Colloids and Surfaces A: Physicochem. Eng.

Aspects 516 (2017) 296–304

Contents lists available at ScienceDirect

Colloids and Surfaces A: Physicochemical and

Engineering Aspects
journal homepage: www.elsevier.com/locate/colsurfa

Effect of Triton X-100 on the stability of titania nanoparticles against

agglomeration and sedimentation: A masked depletion interaction
Yung-Jih Yang, David S. Corti, Elias I. Franses ∗
Davidson School of Chemical Engineering, Purdue University, 480 Stadium Mall Drive, West Lafayette, IN 47907, USA

h i g h l i g h t s g r a p h i c a l a b s t r a c t

• Nonionic surfactant micelles intro-

duce significant depletion forces
causing flocculation of titania
particles.
• Micellar depletion effects can be
masked by the higher solution viscos-
ity caused by the micelles.
• Depletion effects cause fast floccula-
tion and then faster sedimentation of
dense particles.
• A new model of agglomeration and
sedimentation establishes the basis
of the hidden depletion effect.

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

Article history: The sedimentation half-times ts of suspensions of polydisperse titania particles in aqueous Triton X-100
Received 17 October 2016 (TX100) solutions with concentrations cTX100 ranging from 0.05 to 250 mM were measured. Three regions
Received in revised form of TX100 concentrations with different settling behavior were observed. In region A, or for cTX100 below
12 December 2016
its cmc (ca. 0.24 mM in water at 25 ◦ C), the ts -values increase with increasing surfactant concentration. In
Accepted 15 December 2016
region B, or for cmc ≤ cTX100 < c ∗∗ , where c ∗∗ was determined to be 30 mM experimentally, the ts -values do
Available online 19 December 2016
not vary with the surfactant concentration. In region C, or for cTX100 ≥ c ∗∗ , as cTX100 increases, the ts -values
increase. In regions A and B, little or no viscosity effect on the sedimentation was observed. By contrast, in
Keywords:
Particle sedimentation
region C, the viscosity of the micellar solution increases with increasing cTX100 and accounts partially for
Depletion effect the slower particle sedimentation rate. The surface densities of adsorbed TX100 molecules on the particle
Triton X-100 micelles surfaces were measured to probe their effects on the steric interactions, which are responsible for the
Titania increase in the ts -values in region A. The maximum adsorbed densities on the particle surfaces, which are
Viscosity presumed to be hydrophilic, were found to be much smaller than those on other hydrophobic surfaces,
as expected. Nonetheless, the adsorbed densities are high enough to induce strong steric interactions in
region B, preventing the particles from agglomerating. In region C, the increase in viscosity slows down
not only the particle sedimentation, but also the particle agglomeration, resulting in an apparently weaker
micellar depletion effect. Our inference is that depletion-induced flocculation in region C is masked by the
high solution viscosity. This inference is supported by several different lines of evidence. First, the once
corrected half-times, tsI , which are the measured half-times divided by the relative viscosities, start to

∗ Corresponding author.
E-mail address: franses@purdue.edu (E.I. Franses).

http://dx.doi.org/10.1016/j.colsurfa.2016.12.026
0927-7757/© 2016 Elsevier B.V. All rights reserved.
 Y.-J. Yang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 516 (2017) 296–304 297

decrease at cTX100 ∼
= 160 mM, suggesting the formation of particle clusters, and reveals a micelle-induced
depletion effect. The twice corrected half-times, tsII , which are tsI divided by the viscosity of the solution
to account for the effect of viscosity on the agglomeration rate, are consistent with these inferences. In
addition, the DLS data support our inference that the clusters in region C are flocs, which can be de-
agglomerated with some mild agitation. A recently published model of agglomeration and sedimentation
(Yang et al., 2016) is used to fit the data for estimating the effective Fuchs-Smoluchowski stability ratio
W for flocculation. W -values are inferred to be over 106 at cTX100 <160 mM, but only ca. 500–1000 at
higher concentrations, supporting the hypothesis of a depletion-induced flocculation that is masked by
the solution viscosity.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction agglomeration and sedimentation of suspensions with polydis-

perse titania particles (d ∼= 280 ± 100 nm) in water or in 100 mM
In colloidal suspensions, one may add surfactants or polymers NaCl solution. The main motivation was to stabilize white pig-
as dispersants, or stabilizers, to keep the colloidal particles from ments, i.e. titania particles, against sedimentation, which would
agglomerating, and then from settling more rapidly, by provid- improve the use of white inks in ink-jet printers. Based on the
ing either electrostatic stabilization, or steric stabilization, or both. relative agglomeration and sedimentation rates, the suspensions
An excess concentration of colloidal-size additives may, however, were categorized into three regions based on the SDS concentra-
induce strong depletion interactions between the larger colloidal tion cSDS : (1) region 1, where cSDS < cmc, (2) region 2, where cmc ≤
particles, resulting in significant flocculation (easily reversible cSDS < c ∗ , where c ∗ = 60 mM in water and 105 mM in 100 mM
agglomeration), and hence in faster particle settling [1]. Attrac- NaCl solution, and (3) region 3, where cSDS ≥ c ∗ . In region 1, with
tive depletion forces between the larger particles are caused by increasing surfactant concentration, the particles become more
the exclusion of nonadsorbing colloidal depletants from the gap stable against agglomeration, and in turn, more stable against sed-
between the two approaching particles. Then, the concentration imentation, due to the increasing strength of electrostatic, and
difference of depletants between the gap and the bulk solution pro- probably also steric, repulsive interactions. In region 2, the parti-
duces an unbalanced osmotic pressure difference, which yields a cles settle with their intrinsic settling rates and the sedimentation
net attractive force between the large particles [1]. These forces half-time ts , defined as the time for nearly all the particles to settle
can be important when the concentration of the depletants, which by the half height (ca. 0.5 cm) of the suspension, is about 45 h. In
can be nanoparticles, micelles, or polymers, is high enough. this region, the ts -values are independent of the surfactant con-
Depletion effects have been studied for over 100 years, both the- centration, because as the surfactant concentration reaches the
oretically and experimentally [1–6]. Various depletants have been cmc, at which the adsorption of the surfactant monomers on the
used, including spherical rod-like or disk-like nanoparticles, poly- particle surfaces is maximum, no agglomeration occurs in the sus-
mers, polyelectrolytes, and micelles [7–18]. The depletion potential pensions, since the interparticle repulsive interactions overcome
interaction (dep ) between two uncharged spherical particles with the attractive van der Waals (vdW) attractive interactions. The
uncharged spherical depletants was first modeled by Asakura and monomers and micelles neither slow down further the sedimen-
Oosawa [2,19] and is given by tation, nor do they induce significant particle agglomeration that
⎧ would have caused faster sedimentation. In region 3, the high con-
⎪ ∞ h<0
⎪
⎨ (a − h)2  3d 
centration of the micelles induces significant depletion interactions
h between the particles, yielding fast flocculation. Those flocs, or eas-
dep = − +a+ ∞ kB T 0 < h <= a (1)
⎪
⎪
6 2 2 ily reversible agglomerates, sediment in much shorter times of
⎩ about 1–2 h.
0 h >= a In this study, we examine the effect of a nonionic surfactant,
where a is the diameter of the depletant, d is the diameter of Triton X-100 (TX100), on the stability of the same titania parti-
the particle, h is the surface-surface separation distance between cles in water against sedimentation and agglomeration. A TX100
the particles, ∞ is the depletant bulk number density, kB is the molecule has a hydrophilic polyethylene oxide chain and an aro-
Boltzmann constant, and T is the absolute temperature. For disper- matic hydrocarbon hydrophobic group. It is one of the most widely
sions of charged colloidal particles and with charged depletants, used nonionic surfactants in many household and industrial prod-
the depletion interactions are found to be much stronger than the ucts, as a dispersant or a wetting agent, due to its biodegradability
ones for uncharged particles and depletants [4,6–8,10,13,19–21]. and its good performance across the 15–45 ◦ C temperature range
Much of the literature on depletion effects has focused on mod- [27–29]. By comparing the behavior of this nonionic surfactant
els for describing the depletion forces quantitatively, the effects with SDS, we observe many qualitative and quantitative differ-
of the types of the particles and depletants used, and their appli- ences in several properties, including different sedimentation rates
cations. However, much remains to be understood, qualitatively with increasing surfactant concentration, larger viscosity effects,
and quantitatively, of their effects on agglomeration and how they smaller surface densities of the surfactant adsorbed on the titania
may induce faster sedimentation. Sedimentation caused by floccu- particles, and a worse initial dispersibility of the particles. Most
lation is used widely in particle separation processes [22–26]. In one importantly, for TX100, the depletion effect is not immediately
example, the use of flocculation as a low-cost method for harvest- apparent, because it is masked by the high solution viscosity. How-
ing microalgae for bulk biomass production has been reported [22]. ever, it still causes some flocculation. To better understand the
To our knowledge, little work has been reported focusing specifi- sedimentation time data and establish this hidden depletion effect,
cally on how those flocs affect qualitatively the sedimentation rate we use the model of sedimentation and agglomeration, which we
in depletant solutions [12]. published previously [30], to predict the half-times of sedimenta-
In our previous study, we examined the effect of an anionic tion, and compare the predictions to the data.
surfactant, sodium dodecylsulfate (SDS), on the stability against
298 Y.-J. Yang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 516 (2017) 296–304

