PREPARATION AND CHARACTERIZATION OF SOLID STATE NANOPORES FOR DNA SENSING APPLICATIONS

Transcription

1 PREPARATION AND CHARACTERIZATION OF SOLID STATE NANOPORES FOR DNA SENSING APPLICATIONS By KAAN KECECI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA

2 2010 Kaan Kececi 2

3 To my parents, Burhan and Serpil Kececi 3

4 ACKNOWLEDGMENTS I would like to thank my advisor, Dr. Charles Martin, for his guidance and support during my time at the University of Florida. I am grateful to have had an advisor who allowed for independent thinking and scientific creativity. I also thank Dr. Martin for teaching me how to be a professional scientist. Working in the Martin group has truly been a pleasure. I acknowledge the entire Martin group for their support, assistance, and friendship. I would like to thank Drs. Zuzanna Siwy, Lane A. Baker, Youngseon Choi and Hitomi Mukaibo for their insightful advice and experimental ideas. I thank, Lindsay T. Sexton, Pu Jin, Lloyd Horne, Funda Tongay, Otonye Braide, Jillian Perry, Mario Caicedo, Fatih Buyukserin, Gregory Bishop, Ramiro Palma, Stefanie Sherill, Matt Porter, John Warton, JaiHai Wang, Fan Xu, Heather Hillebrenner and Damian Odom for sharing their ideas and offering support on my projects. Last but not least, I thank my dear friends, Basri Gulbakan, Cem Demiroglu, Ibrahim Sookur, Nezih Turkcu, Enes Eryarsoy, Atay Kizilarslan and Tahir Bayrac for their support for helping me to have an enjoyable time. I would like to thank Neslihan Fidan for her endless support. I am also grateful to Karen Kelly from ICBR for her help with SEM, and to Eric Lamber and Kerry Siebein from MAIC for their help with XPS and TEM. 4

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS... 4 LIST OF TABLES... 8 LIST OF FIGURES... 9 ABSTRACT CHAPTER 1 INTRODUCTION AND BACKGROUND Introduction Fabrication of Conical Nanopores The Track-Etch Method Formation of Latent Ion Tracks Ion Track Etching Materials Conical Nanopore Characterization and Properties Electron Microscopy Electrochemical Measurements Electric Field Focusing Ion Current Rectification Tailoring Surface Chemistry of Conical Nanopores Electroless Gold Deposition EDC/Sulfo-NHS Chemistry Biological Nanopores Resistive Pulse Sensing Other Nanopore-Based Sensing Strategies Dissertation Overview PREPARATION AND CHARACTERIZATION OF CONICAL PET NANOPORE Introduction Experimental Materials UV Irradiation EDC Modification Pore Etching and Nanopore Fabrication Electrochemical Measurements and Data Analysis XPS Analysis Results and Discussion Effect of UV Irradiation on Breakthrough Effect of Electrolyte Concentration on Tip Size

6 Rectification Behavior of PET Nanopore at Different ph Surface Modification and Analysis of PET Nanopore Conclusion RESISTIVE-PULSE SENSING OF DNA USING CONICAL PET NANOPORE Introduction Materials & Methods Materials Pore Etching EDC Modification Electrochemical Measurements Results and Discussion Current-Pulse Data Effect of Concentration on Current-Pulse Frequency (ƒ b ) Effect of Potential on Current-Pulse Frequency and Duration Effect of Tip Diameter on Current-Pulse Frequency, Amplitude and Duration.. 73 Effect of Electrolyte Concentration on DNA Translocation Effect of High Potential on Current-Pulse Amplitude and Duration Conclusions DETECTION AND DIFFERENTIATION OF DNA USING CONICAL PET NANOPORE Introduction Materials & Methods Materials Pore Etching EDC Modification Electrochemical Measurements Results and Discussion Effect of Tip Diameter on DNA Current-Pulse Amplitude and Duration Differentiation of Analytes Using Current-Pulse Signatures Conclusions NANOCONE FABRICATION USING ASYMETRIC ETCHING CONDITIONS Introduction Material & Methods Materials Chemical Etching of Conical Nanopores Electroless Gold Plating Liberating the Gold Nanocones Results and Discussion Effect of Strong Acid on Cone Angle Effect of Weak Acid on Cone Angle Effect of Potential on Cone Angle

7 Conclusion CONCLUSION LIST OF REFERENCES BIOGRAPHICAL SKETCH

8 LIST OF TABLES Table page 2-1 Calculated tip diameters in 0.1 M KCl and 1M KCl Percent atomic composition of PET membrane before (bare) and after modification with ethanolamine

9 LIST OF FIGURES Figure page 1-1 Diagram of the ion track-etching method Diagram of the electrochemical cell used for chemical etching and all electrochemical measurements Schematic of a conical nanopore in a polymer membrane showing the base diameter and tip diameter Plot of current versus time recorded during the first step of conical nanopore formation Track-etching of a conical pore, showing bulk etch rate, B, track etch rate, T, and cone half angle, Chemical structures of polymers typically used for track-etching Scanning electron micrographs of nanopores track-etched in various polymer materials Scanning electron micrograph of conical gold nanocones formed by deposition into a conical nanopore membrane A typical current-voltage curve used to calculate the tip diameter of a conical nanopore Distribution of the electric field across a conical nanopore Chemical structure of PET and formation of chain ends after chemical etching Model of ion current rectification Schematic of Au electroless plating procedure Reaction mechanism for EDC chemistry. Formation of a stable amide bond occurs between a carboxylate molecule and a molecule with a terminal primary amine group Illustration of the resistive-pulse sensing method (drawing of pore not to scale) hemolysin protein nanopore embedded in a lipid bilayer support [140]

10 2-1 Current vs time plots for single-track PET membranes exposed to UV for the stated times Schematic of a double layer in an ionoc solution in contact with a solid electrode Current-potential curves for bare PET nanopores at different ph s Reaction scheme for the attachment of ethanolamine to surface carboxylate groups XPS spectrum for a bare PET membrane XPS spectrum for an ethanolamine-modified PET membrane Current-voltage curves in 10 mm KCl (ph 7) before and after ethanolamine modification SEM images Current-voltage curves in 1M KCl (PBS buffer, ph 7) before and after ethanolamine modification Representative current-time transients for the single nanopore membrane with 100 bp DNA in solution and expanded view of typical current-pulse signature associated with DNA translocation Current-time transient for ethanolamine modified conical nanopore sensor.at 1000mV DNA current-pulse frequency versus DNA concentration DNA current-pulse frequency versus transmembrane potential Current-pulse duration ( ) versus potential DNA current-pulse frequency versus tip diameter Plot of current-pulse ratio ( i/i) versus tip diameter (D tip ) Plot of current-pulse duration ( ) versus tip diameter. Transmembrane potential = 1000 mv Current-time transient for 100 bp DNA in 0.5 M KCl Scatter plot of current-pulse-amplitude ( i/i) vs duration ( ) for 100 bp DNA in 0.5 M (black) and 1 M KCl (red) in TE buffer

11 3-13 Scatter plot of current-pulse-amplitude ( i/i) vs duration ( ) for 100 bp DNA at 1 V(black) and 1.5 V (red) SEM images of conical PET nanopore (top view) Scatter plot of current-pulse-amplitude vs duration for 50 bp DNA Histograms of 50 bp DNA current-pulse-amplitude data for three different tip diameters Current-time transients from an ethanolamine-modified conical nanotube sensor Histograms of DNA current-pulse-amplitude data for 50 (red) and 100 (green) bp DNAs Histograms of DNA current-pulse-duration data for 50 (red) and 100 (green) bp DNAs Scatter plot of current-pulse-amplitude vs duration for 50 bp (red) and100 bp (green) Histograms of DNA current-pulse-amplitude data for 25 bp DNA (red) and 100 bp DNA (green) Histograms of DNA current-pulse-duration data for 25 bp DNA (red) and 100 bp DNAs (green) Scatter plot of current-pulse-amplitude vs duration for 25 bp (black) and100 bp (red) Histograms of DNA current-pulse-amplitude data for ss-dna (red) and ds- DNA (green) Histograms of DNA current-pulse-duration data for ss-dna (red) and ds-dna (green) Scatter plot of current-pulse amplitude versus duration for ss-dna (black) and ds-dna (red) DNA current-pulse frequency versus transmembrane potential Schematic diagram of the method used to prepare gold replicas Representetive SEM image of multipore PET membrane (base side) Diagram of a conical nanopore with notations (drawing of the pore not to scale)

12 5-4 Representetive SEM image of gold nanopore replicas etched with 0.1 M, 1M, 10 M HCl Representetive SEM image of gold nanopore replicas etched with 0.1 M, 1M, 10 M HCOOH Direction of electro-migration and diffusion during breakthrough under stopping electric field conditions (drawing of the pore not to scale) Representetive SEM image of gold nanopore replicas etched at 10, 20, 30 volts

13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PREPARATION AND CHARACTERIZATION OF SOLID STATE NANOPORES FOR DNA SENSING APPLICATIONS Chair: Charles R. Martin Major: Chemistry By Kaan Kececi August 2010 The goal of this research is to develop a DNA sensing device from artificial conical nanopores. In the first part of this work, etching conditions, surface characteristics and modification of the PET (poly (ethylene terephthalate)) membranes were studied. Single conical nanopore membranes were used as resistive-pulse sensing devices. Due to the challenges in the sensing paradigm, the surfaces of the PET nanopores were modified to favor the translocation of analyte through the nanopore. It is demonstrated that surface charge can be modulated using 1-ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride (EDC) and N-hydroxysulfosuccinimide (Sulfo-NHS) coupling chemistry. The second part of the research focused on the application of single nanopore membranes to detect 100-base-pair (bp) DNA. A key challenge for this sensing paradigm is finding the optimal conditions for the translocation of the analyte (i.e., DNA). After the surface modification of the nanopore, the sensing paradigm worked as expected and current-pulses were observed. The effects of tip diameter, analyte concentration and applied potential on current-pulse frequency, as well as other experimental conditions (e.g., electrolyte concentration, high transmembrane potential), were investigated. 13

14 In the third part, It was demonstrated that differentiation of analytes can be accomplished. Due to the differences in size and volume of the analytes, the currentpulse signatures could be used to differentiate different analytes. The effect of DNA size on translocation through a narrow nanopore tip was examined. The size of DNA was found to have an effect on current-pulse amplitude. In the fourth part of this work, template synthesis was used to fabricate nanocones in different geometries. Basically, different etching conditions and their effect on nanocone shape and cone angle were investigated. Applied potential and type and concentration of the stopping solution were used as etching parameters. 14

15 CHAPTER 1 INTRODUCTION AND BACKGROUND Introduction In recent years, there has been an increasing interest in nanofabrication. The importance of nanoscience was first pointed out 40 years ago in Richard Feynman s lecture, There is plenty of room at the bottom [1]. Although there is no universal consensus on the definition of what constitutes a nanomaterials, it is considered that at least one dimension must be less than or equal to 100 nm. These materials often have unique properties, which have potential applications in a wide variety of sectors, including biotechnology, energy, electronics and environmental monitoring, to name a few [2-12]. Among nanostructured materials, nanopores have gained great attention. This attention stems from the critical roles that biological nanopores play in almost all life processes [13, 14]. These nanopores are the principle nanodevices mediating the communication of a cell with other cells via exchange of ions and neutral molecules [13, 14]. There is a variety of different types of biological nanopores, but, they all consist of a membrane-bound protein pore that spans the thickness of the cellular membrane [14]. Fabrication of solid state nanopores has gained, growing interest because these systems can combine the advantages of a robust synthetic system with the specificity and precision of biological channels. Artificial nanopores have been fabricated in various materials using a variety of techniques [15-59]. In the Martin group, we have been exploring conically-shaped artificial nanopores in polymer membranes prepared by the track-etch method [15, 51-57, 59-61]. The conical shape is particularly advantageous for certain applications (vide infra) [58, 62-71]. Artificial conical 15

16 nanopores have been used to study transport properties [62, 63, 65-76], for bioseperations [77, 78], as templates for the deposition of other materials [79, 80], and as sensing devices [81-87]. The research presented here explores sensing-based applications of these conical nanopores. In particular, this research has focused on the development of DNA sensing devices. This chapter is divided into seven additional sections providing the background information for the research. The fabrication of conical nanopores and its characteristics will be discussed. Sensing paradigms using biological nanopore sensors will also be reviewed. Fabrication of Conical Nanopores The Track-Etch Method The track-etch method entails the chemical etch of latent ion tracks formed during the irradiation of a membrane sample with a high energy, heavy-ion beam. This method allows the fabrication of nanopores with various dimensions and geometries. The track-etch method has been practiced commercially for decades to prepare multipore membranes that are used, for example, for filtration applications [88-91] or as templates for the deposition of other materials [79, 92, 93]. This method has also been used to create single-pore membranes that have been implemented as sensing devices [81-87]. Formation of Latent Ion Tracks Membranes prepared by the track-etch method are created by first bombarding the membranes with a beam of high-energy particles (>1 MeV/nucleon) from a nuclear reactor or cyclotron (Figure 1-1A). Every ion that potentially penetrates the membrane forms a linear damage track (Figure 1-1B). The number of latent damage tracks formed 16

