Integrated biorefinery model based on production of furans using open-ended high yield processes

Article abstract of DOI:10.1039/c3gc42424c

The biodetoxification pathway for the reduction of the fermentation inhibitor furfural was utilized to produce furfuryl alcohol using both a commercial Bakers' yeast and six other native strains, selected for their high tolerance towards the inhibitory effects of furfural. This study explores the potential of the microbial method as an environmentally-benign alternative to the conventional catalytic hydrogenation process for producing furfuryl alcohol used extensively in industry. The microbial method for furfuryl alcohol production provides the benefit of a homogeneous biochemical conversion, devoid of chemical catalysis, in conjunction with other carbohydrate-based processes (e.g. production of ethanol). Results showed that the yields of furfuryl alcohol using the laboratory yeast strains exceeded 90% of the theoretical yield at a furfural concentration of 25 g l^-1, which are comparable to yields obtained using the catalytic process. Furfuryl alcohol yields progressively declined as the furfural concentration was increased up to 65 g l^-1, where the yields averaged over 37%. Piecing together novel high-yield conversion processes for furfural and furfuryl alcohol, an integrated biorefinery model based on the production of furans has been envisioned. Such a facility bypasses the need for high pressure hydrogenation using copper chromite catalysts and hydrogen and azeotropic distillation of furfural to produce dilute streams of both notable platform chemicals. the Partner Organisations 2014.

Full text of DOI:10.1039/c3gc42424c

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Integrated biorefinery model based on production
of furans using open-ended high yield processes†
Anurag Mandalika,a Li Qin,b Trey K. Satob and Troy Runge*c
Cite this: Green Chem., 2014, 16,
2480
The biodetoxification pathway for the reduction of the fermentation inhibitor furfural was utilized to
produce furfuryl alcohol using both a commercial Bakers’ yeast and six other native strains, selected for
their high tolerance towards the inhibitory effects of furfural. This study explores the potential of the
microbial method as an environmentally-benign alternative to the conventional catalytic hydrogenation
process for producing furfuryl alcohol used extensively in industry. The microbial method for furfuryl
alcohol production provides the benefit of a homogeneous biochemical conversion, devoid of chemical cata-
lysis, in conjunction with other carbohydrate-based processes (e.g. production of ethanol). Results showed
that the yields of furfuryl alcohol using the laboratory yeast strains exceeded 90% of the theoretical yield
at a furfural concentration of 25 g l−1, which are comparable to yields obtained using the catalytic
process. Furfuryl alcohol yields progressively declined as the furfural concentration was increased up to
65 g l−1, where the yields averaged over 37%. Piecing together novel high-yield conversion processes for
furfural and furfuryl alcohol, an integrated biorefinery model based on the production of furans has been
envisioned. Such a facility bypasses the need for high pressure hydrogenation using copper chromite
catalysts and hydrogen and azeotropic distillation of furfural to produce dilute streams of both notable
platform chemicals.
Received 27th November 2013,
Accepted 4th February 2014
DOI: 10.1039/c3gc42424c
www.rsc.org/greenchem
with biomass-derived fuel sources might seem overwhelming
Introduction
Biorefineries, furfural and derivatives
due to the extremely large volumes of oil utilized (∼90 million
barrels per day globally)2 and the commodity nature of the
It is envisioned that bio-based societies can become a reality pricing, other products from the refining process may be more
when biomass-derived renewable substitutes have been found easily replaced with renewable sources. The production of syn-
for the vast array of products currently derived from the proces- thetic organic materials such as fibers, rubbers, adhesives,
sing of crude petroleum and other fossil sources. The design coatings, plastics, etc. (currently produced primarily from pet-
of future biorefineries may seek inspiration from the utility of roleum and natural gas), requires a carbon source, and
current petroleum refineries, producing a variety of fuels and biomass is the only renewable feedstock that fulfills this
value-added chemicals by processing a wide array of biomass requirement.3
feedstock using physical, chemical, and biological unit
To identify the most valuable chemicals amidst a cornuco-
operations. The integrated biorefinery has been defined as pia of available candidates for production in the integrated
‘a facility that integrates biomass conversion processes and biorefinery, the concept of the platform chemical has been
equipment to produce fuels, power, and chemicals from introduced in a study conducted by the United States Depart-
biomass’,1 and there is a consensus among researchers in this ment of Energy (US DOE) in 2004,4 which included com-
area that these are the essential tools for the realization of a pounds such as levulinic acid, glycerol, sorbitol. The study
bio-based society. While the replacement of fossil-origin fuels recommended that research into their production and
implementation would be a promising first step towards the
development of biorefineries. This study was revisited in 2010
by Bozell and Petersen,5 in which they refined the criteria for
the identification of chemicals as ‘new top chemical opportu-
nities’, and this included chemicals such as ethanol, lactic
acid, and furfural.
aLouisiana State University Agricultural Center, Baton Rouge, LA 70803, USA
bDOE Great Lakes Bioenergy Research Center, University of Wisconsin-Madison,
Madison, WI 53726, USA
cBiological Systems Engineering, University of Wisconsin-Madison, Madison,
WI 53706, USA. E-mail: trunge@wisc.edu
Furfural, identified as one of the most promising chemicals
by Bozell and Peterson,5 is the natural dehydration product of
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3gc42424c
2480 | Green Chem., 2014, 16, 2480–2489
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Fig. 1 Possible furan-based chemicals with furfural as a precursor. Uses of some of the chemicals are provided in the brackets (adapted from Sain
et al.7).
five-carbon sugars, arabinose and xylose. As of 2007, the made from cross-linked polymers of furfuryl alcohol with
market price of furfural was reported to be $1450 per ton.6 itself and other products (furfural, formaldehyde, phenolic
Apart from being a valuable platform chemical derived from compounds, urea, etc.) were shown to possess excellent
renewable biomass feedstocks, furfural is the precursor for chemical, thermal and mechanical properties, in addition
many furan-based chemicals, and finds itself inextricably to withstanding corrosion and solvent action.8 Furfuryl
involved in largely catalytic transformations leading to higher- alcohol has also been used in the manufacture of furan fiber-
value chemicals and solvents.
