Vol.4, No.9B, 51-55 (2013) Agricultural Sci ences
http://dx.doi.org/10.4236/as.2013.49B009
Integratio n of p hy s ic al a nd chemical t re a tm e nt on th e
extraction of starch from Canna edulis Ker. rhizome
Judy R. B. Witono*, Herry Santoso, Y. I. P. Arry Miryanti, Daniel Tan
Chemical Engineering Department, Parahyangan Catholic University, Bandung, Indonesia;
*Corresponding Author: [e-mail]
Received August 2013
ABSTRACT
The extraction of Canna edulis Ker. starch from
its rhizome was performed using 2 different
types of press (hy draulic press and scre w press)
and with the addition of Na-metabisulphite and
NaOH (in the range of concentration 100 - 5000
ppm each). The optimum condition for this
process was determined by Central Composite
Design of experiment and the statistical calcula-
tion was solved by Design-Expert 7.0.0. The
targets of the observed responses were high
starch yield, low ash, low fiber, and high carbo-
hydrate content. The results showed that the
starch yield and the reduction of fiber were only
influenced by the physical treatment whereas
ash content in the product was influenced by
both the NaOH concentration and physical treat-
ment. The carbohydrate content in the extrac-
tion product was affected by NaOH, by the inte-
raction bet ween the concentrations of NaOH and
Na2S2O5 and also by the physical treatment. The
hydraulic press gives much better responses
compared to the screw press. But in the se-
lected range of additives concentrations, the
screw press gives a higher starch yield (30% -
52%).
Keywords: Canna edulis Ker.; Central Composite
Design of Experim ent ; Hydraulic Press; Screw
Press; Starch Extraction
1. INTRODUCTION
One of the tropical starch resources which have not
been utilized for industrial application is Canna edulis
Ker. rhizome, partly because of the difficulties in the
extraction processing [1]. The high content of fiber and
other trace elements are a major constraint in producing
pure starch. A sample of 100 g of Canna edulis rhizome
contains 125 mg phosphorus (P), 84 mg calcium (Ca) and
1.5 mg iron (Fe) mineral [2]. The presence of those ele-
ments in food products can be categorized as a nutritional
value but it will be a disadvantage for a chemical starch-
based pr od uc t.
Salt solutions like NaCl, Na-bisulphite, and Na-meta-
bisulphite are commonly used during extraction of starch
from its natural source, to inhibit microbial growth and
deactivate plant enzyme (amylase). The salt solution can
dissolve the surface starch granule protein as well, but
for the breakdown of the integral starch granule protein,
stronger solutions are required, e.g. sodium dodecyl sul-
phate (SDS) [3 ,4] or alkaline solutions [5].
Lim et al. [6] have investigated that the use of 0.2% of
NaOH as an extraction solution for rice starch could re-
duce more than 80% of the flour protein. This was con-
firmed also by Radosavljevic et al. [7] on the extraction
of Amaranth starch; Mistry et al. [8] on the extraction of
corn flour using 0.1% and 0.4% of NaOH. It was inves-
tigated also that sodium hydroxide (NaOH) can remove
phosphorus up to 70% - 90% fr om wheat starch [9].
Since the properties of starch are, to some extent, dif-
ferent from the fiber (cellulose), a relatively low cost
method for isolation of the starch is physical treatment.
The goal of the research reported here is to determine the
optimum condition of the integration of physical and
chemical treatments in producing pure starch. For this
purpose, a Central Composite Experimental Design me-
thod (CCD) was applied, statistical calculations were made
using Design expert 7.0.0 soft wa re.
2. MATERIALS & METHODS
2.1. Materials
Freshly harvested Canna edulis Ker. rhizome (locally
known as Ganyong) was supplied by the farmers union
“Mekar Sari” at the Kulon Progo region (Central Java)
Indonesia . Analytical grade of sodium metabisulphite
and NaOH were purchased from Sigma Aldrich. α-amy-
lase was supplied by N o voz yme and Anthrone rea gent by
Merck.

