Documente Academic
Documente Profesional
Documente Cultură
Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech
h i g h l i g h t s g r a p h i c a l a b s t r a c t
a r t i c l e i n f o a b s t r a c t
Article history: A novel integrated immobilized enzyme-reactor system involving a continuous stirred tank reactor with
Received 11 June 2013 two packed bed reactors in series was developed for the continuous production of biodiesel. The problem
Received in revised form 31 July 2013 of methanol solubility into oil was solved by introducing a stirred tank reactor to dissolve methanol into
Accepted 5 August 2013
partially converted oil. This step made the process perfectly continuous without requiring any organic
Available online 12 August 2013
solvent and intermittent methanol addition in the process. The substrate feeding rate of 0.74 mL/min
and enzyme loading of 0.75 g per reactor were determined to be optimum for maximum biodiesel yield.
Keywords:
The integrated continuous process was stable up to 45 cycles with biodiesel productivity of 137.2 g/L/h,
Biodiesel
Green continuous process
which was approximately 5 times higher than solvent free batch process. In comparison with the pro-
Integrated CSTR-PBR system cesses reported in literature using expensive Novozyme 435 and hazardous organic solvent, the present
Immobilized lipase process is completely green and perfectly continuous with economic and environmental advantages.
Solvent free system Ó 2013 Elsevier Ltd. All rights reserved.
0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.biortech.2013.08.023
396 S. Chattopadhyay, R. Sen / Bioresource Technology 147 (2013) 395–400
known as Novozyme 435, was used for most of the reported liter- separate glycerol completely. Glycerol layer was collected from
atures (Severac et al., 2011; Sakai et al., 2010; Halim et al., 2009). bottom and upper biodiesel layer was analyzed.
This immobilized proparation showed efficient transesterification
property in wide variety of lipid sources. The cost of the Novozyme 2.2.3. Quantitative estimation of the product
435 is significantly high and that makes the overall process expen- The product was analyzed by high performance thin layer chro-
sive. Some relatively inexpensive immobilized system is needed to matography (HPTLC) method. The detail description of the method
be developed to produce biodiesel in economically feasible way. To was described elsewhere (Chattopadhyay et al., 2011b). In brief,
avoid methanol inhibition in solvent free process, a three step different glycerides were separated by thin layer chromatography
methanolysis was described (Lee et al., 2011). In the first step, 1/ (TLC) using hexane, ethyl acetate and acetic acid (90:10:1) as mo-
3 of the stoichiometric amount of methanol was added into the bile phase. The TLC plate was scanned at 203 nm under HPTLC
oil and the product produced after reaction was kept in separator scanner and the percentage yield was calculated from the peak
for the complete separation of glycerol from partially converted areas of different glycerides obtained from HPTLC chromatogram
oil. The next step was performed by adding another 1/3 of the stoi- (Chattopadhyay et al., 2011b).
chiometric amount of methanol into the partially converted oil and
the same process was continued for the third step. The stepwise
2.2.4. Systematic development of a green continuous process for
processes described in the literatures were not truly continuous biodiesel
and sometimes involved environmentally harmful organic solvents
2.2.4.1. Continuous process development in a single packed bed
and expensive enzyme Novozyme 435 in their systems. reactor. To eliminate methanol inhibition problem, t-butanol was
The present study thus aimed at producing biodiesel in a green,
used in the continuous system. The IIT-SARKZYME was used in a
solvent free and economical way using a less expensive immobi- single packed bed reactor (PBR) (inner diameter 2 cm, height
lized lipase, IIT-SARKZYME, developed by Chattopadhyay and Sen
45 cm, volume 140 mL) with a bed volume of 100 mL. Reaction
(2012). For that purpose, an integrated continuous stirred tank mixture containing oil, t-butanol and methanol was fed continu-
reactor with two packed bed reactors (CSTR-PBR) in series was
ously into the PBR from bottom through a peristaltic pump with
used first time for the simultaneous reaction-cum-solubilization a flow rate of 1 mL/min. The enzyme loading was kept as one gram.
