See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/236969360

Agglomeration and sedimentation of titanium

dioxide nanoparticles (n-TiO2) in synthetic and
real waters

Article in Journal of Nanoparticle Research · May 2013

DOI: 10.1007/s11051-013-1684-4

CITATIONS READS

39 383

4 authors:

Andrea Brunelli Giulio Pojana

Università Ca' Foscari Venezia Università Ca' Foscari Venezia
20 PUBLICATIONS 327 CITATIONS 92 PUBLICATIONS 2,874 CITATIONS

SEE PROFILE SEE PROFILE

Sarah Callegaro Antonio Marcomini

7 PUBLICATIONS 49 CITATIONS Università Ca' Foscari Venezia
325 PUBLICATIONS 7,909 CITATIONS
SEE PROFILE

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

China’s urbanization. Current trends in policy and implications for sustainability. View project

Ecological Risk Assessment of Transitional Waters View project

All content following this page was uploaded by Andrea Brunelli on 16 May 2014.

The user has requested enhancement of the downloaded file.

J Nanopart Res (2013) 15:1684
DOI 10.1007/s11051-013-1684-4
RESEARCH PAPER
Agglomeration and sedimentation of titanium dioxide
nanoparticles (n-TiO2) in synthetic and real waters
Andrea Brunelli • Giulio Pojana •
Sarah Callegaro • Antonio Marcomini
Received: 5 February 2013 / Accepted: 29 April 2013
Ó Springer Science+Business Media Dordrecht 2013
Abstract The recent detection of titanium dioxide
nanoparticles (n-TiO2) in wastewaters raised concerns
about its fate in the aquatic environment, which is
related to its mobility through water bodies. Labora-
tory experiments of n-TiO2 (particle size distribution:
10–65 nm) dispersed into both synthetic and real
aqueous solutions under environmentally realistic
concentrations (0.01, 0.1, 1 and 10 mg/l) were
conducted over a time of 50 h to mimic duration of
ecotoxicological tests. Agglomeration and sedimenta-
tion behaviour were measured under controlled con-
ditions of salinity (0–35 %), ionic composition and
strength, pH and dissolved organic carbon (DOC).
Physico-chemical parameters and particle agglomer-
ation in the dispersions were investigated by trans-
mission electron microscopy, Brunauer, Emmett and
Electronic supplementary material The online version of
this article (doi:10.1007/s11051-013-1684-4) contains
supplementary material, which is available to authorized users.
A. Brunelli  G. Pojana  S. Callegaro 
A. Marcomini (&)
Department of Environmental Sciences, Informatics and
Statistics, University Ca’ Foscari Venice, Calle Larga S.
Marta 2137, 30123 Venice, Italy
e-mail: marcom@unive.it
Present Address:
G. Pojana
Department of Philosophy and Cultural Heritage,
University Ca’ Foscari Venice, Malcanton Marcorà-
Dorsoduro 3484/D, 30123 Venice, Italy
Teller method and dynamic light scattering. A fluores-
cence spectrophotometer operating in the nephelometric
mode was employed to obtain the sedimentation rates of
n-TiO2. The overall results showed that agglomeration
and sedimentation of n-TiO2 were affected mainly by
the initial concentration. Sedimentation data fitted
satisfactorily (R2 in the range of 0.74–0.98; average
R2: 0.90) with a first-order kinetic equation.The settling
rate constant, k, increased by approx. one order of
magnitude by moving from the lowest to the highest
concentration, resulting very similar especially for all
dispersions at 1(k = 8 9 10-6 s-1) and 10 mg/l
(k = 2 9 10-5 s-1) n-TiO2, regardless the ionic
strength and composition of dispersions. The implica-
tion of these results on toxicological testing is discussed.
Keywords Titanium dioxide nanoparticles 
Agglomeration  Sedimentation  Nephelometry 
Rate constant
Introduction
Engineered nanoparticles (ENP) such as titanium
dioxide (n-TiO2) are nowadays a common component
of many industrial products and commercial formula-
tions, and their number is expected to rise further in the
near future (Chen and Mao 2007; Colvin 2003).
During the life cycle of consumer products, n-TiO2 is
discharged into wastewaters and can reach natural
123
Page 2 of 10
waters (Barrena et al. 2009; Kahru and Dubourguier
2010; Klaine et al. 2008; Labille et al. 2010). n-TiO2 in
the 20–300 nm size range was detected in freshwater
up to 10 lg/l after runoff from white-painted building
facades (Kaegi et al. 2008). Moreover, n-TiO2 was
shown to be only partly removed (ca. 42 %) from
sewage treatment plants, thus entering the natural
waterways (Gottschalk et al. 2009; Kiser et al. 2009).
An accurate assessment of n-TiO2 behaviour over the
whole lifetime under fresh and seawater-like condi-
tions is therefore necessary in order to define the
potential exposure pathways and to foresee the
expected amounts of n-TiO2 in natural waters.
Agglomeration and sedimentation are the main phys-
ico-chemical processes affecting the behaviour of
n-TiO2 in the aquatic environment and were proven to
result from the combined effects of pH, salinity, ionic
strength (IS), salt composition and concentration of
natural organic matter (Dunphy Guzman et al. 2006;
Keller et al. 2010; Phenrat et al. 2007). Agglomeration
of n-TiO2 was reported to occur under conditions of IS
(0.1 M) and pH (i.e. from 7.5 to 8.4, typical of
seawater), where n-TiO2 was close to its point of zero
charge (PZC) (Domingos et al. 2009; Jiang et al. 2009;
Proskurina et al. 2008). Naturally occurring cations
such as Ca2? and Mg2? were found to enhance
agglomeration and sedimentation of n-TiO2 in surface
and natural waters (French et al. 2009; Zhang et al.
2009), while natural organic substances such as fulvic
and humic acids stabilise the dispersions at low IS
conditions, due to the combined effects of increased
electrostatic and steric repulsions (Domingos et al.
2009; Thio et al. 2011; Zhang et al. 2009).
Investigation of the sedimentation behaviour of
inorganic ENP was carried out by using various
techniques, such as UV–vis spectrophotometry,
X-ray absorption and small angle neutron scattering
(Gustafsson et al. 2003; Jarvie et al. 2009; Phenrat et al.
2007; Tiraferri et al. 2008). Recently, the stability of
n-TiO2, n-ZnO, and n-CeO2, at relatively high concen-
tration levels (10–200 mg/l), was recorded in natural
aqueous matrices and in an artificial seawater over time
periods of up to 7 h: a mass-dependent sedimentation
process was revealed in seawater while invariable
sedimentation profiles were observed in freshwater
(Keller et al. 2010). The existing information about
sedimentation and dissolution rates of various inorganic
ENP in environmentally relevant systems has been
123
J Nanopart Res (2013) 15:1684
recently reviewed (Quik et al. 2011). The lack of data
on agglomeration and sedimentation rates of n-TiO2
dispersed in water samples under realistic concentra-
tion levels and over significatively long times was
outlined.
The aim of this work was to investigate the
agglomeration and sedimentation behaviour of com-
mercially available pristine n-TiO2 in synthetic solu-
tions mimicking fresh and seawater aqueous
composition, as well as in two salt real water samples.
Tested concentrations were 0.01, 0.1, 1 and 10 mg/l,
which include those likely to be found in natural
waters; experiments were carried out over 50 h, as
typically required by ecotoxicological tests. A neph-
elometry-based method was used to obtain sedimen-
tation data.
Materials and methods
n-TiO2 characterization
The commercial AeroxideÒ P25 Titanium Dioxide
powder (declared particle size: 21 nm), obtained from
Evonik Degussa (Essen, Germany), is a mixture of
approx. 80 % anatase and 20 % rutile. It was stored in
the dark at room temperature until use. Particle
morphology and size were examined by transmission
electron microscopy (TEM) with a Jeol 3010 (Tokyo,
Japan) operating at 300 kV. TEM samples were
prepared by deposition of a drop of n-TiO2 dispersion
in ultrahigh-purity water on a carbon-coated copper
TEM grid, followed by water evaporation at room
temperature prior to analysis. Surface area and pore
volume were obtained by the method of Brunauer,
Emmett and Teller (BET) (Brunauer et al. 1938) by
nitrogen adsorption on a Micromeritics (Norcross,
GA, USA) ASAP 2000 instrument at an adsorption
temperature of -196 °C, after pre-treating the sample
under high vacuum at 300 °C for 2 h.
Preparation of aqueous solutions and n-TiO2
dispersions
Ultrahigh-pure water (minimum resistivity: 18.2 MX
cm) was produced by a MilliQ water purifier system
(Millipore, Bedford, MA, USA). The artificial fresh and
J Nanopart Res (2013) 15:1684 Page 3 of 10

