ArticlePDF Available

Abstract and Figures

The basic rheological properties of two Persian wheat flours - Tajan (11% protein) and Back Cross Roshan (8% protein) and two Australian wheat flours-JANZ (12.9% protein) and Rosella (8.6% protein) have been characterized. These properties have been interpreted via a damage function model. All samples could be reasonably well described by the damage function model with a power-law relaxation spectrum. Although the shear stresses in the Australian samples were higher, the relaxation parameter G(1) and power-law exponent p for the Australian varieties were lower than those for the Persian samples and the damage functions were different. Since protein contents were different, this indicates that the amount of protein is not the sole determinant of softness in the samples. The damage function f was also calculated for all samples. This function gives a measure of the softening due to working or kneading of the samples at a given strain level.
Content may be subject to copyright.
34305-1 Applied Rheology
Volume 19 · Issue 3
The basic rheological properties of two Persian wheat flours - Tajan (11 % protein) and Back Cross Roshan (8 %
protein) and two Australian wheat flours-JANZ (12.9 % protein) and Rosella (8.6 % protein) have been charac-
terized. These properties have been interpreted via a damage function model. All samples could be reasonably
well described by the damage function model with a power-law relaxation spectrum. Although the shear stress-
es in the Australian samples were higher, the relaxation parameter G(1) and power-law exponent pfor the Aus-
tralian varieties were lower than those for the Persian samples and the damage functions were different. Since
protein contents were different, this indicates that the amount of protein is not the sole determinant of soft-
ness in the samples. The damage function fwas also calculated for all samples. This function gives a measure
of the softening due to working or kneading of the samples at a given strain level.
Die grundlegenden rheologischen Eigenschaften von zwei persischen Weizenmehlsorten, Tajan (11 % Eiweiß)
und Back Cross Roshan (8 % Eiweiß) sowie zwei australische Weizensorten, Flours-JANZ (12.9 % Eiweiß) und
Rosella (8.6 % Eiweiß) werden im Rahmen dieses Beitrags geschildert. Das damage-Funktionsmodell interpre-
tiert die Eigenschaften der Proben. Alle Proben konnten durch die damage-Funktion mit einem Potenzgesetz-
Relaxationszeitspektrum recht gut beschrieben werden. Obwohl die Schubspannungen in den australischen
Proben höher waren, waren die Entspannungsparameter G(1) und Potenzgesetz-Exponenten für die australi-
schen Sorten niedriger als die für die persischen Proben. Weiterhin wiesen die damage-Funktionen ebenfalls
Unterschiede auf. Die damage-Funktion fwurde ebenfalls für alle Proben berechnet. Diese Funktion gibt ein
Maß für die Anpassung der Verarbeitung sowie des Knetens der Proben bei einem bestimmten Belastungsni-
veau an.
Les propriétés rhéologiques de deux farines de blé persanes - Tajan (11 % de protéines) et Back Cross Roshan (8 %
de protéines) et de deux farines de blé australiennes-JANZ (12,9 % de protéines) et Rosella (8,6 % de protéines)
ont été caractérisées. Ces propriétés ont été interprétées par le biais d'un modèle d'endommagement. Tous les
échantillons ont pu être raisonnablement bien décrits par le modèle d'endommagement avec une loi puissan-
ce du spectre de relaxation. Bien que les contraintes de cisaillement soient plus élevées pour les échantillons
australiens, le paramètre de relaxation G(1) et l'exposant p de la loi de puissance des échantillons australiens
sont inférieurs à ceux des échantillons persans et les fonctions d'endommagement sont différentes. Puisque la
quantité de protéines est différente, cela indique que celle-ci n'est pas le seul critère déterminant le comporte-
ment des échantillons. La fonction d'endommagement fa également été calculée pour tous les échantillons.
Cette fonction donne une mesure de l'assouplissement dû au malaxage de l'échantillon pour une déformation
Key words: dough rheology, wheat flour, relaxation, steady shear, elongation, damage function, power-law
A Comparison of the Rheology of Four Wheat Flour Doughs
via a Damage Function Model
Shiva Amirkaveei1*, ShaoCong Dai2, Marcus Newberry3, Fuzhong Qi2,
Mohammad Shahedi1and Roger I.Tanner2
1Department of Food Science, College of Agriculture, Isfahan University of Technology,
Isfahan, Iran
2Department of Aerospace, Mechanical and Mechatronic Engineering, University of Sydney,
Sydney, NSW 2006, Australia
3CSIRO, Food Futures National Research Flagship and Division of Plant Industry, Canberra,
ACT 2601, Australia
* Email:
Fax: x98.311.3912254
Received: 10.9.2008, Final version: 23.2.2009
© Appl. Rheol. 19 (2009) 34305
DOI: 10.1515/arh-2009-0012
Download Date | 6/14/19 2:34 PM
Dough is made by combining flour, water and ener-
gy. The addition of sufficient mechanical energy
produces the distribution and hydration of flour
particles, allowing the formation of a unique vis-
coelastic material that exhibits both solid and liq-
uid properties [1, 2]. A small amount of other ingre-
dients such as yeast, salt, and preservatives is often
added. Dough is a complex system. It is basically a
cohesive three-dimensional cross-linked network
of gluten in which starch granules (of size ~ 10 mm)
are embedded [3]. Not only the hydrated protein
aggregates, but also the starch matrix and starch-
protein interaction give rise to viscoelastic proper-
ties and all these interactions affect the funda-
mental rheology [4]. In addition, dough rheology
plays an important role in the quality of the final
baking products and its understanding will lead to
progress in the food processing industries [3, 4].
Experimental rheological characterization
of wheat flour dough is desirable. It gives valu-
able information concerning the quality of raw
material, the textural characterization of fin-
ished products and properties needed for the
design and development of new equipment
[5 7]. During the past decade many attempts
have been made to apply fundamental rheo-
logical methods to dough. The early work
(1932 1937) of Schofield and Scott Blair estab-
lished the solid-like behaviour of dough and since
then there have been many investigations [2].
Bloksma (1962) was a pioneer in the measure-
ment of viscoelastic properties of wheat flour
doughs using creep tests [6].
Quantitative characterization of the rheolog-
ical properties of wheat flour and gluten doughs
has been difficult because of their complicated
strain dependent (non-linear) visco-elastic behav-
iour [6]. Bagley et al. [7] simulated dough rheo-
logical properties in uniaxial compression using
the upper convected Maxwell model [7, 8].