of the immiscible phases in the solvent free system for the im-
Reactor temperature was maintained at 37 °C by water circulation
proved conversion in PBR. The stability of the integrated process through jacket. The reactor set up comprises single PBR is
was determined and its efficacy for different feedstock with vari-
presented in Fig. 1a. The product was collected in a separating fun-
ous free fatty acids (FFA) was checked to find out the robustness nel. The upper biodiesel layer was distilled at 84 °C for 30 min to
of the system.
separate solvent and unutilized methanol from the product. The
lower layer comprises mostly glycerol was collected from bottom
2. Methods into the separating funnel. Some important reactor parameters like
void fraction (e) and packing density were determined. The small
2.1. Materials used void fraction ensured better contact of enzyme to substrate and
useful to calculate the residence time of the reaction. The packing
Oils were procured from N.K. Proteins Ltd. (Nagpur, India). density determines the flow characteristics of the substrates
Crude Steapsin (pancreatic lipase Ex-micro organism), and methyl through the PBR. The parameter values were determined using fol-
linoleate standard were purchased from SRL (Mumbai, India), and lowing equations (Vikbjerg et al., 2005).
the standard is chromatographically pure. HPTLC plates, SDS, and
Void fractionðeÞ ¼ VS =V ð1Þ
other chemicals were obtained from MERCK (Germany), and were
of analytical grade. Bradford protein estimation kit was purchased
from Bangalore Genie (Bangalore, India).
Packing density ¼ M=VR ð2Þ
(a) (b)
WB WB
L L L SF
SF BD
BD
G
G
J
J
S
S
P P
Fig. 1. Continuous biodiesel production in (a) single PBR and (b) two PBRs connected in series. S: substrate mixture, P: pump, J: water circulating jacket, L: immobilized lipase,
WB: water bath, PBR: packed bed reactor, D: distillation unit, SF: separating funnel, BD: biodiesel, G: glycerol layer.
focus was to produce biodiesel with higher productivity in solvent in Fig. 2. The final product was separated in separating funnel and
free system. analyzed.
Stability of IIT-SARKZYME in continuous process was deter-
mined by calculating the percentage conversion values in each cy-
cle. The total residence time for the two PBRs in series was
2.2.4.3. Development of a green continuous process based on an calculated and was considered as one cycle.
integrated CSTR-PBR system for solvent free biodiesel produc-
tion. With a view to developing a completely green process, a
2.2.5. Mass balance and product productivity
CSTR-PBR based integrated enzymatic reactor system employing
The mass balance of the reactant and the product was calcu-
our indigenously formulated immobilized enzyme (IIT-SARKZYME)
lated from their volume and specific gravity values. The oil to
was designed for continuous biodiesel production in solvent free
methanol molar ratio was determined by considering an average
system (Patent Application No: 404/KOL/2011). In the integrated
molecular weight of oil.
system, oil and methanol was mixed (1:1 molar ratio) and allowed
The productivity (P) of the entire process at steady state for the
to react for a small period of time (approximately 4 h) in batch
production of 100 mL product was calculated by the following
mode, which partially converted oil to its methyl ester. The en-
equations:
zyme loading in the batch process was kept as 1% (w/v of oil).
The reactor was then operated in continuous mode (as CSTR) with Overall dilution rateðDÞ ¼ F=ðn VÞ ð4Þ
this partially converted oil and fresh feed, where a small fraction of
fresh oil was converted into biodiesel. This initial step involving
ProductivityðPÞ ¼ C D 60 ð5Þ
partial conversion increased the solubility of methanol in oil and
in turn, eliminated the inhibitory effect of methanol as well as where, D is the dilution rate (per min), F is the volumetric flow rate
the requirement of organic solvent. The mixture was then fed into (mL/min), n is the number of reactors used, V is the volume of each
two PBRs as described previously. The reaction scheme is depicted reactor (mL), C is the product concentration (g/L) and P is the pro-
ductivity (g/L/h).
