Article

pubs.acs.org/Biomac

Microuidics-Assisted Diusion Self-Assembly: Toward the Control

of the Shape and Size of Pectin Hydrogel Microparticles
Melanie Marquis, Joelle Davy, Aiping Fang, and Denis Renard*
INRA, UR1268 Biopolymeres Interactions Assemblages, F-44300 Nantes Cedex, France
S Supporting Information
*

ABSTRACT: We demonstrated the generation of pectin hydrogel

microparticles having complex shapes either by combining the
phenomenon of gelation and water diusion-induced self-assembly
in microuidic channels (on-chip) or by the deformation of the
pregelled droplets outside the channels (o-chip) at a uiduid
interface. We proved that by tuning the mode of pectin crosslinking (CaCl2 vs CaCO3) and the degree of shrinking (water
content in the dimethyl carbonate (DMC) organic continuous
phase) we can control the shape of the nal particle. Sphere,
doughnut, oblate ellipsoid, or mushroom-type morphologies were
thus produced, demonstrating the ability to control the formation
of anisotropic biopolymer-based hydrogel microparticles using
microuidics. Shape changes were explained by the redistribution
of calcium ions in combination with the local Peclet number
experienced by the microdroplets during the on-chip process. Moreover, during the o-chip process, the interplay between elastic
and viscous forces for microdroplets entering the CaCl2DMC interface caused deformation of the pregelled droplets to occur
and therefore resulted in the formation of microparticles with a mushroom-like morphology.

generating nonspherical particles make use of a capillary owbased approach, such as microuidics, which often occurs in
combination with lithography and photopolymerization.1618
The capillary ow-based approach oers a number of
advantages over conventional ow-control technology or
manipulation techniques because it ensures highly versatile
geometry and can be used to produce monodisperse nonspherical polymeric microparticles with a diverse range of
shapes. For instance, Xu et al. and Dendukuri et al. formed
liquid droplets with a microuidic device, shaped the droplets
in a microchannel, and polymerized them to form solid
particles of several nonspherical geometries.19,20 Unfortunately,
the UV-polymerization process is very hazardous to biological
applications and UV-sensitive materials. Other groups have
developed shape-controlled microparticles assisted by microuidics using natural polymers such as polysaccharides.12,2126
Two strategies were adopted to control the shape of the
hydrogel microparticles: on-chip by tuning ow rates,
composition, and channel geometries15,26 and o-chip at uid
interfaces by tuning interfacial tension between the on-chip
synthesized microparticles and the collecting bath that generally
contains the cross-linker.2123 Alginate-based rods, threads,
plugs, disks, pears, and tadpoles or chitosan-based hemispherical and discoidal microparticles were thus obtained. In

INTRODUCTION
Microbeads or hydrogel microparticles have attracted great
attention over the past 2 decades owing to their wide
applications in drug delivery, adsorbents, and optical
devices.14 The performance and application of microbeads
generally depend on their size and shape. Nonspherical
microbeads are more desired because of their unique properties
such as anisotropic responses to external force and large surface
area as well as their use as building blocks for hierarchical
superstructures formation. Shape-controlled microbeads have
found applications in several elds, such as uses as the
components of new ow measurement devices, as regularly
sized beads in liquid crystal displays,5 or as model systems to
mimic anisotropic biological cells and to synthesize materials
with unique crystal structures.6 Moreover, for the biological and
medical areas, shape-controlled biocompatible and biodegradable hydrogel microparticles can be applied to drug delivery
systems79 and cell immobilization10 as well as cell cultures and
biosensing bers.11,12 However, methods to prepare shapecontrolled microparticles are limited. Conventional emulsion
and suspension polymerization to generate shape-controlled
nonspherical microparticles is unpractical because the minimization of interfacial energy leads to the formation of a
spherical shape during most syntheses.13,14 In recent years, a
number of reports have emerged that directly address this
bottleneck through the manipulation of previously fabricated
spherical particles into nonspherical geometries and ab initio
synthesis of nonspherical particles.15 Synthesis methods for
2014 American Chemical Society

Received: October 30, 2013

Revised: January 30, 2014
Published: March 27, 2014
1568

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

Figure 1. Schematic of the microuidic ow-focusing geometry used for pectin hydrogel microparticle formation. A pressure-driven ow controller
was used to supply the three liquids to the device: pectin or pectin + carbonate calcium (CaCO3) (inlet 1), anhydrous or water-saturated dimethyl
carbonate (DMC) (inlet 2), and DMC + calcium chloride (CaCl2) or DMC + acetic acid (inlet 3). An outlet capillary tubing allowed the hydrogel
microparticles to be collected in a CaCl2 bath. Zone A: pectin microdroplet generation zone resulting from the ow-focusing geometry. Zone B:
solvent diusion zone resulting from the anhydrous dimethyl carbonate (DMC). Zone C: cross-linking zone resulting from the diusion of calcium
ions within the pectin microdroplets. Micrographs of droplets in microchannels owing in the three zones are also reported.

with divalent ions such as calcium and to form a self-supporting

gel via the egg-box model, a model that is also valid for
alginate gelation.32 In addition, it was surprising that the
formation of pectin microbeads using microuidics has been
reported only recently through the production of spherical
hydrogel microparticles as core materials for the encapsulation
of gold and silver nanoparticles.33 In the same way, anisotropic
Janus microparticles made of pectinpectin and pectin
alginate were also recently produced using microuidics.26
The aim of our study was to extend the f DISA process to a
biopolymer through the combined strategy of biopolymer
gelation and shrinking within a microuidic chip. This approach
combined the eect of a cross-linker to solidify the droplets
with the eect of a diusion-induced self-assembly process that
used an organic solvent, dimethyl carbonate (DMC), to shrink
and deform the pregelled droplets because of its water
immiscibility and solubility properties. Pectin, as a renewable
resource, and CaCl2 or CaCO3, as a source of calcium, were
used to form hydrogel microparticles, and their shape was
controlled on-chip by solvent diusion or o-chip by
deformation of the pregelled droplets at a uiduid interface.

