Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
High-throughput nanoparticle fabrication and nano-biomembrane interactions
(USC Thesis Other)
High-throughput nanoparticle fabrication and nano-biomembrane interactions
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
HIGH-THROUGHPUT NANOPARTICLE FABRICATION AND
NANO-BIOMEMBRANE INTERACTIONS
by
Lu Wang
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMICAL ENGINEERING)
May 2020
Copyright 2020 Lu Wang
ii
Acknowledgements
I would like to express my sincere appreciation to those who have supported me throughout my Ph.D.
program. First, I would like to thank my advisor, Dr. Noah Malmstadt, for his invaluable insights and
his belief in my abilities. Without his guidance and support, I could not possibly complete my
dissertation with such passion and independent thinking. I would also like to thank my committee
members, Dr. Richard Brutchey and Dr. Nicholas Graham. It was through collaborations with them
that I was encouraged to broaden my mind and explore more opportunities.
I am grateful for my mentors, Dr. Shalene Sankhagowit, Dr. Carson Riche, Dr. Kaixuan Ren,
Dr. Gertrude Gutierrez, and Dr. Krisna Bhargava, for their generous help and patience that cannot be
overestimated. I also had great pleasure working with my labmates Dr. Bryant Thompson, Nareh
Movsesian, Ahmed Elbaradei, Justin So, Sepehr Maktabi, Majed Madani, and Lucia Dalle Ore for
their heartfelt encouragement and insightful discussions. I would also like to extend my gratitude to
my collaborators Dr. Emily Robert, Dr. Lucia Mora, Lanja Karadaghi, and Nicholas Hartel, for
working with me and pushing forward through the challenges.
I am also thankful for the staff members at USC who have provided me assistance, especially for
Phillip Sliwoski from glass shop and Dr. Shuxing Li from NanoBiophysics Center, whose exceptional
skill and knowledge had made my research stronger. And many thanks to Andy Chen, Martin
Olekszyk, and Veronica Elizarraras for their valuable help that have made life of a graduate student
easier.
Last but not least, I am extremely grateful to my family, for their warm love and unwavering
support with me in whatever I pursue. I would also like to thank my husband, Dr. Weihua Guo, who
has been my pillar of strength and source of inspiration. I am truly grateful.
iii
Table of Contents
Acknowledgements ..................................................................................................................... ii
List of Tables . ............................................................................................................................. vi
List of Figures ............................................................................................................................ vii
Abbreviations............................................................................................................................. xii
Abstract . ........................................................................................................................... xiii
Chapter 1 . Background ............................................................................................................. 1
1.1 The Rise of Nanotechnology .................................................................................... 1
1.2 Scale up Nanoparticle Synthesis in Flow .................................................................. 1
1.2.1 Superior Characters of Micro- and Milli-fluidic Systems for Colloidal
NPs Synthesis............................................................................................ 2
1.2.2 Continuous Single-Phase and Droplet Micro/Millifluidic Reactors .............. 3
1.3 Ubiquitous Existence of Nanoparticles in the Environment ....................................... 5
1.3.1 Pervasive Exposure of Nanoparticles .......................................................... 5
1.3.2 Underestimated Threat of Micro- and Nano-Plastics ................................... 6
1.4 Exploring the Health Effects of Nanoparticles with Biomembranes ............................ 7
1.4.1 Nanoparticle Toxicity Studies .................................................................... 7
1.4.2 Model Biomembranes in Nano-Membrane Interaction Studies .................... 8
1.4.3 Electrostatic Forces in Non-Specific Nano-Biomembrane Interactions ....... 11
Chapter 2 . Scale-up Synthesis of Transition Metal Nanoparticle Catalysts ................................ 13
2.1 Challenges in Continuous Flow Synthesis of Nanoparticle Catalysts ........................ 13
2.2 Continuous Synthesis of Nickel Nanoparticles for the Catalytic
Hydrodeoxygenation of Guaiacol* ......................................................................... 15
2.2.1 Motivation .............................................................................................. 15
2.2.2 Millifluidic Synthesis of Ni-NPs ............................................................... 16
2.2.3 Characterization and Catalytical Performance of B-Ni-NPs and mF-Ni-
NPs ........................................................................................................ 17
2.3 Continuous Synthesis of Molybdenum Carbide Nanoparticles for
Thermocatalytic CO 2 Hydrogenation* .................................................................... 19
2.3.1 Motivation .............................................................................................. 19
2.3.2 Millifluidic Synthesis of NP-MoC 1-x
.......................................................... 20
2.3.3 Catalytical Performance of NP-MoC 1–x
..................................................... 24
2.4 Conclusion ............................................................................................................ 24
iv
Chapter 3 . Self-Optimizing Parallel Millifluidic Reactor for Scaling Nanoparticle Synthesis* ..... 26
3.1 Abstract ................................................................................................................ 26
3.2 Motivation ............................................................................................................ 26
3.3 Results and Discussion .......................................................................................... 29
3.3.1 Design of the Reactor .............................................................................. 29
3.3.2 Parallel Slug Behaviors and Photoluminescent Spectra .............................. 32
3.3.3 Self-Optimizing Feedback Loop ............................................................... 34
3.3.4 Extended Operation................................................................................. 36
3.4 Experimental ........................................................................................................ 38
3.4.1 Precursor Preparation and Product Work-up ............................................ 38
3.4.2 Product Characterization ......................................................................... 39
3.4.3 Equipment and Devices in the Millifluidic Reactor ................................... 39
3.4.4 Example Spectra Collection ..................................................................... 40
3.4.5 Feedback-Control System ........................................................................ 41
3.5 Conclusions .......................................................................................................... 42
Chapter 4 . Interactions between Charged Nanoparticles and Giant Vesicles Fabricated from
Inverted-Headgroup Lipids* .................................................................................. 43
4.1 Abstract ................................................................................................................ 43
4.2 Motivation ............................................................................................................ 43
4.3 Result and Discussion............................................................................................ 45
4.3.1 Influence of Nanoparticle Charge on Particle-Bilayer Interactions .............. 45
4.3.2 Influence of Nanoparticle Size on Particle-Bilayer Interaction .................... 49
4.4 Experimental ........................................................................................................ 53
4.4.1 Chemicals ............................................................................................... 53
4.4.2 GUV Formation ...................................................................................... 53
4.4.3 Observation Chamber Preparation ........................................................... 54
4.4.4 Microscopy ............................................................................................. 54
4.4.5 Data Analysis .......................................................................................... 54
4.5 Conclusions .......................................................................................................... 55
Chapter 5 . Effect of Protein Corona on Nanoparticle-Plasma Membrane and Nanoparticle-
Biomimetic Membrane Interactions* ...................................................................... 56
5.1 Abstract ................................................................................................................ 56
5.2 Motivation ............................................................................................................ 57
5.3 Results and Discussion .......................................................................................... 59
5.3.1 Protein Corona Characterization .............................................................. 59
5.3.2 Cell-PNP Interactions .............................................................................. 63
5.3.3 Biomimetic Membrane-PNP Interaction ................................................... 69
5.3.4 Leakage of Biomimetic Membranes with PNPs......................................... 72
5.4 Experimental ........................................................................................................ 77
5.4.1 Materials ................................................................................................. 77
5.4.2 Protein Corona Preparation and Quantification ........................................ 78
v
5.4.3 Protein Corona Elution and SDS-PAGE .................................................. 78
5.4.4 Proteomic Analysis .................................................................................. 78
5.4.5 DLS and Zeta Potential Measurement ...................................................... 79
5.4.6 Cell Culture and Imaging ......................................................................... 79
5.4.7 Cell Viability MTT Assay ........................................................................ 80
5.4.8 Lactate Dehydrogenase (LDH) Release Assay .......................................... 80
5.4.9 GPMV Preparation ................................................................................. 81
5.4.10 GUV Preparation .................................................................................... 81
5.4.11 Microscopy Imaging of Model Vesicles and Quantification ....................... 81
5.5 Conclusion ............................................................................................................ 82
Chapter 6 . Conclusions and Future Outlook ............................................................................ 84
References . ............................................................................................................................ 86
vi
List of Tables .
Table 3.1. CsPbBr 3 QD synthesis was carried out in the parallel reactor (see Section 3.4.4). Example
emission peak wavelength and FWHM of the PL spectrum in each of the 16 parallel
channels. ................................................................................................................................. 33
Table 3.2. Communication and control approaches for the devices used in the reactor. ............................ 41
Table 4.1. Wilcoxon rank-sum test results for comparing circularity before and after GUVs interacting
with no PNPs, positively charged PNPs, and negatively charged PNPs. Z is the Z-ratio for
the test. .................................................................................................................................... 46
Table 4.2. Wilcoxon rank-sum test data of GUVs circularity distribution after experiments with 500
nm positively/negatively charged PNPs, compared with the distribution before each
experiment. Z is the Z-ratio for the test. .................................................................................... 47
Table 4.3. Wilcoxon signed rank test data of GUVs circularity distribution after experiments with
positively/negatively charged PNPs at three different sizes, compared with the distribution
before each experiment. Z is the Z-ratio for the test. ................................................................. 51
Table 5.1. Properties of the polystyrene nanoparticles with and without protein corona. All
measurements were performed in PBS buffer; protein mass was determined via BCA assay.
................................................................................................................................................ 60
Table 5.2. Representative proteins identified in the protein corona formed on uncharged, carboxyl-,
and amine-polystyrene nanoparticles by LC-MS-MS. ............................................................... 62
Table 5.3. Hydrodynamic diameters and Zeta potential of PNPs in FBS-absent cell culture media
DMEM and complete media cDMEM. ................................................................................... 66
Table 5.4. Unpaired t-test result of viability MTT assay. ........................................................................... 67
Table 5.5. Unpaired t-test result of viability LDH assay. ........................................................................... 68
vii
List of Figures
Figure 1.1. Transmission electron microscopy images of Ag nanoparticles synthesized in droplet
millifluidic reactor through seed-mediated growth. The shape and size distributions of Ag
nanoparticles depend the species of the carrier fluid as well as the speed of the carrier fluid.
Image taken from Zhang, et al. Langmuir (2013).
27
..................................................................... 5
Figure 1.2. Model membranes. Among spherical vesicles (a), supported planar bilayers (b), interfacial
monolayers (c), vesicles on planar supported (d) and unsupported planar bilayers (e), (b)
and (d) are supported and (c) and (e) are unsupported. Spherical vesicles include
freestanding vesicles in aqueous solution, supported bilayers on nanoparticles or
microparticles, and vesicles on planar support. Image taken from Chen, et al. Environ. Sci.
Technol. (2014).
64
........................................................................................................................ 9
Figure 1.3. Schemes illustrating three stages of giant vesicle formation. (a) Orientation of self-assembled
lamellar lipid bilayers. (b) Growth of liposomes promoted by forces normal to the bilayers.
(c) Fusion of adjacent liposomes into giant vesicles. The black lines in (b) and (c) represent
lipid bilayers. Image is taken from Horger, et al. J. Am. Chem. Soc. (2009).
81
............................. 10
Figure 1.4. Molecular structure of DPPG and DPTAP lipids (a). Cryo-TEM images of (B)
DPPC/DPPG (3:1) liposomes with Ag–NH nanoparticles (b, c) and DPPC/DPTAP (3:1)
liposomes with Ag–COOH nanoparticles after centrifugation (d, e). (b) and (d) display the
supernatants, (c) and (e) display the sediments. The scale bars represent 200 nm. Image is
adapted from Xi, et al. Analyst (2014).
94
................................................................................... 12
Figure 2.1. Reaction conditions of colloidal transition metal nanoparticle synthesis and typical flow
chemistry. Higher temperature and longer residence time are required for transition metal
nanoparticle synthesis in solution phase. .................................................................................. 14
Figure 2.2. Scheme of millifluidic reactor system for the continuous flow production of Ni-NPs. The
injection flow rate of metal precursor is controlled by a feedback loop based on mass loss
from the precursor jar. Precursor flows continuously through a PTFE tubing in an oven at
220 °C, Ni NPs are synthesized in a time course of 16 min. Back flow induced by gas
evolution in the reactor was prevented by a one-way check valve at the inlet. ........................... 17
Figure 2.3. (a,b) XRD patterns with reference peaks for Ni (PDF# 99-000-2639), (c,d) TEM
micrographs, and (e,f) corresponding histograms of colloidal mF-Ni-NPs (11.1 ± 3.1 nm;
σ/d = 27%) and B-Ni-NPs (8.8 ± 2.4 nm; σ/d = 27%), denoted by green and purple colors,
respectively. ............................................................................................................................. 18
Figure 2.4. Scheme of the millifluidic reactor system for the continuous flow synthesis of NP- MoC 1–x.
Metal precursor is delivered into the glass millifluidic reactor in sand bath at 320 C, a check
valve at the inlet and a back pressure regulator at the outlet ensure a 40 psig pressure inside
the system, facilitating the control over sublime of the metal precursor. An analytical
balance and a flow meter are incorporated in the system to determine the residence time of
the reaction. ............................................................................................................................ 20
Figure 2.5. Powder XRD patterns for the -MoC 1–x NPs synthesized in flow at four different precursor
concentrations. ........................................................................................................................ 21
Figure 2.6. TEM images of the NP-MoC 1–x synthesized in flow using (a) 625 mM, (b) 312 mM, (c) 156
mM, and (d) 78 mM Mo(CO) 6 precursor solutions. .................................................................. 22
viii
Figure 2.7. (a) Mass of product collected over time for 625 mM, 312 mM, 156 mM, and 78 mM
Mo(CO) 6 precursor solutions. The linearity indicates that steady state was reached in the
millifluidic reactor. (b) Representation of gas and liquid flow in the millifluidic reactor and
the data generated by the dissipation flow sensor to distinguish the liquid and gas phases
and determine the residence time for fluid plugs in the reactor. ................................................ 23
Figure 2.8. Reaction residence time and NP-MoC 1–x product yield in millifluidic reactor as a function
of [Mo(CO) 6]. .......................................................................................................................... 23
Figure 3.1. Schematic of the reactor system. The CsPbBr 3 perovskite QD synthesis is distributed in 16
parallel channels and monitored by feedback control based on in-situ measured PL
characteristics. ......................................................................................................................... 29
Figure 3.2. Photograph of the 16-channel parallel millifluidic reactor. Blue arrows indicate the flow
streams. ................................................................................................................................... 30
Figure 3.3. Diagram and photograph of the IR sensor apparatus ................................................................ 31
Figure 3.4. In-situ detection modules. (a) Photograph of the detection system monitoring slugs and PL
spectra. (b) Slug status was acquired via IR sensors. Diagrams of the UV-vis/PL detection
unit in (c) perspective view and (d) side view. .......................................................................... 31
Figure 3.5. Box plot (top) and histogram (bottom) displaying the distribution of slug frequencies in 16
channels. The average slug frequency is 11 Hz and the coefficient of variance is lower than
22%. ........................................................................................................................................ 32
Figure 3.6. Parallelization of CsPbBr 3 QD synthesis in the 16 channels. (a) Box plot and histogram
showing size distributions of slugs recorded across 16 parallel channels (grey, top) and in a
single channel (green, bottom) (b) Normalized offset UV-vis (dotted lines) and PL spectra
(solid lines) of the CsPbBr 3 QD product in the 16 channels were highly similar. ....................... 33
Figure 3.7. Self-optimization performance. The goal of self-optimization was set to minimize the upper
endpoint of the 95% confidence interval (UCI) in the distribution of FWHMs from the 16
channels. The driving pressures of Cs
+
/Pb
2+
precursors, Br
–
precursor, and N 2 segmentation
gas are denoted as P[Cs
+
/Pb
2+
], P[Br
–
], and P[gas], respectively. (a) The search for minimal
UCI reached convergence within 8-10 iterations (2 min per iteration). (b) Three variable
input pressures across 11 optimization iterations (labeled 1-11) are plotted with a colormap
indicating the corresponding UCI. (c) The optimization system stabilized the reactor after
a flow disturbance was introduced, and the target UCI below 35 nm was reached again by
feedback control despite the disturbance. ................................................................................. 35
Figure 3.8. TEM images and inset size distributions of CsPbBr 3 QDs from 1, 2, and 3 h time points in
the 4-h automated flow synthesis, with average sizes of 9.8 ± 1.6 nm, 10.2 ± 1.7 nm, and
9.7 ± 1.7 nm, respectively. ....................................................................................................... 36
Figure 3.9. Reactor performance in an extended automated operation. (a) Photos of collected product
in a 5 L bottle from a 4 h synthesis under indoor light (left) and UV light (right). (b) Off-line
UV-vis spectra (dotted lines) and PL spectra (solid lines) of product collected from 1-4 h
time points. In-situ (c) slug sizes, (d) slug frequencies, and (e) FWHMs along the 16 channels
are stable throughout the 4 h reaction. The mean values and 95% confidence intervals are
plotted in the graphs. (e) Powder XRD of CsPbBr 3 QDs obtained at each hour. (f) TEM of
CsPbBr 3 QDs at 4 h time point, with size distribution given as an inset. ................................... 37
Figure 3.10. Flowcharts of the feedback-control process. (a) Self-optimization procedure in the parallel
millifluidic reactor. UCI denotes the upper endpoint of 95% confidence interval of FWHMs
ix
in 16 channels, and P liq corresponds to the driving pressure of either liquid precursor. (b)
The procedure of collecting UV-vis and PL spectra through 16 channels. The spectra were
read through channel = 0 to channel = 15, i denotes the location of the linear stage and j
denotes the channel to be read in the 4-channel detection module. ........................................... 42
Figure 4.1. DOPC lipids and DOCP lipids share identical tailgroups while their headgroups have
inverted orientation of charge distribution. GUVs are formed with DOPC lipids and DOCP
lipids respectively, they are put in the presence of positively charged NH 3
+
-PNPs (left panel,
orange) or negatively charged COO
-
-PNPs (right panel, blue). Differences are observed
between the behavior of DOPC GUVs and DOCP GUVs. The sizes of GUVs and PNPs
are not to scale......................................................................................................................... 44
Figure 4.2. The fluorescent microscopic image of DOPC GUVs (red) and DOCP GUVs (green) 10 min
after introducing no PNPs, positively charged PNPs or negatively charged PNPs (a).
Circularity distribution of DOPC GUVs (b) and DOCP GUVs (c) before and after
corresponding experiments. Circularity distribution is presented as a box plot. The central
box spans interquartile range, with the segment inside indicating the median of the data,
and the fences above and below the box representing the maximum and minimum
circularity values. *p<0.005, **p<0.001, ***p<0.0001 in Wilcoxon signed rank test. Scale
bar: 20 µm. .............................................................................................................................. 46
Figure 4.3. Circularity distribution of DOPC and DOCP GUVs in experiments with 500 nm NH 3
+
-
PNPs (a, c) and with 500 nm COO
-
-PNPs (b, d) categorized by GUVs sizes. GUVs are
categorized in 6-12 m radius (small GUVs), 12-18 m radius (medium GUVs), and >18
m radius (large GUVs). .......................................................................................................... 47
Figure 4.4. Structured water around choline and phosphate group. A water molecule anchoring
clathrate around choline (top) moiety and a phosphate (bottom) oxygen atom
(intramolecular water anchor). Carbon atoms are represented in dark green, oxygen atoms
in red, hydrogen atoms in white, nitrogen atoms in blue and phosphorous atoms in light
green in this molecular model. Image taken from A. Liwo. (Springer Science & Business
Media, 2013).
163
....................................................................................................................... 48
Figure 4.5. Fluorescent microscopy images of DOPC GUVs (red) and DOCP GUVs (green) 10 min
after addition of positively charged PNPs at three different sizes (a, b, c), and 10 min after
addition of negatively charged PNPs at three different sizes (d, e, f) respectively. Scale bar:
50 µm. ..................................................................................................................................... 49
Figure 4.6. Circularity distribution of DOPC and DOCP GUVs in experiments with positively charged
NH 3
+
-PNPs (a, c) and with negatively charged COO
-
-PNPs (b, d) at PNPs diameters of 750
nm, 500 nm, and 50 nm. *p<0.005, **p<0.001, ***p<0.0001 in Wilcoxon signed rank test....... 50
Figure 4.7. Fluorescence images of DOPC GUVs (red) and DOCP GUVs (green) 10 min after the
addition of positively charged nanoparticles (a,b) or negatively charged nanoparticles (c,d)
of 500 nm and 50 nm diameters. The mass concentration of nanoparticles was kept
consistently 0.9 mg/mL. Scale bars 50 m. .............................................................................. 52
Figure 5.1. Size distribution of polystyrene nanoparticles with and without the presence of a protein
corona. Sulfate-PNPs, carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and
APNP. .................................................................................................................................... 60
Figure 5.2. Comparison of protein corona composition on sulfate-, carboxyl-, and amine-polystyrene
nanoparticles (SPNPs, CPNPs, and APNPs respectively). (a) Coomassie blue-stained SDS-
PAGE gel of human plasma proteins obtained from corona on SPNPs, CPNPs, and APNPs.
(b) LC-MS-MS result of proteins identified in the corona formed on SPNPs, CPNPs, and
x
APNPs. This Venn diagram reports the number of unique proteins identified from each of
three nanoparticles as well as proteins common to two or all three nanoparticle populations.
(c) Classification of corona proteins identified by LC-MS-MS according to their calculated
isoelectric point (pI); relative percentages are shown. ............................................................... 61
Figure 5.3. Effect of the protein corona on cellular adhesion of nanoparticles and cell viability. Sulfate-
PNPs, carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP. (a)
Confocal microscopy images of 293T cells show that adsorption of nanoparticles was
reduced with the presence of a protein corona. Images were taken after 4 h incubation of
the cells with nanoparticles in FBS-free culture media. The green channel corresponds to
the fluorescently labeled nanoparticles, blue channel corresponds to DAPI stained nuclei,
and red fluorescence signal comes from CF633-WGA stained cell membranes. The scale
bars are 30 µm. (b, c) Cell viability of 293T cells exposed to nanoparticles. Cells were
incubated with nanoparticles for (b) 4 h and (c) 15 h, under conditions of presence or
absence of protein corona as well as FBS included or excluded from the culture medium.
Cell viability was evaluated using the MTT assay, the viability is normalized based on the
control group where no PNPs were added. LDH leakage of 293T cells exposed to
nanoparticles for (d) 4 h and (e) 15 h were assessed, under conditions of presence or absence
of protein corona as well as FBS included or excluded from the culture medium, the leakage
percentage is normalized based on the negative control group (0%) where no PNPs were
added and the positive control group (100%) where cells were treated with lysis buffer.
(Unpaired t-test, * significant at p < .05, ** significant at p < .01, *** significant at p < .001,
detailed test results are listed in Table 5.4 and Table 5.5) .......................................................... 64
Figure 5.4. Fluorescent intensity of adsorbed nanoparticles on 293T cells after 4 h and 15 h incubation.
Medians and interquartile ranges of calibrated fluorescence intensity were demonstrated
along with individual values in graphs. There is no significant increase of fluorescent
intensity between 4 h and 15 h. Sulfate-PNPs, carboxyl-PNPs, and amine-PNPs are denoted
as SPNP, CPNP and APNP. (Unpaired t-test, * significant at p < .05, ** significant at p
< .01, *** significant at p < .001). ............................................................................................ 65
Figure 5.5. Effect of the protein corona on nanoparticle binding to biomimetic membranes. Sulfate-
PNPs, carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP. (a)
Confocal microscopy image of DiD-stained brain lipid GUVs (640 nm excitation) and green
fluorescent nanoparticles (491 nm excitation) after 4 h incubation. (scale bar: 30 µm). (b)
Fluorescent intensity of adsorbed nanoparticles on lipid membranes of GPMVs and GUVs
after 4 h incubation. Medians and interquartile ranges of calibrated fluorescence intensity
were demonstrated along with individual values in graphs. The adsorption of amine-PNPs
(APNPs) was significantly decreased by protein corona (Unpaired t-test, * significant at p
< .05, ** significant at p < .01, *** significant at p < .001) ........................................................ 70
Figure 5.6. Confocal microscopy images of DiD-stained GUVs and GPMVs with green fluorescent
PNPs after 4 h incubation. Scale bar in GUVs panels: 30 µm, in GPMVs panels: 10 µm.
Sulfate-PNPs, carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP. ...... 71
Figure 5.7. Fluorescent intensity of adsorbed nanoparticles on lipid membranes of GPMVs and GUVs
after 15 h. Medians and interquartile ranges of calibrated fluorescence intensity were
demonstrated along with individual values in graphs. Sulfate-PNPs, carboxyl-PNPs, and
amine-PNPs are denoted as SPNP, CPNP and APNP. (Unpaired t-test, * significant at p
< .05, ** significant at p < .01, *** significant at p < .001). ....................................................... 72
Figure 5.8. Effect of PNPs and protein corona on model membrane integrity. Sulfate-PNPs, carboxyl-
PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP. (a) Confocal microscopy
image of DiD-stained model membrane vesicles (red fluorescence) in 0.5 mg/mL calcein
(green fluorescence) buffer after 15 h exposure to nanoparticles. White arrows point at the
xi
vesicles that had calcein leakage through membranes (scale bars in GUV panels: 60 µm;
scale bars in GPMV panels: 30 µm). (b-c) Population of leaked vesicles after treatment of
PNPs. Percentages of leaked vesicles after 4 h (b) and 15 h (c) incubation with PNPs are
presented in graphs companied with control groups where PNPs were absent. (Unpaired t-
test, * significant at p < .05, ** significant at p < .01, *** significant at p < .001) ....................... 73
Figure 5.9. Effect of PNPs and protein corona on membrane integrity of pure lipid GUVs. DOPC
GUVs and POPC GUVs were incubated with PNPs in 0.5 mg/mL calcein buffer for 4 h
and 15 h. Inflow of calcein was observed after incubation with PNPs, percentages of leaked
vesicles are presented in graphs companied with control groups where PNPs were absent.
(a) Confocal microscopy image of DOPC GUVs (DiD stained, red fluorescence) in 0.5
mg/mL calcein (green fluorescence) buffer for 15 h in the absence of PNPs. (scale bar: 30
µm). (b-e) Percentages of leaked DOPC GUVs after 4 h (b) and 15 h (d) incubation with
PNPs as well as relative leaked population of POPC GUVs after 4 h (c) and 15 h (e)
incubation with PNPs. The control groups in all the leakage assays showed 0% leaked
population. Sulfate-PNPs, carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP
and APNP (Unpaired t-test, * significant at p < .05, ** significant at p < .01, *** significant
at p < .001) .............................................................................................................................. 75
Figure 5.10. Effect of PNPs and protein corona on model membrane integrity. GPMVs and GUVs were
incubated with PNPs in 0.5 mg/mL 10 kDa rhodamine-dextran buffer for 4 h and 15 h.
Inflow of dextran was observed after incubation with PNPs, percentages of leaked vesicles
are presented in graphs companied with control groups where PNPs were absent. Sulfate-
PNPs, carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP.
(Unpaired t-test, * significant at p < .05, ** significant at p < .01, *** significant at p < .001)
................................................................................................................................................ 76
xii
Abbreviations
BSA Bovine serum albumin
cDMEM Complete cell culture medium
DMEM Dulbecco's modified Eagle's medium
DOCP 2-((2,3-bis(oleoyloxy)propyl)-dimethylammonio) ethyl hydrogen phosphate
DOPC 1,2-dioleoyl-sn-glycero-3-phosphocholine
FWHM Full width at half maximum
GPMVs Giant plasma membrane vesicles
GUVs Giant unilamellar vesicles
IR Infrared radiation
IW Incipient wetness
NPs Nanoparticles
PBS Phosphate buffered saline
PEEK Polyether ether ketone
PL Photoluminescence
PNPs Polystyrene nanoparticles
POPC 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine
PTFE Polytetrafluoroethylene
QDs Quantum dots
SMRs Segmented-flow reactors
TEM Transmission electron microscopy
TMCs Transition metal carbides
UV-vis Ultraviolet–visible
XRD Powder X-ray diffraction
xiii
Abstract .
The nanotechnology revolution arising in the last century has led to the synthesis and understanding
of a variety of nanoparticles, bringing about an extraordinary potential in sustainable energy, health
care, and materials development. To cope with the extensive demand of nanoparticle production,
there have been various industrial-scale synthesis techniques facilitated by gas phase process.
Nanoparticles are also commonly synthesized in liquid phase for more uniform size and shape.
However, scale-up of liquid-phase synthesis suffers from irregularities in large-scale batch process,
which can have huge negative effects on nanoparticle products, compromising their efficiency in later
applications. High-throughput and well-controlled continuous flow reactor systems are promising
approaches for colloidal nanoparticle fabrication to be transferred into industry. On the other hand,
the development of nanotechnology has also created environmental contamination and unnecessary
exposure to human health. A better understanding of nano-bio interfaces and interactions can provide
more fundamental insight for the health effects of nanoparticles.