17 is determined by the exposure time to the particle beam. Multi-pore membranes with pore densities ranging from 10 5 to 10 9 pores per cm 2 are commercially available [94]. Gesellschaft für Schwerionenforschung (GSI, Darmstadt, Germany) has developed a technique to create a single track through the membrane by defocusing the ion beam and irradiating the polymer membrane with a single ion [53, 54]. Single-ion irradiation is achieved by placing a shutter between the ion beam and the membrane and an ion detector behind the membrane. The shutter is closed when the ion detector senses the penetration of single ion and prevents the further exposure of the membrane to ion the beam. The efficiency of this irradiation process depends on the type and energy of the irradiating ions, the radiation sensitivity of the material and the storage conditions of the membrane after irradiation [58]. Ion Track Etching After irradiation, the latent damage tracks can be chemical etched to create pores (Figure 1-1C). The commercial nanoporous membranes are simply immersed into the etching solution, and the damage tracks are etched from both faces of the membrane (isotropic etching). This process yields cylindrical pores through the membrane. The pore size can be controlled by the concentration of etchant, etching time and etchant temperature. There is an increasing interest in using conically shaped nanopores because of their interesting characteristics that make them well suited for certain applications, such as sensing. The etching process for preparing such conically-shaped nanopores was first developed by Apel et al. [60] In order to fabricate a conical nanopore, the iontracked membrane is first placed between two halves of an electrochemical cell (Figure 17

18 1-2) [60]. An etching solution is placed on one side of the cell and a stopping solution is added to the other side. When the etchant breaks through to the other side of the membrane, the stopping solution neutralizes the etchant and the etching is stopped by placing the nanopore membrane in water or the stopping solution. The fabricated nanopore is conical in shape with the large-diameter (base) opening and the smalldiameter (tip) opening (Figure 1-3). During the chemical etching, Pt electrodes are placed on either side of the membrane and a potential is applied. The breakthrough moment is monitored by an increase in the ionic current (Figure 1-4). The electrodes are arranged so that the anode is in the half-cell containing the etch solution. After breakthrough, an electro-stopping process occurs [60]. To understand this process, etching of poly(ethylene terephthalate) (PET) is used an example with 9 M NaOH as the etching solution and 1 M formic acid in 1M KCl as the stopping solution. Placing the anode in the etch solution causes the OH - etchant to be electrophoretically driven away from the nanopore tip opening. This process, called electro-stopping. results in conical nanopores with very small tip diameters (<5 nm) [60]. Synthetic nanopore fabrication reproducibility has been a challenging process [17, 26, 30-32, 95, 96]. The anisotropic chemical etching of conical nanopores gives good reproducibility and control in the base diameter of conical track-etched nanopores [59, 61, 73]. Reproducibility of the base diameter is achieved by stopping the first etch step after a predetermined amount of time. However, reproducibility of the tip diameter is lower than reproducibility of the base diameter. The mixing of etching and stopping solutions in the nanopore tip makes it difficult to control the etch rate in this region [57]. 18

19 The Martin group developed a second step etching technique to fine-tune the tip opening [57]. The second step etching involves a setup analogous to that for anisotropic etching, except that a more dilute etchant is placed on both sides of the membrane. This isotropic chemical etching process helps us to etch uniformly and more slowly with more controllable rate. Instead of stopping the second etch at a predetermined time, the process is stopped at a predetermined current value. If the second etch is stopped after a certain amount of time, the variability in tip diameter resulting from the first etch step, is retained. However, it has been demonstrated that the current flowing through the nanopore can be correlated to the tip opening diameter [57]. The quality of the track-etch process is distinguished by the track-etch-ratio [52, 55]. The track-etch-ratio is defined as the ratio of the track etch rate, v, to the bulk etch T rate, v. The parameter v is influenced by the concentration of the etchant, and the B B etchant temperature [55]. The parameters v and v determine the shape of track-etch B T pores as well (Figure 1-5). For high track-etch ratio, the cone half-angle, α, is defined as the ratio v /v. For materials with a high track-etch velocity, the etch cone becomes B T cylindrical. Materials A variety of membrane materials are well-suited for the track-etch technique. However, polymer membranes have seen the greatest use due to their chemical and mechanical robustness and high susceptibility to selective ion track etching.[58] A number of polymer materials are suitable for preparing ion track-etched conical nanopores, including poly(carbonate) (PC), poly(ethylene terephthalate) (PET), 19

20 poly(propylene) (PP), poly(vinylidenefluoride) (PVDF), and poly(imide) (PI). The chemical structures of three commonly used polymers are shown in Figure 1-6, along with scanning electron micrograph (SEM) images of pores etched in these materials in Figure 1-7. The ideal etching parameters, such as etchant composition and etching temperature, differ for each material. For example, ion tracks in PET membranes are typically etched with a 9 M NaOH etchant and a 1 M formic acid/1 M potassium chloride (KCl) stopping solution at room temperature [15, 60]. Upon nanopore breakthrough, hydroxide etchant is simply neutralized by the formic acid. The NaOH hydrolyzes the ester bonds in PET resulting in the formation of carboxylate and hydroxyl groups inside the pore [76]. In contrast, ion tracks in Kapton membranes are etched with a NaOCl etchant with an active chlorine content of 13% and a 1 M potassium iodide (KI) stopping solution, at a temperature of 50 o C [15, 55, 63]. Upon etchant breakthrough at the nanopore tip opening, an oxidation-reduction reaction occurs, whereby iodide ions catalyze the reduction of hypochlorite ions to produce chloride ions. This stop-etch reaction yields iodine, yellow in color, which provides a colorimetric indicator of breakthrough. Etching of Kapton results in the formation of carboxylate groups inside the pore via hydrolysis of imide bonds [63, 97]. Etching rates also differ from one material to another. The two etching properties that have the greatest influence on the shape of conical nanopores are the bulk etch rate, v, and the track etch rate, v, of the material. The bulk etch rate for PET has been B T determined to be ~2.17 nm/min.[15, 60] The track etch rate is ~10 μm/hour for 12 μm 20

21 thick PET [60]. Kapton has a bulk etch rate of 0.42±0.04 μm/hour and a track etch rate of 3.12 ± 0.65 μm/hour for 12 μm thick Kapton [63]. As described above, for a conically shaped pore, the ratio of v /v determines the cone angle of the pores. Kapton has a B T much higher v /v ratio than PET, resulting in conical pores with much larger base B T diameters and cone angles [15, 55, 60, 63]. Conical Nanopore Characterization and Properties Electron Microscopy The three-dimensional shape of conical nanopores is generally characterized from field-emission scanning electron microscopy (FE-SEM) images of gold nanopore replicas (Figure 1-8). The gold replicas are obtained by electrolessly depositing gold inside the empty nanopores. The electroless deposition method (vide infra) also leaves a layer of gold along both faces of the nanopore membrane surface. The nanopore replicas can be liberated by removing one or both layers of gold from the membrane surface and then dissolving away the membrane material. The base diameters, d, of conical nanopores obtained after the first etch step b have also been characterized and measured with FE-SEM. Electron micrographs of the base openings of conical pores (Figure 1-7) in track-etched multi-pore membranes can be used to determine the bulk etch rate, v, of the material being etched. The bulk etch B rate of the material determines the diameter of the base opening in conical nanopores. If the bulk etch rate is known, then the base diameter of single conical nanopores can be calculated by multiplying v by the total etching time of the first step etching. For B example, the bulk etch rate of PET was determined to be 2.17 nm/min from FE-SEM 21

22 images [57, 60]. If the anisotropic etching process is stopped after 2 hours, then the base diameter of a conical nanopore in PET will be ~520 nm. Electrochemical Measurements Current-voltage curves are typically used to determine the tip opening diameter, d, of single conical track-etched nanopores. This electrochemical method for tip size t determination entails mounting the membrane containing the conical nanopore in the same cell setup used for the etching process (Figure 1-2). An electrolyte solution of known ionic conductivity is introduced into both sides of the cell along with electrodes. A current-voltage curve for the electrolyte-filled nanopore is then obtained via a linear scan of the transmembrane potential, while measuring the resulting ion current flowing through the nanopore (Figure 1-10). The slope of this current-voltage curve is equal to the ionic conductance, G (in Siemens, S), of the electrolyte-filled nanopore. Since conical nanopores with small tip diameters rectify ion current (vide infra), the linear portion of the current-voltage curve (between -200 and +200 mv) is used to calculate the nanopore tip diameter [82]. The equation for the ionic conductance of a conical pore is, [57, 60, 81, 82] G = σπd td b 4L (1-1) Where σ is the specific conductivity of the electrolyte solution (S cm -1 ), L is the length of the nanopore (thickness of the membrane), d is the experimentally determined base b opening diameter, and d is the diameter of the tip opening. Since all other parameters t except for d in Equation 1-1 are known, d can be calculated. Post-anisotropic etching t t tip opening diameters typically range from 1 7 nm [57]. It is important to note that this 22

23 equation can only be rigorously applied to conical nanopores after the first etching step [57]. In order to determine the tip diameter after the second, isotropic, etch step one must take into account the change in both base and tip diameters that occur during the second etch step. A detailed mathematical model has been developed that takes into account this change and allows one to calculate the single conical nanopore tip diameter from current voltage curves after the second etch [57]. However, this model predicts that for final tip diameters under ~50 nm the change in the base diameter is negligible compared to the change in the tip diameter [57]. This allows Equation 1-1 to be used to calculate tip diameters after the second step etch for conical nanopores with tip diameters less than 50 nm. Electric Field Focusing An important feature of the conical nanopore sensor is that the voltage drop caused by the ion current flowing through the nanopore is focused at the nanopore tip [64, 85]. Indeed, calculations done by Lee et al. indicate that even when a modest transmembrane potential is applied across an electrolyte-filled nanopore, the electric field strength in the nanopore tip is enormous [64]. For example, with an applied transmembrane potential of 1 V and a conical pore with tip diameter = 60 nm, base diameter = 2.5 mm and length = 6 mm, the magnitude of the electric field at the nanopore tip was modeled to be on the order of 1.5 MV per m (Figure 1-10) [64]. Furthermore, finite element simulations of a nanopore at an applied transmembrane potential of 1 V, with a base opening diameter of 520 nm, a length of 12 mm and tip opening diameters ranging from nm, have been done [85]. These studies have shown that the electric field strength inside the nanopore tip increases with decreasing 23

24 tip opening diameter. A consequence of this focusing effect is that the ion current is extremely sensitive to analyte species present in or near the nanopore tip. That is, there is an analyte sensing zone just inside the tip [64, 81, 82]. This focusing effect makes conically shaped nanopores better suited for sensing applications than cylindrical nanopores. Because of this electric-field focusing effect, it is not the total length of the pore (membrane thickness) which is relevant to the sensing applications but rather the effective length. This effective length is the distance from the tip opening at which most of the electric field has decreased; put another way, the length of the analytesensing zone. If we use the criterion that the effective length is that length over which 80% of the voltage is dropped, then conical nanopores with bases of 5 μm and tips of 20 nm have an effective length of 50 nm, as determined by the finite-element method [64]. Furthermore, such simulations show that the effective length can be controlled by varying the cone angle of the conical nanopore [64]. This is important because it allows for the length of the sensing zone to be tailored to the size of the analyte species to be detected. Ion Current Rectification Another important characteristic exhibited by conical nanopores is the rectification of ion current. Rectification is a term that comes from electronics, referring to devices that conduct electrons only in one direction [58]. Ion current rectifiers show nonlinear current-voltage curves, and have a preferential direction of ion flow through their channels [13, 14]. Single conical nanopores in polymer membranes show both ion selectivity, with a preferential direction of cation flow from the tip opening of the pore to the base, and exhibit nonlinear current-voltage curves [66-69]. 24

25 Several models have been developed to explain ion current rectification in artificial nanopore systems. In the model developed by Siwy et al [68], there are three requirements for ion current rectification: (i) an asymmetric pore shape, (ii) a tip diameter with dimension comparable to the Debye layer at the nanopore wall, and (iii) a surface charge on the pore wall. Artificial conical nanopores in polymer membranes possess all of these characteristics, and thus allow for ion current rectification to occur. The chemical etching of the ion tracks in PET (Figure 1-11) and Kapton membranes leaves carboxylate groups on the nanopore surface [15]. These carboxylate groups are deprotonated at phs above the polymer s isoelectric point (~3 for both polymers) [15], which results in a negative surface charge. Therefore, current-voltage curves taken with single conical nanopores show ion current rectification at phs above the polymer s isoelectric points (i.e., when negative surface charge is present) [15, 63, 66-68]. It is important to note that ion current rectification is only observed in conical nanopores when the tip diameter opening is small. This is because in order for rectification to occur, the thickness of the Debye layer at the nanopore wall must be of comparable dimensions to that of the pore tip. The Debye layer forms in solution close to the pore surface in order to compensate for the negative surface charge. As long as the thickness of this layer is comparable to the tip radius, the negatively charged conical nanopore will preferentially transport cations and reject anions [68]. The asymmetric shape of conical nanopores has also been proven necessary in order to observe rectification. Studies on the transport properties of cylindrical pores [98] have revealed that track-etched cylindrical nanopores of the same limiting diameter do not rectify ion current. 25

26 The model proposed by Siwy et al. is then based on an electrostatic ratchet in which the asymmetry in negatively charged nanopores creates an electrostatic trap for cations at positive applied potentials [66]. The electrostatic trap effectively forms the off state for the artificial nanopores. At negative applied potentials the electrostatic trap is eliminated and the on state is observed. For nanopores with a positively charged surface the ratchet model will be reversed. Another model, introduced by White et. al. [99], also requires an asymmetric pore shape and a pore wall surface charge. In this model, ions of charge opposite to that of the pore walls are able to traverse the pore with much greater ease than those of charge similar to that of the pore walls. For example, when a transmembrane potential is applied so that ions of charge opposite to that of the pore walls are driven through the conical pore from tip to base, ions of charge similar to that of the pore walls will be driven towards the tip from the base side (Figure 1-12A). However, since these ions have the same charge sign as the pore surface, they will be less likely to exit the pore through the tip (depending on the size of the tip and the surface charge density) due to electrostatic interactions. This is expected to result in a local buildup of ions within the nanopore [99]. Therefore, in this case, the nanopore is highly conductive, and large ionic currents result (Figure 1-12B). When a transmembrane potential is applied so that ions of charge opposite to that of the pore walls are driven through the conical pore from base to tip, ions of charge similar to that of the pore walls will be driven towards the base from the tip side (Figure 1-12C). However, again electrostatic interactions will prevent these ions of like charge from entering the tip of the pore. Thus, a local depletion of ions within the nanopore will 26