addition, signicant dierences in the release behavior of drugloaded alginate hydrogel microparticles that were dependent on
their morphology (spherical, pear-like, or mushroom shape)
were clearly demonstrated.23 An alternative strategy was
recently adopted to control the shape of microparticles,
which is called microuidic diusion-induced self-assembly
(f DISA).27,28 By exploring the phenomenon of water
diusion-induced self-assembly of silica particles in microuidic
channels using dimethyl carbonate as the organic phase, Fang et
al. demonstrated that both the geometric connement
experienced by the droplet and the local Peclet number are
responsible for the nal particle shape.27,28 Two shape zones, a
doughnut or bowl-like morphology, were clearly distinguishable
depending on the geometric connement and the Peclet
number. The authors argued that these shape-controlled
morphologies, where the solgel transition of silica was
assisted by a f DISA process, could be extended to other
complex uids such as nanoparticles or biopolymers. It is
therefore interesting to note that the dimethyl carbonate
organic phase, because of its water immiscibility and solubility
properties, was also previously used to shrink alginate
biopolymer microparticles within a microuidic device to
reduce the size of the droplets.29 Such a process that is able to
control both shape and size was, however, never demonstrated
on polysaccharides such as, for example, pectin. The advantages
of pectin compared to other polysaccharides come from its
structural diversity, particularly charge density and branching,
that depends on its botanical source. The charge density
resulting from the presence of carboxylic groups, either
randomly or blockwise distributed, as well as the branches
present along the pectin chains will give unique cross-linking
properties (i.e., kinetics and density of cross-linking) to pectin
compared to other polysaccharides such as alginate.30,31 The
carboxylic groups of galacturonic acid in pectins can be methylesteried, with the degree of esterication (DE) distinguishing
high-methoxylated (HM) pectin from low-methoxylated (LM)
pectin. LM-pectins (with a DE < 50%) are able to cross-link

EXPERIMENTAL SECTION

Fabrication of the Microuidic Flow-Focusing Device.

Microuidic devices were prepared using poly(dimethylsiloxane)
(PDMS) (RTV 615, Elecoproduit, France) and a soft lithography
technique.34 SU-8 2100 (CTS, France) positive-relief structures were
produced on silicon wafers (Si-mat, Germany). PDMS polymer (in a
mixture of 10:1 base polymer/curing agent) was cast from this mold,
and access holes were punched on the PDMS layer. The PDMS layer
(in a mixture of 10:1 base polymer/curing agent) was then placed in
contact with a thin PDMS layer (in a mixture of 20:1 base polymer/
curing agent), which was previously poured and xed in a Petri dish
(to avoid PDMS deformation resulting from anhydrous DMC), to
generate the microchip. The cross-linker diused as a result of the
gradient from PDMS (10:1) to PDMS (20:1). The chip was then
oven-treated at 60 C for 24 h to strengthen the cross-linking. The
microchannels were rectangular in shape with a uniform height (h) of
1569

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

Figure 2. Mode of pectin cross-linking in zone C. External gelation occurred by the diusion of calcium ions (CaCl2) from the DMC continuous
phase to the pectin microdroplets; internal gelation occurred by dissolution of CaCO3 particles through the diusion of acetic acid from the DMC
continuous phase to the pectin microdroplets.
100 m. The microsystem was made of three zones: a ow-focusing
junction where droplets of pectin were formed (zone A), a rst
serpentine channel where droplets shrank by solvent diusion (zone
B), and a second serpentine channel where cross-linking within pectin
microdroplets took place through calcium bridging (zone C) (Figure
1). The widths and lengths of serpentine channels in zones B and C
were, respectively, 200 m and 52 cm and 450 m and 98 cm.
Dimensions of the channels in the microuidic device and at the
droplet formation zone were determined by prolometry (Figure S1,
Supporting Information).
Chemicals and Solutions. Low-methoxyl citrus pectin (Mw = 169
250 g/mol, Mw/Mn = 2.03) was purchased from Cargill France SAS,
had a DE value of 30%, and contained 78.5% galacturonic acid. Pectin
was solubilized in 50 mM MES buer pH 6 at 40 C or in deionized
water at room temperature under magnetic stirring for 2 h and was
then stored at 4 C. The pectin solutions were degassed under vacuum
and ltered through a 5 m lter (Puradisc lter, FP 30 mm 5.0CN,
Whatman). The pH value of pectin solutions prepared in deionized
water was then adjusted to around 7 with 1 M NaOH. Anhydrous
dimethyl carbonate 99% (DMC) (no. D152927-1L), calcium chloride,
carbonate calcium, and acetic acid were purchased from Sigma-Aldrich
(France). N-(3-(Dimethylamino)propyl)-N-ethyl-carbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) (SigmaAldrich, France) were used for covalent coupling of uoresceinamine
isothiocyanate (FITC; Sigma-Aldrich France; exc 488 nm and em
500530 nm) to citrus pectin through activation of the polysaccharide
carboxyl groups, as described by Ogushi et al.35
Preparation of Pectin Hydrogel Microparticles. Emulsication and Shrinkage of Pectin Microdroplets. Pectin microdroplets
were achieved by emulsifying a 5, 10, or 20 g/L aqueous solution of
pectin in DMC as the continuous oil phase in the microuidic devices.
The aqueous liquid phase broke up in the continuous oil phase to
generate W/O emulsion microdroplets in zone A (Figure 1). The two
immiscible liquids were supplied to the microchannels using a
Microuidic Flow Control Systems MFCS-4C-1000 (Fluigent,
France) as a pressure-driven ow controller. The ow rates varied in
the range of 12 L/min for inlet 1 and 33120 L/min for inlet 2.
The PDMS device was linked to the MFCS through a poly(tetrauoroethylene) (PTFE) tube (0.3 mm i.d. 0.76 mm o.d.
with a length of 350 mm) (Fisher Scientic, no. 39240). The size of
the microdroplets was controlled by varying the pressure of the two
immiscible liquids. DMC with water content in the range of zero to
saturation was used as the continuous phase, which allowed the

shrinkage in zone B to occur by solvent diusion. Water-saturated

DMC was prepared by magnetically stirring DMC in the presence of a
large quantity of water for at least 2 h. The bottom phase was used as
water-saturated DMC. DMC with water content between zero and
saturation was prepared stoichiometrically by mixing DMC with its
water-saturated counterpart. Images of microdroplets formation and
shrinkage were captured using an Olympus IX51 inverse microscope
(Olympus, France) equipped with a digital camera (Hamamatsu
C4742-95).
Pectin Hydrogel Microparticles Formation. Pectin cross-linking
occurred in zone C by an internal or external gelation mode (Figure
2). The ow rate varied in the range of 150370 L/min for inlet 3.
The internal gelation mode was induced by a pH decrease inside the
microdroplets. The microdroplets contained pectin (5, 10, or 20 g/
L) and CaCO3 (2.5, 5, or 10 g/L) as the cross-linking agent in an
inactive form. The continuous phase (DMC) was saturated by an
aqueous acetic acid solution (0.5 wt %) that diused into the
microdroplets and triggered the release of Ca2+ ions, resulting in the
cross-linking of pectin chains and the formation of a network. The
external gelation mode was performed by the diusion of a solubilized
cross-linker from the continuous phase to the microdroplets. The
continuous phase (DMC) was saturated by an aqueous calcium
chloride (CaCl2) solution (9 g/L). Cross-linking was induced by
the rapid diusion of calcium ions inside the pectin microdroplets,
which resulted in the formation of a network. In both cases, pectin
hydrogel microparticles were collected in a bath of CaCl2 (20 g/L)
solution.
Drying and Swelling Behavior of Pectin Hydrogel Microparticles.
The swelling behavior of hydrated and driedrehydrated microparticles was studied for three morphologies: spherical, mushroomlike, and doughnut-like. First, we exchanged the CaCl2 solution
corresponding to the collected medium of the hydrogel microparticles
with the three dierent buers under investigation: 50 mM NaCl/HCl
buer, pH 1.2; 500 mM MES buer, pH 5.0; and 500 mM Tris/HCl
buer, pH 7.4. This solvent exchange was done by centrifugation at
1000g for 5 min.
One fraction of the three types of morphologically distinct hydrated
microparticles was thereafter dried using anhydrous DMC by three
centrifugation steps at 1000g for 5 min each. DMC was then
evaporated under dried air for 3 to 4 days (under a gloves box). Dried
microparticle samples were rehydrated under saturated vapor pressure
at 20 C using NaBr (Aw = 0.57, near-ambient temperature) by adding
buer solutions at pH 1.2, 5.0, and 7.4. These microparticles will be
1570

