Full text of DOI:10.1046/j.1365-2621.2000.00360.x

International Journal of Food Science and Technology 2000, 35, 141–154
141
Extrusion of cassava starch with either variations in
ascorbic acid concentration or pH
Pensiri Sriburi & Sandra E. Hill*
Division of Food Sciences, School of Biological Sciences, University of Nottingham, Sutton Bonington Campus,
Loughborough LE12 5RD, UK
(Received 12 June 1999; Accepted in revised form 6 September 1999)
Summary
Thermal mechanical extrusion is used to make cassava products that are later expanded
by frying. During extrusion the native starch granules are broken down and starch poly-
mers degraded. This continuum of change to the starch has been termed starch conver-
sion. Inclusion of ascorbic acid promoted starch conversion as measured by a range of
techniques including measurement of water solubility, water absorption, hot paste and
alkaline viscosities. It was found that the Rapid Visco Analyser (RVA, Newport Scientific,
NSW, Australia) was particularly sensitive in the monitoring of the changes owing to
extrusion. Expansion of the extruded product was increased at the higher starch conver-
sions caused by ascorbic acid addition. Changes in starch conversion were greater in the
presence of ascorbic acid compared with general acid treatment at equivalent pHs.
Although the general acidification did encourage more starch conversion there was no
measured change in the expansion properties.
Keywords Expansion, pH, RVA, water absorption index, water solubility index.
1993) and the linear expansion often correlates
Introduction
with a measure of crispness. Vasanti Nair et al.
(1996) have proposed that the expansion of tapi-
oca chips takes place as a result of rapid water
vaporization inside the amorphous starch matrix
at the initial stages of frying, but reaches a
plateau before maximum moisture loss. During
expansion the starch matrix must accommodate
the water vapour within the bubble. Hence starch
structure and molecular weight must be of
importance as it will govern the viscosity of the
matrix.
Cooking alters native starch. The granular
structure is disrupted and the amylose and amy-
lopectin chains can be broken. The term starch
conversion has been used to describe the contin-
uum of changes that occur from the native starch
granule to depolymerized starch molecules
(Mitchell et al., 1997). Important parameters that
govern starch conversion are water levels, tem-
perature, shear and the chemical environment.
Starch conversion during extrusion has been well
Cassava is one of the most important metabolic
sources of energy for millions of people in the
tropics and is also an important raw material for
other industrial uses (Rickard et al., 1991;
Blanshard, 1994). There are many expanded
snack foods commonly eaten in ASEAN coun-
tries that are based on cassava starch (Yu &
Low, 1992). These products are often formed as
pellets and are then expanded into a porous low-
density product in an oven or fried in oil before
consumption. The pellets can also be formed by
extrusion cooking (Matz, 1993). Cassava starch
has been reported to have a particularly high
degree of expansion (Yu, 1991) and this is very
important to the eating quality of snack foods.
For example, crispness is said to be the most
important sensory attribute in fish crackers (Yu,
*Correspondent: Fax: +44 115 9516142;
e-mail: sandra.hill@nottingham.ac.uk
© 2000 Blackwell Science Ltd

142
Extrusion of cassava starch P. Sriburi & S. E. Hill
documented but is typically considered only in
terms of thermal and mechanical changes
(Colonna & Mercier, 1983; Colonna et al., 1984;
Colonna et al., 1989; Kokini et al., 1992;
Myllymäki et al., 1997). It is sometimes consid-
ered that most of the depolymerisation of the
starch that occurs is owing to mechanical action
(Diosady et al., 1985).
Work at the University of Nottingham has
indicated that starches can be also be depoly-
merised by the action of redox agents and cassa-
va starch is particularly liable to undergo
oxidative reductive depolymerisation compared
with other tuber and cereal starches (Paterson
et al., 1996). The agents sulfite and ascorbic acid
have been shown to cause changes in viscosity of
a range of starches when heated in excess water
(Vallès-Pàmies et al., 1997). We have now shown
that the inclusion of ascorbic acid with cassava,
when it is cooked with limited water at tempera-
tures above of 100 °C, changes the amount of
starch conversion which occurs (Sriburi et al.,
1999a).
basis) was supplied by National Starch and
Chemical Company (Manchester, UK). The speci-
fications from the supplier listed that the starch
contained 0.5% protein, 0.2% fat and 97% carbo-
hydrate on a dry basis.
L-Ascorbic acid, sodium hydroxide, metaphos-
phoric acid and 2,6-dichlorophenolindophenol
(sodium salt) were purchased from Sigma
Chemical Co. (Poole, UK). Potassium hydroxide,
hydrochloric acid and lithium chloride were
obtained from BDH Laboratory Supplies (Poole,
UK). Glacial acetic acid and sodium bicarbonate
were purchased from Fisher Scientific UK
(Loughborough, UK). Ethyl alcohol, potassium
hydrogen phthalate and the buffer salts (di-sodium
hydrogen
phosphate:sodium-di-orthohydrogen
phosphate) were purchased from Fisons
Equipment (Loughborough, UK). All reagents
were of analytical grade. All measurements were
carried out in a pH 6.5 phosphate buffer unless
otherwise stated. Pure vegetable oil (Asda, UK)
was used for frying of cassava extrudates.
Ascorbic acid is often used in wheat products,
especially bread, as it is said to have a role in
increasing loaf volume (Tsen, 1965; Thewlis,
Extrusion
We manufactured the samples using a Clextral
BC-21 intermeshing twin screw extruder (Clextral,
Firminy, France). Cassava starch was extruded
under the following conditions of temperatures in
the different barrel zones (40, 90, 110 and 75 °C).