This report begins with a general background introduction of nanotechnology development and
its effect on the environment and human health. The research work in the following chapters focuses
on nanoparticle synthesis in millifluidic reactors and interactions between nanoplastics and model
biomembranes. In chapter 2, millifluidic syntheses of transition metal nanoparticle catalysts at high
temperatures are presented. The superior heat and mass transfer in millifluidic reactors offered
superior catalytical properties of nanoparticles with shorter reaction times. In chapter 3, parallelization
scale-up of millifluidic reactors is introduced. With self-optimizing feedback control, the automation
of this system is readily achieved. Chapter 4 and chapter 5 discuss the interaction of nano-bio interfaces,
utilizing polystyrene nanoplastics and model biomembranes. Using model membranes composed of
lipid molecules with different charge properties, the role of electrostatic forces in the non-specific
interactions between nanoplastics and biomembranes are elucidated. Furthermore, nanoparticles are
xiv
known to gather protein molecules upon contact with biological milieu, forming a protein corona on
their surface, which affects their biological identity and fate. In the presence of a protein corona, we
have observed a general decrease of electrostatic interactions between nanoparticles and
biomembranes and membrane perturbation across in-vitro and model membranes.
1
Chapter 1 . Background
1.1 The Rise of Nanotechnology
Nanotechnology is defined as manipulation of various types of materials at the nanoscale level.
Amongst materials at the nanoscale, nanoparticles (NPs) are wide class of tiny particulate substances,
which have at least one dimension ranges from 1 nm to 100 nm.
1
When material size goes down to
the nanoscale, the physical and chemical properties of the material start to differ from its bulk form.
Attributing to their nano-specific properties such as high surface-to-mass ratio, nanoparticles have
significant potential in catalysts enabling sustainable chemistry. Taking advantage of the quantum size
effect and electronic properties, nanoparticles are readily applied in optics, sensors, and solar cells.
Owing to other properties such as the high surface energy and the ability to circulate inside the human
body, medical studies have made breakthroughs in drug delivery via nanoparticles.
In recent decades, nanotechnology has developed towards flexible design and commercialization
of nanoparticles, aiming for efficient and broad applications, while at the same time facilitating a better
balance between nano-industry development and environmental sustainability.
1.2 Scale up Nanoparticle Synthesis in Flow
Top-down and bottom-up strategies are used to manufacturing the nanoparticles. Top-down approach
involves breaking bulk materials into nanoscale powders, such as milling process. Despite its wide
application in industry, it lacks control over morphology features of nanoparticles. Bottom-up
approach assembles the nanoparticles from atoms and molecules through nucleation and growth.
2
More complex structures, better control over sizes and shapes can be achieved in bottom-up strategy.
Gas phase processes are commonly applied at industrial scale, such as vapor deposition and aerosol
processes.
2
Liquid phase syntheses typically take place at lower temperatures, and it is better suited for
manufacturing nanoparticles with uniform size and shape.
3
In conventional batch process systems, a
great variety of nanoparticles are synthesized as colloidal particles in solution by hot injection,
chemical precipitation, micro emulsion, and reduction methods.
4
Typical batch reactors on the
laboratory scale (10-100 mL) are not competent to fulfill the demands of industrial production. To
scale up nanoparticle synthesis, increasing volume or reagent concentrations in batch reactors can lead
to inhomogeneities in mass and thermal transport, which leads to poor reproducibility between
batches. Alternatively, continuous microfluidic and millifluidic systems have been recognized as one
of the most promising solutions addressing the issues of scalability as a new frontier of synthetic
nanoscience.
5
1.2.1 Superior Characters of Micro- and Milli-fluidic Systems for Colloidal NPs Synthesis
Microfluidic and millifluidic reactors afford superior reaction control for the production of large
amounts of nanoparticles due to the small dimensions of their channels. Quantitative comparisons of
transport phenomena in batch reactors and microfluidic reactors have been addressed with analytical
approaches.
6, 7
Convective mass and heat transfer dominate in batch reactors. Thorough mixing in
batch reactors requires high Reynolds numbers and optimized impeller geometries, and mixing
becomes more difficult with increasing volume in the scale-up process. Diffusive mass transfer and
conductive heat transfer are dominant in continuous flow reactors. These transfer rates are inversely
proportional to the square of the characteristic channel size, making heat and mass transfer in micro-
and millifluidic reactors inherently advantageous. The ratio of reaction rate to mass-transfer rate in a
reactor is described by the Damköhler number (Da). Da < 1 indicates a kinetically reaction limited
system, where mixing is sufficiently fast, while Da > 1 indicates that mixing is limiting and may result
in undesired products and lower yield. When batch reactions are scaled, Da increases, but micro- and
3
millifluidic reactors can be scaled (either by increasing run time or by parallelization) without
increasing Da.
8
The superior heat and mass transport stemming from the high surface-area-to-volume
ratios of the channels allows these large quantities to be attained without compromising control over
their resulting morphology and polydispersity.
The reduced reaction volume promotes fast heating and cooling, which minimizes energy inputs
and reaction times, leading to more sustainable processes. Small operating volumes and the enclosed
nature of the system also reduce potential risks of exposure to hazardous and volatile compounds.
9
In
addition, online analytical monitoring systems can be conveniently introduced to inspect the progress
of NP synthesis in micro- and milli-fluidic reactors. This real-time information about the product
allows for quick feedback control and optimization of flow reactor parameters, which ensures product
quality with minimized waste of reaction reagent.
10, 11
High mixing efficiency as well as rapid heat and mass transfer afford high quality NPs with
increased yield and lowered reaction time.
12, 13
The amenability to parallelization, the compatibility
with in-situ analysis and precise parameter control allows for automation with reproducibility with
high throughput and high fidelity.
14
All these qualities provide micro- and milli-fluidic systems with
great potential for industrial production and commercialization of colloidal NPs.
1.2.2 Continuous Single-Phase and Droplet Micro/Millifluidic Reactors
Continuous single-phase flow reactors, in which the homogeneous reaction reagent flows
uninterrupted through the channel, can either be conducted with a single precursor stream or with
multiple miscible precursor streams that are mixed through diffusion. Single-phase flow micro- and
millifluidic reactors have been successfully employed in colloidal nanoparticle syntheses.
15-17
However,
scale-up production in the continuous flow system is hindered by its flow characteristics. Mixing
efficiency of passive diffusion in the low-Reynolds-number regime often must be enhanced with
additional active mixers.
18
The non-slip boundary condition at the fluid-channel interface causes a
parabolic velocity profile across the channel’s cross section, leading to broad residence time
4
distribution as well as inhomogeneities in reagent concentration.
19
Moreover, the slow velocity near
the channel wall can result in adhering of nanoparticles, which might impair the product quality and
even clog the channel.
Continuous flow can be improved by segmenting the reaction stream into droplets. Droplet flow
can be generated by introducing an immiscible carrier fluid to divide the reaction stream into discrete
segments. The recirculation inside each droplet intrinsically introduces convective flows that facilitate
rapid reagent mixing and low dispersion without help of additional mixers.
20
Low-cost mineral oil and
silicone oil are often used as carrier fluid to segment reagent stream in synthesis at lower temperature.
21
Fluorinated oil is used in more harsh conditions, owing to its exceptional chemical inertness and
higher decomposition temperature compared to mineral oil and silicone oil.
22
Another approach to
generate segmented flow is to apply gas as a segmenting phase. Inert gases such as nitrogen and argon
can be employed to avoid undesired byproducts.
23
On the other hand, gas generated in-situ during
precursor decomposition can serve as natural segmenting phase.
24
Droplet micro- and millifluidic reactors allow for excellent reproducibility, because droplets serve
as identical miniaturized reactor vessels in terms of reagent mixing and reaction time. The droplets
are geometrically confined and separated from each other, eliminating the parabolic flow profile and
allowing for more uniform residence time. The carrier fluid often isolates the droplets from contacting
the channel wall while internal circulation in the droplets facilitates reagent mixing. In this way,
fouling on channel walls can be mitigated.
25, 26
5
Figure 1.1. Transmission electron microscopy images of Ag nanoparticles synthesized in droplet millifluidic reactor
through seed-mediated growth. The shape and size distributions of Ag nanoparticles depend the species of the carrier fluid
as well as the speed of the carrier fluid. Image taken from Zhang, et al. Langmuir (2013).
27
Additionally, flexible control over morphology of particles can be achieved with the assistance of
the carrier fluid.
28
Zhang and coworkers have managed to control shape and size distribution of the
silver nanocrystals in droplet millifluidic synthesis (Figure 1.1).
27
In this seed-mediated synthesis, they
found out that air served as a better segmenting phase compared to silicone oil, because it provided
O 2 for the generation of reducing agent while also serving as a buffer space for the diffusion of the
byproduct NO gas. Additionally, faster flow rate of air offered adequate O 2 and improved mixing in
the droplets, which led to narrow distribution of both shape and size.
1.3 Ubiquitous Existence of Nanoparticles in the Environment
1.3.1 Pervasive Exposure of Nanoparticles
The rapid nanotechnology revolution not only amazes us with numerous functional nanoparticles,
but also leads us to better understanding of matter at the nanoscale. We have started to realize that
nanoparticles exist everywhere, including naturally occurring nanoparticles such as volcano ashes,
fine sand and dust, and mineral particles in mines.
29
Aside from the natural sources, massive amounts
of nanoparticles have also been artificially produced and distributed. It has been found that inevitable
contamination of nanoparticles has started to enter ecosystems, bringing up concerns about its
influence on environment and human health.
30
6
The health concern might apply to the exposure of nanoparticles from medical diagnostics and
drug delivery, cosmetics, and daily products that contain engineered nanoparticles.
31
Exposure not
only arises from engineered nanoparticles with intentional release, but also unintentionally produced
nanomaterials such as ultrafine soot from combustion, as well as nanoparticle emission in the
environment such as leaching from landfill. Humans can be exposed to nanoparticles through direct
exposure to air, soil or water. Furthermore, indirect consumption of nanoparticles through
accumulation in the food chain is also possible.
32
1.3.2 Underestimated Threat of Micro- and Nano-Plastics
Plastic debris, especially micro- and nanoplastics, are excessively widespread pollution in terms of
quantity and variety. However, their impact has long been underestimated, due to lack of systematic
statistics, besides, micro- and nanoplastics have traditionally been considered inert compared to other
nanomaterials. There is increasing focus on microplastics with mechanisms of their generation well
explored.
33
With the rapid evolving of science in this field, nanoplastics are also recognized very
recently as a result of plastic fragmentation down to lower scales.
34
Micro- and nanoplastics are
released into environment during production of raw materials, usage of products, as well as waste
management after the disposal of plastics. Emission during production and applications includes
product categories in waterborne paints, adhesives, coatings, electronics, magnetics and
optoelectronics.
35
Processes involving thermal treatment of plastics, even 3D printing, can produce
nanoplastics.
36
Degradation and fragmentation after depositing in landfills, as well as weathering in
marine waste disposal, are also major sources of nanoplastics.
37
It has been estimated that as of 2015,
approximately 6300 million metric tons of plastic waste had been generated, while only 2500 million
metric tons of plastics are currently in use.
38
Micro- and nanoplastics are readily produced in our daily life, such as exfoliating particles in
cleaning products,
39
plastic microfibers shedding from synthetic textiles during washing cycles,
40
and
debris from tire abrasion while driving.
41
Before we realize it, these micro- and nanoplastics are emitted
7
into the aquatic and terrestrial systems. And it is possible that with the assistance of atmospheric
transport and deposition, micro- and nanoplastics are causing planetary-scale exposure. Microplastics
are even found with in rain fall and snow at places where anthropogenic activities are rare, such as
national parks, the Alps mountains, and even the Arctic.
42, 43
Evidence of the existence of nanoplastics
in marine organisms has been found,
44
and scientists have found microplastics in human stool.
45
While
due to lack of standardized assessment, the accumulation of micro- or nanoplastics through food chain
has not been confirmed,
46
the potential health risk that micro- and nano-plastics pose to biological
organisms has become a serious issue.
With smaller size, plastic debris might present more potential hazard toward biological organism
as the retention time might be longer in circulation. Therefore, this emerging exposure of nanoplastics
call for a better understanding of their health effects on organisms.
47, 48
1.4 Exploring the Health Effects of Nanoparticles with Biomembranes
1.4.1 Nanoparticle Toxicity Studies
Exposure to nanoparticles has become inevitable as a result of their numerous commercial
applications, such as surface coatings, textiles, electronics, and personal care.
49
Nanoparticles also
have applications in diagnostics and therapeutics,
50, 51
in which direct exposure of nanoparticles to
organs and cells is required. The potential for nanoparticles inducing acute and cumulative toxicity,
through inhalation, dermal absorption, ingestion, or medical treatment has raised scientists’
attention.
52
There have been numerous studies on nanoparticle cytotoxicity.
53
Lactate dehydrogenase
(LDH) release (indicating compromised plasma membrane integrity),
54
reactive oxygen species (ROS)
generation,
55
and nuclear DNA damage
56
have been observed for a wide range of cell types interacting
with a variety of nanoparticles.
Once entering the biological milieu, nanoparticles will rapidly absorb biomolecules, forming a
biomolecular corona. This biomolecular corona can modify the intrinsic NP surface properties,
8
thereby critically affecting the biological identity of the nanoparticles, leading to altered biological
process and results.
57
There has not been a comprehensive understanding of how biomolecule coronas
determine the fate of nanoparticles in biological systems.
58
The presence of the biomolecule corona
creates more complexity in the nanoparticle toxicity studies.
For most nanoparticles, the major pathways of cellular uptake are phagocytosis and pinocytosis.
Phagocytosis internalizes solid particles, and the process is usually achieved by specialized cells.
Pinocytosis engulfs the fluid phase, by enclosing membrane protrusions, or by the mediation of
clathrin or caveolae.
59
Chung et al.
60
found out silica nanoparticles at lower charge density undergo
clathrin-dependent endocytosis; while at high charge density, a new mechanism that does not depend
on clathrin happened in certain cells. Also, the correlation between strong nanoparticle adhesion and
high cellular uptake has been observed in several studies.
61, 62
Non-specific interactions between
nanoparticles and biomembranes play an important role in cellular uptake. Since the plasma
membrane is the site of initial contact between nanoparticles and cells, and the architecture of
biological membranes is provided by lipid bilayers, it is advantageous to utilize simplified model
membranes to elucidate the mechanism of non-specific interaction between nanoparticles and
biological membranes.
63
1.4.2 Model Biomembranes in Nano-Membrane Interaction Studies
Model membranes are composed of lipid bilayers, generally formed by self-assembly in aqueous
solution or on solid supports (Figure 1.2).
64
Unsupported or freestanding lipid bilayers consist of
spherical vesicles dispersed in aqueous solution,
65
suspended bilayers across orifices,
66
and
monolayers
67
or micelles
68
at phase interfaces. Supported bilayers usually are deposited on a planar
substrate,
69, 70
or on the surface of nanoparticles or microparticles.
71, 72
9
Figure 1.2. Model membranes. Among spherical vesicles (a), supported planar bilayers (b), interfacial monolayers
(c), vesicles on planar supported (d) and unsupported planar bilayers (e), (b) and (d) are supported and (c) and (e) are
unsupported. Spherical vesicles include freestanding vesicles in aqueous solution, supported bilayers on nanoparticles or
microparticles, and vesicles on planar support. Image taken from Chen, et al. Environ. Sci. Technol. (2014).
64
Depending on their geometries and accessibilities, these model membranes are employed for a
variety of experimental applications. Supported bilayers are utilized in surface science and their
landscape is usually examined. Non-supported bilayers usually have higher lipid packing density and
higher fluidity; their electrophysiology and morphology are commonly studied. And especially for
unsupported freestanding vesicles, they are free of restriction from solid supports and geometry
confinements, hence their lipid organization and elastic membrane deformation are best employed in
studying simultaneous nanoparticle binding.
Giant unilamellar vesicles (GUVs) represent one of the most widely utilized freestanding model
membranes to investigate membrane-nanoparticle interactions. GUVs are freestanding spherical
vesicles composed of a single lipid bilayer, with size ranging from 10 to 100 µm, which encompasses
the size range of most human cells. The size of GUVs makes changes in bilayer morphology visible
via optical microscopy. Therefore, GUVs are appropriate for achieving convincing biomimetic results.
Formation of GUVs is based on self-assembly of lipid bilayers driven by hydrophobic forces.
Several methods for producing giant vesicles have been established, including freeze-thaw,
73
inverted-
emulsion methods,
74, 75
gentle hydration,
76, 77
and electroformation.
78-80
The latter two methods are both
hydration of lamellar lipid bilayers on a planar substrate (Figure 1.3). A thin film of lipids is spread on
10
bare glass in the gentle hydration method, or on indium tin oxide (ITO)-coated glass in the
electroformation method, and the lipid film is rehydrated with aqueous buffer. In electrofomation, the
lipid film is rehydrated in the presence of an electric field. The growth of liposomes is promoted by
forces normal to the bilayers, and with further crowding and associated mechanical forces, adjacent
liposomes fuse into giant vesicles. As an alternative to gentle hydration, the agarose hydration method
was developed by Horger et al. to minimize disruption of giant vesicles on bare glass surfaces.
81
The
lipid solution is spread and dried on porous agarose gel rather than bare glass, providing rapid
fabrication with higher yield of GUVs.
Figure 1.3. Schemes illustrating three stages of giant vesicle formation. (a) Orientation of self-assembled lamellar
lipid bilayers. (b) Growth of liposomes promoted by forces normal to the bilayers. (c) Fusion of adjacent liposomes into
giant vesicles. The black lines in (b) and (c) represent lipid bilayers. Image is taken from Horger, et al. J. Am. Chem. Soc.
(2009).
81
Besides the bottom-up approaches of fabricating simplified freestanding biomembranes, top-
down approaches extract membrane components from plasma membranes. One popular model is
giant plasma membrane vesicles (GPMVs). GPMVs are blebs isolated from cells that contain
substantial lipid and protein composition of the plasma membranes, they largely preserve membrane’s
physical properties while being free from contamination of membranes from internal organelles.
82
The
prevalent vesiculation method for GPMV fabrication requires formaldehyde and dithiothreitol, which
might introduce artifacts in the following experiment results.
83
Less harmful vesiculation methods
have been developed recently, such as substitution of the vesiculation buffer with hypertonic chloride
salts buffer.
84
Vesiculation by stimulation of white light or laser have also been proved to produce
GPMVs with similar sizes and yields.
85, 86
The nanoparticle-GPMV interactions can be compared with
other model membrane systems to have comprehensive understanding of the interplay.
87
11
Another example is red blood cell (RBC) ghosts, which are membrane and cytoskeletal structures
of erythrocytes left after hemolysis, where hemoglobin and cytoplasmic are removed.
88
Fusion of RBC
membranes with nanoparticles creates a camouflage coating on the NP surface, which allows for
bypass of macrophage uptake and systemic clearance in vivo, facilitating long-circulating cargo
delivery.
89
1.4.3 Electrostatic Forces in Non-Specific Nano-Biomembrane Interactions
A few groups have focused on investigating nanoparticle-induced membrane perturbation in model
membranes, or artificial cells. They have detected nanoparticle-induced defects in supported lipid
bilayers with atomic force microscopy,
90, 91
as well as vesicle rupture and nanoparticle binding with
fluorescence imaging
92, 93
and cryo-electron microscopy.
94
Nanoparticle-induced changes in membrane
fluidity and permeability have been studied with fluorescence recovery after photo bleaching
(FRAP),
95
leakage assays
96
and electrophysiological techniques.
68
Simulation and experimental studies have suggested several mechanisms of nanoparticle-bilayer
interplay, including hydrophobic mismatch leading to embedding of nanoparticles,
97, 98
lipid packing
density theory for bilayer phase change,
99
membrane wrapping for engulfment of adsorbed
nanoparticles as well as pore formation,
100-102
steric crowding of attached nanoparticles for pore
formation, protrusion and pearling behavior.
92, 103
Interfacial interactions depend on nanoparticle surface functionalization, solution ionic strength,
and membrane lipid composition,
104
indicating that electrostatic potential plays an important role in
the interaction. Electrostatic interactions are fundamental in biology, affecting processes such as
buckling of erythrocyte membranes
105
and lumen formation within vascular endothelial cells.
106
12
Figure 1.4. Molecular structure of DPPG and DPTAP lipids (a). Cryo-TEM images of (B) DPPC/DPPG (3:1)
liposomes with Ag–NH nanoparticles (b, c) and DPPC/DPTAP (3:1) liposomes with Ag–COOH nanoparticles after
centrifugation (d, e). (b) and (d) display the supernatants, (c) and (e) display the sediments. The scale bars represent 200
nm. Image is adapted from Xi, et al. Analyst (2014).
94
Electrostatic interactions with lipid bilayers are dominated by lipid headgroups. The electrostatic
interactions with lipid membranes that have net charges such as those formed by anionic
phosphatidylglycerol (PG) lipids and cationic trimethylammonium-propane (TAP) lipids are
straightforward. Xi and Bothun found that DPPG [1,2-dipalmitoyl-sn-glycero-3-phospho-(1'-rac-
glycerol)] and DPTAP [1,2-dipalmitoyl-3-trimethylammonium-propane] lipids aggregate with
counter-ionic nanoparticles and are electrostatically repulsed from co-ionic nanoparticles (Figure
1.4).
94
Cha and coworkers also found that the interaction between charged vesicles and a counter-
charged surface immediately forms supported bilayers.
107
This effect has been utilized in encapsulating
mesoporous nanoparticles and assembling stabilized drug vehicles.
108, 109
13
Chapter 2 . Scale-up Synthesis of Transition Metal
Nanoparticle Catalysts
2.1 Challenges in Continuous Flow Synthesis of Nanoparticle Catalysts
Consistencies in reaction parameters can have strong effects on the morphologies of NPs synthesized.
The performance of nanoparticulate catalysts is largely governed by the NP morphologies.
110, 111
Hence in the large-scale production of NP catalysts, inhomogeneities resulting from the simple
volumetric scaling of batch synthesis might lead to altered nucleation and growth rates, resulting in
compromised catalytic properties of the resultant NPs. In contrast to volumetric scaling of batch
reactions, continuous flow synthesis has been identified as an alternative approach for the large-scale
production of NP catalysts with efficient heat and mass transfer as well as superior reproducibility.
112
Microfluidic reactors have recently been used in the synthesis of colloidal NPs, especially those
made from various semiconducting materials.
16, 113-116
While both microfluidic and millifluidic systems
can achieve superior heat and mass transport with their high surface-area-to-volume ratios, the
demand for high throughput favors millifluidic systems as no lithography is required for their design,
and the relatively larger channel size provides lower flow resistance, resulting in lower cost at fast
flows in a large-scale production. In addition, possible solid buildup on channel walls during
nanoparticle nucleation can often result in flow instability and even channel clogging in micro-scale
channels. Millifluidic systems have been more broadly used in nanoparticle syntheses recently.
15, 117-120
14
Figure 2.1. Reaction conditions of colloidal transition metal nanoparticle synthesis and typical flow chemistry.
Higher temperature and longer residence time are required for transition metal nanoparticle synthesis in solution phase.
The representative temperature of solution-phase transition metal NP syntheses (200-350 °C) is
far above the limit of typical micro- or milli-fluidic reactor materials such as polydimethylsiloxane
(PDMS) and polymethyl methacrylate (PMMA). Within the aforementioned temperature range,
polytetrafluoroethylene (PTFE) tubing is usually utilized when synthesis temperature is below 250 °C,
and glass is more widely employed due to its temperature and chemical compatibility, as well as its
excellent optical transmittance.
121
Stainless steel is not an advantageous option as transition metal
elements might leach out during the redox reaction at this elevated temperature.
At the same time, gas evolution can occur due to dissolved gasses in the solvents, gas-evolving
reducing agents, and decomposition of precursor. As the evolved gas displaces the liquid in the
confined channel, it pushes the liquid backward to the reactor inlet and speeds up the flow of liquid
forward to the outlet. The back flow of reagents with nanoparticle nuclei might contaminate the
precursor reservoir. On the other hand, the in-situ gas evolving reaction can accelerate the liquid flow
and reduce the residence time. In addition to that, the required reaction time for colloidal transition
metal NP synthesis is usually above 30 min (Figure 2.1). It is critical to meet the requirement for a
larger reactor size and better control over gas evolution. Herein, we have designed flow reactors for
continuous NP synthesis at high temperature with long residence time. Two proof-of-concept
15
transition metal nanoparticle catalysts are produced in our reactors with comparable catalytic
properties and superior yields.
2.2 Continuous Synthesis of Nickel Nanoparticles for the Catalytic
Hydrodeoxygenation of Guaiacol*
*This section is summary of the author’s contribution to work published in ACS Sustainable Chem. &
Eng. 2017, 5, 632-639.
122
Images in this section are taken from this publication accordingly.
2.2.1 Motivation
While biomass derived liquid fuels have the potential to supplement crude oil, the bio-oil must be
upgraded before it is suitable for use as a liquid hydrocarbon transportation fuel.
123
Deoxygenation is
critical in the upgrade as the high oxygen content of bio-oil leads to a number of undesirable
characteristics including low heating value, chemical instability, and high viscosity.
124
One promising
process for bio-oil deoxygenation is catalytic fast pyrolysis in which the pyrolysis vapors are
catalytically deoxygenated.
123, 125
As an alternative to the traditional noble metal catalysts, nickel
nanoparticles (Ni-NPs) offer excellent oxygen reduction and higher earth-abundance thus lower
cost.
126
My collaborators and I developed a high-throughput, millifluidic synthesis of well-defined,
colloidal Ni-NPs for the hydrodeoxygenation of guaiacol (2-methoxyphenol), a model compound
representative of the major lignin-derived products found in the fast pyrolysis vapor from
lignocellulosic biomass. The catalytical performance of the millifluidic-synthesized Ni-NPs in the
hydrodeoxygenation of guaiacol was compared to Ni-NPs synthesized by a batch reaction, and a Ni
catalyst prepared by the traditional incipient wetness (IW) impregnation. Overall, the millifluidic
synthesis of colloidal Ni-NPs produced catalyst with high throughput, increased yield, and maintained
comparable activity to the batch-synthesized Ni-NPs.
16
2.2.2 Millifluidic Synthesis of Ni-NPs
The high-throughput, flow synthesis of Ni-NPs was performed in a millifluidic reactor (Figure 2.2)
based on a literature synthesis by Carenco et al.
127
A constant precursor flow rate is achieved through
a feedback loop between an analytical balance and pressurized gas. This balance monitors the flux of
the precursor solution (Ni(acac) 2, oleylamine, octadecene, and trioctylphosphine) in real time, and the
computer-controlled system adjusts the pressure to maintain a constant flow rate. Before introduction
of the reactant stream into the convection oven, the precursor solution passes through an integrated
one-way valve designed to prevent backflow caused by evolution of volatile products downstream.
Upon entering the convection oven preheated to 220 °C, rapid NP nucleation occurs as a result of the
fast temperature ramp rate, yielding a dark-colored suspension.
6
At the reaction temperature, gas is
evolved from the system at a rate of 317 mL h
–1
, causing the stream to separate into discrete plugs.
The plugs are isolated from one another, preventing axial dispersion, while internal circulation within
each plug in the coiled tubing facilitates thorough mixing.
128, 129
Upon exiting the PTFE reactor (1.59
mm ID, 3.18 mm OD, 200 ft), Ni-NP growth is halted by rapid heat dissipation from the reaction
stream. The flow rate was maintained at 133 mL h
–1
of solution (measured residence time of 16 min)
to yield Ni-NPs denoted as mF-Ni-NPs. The yield of the mF-Ni-NPs was estimated to be 62% relative
to the Ni(acac) 2 precursor by gravimetric analysis. Based on the flow rate and yield, the throughput of
a single channel millifluidic device is >27 g of Ni-NPs per day, which would equate to >0.5 kg of 5
wt % SiO 2 supported Ni-NP catalyst. In an effort to synthesize Ni-NPs in a batch reaction under
analogous conditions to the millifluidic system, a batch reaction was performed at 220 °C and
thermally quenched after 16 min, to yield NPs denoted as B-Ni-NPs in a lower yield of 45%. We
attribute the higher yield of the continuous flow synthesis to result from the superior heat transfer and
mixing conditions compared to batch.
17
Figure 2.2. Scheme of millifluidic reactor system for the continuous flow production of Ni-NPs. The injection flow
rate of metal precursor is controlled by a feedback loop based on mass loss from the precursor jar. Precursor flows
continuously through a PTFE tubing in an oven at 220 °C, Ni NPs are synthesized in a time course of 16 min. Back flow
induced by gas evolution in the reactor was prevented by a one-way check valve at the inlet.
2.2.3 Characterization and Catalytical Performance of B-Ni-NPs and mF-Ni-NPs
Characterization of the crystal and catalytical properties was carried out by my collaborators. Powder
X-ray diffraction (XRD) patterns of the resulting mF- and B-Ni-NPs exhibited the diffraction peaks of
face-centered cubic structure Ni (Figure 2.3a, b). Three peaks at 44.6°, 51.9°, and 76.4° 2 θ are in
agreement with bulk Ni (PDF# 99-000-2639). Notably, no crystalline impurity phase of oxidized
nickel compounds was observed, even though nickel is more vulnerable to oxidation than noble metals.
