When EPF is dominant, the capture zone surrounding the pore opening demonstrates an almost spherical shape. In lower ionic concentrations, the EOF can dominate the capturing process and drive the DNA through the nanopore. We investigated the capture zone in glass nanopores under high and low ionic strength (KCl salt) using finite element analysis. We modeled a conical shape glass shape nanopore in two distinct ionic strengths (10 mM and 1 M KCl salt). DNA first diffuses around the ionic solution, once it reaches the EOF or EPF capture zone (mostly dependent on ionic strength), it is translocated through the pore. Figure 1 shows the simulation results for the capture zone of a 70 nm pore in different salt concentrations. Under low ionic strength conditions, the fluid flow map can reveal the capture zone as the DNA can be captured in the opposite direction of EPF. Additionally, the velocity field direction indicates a confined shape along the nanopore walls (Figure 1A). Under high ionic strength conditions, DNA translocates through the pore via EPF. Therefore, the electric field map can reveal the capture zone. The electric field direction indicates a spherical shape surrounding the pore mouth in EPF dominant capture zone at high salt (Figure 1B).

We investigated both contributed factors at low ionic strength (10 mM KCl) using finite element analysis. Figure 2A determines the mean fluid flow velocity at various voltages ranging from −800 mV to +800 mV. The simulation results revealed that by increasing the applied voltage from 0 mV, the mean fluid velocity also increases. For negative applied voltages, by increasing the applied voltages toward 0 mV, the mean fluid velocity decreased. The direction of fluid flow velocity was toward the pore in negative applied voltages, which resulted in EOF-driven DNA captures. In contrast, by applying a positive voltage, the fluid flow direction shifted away from the nanopore. Figure 2B schematically represents the DNA capture passing through the pore by EOF dominating the EPF transport at low ionic strength. We performed simulations for 70 nm conical glass nanopore in 10 mM KCl and calculated the net force considering both EOF and EPF dragging in opposite directions through the pore. The mean fluid velocity was used in EOF simulation and EPF was calculated based on multiplying the electric field (E) by DNA electrophoretic mobility (μ). DNA electrophoretic mobility is a function of salt concentration and experimental conditions. Therefore, to investigate the turning point from EPF to EOF dominant capture zone, four different µ values ranging from 4.25 × 10^–4 to 3.0 × 10^–4 (cm² V⁻¹s⁻¹) were selected. We chose these values according to literature reports at similar conditions and salt concentrations to our experimental conditions. Figures 2C–F demonstrate the net forces through the pore for different µ values of DNA. Initially, the direction of the net force was toward the EPF, but by reducing the µ to 3.25 × 10^–4 (cm² V⁻¹s⁻¹), the net force direction shifted toward the EOF, indicating that EOF dominated the EPF transport. The simulations demonstrated that the EPF drift velocity is extremely sensitive to µ and tuning the actual value of µ drastically affects the nanopore transport properties. Our findings indicate that the EPF decayed faster in comparison to EOF, which results in the dominance of the EOF inside the nanopore. To further explore the DNA capture mechanism at the pore, we investigated the EPF and EOF-driven net forces at various ranges of ionic strength from 10 mM to 1 M KCl for two distinct pore sizes. Simulation results for 70 and 600 nm pore size has shown in Supplementary Figures S4, S5, respectively. Our results indicated that the net force is dominated by the EOF solely at very low ionic strength (less than 20 mM KCl concentration). We determined the net force turning point from EPF dominant zone to EOF at 10 mM salt concentration. This trend was consistent for both nanopores, however, there was an increase in EOF by increasing the pore size.

Since there was a volumetric fluid flow in low ionic strength, the fluid flow rate through the nanopore can be achieved by integrating the fluid velocity over the pore area. In addition, the fluid flow rate can be obtained more precisely using a line integration method by evaluating the velocity magnitude over a cutline 2D data line spanning the width of the nanopore opening. Therefore, we determined the fluid flow rate value equal to 9.49*10^–18 m³/s through the line integration method in our developed finite element model.

We further investigated our simulation results experimentally by capturing λ-DNA at low salt using different DNA concentrations. The details of the experiments and results will be discussed in the following sections.

