Tuning Intermolecular Interactions to Enhance Solid-State Nanopore Force Spectroscopy

Nanopores, nanometer sized holes in membranes, have recently come into prominence as tools for single molecule sensing. A technique called nanopore force spectroscopy uses the nanopore to probe energy landscapes between molecules. With the development of this technique, it will be possible for molecular recognition in complex fluids, such as blood. However, before that can be possible, solid-state nanopores, commonly fabricated in silicon nitride membranes, and having very confined sizes and charged surfaces, need to be optimized to minimize unwanted interactions between solution-phase molecules and the surface. DNA, for example, a crucial part of nanopore force spectroscopy, frequently sticks to the nanopore surface. Surface functionalization techniques, both on the nanopore and molecular surface, were attempted in this thesis work. These surface functionalization methods aimed to reduce surface charge or alter molecular properties in order to minimize the unwanted surface interactions, and they include silane modification, fluid lipid bilayer coating, and surfactant self-assembly on the DNA phosphate backbone. Results from some of these methods yield insights to improve nanopore force spectroscopy performance that will minimize the unwanted surface interactions and deliver on the promise of nanopore sensing.

is placed between two silicone elastomer gaskets (2), and placed between the two halves of an electrochemical cell (3) in the channel connecting the two liquid reservoirs (4). In the assembled electrochemical cell (5) (1) with Ag/AgCl electrodes (2) on either side, which is placed in (B) a primary aluminum (3) and (C) secondary copper Faraday cage (4), and attached to a patch clamp amplifier (5) Figure A shows a two-step blockage. The deeper blockage at the end is a signature of 5-(and-6)-carboxyfluorescein dye in this pore, and could indicate that the molecule may stay around the entrance of the pore before inserting all the way. Figure B         interfere with a molecule's function, as well as requiring time and training to attach them, in likely less than quantitative yield; with a nanopore, detection can be of the molecule directly, label-free, thereby eliminating such challenges. Nanopore sensing has received the most attention for DNA sequencing applications, promising potential capabilities such as single molecule sequencing with single-nucleotide resolution. 9,10,11 Nevertheless, substantial technical barriers remain. A technique called nanopore force spectroscopy (NFS), though, provides a means for identifying nucleotide sequence in short nucleic acid oligomers by experimentally probing energy landscapes between molecules. 12,13 This method has achieved preliminary success in a laboratory setting, but the extension of the method to a robust, point-of-use platform for detection of low concentrations of molecules remains challenging. To harness this sensitive tool requires a great deal of optimization. Preliminary results and literature demonstrate that, while fragments of DNA, for example, can be detected and identified with a nanopore, 14,15 molecular surface interactions may cause the DNA to stick to the pore wall, affecting device lifetime, ease-of-use and reliability. 16 Previous studies 17,18,19,20 have shown that the surface of the pore and DNA, though, can be modified, as can the physicochemical properties of the pore environment, such as pH and ionic concentrations, 21 and these modifications can affect the interactions between the pore surface and the molecule. Thus, judicious design of nanopore detection chemistry is one approach to overcome many of the challenges to nanopore point-of-care sensing. 4

Types of Nanopores
There are two types of nanopores, biological and solid-state. The most frequently used biological nanopore, α-hemolysin, is taken from nature and selfassembles in a lipid bilayer which can be mechanically, and frequently manually, formed across a nanometer-to micron-sized hole. 22,23 While subtle modification of a biological nanopore is feasible using the techniques of molecular biology, 24 millennia of evolution have fine tuned biological nanopores making some modifications, such as size and shape, difficult. Factors such as the pore internal diameter and the range of operating conditions in which the pores are stable cannot be easily modified and another biological nanopore or lipid bilayer compatible with the desired conditions would then need to be used instead, if a suitable combination can be found. 13 Despite the limitations of working with biological nanopores, they have been used to detect proteins, 24 small molecules, 25 metal ions, 26 and they have served as a useful test platform for developing and testing new nanopore methods. 10,12 The second, and more recently adopted, type of nanopore is solid-state. Solidstate nanopores are man-made and more tunable, allowing for a greater range of applications and a larger variety of molecules to be detected. Solid-state nanopores have already been shown to be amenable to the detection of target analytes in pure samples. 8,27,28 They can be polymer-based (i.e. polycarbonate, polyethylene terephthalate), 29 or fabricated from other materials such as silicon nitride, silicon oxide, graphene, or aluminum oxide. 8 These materials are, unlike biological nanopores in lipid bilayers, more robust 30 and stable over a larger range of conditions (pH, ionic concentration, and temperature). 13 These properties are more amenable to 5 the demands of creating a tool for medical diagnostics, which is why solid-state nanopores, in particular fabricated from silicon nitride, were chosen to be used in this study.

Solid-State Nanopore Fabrication
The silicon nitride nanopores for this study were fabricated by drilling either with an electron or ion beam. 8,31 The techniques could each readily produce nanopores of different radii, and the particular fabrication conditions were expected to produce different nanopore shapes. For the electron-beam-milled nanopores, a 200 keV transmission electron microscope was used to mill nanopores into a 30 nm thick silicon nitride chip; these pores were expected to be conical-cylindrical in shape ( Figure 1A). The helium-ion-milled pores were expected to be cylindrical in shape Nanopores are able to detect molecules through resistive-pulse sensing. 32 In the simplest sense, when a molecule inside the nanopore is able to change the nanopore conductance, it can be sensed. In general practice, the nanopore is immersed in electrolyte and a voltage applied across the membrane gives an ionic current through the pore. When the nanopore is occupied by nothing other than electrolyte, the resulting current is known as the "open pore current". However, when a molecule in the pore changes the nanopore conductance, the measured current will change.
Frequently molecules will reduce the measured ionic current magnitude, so that the signal can be described as the "blocked pore current" or "current blockage". Each type of molecule interacts with the pore chemistry differently, as some molecules will pass through the nanopore, or translocate, faster, leading to a shorter current blockage.
Other molecules are bigger leading frequently to a larger current blockage. The detailed interactions between the molecule and the pore environment will be reflected in changes in the nanopore conductance, frequently to its duration and magnitude.
The characteristic time and depth of this blockage can, in principle, be used to identify the molecule.
To use a nanopore as a sensor, first, the desired pore radius that will be suitable to the analyte molecule is chosen. The pore size needs to be large enough for the analyte molecule to travel through, yet small enough, so that there is a high signal-tonoise ratio. Small nanopore diameters place strong technical demands on the methods and experiments. The silicon nitride chip, with a nanopore drilled in it, is mounted between two silicone elastomer gaskets, and placed in an electrochemical cell ( Figure   8 2). Silver/silver chloride electrodes are placed on either side of the chip, and a voltage is applied, which will force an ion current to pass through the nanopore. Passage of a single molecule through the pore may perturb the electrolyte-only current (Figure 3), allowing detection and identification of the molecule. For complex samples, such as blood, that contain a large number of molecules of many different types, recognizing a single, low abundance protein, for example, by its current blockage characteristics, alone, is incredibly challenging. To deal with the complex samples with many molecular components expected in clinical diagnosis, a more sophisticated implementation of nanopore sensing, that leverages molecular recognition, was developed.