A screw speed of 200 r.p.m. and screw torque
between 10 and 11 N.m was used. Cassava starch
was fed into the extruder at 5 kg.h . The extrud-
er feed solution was normally 0.1 M phosphate
buffer (pH 6.5) and was added at 3 kg.h , in order
to obtain a moisture content inside the extruder of
around 37.5% (wet basis). Ascorbic acid was dis-
solved in the buffer immediately before extrusion.
A range of extrudates was obtained with ascorbic
acid level of 0, 0.1, 0.3, 0.5, 0.7 and 1.0% (starch
basis). The specific mechanical energy (SME) of all
1974;
Chamberlain, 1981; Lillard et al., 1982; Pachla &
Reynolds, 1985; Grosch, 1986; Halliwall
Elkassabany
&
Hoseney,
1980;
&
Gutteridge, 1995). The action of the ascorbic and
its oxidised form of dehydroascorbic acid
(DHAA) is thought to be on the gluten in the
flour promoting changes in the disulfide bonds
(Fitchett & Frazier, 1986; Grosch & Wieser,
1999). However, the initial work of cassava starch
could indicate that the starches may also be
altered in some circumstances by the presence of
ascorbic acid.
The objectives of the work reported in this
paper are to demonstrate if ascorbic acid pro-
motes depolymerisation of cassava starch when
extruded and if additional depolymerisation
occurs, does it affect changes in the expansion of
the product.
ꢀ1
ꢀ1
ꢀ1
experiments was around 55.5 w.h.kg , as calculat-
ed below (Fan et al., 1996):
ꢀ1
SME(w.h.kg ) ꢁ
ꢀ1
Materials and methods
screw torque (N.m)ꢂscrew speed (s )ꢂno. of scre
ws
Materials
ꢀ1
mass flow rate (kg.h )
Commercial cassava starch (Target Brand, product
of Thailand) with a moisture content of 13% (dry
A ribbon type die with a 1 mm ꢂ 30 mm dis-
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

Extrusion of cassava starch P. Sriburi & S. E. Hill
143
charge slit was used to produce nonexpanded
strips. We did two sets of experiments with ascor-
bic acid with a one month gap between runs.
Extrusion for the pH control experiment was car-
ried out on a third occasion using the same
method.
and are often important in predicting how the
materials may behave if further processed. These
methods are simple, easy to apply and the results
are consistent when precise methodology is
applied. It must be noted that the WAI and WSI
measurements give results which depend specifical-
ly on the amount of sample being used in the
assay as well as the particle size of the milled
extrudates. The milling procedure might also alter
the starch. It is therefore important to consider
these features in order to provide reproducible
data and when comparing these values. An
increase in WSI denotes increasing starch conver-
sion while WAI shows a maximum level at a cer-
tain amount of starch conversion. This maximum
is when the starch granules are sufficiently dam-
aged for these to imbibe water without disintegra-
tion (Mitchell et al., 1997).
Extruder feed solutions to regulate pH were as
follows (Shugar et al., 1981):
ꢀ
ꢀ
ꢀ
pH 1 ꢁ 0.1 M HCl;
pH 2 ꢁ 0.01 M HCl;
pH 3 ꢁ 0.1 M potassium hydrogen phthalate
(KHP) (750 ml) ꢃ 0.1 M HCl (334.5 mL);
ꢀ
ꢀ
ꢀ
pH 4 ꢁ 0.1 M KHP (1000 mL) ꢃ 0.1 M HCl
(20 mL);
pH 5 ꢁ 0.1 M KHP (750 mL) ꢃ 0.1 M NaOH
(339 mL);
pH 6.5 (control) ꢁ 0.1 M sodium phosphate
Typically the method used was to mix 0.3 g of
dried powdered sample with 10 mL of distilled
water and allow to hydrate and sediment
overnight. The suspensions were then subjected to
centrifugation (Multex, MSE Centrifuge, UK) at
3000 r.p.m. (2045g) for 10 min and WAI expressed
as the weight of pellet per gram of sample. The
supernatant was evaporated to dryness at 105 °C
until constant weight. WSI was the weight of dry
solids in the supernatant expressed as a percentage
of the original weight of sample (Sriburi et al.,
1999a).
buffer.
Moisture content equilibration
On extrusion the strips of cassava product were
cut into pieces approximately 2.5 cm square,
cooled at room temperature for 12 h and dried in
a 70 °C oven (Carbolite, Sheffield, UK) for 10 h.
They were then equilibrated over saturated LiCl
salt solution in air-tight plastic boxes for one week
at room temperature to attain a final moisture
content of 10–11% (dry basis). Apart from the
examination of degree of expansion, samples for
all other methods were ground using a coffee
grinder (Moulinex, Ireland). The powder produced
was sieved and fractions collected between the 106
and 422 ꢄm sized sieves were used. Samples were
stored in small plastic containers at ambient tem-
perature. The moisture content was determined in
two replicates by drying 1–2 g of samples using
Infrared dryer (Sartorius Thermo Control YTC
O1L, Germany) at 160 °C until constant weight.
Viscosity measurement of alkaline solutions
Viscosity of extruded samples gives information
on the state of starch conversion as a result of the
extrusion process. This estimate used in this study
was based on that of Sriburi et al., (1999a). We
mixed powdered extruded samples (0.2 g) with
ethyl alcohol (0.5 mL). To this suspension was
added 4 mL of 5 M KOH. The samples were
stirred overnight and a clear solution was then vis-
ible. These solutions were diluted to 40 mL with
water (0.5% starch) and the viscosities of aliquots
(30 mL) of the resulting solutions were measured
using a Bohlin CS10 Rheometer (Bohlin, Lund,
Sweden) using a double gap concentric cylinder
measuring geometry (DG 40/50) over the increas-
Water absorption and water solubility indices
The water absorption index (WAI) measures the
volume occupied by the granule or starch polymer
after swelling in excess water. While water solubil-
ity index (WSI) determines the amount of free
polysaccharide or polysaccharide released from the
granule on addition of excess water. The WAI and
WSI can be used to characterize extruded products
ꢀ1
ing shear rate range 1–100 s . Owing to its large
surface area, this particular measuring geometry is
suitable for measuring low viscosity fluids. The
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

144
Extrusion of cassava starch P. Sriburi & S. E. Hill
the sample. The sample was then inserted into the
RVA Series (Newport Scientific, NSW,
Australia) which was used along with the accom-
panying software (Thermocline). The analysis
used the standard 1 profile and the temperature of
the system is shown in Fig. 1.
samples were equilibrated at 25.0 ꢅ 0.1 °C for 10
min in the rheometer before the first measurement
was taken. The results are reported at a shear rate
4
ꢀ1
of 10 s .
