Article

pubs.acs.org/cm

Bulk Scale Synthesis of Monodisperse PDMS Droplets above 3 μm

and Their Encapsulation by Elastic Shells
Nina A. Elbers,*,† Jissy Jose,† Arnout Imhof, and Alfons van Blaaderen*
Soft Condensed Matter, Debye Institute for NanoMaterials Science, Utrecht University, Princetonplein 1, 3584 CC, Utrecht, The
Netherlands
*
S Supporting Information

ABSTRACT: We report several facile, surfactant-free methods

to prepare monodisperse polydimethylsiloxane (PDMS)
droplets in the size range 3−8 μm in water. These methods,
of which the pros and cons are discussed, are extensions of a
procedure described before by our group which focused on
smaller droplet sizes. The PDMS oil droplets are formed by
ammonia catalyzed hydrolysis and condensation of the
monomer dimethyldiethoxysilane (DMDES) in water. One of
the methods entails a seeded growth procedure in which other
oils, such as lower molecular weight hydrocarbons, were found
to be able to swell the PDMS droplets if their solubility in water
was higher than that of the seed droplets. This way, larger droplets with mixed composition could be prepared. It also turned out
to be possible to load the monodisperse droplets with an oil soluble dye. The droplets could be coated with an elastic, partially
permeable, shell formed by cross-linking the PDMS with tetraethoxysilane (TES) in the presence of poly(vinylpyrrolidone)
(PVP) that provided colloidal stability. Besides, the liquid interior of these shells could be changed by solvent exchange.

■ INTRODUCTION
Microcapsules, here loosely defined as partially permeable shells
ques,28−30 emulsion droplets are prepared in a large amount
but with a size distribution above 10% in width. The reverse is
around a gaseous, liquid or solid material and with a size in the true for droplets prepared using microfluidics, which allows for
colloidal and granular domain, represent an active field of an exquisite degree of control over droplet size and
research because they can potentially encapsulate, protect, and composition, but for which production rate, and thus yield, is
distribute materials. As a result, there is a compelling low.31−35 In general, a good control over the capsule size
fundamental as well as industrial interest for which the latter distribution is desirable as the particle’s shell thickness and size
field is as diverse as food science, pharmaceutics, cosmetics, sets the release rate as well as the mechanical breakup of shells
biotechnology, the paint industry, and advanced materials in with, for example, ultrasound.36−40 As a result, there remains a
general.1−8 The most versatile approach for capsule synthesis is need for efficient methods that can produce monodisperse
a templating technique.9 In this technique, a shell is formed template drops with well-controlled sizes and size distributions.
around a sacrificial particle via methods like interfacial Although the focus of our research is to use monodisperse
polymerization,10 polymer precipitation (by phase separa- emulsions droplets to arrive at monodisperse microcapsules,
tion),6,11,12 layer-by-layer self-assembly,8,13 or via locking of there are many applications, in science and industry, of
interfacial colloids14 followed by removal of the core. For the monodisperse emulsions that do not involve shell growth.41,42
sacrificial core, solid particles as well as liquid droplets have For instance, monodisperse emulsions are used in food science,
been used. An important advantage of the use of solid template cosmetics, microreaction technology, as delivery vehicles in
particles (like silica, latex, or gold colloidal particles)15−20 is that
pharmacy, and even for the synthesis of bicolored droplets for
it allows for a large scale synthesis of monodisperse cores in a
electronic paper.43−46 Many important characteristics like
wider range of sizes. However, after deposition of the shell
material, relatively harsh dissolution or heating conditions are rheology, turbidity, texture, interparticle interactions, and
needed to remove the template. This complicates the shelf life are controlled by the particle size (distribution).30,47
production process and adversely affects the quality and In addition, from a fundamental perspective, the interpretation
robustness of the as-prepared particles.21,22 On the contrary, of experimental results is easier for narrowly distributed
liquid template droplets allow for a more facile removal of the emulsions.47
core under milder conditions.23,24 However, a combination of
bulk scale synthesis and monodispersity is scarce when using Received: December 8, 2014
liquid templating droplets.21,25 In conventional methods like Revised: February 9, 2015
ultrasonication26,27 and high-shear emulsification techni- Published: February 9, 2015

© 2015 American Chemical Society 1709 DOI: 10.1021/cm504504t

Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

Figure 1. Schematic illustration of different synthesis procedures for PDMS droplet formation and subsequent shell coating.

An exception, where liquid droplets of high monodispersity, Cross-linked-PolyDiMethoxySilane (TC-PDMS) shells,51,52

high yield, and without the use of surfactants can be prepared, polydopamine capsules,53 colloidosomes,54,55 and even for
has been described by Vincent and co-workers.48 The process polymer microspheres.56,57 In all studies, the accessible size
to arrive at monodisperse polydimethylsiloxane (PDMS) range was limited to about 0.6−3 μm.51−53 However, a wider
emulsion droplets bears a strong resemblance to the famous size range is preferred to be able to fully match the particle size
Stöber process, in which monodisperse silica particles are with the application in mind. Moreover, the only alternative to
generated from tetra-alkoxysilanes, usually tetraethoxysilane monodisperse PDMS droplets is microfluidics which usually
(TES). Stöber synthesis is performed in a mixture of an alcohol does not produce particles with sizes less than 10 μm. Related
and water to which ammonia is added as a catalyst for both the to the size dependency in the final applications, larger (micron-
hydrolysis and condensation reactions. Nucleation in combi- sized) particles allow higher loading capacity for in vitro
nation with early aggregation of siloxane substructures and studies.58 Besides in drug delivery, the size of the particles (1−
subsequent growth, when colloidal stability has been achieved, 20 μm) controls the accessibility of certain types of cells or
result in monodisperse silica sphere (see for example ref 49). tissues that form the target location.14,36 Furthermore, the size
To create the PDMS droplets, the bifunctional monomer of the microcapsules dictates the frequency of the ultrasound
dimethyldiethoxysilane (DMDES) is used for which the when release of the payload is induced by acoustic cavitation.40
hydrolysis and condensation reactions are, again, catalyzed by Finally, the particle size determines the detection strategy that
the base ammonia. Because alcohol is not added to the reaction will be used during encapsulation and release studies. Ensemble
medium, the starting monomer does not completely dissolve in assays are required for small particles based on for example UV-
the basic water and the resulting nucleation, and, most likely,
spectrophotometry,59 whereas detection on a single particle
early aggregation steps of siloxane substructures before growth
level is possible for larger (micron sized) particles in real time
are more complicated and less well investigated compared to
and real space with a confocal microscope.60−63 Attempts to
the Stöber process. Hydrolysis generates slightly acidic silanol
further increase the size of the PDMS droplet template, which
groups which make the monomers water-soluble, while
subsequent condensation reactions reverse this. It has been predetermines the final size of the capsule, has not produced
well established that the name PDMS is actually somewhat of a promising results yet. In principle, the droplet size can be varied
misnomer as the vast majority of the condensed species are up to 5 μm by increasing the ammonia solution concentration.
small oligomers, composed of ‘chains’ of only 4 units.48,50 The However, dynamic light scattering measurements on systems in
fact that an important part of the monomer is initially not which the droplet size was greater than 1 μm showed them to
dissolved in the water phase, but is present as droplets, results be more polydisperse.48 Another way to tune the droplet size
in a dependency of the synthesis conditions on the surface area up to several micrometers is by increasing the DMDES
of these monomer droplets. The surface area of the initial monomer concentration.48,50,53 However, it was found to be
droplets is expected to determine the rate of hydrolysis of the difficult to prepare monodisperse, intact capsules for droplet
monomer. Via hydrolysis of silanol groups, the surface area diameters above 2.4 μm, for which the initial monomer
dependency indirectly controls the increase in ionic strength concentration had to exceed 3% v/v.48,50,53 Another attempt to
and hence the final interdroplet charge stabilization. As a result, increase the capsule diameter was described in ref 64, in which
the final PDMS droplet size and polydispersity is, unfortu- both DMDES and ammonia concentrations were higher than
nately, affected by the total surface area of the initial monomer what was used in ref 51. However, this synthesis procedure was
droplets and therefore to stirring conditions. not found to be robust as it sometimes resulted in a
Due to the high monodispersity, large scale production, and polydisperse system. In short, there is a lack of a reproducible
facile drop removal, the liquid PDMS droplets have been used synthesis procedure for monodisperse PDMS templating
as sacrificial templates for the production of Tetraethoxysilane- droplets, and hence capsules, with diameters exceeding 3 μm.
1710 DOI: 10.1021/cm504504t
Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