Transmission electron microscopy (TEM) revealed a spherical morphology for both systems (Figure
2.3c, d). The mF-Ni-NPs possessed slightly larger sizes (11.1 ± 3.1 nm) than the analogous batch
prepared samples (8.8 ± 2.4 nm). Both methods gave a polydispersity with standard deviations about
the mean diameter ( σ/d) of 27%.
18
Figure 2.3. (a,b) XRD patterns with reference peaks for Ni (PDF# 99-000-2639), (c,d) TEM micrographs, and (e,f)
corresponding histograms of colloidal mF-Ni-NPs (11.1 ± 3.1 nm; σ/d = 27%) and B-Ni-NPs (8.8 ± 2.4 nm; σ/d = 27%),
denoted by green and purple colors, respectively.
Both colloidal B-Ni-NPs and mF-Ni-NPs can be successfully supported on amorphous SiO 2 for
further application in catalysts. The catalytical performance was compared between the supported
colloidal Ni-NPs and IW impregnation prepared Ni catalyst. The colloidal Ni-NPs displayed
improved catalytic stability during the conversion of guaiacol, and lower selectivity for the pathway
that might cause the catalyst deactivation. B-Ni-NPs and mF-Ni-NPs were nearly identical in terms
of the size, morphology, and even catalytical performance. The slight difference in catalytical
performance is proposed to be caused by residue on the surface of the mF-Ni-NPs from stabilizer
and/or the fluorine in PTFE tubing.
19
Importantly, the millifluidic synthesis of colloidal Ni-NPs successfully produced catalyst at
increased scale, while maintaining the superior catalytical performance observed in batch prepared
Ni-NP catalysts. These features establish millifluidic synthesis as a reliable and scalable route for
potential industrial applications.
2.3 Continuous Synthesis of Molybdenum Carbide Nanoparticles for
Thermocatalytic CO
2
Hydrogenation*
*This section is a summary of the author’s contribution to work published in J. Am. Chem. Soc. 2020,
142, 1010-1019.
130
Images in this section are taken from this publication accordingly.
2.3.1 Motivation
In recent years, transition metal carbides (TMCs) have garnered significant research attention due to
their excellent catalytic performance for a wide range of transformations including CO 2 reduction,
ammonia synthesis, and hydrodeoxygenation reactions.
131-133
In particular, TMCs have been identified
as potential low-cost candidates for CO 2 reduction because of their multifunctionality that enables H 2
dissociation, C=O bond scission, and reducible-oxide-like behavior, all of which are crucial to the
conversion of CO 2 to CO and higher hydrocarbons.
131
In addition to this multifunctionality (i.e.,
metallic sites, acidic sites, and basic sites ), TMCs have been found to exhibit noble-metal-like behavior,
resist sintering, and offer opportunities to tune catalytic reactivity based on the selection of metal,
stoichiometry, and crystal structure.
123, 134
In this study, my collaborators developed an exceptionally mild solution-phase synthesis of phase-
pure molybdenum carbide (-MoC 1-x
) NPs, a stable and efficient TMC catalyst for thermochemical
CO 2 hydrogenation. I translated the synthesis into a millifluidic system and obtained the catalyst with
higher limit of precursor concentration, improved yield, and shortened reaction time, while crystal
and catalytical properties of the -MoC 1-x
NPs were well maintained. The high efficiency and
20
scalability of the millifluidic synthesis has proved its feasibility for scale-up production of TMC
catalysts.
2.3.2 Millifluidic Synthesis of NP-MoC 1-x
The facile small-scale batch synthesis of NP- MoC 1-x
was developed by my collaborators. This
thermolytic decomposition of Mo(CO) 6 in the presence of oleylamine (OAm) and 1-octadecene at 300
C yields phase-pure -MoC 1-x
NPs over the course of 1 h reaction time, leading to a yield around 40-
50 %. The low throughput due to the small-scale batch process and low yield as a result of the metal
precursor sublimation might limit the industrial application of the NP -MoC 1-x
.
Figure 2.4. Scheme of the millifluidic reactor system for the continuous flow synthesis of NP- MoC1–x. Metal
precursor is delivered into the glass millifluidic reactor in sand bath at 320 C, a check valve at the inlet and a back pressure
regulator at the outlet ensure a 40 psig pressure inside the system, facilitating the control over sublime of the metal
precursor. An analytical balance and a flow meter are incorporated in the system to determine the residence time of the
reaction.
In order to circumvent the issues inherent with the small-scale batch synthesis and enable the high
throughput required for catalytic evaluation of the resulting NPs, we developed a continuous
millifluidic process to execute the thermolytic decomposition of Mo(CO) 6 under conditions similar to
those of the batch synthesis. The continuous millifluidic synthesis of NP-MoC 1-x
was performed using
the reactor configuration illustrated in Figure 2.4. Briefly, a preheated Mo(CO) 6 precursor solution
was injected at a constant flow rate of 12 mL h
–1
into a borosilicate millifluidic reactor (I.D. = 1.8 mm)
21
heated to 320 C. Upon heating to the reaction temperature, the Mo(CO) 6 precursor decomposed,
liberating CO and separating the liquid phase into isolated plugs in the reactor channel. The system
was fitted with a check valve upstream of the heated reactor to prevent the backflow of precursor
solution caused by gas evolution during precursor decomposition. In addition, the millifluidic reactor
was operated at 40 psig to further suppress in situ gas evolution and maintain longer and more
consistent residence times. As the reaction mixture passed through the outlet of the heated segment of
the reactor, the reaction was rapidly thermally quenched as heat dissipated from the narrow diameter
reaction channel upon exposure to ambient temperature air.
Synthetic throughput was optimized by increasing the precursor concentration 8-fold from 78
mM (from batch reaction) to 625 mM with a fixed OAm:Mo ratio of 4, and analysis by my
collaborators with XRD (Figure 2.5) and TEM (Figure 2.6a-d) demonstrated consistent particle
morphology and crystallinity that were analogous to the small-scale batch product.
Figure 2.5. Powder XRD patterns for the -MoC1–x NPs synthesized in flow at four different precursor
concentrations.
20 30 40 50 60 70
2q ()
Intensity (a.u.)
625 mM
312 mM
156 mM
78 mM
MoC
1-x
PDF Card No. 03-065-8092
22
Figure 2.6. TEM images of the NP-MoC1–x synthesized in flow using (a) 625 mM, (b) 312 mM, (c) 156 mM, and (d)
78 mM Mo(CO)6 precursor solutions.
The residence time was estimated by an analytical balance and a flow sensor. The linearity of
product mass measured over time with the analytical balance indicated a steady state condition in the
millifluidic reactor for all measured concentrations (Figure 2.7a). The average density of the product
(𝜌 ) was measured after each run to convert mass flow rate (𝑚 ̇ ) to liquid volume flow rate (Q
L
)
(Equation 2.1). The flow sensor detected the linear flow rates of the liquid and gas plugs (Figure 2.7b).
The ratio between the residence times of each phase measured by the in-line sensor allowed the relative
volumes of the two phases to be calculated. Therefore, the gas volumetric flow rate (Q
G
) was obtained
based on the measured liquid volumetric flow rate. Flow rates of liquid and gas were corrected for
thermal expansion. The residence time ( τ) was determined by dividing measured volume of the reactor
(V) with total volumetric flow rate (Equation 2.2).
𝑄 = 𝑚 ̇ 𝜌 ⁄ (Eq. 2.1)
𝜏 =
𝑉 𝑄 + 𝑄 (Eq. 2.2)
23
Figure 2.7. (a) Mass of product collected over time for 625 mM, 312 mM, 156 mM, and 78 mM Mo(CO)6 precursor
solutions. The linearity indicates that steady state was reached in the millifluidic reactor. (b) Representation of gas and
liquid flow in the millifluidic reactor and the data generated by the dissipation flow sensor to distinguish the liquid and gas
phases and determine the residence time for fluid plugs in the reactor.
Figure 2.8. Reaction residence time and NP-MoC1–x product yield in millifluidic reactor as a function of [Mo(CO)6].
Over the concentration range from 78 mM to 312 mM, it was observed that the residence time
remained nearly constant; however, there was a significant decrease in the residence time (Figure 2.8)
for the highest precursor concentration (625 mM) resulting from increased gas evolution. The NP-
MoC 1–x
product yields as determined by thermogravimetric analysis for Mo(CO) 6 precursor
concentrations of 78 mM, 156 mM, 312 mM, and 625 mM, were 74%, 89%, 96%, and 99%,
24
respectively, demonstrating near quantitative yields of NP-MoC 1-x
using this high-throughput
millifluidic synthesis approach. This represents a significant improvement in throughput over the
small-scale batch reaction, with higher precursor concentrations (625 mM vs. 78 mM), reduced
reaction times (20 min vs. 1 h), and greater yields (99% vs. 40-50%), and translates to a single-channel
reactor throughput of NP-MoC 1–x
of 18.6 g per 24 h of reactor operation, equating to > 450 g of ~4
wt% carbon supported NP-MoC 1–x catalyst.
2.3.3 Catalytical Performance of NP-MoC 1–x
Catalytical performance of the carbon supported NP-MoC 1–x catalyst (NP-MoC 1−x
/C) was accessed
with CO 2 hydrogenation reactions and compared to a traditionally synthesized bulk α-MoC 1−x
catalyst.
The performance of NP-MoC 1−x
/C catalysts represents a 2-fold improvement on a per-site basis as
compared to the bulk α-MoC 1−x
analogue, emphasizing the superior Mo utilization of the
nanostructured catalyst. And increased selectivity toward C 2+ products compared to bulk α-MoC 1−x
suggests the approach to tune the product slate through controlling the nano-structure of TMCs via
the millifluidic method.
2.4 Conclusion
In summary, we demonstrated that the solution phase synthesis of transition metal nanoparticle
catalysts can be achieved using continuous flow millifluidic systems with comparable crystal and
catalytic properties to NPs prepared under analogous batch conditions. The nanoparticles produced
under millifluidic conditions often result at a higher yield than those produced in batch, which may
be attributed to the more rapid and uniform heating profile, as well as superior mixing, when
compared to batch conditions.
With a one-way check valve added in-line at reactor inlet to prevent back flow, and a back-
pressure regulator at reactor outlet maintaining the reactor at a higher pressure, the flow rates as well
25
as the gas evolution were better controlled. Especially when the gas is involved in the synthesis,
increasing the partial pressure of the gaseous reaction species can also facilitate NP fabrication with
faster reaction rate. The higher precursor conversion, more efficient, scalable, and automatable
production with comparable catalytic functionality, make millifluidic synthesis an ideal route to
catalyst scale up.
26
Chapter 3 . Self-Optimizing Parallel Millifluidic Reactor for
Scaling Nanoparticle Synthesis*
* Published in Chem Commun., 2020, DOI: 10.1039/D0CC00064G.
135
3.1 Abstract
Micro- and millifluidic reactors are promising tools for synthesizing colloidal inorganic nanoparticles;
however, the application of these reactors to industrial-scale manufacturing is limited by the volume
of product that can be produced. To adapt millifluidic reactors to larger-scale synthesis, we developed
a high-throughput 16-channel parallel millifluidic reactor with automated feedback control for self-
optimization. This reactor uses a multiphase gas-liquid flow to continuously produce colloidal
CsPbBr 3 quantum dots. The optical properties of the product were monitored, and the Nelder-Mead
simplex algorithm was used to optimize reaction conditions in real time based on the in-situ
photoluminescence characteristics of a model CsPbBr 3 quantum dot system. The automated scaled-
up synthesis was operated continuously for 4 h with a throughput of ~1 L h
–1
using the optimization
algorithm to guarantee the quality and fidelity of the CsPbBr 3 quantum dots throughout the process.
As the first example of high-throughput nanoparticle synthesis in a parallel reactor with self-
optimization via feedback control, this approach could facilitate the fast and precise fabrication of
nanoparticles for optoelectronic applications.
3.2 Motivation
The demand to scale the manufacturing of nanomaterials has been growing, driven by expanding
applications in a multitude of industries, such as quantum dot (QD)-based displays.
136
These
27
nanoparticles are typically produced in small-scale batch processes, which are time- and capital-
intensive to successfully scale volumetrically. Multiphase segmented-flow micro- and millifluidic
reactors (SMRs) are promising platforms for reaction screening processes and the continuous
manufacturing of nanomaterials.
28, 137
SMRs possess several advantages over batch reactors, including
superior heat and mass transfer inherent in the micro- or milli-dimension channels, excellent
reproducibility and control over product quality, reduced environmental health and safety risks, and
the capacity for automating reactions.
112
Unfortunately, the throughput of SMRs in a laboratory
setting is usually several orders of magnitude lower than the demand for industrial production,
hindering the expansion of their application in industry.
138
Moreover, pursuing high-throughput
synthesis in SMRs often results in compromised reaction efficiency and quality. For example,
enlarging the channel size impairs heat and mass transfer,
6
and increasing the flow rate can result in
insufficient residence times, droplet instability, and/or extreme pressure drops in the channels.
Alternatively, increasing the number of channels via parallelization is an appealing approach towards
achieving a high-throughput synthesis in SMRs as it maintains the reactor geometry for stable
production while providing for higher throughput via linear scaling. That is, an n-channel parallel
network leads to an n-fold increase in droplet number and an n-fold increase in throughput.
There have been previous attempts to parallelize droplet generation in microfluidic systems.
139, 140
Specifically, efforts in synthesizing nanoparticles in parallel reactors have been shown to successfully
produce nanoparticles under fixed reaction conditions with a throughput of 10-100 mL h
–1
,
12, 141
which
is still at least an order of magnitude lower than the throughput needed for industrial production.
138
Moreover, these parallel reactors did not adjust reaction conditions in real-time to guarantee product
quality and fidelity, which is necessary for an automated scaled-up process. It is therefore essential to
integrate in-situ process monitoring and feedback control across the parallel network to adjust the
reaction conditions accordingly in real time. To accomplish this, we applied in-situ spectroscopic
detection of product characteristics and chose a “black box” optimization method for feedback control.
28
Black box optimization is faster and more flexible to implement than a more comprehensive modeling
approach, facilitating the straightforward implementation of self-optimization across the channel
network.
142
The reactor described here uses a gas carrier phase to avoid the cost and difficulty of separating
and recycling a liquid carrier. Gas/liquid segmented flows offer heat transfer and mixing efficiency
advantages similar to the more standard liquid-liquid systems that have dominated the literature.
26
Gas phase segmentation has also been proven to remain stable at high temperatures compared to
typical carrier oils, thereby avoiding carrier phase degradation and potential product contamination.
23,
122
However, gas-liquid slug flow in parallelized channels can be challenging, because the
compressibility of the gas phase affects slug dynamics,
143
and the desired plug flow regime requires
well-balanced flow rates of liquid and gas.
26, 144
Here, these shortcomings are addressed by real-time
in-situ monitoring of slug size and frequency coupled with feedback control of product quality as
described above.
We used this SMR to synthesize CsPbBr 3 perovskite QDs as a model system. The tunable band
gap, high photoluminescence (PL) quantum yields, narrow emission linewidths, suppressed blinking,
and high defect tolerance have led to intense interest in these QDs for applications in lasers, LEDs,
and photovoltaics.
145
Previous QD flow reactors have used parametric mapping to obtain optimal
reaction conditions, but they were not parallelized and they delivered less than 100 mL h
–1
throughput.
8, 146
With the guidance of self-optimized feedback control of in-situ PL characteristics, we
demonstrated a successful synthesis of CsPbBr 3 QDs under optimized conditions in a parallel 16-
channel millifluidic reactor yielding a throughput of ~1 L h
–1
.
29
3.3 Results and Discussion
3.3.1 Design of the Reactor
The batch synthesis of CsPbBr 3 QDs was originally achieved by swiftly adding a solution of
tetraoctylammonium bromide (TOAB) dissolved in toluene into a pre-mixed Cs(oleate) and
Pb(oleate) 2 precursor solution in toluene at room temperature with rapid stirring. Under these
conditions, cuboidal CsPbBr 3 QDs are formed within several seconds.
147
This synthesis was adapted
for our reactor such that the Cs
+
/Pb
2+
precursor solution (1:1 mol/mol) in toluene was delivered into
the reactor with nitrogen gas segmentation, and the Br
–
precursor solution in toluene was merged with
the Cs
+
/Pb
2+
precursor slugs (to give a mole ratio of Cs
+
:Pb
2+
: Br
–
near 1:1:3) to allow for mixing,
nucleation, and particle growth while flowing downstream through the millichannels (Figure 3.1).
Figure 3.1. Schematic of the reactor system. The CsPbBr3 perovskite QD synthesis is distributed in 16 parallel
channels and monitored by feedback control based on in-situ measured PL characteristics.
A pressure-driven method for precursor delivery was chosen because it allows for arbitrarily large
reagent reservoirs and is appropriate for large-scale processes. The 3D-printed manifold has one inlet
and four outlets, evenly distributing fluid with a binary tree structure. Liquid or gas was delivered into
one manifold and then distributed into four downstream manifolds, ultimately splitting the flow into
16 channels. To keep comparable flow resistances across the branch channels, identically large flow
resistances in each channel were applied to make resistance variance negligible.
148
Polyether ether
30
ketone (PEEK) tubing with inner diameters below 200 µm were used to create large flow resistances
at each outlet of the secondary manifolds (Figure 3.2). Inlet filters were installed to remove solids.
Gas/liquid slug generation and stream merging took place in PEEK T-junctions and the reaction
occurred in translucent polytetrafluoroethylene (PTFE) tubing (inner diameter = 794 µm).
Figure 3.2. Photograph of the 16-channel parallel millifluidic reactor. Blue arrows indicate the flow streams.
Slug behavior and in-situ PL spectra of the QD product were monitored through this PTFE tubing
(Figure 3.4). An infrared (IR) emitter and receiver were clamped across each PTFE channel to monitor
slug flow (Figure 3.3), with changes in receiver voltage corresponding to the fluid passing by the
emitter/receiver pair and gas having a greater IR transmittance than liquid (Figure 3.4b).
149
31
Figure 3.3. Diagram and photograph of the IR sensor apparatus
Figure 3.4. In-situ detection modules. (a) Photograph of the detection system monitoring slugs and PL spectra. (b)
Slug status was acquired via IR sensors. Diagrams of the UV-vis/PL detection unit in (c) perspective view and (d) side
view.
UV-vis absorbance and PL spectra were collected in-situ and used to evaluate the quality of
CsPbBr 3 QDs. A 405 nm LED and a deuterium-halogen lamp routed to the flow tubing were
alternatively shuttered on and off. Light emitted from the liquid in the tubing upon LED illumination
was then analyzed to produce a PL emission spectrum while the light transmitted upon lamp
illumination was used to produce a UV-vis spectrum (Figure 3.4d). The PTFE tubing was bent in the
detection module by 90 degrees in the x-direction, so that light passes through the radial direction of
32
the tubing, minimizing optical interference from the PTFE (Figure 3.4c, d). The detection module
consisted of four optical channels and was translated through the 16 reactor channels via a linear stage.
Within the detection module, an optical switch connected light signals from one channel to the
spectrometer, while physically blocking the light from the other three channels that are not being read.
The channels were read out serially.
3.3.2 Parallel Slug Behaviors and Photoluminescent Spectra
The slugs formed across the 16 channels were similar in both frequency and size. At driving pressures
of 400 mbar (Cs
+
/Pb
2+
), 400 mbar (gas), and 300 mbar (Br
–
), slug frequencies throughout the 16
channels had a mean value of 11 Hz and a coefficient of variance lower than 22% (Figure 3.5).
Figure 3.5. Box plot (top) and histogram (bottom) displaying the distribution of slug frequencies in 16 channels.
The average slug frequency is 11 Hz and the coefficient of variance is lower than 22%.
Under these fixed driving pressures, 20 slugs were recorded in a single channel and 360 slugs were
recorded in all parallel channels with the IR sensors. Comparing the size distributions, slugs had an
average residence time in the detector around 50 ms with a coefficient of variance ranging from 29%
in a single channel to 35% across the parallel channels. Since the slug size distribution from a single
channel does not significantly differ from that measured across all 16 parallel channels (as evaluated
by a Kolmogorov-Smirnov test, Figure 3.6a), parallelization does not introduce any significant
33
irregularities in slug formation in this parallel network, which is borne out by consistent QD quality
across the parallel network (vide infra).
Figure 3.6. Parallelization of CsPbBr3 QD synthesis in the 16 channels. (a) Box plot and histogram showing size
distributions of slugs recorded across 16 parallel channels (grey, top) and in a single channel (green, bottom) (b)
Normalized offset UV-vis (dotted lines) and PL spectra (solid lines) of the CsPbBr3 QD product in the 16 channels were
highly similar.
Table 3.1. CsPbBr3 QD synthesis was carried out in the parallel reactor (see Section 3.4.4). Example emission peak
wavelength and FWHM of the PL spectrum in each of the 16 parallel channels.
Peak Wavelength λmax (nm) FWHM (nm)
Channel 1 496.2006 34.6923
Channel 2 494.4658 33.6668
Channel 3 495.1598 33.6624
Channel 4 491.6879 34.7460
Channel 5 492.3826 34.0423
Channel 6 493.0772 32.9896
Channel 7 492.7299 33.3433
Channel 8 495.8537 32.2618
Channel 9 496.8941 31.5599
Channel 10 496.2006 32.2618
Channel 11 494.8128 32.9726
Channel 12 494.8128 34.0117
Channel 13 491.6879 35.0912
Channel 14 494.4658 34.7147
Channel 15 496.5474 33.6450
Channel 16 499.6667 31.8798
34
During the continuous CsPbBr 3 QD synthesis, PL emission maxima ( λ max) were typically detected
in the range of 495-520 nm with full widths at half maximum (FWHMs) varying from 25-45 nm,
depending on flow rates and flow rate ratios of precursors and segmenting gas. The emission peak
wavelengths across the 16 channels were typically within a variance of 10 nm and FWHMs were
within 5 nm difference. At driving pressures of 400 mbar for the Cs
+
/Pb
2+
precursor, 300 mbar for the
Br
–
precursor, and 400 mbar for segmenting gas, an example of the UV-vis and PL spectra across 16
channels is presented in Figure 3.6b (Table 3.1).
3.3.3 Self-Optimizing Feedback Loop
To automate CsPbBr 3 QD production with desired quality, a feedback control scheme was established
based on the Nelder-Mead black box optimization algorithm.
150
Our optimization was based on the
PL characteristics of the CsPbBr 3 QDs, with the goal of maintaining the characteristic narrow FWHM
of the CsPbBr 3 QDs, as well as minimizing the FWHM deviation across the 16 channels. To meet the
criteria of Nelder-Mead method, where only one optimization goal is sought, we characterized the
distribution of FWHMs in 16 channels and set the upper endpoint of the 95% confidence interval
(UCI) as the minimizing goal. The input variables are the driving pressures of Cs
+
/Pb
2+
precursors,
Br
–
precursor, and segmenting N 2 gas. The optimization was accomplished with the boundary
conditions being that pressures should not exceed 1 bar (to maintain slug stability), and that driving
pressures for the two precursor solutions should avoid extreme differences in stoichiometry (less than
a 6:1 mole ratio of Br
–
to Cs
+
/Pb
2+
precursor). Within 20 min (allowing for 10 iterations), the
optimization reached convergence around 35 nm (Figure 3.7a, b), and a threshold of 35 nm was set
to terminate the optimization.
35
Figure 3.7. Self-optimization performance. The goal of self-optimization was set to minimize the upper endpoint of
the 95% confidence interval (UCI) in the distribution of FWHMs from the 16 channels. The driving pressures of Cs
+
/Pb
2+
precursors, Br
–
precursor, and N2 segmentation gas are denoted as P[Cs
+
/Pb
2+
], P[Br
–
], and P[gas], respectively. (a) The
search for minimal UCI reached convergence within 8-10 iterations (2 min per iteration). (b) Three variable input pressures
across 11 optimization iterations (labeled 1-11) are plotted with a colormap indicating the corresponding UCI. (c) The
optimization system stabilized the reactor after a flow disturbance was introduced, and the target UCI below 35 nm was
reached again by feedback control despite the disturbance.
The robustness of the self-optimization process was tested in terms of the ability to recover to
optimum conditions following a purposefully introduced disturbance. The flow resistance in the Br
–
precursor stream was manually increased to simulate a situation in which debris is caught by the inlet
filter. As shown in Figure 3.7c, the algorithm initiated a new optimization round when the UCI
exceeded 35 nm, and the UCI returned below 35 nm within 3 iterations.
36
3.3.4 Extended Operation
To further demonstrate the reactor stability, we operated the system for an extended duration. The
throughput varied from 700 mL h
–1
to 1200 mL h
–1
depending on the size of precursor bottles and flow
resistance. In a 4 h reaction, approximately 2.8 L of product was obtained (Figure 3.9a). Based on a
calculated isolated yield of 80% (based on Cs
+
/Pb
2+
), we can obtain a throughput 1.9-3.2 g h
–1
of
CsPbBr 3 QDs in this 16-channel reactor. The slug behaviors and FWHMs in the 16 channels were
maintained throughout the entire process (Figure 3.9c-e). The resulting CsPbBr 3 QD product from the
16 channels was pooled and collected for characterization after each hour. The UV-vis and PL spectra
showed excellent consistency, with narrow PL FWHMs ranging from 21-23 nm (Figure 3.9b). Powder
X-ray diffraction (XRD) analysis from each time point agreed with the expected orthorhombic
structure for CsPbBr 3 (PDF# 00-054-0751), which confirms a consistent crystal structure of the
resulting QDs.
151
The TEM images of the pooled product from 1, 2, 3, and 4 h showed consistent
average sizes of 9.9 ± 1.6 nm, 10.2 ± 1.7 nm, 9.7 ± 1.7 nm, and 9.4 ± 1.5 nm, respectively (Figure 3.8,
Figure 3.9g), which match the QD sizes obtained from PL data.
Figure 3.8. TEM images and inset size distributions of CsPbBr3 QDs from 1, 2, and 3 h time points in the 4-h
automated flow synthesis, with average sizes of 9.8 ± 1.6 nm, 10.2 ± 1.7 nm, and 9.7 ± 1.7 nm, respectively.
37
Figure 3.9. Reactor performance in an extended automated operation. (a) Photos of collected product in a 5 L
bottle from a 4 h synthesis under indoor light (left) and UV light (right). (b) Off-line UV-vis spectra (dotted lines) and PL
spectra (solid lines) of product collected from 1-4 h time points. In-situ (c) slug sizes, (d) slug frequencies, and (e) FWHMs
along the 16 channels are stable throughout the 4 h reaction. The mean values and 95% confidence intervals are plotted in
the graphs. (e) Powder XRD of CsPbBr3 QDs obtained at each hour. (f) TEM of CsPbBr3 QDs at 4 h time point, with size
distribution given as an inset.
38
3.4 Experimental
3.4.1 Precursor Preparation and Product Work-up
Small-Scale (1 h reaction): Cs 2CO 3 (99.9%), PbO (99.9%), and oleic acid (90%) were purchased from
Sigma-Aldrich. Tetraoctylammonium bromide (TOAB) was purchased from Beantown Chemical. In
a typical procedure, a 10 mM Cs
+
/Pb
2+
precursor solution was prepared by dissolving 1.0 mmol of
Cs 2CO 3 and 2.0 mmol PbO in 20 mL of oleic acid under vacuum at 120 °C for 30 min. Upon cooling
the mixture to room temperature, the solution was added to 200 mL of toluene. Separately, a 40 mM
Br
–
precursor solution was prepared by dissolving 6 mmol of TOAB in 12 mL of oleic acid and 150
mL of toluene by stirring at 1,000 rpm at room temperature. The precursor solution bottles were then
assembled for use in the flow reactor, with the PTFE inlet tubing inserted directly into each precursor
bottle.
Large-Scale (4 h reaction): The aforementioned precursor solutions were scaled up 12 in 5-L glass
bottles for the large-scale reaction. Upon preparation of the precursor solutions, the 5-L glass bottles
were assembled for use in the flow reactor. After 1 h of the reaction, a 10 mL aliquot of CsPbBr 3 QDs
was collected and split evenly between two 50-mL centrifuge tubes. The product was precipitated with
4 mL of isopropyl alcohol (IPA) in each tube and briefly vortex mixed, followed by centrifugation
(4,000 rpm, 5 min). The supernatant had a bright yellow tint and was discarded, leaving the
yellow/green solid to be redispersed in exactly 1 mL of hexanes in each tube and briefly vortexed
mixed and bath sonicated. From the 1 mL CsPbBr 3 QD dispersion in the first centrifuge tube, exactly
60 µL was added to a cuvette with 3 mL of hexanes for UV-vis and PL characterization. The remaining
dispersion was used for XRD characterization. The second centrifuge tube with 1 mL CsPbBr 3 QD
dispersion was centrifuged (6,000 rpm, 5 min). The bright yellow/green supernatant was used for
TEM characterization.
39
3.4.2 Product Characterization
Powder X-ray Diffraction (XRD): XRD patterns were acquired on a Rigaku Ultima IV diffractometer
operating at 40 mA and 44 kV with a Cu K α X-ray source ( λ = 1.5406 Å).