Leonard et al. [9] characterized the non-linear vis-
coelastic properties of hard and soft wheat flour
doughs using the Bird-Carreau constitutive mod-
el. Sweep strain experiments and stress relaxation
tests on Australian strong flour doughs indicate
that tests at higher shear strains can differentiate
flour types , and small and large strain rheology
of wheat gluten and comparison with the parent
flours have been investigated by Uthayakumaran
et al. [8]. Also it has been shown that stress relax-
ation tests were clearly useful in distinguishing
between long- and short-mixing flour [9]. Relax-
ation behaviour of dough and its gluten and
gluten fractions in both strong and wheat flours
has also been considered [10].
A striking feature of dough rheology is the
power-law behaviour of stress relaxation and
oscillatory response in the linear range. Gabriele
et al. [11] discussed the use of a power-law mem-
ory function for foods, including doughs, in the
linear range. Venkataraman and Winter [12] had
previously considered crosslinking polydi-
methylsiloxane (PDMS) near its gel point and
used a power-law memory function for the lin-
ear viscoelastic case and for the PDMS start-up
of simple shearing in finite strain. We note that
PDMS is much less shear-thinning than dough
and no “damping function” was used. Ng et al.
[13] used a Lodge-type model with a power-law
memory and a “damping function” of the KBKZ
type pioneered by Wagner [14], but this does not
equate to the “damage function” introduced in
[15] in 2007. Subsequent work by Ng et al. [16],
published in 2008, also uses the KBKZ damping
function. It does consider both shear and elon-
gation, but it is clear that the “damping function”
used is different to the damage function used in
[15] and [4]. Here we will use the damage func-
tion concept. We also note the recent work of
Berzin et al (2007) on In-line property measure-
ment of wheat starch during extrusion, which
will be of interest to industry [17].
Tanner et al. [15, 4] presented the “damage
function' model to describe multiple rheological
tests in 2008. To test this new model a new set
of experiments on bread dough included small
strain oscillatory behaviour, larger strain oscilla-
tory behaviour, simple shearing beginning from
rest, uniaxial elongation beginning from rest,
relaxation after sudden shear and recoil from
elongation. Wheat flour from various classes and
cultivars of wheat display great diversity in their
functional properties. The variation in function-
al properties of wheat cultivars are attributed
largely to its gluten quality and quantity [9]. Also
physical and chemical properties of wheat flours
are influenced by environmental factors and the
protein quality of hard wheat is more stable than
that of soft wheat and all of these factors affect
bread quality [18] so that bread making potential
is highly influenced by cultivar and environmen-
tal interactions.
Applied Rheology
Volume 19 · Issue 3
Download Date | 6/14/19 2:34 PM
Two Persian wheat varieties have been cho-
sen; Tajan is known as a variety with good bread-
making quality while Back Cross Roshan flour
cannot be used for making good quality bread.
The objective of this study was rheological char-
acterization of Persian wheat flour doughs using
the damage function model (if applicable) and
comparison of their characteristics with some
Australian varieties. It seems that the damage
function model, for the tests made, is a reason-
able way to interpret and compare different flour
doughs from different wheat breeds, as will be
discussed below.
Materials used for this study were two brands of
Australian commercial flour (JANZ and Rosella)
and two brands of Persian pure cultivars (Back
Cross Roshan and Tajan). Samples of Australian
wheat prepared by CSIRO plant industry, ACT and
samples of Persian wheat cultivars (Back Cross
Roshan and Tajan) were obtained from the Iran
Agriculture Research Center. All wheat samples
were tempered for 18 hours to 15.5 % moisture
and milled into straight grade flour using a Buh-
ler laboratory mill [19]. Table 1 shows some of the
chemical analyses and the mixograph values for
these four samples. Then the doughs were pro-
duced in a 10 g mixograph by mixing flour, dis-
tilled water and 200 mg of salt up to optimum
dough development time
The dough used for relaxation, oscillatory and
steady shear measurements was stored in a
sealed bag after mixing. All of the shearing ex-
periments were carried out on a Paar Physica
MCR 301 rheometer. Here the parallel plates had
a diameter of 25 mm and the gap was set to 2 mm
for measurements. Slippage during testing was
prevented by two pieces of sandpaper that had
been glued to the parallel plates. Before testing
calibration of the rheometer was performed,
then the sample was mounted on the lower plate
and compressed between the plates by moving
down the upper plate to a set gap. Excess dough
was trimmed and the edge of the sample was
coated with petroleum jelly to prevent moisture
loss. After that the sample was allowed to relax
for 40 minutes.
Steady shear measurements have been
done at shear rates of 0.001, 0.01, 0.1 and 1.0 s-1.
For oscillatory shear measurements, frequency
sweeps in the frequency range of 0.01 – 30 Hz
were conducted at strain amplitudes of 0.1, 1, 5
and 10 %, respectively. After each test the sam-
ple was unloaded and a new sample was up-
loaded for the new frequency sweep. Stress
relaxation tests after application of a sudden
shear strain at time zero were done at four dif-
ferent initial shear strains: 0.1, 1, 5 and 10%.
For the elongation measurements, the sample
after mixing was first formed in an aluminum
cylinder with an inside diameter of 30.6 mm and
stored in a sealed bag to relax for 40 minutes. The
sample then was transferred to an Instron 5564
rheometer to perform the elongation tests. The
sample was confined between two parallel
plates. One is a fixed lower plate with a diame-
ter of 31.0 mm, the other is a moving upper plate
with a diameter of 30.3 mm. The force in the sam-
ple was found using a sensitive load cell which
had a measuring range of 10 N. Before testing the
rheometer was calibrated without any loading.
The aluminum cylinder containing the dough
was then fitted to the upper plate. Simultane-
ously, the upper plate was brought down gently
until the sample was compressed properly
between the plates. Then the excess dough after
compressing was wiped upwards or downwards
evenly to wrap the upper and lower plates,
respectively, so that the sample would be well
attached to the plates during elongation. In addi-
tion, to prevent moisture loss, a thin layer of Shell
petroleum jelly was applied to the outside of the
34305-3 Applied Rheology
Volume 19 · Issue 3
Table 1:
Chemical analysis and mix-
ing properties of wheat
flour for four varieties.