The PDMS droplets can be coated with a tetraethoxysilane- The QD’s contained an oleic acid capping layer and were stored in
cross-linked-polydimethoxysilane shell to form TC-PDMS cyclohexane. All chemicals were used as received. Demineralized water
microcapsules. These capsules were synthesized for the first (resistivity 18 MΩcm) was used in all reactions and also for cleaning of
time by Zoldesi and Imhof,51 with sizes in the range 0.6−2 μm. glassware.
Synthesis Methods. Efficient Initial Mechanical Mixing and No
One advantage of TC-PDMS capsules, next to their Agitation during Growth. (Large scale: 750 mL) We prepared
monodispersity and size control, is the possibility to tune monodisperse PDMS emulsion droplets by hydrolysis and con-
their shell thickness, which is beneficial for potential densation of the monomer DMDES, by modifying the original
encapsulation and release studies. In addition, the capsules procedure of Vincent et al.48 An aqueous dispersion (total volume 750
were found to form well-defined anisotropic shapes upon mL) of 22.7% v/v ammonia and 6.8% v/v of DMDES monomer
drying, or via dissolution of the PDMS oil through the shell (for (assuming additivity of volumes) was prepared in a 2 L beaker.
instance in ethanol), by simply tuning the ratio of shell Another typical sample (total volume 720 mL) was prepared at about
thickness to droplet diameter.50,51 From an applied point of the same ammonia concentration (23.7% v/v) but at a reduced
DMDES concentration of 2.9% v/v. Directly afterward, each sample
view these capsules have already shown promise because they
was vigorously mixed using a Turrax homogenizer (IKA’s Ultra-Turrax
are permeable to small polar dye molecules.50 Recently, our with a S25N 10G dispersing element) at a speed of 11,200 rpm.
group further extended the use of these capsules as containers Mixing typically continued for 8−14 min, until most of the DMDES
by efficiently encapsulating and releasing apolar liquids, which was dispersed, and small droplets in a narrow(er) size range were seen
can contain fluorescent tracer molecules, via controlled and under a light microscope. We do want to stress that all samples,
reversible buckling.64 The capsules could even be loaded with especially at the high 6.8% v/v DMDES monomer concentration,
surfactant molecules.65 In addition, these capsules were shown always contained a layer of excess oil at the top after finishing the
to be an interesting new model system to study the physics of Turrax procedure. Despite the efficient mixing, it is therefore most
jammed matter.66,67 As these capsules are highly flexible and likely that not all DMDES was completely dispersed. The turbid
can be fluorescently labeled and index matched easily, their solution, including the excess layer, was poured into a 1 L flask and
allowed to stand undisturbed for at least 2 days before the coating step.
jamming behavior can be quantitatively studied using real space (Small scale: 30 mL) Emulsion droplets were also prepared with the
confocal microscopy in 3D.68 Turrax homogenizer on a smaller scale as compared to the above-
Here, we introduce several methods to make monodisperse, described procedure. An aqueous solution (total volume 29.9 mL) of
stable PDMS droplets with sizes above 3 μm in a high yield. An 22.7% v/v of ammonia and 6.7% v/v of DMDES monomer was
overview of all droplet sizes and polydispersities, obtained using prepared in a 40 mL glass vial. Samples were vigorously mixed using
the various methods, can be found in Table S1 in the the Turrax homogenizer at a speed of 11,200 rpm for about 4−5 min.
Supporting Information, and a schematic illustration of the Droplets were allowed to nucleate and grow while keeping the sample
various synthesis procedures is depicted in Figure 1. The first static, for at least 2 days.
method to make larger monodisperse emulsions is based on a Based on the work described in ref 51, fluorescent shells were
grown around the as-prepared emulsion droplets. Before incorporation
modification of the PDMS synthesis as described in ref 48. We of the RITC dye into the shells, the silane coupling agent APS and dye
used higher DMDES concentrations in combination with more RITC were covalently coupled by mixing 6.5 mg of the dye and 40 μL
efficient mechanical mixing. The second method is based on of APS with 1 mL of anhydrous ethanol, after which this mixture was
hydrolyzing the monomer under conditions where the stirred for at least 12 h in the dark. The shell growth itself was
condensation reactions are not yet important. The third modified as compared to ref 51 by introducing the steric stabilizer
procedure consists of a seeded growth technique, in which polyvinylpyrrolidone (PVP), which adsorbs onto the particle surface,
primary PDMS droplets grow through diffusion of water in accordance with previous work by Graf et al.20 An aqueous solution
dissolved monomer (DMDES) molecules or oil molecules like (total volume 1725 mL) of 2.3 wt % PVP was prepared in a 2 L flask.
hydrocarbons and silicone oil. We further show that oil droplets Hereafter, 550 mL of the as-prepared emulsion (still containing
ammonia) was added while gently stirring the sample. Inspired by the
of all sizes can be used as templates for the growth of a
postaddition method introduced by O’Sullivan et al.,52 both TES (8
tetraethoxysilane-cross-linked-polydimethoxysilane shell. In this mL) and DMDES (8 mL) were added simultaneously during the shell
way, we have extended the range of particle diameters up to growth step that was performed around templating droplets that were
more than 5 μm while maintaining the high yield, the high (at least) 2 days old. The monomers were added slowly over time
monodispersity, the surfactant free synthesis, and facile drop using a syringe pump (KD Scientific, 10 μL/min). The APS/RITC dye
removal. The size range now nicely extends up to the micron solution (0.330 mL) was typically added 5 h after start of the shell
sized capsules that can be formed using microfluidic devices.69 growth. Shells were allowed to grow for at least 1−2 days, while gently

■
stirring the sample, after which the PDMS template was removed by
following the procedure below.
EXPERIMENTAL SECTION The dispersion containing the fluorescent shells was divided over 40
Materials. Dimethyldiethoxysilane (DMDES, 97.0%), the triblock mL sample vials and centrifuged (800 g for 30 min) after which the
copolymer Pluronic P123 (poly(ethylene glycol)20-poly(propylene concentrated supernatant was collected. The concentrated samples
glycol)70-poly(ethylene glycol)20, Mw - 5800 g/mol), Triton X-100, were dispersed in ethanol, which is known to penetrate the shells and
tetraethoxysilane (TES, 98.0%), 3-aminopropyl-triethoxysilane (APS, to dissolve the liquid cores.50,51 Shells were found to remain spherical
99%), rhodamine B-isothiocyanate (RITC), polyvinylpyrrolidone in shape upon transferring about 100 mL of the concentrated
(PVP, Mw-58000 g/mol), ammonia (25 wt % NH3), dimethyl dispersion to 2 L of (anhydrous) ethanol but started to buckle when
sulfoxide (DMSO), octamethylcyclotetrasiloxane (OMCTS, the same amount of sample was dispersed in 1 L of ethanol. At this
C8H24O4Si4), cyclohexane (C6H12, 99.8%), iso-octane (99%), and point we do not know the reason for this buckling process. In
hexadecane (C16H34, 99.0%) were purchased from Sigma-Aldrich. The principle, buckling induced by dilution with ethanol was also reported
dye pyrromethene-567 was purchased from Exciton. Ethanol (100%, by Zoldesi71 for these shells when their ratio between the shell
technical grade) was obtained from Interchema and n-pentane (>99% thickness and the total radius of the particle was less than 0.17. This is
for spectroscopy) from Acros. Organically stabilized CdSe/CdS/ also the regime in which the current study is performed, as shown in
CdZnS/ZnS nanoparticles (QD’s) with a diameter of 6.9 nm were the results section. However, the concentration dependency is
prepared by using the method of ref 70. Light is absorbed by the QD’s indicative for an additional factor. Possibly, this could be electrostatics,
in the 350−544 nm range, and the emission peak is located at 612 nm. as it was shown that the electrolyte concentration in the medium can