Viscosity measurement using the Rapid Visco
Analyser (RVA)
X-ray diffractometry
The RVA is an instrument for determining the vis-
cous properties of starch under controlled heating
conditions. When starch granules are heated in
water beyond a certain temperature, the granules
absorb a large amount of water and swell to many
times their original size. When the majority of
granules become swollen, a rapid increase in vis-
cosity occurs as the fragile swollen granules break
down. The peak viscosity occurs at the equilibri-
um point between swelling causing an increase in
viscosity, and rupture of the swollen granules
causing viscosity decrease. As the mixture is sub-
sequently cooled, re-association between starch
molecules can occur. In sufficient concentration
this causes the formation of gel, and viscosity will
normally increase to a final viscosity. An example
of the RVA profile for native cassava starch is
shown in Fig. 1.
X-ray measurement detects the semicrystalline
character, which reflects the presence of both the
ordered and amorphous regions within the starch
granule. These will be lost in starch gelatinization
or high levels of starch conversion.
The pattern of crystallinity of the extruded
product was assessed by X-ray diffraction. The
powered samples were put into an aluminium
holder, and were analyzed using Siemens D5005
X-ray
diffractometer
(Bruker
Analytical,
Germany). The sample was subjected to the fol-
lowing conditions: target, CuKꢆ; a graphite crys-
tal monochromator; angular range, 4–32º (2ꢇ);
angular interval, 0.05º (2ꢇ); copper tube; 40 kV;
50 mA; wavelength, 0.154 nm. Data were collect-
ed by a microcomputer and duplicate scans were
made for each sample.
Samples for RVA analysis were based on 3 g
(dry weight basis) of samples plus buffer of 25 g.
The paddle was placed into the canister contain-
ing the sample and buffer. The paddle blade was
vigorously jogged up and down 10 times through
Measurement of linear expansion
Each extrudate piece with a weight between 1.5
and 2.0 g was marked using a marker pen (Edding
Figure 1 An example of the RVA
profile for native cassava starch
(3 g) in the presence of phosphate
buffer (25 g). The Standard 1
temperature (°C) profile during
the RVA measurement is shown.
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

Extrusion of cassava starch P. Sriburi & S. E. Hill
145
Figure 2 Sample preparation for
measurement of linear expansion.
The length of each line was
measured before and after frying.
780, Ahrensburg, Germany) as shown in Fig. 2.
Linear expansion was calculated by measuring
the 4 lines shown in Fig. 2 before and after frying
in hot vegetable oil at 185 °C for 15 sec.
Measurements were made on 10 replicates from
each run for each treatment. The percentage lin-
ear expansion was calculated as follows (Yu et al.,
1981):
was based on the AOAC standard method
(Helrich, 1990) and that of Kirk & Sawyer (1991),
but a further two-fold dilution of the indophenol
standard solution was made. The ascorbic acid
content of samples was estimated by mixing 0.5 g
of sample with 10 mL of extraction solution
(metaphosphoric acid-acetic acid solution) for 1 h
and titrating 5 mL of extract supernatant with
2,6-dichlorophenolindophenol and correcting for
the blank using an equivalent amount of extrac-
tion solution. The percentage of ascorbic acid was
calculated as follows:
% Linear expansion ꢁ
(length after fryingꢀlength before drying)ꢂ100
length before frying
ꢀ2
% ascorbic acid (g/100 g sample) ꢁ 2.46 ꢂ 10
Determination of ascorbic acid
(A ꢀ B)
Ascorbic acid levels were determined to find out
how much ascorbic acid was used during extru-
sion cooking. The amount of ascorbic acid in the
milled samples after extrusion was estimated by
the method of titration with indophenol. There
are advantages in using indophenol as it reacts
almost instantaneously in a fixed stoichiometry
with ascorbic acid at all concentration levels; it
acts as a self-indicator; and because of the deep
blue colour of its solutions, a very dilute titrant
can be used without hampering visual detection of
the end point. The last advantage accounts for the
sensitivity of indophenol titration, which is much
greater than those of other titrations (Verma
et al., 1996).
where A and B are average volume (mL) of
indophenol used for sample titration and blank
titration, respectively.
Two samples were taken from each extruder
run and titrated in duplicate. Results are averaged
from the eight values thus obtained.
pH
The pH values were measured on 10% (w/v) sus-
pensions of the powdered samples using a pH
meter (Jenway 3320, Jenway, Essex, UK).
Statistical analysis
The ascorbic acid powder used in the prepara-
tion of the extrudates was the same as that used
as the standards in the titrations. Care had been
taken to use a new batch material at the start of
the work. This batch was then subdivided and
stored at 4 °C before use.
We have used either Analysis of Variance
(ANOVA) or paired t-test where appropriate to
analysis the data. Statistical differences are indi-
cated in the text within the results’ section. Most
assays were replicated three times and mean val-
ues are given. Typically for the ascorbic acid stud-
ies, where two separate extruder trials had been
Determination of the amount of ascorbic acid
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

146
Extrusion of cassava starch P. Sriburi & S. E. Hill
Table 1 The variations in the measures for the control cassava extrudates obtained for two runs using ascorbic acid and
the control for the pH adjusted extruder run
Run no.
