Full text of DOI:10.1111/j.1365-2621.2002.tb09451.x

JFS: Food Engineering and Physical Properties
Effect of Moisture Content on the
Thermomechanical Behavior of Concentrated
Waxy Cornstarch–Water Preparations—
A Comparison with Wheat Starch
A. ROLÉE, E. CHIOTELLI, AND M. LE MESTE
ABSTRACT: The rheological behavior of waxy cornstarch preparations at intermediate moisture contents (30 to 60%
w/w) was studied by dynamic mechanical thermal analysis (DMTA). Differential scanning calorimetry (DSC) and
electron spin resonance (ESR) experiments were also performed in parallel. The results were compared to those
obtained previously for wheat starch. DMTA results evidenced a critical moisture content (between 50 and 55%) for
waxy cornstarch that delimited a radical change in the rheological behavior both at room temperature and during
heating. This critical water content was around 45% for wheat starch.
Keywords: starch, amylose, moisture content, rheological behavior
Introduction
would be governed primarily by the volume fraction of swol-
HE AMOUNT AND DISTRIBUTION OF WATER WITHIN STARCH len particles. However, several authors reported results
granules is of utmost importance to their physical and showing that amylose could leach out from a temperature as
chemical properties. An understanding of the role of water low as 50 ЊC for wheat starch in excess water (Eliasson 1986;
and its interplay with heat treatments is crucial in the industri- Tester and Morrison 1990). At intermediate water contents,
al processing of starch and starch-containing materials as well the granules can only partially swell during heating because
T
as in its utilization in many technical and food applications.
of a limited available volume. Amylose and amylopectin
When dispersed in excess cold water, starch undergoes a would be only partially separated, the swollen granules in-
limited swelling (around 5% for wheat starch) (Hoseney and cluding consequently both amylose and amylopectin. In that
Rogers 1994), but there is no discernible rheological effect on case, the 2 key variables of the rheological behavior of con-
relatively dilute suspensions of starch granules. With suffi- centrated starch preparations during heating would thus be
cient heat, a starch-water system or a starch-containing food the volume occupied by the granules (closely related to
material (dough for example) undergoes a series of dramatic swelling) and their deformability (Rolée and Le Meste 1997,
changes referred to as gelatinization and pasting. These 1999); the amylose is expected to play only a minor role
changes occur within a relatively wide temperature range. (Keetels 1995).
The gelatinization temperature range is not a constant, but
Recently, Rolée and Le Meste (1999) studied the rheologi-
greatly depends upon the characteristics (water content, dis- cal behavior of wheat starch preparations at intermediate
solved solutes) of the medium (Doublier and others 1987). In moisture contents (25 to 60% w/w) by dynamic mechanical
excess water, heating results in a suspension of swollen parti- thermal analysis (DMTA). Differential scanning calorimetry
cles composed mainly of amylopectin molecules trapped in- (DSC) and electron spin resonance (ESR) experiments were
side the granules. The continuous phase is a solution of amy- also performed in parallel. The DMTA results suggested that
lose that may also contain amylopectin. Hence, the swelling would occur mainly in the range of temperature
rheological behavior of such a composite system depends corresponding to the first endothermic peak observed with
upon diverse parameters. Most of these parameters them- DSC. In the same temperature range, the ESR results sug-
selves depend upon the botanical origin of starch and the gested an increase in the viscosity of the aqueous phase at-
preparation procedure. Mainly based on microscopical ob- tributed to improved starch-water interactions and/or to
servations, Miller and others (1973) suggested that the rheo- leaching of soluble starch molecules. The increase in volume
logical behavior of starch is mainly controlled by the organi- fraction (␾) of the starch granules was assumed to be re-
zation of dissolved and entangled macromolecules leached sponsible for the increase in storage modulus during heating,
out of the granules during gelatinization. This was interpret- when ␾ reached a volume fraction close to the close-packing
ed as evidence of the major importance of the soluble mate- volume fraction ␾m. The 45% moisture content level ap-
rial. In contrast, several authors (Evans and Haisman 1979; peared to be critical for wheat starch preparations, because
Wong and Lelievre 1981) ascribed a major role to the dis- of the radical change in thermomechanical behavior around
persed phase in determining overall properties and postulat- this concentration. Indeed, lower moisture contents would
ed that the continuous phase was of minor importance. Bag- allow the wheat starch granules to reach the close-packing
ley and Christianson (1982) assumed that, due to the absence volume fraction at room temperature. Whatever the mois-
of soluble material in pastes cooked below 85 ЊC, viscosity ture content in the range 25 to 60%, the storage modulus G'
© 2002 Institute of Food Technologists
Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE 1043

Effect of moisture content on cornstarch . . .
would be closely related to the deformability of the granules the stress.
once the volume fraction had reached ␾m.
