The Pennsylvania State University
The Graduate School
Department of Chemistry
CREATING AND PROBING MOLECULAR ASSEMBLIES FOR
SINGLE-MOLECULE DEVICES
A Dissertation in
Chemistry
by
Amanda Michelle Moore
© 2008 Amanda Michelle Moore
Submitted in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
August 2008
ii The dissertation of Amanda Michelle Moore was reviewed and approved* by the following:
Paul S. Weiss Distinguished Professor of Chemistry and Physics Dissertation Advisor Chair of Committee
Thomas E. Mallouk DuPont Professor of Materials Chemistry and Physics
David L. Allara Professor of Polymer Science and Chemistry
Mary Jane Irwin Evan Pugh Professor of Computer Science and Engineering A. Robert Noll Chair of Engineering
Ayusman Sen Professor of Chemistry Head of the Department of Chemistry
*Signatures are on file in the Graduate School
iii ABSTRACT
We explored the creation and conductance of molecular assemblies by looking at both the larger body of work that has been performed to characterize molecular devices, and through probing isolated single-molecule assemblies and creating cluster tether schemes.
We have characterized the bistable conductance switching exhibited by oligo(phenylene-ethynylene) molecules. This conductance switching has been hypothesized to occur through a variety of interactions including reduction, rotation, neighboring molecule interactions, bond fluctuations and changes in hybridization. Using molecular design, we tailored the switch molecules to enable the testing of each mechanism. The hypothesis most consistent with our and others’ data is that of hybridization change at the molecule-substrate interface.
The oligo(phenylene-ethynylene) switches also exhibited motion within the host self-assembled monolayers at substrate step edges. This was observed as three apparent heights in our analyses of scanning tunneling microscope data. We have characterized the
‘third’ apparent height as arising from the switch molecules place-exchanging up and down substrate step edges. Only switches residing at substrate step edges have the ability to exhibit three apparent heights, as compared to those isolated at domain boundaries, which only exhibit two possible apparent heights. The oligo(phenylene-ethynylene) molecules were further characterized to show that the switching and motion events occur on time scales faster (ms) than those of scanning tunneling microscopy imaging (min) using height vs. time measurements. iv We have expanded the capabilities of scanning tunneling microscopy to include
measurements using microwave frequencies. Using AC measurements, we were able to
compare the polarizabilities of several self-assembled monolayers. We applied the
microwave frequencies to the systems of the oligo(phenylene-ethynylene) switches and
gained predicative abilities of which molecules were likely to exhibit conductance
switching or motion.
In addition to our studies of isolating and studying switch molecules, we have
developed capture surfaces for superatom clusters, which have element-like properties.
We used barium ions to model the capture for alkaline-earth-metal-like clusters. We are
developing the captures surfaces both to identify the cluster properties and to tether them for measurement with scanning tunneling microscopy.
These studies demonstrate our abilities to capture, to isolate and to measure
properties of molecules on the single-molecule scale.
v TABLE OF CONTENTS
LIST OF FIGURES ...... viii
LIST OF TABLES...... xii
LIST OF ABBREVIATIONS...... xiii
ACKNOWLEDGMENTS...... xv
CHAPTER 1 CREATING AND IMAGING MOLECULAR ASSEMBLIES...... 1 1.1 Introduction...... 1 1.2 Self-Assembled Monolayers...... 2 1.2.1 Insertion into Self-Assembled Monolayers...... 5 1.2.2 Mixed Alkanethiolate Monolayers...... 7 1.3 Scanning Tunneling Microscopy...... 8 1.3.1 Scanning Tunneling Microscopy Operation ...... 10 1.4 Microwave-Frequency Alternating Current Scanning Tunneling Microscopy...... 13 1.5 Dissertation Overview...... 19
CHAPTER 2 MOLECULAR DEVICES ...... 22 2.1 Introduction...... 22 2.1.1 The Drive Towards Single-Molecule Electronics ...... 22 2.2 Molecules and Measurements at the Single-Molecule Scale...... 31 2.2.1 Scanning Tunneling Microscopy...... 31 2.2.2 Conductive Probe Atomic Force Microscopy...... 37 2.2.3 Mechanically Controlled Break Junctions...... 39 2.2.4 Scanning Tunneling Microscopy Tip Break Junctions...... 41 2.2.5 Electromigration Junctions...... 42 2.3 Converting to the Nanoscale...... 45 2.3.1 Mercury Drop Junction...... 45 2.3.2 Nanopores...... 46 2.3.3 Particle Bridge...... 47 2.3.4 Crossed-Wire Junction...... 48 2.3.5 Nanorod Junction...... 50 2.3.6 Tip-End Junction...... 52 2.4 Connecting to the Outside World...... 54 2.4.1 Molecular Rulers...... 55 2.4.2 Directed Diblock Copolymers...... 57 vi 2.4.3 Imprint Lithography...... 58 2.4.4 On-Wire Lithography...... 59 2.5 Conclusions and Future Molecular Devices Direction...... 60
CHAPTER 3 TESTING HYPOTHESIZED SWITCHING MECHANISMS FOR SINGLE OLIGO(PHENYLENE-ETHYNYLENE) MOLECULES. . . 62 3.1 Introduction...... 62 3.2 Experimental Procedure...... 67 3.2.1 Sample Preparation...... 67 3.2.2 Scanning Tunneling Microscopy...... 70 3.2.3 Apparent Height Determination...... 70 3.3 Results and Discussion...... 73 3.3.1 Reduction and Functional Group Rotation Mechanisms...... 74 3.3.2 Backbone Ring Rotation Mechanism...... 75 3.3.3 Neighbor Molecule Interactions Mechanism...... 76 3.3.4 Bond Fluctuations...... 79 3.3.5 Hybridization Change...... 84 3.3.6 Experiments with Restricted Motion...... 86 3.4 Conclusions and Future Direction...... 89
CHAPTER 4 MOTION UP AND DOWN SUBSTRATE STEP EDGES BY OLIGO(PHENYLENE-ETHYNYLENE) MOLECULES...... 91 4.1 Introduction...... 91 4.2 Experimental Procedure...... 94 4.2.1 Sample Preparation...... 94 4.2.2 Scanning Tunneling Microscopy...... 96 4.2.3 Apparent Height Determination...... 96 4.3 Results and Discussion...... 99 4.4 Conclusions and Future Directions...... 110
CHAPTER 5 REAL-TIME MEASUREMENTS OF CONDUCTANCE SWITCHING AND MOTION OF SINGLE OLIGO(PHENYLENE-ETHYNYLENE) MOLECULES...... 111 5.1 Introduction...... 111 5.2 Experimental Procedure...... 115 5.2.1 Sample Preparation...... 115 5.2.2 Scanning Tunneling Microscopy...... 115 5.2.3 Height vs. Time Spectroscopy Acquisition...... 115 5.3 Results and Discussion...... 117 5.4 Conclusions and Future Direction...... 125
CHAPTER 6 IMAGING SINGLE-MOLECULE POLARIZABILITY AND BURIED INTERFACE DYNAMICS...... 126 6.1 Introduction...... 126 6.1.1 Sample Preparation...... 129 vii 6.1.2 Scanning Tunneling Microscopy...... 131 6.1.2.1 Alternating Current Scanning Tunneling Microscopy Imaging and Spectral Acquisition...... 132 6.1.3 Single-Molecule and Microwave Magnitude Extraction...... 132 6.2 Results and Discussion ...... 133 6.2.1 Probing the Polarizability of Self-Assembled Monolayers ...... 133 6.2.2 Buried Interface Dynamics of Oligo(phenylene-ethynylene) switches ...... 142 6.3 Conclusions and Future Direction ...... 155
CHAPTER 7 DEVELOPMENT OF CLUSTER CAPTURE SURFACES ...... 156 7.1 Introduction ...... 156 7.2 Experimental Procedure ...... 158 7.2.1 Sample Preparation...... 158 7.2.2 Time-of-Flight Secondary Ion Mass Spectrometry ...... 161 7.2.3 X-Ray Photoelectron Spectroscopy ...... 161 7.3 Results and Discussion ...... 163 7.3.1 Electrostatic Capture Surfaces...... 163 7.3.2 Minimizing Nonspecific Adsorption...... 167 7.4 Conclusions and Future Direction ...... 178
CHAPTER 8 CONCLUSIONS AND FUTURE PROSPECTS ...... 180 8.1 Summary ...... 180 8.2 Single-Molecule Switches ...... 181 8.3 Expanding the Capabilities of Scanning Tunneling Microscopy ...... 185 8.4 Cluster Capture Surfaces ...... 186 8.5 Final Thoughts ...... 189
REFERENCES ...... 190
viii LIST OF FIGURES
Figure 1.1 Scanning tunneling microscopy image of a dodecanethiolate self- assembled monolayer and schematic of alkanethiolate self-assembly on Au{111} ...... 4
Figure 1.2 Schematic and scanning tunneling microscopy images of insertion into alkanethiolate self-assembled monolayers ...... 6
Figure 1.3 Dodecanethiolate and octanethiolate solution coadsorption and vapor annealed separated monolayer schematic and scanning tunneling microscopy images ...... 9
Figure 1.4 Energy level diagram for a one-dimensional electron tunneling junction, scanning modes for scanning tunneling microscopy and a scanning tunneling microscopy image of a Au{111} substrate ...... 11
Figure 1.5 Alternating current scanning tunneling microscopy schematic ...... 15
Figure 1.6 Picture of the bias network for the alternating current scanning tunneling microscope ...... 16
Figure 1.7 Reflectance spectra for the alternating current scanning tunneling microscope tip assembly...... 18
Figure 2.1 Molecular rectifier proposed by Aviram and Ratner ...... 24
Figure 2.2 Molecular substituent groups and molecular device components...... 25
Figure 2.3 Schematic of molecular conductance with two contacts and with one contact...... 28
Figure 2.4 Molecular device testbeds...... 30
Figure 2.5 Schematic and extracted scanning tunneling microscopy images of matrix-mediated bias controlled switching...... 33
Figure 2.6 Schematic of Au11 cluster attachment, schematic RC circuit for the junction, scanning tunneling microscopy image with a cluster attached, and scanning tunneling I/V and dI/dV spectroscopy...... 35
Figure 2.7 The Kondo effect in electromigration junctions...... 43
Figure 2.8 Schematic of the nanorod junction assembly...... 51
ix Figure 2.9 Schematics of molecular rulers processing...... 56
Figure 3.1 Schematics of proposed mechanisms for single-molecule conductance switching...... 65
Figure 3.2 Schematics of molecules used to test hypothesized mechanisms for conductance switching, extracted scanning tunneling microscopy images of conductance switching, and associated apparent height changes. . . . . 68
Figure 3.3 Scanning tunneling microscopy images displaying the tracking and extraction of single-molecule switches...... 72
Figure 3.4 Scanning tunneling microscopy images of thiolates inserted as thiols and thiolates inserted as disulfides, with linescans of the inserted molecules topographic height...... 78
Figure 3.5 Scanning tunneling microscopy images of non-covalently tethered Au nanoparticles, which are swept across the surface by the scanning tunneling microscope tip...... 81
Figure 3.6 Scanning tunneling microscopy images of covalently bound Au11 clusters with intermittent appearance...... 83
Figure 3.7 Schematic of nitro-functionalized oligo(phenylene-ethynylene) and caltrop molecules both normal to the surface and at a 30° tilt...... 88
Figure 4.1 Schematics of thiol and disulfide oligo(phenylene-ethynylene) molecules used to study inserted molecule motion...... 95
Figure 4.2 Schematic illustrating the apparent height calculation at a Au{111} substrate step edge...... 97
Figure 4.3 A single extracted molecule from a series of scanning tunneling microscopy images with their apparent height vs. time and the occurrences vs. apparent height for all inserted molecules in the series of images...... 100
Figure 4.4 Scanning tunneling microscopy images of switch molecules inserted into different chain length self-assembled monolayer domains and corresponding occurrences vs. apparent heights for all of the molecules in each image...... 103
Figure 4.5 Apparent height vs. time plot with corresponding extracted frames for a switch molecule at a substrate step edge...... 106
x Figure 4.6 Scanning tunneling microscopy images and extracted images of switch molecules at a substrate domain boundary and a substrate step edge...... 108
Figure 5.1 Schematic of a molecular switch exhibiting switching and real-time height vs. time data for a single molecular switch...... 113
Figure 5.2 Scanning tunneling microscopy image of inserted molecules over which height vs. time data were recorded and data displaying the correction applied for tip drift during spectral acquisition ...... 116
Figure 5.3 Height vs. time data displaying the dependence of the spectra on the location of the switch in the host self-assembled monolayer matrix ...... 118
Figure 5.4 Scanning tunneling microscopy image of inserted molecules over which height vs. time data were recorded. Data displaying the molecule exhibits both switching and motion at the substrate step and an occurrence vs. height plot for the heights represented in the data...... 121
Figure 5.5 Real-time height vs. time data...... 123
Figure 6.1 Scanning tunneling microscopy images of 3-mercapto-N- nonylpropionamide assembled from solutions held at room temperature and 80 °Celsius...... 130
Figure 6.2 Schematics of molecules used for the polarizability studies and the alternating current scanning tunneling microscopy junction...... 134
Figure 6.3 Microwave spectra recorded over self-assembled monolayers...... 139
Figure 6.4 Schematics, scanning tunneling microscopy images and microwave difference frequency images of self-assembled monolayers...... 141
Figure 6.5 Scanning tunneling microscopy image, microwave difference frequency image and extracted switch molecules...... 143
Figure 6.6 Ratio of switch molecule microwave difference frequency to dodecanethiolate microwave difference frequency vs. time plots and occurrence vs. ratio summary...... 146
Figure 6.7 Extracted topography and microwave difference frequency frames for several different switch molecules ...... 148
xi Figure 6.8 Scanning tunneling microscopy images and microwave difference frequency images of a molecule that was switched from ON to OFF using the scanning tunneling microscope electric field...... 152
Figure 6.9 Scanning tunneling microscopy images and microwave difference frequency images of a molecule that switched from OFF to ON...... 154
Figure 7.1 Schematic of capture surface assembly...... 159
Figure 7.2 Imaging time-of-flight secondary ion mass spectrometry image for the electrostatic capture surface, mass selected for the 138Ba+ ion and corresponding counts vs.distance linescan...... 165
Figure 7.3 X-ray photoelectron spectroscopy survey spectrum for the electrostatic capture surface, and high-resolution X-ray photoelectron spectrum for the barium 3d5/2 peak...... 166
Figure 7.4 Imaging time-of-flight secondary ion mass spectrometry image for the electrostatic capture surface minimizing nonspecific adsorption, mass selected for the 138Ba+ ion and corresponding counts vs. distance linescan...... 170
Figure 7.5 X-ray photoelectron spectroscopy survey spectrum for the electrostatic capture surface minimizing nonspecific adsorption and high-resolution X-ray photoelectron spectrum for the phosphorus 2p peak...... 172
Figure 7.6 High-resolution X-ray photoelectron spectra for the barium 3d5/2 peak and the oxygen 1s peak for the electrostatic capture surfaces. . . . .174
Figure 7.7 High-resolution X-ray photoelectron spectra for the carbon 1s peak for the electrostatic capture surfaces ...... 176
xii
LIST OF TABLES
Table 5.1 Physical and measured heights of molecular features on a Au{111} substrate with an adsorbed self-assembled monolayer...... 120
Table 6.1 Molecular polarizabilities along the molecular axis with excitation along the same axis...... 136
Table 6.2 Molecular polarizabilities calculated by Yeganeh and Ratner...... 137
Table 7.1 Atomic concentrations and ratios measured from high-resolution X-ray photoelectron spectroscopy recorded for the electrostatic capture surfaces...... 168
Table 7.2 Atomic concentrations and ratios measured from high-resolution X-ray photoelectron spectroscopy recorded for the electrostatic Capture surfaces minimizing nonspecific adsorption ...... 177 xiii LIST OF ABBREVIATIONS
Acronyms and Abbreviations Molecules
1ATC9 3-Mercapto-N-nonylpropionamide Au11 Undecagold cluster BaTO Barium tetrahydrofurfuryl oxide C8 Octanethiolate C10 Decanethiolate C12 Dodecanethiolate MUDA 11-Mercaptoundecanoic acid MUPA 11-Mercaptoundecylphosphoric acid OPE Oligo(phenylene-ethynylene) OPV Oligo(phenylene-vinylene) PDMS Polydimethylsiloxane PMMA Poly(methyl-methacrylate) PS Polystyrene THF Tetrahydrofuran
Acronyms and Abbreviations Characterization Tools
CP-AFM Conductive probe atomic force microscopy MCB Mechanically controlled break (junction) STM Scanning tunneling microscopy ToF-SIMS Time-of-flight secondary ion mass spectrometry XPS X-ray photoelectron spectroscopy
Acronyms and Abbreviations Scanning Tunneling Microscopy
dI/dV Differential conductance IETS Inelastic tunneling spectroscopy Itunnel Tunneling current set point I/V Current versus voltage I vs. t Current versus time spectroscopy Vsample Applied sample bias z vs. t Height versus time spectroscopy
Acronyms and Abbreviations Other Acronyms
AC Alternating current DC Direct current DFT Density functional theory Eb Binding energy EF Fermi level
xiv
Acronyms and Abbreviations Other Acronyms Continued
FET Field effect transistor FBL Feedback loop HOMO Highest occupied molecular orbital LDOS Local density of states LIA Lock-in amplifier LOR Lift-off resist LUMO Lowest unoccupied molecular orbital MDF Microwave difference frequency NDR Negative differential resistance OWL On-wire lithography PR Photoresist SAM Self-assembled monolayer T Temperature μCIP Microcontact insertion printing ω Frequency
xv ACKNOWLEDGMENTS
I was fortunate to meet Paul Weiss on his visit to Pacific University in 2001 as I was starting to explore the possibility of graduate school. He had been invited by my undergraduate advisor, Kevin Johnson, to share the exciting research being performed in his group. I thank Kevin Johnson for introducing me both to the world of scanning tunneling microscopy and to an expert in this field. I also thank Kevin for his constant encouragement throughout my graduate school career.
Many heartfelt thanks to my graduate advisor, Paul Weiss. Thank you so much for visiting Pacific and showing me what wonderful opportunities would be available for me in your group. Thank you for welcoming me into your group, and for all of the research opportunities that have been afforded to me. I am so grateful for the encouragement and opportunities you have given me to develop and grow as a scientist.
I thank my committee members Tom Mallouk, David Allara, Sharon Hammes-
Schiffer, and Mary Jane Irwin. I appreciate the time and effort you have put into being on my committee. I gratefully acknowledge support from the Air Force Office of Scientific
Research, the Army Research Office, the Defense Advanced Research Projects Agency, the National Science Foundation, the National Institute of Standards & Technology, and the Office of Naval Research. Chapters 3, 4 and 5 have been reproduced in part with permission. Copyrights 2005, 2006 and 2007, the American Chemical Society.
Beth Anderson, you were there to recruit me. I will always be grateful for your friendship throughout graduate school. I have looked to you as a source of inspiration, strength and fellowship. Arrelaine Dameron and Penelope Donhauser, thank you for xvi always making me laugh in lab. Zach Donhauser, Brent Mantooth, Kevin Kelly and
Lloyd Bumm, thank you for all of your help on the molecular switches project and all of
the discussions, even after each of you had moved on from Penn State. Thanks to
everyone who was here when I arrived and welcomed me into the group: Terry
D’Onofrio, Dan Fuchs, Jason Monnell, Julia Heetderks, Sanjini Nanyakkara, and Rachel
Smith, and thanks to the postdocs who have been here and have given me scientific
advice: Anne Counterman, Luis Fernandez, Susan Gillmor, Patrick Han, Dongbo Li,
Charlie Sykes, Tao Ye, and Peng-Peng Zhang.
Thank you to the people who have experienced graduate school with me. To
Hector Saavedra who entered graduate school with me- thanks for sharing this
experience! To Adam Kurland, you are a great friend to talk with and will be missed! To
T. J. Mullen, it has been a pleasure to go through the thesis writing process with you! To
the rest of the Weiss group who have come after me, I have enjoyed all of my interactions with you, and I wish you the best on your future endeavors, Meaghan Blake, Sarawut
Cheunkar, Daniel Dewey, Nate Hohman, Moonhee Kim, Ajeet Kumar, Bala Pathem,
Mitch Shuster, Amit Vaish and Rong Zhang.
My parents, Carol and David Moore, deserve much praise. I could not have
completed graduate school without the constant encouragement that has been provided
from them. Thank you for always encouraging me and allowing me to follow my dreams.
To my husband, Nathan Shrefler, you are the best reason for having moved to
Pennsylvania! I love to laugh with you, and you seem to get all my jokes that no one else
can follow. You are so encouraging; you give me strength when I need it most. You are a
wonderful shoulder to cry on when things don’t go as expected. Ich liebe dich! 1
Chapter 1
CREATING AND IMAGING MOLECULAR ASSEMBLIES
1.1 Introduction
Gaining an understanding of how to create and to control nanoscale device
assemblies, has been of increasing interest in recent years [1,2]. Creating devices at the
nanoscale has largely been driven by the prediction that the number of transistors on a
processor chip doubles about every two years [3]. This trend, predicted in 1965 by Intel
co-founder Gordon Moore and commonly known as Moore’s law, has directed processor
technology to reach ever smaller scales. However, as semiconductor processor devices
decrease in size, statistical variations become unavoidable due to small numbers of
dopant atoms in small volumes of semiconductors [4-6]; thus, increasing the drive
towards single molecules acting as device components.
Richard Feynman, in his 1959 lecture “There’s plenty of room at the bottom” [7],
was one of the first to propose computing on the molecular scale; he envisioned wires to
be tens of atoms in diameter and circuits to be a few thousand Ångstroms across [7].
Fifteen years later, Aviram and Ratner proposed that a single organic molecule placed
between two electrodes could act as an electronic rectifier [8]. While single-molecular 2 devices are not yet produced as a commercial technology, other devices based on organic materials, such as organic light-emitting diodes [9-11], have found commercial applications and success. Continued progress and interest in molecular devices has made the field a possible alternative to conventional solid-state semiconductor processor technology.
This dissertation explores the use of molecular design and novel instrumentation to understand conductance in single molecules. We will discuss how we have gained an understanding of the mechanism of conductance switching occurring for fully conjugated organic molecules, how motions occur at the single-molecule level, and how to assemble surfaces to capture molecules with specific chemical properties. The remainder of this chapter will introduce self-assembly techniques and discuss scanning tunneling microscopy (STM) instrumentation including our development of alternating current
STM (ACSTM).
1.2 Self-Assembled Monolayers
Alkanethiolate self-assembled monolayers (SAMs) are used in experiments detailed throughout this dissertation. Self-assembled monolayers are molecular structures that organize spontaneously without external control [12,13]. The most widely characterized SAMs are those of n-alkanethiolates assembled on Au{111} substrates, which are formed from solutions of alkanethiols or disulfides in which the sulfur head group chemisorbs to the gold substrate [14,15]. Alkanethiolate SAMs are well ordered and highly stable due to the strong S-Au bond (~1.9 eV) and the attractive van der Waals 3 forces between adjacent alkyl chains in a predominately all trans conformation
(~0.4-0.8 eV, depending on length) [13,16,17]. Alkanethiolate SAMs have been widely used due to their ease of preparation and high structural order [1,2,12,13]. Since SAMs can be synthesized to contain a variety of functionalities and be a range of thicknesses, they have been used for many types of molecular assemblies, including the molecular devices discussed here.
Figure 1.1A shows an example of a typical STM image of a dodecanethiolate
(C12) SAM. This imaged monolayer includes examples of substrate defects, including substrate step edges (red arrows), which are boundaries between terraces of the gold substrate that differ in height by one Au-atom, and substrate vacancy islands (teal arrows), which are one Au-atom deep substrate defects formed through Au{111} substrate reconstruction. This reconstruction occurs during solution deposition by the ejection of Au atoms during molecular exchange between the solution and the substrate
[18]. Also present in this image are boundaries between the SAM domains (green arrows). These boundaries are formed where domains of the SAMs converge [16,19-21].
The SAMs assemble via a Au-S bond with a ~30° tilt from surface normal [13,17,22] to maximize the van der Waals interactions between the alkyl chains (schematically,
Figure 1.1B). Alkanethiolate SAMs form a close-packed (√3 × √3)R30° lattice with respect to the underlying Au{111} lattice and related superstructures, shown schematically in Figure 1.1C. The lattice spacing for the alkanethiolate SAMs is 4.99 Å
[13,22].
4
AB
11 Substrate Step Edges Å ngstroms
Domain Boundaries
Substrate C Vacancy 0 Sites 50 Å
Figure 1.1: A. Scanning tunneling microscopy image of a C12 SAM assembled on a Au{111} substrate. Defects within the monolayer are inherent to the assembly and include substrate step edges (red arrows), substrate vacancy islands (teal arrows) and boundaries between of the alkanethiolate domains (green arrows). B. Schematic of the packing of a C12 monolayer showing the tilt of the molecules. Carbon is represented by black circles, H by white, S by purple and Au by yellow. The alkanethiolate molecules are bound to the surface through a S-Au bond. C. The SAMs pack on the surface in a (√3 × √3)R30° lattice (red) relative to the underlying Au lattice (black).
