close

Вход

Забыли?

вход по аккаунту

?

Biodiesel production through microwave assisted transesterification of microbial cells

код для вставкиСкачать
BIODIESEL PRODUCTION THROUGH MICROWAVE ASSISTED
TRANSESTERIFICATION OF MICROBIAL CELLS
by
Yi Cui
B.E., Xian University of Science & Technology, P.R.China, 2002
M.S., Northwest University, P.R.China, 2007
A Dissertation
Submitted in Partial Fulfillment of the Requirements for the
the Doctor of Philosophy Degree
Department of Civil & Environmental Engineering
Southern Illinois University CarbondaleAugust 2013
UMI Number: 3604348
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMI 3604348
Published by ProQuest LLC (2013). Copyright in the Dissertation held by the Author.
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
DISSERTATION APPROVAL
BIODIESEL PRODUCTION THROUGH MICROWAVE ASSISTED
TRANSESTERIFICATION OF MICROBIAL CELLS
By
Yi Cui
A Dissertation Submitted in Partial
Fulfillment of the Requirements
for the Degree of
Doctor of Philosophy Degree
In the field of Engineering Science
Approved by:
Yanna Liang, Chair
Ruplal Choudhary
Bruce DeVantier
Xingmao Ma
Manoj Mohanty
Graduate School
Southern Illinois University Carbondale
May 8th, 2013
AN ABSTRACT OF THE DISSERTATION OF
Yi Cui, for the Doctor of Philosophy in ENGINEERING SCIENCE, presented on
May 8th ,2013, at Southern Illinois University Carbondale
Title: BIODIESEL PRODUCTION THROUGH MICROWAVE ASSISTED
TRANSESTERIFICATION OF MICROBIAL CELLS
MAJOR PROFESSOR: DR. YANNA LIANG
One strain of oleaginous yeasts, Cryptococcus curvatus (ATCC 20509) has been studied to
grow on several substrates including
biodiesel production byproduct crude glycerol and sweet
sorghum juice. After cultivation, yeast cells were heated under microwave radiation to extract
lipid and produce biodiesel through in-situ transesterification.
Firstly, the yeast growth with crude glycerol was studied. When cultured in a one-stage fedbatch process wherein crude glycerol and nitrogen source were fed intermittently for 12 days, the
final biomass density and lipid content were 31.2 g/L and 44.2%, respectively. When cultured in
a two-stage fed-batch operation wherein crude glycerol was supplemented at different time points
while nitrogen source addition was discontinued at the middle of the experiment, the biomass
density was 32.9 g/L and the lipid content was 52% at the end of 12 days.
On the second step, an optimization of yeast fermentation with crude glycerol was
conducted. Through Box–Behnken design and response surface methodology, the optimal
temperature, pH, and glycerol concentration for yeast growth on pretreated crude glycerol was
identified as 30.2 ◦C, 6.0, and 19.8 g/L, respectively. Adopting these optimal parameters, the
i
biomass density and lipid content obtained were 7.11 ± 0.36 g/L and 38.53 ± 1.88%, respectively,
which matched well with the model predicted values of 6.98 g/L and 41.31%.The resulting
parameters of the response surface method optimization were used in a fed-batch fermentation
where crude glycerol was automatically pumped in responding to exhausted oxygen levels in the
fermentor. At the end of 12 days, the biomass density and lipid content were 44.53 g/L and
49%,respectively. Compared with our fed-batch experiment which was conducted under
un-optimized condition, the yield of biomass and lipid increased 35.26% and 25.29%.
When cultured in a fed batch process where sorghum juice was supplemented at different
time points for 3 days, the final biomass density was 23.6 g/L with a lipid content of 51%. To
extract lipids from cells in an effective and fast fashion, a domestic microwave oven was used
with different solvents. With only methanol, a lipid yield of 33.2% of yeast cells was obtained in
4 min. This was comparable with a lipid content of 51% attained through using a traditional
solvent extraction approach.
In the end, to convert yeast lipids to biodiesel directly without the step of lipid extraction,
the in-situ transesterification method used microwave irradiation on the simultaneous extraction
and transesterification of wet yeast biomass to biodiesel. Response surface methodology was
used to analyze the influence of the process variables (solvent to biomass (v:w) ratio, catalyst
concentration, and reaction time) on the fatty acid methyl ester conversion. Based on the
experimental results and RSM analysis, the optimal conditions for this process were determined
as: methanol to yeast biomass (v:w) ratio of around 50:1, catalyst concentration about 5 wt.%,
and reaction time of 2 min. The biodiesel samples were analyzed with GC and the FAME content
in biodiesel is about 50%.
ii
ACKNOWLEDGEMENTS
I would like to thank my advisor Dr. Yanna Liang deeply for her guidance, encouragement
and financial support. I want to acknowledge Drs. James Blackburn, Ruplal Choudhary, Bruce
Devantier, Xingmao Ma, Manoj Mohanty for serving on my committee and Mr. John Hester as
our technician. I would also like to thank my former and current group mates for their kind help.
My dissertation is dedicated to my family, who always stand behind me and support me.
iii
TABLE OF CONTENTS
CHAPTER
PAGE
ABSTRACT..................................................................................................................................... i
ACKNOWLEDGMENTS…………………………………………………………… …..………iii
LIST OF TABLES . ……………………………………………………………………………..viii
LIST OF FIGURES........................................................................................................................ ix
CHAPTER 1
CONVERTING CRUDE GLYCEROL DERIVED FROM YELLOW GREASE
TO LIPIDS THROUGH YEAST FERMENTATION……………….………………..…… ……1
1.1 Introduction…………………………………………………………………….……… ……1
1.2 Materials and Methods…… …………..……………… ……………………..……… ……. 3
1.2.1 Yeast Culture………………………….……………………………………… ……….3
1.2.2 Crude Glycerol Characterization…….………………………………. …… ………….3
1.2.3 Batch Stage Study with Crude Glycerol….…….…… ……… …………… ………….4
1.2.4 Fed-batch Study with Crude Glycerol…….……………..………………… ………….5
1.2.5 Analysis………. …. …. …. …. …. …. …. ……. …. …. …. …. …. …. … . ………..6
1.3 Result and Discussion……… ………………………..….………………………… ……….8
1.4 Conlusion…………………………………………………...…………………….. ……….19
CHAPTER 2
FERMENTATION OPTIMIZATION FOR THE PRODUCTION OF LIPID BY
CRYPTOCOCCUS CURVATUS:
USE OF RESPONSE SURFACE METHODOLOGY..................................................................20
iv
2.1 Introduction………………………………………………………………………………20
2.2 Materials and Methods………..…………………………………….………………….22
2.2.1 Yeast Culture……………………… ……………………………...……………….22
2.2.2 Crude Glycerol Pretreatment…………………..….………………. ………………22
2.2.3 Experimental Design..............................................................................................23
2.2.4 Optimized Fed-batch Fermentation Experiment…………..…...…..……………….24
2.2.5 Analysis…………………………………………..…………………………………26
2.3 Result and Discussion……………………………………………………………………27
2.3.1 Development of A Regression Model………………………… ……..…….. ..…….27
2.3.2 Effects of Process Parameters on Optimization…… …....……… …...…….………29
2.3.3 The Influence of Optimal Parameters on Fed-batch Fermentation………………….34
2.4 Conlusion………………………………………………...……………………..……….40
CHAPTER 3
CULTIVATION AND LIPID PRODUCTION OF YEAST CRYPTOCOCCUS
CURVATUS USING SWEET SORGHUM SYRUP .....................................................................41
3.1 Introduction………………………………………………………………………………41
3.2 Materials and Methods…………………… ..………… …………………..…………….43
3.2.1 Source of Sweet Sorghum Syrup………….….……………….…...……………….43
3.2.2 Microorganism and Inoculum Preparation……...……………...….………..………43
3.2.3 Batch Stage Growth with Sweet Sorghum Syrup….…….........................................44
3.2.4 Fed-batch Growth with Sweet Sorghum Syrup…….….……...…..….…………….44
3.2.5 Analysis…………………………………………..…………………………………45
3.2.6. Microwave Assisted Lipid Extraction……………………......…………………….46
v
3.3 Result and Discussion……………………………………………………………………47
3.3.1 C. curvatus Growth on Sorghum Syrup……….…….………………..…………….47
3.3.2 Microwave Assisted Lipid Extraction………………….... ………………..…....54
3.4 Conlusion…………………………………….…………...……………………..……….56
CHAPTER 4
IN-SITU DIRECT TRANSESTERIFICATION OF THE YEAST
CRYPTOCOCCUS CURVATUS LIPID TO BIODIESEL BY USING MICROWAVE
RADIATION..................................................................................................................................58
4.1 Introduction…………..……….…………………………………………………………58
4.2 Materials and Methods………………..……………… …………………..…… ………60
4.2.1 Yeast culture and biomass preparation………….………………...……………….60
4.2.2 Transesterification Process……………….…..…...………………. ………………60
4.2.3 Screening Test Experimental Design….................................... ...............................62
4.2.4 Optimization Experimental Design………….………………...…..……………….63
4.2.5 FAME Analysis……………….……………………………………………………64
4.3 Result and Discussion……………………………………………………………………65
4.3.1 Significant Parameters Screening by Plackett–Burman Design………….….…….65
4.3.2 Development of A Regression Model…… ………………..… …………………67
4.3.3 Effects of Process Parameters on Optimization………… ……..………………….68
4.3.4 Analysis of Yeast Biodiesel…………..…………………..……………………..73
4.4 Conclusion………………………………………………..……………………..……….75
vi
CHAPTER 5 FUTURE RESEARCH............................................................................................76
5.1 Conversion to Bio-oil or Syngas……..…..……………………………………………….76
5.2 Fermentation and Transesterification Scale up…..……….………………………………76
REFERENCES ..............................................................................................................................78
vii
LIST OF TABLES
TABLE
PAGE
Table 1-1 Comparison among oleaginous species on crude glycerol ............................................15
Table 1-2 Yield calculation............................................................................................................16
Table 1-3 Methanol mass balance..................................................................................................17
Table 1-4 Fatty acid comparison with others.................................................................................18
Table 2-1 Process variables and their levels used in the design ....................................................24
Table 2-2 The Box-Behnken design of the variables with biomass yield and lipid content as
responses ......................................................................................................................26
Table 2-3 Analysis of variance (ANOVA) of experimental data...................................................29
Table 2-4 Illustration of calculations for crude glycerol utilization
and biomass yield.........................................................................................................38
Table 3-1 Microwave assisted lipid extraction yields by using different solvents ........................56
Table 4-1 Variables and their levels employed in the Plackett–Burman design............................62
Table 4-2 Plackett–Burman experimental design matrix and the response values........................63
Table 4-3 Process variables and their levels used in the design ....................................................64
Table 4-4 The Box-Behnken design of the variables with biodiesel
yield as response ..........................................................................................................68
Table 4-5 Analysis of variance (ANOVA) ....................................................................................72
Table 4-6 GC peak of crude biodiesel obtained from yeast biomass.............................................74
viii
LIST OF FIGURES
FIGURE
PAGE
Figure 1-1 With crude glycerol at different concentrations, C. curvatus biomass
changing with time.....................................................................................................10
Figure 1-2 During fed-batch 1 which was one-stage, biomass density and glycerol
concentration changing with time ..............................................................................11
Figure 1-3 During fed-batch 2 which contained two stages, biomass density and glycerol
concentration changing with time. .............................................................................12
Figure 1-4 Methanol concentration changing with time for the two
fed-batch processes ....................................................................................................13
Figure 1-5 Major fatty acids changing with time
during fed-batch 1 ........................................................................................................19
Figure 2-1a Three-dimensional response surface plot of biomass yield and lipid
content as a function of different parameters.............................................................31
Figure 2-1b Three-dimensional response surface plot of lipid productivity as a function
of different parameters ...............................................................................................32
Figure 2-2a During the fed-batch experiment, biomass density and glycerol concentration
changing with time.....................................................................................................35
Figure 2-2b Glycerol concentration changing with time during the
fed-batch experiment .................................................................................................36
Figure 2-3 Comparison of biomass density changing with time
between two fed-batch studies .....................................................................................37
ix
Figure 3-1 With sweet sorghum juice sugars at different concentrations, C. curvatus biomass
changing with time.....................................................................................................48
Figure 3-2 C. curvatus total sugars utilization for different juice concentrations .........................49
Figure 3-3 C. curvatus sugar utilization profiles for different juice concentrations ......................50
Figure 3-4 During fed-batch, biomass density and sugars concentration
changing with time.....................................................................................................53
Figure 3-5 During fed-batch, fructose, glucose and sucrose concentration
changing with time.....................................................................................................53
Figure 4-1 The statistical plots identifying the key operating parameters for
biodiesel yield ............................................................................................................66
Figure 4-2 Three-dimensional response surface plot of biodiesel yield as a function of different
parameters ..................................................................................................................70
x
1
CHAPTER 1
CONVERTING CRUDE GLYCEROL DERIVED FROM YELLOW GREASE TO
LIPIDS THROUGH YEAST FERMENTATION
1.1. Introduction
During recent decades, oils derived from microorganisms, which are often
referred to as microbial oils or single cell oils, have emerged as alternatives for
biodiesel production. Microbes such as yeasts, microalgae, fungi, and bacteria have
been shown to possess oil-producing capabilities. Compared to traditional vegetable
and oilseed sources, production of microbial oils offers many advantages: shorter life
cycle, less labor required, less affected by venue, season and climate, and easier to
scale up (Li and Wang, 1997; Li et al., 2008). Therefore, microbial oils may become
one of the potential oil feedstocks for biodiesel production in the future.
One yeast species, Cryptococcus curvatus is of special interest as an oil producer.
Firstly, this yeast only requires minimal nutrients for growth. Secondly, it can
accumulate up to 60% of its cellular dry weight (DW) as intracellular lipid (Hassan et
al., 1996). Thirdly, the accumulated oil resembles plant seed oils such as Palm oil in
terms of fatty acid composition (Davies, 1988) and is mainly triglyceride (80- 90%)
(Ratledge and Cohen, 2008; Ykema et al., 1988). Fourthly, it grows on a broad range
of substrates including, but not limited to: 1) sugars, such as glucose, xylose,
galactose, mannose, fructose, ribose, maltose, cellobiose, sucrose, and lactose (Glatz
et al., 1984); 2) glycerol (Meesters et al., 1996); and 3) whey concentrate or permeate
2
(Daniel et al., 1999; Moon et al., 1978; Ykema et al., 1988).
As previously mentioned, the prospect of producing biofuels from microbial
feedstocks is attractive.
To be economically viable, however, the high substrate
expenses associated with the heterotrophic growth mode need to be addressed to bring
down overall production expenditures.
Ideally, microbial oil production should
focus on substrates that have zero or negative costs. Crude glycerol, a by-product of
the biodiesel production line, may be an excellent match in this context.
Crude glycerol, produced in 1:10 ratio to biodiesel during production, is a
mixture of glycerol (65-85%, w/w), methanol, and soap (Gonzalez-Pajuelo et al.,
2005; Mu et al., 2006). Presently, in various European countries, crude glycerol is
simply treated as a new kind of industrial wastewater owing to the obligatory
requirement for biodiesel production and the huge excess that results (Mu et al., 2008).
