Journal of Food Engineering 83 (2007) 173–185

[Link]/locate/jfoodeng

Advanced food process engineering to model real foods

and processes: The ‘‘SAFES” methodology
P. Fito *, M. LeMaguer, N. Betoret, P.J. Fito
Institute of Food Engineering for Development, Polytechnic University of Valencia, Spain

Available online 21 February 2007

Abstract

Food quality and safety are the main concerns of consumers and the principal target of the food industry processes. The concept of
‘‘food process engineering for product quality” has been arising in the last years with the aim of designing and controlling processes to
produce food products with very specific properties of quality and safety, previously defined on the basis of market opportunities
analysis.
The final properties of food products are the result of the changes in raw material as a consequence of process conditions. In the case
of colloidal or cellular foods, these changes may be observed as differences in quality factors as food composition, nutritional facts, taste
and flavour, aspect, shape and size, colour, texture, etc. These changes in food properties may be explained because physical and chem-
ical phenomena produced in line with the process progression, as structure deformations, chemical or enzymatic reactions, phase tran-
sitions, etc.
The models currently used in food process engineering simplify too much both the food system description and the mechanisms and
rate equations of changes. The food system is supposed to be homogeneous and continuous. In this way, thermodynamic and kinetic
equations deduced for ideal gas or liquids, in conditions close to equilibrium are applied to cellular solid foods, in conditions far away
from the equilibrium.
It is necessary to develop advanced concepts and methodologies in food process engineering. The new models for food and processes
development must incorporate information about all these aspects (thermodynamic, structural, chemist and biochemist, and even
mechanics). Only in this way they would be able to calculate and predict the real changes in the whole quality of food product in line
with the process progression. The SAFES methodology (systematic approach to food engineering systems) recognizes the complexity of
food system and allows coordinating the information about food structure, composition, quality, thermodynamic, etc. in adequate tools
to develop real food and processes models.
‘‘In a complete theory there is an element corresponding to each element of reality” [Einstein, A., Podolsky, B., & Rosen, N., (1935).
Can quantum-mechanical description of physical reality be considered complete? Physical Review, 47, pp. 777–780].
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Food process engineering; Food product engineering; Models of real foods; SAFES

1. Introduction food products with very specific properties of quality and

safety, previously defined on the basis of market opportu-
Food quality and safety are the main concerns of con- nities analysis (ETP ‘‘Food for Life, 2006).
sumers and the principal target of the food industry pro- The final properties of food products are the result of
cesses. The concept of ‘‘food process engineering for the changes produced in raw materials as a consequence
product quality” has been arising in the last years with of process treatments (Aguilera & Stanley, 1999). These
the aim of designing and controlling processes to produce changes may be observed as differences in quality factors
such as food composition, nutritional facts, taste and fla-
*
Corresponding author. vour, aspect, shape and size, colour, texture, etc. (Aguilera,
E-mail address: pfito@[Link] (P. Fito). Chiralt, & Fito, 2003a, 2003b). In the case of colloidal or

0260-8774/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/[Link].2007.02.017
174 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185

Nomenclature

Latin symbols S Entropy (J/K)

ai activity of i-component (dimensionless) si molar entropy (J/K mol of i)
E internal energy (J) T temperature (K)
Ei molar internal energy (J/mol of i) t time (s)
G Gibbs’ free energy (J) V volume (m3)
Gi molar Gibbs’ free energy for i-component Vi partial molar volume of i-component
(J/mol of i) (m3/mol of i)
Gi;j molar Gibbs’ free energy for i-component in the Xi molar fraction of i-component
j phase (J/mol of i) (mol of i/total mol)
H enthalpy (J) xi mass fraction of i-component (kg of i/total kg)
hi molar enthalpy (J/mol of i)
Ji molar flux of i-component (Mol of i/s m2) Greek symbols
Li phenomenological coefficient for i-component li chemical potential of i-component (J/mol of i)
flux (Mol of i/s m2 Dli) lext
i extended chemical potential of i-component
Mn, n descriptive matrix of product flowing out of the (J/mol of i)
stage of change n considering the overall mass in lext;0
i extended chemical potential of i-component at
the system equal to mn/mn = 1 reference conditions (J/mol of i)
Mn, n1 descriptive matrix of product flowing out of the ui ¼ pi =P oi (relative concentration of i-component in a
stage of change n referred to the strain n1 gas mixture referred to saturation;
(considering the overall mass in the system equal 0 6 ui 6 1) (dimensionless)
to mn/mn1 6¼ 1) ci Activity coefficient of i-component (dimension-
MCn1, n matrix of changes produced when the strain less)
mn1 passes through the stage of change n r surface tension (N/m)
MPN descriptive matrix of one process with N stages q density (kg/m3)
of change
mn mass flow rate corresponding to strain n (kg/s) Subscripts
Dmn mass gain in the stage of change n (kg/s) i some property referred to i-component
n number of moles (dimensionless) w some property referred to water
ni number of moles of i-component (dimension- s some property referred to solutes
less)
P pressure (Pascal; N/m2) Superscripts
pi partial pressure of i-component in a gas (Pascal; o referred to some condition where the value of
N/m2) the function is supposed to be arbitrarily equal
R universal gases constant (N m/mol K) to zero.
P oi vapour pressure of i-component (Pascal; N/m2)

