Journal of the Taiwan Institute of Chemical Engineers 66 (2016) 210–216

Contents lists available at ScienceDirect

Journal of the Taiwan Institute of Chemical Engineers

journal homepage: www.elsevier.com/locate/jtice

Kinetic models for batch and continuous ethanol fermentation from

sweet sorghum juice by yeast immobilized on sweet sorghum stalks
Pongthep Ariyajaroenwong a, Pattana Laopaiboon b,c, Apilak Salakkam b,
Penjit Srinophakun d, Lakkana Laopaiboon b,c,∗
a
Graduate School, Khon Kaen University, Khon Kaen 40002, Thailand
b
Department of Biotechnology, Faculty of Technology, Khon Kaen University, Khon Kaen 40002, Thailand
c
Fermentation Research Center for Value-Added Agricultural Products, Khon Kaen University, Khon Kaen 40002, Thailand
d
Department of Chemical Engineering, Faculty of Engineering, Kasetsart University, Bangkok 10900, Thailand

a r t i c l e i n f o a b s t r a c t

Article history: Kinetic models for batch and continuous ethanol fermentation from sweet sorghum juice by Saccha-
Received 13 January 2016 romyces cerevisiae NP 01 immobilized on unpeeled sweet sorghum stalk pieces were developed. The
Revised 26 April 2016
models accounted for substrate limitation, substrate inhibition, ethanol inhibition and cell death. Batch
Accepted 21 June 2016
ethanol fermentations were done from juice containing various initial sugar concentrations (120–280 g/L).
Available online 4 July 2016
The estimated values of the maximum specific growth rate (μmax ) and Monod constant (Ks ) were found
Keywords: to be 0.313 h−1 and 47.51 g/L, respectively, using a Lineweaver–Burk plot. These data were used to develop
Kinetic models models for batch and continuous ethanol fermentation. For the batch fermentation, it was found that the
Immobilized yeast models could be used to satisfactorily fit the experimental data for initial sugar concentrations ranging
Sweet sorghum from 130 to 225 g/L. However, for the continuous fermentation, only the data for substrate consumption
Batch ethanol fermentation and ethanol production were well fitted by the developed models.
Continuous ethanol fermentation
© 2016 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1. Introduction
Due to the world’s volatile energy market and environmental
concerns, alternative fuels such as bioethanol have received much
attention as potential replacements for fossil fuels [1]. In Thailand,
a government 15-year plan (2008–2022) is in place to increase
bioethanol production capacity to 9 million liters day−1 by 2022
[2]. Currently, the most widely used substrates for bioethanol pro-
duction in Thailand are cassava and sugarcane molasses. Increasing
the production of bioethanol will eventually result in shortages of
these materials. It is therefore necessary to identify promising al-
ternative materials when these substrates are fully utilized. Sweet
sorghum (Sorghum biocolor (L.) Moench) is potentially such a sub-
strate. It yields high amounts of biomass and sugar. The stalk of
sweet sorghum contains large amounts of soluble sugars (glucose
and sucrose) and insoluble carbohydrates (holocellulose). Also, its
juice contains many trace elements that are essential for microbial
growth and ethanol production [3].
In many studies, batch, repeated-batch, fed-batch and con-
tinuous fermentation processes were used to produce ethanol
∗
Corresponding author at: Department of Biotechnology, Faculty of Technology,
Khon Kaen University, Khon Kaen, Thailand. Tel./fax: +66 4336 2121.
E-mail address: lakcha@kku.ac.th (L. Laopaiboon).
from sugars by free yeast cells [4–6]. However, the disadvan-
tages of using free cells for ethanol fermentation are substrate and
product inhibition [7], extra time needed for inoculum prepara-
tion and cleaning the reactor between batches resulting in longer
turnover times. To overcome these limitations, cell immobiliza-
tion was introduced in fermentation processes. Cell immobiliza-
tion is the limitation of cell mobility by isolation within a car-
rier. Commercially available materials, e.g., alginate and carrageen,
are widely used for this purpose. However, these materials are
costly and much research has been directed at finding other low
cost alternative natural materials. These include sorghum bagasse
[8], sugarcane pieces [9], corn cobs [10], thin-shell silk cocoons
[11] and sweet sorghum stalks [12].
Study of fermentation kinetic parameters is important for
understanding the impacts of environmental factors on ethanol
production. These factors include temperature [13] and substrate
concentration [14]. Moreover, kinetic parameters coupled with
mathematical models can be used to predict the dynamics of cell
concentration, substrate utilization and ethanol production rate
[15,16]. In many cases, optimal conditions for product formation
can be predicted using mathematical models without experimen-
tation [17]. However, preliminary studies revealed that most re-
search on kinetic and mathematical models for ethanol production

http://dx.doi.org/10.1016/j.jtice.2016.06.023
1876-1070/© 2016 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
 P. Ariyajaroenwong et al. / Journal of the Taiwan Institute of Chemical Engineers 66 (2016) 210–216 211

