Investigation of Shielding Strategies to Mitigate Noise in Nanopore Recordings

Abstract

Shielding against electromagnetic interference (EMI), vibration, and mechanical noise is crucial for the reliable performance of sensitive electronics, especially in high-disturbance environments. While Faraday cages are ubiquitously used for this purpose, they are often implemented simply as metal boxes enclosing the measurement system without accounting for critical design factors such as skin depth. In this study, we investigated a range of factors to understand noise sources, coupling pathways, and systematic approaches for effective shielding beyond the common metal box approach, including the evaluation of power-line quality, grounding strategies for the Faraday cage, connectivity solutions to maintain Faraday cage integrity, and the relative positioning of the components in the measurement setup. Our findings have rendered the noise in the current measurements comparable to that of biological nanopores in the more ubiquitous 10 kHz lowpass filter setting. This low-noise floor has allowed us to detect DNA translocations under challenging low-electrolyte conditions (e.g., 10 mM KCl) using the highest lowpass filter setting of the Axopatch 200B (i.e., 100 kHz). Moreover, with necessary modifications, these findings were extended to the megahertz-level bandwidth regime, where sampling rates up to 10 MHz yielded extractable events with unfiltered traces. Although our findings broadly cater to solid-state nanopore (SSN) sensing, their applicability should extend beyond SSNs to other sensitive electronic systems.

Introduction

Measurement noise is inevitable and inextricably linked with the performance metrics of the device/system undergoing testing. Nanopore sensors, like any other measurement method, are not immune to inherent noise and stochastic fluctuations and require careful design considerations to shield against them, given the demand for low-noise, picoamp-level ionic current measurements for the interpretation of molecular-level signatures. − A common metric for quantifying the noise of a nanopore system is the RMS (root-mean-square) current (I _rms). Electromagnetic Interference (EMI) is a primary source of interference when performing nanopore experiments. − EMI is omnipresent and unavoidable, given the presence of electronic devices and processes capable of electromagnetic generation in a working area. To minimize exposure to EMI, Faraday cages are used. From a simplistic view, when an electromagnetic wave strikes the surface of a metal, the electric charges on the surface are excited, canceling the electric field at the surface and thus removing the electric field inside the cage. An important criterion for material selection for Faraday cage construction is the skin depth (δ). The skin depth is defined as the depth at which the wave’s intensity decreases to about 37% (1/e) of its original value at the surface: $\delta =1/\sqrt{\pi f\mu \sigma }=\sqrt{(2\rho /\omega \mu )}$ , where f, μ, σ, ρ, and ω are the frequency of operation, material permeability, material conductivity, material resistivity, and angular frequency of the wave, respectively. The minimum Faraday cage thickness can be calculated by considering that 98% of the current flow is within 4δ. Thus, multiple skin depths are needed for substantial attenuation of incident electromagnetic waves. Decreasing the frequency (f) increases δ, allowing the electromagnetic field and induced currents to penetrate more deeply into the material. At 60 Hz line frequency, the δ for aluminum is ∼10.9 mm, whereas that of carbon steel is ∼0.8 mm (lower because of its higher permeability). Given that the typical thickness of aluminum foil is less than 0.2 mm, it is not surprising that a single layer of aluminum foil yields ineffective shielding against line-frequency noise.

This work primarily focuses on noise mitigation and management and on developing shielding strategies against mechanical and electrical noise sources. While highly vibration-sensitive techniques such as electron microscopy use acoustic dampeners, the same is not extended to nanopore sensing, even though incidental vibrational noise, such as knocking on a table, can translate into measurable current fluctuations during (sub)­nA-level current measurements. Conventional isolation platforms or tables (active or passive) are often expensive and are bulky. We found acoustic isolation pucks (hereafter termed vibration isolators) and acoustic foam (AF) (referred to by its commercial use despite its key role in suppressing the propagation of mechanical shocks) used in the music industry to be powerful solutions for this endeavor. A nested Faraday cage equipped with acoustic-damping panels mounted on vibration isolators was developed in this study. This design enabled us to perform experiments without requiring expensive, bulky, active vibration-isolation platforms. Thereafter, we focused on good grounding practices, strategic positioning of the Faraday cage, amplifier, digitizer, and cables in the lab space, and expanding the findings to high-bandwidth experiments. The noise reduction allowed us to perform DNA translocation experiments under challenging conditions, such as 10 mM KCl, using the highest lowpass filter (LPF) of the Axopatch 200B (i.e., 100 kHz). The strategies explored in this study would be useful beyond nanopore sensing, where mechanical, vibrational, and electrical noise features play a key role.