Annular pumping. About 1.5 ml of sample was poured
The present work has been conducted with similar meth- into a cylindrical cell. A piston was oscillating with small am-
ods and techniques on preparations containing a starch plitude in the center of the cell, which was glued with cy-
from another botanical origin, the waxy cornstarch. The anoacrylate glue (Amatron, Provins, France) onto the sensor
main interest of this starch was its very low ratio of amylose registering the stress. The piston was screwed into the sen-
(about 1%). A comparison with the results obtained for sor, which regulated the amplitude and the frequency of the
wheat starch should permit us to discuss the interpretations strain.
previously suggested and to estimate the possible influence
Annular shearing. This device consisted of coaxial cylin-
of amylose on the rheological behavior and structural chang- ders connected to the sensors. The gap between the 2 cylin-
es of starch and on the properties of the aqueous phase at ders was 2 mm. The sample dispersion, poured in the cell in
intermediate moisture contents during heating.
the liquid state before heating, was held in the annular space
by capillary force.
Materials and Methods
To prevent drying during the thermal treatment, starch
samples were coated with silicone grease (Rhône Poulenc,
Lyon, France) with the aid of a small brush for plan shearing.
Sample preparation
Waxy cornstarch (waxilys 200; Roquette Frères, Lestrem, For annular pumping and annular shearing, a mineral oil
France) was used for all investigations. Preparations were (Nachet, Dijon, France) was used. The strain and frequencies
made with 18 g of starch (dry matter: 16.06 g) and distilled were set at 3 ␮m and 5 Hz, respectively. First, strain sweep
water was added until moisture contents of 30, 35, 40, 45, 50, tests were performed at different temperatures in order to
55, and 60% (w/w wb) were obtained. The manual blending confirm that measurements were run in the linear range of
was continued until a homogeneous mixture was obtained viscoelasticity. Then, starch samples were heated from 30 to
then the mixture was allowed to rest for at least 1 h at room 85 ЊC (1.5 ЊC/min) during the mechanical analysis. The high-
temperature in a closed environment; the more liquid-like est temperature was 85 ЊC, beyond which starch granules
samples (moisture contents above 50%) were magnetically might be damaged (Tester and Morrison 1990). All tests were
stirred.
performed at least in triplicate. At 30% moisture content, nu-
merous tests were needed because of the difficulty to obtain
samples with the required shape.
Differential scanning calorimetry
The VA2000 software package provided by Metravib R.D.S.
allowed calculation of rheological parameters including stor-
age modulus (G'). The average standard deviation for all the
tests was calculated to be 10.2%.
Thermograms were obtained with a Perkin-Elmer DSC-7
differential scanning calorimeter, equipped with a TAC/7 DX
thermal analysis data station (Perkin-Elmer, St Quentin en
Yvelines, France), calibrated with azobenzene and indium in
the positive temperature range. Fractions of starch prepara-
tions (40 to 85 mg) were weighed and hermetically sealed in
stainless steel DSC pans. All traces were normalized to 1 mg
Electron spin resonance
Hydrated starch has no paramagnetic activity, so the spin-
of starch. The scanning temperature range and the heating probing technique was employed, in which a compound with
rate were 25 to 140 ЊC and 10 ЊC/min, respectively. An empty a nitroxyde radical, possessing a stable free electron, is add-
pan was used as an inert reference. All tests were performed ed to the system. The ESR spectra reflect the motion of the
at least in triplicate. The partial melting enthalpy was calcu- small paramagnetic probe that depends on the probe size
lated from the onset of the endotherm to 85 ЊC (every 1 ЊC) and on the solvent viscosity. The size and polarity of the
to plot the curve representing the cumulated enthalpy values probes influence their accessibility to microenvironments
as compared with temperature. The average standard devia- and their behavior in a network: a smaller probe may stay
tion was calculated to be 8.9%.
relatively mobile where, for steric reasons, a larger molecule
may have a reduced mobility. The 4-hydroxy,2,2,6,6-tetrame-
thyl-piperidine N-oxyl (TEMPOL) radical (MW 172.2g mol–1)
was purchased from Aldrich Chemicals (Strasbourg, France).
Dynamic mechanical thermal analysis.
The small amplitude oscillatory rheological measurement
was performed with a viscoanalyzer (Metravib R.D.S., Limon-
est, France), equipped with a thermocontrol unit. The tem-
perature was monitored by a thermoprobe at Ϯ 0.5 ЊC. Plan
shearing was used for the more solid-like samples (moisture
contents lower than 50%). For the more liquid-like samples, 2
different devices were used depending on the range of tem-
perature: annular pumping up to approximately 63 ЊC, then
annular shearing up to 85 ЊC. This change in device was neces-
sary because during the thermal treatment, liquid-like sam-
ples became more rigid and annular pumping was no longer
efficient. These different modes are showed in Figure 1.
Plan shearing. The main characteristic of this device was
that 2 cylindrical samples of the same size were needed.
Starch samples were 3 mm high ϫ 15 mm dia or 4 mm high
ϫ 20 mm dia. They were vertically glued with cyanoacrylate
glue (Amatron, Provins, France) on outside and inside plates.
The inside plates were connected to a sensor, which regulat-
ed the amplitude and the frequency of the strain, whereas
the outside plates were connected to a sensor that registered
Figure 1—Schematic cross sections of the different devices
used for DMTA.
1044 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 3, 2002

Effect of moisture content on cornstarch . . .
