OPEN Charging and discharging of single

SUBJECT AREAS:
colloidal particles at oil/water interfaces
APPLIED PHYSICS Peng Gao2*, XiaoChen Xing5*, Ye Li4, To Ngai5 & Fan Jin1,2,3
CHEMICAL PHYSICS
1
Hefei National Laboratory for Physical Sciences at Microscale, University of Science and Technology of China, Hefei, P. R. China
Received 230026, 2Department of Polymer Science and Engineering, University of Science and Technology of China, Hefei, P. R. China
230026, 3CAS Key Laboratory of Soft Matter Chemistry, University of Science and Technology of China, Hefei, P. R. China
17 December 2013
230026, 4Department of Modern Physics, University of Science and Technology of China, Hefei, P. R. China 230026, 5Department
Accepted of Chemistry, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong.
7 April 2014
Published The physical behavior of solid colloids trapped at a fluid-fluid interface remains in itself an open
2 May 2014 fundamental issue. Here, we show that the gradients of surface tension can induce particles to jet towards the
oil/water interface with velocities as high as < 60 mm/s when particle suspensions come in contact with the
interface. We hypothesize that rubbing between the particles and oil lead to the spontaneous accumulation
of negative charges on the hemisphere of those interfacial particles that contact the oil phase by means of
Correspondence and triboelectrification. The charging process is highly dependent on the sliding distances, and gives rise to
requests for materials long-ranged repulsions that protect interfacial particles from coagulating at the interface by the presence of
should be addressed to electrolyte. These triboelectric charges, however, are compensated within several hours, which affect the
F.J. (fjinustc@ustc.edu. stability of interfacial particles. Importantly, by charging different kinds of colloidal particles using various
cn) or T.N. (tongai@
spreading solvents and dispersion methods, we have demonstrated that charging and discharging of single
colloidal particles at oil/water interfaces impacts a broad range of dynamical behavior.
cuhk.edu.hk)

S
olid particles adsorbed at fluid-fluid interfaces have been traditionally exploited in emulsification1,2, mineral
* These authors and crude oil recovery as well as in food industry3,4. More recently they were found in a rapidly increasing
contributed equally to range of cutting-edge applications5–7, including the creation of nanostructured membrane for filtration and
this work. biphasic catalysis8–10, and the fabrication of nanocomposites with tunable electrical or optical properties11–13. For
the purpose of supporting these emerging applications, an improved understating of interfacial behavior of
particles14–16 and their dynamics at a fluid interface17,18 is a key issue. Many earlier experiments19–22 were carried
out on flat, planer fluid interfaces and various forces acting between the particles trapped at interface were
proposed. However, to date, the behavior of solid particles at interfaces is still a very dynamic area of research
with many open challenges that need to be met. For instance, it has been known that colloidal particles trapped at
interface between water and a medium of low relative permittivity (air or oil) would form interfacial dipoles
because of asymmetrical dissociation of surface charges around the particle17, which give arise to a long-ranged
repulsion between particles. It is thereby expected that adding salts can screen these interfacial dipoles and repress
the repulsion. However, Aveyard et al.23,24 showed that charged colloidal particles produce extremely stable
monolayers at an octane/water interface even at high electrolyte concentration, which cannot be addressed by
dipole-dipole repulsions alone. They proposed that this repulsive force lies in charge-charge repulsion, likely
arising from a small fraction of dissociated (residual) charges trapped by tiny surface water droplets in the pores of
the particle’s rough surfaces and in the oil phase, which result in unscreened charge-charge repulsion. In this
model, the presence of tiny water droplets in the oil phase is essential because these polar droplets can provide
high dielectric ‘‘nests’’ for trapping charged ions and subsequently causing the repulsion. However, we argued that
the proposed existence of these water droplets may not account for the presence of residual charges at a nonpolar
oil, especially one that has a low dielectric constant (eoil , 4.0), such as octane (eoil 5 2.0).
All the points put forth above provide us a standing point to revisit this unresolved problem. By using various
state-of-the-art microscopy techniques, here we show that the residual charges do exist in the hemisphere of
particles immersed in nonpolar oil (octane) and dominate the long-ranged repulsions between interfacial part-
icles when the typical dipole-dipole repulsions are screened by added electrolytes. In addition, our results
explicitly indicate that these residual charges are actually unstable, and are compensated in several hours, which
could in turn attenuate the strength of the charge-charge repulsions and eventually cause particles aggregation at
oil/water interfaces. Furthermore, we demonstrate that these unstable charges can be built spontaneously,
especially common approach relies heavily on using a spreading solvent to facilitate dispersion of particles at

