Bioresource Technology 90 (2003) 95–100

Ethanol fermentation of acid-hydrolyzed cellulosic pyrolysate

with Saccharomyces cerevisiae
Zhisheng Yu, Hongxun Zhang *

Laboratory of Environmental Biotechnology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences,
P.O. Box 2871, Beijing 100085, PeopleÕs Republic of China
Received 11 December 2002; received in revised form 5 February 2003; accepted 28 February 2003

Abstract
The acid hydrolysis of cellulosic pyrolysate to glucose and its fermentation to ethanol were investigated. The maximum glucose
yield (17.4%) was obtained by the hydrolysis with 0.2 mol sulfuric acid per liter pyrolysate using autoclaving at 121 °C for 20 min.
The fermentation by Saccharomyces cerevisiae of a hydrolysate medium containing 31.6 g/l glucose gave 14.2 g/l ethanol in 24 h,
whereas the fermentation of the medium containing 31.6 g/l pure glucose gave 13.7 g/l ethanol in 18 h. The results showed that the
acid-hydrolyzed pyrolysate could be used for ethanol production. Different nitrogen sources were evaluated and the best ethanol
concentration (15.1 g/l) was achieved by single urea. S. cerevisiae (R) was obtained by adaptation of S. cerevisiae to the hydrolysate
medium for 12 times, and 40.2 g/l ethanol was produced by S. cerevisiae (R) in the fermentation with the hydrolysate medium
containing 95.8 g/l glucose, which was about 47% increase in ethanol production compared to its parent strain.
Ó 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Acid-hydrolysis; Cellulosic pyrolysate; Ethanol; Fermentation; S. cerevisiae

1. Introduction able sugar. However, the process has been hampered by

Cellulosic biomass is an alternate source of energy
because it is both renewable and available throughout
the globe in large quantities. Ethanol, as an alternate
energy, is one of the largest volumes that can be pro-
duced from cellulosic biomass. Nowadays, phasing out
of lead from gasoline due to environmental concerns has
promoted new applications and markets for alcohol as
octane enhancer. In this field, ethanol has gained in-
creased importance as engine fuels, since it is easily
blended with gasoline. Thus, with the expectation of
supply shortfalls in future from non-renewable fossil
fuels, the production of fermentatively produced bioal-
cohols from low-cost biomass such as cellulosic wastes
to meet energy demands is an attractive alternative.
Cellulosic biomass for ethanol production needs
pretreatments via liquefaction and saccharification.
Over the past decades, emphasis has been placed on
enzymatic hydrolysis of cellulosic biomass to ferment-
*
Corresponding author. Tel.: +86-10-62849155; fax: +86-10-
62923563.
E-mail address: hxzhang@mail.rcees.ac.cn (H. Zhang).
economic problems such as high costs of biomass pre-
treatment and cellulase production (Kumakura, 1997).
Advances in thermal processing of cellulose may offer a
new alternative for biomass pretreatment and sacchari-
fication (Shafizadeh and Stevenson, 1982; Piskorz et al.,
1989; Radlein et al., 1991). Cellulosic materials such as
pure cotton, woody and herbaceous biomass all can be
efficiently and rapidly converted into a high yield of
pyrolysate with levoglucosan (1,6-anhydro-b-D -gluco-
pyraloses) in high concentration under appropriate
pyrolysis conditions. The yields of levoglucosan range
from 38% to 58% in terms of the initial cellulose content
of woody feedstocks (Prosen et al., 1993). Recently,
Brown et al. (2001) reported that levoglucosan in 41–
72% pyrolytic yield is produced from switchgrass, an
herbaceous biomass. The levoglucosan has attracted
many investigators because its production and potential
use as a fermentative carbon and energy source in the
fermentation industry would lead to possible commer-
cial utilization of large quantities of cellulosic materials
(Nakagawa et al., 1984; Prosen et al., 1993; Zhuang et al.,
2001). Unfortunately, up to now, no microorganisms
have been found directly to convert the levoglucosan to