reinforced plastics (FRP) for use in piping, and is rec-
Furfural offers enormous prospects for the development of ommended for use for high performance chemical processes
a biorefinery geared towards the production of furfuryl alcohol when chlorinated aromatics, oxygenated organic solvents, etc.,
and its derivatives. Importantly, it promises to open up a are used, owing to its anti-corrosion properties. Such cor-
whole new class of furanic chemicals that can be derived from rosion-resistant off-the-shelf piping has been available since
biomass feedstock, with very well established chemistry that 1977.8,9 Recently, applications of furfuryl alcohol have also
has been comprehensively researched since furfural was first been explored for impregnating wood (furfurylation) to
produced in large quantities.
improve its physical and mechanical properties (hardness,
The greatest application of furfural, among many, is its conver- improvement of the moduli of rupture and elasticity, and
sion into derivatives such as furfuryl alcohol, furan, furoic acid, enhanced dimensional stability.10,11 Furfurylation has also
etc., and the potential applications for these chemicals. Fig. 1 been shown to impart to the wood-polymer composites excep-
illustrates a snapshot of some of the potential furan-based chemi- tional resistance to microbial decay and attack by insects,11
cals that can be derived from a common precursor, furfural.
with the advantage of being non-toxic over other chemical
treatment agents. Polymerized furfuryl alcohol was also used
as a source of carbon for the preparation of molecular
Furfuryl alcohol
Around 62% of the furfural produced globally each year is sieves for potential applications as confined nanopore chemi-
converted into furfuryl alcohol, which finds its largest appli- cal reactors12 and for the preparation of carbon membranes
cation in the manufacture of foundry resins. Resins with gas separation capabilities.13
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 2480–2489 | 2481

View Article Online
Paper
Green Chemistry
nickel and platinum counterparts, which explains their wide-
spread use in industrial hydrogenation reactions.25
Despite the promise of these catalytic reactions, they lack
conformity towards green chemistry principles and are not
suitable for integration into lignocellulosic biorefineries. If the
value of furfural as a platform chemical must be effectively
realized, greener and much milder alternative routes of fur-
furyl alcohol production need to be explored and established,
which this study attempts to undertake.
Fig. 2 Copper chromite catalyzed conversion of furfural into furfuryl
alcohol (adapted from Wojcik16).
Manufacture of furfuryl alcohol
Microbial conversion of furfural. A relatively understudied
area of furfuryl alcohol production involves the microbial
reduction of furfural. Much of the literature published in this
area pertains to conversion of furfural as a fermentation
inhibitor into the less toxic furfuryl alcohol. Lignocellulose, by
virtue of its structure comprising of the crystalline cellulose
structure enclosed in a lignin shell bonded by the hemicellu-
lose (acting as glue), is notoriously resistant to chemical and
biological attacks that seek to break it down, a phenomenon
referred to as biocalcitrance.26,27 Following an acid-catalyzed
pretreatment of lignocellulosic biomass during the production
of cellulosic ethanol, inhibitors such as furfural, 5-hydroxy-
methylfurfural (HMF), acetic acid, formic acid, vanillin,
levulinic acid, 4-hydroxybenzaldehyde, etc. are produced.27 Of
these, furfural and HMF have been identified to be potent
inhibitors to microbial growth and activity during fermenta-
tion processes. It should be noted that alkaline pretreatments
do not form these moieties due to the absence of the acid
catalyst, but this is not the topic of this paper.
Catalytic hydrogenation of furfural. The catalytic hydrogen-
ation of furfural (Fig. 2) is thought to occur by the selective
hydrogenation of the carbonyl group (CvO) on furfural,
through an alkoxide intermediate producing furfuryl alcohol.14
Hydrogenation of the second active group, the unsaturated
CvC bond, results in the production of tetrahydrofurfural.15
Several catalysts have been evaluated for the hydrogenation
of furfural into furfuryl alcohol. Copper is a highly selective
catalyst for the hydrogenation reaction and is widely used
in its chromite form, while decarbonylation becomes the
dominant path when using palladium, leading to the pro-
duction of 2-methylfuran in the latter case.
Commercial production of furfuryl alcohol in the past
was carried out using 1–2% copper chromite catalyst at
1000–1500 psi by hydrogenating technical-grade furfural in
110-gallon autoclaves at 175 °C, with reported furfuryl alcohol
yields of 96–99%.16 Today, this process is accomplished
almost exclusively in the vapor phase by the hydrogenation of
furfural using copper chromite catalysts to produce furfuryl
alcohol with yields exceeding 92%.17 While the use of copper
chromite catalysts guarantees exceptional yields, the moderate
activity18,19 and the toxicity20,21 associated with chromite
have generated interest in the evaluation of other possible
candidates to catalyze the hydrogenation reaction. In addition,
stringent environmental regulations to prevent the disposal of
deactivated copper chromite catalysts in landfills have been
enacted, and have generated interest in the development of
catalysts that do not contain chromium, as replacements.22
Research has been done on the use of highly selective hetero-
geneous catalysts such as carbon-supported copper, studied at
150–200 °C and at a hydrogen pressure of ∼14 psi.22,23
RANEY® nickel, modified with the salts of heteropolyacids
having Keggin-type structures, has been utilized as a catalyst
in a liquid-phase process (∼145 psi) to accomplish the pro-
duction of furfuryl alcohol with high selectivity (98.5%).21 Lee
and Chen15 worked with ultrafine (0.1–0.01 sq. m g−1 surface
area) amorphous catalysts prepared using Ni, P and B and
found a greater selectivity (over 80% at furfural conversion
greater than 80%) in comparison to conventional RANEY®
nickel catalyst. The hydrogenation reactions in this study were
conducted at 80 °C and at 250 psi., Chromium in the copper
chromite catalyst was substituted with MgO to reduce furfural
on coprecipitated Cu–MgO catalysts in the vapor phase
(250 °C), resulting in the production of furfuryl alcohol at a
high selectivity of 98%.24 Yet catalysts containing copper have
the benefit of being milder and are more selective than their
The Saccharomyces cerevisiae yeast strain 354 has been
reported to convert pure furfural to furfuryl alcohol28 and
tolerate furfural concentrations up to 3%. Using molasses as
the sugar source, the authors report furfuryl alcohol yields of
96% at conditions much milder than the catalytic process.