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

Figure 3. Scanning electronic microscopy images of pectin microparticles obtained by (A) solvent diusion resulting from anhydrous DMC without
a gelation step, (B) solvent diusion with external gelation for (b1) an initial Cpectin = 5 g/L or (b2) initial Cpectin = 10 and 20 g/L, (C) solvent
diusion with internal gelation for (c1) an initial Cpectin = 510 g/L or (c2) initial Cpectin = 20 g/L, and (D) without solvent diusion and (d1)
external gelation for an initial Cpectin = 10 g/L or (d2) internal gelation for initial Cpectin = 20 g/L. Scales bars: 10 m.

denoted hereafter as driedrehydrated microparticles and will be

compared with the initial hydrated microparticles in the three dierent
buers.
Structural Characterization of Pectin Hydrogel Microparticles by Phase Contrast, Confocal, and Scanning Electron
Microscopies. Freshly Prepared Pectin Hydrogel Microparticles.
The microparticles collected o-chip were analyzed using an Olympus
IX51 inverse microscope (Olympus, France) equipped with phasecontrast and bright-eld modes of illumination and a digital camera
(Sony, XCD-SX90). Some hydrogel microparticles with nonspherical
shapes were also observed using a Nikon Ti-E with a C1si scanning
laser confocal microscope (Nikon, France) and a Nikon Eclipse Ti
inverse microscope (Nikon, France) after pectin labeling using FITC.
FITC emission uorescence was recorded between 500 and 530 nm
after excitation at 488 nm.
For scanning electron microscopy (SEM), hydrogel microparticles
were prepared by the supercritically dried method, which is known to
retain most of the volume of the parent hydrogel and to present high
surface areas.36 Liquid CO2 and water are not directly miscible. As a
consequence, to prepare the samples for supercritical drying, the
hydrogel microparticles had to be dehydrated by immersion for 15 min
in hydroalcoholic solutions of increasing ethanol concentration (10,
30, 50, 70, 90, and 100%) and then in anhydrous ethanol and acetone.
Acetone was later exchanged by liquid CO2 in a Balzers Union
CPD030 chamber in which the samples were heated beyond the CO2
critical point (20 C, 50 bar) and outgassed by successive purges of the
chamber to obtain the aerogel microparticles. The temperature and
nal pressure were then increased to 40 C and more than 73.8 bar,
respectively, to be under CO2 supercritical conditions. Samples were
then stored in a desiccator containing anhydrous P2O5. Samples
observations were made using a JEOL 6400F microscope operating at
3 kV on samples mounted on double-sided carbon grids after a gold/
palladium thin coating was applied for 5 min. First, we checked bulk
spherical gels (D 2 cm, C = 10 g/L) to ensure that the
supercritical drying process did not impact the shape of the aerogel
and that only size changes occurred (a 40% reduction in size was
noticed; data not shown).
Hydrated and DriedRehydrated Pectin Hydrogel Microparticles.
The swelling behavior of hydrated and driedrehydrated microparticles, with spherical, mushroom, or doughnut-like shapes, was
followed after 24 and 48 h of incubation at 37 C using a phasecontrast optical microscope (Olympus, France) equipped with a digital
camera (Sony XCD-SX90, Japan). A comparison was made between
the hydrated and driedrehydrated states in the three dierent buers
(pH 1.2, 5.0, and 7.4) by measuring the diameters (or equivalent
diameter when nonsphericity occurred) of the microparticles using
ImageJ v1.35c.

RESULTS AND DISCUSSION

Two alternative processes coupled to two modes of gelation
were used to produce pectin hydrogel microparticles. One
process involved a shrinkage step through a solvent-diusion
process before internal or external gelation took place, and the
other process allowed the production of pectin hydrogel
microparticles by internal or external gelation without a
preliminary shrinkage step. Contrary to the work of Rondeau
and Cooper-White29 on alginate production by microuidics
where solvent diusion and gelation occurred simultaneously,
pectin hydrogel microparticles formation in our study was
followed either by a change in their size resulting from the
solvent-diusion process or by modications of both their size
and shape resulting from the elapsed time between the solventdiusion and gelation processes (Figure 1). In addition, the
properties of the DMC continuous phase toward water
solubility allowed the solvent diusion either to proceed
(when DMC was anhydrous) or not (when DMC was watersaturated). It is important to note that the solubility of pure
water in pure DMC is 3 wt % at room temperature, whereas the
solubility of pure anhydrous DMC in pure water is 12.7 wt %.37
In addition, it was recently demonstrated that the carbonyl
oxygen of DMC serves as the proton acceptor for the
hydrogen-bonded complex with water.38 The formation of
these complexes would allow for a partial solubility of DMC in
water. This partial solubility of the DMC continuous phase
ensured that the exchange was very slow, allowing any on-chip
precipitation event to occur.
Solvent Diusion without a Gelation Process. First, we
looked at the reproducibility of our approach by applying only
solvent diusion to the pectin microdroplets without internal or
external gelation to evaluate the propensity of DMC both to
reduce the size of the pectin microdroplets and to condense
them further to eventually produce solid microparticles without
the help of a cross-linker. Two immiscible liquids were thus
supplied to the ow-focusing device (FFD) by inlet 1 with a 1
wt % pectin solution and inlet 2 with anhydrous DMC,
respectively, whereas water-saturated DMC was supplied by
inlet 3. In this way, spherical microparticles with smooth
surfaces were obtained with an average diameter of 42 m
(Figure 3A). The microparticles formation was due to the slow
extraction of water from the microdroplet, which inhibited any
precipitation event resulting from the contact between the
apolar DMC solvent and the biopolymer. A 3-fold reduction in
1571

