Abstract Using a continuous enrichment technique, a
bacterial consortium capable of degrading 4-chlorophenol
(4-CP) was obtained from the rhizosphere of Phragmites
australis. A granular activated carbon (GAC) biofilm
reactor was established using this consortium, and the
degradation of 4-CP was investigated under continuous
flow operation using a feed of 20–50 mg l
–1
with a
hydraulic residence time of 17 min over a 6-month period.
Chloride liberation occurred throughout the operation,
and the reactor had 4-CP removal efficiencies of
69–100%. Periods of lower performance were attributed
to clogging of the column with biomass and the formation
of channels. Subsequently, the immobilized biofilm was
subjected to a starvation period of 5 months, after which its
degradative capacity was still maintained. The microbial
consortium was characterized during the continuous
flow experiment and dynamic population changes were
observed throughout. One isolate recovered from the
biofilm was shown to be capable of degrading 4-CP as a
sole carbon and energy source.
Introduction
Chlorinated organic compounds form one of the most
important groups of xenobiotic chemicals that enter the
environment. These include some of the most useful and
economically important chemicals available to industry
and agriculture, being extensively used as herbicides,
insecticides, fungicides, heat transfer media, insulators and
lubricants (Keith and Telliard 1979; Barbash and Roberts
1996). Due to their extensive use, these compounds are
common soil and water contaminants (Mason 1991;
Chapelle 1993). Chloroaromatic compounds are susceptible
to biodegradation, and aerobic degradation often
involves their mineralization to innocuous products,
thereby making the utilization of these reactions especially
attractive in remediation applications. It has been reported
that 4-chlorophenol (4-CP) can be partially or completely
degraded aerobically by several bacteria including
Pseudomonas (Knackmuss and Hellwig 1978), Azotobacter
(Wieser et al. 1994), Alcaligenes (Hill et al. 1996), and
Rhodococcus (Finkel’shtein et al 2000), and via different
biodegradation pathways (Bae et al. 1996; Hollender et
al. 1997).
Biological waste treatment techniques for water
and effluents contaminated with chlorinated organic
compounds often include the utilization of biofilm
reactors. Adsorption and biological treatment are two
common approaches used to treat such compounds (Jonge
et al. 1996; Caldeira et al. 1999). Granular activated
carbon (GAC) biofilm reactors can combine these two
features; the adsorptive capacity and irregular shape of
GAC particles provide niches for bacterial colonization
protected from high fluid forces (Christensen and
Characklis 1990), while the variety of functional groups
on the surface can enhance the attachment of micro-
organisms (Weber et al. 1979). In these systems, biomass
is active even at very low concentrations of target organic
chemicals, making it less sensitive to the presence of
toxic and inhibitory materials and more resistant to
shock loading of toxics than dispersed growth systems
(Lee et al. 1994; Fauzi 1995; Shi et al. 1995). In the
fermentation field, biofilm systems have also been
shown to lead to increased bacterial lactic acid produc-
tivities, while protecting bacteria from the effect of anti-
biotics used to subdue yeast contaminations (Velásquez
et al., 2001). The continuous removal of pollutants in
biofilm reactors has been reported by several authors.
Feakin et al.(1995) used a GAC fixed bed reactor for the
treatment of 1–10 µg l
–1
atrazine and simazine in surface
water and achieved removal efficiencies of 53% and
58%, respectively. Massol-Deya et al. (1995) reported a
M.F. Carvalho · I. Vasconcelos · P.M.L. Castro (
✉
)
Escola Superior de Biotecnologia,
Universidade Católica Portuguesa,
Rua Dr. António Bernardino de Almeida, 4200–072 Porto, Portugal
e-mail: plc@esb.ucp.pt
Tel.: +351-22-5580059, Fax: +351-22-5090351
M. F. Carvalho · I. Vasconcelos · A. T. Bull
P. M. L. Castro
A GAC biofilm reactor for the continuous degradation
of 4-chlorophenol: treatment efficiency and microbial analysis
98% reduction of a 3 mg l
–1
influent concentration of
toluene in a fluidized bed reactor-fed toluene-amended
groundwater. Similarly, treatment of groundwater con-
taminated with chlorobenzene at concentrations up to
170 mg l
–1
in a GAC fluidized-bed reactor was achieved
with more than 99% efficiency (Klecka et al. 1996).
Compact matrixes, based on wheat husk and wheat bran,
have been successfully used to immobilize the fungus
Coriolus versicolor to treat effluents contaminated with
dichlorophenol and pentachlorophenol at 50 mg l
–1
, for
which removal rates of 75–80% and 100%, respectively,
were observed within 24 h of batch cultures (Ullah et al.