Ultraviolet-Visible Spectroscopy (UV-vis): UV-vis absorption data was obtained on a Perkin-Elmer
spectrophotometer. The absorption data was normalized to 452 nm. The cuvette path length is 10 mm.
Photoluminescence Spectroscopy (PL): PL spectra were collected on a Horiba Jobin Yvon Nanolog
spectrofluorimeter, with a photomultiplier tube detector and a 450 W Xe lamp for excitation. All
samples were excited at 405 nm.
Transmission Electron Microscopy (TEM): TEM images were acquired with a JEOL JEM2100F (JEOL
Ltd.) microscope operating at 200 kV. The samples were drop-cast on 400 mesh Cu grids coated with
a lacey carbon film (Ted Pella, Inc.) and dried overnight under vacuum at room temperature. The
average edge lengths of the cuboidal nanoparticles were determined using ImageJ, a pixel-counting
software (N = 300).
Yield Calculation: The yield was calculated based on thermal gravimetric analysis (TGA) of the
CsPbBr 3 product. An isolated yield of 80% is calculated.
CsPbBr
3
Sizes from Photoluminescence Data: The resulting sizes of the pooled CsPbBr 3 QDs from 1, 2,
3, and 4 h were calculated using the PL emission peaks at λ max = 515, 514, 513, and 513 nm,
respectively.
152
From the PL data, the sizes of the QDs from 1, 2, 3, and 4 h were determined to be
11.5, 11.1, 10.8, and 10.8 nm, respectively.
3.4.3 Equipment and Devices in the Millifluidic Reactor
Flow Distribution: 3D-printed manifolds (inner diameter = 1.6 mm, Somos WaterShed XC 11122
photoresin) were custom-manufactured by Protolabs. Small inner diameter (ID) PEEK tubing (0.007-
inch (178 µm) ID 3-inch length, and 0.005-inch (127 µm) ID 9-inch length) and the main channel
40
PTFE tubing (1/16-inch (794 µm) ID 1-foot length) were purchased from McMaster. PEEK T-
junctions with 0.020 inch (508 µm) through hole (P-712) were purchased from IDEX Health & Science.
Inlet filters for the liquid streams were stainless steel mesh (# 100, 0.1-mm hole, 150-µ wires, 30%
open area), and inlet filter for the gas stream was PTFE membrane filter (1.0-µm pore size, Pall
Corporation). Compressed N 2 was used to apply positive pressure and displace liquid precursors, the
nitrogen gas was sent into a multichannel pressure regulator (MFCS
TM
-EZ, Fluigent), and dispensing
pressures were controlled by the regulator.
In-situ Monitoring: The 405 nm LED light source (M405FP1) was purchased from Thorlabs. The
deuterium/halogen light source (DH-2000) and the spectrometer (FlameS) were obtained from Ocean
Optics. The IR sensor (SEN-00241, 940-nm, 75-mW Emitter/Detector kit) was purchased from
SparkFun. 3D-printed detection modules (RenShape SL 7820 photoresin) were custom-manufactured
by Protolabs. SMA-SMA fiber patch cable (M92L01) and 1-to-4 fan-out fiber patch cable bundle
(BF42HS01) were obtain from Thorlabs. The optical switch was fabricated with optical cage systems
from Thorlabs, solenoids (ROB-11015, SparkFun) were integrated in the cage systems to block light
from the channels that are not being read. The linear stage (MOX-06-400) was purchased from Optics
Focus. Arduino boards (Uno R3, Mega 2560 R3) were purchased from SparkFun.
3.4.4 Example Spectra Collection
The example results for parallel slugs and parallel synthesis without optimization (corresponding to
Figure 3.6 and Table 3.1) were recorded with a driving pressure of 400 mbar for the Cs
+
/Pb
2+
precursor,
300 mbar for the Br
-
precursor, and 400 mbar for nitrogen. Slug frequency and slug size (ms/drop)
were processed according to the voltage signals from the IR sensors installed on the 16 channels. UV-
vis and PL spectra were acquired from 16 channels serially.
41
3.4.5 Feedback-Control System
The control system was built with Python 3.6. The integrated code of this feedback-control process is
on GitHub: https://github.com/LuWang04/python-mfr-feedback-control.
Table 3.2. Communication and control approaches for the devices used in the reactor.
Devices Functions Methods to achieve the functions
405 nm LED Turn on/off Arduino digital output
Deuterium-halogen lamp shutter Turn on/off Arduino digital output
Solenoids in optical switches Turn on/off
Arduino digital output with supplementary
circuit
IR sensors
Read IR receiver
voltage
Arduino analogue reading with
supplementary circuit
Linear stage
Transport stage to
designated locations
Driven by a NEMA-23 stepper motor driver;
Arduino digital signals were sent to the driver
Arduino boards
Send digital signals and
read analogue signals
Communicated with a python interface
pyFirmata
https://pypi.org/project/pyFirmata/
Fluigent MFCS
TM
-EZ regulator Set and read pressures
The equipment communicates with PC
through a USB cable, Python package is
available in MFCS series Software
Development Kit (SDK)
FlameS spectrometer Read spectra
The equipment communicates with PC
through a USB cable, Python package was
obtained from https://github.com/ap--
/python-seabreeze
The self-optimization procedure is illustrated in Figure 3.10a. The initial pressure sets were
defined in the Nelder-Mead.py file. The PL and UV-vis spectra were processed in real time with
Python. Emission peak wavelengths and full widths at half maximum (FWHMs) of the PL spectra
were used in feedback control. The upper endpoint of 95% confidence interval (UCI) of the 16
FWHMs was minimized with the Nelder-Mead simplex method, and the optimization stopped when
UCI < 35 nm. The driving pressure for either liquid precursor (P
liq
) was ensured to be larger than 400
mbar to maintain a high throughput.
The procedure of collecting spectral information through the 16 channels is presented in Figure
3.10b. The slug feature data were collected from IR sensors simultaneously through multithreading.
Voltages across 16 IR receivers were selected by a multiplexer in sequence and read by Arduino
analogue pin. IR receivers were recorded with a time interval of 1 ms. Slug frequencies were acquired
42
and the time lapse for each liquid slug passing the sensor was regarded as slug size in terms of time
(ms/drop). IR signals were processed off-controller in real time with Python.
Figure 3.10. Flowcharts of the feedback-control process. (a) Self-optimization procedure in the parallel millifluidic
reactor. UCI denotes the upper endpoint of 95% confidence interval of FWHMs in 16 channels, and P liq corresponds to the
driving pressure of either liquid precursor. (b) The procedure of collecting UV-vis and PL spectra through 16 channels. The
spectra were read through channel = 0 to channel = 15, i denotes the location of the linear stage and j denotes the channel to
be read in the 4-channel detection module.
3.5 Conclusions
In summary, we have successfully developed a parallelized 16-channel automated reactor for the
scaled-up synthesis of colloidal CsPbBr 3 QDs. The reactor self-optimizes for product quality based on
in situ monitoring of the optical properties of QDs as they are synthesized. This represents the first
demonstration of high-throughput nanomaterial synthesis in a feedback-controlled continuous flow
parallel reactor. With excellent chemical resistance, highly parallel slug generation, robust
optimization algorithms, and a high throughput of approximately 1 L h
–1
, this reactor can be employed
broadly towards the scaled-up production of photoluminescent QDs.
43
Chapter 4 . Interactions between Charged Nanoparticles and
Giant Vesicles Fabricated from Inverted-Headgroup Lipids*
*Published in J. Phys. D: Appl. Phys. 2017, 50, 415402-415407.
153
4.1 Abstract
The surface chemistry of the cell membrane plays an important role in how cells interact with
particulate species. These interactions are dictated in large part by lipid headgroup charge. To
investigate the nature of electrostatic interactions between lipid bilayers and nanoparticles in solution,
we studied nanoparticles interacting with the zwitterionic lipid 1,2-dioleoyl-glycero-3-phosphocholine
(DOPC), and its inverted-headgroup analog DOCP. These interactions were investigated by
fabricating giant unilamellar vesicles (GUVs) with DOPC lipids and DOCP lipids respectively, and
introducing nanoparticles to suspensions of both. GUVs displayed various deformational modes
depending on the charge and size of the nanoparticles as well as the compositions of the GUVs. The
differences in the responses of the two lipid species illuminate how the phosphate and choline groups
on the lipid interact with charged nanoparticles. This study suggests that the phosphate group
dominates the lipid-nanoparticle electrostatic interaction. We speculate that the formation of water
clathrate structures around the choline group inhibits interactions between negatively charged
nanoparticles and the positively charged choline.
4.2 Motivation
Zwitterionic phosphocholine lipids are predominant in most biological membranes; it’s commonly
believed phosphocholine lipids bear no net charge as the positive and negative charges cancel out.
44
However, it has been demonstrated by several studies that positively charged particle species interact
electrostatically with zwitterionic phosphocholine liposomes and adsorb on their surface.
93, 154
And
negatively charged nanoparticles have been found with the ability to stabilize zwitterionic
liposomes.
155
To better understand this phenomenon, we here investigate the role that headgroup orientation
plays in this electrostatic interaction. We studied DOPC [1,2-dioleoyl-sn-glycero-3-phosphocholine]
and its inverted-headgroup analog DOCP [2-((2,3-bis(oleoyloxy)propyl)-dimethylammonio) ethyl
hydrogen phosphate] (Figure 4.1), which has shown potential in drug encapsulation and delivery.
156-
158
Both lipids share identical tailgroups, but the positions of the cationic choline and the anionic
phosphate in CP headgroups are inverted relative to their positions in PC lipids, extending the anionic
phosphate into aqueous interfacial region.
159, 160
Figure 4.1. DOPC lipids and DOCP lipids share identical tailgroups while their headgroups have inverted
orientation of charge distribution. GUVs are formed with DOPC lipids and DOCP lipids respectively, they are put in the
presence of positively charged NH3
+
-PNPs (left panel, orange) or negatively charged COO
-
-PNPs (right panel, blue).
Differences are observed between the behavior of DOPC GUVs and DOCP GUVs. The sizes of GUVs and PNPs are not
to scale.
By introducing charged polystyrene nanoparticles with varied sizes to aqueous suspensions of
unilamellar lipid vesicles, vesicle deformation and size changes have been observed. A comparison in
the responses of PC lipids versus CP lipids helps to clarify the electrostatic contribution from either
charge in the zwitterionic headgroups, illuminating the mechanism of nanoparticle-membrane
interplay.
45
4.3 Result and Discussion
4.3.1 Influence of Nanoparticle Charge on Particle-Bilayer Interactions
Positively charged 500 nm NH 3
+
-PNPs and negatively charged 500 nm COO
-
-PNPs were added to
the GUV sample at a concentration of 4×10
10
#/mL, corresponding to a total nanoparticle surface
area of 3.3×10
10
µm
2
/mL. The experiment with positively charged NH 3
+
-PNPs showed significant
deformation of DOPC GUVs and complete rupture of most DOCP GUVs. In the experiment with
negatively charged COO
-
-PNPs, neither the DOPC GUVs nor the DOCP GUVs underwent
significant deformation, as shown in Figure 4.2a.
We quantified this shape change to measure the extent of the nanoparticle-bilayer interaction.
The contour of GUVs was traced with ImageJ and the circularity of GUVs was determined according
to Equation 4.1:
Circularity = 4π×
Area
Perimeter
2
(Eq. 4.1)
GUVs with perfect circular shape would have the maximum circularity at 1, and as GUVs deform
from the perfect circular shape, the circularity would decrease, reporting on the extent of deformation.
The distributions of circularity were then compared before and after each PNPs experiment using
Wilcoxon signed rank test in JMP (SAS Institute Inc. NC).
The decrease in circularity of DOPC GUVs was more significant with positively charged PNPs
than with negatively charged PNPs (Figure 4.2, Table 4.1). This is consistent with a relatively strong
electrostatic interaction between positively charged nanoparticles and zwitterionic membranes made
from physiological lipids in previous studies.
93, 154
The interaction between DOCP GUVs and
positively charged PNPs was more vigorous as they lost their membrane integrity instantly. The
ruptured DOCP lipid patches presented amorphous shapes and significantly low circularity in the
circularity measurement. Hence the distribution of circularity shifted further compared with DOPC
46
GUVs (Figure 4.2b and 4.2c). On the other hand, circularity of DOCP GUVs was not significantly
changed with negatively charged PNPs. These changes are independent of GUV size (Figure 4.3,
Table 4.2).
Figure 4.2. The fluorescent microscopic image of DOPC GUVs (red) and DOCP GUVs (green) 10 min after
introducing no PNPs, positively charged PNPs or negatively charged PNPs (a). Circularity distribution of DOPC GUVs
(b) and DOCP GUVs (c) before and after corresponding experiments. Circularity distribution is presented as a box plot.
The central box spans interquartile range, with the segment inside indicating the median of the data, and the fences above
and below the box representing the maximum and minimum circularity values. *p<0.005, **p<0.001, ***p<0.0001 in
Wilcoxon signed rank test. Scale bar: 20 µm.
Table 4.1. Wilcoxon rank-sum test results for comparing circularity before and after GUVs interacting with no
PNPs, positively charged PNPs, and negatively charged PNPs. Z is the Z-ratio for the test.
Level - Level
DOPC GUVs
DOCP GUVs
Z p-Value Z p-Value
After (No PNPs) Before (No PNPs) 0.6273 0.5305 -0.65355 0.5134
After (NH 3
+
-PNPs) Before (NH 3
+
-PNPs) -16.6379 <.0001 -9.62266 <.0001
After (COO
-
-PNPs) Before (COO
-
-PNPs) -3.1839 0.0015 -1.39829 0.1620
47
Figure 4.3. Circularity distribution of DOPC and DOCP GUVs in experiments with 500 nm NH3
+
-PNPs (a, c) and
with 500 nm COO
-
-PNPs (b, d) categorized by GUVs sizes. GUVs are categorized in 6-12 m radius (small GUVs), 12-18
m radius (medium GUVs), and >18 m radius (large GUVs).
Table 4.2. Wilcoxon rank-sum test data of GUVs circularity distribution after experiments with 500 nm
positively/negatively charged PNPs, compared with the distribution before each experiment. Z is the Z-ratio for the test.
PNPs Type GUVs Radius
DOPC GUVs
DOCP GUVs
Z p-Value Z p-Value
Positively Charged
6-12 µm -12.8558 <.0001 -7.4090 <.0001
12-18 µm -5.8652 <.0001 -3.9237 <.0001
>18 µm* 0.1132 0.9099 -0.8874 0.3749
Negatively Charged
6-12 µm -2.3359 0.0195 1.3446 0.1788
12-18 µm 0.5705 0.5684 0.4037 0.6864
>18 µm* 0.7136 0.4755 0.0000 1.0000
*For GUVs in the size range of 6-18 m radius, the p-values agree well with p-values of all-size-range, the reason for p-
value disagreement in GUVs larger than 18 m is that small sample size (10-40 GUVs) lead to reduced power of Wilcoxon
rank-sum test.
161
Given the fact that both types of lipids interacted more strongly with positively charged
nanoparticles and the fact that the DOCP lipids interact more vigorously, it is apparent that the
phosphate group dominates the electrostatic interaction between lipid bilayers and charged
nanoparticles. Since both phosphate and choline groups carry formal charges, we must consider the
underlying mechanism for the dominance of the phosphate-nanoparticle interaction. According to
48
molecular simulation studies, water molecules associate with
phosphatidylcholine through hydrogen bonding with the
phosphate group and clathrate hydrate formation around the
choline (Figure 4.4).
162
The hydrogen bond to phosphate oxygen has a distance of
3.25 Å from phosphate oxygen to the oxygen atom of water
molecule, while the clathrate hydrate surrounding the choline
group has a distance of 4.75 Å from methyl group to the
oxygen atom of water molecules.
164
And by the nature of the
clathrate hydrate, that oxygens and hydrogens are
approximately on the same spherical surface of the water shell,
the choline nitrogen is not capable of forming effective
hydrogen bonds with bulk water molecules.
162
Since the
choline group takes more space and has more unfavorable interaction for hydration access than
phosphate group, it has a weaker contribution than phosphate group in the electrostatic interaction.
When this bulky choline group is further buried in bilayer as in DOCP lipids, the steric effect for
negatively charged PNPs approaching choline core becomes larger, while the phosphate group obtains
more mobility, facilitating electrostatic interactions with nanoparticles in solution. This provides a
mechanism by which positively charged nanoparticles interact more vigorously with the lipid bilayer
of the plasma membrane than negatively charged nanoparticles, explaining observations that cationic
nanoparticles are both more toxic and more likely to be taken up by cells than their anionic
counterparts.
165, 166
Note that the phosphate group in DOCP lipids has two free hydroxyl groups and
that while one of these groups certainly carries a formal charge, the second pK a in this phosphate
group may also be lower than the pH of PBS buffer.
159
Therefore, the second hydroxyl group might be
partially ionized, contributing partly to the enhancement of the electrostatic interaction.
Figure 4.4. Structured water around
choline and phosphate group. A water
molecule anchoring clathrate around choline
(top) moiety and a phosphate (bottom)
oxygen atom (intramolecular water anchor).
Carbon atoms are represented in dark green,
oxygen atoms in red, hydrogen atoms in
white, nitrogen atoms in blue and
phosphorous atoms in light green in this
molecular model. Image taken from A.
Liwo. (Springer Science & Business Media,
2013).
163
49
4.3.2 Influence of Nanoparticle Size on Particle-Bilayer Interaction
We also examined the influence of nanoparticle size on lipid bilayer-nanoparticle interactions. The
dependence on particle size may vary greatly from one system to another.
167
For J774 A1 macrophages,
gold nanoparticles present higher toxicity with larger particle size, while silver nanoparticles show no
size-dependence of toxicity within the same size range of 2-40 nm.
168
Nevertheless, in another study
of the silver nanoparticle toxicity on the crustacean Daphnia magna and the murine fibroblast cell line
Balb/3T3, the toxicity increases with decreasing particle size in the range of 10-80 nm.
169
These
differences of size dependence sometimes indicate different mechanisms of the interaction and
cytotoxicity. So we varied the size of charged PNPs among three different diameters: 750 nm, 500 nm,
and 50 nm. As this electrostatic interaction is a surface phenomenon, we kept the total surface area of
PNPs in the GUVs suspension constant at 3.3×10
10
µm
2
/mL.
Figure 4.5. Fluorescent microscopy images of DOPC GUVs (red) and DOCP GUVs (green) 10 min after addition
of positively charged PNPs at three different sizes (a, b, c), and 10 min after addition of negatively charged PNPs at three
different sizes (d, e, f) respectively. Scale bar: 50 µm.
50
In the experiments with positively charged NH 3
+
-PNPs of three different sizes, DOPC GUVs
deformed with all the sizes, while DOCP GUVs were ruptured with 500 and 50 nm ones, but remained
unchanged with 750 nm NH 3
+
-PNPs. Representative images showing these results are presented in
Figure 4.5; they are quantified with a circularity distribution in Figure 4.6. According to the Wilcoxon
signed rank test in Table 4.3, the 500 nm NH 3
+
-PNPs were most destructive for both DOPC GUVs
and DOCP GUVs. As for negatively charged COO
-
-PNPs, DOPC GUVs presented slight deformation
with all the sizes, and 50 nm diameter PNPs were most destructive. On the other hand, DOCP GUVs
remained not reactive with COO
-
-PNPs throughout all the sizes, as is shown in Table 4.3 and Figure
4.6, indicating the dominant role of the phosphate group despite the size change of PNPs. In general,
positively charged PNPs underwent a stronger interaction with both types of GUVs.
Figure 4.6. Circularity distribution of DOPC and DOCP GUVs in experiments with positively charged NH3
+
-PNPs
(a, c) and with negatively charged COO
-
-PNPs (b, d) at PNPs diameters of 750 nm, 500 nm, and 50 nm. *p<0.005,
**p<0.001, ***p<0.0001 in Wilcoxon signed rank test.
51
Table 4.3. Wilcoxon signed rank test data of GUVs circularity distribution after experiments with
positively/negatively charged PNPs at three different sizes, compared with the distribution before each experiment. Z is the
Z-ratio for the test.
PNPs Type PNPs Size
DOPC GUVs
DOCP GUVs
Z p-Value Z p-Value
Positively Charged
750 nm -5.5252 <.0001 -0.24349 0.8076
500 nm -16.6379 <.0001 -9.62266 <.0001
50 nm -7.4288 <.0001 -9.34566 <.0001
Negatively Charged
750 nm -2.87318 0.0041 -0.58084 0.5614
500 nm -3.18392 0.0015 -1.39829 0.1620
50 nm -3.61245 0.0003 -0.47625 0.6339
But size dependence of this interaction was not strong, as the results of Wilcoxon signed rank test
vary little in Table 4.3. This trend was also observed in experiments maintaining same volume fraction
of nanoparticles (Figure 4.7). So in our system, both GUVs were more vulnerable with smaller charged
PNPs, but PNP surface charge was more crucial than PNP size in the interaction with either DOPC
or DOCP GUVs. This relatively weak dependence on size indicates that, at least in the size range
studied, a size-independent nanoparticle-bilayer mechanism interaction dominates. In contrast to
previous studies
170-172
on membrane fusion on nanoparticles or translocation of nanoparticles,
membrane wrapping does not apply to our study as this mechanism has a strong dependence on the
particle size. With increasing size of nanoparticles, the membrane wrapping and deformation becomes
more favorable
173
because the opposing bending energy is independent of particle size, while favoring
adhesion energy increases with size of nanoparticles. Therefore, the mechanism might be steric and
electrostatic repulsion between adsorbed nanoparticle, resulting in enhanced membrane curvature to
create more space between the adsorbed nanoparticles.
103, 174
52
Figure 4.7. Fluorescence images of DOPC GUVs (red) and DOCP GUVs (green) 10 min after the addition of
positively charged nanoparticles (a,b) or negatively charged nanoparticles (c,d) of 500 nm and 50 nm diameters. The mass
concentration of nanoparticles was kept consistently 0.9 mg/mL. Scale bars 50 m.
Another interesting behavior we noticed is that DOCP GUVs only ruptured in the presence 500
nm and 50 nm NH 3
+
-PNPs, and retained a relatively circular shape for all the other experiments
(Figure 4.5). No elastic local bending of DOCP bilayers
175
was observed in any of the experiments.
Our hypothesis for the mechanism is that the bulky choline group is located adjacent to tailgroups in
DOCP lipids, thus stronger repulsive forces between neighboring clathrates lead to less headgroup
mobility compared with DOPC.
176
In membrane bending, lipids are locally crowding at the inner
leaflet to achieve higher curvature. Once the electrostatic attraction repulsion between PNPs on the
outer leaflet overcomes bending rigidity and membranes start to form protrusions, DOCP lipid bilayer
would fail to keep its integrity due to the rigidity of headgroups. To further verify our conclusion, more
information about the packing parameter of DOCP lipids and bending modulus of the bilayer
membranes formed from DOCP must be investigated.
53
4.4 Experimental
4.4.1 Chemicals
1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) lipids, 2-((2,3-bis(oleoyloxy)propyl) -
dimethylammonio)ethyl hydrogen phosphate (DOCP) lipids, and 1,2-dipalmitoyl-sn-glycero-3-
phosphoethanolamine-N-(cap biotinyl) (Biotin-DPPE) lipids were purchased from Avanti Polar
Lipids (Alabaster, AL). Atto-488-1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine (Atto-488-
DPPE, λ ex 507 nm, λ em 527 nm), Atto-550-1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine (Atto-
550-DPPE, λ ex 556 nm, λ em 578 nm), agarose (ultralow gelling temperature), phosphate-buffered saline,
sucrose, and D-(+)-glucose were obtained from Sigma-Aldrich (St. Louis, MO). Avidin (egg white)
was from Thermo Fisher Scientific (Waltham, MA). Biotin-PEG-silane (M w = 3400) was from Laysan
Bio (Arab, AL). Poly(dimethylsiloxane) (PDMS) was from Dow Chemical (Midland, MI). Amine
functionalized polystyrene nanoparticles (NH 2-PNPs) (Diameters = 50 nm, 500 nm, 750 nm) and
carboxyl functionalized polystyrene nanoparticles (COOH-PNPs) (Diameters = 50 nm, 500 nm, 750
nm) were from Polysciences (Warrington, PA). All buffers and solutions in this work were prepared
with 18.2 M Ω•cm deionized water (Millipore).
4.4.2 GUV Formation
GUVs were formed from DOPC lipids and DOCP lipids respectively using the agarose hydration
method.
81
2 mg/mL DOCP solution was prepared with 0.8 mol% Atto 550 DPPE in chloroform. 1.5
mg/mL DOCP solution was prepared with 0.8 mol% Atto 488 DPPE in chloroform. Both solutions
had 0.2 mol% biotin-DPPE incorporated. 10 µL DOCP solution was added on agarose-coated #1
glass coverslip dropwise, and each drop was spread with gentle air to form a thin layer lipid film.
When the lipid film was dried, 400 µL 200 mM sucrose in 15 mM PBS solution was added on top to
rehydrate the DOCP lipid film. After 1 h of hydration, the DOCP GUVs were formed and the
suspension was transferred into 600 µL 200 mM glucose in 15 mM PBS solution to let GUVs deposit
and enrich at bottom. DOPC GUVs were formed and concentrated with the same procedure.
54
4.4.3 Observation Chamber Preparation
PDMS film with 1 cm height was cured from resin and holes (0.5 cm diameter) were punched through
the film. Then the PDMS film was bound to #1 glass coverslip via plasma treatment. The well-shape
observation chambers usually hold maximum 50 µL corresponding to the size of the hole. Observation
chamber were filled with 10 mg/mL biotin-PEG-silane solution in ethanol-water 9:1 mixture for 1h
then washed with water. Silane binding was cured in oven (60 °C) subsequently, so that biotin-PEG
was bound to the bottom of the chamber. Before each experiment, the chambers were rinsed with 2
mg/mL BSA solution then 1 mg/mL avidin solution and excessive BSA or avidin was washed with
water.
4.4.4 Microscopy
15 µL of DOPC GUV suspension and 15 µL of DOCP GUV suspension were collected from the
bottom of enriched GUV suspension, and were put into the pretreated PDMS observation chamber.
The chamber was then left sitting for 20 min to ensure biotin in the bilayer was anchored to avidin at
the bottom of the chamber. When GUVs were settled at the bottom, 4-by-4 tile images were taken by
Zeiss microscope with definite focus strategy. Then 3 µL nanoparticle solution was added gently to
the GUV suspension. The concentrations of nanoparticles were varied as to keep the surface area of
charged PNPs in GUs suspension at 3.3 10
10
µm
2
/mL. Another set of 4-by-4 tile images was taken
after 10 min of interaction. Each tiles image yielded a DOCP GUV count of around 200 and a DOPC
GUV count of around 1000 (yield was better for DOPC GUVs).
4.4.5 Data Analysis
The tile images were processed in MATLAB with white top-hat transformation to obtain even
background,
177
and the 4-by-4 tiles were stitched together as one image. The image was then
transformed to binary and circularity of the GUVs was detected by particle analyzer tool in ImageJ.
55
4.5 Conclusions
We have used two lipid species with inverted headgroups as a probe in the study of the interaction
between nanoparticles and lipid bilayers. Through the different interactions with charged polystyrene
nanoparticles, we understand the contribution of each group in headgroups. And the study on the
particle size dependence further lead us to plausible mechanism of this nanoparticle-bilayer interaction.
The inversion of phosphatidylcholine in the DOPC headgroup results in a series of differences in
properties including molecular geometry, magnitude of the lipid bilayer-nanoparticle interaction, and
even bending dynamics of the bilayer. The clathrate hydrate around choline group weakens its role in
electrostatic interactions, therefore the phosphate group dominates in the electrostatic interaction
between zwitterionic lipid bilayers and charged nanoparticles. In DOCP membranes, higher activity
of the phosphate group in the electrostatic interaction arises from the deeper insertion of bulky choline
group into lipid bilayers. Furthermore, the change in the position of choline group in DOCP also
induces larger surface area occupation per lipid molecule and more rigidity for membrane bending.
Meanwhile, the similarity in the weak particle size dependence of this electrostatic interaction suggests
that the mechanism is common between the two lipid types. This work provides fundamental insights
regarding the nature of the interaction between nanoparticles and phosphatidylcholine lipids, which
has implications for understanding nonspecific interactions of cell membrane and nanoparticles,
understanding toxicity of nanomaterials in the environment as well as the development of new drug
delivery vehicles.
56
Chapter 5 . Effect of Protein Corona on Nanoparticle-Plasma
Membrane and Nanoparticle-Biomimetic Membrane Interactions*
* Published in Environmental Science: Nano, 2020, DOI: 10.1039/D0EN00035C.
178
5.1 Abstract
Nanomaterial contamination in the environment poses severe threats to public health and wellness.