27 result [99]. This leads to a less conductive state and, consequently, smaller ionic currents (Figure 1-12D). Tailoring Surface Chemistry of Conical Nanopores A key function of biological nanopores is their ability to selectively detect and transport specific analytes. It is likewise important to be able to control the surface chemistry of artificial nanopores in order to create both biocompatible and analyte selective devices. There are several strategies currently used to control artificial nanopore surface chemistry. One method employs the electroless deposition of gold onto the nanopore walls followed by modification with simple thiol-based chemistry. Another method takes advantage of surface carboxylate groups produced during chemical etching to attach amine terminated molecules using 1-ethyl-3-[3- dimethylaminopropyl] carbodiimide hydrochloride/n-hydroxysulfosuccinimide (EDC/sulfo-NHS) chemistry. Electroless Gold Deposition The electroless gold deposition method is a form of template synthesis. The template synthesis method [2, 79, 92, 93, ] for the preparation of nanomaterials was pioneered by the Martin group, and involves depositing a desired material into the nanopores of a solid host. This method is useful for preparing hollow nanotubes and solid nanowires. In the electroless deposition of metals [119], a reducing agent is used to plate metals from solution onto a solid surface. A catalyst is needed to accelerate the rate of the plating reaction on the surface. The procedure that is used to deposit Au along the walls of conical polymer nanopores first involves sensitizing the membrane with Sn(II). This is done by immersing the nanopore membranes in methanol for five minutes and then into a 50/50 27

28 water/methanol solution that is 0.026M in SnCl and 0.07M in trifluoroacetic acid for 45 2 min. The Sn(II) sensitizer will bind to the pore walls and membrane surface via electrostatic complexation with the negatively-charged surface functional groups of the polymer formed during chemical etching [119]. The membrane is then washed again for 5 min. in methanol and placed in an aqueous ammoniacal solution that is 0.029M in AgNO for 7.5 min. During this step a surface redox reaction occurs and the surface 3 bound Sn(II) is oxidized to Sn(IV) and Ag(I) is reduced to elemental Ag. This deposits silver nanoparticles on the pore walls [119]. The membrane is again washed in methanol for 5 min. before being placed in a gold-plating bath that is 7.9x10-3 M in Na Au(SO ), M Na SO, 0.025M in NaHCO and M in formaldehyde at o C. The ph of this solution is first adjusted to 10 by dropwise addition of 1 M H SO. 2 4 This ph, as well as the lower temperature, are necessary for uniform plating. While in this gold-plating bath another surface redox reaction occurs. Since the standard reduction potential of gold is more positive than that of silver, gold galvanically displaces the silver to yield gold nanoparticles on the pore surface. These nanoscopic gold particles catalyze the subsequent reduction of Au(I) to Au(0) using formaldehyde as the reducing agent [119]. 2 Au + + HCHO + 3OH - HCOO - + 2H2 O + 2 Au (s) (1-2) This process occurs spontaneously and produces elemental gold via redox chemistry without using electrodes, hence the name electroless. Electroless deposition yields conical gold nanotubes lining the pore walls as well as gold surface layers that cover both faces of the polymer membrane. These surface layers are 28

29 typically too thin to block the openings of the conical gold nanotube at the membrane surfaces. The surface layers can be removed via a tape-peel method or by swabbing the surfaces with an ethanol-wetted swab. A schematic of the plating process is shown in Figure The thickness of the walls of the gold nanotube can be varied by varying the gold plating time, and this provides another means for controlling the diameter of the tip opening of the nanotubes. Gold nanotubes with tip diameters of molecular dimensions (1 nm) can be obtained. Plating for longer periods of time will result in the nanopores being completely filled with gold to yield solid gold nanocones. As previously mentioned the resulting nanocones can be liberated from the membrane and imaged via SEM. EDC/Sulfo-NHS Chemistry EDC/sulfo-NHS chemistry is used to couple carboxylate groups to primary amines (Figure 1-14). The EDC/Sulfo-NHS procedure is commonly used for protein conjugation and immobilization of proteins to a surface [ ]. Many of the polymers used for preparing track-etched nanopores have carboxylate groups present after etching. Therefore, the EDC/sulfo-NHS method can be used to attach proteins and other molecules having amine terminal groups to the nanopore walls [83, ]. EDC can be used alone to couple carboxylate and primary amine groups. EDC reacts with a carboxylate group to form an amine-reactive O-acylisourea intermediate. If the intermediate encounters a primary amine group, a stable amide bond will join the two molecules. However, the O-acylisourea intermediate is unstable and short lived the reaction of the intermediate with an amine does not occur quickly. Therefore EDC alone is not very efficient in coupling carboxylate and amine groups. In the presence of water the O-acylisourea intermediate normally hydrolyzes and the carboxyl group is 29

30 regenerated. For this reason sulfo-nhs is often used to increase the efficiency of EDCmediated coupling. The addition of sulfo-nhs stabilizes the amine-reactive intermediate by converting it to an amine-reactive sulfo-nhs ester. The sulfo-nhs ester intermediate has sufficient stability to allow for the primary amine groups to be coupled with the carboxylate groups again via a stable amide bond. Biological Nanopores The growing interest surrounding nanopores arises in part because of the critical roles biological nanopores play in many physiological processes of living organisms [13, 14]. Biological nanopores and nanochannels are present in the cellular membranes of all living cells. These channels are formed by membrane proteins that span the entire thickness of the cell membrane (~4 nm). They are the primary devices used by cells to communicate with other cells [13, 14]. Intercellular communication occurs through transport of ions and neutral molecules through the channels [13, 14]. The transport of ions by way of the protein channels is often a highly selective and controlled process. Controlled transport can be achieved because biological ion channels are often selective for certain ions and open and close through gated mechanisms. Gating of ion channels occurs when they open and close in response to certain stimuli, such as deformations in the cell membrane, the presence of a ligand or some other signaling molecule, or changes in membrane potential [13, 14]. For example, with voltage-gated ion channels, the opened and closed states are dependent on membrane potential [126]. When ion channels are in the opened or on state, ions are allowed to pass through the channels, and when in the closed or off state, ion transport is blocked. Many of the transport properties of these biological ion channels are not well understood. Investigating these properties can often be a challenge with biological 30

31 systems due to their fragile nature. Constructing artificial nanopore devices with transport properties similar to those of biological ion channels could give new insight into the physical and chemical principles of biological ion channel operation.[58] Already, very recent advances in single artificial nanopore design are shedding light on the mechanisms by which naturally occurring ion channels function [62-64, 66-71]. Another motivation for studying transport in nanopore membranes comes from the possible implementation of these membranes into single-molecule chemical and biochemical sensing devices. In previous work, it has been shown that single, biological transmembrane protein nanopores embedded in lipid bilayer membranes can function as a single-molecule sensing device using the resistive-pulse sensing method (vide infra) [ ]. Biological nanopores have proven to function as extremely versatile and selective resistive-pulse sensors. Recently, it has been shown that artificial single nanopore systems can also be used as platforms for single-molecule resistive-pulse sensing devices [16-22, 24-50, 81-85, 95, 96, 155, 156]. Currently, there is a large research effort focused towards fabricating single nanopores in synthetic materials [15, 51-57, 59-61]. Resistive Pulse Sensing The resistive pulse sensing method [95, 96, 99, 155, 156], which is sometimes referred to as stochastic sensing [ ], entails mounting a membrane containing a single nanopore between two halves of an electrochemical cell filled with an electrolyte solution. A transmembrane potential is applied, and the resulting ion current flowing through the electrolyte-filled nanopore is recorded versus time (Figure 1-15A). As an analyte with dimensions comparable to the nanopore diameter is driven through the pore, a momentary block in the ion current is observed, yielding a downward current- 31

32 pulse (Figure 1-15B). The concentration of the analyte can be determined from the frequency of these current-pulse events and the identity of the analyte is encoded in the magnitude and duration of the current pulse [95, 96, ]. Current work in the field of resistive-pulse sensing is aimed at the detection and characterization of molecules, ions and biopolymers [72-85, 87, 99, 124, 158][15-17, 22-68, 71][ ] [99, , 159, 160]. Sensing of such molecular-sized analytes is possible if the diameter of the nanopore sensor element is of molecular dimensions. Nanopores in both biological and artificial membranes have been used to sense such analytes. In particular, a number of prototype resistive-pulse sensing devices has been developed from biological nanopores. The most commonly used protein nanopore is α-hemolysin (Figure 1-16) [ ]. α-hemolysin nanopores, either in their wild state or engineered form, have been used to detect DNA [129, 131, 144, ] nitroaromatic compounds [132], metal ions [133], small organic molecules [136], anions [135], proteins [137, 138], polymers [142] and enantiomers of drug molecules [130]. There are two main advantages that are offered from this biological nanopore. The first is that the pore size is reproducible from sample to sample and measurement to measurement. The second advantage offered by the α-hemolysin nanopore is the selectivity imparted through engineering of the pore. Bayley and coworkers have performed numerous modifications to the α-hemolysin pores through genetic engineering and chemical modification, which has allowed for highly selective sensors to be developed [ ]. The work that has been done with biological nanopores has been extremely influential in the development of molecular resistive-pulse sensors. This work currently 32

33 stands as the benchmark by which all other resistive-pulse sensing devices are evaluated. However, it seems unlikely that a practical sensing device will ever be developed from biological nanopores due to the fragility of the lipid bilayer membrane support [128, 156]. The planar lipid bilayers cannot endure a wide range of phs, temperatures, applied transmembrane potentials, or solvents, and they are sensitive to vibrations [128, 156]. As a result, there has been a significant amount of research focused towards developing artificial analogues of biological nanopores to be used as resistive-pulse sensors [15-17, 22-68, 71][ ]. Ideally, an artificial nanopore sensor would have the same sensing capabilities and offer the same advantages of the biological nanopores, but show chemical and mechanical stability over a wide range of conditions. A variety of techniques has been used to fabricate single, artificial nanopores in synthetic materials. The most widely used methods employ ion or electron beams to create single pores in silicon nitride and silicon oxide membranes [17, 22-36]. These nanopores have been used to study mainly DNA [17, 25, 30-32, 34], but some protein sensing [35, 36] work has also been reported. Single pores have also been prepared using soft lithographic techniques [37-42], and have been used to sense colloidal particles [38] and DNA.[41]. Nanopores created with this technique have also been used for direct detection of the binding of antigens to antibody-coated colloidal particles [39]. Single carbon nanotubes embedded in an epoxy membrane [43-47, 95] have been used as the sensing element for the detection of nanoparticles [44, 46, 47]. A femtosecond-pulsed laser-based technique has been developed to create single pores in glass, which have been used to examine immune complexes [49]. Base etching of 33

34 silicon wafers [16] and track-etching of nanopores in polymer membranes [15, 51-57, 59-61] are other methods that have been used to fabricate single nanopores. The Martin group, along with others [15, 51-57, 59-61], has utilized the tracketching method to prepare single, conically-shaped nanopores in polymer membranes. The Martin group works with conical nanopores because the geometry has proven to be especially advantageous for resistive-pulse sensing applications [81-85]. Compared with cylindrical pores of the same limiting diameter, the conical pore has a smaller resistance and can therefore generate higher ion currents for a given voltage [63]. The conical pores are also less susceptible to clogging and allow for single molecule transport [63, 84]. As previously discussed, the conical shape causes the electric field through the pore to be focused at the pore tip opening, which creates a highly sensitive detection zone within the tip [64, 85]. Conically shaped nanopores in polymer membranes also exhibit many of the transport properties seen in biological ion channels [15, 66, 68, 69] and can act as mimics of voltage-gated ion channels [15, 62, 70, 71, 161]. Other Nanopore-Based Sensing Strategies Other sensing strategies have also been used with single conical nanopores etched in polymer membranes. A single conical Au nanotube in a PET membrane was used to design a new type of protein biosensor [86]. The nanotubes were modified with various biochemical molecular recognition agents (MRAs) to detect analytes in solution with an on/off response. Like the resistive-pulse sensing method, this sensing protocol also involves passing an ion current through the single nanotube. However, currentpulse translocation events were not observed in this case. Instead, as the analyte interacted with the surface bound MRAs, the current flowing through the nanopore was 34

35 permanently shut off. Blockage of the ion current occurred because the diameter of the analyte was of comparable dimensions to that of the nanotube tip. This sensor has been shown to be a highly sensitive and selective type of biosensor, and it should be possible to modify the Au nanotube surface with a wide range of MRAs to selectively detect a wide variety of analytes. Conical nanopore sensors which make use of the ion current rectification phenomenon have also been developed [87]. These devices consisted of a single conical nanopore in a Kapton membrane, and were used to detect drug molecules. Prior to addition of a cationic drug molecule, the single nanopore exhibited ion current rectification. However, after exposure of the nanopore to the drug molecule the extent of ion current rectification began to change. The change in rectification was due to the cationic molecule adsorption of the onto the nanopore walls, resulting in a change in the surface charge of the pore. It was found that the magnitude of the change in rectification scaled with the concentration of the drug molecule. Dissertation Overview The goal of this research is the design of a single nanopore sensor. In particular, the research was focused on developing a DNA sensing device. Chapter 1 has reviewed necessary background information for this dissertation, including tracketching, tailoring the nanopore surface, ion current rectification, biological nanopores, and resistive-pulse sensing. Chapter 2, describes the investigation of the surface characteristics of PET membranes were investigated by using electrochemical methods. The negative surface charge of PET membranes was partially neutralized by EDC/NHS chemistry. The surface modification was confirmed by XPS and current-voltage curves. 35

