Chemical Engineering Science 152 (2016) 678–688

Contents lists available at ScienceDirect

Chemical Engineering Science

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

Butanol production by Clostridium acetobutylicum in a series of packed

bed biofilm reactors
F. Raganati a,n, A. Procentese a, G. Olivieri a, M.E. Russo b, P. Gotz c, P. Salatino a,
A. Marzocchella a
a
Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale - Università degli Studi di Napoli Federico II, P.le V. Tecchio 80, 80125
Napoli, Italy
b
Istituto di Ricerche sulla Combustione – Consiglio Nazionale delle Ricerche, P.le V. Tecchio 80, 80125 Napoli, Italy
c
Department of Life Sciences and Technology/Bioprocess Engineering, Beuth University of Applied Sciences Berlin, Seestrasse 64, 13347 Berlin, Germany

H I G H L I G H T S

Butanol production was performed by immobilized cells of C. acetobutylicum.

The process was performed in 4 biofilm packed bed reactors connected in series.

The biofilm was assumed to be formed of an active shell and an inner inactive layer.

art ic l e i nf o a b s t r a c t

Article history: The continuous production of Acetone, n-Butanol and Ethanol (ABE) by immobilized cells of Clostridium
Received 10 December 2015 acetobutylicum DSM 792 using glucose and lactose as carbon source is presented in this paper. The
Received in revised form conversion process was successfully carried out for more than three months in 4 packed bed biofilm
6 June 2016
reactors (PBBRs) connected in series. The first PBBR of the series (fed with fresh medium) was kept under
Accepted 27 June 2016
Available online 27 June 2016
acidogenesis conditions and the three other PBBRs were kept under solventogenesis conditions.
Each PBBR was a glass tube (4 cm ID, 8 cm high) with a 4 cm-bed of 3 mm-Tygon rings as carriers. The
Keywords: PBBR system was fed with 100 g/L of lactose medium. The fermentation process was characterized in
Continuous fermentation terms of metabolite production (butyric and acetic acids, acetone, butanol, and ethanol), sugar conver-
Butanol
sion and mass of biofilm. The overall dilution rate (DTOT) was varied between 0.15 h  1 and 0.9 h  1 to
Fixed bed biofilm reactor
assess the PBBR system performance as a function of DTOT. The best PBBR system performance under
ABE
Clostridium acetobutylicum optimized conditions was: butanol productivity 9.2 g/Lh, butanol concentration 10.8 g/L, acetone con-
Lactose centration 2.4 g/L, ethanol concentration 1.8 g/L, selectivity of butanol with respect to all solvents 72%w.
To the authors’ knowledge, these butanol productivity and concentration values are the highest in the
literature on lactose/(cheese whey) fermentation.
An interpretation of the biofilm structure in the PBBR was put forward.
& 2016 Elsevier Ltd. All rights reserved.

1. Introduction
The development of biotechnological processes to produce
butanol – a second-generation biofuel and a chemical building
block – from renewable sources to find eco-sustainable alter-
natives to the petrochemical routes (Kumar and Gayen, 2011) is
still an open challenge. ABE fermentation by clostridia is drawing
new interest as a way to turn renewable resources into valuable
base chemicals and liquid fuels (Sarchami and Rehmann, 2014;
n
Corresponding author.
E-mail address: francesca.raganati@unina.it (F. Raganati).
Dürre, 2007; Friedl, 2012). Butanol offers several advantages over
ethanol for gasoline-alcohol blending: high energy-content, low
miscibility with water and low volatility (Bohlmann, 2007; Cas-
cone, 2008). In addition, butanol can replace gasoline with no
need to modify the current vehicle and engine technologies
(Cascone, 2008).
Nevertheless, its low yield, the consequent acid-solvent pro-
duction and low concentration of butanol due to its inhibiting
effect on fermentation have hindered its success on an industrial
scale so far. The most preferred substrates used in the traditional
batch fermentation – starch and molasses – result in a total yield of
30–32% and butanol concentration of about 10–15 g/L (Kumar
et al., 2012).

