Supporting information ACS Nano

Full text

1
Supporting Information:
Simultaneous Ionic Current and Potential Detection of
Nanoparticles by A Multifunctional Nanopipette
Namuna Panday,+ Gongming Qian,‡ Xuewen Wang,+ Shuai Chang,§ Popular Pandey,+ Jin He*+#
+ Physics Department, Florida International University, Miami FL 33199, United States
‡College of Resources and Environmental Engineering, Wuhan University of Science and
Technology, Wuhan, Hubei, 430081, China
§College of Materials and Metallurgy, Wuhan University of Science and Technology, Wuhan,
Hubei, 430081, China
# Biomolecular Science Institute, Florida International University, Miami FL 33199, United
States
S1. The nanopore size estimation
As reported previously,1, 2 the surface charge effect of a conical shape nanopore can be ignored at
the small bias range (i.e., V<kBT/e, where kB is the Boltzmann constant, T is the temperature,
and e is the elementary electron charge). A simple analytical equation shown below is used for
nanopore estimation:
  

 (S1)

2
Rp is the nanopore resistance and κ is the conductivity of the electrolyte. The Rp is determined
from the IV measurements at 10mM PBS. The κ for 10 mM PBS (at pH 7.4) is determined to be
1312µS/cm from conductivity measurements. The mean Rp before carbon deposition was 2.22 ±
0.53 GΩ and the value was increased to 3.00±1.16 GΩ after carbon deposition. The half cone angle
θ of the nanopipette was measured to be θ=6.5±0.5o based on five SEM images like Figure 1b. It
should be noted that the inner half cone angle is smaller and this leads to an underestimation of the
nanopore size. Therefore Equation S1 only gave a crude estimation for the nanopore size.
S2. CNE fabrication and electrochemical characterization
The setup is shown in Figure S1a. The back of one barrel is blocked with a removable
plug (Blu-tack) to prevent carbon deposition. Home-built precision pressure meters are used to
monitor the argon and butane gas flow. To prepare the nanopipette geometry described in Figure
1, the pressures for argon and butane gas flow are 0.5 kPa and 25kPa respectively. The carbon
deposition is very fast and normally less than 1 minute. The success yield is high. We also tested
other pressures for butane gas. When the pressure for butane flow is low (<20 kPa), there are
often a large number of voids in the CNE and the CNEs are highly resistive. At pressure between
20 kPa and 25 kPa, the CNE often does not protrude out of the nanopipette tip. When the
pressure for butane flow is too high (>30kpa), the CNE is significantly overgrown at the
nanopipette tip and the nanopore is often entirely blocked. At pressure 25kPa, the carbon can
fully fill the tip region barrel (Figure S1b) and also protrudes out of the tip slightly (Figure S1c-
d).

3
Figure S1. (a) The Schematic setup for the fabrication of CNE from the theta nanopipette. (b) The cross-
section SEM image of a broken nanopore/CNE nanopipette. (c) The SEM image for a side view of the
nanopipette in Figure 1(c). (d) The SEM image of the tip section of another nanopipette
We typically observed sigmoidal shaped steady-state CVs for the fabricated CNEs as
shown in Figure S2. The diffusion limited current id of the CV was used to evaluate the CNE
size. The CV of the CNE was collected in 1x PBS solution containing 1mM Ru(NH3)6+ ions by
cycling the electrode potentials at 20 mV/s using a potentiostat (CHI760D, CH Instruments, Inc.,
USA).
The size of CNEs can be calculated:
 = mFDC , (S2)
where m is a geometry factor, F is the Faraday constant (96485 C/mol), D and C are the diffusion
constant (7.4x10-6 cm2/s) and the bulk concentration of Ru(NH3)6+ ions. The geometry factor m=1
if the electrode is hemispherical and m changes slightly (normally less than 10%) for other
geometries. For example, m=1.1 if the aspect ratio of an oblate hemispheroid is 6. The SEM images
revealed the aspect ratios of CNEs were in the range of 2-4. We, therefore, used m=1 for the
estimation of Aeff.

4
Figure S2 The steady-state CVs (at a sweep rate 20 mV/s) for 18 CNEs in 1x PBS solution containing
1mM Ru(NH3)6+ ions.
S3. Characteristics of P1, P2 and P3
Figure S3. IVs in a small bias range (a) and CVs (b) for multifunctional nanopipettes P1, P2 and P3. IVs
were recorded in 10 mM PBS (pH 7.4). The square symbol represents the experimental data and the
straight lines are fitted lines. CVs were recorded in 1x PBS solution containing 1mM Ru(NH3)6+ ions at a
sweep rate 20mV/s. The black, red and green curves represent IVs and CVs for P1, P2 and P3
respectively.
Based on the IV measurements of nanopores (after carbon fabrication) and CV
measurements of CNEs (Figure S3), we derived the nanopore diameter is about 85±7 nm, 64±5
nm and 60±5 nm and the CNE effective surface area is 0.30±0.02µm2, 1.33±0.09µm2 and
0.30±0.02µm2 for P1, P2 and P3 respectively.