Download Date | 6/14/19 2:34 PM
sample. Then the mounted sample was com-
pressed to the set gap of 9 mm and allowed to
relax for a further 10 minutes. The tests were con-
ducted at elongation rates of 0.001, 0.01 and 0.1
s-1, respectively. The samples were stretched until
they were broken. At the same time a digital cam-
era was used to capture the diameter of the cylin-
drical dough specimen at a rate of 25 fps and the
results were downloaded to a computer as a
movie. The circularity of the cross section was
generally excellent, as judged from specimen
after cutting in half. Hence, the elongation stress
s= F/Acould be calculated, where Fis the elon-
gation force measured by the load cell and Ais
the minimum cross sectional area of the sample.
The actual rate of the extension at the centre of
the sample was found from the variation of
diameter in the sample with time [4].
A modified Lodge elastic fluid model with a dam-
age function has been used by Tanner et al. [4, 15]
to describe the mechanics and rheology of bread
dough. It has been shown that the rheology of
this soft, starch-filled solid could be described by
a constitutive equation of the Lodge form:
where s(t) is the stress tensor at time t, Pthe pres-
sure, Ithe unit tensor and C-1(t’) is the Finger strain
tensor at time t’ computed relative to the con-
figuration at the present time t. The memory
function m is assumed to be of the power-law
form [4, 15]:
where the constants pand G(1) can be found from
relaxation testing or by oscillatory testing at
small amplitudes [15]. Note that the constant G(1)
here signifies the numerical value of the stress
relaxation G(t) at t= 1 s – the constant does not
have the dimensions of stress. The “damage
function” ƒis assumed to be a function of the
Hencky strain at time t, computed relative to the
state of rest which is assumed to occur for times
less than zero.
In the relaxation test we also used the fol-
lowing power-law equation for the relaxation
function, which follows from the use of Eq. 2 [4]
Results for storage and loss moduli G’(w), G’’(w)
as functions of the applied frequency w(rad/s) are
also described by a power-law relation [15],G’(w)
= G’(w)w-p, with a similar relation for G’’. Also it is
known that the phase angle G’’ is given, for a
power-law material by the constant value
So the phase angle is defined by the exponent p;
only two constants, G(1) and pare needed to
describe the linear viscoelastic behaviour of
these solids, which are assumed to be incom-
pressible [20].
From Eqs. 1 and 2 the shear stress sis a func-
tion of p, G(1), g
·and t; G(1) has the dimension of
Pa·sp. It was shown [15] that with the model one
can test if it is reasonably concordant with exper-
iment by plotting s/g
·pas a function of strain. The
stress function can be evaluated for various pval-
ues. Instead of plotting against g, we use the
Hencky strain (see Figure 5) which is related to g
by [21]:
Elongational stress divided by e
·pwas also plot-
ted as a function of eH= e
·tfor our elongation mea-
surements (see Figures 7 - 8). The superposition
is reasonable and fracture usually occurs at eH~
3. We can again use Eqs. 1 and 2 to compute the
response to a steady elongation (e
imposed at t= 0 and we find the following equa-
tion for calculating the stress (when f= 1.0)
Applied Rheology
Volume 19 · Issue 3
Download Date | 6/14/19 2:34 PM
Thus the model predicts for a given strain eHthat
the tensile stress sis also proportional to e
indicated by the experimental data (Figures 7 –
8). For a shearing flow of shear rate gbeginning
at t= 0, one can evaluate the integrals explicitly,
finding the viscometric functions with the same
form of shear rate and strain dependence as in
elongation [15, 4].
where g= g
·t. To evaluate the damage function f
we divide the average measured stresses at a giv-
en strain with those calculated from Eqs. 6 and 7.
In the two Persian samples the rheological char-
acterizations of flour were conducted by using
different rheological tests including stress relax-
ation, steady shear, oscillation and elongation.
The elastic nature of dough is quite apparent in
a large deformation, such as simple shearing
flow. Since the total shear strain acquired in the
deformation is g= g
·t, an increase in stress with
gis consistent with the idea that a part of the
stress response is elastic in nature (stress ~
strain). It may be argued that the initial part of
the response is independent of the strain rate, up
to a strain of O(10-1) [22]. After a peak strain, the
shear stress quickly decreases, indicating a par-
tial failure of the elastic network. The peak strain
is an increasing function of the strain rate. In Fig-
ure 1, the shear stresses are plotted versus strain
for both Persian varieties and the Australian sam-
ples which will be considered in the following at
shear rates of 1 and 0.01 s-1. Values show that
Tajan has a higher shear stress and fracture point
than Back Cross. It seems that its protein network
is stronger than Back Cross. However, despite dif-
ferent protein contents, the shear values are very
close and there is no significant difference
between the two varieties.
The time dependence of the relaxation mod-
ulus in the linear limit (with strain magnitude of
0.1 %) for both Persian wheat flours and the Aus-
tralian samples is shown in Figure 2. The rheome-
34305-5 Applied Rheology
Volume 19 · Issue 3
Figure 1 (left above):
Shear stress as a function of
strain for Persian doughs
and Australian doughs at
strain rates 1 and 0.01 s-1.
Figure 2 (right above):
The stress relaxation of Per-
sian doughs and Australian
doughs at an applied strain
amplitude of 0.1 %.
Figure 3 (left below):
Load versus elongation for
Persian flour doughs in
elongation rates 0.1 and
0.001 s-1.
Figure 4 (right below):
Effect of variety and its
protein content on uniaxial
elongation and rupture
strain. Protein content was
11 % for Tajan and 8 % for
Back Cross Roshan. Elonga-
tion test was conducted at
elongation rate e
·= 0.01 s-1.
Download Date | 6/14/19 2:34 PM
ter in the stress relaxation mode gives the stress
relaxation modulus, G(t), as a function of time and
as the quota of the measured stress, s(t), to the
initial strain [4]. Despite different protein con-
tents, these two varieties showed very similar rhe-
ological behaviour as measured by the stress
relaxation modulus. The initial value for Tajan is a
little higher but after a while (t= 1 s), it becomes
lower than Back Cross (Figure 2). The linear relax-
ation function is assumed to be described by Eq. 3.
The slope of G(t) is - p, and the value of Gat t= 1 s
gives G(1). Relaxation data show that at higher
strains the rheological behaviors for both flours
are almost the same but in the linear region
(strain = 0.1 %) a difference of behaviour with a
difference in power law index has been observed
(pfor Tajan = 0.38, pfor Back Cross = 0.32).