Methods

Nanopipette Fabrication

Nanopore sensing was performed using quartz capillaries (Sutter Instruments, QF100-70-10 or QF100-50-10) that were air plasma-cleaned (Harrick Plasma, PDC-32G (115 V)) at a low setting for ∼1 min and pulled using a CO₂ laser puller (Sutter P-2000). The pipettes were pulled at heat 630, fill 4, and delay 130. The pull parameter was varied from 150 to 220.

Bilayer and Biological Nanopore Formation

A detailed account is provided in Supporting Information Section 1.

Data Acquisition

Readouts from the Axopatch 200B (Molecular Devices) amplifier were digitized using either a Digidata 1550B (Molecular Devices) or a PCIe-6351 (National Instruments) connected to a BNC 2110 (National Instruments) block connector. Nanopore traces were also acquired using an eNPR-10 MHz amplifier (Elements srl).

Data Analysis

Events were extracted using the HyperXtract platform. Custom MATLAB scripts (R2023b) were used for data processing.

Faraday Cage Material

A detailed account is provided in Supporting Information Section 2 and Table S1.

Electrolyte Preparation

All electrolytes were prepared by adding solid powdered salts to ultrapure water (VWR, 76777-142, resistivity ≥18 MΩ cm) and filtered using 0.20 μm PES membranes (VWR, 97066-204). The pH was adjusted by adding concentrated drops of KOH (Sigma-Aldrich, 306568) or HCl (Sigma-Aldrich, H1758) and measured using a pHenomenal MU 6100 L Multi-Parameter Meter (VWR, 76460-494). LiCl (4 M) and 100 mM KCl were buffered using 10 mM Tris-HCl buffer (Fisher Scientific, 50-101-8924), while the 10 mM KCl solution was buffered using 1 mM Tris-HCl buffer.

Translocation Experiments

DNA (1kb ladder, Fisher Scientific, 10-787-018) was added to the bath side to a final concentration of 0.5–25 μg/mL. All DNA experiments were conducted at pH ≈ 8 using a variety of electrolytes as noted in the text.

Results and Discussion

Faraday Cage Construction

Components, materials, functions, and purposes of each component of the nested/dual Faraday cage are outlined in Table . A brief account is also provided in Supporting Information Section 2.

1. Components, Material, Function, and Purpose of Each Component of the Nested/Dual Faraday Cage.

component	layer/material	purpose	figure(s)
outer cage	metal box with Faraday cloth lining	primary EMI/RFI shielding (RFI = radio frequency interference)	Figure a
inner cage	extruded aluminum	physical skeleton for the nested layers	Figures b and S1
metallic layers (L) of the inner cage	stainless steel mesh (L1), copper mesh (L2), Faraday cloth (L3), copper mesh (L4), Faraday cloth (L5)	secondary EMI/RFI shielding	Figures S1b–S1f
internal padding	high-density antistatic foam and anti-electrostatic discharge (ESD) foam	prevents electrostatic discharge (ESD) damage to headstage/amplifier components	Figures S1g and S1h
door and seam	copper mesh/cloth and copper tape	ensure electrical continuity across the opening	Figure S1i
acoustic control	acoustic foam panels	protection from mechanical noise	Figures S1i, S1j, S1l, and S1m
vibration isolation	eight vibration isolators (i.e., acoustic isolation pucks)	passive dampening; allows for benchtop operation	Figure a
analytical imaging		morphology and composition of Faraday cloth, Cu mesh, and wire mesh through SEM and EDX micrographs	Figures c and S2

Breach Points and Grounding

Axopatch 200B

Insertion of the headstage requires breaching the Faraday cage. The opening in the inner Faraday cage for inserting the Axopatch 200B headstage is kept to a minimum and sealed with conductive copper tape (Figure d). The nested Faraday cage system is grounded to the signal ground of the Axopatch 200B (Figures S1k and S3a). Both the inner and outer Faraday cages are electrically connected. This uninterrupted electrical connection provides a uniform grounding potential throughout the shielding structure, preventing localized charge buildup that could otherwise cause EMI or damage to sensitive electronic components. The implications of other grounding methods are discussed later.

1.

Faraday cage design. (a,b) Schematics outlining the parts of the nested Faraday cage. (c) SEM images of metallic components of the inner cage. (d) Photograph of the point where the headstage cable breaches the inner Faraday cage. (e) Photograph of the keystone coupler used to connect the analog module of the eNPR-10 MHz amplifier to its digital module. (f) Photographs of the eNPR-10 MHz analog module connected to a custom 3D-printed grounding plate.

Elements eNPR-10 MHz

A metallic keystone coupler is inserted to connect the analog module placed inside the inner Faraday cage to the digital module residing outside the Faraday cage system (Figures e and f; digital module not photographed). The keystone coupler, inner Faraday cage, and outer Faraday cage are electrically connected and grounded using a grounding hub (Figure S3b).