SCIENTIFIC REPORTS | 4 : 4778 | DOI: 10.1038/srep04778 1

 www.nature.com/scientificreports

the oil/water interface. Importantly, we find that charging or dischar- Surprisingly, we found that at the octane/H2O interface, the particles
ging of colloidal particles at oil/water interfaces is a general phenom- initially were well separated at the interface, but the tendency for
enon, that thus provides an novel insight in understanding their interfacial particles accessing to their neighbors increased gradually
interactions and assembly at the interfaces, including the significant within the first 6 hours and then reached a stable state; this can
heterogeneity of interfacial interactions that have been observed be quantitatively characterized by the time evolution of radius dis-
between particle pairs trapped at fluid interfaces25,26. tribution functions (g(r)) of interfacial particles28, as shown in
Figure 1a. The time-dependent feature of g(r) implies that interac-
tions among interfacial particles, actually, are varying slowly. Similar
Results
phenomena have also been observed at an octane/10 mM NaCl
Time evolution of colloidal particles trapped at oil/water inter- interface, as shown in Figure 1b, indicating that the time-dependent
faces. To our best knowledge, direct measurements of the interac- feature of g(r) is inherent at oil/water interfaces, and is independent
tions between colloidal particles trapped at oil/water interfaces, for of whether presence of salt in water or not. In contrast to the micro-
example, by microscopy techniques, often encounter two technical spheres trapped at an octane/H2O interface, we find that those
problems: 1) Slow drifting of interfacial particles with a certain direc- microspheres form small clusters at octane/10 mM NaCl interface
tion along the interface, and therefore the prolonged perturbation after 2 hours (Fig. 1c), which lead to appearing peaks in their g(r), as
that alter the original spatial distribution of particles is expected to shown in Figure 1B. Our results indicate that microspheres would
occur; 2) Introducing the spreading solvents, such as methanol or become unstable at oil/water interfaces over time if the interfacial
isopropanol, has been demonstrated to affect greatly the particle dipoles are screened. To further validate that, we conducted a series
properties at the fluid interfaces27. Here, we first produce a flat and of experiments with the presence of different amounts NaCl in the
horizontal oil/water interface in a customized microscopic cell that water phase ranging from 0.01 to 10 mM. In all tested conditions, we
can be double sealed (Fig. S1 and Fig. S3), thereby inhibiting any consistently found that g(r) always evolved over time, i.e., increasing
possible evaporation of oil or water that could cause particle drift at the tendency of interfacial particles to access their neighbors, while
the interface; Secondly, unlike common dispersion approach, we the presence of sufficient salts in the water phase would greatly affect
directly dispersed microspheres at the interface using microinjec- the stability of interracial particles and eventually form colloidal
tion (Fig. S2), in which the quantity of spreading solvent can be aggregates at interfaces, as shown in Figure 1d.
reduced to as low as 5 3 1026 (v5v) in the system. These two experi-
mental improvements enable us to produce a highly stable and clean Attenuation of repulsions at oil/water interfaces. We have clearly
oil/water interface, allowing the long-term (.36 hours) measure- shown that the g(r) of interfacial particles possess a time-dependent
ments that are essential to resolve precisely weak interactions be- feature at octane/water interfaces either in absence or presence of a
tween interfacial particles. monovalent salt in the water phase. Assuming that the system stays at
Next, we monitored time evolution of monodispersed polystyrene quasi-equilibrium state in the every fast observing time (see the
microspheres (1.0 mm) trapped at an octane/H2O interface in methods), we can quantitatively evaluate the pair potentials (U(r))
absence or presence of NaCl in the water phase over 10 hours. of interfacial particles at an octane/H2O or octane/NaCl interface by

Time [hr]
A B 8.0

6.0
1.0 1.0
g(r)

g(r)

4.0
0.5 0.5
2.0

0.0 0.0 0.0

0 5 10 15 20 25 0 5 10 15 20
r (μm) r (μm)
C D
l
aC
H2 O

N
M
m

l
aC
10

20 μm
M
m

l
aC
1
10 mM NaCl

N
M
m

l
aC
1
0.

N
M
m

O
01

2
H
0.

0 2 4 6 8
0.7 hr 2.5 hr 6.0 hr Time (hr)
Figure 1 | Time evolution of single colloidal particles trapping at oil/water interfaces in absence or presence of a monovalent salt (NaCl) in the
water phase. Time evolutions of radius distribution functions (g(r)) of polystyrene microspheres (1.0 mm) trapping at an (a) octane/H2O interface
or (b) an octane/10 mM NaCl interface as well as their bright-field images (c), where 5% isopropanol was used as the spreading solvent, colors represent
time lapse. (d) Stability of polystyrene microspheres trapping at oil/water interfaces with presence of different amounts NaCl in the water phase ranging
from 0.01 to 10 mM, where solid symbols in the shadow represent emerging of colloidal aggregates at oil/water interfaces.

SCIENTIFIC REPORTS | 4 : 4778 | DOI: 10.1038/srep04778 2

 www.nature.com/scientificreports

their g(r) using hypernetted chain (HCN) approximation, which was dissolved in it. We find that using methanol as alternative spreading
proved to be a better choice to resolve the long-ranged potentials28. solvent also lead to the time-dependent feature of g(r) or U(r) either
Note that in the case of octane/10 mM NaCl interface, we evaluate at octane/H2O or at octane/10 mM NaCl interface, which are similar
U(r) of the trapped microspheres before the emerging of colloidal to that by using isopropanol as the spreading solvent (Fig. S4), sug-
aggregates, which ensure that the system stays at quasi-equilibrium gesting that the swelling effect cannot be responsible for the attenu-
state. Figure 2a and 2b clearly indicated that 1) exclusive repulsions ation of U(r). Moreover, we noticed h 5 116u 6 10 at octane/10 mM
are detected at the octane/H2O or the octane/10 mM NaCl interface, NaCl, which was slightly larger than h 5 100u 6 7 measured at
in which U(r) monotonically decay with increasing separation of octane/H2O interface. This difference may result from the surface
microspheres; 2) The magnitude of evaluated U(r) were attenuated charges locating at hemisphere immersed in the water phase because
gradually over time, which is independent of whether presence of their image charges have same sign that would push interfacial part-
NaCl in water. It is desirable to note that the long-ranged attractions, icles toward water phase. In this way, screening these surface charges
which likely arise from capillary forces19,29 and have been often obser- by adding salts would lead to more proportion of microspheres res-
ved among the microspheres trapped at air/water interface30,31, have iding in the oil phase and less amount protruding into the aqueous
not been detected in our experiments. phase10,33.
A number of factors can in principle lead to the time-dependence
of U(r), including 1) the gradual change of particle wettability stem- Discharging of interfacial particles at oil/water interfaces. We next
ming from physical aging of contact line22; 2) the local deformation of ask whether the attenuation of residual charges can be responsible for
interfaces attributed by electrodipping forces32 and 3) the spreading the attenuation of U(r). First, we neutralized the charges that were
solvent (isopropanol) may swell these microspheres. To address dissociated from the surface of carboxylate-modified microspheres
these factors, a high-resolution confocal microscope was used to trapped at octane/H2O interface by adjusting pH value in water to
investigate the microstates of these microspheres trapped at a vertical 4.85, as indicated by the Zeta potential (j) of these particles
octane/water interface. Confocal images clearly show that i) three approaching zero (j 5 21 mV 6 3) in the bulk condition (pH 5
phase contact angles (h) of interfacial particles were nearly independ- 4.85); second, we utilized microelectrophoresis (Fig. S3c) to monitor
ent of time either at octane/H2O or at octane/10 mM NaCl interface the electrophoretic mobility of these interfacial particles. By carrying
(Figs. 2c and 2d), which thereby cannot be responsible for the attenu- out theoretical calculation, we find that our microelectrophoresis
ation of U(r); ii) No significant local deformations have been device can generate the static electric filed as high as 2500 to
observed at interfaces (Fig. 2d), which may explain why we did not 5200 V/m near the interface when a 60 V DC was applied, as
detect long-ranged attractions among these microspheres; iii) No shown in Figure S5, which can push the interfacial particles that
significant swelling of those interfacial particles have been observed. carried around 100 e charges toward the anode with a velocity
To further address this issue, we dispersed those microspheres at ranging from 5 to 11 mm/s. In our experiments, we find that
octane/H2O or at octane/10 mM NaCl interface by using methanol initially the particles always migrated to anode when a 60 V DC
as the spreading solvent because polystyrene is unlikely to be was applied (Mov. S1) with an average velocity of 3 mm/s, but no