De Villegas et al.29 investigated the effects of aeration and
stirring on the production of furfuryl alcohol using S. cerevisiae
354, and obtained a conversion of 70% with a final furfural
concentration of 35% by slow addition. Palmqvist et al.30 con-
ducted small-scale fermentation (150 ml) studies by growing
Baker’s yeast in the presence of up to 53 mM (∼5.1 g l−1
)
furfural with glucose as the carbon source. Larger-scale fer-
mentations (3 l) were also carried out in this study at a furfural
concentration of 29 mM (∼2.8 g l−1) and furfuryl alcohol yields
of ∼97% were reported. A corresponding decrease in cell mass
yields and a slight increase in the ethanol yields on glucose
were also found in this study. Higher concentrations of fur-
fural in the reaction mixture were shown to have adverse
effects on the specific growth rate of the organisms. Studies
have also been conducted using diverse organisms such as
Methanococcus deltae ΔLH,31 Coniochaeta lignaria,32 engin-
eered strains of Escherichia coli and Klebsiella oxytoca,33 etc.,
with the specific intent of abating the toxicity of furfural by
converting it into furfuryl alcohol. It is, therefore, evident that
furfural can be converted into furfuryl alcohol in very high
yields by employing microbial processes, which offer the
advantages of milder, non-catalytic conditions. While studies
have focused exclusively on biodetoxification of lignocellulosic
2482 | Green Chem., 2014, 16, 2480–2489
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
hydrolysates for fermentation, use of this pathway for the pro- oxidized to NAD+ during various pathways. Furfural, by the
duction of furfuryl alcohol is resonant of a truly green process action of ADH, can be reduced to furfuryl alcohol, with the H+
that is in line with the goals of the integrated biorefinery.
derived from the oxidation of NADH into NAD+. An interesting
observation from this study was that the presence of furfural
reduced glycerol production, leading to the oxidation of an
equivalent amount of NADH (corresponding to the amount of
glycerol that would have been produced in the absence of fur-
fural). The reduction of furfural into furfuryl alcohol acts as a
redox sink for the regeneration of NAD+, reducing the need for
glycerol production to serve this purpose.39 The excess NADH
formed as a result of biosynthesis was oxidized by both the
reduction of furfural, and glycerol production, but the latter
occurred only after all the furfural was depleted.30
In addition, the presence of furfural led to the production
of significant amounts of extracellular acetaldehyde, which
can limit growth, but not metabolic activity. The depletion of
intracellular acetaldehyde (due to excretion) is reflected in the
reduced amounts of acetate observed (which was reported to
be low in the beginning of the reaction due to the inhibition
of aldehyde dehydrogenase (ALDH) in the presence of high
concentrations of acetaldehyde). The excretion of acetaldehyde
is also reflected in the reduced amounts of ethanol produced
(by the reduction of acetaldehyde by the action of ADH) in the
presence of furfural. The formation of furoic acid as a bypro-
duct has been shown to be the effect of oxidation of furfural
by ALDH, even though its activity towards furfural is low.30
Mechanism
Researchers have sought to elucidate the mechanism by which
S. cerevisiae accomplishes the reduction of furfural into fur-
furyl alcohol, and to identify the enzymes involved. Banerjee
et al.34 assayed several glycolytic enzymes in the presence of
furfural and reported that dehydrogenases were the most sus-
ceptible to inhibition by furfural. In particular, the enzymes
glyceraldehyde 3-phosphate dehydrogenase and alcohol dehy-
drogenase (ADH) were observed to have been the most inhib-
ited due to the presence of furfural. In another study, these
researchers concluded that the activity of ADH II, responsible
for the oxidation of ethanol, seemed to have been stimulated,
leading to an increase in total ADH activity.35 This was aug-
mented by the work of Weigert et al.,36 who concluded that
ADH was the enzyme most responsible for the conversion of
furfural into furfuryl alcohol in the organism Scheffersomyces
stipitis. In batch culture, the specific growth rate of S. cerevisiae
was found to decrease with increasing furfural concentration
(0 to 2 g l−1) and inoculum size.37 It was observed that the
inhibitory effect of furfural on growth was much smaller than
its effect on ethanol production, because growth may be influ-
enced by several metabolic cycles, while furfural was thought
to stimulate some processes which provided energy for growth.
This is likely because during anaerobic fermentation, organisms
utilize glucose mostly for the production of energy in the form
of the energy carrier molecule, ATP (adenosine triphosphate)
and very little (∼5%) of this energy goes towards growth.
Glycolysis and oxidative phosphorylation are seen as the
two catabolic pathways that lead to the production of ATP in
S. cerevisiae, where glucose is consumed in order to produce
ATP and ethanol.38
Experimental
Materials
Initial experiments performed to establish growth and fermen-
tation conditions utilized generic Bakers’ yeast (Red Star®
Active dry yeast, Lesaffre Yeast Corp., Milwaukee, WI). These
served as a follow-up for further fermentation experiments
using yeast strains that exhibited tolerance to furfural.
