Chemical Engineering Science 235 (2021) 116468

Contents lists available at ScienceDirect

Chemical Engineering Science

journal homepage: www.elsevier.com/locate/ces

Insight into drug encapsulation in polymeric nanoparticles using

microfluidic nanoprecipitation
Wei Li, Qiaoli Chen, Thejus Baby, Song Jin, Yun Liu, Guangze Yang, Chun-Xia Zhao ⇑
Australian Institute for Bioengineering and Nanotechnology, The University of Queensland, St. Lucia, QLD 4072, Australia

h i g h l i g h t s

 A combined computational fluid dynamics (CFD) and experimental approach to illustrate microfluidic nanoprecipitation.
 Mixing times of a polymer and a drug are determined using a CFD method.
 Drug loading is dependent on the mixing time of the polymer and drug.
 The precipitation time of polymer and drug should be matched for drug encapsulation.

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

Article history: Synthesis of polymeric nanoparticles (NPs) through self-assembly of di-block copolymers have attracted
Received 7 November 2020 substantial interest in the past decades for drug delivery and controlled release. Microfluidics offers a
Received in revised form 5 January 2021 facile approach for making such NPs and drug encapsulation. However, a fundamental understanding
Accepted 17 January 2021
of the drug encapsulation process is lacking. In this paper, we report a combined computational fluid
Available online 28 January 2021
dynamics (CFD) and experimental approach to illustrate the fundamental principle that governs the
encapsulation of a drug in polymeric NPs through microfluidic nanoprecipitation. Taking a drug curcumin
Keywords:
and a polymer poly (ethylene glycol)-block-poly (D, L-lactide-co-glycolide) (PEG-PLGA) as a model sys-
Microfluidic
Self-assembly
tem, we demonstrated the different precipitation times of curcumin and PEG-PLGA as well as their mix-
Nanoparticles ing times in the microfluidic device. The big difference in their mixing times led to very low drug loading.
mPEG-PLGA This study provides a new perspective in understanding and controlling the formation of drug-loaded
Polymer nanoparticles polymeric NPs and offers a new design rule for selecting the right combinations of drugs, polymers, sol-
Nanoprecipitation vents, and devices.
Drug encapsulation Ó 2021 Elsevier Ltd. All rights reserved.

1. Introduction agents owing to its unique properties like versatile degradation

kinetics, nontoxicity, biocompatibility, and FDA approval for
Polymeric NPs are colloidal particulate systems with particle human use. However, upon injection into the blood PLGA NPs are
sizes ranging from 10 to 1000 nm (Soppimath et al., 2001), which rapidly recognized and cleared by the mononuclear phagocyte sys-
are promising drug delivery carriers owing to their versatile prop- tem (MPS), due to their adsorption of blood proteins (opsonins)
erties. Polymeric NPs based on biodegradable poly (D, L-lactic-co- onto their surface (Stella et al., 2000). This leads to the fast NP
glycolic acid) (PLGA) have been extensively explored for various accumulation in the MPS organs, such as liver and spleen (Gref
applications such as controlled release and drug delivery et al., 1995; Stolnik et al., 1995; Storm et al., 1995). Consequently,
(Danhier et al., 2010; Jain, 2000; Liu et al., 2020a; Liu et al., the surface of PLGA has to be modified using hydrophilic, flexible,
2020b; Lu et al., 2009). PLGA is an aliphatic polymer with a polye- polymeric linkers, such as poly (ethylene glycol) (PEG), poloxamer,
ster backbone that is formed through the copolymerization of lac- polyethylene oxide, polysorbate, lauryl ethers, etc. to achieve long-
tic acid and glycolic acid monomers (Bala et al., 2004; Hines and circulating time (Gref et al., 1995; Kim et al., 2005; Storm et al.,
Kaplan, 2013; Makadia and Siegel, 2011). It is considered as a ‘‘gold 1995).
standard” for intravenous administration of therapeutic anticancer Traditional methods for the preparation of PEG-PLGA based
nanocarriers include single-emulsion-solvent evaporation,
double-emulsion-solvent evaporation, salting-out, emulsification-
⇑ Corresponding author. diffusion, solvent displacement/nanoprecipitation, and emulsion-
E-mail address: [email protected] (C.-X. Zhao).