http://dx.doi.org/10.1016/j.ces.2016.06.059
0009-2509/& 2016 Elsevier Ltd. All rights reserved.
 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688
Nomenclature
AA Acetic Acid
AAUP Concentration of the converted Acetic Acid
ABE Acetone-Butanol-Ethanol
BA Butyric Acid
BAUP concentration of the converted Butyric Acid
CSTR Continuous stirred tank reactor
D Overall Dilution rate
YE Yeast Extract
PBBR Packed Bed Biofilm Reactor
PFR Plug flow reactor
rB Specific butanol production rate
S Sugar concentration
Reactor design and operating conditions play a key role in
fermentative productions (Schugerl, 1997). The main factors that
hinder the commercial development of the traditional batch fer-
mentation processes include low cell density, low reactor pro-
ductivity, high down-times, nutritional limitations and severe
product inhibition (Chen and Blaschek, 1999). Improved perfor-
mance can be obtained by increasing cell concentration in cell
immobilized reactors and retention membrane reactors (Qureshi
et al., 2005). Reactor performance can also be enhanced with
continuous ABE production in reactors operated with clostridium
cell-confinement options: cell immobilisation (Gapes et al., 1996;
Yen et al., 2011; Viikilää et al., 2013; Lee et al., 2008; Qureshi et al.,
2000) or cell recycling (Meyer and Papoutsakis, 1989; Tashiro
et al., 2005; Procentese et al., 2015a, 2015b). In cell immobilization
biofilm reactors, their high cell density allows for better butanol
yield and recovery. Continuous bioconversion presents several
advantages over batch cultures in biofilm reactors (Qureshi et al.,
2000). The main advantages are related to the high cell con-
centration and to the reactor operating at high dilution rates
without cell washout (Welsh et al., 1987). Moreover, the biofilm
support can be reused (Krouwel et al., 1980).
Fig. 1 is the map of butanol productivity and concentration in
the final product proposed by Setlhaku et al. (2013) for continuous
bioreactor systems, as reported in the open literature.
Butanol productivity and concentration are benchmarked
against WB,min ¼ 0.24 g/Lh and Bmin ¼ 13 g/L, which are reported
as the best batch fermentations with clostridia in the literature
(Jones and Woods, 1986). A butanol productivity value of 5 g/ Lh
Fig. 1. Window of operation for ABE continuous, two-stage reactors, immobilized
and product integrated fermentations (Setlhaku et al., 2013). The full circle re-
presents the best performance obtained in the test herein.
679
WABE ABE productivity
WB Butanol productivity
WUP
Acids
Rate for volume unit of converted acids
Net
W Acids Rate for volume unit of produced acids
TOT
WAcids Total acids production rate
XA,active Acidogenic active cells concentration in the biofilm
layer
XS,active Solventogenic active cells concentration in the biofilm
layer
Xinactive Inactive cells concentration in the biofilm layer
XTOT Total cell concentration
Yi/L Sugar-to-“i-species” fractional yield coefficient
ξS Sugar conversion degree
Φ Butanol to ABE selectivity
(WB,industry,min) is suggested as the industrial hurdle rate. The fig-
ure also indicates the minimum butanol target (BDSP,min) of 36 g/L
calculated as the threshold for successful butanol recovery by an
energy optimized distillation unit (Mariano and Filho, 2012). To
the authors’ knowledge and as confirmed in Fig. 1, with the biofilm
reactors described in the literature the final concentration of sol-
vents obtained is too low to get an efficient butanol recovery, in
particular by a distillation unit. Such a low butanol concentration
is likely to depend on the impossibility to keep the two phases of
the clostridia fermentation (acidogenesis and solventogenesis)
separate when a single-stage reactor is used.
The study reported in this paper is about continuous bio-bu-
tanol production by lactose conversion in an innovative im-
mobilized cell reactor system. Lactose was chosen as carbon source
because it is the main component of cheese whey, a very common
by-product of the dairy industry (Raganati et al., 2013). The
anaerobic solventogenic bacterium Clostridium acetobutylicum
DSM 792 was used for the fermentation process. The conversion
was carried out in 4 packed bed biofilm reactors (PBBRs) con-
nected in series: the first reactor (fed with the carbon source) was
kept under acidogenesis conditions, and the three other reactors
were kept under solventogenesis conditions. This series config-
uration allowed to keep the two phases of the ABE fermentation
separate: acids were produced in a section of the system and then
the acids and the residual sugar were converted in solvents in the
following section. The PBBR system performance was character-
ized in terms of final butanol concentration and productivity and
measured as a function of the dilution rate. A possible inter-
pretation of the biofilm structure, based on the non-homogenous
nature of the substrate/metabolite concentration across the bio-
film is also put forward.
2. Materials, methods and procedures
2.1. Microorganism
Clostridium acetobutylicum DSM 792 was supplied by DSMZ
(Deutsche Sammlung von Mikroorganismen und Zellkulturen
GmbH). Stock cultures were reactivated according to the proce-
dure indicated by the supplier. The reactivated cultures were
stored at  80°C. After thawing, the reactivated cultures were in-
oculated in 15 mL Hungate tubes containing 12 mL synthetic
medium (see the following paragraph) supplemented with 30 g/L
lactose and 5 g/L yeast extract (YE). The cells were grown under
anaerobic conditions for 48 h at 37 °C.
680 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688
Fig. 2. Outline of the apparatus used for the continuous process: A) PBBRs con-
nected in parallel during the start-up phase; B) PBBRs connected in series during
the butanol production phase. b: pH measure/control device.
2.2. Medium
The synthetic medium used to feed the system consisted of
sugar (100 g/L glucose and/or lactose), yeast extract (5 g/L YE) and
P2 stock solution. The P2 solution was (Qureshi et al., 2000): buffer
(0.25 g/L KH2PO4, 0.25 g/L K2HPO4, 2 g/L ammonium chloride) and
minerals (0.2 g/L MgSO4.7H2O, 0.01 g/L MnSO4.H2O, 0.01 g/L
FeSO4.7H2O).
The sugar and YE solutions were separately autoclaved (20 min
at 120 °C) and cooled at room temperature. The stock solutions
were sterilized through a 0.2 μm filter and then added aseptically
to the sugar-YE solution.
2.3. Apparatus and operating conditions
The apparatus was: reactor system, liquid pumps, heating ap-
paratus, device for pH control and on-line diagnostics (sketch in
Fig. 2). The reactor system was made of four fixed beds. Each bed
was at the bottom of a 100 mL glass lined pipe (4 cm ID, 8 cm high)
jacketed for the heat exchange. Water from an external circulating
water bath (Julabo heating circulator MA4) was fed into the jacket
of each reactor to keep the operating temperature at 37 °C. The
liquid head was controlled by the overflow duct in each reactor:
the working volume of each reactor was set at 40 mL. Nitrogen
was sparged at the bottom of each reactor to ensure anaerobic
conditions.
The pH control device in each reactor included a pH-meter, a
peristaltic pump, a bottle containing NaOH 0.3 M solution and a
controller.
The reactors loaded with the carriers were autoclaved at 121 °C
for 20 min. The nitrogen stream was sterilized by filtration. The
sterile medium was fed at the bottom of each reactor by means of
a peristaltic pump.
Two configurations were used.
Parallel configuration) The four fixed beds were operated in
parallel (Fig. 2A). This configuration was used during start-up to
promote biofilm formation.
Series configuration) The four fixed beds were connected in
series (Fig. 2B). This configuration was used during butanol pro-
duction after a biofilm layer had formed in each unit in parallel
mode.
Tygon rings (3/1 mm OD/ID) were chosen as biofilm carriers
based on the results of previous investigations (Napoli et al., 2010).
39.2 g of Tygon rings were loaded in each reactor.
2.4. Analytical procedures
Broth samples (1.5 mL) were periodically taken from the re-
actors. Each sample was centrifuged at 6700 g for 10 min at 4 °C to
remove any suspended biomass. The liquid phase was character-
ized in terms of sugar and metabolite concentrations. Sugar con-
centration was measured by high performance liquid chromato-
graphy (HPLC) using an Agilent 1100 Series (Agilent Technologies,
United States). Glucose and/or lactose were separated at room
temperature by means of an 8 μm Hi-Plex H (Agilent Technologies,