Crude glycerol has substantial commercial value if purified to United States
Pharmacopoeia grade. However, the cost to refine this product is approximately
$0.20/lb (Chi et al., 2007). For small and medium-sized biodiesel operations, this
process is cost-prohibitive (Haas et al., 2005).
Considering the challenges that soy bean oil faces — unstable prices and
competition with feed- and food-grade end uses, biodiesel feedstocks such as recycled
restaurant oil (yellow grease) and animal fats (white grease) are on the rise. However,
crude glycerol streams derived from these two sources are less marketable due to the
presence of color, odor contaminants as well as other minor contaminating
compounds that are most costly to refine (Tyson et al., 2004). However, if those
3
low-value crude glycerol streams could be reintegrated as substrates for microbial oil
production, the overall productivity of biodiesel production from yellow or white
grease could be enhanced.
Therefore, in this study, we aim to examine the growth
and lipid yield of C. curvatus on crude glycerol derived from yellow grease-based
biodiesel production in batch and fed-batch modes.
1.2. Materials and Methods
1.2.1. Yeast culture
C. curvatus (American Type Culture Collection (ATCC) 20509) grown in liquid
medium containing 2% peptone, 1% yeast extract, and 16 g/L of pure glycerol was
utilized as inocula for batch and fed-batch experiments adopting a minimal medium
(Meesters et al., 1996). The minimal medium contained (per liter): 2.7 g KH2PO4;
0.95 g Na2HPO4; 0.2 g MgSO4 ·7H2O; 0.1 g yeast extract; and 0.1 g EDTA. After the
pH was adjusted to 5.5, it was supplemented with a 100x spores stock solution
consisting of (per liter): 4 g CaCl2 ·2H2O; 0.55 g FeSO4.7H2O; 0.52 g citric acid; 0.10
g ZnSO4.7H2O; 0.076 g MnSO4.H2O; and 100 µL 18 M H2SO4.
1.2.2. Crude glycerol characterization
A crude glycerol sample from BioVantage (Belvidere, IL, USA) was used as the
experimental substrate in this study. This sample was generated from biodiesel
production using yellow grease as the feedstock. The crude glycerol had a pH of 11.2
resulting from the alkali-catalyzed transestrification reaction occurring during the
4
biodiesel production process.
Without any methanol recovery, the crude glycerol
appeared as a dark brown liquid.
Composition analysis of this sample was conducted as described previously
(Liang et al., 2010a). Briefly, the original sample was diluted 10-fold and passed
through 0.2 µm filter. Glycerol and methanol concentrations were determined by
HPLC, which is detailed below. To estimate soap content, pH of the sample was
adjusted to 1.0. Then the whole content was centrifuged at 4,000 g for 20 min. The
top, dark red layer was free fatty acids (FFAs). This layer was removed and weighed
for calculating soap content.
1.2.3. Batch stage study with crude glycerol
To prepare the crude glycerol for use in culture experiments, the sample pH was
lowered to 1.0. After FFAs were removed, the remaining sample was adjusted to a pH
of 7.0 and autoclaved. This sample was referred to as treated stock. Based on the
glycerol concentration in this stock, different volumes of this sample were added to C.
curvatus cultures to yield final glycerol concentrations of 20, 40, 60, and 80 g/L.
Nitrogen was supplied in the form of NH4Cl at 2.5 g/L. The inoculum size was 10%
of the total volume. All cultures were incubated at 30 oC with shaking at 150 rpm.
After 72 h, a 50 mL sample from each culture was taken, centrifuged, washed with
distilled and deionized water (DDW) twice, and freeze-dried to obtain biomass DW.
5
1.2.4. Fed-batch study with crude glycerol
Two fed-batch experiments were conducted using a 2-liter fermentor with a
bladed stirrer (New Brunswick, Edison, New Jersey, USA). The fermentor was
autoclaved at 121oC for 45 min after 1.0 liter of minimal medium was filled in and pH
and DO probes were connected and calibrated. The treated glycerol stock was then
supplemented together with the inoculum. For both of the experiments, pH was
maintained at 5.5 by automatic pumping of NaOH. Temperature was controlled at 28
o
C. Incoming air was filtered through 0.2 µm filter. The whole culture volume was
maintained between 1.5 and 1.6 liters. When needed, silicone antifoam emulsion (J. T.
Baker, Phillipsburg, NJ, USA) was added to prevent excess foam formation.
For the first fed-batch fermentation, the starting concentrations of glycerol and
NH4Cl were 25.8 and 1.25 g/L, respectively. The C/N ratio (g/g) was 30.
To keep
this constant C/N ratio, different volumes of treated stock, NH4Cl, and medium were
added to the fermentor at different time points, such as 3, 6.25, 7, 9.1, and 10 days.
Starting from day 3, the air flow rate was increased from 0.8 to 1.0 l/min. In addition,
the stirring speed was changed to 900 rpm from the initial 720 rpm on day 5. On a
daily basis, a 100 mL sample was withdrawn from the bioreactor and processed for
biomass DW, glycerol concentration, and other measurements as described below.
The culture growth was sustained for 12 days.
For the second fed-batch investigation, the beginning glycerol concentration was
32 g/L.
As before, the initial C/N ratio was 30. This C/N ratio was maintained for
the first 6 days, after which the nitrogen addition was terminated while the glycerol
6
and medium supplements continued. Air flow rate and stirring speed were kept at 1.0
l/min and 900 rpm, respectively. Daily subsamples were collected and the growth
period lasted 12 days.
Samples harvested at different times were centrifuged first at 4,000×g for 10 min.
The supernatant was filtered through 0.2 µm nylon filter for HPLC determination of
glycerol and methanol concentrations. The pellets were then washed with DDW twice
and freeze-dried overnight to get biomass DW. The dried biomass was reserved for
cellular lipid and fatty acid analyses.
1.2.5. Analysis
Glycerol and methanol concentrations were determined by high-performance
liquid chromatography (HPLC) (Shimadzu Scientific Instrument, Inc. Columbia, MD,
USA) with a refractive index detector. A Supelcogel AG1 column (5 µm, 30 cm×4.6
mm, Supelco, Bellefonte, PA, USA) was used in an oven set at 83 oC. HPLC grade
water was used as the mobile phase with a flow rate as 0.5 mL/min. The injection
volume was 10 µL. Concentrations of glycerol and methanol were calculated based on
calibration curves built for these two compounds using external standards.
Cellular lipid content was determined following a procedure developed in our
laboratory (Liang et al., 2010b). Briefly, 0.5 g dried cell pellet was transferred to a
7-mL chamber of a bead-beater (BioSpec Products, Bartlesville, OK, USA). This
chamber was filled with 0.5 mm zirconium beads to approximately 5 mL. Methanol
was then added to fill the rest of the chamber. After cells were disrupted by
7
bead-beating for 2 min, the entire content was transferred to a 50-mL glass centrifuge
tube. The chamber was washed twice using methanol (total 10 mL) to collect the
yeast residue, and the washes were added to the primary methanolic extract.
Chloroform was then added to make a 2:1 (v/v) chloroform/methanol ratio (v/v). The
tube was vortexed for 5 min and was allowed to stand for 24 h. Afterwards, the tube
was centrifuged at 4,000 × g for 15 min to remove the zirconium beads and yeast
solids. The supernatant was collected and the solvent was vaporized using rotovap.
The resultant crude lipid was weighed to calculate oil content.
Crude lipid samples were subjected to acid-catalyzed transmethylation
performed overnight at 50 ºC as described previously by Christie (1982). The
resultant fatty acid methyl esters (FAME) were separated using a Shimadzu GC-17A
gas chromatograph (Shimadzu Scientific Instruments) equipped with a flame
ionization detector (FID) fitted with a permanently bonded polyethylene glycol, fused
silica capillary column (Omegawax 250, 30 m×0.25 mm i.d., 0.25 µm film). The
injection volume was 1.0 µL, helium was the carrier gas (30 cm/s, 205 ºC), and the
injector temperature was 250 °C. A split injection technique (100:1) was used, and the
temperature program was as follows: 50 °C held for 2 min, increased to 220 °C at
4°C/min, and held at 220 °C for 15 min. Individual FAME was identified by reference
to external standards (Supelco 37 Component FAME Mix, PUFA-1, and PUFA-3;
Supelco, Bellefonte, PA, USA). All solvents used were of HPLC grade and obtained
from Sigma Diagnostics Inc. (St. Louis, MO, USA).
8
1.3. Results and Discussion
To evaluate whether C. curvatus can grow on crude glycerol, we first examined
the growth of this yeast on pure glycerol and have confirmed Meesters’ study
(Meesters et al., 1996) that: 1) C. curvatus can utilize glycerol and 2) glycerol at high
concentration suppresses cell growth. A C/N (g/g) ratio of 30 was determined to be
optimal for cell growth and lipid production, which was in agreement with those
reported (Hassan et al., 1996; Meesters et al., 1996).
The crude glycerol sample that we used in this study was composed of glycerol,
soap, methanol, and water in the percentages of 48.7, 3.0, 22.7, and 25.6%,
respectively. Using treated stock solution, we have observed different biomass
densities with different glycerol concentrations. Apparently, high substrate
concentrations inhibited cell growth (Fig. 1-1). Within 72 h, basically there was no
biomass increase for glycerol concentrations of 60 and 80 g/L. The lowest glycerol
dose of 20 g/L produced 2.5 times more biomass compared to that of 40 g/L.
Comparing the biomass densities between crude glycerol at 20 g/L and pure glycerol
at 16 g/L (data not shown), crude glycerol resulted in 67% of that from pure glycerol.
Therefore, C. curvatus can utilize glycerol in the crude samples, but the biomass
productivity was inferior to that achieved with pure glycerol. Additionally, the impact
of high substrate concentration was significantly greater in crude glycerol treatments
compared to pure glycerol treatments. As reported by Meesters et al. (1996), glycerol
concentrations from 8-64 g/L have little effect on growth rate. However, with crude
glycerol, the biomass produced was significantly less within concentration range of
9
20-80 g/L, between 60 and 80 g/L with which little growth was observed.
Greater
suppressive effects we observed using crude glycerol may also be associated with
impurities in the crude glycerol samples.
Considering substrate inhibition, we employed a fed-batch process to avoid high
glycerol concentration while to improve the biomass productivity. As indicated by Fig.
1-2, the one-stage fed-batch fermentation generated a biomass DW of 31.2 g/L in 12
days. At day 3, the biomass density was 9 g/L, which was similar to that from pure
glycerol concentration at 16 g/L at batch mode. However, due to continuous supply of
glycerol, the growth continued. During the whole experimental period, the average
biomass productivity was 2.6 g/L-day which was half of the maximum biomass
productivity 5.7 g/L-day taking place from day 5 to day 8. Increased air flow rate and
stirring speed might contribute to the increased growth rate starting from day 5. In
terms of glycerol utilization, during the initial 24 h, glycerol utilization rate was low
accompanied by the lag phase cell growth. But after the cells acclimated to crude
glycerol, the substrate uptake rate was faster than we expected. The maximum
substrate utilization rate was 19.2 g/L-day taking place during day 7 and 8.
10
Fig. 1-1. With crude glycerol at different concentrations, C. curvatus biomass
changing with time. ♦: 20 g/L; ■: 40 g/L; ▲: 60 g/L; and ×: 80 g/L.
11
Fig.1-2. During fed-batch 1 which was one-stage, biomass density (♦) and glycerol
concentration (■) changing with time. Solid arrows indicate addition of both
treated stock and NH4Cl. Triangle indicates stirring speed increase from 720 to
900 rpm.
In terms of the second fed-batch, more cells (3 g/L) were added to the fermentor
at the beginning. Thus, the maximum glycerol uptake rate (Fig. 1-3) was higher than
that of the first fed-batch run, 28.3 versus 19.2 g/L-day and it took place earlierbetween day 1 and day 2. But the profile of glycerol uptake was similar between these
two fed-batch processes. After day 6, even when nitrogen was not added to the culture
any more, glycerol was still being utilized and the biomass kept increasing though not
too much. At the end of the 12 days, the biomass density was 32.9 g/L, which was a
12
little higher than that from the first fed-batch process.
Fig. 1-3. During fed-batch 2 which contained two stages, biomass density (♦) and
glycerol concentration (■) changing with time. Solid arrows indicate addition of
both treated stock and NH4Cl. Stars indicate treated glycerol addition only.
Methanol was present in the treated crude glycerol stock. At the start of the
fed-batch, methanol concentration was 3.7 g/L (Fig. 1-4). During fermentation,
methanol concentration decreased to 1.7 g/L by day 3. Since new crude glycerol was
added at day 3 after sampling, methanol concentration was increased to 4.1 g/L. After
that time, the decreasing trend was observed again until another crude glycerol
addition was made on day 6.25. Overall, a pattern of increase then decrease was
13
repeated during the first 10 days of the experiment. After day 10, methanol
concentration decreased with time continuously since no more glycerol
supplementation was attempted.
Fig. 1-4. Methanol concentration changing with time for the two fed-batch processes,
fed-batch 1 (▲) and fed-batch 2 (■). Hollow arrows and solid arrows indicate
addition of both treated stock and NH4Cl for fed-batch 1 and 2, respectively. Stars
indicate treated glycerol addition only.
Compared with data from the first fed-batch, adding more treated stock at the
beginning also resulted in a higher methanol concentration (Fig. 4). However, the
trend of methanol concentration changing with time was identical to that of the
first-fed batch. Whenever glycerol stock was added, methanol concentration increased.
14
During the 12 days, methanol concentration was relatively stable between day 6 and
12.
Similar to the results demonstrated by Meesters et al. (1996) with pure glycerol,
the fed-batch approach was proven effective in our study to overcome substrate
inhibition and improve biomass growth. Using crude glycerol, we did not achieve
biomass densities as high as those reported by Meesters and colleagues. However, the
biomass productivity we demonstrated was definitely higher than that achieved with
C. curvatus provided with different sugars (glucose, xylose, sucrose, and lactose)
either in batch or continuous culture modes (Evans and Ratledge, 1983). In addition,
compared with other oleaginous species (microalgae, yeast, or mold) grown on crude
glycerol, C. curvatus produced the highest biomass density, biomass yield, and
cellular lipid content (Table 1-1). Furthermore, based on the observations from the
two fed-batch operations, the biomass DW could be significantly enhanced through
feeding treated stock more frequently to maintain continuous growth and eliminate
substrate depletion. In comparison with the overall yield reported for pure glycerol
(0.4 g biomass per g glycerol; Meesters et al., 1996), the yields obtained in this study
were either higher or identical for the first and second fed-batch process, respectively.
Moreover, the lipid contents in the final biomass samples were higher than 25%
documented using pure glycerol. Further optimization of the fed-batch process to
yield higher biomass growth and lipid production is ongoing in our laboratory.
15
Table 1-1. Comparison among oleaginous species on crude glycerol.