Figure 2. The silicon nitride chip (1) is placed between two silicone elastomer gaskets
(2), and placed between the two halves of an electrochemical cell (3) in the channel connecting the two liquid reservoirs (4). In the assembled electrochemical cell (5), the electrodes are placed in the liquid reservoirs that hold the electrolyte (6). pA blockage is shown as a rise towards 0 pA.

Nanopore Force Spectroscopy (NFS)
In 2004, Nakane et al. 10 demonstrated the new technique of nanopore force spectroscopy (NFS). Using a single, α-hemolysin nanopore, they were able to demonstrate DNA sequence discrimination with a judiciously located single nucleotide mismatch out of fourteen nucleotides causing a two-order of magnitude change in the signal magnitude. Later, in 2007, Tropini et al. 12 extended the method by performing nanopore force spectroscopy using about 100 α-hemolysin pores in parallel. The particular sensing method used, nanopore force spectroscopy (NFS), is a technique that can be used to probe the energy landscape between two interacting molecules. In these DNA genotyping experiments, it was used to probe the energy between two strands of DNA, a known probe oligomer and an unknown analyte oligomer, in order to identify the nucleotide sequence through the energetics underlying molecular recognition by the probe oligomer. 10,12,13 The energy of duplex formation between any two DNA strands is determined by the particular sequence of base pairs and the degree of complementarity between the strands, making recognition of a specific molecule possible through knowledge of one strand sequence and measurement of the duplex interaction energy. 10,12 In NFS-based profiling of a DNA sequence, special molecular constructs are required. A DNA oligomer, the "probe" molecule, possessing the complementary sequence to the target analyte oligomer is covalently linked via a longer DNA strand, the "spacer", to a biotin molecule which is subsequently bound to avidin; this is known as the probe DNA construct ( Figure 4).
The steps of NFS-based profiling of DNA ( Figure 5) are as follows. A voltage is applied so that nucleic acid components of the probe DNA construct will be electrophoretically inserted into the pore, causing a decrease in current that reports on the successful insertion. The avidin protein is larger than the pore diameter (5 nm), thereby holding the probe DNA construct in the pore without allowing it to translocate. On the other side of the pore, an analyte will hybridize to the probe DNA, with a stability determined by the degree of probe-analyte sequence complementarity.
A reversed voltage will drive the DNA duplex back in the direction of the original side. However, the pore is too small for the DNA duplex to go through, so the duplex unravels as it gets pulled through the pore. How well the analyte sequence matched the probe DNA sequence, that is the duplex binding energy, will affect the time it takes for the probe DNA to return to its original side. 10    These prior studies demonstrated the feasibility and promise of nanopore force spectroscopy-based detection. However, the limitations of the α-hemolysin platform, the inability to easily change the size of the nanopore to accommodate other sensing molecules and the structural fragility of the supporting lipid bilayer, prevent ready clinical use outside the nanopore expert user community. In this study, we therefore replace the α-hemolysin pore platform with a more flexible and more robust solid-state nanopore platform. While prior work 13 has demonstrated that the change is possible, it remains extremely challenging. Surface modification of the pore or molecules is essential to prevent unwanted interfacial effects such as DNA sticking to solid-state nanopores. Various surface chemistry techniques on DNA and silicon nitride can be found in the literature, but require vetting for the nanopore milieu. 8,17,18,19,28 The development of surface functionalization methods to engineer a more robust version of NFS for DNA recognition is the focus of this study. By measuring the interaction time between a DNA probe and DNA analyte, NFS can allow for enhanced chemical selectivity in nanopore sensing to cope with the many molecules in blood and other human fluids. More generally, NFS will be a useful technique for single molecule sensing once fully optimized.. Before a nanopore can be mounted as detailed in chapter 1, the nanopore must be properly prepared. A variety of methods are available to make the surface hydrophilic, principally through removal of organic contaminants from the nanopore surface. Piranha solution cleaning is the most common approach in the semiconductor industry to clean silicon and silicon nitride surfaces. 33 Since the solid-state nanopores used for this thesis are made in silicon nitride membranes on silicon wafers, they can be piranha cleaned. In the process, the desired surface for cleaning, in this case, a nanopore, is soaked in piranha solution (3:1 95-98% sulfuric acid (Sigma Aldrich #320501-2) to 30 wt% hydrogen peroxide (Sigma Aldrich #216763)) at 70°C for 15-30 minutes. 34,35 Nanostrip (90% sulfuric acid, 5% peroxymonosulfuric acid, <1% hydrogen peroxide, and 5% water, OMG Cyantek #539200) is a stabilized alternative to piranha solution having a longer effective lifetime allowing extended cleaning times; the duration used for nanostrip cleaning in this thesis was 2-6 hours at 70°C, depending on the particular application. After cleaning with either solution, chips were rinsed with UV-treated 18 MΩ.cm at 25°C deionized water (Millipore Synergy UV #SYNSV0000) for 5-10 minutes afterwards to remove possible residues. Both of these approaches necessitate the use of extremely corrosive solutions, that introduce operational and disposal challenges. An alternative and effective approach is to use 21 oxygen plasma. The standard approach used in these experiments was to use a Glow Research AutoGlow plasma cleaner to plasma clean the chip at 50 W for 1-3 minutes using O 2 plasma at 0.8-1.0 Torr introduced into a chamber with a base pressure of 0.2 Torr. After cleaning, the chip was mounted into the electrochemical cell ( Figure 2, Chapter 1), and filled with spectroscopic-grade ethanol (Sigma-Aldrich #245119) for about 5 minutes to allow for wetting. The ethanol was then exchanged with 9 mL of the desired electrolyte solution in both wells. These three methods for chip cleaning are explained more in section 2.5.
In the laboratory set up, the electrochemical cell was placed in an aluminum    (1) with Ag/AgCl electrodes (2) on either side, which is placed in (B) a primary aluminum (3) and (C) secondary copper Faraday cage (4), and attached to a patch clamp amplifier (5) and placed on a vibration isolation table. 24

Current Characterization of Nanopores
Two types of measurements are typically used to characterize the nanopore.
Both of these are in situ measurements, where the nanopore is mounted in an electrochemical cell, rather than an ex situ measurement such as TEM imaging, where the nanopore chip needs to be unmounted to be measured. The advantages of having the in situ measurements of the pore size is that it allows for real-time characterization of the nanopore and removes the hassle and hazards of disassembling the electrochemical cell, an occasionally rough process which can cause the nanopore to break.