Different feedstock of varying FFA compositions like palm, sun-
WB flower, soybean, rice bran and waste cooking oil along with cotton-
seed oil were transesterified at same conditions using the
integrated continuous process to check the robustness of the
PBR D process.
PBR
Immobilized Cost Activity of Amount of Total cost (Rs) The total enzyme loading was varied from 0.5 to 2 g and was
lipase (Rs/g of immobilized immobilized equally distributed into two packed bed reactors connected in ser-
lipase) lipase (U/g) lipase with ies. It was observed that higher than 1.5 g of enzyme did not able to
same activity (g)
improve percentage conversion further (Fig. 4b). With increase in
IIT-SARKZYME 50 3557 1 50 enzyme amount, the interaction between enzyme and substrates
Novozyme 435 984 10,000 0.36 358
increased. But after that the enzyme-substrate reaction reached
its saturation level and hence, no additional increment in percent-
analysis, it is clear that IIT-SARKYME as an immobilized enzyme age conversion was found. Similar enzyme loading of 2 and 1 g and
system is approximately 7.2 times cheaper than Novozyme 435. was reported to be used to achieve maximum percentage conver-
sion (Hajar et al., 2009; Royon et al., 2007). The enzyme loading re-
3.2. Batch reaction ported in the various recent literatures is summarized in Table 2.
As cost of the enzyme is the major concern in enzymatic biodiesel
As compared to single step process with maximum conversion production, minimum amount (1.5 g) that was required for effec-
of 20% after 48 h, the percentage conversion in two step process tive transesterification was used for further experiments.
was estimated as 40% after 10 h of reaction (with oil to methanol
molar ratio 1:1). With further methanol addition at 10 h, the con- 3.5. Determination of important reactor parameters
version increased and reached to a saturation level of 72% after
24 h (Fig. 3). The inhibitory effect of methanol in solvent free single From Eqs. (1)–(3), different reactor parameters were calculated.
step process was eliminated by step wise methanol addition at The void fraction was found to be 0.28 whereas packing density
10 h, which helped to solubilize methanol into oil and in turn en- was 0.88 g/mL. The packing density was calculated and compared
hanced percentage conversion (Hajar et al., 2009). with literature reports (Table 3). From the table it is clear that
the packing density was found to be higher as compared to most
3.3. Standardization of substrate flow rate of the literatures but comparable with some (Rosa et al., 2009;
Shaw et al., 2008). Small void fraction with relatively higher pack-
The percentage conversion of oil into biodiesel increased with ing density increases enzyme–substrate interaction and reduces
increasing flow rate, which may be due to the consequence of high channelling effect inside PBRs. The residence time at flow rate of
rate of mass transfer. It was observed that, percentage conversion 0.74 mL/min was calculated as 53 min. As void fraction was not de-
reached its maxima at a flow rate of 0.74 mL/min, which was con- scribed by the available literature reports, comparison of residence
sidered as optimal for further reaction (Fig. 4a). After optimum time was not possible with our value.