2000).
The present report extends our previous research on
GAC biofilms (Caldeira et al. 1999) by evaluating 4-CP
removal in a continuous flow bioreactor, in order to
determine whether steady-state conditions could be
achieved, and by analysing the robustness of the system
to withstand long periods of contaminant starvation.
Such reactor systems closely mimic those most
commonly deployed for effluent and water biotreatment,
which often face variations in concentrations of the
contaminant feed. The efficiency of the GAC biofilm
reactor to continuously biodegrade 4-CP from an aqueous
feed, supplied at different concentrations over a period
of 6 months, and the stability of the degrading activity
after a prolonged period of starvation are discussed. In
addition, microbiological characterization of the biofilm
bacteria is described.
Materials and methods
Preparation of the bacterial inoculum
The bacteria used in this study were isolated from soil samples
collected from the rhizosphere of the reed Phragmites australis,
obtained from a chemically contaminated industrial site in Northern
Portugal (Estarreja), which has received discharges of liquid
effluents for over 50 years. The bacteria were recovered from the
soil by a dispersion and differential centrifugation technique based
on that described by Hopkins et al. (1991). Rhizosphere soil
samples (10 g) were blended mechanically with 10 ml sodium
cholate solution (0.1% w/v) for 1 min, after which 10 ml sodium
cholate solution, 10 g chelating resin (sodium form of iminodiacetic
acid) and approximately 30 glass beads (~5 mm diameter) were
added. The resultant preparation was shaken for 2 h at 5°C and
centrifuged at 500 g for 2 min. The supernatant (a1) was kept for
subsequent inoculum formulation. The same procedure was used
to obtain a further series of supernatants by successively extracting
the soil pellet with the following series of solutions: 10 ml Tris
buffer pH 7.4 (a2), 30 ml sodium cholate solution followed by
ultrasonication for 1 min (b1), 10 ml Tris buffer (b2), and twice
40 ml of distilled water (c1+c2). The final supernatants consisted
of: A (a1+a2), B (b1+b2) and C (c1+c2). A mixture of supernatants
A (5 ml), B (7.5 ml), and C (10 ml) was used to inoculate a batch
flask containing 300 ml of minimal salts medium (Caldeira el al.
1999). This composite preparation was used as the inoculum for
biofilm reactor studies.
Biofilm development in a packed bed GAC column
The composite bacterial preparation obtained from the rhizosphere
soil samples was used to inoculate a tubular glass column
(26 cm×2.6 cm), packed with 15 g of thermally activated
commercial GAC (12–18 mesh). Glass wool was placed in the
bottom of the column to hold the GAC in place. The reactor was
maintained at room temperature and operated in a downflow
mode. The inoculum was recirculated through the packed column for
5 days to allow colonization of the support material. The decrease
(>90%) in the optical density of the circulating inoculum was used as
an indication of biomass loading onto the GAC column. Thereafter,
an enrichment of 4-CP-degrading bacteria proceeded over a
50 day period. During this regime, the column was periodically
supplied (1–2 times per week) with 300 ml minimal salts medium
supplemented with 4-CP and phenol (10 mg l
–1
each), which was
recirculated through the column with a hydraulic residence time
(HRT) of 25 min. The biofilm activity was monitored by chloride
release due to the biodegradation of 4-CP. After 30 days of biofilm
enrichment, the bacteria attached to the GAC were enumerated.
The biomass was extracted by shearing duplicate samples (1 g
GAC particles) on a vortexing machine operating at maximum
speed for 30 s. Diluted suspensions (0.1 ml) were plated onto
nutrient agar (NA) and minimal agar supplemented with 4-CP.
Continuous operation of the biofilm reactor
Aerated minimal medium containing 4-CP was continuously
passed through the GAC reactor, during a 6-month period, with an
HRT of 17 min, based on the GAC bed volume of 33 cm
3
used for
the experiments. Aeration was supplied to the feeding vessel by a
compressor working at 0.6 bar. Air was filtered (Nalgene SFCA
0.2 µm) before entering the vessel, and was fed into the mineral
medium by means of a submerged silicone tube (internal diameter
2.8 mm, 0.8 mm thick, approximately 0.75 m length). At different
stages, the 4-CP concentration in the feed was varied as follows: on
days 0–51, 52–87, 88–113, 114–166 and 167–200, 4-CP concentra-
tions in the feed were 50 mg l
–1
, 25 mg l
–1
, 0mgl
–1
, 25 mg l
–1
and
20 mg l
–1
, respectively. Chlorophenol and chloride concentrations
in the column effluent were analysed as described previously
(Caldeira et al. 1999); the presence of phenol could be detected
using the same analytical procedure as for 4-CP, which was verified
by analysing the separation of standard mixtures of 4-CP and
phenol. Bacteria eluted from the biofilm column were enumerated
every month. Effluent samples were diluted in 0.85% (w/v) NaCl
solution, and 0.1 ml of appropriate dilutions were spread plated
onto NA and minimal salts agar supplemented with 4-CP.
