a A 3D finite element model of heat transfer in food material
3D domain with temperature-dependent thermo-physical prop- during RF heating was developed by Alfaifi et al. (2014).
erties (TPs) and DPs to improve the heating uniformity and to Experiments were conducted using raisins as test material
minimize runaway heating in a 50-V and a free running oscilla- packed in a rectangular plastic container (25.5 £ 15.0 £
tor RF system. The boxed frozen lean beef sample (3.84 kg), 10.0 cm3). Both results showed that the temperatures were
shaped as a parallelepiped (200 £ 200 £ 100 mm3) was placed highest in the middle layer (45–76 C), whereas they were lower
in a polypropylene container (not immersed in water) on the in the top and the bottom layers (41–62 C and 40–57 C).
center of the bottom electrode during RF experiments. Nonuni- Overheating was observed at the edges and corners of the top
form temperature distribution during thawing, especially high (51 and 62 C), middle (65 and 76 C), and bottom (48 and
temperatures encountered along the surface and corners of the 57 C) layers (Fig. 5). According to Alfaifi et al. (2014), over-
product, is a major disadvantage of a RF thawing system. As heating behavior at the edge and corners could be attributed to
demonstrated by Uyar et al. (2015), during defrosting, thawing of the refraction and reflection of the electric field at interfaces,
outer surfaces occurs at the initial stages, while still frozen inside resulting in a higher volumetric power density at these posi-
as parts are being surrounded by a low thermal conductivity layer tions. The temperature distribution patterns were similar to
leading to longer thawing times. Therefore, the thawing times are those reported by Tiwari et al. (2011b) and Huang et al.
always higher for the larger sizes, and this issue makes the RF (2015c). Tiwari et al. (2011b) claimed that the temperature val-
processes an unavailable industrial process. Therefore, it is neces- ues were higher at the middle layers and temperature at the top
sary to systematically study the RF thawing characteristics and surface was lower than that in the middle layer because of a
evaluate treatment parameters to improve the RF thawing unifor- large evaporation at the top surface. It is clear from the litera-
mity in food industry based on the validated computer model. ture that with the help of simulations, it may be possible to find
new ways to overcome nonuniform RF heating and speed up
the incorporation of RF heating in the food industry with
3. Computer simulation for evaluating RF heating acceptable heating uniformity and product quality.
uniformity in food processing
3.1. The nonuniform RF heating characteristics
The sheer number of research on using RF heating in food appli- 3.2. Thermo-physical and dielectric properties of food
cations gives a wealth of information for understanding the pro- products
cess better (Fu, 2004; Hou et al., 2016; Marra et al., 2009; Piyasena
To model the RF heating process, it is necessary to know the
et al., 2003). However, wide temperature variations were observed
DPs and TPs of the food material. A review of DPs and TPs
within heated samples during RF heating of different kinds of
measurement techniques was given by Venkatesh and Raghavan
fruits (Alfaifi et al., 2014), grains (Jiao et al., 2015a), meat (Llave
(2005). DPs and TPs data for foodstuffs used in computer simu-
et al., 2015; Zhang et al., 2004), model foods (Birla et al., 2008a),
lation at RF frequencies over the range of treatment tempera-
ready-to-eat meals (Orsat et al., 2001), vegetables (Liu et al.,
tures were taken from the literature and summarized in Table 2.
2015), and various agricultural products (Fu, 2004; Gao et al.,
2010). Moisture accumulation at the surface of food during RF
heating was studied by Datta and Ni (2002). The temperature dif-
ference between the hottest and the coldest spots in mashed pota- 3.2.1. For pulverized or solid foods
toes (295 £ 235 £ 42 mm3) was greater than 30 C, after 320 s of For pulverized or bulk materials, the overall volume of samples
RF heating (Wang et al., 2008a). Cold and hot spots occur in food was simplified as a whole in the simulation model, which ignor-
load during dielectric heating due to nonuniform electrical field ing the influence of air gap between the food samples. DPs and
distribution. Temperature variation over the cross section of apple TPs of food products were assumed to be homogeneous and iso-
was distinct with contour of maximum temperature (80 C) at tropic, the density was assumed as temperature independent and
fruit core and minimum temperature (55 C) at surface from the other properties were temperature dependent. It should be noted
simulation results (Birla et al., 2008b). Both simulated and experi- that this method is complementary to ensure good model accu-
mental results showed that immersion of the center placed model racy and maintain convergent simulation results. A considerable
fruit in water slightly shifted the hot spot toward the core of the amount of work on DPs of food and agricultural products has
modelfruit. Moreover, the presence of water radically enhanced been published at RF frequencies (Hou et al., 2016; Nelson, 2008;
power coupling as it took only half time (7 min) what required Piyasena et al., 2003). These data were subjected to linear regres-
for RF heating of the fruits in air to reach »50 C temperature. sion analysis for using these properties in simulation model over
Under similar conditions, very high electric field concentrations the treatment temperatures range (Alfaifi et al., 2014).
occurred at the corners and edges of the cube, edges and the mid- For some solid foods, due to the fact that phase change in
dle of the cylinder, whereas maximum E-field was at the bottom some food substances occurs during RF thawing over a range
of the sphere (Birla et al., 2008a). Tiwari et al. (2011b) reported of temperature and large latent heat evolves over this range,
that nonuniform temperature distribution may cause quality loss special techniques were required to deal with such problems
due to over or underheating in different parts of a food product. (Pham, 2006). Liu et al. (1999) reported the possible techniques
Nonuniform temperature distribution resulting in cold spot has to determine the specific heat, density, and thermal
12 Z. HUANG ET AL.

Figure 5. Simulated (a) and experimental (b) temperature distributions of top, middle, and bottom layers of raisin samples placed in a polypropylene container
(255£150£100 mm3) in the center and middle between the top and bottom electrodes of the RF system after 4 min heating at a fixed electrode gap of 13.6 cm and
initial temperature 23 C (Alfaifi et al., 2014).

conductivity of food products by the following equations: is the changes in thermal conductivity per unit change in tem-
8 9 perature (W/mK2), re (kg/m3), ce (J/kgK), ke (W/mK), D (J/
< r1 ; TTsh = Knkg), and n are the empirical constants in Eqs. (31), (32),
rD .Tsw ¡ Tsh / (33), (32) and (32), respectively.
: re C Sd .Tsh ¡ T/ C .r1 ¡ re / ; T < Tsh ;
.Tsw ¡ T/
(31) 3.2.2. For granular and porous foods
8 9 Both granular and porous foods are two-media systems com-
< c1 ; TTsh = posed of air and solids. Air has different DPs values from those
cp D D (32)
: ce C ; T < Tsh ; of a solid food, and a change in bulk density may lead to a
.Tsw ¡ T/n change in the DPs of food. The higher the bulk density, the
8 9 higher the dielectric constant and loss factor observed, which
< k1 ; TTsh = means more amount of mass interacting with the electromag-
kD .Tsw ¡ Tsh / (33) netic fields (Nelson and Trabelsi, 2012). This is especially
: k D ke C Sk .Tsh ¡ T/ C .k1 ¡ ke / ; T < Tsh ;
.Tsw ¡ T/ important for particulate materials and porous-packed samples.
where Tsw is the normal freezing temperature for pure water It is difficult to directly measure the bulk DPs and TPs for such
(K), Tsh is the initial freezing point of the food (K), Sd is the materials (Içier and Baysal, 2004a, b). Nelson (2008) proposed
changes in density per unit change in temperature (kg/m3K), Sk several dielectric mixture equations to calculate the DPs of air
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 13

particle mixtures. The Landau and Lifshitz, Looyenga equation density (Q, W/m3) divided by material volume (Vvol, m3).
(LLLE) is described as (Liu et al., 2009):
1 R
Qav D QdVvol (37)
1 1 1 Vvol Vvol
e3 D va .ea /3 C vs .es /3 (34)
Besides the changes in the processed and absorbed powers,
where e is the complex permittivity of the mixture, ea is the temperature-related calculations were other approaches to
complex permittivity of the air, es is the complex permittivity determine the temperature distribution uniformity in the
of solid, va is the volume fraction of the air, and vs is the volume treated products. The heating uniformity of the treated samples
fractions of particles, where va C vsD 1. Dielectric mixture was evaluated using the heating uniformity index (λ). It is
equations have been used successfully to predict the DPs of var- defined as the ratio of the rise in standard deviation of sample
ious bulk materials that consist of air voids and solid particles temperature to the rise in average sample temperature during-
(Kim et al., 1998; Liu et al., 2009). The calculated DPs of bulk treatment and can be calculated by the following equation
materials were reduced as compared to these properties at par- (Wang et al., 2005):
ticle density (0% air) due to the presence of air (Alfaifi et al., pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2014; Jiao et al., 2015a). Ds s 2 ¡ s 20
The Kopelman model mixture equation was used to estimate λD D (38)
Dm m ¡ m0
the thermal conductivity for granular and porous foods (Alfaifi
et al., 2014):
where s 0 and s are the standard deviations of the initial and
final temperature distributions, respectively. m0 and m are the
2=
! means of the initial and final temperature distributions, respec-
1 ¡ va 3 .1 ¡ .ka 6 ks // tively. The smaller the λ value, the better the RF heating unifor-
ke D ks 2= 2=
(35)
1 ¡ va 3 .1 ¡ .ka 6 ks //.1 ¡ va 3 / mity. Heating uniformity index was used for evaluating RF
heating uniformity in almond (Gao et al., 2010), corn (Zheng
et al., 2016), coffee bean (Pan et al., 2012), legumes (Jiao et al.,
where ke is the effective thermal conductivity for the bulk mass 2012; Wang et al., 2010), rice (Zhou and Wang, 2016), and wal-
(W/m C), ks and ka are the thermal conductivity of solid and nut (Wang et al., 2007a).
air (W/m C), and va is the volume fraction of the air. This Another formulation of a temperature uniformity index
equation has been successfully used to determine the thermal (UI) of the treated samples in simulation is given by the follow-
conductivity of tortilla chips (Moreira et al., 1995) with an error ing equation (Alfaifi et al., 2014):
range of §4.3–9.1%, and that of French fries (Sahin et al., 1999)
with an error range of §2.5–8.3%. R pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
Vvol Vvol .T ¡ Tav /2 dVvol
UI D (39)
Tav ¡ Tinitial
3.3. Criterion to evaluate RF heating uniformity
The well-developed simulation model makes it possible to where Vvol is the material volume (m3), T and Tav are local and
obtain convergent results from various agricultural products. average temperatures (K) inside the dielectric material over the
Different criteria and indexes have been used to study, evaluate, volume (Vvol, m3). This index is a useful tool to evaluate the
and compare the RF heating uniformity in food products. The heating uniformity when using a fixed configuration and a spe-
uniformity of temperature distribution in an RF processed sam- cific RF unit. The smaller UI values indicate an improved RF
ple is a significant parameter and especially affected by the heating uniformity.
absorbed power of the sample. Analysis of the quantity of the A new temperature uniformity index (TUI) was proposed
absorbed power by the sample during a RF heating process was based on Alfaifi’s UI by replacing the average temperature (Tav)
conducted by Neophytou and Metaxas (1998). Tiwari et al. with the target temperature (Tt) in the following form (Jiao
(2011a) defined the power uniformity index (PUI) where the et al., 2015b):
smaller PUI value demonstrates an improved power uniformity
in the processed sample, and a well-designed RF system was R
Vvol j T ¡ Tt j dVvol
reported to result in lower PUI values: TUI D (40)
.Tt ¡ Tinitial /Vvol
R
Vvol sqrt..Q ¡ Qav / /dVvol
1 2
Vvol where Tt is the target heating temperature (K). Also in this case,
PUI D (36) a smaller index corresponds to better heating uniformity. The
Qav
new TUI might be more suitable for describing the heating uni-
formity of a pasteurization/sterilization process, which requires
where the minimum value of PUI is zero, which indicates uni- the cold spot location to reach a certain target temperature
form RF power in dielectric materials. needed for controlling pathogens. It would reflect the degree to
The average RF power density (Qav, W/m3) in a dielectric which temperature in the volume deviated from the target
material is defined as the volume integral of the RF power temperature.
 14