Fig. 3 shows a simplified scheme of pathways for glycolysis,
leading to growth and the production of ethanol, in the pres-
ence of furfural, constructed by Palmqvist et al.30 as part of a
carbon mass-balance. The coefficients α, β and γ are part of
this carbon balance with regard to the quantity of NADH
For medium preparation and sterilization, 20 g l−1 of
glucose (Acros Organics, New Jersey) was sterilized separately
in distilled water (with furfural added depending on whether
the reactor was not the control), and 5X YP (50 g L−1 yeast
extract, 100 g L−1 peptone) was sterilized separately by mixing
100 g l−1 bacto-peptone and 50 g l−1 yeast extract in distilled
water, so that when added back to the glucose–furfural
mixture, the YP concentration would be reduced to 10 g l−1
yeast extract and 20 g l−1 bacto-peptone. This step of perform-
ing the sterilization separately was done to prevent the occur-
rence of caramelization40 when the three chemicals were
sterilized together. Yeast extract and bacto-peptone were pro-
cured from Fisher Scientific (Fair Lawn, New Jersey).
Selection of yeast species
96-well plate growth phenotyping of S. cerevisiae strains. To
identify S. cerevisiae strains with growth tolerance to furfural,
57 S. cerevisiae strains were prepared as previously described.41
In brief, strains were individually inoculated into a 96-deep
well block (NUNC) containing 500 µl of YPD media (10 g L−1
Fig. 3 Proposed pathway for the transformation of furfural into furfuryl
alcohol in conjunction with the fermentation of glucose (adapted from
Palmqvist et al.30); ‘Int’ and ‘Ext’ refer to intracellular and extracellular,
respectively.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 2480–2489 | 2483

View Article Online
Paper
Green Chemistry
yeast extract, 20 g L−1 peptone, 20 g L−1 dextrose) with a multi- Later experiments confirmed the latter as the reason for the
channel pipettor. After inoculation, the deep well block was decreased concentration of furfural, so the results from these
sealed with breathable tape (Axygen), covered with a lid and experiments factor in that loss.
incubated in 30 °C platform shaker. After 48 h of growth, 10 µl
Identification of furfural tolerant S. cerevisiae strains
of saturated cultures from the deep well block were used to
inoculate a standard 96-well plate (NUNC) containing 190 µl of Recent studies have determined that wild S. cerevisiae strains
YPD + 50 mM Furfural (Sigma-Aldrich).
isolated from diverse ecological niches display a range of phe-
The inoculated 96-well plate was placed in Tecan F500 plate notypic responses to environmental stress conditions43,44 due,
reader maintaining an interior chamber temperature of 30 °C. in part, to their genetic differences. Given this, it was specu-
Plates were shaken for 10 s and optical density at 595 nm lated that wild S. cerevisiae strains with wide-ranging genetic
(OD595) measured from each well every 10 minutes for approxi- variation might also display a range of abilities to detoxify fur-
mately 24 h with no shaking. The total change in cell density fural, including differences in enzymatic activities for conver-
for each strain was calculated by subtracting the initial cell sion into furfuryl alcohol and abilities to regenerate cellular
density (OD595 at t = 0 h) from the final cell density (OD595 NADH. Given that yeast cell growth is limited by furfural, it
at t = 24 h). Average changes in cell density and standard devi- was speculated that diverse wild strains would grow at
ations were determined from three independent biological different rates and to different cell densities in the presence of
replicates.
furfural. To test for this, a panel of 57 wild and domesticated
S. cerevisiae strains were screened for those displaying the
greatest change in cell density on lab media (YPD) containing
50 mM furfural within 24 h (Fig. 4 and ESI Table 1†). Indeed, a
range of growth phenotypes were observed; very little change
in growth was seen with commonly-used lab strains BY4741,
S288c, CEN.PK2-1D, and W303 while large increases in growth
were observed by clinical isolates 273614N and YJM421,
natural isolates UWOPS87-2421 and DBVPG6765, fermentation
strains DBVPG6040, DBVPG6044, NC-02 and UC5, and
NCYC361. Interestingly, one lab strain, FL100, grew robustly in
the presence of furfural, while a number of other fermentation
strains and clinical isolates did not show similar growth. This
suggests that specific environmental niches that the strains
were isolated from do not necessarily correlate with furfural
tolerance.
From this work six strains with higher growth in the pres-
ence of furfural compared to common lab strains were selected
for further analysis. The six strains were cultured in YPD
media with 50 mM furfural with their growth rate plotted with
time. Growth rate was measured as the OD at 595 nm with the
background absorbance subtracted. The results indicate that
the 6 strains showed similar growth rates, with an initial lag
before exponential growth after 12 hours (Fig. 5). Based on
these results all six strains displayed similar capacity to re-
produce in the presence of furfural and were suitable for
fermentation experiments.
Aerobic growth and fermentation
Cultures were first grown as starter solutions for a period of
24 hours aerobically with stirring. Inoculation was performed
by suspending yeast grown on YP agar (10 g l−1 yeast extract,
20 g l−1 bacto-peptone and 20 g l−1 agar) media plates
into YPD liquid medium (10 g l−1 yeast extract, 20 g l−1 bacto-
peptone and 20 g l−1 dextrose)42 in sterile conditions.
Data analysis
The furfural and furfuryl alcohol concentrations in the hydro-
lysate were analyzed through a Dionex ICS-3000 Chromato-
graphy System. The samples were centrifuged, decanted and
then filtered through a PTFE 0.20 µm filter (Millipore) to
remove solids. The samples were further diluted and injected
into a Supelcogel C-610H HPLC column (particle size 9 µm,
Sigma-Aldrich) maintained at 30 °C. Detection was performed
by an Ultraviolet (UV) detector with maximum absorbance
at 220. Methanol solution of 15 vol%, which was adjusted to
pH 4.0 by adding 4 wt% H2SO4, was used as the eluent
at a rate of 1.0 ml min−1. HPLC grade furfural, furfuryl alcohol
and furoic acid were purchased from Sigma-Aldrich
(Milwaukee, WI, USA) and used to generate standards for
analysis.