Reactor operation under starvation conditions
After 6 months of continuous operation, the biofilm reactor was
starved of 4-CP and the liquid flow suspended for a further
5 months. After that period the column was operated again under a
recirculating regime (9 days), during which the column was fed
with 2 l of minimal salts medium supplemented with 4-CP
(50 mg l
–1
, on days 0 and 4), with an HRT of 17 min. This recircu-
lating mode was followed by continuous operation, during which
the column was fed with 4-CP (50 mg l
–1
) with the same retention
time. The column was maintained at room temperature.
Biofilm microbial community
Bacteria attached to the GAC were enumerated and characterized
at different stages of bioreactor operation: at the beginning (day 0),
in the middle (day 90) and at the end (day 200) of the continuous
experiment; bacteria were also enumerated after the starvation
period. GAC biomass loading (CFUs g
–1
GAC) was determined
after extracting the biomass as described above. Samples of
approximately 0.5–1 g were taken in duplicate, collected from the
inner middle section of the bed; after each sampling the amount of
GAC extracted was replaced with fresh GAC. The bacterial
suspensions were spread plated onto NA and minimal salts agar
supplemented with 4-CP and, at the same time, used as a 20%
inoculum for batch flasks containing 30 ml of minimal salts medium
supplemented with 4-CP (50 mg l
–1
) in order to verify 4-CP
biodegradation.
Bacteria recovered from the GAC matrix were purified by
repeated subculturing on NA and minimal salts agar supplemented
with 4-CP. A preliminary characterization was based on colony
and cell morphology, presence or absence of cytochrome c oxidase
and Gram staining. Gram-negative isolates were identified using
the API 20 NE system (Biomérieux).
All microorganisms recovered from the GAC were analysed
for their individual capacities to degrade 4-CP. Each colonial type
was suspended in 10 ml of minimal salts medium supplemented
with 25 mg l
–1
4-CP as the sole carbon and energy source. Cultures
were grown aerobically in 100 ml flasks incubated at 25°C on a
rotatory shaker (Julabo-SW-20C) at 100 rpm. Growth was moni-
tored by measuring the optical density of the cultures at 600 nm.
Biodegradation of 4-CP was measured by chloride release to the
medium; residual 4-CP concentration was also determined. When
growth on 4-CP was evidenced by an increase in optical density
and by liberation of chloride, samples of culture were plated on
NA in order to verify whether single species were present. When
pure colonies were obtained, this procedure was repeated several
times.
Reagents
All chemicals were of the highest purity grade available (Sigma,
St. Louis, USA; Aldrich, Dorset, UK; Merck, Darmstadt, Germany).
GAC (12–18 mesh) was obtained from BDH, Dorset, UK. Prior to
use, GAC was washed several times with deionized water to
remove carbon fines, dried in an oven at 105°C, and sterilized by
autoclaving.
Results
Establishment of the 4-CP degrading biofilm
After inoculating the reactor and operating it in a closed
recirculating flow mode, biodegradation of 4-CP was
observed after 3 weeks as the accumulation of chloride
ion in the recirculating vessel. During the subsequent
1-month period, stoichiometric release of chloride
indicated that, under this recirculation mode of operation,
all the supplied 4-CP could be dechlorinated. Thereafter,
the column was maintained for a further 2 months by
recirculating minimal medium, supplemented weekly
with 4-CP.
Biodegradation of 4-CP in the continuous flow
biofilm reactor
Following the establishment of an active 4-CP degrading
biofilm, the performance of the reactor for the continuous
removal of 4-CP was evaluated over a 6-month period.
Chloride and 4-CP in the column outlet were measured
The presence of phenol was also investigated, but this
compound was never found in the column outflow.