0960-8524/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-8524(03)00093-2
96 Z. Yu, H. Zhang / Bioresource Technology 90 (2003) 95–100
ethanol. Moveover, during the pyrolysis of cellulosic
materials, various toxic compounds, including aromatic
species, aldehydes, furan and furfuryl derivatives (Bonn,
1985), are formed, which inhibit the growth and fer-
mentation of microorganisms.
According to the report of Prosen et al. (1993), when
wood pyrolysate is hydrolyzed with sulfuric acid, levo-
glucosan is hydrolyzed to glucose and the toxic materi-
als are converted to inactive materials. Furthermore, the
acid-hydrolyzed pyrolysate is utilized by microorgan-
isms very well. This leads us to investigate the
application of sulfuric acid-hydrolyzed cellulosic pyrol-
ysate to produce substrates for fuel ethanol production.
In the present paper, we report the preliminary results
of ethanol production by a process that combined cell-
ulosic pyrolysis, acid hydrolysis and fermentation with
the yeast, Saccharomyces cerevisiae.
2. Methods
2.1. Preparation and hydrolysis of the pyrolysate
The pyrolysate was prepared by pyrolysis of waste
cotton with the method of Zhuang et al. (2001). About
70–80 g of highly viscous pyrolysate per 100 g dry waste
cotton was obtained from the pyrolysis reactor.
The viscous pyrolysate was diluted with four-fold
water, concentrated sulfuric acid was added, and the
mixture was hydrolyzed on a boiling waterbath for an
hour or by autoclaving at 121 °C for 20 min.
The hydrolysate obtained was neutralized to pH 6.0
with solid Ca(OH)2 and the formed precipitate was then
removed by filtration through a 0.45 micron membrane.
The filtrate was further diluted with distilled water be-
fore it was used as fermentable substrate. The glucose
concentrations used are given in the text.
2.2. Microorganism and medium
S. cerevisiae 2.399 was obtained from China General
Microbiological Culture Collection Center (CGMCC,
China). It was maintained on a medium containing 20.0
g/l glucose, 20.0 g/l peptone and 10.0 g/l yeast extracts at
4 °C, and subcultured every month at 30 °C.
The growth medium of the yeast consisted of 10.0 g/l
yeast extracts, 6.4 g/l urea, 2.0 g/l KH2 PO4 , 1.0 g/l
MgSO4
7H2 O and 20.0 g/l glucose at pH 5.5. The fer-
mentation medium of the yeast consisted of the hydro-
lysate containing 31.6 g/l glucose, 2.0 g/l KH2 PO4 , 1.0 g/l
MgSO4
7H2 O, 10.0 g/l yeast extracts and 6.4 g/l urea at
pH 5.5, unless otherwise mentioned. The adaptation
medium of the yeast was identical to the fermentation
medium except that the glucose concentration was 60.0
g/l. Mineral salts, urea and yeast extracts were auto-
claved separately at 121 °C for 30 min, and then added
to the hydrolysate, aseptically.
2.3. Inoculum preparation and adaptation and fermenta-
tion
S. cerevisiae was initiated in the maintenance medium
at 30 °C. The yeast was grown for 48 h at 150 rev min 1
on a rotary shaker at 30 °C. A 10% (v/v) inoculum was
used for subsequent subcultures.
S. cerevisiae, as a parent strain, was adapted to the
hydrolysate medium by semiaerobic cultivation for 24 h,
centrifuging and recycling the yeast to fresh hydrolysate
medium. The strain was recycled 12 times and referred
to here as S. cerevisiae (R).
Ethanol fermentations by S. cerevisiae and S. cere-
visiae (R) were evaluated at 30 °C in 150-ml Erlenmeyer
flasks having 100 ml appropriate media on a shaker at
150 rev min 1 . The flasks were sealed with a one-hole
rubber stopper, in which a glass tube was connected to
an air lock filled with 40% sulfuric acid solution.
2.4. Analytical methods
Samples were regularly taken from the fermentation
media. Glucose, levoglucosan and ethanol were ana-
lyzed on an HPLC system (GRE-3A, Shimadzu Cor-
poration, Kyoto, Japan) equipped with a Waters Model
401 refractive index detector and a Transgenomic ICSep
ICE-ORH-801 column (300  6.5 mm) (Transgenomic
Inc., San Jose, CA, USA). The mobile phase was 0.0025
N sulfuric acid at 0.6 ml/min, the injection volume was
10 ll and the column temperature was maintained at 30
°C. The known standards were used as controls. All
chemicals were of reagent grade and obtained from
commercial sources. Mean values of three separate ex-
periments and for three replicate flasks per experiment
are presented in the text. Variations of 4–7 and 65%
were seen between individual experiments and replicate
flasks, respectively.
3. Results and discussion
3.1. Acid hydrolysis of cellulosic pyrolysate
In our preliminary study, the pyrolysate was hydro-
lyzed with 0.3 mol/l H2 SO4 in a boiling waterbath for 1 h
or autoclaving at 120 °C for 20 min, respectively. The
results are shown in Fig. 1. By autoclaving at 120 °C for
20 min, over 100% of levoglucosan in the pyrolysate was
converted to glucose. This implied that besides levo-
glucosan, other components in the pyrolysate might have
a contribution to glucose. Bonn (1985) reported that
there existed small amounts of oligomers such as cello-
biosan, a dimeric form of levoglucosan in cellulosic
 Z. Yu, H. Zhang / Bioresource Technology 90 (2003) 95–100 97