Cassava
starch
batch
pH of
extrudate
suspensions
WAI
(g/g)
WSI
(%)
RVA peak
viscosity
(mPa.s)
RVA final
viscosity
(mPa.s)
Alkaline
viscosity
(mPa.s)
Overall
linear
expansion
(%)
Ascorbic
Acid
No.1
No.2
No.2
6.50
6.75
6.56
(5.78
(0.05)
60.7
(0.81)
1384
(43)
393
(30)
(1.69
(0.01)
77.3
(5.7)
RUN 1
Ascorbic
Acid
(8.94
(0.09)
32.6
(0.77)
2072
(7)
495
(10)
(1.94
(0.01)
75.2
(5.8)
RUN 2
pH adjusted
(7.55
(0.05)
54.5
(0.45)
909
(38)
329
(10)
(1.90
(0.00)
73.2
(4.3)
RUN 3
All extrudates were formed at pH 6.5. Values are mean with standard deviation, brackets, of three values for all assessment
except linear expansion where 10 replicates were used.
done, the means from both runs were averaged.
Numbers of replicates are shown on the tables
and graphs.
differences (ANOVA, P ꢈ 0.001) between all the
measures of starch conversion. This is not untypi-
cal of work using extrusion as it is very difficult to
replicate exactly the performance of the equipment
(El-Dash, 1981; Sheard et al., 1986; Guha et al.,
1998). Another factor is that a different batch of
cassava starch from another year was used for the
second run for the ascorbic acid study compared
to the first run. The second batch of starch was
used for the pH study. This may have a significant
effect on the levels of starch conversion detected as
cassava starch is known to be very variable
Results and discussion
Cassava samples in ascorbic acid-buffer systems
were produced by a twin-screw extrusion through
a slit die to form nonpuffed samples. The extrud-
er conditions used were as identical as possible.
However, the results for the control samples, as
shown in Table 1, indicate that there are significant
Figure 3 Indicators of starch conversion in cassava extrudates formed by inclusion of ascorbic acid. Cassava starch was
extruded under conditions of barrel temperatures (40, 90, 110 and 75 °C), screw speed (200 r.p.m.), starch feed rate (5
ꢀ1
ꢀ1
kg.h ), buffer feed rate (3 kg.h ) and moisture content (37.5% wet basis). Water absorption index (WAI) is shown in
(a) and water solubility index (WSI) in (b). Two extruder runs were carried out; run 1 (ꢁ) and run 2 (ꢂ). Bars indicate
ꢅ one standard deviation of three replicates.
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

Extrusion of cassava starch P. Sriburi & S. E. Hill
147
(Asaoka et al., 1991, 1992, 1993; Mat Hashim
et al., 1992; Fernandez, 1996; Sriburi et al., 1999b).
However, the differences we have found between
runs were far greater than the differences between
the two batches of cassava starch.
Water absorption and solubility indices
A high WSI and low WAI would indicate that the
starch has undergone extensive conversion. The
results obtained for these two parameters after the
addition of ascorbic acid are illustrated in
Fig. 3(a), (b). WAI and WSI results show signifi-
cant changes (ANOVA, P ꢈ 0.001) owing to ascor-
bic acid addition and therefore it could be
supposed that ascorbic acid encourages gelatiniza-
tion and molecular breakdown of the starch. The
higher the level of ascorbic acid the greater the
amount of starch conversion, but our results show
that even very low additions of ascorbic acid cause
a difference in the behaviour of the starch.
Extrudate at 0.1% ascorbic acid concentration, the
lowest level used, showed a significantly (paired t-
test) higher WSI (P ꢈ 0.001) and lower WAI (P ꢈ
0.002) than the milled controls.
Figure 4 RVA profiles for cassava extrudates formed in
the presence of ascorbic acid (extruder conditions given in
Fig. 3). Milled extrudates (3g) were mixed with 25g
buffer. Lines represent the average of the two extruder
runs, with each extruder run representing three replicated
assays. The temperature (°C) profile during the RVA
measurement is shown.
molecular starch polymers. This final viscosity
also decreases with ascorbic acid addition (Fig. 5).
Starches can be solubilized in potassium hydrox-
ide and therefore viscosity estimates of these solu-
tions can be assumed to be related to the actual
average molecular size of the starch polysaccha-
rides (Hoseney et al., 1992). This should be a bet-
ter measure of molecular size compared to the
RVA final viscosities where starch aggregation
could occur. The measurements on the extrudate
material for alkaline viscosity were made on a
Viscosities measured by the Rapid Visco
Analyser in alkaline solutions
The RVA viscosity profiles of all samples were
examined using the Standard 1 temperature pro-
file at the minimum and maximum temperatures
of 50 and 95 °C. The viscosity profiles are shown
in Fig. 4. This figure gives the average of the
results from both extruder runs. The results sug-
gest that the milled extrudate appear to swell
when added to water at a temperature below that
necessary for cassava starch gelatinization. Also,
no peak appears at the higher temperature which
one would expect if native starch was present
(Fig. 1). This indicates that the starch granules
must be extensively broken down within the
extruder.
Clearly, all addition levels of ascorbic acid have
a significant effect on cassava starch during extru-
sion as shown by the RVA viscosities (Fig. 4). The
swelling and the breakdown of the starch granules
seems to be different, as demonstrated by the peak
viscosity. The materials at the end of the temper-
ature cycle should reflect the size of the macro-
Figure 5 The effects of different levels of ascorbic acid on
cassava extrudates (see legend to Fig. 3 for details of
extrusion condition) on: (ꢃ) RVA final viscosity and
(ꢄ) alkaline viscosity. Each value is the mean of the two
extruder runs with data for each run being replicated
three times.