5 1.2.1 Insertion into Self-Assembled Monolayers
We employ insertion strategies to isolate single molecules within host SAM matrices. We are able to measure the properties of more conductive molecules when they are isolated within an insulating host [23], and we can understand how molecules act individually and compare these single-molecule properties to measurements of molecules assembled as bundles. Thus, we gain an understanding into properties that are intrinsic to single molecules, vs. to bundles of molecules or to the creation of the assemblies. The two insertion strategies used in the work described in this dissertation are shown schematically in Figure 1.2. Specific molecules used for insertion and their concentrations will be given within each chapter. The method used most frequently was insertion from solution. This insertion is performed by first creating alkanethiolate matrices where the ordering of the matrix is dependent on the time allowed for assembly
(schematically, Figure 1.2A). Longer assembly times lead to higher degrees of order while shorter assembly times result in more defect sites, and enable a higher degree of insertion [24,25]. Next, we placed the preformed SAMs (STM image, Figure 1.2B) in a solution containing the molecules to be inserted (schematically, Figure 1.2C, top).
Insertion occurs into the defect sites in the host SAM (discussed above), as shown schematically Figure 1.2D. When the inserted molecules are imaged with STM, they may appear as protrusions from the host SAM, as shown in Figure 1.2E.
We have also used microcontact insertion printing (μCIP) to isolate molecules within a host SAM [26]. This method uses a polydimethylsiloxane (PDMS) polymeric
6
A SAM C solution formation insertion Au{111}
Au{111} D inserted SAM
Au{111} Au{111} contact insertion PDMS
Au{111} Au{111} B SAM E inserted SAM 6 Ångstroms
100 Å 100 Å 0 Figure 1.2: A. Schematic of the SAM solution assembly. Gold substrates are placed into ethanolic solutions containing alkanethiol molecules resulting in an alkanethiolate SAM deposited on the substrates. B. Scanning tunneling microscopy image of an alkanethiolate SAM, Vsample = -1.0 V, Itunnel = 1 pA. C. Schematic of insertion with the upper arrow indicating solution insertion, which is performed by placing the Au{111} substrate with the preformed SAM into a solution containing the molecules to be inserted. The lower arrow indicates a schematic of μCIP where a PDMS stamp is inked with the molecules to be inserted. This stamp is brought into conformal contact with the Au{111} substrate containing the preformed SAM. Molecules are able to insert only where the stamp contacts the substrate. D. Both solution insertion and microcontact insertion result in molecules being inserted into the host SAM at defect sites. Inserted molecules may appear as protrusions from the host SAM. E. Scanning tunneling microscopy image of an alkanethiolate SAM with molecules inserted at defect sites. In this case, inserted molecules appear as protrusions from the host SAM, Vsample = -1.0 V, Itunnel = 1 pA. 7 stamp to insert molecules into a host SAM matrix. The PDMS stamp is “inked” with a solution containing the molecules to be inserted. The stamp is then blown dry with a stream of nitrogen to remove the solvent molecules. The PDMS stamp is brought into conformal contact with a substrate containing a preformed SAM (Figure 1.2C, bottom).
Molecules from the PDMS stamp are able to insert only into areas of the SAM where the
PDMS stamp is in contact with the substrate. The insertion depends on the concentration of the ink molecules, the contact time between the PDMS stamp and the substrate, and the quality of the host matrix [26].
1.2.2 Mixed Alkanethiolate Monolayers
Monolayers can be formed from mixtures of SAM species. The ordering of the resulting matrix after solution deposition of mixed species can have varied results depending on the nature of the molecules being assembled. For example, using mixed monolayers formed from two types of alkanethiolate chains with similar lengths but different pendant functionality form phase-segregated SAMs separated as domains of common pendant functionality [27,28]. After solution deposition of the mixed SAMs, small domains of one pendant functionality coalesce into larger domains. The observed coalescence of different molecular species through the monolayer occurs mainly at defects, step edges, and other areas where surface coverage is lower because motion in a tightly packed monolayer is available only through collective hindered motions [27,28].
When alkanethiolate monolayers with C12 and octanethiolate (C8) are assembled from solution, the different chain length species intermix since both species are able to 8 form favorable van der Waals interactions (Figure 1.3A). To create separated monolayers of C12 and C8 species, we used vapor-phase annealing [29]. For this process, we created a C12 SAM. This preformed SAM was placed into a 1 mL v-vial (Wheaton, Millville,
NJ) above 10-20 μL of neat octanethiol. The vapor-annealed samples were held at 80 °C for 1-2 hr, resulting in vapor deposition of the second SAM species at defect sites in the original monolayer, creating separated chain lengths shown schematically and in the
STM image in Figure 1.3B. Our group and others have also created separated SAMs through selective electrochemical desorption [30,31], internal functionality [32,33] and microdisplacement printing [34].
1.3 Scanning Tunneling Microscopy
The drive of analytical chemistry to measure compositions, morphologies, and concentrations of molecules from the bulk to the atomic scale has led to the development of techniques to characterize molecular and atomic arrangements on surfaces. Scanning tunneling microscopy, invented by Binnig and Rohrer in 1981 [35], for which they won the Nobel prize in 1986, is able to determine atomic arrangements by rastering an atomically sharp tip across a flat, conducting or semiconducting surface, thus rendering real-space images. This ability has made STM a widely used tool for surface characterization. Many books [36-38] and reviews [39-41] have been written on STM development, theory, and use. Here, we will introduce the theory of STM imaging and discuss how we have coupled microwave frequencies into the STM tunneling junction. 9
A Coadsorption B Vapor Anneling
4 4 Ångstroms Ångstroms
50 Å 0 50 Å 0 Figure 1.3: Octanethiolate and C12 molecules coassembled on Au{111} substrates. A. Schematic and STM image of codeposition from solution of C8 and C12, Vsample = -1.0 V, Itunnel = 1.0 pA. The different chain length alkanethiolate species intermix. B. Schematic and STM image of a separated alkanethiolate monolayer formed by vapor-phase annealing of octanethiol into a preformed C12 SAM, Vsample = -1.0 V, Itunnel = 1.0 pA.
10 1.3.1 Scanning Tunneling Microscopy Operation
An overview of STM operation is detailed schematically in Figure 1.4 presenting a one-dimensional (1-D) metal-vacuum-metal STM tunnel junction. An atomically sharp conducting metal tip, typically tungsten or a platinum/iridium alloy, is brought within short distances (~3-10 Å) of a conducting or semiconducting sample using piezoelectric ceramic materials for probe-tip placement and motion. An applied bias voltage (V) offsets the Fermi levels (EF) of the tip and sample; the polarity of the bias voltage
determines the direction of the electron flow, i.e., if EF,tip is greater than EF,sample (as drawn
in Figure 1.4A), this shift allows electrons to tunnel from the Fermi level of the tip into
unoccupied states within the sample. By changing the bias polarity, electron flow occurs
in the opposite direction (from sample to tip).
In classical mechanics, an electron traveling between the tip and sample would
require an energy greater than the work function of the tip/sample material to overcome
the barrier between them. However, quantum mechanically, electrons are able to tunnel
through the barrier. The state of an electron in a 1-D junction with a rectangular barrier
is:
ψ(z) = ψ(0) e-κz, 1.1
where
2m(V − E) κ = , 1.2 h
ψ is the electronic wavefunction, z is the tip-sample separation, m is the electron mass, V
11
A B C tip sample z
Φ D ~tunneling ~ electron E F,tip V=EF,tip - E F,sample
Energy E F,sample
LDOS z tip LDOSsample 50 Å
Figure 1.4: A. An energy level diagram for a 1-D electron tunneling junction. The EF of the tip and sample are offset by the applied bias voltage, V. The resultant current is exponentially proportional to the distance between the sample and the tip (z). LDOS = local density of states, Φ = work function of the metal. B. Schematic of a STM tip rastering across a metal surface in constant current mode, where the tip is extended and retracted maintaining a constant tunneling current between tip and sample. C. Schematic of a STM tip rastering across a metal surface in constant height mode, maintaining a constant tip-sample separation, measuring current. D. Atomically resolved topographic STM image operating in constant current mode of a Au{111} surface with herringbone reconstruction, Vsample = -0.05 V; Itunnel = 200 pA.
12 is the potential in the barrier, E is the energy of the tunneling electron, and ħ is Planck’s constant. Through the Born interpretation of the wavefunction, the square of the wavefunction is proportional to the probability distribution of the electron, and thus, the probability of tunneling yielding a tunneling current I:
2 -2κz z|Ψ(z)| ∝ e ∝ I(z). 1.3
For small biases, the quantity (V-E) can be approximated as the work function (Φ) of the metal (~5 eV) [42]; thus, the tunneling current decreases about an order of magnitude for every 1 Å change in z. The molecular and atomic resolution of STM is enabled through this exponential current decay; therefore, the tunneling current is extremely localized, and STM is sensitive to both lateral and vertical changes in topography (often tenths of Ångstroms or better) [43,44].
The microscope tip is rastered across a surface in one of two modes: constant current (Figure 1.4B) or constant height (Figure 1.4C). In constant-current mode (most common) the feedback loop (FBL) maintains a set current by adjusting the tip-sample separation. The recorded topography is dependent on both the geometric structure and the local density of states (LDOS) of the tip and sample [43,45,46]. Constant-current mode is able to measure surface features with higher precision (atomic resolution and better) as described above, but is generally scanned at a slower rate due to the mechanical manipulation of the scanner in the z direction, controlled through the FBL. Figure 1.4D is an atomic-resolution image of a Au{111} herringbone-reconstructed substrate obtained in constant-current mode. In constant-height mode, the tip is maintained at a set distance 13 from the sample during scanning, recording the current and enabling much faster measurements; constant-height mode is most useful only for relatively smooth surfaces.
Scanning tunneling microscopy can also be used to measure spectroscopic information for single molecules through scanning tunneling spectroscopy. To perform spectroscopy using the STM tunneling junction, the tunneling gap is held at a constant set point by interrupting the FBL while a voltage ramp is applied and the current is measured as a function of the applied voltage. Current vs. voltage (I/V) spectra give the conductivity of the sample or adsorbate measured. The first and second derivatives of the conductivity can be measured using lock-in amplifier (LIA) detection resulting in differential conductance (dI/dV), which approximates the LDOS when normalized by the conductance (I/V), and inelastic scanning tunneling spectroscopy (IETS, d2I/dV2), peaks of which are assigned as vibrational mode of adsorbed molecules [47-50].
1.4 Microwave-Frequency Alternating Current Scanning Tunneling Microscopy
Above, we have described STM, which is a well-established surface probe that is routinely capable of spatial resolution on the atomic scale. We are working towards making it an even more powerful technique by incorporating the imaging capabilities of
STM with the capacity to record the response of microwave frequencies applied to the tunneling junction. We have custom built an AC scanning tunneling microscope, which is tunable over a wide frequency range (0.5–20 GHz). Lower frequencies can be applied in our system by changing our mixing hardware (described below). Higher frequencies could be reached with different sources; however, care would need to be taken to bring 14 higher frequencies into the STM tunneling junction due to the loss caused by impedance mismatch. Alternating current STM has been in development in our group for several years [23,51-54].
Our ACSTM setup is shown schematically in Figure 1.5. Signal detection is achieved through the use of difference frequency mixing where two microwave frequencies are generated from tunable waveform generators (Hewlett Packard, 83623B).
The applied frequencies are offset by a small difference frequency (5 kHz), which is the frequency at which we detect our signal. The microwave signals are mixed (Hewlett
Packard, 87302C) and a small signal (~5%) is split off by a directional coupler (Hewlett
Packard, 87300C) for reference before the AC signals are combined with the STM direct current (DC) bias voltage. The reference portion of the signal is sent through a detector diode (Narda, 4503A), where the nonlinear diode produces a well-defined signal at the difference frequency of the two applied microwave frequencies. This lower frequency signal was amplified (Frequency Devices, Inc., ASC-50) and sent to a LIA (Stanford
Research Systems, SR850) for detection. The microwave signals were combined with the
DC bias voltage using a bias tee (Hewlett Packard, 11612A) and were sent to the STM tip. A picture of this bias network is shown in Figure 1.6. Using the combination of the
AC and DC bias inputs allowed us to monitor the topography using the constant current imaging described above, and to map the difference frequency signal spatially simultaneously. This allows us to correlate topographic features with the spectroscopic signature revealed by the difference frequency, since the molecules within the tunneling junction respond to the microwave frequencies applied. The nonlinearity of the tunneling
15
source 1 ω 1 power combiner e¯ e¯ directional external coupler reference bias tee source 2 ω detector 2 diode DC bias amp tip bias
STM control electronics Δω reference
tunnel Phase & lock -in current magnitude amplifier amp
computer Δω high pass filtered tunnel current
Figure 1.5: Schematic of the difference frequency detection scheme. Two frequencies, ω1 and ω2 are generated, synchronized with an external source, and combined. A small portion (~5%) of the signal is passed through a detector diode, resulting in a difference frequency (Δω) signal, which is used as the reference signal for our LIA. The remainder of the mixed signal is combined with the DC bias voltage using a bias tee and sent to the ACSTM tip. The nonlinearity of the STM tunneling junction results in mixing, including signals at the difference frequency. This is carried with the tunneling current and is able to pass through our current preamplifier. The difference frequency signal is extracted from the tunneling current and is compared to the reference signal using the LIA.
16
DC Bias
Detector diode narda 4503A
Power Combiner Signal to HP87302C Scanning Signal from Tunneling Function Generators Directional coupler Bias tee Microscope AC Bias HP7300C HP11612A Figure 1.6: The bias network setup for ACSTM. Two tunable waveform generators create the microwave frequencies (Hewlett Packard, 83623B). These signals are mixed (Hewlett Packard, 87302C) and a small reference signal is split off by a directional coupler (Hewlett Packard, 87300C) for reference. The reference signal is sent through a detector diode (Narda, 4503A) to create the signals at the mixed frequencies, including the difference frequency used for reference. The AC signal is combined with the STM DC bias voltage using a bias tee (Hewlett Packard, 11612A) and sent to the STM tip.
17 junction, similar to the detector diode, mixes frequencies, thereby heterodyning the microwave frequencies to the lower difference frequency. The difference frequency signal was measured as a modulation of the tunneling current and was able to pass through the current preamplifier (bandwidth ~30 kHz) and was compared to the reference frequency signal using a LIA.
Microwave frequencies are susceptible to standing waves caused by reflections and impedance mismatches as the bias is carried to the STM tip-sample junction. Since we are unable to run microwave compatible cables directly to the STM tip, we have developed a tip assembly to shield the microwaves as they are brought into the scanning tunneling microscope. Shown schematically in Figure 1.7, the microwave cables are connected to a UT-34 coaxial cable (Microstock, Inc., West Point, PA) using an SMA connector (Microstock, Inc.). This leads to impedance mismatches that appear in our tip assembly reflection spectra (Figure 1.7) at ~17 GHz. The UT-34 cable is connected to a stainless steel syringe tube (Small Parts, Inc., Miramar, FL), which holds the STM tip.
The tip holder syringe tube is shielded by a UT-70 coaxial cable (Microstock, Inc.) that has had the inner conductor removed. This junction leads to further impedance mismatch and spectral reflections at ~12 GHz. The reflections are measured by outputting the microwave frequency signals, and comparing the output to what is reflected by the assembly. These reflections are inherent to our system, and thus when we analyze microwave spectra, we compare spectra for differences in magnitude rather than specific peaks. Two different assemblies were tested and they are offset for clarity in Figure 1.7.
The spectrum for the junction shown in red show less spectral loss than the spectrum
18
S11 Reflectance
0
-5 UT-70
-10
-15
-20
-25 UT-34 S11 Reflectance (dB) -30
-35 Blue spectra offset by -8 dB -40 0 5 10 15 20 25 SMA Frequency (GHz)
Figure 1.7: Reflection spectra measured for two ACSTM tip assemblies. The spectrum shown in blue is offset by -8 dB. Standing waves are present in our image due to our coupling of the microwave frequencies into the tunneling junction. Impedance mismatch between the SMA connector and the UT-34 coaxial cable as well as between the UT-34 coaxial cable and the UT-70 coaxial cable shielding the tip holder cause reflections in the spectra.
19 shown in blue. These spectral loss and higher reflections lead to less signal and therefore shows the importance of taking care in making microwave-compatible tip assemblies.
1.5 Dissertation Overview
To put the work we have done into perspective in terms of progress in understanding molecular conductance, Chapter 2 has been included to give an overview of the different techniques that are being utilized to study small bundles of molecules down to the single-molecule level. Chapter 2 also gives an overview on advances that have been made to create nanometer-scale gaps in which single molecules could be place for future device development.
The electron transport and conductance switching of oligo(phenylene-ethynylene)
(OPE) molecules has been studied extensively within the Weiss group with working originating from Bumm et al. [23] where OPE was shown to be more conductive than alkanethiolate SAMs and Donhauser et al. [24] who showed that the OPE molecules exhibited bistable conductance states when imaged with STM. This observed conductance switching has been attributed to many possible sources within the molecules themselves and the junction between the molecule and the substrate. Chapter 3 discusses how we have tested these possible switching mechanisms through molecule design of the
OPE molecules.
Chapter 4 continues our studies of the OPE molecule switches. In this chapter, we explore other interactions between the molecule and the substrate in our system. Upon analysis of our OPE bistable conductance switching, we observed some samples in 20 which three conductance states were observed. Chapter 4 details how we have attributed this third conductance state to motion of the OPE molecules within the host SAM. The
OPE molecules are able to exhibit place-exchange up and down the substrate step edges leading to different apparent heights of the inserted switch molecules. These apparent height differences are equal to those of substrate step edges.
The OPE switching has mainly been characterized on an image by image basis.
However, this means for the majority of each image, we do not capture the activity of the switch molecules. In Chapter 5, we measure the topographic height of the OPE switches on the millisecond time scale. We relate the activity on this millisecond time scale to switching and motion that we have observed on the imaging time scale (min), and we are able to observe smaller motions that we attribute to substrate rearrangements.
In Chapter 6, we couple microwave frequencies into our STM junction to gain further chemical information from our systems. Using the microwave frequencies, we are able to measure the polarizability of the molecules within our tunneling junction. We compare the relative response of different SAMs as well as our inserted OPE molecules.
We are able to correlate the polarizability of the switch molecules to their ability either to switch or to exhibit motion within the host SAM matrix.
To measure the electronic properties of molecules within a STM junction, we need the ability to tether molecules to our substrates so that they will not be displaced by the microscope tip. Chapter 7 details our progress on creating captures surfaces for nanoscale clusters being made by the Castleman group [55]. These clusters have electronic properties similar to periodic elements. Thus, we have created surfaces able to 21 capture barium ions and should have the capabilities to capture clusters with alkaline- earth properties.
Finally, Chapter 8 is a summary of this dissertation including references to publications that have come out of this work. It also includes further experimental ideas for the instruments and systems presented here.
22
Chapter 2
MOLECULAR DEVICES
2.1 Introduction
This dissertation details studies of single-molecule electronic properties as characterized primarily by STM. To put this work in perspective, I have included this chapter to provide an overview of a variety of techniques that have been utilized to study small bundles of molecules down to single molecules to characterize their electronic properties, and fabrication techniques creating nanometer-scale gaps in which these molecules may be incorporated into devices.
2.1.1 The Drive towards Single-Molecule Electronics
Smaller, faster, lighter, cheaper – these are among the driving forces leading device fabrication to reach the ultimate limit of single-molecule devices. One of the advantages of the single-molecule device approach is that synthetic chemistry can produce large quantities of molecules all containing precisely the same useful electronic and structural properties. However, as devices become smaller, the need to develop an understanding of how single molecules behave alone and in bundles, how to assemble these molecules into addressable devices, and how our fabrication techniques affect the 23 measurements performed becomes ever greater. In this chapter, we will explore select molecules proposed for devices, the testbeds utilized to characterize these molecules, and finally how these molecules are being incorporated into devices and device fabrication.
Current device production techniques utilize “top-down” approaches where features are progressively pared down from well-defined structures. Fabrication techniques are continually being developed to create ever smaller features through top- down assembly; however, there is an increasing drive towards “bottom-up” assembly where molecular and other nanoscale building blocks are assembled into larger structures.
Bottom-up assembly has been enabled through synthetic chemists’ ability to control shapes, functionality and electronic properties of molecules, and to a degree, their supramolecular assembly with high throughput. The drive towards ever smaller and faster devices requires new approaches for the fabrication and testing of single-molecule components. The feasibility of manufacturing devices on the single-molecule scale lies in our ability to understand the electronic characteristics of the molecular components and to exploit these properties into the creation of a new class of devices.
The initial proposal for a single molecule to act as a device component was developed by Aviram and Ratner, who designed a molecule that could act as a rectifier, containing both a donor and an acceptor π system separated by a tunneling bridge
(Figure 2.10H ) [8]. This proposal has led to a diverse field of research where many
molecular components have been suggested for use in devices; a few of these, displayed
in Figure 2.21H , will be referred to throughout this chapter. This, however, is not an
extensive review of all of the molecules that have been studied theoretically and 24
NC CN
H2C CH2
H 2 S S C
H2C
S S
H2C CH2 NC CN
Figure 2.1: Schematic of the initial molecular rectifier proposed by Aviram and Ratner containing donor and acceptor regions separated by a tunneling barrier [8].
25
Functional Group Name
CH3 , CH2 Alkyl COOH, COO- Carboxylic Acid, Carboxylate
CONH2 , CONH Amide
CN Nitrile
NC Isonitrile
+ NH 2 , NH3 Amine
NO2 Nitro
HS n-Alkanethiol n
Aryl
S Thiophene
Oligo(phenylene- ( ) ethynylene), OPE
Oligo(phenylene- ( ) vinylene), OPV
Figure 2.2: Molecular substituent groups and molecular device components. 26 experimentally as molecular device components. Other molecules proposed include, but are not limited to, donor-acceptor dyad molecules [56,57], nucleic acids [58-60], phthalocyanines and other organometallic species [61,62], porphyrins [63-65], rotaxanes
[66], and others that have been synthesized for devices inclusive of electrical wires [67-
69], photochromic switches [70-72], rectifiers and diodes [73-75], amplifiers [76,77], spintronics devices [78-83], logic gates [84-86] and transistors [87].
Many of the suggested molecules to be utilized as molecular components have been selected and synthesized due to their conjugation (leading to low barriers for electron transport) as well as for their select substituent groups, which may be electron donating or electron withdrawing. Central to the study and creation of single-molecule devices is how electrons traverse metal-molecule-metal interfaces. Two of the possible mechanisms for electron transport are coherent nonresonant tunneling, where the electronic energies of the molecule are far from the energy of the tunneling electrons, and coherent resonant tunneling, where the energy of the tunneling electrons are resonant with the conduction band through the molecular system [2,88-90]. For coherent nonresonant tunneling, the electron transport rate depends exponentially on the length of the molecule, whereas in coherent resonant tunneling, the electron transport rate is dominated by contact scattering [2,88].
For single-molecule devices, models for bulk electron transport no longer apply.
Instead, the electron transport is characterized in terms of the ratio of the current to the
bias voltage, or conductance (Figure 2.32H ). The conductance (G) of a single molecule
attached to two electrodes (Figure 2.3A3H ) is affected by the conductance of the contacts at
each end (G0α and G0ω) [88,91-93], and the conductance of the molecule (Gm). If we 27 combine the contact resistances into a single term (Go) and use the coherent nonresonant
tunneling regime for the conductance of the molecule, so that
Gm = exp(-βl), 2.1
where β is the decay constant for the molecule and l is the length of the molecule, then
the total conductance is given by:
G = G0 exp(-βl). 2.2
If the molecule is attached at only one electrode, as when measurements are
performed using STM, the total conductance becomes
G = G0 exp(-βl) exp(-αz), 2.3
where α is the decay constant for the gap and z is the length of the gap (Figure 2.3B)
[88,91]. Components of the molecules that affect electron transfer include: the nature of
the molecular orbitals, the alignment of the molecular orbital levels with the continuum
states of the metal, the bonding within the molecule, and bonding between the molecule
and substrate [88,94-97]. Many of the molecules chosen as candidates for molecular
electronic devices contain π-conjugation, which delocalizes the molecular orbitals,
leading to greater electron transport in these conjugated regions. If the orbitals are not
delocalized over regions of the molecules, these regions are considered nonconducting
[98]. Typically, if electrons traverse through the molecular orbitals, electrons travel
28
A G Gm molecule Electrode Electrode G G 0a 0w B G Gm molecule Electrode (STM tip) Electrode G G 0a gap
Figure 2.3: Conductance (G) for molecules A. with two contacts where Gm is the conductance of the molecule, G0α and G0ω are the contact conductance between the molecule and electrodes and B. with one contact where Ggap is the space between the molecule and the electrode. Adapted from ref. 88.
29 through the lowest unoccupied molecular orbitals (LUMOs) [96]; however, it is possible for hole transport to occur through the highest occupied molecular orbitals (HOMOs)
[95,99,100].
One of the most important factors in determining the conductance of single molecules is the interface between the molecule and the metal. Typically, the EF of the metallic contact does not align with the HOMO or LUMO levels of the molecule, thereby reducing the electron transport at the junction [97,101,102]. For thiolate-gold interactions, common for SAMs, the electronic characteristics at the junction are dominated by σ* (sigma antibonding) orbitals on the carbon-sulfur-gold junction. In this coupling, a molecule with a high degree of delocalization cannot couple well to localized
σ states due to orbital symmetry [103]. Other contact systems have been explored, including isonitriles, selenolates, and group 10 metals [95,103-105].
The need to test the electronic characteristics of single molecules or small bundles
of molecules has led to the development of a variety of testbeds (Figure 2.44H ) [2]. Each
testbed listed in Figure 2.45H will be described in detail later in this chapter. Some of the
testbeds presented here have come under much scrutiny due to the nature of the contacts,
leading to further research and greater understanding of the variety of interactions
occurring at the molecular scale. Self-assembly is utilized in the fabrication of many of
these device testbeds; for further information on self-assembly, refer to Chapter 1.