During the first 51 days, 4-CP was fed to the column
with a concentration of 50 mg l
–1
(Fig. 1, A). Initially
(days 0–5), 4-CP was not detected in the reactor effluent,
but thereafter (days 6–36) its concentration increased
in the outlet, reaching a mean concentration of
8.9±1.5 mg l
–1
. Chloride release was observed during
this period, increasing steadily until day 14, and reaching
an average concentration thereafter of 35.1±3.6 mg l
–1
until day 50, indicating that the biofilm was active in
spite of not degrading all the 4-CP supplied to the reactor.
After decreasing the 4-CP loading to 25 mg l
–1
(days
52–87), a gradual decrease in the outlet 4-CP concentration
was observed, and from day 68 4-CP was no longer
detected in the effluent (Fig. 1, B), while the level of 4-CP
degraded was maintained at 25.1±3.1 mg l
–1
. Due to
a disruption of the reactor (fractured glass), which
occurred at day 75, the colonized GAC was transferred
to a new column (same dimensions); care was taken so
that the biofilm matrix was disturbed to a minimum.
After restarting the feed, 4-CP was detected in the reactor
effluent (days 75–87), despite the fact that the biofilm
was active during that period, as shown by chloride
liberation. Feeding of 4-CP to the reactor was stopped
temporarily from day 88 to day 113, during which time
4-CP was undetectable in the effluent (Fig. 1, C). Chloride
release continued during this period of suspended feeding,
Fig. 1 Degradation of 4-chloro-
phenol (4-CP) in the biofilm
column during continuous
mode operation. Horizontal
bars 4-CP feed concentrations
during open mode operation
during days A 0–51, B 52–87,
C 88–113, D 114–166,
E 167–200. Open circles 4-CP
degraded based on chloride
release, black circles 4-CP
detected in the column effluent
a result indicating that 4-CP adsorbed to the GAC could
be assimilated by the biofilm community. Between days
114 and 166, 4-CP was reloaded on the column at a
concentration of 25 mg l
–1
. 4-CP was detected in the
effluent at concentrations varying between 0–7.7 mg l
–1
.
Chloride release showed an average concentration of
degraded 4-CP of 15.3±2.0 mg l
–1
. 4-CP concentration
in the outlet decreased to an average level of
1.2±0.5 mg l
–1
when the influent concentration was
reduced to 20 mg l
–1
(days 167–200); 4-CP was not
detected in the effluent at several sampling points
(Fig. 1, E). From day 167 until the last day of operation,
chloride liberation indicated a 4-CP degradation of ca.
16.6±2.0 mg l
–1
.
The quantity of 4-CP degraded during each phase of
the biotreatment is shown in Table 1. During the feeding
period between days 52 and 87, total 4-CP degradation
was higher than the amount supplied for that period,
indicating that 4-CP previously adsorbed to the column
was being biodegraded. Of the total amount of 4-CP
supplied to the reactor, approximately 85% was biode-
graded. The activity of the biofilm was highest during
the first half of bioreactor operation, as shown by the
average values for removal capacity (Table 1).
During the continuous flow operation of the reactor,
the detachment of bacteria from the GAC particles into
the effluent was monitored. Samples were taken every
month and values between 1.4×10
6
cfu ml
–1
and 2.0×10
7
cfu ml
–1
were observed. The organisms present in the
effluent were not characterized.
Biofilm activity after 4-CP starvation
At the end of the continuous flow operation the biofilm
reactor was starved of 4-CP for 5 months, after which
the degrading capacity of the biofilm was tested both
under recirculating and continuous flow modes. During
the recirculating regime, 4-CP was fed twice to the reactor,
and chloride release was immediately observed (Fig. 2),
indicating that the biofilm consortium was still exhibiting
4-CP degrading activity. The levels of chloride release
also indicated that most of the 4-CP was degraded (ca.
85%). The fact that 4-CP was not detected in the
feed recirculation vessel, while chloride was released,
suggested that biodegradation occurred after 4-CP
adsorption to the GAC. After switching to continuous
flow mode, 4-CP degradation stabilized after 20 h, as
indicated by chloride release (Fig. 3). The presence of
4-CP was only detected at low levels in the reactor effluent
in the first 8 h. Chloride release data indicated that 78%
of the total amount of 4-CP added during continuous
mode was biodegraded, while 1% was recovered at the
outlet of the reactor. The activity of the biofilm recovered
completely after 4-CP starvation; the removal capacity
was ca. 3.34 g l
–1
day
–1
, which accords closely with the
values initially obtained for the continuous operation
(Table 1).