Effectiveness of the Faraday Cage

Note: Supporting Information Section 3 provides a detailed account of the literature-reported I _rms values and how they compare with the values reported in the current work. As shown later, the I _rms of quartz nanopipettes are now comparable to those of α-hemolysin due to the effective noise suppression provided by the nested Faraday cage system. Since the waves (we are concerned with) are broadly electromagnetic and mechanical in nature, to evaluate the shielding effectiveness (SE), we used the tests outlined in Table S5 (Supporting Information Section 4).

Nomenclature for Figure

2.

Effectiveness of the Faraday cage design against EMI, vibrational, and acoustic noises. See the main text for the nomenclature of the Faraday cage system shown in the figure. (a) Different stages of the Faraday cage construction, with the table on the left outlining the key for each abbreviation. (b) Shielding effectiveness of the dual Faraday cage (i.e., the full Faraday cage system; magenta), the inner Faraday cage (brown), and the outer Faraday cage (cyan). (c) Spectral analysis of frequencies from 200 to 1800 MHz corresponding to the ambient environment (black), outer Faraday cage (cyan), inner Faraday cage (brown), and dual Faraday cage (magenta). (d) Gyroscope readings in response to the drop tests corresponding to different stages of the Faraday cage construction with and without acoustic foam. Current readouts from the Axopatch 200B in response to drop tests corresponding to different stages of the Faraday cage construction (e) without and (f) with acoustic foam. (g) Current readouts from the Axopatch 200B in response to acoustic tests (AN = ambient noise, WN = white noise) corresponding to different stages of the Faraday cage construction with (C3 + VI + AF) and without (C3 + VI – AF) acoustic foam. (h) Acoustic levels in different stages of the Faraday cage with and without the acoustic foam. All current traces shown in this figure were lowpass-filtered at 100 kHz (using the inbuilt LPF option of the Axopatch 200B) and sampled at 500 kHz.

The nomenclature of the Faraday cage system shown in Figure a is as follows: A + B indicates A with B, while A – B indicates A without B. For example, C1-AF stands for C1 without acoustic foam (AF).

EMI Test

Shielding effectiveness (SE) of a Faraday cage (based on receiver-transmitter powers) can be expressed as $\mathrm{S}\mathrm{E}=10{\log }_{10}\left(\frac{P_{\mathrm{u}\mathrm{n}\mathrm{s}\mathrm{h}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d}\mathrm{e}\mathrm{d}}}{P_{\mathrm{s}\mathrm{h}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d}\mathrm{e}\mathrm{d}}}\right)$ , where P _unshielded and P _shielded are receiver power without and with the Faraday cage, respectively. We used RF frequencies in the range of 1 GHz–1.8 GHz to evaluate SE (Figure b). More frequently used MHz-level frequencies were unusable due to environmental features (Figure c, top left black trace). As seen in Figure b, we observed a sinusoidal-like behavior for SE: this is typically expected due to various factors such as cavity effects and standing waves. As seen, at minimum, the SE is > 30 dB, which means the Faraday cage reduces the electromagnetic signal strength by 1,000×. We then examined the Faraday cage’s ability to shield against ambient MHz-level frequencies detected in the lab space (Figure c). For this, the spectral analysis from 200 to 1800 MHz was carried out for four cases:

• IAmbient environment (black trace).
• IIOuter Faraday cage only (cyan trace).
• IIIInner Faraday cage only (brown trace).
• IVDual Faraday cage (magenta trace).

As seen in Figure c, the outer Faraday cage alone is insufficient to shield against ambient EMI (cyan trace). While the inner Faraday cage alone (brown trace) is more effective than the outer Faraday cage, implementing both together is more effective than either individually (magenta trace). This is evident through the presence of a small peak at ∼600 MHz (red arrow in brown trace of Figure c), which disappears with the dual Faraday cage implementation.

Drop Test

Experimental details are outlined in Supporting Information Section 4. This test represents the effectiveness of the Faraday cage system against vibrational noises. Each drop also generated acoustic noise in the 100–120 dB range (beyond the acoustic levels perceived in the lab). These drop tests mimic an extreme scenario of a blow/knock directly on the table supporting the Faraday cage. The mean and standard deviation of the 5 highest absolute changes in angle per unit time (°/s; in the x, y, or z direction) from 15 drops were used as the metric for vibrational noise (Table S6). The control was performed by placing the accelerometer on the table without any weight drops (C0, Figure d). Critical observations are noted below:

• iThe highest noise level is observed with an acoustic foam-free Faraday cage placed directly on the table without vibration isolators (∼0.8 ± 0.1°/s; C1 of Figure d; purple bar).
• iiInner stages (i.e., stages 1 and 2) with Sorbothane feet help minimize noise levels by as much as ∼8–10× (Table S6 and Figure d).
• iiiThe acoustic foam provides shielding against noise propagated through vibrations: a ∼4× reduction was seen in C1, where the Faraday cage is placed directly on the table (Figure d, green vs purple bars).
• ivAlthough the data in Table S6 suggest that stages with Sorbothane feet provide adequate shielding against vibrational noises and vibration isolators may not be needed, the Axopatch 200B readings show otherwise (Figure e,f, Table S7). The drop test measured with the Axopatch 200B was limited to four cases, as seen in Table S7. The headstage was always placed inside the inner Faraday cage (i.e., on stage 2). Notably, as seen in Figure S5 (results from 15 drop tests), noise spikes did not always show up in readings in the presence of vibration isolators, unlike when the Faraday cage was directly placed on the table.
• vAs seen in Table S7, when the Faraday cage is placed on the vibration isolator, the noise values are within error, irrespective of the presence of the inner stage 1 and the acoustic foam. However, in the absence of vibration isolators (i.e., the Faraday cage placed directly on the table/support), the Faraday cage with the acoustic foam performs better.

These results indicate the utility of vibration isolators and acoustic foams for shielding against vibrational noise.

Acoustic Noise Test

This test provides insight into the effectiveness of the Faraday cage system against acoustic noise. Interestingly, as seen in Table S8 and Figure g, acoustic damping does not appear to correlate strongly with the presence or absence of acoustic foam when the Faraday cage is present. In contrast, the vibrational noise appears to be more significantly damped by the presence of acoustic foam than the acoustic noise (vide supra). Furthermore, as shown in Table S9 and Figure h, the I _rms values appear to be independent of the presence or absence of an acoustic foam. The door of the outer Faraday cage has a rubber gasket on the mating surface, which may provide the observed shielding against acoustic noise. The interior of the secondary Faraday cage, with and without acoustic foam layers, is substantially quieter than the surrounding environment (Figure h). Thus, the current nested Faraday cage is well suited to operate in spaces plagued with considerable acoustic noise characteristics.

Power Line Characteristics and Their Influence on the Quality of Traces

Can the Axopatch/Digitizer be plugged into any power outlet? The simple answer is no. This is because unwanted noise components in power lines can directly or inductively couple into the measurements. Three cases were mainly investigated:

• IAxopatch on a noisy power line.
• IIAxopatch and the PC were on the same (clean) power line.
• IIIAxopatch on a clean power line.

To understand the frequency components present in alternating current (AC) lines, a differential probe connected to an oscilloscope was used to acquire time domain traces (Figure a), which were then fast Fourier transformed (FFT) into the frequency domain (Figure b,c). As expected, we could see the AC lines’ principal 60 Hz noise (Figure b; magenta arrow). The frequency spectra from all three power sources were indistinguishable in the low-frequency regime, as seen in Figure b. Apart from the 60 Hz peak, a noticeable subharmonic peak at ∼20 Hz was present (black arrow; discussion to follow). Interestingly, unlike the other two cases, the frequency spectrum corresponding to the noisy power line shows a noticeable peak at ∼17 kHz (Figure c, black arrow). As seen in Figure d,e, when the Axopatch is plugged into this noisy power line, unwanted noise spikes appear, which are more frequent when the PCIe-6351 multifunctional DAQ, coupled to the BNC 2110 block connector, is used to digitize signals (zoomed-in views in Figure S6). However, unlike the PCIe-6351/BNC-2110 combination, the noise spikes remain prevalent even after lowpass filtering the trace at 10 kHz with the Digidata 1550B as the digitizer (black traces in Figure d,e). As seen in Figure f, the noise spikes in the green trace start to diminish with lower LPF cutoff frequencies. The disappearance of the noise spike with the 10 kHz LPF in contrast to its persistence with the 20 kHz filter suggests that the noise feature is primarily coupled within the 10–20 kHz frequency range. Figure S7 shows a series of images in which the green trace from Figure f is lowpass-filtered with progressively decreasing cutoff frequencies from 20 kHz to 10 kHz in 1 kHz intervals. As observed, the periodic jagged spikes become less prominent at LPF settings <16 kHz, in good agreement with the spike observed at ∼17 kHz in Figure c for the noisy power line.

3.