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

148
Extrusion of cassava starch P. Sriburi & S. E. Hill
0.5% starch (dry weight basis). Under these condi-
tions a low viscosity Newtonian solution was
obtained.
starch can be designated as possessing an ‘A’ type
pattern (Banks & Greenwood, 1975; Rickard
et al., 1991; Fernandez, 1996). Native cassava
starch shows peaks at approximately 15 and 23º
(2ꢇ) which are characteristics of the initial starch
A-pattern. These features are not apparent in the
extruded samples. It would appear that the native
cassava granule order is completely lost when
extrusion is carried out in both the absence and
presence of ascorbic acid. The level of ascorbic
acid does not seem to affect the X-ray pattern
observed of these extruded samples which had
been stored more than 18 h before assessment.
The results for the RVA final viscosities and
alkaline viscosities with the range of ascorbic acid
concentration used are displayed in Fig. 5. The vis-
cosities for both estimates are lower on addition of
ascorbic acid. Even the addition of 0.1% ascorbic
acid causes a significant change (paired t-test, P ꢈ
0.003) for both these estimates. These results could
be explained by the concept that during the extru-
sion ascorbic acid has affected the starch polymer
so that depolymerization of the starch macromole-
cules has occurred.
Determination of ascorbic acid
X-ray diffractometry
Figure 7 shows that the greater the level of ascor-
bic acid addition to form the extrudate the more
it is consumed. However, it was found that at the
higher rates of ascorbic acid inclusion of 0.7 and
1.0%, only 25% of the ascorbic acid was used
while at the addition of 0.1%, 50% was utilized.
There has been no attempt to find the products of
the reaction and it is not known if the compound
that initiates the changes in the starch is the
ascorbic acid or the dehydro materials.
Using the X-ray diffraction technique, the data
obtained in Fig. 6 suggests that native cassava
Linear expansion
Figure 8 shows the percentages of the linear
expansion of cassava extrudates, which had been
Figure 7 Estimates of ascorbic acid content of extruded
samples (see legend to Fig. 3 for preparation). Values are
given as the weight of ascorbic acid used per 100 g of
starch (ꢀ) and as the percentage (ꢅ) of the starting level.
Each value represents duplicate readings from two
samples taken from the two extruder runs.
Figure 6 (X)-ray diffraction patterns of cassava extrudates
formed by inclusion of ascorbic acid at various levels
(extruder conditions given in Fig. 3). The pattern for
native cassava starch is also shown.
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

Extrusion of cassava starch P. Sriburi & S. E. Hill
149
prepared with a varying levels of ascorbic acid.
Under frying conditions of 185 °C for 15 s and an
initial moisture content between 10 and 11% (dry
basis), four measures of expansion were observed.
As can be seen from Fig. 8 all linear expansion
measures increased with increasing ascorbic acid
content. As the level of ascorbic acid increases,
the degree of expansion during frying significant-
ly increases (ANOVA, P ꢈ 0.0001).
According to Siaw et al. (1985), a linear expan-
sion greater than 77% is required for an accept-
able level of crispness. From this work, it can be
said that all samples were satisfactory expanded,
although the expansion in the direction in which
extrusion took place (line 4) appeared to be rel-
atively less than for the other directions. From
the overall percentages of linear expansion, it
seems possible that the linear expansion could be
affected by the chemical action from ascorbic
acid.
Figure 9 The pH of milled cassava extrudates
(10% suspension; w/v taken from extruder run 2 and 3).
The extrudates had been formed in the presence of
ascorbic acid (ꢇ) or by changing the hydrogen ion
concentration of the extruder feed solution (ꢈ).
See legend to Fig. 3 for further details of the extrusion
condition.
4.3 and it could be expected that at this level acid
hydrolysis of the starch could play an important
role in altering the behaviour of the starch during
and subsequent to extrusion. We therefore con-
sidered it necessary to carry out a series of exper-
iments where the samples would experience a
range of pHs similar to those shown by the ascor-
bic acid samples. Solutions were used in the liq-
uid stream during extrusion that had pHs ranging
between pH 1 to pH 6.5 and the resultant extru-
dates had pHs from 2.15 to 6.63 (see Fig. 9).
Similar assessments to those carried out on the
ascorbic acid treated samples were done on the
extrudates formed by altering their pH.
Influence of pH during extrusion
A factor that must be considered is the pH of the
samples. Milled extrudates formed by the addi-
tion of ascorbic acid were suspended in the dis-
tilled water and the pH measured. Figure 9
represents data from the second extruder run and
shows that the greater the addition of ascorbic
acid, the lower the pH value. The lowest pH was
The sample using 0.1 M HCl (starting pH of
1.0, extrudate pH of 2.25) did not form a ribbon
of material, but a powderlike product was col-
lected. Samples representing starting pH values 2
to 6.5 all formed ribbons of materials, which were
similar in appearance to those formed by the
inclusion of ascorbic acid. However, the ascorbic
acid samples were lighter in colour compared to
the pH adjusted samples.
Figure 8 Variations in linear expansion of cassava
extrudates (see legend to Fig. 3 for extruder details), after
frying in vegetable oil at 185 °C for 15 sec. The
orientations measured are described by Fig. 2;
(ꢆ) diagonal lines, (ꢇ) line 3, (ꢂ) line 4 and (ꢁ) overall.
Values are mean standard deviation of 20 samples
(10 from each run).