Current-Voltage Plots
The first type of measurement is using a measurement of ionic current (I) versus applied voltage (V), producing an IV plot ( Figure 2). The slope of this plot gives the total conductance of the nanopore. The conductance is based on the pore size and shape, surface chemistry, and solution conductivity.
In the model used for this thesis, and supported by experimental measurements, 36 the total nanopore conductance can be expressed as the sum of two parts-the bulk and surface conductances inside the nanopore (Figure 3), 37 such that where G is the conductance.  Given the small volume and large surface-area-to-volume ratio of the nanopore, the surface conductance, arising from the effects of the charged nanopore surface in solution, can dominate the conductance at low electrolyte concentrations.
The silicon nitride surface is terminated by a mixture of Si-NH and Si-OH surface functional groups, and, thus, can be charged in solution depending on the pH. This surface charge causes counter ions to accumulate at the nanopore surface. Solution ions can then be divided between "surface" and "bulk" ions. The movement of the surface-proximal counter ions produces a noticeable conductance, described as the surface conductance. There are thus two conductive regions: the surface and the bulk solution. Knowledge of the nanopore size, shape, surface chemistry, and key solution physiochemical properties is sufficient to calculate the nanopore conductance. In the bulk solution of the pore, the conductance is given by where K is the electrolyte conductivity and r(z) is the nanopore radius as a function of distance, z, from the membrane surface (that is, the nanopore shape). In this model, the bulk conductance in the pore is the electrolyte conductivity multiplied by the volume of the bulk solution inside the nanopore. The surface conductance of the nanopore is given by where σ is the surface charge density and µ is the mobility of the counter ions. This shows that the surface conductance is the surface charge multiplied by the surface area 28 of the inside of the pore. The total conductance is the sum of the bulk and surface terms as shown by, where A bulk and B surface are functions of geometric parameters, only, accounting for the shape of the nanopore (e.g. cylindrical, conical; compare Figure 1 of Chapter 1). It is possible to determine the conductance of the nanopore at a given electrolyte concentration by measuring an IV curve at that electrolyte concentration. The data of the conductance at varying electrolyte concentrations can then be fitted to determine the values of A bulk and B surface . From these values it is possible to obtain two pore size parameters (e.g. radius and nanopore length) by solving the functional form of A bulk and B surface under the assumption of a reasonable nanopore shape. However, an exact determination of the size and shape cannot be obtained just from the conductance, because any shape with at least two parameters can yield the same A bulk and B surface .
Similarly any shape comprised of more than two unknown size parameters will result in an infinite set of possible solutions, so that any parameters above two unknowns will have to be approximated. 37,37 For characterizing a nanopore with a reasonably well-known shape, as in this thesis work, this two-parameter model with A bulk and B surface is adequate.
Besides giving the conductance, the IV curve can also show if there is contamination in the pore ( Figure 4).

Current Trace Measurements
The second measurement that can be done to analyze the nanopore in its standard laboratory configuration is to apply a constant voltage and measure the current as a function of time. In this thesis, this type of data will be known as a current trace. As described in chapter 1, the current may decrease when a molecule is in the pore because the ion flow in the pore is perturbed by the molecule in the pore.   where ΔI is the change in the current distribution peaks between the 30 minute scan and the 1 minute scan (from Figure 6), is the centerpoint of the current distribution peak of the 30 minute scan, and Δt is 30 minutes.. For the continuous voltage scan, was 1.25×10 -5 /s and for the scan every 3 seconds was -3.67×10 -6 /s.
The order of magnitude of the scan every 3 seconds (10 -6 ) was the same as the tests when the every 3 seconds scan was done first ( Figure 6B); similarly, the order of magnitude of the continuous applied voltage (10 -5 ) was the same as when the continuous applied voltage was applied after the intermittent (every 3 second) scan ( Figure 6B). When the intermittent scan was applied first, there was some drifting on the order of magnitude of 10 -6 as was with the previous scan for the application every 3 seconds, while the drifting with a continuous applied voltage was also on the same order of magnitude as the previous continuous voltage scan (10 -5 ). In Figure 6B, the data for the intermittent (3 second) scan was fit was a Gaussian curvature to smooth the data and was the center of the fit curve. Regardless of order of the measurements, the current drift following continuous current measurement was an order of magnitude higher than the drift following intermittent measurements. This may be because of resistive heating of the solution in the former case increasing the solution conductivity, in addition to any physiochemical changes to the electrodes. Finally, to control for the possible effects of electrode wetting, the electrodes were left soaking in the cell without any voltage being applied between two 1 minute current traces at 200 mV ( Figure 6C). The data shows that even when no voltage was being applied, there was still some drifting .  There did not appear to be a correlation between pH and conductance value, but each solution was independently adjusted to the desired pH and the adjustment varied from solution to solution.

Detection of 5-(and-6)-Carboxyfluorescein Dye
As an example of how a nanopore detects molecules, two types of molecules, a mixture of 5-(and-6)-carboxyfluorescein dye (M.W 376.32, Invitrogen #C194, Figure   7), were able to be detected by the nanopore. As stated in chapter 1, a molecule is detected by the change in current, a blockage, when a molecule enters a nanopore and perturbs the current. These current blockages are sometimes referred to as "events" in the nanopore literature and "blockage" and "event" may be used interchangeably. 38,39 37 The dye molecules were able to be detected in a conical-cylindrical shaped 11 nm sized pore in a 30 nm thick silicon nitride membrane.