Effect of power lines on the quality of current readouts. (a) Oscilloscope reading of the attenuated voltage from noisy (green), clean (magenta), and clean power lines with a PC plugged in (brown). Fast Fourier transform (FFT) of (a) to identify (b) low-frequency and (c) high-frequency noise components. Current readings from an Axopatch 200B plugged into a noisy power line (green), clean power line with a PC plugged in (magenta), and clean power line (purple), with (d) Digidata 1550B and (e) PCIe-6351 as the digitizer. Black traces of (d) and (e) correspond to lowpass-filtered traces at 10 kHz. (f) Zoomed-in view of the noisy power trace of (e) with various lowpass filter settings as indicated in the plot label. Current traces from an Axopatch 200B when an auxiliary power conditioner strip is plugged into the same circuit as the Axopatch, with (g) Digidata 1550B and (h) PCIe-6351 as the digitizer. Green and gray traces of (g) and (h) corresponds to 100 kHz and 10 kHz lowpass filtering while black trace corresponds to the absence of the power conditioner strip (100 kHz lowpass filtering) (i) I _rms corresponding to Axopatch (100 kHz lowpass filtering) on a noisy power line (case 1), Axopatch and the PC on the same (clean) power line (case 2), Axopatch on a clean power line (case 3), and an auxiliary power strip on the same power line as the Axopatch (case 4). (j) Lowpass-filtered trace at 5 kHz and its (k) FFT corresponding to low frequencies. (l) Trace corresponding to highpass filtering the trace in (j) with a cutoff frequency of 25 Hz.

Interestingly, noise coupling can also come through an auxiliary power strip plugged into the same circuit as the Axopatch. For example, Figure g,h was obtained when a power line conditioner (B08QZ8ZXF7, Amazon) was plugged into the same power line as the Axopatch (green trace at 100 kHz lowpass filtering and gray trace at 10 kHz lowpass filtering). Although no equipment is plugged into this auxiliary power strip, we observe periodic noise issues when the BNC2110/PCIe-6351 combination is used instead of the Digidata 1550B. This data also strengthens the previously observed resilience of the Digidata 1550B to noise features present in the power line, compared to the BNC2110/PCIe-6351 combination. Figure i and Table show the I _rms of the two systems corresponding to the four cases shown in Figure d–h. As seen, with a clean power line, the BNC2110/PCIe-6351 combination shows a lower I _rms compared to that of the Digidata 1550B. The latter also shows consistent I _rms performance across the four cases, further demonstrating the resilience of the Digidata 1550B system against noise sources in power lines.

2. I _rms Values Corresponding to Figure i (4 M LiCl as the Electrolyte) .

case	description	I rms with BNC2110/PCIe-6351	I rms with Digidata 1550B
1	Axopatch on a noisy power line	34.2 ± 0.4 pA	27.5 ± 0.3 pA
2	Axopatch and the PC on the same (clean) power line	25.3 ± 0.6 pA	27.0 ± 0.2 pA
3	Axopatch on a clean power line	20.9 ± 0.3 pA	26.7 ± 0.1 pA
4	Auxiliary power strip on the same power line as the Axopatch	28.3 ± 0.4 pA	26.6 ± 0.3 pA

^aThe sampling rate, LPF (inbuilt Bessel filter of the Axopatch 200B), and the voltage bias were 500 kHz, 100 kHz, and 500 mV, respectively

Low-Frequency AC Noise

While most attention is garnered to the 60 Hz noise, its subharmonics, such as the 20 Hz noise (black arrow of Figure b), should not be neglected. To demonstrate this, we lowpass-filtered the purple trace in Figure e at 5 kHz: when zoomed in, a wavy current trace is visible (Figure j). The FFT spectrum (Figure k) shows a strong peak at ∼20 Hz. The wavy behavior disappears completely when the 5 kHz lowpass-filtered trace is highpass-filtered at 25 Hz (Figure l). While the skin depth of carbon steel, with a conductivity of 6.99 × 10⁶ S/m and a relative permeability of 280, is calculated to be 1.5 mm at 60 Hz, it increases to 2.5 mm at 20 Hz, which exceeds the thickness of the outer Faraday cage. Thus, it is unsurprising to see the 20 Hz subharmonic present in the current trace. Thus, an improvement to the current Faraday cage setup would be to use outer cage material thick enough to shield against sub-60 Hz noise levels.

Primary Grounding Considerations for a Low-Noise Station

We have identified major factors contributing to noise penetration and have outlined suboptimal conditions in Table to help enhance the shielding effectiveness and minimize signal contamination.