30
A STM BCSTM tip L tip L L L L AFM L tip L L L
DEFSTM tip
GHI
JKL
B Deflection Current
Figure 2.4: Testbeds used for studying molecular devices: A. scanning tunneling microscopy, B. nanoparticle junction, C. conductive probe atomic force microscopy, D. mechanical break junction, E. scanning tunneling microscopy tip break junction F. electromigration junction, G. mercury drop junction, H. nanopore junction, I. particle bridge, J. crossed-wire junction, K. nanorod junction, and L. tip-end junction. Adapted from refs. 1,2. 31 2.2 Molecules and Measurements at the Single-Molecule Scale
With single-molecule devices being the ultimate goal, it is important to analyze the properties of individual molecules rather than averaging over ensembles of molecules where single-molecule properties may be masked.
2.2.1 Scanning Tunneling Microscopy
Since the invention of STM in the early 1980’s by Binnig and Rohrer at IBM
Zurich [39], its development has allowed for extensive study of electronic properties of surfaces, individual atoms adsorbed on surfaces, and single molecules adsorbed on surfaces through real-space imaging. For further review on the theory and operation utilized for STM, refer to Chapter 1.
For measuring single-molecule devices, STM can be considered a two-contact system, where the tip acts as one electrode and the substrate acts as the second electrode.
This, however, is complicated by the fact that the tunneling barrier must be accounted for in these measurements [91]. Scanning tunneling microscopy has been utilized for the study of molecular devices to measure the electronic properties of single molecule wires and switches [24,25,106-113]. To be able to address a single molecule or small bundles of molecules, alkanethiolate SAMs were formed on Au{111} substrates. The substrate was then placed in a second solution containing OPE molecules, which inserted into defect sites of the host SAM matrix. By controlling deposition times, both the quality of the host matrix and the number of molecules inserted can be tailored. These matrices with
inserted OPE molecules have been studied with STM (Figure 2.4A6H ), where it was 32 determined that single and small bundles of OPE molecules are more conducting than alkanethiolate molecules [23]. Furthermore, though the acquisition of hundreds of images containing several to tens of molecules per image, using automated tracking during acquisition, and post-acquisitional digital extraction of the molecules [109,114], it was observed that individual OPE molecules are capable of stochastic switching and, if
properly functionalized, controlled switching of their conductance states (Figure 2.57H )
[24,25,106-109]. This change in conductance states for the OPE molecules has been
attributed to many possible mechanisms including: reduction of functional groups [96],
rotation of functional groups [115], backbone phenyl ring rotations [116], neighboring
molecule interactions [102,117], bond fluctuations [110], and changes in bond
hybridization [24,25,106,107,109,118,119]. Each of these possible mechanisms has been
tested independently through specifically synthesizing molecules that contain the ability
to test each mechanism, and thereby, it was determined that changes in hybridization at
the sulfur-gold interface are responsible for the observed changes in conductance; please
refer to Chapter 3 for further information on the switching mechanism [109], and Chapter
6 for predicting switching using the polarizability of the OPE molecules with ACSTM.
In initial studies, when OPE molecules were inserted into a host alkanethiolate
matrix, the OPE molecules exhibited stochastic conductance switching with limited
control over the switch using an applied electric field [24,25]. Further control was gained
over functionalized OPE molecules when inserted into a host SAM containing an amide
functional group buried within the monolayer (Figure 2.58H ). It was hypothesized that the
OPE molecules switch conductance states due to the electrostatic interactions between 33
Figure 2.5: A. A nitro-functionalized OPE molecule inserted into an amide-containing SAM on a Au{111} substrate. The molecule is in the ON conductance state when the sample is positively biased. Inset: STM topography image (200 Å х 200 Å, Vsample = 1 V, Itunnel = 5 pA) of the ON conductance state of an OPE molecule inserted into a SAM. B The nitro-functionalized OPE molecule switches to the OFF conductance state when the sample is negatively biased. Inset: STM topography (200 Å х 200 Å, Vsample = -1 V, Itunnel = 5 pA) image of the same molecule in the OFF conductance state. Adapted from refs. 106,107.
34 the dipole moment of the molecule and the bias polarity between the tip and substrate in
STM. When the sample is biased positively (negative polarity on the tip) the molecule aligns normal to the surface. This upright conformation is defined as the ON conductance state of the molecule, observed in STM images as the molecule protruding from the host
SAM matrix (Figure 2.5A9H , inset). Conversely, when the sample has a negative bias
(positive polarity on the tip) the molecule is in a tilted conformation where it is not
observed with STM to protrude from the host matrix. This tilted conformation is defined
as the OFF conductance state. In the example shown in Figure 2.510H , hydrogen bonding
between the amide group of the host monolayer and the nitro-substituent on the OPE
molecule helps maintain the molecule in the OFF conductance state. This dependence on
the electrostatic and matrix interactions was further tested using molecular design to select the direction and magnitude of the dipole moment of the inserted molecule and the host [106,107].
Single-molecule switches can be regulated further using molecules specifically designed to act as photo-induced switches. Azobenzene molecules tethered by an ether- alkanethiolate were synthesized to create rigid assemblies, thus suppressing excited-state
quenching from the metal substrate and limiting the stochastic conductance switching
described above. The azobenzene-functionalized molecules were isolated within C12
SAMs and imaged with STM. Reversible photo-isomerized switching between trans and
cis conformations was controlled by cycling exposures of visible and UV light [120].
A final example of utilizing STM to perform single-device measurements is in
measuring single ligand-stabilized gold nanoparticles and clusters attached to a substrate 35
AB STM tip 20 1.0
10 CR11 L Signal dI/dV Raw L L 0.5 L L L 0 L L L Current (pA) Current -10 CR22 0.0
-20 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 Voltage (V)
Figure 2.6: A. Schematic of a Au11 cluster attachment and corresponding schematic RC circuit. B. I/V and dI/dV signal for a single spectral sweep over the Au11 cluster. Inset: STM topography image of a single immobilized Au11 cluster, (Vsample = -1.5 V; Itunnel = 20 pA; 123 Å x 123 Å). Adapted from ref. 121.
36
through a dithiol molecular linker (Figure 2.4B11H ). The unique properties of nanoparticles
and clusters, stemming from electronic band structures that are neither single atom nor
bulk band structures, have contributed to a wide range of proposed uses for these
materials including exploiting their ability to suppress current flow at low bias voltages
and to accept single electrons for use as single-electron transistors [122-124]. For information on nanoparticle synthesis, see references [125-127].
When a single nanoparticle is addressed with STM, the phenomena of Coulomb
blockade, which is the increase of the differential resistance around zero bias, and
Coulomb staircase, which is quantized electron charging into the particle, are observed
[121,122,124,128]. The potential around the Coulomb blockade region where current
begins to flow is know as the ‘threshold voltage’. Coulomb blockade occurs in tunneling
junctions and is described as RC circuits in series modeling two tunneling barriers
controlling the electron transport through this junction (Figure 2.6A12H ); the first barrier is between the probe tip and the cluster, and the second barrier is between the cluster and the substrate. The energy required to charge a metallic nanoparticle with a single electron
is inversely proportional to the particle’s capacitance,
e2 Echarge = , 2.4 2C
where e = 1.602 х 10-19 C. The particle’s capacitance can be calculated as
C = 4πєє0r, 2.5 37 where є is the dielectric constant of the ligand shell, є0 is the dielectric constant of
vacuum and r is the radius of the particle. When the particle is placed in a junction, the
thickness of the tunneling junction (L) must also be accounted for [122,124,128]:
r C = 4πєє0r(1 + ). 2.6 2L
Weiss and coworkers have analyzed precise structures of isolated, ligand stabilized undecagold (Au11) clusters stabilized by triphenylphosphine ligands
immobilized on C8 SAMs via α,ω-decanedithiolate tethers [121]. Using ultra-stable
ultrahigh vacuum STM operating at cryogenic temperatures (4.2 K), they were able to
record high-resolution images, as well as dI/dV and I/V spectroscopy over the clusters
(Figure 2.6B13H ). Interestingly, for a single tethered Au11 cluster, as well as across many
clusters, significant spectral diffusion was observed, including through the Coulomb
blockade region. This study demonstrates that even in a well-defined, carefully controlled
system, the electronic structure is still complex, and thus difficult to use as input for
engineering devices and device architectures [121]. For further information on tethering
molecules and clusters with specific atom-like properties, please refer to Chapter 7.
2.2.2 Conductive Probe Atomic Force Microscopy
Conductive probe atomic force microscopy (CP-AFM) is able to address electron
transport in single and small bundles of molecules. In this instrument, a cantilever with a
sharp metal tip is brought into contact with the surface under a controlled load, typically
2 nN [93,104,129-132]. Since the tip is in contact with the sample, the area addressed is 38 estimated to be about 15 nm2 addressing around 80 molecules in a SAM [130]. In general, images are created by rastering the sample under the cantilever tip. A laser is reflected off the back of the cantilever to a detector; therefore, changes in the surface topography deflect the cantilever, causing a deflection of the laser on the detector. Since
AFM relies on the detection of forces between the tip and sample, a feedback mechanism is used to maintain a constant force on the cantilever by measuring the voltage applied to the piezoelectric transducer [36].
Similar to the STM studies, nanoparticles coupled to a Au substrate using a dithiol linker molecule inserted into an insulating SAM have been addressed by CP-AFM
(Figure 2.4C14H ) [131-135]. One advantage of using CP-AFM over STM is that the measurement can be performed without a tunneling junction. At room temperature, electron transfer is not limited to single electrons for larger particles because the thermal energy at room temperature (kBT; T ~300 K) is larger than the Coulomb blockade energy.
Lindsay and coworkers have used CP-AFM to study such nanoparticle assemblies to test the conductivity of the linker binding the nanoparticle to the Au surface rather than addressing the properties of the nanoparticle itself. When the linker molecule was an alkanedithiolate molecule, they observed integer quantized I/V characteristics caused by the number of linker molecules forming attachments between the surface and the nanoparticle [133,136]. Dithiolated OPE molecules were also tested as the linker molecules in the assembly [137]. When the OPE molecules were unfunctionalized, the
I/V characteristics were similar to those of alkanedithiolates. However, when the OPE molecules were nitro-functionalized, the I/V characteristics displayed negative 39 differential resistance (NDR) peaks (current magnitude decrease with increased voltage magnitude) [137].
In addition to measuring single to few nanoparticle linker conductivities, AFM has also been used to address the electron transport properties of small bundles of molecules, including their conductance decay constants as a function of length (β) and
contact resistance (Figure 2.4C15H ). Decay constants were determined as a function of molecular length for alkanethiols (0.94 Å-1) and for oligophenylene thiolates (0.42 Å-1).
The lower β values for oligophenylene thiolates indicates these molecules are more conductive [93]. To test the contact resistance of the molecule-metal interface, the metals and molecules used were changed between Au, Ag, Pd and Pt metals, contacted to the molecules linked by S or CN contacts [104]. Changing the metal can influence the magnitude of the contact resistance by more than two orders of magnitude, and for aliphatic thiolate molecules, contact resistance decreases with increasing metal work function, indicating that transport is through hole transport [104].
2.2.3 Mechanically Controlled Break Junctions
In addition to scanning probes, single to small bundles of molecules are
addressable by mechanically controlled break (MCB) junctions (Figure 2.4D16H ) [138]. To create a junction small enough for few to single molecules to bridge, Reed and coworkers created a MCB junction by attaching a notched Au wire to a flexible substrate in a tetrahydrofuran (THF) solution of the molecule desired for study. The substrate was then bent using a piezoelectric actuator until the Au wire fractured at the notch. The distance 40 of the gap was adjusted using the piezoelectric actuator. After fracturing, molecules were deposited into the gap on each broken end of the wire, and the THF solution was evaporated. The ends of the Au wire were brought closer together until a single or few molecules bridged the junction [139-143]. One significant disadvantage of this technique is that it is “blind” – there is no means to look into the gap [144].
Initial measurements with the MCB junction system studied benzene-1,4-dithiol in which symmetric I/V data were acquired with an apparent bandgap of 0.7 V arising from the alignment of the molecular orbitals with the Fermi level of the electrodes [143].
Theoretical work analyzing molecules in these junctions have found that the conductance of a single molecule would be greater than the values measured [144]. It is more likely, rather than a single molecule bridging the gap that two or more molecules were overlapping in the junction with each molecule of the overlapped pair attached to opposite electrodes or the molecules attached with different conformations to the electrodes. Similar reduced conduction has been theoretically calculated for other systems with neighboring molecule interactions, and for molecules with varied attachments to electrodes [102,117,145].
More recent work performed by Weber and coworkers has looked at dithiolate
OPE-derived molecules assembled in the MCB junction with data recorded at low temperatures (~30 K) for reduced noise, higher stability, and enhanced spectral resolution
[73,141,142]. They were able to compare the transport of molecules with asymmetric contacts (only one side attached to a gold electrode) to the transport when both ends of the molecules were attached (symmetric contact), thus gaining information on the importance of the metal-molecule contacts in these systems [73,141,142]. 41 2.2.4 Scanning Tunneling Microscopy Tip Break Junctions
The STM tip break junction Figure 2.4E17H is a technique that creates a metal- molecule-metal junction, similar to the MCB junction, using a Au scanning tunneling microscope tip that is moved in and out of contact with a Au substrate in a solution containing molecules of interest [146-150]. The conductance of a single molecule is measured when it is connected between the substrate and tip electrodes as the tip is pulled away from the substrate. This connection results in conductance steps, and junction can be repeatedly formed thousands of times.
Xu and Tao found that, after contacting the tip to the sample, followed by pulling the tip away from the sample, the initial conductance steps correspond to the formation and breaking of a chain of Au atoms in the junction [150]. After the atomic chain is broken, a second set of sequential steps at lower conductance were observed due to attachment of a molecule within the junction. They measured the conductance of alkanethiolate molecules and found that the resistances were about an order of magnitude less than those found by Cui et al. using CP-AFM [133,150]. This difference was attributed to the resistance between the AFM probe and the gold nanoparticles used in the
CP-AFM experiment [150].
Venkataraman et al. found that amine attachment groups resulted in more reproducible conductance values using the STM tip break junction than using either thiols or isonitriles as the contacts for the molecular junctions [148]. When they studied a series of biphenyl derivatives using amide attachment groups, it was found that the conductance was dependent on the conformational twist angle of the molecule [147]. The variability 42 of the conductance in the STM tip break junctions has been attributed to geometric variations in the junctions [146]. Specifically, theoretical calculations from Ratner and coworkers have indicated that Au-Au and Au-molecule fluctuations provide the geometric freedom for a broad distribution of conductance values [145].
2.2.5 Electromigration Junctions
Another junction assembled to analyze few to single molecules is the
electromigration junction (Figure 2.4F18H ). To prepare this junction, 30 nm of SiO2 is grown on a degenerately doped Si substrate. Gold wires (<200 nm width, 200–400 nm length,
10–15 nm thick) were then fabricated using electron beam lithography onto the SiO2.
After cleaning, the assembly was placed into a solution of the molecules under study, allowing at least 24 h for self-assembly. The wires were then broken by ramping to large voltages (>0.5 V) at cryogenic temperatures until only a tunneling current was present and a single molecule remained in the junction. Typical yields for these devices are quite low ~10% [151-156].
Many different molecules have been examined using electromigration junctions, including C60 [152], Co ions bonded to polypyridyl ligands with insulating tethers [153], tethered divanadium molecules [154], and OPE molecules [155,156]. With the Co ions and divanadium molecules acting as spin impurities, the Kondo effect was observed. The
Kondo effect is the interaction between a localized spin impurity and the conduction of electrons in the system. For the Co ions and divanadium molecules, where both the spin
43
A
0.5
0 I (nA)
-0.5
-1.0 -100 -50 0 50 100 V (mV)
8 B 4 0
V (mV) V -4 -8 -0.50-0.45 -0.40 -0.35
Vg (V)
Figure 2.7: Testing for the Kondo effect in electromigration junctions using a Co ion bonded to polypyridyl ligands with insulating tethers. A. Current versus voltage plots at different gate voltages (Vg from -0.4 V (red) to -1.0 V (black) with ΔVg = -0.15 V). Upper inset: topographic AFM image of electrodes (scale bar 100 nm). Lower inset: schematic diagram of the device. B. Differential conductance plots at zero magnetic field with black representing zero conductance and white representing maximum conductance (5 nS). Reprinted by permission from Macmillan Publishers Ltd: Nature [153], copyright (2002). 44 and orbital degeneracies were controlled, electron transport occurs through well-defined charge states [153,154]. This is shown by the current-voltage plots, dependent on the gate voltage, and the differential conductance, plotted as a function of both bias voltage and gate voltage [153,154]. The I/V characteristics show that current is suppressed up to a threshold voltage, dependent on the gate voltage, after which the current displays a
stepwise increase (Figure 2.7A19H ) [153]. The differential conductance plots, dependent on both the bias voltage and gate voltage, display bias and gate voltage combination regions
of no current, and bias and gate voltage lines displaying the current steps (Figure 2.7B20H )
[153]. Since the charge states are tunable using the gate voltage, this indicates that a single molecule attached to the electrodes through a tunneling barrier is responsible for the observed device characteristics [153,154]. For an in-depth discussion of charge transport in these systems refer to [89].
Further research using electromigration junctions analyzed the temperature effects of charge transport through nitro-functionalized OPE molecules [155,156]. By acquiring
I/V spectra as a function of temperature (T), Allara and coworkers used an Arrhenius analysis (lnI vs. 1/T) for several bias voltages and found coherent temperature- independent charge transport at low temperature, and incoherent temperature-dependent hopping behavior at high temperature. Measurement of these molecules indicated that the charge transport through a given molecular junction is highly dependent on the specific molecular structure that allows conduction through the molecule [155,156]. 45 2.3 Converting to the Nanoscale
While single-molecule devices may be the ultimate limitation, there is yet no known way to fabricate a large quantity of these devices with high yield nor attachment schemes to make several individual single-molecule devices independently addressable.
Much research is focused on creating structures in the 10-100 nm range. This has led to studies of the electronic properties of molecules in bundles.
2.3.1 Mercury Drop Junction
Mercury drop junctions are assembled by forming SAMs on each of two electrodes. One electrode is a mercury drop, since Hg can form thiol-based SAMs similar to those on Au, and the other is a metal-film electrode (typically Ag, Au or Cu)
(Figure 2.4G21H ) [157]. These two electrodes are brought into contact, the metal film is held in place while a micromanipulator moves the mercury drop, and measurements are obtained for electron transport through these thin organic films forming a metal- molecule-molecule-metal junction [157,158]. With this method, electron tunneling characteristics have been studied for several SAMs, including alkanethiolates and polyphenylenes [158-162]. Unique to this measurement is the ability to measure the junction formed at the interface of the two SAMs. By changing the van der Waals interactions or by changing the bonding (covalent, hydrogen, or ionic bonds), the conductivity was observed to change by more than 4 orders of magnitude [163].
Further study using the mercury drop junctions coupled the SAM on the Hg drop to a SAM formed on a semiconductor surface [94,99]. The SAMs bound to the 46 semiconductor surface were functionalized with alkyltrichlorosilane, forming a siloxane bond to a silicon surface. Here, the silicon was p-doped and analyzed at low bias so that the transfer mechanism was through hole transport [94,99,158,160].
2.3.2 Nanopore
In this method, the nanopore junction (Figure 2.4H22H ) is created through electron- beam lithography followed by plasma etching opening up a pore in a silicon nitride substrate into which metal (Au) is evaporated (diameter ~30-50 nm) [164-166]. A SAM is deposited from solution onto the gold followed by evaporation of a top contact onto the
SAM (initially 30 Å of Ti and 800 Å of Au) at low temperature to try to avoid damage to the SAM [167]. Since Ti reacts with and decomposes organic molecules (vide infra), this binding layer was eliminated in later studies. In this junction, the conduction of
~1000 organic molecules can be studied [75,167-171].
The advent of the nanopore junction created much excitement for molecular device fabrication due to the NDR that was measured. [169]. This led to many subsequent studies to understand what characteristics of the molecules and the junction resulted in
NDR. Much debate has ensued over the origin of NDR. Since this peak was initially observed for functionalized OPE molecules but not for unfunctionalized OPE molecules, it has been proposed to be a function of the electronic charge states of the molecule dependent on the functionalization [96]. Others proposed that the fabrication of the junction causes NDR. One important step in fabricating molecular junctions is in creating an ohmic contact to the molecules. Since typically little is known about the chemical, 47 topographical and morphological character of vapor-deposited metal-molecule junctions and the metal overlayer, it is difficult to directly interpret I/V characteristics in terms of the charge transport across the bridge molecule. For example, research by Allara and coworkers has found that vacuum deposition of Ti on methyl-ester- and methyl- terminated alkanethiolate/Au SAMs and on OPE monolayers is not uniform and causes formation of degradation products such as carbides that penetrate deep into the monolayer, thereby providing complex metal-molecule junctions [172-176]. In the case of methyl-terminated SAMs, in the early stages of deposition a large fraction of impinging Ti atoms are scattered off the surface and the fraction that sticks forms clusters with accompanying degradation pits across the surface, rather than a uniform layer
[174,175]. Finally, when Au was thermally evaporated on K-modified SAMs, it was found that ionic interactions block Au penetration, thus, the Au atoms form islands at the vacuum interface [177].
2.3.3 Particle Bridge
This configuration uses a particle to bridge the gap between two metal electrodes
that are functionalized with test molecules (Figure 2.4I23H ). This testbed has been fabricated via two methods. The first method uses photolithography and electron-beam lithography to grow electrodes with a ~40-100 nm gap between them. These electrodes can be functionalized with a SAM. After the electrodes are functionalized, a nanoparticle is placed in the gap using an AC electric field across the electrodes, which traps the particle between the electrodes and closes the circuit leading to a metal-molecule- 48 (metal)nanoparticle-molecule-metal test system [178,179]. The second method creates a particle bridge using permalloy metal (Ni80Fe20) electrodes coated by electroless deposition of Au allowing for thiolate molecule attachment [180]. The particles used were silica spheres with 50 nm of Ni and 10 nm of gold evaporated onto them. After the electrodes were coated with the desired SAM, the particles were trapped with magnetic fields locally intense in the electrode gaps [180].
These particle junctions have been utilized to study alkanethiolate, OPE, and oligo(phenylene-vinylene) (OPV) molecules [179,180]. It was found that the conductance follows the trend that alkanethiolates are the least conductive, OPE molecules are more conductive than alkanethiolates, and the OPV molecules are still more conductive. These results scale similarly to results gained from crossed-wire junctions [181], nano-rod junctions [182], and scanning probe microscopy [112,113,183].
Furthermore, when the OPE molecules were studied using this junction, two NDR peaks were observed. This result is expected since the particle junction configuration of Au-
OPE-particle-OPE-Au is analogous to two resonant tunneling diodes connected in series, thereby giving rise to two NDR peaks [179].
2.3.4 Crossed-Wire Junction
The crossed-wire tunneling junction eliminates problems associated with creating two contacts, including damage to or diffusion through the self-assembled test molecules, and it does not require advanced fabrication techniques to create a nanometer-scale
junction (Figure 2.4J24H ). In this method, the first of two wires is functionalized with the 49 molecules under study. The second wire, in a crossed geometry, perpendicular to an applied magnetic field, is translated towards to the first wire. The wire spacing is controlled through a Lorentz force, the deflection of the wire perpendicular to the magnetic field controlled by a DC current [184-187].
This junction has been used to study the conductance of alkanethiolates, OPEs, and OPVs, similar to the particle bridge junctions. Here the conductance scaled as in the particle bridges where alkanethiolates were the least conductive and OPVs were the most conductive [181,185]. The OPV molecules were about three times more conductive than the OPE molecules, and it was proposed through theoretical modeling that this enhancement in conductivity results from the regular periodicity of the OPV molecules leading to smaller HOMO-LUMO gaps, thereby lowering the contact potential at the junction [181].
Using the crossed-wire junctions, Kushmerick et al. were able to show the importance of contacts in molecular electronics systems by changing the functionality of the α and ω termini of the molecules. Using OPE molecules, the α terminus was functionalized with a thioacetyl group (allowing for a thiolate coupling to the wire) and at the ω terminus, several different functional groups were tested including hydrogen, thiolate, pyridine, and nitro groups. The ω terminus functionalized with hydrogen led to the largest rectification (asymmetric I/V character) with large current onset at only positive bias. The nitro-functionalized ω terminus had the second highest rectification, followed by pyridine with little rectification, and finally, the symmetric molecule
(thiolated on both ends) showing no rectification and current onsets of ±0.5 V [184,188]. 50 The cross-wire tunneling junction has been used not only to study the conductance of molecules within the junction, but also to create electronic switching elements by forming Ag filaments within the junction, similar to filaments formed when vapor depositing metal as discussed with the nanopore junctions above [172-176]. By forming a monolayer of monothiolated unfunctionalized OPE molecules on a Ag wire in the cross-wire junction, at negative bias, silver filaments bridge the gap between the electrodes resulting in metal-metal contact, which increases the current several orders of magnitude. The filaments dissolve at positive bias, returning the switch to the lower current state [189]. Metal filament formation has been shown to be the primary cause of electrical switching in several systems [190-193].