Results
Fermentation: furfural-tolerant strains
Fermentation: Bakers’ yeast
The screening experiments for furfural tolerance were con-
Active dry Bakers’ yeast was used to establish experimentation ducted at a furfural concentration of 50 mM (∼4.8 g l−1) to
criteria initially and to assess the conversion potential and tol- identify the yeast strains capable of tolerating high concen-
erance of generic yeast towards furfural. Fermentation flasks trations of furfural, possibly through greater abilities to reduce
were dosed with furfural at concentrations of 5, 10, 15, 30 and furfural. The screening was followed up with fermentation
40 g l−1, while the concentration of yeast was kept constant at experiments (in duplicates) using the strains that exhibited the
10 g l−1 (∼2 × 108 cells ml−1). Samples taken immediately after greatest furfural tolerance, and the furfuryl alcohol yields for
inoculating the media with yeast indicated a much lower con- the various strains are shown in Fig. 6.
centration of furfural presented in solution, leading to infer
The greatest furfuryl alcohol yield (∼83%) was obtained
that the yeast readily took up furfural and/or there was a loss using the strain UWOPS87-2421, while the lowest yield (∼73%)
of furfural due to condensation reactions during sterilization. was obtained using the strain UC5. Results did not indicate
2484 | Green Chem., 2014, 16, 2480–2489
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Fig. 5 Representative growth curves of indicated strains cultured in
YPD media with 50 mM furfural are plotted as cell density (background
subtracted OD595) over time in hours. Cell density readings every
30 minutes are shown for clarity.
Fig. 4 The average change in total cell density (final OD595 − initial
OD595) for each strain is plotted with standard deviations from 3 inde-
pendent biological replicates. Strains identified for growth tolerance to
furfural and selected for additional study are shown in bold.
Fig. 6 Yield of furfuryl alcohol and the optical density of the yeast
strains following fermentation in the presence of 2.4 g l−1 furfural (with
standard errors for duplicate experiments).
any obvious and meaningful correlations between furfuryl
alcohol yields and the optical density of the organisms follow-
ing fermentation. However, prior research has established
that the inhibition of cell growth by furfural is a function of
furfural concentration, as well as the cell density.30 Navarro
et al.45 have observed that the inhibition effects of furfural
were nullified at high initial cell concentrations as the furfural
was taken up and reduced quickly (up to 5 g l−1 furfural).
Similar results with regard to faster uptake of furfural
with higher inoculum levels are reported by Boyer et al.37 for
25 g l−1 across individual strains, with a steady decline as the
concentration is increased. A logarithmic relationship pro-
vided with the best fit for the decline in yields with increasing
furfural concentration (for average yields for all six strains
varied across the furfural concentration levels). Although
UWOPS87-2421 showed the highest yield across the entire
range of furfural concentrations tested, the overall behavior of
the six strains was quite similar at the furfural concentrations
tested. The similar behavior between the six strains suggests
that being better at growth in the presence of furfural does not
mean greater reduction of furfural. It is likely that growth tole-
rance is furfural is not just by reduction to the alcohol, but
other molecular mechanisms.
furfural concentrations up to 2 g l−1
.
Following favorable results, the six strains were dosed with
furfural at much higher concentrations (25 to 65 g l−1), and
the yields of furfuryl alcohol ranged between ∼38 and ∼94%
(Fig. 7). From the results, it appears that the greatest furfuryl
alcohol yield is obtained at the lowest furfural concentration of
De Villegas et al.29 have observed furfuryl alcohol yields up
to 70% using S. cerevisiae at furfural concentrations of 35 g l−1
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 2480–2489 | 2485

View Article Online
Paper
Green Chemistry
suffered, indicating the inability of the yeast to effectively
convert at higher concentrations. Good selectivity (∼88%) at a
furfural concentration of 25 g l−1, however, indicates efficient
conversion at lower concentrations of furfural in batch mode.
Selectivity towards the production of furfuryl alcohol at low
furfural concentrations (25%) is comparable to that achieved
using novel chemical catalysts (Ni–P–B15), having been
accomplished at ambient conditions as opposed to high
temperature and pressure hydrogenation reactions.
Furoic acid was also produced as a byproduct in small
amounts from all of the strains but in very small yields
(1.7–5.1 mol%). The yields and selectivity towards the gene-
ration of furoic acid, as a byproduct are not substantial, and
little furfural is scavenged towards this end. Researchers have
observed the production of small amounts of furoic acid,28
furoin and furil46,47 alongside furfuryl alcohol during the bio-
transformation process at low furfural concentrations. Use
of HPLC with an ultraviolet detector to study the compounds
produced did not reveal peaks that suggested the existence
of any other compounds aside from furfural, furfuryl
alcohol and furoic acid. It was also not possible to account for
the fate of all of the furfural in solution at higher doses.
Apart from the production of furfuryl alcohol and furoic acid,
it is thought that some furfural is taken up and retained
within unviable yeast cells. Chung and Lee48 have documented
the effects of furfural concentration and initial inoculum level
on cell viability, and have identified a yeast cell concentration
of 108 initial cells mL−1 as the threshold above which
the organisms were able to overcome cell death, use up the
furfural in solution and proceed towards fermentation. Extra-
polating from these observations, it is speculated that at high
furfural concentrations, the organisms were able to convert
small amounts of furfural into furfuryl alcohol and some
furoic acid, but were not able to release the remainder of the
furfural back into solution before furfural toxicity led to cell
death.
Fig. 7 Chart showing the furfuryl alcohol yield obtained using individ-
ual yeast strains and average yield for all six strains at various concen-
trations of furfural (with standard errors for duplicate experiments).