Species
Cryptococcu
curvatus
Cryptococcu
curvatus
Schizochytrium
limacinum
SR21
Schizochytrium
limacinum
SR21
Mortierella
isabellina
Yarrowia
lipolytica
Maximum Biomass
Yield Glycerol Lipid Lipid
Growth
DW (g)
productivity (g/g) uptake
(%)
productivity mode
(g/L-day)
rate
(g/L-day)
(g/L-day)
a
31.2
2.6
0.6
19.2b
44.6 1.2
Fed-batch
Reference
This study
32.9
2.7
0.4a
28.3c
52.9
1.5
Fed-batch
This study
11.5
1.9
0.2
8.3
50.4
1.0
Batch
Pyle (2008)
18.0
3.1
0.3
5.9
50.5
1.5
Batch
Chi (2007)
8.5
0.5
0.2
3.1
51.7
0.3
Batch
8.1
NA
0.2
NA
43.0
2.6
a
Calculation is provided in Table 2.
b
Maximum substrate uptake rate took place between 168 and 192 h.
c
Maximum substrate uptake rate took place between 24 and 48 h.
Papanikolaoua
(2008)
Continuous Papanikolaoua
(2002)
Comparing the two fed-batch operations, the biomass densities produced were
similar, but when nitrogen addition was terminated after 6 days, a higher cellular lipid
content was attained. This is in agreement with other studies reporting that lipid
accumulation increases with nitrogen limitation (Hassan et al., 1996; Meesters et al.,
1996; Ykema et al., 1988).
As indicated by Table 1-1, the biomass yields per gram of glycerol were 0.6 and
0.4 (g/g) for the first and second fed-batch process, respectively. Total biomass
included the amount present at day 12 as well as those withdrawn during daily
16
sampling. Total mass of glycerol utilized was calculated based on total glycerol added
and the amount that was lost in the samples taken daily. Table 1-2 provided the brief
yield calculations for the two fed-batch processes.
Cellular lipid contents of day 12
biomass samples were 44.6±1.5% and 52.9±2.0% for the first and second fed-batch
operation, respectively.
Table 1-2. Yield calculation
Operation
Fed-batch 1
Fed-batch 2
a
Glycerol
Total added
(g)
134.3
214.2
Total lost
(g)
6.0
4.6
a
Biomass
Biomass
losta (g)
21.6
21.9
yield (g/g)
Biomass at
day 12 (g)
50.0
54.6
0.6
0.4
Glycerol and biomass lost due to daily sampling
One of the issues related to usage of crude glycerol is the presence of methanol.
This chemical has been shown to affect microalgal growth negatively at high
concentrations (Chi et al., 2007; Pyle et al., 2008). In this investigation, methanol was
added to the culture through adding treated stock. As indicated in Table 1-3, the
observed methanol concentrations were lower than expected values for the two
fed-batch processes. Additionally, methanol concentrations always decreased with
time after supplementation during the first few days of experiments. At 28 oC,
methanol cannot be evaporated from the bioreactor, thus the loss of methanol must be
the result of utilization by the yeast cells. This yeast has been demonstrated to convert
ethanol to lipids as efficiently as other substrates, such as glucose, xylose, lactose, and
sucrose (Evans and Ratledge, 1983). Thus, it is reasonable to hypothesize that C.
17
curvatus can uptake methanol and utilize it similarly. Further confirmation and
investigation of the methanol concentration effects observed in the present study are
in progress.
For samples taken at day 3, 6, 9, and 11 during the first fed-batch operation, the
fatty acid profile was revealed. As shown in Fig. 1-5, the four major fatty acids that
constituted 94% of all fatty acids were 16:0, 18:0, 18:1n-9, and 18:2n-6. We have
observed that, 1) the percentage of 16:0 was 15% and 16% at day 3 and 6,
respectively, but increased to 23% by day 9 and was stable thereafter;
2) percentage
of 18:0 was increased from 11% at day 3 to 21% at day 6, but then decreased to 15%
based on measurements at day 9 and 11; 3) percentage of 18:1n-9 was basically
unchanged during the experimental period; and 4) percentage of 18:2n-6 decreased
from 23% at day 3 to 13% at day 6 and remained the same after that time. After 11
days, the lipids in yeast biomass contained 16:0, 18:0, 18:1n-9, and 18:2n-6 in the
percentages of 23.0, 16.7, 39.6, and 15.2%, respectively.
Table 1-3. Methanol mass balance
Operation
Total addedc (g)
Total lostd (g)
Fed-batch 1a
Fed-batch 2b
23.6
36.9
4.8
6.1
Anticipated
conc. (g/L)
11.7
18.5
a
Treated stock had a methanol concentration as 135.8 g/L.
b
Treated stock had a methanol concentration as 144.9 g/L.
c
Methanol added was those that were in treated stock.
d
Methanol lost was those that were withdrawn during daily sampling.
Observed
conc. (g/L)
4.7
4.7
18
Considering the biodiesel properties such as freezing point, oxidative stability,
octane number, and NOx emissions associated with various feedstocks, the ideal
material for biodiesel purpose is one composed of 100% monounsaturated fatty acids,
including 16:1, 18:1, 20:1, or 22:1 (Tyson et al., 2004). Comparing the fatty acid
profiles of different oil feedstocks (Table 1-4), yeast oil together with Palm oil and
Jatropha oil have the highest percentages of monounsaturated fatty acids, which make
them better sources for biodiesel production than soybean oil.
Table 1-4. Fatty acid comparison with others
Fatty acid
Jatrophaa
Soybeanb
Palm oilb C.curvatusc
This study
Palmitic acid C16:0
10–13
7–11
32–47
25-30
23
Palmitoleic acid C16:1
0.3–0.5
0–1
0
0
0.9
Stearic acid C18:0
2–3
3–6
1–6
10–15
16.7
Oleic acid C18:1
41–49
22–34
40–52
45–50
39.6
Linoleic acid C18:2
34–44
50–60
2–11
5–7
15.2
Linolenic C18:3
0.1–0.2
2–10
0
0
0.66
Saturated
10–16
10–16
33–53
35–45
39.7
Monounsaturated
41–49
22-35
40–52
45-50
40.5
Polyunsaturated
33–44
52–70
2–11
5-7
15.8
a
Martinez-Herrera et al. (2006).
b
Tyson et al. (2004).
c
Davies (1988).
19
Fig.1-5. Major fatty acids changing with time during fed-batch 1
1.4. Conclusion
This study demonstrated that: 1) C. curvatus can grow on crude glycerol; 2)
fed-batch is a better process than batch for enhanced glycerol uptake and biomass
production; 3) methanol did not pose a significant inhibitory effect even though it was
existent in the bioreactor; and 4) lipid produced from yeast fermentation can serve as
a good biodiesel feedstock. With further process optimization, the crude
glycerol-to-lipid culture model will provide additional oil feedstock for fuel purposes
while eliminating the problem of crude glycerol disposal.
20
CHAPTER 2
FERMENTATION OPTIMIZATION FOR THE PRODUCTION OF LIPID BY
CRYPTOCOCCUS CURVATUS: USE OF RESPONSE SURFACE
METHODOLOGY
2.1. Introduction
The global energy consumption has increased 16-fold while the human
population has quadrupled in the twentieth century (Hoffert et al., 2002). Increased
use of fossil fuels has resulted in accelerated release of CO2, which is now generally
accepted as a major factor contributing to the green house effect (Houghton, 2001).
The Stern-Review on the Economics of Climate Change has publicized the economic
necessity to limit global warming (Stern, 2006). Thus, decreasing usage of fossil fuels
and protecting the environment are becoming overriding challenges.
Biodiesel, which is mainly derived from triglycerides in vegetable oils through
transesterification reaction, has attracted considerable attention in recent decades.
Biodiesel is generally recognized as a renewable and biodegradable fuel that can
replace petrodiesel. But considering the “Fuel vs. Food” debate, oil feedstocks rather
than those from plants are highly sought. In addition, due to rapid expansion of
biodiesel production, the global market has been flooded with crude glycerol, the
major by-product from transesterification reaction. Proper utilization of this material
has become a hot research area. Thus, to maintain a sustainable development of
biodiesel, two issues in terms of limited oil feedstocks and over-supplied crude
21
glycerol must be solved properly. During recent years, microbial oils produced by
various microorganisms have been intensively investigated and are believed to be an
alternative oil source for sustaining biodiesel production. In particular, those
microorganisms that can consume crude glycerol while accumulating lipids have been
placed under the spotlight. Such microorganisms include microalgae, yeasts, bacteria,
and fungi. Among these microbial species, Cryptococcus curvatus (ATCC 20509) is
of great interest to the scientific community. When C. curvatus is grown on cheap
carbon sources like whey permeate (Ykema et al., 1988) and other carbohydrate-rich
agricultural or food processing wastes (Vega et al., 1988; Bednarski et al., 1986), it is
able to accumulate 60% of cell dry weight as lipids (Ratledge, 1991). The yeast oils
produced by C. curvatus resemble plant seed oils like palm oil (Davies, 1988) and can
certainly serve as an excellent feedstock for biodiesel production.
Previous publication by our research group has shown that C. curvatus has
reasonable growth and lipid yields when it is fed with pretreated crude glycerol
derived from yellow grease (Liang et. al., 2010a). To further improve the biomass and
lipid productivities, this study aims to identify the optimal condition for yeast growth
through use of statistical design tools. A Box–Behnken statistical design of
experiments was used. This design comprised three factors at three levels of variation
to permit an un-confounded estimation of the regression coefficients. The responses
were biomass density and cellular lipid content. Optimal parameters obtained from the
optimization study were further verified in one fed-batch experiment.
22
2.2. Materials and Methods
2.2.1. Yeast culture
C. curvatus (ATCC 20509) grown in liquid medium containing 2% peptone, 1%
yeast extract, and 16 g/L of pure glycerol was utilized as an inoculum for all
experiments described in this study adopting a minimal medium (Meesters et al.,
1996). The minimal medium contained (per liter): 2.7 g KH2PO4; 0.95 g Na2HPO4;
0.2 g MgSO4 ·7H2O; 0.1 g yeast extract; and 0.1 g EDTA. After the pH was adjusted
to 5.5, it was supplemented with a 100x spores stock solution consisting of (per liter):
4 g CaCl2 ·2H2O; 0.55 g FeSO4.7H2O; 0.52 g citric acid; 0.10 g ZnSO4.7H2O; 0.076 g
MnSO4.H2O; and 100 µL 18 M H2SO4. Nitrogen was supplied in the form of NH4Cl
at 2.5 g/L with an initial C/N ratio of 30:1.
2.2.2. Crude glycerol pretreatment
Crude glycerol was obtained from Midwest Biodiesel Products, LLC (Caseyville,
IL, USA). This refinery used alkali-catalyzed transesterification to produce biodiesel
from animal fats. The crude glycerol stream was processed for complete methanol
recovery. As a result of that, the crude glycerol sample appeared as a gel-like
semi-solid at room temperature with a pH value of 11.2. To prepare this material for
yeast fermentation, the crude glycerol sample was heated to 60 ◦C in a water bath to
turn it into a liquid form. Following pH adjustment to 1.0, the sample was transferred
into a separatory funnel and allowed to stand for few hours for phase separation. The
bottom layer with a reddish-brown color was the pretreated crude glycerol which was
23
used in the study presented here. Concentration of glycerol in this layer was
determined by HPLC (Shimadzu Scientific Instrument Inc., Columbia, MD, USA) as
described below.
2.2.3. Experimental design
To identify the optimal conditions for yeast growth on pretreated crude glycerol,
Box–Behnken design was employed. This design has been used to examine the
relationship between one or more response variables and a set of quantitative
experimental parameters based on response surface methodology (Box and Behnken,
1960). Three variables, temperature (27-33oC), pH (5.0-6.0), and glycerol
concentration (10-30 g/L) were selected since they were considered as having the
most significant effects on biomass yield and lipid production in general (Evans and
Ratledge, 1983; Meesters et al., 1996; Ykema et.al, 1988). Biomass yield and cellular
lipid content were set as analytical responses. A three-factor and three-level
experiment was designed using the Design-Expert (Stat-Ease Inc. Minneapolis, MN,
USA) program. Each variable was tested in three different coded levels: low (−1),
middle (0), and high (+1) (Table 2-1). Based on experimental results, a second-order
polynomial model for the variables was obtained:
Y=β0+Σβixi+Σβixi2+Σβijxixj
where Y is the predicted response, β is the coefficient of the equation, and xi and xj are
the coded levels of variables i and j, respectively. The statistical analysis of the model
was performed in the form of analysis of variance (ANOVA), the second-order model
24
equation and significance of variables were determined by Fisher’s F-test. This design
consists of replicated center points and the set of points lying at the midpoints of each
edge of the multidimensional cube that defines the region of interest.
The experiments were conducted in 250-mL Erlenmeyer flasks with a total
volume of 100 mL which included minimal medium, different volumes of pretreated
crude glycerol for achieving different concentrations, and 10% yeast inoculum. The
initial C/N ratio (g/g) was set to 30 as the optimal condition for cell growth and lipid
production. A total of 17 flasks were set up based on the design shown in Table 2-2
and incubated at different temperatures with shaking at 150 rpm. pH of the culture
were kept constant during the 72 h experimental period. To obtain yeast cell dry
weight, samples were centrifuged, washed with distilled and deionized water (DDW)
twice, and freeze-dried. The freeze-dried samples were also used for analysis of lipid
contents.
Table 2-1. Process variables and their levels used in the design
FACTORS
A—Temperature/℃, X1
B—pH, X2
C—Glycerol conc. g/L, X3
-1
27
5.0
10
LEVELS
0
30
5.5
20
1
33
6.0
30
2.2.4. Optimized fed-batch fermentation experiment
In this experiment, the obtained optimal parameters from the response surface
method were used. The fed-batch fermentation experiment was conducted in a 2-l
fermentor with a bladed stirrer (New Brunswick, Edison, NJ, USA). The fermentor
25
was autoclaved at 121 oC for 1 h after 1.0 liter of minimal medium was filled in and
pH and DO probes were connected and calibrated. The treated glycerol stock was then
supplemented together with the inoculum. The starting concentrations of glycerol and
NH4Cl were 19.8 and 0.96 g/L, respectively to maintain a C/N ratio (g/g) as 30. pH
was controlled at 6.0 by automatic pumping of NaOH solution. Temperature was kept
at 30.2 oC. Incoming air was filtered through a 0.2 µm filter, and air flow rate was
upheld at 0.6 l/min. The stirring speed was set at 900 rpm. When needed, silicone
antifoam emulsion (J. T. Baker, Phillipsburg, NJ, USA) was added to prevent excess
foam formation. Furthermore, to avoid frequent exhaustion of crude glycerol in the
fermentor and eliminate the tedious addition of the substrate and nitrogen to the
reactor manually, an automatic pump (Cole Parmer, Vernon Hills, IL,USA) controlled
by dissolved oxygen (DO) concentration monitor (B&C, Tampa, FL, USA) was used.
When the DO concentration rose above 0%, the pump would start pumping in mixture
of pretreated crude glycerol and NH4Cl during the first 6 days. After that time, the
supplementation of nitrogen was terminated while the glycerol addition was continued.
When the DO concentration was lower than 0%, the pumping activity was
automatically stopped. In such a way, an uninterrupted nutrient supply was achieved.
To monitor the fermentation process, a 61 mL sample was taken daily for
determining cell dry weight (DW) and glycerol concentration. At the same time, a
total volume of about 60 mL of concentrated minimal medium and spores stock
solution were added to maintain a similar culture volume between 1.5 and 1.6 liters.
The fermentor was allowed to run for 12 days. Samples harvested at different time
26
intervals were centrifuged first at 4,000g for 10 min. The supernatant was then filtered
through a 0.2 µm nylon filter for determining glycerol concentration by HPLC. The
pellets were washed with DDW twice and freeze-dried overnight to attain biomass
DW. The dried biomass was reserved for cellular lipid analysis.
