See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/222037983

Thermophysical properties evolution of French partly baked bread during

freezing

Article  in  Food Research International · August 2004

DOI: 10.1016/j.foodres.2004.02.017

CITATIONS READS

65 969

3 authors:

Nasser Hamdami Jean-Yves Monteau

Isfahan University of Technology GEPEA CNRS, Oniris
86 PUBLICATIONS   1,478 CITATIONS    43 PUBLICATIONS   733 CITATIONS   

SEE PROFILE SEE PROFILE

Alain Le-Bail
École Nationale Vétérinaire, Agroalimentaire et de l'Alimentation Nantes-Atlantique
260 PUBLICATIONS   5,041 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

FreezeWave View project

Modeling and measurement of deep–frying, blanching and predrying effect on the quality of French fries View project

All content following this page was uploaded by Jean-Yves Monteau on 09 December 2015.

The user has requested enhancement of the downloaded file.

 Food Research International 37 (2004) 703–713
www.elsevier.com/locate/foodres

Thermophysical properties evolution of French partly baked

bread during freezing
a,b b,* b
Nasser Hamdami , Jean-Yves Monteau , Alain Le Bail
a
Department of Food Science and Technology, The University of Tabriz, 51664 Tabriz, Iran
b
UMR GEPEA (UA CNRS 6144 – SPI), ENITIAA, Rue de la G
eraudiere BP 82225, F-44322 Nantes Cedex 03, France
Received 1 October 2003; accepted 26 February 2004

Abstract

Few data are available on the thermophysical properties of the frozen partly baked breads. In this paper, thermophysical
properties, including apparent and true densities, specific heat, enthalpy and effective thermal conductivity were determined sepa-
rately for crumb and crust of partly baked bread. Total enthalpy of fusion, unfrozen water and solid specific heat were determined
by differential scanning calorimetry. The apparent specific heats were estimated in base of the unfrozen water at )40 °C and initial
freezing point. The effective thermal conductivity was measured with a line source probe in the range )35 to 25 °C. Four predictive
models of the effective thermal conductivity of porous food were developed (parallel, series, Krischer and Maxwell models). The
effective thermal conductivity predicted by Krischer model was in good agreement with the experimental data.
Ó 2004 Elsevier Ltd. All rights reserved.

Keywords: Bread; Thermophysical properties; Porous media; Freezing

1. Introduction studies on thermophysical properties of bakery products

are available in the literature. However, data on ther-
Partly baked bread has developed a lot in the past mophysical properties of dough and bakery products
decade. The objective of the part baking is to carry out during freezing are scarce as compared to other food
gelatinization and coagulation of gluten, without products, and no published information is available
reaching the coloring reactions of the crust (Roussel & concerning the thermal and physical properties of partly
Chiron, 2002). To retard staling and to extend shelf life baked bread during freezing. The methods available for
of bread, the part baked breads are often frozen. The measuring and modeling the thermophysical properties
frozen part baked breads represent half of the volume of of bakery products have been reviewed by different re-
frozen breads exported by the French industrialists to- searchers. Rask (1989) reviewed data and prediction
wards Northern Europe (Millet & Dougin, 1994). models of thermal properties of bakery products. Lind
Twenty two percent of the industrial bread was part (1991) presented measurement techniques and models of
baked frozen bread in 2001 with a growing market thermal properties of dough during freezing and thaw-
share. ing. More recently, Baik, Marcotte, Sablani, and Cas-
The study of the heat and mass transfers during taigne (2001) reviewed measurement techniques,
freezing, for the assessment of the product behavior prediction models and published data on thermophysical
under proposed process conditions and for the efficient properties of bakery products.
design of process and equipment, requires accurate val- While several modeling approaches have been pro-
ues of thermal and physical properties of bread. Many posed to predict thermal properties of a material at de-
sired conditions, none of them can be used over a wide
*
Corresponding author. Tel.: +33-251-785-481; fax: +33-251-785-
range of foods. The most promising approach is based
467. on chemical composition, temperature and physical
E-mail address: [email protected] (J.-Y. Monteau). characteristics. Besides, most structural models neglect

0963-9969/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodres.2004.02.017
704 N. Hamdami et al. / Food Research International 37 (2004) 703–713

Nomenclature

aw water activity keva–con equivalent thermal conductivity due to eva–

Cp specific (sensible) heat (J kg
1 °C
1 ) con phenomenon (W m
1 K
1 )
D diffusivity of water vapor in air, (m2 s
1 ) h temperature (°C)
E Euler’s constant, 0.5772157 hi initial freezing point (°C)
feva–con resistance factor against vapor transport hr reference temperature ðH ¼ 0Þ (°C)
fk structural factor or distribution factor of the e volume fraction (m3 /m3 product)
Krischer model ea porosity
H enthalpy (J kg
1 ) q density (kg m
3 )
Hfw enthalpy of water fusion (J kg
1 )
Subscripts
I current intensity through the heater wire
a correspond of pores (voids)
(A)
air air
L latent heat of fusion (J kg
1 )
app apparent
L0 latent heat of evaporation (J mol
1 )
u unfrozen water at low temperature (unfree-
P total pressure (Pa)
zable water)
Q power generated by the probe heater
c continue
(W m
1 )
d dispersed
r correlation coefficient
eva–con evaporation–condensation
r distance from the line heat source (m)
i ice or component i
R electric resistance (X m
1 )
Kri Krischer
R perfect gas constant, 8.314510 (J mol
1 °C
1 )
Max Maxwell
S slope of the linear portion of the plot of the
pa parallel
temperature vs. In(time)
s solid
T temperature (K)
se series
x mass fraction (kg/kg product)
sat saturated
t true
Greek letters tw total water
a thermal diffusivity (m2 s
1 ) w liquid water
k thermal conductivity (W m
1 °C
1 ) 0 at 0 °C