Comparison of extrusion runs
All the pH changes were achieved in one extru-
sion run. As demonstrated for the two runs for
the ascorbic acid addition, it is difficult to repli-
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

150
Extrusion of cassava starch P. Sriburi & S. E. Hill
Figure 10 Indicators of starch
conversion; (ꢆ) WAI and (ꢇ)
WSI, in cassava extrudates
formed using extruder feed
solutions of differing pH. Values
are the mean of three replicates.
cate results between runs. The control samples
used for the pH adjusted extrudates should have
the same values as the zero ascorbic acid level
samples. Table 1 shows the extent of the variation
that occurred despite the care in getting the con-
dition the same. This variation between runs
needs to be considered when comparing the data
between the pH adjusted samples and the ascor-
bic acid samples.
in Fig. 13. There seems to be no clear relationship
between the pH and the expansion values.
pH and product quality
Adjustment of the starch buffer system to achieve
a range of pH values does seem to cause a varia-
tion in the rheological properties of the product,
lower pHs causing
a
decreased viscosity.
However, we found that the other measures such
as WAI and expansion are not affected by mod-
erate changes in pH.
WAI and WSI for pH adjustment
The WAI and WSI values show no great depen-
dence on the pH of the sample except the samples
at pH 1.0 (nominal). At this very low pH value
there is a marked increase in the amount of solu-
ble material and the ability of the particles to
swell, as indicated by WAI, is much decreased
when compared to the other samples (Fig. 10).
Comparisons of pH values and ascorbic acid
samples
To help negate the effects of the variations
between runs, thereby elucidating if there were
RVA and alkaline viscosity
The RVA curves representing the extrudates
formed at a range of pH values are shown in
Fig. 11. Clearly decreasing the pH of the extrud-
er feed solution decreases the RVA viscosities
(Fig. 12). The alkaline viscosity is also decreased
when the pH values are decreased (Fig. 12).
Linear expansion
Figure 11 RVA profiles for cassava extrudates formed
using extruder feed solutions of differing pH (see legend
to Fig. 3 for details of extruder condition).
All the extruded ribbon products were fried and
their expansion measured. The results are shown
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

Extrusion of cassava starch P. Sriburi & S. E. Hill
151
Figure 12 The effects of using
extruder feed solutions of
differing pH (see legend to Fig. 3
for details of extruder condition)
on cassava starch on: (ꢈ) RVA
peak viscosity, (ꢃ) RVA final
viscosity and (ꢄ) alkaline
viscosity.
specific ascorbic acid effects above the changes
caused by lowering the pH, we directly compared
six samples. These were the three control samples
from the three runs as shown in Table 1 and a
test sample from each run. The test samples from
run 1 and 2 were the ascorbic acid additions of
1.0% (starch bases) which produced extrudates of
pH 4.3. The pH adjusted sample made in run 3
from buffer of pH 3 was the third test sample.
This material had a pH of 3.74, i.e. lower than for
the ascorbic acid samples. Each of the assess-
ments on these three samples was calculated as a
percentage of the result obtained for their corre-
sponding control, with positive values represent-
ing greater starch conversion. These results are
shown in Fig. 14 and demonstrates that over the
range of assessments undertaken the inclusion of
1% ascorbic acid or the lowering of pH to 3.74
does promote starch conversion. It is also appar-
ent that the magnitude of the affects is greater by
the inclusion of ascorbic acid than by addition of
phthalate and hydrogen chloride to produce an
extrudate of lower pH than that of the ascorbic
acid test samples.
Conclusions
The work described in this paper shows that the
inclusion of ascorbic acid or a reduction in pH
during extrusion alters cassava starch. Higher lev-
els of conversion are observed, as measured by
solubility, water absorption and changes in vis-
Figure 13 Variations in linear
expansion of pH adjusted cassava
extrudates. Other details as for
Fig. 8.
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

152
Extrusion of cassava starch P. Sriburi & S. E. Hill
between the three sets of extrusion runs per-
formed on different occasions. There was a time
lapse of approximately one month between the
sets of runs. It must be emphasized that each set
of extrusion experiments was carried out with the
outmost care with same experienced operators,
yet there were differences between the control val-
ues obtained. The first set of work was done with
one batch of cassava starch, but the second ascor-
bic acid and the pH extrusion runs were carried
out using another batch of material. It is well
known (El-Dash, 1981; Guha et al., 1998) that
reproducibility between extrusion runs is often
poor and that often comparisons should only be
made within one set of extruder runs.
The other and major consideration necessary
for the interpretation of the ascorbic acid data is
the lowering of the pH that the inclusion of the
ascorbic acid causes. It could be considered that
acid hydrolysis could account for the starch’s
behaviour in the presence of the vitamin C. By
matching the results at similar pH values caused
by the inclusion of ascorbic acid or by altering
the solution using acids or buffers it is clear that
much greater effects are seen with the ascorbic
acid. However, acidic conditions do encourage
starch conversion and this is marked at very low
pH values, for example at pH values less than 2
it was not possible to obtain a ribbon of extru-
date. What was surprising was that there was no
suggestion that the expansion of the acid prod-
ucts was altered, even though the viscosity of the
materials had been affected by the acid environ-
ment. Cabrera (1978 as quoted by Matz, 1993)
found that wheat extrudates formed at pH values
between 4.4 and 9.0 did not vary in their expan-
sion ratio. In these experiments cassava products
formed at pH values in the range 6.5 to 4.5
achieved by the inclusion of ascorbic acid, indi-
cate a marked change in the expansion ratio. It is
therefore possible that the mechanisms for the
depolymerization brought about by acid hydroly-
sis and free radical attack are different and this
affects the ability to expand. This is something
that will be explored in later experiments.
Figure 14 Comparison of test samples for three extruder
runs. Test values have been calculated as a % of the
control for the corresponding extruder run. A positive
value has been used to indicate promotion of starch
conversion. The runs are labelled (1), (2) and (3). Runs
(1) and (2) used 1% ascorbic acid (starch basis) as the
test sample and run (3) used a sample produced using an
extruder feed solution of pH 3. The pH of all the control
samples were between pH 6.50 and 6.75. The pH of the
test samples for runs (1), (2) and (3) were respectively
4.44, 4.27 and 3.80. Estimates of starch conversion were
WAI, WSI, RVA peak and final viscosities, alkaline
viscosity and linear expansion.
cosity. Although needing careful interpretation
the RVA viscosity profiles was shown to be an
easy and sensitive way of monitoring changes that
had occurred during extrusion.