2. Experimental methods

2.1. Materials and some key properties

Rutile titania particles were obtained from Huntsman, and were

used as received. They have a purity of 95 wt%, a nominal parti-
cle diameter of 260 nm and a density of 4.1 g/cm3 . The particles
were characterized as detailed previously [12]. They were quite
polydisperse with diameters d = 280 ± 100 nm, as determined from
SEM (scanning electron microscope) images. Ultrapure water was
obtained from a Milli-Q water system (from Millipore), which uses
distilled water as input. Triton X-100 (TX100), which is a mix-
ture of 4-t-octylphenoxyethoxyethanols with an average molecular
weight of 625 Da (g/mole), was purchased from Sigma-Aldrich (MO, Fig. 1. Effect of Triton X-100 concentration on the half-time of sedimentation:
USA) and was used without purification. Its chemical formula is region A, where 0.24 mM; region B, where 30 mM; region C, for . Filled squares:
(CH3 )3 CCH2 C(CH3 )2 C6 H4 O− (CH2 CH2 O)p H with an average num- actual data of sedimentation half-times ; open blue triangles: corrected data of sed-
imentation half-times for bulk solution viscosity; open red diamonds: data corrected
ber p of ethylene oxide (EO) groups of 9.5. Its density is 1.065 g/cm3 . a second time, for the effect of viscosity on agglomeration rates; see text for details
The value of p has a range of 3–18, and the most abundant compo- and Fig. 2 for viscosity data. The lines are aimed to aid the eye. (For interpretation
nents have a p-value of 7–10 [31]. The critical micelle concentration of the references to colour in this figure legend, the reader is referred to the web
(cmc) in water at 25 ◦ C is 0.23 ± 0.01 mM, and the (average) micelle version of this article.)

aggregation number has been reported to be about 105 [32,33].

The partial specific volume in water is 0.9125 mL/g at 20 ◦ C (par- 2.5. Dynamic light scattering (DLS) measurements
tial specific density = 1.096 g/cm3 ) [34]. The micelles have been
shown with SANS (small angle neutron scattering) to have an oblate The average hydrodynamic diameter, dh , of the titania particles
spheroidal shape with two equal half axes with a = b = 5.2 nm, and was measured at 25 ◦ C, with a Brookhaven ZetaPALS dynamic light
with the third axis c = 2.9 nm [35]. These values include a degree of scattering instrument. The wavelength used in the measurements
hydration of 1.18 g of water per gram of surfactant. The hydrody- was 659 nm, and the scattering angle was 90◦ . The suspension
namic diameter is about 8 nm in water at 25 ◦ C for cTX100 ≤ 50 mM. was first homogenized gently by hand, and then diluted with the
The micelle size grows with increasing concentration, temperature, respective suspension medium or with water to a weight fraction of
and ionic strength; in addition, micelle clusters may form at the 5 × 10−6 (0.0005 wt% or 5 ppm). The dilution was needed to avoid
higher surfactant concentrations [35–38]. excessive light scattering. The measurements were completed in a
short time after dilution to minimize any changes in the hydrody-
namic radius of the agglomerates upon the dilution.
2.2. Method for preparing aqueous suspensions
2.6. Surfactant adsorption studies
The suspension media were aqueous TX100 solutions with
various surfactant concentrations cTX100 , from 0.05 to 250 mM The adsorbed densities of TX100, TX100 , on the titania particles
(0.003–15.6 wt%), or equivalently 0.22–1087 times the cmc. Titania at 25 ◦ C were determined as follows. The particles were equili-
suspensions were prepared with 1 wt% titania particles in aqueous brated with a TX100 solution, which had surfactant concentrations
TX100 solutions by adding the pigment particles to each suspen- ranging from 0.02 to 0.98 mM, for at least 24 h. Then, the suspen-
sion medium. Each suspension was first shaken by hand, and then sions were centrifuged at ca. 18,000g0 , where g0 = 9.8 m/s2 , for
stirred magnetically for 30 min, followed by sonicating for 3 h in 30 min with a Beckman Coulter Microcentrifuge until there were
a Branson 3510 sonication bath at a frequency of 40 kHz. All the few or no particles left in the supernatant, as confirmed visually
samples were prepared at room temperature, which was 25 ± 2 ◦ C. and with DLS data. The final TX100 concentration, cTX100,f , in each
supernatant solution was determined by using a Cary 300 UV–Vis
spectrophotometer at a wavelength of 274 nm. The cTX100,f -values
were then used to calculate the steady-state adsorption surface
2.3. Sedimentation half-times of titania particle suspensions densities of TX100 with specific surface area of 12.2 m2 /g obtained
from the BET (Brunauer-Emmett-Teller) method.
The stability against sedimentation was monitored visually by
measuring the height h (t) of the moving front of a nearly-clear
supernatant top layer against a lower white suspension, as the par- 3. Experimental results and discussion
ticles settled. The half-time of sedimentation (ts ) was determined,
for a standard sample with an initial height of 1 cm, at a front height 3.1. Effect of Triton X-100 concentration on sedimentation
of h = 0.5 cm [12]. All experiments were done at room temperature. half-times