Understanding interactions between nanoparticles and biomembranes is pivotal to understanding the
physiological effects of nanomaterials. The prevailing understanding is that a protein corona forms
around nanoparticles upon their entering biological systems. The effect of the protein corona on the
membrane-nanoparticle interaction has not been comprehensively investigated. Here, we report a
systematic study to better understand the effects of the protein corona on nanoparticle-biomembrane
interactions with both plasma membranes (293T cell line) and biomimetic membranes. Giant plasma
membrane vesicles (GPMVs) and giant unilamellar vesicles (GUVs) fabricated from organ lipid
extracts (brain, heart, and liver) served as biomimetic models in our study. Reduced charged-
nanoparticle adhesion to both plasma and biomimetic membranes with the presence of the protein
corona suggests that the protein corona interferes with the electrostatic interaction between
nanoparticles and biomembranes. These similar trends of nanoparticle adhesion among the
membranes indicated that model membranes can capture this electrostatic interaction with similar
responses as plasma membranes. However, the membrane integrity subsequent to the interaction was
different between the two systems, indicating the limitations of model membranes in recreating the
complexity and dynamics of plasma membranes. As the first systematic study correlating nanoparticle
interactions with cell membranes, isolated cell membranes, and synthetic vesicles from natural lipid
57
extracts, we demonstrated that biomimetic membranes can serve as excellent analogues to cell
membranes in providing fundamental insights regarding the electrostatic interaction between
nanoparticles and biomembranes.
5.2 Motivation
Nanomaterial contamination in the environment is present in forms such as ultrafine soot and
nanoplastics. Ultrafine soot with adverse respiratory health effects is heavily emitted from diesel and
gasoline exhaust. The hazardous accumulation of nanoplastics in aquatic and terrestrial environment
originates not only from production and usage lifecycle degradation of extensively used plastics, but
also from the fragmentation in landfills. Engineered nanomaterials, with applications in biosensors,
bioimaging and drug delivery, are also becoming increasingly deployed. Given the increasing presence
of nanomaterials in day-to-day experience, the potential hazards posed to biological systems by
nanoparticles have become a notable concern. The interplay between nanoparticles and cells can lead
to cellular accumulation of nanoparticles, compromised plasma membrane integrity, as well as
mitochondrial and lysosomal damage.
53, 179
These potentially cytotoxic effects are determined by
nanoparticle characteristics such as size, shape, charge, and surface chemistry.
180
Nanoparticles adsorb proteins and other biomolecules upon entering biological fluids due to their
high surface energy. The associated proteins, called the protein corona, modify the surface properties
of the nanoparticles, providing them with biological properties distinct from those they would have in
their pristine state, thereby altering the fate of nanoparticles in biological systems.
57, 181
There remain
many open questions regarding the role of the protein corona.
182
It has been generally believed that
the protein corona protects cells against reactive surfaces of nanoparticles and increases the safety of
nanomedicines.
183, 184
But Obst and her colleagues found that the protein corona does not significantly
decrease cellular uptake of nanoparticles into macrophages.
185
In some cases, the protein corona can
even activate surface receptors and lead to undesired immune responses.
186, 187
It is well known that
58
the initial step in cellular uptake of nanoparticles is dominated by interfacial interactions between the
plasma membrane and nanoparticles, and the cytotoxicity of nanoparticles has been related to this
interaction.
188, 189
Therefore, a systematic knowledge of the nanoparticle-plasma membrane interaction
is the key to understanding this nano-bio interfacial phenomena and the impact of the protein corona
on cells.
To date, the interpretation of nanoparticle-membrane behaviors in in vitro experiments is still not
well established due to the complex and dynamic nature of cell membranes. Simplified biological
model membranes are advantageous to perform focused studies and systematic investigations of the
nanoparticle-biomembrane interface.
64
Giant unilamellar vesicles (GUVs) and giant plasma
membrane vesicles (GPMVs) are representative free-standing model biomembranes: they are bottom-
up and top-down approaches for mimicking plasma membranes, respectively. GUVs are fabricated
from tunable lipid ingredients and can be designed to present representative lipid compositions in
plasma membranes.
190, 191
GPMVs are blebs isolated from cells that have a composition similar or
identical to that of the plasma membrane; they largely preserve the plasma membrane’s physical
properties while being free from contamination of organelle membranes.
83, 192
These simple and stable
model membranes have shown clear similarities with in vitro studies in terms of non-specific
interactions with nanoparticles,
193
particularly not only validating pathways of nanoparticle
internalization but also strongly correlating membrane distortion with cell viability.
87, 194
Our previous
studies utilizing GUVs have further unveiled the an adhesion-based mechanism contributing to
toxicity of charged nanoparticles.
93, 153
It is important to correlate cellular process with biophysical phenomena to identify general
mechanisms underlying the cellular process.
63
Here we report a comprehensive attempt to investigate
the impact of the protein corona on non-specific interactions between charged nanoparticles and
plasma membranes by establishing a correlation between plasma membranes and biomimetic
membranes. We selected representative nanoplastic polystyrene nanoparticles (PNPs) for our study.
59
In addition to examining charged PNP interactions with cell surfaces, we used GUVs fabricated from
natural lipid extracts and GPMVs from 293T cells. We observed interactions between the membranes
and PNPs with and without a protein corona, and further compared PNP adsorption to membranes
as well as membrane integrity upon PNP interaction. Through this study, we have confirmed the
general protective effect of protein corona in non-specific electrostatic nanoparticle-biomembrane
interactions. This systematic study also suggests that model membranes are reliable platforms to
explore the nano-bio interface, providing fundamental information for nanomaterial design in clinical
and environmental applications.
5.3 Results and Discussion
5.3.1 Protein Corona Characterization
Electrostatics can play fundamental roles in nanoparticle-biomembrane interactions and in the fate of
nanoparticles in biological systems.
93, 153, 179
Polystyrene nanoparticles (PNPs) at 100 nm diameter with
native surface sulfate groups (sulfate-PNPs), with surface functionalization of negatively charged
carboxyl groups (carboxyl-PNPs), and positively charged amine groups (amine-PNPs) were selected
in this study. PNPs were all labeled with an encapsulated green fluorophore (491 nm excitation, 509
nm emission). After treatment with human serum at 37 °C for 30 min with gentle shaking, protein
coronas were formed on these three types of PNPs. The protein corona composition with this
prolonged incubation time should be more equilibrated compared to the rapid corona formation at
early exposure.
184
The hydrodynamic diameters of PNPs after treatment with human serum increased
(Table 5.1 and Figure 5.1), indicating the existence of proteins bound to the surfaces of the PNPs.
Sulfate-PNPs and positively charge amine-PNPs with protein corona displayed upward shifts (~20
nm) in mean sizes. Negatively charged carboxyl-PNPs had a broadened size distribution and the
largest increase of mean size. This might be attributed to aggregation of the particles, but since the size
distribution was below 600 nm, the aggregation clusters might only consist of a few particles. Zeta
60
potential measurement showed the expected negative surface charges for the sulfate-PNPs and
carboxyl-PNPs, as well as positive surface charge of the amine-PNPs. After incubation with human
serum, the surface zeta potential for all three of PNPs became close to the value for human serum.
Proteins from human serum not only covered the surface of PNPs, but also altered charge properties
of the PNPs.
Figure 5.1. Size distribution of polystyrene nanoparticles with and without the presence of a protein corona.
Sulfate-PNPs, carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP.
Table 5.1. Properties of the polystyrene nanoparticles with and without protein corona. All measurements were
performed in PBS buffer; protein mass was determined via BCA assay.
Hydrodynamic diameter ± s.d. (nm)
Zeta potential ± s.d. (mV)
mg protein/
mg PNPs
No corona With corona No corona With corona
SPNP* 94.27 ± 2.66 118.50 ± 10.4
-45.22 ± 4.89 -9.56 ± 3.00 0.212
CPNP* 87.53 ± 2.8 198.21 ± 71.68
-37.99 ± 4.42 -10.16 ± 2.82 0.575
APNP* 101.44 ± 14.94 122.29 ± 11.66
20.31 ± 3.29 -9.50 ± 3.48 0.096
*SPNP: 100 nm sulfate-functionalized polystyrene nanoparticles; CPNP: 100 nm carboxyl-functionalized polystyrene
nanoparticles; APNP: 100 nm amine-functionalized polystyrene nanoparticles.
The characterization of the protein corona was achieved by elution, quantification, separation,
and identification. Among the three types of PNPs, carboxyl-PNPs were eluted with the most
abundant proteins, which echoed the change in carboxyl-PNP size distribution with protein corona:
the broadened distribution and relatively extreme size increase might be caused by large amounts of
protein bound to carboxyl-PNPs. According to SDS-PAGE (Figure 5.2a), eluted protein coronas from
all types of particles shared a strong band at around 25 kDa, and major differences were observed in
the range of 50-100 kDa and below 17 kDa.
61
Figure 5.2. Comparison of protein corona composition on sulfate-, carboxyl-, and amine-polystyrene nanoparticles
(SPNPs, CPNPs, and APNPs respectively). (a) Coomassie blue-stained SDS-PAGE gel of human plasma proteins
obtained from corona on SPNPs, CPNPs, and APNPs. (b) LC-MS-MS result of proteins identified in the corona formed
on SPNPs, CPNPs, and APNPs. This Venn diagram reports the number of unique proteins identified from each of three
nanoparticles as well as proteins common to two or all three nanoparticle populations. (c) Classification of corona proteins
identified by LC-MS-MS according to their calculated isoelectric point (pI); relative percentages are shown.
For a better understanding of protein identities in the coronas, proteomic analysis of eluted
protein coronas was carried out with LC-MS-MS. Proteins identified with at least two peptides are
listed in Table 5.2. We found out that proteins existed in coronas of all three types of PNPs were
mainly apolipoproteins with molecular weights corresponding to the bands around 25 kDa on SDS-
PAGE. Apolipoproteins have been discovered previously as a major group of proteins in the corona
formed around nanoparticles of different materials upon contact with plasma.
195, 196
Apolipoproteins
are involved in the transportation of lipids and cholesterol in the bloodstream, thereby they could
greatly affect the intracellular trafficking and fate of nanoparticles in biological environments.
197
Highly abundant human serum albumin (HSA) was not substantially found in the gel analysis or in
the proteomic study, this might be due to it being replaced by the higher-affinity and slower-
exchanging apolipoproteins.
198
While the abundant proteins were observed in the coronas of all types
of PNPs and there were many proteins shared between PNPs (Figure 5.2b), protein corona
composition varied slightly depending on the surface charge of the PNPs. Through classifying the
62
proteins by their isoelectric point (pI), negatively charged carboxyl-PNP and sulfate-PNP coronas
were enriched with proteins with pI higher than 6, and positively charged amine-PNP coronas were
enriched with proteins with pI values lower than 6 (Figure 5.2c). This difference in pI values can be
explained by the attraction between oppositely charged species.
184
In conclusion, these three types of
PNPs formed protein coronas with different quantities and diverse identities of proteins, while they
shared the dominant proteins as observed in SDS-PAGE results. They showed different preferences
in protein charge due to electrostatic attraction such that all corona-coated particles bore similar
surface charges.
Table 5.2. Representative proteins identified in the protein corona formed on uncharged, carboxyl-, and amine-
polystyrene nanoparticles by LC-MS-MS.
SPNP** CPNP** APNP**
Acc. no.* Protein identity Acc. no.* Protein identity Acc. no.* Protein identity
P06727 Apolipoprotein A-IV P02647 Apolipoprotein A-I P02647 Apolipoprotein A-I
P02768 Serum albumin P02649 Apolipoprotein E P02649 Apolipoprotein E
P02647 Apolipoprotein A-I P06727 Apolipoprotein A-IV P04004 Vitronectin
P02649 Apolipoprotein E P01024 Complement C3 P02652 Apolipoprotein A-II
P01024 Complement C3 P01008 Antithrombin-III P10909-5 Isoform 5 of Clusterin
P04114 Apolipoprotein B-100 P10909-4 Isoform 4 of Clusterin
P10909-4 Isoform 4 of Clusterin P02768 Serum albumin
P02652 Apolipoprotein A-II P04004 Vitronectin
P01009 Alpha-1-antitrypsin P00734 Prothrombin
P04004 Vitronectin P04196 Histidine-rich glycoprotein
P02654 Apolipoprotein C-I P02652 Apolipoprotein A-II
P01008 Antithrombin-III P02654 Apolipoprotein C-I
P02774-3 Isoform 3 of Vitamin D-binding protein P04114 Apolipoprotein B-100
P02655 Apolipoprotein C-II P55056 Apolipoprotein C-IV
P00338-4 Isoform 4 of L-lactate dehydrogenase
A chain
Q03591 Complement factor H-related
protein 1
P05546 Heparin cofactor 2 P05546 Heparin cofactor 2
P19827 Inter-alpha-trypsin inhibitor
heavy chain H1
P19823 Inter-alpha-trypsin inhibitor
heavy chain H2
P30101 Protein disulfide-isomerase A3 O14791-3 Isoform 3 of Apolipoprotein L1
P00738 Haptoglobin
P31146 Coronin-1A
P06703 Protein S100-A6
P17661 Desmin
P19823 Inter-alpha-trypsin inhibitor heavy chain H2
P05090 Apolipoprotein D
P16402 Histone H1.3
P07195 L-lactate dehydrogenase B chain
P00734 Prothrombin
* Uniprot accession number.
** SPNP: 100 nm sulfate-functionalized polystyrene nanoparticles; CPNP: 100 nm carboxyl-functionalized polystyrene
nanoparticles; APNP: 100 nm amine-functionalized polystyrene nanoparticles.
63
5.3.2 Cell-PNP Interactions
We have studied the perturbation of cell membranes induced by PNPs. After incubation with PNPs
for 4 h in protein-free culture media, 293T cells were fixed and fluorescently stained. The fluorescence
images of nuclei (DAPI, blue channel), cell membranes (CF633-WGA, red channel) and PNPs (green
channel) were merged in Figure 5.3a. In the condition where protein corona was absent, the sulfate-
PNPs showed almost no colocalization with cells, carboxyl- and amine PNPs were adsorbed onto cell
membranes. The adhesion of carboxyl-PNPs onto cell membranes might suggest binding of negatively
charged carboxyl-PNPs with the rare positively charged domains on cell membranes.
59, 199
Moreover,
the positively charged moiety of the zwitterionic lipid headgroups can attract negatively charged
carboxyl-PNPs.
200
While regarding positively charged amine-PNPs, not only strong binding to cell
membranes was observed, but cellular damage was also discovered with shrinkage of the cell volume
and loss of nuclear boundaries. All these adhesions and damages were alleviated with the presence of
a protein corona. A similar trend was observed after 15 h incubation. In general, there was no notable
increase in fluorescence intensity from 4 h (Figure 5.4), indicating the PNP adhesion had reached
equilibrium before or around 4 h.
64
Figure 5.3. Effect of the protein corona on cellular adhesion of nanoparticles and cell viability. Sulfate-PNPs,
carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP. (a) Confocal microscopy images of 293T cells
show that adsorption of nanoparticles was reduced with the presence of a protein corona. Images were taken after 4 h
incubation of the cells with nanoparticles in FBS-free culture media. The green channel corresponds to the fluorescently
labeled nanoparticles, blue channel corresponds to DAPI stained nuclei, and red fluorescence signal comes from CF633-
WGA stained cell membranes. The scale bars are 30 µm. (b, c) Cell viability of 293T cells exposed to nanoparticles. Cells
were incubated with nanoparticles for (b) 4 h and (c) 15 h, under conditions of presence or absence of protein corona as
well as FBS included or excluded from the culture medium. Cell viability was evaluated using the MTT assay, the viability
is normalized based on the control group where no PNPs were added. LDH leakage of 293T cells exposed to nanoparticles
for (d) 4 h and (e) 15 h were assessed, under conditions of presence or absence of protein corona as well as FBS included or
excluded from the culture medium, the leakage percentage is normalized based on the negative control group (0%) where
no PNPs were added and the positive control group (100%) where cells were treated with lysis buffer. (Unpaired t-test, *
significant at p < .05, ** significant at p < .01, *** significant at p < .001, detailed test results are listed in Table 5.4 and
Table 5.5)
65
Figure 5.4. Fluorescent intensity of adsorbed nanoparticles on 293T cells after 4 h and 15 h incubation. Medians
and interquartile ranges of calibrated fluorescence intensity were demonstrated along with individual values in graphs.
There is no significant increase of fluorescent intensity between 4 h and 15 h. Sulfate-PNPs, carboxyl-PNPs, and amine-
PNPs are denoted as SPNP, CPNP and APNP. (Unpaired t-test, * significant at p < .05, ** significant at p < .01, ***
significant at p < .001).
Aside from the morphological changes of the cells and estimation of PNP adhesion, the acute
toxicity of the PNPs with and without protein corona was assessed by MTT and LDH assays. The
MTT assay evaluates the mitochondrial activity which is related to cell viability, and the LDH assay
evaluates the release of the cytoplasmic enzyme LDH as a consequence of membranes leaking in
damaged or dead cells.
201, 202
Since the presence of fetal bovine serum (FBS) in cell culture media might
contribute to formation of a protein corona on the PNPs, assays were carried out also with conditions
where FBS was omitted.
203
We assessed cell viability and membrane integrity over 4 h and 15 h. At
the 4 h time point (Figure 5.3b), cell viability decreased to 80% with native amine-PNPs, but in the
presence of FBS or protein corona the damage from these positively charge amine-PNPs was mitigated,
which is in accordance with the microscopy images in Figure 5.3a. The viability at the 15 h time point
was more drastically affected by amine-PNPs: direct contact between amine-PNPs and cells induced
around 90% viability loss, cell viability was decreased to 80% even with FBS or protein corona
reducing the damage (Figure 5.3c). As for LDH assays (Figure 5.3d, e), 293T cells had the most
significant cytoplasm leakage with positively charged amine-PNPs, for both 4 h and 15 h time points.
Again, the presence of proteins in the cell culture medium and proteins on the surface of PNPs both
alleviated the damage to membrane integrity. This can be explained by the high surface energy of
nanoparticles. Nanoparticles tend to form coronas if not from proteins and other biomolecules in the
medium, then from cellular components. Therefore, without proteins in the media, PNPs with their
66
pristine surface would likely rupture the cell membranes and extract biomolecules to reduce their
surface energy. Once the PNP surface was pre-treated with protein and surface energy reduced by the
protein corona, the damage to the cell membranes was alleviated.
183
In addition to the surface energy,
electrostatic interactions with negatively charged cell membranes can play an important role as well.
204
Zeta potential has been measured for all three types of PNPs in all the conditions tested in the viability
or leakage assay (Table 5.3). Both FBS in culture media and protein corona treatment maintained
negative charges on initially negatively charged PNPs, while the native amine-PNPs in FBS-absent
culture media presented slight aggregation and unstable surface charges including both negative and
positive charges. The zeta potential result was in line with the cell viability and leakage data, as
positive charge is prone to be more cytotoxic. The adhesion of PNPs was established and stabilized
within the first 4 hours, while the cytotoxicity in MTT assay and LDH assay were majorly developed
from 4 h to 15 h. This suggests that the cytotoxic effects are the result of events that occur subsequent
to PNP-membrane interactions.
Table 5.3. Hydrodynamic diameters and Zeta potential of PNPs in FBS-absent cell culture media DMEM and
complete media cDMEM.
In DMEM
In cDMEM
Without
corona
With
corona
Without
corona
With
corona
Hydrodynamic
Diameter (nm)
Sulfate-PNP 85.86 ± 4.15 114.63 ± 12.78 104.85 ± 10.00 106.14 ± 20.58
Carboxyl-PNP 95.62 ± 10.22 136.04 ± 28.79 101.64 ± 12.72 124.03 ± 35.92
Amine-PNP 645.96 ± 107.95 135.13 ± 18.53 123.53 ± 31.81 128.26 ± 35.42
Zeta potential
(mV)
Sulfate-PNP -30.39 ± 1.43 -12.14 ± 2.56 -7.00 ± 4.12 -6.65 ± 5.53
Carboxyl-PNP -25.70 ± 1.64 -8.78 ± 6.11 -11.45 ± 5.05 -8.97 ± 4.98
Amine-PNP -5.30 ± 12.79 -10.45 ± 3.65 -10.78 ± 0.76 -12.39 ± 4.38
67
Table 5.4. Unpaired t-test result of viability MTT assay.
PNP Time Level 1 Level 2 Mean 1 Mean 2
Mean
Difference
P-
value
P value
summary
SPNP
4h
-FBS/-corona -FBS/+corona 109.177 86.282 22.895 0.0571 ns
-FBS/-corona +FBS/-corona 109.177 85.978 23.199 0.0133 *
-FBS/-corona +FBS/+corona 109.177 90.923 18.254 0.0252 *
-FBS/+corona +FBS/-corona 86.282 85.978 0.304 0.9747 ns
-FBS/+corona +FBS/+corona 86.282 90.923 -4.641 0.6118 ns
+FBS/-corona +FBS/+corona 85.978 90.923 -4.945 0.3933 ns
15h
-FBS/-corona -FBS/+corona 128.520 102.928 25.592 0.0195 *
-FBS/-corona +FBS/-corona 128.520 101.062 27.458 0.0123 *
-FBS/-corona +FBS/+corona 128.520 94.193 34.327 0.0065 **
-FBS/+corona +FBS/-corona 102.928 101.062 1.866 0.6374 ns
-FBS/+corona +FBS/+corona 102.928 94.193 8.736 0.0311 *
+FBS/-corona +FBS/+corona 101.062 94.193 6.870 0.1405 ns
CPNP
4h
-FBS/-corona -FBS/+corona 108.325 84.957 23.368 0.0768 ns
-FBS/-corona +FBS/-corona 108.325 92.251 16.075 0.0564 ns
-FBS/-corona +FBS/+corona 108.325 97.417 10.908 0.1441 ns
-FBS/+corona +FBS/-corona 84.957 92.251 -7.293 0.4797 ns
-FBS/+corona +FBS/+corona 84.957 97.417 -12.460 0.2591 ns
+FBS/-corona +FBS/+corona 92.251 97.417 -5.166 0.14 ns
15h
-FBS/-corona -FBS/+corona 130.606 112.074 18.532 0.0579 ns
-FBS/-corona +FBS/-corona 130.606 103.329 27.277 0.0014 **
-FBS/-corona +FBS/+corona 130.606 86.544 44.062 0.0092 **
-FBS/+corona +FBS/-corona 112.074 103.329 8.745 0.2733 ns
-FBS/+corona +FBS/+corona 112.074 86.544 25.530 0.0567 ns
+FBS/-corona +FBS/+corona 103.329 86.544 16.785 0.1406 ns
APNP
4h
-FBS/-corona -FBS/+corona 79.281 95.553 -16.272 0.0307 *
-FBS/-corona +FBS/-corona 79.281 89.520 -10.239 0.1827 ns
-FBS/-corona +FBS/+corona 79.281 95.867 -16.586 0.0071 **
-FBS/+corona +FBS/-corona 95.553 89.520 6.033 0.443 ns
-FBS/+corona +FBS/+corona 95.553 95.867 -0.314 0.9537 ns
+FBS/-corona +FBS/+corona 89.520 95.867 -6.347 0.3616 ns
15h
-FBS/-corona -FBS/+corona 7.461 82.070 -74.609 0.0026 **
-FBS/-corona +FBS/-corona 7.461 79.603 -72.143 0 ***
-FBS/-corona +FBS/+corona 7.461 77.620 -70.160 0 ***
-FBS/+corona +FBS/-corona 82.070 79.603 2.466 0.7977 ns
-FBS/+corona +FBS/+corona 82.070 77.620 4.449 0.659 ns
+FBS/-corona +FBS/+corona 79.603 77.620 1.983 0.6364 ns
68
Table 5.5. Unpaired t-test result of viability LDH assay.
PNP Time Level 1 Level 2 Mean 1 Mean 2
Mean
Difference
P-value
P value
summary
SPNP
4h
-FBS/-corona -FBS/+corona 2.027 4.817 -2.790 0.1436 ns
-FBS/-corona +FBS/-corona 2.027 2.761 -0.734 0.5381 ns
-FBS/-corona +FBS/+corona 2.027 4.936 -2.909 0.0609 ns
-FBS/+corona +FBS/-corona 4.817 2.761 2.056 0.2105 ns
-FBS/+corona +FBS/+corona 4.817 4.936 -0.120 0.9330 ns
+FBS/-corona +FBS/+corona 2.761 4.936 -2.175 0.0038 **
15h
-FBS/-corona -FBS/+corona 1.886 3.119 -1.233 0.3405 ns
-FBS/-corona +FBS/-corona 1.886 2.933 -1.047 0.3891 ns
-FBS/-corona +FBS/+corona 1.886 1.539 0.347 0.7606 ns
-FBS/+corona +FBS/-corona 3.119 2.933 0.186 0.7614 ns
-FBS/+corona +FBS/+corona 3.119 1.539 1.580 0.0512 ns
+FBS/-corona +FBS/+corona 2.933 1.539 1.394 0.0042 **
CPNP
4h
-FBS/-corona -FBS/+corona 1.172 4.333 -3.162 0.0497 *
-FBS/-corona +FBS/-corona 1.172 1.991 -0.820 0.2184 ns
-FBS/-corona +FBS/+corona 1.172 4.970 -3.798 0.0038 **
-FBS/+corona +FBS/-corona 4.333 1.991 2.342 0.1089 ns
-FBS/+corona +FBS/+corona 4.333 4.970 -0.637 0.6276 ns
+FBS/-corona +FBS/+corona 1.991 4.970 -2.979 0.0110 *
15h
-FBS/-corona -FBS/+corona 2.067 3.083 -1.015 0.0900 ns
-FBS/-corona +FBS/-corona 2.067 1.365 0.702 0.0318 *
-FBS/-corona +FBS/+corona 2.067 1.568 0.499 0.1529 ns
-FBS/+corona +FBS/-corona 3.083 1.365 1.718 0.0202 *
-FBS/+corona +FBS/+corona 3.083 1.568 1.514 0.0278 *
+FBS/-corona +FBS/+corona 1.365 1.568 -0.203 0.5045 ns
APNP
4h
-FBS/-corona -FBS/+corona 10.954 5.114 5.840 0.0035 **
-FBS/-corona +FBS/-corona 10.954 5.070 5.884 0.0001 ***
-FBS/-corona +FBS/+corona 10.954 5.539 5.415 0.0011 **
-FBS/+corona +FBS/-corona 5.114 5.070 0.044 0.9679 ns
-FBS/+corona +FBS/+corona 5.114 5.539 -0.424 0.7267 ns
+FBS/-corona +FBS/+corona 5.070 5.539 -0.469 0.5980 ns
15h
-FBS/-corona -FBS/+corona 44.606 5.766 38.840 0.0001 ***
-FBS/-corona +FBS/-corona 44.606 4.588 40.017 0.0007 ***
-FBS/-corona +FBS/+corona 44.606 4.472 40.133 0.0008 ***
-FBS/+corona +FBS/-corona 5.766 4.588 1.178 0.4884 ns
-FBS/+corona +FBS/+corona 5.766 4.472 1.294 0.4465 ns
+FBS/-corona +FBS/+corona 4.588 4.472 0.116 0.8071 ns
69
5.3.3 Biomimetic Membrane-PNP Interaction
Since lipid bilayers are the fundamental architectural structures in cell membranes, the non-specific
interactions and adhesions of nanoparticles with minimal models of lipid bilayers can be correlated to
the nanoparticle interactions with cell membranes. Four types of biomimetic membrane vesicles--
GPMVs harvested from 293T cell line, GUVs fabricated from brain, heart, and liver lipid extract were
incubated with PNPs for 4 h and 15 h. The results among all four model membranes were similar
(Figure 5.5, Figure 5.6), and it was worth noting that no penetration of PNPs through membranes was
observed. There was a clear colocalization of fluorescence signal between positively charged amine-
PNPs and model membranes, while the negatively charged sulfate-PNPs and carboxyl-PNPs showed
some aggregation and occasional adhesion to the membranes. Yet all the adhesion was strongly
diminished in the presence of protein corona. The amount of PNP adhesion was evaluated through
quantification of the green fluorescence overlaying with the membranes (Figure 5.5b), the fluorescence
intensity was normalized by circumference of the membranes at the equatorial plane facilitating
comparison between vesicles of different sizes. The fluorescent intensity of bound amine-PNPs with
protein corona (amine-PNP/corona) was significantly lower than that of amine-PNPs without corona.
Meanwhile, sulfate-PNPs and carboxyl-PNPs showed low fluorescent intensity regardless of corona
formation. This universal trend among four model membranes was in line with the toxicity of amine-
PNPs and strong electrostatic binding of amine-PNPs to 293T cells, as well as the effect of the protein
corona revealed with 293T cells in the previous section.
70
Figure 5.5. Effect of the protein corona on nanoparticle binding to biomimetic membranes. Sulfate-PNPs, carboxyl-
PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP. (a) Confocal microscopy image of DiD-stained brain
lipid GUVs (640 nm excitation) and green fluorescent nanoparticles (491 nm excitation) after 4 h incubation. (scale bar: 30
µm). (b) Fluorescent intensity of adsorbed nanoparticles on lipid membranes of GPMVs and GUVs after 4 h incubation.
