Page 1
ORIGINAL ARTICLE
Transport and mobilization of multiwall carbon nanotubesin quartz sand under varying saturation
Abenezer Mekonen • Prabhakar Sharma •
Fritjof Fagerlund
Received: 26 March 2013 / Accepted: 27 August 2013 / Published online: 19 September 2013
� Springer-Verlag Berlin Heidelberg 2013
Abstract In this study, a series of sand packed columns
were used to investigate the mobility of multiwall carbon
nanotubes (MWCNTs) in unsaturated porous media under
unfavorable conditions for deposition. The flow through
column experiments were designed to assess water content,
flow rate, and grain size effect on the mobility of
MWCNTs. It was found that variation in water content had
no significant effect on retention of MWCNTs until it was
lowered to 16 % effective saturation. Thick water films,
high flow rate, and repulsive forces between MWCNTs and
porous media made MWCNTs highly mobile. Different
porous media grain sizes (D50 = 150–300 lm) were used
in this study. The mobility of MWCNTs slightly decreased
in finer grain sands, which was deemed to be an effect of
increase in surface area and the number of depositional
sites, in combination with low-pore water velocity. How-
ever, physical straining was not observed in selected fine-
grain sands and aspect ratio of MWCNTs had low impact
on mobility. Variations in pore-water velocity were pro-
duced by both changes in water saturation and in flow rate.
At high pore-water velocities, the MWCNTs were gener-
ally mobile. However, for the combination of low-pore
water velocity with either low water saturation or small
grain size, some retention of MWCNTs was observed.
Hence, low velocity in combination with flow through
smaller pores increased MWCNT deposition.
Keywords MWCNTs � Mobility � Unsaturated
porous media � Water content � Flow rate
Introduction
Carbon nanotubes (CNTs) are cylindrical tubes of graphene
sheet made from carbon atoms arranged in hexagonal rings
(Monthioux et al. 2007; Petersen et al. 2011a). Single wall
carbon nanotubes (SWCNTs) and multiwall carbon nano-
tubes (MWCNTs) are the main types of CNTs. They have
structure-dependent chemical, mechanical, electrical,
optical, thermal properties and possess unique size (Mauter
and Elimelech 2008). Due to the unique properties of
CNTs, they are widely used in various fields such as,
energy conversion, energy storage, wastewater treatment,
environmental monitoring, green nano-composite design,
and nano-medicine (Flahaut 2011; Tan et al. 2012). As a
result, the use and production of CNTs are continuously
increasing. This can also increase the entrance of CNTs
into the environment either ‘‘accidentally’’ or ‘‘deliber-
ately’’ due to improper disposal during production, usage,
and transport (Klaine et al. 2008; Auffan et al. 2009; Pet-
ersen et al. 2011a). CNTs in the environment have high
potential to contaminate soil, enter into surface water,
groundwater and food chains of living organisms (Klaine
et al. 2008; Perez et al. 2009). In various fields such as
biomedical, CNT surfaces are modified to increase their
stability in solution (Shen et al. 2009) or to change surface
charges, which can affect their behavior in the environment
(Petersen et al. 2011b). A number of studies have been
conducted to assess the risk of CNTs and other carbon-
based nanomaterials on living organisms (such as E. coli)
and ecosystems (Petersen et al. 2011b; Wang et al. 2012).
Though these studies reveal the toxic effects of CNTs to
Electronic supplementary material The online version of thisarticle (doi:10.1007/s12665-013-2769-1) contains supplementarymaterial, which is available to authorized users.
A. Mekonen � P. Sharma (&) � F. Fagerlund
Department of Earth Sciences, Uppsala University,
Villavagan 16, 75236 Uppsala, Sweden
e-mail: [email protected]
123
Environ Earth Sci (2014) 71:3751–3760
DOI 10.1007/s12665-013-2769-1
Page 2
human cells (Monteiro-Riviere et al. 2005), aquatic
organisms (Kennedy et al. 2008) and reproduction of
earthworms (Scott-Fordsmand et al. 2008), there is cur-
rently no general agreement about toxicity of CNTs
towards living organisms due to variation of CNTs in
shape, size, and composition (Flahaut 2011). However, the
capacity of CNTs to adsorb divalent metals (Rao et al.
2007) and organic chemicals (Pan and Xing 2008) shows
their potential for environmental applications such as waste
water treatment.
To assess the impact of CNTs on ecosystems and human
health, understanding their toxicity and transport behavior
in soil is crucial (Cullen et al. 2010). Several column
experiments have been conducted to investigate the
mobility of CNTs in saturated porous media. According to
those recent studies, surface properties of CNTs (Wang
et al. 2008; Jaisi et al. 2008; Liu et al. 2009; Tian et al.
2011, Tian et al. 2012b) and porous media (Jaisi and
Elimelech 2009; Mattison et al. 2011; Tian et al. 2012a)
control their mobility and deposition. In addition, physical
straining has also been suggested as a mechanism which
can govern the mobility and deposition of CNTs in satu-
rated porous media (Jaisi et al. 2008; Jaisi and Elimelech
2009). However, physical straining was not reported in
several other studies on CNTs mobility (Liu et al. 2009;
Tian et al. 2011, Tian et al. 2012a; Mattison et al. 2011).
Even though a number of studies have been conducted, the
different factors which govern the transport and fate of
CNTs in saturated porous media are not yet fully under-
stood and require further investigation (Jaisi and Elimelech
2009).
Understanding the mobility and fate of CNTs in unsat-
urated soil is crucial to determine the risk they impose on
ecosystems, and improve their environmental application.
Although there are a number of studies focused on the fate
and transport of relatively spherical particles in the unsat-
urated zone (e.g., Shang et al. 2008; Sharma et al. 2008a;
Torkzaban et al. 2008; Tiraferri et al. 2011), much fewer
studies have been made on the transport of cylindrical
particles (such as CNTs) in unsaturated zone (Tian et al.
2012b). It has been suggested that studies on nanomaterials
can be based on existing knowledge of the behavior and
toxicology of colloids (Klaine et al. 2008). Although col-
loids (i.e., bigger size and spherical shape) differ from
CNTs, knowledge of their transport and fate can be used as
a reference to study fate and mobility of CNTs in unsatu-
rated porous media. Studies on colloid transport in unsat-
urated porous media have found that deposition/attachment
to solid–water interfaces (SWI) (Schijven and Hassani-
zadeh 2000), air–water interfaces (AWI) (Wan and Wilson
1994a; Sharma et al. 2008a; b), air–water–solid interfaces
(AWSI) (Gao et al. 2008), physical straining (Bradford
et al. 2007; Torkzaban et al. 2008) and film straining (Wan
and Tokunaga 1997) are the possible retention mechanisms
for colloids in unsaturated porous media. These studies
have shown that the mobility and fate of colloids in
unsaturated porous media can be complicated by additional
factors and are different from saturated porous media due
to the existence of an air phase in the porous medium.
However, specific studies on MWCNTs in unsaturated
porous media are lacking. Therefore, the aim of this study
was to investigate how different physical and chemical
factors affect the mobility of cylindrical nanoparticles (e.g.,
MWCNTs) in unsaturated porous media. One-dimensional
column experiments were designed to specifically assess
the impact of moisture content variation, porous media
grain size and flow rate on the mobility and retention of
MWCNTs. In addition, aspect ratio and functionalization
effect on the mobility of MWCNTs were investigated.
Experiments were conducted under unfavorable conditions
for deposition of MWCNTs and using clean quartz sand.
Materials and methods
Porous media
Quartz sands used in this study were obtained from Sibelco
Nordic, Baskarp, Sweden and sieved into three size dis-
tributions: fine (125–177 lm), medium (177–250 lm), and
coarse (250–350 lm) to achieve uniform grain size distri-
bution based on their supplied particle size analysis. The
quartz sand was washed sequentially by hydrochloric acid
(0.1 M) and hydrogen peroxide (5 %) to remove any
impurities from the surfaces of the sand (Mattison et al.
2011). The sand was rinsed at least five times with
deionized water to obtain neutral pH and remove any
remaining impurities. Then, the sand was oven dried
(105 �C) for 24 h and stored. The bulk densities of the
sieved sands were in the range of 1.48–1.58 g/cm3, and
porosities were in the range of 0.40–0.44 (see Table 1).
Soil moisture characteristics of packed sands were deter-
mined using the capillary fringe method, and hydraulic
conductivity was determined by the falling head method
(Dane and Hopmans 2002; Sharma et al. 2008c). The zeta
potential of the sand was -53.4 mV at pH 7 (SurPASS
Electrokinetic Analyzer, Anton Paar). The surface area of
the sand was calculated by assuming an average spherical
diameter of packed sand grains in the column as shown in
Table 1.
Multiwall carbon nanotubes
The MWCNTs, used in this study, were obtained from
Cheap Tubes Inc. The diameter and length of MWCNT
3752 Environ Earth Sci (2014) 71:3751–3760
123
Page 3
used in most of these experiments were 20–30 nm and
0.5–2 lm, respectively. MWCNTs of diameter 30–50 nm
and length 10–20 lm were also used for comparison. To
increase the hydrophilicity and stability of MWCNTs in
aqueous solution, MWCNTs were functionalized using a
3:1 (v/v) ratio of sulfuric acids (97 % purity) and nitric
acids (70 % purity) (Mattison et al. 2011). The MWCNTs
were then filtered and washed from the acid mixture.
The MWCNT solution was prepared by adding 10 mg
of MWCNTs in 1 liter of 0.1-mM background solution.
The background solution was prepared with pH 7 (adjusted
using 0.1-M HCl and 0.1-M NaOH solution) and ionic
strength of 0.1 mM (using 0.1 mM NaCl) to produce
unfavorable conditions for the deposition of MWCNTs. A
mono-dispersed and stable MWCNT solution was prepared
for the column experiments using ultra-probe sonication
(ultrasonic homogenizer, Biologics Inc. Model 3000) for
30 min of sonication (at 40 % of 300 W power), which
provided stable solution for more than double the time
period of our experiments. The zeta potential of the func-
tionalized and less functionalized MWCNTs was -45.42
and -47.36 mV, respectively, at the chosen solution
chemistry (Zeta-sizer, Malvern Instruments). Tracer
experiments were conducted using a conservative tracer
(Brilliant blue FCF, Dr. Ehrenstorfer GmbH).
Experimental setup
Quartz sands were packed in a custom made plexiglass
column of 5-cm diameter and 10-cm length. A sprinkler
made of 12 needles (22 gauges, 0.16 mm inner diameter)
and a water pump (IPC8, Ismatic, Peristaltic pump) were
used to supply inflow solutions from the top of the unsat-
urated column (similar to Sharma et al. 2008b). At the
outlet of the column, a 25-lm membrane filter of 55-cm
bubbling pressure (Nylon net filter, Merck KGaA,
Darmstadt, Germany) was used to support the sand and
create an appropriate suction (Fig. 1).
Before the MWCNTs experiments, tracer experiments
were conducted to determine the hydraulic conditions in
the column at different water contents, flow rates, and sand
sizes. The effluent samples from the column were collected
by a fraction collector (CF-2, Spectrum Labs), and the
absorbance of the samples was measured by a
spectrophotometer (DR 5000, Hach Lange AB) at a
wavelength of 400 nm for MWCNTs and 300 nm for the
tracer.
A series of column experiments were conducted under
unfavorable condition (pH = 7 and 0.1-mM ionic strength)
for deposition. Prior to packing the sand, suction was
created at chosen height to achieve the desired water sat-
uration in the column. Following this, the packed sand was
flushed with 150 ml (3–5 pore volume) of background
solution before MWCNT solution was injected into the
column. Then the sands were flushed with 150 ml of
MWCNT solution (phase I), background solution (phase
II), and deionized water (phase III), respectively, at a
chosen flow rate.
All the different experiments were repeated two times.
Eight column experiments (Experiments 1–4) were con-
ducted at four different saturation conditions (16–100 %)
using coarse sand (D50 = 300 lm) and at the same flow
rate (2.5 ml min-1) to assess the effect of porous media
water content on mobility of MWCNTs (Table 2). In
addition, flow rate effects were investigated by testing three
different flow rates (Experiments 3, 5, and 6) at 25-cm
suction (31 ± 5 % effective saturation) and using coarse
sand (D50 = 300 lm). To compare the mobility of
Fig. 1 Experimental setup of column studies
Table 1 Hydraulic conductivity, porosity, and surface area of sands used in the column experiments
Sand average diameter (lm) Porosity (-) Bulk density (g/cm3) Hydraulic conductivity (cm/sec) Total surface area (cm2)
300 (Coarse) 0.4 1.58 0.012 23,560
211 (Medium) 0.44 1.48 0.0077 31,265
150 (Fine) 0.44 1.48 0.0056 43,980
Surface area of sand calculated by assuming all sands are the same in size and spherical in shape
Environ Earth Sci (2014) 71:3751–3760 3753
123
Page 4
MWCNTs in different sand sizes, six column experiments
using three different grain size distributions (Experiments
2, 7, and 8) were conducted at constant water content
(70 ± 2 %) and using the same flow rate (2.5 ml min-1).
Finally, low moisture content (17 ± 1 %), high flow rate
(2.5 ml min-1), and coarse sand (D50 = 300 lm) were
chosen to compare MWCNTs of different aspect ratios
(Experiments 4 and 9) and functionalization impact on
mobility of MWCNTs (Experiments 4 and 10) (see Table 2
for a summary of experimental conditions).
Water flow modeling
HYDRUS-1D was used to simulate one-dimensional ver-
tical water flow through unsaturated porous media (Simu-
nek et al. 2008; Elmi et al. 2012). A constant head upper
boundary condition and a suction-controlled lower
boundary condition were applied for modeling water flow
across the column. The simulations were done for all flow
rates, sand sizes, and suction heads used in the column
experiments. Water flow parameters used for simulations
were determined using the RETC-code (Van Genuchten
et al. 1991) and the simulation ran until steady-state flows
were reached.
Both experimental and simulation results showed that
the water content of the porous media decreased as the
applied suction increased (Fig. 2a). For all experimental
conditions, the top of the column was more saturated than
the bottom except for the lowest flow rate at the top with
10-cm suction at the bottom of the column (Pfletschinger
et al. 2012). The average water content determined by
HYDRUS-1D was different from measured values except
for experiments conducted at no suction (Table 2). Water
flow simulations in finer sands at a given suction and flow
rate showed that the moisture distribution in both sands
was approximately similar. At the top of the column, the
finer sand (D50 = 150 lm) was more saturated, whereas at
the bottom it was less saturated than the medium sand
(D50 = 211 lm) (as shown in Fig. 2b).
DLVO energy calculation
Derjaguin, Landau, Verwey, and Overbeek (DLVO) theory
was used to calculate interaction energy between
MWCNTs and porous media. Originally, DLVO theory
was established to calculate the interaction energies of
spherical particles. Since the majority of nanoparticles
(NPs) are non-spherical (e.g., MWCNTs are cylindrical),
change in shape can affect the total interaction force
between nanoparticles and porous media (Phenrat et al.
2010). To calculate the interaction energy between
MWCNTs and porous media, either length or diameter has
been used as an effective size in the recent studies (Liu
et al. 2009; Tian et al. 2011, 2012a, b). In this study, both
length and diameter of MWCNTs were used as effective
sizes to calculate a range for the interaction energy. In this
calculation, the zeta potential of sand was -53.4 mV, and
the Hamaker constant between sand and MWCNT was
9.80 9 10-21 and the zeta potential of MWCNTs were
listed in Table 2.
The total interaction energy calculation showed that the
repulsive interaction between MWCNTs and porous media
exists when either diameter or length of MWCNTs was
used, but the primary energy barriers were higher when
length was used in the calculation and relatively low when
diameter was used (Fig. 3). This indicates that the experi-
mental conditions were unfavorable for deposition and that
deposition in primary energy minima on the sand surfaces
was not significant due to a very high primary energy
barrier. A secondary energy minimum was not observed
between the MWCNTs and porous media for both effective
sizes (length and diameter) within 60-nm separation
Table 2 Experimental conditions selected in this study
Exp Average sand
size (lm)
Suction
(-cm of H2O)
Flow rate
(mL/min)
Pore-water
velocity
(cm/min)
Water content (-) Saturation
(%)
MWCNT
n-potential
(mV)
Water-film
thicknessa
(lm)Measured Modeled
1 300 0 2.5 0.32 0.42 0.45 100 -45.42 –
2 300 10 2.5 0.45 0.284 0.43 70.06 -45.42 20.35
3 300 25 2.5 0.88 0.145 0.283 35.91 -45.42 11.217
4 300 40 2.5 1.7 0.075 0.26 18.59 -45.42 5.209
5 300 25 1.5 0.65 0.117 0.256 28.98 -45.42 8.17
6 300 25 0.5 0.23 0.109 0.21 27.72 -45.42 7.964
7 211 29 2.5 0.43 0.295 0.21 66.91 -45.42 20.34
8 150 39 2.5 0.42 0.287 0.21 68.82 -45.42 20.34
9 300 40 2.5 2.1 0.061 0.26 15.12 -47.36 5.209
10 300 40 2.5 1.88 0.069 0.26 17.01 -45.42 5.209
a Calculated by assuming sand is spherical in shape and covered uniformly to their surface area
3754 Environ Earth Sci (2014) 71:3751–3760
123
Page 5
distance, indicating that deposition at secondary minima
was also not feasible. (See supporting material for DLVO
theory and interaction energy calculation).
Results and discussion
Effect of porous media water content on mobility
of MWCNTs
To assess the impact of water content on mobility of
MWCNTs, four column experiments were conducted
(Experiments 1–4). For all the different water contents,
MWCNTs were detected in the effluent after less than one
pore volume of flow, and increased quickly to make a
plateau (phase I, in Fig. 4). For all moisture conditions, the
shape of the breakthrough curves for MWCNTs was sim-
ilar, and their maximum relative concentrations (C/C0)
were close to 0.94 ± 0.02. After steady-state concentration
had been reached, the influent was switched to background
solution, as a result, C/C0 decreased to a minimum value
for all moisture conditions (phase II). Finally, to flush out
the retained (5–8 %) MWCNTs, the influent was switched
to deionized water (phase III). The fact that no significant
changes were observed in the effluent concentrations dur-
ing this phase indicates that a decrease in ionic strength had
no significant effect on retained MWCNTs in the column
(Fig. 4).
These findings indicate that MWCNTs were highly
mobile ([92 %) in unsaturated porous media, and the
presence of an air phase in porous media had minimal
impact on their transport even at the lowest water satura-
tion (16 % of saturation) tested in this study. The high
mobility of MWCNTs in unsaturated porous media can be
explained by repulsive forces between MWCNTs and
porous media and is supported by DLVO calculations
showing that surface deposition of MWCNTs was unlikely
at secondary energy minima.
However, water-film straining of MWCNTs in unsatu-
rated porous media may cause retention of MWCNTs.
According to film straining theory of colloids (Wan and
Tokunaga 1997), straining of colloids in unsaturated por-
ous media occurs only if the moisture content is lower than
a ‘‘critical saturation’’ at which the linkage between pen-
dular rings is broken, or when particle size is bigger than
the water-film thickness. In this study, all column experi-
ments were conducted in a water saturation range between
16 and 100 % (see Table 2). Although the water content
Fig. 2 HYDRUS-1D
simulation of water content
distribution in column depth for
different suctions and flow rates
in a coarse sand (300 lm) at
different suction and flow rates
b course sand at 10-cm suction,
medium sand at 29-cm suction,
and fine sand at 39-cm suction
with flow rate of 2.5 ml min-1
and 70 % average water content
Fig. 3 MWCNTs and porous
media interaction energy as a
function of separation distance
at the solution chemistry used in
this experiment: a fully
functionalized and b less
functionalized
Environ Earth Sci (2014) 71:3751–3760 3755
123
Page 6
varies with depth over the column (Fig. 2), we used aver-
age water content for comparison. For all experimental
conditions, the calculated average water-film thickness
around a single sand grain was thicker than both diameter
and length of the MWCNTs except Experiment 9 (see
Table 2), as discussed in ‘‘Effect of MWCNTs size of their
mobility’’ section. The water-film thickness over the sand
surface was calculated by assuming spherical sand grains
uniformly covered with pore water. With this water-film
thickness, effects of film straining on the mobility of
MWCNTs were unlikely to be observed. The results of this
study indicate that variation of the water saturation in the
range of 16–100 % had negligible effect on retention of
MWCNTs. This is in accordance with preliminary findings
by Tian et al. (2011), indicating that retention of SWCNTs
may occur only at very low (\10 %) water saturation due
to water-film straining.
Corapcioglu and Choi (1996) reported high affinity of
hydrophilic particles toward AWIs, but Tian et al. (2011),
on the other hand, observed repulsive forces between
anionic surfactant-coated SWCNTs and AWI. In this study,
deposition of MWCNTs on AWI was not significant,
because observed retention of functionalized MWCNTs
was very low. Particle deposition/removal by moving
liquid–gas interfaces are more prominent in drying/wetting
events specially when the interface can move over initially
dried surfaces of deposited particles (Sharma et al. 2008b;
Aramrak et al. 2013), but MWCNT-contained suspensions
were replaced by particle-free solution without further
changing the total wet surface area in this study. To con-
firm this finding for MWCNTs, additional studies at much
lower moisture content (\16 %) are needed to further
investigate the effect of pore straining (Jaisi and Elimelech,
2009; Jaisi et al. 2008), AWSI attachment (Crist et al.
2005; Gao et al. 2008), and air–water interface attachment
(Chen et al. 2008).
In addition, the transport of colloids in unsaturated
(Sharma et al. 2008a; Wan and Tokunaga 1997) and
MWCNTs in saturated porous media (Liu et al. 2009) can
depend on water velocity. Since Experiments 1–4 were
conducted at the same flow rate (2.5 mL/min) but at dif-
ferent water saturations, the pore-water velocity increased
with decreasing saturation, as shown in Table 2. Higher
pore-water velocity should support the mobility of
MWCNTs. However, at the same time, lower water satu-
ration means that the water flow takes place through a
subset of smaller sized pores, since water resides in the
smallest pores in a partially saturated medium. In smaller
pores, the particle mobility may be reduced and hence, this
is a competing effect to the increased velocity at lower
water saturation. The results from Experiments 1–4 show
that the velocities are large enough to keep MWCNTs
mobile. Similarly, Liu et al. (2009) found that the mobility
of MWCNTs in saturated porous media was affected by the
pore-water velocity, and that MWCNTs were highly mobile
at pore-water velocities [4 m/day. This is consistent with
our experimental observations and also in agreement with
other colloid literature. The following experiments further
investigate the effect of velocity and pore size by varying,
first, the flow rate and then the grain size.
Effect of flow rate on mobility of MWCNTs
Three column experiments (Experiments 3, 5, and 6) were
conducted to investigate the effect of flow rate on mobility
and retention of MWCNTs in unsaturated porous media
(Fig. 5). After MWCNTs were injected into the column,
they were detected in the effluent after less than one pore
volume, which suggests that some dispersion occurred
during their movement through the column. For all flow
rates, the relative concentration in the effluent increased
quickly and reached a steady-state/maximum value (phase
I), then after the influent switched to background solution,
they decreased to a minimum value (phase II). Deionized
water was used to remobilize retained MWCNTs, but no
changes were observed in the effluent (phase III). At high
flow rate (2.5 mL min-1), the mobility of MWCNTs was
high, and the maximum/steady-state relative concentration
in the effluent was close to 0.95. For low flow rates, 1.5 and
0.5 mL min-1, retention of MWCNTs increased as flow
rate decreased, and the effluent maximum relative con-
centration decreased to 0.88 and 0.75, respectively (Fig. 5).
This shows that the mobility of MWCNTs was a function
of flow rate.
Fig. 4 MWCNTs breakthrough curves in coarse sand (300 lm) at pH
7 and ionic strength 0.1 mM for effective saturation of a 16 %,
b 36 %, c 70 %, d 100 %
3756 Environ Earth Sci (2014) 71:3751–3760
123
Page 7
In this study, the variations in flow rate also resulted in
moderate changes in water content in the column, from
0.145 at 2.5 mL/min to 0.109 at 0.5 mL/min. However, the
pore-water velocity still deceased with decreasing flow rate
(as shown in Table 2). A possible explanation to the higher
retention observed for lower flow rates (e.g., Experiment 6)
in Fig. 5 is a possible combination of a low-pore water
velocity (0.23 cm/min) with a low water saturation, where
the water flow is restricted to a subset of smaller pores. It
can be noted in Experiment 1 with low-pore water velocity
(0.32 cm/min) that the medium was fully saturated and the
water flow had full access of large pores as well.
Another possible explanation to the retention and low
mobility of MWCNTs can be slow exchange between
immobile and mobile zones of the pore water occurring at
the low flow rates. Findings of Gao et al. (2006) suggest the
exchange between the zone of immobile and mobile water
under steady water flow is a mechanism that governs
mobility of colloids at low flow rates. In this study, high
retention of MWCNTs was observed at low flow rate. At
high flow rates, higher mobility of MWCNTs would be
favored by higher hydrodynamic forces increasing the drag
force on MWCNTs.
Grain size effect on mobility of MWCNTs
Three column experiments were conducted (Experiments
2, 7, and 8) to investigate the impact of sand grain size on
mobility and retention of MWCNTs in unsaturated porous
media (Fig. 6). Breakthrough curves of MWCNTs with
three types of porous media follow a similar trend on both
sides of the limb. For all three cases (D50 = 300, 211,
150 lm), MWCNTs were detected in the effluent after less
than one pore volume. In the coarse sand (50 = 300 lm),
MWCNTs were highly mobile and the maximum relative
concentration was close to 1 (*0.95), whereas for the
medium (D50 = 210 lm) and fine (D50 = 150 lm) sands,
the maximum relative concentration of MWCNTs
decreased to 0.90 and 0.86, respectively (Fig. 6).
According to Mattison et al. (2011) for water saturated
conditions, an increase in surface area and number of
depositional sites increase the retardation of MWCNTs in
fine sands. Also in the unsaturated experiments conducted
here, the surface area increased as the sand sizes were
decreased (see Table 1), which in turn increased the
number of depositional sites and adsorption capacity of the
media (Canales et al. 2013). In addition, even though all
column experiments were conducted at the same flow rate
(2.5 mL/min), pore-water velocities in the medium and fine
sands (0.42 cm/min) were slightly lower than that in coarse
sand (0.44 cm/min) because of a slight difference in
porosity. The experimental results indicated that MWCNTs
were mobile in both sand types. The difference between the
breakthrough curves indicating somewhat more deposition
of MWCNTs in the finer sands (Fig. 6) was deemed to
occur mainly due to differences in sand surface area and
number of depositional sites although a slight effect of the
lower pore-water velocity cannot be excluded.
In previous studies, physical straining (grain-to-grain
straining) has been suggested as a factor that affects
deposition and transport of CNTs (Jaisi and Elimelech
2009; Jaisi et al. 2008). Findings by Bradford et al. (2007)
indicate that colloid physical trapping increases with
increasing colloid size to grain size (dp/dg) ratio and sug-
gest that straining could occur at even low dp/dg ratio
(0.003). In another study using SWCNTs, adp/dg ratio of
0.0008 was reported as the effective diameter ratio for
straining (Jaisi et al. 2008). In this study, the calculated
MWCNT length to sand-grain-size ratio was much greater
than the critical limit, but MWCNT diameter to sand-grain-
size ratio was less than this critical limit (see Table 3).
Fig. 5 MWCNTs breakthrough curves in coarse sand (300 lm) at pH
7 and ionic strength 0.1 mM for different flow rate. Data point shows
the average value from replicate experiments
Fig. 6 Breakthrough curves comparison of MWCNTs (diame-
ter = 20–30 nm and length = 0.5–2 lm) in coarse (D50 =
300 lm), medium (D50 = 211 lm) and fine (D50 = 150 lm) sand
at pH 7, ionic strength 0.1 mM, effective saturation 70 %, and flow
rate 2.5 ml min-1. Data point shows the average value from replicate
experiments
Environ Earth Sci (2014) 71:3751–3760 3757
123
Page 8
According to Auset and Keller (2006), straining effects
could be reduced if cylindrical nanoparticles can be ori-
ented parallel to the stream lines in the porous media. The
results of this study may not indicate straining because
retention of MWCNTs was only observed to a relatively
small degree with decreasing sand size. For the case of this
study, MWCNTs may be retained in both sand types if they
were oriented perpendicular to stream channel. However,
the high flow rate and also unfavorable conditions for
deposition have likely reduced the straining effect. The
results in this study suggest that a combination of low-pore
water velocity and small pore sizes available for flow
(Experiments 6 and 8) increases retention of MWCNTs.
This is in agreement with previous studies suggesting that a
combined effect of water content, DLVO, and hydrody-
namic forces influences the straining of colloids (Torkz-
aban et al. 2008), which should be further investigated for
CNTs. However, we performed two separate experiments
with different set of MWCNTs size to confirm the straining
effect of retention and two types of MWCNTs for the effect
of their interaction energies (DLVO energy) in the fol-
lowing sections.
Effect of MWCNTs size on their mobility
To further investigate the possible effects of straining on
retention, we also investigated the effect of MWCNT
aspect ratio on mobility; two different types of MWCNTs
were used in this case, and two column experiments were
conducted all together (Experiments 4 and 9). The calcu-
lated water-film thickness was approximately 5.2 lm in the
selected experiments (Table 2). Although the diameter of
MWCNTs was much smaller for both cases, but the aver-
age length of MWCNTs in Experiment 9 was three times
higher than the water-film thickness. Breakthrough curves
of both MWCNTs are similar in shape, and the maximum
relative concentration (C/C0) was varied only by 4 % from
each other (Fig. 7). The results showed that the shorter
MWCNTs were more mobile in porous media under the
experimental conditions of this study, as possible cause of
physical straining. The two types of MWCNTs only differ
by their dp/dg ratios (MWCNTs length/diameter to sand
size ratio) (Table 3). According to previous studies, aspect
ratio of CNTs and collector properties can cause physical
straining (Jaisi et al. 2008; Jaisi and Elimelech 2009).
However, the observed result shows minimal sign of
physical straining. It is anticipated that the difference in the
mobility of MWCNTs due to aspect ratio occurs if an
experiment can be conducted at very low moisture condi-
tions and/or in very fine porous media.
Effect of MWCNT functionalization on mobility
To investigate the effect of functionalization of MWCNTs
on their mobility, two experiments were conducted
(Experiments 4 and 10) using both fully functionalized and
less functionalized (as received from company) MWCNTs
of 20–30 nm diameter and 0.5–2 lm length. The maximum
relative concentration (C/C0) of fully functionalized
MWCNTs was approximately 0.95, while that of less
functionalized MWCNTs was only 0.65 (Fig. 8). The less
functionalized MWCNTs were detected in the effluent later
than the fully functionalized MWCNTs and their break-
through curve is also distorted. This indicates that the fully
functionalized MWCNTs were highly mobile, whereas the
less functionalized MWCNTs were less mobile and
retained in the column.
Since these studies were conducted under the same
conditions and their DLVO profile did not show significant
difference (Fig. 3), the observed differences were attrib-
uted to the functionalization and particularly the addition of
carboxyl (-COOH) groups to the surfaces of the MWCNTs
Table 3 Ratio of MWCNTs
length and diameter with
different types of porous media
Average diameter and length
used to calculate MWCNTs to
sand size ratio
Sand types MWCNT (D = 25 nm and L = 1.25 lm) MWCNT (D = 40 nm and L = 15 lm)
Diameter/sand size Length/sand size Diameter/sand size Length/sand size
Coarse (300 lm) 8.3 9 10-5 41.6 9 10-4 1.3 9 10-4 0.05
Medium (211 lm) 1.2 9 10-4 59.2 9 10-4 1.9 9 10-4 0.07
Fine (150 lm) 1.6 9 10-4 83.3 9 10-4 2.6 9 10-4 0.1
Fig. 7 Different size of MWCNTs in coarse sand (D50: 300 lm), pH
7, ionic strength 0.1 mM, effective saturation 16 %, and flow rate
2.5 ml min-1. Data point shows the average value from replicate
experiments
3758 Environ Earth Sci (2014) 71:3751–3760
123
Page 9
increasing their hydrophilicity. It has been found that the
addition of surfactants to NP’s surfaces stabilizes their
suspension and reduces aggregation (Phenrat et al. 2010).
According to other studies, aggregation, due to the absence
of energy barriers between the particles, increases the size
and deposition rate of NP aggregates and contributes to the
occurrence of physical straining (Crist et al. 2005; Jaisi
et al. 2008; Jaisi and Elimelech 2009; Phenrat et al. 2010).
Therefore, aggregation is likely to have increased the
deposition and retention of the less functionalized (more
hydrophobic) MWCNTs in this study. According to the
previous studies, retention of hydrophobic colloids in
porous media is higher than that of hydrophilic colloids,
and solid–water and AWSI are suggested as the main site
of desposition (Corapcioglu and Choi 1996; Crist et al.
2005; Wan and Wilson 1994b). In addition, the retention of
colloids increases with surface hydrophobicity (Wan and
Wilson 1994a) which agrees with our experimental
observations, since less functionalized MWCNTs are more
hydrophobic compared to (fully) functionalized MWCNTs.
Conclusions
This study demonstrates that the functionalized MWCNTs
were highly mobile in unsaturated porous media under the
investigated experimental conditions, and the variation in
water saturation had negligible effect on their mobility and
retention for the range of investigated water saturation
(16–100 %). The high mobility of MWCNTs was favored
by thick water-film coating over the sand grains, electro-
chemically unfavorable conditions for deposition, and high
flow rate under which these experiments were conducted.
The results also suggest that the mobility and retention of
MWCNTs can vary with pore-water velocity. Variations in
pore-water velocity were produced by both variation of the
flow rate and variation of the water saturation at a given
flow rate. At high pore-water velocities, the MWCNTs
were highly mobile, and their retention was very low,
whereas at low velocities in combination with either low
water saturation or small grain size, the retention of
MWCNTs was higher. It is possible that MWCNTs
deposited in immobile pore-water zones, and slow
exchange between mobile and immobile zones makes
MWCNTs less mobile at slow flow rates. This study also
showed that MWCNTs were mobile in both the coarse and
fine sands used in this study. However, an increase in
surface area and the number of depositional sites made
MWCNTs less mobile in finer sands. Effects of physical
straining were not observed in fine sands size used in this
study and MWCNTs were highly mobile in unsaturated
porous media regardless of their aspect ratio. This study is
aimed at improving understanding of the transport mech-
anisms for MWCNTs in the vadose zone after their release
into the environment. The study presents an evidence that,
functionalized MWCNTs can be fully mobile in the
unsaturated zone, and thereby can be transported with
infiltrating water to the saturated zone contaminating soil
and groundwater resources.
References
Aramrak S, Flury M, Harsh JB, Zollars RL, Davis HP (2013) Does
colloid shape affect detachment of colloids by a moving air-
water interfaces? Langmuir 29:5770–5780
Auffan M, Rose J, Bottero JY, Lowry GV, Jolivet JP, Wiesner MR
(2009) Towards a definition of inorganic nanoparticles from an
environmental, health and safety perspective. Nature nanotech-
nol 4:634–641
Auset M and Keller AA (2006) Pore-scale visualization of colloid
straining and filtration in saturated porous media using micro-
models. Water Resour Res 42: W12S02
Bradford SA, Torkzaban S, Walker SL (2007) Coupling of physical
and chemical mechanisms of colloid straining in saturated
porous media. Water Res 41:3012–3024
Canales RM, Guan H, Bestland E, Hutson J, Simmons CT (2013)
Particle-size effects on dissolved arsenic adsorption to an
Australian laterite. Environ Earth Sci 68:2301–2312
Chen L, Sabatini DA, Kibbey TC (2008) Role of the air-water
interface in the retention of TiO2 nanoparticles in porous media
during primary drainage. Environ Sci Technol 42:1916–1921
Corapcioglu MY, Choi H (1996) Modeling colloid transport in
unsaturated porous media and validation with laboratory column
data. Water Resour Res 32:3437–3449
Crist JT, Zevi Y, McCarthy JF, Throop JA, Steenhuis TS (2005)
Transport and retention mechanisms of colloids in partially
saturated porous media. Vadose Zone J 4:184–195
Cullen E, O’Carroll DM, Yanful EK, Sleep B (2010) Simulation of
the subsurface mobility of carbon nanoparticles at the field scale.
Adv Water Resour 33:361–371
Dane JH, Hopmans JW (2002) Hanging water column. In: Dane JH,
Topp GC (eds) Methods of Soil Analysis, Part 4 Physical
Methods. Soil Science Society of America, Madison,
pp 680–683
Fig. 8 Fullsy functionalized and less functionalized MWCNTs
breakthrough curves comparison in coarse sand, 16 % effective
saturation, and 2.5 ml min-1 flow rate. Data point shows the average
value from replicate experiments
Environ Earth Sci (2014) 71:3751–3760 3759
123
Page 10
Elmi A, Abou Nohra JS, Madramootoo CA, Hendershot W (2012)
Estimating phosphorus leachibility in reconstructed soil columns
using HYDRUS-1D model. Environ Earth Sci 65:1751–1758
Flahaut E (2011) Toxicity and environmental impact of carbon
nanotubes carbon nanotubes for biomedical applications. Carbon
Nanostruct 3:211–219
Gao B, Saiers JE, Ryan J (2006) Pore-scale mechanisms of colloid
deposition and mobilization during steady and transient flow
through unsaturated granular media. Water Resour Res
42:W01410
Gao B, Steenhuis TS, Zevi Y, Morales VL, Nieber JL, Richards BK,
Parlange JY (2008) Capillary retention of colloids in unsaturated
porous media. Water Resour Res 44:W04504
Jaisi DP, Elimelech M (2009) Single-walled carbon nanotubes exhibit
limited transport in soil columns. Environ Sci Technol
43:9161–9166
Jaisi DP, Saleh NB, Blake RE, Elimelech M (2008) Transport of
single-walled carbon nanotubes in porous media: filtration
mechanisms and reversibility. Environ Sci Technol
42:8317–8323
Kennedy AJ, Hull MS, Steevens JA, Dontsova KM, Chappell MA,
Gunter JC, Weiss CA Jr (2008) Factors influencing the
partitioning and toxicity of nanotubes in the aquatic environ-
ment. Environ Toxicol Chem 27:1932–1941
Klaine SJ, Alvarez PJJ, Batley GE, Fernandes TF, Handy RD, Lyon
DY, Lead JR (2008) Nanomaterials in the environment: behav-
ior, fate, bioavailability, and effects. Environ Toxicol Chem
27:1825–1851
Liu X, O’Carroll DM, Petersen EJ, Huang Q, Anderson CL (2009)
Mobility of multi-walled carbon nanotubes in porous media.
Environ Sci Technol 43:8153–8158
Mattison NT, O’Carroll DM, Kerry Rowe R, Petersen EJ (2011)
Impact of porous media grain size on the transport of multi-
walled carbon nanotubes. Environ Sci Technol 45:9765–9775
Mauter MS, Elimelech M (2008) Environmental applications of
carbon-based nanomaterials. Environ Sci Technol 42:5843–5859
Monteiro-Riviere NA, Nemanich RJ, Inman AO, Wang YY, Riviere
JE (2005) Multi-walled carbon nanotube interactions with
human epidermal keratinocytes. Toxicol. Letters 155:377–384
Monthioux M, Serp P, Flahaut E, Razafinimanana M, Laurent C,
Peigney A, Broto JM (2007) Introduction to carbon nanotubes.
Springer Handbook of Nanotechnology, ISBN: 978-3-540-
29855-7. Springer, Berlin
Pan B, Xing B (2008) Adsorption mechanisms of organic chemicals
on carbon nanotubes. Environ Sci Technol 42:9005–9013
Perez S, Farre M, Barcelo D (2009) Analysis, behavior and
ecotoxicity of carbon-based nanomaterials in the aquatic envi-
ronment. Trends Anal Chem 28:820–832
Petersen EJ, Zhang L, Mattison NT, O’Carroll DM, Whelton A,
Uddin N, Holbrook RD (2011a) Potential release pathways,
environmental fate, and ecological risks of carbon nanotubes.
Environ Sci Technol 45:9837–9856
Petersen EJ, Pinto RA, Zhang L, Huang Q, Landrum PF, Weber WJ Jr
(2011b) Effects of Polyethyleneimine-mediated functionalization
of multi-walled carbon nanotubes on earthworm bioaccumulation
and sorption by soils. Environ Sci Technol 45:3718–3724
Pfletschinger H, Engelhardt, Piepenbring M, Koniger F, Schuhmann
R, Kallioras A, Schuth C (2012) Soil column experiments to
quantify vadose zone water fluxes in arid settings. Environ. Earth
Sci 65:1523–1533
Phenrat T, Lowry GV, Hotze EM (2010) Nanoparticle aggregation:
challenges to understanding transport and reactivity in the
environment. J Environ Qual 39:1909–1924
Rao GP, Lu C, Su F (2007) Sorption of divalent metal ions from
aqueous solution by carbon nanotubes: a review. Sep Purif
Technol 58:224–231
Schijven JF, Hassanizadeh SM (2000) Removal of viruses by soil
passage: overview of modeling, processes, and parameters. Crit
Rev Environ Sci Technol 30:49–127
Scott-Fordsmand JJ, Krogh PH, Schaefer M, Johansen A (2008) The
toxicity testing of double-walled nanotubes-contaminated food to
Eisenia veneta earthworms. Ecotoxicol Environ Safety 71:616–619
Shang J, Flury M, Chen G, Zhuang J (2008) Impact of flow rate, water
content, and capillary forces on in situ colloid mobilization
during infiltration in unsaturated sediments. Water Resour Res
44:W06411
Sharma P, Abdou HM, Flury M (2008a) Effect of the lower boundary
condition and flotation on colloid mobilization in unsaturated
sandy sediments. Vadose Zone J 7:930–940
Sharma P, Flury M, Zhou J (2008b) Detachment of colloids from a
solid surface by a moving air-water interface. J Colloid Interface
Sci 326:143–150
Sharma P, Flury M, Mattson E (2008c) Studying colloid transport in
porous media using a geocentrifuge. Water Resour Res
44:W07407. doi:10.1029/2007WR006456
Shen M, Wang SH, Shi X, Chen X, Huang Q, Petersen EJ, Weber WJ
Jr (2009) Polyethyleneimine-mediated functionalization of
multi-walled carbon nanotubes: synthesis, characterization, and
in vitro toxicity assay. J Phys Chem C 113:3150–3156
Simunek J, van Genuchten MT, Sejna M (2008) Development and
applications of the HYDRUS and STANMOD software pack-
ages and related codes. Vadose Zone J 7:587–600
Tan C, Tan K, Ong Y, Mohamed A, Zein S, Tan S (2012) Energy and
environmental applications of carbon nanotubes. Environ Chem
Lett 10:265–273
Tian Y, Gao B, Ziegler KJ (2011) High mobility of SDBS-dispersed
single-walled carbon nanotubes in saturated and unsaturated
porous media. J Hazard Mater 186:1766–1772
Tian Y, Gao B, Wang Y, Morales VL, Carpena RM, Huang Q, Yang L
(2012a) Deposition and transport of functionalized carbon nanotubes
in water-saturated sand columns. J Hazard Mater 213–214:265–272
Tian Y, Gao B, Morales VL, Wang Y, Yang L (2012b) Effect of
surface modification on single-walled carbon nanotube retention
and transport in saturated and unsaturated porous media. J Hazard
Mater 239–240:333–339
Tiraferri A, Tosco T, Sethi R (2011) Transport and retention of
microparticles in packed sand columns at low and intermediate
ionic strengths: experiements and mathematical modeling.
Environ Earth Sci 63:847–859
Torkzaban S, Bradford SA, Van Genuchten MT, Walker SL (2008)
Colloid transport in unsaturated porous media: the role of water
content and ionic strength on particle straining. J Contaminant
Hydrol 96:113–127
Van Genuchten MT, Leij F, Yates S (1991) The RETC code for
quantifying the hydraulic functions of unsaturated soils. U.S.
Salinity Laboratory, USDA, Riverside
Wan J, Tokunaga TK (1997) Film straining of colloids in unsaturated
porous media: conceptual model and experimental testing.
Environ Sci Technol 31:2413–2420
Wan J, Wilson JL (1994a) Visualization of the role of the gas-water
interface on the fate and transport of colloids in porous media.
Water Resour Res 30:11–24
Wan J, Wilson JL (1994b) Colloid transport in unsaturated porous
media. Water Resour Res 30:857–864
Wang P, Shi Q, Liang H, Steuerman DW, Stucky GD, Keller AA
(2008) Enhanced environmental mobility of carbon nanotubes in
the presence of humic acid and their removal from aqueous
solution. Small 4:2166–2170
Wang C, Wang L, Wang Y, Liang Y, Zhang J (2012) Toxicity effects
of four typical nanomaterials on the growth of Escherichia coli,
Bacillus subtilis and Agrobacterium tumefaciens. Environ Earth
Sci 65:1643–1649
3760 Environ Earth Sci (2014) 71:3751–3760
123