cellular foods, these changes in food properties may be lular structures and, of course, in conditions far away from
explained as the result of physical and chemical phenom- the equilibrium (Bird, Stewart, & Lightfoot, 2002). The
ena produced in line with the process progression such as typical approach used to model food drying operations is
structure deformations and relaxations (shrinking and a good example of this methodology. Frequently, the mod-
swelling), chemical or enzymatic reactions, phase transi- els obtained in this way cannot accurately predict the
tions, mass and energy transport phenomena, etc. (Aguil- changes in the food properties through the industrial pro-
era & Baffico, 1997; Barat, Fito, & Chiralt, 2001). cesses when the range of values for the industrial process
The models currently used in food process engineering variables is different to those used in the experiments. In
usually simplify too much both the food system description this sense it is remarkable as to how scattered reported val-
and the mechanisms and rate equations of changes. Typi- ues in the literature for the experimental effective diffusivi-
cally the food system is supposed to be isotropous, homo- ties are (Martinez, Andres, Chiralt, & Fito, 1998).
geneous and continuous, with only two or three Based on this remark, it is necessary to develop
components distributed in one or two phases (Crank, advanced concepts and methodologies in food process
1985). In this approach, thermodynamic and kinetic equa- engineering (Marcotte & Le Maguer, 1991; Zhiming &
tions deduced for ideal gas or liquids in conditions close to Le Maguer, 1996). The new models for food and processes
equilibrium are applied to model food with colloidal or cel- development must incorporate enough information about
 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185 175

all these aspects (thermodynamic, structural, chemist and juices) or crystals (like sodium chloride or sucrose) but
biochemist, and even mechanics). Only in this way they most of them belong to the more complexes categories: col-
would be able to calculate and predict the real changes in loidal or cellular and multi-cellular structures. Therefore,
the quality of food product in line with the process progres- in foods systems the structure concerns their molecular,
sion. The SAFES methodology (systematic approach to polymeric, colloidal and cellular complexity and conse-
food engineering systems) recognizes the complexity of quently their functionality concerns their chemical, physi-
food system and allows coordinating the information cal, biological and sensorial properties (Fito & Chiralt,
about food structure (Aguilera, 2006), composition, ther- 2001). Some of these properties are used by the consumers
modynamic, etc. in adequate tools to develop real food to define the food quality. For this reason it is very impor-
and processes. tant to be able to analyze and evaluate, in any food system,
its structure-properties ensemble and its concerns with the
2. Food product engineering: modelling of food and biological food quality and safety factors as appreciated by
systems consumers.

2.1. Structure of food system: the structure-properties 3. The SAFES simplified space of the structured phases and
ensemble components

The food system is a volume of universe full of matter 3.1. Components, structured phases and aggregation states
and energy. The structure concerns the way in which mat-
ter and energy organize themselves inside the system vol- Going ahead with the analysis of the structure-properties
ume. Some levels of complexity in matter condensation ensemble, the first step would be to define a simplified space
inside the system may be observed in Table 1 (in order of of components and structuredCardiff, phases. Now the
increasing complexity). relation between the concepts of food quality and safety
These levels of complexity in the structure of condensed (FQS) factors and the structure-properties ensemble (SPE)
soft matter are the results of the universe evolution. Each plays an important role. From the knowledge of FQS fac-
new level of complexity incorporates an improved level of tors, defined by consumer’s needs and preferences, the
functionality and results from the rearrangement of the more significant components and phases in the system must
previous less complex structures. In this way, more com- be defined. We may obtain information about components
plexity does imply new and more powerful functionality. from chemical analysis and we also have information about
In the same way, the system functionality is appreciated phases from both the thermodynamic properties of compo-
by an observer as the system properties (ETP ‘‘Food for nents (aggregation state of components) and from the food
Life, 2006). As a consequence complexity refers to the level micro and macrostructure (the main structural features in
of structure as well as functionality refers to the system the system are frequently phases or even multi-phases);
properties. Therefore, the concept of complexity-function- for this reason, in the future, we will always use the term
ality is equivalent to those of structure-properties. The of structured phases. The decision about whether a real
structure-properties ensemble refers to the knowledge of component or structured phase must be considered alone
structure and properties in a system as well as the under- as it is, or whether it should be included with others in a
standing of the relationships existing between the two con- simplified component or phase depends on how important
cepts and, of course, the capability to predict the changes in it is to explain some FQS factor. The arrangement of sim-
food properties, produced when any change in food struc- plified components (i) and structured phases (j) constitutes
ture occurs (Aguilera et al., 2003a). the space of structured phases and components. It is a two-
As food systems structure concerns, some foods are dimension space (i, j), but it would be very useful to include
monophasic; liquids (water solutions as oils or clarified also the information about the aggregation state (k) of each