interactions between components although they can be of the apparent density of baked products, such as
significant. cookies and bread rolls, as a function of temperature,
The properties that are concerned by this study are the moisture and baking time. The true density is usually
density, the porosity, the unfrozen water mass fraction, determined with a pychnometer.
the specific heat, the enthalpy, and the thermal conduc- If the material is soft, the porosity can be determined
tivity. Two densities can be distinguished, namely the from volumes of compressed and non-compressed ma-
bulk and the true density. For the bulk or apparent terial (Zanoni, Peri, & Gianotti, 1995). More often it is
density, the pores volume is considered as a part of the calculated from the real and apparent densities (Miles
material volume. For the second one, the material vol- et al., 1983).
ume is the volume without the pores (McDonald & Sun, Water can be present in different states in a food
2001). The bulk density can be determined simply by during freezing. We arbitrarily choose that 100% of the
dividing the mass of the product by its volume. Volume frozenable water will frozen at a temperature of )40 °C.
can be measured with various displacement methods Therefore, the amount of unfrozen water at this tem-
(Bakshi & Yoon, 1984; Hwang & Hayakawa, 1980). The perature will be named ‘‘unfreezable water’’. The un-
volume can also be calculated from sample dimensions if freezable water mass fraction can be determined by
the object has a regular geometry (Christenson, Tong, & differential scanning calorimetry (DSC): it is the differ-
Lund, 1989). General models reported in the literature ence between the total water content and the amount of
allow the apparent density of any food product to be water detected by the fusion endotherm (Ross, 1978).
calculated as a function of the chemical composition, Another method was proposed by Pham (1987) using
moisture, porosity and temperature (Miles, van Beek, & enthalpy-temperature data and the Schwartzberg model
Veerkamp, 1983). In addition, semi-empirical (Hwang & of the apparent specific heat: the unfreezable water mass
Hayakawa, 1980) and empirical models (Bakshi & Yoon, fraction was expressed by equaling enthalpy expressions
1984) are available. These models describe the variation below and above the freezing point.
 N. Hamdami et al. / Food Research International 37 (2004) 703–713 705