J. R. B. Witono et al. / A gricultural Sciences 4 (2013) 51-55
52
2.2. Design of the Experiment
To determine the optimum condition from 2 numerical
factors (the Na-metabisulphite and NaOH concentrations)
and 1 categorial factor (type of mechanical treatment or
type of press instrument), the Central Composite Design
was applied. This results in 26 runs with variations as
shown in Table 1.
2.3. Starch Extraction
Fresh Canna edulis was peeled and washed, then milled
with a cross beater mill. The pulp was mixed with water
at a weight ratio 1:10. Half of mixture was filtered with
the hydraulic press and the rest with the screw press, to
separate the crude fiber.
The free crude fiber mixture which consists of the fil-
trate from each press was divided into 13 portions that
were mixed with different concentrations of Na-metabi-
sulphite and NaOH solutions as stated Tab le 1. After 12
hours the precipitated starch was separated from the liq-
uor and dried in a tray drier for about 12 hours at 45˚C
until constant weight. The dried starch was kept in closed
containers to be analyzed further.
Table 1. Overview of experimental runs.
Factor 1 Factor 2 Factor 3
Run [Na2S2O5] [NaOH] Type of treatment*
ppm ppm
1 2550 2550 Level 2
2 2550 2550 Level 2
3 100 100 Level 2
4 2550 2550 Level 1
5 2550 2550 Level 2
6 5000 100 Level 2
7 100 5000 Level 1
8 2550 2550 Level 2
9 5000 5000 Level 2
10 5000 5000 Level 1
11 2550 2550 Level 1
12 100 2550 Level 1
13 100 2550 Level 2
14 2550 2550 Level 1
15 5000 2550 Level 1
16 5000 2550 Level 2
17 2550 2550 Level 1
18 2550 2550 Level 1
19 2550 5000 Level 2
20 2550 2550 Level 2
21 100 5000 Level 2
22 100 100 Level 1
23 2550 100 Level 1
24 2550 5000 Level 1
25 2550 100 Level 2
26 5000 100 Level 1
*Level 1—hydraulic press and level 2—screw press.
2.4. Fiber Content
The fiber content was measured by heating a mixture
of 5 g of starch and 50 mL of water until 90˚C. Then 4
mL of α-amylase enzyme was added and the total was
kept at 90˚C for a further 30 minutes. After that 50 mL of
water was added and the sample was cooled to room
temperature.
This method was based on the characteristic property
of the α-amylase enzyme which degrades carbohydrates
to produce shorter chain molecules such as glucose, which
is soluble in water. The cooled mixture was filtered using
Whatman 42 filter paper under vacuum conditions to
separate the fine fiber. The Fiber content (FC) was cal-
culated using Eq.1.
( )
1
2
w
FC %=100%
w×
(1 )
Where w1 is the weight of fine fiber and w2 is the
weight of starch sample (dry basis).
2.5. Ash Content
Ash content measurement was also based on a gravi-
metric technique. 1 g of star ch sample was combusted in
a furnace at 600˚C for 1 hour then cooled in desiccator
and weight. This step was repeated until the weight was
constant. Ash content (AC) was calculated using Eq.2.
( )
3
4
w
AC %=100
w×
(2)
Where w3 is the weight of ash and w4 is the weight of
starch sample (dry basis).
2.6. Carbohydrate Content
The carbohydrate content was analyzed by anthrone
reagent [10]. The anthrone method was started by mak-
ing a standard graph which was made by diluting 10 mg
of glucose into 100 mL of water. The standards solution
was taken 0, 0.2, 0.4, 0.6, 0.8 and 1 mL (0 serves as
blank) and made up to 1 mL with distilled water. Then 4
mL of anthrone reagent was added.