2.3.5 Nanorod Junction
The nanorod junction is unique in that the investigated molecules are incorporated into a metallic nanowire through a templated growth regime leading to a large number of
test wires that could be utilized in future molecular devices (Figure 2.4K25H ). The nanowires have diameters on the nanometer scale yet can be grown to micrometers in length, thereby easing the connection to the outside world. The nanowires are grown in a polycarbonate (or other) membrane by first electrochemically depositing the initial metal of the wire into the pores. This metal is typically single crystal Au or Pd. The studied molecules are then self-assembled into the pores. The metal forming the top contact to the SAM is then electrolessly plated from a solution utilizing a reducing agent
51
Assemble Electroless Electroplate Dissolve Align in-wire monolayer seeding top segment template junctions
Figure 2.8: Schematic for assembling molecules in nanorod juctions. Adapted from ref. 194.
52 until the SAM is capped with a seed layer (typically Ag or Pd) for further deposition. This capping is much less likely to damage the SAM, as opposed to an
evaporated contact used in the nanopore junction. After capping the SAM, the remainder of the wire is grown (usually Ag or Pd) through electrodeposition. Finally, the membrane is dissolved and the nanowires are aligned on preformed contacts by dispersing a
nanowire suspension onto an array of contacts with an AC voltage (Figure 2.826H ) [182,195-
200]. Molecules studied in these junctions have included alkanedithiolates,
16-mercaptohexadecanoic acid (carboxylic acid terminus), OPE (unfunctionalized and nitro-functionalized) dithiolates, oligoaniline dithiolates, and OPV dithiolates
[182,195,197,198]. Interesting properties observed for these molecules with this junction include: for the nitro-functionalized OPE molecules, room temperature NDR (10% of the devices exhibited NDR with similar peak shape and position, the other 90% exhibited ohmic responses attributed to the nanorod preparation) [197], for oligoaniline, bistable switching [195], and the alkanedithiolates, OPE and OPV molecules conductivity scale similarly to the crossed-wire junctions [194].
2.3.6 Tip-End Junction
Similar to the electromigration junction, the tip-end junction is fabricated to include a gate in addition to the source and drain to analyze conducting molecules
(Figure 2.4L27H ). The tip-end junction was fabricated by pulling a quartz rod to a desired size. On the top of this tip, a 15-nm Al gate was deposited followed by a 50-nm SiO2 gate insulator. The right face of the tip was then coated with (after a Cr adhesion layer) 15 nm 53 of Au, forming the drain electrode to which wires were attached. On the drain, a SAM of the molecules to be studied was formed, after which the source electrode was deposited; first, 15 nm of Au was deposited from the left, then, Au was deposited at the tip end and stopped when tunneling conductance was detected in the range of 10-7–10-6 Ω-1. These junctions have a yield on the order of 80-90% [201].
Using this junction, the conductance of oligothiophene dithiolates containing between 2 and 4 thiophene units has been studied [201]. At low temperature, periodic conductance steps were observed that depended on the source-drain voltage; the steps could be shifted using the gate voltage. The spacing of these conductance steps has been attributed to molecular vibrations and the transport in these systems to resonant tunneling
[201].
In a device similar to the tip-end junction, fabrication was performed to be compatible with integrated systems [202,203]. Here, silicon substrates were patterned using photolithography and plasma-etching to define 2-3 μm high mesas. On these raised areas, SiO2 was deposited followed by Au deposition (both with Ti adhesion layers) and finally a second layer of SiO2. Each deposition step was done through evaporation normal to the surface; thus, the Au deposited was covered by SiO2 except at the edges of the mesas. Organic multilayers, consisting of 11-mercaptoundecanoic acid (MUDA) alternating with Cu ions from Cu(ClO4)2 were deposited on the exposed regions of the Au at the mesa steps. To create a second electrode to the multilayer, a second Au layer was deposited onto the sample 60° from normal using a shadow mask to limit the amount of
Au deposited on the top of the mesa. 54 For the multilayer devices fabricated through this method, devices containing between 3 and 7 MUDA layers were insulating with the current density decreasing with each additional MUDA layer [202]. This low current density could be increased after the voltage of the device was ramped above 6.5 V becoming “burned-in”. These “burned-in” devices displayed increased conductivity by 102 to 104 and could be used as programmable memory [203]. Using a 1 V pulse for the device “read”, a 4 V pulse for
“write”, and a 10 V pulse for “erase”, the device could be operated. The 4 V pulse allows for the current to increase into a high state whereas the 10 V pulse returned the current to the low state. It is hypothesized that the measured reversible conductivity was caused by
3-5 V pulses enabling the device to form metallic conduction paths in the film leading the high current state. The higher voltages (10 V) destroy these conduction paths returning it to a low current state. Furthermore, a NDR peak is observed if a voltage sweep was applied to the device [203].
2.4 Connecting to the Outside World
One of the greatest challenges, and a key area of research, is in interfacing the molecular device components created into hierarchical structures with addressability on the 1-100 nm range using micro-scale structures. The development of hybrid strategies where conventional lithographic techniques are combined with chemical processes moves towards this goal [204]. Presented here are a few techniques developing towards connections on the nanoscale. 55 2.4.1 Molecular Rulers
Weiss and coworkers have developed a molecular-ruler process to fabricate well- defined gaps between structures using organic multilayers as a patterning resist
(Figure 2.928H ) [204-206]. The organic multilayers are grown via self-assembly on a preformed metal electrode (“parent” structure) through alternating layers of a SAM, typically a bifunctional mercaptoalkanoic acid such as 16-mercaptohexadecanoic acid,
+ and a coordination ion, typically Cu obtained from copper(II) perchlorate (Figure 2.9A29H ).
The SAM component of the layer can be tuned using molecules of varied length to
selectively tune the length of the gap (Figure 2.9B30H ). After the formation of the multi- layers, a second metal electrode (“daughter” structure) is deposited. The multilayers, which act as a lift-off resist, are then removed leaving behind a well defined gap, defined
by the multilayers, between the parent and daughter structures (Figure 2.931H ) [204-208].
This process has been further developed to combine its ability to create precise nanometer-scale spacings with commonly used photolithographic techniques. Here, photolithography is used to define structures, while the ruler resists are used to define the spacings between structures precisely. Several methods have been developed; one is
highlighted in Figure 2.9C32H [204]. In this process, an initial parent structure is defined using photolithography followed by metal deposition. A bilayer resist stack consisting of a lift-off resist (LOR) and a photoresist (PR) were deposited onto the sample. The bilayer stack was photolithographically patterned so that part of the parent structure was exposed. On the exposed parent structure, the molecular ruler deposition was performed
56
A C Au SiO2 Photoresist
Lift-off resist
B O OH O OH O OH O OH
(H C) (H C) (H C) n-6 2 n-6 2 n-6 2 n-6(H2C)
S S S S
x+ x+ x+ x+ Cu Cu Cu Cu Multilayers O OH O OH O OH O OH
(H C) (H C) (H C) n-6 2 n-6 2 n-6 2 n-6(H2C)
S S S S
Figure 2.9: Combining photolithography and the molecular-ruler process. A. Au parent structure with multilayer assembly. The blue box highlights the molecules shown in B. B. 16-mercaptohexadecanoic acid alternating layers with copper ions. C. Combining photolithography with molecular rulers to define structures on the micron scale while defining the gap between structures on the nanometer scale. Adapted from refs. 204-207.
57 and daughter metal was evaporated. By removing the molecular rulers and the remaining resists, the resulting structure has been defined on the micron scale through photolithography, and on the nanometer scale between the structures by the molecular rulers process [204,208].
2.4.2 Directed Diblock Copolymers
A second hybrid technique combining conventional lithography with chemical patterning creates nanoscale gaps through the directed assembly of diblock copolymers
[209,210]. Using a cylindrical-phase diblock copolymer, which has an intrinsic distance between cylinder rows, a lithographically defined trench of a specific width is divided into an integer number of subunits by the polymer. The walls of the trench initiate the polymer domain assembly. Using a diblock copolymer composed of 70% polystyrene
(PS) and 30% poly(methyl-methacrylate) (PMMA). This copolymer forms hexagonal lattices of 20 nm diameter cylindrical PMMA domains contained in a matrix of PS. These domains register to lithographic structures and can be improved through annealing
[209,210].
Field effect transistors (FETs) have been assembled using the diblock copolymer technique [209]. Lithographically defined gaps with widths of 200 nm, 280 nm and
600 nm were created and the PS:PMMA diblock copolymer was assembled into each gap, creating linear polymer arrays with periods of 5, 7 and 15, respectively. The PMMA portion of the diblock copolymer was removed from the array using acetic acid, leaving a striped pattern of PS. This PS template was converted into a silicon nanowire array in 58 three steps. First, an O2 plasma etch removed enough of the PS to expose the silicon layer in the areas where the PMMA was removed. Second, a plasma etch (SF6 + O2) removed the silicon not masked by the remaining PS, leaving behind the silicon nanowire array. Finally, a second lithographic level was aligned to the trench defined source and drain contacts, creating a FET device. These devices have been tested and have been found to have current drives of ~5 μA/wire and current on/off ratios of ~105.
2.4.3 Imprint Lithography
Imprint lithography can be used to define test structures for molecular devices
[211-218]. In imprint lithography, a mold is created through lithography where electron- beam lithography is used to make features in the mold on the tens of nanometers scale.
After a mold with desired features is created, the mold can be transferred to the substrate through two different methods. In the first of these methods, the mold is brought into close proximity with a substrate. A photopolymer is introduced into the gap between the mold and substrate. The mold is then contacted to the substrate, thereby trapping the photopolymer solution in the relief of the mold. The photopolymer is polymerized after which the mold is removed [211-215]. In the second method, the mold and a substrate
(SiO2) with a thermoplastic polymer (PMMA) spin cast onto it are heated up to the glass transition temperature of the polymer. The mold is then pressed against the sample, compressing the polymer in the areas where the mold contacts the substrate. The mold is held to the substrate until the temperature drops below the glass transition temperature of the polymer. Anisotropic etching processes are then used to remove the remaining resist 59 in the compressed areas. Metal can then be deposited over the entire surface where the remaining PMMA (not contacted by the mold) acts as a lift of resist [211-218].
Imprint lithography can be used to create contacts for molecular devices as demonstrated in contacts used in electromigration junctions [216]. Using imprint lithography, gold nanowires with 20 nm width and 70 nm height can be deposited on
SiO2 substrates. The advantage here is that after a mold is created, it can be used to pattern several substrates, rather than patterning each individual substrate with electron- beam lithography, a much slower and more expensive process. A second example of a molecular device created through imprint lithography is a nanometer-scale molecular switch, where the molecules under study are sandwiched between 40 nm imprinted metal electrodes [217].
2.4.4 On-Wire Lithography
Mirkin and coworkers have developed a method for lithographically processing one-dimensional nanowires called on-wire lithography (OWL) [219,220]. These OWL gaps were created down to the 2.5-nm length scale [219,220]. The nanowires were created along similar lines to the nanorod junctions described above. A porous alumina template enables growth of nanowires within the pores using electrochemical deposition.
For the OWL junctions, Au-Ag and Au-Ni wires were grown with each material segment controlled by the charge passed during electrodeposition. Once grown, the wires were released from the template through template dissolution, cast onto a microscope slide and dried before a 50-nm layer of silica was deposited on the substrate. The substrate was 60 sonicated to release the wires, and finally the wires could be selectively, chemically etched to removed a chosen segment, i.e., for the Au-Ni wires, the Ni was removed by wet-chemical etching with concentrated nitric acid, generating Au nanowires containing gaps precisely the length of the original Ni segments.
The transport properties of the OWL wires were tested by evaporating a suspension containing the wires onto a microelectrode array fabricated by conventional lithography containing 3-μm electrodes separated by 2-μm gaps. Some of the OWL wires were able to bridge the gap between microelectrodes. The I/V curves for the nanowires bridging the gaps displayed insulating behavior [219,220]. Using dip-pen nanolithography [221], the OWL gaps were functionalized with a mixture of polyethylene oxide and self-doped polypyrrole. The I/V curves for the filled gaps displayed a linear response between -1.0 V to 1.0 V [220]. These OWL wires have shown high-yield, controllable synthesis and do not require expensive fabrication facilities [219,220].
2.5 Conclusion and Future Molecular Devices Direction
This chapter has discussed how the properties of individual molecules as well as the systems in which they are assembled play roles in their electronic behavior. The nature of the metal-molecule junction can greatly affect the overall device behavior. One of the difficulties of performing measurements on the nanometer scale is the intricacy associated with knowing the chemical nature (between molecules and contacts) and order of the molecules. Each testbed measurement presented here has intrinsic advantages and 61 disadvantages. The single-molecule measurements yield information on local structure and identity but are difficult to fabricate and to analyze, whereas the ensemble measurements provide simple rapid analysis but lose some of the information gained in testing single molecules.
One of the challenges that lies ahead is combining what research has gained in understanding of molecular conductivity and exploiting this understanding in combination with fabrication techniques, reaching the nanometer scale to create functional devices. The devices that will be used for future technology must provide resilience to repeated cycling, and have the ability to address few to individual molecules.
A main caution in creating these junctions is the required ability to understand and to control the junction between the molecules and the contacts.
62
Chapter 3 T
TESTING HYPOTHESIZED SWITCHING MECHANISMS FOR SINGLE OLIGO(PHENYLENE-ETHYNYLENE) MOLECULES
3.1T T Introduction
Chapter 2 introduced testbeds used to measure properties down to the single- molecule level, including conductance switching measured by STM. In this chapter, we look closely at this conductance switching by testing theoretically and experimentally predicted switching mechanisms. Using molecular engineering, we are able to tailor specific bonds, functionalities and interactions of OPE molecules to test each proposed mechanism through molecular design.
With the physical limitations of semiconductor microelectronic fabrication in sight, new approaches are needed to obtain ever smaller and faster devices. The ultimate limitation lies in single-molecule electron transport through controllable, connectable systems. The feasibility of these molecular devices will be determined by our ability to understand specific properties of molecules acting individually or in bundles and to exploit these properties in the creation of new classes of devices [1,2]. Molecules with full π-conjugation have been of great interest for their potential use in fabricating molecular electronic devices due to their electron delocalization leading to low barriers 63 for electron transport [92,222,223]. Conjugated molecules studied have included polythiophenes, polyphenylenes, polyanalines, OPVs and OPEs, which have shown interesting properties including conductance switching [24,25,106-110,224-226].
One of the most useful aspects of using single molecules for device fabrication is the ability to synthesize and to modify each molecule to have controllable and reproducible properties (i.e., each molecule is precisely the same) [222,227]. In addition, using single molecules eliminates the statistical variations inherent in having small numbers of dopant atoms in small volumes of semiconductors [4-6]. Fully conjugated molecules have been studied by a variety of techniques analyzing ensembles of many molecules down even to single molecules. Briefly, using nanopore junctions, bundles of functionalized OPE molecules have been shown to exhibit hysteretic conductance switching [167,171,228]. In break junctions [139,140,142,143] and crossed-wire tunneling junctions [184] few to single molecules have been analyzed for their conductance and contact dependence. Finally at the single-molecule level, we and others have used scanning probe techniques including STM to study conductance switching for
OPE molecules [24,25,106-110,224-226]. Please refer to Chapter 2 for further details on these methods.
Stochastic conductance switching is an interesting, useful and reproducible property observed by STM for single OPE and other molecules inserted into host n-alkanethiolate SAMs [24,25,229]. A molecule observed in a high conductance or ON state appears to protrude several Ångstroms from the host SAM in STM images, while a molecule in the low conductance or OFF state appears to protrude only slightly or not at all. The inserted molecules stochastically switch or in some cases can be driven between 64 conductance states and are stable within each state for time periods of milliseconds to many hours depending on the packing of the host matrix around the inserted molecule and the nature of the surrounding matrix (Chapter 2) [24,25,106,107,225]. These lifetimes are measured by acquiring several successive images over the same area and extracting the apparent height for individual molecules or by recording the time history profile over a single inserted molecule, as will be discussed in Chapter 5 [225,230,231].
The time spent in each conductance state has been shown to be strongly dependent on the rigidity of, and interactions with, the surrounding matrix [24,25,106,107]. In addition, a limited ability to control the conductance switching of these molecules has been demonstrated. We have used an applied electric field to control switching
[24,25,106,107], and we have demonstrated bias-dependent, matrix-mediated switching.
For the bias-dependent, matrix-mediated switching we used amide-containing alkanethiols in which the molecule’s dipole mediates the bias-dependent switching, and hydrogen-bonding interactions between the host matrix and the switch molecule play a role to stabilize the OFF conductance state [106,107].
We note that conductance switching imaged by STM is a process different from
NDR observed in nanopores and some other testbeds, but remains a subject of controversy [137,169,170,179,197,227,232]. Nevertheless, similar hypotheses have been suggested to describe both conductance switching and NDR. Both theoretical and experimental works have attempted to explain conductance changes through a variety of
mechanisms shown schematically in Figure 3.1X33HT ,TX including reduction of functional groups
[96], rotation of functional groups [115], backbone phenyl-ring rotations [116],
65
A. Reduction B. Rotation C. Neighbor Interactions e-
reduced state e- O O O O N N NO N N O O O O O
S S S S S S S S
D. Bond Fluctuation E. Hybridization Change
O N O N O N O O O N O O
S S S S
T
Figure 3.1T :T SeveralSS mechanisms have been used to describe conductance switching. Suggested mechanisms include A. functional-group reduction, B. ring rotation, C. concerted motions, D. bond fluctuations and E. hybridization changes. Hybridization change is boxed since it is the only mechanism consistent with all data.
66 neighboring molecule interactions [102,117], bond fluctuations [110] and changes in bond hybridization [24,25,118,119]. In the reduction scheme, Seminario et al. suggested that applying a reducing potential to nitro- and amino-functionalized molecules can reduce the molecule, thereby increasing conduction [96]. Di Ventra et al. suggested that the rotational plane of the ligands (i.e. nitro-group on functionalized OPEs), not reduction, affects molecular conductance [115]. Conversely, Brédas and coworkers, analyzed the rotation of the middle OPE phenyl ring concluding that constricting the rotation of the ring inhibits changes in conductance [116]. Further theoretical work has discussed the possibility that motion and interactions with neighboring molecules affect the conductance of the switches [102,117].
For bond fluctuations, Ramachandran et al. suggested that conductance switching is caused by bond-breaking at the S-Au interface where the molecule is in the OFF state
(the molecule or nanoparticle is not imaged by STM) when it is not covalently attached to the Au substrate [110]. We have previously suggested that a change in hybridization between the conjugated molecules and the substrate results in our observed stochastic changes in conductance [24,25]. The hybridization change can occur through surface reconstruction or a change in the alignment of the molecule with the surface. Sellers et al. suggest that S-Au interactions can bind at 180° (normal to the surface) having sp hybridization, while those that are at 104° (tilted) have sp3 hybridization; sp3 hybridization is favored for alkanethiols [118]. The barrier for this change in molecular tilt has been calculated to be 0.11 eV for alkanethiols, an energy accessible in our system
[118,119]. Bauschlicher et al. have calculated conductance changes for benzene-1,4-dithiol molecules dependent on the conformation of the molecules with 67 respect to the gold electrodes to which they are attached, and their bonding coordination
[233], while Ratner and coworkers have shown that conductance fluctuates in molecular junctions due to geometric changes of the Au-Au and Au-molecule structure [145].
Here, we show STM data for conductance switching for a variety of conjugated
molecules, shown in Figure 3.2X34HT .TX By specifically engineering the molecular design, we experimentally test each of the proposed switching mechanisms.
3.2T T Experimental Procedure
3.2.1T T Sample Preparation
Host matrices of n-alkanethiolate SAMs were prepared on commercially obtained
Au{111} on mica (Molecular Imaging, Phoenix, AZ). The Au{111} substrates were annealed using a hydrogen flame followed immediately by SAM deposition. The host
SAM was prepared by immersing the substrate in a 1 mM ethanolic solution of n-alkanethiol for 24 h. The length of the alkanethiolate chains presented here vary between 8 and 12 carbons, using only even numbers of carbons due to the more favorable van der Waals interactions and the orientation of the chain terminal-methyl group [22].
Following the insertion procedure (described below), vapor annealing (described in
Chapter 1) was sometimes performed to tighten the host matrix, and to favor insertion of single, conjugated molecules. This was done by vapor annealing with a neat n-alkanethiol (the same chain length as was used to create the original SAM) at 80 °C for
2 h in a sealed v-vial (Wheaton, Millville, NJ). Vapor annealing was used to add
68
Apparent height Molecules ONON state OFFOFF state change (Å)
A NO2
SH 3.6 ± 0.7
1 86 Å x 86 Å B
SH 3.3 ± 1.5
2 86 Å x 86 Å C 5.1 ± 1.2 SAc
3 28 Å x 28 Å D NO2 S 5.8 ± 0.7 S
O N 4 2 80 Å x 80 Å E
1.7 ± 1.3
86 Å x 86 Å
SAc 5 SAc F
AcS
Si SAc 3.4 ± 1.9
86 Å x 86 Å
6 SAc
Figure 3.2T :T Molecular structures, extracted images of molecules isolated in alkanethiolate SAMs in both ON and OFF conductance states, and the average measured differences in apparent heights between states (matrix dependent) for each type of inserted molecule. Apparent heights listed for A and B are from previously published works for these molecules inserted in C12 SAMs [25]. A. Nitro-functionalized OPE molecule [24,25,234] 1. B. Unfunctionalized OPE molecule [24,25,234] 2. C. Phenanthrene-based molecule [108,235] 3 in an C8 SAM. D. Disulfide OPE molecules 4, prepared by oxidative homocoupling of the thiol from 1 in an C8 SAM. E. Two-contact OPE molecule [236,237] 5 in a C12 SAM. F. Caltrop [238] molecule in a C12 SAM. Imaging conditions: VsampleB B = -1.0 V, ItunnelB B = 1-3 pA. 69 molecules to the matrix from the gas phase, while minimizing the exchange processes that occur during solution phase annealing [24,25,29]. Following SAM deposition, the substrates were rinsed with neat ethanol and dried under a stream of nitrogen.
To insert each conjugated molecule (Figure 3.X35HT 2TT ) into the host SAMs, 0.1 µM solutions of each conjugated molecule were prepared in anhydrous THF in a nitrogen environment due to the reactivity of the OPE molecules in air. For molecules synthesized
with a thioacetyl protecting group (indicated in Figure 3.X36HT 2TT TX with the abbreviation SAc), aqueous ammonia was added to hydrolyze the acetate, generating a thiol in situ, allowing the conjugated molecules to adsorb to the Au surface at SAM defect sites [68,239]. The
SAMs were exposed to insertion solutions for times ranging from a few seconds to 0.5 h.
In general, more time was allowed for thioacetyl tether groups than thiol tether groups and more time was allowed for larger, branched molecules to insert. No significant differences were observed with STM between the protected and unprotected thiols; some deprotection is expected from direct surface reaction. The substrates were dried under a stream of nitrogen after insertion and stored at room temperature in a desiccator until imaging.
Insertion occurs at defect sites such as domain boundaries, step edges and substrate vacancy islands in the host matrix, as described in Chapter 1. Molecules were shown to be isolated in the host SAM using STM images recorded with molecular resolution of the surrounding matrix; the inserted molecules appear in the SAM each with the same shape, indicative of a feature smaller than the tip being imaged [229]. The STM topography images recorded for these molecules are convolutions of the geometries and conductivities of the molecules; therefore, because these molecules are both more 70 conductive and physically thicker than the host SAMs, they appear as protrusions in STM images [23,229]. Larger bundles of these molecules are also imaged and appear broader, indicating insertion of both isolated single molecules and bundles of molecules within a single sample [23]. The number of molecules inserted together can be controlled to some extent by controlling the tightness (quality/degree of order) of the SAM matrix.
3.2.2T T Scanning Tunneling Microscopy
Please refer to Chapter 1 for a description of STM operation. All STM images presented here were obtained in either of two ambient-condition, custom-built microscopes with high mechanical and temperature stability, and high current sensitivity, as described previously [240]. Successive STM images were acquired over the same area for several (up to 30) hours to determine the behavior of the molecules inserted into host matrices. Relatively large scan areas were acquired so that the activity of tens of isolated molecules could be recorded simultaneously, and, in addition, many high-resolution time- lapse series of images have been recorded.
3.2.3T T Apparent Height Determination
Piezoelectric drift from thermal fluctuations and piezoelectric creep can change the area imaged, so a tracking algorithm was developed to correct for drift during acquisition and for post-acquisition analyses [114]. To generate sufficient statistics for the hundreds of molecules analyzed over hundreds of frames, automated analyses are 71 required. In our previously reported method [114], variations in the calculated apparent height occurred due to scan size (i.e., how many image pixels represent the molecule of interest) and variations of the substrate (i.e., Au vacancy islands and Au step edges within the extracted frame). Thus, the local environment of the molecule influenced the apparent height using this algorithm. To circumvent this problem, we have developed a new background subtraction routine to obtain more accurate and reproducible apparent heights for inserted molecules, thereby reducing the influence of the local environment in our measurements.
Several inserted molecules were present in the STM images at domain
boundaries, step edges and substrate vacancy defects. Figures 3.3X37HT ATX and 3.3TX38H BX T show a
1000 Å × 1000 Å scan area with a single molecule (the same molecule in both images) highlighted (white box) to illustrate our data extraction procedure. To maintain a consistent background, a second box (adjacent, red box) was extracted proximate to the molecule of interest. The median of the top five pixel values for each extracted frame of the inserted molecule determined the raw imaged height of the inserted molecule. The smaller extracted SAM background area was selected to contain neither substrate vacancies nor step edge sites. The mean value of this extracted area was used as the background. The difference between the raw extracted height of the molecule and the background yielded the apparent height, thereby alleviating background variations in the apparent height due to the local environment of the inserted molecule.