1711 DOI: 10.1021/cm504504t

Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

alter the shape of elastic, closed shells.72 Once shells were successfully The preformed silicone oil droplets were grown larger using a
transferred to ethanol, several washing steps with ethanol were seeded growth procedure with a second oil. The procedure for samples
performed in order to remove all PDMS oil during which no ethanol A and B was the following. Aqueous solutions (total volume 50 mL,
volume dependency was observed anymore. 2.7 wt %) of the steric stabilizer PVP were prepared after it was found
Prehydrolyzed DMDES. In this method, DMDES monomer was that increased particle stability was necessary during shell growth.
first hydrolyzed prior to emulsion preparation by adding 2 mL of Hereafter, 20 mL of the original emulsion (still containing ammonia)
deionized water to 5 mL of DMDES monomer in a 20 mL glass vial was added, as well as 10 mL of the oil, which formed a layer on top of
and then stirring the solution in a vortex mixer (IKA minishaker MS2) the emulsion. The oils used were cyclohexane, pentane, iso-octane,
at 2500 rpm for 5 min and thereafter on a rollerbank (Stuart SRT9D, hexadecane, and octamethylcyclotetrasiloxane (OMCTS). Swelling
solution was tumbled at a speed of 35 rpm) to gently homogenize the was found to be most successful with cyclohexane and pentane,
mixture. This resulted in a clear solution. By prehydrolysis, the followed by iso-octane, whereas deswelling and unstable systems were
solubility of the monomer in the reaction medium is enhanced. Two observed for the oils hexadecane and OMCTS, a point we will come
batches of monodisperse emulsion (total volume 11 mL) of PDMS oil back to later. The droplets in sample C were exposed to cyclohexane
droplets were prepared from DMDES that was prehydrolyzed for half containing 0.28 mg/mL pyrromethene-567. For this, an aqueous
an hour and 24 h, respectively, by adding 1 mL of prehydrolyzed solution (total volume 25 mL) of 2.7 wt % PVP was prepared.
DMDES (final monomer concentration 6.5% v/v) to a 22.7% v/v Hereafter, 5 mL of the original emulsion was added, as well as 3.5 mL
aqueous ammonia solution in a 20 mL glass vial. At this point the of the fluorescently labeled oil phase. Droplets were allowed to grow,
solution containing half an hour prehydrolyzed DMDES appeared while keeping all samples slowly tumbling on a rollerbank for about 22
clear, whereas the solution containing 24 h prehydrolyzed DMDES h (samples A and B) or 6 h (sample C).
slightly turned turbid as a result of nucleation of oil droplets. All The droplets resulting from the seeded growth in samples A and C
samples were immediately mixed with a vortex mixer for 3−4 min to were encapsulated by a solid shell by making use of the coating
completely dissolve prehydrolyzed DMDES and placed on a procedure described above for the method based on efficient initial
rollerbank during droplet growth. After 24 h of droplet growth, the mechanical mixing with the Turrax homogenizer. However, compared
droplet sizes from all samples were measured using static light to this latter method, the current dispersion was already diluted with a
scattering (SLS). PVP solution (as described above) prior to the start of the
Seeded Growth - DMDES. A suspension of monodisperse PDMS encapsulation process and even before addition of the extra oil. The
seed droplets (total volume 31 mL) was prepared by shaking a mixture dilution step with the steric stabilizer at such an early stage ensured
of 2.8 mL of DMDES (final concentration 9.0% v/v) and 28.2 mL of that the continuous phase was presaturated with oil at all times and
an aqueous solution of ammonia (concentration of ammonia 22.7% v/ thereby facilitated a constant droplet size. Shell growth started by
v) in a 40 mL glass vial for 2 min by hand. The sample was placed on a keeping the sample static for 5 min such that the excess oil phase
rollerbank during droplet growth. The seeds were allowed to grow for creamed up. Subsequently, the turbid subnatant of the emulsion was
48 h after which the emulsion was transferred to a new vial (40 mL). transferred to a new vial such that the top oil phase would not be
We would like to note here that (as an exception) this sample was present during the coating reaction. Hereafter, the APS/RITC solution
prepared in a glass vial that contained a styrene−butadiene rubber was added (4 μL for sample A and 2 μL for sample C) while gently
capliner, manufactured by Wheaton. Synthesis of emulsions in all other stirring the dispersion. Then, equal amounts (0.34 mL for sample A
methods were performed in glass bottles/vials without a capliner. and 0.85 mL for sample C) of the monomers DMDES and TES were
Strong indications were found that chemicals, dissolved from these added using a syringe-pump (10 μL/min). The sample was transferred
capliners, favorably affected the particle size distribution, most to the rollerbank after all monomer had been added. Shells were
probably by acting as surface active agents, as briefly discussed in allowed to grow for a maximum of 24 h (sample A) or 3 h (sample C),
the Supporting Information. However, we do not think that the a time-scale set by the onset of buckling, which is a point we will come
presence of the capliner influenced the droplet growth and hence back to later. The coated emulsion droplets in sample A were
believe that the results would be the same otherwise. For the seeded transferred to ethanol within 24 h after the start of the shell growth.
growth step, 2.8 mL of fresh DMDES was added at once to this seed Shells remained stable and spherical when adding the sample drop-by-
suspension, and the sample was kept on the rollerbank (at a speed of drop to (anhydrous) ethanol in a ratio of 1:8 mL, under vigorous
35 rpm) for gentle mixing. The addition of DMDES was repeated stirring. Strangely enough, and unfortunately for unclear reasons as
again at 120 h, 160 h, and 216 h after the emulsion synthesis using the described above in the section for droplets prepared with the efficient
same amount of DMDES. Each time an excess layer of oil mechanical mixing procedure, particles were found to buckle when
(approximately 0.3 mL), which appeared in a matter of minutes using less ethanol.
after a sample was taken from the rollerbank, was discarded by Characterization. Static Light Scattering (SLS). The size and
transferring the emulsion to a new vial. polydispersity of the PDMS oil droplets were determined by Static
Seeded Growth - Hydrocarbons and Silicone Oil. For the seeded Light Scattering (SLS) experiments. The measurements were carried
growth of PDMS droplets with a second oil, several seed emulsions out in a home-built equipment using a HeNe laser as light source
were prepared, this time using the Turrax homogenizer, as described (632.8 nm, 10 mW). The measurements were performed on dilute
previously. Two samples were prepared according to the large scale suspensions in demineralized water, containing about 0.1% v/v of the
mechanical mixing procedure with the Turrax but with reduced original sample. The angular distribution of intensity of the scattered
DMDES concentrations. One of these samples (sample A) was light was measured at scattering angles in the range 14° to 135°
exposed to one type of oil, during which the particle size was relative to the transmitted beam, with a photomultiplier detector
monitored in time. The other sample (sample B) was used to mounted on a goniometer. The data were plotted against the
investigate the change in droplet size for various types of oil. Sample A scattering vector k = 4π nsin(θ/2)/λ, where n is the solvent refractive
(total volume 728.9 mL) was prepared in a 2 L beaker with an index and λ is the wavelength in vacuum. By comparing the scattering
ammonia concentration of 23.4% v/v and DMDES concentration of profiles with theoretical ones calculated with the full Mie solution for
4.1% v/v. Sample B (total volume 743.9 mL) contained 22.9% v/v the scattering factor,73 the radius and polydispersity were determined
ammonia and 6.0% v/v DMDES. The third monodisperse emulsion by fitting (by hand). The positions of the interference minima in the
(sample C) was prepared according to the small scale mechanical scattering curves give the size of the droplets, whereas the
mixing method based on the Turrax as described above (for an polydispersity is given by the depth of the minima: the lower the
overview of all final sizes, see Table S1 in the Supporting Information). polydispersity, the deeper the minima. During the fitting, the last
This sample was used during swelling studies with fluorescently minima in the scattering profile were used to determine the
labeled cyclohexane. In all preparations the droplets were allowed to polydispersity. By doing so, the effect of multiple scattering at low
grow, while keeping the sample without stirring or homogenization, angles (the first minima) could be avoided for these micron-sized
for at least 2 days before starting the seeded growth procedure. particles. The value of the refractive index used to fit the experimental

1712 DOI: 10.1021/cm504504t

Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

Figure 2. (A) SLS experimental data (scatter) fitted by theoretical calculations with full Mie solution of the form factor (lines) for different series of
PDMS droplets prepared using the Turrax homogenizer. The values of droplets size and polydispersity are obtained from the fit (described in the
Characterization section). (B) Confocal micrograph of the 5.36 μm microcapsules in an index matching solvent (mixture of ethanol and 52.6% v/v
DMSO), the scale bar represents 5 μm. (C) AFM image of a dried capsule, after removing the PDMS core by washing with ethanol. (D) The height
profile taken along the line drawn through the collapsed capsules from which we obtained a shell thickness of d = 56 nm.