Table 1
Some levels of complexity in matter condensation
Level of complexity Name of the structure Structural features Branch of science
Sub-atomic Elemental particles Subatomic particles Physics
Atomic Atomic structure Nucleus and electrons Physics
Molecular Molecular structure Covalent or ionic bonds; ions Chemistry
Polymeric Macromolecular Tertiary and quaternary structures Science of polymers. Proteomics,
structure genomics
Colloidal Colloidal structure Colloids (suspensions, emulsions, foams, Colloids science
gels)
Cellular Cellular structure Cells Cellular biology
Multi-cellular plants and Live plants and animals Tissues, organs, individuals Biology
animals
176 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185

component (i) in every structured phase (j). In this way, we Table 2

have finally a three-dimension space (i, j, k). Elements included in the three-dimension (i, j, k) simplified space
describing an industrial whipped cream
Further on, we will use this three-dimension space to
develop mathematical tools to describe the food system Components (i) Structured phases (j) Aggregation state (k)
and to perform mass and energy balances inside the system Water Gas Solid
and between the food system and the surroundings. It will Fat Insoluble dispersed solids Adsorbed
Air Liquid fat Vitreous
also be used to analyze the equilibrium situations or the Seroprotein Solid fat Rubbery
chemical reactions, phase transitions or transport phenom- Caseinate Aqueous liquid solution Liquid
ena suffered by the food throughout any operation or process Sucrose Gas
(Fito, Ortega-Rodriguez, & Barbosa-Canovas, 2001). The Lecithin
three-dimension space may provisionally be represented in Modified starch
Carragenate
a table with three rows: components (i), structured phases Xantan gum
(j) and aggregation states (k). Let us proceed with some
examples:
A. In Fig. 1, the microstructure of a sample of industrial
whipped cream containing water, fat, air, seroprotein, case-
inate, sugar, lecithin, modified starch, carragenate and
xantan gum may be observed.
In this food system, a foam structure with a phase of gas
bubbles dispersed in an emulsion (oil–water) containing fat
granules and an aqueous solution as continuous phase may
be observed. The description of such a complex product as
well as the analysis of its properties and changes through-
out any process implies the recognition of the phases that
play some role in the product behaviour. As a first
approach, it would be convenient to include the following
simplified ‘‘structured phases”: gas, insoluble dispersed
solids, liquid fat, solid fat and aqueous solution. The aggre-
gation states, depending on temperature and concentra-
tion, may be solid, adsorbed, vitreous, rubbery, liquid
and gas. Additionally, some characteristic dimensions of
the different structural features must be included, like the Fig. 2. Parenchyma tissue of apple ‘‘Granny Smith” observed by Cryo-
average diameter (or even diameter distribution) of gas SEM.

Fig. 1. Sample of an industrial whipped cream observed by Cryo-SEM (right) and TEM (left) (Dickinson & Stainsby, 1982).
 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185 177

bubbles and fat globules. Table 2 summarizes the elements Table 4

included in the three-dimension space. Elements included in the three-dimensional simplified space (i, j, k)
describing a typical salted dried meat
B. Fig. 2 shows a magnification of parenchyma of apple
‘‘Granny Smith” observed by Cryo-SEM (Fito et al., 2001). Components (i) Structured phases (j) Aggregation state (k)
Several cells may be observed (diameter ranging from 200 Water Gas Solid
to 300 lm) showing their cell wall, cell membrane and med- Native solutes Solid matrix Adsorbed
Solid matrix Intracellular liquid Vitreous
ium lamella. All these elements are insoluble in water and NaCl Extracellular liquid Rubbery
may be simplified as a food insoluble matrix. Inside the bro- Gas Solid solutes (crystals) Liquid
ken cells a simplified internal liquid phase may be observed. Gas
Also several intercellular spaces are filled both with some
extracellular liquid phase and/or with a gas phase. In some
cases (if the product will be air dried) a solid solutes phase
may appear when the liquid phases in the food become sat-
urated. Table 3 shows the different elements included in the
three-dimension simplified space of structured phases and
components. Additional information is the fact that the cell
membrane is semi-permeable, allowing only the water
transport by osmotic mechanisms (Barat et al., 2001).
Structural features may include the average diameters of
cells and intercellular spaces, cell wall thickness, etc.
C. Fig. 3 shows a magnification of a typical salted dried
meat named ‘‘tasajo” in some countries of Latin-America.
Sodium chloride crystals produced because the internal
liquid phase in the salted meat became saturated
(aw = 0.75 and 0.245 kg salt/kg solution) throughout the
drying operation may be observed. From this point onward
any additional loss of water produces more NaCl crystalli-
zation. In Table 4 the different components, structured Fig. 4. Starch in cassava cells.
phases and aggregation states included in the simplified
space of phases and components may be observed. D. Fig. 4 shows a Cryo-SEM observation of a cell of
dried cassava. A typical feature of this product is the great
Table 3 deal of starch granules that may be observed inside the cell
Elements included in the three-dimensional simplified space (i, j, k) membrane. In such a product an interesting question is the
describing the parenchyma of apple ‘‘Granny Smith” knowledge of the shared-out water between the starch
Components (i) Structured phases (j) Aggregation state (k) granules and the liquid phase inside the cell. Different ways
Water Gas Solid of sharing-out may result in different aw values for the
Native solutes Solid matrix Adsorbed same water content due to the different isotherms for liquid
Solid matrix Intracellular liquid Vitreous phase and for the solid matrix.
Added solutes Extracellular liquid Rubbery
Gas Solid solutes Liquid
Gas 4. The descriptive matrix: a mathematical tool to describe the
food system