18 atives) in the hydrolysate would inhibit the growth of

16 yeast, the dilution containing low concentration of glu-
14
cose (31.6 g/l) was firstly used for the investigation of the
Sugar (%)

12
course of ethanol fermentation by S. cerevisiae. Fur-
10
8
thermore, yeast extracts and urea were used as nitrogen
6 sources because they could stimulate the growth of yeast
4 and alleviate the inhibition of toxic materials to it
2 (Bafrncov
a et al., 1999). The time course of ethanol
0 production, residual glucose and pH changes was de-
control 1 sample 1 sample 2 control 2 termined (Figs. 3–5). A control culture was also carried
Fig. 1. Effects of boiling waterbath and autoclaving on hydrolysis of out with the medium containing 31.6 g/l of pure glucose.
cellulosic pyrolysate to glucose. Control 1 and control 2 were the Fig. 3 shows the time course of ethanol production
samples without adding acid, hydrolyzed with boiling waterbath for 1 using the hydrolysate as the substrate. The ethanol
h and autoclaving at 121 °C for 20 min, respectively. Sample 1 and production rate in the early phase of the culture was
sample 2 were the hydrolysates obtained with 0.3 mol/l H2 SO4 with
boiling waterbath for 1 h and autoclaving at 121 °C for 20 min,
relatively slow but rapidly increased after 12 h. Fer-
respectively, (j) levoglucosan and ( ) glucose. mentation was completed after 24 h and the final con-
centration of ethanol was 14.2 g/l. For the control, the
ethanol production rate increased rapidly after 6 h. The
fermentation was completed after 18 h and the final
concentration of ethanol was 13.7 g/l.
Fig. 4 shows the changes of residual glucose in each
fermentation. In the fermentation using the hydrolysate,
the glucose was exhausted after 24 h, whereas it was
exhausted after 18 h in the control. This indicated that
the glucose consumption was consistent with the time

Fig. 2. Effect of acid concentration on hydrolysis of cellulosic pyrol-

ysate to glucose by autoclaving at 121 °C for 20 min: (N) glucose and
(r) levoglucosan.

pyrolysate. In this experiment, the similar oligomers in

the pyrolysate might be converted to glucose. However,
with boiling waterbath for 1 h, only 81% of levoglucosan
was hydrolyzed to glucose. So, the hydrolysis efficiency
for the pyrolysate by autoclaving was higher than that
with boiling waterbath.
Fig. 3. Time course of ethanol production in the hydrolysate medium
To maximize glucose yield from the pyrolysate, sul- and control medium: (r) hydrolysate medium and (N) control me-
furic acid concentration was optimized by autoclaving at dium.
121 °C for 20 min. As shown in Fig. 2, the optimal acid
concentration for hydrolysis of the pyrolysate to glucose
was 0.2 mol/l, with which the maximal glucose yield
(17.4%) was obtained. A further increase in acid con-
centration (higher than 0.2 mol/l) decreased glucose
yield (lower than 17.4%). This suggested that small
amounts of glucose in the solution were not stable in the
presence of more sulfuric acid when autoclaving at 121
°C for 20 min.