The load in elongation versus Hencky strain
is shown in Figure 3. At an elongation rate of
0.001 s-1 the rupture strain for Tajan is nearly the
same as for Back Cross Roshan but the load is
higher, while at 0.1 s-1 the rupture strain for Tajan
is higher and it is around 2.8 since that of Back
Cross Roshan is 2.4. A possible reason may be the
difference in protein content, so that the sample
with the lower protein content, at higher strains,
has a lower resistance to rupture because of the
weak protein matrix.
Stress increases slowly at low strains or
extension level (Figure 4). The sharp increase in
stress with increasing strain level is known as
strain hardening, and then the dough sample rup-
tured. The strain and viscosity at which the dough
samples broke or ruptured were simple measures
of dough strain-hardening properties and corre-
spond to extensibility and maximum resistance to
extension measured in traditional dough testing.
Increasing protein content increases the strain
hardening properties of dough, as measured by
the elongational rupture and rupture strain; Tajan
(11 % protein content) has a higher stress value
than Back Cross (8 % protein content).
We considered differences in rheological behav-
iour of two Australian wheat flours (JANZ and
Rosella) in comparison to the Persian samples
(Tajan and Back Cross Roshan) in various rheo-
logical tests including steady shear, stress relax-
ation, shear oscillation and elongation. Going
back to Figure 1, the results of steady shear at two
shear rates 1 and 0.01 s-1 can be compared for all
samples. The results show that the fracture point
happens at higher stress values for both Aus-
tralian flours than for the Persian wheat flours.
It seems that elasticity in the protein matrix for
them is higher than for the Persian samples. Cal-
culating p for different samples shows that p val-
ues for Australian wheat samples are between
0.2 – 0.3 while they are higher for Persian wheat
flour - between 0.3 – 0.4 (Figure 2). Bearing in
mind that each of the shear rates used a new
sample, with some inevitable variability, the
agreement is reasonable. Estimates of errors in
the slop pare given; we compared different sam-
ples and relaxation oscillatory data (only the
mean dis given here).
Table 2 shows calculated values for these
four varieties, data show that power-law para-
meters for Australian wheat samples are lower
than for Persian wheat, however G(t) is lower for
the Australian samples. Uthayakumaran et al. [8,
23] have shown that the pvalue is not apprecia-
bly affected by changes in water content for a
strong flour dough, although G(t) changes. Also
Applied Rheology
Volume 19 · Issue 3
Figure 5 (left):
Shear stress data for Tajan
dough plotted t/g
·pas a
function of the Hencky
strain. Fracture occurs at
eH~ 2.9.
Figure 6:
Shear stress data for JANZ
dough plotted t/g
·pas a
function of the Hencky
strain. Fracture occurs at
eH~ 3.
Table 2:
Calculated constants for
wheat flour samples
(din degrees).
Download Date | 6/14/19 2:34 PM
with comparable protein contents and p values,
there is no correlation between protein content
and pvalue for Persian wheat samples while for
Australian wheat flour there is. Comparing pfor
Persian wheat shows that Tajan with the higher
protein content has a higher p. On the other
hand, for Australian samples, a lower protein
content sample shows a slightly higher value of
p. By using Eqs. 1 and 2 and plotting s/g
·pand s/e
as a function of Hencky strain, it can be clearly
observed for the four varieties that the damage
function model correlation is reasonable (Figures
5 – 6 for shear and Figures 7 – 8 for elongation of
one Persian (Tajan) dough and one Australian
(JANZ) dough).
Strain hardening, a rapid increase in viscos-
ity at higher strain level, is thought to be respon-
sible for the ability of dough to expand and retain
the gas involved during fermentation and bak-
ing. Studies have shown that flours with good
baking quality tend to have much greater strain
hardening than flours that perform poorly in bak-
ing [24]. Figures 9 and 10 show plots of stress ver-
sus Hencky strain at elongation rates of 0.1 and
0.001 s-1. Result shows that at 0.1 s-1 the rupture
for Persian samples and Rosella happened at eH~
2.7 while for JANZ rupture happens at eH~ 3.1. In
0.001 s-1 for Rosella and Back Cross rupture hap-
pens in eH~ 2.5 but for Tajan and JANZ it happens
at a higher strain value eH~ 3, while rupture
'stress' for Persian samples is higher than that for
Australian samples. This increase in rupture
stress and rupture strain with increasing protein
content was confirmed by Uthayakumaran et al.
(1999) using a small scale extension tester. Also
increasing the glutenin to gliadin ratio increases
the elongational rupture viscosity but decreases
the rupture strain [23].
For steady elongation and shear discussed
above, we can then find f for elongation and
shear by using the result of Table 2 and Figures 5
- 8, plus Eqs. 1 and 2. The damage stress reduction
factor fcan be plotted as a function of Hencky
strain, Figure 11, for elongational data of four vari-
eties. (Formulas for fare given in the caption of
Figure 11). Note that the drop from f=1 at eH= 0
is very fast for all samples while the decrease pat-
tern in both sets of samples is seen to be differ-
ent. The damage function fmust be 1.0 at small
enough strains; however, by a strain magnitude
eHof only 0.01 fdrops abruptly to the values
shown in Figure 11, and the rapid rise for eH< 0.01
34305-7 Applied Rheology
Volume 19 · Issue 3
Figure 7 (left above):
Elongational stress data for
Tajan plotted s/e
·pas a
function of the Hencky
strain (p = 0.38). Fracture
happens when eH~2.9.
Figure 8 (right above):
Elongational stress data for
JANZ plotted s/e
·pas a func-
tion of the Hencky strain
(p = 0.27). Fracture happens
when eH~3.2.
Figure 9 (left below):
Elongational stress versus
Hencky strain for all sam-
ples at an elongation rate
·= 0.1 s-1.
Figure 10 (right below):
Elongational stress versus
Hencky strain for all sam-
ples at an elongation rate
·= 0.001 s-1.
Download Date | 6/14/19 2:34 PM
is not shown in this figure. This rapid change is
discussed in [15]. The sudden drop is possibly due
to reduced starch particle-protein network inter-
action at quite low stresses. Figure 11 shows that
the Persian samples are softer at Hencky strains
less than unity.