United States), 30 cm 7.7 mm, and detected with a refractive index
detector. Deionized water was used as mobile phase at a 0.6 mL/
min flow rate. A GC apparatus Agilent 6890 Series (Agilent Tech-
nologies, United States) equipped with a capillary column poraplot
Q (25 m  0.32 mm) (Agilent Technologies, United States) and a
FID was used for the metabolite analysis. Hexanoic acid was used
as an internal standard to measure acid and alcohol concentration.
2.5. Operating procedures
300 microliters of stock culture were transferred into sixteen
15-mL Hungate tubes containing the culture medium (30 g/L of
glucose). The batch cultures were incubated for 1 day under
anaerobic sterile conditions at 37 °C.
2.5.1. Biofilm start-up
40 mL of active cultures (0.5 gDM/L) from the Hungate tubes
were introduced in each fixed bed reactor. The reactor system was
operated in parallel mode. The pH in each reactor was controlled
and kept at about 5.5 (acidogenesis conditions) to promote biofilm
formation. The biofilm was grown using a glucose-based medium.
The main steps of the PBBR system start-up were:
1. Inoculation: Each PBBR was inoculated at t¼0.
2. Batch cultures: Each PBBR was operated in batch mode for 24 h.
3. Continuous cultures: at t ¼24 h the PBBRs were switched to
continuous mode. The dilution rate was set at 0.40 h  1; the pH
was gradually increased up to 5.5 to operate fermentation under
acidogenic conditions (Napoli et al., 2011). A visible biofilm layer
formed on the carriers in about one week and at t ¼ 7 days the
dilution rate was increased up to 0.8 h  1 to promote the biofilm
growth with respect to suspended cell growth.
At t ¼20 days the carriers were covered with abundant biofilm
and steady state conditions had established in all the reactors.
Altogether, the system start-up took about 20 days.
 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688
2.5.2. Butanol production
Butanol production tests were carried out with the 4 PBBRs
connected in series (Fig. 2B) and operated at pre-set conditions.
The pH of the first reactor (reactor 1, Fig. 2B) was set at 5.5 to
promote acidogenesis conditions. The pH of the three other re-
actors (reactors 2, 3, and 4, Fig. 2B) was set at 4.7 to promote
solventogenesis conditions. Butanol production was investigated
using lactose-containing solutions.
The overall dilution rate (DTOT) – ratio between the feeding flow
rate and the total volume of the 4 fixed beds – ranged between
0.15 and 0.9 h  1. After setting the dilution rate, the reactor system
was operated until steady state conditions were reached: meta-
bolite and lactose concentration in each reactor staying constant
for at least 5 times the residence time of the reactor.
The mass of biofilm in the reactors was measured at the end of
the test program sacrificing the reactors (Qureshi and Maddox,
1987). The mass and the concentration of biofilm (XTOT) in the
PBBRs were assessed at the end of the run according to the fol-
lowing procedure: (i) the dry carriers were weighted before re-
actor loading; (ii) at the end of the test, the reactors were rinsed
with sterile water to remove substrates and metabolites; (iii) the
carriers with the biofilm were harvested and dried at 40 °C until
constant weight was reached (typically less than 1 day). The dry
weight of the biofilm in each reactor was calculated as the dif-
ference between the dry weight of the carrier-biofilm and that of
the carrier. The XTOT in each reactor was calculated as the ratio
between the dry weight of the biofilm and the volume of liquid in
each reactor.
2.6. Data processing
To assess the PBBR system performance the concentration of
sugar and metabolites was measured under steady-state condi-
tions (concentration staying constant for at least 5 times the mean
residence time, 1/DTOT). The system performance is reported in
terms of: sugar conversion (ξS), sugar-to-“i-species” fractional yield
coefficient (Yi/S), butanol productivity (WB), ABE productivity
(WABE), butanol to ABE selectivity (Φ).
The performance variables ξS, Yi/L, WB, WABE, and Φ were as-
sessed assuming that: i) the feeding of PBBRs was aseptic and free
of metabolites, and ii) the gas stripping of metabolites was negli-
gible. The variables were assessed according to the following re-
lationships:
( SIN− SOUT )
ξS =
SOUT
iOUT
Yi / L =
(S IN
− SOUT ) (2)
WB = DTOT ⋅BOUT
WABE = DTOT ⋅ABE OUT
BOUT
Φ=
( AOUT + BOUT + E OUT ) thickness of the biofilm characterized by high lactose con-
where S, A, B, E and ABE are the concentrations of sugar, acetone,
butanol, ethanol and total solvents, respectively, as measured at
the inlet (superscript IN, stream S1 in Fig. 2B) and at the outlet of
the reactor system (superscript OUT, stream S5 in Fig. 2B).
681
2.7. Interpretative model of the PBBR system performance
The PBBR system performance was interpreted taking into ac-
count the behaviour of a series of N equal-size mixed flow reactors
(Levenspiel, 1999) and the cell population of the biofilm coupled
with the expected pH and lactose concentration profiles in the
biofilm layer.
2.7.1. The reactor
The high substrate conversion and butanol concentration ob-
tained at the end of batch fermentations can also be obtained at
the outlet of a continuous plug flow reactor (PFR) with continuous
inoculum and similar space-time. The alternative to continuous
inoculum is microorganism immobilization in the PFR. The uni-
form microorganism concentration in the PFR ensures constant
conversion rate in the reactor under constant conditions. A packed
bed biofilm reactor with constant biofilm concentration behaves
as a PFR provided that there is negligible axial dispersion. Since
negligible axial dispersion cannot be assumed at such a low fer-
mentation flow rate then a series of N equal-size mixed flow re-
actors can be used to approximate the PFR behaviour (Levenspiel,
1999). In this research work, four PBBRs were used to approximate
a series of 4 equal-size mixed flow reactors. A PBBR system is
expected to behave like a PFR with the substrate conversion and
the final product concentration increasing with the number of
stages. Product inhibition is expected only in the last stages of the
PBBR series. It is well known that more than 4 equal-size mixed
flow reactors are required to approximate a PFR, but the choice
made in this work is a compromise between the PFR approxima-
tion and the operability of a lab-scale plant.
2.7.2. Biofilm structure
Lactose, metabolite and pH concentration in the biofilm is the
result of a complex interaction of several phenomena involving
external/internal mass transport and local conversion rate. The
effects of the dilution rate on the concentration profiles in the
biofilm are reported hereinafter.
The pH and lactose concentration at the liquid-biofilm interface
was assumed to be equal to the bulk values (measured at the exit
of the reactor) and to decrease in the biofilm with the biofilm
depth (Fig. 3A). The pH profile through the biofilm thickness was
assumed to be as follows: at the liquid-biofilm interface it was
equal to the pH set in the liquid bulk (biofilm external transport
rate negligible with respect to the internal transport rate); a gra-
dient that was the result of the equilibrium between transport
phenomena and microbiological conversions (Olivieri et al., 2011;
(1)
Rai et al., 2012). A pH value lower than 4.7 allows identifying two
regions: i) an external region characterized by local value of
pH 44.5 and presence of acidogenic cells; ii) an internal region
characterized by local value of pH o4.5 and presence of solven-
togenic cells. As reported in previous investigations (Raganati
et al., 2013), the thickness of the acidogenic cell region is expected
to increase with D.
(3) The lactose concentration profile through the biofilm thick-
ness (Fig. 3A) was as follows: the value at the liquid-biofilm in-
terface was equal to the lactose concentration in the liquid bulk
(4) (biofilm external transport rate negligible with respect to the
internal transport rate); a gradient that was the result of the
combined effect of the diffusion in the biofilm and the local
conversion to acid/cells and solvents. Fig. 3A shows that the
(5)
centration increases with D.
Based on the pH and lactose concentration profiles, the biofilm
in each PBBR operated under solventogenic conditions was as-
sumed to have two layers (Fig. 3B): the inner layer (close to the
carrier surface) consisting of inactive cells (Xinactive); the external
682 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688