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

Table 1. Diameter (D; m) of Pectin Microdroplets and Hydrogel Microparticles during the Dierent Steps of the On-Chip
Process and the O-Chip Microscopic Observationsa
diameter (m)
external gelation (CaCl2)
and solvent diusion
C = 5 g/L
C = 10 g/L
C = 20 g/L
external gelation (CaCl2)
without solvent diusion
C = 10 g/L
internal gelation (CaCO3)
and solvent diusion
C = 5 g/L
C = 10 g/L
C = 20 g/L
internal gelation (CaCO3)
without solvent diusion
C = 5 g/L
C = 10 g/L
C = 20 g/L
no gelation and solvent diusion

zone A zone B

zone C

phase contrast
microscopy

scanning electron microscopy

morphological transition

130
135
135

30
45
48

30
45
45

42
53
57

22
24
42

spherical doughnut
spherical spherical
spherical spherical

130

130

130

175250b

120200b

spherical anisotropic

120
120
125

30
30
40

30
30
40

40
40
53

24
31
42

spherical doughnut
spherical doughnut
spherical doughnut

120

120

120

115

130
125

130
40

140
40

140
44

18 (sphere)23 (doughnut)
45 (sphere)32 (doughnut)
46
42

spherical
spherical
spherical
spherical

spherical and doughnut

spherical and doughnut
ellipsoidal (oblate)
spherical

The dierences in size between the on-chip and o-chip optical microscopy observations arose from some dehydration processes resulting from
osmotic pressure dierences because of the high calcium concentrations on-chip, thus reducing the size of the on-chip microdroplets and
microparticles. For the o-chip phase-contrast and scanning electron microscopy observations, discrepancies came from the dehydration process
mandatory for SEM sample preparation (despite the supercritically dried method used). The on-chip/o-chip morphological transition was,
therefore, conrmed by both optical and electron microscopies, giving strong condence to the absence of morphological changes induced by SEM
sample preparation. bMaximum length corresponding to the sum of the two major axes (B + D) of mushroom-like microparticles (see Table 2).

size was obtained from the initial microdroplets (D 125 m)

(Table 1). This size reduction provided strong evidence of
signicant losses of water to the DMC phase during the
transport of the microdroplets downstream. In this more simple
case, water diused out of the pectin microdroplets to the
DMC phase in zone B, whereas no chemical reactions occurred
in zone C. The reaction conditions were thus unmodied by
the competing gelation process and led to condensed pectin
microparticles with homogeneous smooth surfaces. In addition,
the condensed biopolymeric state allowed the spherical shape
to be preserved similar to that which was observed in alginate
microparticles formation using anhydrous DMC as the
continuous phase.25 An estimation of the pectin concentration
from the volume of the microparticles gave 204 g/L (from
an initial pectin concentration of 11 g/L) (Table 2). The
high pectin concentration obtained after solvent diusion that
led to a very condensed phase within the microdroplets was
conrmed by the unchanged of size of the microparticles
observed by phase-contrast optical and scanning electron
microscopies (Figure S2, Supporting Information). It was
therefore demonstrated that a 40% reduction in size was
observed by SEM after the supercritical drying process on bulk
spherical gels at low pectin concentration.
Solvent Diusion and (External or Internal) Gelation.
A two-step process was adopted through a combination of a
solvent-diusion process, to shrink the pectin microdroplets
and to reduce their size, and a gelation process, to cross-link
pectin chains within microdroplets by cation bridging through
external or internal gelation (Figure 2). Two immiscible liquids
were thus supplied to the FFD by inlet 1 with a 5, 10, or 20 g/
L pectin solution (or pectin + CaCO3) and inlet 2 with
anhydrous DMC, respectively, whereas CaCl2-saturated DMC
(or acetic acid-saturated DMC) was supplied by inlet 3. It is
important to note that because of the solubility properties of

DMC in water (3 wt %) it was necessary, for external gelation,

to solubilize calcium chloride in DMC at a concentration of 300
g/L, giving a nal soluble CaCl2 concentration in DMC of 9
g/L. For the internal gelation process, acetic acid was
solubilized in DMC at 50 g/L, giving a nal soluble acetic
acid concentration in DMC of 1.5 g/L. As a consequence,
these solubility conditions could induce some osmotic pressure
balance between the two immiscible liquids within the
microchannels, particularly for the external gelation process,
where the CaCl2 concentration was high. These dierences, if
they exist, will be evaluated by microscopy through the
comparison of the on-chip and o-chip microdroplet and
microparticle sizes.
Solvent Diusion and External Gelation. Hydrogel microparticles formed by solvent diusion and external gelation
through CaCl2 ion bridging displayed either a doughnut-like or
spherical-like shape depending on the initial pectin concentration (Figure 3B and Figure S3, Supporting Information).
Deformation of pregelled microdroplets was seen by confocal
microscopy (Figure S3, b1, Supporting Information), highlighting both the eect of anisotropic solvent diusion on the
formation of doughnut-like particles and the nonequilibrium gel
state of microparticles at the end of the on-chip process. The
nonequilibrium gel state hypothesis is based on the spherical
shape of microparticles obtained when only solvent diusion
proceeds, and it was previously demonstrated that the binding
of anionic polysaccharides such as pectin to their adapted ion of
choice takes place on the order of milliseconds to seconds.39
The anisotropic shape of microparticles was, however, not
maintained at pectin concentrations of 10 and 20 g/L,
presumably because of the higher pectin/calcium molar ratio,
which led to an equilibrium gel state of the hydrogel
microparticles at the end of the on-chip process. A 3-fold
reduction in size was also noticed for microparticles formed
1572

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

Table 2. Volumes of Pectin Microdroplets and Hydrogel Microparticles and Evaluation of Their Pectin Concentrations during
the Dierent Steps of the On-Chip Process and the O-Chip Optical Microscopic Observations

a
Zones A and C: the volume when droplets were slightly deformed was calculated assuming an oblate ellipsoid with a volume V = (4/3)(abc), with
a = b = l/2 and c = h/2. The same calculation was applied for microparticles issued from internal gelation without solvent diusion (C = 20 g/L).
b
The equation of a sphere was used for the doughnut-like and ellipsoidal shapes, which were not clearly distinguishable by optical microscopy but
only by SEM.