4.1. The state variables: the share-out of matter among

components and structured phases

After defining the three-dimension space of phases and

components (i, j, k), it is possible to include the information
about the state variables of the system in an ordered way.
(see Table 5) The state variables are P, T, V and the mass
of any component in any phase and in any aggregation
state. Usually the chemical complexity of real food systems
leads to the use of mass content (m) or mass fractions (x)
instead of the number of moles (n). Then the needed state
variable information may be integrated as follows:

 The share-out of matter among phases, components and

Fig. 3. Cryo-SEM observation of dried salted meat ‘‘Tasajo”. aggregation states (the composition matrix: xki;j Þ.
178 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185

Table 5  when m 6¼ 1, values in xki;j are just the value of the

Elements included in the three-dimensional (i, j, k) simplified space mass of component i in phase j and in an aggregation
describing a casaba
state k.
The ratio P Pj xk is the specific volume of phase j.
Components (i) Structured phases (j) Aggregation state (k) V
–
Water Gas Solid Pi Pk i;jP k
Solutes Solid matrix Adsorbed x
i Pj k i;j
Solid matrix Intracellular liquid Vitreous
– The ratio Vj
¼ q is the food system average
Starch Extracellular liquid Rubbery density
P
j

Gas Starch granules Liquid – Pj V j is the volume of food system when m 6¼ 1

Gas – j V j is the average specific volume of food system when
m = 1.

 The share-out of the system volume among phases (the

volume vector: Vj) 4.2. Mass and volume balances inside the product
 The values of P and T in the phases (the pressure and
temperature vectors: Pj and Tj)  Mass balance for each component: vector of mass
shared-out for each component among the phases. For
Then, the descriptive matrix of the product will be component 1:
provisionally XX
xi¼1 ¼ xk1j ð2Þ
Mðxki;j ; V j ; P j ; T j Þ ð1Þ k j

The composition matrix xki;j

may be represented as shown in  Mass balance for each phase: vector of overall mass
Fig. 5. The components, phases and aggregation states are shared-out among phases (last row of the concentration
located on the three axes (i, j, k), respectively. In this matrix). For phase 1:
arrangement, the dotted arrows show the possible changes XX
in local values of xki;j produced by mass transport, phase xj¼1 ¼ xki1 ð3Þ
k i
transitions and/or chemical reactions inside the system.
Additionally, it is also possible to show the three-dimen-  Overall mass balance
sional data in a plane. In this way, the values of aggrega- XXX
x¼ xkij ¼ 1 ð4Þ
tion states (k) of each component are included in the i j k
plane (i, j) as described in Fig. 6. Now it is possible to show
together both the composition matrix and the PVT vectors  Overall volume balance
X
in a unified matrix: the descriptive matrix. V ¼ Vj ð5Þ
Some properties of the descriptive matrix are j

PPP
– The value i j k xki;j ¼ m where m is the overall mass
of the system:
4.3. The energy inside the system: the Gibbs free energy
 when m = 1; values in xki;j are mass fractions
(concentrations);
The thermodynamic analysis of the energy, the equilib-
rium and fluxes analysis in food systems is usually made
in terms of the Gibbs free energy (Nicolis & Prigogine,
1977a, 1977b).
For a closed system:
G ¼ H  TS ¼ E þ PV  TS ð6Þ
and so
dG ¼ dH  T dS  SdT ¼ V dP  SdT ð7Þ

In the food system the state variables have been included in

the descriptive matrix as follows:
Mðxki;j ; V j ; P j ; T j Þ ð8Þ
But we may change the units to describe the mass share-out
in the system using moles instead of kg and then Eq. (8)
becomes:
Mðnki;j ; V j ; P j ; T j Þ ð9Þ
Fig. 5. Composition matrix of food system.
P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185 179

Fig. 6. Descriptive matrix of food system developed on a plane.

180 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185
 
Eqs. (8) and (9) are completely equivalents, depending only oG
¼V ð17Þ
on the units used to define the mass in the system. Then, in oP T ;n
any phase j of the system it may be assumed that  
oG
  ¼ S ð18Þ
oT P ;n
Gj ¼ Gj T j ; P j ; nki;j ð10Þ  
oG
It is supposed that the function Gj represents the total free ¼ li ð19Þ
oni P ;T ;n6¼ni
energy in phase j, including of those corresponding to the
interactions of phase j with its interphases with the neigh- Eqs. (12) and (13) became, respectively,
bours phases. Function (10) allows the possibility that XX
dGj ¼ S j dT j þ V j dP j þ klki;j dnki;j ð20Þ
the values of G in the same phase would be different for i
one i-component at the different aggregation states k. X
dGj ¼ S j dT j þ V j dP j þ li;j dni;j ð21Þ
That means there is a possibility of a non-equilibrium
i
situation inside the same phase. This is possible only
because we are working with simplified structured phases We may use either Eq. (12) or Eq. (13) to calculate the
where sometimes we have included two or more real phases Gibb free energy in any j-phase depending on the situation
in the same simplified one. Anyway, if internal equilibrium of internal equilibrium inside the phase. From both equa-
in the phase is assumed, Eq. (10) simplifies as tions, the well known Gibbs–Duhem equation may be eas-
  ily obtained. In the same way, for the whole system the two
Gj ¼ Gj T j ; P j ; ni;j ð11Þ equations became, respectively,
X X X
Consequently, any differential change of G in the j-phase dG ¼ dGj ¼  SdT j þ V dP j
may be calculated by the equations: from Eq. (10): j j j
    XXX
oGj oGj þ lki;j dnki;j ð22Þ
dGj ¼ dT j þ dP j
oT P j ;nk oP T j ;nk i j k
i;j i;j X X X
X X  oGj  dG ¼ dGj ¼  SdT j þ V j dP j
þ dnki;j ð12Þ j j j
i k
on i;j P j ;T j ;n6¼nk XX
i;j
þ li;j dni;j ð23Þ
from Eq. (11): i j
   