Fig. 3. Sketch of the biofilm layer in the biofilm unit operated under solventogenic conditions.

layer (in contact with the broth-liquid bulk) consisting of active The acids produced were partially converted into solvents ac-
cells (acidogenic and solventogenic cells, XA,active and XS,active) to cording to the reactions:
convert the carbon source (lactose/acids) into solvents. The con- +4ATP
centration of active acidogenic cells (XA,active) and active solven- C12H22 O11⋅H2 O + 2CH3 COOH ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ 2CH3 COCH3 + 2C2H5 OH + 6CO2 + 4H2 (11)
togenic cells (XS,active) in the biofilm of PBBR # 2, # 3, and # 4 was
+4ATP (12)
estimated according to the theoretical framework reported here- C12H22O11⋅H2O + 2C3H7COOH ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ 2CH3COCH3 + 2C 4H 9OH + 6CO2 + 4H2
inafter. The main assumptions were: a) coexistence of acidogenic
According to the stoichiometry of reactions (11) and (12), the
and solventogenic cells in the biofilm; b) acid production/conver-
amount of acetone produced was equal to the sum of the acetic
sion reactions put forward by Raganati et al. (2013); c) the total
and butyric acid converted:
cell concentration of the biofilm (XA,active þ XS,active þ Xinactive)
during the test for all the steady-state conditions investigated. A = AAUP + BAUP (13)
The total acid production rate by the acidogenic cells ( WTOT
Acids
) in UP
the biofilm was estimated taking into account the yields of butyric where A is the acetone concentration and j the concentration of
and acetic acids (BA and AA) with respect to the sugar (YBA/S and the converted acid “j” (j ¼AA, BA). The acids converted were as-
YAA/S). YX/S, YBA/S and YAA/S were assessed as a function of the sumed to confirm the balance put forward by Desai et al. (1998):
specific growth rate as in Napoli et al. (2012). The biomass-to-acids BAUP BA
fractional yield coefficient (YAcids/X) was: = 0.315
AAUP AA (14)
YAA/S + YBA/S where AA and AB are the concentrations of acids in the reactor.
YAcids/X =
YX/L
( gAcids/gDM).
(6) Given the dilution rate at which the reactor was operated, the rate
of the acids ( WUP
Acids ) converted can be assessed by Eqs. (13) and
The total acid production rate per volume unit by active
(14).
acidogenic cells ( WTOT
Acids
) was:
The concentration of active acidogenic cells in the biofilm was
WTOT
Acids = D i ⋅X A,active⋅YAcids/X (7) calculated by Eqs. (7) and (8), resulting in:

where Di is the dilution rate of each PBBR (DTOT*4). WTOT Acids is also
WTOT
Acids
XA,active =
the sum of the rate per volume unit of acids converted ( WUP Acids ) and
Di⋅YAcids/X (14)
of the rate per volume unit of acids produced ( WNet
Acids
):
The concentration of active solventogenic cells in the biofilm
WTOT (
Net UP
Acids = W Acids + W Acids ) (8) (XS,active) of each reactor of the PBBR system operated under sol-
ventogenic conditions was estimated based on the mass balance of
where butanol extended to the PBBR volume:
WNet
Acids = D i ⋅( AA + BA) (9) Di⋅(BOUT –BIN ) = r B⋅X S,active (15)

where BOUT and BIN are the butanol concentrations in the outlet
WUP (
Acids = D i ⋅ AA
UP + BAUP
) (10) and in the inlet stream of each reactor, and rB is the specific
 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688
butanol production rate estimated according to the model re-
ported by Procentese et al. (2015a):
⎛ S ⎞⎛ AA ⎞⎛ BA ⎞ ⎛ KB ⎞
r B = r B,max⋅⎜ ⎟⋅⎜ ⎟⋅⎜ ⎟⋅⎜ ⎟ Mixture
⎝ S,B
K + S ⎠ ⎝ K AA + AA ⎠ ⎝ K BA + BA ⎠ ⎝ KB + B ⎠
where S is the concentration of the sugar.
The concentration of inactive cells (Xinactive) was estimated
using the following relationship:
Xinactive = XTOT –XA,active–XS,active
were XTOT is the total cell concentration measured in each reactor
operated under solventogenic conditions. The assumption was
that XTOT was constant during the test campaign.
3. Results and discussion
3.1. Biofilm system start-up
Fig. 4 reports the time series of the sugar and metabolite con-
centration and the pH measured in one of the four packed bed
reactors (the data for the other reactors were the same as those
reported in Fig. 4, within the experimental error). Altogether, the
system start-up took about 20 days. The total acid (acetic þ bu-
tyric) yield at the end of the start-up procedure was about 0.45
gacid/gglucose.
3.2. Butanol production
At t¼ 22 days the PBBRs were switched to series mode
(Fig. 2B). In particular, the pH of the first reactor (reactor 1, Fig. 2B)
was kept at 5.5 to have it operate under acidogenic conditions. The
pH of the other reactors of the series (reactors 2, 3 e 4, Fig. 2B) was
Fig. 4. Main data measured during PBBR system start-up. The vertical line marks
the instant at which the continuous operation starts. The vertical dashed line marks
the instant at which the D was increased from 0.4 to 0.8 h-1.
683
Table 1
Composition of the feeding stream set during the switch from glucose to lactose
carbon source. Total sugar concentration: 100 g/L.
Glucose % Lactose %
(16)
a 100 0
b 75 25
c 50 50
d 25 75
e 0 100
(17)
kept at 4.7 to have them operate under solventogenic conditions.
The DTOT was set at 0.15 h  1 and the feeding stream was gradually
switched from glucose-based medium to lactose-based medium
stream to let cells get accustomed to the new carbon source. The
total sugar concentration of the feeding was set at 100 g/L and the
lactose fraction was gradually increased as reported in Table 1.
The feeding composition was kept constant until a steady state
regime was reached, then the lactose fraction was increased again.
Figs. 5 and 6 show the concentration of sugars (glucose and lac-
tose) and metabolites (acetic acid, butyric acid, and butanol)
measured in the four reactors under steady state conditions for the
feeding composition reported in Table 1. The sugar concentration
at reactor # 0 is that of the feeding. It can be noticed that:

mixture b (25% lactose): the glucose was completely converted
in the PBBR series. Lactose conversion was negligible. Butanol
concentration was about 8 g/L in reactor # 4 and butanol pro-
ductivity was 1.2 g/Lh;

mixture c (50% lactose): glucose conversion was almost com-
plete already in reactor # 2. Lactose began to be converted after
glucose depletion (reactor 2). The concentration of butanol in
the PBBRs was 7.2 g/L and butanol productivity was 1.08 g/Lh;

mixture d (75% lactose): both sugars were converted right from
the start of the PBBRs series. The concentration of butanol at the
exit of the PBBRs was 6.8 g/L and the butanol productivity was
1.03 g/Lh;

glucose-free feeding (mixture e): about 80 g/L of lactose were
converted in the PBBRs series. The concentration of butanol at
the exit of PBBRs was about 6 g/L and the butanol productivity
was 0.9 g/Lh.
The analysis of the results confirms that the microorganism
prefers glucose to lactose as carbon source (Raganati et al., 2015;
Yu et al., 2007). Only after glucose had depleted lactose was con-
verted by the biofilm cells in the PBBR series. This sugar preference
was confirmed when the PBBR series performance – expressed in
terms of sugar conversion and butanol concentration, productivity
and yield – was compared in the tests carried out using 100%
glucose (mixture a) and 100% lactose to feed the reactor at DTOT ¼
0.15 h  1. The main results of these two tests may be summarized
as follows:

Hundred percent glucose feed (mixture a), 88% sugar converted
and about 8 g/L of butanol produced. Butanol productivity and
butanol-to-sugar yield were 1.22 g/Lh and 0.09 g/g, respectively.

Hundred percent lactose feeding (mixture e), the PBBRs suc-
ceeded in converting 77% of the sugar and produced about 6 g/L
of butanol. Butanol productivity and butanol-to-sugar yield
were 0.9 g/Lh and 0.08 g/g, respectively.
The comparison suggests that the PBBR system performance
assessed in the tests carried out with lactose is the low limit of the
performance of the system when glucose is used. Although glucose
is preferable as carbon source, the results reported in the following
concern the tests carried out with lactose to compare this PBBR
684 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688
Fig. 5. Sugar (glucose and lactose) and butanol concentrations measured in the
PBRs (reactor #0 in this figure and the following figures is the feeding composi-
tion). Feeding: sugar mixture reported in Table 1. Dilution rate: 0.15 h-1.
system with the PBBR reported by Napoli et al. (2010).
Figs. 7 and 8 report the concentration of metabolites (acetic
acid, butyric acid and butanol) and lactose in the four PBBRs
measured under steady state conditions for different dilution
rates. Table 2 reports the acetone and ethanol concentrations at
the exit of reactor # 4 under steady state conditions for different
dilution rates. The DTOT was increased between 0.15 and 0.9 h  1.
Butanol productivity and butanol-to-ABE selectivity under
steady state conditions for the DTOT investigated are reported in
Fig. 9. During the tests DTOT was not gradually increased but was
selected randomly within the investigated range to check for any
progressive change in the biofilm structure over time. Indeed, the
smooth gradual change of the data reported vs. DTOT is an intrinsic
measure of the consistency of the results.
Once the value of DTOT has been set, it is interesting to analyse
the substrate and metabolite concentration in the four reactors
from the PFR behaviour point of view. The gradual decrease in
lactose concentration (Fig. 7A), the acid concentration peak (Fig. 8)
and the gradual increase in solvent concentration (Fig. 7B) along
Fig. 6. Acetic acid and butyric acid concentrations measured along the PBBR series
during Tests a, b, c and d.
the PBBR series reproduce the typical behaviour of batch fer-
mentation. As expected, the formation of a stable biofilm in the
four units and the choice of the proper operating conditions as
reported above (§ 2.5) lead to a successful reactor system that
behaves like a PFR with continuous inoculum and constant biofilm
concentration.
Some remarks on the results reported in Figs. 6 through 8 and
Table 2 can be found in the following:

lactose conversion decreased with DTOT (Fig. 9). This result is in
agreement with Qureshi and Maddox (1987) who operated a
packed bed reactor fed with whey permeate;

acid production (acetic and butyric) increased with DTOT (Fig. 8A
and B) in reactor # 1 (operated under acidogenesis conditions);

butanol concentration increased with DTOT in all reactors (Fig. 7B).
At DTOT ¼ 0.9 h  1 butanol concentration in reactor # 3 and # 4 was
constant. In particular, butanol concentration in reactor # 4 slightly
decreased when DTOT was increased from 0.85 h  1 to 0.90 h  1;
 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688 685

Table 2
Acetone and ethanol concentrations at the exit of reactor #4 under steady state
conditions for different dilution rates and initial lactose concentrations in the
100 g/L feed.

DTOT h  1 Acetone g/L Ethanol g/L WA g/Lh WA g/Lh

0.15 0.93 0.46 0.14 0.07

0.25 1.21 0.68 0.30 0.17
0.45 1.67 0.99 0.75 0.45
0.65 2.01 1.33 1.31 0.86
0.75 2.21 1.51 1.66 1.13
0.85 2.44 1.79 2.07 1.52
0.9 2.11 1.65 1.90 1.48

Fig. 7. Butanol and lactose concentration measured along the PBBR series at dif-
ferent DTOT.

Fig. 9. Butanol productivity, lactose conversion degree and butanol selectivity

measured during the operation of the PBRs series in function of the DTOT. Feed-
stock: 100 g/L Lactose.

approached a constant value of about 10.8 g/Lh;

selectivity of butanol Φ decreased from 0.8 to about 0.7 with
DTOT (Fig. 9);

at DTOT ¼ 0.9 h  1 the reactor # 4 did not contribute to the
conversion process. Indeed, the concentration of lactose and
metabolites in reactor # 4 was the same as in reactor # 3
(Figs. 7 and 8). The absence of further conversion was probably
due to the high concentrations of butanol reached in PBBR # 3
(about 10 g/L) close to the inhibiting level for the microorgan-
ism. Indeed, Papoutsakis (2008) reported that butanol titers
rarely exceed 12–13 g/L when the wild type C. acetobutylicum is
used.