Alkaline viscosity results show a decrease as
the ascorbic acid levels increase and therefore it is
suggested that the ascorbic acid is causing a
depolymerization of the starch macromolecules.
Earlier work (Paterson et al., 1996; Vallès-Pàmies
et al., 1997) has indicated that the reaction is a
free radical attack and that the amylose and amy-
lopectin are depolymerized.
A critical quality parameter for cassava starch
products is its ability to puff. Expansion of the
products was increased by the addition of ascor-
bic acid and therefore addition of this reducing
agent could be directly related to the quality of
final product.
Although the case for ascorbic acid addition
changing the starch during extrusion seems well
proven, two other factors needed to be considered
when interpreting the results from the practical
work. Firstly, there is a marked difference
Acknowledgments
Pensiri Sriburi would kindly thank the Royal
Thai Government for financial support. The
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

Extrusion of cassava starch P. Sriburi & S. E. Hill
153
Fernandez, A.Q. (1996). Effects of processing, procedures
and cultivar on the properties of cassava flour and
starch. PhD Thesis. Nottingham, UK: University of
Nottingham.
authors would like to thank Prof. John Mitchell
for helpful discussions.
Fitchett, C.S. & Frazier, P.J. (1986). Action of oxidants
and other improvers. In: Chemistry and Physics in
Baking: Materials, Processes, and Products (edited by J.
M. V. Blanshard, P. J. Frazier & T. Galliard). Pp.
179–198. London: The Royal Society of Chemistry.
Grosch, W. (1986). Redox systems in dough. In: Chemistry
and Physics in Baking: Materials, Processes, and
Products (edited by J. M. V. Blanshard, P. J. Frazier &
T. Galliard). Pp. 155–169. London: The Royal Society
of Chemistry.
Grosch, W. & Wieser, H. (1999). Redox reactions in wheat
dough as affected by ascorbic acid. Journal of Cereal
Science, 29, 1–16.
Guha, M., Ali, S.Z. & Bhattacharya, S. (1998). Effect of
barrel temperature and screw speed on rapid
viscoanalyser pasting behaviour of rice extrudate.
International Journal of Food Science and Technology,
33, 259–266.
Halliwall, B. & Gutteridge, J.M.C. (1995). Free Radicals in
Biology and Medicine. Oxford: Clarendon Press.
Helrich, K. (ed). (1990). AOAC Official Methods of
Analysis. vol 2, 15th ed. Pp. 1058–1059. Virginia:
Association of Official Analytical Chemists.
Hoseney, R.C., Mason, W.R., Lai, C.-S. & Guetzlaff, J.
(1992). Factor affecting the viscosity and structure of
extrusion-cooked wheat starch. In: Extrusion Science and
Technology (edited by J. L. Kokini, C. T. Ho & M. V.
Karwe). Pp. 277–305. New York: Marcel Dekker Inc.
Kirk, R.S. & Sawyer, R. (1991). Pearson’s Composition and
Analysis of Foods. Pp. 243–244. Singapore: Longman
Scientific & Technical.
Kokini, J.L., Lai, L.S. & Chedid, L.L. (1992). Effect of
starch structure on starch rheological properties. Food
Technology, 46, 124–139.
Lillard, D.W., Sieb, P.A. & Hoseney, R.C. (1982).
Isomeric ascorbic acids and derivatives of L-ascorbic
acid: Their effect on the flow of dough. Cereal
Chemistry, 59, 291–296.
References
Asaoka, M., Blanshard, J.M.V. & Rickard, J.E. (1991).
Seasonal effects on the physico-chemical properties of
starch from four cultivars of cassava. Stärke, 43,
455–459.
Asaoka, M., Blanshard, J.M.V. & Rickard, J.E. (1992).
Effects of cultivar and growth season on the
gelatinisation properties of cassava (Manihot esculenta)
starch. Journal of the Science of Food and Agriculture,
59, 53–58.
Asaoka, M., Blanshard, J.M.V. & Rickard, J.E. (1993).
The effects of pre-harvest pruning on the quality of
cassava starch. Annals of Applied Biology, 122, 337–344.
Banks, W. & Greenwood, C.T. (1975). Starch and its
Components. Pp. 242–243. Great Britain: The University
of Edinburgh.
Blanshard, J.M.V. (1994). Cassava starch, structure,
properties and implications for contemporary processing.
In: Proceedings at the 2nd International Scientific
Meeting ‘the Cassava Biotechnology Network’. Pp.
625–638. Bogor, Indonesia.
Chamberlain, N. (1981). The use of ascorbic acid in
breadmaking. In: Vitamin C (Ascorbic Acid) (edited by
J. N. Cousell & D. H. Horning). Pp. 87–104. London:
Applied Science Publishers.
Colonna, P. & Mercier, C. (1983). Macromolecular
modifications of manioc starch components by
extrusion-cooking with and without lipids. Carbohydrate
Polymers, 3, 87–100.
Colonna, P., Doublier, J.L., Mercion, J.P., de Monredon,
F. & Mercier, C. (1984). Extrusion cooking and drum
drying of wheat starch. Cereal Chemistry, 61, 538–554.
Colonna, P., Tayed, T. & Mercier, C. (1989). Extrusion
cooking of starch and starchy products. In: Extrusion
Cooking (edited by C. Mercier, P. Linko & J. M.
Harper). Pp. 247–319. St. Paul, MN: American
Association of Cereal Chemists.