Time (hr)
(a) (b) 8.0
6.0 6.0
6.0
U(r) / kBT

U(r) / kBT

4.0 4.0
4.0
2.0 2.0

2.0
0.0 0.0
0.0
5 10 15 20 25 5 10 15 20 25
(c) r (μm) (d) r (μm)
1.0 μm
H 2O

100
θ

10 mM NaCl

60

20

0.7 hr 2.5 hr 6.0 hr 0.7 hr 6.0 hr

Figure 2 | Attenuation of repulsions among single colloidal particles at oil/water interfaces in absence or presence of NaCl in the water phase. Time
evolutions of pair potentials (U(r)) of polystyrene microspheres trapping at an (a) octane/H2O interface or (b) an octane/10 mM NaCl interface, where
symbols or lines represent U(r) or their fitting curves based on equation (1) and (2), colors represent time lapse. (c) Time dependence of three phase
contact angles (h) of polystyrene microspheres trapping at an octane/H2O (red bars) or at an octane/10 mM NaCl (cyan bars) interfaces, and their
confocal images (d), where blue colors or green lines respectively represent emissions of coumarin in aqueous phase or boundaries determined by finding
maximum gradients of intensities.

SCIENTIFIC REPORTS | 4 : 4778 | DOI: 10.1038/srep04778 3

 www.nature.com/scientificreports

(a) (b)
120
3.5

200 μm
τ (hr)
2.5

1.5
80
Qres (e)

5 ms 10 ms 15 ms
(c)

l
aC
O

aC

l
l
aC
aC
2

N
H

N
N
* * *

% % % % % % % %
M
M

Isopropanol %
m

20 15 10 7.5 5 2.5 1 0
m
m
01

10
1
0.
0.
40

0
0 2 4 6 8 0 1.0 0 20 40 60 101 102 103
Time (hr) Rc (mm) vj (mm/s) d (mm)
2.0
Figure 3 | Discharging and spreading of single colloidal particles at oil/water interfaces. (a) Fading of residual charges on polystyrene microspheres
trapping at oil/water interfaces in presence of different amounts NaCl in the water phase, where solid lines represent their fitting by single exponentially
decay, and inset of (a) shows their characteristic decay time (t). Dispersion of particle suspensions containing different amounts of isopropanol
ranging from 1 to 20% at an octane/H2O interface: (b) Spinning-disk confocal images acquired with a certain exposure time ranging from 5 to 20 ms
(containing 2.5% isopropanol in the suspension), (c) sliding velocities (vj) or radius convection zone (Rc) or sliding distance (d), where the
symbol (*) presents that no jets had been observed at the interface.

significant electrophoretic mobility was observed after 4 hours, (Fig. 3a). To gain insight into the discharging effect, we carried out
demonstrating that the residual charges fade out in several hours. a series of experiments with the presence of different amount NaCl in
To quantify the discharging of interfacial particles, we fitted the water. In all tested conditions, we found that Qres always decayed over
repulsive potentials measured at the octane/10 mM NaCl interfaces time (Fig. 3a), and can be fitted with a single exponential decay
using a theoretically pair potential (Ures(r)) that accounts for a point Q0res expð{t=tÞ, where Q0res is the Qres at t R 0, t is the characteristic
charge (Qres) located in the oil phase at a distance (a)23, decay time. Inset of Figure 3a clearly indicated that increasing ionic
o  strength in water phase would lead to much faster discharging of
Ures ðrÞ l lBo microspheres at the oil/water interface, suggesting that the aqueous
^4pe{2 Q2res B { pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi ð1Þ
kB T r 4a2 zr2 phase provides a discharging path for the interfacial particles.