Table 1 Chemical composition and physico-chemical characteristics of the tested aqueous solutions
AFW AEW ASW1 ASW2 LW SW

Ionic composition (mg/l)

Ca2? \0.1 5.61 274.53 403.99 413.12 427.04
K? \0.1 5.86 344.06 402.32 413.70 466.97
Mg2? \0.1 18.47 1,167.13 1,296.43 1,341.07 1,405.15
Na? 54.72 156.10 10,204.80 10,922.32 11,287.76 11,647.63
Sr2? \0.1 \0.1 \0.1 6.13 \0.1 \0.1
Cl- \0.1 274.41 17,650.28 19,203.47 19,255.87 21,052.65
BO33- \0.1 0.59 21.76 28.81 54.83 71.89
Br- \0.1 0.80 \0.1 67.12 141.92 153.15
SiO32- \0.1 \0.1 \0.1 10.65 \0.1 \0.1
SO42- \0.1 46.11 2,986.54 3,227.65 2,933.41 2,987.76
HCO3- 145.22 1.83 145.20 145.20 \0.1 \0.1
pH 8.14 8.06 8.10 8.08 8.14 8.17
DOC (mg/l) \0.1 \0.1 \0.1 \0.1 1.80 2.16
Specific conductance at 22 °C (mS) 0.213 0.671 42.8 47.3 41.8 47.8
Salinity (%) \1 \1 33 35 32 35
IS (mmol/l) 2.0 9.8 630 690 932 984
Divalent cation concentration (mmol/l) \0.1 1.0 55 63 92 102
AFW artificial freshwater, AEW artificial estuarine water, ASW1 artificial seawater 1, ASW2 artificial seawater 2, LW lagoon water,
SW seawater, IS Ionic strength