3.2. Ethanol production from the hydrolysate by S.

cerevisiae

Due to the concern that the residual toxic materials Fig. 4. Changes of glucose in the hydrolysate medium and control
(aromatic species, aldehydes, furan and furfuryl deriv- medium: (r) hydrolysate medium and (N) control medium.
98 Z. Yu, H. Zhang / Bioresource Technology 90 (2003) 95–100

5.6 growth in lignocellulosic hydrolysates is strongly de-

5.5 pendent on pH due to the large concentration of dis-
5.4
5.3 sociated weak acids at low pH. In studies to determine
5.2 the optimal pH for batch fermentation of spruce enzy-
pH

5.1
5
matic hydrolysate, no cell growth at pH 4.6 was ob-
4.9 served during 17 h. However, when the pH was adjusted
4.8 to 5.0, both cell growth and ethanol productivity in-
4.7
0 3 6 9 12 15 18 21 24 27 30 creased. Additionally, Roberto et al. (1996) also ob-
Time (h) served a lag phase during cultivation of Candida
guilliermondii in rice straw hydrolysate at pH 4.5, at-
Fig. 5. Changes of pH in hydrolysate medium and control medium:
(r) hydrolysate medium and (N) control medium. tributed to the inhibitory effect of acetic acid, but this
was overcome by increasing initial pH to 5.3 or 6.0.

period of ethanol production. In spite of the complex

3.3. Effects of nitrogen sources
property of the substrate derived from the hydrolysate,
no strong diauxie pattern of sugar utilization was ob-
Nitrogen is necessary for the growth and multiplica-
served here. In the fermentation of the hydrolysate, the
tion of yeasts and it also influences the ethanol tolerance
productivity of ethanol was 0.59 g/l/h, lower than that of
of yeasts and ethanol productivity (Bafrncov
a et al.,
the control (0.76 g/l/h), which suggested that the toxic
1999). In traditionally industrial ethanol fermentation,
materials in the hydrolysate retarded the ethanol pro-
(NH4 )2 SO4 and urea are usually used as nitrogen sour-
duction. The ethanol yield in the fermentation of the
ces and yeast extracts are occasionally used as stimula-
hydrolysate was 0.45 g ethanol g 1 glucose, which cor-
tive factor for the growth of the yeast. In this study, the
responded to 88.2% of the theoretical yield. However,
effects of yeast extracts, (NH4 )2 SO4 , urea and their dif-
the ethanol yield of the control was only 0.43 g ethanol
ferent combinations on ethanol production were exam-
g 1 glucose, which corresponded to 84.3% of the theo-
ined by batch culture (Table 1). To observe easily the
retical yield. The observation implied some other un-
effects of different nitrogen sources on ethanol produc-
known materials together with glucose in the
tion, the relatively low concentration of hydrolysate
hydrolysate were converted to ethanol. Prosen et al.
containing 31.6 g/l glucose was used. The results showed
(1993) also mentioned the similar phenomenon in the
that single urea was the best nitrogen source. The eth-
report on the growth of microorganisms using wood
anol production from the fermentation was 15.1 g/l, up
pyrolysate as sole carbon source. In this study, the cell
to 93.7% of the theoretical yield. Although the ethanol
growth of S. cerevisiae was not monitored by the
yield using both yeast extracts and (NH4 )2 SO4 as ni-
method of optical absorbance due to very deep color of
trogen source was almost identical to that using urea,
the hydrolysate, but the fermentation results suggested
yeast extracts were not economic in industrial fermen-
that S. cerevisiae could grow well in the hydrolysate
tation. In the fermentation using both yeast extracts and
medium.
urea as nitrogen source, the ethanol yield was slightly
Fig. 5 shows the changes of pH in the hydrolysate
decreased. The reason might be that urea together with
and control medium during the batch culture. For the
yeast extract excessively stimulated the growth of the
control medium, the pH decreased rapidly from 5.50 to
yeast and decreased its ethanol biosynthesis.
4.77 after 9 h of cultivation. But for the hydrolysate
medium, it decreased slowly and remained above 5.05
throughout the fermentation. Silva and Roberto (2001) 3.4. Recycling of S. cerevisiae
supposed that yeasts have an ability to maintain the
relatively stable pH, which in turn leads to the inacti- Compared with a variety of physico-chemical treat-
vation of the toxic compounds in the hydrolysate. ments, the preadaptation of the yeast strain to the hy-
Palmqvist and Hahn-Hagerdal (2000) reported that cell drolysate medium was a low-cost technique by which