Table 1 Degradation of 4-chlorophenol (4-CP) in the granular activated carbon (GAC) reactor during continuous operation
Feed period 4-CP influent load 4-CP in influent 4-CP in effluent 4-CP degraded Removal capacity
(days) (g l
–1
d
–1
) (g) (g) (g) (g l
–1
d
–1
)
A. 0–51 4.24 5.37 0.73 3.72 2.89
a
B. 52–87 2.12 1.86 0.46 2.02 2.27
C. 88–113 0.00 0.00 0.03 0.63 0.99
D. 114–166 2.12 2.84 0.47 1.95 1.47
E. 167–200 1.69 1.28 0.08 1.24 1.47
0–200 11.35 1.76 9.55
a
The values for period A refer only to days 14–51
Fig. 2 Degradation of 4-chlorophenol (4-CP) in the biofilm reactor
after the starvation period (recirculating operation mode). Black
triangles 4-CP concentration in the recirculating vessel, open circles
4-CP degraded based on chloride release
Fig. 3 Degradation of 4-chlorophenol (4-CP) in the biofilm
reactor after the starvation period (continuous operation mode).
Horizontal bar 4-CP concentration in the feed, open circles 4-CP
degraded based on chloride release, black circles 4-CP in the
effluent
Analysis of the biofilm community
GAC biomass loading was determined at different stages
of continuous flow operation. The bacterial population
sizes adsorbed onto GAC at days 0, 90 and 200 were
1.01±0.3×10
8
cfu g
–1
GAC, 2.09±0.6×10
9
cfu g
–1
GAC
and 1.65±0.2×10
9
cfu g
–1
GAC, respectively. The bacteria
recovered from the GAC reactor were investigated for
their 4-CP-degrading capacity; 50 mg l
–1
4-CP was
degraded within 24–29 h, in batch cultures.
During continuous flow operation of the reactor, a
preliminary characterization of the biofilm microbial
community was made. Five different morphological colony
types were obtained from GAC at the beginning of the
continuous flow period, three were recovered on NA, one
on minimal salts medium and one in both media. Prelimi-
nary characterization of the five bacterial strains showed
that all were Gram-negative rods. The isolates were iden-
tified by the API 20 NE system and the results for the
ones recovered in NA are shown in Figure 4. Due to the
small size of the colonies obtained on minimal salts medi-
um agar, bacterial counts were not reliable. The isolation
of the morphologically different colonial types recovered
on that medium and identification by API revealed the
presence of two different bacteria: Burkholderia cepacia
(identification probability: 99.5%) and Agrobacterium
radiobacter (identification probability: 95.7%), which was
also recovered on NA medium. At the mid-stage of the
continuous experiment (day 90), five Gram-negative, rod
shaped, bacterial strains were recovered in NA (Fig. 4). In
minimal salts medium, two types were recovered, identi-
fied by API 20 NE system as B. cepacia (identification
probability: 99.5%) and Sphingobacterium multivorum
(identification probability: 99.9%), which were also present
on NA. Characterization of the bacteria extracted at the
end of the continuous experiment (day 200) and recovered
in NA revealed five Gram-negative rods (Fig. 4). Visual
observation of the bacterial populations recovered in the
effluent during the first 30 days of continuous operation
showed mainly four morphologically different colonial
types, with a predominance (65–90%) of a morphological
type similar to the one identified as A. radiobacter in the
initial GAC extracts. From then on, four to five morpho-
logically different colonial types were recovered at each
sampling stage, with a predominance (55–85%) of a
morphological type similar to the one identified as
S. multivorum in the subsequent GAC extracts.
After the starvation period, bacterial communities of
the biofilm were also extracted from the GAC particles
and enumerated. Bacterial populations were found in the
order of 1.4±0.3×10
7
cfu g
–1
GAC, a lower value than
during continuous flow operation. Seven morphologically
different organisms were recovered on NA media and
two were recovered on minimal agar plates supplemented
with 4-CP. These organisms were not characterized.
The degradation of 4-CP by individual species was
tested in liquid medium. In control experiments, without
inoculation, no chloride release was observed. Only one
organism, identified by the API 20 NE system with low
discrimination between Chryseomonas luteola, B. cepacia
or Sphingomonas paucimobilis, was capable of degrading
4-CP in a mono-species culture at 4-CP concentrations
of 25 and 50 mg l
–1
. In the latter case, 4-CP was
completely biodegraded after 30 h.
Discussion
Microbial populations in the rhizosphere are one to two
orders of magnitude larger than those in adjacent soil
Fig. 4 Characterization of the
bacteria consortia isolated from
the biofilm reactor throughout
the continuous flow operation,
recovered in NA. The values in
parenthesis show API identifi-
cation probability (LD – low
discrimination)