Because of its small size, this probe has the possibility to dif- cornstarch was similar to that of wheat starch (Rolée and Le
fuse into the starch granules, so it can be dispersed in the Meste 1999). However, the curves appeared somewhat less
aqueous phase inside and outside the granules. When pre- complicated because of the absence of contributions from
paring the samples, 300 ␮l of a TEMPOL aqueous solution (2 amylose, including the amylose-lipid complexes. As the wa-
mg/ml) were added to distilled water, then manually mixed ter content was lowered, several transitions became appar-
with starch (approximate final TEMPOL concentration: ent. The endotherm (G) observed at high moisture content
2.44 ϫ 10–7 mol/g of dry starch). Sealed capillary tubes con- developed a trailing shoulder (M1), which shifted to higher
taining aliquots of the samples were placed in 3 mm-dia ESR temperatures and became predominant when the water
sample quartz tubes, and ESR spectra were collected using a content was decreased. On the opposite, the first peak (G)
Bruker EMX spectrometer (Bruker, Nissembourg, France) remained located in the same temperature range and its area
with a nitrogen-flow temperature control. The operating fre- progressively decreased when the water content was de-
quency and center field were respectively at about 9.42 GHz creased. At 30% moisture content, the first endothermic
and 3357 G. The spectra were recorded at a microwave pow- peak almost disappeared. Total enthalpies of 16 J/g (db) were
er of 10 mW. Any saturation phenomenon was avoided. The estimated for the highest moisture contents studied. All of
scan rate (20.97 s for 100 G), time constant (81.92 ms), and these results are in agreement with previous studies (Dono-
modulation amplitude (0.40 G) were adjusted so that distor- van 1979; Ghiasi and others 1982; Maurice and others 1985;
tion of the spectra was avoided. For all experiments, the Gidley and Cooke 1991).
temperature was varied stepwise, every 2 ЊC, between 25 and
Although several contradictory interpretations have been
85 ЊC and the sample was stabilized for 3 min before record- suggested for the biphasic profile at intermediate moisture
ing the spectra. All ESR experiments were carried out in trip- contents, it is now widely admitted that the redistribution of
licate. The average standard deviation for all the tests was water within the sample during heating plays a key role
calculated to be 8.1%.
(Donovan 1979; Evans and Haisman 1982; Liu and others
The rotational correlation time (␶c) was determined from 1991; Liu and Lelievre 1992; Beleia and others 1996; Garcia
the relation:
␶c = 6.65 ϫ 10–10 (⌬HI+1) ϫ [ (I+1/I–1)1/2 – 1 ]
and others 1996; Rolée and Le Meste 1999). The similarities in
thermal curves for waxy corn and wheat starches indicate
that the melting endotherms may be largely accounted for
by the amylopectin portions of the starch granules. Even if
deduced from the Freed and Fraenkel (1963) theory where amylose seems to play a minor role, its contribution remains
⌬HI+1 is the width of the I+1 line, I+1, and I–1 are respectively unclear. The endothermic peak present at high temperature
the height of the lines I+1 and I–1 (Figure 2).
for wheat starch corresponding to the melting of amylose-
The conventional ESR method was used, allowing mobility lipid complexes was, of course, not observable for waxy
measurements in the range 10–11 Ͻ ␶c Ͻ 10–7 s. The rotation- cornstarch. As the gelatinization endotherm would represent
al diffusion coefficient (Drot) was evaluated from the rota- essentially the difference between the endothermic energy,
tional correlation time (␶c) (Nordio 1976):
associated with melting of crystallites, granule swelling and
denaturation, and the exothermic energy associated with hy-
dration of starch and formation of the amylose-lipid com-
plexes (Kugimiya and others 1980), the absence of amylose in
waxy starches may generate a higher endothermic energy.
The results obtained with the waxy cornstarch seem to con-
firm this hypothesis since our values of enthalpies of gelatini-
Drot = 1/(6␶c)
Results and Discussion
Thermal disordering
The DSC curves for waxy cornstarch were strongly affect- zation, as those currently found in literature, are high com-
ed by the mass fraction of water (Figure 3), over the 30 to pared to most of the non-waxy starches. In the same way,
60% water content range.
potato and cassava starches contain no lipids and, hence, no
The water dependence of the thermal behavior of waxy amylose-lipid complexes can form; as for waxy cornstarch,
Figure 3—Differential scanning calorimetry thermograms
of waxy starch preparations at intermediate moisture con-
Figure 2—Typical ESR spectrum of a nitroxyde free radi- tents (30 to 60% w/w, wb). G and M1 indicate the 2 suc-
cal (TEMPOL).
cessive endothermic peaks.
Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE 1045

Effect of moisture content on cornstarch . . .
enthalpies of gelatinization currently found in literature for 30 to 50% moisture content (5 ϫ 105 – 1.2 ϫ 106 Pa).
both starches are relatively high, tending to reinforce the
previous hypothesis.