Enthalpy and apparent specific heat are two key ture range using a specific model is sometimes more
thermal properties used to solve the non-linear heat accurate than measurements (Wang & Kolbe, 1991).
diffusion equation when taking into account phase Bakshi and Yoon (1984) used the line heat source probe
transition. Enthalpy is a state function for which a ref- method to measure thermal conductivity of bread
erence temperature has to be defined ()40 °C is often products at various moisture contents. They showed
chosen) (Heldman, 1982; Wang & Kolbe, 1991). Specific that average thermal conductivity of bread dough is
heat represents the rate of enthalpy change with the about 0.4 W m
1 °C
1 . Rask, in its review of published
temperature, dH =dh. Since latent heat removal occurs data (Rask, 1989), showed that the thermal conductivity
over the freezing temperature domain, it is usual to in- for crumb and crust are respectively about 0.3 and 0.05
clude the latent heat contribution in the specific heat, W m
1 °C
1 . These values depend on the fact that
which is then called the apparent specific heat, Cpapp . thermal conductivity is a function of moisture, apparent
Both experimental and mathematical modeling ap- density and therefore porosity. Thermal conductivity
proaches have been used to determine enthalpy and has been presented and modeled as a function of mois-
apparent specific heat of frozen foods. Several methods ture content or process temperature. It has now been
are used to measure the apparent specific heat and the recognized that thermal conductivity of foods, including
enthalpy as functions of temperature. These methods are bakery products, is also influenced by the amount of air
based on experimental results obtained from calori- fraction in the food material. A great variety of more or
metric measurements (Lind, 1991; Mannheim, Stein- less complex equations for calculating thermal conduc-
berg, Nelson, & Kendall, 1957; Tocci, Flores, & tivities exists in the literature. The most well known
Mascheroni, 1997). Bakshi and Yoon (1984) developed physical models are the parallel model, the perpendicu-
equations to estimate the specific heat of bread rolls by lar (or series) model, the dispersed phase model (or
considering it as a two-component material, water and Maxwell model) and the Krischer model (Miles et al.,
bread solids. They found the specific heat of bread solids 1983; Rahman, 1995).
to be lower than the estimated value by Mannheim et al. This research is a part of a study on freezing modeling
(1957) (1130.44 to compare with 1557.49 J kg
1 °C
1 ). of the part baked breads. It was designed to collect in-
Zanoni et al. (1995) showed that the average specific formation about thermal and physical properties of
heat of dough and crumb is about 2900 J kg
1 °C
1 , these products during freezing. The specific objectives
whereas the average specific heat of the crust is about were:
1600 J kg
1 °C
1 . By analyzing the data presented they 1. To determine the thermophysical properties of part
concluded that moisture is the major variable that in- baked bread as a function of moisture content and
fluences the specific heat of food products, and that temperature.
temperature has little effect provided there is no phase 2. To relate these properties to the samples composi-
change. Thus, there are semi-empirical models describ- tions using different mathematical model.
ing the variation of the specific heat of baked products Since there are structural differences between crumb
with moisture, such as the model of Bakshi and Yoon and crust, the properties mentioned above are evaluated
(1984) for bread rolls, and that of Christenson et al. separately for each part.
(1989) for bread, muffins and biscuits. Johnsson and
Skj€oldebrand (1983) setup semi-empirical models for
specific heat of bread as functions of moisture and 2. Theoretical section
temperature.
One of the most important thermal properties is This section is intended to document the physical
thermal conductivity. Measurement methods of thermal phenomena that determine differences in thermophysical
conductivity can be divided in two categories: steady- properties between the crumb and the crust, and to give
state methods (SSM) and unsteady state or transient some theoretical elements allowing the prediction of
methods (TM). Most researchers recommended TM these properties, elements used to obtain the results
(Nesvadba, 1982). Among TM, the simplest and most below.
widely used is the line heat source thermal conductivity
probe method. Advantages of this method are: (i) a 2.1. Differences between crust and crumb
short duration of the experiments (range 5–20 s) and (ii)
a reduced temperature rise of the sample (a few °C). By When dough is exposed to a high temperature in an
its theory assuming purely conductive–diffusive heat oven, the temperature of the dough surface rises and
transfer, this method is not appropriate for the deter- water from the outer layer evaporates. The moisture
mination of thermal conductivity at temperatures content of bread dough will thus change with time
slightly below the initial freezing temperatures because during baking and the crust will contain less water than
of the associated variation of the ice content. The pre- the crumb which represents the internal part of the
diction of thermal conductivity values in this tempera- bread (Johnsson & Skj€ oldebrand, 1983; Rask, 1989).
706 N. Hamdami et al. / Food Research International 37 (2004) 703–713

dxw ðhÞ
Another characteristic change is the increase in volume Cpapp ðhÞ ¼ CpW ðhÞxW ðhÞ þ Cpi ðhÞxi ðhÞCps ðhÞxs þ L :
of the product. This occurs in the initial stage of baking dh
and is mainly due to the expansion of the gas enclosed in ð3Þ
the porous dough structure. The pore size in the crust The ice content may be estimated with
may be very different from that in the crumb, the crust

hi
having a more dense structure than the crumb. Ac- xi ¼ ðxtw
xu Þ 1
if h < hi ; ð4Þ
cordingly, the apparent density of bread varies due to h
both the increase in the product volume and the de- xi ¼ 0 if h P hi : ð5Þ
crease in the water content during baking. At the end of
baking, the crust has low water content and a firm and
less porous structure; as a result, its apparent density 2.6. Enthalpy
values are not similar to those for the crumb, which has
a higher water content and is more porous (Rask, 1989; The relation between the enthalpy and the apparent
Zanoni et al., 1995). Consequently, the other thermal specific heat is given by Eq. (6), with H ¼ 0 for the
properties are different in the two parts. de Vries, Slui- reference temperature hr .
mer, and Bloksma (1989) proposed mathematical mod- Z h
els for heat and mass transfer in dough and crumb with H¼ Cpapp ðhÞdh: ð6Þ
hr
consideration of evaporation–condensation mechanism
in pores. Evaporation–condensation in the crumb
seemed to be a significant mechanism for heat transport
2.7. Evaporation–condensation phenomenon and thermal
in bread in comparison with conductive or radiative
conductivity
heat transfer. Nevertheless this transport might not to
be applied to bread crust according to Zanoni and Peri
If a temperature gradient is applied to a porous food
(1993).
material, moisture migration as vapor in the pore space
occurs. Water evaporates at the high temperature side,
2.2. Porosity diffuses in the pore space according to the vapor pres-
sure gradient caused by the temperature gradient, and
Using porosity definition and real and apparent condensates at low temperature side (de Vries et al.,
densities definition, Eq. (1) can be setup to calculate the 1989). Thus latent heat is transported through the pores.
porosity (Miles et al., 1983) By considering the effect of the latent heat transport, the
qapp effective thermal conductivity in pores was given by the
ea ¼ 1
: ð1Þ following equation:
qt
ka ðhÞ ¼ kair ðhÞ þ keva–con ðhÞfeva–con ; ð7Þ
2.3. Initial freezing point where keva–con is the equivalent thermal conductivity
due to the latent heat transport (evaporation–con-
The initial freezing point was defined as the temper- densation):
ature at which the rate of temperature drop of the DðT Þ P dPsat
freezing curve is slowed down (freezing plateau). keva–con ðT Þ ¼ L0 ðT Þaw ðT Þ: ð8Þ
RT P
aw Psat ðT Þ dT