https://doi.org/10.1016/j.ces.2021.116468
0009-2509/Ó 2021 Elsevier Ltd. All rights reserved.
W. Li, Q. Chen, T. Baby et al. Chemical Engineering Science 235 (2021) 116468

diffusion-evaporation (Bala et al., 2004; Hines and Kaplan, 2013; 2. Materials and methods
Liu et al., 2020c; Liu et al., 2020d). However, these approaches
are often based on traditional ‘‘beaker methods’’ which lack precise 2.1. Chemicals and reagents
control over mixing, nucleation, and growth processes, thus
adversely affecting the final NP properties, such as particle size, All chemicals were of analytical grade and used as received
size distribution, particle structure, etc. In contrast, microfluidic without further purification. Methoxy poly(ethylene glycol)-b-pol
technology, which uses reactors of micrometer length scale for y(lactide-co-glycolide) MW ~ 5,000:55,000 Da (mPEG5k-PLGA55k)
the synthesis of NPs (Ran et al., 2017; Zhao et al., 2011a; Zhao, was purchased from PolySciTech (Akina, West Lafayette, IN, USA).
2013), offers enormous advantages over conventional technolo- Other chemicals were purchased from either Sigma-Aldrich or
gies, including, (1) small volume of reagents allowing efficient Merck. Water was obtained from a MilliQ system (Millipore, North
use of expensive and toxic chemicals, (2) large surface area to vol- Ryde, Australia) equipped with a 0.22 lm filter and had resistivity
ume ratio thus offering enhanced heat and mass transfer, (3) rapid larger than18.2 MX
cm.
and efficient mixing, (4) homogeneous reaction environments, and
(5) possibility of automatic multi-step processing, such as by com-
2.2. Microfluidic device
bining analysis, reactions and purification in a single microchip.
Drug loaded PEG-PLGA NPs have been successfully developed Experiments were performed in a flow-focusing microfluidic
using the microfluidic hydrodynamic flow-focusing method, by
device, which was fabricated from poly(dimethyl-siloxane) (PDMS)
rapid mixing (mixing time: 0.04–0.4 ms) (Karnik et al., 2008). using standard SU-8 photolithography and soft-lithography proce-
This microfluidic approach allows the facile preparation of
dures, and bonded to PDMS-coated glass coverslips to provide uni-
libraries of various NP formulations with tunable properties form surface properties on all the walls of the microchannel (Zhao
(Baby et al., 2017; Liu et al., 2018; Ran et al., 2016; Ran et al.,
et al., 2011b). Fig. 1 shows the design of our flow-focusing
2017; Ran et al., 2018). Several parameters such as flow rate, microfluidic device. It has two inlets and one outlet. The acetoni-
polymer composition, and polymer concentration affect the prop-
trile (ACN) solution containing the dissolved polymer was intro-
erties of the synthesized NPs, including their particle size, poly-
duced from the central channel (red, Fig. 1c), and water was
dispersity, and drug release (Karnik et al., 2008). Despite the
pumped in through the two side perpendicular channels (blue,
great potentials of this microfluidic approach, the precise control
Fig. 1c). The filter structure (shown in Fig. 1b,d) in the two inlets
of drug encapsulation in PEG-PLGA NPs formation using microflu-
(shown in Fig. 1d) was designed to prevent any debris or particles
idics remains largely unexplored. Therefore, it is of great impor-
from entering the microchannel which can easily cause the block-
tance to fundamentally understand the key factors that affect
age of the channel. All the microchannels have a width of 20 lm
drug encapsulation in polymeric NPs using microfluidic
and a depth of 60 lm, and the central mixing channel is 1 cm long.
nanoprecipitation.
Syringes (Terumo, Binan Laguna, Philippines) were used for water
One of the key parameters controlling the formation of NP is
and polymer ACN solutions. Flow rates of the water and polymer
the mixing time of the nanoprecipitation process (Baby et al.,