We also conducted G' and G'’ measurements
at strain amplitudes of 0.001, 0.01, 0.05 and 0.1 for
all samples. As it has been shown in different stud-
ies the range of maximum strain that results in a
linear response is very small for dough [8]. For
wheat flour, the maximum shear strain permitted
in the linear range is around 0.001 (0.1 %). We used
this value in our (linear) oscillatory tests. Results
for storage module G'(w) as a function of applied
frequency w(rad/s) are shown in Figure 12. The
slope of the curves here is again pfor small ampli-
tudes (Eq. 3). Exponent pis close to the value of p
from relaxation as it has been shown in Figure 12.
It is in the range of 0.2 - 0.3 for Australian samples
and 0.3 - 0.4 for Persian samples and G' > G'’' indi-
cating soft solid behavior.
We have shown that it is possible to differenti-
ate between different flours by subjecting the
doughs to different rheological tests. It seems
that stress relaxation experiments at a strain of
0.1 % and determination of power-law pand G(1)
could be good indices for dough behaviour in that
stronger and more elastic flours show lower G(1)
and p, but it seems that these differences are not
solely related to protein content. The damage
function fis also clearly different for the various
flours. Another relatively high strain deforma-
tion that brings out subtle differences between
different flour types is the constant shear rate
flow between two parallel plates [22], where the
strain increases linearly with the time of defor-
mation. Typically Australian flours have higher
values of stress than Persian samples, so it seems
that they have stronger matrices. Of course in
shear flow, it should be mentioned that the val-
ue of the stress peaks are strain-rate dependent.
They are consistent with the assumption of a
finite elastic network in gluten.
In oscillation, with a strain amplitude of
0.1 %, the power-law index is very close to the
value in relaxation but we believe oscillation
tests are not as convenient as data from relax-
ation tests, so that they were not used as a crite-
rion for discriminating flours with similar protein
qualities. We note that stress relaxation tests at
various strain amplitudes are a convenient way
of finding the damage function [4]. Testing of the
samples by extensional testing seems to be a
good way for determination of sample differ-
ences since extensional properties of the flours
in constant elongation rate are quite different.
Elongation rate is very important in dough reac-
tions and baking so that at lower elongation
rates the reaction of dough is related to the pro-
tein matrix strength, protein quantity and sam-
ple behaviour.
Typically, the wheat breed has a significant
effect on dough rheology so that in all experi-
ments Australian flours showed stronger resis-
tance and a more elastic protein network. It
should be noticed that determination of opti-
mum water absorption and working are very
important factors for producing reliable and
reproducible data in experiments.
Comparing four rheological tests for these four
samples reveals that the breed has a remarkable
effect on the protein quality but differences in
protein quantity cannot be the sole reason for
having different rheological behaviour. It seems
that using chemical and analytical information
about the wheat protein besides rheological
characteristics can give a good perspective of the
protein efficiency for producing special products.
Finding a correlation between rheological and
Applied Rheology
Volume 19 · Issue 3
Figure 11 (left):
Damage function f as a
function of Hencky strain
for elongation measure-
ments of four varieties.
The f function for JANZ:
f = - 0.0958logeH+ 0.0669,
Rosella: f = - 0.0841logeH+
0.0542, Tajan:
f = - 0.0144logeH+ 0.0561
and for Back Cross:
f = - 0.0695logeH+ 0.0025
(eH £0.1) and
f = - 0.0303logeH+ 0.0417
(eH> 0.1). The rise in f for eH
is not shown;
clearly f(0) = 1.0.
Figure 12:
G’ versus frequency at two
strain amplitudes (0.1 and
10 %) for Australian and Per-
sian samples. The slope of
power-law fitted curves has
been shown for each flour.
Download Date | 6/14/19 2:34 PM
analytical data may permit one to define some
indices for each variety as a control factor,
although it might be very hard to characterize
such flours. As data showed the damage model
seems quite useful for the description and com-
parison of the rheological behaviour of different
doughs. With an overall look at different experi-
ments, it seems that relaxation and elongation
and steady shear are reliable tests for ranking
flours but oscillatory tests also give useful com-
plementary information.
The authors would like to thank you Iranian gov-
ernment for financial support and also Mr. Erwan
Bertevas, school of Aerospace, Mechanical and
Mechatronic Engineering at the University of
[1] Lee L, Ng PKW, Whallon JH, Steffe JF: Relationship
between rheological properties and microstruc-
tural characteristics of non developed, partially
developed and developed dough. Cereal Chem.
78 (2001) 447.
[2] Campos DT, Steffe JF, Ng PKW: Rheological
behavior of undeveloped and developed wheat
dough. Cereal Chem. 74(1997) 489.
[3] Phan-Thien N, Newberry M, Tanner RI: Nonlinear
oscillatory flow of a soft solid-like viscoelastic
material. J. Non Newt. Fluid Mech. 92 (2000) 67.
[4] Tanner RI, Dai Sh C, Qi F: Bread dough rheology
and recoil,2 Recoil and relaxation. J. Non Newt.
Fluid Mech. 143 (2007) 107.
[5] Li W, Dobraszczyk BJ, Scofield JD: Stress relax-
ation behavior of wheat dough, gluten and
gluten protein fractions. Cereal Chem. 80 (2003)
[6] Rao VK, Mulvaney SJ, Dextert JE: Rheological
characterization of long- and short- mixing flour
based on stress relaxation. J. Cereal Sci. 31 (2000)
[7] Bagley B, Christianson DD, Martindale JA: Uni-
axial compression of a hard wheat flour
dough:data analysis using the upper convected
Maxwell model. J. Texture Stud. 29 (1988) 289.
[8] Uthayakumaran S, Newberry M, Phan-Thien N,
Tanner RI: Small and large strain rheology of
wheat gluten dough. Rheol. Acta 41 (2002) 162.
[9] Leonard A-L, Cisneros F, Kokini JK: Use of the rub-
ber elasticity theory to characterize the vis-
coelastic properties of wheat flour doughs. Cere-
al Chem. 76 (1999) 243.
[10] Mikhaylenko GG, Czuchajawska Z, Baik B-K, Kid-
well KK: Environmental influence on flour com-
position, Dough rheology and baking quality of
spring wheat. Cereal Chem. 77 (2000)507.
[11] Gabriele D, de Cindio B, D'Antona P: A weak gel
model for foods, Rheol. Acta 40( 2001) 120.
[12] Venkataraman SK, Winter HH: Finite shear strain
behaviour of a crosslinking polydimethylsiloxane
near its gel point. Rheol. Acta 29( 1990) 423.