2.4. Unfreezable water This equation is assumed to be applicable to the esti-

mation of the effective thermal conductivity of porous
The unfreezable water at low temperature can be food in our conditions.
measured by DSC, simply by difference between total
water content and the amount of water detected by the 2.8. Line heat source probe
fusion endotherm:
Hfw For
pffiffiffiffipoints close to the line heat source such that
xu ¼ xtw
: ð2Þ r=ð2 atÞ P 0:16, the temperature variation initiated by
L
a line heat source is given to better than 1% error
2.5. Apparent specific heat (McGinnis, 1987):


Q 4at
The apparent specific heat may be calculated by DT ðtÞ ¼ ln 2
E : ð9Þ
4pk r
summing specific heats of each component weighed
by their mass fractions, and by including energy released The conductivity may then be determined from the slope
by the ice fusion (Miles et al., 1983) S of the straight line DT vs. ln t:
 N. Hamdami et al. / Food Research International 37 (2004) 703–713 707

Q 3. Materials and methods

k¼ ; ð10Þ
4pS
where the power generated by the probe is 3.1. Partly baked bread preparation

Q ¼ I 2 R: ð11Þ Flour was purchased from a local milling plant

(Nantes, France). Flour analysis and dough recipe are
2.9. Thermal conductivity models provided in Table 1. Mixing was done in a spiral mixer
(VMI, Montaigu, France) and comprised two steps: a
Assuming that the food material is a four-phase low-speed mixing step (4 min, 100 rpm), and a high-
system consisting of air, water, ice and solids, the par- speed mixing step (11 min, 200 rpm). The temperature of
allel and series models are respectively represented by the dough after kneading was 23  0.5 °C. The dough
the following equations (Miles et al., 1983): was divided into round pieces (140  1 g) and fermented
for 30 min at ambient temperature. After first proofing
kpa ðhÞ ¼ ea ðhÞka ðhÞ þ es ðhÞks ðhÞ þ ew ðhÞkw ðhÞ þ ei ðhÞki ðhÞ; the dough was formed into 30 cm long cylinders and
ð12Þ proofed for 90 min at 27 °C and at a relative humidity

 greater than 90% (final proof). Then, the scarified pieces
ea ðhÞ es ðhÞ ew ðhÞ ei ðhÞ of dough were pre-baked for 12 min in a convection
kse ðhÞ ¼ 1 þ þ þ ; ð13Þ
ka ðhÞ ks ðhÞ kw ðhÞ ki ðhÞ oven at 150 °C. During the first minutes, water was in-
where ki , kw and ks are respectively the thermal con- jected in oven for decreasing the weight loss of samples.
ductivities of ice, water and solid (carbohydrate). The The final diameter of the bread cylinders was approxi-
volume fraction of any component, ei as a function of mately 57  2 mm. After pre-cooling at 10 °C and at a
temperature is obtained from the following equation relative humidity greater than 85% for 40 min, the partly
baked breads were packaged in two moisture imper-
qapp xi ðhÞ meable films, frozen at )30 °C and stored at )20 °C.
ei ðhÞ ¼ ; ð14Þ
qi
where xi and qi are respectively the mass fraction (kg/kg 3.2. Sample preparation for analysis
product) and the density of the components.
Krischer’ s model is based on two extremes in thermal In this study, cylindrical breads for density and po-
conductivity values, one being derived from the parallel rosity, crumb samples for unfreezable water and thermal
model and the other being derived from the series conductivity, and crust slices for unfeezable water,
model, whilst the real value of thermal conductivity conductivity, density and porosity were used. They were
should be somewhat in between these two extremes. The obtained according to the following procedure
distribution factor fk is a weighing factor. (Fig. 1(a)): firstly, the frozen partly baked bread was
1 divided in three parts. Then, the cylindrical bread sam-
kKri ðhÞ ¼ : ð15Þ
1
fk
þ ksefðhÞ
k ple was obtained from central part (15 cm long). Finally,
kpa ðhÞ
the crust was removed from cylindrical bread sample
The thermal conductivity of a food by the Maxwell and the interior portion remained as crumb. The so-
equation is defined as (Miles et al., 1983): called crust was characterized by a dense structure with
a thickness of about 3 mm. Separating the crust frac-
2kc þ kd
2ed ðkc
kd Þ
kMax ¼ kc ; ð16Þ tions in this way can be done with good reproducibility.
2kc þ kd þ ed ðkc
kd Þ All types of sample (bread, crumb and crust) were
where kc and kd are respectively the thermal conduc- placed in moisture impermeable polyethylene films. The
tivities of the continuous and discontinuous phase, and packaged samples were thawed at 20 °C for 2 h before
ed the volume fraction of the dispersed phase. the experiments.