Frames 1-50 extracted for this single molecule appear in Figure 3.3TX39H CX .T Each
molecule in Figure 3.3TX40H AX T underwent the extraction process, resulting in the histograms of
72
A B 9 Ångstroms
250 Å 250 Å 0 C
Figure 3.3T :T Molecular tracking process from a series of images (Vsample = -1.0 V, Itunnel = 2 pA) of an C8 SAM with inserted nitro-functionalized OPE molecules 1, with background correction. A. and B. show two frames from this series of images. For each molecule, extracted (white box) and background (red box) regions are automatically selected for each frame. The background corresponds to the SAM apparent height near the extracted molecule. C. Extracted frames 1-50 from the molecule highlighted in A and B with corrected background.
73 the apparent heights. The differences in apparent heights between the ON and OFF states
for each OPE derivative are listed in Figure 3.2X41HT .TX To determine these values, the number of molecules analyzed ranged from 13-26 molecules analyzed over 153-200 frames. Also
displayed in Figure 3.2X42HT TX are example extracted images of the ON and the OFF conductance states for each type of molecule. For all molecules studied, switching was reversible and stochastic. Molecules in the OFF state, depending on the thickness of the host matrix, still appeared as slight protrusions, indicating that switching is not occurring through molecular desorption. The OFF state can also be imaged using the polarizability signal from the molecule when the molecule is not present in topography as will be discussed in Chapter 6.
2 For switching analyses, we typically recorded image areas of 1000-2000P ÅP P to characterize as many molecules as possible, while retaining sufficient resolution to treat each individually. At this scale and at our typical scan line density, the STM tip scanned over a molecule 2-4 times per image (accumulating a total of 5-10 image pixels to represent the molecule, where each pixel represents sampling on a millisecond time scale); thus, the switching activity of a single molecule was not recorded during the majority of each image, but multiple measurements of each molecule were collected in each image.
3.3T T Results and Discussion
Figure 3.2X43HT TX presents the conjugated molecules we have inserted into host SAMs to test the proposed switching mechanisms using STM. Molecules 1 and 2 have previously 74 been analyzed and discussed elsewhere [24,25]. However, they are included here to complete the analysis and testing of the proposed mechanisms. We will discuss how each
molecule listed in Figure 3.2X44HT TX tests proposed mechanisms for conductance switching when inserted into host n-alkanethiolate matrices.
3.3.1T T Reduction and Functional Group Rotation Mechanisms
Seminario and coworkers have suggested that applying a reducing potential to functionalized OPE molecules with both nitro- and amino-groups can, in principle, reduce the molecule, thereby extending the LUMO over the molecule and increasing
conduction [96]. This scheme (Figure 3.1TX45H AX ) requires potentials high enough (calculated at 0 K to be 1.74 V) to reduce the molecule for the conductance change to occur [96].
Similarly, for the functional group rotation scheme hypothesized by Lang and coworkers, a change in the molecular orbitals changes the conductance. Here, the HOMO to LUMO gap would be changed modulating conduction through the molecule, dependent on the rotation of a nitro functional group with respect to the phenyl ring to which it is bonded
[115]. According to this scheme, for the conductance to change, it would be necessary to have substituent groups on the conjugated molecule.
These first two hypotheses of reduction of the molecule and functional group rotation are not consistent with our observations [24,25,241]. Stochastic conductance switching occurs spontaneously at moderate scanning potentials of ±1 V sample bias, which are non-reductive potentials, refuting the possibility of reduction causing the observed changes in conductance. At higher biases, Donhauser et al. have shown limited 75 ability to control switching events from the ON to the OFF states, even out of tunneling range where electrons are not supplied for reduction (however, these higher biases can also cause disorder to the surrounding SAM) [24,25]. For 1, Lewis et al. were able to induce switching at low bias by stabilizing the conductance states via intermolecular interactions with the surrounding matrix [106,107]. Furthermore, as shown in
Figures 3.2X46HT ATX and 3.2TX47H XB,T both the functionalized OPE molecule as well as the unfunctionalized (no substituent groups on the phenyl rings) exhibit the same stochastic conductance switching (i.e., exhibit both ON and OFF conductance states), eliminating functional groups as being required for conductance switching.
3.3.2T T Backbone Ring Rotation Mechanism
Brédas and coworkers have suggested that inhibition of rotation for the middle
OPE phenyl rings would be required for changes in conductance to occur. When all three rings of the OPE molecules are parallel, the conductance would be maximized due to optimal π-orbital overlap; conversely, when the rings are rotated with respect to one
another (Figure 3.1TX48H BX ),T conductance would be lower [116]. This theoretical work relies on substituent groups to inhibit rotation; the nature of the substituent group should not matter as long as it is bulky and can restrict rotation at room temperature. The barrier for ring rotation for one phenyl ring in isolated OPE molecules has been theoretically calculated to be 0.037 eV. This energy is available in our system. If a tolane molecule
complexes with undecane, the barrier for ring rotation can be as high as 1.6 eV [96,241]. T
To test this mechanism, the phenanthrene molecule (2-thioacetyl-phenanthrene) 76 [242] 3 was synthesized to fuse aromatic rings together and thus into the same plane, precluding rotation about the phenyl-phenyl bond (an energy in excess of 6 eV would be required for this to occur, and would not be accessible at room temperature). Several
2 series of images of areas ranging from 1200-1500 Å P over time periods of 10-12 h revealed that the phenanthrenethiolate molecule exhibited two conductance states similar to that observed for previously inserted molecules [108]. The apparent height differences between the ON and the OFF states were determined to be 5.1 ± 1.2 Å in an C8 matrix
(VsampleB =B -1 V, I tunnelB B = 1 pA), analyzing 13 molecules over 300 frames. This observation of two conductance states is inconsistent with the model that internal ring-rotation is the mechanism for the conductance switching observed.
3.3.3T T Neighbor Molecule Interactions Mechanism
Lang and Avouris have calculated a reduction in the conductance of a pair of
“carbon wires” at low bias (0.01 V) when the wires were closely packed (i.e., between threefold hollow sites on Ni{111}) as compared to a single isolated wire [102]. This was attributed to two possible sources. The first was the overlap of π-orbitals between adjacent carbon wires, which is predicted to decrease the effective bond order of the
π-bonds via creation of σ-like bonds linking the wires. The second was that of through- electrode (substrate) binding; the electronic charge transferred to the molecules from the electrodes is predicted to decrease [102]. In related work, Seminario et al. calculated that different conformations of a tolane dimer have different conductivities depending on the ring alignment of neighboring rings [117], while Emberly and Kirczenow have made 77 theoretical calculations comparing single benzenedithiol molecules bridging break junction gaps to overlapping benzenedithiol molecules in break junctions [144]. The magnitude of the theoretically calculated current for a single molecule in the break junction exceeds experimental values measured by Reed, Tour and coworkers [143], which Emberly and Kirczenow attempted to explain by suggesting that the current can flow through overlapping benzenedithiol molecules with only one molecule attached at each end of the fractured Au wire [144] .
Two molecules in close proximity have the potential to interact, thereby changing
the observed conductance (Figure 3.1TX49H CX ).T To test how this would affect the observed conductance in our systems, disulfide forms of the nitro-functionalized OPE molecules
(4[4-(2-nitro-4-phenylethynyl-phenylethynyl)-benzenethiol]disulfide) 4 were inserted into host C8 SAM matrices. Insertion into a thicker matrix was attempted, but unsuccessful; we attribute this to steric hindrance where the disulfide bond was unable to access the substrate, and thus to chemisorb dissociatively. Nuzzo et al. have shown that disulfide molecules dissociatively chemisorb as thiolates on the Au{111} surface [243].
We also believe that after surface-induced dissociation takes place one or both halves of the molecule adsorb, because we can subsequently image inserted molecules; however, due to the higher conductance and close proximity of the inserted molecules, we could
not typically resolve pairs from single molecules. Figure 3.4X50HT TX presents imaged heights of inserted OPE molecules where the host C8 SAM matrix has been normalized
to 1 Å. Figures 3.4TX51H AX and 3.4TX52H BX T present unfunctionalized OPE molecules 2 chemisorbed as thiolates from thiols, where the leftmost inserted molecule is in the ON state in
Figure 3.4TX53H AX T and in the OFF state in Figure 3.4TX54H B,X T while the rightmost molecule remains 78
7 7 AB6 6 5 5 4 4 3 3 2 2 Imaged Height (Å) 1 Imaged Height (Å) 1 25 Å 0 25 ÅÅ 0 20 40 60 80 100 00 20 40 60 80 100 Distance Along Profile (Å) Distance Along Profile (Å) 7 7 CD6 6 5 5 4 4 3 3 2 2
Imaged Height (Å) 1 Imaged Height (Å) 1 25 Å 25 Å 00 20 40 60 80 100 00 20 40 60 80 100 Distance Along Profile (Å) Distance Along Profile (Å)
Figure 3.4T :T Extracted inserted frames for inserted molecules and corresponding line scans (lines through images correspond to associated line scans shown) for A and B unfunctionalized OPE, thiolates inserted as thiols 2 (Vsample = -1.0 V, Itunnel = 2 pA) and C and D dimerized OPE molecule thiolates inserted from disulfides 4 (presumably as pairs) (Vsample = -1.0 V, Itunnel = 2 pA). A. Both unfunctionalized OPE molecules appear in the ON state with apparent heights of ~4.5 Å. B. The leftmost unfunctionalized OPE molecule has switched into the OFF conductance state with an apparent height of ~0 Å. C. Both dimerized pairs of OPE molecules appear in the ON state with apparent heights averaging ~4.5 Å. D. The rightmost dimerized pair of OPE molecules has switched into the OFF conductance state with an apparent height of ~0 Å.
79
ON in both images. Figures 3.4TX55H CX T and 3.4TX56H DX T show nitro-functionalized OPE molecules chemisorbed as thiolates from disulfides 4, presumably inserted as pairs. The rightmost molecules exhibit conductance switching from the ON state to the OFF state from
Figure 3.4TX57H CX T to Figure 3.4TX58H DX ,T while the leftmost molecules remain in the ON conductance state. The apparent heights (the imaged height minus the background height, here ~4.5 Å) for both the dimer thiolate molecules and single thiolate molecules that switch have no significant variation. The average apparent heights for molecules adsorbed as pairs fall within the statistical distributions of our previous measurements of single molecules. As a general observation, molecules inserted from the dimerized version of the nitro-functionalized OPEs exhibit switching events more frequently than do single molecules. This is likely a consequence of disulfide molecules requiring relatively larger defects in the matrix to enable access and dissociative chemisorption (vs. thiols) on the substrate. We have previously shown that the rigidity of the matrix around the switches determines the frequency of switching. No significant differences in conductances were observed for the molecules inserted as pairs, spontaneously formed pairs, or those existing as individuals, thus eliminating neighbor molecule interactions as a mechanism for conductance switching. The observance of not requiring neighboring molecule interactions is consistent with other experimental data [133,150,185].
3.3.4T T Bond Fluctuations
Scanning tunneling microscopy measurements performed by Lindsay and coworkers analyzed conductance switching for inserted α,ω-dithiol OPE molecules with 80 nanoparticles attached to the pendant thiols at the film interface. They hypothesized that
these OPE molecules detached from the surface (Figure 3.1TX59H DX )T causing the observed conductance changes where the nanoparticle was not present in the image [110]. This hypothesis assumed that the nanoparticle must be covalently attached to the surface to be imaged in the ON conductance state, and, conversely, detached from the substrate to be in the OFF conductance state.
Figure 3.5X60HT TX shows images of 1.5-nm C12-passivated Au nanoparticles deposited onto a C12 SAM formed on a Au{111} substrate [244]. No covalent attachment was possible between these nanoparticles and the underlying Au{111} surface; however,
Figure 3.5TX61H AX T shows that STM imaging of these particles was nevertheless possible.
Nanoparticles that are not covalently attached to the underlying substrate are easily
displaced by the STM tip and appear as streaks in STM images. Figures 3.5X62HT ATX and 3.5TX63H BX T
show successive images where the area for Figure 3.5TX64H BX isT indicated by the white box in
Figure 3.5X65HT ATX . These sequential images show that a region that initially had a high coverage of non-covalently attached nanoparticles had a relatively low coverage after the
STM tip imaged the area, thereby sweeping away the nanoparticles. Figure 3.5TX66H CX T shows that a non-covalently attached nanoparticle (indicated by the arrow) may be immobilized
(stable) on a SAM surface (presumably through interdigitation of the C12 ligand shell with the SAM). The presence of a covalent bond with the underlying surface is not necessary to image the nanoparticle with STM. This is contrary to the assumption made in the assignment of the bond fluctuation mechanism by Lindsay and coworkers [110].
Instead, it appears that under the harsher tunneling conditions used (i.e., smaller tip- 81
A 37 B 15 C 13 Ångstroms Ångstroms Ångstroms
375 Å 0 250 Å 0 250 Å 0
Figure 3.5T :T Scanning tunneling microscopy images of 1.5-nm C12-stabilized nanoparticles [244] deposited onto a C12 SAM (VsampleB B = -1.0 V, ItunnelB B = 10 pA). Images A and B are sequential images, and the area of B is indicated by the white box in A. A. Non-covalently attached nanoparticles imaged on a C12 SAM. The streaks in the horizontal direction indicate nanoparticle displacement by the STM tip. B. Particles swept away by the STM tip no longer appear in the image. C. A single nanoparticle (arrow) sufficiently immobilized on the SAM to be imaged.
82 sample separations), Lindsay and coworkers mechanically disrupted the tethered nanoparticles [109].
In addition, we have imaged Au11B B clusters adsorbed for 20 minutes on a
decanethiolate (C10) SAM containing inserted decanedithiol tethers (Figure 3.6X67HT ).TX
Figure 3.6TX68H BX T depicts the intermittent appearance (imaged for a few lines, disappears and
reappears) of a cluster (indicated by arrow) that was not present in Figure 3.6TX69H AX .T This
cluster remained stable through Figure 3.6TX70H CX ,T and then “disappeared” after being imaged
for a few lines in Figure 3.6TX71H DX .T The streak in Figure 3.6TX72H DX T (indicated by arrow) is indicative of the cluster being pushed about by the tip or that the cluster hopped from the surface to the tip.
Finally, in Figure 3.6TX73H EX ,T the cluster was no longer present. We conclude that in the system studied by Lindsay and coworkers, when the nanoparticles disappeared
(reappeared), they had detached from (reattached to) the tethering molecule and either transferred to (released from) the tip, been swept across the surface, or released into
(attached from) the toluene solution. Lindsay and coworkers believed nanoparticle detachment/reattachment was not possible because the incubation time for nanoparticle reattachment was too long; however, incubation times on the minute time scale are adequate to observe many nanoparticles bound to dithiol tethers. We have observed that exposing dithiol-functionalized SAMs to thiol-functionalized nanoparticles for deposition times as short as 20 min produces surfaces with substantial coverages of nanoparticles
(Figure 3.6X74HT ).TX Finally, Lindsay and coworkers have previously reported that nanoparticles may attach to the Au surface through multiple alkanedithiol tethers [133]. For the
83
A B C
250 Å 250 Å 250 Å D E 30 Ångstroms
250 Å 250 Å 0
Figure 3.6T :T A-E Series of STM images of Au11B B clusters adsorbed for 20 min on a C10 SAM with decanedithiol tethers. These images display the mobility of (initially) dithiolate tethered clusters. The boxed area shows A. no cluster attached, B. intermittent appearance of a cluster that was imaged for a few lines, “disappeared”, and then reappeared for the remainder of the image (indicated by the arrow), C. a stable cluster, D. a streak in the image before the cluster was imaged (indicated by the arrow) indicates that the cluster was being pushed by the STM tip, or that the cluster hopped from the surface to the tip, and E. no cluster attached (VsamB ple B = -0.5 V, ItunnelB B = 5 pA).
84 proposed bond-fluctuation mechanism, multiple tethering molecules should influence the switching rate; if a nanoparticle has to break all tethers before switching OFF, this would slow switching. In addition, the presence of multiple tethers should influence conductivity; i.e., the apparent height of the nanoparticle may scale with the number of tethers. Neither of these observations were reported [110].
3.3.5T T Hybridization Change
We have previously suggested that a change in hybridization between the conjugated molecules and the substrate results in the observed changes in conductance
(Figure 3.1E) [24,25]. The contacts of organic molecules to electrodes can have a large influence in the conductance observed through the molecule. For self-assembly techniques, covalent attachment of organic molecules to metal substrates, commonly achieved through thiolate end-groups to Au{111} substrates, form well-ordered SAMs
[15,68]. Other attachment schemes have been realized including various combinations of transition metal and semiconductor substrates with other molecular head-groups such as selenolates [105], and isonitriles [227].
We have experimentally observed conductance switching using STM for each of
the molecules in Figure 3.2X75HT .TX Each molecule has been engineered to test various hypotheses for conductance switching. What has not changed in our molecular design is the S-Au bonding scheme. In the hybridization change mechanism, the metal-molecule system must have enough degrees of freedom to change the structure of the metal- molecule bond. The packing of the matrix around the inserted molecule inhibits the 85 motion of the molecules. We previously showed that tightening or using more rigid matrices leads to longer persistence times in a particular conductance state
[24,25,106,107]. Also, if a particular hybridization is favored, a larger ratio of that conformation should be observed. This is difficult to measure quantitatively with STM since some molecules in the OFF conductance state are not easily imaged, and thus the measurements are biased towards the ON state. The hypothesized change in hybridization
(i.e., change in the tilt or substrate reconstruction) would be concurrent with changing interactions of the molecular π-electrons with the gold substrate, as predicted theoretically [245,246]. A higher degree of overlap was predicted when the molecule is tilted rather than normal with the surface and therefore higher conductance was predicted for a tilted molecule [118,119]. However, we have no direct experimental information as to the tilt angles in the two conductance states. From the electric-field-driven switching data acquired by Lewis et al., we infer that the tilted conformation is the OFF conductance state [106,107], and thus, such increased overlap is not responsible for the observed switching.
Note that we cannot rule out substrate atom rearrangements that occur simultaneously with bond rehybridization as possible contributors to changes in conductance. Even for simple alkanethiolate SAMs, there is significant controversy over
S-Au bonding and binding sites [247,248]. When thiolates and other electronegative species bind to Au and other coinage metals, the substrate surface atoms are known to relax and to become more mobile [249,250]. Alkaneselenolates induce substantial substrate rearrangements on Au{111} and adopt a variety of binding sites [105]. Also, atoms moved from the equilibrium positions change conductance [251]. Determining the 86 role of substrate rearrangement remains central to understanding self-assembly and conductance switching [145].
3.3.6T T Experiments with Restricted Motion
Molecules with two or three possible surface contacts, (thioacetic acid S-{4-[3”-
(4-acetylsulfanyl-phenylethynyl)-4’,6’-bis-phenylethynyl-[1,1’,3’,1”]terphenyl-3- ylethynyl]-phenyl}ester) “two-contact” 5 [237] or the caltrop 6 [238], respectively, would have less ability to move and thus reduced switching activity if each contact were bound. Also, if multiple thiols attach to the gold, this would inhibit switching unless hybridization changes occur at each attachment. Due to the rigidity in the molecular design, tilt would be limited to one plane if both thiols attached, reducing the observed switching events. Multiple attachments could also give insight into how the conductance is affected by having multiple conduction pathways, similar to neighboring molecule interactions.
In our experiments, we infer that it is most likely that only one attachment was able to form, because the spacing between the thiols does not adequately match the binding site spacing between gold substrate atoms. We would expect the registry to be close to that of the host alkanethiolate monolayer preassembled on the Au{111} substrate. For similar “two-contact” molecules, Maya et al. found these molecules had two energy minima: one in a “U-shaped” conformation, and the second in a zig-zag conformation due to free rotation of the terphenyl backbone [236]. Due to steric hindrance caused by the preassembled host SAM, it is most likely that the molecules we 87 observe have inserted in the zig-zag conformation, although it is possible for two contacts to bind to the surface [236]. Furthermore, it is unlikely (but in principle possible) to have a defect in the host SAM large enough to accommodate molecule 6 with several contacts to the substrate. Design and syntheses of new molecules with two or more contacts matching the gold lattice spacing would need to be performed.
The switching frequency of the singly contacted, but complex molecule with two possible contacts was similar to that observed for the nitro-functionalized OPE 1 when in a C12 host SAM. We reported 32% of nitro-functionalized OPE molecules 1 switching
-6 2 with an ON/OFF ratio of 4:1 for a surface coverage of 12.4 × 10P Pmolecules/ÅP P when in a
C12 host matrix that had been deposited for 24 h [24]. For the two-contact OPE molecules 5, we observed 20% of molecules switching with an ON/OFF ratio of 21:5 for
-6 2 a surface coverage of 11 × 10P P molecules/ÅP P in a similar host matrix. The similarity of these results for completely different molecules inserted into a similar host matrix leads us to believe the switching is dependent on something common in both molecules, that is, their attachment to the surface. The size of the more complex molecule does not appear to affect the measured ON/OFF ratios. We expect that the molecules only insert in areas where there is enough space for insertion of a bulky molecule (i.e., large SAM defect sites).
Under similar conditions, the caltrop molecule 6 displays similar switching
behavior to theT nitro-functionalized OPE, 1 and two-contact molecules 5
-6 2 (3.7×10P P molecules/ÅP ;P 4:1 ON/OFF ratio; and 53% exhibiting switching). An interesting observation with the caltrop molecules deals with the issue of the appearance
88
A C SR
Si RS
NO2 22.8 Å 19.0 Å
S S S S S S S S S S S S S S S S S S
Au{111} Au{111}
B D RS SR
Si
NO2 21.7 Å 18.6 Å
S S S S S S S S S S S S S S S S S S
Au{111} Au{111} R = Ac or H depending on whether acetate removal was complete
Figure 3.7T : Schematic of nitro-functionalized OPE 1 and caltrop 6 molecules inserted into C12 SAMs. A. and B. Nitro-functionalized OPE molecule in tilted and normal configurations, respectively. C. and D. Caltrop molecule in tilted and normal configurations, respectively. Note that the nitro-functionalized OPE molecule protrudes further from the SAM in the normal conformation, yet, for the caltrop, protrusion from the SAM is largest for the tilted conformation. The non-surface-bound sulfur groups (shown as SR) could exist as free thiols or as protected thioacetyl groups depending on whether acetate removal was complete.
89 in STM images being a convolution of electronic states and geometric shape. For the nitro-functionalized OPE molecules 1, a molecule oriented normal to the surface will
naturally have a topographic height greater than that of its tilted form (Figures 3.7TX76H AX T
and 3.7TX77H BX ).T Conversely, if only one leg of the caltrop molecule is attached to the gold substrate, a molecule oriented normal to the surface is physically shorter than when the
molecule is tilted due to the tetrahedral geometry (Figures 3.7TX78H CX and 3.7TX79H DX ).T If the observed conductance change were due to physical height alone, the molecules would
always protrude from the SAM, and the average apparent height changes (Figure 3.2X80HT )TX for different molecules in similar matrices would not be in the same range. We would have expected to observe a trend opposite from that observed for the other molecules inserted; that is, we would see a higher ratio of molecules in the OFF state. Because this trend was not observed, we look for similarities between the molecules inserted; this again points toward the interaction of the thiol and substrate enabling the observed switching events.
3.4 Conclusions and Future Direction
We have designed and studied several different molecules exhibiting stochastic conductance switching. Previous work has demonstrated how the surrounding matrix can influence the frequency of stochastic switching events [24,25]. We have studied a number of proposed mechanisms through engineering the molecular structure to test possible conductance switching mechanisms. The only mechanism consistent with all our data is that switching is caused by a change in hybridization of the molecule that occurs with a change in the molecule-substrate contact. We have given examples of switching events in 90 molecules that do not allow internal ring rotation, eliminating this as the general mechanism for the conductance changes in these systems. We have also shown conductance switching in several molecules not containing substituent groups, indicating that the stochastic conductance switching is not attributable to the roles of the functional groups. All examples given have used fully conjugated molecules adsorbed through S-Au interactions. Further work can be performed to analyze different contact interactions, including S-Pd and CN-Au. At this time, all conjugated thiol molecules we have inserted into n-alkanethiolate matrices exhibit bistable conductance switching.
91
Chapter 4
MOTION UP AND DOWN SUBSTRATE STEP EDGES BY OLIGO(PHENYLENE-ETHYNYLENE) MOLECULES
4.1 Introduction
To control our molecular switch system, we want to understand all the apparent height changes that occur for the OPE molecules. Thus far, I have described OPE conductance switching and the mechanism most consistent with our and others’ data.
Further analysis of this system has yielded a third conductance state, which we have attributed to motion of OPE molecules at the Au-substrate step edges. This chapter describes how we have both characterized these motions, and compared them to the previously observed conductance switching.
If future devices are to operate on a single-molecule basis, their design and operation will require understanding of the electronic properties, dynamics and interactions of the molecules utilized. We have examined these interactions for isolated, conjugated molecules inserted into insulating host SAMs using STM. Conductance switching of inserted conjugated molecules has been well characterized by our group and others, and the mechanism for the observed switching is discussed in Chapter 3
[2,24,25,106-111,252]. In this chapter, we show that inserted molecules can diffuse up 92 and down substrate step edges in addition to exhibiting conductance switching [224].
Various mechanisms have been suggested for conductance switching [2,24,25,109], yet the motion discussed in this chapter occurs independently and is uncorrelated with switching events observable on the acquisition time scale of these STM images.