Anthrone reagent was made by diluting 200 mg anth-
rone in 100 mL of ice-cold 95% H2SO4. All of the stan-
dard solutions were heated for eight minutes in a boiling
water bath and then cooled rapidly. Carbohydrate content
was measured with a spectrophotometer at 595 nm wa-
velength. A calibration graph was made by plotting con-
centrations of the standard solution on the X-axis against
and the absorbance on the Y-axis.
Starch must be hydrolyzed prior to the treatment with
anthrone solution. Therefore, 100 mg of a sample of
starch with 5 mL of 2.5 N HCl added, was boiled in a
water bath for three hours. When sample temperature

J. R. B. Witono et al. / A gricultural Sciences 4 (2013) 51-55
53
was back to room temperature, it was neutralized with
solid sodium carbonate (Na2CO3) until the effervescence
ceases. The solution was then made up to 100 mL and
centrifuged and 0.1 mL of the supernatant was used for
analysis. The rest of the procedure is the same for the
standard solution. The amount of carbohydrate (CC) in
the sample tube was then calculated from the absorbance
in the spectrophotometer and the calibration graph by
Eq.3.
( )
[ ]
5
0.9C 1000
CC %=100
w
×× ×
(3)
where [C] is the glucose concentration measured from
the calibration graph whereas w5 is the weight of starch
sample (dry basis).
2.7. Starch Yi eld
Starch yield was measured by comparing the weight of
obtained starch (dry basis.) with the weight of dry matter
sample (Canna edulis rhizome). The starch has been free
from dirt 100. Starch Yield (SY) was determined by
Eq.4.
( )
6
7
w
SY %=100
w×
(4)
where w6 is the weight starch (d.b.) whereas w7 is the
weight of the dry matter original sample.
3. RESULTS AND DISCUSSION
The analysis results from every run can be seen in Ta-
ble 2.
3.1. Starch Yi eld
Based on the ANOVA of the values obtained for the
starch yield (Table 3) it can be seen that the additives
concentrations, both Na2S2O5 and NaOH do not affect
the yield of starch. This is also proved by the probability
value (P value) from both additives which are above
0.005. The yield obtained is apparently more determined
by the method of physical treatment.
Between the two types of press used, the higher starch
yield was found when the screw press was used for se-
paration (30% - 52%) compared to the hydraulic press
(25% - 41%). However, this fraction contains 79% ± 6%
carbohydrate which is slightly lower than the product
from the hydraulic press 82% ± 8%).
3.2. Carbohydrate Content
Based on the ANOVA of carbohydrate content (Table
4) it can be seen that the NaOH concentration and the
interaction between the concentrations of Na2S2O5-
NaOH affect the amount of carbohydrate obtained. Since
Table 2. Results of the experimental runs.
Run Fiber Ash Carbohydrate Yield
% % % %
1 3.7 4.0 76.0 14.1
2 2.4 1.9 76.4 33.7
3 2.9 0.7 76.9 49.7
4 1.6 1.0 82.2 35.0
5 3.3 1.5 82.0 40.7
6 2.7 1.5 74.5 44.0
7 2.0 2.7 88.9 31.7
8 2.6 1.6 82.1 40.9
9 3.8 3.8 92.0 64.9
10 2.4 3.5 88.7 47.0
11 1.9 1.3 94.0 31.5
12 2.5 1.3 81.3 35.1
13 3.2 2.0 82.3 43.2
14 2.3 2.9 76.7 25.2
15 2.8 1.7 73.8 33.9
16 3.2 2.3 77.0 57.6
17 2.5 1.2 74.2 35.0
18 2.9 3.0 81.1 11.1
19 2.7 4.3 66.5 30.0
20 2.7 2.7 81.8 34.6
21 3.2 3.6 75.4 45.7
22 1.4 0.4 82.7 39.0
23 2.6 0.5 78.7 37.4
24 2.4 2.6 93.6 30.4
25 3.2 1.0 77.7 39.2
26 1.4 0.7 65.1 38.7
Table 3. ANOVA for starch yield response.