where kB is the Boltzmann constant, T is the absolute temperature, Spreading of particles at oil/water interfaces. It is reasonable to
lBo ~e2 4peoil e0 kB T is the Bjerrum length in octane, e is the element- anticipate that these charges are generated when the microspheres
ary charge, and e0 is the vacuum permittivity. Figure 2B indicated arrive at the oil/water interface, giving the fact that microspheres or
that equation (1) can fit the experimental data perfectly, further their suspensions could not carry any unstable charges before the
demonstrating that these long-ranged repulsions found at octane/ dispersion. Therefore, a high-speed spinning-disk confocal micro-
10 mM NaCl interface is a direct result from electrostatic interac- scope was applied to monitor the dispersion of fluorescent micro-
tions and can be solely attributed to the residual charges. spheres at an octane/H2O interface, which ensures that only
Importantly, the quantity of residual charges decreases gradually interfacial particles can been seen. Surprisingly, we observed that
over time, as shown in Figure 3a. In addition to residual charges these microspheres jet along the interface with extremely high
leading to the charge-charge repulsion acting through unscreened speeds (20–60 mm/s) at the moment of contact with the interface
oil phase, the interfacial dipoles would provide stable repulsions (Mov. S2). The sliding velocity (vj) were estimated by measuring the
(Udip(r)) at the octane/H2O interface, which can be theoretically travel distances of the corresponding flight paths in images that were
described as20 captured by an EMCCD camera with specific exposure times, as
 w  2 shown in Figure 3b. Note that the interfacial jet is not a result of
Udip ðrÞ l lD microinjection due to the fact that particles injected to either H2O or
^8pe{2 Q2dip B ð2Þ
kB T r r octane do not jet. Our findings are similar to the case of dispersion of
small particles at air/water interfaces that has been reported by Singh
where Qdip is the electric quantities of dissociated surface charges, lBw et.al, in which they34 proposed that the capillary force pull particles
is the Bjerrum length in water, and lD is the Debye screening length. into the interface, causing them to accelerate to a high velocity. How-
To evaluate Qres at octane/H2O interfaces, equation (2) is used to first ever, for the oil/water interface, we noticed that 1) trace amounts of
fit the experimental data obtained after 8 hours to find the Udip(r), spreading solvents in the suspensions, such as isopropanol, was
subsequently equation (1) is used to resolve the Qres by excluding essential to induce the interfacial jets; 2) vj was largely correlated
Udip(r) in the original data. The corresponding fitting curves are with the quantity of spreading solvent, as shown in Figure 3c.
shown in Figure 2a. Similar to the octane/10 mM NaCl interface, These finding suggest that spreading solvents play a significant role
we found that Qres decayed monotonically over time, but their in facilitating the spreading of particles at the interface; this is known
discharging is slower than that found for octane/10 mM NaCl as the Marangoni effect35.

SCIENTIFIC REPORTS | 4 : 4778 | DOI: 10.1038/srep04778 4

 www.nature.com/scientificreports

To further demonstrate that the Marangoni effect can lead jetting

of particles at the interface, we captured the larger-scale images of (a) (b)
these interfacial jets during the microinjections. We find that 1) the 250

PMMA
20% 250
microinjection of isopropanol (aq.) at the interface can strongly 10%

methanol
7.5%

Q0res (e)
Q0res (e)

ethanol
induce the convections, as shown in Figure S6a, which is in a good

pipetting
agreement with the Marangoni effect that also can induce convection 150 5% 150
at the interface; 2) the characteristic radius (Rc) of the convection
2.5%
zone are insensitive with the quantity of isopropanol ranging from

Gold
50 1% 50
2.5%–10%, as shown in Figure 3c. Next, we estimated the sliding
velocity of particles at the interface in different conditions by using *
a simple model (see the method) that only accounts the surface Isopropanol %
tension driven phenomena, as

vj ^
½cðcb Þ{cð0Þdp
ð3Þ
(c) Oil

es
gRc

arg
where c(cb) or c(0) is the surface tension of isopropanol (aq.)/octane

Ch
or water/octane interface, cb is the concentration of the isopropanol

of
in particles suspensions, dp is the particle size and g is the interfacial

er
viscosity. Figure S6b clearly indicated that the predicted sliding velo-

mb
cities (vj0 ) agree well with the experimental observations (Fig. 3c) at

Nu
the conditions of containing 2.5–10% isopropanol in the particle
suspensions, demonstrating that jetting of particles at the interface
mainly arose from the gradient of surface tensions.

Charging particles at oil/water interfaces. Previous findings indi- Time

cated that the gradients of surface tension would drive particles
rubbing to the interface, which inspire us to conjecture that fast Figure 4 | Charging of single colloidal particles at oil/water interfaces.
rubbing may induce the electrons transfer from one insulator Charging polystyrene microspheres (cyan bars) at an octane/10 mM NaCl
(octane) to another (polystyrene microspheres). If yes, the quantity interface using distinct suspensions: (a) containing different amounts of
of residual charges building at interfacial particles should highly isopropanol ranging from 1 to 20%, (b) containing 5% methanol or
depend on the dispersion processes. Thus, we utilized aforemen- ethanol or using pipetting for dispersions. Charging
tioned methods to evaluate the Q0res with distinctive dispersion con- polymethylmethacrylate (red bar) or gold microspheres at an octane/
ditions, including varying the isopropanol content in particles 10 mM NaCl interface, where the symbol (*) presents that the long-ranged
suspensions during the microinjection or directly pipetting particle repulsion had not been observed. (c) Cartoon schematically shows
suspensions to the interfaces (Fig. S7). Note that changing of charging and discharging of single colloidal particles at an oil/water
isopropanol content from 1% to 20% in suspensions only slightly interface.
alters the wettability of particles at the interface (Fig. S8) because the
isopropanol concentrations in our systems are very low27. Figure 4a well with Lowell’s experimental results38. Moreover, we find that
indicated that Q0res are positively correlated to the content of isopro- dispersion particles by pipetting lead particles much faster jetting
panol when the particles were injected to the interface, confirming on the interface (160 mm/s) compared with the microinjection,
that dispersion conditions strongly affect Q0res . Moreover, we find however the interfacial flows can only be lasted for 0.3 s. As the
that the dispersion of particles by direct pipetting leads to less Q0res result, d is around ,50 mm, which is well explained why dispersion
(Fig. 4b) comparing to that by microinjections. particles by pipetting result in less Q0res comparing to that by
To further elucidate these observations, we quantified the triboe- microinjection.
lectrification effect by using two surface states theory (TSST). In the To further validate that these residual charges are transient
TSST36, rubbing one insulator to another can induce electrons trans- charges, we dispersed gold particles at the oil/water interface in pres-
fer from one surface to the other, in which the direction of transfer ence of 10 mM NaCl, and we did not observe the long-ranged repul-
depends on triboelectric series37; and the net charges can be quan- sions. This indicates that no residual charges exist in the hemisphere
tified as (see the methods), of gold particles immersed in oil, as shown in Figure 4b. We also
investigated factors that affect the charging of particles at interfaces,
Q0res ^Qmax ½1{ expð{d=lÞ ð4Þ by dispersing colloidal particles at oil/octane interfaces with the
where Qmax is the maximum charge quantities, d is the sliding dis- presence of 10 mM NaCl in water by using various experimental
tance and l is the characteristic distance. Equation (4) clearly indi- conditions, e.g.; using methanol or ethanol as spreading solvents or
using polymethylmethacrylate (PMMA) microspheres. Figure 4b
cated that Q0res are positively correlated to the sliding distance rather
than the sliding velocity. Therefore, we estimated the distances of clearly indicated that the value of Q0res depends on experimental
particles sliding on the interface by measuring the time that particles conditions. However, the charging of colloidal particles at oil/water
stay at the convection zone because we have known the sliding velo- interfaces is quite general.
city. We find that the time is ranged from 2 to 8 s during the micro-
injection, in which the characteristic time (tc) is around ,5 s that is Discussion
insensitive to the content of isopropanol, d thus can be estimated by The key results extracted from the data presented above provide a
vjtc, as shown in Figure 3c. These results readily explained why Q0res new insight into interactions and assembly of solid particles at oil/
are positively correlated to the content of isopropanol in the particle water interface. Moreover, our result can be used to explain some
suspensions (Fig. 4a). Most importantly, our results displayed that physical behaviors that are not well understood. For instance, Park
changing d from 60 to 250 mm can strongly affect the triboelectri- et al. directly measured the repulsive forces between particle pairs
fication, which suggests that l is in the range of ,100 mm that agrees trapped at oil/water interface using optical tweezers. They found that