Fig. 8 Average selectivity of the six strains towards the production of
primary product furfuryl alcohol, and the byproduct, furoic acid, with
standard error.
Conclusions
in fed-batch mode, inferring that uptake and assimilation of
small amounts of furfural occurred rapidly, allowing for the
addition of furfural at various intervals. Similar results have
been reported by Villa et al.28 for furfural concentrations up to
30 g l−1 in fed-batch mode, and furfuryl alcohol yields of 96%
using S. cerevisiae, without any apparent effects on the growth
or morphology of yeast cells. The lower yields in the current
study were possibly due to the experiments performed in batch
mode, causing the organisms to be overwhelmed by the high
initial concentration of furfural in solution, unlike in the fed-
batch mode where limited amounts of furfural were fed to the
yeast at specific intervals.
Sensitivity towards furfural production was also similar
between strains, and fell sharply as the concentration of
furfural was increased (Fig. 8). While furfural conversion
was quite high even at the highest concentration (85.2 to
88.2 mol% and higher), the selectivity towards furfuryl alcohol
Furfural was converted into furfuryl alcohol, along with furoic
acid as a byproduct, using both generic Bakers’ yeast and
six furfural-resistant wild S. cerevisiae strains, with yields from
the latter ranging from ∼93 to ∼37% with increasing furfural
concentrations from 25 to 65 g l−1. All six strains exhibited
similar yields at the various furfural concentrations tested for,
with strain UWOPS87-2721 producing furfuryl alcohol at the
highest yield.
It was found that selectivity towards the production
of furfuryl alcohol decreased with increasing furfural con-
centrations, while that of furoic acid remained relatively
constant. All of the furfural introduced into the solution could
not be accounted for at higher concentrations, and it is
presumed that some of it taken up by the cells and not
released back into solution as either product due to toxicity-
associated death.
2486 | Green Chem., 2014, 16, 2480–2489
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Fig. 9 Proposed model biorefinery for the production of furans in an aqueous, green high-yield process.
The biodetoxification pathway for the reduction of furfural hydrothermal pretreatment techniques (both liquid hot water
was shown to be of potential use for the production of furfuryl and steam explosion) are effective towards enhancing the
alcohol in relatively high yields at low concentrations of fur- enzymatic hydrolysis of the extracted solids (using poplar).
fural of 25 g L−1 or less.
In this regard, the water-extracted solids present a very
The work shows that the pathway resonates with the stated suitable substrate for fermentation and production of
goals of biocatalysis of safety in lieu of the mild conditions cellulosic ethanol using the SSF (simultaneous saccharifica-
needed (in comparison to catalytic hydrogenation), high tion and fermentation) configuration.
selectivity exceeding 85%, and the natural origin of the
‘Spent yeast’ refers to yeast that has been obtained from
organisms meaning the biocatalysts are reproducible and recycling of the organisms from fermentation of the extracted
biodegradable.49
solids. Further research may be directed at exploring the feasi-
bility of using this yeast as a low cost ‘catalyst’ at high loading
rates (as was done in the experiments in this study) to convert
the BRD furfural product into furfuryl alcohol instead of cul-
turing an expensive, inhibitor-resistant yeast strain. Recycling
of yeast has been shown to be beneficial as it allows for adap-
tation of the organisms towards inhibitors,57 so it is plausible
that the yeast recycled from fermentation of the extracted
solids may display more resistance to inhibitors. Research-
ers58,59 have used recycled yeast (S. stipitis) to produce ethanol
by the fermentation of red oak hydrolysates and it has been
suggested60 that the use of recycled yeast offers the benefits of
adaptation and can aid fermentation (for galactose fermenta-
tion using several organisms in this particular study).
The furfuryl alcohol produced from the microbial conver-
sion can then be distilled (more economical than the azeotro-
pic distillation of furfural) to obtain a purified product, and
the model biorefinery is thus capable of producing cellulosic
ethanol and furfuryl alcohol in separate streams, following
liquid hot-water extraction. Production of the furan com-
pounds will thus add value to an existing cellulosic ethanol
plant and diversify the range of products that can be produced
at such a facility. Advantages conferred over the conventional
catalytic process include extremely benign reaction conditions
(temperature and pressure) at the expense of time, bypassing
the need for heterogeneous catalysts or for azeotropic distilla-
tion (for furfural) as furfuryl alcohol can be directly distilled
and purified following conversion, along with ethanol. This
biorefinery model also accords other advantages compared to
some of the newer technologies described in literature6,61–63
Next steps
Based on the favorable results obtained for the production of
furfuryl alcohol in high yields using a non-catalytic aqueous
process that can be implemented at ambient conditions,
the concept of a biorefinery is proposed for the production of
furfural and furfuryl alcohol in relatively high yields (Fig. 9).
Earlier work50–52 demonstrated that pentose sugars may be
recovered in substantial amounts (19.5 to 33% of initial xylose
present in the biomass) in the liquid fraction from a range of
biomass feedstock (hybrid poplar, miscanthus, switchgrass
and corn stover) using liquid hot water extraction without
degrading the cellulose fraction by reacting the feedstock at
170 °C for one hour. These hot water extracts contained
pentose sugars on the order of 14 to 21 g l−1, characterized as
total pentose available for conversion into furfural, while only
3 to 5 g l−1 of the hexose sugars (as total hexose) was extracted,
indicating that cellulose degradation was minimal. Batch reac-
tive distillation (BRD) was employed to convert these diluted
(to 5 g l−1) pentose-rich streams into furfural in high yield
(85 to 94%), albeit at very low concentrations.51,52 Furfural
solutions produced by the BRD process (3.86 g l−1) were
converted into furfuryl alcohol at lower yields approaching
60% as reported earlier (due to losses encountered during
sterilization).