ACN solutions were controlled using syringe pumps (Harvard
2017; Karnik et al., 2008). It is generally believed that a longer
Pump 11 Plus, Harvard Apparatus, MA, USA). PTFE tubing (Cole-
mixing time correlates with a larger NP size. But due to technical
Parmer, IL, USA) was used to connect the syringe and the microflu-
difficulties in direct experimental measurement, only empirical
idic device. Microscope images of fluid flow-focusing in the
formula that is based on an ideal 2-D Poiseuille flow, was pro-
microchannel are shown in Fig. 1d,e. The inlets and fluid flow-
posed to estimate an approximate mixing time (Karnik et al.,
focusing (Fig. 1d,e) were captured by a video camera (Canon,
2008). It is also assumed that the interaction between water
Tokyo, Japan) mounted on top of an inverted optical microscope
streams and solution streams is so weak that the flow profile is
(Nikon, Tokyo, Japan).
not significantly changed. A two-dimensional analytical model
was also developed for convective and diffusive mixing in a
two-phase hydraulic focusing microchannel (Wu and Nguyen, 2.3. Synthesis of mPEG-PLGA NPs
2005). In fact, the actual flow is 3-D and the solution-water inter-
action is strong in the flow-focusing microfluidic device. Further- The NPs were synthesized in the microfluidic device using a
more, geometrical factors also affect mixing times, such as device nanoprecipitation method. (Karnik et al., 2008) Briefly, the ACN
design, microchannel width and depth, etc. In this context, com- solution (containing dissolved mPEG-PLGA or mPEG-PLGA/drug)
putational fluid dynamics (CFD) provides a useful tool to numer- at a certain flow rate (0.3, 0.5, 1, 2, 3 lL/min) was introduced in
ically study the mixing process, NP formation, and drug the central channel, while water at a flow rate of 10 lL/min was
encapsulation. pumped through the two sides channels by Harvard Pump. Then
To analyze the hydrodynamic flow focusing mechanism under- the ACN solution stream was focused into a narrow stream by the
lying NPs formation, CFD has been used to obtain the full temporal- two adjacent water streams at the crossing section. Thus, rapid
spatial distribution of fluid–structure and NPs concentration mixing occurred through molecular diffusion, and mPEG-PLGA
(Andreas Jahn et al., 2013; Jahn et al., 2010; Mamet et al., 2017). self-assembled into NPs. The NPs collected from the outlet were
Then optimized parameters obtained from above CFD-modeling washed three times by centrifugation (4000 g, 15 mins, room tem-
can be used to assist the design of microfluidic approaches. To perature) using Amicon ultra centrifugal filter devices with 100 kDa
our knowledge, CFD-assist microfluidic modeling has neither been NMWL (Millipore, Billerica, MA, USA). Three parameters were
used to estimate the mixing times, nor used to study drug encap- investigated to control the size of mPEG-PLGA NPs, including five
sulation. In this study, we firstly develop and validate our CFD different flow ratios (0.03, 0.05, 0.1, 0.2, 0.3) (flow ratio is the ratio
modeling, then use it to calculate the mixing time, and study of the flow rate of the ACN solution with polymer dissolved to that
numerically the effects of solution/water flow rate ratios and of water), three concentrations (10 mg/mL, 30 mg/mL, 50 mg/mL).
microfluidic device geometry on mixing time. We next investigate
experimentally the effect of mixing time on the NPs size distribu- 2.4. Synthesis of drug-loaded mPEG-PLGA NPs
tion to further verify the numerical results. Finally, we use a model
hydrophobic drug curcumin to explore drug encapsulation at dif- The drug-loaded mPEG-PLGA NPs were synthesized in a similar
ferent initial drug concentrations. way to the mPEG-PLGA NPs. The model drug curcumin was
2
W. Li, Q. Chen, T. Baby et al. Chemical Engineering Science 235 (2021) 116468