Factor SS DF MS F Value p-Value
[Na2S2O5] 144.47 1 144.47 3.38 0.18
[NaOH] 0.24 1 0.24 2 0.95
Treatment 442.65 1 442.65 3.37E-003 0.02
Tabl e 4. ANOVA for carbohydrate content response.
Factor SS DF MS F Value p-Value
A:[Na2S2O5] 22.39 1 22.39 0.69 0.42
B:[NaOH] 204.67 1 204.67 6.28 0.02
Treatment 62.86 1 62.86 1.93 0.18
AB 165.91 1 165.91 5.09 0.04
BC 132.74 1 132.74 4.07 0.06
the p-value of BC ≈ 0.05, so it can be predicted that there
is a relation between the physical treatment and the addi-
tive (NaOH).
From Figure 1, it can be seen that at the high concen-
tration of Na2S2O5 and the high concentration of NaOH,
within the selected ranges of this research the carbohy-
drate content reached the maximum value.
However, if the 3D surface plot is observed for each
physical treatment (Figures 2 and 3), the effect of con-
centration of additives on the carbohydrate content gives
a different profile. The separation process using the screw

J. R. B. Witono et al. / A gricultural Sciences 4 (2013) 51-55
54
Figure 1. 3D plot of carbohydrate content.
Figure 2. 3D surface of carbohydrate content using the hydrau-
lic press.
Figure 3. 3D surface of carbohydrate content using the screw
press.
press is slower than the hydraulic press. This results in
the need for the higher concentration of Na-metabisul-
phite (5000 ppm) to achieve high carbohydrate content
as well as to prevent microbial growth in the slurry
which can destroy the carbohydrate. This is a different
result compared to the hydraulic press in which it is
enough to use 100 ppm to gain a high carbohydrate yield.
The models resulting from the statistical methods ill u-
strate the relation between the carbohydrate content and
the significant factors as followed:
For the hydraulic press
[ ] [ ]
[ ] [ ]
-3 -3
22 5
-7
22 5
CC = 83.23.710NaSO+1.110NaOH
+7.610NaS ONaOH
−× ×
×
(5)
For the screw press
[ ] [ ]
[ ] []
-3 -3
22 5
-7 22 5
CC=81.11.310NaSO+1.610 NaOH
+7.610NaSONaOH
−× ×
×
(6)
3.3. Fiber Content
Based on the ANOVA calculations (see Table 5), the
additives concentration, both Na2S2O5 and NaOH do not
affect the reduction of fiber content. This is also proved
by the P values for both additives (>0.05). So, the physi-
cal treatment determines the amount of fiber in starch
after extraction. This result is the same as for the starch
yield.
3.4. Ash Content
Ash content shows the presence of inorganic compo-
nent in the starch. These can originate from the rhizome,
but also from the chemicals added during processing.
From the ANOVA calculations (see Table 6) it can be
seen that the concentration of NaOH and the physical
treatment affect the amount of inorganic material left in
the starch. This seems due to the solubility level of
Na2S2O5 in water which is hi gher than Na OH.
On Figure 4, it is shown that although Na2S2O5 con-
centration does not affect the ash content significantly,
there is a tendency that the higher concentrations of ad-
ditives, both Na2S2O5 and NaOH, result in a higher ash
content in the starch.
4. CONCLUSIONS
This study showe d that the integration of physical and
chemical is a promising technology for the extraction of
starch from Canna edulis. It is a relatively simple and
low cost pr oc ess and it produces a good quality st a rc h.
Tabl e 5. ANOVA for fiber content response.
Factor SS DF MS F Value p-Value
[Na2S2O5] 0.089 1 0.089 0.41 0.5261
[NaOH] 0.43 1 0.43 2.00 0.1717
treatment 4.54 1 4.54 21.18 0.0001
Tabl e 6. ANOVA for ash content response.