The biorefinery loop is closed by fermenting the solid
fraction biomass following hot-water extraction (research on
this aspect has not been performed in this study, but has
been previously reported53–55). Negro et al.56 have shown that
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 2480–2489 | 2487

View Article Online
Paper
Green Chemistry
due to its completely aqueous nature as opposed to the use of 11 S. Lande, M. Westin and M. Schneider, Scand. J. Forest Res.,
expensive and potentially hazardous solvents (toluene, γ-valero- 2004, 19, 22–30.
lactone, etc.), circumvent the need for expensive solid-acid and 12 S. H. Joo, S. J. Choi, I. Oh, J. Kwak, Z. Liu, O. Terasaki and
zeolite catalysts, giving rise to the possibility of retrofitting R. Ryoo, Nature, 2001, 412, 169–172.
conventional liquid acid reactors to the BRD configuration 13 H. Wang, L. Zhang and G. R. Gavalas, J. Membr. Sci., 2000,
to improve furfural yields, and finally, a system that is truly 177, 25–31.
green in concept and open-ended to be readily integrated into 14 S. Sitthisa, T. Sooknoi, Y. Ma, P. B. Balbuena and
existing bioprocessing facilities (ethanol and furfural plants).
D. E. Resasco, J. Catal., 2011, 277, 1–13.
15 S.-P. Lee and Y.-W. Chen, Ind. Eng. Chem. Res., 1999, 38,
2548–2556.
16 B. H. Wojcik, Ind. Eng. Chem., 1948, 40, 210–216.
17 K. J. Zeitsch, The Chemistry and Technology of Furfural and
its Many By-Products, Elsevier, Amsterdam, 2001.
Acknowledgements
This material is based upon work support by the National
Institute of Food and Agriculture, United States Department of 18 J. Kijeński, P. Winiarek, T. Paryjczak, A. Lewicki and
Agriculture, under ID number WIS01521, and the DOE Great A. Mikołajska, Appl. Catal., A, 2002, 233, 171–182.
Lakes Bioenergy Research Center (DOE BER Office of Science 19 J. Wu, Y. Shen, C. Liu, H. Wang, C. Geng and Z. Zhang,
DE-FC02-07ER64494). The authors would like to acknowledge Catal. Commun., 2005, 6, 633–637.
Zhouyang Xiang and Yi-Kai Su for development of the 20 B. M. Nagaraja, V. Siva Kumar, V. Shasikala,
analytical and fermentation methods utilized in this work. The
authors are also grateful to the Wisconsin Energy Institute
A. H. Padmasri, B. Sreedhar, B. David Raju and K. S. Rama
Rao, Catal. Commun., 2003, 4, 287–293.
at the University of Wisconsin, for support of this work, and 21 L. Baijun, L. Lianhai, W. Bingchun, C. Tianxi and
assistance provided by Drs JunYong Zhu and Carl Houtman of
the USDA Forest Products Lab in Madison, WI.
I. Katsuyoshi, Appl. Catal., A, 1998, 171, 117–122.
22 R. Rao, R. T. Baker and M. A. Vannice, Catal. Lett., 1999,
60, 51–57.
23 R. Rao, A. Dandekar, R. T. K. Baker and M. A. Vannice,
J. Catal., 1997, 171, 406–419.
24 B. M. Nagaraja, A. H. Padmasri, B. David Raju and
K. S. Rama Rao, J. Mol. Catal. A: Chem., 2007, 265,
90–97.
25 M. V. Twigg and M. S. Spencer, Appl. Catal., A, 2001, 212,
161–174.
26 M. E. Himmel, S.-Y. Ding, D. K. Johnson, W. S. Adney,
M. R. Nimlos, J. W. Brady and T. D. Foust, Science, 2007,
315, 804–807.
References
1 NREL, Biomass Research-What is a Biorefinery?, http://
www.nrel.gov/biomass/biorefinery.html, Accessed June 24,
2012.
2 EIA, Short-Term Energy Outlook, Energy Information Admin-
istration, Washington, DC, 2013.
3 I. Goldstein, Biomass Availability and Utility for Chemicals,
in Organic Chemicals from Biomass, ed. I. Goldstein, CRC
Press, Boca Raton, FL, 1981, pp. 2–7.
27 H. Dong and J. Bao, Nat. Chem. Biol., 2010, 6, 316–318.
4 T. Werpy, G. Petersen, A. Aden, J. Bozell, J. Holladay, 28 G. P. Villa, R. Bartroli, R. López, M. Guerra, M. Enrique,
J. White and A. Manheim, Top Value Added Chemicals from
Biomass Volume I – Results of Screening for Potential Candi-
M. Peñas, E. Rodríquez, D. Redondo, I. Jglesias and
M. Díaz, Acta Biotechnol., 1992, 12, 509–512.
dates from Sugars and Synthesis Gas, U. S. D. o. Energy, 29 M. E. Diaz De Villegas, P. Villa, M. Guerra, E. Rodriguez,
Pacific Northwest National Laboratory and National Renew-
able Energy Laboratory, Oak Ridge, TN and Springfield, VA,
2004.
5 J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12, 539–
554.
6 R. Xing, W. Qi and G. W. Huber, Energy Environ. Sci., 2011,
4, 2193–2205.
7 B. Sain, A. Chaudhuri, J. Borgohain, B. Baruah and
J. Ghose, J. Sci. Ind. Res., 1982, 41.
D. Redondo and A. Martinez, Acta Biotechnol., 1992, 12,
351–354.
30 E. Palmqvist, J. S. Almeida and B. Hahn-Hägerdal, Bio-
technol. Bioeng., 1999, 62, 447–454.
31 N. Belay, R. Boopathy and G. Voskuilen, Appl. Environ.
Microbiol., 1997, 63.