The sedimentation half-times, ts , were measured for the sus-

pensions of 1 wt% titania particles as a function of the TX100
2.4. Viscosity measurements concentration, cTX100 ; see Fig. 1. Three regions were identified. In
region A, for cTX100 < cmc, the half-time increased with increas-
The viscosities of the suspension media were measured with ing TX100 concentration from 2 ± 1 h to about 30 ± 7 h. In this
a series of calibrated Cannon-Fenske viscometers (no. 25, no. 50, region, the surfactant played a major role in slowing down the
and no. 100) at 25 ◦ C. The series of viscometers are suitable for sedimentation, probably by stabilizing the particles against some
viscosity values of 0.5–15 cP. At least five measurements were done agglomeration [12,30]. We previously showed in the SDS study,
for each sample, and the averages and the standard deviations were however, that if these polydisperse titania particles were non-
calculated. agglomerating, they will have a maximum half-time of around
 Y.-J. Yang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 516 (2017) 296–304 299

observed increase in the sedimentation half-times in this region is

not due to any viscosity effect. In region B, the viscosity increased
by up to 7%, compared to that of water, at cTX100 = 30 mM. No corre-
sponding half-time increase was observed, and the slight viscosity
effect could be masked by the uncertainty, or the standard devi-
ation, of the data. But, for cTX100 ≥ 30 mM (region C), the viscosity
increased significantly, from 0.94 mPa s (0.94 cP) at cTX100 = 30 mM
to 6.55 mPa s at cTX100 = 250 mM. This substantial increase in vis-
cosity should lead by itself to a larger sedimentation half-time, since
ts for non-agglomerating particles (at dilute condition) is given by
the following equation [40],

ts = L/2vsed = 9L/d2 g (2)

where vsed is the settling velocity predicted from Stokes law, L is

Fig. 2. Relative viscosities (black circles), of aqueous Triton X-100 solutions as a
function of the surfactant concentration, and predictions from various model equa-
the initial height of the sample suspension, d is the diameter of
tions for spheres with intrinsic viscosity 2.5 or for oblate spheroids with intrinsic the particles,  is the viscosity of the suspension medium, and 
viscosity 3.2. Line 1: Einstein’s eq. for spheres; Line 2: Einstein’s eq. for spheroids; is the density difference between the particle and the suspension
Line 3: Eq. (6) for spheres; Line 4: Eq. (6) for spheroids; Line 5: Krieger-Dougherty medium. The density of the suspension medium changes little with
Eq. for spheres; Line 6: Krieger-Dougherty Eq for spheroids.
surfactant concentration. Therefore, the density difference can be
presumed to be constant for a non-agglomerating particle. If L, d,
45 ± 2 h. The smaller maximum half-time and its larger standard and  are constant, then Eq. (2) indicates that the value of ts /
deviation observed with TX100 may be due to a poorer dispersibil- should be constant as well.
ity provided by the TX100 molecules. Some evidence from DLS In order to examine quantitatively the viscosity effect on the
data support this inference; see Section 3.4. Thus, for this surfac- sedimentation half-time, we corrected the half-times by dividing
tant, the reference maximum half-time is about 30 h. In region B, them by the ratio of the viscosity of the respective suspension
for cmc ≤ cTX100 < c ∗∗ , where c ∗∗ was found to be around 30 mM, medium to that of water w , and then the corrected half-times tsI
the value of ts varied little with surfactant concentration, remain- are given by
ing constant at about 30 h. The TX100 monomers and the micelles  
w
therefore played no role in either slowing down or accelerating the tsI = ts (3)

particle sedimentation, as observed and identified in region 2 in the
SDS study. For regions A and B, the viscosity effect is negligible, as tsI ≈ ts ;
In region C, for cTX100 ≥ c ∗∗ , the half-time increased with increas- see Fig. 1. In region C, the corrected half-times are about 34 h for
ing surfactant concentration from about 30 h at cTX100 = 30 mM to cTX100 up to about 160 mM, suggesting that no agglomeration may
about 142 h at cTX100 = 250 mM. The presence of excessive con- have occurred and that the slower sedimentation is mainly due
centrations of TX100 micelles seems to provide a strong resistance to a viscosity effect. At higher concentrations, ranging from 160
against particle sedimentation. This behavior differs from what was to 250 mM, tsI decreases to about 20 h as cTX100 increases, imply-
observed at high micelle concentrations, or region 3, in the SDS ing the formation of agglomerates, which settle faster. Some type
study, in which we inferred that the SDS micelles induced strong of additional attractive interaction should be responsible for their
attractive depletion interactions and destabilized the particles to formation. We postulate a micelle-induced depletion interaction,
form large flocs, which settled in 1–2 h. By contrast, there was no which is evidently masked by the higher viscosity, and that the
readily obvious depletion effect arising from the TX100 micelles at agglomerates are flocs, which can be easily de-agglomerated with
high micelle concentrations. some mild agitation. The effect of such an interaction may be, how-
Initially, and given the observed behavior in region C, one may ever, less pronounced than the one observed with the SDS micelles
infer that depletion effects are non-existent, or at least unim- [12].
portant, for these nonionic surfactant micelles. This conclusion, The increase in viscosity, however, can slow down not only
however, contradicts various literature results for this nonionic the particle sedimentation, but the particle agglomeration, since
surfactant or similar surfactants with other types of particles. Ma the agglomeration rate constant in the standard Smoluchowski
[5] reported a depletion destabilization of the suspensions with equation is inversely proportional to the viscosity. In turn, the

polystyrene particles with diameters of 795, 562, or 300 nm at agglomeration time is proportional to the viscosity  or to w w
cTX100 ∼= 20, 40, and 50 mM, respectively, at 30 ◦ C, as determined [30]. Since for region C, the sedimentation time is controlled
with UV–vis spectrophotometry. Degiorgio et al. [39] also observed substantially by the rate of agglomeration, which produces increas-
settling of some polystyrene latex particles, dispersed in aque- ingly larger particles, as the agglomeration rate decreases, the
ous TX100 solutions, resulting from the micelle-induced depletion effective agglomerate size for sedimentation would be smaller.
effects. It is noted, however, that our initial analysis of the sedimen- Hence, the sedimentation time will  be

longer, and will increase
tation times does not take into account the effects of viscosity. We approximately by a factor close to w . Hence, a second correc-
now examine possible viscosity effects on sedimentation times by tion of the half-times is needed, to account for the viscosity effect on
considering the viscosity of the suspension media. the agglomeration rate, and the sedimentation rate. We therefore
define the secondarily-corrected half-times tsII as follows,
3.2. Viscosity measurements of aqueous Triton X-100 solutions     2
w w
tsII = tsI = ts (4)
 