Factor SS DF MS F Value p-Value
[Na2S2O5] 0.63 1 0.63 1.54 0.2284
[NaOH] 20.18 1 20.18 48.80 <0.0001
Treatment 2.55 1 2.55 6.18 0.0210

J. R. B. Witono et al. / A gricultural Sciences 4 (2013) 51-55
55
Figure 4. 3D surface of ash content.
The use of the screw press in the separation process of
the fibers after the chemical extraction produce gives a
higher starch yield compared to the use of hydraulic press,
but the purity is lower. Therefore, the hydraulic press is
perhaps the most suitable method for preparing starch
that has to be a feed material for chemical modification
processing.
5. ACKNOWLEDGEMENTS
The authors would like to express their thanks to the Indonesian Di-
rectorate General of Higher Education—Ministry of Education which
finances this research and also the Chemical Engineering fresh gra-
duates of Parahyangan Cath olic University: Ricky Gunadi, Pamela and
David who assisted the authors in the preparation of this research.
Special thanks are addressed to the leader of “Mekar Sari” farmers
union Bapak Kemin, who facilitated the collection of the Canna edulis
Ker. Rhizome, and Indonesian Institute of Sciences (LIPI, Subang)
which allowed us to use their equipment for extracting the starch.
REFERENCES
[1] Moorthy, S.N. (2002) Physicochemical and functional
properties of tropical tuber starches: A review. Starch-
Stärke, 54, 559-592.
http://dx.doi.org/10.1002/1521-379X(200212)54:12<559:
:AID-STAR2222559>3.0.CO;2-F
[2] Ganyong (2013) Tanaman Pangan.
http://tanamanpangan.deptan.go.id/
[3] Park, S.H., Bean, S.R., Wilson, J.D. and Schober, T.J.
(2006) Rapid isolation of sorghum and other cereal
starches using sonication. Cereal Chemistry, 83, 611-616.
http://dx.doi.org/10.1094/CC-83-0611
[4] Wang, S., Hassani, M.E., Crossett, B. and Copeland, L.
(2013) Extraction and identification of internal granule
proteins from waxy wheat starch. Starch-Stȁrke, 65, 186-
190.
[5] Wang, C., Tian, Z., Chen, L.,Temelli, F., Liu, H. and
Wang, Y. (2010) Functionality of barley proteins ex-
tracted and fractionated by alkaline and alcohol methods.
Cereal Chemistry, 87, 597-606.
http://dx.doi.org/10.1094/CCHEM-06-10-0097
[6] Lim, S., Lee, J., Kyonggi-do, Shin, D. and Lim, H.S.
(1999) Comparison of protein extraction solutions for rice
starch isolation and effects of residual protein content on
starch pasting properties. Starch-Stȁrke, 51, 120-125.
http://dx.doi.org/10.1002/(SICI)1521-379X(199904)51:4
<120::AID-STAR120>3.0.CO;2-A
http://dx.doi.org/10.1002/(SICI)1521-379X(199904)51:4
<120::AID-STAR120>3.3.CO;2-1
[7] Radosavljevic, M., Jane, J. and Johnson, L.A. (1997)
Isolation of amaranth starch by diluted alkaline-protease
treatment. Cereal Chemistry, 75, 212-216.
http://dx.doi.org/10.1094/CCHEM.1998.75.2.212
[8] Mistry, A.H., Schmidt, S.J. and Eckhoff, S.R. (1992) Al-
kali extraction of starch from corn flour. Starch-Stärc h,
44, 284-288. http://dx.doi.org/10.1002/star.19920440803
[9] Matsunaga, N. and Seib, P.A. (1997) Extraction of wheat
starch with a queous sodium hydroxide. Cereal Chemistry,
74, 851-857.
http://dx.doi.org/10.1094/CCHEM.1997.74.6.851
[10] Hedge, J.E. and Hofreiter, B.T. (1962) Carbohydrate
chemistry 17. Whistler, R.L. and Be Miller, J. N., Eds.,
Academic Press, New York.