data for the oil droplets was n25 D = 1.394, corresponding to a low of ref 48. The droplets were then encapsulated by a thin
molecular weight silicone oil,74 and for water was n20
D = 1.333.
75
tetraethoxysilane-cross-linked-polydimethoxysilane (TC-
Confocal and Optical Microscopy. The fluorescently labeled shells PDMS) shell. A schematic illustration of the newly developed
were imaged using an inverted Leica, TCS-SP2 Confocal Scanning
Laser Microscope (63 × NA 1.4 oil immersion confocal objective). A
synthesis procedures which are presented in this paper is shown
543 nm green HeNe laser was used for the excitation of rhodamine in Figure 1.
(RITC) labeled shells, and a 488 nm blue argon laser was used for the Efficient Initial Mechanical Mixing and No Agitation
excitation of pyrromethene-567 dye and of CdSe/CdS/CdZnS/ZnS during Growth. We modified the synthesis procedure, by
QD’s. Samples were put in a capillary (Vitrocom) either 0.1 × 1 mm introducing a Turrax homogenizer for mixing and by using
or 0.1 × 2 mm for imaging. Unlabeled oil droplets were observed using higher ammonia and DMDES monomer concentrations than
a Leica optical microscope (63× NA 0.7 air objective). what was reported before by Zoldesi et al.50,51 The Turrax
Atomic Force Microscopy (AFM). The thickness of the micro-
homogenizer helped to speed up the hydrolysis of the added
capsules was measured using Atomic Force Microscopy (AFM, Digital
Instrument, Nanoscope) in tapping mode. Samples for AFM were monomer by increasing the surface area. As a result, the
prepared by applying a drop of the hollow shells in ethanol onto a nucleation time for PDMS droplets was reduced, so that the
glass cover slide. The collapse of the shells, due to drying, leads to droplets became more monodisperse. Samples prepared in this
plateaus in the height profile that correspond to twice the thickness of way became turbid about 8−14 min after start of the mixing,
the shell. indicative for the formation of PDMS droplets. This happened
Refractive Index of the Shells. The refractive index of the solid faster when compared to the 1 h time-scale described in refs 48
shells (after removing the PDMS oil template) was obtained by adding
and 51. The position of the homogenizer-probe in the flask was
dimethyl sulfoxide (DMSO, n20 D = 1.47) drop by drop to a dispersion
of shells in ethanol (n20 found to strongly affect the monodispersity of the droplets,
D = 1.36) up to a point where the solution
became almost transparent. Initially, shells in pure ethanol scattered because mixing variations affect the homogenization and hence
light as the solution was slightly turbid, and the particles were also the nucleation and growth (by early aggregation of siloxane
visible with an optical microscope. The concentration of DMSO in the substructures and growth with monomer) details. A probe
final volume was approximately 52.6% v/v, and the refractive index of position close to the interface of the two immiscible liquids was
the solvent mixture (without shells, shells were removed by found to be most successful. Despite the efficient mixing with
centrifugation) was n20 D = 1.42, measured using a refractometer the Turrax homogenizer, all samples contained an excess layer
(Abbe Refractometer Atago NAR-3 T).

■
of oil directly after stopping the homogenization. It was also
found out that the samples had to be kept static after mixing.
DISCUSSION Samples that were kept on the rollerbank became polydisperse.
A number of approaches was used to prepare monodisperse, This problem can be attributed to the unreacted DMDES
large polydimethylsiloxane (PDMS) oil droplets by hydrolysis monomer which would be mixed in during the tumbling
and condensation of dimethyldiethoxysilane (DMDES) in an motion on the rollerbank and could, upon hydrolysis, increase
aqueous ammonia solution, based on variations of the method the ionic strength and hence affect the charge stabilization of
1713 DOI: 10.1021/cm504504t
Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

the as formed emulsion. It is known,76 that these PDMS difference. As mentioned already, the consequence of using a
droplets are typically stabilized both by surface ionized groups pure water phase which does not dissolve the unhydrolyzed
(resulting in a zeta potential of −69 mV) and by the linear monomer makes that the latter changes in the effectiveness of
chains in the synthesized oil that act as surfactants. In addition, the mixing will alter the surface area of the (unhydrolyzed)
sensitivity to the way the system was agitated could originate monomer and hence also alter the size of the final droplets. A
from the increased surface area of the dispersed DMDES strong influence of the emulsion amount and the mixing
droplets by using the Turrax homogenizer, which leads to an procedure on the particle size was also reported in ref 50.
increase of the hydrolysis rate. For both reasons, the higher Tetraethoxysilane-cross-linked-polydimethoxysilane shells
concentration of silanol groups by hydrolysis would result in a were grown around the monodisperse, template droplets
higher ionic strength which reduces the charge stabilization of prepared with the large scale Turrax procedure. The cross-
droplets present and makes them less stable against shear linked network of silica and siloxane units formed upon
induced aggregation. Apparently, the tumbling motion caused addition of both the monomers DMDES and TES, in
by the rollerbank already produced enough shear to induce accordance with previous work done by O’Sullivan et al.52
aggregation compared to unagitated growth. A further However, shells were found to form aggregates consisting of a
indication that the system is indeed really sensitive to changes small number of up to 10 particles during the coating step,
in the amount of hydrolyzed DMDES is that both methods likely caused by van der Waals attractions between these
(agitated/unagitated) resulted in polydisperse systems when micron sized particles. In order to prevent this, we modified the
the overall monomer concentration was increased to higher procedure slightly by diluting the emulsion with a solution of
values than what has been reported in the Experimental the steric stabilizer PVP prior to addition of the monomers as
Section. When taking this restriction into account, the method described in the Experimental Section. In addition, PVP is
resulted in a high yield of highly monodisperse and micron known to facilitate silica growth on surfaces.20 A concentration
sized oil droplets. Four batches of droplets were prepared at an of PVP of 2.3 wt % was found to reduce aggregation
ammonia concentration of 23% v/v and DMDES concentration considerably.
of 6.8% v/v, and one more batch was prepared at a reduced Shell growth typically continued for about 2 days, after which
DMDES concentration of 2.9% v/v. The many well-defined the liquid templating drop could be removed by transferring the
minima in the static-light-scattering (SLS) curves in Figure 2A coated droplets to ethanol. In pure ethanol, the TC-PDMS
are indicative of the high monodispersity in these large scale microcapsules were found to have a zeta potential of −40.5 mV
(750 mL) suspensions. The size and polydispersity of the (Zetasizer Nano ZS, Malvern), indicating that these dispersions
droplets could be obtained by comparing the scattering profiles are stabilized by surface charges. A confocal micrograph of
with theoretical ones calculated using the full Mie solutions for shells formed around 5.36 μm templating droplets, dispersed in
the scattering form factors.73 The fits showed that the droplet an index matched solvent mixture of ethanol and DMSO (n20 D =
diameters and polydispersities varied slightly between the four 1.42), is given in Figure 2B. The shell thickness of the capsules
different batches (with a DMDES concentration of 6.8% v/v) in this sample was measured by taking cross-sectional profiles
after 3 days of droplet growth, in the range 4−5.4 μm and 2− through Atomic Force Microscopy (AFM) height images, as
2.5%, respectively. See Table S1 in the Supporting Information shown in Figure 2C and D. Using this procedure, the shells
for an overview of all droplet diameters and polydispersities for were found to be about 56 nm thick. We can therefore
this and all other synthesis methods. On a side note, we would conclude that we can grow stable TC-PDMS capsules around
like to mention that these droplets typically became 3 μm PDMS template droplets with diameters exceeding the limit of
already within the first 24 h. Hence, the synthesis procedure about 3 μm reported before.50,52
generates droplets that are similar in size to the upper limit that Prehydrolysis. Another method that we used to synthesize
was achieved so far48(∼5 μm). However, the deep minima in monodisperse PDMS oil-in-water emulsions is using prehy-
the SLS profiles prove that the polydispersity remains drolyzed DMDES. A similar procedure has been reported in
significantly lower in the current system. When reducing the refs 77 and 78 for the preparation of lock and key colloids,
DMDES monomer concentration from 6.8 to 2.9% v/v, the where the trialkoxysilane 3-methacryloxypropyltrimethoxysilane
droplet diameter decreased to 2.4 μm, Figure 2A and (TPM), instead of DMDES, was solubilized in water and used
Supporting Information Table S1. Earlier studies have also as a precursor of monodisperse silicone oil droplets.
shown a similar relationship between droplet size and Prehydrolysis of DMDES makes the monomer readily soluble
monomer concentration.48,50,53 This decrease is probably in the reaction medium, so that the nucleation time may be
affected not only by the total amount being less but also by reduced, leading to monodisperse droplets. Prehydrolysis, prior
the reduced ionic strength during nucleation and early to emulsion preparation, was done by mixing 28.5% v/v of
aggregation of the siloxane subunits, as was observed before deionized water and 71.5% v/v DMDES. Before mixing, the
for the Stöber growth.49 solution formed two phases, as DMDES is less soluble in water;
All samples prepared with the smaller scale method also but within half an hour it became a clear single phase, indicating
resulted in monodisperse emulsions. Using SLS, the droplet the hydrolysis of the monomer. Note that the formation of a
sizes were determined for two batches after 2 days of droplet clear single phase did not necessarily mean that the monomer is
growth. The droplet diameter and polydispersities were found completely hydrolyzed in water. Previously, Zoldesi et al.50
to be 2.8 and 1.7 μm and 1.9% and 1.8% (see Supporting already confirmed from NMR studies on solutions of PDMS oil
Information Table S1), respectively. One can conclude that in water, after 10 h of emulsion preparation, that a significant
these sizes were reduced when compared to the large scale amount of DMDES was only hydrolyzed at one end. Besides,
procedure that uses the same concentration of ammonia and we also checked the conductivity (Radiometer Analytical,
DMDES. The absence of an excess DMDES phase after mixing CDM230) of the prehydrolyzed solution after half an hour and
as well as a more effective homogenization in general for 24 h of sample preparation. We measured an increase in the
smaller amounts of emulsions are most likely the cause of this conductivity value from 1.0 ± 0.1 μS/cm to 2.2 ± 0.1 μS/cm in
1714 DOI: 10.1021/cm504504t
Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