oGj oGj where the physical sense of li, j (the chemical potential of i-
dGj ¼ dT j þ dP j component in j-phase) is the change in Gj produced by the
oT P j ;ni;j oP T j ;ni;j
X  oGj 
loss () or gain (+) of one mol of one specific i-component
þ dni;j ð13Þ (J/mol) in conditions of Pj, Tj and nj 6¼ ni, j constants. In
i
oni;j P j ;T j ;n6¼ni;j conclusion, the value of G in the system is the sum of the
values of G in every phase. It is very important to realise
Obviously, for the whole food system, the change in the
if the phases are in internal equilibrium and also if there
Gibbs free energy may be calculated as
are equilibria among them. Obviously, a non-equilibrium
X situation, inside the phases or among them, implies the
dG ¼ dGj ð14Þ
j spontaneous appearance of internal transport of energy
and/or mass.
Thus, obtaining Eqs. (15) and (16), depending on the equi-
librium situation inside the phases, 4.4. Equilibrium and driving forces
" # " #
X oGj  X oGj  The Gibbs energy defines the equilibrium and the events
dG ¼ dT j þ dP j
j
oT P j ;nk j
oP T j ;nk in the system.
i;j i;j
Gradients of the intensive variables (T, P and lki;j Þ repre-
X X X  oGj  sent driving forces responsible for fluxes of energy and/or
þ dnki;j ð15Þ
i j k
on i;j P j ;T j ;n6¼nk matter (Fito & Pastor, 1994). Gradients of T usually result
i;j
"  # " # in heat transfer, gradients of P result in deformations or/
X oGj X oGj  and bulk fluxes depending on the viscoelastic properties
dG ¼ dT j þ dP j of the system (Fito, 1994; Fito, Andrés, Chiralt, & Pardo,
j
oT P j ;ni;j j
oP T j ;ni;j
1996), whereas gradients of chemical potential produce
X X  oGj  mass transfer phenomena (Barat, Chiralt, & Fito, 1998).
þ dni;j ð16Þ
i j
oni;j P j ;T j ;n6¼ni;j Any change in the system which occurs at constant temper-
ature (isothermal) and constant pressure (isobar) must
Remembering the values of the first derivatives of function result in a decrease in G if the process occurs naturally,
G, we have or no change in G if the process is reversible. Taking into
 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185 181

account Eqs. (20) and (21), we have the conditions for the mechanical energy, even against the activities gradients
equilibrium: (as is in the case of inverse osmosis) or, in other cases,
chemical potentials gradients produce simultaneously mass
dG ¼ 0; dT ¼ 0; dP ¼ 0 and dlki;j ¼ 0 ð24Þ
transfer and mechanical deformations (as in cellular sys-
tems drying).
4.5. Transport mechanisms and rate equations The definition of descriptive matrix may be defined
finally as
The structure features in food may produce coupling of Mðxki;j ; V j ; P j ; T j ; lki;j Þ: ð31Þ
some of these transport phenomena (Gras, Vidal, Betoret,
Chiralt, & Fito, 2003). For instance, in cellular structures If it is assumed that there are not gradients of P, T, li in-
the loss or gain of water by any individual cell is always side any phase of the system, then Eq. (24) becomes
coupled with deformations and relaxations phenomena in
Mðxki;j ; V j ; P j ; T j ; li;j Þ ð32Þ
the insoluble matrix (cellular wall and membrane) promot-
ing the use of an extended definition of chemical potential The equilibrium condition arises when: grad P = 0; grad
(Seguı́, Fito, Albors, & Fito, 2006): T = 0; grad li, j = 0, in any phase of the system.
lext ext;0 Finally, the corresponding vectors, li, j, applied to the
i  li ¼ RT ln ai þ V i ðP  P 0 Þ ð25Þ
components affected by mass transport phenomena may
In these cases, it is possible to calculate the changes of G be included in the descriptive matrix (Gekas, González,
due to couplings of pressure and activities gradients, as well Sereno, Chiralt, & Fito, 1998). In a situation with only
as the contribution of the different gradients (pressure or water and solutes moving throughout the system, a com-
activities) to the global change of energy (Martı́nez-Mon- plete descriptive matrix will result as follows:
zó, Martı́nez-Navarrete, Chiralt, & Fito, 1998). These situ-
ations are usual in drying (both air-drying and osmotic Mðxki;j ; V j ; P j ; T j ; lw;j ; ls;j Þ ð33Þ
dehydration) of colloidal or cellular foods. In drying, for
instance, the water flux between two points (1 and 2) inside where lw, j and ls, j are the chemical potential of water and
the food system may be calculated by Eq. (19) (Wesseling solutes in all the phases of the system (where they may ex-
& Krishna, 1990): ist), respectively. In the case of air-drying or hydration, the
simplification that only water may move inside the system
J w ¼ Lw ðlw1  lw2 Þ ¼ RT ðln aw1  ln aw2 Þ ð26Þ
may be considered. In this case, the descriptive matrix may
In this equation the driving force is (ln aw1ln aw2) and may be organized as
be simplified as
Mðxki;j ; V j ; P j ; T j ; aw;j ; lw;j Þ ð34Þ
ðln aw1  ln aw2 Þ ffi xw1  xw2 ð27Þ
only if and when the requirements below can be accepted as The descriptive matrix has been presented again in Fig. 7
true: removing the information about aggregation states (dimen-
sion k) and including the vectors aw, j and lw, j, as well as
 The activity coefficient (cw) for water is close to one: the equations of mass balances.
aw ¼ cw xw  xw ð28Þ
5. Food process engineering: modelling of food operations
 The values of aw1 and aw2 are very close (the system is and processes
close to equilibrium):
aw1  aw2 ð29Þ We have introduced above the descriptive matrices (M)
as a useful tool both to describe food products recognizing
their structural features (physicochemical, thermodynami-
Only in this way Eq. (26) becomes similar to the well cal and biological) and for explaining and understating
known Fick equation, and these two conditions mentioned their properties. In food processes the food products suffer
above determine the conditions for its use in a rigorous controlled changes affecting their structure and properties.
way. The comparison between two descriptive matrices corre-
When we apply Eq. (26) to a system where points 1 and sponding to products in two different points of a process
2 are separated by a membrane (that allows water to flow permits one to identify when and where a change appears
across it but not to the solutes), it is necessary to use the as well as to analyze the mechanism(s) responsible for this
extended definition of chemical potential (Eq. (25)). In this change. This use of properties of M in the analysis of food
case Eq. (26) becomes processes will be discussed in the following points.
 