For samples with 55 and 60% water, G' increased very
strongly from 55 to 60 ЊC and reached a maximum value at
Though similarities existed in the DSC traces for waxy around 70 ЊC, before decreasing slightly up to 85 ЊC (Figure
corn and wheat starches, a clear difference could also be ob- 5). In such dispersions with quite high moisture contents, an
served at the highest moisture contents studied. Indeed, for increase in volume fraction of starch granules has been
wheat starch, a main peak and a shoulder were observed, shown to be responsible for this G' increase during heating
but for waxy cornstarch, 2 partly superimposed peaks were (Bagley and Christianson 1982; Doublier and others 1987;
visible. This difference is highlighted in the partial melting Rolée and Le Meste 1997). The dominant effect of the pro-
enthalpy plot (Figure 4a and 4b).
gressive swelling of starch granules, subsequent to the crys-
For wheat starch, at the highest moisture contents stud- tallite melting in this range of temperature, led to a maxi-
ied, the melting enthalpy grew up steeply with temperature mum occupancy of the available space by the swollen
at the beginning of melting (in an approximate range of granules. Due to the resulting increase in the granule-granule
10 ЊC), then an inflection point was observed around 65 ЊC interactions, the system formed a «gel» with a crystal-like
and the rate of enthalpy increase slowed down. For waxy package of the granules. This state has also been assimilated
cornstarch, no inflection point was observed, the enthalpy to a transient network formed by the granules, which could
increased almost linearly up to 85 ЊC. This difference could eventually disrupt during further heating (Champenois and
be due, at least partly, to the exothermic formation of amy- others 1998). Wheat starch dispersions in the 50 to 60% mois-
lose-lipid complexes, thus inducing above 65 ЊC an underes- ture content range were shown to behave in a similar way
timation of the enthalpy of melting for wheat starch. Normal (Rolée and Le Meste 1999). However, the G' increase was
cornstarch, with approximately the same amylopectin/amy- shifted towards higher temperatures for the waxy corn-
lose ratio as wheat starch, was observed to exhibit a similar 2 starch. This result is in agreement with DSC results showing
step melting process with an apparent lower rate of melting the higher thermal stability of the crystals of waxy corn-
above 75 ЊC (Rolée and Le Meste 1997).
starch. Moreover, we noticed that waxy cornstarch disper-
The disorganization of crystallites seems to be facilitated sions reached their maximum G' value at about 70 ЊC after a
in wheat starch compared to cornstarch (waxy and normal). continuous increase, whereas for wheat starch dispersions a
Indeed, melting started at a lower temperature for wheat very steep increase in modulus was first observed up to
starch. This difference between wheat and corn could not be about 55 ЊC, followed by a more reduced increase to approx-
attributed to amylose, as the onset temperature for the melt- imately 65 ЊC (Figure 5 and 6).
ing of normal cornstarch crystallites was similar to that of
At least 2 events could be responsible for this difference.
the waxy cornstarch (Rolée and Le Meste 1997). This appar- First, this could be due to the polydispersed granules (2 pop-
ent higher thermal stability of the crystallites in corn and ulations) of native wheat starch, whereas waxy cornstarch
waxy cornstarches might reflect a different organization of has only one population of granules, as described in the liter-
the molecules, either within the crystalline domains or be- ature (Galliard and Bowler; 1987 Rolée and Le Meste 1997).
tween the crystalline and the amorphous domains.
For wheat starch, the straight increase in G' would reflect the
rapid packing of the large population of granules during
heating. Some available spaces between these large packed
granules could allow the smallest granules to swell further
and reinforce gently the rigidity of the system. Secondly,
when the granules get closed-packed, contact between the
granules might occur either directly, which might be the case
for waxy cornstarch, or, as suggested by Eliasson (1986) for
wheat starch, the contact might also involve amylose that
Rheological behavior
Storage modulus (G') as compared with temperature was
plotted for all moisture contents studied (Figure 5).
As for wheat starch (Figure 6), the initial storage modulus
(G'i = G' at 30 ЊC) increased as moisture content decreased.
For dispersions with 55 and 60% water content, G'i was low
(10 to 100 Pa) whereas G'i was very high for dispersions with
Figure 4a—Partial melting enthalpy as a function of tem- Figure 4b—Partial melting enthalpy as a function of tem-
perature for waxy cornstarch preparations (moisture con- perature for wheat starch preparations (moisture contents
tents given on graph expressed on a wb).
given on graph expressed on a wb) (Rolée and Le Meste
1999).
1046 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 3, 2002

Effect of moisture content on cornstarch . . .
has leached out, at least partly, from the granules. A thin lay- Stokes-Einstein equation:
er of leached amylose wrapping the granules could generate
indirect granule-granule contacts, and thus explain the gentle
Drot = kT/8␲␩r3C
G' increase preceding the maximum G' value.
Once granules are in close contact, G' would then be where k is the Boltzmann constant, T the absolute tempera-
sensitive to the intrinsic softness, that is, deformability of ture, ␩ the viscosity of the medium, r the radius of the diffus-
the granules, closely related to the swollen state and the re- ing molecule, and C the coupling parameter representing the
maining ordered zones. The values of the modulus for amount of solvent that is dragged with the molecule when it
closed packed suspensions of both waxy corn and wheat moves (Kowert and Kivelson 1976). Rolée and Le Meste
starch suggest that the granular material is in the rubbery (1999) emphasized that, in heterogeneous systems such as
state. Heating above the temperature where the maximum concentrated starch-water dispersions, the relevant parame-
in G' is observed provides energy to break down the residu- ter should be the viscosity of the diffusion medium (that is,
al crystalline structure of starch, causing G' to drop down aqueous phase). This viscosity is expected to be sensitive to
(Figure 5b) (Lii and others 1996). As for wheat starch, DSC the interactions between water and the starch molecules and
results on waxy cornstarch (Figure 4a and Figure 4b) thus to starch structural disorganization such as the disrup-
showed evidence of melting/dissociation of ordered zones tion of low energy starch-starch interactions (melting) and
in this range of temperature.
subsequent increase in starch-water hydrogen bonding.