seawater solutions (Allen et al. 2010) were prepared by
dissolving analytical reagent grade inorganic salts from
Sigma-Aldrich (St. Louis, MI, USA) in MilliQ water by
magnetic stirring; the solutions were then aerated
overnight with filtered laboratory air in order to guaran-
tee air saturation (pH 8.2 ± 0.2). A common growth
medium for freshwater crustaceans (e.g., Daphnia
magna) was prepared according to ASTM E729-E796
protocol (ASTM E729-96, 2004 (2007)) and used as
synthetic freshwater (AFW); two artificial seawater
solutions (ASW1, ASW2), usually employed as culture
media for marine algae and marine invertebrates such as
Phaeodactylum tricornutum and Artemia franciscana,
respectively, were prepared according to ASTM E724-
E798 protocol (ASTM E724-98, 2004 (2012)) and
utilized as synthetic seawaters: ASW1 and ASW2
slightly differed in the ionic composition, salinity and
IS. The synthetic solution AEW (0.5 % salinity) was
instead obtained by appropriate dilution with MilliQ
water of the salt solution ASW2. Samples of natural
waters were collected in the Venice Lagoon (Italy)
(lagoon water, LW) and in the Northern Adriatic Sea,
6 km far from the coast (seawater, SW), respectively. All
solutions were stored at 4 °C in the dark and filtered at
0.22 lm before use. The solution salinity of natural salt
waters was measured at 22 °C with a salinometer PCE-
0100 (Atago, Tokyo, Japan), while their solutions
specific conductance was determined at 22 °C with an
electrical conductivity meter HD 8706 (DeltaOhm,
Padua, Italy). The organic carbon content was deter-
mined by using a Total Organic Carbon analyser TOC-
5050A (Shimadzu, Kyoto, Japan). Samples were previ-
ously acidified with H3PO4 (25 % w/w) and the
inorganic carbon was then removed by heating and
ventilation for 5 min (Nieuwenhuize et al. 1994).
Compositions and physico-chemical parameters of
prepared aqueous solutions are reported in Table 1.
The salinity value of the artificial solution AEW was
selected on the basis of the thalassic series, which
identifies 0.5 % salinity as one of the boundary values of
oligohaline waters (Mantyla 1987). The AFW solution
contained only 0.2 g/l of the naturally occurring pH
buffer NaHCO3 and represented the electrolyte back-
ground reference.
The n-TiO2 aqueous dispersions were obtained by
probe sonication (20 min at 100 W) of pristine powder
(i.e. as received from the supplier) in 500 ml of the
selected aqueous solutions by means of an UP-200H

123
Page 4 of 10
Hielscher (Hielscher Ultrasonics GmbH, Teltow, Ger-
many) ultrasonic probe. After sonication, 3 ml of each
dispersion was placed in a 1 cm quartz cuvette ready to
be analysed.
Agglomeration and sedimentation experiments
The sedimentation experiments were carried out by
using a fluorescence spectrophotometer LS55 (Perkin
Elmer, Waltham, MA, USA) employed as nephelom-
eter, by setting kex = kem (Lawler and Townshend
2005), which measures the light intensity scattered by
the dispersed TiO2. Unlike UV measurements, the light
scattered by particles is measured at 90° with respect to
the light source. This technique is appropriate when:
(1) sample contains few scattering particles because
the intensity of the transmitted radiation is very similar
to the intensity of the radiation source; (2) the particles
into the sample are small enough to observe the
scattered radiation, causing therefore Rayleigh scatter-
ing (Quik et al. 2012; Sene et al. 2009).
The kex wavelength was optimized at 625 nm in
order to obtain best signal/noise (S/N) ratio. Emission
filter opening, excitation and emission slit widths were
2.5 and 20 nm, respectively, adjusted according to the
investigated n-TiO2 initial concentrations. The back-
ground signal of each aqueous dispersion was first
acquired (i.e. before adding n-TiO2) and instrumen-
tally subtracted from each subsequent measurement.
Linear correlations between scattered light intensity
and concentration of dispersed n-TiO2 were obtained
for each sample, exhibiting R2 [ 0.997 in all cases.
The sedimentation process was monitored (30 s mea-
surement every 15 min) for 50 h immediately after
preparation by keeping the cuvette inside the fluores-
cence spectrophotometer during the whole analysis
time in order to avoid any perturbation. The tested
n-TiO2 concentrations were 0.01, 0.1, 1, and 10 mg/l
for all samples. All experiments were conducted at
22 °C. Each analysis was performed in triplicate, and
reported results were the average of the intensity of the
scattered radiation from the dispersed n-TiO2. Relative
standard deviation among replicates was \7 % in all
experiments. Sedimentation profiles were obtained as
relative variation with time of the scattered light
normalised with respect to the initial scattering value
(I/I0). As far as the light intensity is directly propor-
tional to the dispersed particle concentration, the
sedimentation profile was obtained by reporting
123
J Nanopart Res (2013) 15:1684
C/C0, where C is the concentration at a defined time
and C0 is the initial concentration.
The size distribution (duplicate determination) of
n-TiO2 in the aqueous dispersions was measured by
dynamic light scattering (DLS) at 0.2, 25 and 50 h from
start by means of a Nicomp Submicron Particle Sizer
AutodiluteÒ Model 370 (Particle Sizing Systems, Port
Richey, FL, USA), with a 632.8 nm laser wavelength at a
detector set at 90°. The employed instrument can
automatically recognize, through a patented software
algorithm, a three modal size distribution of particles in
the 0.5–6,000 nm range. The size distributions of n-TiO2
dispersions at 0.01 mg/l could not be determined due to
the very low scattering signal exhibited. Relative standard
deviation among replicates was\13 % in all experiments.
Data processing and analysis
In order to describe the sedimentation behaviour of n-
TiO2 in examined aqueous dispersions, the sedimen-
tation rates of n-TiO2 were calculated according to a
first-order decay equation, that was applied to fit the
obtained sedimentation profiles, as reported in recent
literature describing particle settling in the environ-
ment (Joss et al. 2006; Keller et al. 2010; Phenrat et al.
2007; Quik et al. 2011, 2012; Starchevskii et al. 2009).
Considering the first-order linear equation below:
dC
¼ kC ð1Þ
dt
where C is concentration at time t and k is first-order
rate constant,
Equation 1 can be analytically integrated as follow:
dC
¼ k dt ð2Þ
C
ln C ¼ kt þ p ð3Þ
where p is a constant.
By introducing the initial condition C(t0) = C0 and
setting t0 = 0, the following results:
C
ln ¼ kt ð4Þ
C0
ln C ¼ ln C0  kt ð5Þ
C ¼ C0 ekt ð6Þ
From Eq. 4, the first-order rate constant (k) was
calculated as follow:
J Nanopart Res (2013) 15:1684
Fig. 1 Representative TEM image of AeroxideÒ n-TiO2 P25
 