Figure 3. Increase of the droplet diameter with time using a seeded growth procedure with (A) DMDES (data points marked in red denote the time
at which more of the monomer was added), and (B) cyclohexane or pentane. Please note that the error bars in these figures are not the errors in the
measurement of the particle size but indicate the polydispersity of the size distributions. Sizes and polydispersities were determined with SLS (A) and
from optical micrographs (B), making the systematic deviation in B larger due to the poor resolution. Inset pictures are the optical micrographs of
droplets before and after addition of DMDES and hydrocarbons. Scale bars represent 5 μm.

24 h of prehydrolysis, caused by an increase in the during the preparation of the seed/original suspension. Seeded
concentration of acidic silanol groups in the solution as a growth was performed for 9 days during which the monomer
result of hydrolysis of DMDES. was added four times with time intervals of 48 h, except for the
We prepared four batches of emulsions using this hydrolyzed second addition which was done after 72 h. The sample was
DMDES solution, two batches each, after half an hour and 24 h placed on a rollerbank (at a speed of 35 rpm), after each
of prehydrolysis. Emulsions were prepared by adding 9% v/v of addition, during droplet growth. In previous systems with an
prehydrolyzed DMDES (final concentration of DMDES was excess DMDES phase in the reaction medium even the slight
6.5% v/v) to an aqueous solution of 22.7% v/v of ammonia, agitation caused by a rollerbank often resulted in a polydisperse
followed by shaking in a vortex mixer for 3−4 min to dissolve emulsion. This was probably due to an increase in ionic
prehydrolyzed DMDES in the solution. Thereafter, the samples strength (by the hydrolysis of excess monomer) and hence a
were placed on a rollerbank to gently tumble the mixture decrease in stability of droplets against shear induced
during droplet growth. After 24 h of synthesis, two batches aggregation. However, in the present system PDMS droplets
prepared from half an hour prehydrolyzed DMDES contained remained stable and monodisperse, even in the presence of
monodisperse PDMS oil droplets of diameter 5.6 and 5.0 μm, excess unreacted DMDES. Apparently, the stability of the
respectively, with polydispersities 2% and 1.9% (see Supporting droplets here can be attributed to factors like the increase in
Information Table S1). These sizes were comparable to those surface area and hence surface charge,48 as the seed droplets are
obtained using the Turrax homogenizer. However, two batches a lot larger than the PDMS nuclei formed during emulsion
of emulsions, each prepared using a 24 h old prehydrolyzed synthesis, and the surfactant-like behavior of linear PDMS
DMDES, contained droplets of only 1.7 and 2.0 μm diameter chains.76 The growth profile of the DMDES seeded growth
with polydispersity 2.9% (see Supporting Information Table study is shown in Figure 3A, where the points marked in red
S1). Although still monodisperse, the average droplet size was represent the time of addition of monomer. The data point at 0
reduced approximately by 60% compared to the droplets h represents the time of preparation of the seed emulsion. A
prepared from a half hour old prehydrolyzed DMDES. This major increase in size of the seed droplets, approximately 1 μm
could be due to the fact that the aging of prehydrolyzed in diameter occurred during the first 24 h of growth; in the next
monomer results in a gradual increase in the concentration of 24 h the droplets grew approximately another 0.5 μm. The
hydrolyzed monomer, confirmed from the conductivity droplets were found to remain stable and monodisperse up to
measurements. As a result the nucleation rate is increased in 8.00 μm (the SLS curve is shown in the graphical abstract).
the sample, leading to smaller droplets. However, after the last addition of monomer the emulsion
On the contrary, we found that emulsions made without became unstable and underwent coalescence. Thus, although
prehydrolysis of monomer under otherwise identical conditions the size of the droplets can be increased by feeding with
often became polydisperse. This result is therefore also in DMDES monomer, the size increase occurred relatively slowly
agreement with the adverse effect of the excess unreacted due to the slow condensation reaction of hydrolyzed
DMDES phase that was described above for samples prepared monomer.50 So it would be interesting to search for oils
with the Turrax homogenizer, which were gently tumbled where a similar, or even larger, size increase can be achieved in
during droplet growth instead of being kept static. Apparently, a shorter time period.
prehydrolysis reduces the amount of unreacted DMDES in the Seeded Growth - Hydrocarbons and Silicone Oil. It
sample that would otherwise adversely affect the size was found that PDMS emulsions could be grown larger, at a
distribution of droplets. faster rate and without the need of successive addition steps, by
Seeded Growth - DMDES. Multiple additions of DMDES using a seeded growth procedure with hydrocarbon oils instead
were used as an alternative means to increase the size of the of DMDES. Preformed PDMS emulsions were used that were
PDMS droplets, in a way similar to the seeded growth of silica at least 2 days old as the droplets are known to increase slowly
colloids.79 The total amount of DMDES that was added during in time within the first 48 h.50 An elapse of at least 2 days
the seeded growth step was the same as what had been used therefore allowed us to attribute any further changes in droplet
1715 DOI: 10.1021/cm504504t
Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