At room temperature, like over the entire temperature
The dispersions having the highest G'i (moisture content
ranging from 30 to 50%) did not show pronounced viscoelas- range studied, 3 line spectra similar to those of probes in wa-
tic changes up to 65 to 70 ЊC (Figure 5). However, further heat- ter were obtained (not shown). This behavior is generally in-
ing up to 85 ЊC induced a G' decrease. For such moisture con- terpreted in terms of isotropic fast motion. The calculated
tents, the granules were already interacting at room rotational diffusion coefficients (Figure 7) were found to be
temperature. The high initial modulus of the dispersions could lower than the values for the probe in water. The observed
be explained by the tight packing of compressed granules. Not slower motion in the presence of starch suggests that the
enough space was available for the granules to further swell probe experienced an environment of higher viscosity. A
and consequently no G' increase was observed, unlike the previous work (Rolée and Le Meste 1999) suggested that this
higher moisture contents studied. These concentrated disper- higher viscosity would be due to a reduced mobility of water
sions composed of packed granules were then sensitive to the molecules involved in the starch-water interactions. Indeed,
intrinsic softness of the granules during the whole heating. The starch molecules become partly hydrated as soon as the
G' decrease observed from 65 to 70 ЊC to 85 ЊC was then ex- starch-water dispersion is prepared, and the granules swell
plained by the same reasons as previously accounted for the reversibly to some extent. This would decrease the mobility
dispersions with 55 and 60% water contents, that is, crystal of probe molecules. It can be noticed that only one popula-
melting and subsequent decrease in the granule rigidity.
tion of probes was observed, that is, probes inside and out-
side the granules could not be distinguished. Studies by solid
state 2H and 1H NMR on waxy cornstarch have already
shown that water was mobile in the starch matrix, despite
the extremely high rigidity of starch molecules in the solid
and semi-solid state (Li and others 1998).
In comparison to wheat starch and for similar concentra-
tions (Figure 8), waxy cornstarch dispersions showed lower
Drot values at room temperature, suggesting starch-water in-
teractions were favored. Compared to wheat starch, only
slight changes occurred during the thermal treatment of
waxy cornstarch. Up to 60 to 65 ЊC, Drot increased slightly for
Motional behavior of probes (TEMPOL) dispersed in
the aqueous phase studied by ESR
ESR proved to be an appropriate technique to detect
changes in the properties of the water phase within starch
dispersions (Biliaderis and Vaughan 1987; Rolée and Le Meste
1999). In this case, ESR is not a direct method, but is used to
measure the rotational mobility of a water-soluble probe
(TEMPOL) dispersed in the aqueous medium of the suspen-
sion. The rotational diffusivity of a probe dispersed in a ho-
mogeneous liquid medium can be described by the modified
(a)
(b)
Figure 5—(a) Storage modulus changes for waxy starch dispersions at intermediate moisture contents (30 to 60% w/
w, wb), as a function of temperature during heating. (b) Expansion of curves in Figure 5a.
Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE 1047

Effect of moisture content on cornstarch . . .
dispersions having 30 and 40% moisture contents, was stable
for 50% moisture content, and slightly decreased for 60%
␾
m = 1-0.47(d/D)0.2
moisture content. From 60 to 65 to 85 ЊC, Drot increased where d and D are respectively the lower and the upper lim-
slightly for the lowest water contents (30 and 40%) and more its of the size distribution of the particles. A previous study
sharply for 50 and 60% moisture contents.
using laser-light diffraction (Rolée and Le Meste 1997)
During heating wheat starch dispersions with 50 and 60% showed that wheat starch had a larger size distribution than
moisture, a clear decrease in Drot was observed from 47 to waxy cornstarch. Using this relation and the published values
50 ЊC to a minimum value at 57 to 60 ЊC (Figure 8). This de- of granule size distribution, ␾ could be estimated to be 0.73
m
crease occurred in the same temperature range as does the for wheat starch and 0.66 for waxy cornstarch. The ␾ value
m
G' increase and the onset of the endothermic events. There- calculated for wheat starch must be even higher because of
fore, the Drot decrease appeared to be related to the starch- the polydispersity of the wheat granules. Therefore, at room
water interactions that could be improved by amylopectin temperature, ␾ would be reached at a lower moisture con-
m
melting and granule swelling. However, this hypothesis tent (higher granule concentration) for wheat than for waxy
seems still open to question. Indeed, if we carefully pay at- cornstarch.
tention to the onset of the different phenomena, it seems
Below ␾m, the modulus of the suspensions of both type of
that the decrease in probe mobility occurs slightly before the starch is similar and controlled by the continuous phase.
onset of starch melting (Figure 4b). Furthermore, waxy corn- Above ␾ the modulus of the suspensions is controlled by
m
starch dispersions with 50 and 60% moisture content, which the rigidity of the granules acting as a continuous phase. The
also exhibit a similar melting process, did not show such a main difference in the rheological behavior of the suspen-
decrease in Drot
.
sions obtained with the 2 starches can be at least partly at-
tributed to the hydration properties (or granule concentra-
tion), and swelling capacity, allowing the granule volume
General discussion
fractions ␾ to get close to their respective ␾
m.