ln CC0
k¼ ð7Þ
t
Results and discussion
Characterization of pristine n-TiO2
The AeroxideÒ n-TiO2 P25 pristine sample was
characterized by TEM and BET. Figure 1 shows a
TEM image of the n-TiO2 powder, as received from
the supplier, which exhibited a size distribution
ranging approx. from 10 to 65 nm, and a shape partly
irregular and semi-spherical. The main crystallo-
graphic phases were confirmed to be anatase and
rutile (4:1 ratio), according to the manufacturer. BET
analysis allowed to obtain a specific surface area of
61 m2/g, a pore size of 0.5 ml/g and a bimodal pore
size distribution in the 2–4 and 10–90 nm size range,
respectively. According to these results, the selected
n-TiO2 sample can be classified as mesoporous NP.
Characterization of aqueous solutions
Table 1 shows the ionic composition, pH, dissolved
organic carbon (DOC), conductance, salinity, ionic
strength (IS) and divalent cation concentration of the
aqueous solutions before dispersing the n-TiO2
powder. Samples AFW, AEW, ASW1, and ASW2
were artificial and prepared according to standard
Page 5 of 10
protocols for ecotoxicological testing (ASTM E729-
96 for the freshwater sample AFW, and ASTM E724-
98 for the seawater samples ASW1 and ASW2, and the
estuarine water sample AEW). LW and SW were real
samples collected from the lagoon (LW) of Venice and
the Northern Adriatic Sea (SW).
Sodium bicarbonate, a naturally occurring buffer
commonly added to culture media for ecotoxicological
studies, was added to all artificial solutions in order to
ensure the pH stability in the 8.0–8.2 range. DOC was
negligible (\0.2 mM) in all artificial samples thus
preventing stabilization of the dispersed n-TiO2 by
interaction with organic matter. Ionic composition, IS,
specific conductance, and divalent cation concentration
were already reported to greatly affect n-TiO2 stability
in solution (French et al. 2009). These parameters varied
widely in the prepared dispersions (Table 1) and can be
regarded as the key variables affecting the n-TiO2
behaviour under the tested physico-chemical conditions.
The artificial estuarine like sample, AEW, exhibited
values of most parameters very similar to those of the
artificial freshwater, AFW.
Agglomeration
The results of the n-TiO2 agglomeration recorded at
selected times (0.2, 25 and 50 h) of each experiment
are summarized in Table 2 where DLS measurements,
i.e. the average hydrodynamic diameters of agglom-
erates, are reported. The most diluted dispersions, i.e.
0.01 mg/l, could not be investigated by the available
DLS because of its sensitivity (too low signal, i.e.
\60 kHz). From Table 2, the first DLS measurement
(after 0.2 h, i.e. 12 min, the start of each experiment)
revealed the formation of agglomerates regardless of
the initial concentrations of n-TiO2. The average size
of agglomerates at 0.2 h was found in the range
164–405 nm for the artificial samples (i.e. AFW,
AEW, ASW1, ASW2), and 492 and 598 nm for the
real samples (i.e. LW and SW), respectively. Then, a
general growth of the agglomerates size, up to
800–880 lm, was recorded at 25 h, with some excep-
tions (e.g. AFW at 1.0 mg/l and LW at 0.1 mg/l). The
average sizes of agglomerates after 50 h were both
higher or lower, as well as similar, to those at 25 h. In
detail, AFW (i.e. the n-TiO2 freshwater dispersion)
was the only aqueous dispersion presenting stable
agglomerates (around 200 nm size) throughout the
entire experiment time (i.e. 50 h), both at 0.1 and
123
Page 6 of 10 J Nanopart Res (2013) 15:1684

Table 2 Average DLS hydrodynamic diameters of n-TiO2 dispersed in the tested aqueous solutions at different concentration levels
and times
n-TiO2 (mg/l) Time (h) AFW AEW ASW1 ASW2 LW SW