Table 2. DPs and TPs of various food products and materials used in mathematical modeling.

Product Process Specific heat Thermal conductivity Density Electrical

conditionT- ( C) Dielectric constant Loss factor (J/kg C) (W/m C) (kg/m3) voltage (V) Reference
2 0.465 ¡6 3 ¡6 2 ¡4
Albumen TD20–70 0.0053T -0.45TC 90.67 205.6T - 326.7 6.53£10 T 2.50£10 T C 3.63£10 T C 1045 8000 Lau et al. (2016)
Z. HUANG ET AL.

¡7.87£10¡4T2 0.54
C2.14£10¡2TC3.42
Apple (green) TD20–60 ¡0.152T C 75.7 2.857T C 60.2 3700 0.422 790 ¡ Birla et al.(2008b)
Apple (red) TD20–60 ¡0.196T C 78.52 2.179T C 45.68 3600 0.513 840 ¡ Birla et al. (2008b)
Avocado TD20–60 0.63T C 104.36 14.47T C 393.94 3380 0.42 1060 9500 Birla et al. (2008b)
Beef meats TD-18-(C10) ¡18T<-10:-0.108T2-1.91TC19.75 ¡18T<-10:-1.08T2C0.09TC61.75 TTm1: 1935.2 0.47–0.152 T<Tm1: 961 141.42 Uyar et al. (2015)
-10T1:0.699T2C12.33TC81.21 -10T- Tm1TTm2:153016.3 Tm1TTm2:
1<T10:-0.189T2C2.437TC72.37 1:3.647T2C62.15TC307.15 T>Tm2:3497.4 1007
-1<T10:- T>Tm2: 1053
0.112T2C7.092TC255.73
Butter TD20–60 2.3 1.0 2720 0.29 911 ¡ Chen (2015a)
Cherry TD20–60 ¡0.046T C 92.44 6.836T C 159.2 3643 0.511 1010 ¡ Birla et al. (2008b)
Cheese TD 20–80 54.6 C 0.28T 160.6 C 11.2T 3700 0.48 1013.2 ¡ Wang et al. (2012)
Gellan gel TD20–60 ¡0.21T C 86.76 4.36T C 129.4 4160 0.53 1010 9500 Birla et al. (2008b)
Grapefruit TD20–60 0.17T C 85 5.01T C 95.21 3703 0.54 950 9500 Birla et al. (2008b)
Lean tuna TD ¡20-(C10) ¡20T<-5:0.06T2C2.62TC31.5 ¡20T<- TTsh:3650 TTsh:0.523 1070 ¡ Llave et al. (2015)
2
-5T1:-11.364T -2.454TC291.64 5:0.338T2C10.886TC84.65 T<Tsh:cpD1179 T<Tsh:kDkeC0.01518(-1-T)C
1<T10:-2.23TC278.56 -5T1:-20.42T2- C173900/(Tsw-T)¡1.628 (k1-ke)¢(TswC1)/(Tsw-T)
2.026TC540.55
1<T10:-4.23TC524.93
Mashedpotato TD 20–80 83.3–0.14T 173.2–16.4T 3763 0.548 1000 ¡ Chen et al. (2013);
Wang et al. (2008a)
Meat batters TD 20–80 0.0011 T2- 0.5531T ¡0.0058 T3 C 5.32 T2 ¡1609.1 T C 3600 0.01088T-0.0848 1057 173.21 Marra et al. (2007);
C 109.43 161643 Uyar et al. (2016);
Romano and Marra
(2008)
Meatballs TD 20–80 62.6 C 0.29T 198.4 C 12.6T 3600 0.48 1144.7 ¡ Wang et al. (2012)
Mung beans TD20–60 0.021T C 1.82 0.0028T C 0.036 18T C 1165 0.102 953 8400 Huang et al. (2015b)
Noodles TD 20–80 98.9–0.34T C2£10¡3T2 203.1 C 13.0T 3690 0.52 977 ¡ Wang et al. (2012)
Orange (pulp) TD20–60 ¡0.22T C 88.6 4.9T C 122.6 3661 0.580 1030 9500 Birla et al. (2008b)
Orange (peel) TD20–60 3.94T C 58.2 ¡0.16T C 82.53 3300 0.40 800 9500 Birla et al. (2008b)
Pasta TD20–60 46.0 15.4 2460 0.50 1050 ¡ Chen et al. (2015a)
Peanut butter TD20–80 4.03 0.4 2030 0.209 1115 12,100 Jiao et al. (2014a,
2015b)
Polystyrene TD20 2.6 0.0003 1300 0.036 25 ¡ Huang et al. (2016b)
Polypropylene TD20 2.0 0.0023 1800 0.2 900 ¡ Alfaifi et al. (2014)
Polyetherimide TD20 3.15 0.0025 2000 0.122 1270 ¡ Jiao et al. (2014a)
Potato TD 20–70 50 15 3900 0.4 1000 ¡ Geedipalli et al. (2007)
Raisins TD20–60 0.08T C 6.3 0.02T C 1.9 10.9T C 1831 0.001TC 0.15 784 4100 Alfaifi et al. (2014)
Sauce TD 20–80 101.8–0.932T C 1.16£10¡2T2- 418.3 C 25.9T 3730 0.51 905.6 ¡ Wang et al. (2012)
4£10¡5T3
Sauce TD20–60 69.4 25.2 3000 0.50 1050 ¡ Chen et al. (2015a)
Shell TD20–70 3.081T0.3072 C 4.45 216.9T0.04521- 236.8 8.52£10¡7T3- 8.50£10¡5T2 ¡1.03£10¡3T C 2300 8000 Lau et al. (2016)
1.30£10¡4T2 0.68
C8.64£10¡3TC0.90
Soybeans TD20–80 3.6 0.26 1737 0.11 748 5000 Huang et al. (2015c)
Soybean flour TD20–60 3.96 0.38 5.8T C 1614 0.0007T C 0.083 380 4050 Huang et al. (2016a)
Tap water TD20–60 ¡0.48T C 84.74 0.33T C 11.1 4180 0.56 1000 9500 Birla et al. (2008b)
Walnut TD 20–60 0.01T C 4.7 0.0003 T2 ¡ 0.0289T C 1.08 2510 ¡ 900 ¡ Wang et al. (2003a)
Whey protein TD20–60 0.28T C 93.56 17.51T C 468.23 3850 0.55 1050 9500 Chen et al.(2013)
gelb
Wheat TD20–60 4.30 0.11 2670 0.15 860 15,000 Chen et al. (2015b)
Wheat kernels TD20–80 0.002T2-0.16TC5.097 0.001T2¡0.091TC1.695 13.158TC1537.4 0.0031TC0.0506 803 5100 Jiao et al.(2015a)
Wheat flour TD20–70 0.0007T2-0.0345TC3.72 0.33 23T C 757 1.36£10¡4T2-0.0094T C 0.2819 800 13,000 Tiwari et al. (2011b)
Yolk TD20–70 12.82T0.1975 C 28.24 278.4T0.2599- 431.9 4.12£10¡6T3- 1.07£10¡6T2 C1.41£10¡4T C 1148 8000 Lau et al. (2016)
5.75£10¡4T2 0.35
C2.37£10¡2TC2.50
CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION
15
16 Z. HUANG ET AL.