2.2. Raw material

Nomenclature
Sweet sorghum juice extracted from its stalks (cv. KKU 40) was
D dilution rate, h−1 obtained from the Division of Agronomy, Faculty of Agriculture,
KIP substrate inhibition constant for ethanol formation, Khon Kaen University, Thailand. The juice initially had total soluble
g/L solids of 17ºBrix. Then it was concentrated to 65ºBrix and stored at
KIS substrate inhibition constant for growth, g/L 4 °C until use.
KS Monod constant for growth, g/L
KSP ethanol saturation constant, g/L
2.3. Ethanol production medium
m cell maintenance coefficient, h−1
PE ethanol concentration, g/L
Ethanol production (EP) medium was prepared by diluting the
Pi initial ethanol concentration, g/L
concentrated juice with distilled water to total sugar concentra-
PP. max maximum ethanol concentration for ethanol
tions of 120–280 g/L. Then, it was supplemented with 6 g/L of yeast
fermentation, g/L
extract and autoclaved at 110 °C for 28 min [3]. The medium was
PX. max maximum ethanol concentration for growth, g/L
left at room temperature to cool prior to use.
QP ethanol productivity, g/L h
qP specific ethanol production rate, g/g h
qmax maximum specific ethanol production rate, g/g h 2.4. Cell immobilization on sweet sorghum stalk
S substrate concentration, g/L
SC sugar consumption (%) Cell immobilization was conducted by adsorption of cells on
Si initial substrate concentration, g/L sterile sweet sorghum stalk (SSS) pieces (6–20 mm in diameter
t time, h and 6 mm in thickness). SSS pieces were transferred into sweet
t0 incubation time at the commencement of the log sorghum juice containing 100 g/L of total sugar with active yeast
phase, h cells at a concentration of ∼1 × 108 cells/mL and incubated at 30 °C
tL incubation time at the end of the log phase, h for 18 h. After that, the SSS pieces were washed with sterile EP
V working volume, L medium before use as an inoculum (immobilized cells) for ethanol
X total cell concentration, g/L production [23].
X0 cell concentration at the commencement of the log
phase, g/L 2.5. Batch fermentation
XL cell concentration at the end of the log phase, g/L
Xf free cell concentration in broth, g/L The SSS pieces containing immobilized yeast cells at 50% of
Xi initial total cell concentration, g/L working volume (350 mL) were transferred into sterile EP medium
Xim cell concentration in carrier, g/L in a 500-mL air-locked Erlenmeyer flask. The fermentation was car-
YP/S ethanol yield, g/g ried out at 30 °C for 72 h under static conditions to prevent the de-
YX/S biomass yield, g/g tachment of the immobilized cells from the carriers. Samples were
yA actual data taken for analysis at regular time intervals.
yav average actual data
yP predicted data
μ specific growth rate, h−1 2.6. Continuous fermentation
μmax maximum specific growth rate, h−1
α, β ethanol inhibition constant, g/L The immobilized yeast cells were packed into a single-tubular
packed bed bioreactor (working volume of 0.78 L) at 50% of the
column height [23]. The process was started as a batch fermenta-
focused on free cell systems [13,18,19]. Little information is avail- tion at 30 °C. When the sugar concentration in the broth decreased
able on the ethanol production by immobilized cells [15,20,21]. to approximately 20% of its initial value, a continuous system was
The aim of this research was to study the kinetics of ethanol started by feeding sterile sweet sorghum juice into the bottom of
production from sweet sorghum juice using yeast immobilized on the bioreactor at a dilution rate of 0.023 h−1 . During the fermenta-
sweet sorghum stalk pieces under batch and continuous fermenta- tion, samples were taken for analyses.
tion. Then mathematical models were developed to predict ethanol
production. For this purpose, a modified Monod’s equation was 2.7. Analytical methods
used to account for substrate and product inhibition.
Ten grams of SSS pieces containing immobilized yeast cells
2. Materials and methods were blended with 90 mL of a 0.85% NaCl solution [24]. The sus-
pension was subjected to serial dilution. Cell concentration in the
2.1. Microorganism and inoculum preparation suspension was determined by a direct counting method using a
hemacytometer and a methylene blue staining technique. The dry
S. cerevisiae NP 01 was isolated from a dried starter culture weight of viable cells and the SSS pieces were measured using a
for making Thai rice wine [22]. It was inoculated into a 250-mL gravimetric method [23]. Correlation between viable cell concen-
Erlenmeyer flask containing 150 mL of yeast extract malt extract tration and dry cell weight were determined by linear regression.
(YM) medium [12]. The flask was incubated on a rotating shaker Then, the viable cells in the SSS pieces were determined in terms
at 200 rpm, 30 °C for 18 h. A 10% inoculum (v/v) of the culture was of cells/g dry weight of SSS pieces. Sugar, ethanol (PE ) and viable
added into 350 mL of sweet sorghum juice containing 100 g/L of to- cell concentrations were determined, and the batch ethanol pro-
tal sugar to yield an initial cell concentration of ∼5 × 106 cells/mL. duction efficiencies in terms of ethanol yield (YP/S ) and ethanol
After being further incubated for 18 h, the cells were harvested by productivity (QP ) were calculated as described by Laopaiboon et al.
centrifugation at 60 0 0 rpm for 10 min, and used for cell immobi- [3]. Additionally, the QP of continuous ethanol fermentation was
lization. estimated by the of P value multiplied by its dilution rate (D).
212 P. Ariyajaroenwong et al. / Journal of the Taiwan Institute of Chemical Engineers 66 (2016) 210–216