Table 2-2. The Box-Behnken design of the variables with biomass yield and lipid
content as responses
Runs
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Temperature
o
C
30.0
33.0
30.0
30.0
27.0
27.0
33.0
30.0
27.0
30.0
33.0
30.0
33.0
30.0
30.0
30.0
27.0
pH
5.5
5.5
5.5
5.5
5.5
5.5
5.0
6.0
6.0
6.0
6.0
5.5
5.5
5.5
5.0
5.0
5.0
Glycerol
Conc. (g/L)
20.0
10.0
20.0
20.0
30.0
10.0
20.0
10.0
20.0
30.0
20.0
20.0
30.0
20.0
10.0
30.0
20.0
Biomass
yield (g/L)
6.6
5.0
5.8
7.8
3.4
6.3
2.1
6.8
4.9
4.2
5.9
9.0
3.0
7.4
4.8
3.6
6.8
Lipid
content (%)
32.77
29.03
33.09
26.72
15.27
36.05
26.32
47.11
44.24
18.23
42.40
26.66
18.53
34.85
24.81
21.99
33.30
2.2.5. Analysis
Glycerol concentration was determined by HPLC (Shimadzu Scientific
Instrument, Inc. Columbia, MD, USA) with a refractive index detector. A Supelcogel
AG1 column (5 µm, 30 cm ×4.6 mm, Supelco, Bellefonte, PA, USA) was used in an
oven set at 83 oC. HPLC grade water was used as the mobile phase with a flow rate as
27
0.5 mL/min. The injection volume was 10 µL. Concentration of glycerol was
calculated based on a calibration curves built for this compound using an external
standard.
Cellular lipid content was determined following a procedure developed in our
laboratory (Liang et al., 2010b). Briefly, 0.1 g dried cell pellet was transferred to a
7-mL chamber of a bead-beater (BioSpec Products, Bartlesville, OK, USA). This
chamber was filled with 0.5 mm zirconium beads to approximately 5 mL. Methanol
was then added to fill the rest of the chamber. After cells were disrupted by
bead-beating for 2 min, the entire content was transferred to a 50-mL glass centrifuge
tube. The chamber was washed twice using methanol (total 10 mL) to collect the
yeast residue, and the washes were added to the primary methanolic extract.
Chloroform was then added to make a 2:1 (v/v) chloroform/methanol ratio (v/v). The
tube was vortexed for 5 min and was allowed to stand for 24 h. Afterwards, the tube
was centrifuged at 4,000 × g for 15 min to remove the zirconium beads and yeast
solids. The supernatant was collected and the solvent was vaporized using rotovap.
The resultant crude lipid was weighed to calculate oil content.
2.3. Results and discussion
2.3.1 Development of a regression model
The experimental runs and results for the Box–Behnken design are shown in
Table 2-2. The 17 runs in a single block were used to study the effects of three factors
on two responses. For all combinations tested, biomass concentration ranged from 2.1
28
g/L to 9.0 g/L and lipid content varied from 15.3% to 44.2%.
The application of response surface methodology yielded the following
regression equation models which are empirical relationships between the biomass
yield, cellular lipid content values and the test variables in coded units. The relation
among the variables (as coded values) temperature (X1), pH (X2), and pretreated crude
glycerol concentration (X3) was fitted by second-order polynomial equations (1) and
(2):
Biomass yield =7.32-0.67X1 + 0.56X2 - 1.09X3 + 1.42X1X2 + 0.22X1X3 - 0.35X2X3
2
2
- 1.41X1 - 0.99X2 - 1.48X3
2
(1)
Lipid content =30.82 - 1.57X1 + 5.69X2 - 7.87X3 + 1.29X1X2 + 2.57X1X3 - 6.51X2X3
2
2
+ 1.22X1 + 4.53X2 - 7.31X3
2
(2)
Analysis of variance (ANOVA) is required to test the significance and
adequacy of the model. ANOVA for response surface quadratic model for the
biomass yield and lipid content is indicated in Table 2-3. The regression models
accurately described the experimental data, which indicated successful correlation
among the three fermentation process parameters that affected both biomass yield and
lipid content. This was supported by the values of correlation coefficients R2 as 0.88
and 0.94 for biomass yield and lipid content, respectively. These R values suggested a
satisfactory representation of the process model and a good correlation between the
experimental results and the theoretical values predicted by the model equation. The
P value of the models of 0.0156 and 0.0011 indicated the significance of the
coefficients. The Lack of Fit F-values of 0.16 and 0.28 implied the lacks of fit were
29
not significant relative to the pure error. The P-values of the two F-values indicated
that there were 91.71% and 84.12% chance that the lack of fit F-value this large could
occur due to noise. The response surfaces were fitted using process variables that
were found to be significant after the analysis. With the established models, different
combinations of variables (pH, temperature, and glycerol concentration) are able to
lead to desired yields of biomass and lipid or result in maximal glycerol consumption.
Table 2-3. Analysis of variance (ANOVA) of experimental data
Source
Model
X1
X2
X3
X1X2
X1X3
X2X3
X12
X22
X32
Residual
Lack of fit
Pure Error
Cor Total
Degree
of
freedom
9
1
1
1
1
1
1
1
1
1
7
3
4
16
Biomass yield
F-value P-value
Lipid Content
F-value
P-value
5.73
3.87
2.68
10.03
8.61
0.21
0.52
8.88
4.33
9.85
0.0156
0.09
0.1453
0.0158
0.0219
0.6571
0.4944
0.0205
0.0759
0.0164
13.92
1.94
25.39
48.54
0.65
2.59
16.61
0.61
8.46
22.02
0.0011
0.2062
0.0015
0.0002
0.4478
0.1517
0.0047
0.4608
0.0227
0.0022
0.16
0.9171
0.28
0.8412
2.3.2. Effects of process parameters on optimization
From the variance analysis, it could be concluded that the glycerol concentration
and temperature had more significant effects on biomass yield than those from pH.
But, in terms of lipid content, the effects from glycerol concentration and pH were
30
more significant. Three-dimensional surface responses were plotted to illustrate the
relationships between the responses and variables (Fig. 2-1). As shown by this figure,
when pH was fixed at 6.0, with the increase of temperatures, the increase of glycerol
concentration led to a decrease in biomass yield; for lipid content,when temperature
was fixed at 30 ◦C, at middle glycerol concentration (20 g/L), the increase of pH
resulted in a better accumulation of cellular lipids.
31
Fig. 2-1a. Three-dimensional response surface plot of biomass yield (top) and lipid
content (bottom) as a function of different parameters.
32
Fig. 2-1b. Three-dimensional response surface plot of lipid productivity as a function
of different parameters.
With regard to biomass yield, four effects had P-values less than 0.05, indicating
that they were significantly different from zero at the 95% confidence level (Table
2-3). These effects were the glycerol concentration, the quadratic effect of glycerol
concentration and temperature, and the interaction between temperature and pH.
Considering the F-ratio statistic, it might be concluded that a change in glycerol
concentration caused the major variation in biomass density. This can be explained
that carbon from glycerol was the major nutrient for yeast biomass growth, which
constituted the structural backbone of living cells. In this experiment, nitrogen
feeding through a fixed ratio with carbon helped the synthesis of DNA and proteins,
hence resulted in cells proliferation as well. On the other hand, high concentration of
33
glycerol would inhibit the growth of yeast cells, which was also in agreement with
those reported (Hassan et al., 1996; Meesters et al., 1996).
In the case of lipid content, five effects: pH, glycerol concentration, the
interaction between pH and glycerol concentration, the quadratic effect of pH and
glycerol concentration, had P-values less than 0.05. Similarly, glycerol concentration
was the main source of variation in lipid accumulation. The effect from pH was also
statistically significant. pH appeared as an important factor through its quadratic
effect and the interactions with glycerol concentration. Temperature had no effect in
comparison with the other variables. According to the F-ratio statistics, glycerol
concentration was the principal factor that influenced cellular lipid accumulation. An
excessively low or high glycerol concentration would both reduce the lipid content.
The cellular lipid accumulation process generally requires the exhaustion of nitrogen
to allow the excess carbon to be converted into lipids. Under such conditions, the
increase in intracellular lipid content results essentially from the synthesis of
saturated and monounsaturated fatty acids in most oleaginous microorganisms
(Hassan et al., 1996). The F-ratio also demonstrated that pH had a secondary
important role and lipid production capacity was affected by pH values of the culture
medium. Similar to other yeast species, the optimum pH values varied from 5.0 to 6.0
(Karatay and Dönmez, 2010).
In this study, our aim was to maximize the biomass yield and cellular lipid
content by finding optimal conditions. According to the models identified above, the
optimal conditions for achieving maximal biomass density and lipid content were 30.2
34
◦
C, pH 6.0, and 19.8 g/L of glycerol concentration. The predicted values of biomass
yield and lipid content were 6.98 g/L and 41.31%, respectively. All of the optimal
parameters were verified by comparing the experimental data obtained under these
conditions with the predicted numbers. The verification experiment provided a
biomass yield and lipid content as 7.11 ± 0.36 g/L and 38.53% ± 1.88%, respectively.
The small deviations between the experimental and predicted data suggested that the
experimental designs used in this work were effective for accomplishing our purpose.
Comparing with our previous batch experimental results of 5.59 g/L for biomass yield
and 35.43% for cellular lipid content obtained when C. curvatus was grown on 20 g/L
of crude glycerol at 30 oC with a pH of 5.5 (Liang et al., 2010a), the percentages of
increase of 27.19 and 8.75 were attained for biomass density and lipid content,
respectively. Therefore, the optimal temperature, pH, and glycerol concentration
acquired from this study are valuable for our future efforts toward further enhancing
lipid productivity in fed-batch or continous culture mode.
2.3.3 The influence of optimal parameters on fed-batch fermentation
For the fed-batch experiment described above, adopting the optimal parameters
obtained, the biomass yield was 44.5 g/L (Fig. 2-2a). During the whole experimental
period, the average biomass productivity was 3.71 g/L-day which was half of the
maximum biomass productivity of 8.91 g/L-day taking place from day 0 to day 2.
With regard to lipid content, it was 45.6 ± 1.2% for day 6 cells and 49.0 ± 1.0% for
day 12 samples. As mentioned above, nitrogen was no longer supplied after day 6.
35
Thus, limiting nitrogen concentration in the culture did promote lipid accumulation.
Fig. 2-2a: During the fed-batch experiment, biomass density (▲) and glycerol
concentration (■) changing with time.
As shown in Fig. 2-2a, glycerol concentration in the fermentor was almost zero.
But a closer look indicated that it was not absolutely zero (Fig. 2-2b). After day 6,
glycerol concentration was maintained around 0.06 g/L. This low concentration was
resulted from the balance between active yeast growth and intermittent feeding of
concentrated crude glycerol. A similar approach was adopted by Schmidt (Schmidt et
al., 2005) to obtain a high cell-density culture for red and acidophilic microalga
Galdieria sulphuraria. During the fed-batch operation, pulsed addition of feed
medium which contained high concentration of glucose and ammonium was
controlled by the DO tension. Rapid consumption of glucose resulted in fast decrease
of DO. After glucose was depleted, DO tension was increased due to diminished
36
respiration rate which triggered the intermittent feeding of glucose and ammonium to
the reactor. Throughout the fed-batch fermentation, the glucose concentration was
kept sufficiently low (lower than 0.3 g/L) to serve as the growth-limiting factor,
although the total amount of added glucose was significant.
Fig. 2-2b: Glycerol concentration changing with time during the fed-batch
experiment (Note: small glycerol concentrations on the y-axis).
Comparing the growth curve from this study and one from our previous
fed-batch experiment (Fig. 2-3), the biomass density of 44.5 g/L from this study was
higher than 32.9 g/L from the previous one that was conducted at 28 oC with a pH of
5.5 and lasted also for 12 days (Liang et al., 2010a). Similarly, at the end of the
fed-batch experiment, the lipid yield of 21.8 g/L was higher than the previously
obtained 17.4 g/L. For the study presented here, at day 1, the biomass density was
4.79 g/L, which was a 36.86% increase from the previous one (3.5 g/L at day 1) at the
37
same time point. This increase demonstrated the individual effect of the optimized
parameters, since the intermittent feeding has not yet begun. At day 6, the cell yield
almost reached the day 12 value from the previous study. The increase of both
biomass and lipid yields could be attributed to optimal parameters adopted and the
automatic pumping of crude glycerol instead of intermittent manual addition.
Fig. 2-3: Comparison of biomass density changing with time between two fed-batch
studies. Optimized (▲) and Unoptimized (■).
Regarding glycerol utilization, during the initial 24 h, the glycerol utilization rate
of 3.17 g/L-day was low accompanied by the lag phase of cell growth. But after the
cells acclimated to the conditions in the fermentor, the fast substrate uptake rate of
around 40 g/L-day resulted in the exponential growth stage of yeast cells. The average
substrate utilization rate for the whole experiment was 31.44 g/L-day, while in the
38
latter 6 days the average utilization rate was 34.03 g/L-day.
The biomass yield (g cells/g glycerol consumed) was calculated for the
optimized fed-batch process (Table 2-4). Total glycerol added included those that
were added to the fermentor at the beginning of the experiment and those that were
pumped in intermittently during the fermentation. Lost glycerol was referred to those
that were lost in the daily withdrawn samples. Similarly, the daily taken samples also
contributed to the biomass loss. Comparing with our previous fed-batch experiment
which ended with a cell yield of 0.36 g/g under the condition of manual addition of
crude glycerol periodically (Liang et al., 2010a), the overall biomass yield of 0.26
from this study was smaller. This lower cell yield could be contributed by the absence
of nitrogen during the second stage of fermentation. In addition, considering the
nature of crude glycerol as an industrial waste, the above experiments were designed
to
maximize cellular lipid yield while achieving maximal glycerol consumption.
Thus, this result demonstrated the effectiveness of the polynomial regression models
of the optimization.
Table 2-4. Illustration of calculations for crude glycerol utilization and biomass yield
Operation
Unoptimized
fed-batch
Optimized
fed-batch
Glycerol
Total added
Total lost
(g)
(g)
214.2
4.57
384.8
1.44
Biomass
Biomass
Biomass at
lost (g)
day 12 (g)
21.9
54.6
18.77
80.19
yield (g/g)
0.36
0.26
39
Using a similar C. curvatus strain (ATCC 20508), Thiru et al. (2011) reported a
biomass density and lipid content of 69.2 g/L and 48%, respectively from a fed-batch
culture. The optimum medium used was comprised of crude glycerol (10 g/L), corn
steep liquor (20 g/L), and deoiled C. curvatus lysate (5 g/L). Based on glycerol, the
biomass and lipid yield was 0.77 and 0.53 g/g, respectively. Though these yields were
impressive, they were much higher than the maximum theoretical yield that is
possible. For example, the maximum theoretical yield for lipid is 0.30 g lipid/g
glycerol (Ratledge, 1988). The overestimated yield could be due to the organic
carbons available in corn steep liquor (CSL) and deoiled C. curvatus lysate.