Table 1
Effects of different nitrogen sources on fermentation of the hydrolysate to ethanol by S. cerevisiae at 30 °C for 24 h
Fermentation parameters (NH4 )2 SO4 Yeast extracts + (NH4 )2 SO4 Urea Yeast extracts + urea
Final ethanol concentration (g/l) 13.3 15.0 15.1 14.1
Ethanol yield coefficient (%) 42.1 47.5 47.8 44.6
Ethanol yield of theoretical maximum (%) 82.5 93.1 93.7 87.5
Note. (NH4 )2 SO4 : 5.0 g/l; yeast extracts: 10.0 g/l; urea: 6.4 g/l.
Initial glucose concentration was 31.6 g/l in the hydrolysate.
 Z. Yu, H. Zhang / Bioresource Technology 90 (2003) 95–100
Table 2
Fermentations of the hydrolysate to ethanol by S. cerevisiae (S. c.) and S. cerevisiae (R) (S. c. R)
Glucose in the hydrolysate (g/l) Strain Substrate utilized (%)
a
37.8 S. c. 100
S. c. R 100
58.6b S. c. 85
S. c. R 100
95.8b S. c. 73
S. c. R 95
Note: a and b represent the results were taken at 48 and 72 h, respectively.
hydrolysate fermentation could be significantly im-
proved (Silva and Roberto, 2001). In this study, the
comparison experiments with S. cerevisiae (R) using
urea as nitrogen source were carried out. The results in
Table 2 showed that S. cerevisiae (R) gave an im-
provement in substrate utilization and ethanol produc-
tion, especially when higher concentrations of
hydrolysates were used as substrate. For the fermenta-
tion of the hydrolysate containing 95.8 g/l glucose, the
recycled strain S. cerevisiae (R) showed about 30% im-
provement in substrate utilization and 47% increase in
ethanol production compared to its parent strain. The
results clearly implied that relatively low conversion
yields and ethanol concentrations found with unadapted
fresh cultures could be easily overcome by cell recycling
technique. It should be pointed that S. cerevisiae was a
robust yeast and had a good adaptability to unfavorable
environment. The adaptation technique for it may not
be very remarkable in improving the fermentability of
the hydrolysate. In previous studies, the strains like
Candida shehatae and Pichia stipitis improved dramati-
cally ethanol production in the fermentation of ligno-
cellulosic hydrolysate after they were adapted (Parekh
et al., 1986). If the strains were used in the fermentation
of the pyrolysate-derived hydrolysate, the adaptation
technique would be very helpful for them.
In this context, results of the investigation clearly
showed that the process by hydrolysis of cellulosic
pyrolysate to glucose and subsequent fermentation to
ethanol with S. cerevisiae was feasible. It should be
noted that the use of such process could allow elimi-
nation of the enzymatic cellulose hydrolysis step used in
many commercial processes for ethanol production
from the cellulosic biomass. As a typical example, the
cellulase used in simultaneous saccharification and
fermentation (SSF) was easily inhibited by fermentation
products such as ethanol and fermentation conditions
(pH and temperature), which would decrease the effi-
ciency of ethanol production from cellulosic materials
and increase the cost (Ooshima et al., 1984; Mamma
et al., 1995; Wu and Lee, 1997). Furthermore, some
cellulosic materials such as hard wood were very hard
in hydrolyzing to fermentable sugar with enzyme pro-
duced by microorganisms. For these problems, they
99
1
Ethanol (g/l) Yield (g ethanol g glucose)
16.3 0.43
18.1 0.48
20.1 0.40
27.6 0.47
27.4 0.39
40.2 0.44
could be overcome by the process presented in this
paper.
Acknowledgements