SCIENTIFIC REPORTS | 4 : 4778 | DOI: 10.1038/srep04778 5

 www.nature.com/scientificreports

the repulsions actually are heterogonous26, highly depending on the bright-field microscope (Olympus IX-81) with a 310 objective was used to monitor
the microinjection (Fig. S2c); Octane/NaCl interface was produced by further adding
individual pairs to be selected. In addition, the amount of sodium
NaCl in water via microinjection (Fig. S2b), where concentration of NaCl was
dodecyl sulfate (SDS) presence in the water phase can gradually determined with the volume ratio of NaCl (100 mM aq.) and H2O. Finally, the small
attenuate the strength of this repulsion at the interface25. The finding chamber containing the octane/H2O or octane/NaCl interface was sealed with a (20 3
of charging of single particles at interface can readily explain the 20 3 1.1 mm) glass slide and the whole microscopic cell was further sealed with a
heterogonous feature of these repulsive forces because these pro- polystyrene lid, as shown in Figure S3A. Vertical octane/H2O or octane/NaCl
interface was prepared in a device, as shown in Figure S3b, composing a 30 mm PTFE
duced charges largely relate to the sliding distances and which are spacer and two hydrophobic modified coverslips whose water contact angles
varied for individual particles, whereas attenuation of this repulsion approach to ,90u. Microspheres were dispersion at the vertical oil/water interface via
over time are most likely related the discharging of these particles at the same microinjection procedure described above.
interface.
In summary, we have shown that single colloidal particles can be Tracking of microspheres at horizontal oil/water interfaces. An Olympus IX81
inverted-microscope equipped with a 403 objective and a sCMOS camera (Andor
spontaneously charged when spreading solvents are used to disperse
Zyla), was used to acquire bright-field images of microspheres trapped at horizontal
them at oil/water interfaces. This is due to the gradients of surface octane/H2O or octane/NaCl interface with a frame rate 0.5 fps. A typical dataset,
tension driving particles sliding at the interface, which in turn lead to usually containing 2400 images with a field size of 409 3 346 mm2, has recorded
triboelectric charges transferring from oil to the colloidal particles. trajectories of single microspheres (ri(t)) at oil/water interface. The datasets were
These residual charges dominate the long-ranged electrostatic repul- continuously collected every 20 minutes over 10 hours, starting from the beginning of
dispersions. These experiments were repeated multiple times in each identical
sions, and prevent the interfacial particles against aggregation even conditions, thus over 300 datasets (.720,000 images) were acquired and analyzed in
when the conventional dipole-dipole repulsions are screened by add- total, in which the 2560 3 2160 16-bit greyscale images were firstly converted to
ing monovalent salts in the aqueous phase. However, these tribo- binary images for detection of microsphere with a standard image processing
electric charges are unstable, and can fade out in several hours, thus algorithm coding by MATLAB, and the x-y positions of microsphere centroids were
gradually attenuating the repulsions among interfacial particles. As subsequently determined and linked individually over time by using a particle
tracking algorithm39.
the interfacial particles are slowly discharged, they become unstable
when their dipoles are screened. Charging and discharging of single Measurements of radius distribution functions and pair potentials of
colloidal particles at oil/water interfaces were graphically summar- microspheres at horizontal oil/water interfaces. The trajectories of single
ized in Figure 4c. Charging of colloids at oil/water interfaces plays microspheres (ri(t)) in each dataset was used to calculate the radius distribution
remarkable roles in interfacial interactions of colloids, and our stud- function (g(r)) by time-averaging pair correlations (Si?jd(r1 2 ri)d(r2 2 rj)/n) in
every moment over 20 minutes, where r~jr1 {r2 j is the spatial separation and n is the
ies suggest that such phenomenon is general. It may prove to be a surface density of microspheres. The pair potentials U(r) was numerically evaluated
novel mechanism to induce transient charges on biological macro- from g(r) using the hypernetted chain (HCN) approximation, in which U(r) 5 2kBT
molecules, such as DNAs, proteins and viral particles at fluid ln[g(r)] 1 nkBTI(r) was applied, where kB is the Boltzmann constant, T is the absolute
interfaces. temperature, and convolution integral I ðr Þ~½g ðr 0 Þ{1{nI ðr Þ½g ðjr0 {rjÞ{1d2 r 0
can be obtained iteratively, starting with I(r) 5 0 28. Note that U(r) can be evaluated
from g(r) only when the system stay at the equilibrium state or at quasi- equilibrium
Methods state. In our case, the interfacial particles can be either at stable or at unstable state,
Materials. The coverslips (25 3 25 3 0.17 mm, Fisher) were pre-washed by which depends on time and whether existing sufficient salts in water phase (Fig. 1d).
sonication (5 min) in acetone, ethanol and ultrapure water (18.2 MV?cm Millipore), Therefore, only the g(r) at the stable region can be used to evaluate U(r) by assuming
respectively. Afterward, they were rinsed with 5% (aq.) hydrogen fluoride for 10 s, that the system stays at quasi- equilibrium state.
and then washed with water immediately. The cleaned coverslips were finally dried
with nitrogen and further treated (5 min) with a plasma cleaner before assembly on Microelectrophoresis. Two platinum-wire electrodes (100 mm in diameter) with a
the microscopic sample cell. The resultant static contact angles of water on these 0.3 mm separation were placed in the octane, ,110 mm above from the horizontal
surfaces almost approach to 0u. The hydrophobic modified coverslips were prepared oil/water interface, as shown in Figure S3c. A 60 V DC was applied to generate a static
by treating the cleaned coverslips with [(chloromethyl)phenylethyl]- electric field along the interface, and bright-field microscope equipped with a 203
Trimethoxysilane (CTMS) in toluene. Carboxylate-modified fluorescent objective was used to monitor the electrophoretic mobility of microspheres (Mov. S1)
microspheres (polystyrene 1.0 mm in diameter, Invitrogen) were purified by trapped at a horizontal octane/H2O, where the pH of aqueous phase was adjusted to
resuspending 2 mL concentrated suspensions (,2%) in 1.5 mL ultrapure water, then 4.85 by HCl (aq.). Note that, in this set up, the electric field near the interface will
washed by centrifugation to remove stabilizer. The stabilizer-free microspheres were greatly reduce by the field that was induced by their image charges. To assess the
finally dispersed in spreading mediums, containing different amounts (1 to 20%) of effective electric field near the interface, we have carried out a theoretical calculation
isopropanol, methanol or ethanol in water, to a fixed volume fraction ,1 3 1024, by using COMSOL, as shown in Figure S5a. Our results indicate that the resultant
which were monitored by their optical densities at 600 nm (O.D.600). The averaged electric field is ranged from 2500 to 5200 V/m, depending on the distance apart from
surface charge density of the microsphere was around 1.3 mC?cm22, estimating by the interface, as shown in Figure S5b.
their Zeta potentials (j 5 255 mV 6 3) measured in a buffer containing of 10 mM
NaCl. Note that none of cross-linked polystyrene microspheres may likely contain
short PS chains that dissolve in the spreading mediums. Therefore, these PS Measurements of contact angles of microspheres at vertical oil/water interfaces.
microspheres were directly dissolved in the tetrahydrofuran (THF) to determine their 15 mM coumarin was added in water for imaging of oil/water interface. Fluorescent
molecular weight distribution. Next, their average molecular weight (Mn 5 1.5 3 microspheres trapped at vertical oil/water interface were examined with a confocal
105 g/mol, Mw 5 6.8 3 105 g/mol) as well as polydispersity (P.DI 5 4.2) were microscope (Olympus FV1000) equipped with a 1003 oil objective, in which
measured by using gel permeation chromatography (GPC). The GPC result indicated coumarin and microspheres were excited simultaneously with 405 and 543 nm lasers
that these PS microspheres do not contain short PS chains, in which the molecular to acquire confocal images via two distinct emission channels (425–475 nm and 555–
weight of smallest fraction is ,1.0 3 104 g/mol. Polymethylmethacrylate (PMMA) 655 nm). The boundaries of the oil/water interface and microspheres were
microspheres (1.0 mm in diameter) were a gift from Xu’s lab in The Chinese respectively evaluated by finding the maximum intensity gradients from their images.
University of Hong Kong. Gold colloids (300 nm in diameter) were obtained from Three phase contact angle (h) of single interfacial particle was geometrical measured
Xiong’s Lab in University of Science and Technology of China. The ultrapure water or using their determinate boundaries. These experiments were repeated multiple times
ultrapure octane (electronic grade, .99.999%, Sigma-Aldrich) was filtrated with a in each identical conditions, thus over 500 confocal images, typically containing 2 , 5
syringe filter (PTFE 0.2 mm Millipore) to remove dusts before the preparation of interfacial particles, were acquired and analyzed in total, enabling us to actually
octane/H2O interfaces. evaluate the h in each conditions by averaging twenty single results at least.