3. Descriptions of Suboptimal Grounding Conditions, I _rms, and the Associated Figures .

grounding conditions	I rms	figure
optimal grounding condition (i.e., Faraday cage grounded to the signal ground of the Axopatch 200B)	∼26.5 ± 0.1 pA	a
Faraday cage grounded to an optical table	∼367.1 ± 44.2 pA	b
Faraday cage grounded to a grounding hub	∼369.1 ± 18.3 pA	c
signal ground of the Axopatch grounded to an optical table	∼656.0 ± 4.6 pA	d
signal ground of the Axopatch grounded to a grounding hub	∼1310.0 ± 49.04 pA	e
Faraday cage and the Axopatch signal ground grounded to an optical table	∼96.8 ± 6.5 pA	f
Faraday cage and the Axopatch signal ground connected to a grounding hub	∼124.6 ± 9.0 pA	g
the Axopatch headstage cable adjacent to a power cable	∼30.8 ± 1.1 pA	h and S8a
ground cable from the outer Faraday cage connecting to the signal ground of the Axopatch 200B lying close to a power cable	∼35.6 ± 2.3 pA	i
ground cable from the outer Faraday cage overlapping with the headstage cable of Axopatch 200B	∼28.4 ± 0.5 pA	j and S8b

^aThe sampling rate, lowpass filter frequency (inbuilt Bessel filter of the Axopatch 200B), and the voltage bias were 500 kHz, 100 kHz, and 500 mV, respectively.

Based on Table , the following recommendations are made:

• iAll cables exposed to the ambient environment are wrapped with a copper-braided cable, followed by a cable sleeve for additional shielding.
• iiSignal and grounding cables should not overlap and should be kept away from the power cables.
• iiiThe Faraday cage should be grounded via the signal ground of the Axopatch amplifier.
• ivAlthough not shown, signal and grounding cables should be isolated from metallic structures, such as shelves, which could be magnetized by nearby power strips and other sources.

Figure k shows a head-to-head comparison of the cases in Table , with Figure l showing selected cases that are hard to see in Figure k because of the scaling. Other critical factors to consider are included in Supporting Information Section 5.

4.

Important grounding considerations for a nanopore station with Axopatch 200B and Digidata 1550B. All measurements in this subsection were carried out using a ∼9 nm diameter nanopipette immersed in 4 M LiCl. The sampling rate, lowpass filtering (inbuilt Bessel filter of the Axopatch 200B), and the voltage bias were 500 kHz, 100 kHz, and 500 mV, respectively. (a–j) Figure captions outline each of the grounding conditions investigated, with (k) and (l) showing the corresponding I _rms values. See Table in the main text for more details. FC stands for the Faraday cage.

5.

Important grounding considerations for a nanopore station with Axopatch 200B and BNC2110/PCIe-6351. All measurements in this subsection were carried out using a ∼17 nm diameter nanopipette immersed in 4 M LiCl. The sampling rate, lowpass filter frequency (inbuilt Bessel filter of the Axopatch 200B), and the voltage bias were 500 kHz, 100 kHz, and 500 mV, respectively. (a–e) Figure captions outline each investigated grounding condition, with (f) showing the corresponding I _rms values for the PCIe-6351/BNC 2110 digitizer system (i.e., DS-II).

Grounding Considerations for Block Connector-Based Digitizers

We investigated whether the noise characteristics of the PCIe-6351, coupled to the BNC 2110 block connector (referred to as digitizer system II (DS-II)), would differ from those of the Digidata 1550B. The block connector also has analog and digital input grounding ports (AIG and DIG, respectively, located in the digital and timing I/O section). Table lists power line and grounding considerations critical for each digitizer system.

4. Influence of Power Line and Grounding Conditions on the Noise Performance of the Digitizer Systems, as Shown in Figure .

power line/grounding conditions	I rms	figure
DS-II system on a noisy power line	∼33.0 ± 16.7 pA	a
Digidata 1550B on a noisy power line	∼26.5 ± 0.1 pA	b (gray trace)
DIG of DS-II grounded	∼21.1 ± 0.3 pA	c
AIG of DS-II grounded	∼23.4 ± 0.3 pA	d
DIG and AIG of DS-II grounded	∼17.2 ± 0.2 pA	e

^aThe sampling rate, lowpass filter frequency (inbuilt Bessel filter of the Axopatch 200B), and voltage bias were 500 kHz, 100 kHz, and 500 mV, respectively.

^bNote: while the I _rms of the DS-II system is higher on a noisy power line compared to the Digidata 1550B, simultaneous grounding of AIG and DIG leads to a lower I _rms, further emphasizing the importance of grounding.