36 Chapters 3 and 4 describe use of surface modified nanopores for resistive pulse sensing with 100 bp DNA as the analyte to study the translocation dynamics. In order to confirm the results as translocation, different parameters were investigated. DNA molecules with different lengths were used for sensing and resistive-pulse sensing allowed them to be discriminated. Chapter 5 describes the synthesis of gold nanocones using a template method. Different etching conditions were used and their effects on nanocone shape were investigated. 36

37 Figure 1-1. Diagram of the ion track-etching method. A) Irradiation of a membrane with swift, heavy ions results in B) the formation of latent damage tracks along the ions path. C) Chemical etching selectively removes the damage tracks resulting in the formation of pores. 37

38 Figure 1-2. Diagram of the electrochemical cell used for chemical etching and all electrochemical measurements Figure 1-3. Schematic of a conical nanopore in a polymer membrane showing the base diameter and tip diameter (drawing to scale). 38

39 Current (na) Breakthrough Time (min) Figure 1-4. Plot of current versus time recorded during the first step of conical nanopore formation. Figure 1-5. Track-etching of a conical pore, showing bulk etch rate, B, track etch rate, T, and cone half angle,. 39

40 Figure 1-6. Chemical structures of polymers typically used for track-etching. A) poly(carbonate) (PC), B) poly(ethylene terephthalate) (PET) and C) poly(imide) (PI) Figure 1-7. Scanning electron micrographs of nanopores track-etched in various polymer materials. A) PC [58], B) PET [162] and C) PI [73] 40

41 Current (na) Figure 1-8. Scanning electron micrograph of conical gold nanocones formed by deposition into a conical nanopore membrane. The nanocones have been liberated from the polymer membrane mold Applied potential (V) Figure 1-9. A typical current-voltage curve used to calculate the tip diameter of a conical nanopore. 41

42 Figure Distribution of the electric field across a conical nanopore. The nanopore used for simulations had a base diameter opening of 2.5 mm, a tip opening diameter of 60 nm and a thickness of 6 mm with 1V applied across the nanopore in 1M KCl. White hash marks are added to the section of the nanopore where most of the electric field is focused. [64] 42

43 Figure Chemical structure of PET and formation of chain ends after chemical etching 43

44 Figure Model of ion current rectification. Model proposed by White et al. [99] (drawing of pore not to scale) 44

45 Figure Schematic of Au electroless plating procedure [119]. 45

46 Figure Reaction mechanism for EDC chemistry. Formation of a stable amide bond occurs between a carboxylate molecule and a molecule with a terminal primary amine group via EDC/Sulfo-NHS chemistry. [adapted from Pierce Biotechnology, Figure Illustration of the resistive-pulse sensing method (drawing of pore not to scale). 46

47 Figure hemolysin protein nanopore embedded in a lipid bilayer support [140]. 47

48 CHAPTER 2 PREPARATION AND CHARACTERIZATION OF CONICAL PET NANOPORE Introduction Nanopores have been used to detect and characterize molecules and have become an important research component within modern biochemistry, biophysics and chemistry [163]. The demand for rapid and reliable individual molecule detection has motivated development of new bioanalytical tools. In this regard, remarkable progress has been made by using nanopore-based methods [164]. The simplicity of this approach also makes it popular and potential as a new sensing tool of individual molecules. Nanopore-based detection is called resistive pulse sensing in which the sensing paradigm entails mounting the synthetic or biological membrane containing the nanopore between two electrolyte solutions, applying a transmembrane potantial difference, and measuring the resulting ion current flowing through the electrolyte filled nanopore. In simplest terms, when the analyte enters and translocates through the nanopore, it transiently blocks the ion current, resulting in a downward current pulse. This current pulse is used to obtain information about the analyte by analyzing characteristics such as shape, magnitude and duration. Most of the resistive-pulse sensing data has been obtained using a biological nanopore, -hemolysin ( -HL), which is embedded in a supported lipid bi layer membrane[136, 165]. Due to the fragile nature of the lipid bilayer, synthetic nanopore fabrication techniques were developed [166]. The track-etch method has been practiced commercially for decades which is one of these techniques for synthetic nanopore fabrication and [51]. It is based on 48

49 bombardment of polymeric materials with heavy ions to induce the formation of tracks. Then these tracks are expanded into pores by chemical etching [51]. This process can be applied to produce single-track [18, 20] and multi-track [55] materials. In order to develop a sensor system, the fabrication parameters need to be characterized and optimized. In this chapter, basic characteristic of the track-etch method for poyl(ethylene terephthalate) (PET) are determined and optimized for resistive-pulse studies. Experimental Materials Poly(ethylene terephthalate) (PET) membranes, 12 um thick, which contained single and multi heavy-ion induced damage tracks were obtained from GSI (Darmstadt, Germany). Purified water (obtained by passing house-distilled water through a Barnstead, E-pure purification system) was used to prepare all solutions. UV Irradiation Membranes were irradiated with UV light (Entela, UVGL 58) with wavelength of about 320 nm for various amount of time on both sides prior to etching. The effect of UV irradiation on etching rate was monitored by asymmetric chemical etching [15, 60]. Membranes irradiated for various times were sandwiched between the cells and etched under 1V potential. EDC Modification The pore-etching procedure generates carboxylate groups on the pore walls and membrane faces [76]. EDC chemistry was used to attach ethanolamine to these -COO - groups via amide bond formation [186]. This was accomplished by first immersing the membrane into a solution that was 1 mm in EDC and 1 mm in sulfo-nhs dissolved in 10 49

50 mm MES-buffered saline. The membrane was immersed into this solution for ~2 hr and then immersed over night in a solution that was 1mg/ml in ethanolamine. HCl dissolved in purified water (Figure 2-4). Pore Etching and Nanopore Fabrication Conically shaped nanopores were etched into the single-track PET membrane by anisotropic chemical etching. As per the procedure developed by Apel et. al. [60], The irradiated membrane was placed between two half-cells and etch solution (9M NaOH) was added to one half-cell and a stop solution (1M KCl in 1M formic acid) to the other half-cell. The damaged track is preferentially etched from the face in contact with the etch solution, so that the base opening is etched into this face of the membrane. To detect breakthrough, each half-cell contained a Pt wire, and a transmembrane potential difference of 1 V was applied during etching using a Keithley 6487 picoammeter/voltage source (Keithley Instruments, Cleveland, OH). The electrodes were configured such that the anode was on the side of the cell containing the etch solution and the cathode was on the side containing the stop solution. The current was initially zero, and breakthrough was signaled by a sudden increase in current, typically after period of 1.5h to 2h. After etching, the membrane was removed from the cell and stored in water. Electrochemical Measurements and Data Analysis To obtain current-voltage curve, the membrane sample was mounted between the two halves of the conductivity cell [57], and each half-cell was filled with 3.5 ml of electrolyte solution. A Ag/AgCl electrode was inserted into each half-cell solution, and a Keithley 6487 picoammeter/voltage source was used to step the potential between -1V and 1V. 50

51 XPS Analysis Ethanolamine attachment (Figure 2-4) was confirmed using X-ray photoelectron spectroscopy (XPS, Perkin-Elmer PHI 5100). Multi-track (10 8 tracks per cm 2 ) PET membranes were used for XPS studies. The X-ray photoelectron spectroscopy system (XPS, Perkin Elmer PHI 5100) was equipped with a magnesium anode (15 kv, 20 ma). For each sample, detailed scans of the O1s, C1s, N1s and Si2p lines were performed with a 10 mm x 4 mm spot size and pass energy of ev. This was followed by a wide scan performed under the same conditions with pass energy of ev. During the analyses, the vacuum in the analysis chamber was 3x10-9 torr. Results and Discussion Effect of UV Irradiation on Breakthrough Treatment of the irradiated polymer membranes with UV light [167] can enhance the etchability of the latent tracks [168, 169]. Exposing the tracked membrane to UV prior to chemical etching can be seen as an accelerated aging of the material. This results overall in a higher track etch rate and a narrow size distribution of the channels. The etch rate is also less sensitive to the time elapsed since the actual ion irradiation and the storage conditions. It is obvious that UV has strong effects on the etching rate. This has been proven by irradiating PET membranes by UV light for various times (i.e., 2, 6, 10 hours). A non-irradiated membrane was used as control and the other membranes were etched for 2h. As seen in Figure 2-1, the non irradiated membrane did not breakthrough after 2h (black line), the membrane irradiated for 2h broke through after 45 min and other irradiated membranes, 6 and 10h, broke through after 33 and 42 min, respectively. 51

52 The track-etching rates of PET membranes (12 um in thickness) in 9 M NaOH solution can be calculated by dividing thickness of membrane by the break through times. The track-etching rates were calculated to be 267 nm/min, 363 nm/min, 286 nm/min for 2, 6, 10 hours of UV irradiated membranes, respectively. It is obvious that the accelerating effect of UV irradiation is not constant along the ion track. The difference in etching rates can be explained by the chemical structure of PET. Several studies reported the structure of PET in detail and pointed out that it has crystalline and amorphous domains [80, 205, 206]. Lueck et al. [205] reported the difference in etching rates of amorphous and crystalline domains, which can explain the difference in etching rates. Since the distribution of these domains cannot be predicted or determined, UV irradiation can yield different track-etching rates. Effect of Electrolyte Concentration on Tip Size The diameter of the tip was determined using an electrochemical method [57, 60]. Briefly, the membrane was mounted between the two halves of the conductivity cell as described [57] and the electrolyte solutions (0.1 and 1 M KCl) were placed on either side of the membrane. A Ag/AgCl electrode immersed into each half solution was used to obtain a current-voltage curve for the electrolyte-filled nanopore. The slope of this curve is the ionic conductance of the electrolyte-filled pore. The conductance was used to calculate the diameter of the tip opening [57, 60] as indicated in Chapter 1. Due to the surface charge of the substrate, ions of the same charge polarity as the substrate material (referred as co-ions) are depleted from the solution near the substrate. The region near the surface in which co-ions are excluded and counter-ions are accumulated is referred to as the electrical double layer [170, 171]. The thickness of 52

53 the electric double layer is characterized by the Debye screening length. In aqueous solutions of a monovalent salt, equation 2-1 describes the Debye length, -1 (m): κ 1 = ε rε 0 RT 2C 0 F 2 (2-1) where R (J.K -1.mol -1 ) is gas constant, F (C.mol -1 ) is the Faraday constant, C 0 (mol.m -3 ) is the bulk concentration of salt, ε 0 (C 2.N -1.m -2 ) is the permitivity constant, ε r (unitless), is the dielectric constant, T (K) is the absolute temperature. For a monovalent salt like KCl the Debye length (in nm) is given ( -1 ), κ 1 = 0.303/ C (2-2) In aqueous solutions that contain a monovalent salt, the Debye-length ranges from less than 1 nm at a concentration of 1.0 M to nearly 10 nm at concentration of 1mM. In nanopores with pore diameters on the order of the Debye-length, the electrical double layer can occupy a significant volume of the nanopore and may even overlap within the pore [172]. The effect of double-layer thickness on tip diameter is given in Table 1. Tip diameter was calculated based on the ionic conductance of the nanopore in 0.1 M and 1 M KCl at different ph values. The electrochemical measurements provided an average tip diameter of nm in 0.1 M KCl and nm in 1 M KCl. The accuracy of the electrochemical measurement of tip diameter was also discussed in our previous studies by the Martin group [58]. The tip diameters of tracketched membranes were calculated by both SEM images and electrochemically, and it was shown that the electrochemically determined value is within the standard deviation of the SEM value, indicating that the electrochemical measurement of tip diameter is accurate and precise. 53

54 Rectification Behavior of PET Nanopore at Different ph It is known that chemical etching of PET yields carboxylate groups due to cleavage of polymeric chains. These carboxylate groups create excess negative surface charge on the nanopore wall with estimated density of ~1e/nm 2 [60, 173]. The degree of protonation and deprotonation of carboxylate groups determines the rectification behavior, which is regulated by the ph of the electrolyte solution. Lowering the ph value results in protonation of the groups and a decrease in the surface charge and the opposite behavior can be observed as the ph is increased. The ph at which the net surface charge is zero also called the isoelectric point. The issue of surface charge is very important because when ions pass through a narrow pore, they interact with charges on the pore wall, influencing the pore transport properties. We examined the dependence of the ion current rectification at different ph values (Figure 2-3). As mentioned before, carboxyl groups formed after etching are protonated or deprotonated, depending on the ph of the solution. At ph lower than 3, the rectification plot in asymmetric PET is stepper due to the positive surface charge. All surface charges are neutralizes at ph 3, which yields a linear I-V curve. At this ph, the asymmetric pore acts as if it were cylindrical. On the other hand, over ph 3, the I-V curve rectifies strongly again due to the negative surface charge at ph s above the isoelectric point. These results show that surface charge on the pore walls plays a critical role in the rectification behavior. As a result, we found that especially for the neutral and basic ph values at which DNA is stable, the surface of the PET nanopore is negatively charged and shows non-linear I-V curves. In order to understand the nonlinear behavior, we need to determine, which ions contribute to the recorded current. Ion selectivity is usually expressed by the 54