Depending on the variety of CSL, the organic carbon content can be as high as 42.9%
(Obayori et al., 2010). The same is true for deoiled C. curvatus lysate. Since only oils
were eliminated from the yeast biomass, the lysate surely contained a variety of
organic carbons that can be consumed by the yeast. Thus, the biomass and lipid
produced from the fed-batch were from glycerol and other organic carbons
present in the medium. For the purpose of maximizing the consumption of crude
glycerol by C. curvatus, the medium proposed in that study is not appropriate.
Additionally, the yield reported should be reevaluated.
Crude glycerol has also been tested as a substrate for other oleaginous
microorganisms, such as microalgae, yeast and fungi. When Yarrowia lipolytica
was cultivated in a continuous growth mode, a maximum cell dry weight of 8.1 g/L
with the lipid content of 43% was identified (Papanikolaou et al., 2002). Using a
similar continuous culture, microalga Schizochytrium limacinum SR21 produced a
40
biomass density and docosahexaenoic acid (DHA) yield as 11.78 g/L and 1.74 g/L ,
respectively (Ethier et al., 2011). In terms of fed-batch culture mode, the yeast
Rhodotorula glutinis provided 6.10 g/L lipid from 10.05 g/L of dry cells (lipid content
60.7%) in a 3-day experiment (Saenge et al., 2011). For a green microalga, Chlorella
protothecoides, cell density of 45.2 g/L and lipid concentration of 24.6 g/L were
achieved in 8.2 days (Chen and Walker, 2011). The biomass and lipid yield from our
work is comparable to those listed here, but can still be improved. Currently, we are
raising the oxygen threshold to keep a higher glycerol concentration in the fermentor
to prevent substrate limitation on cell growth and lipid accumulation.
2.4. Conclusion
The response surface methodology allowed the development of empirical polynomial
models for predicting biomass production and cellular lipid content for oleaginous
yeast C. curvatus grown on crude glycerol. The derived equations and contour plots
allowed the identification of optimal parameters for obtaining maximal biomass
density and lipid content. Verification by an experiment using the optimal temperature,
pH, and glycerol concentration resulted in similar values for biomass yield and lipid
content as those predicted by the models. A fed-batch process adopting the optimized
variables led to improved yields of biomass and lipids compared to those from our
previous study.
41
CHAPTER 3
CULTIVATION AND LIPID PRODUCTION OF YEAST CRYPTOCOCCUS
CURVATUS USING SWEET SORGHUM SYRUP
3.1. Introduction
Sweet sorghum has been recently selected by USDA/DOE as a suitable biomass
feedstock for South Central and Southeast regions considering its several advantages,
such as: its ability to withstand dry conditions, a low requirement for fertilizer, a rapid
growth rate, ease of planting and a low cost of total fermentable sugars (Bulawayo et
al., 1996). Sugars in the stalk juice are mainly sucrose, fructose and glucose.
(Sipos et al., 2009). Traditionally, sorghum juice has been used for making white
sugar (Gnansounou et al., 2005) and syrup. During the past decades, due to rapidly
growing energy demand, sweet sorghum juice has been examined for bioethanol
production through yeast fermentation (Laopaiboon et al., 2007; Liu et al., 2008; Liu
and Shen, 2008; Mamma et al., 1995; Sipos et al., 2009). However, as US hits the
ethanol blend wall, novel uses of the sorghum juice needs to be researched and
developed. So far, only a few researches have attempted on converting sorghum juice
to microbial lipids. Grown on sorghum juice treated by sucrose invertase, microalga
Chlorella protothecoides produced 35.7% of more lipids compared to that from
glucose alone (Gao et al., 2010). Another microalga, Schizochytrium limacinum SR21,
was also demonstrated to accumulate lipids on sorghum juice (Liang et al., 2010).
42
However, both microalgal strains tested so far do not utilize the juice sugars
efficiently. In terms of C. protothecoides, an invertase enzyme was added to convert
sucrose to glucose and fructose which could then be utilized by the algae. Regarding S.
limacinum SR21, sucrose was not used at all while fructose utilization was very
limited. Thus, to continue research on producing microbial lipids on sorghum juice, a
better microbial strain need to be adopted.
Oleaginous yeast strains typically use a much broader range of substrates.
Among different lipid-accumulating yeast species, Cryptococcus curvatus is
exceptional. Research in our lab has shown that this yeast can accumulate lipids when
grown on glycerol and hydrolysates of sorghum bagasse and corn fiber (Liang et al.,
2009, 2010). Compared with other oleaginous yeasts, this strain produced the highest
lipid productivity that has been report so far. Thus, in this study, we aim to evaluate
growth and lipid productivity of this yeast on sorghum juice in batch and fed-batch
modes.
Besides evaluating lipid productivity, we also seek to develop a fast and effective
method for extracting microbial lipids. Unlike pressing oil out of seeds, extracting
lipids out of microbial cells involves two steps: cell wall disruption and lipid
extraction. To disintegrate microbial cells, various approaches, such as: ultrasonic,
bead beating, French press and microwave irradiation have been investigated. Among
these methods, microwave irradiation has been reported to be the most efficient since
heat and pressure generated rapidly within the biological system can disrupt the cell
wall structures and force compounds out of biological matrix, producing good quality
43
extracts in a very short time (Young, 1995). Following cell disruption, most studies
employed the conventional two-solvent (chloroform and methanol) system to extract
lipids. Some protocols avoid the use of organic solvents by adopting supercritical
fluid. But from the perspective of producing microbial lipids for cheap biofuel
production, lipid extraction conducted under supercritical conditions is just too
expensive to be practical. Thus, for this study, we evaluated the efficiency of lipid
extraction from C. curvatus cells by using microwave-assisted extraction (MAE). In
addition, we identified the best solvent for this process.
3.2. Materials and methods
3.2.1. Source of sweet sorghum syrup
Fresh sweet sorghum crops (KN Morris) grown at Heil’s Farm at Cobden,
Illinois, were harvested in early October 2011. After the heads were cut and leaves
removed, the stalks were stored in shade for seven days before they were squeezed by
a mill to obtain fresh juice, the fresh juice was cooked to syrup and kept in
refrigerator at 4 ◦C for long time preservation. Before use in this experiment, the
sorghum syrup was diluted and sterilized at 121 ◦C for 45 min.
3.2.2. Microorganism and inoculum preparation
C. curvatus (ATCC 20509) grown in liquid medium containing 2% peptone, 1%
yeast extract, and 20 g/L of glucose was utilized as an inoculum for all experiments
described in this study adopting a minimal medium (Meesters et al.,1996). The
44
minimal medium contained (per liter): 2.7 g KH2PO4; 0.95 g Na2HPO4; 0.2 g
MgSO4 ·7H2O; 0.1 g yeast extract; and 0.1 g EDTA. After the pH was adjusted to 5.5,
it was supplemented with a 100x spores stock solution consisting of (per liter): 4 g
CaCl2 ·2H2O; 0.55 g FeSO4.7H2O; 0.52 g citric acid; 0.10 g ZnSO4.7H2O; 0.076 g
MnSO4.H2O; and 100 µL 18 M H2SO4. The medium was autoclaved at 121◦C for 15
min before use. The C. curvatus culture was maintained at room temperature on a
rotary shaker set at 150 rpm.
3.2.3. Batch stage growth with sweet sorghum syrup
C. curvatus cultured to log phase was used to inoculate medium with syrup at
different concentrations diluted by different volumes of medium. Total sugars
concentrations as 27, 36, 54, and 72 g/L were tested in this study. Size of inoculum
was always 10% of the total volume. All cultures were grown in 250-mL Erlenmeyer
flasks with 100 mL total volume and shaken at 150 rpm on a rotary shaker. At
different time points, 10 mL samples were taken under sterile conditions and analyzed
as stated below. The same experiment was conducted two times to obtain repeatable
results which were described here.
3.2.4. Fed-batch growth with sweet sorghum syrup
Fed-batch experiments were conducted using a 2-L fermentor with a bladed
stirrer (New Brunswick, Edison, NJ, USA). The fermentor was autoclaved at 121 ◦C
for 45 min after 1.0 L of minimal medium was filled in and pH and DO probes were
45
connected and calibrated. Diluted sorghum syrup was then supplemented together
with the inoculum, pH was maintained at 5.5 by automatic pumping of NaOH.
Temperature was controlled at 28 ◦C. Incoming air was filtered through 0.2 µm filter.
The whole culture volume was maintained between 1.5 and 1.6 L. When needed,
silicone antifoam emulsion (J.T. Baker,Phillipsburg, NJ, USA) was added to prevent
excess foam formation.
The starting concentration of sugars was 47 g/L. On a daily basis, a 50 mL
sample was withdrawn from the bioreactor and processed for biomass DW, sugars
concentration, and other measurements as described below. The culture growth was
sustained for 3 days.
3.2.5. Analysis
Samples harvested at different times were centrifuged first at 5000 g for 5 min.
The supernatant was filtered through 0.2 µm nylon filter for HPLC determination of
sugar concentrations. The pellets were then washed with distilled and deionized water
(DDW) twice and freeze-dried overnight to obtain biomass dry weight. The dried
biomass was also used for lipid content analysis.
Sugar concentrations were determined by HPLC (Shimadzu Scientific Instrument,
Inc., Columbia, MD, USA) with a refractive index detector. A Supelcosil LC-NH2
column (5 µm, 25 cm × 4.6 mm) was used with 75% acetonitrile in DDW as the
mobile phase. The flow rate was 1 mL/min. The injection volume was 10 µL.
Concentrations of glucose, fructose, and sucrose were calculated based on calibration
46
curves built for these three sugars using standard chemicals.
Cellular lipid content was determined following a procedure developed in our
laboratory (Liang et al., 2009). Briefly, 0.5 g dried cell pellet was transferred to a
7-mL chamber of a bead-beater (Bio-Spec Products, Bartlesville, OK, USA). This
chamber was filled with 0.5 mm zirconium beads to approximately 5 mL. Methanol
was then added to fill the rest of the chamber. After cells were disrupted by
bead-beating for 2 minutes, the entire content was transferred to a 50-mL plastic
centrifuge tube. The chamber was washed twice using methanol (total 10 mL) to
collect the yeast residue. Chloroform was then added to the tube to make the
chloroform/methanol (2:1, v/v). The tube was vortexed for 5 min and was allowed to
stand for 24 h. After that, the liquid layer was transferred to another tube and
centrifuged at 5000 g for 5 minutes to remove the zirconium beads and yeast residuals.
The supernatant was collected and the solvent was vaporized using rotovap. Oil left in
the flask without solvent was weighed to calculate oil content.
3.2.6. Microwave assisted lipid extraction
A domestic microwave oven with exiting power of 900W was modified for this
purpose. The roof of the oven was drilled with three holes to pass through a
temperature reader, a water-cooled reflux condenser and a motor driven stirring bar,
which could ensure uniform mixing of the reaction mixture. Microwave-transparent,
100 mL three-neck round bottomed flasks were used as sample vessels. Test results
obtained from the average of duplicate tests for each run were analyzed to evaluate
47
the reproducibility of microwave effect. After each test, a time interval was spent to
let the reactor cool down and return to original conditions.
Freeze dried C. curvatus pellet was grinded to powder and passed through 500
µm mesh, then 0.5 g powder was mixed with different solvents such as
chloroform/methanol (2:1, v/v) mixture, pure methanol, ethanol or distilled water. The
reaction mixture was heated using microwave irradiation, meanwhile the stirring bar
moved at
a rotating speed of 850 rpm, for different time intervals from 4 to 24 min.
Upon completion of the extraction, the samples were centrifuged (5000 rpm) for 5
min and filtered through a 0.45 µm filter to separate the liquid phase that contained
the lipid from the yeast powder residuals. The supernatant was then transferred into a
50 mL round-bottom flask and solvents were evaporated in a rotovap. Oil left in the
flask without solvent was weighed to calculate lipid yield. During the extraction
process, the temperature was measured by placing a digital thermocouple in the
reaction flask. The temperature was found to be around 90 oC for the lipid extraction
conducted in this study.
3.3. Results and discussion
3.3.1. C. curvatus growth on sorghum syrup
Sorghum syrup is the highly concentrated form of sorghum juice. Based on
HPLC analysis, the syrup that we used contained
fructose and
327 g/L of glucose,
806 g/L of
810 g/L of sucrose. Since high concentration of substrate can inhibit
growth of C. curvatus, we first valuated cell growth and sugar utilization profile of
48
this yeast at different initial sugar concentrations.
As demonstrated by Fig. 3-1, a total sugar concentration of 72 g/L produced
similar biomass density as that by 54 g/L in 3 days, while 27 g/L and 36 g/L of sugars
generated less dense cultures.
However, on day 2, yeast grew with sugars
concentration of 27 g/L and 36 g/L had higher biomass dry weight. The possible
reason for these results may be there were influence of substrate inhibition: lower
substrate concentration could stimulate cell growth whereas higher concentration has
an inhibitory effect.
Biomass yield (g/L)
14
12
10
27
36
54
72
8
6
4
g/L
g/L
g/L
g/L
2
0
0
12
24
36
48
Time (hr)
60
72
84
Fig.3-1. With sweet sorghum juice sugars at different concentrations, C. curvatus
biomass changing with time. : ♦:27 g/L; ■: 36 g/L; ▲: 54 g/L; and ×: 72 g/L. Error
bars represent standard deviations among duplicates.
We used HPLC to track the three sugars concentration change over time for all
the samples, and we observed the preference of this yeast strain. As shown in Fig. 3-2
49
and Fig. 3-3, for all the four doses: (1) when there were abundant glucose present, it
was the first sugar to be utilized by yeast; (2) fructose concentration almost did not
change until glucose concentration decreased to around 5 g/L; and (3) sucrose was the
last sort to be utilized compare with the other two. Regarding glucose and fructose,
these observations were in agreement with Hassan’s study which reported that C.
curvatus had preferential utilization of glucose (Hassan et al., 1996). The explanation
for this observation may be the different energy requirements for yeast to produce
enzyme utilizing different sugars. When glucose, fructose and sucrose were all in the
medium, glucose was utilized preferably to fructose and sucrose.
80
Sugars conc. (g/L)
70
60
50
27g/L
36g/L
54g/L
72g/L
40
30
20
10
0
-10 0
12
24
36
48
60
72
84
Time (hr)
Fig.3-2. C. curvatus total sugars utilization for different juice concentrations.
♦: 27 g/L; ■:36 g/L; ▲:54 g/L and ×: 72 g/L. Error bars represent standard
deviations among duplicates.
50
Fructose
Glucose
Sucrose
14
Sugar Conc. (g/L)
12
10
8
6
4
2
0
0
12
24
36
48
60
72
84
-2
Time (hr)
Fructose
Glucose
a
Sucrose
16
Sugar Conc. (g.L)
14
12
10
8
6
4
2
0
-2
0
20
40
Time (hr)
60
80
b
51
Fructose
Glucose
Sucrose
Sugar Conc. (g/L)
25
20
15
10
5
0
0
12
24
36
48
60
72
84
Time (hr)
Fructose
Glucose
c
Sucrose
30
Sugar Conc. (g/L)
25
20
15
10
5
0
0
20
40
Time (hr)
60
80
d
Fig.3-3. C. curvatus sugar utilization profiles for different juice concentrations. (a)
27g/L (b) 36g/L; (c) 54g/L and (d) 72g/L. (from upper left to lower right a,b,c and d)
fructose (♦), glucose (■) , sucrose (▲) Error bars represent standard deviations among
duplicates.