J w ¼ Lw lext ext
w1  lw2
5.1. Definition of unit operation and stage of change
¼ RT ðln aw1  ln aw2 Þ  V w ðP 1  P 2 Þ ð30Þ
The two terms on the right hand side of Eq. (30) let us The concept of unit operation came from the chemical
understand how the mass transport may be driven by engineering processes theory and concerns some stage in
 182
P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185
Fig. 7. Descriptive matrix of food system developed on a plane, including the vectors aw,j and lw,j.
 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185 183

a process where the product suffers one change in one phys- 5.2. Mass balances and transformed matrices: matrix of
ical or chemical property. Several unit operations consti- changes
tute a process where the raw material becomes the
desired final product throughout the successive changes. Considering one process with N stages of change and
Very often in a real process every unit operation is per- N + 1 different products flowing in the N + 1 strains (being
formed in a specific type of equipment (vs. drying opera- m0 the raw material and mN the final product), it must be
tion is performed in a drier). In this way, it is frequent to considered that
use both concepts (unit operation and the related equip-
N stages : f1; 2; 3; . . . ; n  1; n; n þ 1; . . . ; N g
ment) as similar or even equivalents.
In food processes, the complexity of the food systems N þ 1 strains fm0 ; m1; ; . . . ; mn1 ; mn ; m;nþ1 ; . . . ; mN g
usually results in some different changes suffered by the In Fig. 8, three steps of change of this process (n1, n,
food in the same equipment. Sometimes the major macro- n + 1) may be observed.
scopic change identified in the product (vs. loss of water in The subscripts of any matrix (vs. Ma, b) indicate that the
drying) results from several important changes in the sys- first one (a) is the subscript of the corresponding strain and
tem (in the state variables, in the structure, etc.). Sometimes the second one (b) is the strain of reference. Therefore, the
these changes occur coupled, but in other cases they are value of the overall mass of the system is always fixed as
consecutive in time. Then it is convenient to introduce a " #
new concept, stage of change, defined as one stage of the XXX b
xa ¼ xki;j ¼ ð35Þ
process where the product suffers only a significant change i j k
a
a
in its properties. According to this definition, one unit
operation may refer to one or more stages of change and The overall mass balance for the n stage will be
the same applies to the specific equipment used to perform mn  mn1 ¼ Dmn ð36Þ
this operation. Some examples of phenomena that demand
defining a stage of change may be If the value of Dmn is negative, it means a loss of matter by
the system. As a consequence, the values of the ratio ba
 When a phase of the system disappears or a new one would be 1 for any descriptive matrix, but to compare with
arises (as for instance when in drying operations the loss any other should be necessary to calculate one transformed
of water by the product causes the disappearance of the matrix with different values for a and b.
food-liquid phase, and only adsorbed water remains in As an example, for the stage n may be defined as
the solid). The descriptive matrix for strain mn1 Mn1, n1
 When a phase transition occurs (because the physical The descriptive matrix for strain mn Mn, n
and transport properties of the food changes, sometimes The transformed matrix of strain mn Mn1, n
dramatically). referred to the strain mn1
 When the pattern of the mass and energy transport phe-
nomena changes (vs. change of the mechanisms respon-
sible for these phenomena). The descriptive matrices of products corresponding to
 When the pattern of the chemical or microbiological these strains will be as follows:
" #
reactions changes (because of changes in temperature, XXX mn
pH, activation or inactivation of enzymes, etc.). xn;n ¼ xkij ¼ ¼1 ð37Þ
i j k
mn
n;n
" #
In conclusion, one process will be represented as a series XXX mn1
of stages of change in order to analyse the changes under- xn1;n1 ¼ xkij ¼ ¼1 ð38Þ
i j k
mn1
gone by the product after any one of them. n:1;n1