Medians and interquartile ranges of calibrated fluorescence intensity were demonstrated along with individual values in
graphs. The adsorption of amine-PNPs (APNPs) was significantly decreased by protein corona (Unpaired t-test, *
significant at p < .05, ** significant at p < .01, *** significant at p < .001)
71
Figure 5.6. Confocal microscopy images of DiD-stained GUVs and GPMVs with green fluorescent PNPs after 4 h
incubation. Scale bar in GUVs panels: 30 µm, in GPMVs panels: 10 µm. Sulfate-PNPs, carboxyl-PNPs, and amine-PNPs
are denoted as SPNP, CPNP and APNP.
Interestingly, negatively charged sulfate-PNPs and carboxyl-PNPs were found to adhere less
aggressively to model membranes compared to cell membranes, while the adhesion of positively
charged amine-PNPs appeared to be more aggressive. This can be explained by the lipid composition
of model membranes. The formation of GPMVs is accompanied by a significant enrichment of
negatively charged phosphatidylserine (PS) lipids to the outer membrane leaflet and degradation of
negatively charged phosphatidylinositol (PI) lipids, making it unlikely for GPMVs to bear positive
charges.
205
As for the lipid extracts composing GUVs, the lipid headgroups are mostly negatively
charged according to the manufacturer. We believed that electrostatic interaction is an essential part
of the nanomaterial-biomembrane interaction. In this regard, model membranes can essentially
capture nanoparticle interaction as expected in cells, especially when electrostatic interactions
dominate. However, model membranes cannot fully reflect plasma membranes, as they do not
necessarily recreate the asymmetry of charge on plasma membranes. This effect can be investigated
72
by fabricating asymmetric vesicles via microfluidic technique or fabricating vesicles from purely inner
leaflet or outer leaflet lipid compositions.
206, 207
The development of PNP adhesion on model membranes was further studied after 15 h
incubation; the results from 15 h incubation showed similar behaviors, where amine-PNP binding
significantly decreased in the presence of a protein corona (Figure 5.7). Similar to plasma membranes,
there was no apparent increase of PNP fluorescence between the two time points in model membranes,
suggesting that the quantity of PNPs on membranes had reached equilibrium before 4 h incubation.
Figure 5.7. Fluorescent intensity of adsorbed nanoparticles on lipid membranes of GPMVs and GUVs after 15 h.
Medians and interquartile ranges of calibrated fluorescence intensity were demonstrated along with individual values in
graphs. Sulfate-PNPs, carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP. (Unpaired t-test, *
significant at p < .05, ** significant at p < .01, *** significant at p < .001).
5.3.4 Leakage of Biomimetic Membranes with PNPs
As a comparison to the LDH assay of the 293T cells, the leakage of the model membranes was also
studied. Calcein release assays are well-established tools for assessing membrane damage.
208, 209
The
poly-anionic nature of calcein molecule makes it membrane impermeable under normal physiological
conditions, therefore the flux of calcein across membranes indicates compromised membrane integrity.
To achieve comparable experimental conditions with LDH assays of 293T cells, we incubated the
GPMVs and GUVs with PNPs for the same time scales in calcein buffer, and we observed the influx
of calcein from the surrounding medium into the lumen. Sample images of model membrane vesicles
in the presence of positively charged amine-PNPs and amine-PNP/corona are shown in Figure 5.8a.
73
Although both PNPs and calcein appear in the green fluorescence channel, the fluorescence intensity
inside the vesicles can be considered solely coming from calcein inflow, as PNPs were found not
penetrating across model membranes in previous experiments. There was large vesicle-to-vesicle
variation of leakage behavior within each sample. The vesicle-to-vesicle heterogeneity of GUVs may
be caused by demixing of lipids in the dry film before rehydration.
210
And GPMVs can have varied
compositions depending on local surface density of cells they derive from.
211
Therefore, integrity of
model membranes with complex lipid compositions should be investigated based on population.
Figure 5.8. Effect of PNPs and protein corona on model membrane integrity. Sulfate-PNPs, carboxyl-PNPs, and
amine-PNPs are denoted as SPNP, CPNP and APNP. (a) Confocal microscopy image of DiD-stained model membrane
vesicles (red fluorescence) in 0.5 mg/mL calcein (green fluorescence) buffer after 15 h exposure to nanoparticles. White
arrows point at the vesicles that had calcein leakage through membranes (scale bars in GUV panels: 60 µm; scale bars in
GPMV panels: 30 µm). (b-c) Population of leaked vesicles after treatment of PNPs. Percentages of leaked vesicles after 4 h
(b) and 15 h (c) incubation with PNPs are presented in graphs companied with control groups where PNPs were absent.
(Unpaired t-test, * significant at p < .05, ** significant at p < .01, *** significant at p < .001)
74
Figure 5.8b and c compare the fraction of vesicle population that leaked under various conditions.
The control groups of GPMVs, heart lipid GUVs, and liver lipid GUVs have notable high population
of leakage even when no PNPs were added. One might relate this with the diffusivity of the
membranes, as previous studies suggest that calcein can be facilitated by membrane characteristics
such as high membrane fluidity and low packing density.
212
Liver lipid extract might have the highest
fluidity due to the presence unsaturated lipids or short lipid tails, since it has the lowest phase transition
temperature among the three lipid extracts.
213
Based on this premise, we conducted a control leakage
assay with GUVs fabricated from pure DOPC, which possess phase transition temperature as low as
-2 °C, and relative high diffusivity among the common phospholipids.
214
The DOPC GUVs did not
show leakage unless nanoparticles were added (Figure 5.9), the leakage of control group maintained
0% even after 15 h. This is in line with our previous study showing membrane pore formation happens
when nanoparticles adhere and impose surface tension onto the membrane surface.
93
While this result
suggests against the hypothesis that high diffusivity leads to leakage in control groups, there have been
studies demonstrating that oxidized lipid in the lipid bilayers can lead to pore formation, where these
transient pores can have sizes above 545 Å.
215, 216
We speculate that lipid oxidation might have
occurred during lipid extraction and GPMV isolation. Due to control group leakage in other model
membranes, only brain lipid GUVs can be used to assess the effect of protein coronas in the calcein
leakage assay. The differences in the three types of PNPs were not as obvious as the differences in
membrane adhesion, so we grouped the three types of the PNPs to evaluate the effect of the protein
corona. The group in the absence of protein corona had more leaked population. This result is in line
with the aforementioned mechanism: protein coronas minimize PNP adhesion, and subsequent pore
formation is hence reduced.
75
Figure 5.9. Effect of PNPs and protein corona on membrane integrity of pure lipid GUVs. DOPC GUVs and
POPC GUVs were incubated with PNPs in 0.5 mg/mL calcein buffer for 4 h and 15 h. Inflow of calcein was observed
after incubation with PNPs, percentages of leaked vesicles are presented in graphs companied with control groups where
PNPs were absent. (a) Confocal microscopy image of DOPC GUVs (DiD stained, red fluorescence) in 0.5 mg/mL calcein
(green fluorescence) buffer for 15 h in the absence of PNPs. (scale bar: 30 µm). (b-e) Percentages of leaked DOPC GUVs
after 4 h (b) and 15 h (d) incubation with PNPs as well as relative leaked population of POPC GUVs after 4 h (c) and 15 h
(e) incubation with PNPs. The control groups in all the leakage assays showed 0% leaked population. Sulfate-PNPs,
carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP (Unpaired t-test, * significant at p < .05, **
significant at p < .01, *** significant at p < .001)
To investigate this mechanism further, we performed the calcein leakage assays with GUVs
fabricated from pure DOPC lipid and pure POPC lipid (Figure 5.9). The trend observed here was
similar to that observed for brain lipid GUVs. This common result between complex lipid extract and
single lipid GUVs suggest that the pore formation on model membranes is not dependent on the
surface charge of the PNPs and strong adhesion as observed for positively charged PNPs might not be
required for pore formation, while the hydrophobic particle surface may be the dominant factor.
217
Therefore, the protein corona can reduce leakage by minimizing the possible contact of the
hydrophobic surface to biomembranes.
76
Figure 5.10. Effect of PNPs and protein corona on model membrane integrity. GPMVs and GUVs were incubated
with PNPs in 0.5 mg/mL 10 kDa rhodamine-dextran buffer for 4 h and 15 h. Inflow of dextran was observed after
incubation with PNPs, percentages of leaked vesicles are presented in graphs companied with control groups where PNPs
were absent. Sulfate-PNPs, carboxyl-PNPs, and amine-PNPs are denoted as SPNP, CPNP and APNP. (Unpaired t-test, *
significant at p < .05, ** significant at p < .01, *** significant at p < .001)
We also conducted a leakage assay using 10 kDa rhodamine-dextran (Figure 5.10). The control
groups of brain lipid GUVs and heart lipid GUVs showed lower leakage population with increasing
leakage molecular size, indicating their pore sizes might be smaller than other model membranes.
Similar to the calcein leakage assays mentioned previously, there are significant differences between
the groups without and with coronas, however, the differences between membrane types were more
obvious. Taken together, leakage assay results suggest that the pore formation is not dependent on the
charge of the PNPs, that the presence of a protein corona can alleviate pore formation, but the major
factor affecting membrane integrity in contact with the PNPs lies in the intrinsic properties of the
model membranes.
The LDH leakage of 293T cells occurred a relatively lower extent, and significant leakage from
positively charged amine-PNPs can be distinguished from control group and groups without strong
77
PNP adhesion. This suggests that cellular plasma membranes were more stable in the presence of
nanoparticles, and that membrane integrity disruption might be due to a different mechanism. Plasma
membranes are supported and tethered by the cytoskeleton, and the lateral diffusivity of lipid
molecules in plasma membrane can be one order of magnitude lower than it in free-standing GUVs.
218,
219
The cytoskeleton not only restrains the diffusion of lipid molecules, but it can also work as a
physical diffusion barrier in the influx and efflux transport of charged molecules and
macromolecules.
193, 220
We can conclude that biomimetic membranes cannot fully recreate transport
phenomena (particularly leakage) across cell membranes, as in terms of passive transport, lateral
diffusivities of lipid molecules are different in the two systems. Furthermore, the cellular uptake and
removal of charged or large molecules is mostly regulated by active transport. While at the same time,
the simplified compositions make model membranes excellent for studying nano-bio interfacial
phenomena on the lipid level without interference from other factors.
5.4 Experimental
5.4.1 Materials
Green fluorescent 100 nm diameter polystyrene nanoparticles were purchased from Magsphere, CA.
1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine
(POPC), cholesterol as well as brain, heart, and liver total lipid extract were purchased from Avanti
Polar Lipid, AL. 1,1′-dioctadecyl-3,3,3′,3′- tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate
salt (DiD) and CF633 labeled wheat germ agglutinin (CF633-WGA) were purchased from Biotium,
CA. Phosphate buffered saline (PBS), Dulbecco’s phosphate-buffered saline with calcium and
magnesium (DPBS), penicillin/streptomycin, L-glutamine, and trypsin were obtained from Corning,
NY. Dulbecco's modified Eagle's medium (DMEM) and fetal bovine serum (FBS) were purchased
from Gibco, MA. DAPI stain, 10 kDa rhodamine-dextran, BCA assay kit and LDH assay kit were
purchased from ThermoFisher Scientific, MA. Human male type AB serum (H4522) was purchased
78
from Sigma Aldrich, MO. 293T cell line was obtained from the American Type Culture Collection
(ATCC), VA. Other reagents were purchased from Sigma Aldrich, MO.
5.4.2 Protein Corona Preparation and Quantification
5 mg/mL PNPs (15 nM) were incubated in human male type AB serum (Sigma Aldrich, MO) for 30
min at 37 °C under gentle shaking. Unbound proteins were separated from PNP–protein complexes
with centrifugation (16,100 g, 20 min). Pellets were then washed three times with PBS buffer and then
resuspended in PBS buffer; PNPs with protein corona (PNP/corona) were hereby obtained. Unbound
proteins in the supernatant were quantified by a bicinchonicic acid (BCA) assay for each wash. Using
a PNP-absent control sample with the same starting concentration of human serum, the amount of
proteins extracted by PNPs can be calculated through the BCA assay results.
5.4.3 Protein Corona Elution and SDS-PAGE
Proteins were eluted from PNP/corona particles by adding elution buffer (95% 2X Laemmili buffer,
5% beta-mercaptoethanol, BioRad) and heating at 95 °C for 5 min. Then eluted proteins were
separated by centrifuging out the PNPs (16,100 g, 25 min). Proteins harvested from 0.1 mg (0.3 pmol)
PNPs were analyzed by SDS-PAGE using precast 4%-20% Mini-PROTEAN TGX polyacrylamide
gels (BioRad). Color prestained protein standard (11-245 kDa) (BioLabs) was used as a molecular
weight marker and the gels were run for 30 min at 200 V in Tris-Glycine-SDS buffer. Gels were then
stained using Coomassie Blue Protein stain.
5.4.4 Proteomic Analysis
To avoid noise introduced by surfactant in the LC-MS system, Laemmili buffer cannot be used for
protein elution, so here we used a paramagnetic bead isolation method separate from the one for SDS-
PAGE.
221
Protein corona was eluted from 0.5 mg (1.5 pmol) PNPs by adding 500 µL 8M urea buffer,
heating at 95 °C for 20 min. Proteins were then reduced by incubating with 5 mM dithiothreitol (DTT)
at 45 °C for 30 min. Alkylation was induced by incubating with 25 mM iodoacetamide (IAA) at room
temperature for 30 min in dark followed by quenching with 10 mM DTT. 30 µL each of Sera-Mag
79
Beads A (Thermo CAT No. 09-981-121) and Sera Mag Beads B (Thermo CAT No. 09-981-123) were
combined and washed with 200 µL of water 3 times. 500 µg of beads were added to each 500 µL
protein sample. 500 µL of ethanol was added and samples were incubated for 8 min at room
temperature. Supernatant was then removed. 200 µL 70% ethanol was added and incubated for 30
min twice, and the supernatant was removed each time. 180 µL of ethanol was added and the
supernatant removed. Samples were reconstituted in 100 µL of digestion buffer (50 mM HEPES, pH
8, 10 µg trypsin) and incubated overnight at 37 °C. The samples were sonicated to improve peptide
recovery. The supernatant was collected and dried in a SpeedVac. Samples were resuspended in 100
µL of 0.1% Trifluoroacetic acid (TFA) and desalted on C18 STAGE tips, and eluted with 30%
Acetonitrile, 0.1% TFA. Eluates were dried, resuspended in 10 µL of 0.1% Formic Acid and injected
onto an EasynLC1200 which was directly electrosprayed into a Q-Exactive Plus Mass Spectrometer.
RAW data was processed on Proteome Discoverer 2.2 with human FASTA file downloaded from
UniProt.
5.4.5 DLS and Zeta Potential Measurement
DLS and zeta potential measurements were taken on a Wyatt Mobius mobility instrument. Samples
of PNPs in PBS were prepared at a concentration of 0.5 mg/mL (1.5 nM). Samples of PNPs in cell
culture media were prepared by adding PNPs (1 mg/mL in PBS, 3 nM) to the culture media at a
concentration of 0.1 mg/mL (0.3 nM), the samples were equilibrated for 5 min before the
measurements. Measurements were taken at 25 °C with 5 s run time. Diameters and zeta potentials
are reported as averages and standard deviations of ten and three acquirements, respectively.
5.4.6 Cell Culture and Imaging
293T cells were cultured in complete cell culture medium (cDMEM), consisting of DMEM
supplemented with 10% FBS along with 1% penicillin/streptomycin and 2 mM L-glutamine. Cells
were grown in a 5% CO 2 incubator at 37 °C, passaged using trypsin. Cells were treated with PNPs by
replacing culture medium with DMEM containing 0.1 mg/mL PNPs (0.3 nM). For imaging, cells
80
were fixed with 4% paraformaldehyde, and stained with 1 µg/mL DAPI and 5 µg/mL CF633-WGA,
then observed with a spinning disk confocal microscope (Nikon Eclipse TiE equipped with a
Yokogawa confocal head). Quantification of PNP adhesion was achieved by measuring the intensity
of green fluorescence colocalized with the cells, and this intensity was calibrated versus a control group
where PNPs were absent to eliminate background fluorescence. Mean fluorescence was normalized
based on the cell area.
5.4.7 Cell Viability MTT Assay
293T cells were cultured in 96-well plates overnight with a seeding density of 1×10
5
cell/mL in 0.1
mL cDMEM, followed by incubation with 0.1 mg/mL PNPs (0.3 nM) in DMEM or cDMEM at
37 °C for 4 h and 15 h. Untreated cells were used as a negative control. The incubation media was
then replaced with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Sigma
Aldrich, MO) dissolved in cDMEM (0.5 mg/mL) to start the assay. Formazan was allowed to form
during 4 h incubation at 37 °C. The formed formazan was dissolved in DMSO and absorbance at 550
nm was measured with a microplate reader (Synergy H1; BioTek). Each sample was analyzed in four
replicates. No interference of the PNPs present in solution with MTT was found in the absorbance
measurement.
5.4.8 Lactate Dehydrogenase (LDH) Release Assay
293T cells were cultured in 96-well plates overnight with a seeding density of 1×10
5
cell/mL in 0.1
mL cDMEM, and then treated with 0.1 mg/mL PNPs (0.3 nM) in DMEM or cDMEM, incubating
at 37 °C for 4 h and 15 h. LDH assays were carried out according to the manufacturer's instruction.
The percentage of released LDH was normalized by the amount of LDH from complete lysis of
control cells. Four replicates were used for each condition. And no noticeable assay activity between
PNPs and LDH assay buffer was found.
81
5.4.9 GPMV Preparation
At 70%-80% cell confluence, GPMVs were isolated by chemical induced cell blebbing with 25 mM
paraformaldehyde and 2 mM dithiothreitol in DPBS buffer for 1 h at 37 °C. The membrane dye DiD
was added to the GPMV suspension at a concentration of 5 µg/mL.
5.4.10 GUV Preparation
The GUVs were prepared by the agarose rehydration method.
81
The agarose hydration method was
selected as it avoids oxidative degradation of lipids as in the alternative electro-formation method, and
it is more capable of incorporating charged lipids thus preserving most of the natural lipid
compositions.
222
Despite the possible existence of agarose residue encapsulated in the GUVs,
223
our
study focuses on the interplay on the surface of the membranes, agarose hydration method is preferred
in our study. Our previous work with agarose hydration shows that vesicles made via this method
maintain expected liquid-liquid phase segregation behavior, suggesting that any agarose present
introduces minimal biophysical artifacts.
224, 225
Brain, heart and liver total lipid extract were dissolved
in chloroform; heart total lipid extract requires addition of 10 wt% cholesterol to form GUVs. The
lipid solution was deposited on 2 wt% agarose-coated coverslips. After chloroform evaporation, PBS
buffer was added to rehydrate the lipid film. The membrane dye DiD was incorporated in the
rehydration buffer with a final concentration of 5 µg/mL. The lipid film was rehydrated at 37 °C for
30 min and GUVs were harvested afterwards.
5.4.11 Microscopy Imaging of Model Vesicles and Quantification
Microscopy images were taken with a spinning disk confocal microscope (Nikon Eclipse TiE equipped
with a Yokogawa confocal head). PNPs and calcein can be captured with 491 nm laser excitation, 10
kDa rhodamine dextran can be captured with 561 nm laser excitation and DiD dye with 640 nm. To
avoid crosstalk between different dyes, emission signals were collected independently in serial mode.
Images were acquired at constant laser power and exposure time. Model membrane samples were
82
held in glass-bottom multiwell plates treated with bovine serum albumin (BSA) and rinsed three times
with PBS. Images were taken at the equatorial plane of each vesicle.
For PNP adhesion observation, the GPMV or GUV suspension was loaded into BSA-treated
wells and then incubated with 0.1 mg/mL (0.3 nM) PNPs. PNP adhesion was observed and recorded
after incubation at room temperature for 4 h and 15 h. Control groups left out PNPs, but same volume
of PBS buffer was added instead. Quantification of PNP adhesion was achieved by measuring the
intensity of green fluorescence colocalized with the membranes (the outer contour and inner contour
of each vesicle were identified, and the intensity of nanoparticle fluorescence was measured only
between these two contours). The fluorescence was then normalized based on the circumference of
the membrane at the equatorial plane to allow for comparison between vesicles with different sizes.
This normalized fluorescence intensity was further calibrated by subtracting the normalized intensity
from the control group to eliminate background fluorescence.
For the membrane integrity study, the GPMV or GUV suspension was loaded into BSA-treated
wells with 1mg/mL calcein in PBS buffer (or 1mg/mL 10 kDa rhodamine-dextran in PBS buffer) at
a 1:1 ratio, followed by PNP addition to a final concentration of 0.1 mg/mL (0.3 nM). The osmolarity
of calcein buffer or dextran buffer was balanced with the vesicle suspensions. Control groups were
identical with the exception of PNP addition; the same volume of PBS buffer was added instead.
Quantification was carried out by calculating the fractional population of vesicles with calcein leaked
into the lumen. We set the threshold to be ten percent of the background fluorescence intensity,
vesicles whose fluorescence intensity differences across membranes were less than this threshold were
categorized as leaked vesicles.
5.5 Conclusion
We have studied the effect of the protein corona on nanoparticle-biomembrane interactions. Through
investigating these non-specific nanoparticle-biomembrane interactions by establishing a correlation
83
between plasma membranes and biomimetic membranes, we have made the following conclusions:
Protein corona composition depends on the surface charge of nanoparticles, but in general, it reduces
nanoparticle adhesion and damage to the biomembranes. This is possibly due to surface energy
stabilization and charge modification that comes with the corona. As a crucial part of this interplay,
electrostatic interaction between nanoparticles and plasma membranes can be correlated with
cytotoxicity of the nanoparticles. It is advantageous that model membranes such as GPMVs and
GUVs can relate with plasma membranes through this fundamental interaction with similar responses.
However, the model membranes have their limitations as their simplified composition does not mimic
the complexity and dynamics in plasma membranes, such as differences in fluidity and tethering from
cytoskeleton.
This study recognized the toxicity of positively charged nanoparticles and the general protective
effect of the protein corona in the interplay between nanomaterials and biomembranes, providing
insights about the relationship between electrostatic interactions and biological system perturbation
caused by nanoparticles. Furthermore, defining the limits of the correlation between plasma
membranes and biomimetic membranes has revealed promising applications of model membranes in
studying nano-biomembrane interface interactions. It also provides an approach for studying these
phenomena by reducing them into simplified models that isolate individual biophysical aspects of the
system.
84
Chapter 6 . Conclusions and Future Outlook
We have achieved high-throughput synthesis of functional nanoparticles, taking advantage of the
millifluidic systems. The nanoparticle synthesis took place with shorter reaction time and higher yield
compared to their batch analogue, owing to the superior heat and mass transfer inherent in the high
surface-to-volume ratio milli-channels. Further scale up of the millifluidic reactor has been
accomplished in parallel with feedback-monitored automation, which advances the millifluidic
nanoparticle synthesis towards industrial application. Currently we focus on the large-scale synthesis
of nanoparticles. However, preparation of the reaction reagents as well as purification of the product
are still left in batch scale. To accomplish scale-up and automation of the entire cycle of nanoparticle
production, it is necessary to consider adapting preparation and follow-up processes into continuous
systems.
On the other hand, excessive exposure of nanoparticles, especially nanoplastics, resides in the
environment and can pose potential health risks to human. Our studies on the biophysical aspects
have proved the dominant role of electrostatic force in the non-specific interactions between
nanoplastics and biomembranes. In addition, the presence of protein corona which are rapidly formed
in biological environment and modifies particle surfaces can mitigate the electrostatic interactions,
minimizing the non-specific adhesions and subsequent membrane perturbation. Furthermore,
systematic comprehension was obtained through comparison across in vitro experiments and
experiments with various models. By far, a comprehensive understanding on the health effect of
nanoparticles has not been reached, partially due to the complexity of nanoparticle species and
biological systems. Safe handling of nanoparticles with fully understood mechanism would help
85
minimize health effect to human in the future. Anyway, better understanding, enhanced designs, and
more effective regulation over the nanoparticle production, application as well as disposal should be
taken into effect to minimize unnecessary the exposure and leakage into the environment. After all, it
is for us to decide that whether nanoparticles become an environmental problem or a problem solver.
86
References .
1. Khan, I.; Saeed, K.; Khan, I., Nanoparticles: Properties, applications and toxicities. Arabian
Journal of Chemistry 2019, 12 (7), 908-931.
2. Charitidis, C. A.; Georgiou, P.; Koklioti, M. A.; Trompeta, A.-F.; Markakis, V.,
Manufacturing nanomaterials: from research to industry. Manufacturing Rev. 2014, 1.
3. Raab, C.; Simko, M.; Fiedeler, U.; Nentwich, M.; Gazso, A., Production of nanoparticles
and nanomaterials. Nano Trust Dossier 2011, 6, 1998-7293.
4. Burda, C.; Chen, X.; Narayanan, R.; El-Sayed, M. A., Chemistry and Properties of
Nanocrystals of Different Shapes. Chem. Rev. 2005, 105 (4), 1025-1102.
5. Skrabalak, S. E.; Brutchey, R. L., Going with the Flow: Continuous Flow Routes to Colloidal
Nanoparticles. ACS Publications: 2016.
6. Hartman, R. L.; McMullen, J. P.; Jensen, K. F., Deciding Whether To Go with the Flow:
Evaluating the Merits of Flow Reactors for Synthesis. Angew. Chem. Int. Ed. 2011, 50 (33),
7502-7519.
7. Zhang, J.; Wang, K.; Teixeira, A. R.; Jensen, K. F.; Luo, G., Design and Scaling Up of
Microchemical Systems: A Review. Annual Review of Chemical and Biomolecular Engineering
2017, 8 (1), 285-305.
8. Epps, R. W.; Felton, K. C.; Coley, C. W.; Abolhasani, M., Automated microfluidic platform
for systematic studies of colloidal perovskite nanocrystals: towards continuous nano-
manufacturing. Lab on a Chip 2017, 17 (23), 4040-4047.
9. Mason, B. P.; Price, K. E.; Steinbacher, J. L.; Bogdan, A. R.; McQuade, D. T., Greener
approaches to organic synthesis using microreactor technology. Chem. Rev. 2007, 107 (6), 2300-
2318.
10. Rubens, M.; Vrijsen, J. H.; Laun, J.; Junkers, T., Precise Polymer Synthesis by Autonomous
Self-Optimizing Flow Reactors. Angew. Chem. Int. Ed. 2019, 58 (10), 3183-3187.
11. Krishnadasan, S.; Brown, R. J. C.; deMello, A. J.; deMello, J. C., Intelligent routes to the
controlled synthesis of nanoparticles. Lab on a Chip 2007, 7 (11), 1434-1441.
12. Riche, C. T.; Roberts, E. J.; Gupta, M.; Brutchey, R. L.; Malmstadt, N., Flow invariant
droplet formation for stable parallel microreactors. Nature Communications 2016, 7, 10780.
13. Ahrberg, C. D.; Choi, J. W.; Chung, B. G., Droplet-based synthesis of homogeneous
magnetic iron oxide nanoparticles. Beilstein Journal of Nanotechnology 2018, 9, 2413-2420.
14. Liu, Y.; Jiang, X., Why microfluidics? Merits and trends in chemical synthesis. Lab on a Chip
2017, 17 (23), 3960-3978.
87
15. Lin, X. Z.; Terepka, A. D.; Yang, H., Synthesis of Silver Nanoparticles in a Continuous Flow
Tubular Microreactor. Nano Lett. 2004, 4 (11), 2227-2232.
16. Yen, B. K. H.; Stott, N. E.; Jensen, K. F.; Bawendi, M. G., A Continuous-Flow
Microcapillary Reactor for the Preparation of a Size Series of CdSe Nanocrystals. Adv. Mater.
2003, 15 (21), 1858-1862.
17. Abou-Hassan, A.; Bazzi, R.; Cabuil, V., Multistep Continuous-Flow Microsynthesis of
Magnetic and Fluorescent γ-Fe2O3@SiO2 Core/Shell Nanoparticles. Angew. Chem. Int. Ed.
2009, 48 (39), 7180-7183.
18. Lee, C.-Y.; Chang, C.-L.; Wang, Y.-N.; Fu, L.-M., Microfluidic mixing: a review. Int J Mol
Sci 2011, 12 (5), 3263-3287.
19. Kamholz, A. E.; Yager, P., Theoretical Analysis of Molecular Diffusion in Pressure-Driven
Laminar Flow in Microfluidic Channels. Biophys. J. 2001, 80 (1), 155-160.
20. Song, H.; Tice, J. D.; Ismagilov, R. F., A microfluidic system for controlling reaction
networks in time. Angew. Chem. 2003, 115 (7), 792-796.
21. Zhang, L.; Niu, G.; Lu, N.; Wang, J.; Tong, L.; Wang, L.; Kim, M. J.; Xia, Y., Continuous
and scalable production of well-controlled noble-metal nanocrystals in milliliter-sized droplet
reactors. Nano Lett. 2014, 14 (11), 6626-6631.