2.8. Theories and kinetic models 2.8.2. Continuous models

In the continuous ethanol fermentation equations, the dilution
2.8.1. Batch models rate (D) parameter was added into the biomass, product and sub-
Mathematical models were developed to determine the quan- strate models as shown in Eqs. (10), (11) and (12), respectively.
titative interaction of environmental parameters and cell kinetics.
dX
In general, the Monod equation (Eq. 1) is used to describe the re- For biomass: = DXi + (μ − D )X (10)
lationship between growth rate and substrate concentration. How- dt
ever, this model does not account for inhibition caused by high
substrate and product concentrations. dPE
For product: = DPi − DPE + qP X (11)
μmax S dt
μ= (1)
KS + S
 
During ethanol fermentation, the primary inhibitory factors are dS 1 dX 1 dPE
high concentrations of sugar and ethanol. Both inhibitors affect For substrate: − = + + mX
dt YX/S dt YP/S dt
yeast growth in a non-competitive inhibitory manner [25]. One
model that includes terms for substrate and product inhibition is − D ( Si − S ) (12)
the Andrew and Levenspiel model (Eq. 2).
Substituting Eqs. (2) and (3) into Eqs. (10) and (11), respectively,
 
μmax S  PE
α models for continuous ethanol fermentation were:
μ= 1− (2)   
KS + S + S2
KIS
PX. max
dX μmax S  PE
α
For biomass: = DXi − X 2
1− −D
Ethanol is a primary metabolite of yeast growth under anaer- dt KS + S+ KSIS PX. max
obic conditions and its formation is associated with cell growth. (13)
Using the same philosophy based on the Andrew and Levenspiel
equation, a product formation model was proposed as given in
Eq. (3).  
   β
qmax S
 PE
β For product:
dPE
= D(Pi − PE ) + X
qmax S
1−
PE
dt S2 PP. max
qP = 1− (3) KSP +S + KIP
S2 PP. max
KSP + S + KIP
(14)
In the case of cell immobilization by an adsorption technique,
most yeast cells are immobilized on the surface of the carriers,
while the rest remain suspended in the broth. For this reason, cell 2.9. Determinations of the maximum specific growth rate (μmax ) and
concentration (X) is calculated in terms of total cell concentration Monod constant for growth (KS )
which includes cell concentrations both in the carrier and in the
broth (Eq. 4). Batch ethanol fermentations from the sweet sorghum juice con-
   taining different initial sugar concentrations (120–280 g/L) were
(Xim × gcarrier ) + X f × V carried out using the immobilized cells on SSS pieces. Samples
X= (4) were collected at 3 h intervals. The viable cell concentrations both
V
in the carriers and the broth were determined and used for μ de-
To describe fermentation dynamics, the rates of cell growth and termination using Eq. (15). The μ values were used to evaluate
ethanol production are as follows. μmax and KS by Lineweaver–Burk plots.
dX 
For biomass : = μX (5) ln XL − ln X0
dt μ= (15)
tL − t0
dPE
For product : = qP X (6)
dt 2.10. Calculation of kinetic parameters
Substituting Eq. (2) into (5) and (3) into (6) gives Eqs. (7) and
(8) for biomass and product formation, respectively, as follows. Other kinetic parameters (KIS , KIP , KSP , PX. max , PP. max , YX/S , α ,
  β , qP and m) were determined using the non-linear least squares
dX μmax S  PE
α method of Marquardt [17]. Simultaneous numerical integration of
For biomass : =X S2
1− (7) the model equations was done using the ODE23 routine of MATLAB
dt KS + S + PX. max
KIS (Version 7.0) [17]. The accuracy of the models was evaluated as a
function of their coefficients of determination (R2 ). Well fit models
 