Behavior at room temperature
Thermomechanical behavior
Rolée and Le Meste (1999) showed that for wheat starch
dispersions, the 45% water content seemed critical, since
there was a radical change in the G' modulus of the starch
preparation and of the thermomechanical behavior around
this concentration. Beyond this critical volume fraction,
granules are expected to be connected on a large scale,
hence they act as a continuous phase. For waxy cornstarch,
this radical change in modulus and thermomechanical be-
havior occurs at a higher moisture content (thus less con-
centrated suspensions), around 50 and 55% moisture. This
observation is in agreement with the ESR results that sug-
gested a higher affinity of waxy cornstarch for water at room
temperature, in comparison to wheat starch. Thus, an im-
proved swelling capacity at room temperature is expected
for waxy cornstarch. The different values of the critical water
content could also be partly explained by differences in the
granules size distribution. Indeed, ␾m could be determined
by a simple equation (Alberola and Mele 1996):
For the highest moisture contents studied, the increase in
rigidity observed by DMTA occurred at a lower temperature
for wheat starch than for waxy cornstarch (Figure 5 and 6).
As previously, the increase was attributed to the fact that the
granule volume fraction gets close to ␾m, as a consequence
of the heat-induced swelling. As ␾ would be higher for
m
wheat starch granules than for the other starch, the DMTA
results suggest that granule swelling upon heating started at a
lower temperature for wheat starch. DSC results, showing
that wheat starch crystallites start melting at a lower temper-
ature, are in agreement with this proposition. These struc-
tural changes have been described as being accompanied by
a redistribution of water within the material. Indeed the
amorphous domains are expected to exhibit a higher affinity
for water than the crystalline zones. NMR spectroscopic
studies have shown that water molecules in starch are highly
mobile even at low water contents, thus they can diffuse
Figure 7—Rotational diffusion coefficient of spin probe
(TEMPOL) dispersed in the aqueous phase of waxy corn-
starch dispersions at intermediate moisture contents (w/
w, wb), as a function of temperature.
Figure 6—Storage modulus changes for wheat starch dis-
persions at intermediate moisture contents (25 to 60% w/
w, wb), as a function of temperature during heating (Rolée
and Le Meste 1999).
1048 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 3, 2002

Effect of moisture content on cornstarch . . .
rather rapidly into the starch samples. As a result, water can a structural reorganization. Hydrothermal treatments, such
be heterogeneous with respect to mobility, depending on its as storage performed at a certain moisture level during a
partitioning among domains, such as in a gel phase, in dis- certain period of time at a temperature above the glass tran-
solved starch (solution domains), in crystalline domains, or sition temperature but below the gelatinization temperature,
in the granular amorphous regions. Changing the initial wa- are known to cause physical modifications, such as the de-
ter content and temperature could thus greatly influence the velopment of new crystals in the amorphous regions, or
balance or the distribution of the water within different crystallite growth or perfection of already-existing crystals.
phases or domains (Li and others 1996). The decrease in the Such hydrothermal treatments are called «annealing» when
rotational diffusivity of the water soluble probes dispersed in storage is in excess of water, while the term «heat moisture
the high moisture wheat starch dispersions in the same tem- treatment» is used when low moisture levels are applied.
perature range as the onset of melting, has been attributed Storage in the gelatinization temperature range can also lead
to the improved starch-water interactions associated with to structural reorganization. Recrystallization or polymor-
melting (Rolée and Le Meste 1999). However, the ESR results phic transition processes, which entails melting followed by
obtained for waxy cornstarch seem to exclude the latter hy- renewed crystallization, have been reported (Yost and
pothesis. Indeed, in that case starch melting was not associ- Hoseney 1986; Zobel and others 1988). However, this phe-
ated to a decrease in Drot
.
nomenon is more noticeable at higher temperatures, per-
Another hypothesis relies on an increase in starch-water haps because at lower temperatures, the presence of the
interactions not directly related to melting and swelling of greater number of crystallites tends to constrain the amor-
starch granules. This could well be the reason why ESR results phous regions into conformations that are incompatible with
for wheat and waxy cornstarches are different. A high inher- crystal formation (Liu and others 1991). Although no storage
ent amylose content has been reported to enhance the rigidity was performed, the slow heating of the samples during our
of starch granular structures (Lii and others 1996). It can, thus, experiments could have caused crystal improvement and/or
be suggested that amylose could prevent the penetration of more orderly arrangements between chains segments in
water molecules into some sites of starch. The presence of amorphous regions. This strengthening of the biopolymer
amylose in wheat starch would then explain the higher Drot at matrix could generate a decrease in Drot. However, it must
room temperature and during heating up to 50 ЊC. Eliasson be noticed that waxy cornstarch did not show the Drot de-
(1986) and Tester and Morrison (1990) reported results show- crease around the onset of melting. This could be due to the
ing that amylose could leach out from temperatures as low as presence of apparently more stable crystals in native waxy
50 ЊC for wheat starch. When the amylose leaching starts it cornstarch as suggested by DSC results.