Leveraging Low Noise for Low-Electrolyte and Short Single-Stranded DNA Sensing

High-Bandwidth DNA Detection in Low Electrolyte Conditions

Low signal magnitude is a distinct challenge when operating in low electrolyte solutions (e.g., physiological levels). In such cases, operating at a low bandwidth is common practice to minimize open-pore noise. Herein, we leveraged the performance of our low-noise Faraday cage system to demonstrate translocation of DNA in challenging 100 mM KCl (Figure b top trace) and 10 mM KCl (Figure b middle trace) using unmodified quartz nanopipettes (500 mV applied bias, 500 kHz sampling rate, and 100 kHz inbuilt lowpass filtering of the Axopatch 200B). The latter might be the lowest salt concentration attempted for DNA with the 100 kHz built-in LPF of the Axopatch 200B. The ability to operate under low electrolyte conditions would be crucial for the field going forward, with the strong growing interest in probing biological systems. The corresponding scatter plots are shown in Figure c, where we see a compact translocation profile for 10 mM KCl compared to 100 mM KCl.

6.

Leveraging low noise for low-electrolyte and short single-stranded DNA sensing. (a) Schematic of nanopipette sensing where DNA is loaded into the pipette. (b) 30 s traces of double-stranded DNA translocations corresponding to 100 mM KCl and 10 mM KCl as the electrolyte with 100 kHz (black) and 10 kHz (magenta) lowpass filtering (as expected, both yielded conductive pulses), with (c) showing the corresponding color-weighted scatter maps for the 100 kHz lowpass filtering scenario. (d) Resistive current trace corresponding to single-stranded DNA (ssDNA) homopolymer composed of 15 adenine bases (recorded at the 100 kHz lowpass filter setting of the Axopatch) and (e) the corresponding color-weighted scatter maps. KCl (1 M) was used as the electrolyte, with ssDNA (132.6 μg/mL) in the bath and 0.01% w/v Tween-20 in the pipette.

High-Bandwidth Short-Strand DNA Detection

Detection of short ssDNA using solid-state nanopores is particularly challenging because their small molecular size results in extremely fast translocation events (∼10–10,000 ns/base). The rapid passage leads to limited interaction time with the sensing region, producing low signal amplitudes and poor SNR. Hu et al. recently demonstrated the successful detection of 20-nucleotide-long single-stranded DNA (ssDNA) homopolymer types using a quad-pore architecture. In this work, we used a 1 M KCl system (with 0.01% w/v Tween-20 in the pipette) and leveraged the low-noise floor of the system to detect ssDNA homopolymer consisting of 15 adenines (poly_15A) using 0.5 mm ID quartz nanopipettes. As shown in Figure d, we detected resistive translocation events from this homopolymer. The corresponding scatter plot is shown in Figure e: a single population was observed due to the small pore size and the homopolymer’s strand length. These findings further strengthen the utility of our noise management strategies to detect historically challenging analytes with minimal modifications to the pore system.

Comparison with Biological Nanopores

Biological nanopores are considered the gold standard in the field for noise performance, typically outperforming solid-state nanopores. Biological nanopores also benefit from being embedded in lipid bilayers, which are natural insulators with a low membrane capacitance (≲150 pF), thereby suppressing amplification of high-frequency noise. In this section, we compare the improved noise performance of quartz nanopipettes with that of α-hemolysin (see Supporting Information Section 1 for experimental details). The Elements 100 kHz amplifier, operating at an 80 kHz sampling rate with the ultralow-noise (ULN) chip, was used for these experiments (1 M KCl, pH ≈ 8, applied voltage of 100 mV). Figure a shows the unfiltered traces corresponding to α-hemolysin and a quartz nanopipette (0.7 mm ID), and Figure b shows a trace corresponding to the translocation of poly_15A. The trace in Figure b is provided as evidence for the presence of the α-hemolysin pore across the lipid bilayer. The I _rms values for α-hemolysin and quartz were 3.9 ± 0.3 pA and 2.3 ± 0.2 pA, respectively (Figure c). These are in close agreement with values found in the literature (Table S2). This further supports the superior noise shielding our nested Faraday cage provides, since solid-state nanopores can now achieve noise levels comparable to the gold standard in nanopore sensing at the more ubiquitous 10 kHz lowpass filter setting.

7.

Comparison of α-hemolysin and the quartz nanopipette. (a) Unfiltered current traces corresponding to α-hemolysin and the quartz nanopipette tip filled with 1 M KCl under an 80 kHz sampling rate at a 0 mV applied bias. (b) An unfiltered current trace corresponding to the translocation of poly_15A through the α-hemolysin pore at an applied bias of 100 mV. (c) Bar graph corresponding to the I _rms values of α-hemolysin (α-H) and quartz (Q) nanopipettes (three runs each; 10 kHz lowpass filtering).