Table 1
Chemical components of the flour and dough recipe used in experiments
Flour component Content Method Dough recipe Content (g)
Moisture (% DM) 15.25 NFa ISOb 712 Flour 100
Ash content (% DM) 0.61 NF V 03720 Water 60
Protein content (% DM) 10.80 NF 5 7/MS Salt 2.2
Hagberg falling number (s) 284 NF V 03703 Compressed yeast 3
Z
el
eny (%) 33 NF ISO 5529 Baking aid mix 0.7
a
NF, French Standard.
b
ISO, International Standard Organization.
708 N. Hamdami et al. / Food Research International 37 (2004) 703–713

Fig. 1. (a) Preparation scheme of the samples (part baked bread, crumb and crust samples) and (b) probe positions in the crumb and crust samples.

3.3. Bulk and true densities and porosity estimated as the area under the DSC base line and ex-
pressed in J kg
1 of sample (Tocci et al., 1997). The cor-
Firstly, the cylindrical bread sample was weighed and responding total enthalpy of fusion and specific heat were
measurements of apparent volume, of length and of evaluated with the software of the calorimetric apparatus.
diameter were taken. Secondly, the crumb portion was Because of the span in temperature of the calori-
removed, and the apparent volume and the weight of metric peak, the experimental curve was not properly
crust sample were measured. The apparent volumes of located on the temperature scale. In order to have an
bread and crust were measured by rape seed displace- acceptable correspondence between the initial freezing
ment method. Finally, the volume of crumb was calcu- point and the apparent specific heat function Cpapp , the
lated by subtraction of the crust volume from bread apparent specific heat of samples was estimated from the
volume. The true density of samples was measured using Eq. (3). The temperature dependence of the amount of
a helium pychnometer (AccuPyc 1330, Micromeritics, ice and unfrozen water were taken into account by using
USA). Eq. (1) was then used to calculate the sample Eqs. (4) and (5). The initial freezing point, hi , was de-
porosities. duced from the temperature history of sample. En-
thalpy, H , was determined using Eq. (6), with the
3.4. Total water content reference temperature selected as )40 °C.

The total water content of samples was determined in 3.7. Effective thermal conductivity
triplicate by drying 5 g of sample in a forced convection
oven at 103 °C for 16 h. The samples were cooled in a 3.7.1. Thermal conductivity measurement
desiccator and weighed by an analytical balance (sensi- Thermal conductivity was determined with a line
tivity 0.01 mg). source probe, purchased from V. E. Sweat (Texas,
A and M University). The thermal conductivity probe
3.5. Unfreezable water mass fraction was similar to the one described by Sweat and Parmelee
(1978). A version of the thermal conductivity measure-
The unfreezable water weight fraction was deter- ment system developed by McGinnis (1987) was used to
mined using DSC, according to the method presented in supply power, measure current and record temperatures
theoretical section (Eq. (2)). and voltage (Fig. 2). Firstly, the probe was inserted into
the crumb and crust samples according to the Fig. 1(b).
3.6. Apparent specific heat and enthalpy To prevent the crust samples compression during con-
ductivity measurement, the crust slices were cut into
A differential scanning calorimeter (DSC 92, Seta-
ram, Caluire, France) was used to evaluate the total
enthalpy of the phase change and the specific heat of dry
matters of the samples at atmospheric pressure. Samples
(100–140 mg) were removed from different locations in
crumb and crust, and hermetically closed in aluminum
pans. They were frozen in situ in the calorimeter by the
way of liquid nitrogen. Experiments were realized five
times from )50 °C and up to 40 °C at a heating rate of
2 °C/min. An empty pan was used as a reference. The
base line was obtained from a scan realized with two
empty pans. Total enthalpy of the phase change was Fig. 2. Setup of the thermal conductivity measurement system.
 N. Hamdami et al. / Food Research International 37 (2004) 703–713 709

