Supplemental Information
Supplemental Figure 1. (a) Centrifugal purification approach to exclude
microparticles, (b) Zeta-sizer PLL core-shell nanoparticle size distributions before
and after centrifugal pretreatment, (c) Representative post-centrifugal size
distributions for 220 nm and 400 nm PLL CSnps suspended in E-3 S/m solution,
and (d) Representative size post- centrifugal distributions in E-3 S/m and E-5 S/m
medium conductivities for 220 nm PLL CSnps.
To exclude extraneous microparticles that arose during synthesis, centrifugal separation
was compared to, then chosen over, filtration and sonication methods. Criteria used for
selection included shell integrity via zeta-sizer and TEM results (data not shown).
Centrifugal conditions including time and relative centrifugal force (rcf) were optimized
separately for each shell material. Supplemental Figure 1a and 1b illustrate the settling of
PLL microparticles (> 1um) upon centrifugation; the supernatant segment containing
nanoparticles was used for all DEP experiments. PLL CSnps were able to withstand 2500
rcf while chitosan CSnp were only able to withstand 100 rcf likely due to density and
sample viscosity differences. Both chitosan and PLL shells were relatively stable with
average zeta potentials above ±20 mV, which is sufficient to repel neighbouring
nanoparticles (Table 1).
In the uncorrected circuit, only nDEP was observed over all frequencies tested from
100kHz to 80MHz. The measured nDEP magnitudes increased from 100kHz up to
~62MHz and then nDEP decreased between ~62MHz and 80MHz. To verify the
observed phenomena, we tested CSnp (400nm PLL shell) at 60MHz and then turned off
the field to see natural particle dispersion. In addition, the frequency was switched from
60MHz to 100kHz, which illustrated weak nDEP, and the results are shown in
Supplemental Figure 2.
Supplemental Figure 2. Comparison between the free diffusion (a-c) and low
frequency at 100kHz (d-f) of core-shell nanoparticle polarization after 20 second of
field-on at 60MHz. 5Vpp at 60MHz was applied for 20 second and then free diffusion
at field-off (c) and low frequency field-on (f) were observed for 80 second after.
Qualitatively, the nDEP phenomena was verified by selectively changing the applied
frequency. These experiments were done because a single nanoparticle’s size is smaller
than the microscope optical limitation making pDEP harder to be distinguished next to
the light blocking metal electrodes. Therefore, the electric field was applied at 60 MHz,
which induced the maximum nDEP, and then switched to either (1) turn off the electric
field altogether (Figure 2 a, b, and c), or (2) change the applied frequency to 100kHz
(Figure 2 d, e, and f), whereby a minimum nDEP was observed. By comparing these two
cases and other frequencies, we qualitatively determined that pDEP was inadvertently
overlooked in the experimental data. If the CSnp did experience pDEP, then the CSnp
cluster at the quadrapole center (see b and e) would disperse more rapidly that via
diffusion (Figure 2 a, b, and c) to the electric field maxima (electrode edge) region. The
field off condition was tested because in this case, only free diffusion acted upon the
particles. In Figure 2 c, once the electric field was off, particle clusters became more
dispersed compared to the 60 MHz in Figure 2b. A slight hydrodynamic force resulted in
the system after turning off the electric field; this small flow caused the cluster to move
slightly to the right from the center. However, in Figure 2f, the cluster still remain at the
center which indicates no pDEP forces and possibly weak nDEP at 100 kHz. In fact,
cluster position remained at the center and displayed slightly reduced cluster size
compared to 60 MHz.
Device geometry effect was additionally explored with the uncorrected circuit. Three
device geometries were tested including quadrapole electrodes with (a) 25 μm spacing,
(b) 200 μm spacing, and (c) T-configuration electrodes with 100 μm spacing. The gas
core-PLL shell nanoparticles with mean diameter of 400nm displayed similar maximum
peak responses near 60MHz as shown in Supplemental Figure 4. This indicates that the
CSnps’ unusual maximum peak nDEP responses are due to CSnp properties, not due to
the device geometry (25 μm vs. 200 μm) nor due to device geometry (quadrapole vs. Tconfiguration electrodes).
0.0
Arbitary Value
-0.2
-0.4
25um gap quadrapole
electrodes
200um gap quadrapole
electrodes
-0.6
T-device w/ 100um gap
-0.8
-1.0
1.E+05
2.E+07
4.E+07
6.E+07
Frequency (Hz)
8.E+07
Supplemental Figure 3. Device effect on CSnp (gas core-PLL shell nanoparticles) on
measured nDEP responses as a function of frequency at 5Vpp.
In Supplemental Figure 5, a control experiment was conducted by a reviewer’s comment
to conclude further on the effect of the gas core. In order to compare the gas core effect
with solid core material, we chose polystyrene beads (~300nm) coated with PLL shell
materials. The measured nDEP from solid core displayed similar responses that included
maximum nDEP peak, but the maximum frequency switched to 70MHz. Because DEP
responses are determined by dielectric properties (permittivity & conductivity) and both
gas core and the solid PS beads are insulator. Thus, the resulting phenomena were
similar. However, the altered maximum peak might be an indicator that showed
difference between the solid PS and the gas core.
0.0
Arbitary Value
-0.2
-0.4
400nm PLL
-0.6
300nm PS beads coated with
PLL
-0.8
-1.0
1.E+05
2.E+07
4.E+07
6.E+07
Frequency (Hz)
8.E+07
Supplemental Figure 4. Comparison between the gas core and solid polystyrene (PS)
core materials within similar size ranges. Measured nDEP responses as a function of
frequency at 5Vpp within 25 μm gap quadrapole electrodes .