55 transference numbers of cations (t + ) and anions (t - ) in the membrane. The sum of the transference number is 1, t + with t - indicating the percentage of the current signal carried by cations and anions, respectively. The t + and t - can be determined in a conductivity cell in which the membrane separates two electrolyte solutions of the same salt but different concentrations. The resulting potential difference for 1:1 salt is given by, E m = 2.303RT nf t + t log c 1 c where F is the Faraday constant equal to C/mol, and n is the charge of ions [174]. For the PET membrane and tenfold concentration gradient of KCl (C 1 /C 2 =10), E m ~ 50 mv, which gives t + ~0.9 (and t - ~0.1). Thus, 90% of current signal is therefore due to potassium ions [68], indicating that negatively charged PET nanopore is cation selective. Surface Modification and Analysis of PET Nanopore When we attempted to detect DNAs with an unmodified PET nanopore, no current-pulses were observed. We hypothesized that this was due to electrostatic repulsion between the anionic DNA molecules and the carboxylate groups on the pore walls [26]. In order to overcome this problem, ethanolamine was attached to the carboxylate groups to remove the anionic surface charge. This EDC chemistry [175] is shown schematically in Figure 2-4 [124]. Ethanolamine attachment was confirmed by using XPS. The unmodified PET membrane exhibits strong signals for C and O and a weaker signal for Na (Figure 2-5). The Na signal is due to Na +, which is present on the membrane surfaces after the NaOH etch as the counterion for the carboxylate groups. This is confirmed by the fact that the XPS spectrum for a membrane that had been reacted with ethanolamine shows 55

56 no Na signal (Figure 2-6), because the carboxylate groups have been converted to amide linkages to the amine. In addition, this membrane shows a signal due to N, confirming the presence of ethanolamine on the surface. The peaks at 720 ev and 990 ev are Auger peaks for oxygen and carbon, respectively. The surface atomic percentages [119] obtained from the XPS spectra are shown in Table 2-2. Considering the bare (not reacted with ethanolamine) membrane first, we see that the percentages of C and O are in good agreement with values calculated from the empirical formula of the PET monomer, C 10 O 4 H 8. It is worth mentioning, however, that since XPS does not detect H, the atomic percentages calculated from the monomer would be 71.4% and 28.6% for C and O, respectively. The discrepancy may be related to the fact that the empirical formula does not include Na. After reaction with ethanolamine, we see a decrease in the oxygen content of 1.4% and a concomitant increase in N from zero to 1.6%. It might be expected that the Na % of the bare membrane would be identical to the N % of the modified membrane, because for each amide link that is formed, one Na + -carboxylate should be lost. However, some of the carboxylates are undoubtedly present in the acid form (proton attached instead of Na + ). Furthermore, it is possible that the ionic groups are depleted from the surface because in so doing the polymer lowers its surface free energy. For these reasons, it is not surprising that the XPS data provide a Na % for the bare membrane that is lower than the N % for the modified membrane (Table 2-2). Current-voltage curves can also provide information about the surface charge on the pore walls. It is well known that when the radius of the tip opening of the conical pore is comparable to the thickness of the electrical double layer associated with fixed 56

57 surface charge on the pore wall, the nanopore will show a non-linear current-voltage curve. The reasons for this ion-current rectification have been discussed in literature [158, 176] and in Chapter 1. If, however, the surface charge is removed, then the pore will show a linear current-voltage curve [158]. In agreement, with this prior work, before modification with ethanolamine, the PET nanopore shows a nonlinear current-voltage plot, which becomes more linear after reaction with ethanolamine (Figure 2-7). The fact that the curve is not completely linear after modification indicates that less than 100% of the carboxylate sites have been removed. Conclusion This chapter has described fabrication of PET nanopores and characterization of their surface properties and behavior. First, the effect of UV irradiation on membrane breakthrough was investigated. We found that UV irradiation plays an important role in sensitizing track etching. Two hours of UV irradiation is sufficient for breakthrough within 45 min of etching and the breakthrough time can be adjusted by changing the irradiation time. Second, the rectification behavior as a function of solution ph was investigated. Our results showed that surface charge determines the rectification behavior of PET nanopores, which are cation selective in their bare form. The effect of electrolyte concentration on tip size was also investigated. It was found that the double layer thickness is comparable with the tip diameter and that high electrolyte concentration plays a critical role in tip size calculations. Furthermore, surface modification of PET nanopores was also performed using EDC/sulfo-NHS chemistry and the modification was confirmed by XPS and currentvoltage curves. For a bare PET membrane, carbon and oxygen content was confirmed 57

58 and after modification with ethanolamine nitrogen signal was observed in XPS spectra. Ethanolamine modification was also confirmed using current-voltage curves. A decrease in rectification was observed, indicative of less negative surface charge on pore walls after modification. 58

59 Current (na) Control 2h 6h 10h Time (min) Figure 2-1. Current vs time plots for single-track PET membranes exposed to UV for the stated times. 59

60 Figure 2-2. Schematic of a double layer in an ionoc solution in contact with a solid electrode. 60

61 Figure 2-3. Current-potential curves for bare PET nanopores at different ph s 61

62 Figure 2-4. Reaction scheme for the attachment of ethanolamine to surface carboxylate groups Figure 2-5. XPS spectrum for a bare PET membrane. 62

63 Figure 2-6. XPS spectrum for an ethanolamine-modified PET membrane. 63

64 Figure 2-7. Current-voltage curves in 10 mm KCl (ph 7) before ( ) and after ( ) ethanolamine modification. 64

65 Table 2-1. Calculated tip diameters in 0.1 M KCl and 1M KCl ph Tip Diameter in 0.1 M KCl (nm) Tip Diameter in 1M KCl (nm) Average Table 2-2. Percent atomic composition of PET membrane before (bare) and after modification with ethanolamine. Atom Percent Bare Ethanolamine modified C O Na N

66 CHAPTER 3 RESISTIVE-PULSE SENSING OF DNA USING CONICAL PET NANOPORE Introduction There is increasing interest in using nanopores in synthetic [16, 23-25, 39, 49, 81, 82, 86, 95, 157, 159, 166, 177, 178] or biological [128, 129, , , 144, 146, , 177] membranes as resistive-pulse sensors for molecular and macromolecule analytes. The resistive-pulse method [157], which when applied to such analytes is sometimes called stochastic sensing [128, 129, , ], entails mounting the membrane containing the nanopore between two electrolyte solutions. A transmembrane potential difference is applied and the resulting ion current flowing through the electrolyte-filled nanopore is measured. In the simplest terms, when the analyte enters and translocates through the nanopore, it transiently blocks the ion current, resulting in a downward current pulse. The current-pulse frequency is proportional to the concentration of the analyte, and the identity of the analyte is encoded in the current-pulse signature, as defined by the average magnitude and duration of the current pulses. DNA translocation through synthetic nanopores has been of particular interest [166, 177, 178]. With synthetic nanopores it is sometimes the case that the surface chemistry of the pore must be modified in order to detect DNA translocation events[26, 179]. For example, the silicon nitride pores investigated by Chen et al., had fixed negative charges on the pore wall, and this resulted in electrostatic repulsion of DNA from the pore [26]. This problem was solved by using an atomic-layer deposition method to uniformly coat the pore walls with uncharged Al 2 O 3 [26]. There have been numerous other reports of chemical functionalization of nanopore sensors [34, 124, 166, 66

67 175, ]. However, surface modification is not always necessary. For example, we have found that very large DNAs (e.g., a plasmid DNA of 6,600 base pairs (bp)) could be driven electrophoretically through conical nanopores in polycarbonate, in spite of the fact that this polymer has negative surface charge density [183, 184]. However, these nanopores had larger tip diameters than the pores described here, and the higher charge on these longer DNAs yields a higher net electrostatic force acting on the molecule, which may also aid in translocation. More recently, we and others [185] have been attempting to drive much smaller DNAs through nanopore sensing elements. In our case, the sensor is a conical nanopore in a PET membrane. PET also has fixed negative surface charge, due to carboxylate groups on the pore walls and membrane faces [15]. In contrast to the large DNAs sensed with pores in polycarbonate membranes [82], we found that we could not observe current pulses associated with translocation of the short DNAs through the PET nanopores. It seemed possible that this was due to electrostatic repulsion of the short DNAs from the negatively charged PET surface, and this led us to investigate membranes that were chemically modified to remove the surface charge. It has been shown that well-known EDC chemistry can be used to attach chemical functional groups to the carboxylate sites on PET [123, 175]. This yields an amide link to the -COO - groups on the pore walls and membrane faces, thus removing the negative charge. To test whether our inability to detect the short DNAs with the PET nanopores was due to electrostatic repulsion, we used EDC chemistry to attach a hydrophilic yet electrically neutral molecule, ethanolamine, to the -COO - sites on the PET. We have found that current pulses for 100-bp DNAs can be detected by the 67

68 resistive-pulse method with these chemically modified nanopores. In addition, we have found that the frequency of current-pulse is proportional to analyte concentration and applied potential. The results of these investigations are reported here. Materials & Methods Materials The double-stranded (ds) DNAs were obtained from Integrated DNA Technologies, as was Tris-EDTA (TE) buffer (10 mm Tris, 0.1 mm EDTA, ph = 7.5). The sequence of 100 bp DNA is 5'-AAT TCG AGC TCG GTA CCC GGG GAT CCT CTA GAG TCG ACC TGC AGG CAT GCA ATT CGA GCT CGG TAC CCG GGG ATC CTC TAG AGT CGA CCT GCA GGC ATG C-3' with its complementary strand. PET membranes, 12 μm thick, which contained a single heavy-ion induced damage track were obtained from GSI (Darmstadt, Germany). 1-Ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride (EDC), N-hydroxysulfosuccinimide (Sulfo-NHS), and 2-(Nmorpholino) ethanesulfonic acid buffered saline (MES) were obtained from Pierce. Ethanolamine hydrochloride was purchased from Aldrich. All other chemicals were of reagent grade and used as received. Purified water, obtained by passing housedistilled water through a Barnstead E-pure water-purification system, was used to prepare all solutions. Pore Etching The damage track in each PET membrane was chemically etched into a conically shaped pore (Figure 3-1) by using the two-step etching method described in detail previously [57]. The large-diameter (base) opening at one face of the membrane was ~520 nm. The small diameter (tip) opening at the opposite face was tailored to a specific size in each membrane using the two-step etching method. The diameter of the tip 68

69 opening was determined using an electrochemical method [57, 60]. Briefly, the membrane containing the single conical nanopore was mounted between the two halves of the conductivity cell described in [57], and an electrolyte solution was placed on both sides of the membrane. A Ag/AgCl electrode immersed into each solution was used to obtain a current-voltage curve for the electrolyte-filled nanopore. The slope of this linear curve is the ionic conductance of the electrolyte-filled pore. The conductance was used to calculate the diameter of the tip opening [57, 60]. The electrolye solution for these experiments was 1M KCl in 10 mm phosphate-buffered saline (ph 7). A representative current-voltage curve is given in Figure 3-2. EDC Modification The pore-etching procedure generates carboxylate groups on the pore walls and membrane faces [76]. EDC chemistry was used to attach ethanolamine to these -COO - groups via amide bond formation [186]. This was accomplished by first immersing the membrane into a solution that was 1 mm in EDC and 1 mm in sulfo-nhs dissolved in 10 mm MES-buffered saline. The membrane was immersed into this solution for ~2 hr and then immersed over night in a solution that was 1mg/ml in ethanolamine. HCl dissolved in purified water. Electrochemical Measurements To obtain current-pulse data, the membrane sample was mounted between the two halves of the conductivity cell [57], and each half-cell was filled with 3.5 ml of 1M KCl in 10 mm Tris, 0.1 mm EDTA. A Ag/AgCl electrode was inserted into each half-cell solution, and an Axopatch 200B (Molecular Devices Corporation) was used to apply a transmembrane potential of 1000 mv in the voltage-clamp mode with a low-pass Bessel filter at 2 khz bandwidth. The Axopatch measured the resulting ion current flowing 69

70 through the nanopore. The polarity was such that the anode was in the solution facing the base opening of the conical nanopore. The 100-bp DNA was added to the solution facing the tip opening and driven through the pore by electrophoresis [82, 85]. The signal was digitized using a Digidata 1233A analog-to-digital converter (Molecular Devices Corporation), at a sampling frequency of 10 khz. Data were recorded and analyzed using pclamp 9.0 software (Molecular Devices Corporation). Results and Discussion Current-Pulse Data The duration of the current-pulse ( ) is defined as the time interval between the precipitous drop and the time when current returns to the baseline value. The current pulse amplitude (Δi) is the difference in current between the baseline value and the lowest current within the pulse (Figure 3-3). Each translocation event is characterized by these two parameters. Effect of Concentration on Current-Pulse Frequency (ƒ b ) Applying the transmembrane potential resulted in steady-state current, and no current-pulse was observed until the 100 bp DNA was added to the tip side. Upon addition, numerous current-pulses were observed. The current-pulse frequency (ƒ b ) was determined by counting the number of pulses in 3 minute windows. The counts from three such windows were averaged to obtain current-pulse frequency (Figure 3-4). Figure 3-5 shows that there is a linear relationship between the frequency of these current-pulses (ƒ b ) and the concentration of the ds-dna. This can be explained by the electrophoretic flux (J mol s -1 cm -2 ) of ds-dna through the nanopore tip which is linearly related to the concentration of ds-dna (C) via equation 3-1. J = zfd t CE/RT