dPE qmax S
 PE
β have R2 values close to 1.
For product : =X S2
1− (8)
dt KSP + S + PP. max
KIP
2.11. Simulation of continuous ethanol fermentation
In batch ethanol fermentation, substrate is used for cell growth
and maintenance as well as for ethanol production [17]. The sub- Eqs. (12), (13) and (14), with suitable kinetic parameters, were
strate utilization rate equation can be written as shown in Eq. (9). used for calculation by the ODE 45 routine [17] with the initial val-
ues of biomass, sugar and ethanol concentration used in the exper-
 
dS 1 dX 1 dPE iments. The fermentation was started in batch mode (zero dilution
− = + + mX (9) rate). After 36 h, the continuous fermentation simulation was run
dt YX/S dt YP/S dt
at various dilution rates (0.01–0.15 h−1 ).
 P. Ariyajaroenwong et al. / Journal of the Taiwan Institute of Chemical Engineers 66 (2016) 210–216 213

.30 Table 1
Kinetic parameters of batch ethanol fermentation
.28
Specific growth rate (h -1)
from sweet sorghum juice by S. cerevisiae NP 01 im-
.26 mobilized on sweet sorghum stalk pieces compared
with those of other studies.
.24
Parameters This study Other studies
.22
.20 PX. max , g/L 83.35 112 [32], 107 [33]
PP. max , g/L 107.79 125 [19]
.18 KSP , g/L 28.39 20.016 [17]
.16 KIS , g/L 308.13 290.003 [17]
KIP , g/L 299.67 366.7 [19]
.14 m, h−1 0.001 0.031 [17]
.12 qmax , g/g h 3.69 1.9157 [29]
α 1.53 3.68 [19]
.10 B 1.53 1.72 [19]
120 160 200 240 280 YX/S , g/g 0.48 0.50 [28], 0.235 [17]
-1
Sugar concentration (g l )
Table 2
Fig. 1. Specific growth rates (μ) of S. cerevisiae at different initial sugar concentra- Applicability of the models, in terms of coefficient of determination (R2 )
tions. at different initial sugar concentrations in batch fermentation.

Initial sugar concentrations (g/L) Models

3. Results and discussion Biomass Product Substrate

130 0.9074 0.9881 0.9907

3.1. μmax and KS of batch ethanol fermentation 170 0.9509 0.9916 0.9908
220 0.9957 0.9889 0.9981
The μ values of the batch ethanol fermentation from sweet 225 0.9809 0.9934 0.9972
sorghum juice (120-280 g/L of total sugar) by the immobilized 260 0.5385 0.9826 0.9818
yeast cells increased from 0.224 to 0.278 h−1 when sugar concen-
tration was increased from 120 to 240 g/L (Fig. 1). Similar results
were reported by Laopaiboon et al. [26], who studied batch ethanol respectively, and these values were lower than S2 (Eqs. (7) and
fermentation from yeast extract malt extract (YM) broth by free (8)), indicating that substrate at higher concentrations could inhibit
cells of S. cerevisiae NP 01. They found that the μ values increased both cell growth and ethanol formation. The value of the cell main-
when glucose concentration in YM broth was increased from 10 to tenance coefficient, m, was fixed at 0.001 h−1 because this value
240 g/L. However, a higher sugar concentration of 280 g/L gave a had a little effect on model prediction of cell growth [28]. The
lower μ value (0.219 h−1 ). This might be due to osmotic stress or specific ethanol production rate (qmax ) was 3.69 g/g h, which was
substrate inhibition [27]. higher than that (1.92 g/g h) reported by Jin et al. [29] using the
The μ values were used to calculate μmax and KS using same raw material. This might be due to the higher sugar concen-
a Lineweaver–Burk plot (Fig. 2). It was found that μmax was trations used in our experiment.
0.313 h−1 , which was in the range (0.259–0.463 h−1 ) reported in The value of parameters in Table 1 gave excellent fit with high
other studies using S. cerevisiae and sucrose–glucose based sub- coefficients of determination (R2 > 0.98) (Fig. 3). By substituting
strates [17,18,28,29]. KS was 47.51 g/L, which was higher than these values in Eqs. (7), (8) and (9), Eqs. (16), (17) and (18) were
those of an ethanol fermentation by immobilized yeast cells re- obtained.
ported by Wöhrer [15] (KS = 1.86 g/L) and Manikandan et al.   1.53
[30] (KS = 25 g/L). The differences in KS values may depend on fac- dX 0.313S PE
For biomass: =X 1−
tors such as types and concentration of substrates, yeast concentra- dt 47.51 + S + S2 83.35
308.13
tions and strains or fermentation processes which affected growth
(16)
mechanisms [31]. The μmax and KS values were used to develop
mathematical models in the subsequent experiments.   1.53
dPE 3.69S PE
For product: =X 1−
3.2. Kinetic parameters and mathematical models for batch dt 28.39 + S + S2 107.79
299.67
fermentation
(17)