32 N. N. Nichols, L. N. Sharma, R. A. Mowery,
C. K. Chambliss, G. P. van Walsum, B. S. Dien and
L. B. Iten, Enzyme Microb. Technol., 2008, 42, 624–630.
8 R. H. Kottke, in Kirk-Othmer Encyclopedia of Chemical 33 T. Gutiérrez, M. Buszko, L. Ingram and J. Preston, Appl.
Technology, ed. A. Seidel, John Wiley & Sons, Inc., Hoboken,
NJ, 5 edn, 2000, vol. 12, pp. 259–286.
9 W. J. McKillip, G. Collin, H. Höke and K. J. Zeitsch, in
Biochem. Biotechnol., 2002, 98–100, 327–340.
34 N. Banerjee, R. Bhatnagar and L. Viswanathan, Enzyme
Microb. Technol., 1981, 3, 24–28.
Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH 35 N. Banerjee, R. Bhatnagar and L. Viswanathan, Appl. Micro-
Verlag GmbH & Co. KGaA, 2000. biol. Biotechnol., 1981, 11, 226–228.
10 E. Baysal, S. K. Ozaki and M. Yalinkilic, Wood Sci. Technol., 36 B. Weigert, C. Klein, M. Rizzi, C. Lauterbach and
2004, 38, 405–415.
H. Dellweg, Biotechnol. Lett., 1988, 10, 895–900.
2488 | Green Chem., 2014, 16, 2480–2489
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
37 L. J. Boyer, J. L. Vega, K. T. Klasson, E. C. Clausen and 50 P. J. Wipperfurth, Master of Science, University of Wiscon-
J. L. Gaddy, Biomass Bioenergy, 1992, 3, 41–48. sin, Madison, 2011.
38 R. L. Fireoved and R. Mutharasan, Ann. New York Acad. Sci., 51 A. Mandalika and T. Runge, Green Chem., 2012, 14, 3175–
1986, 469, 433–446. 3184.
39 I. S. Horváth, M. J. Taherzadeh, C. Niklasson and G. Lidén, 52 A. Mandalika, Master of Science, University of Wisconsin,
Biotechnol. Bioeng., 2001, 75, 540–549. Madison, 2012.
40 S. R. Green and C. M. Moehle, in Current Protocols in Cell 53 M. Laser, D. Schulman, S. G. Allen, J. Lichwa, M. J. Antal Jr.
Biology, John Wiley & Sons, Inc., 2001.
and L. R. Lynd, Bioresour. Technol., 2002, 81, 33–
41 M. Jin, C. Sarks, C. Gunawan, B. D. Bice, S. P. Simonett,
44.
R. Avanasi Narasimhan, L. B. Willis, B. E. Dale, V. Balan 54 G. van Walsum, S. Allen, M. Spencer, M. Laser, M. Antal
and T. K. Sato, Biotechnol. Biofuels, 2013, 6, 108.
42 B. Laboratory, Media, http://lfbisson.ucdavis.edu/PDF/
media.pdf, Accessed March 2012.
and L. Lynd, Appl. Biochem. Biotechnol., 1996, 57–58, 157–
170.
55 Y. Kim, N. S. Mosier and M. R. Ladisch, Biotechnol. Prog.,
2009, 25, 340–348.
43 J. Warringer, E. Zorgo, F. A. Cubillos, A. Zia, A. Gjuvsland,
J. T. Simpson, A. Forsmark, R. Durbin, S. W. Omholt, 56 M. Negro, P. Manzanares, I. Ballesteros, J. Oliva,
E. J. Louis, G. Liti, A. Moses and A. Blomberg, PLoS Genet.,
2011, 7, e1002111.
A. Cabañas and M. Ballesteros, Appl. Biochem. Biotechnol.,
2003, 105, 87–100.
44 G. Liti, D. M. Carter, A. M. Moses, J. Warringer, L. Parts, 57 E. Palmqvist and B. Hahn-Hägerdal, Bioresour. Technol.,
S. A. James, R. P. Davey, I. N. Roberts, A. Burt, 2000, 74, 17–24.
V. Koufopanou, I. J. Tsai, C. M. Bergman, D. Bensasson, 58 A. V. Tran and R. P. Chambers, Biotechnol. Lett., 1985, 7,
M. J. O’Kelly, A. van Oudenaarden, D. B. Barton, E. Bailes, 841–845.
A. N. Nguyen, M. Jones, M. A. Quail, I. Goodhead, S. Sims, 59 A. V. Tran and R. P. Chambers, Enzyme Microb. Technol.,
F. Smith, A. Blomberg, R. Durbin and E. J. Louis, Nature,
2009, 458, 337–341.
45 A. Navarro, Curr. Microbiol., 1994, 29, 87–90.
1986, 8, 439–444.
60 T. Lindén and B. Hahn-Hägerdal, Enzyme Microb. Technol.,
1989, 11, 583–589.
46 S. Morimoto, T. Hirashima and N. Matutani, J. Ferment. 61 R. Weingarten, J. Cho, W. C. Conner Jr. and G. W. Huber,
Technol., 1969, 47, 486–490. Green Chem., 2010, 12, 1423–1429.
47 M. J. Taherzadeh, L. Gustafsson, C. Niklasson and 62 J. Lessard, J.-F. Morin, J.-F. Wehrung, D. Magnin and
G. Lidén, J. Biosci. Bioeng., 1999, 87, 169–174. E. Chornet, Top. Catal., 2010, 53, 1231–1234.
48 I. S. Chung and Y. Y. Lee, Biotechnol. Bioeng., 1985, 27, 308–315. 63 E. I. Gürbüz, J. M. R. Gallo, D. M. Alonso, S. G. Wettstein,
49 K. Nakamura, R. Yamanaka, T. Matsuda and T. Harada,
W. Y. Lim and J. A. Dumesic, Angew. Chem., Int. Ed., 2013,
52, 1270–1274.
Tetrahedron: Asymmetry, 2003, 14, 2659–2681.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 2480–2489 | 2489