[13] Ng TSK, McKinley GH, Padmanabhan M: Linear to
nonlinear rheology of wheat flour dough. Appl.
Rheol. 16 (2006) 265.
[14] Tanner RI, Engineering Rheology, Oxford Univer-
sity Press (2000).
[15] Tanner RI, Qi F, Dai Sh C: Bread dough rheology
and recoil,1.Rheology. J. Non Newt. Fluid Mech.
148 (2008) 33.
[16] NgTSK, McKinley GH:Power law gels at finite
strains: the nonlinear rheology of gluten gels. J.
Rheol. 52 (2008) 417.
[17] Berzin F, Tara A, Tighzert L: In-line measurement
of the viscous behaviour of wheat strach during
extrusion. Application to starch cationisation.
Appl. Rheol. 17 (2007) 21222.
[18] Safari-Ardi M, Phan-Thien, N: Stress relaxation
and oscillatory tests to distinguish between
doughs prepared from wheat flours of different
varietal origin. Cereal Chem. 75 (1998) 80.
[19] Miller KA, Hoseney RC: Dynamic rheological prop-
erties of wheat starch-gluten doughs. Cereal
Chem. 76(1999) 105.
[20] Wang C, Dai Sh C, Tanner RI: On the comperssi-
bility of bread dough. Korea-Australia Rheo. J. 18
(2006) 127.
[21] Kitoko V, Keentok M, Tanner RI: Study of shear
and elongational flow of solidifying propylene
melt for low deforametion rate. Korea-Austalia
Rheo. J. 15 (2003) 63.
[22] Phan-Thien N, Safari-Ardi M, Morales-Patino A:
Oscillatory and simple shear flows of a flour
water dough: a constitutive model. Rheol. Acta
36 (1997) 38.
[23] Uthayakumaran S, Newberry M, Keentok M,
Stoddard FL, Bekes F: Basic rheology of bread
dough with modified protein content and
glutenin-to-gliadin ratios, Cereal Chem. 77
(2000) 744.
[24] van Vliet T, Janssen AM, Bloksma AH Walstra P:
Strain hardening of dough as a requirment for
gas retention. J. Texture Stud. 23(1992) 439.
34305-9 Applied Rheology
Volume 19 · Issue 3
Download Date | 6/14/19 2:34 PM
... Low availability of water and lipids causes starch granules to achieve only partial gelatinization and form a solid network of interconnected and swollen starch grains [13]. The rheological properties of the dough are complex and are viscoelastic systems showing the properties of multiphase liquids or solids [12]. Sucrose is the most frequently applied sweetener in the confectionery industry, and it is also applied as a texture-forming agent. ...
... The rheological properties of the dough and the texture of the sponge cake are significantly affected by the ingredients, the dominant ones being protein and sugar [2,[9][10][11]. The development of the dough structure depends on the flour used and, above all, on the content of starch and gluten [12]. Starch is responsible for the viscoelastic properties and determines the textural characteristics of the sponge cake. ...
Full-text available
Changes in the rheological properties of dough, as well as the microstructural, mechanical, and sensory properties of sponge cakes, as a function of the substitution of sucrose in a formulation with maltitol, erythritol, and trehalose are described. Moreover, the relationship between the examined properties was investigated. The replacement of sucrose with maltitol or trehalose did not affect the consistency index, whereas erythritol caused a decrease in its value. X-ray tomography was used to obtain the 2D and 3D microstructures of sponge cakes. All studied sweeteners caused the sponge cakes to have a typical porous structure. Erythritol and maltitol resulted in about 50% of the pores being smaller than 0.019 mm2 and 50% of the pores being larger than 0.032 mm2. Trehalose resulted in a homogeneous microstructure, 98% of whose pores were similar in size (0.019 to 0.032 mm2). The sponge cakes with polyols had a higher structure index than did the trehalose and sucrose samples. There were also significant differences in color parameters (lightness and chromaticity). The crust of the sponge cake with sweeteners was lighter and had a less saturated color than the crust of the sponge cake with sucrose. The sponge cake with maltitol was the most similar to the sponge cake with sucrose, mainly due to the mechanical and sensory properties. Trehalose led to the samples having high adhesiveness, which may limit its application as a sucrose substitute in sponge cake. Sensory properties were strongly correlated to cohesiveness, adhesiveness, and springiness and did not correlate to the 2D and 3D microstructures. It was found that 100% replacement of sucrose allows for a porous structure to be obtained. These results confirm that it is not the structure, but most of all the flavor, that determines the sensory perception of the sponge cakes.
... The protein network gives the dough most of its rheological characteristics, while water acts as a solvent and the starch granules act as modulus modifiers, increasing the modulus as their percentage increases. The protein network and starch granules make a gel-like structure, and as such, the linear rheology of dough is well described by the critical gel model, also called the power law model (Ng et al. 2008, Tanner et al. 2008, Amirkaveei et al. 2009, Tanner et al. 2011. The simplicity of the power-law model allows for an easy characterization of the dough linear rheology through two main parameters: the gel strength and gel exponent. ...
Full-text available
Doughs made from blends of white wheat flour and wholegrain wheat flour mixed with chickpea flour were studied rheologically and morphologically in an effort to understand the effect of chickpea flour on the rheology of doughs. The doughs were subjected to a variety of rheological tests to understand how chickpea flour substitution affects wheat dough rheological behavior. Strain sweep, frequency sweep and temperature sweep tests were carried out. Strain sweep tests show that chickpea flour dough has a significantly strong nonlinear behavior compared to white wheat flour dough and wholegrain wheat flour dough. Frequency sweep tests show the solid like behavior of the doughs. From the theoretical analysis of strain sweep and frequency sweep experimental results, we conclude that up to 10% substitution of chickpea flour to wheat flour does not alter the doughs rheological behavior. Temperature sweep tests show that the chickpea flour dough follows zero-order gelatinization kinetics, whereas white wheat flour dough and wholegrain wheat flour doughs follow first-order gelatinization kinetics.
... Bread dough rheology has been modeled by several researchers. Bird-Carreau constitutive equation was used to describe the viscoelastic behavior of water-wheat flour bread dough (Amirkaveei et al., 2009). Wagner constitutive model was also used to describe bread dough by fitting small amplitude oscillatory shear data to a Maxwell relaxation model and then fitting damping functions using step-strain experiments (Wang and Kokini, 1995). ...