71 where z is the charge of the double-stranded DNA, D t is the diffusion coefficient associated with transport of the ds-dna through the tip opening, and E is the electrical field strength in the tip. The relationship between current-pulse frequency, ƒ b, and concentration can be seen more clearly by multiplying both sides of equation 3-1 by the cross sectionol area of the tip opening ( r 2 tip ) and then by Avogadro s number (A). This converts the left hand side of the equation 3-1 to molecules of ds-dna translocating the tip per second, which is the current-pulse frequency (ƒ b ). molecules s = f b = zfd t CE πr tip 2 A/RT 3-2 The charge z, on the ds-dna is equivalent to the number of bases, since there is one phosphate per base r tip was measured as described above. The concentration dependence was analyzed with 100-bp DNA concentrations ranging from 0.25 nm to 10 nm in our experiments. As given in equation 3-2, the current-pulse frequency has a linear relationship with concentration. This relationship had been observed in our previous studies [58, 82] and reported by other groups [146, 208]. Figure 3-5 shows the linear relationship between current-pulse frequency and the concentration of 100-bp DNA (red line, R 2 =0.9938). However, when the data were analyzed in detail, we found that the data obtained under 2 nm (blue line) gave a more linear fit (R 2 =0.9945) with a different slope. The slope value obtained from the low concentrations is half of the slope obtained from the total data, indicating lower translocation rates at low concentrations. We think such a difference in translocation rate may be due to the increased analyte-analyte and analyte-surface interaction at the tip opening of the nanopore. 71

72 Effect of Potential on Current-Pulse Frequency and Duration In addition to concentration dependence, the effect of potential on current-pulse frequency was also investigated. As given in equation 3-2, at fixed concentration of ds- DNA, the frequency of the current-pulses should increase linearly with electrical field strength (E). Figure 3-6 displays the current-pulse frequency in a 21 nm pore as a function of applied potential for 100 bp DNA. The current-pulse frequency shows a linear increase with applied potential and no current-pulses were observed under 600 mv. The linear relation between frequency and potential is consistent with other solidstate nanopore studies [25] and has been supported with theoretical studies [187, 188]. Figure 3-7 shows the duration of current-pulses ( ) for a range of voltages from 600 to 1000 mv. The results show that the average duration ( ) decreases at higher voltages. The reason for 1/V can be explained by the relation between the electrophoretic velocity and electrical field strength. The electrophoretic velocity ( ) of an ion is proportional to electric field strength (E) via equation 3-3 [189]. ν = z ee/6πηr 3-3 Where r is radius of ion, e is the electronic charge and is the viscosity of the solution. Taking the inverse of both sides of equation 3-3 yields 1 ν = 6πηr/ z ee 3-4 Multiplying both sides of equation 3-4 by the length of the detection zone, l d, makes the left-hand side into the translocation time through the detection zone,, which is equivalent to the current-pulse duration. τ = 6πηrl d / z ee

73 As seen in equation 3-5, the duration ( ) is inversely proportional to the electric field strength. Figure 3-7 illustrates the duration of 100-bp DNA as a function of potential. The result agrees with equation 3-5 in that the duration for 100-bp DNA increases when the potential is decreased. This result also agrees with our previous observations [50, 82] and with results using a biological protein channel [146]. Effect of Tip Diameter on Current-Pulse Frequency, Amplitude and Duration The effect of tip diameter on the current-pulse frequency was also investigated. 100 bp DNA was used for resistive pulse-sensing, and the same experiment was performed for nanopores with different diameters. The current-pulse frequency (ƒ b ) was determined by counting the number of pulses in three minute windows and averaging the counts from three such windows. The nanopores used throughout experiments have diameters comparable to the size of the 100-bp DNA. The double-stranded DNA molecule is 2.2 nm in diameter and the distance between the bases is ~0.33 nm [14]. So, 100 bp DNA is ~33 nm in length, which is below the reported persistence length of DNA (i.e., ~50 nm) [190, 191]. This property of 100-bp DNA eliminates the coiling of DNA inside the solution and gives a more rigid structure. The effect of tip diameter on current-pulse frequency was already reported in one of our previous studies [85]. The BSA molecule, which is shaped roughly like an American football with long axis of ~ 14 nm and a short axis of ~ 4 nm, was used as the analyte, [209]. The BSA molecules were translocated through the PET nanopore and results showed that when the tip diameter was smaller than the 14 nm long axis, causing ƒ b to be low. However, there was an exponential increase in ƒ b for the tips with 73

74 diameters larger than the long axis. This exponential increase was explained by the entropic penalty paid by the molecule when it enters the tip. The penalty is higher for the tip diameter smaller than the 14 nm BSA long axis. This is because the BSA molecule loses a degree of rotational freedom as it enters a tip with a small diameter. We investigated the effect of tip diameter on current-pulse frequency for 100-bp DNA and observed the same phenomenon as for BSA. As shown in Figure 3-8, when the tip diameter is smaller than length of 100-bp DNA (i.e., ~33 nm), the current-pulse frequency is low. However, there is a jump in frequency when the tip diameter is close to or larger than the length of 100-bp DNA (Figure 3-8). As mentioned above, this can be explained by the entropic penalty paid by the molecule when it enters the tip. The penalty is higher for small tips due to the loss of a rotational degree of freedom. In addition to current-pulse frequency, the change in current-pulse amplitude as a function of tip diameter was also investigated. As mentioned above, the resistive-pulse technique is based on a particle passing through a pore and displacing the fluid inside the pore. This causes a transient increase in pore resistance, which is measured as a decrease in current. For a particle smaller than the pore, the relative current change of the particle flowing through the pore can be given by: Δi I = D L arcsin ( d D ) 1 d D 2 d D 3-6 as predicted by DeBlois and Bean [19]. Here, d is the particle diameter, D is the pore diameter, and L is the pore length. For simplicity, equation 3-6 can be approximated as the volume ratio of particle to pore [37, 192, 193]: Δi I V particle V pore

75 In simple words, equation 3-7 gives the correlation between the current-pulse amplitude and the solution volume displaced by the particle. In order to calculate the relative current-pulse amplitude ( i/i), a Gaussian function was fitted to the background data to find the mean current (I) value. Then i values were divided by the calculated mean current value. For example, the background current value for 21 nm tip was 6.3 na from the Gaussian distribution, and i values were divided by this average current value to give i/i = 0.13± 0.03 The values of i/i for different tip diameters are plotted in Figure 3-9. The results agree with the equation 3-7. As expected i/i decreased as the tip diameter increased, which means that current-pulse amplitude is correlated with the ratio between the volume occupied by the 100-bp DNA and pore volume. However, when the tip diameter was 30 nm and over, this trend was not observed. This can be explained by the higher ratio between the DNA volume and the pore volume. Figure 3-10 shows a plot of duration ( ) data obtained from 100 bp DNA for nanopores with tip diameters ranging from 21 to 42 nm. Duration values of 100-bp DNA ranged from ms to ms when the tip diameter increased from 21 to 42 nm. Duration ( ) of 10 kb DNA as a function of pore diameter was also studied by Li et al.[22] who reported that is independent of pore diameter. A further investigation was made in our previous BSA sensing studies [85, 162] which showed that the duration of the BSA molecule is also independent of tip diameter. Effect of Electrolyte Concentration on DNA Translocation To study the effect of electrolyte concentration on DNA translocation, 0.5 M and 1 M KCl in Tris-EDTA were used for the translocation of 100-bp DNA. Current-time 75

76 recordings of DNA translocation at different electrolyte concentrations are given in Figure Duration and current-pulse amplitude data were collected and summarized in Figure The magnitude of current-pulse amplitude is formulated by Fologea et al. [194, 195]. Since current-pulse amplitude (Δi) is a function of solution conductivity ( ) (see equation 3-8), relative current-pulse amplitudes ( i/i) were calculated as explained in previous section, and results from both concentrations were compared. i = σva DNA /H 3-8 where V is applied potential, A is hydrodynamic cross section of the translocation molecule and H is the effective thickness of the nanopore. The duration values as a function of electrolyte concentration were found to be ms and ms for 0.5 M and 1M KCl, respectively. Duration results in 0.5 M KCl gave broader distribution than in 1 M KCl. The narrower duration and currentpulse amplitude distributions obtained with 1 M KCl are desirable. Therefore, all other resistive-pulse sensing experiments for differentiating short DNAs were carried out with 1 M KCl as the supporting electrolyte. The effect of electrolyte concentration on electrophoretic velocity (see equation 3-3) and its effect on translocation have been discussed in different studies [159, ]. Ghosal [210] calculated the translocation time as a function of KCl concentration and compared his theoretical results with the experimental data of Smeets et al. [159]. Their calculations showed that an optimal salt concentration exists for which the translocation speed is maximum and at 0.5 M KCl the translocation time is longer than at 1 M KCl. Their results showed good agreement with the experimental results in low electrolyte concentrations. They also pointed out the electrical double layer and electro- 76

77 osmotic flow oppose the translocation of DNA through the pore. To understand the effect of electrolyte on translocation, further studies must be made to understand the dynamics of translocation and optimize the conditionsfor translocation. Effect of High Potential on Current-Pulse Amplitude and Duration Since lipid bilayer supports can only withstand a few hundreds of mv, biological nanopore sensors have a limited range of potential [128, 156]. However, track etched membranes have physical durability and stability. All the previously reported sensing work has been done at the range of 1V. Restriction to this applied potential range was limited by the instrumentation. In order to overcome the instrument s limited potential range and apply higher potential, a homemade device was used. The device consists of a 9V battery placed in series with the head stage of the Axaopatch 200B. The main motivation was to improve the distribution in current-pulse durations and to obtain better differentiation among the sample molecules. In this regard, 100 bp DNA was translocated through the nanopore under 1 and 1.5 V. The current-pulse amplitude values were normalized as described in the previous section and the 100-bp DNA translocations were characterized by the current-pulse amplitude ( i) and duration ( ). Figure 3-13 shows a scatter plot of current-pulse amplitude versus duration for the 100-bp DNA translocation at 1V and 1.5 V. The scatter plot shows that no significant difference was observed in the duration or current-pulse amplitude. The duration values are calculated as ms and ms for 1 V and 1.5 V, respectively. Under both potentials, the duration and the current-pulse amplitudes gave a similar spread and distribution, which means that higher potential did not improve our ability to differentiate short DNA molecules. 77

78 On the other hand, Golovchenko and coworker showed the effect of potential on DNA translocation by increasing the potential from 60 mv to 120 mv. They observed a decrease in average duration value and standard deviation. However, by optimizing the surface characteristics of the nanopore and experimental conditions, better distribution in duration and current-pulse amplitude can be achieved. Conclusions In this chapter, it was shown that a single conical nanopore in a PET membrane could be prepared with good reproducibility. Ethanolamine modified PET nanopores were applied in resistive-pulse sensing of 100-bp DNA. As the focus of this chapter, 100-bp DNA was successfully translocated through the surface modified PET nanopore. The dynamics of the translocation phenomena were shown as a function of potential, concentration and tip diameter. The results showed that the current-pulse frequency has a linear relation with DNA concentration. By increasing the tip diameter, we observed an increase in current-pulse frequency due to the increase in the degree of freedom of DNA. Furthermore, we have shown that the frequency of current-pulses is linearly related to the applied potential and the duration of these translocations are inversely proportional to applied potential. The effect of electrolyte concentration was also investigated. The lower concentration electrolyte (0.1M KCl) yielded higher duration and broader distributions than the higher electrolyte concentration (1M KCl). We also investigated the effect of high potential. When 1.5 V was applied to translocate the DNA, the distributions of duration and current-pulse amplitude were observed to be similar to those obtained using 1 V. 78

79 Current (na) Figure 3-1. SEM images (A) conical PET nanopore (top view). (B) Au replica of conical nanopore Potential (V) Figure 3-2. Current-voltage curves in 1M KCl (PBS buffer, ph 7) before ( ) and after ( ) ethanolamine modification. 79

80 Figure 3-3. Representative current-time transients for the single nanopore membrane with 100 bp DNA in solution and expanded view of typical current-pulse signature associated with DNA translocation 80

81 Figure 3-4. Current-time transient for ethanolamine modified conical nanopore sensor.at 1000mV. Buffer only (A). 1 nm DNA in buffer (B). 5 nm DNA in buffer (C). 81

82 Current-Pulse Frequency (min -1 ) Current-Pulse Frequency (min -1 ) y=26x-5.4 R 2 = y=13x-0.3 R 2 = Concentration (nm) Figure 3-5. DNA current-pulse frequency versus DNA concentration. Transmembrane potential = 1000 mv. Tip diameter = 21 nm. Error bars represent the standard deviation obtained from the average number of pulses in three 3-min windows of the current-pulse data Potential (mv) Figure 3-6. DNA current-pulse frequency versus transmembrane potential. Tip diameter = 21 nm. [DNA] = 5 nm. Error bars represent the standard deviation obtained from the average number of pulses in three 3-min windows of the current-pulse data. 82

83 (msec) Current-Pulse Frequency (min -1 ) Potential (mv) Figure 3-7. Current-pulse duration ( ) versus potential. Tip diameter = 21 nm. Error bars represent the standard deviation obtained from the average number of pulses in three 3-min windows of the current-pulse data Tip Diameter (nm) Figure 3-8. DNA current-pulse frequency versus tip diameter. Transmembrane potential = 1000 mv. Error bars represents standard deviation obtained by averaging the number of pulses in three 3-min windows of the current-pulse data. 83

84 i/i Tip Diameter (nm) Figure 3-9. Plot of current-pulse ratio ( i/i) versus tip diameter (D tip ). Transmembrane potential = 1000 mv (msec) Tip Diameter (nm) Figure Plot of current-pulse duration ( ) versus tip diameter. Transmembrane potential = 1000 mv. 84

85 I/I Figure Current-time transient for 100 bp DNA in 0.5 M KCl (A). 1 M KCl (B) Tip diameter =32 nm. Transmembrane potential = 1000 mv M KCl 1 M KCl (msec) Figure Scatter plot of current-pulse-amplitude ( i/i) vs duration ( ) for 100 bp DNA in 0.5 M (black) and 1 M KCl (red) in TE buffer. Tip Diameter = 32 nm 85

86 i/i (pa) V 1.5 V (msec) Figure Scatter plot of current-pulse-amplitude ( i/i) vs duration ( ) for 100 bp DNA at 1 V(black) and 1.5 V (red). Tip Diameter = 27 nm 86