In batch fermentation at 225 g/L of total sugar, it was found that  

the Xf value was very low (2.58 × 10−5 g/L), whereas the Xim value dS 1 dX 1 dPE
For substrate: − = + + 0.001X
was significantly higher at 0.3660 g/L. For this reason, the Xim value dt 0.48 dt 0.50 dt
was considered to be Xi . The results from the batch fermentation
(18)
(Fig. 3) showed that P, QP and YP/S values were 101.56 g/L, 1.41 g/L
h, and 0.50 g/g, respectively. The data from this experiment were Eqs. (16), (17) and (18) were validated at other initial sugar con-
used to estimate the other parameters in Eqs. (7), (8) and (9) (KIS , centrations ranging from 130 to 260 g/L to confirm the applicability
KIP , KSP , PX .max , PP. max , YX/S , α , β , qmax and m) using a least squares of the models. The initial values in each experiment, i.e. biomass,
method. These results are shown in Table 1. The maximum ethanol sugar and ethanol concentrations were input into the software, and
concentration for growth (PX .max ) was 83.35 g/L, which was close parameter predictions were generated to compare with the exper-
to those of other reports [32,33]. The maximum ethanol concentra- imental data. Except for biomass at the initial sugar concentra-
tion (PP. max ) and ethanol saturation constant (KSP ) for ethanol fer- tion of 260 g/L, all the R2 values exceeded 0.75 (lowest acceptable
mentation were 107.79 g/L and 28.39 g/L, respectively, which were value) (Table 2), reflecting the models’ performance [34]. The low
similar to those reported by Ge and Bai [19] and Thangprompan R2 value at 260 g/L of total sugar may be due to higher osmotic
et al. [17]. The KIS and KIP parameters were 308.13 and 299.67 g/L, stress to cell growth in the system than the base model (225 g/L)
214 P. Ariyajaroenwong et al. / Journal of the Taiwan Institute of Chemical Engineers 66 (2016) 210–216

Fig. 2. Lineweaver–Burk plot estimating μmax and KS values in batch ethanol fermentation.

A very well with the R2 for biomass, product and substrate of 0.8611,
6 0.9941 and 0.9872, respectively (data not shown).
Additionally, in spite of the differences in substrate and initial
Cell concentration (g l-1)

5 substrate concentration, the models were found to describe satis-

factorily the data extracted from literatures (R2 greater than 0.81,
4
data not shown). In doing this, the data from the systems with
3 sugarcane molasses (131 g/L) [35] and sugar beet juice (136 g/L)
[36] were used. This helped confirm the usability of the models,
2
constants or kinetic parameters obtained in this work.
1

0
R2= 0.9809 3.3. Continuous models determination
B 0 20 40 60 80
250 Continuous ethanol fermentation was carried out on the sweet
Sugar concentration (g l-1)

sorghum juice containing 225 g/L of total sugar by the immobilized

200 yeast cells in the single-tubular packed-bed bioreactor at the di-
lution rate of 0.023 h−1 . The continuous fermentation system was
150 started after 36 h of batch operation (Fig. 4). The P, QP and YP/S
values were 71.42 ± 1.55 g/L, 1.64 ± 0.01 g/L h and 0.50 ± 0.01 g/g,
100
respectively. The kinetic parameter values obtained from batch ex-
periments were used to construct continuous models for biomass
50 R2= 0.9972 (Eq. 19), product (Eq. 20) and substrate (Eq. 21) at a dilution rate
of 0.023 h−1 . The results showed that the models for substrate
0
C 0 20 40 60 80
consumption and ethanol production fit the data reasonably well,
120 as they gave high R2 values of 0.9322 and 0.9574, respectively
Ethanol concentration (g l-1)