may open up more sites for hydration in domains that water
Conclusions
MTA RESULTS EVIDENCED A CRITICAL MOISTURE CONTENT
(approximately 45%) for wheat starch that delimited a
molecules were not able to reach before. New starch-water
interactions could then form, that can be evidenced by the
clear decrease in probe mobility from 50 ЊC to a minimum
value at approximately 60 ЊC for wheat starch dispersions with
50 and 60% moisture contents (Figure 8). Lower moisture
contents should not allow amylose solubilization and leaching
out and hence no further Drot decrease was observed. Since
waxy maize starch contains only a little proportion of amylose
(about 1%), water could absorb more freely onto the starch
molecules, whatever the temperature, inducing a lower probe
mobility, than for wheat starch, with no clear decrease during
heating as no amylose can leach out.
D
radical change in the rheological behavior. This critical water
content was between 50 and 55% for waxy cornstarch. These
critical water contents were interpreted as corresponding to
the concentrations where the volume fractions occupied by
the partly hydrated starch granules get close to their respec-
tive close-packing volume fraction ␾m. For less concentrated
dispersions, this critical volume fraction is reached upon
heating, through the swelling subsequent to the melting pro-
cess.
Whatever the moisture content from 30 to 60%, the stor-
age modulus G' appears closely related to the rigidity of the
The decrease in probe mobility could also be attributed to
granules once the volume fraction had reached ␾ and to
m
the viscosity of the continuous aqueous phase when ␾ Ͻ ␾m.
Differences in the rheological behaviors of wheat and waxy
cornstarches have been partly attributed to the polydisper-
sion and size distribution of the granules, but also to differ-
ences in the accessibility of water to starch hydration sites
attributed to the presence or not of amylose.
DSC results suggested that amylose could affect the enthalpy
of melting of ordered regions in wheat starch. Low values of
enthalpy found for the melting of wheat starch, compared with
waxy cornstarch, could be explained, at least partly, by an un-
derestimation of the enthalpy of melting, due to the concomi-
tant exothermic formation of amylose-lipid complexes.
At the highest moisture contents studied, leaching out of
amylose occur during the heating process of non-waxy
starches. The decrease in probe mobility observed in the
range 50 to 60 ЊC for wheat starch was attributed to either a
structural reorganization, or more probably to starch-water
interactions which should form when amylose starts to leach
out, opening up more sites of hydration in domains where
water molecules were not able to reach before. In preventing
Figure 8—Rotational diffusion coefficient of spin probe
(TEMPOL) dispersed in the aqueous phase of wheat starch
dispersions at intermediate moisture contents (w/w, wb),
as a function of temperature (Rolée and Le Meste 1999).
Vol. 67, Nr. 3, 2002—JOURNAL OF FOOD SCIENCE 1049

Effect of moisture content on cornstarch . . .
Kowert B, Kivelson D. 1976. ESR linewidths in solution. VIII.Two component dia-
magnetic solvents. J Chem Phys 64(12):5206-5217.
the penetration of water into some regions of the granules,
up to 50 ЊC, the presence of amylose could well explain why
wheat starch reaches the close-packing volume fraction, at
room temperature, at a lower moisture content (that is, criti-
cal moisture content) than waxy cornstarch.
Kugimiya M, Donovan JW, Wong RY. 1980. Phase transitions of amylose-lipid
complexes in starches: A calorimetric study. Starch/Stärke 32(8):265-270.
Li S, Tang J, Chinachoti P. 1996. Thermodynamics of starch-water systems: an
analysis from solution-gels model on water sorption isotherms. J Polym Sci:
Part B: Polymers Physics 34:2579-2589.
Li S, Dickinson LC, Chinachoti P. 1998. Mobility of «unfreezable» and «freezable»
water in waxy cornstarch by 2H and 1H NMR. J Agric Food Chem. 46:62-71.
Lii CY, Tsai ML, Tseng KH. 1996. Effect of amylose content on the rheological
property of rice starch. Cereal Chem 73(4):415-420.
Liu H, Lelievre J, Ayoung-Chee W. 1991. A study of starch gelatinization using
differential scanning calorimetry, X-ray, and birefringence measurements.
Carbohydr Res 210:79-87.
Liu H, Lelievre J. 1992. Differential scanning calorimetric and rheological study
of the gelatinization of starch granules embedded in a gel matrix. Cereal Chem
69(6):597-599.
References
Alberola ND, Mele P. 1996. Viscoelasticity of polymers filled by rigid or soft par-
ticles: theory and experiment. Polym Compos 17(5):751-759.
Bagley EB, Christianson DD. 1982. Swelling capacity of starch and its relation-
ship to suspension viscosity-effect of cooking time, temperature and concen-
tration. J Texture Stud 13:115-126.