3.2.1. Effects of viscosity on the sedimentation half-times
The viscosity  of aqueous TX100 solutions was found to For regions A and B, tsII ≈ tsI , suggesting that the viscosity has lit-
increase significantly with increasing concentration of TX100, tle or no additional effect on the particle agglomeration. In region
cTX100 ; see Fig. 2. For cTX100 < cmc (region A), the viscosity C, tsII decreases with increasing surfactant concentration. At such
increased by less than 2% compared to that of water. Hence, the high concentrations, particle agglomeration occurs at a slower rate,
300 Y.-J. Yang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 516 (2017) 296–304

because of the relative high viscosities. If the particle agglomera-

tion had not been impeded by the relative high viscosities, then, as
explained earlier, larger flocs would have formed and settled faster.
We infer that the depletion effects induced by the TX100 micelles
are mostly masked by the large increase of the suspension medium
viscosity, which slows down the rate of sedimentation once, and
then twice because of the slower agglomeration.

3.2.2. Comparison of viscosity data to various model equations

For examining the shapes, the sizes and the volume fraction of
the micelles, the relative viscosity r = /s is calculated and com-
pared to various model equations, where s is the viscosity of water,
w , for concentrations below the cmc, or the viscosity of the solu-
tion at the cmc, for concentrations above the cmc. Since cmc is
Fig. 3. Surface densities of Triton X-100 surfactant adsorbed on the titania parti-
small, s ≈ w . The model equations include the generalized Ein-
cles as a function of the final equilibrium surfactant concentration after reaching
stein relation for low volume fractions, the generalized equation adsorption equilibrium; the vertical dashed line indicates the CMC.
for high volume fractions, and the Krieger & Dougherty equation,
which is valid for even higher volume fractions. In these calcula-
tions, the TX100 micelles are treated as “dispersed particles.” To the micelles, or to the effects of increasing micelle sizes or aggre-
estimate the volume fraction of the micelles m , we used the liter- gation numbers at higher concentrations. One may infer that the
ature values of the average aggregation number na of 105 and the TX100 micelles should indeed have an oblate spheroidal shape with
hydrodynamic diameter of 8 nm, in which the substantial hydration an axial ratio of 2.0, as reported [35], and induce stronger depletion
of the micelles head group is included [34,37]. These values were interactions than spherical micelles [1].
used as a first approximation for all concentrations, even though
the actual values may increase with the surfactant concentration. 3.3. Surface densities of Triton X-100 adsorbed on the titania
Using the measured partial specific density of TX100, we find particles
m = Vm NA (cTX100 − cmc) /na (5)
The surface densities of TX100 molecules, TX100 , were mea-
where Vm represents the volume of a single spherical micelle, and sured for initial concentrations, cTX100 , ranging from 0.02 to
NA is Avogadro’s number. The micellar volume fraction is 0.1 at 0.98 mM, which led after adsorption to final concentrations,
cTX100 ∼
= 65 mM, and 0.38 at cTX100 ∼ = 250 mM. The corresponding cTX100,f , ranging from 0.0001 to 0.89 mM; see Fig. 3. For cTX100,f <
molar concentration of the micelles at 250 mM is estimated to be 0.1 mM, the surface density remains low, with values below
2.4 mM. 0.002 ␮mol/m2 , suggesting that little surfactant adsorbs on the
The generalized Einstein relation, Eq. (6), is used as the first particle surfaces. The TX100 -value increases to 0.015 ␮mol/m2 at
model for predicting the relative viscosity. cTX100,f = 0.14 mM, and then remains constant forming a plateau,
when cTX100,f is close to the cmc. Above the cmc, the surface
r = 1 + [] m (6)
density increases and reaches a second plateau with a value
where [] is the intrinsic viscosity, which for spherical particles is of 0.64 ␮mol/m2 at cTX100,f = 0.40 mM. The observed adsorption
equal to 2.5, and generally is a fitting parameter based on the par- isotherm is like a Type IV isotherm, which has been reported
ticle shape; for an oblate spheroid with an axial ratio of 2.0, it is for adsorption of other nonionic surfactants from solution onto
taken to be 3.2, as calculated from Simha theory [40]. Equation (6) hydrophilic solid surfaces [42,43].
is valid for particles with relatively low volume fractions, below 0.1, The minimum area per molecule Am is calculated from the max-
at which there are no significant interparticle hydrodynamic inter- imum surface density of 0.64 ␮mol/m2 , where Am ≡ 1/ (NA max ),
actions [40,41]. For higher volume fractions, the effect of particle to be about 250 Å2 /molecule. This is larger than the reported values
“crowding” should be considered. A general empirical equation for for hydrophilic particles [5,31,44]. Dong et al. reported a value of
the relative viscosity vs. m is [40] Am = 65 Å2 /molecule for hydrophobic CuPc particles, and a value of
−[] 121 Å2 /molecule for less hydrophobic CuPc particles, which have
r = (1 − m ) (7)
some sulfonate groups on the surfaces [31]. Cano et al. reported val-
Eq. (7) reduces to the Einstein relation for m 1. The upper ues of about 50–75 Å2 /molecule for various types of hydrophobic
limit on m in Eq. (7) is m =1 [40]. To account for the maximum polystyrene particles with different functionalities [44].
allowed value of m , or max , due to excluded volume packing The smaller Am -values, or larger max -values, for hydropho-
effects, another empirical equation has been proposed by Krieger bic surfaces, could be explained as follows (Fig. 4(b)). The phenyl
and Dougherty [40,41], group of the TX100 molecules may adsorb on the hydrophobic
 m
−[]max surfaces with the hydrophilic EO chains extended in the aqueous
r = 1− (8) phase, resulting in a packed monolayer. This conformation of TX100
max
molecules would yield a max -value of about 2.5–3.3 ␮mol/m2 and
where max is taken to be 0.64 for a random close-packing of a Am -value of about 50–75 Å2 /molecule, close to the Am value of
spheres [40]. 48–55 Å2 /molecule reported for the air-water interface. By con-
The predictions from Eqs. (5)–(7), for [] = 2.5 or for [] = 3.2, trast, for hydrophilic surfaces, such as those in this study, the high
are shown in Fig. 2. The predictions from Eq. (6) fit the data for flexibility of the EO chains may yield different surfactant layer
cTX100 up to around 50 mM for either shape, and underestimate conformations, in which the EO chains may be in direct contact
the data significantly at higher concentrations, at which the data with the surfaces. Then, these chains would no longer extend into
follow more closely the predicting of Eqs. (7) and (8). The Krieger- the aqueous phase but lie flat on the surface, resulting in smaller
Dougherty equation, Eq. (8), for [] =3.2 fits the data the best, surface densities; see Fig. 4(a). A cooperative effect, of a surface-
especially at concentrations above 160 mM, where m > 0.25. The aggregation-type, may also occur, preventing the hydrophobic
small discrepancies may be attributed to some interactions among phenyl groups of the first layer from being in contact with water.
 Y.-J. Yang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 516 (2017) 296–304 301

