Waste Management 32 (2012) 327–334

Contents lists available at SciVerse ScienceDirect

Waste Management
journal homepage: www.elsevier.com/locate/wasman

Fate and transport of phenol in a packed bed reactor containing simulated

solid waste
Jovita M. Saquing ⇑, Detlef R.U. Knappe, Morton A. Barlaz
Department of Civil, Construction, and Environmental Engineering, Campus Box 7908, North Carolina State University, Raleigh, NC 27695-7908, United States

a r t i c l e i n f o a b s t r a c t

Article history: An assessment of the risk to human health and the environment associated with the presence of organic
Received 4 March 2011 contaminants (OCs) in landfills necessitates reliable predictive models. The overall objectives of this
Accepted 8 September 2011 study were to (1) conduct column experiments to measure the fate and transport of an OC in a simulated
Available online 19 October 2011
solid waste mixture, (2) compare the results of column experiments to model predictions using HYDRUS-
1D (version 4.13), a contaminant fate and transport model that can be parameterized to simulate the lab-
Keywords: oratory experimental system, and (3) determine model input parameters from independently conducted
Organic contaminant
batch experiments. Experiments were conducted in which sorption only and sorption plus biodegrada-
Fate
Transport
tion influenced OC transport. HYDRUS-1D can reasonably simulate the fate and transport of phenol in
Landfill an anaerobic and fully saturated waste column in which biodegradation and sorption are the prevailing
HYDRUS-1D fate processes. The agreement between model predictions and column data was imperfect (i.e., within a
factor of two) for the sorption plus biodegradation test and the error almost certainly lies in the difficulty
of measuring a biodegradation rate that is applicable to the column conditions. Nevertheless, a biodeg-
radation rate estimate that is within a factor of two or even five may be adequate in the context of a land-
fill, given the extended retention time and the fact that leachate release will be controlled by the
infiltration rate which can be minimized by engineering controls.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction using chemical warfare agents (CWAs) can result in contaminated

With 54.2% of municipal solid waste (MSW) discarded in land-
fills in 2008, landfill disposal remains a major waste management
option in the United States (US EPA, 2008). An assessment of the
risk to human health and the environment associated with the
presence of organic contaminants (OCs) in landfills necessitates
the use of reliable predictive models (Kjeldsen et al., 2002; Staley
et al., 2006; Oman and Junestedt, 2008; Eklund et al., 1998; Paxeus,
2000; Schwarzbauer et al., 2002). Recent US government guidance
that recommends the disposal of discarded medications in MSW
(California Integrated Waste Management Board, 2010; Office of
National Drug Control Policy, 2007; Florida Department of Environ-
mental Protection, 2006) may lead to an increased presence of OCs
such as endocrine-disrupting chemicals in landfill leachate (Mus-
son et al., 2008). For example, pharmaceuticals (propyphenazone,
ibuprofen), pesticides (N,N-diethyl-3-toluamide (DEET), meco-
prop), and plasticizers (di(2-ethylhexyl) phthalate (DEHP), N-
butylbenzenesulfonamide (NBBS), bisphenol A) were detected in
a recently excavated 25-year-old landfill waste deposit (Schwarz-
bauer et al., 2006). Further, an attack on a building or outdoor area
⇑ Corresponding author. Tel.: +1 919 760 0994; fax: +1 919 515 7908.
E-mail addresses: jmsaquing@gmail.com (J.M. Saquing), knappe@ncsu.edu
(D.R.U. Knappe), barlaz@ncsu.edu (M.A. Barlaz).
debris requiring disposal. Given the potential for large quantities of
debris to be generated, it is important to evaluate the safety of a
landfill as a disposal alternative.
The development of plans for the safe disposal of large quanti-
ties of contaminated debris in MSW requires information on the
behavior of such debris in landfills. Predictive contaminant trans-
port models can describe quantitatively the fate and transport of
specific OCs and generate input data to estimate the potential risks
associated with the disposal of specific compounds.
The fate and transport of OCs is governed by a number of pro-
cesses including sorption/desorption, volatilization, biodegrada-
tion, hydrolysis, diffusion, and advective flow. Some landfill
contaminant transport models were developed to examine emis-
sions through cover and liner systems (Foose et al., 2002; Jury
et al., 1990; Lin and Hildemann, 1995) while other models incorpo-
rated different fate routes (Reinhart et al., 1991; Kjeldsen and
Christensen, 2001; Scheutz et al., 2007; Lowry et al., 2008). In
the model for organic chemicals in landfills (MOCLA), the distribu-
tion of OCs in leachate, gas, and solid waste components and re-
moval of OCs through leachate and gas transport were estimated.
MOCLA assumes a homogeneous and completely mixed landfill
with constant gas and leachate production rates (Kjeldsen and
Christensen, 2001). MOCLA-FOAM is a modified version to account
for time-dependent release patterns of blowing agents from foam

0956-053X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.wasman.2011.09.017
328 J.M. Saquing et al. / Waste Management 32 (2012) 327–334
present in landfills (Scheutz et al., 2007). During a recent numerical
evaluation of CWA fate in landfills, MOCLA was modified to include
contaminant diffusion through the landfill liner (Bartelt-Hunt
et al., 2006). Following the MOCLA approach, a landfill coupled-
reactor (LFCR) model was developed, in which the daily placement
of waste is simulated by a series of completely mixed reactors
(Lowry et al., 2008). Since the LFCR model is a numerical model,
it can simulate transient gas and leachate flow. One drawback of
the LFCR model that also applies to MOCLA is the lack of model
confirmation. Model confirmation for LFCR would entail a con-
trolled and well-monitored landfill in which required information
such as initial concentration of the contaminant, water infiltration,
gas production and operating parameters are available (Lowry
et al., 2008).
The overall objectives of this study were to (1) conduct column
experiments to measure the fate and transport of an OC in a simu-
lated solid waste mixture, (2) compare the results of column
experiments to model predictions using HYDRUS-1D (version
4.13), a contaminant fate and transport model that can be param-
eterized to simulate the laboratory experimental system used in
this research, and (3) determine model input parameters from
independently conducted batch experiments. Experiments were
conducted in which both sorption only and sorption plus biodegra-
dation influenced OC transport.
2. Materials and methods
The batch and column tests are described in this section. Batch
tests were conducted to independently measure the partitioning of
phenol to the simulated waste and the biodegradation rate of phe-
nol by a mixed consortium of anaerobic bacteria. Phenol transport
in a column was measured under the influence of sorption only
and sorption plus biodegradation. To evaluate the rate of phenol
biodegradation in the column more accurately, biodegradation rate
batch tests were conducted with the column contents at the end of
the column test, in which both sorption and biodegradation fate
processes were tested. As described below, the column contents
consisted of a mixture of glass beads and decomposed newsprint
that was designed to allow the study of biodegradation plus
sorption.
2.1. Model contaminant and MSW
Phenol was selected as the model hydrophobic organic contam-
inant (HOC) because it is frequently detected in landfill leachates,
sorbs to refuse, and biodegrades in decomposing refuse (Wang
and Barlaz, 1998). Phenol (99.2% purity, ACROS) stock solutions
were prepared in anaerobic ultrapure water and stored at 4 °C.
Degraded newsprint (DNP) was selected as a representative
MSW component because it is biodegradable and will sorb phenol.
Moreover, paper constitutes as much as 35.5% of landfill discarded
MSW in the US (Staley and Barlaz, 2009). Anaerobically degraded
forms of newsprint were prepared in laboratory reactors using
leachate recycle and neutralization as described (Eleazer et al.,
1997), except that substrates were seeded with anaerobic sewage
sludge instead of well-decomposed refuse. Reactors containing
shredded newsprint were operated over a period of about
9 months to obtain materials that, based on their methane yield,
were significantly decomposed relative to the starting materials.
DNP was previously analyzed for cellulose, hemicellulose, lignin,
lipophilic extractives, crude protein, ash content, organic carbon,
fraction and Brunauer, Emmett, and Teller (BET) surface area. De-
tailed characterization of DNP is presented elsewhere (Wu et al.,
2001).
2.2. Batch experiments
2.2.1. Batch sorption isotherm
A batch isotherm experiment with DNP was conducted at initial
phenol concentrations of 8–25 mg/L and three aqueous solution (1)
23.7 mM phosphate buffer, (2) biochemical methane potential
(BMP) assay medium, and (3) BMP medium inoculated with a
methanogenic consortium of bacteria that was subsequently auto-
claved. Established methods were followed for BMP medium prep-
aration (Wang et al., 1994) and batch isotherm experiments (Wu
et al., 2001). The BMP medium is a phosphate buffered medium
that includes trace minerals and vitamins and is reduced with cys-
teine hydrochloride. The batch isotherm experiment was con-
ducted in triplicate by adding 35 mL of liquid, pre-spiked with
phenol, to 160-mL serum bottles that contained 1.9 g of DNP. After
the serum bottles were capped, a 1-mL liquid sample was immedi-
ately withdrawn and filtered through a 13-mm diameter syringe
filter (0.2-lm PTFE, Millipore Co.) prior to analysis by high perfor-
mance liquid chromatography (HPLC, Dionex Corp.) to measure the
initial aqueous concentrations of phenol. The final aqueous con-
centrations of phenol were measured after 5 days of sorption
equilibration at room temperature.
The partition coefficient (Kp) describing phenol sorption to DNP
was determined from batch isotherm data. Solid phase phenol con-
centrations were calculated by means of a mass balance as shown
in Eq. (1):
ðC 0  C e Þ V
q¼ ð1Þ
m
where q is the equilibrium solid phase concentration of phenol, C0 is
the initial aqueous phenol concentration, Ce is the aqueous phenol
concentration at equilibrium, V is the volume of aqueous solution,
and m is the mass of DNP. The single-solute isotherm data were de-
scribed with a linear partitioning model (Eq. (2)):
q ¼ K pCe ð2Þ
Kp was determined by conducting a linear regression analysis
with the isotherm data.
2.2.2. Batch biodegradation kinetics
After completion of the column test in which both sorption and
biodegradation fate pathways were evaluated, a batch kinetic test
was conducted to determine the biodegradation rate constant, kb,
for phenol. The whole column was placed inside a N2-sparged
glove box to extract solid samples from the top, middle, and bot-
tom column sections under anaerobic conditions. The biodegrada-
tion batch test was conducted in 260-mL serum bottles containing
150 g of wet solids (DNP + glass beads (GBs)), and 90 mL of BMP
medium that was pre-spiked with phenol at a concentration of
27.7 mg/L. Since the solid samples were wet, any dilution of the
phenol concentration was accounted for by measuring the initial
aqueous phenol concentration in each bottle immediately after
adding the phenol-spiked BMP medium. Triplicate bottles were
setup for each 20-cm column section and 1-mL liquid samples
were withdrawn for phenol analysis after 1, 2, 3, 5, 7, 10, 16.6,
24, and 48 h of incubation at 37 °C. Serum bottles containing only
phenol-spiked BMP medium were set up as negative controls. The
first order rate constant kb was estimated by plotting ln(C/C0) as a
function of time (t) and conducting a linear regression analysis. The
C0 value used in the calculation was based on the aqueous phenol
concentration measured after 1 h of incubation to account for phe-
nol loss due to sorption to DNP. Using the Kp value from batch iso-
therm experiments, the calculated sorption equilibrium
concentrations of phenol and those measured after 1 h of incuba-
tion differed by 3–8%.
 J.M. Saquing et al. / Waste Management 32 (2012) 327–334 329

2.3. Column setup and experiments Table 1

Experimental and modeling parameters for phenol sorption column test.

Column experiments were conducted using the configuration Column section

shown in Fig. 1. The glass column was equipped with three sam- 0–20 cm 0–40 cm 0–60 cm
pling ports located 20 cm apart. The flow through the column
Kp phenol–DNP (L/kg) 4.51 4.51 4.51
was upward so the first sampling port was located 20 cm from Kp column (L/kg) 0.21 0.21 0.21
the bottom of the column (Fig. 1A). The glass column was packed Darcy flux, q (cm/day) BMP medium 9.25 9.25 9.25
with a mixture of shredded DNP (1  2 cm) and GBs (3 mm diam- HRT (days) 1.13 2.25 3.38
eter, Fisher), and the masses of DNP and GBs contained in the col- Glass beads mass (g) 1747.6 3495.1 5242.7
DNP mass (g) 87.1 174.2 261.3
umn from the influent end to each of the sampling ports are shown
Weighted particle density (kg/L) 2.49 2.49 2.49
in Table 1. These masses were selected to obtain phenol break- Bulk density (kg/L) 1.05 1.19 1.22
through curves in a manageable time frame. Preliminary model Porosity (h) 0.58 0.52 0.51
runs using HYDRUS-1D were conducted to develop sampling Dispersivity (cm) 0.32 0.36 0.44
schedules for each column section.
DNP and GBs were separately dried at 105 °C for 24 h and same amount of DNP and GB. Prior to the column experiments,
weighed. GBs were thoroughly washed with soap and rinsed with the column environment was made anaerobic by flushing the sys-
tap water and then deionized water prior to drying. To achieve sat- tem with anaerobic BMP medium. To prevent clogging of the HPLC
urated conditions and minimize air in the column, the DNP sam- pump used for column testing, the BMP medium had a reduced
ples were soaked for about 5 days in nine 500-mL mason jars amount of phosphate buffer (6.9 mM) and was filtered through a
containing 23.7 mM phosphate buffer (pH = 7.1) before packing. 0.22 lm PTFE membrane. The modified BMP medium was pre-
The column was packed by simultaneously filling the column pared according to (Wang et al., 1994) except that resazurin (an
with phosphate buffer solution (23.7 mM at pH 7.1) in an upward indicator for O2 presence) was added after reboiling the filtered
mode and packing it with DNP and GBs. To facilitate homogeneity, BMP medium. All column tests were performed under the same
small amounts of wet DNP and GB were placed in layers that were experimental conditions (anaerobic, phosphate buffered solution
compacted with a metal rod. In addition, the column was divided at pH 7.1) in a room maintained at 37 °C. Column experiments
into nine sections, and each section was packed with about the were completed in three consecutive phases: (1) tracer test, (2)

A Screw Thread tube, 14 thread size PRG-201

5 cm

20 cm Reaction Kettle Lid –PRG-188, I.D. 100mm

Flange Clamp –PRG-188, I.D. 100mm

20 cm

Flange Clamp –PRG-188, I.D. 100mm

80 cm
20 cm
Sampling ports, size 1 black butyl rubber stoppers
Length = 4 cm, ID = 15.85mm

Flange Clamp –PRG-188, I.D. 100mm

20 cm

Perforated plug screw thread (Teflon) with Teflon-

coated o-ring –thread size 100 mm
Hole size: 14 mm
10 cm

B Water
Effluent Trap
reservoir
Gas
bag

Glass
column

Sampling ports
Phenol
feedstock
reservoir

Magnetic
stir plate

HPLC pump

Fig. 1. Schematic of (A) column and (B) experimental setup used in column tests.
330 J.M. Saquing et al. / Waste Management 32 (2012) 327–334
sorption column test, and (3) biodegradation–sorption column test
that was done simultaneously with the second tracer test.
First, a bromide tracer test was performed to obtain the param-
eters describing the hydrodynamic column properties, which are
the dispersivity (a) and the porosity (h). BMP medium containing
66 mg/L bromide was continuously fed by an HPLC pump into
the column at a flow rate of 0.48 mL/min until the column effluent
concentration reached the concentration in the influent (6 days).
After the tracer test, the column was flushed with blank BMP med-
ium to prepare the column for the phenol sorption column test,
which was initiated when the aqueous bromide concentrations in
all column sections were below the analytical detection limit
(<0.5 mg/L). The phenol sorption column test was conducted using
the experimental parameters shown in Table 1. BMP medium with
28.1 mg/L of aqueous phenol concentration was continuously fed
into the column at a flow rate of 0.48 mL/min until sorption equi-
librium was attained (8 days). After the sorption test, the column
was flushed with blank BMP medium to desorb phenol completely
in preparation for the biodegradation–sorption test. To initiate bio-
degradation, 6.36 L of anaerobic phenol-enriched consortium was
fed into the column at a flow rate of 3.5 mL/min in two install-
ments using a Masterflex L/S peristaltic pump equipped with PTFE
tubing (Cole-Parmer). The first 3.2 L of anaerobic consortium was
fed into the column but wasted. The second batch of anaerobic
consortium (3.16 L) was fed into the column and recirculated twice
to enable attachment of culture to the solid media in the column
(i.e., DNP and GB). Following inoculation, the biodegradation–
sorption column test was conducted in parallel with the second
tracer test. BMP medium spiked with 27.4 mg/L phenol and
56 mg/L bromide was pumped into the column at a flow rate of
1.45 mL/min. Table 2 provides the experimental parameters used
for the biodegradation–sorption test. The biodegradation–sorption
test was terminated when the breakthrough curve was completed
and a pseudo-steady state condition was observed.
Before each column experiment, the flow rate from the HPLC
pump was checked by diverting the flow to a 40-mL glass vial for
a set time (20–60 min). The amount of liquid in the vial was
weighed and the pump was adjusted if the flow rate was not with-
in 3% of the target rate. Also, the concentrations of phenol and bro-
mide in the feed reservoir were measured before and after each
column experiment to ensure that the mass flux of the studied
chemicals was constant throughout each column experiment.
2.4. Sampling and analytical procedures
Samples for analysis of phenol and bromide were collected at
strategic times during the course of the column experiments from
all three sampling ports. Preliminary modeling was performed
Table 2
Experimental and modeling parameters for phenol biodegradation–sorption column
test.
Column section
0–20 cm 0–40 cm
Biodegradation t1/2 (day) 0.58 0.68
kb (day1) 1.19 1.02
Kp phenol–DNP (L/kg) 4.51 4.51
Kp column (L/kg) 0.21 0.21
Darcy flux, q (cm/day) BMP medium 27.75 27.75
HRT (days) 0.41 0.82
Glass beads mass (g) 1747.6 3495.1
DNP mass (g) 87.1 174.2
Weighted particle density (kg/L) 2.49 2.49
Bulk density (kg/L) 1.05 1.19
Porosity (h) 0.58 0.52
Dispersivity (cm) 4.98 9.46
prior to each column test to develop sampling schedules that
would assure that the entire breakthrough curve would be cap-
tured at each column section. At each sampling time, duplicate
2.5-mL samples were withdrawn by inserting a syringe through a
septum at each sampling port. The first 2.5 mL was wasted and
not filtered. Collected effluent samples for phenol and bromide
analyses were immediately filtered with 13-mm diameter syringe
filters (0.2-lm PTFE, Millipore Co.) and (0.45-lm PSF, Acrodiscs),
respectively. Filtered effluent samples were stored at 4 °C if HPLC
and IC analyses could not be performed immediately after
sampling.
Phenol analysis was performed using a HPLC (Shimadzu Prom-
inence) equipped with a diode array detector set at 210 nm wave-
length, and a Kinetex C18 (100A, 4.6  100 mm) column. The
mobile phases were a mixture of methanol (30%) and 7.4 mM
phosphoric acid (70%). The mobile phase flow rate was 1 mL/min
and the column oven was set at 55 °C.
Bromide analysis was conducted using a Dionex IC (ICS2500)
equipped with an Ion PAC AS19 (4  250 mm) analytical column.
The mobile phases were a mixture of sodium carbonate and so-
dium bicarbonate. The concentrations of bromide were detected
using a conductivity detector and quantified using Dionex AI-450
analytical software.
2.5. Transport models
HYDRUS-1D (version 4.13) is a numerical model that has been
extensively used in soils to simulate one-dimensional variably-sat-
urated water flow and transport of OCs (Simunek et al., 2009;
Casey et al., 2004; De Wilde et al., 2009). HYDRUS-1D was selected
for use in this study because it could be readily adapted to the col-
umn system described above. The transport of OCs is described by
the advection–dispersion (A–D) equation in the liquid phase. The
model is capable of incorporating equilibrium and non-equilibrium
sorption processes as well as first-order degradation reactions
(Simunek et al., 2009). In addition, HYDRUS has an inverse model-
ing feature that relies on the Levenberg–Marquardt nonlinear least
squares optimization algorithm (Simunek et al., 2009) to search for
model input parameters that produce the best model fit of the
experimental data. In this study, sorption and biodegradation
parameters derived from inverse modeling (HYDRUS fit) were
compared to the parameters determined from batch tests.
The porosity and dispersivity that apply for column experiments
were determined from bromide tracer data. The porosity was de-
fined by the time at which 50% of the bromide tracer was recovered,
and the dispersivity was defined by the slope of the S-shaped
bromide breakthrough curve. To determine the porosity and the
dispersivity, the root mean square error between the bromide tracer
data and the output from the A–D model was minimized.
The analytical solution of the one-dimensional (1D) A–D equa-
tion with a step input function of tracer is shown in Eq. (3)
       
C0 Rx  v t vx Rx þ v t
C¼ erfc pffiffiffiffiffiffiffiffiffiffiffi þ exp erfc pffiffiffiffiffiffiffiffiffiffiffi ð3Þ
2 2 Dh Rt Dh 2 Dh Rt
0–60 cm Eq. (3) applies when the initial and boundary conditions are as
0.70 follows (Schnoor, 1996):
1.00
Initial condition: C(x,0) = 0, for x P 0.
4.51
0.21 Boundary condition 1: C(0,t) = C0, for x = 0 and all time t.
27.75 @C
1.11 Boundary condition 2: ¼ 0, for x = 1.
5242.7
@x
261.3 In Eq. (3), C is the effluent concentration (mg/L); C0 is the influ-
2.49 ent concentration (mg/L); x is column length (cm); t = elapsed time
1.22 (days); v is the pore water velocity (cm/day); R is the retardation
0.51
factor (dimensionless) which is equal to 1 for a tracer, and erfc is
8.95
the complementary error function. The pore velocity can be
 J.M. Saquing et al. / Waste Management 32 (2012) 327–334 331

calculated from the known Darcy velocity (q) and the porosity (h)
as shown in Eq. (4):
q
v¼ ð4Þ
h

3. Results and discussion

3.1. Batch experiments

Sorption of phenol to DNP was well described by a linear iso-

therm with a Kp value of 4.5 L/kg (Fig. 2, Table 3). The Kp value of
4.5 L/kg is comparable to the predicted value of 6.1 L/kg that was
obtained from a linear free energy relationship that relates the or-
ganic-carbon-normalized partition coefficient for organic contam-
inant sorption to DNP and the octanol–water partition coefficient
(Wu et al., 2001; Saquing et al., 2010). Because sorption of HOCs
to newsprint is controlled by lignin content, variation in Kp values
is expected (Wu et al., 2001). Fig. 2 also shows that constituents of
BMP medium and biomass did not affect phenol sorption to DNP.
The phenol biodegradation data obtained with the phenol
enrichment culture that was developed during this research are
shown in Fig. 3. Results from the linear regression indicate that
the mean anaerobic biodegradation rate of phenol was
1.0 ± 0.24 d1 (Table 4, ± values represent one standard deviation),
resulting in an average half-life of 16.7 ± 3.1 h. The measured first
order biodegradation rates from all column sections are statisti-
cally similar at the 95% confidence level (Table 4).

3.2. Bromide tracer test

Dispersivity and porosity are required input parameters in con-

taminant transport models based on the A–D equation (Eq. (3)).
These parameters describe the hydrodynamic characteristics of a
particular column matrix and were derived from bromide tracer

Fig. 3. Linearized first-order biodegradation of phenol.

breakthrough curves. In Fig. 4, relative Br concentrations are

shown as a function of time for each of the three column sections.
The bromide tracer breakthrough curves in Fig. 4 were obtained
from two tracer tests that were conducted at Darcy fluxes that dif-
fered by a factor of three.
Parameters derived from the first tracer test are provided in Ta-
Fig. 2. Single-solute phenol isotherm for degraded newsprint. Solid line indicates ble 1. The initial porosity ranged from 0.51 to 0.58 which is within
best fit of the linear partition model and partition coefficient is shown in Table 3. measured porosity values of landfilled MSW (0.12–0.62 (Bendz and
Singh, 1999; Rosqvist and Destouni, 2000; Bou-Zeid and El-Fadel,
2004)). The dispersivity of the column varied from 0.32 to
0.44 cm. Dispersivity for a variety of soils ranges from <0.5 cm
Table 3
for loamy sand to >20 cm for silty clay (Perfect et al., 2002). Very
Sorption parameters for phenol.
few measured values of MSW dispersivity are available. Very high
Column section Partition coefficienta Kp (mg Kg1/mg L1) dispersivity values (100–320 cm) were obtained in an undisturbed
Batch Kp fitb HYDRUS fitc well-degraded waste sample (Rosqvist and Destouni, 2000) and
0–20 cm 4.51 (2.5E-01) 5.53 (2.3E-01) the authors suggested that preferential flow dominated the solute
0–40 cm 4.51 (6.3E-01) 6.03 (4.6E-01) transport.
0–60 cm 4.51 (2.9E-01) 4.01 (2.0E-01) The second tracer test was conducted concurrently with the
a
Values in parentheses represent standard error. sorption–biodegradation test. As shown in Fig. 4, the second tracer
b
Batch Kp value based on batch sorption isotherm experiment. test indicated that dispersion had become more important with a
c
HYDRUS fit Kp value derived from sorption column test. dispersivity about 10 times the initial dispersivity. The addition
332 J.M. Saquing et al. / Waste Management 32 (2012) 327–334

Table 4
Biodegradation parameters for phenol.

Column section Batch test Column test

measured biodegradation ratea,b kb (d1) fitted biodegradation rate kb (d1)
HYDRUS Fitc Lowerd Upperd
0–20 cm 1.19 (0.32) 2.32 (1.6E-02) 2.28 2.35
0–40 cm 1.02e (0.28) 2.23 (7.6E-02) 2.11 2.46
0–60 cm 1.00f (0.24)
a
Average of replicate analyses, based on linearized first order anaerobic biodegradation of phenol illustrated in Fig. 3.
b
Values in parentheses represent one standard deviation.
c
Values in parentheses represent standard error.
d
Values at 95% confidence interval.
e
Average value for 0–40 cm column section includes 20 cm data.
f
Average value for 0–60 cm column section includes 20 and 40 cm data.

positively skewed with a prolonged right-hand tail. The early

appearance of bromide tracer in Fig. 4B compared to Fig. 4A is
indicative of rapid solute transport in preferential flow paths and
the prolonged tails suggest slow water flow in less mobile domains
(Rosqvist and Destouni, 2000).

3.3. Sorption column test

Sorption of HOCs to sorbent organic matter (OM) retards their

transport and limits their availability for biological and chemical
degradation. The retardation effect of phenol sorption to DNP is de-
picted in Fig. 5. HYDRUS simulations with the batch Kp value
(4.5 L kg1) successfully estimated phenol breakthrough to at least
50% of the influent concentration at both the 0–20 and 0–40 cm
column sections. For the 0–60 cm column section, the HYDRUS
prediction for the onset of phenol appearance was approximately
4 days (Fig. 5), and the observed initial appearance of phenol dif-
fered by about 8% (i.e., it occurred 8 h earlier). Table 3 shows that
the Kp values that were obtained with HYDRUS (4.0–6.0 L kg1) are
comparable to the batch Kp value (4.5 L kg1). Thus, the effect of
sorption on the fate and transport of HOC can be simulated using
HYDRUS-1D using the Kp parameter developed from batch iso-
therm experiments.
As shown in Fig. 5, sorption equilibrium was attained but the
effluent equilibrium concentrations of phenol at 0–40 and 0–
60 cm column sections were 6–8% lower than the influent phenol
concentrations. We suspect that some phenol biodegradation oc-
curred about 4.5 days into the column test. Separate batch tests
were conducted with samples from the 0–60 cm column section
to test whether phenol biodegradation occurred in the aqueous
phase, but the results did not show a statistically significant

Fig. 4. Observed and simulated relative concentrations of bromide as a function of

time in (A) first and (B) second tracer tests. Model fits for (A) used porosity and
dispersivity parameters derived for each column section. Model fits for (B) used
porosity from the first tracer test and derived dispersivity for respective column
sections.

and recirculation of phenol enrichment culture at a higher flow

rate (3.5 mL/min) prior to the biodegradation–sorption column
test most likely disturbed the column and changed its hydrody-
namic properties. Since the porosity most likely did not change
considerably, the breakthrough curves obtained in the second tra-
cer test were used to search for the dispersivity parameters only,
while keeping the porosity values obtained from the first tracer
test (Table 2). The shape of the breakthrough curves from the sec-
ond tracer test is consistent with breakthrough curves of soil and
waste samples that have considerable preferential flow paths
(Rosqvist and Destouni, 2000; Perfect et al., 2002). The presence Fig. 5. Relative concentrations of phenol as a function of time in sorption column
of preferential flow paths causes the breakthrough curves to be test with model fits using porosity and dispersivity parameters from first tracer test.
 J.M. Saquing et al. / Waste Management 32 (2012) 327–334
Fig. 6. Relative concentrations of phenol as a function of time in biodegradation
plus sorption column test with model fits using (A) batch Kp and kb (B) batch Kp and
fitted kb. Model fits used porosity from first tracer test and derived dispersivity for
each respective column section. Dotted lines represent one standard deviation of
batch kb.
decrease (p = 0.01) in the phenol concentration after 5 days of incu-
bation at 37 °C. Attached growth in the column may have been
responsible for the degradation of phenol. Overall, the phenol loss
during the sorption column test was surprising because in preli-
minary batch tests, phenol did not degrade using leachate from a
reactor containing mixed MSW. Moreover, phenol biodegradation
using leachate from a reactor containing hardwood was not de-
tected for 41 days and only a 50% reduction in phenol concentra-
tion was observed after 69 days.
3.4. Biodegradation–sorption column test
The combined effects of biodegradation and sorption on phenol
transport were studied in a second column test. Experimentally
determined phenol breakthrough curves at the 0–20, 0–40 and
0–60 cm column sections were compared with HYDRUS model
predictions in Fig. 6A. HYDRUS predictions were obtained with
Kp and kb values that were obtained from batch experiments. The
model predictions obtained with the batch test parameters did
not match the experimentally determined breakthrough curves
well (Fig. 6A). However, HYDRUS simulation using the batch Kp va-
lue and a fitted kb value described well the experimentally deter-
mined breakthrough curves for 0–20 and 0–40 cm column
sections (Fig. 6B). The fitted kb values were approximately twice
those obtained from the batch kinetic tests (Table 4). Breakthrough
curve data obtained at the 0–60 cm column section could not be
333
Fig. 7. Relative concentrations of phenol as a function of time in biodegradation
plus sorption column test with model fits using batch Kp and kb and derived
porosity and dispersivity for each respective column section from the first tracer
test.
described by a HYDRUS fitted kb value and best fit hydrodynamic
parameters from the second tracer test (Table 2). Apparently, the
inoculation of the column with microorganisms at a higher flow
rate (i.e., 7 higher than in the sorption column test) significantly
altered the flow patterns such that phenol and tracer transport be-
came more non-ideal, especially in the 40–60 cm column section.
To the extent that the second tracer test was able to characterize
the change in the hydrodynamic properties of the column,
HYDRUS-1D was able to reasonably simulate the breakthrough
curves obtained from the biodegradation–sorption column test
(Fig. 6). HYDRUS simulations using batch Kp and kb values and
hydrodynamic parameters prior to inoculation (i.e., first tracer test)
poorly described the phenol breakthrough curve data (Fig. 7). The
predominance of preferential flow paths during the biodegrada-
tion–sorption column test resulted in the early appearance of phe-
nol in various column sections. Given the heterogeneity of landfill
discarded MSW, preferential flow paths will likely influence the
fate and transport of contaminants in landfills. Because both MO-
CLA and LFCR assume uniform flow and lack capabilities to simu-
late preferential flow regimes, contaminant fate and transport
predictions using these models are likely to have some error.
4. Conclusions and implications for landfill disposal
Simulating the effects of various fate processes during transport
of OCs is complex. A contaminant transport model (LFCR) that can
quantitatively describe the fate of OCs in landfills has been de-
scribed, but data to confirm the model were lacking. The results
of the column experiments described in this study demonstrated
that HYDRUS-1D can reasonably simulate the fate and transport
of phenol in an anaerobic and fully saturated waste column, in
which biodegradation and sorption are the prevailing fate pro-
cesses. The model description of biodegradation and sorption in
HYDRUS, MOCLA and LFCR are similar but the experimental system
used here was more readily simulated using HYDRUS. Moreover,
the study showed that the incorporation of hydrodynamic param-
eters that describe prevailing flow regimes are essential to improve
predictions of contaminant fate and transport in landfills. The
agreement between model predictions and column data was
imperfect for the sorption plus biodegradation test and the error
almost certainly lies in the difficulty in measuring a biodegradation
rate that is applicable to column conditions. The difference be-
tween the independently measured biodegradation rate (kb) and
the value obtained by inverse modeling in HYDRUS was about a
334 J.M. Saquing et al. / Waste Management 32 (2012) 327–334
factor of two. While this difference could be considered as 100% er-
ror, a rate estimate that is within a factor of two or even five may
be adequate in the context of a landfill given the extended reten-
tion time and the fact that leachate release will be controlled by
the infiltration rate which can be minimized by engineering
controls.
The LFCR model represents the most rigorous description of
landfill contaminant fate and transport to date. An appropriate
next step in work to predict the fate and transport of organic con-
taminants in landfills would be to modify the LFCR model so that it
could account for preferential flow processes and also be evaluated