22. Chan, E. M.; Alivisatos, A. P.; Mathies, R. A., High-Temperature Microfluidic Synthesis of
CdSe Nanocrystals in Nanoliter Droplets. J. Am. Chem. Soc. 2005, 127 (40), 13854-13861.
23. Yen, B. K. H.; Günther, A.; Schmidt, M. A.; Jensen, K. F.; Bawendi, M. G., A
Microfabricated Gas–Liquid Segmented Flow Reactor for High-Temperature Synthesis: The
Case of CdSe Quantum Dots. Angew. Chem. Int. Ed. 2005, 44 (34), 5447-5451.
24. Khan, S. A.; Duraiswamy, S., Controlling bubbles using bubbles-microfluidic synthesis of
ultra-small gold nanocrystals with gas-evolving reducing agents. Lab on a Chip 2012, 12 (10),
1807-1812.
25. Poe, S. L.; Cummings, M. A.; Haaf, M. P.; McQuade, D. T., Solving the Clogging Problem:
Precipitate-Forming Reactions in Flow. Angew. Chem. Int. Ed. 2006, 45 (10), 1544-1548.
26. Günther, A.; Jhunjhunwala, M.; Thalmann, M.; Schmidt, M. A.; Jensen, K. F.,
Micromixing of Miscible Liquids in Segmented Gas−Liquid Flow. Langmuir 2005, 21 (4),
1547-1555.
27. Zhang, L.; Wang, Y.; Tong, L.; Xia, Y., Seed-Mediated Synthesis of Silver Nanocrystals
with Controlled Sizes and Shapes in Droplet Microreactors Separated by Air. Langmuir 2013,
29 (50), 15719-15725.
28. Niu, G.; Ruditskiy, A.; Vara, M.; Xia, Y., Toward continuous and scalable production of
colloidal nanocrystals by switching from batch to droplet reactors. Chem. Soc. Rev. 2015, 44
(16), 5806-5820.
88
29. Griffin, S.; Masood, M. I.; Nasim, M. J.; Sarfraz, M.; Ebokaiwe, A. P.; Schäfer, K.-H.;
Keck, C. M.; Jacob, C., Natural Nanoparticles: A Particular Matter Inspired by Nature.
Antioxidants (Basel) 2017, 7 (1), 3.
30. Bundschuh, M.; Filser, J.; Lüderwald, S.; McKee, M. S.; Metreveli, G.; Schaumann, G.
E.; Schulz, R.; Wagner, S., Nanoparticles in the environment: where do we come from, where
do we go to? Environ Sci Eur 2018, 30 (1), 6-6.
31. Nazarenko, Y.; Han, T. W.; Lioy, P. J.; Mainelis, G., Potential for exposure to engineered
nanoparticles from nanotechnology-based consumer spray products. Journal of Exposure Science
& Environmental Epidemiology 2011, 21 (5), 515-528.
32. Nowack, B.; Bucheli, T. D., Occurrence, behavior and effects of nanoparticles in the
environment. Environ. Pollut. 2007, 150 (1), 5-22.
33. Andrady, A. L., Microplastics in the marine environment. Marine Pollution Bulletin 2011, 62
(8), 1596-1605.
34. Gigault, J.; Halle, A. t.; Baudrimont, M.; Pascal, P.-Y.; Gauffre, F.; Phi, T.-L.; El Hadri,
H.; Grassl, B.; Reynaud, S., Current opinion: What is a nanoplastic? Environ. Pollut. 2018,
235, 1030-1034.
35. Koelmans, A. A.; Besseling, E.; Shim, W. J., Nanoplastics in the Aquatic Environment.
Critical Review. In Marine Anthropogenic Litter, Bergmann, M.; Gutow, L.; Klages, M., Eds.
Springer International Publishing: Cham, 2015; pp 325-340.
36. Stephens, B.; Azimi, P.; El Orch, Z.; Ramos, T., Ultrafine particle emissions from desktop
3D printers. Atmos. Environ. 2013, 79, 334-339.
37. Jahnke, A.; Arp, H. P. H.; Escher, B. I.; Gewert, B.; Gorokhova, E.; Kühnel, D.;
Ogonowski, M.; Potthoff, A.; Rummel, C.; Schmitt-Jansen, M.; Toorman, E.; MacLeod,
M., Reducing Uncertainty and Confronting Ignorance about the Possible Impacts of
Weathering Plastic in the Marine Environment. Environmental Science & Technology Letters 2017,
4 (3), 85-90.
38. Geyer, R.; Jambeck, J. R.; Law, K. L., Production, use, and fate of all plastics ever made.
Science Advances 2017, 3 (7), e1700782.
39. Hernandez, L. M.; Yousefi, N.; Tufenkji, N., Are There Nanoplastics in Your Personal Care
Products? Environmental Science & Technology Letters 2017, 4 (7), 280-285.
40. Salvador Cesa, F.; Turra, A.; Baruque-Ramos, J., Synthetic fibers as microplastics in the
marine environment: A review from textile perspective with a focus on domestic washings.
Sci. Total Environ. 2017, 598, 1116-1129.
41. Kole, P. J.; Löhr, A. J.; Van Belleghem, F. G. A. J.; Ragas, A. M. J., Wear and Tear of Tyres:
A Stealthy Source of Microplastics in the Environment. Int J Environ Res Public Health 2017,
14 (10), 1265.
42. Wetherbee, G. A.; Baldwin, A. K.; Ranville, J. F. It is raining plastic; 2019-1048; Reston, VA,
2019.
89
43. Bergmann, M.; Mützel, S.; Primpke, S.; Tekman, M. B.; Trachsel, J.; Gerdts, G., White
and wonderful? Microplastics prevail in snow from the Alps to the Arctic. Science Advances
2019, 5 (8), eaax1157.
44. Alimi, O. S.; Farner Budarz, J.; Hernandez, L. M.; Tufenkji, N., Microplastics and
Nanoplastics in Aquatic Environments: Aggregation, Deposition, and Enhanced
Contaminant Transport. Environmental Science & Technology 2018, 52 (4), 1704-1724.
45. Schwabl, P.; Köppel, S.; Königshofer, P.; Bucsics, T.; Trauner, M.; Reiberger, T.; Liebmann,
B., Detection of Various Microplastics in Human Stool: A Prospective Case Series. Annals of
Internal Medicine 2019, 171 (7), 453-457.
46. Toussaint, B.; Raffael, B.; Angers-Loustau, A.; Gilliland, D.; Kestens, V.; Petrillo, M.;
Rio-Echevarria, I. M.; Van den Eede, G., Review of micro- and nanoplastic contamination in
the food chain. Food Additives & Contaminants: Part A 2019, 36 (5), 639-673.
47. Lehner, R.; Weder, C.; Petri-Fink, A.; Rothen-Rutishauser, B., Emergence of Nanoplastic in
the Environment and Possible Impact on Human Health. Environmental Science & Technology
2019, 53 (4), 1748-1765.
48. Wright, S. L.; Kelly, F. J., Plastic and Human Health: A Micro Issue? Environmental Science &
Technology 2017, 51 (12), 6634-6647.
49. Vance, M. E.; Kuiken, T.; Vejerano, E. P.; McGinnis, S. P.; Hochella, M. F., Jr.; Rejeski,
D.; Hull, M. S., Nanotechnology in the real world: Redeveloping the nanomaterial consumer
products inventory. Beilstein Journal of Nanotechnology 2015, 6, 1769-1780.
50. Brigger, I.; Dubernet, C.; Couvreur, P., Nanoparticles in cancer therapy and diagnosis.
Advanced Drug Delivery Reviews 2002, 54 (5), 631-651.
51. Sheng, R. L.; Zhuang, X. Q.; Wang, Z.; Cao, A.; Lin, K. L.; Zhu, J., Cationic Nanoparticles
Assembled from Natural-Based Steroid Lipid for Improved Intracellular Transport of siRNA
and pDNA. Nanomaterials 2016, 6 (4), 18.
52. Kettiger, H.; Schipanski, A.; Wick, P.; Huwyler, J., Engineered nanomaterial uptake and
tissue distribution: from cell to organism. Int J Nanomedicine 2013, 8, 3255-3269.
53. Lewinski, N.; Colvin, V.; Drezek, R., Cytotoxicity of nanoparticles. small 2008, 4 (1), 26-49.
54. Napierska, D.; Thomassen, L. C. J.; Rabolli, V.; Lison, D.; Gonzalez, L.; Kirsch-Volders,
M.; Martens, J. A.; Hoet, P. H., Size-Dependent Cytotoxicity of Monodisperse Silica
Nanoparticles in Human Endothelial Cells. Small 2009, 5 (7), 846-853.
55. Carlson, C.; Hussain, S. M.; Schrand, A. M.; K. Braydich-Stolle, L.; Hess, K. L.; Jones, R.
L.; Schlager, J. J., Unique cellular interaction of silver nanoparticles: size-dependent
generation of reactive oxygen species. The journal of physical chemistry B 2008, 112 (43), 13608-
13619.
56. Ding, L.; Stilwell, J.; Zhang, T.; Elboudwarej, O.; Jiang, H.; Selegue, J. P.; Cooke, P. A.;
Gray, J. W.; Chen, F. F., Molecular Characterization of the Cytotoxic Mechanism of
90
Multiwall Carbon Nanotubes and Nano-Onions on Human Skin Fibroblast. Nano Lett. 2005,
5 (12), 2448-2464.
57. Monopoli, M. P.; Åberg, C.; Salvati, A.; Dawson, K. A., Biomolecular coronas provide the
biological identity of nanosized materials. Nature nanotechnology 2012, 7 (12), 779.
58. Docter, D.; Westmeier, D.; Markiewicz, M.; Stolte, S.; Knauer, S. K.; Stauber, R. H., The
nanoparticle biomolecule corona: lessons learned – challenge accepted? Chem. Soc. Rev. 2015,
44 (17), 6094-6121.
59. Verma, A.; Stellacci, F., Effect of Surface Properties on Nanoparticle–Cell Interactions. Small
2010, 6 (1), 12-21.
60. Chung, T.-H.; Wu, S.-H.; Yao, M.; Lu, C.-W.; Lin, Y.-S.; Hung, Y.; Mou, C.-Y.; Chen,
Y.-C.; Huang, D.-M., The effect of surface charge on the uptake and biological function of
mesoporous silica nanoparticles in 3T3-L1 cells and human mesenchymal stem cells.
Biomaterials 2007, 28 (19), 2959-2966.
61. Peetla, C.; Labhasetwar, V., Effect of Molecular Structure of Cationic Surfactants on
Biophysical Interactions of Surfactant-Modified Nanoparticles with a Model Membrane and
Cellular Uptake. Langmuir 2009, 25 (4), 2369-2377.
62. Leroueil, P. R.; Berry, S. A.; Duthie, K.; Han, G.; Rotello, V. M.; McNerny, D. Q.; Baker,
J. R.; Orr, B. G.; Banaszak Holl, M. M., Wide Varieties of Cationic Nanoparticles Induce
Defects in Supported Lipid Bilayers. Nano Lett. 2008, 8 (2), 420-424.
63. Rascol, E.; Devoisselle, J.-M.; Chopineau, J., The relevance of membrane models to
understand nanoparticles–cell membrane interactions. Nanoscale 2016, 8 (9), 4780-4798.
64. Chen, K. L.; Bothun, G. D., Nanoparticles Meet Cell Membranes: Probing Nonspecific
Interactions using Model Membranes. Environmental Science & Technology 2014, 48 (2), 873-
880.
65. Moghadam, B. Y.; Hou, W.-C.; Corredor, C.; Westerhoff, P.; Posner, J. D., Role of
Nanoparticle Surface Functionality in the Disruption of Model Cell Membranes. Langmuir
2012, 28 (47), 16318-16326.
66. Corredor, C.; Hou, W.-C.; Klein, S. A.; Moghadam, B. Y.; Goryll, M.; Doudrick, K.;
Westerhoff, P.; Posner, J. D., Disruption of model cell membranes by carbon nanotubes.
Carbon 2013, 60, 67-75.
67. Peetla, C.; Labhasetwar, V., Biophysical Characterization of Nanoparticle−Endothelial
Model Cell Membrane Interactions. Molecular Pharmaceutics 2008, 5 (3), 418-429.
68. de Planque, M. R. R.; Aghdaei, S.; Roose, T.; Morgan, H., Electrophysiological
Characterization of Membrane Disruption by Nanoparticles. ACS Nano 2011, 5 (5), 3599-3606.
69. Richter, R.; Mukhopadhyay, A.; Brisson, A., Pathways of Lipid Vesicle Deposition on Solid
Surfaces: A Combined QCM-D and AFM Study. Biophys. J. 2003, 85 (5), 3035-3047.
91
70. Richter, R. P.; Lai Kee Him, J.; Tessier, B.; Tessier, C.; Brisson, A. R., On the Kinetics of
Adsorption and Two-Dimensional Self-Assembly of Annexin A5 on Supported Lipid Bilayers.
Biophys. J. 2005, 89 (5), 3372-3385.
71. Hou, W.-C.; Moghadam, B. Y.; Corredor, C.; Westerhoff, P.; Posner, J. D., Distribution of
Functionalized Gold Nanoparticles between Water and Lipid Bilayers as Model Cell
Membranes. Environmental Science & Technology 2012, 46 (3), 1869-1876.
72. Hou, W.-C.; Moghadam, B. Y.; Westerhoff, P.; Posner, J. D., Distribution of Fullerene
Nanomaterials between Water and Model Biological Membranes. Langmuir 2011, 27 (19),
11899-11905.
73. Castile, J. D.; Taylor, K. M. G., Factors affecting the size distribution of liposomes produced
by freeze–thaw extrusion. Int. J. Pharm. 1999, 188 (1), 87-95.
74. Kim, S.; Martin, G. M., Preparation of cell-size unilamellar liposomes with high captured
volume and defined size distribution. Biochimica et Biophysica Acta (BBA) - Biomembranes 1981,
646 (1), 1-9.
75. Pautot, S.; Frisken, B. J.; Weitz, D. A., Production of Unilamellar Vesicles Using an Inverted
Emulsion. Langmuir 2003, 19 (7), 2870-2879.
76. Akashi, K.; Miyata, H.; Itoh, H.; Kinosita, K., Preparation of giant liposomes in
physiological conditions and their characterization under an optical microscope. Biophys. J.
1996, 71 (6), 3242-3250.
77. Reeves, J. P.; Dowben, R. M., Formation and properties of thin ‐walled phospholipid vesicles.
J. Cell. Physiol. 1969, 73 (1), 49-60.
78. Angelova, M. I.; Dimitrov, D. S., Liposome electroformation. Faraday discussions of the
Chemical Society 1986, 81, 303-311.
79. Kuribayashi, K.; Tresset, G.; Ph, C.; Fujita, H.; Takeuchi, S., Electroformation of giant
liposomes in microfluidic channels. Meas. Sci. Technol. 2006, 17 (12), 3121.
80. Micheletto, Y. M. S.; Marques, C. M.; Silveira, N. d. P. d.; Schroder, A. P., Electroformation
of giant unilamellar vesicles: investigating vesicle fusion versus bulge merging. Langmuir 2016,
32 (32), 8123-8130.
81. Horger, K. S.; Estes, D. J.; Capone, R.; Mayer, M., Films of Agarose Enable Rapid
Formation of Giant Liposomes in Solutions of Physiologic Ionic Strength. J. Am. Chem. Soc.
2009, 131 (5), 1810-1819.
82. Scott, R. E., Plasma membrane vesiculation: a new technique for isolation of plasma
membranes. Science 1976, 194 (4266), 743.
83. Sezgin, E.; Kaiser, H.-J.; Baumgart, T.; Schwille, P.; Simons, K.; Levental, I., Elucidating
membrane structure and protein behavior using giant plasma membrane vesicles. Nature
protocols 2012, 7 (6), 1042.
92
84. Del Piccolo, N.; Placone, J.; He, L.; Agudelo, S. C.; Hristova, K., Production of Plasma
Membrane Vesicles with Chloride Salts and Their Utility as a Cell Membrane Mimetic for
Biophysical Characterization of Membrane Protein Interactions. Anal. Chem. 2012, 84 (20),
8650-8655.
85. Liu, Q.; Bi, C.; Li, J.; Liu, X.; Peng, R.; Jin, C.; Sun, Y.; Lyu, Y.; Liu, H.; Wang, H.;
Luo, C.; Tan, W., Generating Giant Membrane Vesicles from Live Cells with Preserved
Cellular Properties. Research 2019, 2019, 9.
86. Kelly, C. V.; Kober, M.-M. T.; Kinnunen, P.; Reis, D. A.; Orr, B. G.; Banaszak Holl, M.
M., Pulsed-laser creation and characterization of giant plasma membrane vesicles from cells.
Journal of Biological Physics 2009, 35 (3), 279-295.
87. Dubavik, A.; Sezgin, E.; Lesnyak, V.; Gaponik, N.; Schwille, P.; Eychmüller, A.,
Penetration of amphiphilic quantum dots through model and cellular plasma membranes. Acs
Nano 2012, 6 (3), 2150-2156.
88. Hoffman, J. F., On Red Blood Cells, Hemolysis and Resealed Ghosts. In The Use of Resealed
Erythrocytes as Carriers and Bioreactors, Magnani, M.; DeLoach, J. R., Eds. Springer US: Boston,
MA, 1992; pp 1-15.
89. Hu, C.-M. J.; Zhang, L.; Aryal, S.; Cheung, C.; Fang, R. H.; Zhang, L., Erythrocyte
membrane-camouflaged polymeric nanoparticles as a biomimetic delivery platform.
Proceedings of the National Academy of Sciences 2011, 108 (27), 10980.
90. Roiter, Y.; Ornatska, M.; Rammohan, A. R.; Balakrishnan, J.; Heine, D. R.; Minko, S.,
Interaction of Lipid Membrane with Nanostructured Surfaces. Langmuir 2009, 25 (11), 6287-
6299.
91. Roiter, Y.; Ornatska, M.; Rammohan, A. R.; Balakrishnan, J.; Heine, D. R.; Minko, S.,
Interaction of Nanoparticles with Lipid Membrane. Nano Lett. 2008, 8 (3), 941-944.
92. Yu, Y.; Granick, S., Pearling of Lipid Vesicles Induced by Nanoparticles. J. Am. Chem. Soc.
2009, 131 (40), 14158-14159.
93. Li, S.; Malmstadt, N., Deformation and poration of lipid bilayer membranes by cationic
nanoparticles. Soft Matter 2013, 9 (20), 4969-4976.
94. Xi, A.; Bothun, G. D., Centrifugation-based assay for examining nanoparticle-lipid membrane
binding and disruption. Analyst 2014, 139 (5), 973-981.
95. Grossutti, M.; Seenath, R.; Noel, J. A.; Lipkowski, J., Infrared and fluorescence
spectroscopic studies of a phospholipid bilayer supported by a soft cationic hydrogel scaffold.
J. Colloid Interface Sci. 2016, 473, 162-171.
96. Alkhammash, H. I.; Li, N.; Berthier, R.; de Planque, M. R. R., Native silica nanoparticles
are powerful membrane disruptors. PCCP 2015, 17 (24), 15547-15560.
97. Rasch, M. R.; Rossinyol, E.; Hueso, J. L.; Goodfellow, B. W.; Arbiol, J.; Korgel, B. A.,
Hydrophobic gold nanoparticle self-assembly with phosphatidylcholine lipid: membrane-
loaded and janus vesicles. Nano Lett. 2010, 10 (9), 3733-3739.
93
98. Schmidt, U.; Guigas, G.; Weiss, M., Cluster Formation of Transmembrane Proteins Due to
Hydrophobic Mismatching. Phys. Rev. Lett. 2008, 101 (12), 128104.
99. Wang, B.; Zhang, L.; Bae, S. C.; Granick, S., Nanoparticle-induced surface reconstruction
of phospholipid membranes. Proceedings of the National Academy of Sciences 2008, 105 (47),
18171-18175.
100. Chen, X.; Tian, F.; Zhang, X.; Wang, W., Internalization pathways of nanoparticles and
their interaction with a vesicle. Soft Matter 2013, 9 (31), 7592-7600.
101. Šari ć, A.; Cacciuto, A., Mechanism of Membrane Tube Formation Induced by Adhesive
Nanocomponents. Phys. Rev. Lett. 2012, 109 (18), 188101.
102. Yue, T.; Zhang, X.; Huang, F., Membrane monolayer protrusion mediates a new
nanoparticle wrapping pathway. Soft Matter 2014, 10 (12), 2024-2034.
103. Stachowiak, J. C.; Schmid, E. M.; Ryan, C. J.; Ann, H. S.; Sasaki, D. Y.; Sherman, M. B.;
Geissler, P. L.; Fletcher, D. A.; Hayden, C. C., Membrane bending by protein–protein
crowding. Nat Cell Biol 2012, 14 (9), 944-949.
104. Lu, B.; Smith, T.; Schmidt, J. J., Nanoparticle-lipid bilayer interactions studied with lipid
bilayer arrays. Nanoscale 2015, 7 (17), 7858-7866.
105. Gilbert, D. L.; Ehrenstein, G., Effect of Divalent Cations on Potassium Conductance of Squid
Axons: Determination of Surface Charge. Biophys. J. 1969, 9 (3), 447-463.
106. Strili ć, B.; Eglinger, J.; Krieg, M.; Zeeb, M.; Axnick, J.; Bab ál, P.; M üller, D. J.; Lammert,
E., Electrostatic Cell-Surface Repulsion Initiates Lumen Formation in Developing Blood
Vessels. Current Biology 2010, 20 (22), 2003-2009.
107. Cha, T.; Guo, A.; Zhu, X. Y., Formation of Supported Phospholipid Bilayers on Molecular
Surfaces: Role of Surface Charge Density and Electrostatic Interaction. Biophys. J. 2006, 90 (4),
1270-1274.
108. Liu, J.; Jiang, X.; Ashley, C.; Brinker, C. J., Electrostatically Mediated Liposome Fusion
and Lipid Exchange with a Nanoparticle-Supported Bilayer for Control of Surface Charge,
Drug Containment, and Delivery. J. Am. Chem. Soc. 2009, 131 (22), 7567-7569.
109. Ashley, C. E.; Carnes, E. C.; Phillips, G. K.; Padilla, D.; Durfee, P. N.; Brown, P. A.;
Hanna, T. N.; Liu, J.; Phillips, B.; Carter, M. B.; Carroll, N. J.; Jiang, X.; Dunphy, D. R.;
Willman, C. L.; Petsev, D. N.; Evans, D. G.; Parikh, A. N.; Chackerian, B.; Wharton, W.;
Peabody, D. S.; Brinker, C. J., The targeted delivery of multicomponent cargos to cancer cells
by nanoporous particle-supported lipid bilayers. Nat Mater 2011, 10 (5), 389-397.
110. Cuenya, B. R., Synthesis and catalytic properties of metal nanoparticles: Size, shape, support,
composition, and oxidation state effects. Thin Solid Films 2010, 518 (12), 3127-3150.
111. Narayanan, R.; El-Sayed, M. A., Shape-Dependent Catalytic Activity of Platinum
Nanoparticles in Colloidal Solution. Nano Lett. 2004, 4 (7), 1343-1348.
94
112. Roberts, E. J.; Karadaghi, L. R.; Wang, L.; Malmstadt, N.; Brutchey, R. L., Continuous
Flow Methods of Fabricating Catalytically Active Metal Nanoparticles. ACS Applied Materials
& Interfaces 2019, 11 (31), 27479-27502.
113. Nakamura, H.; Yamaguchi, Y.; Miyazaki, M.; Maeda, H.; Uehara, M.; Mulvaney, P.,
Preparation of CdSe nanocrystals in a micro-flow-reactor. Chem. Commun. 2002, (23), 2844-
2845.
114. Wang, H.; Nakamura, H.; Uehara, M.; Miyazaki, M.; Maeda, H., Preparation of titania
particles utilizing the insoluble phase interface in a microchannel reactor. Chem. Commun. 2002,
(14), 1462-1463.
115. Shestopalov, I.; Tice, J. D.; Ismagilov, R. F., Multi-step synthesis of nanoparticles performed
on millisecond time scale in a microfluidic droplet-based system. Lab on a Chip 2004, 4 (4),
316-321.
116. Chan, E. M.; Mathies, R. A.; Alivisatos, A. P., Size-Controlled Growth of CdSe Nanocrystals
in Microfluidic Reactors. Nano Lett. 2003, 3 (2), 199-201.
117. Lohse, S. E.; Eller, J. R.; Sivapalan, S. T.; Plews, M. R.; Murphy, C. J., A Simple Millifluidic
Benchtop Reactor System for the High-Throughput Synthesis and Functionalization of Gold
Nanoparticles with Different Sizes and Shapes. ACS Nano 2013, 7 (5), 4135-4150.
118. Vikram, A.; Kumar, V.; Ramesh, U.; Balakrishnan, K.; Oh, N.; Deshpande, K.; Ewers,
T.; Trefonas, P.; Shim, M.; Kenis, P. J. A., A Millifluidic Reactor System for Multistep
Continuous Synthesis of InP/ZnSeS Nanoparticles. ChemNanoMat 2018, 4 (9), 943-953.
119. Jin, H. D.; Chang, C.-H., Continuous synthesis of SnTe nanorods. J. Mater. Chem. 2011, 21
(33), 12218-12220.
120. Gottesman, R.; Tangy, A.; Oussadon, I.; Zitoun, D., Silver nanowires and nanoparticles
from a millifluidic reactor: application to metal assisted silicon etching. New J. Chem. 2012, 36
(12), 2456-2459.
121. Solsona, M.; Vollenbroek, J. C.; Tregouet, C. B. M.; Nieuwelink, A. E.; Olthuis, W.; van
den Berg, A.; Weckhuysen, B. M.; Odijk, M., Microfluidics and catalyst particles. Lab on a
Chip 2019, 19 (21), 3575-3601.
122. Roberts, E. J.; Habas, S. E.; Wang, L.; Ruddy, D. A.; White, E. A.; Baddour, F. G.; Griffin,
M. B.; Schaidle, J. A.; Malmstadt, N.; Brutchey, R. L., High-Throughput Continuous Flow
Synthesis of Nickel Nanoparticles for the Catalytic Hydrodeoxygenation of Guaiacol. ACS
Sustainable Chemistry & Engineering 2017, 5 (1), 632-639.
123. Ruddy, D. A.; Schaidle, J. A.; Ferrell Iii, J. R.; Wang, J.; Moens, L.; Hensley, J. E., Recent
advances in heterogeneous catalysts for bio-oil upgrading via “ex situ catalytic fast pyrolysis”:
catalyst development through the study of model compounds. Green Chemistry 2014, 16 (2),
454-490.
124. Bridgwater, A. V., Upgrading biomass fast pyrolysis liquids. Environmental Progress &
Sustainable Energy 2012, 31 (2), 261-268.
95
125. Venkatakrishnan, V. K.; Delgass, W. N.; Ribeiro, F. H.; Agrawal, R., Oxygen removal from
intact biomass to produce liquid fuel range hydrocarbons via fast-hydropyrolysis and vapor-
phase catalytic hydrodeoxygenation. Green Chemistry 2015, 17 (1), 178-183.
126. Griffin, M. B.; Baddour, F. G.; Habas, S. E.; Ruddy, D. A.; Schaidle, J. A., Evaluation of
silica-supported metal and metal phosphide nanoparticle catalysts for the hydrodeoxygenation
of guaiacol under ex situ catalytic fast pyrolysis conditions. Top. Catal. 2016, 59 (1), 124-137.
127. Carenco, S.; Boissière, C.; Nicole, L.; Sanchez, C.; Le Floch, P.; Mézailles, N., Controlled
Design of Size-Tunable Monodisperse Nickel Nanoparticles. Chem. Mater. 2010, 22 (4), 1340-
1349.
128. Yen, B. K. H.; Günther, A.; Schmidt, M. A.; Jensen, K. F.; Bawendi, M. G., A
microfabricated gas –liquid segmented flow reactor for high ‐temperature synthesis: the case of
CdSe quantum dots. Angew. Chem. Int. Ed. 2005, 44 (34), 5447-5451.
129. Song, H.; Tice, J. D.; Ismagilov, R. F., A microfluidic system for controlling reaction
networks in time. Angew. Chem. Int. Ed. 2003, 42 (7), 768-772.
130. Baddour, F. G.; Roberts, E. J.; To, A. T.; Wang, L.; Habas, S. E.; Ruddy, D. A.; Bedford,
N. M.; Wright, J.; Nash, C. P.; Schaidle, J. A.; Brutchey, R. L.; Malmstadt, N., An
Exceptionally Mild and Scalable Solution-Phase Synthesis of Molybdenum Carbide
Nanoparticles for Thermocatalytic CO2 Hydrogenation. J. Am. Chem. Soc. 2020, 142 (2), 1010-
1019.
131. Porosoff, M. D.; Yan, B.; Chen, J. G., Catalytic reduction of CO2 by H2 for synthesis of CO,
methanol and hydrocarbons: challenges and opportunities. Energy & Environmental Science
2016, 9 (1), 62-73.