Dough blended with rocket leaves powder was subjected to small and large amplitude oscillatory shears. Small amplitude oscillatory shear data were fitted to a discrete relaxation model of elastic solids and to a critical gel model. The small amplitude relaxation spectrum was thereafter used to calculate the LAOS predictions of various large deformation models. The LAOS theoretical calculations using the Phan-Thien model showed good agreement with the first harmonic stress data, and only qualitative agreement with the third and the fifth harmonic stress values. Lissajous curves showed dissimilarity in shape between the experimental data and Phan-Thien model. The network model of Sim et al. (2003). Did not have the butterfly shape displayed in the Phan-Thien model, but it provided a worse fit to stress harmonics than the Phan-Thien model. An improved damage function was proposed, where time effect on network damage was taken into consideration, and fits to stress harmonics and to Lissajous stress-strain curves were significantly improved.
... Dough can be seen as a starch particle suspension dispersed in a concentrated biopolymer solution where interactions on the molecular to the micron length scale determine the overall rheological properties, baking process, and final bread quality. In dough rheology the strong link to bread-making and the instantaneous measurement of dough quality prior the bread baking rationalized special characterization methods [72][73][74][75][76][77]. A considerable amount of publications address the characterization of the expanding bubble technique (biaxial extension flow) and its numerical simulation [78][79][80][81]. ...
Food rheology focuses on the flow properties of individual food components, which might already exhibit a complex rheological response function, the flow of a composite food matrix, and the influence of processing on the food structure and its properties. For processed food the composition and the addition of ingredients to obtain a certain food quality and product performance requires profound rheological understanding of individual ingredients their relation to food processing, and their final perception.
... Dough can be seen as a starch particle suspension dispersed in a concentrated biopolymer solution where interactions on the molecular to the micron length scale determine the overall rheological properties, baking process, and final bread quality. In dough rheology the strong link to bread-making and the instantaneous measurement of dough quality prior the bread baking rationalized special characterization methods [72][73][74][75][76][77]. A considerable amount of publications address the characterization of the expanding bubble technique (biaxial extension flow) and its numerical simulation [78][79][80][81]. ...
Food rheology focuses on the flow properties of individual food components, which might already exhibit a complex rheological response function, the flow of a composite food matrix, and the influence of processing on the food structure and its properties. For processed food the composition and the addition of ingredients to obtain a certain food quality and product performance requires profound rheological understanding of individual ingredients their relation to food processing, and their final perception.
With the emergence of large amplitude oscillatory shear (LAOS) as a powerful tool to characterize the viscoelastic behavior of materials, it is evident that LAOS is even more appropriate to study the rheology of food materials as they show extreme nonlinear behavior at exceedingly small deformation. This chapter examines many constitutive equations using a hierarchical approach. A successful constitutive model must pass through stringent tests of describing stresses during linear oscillatory shearing flow as well as during nonlinear oscillatory shearing flow. Specifically, the Johnson-Segalman model, the Giesekus model, the K-BKZ model, the Larson model, the Phan-Thien Tanner model, and the Lodge-damage model will be explored. Recommendations for the best model/models to use in describing linear and nonlinear shearing behavior of dough are discussed.
We describe an improved damage function model for bread dough rheology. The model has relatively few parameters, all of which can easily be found from simple experiments as discussed in this paper. Small deformations in the linear region are described by a gel-like power-law memory function. Then, we consider a set of large non-reversing deformations—stress relaxation after a step of shear, steady shearing and elongation beginning from rest and biaxial stretching. With the introduction of a revised strain measure which includes a Mooney–Rivlin term, all of these motions can be well described by the damage function described previously. For reversing step strains, larger amplitude oscillatory shearing and recoil we present a discussion which shows how the damage function model can be applied in these cases. KeywordsDough rheology–Biaxial strain–Damage function–Viscoelasticity
Full-text available
A specific twin channel slit die was used to measure in-line the viscous behaviour of an extruded wheat starch. This allows to put in evidence the influences of temperature, water content and specific mechanical energy (SME). The proposed rheological law permits to satisfactorily predict the viscosity of a wheat starch for any processing condition. Original results are presented for the behaviour of cationic starches obtained by reactive extrusion.
Full-text available
A new set of experiments on a bread dough includes small-strain oscillatory behaviour, larger-strain oscillatory behaviour, simple shearing beginning from rest, uniaxial elongation beginning from rest, relaxation after sudden shear and recoil from elongation. We believe this is the most complete set of rheological data yet reported for a bread dough. Analysis of these soft-solid experiments proceeds from a Lodge-type rubberlike material with a power-law memory function. The model suggests that the response to steady shear and elongational flows may be described as a product of (strain rate) p times a function of strain; the exponent p is found to be about 0.2–0.3 from small-strain oscillatory measurements. Experiments confirm this finding. The model overestimates stresses, and in order to improve predictions, the use of a KBKZ model and a damage function model are investigated. Due to the eventual fracture of the soft-solid material, the idea of a “damage function” was adopted to produce a simple accurate, integral-type constitutive model for small-strain oscillations, simple shearing and elongation. Further analysis of reversing strains, for example, larger-strain oscillatory flows and recoil, is needed.
Full-text available
SYNOPSIS Many complex fluids exhibit power-law responses in their relaxation modulus; examples include foods, soft solids, fractal gels and other polydisperse systems. In the present study we investigate the rheological characteristics of such materials beyond the linear regime using a gluten-water gel as a prototypical system. The material functions of gluten dough under finite strains can be described by combining the linear viscoelastic response of a critical gel (Chambon and Winter 1987) with a Lodge rubber-like network to develop a frame invariant constitutive equation (Winter and Mours 1997). This generalized gel equation is a simple but accurate description of the material functions in the linear regime and also at large strains, using only two parameters. We compare the model predictions with experimental measurements in transient shear and elongational flows of gluten gels over a wide range of deformation rates. An essential feature of both the experimental data and the generalized gel model is a strain/rate separability in the system response. Further modifications to the generalized gel equation can be made by incorporating a damping function to include non-linear strain softening effects seen in more complex gels such as wheat flour doughs. From the rheological data, we find compelling evidence that indicates gluten to be a polymeric network consisting of flexible or semi-flexible chains between junction points and has a typical mesh size of approximately 20 nm.