compared to the ones in SDS solutions. Moreover, we compared the

DLS data to an average effective particle size, deff , estimated from
the measured values of ts by using the polydisperse model based
on the Mason-Weaver equation, which was published previously
in reference [12]. In the model, the particle sizes are assumed to
be normally distributed, with an average diameter of d̄ and a stan-
dard deviation d. Then the 95% confidence interval of the particle
sizes ranges from d̄−2d to d̄+2d, and the corresponding weight
fraction of the smallest particle is 0.05. From the model, we obtain
deff = 310 ± 170 nm, which is in fair agreement with the DLS data.
The discrepancy of the uncertainties might be due to the different
descriptions of the polydispersity used in the DLS measurement
and in the model.
At 96h, for cTX100 = 0.20 mM (in region A), the particles agglom-
erated, with cluster sizes exceeding the measurable size range of
the DLS technique. For cTX100 = 0.96 and 1.5 mM (in region B),
the particles underwent some coagulation after they settled, and
formed clusters with diameters of about 620 and 553 nm. The DLS
data suggest that the stability against coagulation slightly improved
Fig. 4. Schematic illustration of the presumed conformations of Triton X-100
molecules (a) on a hydrophilic surface, with a mushroom-type conformation, or with increasing surfactant concentration in region B. However,
(b) on a hydrophobic surface; see text for details. the half-time remains constant with increasing surfactant concen-
tration in this region, suggesting that the particles settled before
significant agglomeration had a chance to occur [30]. At cTX100 =
45 and 90 mM (in region C), the hydrodynamic diameters were
about 320 nm and 290 nm, respectively, and they did not change at
96 h, indicating good stability against coagulation in this time scale.
The results imply that the presence of the micelles plays an impor-
tant role in slowing down particle coagulation, which may arise
from the enhancement of the repulsive steric interactions or from
the increase of the viscosity. The viscosity increase slows down the
Brownian motion of the particles and therefore reduces the prob-
ability of the particle collisions. At cTX100 = 250 mM, the dh -value
was about 300 nm and remained constant at 96 h, suggesting that
the particles are stable against coagulation, as observed at other
concentrations in region C. However, the decrease in the corrected
half-times reveals that some type of agglomeration occurred while
Fig. 5. Hydrodynamic diameter of titania particles dispersed at 1 wt%, for 5 min or the particles were settling (Section 3.2). The DLS data confirmed
98 h, in aqueous solutions of various TX100 concentrations. For the DLS measure- that the agglomeration was reversible. Hence, the agglomeration
ments, each sample was diluted in water to about 5 ppm, and the measurement was
was flocculation, rather than coagulation.
done in less than 10 s after dilution, to minimize any change in the agglomerate size.
Agglomeration occurred after 96 h only for 0.20, 0.96, and 15 mM. On the basis of the available evidence discussed above, we can
conclude that: (i) in region A, the particles are unstable against
coagulation, and the agglomeration and the sedimentation time
The EO chains in the second layer might also have a “mushroom- scales are comparable; (ii) in region B, the particles undergo no
type” conformation (see Fig. 4), since the water is a good solvent coagulation or flocculation before they settle out, but eventually
for them. The adsorbed TX100 molecules, either monomers or coagulate after they settle out; (iii) in region C at cTX100 <160 mM,
hemi-micelles, could provide some steric protection against par- the suspensions are also stable against coagulation and flocculation,
ticle coagulation, and allow for the behavior observed in region B and (iv) in region C at cTX100 ≥ 160 mM, the suspensions are stable
in Fig. 1. against coagulation, but unstable against flocculation.
To examine the effects of the TX100 micelles on the formation
3.4. DLS results on dispersibility and stability against coagulation of the flocs at the higher concentrations in region C, we diluted
or flocculation the suspensions with cTX100 = 250 nm in its suspension medium at
5 min after sample preparation, and kept the diluted sample still
DLS measurements were used to assess the dispersibility of the in the DLS instrument. The hydrodynamic diameter was tracked
particles for the mixing protocol used and the stability of the sus- for about 66 h (Fig. 6), and remained the same in this time scale,
pension against coagulation for cTX100 = 0.20, 0.96, 1.5, 45, 90, and suggesting that no significant flocculation occurred in the diluted
250 mM. The hydrodynamic particle sizes, dh , in aqueous TX100 suspension. At the lower particle concentrations, the rate of the
solutions for all the concentrations examined were measured 5 min particle collisions is smaller, leading to a weaker depletion effect
and 96 h after sample preparation. These measurements were done than that for the more concentrated suspension.
after dilution in water.
At 5 min, for all the concentrations examined, the initial particle
sizes were around 310 ± 100 nm, slightly larger than the primary 4. Comparison of the depletion interactions of TX100 to
particle size (Fig. 5). The reported uncertainty represents size poly- those of SDS micelles
dispersity, rather than errors of the averages, and it is the peak
width at half its height. The dispersibility in aqueous TX100 is bet- The doubly corrected sedimentation half-times, tsII , in this study
ter than in water but worse than that in aqueous SDS solutions [12]. are compared to the actual sedimentation half-times in the previ-
The slightly larger initial size accounts for the smaller half-time ous SDS work, ts , for revealing the relative strength of the depletion
302 Y.-J. Yang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 516 (2017) 296–304

SDS, which interact with long-range electrostatic repulsive forces

∗ is at the order of 1 for these
rather than steric forces. However, cm
two different surfactants. The generality of this critical concen-
tration for globular micelles needs to be established with more
surfactant structures.