87 CHAPTER 4 DETECTION AND DIFFERENTIATION OF DNA USING CONICAL PET NANOPORE Introduction There is increasing interest in using nanopores in synthetic [16, 23-25, 39, 49, 81, 82, 86, 95, 157, 159, 166, 177, 178] or biological [128, 129, , , 144, 146, , 177] membranes as resistive-pulse sensors for molecular and macromolecule analytes. The resistive-pulse method [157], which when applied to such analytes is sometimes called stochastic sensing [128, 129, , ], entails mounting the membrane containing the nanopore between two electrolyte solutions. A transmembrane potential difference is applied and the resulting ion current flowing through the electrolyte-filled nanopore is measured. In the simplest terms, when the analyte enters and translocates through the nanopore, it transiently blocks the ion current, resulting in a downward current pulse. The current-pulse frequency is proportional to the concentration of the analyte, and the identity of the analyte is encoded in the current-pulse signature, as defined by the average magnitude and duration of the current pulses. DNA translocation through synthetic nanopores has been of particular interest [166, 177, 178]. With synthetic nanopores it is sometimes the case that the surface chemistry of the pore must be modified in order to detect DNA translocation events[26, 179]. For example, the silicon nitride pores investigated by Chen et al., had fixed negative charges on the pore wall, and this resulted in electrostatic repulsion of DNA from the pore [26]. This problem was solved by using an atomic-layer deposition method to uniformly coat the pore walls with uncharged Al 2 O 3 [26]. There have been numerous other reports of chemical functionalization of nanopore sensors [34, 124, 166, 87

88 175, ]. However, surface modification is not always necessary. For example, we have found that very large DNAs (e.g., a plasmid DNA of 6,600 base pairs (bp)) could be driven electrophoretically through conical nanopores in polycarbonate, in spite of the fact that this polymer has negative surface charge density [183, 184]. However, these nanopores had larger tip diameters than the pores described here, and the higher charge on these longer DNAs yields a higher net electrostatic force acting on the molecule, which may also aid in translocation. More recently, we and others [185] have been attempting to drive much smaller DNAs through nanopore sensing elements. In our case, the sensor is a conical nanopore in a PET membrane. PET also has fixed negative surface charge, due to carboxylate groups on the pore walls and membrane faces [15]. In contrast to the large DNAs sensed with pores in polycarbonate membranes [82], we found that we could not observe current pulses associated with translocation of the short DNAs through the PET nanopores. It seemed possible that this was due to electrostatic repulsion of the short DNAs from the negatively charged PET surface, and this led us to investigate membranes that were chemically modified to remove the surface charge. It has been shown that well-known EDC chemistry can be used to attach chemical functional groups to the carboxylate sites on PET [123, 175]. This yields an amide link to the -COO - groups on the pore walls and membrane faces, thus removing the negative charge. To test whether our inability to detect the short DNAs with the PET nanopores was due to electrostatic repulsion, we used EDC chemistry to attach a hydrophilic yet electrically neutral molecule, ethanolamine, to the -COO - sites on the PET. We have found that current pulses for 100-bp DNAs can be detected by the 88

89 resistive-pulse method with these chemically modified nanopores. In addition, we have found that the frequency of current-pulse is proportional to analyte concentration and applied potential. The results of these investigations are reported here. Materials & Methods Materials The double-stranded (ds) DNAs were obtained from Integrated DNA Technologies, as was Tris-EDTA (TE) buffer (10 mm Tris, 0.1 mm EDTA, ph = 7.5). The 25 bp DNA consisted of a duplex of 5'- AAT TCG AGCTCG GTA CCC GGG GAT C-3', 50 bp DNA consisted of a duplex of 5'-AAT TCG AGC TCG GTA CCC GGG GAT CCT CTA GAG TCG ACC TGC AGG CAT GC-3' and 100 bp DNA consisted of a duplex of 5'-AAT TCG AGC TCG GTA CCC GGG GAT CCT CTA GAG TCG ACC TGC AGG CAT GCA ATT CGA GCT CGG TAC CCG GGG ATC CTC TAG AGT CGA CCT GCA GGC ATG C-3' with their complementary strand. Single stranded DNA had the same sequence with 100 bp DNA given above without its complimentary strand. The PET membranes, 12 μm thick, which contained a single heavy-ion induced damage track were obtained from GSI (Darmstadt, Germany). 1-Ethyl-3-[3- dimethylaminopropyl] carbodiimide hydrochloride (EDC), N-hydroxysulfosuccinimide (Sulfo-NHS), and 2-(N-morpholino) ethanesulfonic acid buffered saline (MES) were obtained from Pierce. Ethanolamine hydrochloride was purchased from Aldrich. All other chemicals were of reagent grade and used as received. Purified water, obtained by passing house-distilled water through a Barnstead E-pure water-purification system, was used to prepare all solutions. 89

90 Pore Etching The damage track in the PET membrane was chemically etched into a conically shaped pore by using the two-step etching method described in detail previously [57]. The large-diameter (base) opening at one face of the membrane was ~520 nm, and the small diameter (tip) opening at the opposite face was prepared in different sizes. The diameter of the tip opening was determined using the electrochemical method [57, 60]. Briefly, the membrane containing the single conical nanopore was mounted between the two halves of the conductivity cell described in [57], and an electrolyte solution was placed on either side of the membrane. A Ag/AgCl electrode immersed into each solution was used to obtain a current-voltage curve for the electrolyte-filled nanopore. The slope of this linear curve is the ionic conductance of the electrolyte-filled pore. The conductance was used to calculate the diameter of the tip opening [57, 60]. EDC Modification The pore-etching procedure generates carboxylate groups on the pore walls and membrane faces [76]. EDC chemistry was used to attach ethanolamine to these -COO - groups via amide bond formation [186]. This was accomplished by first immersing the membrane into a solution that was 1 mm in EDC and 1 mm in sulfo-nhs dissolved in 10 mm MES-buffered saline. The membrane was immersed into this solution for ~2 hr and then immersed over night in a solution that was 1mg/ml in ethanolamine. HCl dissolved in purified water. Electrochemical Measurements To obtain current-pulse data, the membrane sample was mounted between the two halves of the conductivity cell [57], and each half-cell was filled with 3.5 ml of 1M KCl in 10 mm Tris, 0.1 mm EDTA. A Ag/AgCl electrode was inserted into each half-cell 90

91 solution, and an Axopatch 200B (Molecular Devices Corporation) was used to apply a transmembrane potential of 1000 mv in the voltage-clamp mode with a low-pass Bessel filter at 2 khz bandwidth and measure the resulting ion current flowing through the nanopore. The polarity was such that the anode was in the solution facing the base opening of the conical nanopore. The DNA was added to the solution facing the tip opening and driven through the pore by electrophoresis [82, 85]. The signal was digitized using a Digidata 1233A analog-to-digital converter (Molecular Devices Corporation), at a sampling frequency of 10 khz. Data were recorded and analyzed using pclamp 9.0 software (Molecular Devices Corporation). Results and Discussion We are interested in resistive-pulse sensing of short ds-dnas because they are ideal model systems for studying how molecular weight within a homologous series of analytes affects the current-pulse signature. A key question is: how similar in molecular weight can two homologous analytes be and yet still be differentiable by the resistivepulse method? Conically shaped nanopores were used because we have shown that a sensitive analyte detection zone is present just inside the tip opening of conical pores [85]. PET was used for these studies because we have shown that conically shaped nanopores can be prepared with good reproducibility in PET [57]. Effect of Tip Diameter on DNA Current-Pulse Amplitude and Duration Before differentiation experiments, the effect of tip diameter on current-pulse amplitude and duration was studied. The histogram of DNA current-pulse amplitude data were obtained from three different tip diameters. The 50 bp DNA was used for the sensing experiment due to its rigid structure compared to 100 bp DNA. Each histogram was fitted to a Gaussian distribution, which provided the average pulse amplitude and 91

92 standard deviation. Tip diameters of 10, 16 and 35 nm gave current-pulses with average i values , and pa and durations 1 1.5, , ms respectively. The difference in suggests translocation dynamics are weakly dependent on pore size. In Figure 4-2, a summary of translocation data for 50-bp DNA using three different tip diameters is given. The scatter plot of i versus showed a broad distribution for the 10 nm nanopore. In contrast, 16 and 35 nm nanopores gave narrow distributions. This distribution behavior again can be explained by the size match between the tip diameter and length of the molecule, as discussed in the previous chapter. When the tip diameter is decreased from 35 nm to 10 nm, 50-bp DNA has less degree of freedom for translocation. The length of 50-bp DNA is ~15 nm and the decrease in the degree of freedom results in a broad distributions in duration and current-pulse amplitude. The data obtained from these three different sized tips are shown on the histogram of current pulse amplitude (Figure 4-3). The results show that a better differentiation can be established by increasing the tip diameter and using the mean current-pulse amplitude. Differentiation of Analytes Using Current-Pulse Signatures In the absence of DNA, a steady-state ion current (no current-pulses) of 11.6 na was observed for the ethanolamine-modified nanopore with tip diameter of 35 nm. When either the 50- or 100-bp DNA was added to the tip side, current-pulse events were observed (Figure 4-4). These current-pulse events can be characterized by the current-pulse amplitude ( i) and its duration ( ). As per our prior work [85], i is defined as the difference in current between the baseline and the lowest current within a pulse, 92

93 and is defined as the time interval between the departure and return of the current from/to the baseline value (see insets Figure 4-4). Histograms of the i data for the 50- and 100-bp DNAs are shown in Figure 4-5. The histograms were fit to a Gaussian distribution (solid curves), which provided the average current-pulse amplitude. Average i values of pa and pa were observed for the 50- and 100-bp DNAs, respectively. These data show that the ethanolamine-modified nanopore sensor can discriminate between the 50- and 100-bp DNAs on the basis of the current-pulse amplitude. This is the first report using a synthetic nanopore sensor to discriminate between such short DNAs. The currentamplitude is larger for the larger 100-bp DNA. This is consistent with a simple Coultercounter-based explanation, in which the magnitude of the current pulse is related to the volume element of electrolyte solution displaced as the molecule traverses the sensing zone [95, 196]. Analogous histograms of the current-pulse duration data yielded average values of 1±0.4 ms and 1±0.3 ms for the 50- and 100-bp DNAs, respectively (Figure 4-6). These data show that current-pulse duration cannot be used to differentiate between these two short ds-dnas. This is because, as described in our prior work [81, 82, 85] and work other labs [32, 149], the distribution in pulse duration is much broader than the distribution in pulse amplitude. This can be seen more clearly in Figure 4-7. In contrast to our results, Heng et al., could distinguish between various ds-dnas on the basis of current-pulse duration [147], but the DNAs were 100 bp, 600 bp and 1500 bp, a much larger spread in size than those investigated here. 93

94 The same experiment was performed to differentiate 25 bp DNA from 100 bp DNA using ethanolamine modified nanopore with tip diameter of 21 nm. Since 25 bp DNA is ~7-8 nm, a smaller tip diameter was used for better differentiation. A steadystate ion current (no current-pulses) of 6.3 na was observed. When either the 25- or 100-bp DNA was added to the tip side, current-pulse events were observed. Histograms of the i data for the 25- and 100-bp DNAs are shown in Figure 4-8. The histograms were fit to a Gaussian distribution (solid curves), which provided the average currentpulse amplitude. Average i values of pa and pa were observed for the 25- and 100-bp DNAs, respectively. In addition, histograms of the current-pulse duration data yielded average values of 1.4±0.5 ms and 1.7±0.5 ms for the 25- and 100-bp DNAs, respectively (Figure 4-9). The distribution can be seen clearly in Figure Differentiation of single stranded DNA (ss-dna) and double stranded DNA (ds- DNA) was also demonstrated using DNA s containing 100 bases and 100 base pairs. For ss-dna, same base sequence of 100-bp DNA was used without its complementary strand. An ethanolamine modified nanopore with tip diameter of 37 nm was used. Histograms of the i data for the ss- and ds-dna are shown in Figure The histograms were fit to a Gaussian distribution (solid curves), which provided the average current-pulse amplitude. Average i values of 63+9 pa and pa were observed for the ss- and ds-dnas, respectively. Analogous histograms of the current-pulse duration data yielded average values of 1.2±0.4 ms and 1.6±0.5 ms for the ss- and ds- DNAs, respectively (Figure 4-12).These results show that our nanopore sensor can be 94

95 used to differentiate the single stranded and double stranded DNAs using current-pulse amplitude (Figure 4-13). Potential dependence of ss-dna was also investigated and compared with ds- DNA (Figure 4-14). As given in equation 3-2 and 3-3, the current-pulse frequency and electrophoretic velocity are directly proportional to the electrical field and the charge of particle. However, our results did not show the charge correlation between ss- and ds- DNA. Double stranded DNA has 2 electron charges per basepair (or 0.98 nc/m) coming from one negative charge for each phosphate group. The expected charge of the ss- DNA used in our experiments is half of the ds-dna. This should be reflected in the translocation frequency, but the translocation frequencies of ss- and ds-dna are not correlated with their theoretical charge values. The translocation frequency of ss-dna was nearly ten times less than the translocation frequency of ds-dna. This result can be attributed to the effective charge of DNA in solution. Recent studies showed that the effective charge of double-stranded DNA can be less than the theoretical charge value. Hall et. al. [197] and Smeets et. al [159] reported the effective charge of double stranded DNA as and electron charges per base pair, respectively. Both of these values are less than the expected charge of 2 electron charges per base pair. On the other hand, the effective charge of single stranded DNA was studied by Fogarty et. al. [213], who pointed out that counterions in the background solution can quench the ss-dna charge and can even neutralize the total charge of ss-dna. Their results showed that the effective charge of 40 bases of ss-dna is nearly 3 electron charges per base which is only 8% of the theoretical value (i.e.,40 e - ). The neutralization 95