only at the periphery of the microdroplets during the gelation

step, inducing the formation of doughnut-like shapes in the
nal hydrogel microparticles. The doughnut-like shape results
from the limited diusion of protons during the fast transition
of the droplets into channels from the continuous phase, which
releases calcium ions at the periphery of the droplets.
Internal vs External Gelation Comparison. The sizes of the
hydrogel microparticles determined from optical and electron
microscopy images were similar regardless of the mode of
gelation (Table 1). However, on the basis of the doughnut-like
shape generated from the internal CaCO3 gelation and solventdiusion process, the pectin concentrations were lower
compared to the previous external gelation process involving
CaCl2 ion bridging. In addition, it was explicitly demonstrated
that the pectin did not signicantly inuence the dissolution
rate of CaCO3.40 This argument means that regardless of the
type of calcium used in the gelation process, CaCl2 versus
CaCO3, the morphological transitions would be governed only
by the competing phenomenon of calcium diusion and
redistribution during the solvent-diusion step and by the
pectin concentration. Therefore, an increase in the pectin
concentration, leading to an increase in the viscosity, would
prevent calcium ions from redistributing, as spherical hydrogel
microparticles were observed for the high pectin concentration
used in the case of the external gelation and solvent-diusion
process. Internal gelation combined with solvent diusion
would thus be a good way to control the shape of the hydrogel

after the external gelation and solvent-diusion processes, as

was the case in the absence of gelation (Table 1). In addition,
the size of the hydrogel microparticles was also reduced after
the supercritical drying process used for the SEM observations
(from 30 to 50%), but this occurred without shape changes.
The reduction in size was lower when the pectin concentration
increased, again highlighting the equilibrium gel state reached
for the concentrated pectin microparticles. The pectin
concentrations within the hydrogel microparticles were similar
to those obtained in the absence of gelation (Table 2),
conrming that the solvent-diusion process was not prolonged
during the gelation step.
Solvent Diusion and Internal Gelation. Hydrogel microparticles formed by solvent diusion and internal gelation
through CaCO3 ion solubilization and bridging displayed a
doughnut-like shape regardless of the initial pectin concentration (Figure 3C and Figure S4, Supporting Information).
The 3-fold reduction in size induced by solvent diusion was
again conrmed in this case, but it appears that the gelation
process, induced by the solubilization of CaCO3 particles after
acidication by the diusion of acetic acid from the continuous
DMC phase to the aqueous pectin microdroplets, is the origin
of the anisotropic shape of the hydrogel microparticles. Indeed,
the calcium within the microdroplets at the beginning of the
process is preferentially distributed to the periphery of the
droplets during the solvent-diusion step; the asymmetric
distribution of calcium causes the pectin chains to cross-link
1573

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

surfaces, allowing the oatability and the subsequent

deformation of the alginate drops during gel formation at the
airCaCl2 interface to occur. In our study, the interfacial
properties of the CaCl2DMC interface could be the driving
force for the deformation of the pregelled microdroplets. The
pregelled microdroplets will become oblate because of the
interplay between interfacial tension and gravity during the
falling process. A protrusion from the oblate-shaped microdroplets would allow a tail to gradually grow owing to the
mushroom-like morphology that was formed when the
pregelled microdroplets crossed the CaCl2DMC interface.
However, Rondeau and Cooper-White29 noted that the
interfacial tension for alginate was very close to the value for
water and DMC (water/DMC = 5.31 vs alginate/DMC = 4.95 mN/
m) and did not depend on the alginate concentration. This
nding may call into question the driving force of microdroplet
deformation in our pectin hydrogel microparticles when they
crossed the CaCl2DMC interface. The most probable
hypothesis is that our pectin sample at the concentrations
used (520 g/L) and/or our DMC continuous phase,
because it was CaCl2-saturated, would lead to a much greater
interfacial tension value, allowing for the deformation of the
pregelled pectin microdroplets. This hypothesis is reinforced by
the fact that hydrogel microparticles formed by the same
process but with an additional solvent-diusion step were never
deformed because of their higher pectin concentrations, with
the condensed state of the particles being unaected o-chip by
the CaCl2DMC interface. In conjunction with previous
hypotheses, the nonuniformity of cross-linking during the
gelation process could lead to deformation at the CaCl2DMC
interface (even if apparent uniformity appeared by CSLM on
the micrometric scale).
Internal Gelation without Solvent Diusion. Pectin
hydrogel microparticles formed by internal gelation through
CaCO3 ion solubilization and bridging without solvent diusion
displayed spherical and doughnut-like shapes at initial pectin
concentrations of 5 and 10 g/L, whereas an oblate elliptical
shape was obtained at C = 20 g/L (Figure 3D, d2 and Figure
S6, Supporting Information). These morphologies were,
however, clearly distinguishable only by electron microscopy.
The transition morphology observed during this process was
very similar to what was observed in the process where solvent
diusion occurred, with the main dierence being observed for
pectin hydrogel microparticles formed at C = 20 g/L.
Although a doughnut-like morphology was observed when
solvent diusion occurred, an oblate elliptical shape with a
raspberry-like morphology was visible when solvent diusion
did not occur. Pectin concentrations and sizes determined from
optical microscopy images again conrmed that solvent
diusion was ineective during this process (Tables 1 and 2).
The slight dierences observed between the initial and nal
pectin concentrations are probably due to the fact that the
volumes from the optical microscopy images were calculated
assuming a spherical shape, whereas SEM images clearly
revealed anisotropic shapes in the nal hydrogel microparticles
structure. In addition, the size of the microparticles determined
from SEM observations was quite similar regardless of the onchip process used or the initial pectin concentration (Table 1).
The comparable sizes obtained in the absence of solvent
diusion versus with solvent diusion may call into question
the driving force that is responsible for this important reduction
in size. A unique and most probable hypothesis is that the
sample preparation for SEM observations aects the sizes of the