Combining the results from cell growth and sugar utilization, we concluded that
52
juice sugars concentration as about 50 g/L was the optimal one for the further 3 days
fed-batch experiment. As indicated by Fig.3-4, the fed-batch fermentation generated a
biomass DW of 23.6 g/L in 3 days. At day 2, the biomass density was 21.6 g/L, and
yeast cells almost exhausted all the sugars in medium. However, due to continuous
supply of juice, the growth continued. During the whole experimental period, the
average biomass productivity was 7.9 g/L-day which was lower than the maximum
biomass productivity 10.8 g/L-day taking place from day 0 to day 2. In terms of
sugars utilization as indicated by Fig.3-5, during the initial 12 h, sugars utilization rate
was high accompanied by the immediate exponential growth phase. There was almost
no lag phase observed due to the fast acclimation of cells to glucose in juice. The
maximum substrate utilization rate was 23.92 g/L-day taking place during the first day.
With regard to lipid content, it was 50.8 ± 0.3% for day 3 cells. The fed-batch
approach was proven effective in our study to overcome substrate inhibition and
improve biomass growth.
50
45
40
35
30
25
20
15
10
5
0
25
20
15
10
5
Biomass yield (g/L)
Sugars conc. (g/L)
53
Sugars conc.
Biomass density
0
0
12
24
36
48
Time (hr)
60
72
Fig.3-4. During fed-batch, biomass density (♦) and sugars concentration (■)
changing with time. Two times of juice addition happened at the 48 hr and 60 hr.
30
Sugars conc. (g/L)
25
20
Fructose
Glucose
Sucrose
15
10
5
0
-5
0
12
24
36
48
60
72
Time (hr)
Fig.3-5. During fed-batch, fructose (♦), glucose (■) and sucrose (▲)
concentration changing with time. Two times of juice addition happened at the
48 hr and 60 hr.
54
Sweet sorghum juice has also been tested as a substrate for other oleaginous
microorganisms, such as microalga. When Chlorella protothecoides was cultivated in
a 120 hours test in flasks, a maximum cell dry weight of 5.1 g/L with the lipid content
of 52.5% was identified (Gao et al.,2010). Using a similar growth mode, microalga
Schizochytrium limacinum SR21 produced a biomass yield and lipid content as 9.4
g/L and 73.4% in 5 days , respectively (Liang et al.,2009) . Substrate inhibition was
not as obvious as to the growth of C.curvatus contrast with S. limacinum SR21,
furthermore, S. limacinum SR21 can not grow on sucrose, thus reduces its
applicability on sugar crops. Compared with other oleaginous species grown on sweet
sorghum juice, C. curvatus produced the highest biomass and cellular lipid
productivity of 7.87 g/L-day and 4.01 g/L-day, contrast to the 1.02 g/L-day and 0.53
g/L-day of Chlorella protothecoides, and 1.88 g/L-day and 1.38 g/L-day of
Schizochytrium limacinum SR21.Furthermore, based on the observations from the
fed-batch operations, the biomass DW could be significantly enhanced. Currently, we
are improving the uninterrupted substrate feeding process to keep a constant sugars
concentration in the fermentor to prevent substrate limitation on cell growth and lipid
accumulation. Further optimization of the fed-batch process is ongoing in our
laboratory.
3.3.2. Microwave assisted lipid extraction
In microwave-assisted extraction, heat and pressure are generated in a very short
period of time during microwave assisted extraction, and the rapid heating can cause
55
localized high temperature and pressure gradients that assist in cellular wall
degradation and enhanced mass transfer rates (Boldor et al., 2010). Furthermore, since
the diffusion of lipids across the cell wall is affected by solvent properties, this
parameter and the selectivity of solvent can also determine the extraction efficiency.
Comparative analysis of the lipid extraction with different solvents showed
methanol had a higher yield (% of dry biomass w:w) than the others, as shown in
Table 3-1. This result was out of our expectation that methanol/chloroform mixture
would have higher lipid yield, as this mixture had the highest extracting efficiency in
the Bligh and Dyer method (Bligh and Dyer, 1959). The possible reason for this is the
boiling temperature difference of methanol and methanol/chloroform mixture:
methanol’s boiling point is 64.5 ◦C, which is higher than the methanol/chloroform
mixture’s 53.43 ◦C. Thus, in the microwave radiation process, cellular wall in
methanol can accumulate more localized heat than in methanol/chloroform mixture,
this difference leads to higher temperature and pressure gradients, which improve the
final lipid yield. On the other hand, though ethanol has a higher boiling point than
methanol (relative polarity 0.762), its lower polarity (relative polarity 0.654) leads to
a lower selectivity of yeast cell lipids (comprising of unsaturated fatty acids). In
addition, increasing microwave heating time did not increase lipid yield significantly
for using methanol as solvent. For using methanol/chloroform as solvent, when
heating time increased from 4 minutes to 20 minutes, the yield increased two folds,
but still lower than using methanol alone. The attempt of using distilled water as
solvent totally failed because of the difficulty of separating yeast biomass from water
56
layer. Compare with the traditional Bligh and Dyer method, microwave assisted
extraction has a lower lipid yield, however, considering the further biodiesel
production, it would appear to be the most simple, easy, and efficient method for lipid
extraction from yeast cells.
Table.3-1. Microwave assisted lipid extraction yields by using different solvents
Extracting solvent
Biomass/solvent ratio
(w:v)
1:200
1:200
chloroform/methanol 1:200
(2:1,v/v)
1:200
1:200
1:200
1:50
methanol
1:100
1:100
ethanol
1:100
distilled water
1:100
Heating time (min)
4
8
12
16
20
24
4
4
12
4
4
Lipid yield (%
to biomass)
15.72 ± 4.71
19.15 ± 2.33
26 ± 1.51
28.22
30.26
29.73
27.85
33.18 ±1.66
33.9 ± 2.20
17.52 ± 1.4
0
MAE on other oleaginous microorganisms has also been tested. Lee’s group
reported a lipid yield of 28.6% on microalga Botryococcus sp,about 10% yield on
microalga Chlorella vulgaris and about 10% yield on microalga Scenedesmus
sp
(Lee et al.,2010) . The lipid yield from our work is comparable to those of Lee’s, but
can still be improved by further optimization.
3.4. Conclusion
This study demonstrated that producing lipids from sweet sorghum juice through
57
yeast C. curvatus fermentation is feasible. In a very short time period, significant
amount of lipids can be generated, the extracted lipids may serve as feedstock for
biodiesel production. Compared with the classical solvent extraction methods, though
the microwave assisted extraction procedure had a lower extraction efficiency, for
further up-scale production, this method is quicker and less labor intensive.
58
CHAPTER 4
IN-SITU DIRECT TRANSESTERIFICATION OF THE YEAST CRYPTOCOCCUS
CURVATUS LIPID TO BIODIESEL BY USING MICROWAVE RADIATION
4.1. Introduction
The worldwide energy consumption continues to rise as populations grow, and
the accelerated CO2 release which is caused by using of fossil fuels now generally
accepted as a major factor of global warming . Biodiesel, which is the alkyl (usually
methyl and ethyl) esters of long chain fatty acids derived from vegetable oils or
animal fats, is generally recognized as a renewable and biodegradable fuel that can
replace petrodiesel. Biodiesel is synthesized by the chemical reaction of
transesterification in which the triglycerides in oils or fats react with short chain
alcohol such as methanol in the presence of an acidic or basic catalyst, yielding esters
of fatty acids and glycerol as major byproducts.
During recent years, microbial oils produced by various microorganisms have
been intensively investigated and are believed to be an alternative oil source for
sustaining biodiesel production. Among these microbial species, Cryptococcus
curvatus (ATCC 20509) is of great interest to the scientific community. When C.
curvatus is grown on cheap carbon sources like whey permeate (Ykema et al., 1988)
and other carbohydrate-rich agricultural or food processing wastes (Vega et al.,1988;
Bednarski et al., 1986), it is able to accumulate 60% of cell dry weight as lipids
(Ratledge,1991). The yeast oils produced by C. curvatus resemble plant seed oils like
59
palm oil (Davies, 1988) and can certainly serve as an excellent feedstock for biodiesel
production.
However, in the biodiesel producing process, releasing lipid from yeast cells is
relatively difficult due to the hindrance of rigid cell wall structure, thus it is an
inefficient method of using traditional mechanical crushing to extract lipid (Woo et al.,
2000). On the other hand, microwave assisted extraction (MAE) could be a better
choice due to the penetrating ability of microwave. Microwave irradiation has been
used in the past to extract oils from the biomass, soils and vegetable feedstock
(Barnabas et al., 1995; Kiss et al., 2000; Pan et al., 2002; Lucchesi et al.,2004; Li et
al.,2004). Heat and pressure are generated in a very short period of time during
microwave assisted extraction, and the rapid heating can cause localized high
temperature and pressure gradients that assist in cellular wall degradation and
enhanced mass transfer rates (Boldor et al., 2010).
Compare with conventional extraction and transesterification process, the “in
situ” method, which means oils extraction and transesterification happens
simultaneously, is considered to have the potential of reducing processing units and
fuel production costs. This process can directly convert cell oils to fatty acid methyl
esters (FAME) from the oleaginous biomass, thus skip the lipid extraction step in the
conventional method. In addition, the influence of the moisture content of the yeast
biomass on the conversion process needs to be studied since the avoidance of biomass
drying step could be of importance energetically and cost-wise. With the yeast
biomass containing up to 80% moisture content after culture centrifugation and
60
supernatant separation, this study also aimed to evaluate the extent to which moisture
containing yeast biomass can be effective for the in situ transesterification process in
order to reduce the biomass drying requirements.
In the current work we explored the optimization of biodiesel production from
the yeast Cryptococcus curvatus, set up a routine of in-situ direct transesterification
from the crude wet yeast pellet using microwave radiation.
4.2. Materials and Methods
4.2.1. Yeast culture and biomass preparation
C. curvatus (ATCC 20509) grown in liquid medium containing 2% peptone, 1% yeast
extract, and 16 g/L of pure glycerol was utilized as an inoculum for all experiments
described in this study adopting a minimal medium. The minimal medium contained
(per liter): 2.7 g KH2PO4; 0.95 g Na2HPO4; 0.2 g MgSO4 ·7H2O; 0.1 g yeast extract;
and 0.1 g EDTA. After the pH was adjusted to 5.5, it was supplemented with a 100x
spores stock solution consisting of (per liter): 4 g CaCl2 ·2H2O; 0.55 g FeSO4.7H2O;
0.52 g citric acid; 0.10 g ZnSO4.7H2O; 0.076 g MnSO4.H2O; and 100 µL 18 M H2SO4.
Nitrogen was supplied in the form of NH4Cl at 2.5 g/L with an initial C/N ratio of
30:1. The medium was autoclaved at 121◦C for 15 min before use. The C.curvatus
culture was maintained at room temperature on a rotary shaker (Excella E24,New
Brunswick Scientific, Enfield,CT,USA) set at 150 rpm.
4.2.2 Transesterification process
61
A domestic microwave oven with exiting power of 900W was modified for this
purpose. The roof of the oven was drilled with three holes to pass through a digital
thermocouple, a water-cooled reflux condenser and a motor driven stirring bar, which
could ensure uniform mixing of the reaction mixture. Microwave-transparent, 100 mL
three-neck round bottomed flasks were used as sample vessels. Test results obtained
from the average of duplicate tests for each run were analyzed to evaluate the
reproducibility of microwave effect. After each test, a time interval was spent to let
the reactor cool down and return to original conditions.
Freeze dried C. curvatus pellet was grinded to powder and passed through 500
µm mesh, then 0.5g powder was mixed with distilled water, methanol and KOH
catalyst. The reaction mixture was heated using microwave irradiation, meanwhile the
stirring bar stirred by different rotating speed, for different time interval from 2 to 10
minutes. After the completion of the reaction, the samples were centrifuged at 4500g
for 5 min and filtered by a 0.45 µm filter to separate the methanol phase that
contained the biodiesel from the yeast powder, glycerol and catalyst. The solution of
methanol was transferred into a 50 mL round-bottom flask and evaporated methanol
in a rotovap. The remaining products were taken in hexane and then centrifuged at
4500g for 3 min, the upper layer was transferred to a pre-weighed glass tube, hexane
was evaporated by nitrogen and the mass of biodiesel was determined gravimetrically.
After this, an internal standard, methyl heptadecanoate (C17:0) was added
quantitatively to the hexane redissolved biodiesel and analyzed by gas
chromatography. During the microwave radiation process the temperature was
62
measured by placing a digital thermocouple (Omega 2160A, Omega Engineering
Inc.,Stamford,CT,USA) in the reaction flask. The temperature was found to be 60~80
o
C for the direct transesterification conducted by microwave.
4.2.3. Screening test experimental design
The Plackett–Burman design (Plackett and Burman, 1946) was used to screen
significant factors with respect to their effects on the biodiesel yield, five factors were
selected and further screened. Each variable was represented at two levels, i.e., high
(+) and low (-).The lowest and the highest level of each factor are given along with
the design in Table 4-1. The method was developed using Design Expert System
(version 7.1.6, Stat-Ease Inc, Minneapolis, MN 55413, USA). According to the
Plackett–Burman design, twelve trials were performed with the biodiesel yield (% to
yeast lipid content) as the response are showed in Table 4-2.
Table 4-1.
Variables and their levels employed in the Plackett–Burman design.
Factor
1
Time
2 Solvent/Biomass ratio
3 Stirring speed
4 Catalyst
5 Moisture content
Unit
minute
mL/g
rpm
%
%
Low level
2
10
450
1
80
High level
10
50
1000
5
300
63
Table 4-2. Plackett–Burman experimental design matrix and the response values
Run
1
2
3
4
5
6
7
8
9
10
11
12
Factor 1
2
2
2
10
10
10
10
2
2
2
10
10
Factor 2
10
50
10
10
50
50
10
50
10
50
10
50
Factor 3
450
1000
450
1000
450
450
450
450
1000
1000
1000
1000
Factor 4
1
1
5
5
5
1
1
5
1
5
5
1
Factor 5
80
300
80
80
300
80
300
300
300
80
300
80
Biodiesel Yield %
9.06
11.76
6.25
19.72
15.16
16.33
3.54
17.29
6.04
56.04
0.80
17.71
4.2.4. Optimization experimental design
After the factors that could significantly influence extraction efficiency were
obtained, response surface methodology (RSM) using the Box–Behnken factorial
design (Box and Behnken, 1960) was applied to further develop mathematical
correlations between three independent variables on biodiesel yield. This design has
been used to examine the relationship between one or more response variables and a
set of quantitative experimental parameters based on response surface methodology.
Three variables, solvent/biomass ratio (15-50, v:m), stirring speed (450-1000, rpm),
and catalyst/biomass ratio (1-5, %) were selected since they were determined as
having the most significant effects on biodiesel yield by previous screening test.
Biodiesel yield was set as analytical responses. A three-factor and three-level
experiment was designed using the Design-Expert (Stat-Ease Inc. Minneapolis, MN,
USA) program. Each variable was tested in three different coded levels: low (−1),
middle (0), and high (+1) (Table 4-3). Based on experimental results, a second-order
64
polynomial model for the variables was obtained:
Y=β0+Σβixi+Σβixi2+Σβijxixj
where Y is the predicted response, β is the coefficient of the equation, and xi and xj are
the coded levels of variables i and j, respectively. The statistical analysis of the model
was performed in the form of analysis of variance (ANOVA), the second-order model
equation and significance of variables were determined by Fisher’s F-test. This design
consists of replicated center points and the set of points lying at the midpoints of each
edge of the multidimensional cube that defines the region of interest.