sub-fractions (30  30  3 mm) and 13 laminates of these ki ðhÞ ¼ 2:2196
6:248  10
3 h þ 1:0154  10
4 h2 ;
sub-fractions were placed in a rectangular container
ð17Þ
with internal dimension 30  30  40 mm, and packaged
in moisture impermeable film to prevent the sample kw ðhÞ ¼ 0:57109 þ 1:7625  10
3 h
6:7036  10
6 h2 ;
dehydration. Then, the assembly (probe + sample) was
kept at a given temperature in an air-freezer (Froilabo- ð18Þ
S.A., Ozoir, France) having a temperature stability of
kw h ¼ 0:20141 þ 1:38410  10
3 h þ 4:3312  10
6 h2 ;
0.1 °C. The probe heater was connected to a DC
power supply (Redelec 60430, Noailles, France) and a ð19Þ
multimeter (Fluke, John Fluke MFG. Co., Inc., USA) where ka is obtained from the Eq. (7). In all the
to measure the current intensity. The accuracy of the calculations, the values used for ice, water and
current measurement in the heater circuit was 0.1 mA. solid density were respectively 917, 1000 (Miles
The temperature–time and voltage–time data were et al., 1983), and 1463  1.2 kg m
3 (obtained from
continuously collected by a digital recorder (Datalog 20, experiments).
AOIP, Evry, France) which was bi-directionally inter- The volume fraction of discontinuous phase, ed , is
faced with a PC in order to program measurements. The obtained from Eq. (14).
record of voltage was used to find the initial time of the For Maxwell model application to porous materials
heating. including three or more components, a stepwise proce-
dure is required. At each step the mean thermal con-
3.7.2. Thermal conductivity calculation ductivity of pairs of components is found. In this study a
The samples thermal conductivity was calculated us- three-step Maxwell model, developed by Hamdami,
ing the Eq. (10) where Q is calculated by Eq. (11). Monteau, and Le Bail (2003), was used for thermal
To obtain satisfactory linearity of temperature vs. conductivity prediction. In the first step, the two phases
lnðtimeÞ plot, the procedure was standardized by (1) the considered were (1) continuous water phase, (2) dis-
choice of a power level to increase the temperature up to continuous ice phase. In the second step, the phases were
10 °C (initial temperature basis), (2) using a duration of (1) continuous solid phase, (2) discontinuous water–ice
8 s, and (3) by accepting thermal conductivity values phase. Finally, in the third step, the phases were (1)
measured only when r2 > 0:98. To obtain these condi- continuous solid–water–ice phase, (2) discontinuous air
tions, a current value of 0.09–0.17 A was necessary. phase. The model parameters fk and feva–con were esti-
Seven measurements were made in each sample. For mated by fitting the model calculated values of effective
each sample, the conductivity reported is the mean of thermal conductivity to experimental ones.
seven measurements.
The thermal conductivity experiments were carried
out at selected sample temperature, )35, )30, )25, 1, 15 4. Results and discussion
and 25 °C on the crumb and crust samples, with mois-
ture contents of 45% and 27% (wet basis), respectively. 4.1. Density and porosity

3.7.3. Calibration The weight and apparent volume measurements at

The probe calibration was realized by measuring the room temperature on the partly baked bread samples
thermal conductivity of glycerin and of a gel of 0.5% agar and their crumb and crust, showed that the crust forms
and water at selected temperatures between 5 and 25 °C. 25.0  1.9% of volume and 37.8  1.9% of weight of the
partly baked bread samples.
3.7.4. Thermal conductivity model development Table 2 gives the experimental apparent and true
The crumb and crust of bread can be analyzed as a density values of the different samples (crumb and crust,
multicomponent three-phase system which, below the average of 10 tests). As expected, crust has an apparent
initial freezing point, represents a complicated dynamic and true density greater than those of the crumb, due to
complex of four fractions, continuously changing their the porosity and the total water mass fraction of the
ratios: dry substance (proteins, lipids, carbohydrates, crust lower than for the crumb.
mineral salts, etc.), water, ice crystals and gaseous phase
(including air and water vapor). In this study, four
physical models, parallel, series, Maxwell and Krischer 4.2. Unfreezable water, apparent specific heat and
models, were tested for the thermal conductivity pre- enthalpy
diction of crumb and crust.
The thermal conductivities of ice, water and solid The values of the crumb and crust unfreezable water
(carbohydrates) are calculated with the following equa- mass fraction were estimated by application of the
tions (Singh, 1992): Eq. (2) to 10 samples for which total enthalpy of fusion
710 N. Hamdami et al. / Food Research International 37 (2004) 703–713

Table 2
Apparent and true density, porosity, unfreezable and total water mass fractions of the crumb and crust samples
Property Crust Crumb
Mean value SD Mean value SD

3
Apparent density (kg m ) 332.8 6.0 181.7 2.6
True density (kg m
3 ) 1390 5.9 1291.5 10.5
Porosity 0.760 0.017 0.8592 0.0030
Unfreezable water mass fraction (kg kg DM
1 ) 0.3364 0.0060 0.413 0.049
Total water mass fraction (kg/kg product) 0.273 0.025 0.4529 0.0032

have been previously measured. Table 2 gives the un- Figs. 3 and 4, because of an important freezable water
freezable water mass fraction values ðxu Þ in the crumb mass fraction, crumb has a more intense phase change
and crust samples. The difference in the unfreezable peak and a greater enthalpy than that of crust. When the
water values of the crumb and crust samples can be samples are frozen, the enthalpies for the crust and the
explained by both physical and chemical changes oc- crumb are very close.
curring during the baking process (Lind, 1991).
The average initial freezing point of the crumb and 4.3. Effective thermal conductivity values
crust samples was found to be )5.7 °C 0.2 and )15 to
)24 °C, respectively (average of five experiments). The The values of thermal conductivity measured by the
large variation of the initial freezing point of the crust line source probe method were reproducible (2% un-
samples was due to the great freezable water content certainty ‘‘experimental standard deviation’’).
variation of the samples. Fig. 5 shows the effective thermal conductivity
The following equation gives the specific heat of dry values measured for the crumb at two different probe
bread as a function of temperature. It was obtained positions (axial and radial) for three positive temper-
from eight tests of DSC: atures: 2.5, 15 and 23 °C. This figure shows that there
is not important difference between the results. Thus it
Cps ðhÞ ¼ 4:582h þ 1138: ð20Þ
can be concluded that the effective thermal conduc-
This shows a very good agreement with 1130.44 tivity measurements were independent on the probe
J kg
1 °C
1 for bread solids as calculated by Bakshi and position. Moreover, it can be noticed that the points
Yoon (1984). are along a straight line.
From the Eqs. (3) and (6), the apparent specific heat The Figs. 6 and 7 present the thermal conductivity
and the enthalpy of both frozen and unfrozen products variation vs. temperature for the crumb and the crust.
were adequately calculated as a function of temperature Values given by the models are also represented here.
while knowing total and unfrozen water content, and These figures show that effective thermal conductivity
initial freezing point. is strongly dependent on the porous structure, the
Figs. 3 and 4, show the evolution of the apparent freezable water content and the temperature. The effec-
specific heat and enthalpy values as a function of tem- tive thermal conductivity above the initial freezing point
perature during freezing for a crumb and a crust sample increases with increasing the temperature due to the
in respect of the given data in Table 2. As shown in increase of the thermal conductivity of the components