Figure 7. 5-carboxyfluorescein dye (left) and 6-carboxyfluorescein dye (right).
Based on bond lengths, this molecule can be estimated to be about 1.2 nm in width. 40 Various concentrations were run of these carboxyfluorescein dye samples, and two types of characteristic blockages were observed. Of the total 160 blockages, 14% of the total showed a two-step blockage ( Figure 8); the other 86% were one-step blockages. Figure 8 shows

Detection of λ-DNA
One of the most popular molecules used to characterize a pore is λ-DNA, because its nanopore signal is well-established. 41,42,43,44 λ-DNA has very well defined and studied events, making it a useful molecule to determine if the particular nanopore delivers the expected single-molecule signal. λ-DNA is double stranded DNA that is isolated from bacteriophage lambda from E. coli. It is 48502 base pairs in length. 45 For these experiments, λ-DNA, obtained from New England BioLabs (#N3011L), is Solid-state nanopores used in the experiments for this thesis were able to detect λ-DNA events ( Figure 10), however, electrostatic interactions between the DNA and the pore surface caused the DNA to "stick" to the pore surface, terminating detection 43 until the nanopore could be cleaned or replaced. Long lasting blockages to the nanopore, referred to as clogging, are a severe challenge to the method. The first technique mentioned is applying a high voltage. This is usually the first method that is tried because it does not require disassembling the pore and should not damage the pore. During experiments voltages ranging from 0 to ±1000 mV are applied to the pore to try to capture molecules. When a molecule becomes stuck to a pore for an extended period of time, applying a voltage of ±1000 mV may drive the molecule away from the pore. Applying high voltage is most successful when the molecule is clogging the outside of the pore. When a clog is deep inside the pore, usually indicated by a larger depth of the current blockage, it could be more difficult to remove. A high voltage can be applied from an hour to over a couple days. After each application of voltage, the solution in each well is exchanged to fresh electrolyte solution. This is to ensure that if there are any molecules that were freed, they will not return to the pore and clog again. The possibility of the molecule being electrophoretically driven deeper into the pore or reattaching is a limitation of the high voltage method and other methods will have to be tried.
The second technique is to use Deoxyribonuclease I (DNase I), a polypeptide that will degrade single and double-stranded DNA by cleaving the DNA into small segments. 46,47,48 This method is only applicable when it is DNA that is stuck in the pore; other molecules will not be attacked by DNase. A 2500 U/mL solution of DNase (Thermo Scientific #90083) was diluted with a pH 7 aqueous solution of 1 M KCl and 10 mM HEPES in order to get a 1 U DNase solution. One unit of DNase solution is used to remove 1 mg DNA in 10 minutes, and the effect will decrease with high amounts of K + ions in the solution. 46 While about 1 mg of DNA, requiring 1 U DNase to digest it, had been injected initially into the cell, the electrolyte solution in the wells was exchanged to remove the free solution DNA before injection of 2 U DNase, so that the substantial excess of DNase was to target only the single molecule of DNA stuck to the nanopore over the 15 minute treatment time. The electrolyte was subsequently exchanged for fresh solution to remove the DNase and partially digested DNA. This method worked 1 of the three times it was employed, but its function was not systematically examined.
The use of phosphate buffer (Sigma Aldrich #P5655) is not a very common technique being that hot phosphoric acid is used to etch silicon nitride; the increasing conductance of silicon nitride nanopores has been used to suggest physical pore growth through chemical etching when the pores have been immersed in neutral, room temperature phosphate buffer. 49 A systematic study, however, has not been performed to determine if it was phosphate buffer etching or an artifact of the particular nanopore fabrication conditions leading to apparent size instability. Due to the possibility that the phosphate buffer will etch the silicon nitride, this third technique is mainly used for silicon nitride pores that are closed off, which may happen after drilling a pore, when the silicon nitride may structurally relax to fill the hole again, or if a pore is not thoroughly drilled. A closed off pore is indicated by the current being at or close to zero when a voltage is applied. Trials were performed with a pH 7 aqueous solution of 0.1 M KCl and 10 mM phosphate buffer. The phosphate buffer then replaced the 47 usual HEPES in the solution. Voltages of ±800 and ±1000 mV were applied to the cell for times varying from one to three hours. Of the five trials performed to try to open a closed silicon nitride pore, none of them worked. While the high applied voltage could have heated up the phosphate buffer, especially in the pore given its high-resistance, the concentration might have been too low to etch. Overall, this was not an effective method for opening the pores tested.
As mentioned in section 2.1 of this thesis, O 2 plasma cleaning, the fourth method, will remove organics from the silicon nitride chip. This technique, as well as the following ones, is usually used as an extreme procedure, because it requires disassembling the electrolyte cell setup. Due to the roughness of the handling, it is possible that the nanopore membrane may break during the disassembly process. For cleaning a nanopore clog with oxygen plasma cleaning, the standard initial mounting conditions are used first: 50 W for 1-3 minutes, as outlined in Section 2.1. Afterward, the pore is remounted, allowed to wet with the electrolyte solution overnight, and an IV curve is used to test whether or not the pore opened. Only if this does not work are more intense conditions applied: 100 W for 1-5 minutes or 10 W for 5-10 minutes.
This method has worked 17 times out of 26 in removing clogs.
The final methods are using piranha solution and Nanostrip solution. As mentioned in section 2.1, piranha is a 3:1 mixture of 95-98% sulfuric acid to 30 wt% hydrogen peroxide. This is a strong oxidizing agent, which is why it is very useful to remove unwanted materials that could be inside the pore. While these final methods have not been used to remove clogs, they have been used to try to open pores that were not thoroughly electron-or helium ion milled. Limited stability and lifetime 48 requires that piranha solution be made fresh for each use. In contrast, Nanostrip is a stabilized formulation that allows for storage for extended periods of time and for longer treatment times approaching 2 hours or longer at 70°C. The necessity for extensive dilutions and rinsing when using either piranha or Nanostrip solution introduced nanopore chip handling challenges and frequently resulted in broken nanopores when using tweezers for manipulation. A chemically resistant Teflon holder ( Figure 11) was designed to securely hold the nanopore during Nanostrip cleaning.
While the Nanostrip was unsuccessful to open pores that were not thoroughly electron or helium ion milled, there is still a possibility that it may be useful to remove clogs, because it is a strong oxidizing agent.

Figure 11. A Teflon chip holder designed for Nanostrip solution cleaning to reduce handling damage. The Teflon is stable in the Nanostrip solution so it will not degrade. The silicon nitride chip is sandwiched between two Teflon pieces (A) and they are screwed together with Teflon screws (B). The holder with the chip is
submerged into the Nanostrip solution.

Nanopore Surface Functionalization with Silane
Control over the surface properties of a nanopore will help to direct interactions with various analytes. 50,51 Surface modification of the nanopore can be done to be able to control the chemical functionality, hydrophobicity, and surface charge. Another benefit of coating an inorganic pore, such as silicon nitride, is that it will make the pore more biologically compatible. Wanunu et al. 50 performed a series of surface coatings on silicon nitride nanopores, similar in size and shape to the nanopores used frequently within this thesis. In their work, one of the surface functionalizing molecules they investigated is methoxyethoxyundecyltrichlorosilane (1, Gelest, Inc. #SIM6491.5, Figure 1A), which is a methoxyethylene gycol, or a polyethylene glycol-like (PEG, Figure 1B) molecule. PEG modified surfaces are known to lower interactions between biomolecules and the surface. 52,53 As DNA tends to interact with the silicon nitride surface, previously published literature shows that PEGylation of the surface will minimize this interaction, and that without PEG, longterm clogging of the nanopore will occur. 54 Therefore, coating a nanopore with PEG would be a good option to lower any unwanted interactions and prevent DNA from sticking to the nanopore. It was hypothesized that by surface coating the pore, it would lower any unwanted interactions, allowing for nanopore force spectroscopy (NFS, see chapter 1) to be routinely possible in solid-state nanopores.