As mentioned before, studies indicate that the capture process involves a free energy barrier due to threading DNA ends through the pore. Since DNA is a highly charged molecule, in addition to the DNA confinement, there are some probable unfavorable DNA interactions with the nanopore itself which can form a free energy barrier for the capturing process. These energetic barriers have been investigated theoretically (Zhang and Shklovskii, 2007) and experimentally (Meller and Branton, 2002; Wanunu et al., 2010) for DNA translocation through biological and solid-state nanopores. There is an exponential dependence of the DNA capture rate on the applied voltage due to the free energetic barriers. On the contrary, multiple studies (Wong and Muthukumar, 2007; Gershow and Golovchenko, 2007) reported a linear dependence of the DNA capture rate on the applied voltage associated with a diffusion-based capture mechanism. We investigated the free energy behavior of the capture mechanism by measuring the DNA capture rate as a function of applied voltage through glass nanopores. Glass nanocapillaries were fabricated using a CO₂ laser assist pipette puller using specific pulling protocols to yield two identical nanopores. Supplementary Figure S1A demonstrates the conical shape of the glass nanopore. The taper lengths were measured under an optical microscope. The glass nanopores possessed a mean taper length of 1.5 mm. The fabrication process and nanopore measurement setup are explained in detail in the material and methods section. Briefly, the nanopipettes were filled with a low salt solution (10 mM KCl), and two silver electrodes were placed inside the nanopore as well as the solution chamber containing the same buffer including λ-DNA. The nanopore was secured inside the measurement setup (Axopatch 200B amplifier) for measuring the ionic current. After confirming the nanopore’s stability, the current-voltage (I-V) analysis was performed to estimate the pore diameter. Electrical conductance measurements have been frequently used in nanopore literature to approximate the nanopore sizes. The pore size estimation was calculated using the following equation (Smeets et al., 2006): G = σ π d i d b 4 l Where G is the electrical conductance, σ is the bulk conductivity, l is the taper length, d_i is the diameter of the pore, and d_b is the capillary inner diameter. Supplementary Figure S1B demonstrates a linear I-V relationship that indicates an ohmic electrical property with an average conductance of 2.6 nS. The approximate pore diameters were calculated around 70 nm (±5 nm), consistent with the conductance values. Moreover, the fabricated nanopores were imaged by transmission electron microscope (TEM) to confirm the accuracy of the conductance-based estimations. Supplementary Figure S1C represents the TEM image of a typical glass nanopore fabricated by the laser assist pipette puller with 70 nm pore size. The nanopore size measured by the TEM indicated the same size (±5 nm) as approximate pore sizes by the electrical conductance estimations. The geometry of conical glass nanopore is defined by the pore diameter and length of the nanopore, both be contingent on the pulling parameters. In all the experiments, nanopores were inspected for a stable open-pore current for 10–15 min before the I-V analysis.

We performed the λ-DNA translocation experiment at low ionic strength (10 mM KCl solution) and calculated the mean capture rate under multiple applied voltages. Figure 3A demonstrates the current traces showing EOF-driven DNA events at different applied voltages. The increase of translocation events in the current traces is noticeable by increasing the applied voltages. Figure 3B indicates the sketch of our experimental observation results. Principally, at zero voltage some free energetic barriers exist. These barriers will be higher by applying a negative voltage that is unfavorable in the translocation mechanism. By considering EOF plus the negative voltage, DNA event translocations are caused by some decrease in free energetic barriers. Therefore, taking both voltage and EOF from the pore entrance, the free energy will be favoring the DNA forcing into the pore. Figure 3C demonstrates DNA capture rate as a function of various voltage ranging from −300 to −800 mV. The capture rate showed a linear increase from 2.2 s⁻¹ to 17.8 s⁻¹ by increasing the voltage from −300 to −800 mV, respectively. We found a linear dependence of the capture rate on the applied voltages, which is indicative of the diffusion-limited capture process. Investigation of dwell time and current drop distributions at different applied voltages have been shown in Supplementary Figure S3. At lower voltages, the average dwell time was higher, indicating slower translocation of DNA molecules through the pore. Moreover, the results demonstrated a linear increase of the mean current drop from 40 to 64 pA by increasing the voltage from −300 toward −800 mV applied voltages.