The PBBR system was stopped after more than 3 months of

Fig. 8. Acetic acid and butyric acid concentrations measured along the PBRs series
at different DTOT.

acetone and ethanol final concentration (at the exit of reactor
# 4) increased with DTOT (Table 2). At DTOT ¼ 0.9 h  1 acetone
and ethanol concentration in reactor # 3 and # 4 was constant
(data not reported). In particular, acetone and ethanol concen-
tration in reactor #4 slightly decreased when DTOT was in-
creased from 0.85 h  1 to 0.90 h  1;

specific butanol productivity of the PBBR system vs. DTOT
operation and the overall biomass concentration in each rector
(XTOT) was about 180 gDM/L. The biomass-to-carrier ratio was 0.18
gDM /g. The biomass concentration was larger than the value re-
ported by Qureshi et al. (2005) - bonechar as carriers -
(0.087gDM/g) and by Raganati et al. (2013) - Tygon rings as carriers
- (0.1 gDM/L) using cheese-whey as carbon source. This biofilm
concentration measured for the proposed PBBR system may be
due to the procedure used for the biofilm build-up and to the used
operation strategy.
The analysis of the effect of the dilution rate on the perfor-
mance of the PBBR system points out that two strategies can be
followed to optimise the system: feedstock conversion and buta-
nol concentration. The scenarios are:

complete lactose conversion is possible at DTOT o0.15 h  1;
686 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688

maximum concentration of solvent (acetone, butanol and ethanol)

in stream produced is possible at DTOT 4 0.8 h  1. In particular, at
DTOT ¼ 0.85 h  1 about 60 g/L of lactose are converted (at a rate of
50 gsugar/Lh), and about 1.9 g/Lh of acetone, 9.2 g/Lh of butanol and
1.48 g/Lh of ethanol are produced respectively.

The PBBR performances can also be compared with those of the

single PBBR reported by Napoli et al. (2010). The specific pro-
ductivity of the single PBBR and of the 4 PBBRs in series at the
same DTOT (E0.95 h  1) is about 4.5 g/Lh and 9 g/Lh, respectively.
The PBBR system doubles butanol productivity. This improvement
of the specific productivity is the result of the separation of the
two fermentation phases (acidogenesis and solventogenesis) and
of the progressive conversion of the substrate for a reactive system
characterized by product inhibition. The single stage PBR beha-
viour approaches the CSTR (continuous stirred tank reactor) limit:
the whole system works under the final conditions (high acids and
solvents concentrations), which inhibit the cell growth and the
solvent production.
The best PBBR system performance assessed in this work was
included in the map by Setlhaku et al. (2013) (Fig. 1). The data
point (full circle in the Fig. 1) is closer to the “operating windows”
than the operating point found by Napoli et al. (2010) (7-im-
point). It may be observed that: i) the butanol productivity is al-
most the highest reported in the literature for t C. acetobutylicum
fermentation; ii) the butanol concentration in the exit fermenta-
tion broth is the maximum for the immobilised class of reactors.
Higher butanol concentrations can be obtained using Clostridium
strains with higher butanol critical concentration – Harris et al.
(2001) reported strains that produce up to 17.6 g/L butanol (total
solvents: 28 g/L); Ezeji et al. (2007) reported the C. beijerickii strain
BA101 that produces high total solvent titers (up to 33 g/L) in
batch fermentations –, but can also be obtained with technological
solutions, as pointed out by Setlhaku et al. (2013).
Fig. 10 reports the results of the cell population assessment in
the biofilm of PBBR # 2, # 3 and # 4 according to the procedure
reported in §2.7 “Interpretative model of the PBBR system per-
formance”. The concentration of the active acidogenic cells
(XA,active), active solventogenic cells (XS,active) and inactive cells
(Xinactive) is reported as a function of the total dilution rate. As a
matter of fact, Xinactive is a measure of the thickness of the biofilm
close to the carrier that does not contribute to the acid/solvents
production.
The analysis of the data reported in the figures suggests that:

the biofilm formed mainly during the PBBR system start-up

stage was sufficient to support the conversion process. Indeed,
cells converting the substrates in solvents were a small fraction
of the biofilm (Fig. 10A). For PBBR # 3 and # 4, the active cell
fraction slightly decreased with DTOT and it was lower than 15-
20% of the total biomass. The active cell fraction of PBBR #
Fig. 10. Estimated XA,active, XS,active and XD in each reactor of the series - (Eqs. (12),
2 decreased significantly with DTOT; 13)–(15) - vs. the DTOT. Feedstock: 100 g/L Lactose.

the fraction XA,active was very low ( o2%) and did not change
remarkably with D. The increase in XA,active at high DTOT may be
due to the dominant effect of the lactose concentration increase coupled with a not so high metabolite concentration. The drop
on the acidogenic stage with respect to the increase in the in- of XS,active in PBBR # 4 at the maximum DTOT investigated was
hibitor metabolite concentration. The small increase in XA,active due to the particularly harsh conditions.
along the PBBR series may be due to the decrease in acids. Al-
though the solvent concentration increases and the lactose Altogether, the structure of the biofilm is satisfactory. The
concentration decreases, the beneficial effects of the decrease in structure proposed offers a sound interpretation of the observed
acid concentration may favour acidogenic activity; concentration profiles along the PBBR system.

as expected, the solventogenic cells were the largest fraction of Table 3 summarizes the results of previous works on con-
the active biofilm and they typically increased with DTOT. The tinuous biolfilm reactors reported in the literature. The analysis of
increase in XS,active with DTOT was more marked in PBBR # the results suggests that:
2 than in the others. The high sensitivity of XS,active in PBBR #
2 was probably due to the increase in lactose concentration
the PBBR system proposed outperforms other continuous
 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688 687

biofilm reactors fed with lactose-whey. Indeed, the ABE con-

porous ma-
tygon rings

tygon rings

tygon rings
trix of PVA
wood pulp

wood pulp
centration (15 g/L) is 2.5 times higher that the one reported by

bone char

corn stalk
clay brick
Support

bagasse
Qureshi and Maddox (1988) and Raganati et al. (2013) and
3 times higher than the concentration reported by Napoli et al.
(2010). The ABE productivity (12.8 g/Lh) is more than twice as
ABE Concentration

much as that reported by Qureshi and Maddox (1988), Raganati

et al. (2013) and Napoli et al. (2013);

the PBBR system proposed hits two targets - high butanol/ABE
concentration and high butanol/ABE productivity - under the
same operating conditions. Chang et al. (2016) reported high
5.8

7.9

5.2

7.2

7.2

11.8

19.9
g/L

15
butanol concentration (12.4 g/L) at very low butanol productiv-
ity (0.5 g/Lh). On the contrary, Survase et al. (2012) reported
Butanol Concentra-

high ABE productivity (13.7 g/Lh) at low butanol concentration

(4.5 g/L).

The performance of the PBBR system proposed is among the

tion g/L

highest ever reported in the literature, in terms of both pro-

12.4

10.8
n.a.

n.a.

n.a.
4.6

4.5

4.9

7.8

ductivity and concentration of butanol.