5. Half-times predicted from the agglomeration and

sedimentation model

So far, we have inferred that, at the higher TX100 concentrations

in region C depletion-induced flocculation occurs but is masked by
the high solution viscosity, which lowers both the sedimentation
rate and the agglomeration rate. To model the flocculation induced
Fig. 6. Hydrodynamic diameter of 5 ppm titania particles in 250 mM Triton X-100 by the depletion effects, we need to calculate the total potential
solutions. For such dilute dispersions, no flocculation or coagulation occurred for at energy of interparticle interactions, including the van der Waals
least 70 h. (vdW) potential, the electrostatic potential, the steric potential, and
the depletion interaction potential between the particles. One goal
is to seek for the presence of a secondary potential energy mini-
mum, which may cause flocculation, as was done previously [12],
and calculate the effective Fuchs-Smoluchowski stability ratio W
against flocculation. Several parameters needed for such calcula-
tions are, however, not known accurately or are unknown. Such
parameters are the effective Hamaker constant, the effective sur-
face potential, the effective hydrodynamic micelle diameter, the
thickness of the adsorbed surfactant layer, and the effective volume
fraction of the surfactant in the adsorbed layer, which determines
the steric interaction and may be non-uniform. Small variations of
these values can dramatically affect the predictions of the Fuchs-
Smoluchowski stability ratio and the predictions of the depth and
the location of a secondary energy minimum. Because of these dif-
ficulties, we do not rely on W -calculations, but present a different
Fig. 7. Dimensionless sedimentation times, vs molar concentration of micelles for
TX100 and SDS; see text for details. The broken lines are meant to aid the eye.
approximate approach to search for possible depletion effects that
yield the masked or hidden flocculation at the higher concentra-
tions in region C. We apply the new model of agglomeration and
effects, induced by these two types of micelles in the absence of the sedimentation that we published recently [30].
effects of the solution viscosity, on both sedimentation and agglom- The model is used for predicting an upper bound of the net
eration. In the first comparison, for a 250 mM TX100 solution, with half-time of sedimentation for an initially monodisperse suspen-
a micelle molar concentration cm ∼ = 2.4 mM and a 100 mM SDS solu- sion, where particles settle without or with agglomeration [30].
tion, with the same micellar concentration cm ∼ = 2.5 mM (found by The agglomeration times, tan USS , in the model are calculated based
dividing 100 mM by the aggregation number of 40), the sedimen- on the unsteady-state diffusion (USS) model of agglomeration and
tation times are about 2.5 and 1 h, respectively. In the absence of depend on the stability ratio W . The time tan USS for monomers with
other interactions, the SDS micelles, serving as depletants, would a diameter of d1 and a volume fraction of 1 to form n-mers, or
be expected to affect more strongly the rate of flocculation, proba- specifically 2Nm -mers, was derived in reference [30] and is given
bly due to their effective diameter a being around 12 nm, which is by the equation
larger than that of 8 nm for TX100 micelles; see Eq. (1). The value of
 2
12 nm for the effective micelle diameter of SDS includes a contribu-
p=Nm −1  3p
SS
ta2
12
3p
USS
tion by the double layers of the charged particles and the charged tan =
1 2 df + W 2p −2 2df (9)
p=0
1
micelles, as detailed in Ref. [12]. For SDS, the value of the effective
micelle size is a function of the ionic strength, which affects the
where df represents the fractal dimension of an n-mer agglomerate;
Debye length.
df is taken to be 1.8 for diffusion-limited agglomeration;
1 is the
A more general comparison is given in Fig. 7, where the dimen-
characteristic monomer diffusion time, given by the equation
sionless sedimentation times, s , are plotted against cm (or ∞ in Eq.
(1)). For SDS, s is defined as the ratio of the actual sedimentation 3d13
time to 45 h, which is the time for single unagglomerated titania
1 = (10)
2kB T
particles to settle. For TX100, s is calculated as the ratio of the
SS is the time
where  is the viscosity of the suspension medium; ta2
doubly corrected sedimentation times to the plateau sedimentation
time of 30 h of Fig. 1. for the monomers to form dimers in the steady-state model of
At low micellar concentration, s , is close to 1, because the agglomeration and is equal to
depletion forces are overall weaker. As cm increases, s decreases
∗ of about 0.95 mM for TX100, SS
99␲d13
beyond a critical concentration cm ta2 = (11)
or 1.3 mM for SDS. The value of cm ∗ , over which their effect on 161 kB T
s
is observable, for SDS micelles is slightly larger than the one for The diameter dn of a cluster consisting of n monomers and the
TX100 micelles, even though the depletion forces are predicted to density difference between the cluster and the suspension medium
be stronger for SDS micelles due to the larger effective diameter of n are estimated from the following fractal model,
the SDS micelles. This slightly larger cm ∗ -value might be due to the

sum of the other interparticle potential energies being stronger for dn = d1 (n)1/df (12)
 Y.-J. Yang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 516 (2017) 296–304 303