(Fig. 4B and C). However, a low R2 (0.5362) was observed for the
100 biomass model (Fig. 4A). This might have been due to some non-
80 homogeneity in the system as no agitation was used during the
fermentation and/or the adsorption of cells on the carriers [37]. To
60

40
. confirm this, the total cell concentrations were determined at the
end of the fermentation. The results showed that the cell concen-
tration at this time was 1.76 ± 0.04 g dry cells/L which was close to
20 the predicted value (2.16 g dry cells/L) (Fig. 4A). This could be the
R2= 0.9934 reason why the prediction of sugar and ethanol profiles were very
0 good (R2 > 0.93).
0 20 40 60 80 
Time (h) dX 0.313S
For biomass: = DXi − X S2
Fig. 3. Cell growth (A), sugar consumption (B), and ethanol production (C) in batch dt 47.51 + S + 308.13
ethanol fermentation (dotted lines), fitted with the models developed in this study  1.53
(solid lines). PE
× 1− −D (19)
83.35


which might have affected some kinetic parameters in the mod- dPE 3.69S
els [17]. Nevertheless, the results demonstrate that the models are For product: = D(Pi − PE ) + X S2
dt 28.39 + S +
applicable at a wide range of sugar concentrations with high ac- 299.67
 PE
1.53
curacy. When the models were tested to fit the results of our
× 1− (20)
previous work [23], it was found that the models fitted the data 107.79
 P. Ariyajaroenwong et al. / Journal of the Taiwan Institute of Chemical Engineers 66 (2016) 210–216 215

A A
5 5
concentration (g dried cells l )
-1

concentration (g dry cells l )

-1
4
4
Total cell

Total cell
3
2

1
2

R2 = 0.5362
0 1
B 0 50 100 150 200 250 300
250
0
Sugar concentration (g l-1)

200 B 0 50 100 150 200 250 300

250

Sugar concentration (g l )
150

-1
200
100

50 150
2
R = 0.9322
0
C 100
0 50 100 150 200 250 300
100
Ethanol concentration (g l-1)

50
80

60 0
C 0 50 100 150 200 250 300
40 100
Ethanol concentration (g l )
-1

20 80
R2 = 0.9574
0
0 50 100 150 200 250 300 60
Time (h)

Fig. 4. Cell growth (A), sugar consumption (B), and ethanol production (C) in con- 40
tinuous ethanol fermentation (dotted lines), fitted with the constructed models
(solid lines).
20

0
0 50 100 150 200 250 300
  Time (h)
dS 1 dX 1 dPE
For substrate: − = +
dt 0.48 dt 0.50 dt Fig. 5. Simulation of continuous ethanol fermentation profiles of biomass (A), sub-
strate (B) and ethanol (C) concentrations at different dilution rates; 0.01 h−1 (solid
+ 0.001X − D(Si − S ) (21)
lines), 0.02 h−1 (long dashed lines), 0.05 h−1 (short dashed lines), 0.10 h−1 (dash-dot
lines) and 0.15 h−1 (dotted lines).