Silane Functionalization Methods
Silane functionalization of nanopores was first reported by Wanunu et al.Error! Bookmark not defined.. They pursued two different routes: so-called ex situ coating on large (~10-25nm in diameter) and in situ coating on nanopores smaller than 10nm. In the former approach, nanopores were simply soaked in a bulk solution of the reagents, whereas in the latter approach, the nanopores were exposed to reagent in a typical nanopore sensing configuration and applied voltage was used to drive the reagents into the nanopore. In an extension of this work, coating with 1 was tried in situ on a large pore whose diameter, based on conductance, was estimated to be about 88 nm. It is important to note that 1 will polymerize in water. Therefore, certain steps must be taken to avoid this. Once the pore is mounted into the electrochemical cell conductance was essentially zero, indicating that the nanopore was closed. After extended soaking in the methanol-based TBACl electrolyte and return to the aqueous potassium chloride-based electrolyte, the conductance was no longer zero, but was 57 rather 4 nm. In both instances of in situ silane coating, a measureable reduction in the nanopore conductance, suggesting a reduction in its diameter, was recorded. The inability to detect DNA in the larger pore, coupled with the variability in modeldependent inferred nanopore diameter, suggested the need for additional characterization steps, however. A sample of lambda DNA was used in an attempt to characterize the 7 nm bare to 4 nm coated nanopore, as had been attempted previously for the 88 nm bare to 63 nm nanopore. 58

Figure 2. A) An ideal depiction of what a silane-coated nanopore will look like. B) A
depiction of what could happen when 1 polymerizes the pore.

Probe DNA Construct Preparation
Given the lowered conductance of the 4 nm silane-treated pore, single-stranded probe DNA, biotinylated at the 5' end and coupled with neutravidin (hereafter referred to as the "probe DNA construct"; see section 1.3 for description, and Figure 4 in Chapter 1) was used to characterize the pore. The purpose to using the probe DNA construct instead of the well-characterized λ-DNA is that the probe DNA construct is an important element in NFS and can be held in the nanopore for extended measurements of both the nanopore and the nanopore-DNA interaction. The hypothesis was that surface functionalization would make NFS possible, so characterizing with the probe DNA construct was the next logical step toward NFS.
The probe DNA construct was prepared first by making a neutravidin stock solution. Neutravidin, obtained from Thermo Scientific (#31000), has a lower isoelectric point and a higher specific binding to biotin than avidin. 55  overnight; for long term storage, the probe DNA construct was stored at -20°C.
Usually about 2-5 µL of the probe DNA solution was injected into a 500 µL well to give a final concentration of about 100 nM.

Characterization of Probe DNA Construct in the Silane-Coated Pore
Using the apparently silane-modified 4 nm pore, the probe DNA was run with an applied voltage of -200 mV. Current blockages indicating insertion of the probe into the nanopore were not detected, and after about an hour of applied voltage, the apparent pore size had been reduced to 2.8 nm. The underlying reduction in nanopore conductance could have been caused by DNA or protein sticking to the pore, or due to changes to the silane coating. The combination of no observed current blockages consistent with molecule insertion into the nanopore, the decrease in conductivity of the pore after silane coating, and the possibility of DNA sticking to the silane-treated pore, could support the existence of a silane "coating" as shown in Figure 2. The decrease in nanopore conductivity, in that instance, was caused not by a uniform coating of the nanopore edge, reducing the diameter by a fixed amount, but by the random filling of the nanopore that prevented molecular insertion and reduced the conductance to provide a reduced effective nanopore diameter. This could have happened by silane polymerizing in the bulk of the pore ( Figure 2B), which allowed for ions to pass through to give a measurable current, but not for molecules, such as DNA, to pass through. Further polymerization of the silane in the aqueous electrolyte could have been the reason the pore size decreased after about an hour.
Based on the results obtained here, it can be concluded that coating the outer rim of the nanopore with a monolayer of 1 in situ, so that the nanopore can be used as a single molecule detector, did not successfully work when reproducing the literature procedure. While silane coating of planar surfaces is itself frequently challenging, the conformal silane coating of a nanopore is even more dramatically challenged by its reproducibility. 56 For example, while the literature 50 reports that the coating on the rim of the nanopore should be 2.2 nm, the experiment here showed that the longer 1 stayed in the pore, the more the size decreased. Overall, these challenges make it difficult to control the silane-based functionalization of a nanopore, and alternative approaches are reported to offer the possibility of greater control and reproducibility.

Nanopore Surface Functionalization with a Fluid Lipid Bilayer
Inspired by lipid-coated nanopores in insects, Yusko et al. 57 coated nanopores with a fluid lipid bilayer to prevent clogging. In their work they used a few different types of lipids to coat nanopore walls, including 1,2-dilauroyl-sn-glycero-3phosphocholine (DLPC), 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC). For the work in this thesis, POPC was used, because it is an electrically neutral bilayer over the pH range that was used in this work (pH 7-7.5) and it was shown to have anti-clogging capabilities. 57 However, at very acidic conditions (below pH 4) the bilayer becomes positively charged, and at very basic conditions (above pH 10) the bilayer becomes negatively charged causing the bilayer to fold. 58 The end-to-end length of POPC is 1.9 nm, so that it is expected to physically reduce the nanopore diameter by about 8 nm. in solution for about 20 minutes, and the solution was refreshed. Afterwards, the pore was about 18 nm in diameter ( Figure 3B). This is a much larger change than expected, because the bilayer thickness is reported to be about 3.7 nm in the literature, so an overall 8 nm change in diameter should have occurred. 57 However, besides indicating the decrease in size, the IV curves also showed that something was in or on the pore because it was asymmetrical after the POPC injection ( Figure 3B). It was assumed that POPC had coated the pore, and to test this λ-DNA was used to characterize it. 10 µL of 1.70 nM λ-DNA (see chapter 2.2 for preparation) was added to one side of the well to a final concentration of 33 pM. A total of 182 events recorded in the current trace at -200mV were automatically extracted using a current threshold algorithm. A characteristic event is shown in blockage magnitude, which is closely supported by the observation here. Other methods for coating the nanopore were attempted without demonstrated success. One such method was to allow the POPC to sit in the well for longer than 20 minutes, and instead sat for about 60 minutes; this was tried 12 times without success. Another method was disassembling the electrochemical cell and placing a 150 µL drop of 2 mM POPC on top the nanopore chip directly and allowed to sit for 15-60 minutes; this 66 was tried 3 times without success. Finally, an increase of concentration of the POPC from 2 mM to 20 mM was used instead; this was tried once, and did not work. A reason that this might not have worked is that no pressure was being applied across the membrane (see Section 3.3), and the liposome was expected to float gently onto the pore and collapse to form a bilayer coating the surface. Applying a voltage during this process in the future may allow the liposome to reach the pore. Another challenge of applying voltage is that the fluid lipid bilayer may go through the pore instead of coating it. From the single successful trial, coating a pore with POPC appears to prevent sticking from occurring, and if perfected will be a very good surface functionalization method to stop clogging. It requires a change in the apparatus to allow a pressure difference to be applied across the nanopores, however, and this work is underway.