where the overall mass distributed in the system has been

arbitrarily considered as 1 kg. In this way, the physical
meaning of xn1,n1 and xn,n would be the mass fraction
of total mass corresponding to any component i in any
Δmn-1 Δmn Δmn+1 phase j and in any aggregation state k. In order to compare
the two mass distribution matrices in a balance, Eq.
mn-2 mn-1 mn mn+1 (36), the values of xn, n must be transformed by multiplying
n-1 n n+1 by the ratio mmn1
n
6¼ 1, defining
" #
Mn-1,n-1 Mn,n XXX mn
(base mn-1/mn-1) (base mn/mn) k
xn1;n ¼ xij ¼ 6¼ 1 ð39Þ
i j k
m n1
n1;n
Mn-1,n
(base mn/mn-1)
We obtain the transformed mass share-out matrix of strain
Fig. 8. Different stages of change. mn referred to those of the strain mn1:
184 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185

h i  h i
mn MC9;0 ¼ M0;9  M0;0 ð45Þ
xki;j ¼ xki;j ð40Þ
n1;n mn  1 n;n
In other words, the differences between the final product
(m9) and the raw material (m0) would be represented by this
Therefore, the descriptive matrix and the transformed
matrix of changes. Obviously, these ‘‘overall” values of
descriptive matrix of the product flowing throughout the
change have lost a lot of significant information about
n stage of change of the process will be
what has really happened in each one of the stages. For this
Mn;n ðxki;j ; V j ; P j ; T j ; li;j Þ ð41Þ reason, the analysis of the changes in every stage is of great
h i interest in order to know what elements do change in the
Mn1;n ðmn =mn1 Þxki;j ; V j ; P j ; T j ; li;j ð42Þ food descriptive matrix, after passing through the stage
and to quantify the level of change as well as to identify
And also the matrix of changes: the mechanisms responsible for these changes. In conclu-
sion, the matrix of one process with N stages of change
MCn;n1 ¼ Mn1;n  Mn1;n1 ð43Þ
  would be made up from (N + 1)2 matrices distributed as
MCn;n1 Dxki;j ; DP j ; DT j ; DV j ð44Þ follows:

 N + 1 Descriptive matrices,
5.3. The process matrix  0,5[(N + 1)2(N + 1)] Transformed matrices,
 0,5[(N + 1)2(N + 1)] Matrices of change.
In a process with nine stages of change, there are 10
strains of matter (m0, m1,. . .,m8, m9). The description of
the product of each strain may be represented by the corre- 6. Conclusions
sponding descriptive matrices (m0,0, m1,1, . . ., m9,9).
The matrix of process (MP) has been represented in Nowadays food quality and safety (food properties) are
Fig. 9. The cells in the diagonal of this MP are occupied the major concerns of consumers and food industry and the
by the ten descriptive matrices corresponding to the ten target for product and process design, structured foods are
process strains. The cells corresponding to the right-up side dissipative structures (Nicolis & Prigogine, 1977a) with a
of the diagonal would be occupied by the transformed high level of complexity. Therefore, the mathematical mod-
matrices and the left-down side of the diagonal by the els of food behaviour through operations and processes
matrices of change. must include all the information about the food character-
The matrix MC9,0 would be the result of the accumu- istics as well as those regarding the different phenomena
lated changes suffered by the strain (m0) after passing occurring during the processes: transport phenomena,
through the nine stages of change of the process, and has chemical reactions, phase transitions, structure changes,
been calculated as etc.

M 0,0 M 0,1 M 0,2 M 0,3 M 0,4 M 0,5 M 0,6 M 0,7 M 0,8 M 0,9

Descriptive Matrices
MC 1,0 M 1,1 M 1,2
M i,j

MC 2,1 M 2,2 M 2,3

m n-2 m n-1 mn m n+1
n-1 n n+1
MC 3,2 M 3,3 M 3,4
Mn-1,n- 1 Mn,n
(base mn-1/mn-1) (base mn/mn)
MC 4,3 M 4,4 M 4,5
Mn-1,n
(base mn/mn-1)
MC 5,4 M 5,5 M 5,6

Matrices of changes
MC 6,5 M 6,6 M 6,7
MC i,j

MC 7,6 M 7,7 M 7,8

MCn, n- 1 = Mn-1, n - Mn-1, n- 1
(base(mn-mn-1/mn-1)) (base(mn/mn-1)) (base(mn-1/mn-1))
MC8,7 M 8,8 M 8,9

MC 9,0 MC 9,8 M 9,9

Fig. 9. The matrix of process (MP).