Apart from predicting, the proposed models can be of use in

optimizing the process. In silico simulation of the process carried
out at different dilution rates was conducted. The results shown
in Fig. 5 demonstrate that the dilution factor has a great effect 4. Conclusions
on ethanol production as increasing the dilution rate resulted in
lower ethanol production and this eventually led to wash out phe- Kinetic models for ethanol fermentation by adsorbed immobi-
nomenon. Of all dilution rate tested, the highest PE of 68 g/L was lized yeast were successfully developed. By accounting for sub-
obtained at the dilution rate of 0.01 h−1 , and wash out occurred at strate and product inhibition, the batch models were applicable at
high dilution rates above 0.10 h−1 . Thangphrompan et al. [17] stud- a wide range of initial sugar concentrations (130–225 g/L). The data
ied the effects of dilution rates, initial sugar concentrations and for biomass, sugar consumption and ethanol production could be
continuous start-up times using computer simulation, and they fitted with very high accuracy (R2 > 0.90). In continuous fermenta-
found that these models could be used to describe the effects of tions, the developed models could be used to describe sugar con-
these parameters very well. The use of process simulation can be sumption and ethanol production very well (R2 > 0.93). To the best
beneficial for the design of ethanol production processes as it can of our knowledge, this is the first report of the development of
be used to describe, and better still predict the performance of the mathematical models for fuel ethanol fermentation using adsorbed
system with less labor and expense. immobilized yeast.
216 P. Ariyajaroenwong et al. / Journal of the Taiwan Institute of Chemical Engineers 66 (2016) 210–216
Acknowledgments
This research was financially supported by The Royal Golden Ju-
bilee Ph.D Program (Grant No. PHD/0042/2554), Thailand Research
Fund (TRF), Fermentation Research Center for Value-Added Agri-
cultural Products (FerVAAP) and Center for Alternative Energy Re-
search and Development, Khon Kaen University (KKU). We would
like to thank Asst. Prof. Dr. Paiboon Danviruthai, Faculty of Tech-
nology, KKU for providing the NP 01 strain, Assoc. Prof. Dr. Pra-
sit Jaisil, Faculty of Agriculture, KKU for providing sweet sorghum
stalk and its juice and Mr. Preuk Thangprompan for his sugges-
tions.
References
[1] Sánchez ÓJ, Cardona CA. Trends in biotechnological production of fuel ethanol
from different feedstocks. Bioresour Technol 2008;99:5270–95.
[2] Silalertruksa T, Gheewala SH. Security of feedstocks supply for future
bio-ethanol production in Thailand. Energy Policy 2010;38:7476–86.
[3] Laopaiboon L, Nuanpeng S, Srinophakun P, Klanrit P, Laopaiboon P. Ethanol
production from sweet sorghum juice using very high gravity technol-
ogy: effects of carbon and nitrogen supplementation. Bioresour Technol
20 09;10 0:4176–82.
[4] Alfenore S, Cemeleyre X, Benbadis L, Bideaux C, Uribelarrea JL, Goma G,
et al. Aeration strategy: a need for very high ethanol performance
in Saccharomyces cerevisiae fed-batch process. Appl Microbiol Biotechnol
2004;63:537–42.
[5] Tang YQ, An MZ, Zhong YL, Shigeru M, Wu X-L, Kida K. Continuous ethanol
fermentation from non-sulfuric acid-washed molasses using traditional
stirred tank reactor and the flocculating yeast strain KF7. J. Biosci Bioeng
2010;109:41–6.
[6] Ariyajarearnwong P, Laopaiboon P, Jaisil P, Laopaiboon L. Repeated-batch
ethanol fermentation from sweet sorghum juice by free cells of Saccharomyces
NP 01. Afr J Biotechnol 2011;10:13909–18.
[7] Pátková J, Šmogrovičová D, Dömény Z, Bafrncová P. Very high gravity wort
fermentation by immobilized yeast. Biotechnol Lett 20 0 0;22:1173–7.
[8] Yu J, Zhang X, Tan T. An novel immobilization method of Saccha-
romyces cerevisiae to sorghum bagasse for ethanol production. J Biotechnol
2007;129:415–20.
[9] Liang L, Zhang YP, Zhang L, Zhu MJ, Liang SZ, Huang YN. Study of sugarcane
pieces as yeast supports for ethanol production from sugarcane juice and mo-
lasses. J Ind Microbiol Biot 2008;35:1605–13.
[10] Laopaiboon L, Laopaiboon P. Ethanol production from sweet sorghum juice in
repeated-batch fermentation by Saccharomyces cerevisiae immobilized on corn-
cob. World J Microbiol Biotechnol 2012;28:559–66.
[11] Rattanapan A, Limtong S, Phisalaphong M. Ethanol production by repeated
batch and continuous fermentations of blackstrap molasses using immobilized
yeast cells on thin-shell silk cocoons. Appl Energy 2011;88:4400–4.
[12] Ariyajaroenwong P, Laopaiboon P, Jaisil P, Laopaiboon L. Repeated-batch
ethanol production from sweet sorghum juice by Saccharomyces cerevisiae im-