0.1 0.2 221 ± 3 164 ± 15 180 ± 21 297 ± 47 262 ± 32 229 ± 25

25 290 ± 15 220 ± 18 207 ± 26 381 ± 90 156 ± 9 448 ± 55
50 251 ± 4 270 ± 10 304 ± 38 246 ± 16 161 ± 8 315 ± 41
1.0 0.2 294 ± 37 194 ± 11 161 ± 19 350 ± 41 331 ± 71 375 ± 31
25 225 ± 8 607 ± 10 799 ± 120 466 ± 9 827 ± 166 610 ± 52
50 228 ± 15 663 ± 69 738 ± 84 519 ± 59 514 ± 63 512 ± 51
10 25
0.2 627
290 ± 25
19 226
271 ± 14
34 531
505 ± 51
24 883
401 ± 114
42 608
492 ± 43
53 598
720 ± 41
62
50 548 ± 29 211 ± 11 344 ± 30 1137 ± 48 419 ± 8 588 ± 64
AFW artificial freshwater, AEW artificial estuarine water, ASW1 artificial seawater 1, ASW2 artificial seawater 2, LW lagoon water,
SW seawater

1 mg/l. The values of n-TiO2 agglomerates for salt

solutions after 0.2 h at 10 mg/l were higher (up to
598 ± 41 nm recorded for SW) compared to those at
lower initial concentrations. Over the experiment time
course, the average sizes of agglomerates became
generally larger, reaching 1,137 ± 48 nm for ASW2
at 50 h. The smaller agglomerates of n-TiO2, recorded
in the freshwater dispersion, compared to the other
dispersions, are likely related to the lower salinity and
divalent cation concentration, thus decreasing the
probability of interaction between n-TiO2 and the
surrounding ions.
The overall DLS data of Table 2 indicate that the
agglomeration behaviour of n-TiO2 is affected pri-
marily by the initial concentration and, secondly, by
the time duration of each experiment. While the
average size of n-TiO2 agglomerates generally
increased with time over the initial hours (up to
25 h), the average size of agglomerates of most Fig. 2 Sedimentation profiles of n-TiO2 dispersed into AFW
(synthetic freshwater sample) at the tested concentrations
samples was found at 50 h similar or lower than those
at 25 h. This experimental finding could be related to
the concurrent sedimentation of larger agglomerates,
as confirmed by sedimentation profiles, with conse- freshwater sample (AFW), one synthetic seawater
quent decrease of the average size of agglomerates, sample (ASW1) and the real seawater sample (SW),
due to reduced residual concentrations of particles in by plotting ln(C/C0) versus time. The sedimentation
the aqueous solution. The occurrence of agglomerates profiles of all other tested aqueous dispersions, for
up to 1,500 nm was already observed for river and lake each n-TiO2 concentration tested, are provided in
water samples (Zhang et al. 2009). Online Resource.
The overall sedimentation profiles fitted well with a
Sedimentation first-order kinetic (Eq. 6), and the rate constant k (i.e. the
slope of the graph) for each dispersion was calculated
Figures 2, 3 and 4 display the sedimentation profiles according to Eq. 7. The k values and the correlation
of three representative samples, i.e. the synthetic coefficients (R2) are summarized in Table 3. R2 resulted

123
J Nanopart Res (2013) 15:1684 Page 7 of 10

Table 3 First-order fitting equation parameters for all the

dispersions, calculated from the experimental sedimentation
profiles
Sample n-TiO2 k (s-1) R2
concentration
(mg/l)

AFW 0.01 3 9 10-6 0.981

a
0.1 1 9 10-5 0.932
1 5 9 10-6 0.935
10b 7 9 10-6 0.904
b
AEW 0.01 9 9 10-6 0.904
0.1 5 9 10-6 0.750
1 8 9 10-6 0.989
10 2 9 10-5 0.931
ASW1 0.01 4 9 10-6 0.778
-6
0.1 5 9 10 0.773
1 8 9 10-6 0.986
10 2 9 10-5 0.979
Fig. 3 Sedimentation profiles of n-TiO2 dispersed in ASW1 ASW2 0.01b 7 9 10-6 0.909
(synthetic seawater sample) at the tested concentrations 0.1 4 9 10-6 0.891
-6
1 8 9 10 0.965
10 2 9 10-5 0.978
LW 0.01 3 9 10-6 0.860
0.1 4 9 10-6 0.741
1 1 9 10-5 0.810
10 2 9 10-5 0.926
SW 0.01 3 9 10-6 0.864
0.1 4 9 10-6 0.760
1 8 9 10-6 0.981
10 2 9 10-5 0.983
AFW artificial freshwater, AEW artificial estuarine water,
ASW1 artificial seawater 1, ASW2 artificial seawater 2
a
k and R2 were calculated 5 h from the beginning because of
the non-linear behaviour observed
b
k and R2 were calculated 10 h from the beginning because of
the non-linear behaviour observed

of the straight line behaviour observed. The overall

Fig. 4 Sedimentation profiles of n-TiO2 dispersed in SW (real
seawater sample) at different concentration levels tested
satisfactory for all the tested dispersions, except for
ASW2 and AEW at 0.01 mg/l and AFW at 0.1 and
10 mg/l, respectively, for which R2 was calculated 10 or
5 h from the beginning of the experiment, as a function
results suggest that the observed sedimentation behav-
iour of n-TiO2 in the examined dispersions obeys the
first-order kinetic regardless of the dispersion compo-
sition (i.e. artificial/real seawater or artificial freshwater)
when the initial amount of NP is[0.1 mg/l. The k values
(Table 3) covered approximatively one order of mag-
nitude, ranging from 3 9 10-6 to 2 9 10-5 s-1,
increasing with the increasing of concentration.
From Figs. 2, 3, 4 and Table 3, it follows that there
were small differences among sedimentation rates of