microparticles, whereas modulation of the pectin concentration

during the external gelation process combined with solvent
diusion would be much more appropriate to guarantee the
production of pectin microparticles with highly condensed
phases for heterogeneous catalysis applications.
External (or Internal) Gelation without a SolventDiusion Process. In this second two-step process, the DMC
continuous phase was saturated with water to prevent solvent
diusion, and gelation (external or internal) was then applied
on the basis of calcium ions bridging. It was demonstrated that
the inverse process (i.e., external or internal gelation preceding
the absence of solvent diusion resulting from the watersaturated DMC) did not aect the size and shape of the pectin
hydrogel microparticles (data not shown).
External Gelation without Solvent Diusion. Pectin
hydrogel microparticles formed at C = 10 g/L after external
gelation without solvent diusion were spherical or slightly
elliptical at the end of the on-chip process but were drastically
deformed when they fell into the CaCl2DMC interface
collecting tube. The resulting anisotropic shape of the
microparticles with a mushroom-like morphology was clearly
visible by optical and electron microscopies (Figure 3D, d1). In
addition, confocal microscopy (CSLM) images revealed a
homogeneous pectin distribution within the hydrogel microparticles (Figure S5c, Supporting Information). The maximum
length corresponding to the sum of the two major axes of the
mushroom-like shape of the microparticles ranged from 175 to
250 m by optical microscopy and from 120 to 200 m by
SEM (Figure S5, Supporting Information and Table 1). Pectin
concentrations determined from the volume of these
anisotropic microparticles conrmed that solvent diusion did
not occur during this two-step on-chip process (Table 2). More
precisely, the decrease of the pectin concentration in the
microparticles was mainly due to the o-chip increase of the
volume in the anisotropic mushroom-like microparticles when
they crossed the CaCl2DMC interface, consequently diluting
their initial pectin concentration. The anisotropic shape
obtained in this case was in accordance with previous results,
when solvent diusion occurred, where the hydrogel microparticles were in a nonequilibrium state at the end of the onchip process. The pre-cross-linking reaction of the droplets in
the DMC phase saturated with CaCl2 did not completely
solidify them. Thus, when the droplets diused through the
CaCl2DMC interface, the eect of an impact force led to the
formation of a at bottom of the droplet, and the upper half of
the droplets retained a spherical shape. More interestingly, the
shapes of the hydrogel microparticles were maintained after
supercritical drying (Figure 3D, d1 and Figure S5b, Supporting
Information). A similar morphology was previously demonstrated in the case of alginate microgel particles where the
authors, by combining microuidic and external ionic crosslinking, were able to produce mushroom-like, hemispherical,
and red blood cell-like morphologies.25 The control of shape
was driven by simply varying the gelation conditions (viscosity
of the gelation bath, collecting height, and interfacial tension),
and the authors explained that the mushroom-like morphology
was due to an interplay between the viscous and elastic forces
of prolate microdroplets during the subsequent gelation process
inside the collecting tubing. Another recent paper reported the
production of macroscopic mushroom-like structures made of
alginate.41 The authors controlled the anisotropic morphology
of the hydrogel particles by tuning the viscosity of the alginate
drops, the impact velocity, and the falling height on the liquid
1574

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

Figure 4. SEM micrographs of the surface of pectin hydrogel microparticles. Scale bars: 1 m. (A) No gelation and solvent diusion. Pectin/
anhydrous DMC/water-saturated DMC. (B) External/internal gelation and solvent diusion. Pectin + CaCO3/anhydrous DMC/CaCl2-saturated
DMC. (C) Internal gelation without solvent diusion. Pectin + CaCO3/water-saturated DMC/acetic acid-saturated DMC. (D) External gelation
without solvent diusion. Pectin/water-saturated DMC/CaCl2-saturated DMC.

the droplets was very fast; thus, pectins within the droplets had
insucient time to diuse from the surface to the center of the
droplet. The microuidic drying and condensation can thus be
tuned by varying the parameters of D0/h and Pe, providing a
starting point to control the geometry of a 3D microstructure
by droplet-based microuidics. To the best of our knowledge,
this is the rst test of the ability to accurately engineer
anisotropy in pectin particles by an on-chip microuidics
approach.
To summarize the on-chip/o-chip morphological transitions induced by ionic gelation and the presence or absence of
solvent diusion from initially spherical (or elliptical) microdroplets, two routes of gelation (internal or external) lead to
isotropic and anisotropic hydrogel microparticles. Although the
on-chip process is very powerful and reproducible for tuning
the shape as a function of connement, the pectin
concentration, and the mode of gelation, better control of the
o-chip mushroom-like morphology by designing a new
microuidic device to deform on-chip pregelled microdroplets
would be very promising. This approach is currently under
investigation to create very original biopolymeric hydrogel
microparticles of tunable size for controlled release applications.
Surface Microstructure of Pectin Hydrogel Microparticles. The microstructure of the surface of the pectin
hydrogel microparticles, probed by SEM, was useful for
attempting to correlate microuidic diusion-induced selfassembly and surface porosity. The solvent-diusion process
coupled to gelation (internal or external) or to the absence of
gelation did not have a drastic eect on the surface
microstructure of the pectin hydrogel microparticles (Figure
4). The rather smooth surfaces were uniformly covered because
of condensation and subsequent pectin chain merging was
induced by water uptake from the DMC continuous phase. In
addition, the mode of gelation and the pectin concentration had
no impact on the surface of the microparticles, in contrast to
what was observed for their shape. Without solvent diusion,
the morphological transitions previously observed for the

resulting structures after supercritical drying regardless of their

shape. The doughnut-like shape of the hydrogel microparticles
obtained without a solvent-diusion step would mean that
CaCO 3 particles (or proton ions) would be partially
redistributed during the on-chip process because of the higher
convection phenomena occurring within the microdroplets in
the absence of a solvent-diusion process, leading to slightly
anisotropic structures.
Peclet Number and Shape Changes. The doughnut-like
shape of the hydrogel microparticles obtained by internal or
external gelation and solvent diusion could be rationalized by
considering the Peclet number experienced by the microdroplets. Fang et al. identied two zones corresponding to two
dierent particle shapes using DMC as the continuous phase
for silica particles: either a doughnut-like or bowl-like shape for
Pe 1.28 In addition, the morphological transition from a
bowl-like to a doughnut-like shape occurred for D0/h = 0.8,
where D0 is the initial droplet size and h is the serpentine
microchannel height. In our study, D0/h was equal to 1.3 for
external gelation and 1.2 for internal gelation on-chip processes,
with h = 107 m. The Peclet number was dened by the
equation Pe = (R2/dD), where R is the radius of the initial
microdroplet (R = 1/2D0), D is the pectin diusion coecient,
and d is the time required for a microdroplet to dry. From the
ow rates applied in the solvent-diusion zone and the length
of the serpentine, the drying time, d, was calculated to be
between 6.7 and 10 s. The resulting Pe numbers experienced by
the microdroplets in this zone were between 15 and 30
regardless of the mode of gelation. In conclusion and according
to the rationalization of the shape changes occurring with the
DMC continuous phase, our respective D0/h and Pe number
values determined for internal and external gelation fell in the
doughnut-like regime identied by Fang et al.28 This theoretical
interpretation of particle shape changes resulting from the eect
of connement for high Peclet numbers (where convection
dominates over diusion) was in line with our results obtained
for pectin hydrogel microparticles. For Pe 1, the drying of
1575

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

Figure 5. Optical microscopy images of pectin hydrogel microparticles with spherical (a), doughnut-like (b), and mushroom-like (c) morphologies
after drying (1) and rehydration during 48 h of incubation at 37 C in 500 mM MES buer, pH 5.0 (2). Scales bars: 50 m.