flow rate, where mass transfer rate reached its maxima, the per-
centage conversion decreased, which might be due to the small 3.6. Salient features of the integrated continuous process
residence time for the reaction (Lee et al., 2010). Some earlier re-
ports exemplified such trend of drastic reduction of methyl ester 3.6.1. Productivity
yield at higher flow rates (Chang et al., 2009; Shaw et al., 2008). It was observed that oil was partially converted around 25% in
The optimum flow rates for maximum percentage conversion, as CSTR after 4 h reaction. This partially converted oil was fed into
reported in literature, are summarized in Table 2. In most of the two PBRs connected in series. After the system achieved steady
cases, very low flow rates ranging from 0.1 to 0.5 mL/min were state, the overall productivity of the integrated process was
used to get maximum conversion (Rosa et al., 2009; Wang et al., 137.2 g/L/h, which was higher than most of the reported literature
2011). However, a process with very high flow rate of 6 mL/min values (Wang et al., 2011; Hajar et al., 2009; Shaw et al., 2008). The
was also reported (Hajar et al., 2009), wherein the productivity comparative analysis of the different continuous processes is sum-
was less as the product was recycled for 72 h before collecting marized in Table 4. Though some processes indicated higher pro-
from the reactor. ductivity than ours (Rosa et al., 2009; Halim et al., 2009), use of
organic solvent and expensive enzyme made those processes eco-
80 nomically and environmentally unfavourable. Experimental data
indicated that the productivity of the integrated process was
1 step
70 approximately five times higher than the solvent free batch pro-
2 step cess (26.1 g/L/h). The productivity of the batch processes was less
60 due to the higher reaction time by stepwise methanol addition into
the system. The integrated process may prove to be economically
% conversion
30
3.6.2. Mass balance
The overall mass balance of the entire process was calculated
20
as:
10 100 g oil þ 56:5 g methanol ! 103:5 g BD þ 7:8 g glycerol
0 þ 45:2 g methanol
0 10 20 30 40 50 60
Having considered 100 g oil as the basis, a mass balance of the
Time (h)
transesterification process was carried out by using the actual
Fig. 3. Percentage conversion of oil to biodiesel in single step and two steps batch mass values of the substrates and the products. The average molec-
process. The methanol addition point at 10 h is indicated by an arrow. ular weight of oil was calculated to be 850 g/mole. As indicated by
S. Chattopadhyay, R. Sen / Bioresource Technology 147 (2013) 395–400 399
80 (a) 80 (b)
70 0.74 70
60 1 60
0.33
% conversion
% conversion
50 1.78 50
40 0.12 40
30 30
20 20
10 10
0 0
0 0.5 1 1.5 2 0.25 0.75 1.25 1.75 2.25
Flow rate (ml/min) Amount of enzyme (g)
Fig. 4. Effect of (a) substrate flow rate and (b) different enzyme loading (in two columns) on transesterification.
Table 2 Table 4
Comparative analysis of optimum flow rates and enzyme loading. Comparative analysis of different continuous processes.
Flow rate (mL/min) Enzyme loading (g) Reference Effective Productivity Remarks on the process Reference
reactor (g/L/h)
0.25 10 Wang et al. (2011)
volume
1.50 3.0 Salum et al. (2010)
(mL)
6.00 2.0 Hajar et al. (2009)
0.50 22.5 Rosa et al. (2009) 5 79.3 Low reaction rate, use of Royon et al. (2007)
0.60 2.8 Halim et al. (2009) organic solvent
0.10 NA Chang et al. (2009) 1.2 40.7 Low reaction rate, use of Shaw et al. (2008)
0.10 1.0 Shaw et al. (2008) organic solvent
0.16 1.0 Royon et al. (2007) 18 9.5 Three step methanolysis, Hajar et al. (2009)
0.74 1.5 This work recycled up to 72 h
21 247.8 High enzyme loading and Rosa et al. (2009)
NA: Data not available use of organic solvent
8 151.5 Use of organic solvent and Halim et al. (2009)
expensive Novozyme 435
23 82.5 Two step methanolysis, Salum et al. (2010)
Table 3 methanol addition at 7 h
Comparison of packing density with different reactor volume. 40 28.7 High reaction time, use of Wang et al. (2011)
organic solvent
Reactor Immobilized Packing Reference 100 137.2 High reaction rate, perfectly This work
volume (mL) enzyme (g) density continuous, no organic solvent,
(g/mL) green, no intermittent methanol
40.2 10 0.25 Wang et al. (2011) addition
22.7 3 0.13 Salum et al. (2010)
17.7 2 0.11 Hajar et al. (2009)
21.0 22.5 1.07 Rosa et al. (2009)
8.3 2.8 0.34 Halim et al. (2009)
1.2 1 0.83 Shaw et al. (2008)
5.1 1 0.20 Royon et al. (2007) 80
140 123 0.88 Our work
70
60
initial optimization experiments (Chattopadhyay et al., 2011a),
50
% conversion
100
Acknowledgements
80
% relative yield