mobilized on sweet sorghum stalks. Energies 2012;5:1215–28.
[13] Phisalaphong M, Srirattana N, Tanthapanichakoon W. Mathematical modeling
to investigate temperature effect on kinetic parameters of ethanol fermenta-
tion. Biochem Eng J 2006;28:36–43.
[14] Aguilar-Uscanga MG, Garcia-Alvarado Y, Gomez-Rodriguez J, Phister T,
Delia ML, Strehaiano P. Modelling the growth and ethanol production of Bret-
tanamyces bruxellensis at different glucose concentrations. Lett Appl Microbiol
2011;53:141–9.
[15] Wöhrer W. A horizontal bioreactor for ethanol production by immobilized
cells Part 3: Reactor modeling and experimental verification. Bioprocess Eng
1989;4:105–11.
[16] Ghaly AE, El-Taweel AA. Kinetic modeling of continuous production of ethanol
from cheese whey. Biomass Bioenergy 1997;12:461–72.
[17] Thangprompan P, Thanapimmetha A, Saisriyoot M, Laopaiboon L,
Srinophakun P. Production of ethanol from sweet sorghum juice us-
ing VHG technology: a simulation case study. Appl Biochem Biotechnol
2013;171:294–314.
[18] Chen LJ, Xu YL, Bai FW, Anderson WA, Moo-Young M. Observed quasi-steady
kinetics of yeast cell growth and ethanol formation under very high gravity
fermentation condition. Biotechnol Bioproc Eng 2005;10:115–21.
[19] Ge XM, Bai FW. Intrinsic kinetics of continuous growth and ethanol production
of a flocculating fusant yeast strain SPSC01. J Biotechnol 2006;124:363–72.
[20] Kobayashi F, Nakamura Y. Mathematical model of direct ethanol produc-
tion from starch in immobilized recombinant yeast culture. Biochem Eng J
2004;21:93–101.
[21] Kandylis P, Mantzari A, Koutinas AA, Kookos IK. Modelling of low temperature
wine-making, using immobilized cells. Food Chem 2012;133:1341–8.
[22] Deesuth O, Laopaiboon P, Klanrit P, Laopaiboon L. Improvement of ethanol
production from sweet sorghum juice under high gravity and very high grav-
ity conditions: effects of nutrient supplementation and aeration. Ind Crop Prod
2015;74:95–102.
[23] Ariyajaroenwong P, Laopaiboon P, Laopaiboon L. Capability of sweet sorghum
stalks as supporting materials for yeast immobilization to produce ethanol un-
der various fermentation processes. J Taiwan Inst Chem Eng 2015;49:79–84.
[24] Plessas S, Bekatorou A. Use of Saccharomyces cerevisiae cells immobilized
on orange peel as biocatalyst for alcoholic fermentation. Bioresour Technol
2007;98:860–5.
[25] Aiba S, Shoda M, Nagatani M. Kinetics of product inhibition in alcohol fermen-
tation. Biotechnol Bioeng 20 0 0;67:671–90.
[26] Laopaiboon L, Nuanpeng S, Srinophakun P, Klanrit P, Laopaiboon P. Selection
of Saccharomyces cerevisiae and investigation of its performance for very high
gravity ethanol fermentation. Biotechnol 2008;7:493–8.
[27] Ingledew WM. Alcohol production by Saccharomyces cerevisiae: A yeast primer.
The Alcohol Textbook. UK: Nottingham University Press; 1999.
[28] Oliveira SC, De Castro HF, Visconti AES, Giudici R. Continuous ethanol fer-
mentation in a tower reactor with flocculating yeast recycle: scale-up effects
on process performance, kinetic parameters and model predictions. Bioprocess
Eng 1999;20:525–30.
[29] Jin H, Liu R, He Y. Kinetics of batch fermentations for ethanol production with
immobilized Saccharomyces cerevisiae growing on sweet sorghum stalk juice.
Proc Environ Sci 2012;12:137–45.
[30] Manikandan K, Saravanan V, Viruthagiri T. Kinetics studies on ethanol produc-
tion from banana peel waste using mutant strain of Saccharomyces cerevisiae.
Indian J Biotechnol 2008;7:83–8.
[31] Felix E, Clara O, Vincent AO. A kinetic study of the fermentation of cane sugar
using Saccharomyces cerevisiae. Open J Phys Chem 2014;4:26–31.
[32] Luong JHT. Kinetics of ethanol inhibition in alcohol fermentation. Biotechnol
Bioeng 1985;27:280–5.
[33] Thatipamala R, Rohani S, Hill GA. Effects of high product and substrate inhi-
bitions on the kinetics and biomass and product yields during ethanol batch
fermentation. Biotechnol Bioeng 1992;40:289–97.
[34] Hu R. Food product design: a computer-aided statistical approach. Pennsylva-
nia: Technomic publishing Co. Ltd; 1999.
[35] de Andrade RR, Rivera EC, Atala DIP, Filho RM, Filho FM, Costa AC. Study of
kinetic parameters in a mechanistic model for bioethanol production through a
screening technique and optimization. Bioprocess Biosyst Eng 2009;32:673–80.
[36] Dodić JM, Vučurović DG, Dodić SN, Grahovac JA, Popov SD, Nedeljkovic NM.
Kinetic modelling of batch ethanol production from sugar beet raw juice. Appl
Energy 2012;99:192–7.
[37] Chen Y, Liu Q, Zhou T, Li B, Yao S, Li A, et al. Ethanol production by repeated–
batch and continuous fermentations by Saccharomyces cerevisiae immobilized
in a fibrous bed bioreactor. J Microbiol Biotechnol 2013;23:511–17.