This work was supported by grants from the Chinese
Academy of Sciences. The authors wish to express their
thanks to Prof H.Y. Qi for her advice and Ms. H.Y. Gu,
Mr. Z.H. Bai, Mr. L. Li, Mr. H.F. Luo and Mr. W.T.
Lin for their technical assistance.
References
mogrovicov
a, D., Sl
avikov
a, I., P
atpov
a, J., D€
Bafrncov
a, P., S om
eny,
Z., 1999. Improvement of very high gravity ethanol fermentation
by media supplementation using Saccharomyces cerevisiae. Bio-
technol. Lett. 21, 337–441.
Bonn, G., 1985. Analytical determination of 1,6-anhydro-b-D -gluco-
pyranose and kinetic studies under hydrothermal conditions. J.
Carbohydr. Chem. 4, 405–419.
Brown, R.C., Radlein, D., Piskorz, J., 2001. Pretreatment processes to
increase pyrolytic yield of levoglucosan from herbaceous feed-
stocks. In: Bosell, J.J. (Ed.), American Chemical Society Sympo-
sium, Series no. 784. American Chemical Society, Washington DC,
USA, pp. 123–134.
Kumakura, M., 1997. Preparation of immobilized cellulase beads and
their application to hydrolysis of cellulosic materials. Process
Biochem. 32, 555–559.
Mamma, D., Koullas, D., Fountoukidis, G., Kekos, D., Macris, B.J.,
Koukios, E., 1995. Bioethanol from sweet sorghum: simultaneous
saccharification and fermentation of carbohydrates by a mixed
microbial culture. Process Biochem. 31, 377–381.
Nakagawa, M., Sakai, Y., Yasui, T., 1984. Itaconic acid fermentation
of levoglucosan. Ferment. Technol. 62, 201–203.
Ooshima, H., Ishitani, Y., Harano, Y., 1984. Simultaneous sacchari-
fication and fermentation of cellulose: effect of ethanol on
enzymatic saccharification of cellulose. Biotechnol. Bioeng. 27,
389–397.
Palmqvist, E., Hahn-Hagerdal, B., 2000. Fermentation of lignocellu-
losic hydrolysates, I: inhibition and detoxification. Bioresour.
Technol. 74, 17–24.
Parekh, S.R., Yu, S., Wayman, M., 1986. Adaptation of Candida
shehatae and Pichia stipitis to wood hydrolysates for increased
ethanol production. Appl. Microbiol. Biotechnol. 25, 300–304.
Piskorz, J., Radlein, D., Scott, D.S., Czernik, S., 1989. Pretreatment of
wood and cellulose for production of sugars by fast pyrolysis. J.
Anal. Appl. Pyroly. 16, 127–136.
100 Z. Yu, H. Zhang / Bioresource Technology 90 (2003) 95–100
Prosen, E.M., Radlein, D., Piskorz, J., Scott, D.S., Legge, R.L.,
1993. Microbial utilization of levoglucosan in wood pyrolysate
as a carbon and energy source. Biotechnol. Bioeng. 42, 538–
541.
Radlein, D., Piskorz, J., Scott, D.S., 1991. Fast pyrolysis of natural
polysaccharides as a potential industrial process. J. Anal. Appl.
Proly. 19, 41–63.
Roberto, I.C., Felipe, M.G.A., Mancilha, I.M., Vitolo, M., Sato, S.,
1996. Bioconversion of rice straw hemicellulose hydrolysate for
production of xylitol production: effect of pH and nitrogen source.
Appl. Biochem. Biotechnol. 57/58, 339–347.
Shafizadeh, F., Stevenson, T.T., 1982. Saccharification of douglas-fir
wood by a combination of prehydrolysis and pyrolysis. J. Appl.
Polym. Sci. 27, 4577–4585.
Silva, C.J.S.M., Roberto, I.C., 2001. Improvement of xylitol production
by Candida guilliermondii FTI 20037 previously adapted to rice straw
hemicellulosic hydrolysate. Lett. Appl. Microbiol. 32, 248–252.
Wu, Z., Lee, Y.Y., 1997. Inhibition of the enzymatic hydrolysis of
cellulose by ethanol. Biotechnol. Lett. 19, 977–979.
Zhuang, X.Y., Zhang, H.X., Yang, J.Z., Qi, H.Y., 2001. Preparation
of levoglucosan by pyrolysis of cellulose and its citric acid
fermentation. Bioresour. Technol. 79, 63–66.