3.4. Factors influencing RF heating uniformity factor materials than that in high loss factor materials (Wang
et al., 2008b). Their results also showed that the voltage of the
During RF processing, several interacting factors influence
top electrode is an important factor to influence the electro-
heating uniformity. These factors include the design of RF heat-
magnetic field intensity and final temperature distributions in
ing systems (e.g. the inductance positions and feeding strips,
samples. Huang et al. (2015b) reported that a higher top elec-
electrode shape and electrode configuration), packaging geome-
trode voltage resulted in faster RF heating in mung beans dur-
tries, DPs, and TPs of the treated materials, position of the
ing 6 min RF heating. With the important effect on RF heating
treated materials within the RF units, and the surrounding
rates, it is essential to precisely determine the top electrode volt-
media (Fu, 2004). The following subsection summarizes various
age by direct measurement (Zhu et al., 2014).
factors that influence RF heating uniformity.
3.4.3. Nonuniformity due to electrode configuration
3.4.1. Nonuniformity due to electromagnetic field The effect of the top electrode bending and angles on RF heat-
The nonuniform electromagnetic field distribution is the major ing uniformity of samples was evaluated by Tiwari et al.
factor resulting in nonuniform heating during RF treatments (2011a). Results demonstrated that the top electrode bending
(Wang et al., 2007a). Many factors are included in the design of position greatly affected the sample PUI as magnitudes of the
RF heating systems, such as the voltage of the top electrode, elec- electric fields and their distributions inside the sample changed
trode shape, and power output, shape, TPs and DPs of the sam- with the top electrode bending positions and angles. Optimum
ples, and position of the samples in the RF units, which may RF power uniformity in a particular sample size could be
influence electromagnetic field distributions in RF-treated agricul- achieved with a particular top electrode bending position and
tural products (Fu, 2004; Marra et al., 2009; Tiwari et al., 2011a). angle (Tiwari et al., 2011a). The effect of different combinations
The electric field between two parallel plate electrodes was in of the setback distances (distance between the edge of the sam-
parallel lines uniformly spaced throughout the region between two ple and the bottom base and the top of the new designed elec-
electrodes, and perpendicular to their surfaces with no dielectric trodes) and bend angles on the heating uniformity of raisins
material placed in it (Huang et al., 2016a). Fig. 6 shows a typical packed in a rectangular-shaped container after RF heating to a
electromagnetic field distribution for a rectangular-shaped dielec- central temperature of 52 C was explored by Alfaifi et al.
tric material placed between two parallel plate electrodes. When (2016). These configurations were applied to both the top and
the dielectric material was placed at the center of the bottom elec- bottom electrodes, and were attached to the flat top and bottom
trode with an air gap over it, the presence of the surrounding air electrodes. They reported that treating materials with different
causes an intensification of the heating near the top of the elec- DPs, TPs and different container dimensions may require dif-
trode space at the edges of the material, forming hot spots and ferent electrode configurations. Besides, Wang et al. (2008b)
cold spots. Samples placed in the middle of RF electrodes showed also reported that the asymmetric connections of the inductor
higher electricfield intensity at their central section as electric field and the power supply to the top electrode resulted in an uneven
deflected by both (top and bottom) edges with increased electric standing wave pattern across the applicator. Therefore, a new
field concentration at the central parts of the sample (Uyar et al., symmetric design was developed by installing four conductors
2015). Therefore, it is essential that the electromagnetic field at the four corners on the top electrode and moved the feed
within food geometry should be uniform to ensure even heating. strip to the center. The four corner inductors allowed indepen-
To achieve this, some additional methods (electrode attachments, dent adjustment of the voltage at each corner.
sample movement, and/or mixing) could be used to ensure RF
heating uniformity in industrial applications. According to Alfaifi 3.4.4. Nonuniformity due to electrode geometry and
et al. (2016) and Huang et al. (2016b), the uniformity of electro- position
magnetic field for a rectangular shaped food sample was improved During RF processing, geometrical factors, such as distance
by rounding corners and reducing edges of the package geometry, between electrodes and projection of top electrode on the sam-
and modifying electrode configurations. ples’ exposed surface area (area that electric field passes
through the sample) have a certain effect on heating uniformity
3.4.2. Nonuniformity due to RF power and top electrode in the samples (Uyar et al., 2016). Effect of the size of the pro-
voltage jection of the top electrode on temperature distributions of
Marra et al. (2007) studied temperature distributions in cylin- tuna after RF thawing was evaluated by Llave et al. (2014).
drical meat batters during RF treatments at different output They reported that a more uniform temperature distribution
powers and heating rates by solving the electromagnetic and was found with an electrode of the same size as the sample. The
heat equations using the finite element analysis. The authors electrode size larger than the upper surface of the sample
reported that using lower output power resulted in more even resulted in higher temperatures at the edges of the sample and
temperature distributions. On the other hand, Romano and the use of an electrode of smaller size than the sample showed
Marra (2008), by a mathematical model for simulation of RF an irregular temperature distribution (Llave et al., 2015).
heating using meat batter as the sample food, reported that the Huang et al. (2015c) also reported that when the top plate area
higher the applied power, the more uneven the temperature was changed to 1.32 times of the sample size, the electric field
distribution. Based on vertical temperature measurements, the started entering normally into the sample, and thus the hot
temperature gradient across the 1% CMC solutions was larger spot in corners and edges of sample would be largely reduced.
than that observed in tap water due to RF power decay, suggest- According to Tiwari et al. (2011a), a proportion of oblique
ing that more uniform heating can be achieved in low loss electric field, which caused the RF power nonuniformity,
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 17

Figure 6. The electric field strength between two parallel plate electrodes with (a) no dielectric sample, (b) dielectric sample placed on the center of bottom electrode, (c)
and dielectric sample placed in the center and middle between top and bottom electrodes (Huang et al., 2016a).

decreased when the electrode gap was reduced. However, an Romano and Marra (2008). Among the sample shapes investi-
extremely reduced air gap might lead to runaway heating, which gated by the authors, regular cubes were found to be more suit-
takes place when the warmest part of the product gets more able for RF treatment since cubic-shaped products exhibited a
power compared to colder parts due to higher rates of absorp- fast and more uniform heating, with a good absorption of
tion. Uyar et al. (2014) showed that increased air gap between power (Tiwari et al., 2011a; Uyar et al., 2014). In case of cylin-
electrodes and sample surfaces with decreasing volume led the drically shaped products, authors recommended a vertical ori-
deflection of electric field to reduce its concentration inside the entation during treatment, since horizontally oriented cylinders
sample. Due to the significant effect of air gap as indicated showed a slower heating, characterized by uneven temperature
above, when load sizes got smaller with the increase in the air distribution (Wang et al., 2006b). Eventually, spherical shapes
gap, a major part of the electric field passed through the ground were found to be the less favored to RF heating (Birla et al.,
electrode without absorption by the sample. Therefore, Jiao et al. 2008a).
(2014b) corroborated the fact that the RF heating rate decreased Simulated results conducted by Tiwari et al. (2011a) in dry
as the electrode gap increased. For a fixed sample height, the air food materials showed that the RF power uniformity should be
gap increases and the heating rate decreases with an increasing better for sample sizes, either approaching zero or approaching
electrode gap (Huang et al., 2015b). When the air gap is reduced the maximum possible size that can be kept completely between
to zero, the value of e 0 does not influence the heating rate. As a the RF electrodes. When sizes approached to zero, the major
result, e00 is the dominating factor in RF heating, and the heating part of the electric field passed directly through the ground
mechanism changes from dielectric heating to resistive heating electrode, without entering into samples. With increase in sizes,
(Metaxas, 1996). most of the electric field entered obliquely as sizes were com-
paratively smaller than those of the upper electrode size (Jiao
3.4.5. Nonuniformity due to food geometry and position et al., 2015a). Oblique electric field increased the RF power
Effects of load shape, orientation, and position on the heating nonuniformities in samples (Birla et al., 2008a). With a further
patterns during RF treatments were also investigated by increase in sample size, the electric field started entering
18 Z. HUANG ET AL.