5
S4. Zeta potential of 40 nm GNPs
The zeta potential measurement of 40 nm GNPs was performed by Zetasizer nano–ZS
(Zen 3600, Malvern Instruments Ltd.) using a folded capillary cell (Catalog # DTS1070) at room
temperature. Each sample was analyzed six times. The zeta potential of GNP depends on
solution salt concentration. The zeta potential of 40 nm GNPs in10 mM PBS solution at pH 7.4
was -34.2 mV. The zeta potential was changed to -41 mV at 5mM PBS solution at pH 7.4.
According to Ted Pella Inc. (the supplier of 40 nm GNP), the zeta potential was -44 mV in DI
water.
The measured zeta potential is at the slipping plane of the GNP. At low salt concentration,
we can ignore the stern layer. The slipping plane thickness xSP of the GNP can be calculated with
the following formula:3
 



 , (S3)
where  and  are the zeta potentials, and  and are the Debye lengths of the 5mM and 10
mM PBS solutions respectively. For the GNPs used in the present experiment,  = -41 mV,  =
-34 mV, = 4.02 nm and = 2.84 nm for 5 mM and 10 mM PBS solutions. The slipping plane
thickness was estimated to be ~1.9 nm. Using this slipping plane thickness, the GNP surface
potential  was calculated with the following formula:
 
 , (S4)
where Vsp is the potential of GNP at the slipping plane or the measured zeta potential. We got V0
= -66 mV. Using the Grahame equation, the corresponding surface charge density σ0 of GNP in
10 mM PBS was calculated to be -24 mC/m2.
S5. Noise analysis of ionic current and potential measurements
The potential amplifier is battery powered, and the only x10 gain is used. Therefore the
bandwidth for potential measurement can be much larger than the current measurement. In the
measurements, we used 5 kHz and 40 kHz low-pass filter bandwidth for current and potential

6
measurements respectively. We compared the noise in ionic current and potential data in our
measurements. The noise power spectrum density (PSD) S(f) is obtained by performing Fast
Fourier Transformations (FFT) on a current or potential time trace (one second) at 0.1V. Figure
S4 (a) and (b) showed the normalized PSDs of current (SI/I2) and potential (Sv/V2) for P1, P2 and
P3 before adding GNPs. It is obvious that the noise of potential data is much smaller than that of
ionic current data, especially at high frequency (>100 Hz). The potential noise spectra display
characteristic 1/f noise or flicker noise at low frequency (<100 Hz). The potential noise spectral
density distribution flats out at higher frequency with reduced magnitude. In contrast, the ionic
current noise spectra show capacitance noise, which increases with the frequency. The noise
analysis suggests the potential measurement is better for faster event measurement. We also
compared the potential noise of P1, P2 and P3 before adding GNPs in the bath solution. Overall,
the noise of P1 is the smallest and the noise of P2 is the biggest. The CNE surface area may play
a role here. There should be an optimized CNE geometry and surface area considering the
balance among noise, sensitivity and sensing range. After adding GNPs, we noticed noise
increase in the potential noise spectra. The noise is related to the accumulation of GNPs. As
shown in Figure S4(c), no obvious change is observed in the noise spectra immediately after
adding GNPs. With the accumulation of GNPs near the nanopipette tip, the change in noise
spectra becomes obvious. Low frequency noise (<100Hz) is increased with the GNP
accumulation, which is related to the observed potential baseline change. At high frequency, a
broad bump appears between 1-10 k Hz. This high frequency noise (also see Figure S4 (f)) is
stronger after a longer accumulation time. This phenomenon is interesting, and the exact reason
needs further investigation. Figure S4 (d) compares the noise spectra of P1, P2 and P3 after the
heavy accumulation of GNPs, suggested by event frequency. The same high frequency noise can
be barely found in the noise spectra of P3 but cannot be resolved in the noise spectra of P2. This
is because the noise level of P2 and P3 is higher.
Before data analysis, we normally applied 10 points (0.2ms window size) to both the
current and potential time traces to reduce noise. Figure S4(e) shows one example before and
after the 10 points moving average smoothing. The zoom-in trace in Figure S4(f) illustrates that
the 10 points moving average did not change the time duration and magnitude of a fast current
spike and potential step (<1 ms). This is a nosier trace from P1. The meter picked up spurious
high frequency (>1 k Hz) and 60 Hz powerline noises. Even in this condition, the potential step