Formation of a Lipid Bilayer in a Solid-State Pore with the Nanion
The Nanion Port-a-Patch 60 system affords the opportunity to explore the interaction of pressure-driven lipid bilayer vesicles with micropores before the more demanding interaction of 50 nm vesicles with ~10nm nanopores. The Nanion system applies a pressure gradient across a glass micropore with diameter close to 1 micrometer, and thereby draws in ~30 micrometer giant unilamellar vesicles (GUVs) that have been formed via electroswelling in a Nanion Vesicle Prep 61 Pro system.
This system allows the straightforward investigation of solution and GUV composition and resulting lipid bilayer properties, in a configuration compatible with nanopore sensing. In this testing configuration, the bilayer can be suspended across the open, approximately 1 micrometer borosilicate glass aperture of the disposable micropore. While it is possible to monitor the bilayer stability simply by applying a voltage and measuring to see if a measured current indicates bilayer rupture, it is also convenient to introduce α-hemolysin nanopores in the bilayer so that the trans-pore current can be used to monitor the bilayer stability. A stable, well-characterized ahemolysin conductance will then indicate acceptable bilayer properties. Since the first documented use in 1996 by Kasianowicz et al. 62 α-hemolysin nanopores have been the biological nanopore of choice to analyze single stranded DNA and RNA molecules. 63,64,65,66 3.

GUV and Bilayer Formation
The pores for this work were micropores with a resistance of 3-5 MΩ on glass chips (Nanion NPC-1). Along with pore size, shape, and surface chemistry, the resistance is an indication of whether the micropore is open or closed, and hence, 70 whether a bilayer had formed a seal over the top. IV curves can be used with these micropore chips to measure the resistance, as in the standard approach to characterizing nanopores. Bilayers were formed using giant unilamellar vesicles