size to the exposure to the oil. The original PDMS droplets smaller increase, resulting in diameters of about 4.9 μm, was
(sample A), 3 days old, were found to be 2.96 μm in diameter observed for the sample with iso-octane. Droplets in both
with a polydispersity of 2% as determined with SLS. One batch samples remained stable for a couple of days. With OMCTS
of these droplets was exposed to cyclohexane, and another (result not shown) and hexadecane, the droplet size was found
batch was exposed to pentane. After 22 h, the droplet diameter to decrease and the number of oil droplets to reduce. Besides,
in both samples had increased to 5.4 μm as shown in Figure 3B, these systems became polydisperse. From these observations,
where sizes were measured from optical micrographs instead of one can conclude that the various oils affect both the size and
via SLS as a dilution step was not possible since the reaction the stability of the original PDMS seed droplets.
medium had to be saturated with oil at all times, as discussed The droplet size is most likely related to the solubility of the
later. As a result, the resolution was poor, resulting in a larger added oil in water. The threshold for successful swelling of the
standard deviation than what was typically seen with SLS but preformed emulsion will be set by the solubility of the
which remained almost constant for increasing sizes. The synthesized PDMS oil itself. Oils that are more soluble in water
similar increase of the particle diameter is caused by the fact than PDMS resulted in a droplet size increase. Also note that
that the solubility of pentane in water is similar to that of the largest, and similar, size increase was seen for cyclohexane
cyclohexane,80 a point we will come back to. Compared to the and pentane (Figure 3B) which are equally well soluble in
seeded growth with DMDES (Figure 3A), the droplet size water, whereas iso-octane, for which the solubility is
increase after 24 h was more rapid for cyclohexane and pentane. significantly reduced, but still larger than PDMS, resulted in a
In order to investigate the seeded growth procedure for a smaller size increase. The reverse, a decrease in droplet size,
wider variety of oils, preformed PDMS emulsions were not only occurred for oils poorly soluble in water (like hexadecane81).
exposed to cyclohexane but also to iso-octane, hexadecane, and The stability of the swollen droplets is most likely brought
the silicone oil OMCTS for 22 h. For this experiment, 2 days about by the linear chains in the synthesized PDMS oil which
old PDMS droplets were used with a diameter of 4.2 μm and can act as surfactants.76 The presence of these linear chains
polydispersity of 2.1% (sample B). The optical micrographs of could explain the stability of droplets swollen with cyclohexane,
the original droplets and the droplets exposed to various oils pentane, and iso-octane in the present study even though the
volume fraction of PDMS within the swollen droplets (down to
are shown in Figure 4. A significant size increase was observed
0.12 in the present system) is lower than what has been
in the presence of cyclohexane, resulting in monodisperse
described in ref 76. In this context, it is surprising that droplets
droplets of about 8.2 μm in diameter. This indicates that
exposed to OMCTS were found to deswell and to become
swelling with cyclohexane is a highly efficient process as seeded
unstable as indicated by the polydispersity and the reduced
growth with 100% of the added oil would have resulted in
number of droplets. In principle, stable droplets were expected
droplets of about 8.9 μm in diameter, (roughly) estimated from as the interfacial tension of OMCTS is lower than that of n-
the volume fraction and size of the initial PDMS droplets. A heptane, which is known to result in stable droplets when
mixed with synthesized PDMS.76 About 80% of the total
material in the synthesized PDMS oil is comprised of the cyclic
tetramer OMCTS.48 Deswelling with OMCTS therefore
suggested that the linear oligomers, present in the synthesized
PDMS oil, raise the water solubility of the preformed emulsion
to values above that of OMCTS. Along the same line of
reasoning, it is possible that these linear chains were withdrawn
first from the as-prepared emulsion during the deswelling
process. A reduction in the number of these surface active
molecules could also clarify why droplets did not only become
smaller but also became unstable in the presence of poorly
soluble oils, as indicated by the polydispersity and by the
reduction in the number of droplets. Hence, the interplay
between the solubility of the oils in water and the presence of
linear PDMS oligomers sets the success rate for swelling.
On a final note, we would like to mention that, apart from
swelling, also loading with a fluorescent dye was possible when
using the seeded growth procedure. Such loading of chemicals
is clearly of importance in many applications. Preformed PDMS
droplets (sample C) (3.2 μm in diameter, polydispersity 2%, 5
days old) were swollen to 5.5 μm within 6 h by exposure to
cyclohexane stained with pyrromethene-567. A confocal
micrograph of this dyed emulsion is shown in the inset of
Figure 4B. Unfortunately, emulsion preloading in the present
Figure 4. Optical micrographs of the (A) original PDMS seed droplets system was not found to be successful when using cyclohexane
and of these droplets after 22 h of exposure to (B) cyclohexane, (C)
in which 6.9 nm quantum dots were dispersed (0.5 * 10−7 mol/
iso-octane, and (D) hexadecane. The droplet size was found to vary,
depending on the solubility of the hydrocarbon oil in water. Scale bars mL). Even though the droplets did swell, the quantum dots
denote 10 μm. The appearance of a core−shell structure is caused by were apparently not water-soluble enough and remained
the refractive index mismatch. Inset figure in (B) represents a dispersed in the phase separated (excess) cyclohexane phase.
fluorescence micrograph of pyrromethene-567 loaded droplets, Future work has to show if a slightly more polar capping
swollen with cyclohexane. molecule, making the solubility of the quantum dots in water
1716 DOI: 10.1021/cm504504t
Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

Figure 5. (A) Confocal micrograph of shells formed around cyclohexane swollen PDMS droplets, 24 h after the start of shell growth. (B) The
preformed PDMS emulsion could be swollen with fluorescently labeled cyclohexane. After encapsulation, shell buckling indicated that the volume of
the templating droplets is not conserved anymore, even though the water was initially presaturated with this oil. This suggests that loss of oil possibly
occurred via evaporation. (C) Shells unbuckled, and then budded secondary droplets, in the presence of excess cyclohexane due to overloading,
similarly as reported in ref 64. Scale bars denote 5 μm.

higher, could overcome this negative result. As a result, loading further studies by exposing the aqueous dispersion to oil
with quantum dots such as described by Cui et al.53 could not vapors. Alternatively, the formation of other solutes during the
be achieved in the present study. coating step could affect the solubility of the added oil in the
The hydrocarbon swollen droplets could also be encapsu- continuous phase. Finally, also a reduction in the ionic strength
lated by a solid shell. In principle, we used the same procedure might be the cause,72 due to slow disappearance of silanol
as described for the method based on efficient initial mixing groups in the aqueous phase.
with the Turrax homogenizer, including the addition of PVP. When transferring buckled shells from the reaction medium
Addition of PVP was required to ensure stability of the swollen to ethanol, a deformed capsule was obtained. These capsules
droplets under the conditions of the shell growth step. were slightly dimpled, and their shape did not change 24 h after
However, addition of a PVP solution would undo part of the transfer to ethanol. This therefore suggests that the particles
swelling, as some cyclohexane or pentane would dissolve back had become fixated in this shape, possibly because shell growth
in the added quantity of water. To prevent this, a PVP stock proceeded while the volume of the template had already
solution was already added before swelling with cyclohexane or decreased. To prevent this, shells could be transferred to
pentane. The same effect also made it necessary to determine ethanol before the onset of buckling. Unfortunately, this also
the drop size in the reaction medium using an optical limits the time for shell growth and hence the achievable shell
microscope, rather than using the SLS for which dilution of thickness.
the suspension is needed. With this presaturation step, shells Interestingly, unbuckling of the buckled capsules was
could for example be successfully grown around the 5.4 μm possible by exposing the sample, in water, again to an excess
cyclohexane swollen droplets (sample A, Figure 3B) as can be phase of (dyed) cyclohexane. We visualized this by filling half a
seen in Figure 5A. capillary with the buckled shells from the reaction medium
Spherical, ethanol-filled capsules could be obtained with this (Figure 5B) and the rest of the capillary with the (fluorescent)
seeded growth procedure when shells were transferred to cyclohexane phase. Confocal micrographs of this system were
ethanol according to the procedure described in the taken near the oil−water interface soon after preparation of this
Experimental Section. The shell thickness was found to be sample. As can be seen in Figure 5C, shells fully relaxed back to
only 20 nm with AFM. We expect that by increasing the spheres and even became overloaded with the oil phase64
amount of DMDES and TES during shell growth, thicker shells resulting in the formation of snowman-like particles. The
could be formed.52 However, the present study focused on the relaxation of the buckled capsules indicated that the shells are
increase in the capsule diameter rather than control over the still highly permeable to both oil and dye molecules, even
shell thickness as the latter has already been achieved.50−52 though the particle synthesis is slightly altered when compared
Besides, capsules with high diameter-to-wall thickness ratios can to procedures used in refs 50, 51, 82, and 52 by introducing the
be interesting for potential applications.7 polymer PVP and using the seeded growth procedure.
Unfortunately, shells that were formed around the seeded
grown droplets were found to buckle in time in the reaction
medium. Often this was observed within about 24 h of shell
■ CONCLUSIONS
In the present study, we have shown that we can synthesize
growth. An example of the buckling behavior is shown in Figure monodisperse PDMS droplets with sizes significantly above 3
5B, for shells formed around the droplets swollen with stained μm. In addition, we showed that it is possible to grow an elastic
cyclohexane (sample C, inset Figure 4B). Buckling indicated shell of TES-cross-linked-PDMS around these droplets. The
that the volume of the templating droplet was not conserved templating droplets were prepared via various methods, for
anymore. In principle, such a decrease in volume should be which an overview of the achieved final particle sizes and
prevented if the continuous aqueous phase is kept saturated polydispersities is given in Table S1 in the Supporting
with oil during the whole process of shell growth. These results Information. In one method, a Turrax homogenizer was used
therefore suggest that loss of oil possibly occurred via to enhance mixing of the immiscible monomer DMDES and to
evaporation, even though the samples were sealed with a cap. increase the rate of hydrolysis and solubilization of the as
An indication for this hypothesis is that samples swollen with formed hydrolyzed species. This procedure resulted in droplets
pentane, which has an even higher vapor pressure than (and hence capsules) with diameters above 5 μm on a bulk
cyclohexane, buckled in a more pronounced way and already (750 mL) scale. In the second technique, the monomer
within the first 24 h. Leakage of oil might be prevented in DMDES was prehydrolyzed, creating readily hydrolyzed and
1717 DOI: 10.1021/cm504504t
Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

water-soluble species. By doing so, the nucleation time was Notes

reduced when compared to emulsions prepared under The authors declare no competing financial interest.