132. Oyama, S. T. e., Chapter 1. In The Chemistry of Transition Metal Carbides and Nitrides, Springer:
Netherland, 1996.
133. Lin, Z.; Chen, R.; Qu, Z.; Chen, J. G., Hydrodeoxygenation of biomass-derived oxygenates
over metal carbides: from model surfaces to powder catalysts. Green Chemistry 2018, 20 (12),
2679-2696.
134. Nash, C.; Yung, M.; Chen, Y.; Carl, S.; Thompson, L.; Schaidle, J., Catalysis by Metal
Carbides and Nitrides. In Handbook of Solid State Chemistry, Dronskowski, R.; Kikkawa, S.;
Stein, A., Eds.
135. Wang, L.; Karadaghi, L.; Brutchey, R. L.; Malmstadt, N., Self-Optimizing Parallel
Millifluidic Reactor for Scaling Nanoparticle Synthesis. Chem. Commun. 2020.
136. Pitkethly, M. J., Nanomaterials – the driving force. Mater. Today 2004, 7 (12, Supplement),
20-29.
137. Lignos, I.; Maceiczyk, R.; deMello, A. J., Microfluidic Technology: Uncovering the
Mechanisms of Nanocrystal Nucleation and Growth. Acc. Chem. Res. 2017, 50 (5), 1248-1257.
138. Phillips, T. W.; Lignos, I. G.; Maceiczyk, R. M.; deMello, A. J.; deMello, J. C., Nanocrystal
synthesis in microfluidic reactors: where next? Lab on a Chip 2014, 14 (17), 3172-3180.
96
139. Nisisako, T.; Ando, T.; Hatsuzawa, T., High-volume production of single and compound
emulsions in a microfluidic parallelization arrangement coupled with coaxial annular world-
to-chip interfaces. Lab on a Chip 2012, 12 (18), 3426-3435.
140. Yadavali, S.; Lee, D.; Issadore, D., Robust Microfabrication of Highly Parallelized Three-
Dimensional Microfluidics on Silicon. bioRxiv 2019, 625277.
141. Nightingale, A. M.; Bannock, J. H.; Krishnadasan, S. H.; O'Mahony, F. T. F.; Haque, S.
A.; Sloan, J.; Drury, C.; McIntyre, R.; deMello, J. C., Large-scale synthesis of nanocrystals
in a multichannel droplet reactor. Journal of Materials Chemistry A 2013, 1 (12), 4067-4076.
142. Mateos, C.; Nieves-Remacha, M. J.; Rincón, J. A., Automated platforms for reaction self-
optimization in flow. Reaction Chemistry & Engineering 2019, 4 (9), 1536-1544.
143. Hashimoto, M.; Shevkoplyas, S. S.; Zaso ńska, B.; Szymborski, T.; Garstecki, P.; Whitesides,
G. M., Formation of Bubbles and Droplets in Parallel, Coupled Flow-Focusing Geometries.
Small 2008, 4 (10), 1795-1805.
144. Holland, F. A.; Bragg, R., 7 - Gas–liquid two-phase flow. In Fluid Flow for Chemical Engineers
(Second Edition), Holland, F. A.; Bragg, R., Eds. Butterworth-Heinemann: Oxford, 1995; pp
219-267.
145. Kulkarni, S. A.; Mhaisalkar, S. G.; Mathews, N.; Boix, P. P., Perovskite Nanoparticles:
Synthesis, Properties, and Novel Applications in Photovoltaics and LEDs. Small Methods 2019,
3 (1), 1800231.
146. Lignos, I.; Stavrakis, S.; Nedelcu, G.; Protesescu, L.; deMello, A. J.; Kovalenko, M. V.,
Synthesis of Cesium Lead Halide Perovskite Nanocrystals in a Droplet-Based Microfluidic
Platform: Fast Parametric Space Mapping. Nano Lett. 2016, 16 (3), 1869-1877.
147. Wei, S.; Yang, Y.; Kang, X.; Wang, L.; Huang, L.; Pan, D., Room-temperature and gram-
scale synthesis of CsPbX3 (X = Cl, Br, I) perovskite nanocrystals with 50-85%
photoluminescence quantum yields. Chem. Commun. 2016, 52 (45), 7265-7268.
148. Guermonprez, C.; Michelin, S.; Baroud, C. N., Flow distribution in parallel microfluidic
networks and its effect on concentration gradient. Biomicrofluidics 2015, 9 (5), 054119.
149. Bhargava, K. C.; Thompson, B.; Malmstadt, N., Discrete elements for 3D microfluidics.
Proceedings of the National Academy of Sciences 2014, 111 (42), 15013.
150. Nelder, J. A.; Mead, R., A Simplex Method for Function Minimization. The Computer Journal
1965, 7 (4), 308-313.
151. Cottingham, P.; Brutchey, R. L., On the crystal structure of colloidally prepared CsPbBr3
quantum dots. Chem. Commun. 2016, 52 (30), 5246-5249.
152. Brennan, M. C.; Herr, J. E.; Nguyen-Beck, T. S.; Zinna, J.; Draguta, S.; Rouvimov, S.;
Parkhill, J.; Kuno, M., Origin of the Size-Dependent Stokes Shift in CsPbBr3 Perovskite
Nanocrystals. J. Am. Chem. Soc. 2017, 139 (35), 12201-12208.
97
153. Wang, L.; Malmstadt, N., Interactions between charged nanoparticles and giant vesicles
fabricated from inverted-headgroup lipids. J. Phys. D: Appl. Phys. 2017, 50 (41), 415402.
154. Kwolek, U.; Jamroz, D.; Janiczek, M.; Nowakowska, M.; Wydro, P.; Kepczynski, M.,
Interactions of Polyethylenimines with Zwitterionic and Anionic Lipid Membranes. Langmuir
2016, 32 (19), 5004-5018.
155. Zhang, L.; Granick, S., How to Stabilize Phospholipid Liposomes (Using Nanoparticles).
Nano Lett. 2006, 6 (4), 694-698.
156. Motion, J. P. M.; Nguyen, J.; Szoka, F. C., Phosphatase-Triggered Fusogenic Liposomes for
Cytoplasmic Delivery of Cell-Impermeable Compounds. Angew. Chem. 2012, 124 (36), 9181-
9185.
157. Yu, X.; Liu, Z.; Janzen, J.; Chafeeva, I.; Horte, S.; Chen, W.; Kainthan, R. K.;
Kizhakkedathu, J. N.; Brooks, D. E., Polyvalent choline phosphate as a universal
biomembrane adhesive. Nat Mater 2012, 11 (5), 468-476.
158. Yu, X.; Yang, X.; Horte, S.; Kizhakkedathu, J. N.; Brooks, D. E., A pH and thermosensitive
choline phosphate-based delivery platform targeted to the acidic tumor microenvironment.
Biomaterials 2014, 35 (1), 278-286.
159. Perttu, E. K.; Kohli, A. G.; Szoka, F. C., Inverse-phosphocholine lipids: A remix of a
common phospholipid. J. Am. Chem. Soc. 2012, 134 (10), 4485-4488.
160. Wang, F.; Liu, J., A Stable Lipid/TiO2 Interface with Headgroup-Inversed Phosphocholine
and a Comparison with SiO2. J. Am. Chem. Soc. 2015, 137 (36), 11736-11742.
161. Castellan, N. J., Nonparametric statistics for the behavioral sciences. McGraw-Hill: 1988.
162. Damodaran, K. V.; Merz, K. M., Head group-water interactions in lipid bilayers: a
comparison between DMPC- and DLPE-based lipid bilayers. Langmuir 1993, 9 (5), 1179-1183.
163. Liwo, A., Computational Methods to Study the Structure and Dynamics of Biomolecules and
Biomolecular Processes: From Bioinformatics to Molecular Quantum Mechanics. Springer Science &
Business Media: 2013; Vol. 1.
164. Pasenkiewicz-Gierula, M.; Takaoka, Y.; Miyagawa, H.; Kitamura, K.; Kusumi, A.,
Hydrogen bonding of water to phosphatidylcholine in the membrane as studied by a molecular
dynamics simulation: location, geometry, and lipid-lipid bridging via hydrogen-bonded water.
The Journal of Physical Chemistry A 1997, 101 (20), 3677-3691.
165. Goodman, C. M.; McCusker, C. D.; Yilmaz, T.; Rotello, V. M., Toxicity of gold
nanoparticles functionalized with cationic and anionic side chains. Bioconjugate Chem. 2004,
15 (4), 897-900.
166. Yacobi, N. R.; Malmstadt, N.; Fazlollahi, F.; DeMaio, L.; Marchelletta, R.; Hamm-
Alvarez, S. F.; Borok, Z.; Kim, K.-J.; Crandall, E. D., Mechanisms of alveolar epithelial
translocation of a defined population of nanoparticles. American journal of respiratory cell and
molecular biology 2010, 42 (5), 604-614.
98
167. Shang, L.; Nienhaus, K.; Nienhaus, G. U., Engineered nanoparticles interacting with cells:
size matters. J Nanobiotechnology 2014, 12 (1), 5.
168. Yen, H. J.; Hsu, S. h.; Tsai, C. L., Cytotoxicity and immunological response of gold and
silver nanoparticles of different sizes. Small 2009, 5 (13), 1553-1561.
169. Ivask, A.; Kurvet, I.; Kasemets, K.; Blinova, I.; Aruoja, V.; Suppi, S.; Vija, H.; Käkinen,
A.; Titma, T.; Heinlaan, M., Size-dependent toxicity of silver nanoparticles to bacteria, yeast,
algae, crustaceans and mammalian cells in vitro. PloS one 2014, 9 (7), e102108.
170. Zhang, S.; Nelson, A.; Beales, P. A., Freezing or wrapping: the role of particle size in the
mechanism of nanoparticle–biomembrane interaction. Langmuir 2012, 28 (35), 12831-12837.
171. Liu, J.; Stace-Naughton, A.; Jiang, X.; Brinker, C. J., Porous Nanoparticle Supported Lipid
Bilayers (Protocells) as Delivery Vehicles. J. Am. Chem. Soc. 2009, 131 (4), 1354-1355.
172. Rocha, E. L. d.; Caramori, G. F.; Rambo, C. R., Nanoparticle translocation through a lipid
bilayer tuned by surface chemistry. PCCP 2013, 15 (7), 2282-2290.
173. Bahrami, A. H.; Raatz, M.; Agudo-Canalejo, J.; Michel, R.; Curtis, E. M.; Hall, C. K.;
Gradzielski, M.; Lipowsky, R.; Weikl, T. R., Wrapping of nanoparticles by membranes. Adv.
Colloid Interface Sci. 2014, 208, 214-224.
174. Stachowiak, J. C.; Hayden, C. C.; Sasaki, D. Y., Steric confinement of proteins on lipid
membranes can drive curvature and tubulation. Proceedings of the National Academy of Sciences
2010, 107 (17), 7781-7786.
175. Fischer, T. M., Bending stiffness of lipid bilayers. I. Bilayer couple or single-layer bending?
Biophys. J. 1992, 63 (5), 1328-1335.
176. Magarkar, A.; Róg, T.; Bunker, A., Molecular Dynamics Simulation of Inverse-
Phosphocholine Lipids. The Journal of Physical Chemistry C 2014, 118 (33), 19444-19449.
177. Gonzales, R. C.; Wintz, P., Digital image processing (2nd ed.). Addison-Wesley Longman
Publishing Co., Inc.: 1987; p 503.
178. Wang, L.; Hartel, N.; Kaixuan, R.; Graham, N.; Malmstadt, N., Effect of protein corona
on nanoparticle-plasma membrane and nanoparticle-biomimetic membrane interactions.
Environmental Science: Nano 2020.
179. Fröhlich, E., The role of surface charge in cellular uptake and cytotoxicity of medical
nanoparticles. International journal of nanomedicine 2012, 7, 5577.
180. Behzadi, S.; Serpooshan, V.; Tao, W.; Hamaly, M. A.; Alkawareek, M. Y.; Dreaden, E.
C.; Brown, D.; Alkilany, A. M.; Farokhzad, O. C.; Mahmoudi, M., Cellular uptake of
nanoparticles: journey inside the cell. Chem. Soc. Rev. 2017, 46 (14), 4218-4244.
181. Mahmoudi, M.; Lynch, I.; Ejtehadi, M. R.; Monopoli, M. P.; Bombelli, F. B.; Laurent, S.,
Protein− nanoparticle interactions: opportunities and challenges. Chem. Rev. 2011, 111 (9),
5610-5637.
99
182. Corbo, C.; Molinaro, R.; Parodi, A.; Toledano Furman, N. E.; Salvatore, F.; Tasciotti, E.,
The impact of nanoparticle protein corona on cytotoxicity, immunotoxicity and target drug
delivery. Nanomedicine 2016, 11 (1), 81-100.
183. Lesniak, A.; Fenaroli, F.; Monopoli, M. P.; Åberg, C.; Dawson, K. A.; Salvati, A., Effects
of the presence or absence of a protein corona on silica nanoparticle uptake and impact on
cells. ACS nano 2012, 6 (7), 5845-5857.
184. Tenzer, S.; Docter, D.; Kuharev, J.; Musyanovych, A.; Fetz, V.; Hecht, R.; Schlenk, F.;
Fischer, D.; Kiouptsi, K.; Reinhardt, C., Rapid formation of plasma protein corona critically
affects nanoparticle pathophysiology. Nature nanotechnology 2013, 8 (10), 772.
185. Obst, K.; Yealland, G.; Balzus, B.; Miceli, E.; Dimde, M.; Weise, C.; Eravci, M.; Bodmeier,
R.; Haag, R.; Calderón, M., Protein corona formation on colloidal polymeric nanoparticles
and polymeric nanogels: impact on cellular uptake, toxicity, immunogenicity, and drug release
properties. Biomacromolecules 2017, 18 (6), 1762-1771.
186. Deng, Z. J.; Liang, M.; Monteiro, M.; Toth, I.; Minchin, R. F., Nanoparticle-induced
unfolding of fibrinogen promotes Mac-1 receptor activation and inflammation. Nature
nanotechnology 2011, 6 (1), 39.
187. Yan, Y.; Gause, K. T.; Kamphuis, M. M.; Ang, C.-S.; O’Brien-Simpson, N. M.; Lenzo, J.
C.; Reynolds, E. C.; Nice, E. C.; Caruso, F., Differential roles of the protein corona in the
cellular uptake of nanoporous polymer particles by monocyte and macrophage cell lines. ACS
nano 2013, 7 (12), 10960-10970.
188. Sayes, C. M.; Fortner, J. D.; Guo, W.; Lyon, D.; Boyd, A. M.; Ausman, K. D.; Tao, Y.
J.; Sitharaman, B.; Wilson, L. J.; Hughes, J. B., The differential cytotoxicity of water-soluble
fullerenes. Nano Lett. 2004, 4 (10), 1881-1887.
189. Smith, P. J.; Giroud, M.; Wiggins, H. L.; Gower, F.; Thorley, J. A.; Stolpe, B.; Mazzolini,
J.; Dyson, R. J.; Rappoport, J. Z., Cellular entry of nanoparticles via serum sensitive clathrin-
mediated endocytosis, and plasma membrane permeabilization. International journal of
nanomedicine 2012, 7, 2045.
190. Walde, P.; Cosentino, K.; Engel, H.; Stano, P., Giant Vesicles: Preparations and
Applications. ChemBioChem 2010, 11 (7), 848-865.
191. Fenz, S. F.; Sengupta, K., Giant vesicles as cell models. Integrative Biology 2012, 4 (9), 982-995.
192. Czogalla, A.; Grzybek, M.; Jones, W.; Coskun, U., Validity and applicability of membrane
model systems for studying interactions of peripheral membrane proteins with lipids.
Biochimica et Biophysica Acta (BBA)-Molecular and Cell Biology of Lipids 2014, 1841 (8), 1049-1059.
193. Zhang, M.; Wei, X.; Ding, L.; Hu, J.; Jiang, W., Adhesion of CdTe quantum dots on model
membranes and internalization into RBL-2H3 cells. Environ. Pollut. 2017, 225, 419-427.
194. Montis, C.; Generini, V.; Boccalini, G.; Bergese, P.; Bani, D.; Berti, D., Model lipid bilayers
mimic non-specific interactions of gold nanoparticles with macrophage plasma membranes. J.
Colloid Interface Sci. 2018, 516, 284-294.
100
195. Gessner, A.; Lieske, A.; Paulke, B. R.; Müller, R. H., Functional groups on polystyrene
model nanoparticles: influence on protein adsorption. Journal of Biomedical Materials Research
Part A: An Official Journal of The Society for Biomaterials, The Japanese Society for Biomaterials, and
The Australian Society for Biomaterials and the Korean Society for Biomaterials 2003, 65 (3), 319-326.
196. Hadjidemetriou, M.; Al-Ahmady, Z.; Mazza, M.; Collins, R. F.; Dawson, K.; Kostarelos,
K., In vivo biomolecule corona around blood-circulating, clinically used and antibody-
targeted lipid bilayer nanoscale vesicles. ACS nano 2015, 9 (8), 8142-8156.
197. Lundqvist, M.; Stigler, J.; Elia, G.; Lynch, I.; Cedervall, T.; Dawson, K. A., Nanoparticle
size and surface properties determine the protein corona with possible implications for
biological impacts. Proceedings of the National Academy of Sciences 2008, 105 (38), 14265-14270.
198. Cedervall, T.; Lynch, I.; Foy, M.; Berggård, T.; Donnelly, S. C.; Cagney, G.; Linse, S.;
Dawson, K. A., Detailed identification of plasma proteins adsorbed on copolymer
nanoparticles. Angew. Chem. Int. Ed. 2007, 46 (30), 5754-5756.
199. Zhang, H.; Wei, X.; Liu, L.; Zhang, Q.; Jiang, W., The role of positively charged sites in
the interaction between model cell membranes and γ-Fe2O3 NPs. Sci. Total Environ. 2019, 673,
414-423.
200. Wang, B.; Zhang, L.; Bae, S. C.; Granick, S., Nanoparticle-induced surface reconstruction
of phospholipid membranes. Proceedings of the National Academy of Sciences 2008, 105 (47), 18171.
201. Hussain, S. M.; Hess, K. L.; Gearhart, J. M.; Geiss, K. T.; Schlager, J. J., In vitro toxicity
of nanoparticles in BRL 3A rat liver cells. Toxicology in Vitro 2005, 19 (7), 975-983.
202. Kroll, A.; Pillukat, M. H.; Hahn, D.; Schnekenburger, J., Interference of engineered
nanoparticles with in vitro toxicity assays. Arch. Toxicol. 2012, 86 (7), 1123-1136.
203. Mirshafiee, V.; Kim, R.; Mahmoudi, M.; Kraft, M. L., The importance of selecting a proper
biological milieu for protein corona analysis in vitro: Human plasma versus human serum.
The International Journal of Biochemistry & Cell Biology 2016, 75, 188-195.
204. Forest, V.; Cottier, M.; Pourchez, J., Electrostatic interactions favor the binding of positive
nanoparticles on cells: A reductive theory. Nano Today 2015, 10 (6), 677-680.
205. Keller, H.; Lorizate, M.; Schwille, P., PI (4, 5) P2 degradation promotes the formation of
cytoskeleton ‐free model membrane systems. Chemphyschem 2009, 10 (16), 2805-2812.
206. Nazemidashtarjandi, S.; Farnoud, A. M., Membrane outer leaflet is the primary regulator of
membrane damage induced by silica nanoparticles in vesicles and erythrocytes. Environmental
Science: Nano 2019, 6 (4), 1219-1232.
207. Maktabi, S.; Schertzer, J. W.; Chiarot, P. R., Dewetting-induced formation and mechanical
properties of synthetic bacterial outer membrane models (GUVs) with controlled inner-leaflet
lipid composition. Soft matter 2019, 15 (19), 3938-3948.
208. Alkhammash, H. I.; Li, N.; Berthier, R.; de Planque, M. R., Native silica nanoparticles are
powerful membrane disruptors. PCCP 2015, 17 (24), 15547-15560.
101
209. Bischof, J. C.; Padanilam, J.; Holmes, W. H.; Ezzell, R. M.; Lee, R.; Tompkins, R. G.;
Yarmush, M. L.; Toner, M., Dynamics of cell membrane permeability changes at
supraphysiological temperatures. Biophys. J. 1995, 68 (6), 2608-2614.
210. Baykal-Caglar, E.; Hassan-Zadeh, E.; Saremi, B.; Huang, J., Preparation of giant unilamellar
vesicles from damp lipid film for better lipid compositional uniformity. Biochimica et Biophysica
Acta (BBA)-Biomembranes 2012, 1818 (11), 2598-2604.
211. Gray, E. M.; Díaz-Vázquez, G.; Veatch, S. L., Growth conditions and cell cycle phase
modulate phase transition temperatures in RBL-2H3 derived plasma membrane vesicles. PloS
one 2015, 10 (9), e0137741.
212. Kuboi, R.; Shimanouchi, T.; Yoshimoto, M.; Umakoshi, H., Detection of protein
conformation under stress conditions using liposomes as sensor materials. Sensors and Materials
2004, 16 (5), 241-254.
213. Koynova, R.; MacDonald, R. C., Natural lipid extracts and biomembrane-mimicking lipid
compositions are disposed to form nonlamellar phases, and they release DNA from lipoplexes
most efficiently. Biochimica et Biophysica Acta (BBA)-Biomembranes 2007, 1768 (10), 2373-2382.
214. Mach á ň, R.; Hof, M., Lipid diffusion in planar membranes investigated by fluorescence
correlation spectroscopy. Biochimica et Biophysica Acta (BBA) - Biomembranes 2010, 1798 (7),
1377-1391.
215. Sankhagowit, S.; Wu, S.-H.; Biswas, R.; Riche, C. T.; Povinelli, M. L.; Malmstadt, N., The
dynamics of giant unilamellar vesicle oxidation probed by morphological transitions.
Biochimica et Biophysica Acta (BBA) - Biomembranes 2014, 1838 (10), 2615-2624.
216. Runas, K. A.; Malmstadt, N., Low levels of lipid oxidation radically increase the passive
permeability of lipid bilayers. Soft Matter 2015, 11 (3), 499-505.
217. Xia, Z.; Woods, A.; Quirk, A.; Burgess, I. J.; Lau, B. L. T., Interactions between polystyrene
nanoparticles and supported lipid bilayers: impact of charge and hydrophobicity modification
by specific anions. Environmental Science: Nano 2019, 6 (6), 1829-1837.
218. Jacobson, K., Lateral diffusion in membranes. Cell motility 1983, 3 (5), 367-373.
219. Przybylo, M.; Sýkora, J.; Humpolíčková, J.; Benda, A.; Zan, A.; Hof, M., Lipid diffusion
in giant unilamellar vesicles is more than 2 times faster than in supported phospholipid bilayers
under identical conditions. Langmuir 2006, 22 (22), 9096-9099.
220. Farnoud, A. M.; Nazemidashtarjandi, S., Emerging investigator series: interactions of
engineered nanomaterials with the cell plasma membrane; what have we learned from
membrane models? Environmental Science: Nano 2019, 6 (1), 13-40.
221. Hughes, C. S.; Foehr, S.; Garfield, D. A.; Furlong, E. E.; Steinmetz, L. M.; Krijgsveld, J.,
Ultrasensitive proteome analysis using paramagnetic bead technology. Molecular Systems
Biology 2014, 10 (10), 757.
102
222. Weinberger, A.; Tsai, F.-C.; Koenderink, Gijsje H.; Schmidt, Thais F.; Itri, R.; Meier, W.;
Schmatko, T.; Schröder, A.; Marques, C., Gel-Assisted Formation of Giant Unilamellar
Vesicles. Biophys. J. 2013, 105 (1), 154-164.
223. Lira, Rafael B.; Dimova, R.; Riske, Karin A., Giant Unilamellar Vesicles Formed by Hybrid
Films of Agarose and Lipids Display Altered Mechanical Properties. Biophys. J. 2014, 107 (7),
1609-1619.
224. Hansen, J. S.; Thompson, J. R.; Hélix-Nielsen, C.; Malmstadt, N., Lipid Directed Intrinsic
Membrane Protein Segregation. J. Am. Chem. Soc. 2013, 135 (46), 17294-17297.
225. Gutierrez, M. G.; Malmstadt, N., Human Serotonin Receptor 5-HT1A Preferentially
Segregates to the Liquid Disordered Phase in Synthetic Lipid Bilayers. J. Am. Chem. Soc. 2014,
136 (39), 13530-13533.
Abstract (if available)
Abstract
The nanotechnology revolution arising in the last century has led to the synthesis and understanding of a variety of nanoparticles, bringing about an extraordinary potential in sustainable energy, health care, and materials development. To cope with the extensive demand of nanoparticle production, there have been various industrial-scale synthesis techniques facilitated by gas phase process. Nanoparticles are also commonly synthesized in liquid phase for more uniform size and shape. However, scale-up of liquid-phase synthesis suffers from irregularities in large-scale batch process, which can have huge negative effects on nanoparticle products, compromising their efficiency in later applications. High-throughput and well-controlled continuous flow reactor systems are promising approaches for colloidal nanoparticle fabrication to be transferred into industry. On the other hand, the development of nanotechnology has also created environmental contamination and unnecessary exposure to human health. A better understanding of nano-bio interfaces and interactions can provide more fundamental insight for the health effects of nanoparticles. ❧ This report begins with a general background introduction of nanotechnology development and its effect on the environment and human health. The research work in the following chapters focuses on nanoparticle synthesis in millifluidic reactors and interactions between nanoplastics and model biomembranes. In chapter 2, millifluidic syntheses of transition metal nanoparticle catalysts at high temperatures are presented. The superior heat and mass transfer in millifluidic reactors offered superior catalytical properties of nanoparticles with shorter reaction times. In chapter 3, parallelization scale-up of millifluidic reactors is introduced. With self-optimizing feedback control, the automation of this system is readily achieved. Chapter 4 and chapter 5 discuss the interaction of nano-bio interfaces, utilizing polystyrene nanoplastics and model biomembranes. Using model membranes composed of lipid molecules with different charge properties, the role of electrostatic forces in the non-specific interactions between nanoplastics and biomembranes are elucidated. Furthermore, nanoparticles are known to gather protein molecules upon contact with biological milieu, forming a protein corona on their surface, which affects their biological identity and fate. In the presence of a protein corona, we have observed a general decrease of electrostatic interactions between nanoparticles and biomembranes and membrane perturbation across in-vitro and model membranes.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Applications of continuous flow reactors towards screening catalytically active nanoparticle syntheses
PDF
Imaging molecular transport across and nanomaterial interaction with lipid membranes
PDF
Sustainable continuous flow syntheses of colloidal inorganic nanoparticle catalysts
PDF
Synthesis of high-quality nanoparticles using microfluidic platforms
PDF
Vapor phase deposition of dense and porous polymer coatings and membranes for increased sustainability and practical applications
PDF
Continuous flow synthesis of catalysts with custom made reactor with flow and temperature studies
PDF
Probing the effects of transmembrane domains on the continuum mechanics of lipid bilayers
PDF
Hybrid lipid-based nanostructures
PDF
Stability and folding rate of proteins and identification of their inhibitors
PDF
The physics of membrane protein polyhedra
PDF
Metabolomic and proteomic approaches to understanding senescence and aging in mammary epithelial cells
PDF
Methanol synthesis in the membrane reactor
PDF
Multi-scale biomimetic structure fabrication based on immersed surface accumulation
PDF
Engineering chimeric antigen receptor-directed immune cells for enhanced antitumor efficacy in solid tumors
PDF
Nano-fabricated devices in electrochemistry and cancer therapy
PDF
Formation of polymer gels, films, and particles via initiated chemical vapor deposition onto liquid substrates
PDF
Engineering immunotoxin and viral vectors for cancer therapy
PDF
Development of a novel heterogeneous flow reactor: soot formation and nanoparticle catalysis
PDF
Numerical and experimental investigations of dust-plasma-asteroid interactions
PDF
Fabrication and application of plasmonic nanostructures: a story of nano-fingers
Asset Metadata
Creator
Wang, Lu
(author)
Core Title
High-throughput nanoparticle fabrication and nano-biomembrane interactions
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Chemical Engineering
Publication Date
04/01/2020
Defense Date
02/21/2020
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
biomimetic membranes,electrostatic interactions,flow synthesis,nanoparticles,OAI-PMH Harvest,protein corona
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Malmstadt, Noah (
committee chair
), Brutchey, Richard (
committee member
), Graham, Nicholas (
committee member
)
Creator Email
wang737@usc.edu,wlantoine04@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-278651
Unique identifier
UC11673903
Identifier
etd-WangLu-8238.pdf (filename),usctheses-c89-278651 (legacy record id)
Legacy Identifier
etd-WangLu-8238.pdf
Dmrecord
278651
Document Type
Dissertation
Rights
Wang, Lu
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
biomimetic membranes
electrostatic interactions
flow synthesis
nanoparticles
protein corona