Fig. 1. Flow-focusing microfluidic device. (a) Schematic diagram of the microfluidic device (red represents a stream of acetonitrile solution with polymer dissolved and blue
represents water. In the mixing channel and outlet, the blue color is more obvious since the flow rate of water is much higher than that of acetonitrile), (b) Schematic diagram
of filters, (c) Schematic diagram of the flow-focusing section, (d) Inlets and filters under a microscope, (e) Flow-focusing channels (polymer in acetonitrile solution/water flow
ratio equals to 0.2) representing the mixing of the acetonitrile and water phases (channel width 20 lm). (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article.)

dissolved in ACN along with the mPEG-PLGA and then mixed with to a UV-detector (Shimadzu, Kyoto, Japan). The HPLC was equipped
water in the microfluidic device to obtain the drug-loaded mPEG- with a Jupiter C18 column (5 lm, 300 Å, 150 mm  4.6 mm) (Phe-
PLGA NPs. nomenex, Torrance, CA, USA). The mobile phase buffer A was 0.1%
v/v trifluoroacetic acid (TFA) in water, and the mobile phase B was
2.5. Characterization of NPs 90% v/v ACN, 0.1% v/v TFA in water. An applied gradient from 30 to
60% buffer B in 30 mins followed by 60% buffer B for 10 mins and
Dynamic Light Scattering (DLS). Size distribution and zeta poten- 100% buffer B for another 10 mins at a flow rate of 1 mL/min was
tial of the mPEG-PLGA NPs were measured by DLS using a Malvern used. Under these conditions, the mPEG-PLGA did not interfere
Zetasizer Nano ZS (Malvern Instrument, UK) at a scattering angle of with the drug. The injection volume was 100 lL. The wavelength
173° and a temperature of 25 °C. Samples were diluted by a factor used for curcumin was 425 nm. The concentration of curcumin
of 50 for measurement to avoid multiple scattering effects. was determined based on the peak area at a retention time of
Transmission Electron Microscopy (TEM). Morphology and size of approximately 24.9 min according to the standard curve. The drug
NPs were observed by TEM using a JEOL 1011 microscope (JEOL loading is defined as the mass fraction of drug in the NPs, while
Ltd., Tokyo, Japan) operated at 100 kV. Twenty microliters of the encapsulation efficiency is the mass fraction of encapsulated drug
sample were applied onto 200-mesh carbon-coated grids (Prosci- to the initial drug used for the preparation of NPs. Each experiment
tech, Brisbane, Australia). The grids were then dried in air, and was carried out in triplicate.
stained with 1% (w/v) uranyl acetate for 2 mins. The NPs size
was analyzed using iTEM software (version 3.2, Soft Imaging Sys- 3. Results and discussion
tem GmbH).
Confocal Microscope. Zeiss Confocal Microscope 710 was used to 3.1. Numerical simulation of the microfluidic flow-focusing flow
observe the mixing in the microchannel using DiI.
The Ansys CFX 18.0 software was adapted to conduct the
2.6. Drug loading and encapsulation efficiency numerical simulation of ACN concentration distribution in the
flow-focusing microfluidic device, in which the ACN distribution
Lyophilized NPs were accurately weighed and dissolved in ACN. in the mixing process was governed by 3-D Navier-Stokes govern-
Then the concentration of drug in the solution was measured by ing equations coupled with species transport equation. The finite
HPLC using a Shimadzu system (Shimadzu, Kyoto, Japan) linked volume method was used to solve the governing equations and
3
W. Li, Q. Chen, T. Baby et al. Chemical Engineering Science 235 (2021) 116468