123
Page 8 of 10 J Nanopart Res (2013) 15:1684

all dispersions at the lowest dispersion concentrations,
i.e. 0.01 and 0.1 mg/l. The settled n-TiO2 observed
after 50 h ranged from 33.5 to 52.2 % of the initial
concentrations (Figures S1 and S2, Online Resource).
With regard to the 1 mg/l dispersion concentration, an
intermediate sedimentation profile was observed for
both artificial and real seawater dispersions (Figs. 3, 4)
and for the estuarine like artificial dispersion (Fig. S3,
Online Resource), while the artificial freshwater
dispersion showed a sedimentation profile very similar
to that of lower concentrations (Fig. 2). The settled
n-TiO2 at the end of the experiments (i.e. 50 h) at 1 mg/
l was in the range of 53–79 % initial concentration.
Taking into account the highest initial concentration of
10 mg/l, the n-TiO2 sedimentation behaviour was
basically the same for all the aqueous dispersions
(Fig. S4, Online Resource), and the settled n-TiO2 in all
dispersions reached very similar values after 50 h,
ranging from 94.5 to 98.3 % initial concentration. A
similar behaviour was already reported for n-TiO2 in
real environmental waters but only at much higher
concentration (200 mg/l) over a quite shorter time
(400 min), approximately 80 % being settled after 2 h
(Keller et al. 2010).
As far as the AFW sedimentation profiles (Fig. 2)
are concerned, it can be observed that they were quite
similar for all dispersion concentrations, except at
10 mg/l. At this concentration, three significantly

Table 4 Actual dispersed n-TiO2 concentrations (mg/l) at 1, 5, 10, 15, 25 and 50 h from the start of the experiments
1h 5h 10 h 15 h 25 h 50 h

0.01 mg/l
AFW 0.00974 0.00955 0.00884 0.00836 0.00755 0.00580
AEW 0.00974 0.00806 0.00755 0.00690 0.00614 0.00554
ASW1 0.00989 0.00861 0.00809 0.00721 0.00641 0.00571
ASW2 0.00939 0.00863 0.00784 0.00746 0.00734 0.00665
LW 0.00963 0.00891 0.00835 0.00784 0.00685 0.00594
SW 0.00972 0.00893 0.00859 0.00808 0.00691 0.00591
0.1 mg/l
AFW 0.096 0.081 0.076 0.068 0.067 0.061
AEW 0.094 0.083 0.070 0.065 0.060 0.048
ASW1 0.092 0.081 0.075 0.071 0.058 0.048
ASW2 0.096 0.085 0.079 0.076 0.063 0.048
LW 0.094 0.082 0.071 0.069 0.062 0.050
SW 0.097 0.089 0.078 0.074 0.065 0.052
1 mg/l
AFW 0.954 0.895 0.836 0.717 0.613 0.470
AEW 0.937 0.808 0.736 0.687 0.476 0.235
ASW1 0.959 0.846 0.777 0.696 0.518 0.237
ASW2 0.919 0.812 0.728 0.637 0.453 0.244
LW 0.827 0.684 0.578 0.400 0.383 0.208
SW 0.918 0.795 0.728 0.663 0.465 0.246
10 mg/l
AFW 9.42 9.06 7.36 3.15 1.66 0.55
AEW 9.58 8.73 6.36 2.60 1.17 0.48
ASW1 9.66 8.44 5.68 3.34 0.87 0.17
ASW2 9.59 8.30 6.01 4.01 1.46 0.37
LW 9.72 7.95 3.88 2.22 1.41 0.51
SW 9.02 6.50 4.13 2.77 1.26 0.30
AFW artificial freshwater, AEW artificial estuarine water, ASW1 artificial seawater 1, ASW2 artificial seawater 2