normally (except a small area, which was getting deflected by lethality are achieved there (Jiao et al., 2014a). Huang et al.
sample edges and corners) into the sample (Marra et al., 2007). (2015c) corroborated the fact that the higher moisture contents
This caused increase in the RF power uniformity again. As along with the larger DPs and thermal conductivity, resulting
demonstrated by Uyar et al. (2014), while the load volume in poor RF heating uniformity. Thus, dry products should pro-
changes from 100 to 1%, volume of air gap between electrodes vide a better uniform RF heating because of the smaller DPs
and top-bottom surfaces of the sample increased. The smaller (Tiwari et al., 2011a) as compared to high moisture food.
the load volume, the larger the air gap and the slower the heat-
ing rate. In a fixed air gap between electrodes and sample, the 3.4.7. Nonuniformity due to properties of the surrounding
smaller the sample volume, the faster the heating rate of the medium
sample. So that smaller volumes demonstrated higher tempera- Birla et al. (2008a) reported that the nonuniform temperature
ture increase. According to Uyar et al. (2016), the effect of a distribution in fresh fruit subjected to RF heating was caused
decreasing sample volume was demonstrated to have a negative by the different DPs of food and the surrounding medium
effect on the temperature evaluation if the electrodes were (usually air), which results in an unevenly distributed electric
maintained at a constant distance. As a result, the wider projec- field. It has been reported that altering the DPs of water by
tion area and shorter distance between electrodes led to faster addition of salt minimizes the differential heating between the
heating, in terms of average temperature, but less uniform tem- fruit and water (Birla et al., 2008a). Therefore, Wang et al.
perature distribution. The temperature distribution should be (2008a) conducted a study for pre-packaged mashed potatoes
more uniform when samples were placed in the middle of RF heated in RF systems with circulating water to remove the heat
electrodes as electric field deflected by top and bottom edges accumulation at the edges. Water with various electrical con-
increased net electric field concentration at the central part of ductivities was tested for heating rate, and the highest electrical
samples (Huang et al., 2015c; Jiao et al., 2015a; Marra et al., conductivity (220 S/m) can reduce the hot and cold spot tem-
2007). Similar results have been obtained by Birla et al. (2008a) perature difference from 30.9 to 24.2 C, and 22.4 to 13.6 C on
and Tiwari et al. (2011a) when the small size sample was placed different mashed potato samples. When the electrical conduc-
in the middle of RF electrodes, and the power uniformity would tivity of surrounding water increases, water should absorb most
be clearly improved. energy and food must be absorbing less energy, this could
explain why the heating uniformity was improved. Tiwari et al.
3.4.6. Nonuniformity due to food DPs and TPs (2011a) reported that the PUI was the lowest when the sur-
During RF heating, two physical factors, temperature and elec- rounding material dielectric constant (between 8 and 11) was
tric field intensity, interrelate with each other. In particular, the approaching to the dielectric constant value of wheat flour
dissipated energy provided by electric fields heats those materi- (fixed as 8). Theoretically, the dielectric constant determines
als with temperature-dependent DPs. The variation, caused by the electric field distribution when the loss factor is far smaller
temperature change and heat transfer, influences the electro- than the dielectric constant (Metaxas, 1996).
magnetic field distribution. Hence, TPs and DPs of the product The maximum temperature difference in the peanut butter
are the key factors to influence the uniformity of temperature was reduced from 28 to 18 C after using PEI assisting method,
distribution. Romano and Marra (2008) reported that the loss which has the closest dielectric constant (3.15) to that of peanut
tangent first increased and then decreased with the increase in butter (4.03) and a lower dielectric loss factor (0.0025) (Jiao
temperature of luncheon roll batter. Increasing the loss tangent, et al., 2014a). Huang et al. (2015c) also showed that the temper-
the dielectric absorption increases with the power absorbed ature uniformity could be achieved when the surrounding
within the domain. Wang et al. (2008a) also showed that the material dielectric constant (3) was approaching to the soybean
decreased electric field intensity caused by the increased loss samples (3.6). After that, Huang et al. (2016b) demonstrated
factor not only compensated for the influences brought by the that the heating uniformity was greatly improved by placing
increase in loss factor, but also influenced by the thickness of soybean samples in a polystyrene container, which had the
mashed potato and circulating water and their DPs and electric closest dielectric constant (2.6) to that of soybeans (average
conductivity, respectively. value of 2.5 and 2.7 with moisture contents of 4.64 and 7.86%
Birla et al. (2008b) demonstrated that increasing the e00 value w.b.) and a lower dielectric loss factor value (0.0003). There-
of the model fruit resulted in increasing heating rate only to a fore, the electric field distortion could be reduced by using a
certain value. This was contrary to a general belief that increas- surrounding material with a low loss factor and a similar dielec-
ing loss factor constantly increases heating rate inside the fruit. tric constant, instead of air (Huang et al., 2015c; Jiao et al.,
Beyond that value, increasing loss factor reduced the heating 2014a; Tiwari et al., 2011a). On the other hand, the container
rate inside the fruit. Similar results demonstrated that a loss material, thickness, and corner radius were found to have a sig-
factor equal to the dielectric constant provided fast RF heating nificant effect on heating uniformity of soybeans during RF
(Huang et al., 2015b; Jiao et al., 2014b). Rapid heating rates cor- processes (Huang et al., 2016b).
respond to higher throughputs but may adversely affect the
heating uniformity. Therefore, the differences in DPs and TPs
3.5. Solutions proposed for overcoming nonuniform RF
among the components in a food may cause differential heating
heating
and thus temperature nonuniformity (Jiao et al., 2015b; Nelson,
2008). This may bring about a food safety issue for pasteuriza- With the help of computer simulations, there have been various
tion and sterilization applications, and attention must be paid methods developed, accidental or on purpose, for various food
to find the cold spot(s) and make sure enough heating and products to improve the uniformity of RF heating (Table 3).
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 19

The results of the various studies for various types of food too) to approximately match the DPs of the food. The water
materials, such as agricultural products (Huang et al., 2016b; and package present a very flat surface to the imposed field,
Jiao et al., 2015a), fresh fruits (Birla et al., 2008a; Wang et al., and push the boundary with air away from the food package. A
2006a), low moisture foods (Alfaifi et al., 2016; Chen et al., pressurized vessel was developed to provide an overpressure of
2016), meat and fish (Llave et al., 2015; Uyar et al., 2016) to up to 0.276 MPa gauge (40 psig) that allows foods in large poly-
overcoming nonuniform RF heating are discussed in this meric trays to be heated up to 135 C without bursting. The
section. using of a pressure vessel in which a tray was immersed in cir-
culating water enabled the heating of products to 121 C utiliz-
3.5.1. For meats and seafood ing RF energy (Wang et al., 2003b). The pilot-scale RF system
A number of researches have been conducted for overcoming operated at 27.12 MHz, 6 kW, and the pressure vessel was
nonuniform RF heating of meats and seafood, such as modify- sandwiched between parallel electrodes (Wang et al., 2008a).
ing electrodes, making the size of sample similar to that of the The salt was added to closely match the sodium content of the
top electrode, and placing the sample in the middle of two elec- packaged macaroni and cheese. With the assistance of circulat-
trodes. Most of the recent studies focused on the defrosting and ing water at controlled temperatures, similar heating histories
thawing of meat and fish using RF heating, with successful in model foods (20% whey protein gels, macaroni and cheese)
results compared to those of conventional methods (Uyar et al., were observed at different locations inside a tray. According to
2014). As described by Uyar et al. (2014), a cubic shaped lun- Wang et al. (2008a), the circulating water was pumped through
cheon roll meat was preferred as reference geometry for the the vessel before and during the heating process, which was
load, as it exhibits a faster and more uniform heating with a used to improve the heating uniformity of mashed potato,
good rate of power absorption (Romano and Marra, 2008). including reduction of fringing electric field and edge-heating
Therefore, a considerable number of food products are proc- effects during the RF heating process.
essed in rectangular boxes during RF heating. Uyar et al. (2014)
confirmed that reducing electrode gap by increasing the load 3.5.4. For eggs and egg products
volume improved the RF power uniformity in the sample load. When the eggs were kept static and heated in parallel plate RF
However, Llave et al. (2014) reported that greater uniformity of applicator, nonuniformity in heating led to accumulation of
end-point temperature distribution of tuna muscle was cooked egg white at the top of the eggs (Dev et al., 2012). This
obtained when the top electrode projection was similar in size was because the eggs were kept in a position where the heating
to the sample, especially for samples with high-moisture con- took place from top to bottom in the parallel plate RF applica-
tent. In addition, greater uniformity of end-point temperature tor. The nonuniformity in heating was convincingly reduced
distribution was observed for low-fat content tuna muscle. when the eggs were rotated between the electrodes. As the egg
rotates, different region of the eggs were in close proximity to
3.5.2. For fresh fruits and vegetables the electrodes at different times (Dev et al., 2012). Additional
For high water content food, like fresh fruits and vegetables, processing aids, such as cooling of the air cell, were recom-
researchers used water surrounding combined with a moving mended to ensure successful pasteurization of eggs using RF
or rotating method to enhance the RF heating uniformity. Birla heating within a reasonable timeframe and with minimal qual-
et al. (2008a) reported that model fruit surrounded with air ity issues (Lau et al., 2016). In addition, air bubbles could also
between RF electrodes and placed in the proximity of electrodes be removed from food products using methods, such as ultra-
would not heat uniformly. Immersing the model fruit in water sonic treatment, to achieve better RF heating uniformity.
helped to reduce uneven heating within the model fruit. There-
fore, they suggested that movement and rotation of the spheri- 3.5.5. For low moisture foods
cal object was the only plausible solution for improving heating Several practical means can be used to minimize nonuniform
uniformity. With rotation and movement of apples (Wang RF heating in intermediate/low moisture foods, i.e. adding
et al., 2006a), cherries (Ikediala et al., 2002), mangoes (Sosa- forced hot air to the product surface to increase the energy
Morales et al., 2009), oranges (Birla et al., 2005), and persim- exchange between the product surface and surrounding air,
mons (Tiwari et al., 2008) using a fruit mover, RF heating uni- sample movement and rotation or mixing during RF treatment
formity has been significantly improved. Finally, a saline water (Birla et al., 2008a; Chen et al., 2015b, 2016; Tiwari et al., 2008).
immersion technique or differential startup temperature should Wang et al. (2005) developed an intermittent stirring method
be adopted along with fruit movement and rotation during RF during RF treatments to improve the heating uniformity for in-
heating to minimize differential heating between fruit and shell walnut. Their results showed a minimum of two stirrings
water. As demonstrated by Wang et al. (2006a), preheating in were needed for desired uniformity and insect mortality. Tiwari
hot water established a temperature gradient from the surface et al. (2011a) confirmed that increasing electrode gap by reduc-
(44 C) to the core (36 C) and using preheated apples in RF- ing the load volume and the upper electrode bent to a certain
assisted heating treatments resulted in a fairly uniform temper- angle improved the RF heating uniformity in the sample load.
ature (47.3 § 0.7 C) in fruit. The heating uniformity of RF-treated raisins were signifi-
cantly improved when containers with rounded edges and cor-
3.5.3. For liquid foods ners were used, with a maximum temperature difference of
To reduce fringe effect at the interface between the side of the about 12 C (Alfaifi et al., 2016). Using RF heating with a 2 cm
food package and the air in the RF applicators, low conductivity set back distance and 90 angle of the newly designed electro-
water was used to immerse the food tray (with low conductivity des, followed by forced hot air at 60 C improved the heating
20 Z. HUANG ET AL.

Table 3. Simulation methods for overcoming nonuniform RF heating in the literature.