Preparation of oil/water interfaces and dispersion microspheres at interfaces. Observation of spreading of microspheres at horizontal oil/water interfaces. A
Horizontal octane/H2O or octane/NaCl interface was prepared in a customized spinning-disk confocal microscope (Andor Revolution) equipped with a 103
microscopic cell, which was assembled with a coverslip, a PET spacer with thickness objective was used to monitor the spreading of fluorescent microspheres at horizontal
of 110 mm and a cup-like glass cell, as shown in Figure S1a. First, ultrapure water was oil/water interface using different dispersion methods, including microinjection
gently added in bottom of cell to reach a height of 110 mm, and then ,2.0 mL (Mov. S2) of pipetting (Fig. S7). Fluorescent microspheres were excited with a 561 nm
ultrapure octane was slowly added to form a horizontal octane/H2O interface that was laser. High-speed confocal images were collected through an emission filter (600 6
stuck with the PET spacer exactly (Fig. S1b); Second, 0.2 mL microsphere suspension, 25 nm). The dispersion speeds (sliding velocities) of microspheres at oil/water
typically in 5% isopropanol (aq.), was directly microinjected at the interface with a interface were assessed by measuring the length of their flowing light paths remained
flow rate of 0.1 mL/min using a microinjector (Fig. S2a) composing of a XYZ in images, capturing by an EMCCD (Andor iXon897) with a certain exposed time
manipulator, a syringe pump and a glass capillary (15–30 mm in diameter), in which ranging from 1 to 20 ms.