particles, the increase in size was of at most 13% at pH 1.2,

whereas no size changes occurred at pH 5.0 and 7.4 (data not
shown). These rst results indicate that the swelling capacity of
hydrated pectin microparticles is rather limited regardless of
their shape, pH, and incubation time.
In contrast, the driedrehydrated pectin microparticles
displayed morphological changes rst because of the drying
process and second because of the swelling capacity as a
function of the pH after rehydration. Figure 5 illustrates these
changes for the three morphologies studied (sphere, doughnutlike, and mushroom-like shapes) in 500 mM MES buer, pH
5.0, after 48 h of incubation at 37 C. Although the sphere and
doughnut-like shapes were reduced in size only after drying, 36
and 60%, respectively, without morphological changes, it was
noticed that the mushroom-like microparticles were rather
highly compacted, which is certainly due to buckling and
condensation phenomena occurring during the drying process,
with a decrease in size of 13% compared to the hydrated
microparticles. The rehydration of dried microparticles led to
an important increase in size for the spherically shaped
microparticles of 79.8, 105, and 118% at pH 1.2, 5.0, and 7.4,
respectively. In addition, the driedrehydrated microparticles
had slightly larger sizes after 48 h of incubation at 37 C
compared to those hydrated at pH 5.0 and 7.4. The opposite
was true of the mushroom-like morphology microparticles,
where their size did not change with time after rehydration
regardless of the pH of the three buers. The mushroom-like
shape was, however, partly recovered after rehydration. In
addition, the sizes of the microparticles were always smaller
than the corresponding hydrated ones. This last result indicates
that the drastic changes occurring during the drying process
lead to irreversible changes in the microstructure of the pectin
network, resulting in a limited swelling capacity. Dried
rehydrated pectin microparticles with doughnut-like shapes
displayed an increase in size only at pH 1.2, showing a 29%
increase in length and a 14% increase in width. The behavior of
the driedrehydrated doughnut-like microparticles was similar
to hydrated ones regardless of the pH of the buer, highlighting
the full recovery of the swelling capacity and therefore the
absence of pectin network microstructural changes after
dehydration.
These preliminary results concerning the swelling capacity of
hydrated and driedrehydrated pectin microparticles will open
new perspectives for the potential applications of these

microparticles were accompanied by drastic changes on their

surfaces. In the case of internal gelation, the surfaces were
irregular with an inhomogeneous pectin distribution. An
interconnected pectin chain network was clearly visible on
the surface of the microparticles, with a pore size distribution
ranging from 100 to 600 nm. The increase in the pectin
concentration had a smoothing eect on the surface of the
microparticles, resulting in better homogeneity and no voids
(data not shown). Lastly, in the particular case of the formation
of a mushroom-type morphology (i.e., external gelation without
solvent diusion), the cap surface was made of tightly
interspersed more or less spherical domains, giving rise to an
undulating surface. The tail surface of the mushroom displayed
rather loosely interconnected anisotropic domains (with a starlike shape) with some interspersed voids. The surface topology
of the mushroom-like morphology microparticles reinforces the
hypothesis that the nonuniformity of pectin cross-linking
during the gelation process leads to the deformation of the
pregelled microdroplets at the CaCl2DMC interface. Moreover, it was established that after crossing the DMCCaCl2
liquidliquid interface during their collection the mushroomlike morphology microparticles did not migrate into the
calcium or DMC phase but stayed at the interface (data not
shown). The interfacial positioning of the microparticles was
also noticed after manually shaking them or after a
centrifugation step. This unexpected result calls into question
the relationship that exists between the surface structure and
the liquidliquid interface stabilizing property of these
microparticles.
Swelling Behavior of Hydrated and DriedRehydrated Pectin Hydrogel Microparticles. The swelling
behavior of both hydrated and driedrehydrated pectin
microparticles with a spherical, mushroom-like, or doughnutlike shape was followed during a 48 h incubation at 37 C in
three dierent buers (50 mM NaCl/HCl buer, pH 1.2; 500
mM MES buer, pH 5.0; and 500 mM Tris/HCl buer, pH
7.4) by phase-contrast optical microscopy and image analysis in
order to determine the diameter of the microparticles. It was
rst noticed that the hydrated microparticles incubated in the
three dierent buers displayed an increase in size, without
morphological changes, of at most 12% for spherical-like (until
144 h of incubation) and 7.5% for mushroom-like (total length
= cap + tail) microparticles regardless of the pH of the buer
and the incubation time. For doughnut-like shape micro1576

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

from the Biopolymers Interactions Structural Biology platform

(INRA, Nantes) for guidance with the confocal microscopy
observations.

biopolymer-based materials in drug carrier applications because

drying did not considerably aect their shapes and swelling
capacities except for the mushroom-like shape microparticles,
where their partial loss of shape resulting from buckling and
condensation phenomena after drying and the resulting limited
swelling capacity as a function of pH would need to be
improved, particularly in the drying process.