the High Resolution was adapted in Advection Scheme. The PBS concentration of ACN decreases from 100% to 40%, curcumin
buffer solution was chosen with a density of 997.0 kg/m3 and remains soluble in the ACN-water mixture until it reduces to
dynamic viscosity of 8.899  10 4 kg/m
s and ACN was introduced 37%. So, the critical ACN concentration for curcumin is 37%. And
in the central microchannel with the density of 783.0 kg/m3 and the precipitation is instant, when the concentration is below 37%,
dynamic viscosity of 0.334 kg/m
s. All the flows were low-speed, nearly 100% curcumin precipitates out. In contrast, the mPEG-
so the laminar model was selected. The mass flow rate boundary PLGA polymer shows a quite different precipitation curve. At
in the three inlets and static pressure in outlet were specified. 100% ACN, mPEG-PLGA is soluble, but its precipitation starts from
The mesh independence had been explored and the result dis- the beginning and increases with the amount of water added. At
played that the mesh number about 400,000 in the baseline case 70% ACN and 30% water in volume, 100% mPEG-PLGA precipitates
was reasonable to conduct the simulation. out. So the critical concentration of mPEG-PLGA is 70% (Fig. 3(b))
(Baby et al., 2017). The results indicate that the mPEG-PLGA is less
3.2. Validation of numerical modeling with experimental data soluble than curcumin, so curcumin precipitates out before mPEG-
PLGA.
Fig. 2 displays the comparison of the ACN concentration distri- Fig. 3(c) shows the simulation result of ACN distribution con-
bution in the middle section of microchannel between numerical tour in the middle section of the device vertical to the device depth
simulation and experimental results in five different ACN/water direction. It can be clearly seen that ACN spreads from the central
flow rate ratios (R) of 0.03, 0.05, 0.1, 0.2, and 0.3. It can be clearly part of the mixing channel. Fig. 3(d) shows the quantitative results
seen that the vertical water streams squeeze ACN and make it form of ACN concentration at the central line of the channel as well as
a narrow stream. When ACN flows downstream along the the maximum ACN concentration in the mixing channel and its
microchannel, it gradually mixes with water, which reduces its concentration on the center line in microchannel as a function of
maximum concentration. The increase of the flow rate ratio leads the channel length. The results shown in Fig. 3(d) demonstrate that
to the increase of the width of the focused stream in the central these two concentrations are nearly identical, confirming that the
channel, which agrees well with the experimental data (Fig. 2b). ACN concentration at the center line is the maximum concentra-
In summary, the simulation results of the ACN concentration dis- tion in the channel. The mixing distance for mPEG-PLGA and cur-
tribution in the mixing microchannel agrees well with experimen- cumin can be determined based on the ACN concentration curve
tal data, which validates our numerical modeling. on the central line in Fig. 3(b). Then the mixing time is calculated
based on the ratio of the mixing distance to its velocity. We can see
3.3. Mixing time that the concentration at the central line decreases exponentially
with the channel length. Again, the polymer mPEG-PLGA precipi-
The mixing time in this study is defined as the time interval tates out at the channel length of about 0.2 mm, but curcumin pre-
between the starting point and the critical point where 100% nano- cipitates much later at the channel length of around 1 mm.
precipitation is achieved in the mixing microchannel. Firstly, the
maximum ACN concentration distribution is calculated in the sec-
tions vertical to flow direction along the mixing microchannel. Sec- 3.4. Numerical study on the effects of microchannel geometry and flow
ondly, the critical ACN concentration to achieve 100% precipitation rate ratio on mixing time
is determined experimentally. It is assumed that when the maxi-
mum ACN concentration in the channel center is below the critical Based on the calculation method described above, the mixing
ACN concentration, the nanoprecipitation process is completed times of mPEG-PLGA and curcumin and their time interval (the
within the microchannel. Then based on the critical ACN concen- time difference between these two mixing times) can be calcu-
tration and the maximum ACN concentration on the central line lated. It is known that both the microchannel geometry and ACN/
of the microchannel, we could determine the mixing distance that water flow rate ratio affect the mixing time. Fig. 4(a) displays the
is required to achieve 100% precipitation of either polymers and mixing times and their time interval for five different flow rate
drugs. Finally, the mixing time is calculated based on the mixing ratios in a microfluidic flow-focusing device with a microchannel
distance and the flow velocity in the microchannel. height and width of 60 mm and 20 mm, respectively. It is clear that
Fig. 3 shows the details about the calculation of the mixing the mixing time of mPEG-PLGA doesn’t vary much with the flow
time. The microfluidic device with a channel width of 20 mm and rate ratio, probably due to its fast precipitation. In contrast, the
depth of 60 mm was adapted to conduct the numerical investiga- flow rate ratio has a significant effect on the mixing time of cur-
tion for a flow rate ratio of ACN to water at 0.1. Fig. 3(a) shows cumin, as a higher flow rate ratio means a wide central stream in
the precipitation curve of curcumin. At 100% ACN, curcumin is the microchannel, consequently a slower mixing. It is obvious that
100% soluble, in other words, 0% precipitation. When the volume the bigger the time interval is, the more difficult for drug encapsu-