7
is still resolvable in the raw data. After smoothing, the peak-to-peak potential noise is reduced
from 0.4mV to 0.1mV and the potential step is obvious. We sometimes also used 100 points
moving average (2 ms window size) when the events are slower than 2 ms. The peak-to-peak
potential noise can be smaller than 10µV.
Figure S4. (a-b) The normalized noise power spectra for ionic current (a) and potential (b) for a 1 second
time trace from nanopipettes P1, P2 and P3 at applied bias Vb = 100 mV. No GNPs were added in the
solution. (c) The normalized potential noise power spectra for P1 before (blue) and after adding GNPs in
the solution at various times. A higher bias (Vb = 300 mV) is used during GNP accumulation. (d) The
normalized potential noise spectra for P1, P2 and P3 at Vb=100mV after GNP accumulation. (e) A typical
curve of P1 at 50mV after adding GNPs. The black curve is the raw data for current measurement and the
red curve is the smoothed curve of current after 10 points moving average. Similarly, the gray curve is the
raw data for potential measurement and the blue curve is the smoothed curve of potential after 10 points

8
moving average. (f) The zoomed in current and potential traces after applying moving average smoothing
method. Both raw data and smoothed data are shown. The sampling rate is 50kS/s for both measurements
and the bandwidth is 5k Hz for current and 40k Hz for potential. The bath solution is always 10mM PBS.
S6. Ionic current and potential measurement of a dual-nanopore nanopipette
Figure S5. (a) The setup for the measurement using dual-nanopore nanopipette. (b) The current (black)
and potential (red) time traces (1s) before adding GNPs (Vp=0.5V). (c) The current (black) and potential
(red) time traces (1s) after adding GNPs (Vp=0.5V). (d) The zoom-in trace of the green color shade
region in (c). (e) The zoom-in trace to show individual rectangular shape current spikes and the
corresponding potential change. (f) The histograms for spike height (∆II) and spike width (∆tI) of 583
current spikes. The red curves are the Gaussian fits.
The size of the nanopore is 44 ± 4 nm. For this size of the nanopore, we often observed
upward and square shape current spikes at low salt concentration. Figure S5 (c-e) showed the
typically upward current spikes at 10mM PBS. These current spikes showed uniform magnitude
at 8 ±1 pA. The width of these spikes is 0.62 ± 0.02 ms. The current spike magnitude is bigger
than P1 and P2 nanopipettes in the main text, which is due to the comparable size between GNPs
and the nanopore here. The width of these spikes is shorter than the one of P1, suggesting GNPs
translocate faster through the nanopore. As shown in Figure S5 (c) and (d), we also observed

9
familiar saw-tooth potential changes. These potential changes can roughly correlate to the spike
clusters. However the one-to-one correspondence is lost, as shown in Figure S6e.
We also observed downward current spikes from other measurements. In summary,
upward current spikes are more often observed (1) when the size of the nanopore is close to the
GNP size, (2) at a higher applied bias and (3) at lower salt concentrations. For some nanopores,
we observed downward spikes at 0.1V but upward current spikes at higher bias, which is likely
due to the stronger electroosmotic flow at a higher applied bias. With 25 mM PBS bath solution,
the chance to observe downward current spikes are more often.
S7. Optical microscope images of GNPs inside the nanopipettes after translocation
experiments
Figure S6. Optical microscope images for (i) dual-nanopore nanopipette and (ii) CNE/nanopore
nanopipette after GNP translocation experiments. The GNP aggregates are clearly visible inside the
nanopipette barrel.
S8. Simulation

10
Figure S7. (a) 2D axial symmetric geometry of the nanopipette/CNE used for the FEM simulations. The
figures are drawn to scale (r=0 indicates the axis symmetry line). Insets: (i) the quasi 3D view of the
simulation model near the tip. (ii-iii) zoom-in views of the tip region with (ii) a hemispherical shape CNE
with radius 38.5nm and (iii) a flat CNE. (b) Electric potential distribution near the tip region. (c) The net
ion distribution near the tip region. Only potassium and chlorine ions are considered in the simulation. (d)
The potential changes versus the CNP center position along Z axis at different GNP and CNE polarization
conditions. (e) The potential changes versus the CNP center position along Z axis with CNE geometry (ii)
and (iii) in (a). The inset is the zoom-in near the nanopore mouth. (f) The potential changes versus the
CNP center position along Z axis with one (black) and three GNPs (red) at the nanopore mouth. For 3
GNPs case, the Z position is the center of the first GNP. The inset is the electric field distribution for 3
GNPs.
Multifunctional nanopipette/CNE was modeled using 2D axial symmetric geometry as
shown in Figure S7 (a). The quasi 3D view near the tip was shown in the inset (i) and the CNE