Table 4-3. Process variables and their levels used in the design
FACTORS
A—Solvent/biomass
ratio., X1
B—Stirring speed. rpm,
X2
C—Catalyst. %, X3
LEVELS
-1
15
0
32.5
1
50
450
725
1000
1
3
5
4.2.5. FAME analysis
For the quantification of reaction product, the resultant biodiesel were separated
using a Shimadzu GC-17A gas chromatograph (Shimadzu Scientific Instruments,
Columbia, MD, USA) equipped with a flame ionization detector (FID) fitted with a
permanently bonded polyethylene glycol, fused silica capillary column (Omegawax
250, 30 m × 0.25 mm i.d., 0.25 µm film). The injection volume was 1.0 µL, helium
was the carrier gas (30 cm/s, 205 ◦C), and the injector temperature was 250◦C. A split
injection technique (100:1) was used, and the temperature program was as follows: 50
65
◦
C held for 2 min, increased to 220 ◦C at 4 ◦C /min, and held at 220 ◦C for 15 min.
Individual FAME was identified by reference to external standards (Supelco 37
Component FAME Mix, PUFA-1, and PUFA-3; Supelco, Bellefonte, PA,USA). All
solvents used were of HPLC grade and obtained from Sigma Diagnostics Inc. (St.
Louis, MO, USA).
4.3. Results and discussion
4.3.1 Significant parameters screening by Plackett–Burman design
In the Plackett–Burman design, five factors that influenced the transesterification
reaction, i.e., heating time, solvent/biomass ratio (v:w), stirring speed, catalyst
percentage (w:w), and yeast powder moisture content (w:w) were selected.
The experiment results obtained from the Plackett–Burman design experiments
were utilized to generate half-normal probability plot, which is a graphical tool to test
for significant versus unimportant factors (Filliben, 2002). Typically, these plots are
only partially linear. Unimportant (i.e., near zero) effects manifest themselves as
being near zero forming nearly a straight line while important (i.e., large) effects
manifest themselves by being off the line and well-displaced from zero (Mohanty,
2003). In the present investigation, as illustrated in the half-normal probability plots in
Fig. 4-1, B (solvent/biomass ratio), C (stirring speed),D (catalyst percentage), and E
(moisture content) were found to have significant effect on biodiesel yield response.
On the contrary, A (heating time) is the least important one in the range of values
tested.
66
Fig.4-1. The statistical plots identifying the key operating parameters for biodiesel
yield.
Since biomass drying has been demonstrated to be one of the most important
economic steps in the microalgae production process, according a report that the cost
of drying could be up to 30% of the total production costs (Becker, 1994). With the
yeast biomass pellet containing an average of 80% moisture content after the
harvesting process via centrifugation and supernatant separation, we arranged the
moisture content of this study to be fixed at 80% level. Therefore, the biomass drying
step can be skipped and the total cost reduced as well.
According to the evaluations above, three factors B (solvent/biomass ratio), C
(stirring speed), and D (catalyst concentration) were selected for a further study of
Box–Behnken factorial design. The other two factors were set to be fixed at 2 minutes
67
for heating time and 80% for moisture content in subsequent experiments.
4.3.2 Development of a regression model
The experimental runs and results for the Box–Behnken design are shown in
Table 4-4. The 17 runs in a single block were used to study the effects of three factors
on the response of biodiesel yield. For all combinations tested, crude biodiesel yield
varied from 8.28% to 53.56%.
The application of response surface methodology yielded the following
regression equation models which are empirical relationships between the biodiesel
yield and the test variables in coded units. The relation among the variables (as coded
values) solvent/biomass ratio (X1), stirring speed (X2), and percentage of catalyst (X3)
was fitted by second-order polynomial equations (1):
Biodiesel yield =23.05+11.15X1+1.19X2+6.29 X3+4.60 X1 X2+8.56 X1 X3+3.49X2 X3
(1)
Analysis of variance (ANOVA) is required to test the significance and
adequacy of the model. ANOVA for response surface quadratic model for the
biodiesel yield is indicated in Table 4-5. The regression model accurately described
the experimental data, which indicated successful correlation among the three
transesterification process parameters that affected biodiesel yield. This was
supported by the values of correlation coefficients R2 as 0.93 for biodiesel yield. The
R value suggested a satisfactory representation of the process model and a good
correlation between the experimental results and the theoretical values predicted by
68
the model equation. The P value of the model of less than 0.0001 indicated the
significance of the coefficients. The Lack of Fit F-value of 3.98 implied the lacks of
fit were not significant relative to the pure error. The response surfaces were fitted
using process variables that were found to be significant after the analysis. With the
established models, different combinations of variables (solvent/biomass ratio,
stirring speed, and percentage of catalyst) were able to lead to desired yields of
biodiesel or result in maximal biodiesel production.
Table 4-4. The Box-Behnken design of the variables with biodiesel yield as response
Runs
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Solvent/biomass Stirring
ratio
speed
(rpm)
50
725
32.5
1000
15
450
32.5
725
50
450
32.5
725
15
725
15
1000
32.5
725
32.5
725
15
725
32.5
1000
50
725
50
1000
32.5
450
32.5
450
32.5
725
Catalyst
(%)
Biodiesel
yield (%)
1
5
3
3
3
3
5
3
3
3
1
1
5
3
5
1
3
22.53
33.71
19.64
21.08
26.87
23.27
8.28
12.37
21.45
22.56
11.49
15.49
53.56
38.02
23.92
19.67
18.02
4.3.3. Effects of process parameters on optimization
From the variance analysis, it could be concluded that the solvent/biomass ratio
69
and catalyst percentage had more significant effects on biodiesel yield than those
from stirring speed. Three-dimensional surface responses were plotted to illustrate the
relationships between the responses and variables (Fig. 4-2). As shown by this figure,
when stirring speed was fixed at 725 rpm, with the increase of solvent/biomass ratio,
the increase of catalyst percentage led to an increase in biodiesel yield; when
solvent/biomass ratio was fixed at 32.5, with the increase of stirring speed,the
increase of catalyst percentage led to an increase in biodiesel yield as well.
70
Fig 4-2. Three-dimensional response surface plot of biodiesel yield as a function
of different parameters.
With regard to biodiesel yield, four effects had P-values less than 0.05,
indicating that they were significantly different from zero at the 95% confidence level
(Table 4-5). These effects were the solvent/biomass ratio, the catalyst percentage, the
71
interaction between solvent/biomass ratio and stirring speed, and the interaction
between solvent/biomass ratio and catalyst percentage. Considering the F-ratio
statistic, it might be concluded that a change in the solvent/biomass ratio caused the
major variation in biodiesel yield. This can be explained that the solvent (methanol)
was a good microwave radiation absorption material, in the in-situ reaction process,
methanol acted both as a yeast lipids extraction solvent (Mulbry et al., 2009) as well
as the reactant for tranesterification of esters. During the microwave heating process,
the dipole of methanol molecules quickly varied under the microwave irradiation,
which broke the two-tier structure of the interface of methanol and lipid extracted
from yeast cells (Patil et al.,2012). Thus, the transesterification reaction efficiency
increased due to the improved solubility of methanol and lipid under microwave
irradiation (Yuan et al., 2009). A higher methanol to biomass ratio could get higher
yield of biodiesel, which was indicated by experimental results. The possible reason
is that the higher ratio increased the contact area between methanol and lipid, and
induced the reversible reaction to move forward.
72
Table 4-5. Analysis of variance (ANOVA)
Source
Model
X1
X2
X3
X1X2
X1X3
X2X3
Residual
Lack of fit
Pure Error
Cor Total
Degree
of
freedom
6
1
1
1
1
1
1
10
6
4
16
Biodiesel yield
F-value P-value
25.65
87.54
0.99
27.82
7.47
25.80
4.29
<0.0001
<0.0001
0.3430
0.0004
0.0211
0.0005
0.0650
3.98
0.1011
In this study, our aim was to maximize the biodiesel yield by finding optimal
conditions. According to the models identified above, the optimal conditions for
achieving maximal biodiesel yield were solvent/biomass ratio 49.8,stirring speed 966
rpm and catalyst percentage 4.99% . All of the optimal parameters were verified by
comparing the experimental data obtained under these conditions with the predicted
numbers. The verification experiment provided a biodiesel yield as 56.1% , which
was very close to the predicted value 56.8%,
the small deviations between the
experimental and predicted data suggested that the experimental designs used in this
work were effective for accomplishing our purpose. We performed a second time
transesterification using the yeast residuals of first time opreation, and found that the
total biodiesel yield can be as high as 92%. Therefore, the optimal solvent/biomass
ratio, stirring speed, and catalyst percentage acquired from this study are valuable for
73
our future efforts toward scale up process design.
Compare with the conventional biodiesel production procedure, the microwave
assisted direct transesterification (MAT) method decreases the handling time, labor
intensity and reduces the chemical and energy requirements. The conventional
methanol/chloroform mixture as solvent requires 6 h for biomass drying, 12 h for lipid
extraction and 12h of heating and mixing for transesterification (Liang et al.,2010).
On the contrary, the MAT method only needs 4 to 5 minutes of microwave radiation
time direct from the wet biomass of cultivation. This is possible with MAT method
due to its ability to heat the reaction mixture more rapidly than the conventional
method.
4.3.4. Analysis of yeast biodiesel
The content of the FAME in the biodiesel was calculated quantitatively by
comparing the peak areas of fatty acid methyl esters to the peak area of the internal
standard (methyl heptadecanoate, C17:0) obtained from GC analysis. From the GC
chromatogram, as shown in Table 4-6, it can be noted that this yeast biodiesel
contains a high percentage of mono unsaturated fatty acid methyl esters, which is
fitted for the requirements of ideal biodiesel, such as freezing point, oxidative stability,
octane number, and NOx emissions (Tyson et al., 2004). The total FAME yield is
35.84±0.61% of the yeast lipid (w:w), and the FAME content in biodiesel is 49.97%
(w:w). Other researching groups also reported biodiesel production using in situ direct
transesterification on oleaginous microorganisms. Johnson et al. (2009) reported
a
74
42.05% (of biomass) of FAME yield on dry microalga biomass Schizochytrium
limacinum, but with wet biomass, the yield decreased drastically to 5.20%. Koberg et
al. (2011) reported a biodiesel yield of 37.1% and a very high lipid to biodiesel
conversion of 99.9% on on dry microalga biomass Nannochloropsis sp by MAT . Patil
et al. (2012) also reported using MAT on dry microalga biomass Nannochloropsis
salina and obtained a FAME yield of 40.03% of dry biomass. The FAME yield from
our work is lower than those did direct transesterification on dry biomass, but much
higher than on wet biomass. Considerring about the labor and economic factors in
biofuel production, skipping wet biomass dessication would be a beneficial shortcut.
Table 4-6. GC peak of crude biodiesel obtained from yeast biomass
Peak
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Retention time (min)
21.367
24.185
27.013
27.570
29.915
29.995
32.016
32.526
33.525
34.196
34.909
35.474
36.559
37.086
40.201
42.636
42.947
53.357
54.172
54.548
Area %
0.868
0.1
21.625
1.325
0.047
0.075
12.181
45.223
13.435
0.091
0.327
0.593
0.321
0.172
0.039
0.059
0.255
1.102
0.107
0.044
FAME
C14:0
C15:0
C16:0
C16:1n-7
C16:3n-4
C17:1
C18:0
C18:1n-7
C18:2n-6
C18:3n-6
C18:3n-3
C18:4n-3
C20:0
C20:1n-9
C20:4n-6
C20:5n-3
C22:0
C24:0
C22:6n-3
24:1n-9
75
4.4. Conclusion
The in-situ direct transesterification of yeast biomass lipid by microwave heating was
demonstrated for optimum reaction conditions using a response surface methodology.
This process being a fast and easy method was effective to produce biodiesel from
wet yeast biomass. The one-step direct transesterification process has the potential to
provide energy efficient and economical route for the biodiesel production from yeast
C. curvatus.
76
CHAPTER 5
FUTURE RESEARCH
Further research on the subsequent disposal of yeast extraction residuals such as
the conversion to bio-oil or biogas, the scale up of fermentation and transesterification
process design need to be conducted to obtain more advantages from this yeast strain.
5.1 Conversion to bio-oil or syngas
The thermochemical conversion of biomass (pyrolysis, liquefaction, gasification)
is a promising, non-nuclear form of future energy. Bio-oil and biogas has several
environmental advantages over fossil fuels as a clean fuel. Bio-oils are CO2/GHG
neutral. Therefore, they can generate carbon dioxide credits. Bio-oil fuels generate
more than 50% lower NOx emissions than diesel oil in a gas turbine (Mohan et al.,
2005). In our previous study, the lipid extraction residuals of yeast still have high
percentage of cellulose and protein left, thus they are a rich resource of carbon and
hydrogen. How to make use of this resource can be a project of our future research, by
using efficient catalyst and conversion method, the yeast extraction residuals can be
completely converted to high energy density biofuels.
5.2 Fermentation and transesterification scale up
We have accomplished the bench scale experiments of yeast fermentation and
in-situ transesterification, and obtained optimum parameters for these processes. The
77
future work of these parts will be the scale up study, many problems such as oxygen
transfer, heat transfer, mixing and byproduct formation may happen during
fermentation scale up process. On the other hand, mixing, heat transfer and reaction
kinetics need to be studied in the MAT scale up process. The economical analysis for
the whole routine needs to be done as well to promote the commercial application of
C. curvatus.
78
REFERENCES
Barnabas,I., Dean,J., Fowlis,I., Owen,S., 1995. Extraction
of polycyclic
hydrocarbons from highly contaminated soils using microwave energy.
Analyst. 120, 1897–1904.
Becker,E.,1994. Microalgae biotechnology and microbiology. Cambridge University
Press,Cambridge, United Kingdom.
Bednarski,W., Leman, J., Tomasik, J., 1986. Utilization of beet molasses and whey for
fat biosynthesis by a yeast. Agric. Wastes. 18, 19-26.
Bligh,E., Dyer,W.,1959. A rapid method of total lipid extraction and purification.
Canadian Journal of Biochem. & Physiol. 37, 911-917
Boldor,D.,
Kanitkar,A.,
Terigar,B., Leonardi,C., Lima,M., Breitenbeck,G.,2010.
Microwave Assisted Extraction of Biodiesel Feedstock from the Seeds of
Invasive Chinese Tallow Tree. Environ. Sci. Technol. 44,4019–4025
Boulton,C., Ratledge,C.,1981.Correlation of lipid accumulation in yeasts with
possession of ATP: citrate lyase. Microbiology.127,169-176
Box, G., Behnken, D., 1960. Some new three level designs for the study of
quantitative variables. Technometrics. 2, 455–475.
Bulawayo,B., Bvochora,J., Muzondo,M., Zvauya,R., 1996.Ethanol production by
fermentation of sweet-stem sorghum juice using various yeast strains.World
Journal of Microbiol. & Biotechnol. 12, 357-360
79
Chi,Z., Pyle, D., Wen, Z., Frear, C., Chen, S., 2007. A laboratory study of producing
docosahexaenoic acid from biodiesel-waste glycerol by microalgal
fermentation. Process Biochem. 42, 1537–1545.