SCIENTIFIC REPORTS | 4 : 4778 | DOI: 10.1038/srep04778 6

 www.nature.com/scientificreports

Estimation the sliding velocity of particles at the interface. Marangoni effect 25. Park, B. J. et al. Direct measurements of the effects of salt and surfactant on
produces a slip velocity in the tangential direction on a fluid-fluid interface when interaction forces between colloidal particles at water-oil interfaces. Langmuir 24,
gradients in the surface are present40. As a result, the tangential force per unit area (fs) 1686–1694 (2008).
is given by41 fs 5 =sc, where =s denotes the surface tension (c) gradients. In one hand, 26. Park, B. J., Vermant, J. & Furst, E. M. Heterogeneity of the electrostatic repulsion
the tangential force acting on the particles (Fs) at interface is Fs ^dp2 +s c. On the other between colloids at the oil-water interface. Soft Matter 6, 5327–5333 (2010).
hand, the viscous force is ^dp gvj . Therefore, one can find vj ^dp +s c=g. =sc can be 27. Maestro, A., Bonales, L. J., Ritacco, H., Rubio, R. G. & Ortega, F. Effect of the
further estimated by +s c^½cðcb Þ{cð0Þ=Rc (for the microinjection) or by spreading solvent on the three-phase contact angle of microparticles attached at
+s c^½cðcb Þ{cð0Þ=rd (for the pipetting), where rd is the radius of droplet. c(cb) is fluid interfaces. Phys. Chem. Chem. Phys. 12, 14115–14120 (2010).
expected to close to the surface tension of interface between the isopropanol aqueous 28. Behrens, S. H. & Grier, D. G. Pair interaction of charged colloidal spheres near a
solution and air42. charged wall. Phys. Rev. E 64, (2001).
29. Boneva, M. P., Danov, K. D., Christov, N. C. & Kralchevsky, P. A. Attraction
Quantification of triboelectrification. In the two state surface theory36, rubbing a between particles at a liquid interface due to the interplay of gravity- and electric-
spherical insulator to a insulate surface would induce net charges (Q) on the spherical field-induced interfacial deformations. Langmuir 25, 9129–9139 (2009).
insulator: Q~Q0 l1 ½1{ expð{d=l1 Þ{Q0 l2 ½1{ expð{d=l2 Þ (eq. 5), where 30. Chen, W., Tan, S. S., Ng, T. K., Ford, W. T. & Tong, P. Long-ranged attraction
Q0 jl1 {l2 j is the maximum charges quantity, l1 or l2 are the characteristic distances between charged polystyrene spheres at aqueous interfaces. Phys. Rev. Lett. 95,
for charges transfer into the lower and out of the upper states of the spherical insulator (2005).
respectively. Note that the material of polystyrene locate at the rather negative 31. Chen, W. et al. Measured long-ranged attractive interaction between charged
positions in triboelectric series43, which suggests that charges transfer into the lower polystyrene latex spheres at a water-air interface. Phys. Rev. E 74, (2006).
state is much easier than charges transfer out of the upper state; i.e., l1 ?l2 . 32. Danov, K. D., Kralchevsky, P. A. & Boneva, M. P. Electrodipping force acting on
Therefore, equation (5) can be rewritten as equation (4), where Qmax 5 Q0l1. solid particles at a fluid interface. Langmuir 20, 6139–6151 (2004).
33. Wang, H. Z., Singh, V. & Behrens, S. H. Image charge effects on the formation of
pickering emulsions. J. Phys. Chem. Lett. 3, 2986–2990 (2012).
1. Binks, B. P. Particles as surfactants - similarities and differences. Curr. Opin. 34. Singh, P., Joseph, D. D., Gurupatham, S. K., Dalal, B. & Nudurupati, S.
Colloid Interface Sci. 7, 21–41 (2002). Spontaneous dispersion of particles on liquid surfaces. Proc. Natl. Acad. Sci.USA
2. Leal-Calderon, F. & Schmitt, V. Solid-stabilized emulsions. Curr. Opin. Colloid 106, 19761–19764 (2009).
Interface Sci. 13, 217–227 (2008).
35. Scriven, L. E. & Sternling, C. V. Marangoni effects. Nature 187, 186–188 (1960).
3. Hunter, T. N., Pugh, R. J., Franks, G. V. & Jameson, G. J. The role of particles in
36. Lowell, J. & Truscott, W. S. Triboelectrification of identical insulators 2 theory and
stabilising foams and emulsions. Adv. Colloid Interface Sci. 137, 57–81 (2008).
further experiments. J. Phys. D-Appl. Phys. 19, 1281–1298 (1986).
4. Dickinson, E. Food emulsions and foams: Stabilization by particles. Curr. Opin.
Colloid Interface Sci. 15, 40–49 (2010). 37. Lacks, D. J. & Sankaran, R. M. Contact electrification of insulating materials.
5. Bos, M. A. & van Vliet, T. Interfacial rheological properties of adsorbed protein J. Phys. D-Appl. Phys. 44, 453001 (2011).
layers and surfactants: a review. Adv. Colloid Interface Sci. 91, 437–471 (2001). 38. Lowell, J. & Truscott, W. S. Triboelectrification of identical insulators 1 an
6. Krafft, M. P. & Riess, J. G. Highly fluorinated amphiphiles and colloidal systems, experimental investigation. J. Phys. D-Appl. Phys. 19, 1273–1280 (1986).
and their applications in the biomedical field. A contribution. Biochimie. 80, 39. Jin, F., Conrad, J. C., Gibiansky, M. L. & Wong, G. C. L. Bacteria use type-IV pili to