■
otherwise identical conditions, leading to more monodisperse
droplets. In the third method, preformed PDMS templates ACKNOWLEDGMENTS
were swollen with DMDES monomer or with low-molecular
weight hydrocarbons during a seeded growth procedure, The authors thank Mark Boneschanscher and Jaco Geuchies for
resulting in the largest monodisperse droplets formed in this their help with the AFM measurements on the capsules and
study of up to 8.2 μm. Besides, we have shown that along with Relinde Moes for synthesizing and providing the QD’s. We also
swelling, oil soluble chemicals (like a pyrromethene dye) can be acknowledge Esther van Duijn, Mirjam Damen, Reinout
loaded into the droplets. For swelling with a second oil, the Raijmakers, and Robin Jastrzebski for the mass-spectroscopy
solubility of the added oil in water has to be higher than that of and NMR measurements and for help during spectrum
the synthesized PDMS, which for the oils we tried was valid interpretation. N.A.E. is supported by the Industrial Partnership
when using cyclohexane, pentane, and iso-octane. A drawback Programme (IPP) Innovatie Physics for Oil and Gas (iPOG) of
when using swelling with second oils is that encapsulation of the Stichting voor Fundamenteel Onderzoek der Materie
the mixed oil droplets resulted in buckled capsules even after (FOM), which is supported financially by Nederlandse
presaturating the aqueous phase, possibly because the solubility Organisatie voor Wetenschappelijk Onderzoek (NWO). The
changes over time due to formation of other solutes or due to IPP iPOG is cofinanced by Stichting Shell Research. Within this
evaporation of oil affecting the equilibrium concentration. To program, Esther Vermolen is thanked for useful discussions. J.J.
prevent this, the spherical capsules had to be transferred to was supported by Stichting voor Fundamenteel Onderzoek der
ethanol before buckling started in the reaction medium, limiting Materie (FOM), which is financially supported by the
the duration of shell growth. Nederlandse organisatie voor Wetenschappelijke Onderzoek
In short, this paper presents several robust and surfactant free (NWO).
synthesis procedures that yield highly monodisperse PDMS
droplets above 3 μm. In addition, we have shown that these
large particles can also be used as sacrificial templates and allow
■ REFERENCES
(1) Langer, R. Nature 1998, 392, 5−10.
facile drop removal. When comparing all procedures, it can be (2) Peyratout, C. S.; Dahne, L. Angew. Chem., Int. Ed. 2004, 43,
stated that the efficient initial mechanical mixing with the 3762−3783.
Turrax homogenizer currently allows for the largest scale (3) Johnston, A. P. R.; Cortez, C.; Angelatos, A. S.; Caruso, F. Curr.
synthesis of droplets up to about 5.5 μm in diameter. When Opin. Colloid Interface Sci. 2006, 11, 203−209.
aiming for even larger droplets, the seeded growth procedures (4) Ariga, K.; Hill, J. P.; Lee, M. V.; Vinu, A.; Charvet, R.; Acharya, S.
Sci. Technol. Adv. Mater. 2008, 9, 014109.
with DMDES or hydrocarbons are preferred. The main (5) Vriezema, D. M.; Aragonés, M. C.; Elemans, J.; Cornelissen, J.;
advantages of the hydrocarbon method are that the droplet Rowan, A. E.; Nolte, R. J. M. Chem. Rev. 2005, 105, 1445−1489.
size increases faster, and no successive addition steps are (6) Yow, H. N.; Routh, A. F. Soft Matter 2006, 2, 940−949.
required. Finally, also the droplet size up to 5.6 μm of the (7) Caruso, F. Chem.Eur. J. 2000, 6, 413−419.
prehydrolysis method can compete with the efficient initial (8) Donath, E.; Sukhorukov, G. B.; Caruso, F.; Davis, S. A.;
mechanical mixing of the Turrax homogenizer even though the Möhwald, H. Angew. Chem., Int. Ed. 1998, 37, 2202−2205.
synthesis scale is smaller. Overall, we think that the increased (9) Liu, Y. D.; Goebl, J.; Yin, Y. D. Chem. Soc. Rev. 2013, 42, 2610−
PDMS drop diameters broaden the horizon for potential 2653.
applications and also allow the use of elastic shells for granular (10) Zha, L. S.; Zhang, Y.; Yang, W. L.; Fu, S. K. Adv. Mater. 2002,
jamming/contact force studies not only when the sub-10 μm 14, 1090−1092.
(11) Berg, J.; Sundberg, D.; Kronberg, B. J. Microencapsulation 1989,
PDMS droplets are used as templates for shell synthesis, a topic 6, 327−337.
we focused on, but also in general for synthesis of (12) Tiarks, F.; Landfester, K.; Antonietti, M. Langmuir 2001, 17,
monodisperse emulsions that find applications in food science, 908−918.
cosmetics, and drug delivery. (13) Caruso, F.; Caruso, R. A.; Möhwald, H. Science 1998, 282,

■
1111−1114.
ASSOCIATED CONTENT (14) Rossier-Miranda, F. J.; Schroën, C.; Boom, R. M. Colloids Surf.,
A 2009, 343, 43−49.
*
S Supporting Information (15) Zhang, L.; D’Acunzi, M.; Kappl, M.; Auernhammer, G. K.;
Information and characterization data on the contaminants Vollmer, D.; van Kats, C. M.; van Blaaderen, A. Langmuir 2009, 25,
released by the styrene−butadiene capliner that were found to 2711−2717.
favorably affect the particle distribution in the samples prepared (16) Nagao, D.; van Kats, C. M.; Hayasaka, K.; Sugimoto, M.;
Konno, M.; Imhof, A.; van Blaaderen, A. Langmuir 2010, 26, 5208−
for the seeded growth study with DMDES. An overview of 5212.
PDMS droplets size and polydispersity obtained from all (17) Xu, X. L.; Asher, S. A. J. Am. Chem. Soc. 2004, 126, 7940−7945.
different synthesis methods. This material is available free of (18) Kamata, K.; Lu, Y.; Xia, Y. N. J. Am. Chem. Soc. 2003, 125,
charge via the Internet at http://pubs.acs.org. 2384−2385.

■
(19) Imhof, A. Langmuir 2001, 17, 3579−3585.
AUTHOR INFORMATION (20) Graf, C.; Vossen, D. L. J.; Imhof, A.; van Blaaderen, A. Langmuir
2003, 19, 6693−6700.
Corresponding Authors (21) Lou, X. W.; Archer, L. A.; Yang, Z. C. Adv. Mater. 2008, 20,
*E-mail: N.A.Elbers@uu.nl. 3987−4019.
*E-mail: A.vanBlaaderen@uu.nl. (22) Horecha, M.; Senkovskyy, V.; Stamm, M.; Kiriy, A. Macro-
molecules 2009, 42, 5811−5817.
Author Contributions (23) Tjipto, E.; Cadwell, K. D.; Quinn, J. F.; Johnston, A. P. R.;
†
N.A.E. and J.J. contributed equally to this work. Abbott, N. L.; Caruso, F. Nano Lett. 2006, 6, 2243−2248.