determined for SDS molecules. This difference could account for the
worse dispersibility in TX100 molecules as compared to SDS [12].
For concentrations above the cmc (0.24 mM), but below 30 mM
(region B), the half-times remain constant. In this concentration
range, the TX100 micelles affect neither the rates of sedimentation
nor the rates of agglomeration.
At cTX100 ≥ 30 mM (region C), the half-times increase substan-
tially, from 30 to about 140 h at cTX100 = 250 mM, primarily because
of the higher solution viscosity, which increases about 7-fold. For
examining the viscosity effect on the sedimentation quantitatively,
the half-times were corrected by the viscosity ratio of the sus-
pension medium to that of water, and denoted as tsI . These values
remain constant of about 34 h at 30 mM≤ cTX100 < 160 mM. For
Fig. 8. Comparison of the measured sedimentation half-times in region C with the cTX100 ≥ 160 mM, a small decrease in tsI is observed, suggesting
half-times predicted from the agglomeration/sedimentation model with various the formation of agglomerates, which settle faster. We postu-
Fuchs-Smoluchowski stability ratios as indicated. late that the micelle-induced depletion interaction is responsible
for the formation of these clusters, and that these clusters are
 d df −3 3 easily reversible agglomerates, or flocs. This depletion effect is
n
n = 1 = 1 (n)1− df (13) masked, however, by the effect of the solution viscosity, which
d1
not only slows down the sedimentation, but also decreases pro-
The intrinsic sedimentation half-time of a cluster, tsn , is deter- portionally the agglomeration rate. A twice corrected half-time tsII
mined by Stokes law (Eq. (2) with d and  replaced by dn and n ). was calculated, for considering the viscosity effect on the parti-
Both tan and tsn depend on the number of monomers n in a given cle agglomeration, by dividing tsI by the viscosity ratio. At these
cluster. As n increases, tan increases, while tsn decreases. When tan concentrations, tsII < tsI , indicating that larger flocs would form and
and tsn are about equal, which should occur for some cluster size would settle faster if there were no viscosity effects. Further, as
n = n∗ , the net or measured sedimentation half-time ts∗ should be cTX100 increases, tsII decreases from 34 h to 2.6 h, suggesting that the
roughly equal to tan∗ and tsn∗ . micelle-induced depletion effects would be much stronger if there
As described, the model cannot be applied to suspensions with were no viscosity effects involved. Besides, the viscosity data are
initially polydisperse sizes, because then a wide range of tan and tsn fairly consistent with the predictions from the Krieger-Dougherty
values are predicted. The actual system contains particles with an equation for hydrated micelles with an intrinsic viscosity of 3.2,
average diameter of 310 nm and a volume fraction of  = 0.0025. consistently with oblate spheroidal micelle shape with an axial
As detailed in Section 3.4, the smallest particles in this poly- ratio of 2.0. The spheroidal micelles may induce stronger depletion
disperse system should have a diameter of about 140 nm and a effects than the spherical micelles.
volume fraction of  ∼ = 0.0001. Hence, we consider two simpler, The dispersibility in aqueous TX100 solutions is better than in
and monodisperse, suspensions for estimating the overall agglom- pure water but worse than in aqueous SDS solutions [12], in which
eration/sedimentation behavior and the sedimentation half-times the particles are dispersed well. Their initial hydrodynamic diame-
of the actual system: (a) an initially monodisperse system of hard ters in all regions are 310 ± 100 nm, slightly larger than the primary
spheres with a diameter d1 = 310 nm and a particle volume fraction particle sizes, obtained from SEM. In region B, the particles settle
1 = 0.002; (b) an initially monodisperse system of d1 = 140 nm with little or no agglomeration, but they undergo some coagulation
and 1 = 0.0001. In our previous study, we concluded that the sed- after settling out, probably because of the much higher particle con-
imentation time is mainly controlled by the smallest particles [12]. centration in the sediment. In part of region C, for cTX100 < 160 mM,
Therefore, case (b) seems to be the most relevant for our data. the suspensions are stable against coagulation and flocculation. For
The model predictions of the half-times for various values of the cTX100 ≥ 160 mM, the particles undergo some flocculation, but no
stability ratio are shown in Fig. 8. For 20 mM ≤ cTX100 < 160 mM, coagulation.
the data fit the predictions fairly well for a stability ratio W ≥ 106 , The new model of agglomeration and sedimentation [30]
implying low and insignificant rates of agglomeration. For cTX100 ≥ has been applied for estimating the possible effective Fuchs-
160 mM, the W -values that fit the data are much lower, about Smoluchowski stability ratio W against flocculation for region C. By
100–500, suggesting that agglomeration occurs at a faster rate. fitting the sedimentation half-times to the model, the data are con-
Depletion attractive interactions should therefore be responsible sistent with a large stability ratio W above 106 for cTX100 < 160 mM,
for the reduction of W , since evidently the micelle concentration but a much smaller value, around 100–500, better describes the
should not alter significantly the vdW attractive interactions and data at the higher concentrations. Since the vdW attractive inter-
possible electrostatic interactions between the particles. actions and the possible electrostatic repulsive interaction should
vary little with the micelle concentration, the lower stability ratio
6. Conclusions should be due to additional depletion attractive interactions, which
depend strongly on the surfactant concentration. The depletion
The sedimentation half-times ts of titania particles with diam- effect is hidden, or masked, by the effects of the viscosity of the
eters of 280 ± 100 nm increase, from 1 to ca. 30 h, with TX100 suspension medium. The higher viscosities at the higher micelle
concentrations up to the cmc (region A). This results from the concentrations not only decrease the sedimentation rate, but also
increasing surfactant adsorption, which may produce steric repul- slow down the depletion-induced agglomeration process, which
sive interactions that prevent the particles from agglomerating. would otherwise have accelerated the sedimentation.
The adsorbed densities of TX100 molecules on the hydrophilic
particle surfaces are much smaller than those on hydrophobic
surfaces [5,31,44], probably because the TX100 molecules lay flat Acknowledgements
on hydrophilic surfaces instead of forming close-packed mono-
layer structures. Furthermore, the maximum surface density of the This work was supported by Hewlett-Packard (Palo Alto, CA) and
TX100 molecules on titania surfaces is 17 times smaller than the one a graduate Bilsland Fellowship from Purdue University. The authors
304 Y.-J. Yang et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 516 (2017) 296–304