Diosady, L.L., Paton, D. & Rosen, N. (1985). Degradation
of wheat starch in a single screw-extruder: Mechano-
kinetic breakdown of cooked starch. Journal of Food
Science, 50, 1697–1699, 1706.
Mat Hashim, D.B., Moorthy, S.N., Mitchell, J.R., Hill,
S.E., Linfoot, K.J. & Blanshard, J.M.V. (1992). The
effect of low levels of antioxidants on the swelling and
solubility of cassava starch. Stärke, 44, 471–475.
Matz, S.A. (1993). Snack Food Technology. 3rd edn. Pp.
159–172. New York: AVI Publishing.
Mitchell, J.R., Hill, S.E., Paterson, L.A., Vallès-Pàmies, B.,
Barclay, F. & Blanshard, J.M.V. (1997). The role of
molecular weight in the conversion of starch. In: Starch:
Structure and Functionality (edited by P. J. Frazier, A.
M. Donald & P. Richmond). Pp. 68–76. Cambridge:
The Royal Society of Chemistry.
El-Dash, A. (1981). Application and control of
thermoplastic extrusion of cereals for food and industrial
uses. In: Cereals: A Renewable Resources: Theory and
Practice (edited by Y. Pomeranz & L. Munck). Pp.
165–216. St Paul, Minnesota: The American Association
of Cereal Chemists.
Elkassabany, M. & Hoseney, R.C. (1980). Ascorbic acid as
an oxidant in wheat flour dough. II. Rheological effects.
Cereal Chemistry, 57, 88–91.
Myllymäki, O., Eerikainen, T., Suortti, T., Forssell, P.,
Linko, P. & Poutanen, K. (1997). Depolymerisation of
barley starch during extrusion in water glycerol
mixtures. Lebensmittel-Wissenchaft und Technologie, 30,
351–358.
Fan, J., Mitchell, J.R. & Blanshard, J.M.V. (1996). The
effect of sugars on the extrusion of maize grits: I. The
role of the glass transition in determining product
density and shape. International Journal of Food Science
and Technology, 31, 55–65.
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd

154
Extrusion of cassava starch P. Sriburi & S. E. Hill
Pachla, L.A. & Reynolds, D.L. (1985). Review of ascorbic
acid methodology. Journal of Association Official
Analytical Chemists, 68, 1–12.
Publishers.
Tsen, C.C. (1965). The improving mechanism of ascorbic
acid. Cereal Chemistry, 42, 86–96.
Paterson, L.A., Mitchell, J.R., Hill, S.E. & Blanshard,
J.M.V. (1996). Evidence for sulfite induced oxidative
reductive depolymerisation of starch polysaccharides.
Carbohydrate Research, 292, 143–151.
Rickard, J.E., Asaoka, M. & Blanshard, J.M.V. (1991).
Review: the physico-chemical properties of cassava
starch. Tropical Science, 31, 189–201.
Sheard, P.R., Mitchell, J.R. & Ledward, D.A. (1986).
Extrusion behaviour of different soya isolates and the
effect of particle size. Journal of Food Technology, 21,
627–641.
Shugar, G.J., Shugar, R.A., Bauman, L. & Bauman, R.S.
eds. (1981). Chemical Technicians’ Ready Reference
Handbook. 2nd edn. P. 539. New York: McGraw-Hill.
Siaw, C.L., Idrus, A.Z. & Yu, S.Y. (1985). Intermediate
technology for fish crackers (‘keropok’) production.
Journal of Food Technology, 20, 17–21.
Vallès-Pàmies, B., Barclay, F., Hill, S.E., Mitchell, J.R.,
Paterson, L.A. & Blanshard, J.M.V. (1997). The effects
of low molecular weight additives on the viscosities of
cassava starch. Carbohydrate Polymers, 34, 31–38.
Vasanti Nair, C.K., Seow, C.C. & Sulebele, G.A. (1996).
Effects of frying parameters on physical changes of
tapioca chips during deep-fat frying. International
Journal of Food Science and Technology, 31, 249–256.
Verma, K.K., Jain, A., Sahasrabuddhey, B., Gupta, K. &
Mishra, S. (1996). Solid-phase extraction cleanup for
determining ascorbic acid and dehydroascorbic acid by
titration with 2,6-dichlorophenolindophenol. Journal of
AOAC International, 79, 1236–1243.
Yu, S.Y. (1991). Acceptability of fish crackers (keropok)
made from different types of flour. ASEAN Food
Journal, 6, 114–116.
Yu, S.Y. (1993). Effect of rice starch on the linear
expansion of fish crackers (keropok). Tropical Science,
33, 319–321.
Sriburi, P., Hill, S.E. & Mitchell, J.R. (1999a). Effects of
L-ascorbic acid on the conversion of cassava starch.
Food Hydrocolloids, 13, 177–183.
Sriburi, P., Hill, S.E. & Barclay, F. (1999b).
Depolymerisation of cassava starch. Carbohydrate
Polymers, 38, 211–218.
Yu, S.Y. & Low, S.L. (1992). Utilization of pre-gelatinised
tapioca starch in the manufacture of a snackfood fish
crackers (‘keropok’). International Journal of Food
Science and Technology, 27, 593–596.
Thewlis, B.H. (1974). Vitamin C in breadmaking. In:
Vitamin C: Recent Aspects of its Physiological and
Technological Importance (edited by G. G. Birch & K. J.
Parker). Pp. 150–160. London: Applied Science
Yu, S.Y., Mitchell, J.R. & Abdullah, A. (1981).
Production and acceptability testing of fish crackers
(‘keropok’) prepared by the extrusion method. Journal
of Food Technology, 16, 51–58.
International Journal of Food Science and Technology 2000, 35, 141–154
© 2000 Blackwell Science Ltd