Leveraging Low Noise for High-Bandwidth Sensing

For the high-bandwidth (HBW) experiments, we used the Elements 10 MHz amplifier (eNPR-10 MHz). Since nanopipettes cannot be used as-is within the amplifier’s built-in Faraday cage, a few key modifications were needed to operate with the cage open while mitigating noise, as noted in Supporting Information Section 6. Figures a and b showcase the three grounding permutations that were tested:

• aGrounding pins of the analog and digital modules not connected and not grounded (black, modules not connected (NG)).
• bGrounding pins of the two modules connected but not grounded (green, modules connected (NG)).
• cGrounding pins of the two modules connected and grounded (magenta, modules connected (G)).

8.

Optimizing grounding conditions for high-bandwidth sensing using the eNPR-10 MHz amplifier. Unfiltered current traces corresponding to ∼14 nm diameter nanopipette tips filled with 4 M LiCl under a 5 MHz sampling rate and 500 mV bias (a) without and (b) with the keystone coupler. Colors in (a) and (b) correspond to ground points of digital and analog modules: (black) not connected and ungrounded, (green) connected yet ungrounded, and (magenta) grounded and connected. (c) Unfiltered traces corresponding to DNA translocation under different sampling rates (∼9 nm tip diameter, 4 M LiCl, 500 mV).

Note: A keystone coupler is placed across the inner Faraday cage (Figure e) to connect the analog and digital modules in Figure b, whereas in Figure a, the two modules are connected directly without a keystone coupler. As shown in Figure a,b, the keystone coupler helps preserve the integrity of the Faraday cage, thereby reducing noise.

The I _rms numbers for each case at sampling rates of 1.25 MHz, 2.5 MHz, 5 MHz, 10 MHz, 20 MHz, and 40 MHz are provided in Table S10; the I _rms values are >4× lower with the keystone coupler. While the grounding method appears to have less effect on I _rms up to 10 MHz, its importance becomes apparent at sampling rates above 10 MHz.

For HBW DNA-sensing experiments, we used quartz capillaries with a 0.5 mm ID. Figure c shows the unfiltered DNA translocation traces acquired at sampling rates of 1.25 MHz, 2.5 MHz, 5 MHz, and 10 MHz. Noise above 10 MHz dominated the signal magnitude; no events could be observed without lowpass filtering. The Drndic group has successfully shown the use of silicon nitride membranes on fused-silica (glass) chips for detecting DNA under the highest specifications of the eNPR-10 MHz amplifier. In addition to the superior noise properties of glass, the sub-5 nm effective membrane thickness yields larger signal magnitudes, facilitating the operation in challenging high bandwidth conditions (e.g., with ultrasmall pores (thickness = 1.7 nm and diameter = 3 nm), the event depth can theoretically reach up to 23 nA). Since the sensing length of nanopipettes is typically >100 nm, additional work is needed to increase the signal magnitude to be useable at extremely high bandwidths.

Concluding Remarks

In this work, we have shown the construction of a nested Faraday cage and its resilience against EMI, acoustic, and vibrational noise. Its construction allows it to forego the traditional requirement for passive or active vibration isolation tables, enabling broader adoption. We have also shown the importance of clean power sources for the Axopatch and the Digitizer, without which unwanted noise coupling from the power lines occurs. Furthermore, the importance of not plugging in additional equipment that could cause noise coupling was demonstrated. Thereafter, we showed grounding strategies for both the Axopatch and the digitizer, in which we grounded the Faraday cage system directly to the Axopatch’s signal ground, resulting in the lowest noise. Our findings discourage the grounding via the table or through a grounding hub. Thereafter, the influence of the digitizer type was demonstrated using an alternative to the widely used Digidata 1550B (PCIe-6351 coupled to a BNC 2110 block connector). Upon optimal grounding of the BNC-2110, we observed that this alternative digitizer system outperforms the Digidata system, achieving sub-20 pA I _rms at the highest lowpass filter setting of the Axopatch (i.e., 100 kHz). Taking this noise improvement holistically has enabled us to conduct DNA experiments under very low-electrolyte conditions, such as 10 mM KCl, and to detect 15-nucleotide-long oligonucleotides at the highest lowpass filter setting of the Axopatch. Furthermore, noise-reduction strategies were used to test the performance of quartz nanopipettes at MHz-level bandwidths, where we observed DNA events up to 10 MHz (unfiltered traces). Thus, our findings provide a detailed framework for minimizing noise in nanopore measurements, ranging from wiring orientation to Faraday cage construction.

The data supporting this study’s findings are available from the corresponding author upon reasonable request.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmeasuresciau.6c00125.

• Bilayer and biological nanopore formation; construction of the Faraday cage system; account on literature reported and current I _rms; testing of Faraday cage effectiveness; grounding considerations; operations with high-bandwidth equipment; and references (PDF)

†.

T.L. and J.C. contributed equally.

The authors declare no competing financial interest.