are grateful to Ms. Xilan Zhu for obtaining some experimental data [23] G.V. Duarte, B.V. Ramarao, T.E. Amidon, Polymer induced flocculation and
and to Prof. You-Yeon Won at Purdue University for the use of the separation of particulates from extracts of lignocellulosic materials,
Bioresour. Technol. 101 (2010) 8526–8534.
Brookhaven ZetaPALS instrument. [24] L.M. Nieto, G. Hodaifa, S. Rodríguez, J.A. Giménez, J. Ochando,
Flocculation–sedimentation combined with chemical oxidation process,
References CLEAN Soil Air Water 39 (2011) 949–955.
[25] B. Etter, E. Tilley, R. Khadka, K.M. Udert, Low-cost struvite production using
source-separated urine In nepal, Water Res. 45 (2011) 852–862.
[1] H.N.W. Lekkerkerker, R. Tuinier, Colloids and the Depletion Interaction, [26] M.I. Abro, A.G. Pathan, A.R. Memon, Sirajuddin, Dual polymer flocculation
Springer, Netherlands, 2011. approach to overcome activation of gangue minerals during beneficiation of
[2] S. Asakura, F. Oosawa, Interaction between particles suspended in solutions of complex iron ore, Powder Technol. 245 (2013) 281–291.
macromolecules, J. Polym. Sci. 33 (1958) 183–192. [27] J. Saien, Z. Ojaghloo, A.R. Soleymani, M.H. Rasoulifard, Homogeneous and
[3] S. Asakura, F. Oosawa, On interaction between two bodies immersed in a heterogeneous aops for rapid degradation of triton X-100 in aqueous media
solution of macromolecules, J. Chem. Phys. 22 (1954) 1255–1256. via uv light, nano titania hydrogen peroxide and potassium persulfate, Chem.
[4] B. Vincent, J. Edwards, S. Emmett, A. Jones, Depetion flucculation in Eng. J. 167 (2011) 172–182.
dispersions of sterically-stabilized particles (soft spheres), Colloids Surf. 18 [28] G. Zeng, H. Fu, H. Zhong, X. Yuan, M. Fu, W. Wang, G. Huang, Co-degradation
(1986) 261–281. with glucose of four surfactants, ctab, Triton X-100, sds and rhamnolipid, in
[5] C. Ma, The effect of Triton X-100 on the stability of polystyrene latices, liquid culture media and compost matrix, Biodegradation 18 (2007) 303–310.
Colloids Surf. 28 (1987) 1–7. [29] B. Wyrwas, Ł. Chrzanowski, Ł. Ławniczak, A. Szulc, P. Cyplik, W. Białas, A.
[6] R.I. Feigin, D.H. Napper, Depletion stabilization and depletion flocculation, J. Szymański, A. Hołderna-Odachowska, Utilization of triton X-100 and
Colloid Interface Sci. 75 (1980) 525–541. polyethylene glycols during surfactant-mediated biodegradation of diesel
[7] P. Weroński, J.Y. Walz, An approximate method for calculating depletion and fuel, J. Hazard. Mater. 197 (2011) 97–103.
structural interactions between colloidal particles, J. Colloid Interface Sci. 263 [30] Y.-J. Yang, A.V. Kelkar, D.S. Corti, E.I. Franses, Effect of interparticle
(2003) 327–332. interactions on agglomeration and sedimentation rates of colloidal silica
[8] F. Tardani, C. La Mesa, Attempts to control depletion in the surfactant-assisted microspheres, Langmuir 32 (2016) 5111–5123.
stabilization of single-walled carbon nanotubes, Colloids Surf. A 443 (2014) [31] J. Dong, S. Chen, D.S. Corti, E.I. Franses, Y. Zhao, H.T. Ng, E. Hanson, Effect of
123–128. Triton X-100 on the stability of aqueous dispersions of copper phthalocyanine
[9] C. Durand-Gasselin, R. Koerin, J. Rieger, N. Lequeux, N. Sanson, Colloidal pigment nanoparticles, J. Colloid Interface Sci. 362 (2011) 33–41.
stability of zwitterionic polymer-grafted gold nanoparticles in water, J. [32] G.E. Tiller, T.J. Mueller, M.E. Dockter, W.G. Struve, Hydrogenation of Triton
Colloid Interface Sci. 434 (2014) 188–194. X-100 eliminates its fluorescence and ultraviolet light absorption while
[10] V. Vlachy, Depletion interactions in ionic micellar solutions, Langmuir 12 preserving its detergent properties, Anal. Biochem. 141 (1984) 262–266.
(1996) 2881–2883. [33] A. Patist, S.S. Bhagwat, K.W. Penfield, P. Aikens, D.O. Shah, On the
[11] R. Roth, R. Evans, S. Dietrich, Depletion potential in hard-sphere mixtures: measurement of critical micelle concentrations of pure and technical-grade
theory and applications, Phys. Rev. E 62 (2000) 5360–5377. nonionic surfactants, J. Surfactants Deterg. 3 (2000) 53–58.
[12] Y.-J. Yang, A.V. Kelkar, X. Zhu, G. Bai, H.T. Ng, D.S. Corti, E.I. Franses, Effect of [34] H.H. Paradies, Shape and size of a nonionic surfactant micelle. Triton X-100 in
sodium dodecylsulfate monomers and micelles on the stability of aqueous aqueous solution, J. Phys. Chem. 84 (1980) 599–607.
dispersions of titanium dioxide pigment nanoparticles against agglomeration [35] P.S. Goyal, S.V.G. Menon, B.A. Dasannacharya, P. Thiyagarajan, SANS study of
and sedimentation, J. Colloid Interface Sci. 450 (2015) 434–445. micellar structure and interactions in Triton X-100 solutions, Phys. B 213
[13] A.B. Jódar-Reyes, A. Martín-Rodríguez, J.L. Ortega-Vinuesa, Effect of the ionic (1995) 610–612.
surfactant concentration on the stabilization/destabilization of polystyrene [36] R.J. Robson, E.A. Dennis, The size shape, and hydration of nonionic surfactant
colloidal particles, J. Colloid Interface Sci. 298 (2006) 248–257. micelles. Triton X-100, J. Phys. Chem. 81 (1977) 1075–1078.
[14] G.K. James, J.Y. Walz, Experimental and theoretical investigation of the [37] W. Brown, R. Rymden, J. Van Stam, M. Almgren, G. Svensk, Static and dynamic
depletion and structural forces produced by ionic micelles, Colloids Surf. A properties of nonionic amphiphile micelles: Triton X-100 in aqueous solution,
441 (2014) 406–419. J. Phys. Chem. 93 (1989) 2512–2519.
[15] J.K. Whitmer, E. Luijten, Sedimentation of aggregating colloids, J. Chem. Phys. [38] G.D.J. Phillies, J.E. Yambert, Solvent and solute effects on hydration and
134 (2011). aggregation numbers of Triton X-100 micelles, Langmuir 12 (1996)
[16] A. Trokhymchuk, D. Henderson, A. Nikolov, D.T. Wasan, A simple calculation 3431–3436.
of structural and depletion forces for fluids/suspensions confined in a film, [39] V. Degiorgio, R. Piazza, G. Di Pietro, Depletion interaction and phase
Langmuir 17 (2001) 4940–4947. separation in mixtures of colloidal particles and nonionic micelles, in: C.
[17] J.J. Cerda, T. Sintes, C.M. Sorensen, A. Chakrabarti, Structure factor scaling in Solans, M.R. Infante, M.J. García-Celma (Eds.), Trends in Colloid and Interface
colloidal phase separation, Phys. Rev. E 70 (2004). Science X, Steinkopff, Darmstadt, 1996, pp. 201–205.
[18] A. Tulpar, R.D. Tilton, J.Y. Walz, Synergistic effects of polymers and surfactants [40] P.C. Hiemenz, R. Rajagopalan, Principles of Colloid and Surface Chemistry, 3rd
on depletion forces, Langmuir 23 (2007) 4351–4357. ed., Marcel Dekker, New York, 1997.
[19] J.Y. Walz, A. Sharma, Effect of long range interactions on the depletion force [41] S. Mueller, E.W. Llewellin, H.M. Mader, The rheology of suspensions of solid
between colloidal particles, J. Colloid Interface Sci. 168 (1994) 485–496. particles, Proc. R. Soc. Lond. A 466 (2010) 1201–1228.
[20] R.I. Feigin, D.H. Napper, Stabilization of colloids by free polymer, J. Colloid [42] S. Paria, K.C. Khilar, A review on experimental studies of surfactant adsorption
Interface Sci. 74 (1980) 567–571. at the hydro philic solid–water interface, Adv. Colloid Interface Sci. 110
[21] A. Zaccone, H. Wu, M. Lattuada, M. Morbidelli, Correlation between colloidal (2004) 75–95.
stability and surfactant adsorption/association phenomena studied by light [43] C. Hinz, Description of sorption data with isotherm equations, Geoderma 99
scattering, J. Phys. Chem. B 112 (2008) 1976–1986. (2001) 225–243.
[22] D. Vandamme, I. Foubert, K. Muylaert, Flocculation as a low-cost method for [44] M.S. Romero-Cano, A. Martin-Rodriguez, F.J. De Las Nieves, Adsorption and
harvesting microalgae for bulk biomass production, Trends Biotechnol. 31 desorption of Triton X-100 in polystyrene particles with different
(2013) 233–239. functionality, J. Colloid Interface Sci. 227 (2000) 322–328.