1718 DOI: 10.1021/cm504504t

Chem. Mater. 2015, 27, 1709−1719
Chemistry of Materials Article

(24) Shchukin, D. G.; Köhler, K.; Möhwald, H.; Sukhorukov, G. B. (61) Gao, C. Y.; Donath, E.; Möhwald, H.; Shen, J. C. Angew. Chem.,
Angew. Chem., Int. Ed. 2005, 44, 3310−3314. Int. Ed. 2002, 41, 3789−3793.
(25) Esser-Kahn, A. P.; Odom, S. A.; Sottos, N. R.; White, S. R.; (62) Zhang, J.; Coulston, R. J.; Jones, S. T.; Geng, J.; Scherman, O.
Moore, J. S. Macromolecules 2011, 44, 5539−5553. A.; Abell, C. Science 2012, 335, 690−694.
(26) Abismaïl, B.; Canselier, J. P.; Wilhelm, A. M.; Delmas, H.; (63) Li, C.; Li, Z. Y.; Zhang, J.; Wang, K.; Gong, Y. H.; Luo, G. F.;
Gourdon, C. Ultrason. Sonochem. 1999, 6, 75−83. Zhuo, R. X.; Zhang, X. Z. J. Mater. Chem. 2012, 22, 4623−4626.
(27) Richards, W. T. J. Am. Chem. Soc. 1929, 51, 17241729. (64) Jose, J.; Kamp, M.; van Blaaderen, A.; Imhof, A. Langmuir 2014,
(28) Hsu, P.; Poulin, P.; Weitz, D. A. J. Colloid Interface Sci. 1998, 30, 2385−2393.
200, 182−184. (65) Elbers, N. A.; Jose, J.; van Blaaderen, A. Manuscript in
(29) Bibette, J. J. Colloid Interface Sci. 1991, 147, 474−478. preparation.
(30) Mason, T. G.; Bibette, J. Phys. Rev. Lett. 1996, 77, 3481−3484. (66) Liu, A. J.; Nagel, S. R. Soft Matter 2010, 6, 2869−2870.
(31) Vilanova, N.; Rodriguez-Abreu, C.; Fernandez-Nieves, A.; (67) Jose, J.; van Blaaderen, A.; Imhof, A. Manuscript in preparation.
Solans, C. ACS Appl. Mater. Interfaces 2013, 5, 5247−5252. (68) Jose, J.; Blab, G. A.; van Blaaderen, A.; Imhof, A. Soft Matter
(32) Priest, C.; Quinn, A.; Postma, A.; Zelikin, A. N.; Ralston, J.; 2015, 11, 1800−1813.
Caruso, F. Lab Chip 2008, 8, 2182−2187. (69) Kwon, O. S.; Jang, J.; Bae, J. Curr. Org. Chem. 2013, 17, 3−13.
(33) Engl, W.; Backov, R.; Panizza, P. Curr. Opin. Colloid Interface Sci. (70) Xie, R. G.; Kolb, U.; Li, J. X.; Basché, T.; Mews, A. J. Am. Chem.
2008, 13, 206−216. Soc. 2005, 127, 7480−7488.
(34) Utada, A. S.; Lorenceau, E.; Link, D. R.; Kaplan, P. D.; Stone, H. (71) Zoldesi, C. I.; Ivanovska, I. L.; Quilliet, C.; Wuite, G. J. L.;
A.; Weitz, D. A. Science 2005, 308, 537−541. Imhof, A. Phys. Rev. E 2008, 78, 051401.
(35) Duncanson, W. J.; Lin, T.; Abate, A. R.; Seiffert, S.; Shah, R. K.; (72) Jadhao, V.; Thomas, C. K.; de la Cruz, M. O. Proc. Natl. Acad.
Sci. U. S. A. 2014, 111, 12673−12678.
Weitz, D. A. Lab Chip 2012, 12, 2135−2145.
(73) Bohren, C. F.; Huffman, D. R. Absorption and scattering of light
(36) Berkland, C.; Kim, K. K.; Pack, D. W. J. Controlled Release 2001,
by small particles; John Wiley and Sons: New York, 1983.
73, 59−74.
(74) Smith, A. L The Analytical Chemistry of Silicones; Wiley-
(37) Berkland, C.; Pollauf, E.; Varde, N.; Pack, D. W.; Kim, K. Pharm.
Interscience: New York, 1991.
Res. 2007, 24, 1007−1013. (75) Lide, D. R. CRC Handbook of Chemistry and Physics; CRC Press:
(38) Kohane, D. S. Biotechnol. Bioeng. 2007, 96, 203−209. Boca Raton, FL, Internet version, 2005.
(39) Tran, V. T.; Benoît, J. P.; Venier-Julienne, M. C. Int. J. Pharm. (76) Neumann, B.; Vincent, B.; Krustev, R.; Muller, H. J. Langmuir
2011, 407, 1−11. 2004, 20, 4336−4344.
(40) Pavlov, A. M.; Saez, V.; Cobley, A.; Graves, J.; Sukhorukov, G. (77) Sacanna, S.; Irvine, W. T. M.; Chaikin, P. M.; Pine, D. J. Nature
B.; Mason, T. J. Soft Matter 2011, 7, 4341−4347. 2010, 464, 575−578.
(41) Maan, A. A.; Schroen, K.; Boom, R. J. Food Eng. 2011, 107, (78) Sacanna, S.; Pine, D. J. Curr. Opin. Colloid Interface Sci. 2011, 16,
334−346. 96−105.
(42) van Dijke, K. C.; Veldhuis, G.; Schroen, K.; Boom, R. M. AIChE (79) Bogush, G. H.; Tracy, M. A.; Zukoski, C. F. J. Non-Cryst. Solids
J. 2010, 56, 833−836. 1988, 104, 95−106.
(43) Sugiura, S.; Nakajima, M.; Seki, M. J. Am. Oil Chem. Soc. 2002, (80) Haynes, W. M. CRC Handbook of Chemistry and Physics, 95th
79, 515−519. ed.; 2014−2015. Internet version http://www.hbcpnetbase.com/
(44) Nguyen, N. T.; Wang, C.; Teck, N. W.; Lee, N. L.; Soon, S. H. (accessed Jan 20, 2015).
International MEMS Conference 2006 2006, 34, 130−135. (81) Zhang, Y. M.; Miller, R. M. Appl. Environ. Microbiol. 1994, 60,
(45) Bringer, M. R.; Gerdts, C. J.; Song, H.; Tice, J. D.; Ismagilov, R. 2101−2106.
F. Philos. Trans. R. Soc. London, Ser. A 2004, 362, 1087−1104. (82) Zoldesi, C. I.; Steegstra, P.; Imhof, A. J. Colloid Interface Sci.
(46) Nisisako, T.; Torii, T.; Higuchi, T. Chem. Eng. J. 2004, 101, 23− 2007, 308, 121−129.
29.
(47) McClements, D. J. Food Emulsions: Principles, Practice, and
Techniques; CRC Press: Boca Raton, FL, 1999; p 7.
(48) Obey, T. M.; Vincent, B. J. Colloid Interface Sci. 1994, 163, 454−
463.
(49) van Blaaderen, A.; van Geest, J.; Vrij, A. J. Colloid Interface Sci.
1992, 154, 481−501.
(50) Zoldesi, C. I.; van Walree, C. A.; Imhof, A. Langmuir 2006, 22,
4343−4352.
(51) Zoldesi, C. I.; Imhof, A. Adv. Mater. 2005, 17, 924−928.
(52) O’Sullivan, M.; Zhang, Z. B.; Vincent, B. Langmuir 2009, 25,
7962−7966.
(53) Cui, J. W.; Wang, Y. J.; Postma, A.; Hao, J. C.; Hosta-Rigau, L.;
Caruso, F. Adv. Funct. Mater. 2010, 20, 1625−1631.
(54) Prestidge, C. A.; Barnes, T.; Simovic, S. Adv. Colloid Interface Sci.
2004, 108, 105−118.
(55) Prestidge, C. A.; Simovic, S. Int. J. Pharm. 2006, 324, 92−100.
(56) Ohta, T.; Nagao, D.; Ishii, H.; Konno, M. Soft Matter 2012, 8,
4652−4658.
(57) Nagao, D.; Ohta, T.; Ishii, H.; Imhof, A.; Konno, M. Langmuir
2012, 28, 17642−17646.
(58) Skirtach, A. G.; Yashchenok, A. M.; Möhwald, H. Chem.
Commun. 2011, 47, 12736−12746.
(59) Yaseen, M. A.; Yu, J.; Jung, B. S.; Wong, M. S.; Anvari, B. Mol.
Pharmaceutics 2009, 6, 1321−1332.
(60) Fernandes, P. A. L.; Delcea, M.; Skirtach, A. G.; Möhwald, H.;
Fery, A. Soft Matter 2010, 6, 1879−1883.

1719 DOI: 10.1021/cm504504t

Chem. Mater. 2015, 27, 1709−1719