Sugar Conver-

4. Conclusions
sion -

0.35

0.62

0.68

0.39

0.84

0.61

An innovative continuous reactor system consisting of four

0.3

0.3

0.8

packed bed biofilm reactors (PBBRs) connected in series was

presented. The reactor system was successfully operated for more
Feeding Sugar Con-

than three months to produce butanol by means of Clostridium

centration g/L

acetobutylicum using lactose as carbon source. The reactor per-

formance – butanol productivity and butanol concentration – was
assessed as a function of the reactor system dilution rate. By
tuning the DTOT it was possible: i) to totally convert the carbon
60

60

30

60

28

58

60

71

100

source (DTOT o0.15 h  1); ii) to maximize the concentration of

butanol in the produced stream (DTOT 4 0.8 h  1); iii) to maximize
ABE Productivity

the butanol productivity (about 9.2 g/Lh at DTOT ¼ 0.85 h  1).

The biofilm was assumed to consist of two layers: an active
shell and an inner inactive layer. The extent of these layers de-
0.59
g/Lh

5.8

15.8

13.7

3.2

5.5

0.8

12.8

pended on DTOT. This structure offers a sound interpretation of the

5

concentration profiles observed along the PBBR system.

The reported results suggest that cheese whey - a by-product of
cheese-making that is very rich in lactose (up to about 70 g/L) –
ductivity g/Lh
Butanol Pro-

can be used as a substrate for solvent production, as reported in

previous studies (Raganati et al., 2013; Foda et al., 2010). The PBBR
system is able to completely convert the lactose from the cheese
0.39
n.a.

n.a.

n.a.
4.4

8.6

2.7

0.5

9.2

whey stream without any dilution and supplementation of nutri-

tional microelements thus reducing the pressure on disposal
0.4 gABE

0.3 gABE
Yield g/

0.38 gB
gABE/g

gABE/g

gABE/g

gABE/g

gABE/g

gABE/g

systems.
0.32

0.28

Cheese whey 0.32

0.24

0.32

0.25
/g

/g

/g
g

Corn stalk
D h  1 Substrate

Acknowledgements
permeate

Lactose-
Glucose

Glucose

Glucose

Glucose

Lactose
Summary of the literature on continuous biofilm bioreactors.

Whey

juice

The authors thank the Ministero dello Sviluppo Economico for

their financial support to the project EuroTransBio ETB-2012-16
0.54

0.75

0.05

0.04

0.85
0.97

OPTISOLV (Development, optimization and scale-up of biological

1.9
1

solvent production) (C01/0878/01-03/X21) as well as the BMBF

C. acetobutylicum ATCC

C. acetobutylicum ATCC
C. acetobutylicum DSM

C. acetobutylicum DSM

C. acetobutylicum DSM

(Germany) for supporting the project OPTISOLV (FKZ: 031A231C).

C. acetobutylicum ABE
C. beijerinckii BA101
C. acetobutylicum

C. acetobutylicum
Microorganism

References
792-ADH

1201

Bankar, S.B., Jurgens, G., Survase, S., Ojamo, H., Granström, T., 2014. Enhanced iso-
792
824

824

792

propanol–butanol–ethanol (IBE) production in immobilized column reactor

using modified Clostridium acetobutylicum DSM792. Fuel 136, 226–232.
Maddox (1988)

Bohlmann, G., 2007. Biobutanol. SRI Consulting, California.

Raganati et al.
Qureshi et al.

Survase et al.

Cascone, R., 2008. Biobutanol – a replacement for bioethanol? Chem. Eng. Prog.,
Bankar et al.
Qureshi and

Napoli et al.

Chang et al.
Dolejš et al.

S4–S9.
This work
(2000)

(2012)
(2010)

(2013)

(2014)

(2014)

(2016)

Chang, Z., Cai, D., Wang, Y., Chen, C., Fu, C., Wang, G., Qin, P., Wang, Z., Tan, T., 2016.
Table 3

Effective multiple stages continuous acetone–butanol–ethanol fermentation by

Ref.