Chen,Y., Walker, T., 2011. Biomass and lipid production of heterotrophic microalgae
Chlorella protothecoides by using biodiesel-derived crude glycerol. Biotechnol.
Lett. 33, 1973-1983.
Christie,W., 1982. Lipid Analysis, second ed. Pergamon, Oxford.
Daniel,H., Otto,R., Binder,M., Reuss,M., Syldatk,C., 1999. Production of
sophorolipids from whey: development of a two-stage process with
Cryptococcus curvatus ATCC 20509 and Candida bombicola ATCC 22214 using
deproteinized whey concentrates as substrates. Appl. Microbiol. Biotechnol. 51,
40–45.
Davies, R., 1988. Yeast oil from cheese whey; process development. In: Moreton, R.S.
(Ed.), Single Cell Oil. Longman, London, pp. 99–145.
Ethier, S., Woisard, K., Vaughan, D., Wen, Z., 2011. Continuous culture of the
microalgae Schizochytrium limacinum on biodiesel-derived crude glycerol for
producing docosahexaenoic acid. Bioresour. Technol. 102, 88-93.
Evans,C., Ratledge,C., 1983. A comparison of the oleaginous yeast, Candida curvata,
grown on different carbon sources in continuous and batch culture. Lipids. 18,
623–629.
Filliben,J., 2002.
NIST/SEMATECH e-handbook of statistical methods.
www.itl.nist.gov/div898/handbook/
80
Gao,C., Zhai,Y., Ding,Y., Wu,Q., 2010.
biodiesel
production
by
Application of sweet sorghum for
heterotrophic
microalga
Chlorella
protothecoides.Applied energy.87,756-761
Glatz,A., Hammond,G., Hsu, H., Baehman,L., Bati,M., Bednarski,W., Brown,D.,
Floetenmeyer,M., 1984. Production and modification of fats and oils
by yeast fermentation. In: Ratledge, C., Danson, P., Rattray, J. (Eds.),
Biotechnology for the Oils and Fats Industry. American Oil Chemists’
Society, Champaign, IL. pp. 163–176.
Gnansounou, E., Dauriat, A., Wyman, C.E., 2005. Refining sweet sorghum to ethanol
and
sugar:
economic
trade-offs
in
the
context
of
North
China.
Bioresour.Technol. 96, 985–1002.
Gonzalez-Pajuelo,M., Meynial-Salles,I., Mendes,F., Andrade,J., Vasconcelos,I.,
Soucaille,P., 2005. Metabolic engineering of Clostridium acetobutylicum for
the industrial production of 1,3-propanediol from glycerol. Metabol. Eng. 7,
329–336.
Haas,M., McAloon,A., Yee,W., Foglia,T., 2005. A process model to estimate biodiesel
production costs. Bioresour. Technol. 97, 671–678
Hassan,M., Blanc,P., Granger,L., Pareilleux, A., Goma, G., 1996. Influence of nitrogen
and iron limitations on lipid production by Cryptococcus curvatus grown in
batch and fed-batch culture. Process Biochem. 31, 355–361.
81
Hoffert, M., Caldeira,K., Benford,G., Criswell.D., Green,C., Herzog,H., Jain,A.,
Kheshgi,H., Lackner,K., Lewis,J., Lightfoot,H., Manheimer,W., Mankins,J.,
Mauel,M., Perkins,L., Schlesinger,M., Volk,T., Wigley,T., 2002. Advanced
technology paths to global climate stability: energy for a greenhouse planet.
Science. 298, 981-987.
Houghton J., Ding,Y., Griggs,D., Noguer,M., Dai,X., Maskell,K., Johnson,C., 2001.
Climate change 2001: The scientific basis. contribution of working group I to
the third assessment report of the Intergovernmental Panel on Climate Change.
Cambridge University Press, Cambridge, United Kingdom.
Johnson,M., Wen, Z., 2009. Production of biodiesel from the microalga
Schizochytrium limacinum by direct transesterification of algal biomass.
Energy Fuels. 23, 5179–5183.
Karatay, S., Dönmez,G., 2010. Improving the lipid accumulation properties of the
yeast cells for biodiesel production using molasses. Bioresour. Technol. 101,
7988-7990.
Kiss, G., Forgacs, E., Cserhati, T., Mota, T., Morais, H., Ramos, A., 2000.
Optimization of the microwave-assisted extraction of pigments from paprika
(Capsicum annuum L.) powders. J. Chromatogr. A 889, 41–49.
Koberg,M., Cohen,M., Ben-Amotz,A., Gedanken,A., 2011.Bio-diesel production
directly from the microalgae biomass of Nannochloropsis by microwave and
ultrasound radiation. Bioresour. Technol.102,4265-4269.
82
Laopaiboon,L., Thanonkeo,P., Jaisil,P., Laopaiboon,P., 2007.Ethanol production from
sweet sorghum juice in batch and fed-batch fermentations by Saccharomyces
cerevisiae. World Journal of Microbiol. & Biotechnol.23,1497-1501
Lee,J., Yoo,C., Jun,S., Ahn,C., Oh,H.,2010.
Comparison of several methods for
effective lipid extraction from microalgae. Bioresour. Technol. 101, 75–77
Li, H., Pordesimo,L., Weiss,J., Wilhelm,L., 2004. Microwave and ultrasound assisted
extraction of soybean oil. Trans. ASAE. 47, 1187–1194.
Li,Q., Wang,M., 1997. Use food industry waste to produce microbial oil. Sci. Technol.
Food Indust. 6, 65–69.
Li,Q., Du,W., Liu,D., 2008. Perspectives of microbial oils for biodiesel
production.Appl. Microbiol. Biotechnol. 80, 749–756.
Liang, Y., Cui, Y., Trushenski, J., Blackburn, J., 2010a. Converting crude glycerol
derived from yellow grease to lipids through yeast fermentation. Bioresour.
Technol. 101, 7581-7586.
Liang, Y., Sarkany, N., Cui, Y., 2009. Biomass and lipid productivities of Chlorella
vulgaris under autotrophic, heterotrophic and mixotrophic growth
conditions.Biotechnol. Letters. 31, 1043–1049.
Liang,Y., Sarkany, N., Cui, Y., Blackburn, J., 2010b. Batch stage study of lipid
production from crude glycerol derived from yellow grease or animal fats
through microalgal fermentation. Bioresour. Technol. 101, 6745-6750.
83
Liang,Y., Sarkany, N., Yesuf, J., Cui,Y., Trushenski, J., Blackburn, J., 2010. Use of
sweet sorghum juice for lipid production by Schizochytrium limacinum SR21
Bioresour. Technol. 101, 3623–3627.
Liu, R., Shen, F., 2008. Impacts ofmain factors on bioethanol fermentation from stalk
juice of sweet sorghum by immobilized Saccharomyces cerevisiae (CICC 1308).
Bioresour. Technol. 99, 847–854.
Liu, R., Li, J., Shen, F., 2008. Refining bioethanol from stalk juice of sweet sorghum
by immobilized yeast fermentation. Renewable Energy. 33, 1130–1135.
Lucchesi,M., Chemat,F., Smadja,J., 2004. Solvent-free microwave extraction of
essential
oil
from
aromatic
herbs:
comparison
with
conventional
hydro-distillation. J. Chromatogr. A 1043, 323–327
Mamma,D., Christakopoulos,P., Koullas,D., Kekos,D., Macris,B., Koukios,E., 1995.
An alternative approach to the bioconversion of sweet sorghum carbohydrates to
ethanol. Biomass and Bioenergy. 8, 99–103.
Meesters,P., Huijberts,G., Eggink,G., 1996. High cell density cultivation of the lipid
accumulating yeast Cryptococcus curvatus using glycerol as a carbon
source. Appl. Microbiol. Biotechnol. 45, 575–579.
Mohan D., Pittman C., Steele P., 2006. Pyrolysis of wood/biomass for bio-oil: A
critical review. Energy Fuels. 20,848–889
Mohanty,M.,2003. Fine coal screening performance enhancement using the Pansep
screen.Int. J. Miner. Process. 69, 205–220
84
Moon,N., Hammond, E., Glatz, B., 1978. Conversion of cheese whey and whey
permeate to oil and single-cell protein. J. Dairy Sci. 61, 1537–1547.
Mulbry,W., Kondrad,S., Buyer, J., Luthria, D., 2009. Optimization of an oil extraction
process for algae from the treatment of manure effluent.Journal of the
American Oil Chemists' Society. 86,909-915
Mu,Y., Teng, H., Zhang, D., Wang, W., Xiu, Z., 2006. Microbial production of
1,3-propanediol by Klebsiella pneumoniae using crude glycerol from biodiesel
preparations. Biotechnol. Lett. 28, 1755–1759.
Mu, Y., Xiu, Z., Zhang, D., 2008. A combined bioprocess of biodiesel production by
lipase with microbial production of 1,3-propanediol by Klebsiella pneumoniae.
Biochem. Eng. J. 40, 537–541.
Obayori,O., Adebusoye,S., Ilori,M., Oyetibo,G., Omotayo,A., Amund,O., 2010.
Effects of corn steep liquor on growth rate and pyrene degradation by
Pseudomonas strains. Curr. Microbiol.60, 407-411
Pan,X., Niu,G., Liu,H., 2002. Comparison of microwave-assisted extraction and
conventional extraction techniques for the extraction of tanshinones from
Salvia miltiorrhiza bung. Biochem. Eng. J. 12, 71–77
Patil,P., Reddy,H., Muppaneni,T., Mannarswamy,A., Schuab,T., Holguin,F.,
Lammers,P., Nirmalakhandan,N., Cooke,P., Deng,S.,2012.Power dissipation in
microwave-enhanced in situ transesterification of algal biomass to biodiesel.
Green Chem. 14,809-818.
85
Papanikolaou,S., Aggelis, G., 2002. Lipid production by Yarrowia lipolytica growing
on industrial glycerol in a single-stage continuous culture. Bioresour.
Technol.82, 43–49.
Papanikolaoua, S., Fakasb, S., Ficka, M., Chevalota, I., Galiotou-Panayotoub,
M.,Komaitisb, M., Marca, I., Aggelisc, G., 2008. Biotechnological valorisation
of raw glycerol discharged after bio-diesel (fatty acid methyl esters)
manufacturing process: production of 1,3-propanediol, citric acid and single
cell oil. Biomass and Bioenergy. 32, 60–71.
Plackett,R., Burman,J.,1946. The design of optimum multifactorial experiments.
Biometrika. 33, 305-325
Pyle, D., Garcia, R.A., Wen, Z., 2008. Producing docosahexaenoic acid (DHA)-rich
algae from biodiesel-derived crude glycerol: effects of impurities on DHA
production and algal biomass composition. J. Agric. Food Chem. 56,
3933–3939.
Ratledge, C., 1991. Microorganisms for lipids. Acta. Biotechnol. 11, 429-438.
Ratledge, C., Cohen, Z., 2008. Microbial and algal oils: do they have a future for
biodiesel or as commodity oils? Lipid Technol. 20, 155–160
Ratnavathi,C., Suresh,K., Kumar,B., Pallavi,M., Komala,V., Seetharama,N.,2010.
Study on genotypic variation for ethanol production from sweet sorghum
juice.Biomass and Bioenergy.34,947-952
86
Saenge,C., Cheirsilp,C., Suksaroge,T., Bourtoom,T., 2011. Potential use of oleaginous
red yeast Rhodotorula glutinis for the bioconversion of crude
glycerol from
biodiesel plant to lipids and carotenoids. Process Biochem. 46, 210-218.
Schmidt,R., Wiebe,M., Eriksen,N., 2005. Heterotrophic high cell-density fed-batch
cultures of the phycocyanin-producing red alga Galdieria sulphuraria.
Biotechnol Bioeng. 90, 77-84.
Sipos,B.,Réczey,J.,Somorai,Z.,Kádár,Z.,Dienes,D.,Réczey,K.,2009.Sweet Sorghum as
Feedstock for Ethanol Production: Enzymatic Hydrolysis of Steam-Pretreated
Bagasse.Applied Biochem.& Biotechnol. 153, 151-162
Stern, N., 2006. Stern Review on the Economics of Climate Change. HM Treasury,
London.
Thiru,M., Sankh,S., Rangaswamy,V., 2011.Process for biodiesel production from
Cryptococcus curvatus. Bioresour. Technol.102, 10436-10440
Tyson, K.S., Bozell, J.,Wallace, R., Petersen, E., Moens, L., 2004. Biomass Oil
Analysis: Research Needs and Recommendations. National Renewable Energy
Laboratory, NREL/TP-510-34796.
Vega, E., Glatz, B., Hammond, E., 1988. Optimization of banana juice fermentation
for the production of microbial oil. Appl. Environ. Microbiol. 54, 748-752.
Woo, I., Rhee, I., Park, H., 2000. Differential damage in bacterial cells by microwave
radiation on the basis of cell wall structure. Appl. Environ. Microbiol.
66,2243-2247
87
Ykema, A., Verbree, E.C., Kater, M.M., Smit, H., 1988. Optimization of lipid
production in the olgeaginou yeast Apiotrichum curvatum in whey permeate.Appl.
Microbiol. Biotechnol. 29, 211–218.
Young,J.,1995.Microwave-Assisted Extraction of the Fungal Metabolite Ergosterol
and Total Fatty Acids. J. Agric. Food Chem.43,2904-2910
Yuan, H., Yang, B., Zhu, G., 2009. Synthesis of biodiesel using microwave
absorption catalysts. Energy Fuels. 23, 548–552.
Yu,J., Zhang,X., Tan,T.,2009.
Optimization of media conditions for the production
of ethanol from sweet sorghum juice by immobilized Saccharomyces
cerevisiae.Biomass and Bioenergy.33,521-526
VITA
Graduate School
Southern Illinois University
Yi Cui
lfc4everfalcon@gmail.com
Xian University of Science & Technology, P.R.China
Bachelor of Engineering in Chemical Engineering, July 2002
Northwest University, P.R.China
Master of Science in Chemical Engineering, July 2007
Special Honors and Awards:
Graduate Registration Scholarship, 2004
Excellent Graduate Student of Northwest University, 2006
Dissertation Title:
BIODIESEL PRODUCTION THROUGH MICROWAVE ASSISTED
TRANSESTERIFICATION OF MICROBIAL CELLS
Major Professor: Yanna Liang
Publications:
Cui,Y., Blackburn,J., Liang,Y., 2012. Fermentation optimization for the production of lipid by
Cryptococcus curvatus: Use of response surface methodology. Biomass and Bioenergy.47:410417
Cui,Y., Blackburn,J., Liang,Y., 2010. Identifying the optimal parameters
for yeast fermentation on crude glycerol for lipid production. American Institute of
Chemical Engineers Annual Conference Proceedings,2010
Liang,Y., Cui,Y., Trushenski,J., Blackburn,J., 2010. Converting crude
glycerol derived from yellow grease to lipids through yeast fermentation. Bioresource
Technology. 101: 7581-7586
Cui,Y; Li,W;Tang,X, 2007. Technology study on extracting pigment from
cochineal insect. Chemical Engineering (China).35:66-69
Документ
Категория
Без категории
Просмотров
0
Размер файла
1 439 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа