Embed Size (px)
Citation preview
Examensarbete vid Institutionen för geovetenskaper ISSN 1650-6553 Nr 299
Mobilization and Transport of Different Types of Carbon-based
Engineered and Natural Nanoparticles through Saturated
Porous Media
Mobilization and Transport of Different Types of Carbon-based Engineered and Natural Nanoparticles through Saturated Porous Media
Maryeh Hedayati
Maryeh Hedayati
Uppsala universitet, Institutionen för geovetenskaperExamensarbete E1, 30hp i hydrologi/hydrogeologiISSN 1650-6553 Nr 299Tryckt hos Institutionen för geovetenskaper, Geotryckeriet, Uppsala universitet, Uppsala, 2014.
Carbon –based engineered nanoparticles have been widely used due to their small size and unique physical and chemical properties. They can dissolve in water, transport through soil and reach drinking water resources. The toxic effect of engineered nanoparticles on human and fish cells has been observed; therefore, their release and distribution into the environment is a subject of concern. In this study, two types of engineered nanoparticles, multi-walled carbon nano-tubes (MWCNT) and C60 with cylindrical and spherical shapes, respectively, were used. The aim of this study was to investigate transport and retention of carbon-based engineered and natural nanoparticles through saturated porous media. Several laboratory experiments were conducted to observe transport behavior of the nanoparticles through a column packed with sand as a representative porous media. The column experiments were intended to monitor the effect of ionic strength, input concentration and the effect of particle shape on transport. The results were then interpreted using Derjaguin-Landau-Verwey-Overbeak (DLVO) theory based on the sum of attractive and repulsive forces which exist between nanoparticles and the porous medium. It was observed that as the ionic strength increased from 1.34 mM to 60 mM, the mobility of the nanoparticles was reduced. However, at ionic strength lower than 10.89 mM, mobility of C60 was slightly higher than that of MWCNTs. At ionic strength of 60 mM MWCNT particles were significantly more mobile. It is rather difficult to relate this difference to the shape of particle and further studies are required.The effect of input concentration on transport of MWCNTs and C60 was observed in both mobility of the particle and shape of breakthrough curves while input concentration was elevated from 7 mg/l to 100 mg/l. A site-blocking mechanism was suggested to be responsible for the steep and asymmetric shape of the breakthrough curves at the high input concentration.Furthermore inverse modeling was used to calculate parameters such as attachment efficiency, the longitudinal dispersivity, and capacity of the solid phase for the removal of particles. The inversion process was performed in a way that the misfit between the observed and simulated breakthrough curves was minimized. The simulated results were in good agreement with the observed data.
Supervisor: Prabhakar Sharma
Examensarbete vid Institutionen för geovetenskaper ISSN 1650-6553 Nr 299
Mobilization and Transport of Different Types of Carbon-based
Engineered and Natural Nanoparticles through Saturated
Porous Media
Maryeh Hedayati
Copyright © Maryeh Hedayati and the Department of Earth Sciences Uppsala University Published at Department of Earth Sciences, Geotryckeriet Uppsala University, Uppsala, 2014
I
Abstract
Carbon –based engineered nanoparticles have been widely used due to their small size and unique
physical and chemical properties. They can dissolve in water, transport through soil and reach drinking
water resources. The toxic effect of engineered nanoparticles on human and fish cells has been
observed; therefore, their release and distribution into the environment is a subject of concern. In this
study, two types of engineered nanoparticles, multi-walled carbon nano-tubes (MWCNT) and C60 with
cylindrical and spherical shapes, respectively, were used. The aim of this study was to investigate
transport and retention of carbon-based engineered and natural nanoparticles through saturated porous
media. Several laboratory experiments were conducted to observe transport behavior of the
nanoparticles through a column packed with sand as a representative porous media. The column
experiments were intended to monitor the effect of ionic strength, input concentration and the effect of
particle shape on transport. The results were then interpreted using Derjaguin-Landau-Verwey-
Overbeak (DLVO) theory based on the sum of attractive and repulsive forces which exist between
nanoparticles and the porous medium. It was observed that as the ionic strength increased from 1.34
mM to 60 mM, the mobility of the nanoparticles was reduced. However, at ionic strength lower than
10.89 mM, mobility of C60 was slightly higher than that of MWCNTs. At ionic strength of 60 mM
MWCNT particles were significantly more mobile. It is rather difficult to relate this difference to the
shape of particle and further studies are required.
The effect of input concentration on transport of MWCNTs and C60 was observed in both
mobility of the particle and shape of breakthrough curves while input concentration was elevated from
7 mg/l to 100 mg/l. A site-blocking mechanism was suggested to be responsible for the steep and
asymmetric shape of the breakthrough curves at the high input concentration.
Furthermore inverse modeling was used to calculate parameters such as attachment efficiency,
the longitudinal dispersivity, and capacity of the solid phase for the removal of particles. The inversion
process was performed in a way that the misfit between the observed and simulated breakthrough
curves was minimized. The simulated results were in good agreement with the observed data.
Keywords: Nanoparticle, MWCNT, C60, Transport and mobilization, DLVO theory, Ionic strength, Input concentration, Clean-bed filtration theory, Column test
II
Populärvetenskaplig sammanfattning
Nanopartiklar avser partiklar med en eller flera dimension som understiger 100 nm. De besitter unika
kemiska och fysikaliska egenskaper. Kolbaserade nanopartiklar, som behandlas i denna studie,
används i många typer av produkter, alltifrån sportutrustning till medicinska apparatur, elektroniska
komponenter och vattenrening. Dessa partiklar kan hamna på markyta under tillverkningsprocessen,
transport, användning och avfallshantering. Toxiska effekten av kolbaserade nanopartiklar har
observerats på mänskliga celler och fiskceller. Detta gör att det finns en oro för de potentiella riskerna
för människa och miljö vid användandet av nanomaterial. Få studier har gjorts för att undersöka
transport och deponering av nanopartiklarna i sand och jord under varierande kemiska och fysikaliska
förhållanden. Det har observerats att avsättningen av partiklarna i sanden är starkt beroende av de
kemiska och fysikaliska egenskaperna.
I denna studie har flera experiment i laboratorieskala genomförts för att förstå effekten av
koncentration och jonstyrka på transport och deponering av kolbaserade nanopartiklar. Två typer av
kolbaserade nanopartiklar som har olika former studerades; flerväggiga kolnanorör (MWCNT) och C60
partiklar med cylindriska eller sfäriska former. En kolonn av sand användes som ett representativ
poröst material. Resultaten indikerar att hög jonstyrka och låg koncentration av dessa två typer av
nanopartikelsuspensioner ökar avsättningen av dem i porösa material och hindrar dem att transporteras
till en dricksvattenkälla (t ex en akvifär).
Keywords: Nanopartiklar, Nanokoltuber, MWCNT, C60, Partikeltransport, DLVO teori, Jonstyrka, Kolonntest
III
Contents Abstract .................................................................................................................................................... I
Populärvetenskaplig Sammanfattning ..................................................................................................... II
List of Figures ....................................................................................................................................... IV
List of Tables .......................................................................................................................................... V
Abbreviations ........................................................................................................................................ VI
1 Introduction .......................................................................................................................................... 1
1.1 Background ................................................................................................................................... 2
1.2 Objectives ...................................................................................................................................... 5
2 Theory .................................................................................................................................................. 6
2.1 DLVO Theory ............................................................................................................................... 6
2.2 Modeling ....................................................................................................................................... 8
3 Materials and Methods ....................................................................................................................... 10
3.1 Porous Media ............................................................................................................................... 10
3.2 Carbon-based Nanoparticles ........................................................................................................ 11
3.2.1 Multi-Walled Carbon Nanotubes ......................................................................................... 11
3.2.2 Fullerene C60 ......................................................................................................................... 12
3.2.3 Fire-born particles (FBP) ...................................................................................................... 13
3.3 Column Experiments ................................................................................................................... 13
3.4 DLVO calculations ...................................................................................................................... 15
3.5 Modeling ..................................................................................................................................... 16
4 Results and Discussion .................................................................................................................. 17
4.1 Effect of Solution Chemistry ................................................................................................. 17
4.1.1 Effect of ionic strength .................................................................................................. 17
4.1.2 Effect of input concentration ......................................................................................... 21
4.2 FBP ........................................................................................................................................ 24
4.3 Simulated Results .................................................................................................................. 25
4.4 DLVO energy profiles ........................................................................................................... 28
4.5 Effect of particle shape .......................................................................................................... 32
5 Conclusion ..................................................................................................................................... 35
6 Acknowledgements ....................................................................................................................... 36
7 Reference ............................................................................................................................................ 37
Appendix ............................................................................................................................................... 39
IV
List of Figures Figure 1: Schematic of C60 (a) and MWCNT (b) structure (after Martins-Júnior et al. 2013 and Yadav, 2008)........................................................................................................................................................ 2 Figure 2: Mechanisms responsible for transport of naniparticles to the surfaces of collectors ............... 3 Figure 3: van der Waals attractive energy, electrostatic double layer repulsive energy and net energies (DLVO) have been plotted together as a function of separation distance. The forces are net attractive in secondary and primary minimum ........................................................................................................ 7 Figure 4: Schematic of functionalizing MWCNT using sulfuric and nitric acid .................................. 12 Figure 5: Schematic of the experimental set up. Column sand has been used as representative porous media. A pump was used to inject the solutions into the column and a spectrophotometer was used to measure the absorbance of the outflow every 1 min. ............................................................................ 14 Figure 6: Simulated (solid lines) and observed (dashed lines) BTCs of MWCNT at different ionic strength (phase 1 and 2), at all experiments pH was 7 and input concentration was equal to 7 mg/l. The BTCs are simulated for only phase 1 and 2…………………………………………………………...18 Figure 7: Proportion of the outflow particles during (a) phase 1 and 2 (b) phase3, to the total input particles (MWCNTs) into the column (The values are the average of two or three replication of the experiments ±standard deviation)…………………………………………………………………...…19 Figure 8: Simulated and observed BTCs of C60 at different ionic strength (phase 1 and 2). The BTCs are simulated for only phase 1 and 2……………………………………………………………….….20 Figure 9: Proportion of the outflow particles during (a)phase 1 and 2 (b) phase3, to the total input particles into the column (The values are the average of two of three replication of the experiments ±standard deviation) .............................................................................................................................. 21 Figure 10: Simulated (solid lines) and observed (dash lines) BTCs of MWCNT (ionic strength of 60mM and pH of 7) at different concentrations: 7 mg/l (E9) and 100 mg/l (E11). The BTCs are simulated for only phase 1 and 2 ........................................................................................................... 22 Figure 11: Simulated (solid lines) and observed (dash lines) BTCs of C60 at different concentrations: .7 mg/l (E5) and 100 mg/l (E10). The BTCs are simulated for only phase 1 and 2 .................................. 23 Figure 12: Simulated (solid lines) and observed (dash lines) BTC of FBP at ionic strength 61.12 mM and pH =7. The BTC is simulated for only phase 1 and 2 .................................................................... 24 Figure 13: Relationship between ionic strength, Smax and Attachment efficiency for MWCNTs and C60 .......................................................................................................................................................... 27 Figure 14: DLVO net energy profile: energy distribution versus the distance of particles [(a) MWCNT (Diameter), (b) MWCNT (Length), (c) FBP and (d) C60 and collector under different ionic strength. 29 Figure 15: DLVO energy profile. Secondary energy minimum at different ionic strength for MWCNT. Dashed lines are calculated based on length of particles and solid lines are based on diameter of the particles ................................................................................................................................................. 30 Figure 16: DLVO energy profile. Secondary energy minimum at different ionic strength for C60 particles ................................................................................................................................................. 30 Figure17: Comparing observed BTCs of MWCNT and C60 under various ionic strengths …………..32 Figure 18: Fraction of nanoparticle mass in the effluent to the total injected mass into the column at different ionic strength. The fraction is higher for C60 at lower IS and for MWCNT at higher ionic strength……………………………………………………………………………………………...….33 Figure 19: Comparing observed BTCs of MWCNT and C60 under various ionic strengths. ................ 32
V
List of Tables
Table 1: Chemical composition of the purchased sand (Sibelco Nordic, Baskarp, Sweden)…………10 Table 2: Properties of the sand column and constant experimental condition ...................................... 13 Table 3: Experimental conditions ......................................................................................................... 15 Table 4: The parameters that have been used for DLVO calculations ................................................. 16 Table 5: Input parameters which been used to run the model .............................................................. 16 Table 6: Fitted parameter and their RMSE ......................................................................................... 26 Table 7: Percentage of the retained particles during phase 1 and 2 which remobilized after DI-water injection ................................................................................................................................................. 28
VI
Abbreviations
ENPs Engineered nanoparticles
SWCNTs Single-walled carbon nanotubes
MWCNTs Multi-walled carbon nanotubes
PAH polycyclic aromatic hydrocarbon
DI –water Deionized water
CNTs Carbon nanotubes
CFT Colloid filtration theory
DLVO Derjaguin-Landau-Verwey-Overbeek
IS Ionic strength
BTC Breakthrough curve
RMSE Root mean square error
1
1 Introduction Nanoparticles refer to particles with one or more dimensions less than 100nm. They possess unique
chemical and physical properties because of their small size. Large surface area, high adsorption
capacity, high tensile strength, high electrical and thermal conductivity are some reasons for being
widely used in a range of everyday commercial products. The applications of these particles are
common in medical devices, pharmaceuticals, energy conversion (solar cells, fuel cells, hydrogen
storage, lithium ion batteries and electrochemical super-capacitors), environmental monitoring and
waste water treatments (for the detection and removal of gas pollutants, pathogens, dyes, heavy metals
and pesticides) (Tan et al. 2012), and in lots of consumer products ranging from cosmetics to
electronics, with many upcoming applications (Grassian, 2008).
The engineered nanoparticles may reach the land’s surface during production, transport, use or
disposal. Their migration primarily occurs with water flow. Engineered nanoparticles can be made
from different base minerals, with different structures. The carbon-based nanoparticles, which are
considered in this study, are reported to be hydrophobic and insoluble in water but in a curtain
conditions, when their surface is charged, they are able to form a stable aqueous suspension (Deguchi
et al. 2001). Therefore stable aqueous suspensions of the nanoparticles can migrate through porous
media. Depending on chemical and physical characteristics of the porous medium, the particles may
be retained in the medium. Previous researches have revealed the potential risk of these particles for
human and organisms (Lam et al. 2004; Musee, 2011). In a research study by Musee (2012), which
defines a qualitative quantification of the toxicity level of different nanomaterials based on the
currently available eco-toxicity data, carbon-based nanoparticles are considered as high level
hazardous material. Hence, the capability to determine their fate, mobility and retention helps to assess
the potential risk of these particles to reach the water table and pollute the drinking water resources. In
addition the fundamental knowledge to design and develop effective waste management systems for
industrial, commercial, or household nano-waste streams, can be obtained (Musee, 2011).
Chemical and physical conditions of the environment can affect mobilization of nanoparticles
through porous media. Lots of research has been done in order to investigate the effect of chemistry of
the solution, water content, porous media grain size (Mattison et al. 2011), degree of saturation
(Mekonen et al. 2014), flow rate (Mekonen et al. 2014) (Liu et al. 2009), concentration (Kasel et al.
2013) and the particle diameter (O’Carroll et al. 2013). The effect of shape of carbon nanoparticles on
their transport and retention is still only scarcely reported (Seymour et al., 2013).
In this study I have examined the mobilization and transport of two different carbon-based,
engineered nanoparticles and natural carbon, which is present in a sample from a fire location, by
using several laboratory scale column experiments. Two types of engineered carbon-based
nanoparticles have been studied: multi-walled carbon nanotube (MWCNT) and C60, which
havecylindrical and spherical shapes, respectively, together with natural, fire born particles (FBP).
Particles
through
T
several
concentr
compare
experime
simulate
efficienc
effective
1.1 BaCarbon n
These pa
neighbor
The cylin
(Ngo et
multiwal
(concent
0.7 to 1.0
Fi
T
al. 2008)
nanopart
evaluate
Mattison
transport
s which hav
soil and reac
This researc
laboratory
ration) on tra
ed in order to
ent was spe
ed by an in
cy and maxi
e parameters
ackgroundnanotubes (C
articles are f
ring carbon w
nders of CN
al. 2008). C
ll carbon na
tric tubes) w
0 nm.
igure 1: Schem
The effectiv
) and much
ticles migra
the transpo
n et al. 2011
t behavior; a
e been prod
ch ground wa
ch was plann
scale exper
ansport of M
o find the eff
ecifically des
nversion mo
imum adsorp
of retention
d CNTs) and C
formed by ca
which create
Ts are betwe
CNTs are clas
anotubes (M
while SWCNT
matic of C60 (a) a
e pore diame
greater than
ate through so
ort behavior
; Pennell et
a column tes
duced during
ater and pote
ned to first
riments. Eff
MWCNT and
fect of the sh
signed to un
del to find
ption capacit
and remobil
C60 are two c
arbon molec
es a spherical
een 0.5 to 50
ssified in tw
MWCNT). M
Ts are only o
and MWCNT (b
eter of soil i
the size of
oil and reach
of nanoparti
al. 2008; W
st is an expe
2
g fire accide
entially pose
observe the
fect of solu
d C60 has bee
hape of these
nderstand FB
transport p
ty. DLVO t
lization of th
classes of ca
cules. In thei
l shape in C6
0 micrometer
wo groups, si
MWCNTs co
one rolled up
b) structure (aft
is typically g
C60 and MW
h groundwat
icles qualitat
Wang, 2009).
erimental tes
ents can trav
a human hea
e transport b
ution chemi
en studied fir
e nanoparticle
BP mobilizat
parameters s
theory was a
he particles.
arbon-based e
ir structure, e
60 and a cylin
r in length an
ngle walled
onsist of mu
p sheet of gr
fter Martins-Jún
greater than
WCNT; there
er. Lots of st
tively and qu
For example
sting approac
vel with fire
alth risk.
behavior of
istry (ionic
rst and then
es on their tr
tion. The re
such as disp
also employe
engineered n
each carbon
ndrical shape
nd 1 to 100 n
carbon nano
ultiple rolled
raphene. C60
nior et al. 2013 a
1000nm (exc
efore it is po
tudies have b
uantitatively
e, column te
ch in which
e-extinguishi
the nanopar
strength an
the results h
ransport beha
esults have b
persivity, at
ed to under
nanoparticles
is bonded w
e in CNTs (F
nanometer in
otubes (SWC
d layers of
has a diame
and Yadav, 200
cept clay) (P
ossible that su
been done in
(e.g. Liu et
ests are used
a column o
ing water
rticles by
nd input
have been
avior. An
been also
ttachment
stand the
s (ENPs).
with three
Figure 1).
diameter
CNT) and
graphene
eter about
08)
Pennell et
uspended
n order to
al. 2009;
do study
of sand is
used as
Experim
numerica
P
advectio
laminar
negligibl
suspensi
B
transport
media (F
al. 2008
liquid ph
surfaces
mechani
Verwey-
Chapter
charge h
variation
aqueous
interactio
retention
around a
depend o
Therefor
an impo
the med
potential
ionic stre
T
factors,
particles
spaces o
particles
rate decr
shape. T
length of
a representa
mental data h
al model (W
Particle tran
on, hydrodyn
flow, absenc
le particle re
ion can be us
Based on tra
t of nanopar
Figure 2). Th
). As nanop
hase due to
of the col
isms are exp
-Overbeak (
2.1). Retent
heterogeneiti
n. The surf
media are
on has the k
n of particl
a particle su
on the conce
re chemical
ortant effect
dium (Elime
l is often us
ength and pH
There are so
straining and
s to be remo
of the mediu
s are retained
reases over t
This is only a
f MWCNTs
ative porous
have been us
Westerhoff and
nsport throug
namic disper
ce of particle
elease or dec
sed (Chapter
aditional cle
rticles to the
hese mechan
articles com
o attractive f
llector and
plained by
(DLVO) the
tion depend
ies and parti
face of mo
e charged, t
key control o
les. The di
urface and
entration of
properties o
on the reten
elech et al.
ed as an ind
H (Brant et a
ome other m
d site blocki
oved are larg
um (Mattiso
d onto the co
time. CFT is
a rough appr
have been co
media and
ed in transpo
d Nowack, 2
gh porous m
rsion and pa
e–particle int
cay, the one-
2.2).
an-bed filtra
e surfaces o
nism are sed
me close to th
forces betwe
particle. R
Derjaguin-L
eory (discus
s on porous
icle surface
st particles
therefore el
on aggregati
istribution o
the surface
salt in the so
of the solutio
ntion of part
. 1995). Th
dicator of su
l. 2005a).
echanisms a
ing. Strainin
ger than the
on et al. 20
ollector surfa
s based on t
roximate calc
onsidered as
3
all the phys
ort models i
2013).
media can b
article reten
teractions on
dimensional
ation theory
of collectors
dimentation,
he surface of
een the
Removal
Landau-
ssed in
s media
charge
in an
ectrical
ion and
of ions
charge
olution.
on have
ticles in
he zeta
urface charge
affecting rete
ng is a phys
pore size; th
011). The si
ace, as it bec
this assumpt
culation for
effective siz
ical and che
in order to d
be explained
ntion. Assum
n the surface
advection–
(CFT), four
(sand grain
inertia, diffu
f a collector
e of particle
ention of nan
sical forcing
herefore pa
te blocking
comes cover
ion that coll
using CFT f
ze in previou
Figure 2: Mecnanoparti
emical param
determine par
d mathematic
ming conditio
(i.e., clean b
dispersion e
r mechanism
s) as they m
usion and int
r, they may b
s which vari
noparticles su
mechanism
rticles are re
mechanism
red by partic
lector and pa
for MWCNT
us studies.
chanisms respoicles to the surf
meters are co
rameter valu
cally by co
ons of hom
bed at early ti
equation for
ms are respon
move throug
terception (P
be removed
ies significa
uch as hydro
m, , and occu
etained in sm
m occurs wh
cles, and the
article have
Ts. Both diam
onsible for transfaces of collecto
ontrolled.
ues in the
nsidering
ogeneity,
ime), and
a particle
nsible for
gh porous
Pennell et
from the
ntly with
odynamic
urs when
mall pore
en fewer
retention
spherical
meter and
sport of ors
4
MWCNT and C60 are inherently insoluble in water. The addition of a hydroxyl group to the
structure of particles makes their surface negatively charged and increases their stability in water
(Deguchi et al. 2001). The procedure for functionalizing MWCNTs has been done by adding hydroxyl
groups to their surface using sulfuric and nitric acid. Functionalized MWCNTs have negatively
charged surface such that the particles tend to separate and disperse in water. Two general methods
have been used in previous studies in order to make stable aqueous suspension of C60, the physically
mixing method (Wang et al. 2014) and Solvent Exchange method (Andrievsky et al. 1995; Brant et al.
2006; Wang et al. 2014; ).
Number of studies have been done in the past decades to investigate the effect of chemical and
physical parameter on transport of MWCNTs. O’Carroll et al. (2013) have observed that smaller
diameter MWCNTs are more retained in the column compared with the larger ones, perhaps due to
more collision due to Brownian motion. Although, the results of research by Mekonen et al. (2014)
showed that shorter MWCNTs are more mobile than their longer counterparts. Liu et al. (2009) have
reported that greater pore water velocity significantly mobilizes MWCNT. The combined effect of
ionic strength and pH was examined by Yuan Tian (2012). The results show strong influence of ionic
strength and pH on mobilization of MWCNTs. He has reported significant increases in irreversible
retention of particles with slight increase of ionic strength. Effect of input concentration has also been
studied by Kasel et al. (2013). They observed that the relative concentration of effluent increased by
increasing input concentration and also the shape of the breakthrough curve (BTC) was more
asymmetric. Mekonen et al. (2014) reported slightly lower mobility of MWCNTs in fine sands
(d50=150μm) in comparison with coarse sand (d50= 300 μm).
Few studies have been done in order to understand transport behavior of C60 in porous media
(homogeneous and heterogeneous). Among these few studies, transport properties vary considerably.
It has been observed that mobilization of C60 particles decreases with decreasing the medium grain size
and pore water velocity (Wang, 2009). In research by Wang (2009), the C60 retention was irreversible
and only 6.9% of retained particles remobilized after injecting deionized (DI) water. It was also
concluded that the retention of MWCNT in sand increases with increasing ionic strength independent
of the type of salt (Ca2+ or Na+) and clean-bed filtration theory was not able to describe transport of
C60 particles because the breakthrough curves were not symmetric.
5
1.2 Objectives The primary aim of this study was to examine transport and retention of different types of engineered
and natural carbon-based nanoparticles through saturated porous media.
The following points were specific objectives of this study:
To study the effect of solution chemistry (ionic strength and input concentration) on
transport and mobilization of carbon-based nanoparticle. Therefore two types of
engineered nanoparticles with different shapes (MWCNT and nC60) have been
investigated
To investigate the effect of shape of engineered nanoparticles on their transport and
retention in saturated porous media. Therefore the observed BTCs of MWCNT and C60
were compared to observe the differences and understand the possible effect of shape of
particles on their transport behavior.
To conduct an experiment to understand mobilization of natural carbon present in a
sample of fire born particles (fire- extinguishing water from a single fire location) and
compare its mobilization with the engineered nanoparticles.
To simulate the results of experiments using a 1-D finite element model to find transport
parameters such as dispersivity, attachment efficiency and maximum adsorption capacity
and also examine the ability of the model to describe the observed particle retention.
6
2 Theory
2.1 DLVO Theory Dispersed particles in a solution, sized between 1nm to 1μm, tend to aggregate and deposit out of the
suspension. Derjaguin-Landau-Verwey-Overbeak (DLVO) theory is the most common theory used to
explain aggregation of sub micrometer spherical particle and approximated for non-spherical particles.
Recently lots of studies have been done to understand manufactured nanoparticle aggregation,
transport and retention under different conditions, based on DLVO theory (e.g. Kasel et al. 2013;
Mekonen et al. 2014).
Due to physical processes, particle surfaces interact with each other and this contact can lead
to attachment or repulsion of the particles. Short-range thermodynamic interactions allow for particle-
particle attachment to occur. If the attachment is between two similar particles it is called homo-
aggregation (nanoparticle-nanoparticle attachment) and if it is between two dissimilar particles it is
called hetero-aggregation (nanoparticle-sand particle attachment). Aggregation between manufactured
nanoparticles and natural particles are more probable than homo-aggregation which is due to higher
quantity of the natural particles in the environment. Therefore hetero-aggregation has higher effect on
the retention process. The probability of attachment to occur is defined by ‘attachment efficiency’
(α)(Pennell et al. 2008).
Based on DLVO theory, van der Waals attractive energies (EV) and electrostatic double layer
(Eedl) energies control the interaction between particles and the sum of these two energies (the
interaction energy (Ei) ) can be used to predict the probability of particle attachment (Wang et al.
2008). The resultant force can be negative or positive (i.e., repulsive or attractive). Plot of van der
Waals attractive (Ev) and electrostatic double layer (Eedl) repulsive energy versus separation distances
between particles together with sum of these two energies (Ei) can help to understand the process
(Figure 3). Particles can aggregate at primary or secondary energy minima, where the attractive energy
exceeds the repulsive energy. Aggregation at the primary minimum is irreversible but there is
possibility of attachment at secondary minimum to separate again. Ev and EEDL can be calculated by
Gregory equations (Gregory, 1981).
1.
ln 1.
(2)
(3)
Where, A is Hamaker constant (6.7× 10-21 J for C60-water-C60 is (Chen and Elimelech, 2006) ,
9.8×10-21 J for MWCNTs (Mekonen et al. 2014) and 3.84×10-21 for FBP (Iverfelt, 2014)). λ is the
characte
shown b
k
valence,
sedimen
potential
I
constant
NA is the
Figure 3: Vplotted tog
ristic wavele
by ’a’. ‘d’ is t
k is the Bo
e is the elec
ts (the surfa
l) (Sharma et
If εr is the
t, the inverse
e Avogadro n
Van der Waals gether as a func
ength of the
the separatio
ltzmann con
ctron charge
ace potentia
t al. 2008).
relative diel
of Debye-H
number and I
attractive energction of separati
interaction w
n distance.
nstant (1.38.
(1.6× 10-19
al can be ap
lectric const
Huckel length
Ic is the ionic
gy, electrostaticion distance. Th
7
which assum
10 -23 J/K),
coulombs), Ψ
pproximated
tant of water
h ’κ’ can be c
c strength of
c double layer rhe forces are ne
es to be 100
T is the ab
Ψ is the surf
by the dete
r (78.55) an
calculated fro
f the solution
epulsive energyet attractive in s
nm. The rad
bsolute temp
face potentia
ermination o
nd ε0 is the
om the follow
n (Wang et al
(4)
y and net energiecondary and p
dius of the pa
perature, z is
al of particle
of the surfa
vacuum per
wing formul
l. 2008).
ies (DLVO) haprimary minimu
articles is
s the ion
s and the
ce’s zeta
rmittivity
la. Where
ve been um
8
2.2 Modeling of nanoparticle transport Transport of MWCNTs in porous media can be simulated with a numerical model (one dimensional
(1D) finite element code) (Liu et al. 2009). The model was generated using MATLAB by Denis M.
O'Carroll (Assistant Professor, University of western Ontario), modified by Prabhakar Sharma, to
solve following mass balance equations.
The aqueous phase mass balance is represented as:
0 (5)
Where C is the concentration of nanoparticle in the aqueous phase, t is time, ρbis the solid phase bulk
density, n is porosity, S is the amount of particles associated with the solid phase, v is the pore water
velocity, x is the spatial dimension in the column and D is the dispersion coefficient (D = v * αl, where
α1 is the longitudinal dispersivity).
The solid phase mass balance equation is represented as:
ψCρ
S 0 (6)
Where katt is the removal rate constant associated with mechanisms typically linked with colloid
filtration theory. Diffusion (random movement of particles), sedimentation (due to gravity) and
interception (moving of particles through stream lines close to the collector) are the three mechanisms
that cause the particles to reach the collector’s surface. ψ is an absorption site blocking term and kdet is
the rate constant for the detachment of nanoparticles associated with the solid phase. The adsorption
site blocking term is defined as:
1 (7)
Where Smax is the maximum adsorption capacity of the solid phase for the removal of particles due to
mechanisms typically associated with colloid filtration theory, the removal rate constant katt is defined
as:
kαη (8)
Where α is the attachment efficiency, η0 is the theoretical single collector efficiency and dc is the mean
diameter of the collector (grain). The attachment efficiency (α) typically fits to the experimental data
whereas the theoretical single collector efficiency is based on theoretical considerations.
In this study the theoretical collector efficiency (η0) is assumed to be the sum of three distinct collision
mechanisms, collision due to particle interception (ηI), gravity sedimentation (ηG), and diffusion (ηD).
For MWCNTs, it is assumed that they are uniform cylinders with diameter of dp and length of l,
therefore interception can occur both in side or end of cylinders (Liu et al. 2009).
9
For “end contact”, η0 is defined as follow:
3 (9)
And “side contact” is defined as below:
12
3 10
where
. (11)
and
4.03 ln
1 1
332
1
(12)
For spherical shape C60 particles, the theoretical collector efficiency (η0) is calculated by the following
equation (Logan et al. 1995):
0.003818
4
(13)
14 , 1 15 , 16 and 3⁄ (17)
Where l is the length of colloid and dp is the colloid diameter, ρp is the particle density,
ρ is the fluid density, H is Hamaker constant, v is pore water velocity, n is porosity, D is
diffusion coefficient, μ is the fluid viscosity, T is the absolute temperature, and k is the
Boltzmann constant.
10
3 Materials and Methods In this section, the properties of the materials and the preparation methods of aqueous suspension of
each material have been described. The characteristics of the porous media and the experimental
conditions are also clarified.
3.1 Porous Media The silica sand was purchased from Sibelco Nordic, Baskarp, Sweden to be used as representative
porous media. The chemical composition of the sand showed that it was not pure silica and there were
some impurities that need to be removed (table 1). Hydrochloric acid 0.1 M and hydrogen peroxide
(7%) were used to wash the sand to get rid of the composition other than silica (Huang and Weber,
2004). Hydrochloric acid is a strong acid that react with the impurities in sand and dissolves them.
Hydrogen peroxide (H2O2) is the simplest peroxide and a strong oxidizer. The standard procedure to
wash the sand is based on a method proposed by Huang and Weber (2004). To prepare 0.1 M
hydrochloric acid solution, 12 ml HCl 37% was added to 1 liter of de-ionized water. 1.5 kg sand,
which is previously passed through 250 and 400 μm sieves, was added to the HCl solution and left for
30 min to allow reactions to be completed and impurities to be dissolve in acid. Afterward the liquid
was poured out of the sand flask and the sand was washed by DI water several times.
Table 1: Chemical composition of the purchased sand (Sibelco Nordic, Baskarp, Sweden)
parameter SiO2 Al2O3 Fe2O3 CaO MgO Na2O K2O
Result % weight
90.9 4.9 0.52 0.37 0.1 1.1 1.96
uncertainty ±0.5 ±0.3 ±0.02 ±0.01 ±0.01 ±0.02 ±0.04
Then the sand was washed by H2O2 solution. 240ml of H2O2 (30%) was added to 1050 ml DI
water gradually. The sand was added to the solution and allowed to react for 40 minute to allow the
reactions to be completed. Finally the sand was washed several times with DI water to remove the
impurities completely. The washed sand was dried in the oven at 105˚C for 24 hours and stored in a
plastic bottle.
11
3.2 Carbon-based Nanoparticles Multiwall carbon nanotubes (MWCNTs) and C60 are the two types of engineered carbon-based
nanoparticles that were purchased in order to use in this study. The nanoparticles were inherently
hydrophobic. Therefore laboratory procedures were done to make them hydrophilic and stable in
water. Polycyclic aromatic hydrocarbons (PAH) which are present in the fire extinguishing water were
also studied as natural carbon-based nanoparticles. The details of methods which have been used to
prepare aqueous suspension of each type of nanoparticles in this study are described in the following
section.
3.2.1 Multi-Walled Carbon Nanotubes
Multiwalled carbon nanotubes (MWCNTs) purity >95%, with diameters 20-30 nm and length of 0.5-2
μm were purchased from Cheap Tubes, Inc (Brattleboro, USA). The particles were hydrophobic and
unstable in water, thus functionalization of MWCNTs has been done in order to make them
hydrophilic. A concentrated sulfuric/nitric acid was used to functionalize the nanotubes which resulted
in a large concentration of carboxylic acid (-COOH) and Hydroxyl (-OH) groups on the nanotube
surface (Figure 4). The added groups to the particle surface produce high negative zeta potential
depending on pH of the solution. In order to generate functionalized MWCNTs, a mixture of 240 ml
sulfuric acid (95-97%) and 80 ml nitric acid (70%) was prepared (Mekonen et al. 2014) . Afterward
320 mg untreated MWCNTs were added to the mixture of acids. The solution was then set into the
bath sonicator for 2 hours at room temperature to disperse the particle in the liquid. The solution was
placed on a hot plate at 90˚C and stirred slowly by using magnetic stirrer for 5 hours. The resultant
solution was then filtered through 0.2 μm hydrophilic polypropylene membrane filters and washed
several times using hot DI-water to remove the residual acid. A vacuum was used to accelerate the
filtering process. The functionalized MWCNTs were scraped out of the membranes using spatula and
finally the filters were washed by DI water to remove all the particles. The resulting solution was dried
in the oven and the remaining powder was collected and kept in a small glass bottle.
For column experiments, dispersed solutions of MWCNT with concentrations of 7 mg/l and
100 mg/l were required. To prepare 7mg/l solution, 7 mg of the functionalized MWCNTs was added
to 200 ml de-ionized (DI) water and sonicated for 30 min using an Ultrasonic Homogenizer (Model
3000, BioLogics Inc. Manassas Virginia) with 40% power output. The solution was then diluted by
adding 800 ml of DI- water to achieve the required concentration. The same procedure was used to
prepare higher concentrations. The suspension looked stable after 24 hour and no sedimentation or
aggregations was observed. The pH of the solution was measured using pH meter (METROHM Ltd.
CH-9101, Herisau, Switzerland and its electrical conductivity was measured by a conductivity meter
(Cond 340i/SET, WTW, Germany). A wavelength scan was also done using UV-vis
spectrophotometer (Model DR 5000, Hach Lange Ltd) and a peak was observed at wavelength 300 nm
therefore
absorptio
3.2.2 Fu
Fulleren
this prod
organic
suspensi
are two
exchang
E
natural a
using To
method a
particles
purple co
was dilu
with 60%
the purp
and kept
W
wavelen
solution
solution
approxim
potential
time cou
no chang
e this wavel
on.
F
ullerene C6
ne (C60, sublim
duct in wate
solvents suc
ion has high
methods to
e method (W
Even though
aquatic syste
oluene, follo
also (Wang e
s.
First, 20 mg
olor solution
uted by additi
% of maximu
pose of remo
t at room tem
Wavelength
gth with stab
at this wav
was then d
mated to be 7
l of C60 solu
urse test was
ge in concent
length was c
Figure 4: Schem
60
med 99%) w
er is very lo
ch as Toluen
impact on i
prepare stab
Wang et al. 20
h the extende
ems, the met
owing establ
et al. 2014) b
g of C60 pow
n after severa
ion of 200 m
um power. It
oving the lar
mperature for
scan also h
ble absorptio
velength was
diluted by a
7 mg/l using
ution prepare
done to che
tration durin
chosen to us
matic of function
was purchased
ow (i.e., <10
ne, Tetrahyd
its mobilizat
ble aqueous
014).
ed mixing me
thod that we
lished metho
but it did not
wder was w
al minute of
ml DI water a
t was then fo
ger aggregat
r later use.
has been don
on peak. A fla
s considered
dding 2 lite
the weight o
ed by this m
eck the stabil
ng the test tim
12
se for showi
nalizing MWCN
d from SIGM
0-9 mg/L) (W
drofuran or B
tion and tran
suspensions
ethod is mor
e have used
od given by
t give us a s
weighed and
stirring and
and sonicated
ollowed by o
tes, the solut
ne using spe
at peak was
d as indicato
ers of water.
of removed p
method is abo
lity of the pa
me.
ing the conc
NT using sulfur
MA-ALDRIC
Wang et al. 2
Benzene. Pre
nsport throug
of C60: phy
re indicative
in this resea
y Brant at a
stable disper
added to 20
shaking (Bra
d for 1.5 hou
one hour soni
tion was filt
ectrophotom
observed at
or of C60 co
. The conce
particles from
out -52 ±6 m
articles in th
centration of
ric and nitric ac
CH. Co. The
2008) but it
eparation me
gh porous m
sical mixing
of condition
arch is solve
l. We tried
sed solution
ml of tolue
ant et al. 200
urs using Ultr
ication using
tered using 1
eter to find
344±3. The a
oncentration
entration of
m the solution
mV(Yang et
e solution an
f MWCNT
cid
inherent sol
t is highly s
ethod of C60
media. Genera
g method an
ns that might
ent exchange
the physica
with the sel
ene that resu
06). Then the
trasonic Hom
g bath sonica
11μm nylon
out the app
absorption o
in the solut
diluted solu
n after filtrat
al. 2013). A
nd the result
based on
ubility of
oluble in
0 aqueous
ally there
d solvent
t occur in
e method
al mixing
ected C60
ulted in a
e solution
mogenizer
ator. With
net filter
proximate
f light by
tion. The
ution was
tion. Zeta
A 24 hour
s showed
13
3.2.3 Fire-born particles (FBP)
A sample in aqueous phase has been collected from a fire location, where water had been used for
extinguishing the fire. The collected solution was filtered using an 11 μm nylon filter to remove the
particles larger than 11μm.The chemical analysis of the filtered sample has shown that it contains a
large amount of PAH even when filtered (shown in Appendices, table A1).
The pH of the sample was measured by a pH meter which was 9.98 at 22˚C. The electrical
conductivity (EC) of the sample was measured by a conductivity meter. The concentration of particles
in the sample was estimated by weighing the remained powder from drying up 100 ml of the sample in
oven. To be able to compare the transport behavior of these samples with manufactured nanoparticles
(C60 and MWCNT) the pH of the sample has been adjusted to 7 using Hydrochloric acid (0.1M). The
electrical conductivity of the sample was 4.36 mS/cm at 22˚C. By using the following empirical
equation, the ionic strength (IS) of samples was estimated to be around 60mM (Lind, 1970).
IS (M) = (1.4769.10-5 (M/ μS/cm)). (EC (μS/cm)) + 0.00015 (1)
3.3 Column Experiments A glass column (Chormaflex Inc.) with 2.5 cm diameter and 15 cm length, filled with 120 g silica sand
( 250<d<400 μm), was used as representative porous medium. The properties of sand column are
shown in table 2. Two steel mesh filters (0.2 mm) together with two 100μm nylon filters was used at
both side of the column to prevent the sand from moving out of the column. A Peristaltic pump (IPC8,
Ismatic, Peristaltic pump), which was calibrated previously (Figure A1), was used to inject the
solutions into the sand column. The effluent solution was directed to a UV-vis Spectrophotometer
(Model DR 5000, Hach Lange Ltd) in order to measure the absorbance/concentration of the outflow
sample at one minute interval using auto-sampler. The effluent was then collected in small laboratory
tubes by using fraction collector (CF-2, Spectrum Labs Inc.) every 5 minutes.
Table 2: Properties of the sand column and constant experimental condition
Grain size 250-450 μg
Density 2.65 g/cm3
Weight 120 mg
bulk density 1.63 g/cm3
Porosity 0.39
Saturated Pore Volume 28.35 ml
Temperature 21-22˚C
Pore water velocity 7.5 m/d
T
by spect
figure 5.
T
strength)
medium
starting e
to pH 7
ionic str
ml/min.
and ioni
main so
backgrou
column f
particles
A
flow and
strength
solution
collected
conducti
Figure 5: Sused to inj1 min.
The time shi
trophotomete
To pack the
) was injecte
to be satura
each experim
with sodium
rength of 1.3
The aqueous
c strength as
lution (C60,
und solution
for another 3
s.
A tracer test
d hydrodyna
of the solut
which were
d samples wa
ivity versus t
Schematic of thject the solution
ift between t
er was calcu
e sand colum
ed into the co
ated with the
ment the sand
m phosphate
4, 10.89, 32
s suspension
s the backgro
MWCNT o
n was injecte
3 pore volum
t was condu
amic dispersi
tion increase
e followed by
as measured
time).
he experimentalns into the colum
the measurem
ulated later (
mn, the des
olumn while
background
d column wa
e (NaH2PO4
2.7 or 60 mM
n of C60, MW
ound solutio
r FBP) to th
ed for 3 pore
me (phase 3)
ucted three ti
ion in the co
ed to 2mM.
y three pore
using a cond
l set up. Columnmn and a spectr
14
ments of frac
(Figure A2
sired backgr
e the sand wa
d solution and
as flushed by
4 H2O and N
M for three p
WCNT and F
on. Experime
he column f
re volume (P
to investiga
imes using s
olumn. By ad
The tracer t
e volumes of
ductivity me
n sand has beenrophotometer w
ction collecto
& A3). The
round solutio
as added to t
d avoid any
y the backgro
Na2HPO4.7
pore volumes
BP were use
ent was start
for about 5
Phase 2) and
ate the remob
sodium chlor
dding 116.88
test started w
f DI-water. T
ter to obtain
n used as represwas used to mea
or samples a
experiment
on (with adj
the column g
trapped air i
ound solution
H2O, Appen
s with the re
ed as main so
ed (at time z
pore volume
d DI-water w
bilization beh
ride, in orde
8 mg NaCl t
with five po
The electrica
the breakthr
sentative porousasure the absorb
and the meas
tal setup is s
djusted pH a
gradually to
in the column
n which was
ndix) to yiel
equired flow
olution with
zero) by inje
me (Phase 1)
was introduc
havior of the
er to assess t
to DI water
ore volumes
al conductivi
rough curve
s media. A pumbance of the out
urements
shown in
and ionic
allow the
n. Before
s buffered
ld a final
rate of 1
same pH
ecting the
then the
ed to the
e retained
the water
the ionic
of tracer
ity of the
(Relative
mp was tflow every
15
In order to investigate the effect of ionic strength on C60 mobilization, four C60 solutions with
different ionic strength of 1.34, 10.89, 32.7 and 60mM were prepared. The pH of the main solution for
the first four experiments (E1, E2, E3, and E4, table 3) was equal to 7. The pH and ionic strength was
controlled by using Sodium phosphate buffer solution. A C60 solution with high concentration was also
prepared to use at E5 to observe the effect of concentration on C60 particle transport behavior.
Experiment 6 was performed using FBP. The ionic strength of the sample was estimated to be
60mM using the electrical conductivity value. The pH was adjusted at 7 by adding HCl to the solution
drop by drop. Three experiments were also conducted to study the transport and retention of the
MWCNT solution. At experiments 7, 8 and 9 the injected main solution was MWCNT with same
properties but different ionic strength. The ionic strength of each solution was adjusted by adding
sodium phosphate. In order to observe the effect of input concentration on C60 and MWCNT,
experiment 10 and 11 were conducted.
Table 3: Experimental conditions
3.4 DLVO calculations In this study the values of Zeta potential of C60 and MWCNT at pH= 7 are adapted from Brant et al.
(2005) and Sharma et al. (2014), respectively. Brant et al. (2005) have prepared the C60 solution by the
solvent exchange method and sodium phosphate has been used to adjust the ionic strength. The zeta
potential of sand particles at pH=7 at different ionic strength are adapted from Saiers and Lenhart
Experiment nanoparticle Concentration (mg/l)
pH IS (mM)
E0 (tracer) C60 7 8.08 2
E1 C60 7 7 1.34
E2 C60 7 7 10.89
E4 C60 7 7 32.7
E5 C60 7 7 60
E6 FBP 3.38 7 60
E7 MWCNT 7 7 1.34
E8 MWCNT 7 7 10.89
E9 MWCNT 7 7 60
E10 C60 100 7 60
E11 MWCNT 100 7 60
16
(2003) (Table 4). FBP Particle diameter has been estimated using a SEM image (Figure A4) and zeta
potential and Hamaker constant is a rough estimation adapted from a previous study on same samples
(Iverfelt, 2014). To calculate the interaction energy between MWCNTs and porous media, both length
and diameter were used as effective size.
Table 4: The parameters that have been used for DLVO calculations
Particle Diameter (nm)
Length (μm)
Ionic strength (mM)
Zeta potential of sand
(mv)
Zeta potential of
nanoparticle (mv)
Hamaker constant
(J)
C60 160 - 1.34 -55.7 -44 6.7× 10-21
C60 160 - 10.89 -41.5 -25 6.7× 10-21
C60 160 - 32.7 -33.9 -20 6.7× 10-21
C60 160 - 60 -29.1 -15 6.7× 10-21
MWCNT 25 1.25 1.34 -55.7 -50 9.8×10-21
MWCNT 25 1.25 10.89 -41.5 -40 9.8×10-21
MWCNT 25 1.25 60 -29.1 -20 9.8×10-21
FBP 200 - 60 -29.1 -5 3.84×10-21
3.5 Modeling The known parameters which were included in the model are given by table 5. In order to estimate the
unknown parameters the model was programmed to find the best set of parameters which produce the
simulated breakthrough curve with minimum root-mean-square error (RMSE) from the observed
breakthrough curve using an optimization routine. In this model it was assumed that the mechanisms
associated with traditional colloid filtration, detachment of particle from the collector and site blocking
term are the effective mechanisms.
Table 5: Input parameters which been used to run the model
Particle Length of
column (m)
Bulk density
(kg/m3)
Porosity Darcy velocity
(m/s)
Collector
efficiency (η)
MWCNT
0.15 1630 0.39 3.4.10-5
5.93×10-3
C60 5.4×10-2
FBP 7.37×10-3
17
4 Results and Discussion In order to compare transport and mobilization of MWCNT and C60, several experiments have been
performed to observe transport behavior of these two types of particles at different ionic strengths and
input concentrations, while the other parameters were kept constant (e.g. pore water velocity, grain
size, pH, input concentration). The effect of solution chemistry (ionic strength and input
concentration) on the particle transport was examined for both MWCNT and C60, separately in section
4.1. The comparison of transport behavior of MWCNT and C60 with FBP has been discussed in
section 4.5.
4.1 Effect of Solution Chemistry Effect of ionic strength and input concentration on mobilization and transport of both MWCNT and
C60 have been investigated by conducting several column tests (E1-E10). The observed and simulated
breakthrough curves of all experiments have been plotted and discussed in this part.
4.1.1 Effect of ionic strength
4.1.1.1 MWCNT Three column experiments (E7, E8 and E9) were performed to study the effect of ionic strength on
transport and retention of MWCNTs. The observed and simulated breakthrough curves of these
experiments, at three phases, are illustrated by figure 6. Ionic strength of the aqueous suspension of
MWCNT was adjusted at 1.34, 10.89 and 60 mM in experiment E7, E8 and E9, respectively. The
other chemical and physical conditions maintained constant during experiments as described in tables
2 and 3. After starting injection of the suspension at time zero, the first appearance of MWCNT in the
effluent was observed at 0.67 PV which is approximately at the same time with the conservative
tracer.
The maximum effluent concentration for experiment E7, E8 and E9 was 0.96, 0.83 and 0.69,
respectively, and was observed at about 5.67 PV. The area under the BTCs, which corresponds to the
outflow mass after phase 1 and 2 (total 8 PV), equals to 0.88, 0.79 and 0.60 for E7, E8 and E9
respectively. It shows that at higher ionic strength, more particles have been retained in the column. In
order to see the effect of ionic strength on retention of particle in the sand column, the BTCs of
experiments E7, E8, and E9, are shown together figure 6.
The relative concentration at the end part of BTCs did not reach the background value which
indicates that detachments of particles from the sand grains were happening. This indicates that the
attachment of particles to the sand is reversible even before injecting DI-water to the column.
Therefore detachment coefficient (kdet) was considered in the modeling of these results.
The results specify that mobilization of the particles is decreasing with increasing ionic
strength. Enhancing retention of MWCNTs by increasing ionic strength of the solution has been
observed in previous studies (Tian, 2012). Based on DLVO theory, by increasing ionic strength of
18
suspension, negative surface charge of both particle and sand grains is reduced and repulsive
interaction energy is declined.
Figure 6: Simulated (solid lines) and observed (dashed lines) BTCs of MWCNT at different ionic strength (phase 1 and 2), at all experiments pH was 7 and input concentration was equal to 7 mg/l. The BTCs are simulated for only phase 1 and 2
The total eluted mass from the column at phase 1 and 2 was reduced at higher ionic strength
(Figure 7). At ionic strength 1.34 mM the total eluted mass was about 88% which was lowered to 60%
by increasing ionic strength to 60mM. The amount of remobilized mass at phase 3 was increased with
increasing of ionic strength (Figure 6).
The results of Sharma et al. (2014) suggests almost no BTCs for MWCNTs at ionic strength
higher than 4 mM, where pH was 5 and flow rate was 2 ml/min. The size of particles used in that
research was larger than particles used in current study (30-50nm in diameter and length of 10-20 μm)
and the input concentration was less (5 mg/l). Higher input concentration, higher pH, shorter particle
length in this study may be the reason for higher mobility of the particles.
In other work by Liu et al. (2009) the impact of ionic strength was observed to be only on the
rate at which steady-state is achieved. But in the current study the maximum relative concentration and
the total effluent mass is also changing with ionic strength. The ratio of total outflow mass during
phase 1 and 2 to the total inflow mass shows a decreasing trend by increasing ionic strength (Figure
7). The ratio of remobilized particle at phase 3 to the total injected mass in the column has increased
with ionic strength (Figure 7b). The area under the BTCs was calculated and considered as the mass
out for all the experiments (Table A2). At ionic strength of 1.34mM only 10% of the particles retained
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
0 2 4 6 8 10
C/C
0
Pore Volume
E7, IS=1.34 mM E8, IS=10.89 mM
E9, IS=60 mM Tracer
Simulated E7 Simulated E8
Simulated E9
19
in the column after phase 2 while at ionic strength of 60mM about 40% of total influent particles
remained in the column.
Figure 7: Proportion of the outflow particles during (a) phase 1 and 2 (b) phase3, to the total input particles (MWCNTs) into the column (The values are the average of two or three replication of the experiments ±standard deviation)
The results of the current study are very close to a previous study by Mattison et al. (2011). In
that research maximum effluent of around 0.8 for ionic strength of 7.5 mM after 10 PV was observed.
Which is very close to our results for ionic strength 10.89 (maximum effluent = 0.83 at 5.67PV).
Based on our results, higher maximum effluent is expected for ionic strength of 7.5 mM, the
difference is probably because of lower flow rate in their experiment (half of the flow rate in our
study) that causes lower mobility. The input concentrations (8 mg/l), pH (7), the salt for adjusting the
ionic strength (sodium phosphate) are similar. Their particles were shorter in length but larger in
diameter (around 30 to 50 nm diameter and 200 to 800 nm length) than ours. However the size of the
particles has been used in their study is determined by X-ray photoelectron spectroscopy but we have
used the size information from the product information given by manufacturer.
The simulated breakthrough curves are fitted very well with the observed results from
experiments. These results will be discussed later at Chapter 4.3. Remobilization of the particles at
phase 3 will also be discussed in Chapter 4.4.
4.1.1.2 C60 Nanoparticles Four column experiments were conducted in order to investigate the impact of ionic strength on C60
particles mobilization (E1, E2, E4, and E5). Transport of C60 at ionic strength of 1.34, 10.89, 32.7 and
60 mM has been examined. The resulting BTCs are given in figure 8, which shows the relative
concentration of C60 in the effluent during phase1 (five PV injection of C60 suspension), phase 2 (three
PV injection of background solution) and phase 3 (three PV injection of DI-water).
The first appearance of C60 in the effluent was at 0.67 PV for all the experiments which is at
the same time with the conservative tracer. The effluent relative concentration then increased fast until
it reaches a plateau then the relative concentration increases with a much lower rate. The maximum
relative concentration was achieved at 5.67 PV and was equal to 0.93, 0.89, 0.49 and 0.39 for ionic
0,5
0,7
0,9
0 20 40 60 80
Mas
s ou
t/M
ass
in (
mg/
mg)
Ionic Strength (mM)
a average
Upper limit
Lower limit
0,0
0,2
0,4
0,6
0,8
0 20 40 60 80Mas
s ou
t/M
ass
in (
mg/
mg)
Ionic Strength (mM)
b average
Upper limit
Lower limit
20
strength equal to 1.34, 10.89, 32.7 and 60 mM, respectively. After reaching the maximum value, the
relative concentration decreased fast but did not reach zero. As for the MWCNTs, the reduction of
relative concentration at the end part of the BTCs did not reach the background value which could be
explained by the fact that detachments of particles from the sand grains was happening. For that
reason, a detachment coefficient (kdet) was considered in the modeling of the results.
Comparing BTCs of different ionic strength revealed that retention of C60 particles were
enhanced with increase in ionic strength. The same result has been reported by Brant et al. 2005b and
Zhang et al. 2012. At high ionic strength, the BTCs reached plateau in slower rate than at the lower
ionic strength. But the rate at the plateau is slightly elevated at higher ionic strength (Figure 8).
Our results is consistent with the research by Zhang et al., (2012). They have prepared the
suspension of C60 using solvent exchange method and the experiments were performed at pH of 6.8,
ionic strength was increased from 1 mM to 10 mM (NACl). According to their results, by increasing
ionic strength, the shape of BTCs becomes more asymmetric and retention of the particles increases,
which is similar with our findings.
Figure 8: Simulated and observed BTCs of C60 at different ionic strength (phase 1 and 2). The BTCs are simulated for only
phase 1 and 2
The ratio of eluted mass to the total injected mass into the column is decreasing by increasing
ionic strength (Figure 9). At ionic strength 1.34mM, totally 97% of input mass is out of column after
the three phases. While only 70% of input mass comes out of the column at ionic strength 60 mM after
11 PV (five PV injection of C60 suspension+ three PV background solution+ three PV DI-water).
During the experiments it was observed that the stability of C60 suspension is declining by addition of
salt. At low ionic strength, the suspension was stable for 24 hour, but by increasing ionic strength the
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
1,8
0 1 2 3 4 5 6 7 8 9 10 11
C/C
0
Pore volume
E1, IS=1.34 mM E2, IS=10.89 mM
E4, IS=32.7 mM E5, IS=60 mM
Tracer Simulated E1
Simulated E2 Simulated E4
Simulated E5
21
particles aggregated faster and tended to sediment out of the suspension after few hours. Therefore
physical forcing (straining) may be an effective mechanism for retention of some particles at high
ionic strength. Yang et al. (2013) also reported that size of C60 aggregations increase at high ionic
strength as a result of reducing zeta potential.
Figure 9: Proportion of the outflow particles during (a)phase 1 and 2 (b) phase3, to the total input particles into the column (The values are the average of two of three replication of the experiments ±standard deviation)
4.1.2 Effect of input concentration
Two experiments (E10 and E11) were performed to study the effect of input concentration in transport
of C60 and MWCNTs. In these experiments the ionic strength were adjusted at 60 mM and pH was 7.
The observed results were used in the model to obtain the unknown parameters. The results for both
materials have been discussed as follows:
4.1.2.1 MWCNT The ionic strength of MWCNT suspension with concentration of 100 mg/l was adjusted at 60 mM
using Sodium Phosphate buffer. The column experiment was performed for 11 PV (5 PV injection of
MWCNT suspension followed by 3 PV background solution and end with 3 PV DI-water). The
resulting BTC were compared with BTC of Experiment E9 which had approximately 14 times lower
input concentrations (Figure 10). The early appearance of MWCNT in the effluent was at 0.67 PV, at
the same time with conservative tracer test and experiment E9. The relative concentration then
increased rapidly until they reached a plateau with lower increasing rate. The relative concentration at
this point was higher than for same conditions in experiment E9. In the plateu stage (after approx. 1
PV), the relative concentration increased with a higher rate in experiment E11 than in experiment E9.
The relative concentration of MWCNT in the effluent reached a maximum value of 0.92 at 5.53 PV in
experiment E11, while it reached maximum value of 0.63 at 5.66 PV in E9. The retained mass in the
0,2
0,4
0,6
0,8
1,0
0 20 40 60 80
Mas
s ou
t/M
ass
in (
mg/
mg)
Ionic Strength (mM)
a
average
Upper limit
Lower limit
0,0
0,2
0,4
0,6
0,8
1,0
0 20 40 60 80
Mas
s ou
t/M
ass
in (
mg/
mg)
Ionic Strength (mM)
b
average
Upper limit
Lower limit
22
column during phase 1 and 2 was two times higher when the input concentration was 7 mg/l compared
to when it was 100 mg/l (40% of injected mass retained in E9 and 21.97% of injected mass retained in
E11). These results suggest higher mobility of MWCNTs at high input concentration.
Figure 10: Simulated (solid lines) and observed (dash lines) BTCs of MWCNT (ionic strength of 60mM and pH of 7) at
different concentrations: 7 mg/l (E9) and 100 mg/l (E11). The BTCs are simulated for only phase 1 and 2
On the other hand, the mass calculation at phase 3, when DI-water was injected into the
column, shows that at higher input concentration (E11) only 31.82% of retained particle (retained
particle in the sand column during phase 1 and 2) are remobilized at phase 3 while at lower input
concentration (E9), 72.8% of the retained particles remobilized after injection of DI-water into the
column. It indicates that attachment of more than 70% of deposited particle at higher input
concentration is irreversible and cannot be due to a secondary energy minimum. These may be related
to larger aggregation of the particles at higher concentration that results to promote straining
mechanism or even gravitational sedimentation.
These results indicate that input concentration has significant impact on transport of
MWCNTs and higher input concentration results in less retention and high mobility of the MWCNT
particles. This is consistent with the findings of Kasel et al. (2013). They also observed that shape of
BTCs is much steeper for higher input concentration which could be explained by blocking behavior.
Retention sites fill up over time faster when concentration is higher.
4.1.2.1 C60 Nanoparticles
Experiment E10 was performed to examine the effect of input concentration on transport behavior of
C60 particles. The concentration of C60 particles in the injected suspension was about 100 mg/l and
ionic strength was adjusted to 60 mM using Sodium phosphate buffer. The resulting BTC then was
compared with the results of experiment E5 which had same condition but lower input concentration
(Figure11). The BTCs differ both in shape and height. This indicates that at higher input concentration
0
0,5
1
1,5
2
0 1 2 3 4 5 6 7 8 9 10 11
C/C
0
Pore Volume
E9, C0 = 7mg/l
E11, C0 = 100 mg/l
Simulated E9
Simulated E11
Tracer
23
C60 particles are more mobile. The breakthrough curve at E10 was retarded with upper maximum
relative concentration. During phase 1 and 2, 32.5% of the total injected mass of C60 has been
observed in the effluent at the lower input concentration (7mg/l), while the outflow mass was two
times higher (61.78%) at higher concentration (100mg/l).
.
Figure 11: Simulated (solid lines) and observed (dash lines) BTCs of C60 at different concentrations: 7 mg/l (E5) and 100
mg/l (E10). The BTCs are simulated for only phase 1 and 2
At phase 3 when DI-water was injected into the column, only 13 % of retained particles in the
column remobilized while this value was 63.8% at lower input concentration. This is probably due to
improving effect of straining or physically forcing mechanisms at higher concentration.
The BTC of E10 is retarded in compare with BTC of E5. Retardation at the high concentration
suggests that site blocking is an operative mechanism. As retention sites fill with time rapidly, because
of high concentration of particle, after a while the concentration of particle increases in the effluent.
0
0,5
1
1,5
2
2,5
0 1 2 3 4 5 6 7 8 9 10 11
C/C
0
Pore volume
E10, C0 = 100 mg/l
E5, C0 = 7 mg/l
Tracer
Simulated E10
Simulated E5
24
4.2 FBP A column experiment was performed to investigate the transport of FBP. The ionic strength of the
sample was estimated to be 60 mM (chapter 3.4) and the pH was adjusted at 7 with hydrochloric acid.
The BTC shape was asymmetric and the first appearance of particles in the effluent was observed at
0.66 PV (Figure 12). Relative concentration of FBP particle in effluent increased rapidly until reached
the plateau and then gradually increased to the maximum value of 0.92 at 5.7 PV. At total 11 PV
(phase 1, 2 and 3) 90% of the injected mass was recovered in the effluent. Around 10% of the injected
mass was retained in the sand column (during phase 1 and 2) in our experimental conditions which
indicates that FBPs are highly mobile.
Only 12.32% of the retained mass in the column were remobilized after injection of DI-water
into the column. This indicates that the portion of the particles which were retained in the column was
irreversibly attached or was trapped due to physical forcing (straining).
Figure 12: Simulated (solid lines) and observed (dash lines) BTC of FBP at ionic strength 61.12 mM and pH =7. The BTC is simulated for only phase 1 and 2
0
0,4
0,8
1,2
0 2 4 6 8 10
C/C
0
Pore Volume
E6, IS=60 mM
Simulated E6
Tracer
25
4.3 Simulated Results The model which incorporates traditional colloid filtration theory with a site-blocking term and a
detachment rate was in a good agreement with the experimental results (chapter 2.2).
The observed data of all column experiments were used in the model to simulate breakthrough
curves of MWCNT, C60 and FBP. The conservative tracer test results were used to find dispersivity of
the medium. The values of detachment rate and site blocking term were assumed to be zero while
attachment efficiency and dispersivity were fitted to the observed breakthrough curve. The RMSE for
the resulting fit was 0.044 which is an acceptable fit (Figure A5, in Appendix). The fitted dispersivity
with the observed breakthrough curve of the tracer element was 0.0026 m. Since dispersivity is a
medium dependent parameter which depends on porosity and the uniformity of the grain size (Xu and
Eckstein, 1997),the obtained value (0.0026) was kept constant for the rest of the simulations, because
the porous media characteristics were kept constant during the experiments.
The theoretical single collector efficiency (η) was calculated for MWCNT assuming both end
contact (5.93×10-3) and side contact (5.39×10-3) (chapter 2.2). From these calculations it was
concluded that the dominated mechanism is diffusion, which is independent of orientation, and the
effect of interception and gravitational sedimentation is very small which was consistent with findings
of previous study by (Wang et al. 2008). Therefore the difference between the calculated single
collector efficiencies values for end and side contacts were relatively small. Thus, it was assumed that
the attachment of MWCNTs was through side contact with the collector.
Single collector efficiency calculations for C60 particles also confirm that the dominated
mechanism which is responsible for transport of the nanoparticles to the surface of collector is
diffusion (99.7%). To calculate theoretical single collector efficiency for FBP, it was assumed that the
particles were spherical. The average diameter of FBPs was estimated to be 200 nm from a SEM
image of the particles (Figure A4).
For experiment E1 to E11, α1 (dispersivity) was kept constant at 0.0026 and kdet (detachment
rate), α (attachment efficiency), Smax (maximum adsorption capacity) were fitted to the observed
breakthrough curves. The fitted parameters together with the root mean square error of the fitting are
given by table 6.
The RMSE values show that the model is able to describe the experimental results. However
the model is not capable to predict retardation of C60 particle at experiment E10. Therefore the RMSE
value is relatively high for this experiment simulation. The obtained value of Smax at low ionic strength
(1.34 mM) for C60 in this study were similar in order of magnitude with the results of a study by Li et
al. 2008. Comparing the simulated parameters of all experiments, it was observed that the Smax value
increased with increasing ionic strength. The Smax values for MWCNT were one order of magnitude
greater than previous results for MWCNT particles (Mattison et al. 2011) .
Attachment efficiency (α)and Smax are functions of chemical and hydrodynamic factors such as
pH, ionic strength and pore water velocity(Mattison et al. 2011). Smaller Smax refers to fewer site-
26
blocking occurrences. The results suggest that by increasing ionic strength, Smax have increased and
the site-blocking mechanism become more operative (Figure 13). It is probably due to lowering of the
repulsive energy between sand and particles at higher ionic strength and increase in the particle
attachment at secondary energy minimum. At ionic strength 1.34 mM the value of Smax is greater for
MWCNT than for C60, but the value of Smax becomes greater for C60 than for MWCNTs at higher ionic
strength.
Table 6: Fitted parameter and their RMSE
Experiment Particle Concentration (mg/l)
IS (mM)
α Smax Kdet α1
(m) RMSE
E0 - 2 (NACl) 2 5.6×10-3 0 0 2.60×10-3 0.061
E1 C60 7 1.34 9.15×10-3 9.41×10-5 3.42×10-5 2.60×10-3 0.061
E2 C60 7 10.89 1.47×10-2 3.51×10-4 1.10×10-5 2.60×10-3 0.083
E4 C60 7 32.7 2.82×10-2 6.23×10-4 3.49×10-6 2.60×10-3 0.031
E5 C60 7 60 2.66×10-1 6.30×10-4 3.33×10-6 2.60×10-3 0.021
E6 FBP 3.8 61.12 8.79×10-2 3.15×10-4 3.45×10-5 2.60×10-3 0.302
E7 MWCNT 7 1.34 8.71×10-2 1.80×10-4 2.20×10-5 2.60×10-3 0.268
E8 MWCNT 7 10.89 1.41×10-1 2.47×10-4 3.06×10-6 2.60×10-3 0.072
E9 MWCNT 7 60 3.06×10-1 6.13×10-4 1.26×10-5 2.60×10-3 0.060
E10 C60 100 60 6.47×10-1 1.53×10-4 1.12×10-5 2.60×10-3 0.372
E11 MWCNT 100 60 2.61×10-1 3.1×10-4 1.67×10-6 2.60×10-3 0.075
27
Increasing ionic strength is related to increasing Smax and α, for both C60 and MWCNTs in
our experiments, which has also been observed in previous studies (Zhang et al. 2012). This trend
indicated that increasing ionic strength results in greater attachment rate and more attachment sites.
Figure 13: Relationship between ionic strength, Smax and Attachment efficiency for MWCNTs and C60
0,00
0,10
0,20
0,30
0 20 40 60 80
Att
achm
ent
effi
cien
cy
Ionic strength (mM)
C60
MWCNT
0,E+0
2,E‐4
4,E‐4
6,E‐4
0 20 40 60 80S
max
Ionic strength (mM)
C60
MWCNT
28
4.4 DLVO energy profiles The mechanisms responsible for retention and transport of nanoparticles at different chemical
conditions in the experiments were investigated using DLVO energy profiles. The used parameters
and values for these calculations were reported previously in chapter 2.1. The possible mechanisms
that contributed to retention of particles in the media are aggregation, a primary energy minimum, a
secondary energy minimum, straining or gravitational sedimentation. In order to examine
remobilization of the retained particles in the sand column during phase 1 and 2, three pore volume
DI-water was injected into the column (phase 3). Theoretically, attachment of particles to the collector
due to secondary energy minimum is reversible. Therefore, they tend to release into the water and
transport out of the column at very low ionic strength (zero). Thus it was possible to examine the role
of secondary energy minimum for retention of the particles in the media at different ionic strengths.
All BTCs of the experimental data (MWCNT, C60 and FBP experiments) showed a secondary
peak at 9.2 pore volumes which is 1.2 PV after injection of di-water into the column. The peak went
up with increasing ionic strength (Figure 4, 7 and 12). Mass calculations indicated that by increasing
ionic strength more particles were retained in the column during phases 1 and 2. At low ionic strength
(1.34 mM) 59.96 % of the retained MWCNT particles were remobilized during phase 3, while this
percentage increased to 76.32% at high ionic strength (60 mM). On the other hand, a different trend
was observed for C60 particles. At ionic strength 1.34, about 68% of the retained particles were
remobilized at phase 3, while at higher ionic strengths this value decreased to 54% of total retained
particles. These results show that with increasing ionic strength, a reversible attachment mechanism
becomes dominant for MWCNT retention. While for the C60 particles the dominant mechanism leads
to irreversible attachment.
Table 7: Percentage of the retained particles during phase 1 and 2 which remobilized after DI-water injection
IS (mM) Concentration (mg/l)
MWCNT (%) C60 (%) FBP (%)
1.34 7 mg/l 62.39 68.21
10.89 7 mg/l 71.5 54.20
32.7 7 mg/l 52.39
60 7 mg/l 75.56 54.48
60 ≥100 mg/l 31.82 13.95 12.32
In the DLVO energy profiles of both C60 and MWCNTs (Figure 14, 15 and 16) it was
observed that at low ionic strength the energy barrier is very high and the separation distance is very
close at primary energy minimum so the particle should have relatively high energy and be very close
to the collector to allow irreversible attachment to occur. A secondary energy minimum was not also
observed at low ionic strength. As ionic strength increases a primary energy minimum and secondary
energy minimum become more effective because: first, the energy barrier is declining, second, the
separation distance range where primary energy minimum occurs is increasing, and the secondary
29
energy minimum is happening in a bigger distance range. These observations indicate that the effect
of secondary energy minimum and primary minimum is increasing with increasing ionic strength
(Figure 14 a, b and c). The primary energy minimum, which results in irreversible attachment between
collector and particle, occurred in a separation distance less than 0.5nm (Figure 15 and 16). The fact
that this distance is increasing by increasing ionic strength, means that at higher ionic strength
attachment at primary energy minimum can happen in longer separation distance than lower ionic
strength. For example, for MWCNTs at ionic strength 1.34 mM, irreversible attachment can occur
only if the particle is up to 0.01 nm away from the collector (Figure 15). While at ionic strength
60mM, the particle can irreversibly attach to the collector even if they have 0.41 nm separation
distance. Therefore the chance for both irreversible and reversible (primary and secondary energy
minimum) attachment increases with increasing ionic strength. These observations can explain the
increasing retention of particles with ionic strength in the experimental results.
Figure 14: DLVO net energy profile: energy distribution versus the distance of particles [(a) MWCNT (Diameter), (b) MWCNT (Length), (c) FBP and (d) C60 and collector under different ionic strength.
-15
-5
5
15
25
35
45
55
0 5 10 15 20 25 30
ΔE/KT (‐)
Seperation distance (nm)
aIS=1.34 mM
IS=10.89 mM
IS=60 mM
-1000
0
1000
2000
3000
4000
5000
0 5 10 15 20 25 30
Seperation distance (nm)
bIS=1.34 mM
IS=10.89 mM
IS=60 mM
-4
-2
0
2
4
0 5 10 15 20
ΔE
tot /
KT
(-)
Seperation distance (nm)
c
-100
0
100
200
300
400
0 5 10 15 20 25 30
Seperation Distance (nm)
d
IS=1.34 mM
IS= 10.89 mM
IS=32.7 mM
IS=60 mM
30
The mass calculations (table 7) show that even after 3 PV injection of water, 30 to 50% of
retained particles are still trapped in the column. The observed data from experiments show that after 3
PV injection of water the relative concentration of particles in the effluent does not reach the
background level. Therefore it is rather difficult to confirm the mechanism that causes retention of
these trapped particles. To be sure about the effective mechanism, phase 3 of the experiments should
be prolonged until the relative concentration reaches background. If we assume the decreasing
tendency continues until background concentration between 20 to 40% of the particles would be
Figure 15: DLVO energy profile. Secondary energy minimum at different ionic strength for MWCNT. Dashed lines are
calculated based on length of particles and solid lines are based on diameter of the particles
Figure 16: DLVO energy profile. Secondary energy minimum at different ionic strength for C60 particles
-40
-30
-20
-10
0
10
20
30
40
50
60
0 5 10 15 20 25 30 35 40
ΔE
tot/
KT
(-)
IS=1.34 mM (length)
IS=10.89 mM (length)
IS=60 mM (length)
IS=1.34 mM (diameter)
IS=10.89 mM (diameter)
IS=60 mM (diameter)
-4
-2
0
2
4
6
8
10
0 5 10 15 20 25 30 35 40 45 50
ΔE
tot/
KT
(-)
Seperation Distance (nm)
IS=1.34 mM
IS= 10.89 mM
IS=32.7 mM
IS=60 mM
31
trapped in the sand column. The possible mechanisms can be primary energy minimum and straining.
At high ionic strength, both primary and secondary energy minimum become dominating mechanism
for attachment. On the other hand the stability of particles in solution at high ionic strength was also
critical in this study. It was observed that the MWCNT solution was stable for several days even for
60mM of ionic strength. However, C60 solution at higher ionic strength (10.89 and 60mM) was semi-
stable and aggregated C60 particles were observed in the bottle after few hours after their preparation.
By augmenting the size of aggregates, the straining mechanism and gravitational sedimentation
become more effective mechanisms for retention of particles. We can conclude that straining and
gravitational sedimentation are two mechanisms that cause more irreversible retention of C60 particles.
In a previous study by Mattison et al. (2011), it has been observed that straining is a minor
mechanism for retention of MWCNT particles. The DLVO calculation using Hamaker constant for
C60 particle shows minimum importance of secondary energy minimum in C60 retention in a study by
Li et al. (2008).
The energy profile of FBP shows a very low energy barrier which indicates that if particles
have even very low energy at a distance closer than 0.91nm, an irreversible attachment can happen at
primary energy minima. Therefore only a very small portion of the retained particles remobilized
during phase 3.
32
4.5 Effect of particle shape MWCNT and C60 both are carbon-based nanoparticles with different shapes. One important purpose of
this study was to explore how mobility of MWCNT and C60 differs under similar chemical and
physical conditions. For that reason, two similar sets of column experiments were conducted using
MWCNT and C60 suspensions (Figure 17). Experiment E7 and E1 both were performed at pH of 7 and
ionic strength of 1.34mM. The resulting BTCs were similar and both relative concentration of particle
in the effluents reached the maximum value of 0.88. In experiments E8 and E9, the ionic strength was
increased to 10.89 mM and the other conditions were kept constant. The results showed slightly higher
mobility of C60 in this condition. At ionic strength 1.34 mM and 10.89 mM, the breakthrough curves
for both types of nanoparticles were very similar even though the C60 particles were slightly more
mobile than MWCNTs. At high ionic strength (60 mM), the behavior looked different. Here
MWCNTs were considerably more mobile than C60. Relative concentration of MWCNTs in effluent
reached to maximum value of 0.7 while the maximum was 0.4 for C60 particles. The shape of BTC
was also different at elevated ionic strength. The rate of increasing relative concentration was greater
for MWCNT. As the concentration of salt in the suspension increases, the difference between
breakthrough curves became more visible (Figure 18).
The mass calculations also revealed that at high ionic strength (60mM) 75.56% of retained
MWCNT particles remobilized at phase 3 while this value was only 54% for the C60 particles (table 7).
Extreme amount of irreversible attachment of C60 particles in the media suggests physical forcing
(straining) or primary energy minimum are the responsible mechanism for the retention.
Figure 17: Comparing observed BTCs of MWCNT and C60 under various ionic strengths.
0
0,4
0,8
1,2
0 1 2 3 4 5 6 7 8
C/C
0
Pore volume
C60, IS=1.341 MWCNT, IS=1.34 mMC60, IS=10.89 MWCNT, IS=10.89 mMC60, IS=60 MWCNT, IS=60 mM
33
Figure 18: Fraction of nanoparticle mass in the effluent to the total injected mass into the column at different ionic strength. The fraction is higher for C60 at lower IS and for MWCNT at higher ionic strength
Aggregation of the C60 particles in the sand column which can enhance straining mechanism
could be one of the reasons responsible for more irreversible retention of these particles. Straining can
happen only when the ratio of particle diameter to sand diameter is greater than 0.003(Bradford et al.
2002). Even though the size of MWCNTs and C60 were not measured using appropriate imaging
methods the characteristics of MWCNTs were obtained from product information given by
manufacture (25 nm diameter and 1.25 μm length) and C60 particle size was also adopted from the
published articles which used same preparation method as our study (160 nm diameter). The ratio of
particle diameter to collector diameter (d50 = 325 μm) considering length of MWCNT is 0.0038 and
considering diameter is 0.000076. The ratio is 0.00049 for C60 particles. The ratios are below or close
to the critical ratio. The initial C60 suspension used in this study (before addition of salt) was stable for
24 hours and no aggregation was observed during this period. However after addition of the salt the
particles tended to aggregate. Therefore the salt was added to the suspension just few minute before
conducting each experiment. It was observed that this instability was elevated by increasing the
amount of the added salt.
At high input concentration also MWCNTs were more mobile than C60 particles. The mass
calculation indicates that after injection of DI-water into the column, 31.82% of retained MWCNT
remobilized which is higher in comparison with of 13.95% of remobilized C60 particles. This can be
also as a result of larger aggregation and further straining.
0,0
0,2
0,4
0,6
0,8
1,0
0 10 20 30 40 50 60 70
Mass out/Mass in (mg/mg)
Ionic Strength (mM)
Average (C60)
Average (MWCNT)
34
Figure 19: Comparing observed BTCs of MWCNT, C60 and FBP (pH = 7, ionic strength=60 mM, input concentration≥100 mg/l)
The BTC of Experiments E6, E10 and E11 are illustrated by figure 19. These experiments are
conducted in same conditions (chemical and physical conditions) in order to compare transport
behavior of them. The concentration of MWCNT and C60 in the injected suspension was 100 mg/l.
Concentration of solid phase in the FBP suspention was estimated to be 3.5 g/l.
0
0,4
0,8
1,2
0 2 4 6 8 10 12
C/C
0
Pore Volume
MWCNT
C60
FBP
35
5 Conclusion The aim of this study was to investigate transport behavior of two types of carbon-based nanoparticles
(MWCNT and C60) and compare it with transport of a sample of fire born particles through saturated
porous media. Several column experiments were conducted in order to examine the effect of ionic
strength and input concentration on transport of MWCNTs and C60. All the experiments were
conducted at pH of 7, pore-water velocity of 7.5m/d and the ionic strength was varied from 1.34mM to
60 mM. The effect of input concentration was also examined by conducting experiments at 100 mg/l
as high and 7 mg/l as low input concentration.
After the comparison of the two types of carbon-based nanoparticles (C60 and MWCNT), it
was observed that at low ionic strength (1.34 and 10.89 mM) C60 was slightly more mobile than
MWCNTs. While at high ionic strength, the mobility of MWCNTs was significantly higher than that
of C60 particles. About 50% of the retained C60 particles in the column were not remobilized after
injecting of DI-water. This suggests that irreversible attachment between the nanoparticles and sand
grains became more abundant at higher ionic strength. Straining and gravitational sedimentation are
two possible mechanisms which may have been enhanced due to formation of larger aggregates of C60
nanoparticles at high ionic strength. FBP had higher mobility than engineered nanoparticles.
The experimental observations revealed that the concentration of nanoparticles had an
important role in their transport behavior. Both the mobility of the particles, as inferred from the
observed retention behavior and the shape of the BTCs were affected by changing the input
concentration. Increasing input concentration led to increase in mobility of nanoparticles (both
MWCNT and C60). The observed BTC at higher input concentration had more asymmetric shape
which can be explained by the effect of site blocking. When the concentration of particles was high,
the retention sites filled up faster and caused a high amount of retention during early times. Over time,
when the available sites became full the retention decreased and concentration of particle in the
effluent increased. In the simulations that considered CFT, the inclusion of a site blocking term and a
detachment rate produced good agreement with the observations.
More research is required to establish a quantitative relationship between the effect of ionic
strength and input concentration on transport of MWCNT and C60 through saturated porous media. It
was rather difficult to link the observed differences between transport behavior of MWCNT and C60
nanoparticles to the shape of the particles. Imaging methods may help to monitor the exact shape and
size of the nanoparticles inside the sand column, as well as in the inflow and outflow solution.
The results of this study suggest that MWCNT, C60 and also FBP are highly mobile in sandy
saturated environments. Therefore the risk of their transport to depth and contamination of aquifers is
considerable.
36
6 Acknowledgements First and foremost, I would like to thank my supervisor Dr. Prabhakar Sharma for his patient guidance,
incessant encouragement, and constructive advices and also for everything he taught me throughout
my research. I would also like to show gratitude to my subject reviewer Dr. Fritjof Fagerlund and my
examiner Dr. Roger Herbert for their valuable comments and suggestions to improve the quality of the
thesis. My special thanks to Dr. Deeksha Katyal from GGSIP University of Delhi for her nice advices
during her short visit at Uppsala University. Magnus Anderson is specifically thanked for kindly
correcting the Swedish abstract.
Getting through my thesis required more than academic support, and I have to thank my
family in Iran and my friends in Sweden for their personal and professional support during the time I
spent at Uppsala University. Most importantly I would like to thank my dear husband ‘Omid’ for his
endless love, support and faith in me. Without him, completing my master study would not been
possible.
37
7 Reference Andrievsky, G.V., Kosevich, M.V., Vovk, O.M., Shelkovsky, V.S., Vashchenko, L.A., 1995. On the
production of an aqueous colloidal solution of fullerenes. J. Chem. Soc. Chem. Commun. 1281–1282. doi:10.1039/C39950001281
B C Yadav, R.K., 2008. Structure, properties and applications of fullerenes. Int. J. Nanotechnol. Appl. 1, 15–24.
Bradford, S.A., Yates, S.R., Bettahar, M., Simunek, J., 2002. Physical factors affecting the transport and fate of colloids in saturated porous media. Water Resour. Res. 38, 1327. doi:10.1029/2002WR001340
Brant, J., L E C O A N E T, H.E.L.E.N.E., H O T Z E, M.A.T.T., 2005a. Comparison of electrokinetic properties of colloidal fullerenes (n-C60) formed using two procedures. Environ. Sci. Amp Technol. 39, 6343–51. doi:10.1021/es050090d
Brant, J., Lecoanet, H., Wiesner, M.R., 2005b. Aggregation and Deposition Characteristics of Fullerene Nanoparticles in Aqueous Systems. J. Nanoparticle Res. 7, 545–553. doi:10.1007/s11051-005-4884-8
Brant, J.A., Labille, J., Bottero, J.-Y., Wiesner, M.R., 2006. Characterizing the Impact of Preparation Method on Fullerene Cluster Structure and Chemistry. Langmuir 22, 3878–3885. doi:10.1021/la053293o
Chen, K.L., Elimelech, M., 2006. Aggregation and Deposition Kinetics of Fullerene (C60) Nanoparticles. Langmuir 22, 10994–11001. doi:10.1021/la062072v
Deguchi, S., Alargova, R.G., Tsujii, K., 2001. Stable Dispersions of Fullerenes, C60 and C70, in Water. Preparation and Characterization. Langmuir 17, 6013–6017. doi:10.1021/la010651o
Elimelech, M., Gregory, J., Jia, X., Williams, R. A., Gregory, John, Jia, Xiaodong, Williams, Richard A, 1995. Chapter 1 - Introduction, in: Elimelech, M., Gregory, J., Jia, X., Williams, R. A., Gregory, John, Jia, Xiaodong, Williams, Richard A (Eds.), Particle Deposition & Aggregation. Butterworth-Heinemann, Woburn, pp. 3–8.
Grassian, V.H., 2008. Nanoscience and Nanotechnology: Environmental and Health Impacts. John Wiley & Sons.
Gregory, J., 1981. Approximate expressions for retarded van der waals interaction. J. Colloid Interface Sci. - J COLLOID INTERFACE SCI 83, 138–145. doi:10.1016/0021-9797(81)90018-7
Huang, Q., Weber, W.J., 2004. Peroxidase-Catalyzed Coupling of Phenol in the Presence of Model Inorganic and Organic Solid Phases. Environ. Sci. Technol. 38, 5238–5245. doi:10.1021/es049826h
Iverfelt, U., 2014. Släckvattenpartiklars spridning i mark och grundvatten En studie av brandgenererade partiklars egenskaper och påverkan på föroreningsspridning.
Kasel, D., Bradford, S.A., Šimůnek, J., Heggen, M., Vereecken, H., Klumpp, E., 2013. Transport and retention of multi-walled carbon nanotubes in saturated porous media: Effects of input concentration and grain size. Water Res. 47, 933–944. doi:10.1016/j.watres.2012.11.019
Lam, C.-W., James, J.T., McCluskey, R., Hunter, R.L., 2004. Pulmonary Toxicity of Single-Wall Carbon Nanotubes in Mice 7 and 90 Days After Intratracheal Instillation. Toxicol. Sci. 77, 126–134. doi:10.1093/toxsci/kfg243
Li, Y., Wang, Y., Pennell, K.D., Abriola, L.M., 2008. Investigation of the Transport and Deposition of Fullerene (C60) Nanoparticles in Quartz Sands under Varying Flow Conditions. Environ. Sci. Technol. 42, 7174–7180. doi:10.1021/es801305y
Lind, C.J., 1970. Specific conductance as a means of estimating ionic strength. U.S. Geological Survey Professional Paper 700-D, D272–D280.
Liu, X., O’Carroll, D.M., Petersen, E.J., Huang, Q., Anderson, C.L., 2009. Mobility of Multiwalled Carbon Nanotubes in Porous Media. Environ. Sci. Technol. 43, 8153–8158. doi:10.1021/es901340d
Logan, B.E., Jewett, D.G., Arnold, R.G., Bouwer, E.J., O’Melia, C.R., 1995. Clarification of Clean-Bed Filtration Models. J. Environ. Eng. 121, 869–873. doi:10.1061/(ASCE)0733-9372(1995)121:12(869)
38
Martins-Júnior, P.A., Alcântara, C.E., Resende, R.R., Ferreira, A.J., 2013. Carbon nanotubes: directions and perspectives in oral regenerative medicine. J. Dent. Res. 92, 575–583. doi:10.1177/0022034513490957
Mattison, N.T., O’Carroll, D.M., Kerry Rowe, R., Petersen, E.J., 2011. Impact of Porous Media Grain Size on the Transport of Multi-walled Carbon Nanotubes. Environ. Sci. Technol. 45, 9765–9775. doi:10.1021/es2017076
Mekonen, A., Sharma, P., Fagerlund, F., 2014. Transport and mobilization of multiwall carbon nanotubes in quartz sand under varying saturation. Environ. Earth Sci. 71, 3751–3760. doi:10.1007/s12665-013-2769-1
Musee, N., 2011. Nanotechnology risk assessment from a waste management perspective: Are the current tools adequate? Hum. Exp. Toxicol. 30, 820–835. doi:10.1177/0960327110384525
Ngo, M.A., Smiley-Jewell, S., Aldous, P., Pinkerton, K.E., 2008. Nanomaterials and the Environment, in: Grassian, V.H. (Ed.), Nanoscience and Nanotechnology. John Wiley & Sons, Inc., pp. 1–18.
O’Carroll, D.M., Liu, X., Mattison, N.T., Petersen, E.J., 2013. Impact of diameter on carbon nanotube transport in sand. J. Colloid Interface Sci. 390, 96–104. doi:10.1016/j.jcis.2012.09.034
Pennell, K.D., Costanza, J., Wang, Y., 2008. Transport and Retention of Nanomaterials in Porous Media, in: Grassian, V.H. (Ed.), Nanoscience and Nanotechnology. John Wiley & Sons, Inc., pp. 91–106.
Saiers, J.E., Lenhart, J.J., 2003. Ionic-strength effects on colloid transport and interfacial reactions in partially saturated porous media. Water Resour. Res. 39, 1256. doi:10.1029/2002WR001887
Seymour, M.B., Chen, G., Su, C., Li, Y., 2013. Transport and Retention of Colloids in Porous Media: Does Shape Really Matter? Environ. Sci. Technol. 47, 8391–8398. doi:10.1021/es4016124
Sharma, P., Bao, D., Fagerlund, F., 2014. Deposition and mobilization of functionalized multiwall carbon nanotubes in saturated porous media: effect of grain size, flow velocity and solution chemistry. Environ. Earth Sci. 1–11. doi:10.1007/s12665-014-3208-7
Sharma, P., Flury, M., Mattson, E.D., 2008. Studying colloid transport in porous media using a geocentrifuge. Water Resour. Res. 44, W07407. doi:10.1029/2007WR006456
Tan, C.W., Tan, K.H., Ong, Y.T., Mohamed, A.R., Zein, S.H.S., Tan, S.H., 2012. Energy and environmental applications of carbon nanotubes. Environ. Chem. Lett. 10, 265–273. doi:10.1007/s10311-012-0356-4
Wang, L., Hou, L., Wang, X., Chen, W., 2014. Effects of the preparation method and humic-acid modification on the mobility and contaminant-mobilizing capability of fullerene nanoparticles (nC60). Environ. Sci. Process. Impacts. doi:10.1039/C3EM00577A
Wang, Y., 2009. Transport and retention of fullerene-based nanoparticles in water-saturated porous media [WWW Document]. URL https://smartech.gatech.edu/handle/1853/29782 (accessed 7.8.14).
Wang, Y., Li, Y., Pennell, K.D., 2008. Influence of electrolyte species and concentration on the aggregation and transport of fullerene nanoparticles in quartz sands. Environ. Toxicol. Chem. SETAC 27, 1860–1867.
Westerhoff, P., Nowack, B., 2013. Searching for Global Descriptors of Engineered Nanomaterial Fate and Transport in the Environment. Accounts Chem. Res. 46, 844–853. doi:10.1021/ar300030n
Xu, M., Eckstein, Y., 1997. Statistical Analysis of the Relationships Between Dispersivity and Other Physical Properties of Porous Media. Hydrogeol. J. 5, 4–20. doi:10.1007/s100400050254
Yang, Y., Nakada, N., Nakajima, R., Yasojima, M., Wang, C., Tanaka, H., 2013. pH, ionic strength and dissolved organic matter alter aggregation of fullerene C60 nanoparticles suspensions in wastewater. J. Hazard. Mater. 244–245, 582–587. doi:10.1016/j.jhazmat.2012.10.056
Yuan Tian, B.G., 2012. Effect of solution chemistry on multi-walled carbon nanotube deposition and mobilization in clean porous media. J. Hazard. Mater. 231-232, 79–87. doi:10.1016/j.jhazmat.2012.06.039
Zhang, L., Hou, L., Wang, L., Kan, A.T., Chen, W., Tomson, M.B., 2012. Transport of Fullerene Nanoparticles (nC60) in Saturated Sand and Sandy Soil: Controlling Factors and Modeling. Environ. Sci. Technol. 46, 7230–7238. doi:10.1021/es301234m
39
Appendix
Table A1: Chemical analysis of sample from fire location (filtered from 11 μm nylon filter) (Alcontrol laboratories, Linköping, Sweden)
Method Component concentration(μg/l) STDEV
GC/MS Acenaphthene 1.3 ±0.3
GC/MS Acenaphthylen 12 ±2.4
GC/MS Naphthalene 6.2 ±1.2
calculated PAH-L, total 20
GC/MS Anthracene 11 ±2.2
GC/MS Phenanthrene 55 ±11
GC/MS Fluoranthene 25 ±5
GC/MS Fluorene 5.7 ±1.1
GC/MS Pyrene 22 ±4.4
calculated PAH-M, total 120
GC/MS Benzo (a) anthracene 5.6 ±1.1
GC/MS Benzo (a) pyrene 3.3 ±0.7
GC/MS Benzo (b) Fluoranthene 5.5 ±1.1
GC/MS Benzo (k) Fluoranthene 1.3 ±0.3
GC/MS Benzo (ghi) Perylene 1.6 ±0.3
GC/MS Chrysene / Triphenylene 7.6 ±1.5
GC/MS Dibenzo (a, h) Anthracene 0.4 ±0.1
GC/MS Indeno (1,2,3-cd) Pyrene 1.9 ±0.4
calculated PAH-H, total 27
calculated PAH ,Carcinogenic 26
calculated PAH, Total others 140
ISO 17294 As 12 ±3.0
ISO 17294 Ba 13000 ±3300
ISO 17294 Pb 320 ±80
ISO 17294 Cd 4.9 ±1.2
ISO 17294 Co 95 ±24
ISO 17294 Cu 650 ±160
ISO 17294 Cr 140 ±35
ISO 17294 Ni 140 ±35
40
Calculation of Ionic strength of Sodium phosphate buffer solution
Molecular weight of Na=23, H=1, P=31 and O=16, therefore the molecular weight of
NaH2PO4 H2O is equal to 138 and the molecular weight of Na2HPO4.7 H2O is equal to 268. 13.8 g of
NaH2PO4 H2O has been added to 100ml DI water to prepare 1M solution A. 14.2 g of Na2HPO4 .7H2O
has been added to 100ml DI water to prepare 0.53M solution B. 42.3ml of solution A and 57.7 ml of
solution B has been mixed to prepare 100ml of 0.729M buffer solution. The concentration of NaH2PO4
H2O is 0.423M and the concentration of Na2HPO4 .7H2O is 0.306M in the new solution.
142.3100
0.423
0.5357.7100
0.306
Total concentration of sodium phosphate in DI water is 0.729M. Considering the dissociate reaction,
pH of the solution can be calculated:
H2PO4- <==> HPO4
2- + H+
; At room temperature equals to 7.21
7.21.
. ; PH=7.07
Ionic strength calculation considering following formula:
12
[Na+] = 0.306M (two sodium in Na2HPO4) and 0.423M (one in NaH2PO4).
[HPO42-] = 0.306M
[H2PO4-] = 0.423M
IS=1/2{(2(0.306) +1(0.423)) (1)2 + (0.306) (-2)2 + (0.423) (-1)2} = 1.341 M
Ionic strength of 0.729M solution is 1.341M. If the solution diluted by the factor of 1000 (0.5 ml
/500ml DI water), the IS would be 1.341mM. If diluted by the factor of 125 (4ml/500ml), the IS would
be 10.888.
41
Pump calibration
Using several nominal rate of the pump and pumping water for 10 minute, the actual rate were
calculated by measuring the weight (which is equal to volume) of the pumped water. The calibration
curve and its related equation are given by figure A1. For the experiments in this research the nominal
pump rate was selected to be 0.517ml/min in order to achieve actual rate of 1 ml/min.
Figure: A1 Pump calibration trend line and its related equation
Pipe effect correction
The volume of solution in the pipe between the fraction collector and the spectrophotometer,
cause a time shift between their measurements. Therefore the measurements of both
spectrophotometer and the fraction collector measurements have been plotted to find the time that
should be corrected in tracer test data. After plotting both data a shift of around 10 minute has been
estimated.
Figure A2: Absorption measurements of tracer test effluent, by spectrophotometer, both automatically and manually after passing a pipe to fraction collector
y = 1,9334x
0
2
4
6
8
0 1 2 3 4
Act
ual
rat
e
Nominal rate
0
0,05
0,1
0,15
0,2
0 2 4 6 8 10 12
Absorption
Pore Volume
Manually
Automatically
42
Figure A3: Matched curves of the manually and automatically measurements of tracer test after pipe correction
Table A2: Mass calculation results of all experiments (with repetitions)
Mass out/Mass in %
Experiment Total phase 1&2 phase 3 E1.1 97.12 90.36 6.76 E1.2 96.64 89.43 7.21 E2.1 95.18 84.64 10.55 E2.2 91.94 82.40 9.54 E4.1 65.10 41.36 27.74 E4.2 71.46 40.05 31.41 E4.3 66.47 42.18 24.29 E5.1 63.71 22.02 40.69 E5.2 70.58 20.19 50.39 E5.4 69.27 32.51 36.76 E5.5 76.58 32.69 43.88 E6.1 90.01 88.59 1.42 E6.2 90.16 88.36 1.81 E7.1 95.83 89.06 6.76 E7.2 95 87.59 7.41 E8.1 90.58 78.94 11.64 E8.2 94.47 80.6 13.87 E9.1 87.9 58.15 29.75 E9.2 72.65 44.38 28.27 E9.3 90.76 62.19 28.57 E10 67.16 61.87 5.29 E11 85.3 78.03 7.26
0
0,05
0,1
0,15
0,2
0 2 4 6 8 10 12
Abs
Pore Volume
Manually
Automatically
Tracer
E
and the r
been ob
dispersiv
Figur
r test
Experiment
results of thr
bserved arou
vity. The firs
re A5: Observed
0,0
0,2
0,4
0,6
0,8
1,0
0
Rel
ativ
e co
nd
uct
ivit
y
Figure A
‘E0’ has bee
ree repetition
und one por
st presence o
d and simulated
1
A4: Scanning ele
en done in o
n of the trace
re volume.
f the salt in t
d breakthrough repetiti
2 3
43
ectron microsco
order to asse
er test are giv
The simulat
the effluent w
curves for consions of the expe
4
Pore volum
ope (SEM) imag
ess water flow
ven by figure
ted result g
was at 0.83 P
servative tracer eriments
5 6
me
ge of FBP
w and disper
e A5. The firs
ives a valu
PV.
(NaCl) , E0-1,2
6 7
E0‐
E0‐
E0‐
Sim
ersivity in the
st out flow o
ue of 2.6×10
2,3 are the num
8
‐1
‐2
‐3
mulated
e column
of salt has
0-3 m for
mber of
Tidigare utgivna publikationer i serien ISSN 1650-6553 Nr 1 Geomorphological mapping and hazard assessment of alpine areas in Vorarlberg, Austria, Marcus Gustavsson Nr 2 Verification of the Turbulence Index used at SMHI, Stefan Bergman Nr 3 Forecasting the next day’s maximum and minimum temperature in Vancouver, Canada by using artificial neural network models, Magnus Nilsson
Nr 4 The tectonic history of the Skyttorp-Vattholma fault zone, south-central Sweden, Anna Victoria Engström Nr 5 Investigation on Surface energy fluxes and their relationship to synoptic weather patterns on Storglaciären, northern Sweden, Yvonne Kramer
Nr 288 Unconventional Natural Resources, Mark Wallaker, Juni 2014
Nr 289 A Lithogeochemical Study of Northern Sweden and the Kiruna and Malmberget Iron-apatite Ore Deposits, Jon Lundh, Juni 2014
Nr 290 Denitrification in a LowTemperature Bioreactor System- Laboratory column studies, Albin Nordström, Juni 2014
Nr 291 Inläckage till spillvattennät i östra Knivsta - Volymbestämning med hjälp av beräkningar, modellering och fältarbete, Johan Hamelius, Juni 2014
Nr 292 Single coccolith weight estimates using two birefringence methods on cultured Gephyrocapsa oceanica, Annalena Antonia Lochte, Juli 2014
Nr 293 Kan den hydrologiska modellen S-HYPE användas för att beräkna grundvattennivåer med tillräcklig noggrannhet?, Hanna Lagergren, September 2014
Nr 294 One- and Three-dimensional P- and S-wave Velocity Models of Central and Southern Sweden Based on SNSN Data , Ne Xun Chan, September 2014
Nr 295 Tooth Replacement of Euhelopus zdanskyi (Dinosauria: Sauropoda) and the Evolution of Titanosaurian Tooth Morphology, Seela Salakka, September 2014
Nr 296 Processing and Modeling of Gravity, Magnetic and Electromagnetic Data in the Falkenberg Area, Sweden, Soroor Mohammadi, September 2014
Nr 297 Calving Front Dynamics: External Forces that Lead to Specific Sized Calving Events, Daniel Wainwright, September 2014
Nr 298 Atmospheric Mercury Deposition in Alpine Snow, Northern British Columbia and Southwestern Yukon, Canada, Pia Karlsson, September 2014