Full-text available
Flour-water doughs made from strong and weak flours were tested using a dynamic rheometer with cone-and-plate geometry. Flour was fractionated to determine what component or components were responsible for the dynamic rheological properties (elastic modulus [G'], viscous modulus [G'], and tan δ [G'/G']) values. Doughs made from strong flour had lower tan δ values than medium or weak flours. The isolated starch or gluten fraction was combined with vital wheat gluten or commercial wheat starch. Only Larned starch gave doughs that were significantly different in dynamic theological properties from dough made with other starches. The gluten isolated from strong flours gave doughs that were significantly different from doughs made with gluten isolated from weak flours. Reconstituted flours containing starch, gluten, and various amounts of lyophilized water-solubles were tested. Addition of water solubles decreased the elastic modulus and dramatically shortened optimum mixing time of the reconstituted flour.
Full-text available
Cereal Chem. 80(3):333-338 Relaxation behavior was measured for dough, gluten and gluten protein fractions obtained from the U.K. biscuitmaking flour, Riband, and the U.K. breadmaking flour, Hereward. The relaxation spectrum, in which relaxation times ( ) are related to polymer molecular size, for dough showed a broad molecular size distribution, with two relaxation proces- ses: a major peak at short times and a second peak at times longer than 10 sec, which is thought to correspond to network structure, and which may be attributed to entanglements and physical cross-links of polymers. Relaxation spectra of glutens were similar to those for the corresponding doughs from both flours. Hereward gluten clearly showed a much more pronounced second peak in relaxation spectrum and higher relaxation modulus than Riband gluten at the same water content. In the gluten protein fractions, gliadin and acetic acid soluble glutenin only showed the first relaxation process, but gel protein clearly showed both the first and second relaxation processes. The results show that the relaxation prop- erties of dough depend on its gluten protein and that gel protein is responsible for the network structure for dough and gluten.
Full-text available
The viscoelastic properties of durum wheat flour doughs were measured using the extensigraph in uniaxial extension and the Rheometrics mechanical spectrometer in oscillatory shear. The research examined the effect of increasing density of cross-links on rubber elasticity in these systems. The stress-strain behavior of durum wheat flour dough was not well simulated by Mooney-Rivlin type nonlinear elasticity. Addition of increasing amounts of iodate made the dough show appreciable strain thickening behavior, approximating the behavior of natural rubbers The estimated apparent molecular weight between cross-links ranged from 10,500 to 16,000, much larger than that of rubbers, for which values are in the range of 500-1,000. When the Mooney-Rivlin equation was tested, it appeared to approximate only moderately well the extensional behavior of iodate-added wheat flour doughs at finite but low extensions, where the finite extensibility of chains is not a factor. It is hypothesized that the crosslinked network is highly diluted with hydrogen and hydrophobic bonds that limit the applicability of rubberlike elasticity theories. Increasing the cross-linked density using iodic acid developed matrices that moved the behavior of durum flour doughs closer to Mooney-Rivlin behavior.
An experimental technique was developed to determine the strain-rate in a tensile specimen. Then one can calculate the transient isothermal elongational viscosity. Both shear and elongational viscosities were measured to study the effect of shear and elongational fields on the flow properties. The comparison between these viscosities shows that the onset of rapid viscosity growth as crystallization solidification proceeds occurs at about the same value of time at very small deformation rates (0.0028 and 0.0047 s(-1)). The comparison of these measured viscosities as functions of shear and elongational Hencky strains also reveals that the onset of rapid viscosity growths starts at critical Hencky strain values. The behaviour of steady shear viscosity as function of temperature sweep was also explored at three different low shear rates. Finally, the influence of changing oscillatory frequencies and strain rates was also investigated.
The relaxation properties of flour-water-salt doughs prepared from four different flour types (weak medium, strong, and extra strong) at different water absorption levels from 58 to 66% with protein contents of 10.0, 10.9, 13.2, and 11.8%, respectively, were studied by imposing varying strain amplitudes of 0.1-29%. Oscillatory tests in the linear viscoelastic region of the 66% absorption strong and weak dough cannot distinguish between the two types of dough. The inability to differentiate between dough types also applied to oscillatory tests on 58% absorption weak and 66% absorption strong doughs. However, the relaxation modulus of dough (extending over time) behaved quite distinctively at high strains, where dough samples experience large deformations. At strain amplitudes of less than or equal to 0.1% (i.e., in the linear viscoelastic region), different dough types behaved similarly. Likewise, the relaxation modulus completely relaxed at sufficiently long times. The magnitude of the modulus at intermediate- and high-strain amplitudes were in the order: extra strong > strong > medium > weak, indicating a higher level of elasticity in the extra strong dough samples despite its lower protein content. The relaxation times spectrum of the weak flour, extracted from the relaxation modulus data, reveals a broad relaxation process. The stress relaxation data are very reproducible at high-strain amplitudes (approximate to 1-15% for up to 3 x 10(3) sec). This work demonstrated, for the first time, the consistency in oscillatory and relaxation measurements for dough. It also clearly showed that linear viscoelastic data, although important in the characterization of time scales in dough, are largely irrelevant in differentiating between dough types. Furthermore, without proper care, a false steady-state behavior can be obtained with standard viscometric measurements due to slippage at the dough-plate interface.
Cereal Chem. 78(4):447-452 Farinography and mixography are two commonly used procedures for evaluating dough properties. These procedures, however, cannot separate hydration and energy input during dough development, both of which are critically important for understanding fundamental rheological properties of dough. A rheometer and laser scanning confocal microscopy (LSCM) were used to study the relationship between rheological properties and microstructural characteristics of developed (by farinograph with both shear and extensional deformations), of partially developed (by rheometer with either shear or extensional deformation), and of nondeveloped (no deformation) dough samples of wheat flours. Rheological data revealed that developed dough had the highest G* (most elastic or strong), followed by doughs partially developed with extensional deformation, and then shear deformation, and finally by nondeveloped dough. The LSCM z-sectioning (scanning of different layers of the sample) and the analysis of amount of protein matrix showed that developed dough had the most protein matrix and nondeveloped dough had the least protein matrix. It also showed that the higher the G*, the greater the protein network. Moreover, the type of deformation appeared to contribute to the development of protein matrix and further increase the dough strength. In this study, a combination of shear and extensional deformations by farinograph produced the most pro- tein matrix and the strongest dough, followed by extensional defor- mation, shear deformation, and then no deformation. Wheat flour doughs are viscoelastic, that is they exhibit both solid and liquid properties. The viscoelastic properties of a dough are strongly related to the gluten proteins (Faubion and Hoseney 1990;