96 of charge by cations was also shown by simulation studies which agreed with the experimental results [214]. All these results confirm that the effective charges of ss- and ds-dna are less than the theoretical values and ss-dna is more open to the neutralization effect of cations in solution. Conclusions In this chapter, we have shown that chemically modified nanopore can be used as a resistive-pulse sensor for the detection and differentiation of short ds-dnas. We first demonstrated that tip diameter can change the distribution of current-pulse amplitude and duration. This result was critical because in order to differentiate the current-pulses of different analytes (i.e., short DNAs), duration and/or amplitude of current-pulses should give separate distributions. Our results showed that small tip diameter gives broader distribution in current-pulse amplitude than the large tip diameter. The results also showed that our nanopore sensor can distinguish between a 50- bp DNA and a 100-bp DNA on the basis of the magnitude of the current pulses produced by these DNAs. There are no prior examples of distinguishing between such closely sized DNAs using a synthetic nanopore sensor. In contrast, a biological nanopore sensor based on the protein -hemolysin has been used to distinguish between such short DNAs [129, 149, 150]. However, the -hemolysin-based sensor can detect only ss-dna, not ds-dna as reported here. Smaller DNAs were also investigated and 25 bp DNA was differentiated from 100 bp DNA. Single-stranded DNA was also studied, and it was shown that the conical nanopore sensor can be used to distinguish between ss- and ds-dna. ss-dna showed lower translocation rate than ds-dna due to the charge difference. 96

97 In our experiments, current-pulse amplitude data were used to differentiate the analytes. In the literature, there are examples of use of current-pulse amplitude to differentiate the analytes [48, 131, 195, 198]. The surface charge, geometry or experimental conditions may play a critical role in translocation process and in time resolution. On the other hand, in our experiments, duration values for each analyte were close to each other and gave us limited information about our tranlocation. Less sensitive results in duration can be explained by the speed of DNA in solution. The lengths of DNAs used in our experiments are shorter than those used in studies [24, 82, 198]. Fologea et. al. [195] reported the speed of the translocation as 3 base/μs using solid state nanopore. Although we cannot completely adapt their results to our findings, we can estimate the translocation of 100 bp DNA as ~33 μs, which is under our duration results. 97

98 Figure 4-1. SEM images of conical PET nanopore (top view) (A) Au replica of conical nanopore (B). 98

99 Counts i (pa) nm 16 nm 35 nm (msec) Figure 4-2. Scatter plot of current-pulse-amplitude vs duration for 50 bp DNA. Transmembrane potential = 1000 mv i (pa) 10 nm 16 nm 35 nm Figure 4-3. Histograms of 50 bp DNA current-pulse-amplitude data for three different tip diameters. Transmembrane potential = 1000 mv. 99

100 Figure 4-4. Current-time transients from an ethanolamine-modified conical nanotube sensor. (A) Buffer only. (B) Buffer plus 50 bp DNA (10 nm). (C) Buffer plus 100 bp DNA (10 nm). Applied transmembrane potential was 1000 mv. 100

101 Counts Counts bp 100 bp i (pa) Figure 4-5. Histograms of DNA current-pulse-amplitude data for 50 (red) and 100 (green) bp DNAs bp 100 bp (msec) Figure 4-6. Histograms of DNA current-pulse-duration data for 50 (red) and 100 (green) bp DNAs. 101

102 Counts i (pa) 50 bp 100 bp (msec) Figure 4-7. Scatter plot of current-pulse-amplitude vs duration for 50 bp (red) and100 bp (green) bp 100 bp i (pa) Figure 4-8. Histograms of DNA current-pulse-amplitude data for 25 bp DNA (red) and 100 bp DNA (green). 102

103 i( pa) Counts bp 100 bp (msec) Figure 4-9. Histograms of DNA current-pulse-duration data for 25 bp DNA (red) and 100 bp DNAs (green) bp 100 bp (msec) Figure Scatter plot of current-pulse-amplitude vs duration for 25 bp (black) and100 bp (red). 103

104 Counts Counts ss ds i(pa) Figure Histograms of DNA current-pulse-amplitude data for ss-dna (red) and ds- DNA (green). ss ds (msec) Figure Histograms of DNA current-pulse-duration data for ss-dna (red) and ds- DNA (green). 104

105 Current-Pulse Frequency (min -1 ) i (pa) ss ds (msec) Figure Scatter plot of current-pulse amplitude versus duration for ss-dna (black) and ds-dna (red). 250 ds DNA ss DNA Potential (mv) Figure DNA current-pulse frequency versus transmembrane potential. DNA concentration = 5 nm. Tip diameter = 32 nm. 105

106 CHAPTER 5 NANOCONE FABRICATION USING ASYMETRIC ETCHING CONDITIONS Introduction Template synthesis, which entails synthesis or deposition of a desired material within pores of a nanopore membrane that serves as a template, has been used to fabricate nanomaterials [5, 199]. The size of monodisperse pores can be adjusted from nanometer to microns. In previous studies cylindrical nanostructures were synthesized as either solid nanowires or hollow nanotubes [200]. A number of applications have been reported for conically shaped nanostructures such as in field-emission devices [201], dye-synthesized solar cells [202], and electrochemical supercapacitors [203]. The importance of conically shaped nanopores has led us to develop a method for preparing such pores in polymer membranes. It has been shown that the track-etch method [204] can be used as a starting place for preparing such conical nanopores [15, 60, 62, 66, 69, 98]. This method entails bombarding a thin film of polymer with beam of high energy particles to create parallel damage tracks through the film. To make cylindrical pores, the track etched membrane is immersed into etching solution so that each track is converted into a pore. In order to fabricate conical pores, the tracked membrane is mounted in an etching cell with an etch solution on one side of the tracked membrane and a stop solution on the other side. Since the damaged tracks are etched faster at the face of the membrane exposed to the etch solution than at the face of the membrane exposed to the stop solution, conically shaped nanopores are obtained. While this stop etch approach has been used to make conical nanopores in a variety of different tracked materials, a key issue that remains is fine control over the 106

107 geometry of the resulting conical nanopore. Recently, it has been shown that the cone angle of conical pores etched into poly(ethylene terephthalate) (PET) films can be controlled by varying the chemical composition of the etching solution. In this study, the stop etch chemistry has been investigated and it was found that precise control over pore geometry can be obtained for conical pores etched in PET membranes. Material & Methods Materials Chemicals and materials were used as received, NaOH, KCl and CHCl 3 (certified A.C.S., Fisher Scientific), HCOOH (88% in water, Fisher Biotech), 1, 1, 1, 3, 3, 3-hexa fluoro-2-propanol (HFIP) (99+%, Aldrich), alumina filters (Anodisc, 25 mm, 0.1 μm, Whatman). Ion-tracked PET films (12 μm thick, 10 6 tracks cm -2 ) were obtained from GSI, Darmstadt, Germany. Chemical Etching of Conical Nanopores Ion-tracked poly(ethylene terephthalate) membranes were mounted between etching cells. One side of membrane was filled with etching solution (9 M NaOH) and the other side (as the tip side of membrane) was filled with a stopping solution (HCl and HCOOH in different concentrations). Potential was applied across the membrane by two Pt electrodes and the electrical current was monitored with a Keithley 6487 picoammeter/voltage source. The electrode in the etching solution served as the anode, and the electrode in the stopping solution served as cathode. The current was initially zero, but when the etching solution broke through the membrane to the stop solution, the current value increased gradually. Etching was stopped after 2 hours, and the membrane was immersed in stopping solution for 30 min. Then the etched membrane was kept in 18MΩ water to remove residual salts. 107

108 Electroless Gold Plating Template synthesis via the electroless deposition of gold was performed as previously described in Chapter 1. To prepare the nanomaterials described here, plating was performed overnight in order to fill the pores and obtain complete coverage of all surfaces. Electroless plating causes a gold nanocone to be deposited in each pore; in addition, both faces of the membrane are coated with thin films of gold. Liberating the Gold Nanocones To obtain suspensions of gold nanocones, the gold surface films on both faces of the membrane were removed by applying and then removing a piece of Scoth tape (810), and the membrane was cut from the original tape mask. The membrane was then immersed in a solution of HFIP/CHCl 3 (1:9 by volume) for 30 min. This results in dissolution of the membrane, thus liberating the template synthesized gold nanocones. The nanocones were collected from solution by filtration on an anodisc filter. The anodisc filter and filtrate were then coated with Au/Pd using a Denton Vacuum Desk II sputtering system. Electromicrographs were obtained using a Hitachi S 4000 scanning electron microscope (SEM). Results and Discussion Figure 5-1 shows the schematic for preparation of gold nanocones using template synthesis. During chemical etching of ion-track membranes, the damaged zone of a latent track is transformed into the nanopore. In order to determine the pore shape, the pores were filled with gold and the membrane was dissolved in HFIP/CHCl 3 solution. Conical nanocones were obtained by using an electroless plating method to deposit gold within the pores of the membrane and on the surface layers that coat both faces of the membrane. To liberate the nanocones, gold plated membranes were dissolved in 108

109 hexafluroisopropanol to yield a suspension of gold nanocones. The suspension was filtered through an alumina filter, and the filter surface was imaged using field-emission scanning electron microscopy (FESEM). A representative SEM image of multipore PET membrane (base side) is given in Figure 5-2. Effect of Strong Acid on Cone Angle The effect of stopping solution on nanocone shape was investigated. HCl was used as a stopping solution in different concentrations (i.e., 0.1, 1, 10 M). The cone angles ( ) of the resulting gold nanocone replicas were compared (Figure 5-3). Figure 5-4 shows the gold replicas fabricated using 3 different concentrations of HCl as stopping solution. Cone angles were calculated as 2.4 ±0.6, 2.4 ±0.6, 2.6 ±0.3 for 0.1, 1, 10 M HCl, respectively. Our calculations showed no significant difference in cone angle. This can be due to acid strength of HCl. As mentioned before, after breakthrough the etching solution (9M NaOH) and stopping solution meet at the tip side of the pore, where they neutralized one another. Through this process, the concentration of HCl had no observable effect on cone angle since its dissociation is not dependent on its concentration. Effect of Weak Acid on Cone Angle The effect of weak acid was also investigated. Three different concentrations (i.e., 0.1, 1, 10 M) of formic acid were used as stopping solution and cone angles were calculated to be 1.4 ±0.2, 2.7 ±0.4, 2.6 ±0.5 for 0.1, 1, 10 M HCOOH, respectively. We observed a significant increase in cone angle over 0.1 M HCOOH. On the other hand, the cone angle did not change when the concentration was increased to 10 M. The difference in cone angle can be due to weak acid character of formic acid. 109

110 Effect of Potential on Cone Angle As mentioned in previous chapters, breakthrough of the tracked membrane is monitored by applying potential. When the etch solution penetrates through the membrane, a jump is observed indicating that breakthrough has occuried. Apel et al. [60] reported that the electric field controls the transport of etching ions in the pore, promoting or stopping the process (Figure 5-6). We investigated the effect of electric field on breakthrough using different applied potentials (i.e., 10, 20, 30 V) during chemical etching. Results of the etching process are shown in Figure 5-7. The cone shapes were different from those in previous results. The cone angle in the tip side did not change; however through the rest of the membrane, the cone angle increased dramatically. This can be explained by the electrostopping of the etching process and/or by increase in solution temperature due to resistive heating which can also increase the etching rate [80]. In addition to cone angle change, we also observed change in surface roughness, which reflects the porewall surface. The surface roughness was discussed in various studies, which pointed out that PET has a partial crystalline structure [80, 205, 206], which means PET has two domains, crystalline and amorphous. The difference in etching rates of these domains was discussed by Lueck et al. [205], who concluded that amorphous domains etch faster than crystalline domains. In a recent study, Mukaibo et. al. [80], showed that longer etching time can increase surface roughness due to deeper etching into amorphous regions. Bomko et al. [207] also showed the higher etching rate of PET at high temperatures. We suggest that not only longer etching time but also the increase in solution temperature, due to resistive heating, can increase the etching rate 110

111 of amorphous regions and cause deeper etching in amorphous regions, resulting in rougher surfaces. Conclusion In this chapter, we investigated the effect of different etching conditions on gold nanocone shape and cone angle. For HCl and HCCOH stopping solutions, the results were similar. On the other hand, formic acid gave more promising results than hydrochloric acid. Control of formic acid concentration may be used to manipulate the cone angle and cone shape, though the cone angle may only vary from ~1 to 2. The applied potential can change the shape and cone angle more dramatically than stopping solutions and be used to manipulate the shape of nanocones as well.by changing the conditions, further control of nanostructures are possible and these structures will find applications in nanotechnology, including template synthesis, separations, and chemical sensing. 111

112 Figure 5-1. Schematic diagram of the method used to prepare gold replicas. 112

113 Figure 5-2. Representetive SEM image of multipore PET membrane (base side). Figure 5-3. Diagram of a conical nanopore with notations (drawing of the pore not to scale). 113

114 Figure 5-4. Representetive SEM image of gold nanopore replicas etched with (A) 0.1 M, (B) 1M, (C) 10 M HCl. 114

115 Figure 5-5. Representetive SEM image of gold nanopore replicas etched with (A) 0.1 M, (B) 1M, (C) 10 M HCOOH. 115

116 Figure 5-6. Direction of electro-migration and diffusion during breakthrough under stopping electric field conditions (drawing of the pore not to scale). 116

117 Figure 5-7. Representetive SEM image of gold nanopore replicas etched at (A) 10, (B) 20, (C) 30 volts. 117