(1) Das, M.; Zhang, H.; Kumacheva, E. Annu. Rev. Mater. Res. 2006,
36, 117142.
(2) Zhang, H.; Tumarkin, E.; Sullan, R. M. A.; Walker, G. C.;
Kumacheva, E. Macromol. Rapid Commun. 2007, 28, 527538.
(3) Panda, P.; Ali, S.; Lo, E.; Chung, B. G.; Hatton, T. A.;
Khademhosseini, A.; Doyle, P. S. Lab Chip 2008, 8, 10561061.
(4) Fiddes, L. K.; Young, E. W. K.; Kumacheva, E.; Wheeler, A. R.
Lab Chip 2007, 7, 863867.
(5) Yin, Y. D.; Xia, Y. N. Adv. Mater. 2001, 13, 267271.
(6) Zheng, B.; Tice, J. D.; Ismagilov, R. F. Anal. Chem. 2004, 76,
49774982.
(7) Zhang, X. Z.; Lewis, P. J.; Chu, C. C. Biomaterials 2005, 26,
32993309.
(8) Dai, C. Y.; Wang, B. C.; Zhao, H. W. Colloids Surf., B 2005, 41,
117120.
(9) De Geest, B. G.; Urbanski, J. P.; Thorsen, T.; Demeester, J.; De
Smedt, S. C. Langmuir 2005, 21, 1027510279.
(10) Sugiura, S.; Oda, T.; Izumida, Y.; Aoyagi, Y.; Satake, M.; Ochiai,
A.; Ohkohchi, N.; Nakajima, M. Biomaterials 2005, 26, 33273331.
(11) Yamada, M.; Sugaya, S.; Naganuma, Y.; Seki, M. Soft Matter
2012, 8, 31223130.
(12) Jeong, W.; Kim, J.; Kim, S.; Lee, S.; Mensing, G.; Beebe, D. J.
Lab Chip 2004, 4, 576580.
(13) Berkland, C.; Kim, K. K.; Pack, D. W. J. Controlled Release 2001,
73, 5974.
(14) Carturan, G.; Dal Toso, R.; Boninsegna, S.; Dal Monte, R. J.
Mater. Chem. 2004, 14, 20872098.
(15) Champion, J. A.; Katare, Y. K.; Mitragotri, S. J. Controlled Release
2007, 121, 39.
(16) Zhao, L. B.; Pan, L.; Zhang, K.; Guo, S. S.; Liu, W.; Wang, Y.;
Chen, Y.; Zhao, X. Z.; Chan, H. L. W. Lab Chip 2009, 9, 29812986.
(17) Nisisako, T.; Hatsuzawa, T. Microfluid. Nanofluid. 2010, 9, 427
437.
(18) Lewis, C. L.; Lin, Y.; Yang, C.; Manocchi, A. K.; Yuet, K. P.;
Doyle, P. S.; Yi, H. Langmuir 2010, 26, 1343613441.
(19) Xu, S. Q.; Nie, Z. H.; Seo, M.; Lewis, P.; Kumacheva, E.; Stone,
H. A.; Garstecki, P.; Weibel, D. B.; Gitlin, I.; Whitesides, G. M. Angew.
Chem., Int. Ed. 2005, 44, 724728.
(20) Dendukuri, D.; Tsoi, K.; Hatton, T. A.; Doyle, P. S. Langmuir
2005, 21, 21132116.
(21) Seo, M.; Nie, Z. H.; Xu, S. Q.; Lewis, P. C.; Kumacheva, E.
Langmuir 2005, 21, 47734775.
(22) Liu, K.; Ding, H.-J.; Liu, J.; Chen, Y.; Zhao, X.-Z. Langmuir
2006, 22, 94539457.
(23) Yang, C.-H.; Huang, K.-S.; Wang, C.-Y.; Hsu, Y.-Y.; Chang, F.R.; Lin, Y.-S. Electrophoresis 2012, 33, 31733180.
(24) Dang, T. D.; Joo, S. W. Colloids Surf., B 2013, 102, 766771.
(25) Hu, Y.; Wang, Q.; Wang, J.; Zhu, J.; Wang, H.; Yang, Y.
Biomicrofluidics 2012, 6, 026502-1026502-9.
(26) Marquis, M.; Renard, D.; Cathala, B. Biomacromolecules 2012,
13, 11971203.
(27) Fang, A.; Gaillard, C.; Douliez, J.-P. Chem. Mater. 2011, 23,
46604662.
(28) Fang, A.; Gosse, C.; Gaillard, C.; Zhao, X.; Davy, J. Lab Chip
2012, 12, 49604963.
(29) Rondeau, E.; Cooper-White, J. J. Langmuir 2008, 24, 6937
6945.
(30) Williams, M. A. K.; Cucheval, A.; Nasseri, A. T.; Ralet, M. C.
Biomacromolecules 2010, 11, 16671675.
(31) Morris, G. A.; Ralet, M. C. Carbohydr. Polym. 2010, 88, 1488
1491.
(32) Sikorski, P.; Mo, F.; Skjak-Braek, G.; Stokke, B. T.
Biomacromolecules 2007, 8, 20982103.

CONCLUSIONS
We demonstrated the generation of pectin hydrogel microparticles of complex shapes either by combining the
phenomenon of gelation and water diusion-induced selfassembly in microuidic channels (on-chip) or by the
deformation of the pregelled droplets outside the channels
(o-chip) at a uiduid interface. We proved that by tuning
the mode of pectin cross-linking (CaCl2 vs CaCO3) and the
degree of shrinking (water content in the organic continuous
phase) we can control the shape of the nal particles. Sphere,
doughnut-like, oblate ellipsoid, or mushroom-like morphologies
were thus produced, demonstrating the ability to control the
formation of anisotropic biopolymer-based hydrogel microparticles using microuidics. The swelling capacities of the
driedrehydrated microparticles of dierent shapes (sphere,
doughnut-like, and mushroom-like) as a function of pH will
open opportunities for their use in applications in areas where
dried materials are preferred for economic reasons. Our
approach can also be extended to hydrogel microparticles
formation from other anionic biopolymers, providing new
opportunities to produce biomimetic microgels with various
anisotropic dimensions for applications in drug delivery, optical
devices, and advanced materials formation. For instance, by
incorporating functional materials or active components inside
the biopolymer network, functional doughnut-like or ellipsoidal
hydrogel microparticles can be tailor-made for specic
applications such as magnetic actuation or as a drug carrier.
These promising and original biopolymer-based hydrogel
microparticles could also be good candidates for food
applications where the modulation of avors as a function of
time or temperature would be mandatory.

ASSOCIATED CONTENT

S Supporting Information
*

Microchannel dimensions of the microuidic device; optical

and scanning electron microscopy images of pectin microparticles obtained by solvent diusion; optical and confocal
microscopy images of pectin hydrogel microparticles obtained
by solvent diusion; and optical, scanning electron, and
confocal microscopy images of pectin hydrogel microparticles
obtained without solvent diusion. This material is available
free of charge via the Internet at http://pubs.acs.org.

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*Tel.: 33 2 40 67 50 52; Fax: 33 2 40 67 50 43; E-mail: denis.

renard@nantes.inra.fr.
Notes

The authors declare no competing nancial interest.

ACKNOWLEDGMENTS
We acknowledge N. Stephan for use of MEB facilities at IMN
(Nantes, France) and for expertise. We also acknowledge I.
Capron and H. Bizot for fruitful discussions and B. Bouchet
1577

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578

Biomacromolecules

Article

(33) Ogonczyk, D.; Siek, M.; Garstecki, P. Biomicrofluidics 2011, 5,

013405-1013405-12.
(34) Xia, Y. N.; Whitesides, G. M. Annu. Rev. Mater. Sci. 1998, 28,
153184.
(35) Ogushi, Y.; Sakai, S.; Kawakami, K. J. Biosci. Bioeng. 2007, 104,
3033.
(36) Robitzer, M.; David, L.; Rochas, C.; Di Renzo, F.; Quignard, F.
Macromol. Symp. 2008, 273, 8084.
(37) Stephenson, R.; Stuart, J. J. Chem. Eng. Data 1986, 31, 5670.
(38) Kar, B. P.; Ramanathan, N.; Sundararajan, K.; Viswanathan, K. S.
J. Mol. Struct. 2012, 1024, 8493.
(39) Goodall, D. M.; Norton, I. T. Acc. Chem. Res. 1987, 20, 5965.
(40) Strom, A.; Williams, M. A. K. J. Phys. Chem. B 2003, 107,
1099510999.
(41) Mele, E.; Fragouli, D.; Ruffilli, R.; De Gregorio, G. L.; Cingolani,
R.; Athanassiou, A. Soft Matter 2013, 9, 63386343.

1578

dx.doi.org/10.1021/bm401596m | Biomacromolecules 2014, 15, 15681578