Fig. 2. Comparison between numerical simulation and experimental measurement at five different flow ratios of 0.03, 0.05, 0.1, 0.2, 0.3 with the microchannel width of 20 mm
and depth of 60 mm. (a) Numerical simulation; (b) Experimental measurement.

4
W. Li, Q. Chen, T. Baby et al. Chemical Engineering Science 235 (2021) 116468

Fig. 3. Determination of mixing time (width: 20 mm; depth: 60 mm; flow rate ratio of ACN to water: 0.1). (a) ACN concentration distribution with a flow rate ratio of 0.1; (a)
Curcumin precipitation curve with ACN concentration; (b) Polymer precipitation curve with ACN concentration; (c) Simulation result to show ACN concentration distribution
in the middle section of channel; (d) Simulation result of the ACN concentration along the center line and its maximum value in sections vertical to the x-direction.

Fig. 4. Mixing time and its time interval distribution. (a) Flow rate ratio (width: 20 mm; depth: 60 mm); (b) Microchannel width (Flow rate ratio: 0.1; depth: 60 mm).

lation. Therefore, a smaller flow rate ratio is beneficial for forming microchannel height used was 60 mm and the flow rate ratio was
drug-encapsulated NPs. 0.1. Similar to the effect of flow ratio, the increase of channel
Fig. 4(b) shows the effect of the microchannel width (20, 40, 60, widths leads to a quick increase in the mixing times and time
80, and 100 mm) on mixing times and their time interval. The intervals. But the difference is that the impact of channel width

5
W. Li, Q. Chen, T. Baby et al. Chemical Engineering Science 235 (2021) 116468

Fig. 5. TEM images of mPEG-PLGA NPs at flow ratios of (a) 0.03, (b) 0.05, (c) 0.1, (d) 0.2, (e) 0.3, and (f) DLS results of NPs formed using 30 mg/mL mPEG5k-PLGA55k. (Flow ratio
is defined as the ACN/water flow rate ratio, and the flow rate of water was fixed at 10 lL/min).

Table 1
TEM diameter, DLS number-mean diameter and PDI values for NPs synthesized using mPEG5k-PLGA55k at a concentration of 30 mg/mL.

Flow ratio 0.03 0.05 0.1 0.2 0.3

TEM diameter (nm) 47 ± 2.5 48 ± 2.0 54 ± 1.9 62 ± 2.5 74 ± 2.2
DLS number-mean diameter (nm) 49 ± 2.9 52 ± 1.7 56 ± 2.6 59 ± 1.2 71 ± 1.2
PDI 0.131 0.097 0.080 0.069 0.092