123
J Nanopart Res (2013) 15:1684
different settling rates were recognized: one from 0 to
ca. 7 h, one from ca. 7 to ca. 17 h and the last one from
ca. 17 to the end of the experiment, which resulted in
an overall sigma shaped curve. This would suggest the
slow formation of agglomerates, which, upon achiev-
ing a critical size, settled more rapidly than agglom-
eration, i.e. the agglomeration is the limiting step for
sedimentation. Considering the artificial sea sample
ASW1 (Fig. 3), the sedimentation profiles were very
similar to those of the real seawater sample SW
(Fig. 4). This similarity suggests that the sedimenta-
tion rate recorded in this investigation is not signif-
icatively affected by organic matter, because DOC
was\0.1 mg/l in the ASW1 dispersion and 2.16 mg/l
in the SW dispersion, respectively.
The actual dispersed n-TiO2 concentrations at 1, 5,
10, 15, 25 and 50 h were calculated from sedimentation
data and reported in Table 4. These data show that the
actual concentration of dispersed n-TiO2 is by far lower
than the nominal concentration especially at the highest
dispersion concentrations, i.e. 1 and 10 mg/l. Moreover,
the final (i.e. at 50 h) average concentration of all
dispersions was 60 % of the initial amount of n-TiO2 at
0.01 mg/l, 51 % at 0.1 mg/l and to 27 % at 1 mg/l,
reaching finally 4 % at 10 mg/l. It follows that, for
instance, the difference between the actual and final
concentrations of the n-TiO2 sea water dispersions of 0.1
and 10 mg/l, respectively, was only a factor 6 instead of
a factor 100. These results may have quite significant
implications on the evaluation of n-TiO2 toxicity if the
actual exposure concentration is not estimated.
Conclusions
Agglomeration and sedimentation are known to affect
the environmental behaviour of engineered nanopar-
ticles in natural waters. In this work, we proved it
quantitatively under conditions relevant for ecotoxi-
cological testing, by using both synthetic and real
seawater, as well as synthetic freshwater. The adopted
procedure for investigating sedimentation, based on
nephelometry, provided consistent and reproducible
results. The tested n-TiO2 was found to undergo a fast
agglomeration as soon as it was dispersed in both
synthetic and real aqueous dispersions, while sedi-
mentation was slower. The extent and rate of sedimen-
tation, over the tested 50 h, increased remarkably with
increase in the initial n-TiO2 concentration, thus
Page 9 of 10
decreasing the stability over time of n-TiO2 dispersions
moving from 0.1 to 10 mg/l. Both agglomeration and
sedimentation were observed to depend mostly on n-
TiO2 initial concentration and, to a minor extent, on
salt content, ionic strength, and DOC, under the
investigated conditions. The reported kinetic results
allow estimating the residual concentration of dis-
persed n-TiO2, thus providing the actual exposure to
n-TiO2 in ecotoxicological tests, as well as for modelling
mobility and fate of n-TiO2 in the aquatic environment.
Acknowledgments This research was partly supported by
Veneto Nanotech (Padua, Italy) through a Ph.D. sponsorship
programme. The authors thankfully acknowledge Riccardo
Cossi (Qi srl, Pomezia, Italy) for the valuable technical support.
The support of Michele Gallo, Davide Marchetto, Dagmar
Bilaničová and Alessandra Moccia (University Ca’ Foscari
Venice) for sedimentation experiments, DLS analysis and
mathematical treatment of experimental data, respectively, is
also gratefully acknowledged.
References
Allen HJ, Impellitteri CA, Macke DA, Heckman JL, Poynton
HC, Lazorchak JM, Govindaswamy S, Roose DL, Nad-
agouda MN (2010) Effects from filtration, capping agents,
and presence/absence of food on the toxicity of silver
nanoparticles to Daphnia magna. Environ Toxicol Chem
29:2742–2750. doi:10.1002/etc.329
ASTM E729-96, 2004 (2007) Standard guide for conducting
acute toxicity testing on test materials with fishes, macro-
invertebrates, and amphibians. ASTM International, West
Conshohocken, PA, 2007. doi:10.1520/E0729-96R07
ASTM E724-98, 2004 (2012) Standard guide for conducting static
acute toxicity tests starting with embryos of four species of
saltwater bivalve molluscs. ASTM International, West
Conshohocken, PA, 2012. doi:10.1520/E0724-98R12
Barrena R, Casals E, Colón J, Font X, Sánchez A, Puntes V
(2009) Evaluation of the ecotoxicity of model nanoparti-
cles. Chemosphere 75:850–857. doi:10.1016/j.chemos
phere.2009.01.078
Brunauer S, Emmett PH, Teller E (1938) Adsorption of gases in
multimolecular layers. Am Chem Soc 60:309–319. doi:
10.1021/ja01269a023
Chen X, Mao SS (2007) Titanium dioxide nanomaterials: syn-
thesis, properties, modifications, and applications. Chem
Rev 107:2891–2959. doi:10.1002/chin.200741216
Colvin VL (2003) The potential environmental impact of
engineered nanomaterials. Nat Biotechnol 21:1166–1170.
doi:10.1038/nbt0604-760c
Domingos RF, Tufenkji N, Wilkinson KJ (2009) Aggregation of
titanium dioxide nanoparticles: role of a fulvic acid. Environ
Sci Technol 43:1282–1286. doi:10.1021/es8023594
Dunphy Guzman KA, Finnegan MP, Banfield JF (2006) Influ-
ence of surface potential on aggregation and transport of
123
Page 10 of 10