Method Rationale Food product Reference

Surround the food with a cooling or Reduces temperature variation in the food; Apple Birla et al. (2004)
heating medium Bread Liu et al. (2011, 2013)
In-shell Macadamia nuts Wang et al. (2014)
Allows holding at a desired temperature Lentil Jiao et al. (2012)
Orange Birla et al.(2004)
Roasted peanuts Jiao et al. (2016)
Walnut Wang et al. (2007a)
Electrode modification Reduce the fringe effect from the edges of the Fish Llave et al. (2015)
electrode Raisins Alfaifi et al. (2016)
Soybeans Huang et al. (2015c)
Tuna muscle Llave et al. (2014)
Wheat flour Tiwari et al. (2011a)
Intermittent stirrings and mixing After mixing, the positions of the hot and cold Almonds Gao et al. (2010)
spots mixed evenly; Chestnuts Hou et al. (2014)
Temperature variations between central parts Corn Zheng et al.(2016)
and corner and edge parts were reduced due Pistachios Ling et al. (2016)
to heat loss Rice Zhou et al.(2015, 2016)
Walnut Wang et al. (2006b, 2007a)
Wheat Chen et al. (2015b)
Movement and rotation Redistribute electric field and heat within the Apple Birla et al. (2004)
food product Coffee bean Pan et al. (2012)
Eggs Dev et al. (2012)
Legume Wang et al. (2010)
Mango Sosa-Morales et al. (2009)
Orange Birla et al. (2004)
Walnut Wang et al. (2005, 2007a)
Wheat Chen et al. (2016)
Placing in the middle of the Avoids the electric field concentration at contact Dry soybeans Huang et al. (2015c)
electrodes surfaces; Meat cubes Uyar et al. (2014)
Electric field deflected by top and bottom edges Raisins Alfaifi et al. (2014)
in the middle location Wheat kernels Jiao et al. (2015a)
Using similar dielectric material Reduces the disparity in DPs between the food Apple Birla et al. (2004)
around the samples and the surrounding air, thus evening out the Cherry Ikediala et al. (2002)
electric field distribution Cured ham Bengtsson et al. (1970)
Mango Sosa-Morales et al. (2009)
Mashed potato Wang et al. (2008a)
Orange Birla et al.(2004)
Peanut butter Jiao et al. (2014a, 2015b)
Scrambled egg Luechapattanaporn et al. (2005)
Soybean flour Huang et al. (2016a)
Stone fruit Sisquella et al. (2013)

uniformity of the whole raisin samples with a temperature dif- equipped with an auxiliary hot air system and conveyor belt,
ference of about 5 C. Similarly, based on the heating unifor- choosing an optimum container material could be effectively
mity studies for disinfesting chestnuts conducted by Hou et al. minimize the effect of electric field bending and distortion
(2014), a RF treatment protocol was finally developed to com- within the corners and edges of food products during RF
bine 0.6 kW RF powers with a forced hot air at 55 C, move- treatment.
ment of the conveyor, mixing twice, and holding at 55 C hot
air for 5 min, followed by forced room air cooling through sin- 3.5.6. For agricultural products
gle-layer samples. The heating uniformity of RF treated wheat samples was
Subsequently, a simulation method of adding PEI cylindrical improved with the increasing mixing times based on experi-
blocks on top of and at the bottom of peanut butter samples in ment and simulation results (Chen et al., 2015b). Under mixing
a cylindrical jar has been evaluated, showing effective results to conditions, the maximum temperature of top layer in wheat
improve the RF heating uniformity (Jiao et al., 2015b). Their sample reduced from 66.5 C to 51.0 C in experiment and from
results revealed that the combination of PEI surrounding and 68.3 C to 56.1 C in simulation after three time mixing during
the addition of 8 cm diameter PEI blocks could further reduce 3.5 min RF heating. Chen et al. (2015b) reported that in practi-
the temperature distribution range in peanut butter within cal applications using several RF systems in series with mixings
7.1 C when the peanut butter was heated from 23 to 70 C. in between, forced hot air should be used to minimize the
Huang et al. (2016a) also demonstrated that the RF heating amount of heat loss during the mixing process. The same
uniformity could be improved when the surrounding container results have also been reported for stirring and mixing chest-
dielectric constant was in a comparable range of the sample’s, nut, lentil, soybean, walnuts, and wheat (Hou et al., 2014;
with the loss factor values of surrounding container lying Wang et al., 2007a, b; Wang et al., 2008c). Simulated results
between 0.01 and 0.1% of the sample’s. Therefore, in practical from Chen et al. (2016) demonstrated that moving and turning
applications for an industrial-scale RF system, which was the container by 90 could improve the RF heating uniformity.
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 21

During RF heating with three speeds of conveyor belt (v1 D prune, and walnut) treated in RF systems, each single material
8.57 m/h, v2 D 14.23 m/h, and v3 D 17.14 m/h) at a fixed gap of should be constructed in a realistic geometric model. Thus, the
120 mm, lower speeds of movement had a better heating uni- computer program allows an independent and accurate simula-
formity. To get a better heating uniformity for agricultural tion of solid-air interaction using a single particle approach.
products, they proposed to combine conveyor movement with
adding forced hot air and mixing.
4.3. Shrinkage and deformation during RF heating
Moreover, according to the simulated results of Huang et al.
(2015c), the RF heating uniformity of dry soybeans could be Water in high moisture foods could be lost in an open package
improved by using a suitable container material (with a similar during the RF heating process. This water loss could result in
dielectric constant to that of the sample), a smaller upper plate an important size change that hinders the analysis of heat and
area (similar to the sample size), and placing the samples in the mass transport and obtaining a convergent heating model.
middle of two plate electrodes. In addition to this, the rectangu- Shrinkage and deformation in foods occur due to moisture loss
lar-shaped polystyrene container with rounded edges and cor- during thermal processes. Effects of shrinkage and deformation
ners (inner corner radius of 8 cm) combined with container on the accuracy of RF heating models are sometimes signifi-
thickness of 8 cm could provide good heating uniformity for cant. Therefore, both the water transport model and the
soybeans (Huang et al., 2016b). mechanical deformation model should be taken into account in
the future.
4. Suggestions for future research
4.4. Coupling the heat and mass transfer model with
Although continuous progress has been made in recent years in
other models
improving the accuracy of the modeling, much research work
still needs to be carried out. The following section discusses the One of the critical roles of numerical modeling in the food
possible areas where further research could be performed to industry is to analyze a heating process for producing a food
improve the accuracy of model prediction. More efforts are product with high safety and quality. To obtain better under-
also needed to develop computer-aided engineering for a high standing of the RF heating mechanism and improve the
degree of automation in industrial RF heating processes. heating uniformity of various agricultural products in food
industry, more attempts can be made to combine other mod-
els (biochemical reaction, microbial deactivation, and
4.1. Surface heat and mass transfer coefficients
mechanics models) with the current heat and mass transfer
Heat and mass transfer coefficients are important parameters in models for further evaluating the safety and quality of foods
modeling heating, drying, and cooling processes. The mass and during RF heating processes. Therefore, it is necessary to
momentum transfers of water within the food are commonly develop an advanced simulation model to simulate the
ignored due to a short RF heating time (< 5 min) and the influ- industrial scale RF heating system, which combined the
ence on moisture content reduction is unnoticeable (< 2%) hot air-assisted RF heating with conveyor movement under
based on the reported research. For a long time RF heating (> mixing conditions.
5 min) in high moisture content foods, three phases are consid-
ered in the treatment: solid (skeleton), liquid (water), and gas
4.5. Model library for various food products
(water, vapor, and air). Based on the theory of mixtures (water,
vapor, and air), the coupled electromagnetic and transport There have been a number of publications in the area of com-
model needs to be developed for obtaining the moisture con- puter simulation for the use of RF technology for food process-
tent and mass fraction of vapor. Therefore, more reliable math- ing and preservation. Therefore, a simulation model library
ematical models need to be established by considering the heat and the corresponding heat treatment process should be devel-
and mass transfer coefficients under the given range of operat- oped for various food materials processed in different condi-
ing conditions. tions. When DPs and TPs of any kind of foods were entered to
the model library, the best heat treatment protocol could be
matched intelligently.
4.2. Food properties and geometries
Food properties (DPs and TPs) are one of the most important
4.6. Computer-assisted design and automatic heat
factors determining the accuracy of model predictions. There is
treatment process control
a critical need for the production of more DPs and TPs data on
foodstuffs and potential packaging. This information is the key The application of numerical modeling of RF heating pro-
to improve understanding of temperature distribution but also cesses may benefit the understanding of the physics of a
is important in the design of RF heating systems. To simplify food processing operation and thus aid in design, optimiza-
the model, bulk food products placed in the box or container is tion and control of a processing system. With a large store
simplified as a whole in the simulation model. TPs and DPs of and highspeed of models available in the food industry,
air particles mixtures are calculated by the dielectric and ther- more research should be stressed to practically use those
mal mixture equations. This may be available for the small par- models in design and control of a thermal processing sys-
ticle materials (almond, legume, peanut, raisin, and wheat tem. The treatment parameters can be controlled without
kernel). But for the large particle products (apricot, date, fig, delay by comparing 3D computer simulation results with
22 Z. HUANG ET AL.