(Table 1). As TEM can only give an average number of those limited
number of selected particles in the image, NPs size was also
analyzed using DLS, and Fig. 5(f) represents the resulting size dis-
tributions. We can see that the distribution peaks are quite sharp
for all the NPs obtained at different flow ratios indicating a narrow
size distribution. Table 1 gives detailed information about the DLS
number-mean diameter and PDI. Consistent with our observation
of the size distribution shown in Fig. 5(f), all the NP PDI values
are less than 0.2, which means that very uniform mPEG-PLGA
NPs were synthesized, as also observed from the TEM images. Fur-
thermore, the comparison between these two size measurement
methods (DLS and TEM) demonstrated that the DLS number-
mean diameter is consistent with that obtained from TEM images.
As the DLS measurement is quick and convenient, it was used in
the subsequent studies for monitoring NP size.
From both TEM images and DLS results, we can conclude that
Fig. 6. Drug loading and encapsulation efficiency of curcumin-loaded mPEG-PLGA the NP size increases with the ACN/Water flow rate ratio, which
NPs obtained using different initial drug concentrations.
is consistent with our simulation results that the mixing time
increases with flow rate ratio.
on the mixing time of mPEG-PLGA is more significant, as bigger
channel width results in wider central streams of ACN, thus slower 3.6. Drug loading and encapsulation efficiency
mixing (Karnik et al., 2008).
For drug delivery applications, we chose to use mPEG-PLGA of
3.5. Experimental study on the effect of ACN/water flow rate ratios on 30 mg/mL at the flow ratio of 0.2 (at this operation condition,
NP size mPEG-PLGA NPs size is around 60 nm) to make the drug-loaded
mPEG-PLGA NPs at a relatively higher yield (Fig. 6). Curcumin
Based on the simulation results, the effect of ACN/water flow was used as a model drug for drug-loading experiments.
rate ratio was further examined experimentally using five different Curcumin is very hydrophobic. For curcumin encapsulation, an
flow rate ratios of 0.03, 0.05, 0.1, 0.2, and 0.3. The concentration of initial 3.33% theoretical drug concentration was explored by using
mPEG-PLGA used was 30 mg/mL. Consistent with the simulation 29 mg/mL mPEG-PLGA + 1 mg/mL curcumin. Curcumin loaded
results, bigger mixing time leads to bigger particle size. polymeric NPs were formed with an actual drug loading of 2.6%
Fig. 5(a-e) shows the TEM images of NPs revealing the increase and encapsulation efficiency of 77.3% (Fig. 6). This is expected as
of particle sizes of the spherical mPEG-PLGA NPs with uniform size mPEG-PLGA is precipitated much quicker and earlier than cur-
6
W. Li, Q. Chen, T. Baby et al. Chemical Engineering Science 235 (2021) 116468

Fig. 7. Microfluidic nanoprecipitation of mPEG-PLGA and curcumin.

cumin, so only limited curcumin can be encapsulated. Further ratio and microchannel geometry on the nanoprecipitation pro-
increasing the concentration of curcumin to 1.5 mg/mL cess. These results demonstrate that a good match between the
curcumin + 28.5 mg/mL mPEG-PLGA (5% initial drug concentra- mixing times of the drug and NP material is critical for formulating
tion) only led to a slight increase of drug loading to 3.4% but a drug-loaded polymeric NPs for drug delivery. The mixing time
decreased encapsulation efficiency of 66.0% (Fig. 6). Furthermore, defined in this work is different from previous studies. It offers a
any further increase of the initial drug concentration above 5% quantitative tool to evaluate drug encapsulation of polymer
resulted in the formation of curcumin crystals thus channel block- nanoparticles using microfluidic nanoprecipitation. Furthermore,
age. This is expected as a large amount of mPEG-PLGA precipitates this study provides insights into the synthesis of drug-loaded
earlier and quicker, so when curcumin precipitates and forms cur- polymeric NPs, which is critical for the future design of polymeric
cumin NPs, there is no enough mPEG-PLGA to stabilize them, thus NPs for drug delivery applications.
leading to quick aggregation of curcumin. Consequently, the high-
est drug loading of curcumin we can acheieve in this study was
Author contributions
3.4%, which is similar to those obtained by other groups,
(Valencia et al., 2013) but this drug loading is considered very low.
The manuscript was written through contributions of all
Based on the numerical and experimental studies, the microflu-
authors. All authors have given approval to the final version of
idic nanoprecipitation of polymer mPEG-PLGA and curcumin
the manuscript.
occurs at two different stages (Fig. 7). In the first stage, the major-
ity of the polymer precipitates and forms polymeric NPs followed
by the precipitation of curcumin NPs. Some of them are stabilized CRediT authorship contribution statement
by the mPEG-PLGA thus forming drug-loaded polymeric NPs.
When the concentration of curcumin is increased, no enough Wei Li: Methodology, Writing - original draft, Writing - review
mPEG-PLGA presents in the solution to stabilize the drug NPs, & editing, Software, Validation, Formal analysis, Investigation.
resulting in the aggregation of curcumin NPs forming big aggre- Qiaoli Chen: Methodology, Writing - original draft, Writing -
gates thus blocking the microchannel. To increase the drug loading review & editing, Formal analysis, Investigation. Thejus Baby:
and forming stable drug-loaded polymeric NPs, it would be ideal to Investigation. Song Jin: Investigation. Yun Liu: Investigation, For-
precipitate drug molecules first, so polymers in the solution are mal analysis. Guangze Yang: Investigation, Formal analysis.
enough to cover the drug NPs and render them stable in the solu- Chun-Xia Zhao: Conceptualization, Methodology, Writing - origi-
tion. Therefore, the precipitation time of polymer and drug should nal draft, Writing - review & editing, Supervision, Project adminis-
be matched by using two different approaches. One is to screen a tration, Funding acquisition.
library of polymers with systematic varied properties, so the best
polymer can be identified with a precipitation time close to the Declaration of Competing Interest
drug to be encapsulated. Alternatively, a different strategy can be
adopted by screening different solvent systems. It is very difficult The authors declare that they have no known competing finan-
to tune the precipitation time of polymers or drugs using a cial interests or personal relationships that could have appeared
single-solvent system, but multiple-solvent systems offer a facile to influence the work reported in this paper.
method. For example, a multiple-solvent system could consist of
two or three solvents with distinct properties. One solvent has high
solubility of the polymer or drug, while the other solvent has very Acknowledgements
low solubility. So by tuning the ratio of these two solvents, the pre-
cipitation times of the drug and drug could be matched, thus The project was supported by the Australian Research Council
achieving successful drug encapsulation. projects (FT140100726 and DP200101238). This work was per-
formed in part at the Queensland node of the Australian National
Fabrication Facility. A company established under the National Col-
4. Conclusions laborative Research Infrastructure Strategy to provide nano and
microfabrication facilities for Australia’s researchers. The authors
In summary, the synthesis of drug-loaded mPEG-PLGA NPs was acknowledge the facilities, and the scientific and technical assis-
investigated using a flow-focusing microfluidic device numerically tance, of the Australian Microscopy & Microanalysis Research Facil-
and experimentally. Firstly, a method to calculate mixing times ity at the Centre for Microscopy and Microanalysis, The University
was developed to investigate the effects of solution/water flow rate of Queensland.
7
W. Li, Q. Chen, T. Baby et al. Chemical Engineering Science 235 (2021) 116468