The capture mechanism defines the process of transitioning the molecule from a diffusion-dominated region far from the pore to a voltage and/or pressure dominated region surrounding the pore. As the DNA molecule approaches the vicinity of the nanopore, the magnitude of the voltage gradient increases, however, EOF also increases with electric field. DNA velocity can be explained by the following equation in the case of EOF-driven translocations: v  =  V EOF - μrV ( r ) (1) where µ is the electrophoretic mobility of the DNA. The value of the µ does not depend on DNA length longer than a few persistence lengths. (Olivera et al., 1964; Wanunu et al., 2010) For purely EPF-driven transport, the diffusion-limited flux is equal to J _diff = R _diff c, where c is the concentration of the DNA. The same applies to EOF-driven translocations however the analytical expression for capture rate is different than EPF-driven events and does not depend on EPF drift velocity. Instead, the capture rate is dictated solely by the EOF volumetric flow which depends on the diameter of the pore and the applied voltage (i.e. EOF).

We furthered performed studies on detecting DNA molecules and quantifying the DNA concentration using our glass nanopores. We could capture EOF-driven events at low salt (10 mM KCl solution) in five distinct concentrations of λ-DNA: 100, 500, 1,000, 1,500 and 2,000 pM. Figure 4A demonstrates the current traces for the EOF-driven DNA events at an applied voltage of −700 mV for five different DNA concentrations. Translocations occurred more frequently with an increase in DNA concentration. The capture rates were calculated using two different methods. The first technique was based on calculating the interarrival time using a custom written MATLAB program. The semi-logarithmic distribution of interarrival time for each DNA concentration has been shown in Figure 4B The exponential fitting of these distributions is typically used to achieve the capture rate (Meller and Branton, 2002; Wanunu et al., 2010). We defined the time interval between successive events as delta T ( δ T) and calculated the capture frequencies by the exponential fitting of the time interval distributions: P C ( t ) = A ⁡ exp ( − t / τ ) . The average capture rates were then reported as τ⁻¹ from the fits. We derived values of the capture rates as 2.66, 3.85, 4.49, 5.37 and, 6.49 for 100, 500, 1,000, 1,500 and 2000 pM concentration of DNA, respectively. In the second method, we summed the calculated interarrival times of all the data files (60 s). The total number of captures was then divided by the overall interarrival time to achieve the captures per second units for distinct DNA concentrations. Using this approach, the capture rates were determined equal to 2.07, 3.94, 4.73, 6.21, and 7.02 for 100, 500,1000, 1,500, and 2000 pM concentration of DNA, respectively.

The EOF-driven capture rates were utilized to measure the concentration of DNA using the calculated flow rates specific to EOF in the finite element analysis model. At low ionic strength, there was a volume fluid flow; therefore, quantifying the DNA concentration were achieved by dividing the flow rate (nm³s⁻¹) over the number of DNA events (s⁻¹) captured by the nanopore sensor. The DNA concentration were obtained as the number of DNA molecules per nm³ Table 1 indicates values of the calculated DNA concentration per each of the actual values (100, 500, 1000, 1500, 2000 pM) based on method 2 capture rates. Figure 4C demonstrates a semi-linear relationship between the actual versus the calculated values for five distinct DNA concentrations. Using the capture rates from both methods, we found a linear relationship along with excellent approximation between the actual DNA concentration values and the calculated values based on the EOF-driven capture frequency and volumetric fluid flow rates.

The current drop and dwell time distributions were investigated in different DNA concentrations. Figure 5 demonstrates the current drop and dwell time distributions of EOF-driven captures in three distinct DNA concentrations at −700 mV applied voltage. The current drop peak in 100 and 500 pM concentration of DNA was around 27 pA which showed a bit higher value than 1,000 pM concentration of DNA with a 23 pA distribution peak (Figures 5A,C,E). Results indicated longer dwell time values, with the most probable translocation time at around 6 ms for all DNA concentrations (Figures 5B,D,F). Considering the pore size was around 70 nm, these large dwell time values reveal all DNA configurations, most likely, linear and folded DNA configurations. Moreover, since EOF and EPF are competing in these captured events, the net force would be lower than purely voltage-driven translocation events with no opposing forces. Therefore, as we applied the electroosmotic flow as the opposing force, the DNA transition time through the pore would be slower than in the case with no competing force. The number of EOF-driven translocation events recorded per each analysis was 932, 1445, and 994 for 100, 500, and 1000 pM concentration of DNA, respectively.