77
The bilayer formation process needed optimization. A few changes were made to try to make a more stable bilayer in the micropore. The first was the electrolyte concentration; it was decreased from 1.0 M KCl to 0.1 M KCl, because the higher salt concentration might have been interfering with the stability of the GUVs. The second change was to create a pH gradient from one side of the micropore to the other, where the pH on one side of the chip would be pH 7.0 and on the other side it would be pH 6.0, because this method was previously shown to stabilize the bilayer on the chip. 71 Of the 17 trials performed at 0.1 M KCl with a pH gradient, 6 formed bilayers that lasted at least 5 minutes, with the longest lasting about 20 minutes under an applied voltage of +100 mV. If the bilayer did form, the side with pH 6.0 would be exchanged to pH 7.0 after about 5-10 minutes, so as to run experiments with symmetrical pH on both sides. However, these bilayers were still not stable, and broke during perfusion.
Of the six bilayers formed, four broke before perfusion did take place and one broke when perfused.
One trial did remain stable for about 30 minutes after formation, and αhemolysin-containing solution was injected to form pores. The α-hemolysin solution was injected into solution on top of the micropore so that the final concentration would be 5.0×10 -4 mg/mL. 72 To prepared the α-hemolysin solution, 1.0 mL of 18 MΩ.cm deionized water was mixed with 1 mg of powder α-hemolysin to give a 1.0 mg/mL stock solution of α-hemolysin. Then this stock solution was diluted to make a concentration of 1.5×10 -3 mg/mL.
The incorporation of α-hemolysin into the bilayer was initiated by placing 5.0 µL of the 1.5×10 -3 mg/mL solution on one side of the micropore. If α-hemolysin incorporated into the bilayer, it would be indicated by a characteristic increase in the measured current. However, this did not happen for this particular trial, so it may be that a multilayer formed. A multilayer is also evidenced by how the current remained stable after electroporation and applied pressure up to -200 mB was attempted-that is, how both of these perturbations failed to rupture a multilayer with greater mechanical stability than a bilayer. It is clear from all these trials that further optimization will need to be done for the formation of a stable lipid bilayer in the micropore.
The goal of using the Nanion system is to form a stable lipid bilayer with αhemolysin nanopores. It will be possible to characterize DNA with these nanopores.
While the Nanion will ease the bilayer formation process, there is still some configuration that must be done to be able to create reproducible and stable bilayers. The sticking of λ-DNA to the silicon nitride surface has been a major problem in the majority of the work presented in this thesis and within the nanopore research community, because for the single molecule detector to work, the molecule being detected needs to translocate through the pore, or to interact with it only transiently. If the molecule sticks, it will, firstly, interfere with other molecules from entering the pore, and secondly, the pore might become permanently clogged so that a new pore will be needed. While a number of possible cleaning techniques have been presented in chapter 2, it would be ideal to have a nanopore detection scheme that requires little maintenance while providing high functionality. In chapter 3, different surface functionalization techniques on the silicon nitride nanopore were introduced, and from those results it can be understood that a lot of fine-tuning must be done to create the vision of a robust, easy to use, rapid, and inexpensive tool. In previous chapters, we have explored controlling the molecule-surface interaction through functionalizing the surface of the nanopore. In this chapter, instead of functionalizing the surface of the nanopore for control of the molecule-surface interaction, functionalizing the surface of the analyte molecule, in this case, DNA, will be done. Not only does this approach have the potential to control unwanted molecule-nanopore interactions, it also allows for control over the physicochemical properties, such as charge, of the molecule being sensed. In addition to modifying the surface chemistry, self-assembly approaches will allow us to tune the effective size of the DNA molecule. It is important to note that both molecular size and surface charge are key parameters in determining the measured signal in nanopore single-molecule sensing. In addition, since the NFS method manipulates molecules through their effective charge, using self-assembly of cationic surfactants on the negatively-charge DNA polymer to tune the net macromolecule charge, allows for molecular-level control of the technique. The detection of simple surfactant-driven self-assembled complexes here are also a proxy for the study of more sophisticated molecular assembly approaches, for example, the detection of site-specific binding of proteins to DNA or the measurement of whether or not a drug molecule has bound to a protein surface. In addition to these fundamental aims, this molecular surface functionalization approach serves an extremely practical purpose. Since DNA is the molecule sticking to the surface, it might be possible to reduce the charges on the DNA molecule in order to reduce the interaction between the pore surface and the molecule.
The strong electrostatic interaction between the cationic surfactant cetyltrimethylammonium bromide (CTAB, Sigma #H9151) and DNA has long been exploited to selectively precipitate DNA during DNA extractions. 73,74,75 The interaction received greater scrutiny in the context of DNA-based molecular electronics, in which literature reports revealed the well-ordered molecular structure of this interaction. Previous literature on the molecular-level fabrication of DNA has shown that it is possible to create a film using DNA, which has a negatively charged phosphate backbone, complexed with a positively charged cationic surfactant. The ready formation of films of surfactant:DNA complexes allowed for x-ray diffractionbased studies that revealed a highly ordered structure of surfactant brushes on a DNA polymer. 74 Thus, in addition to charge neutralization of the phosphate backbone, the CTAB may also allow for a rigid structure of DNA (Figure 1). The rigid structure will alter how the complex enters into the pore, compared to the more flexible DNA polymer. As stated in Section 2.4, the flexible DNA polymer can bend and fold in many ways, allowing a variety of conformations of DNA to enter the pore, and each conformation has a different characteristic translocation event through a pore. 76 By making the DNA rigid, it will decrease the variety of conformations. The CTAB:DNA complex will also increase the radius of the structure, as the length of CTAB is about 2.2 nm. 77 Twice this length will add on to the original 2 nm diameter of double stranded DNA. 73 Being able to the change the size of the molecule will possibly allow for a larger range of more easily fabricated pore sizes to be used.   Other ratios were tested, namely 1 CTAB for every 500, 200, 100, and 50 base pairs.
Testing of these latter ratios (1:500, 1:200, 1:100, and 1:50) used a 19 nm nanopore in an aqueous solution buffered to pH 7 with 10 mM HEPES containing 1 M KCl. For these latter ratios, 100 µL of CTAB:λ-DNA complex was injected into the pore, yielding a concentration of 282 pM of λ-DNA in the 0.5 mL well. In the increased pore size, the 6 nm diameter of the CTAB:λ-DNA complex would be about a third of the pore size. This would be a noticeable current blockage in the current trace based on equation 2.5, repeated below, but, in fact, there were no translocation events detected.
Instead, in one trial in the 19 nm nanopore with 282 pM of λ-DNA with 1 CTAB for every 500 base pairs (1:500) in a pH 7 aqueous solution of 1 M KCl and 10 mM 88 HEPES, the current only blocked about a third to a half of the expected blockage for an extended period of time, never unblocking, indicating that the pore became clogged. This may be an indication that the 1:500 ratio was not great enough to sufficiently minimize electrostatic interactions between the DNA and the nanopore surface, and that more CTAB would be needed. It may be that at the higher ratios, the decrease of the molecular charge lowers the applied force on the molecule, so that it is not readily drawn into the pore for translocation; it could also be that the increasing rigidity from the CTAB brush structures also diminishes the likelihood of DNA capture. In the electrostatic sense, the applied force on a molecule depends on factors such as the molecular charge and the applied voltage, such that where F is the applied force, z×e is the effective molecular charge in the pore, where z is the effective charge on each DNA phosphate group and e is electron charge, and V is the applied voltage. From the equation, a decrease in molecular charge will also decrease the applied force. Therefore, since CTAB decreases the molecular charge on the DNA, a higher applied voltage is need to be equivalent to the applied force at lower voltage without CTAB on the DNA.

Probe Escape
Since there were no translocation events that were detected and extended blockages indicating clogging at the ratios used for the number of CTAB to λ-DNA base pairs suggested that the ratio of number of CTAB molecules to base pairs needed more fine-tuning to determine the perfect ratio to prevent the molecule from sticking 89 to the nanopore surface. Another problem could be that since λ-DNA has so many bases, the CTAB molecules might not be as evenly spaced as desired and there could be large gaps between clusters of CTAB molecules. These large gaps would leave bare DNA that could then stick to the nanopore surface. In addition, a stated 1:1 ratio of CTAB to lambda DNA left 50% of phosphates without charge neutralization. In response to this, the shorter 94 nucleotide probe DNA construct was used instead (see section 3.1.2 for probe DNA construct description and preparation, Figure 2), because instead of large gaps of perhaps hundreds of nucleotides in the case of non-random CTAB distribution on the polymer, there would be much smaller gaps, thereby lowering the chance of sticking. Also, the probe DNA construct is used in NFS, which is ultimately the goal of this research: to perfect NFS in solid-state nanopores to a reliable single molecule sensing technique. Using the probe DNA construct to analyze the system is a preliminary step toward this goal.

91
To begin, a mixture of 10 CTAB molecules to 94 bases was used in an experiment called probe escape. Literature reports indicated that probe escape had been previously demonstrated on bare, fully charged, DNA when using both biological and solid-state nanopores. 78,79 In probe escape (Figure 3), a voltage, known as the "capture voltage," is applied to drive the probe DNA construct into the pore. Once the probe is in the pore, the voltage sign is unchanged, but its magnitude is decreased, known as the "hold voltage," and it is held at that value for 2 seconds or until the probe escapes, a process whereby the probe leaves the pore by escaping against the hold potential and returns to the original side. If the probe did not escape after 2 seconds, the voltage was reversed to try and force the probe back to the original side. For the probe escape experiment, the CTAB/probe DNA construct mixture was injected into the reservoir on one side of a 3 nm pore to give a concentration of 100 where f(V) is the applied force in Newtons, z is the formal charge on each base inside the nanopore where e is the electron charge, n is the number of nucleotides in the pore, L is the length of the pore in meters, so that n/L is the inverse of nucleotide spacing (a constant of 0.45 nm -1 ), and V is the applied voltage. By complexing the anionic phosphate groups with cationic CTAB, the average magnitude of z is decreased so that if the section of DNA polymer constrained inside the pore is CTAB-coated, a given applied voltage will exert a lower force on that complex than if the DNA were uncoated, or if the partial CTAB coating were not inside the nanopore. It was expected, therefore, that these hold voltages, while higher in magnitude than the literature results for bare DNA, would nevertheless allow probe escape. The time taken for the CTAB:DNA complex to exit the pore, however, was much faster than for the probe escape experiments published in the literature. 78,79 The timescales were in fact more in keeping with the literature timescales for rupture of the neutravidin-biotin bond 80 via NFS. It was likely, therefore, that the incomplete coverage of the DNA molecule with CTAB, or perhaps voltage-induced stripping or migration of the CTAB molecules, allowed for entrained CTAB:DNA constructs to feel an applied force somewhat comparable to that of bare DNA. Alternatively, a CTAB-coated DNA molecule could possibly be inserted into a nanopore, but come to rest with a bare section in the nanopore, at least some fraction of the time. Based on the pore clearance time suggesting rupture of the biotin-avidin bond instead of probe escape, the experiment was reconfigured to attempt to directly measure this timescale.