the online real-time monitoring. The whole process of RF Sd changes in density per unit change in temperature (kg/
heat treatment can be realized automatically by using the m3 K)
computer control methods. The adjustment of the plate Sk changes in thermal conductivity per unit change in
spacing, conveyor belt movement, and the application of temperature (W/mK2)
the auxiliary measures (hot air) can all be changed automat- t time (s)
ically, which is applicable to industrial online production. T sample temperature (K)
Therefore, it is necessary to develop an advanced control tan d dielectric loss tangent (dimensionless)
system integrated with process optimization for a thermal @T/@t temperature rate of change (K/s)
process. V electric potential (V)
Vvol material volume (m3)
v angular frequency (rad/s)
SD standard deviation
5. Conclusions
d dielectric loss angle (rad)
This paper has systematically reviewed recent developments e permittivity (dimensionless)
in computer simulation for improving RF heating unifor- e0 free space permittivity (F/m)
mity in the processed food. These methods include combi- e0 dielectric constant (F/m)
nation with an external heating or cooling device, enclosing e00 dielectric loss factor (F/m)
in another medium, mixing or rotating food, modifying er relative permittivity (dimensionless)
0
electrode shapes, and sample movement. Mathematical er relative dielectric constant (dimensionless)
00
modeling improves the understanding of interaction of RF er relative dielectric loss factor (dimensionless)
waves with food and is essential to the continued develop- r gradient operator
ment of this novel technology. More research should be λ latent heat of vaporization per unit mass of evaporated
conducted to improve the accuracy of models by finding water (J/kg)
sufficient information on surface heat and mass transfer m magnetic permeability (H/m)
coefficients, food properties, and geometries during RF r density (kg/m3)
processing. More research should also be carried out on the re empirical constant in Eq. (31) (kg/m3)
heat and mass transfer through high moisture foods, defor- s electrical conductivity (S/m)
mation and shrinkage in RF heating processes. More efforts
are also needed to establish a model library, couple the heat Subscripts
and mass transfer model with other models, and develop a air
computer-aided engineering of RF processes on an indus- av average
trial scale. i insect
mat food material
p product
Nomenclature s solid
A!
plate area of an electrode (m2) sh initial freezing point
B magnetic flux density (dimensionless) sw normal freezing point of pure water
c speed of light in free space (3£108 m/s)
C capacitance (F)
cp heat capacity (J/kg K) Acknowledgments
ce empirical constantin Eq. (32) (J/kgK)
This research was conducted in the College of Mechanical and Electronic
d plate distance between the two electrodes (m) Engineering, Northwest A&F University. The authors thank Qian Hao,
dp penetration depth (m) Xiaoxi Kou, Lixia Hou, Rui Li, Shuang Zhang, and Bo Zhang for their help
D empirical constants in Eq. (32) (J/Knkg)
!
in collecting materials.
D!
electric flux density (dimensionless)
E electric field vector (dimensionless)
E electric field intensity (V/m) Funding
h overall external convective heat transfer coefficient (W/ We acknowledge the financial support by research grants from General
!
m2K) Program of National Natural Science Foundation of China (31371853)
H magnetic field intensity (dimensionless) and Program of Introducing International Advanced Agricultural Science
and Technologies (948 Program) of Ministry of Agriculture of China
f frequency (Hz)
(2014-Z21).
IC! electric current (A)
J current density (dimensionless)
k thermal conductivity (W/mK) References
ke empirical constantin Eq. (33) (W/mK)
mw evaporating flux of water (kg/m2 s) Alfaifi, B., Tang, J., Jiao, Y., Wang, S., Rasco, B., Jiao, S. and Sablani, S.
(2014). Radio frequency disinfestation treatments for dried fruit: model
n empirical constants in Eq. (32) (dimensionless) development and validation. J. Food Eng. 120:268–276.
!
n normal vector of the sample surface (dimensionless) Alfaifi, B., Tang, J., Rasco, B., Wang, S. and Sablani, S. (2016). Computer
P power density generated by electric field (W/m3) simulation analyses to improve radio frequency (RF) heating
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 23

uniformity in dried fruits for insect control. Innov. Food Sci. Emerg. Farag, K., Lyng, J., Morgan, D. and Cronin, D. (2008b). Dielectric and ther-
Technol. 37:125–137 mophysical properties of different beef meat blends over a temperature
Awuah, G., Ramaswamy, H., Economides, A. and Mallikarjunan, K. range of ¡18 to C10 C. Meat Sci. 79:740–747.
(2005). Inactivation of Escherichia coli K-12 and Listeria innocua in Farag, K., Duggan, E., Morgan, D., Cronin, D. and Lyng, J. (2009).
milk using radio frequency (RF) heating. Innov. Food Sci. Emerg. Tech- A comparison of conventional and radio frequency defrosting of lean
nol. 6:396–402. beef meats: effects on water binding characteristics. Meat Sci. 83:
Awuah, G. B., Koral, T. and Guan, D. (2014a). Radio-frequency baking and 278–284.
roasting of food products. In: Radio-Frequency Heating in Food Proc- Farag, K., Lyng, J., Morgan, D. and Cronin, D. (2011). A comparison of con-
essing: Principles and Applications, pp. 231–243. George, B. A., Hosa- ventional and radio frequency thawing of beef meats: effects on product
halli, S. R. and Tang, J., Eds., CRC Press, Boca Raton, FL. temperature distribution. Food Bioprocess Technol. 4:1128–1136.
Awuah, G. B., Ramaswamy, H. S. and Tang, J. (2014b). Principles of Fu, Y. C. (2004). Fundamentals and industrial applications of microwave
radio frequency and microwave heating. In: Radio-Frequency Heat- and radio frequency in food processing. In: Food Processing: Principles
ing in Food Processing: Principles and Applications, pp. 3–20. and Applications, pp.79–100. Smith, J. S. and Hui, Y. H., Eds., Black-
George, B. A., Hosahalli, S. R. and Tang, J., Eds., CRC Press, Boca well, Iowa, USA.
Raton, FL. Gao, M., Tang, J., Wang, Y., Powers, J. and Wang, S. (2010). Almond qual-
Baginski, M., Riggs, L., German, F. and Reed, M. (1989). Experimental and ity as influenced by radio frequency heat treatments for disinfestation.
numerical characterization of the radio frequency drying of textile Postharvest Biol. Technol. 58:225–231.
materials (I). J. Microw. Power EE. 24:14–20. Gao, M., Tang, J., Villa-Rojas, R., Wang, Y. and Wang, S. (2011). Pasteuri-
Baginski, M., Broughton, R., Hall, D. and Christman, L. (1990). Experi- zation process development for controlling Salmonella in in-shell
mental and numerical characterization of the radio frequency drying of almonds using radio frequency energy. J. Food Eng. 104:299–306.
textile materials (II). J. Microw. Power EE. 25:104–113. Geedipalli, S., Rakesh, V. and Datta, A. (2007). Modeling the heating uni-
Bengtsson, N., Green, W. and Del, V. F. (1970). Radio-frequency pasteuri- formity contributed by a rotating turn3/17/2017 in microwave ovens. J.
zation of cured hams. J. Food Sci. 35:682–687. Food Eng. 82:359–368.
Birla, S., Wang, S., Tang, J. and Hallman, G. (2004). Improving heating Geveke, D. J., Kozempel, M., Scullen, O. J. and Brunkhorst, C. (2002).
uniformity of fresh fruit in radio frequency treatments for pest control. Radio frequency energy effects on microorganisms in foods. Innov
Postharvest Biol. Technol. 33:205–217. Food Sci. Emerg Technol. 3:133–138.
Birla, S., Wang, S., Tang, J., Fellman, J., Mattinson, D. and Lurie, S. (2005). Guan, D., Cheng, M., Wang, Y. and Tang, J. (2004). Dielectric properties of
Quality of oranges as influenced by potential radio frequency heat mashed potatoes relevant to microwave and radio-frequency pasteuri-
treatments against Mediterranean fruit flies. Postharvest Biol. Technol. zation and sterilization processes. J. Food Sci. 69:30–37.
38:66–79. Ha, J. W., Kim, S. Y., Ryu, S. R. and Kang, D. H. (2013). Inactivation of Sal-
Birla, S., Wang, S. and Tang, J. (2008a). Computer simulation of radio fre- monella enterica serovar Typhimurium and Escherichia coli O157: H7
quency heating of model fruit immersed in water. J. Food Eng. 84:270– in peanut butter cracker sandwiches by radio-frequency heating. Food
280. Microbiol. 34:145–150.
Birla, S., Wang, S., Tang, J. and Tiwari, G. (2008b). Characterization of Hartshorn, L. (1949). Radio frequency heating. George Allen & Unwin Ltd.
radio frequency heating of fresh fruits influenced by dielectric proper- London, UK.
ties. J. Food Eng. 89:390–398. Holland, J. M. (1974). Dielectric post baking in biscuit making. Bak Ind. J.
Brunton, N. P., Lyng, J. G., Li, W., Cronin, D. A., Morgan, D. and 6:169–176.
McKenna, B. (2005). Effect of radio frequency (RF) heating on the tex- Hou, L., Ling, B. and Wang, S. (2014). Development of thermal treatment
ture, colour and sensory properties of a comminuted pork meat prod- protocol for disinfesting chestnuts using radio frequency energy. Post-
uct. Food Res. Int. 38:337–344. harvest Biol. Tech. 98:65–71.
Chan, T., Tang, J. and Younce, F. (2004). 3-Dimensional numerical model- Hou, L., Johnson, J. A. and Wang, S. (2016). Radio frequency heating for
ing of an industrial radio frequency heating system using finite ele- postharvest control of pests in agricultural products: A review. Posthar-
ments. J. Microwave Power EE. 39:87–105. vest Biol. Technol. 113:106–118.
Chen, J., Pitchai, K., Birla, S., Gonzalez, R., Jones, D. and Subbiah, J. (2013). Huang, Z., Chen, L. and Wang, S. (2015a). Computer simulation of radio
Temperature-dependent dielectric and thermal properties of whey pro- frequency selective heating of insects in soybeans. Int. J. Heat Mass
tein gel and mashed potato. Trans. ASAE. 56:1457–1467. Transfer. 90:406–417.
Chen, J., Pitchai, K., Birla, S., Jones, D., Subbiah, J. and Gonzalez, R. (2015a). Huang, Z., Zhu, H. and Wang, S. (2015b). Finite element modeling and
Development of a multi-temperature calibration method for measuring analysis of radio frequency heating rate in mung beans. Trans. ASAE.
dielectric properties of food. IEEE T. Dielect. El. In. 22:626–634. 58:149–160.
Chen, L., Wang, K., Li, W. and Wang, S. (2015b). A strategy to simulate Huang, Z., Zhu, H., Yan, R. and Wang, S. (2015c). Simulation and prediction
radio frequency heating under mixing conditions. Comput. Electron. of radio frequency heating in dry soybeans. Biosyst. Eng. 129:34–47.
Agr. 118:100–110. Huang, Z., Marra, F. and Wang, S. (2016a). A novel strategy for improving
Chen, L., Huang, Z., Wang, K., Li, W. and Wang, S. (2016). Simulation and radio frequency heating uniformity of dry food products using compu-
validation of radio frequency heating with conveyor movement. J. Elec- tational modeling. Innov Food Sci. Emerg Technol. 34:100–111.
tromag. Waves Appl. 30:473–491. Huang, Z., Zhang, B., Marra, F. and Wang, S. (2016b). Computational
Choi, C. and Konrad, A. (1991). Finite element modeling of the RF heating modelling of the impact of polystyrene containers on radio frequency
process. IEEE Trans. Magn. 27:4227–4230. heating uniformity improvement for dried soybeans. Innov Food Sci.
Datta, A. and Ni, H. (2002). Infrared and hot-air-assisted microwave heat- Emerg Technol. 33:365–380.
ing of foods for control of surface moisture. J. Food Eng. 51:355–364. Içier, F. and Baysal, T. (2004a). Dielectrical properties of food materials-1:
Datta, A. (2007). Porous media approaches to studying simultaneous heat factors affecting and industrial uses. Crit. Rev. Food Sci. Nutr. 44:
and mass transfer in food processes. I: Problem formulations. J. Food 465–471.
Eng. 80:80–95. Içier, F. and Baysal, T. (2004b). Dielectrical properties of food materials-2:
Dev, S. R., Kannan, S., Gariepy, Y. and Raghavan, V. G. (2012). Optimiza- Measurement techniques. Crit. Rev. Food Sci. Nutr. 44:473–478.
tion of radiofrequency heating of in-shell eggs through finite element Ikediala, J., Hansen, J., Tang, J., Drake, S. and Wang, S. (2002). Develop-
modeling and experimental trials. Prog. Electromagn. Res. 45:203–222. ment of a saline water immersion technique with RF energy as a post-
Dhall, A. and Datta, A. (2011). Transport in deformable food materials: a harvest treatment against codling moth in cherries. Postharvest Biol.
poromechanics approach. Chem. Eng. Sci. 66:6482–6497. Tech. 24:209–221.
Farag, K., Lyng, J., Morgan, D. and Cronin, D. (2008a). A comparison of Jiao, S., Tang, J., Johnson, J., Tiwari, G. and Wang, S. (2011). Determining
conventional and radio frequency tempering of beef meats: effects on radio frequency heating uniformity of mixed beans for disinfestation
product temperature distribution. Meat Sci. 80:488–495. treatments. Trans. ASAE. 54:1847–1855.
24 Z. HUANG ET AL.