 P. Fito et al. / Journal of Food Engineering 83 (2007) 173–185 185

To simplify foods as continuous (monophasic) and Fito, P., Andrés, A., Chiralt, A., & Pardo, A. (1996). Coupling of
homogeneous systems is a way to lose a lot of relevant hydrodynamic mechanism and deformation-relaxation phenomena
during vacuum treatments in solid porous food–liquid systems. Journal
information about their properties concerning their quality of Food Engineering, 27, 229–240.
and security factors. Fito, P., & Chiralt, A. (2001). Food matrix engineering: the use of
The SAFES methodology provides tools to describe, to the water-structure-functionality ensemble in dried food product
calculate and to develop food products and processes with development. Food Science and Technology International, 9,
enough information to predict their food quality and 151–156.
Fito, P., Chiralt, A., Betoret, N., Gras, M., Cháfer, M., Martı́nez-Monzó,
safety. J., et al. (2001). Vacuum impregnation and osmotic dehydration in
matrix engineering: application in functional fresh food development.
Acknowledgements Journal of Food Engineering, 49, 175–183.
Fito, P., Ortega-Rodriguez, E., & Barbosa-Canovas, G. V. (2001). Food
The authors acknowledge the Grants received from the engineering 2000. Maryland, USA: Aspen Publishers Inc.
Fito, P., & Pastor, R. (1994). On some non-diffusional mechanism
Spanish Ministry of Science and Education. occurring during vacuum osmotic dehydration. Journal of Food
Engineering, 21, 513–519.
References Gekas, V., González, C., Sereno, A., Chiralt, A., & Fito, P. (1998). Mass
transfer properties of osmotic solutions. I. Water activity and osmotic
Aguilera, J. M. (2006). Seligman lecture 2005 food product engineering: pressure. International Journal of Food Properties, 1–2(1), 95–112.
building the right structures. Journal of the Science of Food and Gras, M. L., Vidal, D., Betoret, N., Chiralt, A., & Fito, P. (2003). Calcium
Agriculture, 86(8), 1147–1155. fortification of vegetables by vacuum impregnation. Interactions with
Aguilera, J. M., & Baffico, P. (1997). Structure-mechanical properties of cellular matrix. Journal of Food Engineering, 56(2–3), 279–284.
heat-induced whey protein/cassava starch gels. Journal of Food Marcotte, M., & Le Maguer, M. (1991). Mass Transfer in cellular
Science, 62(5), 1048–1053. tissues. part 1. The Mathematical model. Journal of Food
Aguilera, J. M., Chiralt, A., & Fito, P. (2003a). Special issue on the Engineering, 13, 199.
international conference iberdesh 2002: process, structure and func- Martinez, N., Andres, A., Chiralt, A., & Fito, P., (1998). Termodinámica y
tionality. Food Science and Technology International, 9, 135. Cinética de sistemas alimento – entorno. Editorial UPV. Valencia
Aguilera, J. M., Chiralt, A., & Fito, P. (2003b). Food dehydration and (Spain).
product structure. Trends in Food Science and Technology, 14, 432–437. Martı́nez-Monzó, J., Martı́nez-Navarrete, N., Chiralt, A., & Fito, P.
Aguilera, J. M., & Stanley, D. W. (1999). Microstructural principles of food (1998). Mechanical and structural changes in apple (var. Granny
processing and engineering (2nd ed.). Maryland, USA: Aspen Publish- Smith) due to vacuum impregnation with cryoprotectant agents.
ers Inc. Journal of Food Science, 63, 499–503.
Barat, J. M., Chiralt, A., & Fito, P. (1998). Equilibrium in cellular food- Nicolis, G., & Prigogine, I. (1977a). Self organization in non equilibrium
osmotic systems: the role of the structure. Journal of Food Science, 63. systems. from dissipative structures to order through fluctuations. John
Barat, J. M., Fito, P., & Chiralt, A. (2001). Modelling of simultaneous Wiley and Sons, Inc..
mass transfer and structural changes in fruit tissues. Journal of Food Nicolis, G., & Prigogine, I. (1977b). Die erforschung des komplexen. auf
Engineering, 49, 77–85. dem weg zu einem neuen verständnis der naturwissenschaften. R. Piper
Bird, R. B., Stewart, W. E., & Lightfoot, E. N. (2002). Transport GmbH and Co. KG. München.
phenomena (2nd ed.). USA: John Wiley and Sons. Inc.. Seguı́, L., Fito, P. J., Albors, A., & Fito, P. (2006). Mass transfer
Crank, J. (1985). The mathematics of diffusion. New York: Oxford phenomena during the osmotic dehydration of apple isolated protop-
University Press. lasts (Malus domestica var. Fuji). Journal of Food Engineering, 77, 179.
Dickinson, E., & Stainsby, G., (1982). Colloids in food. Applied Science, Wesseling, J. A., & Krishna, R. (1990). Mass transfer. England: Ellis
London, New York. Horwood Limited.
ETP ‘‘Food for Life, 2006. The vision for 2020 and beyond. Confederation Zhiming, Y., & Le Maguer, M. (1996). Mathematical modelling and
of the food and drink industries of the EU. <[Link]>. simulation of mass transfer in osmotic dehydration processes. Part I:
Fito, P. (1994). Modelling of vacuum osmotic dehydration of foods. conceptual and mathematical models. Journal of Food Engineering,
Journal of Food Engineering, 22, 313–318. 29(3), 349–360.