Previously, we analyzed conductance switching for single OPE molecules inserted into host alkanthiolate matrices using STM. We have found these molecules to switch conductance reversibly and stochastically between discrete states with a bimodal distribution indicating an ON and OFF state for the inserted molecules. Molecules with sufficiently large dipole moments can also be driven from one state to another using and applied electric field (supplied by the probe tip in our experiments). The molecules displayed persistence times in each state varying from fractions of seconds to tens of hours [225,230,231]. The packing of the host matrix around the inserted molecule strongly influences the conductance state stability and its interactions with the surface
[24,25,109,253]. Through changes in the degree of order in the matrix [24,25], as well as through changes in the functionality of the matrices themselves [106], we have shown effects on the rate of stochastic switching exhibited by the inserted molecules. From these results, and by testing each of the proposed mechanisms using molecular engineering
(Chapter 3) we have concluded that conductance switching is the result of a change in hybridization at the substrate-molecule interface [24,25,109].
Motion that leads to Ostwald ripening and domain coalescence has been observed within SAMs [16,18,19,27,28]. During SAM deposition, substrate vacancy islands form, which are attributed to the ejection of Au surface adatoms during the relaxation of the
Au{111} herringbone reconstruction [16,18-20,254-256]. As the SAM develops from a 93 disordered to an ordered state, vacancy islands evolve from many small vacancy islands to few large vacancy islands, consistent with Ostwald ripening. This process occurs through single Au vacancies diffusing from the edges of smaller vacancy islands to the edges of larger vacancy islands. The driving force for this migration is the reduction of the total step energy. Since the relative locations of vacancy islands remains fixed, it is unlikely that coalescence (island diffusion) is involved in large vacancy island formation
[16,18,19]. Furthermore, motion is far more rapid at step edges and at domain boundaries than elsewhere on the surface due to decreased coordination [249].
Our previous work discussed coalescence of SAM components via diffusion using mixed monolayers formed from two types of alkanethiolate chains with similar lengths but different pendant functionality [27,28]. These alkanethiolates formed phase- segregated SAMs with domains of common pendant functionality. After solution deposition of the mixed SAMs, small domains coalesced into larger domains, displaying motion working (enthalpically) towards a local minimum. The observed domain coalescence in the monolayer occurred mainly at defects, steps and other areas where surface coverage was lowered because motion in a tightly packed monolayer is available only through collective hindered motions [27,28].
Of relevance to the discussion below are place-exchange mechanisms at step edges, such as one proposed by Kellogg and Feibelman [257]. Using field-ion microscopy and theory, they determined the activation barrier for diffusion of Pt adatoms on a Pt substrate. Place-exchanging adatoms have two possible mechanisms to reach the lower terrace: over the top of a surface atom or replacement of a surface atom in a concerted motion. Adatoms traveling over the top of the surface atoms would have a 94 relatively high barrier to motion due to decreased coordination with substrate atoms. For the concerted-displacement process, a Schwoebel barrier was calculated. This barrier is the difference of the barrier to motion of the adatom down a step minus the barrier to motion of the substrate atom being displaced [258,259]. Since the activation barrier for self-diffusion of the adatom was low, Kellogg and Feibelman concluded that the adatom- substrate atom exchange process with a small Schwoebel barrier was the only feasible mechanism to explain their data, because hopping over a surface atom would have a significantly higher activation barrier [257].
4.2 Experimental Procedure
4.2.1 Sample Preparation
Both thiol and disulfide versions of the nitro-functionalized OPE molecule were
used for insertion (Figure 4.181H ), as conjugated molecules including OPEs have been of interest due to their electronic properties and function [2,24,25,92,98,106,108-
110,169,222,227,260]. Methods for sample preparation and insertion have been discussed in Chapters 1 and 3 [24,25,109,229]. Briefly, to insert the OPE molecules, 0.1 μM solutions of OPEs were prepared in anhydrous THF in a nitrogen environment. Gold substrates with pre-assembled SAMs (24 hr deposition) were placed in the THF solution for ~3 min allowing the conjugated molecules to adsorb to the Au{111} surface at the
SAM defect sites [23,229]. The host SAM matrix thickness influences the apparent heights of the inserted molecules when imaged by STM; thus, short alkanethiolate 95
A NO2 SH 4-(2-Nitro-4-phenylethynyl-phenylethynyl)-benzenethiol B NO2 SS
O2N 4[4-(2-Nitro-4-phenylethynyl-phenylethynyl)-benzenethiol]disulfide
Figure 4.1: A. Thiol and B. disulfide nitro-functionalized OPE molecules used for insertion and analyzed for switching and motion up and down substrate step edges. Insertion onto the Au{111} surface occurred at existing defect sites in the host SAM matrix.
96 monolayers were used to resolve the lower apparent heights of inserted molecules. We have used C8 and C10 SAMs in the work presented here; C12 SAMs were used in earlier studies [24,25].
4.2.2 Scanning Tunneling Microscopy
Please refer to Chapter 1 for a description of STM operation. Here, nitro- functionalized OPEs were analyzed using a custom-built scanning tunneling microscope operating under ambient conditions [240].
4.2.3 Apparent Height Determination
To analyze the behavior of the inserted molecules, time-lapse series of STM images were acquired over sample areas ranging from 200-2000 Å2. To account for thermal fluctuations and piezoelectric translator creep that cause drift during data acquisition, an active tracking algorithm was applied to keep the molecules of interest within the field of view [114]. As discussed previously, STM imaged heights are convolutions of both the physical and electronic structures of the molecules [37]. Thus, changes in either or both will result in a change in the apparent height; therefore, we use the apparent height to monitor the state of a molecule. The apparent height of an inserted molecule is defined as the difference between the imaged height of the inserted molecule and the imaged height of the SAM (see Chapter 3 for further details on the apparent height calculation). 97
2.4 Å Apparent Height
2.4 Å
Figure 4.2: Schematic illustrating the apparent height calculation at a step edge. The blue dashed line illustrates the path of the scanned probe tip; the position of the red dashed line on the upper terrace corresponds to the calculated height of the C10 SAM used as the background to calculate the OPE apparent height. For an inserted OPE switch located near a substrate step edge, the background was standardized by acquiring the imaged height of the SAM on the upper terrace. Note that the inserted switch molecule will protrude from the SAM even if it is located on the lower terrace of a substrate step, where it appears 2.4 Å shorter than when it is located on the upper terrace of a substrate step with the same background. The tip trajectory is drawn closer to the OPE molecule than to the host SAM, consistent with our previous measurements [23].
98 Bumm et al. have shown that the measured apparent height difference for mixed monolayers of C10 and C12 imaged by STM differ from the actual physical thickness of the SAM molecules due to the imaging mechanism of STM [23,91]. It was also shown that a substrate vacancy defect appears as a ~2.4 Å depression, equal to that of the Au
substrate step height, shown schematically in Figure 4.282H [20,21,91,256].
Correspondingly, molecules adsorbed in identical conformations and environments at the top vs. the bottom of a substrate step will exhibit an apparent height difference equivalent to the substrate atomic step height. The noted variations in the measured height of the background C12 SAM used in the previous OPE studies lead to difficulties in differentiating molecules in the OFF conductance state at the top of substrate step edges from molecules in the ON state at the bottom of step edges (refer to Chapter 3 for further details). This does not contradict the observation of conductance switching, but rather gives an indication that two unrelated mechanisms can influence the measurements of the molecules under study.
As discussed in Chapter 3, we have improved our apparent height calculation to create a consistent background height determination that is not influenced by defect sites or by substrate step edges within the host matrices. The background was determined as the difference between the imaged height of the inserted molecule, and the average of a defect-free area of the SAM near the inserted molecule. Using this improved apparent height calculation, we can easily differentiate between conductance switching and place- exchange up and down substrate step edges. We find that molecules at step edges and substrate defects often exhibit apparent height changes of ~2.4 Å, equal to that of the Au substrate monatomic step height. Therefore, we attribute this to motion up and/or down 99 the step edge. This behavior is in addition to conductance switching, which typically yields a larger apparent height change. The new background measurements yield apparent OPE heights that vary with the thickness of the SAM matrix; for example, an
C8 matrix yields apparent OPE heights greater than those of a C12 matrix due to the difference in matrix thickness [91].
4.3 Results and Discussion
For some inserted OPE molecules, we observed three apparent heights, while for others we found bimodal distributions [24,25]. The three apparent heights for inserted
molecules are highlighted by the colors red, green and blue in Figure 4.383H . In the data presented here, the thickness of the SAM was chosen (C10 and C8) to enable differentiation of these height differences as a function of location on the substrate.
Although the apparent height of the inserted molecules varied with the SAM matrix, we observed the difference in apparent height between the middle and highest value apparent heights always to be ~2.4 Å corresponding to the substrate Au{111} monatomic step height. From these observations we assign the three apparent heights as follows: apparent heights highlighted in red correspond to molecules in the ON conductance state on the terrace above a substrate step or above a substrate vacancy defect, apparent heights highlighted in green correspond to molecules in the ON conductance state on the lower terrace below a substrate step or in a substrate vacancy, and apparent heights highlighted in blue correspond to the molecule in the OFF conductance state. Reasons for these assignments are discussed below. We cannot differentiate the location of the OFF 100
A
Time (h) 200 0 1 2 3 4 5 6 C 4.9 5 B 180 4.5 160 4 140 3.5 120 0.5 3 100 2.5 2 80 Occurrences 1.5 60 2.5 Apparent Height (Å) Height Apparent 1 40 0.5 20 0 -1 7 5 10 15 20 25 30 35 40 45 50 0123456 Frame Number Apparent Height (Å) Figure 4.3: A. Frames 1-50 for an inserted nitro-functionalized OPE molecule in an C8 host matrix. The molecule was located at the edge of a substrate vacancy island. The boxed areas (red, green and blue) correspond to the apparent heights in parts B and C. B. Apparent height vs. time (frame number) for the extracted molecule in part A. Two phenomena are displayed for this molecule. The transition between the red boxed area and the blue boxed area is the molecule switching conductance states. The height difference of the green boxed area and red boxed area is that of a Au{111} step height (2.4 Å), indicating that the molecule was mobile in to and out of the substrate vacancy island. The motion at substrate step edges is reversible between the top and bottom of the step (frames 32-41). C. Histogram of occurrences vs. apparent height for all molecules in this data set. The number of occurrences at each apparent height varied by data set. Additional histograms for different samples are shown in Figure 4.4.
101 conductance state as at the top or bottom of a terrace or vacancy site separately because the apparent height of the OFF state at the bottom of a vacancy site is buried by the host
SAM matrix. The only molecules that exhibited all three apparent heights were those located near step edges or near substrate vacancy islands. Molecules that are located at other sites within terraces, such as at domain boundaries, exhibit only two apparent heights.
Figure 4.3A84H displays the first 50 frames extracted for a nitro-functionalized OPE thiol-terminated molecule inserted at a step edge in an C8 host matrix. Using the color scheme described above, this single molecule exhibits all three apparent heights as
indicated by the colored boxed areas. Figure 4.3B85H is the corresponding apparent height
vs. time (frame number) data for the molecule extracted in Figure 4.3A86H . The color
intensity gradient in Figure 4.3B87H corresponds to the Gaussian fit to the apparent height
distribution histogram (Figure 4.3C88H ) for the cumulative data of all the inserted molecules from this particular data set. We observed a trimodal distribution containing peaks centered at apparent heights of 0.5 Å, 2.5 Å and 4.9 Å. The height difference between the middle and highest peaks is 2.4 Å, again corresponding to a Au step. From this we posit that two of the peaks correspond to different conductance states and that the third peak corresponds to the inserted molecules in one of the conductance states, but on a different terrace. We infer that the lowest apparent height (blue, ~0.5 Å) corresponds to the OFF conductance state when the molecule is located at the top of a Au step. We assign the middle apparent height (green, ~2.5 Å) as that of the ON conductance state when the molecule resides at the bottom of a monatomic Au step. The highest apparent height (red,
~4.9 Å) corresponds to the ON conductance state at the top of a Au step, i.e., on the 102 upper terrace. We do not image the OFF state for the molecule when it is located in a Au vacancy island or at the bottom of a substrate step because the apparent height of the molecule is below that of the host SAM. Following our hypothesis, the OFF state on the lower terrace of the substrate step edge would, in principle, be imaged at an apparent height of approximately -1.9 Å with respect to the SAM at the top of the step edge, and is thus masked by the matrix in these measurements.
The molecule shown in Figure 4.3A89H exhibits all three apparent heights, demonstrating that both conductance switching and motion up and down the step edge are possible for a single inserted molecule. However, switching and motion up and down a substrate step appear to be independent events on the time scale of our images. In
Figure 4.3A90H , the blue box around frames 9-13 indicates the OFF conductance state at the top of the step edge. Frames 8 and 14, which precede and follow the blue boxed area, respectively, are boxed in green, indicating the middle apparent height which is the ON state at the bottom of the step edge. This identifies the molecule as exhibiting both motion and switching between frames. However, each time motion up and down the substrate step occurs, conductance switching does not necessarily occur simultaneously.
A green box around frames 33-40 again indicates the ON state at the bottom of the substrate step edge. The frames preceding and following this green boxed area are boxed in red, indicating the molecule in the ON conductance state at the top of the substrate step edge. This identifies the molecule as exhibiting motion up and down the substrate step without a simultaneous switching event. Therefore, it is possible for a switching event to occur simultaneously with motion up or down a substrate step within the time scale of
103
A B 600 4.3 500
400 0.4 1.9 300 Occurrences 200
100
0 -1- 0 123456 7 8 500 Å Apparent Height (Å) C D 300
250
200 0.40.4 2.32.3 4.84.8 150
Occurrences 100
50
0 -1 0 1 2 3 4 5 6 7 8 250 Å Apparent Height (Å)
Figure 4.4: A. A topographic STM image of nitro-functionalized OPE molecules inserted into a C10 SAM (Vsample= -1.0 V, Itunnel = 2 pA). B. Corresponding histogram for the apparent heights of the inserted molecules extracted from A and other images in this series. Red indicates the ON conductance state at the top of a monatomic substrate step edge. Green indicates the ON conductance state at the bottom of a step edge and blue indicates the OFF conductance state. The OFF state at the bottom of the monatomic substrate step edge is not imaged by STM. The apparent heights are shifted to lower values compared to those of the C8 matrix due to the thickness difference of the matrices. C. Topographic STM image of nitro-functionalized OPE molecules inserted into an C8 SAM (Vsample = 1.0 V, Itunnel = 25 pA). D. Corresponding histogram for the apparent heights of the inserted molecules extracted from C and other images in this series. The color scheme is the same as for A and B.
104 imaging, yet we have shown that motion up or down a substrate step and switching can occur independently. It is important to note that the conductance states and position are determined by several scans across a molecule even within one image (by selecting the imaging resolution appropriately), thereby reducing the possibility that current fluctuations or instabilities lead to erroneous assignments.
The calculations we have developed to determine the apparent heights of the inserted OPE switches are dependent on the thickness of the host SAM molecules. Thus, the calculated apparent height for an OPE inserted into an C8 SAM would differ from that of the apparent height for an OPE molecule inserted into a longer chain C10 SAM.
This is shown in Figure 4.491H where nitro-functionalized OPE molecules have been
inserted into C10 (Figure 4.4A92H ) and C8 (Figure 4.4C93H ) SAMs. The switch molecules have a higher apparent height in the C8 SAM as shown by the ON state at the top of a
substrate step (red) for the C10 SAM (Figure 4.4B94H ) compared to the C8 SAM
(Figure 4.4D95H ). Notable here is that the difference in apparent height between the ON state at the top of a substrate step edge (red) and the ON state at the bottom of the substrate step edge (green) for switch molecules inserted into different thickness host SAMs are both approximately that of the Au{111} substrate step height [20,21,91,256]. This difference has been consistent in our measurements, leading us to define this difference as motion of the OPE molecules at the substrate step edges.
Figure 4.5A96H shows the apparent heights vs. time (frame number) calculated for the first 50 frames for an inserted nitro-functionalized OPE molecule inserted at a Au{111}
substrate step edge in Figure 4.4C97H . The molecule in these extracted frames exhibits three apparent heights as indicated in the histogram by ON at the top of the substrate step-edge 105 (red), ON at the bottom of the substrate step-edge (green) and OFF (blue). The
corresponding extracted frames for this molecule (Figure 4.5B98H ) exhibit the substrate-step edge at which the molecule is located and place-exchanges up and down.
To show that molecular motion is responsible for the trimodal distributions observed, we have imaged areas of our SAMs with the inserted nitro-functionalized OPE disulfide molecule at high resolution. The disulfide molecules require areas of the host matrix to be less densely packed for insertion, and thus they favor step edges and substrate vacancy islands over domain boundaries, relative to thiols. Monolayer coverages are lower at step edges than elsewhere on the surface; domain boundaries favor single-molecule insertion
[249]. Nonetheless, we find occasional insertion of the disulfides into SAM matrix
domain boundaries, but to a substantially lesser extent than for thiols. Figure 4.6A99H is an area (500 Å × 500 Å) from a time-lapse series of 140 frames. The switch molecule boxed in magenta is located at a SAM domain boundary with neither substrate defects nor steps nearby. Note that in this case we do not resolve whether the two nitro-functionalized
OPE molecules remain together on the surface after disulfide cleavage. Throughout imaging, this switch (or pair of switch molecules, as noted above) exhibits only two
apparent heights, as shown in Figure 4.6B100H . The limitation of only two states is attributed to the lack of substrate steps proximate to the molecule; therefore, the molecule or molecules can exhibit only the two apparent heights (switch molecules in the same bundle switch together). This demonstrates a mechanism for conductance switching independent of motion up or down substrate steps, indicating that motion up and down the substrate step is not the cause of conductance switching.
106
A Time (min) 80 30 60 90 120 150 180 210 240 270 300
7
6
(Å) 5
4
3
2 Apparent Height 1
0
-1 0 5 10 15 20 25 30 35 40 45 50 Frame Number B
Figure 4.5: A. Apparent height vs. time (frame number) histogram for frames 1-50 of a molecule from Figure 4.4C101H inserted at a substrate step edge. The location of the molecule and conductance state are indicated by the colors: ON at the top of the substrate step-edge (red), ON at the bottom of the substrate step-edge (green), and OFF (blue). This molecule displayed three apparent heights, consistent with the molecule place-exchanging up and down the substrate step-edge. B. The OPE molecule extracted from frames 1-50 corresponding to the histogram in A. Note the step edge present in these extracted frames.
107
The second extracted area, boxed in cyan in Figure 4.6A102H , encompasses the area around a Au vacancy island. This vacancy island allows the switches, through place- exchange up and down the substrate step, to occupy locations across the Au step within or outside of the vacancy island. The possible locations for molecules descending a step
edge are shown in the extracted frames in Figure 4.6C103H . Each frame displays two extracted molecules, one (white arrows) just outside and thus on the upper terrace of a monatomic substrate layer above the Au vacancy island, the other (yellow arrows) within the vacancy island. The left extracted frame displays the OFF conductance state for the switch at the top of the substrate step edge with the ON conductance state for the switch within the vacancy island. The frame on the right displays the opposite conductance state for each switch at the same location. Note that the switch at the upper terrace of the substrate step protrudes to a greater extent in the ON state than the switch located at the bottom of the substrate terrace in the ON state. These correspond to our two observed ON apparent heights. A switch located on the upper terrace of the substrate step edge in the OFF state protrudes only slightly from the SAM (less than 1 Å), whereas within the vacancy site the switch in the OFF state is not imaged by STM, leading to a calculated apparent height of
0 Å. As discussed above, this is due to the conductance being lower than that of the host matrix. Since the calculated apparent heights for the OFF state are not significantly different on the upper and lower terraces, both locations contribute to the third observed apparent height we label as the OFF state. The apparent heights of the imaged conductance and position states correspond to the trimodal or bimodal distributions, dependent on the molecules’ locations on the substrate. 108
A
125 Å BC
Figure 4.6: A. Topographic STM image of nitro-functionalized OPE disulfide molecules inserted into an C8 matrix (Vsample = -1.0 V, Itunnel = 2.0 pA). The magenta box highlights a switch located at a domain boundary. The cyan box highlights a switch on the upper terrace (white arrow) and a switch on the lower terrace (yellow arrow) of a substrate vacancy island. B. Extracted frames of the ON and OFF states of the switch at a domain boundary. At this location, far away from any substrate steps, switches exhibit only two apparent heights. C. Extracted frames of the two possible conductance states when a switch is located on the upper terrace of the substrate step at the edge (white arrows) switching from OFF to ON (left frame to right frame), and when a switch is located on the lower terrace of the Au step (yellow arrows) switching from ON to OFF (left frame to right frame). Note that the switch on the upper terrace of the substrate step protrudes to a greater extent in the ON state than the switch located at the bottom of the substrate terrace in the ON state. A switch located on the upper terrace of the substrate step edge in the OFF state protrudes only slightly from the SAM, whereas, within the vacancy site, the switch in the OFF state is not imaged by STM.
109
Figures 4.3104H and 4.6105H show the dependence of apparent height both on conductance switching and location relative to substrate steps. Although STM cannot directly probe the molecule-substrate interactions responsible for the observed motion, possible mechanisms for this motion can be related to previously discussed mechanisms for adsorbate motion. As Stranick et al. observed [249], in Au diffusion, covalently bound molecules can remain attached to substrate atoms and move as a complex. If this mechanism describes the step diffusion observed here, a single uncoordinated vacancy would exchange sites with the inserted molecule attached to its substrate atom, allowing it to move reversibly between the top and bottom of the step edge as in the place- exchange mechanism for metals proposed by Kellogg and Feibelman [257]. The inserted rigid aromatic molecules would be more likely to exhibit such motion than SAM (matrix) counterparts because they are not held in place by the same strength of van der Waals interactions that exist between the saturated alkanethiolate molecules. Decreased stability and thus increased mobility are expected at substrate step edges due to the limited coordination of atoms located at these step edges. However, it would be less energetically favorable for the vacancy to have mobility through the SAM matrix because of the alkanethiolate intermolecular interactions that would need to be disrupted for such motion to occur. While it is known that attaching thiolates weakens the bonds of the top- layer Au atoms to the substrate [249,261], the relative strengths of these interactions for alkanethiolates vs. conjugated thiolates are not known.
Having two OPE molecules in close proximity could lower the place-exchange barrier for the Kellogg-Feibelman mechanism [257]. We expect that with OPE disulfide insertion, there is more than one inserted molecule in proximity. However, adjacent 110 molecules are difficult to resolve, especially at Au steps because of the STM imaging mechanism [37]. As noted above, the insertion of the nitro-functionalized OPE disulfides requires more space than their nitro-functionalized OPE thiol analogues. We believe that a single uncoordinated vacancy could exchange sites with the inserted OPE molecule attached to one or more substrate atoms, allowing it to move reversibly between the top and bottom of the step edge. We do observe qualitatively more place-exchange activity up and down steps for insertion performed with the nitro-functionalized OPE disulfides.
This could be due to interactions between these molecules enhancing this motion, but we cannot rule out the preselection of larger void areas required to insert the disulfides that would also lead to more motion.
4.4 Conclusions and Future Direction
We have demonstrated motion of inserted molecules up and down substrate step edges. Previously observed stochastic conductance switching is not dependent on the presence of substrate vacancy islands nor on place-exchange up and down substrate step edges. Our results show that motion up and down step edges is possible even for well- ordered SAM matrices. It should be noted that while motion up and down substrate steps was observed, minimal lateral diffusion within terraces at insertion sites such as domain boundaries was observed. Several mechanisms may play roles in interactions on the molecular level that govern these types of motion and we posit that these interactions will be increasingly important as attempts are made to pattern and to control the function of single molecules. 111
Chapter 5
REAL-TIME MEASUREMENTS OF CONDUCTANCE SWITCHING AND MOTION OF SINGLE OLIGO(PHENYLENE-ETHYNYLENE) MOLECULES
5.1 Introduction
The switching and motion analyses presented in Chapters 3 and 4 were performed on an image by image basis; thus, we could analyze the switching of tens of molecules over hundreds of images. However, each STM image was acquired over several minutes
(~5 min), and for the majority of each image, we did not record the activity of any OPE molecule. This chapter details how we acquired height vs. time (z vs. t) data over single
OPE molecules to understand their ability to switch and to exhibit motion on faster (ms) time scales.
In order to understand the electronic properties involved in the conductance switching of individual molecules, it is important to analyze and to understand the motions of the molecules and the substrate atoms to which they are bound. We and others have studied the conductance switching of OPE molecules isolated in host SAM matrices
[24,25,106-110,225,262], and their place-exchange up and down substrate step edges
[224] using STM. Previously, conductance switching and motion of the OPE molecules were analyzed by determining the apparent height of each molecule in each image over 112 hundreds of frames [114,224]. However, on the time scale of imaging (~5 min), each
OPE molecule was imaged for only 1-25 ms per image, depending on the image resolution. To broaden the dynamic range for measuring the switching of OPE molecules and the motion of the OPE molecules and the host SAM, real-time (data collected at ca.
10 kHz), fixed lateral position topographic measurements have been acquired.
Figure 5.1A106H shows a schematic of a nitro-functionalized OPE molecule switching conductance states. As described in Chapter 3, the reversible, stochastic switching we observed is due to a hybridization change at the substrate-molecule interface via a
molecular tilt. Figure 5.1B107H is an example of the real-time data of an OPE molecule exhibiting conductance switching discussed in this chapter.
Real-time motions have been studied previously using STM for single atoms and molecules using time-dependent tunneling measurements [39,263-265]. These studies, performed in ultrahigh vacuum, analyzed the dynamics of single atoms as they diffused or were laterally manipulated by blanking the scanning tunneling microscope current
FBL and recording the tunneling current vs. time (I vs. t) [47]. Our real-time measurements were performed in ambient conditions; thus, the STM tip was not stable enough to blank the FBL without significant tip drift caused by thermal fluctuations or piezoelectric creep. This drift would cause the tip either to contact the sample or to drift out of tunneling range, so these real-time measurements were performed with the FBL active.