References Liu, Y., Yang, G.Z., Zou, D., Hui, Y., Nigam, K., Middelberg, A.P.J., Zhao, C.X., 2020d.
Formulation of nanoparticles using mixing-induced nanoprecipitation for drug
delivery. Ind. Eng. Chem. Res. 59, 4134–4149.
Andreas Jahn, F.L., Wepf, R.A., Dittrich, P.S., 2013. Freezing continuous-flow self-
Lu, J.M., Wang, X., Marin-Muller, C., Wang, H., Lin, P.H., Yao, Q., Chen, C., 2009.
assembly in a microfluidic device: Toward imaging of liposome formation.
Current advances in research and clinical applications of PLGA-based
Langmuir 29, 1717–1723.
nanotechnology. Expert. Rev. Mol. Diagn. 9, 325–341.
Baby, T., Liu, Y., Middelberg, A.P.J., Zhao, C.-X., 2017. Fundamental studies on
Makadia, H.K., Siegel, S.J., 2011. Poly Lactic-co-Glycolic Acid (PLGA) as
throughput capacities of hydrodynamic flow-focusing microfluidics for
Biodegradable Controlled Drug Delivery Carrier. Polymers 3, 1377–1397.
producing monodisperse polymer nanoparticles. Chem. Eng. Sci. 169, 128–139.
Mamet, V., Namy, P., Dedulle, J.-M., 2017. Numerical modeling of flow focusing:
Bala, I., Hariharan, S., Kumar, M.N., 2004. PLGA nanoparticles in drug delivery: the
Quantitative characterization of the flow regimes. Phys. Fluids 29.
state of the art. Crit. Rev. Ther. Drug Carrier Syst. 21, 387–422.
Ran, R., Middelberg, A.P.J., Zhao, C.X., 2016. Microfluidic synthesis of multifunctional
Danhier, F., Feron, O., Preat, V., 2010. To exploit the tumor microenvironment:
liposomes for tumour targeting. Colloid Surface B. 148, 402–410.
Passive and active tumor targeting of nanocarriers for anti-cancer drug delivery.
Ran, R., Sun, Q., Baby, T., Wibowo, D., Middelberg, A.P.J., Zhao, C.X., 2017. Multiphase
J. Control. Release 148, 135–146.
microfluidic synthesis of micro- and nanostructures for pharmaceutical
Gref, R., Domb, A., Quellec, P., Blunk, T., Muller, R.H., Verbavatz, J.M., Langer, R.,
applications. Chem. Eng. Sci. 169, 78–96.
1995. The controlled intravenous delivery of drugs using PEG-coated sterically
Ran, R., Wang, H.F., Liu, Y., Hui, Y., Sun, Q., Seth, A., Wibowo, D., Chen, D., Zhao, C.X.,
stabilized nanospheres. Adv. Drug Deliv. Rev. 16, 215–233.
2018. Microfluidic self-assembly of a combinatorial library of single- and dual-
Hines, D.J., Kaplan, D.L., 2013. Poly(lactic-co-glycolic) acid-controlled-release
ligand liposomes for in vitro and in vivo tumor targeting. Eur. J. Pharm.
systems: experimental and modeling insights. Crit. Rev. Ther. Drug Carrier
Biopharm. 130, 1–10.
Syst. 30, 257–276.
Soppimath, K.S., Aminabhavi, T.M., Kulkarni, A.R., Rudzinski, W.E., 2001.
Jahn, A., Stavis, S.M., Hong, J.S., Vreeland, W.N., DeVoe, D.L., Gaitan, M., 2010.
Biodegradable polymeric nanoparticles as drug delivery devices. J. Control.
Microfluidic mixing and the formation of nanoscale lipid vesicles. ACS Nano 4,
Release 70, 1–20.
2077–2087.
Stella, B., Arpicco, S., Peracchia, M.T., Desmaele, D., Hoebeke, J., Renoir, M., D’Angelo,
Jain, R.A., 2000. The manufacturing techniques of various drug loaded
J., Cattel, L., Couvreur, P., 2000. Design of folic acid-conjugated nanoparticles for
biodegradable poly(lactide-co-glycolide) (PLGA) devices. Biomaterials 21,
drug targeting. J. Pharm. Sci. 89, 1452–1464.
2475–2490.
Stolnik, S., Illum, L., Davis, S., 1995. Long Circulating Microparticulate Drug Carriers.
Karnik, R., Gu, F., Basto, P., Cannizzaro, C., Dean, L., Kyei-Manu, W., Langer, R.,
Adv. Drug Delivery Rev. 16, 195–214.
Farokhzad, O.C., 2008. Microfluidic platform for controlled synthesis of
Storm, G., Belliot, S.O., Daemen, T., Lasic, D.D., 1995. Surface Modification of
polymeric nanoparticles. Nano Lett. 8, 2906–2912.
Nanoparticles to Oppose Uptake by the Mononuclear Phagocyte System. Adv.
Kim, S.H., Jeong, J.H., Chun, K.W., Park, T.G., 2005. Target-specific cellular uptake of
Drug Delivery Rev. 17, 31–48.
PLGA nanoparticles coated with poly(L-lysine)-poly(ethylene glycol)-folate
Valencia, P.M., Pridgen, E.M., Perea, B., Gadde, S., Sweeney, C., Kantoff, P.W., Bander,
conjugate. Langmuir 21, 8852–8857.
N.H., Lippard, S.J., Langer, R., Karnik, R., Farokhzad, O.C., 2013. Synergistic
Liu, Y., Hui, Y., Ran, R., Yang, G.Z., Wibowo, D., Wang, H.F., Middelberg, A.P.J., Zhao, C.
cytotoxicity of irinotecan and cisplatin in dual-drug targeted polymeric
X., 2018. Synergetic Combinations of Dual-Targeting Ligands for Enhanced In
nanoparticles. Nanomedicine 8, 687–698.
Vitro and In Vivo Tumor Targeting. Adv. Healthc. Mater. 7, e1800106.
Wu, Z., Nguyen, N.-T., 2005. Rapid mixing using two-phase hydraulic focusing in
Liu, Y., Yang, G., Baby, T., Tengjisi, Chen, D., Weitz, D.A., Zhao, C.-X., 2020a. Stable
microchannels. Biomed. Microdevices 7, 13–20.
Polymer Nanoparticles with Exceptionally High Drug Loading by Sequential
Zhao, C.X., 2013. Multiphase flow microfluidics for the production of single or
Nanoprecipitation. Angew. Chem. Int. Ed. 59, 4720–4728.
multiple emulsions for drug delivery. Adv. Drug Deliver. Rev. 65, 1420–1446.
Liu, Y., Yang, G., Jin, S., Xu, L., Zhao, C.-X., 2020b. Development of high-drug-loading
Zhao, C.-X., He, L., Qiao, S.Z., Middelberg, A.P.J., 2011a. Nanoparticle Synthesis in
nanoparticles. ChemPlusChem 85, 2143–2157.
Microreactors. Chem. Eng. Sci. 66, 1463–1479.
Liu, Y., Yang, G., Jin, S., Zhang, R., Chen, P., Tengjisi, Wang, L., Chen, D., Weitz, D.A.,
Zhao, C.-X., Miller, E., Cooper-White, J.J., Middelberg, A.P.J., 2011b. Effects of Fluid-
Zhao, C.-X., 2020c. J-Aggregate-Based FRET Monitoring of Drug Release from
Fluid Interfacial Elasticity on Droplet Formation in Microfluidic Devices. AIChE J.
Polymer Nanoparticles with High Drug Loading. Angew. Chem. Int. Ed. 59,
57, 1669–1677.
20065–20074.