489–514 (1998). slingshot on surfaces. Proc. Natl. Acad. Sci. USA 108, 12617–12622 (2011).
7. Herzig, E. M., White, K. A., Schofield, A. B., Poon, W. C. K. & Clegg, P. S. 40. Levich, V. G. & Krylov, V. S. Surface-tension-driven phenomena. Annu. Rev.
Bicontinuous emulsions stabilized solely by colloidal particles. Nat. Mater. 6, Fluid. Mech. 1, 293–316 (1969).
966–971 (2007). 41. Nikolov, A. D. et al. Superspreading driven by Marangoni flow. Adv. Colloid.
8. Verma, A. et al. Surface-structure-regulated cell-membrane penetration by Interface. Sci. 96, 325–338 (2002).
monolayer-protected nanoparticles. Nat. Mater. 7, 588–595 (2008). 42. Vazquez, G., Alvarez, E. & Navaza, J. M. Surface-tension of alcohol plus water
9. Verma, A. & Stellacci, F. Effect of surface properties on nanoparticle-cell from 20-degrees-c to 50-degrees-c. J. Chem. Eng. Data 40, 611–614 (1995).
interactions. Small 6, 12–21 (2010). 43. Diaz, A. F. & Felix-Navarro, R. M. A semi-quantitative tribo-electric series for
10. Leunissen, M. E., van Blaaderen, A., Hollingsworth, A. D., Sullivan, M. T. & polymeric materials: the influence of chemical structure and properties.
Chaikin, P. M. Electrostatics at the oil-water interface, stability, and order in J. Electrost. 62, 277–290 (2004).
emulsions and colloids. Proc. Natl. Acad. Sci. USA 104, 2585–2590 (2007).
11. Bausch, A. R. et al. Grain boundary scars and spherical crystallography. Science
299, 1716–1718 (2003).
12. Law, A. D., Buzza, D. M. A. & Horozov, T. S. Two-dimensional colloidal alloys. Acknowledgments
Phys. Rev. Lett. 106, (2011). This work was supported by the National Program on Key Basic Research Project
13. Tavacoli, J. W., Thijssen, J. H. J. & Clegg, P. S. Particle-stabilized oscillating diver: a (2012CB933802), the National Natural Science Foundation of China (21104071),
self-assembled responsive capsule. Soft Matter 7, 7969–7972 (2011). (21274141), and the Hong Kong Special Administration Region (HKSAR) General
14. Binks, B. P. & Clint, J. H. Solid wettability from surface energy components: Research Fund (CUHK403210, 2130237).
Relevance to pickering emulsions. Langmuir 18, 1270–1273 (2002).
15. Binks, B. P. & Lumsdon, S. O. Pickering emulsions stabilized by monodisperse
latex particles: Effects of particle size. Langmuir 17, 4540–4547 (2001). Author contributions
16. Oettel, M. & Dietrich, S. Colloidal interactions at fluid interfaces. Langmuir 24, P.G. performed the experiments. F.J. and Y.L. analyzed data. X.C.X., T.N. and F.J.
1425–1441 (2008). contributed jointly to data interpretation and manuscript preparation. All authors reviewed
17. Dickinson, E. Adsorbed protein layers at fluid interfaces: interactions, structure the manuscript.
and surface rheology. Colloids Surf. B-Biointerfaces 15, 161–176 (1999).
18. Ash, P. A., Bain, C. D. & Matsubara, H. Wetting in oil/water/surfactant systems.
Curr. Opin. Colloid Interface Sci. 17, 196–204 (2012). Additional information
19. Nikolaides, M. G. et al. Electric-field-induced capillary attraction between like- Supplementary information accompanies this paper at http://www.nature.com/
charged particles at liquid interfaces. Nature 420, 299–301 (2002). scientificreports
20. Pieranski, P. Two-dimensional interfacial colloidal crystals. Phys. Rev. Lett. 45, Competing financial interests: The authors declare no competing financial interests.
569–572 (1980).
21. Park, B. J. & Furst, E. M. Attractive interactions between colloids at the oil-water How to cite this article: Gao, P., Xing, X.C., Li, Y., Ngai, T. & Jin, F. Charging and
interface. Soft Matter 7, 7676–7682 (2011). discharging of single colloidal particles at oil/water interfaces. Sci. Rep. 4, 4778;
22. Kaz, D. M., McGorty, R., Mani, M., Brenner, M. P. & Manoharan, V. N. Physical DOI:10.1038/srep04778 (2014).
ageing of the contact line on colloidal particles at liquid interfaces. Nat. Mater. 11,
138–142 (2012). This work is licensed under a Creative Commons Attribution-NonCommercial-
23. Aveyard, R., Clint, J. H., Nees, D. & Paunov, V. N. Compression and structure of NoDerivs 3.0 Unported License. The images in this article are included in the
monolayers of charged latex particles at air/water and octane/water interfaces. article’s Creative Commons license, unless indicated otherwise in the image credit;
Langmuir 16, 1969–1979 (2000). if the image is not included under the Creative Commons license, users will need to
24. Aveyard, R. et al. Measurement of long-range repulsive forces between charged obtain permission from the license holder in order to reproduce the image. To view
particles at an oil-water interface. Phys. Rev. Lett. 88, (2002). a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/3.0/

SCIENTIFIC REPORTS | 4 : 4778 | DOI: 10.1038/srep04778 7