113
A
B 4 2 0 -2
Topographic Height (Å) - 4 3 4 5 6 7 8 9 10 11 Time (s) Figure 5.1: A. Schematic of an inserted nitro-functionalized OPE molecule switching conductance states. The switching has been observed as reversible and stochastic. B. Topographic z vs. t data recorded over a single nitro-functionalized OPE switch. The data indicate that the molecule exhibited switching events on time scales faster (milliseconds) than STM imaging (minutes).
114 The bandwidth of the FBL determines the rate at which the tip responds to a change in current. This determines the time resolution that we can observe topographic switching of the OPE molecules. At lower frequencies (below ~1 kHz), the FBL is able to match the rate of change of the bias signal with a change in the z-piezoelectric transducer, resulting in tip motion. As frequencies are increased (above
~1 kHz), oscillations occur faster than the FBL can respond, causing the z-piezoelectric transducer signal to respond to the average bias voltage rather than the high frequency signal [231]. The current preamplifier portion of the FBL has a higher bandwidth (~1 MHz) than the total FBL and is able to transmit the higher frequency signal. This is the same principle we use for ACSTM (Chapter 1) where both the microwave frequencies applied and the signal at the difference frequency are greater than the bandwidth (and thus the response) of the FBL. Furthermore, a similar high frequency signal (typically ~200 Hz) is used in STM to record differential conductance (dI/dV) images (Chapter 1).
The Nyquist theorem states that to observe a given frequency (ω), the sampling rate must be greater than 2ω, so that both the peak and trough of the frequency are sampled (at least) [266,267]. The 10 kHz acquisition for the real-time measurements is well above the bandwidth of the FBL (~ 1 kHz), ensuring that all of the FBL response is detected in our measurements. The duration of data acquisition on a molecule is dependent on the thermal and piezoelectric drift. We acquired data for 20 s at each point, a time in which the tip remained over the molecule as observed in topographic images by viewing the location of the molecule before and after the point was acquired. 115 5.2 Experimental Procedure
5.2.1 Sample Preparation
Sample preparation and insertion have been described previously, please refer to
Chapters 1 and 3 [24,25,109,229]. Here, C12 SAMs were adsorbed from 1 mM solutions onto flame-annealed Au{111} substrates for 5 min. This short adsorption time formed
SAM matrices with less order and containing relatively more defect sites than SAMs deposited for longer adsorption times (typically 24 h), ultimately allowing more stochastic switching and motion [24,25]. After adsorption, the SAMs were rinsed with ethanol and dried under nitrogen. Nitro-functionalized OPE molecules were inserted for
1 min from 1 µM solutions into the preformed SAMs under a nitrogen environment.
5.2.2 Scanning Tunneling Microscopy
Please refer to Chapter 1 for a description of STM operation. Here, the nitro- functionalized OPEs were analyzed using a custom-built scanning tunneling microscope operating under ambient conditions [240].
5.2.3 Height vs. Time Spectroscopy Acquisition
The z vs. t data measuring conductance switching and place-exchange for OPE molecules and the host SAM were recorded with the FBL active, enabling us to perform measurements in ambient conditions by minimizing drift and recording the topographic
116
A 4 Å ngstroms
100 Å 0 -2 B Before z-drift slope correction -4 -6 -8 -10 0 2 4 6 8 10 12 14 16 18 20 Topographic Height (Å) Height Topographic Time (s)
4 C After z-drift slope correction 2
0
-2-
-4
Topographic Height (Å) Height Topographic 0 2 4 6 8 10 12 14 16 18 20 Time (s) Figure 5.2: A. Topographic STM image with colored dots indicating z vs. t data acquisition points recorded over this image (Vsample = -1.0 V; Itunnel = 1.0 pA). The molecule circled in red corresponds to the z vs. t data in B and C. The red arrow indicates z-drift that occurred during acquisition. Data for the green and teal circled molecules are shown in Figure 5.3108H . The points in black were recorded for this image, but data are not shown. B. Uncorrected z vs. t data for the molecule circled in red in A. C. A linear background correction was applied to the data in B to remove the z-drift that occurred during data acquisition. 117 height of the molecule. During image acquisition, z vs. t data were obtained by pausing the tip for 20 s over OPE molecules or the host SAM and recording the FBL response to the molecule within the junction with a data acquisition rate of 10 kHz. Sites for real-time data acquisition were selected by imaging a sample area and then defining acquisition points in that image over the molecules of interest. A subsequent image was acquired where the tip was paused for the real-time measurements at the previously defined
acquisition points. Figure 5.2A109H displays a STM image in which z vs. t data have been recorded at each of the colored dots within the image.
While drift in the z-direction (normal to the surface), was minimized by keeping the FBL active, drift was still present during the real-time data acquisition. This was
observed in the topographic images as horizontal lines (red arrow, Figure 5.2A110H ), which were caused by a difference in the z-position of the tip before and after data acquisition.
The z-drift was also present in the real-time data, as shown in Figure 5.2B111H . These data
were recorded over the molecule circled in red in Figure 5.2A112H . In our postacquisitional analyses, we corrected for this linear z-drift by subtracting a linear background to
produce the z-corrected trace as shown in Figure 5.2C113H .
5.3 Results and Discussion
Real-time data recorded over the molecules circled in green and teal in Figure 5.2114H ,
(Figures 5.3A115H and 5.3B116H over green and Figures 5.3C117H and 5.3D118H over teal) are shown.
Previously, we have shown the OPE molecule switching to be highly dependent
118
4 A 2
0
-2-
-4
Topographic Height (Å) Height Topographic 0 2 4 6 8 10 12 14 16 18 20 Time (s) 4 B 2
0
-2-
-4
Topographic Height(Å) 0 2 4 6 8 10 12 14 16 18 20 Time (s) 4 C 2
0
-2-
-4
Topographic Height(Å) 0 2 4 6 8 10 12 14 16 18 20 Time (s) 4 D 2
0
-2-
-4
Topographic Height (Å) Height Topographic 0 2 4 6 8 10 12 14 16 18 20 Time (s)
Figure 5.3: Real-time data recorded for the molecule circled in green (A. and B.) and in teal (C. and D.) in Figure 5.2119H . Data for each molecule were recorded in separate images. Data acquisition conditions: Vsample = -1.0 V; Itunnel = 1 pA. 119 on the quality of the host SAMs around the inserted switch molecules where a less ordered SAM matrix (assembled for 5 min) showed a higher degree of switching than a
more ordered SAM (assembled for 24 h) [24,25]. The real-time data in Figure 5.3120H are indicative of the environment around the inserted molecules. We observed that data acquired over single OPE molecules in several images tend to have similar features.
Figures 5.3A121H and 5.3B122H display real-time data indicating that the inserted switch molecule was active as a switch during much of the data acquisition (topographic height change
~4 Å). From this, we infer that there is less order in the packing of the host SAM around
this molecule as the data showed activity in several series. Conversely, in Figures 5.3C123H
and 5.3D124H , the data show few and small (~0.5 Å) topographic changes. The origins of these changes are discussed below. Here, we believe that the packing of the host matrix is likely more rigid around the molecule, and thus inhibited switching on the millisecond or faster time scale.
The topographic changes measured in the real-time data should correspond to height changes possible within our system. Table 5.1 displays the physical and measured heights of features associated with SAMs including the Au substrate atomic step height measured by STM [91], the apparent height difference between the OPE and the host
C12 SAM measured by STM [230], the apparent heights between the OPE in the ON and the OFF conductance states, and the height of the Au substrate relaxation when a SAM is assembled on it as measured by X-ray diffraction [268,269].
Figure 5.4125H displays a 500 Å x 500 Å imaged area of a SAM containing inserted nitro-functionalized OPE molecules that appear as protrusions from the SAM [23]. 120
Feature Physical Height (Å) Measured Height (Å) Au atomic step 2.35 2.35* Au{111} three-fold hollow to atop 0.15 n/a Au{111} relaxation with SAM 0.5** n/a ON OPE (normal) to C12† 5.3 4.5 (± 0.7) OFF OPE (30° tilt) to C12† 2.9 0.9 (± 0.6)
Table 5.1: Atomic and molecular features of a SAM with inserted OPE molecules. *Gold atomic step height at a substrate step or substrate vacancy site measured with STM [91]. **X-ray diffraction has measured the relation between the top most layers of the Au atoms [268,269] †The physical heights were calculated from the bonding lengths of the molecules and the apparent height were measured by STM [230].
121
A 13 2
3 a 4 Topographic Height (Å) 1 100 Å 0 B 4 2 0 -2 Height (Å) -4 4 6 8 10 12 14 C Time (s) - 3 15 OFF 10 4 2 .
2 ON 4 4 .
5 0 STEP
0
Occurrences (x10 ) -2 0 2 Height (Å) Figure 5.4: A. Topographic STM image of nitro-functionalized OPE molecules inserted into a host C12 SAM matrix (Vsample = -1.0 V, Itunnel = 1 pA). B. Real-time z vs. t measurements for molecule 1. Height changes indicate the molecule was actively switching conductance and place-exchanging up and down the substrate step edge during the measurement. C. Occurrences vs. height tabulation for the real-time measurements in B displaying ON, STEP edge and OFF topographic heights.
122 Several images were obtained over this area with z vs. t measurements recorded in each
image for each inserted OPE molecule (labeled 1-4 in Figure 5.4A126H ) and over the host
C12 SAM (labeled a). The labels are offset from each molecule to show each molecule’s
location within the host SAM. Figure 5.4B127H shows z vs. t data recorded over molecule 1.
This molecule exhibits topographic changes that correspond to conductance switching height changes (~4 Å apparent height change), as displayed between 4 and 5 s, and topographic changes corresponding to place-exchange up and down the substrate step edge (~2.3 Å apparent height change), as displayed at several points. By tabulating the
occurrences at each apparent height from the time history in Figure 5.4B128H we find three
distinct peaks (Figure 5.4C129H ), in agreement with our previous data where we characterized molecules over a series of STM images at slower time scales, and found place-exchange up and down substrate step edges occurring independently of conductance switching
(Chapter 4) [224]. We assign the peaks as the following: ON for the greatest apparent height, STEP for the middle apparent height and OFF for the lowest apparent height.
Note that we are unable to distinguish between OFF for molecules at the top or bottom of the substrate step edges, because the molecules’ topographic heights when in the OFF conductance state are indistinguishable from the surrounding host SAM matrix (i.e., the molecules do not appear in the image). See Chapter 4 for further discussion of the place- exchange of OPE molecules at substrate step edges [224].
Figure 5.5130H displays a few typical time z vs. t data series as examples of many measurements obtained over OPE molecules and the host SAM, including real-time switching activity, place-exchange up and down the substrate step edge, and inactive
123
4 A 2 0 -2 -4 2 3 4 5 6 7 8 9 10 11 12 4 B 2 0 -2 -4 2 3456789101112 4 C 2 0 -2 Topographic Height (Å) -4 2 3456789101112 4 D 2 0 -2 -4 6 7 8 9 10111213141516 Time (s) Figure 5.5: Real-time z vs. t measurements obtained from the OPE molecules and the SAM matrix at point a, as marked in Figure : A. molecule 2; B. molecule 3; C. inactive molecule 4; D. SAM matrix at point a. 124
molecules from Figure 5.4131H . Each real-time data series was recorded for 20 s; however, only the most active 10 s are displayed here for ease of visualization. Molecule
1 (Figure 5.4B132H ) displayed both conductance switching and place-exchange at a step edge.
Figure 5.5A133H presents z vs. t for molecule 2, which exhibited only stochastic conductance switching between ON and OFF states, measured as topographic height changes of ~4 Å.
Only conductance switching was observed since molecule 2 was located within a SAM domain boundary and not at a step edge.
Figure 5.5B134H shows z vs. t data for molecule 3, which underwent place-exchange up and down a substrate step edge measured as a height change of ~2.3 Å. This motion is only observed at substrate step edges and substrate vacancy sites where molecules are able to place-exchange up and down a substrate step. The observed height change is equal to that of a step in the Au{111} substrate [20,240,256]. Molecule 3 was not active as a switch during these measurements; however, it did exhibit smaller motions, observed as topographic changes of ~0.5 Å, typical and far above the noise floor of our z vs. t measurements. These small topographic height changes were also observed for the host
SAM and are discussed further below. Not all OPE molecules exhibit conductance
switching or place-exchange during all z vs. t measurements. Figure 5.5C135H shows a measurement for molecule 4 where the molecule displays neither topographic height changes corresponding to conductance switching nor motion at a substrate step edge.
Over the host SAM and over the OPE molecules, we observe small topographic
height changes of ~0.5 Å as displayed in Figure 5.5D136H . Our previous data have shown that both OPE molecules and SAM matrices can exhibit motion [28,224]. Experimental and theoretical data for SAMs suggest that the S headgroup of a SAM molecule is bound to a 125 surface Au adatom and that these Au-thiolate species are mobile on the Au{111} surface
[249]. We tentatively assign the small ~0.5 Å topographic height changes observed over the OPE and SAM to substrate or molecule-substrate complex motions [269,270].
Grazing incidence X-ray diffraction performed on alkanethiolate monolayers found that the topmost layer of Au was distorted from the ideal bulk termination and oscillated between maximum and minimum values with an amplitude of 0.5 Å along the surface normal direction [268]. We expect a 0.5 Å topographic height change for a 0.5 Å physical height change of the substrate [91]. Our temporal resolution is limited by the bandwidth of the feedback loop; therefore, some conductance switching and motion events may not be observed. This can be improved either by recording tunneling I vs. t (the current amplifier bandwidth here is ~30 kHz) or by performing frequency-based measurements
(to reach on the order of 10 GHz) [52]. This higher frequency range would be able to probe the slew rate of the molecular switches, as this should correspond to that of a frustrated rotation.
5.4 Conclusions and Future Direction
The real-time measurements presented here show that OPE molecules exhibit conductance switching and place-exchange at substrate step edges faster than the time scale of typical STM image acquisition. Further, we have been able to observe smaller topographic height changes that have not previously been measured in real time, and that we attribute to substrate and/or molecule-substrate motion.
126
Chapter 6
IMAGING SINGLE-MOLECULE POLARIZABILITY AND BURIED INTERFACE DYNAMICS
6.1 Introduction
Scanning tunneling microscopy is a powerful tool for imaging down to the atomic scale, and for understanding electronic properties of conducting and semiconducting surfaces. However, we are interested in expanding the capabilities of the instrument to address single molecules for elucidating their chemical properties. We have developed an
AC scanning tunneling microscope with the ability to couple microwave frequencies
(0.5–20 GHz) into the tunneling junction. Here, we show how we have used these microwave signals to measure the polarizabilities of single molecules and how we have used the microwave signals to gain a better understanding of our OPE switching system.
As knowledge of single-molecule interactions at the nanoscale increases, metrology tools recording topographic information at the single-molecule level can be enhanced by combining an additional signal from which further chemical information can be obtained. Here, we have used microwave-frequency ACSTM [23,51-54] to image the topography and polarizability of single molecules simultaneously. We have imaged a variety of SAMs comparing their relative polarizabilities, we have calculated 127 polarizability values for each molecule, and we have applied the polarizability signal to study single-molecule switches.
Devices based on electronic switches are central to modern information technology. As miniaturization of such devices continues, nanostructures and single molecules that exhibit conductance switching must be understood and controlled within their physical environment [2]. To accomplish this, we must have the capabilities to address single switch molecules both in their ON and OFF conductance states, to predict when each molecule will be active as a switch and to control their switching properties.
By applying ACSTM imaging to systems of isolated OPE switch molecules, we are able to locate and to predict which molecules are likely to be active either through switching or through exhibiting motion, and we are able to switch the conductance state of single molecules while simultaneously mapping the microwave signal to confirm the location of each switch (even when it is not visible in STM topography).
A variety of techniques has been used to study fully conjugated OPE molecules for use as molecular wires and molecular switches (Chapter 2)
[2,24,25,106,107,109,110,141,155,169,182,185,224,225]. When single OPE molecules are isolated within host alkanethiolate SAM matrices and imaged using STM, they exhibit bistable conductance switching, defined as ON when a molecule appears to protrude from the host matrix, and OFF when the molecule is less protruding or no longer visible within the host matrix. The apparent height of the molecule depends on the thickness of the alkanethiolate SAM and the location of the molecule within the matrix
[224]. The observed switching is due to a convolution of the physical height of the molecule and the conductance of the molecule [24,25]. Previously, we showed that 128 conductance switching was regulated by the packing and chemistry of the host matrix
[24,25,106,107], that the OPE molecules exhibited motion within the host matrix
(Chapter 4) [224], and that the switching and motion events occurred on time scales faster than those of STM imaging (Chapter 5) [225]. Several hypotheses for the observed conductance switching were tested by modifying the molecular design of the OPE moieties [109]. The only mechanism consistent with our and others’ results is a change occurring at the substrate-molecule interface, which we posit to be a change in hybridization between the substrate and the molecule via a molecular tilt (Chapter 3)
[24,25,106,107,109,110]. Theoretical data likewise suggest that changes in the Au- molecule bonding and the Au-Au arrangements influence molecular conductance
[145,251].
Molecular polarizability is the relative susceptibility of the electron cloud around a molecule to be distorted by an external electric field. Polarizability (α) is the ratio between the measured induced dipole moment (p) and the applied electric field (E) as:
p = αE. 6.1
The microwave frequencies applied in our ACSTM tunneling junction are able to distort the electron cloud of the molecules present in the junction. By measuring the microwave difference frequency (MDF) signal passed with the tunneling current, compared to the reference MDF signal from the applied microwaves, we posit that we are measuring the relative differences in the polarizability of molecules within our tunneling junction. 129 6.1.1 Sample Preparation
Methods for sample preparation and insertion have been discussed in Chapters 1 and 3 [24,25,109,229]. Here, we created SAMs for polarizability analysis with microwave-frequency ACSTM. Alkanethiolate monolayers (C12 and C8) were adsorbed for 24 h from 1 mM ethanolic solutions. Mixed alkanethiolate monolayers were formed by creating a SAM of C12 by solution deposition for 24 h followed by vapor annealing octanethiol into the preformed SAM for 1 h at 80 °C. Vapor annealing was used rather than coadsorption to create domains of single species rather than mixed domains [29].
The vapor-annealing process is described in Chapter 1.
For 3-mercapto-N-nonylpropionamide (1ATC9) monolayers, we used 1 mM ethanolic solutions and the deposition time was increased to 48 h [32,33]. When the
1ATC9 monolayers were formed at room temperature, domains containing different tilts were observed as different apparent heights (~1.2 Å height difference) in STM images
(Figure 6.1A137H ). This is consistent with surface infrared spectroscopy data from
Whitesides, Nuzzo and coworkers, which suggest that monolayers containing buried amide functionalities assemble with tilts variable from normal to the surface to 18° tilted to maximize hydrogen bonding interactions [271]. When the 1ATC9 monolayers were formed at elevated temperatures (80 °C, 1 mM ethanolic solutions, 48 h), only one tilt
domain was observed (Figure 6.1B138H ). From this we infer that the different tilt domains imaged for the room-temperature samples were not due to sample contamination. Self- assembled monolayers formed for unfunctionalized OPE monolayers were adsorbed for 130
A: Solution Held at Room Temperature B: Solution Held at 80 °C 5 Ångstroms
250 Å 250 Å 0 Figure 6.1: Self-assembled monolayers of 1ATC9 molecules formed from 1 mM ethanolic solutions held at A. room temperature and B. 80 ºC for 48 h. The room temperature samples contained two tilt domains imaged as different apparent heights. The SAM deposition at elevated temperatures resulted in only one tilt domain.
131 24 h under a nitrogen environment from 1 mM anhydrous THF solutions. Since monolayers of nitro-functionalized OPE molecules are difficult to form and to image with STM held at ambient conditions due to their instability in air [230,272], we used insertion techniques to study the nitro-functionalized OPE molecules.
Briefly, to insert the nitro-functionalized OPE molecules, 0.1 μM solutions were prepared in anhydrous THF under a nitrogen environment. Gold substrates with pre- assembled SAMs (24 h deposition from solution) were placed in the THF solution for
~3 min allowing the conjugated molecules to adsorb to the Au{111} surface at the SAM defect sites [23,229]. Dodecanethiolate SAMs were used as the host matrices in this chapter [24,25].
6.1.2 Scanning Tunneling Microscopy
Please refer to Chapter 1 for a description of STM and ACSTM operation. Here,
SAMs and SAMs with inserted nitro-functionalized OPEs were analyzed using a custom- built AC scanning tunneling microscope operating under ambient conditions
[51,53,54,273]. All ACSTM spectra and images presented in this chapter used a difference frequency of 5 kHz. While this value can be changed, we have not observed any variation in our data as a result. 132 6.1.2.1 Alternating Current Scanning Tunneling Microscopy Imaging and Spectral Acquisition
Molecular polarizabilities were imaged using the AC scanning tunneling microscope in which two microwave-frequency AC signals (in the range 0.5–20 GHz), offset by a difference frequency (5 kHz), were mixed and combined with the DC bias voltage. The current used for the STM imaging feedback loop is dominated by the DC bias, with small contributions from the rectified AC signals. Applying two AC frequencies into our nonlinear tunneling junction allowed us to measure the heterodyned
AC signal at the difference frequency. We recorded the signals due to the applied microwave frequencies through the current preamplifier (bandwidth ~30 kHz) [23,274].
To obtain microwave spectra and images, we used a LIA referenced at the difference frequency to analyze the tunneling current, which carried the heterodyned microwave frequency information. The magnitude of the MDF signal is dependent on both the polarizability of the molecules and the STM tip/sample junction; therefore, the polarizability MDF images are, at present, qualitative measurements.
6.1.3 Single Molecules and Microwave Magnitude Extraction
Please refer to Chapter 3 for our procedure for tracking and extracting single- molecule measurements from a series of STM images. We use the same procedure here for extracting single switches from our MDF signal. To create the ratio of the nitro- functionalized OPE MDF signal to the C12 MDF signal, shown as ratio vs. time measurements, we divided the highest pixel value from each extracted image for the 133 inserted nitro-functionalized OPE switches by the mean (average of 25 pixels) of the
C12 polarizability extracted from an area next to the inserted molecule. This differs from our extraction of topographic apparent height, in which we take a median value for the inserted switch apparent height. For MDF images of switch molecules, we typically observe the molecule to have a small number of pixels (< 5) that are a larger value
(~20 fA) than that of the rest of the MDF image of the molecule. This leads to higher variability of the MDF amplitude for molecules in the ON conductance state.
6.2 Results and Discussion
6.2.1 Probing the Polarizability of Self-Assembled Monolayers
We probed the polarizability of the SAM molecules shown in Figure 6.2A139H by forming SAMs of C8, C12, 1ATC9 and OPE molecules. We further related the polarizability to conductance switching by inserting single nitro-functionalized OPE
switch molecules into host C12 SAMs. Figure 6.2B140H displays a schematic of this insertion with a single-molecule switch bound in a defect site of a host C12 SAM on Au{111} within the ACSTM tunneling junction.
The polarizability of each molecule, listed in Table 6.1141H , was calculated by
Gaussian 03 software using density-functional theory (DFT) with a 6-31G* basis set
[275]. The values listed are for the free thiol of each molecule, and we expect these values to change when the molecule is chemisorbed as a thiolate on a Au{111} surface
[272]; however, the values should give a qualitative comparison between the molecules
134
A octanethiol: SH dodecanethiol: SH 3-mercapto-N-nonylpropionamide: O N SH H 4,4'-di(ethynylphenyl)-1-benzenethiol: SH
4,4'-di(ethynylphenyl)-2'-nitro-1-benzenethiol: NO2 SH
B e-+ e-
Figure 6.2: A. Schematics of molecules used in this study. B. Schematic of a STM tip- sample junction containing self-assembled molecules of a C12 host SAM with a single switch molecule inserted at a defect site in the host SAM. The microwave frequencies were applied from the STM tip, measuring the polarizabilities (indicated by the color surrounding the molecules) of the self-assembled molecules.
135 as given by the ratio of each calculated polarizability compared to that of the dodecanethiol. We normalized to the calculated dodecanethiol values since this molecule was used for the host matrix for the inserted nitro-functionalized OPEs.
For comparison to the experiments and calculations we performed, Yeganeh and
Ratner calculated polarizability values for each of our molecules using DFT with a
++ ** 6-31 G basis set (Table 6.2142H ) [276]. They included calculations for molecules aligned normal to the surface and in tilted conformations, as well as for both anion and cation states for the nitro-functionalized OPE molecules. As discussed in Chapter 3, we do not believe we have enough energy in our system to ionize the molecules directly, but it may be that the ground states of (switch) molecules in different tilt conformations have different oxidation states.
We obtained microwave frequency response spectra for our Au{111} substrates
and for the SAM molecules within our ACSTM tunneling junction (Figure 6.3143H ). These spectra are of the MDF amplitude vs. the applied microwave frequency scanned from
0.5–20 GHz in 50 MHz steps. The microwave spectra recorded over the Au{111}
substrate with no SAM molecules (Figure 6.3A144H ) display the standing waves within our
ACSTM tunneling junction. These standing waves are inherent to the instrument and appear in all of our microwave spectra. Please refer to Chapter 1 for further information on these standing waves. When analyzing the microwave spectra presented here, we analyzed the relative differences in the magnitude of the spectra taken over the different
SAM molecules, not specific peaks within the spectra.