using laboratory-scale columns.
Acknowledgments
The US Environmental Protection Agency through its National
Homeland Security Research Center funded the research described
here under (EP-C-05-059) to Eastern Research Group. It has been
subject to an administrative review but does not necessarily reflect
the views of the Agency. We would like to thank David Black for
the bromide analysis.
References
Bartelt-Hunt, S.L., Barlaz, M.A., Knappe, D.R.U., Kjeldsen, P., 2006. Fate of chemical
warfare agents and toxic industrial chemicals in landfills. Environ. Sci. Technol.
40, 4219–4225.
Bendz, D., Singh, V.P., 1999. Solute transport under steady and transient conditions
in biodegraded municipal solid waste. Water Resour. Res. 35, 2333–2345.
Bou-Zeid, E., El-Fadel, M., 2004. Parametric sensitivity analysis of leachate transport
simulations at landfills. Waste Manage. 24, 681–689.
Casey, F.X.M., Hakk, H., Simunek, J., Larsen, G.L., 2004. Fate and transport of
testosterone in agricultural soils. Environ. Sci. Technol. 38, 790–798.
California Integrated Waste Management Board, 2010. Health Care Waste at Home.
Retrieved December 2010, from <http://www.calrecycle.ca.gov/ReduceWaste/
HealthCare/PPCP.htm>.
De Wilde, T., Mertens, J., Simunek, J., Sniegowksi, K., Ryckeboer, J., Jaeken, P.,
Springael, D., Spanoghe, P., 2009. Characterizing pesticide sorption and
degradation in microscale biopurification systems using column displacement
experiments. Environ. Pollut. 157, 463–473.
Eklund, B., Anderson, E.P., Walker, B.L., Burrows, D.B., 1998. Characterization of
landfill gas composition at the Fresh Kills municipal solid-waste landfill.
Environ. Sci. Technol. 32, 2233–2237.
Eleazer, W.E., Odle, W.S., Wang, Y.S., Barlaz, M.A., 1997. Biodegradability of
municipal solid waste components in laboratory-scale landfills. Environ. Sci.
Technol. 31, 911–917.
Florida Department of Environmental Protection, 2006. How to dispose of
unwanted medications. Tallahassee, FL.
Foose, G.J., Benson, C.H., Edil, T.B., 2002. Comparison of solute transport in three
composite liners. J. Geotech. Geoenviron. Eng. 128, 391–403.
Jury, W.A., Russo, D., Streile, G., Elabd, H., 1990. Evaluation of volatilization by
organic-chemicals residing below the soil surface. Water Resour. Res. 26, 13–20.
Kjeldsen, P., Barlaz, M.A., Rooker, A.P., Baun, A., Ledin, A., Christensen, T.H., 2002.
Present and long-term composition of MSW landfill leachate: a review. Crit.
Rev. Environ. Sci. Technol. 32, 297–336.
Kjeldsen, P., Christensen, T.H., 2001. A simple model for the distribution and fate of
organic chemicals in a landfill: MOCLA. Waste Manage. Res. 19, 201–216.
Lin, J.S., Hildemann, L.M., 1995. A nonsteady-state analytical model to predict
gaseous emissions of volatile organic-compounds from landfills. J. Hazard.
Mater. 40, 271–295.
Lowry, M.I., Bartelt-Hunt, S.L., Beaulieu, S.M., Barlaz, M.A., 2008. Development of a
coupled reactor model for prediction of organic contaminant fate in landfills.
Environ. Sci. Technol. 42, 7444–7451.
Musson, S.E., Tolaymat, T., Townsend, T.G., 2008. Determination of the fate of
selected pharmaceutical compounds under anaerobic landfill conditions.
Proceedings of Global Waste Management Symposium, Copper Mountain,
Colorado, September 8–10.
Oman, C.B., Junestedt, C., 2008. Chemical characterization of landfill leachates – 400
parameters and compounds. Waste Manage. 28, 1876–1891.
Office of National Drug Control Policy, 2007. Proper Disposal of Prescription Drugs.
Executive Office of the President of the United States, Washington, DC.
Paxeus, N., 2000. Organic compounds in municipal landfill leachates. Water Sci.
Technol. 42, 323–333.
Perfect, E., Sukop, M.C., Haszler, G.R., 2002. Prediction of dispersivity for
undisturbed soil columns from water retention parameters. Soil Sci. Soc. Am.
J. 66, 696–701.
Reinhart, D.R., Pohland, F.G., Stevens, D.K., 1991. Mathematical fate modeling of
hazardous organic pollutants during codisposal with municipal refuse. Hazard.
Waste Hazard. Mater. 8, 85–97.
Rosqvist, H., Destouni, G., 2000. Solute transport through preferential pathways in
municipal solid waste. J. Contam. Hydrol. 46, 39–60.
Saquing, J.M., Mitchell, L.A., Wu, B., Wagner, T.B., Knappe, D.R.U., Barlaz, M.A., 2010.
Factors controlling alkylbenzene and tetrachloroethene desorption from
municipal solid waste components. Environ. Sci. Technol. 44, 1123–1129.
Scheutz, C., Dote, Y., Fredenslund, A.M., Mosbaek, H., Kjeldsen, P., 2007. Attenuation
of fluorocarbons released from foam insulation in landfills. Environ. Sci.
Technol. 41, 7714–7722.
Schnoor, J., 1996. Environmental Modeling: Fate and Transport of Pollutants in
Water, Air, and Soil. J. Wiley, New York.
Schwarzbauer, J., Heim, S., Brinker, S., Littke, R., 2002. Occurrence and alteration of
organic contaminants in seepage and leakage water from a waste deposit
landfill. Water Res. 36, 2275–2287.
Schwarzbauer, J., Heim, S., Krooss, B., Littke, R., 2006. Analysis of undisturbed layers
of a waste deposit landfill – Insights into the transformation and transport of
organic contaminants. Org. Geochem. 37, 2026–2045.
Simunek, J., Sejna, M., Saito, H., Sakai, M., van Genuchten, M., 2009. The HYDRUS-1D
Software Package for Simulating the Movement of Water, Heat and Multiple
Solutes in Variably Saturated Media, Version 4.08. Department of
Environmental Sciences, University of California Riverside, Riverside, California.
Staley, B.F., Xu, F., Cowie, S.J., Barlaz, M.A., Hater, G.R., 2006. Release of trace organic
compounds during the decomposition of municipal solid waste components.
Environ. Sci. Technol. 40, 5984–5991.
Staley, B.F., Barlaz, M.A., 2009. Composition of municipal solid waste in the United
States and implications for carbon sequestration and methane yield. J. Environ.
Eng. ASCE 2009 (135), 901–909.
Wang, Y.S., Barlaz, M.A., 1998. Anaerobic biodegradability of alkylbenzenes and
phenol by landfill derived microorganisms. FEMS Microbiol. Ecol. 25, 405–418.
Wang, Y.S., Byrd, C.S., Barlaz, M.A., 1994. Anaerobic biodegradability of cellulose and
hemicellulose in excavated refuse samples using a biochemical methane
potential assay. J. Ind. Microbiol. 13, 147–153.
Wu, B.Y., Taylor, C.M., Knappe, D.R.U., Nanny, M.A., Barlaz, M.A., 2001. Factors
controlling alkylbenzene sorption to municipal solid waste. Environ. Sci.
Technol. 35, 4569–4576.
US EPA, 2008. Municipal solid waste in the United States, Facts and figures [EPA-
530-F-009-021].