11
displayed a donut shape. This geometry exaggerated the CNE size. As we will discuss later, the
size of CNE did not play important role though it will slightly affect the access ionic resistance
and the amount of induced charges. Therefore, we still use this geometry because the
computation time is much shorter. The half cone angle θ of the nanopipette was fixed at 6.5˚. -5
mC/m2 surface charge density was typically applied on the quartz walls if not mentioned
otherwise. We also compared the nanopipette surface charge effect to the ionic current and
potential changes. The changes are very small and there are about 7% increases for both current
and potential when the surface charge density of the quartz wall is increased from 0 to -5 mC/m2.
The nanopore diameter and the CNE base size were always 77 nm. The protrusion length of
CNE was changed from 38.5 nm (ii, hemispherical CNE) to 0 nm (iii, flat CNE) in the models of
Figure S7a. The surface area of CNE (ii) is 1.53 time CNE (iii). The simulation was carried out
with one 40 nm GNP or a cluster of three GNPs (1 nm inter-GNP distance) in the model. A
surface charge density -24 mC/m2 was typically applied to the GNP surface if not mentioned
otherwise. The surfaces at the CNE and the GNP were allowed to float. A constant potential
difference 0.1V was applied to the system and the bath solution is grounded. The whole
computation domain was discretized into free triangular elements. The mesh size is much smaller
than the Debye length (~2.8 nm). To simplify the simulation, only two ions, sodium and chloride
ions, are used at 10 mM concentration. Table S1 summarized the boundary conditions and the
related physics equations.
Table S1: Typical Boundary Conditions and Physics
Surface
Poisson’s Equation
Nernst- Plank Equation
KA
Axial symmetry
Axial symmetry
AB (Ag/AgCl electrode)
Ground
Constant concentration
BC
Zero Charge
No flux (insulation)
CD
Zero Charge
No flux (insulation)
DE (quartz)
-5mC/m2 or 0
No flux (insulation)
EFG (CNE surface)
No charge, floating potential
No flux (insulation)
GH (quartz)
-5mC/m2 or 0
No flux (insulation)
HI (quartz)
-5mC/m2 or 0
No flux (insulation)
IJ (quartz)
Zero Charge
No flux (insulation)
JK (Ag/AgCl electrode)
0.1V
Constant concentration
MNO (GNP surface)
-24 mC/m2 (or others),
floating potential
No flux (insulation)

12
Figures S7 (b-c) showed the electric potential and net ion distribution near the
nanopipette tip with the presence of a GNP at the nanopore mouth. The polarization of CNE was
revealed in the net ion distribution (Figure S7(c)) and negative induced charges were near the
nanopore entrance. The results in Figure S7 (d) demonstrated how the induced surface charges
at the CNE and GNP surfaces affected the potential change at the CNE. Without fixed and
induced charges at nearby surfaces, the voltage divider model dominates, and an obvious
potential increase is observed when the GNP is close to the nanopore entrance. After allowing
the CNE to be polarized, the increased negative induced charges at the CNE compete with the
increased access ionic resistance while the neutral GNP approaches the nanopore. The increase
of induced charges at the CNE is due to the increase of the local electric field with the
approaching of the GNP to the nanopore. The polarization of GNP shows a negligible effect on
the potential change at the CNE. Figure S7(e) shows the effect of the CNE geometry to the
potential measurements. As shown in the inset, the protruded CNE can detect the potential
change slightly earlier with a bigger magnitude. However, the difference is only about 5% when
we changed the CNE area 1.53 times. Figure S7 (f) compares the potential changes ∆V when one
and three GNPs near the nanopore mouth. It shows that the CNE can cumulatively measure a
bigger potential change from three charged GNPs near the CNE (~25% change). The potential
increase near Z=1µm in the red plot is attributed to the increased access resistance by the
accumulation of 3 GNPs near the nanopore. This is somewhat exaggerated by the donut shape of
the CNE.
References
1. Wang, Y.; Kececi, K.; Mirkin, M. V.; Mani, V.; Sardesai, N.; Rusling, J. F. Resistive-
Pulse Measurements with Nanopipettes: Detection of Au Nanoparticles and Nanoparticle-Bound
Anti-Peanut Igy. Chem. Sci. 2013, 4, 655-663.
2. Tiwari, P. B.; Astudillo, L.; Miksovska, J.; Wang, X.; Li, W.; Darici, Y.; He, J.
Quantitative Study of Protein–Protein Interactions by Quartz Nanopipettes. Nanoscale 2014, 6,
10255-10263.
3. Shan, X.; Fang, Y.; Wang, S.; Guan, Y.; Chen, H.-Y.; Tao, N. Detection of Charges and
Molecules with Self-Assembled Nano-Oscillators. Nano Lett. 2014, 14, 4151-4157.