Beleia A, Miller RA, Hoseney RC. 1996. Starch gelatinization in sugar solutions.
Starch/St⌬rke 48(7/8):259-262.
Maurice TJ, Slade L, Sirett RR, Page CM. 1985. Polysaccharide-water interac-
tions-Thermal behavior of rice starch. In: Simatos D, Multon JL, editors. Prop-
erties of water in foods. Dordrecht: Martinus Nijhoff. P 211-227.
Miller BS, Derby RI, Trimbo HB. 1973. A pictorial explanation for the increase in
viscosity of a heated wheat starch-water suspension. Cereal Chem 50(3):271-
280.
Nordio PL. 1976. General magnetic resonance theory. In: Berliner LJ, editor.
Spin labeling. Theory and applications. New York: Academic Press. P 5-52.
Rolée A, Le Meste M. 1997. Thermomechanical behavior of concentrated starch-
water preparations. Cereal Chem 74(5):581-588.
Rolée A, Le Meste M. 1999. Effect of moisture content on the thermomechanical
behavior of concentrated wheat starch-water preparations. Cereal Chem
76(3):452-458.
Tester RF, Morrison WR. 1990. Swelling and gelatinization of cereal starches. I.
Effects of amylopectin, amylose, and lipids. Cereal Chem 67(6):551-557.
Wong RBK, Lelievre J. 1981. Viscoelastic behaviour of wheat starch pastes. Rheol
Acta 20:299-307.
Biliaderis CG, Vaughan DJ. 1987. Electron spin resonance studies of starch-wa-
ter-probe interactions. Carbohydr Polym 7:51-70.
Champenois Y, Rao MA, Walker L. 1998. Influence of gluten on the viscoelastic
properties of starch pastes and gels. J Sci Food Agric 78:119-126.
Donovan JW. 1979. Phase transitions of the starch-water system. Biopolymers
18:263-275.
Doublier JL, Llamas G, Le Meur M. 1987. A rheological investigation of cereal
starch pastes and gels. Effect of pasting procedures. Carbohydr Polym 7:251-
275.
Eliasson AC. 1986. Viscoelastic behaviour during the gelatinization of starch. J
Texture Stud 17:253-265.
Evans ID, Haisman DR. 1979. Rheology of gelatinized starch suspensions. J Tex-
ture Stud 10:347-370.
Evans ID, Haisman DR. 1982. The effect of solutes on the gelatinization temper-
ature range of potato starch. Starch/Stärke 34(7):224-231.
Freed JH, Fraenkel J. 1963. Theory of line width in electron spin resonance
spectra. J Chem Phys 39: 326-348.
Gaillard T, Bowler P. 1987. Morphology and composition of starch. In: Gaillard T,
editor. Starch: properties and potential. New York: John Wiley & Sons. P 55-78.
Garcia V, Colonna P, Lourdin D, Buleon A, Bizot H, Ollivon M. 1996. Thermal
transitions of cassava starch at intermediate water contents. J Therm Anal
47:1213-1228.
Yost DA, Hoseney RC. 1986. Annealing and glass transition of starch. Starch/
Stärke 38(9):289-292.
Zobel HF, Young SN, Rocca LA. 1988. Starch gelatinization : an X-ray diffraction
study. Cereal Chem. 65(6):443-446.
MS 20010039 Submitted 1/26/01, Accepted 4/6/01, Received 5/1/01
Ghiasi K, Hoseney RC, Varriano-Marston E. 1982. Gelatinization of wheat starch.
III.Comparison by differential scanning calorimetry and light microscopy.
IV.Amylograph viscosity. Cereal Chem. 59(4):258-262.
Gidley MJ, Cooke D. 1991. Aspects of molecular organization and ultrastructure
in starch granules. Biochem Soc Trans 19:551-555.
Hoseney RC, Rogers DE. 1994. Mechanism of sugar functionality in cookies. In:
Faridi H, editor. The science of cookie and cracker production. New York-Lon-
don: Chapman & Hall. P 203-225.
Keetels CJAM. 1995. Retrogradation of concentrated starch systems; mechanism
and consequences for product properties. [PhD thesis] Wageningen, The Neth-
erlands: Wageningen Agricultural Univ. P 31-53. Available from Wageningen
Agricultural Univ.
The study was conducted with financial support from the Commission of the European
Communities, Agriculture and Fisheries (FAIR) specific RTD program, CT96-1085: Enhance-
ment of Quality of Food and Related Systems by Control of Molecular Mobility. It does not
necessarily reflect its views and in no way anticipates the Commission’s future policy in this
area. The authors would like to thank the Greek Republic Scholarship Foundation (IKY) for
financing Eleni Chiotelli.
The authors are with the Laboratoire d’Ingénierie Moléculaire et Sensorielle
de l’Aliment, Ecole Nationale Supérieure de Biologie Appliquée à la Nutri-
tion et à l’Alimentation, 1 Esplanade Erasme, 21000 Dijon, France. Direct
inquiries to author Le Meste (E-mail: mlemeste@u-bourgogne.fr).
1050 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 3, 2002