immobilized bioreactors: making full use of fresh corn stalk. Bioresour. Technol.
205, 82–89.
688 F. Raganati et al. / Chemical Engineering Science 152 (2016) 678–688
Chen, C.K., Blaschek, H.P., 1999. Acetate enhances solvent production and prevents de-
generation in Clostridium beijerinckii BA101. Appl. Microbiol. Biotechnol. 52, 170–173.
Desai, R., Nielsen, L.K., Papoutsakis, E.T., 1998. Stoichiometric modeling of Clos-
tridium acetobutylicum fermentations with non-linear constraints. J. Biotechnol.
71, 191–205.
Dolejš, I., Krasňan, V., Stloukal, R., Rosenberg, M., Rebroš, M., 2014. Butanol pro-
duction by immobilised Clostridium acetobutylicum in repeated batch, fed-
batch, and continuous modes of fermentation. Bioresour. Technol. 169,
723–730.
Dürre, P., 2007. Biobutanol: an attractive biofuel. Biotechnol. J. 2, 1525–1534.
Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007. Bioproduction of butanol from biomass:
from genes to bioreactors. Curr. Opin. Biotechnol. 18, 220–227.
Foda, M.I., Dong, H., Li, Y., 2010. Study the suitability of cheese whey for bio-butanol
production by clostridia. J. Am. Sci. 6 (8), 39–46.
Friedl, A., 2012. Lignocellulosic biorefinery. Env. Eng. Manag. J. 11, 75–79.
Gapes, J.R., Nimcevic, D., Friedl, A., 1996. Long-term continuous cultivation of
Clostridium beijerinckii in a two-Stage chemostat with on-line solvent removal.
Appl. Environ. Microb. 62 (9), 3210–3219.
Harris, L.M., Blank, L., Desai, R.P., Welker, N.E., Papoutsakis, E.T., 2001. Fermentation
characterization and flux analysis of recombinant strains of Clostridium acet-
obutylicum with an inactivated solR gene. J. Ind. Microbiol. Biotechnol. 27,
322–328.
Jones, D.T., Woods, D.R., 1986. Acetone-butanol fermentation revisited. Microbiol.
Rev. 50 (4), 484–524.
Krouwel, P.G., Van der Laan, W.F.M., Kossen, N.W.F., 1980. Continuous production of
n-butanol and isopropanol by immobilized, growing Clostridium butylicum
cells. Biotechnol. Lett. 2, 253–258.
Kumar, M., Gayen, K., 2011. Developments in biobutanol production: new insights.
Appl. Energy 88 (6), 1999–2012.
Kumar, M., Goyal, Y., Sarkar, A., Gayen, K., 2012. Comparative economic assessment
of ABE fermentation based on cellulosic and non-cellulosic feedstocks. Appl.
Energy 93, 193–204.
Lee, S., Cho, M., Park, C., 2008. Continuous butanol production using suspended and
immobilized Clostridium beijerinckii NCIMB 8052 with supplementary butyrate.
Energy Fuel 22, 3459–3464.
Levenspiel, O., 1999. Chemical reaction engineering, 3rd ed. Wiley, New York.
Mariano, A.P., Filho, R.M., 2012. Improvements in biobutanol fermentation and their
impacts on distillation energy consumption and wastewater generation. Bioe-
nerg. Res. 5, 504–514.
Meyer, C.L., Papoutsakis, E.T., 1989. Continuous and biomass recycle fermentations
of Clostridium acetobutylicum. Part 1. Bioprocess Eng. 4, 1–10.
Napoli, F., Olivieri, G., Russo, M.E., Marzocchella, A., Salatino, P., 2010. Butanol
production by Clostridium acetobutylicum in a continuous packed bed reactor. J.
Ind. Microbiol. Biothechnol. 37, 603–608.
Napoli, F., Olivieri, G., Russo, M.E., Marzocchella, A., Salatino, P., 2011. Continuous
lactose fermentation by Clostridium acetobutylicum – assessment of acidogen-
esis kinetics. Bioresour. Technol. 102, 1608–1614.
Napoli, F., Olivieri, G., Russo, M.E., Marzocchella, A., Salatino, P., 2012. Continuous
lactose fermentation by Clostridium acetobutylicum – assessment of energetics
and product yields of the acidogenesis. Enzym. Microb. Technol. 50, 165–172.
Olivieri, G., Russo, M.E., Marzocchella, A., Salatino, P., 2011. Modeling of an aerobic
biofilm reactor with double-limiting substrate kinetics: bifurcational and dy-
namical analysis. Biotechnol. Prog. 27, 1599–1613.
Papoutsakis, E.T., 2008. Engineering solventogenic clostridia. Curr. Opin. Biotechnol.
19 (5), 420–429.
Procentese, A., Raganati, F., Olivieri, G., Russo, M.E., Salatino, P., Marzocchella, A.,
2015a. Continuous lactose fermentation by Clostridium acetobutylicum – as-
sessment of solventogenic kinetics. Bioresour. Technol. 180, 330–337.
Procentese, A., Raganati, F., Olivieri, G., Russo, M.E., Salatino, P., Marzocchella, A.,
2015b. Continuous xylose fermentation by Clostridium acetobutylicum – as-
sessment of solventogenic kinetics. Bioresour. Technol. 192, 142–148.
Qureshi, N., Maddox, I.S., 1987. Continuous solvent production from whey permeate
using cells of Clostridium acetobutylicum immobilized by adsorption onto
bonechar. Enzym. Microb. Technol. 9, 668–671.
Qureshi, N., Maddox, I.S., 1988. Reactor design for ABE fermentation using cells of
Clostridium acetobutylicum immobilized by adsorption onto bonechar. Biopro-
cess Eng. 3, 69–72.
Qureshi, N., Schripsema, J., Lienhardt, J., Blaschek, H.P., 2000. Continuous solvent
production by Clostridium beijerinckii BA101 immobilized by adsorption onto
brick. World J. Microb. Biotechnol. 16, 337–382.
Qureshi, N., Annous, B.A., Ezeji, T.C., Karcher, P., Maddox, I.S., 2005. Biofilm reactors
for industrial bioconversion processes: employing potential of enhanced reac-
tion rates. Microb. Cell Fact. 4, 24.
Raganati, F., Olivieri, G., Procentese, A., Russo, M.E., Salatino, P., Marzocchella, A.,
2013. Butanol production by bioconversion of cheese whey in a continuous
packed bed reactor. Bioresour. Technol. 138, 259–265.
Raganati, F., Procentese, A., Olivieri, G., Götz, P., Salatino, P., Marzocchella, A., 2015.
Kinetic study of butanol production from various sugars by Clostridium acet-
obutylicum using a dynamic model. Biochem. Eng. J. 99, 156–166.
Rai, V., Upadhyay, R.K., Thakur, N.K., 2012. Complex population dynamics in het-
erogeneous environments: effects of random and directed animal movements.
Int. J. Nonlinear Sci. Numer. Simul. 13, 299–309.
Sarchami, T., Rehmann, L., 2014. Optimizing enzymatic hydrolysis of inulin from
Jerusalem artichoke tubers for fermentative butanol production. Biomass
Bioenergy 69, 175–182.
Schugerl, K., 1997. Three-phase-biofluidization – application of three-phase fluidi-
zation in the biotechnology - a review. Chem. Eng. Sci. 52 (21–22), 3661–3668.
Setlhaku, M., Heitmann, S., Górak, A., Wichmann, R., 2013. Investigation of gas
stripping and ervaporation for improved feasibility of two-stage butanol pro-
duction process. Bioresour. Technol. 136, 102–108.
Survase, S., van Heiningen, A., Granström, T., 2012. Continuous bio-catalytic con-
version of sugar mixture to acetone-butanol-ethanol by immobilized Clos-
tridium acetobutylicum DSM 792. Appl. Microbiol. Biotechnol. 93 (6),
2309–2316.
Tashiro, Y., Takeda, K., Kobayashi, G., Sonomoto, K., 2005. High production of
acetone-butanol-ethanol with high cell density culture by cell-recycling and
bleeding. J. Biotechnol. 4 (120), 197–206.
Welsh, F.W., Williams, R.E., Veliky, I.A., 1987. Solid carriers for a Clostridium acet-
obutylicum that produces acetone and butanol. Enzym. Microb. Technol. 9,
500–502.
Yen, H.W., Li, R.J., Ma, T.W., 2011. The development process for a continuous acet-
one–butanol–ethanol (ABE) fermentation by immobilized Clostridium acet-
obutylicum. J. Taiwan Inst. Chem. E 42, 902–907.
Yu, Y., Tangney, M., Aass, H.C., Mitchell, W.J., 2007. Analysis of the mechanism and
regulation of lactose transport and metabolism in Clostridium acetobutylicum
ATCC 824. Appl. Environ. Microbiol. 73 (6), 1842–1850.
Viikilä, M., Wallenius, J., Ojamo, H., Granström, T., Survase, S., 2013. Impact of
varying lignocellulosic sugars on continuous solvent production in an im-
mobilized column reactor. Bioresour. Technol. 147, 299–306.