Neutravidin-Biotin Rupture
In previous literatureError! Bookmark not defined. the neutravidin-biotin bond rupture was performed using bare DNA in solid-state nanopores and with voltages ranging from 400 to 900 mV. These published experiments were challenged by DNA-nanopore sticking, in keeping with the experimental results observed here.
For this thesis, the results of the neutravidin-biotin bond rupture using DNA coated with CTAB were compared to the literature values using DNA free of CTAB. In the rupture experiment (Figure 4), the probe DNA molecules (CTAB-free in the literature reports and CTAB coated here) are on one side of the pore and a voltage is applied to drive one into the pore. Once the probe DNA molecule enters the pore, the current will decrease. Whereas in the probe escape experiments the capture voltage is switched to the hold voltage once the molecule enters into the pore, in the bond rupture experiments, the voltage is unchanged. After some time, the neutravidinbiotin bond will rupture, which will cause the probe molecule and the neutravidin 96 molecule to go to separate sides of the pore. The time required for this bond rupture can be used directly to uncover the neutravidin-biotin binding energy.
The ratio for the neutravidin-biotin bond rupture experiment was 30 CTAB molecules to 94 bases (30:94). Based on concentrations given in the literature,Error! Bookmark not defined. 5 µL of the 30 CTAB to one probe DNA strand was injected to one side of the well to give a concentration of about 100 nM in the well. The voltage applied for rupture was -500 mV, and 66 ruptures were captured in one experiment ( Figure 5). These 66 rupture times varied from 0.002 to 25 seconds.  In Figure 6, the distribution of the dissociation times for the data with CTAB is

CONCLUSION
In this thesis work, different methods for surface functionalization were tried to control the interaction between solution-phase molecules and silicon nitride nanopore surfaces, advances necessary to deliver the full potential of nanopores as robust, lowcost tools for single-molecule sensing. Solid-state nanopores can be used in a technique called nanopore force spectroscopy (NFS) to probe the energy landscape between molecules for the possibility of molecular recognition. 81,82 This technique can eventually allow for a label-free, robust, and easy-to-use single molecular sensor with little maintenance or replacement of parts. The challenges in working with a solid-state nanopore, especially due to its confined size and charged surface, have led to the need for optimized methods to cope with nanopore surface chemistry for the improvement of NFS performance. This thesis work has highlighted some of the technical issues that must be solved.
Research grade nanopores, which are frequently bare silicon nitride in contact with solution, can be used to sense a wide variety of molecules but typically require a great deal of user intervention, such as cleaning, or frequent replacement of the nanopore. A number of methods exist to clean nanopores. Piranha solution is the most historically used approach, but the work in this thesis found that both Nanostrip and oxygen plasma produced very good results with less risk and hazard. The solution based nature of Nanostrip requires extensive washing steps, and this necessitated used of a Teflon holder to prevent mechanical damage to the nanopore during cleaning and rinsing. Oxygen plasma also used a holder to secure the nanopore, but the gas-phase nature of the process required much less handling and device cleanup. Plasma cleaning was especially useful for rendering nanopores hydrophilic for mounting in the fluid holder.
While the literature 83 contains a report of the use of silane chemistry to coat a nanopore with methoxyethoxyundecyltrichlorosilane (1), in practice this method remains incredibly challenging to reproduce and to implement in controllable fashion.
The results in this thesis suggest that rather than uniformly coating surface of the pore, 1 likely readily polymerizes in the bulk of the pore, so that while still permeable to small ions, larger analyte molecules could not enter the pore and could not, therefore, be detected.
Use of non-covalent lipid bilayers to coat the nanopore surface was more successful, allowing for detection of DNA, whose slow translocation dynamics were consistent with literature reports for molecular transloation 84 through such fluid lipid bilayer coated pores. Yet, the reproducibility of bilayer coating of the nanopore was low, motivating the need for improvement to the nanopore setup, such as the ability to apply a pressure differential to deliver greater control over the bilayer coating process.
These modifications are currently underway.
The use of a glass micropore allowed lipid bilayer pore interactions to be studied without extreme challenges accompanying typical nanopore experiments. The ability to apply a pressure differential across pores allowed for bilayers to be quickly formed across the pore opening, supporting the addition of pressure differential 104 capability in the nanopore experiments. Initial investigations using this micropore platform underscore the complexity of lipid bilayer coating of much smaller pores and show the need to carefully control lipid and electrolyte composition.
Throughout this thesis work a major limitation to achieving the full potential of nanopore methods has been prevalence of unwanted interactions between the charged molecules and charged nanopore surfaces. While this thesis work has provided insight into improved methods for nanopore surface passivation, surface functionalizing a highly constrained nanopore surface remains incredibly challenging. Suppressing unwanted interactions between molecule and pore by decorating the surface of the molecule, which is freely accessible in bulk solution, offers compelling possibilities.
While we used non-specific self-assembly approaches, it is possible to imagine using sophisticated molecular recognition techniques to decorate the molecule and, thereby, simultaneously suppressing unwanted interactions and conferring sensing selectivity.
In this thesis work, we used surfactant self-assembly on the DNA backbone to both change net macromolecule charge and its effective cross-section. This tuning of charge allowed us to adjust at the molecular level the interaction between the applied voltage and the DNA complex, an interaction dependent on the net charge of the DNA. By charge neutralizing a fraction of the DNA oligmer's charges we demonstrated through the NFS measurement of the avidin-biotin rupture time that this molecular tuning was indeed effective. Thus, this decoration of the analyte molecule holds promise for improving NFS performance and allows for performance enhancements that can be delivered by tuning the effective analyte surface by using the powerful tools of molecular design. Once the surface chemistry is optimized, NFS