Fig. 3. Evolution of the apparent specific heat values for the crumb and crust samples.
 N. Hamdami et al. / Food Research International 37 (2004) 703–713 711

Fig. 4. Evolution of the enthalpy values for the crumb and crust samples.

Fig. 5. Effective thermal conductivity values measured for evolution of the apparent specific heat values for crumb at two different probe positions
(radial and axial).

Fig. 6. Experimental and predicted thermal conductivity values for crumb.

(crumb and crust cases) and to the evaporation–con- the state change to ice. Then, it decreases very slightly.
densation effect (crumb case only). In the sub-freezing However, if the freezable water content value is less
temperatures, however, the effective thermal conductiv- important (crust case), the stage of effective thermal
ity is dependent on freezable water content. If its value is conductivity increase cannot be observed with decreas-
important (crumb case), the effective thermal conduc- ing the temperature: for all the temperatures, the points
tivity increases with decreasing the temperature, due to are disposed along a straight line.
712 N. Hamdami et al. / Food Research International 37 (2004) 703–713

Fig. 7. Experimental and predicted thermal conductivity values for crust.

The comparison between the experimental and pre- by Krischer’s model are in good agreement with the
dicted effective thermal conductivity values show that: experimental values in both cases. It can be concluded
(i) Series model gives always under-estimated thermal that Krischer’s model is able to correlate the heat
conductivity values for crumb and crust, (ii) Parallel transfer in the porous structure of bread crumb and
model gives upper-estimated thermal conductivity val- crust better than the others models. These better results
ues for crumb. However the predicted values approach are due to the presence of the empirical curve-fitting
the experimental ones for the crust, (iii) The predicted factor, fk .
effective thermal conductivity values by Maxwell model From Table 3 it can be observed that the identified
agree with the experimental ones only for the crumb values of fk and feva–con for the crust samples are less
samples, (iv) The predicted thermal conductivity values important than for the porous crumb (assuming
aw ¼ 1). Generally, the structural factor fk (which
stands for the part of the series structure) decreases with
Table 3 an increasing moisture content and a decreasing poros-
Identified values of the resistance factor and structural factor
ity (Hamdami et al., 2003; Maroulis, Krokida, & Rah-
Property Crumb Crust man, 2002). The Krischer’s model tends then toward the
Porosity 0.85 0.76 parallel model. By considering that the crust porosity
Total water mass fraction 0.46 0.25 and moisture were lower than those of the crumb, it can
Resistance factor feva–con 0.37 0.09 be deduced from Table 3 that the moisture content effect
Structural factor fk 0.27 0.09
is less important than the porosity effect on the struc-

Fig. 8. Effect of the evaporation–condensation phenomenon on the crumb and crust samples thermal conductivity with xtw ¼ 0:45 and 0.25, and
ea ¼ 0:85 and 0.76.
 N. Hamdami et al. / Food Research International 37 (2004) 703–713 713