Jiao, S., Johnson, J., Tang, J. and Wang, S. (2012). Industrial-scale radio fre- Marchand, C. and Meunier, T. (1990). Recent developments in industrial
quency treatments for insect control in lentils. J. Stored Prod. Res. radio-frequency technology. J. Microwave Power EE. 25:39–46.
48:143–148. Marra, F., Lyng, J., Romano, V. and McKenna, B. (2007). Radio-frequency
Jiao, S., Deng, Y., Zhong, Y., Wang, D. and Zhao, Y. (2015a). Investigation heating of foodstuff: solution and validation of a mathematical model.
of radio frequency heating uniformity of wheat kernels by using the J. Food Eng. 79:998–1006.
developed computer simulation model. Food Res. Int. 71:41–49. Marra, F., Zhang, L. and Lyng, J. G. (2009). Radio frequency treatment of
Jiao, S., Zhu, D., Deng, Y. and Zhao, Y. (2016). Effects of hot air-assisted foods: review of recent advances. J. Food Eng. 91:497–508.
radio frequency heating on quality and shelf-life of roasted peanuts. Marra, F., De Bonis, M. V. and Ruocco, G. (2010). Combined microwaves
Food Bioprocess Technol. 9:308–319. and convection heating: A conjugate approach. J. Food Eng. 97:31–39.
Jiao, Y., Tang, J. and Wang, S. (2014a). A new strategy to improve heating Marshall, M. and Metaxas, A. (1998). Modeling of the radio frequency
uniformity of low moisture foods in radio frequency treatment for electric field strength developed during the RF assisted heat pump dry-
pathogen control. J. Food Eng. 141:128–138. ing of particulates. J. Microwave Power EE. 33:167–177.
Jiao, Y., Tang, J., Wang, S. and Koral, T. (2014b). Influence of dielectric McCormick, R. (1988). Dielectric heat seeks low moisture applications.
properties on the heating rate in free-running oscillator radio frequency Prepared Foods. 162:139–140.
systems. J. Food Eng. 120:197–203. Mermelstein, N. (1997). Interest in radio frequency heating heats up. Food
Jiao, Y., Shi, H., Tang, J., Li, F. and Wang, S. (2015b). Improvement of Technol. 50:94–95.
radio frequency (RF) heating uniformity on low moisture foods with Metaxas, A. (1996). Foundations of electroheat: a unified approach. John
Polyetherimide (PEI) blocks. Food Res. Int. 74:106–114. Wiley & Sons, New York.
Jojo, S. and Mahendran, R. (2013). Radio frequency heating and its appli- Moreira, R., Palau, J., Sweat, V. and Sun, X. (1995). Thermal and physical
cation in food processing: A review. Int. J. Curr. Agric Res. 1:42–46. properties of tortilla chips as a function of frying time. J. Food Process
Jones, P. L. and Rowley, A. T. (1996). Dielectric drying. Drying Technol. Pres. 19:175–189.
14:1063–1098. Moyer, J. and Stotz, E. (1947). The blanching of vegetables by electronics.
Jumah, R. (2005). Modelling and simulation of continuous and intermit- Food Technol. 1:252–257.
tent radio frequency-assisted fluidized bed drying of grains. Food Bio- Nagaraj, G., Purohit, A., Harrison, M., Singh, R., Hung, Y. C. and Mohan,
prod. Process. 83:203–210. A. (2016). Radio frequency pasteurization of inoculated ground beef
Kim, S. Y., Sagong, H. G., Choi, S. H., Ryu, S. and Kang, D. H. (2012). homogenate. Food Control. 59:59–67.
Radio-frequency heating to inactivate Salmonella Typhimurium and Nelson, S. O. (2008). Dielectric properties of agricultural products and
Escherichia coli O157: H7 on black and red pepper spice. Int. J. Food some applications. Res. Agric. Eng. 54:104–112.
Microbiol. 153:171–175. Nelson, S. O. and Trabelsi, S. (2012). Factors influencing the dielectric
Kim, Y. R., Morgan, M., Okos, M. and Stroshine, R. (1998). Measurement properties of agricultural and food products. J. Microw. Power EE.
and prediction of dielectric properties of biscuit dough at 27 MHz. J. 46:93–107.
Microw. Power EE. 33:184–194. Neophytou, R. and Metaxas, A. (1996). Computer simulation of a radio
Kirmaci, B. and Singh, R. K. (2012). Quality of chicken breast meat cooked frequency industrial system. J. Microw. Power EE. 31:251–259.
in a pilot-scale radio frequency oven. Innov Food Sci. Emerg Technol. Neophytou, R. and Metaxas, A. (1997). Characterisation of radio frequency
14:77–84. heating systems in industry using a network analyser. IEE P. Sci. Meas.
Lau, S. K., Thippareddi, H., Jones, D., Negahban, M. and Subbiah, J. Tech. 144:215–222.
(2016). Challenges in radio frequency pasteurization of shell eggs: Neophytou, R. and Metaxas, A. (1998). Combined 3D FE and circuit
Coagulation rings. J. Food Sci. 81:2492–2502. modeling of radio frequency heating systems. J. Microw. Power EE.
Laycock, L., Piyasena, P. and Mittal, G. (2003). Radio frequency cook- 33:243–262.
ing of ground, comminuted and muscle meat products. Meat Sci. Neophytou, R. and Metaxas, A. (1999). Combined tank and applicator
65:959–965. design of radio frequency heating systems, IEEE P-Microw. Anten P.
Ling, B., Hou, L., Li, R. and Wang, S. (2016). Storage stability of pistachios 146:311–318.
as influenced by radio frequency treatments for postharvest disinfesta- Ni, H., Datta, A. K. and Torrance, K. E. (1999). Moisture transport in
tions. Innov Food Sci. Emerg Technol. 33:257–364. intensive microwave heating of biomaterials: a multiphase porous
Liu, C. M., Sakai, N. and Hanzawa, T. (1999). Three dimensional analysis media model. Int. J. Heat Mass Transfer. 42:1501–1512.
of heat transfer during food thawing by far-infrared radiation. Food Orsat, V., Gariepy, Y., Raghavan, G. and Lyew, D. (2001). Radio-frequency
Sci. Technol. Res. 5:294–299. treatment for ready-to-eat fresh carrots. Food Res. Int. 34:527–536.
Liu, Q., Zhang, M., Xu, B., Fang, Z. and Zheng, D. (2015). Effect of radio Pace, M., De Bonis, M. V., Marra, F. and Ruocco, G. (2011). Characteriza-
frequency heating on the sterilization and product quality of vacuum tion of a combination oven prototype: Effects of microwave exposure
packaged Caixin. Food Bioprod. Process. 95:47–54. and enhanced convection to local temperature rise in a moist substrate.
Liu, Y., Tang, J. and Mao, Z. (2009). Analysis of bread dielectric properties Int. Commun. Heat Mass Transfer. 38:557–564.
using mixture equations. J. Food Eng. 93:72–79. Pan, L., Jiao, S., Gautz, L., Tu, K. and Wang, S. (2012). Coffee bean heating
Liu, Y., Tang, J., Mao, Z., Mah, J. H., Jiao, S. and Wang, S. (2011). Quality uniformity and quality as influenced by radio frequency treatments for
and mold control of enriched white bread by combined radio frequency postharvest disinfestations. Trans. ASAE. 55:2293–2300.
and hot air treatment. J. Food Eng. 104:492–498. Pham, Q. T. (2006). Modelling heat and mass transfer in frozen foods: a
Liu, Y., Wang, S., Mao, Z., Tang, J. and Tiwari, G. (2013). Heating patterns review. Int. J. Refrig. 29:876–888.
of white bread loaf in combined radio frequency and hot air treatment. Piyasena, P., Dussault, C., Koutchma, T., Ramaswamy, H. and Awuah, G.
J. Food Eng. 116:472–477. (2003). Radio frequency heating of foods: principles, applications and
Llave, Y., Terada, Y., Fukuoka, M. and Sakai, N. (2014). Dielectric proper- related properties-a review. Crit. Rev. Food Sci. Nutr. 43:587–606.
ties of frozen tuna and analysis of defrosting using a radio-frequency Rincon, A. M., Singh, R. K. and Stelzleni, A. M. (2015). Effects of endpoint
system at low frequencies. J. Food Eng. 139:1–9. temperature and thickness on quality of whole muscle non-intact
Llave, Y., Liu, S., Fukuoka, M. and Sakai, N. (2015). Computer simula- steaks cooked in a radio frequency oven. LWT ¡ Food Sci. Technol.
tion of radiofrequency defrosting of frozen foods. J. Food Eng. 64:1323–1328.
152:32–42. Romano, V. and Marra, F. (2008). A numerical analysis of radio frequency
Luechapattanaporn, K., Wang, Y., Wang, J., Al-Holy, M., Kang, D. H., heating of regular shaped foodstuff. J. Food Eng. 84:449–457.
Tang, J. and Hallberg, L. M. (2004). Microbial safety in radio frequency Ryyn€anen, S. (1995). The electromagnetic properties of food materials: a
processing of packaged foods. J. Food Sci. 69:201–206. review of the basic principles. J. Food Eng. 26:409–429.
Luechapattanaporn, K., Wang, Y., Wang, J., Tang, J., Hallberg, L. M. and Sahin, S., Sastry, S. and Bayindirli, L. (1999). Effective thermal conductivity
Dunne, C. P. (2005). Sterilization of scrambled eggs in military poly- of potato during frying: measurement and modeling. Int. J. Food Prop.
meric trays by radio frequency energy. J. Food Sci. 70:288–294. 2:151–161.
 CRITICAL REVIEWS IN FOOD SCIENCE AND NUTRITION 25