titania nanoparticles. Environ Sci Technol 40:7688–7693.
doi:10.1021/es060847g
French RA, Jacobson AR, Kim B, Isley SL, Penn RL, Baveye
PC (2009) Influence of ionic strength, pH, and cation
valence on aggregation kinetics of titanium dioxide nano-
particles. Environ Sci Technol 43:1354–1359. doi:
10.1021/es802628n
Gottschalk F, Sonderer T, Scholz RW, Nowack B (2009)
Modeled environmental concentrations of engineered
nanomaterials (TiO2, ZnO, Ag, CNT, fullerenes) for dif-
ferent regions. Environ Sci Technol 43:9216–9222. doi:
10.1021/es9015553
Gustafsson J, Nordenswan E, Rosenholm JB (2003) Consolidation
behavior in sedimentation of TiO2 suspensions in the pres-
ence of electrolytes. J Colloid Interf Sci 258:235–243. doi:
10.1016/s0021-9797(02)00177-7
Jarvie HP, Al-Obaidi H, King SM, Bowes MJ, Lawrence MJ,
Drake AF, Green MA, Dobson PJ (2009) Fate of silica
nanoparticles in simulated primary wastewater treatment.
Environ Sci Technol 43:8622–8628. doi:10.1021/es901399q
Jiang JK, Oberdorster G, Biswas P (2009) Characterization of
size, surface charge, and agglomeration state of nanopar-
ticle dispersions for toxicological studies. J Nanopart Res
11:77–89. doi:10.1007/s11051-008-9446-4
Joss A, Zabczynski S, Göbel A, Hoffmann B, Löffler D,
McArdell CS, Ternes TA, Thomsen A, Siegrist H (2006)
Biological degradation of pharmaceuticals in municipal
wastewater treatment: proposing a classification scheme.
Water Res 40:1686–1696. doi:10.1016/j.watres.2006.
02.014
Kaegi R, Ulrich A, Sinnet B, Vonbank R, Wichser A, Zuleeg S,
Simmler H, Brunner S, Vonmont H, Burkhardt M, Boller
M (2008) Synthetic TiO2 nanoparticle emission from
exterior facades into the aquatic environment. Environ
Pollut 156:233–239. doi:10.1016/j.envpol.2008.08.004
Kahru A, Dubourguier HC (2010) From ecotoxicology to
nanoecotoxicology. Toxicology 269:105–119. doi:
10.1016/j.tox.2009.08.016
Keller AA, Wang H, Zhou D, Lenihan HS, Cherr G, Cardinale
BJ, Miller R, Ji Z (2010) Stability and aggregation of metal
oxide nanoparticles in natural aqueous matrices. Environ
Sci Technol 44:1962–1967. doi:10.1021/es902987d
Kiser MA, Westerhoff P, Benn T, Wang Y, Perez-Rivera J,
Hristovski K (2009) Titanium nanomaterial removal and
release from wastewater treatment plants. Environ Sci
Technol 43:6757–6763. doi:10.1021/es901102n
Klaine SJ, Alvarez PJJ, Batley GE, Fernandes TF, Handy RD,
Lyon DY, Mahendra S, McLaughlin MJ, Lead JR (2008)
Nanomaterials in the environment: behavior, fate, bio-
availability, and effects. Environ Toxicol Chem 27:1825–
1851. doi:10.1897/08-090.1
123
View publication stats
J Nanopart Res (2013) 15:1684
Labille J, Feng J, Botta C, Borschneck D, Sammut M, Cabie M,
Auffan M, Rose J, Bottero J-Y (2010) Aging of TiO2
nanocomposites used in sunscreen. Dispersion and fate of
the degradation products in aqueous environment. Environ
Pollut 158:3482–3489. doi:10.1016/j.envpol.2010.02.012
Lawler DM, Townshend A (2005) Spectrophotometry: turbi-
dimetry and nephelometry. In: Encyclopaedia of Analyti-
cal Science, pp 343–351
Mantyla AW (1987) Standard seawater comparisons updated.
J Phys Oceanogr 17:543–548. doi:10.1175/1520-0485
(1987)017\0543:sscu[2.0.co;2
Nieuwenhuize J, Maas YEM, Middelburg JJ (1994) Rapid analysis
of organic carbon and nitrogen in particulate materials. Mar
Chem 45:217–224. doi:10.1016/0304-4203(94)90005-1
Phenrat T, Saleh N, Sirk K, Tilton RD, Lowry GV (2007)
Aggregation and sedimentation of aqueous nanoscale zero-
valent iron dispersions. Environ Sci Technol 41:284–290. doi:
10.1021/es061349a
Proskurina VE, Chichkanov SV, Myagchenkov VA (2008)
Kinetics of flocculation of the suspension of titanium
dioxide in water–salt media in the presence of cationic
SAS. J Water Chem Technol 30:112–120. doi:10.3103/
s1063455x08020082
Quik JTK, Vonk JA, Hansen SF, Baun A, Van De Meent D
(2011) How to assess exposure of aquatic organisms to
manufactured nanoparticles? Environ Int 37:1068–1077.
doi:10.1002/etc.1783
Quik JTK, Stuart MC, Wouterse M, Peijnenburg W, Hendrinks AJ,
Van De Meent D (2012) Natural colloids are the dominant
factor in the sedimentation of nanoparticles. Environ Toxicol
Chem 31:1019–1022. doi:10.1016/j.envint.2011.01.015
Sene JA, Pinheiro MVB, Krambrock K, Barbeira PJS (2009)
Quantification of fullerene nanoparticles suspensions in
water based on optical scattering. Talanta 78:1503–1507.
doi:10.1016/j.talanta.2009.02.029
Starchevskii VL, Kislenko VM, Maksymiv NL, Koval IZ (2009)
Variation kinetics of chemical and bacterial contamina-
tions of water containing yeast cells. Biol Method Water
Treat 31:469–477. doi:10.3103/s1063455x09040109
Thio BJR, Zhou D, Keller AA (2011) Influence of natural
organic matter on the aggregation and deposition of tita-
nium dioxide nanoparticles. J Hard Mater 189:556–563.
doi:10.1016/j.jhazmat.2011.02.072
Tiraferri A, Chen KL, Sethi R, Elimelech M (2008) Reduced
aggregation and sedimentation of zero-valent iron nano-
particles in the presence of guar gum. J Colloid Interf Sci
324:71–79. doi:10.1016/j.jcis.2008.04.064
Zhang Y, Chen Y, Westerhoff P, Crittenden J (2009) Impact of
natural organic matter and divalent cations on the stability
of aqueous nanoparticles. Water Res 43:4249–4257. doi:
10.1016/j.watres.2009.06.005