tural factor. The difference in feva–con for the crumb and Hamdami, N., Monteau, J.-Y., & Le Bail, A. (2003). Effective thermal
crust samples can be explained by the lower value of free conductivity of a high porosity model food at above and sub-
freezing temperatures. International Journal of Refrigeration, 26(7),
water for evaporation and the less porous structure in 809–816.
the crust (Fig. 8 and Table 3). By comparing the pre- Heldman, D. R. (1982). Food properties during freezing. Food
dictions of the Krischer’s model with the experimental Technology, 36, 92–96.
values, with and without considering the evaporation– Hwang, M. P., & Hayakawa, K. (1980). Bulk densities of cookies
condensation phenomenon (Fig. 8), it can be seen that undergoing commercial baking processes. Journal of Food Science,
45, 1400.
evaporation–condensation plays an important part in Johnsson, C., & Skj€ oldebrand, C. (1983). Thermal properties of bread
heat transfer in crumb compared to the crust. This ob- during baking. In Proceedings of third international congress on
servation confirms the hypothesis by de Vries et al. engineering and food (pp. 333–341). Dublin: Elsevier Applied
(1989) that evaporation–condensation in the crumb re- Science Publishers.
sults in more rapid heat transfer. Lind, I. (1991). The measurement and prediction of thermal properties
of food during freezing and thawing – A review with particular
reference to meat and dough. Journal of Food Engineering, 13, 285–
319.
5. Conclusion Mannheim, H. C., Steinberg, M. P., Nelson, A. I., & Kendall, T. W. (1957).
The specific heat content of bread. Food Technology, 7, 384–388.
In this paper, some thermophysical properties, such Maroulis, Z. B., Krokida, M. K., & Rahman, M. S. (2002). A
structural generic model to predict the effective thermal conduc-
as porosity, apparent and true densities, specific heat, tivity of fruits and vegetables during drying. Journal of Food
enthalpy and effective thermal conductivity, of partly Engineering, 52(1), 47–52.
baked bread crumb and crust were estimated. The re- McDonald, K., & Sun, D.-W. (2001). The formation of pores and their
sults showed that: (i) The crust forms 25.0  1.9% in effects in a cooked beef product on the efficiency of vacuum
volume and 37.8  1.9% in weight of the partly baked cooling. Journal of Food Engineering, 47, 175–183.
McGinnis, D. S. (1987). Automated line-heat source system for the
bread samples. (ii) The total and unfrozen water mass measurement of thermal conductivity and diffusivity. Canadian
fractions are respectively 0.45 and 0.23 in crumb, and Agricultural Engineering, 29(2), 201–207.
0.27 and 0.24 in crust. (iii) The Krischer’s model cor- Miles, C. A., van Beek, G., & Veerkamp, C. H. (1983). Calculation of
relates thermal conductivity data in bread crumb and thermophysical properties of foods. In R. Jowitt, F. Escher, B.
crust better than the other models (parallel, series and Hallstrom, H. F. Meffert, W. E. L. Spiess, & G. Vos (Eds.),
Physical properties of food (pp. 269–313). London: Applied Science
Maxwell models) due to the presence of an extra em- Publishers.
pirical curve-fitting factor. (iv) In the crust, the mea- Millet, P., & Dougin, Y. (1994). Boulangerie industrielle francßaise.
sured thermal conductivity lies closer to the parallel L’^age adulte. Industries des C
er
eales (April–June), 45–50.
model than to the series model compared to the crumb. Nesvadba, P. (1982). Methods for the measurement of thermal
(v) Evaporation–condensation has a significant part in conductivity and diffusivity of foodstuffs. Journal of Food Engi-
neering, 1, 93–115.
heat transfer of the crumb. Pham, Q. T. (1987). Calculation of bound water in frozen food.
These results could be used to model coupled heat and Journal of Food Science, 52(1), 210–212.
mass transfer during freezing of the partly baked breads. Rahman, M. S. (1995). Food properties handbook. Boca Raton, FL:
CRC Press.
Rask, C. (1989). Thermal properties of dough and bakery products: A
Acknowledgements review of published data. Journal of Food Engineering, 9, 167–193.
Ross, K. D. (1978). Differential scanning calorimetry of nonfreezable
water in solute-macromolecule-water systems. Journal of Food
Acknowledgements to Jacques Laurenceau, Luc Science, 43(6), 1812–1815.
Guihard and Olivier Rioux for technical support. Roussel, P., & Chiron, H. (2002). Les pains francßais: Evolution, qualit
e,
production. Vesoul: Mae-Erti Editeurs.
Singh, R. P. (1992). Heating and cooling processes for foods. In D. R.
References Heldman & D. B. Lund (Eds.), Handbook of food engineering (pp.
247–276). New York: Marcel Dekker Inc..
Baik, O. D., Marcotte, M., Sablani, S. S., & Castaigne, F. (2001). Sweat, V. E., & Parmelee, C. E. (1978). Measurement of thermal
Thermal and physical properties of bakery products. Critical conductivity of dairy products and margarines. Journal of Food
Review in Food Science and Nutrition, 41(5), 321–352. Process Engineering, 2, 187–197.
Bakshi, A. S., & Yoon, J. (1984). Thermophysical properties of bread Tocci, A. M., Flores, E. S. E., & Mascheroni, R. H. (1997). Enthalpy, heat
rolls during baking. Lebensmittel Wissenschaft und Technologie, 17, capacity and thermal conductivity of boneless mutton between )40
90–93. and +40 °C. Lebensmittel Wissenschaft und Technologie, 30, 184–191.
Christenson, M. E., Tong, C. H., & Lund, D. B. (1989). Physical Wang, D. Q., & Kolbe, E. (1991). Thermal properties of surimi
properties of baked products as functions of moisture and analysed using DSC. Journal of Food Science, 56(2), 302–308.
temperature. Journal of Food Processing and Preservation, 13, Zanoni, B., & Peri, C. (1993). A study of the bread-baking process. I: a
201–217. phenomenological model. Journal of Food Engineering, 19(4), 389–
de Vries, U., Sluimer, P., & Bloksma, A. H. (1989). A quantitative 398.
model for heat transport in dough and crumb during baking. In Zanoni, B., Peri, C., & Gianotti, R. (1995). Determination of the
Proceedings of cereal science and technology in Sweden (pp. 174– thermal diffusivity of bread as a function of porosity. Journal of
188). Ystad, Sweden: Lund University. Food Engineering, 26(4), 497–510.

View publication stats