Shrestha, B. and Baik, O. D. (2013). Radio frequency selective heating of Wang, S., Tang, J., Sun, T., Mitcham, E., Koral, T. and Birla, S.
stored-grain insects at 27.12 MHz: a feasibility study. Biosyst. Eng. (2006b). Considerations in design of commercial radio frequency
114:195–204. treatments for postharvest pest control in in-shell walnuts. J. Food
Sisquella, M., Casals, C., Picouet, P., Vi~ nas, I., Torres, R. and Usall, J. Eng. 77:304–312.
(2013). Immersion of fruit in water to improve radio frequency treat- Wang, S., Monzon, M., Johnson, J., Mitcham, E. and Tang, J. (2007a).
ment to control brown rot in stone fruit. Postharvest Biol. Technol. Industrial-scale radio frequency treatments for insect control in wal-
80:31–36. nuts: I: Heating uniformity and energy efficiency. Postharvest Biol.
Sosa-Morales, M., Tiwari, G., Wang, S., Tang, J., Garcia, H. and Lopez- Technol. 45:240–246.
Malo, A. (2009). Dielectric heating as a potential post-harvest treat- Wang, S., Monzon, M., Johnson, J., Mitcham, E. and Tang, J. (2007b).
ment of disinfesting mangoes, Part II: Development of RF-based proto- Industrial-scale radio frequency treatments for insect control in wal-
cols and quality evaluation of treated fruits. Biosyst. Eng. 103:287–296. nuts: II: Insect mortality and product quality. Postharvest Biol. Tech-
Tang, J., Feng, H. and Lau, M. (2002). Microwave heating in food process- nol. 45:247–253.
ing. In: Advances in Bioprocessing Engineering, pp. 1–43. New York: Wang, S., Luechapattanaporn, K. and Tang, J. (2008b). Experimental
Scientific Press. methods for evaluating heating uniformity in radio frequency systems.
Tang, J., Wang, Y. and Chan, T. (2005). Radio frequency heating in food Biosyst. Eng. 100:58–65.
processing. In: Novel Food Processing Technologies, pp. 501–524. New Wang, S., Yue, J., Chen, B. and Tang, J. (2008c). Treatment design of radio
York, NY: Marcel Dekker. frequency heating based on insect control and product quality. Post-
Tiwari, G., Wang, S., Birla, S. and Tang, J. (2008). Effect of water-assisted harvest Biol. Technol. 49:417–423.
radio frequency heat treatment on the quality of ‘Fuyu’persimmons. Wang, S., Tiwari, G., Jiao, S., Johnson, J. and Tang, J. (2010). Developing
Biosyst. Eng. 100:227–234. postharvest disinfestation treatments for legumes using radio frequency
Tiwari, G., Wang, S., Tang, J. and Birla, S. (2011a). Analysis of radio fre- energy. Biosyst. Eng. 105:341–349.
quency (RF) power distribution in dry food materials. J. Food Eng. Wang, S., Tang, J., Johnson, J. and Cavalieri, R. (2013). Heating uniformity
104:548–556. and differential heating of insects in almonds associated with radio fre-
Tiwari, G., Wang, S., Tang, J. and Birla, S. (2011b). Computer simulation quency energy. J. Stored Prod. Res. 55:15–20.
model development and validation for radio frequency (RF) heating of Wang, Y., Wig, T., Tang, J. and Hallberg, L. (2003b). Sterilization of food-
dry food materials. J. Food Eng. 105:48–55. stuffs using radio frequency heating. J. Food Sci. 68:539–544.
Uyar, R., Erdogdu, F. and Marra, F. (2014). Effect of load volume on power Wang, Y., Li, Y., Wang, S., Zhang, L., Gao, M. and Tang, J. (2011). Review
absorption and temperature evolution during radio-frequency heating of dielectric drying of foods and agricultural products. Int. J. Agric. &
of meat cubes: A computational study. Food Bioprod Process. 92:243– Biol. Eng. 4:1–19.
251. Wang, Y., Zhang, L., Gao, M., Tang, J. and Wang, S. (2014). Pilot-scale
Uyar, R., Bedane, T. F., Erdogdu, F., Palazoglu, T. K., Farag, K. W. and radio frequency drying of macadamia nuts: Heating and drying unifor-
Marra, F. (2015). Radio frequency thawing of food products-a compu- mity. Drying Technol. 32:1052–1059.
tational study. J. Food Eng. 146:163–171. Yang, J., Zhao, Y. and Wells, J. H. (2003). Computer simulation of capaci-
Uyar, R., Erdogdu, F., Sarghini, F. and Marra, F. (2016). Computer simulation tive radio frequency (RF) dielectric heating on vegetable sprout seeds.
of radio-frequency heating applied to block-shaped foods: Analysis on the J. Food Proc. Eng. 26:239–263.
role of geometrical parameters. Food Bioprod. Process. 98:310–319. Zhang, H. and Datta, A. K. (2001). Electromagnetics of microwave heating:
Venkatesh, M. and Raghavan, G. (2005). An overview of dielectric proper- magnitude and uniformity of energy absorption in an oven. In: Hand-
ties measuring techniques. Can. Biosyst. Eng. 47:15–30. book of Microwave Technology for Food Applications, pp. 1–28. Datta,
Wang, J., Olsen, R. G., Tang, J. and Tang, Z. (2008a). Influence of mashed A. K. and Anantheswaran, R. C., Eds., Marcel Dekker, New York.
potato dielectric properties and circulating water electric conductivity Zhang, L., Lyng, J. G. and Brunton, N. P. (2004). Effect of radio frequency
on radio frequency heating at 27 MHz. J. Microw. Power EE. 42:31–46. cooking on the texture, colour and sensory properties of a large diame-
Wang, J., Luechapattanaporn, K., Wang, Y. and Tang, J. (2012). Radio fre- ter comminuted meat product. Meat Sci. 68:257–268.
quency heating of heterogeneous food-meat lasagna. J. Food Eng. Zhao, Y., Flugstad, B., Kolbe, E., Park, J. W. and Wells, J. H. (2000). Using
108:183–193. capacitive (radio frequency) dielectric heating in food processing and
Wang, S., Ikediala, J., Tang, J., Hansen, J., Mitcham, E., Mao, R. and Swan- preservation-a review. J. Food Proc. Eng. 23:25–55.
son, B. (2001). Radio frequency treatments to control codling moth in Zheng, A., Zhang, B., Zhou, L. and Wang, S. (2016). Application of radio
in-shell walnuts. Postharvest Biol. Technol. 22:29–38. frequency pasteurization to corn (Zea mays L.): Heating uniformity
Wang, S., Tang, J., Johnson, J., Mitcham, E., Hansen, J., Cavalieri, R., improvement and quality stability evaluation. J. Stored Prod. Res.
Bower, J. and Biasi, B. (2002). Process protocols based on radio fre- 68:63–72.
quency energy to control field and storage pests in in-shell walnuts. Zhou, L., Ling, B., Zheng, A., Zhang, B. and Wang, S. (2015). Developing
Postharvest Biol. Technol. 26:265–273. radio frequency technology for postharvest insect control in milled
Wang, S., Tang, J., Cavalieri, R. and Davis, D. (2003a). Differential heating rice. J. Stored Prod. Res. 62:22–31.
of insects in dried nuts and fruits associated with radio frequency and Zhou, L. and Wang, S. (2016). Verification of radio frequency heating uni-
microwave treatments. Trans. ASAE. 46:1175–1182. formity and Sitophilus oryzae control in rough, brown and milled rice.
Wang, S., Yue, J., Tang, J. and Chen, B. (2005). Mathematical modelling of J. Stored Prod. Res. 65:40–47.
heating uniformity for in-shell walnuts subjected to radio frequency Zhu, H., Huang, Z. and Wang, S. (2014). Experimental and simulated top
treatments with intermittent stirrings. Postharvest Biol. Technol. electrode voltage in free-running oscillator radio frequency systems. J.
35:97–107. Electromag. Waves Appl. 28:606–617.
Wang, S., Birla, S., Tang, J. and Hansen, J. D. (2006a). Postharvest treat- Zhu, J. F., Liu, J. Z., Wu, J. H., Cheng, J., Zhou, J. H. and Cen, K. F. (2015).
ment to control codling moth in fresh apples using water assisted radio Thin-layer drying characteristics and modeling of Ximeng lignite under
frequency heating. Postharvest Biol. Technol. 40:89–96. microwave irradiation. Fuel Process. Technol. 130:62–70.