136
Molecule α (a.u.) ratio to C12 octanethiol 138 0.69 dodecanethiol 200 1.00 1ATC9 210 1.05 OPE 607 3.04 nitro-OPE 815 4.08 Table 6.1: Calculated polarizabilities (α) in atomic units, along the molecular axis from an electric field applied to the same axis for each self-assembled molecule and the ratio of each polarizability compared to dodecanethiol, since this was the host matrix for the inserted nitro-functionalized OPE molecules. Polarizabilities were calculated by Gaussian 03 software, with DFT using a 6-31G* basis set [275].
137
Molecule α (a.u.) ratio to C12 octanethiol Normal 158 Tilted (30°) 143 0.76 dodecanethiol Normal 199 Tilted (30°) 189 1.00 1ATC9 Normal 241 1.28 Tilted (18°) 232 1.23 OPE Normal 810 4.29 Tilted (30°) 643 3.40 Tilted (45°) 475 2.51 nitro-OPE (neut., neut. geom.) Normal 871 4.61 Tilted (30°) 690 3.65 Tilted (45°) 508 2.69 (neut., anion geom.) Normal 918 4.86 Tilted (30°) 725 3.84 Tilted (45°) 531 2.81 (anion, neut. geom.) Normal 1289 6.82 Tilted (30°) 1005 5.32 Tilted (45°) 722 3.82 (anion, anion geom.) Normal 1127 5.96 Tilted (30°) 882 4.67 Tilted (45°) 639 3.38 (cation, neut. geom.) Normal 2393 12.66 Tilted (30°) 1827 9.67 Tilted (45°) 1262 6.68 (cation, cation geom.) Normal 2220 11.75 Tilted (30°) 1698 8.98 Tilted (45°) 1176 6.22 Table 6.2: Polarizability calculations in atomic units by Yeganeh and Ratner using DFT with a 6-31++G** basis set [276]. The values are normalized to the polarizability of the dodecanethiol molecule tilted 30°. Note the cation and anion calculations for the nitro- functionalized OPE molecules. 138
Figure 6.3B145H shows the microwave spectra for two different chain length alkanethiolate species, C12 and C8. These spectra show small differences (~1-2 fA) in their microwave magnitude with the C12 showing a slightly larger response throughout this frequency range. The microwave spectra for the 1ATC9 monolayers showed different magnitudes for the different tilt domains within the room temperature sample
(Figure 6.3C146H ). Here, the more protruding regions of the 1ATC9 monolayer resulted in the smaller MDF amplitude response when compared to the more depressed 1ATC9 monolayer regions, which resulted in the larger MDF amplitude response. For the unfunctionalized OPE molecules, we expected a higher response from the MDF signal compared to the alkanethiolate and 1ATC9 SAMs since the OPE molecule is fully conjugated with electron delocalization. Since polarizability is the ability to distort the electron cloud around the molecule, we would expect a molecule with (more) delocalized orbitals to have a larger polarizability. The MDF signal for the OPE molecules
(Figure 6.3D147H ) does display a larger amplitude than those of the alkanethiolate and
1ATC9 SAM molecules. This is consistent with the microwave signals measuring the polarizability of the molecules.
Figure 6.4148H displays three different monolayers imaged using ACSTM and the corresponding schematic of each monolayer. To compare the polarizabilities of different
length alkanethiolate SAMs (shown schematically, Figure 6.4A149H ), we formed a C12 monolayer and vapor-annealed it with octanethiol, thereby forming domains of C12 and
C8 [29]. The topographic (Figure 6.4B150H ) and MDF polarizability (Figure 6.4C151H ) images for a C12/C8 separated monolayer were recorded and a small contrast was observed between
139
30
) A Au{111} substrate rms
fA 20
10 Amplitude (fA Amplitude Difference Frequency
00 2 4 6 8 10 12 14 16 18 20 Frequency (GHz) 30
) B dodecanethiolate rms octanethiolate 20
10 Amplitude (fA Amplitude Difference Frequency
00 2 4 6 8 10 12 14 16 18 20 Frequency (GHz) 30
) C 1ATC9 - depressed rms 20 1ATC9- protruding
10 Amplitude (fA Amplitude Difference Frequency
00 2 4 6 8 10 12 14 16 18 20 Frequency (GHz) 30
) D OPE rms 20
10 Amplitude (fA Amplitude Difference Frequency
00 2 4 6 8 10 12 14 16 18 20 Frequency (GHz) Figure 6.3: Microwave spectra (0.5–20 GHz) for A. a nominally bare Au{111} substrate, B. alkanethiolate SAMs on Au{111}, C. 1ATC9 SAM on Au{111}, and D. unfuctionalized OPE SAM on Au{111}. 140 the C12 and C8 molecules in the MDF signal. This small contrast was expected from the
spectra recorded over the different length alkanethiolate chains (Figure 6.3B152H ), and since the C8 calculated polarizability is ~76% that of C12.
When we imaged monolayers formed at room temperature from 1ATC9 molecules containing buried amide functionalities that are able to hydrogen bond, we observed two apparent height domains in the images, presumably due to different tilts of
the 1ATC9 molecules (shown schematically, Figure 6.4D153H ). This observation is consistent with surface infrared spectroscopy data that measured the possible tilt angles between normal and tilted 18º from normal to the surface to enable hydrogen bonding of the amide functionalities [271]. When we imaged these monolayers, the apparent topographically
less protruding (tilted) domains (Figure 6.4E154H ) had the stronger MDF polarizability signal
(Figure 6.4F155H ) than the more protruding (normal to the surface) domains. Since the
1ATC9 molecules within the tilt domains are identical, and since their calculated polarizabilities have very little change when normal to the surface or tilted, we posit the difference in the polarizability between the apparent height tilt domains is due to either differences in the polarizability at the molecule-surface interface or due to monolayer effects not calculated in single molecule measurements [272].
Furthermore, we believe the MDF signal occurs due to the polarizability and is not a function of the tip-sample separation. The topographically less protruding regions of the 1ATC9 SAMs have the stronger MDF signal, while in the mixed alkanethiolate monolayers the more protruding C12 regions display a stronger MDF signal; thus, we posit the signal is due to the molecules within the junction, not the tip-sample separation.
141
4 C 20
A B ) MDF Ångstroms (fA rms
50 Å 0 0 D E 4 F 20 D ) MDF Ångstroms (fA rms
50 Å 0 0 G H 4 I 20 D ) MDF Ångstroms (fA rms
50 Å 0 0 Figure 6.4: A. Schematic, B. topographic image and C. MDF image of a separated SAM containing C12 and C8 formed by vapor annealing octanethiol into a C12 SAM. The topographically more protruding C12 molecules have a larger MDF magnitude. D. Schematic and E. topographic image of a 1ATC9 SAM. Molecules with buried amide functionalities assemble with varied tilts from normal to 18° from surface normal to allow for hydrogen bonding [271]. F. Microwave difference frequency image for 1ATC9 molecules. Note that the topographically less protruding 1ATC9 molecules have a larger MDF signal. G. Schematic of an OPE SAM with structure based on the model proposed by Liu and coworkers who have measured a tilt of less than 5° for the molecules in this monolayer [277]. H. Topographic image, and I. MDF image of an OPE SAM. This molecule has a larger polarizability than alkanethiolate molecules, and thus we observe a larger MDF signal. Imaging conditions: Vsample = +1.0 V, Itunnel = 1.0 pA, applied frequency = 2 GHz, difference frequency = 5 kHz, input power level = 10 dBm. 142 The monolayer packing of unfunctionalized OPE molecules has been described
previously and is shown schematically in Figure 6.4G156H with the phenyl rings on neighboring molecules aligned perpendicular to one another [230,277]. The OPE SAMs have been found to have less than a 5° tilt from surface normal [230,277]. From our polarizability spectra and calculations, we expect the unfunctionalized OPE molecules to have a polarizability ~3-4 times that of C12. When we imaged these monolayers
(Figure 6.4H157H ), we found the polarizability (Figure 6.4I158H ) to be ~3 times that of C12. Thus, we conclude that we are imaging features related to the polarizability of adsorbed molecules.
6.2.2 Buried Interface Dynamics of Oligo(phenylene-ethynylene) Switches
We are unable to create and to image full monolayers of the nitro-functionalized
OPE switch molecules due to their instability in air [230,272]. However, using the insertion scheme detailed above, we are able to relate the polarizability of OPE-based switch molecules to the dynamics of these switches. Time-lapse series of STM images
2 recorded over single areas (500-1000 ÅP ) enabled us to record and to extract switching behavior and polarizability responses for each molecule in each frame (~5 min/frame) over several (8-18) hours. The first frame from one series of images is shown in
Figures 6.5A159H and 6.5B160H , topography and MDF magnitude, respectively. Each inserted
OPE molecule has been numerically labeled to correspond to the extracted frames in
Figure 6.5C161H , which display the most dynamic 25 frames for each molecule. From these
143
AB
2 4 2 4 1 1
3 3 5 5 250 Å
070Ångstroms 17 0 MDF (fA ) C rms 1 2 3 4
5 Figure 6.5: Simultaneously acquired A. topographic, and B. MDF images of a C12 SAM with inserted nitro-functionalized OPE molecules. A series of 200 images was acquired (5 min/frame, 30 s between frames). C. 120 Å × 120 Å areas were extracted for each inserted nitro-functionalized OPE molecule in each image. The numbers for the extracted series correspond to the labels in A and B. The most dynamic 25 sequential frames are shown for each: 1 frames 1-25, 2 frames 90-115, 3 frames 85-110, 4 frames 1-25, and 5 frames 130-155. Molecules that were stable (did not switch or exhibit motion) had a constant MDF profile, as in 1. Molecules that switched OFF showed fluctuations in the MDF images before they switched, shown in 2 and 3. Note that in 3, the OPE molecule still appeared in the MDF images after the molecule switched OFF. Molecules that exhibited motion within the SAM showed fluctuations in the MDF images as in 4 and 5. Imaging conditions: Vsample = -1.0 V, Itunnel = 2.0 pA, applied frequency = 2 GHz, difference frequency = 5 kHz, input power level = 10 dBm. 144 frames, we were able to use the polarizability of each OPE molecule to determine if a molecule was dynamic, with changes occurring at the buried substrate-molecule interface. When a molecule remained stable (no switching or motion) throughout imaging
as in Figure 6.5C162H , molecule 1, we observed a constant height in each of the topographic images and a constant profile in each of the MDF image frames. The constant MDF profiles indicated that the molecule was unlikely to switch or to exhibit motion.
For molecules that exhibited switching from the ON to OFF conductance states during imaging, the MDF images indicated that each molecule was likely to switch or to exhibit motion. This was observed as fluctuations in the profiles of the MDF images before the switching event occurred, even though the topographic images showed no change, as was observed for molecules 2 and 3. Previously, using apparent height vs. time measurements, we were able to show that switching can occur on time scales faster than those of imaging; however, the STM FBL does not respond quickly enough to image these changes as they occur (Chapter 5) [225]. We expect that the frustrated rotations responsible for switching would have slew rates in the GHz range. Here, the heterodyned microwave signal can respond to events occurring at faster time scales, and thus can image motion and switching not captured by the FBL. However, this still was not fast enough to follow the rotations from one conductance state to another. When the MDF signal exhibited fluctuations, we posit that changes were occurring at the buried interface
(Au-thiolate bond) and the molecule was active and likely to switch conductance states or exhibit motion, as observed in topographic images.
We were further able to observe molecules that were initially in the OFF
conductance state, but were imaged using the MDF signal, as shown in Figure 6.5C163H , 4. 145 In the first frame for 4, the molecule did not appear in the topographic image, but a small signal was nevertheless present in the MDF image. After this molecule switched conductance states, it also exhibited motion up and down the substrate step edge at which it was located [224]. Motion of an OPE molecule within the SAM was further
demonstrated by molecule 5 in Figure 6.5C164H , which was not active as a switch, but did exhibit motion within the host SAM. Density-functional theory calculations have found the Au-thiolate complex to have a low barrier for diffusion (~0.07–0.10 eV, between hcp and fcc sites), lower than that of a bare thiolate radical or of a single gold adatom [278].
In the microwave images for molecules 4 and 5, we found fluctuating profiles in the microwave images, similar to those for the molecules that switched. Thus, the fluctuating profiles in the MDF signals indicate an active molecule undergoing dynamic changes at the buried interface.
The ratio of the OPE MDF signal to the C12 MDF signal vs. time for each
molecule in Figure 6.5165H is given in Figure 6.6166H . The total occurrences vs. OPE:C12 ratio in
are summarized in Figure 6.6F167H . The colored lines in Figure 6.6168H correspond to red for the
ON state OPE:12 ratio (6.8, a value similar to that calculated for the nitro-functionalized
OPE as an anion with a neutral geometry), blue for the OFF state OPE:12 ratio (3.3), and black for where the OPE is not distinguishable from the C12 (1.7). Note that we expect a ratio of 1:1 when the OPE polarizability is not above that of the C12; however, since our calculation determines the maximum extracted value divided by the average of
the background, this value is slightly greater than 1. For molecule 1 (Figure 6.6A169H ), while there were fluctuations in the MDF signal, it was relatively constant at the ON state
146
15 A Molecule 1 10
5
0 B 15 Molecule 2 10
5
0 15 C Molecule 3 10
5
0 15 D Molecule 4 10
5 Ratio of OPE MDF Signal to C12
0 15 E Molecule 5 10
5
0 0 100 200 300 400 500 620 700 800 900 1000 Time (min) F 120
100
80
60
40 Occurrence
20
0 1 2 3 4 5 6 7 8 9 10 11 Ratio of OPE MDF Signal to C12Width MDF Signal Figure 6.6: Ratio of the OPE molecule MDF signal to the C12 MDF signal vs. time for each molecule (1-5) in Figure 6.5170H . The red lines in the spectra are the ratio for switches in the ON state, blue lines are for ratios for switches in the OFF state still visible with the MDF signal and the black lines are the amplitude where molecules blend into the background in the image. A. Molecule 1 was stable throughout imaging. B-D. Molecules 2, 3 and 4 have fluctuating MDF signals before they switch to the OFF state. Molecule 3 was visible above the background in the MDF images after it switched. E. Molecule 5 exhibited motion within the host SAM. F. Occurrences vs. ratio of the OPE MDF signal to the C12 MDF signal. 147 OPE:12 ratio. This was the molecule that did not switch or exhibit motion. Molecules 2,
3 and 4 all switched conductance during imaging (Figures 6.6B,171H 6.6C172H and 6.6D173H ). For
molecule 3 (Figure 6.6C174H ), when this molecule switched to the OFF conductance state, it was still observed in the MDF images through frame 180, but was not observed above the
background in the remaining 20 frames. Finally, for molecule 5 (Figure 6.6E175H ), while this molecule exhibited motion within the host SAM, its polarizability ratio fluctuates around
the ON state ratio. Figure 6.6F176H shows the occurrences vs. ratio of the OPE MDF signal to the C12 MDF signal. Note the large variability for the ratio of molecules in the ON conductance state (6.8 ± 3.6). This is mostly likely due to the variability in the signal caused by switching and motion events occurring on faster (GHz) time scales than those passed with our difference frequency (5 kHz). The ratio of the inserted molecules in the
OFF state still observed above the background (2.5 ± 1.0) and those no longer visible above the background (1.7 ± 0.3) are shown.
Figure 6.7177H displays extracted topographic and MDF signal frames for a variety of
inserted nitro-functionalized OPE molecules. In Figure 6.7A178H , two different molecules that remained stable in the ON conductance state are displayed. Molecules that remained stable in the monolayer had a constant MDF profile. For some molecules, when they switched from the ON to the OFF conductance states, they still produced a MDF signal at
a lower magnitude (Figure 6.7B179H ). However, for other molecules, when they switched between conductance states, in some frames the polarizability signal was observed above the background level of the matrix, while in others, the polarizability signal was not
above the background (Figure 6.7C180H ). 148
A: Molecules that remain ON
B: Molecules that switch OFF, polarizability signal remains
C: Molecules that switch, polarizability signal not always present 1
2
3
4
5
6
7
Figure 6.7: Extracted frames for several nitro-functionalized OPE molecules. A. Molecules that remained stable in the ON conductance state. B. Molecules that switched between the ON and the OFF conductance states and the MDF signal remained above the level of the C12 MDF signal in the OFF conductance state. C. Molecules that switched between the ON and the OFF conductance states, but the MDF signal was not always observed above the background level in the OFF conductance state. Imaging conditions: Vsample = -1.0 V, Itunnel = 1.0-2.0 pA, applied frequency = 2 GHz, difference frequency = 5 kHz, input power level = 10 dBm, extracted area = 40 Å × 40 Å. 149
For the molecules displayed in Figure 6.7C,181H 1-4, we observed the molecules to switch between conductance states as observed in topography, but the MDF signal was present in some, but not all of the frames when each of the molecules was in the OFF
conductance state. In molecule 5 in Figure 6.7C182H , we did not observe a molecule in the topography and we only observed the molecule in the MDF signal for the first five frames presented, after which the molecule was no longer in the topography or the MDF
signal. Finally, for molecules 6 and 7 in Figure 6.7C183H , in topography we observed the molecules to be initially in the OFF conductance state and to switch to the ON conductance state. No MDF signal was observed above the level of the C12 until the molecules appeared in the ON conductance state in topography. Note that the position of molecules 6 and 7 changed within each frame changes, indicating motion of the molecule within the host SAM.
At this point, we cannot explain why we do not observe the polarizability signal above the background for the inserted nitro-functionalize OPE switches in every frame when the molecules were in the OFF conductance state. However, we do have hypotheses for our observations. First, it is possible that the molecules are exhibiting motion within the host matrix [27,28,224]. If the molecule moves away from the STM tip, we would not capture the molecule’s polarizability signal. Second, it is possible that the molecule- substrate bond may have broken [110]. If the molecule-substrate bond broke, the molecule would likely desorb from the surface; however, the energy required to break the
Au-S bond (~1.6 eV) [13,279] was not proviced by our scanning bias voltage (1 V).
Since we are able to use the MDF imaging to measure the polarizability of the nitro-functionalized OPE molecules, even at times when the molecules did not appear in 150 the topographic images, we have attempted to switch molecules controllably. In a limited number of attempts, we could controllably switch the OPE molecules from the ON to the
OFF conductance states using the electric field of the STM junction without desorbing
the switches (in contrast to the conclusions in [110]). In Figures 6.8184H and 6.9185H , we show examples where we have used the electric field between the tip and the sample to switch the conductance state of the molecules [24,25]. This measurement has not been reproducible with a single set of conditions (i.e., varied electric field pulses have caused molecules to switch and to remain stable in a particular state) due to the switching activity being highly dependent on the matrix surrounding the monolayer
[24,25,106,107]. A more reliable method for mediated switching has been demonstrated using the electric field and hydrogen-bonding interactions [106,107]; however, for the purposes of this work, we have shown examples using only the electric field.
Controlled switching using the electric field of the scanning tunneling microscope is a slow and serial technique. In many cases, applying a voltage pulse to the STM tip would result in poor image quality after the voltage pulse; thus, a limited number of measurements have been performed. We have attempted to switch molecules from the
ON to the OFF conductance states over eight separate nitro-functionalized OPE molecules and we were successful for five of these attempts. To switch a molecule from the ON to the OFF conductance states, we first move the tip off center from the molecule followed by moving the tip closer (5-10 Å) to the SAM. Next, we used voltage pulses ranging between -2V to 2V with tunneling current between 1 pA to 10 pA with pulses durations ranging from 100 ms to 500 ms. In the unsuccessful attempts, we found that the molecule either did not exhibit switching, or we found the molecule was no longer 151 present in the topographic or microwave images. The inability to switch a molecule could result from steric hindrance around the molecule, consistent with our previous results
[24,25]. No longer observing the molecule using STM could result from desorbing the molecule from the surface, rather than switching the molecule, or the molecule moving through the SAM. We have attempted to switch molecules from the ON to the OFF states with either tip motion or a voltage pulse alone; however, these actions did not result in controlled switching events. We infer that the tip motion loosened the host matrix while the switching was influenced by the applied electric field from the voltage pulse.
Figure 6.8186H exemplifies a molecule switched from the ON to the OFF conductance
states. Here, to switch the molecule controllably (boxed in Figures 6.8A187H and 6.8B188H ), we moved the tip off center from the molecule and induced local disorder in the host matrix by contacting the STM tip to the SAM, followed by a voltage pulse [280]. The same area was imaged after the tip motions and voltage pulses. The manipulated switch molecule
appeared as a weak signal in the topographic images (Figures 6.8C,189H 6.8E190H and 6.8G191H ) and
the corresponding MDF signal magnitude and profiles (Figures 6.8D,192H 6.8F193H and 6.8H194H ) indicate that the molecule was active as a switch during the acquisition of these images.
In Figure 6.8I195H , the molecule signature no longer appeared in the topographic image;
however, we found that the MDF image (Figure 6.8J196H ) indicated that the nitro- functionalized OPE switch was still present after the switch molecule was stabilized in the OFF state with the magnitude of the MDF signal decreased. The topographic and
MDF signals for the other nitro-functionalized OPE switch appearing in the STM images
(Figure 6.8197H , unboxed switch) indicate that this was an inactive switch through imaging.
152
A BDC
100 Å 100 Å E FHG
100 Å 100 Å 6 40 I J ) Magnitude AC Ångstroms (fA rms
100 Å 0 0 Figure 6.8: Switching of a molecule from the ON to the OFF conductance states. Simultaneously acquired topographic and MDF images of C12 with inserted nitro- functionalized OPE molecules. After acquisition of A. and B. the tip was moved off center from the boxed molecule and the STM tip was moved 10 Å towards the sample five times (100 ms for each movement), followed by five voltage pulses from the tip (+2.0 V, 10 pA, 100 ms). This was done to “loosen” the SAM matrix to allow the molecule to switch from the ON to the OFF state. Note that a small topographic signature is observable in the topographic images C. E. and G. before the molecules settles into the OFF conductance state I. The MDF signature of the molecule in D. F. and H. indicates that the molecule is active as a switch. The MDF signal decreases in J. indicating that the molecule is stable in the OFF state. Imaging conditions: Vsample = -1.0 V, Itunnel = 1.0 pA, applied frequency = 2 GHz, difference frequency = 5 kHz, input power level = 10 dBm. Z-scales for all topography and MDF images are given in I. and J. Note that the maximum of the MDF signal has been decreased to display the MDF signal of the switch in the OFF conductance state. 153 Previously, we were limited to studying only molecules that appeared (at some point) in the ON conductance state during imaging, due to the difficulty in locating molecules continually in the OFF state in large topographic images [24,25]. Now, we have the additional capability of imaging molecules that are in the OFF conductance state and do not appear in topography.
Selectively switching a molecule from the OFF to the ON states was less reproducible than the reverse switching from ON to OFF. We attempted to switch ON twenty-one molecules and we were successful in eight attempts. Our conditions for switching from the OFF to the ON state used voltage pulses ranging between -2V to 2V with tunneling current between 1 pA to 10 pA with pulses durations ranging from 1 ms to
500 ms. Typically, it took several voltages pulses to switch a molecule from the OFF to
ON conductance state. No one set of conditions reproducibly resulted in switching from the OFF to the ON conductance state. The conditions appear to be highly dependent on the location of the molecule within the host matrix, again consistent with previous results
[24,25].
Figure 6.9198H displays a molecule that was switched from the OFF to ON
conductance state. In the initial topographic and MDF image (Figures 6.9A199H and 6.9B200H ), the signature for the boxed molecule appeared only in the MDF image. After
several successive frames (Figures 6.9C201H - 6.9F202H ) and voltage pulses slightly off center from the molecule, the molecule switched to the ON conductance state and the MDF
signal increased (Figures 6.9G203H and 6.9H204H ); however, the voltage pulses applied to the tip destabilized the tunneling junction [24,25].
154
AB
100 Å CD
100 Å EF
100 Å GH
100 Å
0 8 070 ( ) Ångstroms ACMagnitude fArms Figure 6.9: Switching from the OFF to the ON states for a switch found in the MDF signal. Sequential images of simultaneously acquired topographic and microwave magnitude images of C12 with inserted nitro-functionalized OPE molecules. After acquisition of each image, the tip was moved off center from the boxed molecule and a voltage pulse from the STM tip was applied (-1.0 V, 7 pA, 100 ms; before G-H, -1.5 V, 2 pA, 100 ms). Imaging conditions: Vsample = 1.0 V, Itunnel = 2.0 pA, applied frequency = 2 GHz, difference frequency = 5 kHz, input power level = 10 dBm. 155 6.3 Conclusions
Using ACSTM to obtain MDF images, we are able to image single-molecule polarizabilities and to correlate them with calculated values. For single inserted switch molecules, instability in the MDF signal was attributed to dynamics at the molecule- substrate interface, and we could predict which molecules were most likely to be active as switches or exhibit motion. From this, we better understand OPE switch molecules within their host environments. We can selectively switch molecules between conductance states and verify their location when in the OFF state though the MDF signal. Through increasing our understanding and imaging capabilities of chemical signals at buried interfaces, we can further control systems at the single-molecule level.
Further theoretical work needs to be performed on this system to understand the influence of the Au{111} substrate on the molecular polarizability. Initial calculations from Yeganeh and Ratner show that polarizability decreases when the molecules were attached to a Au{111} substrate at a 3-fold hollow site [276], and other bonding sites need to be explored. Furthermore, for the 1ATC9 molecules, we found a higher magnitude for the tilted domains, opposite that of the calculation. Thus, we posit the interactions between the molecules and of the molecule and the substrate lead to the differences between the experimental and the calculated values. Further calculations will need to be performed to address these issues.
156
Chapter 7
DEVELOPMENT OF CLUSTER CAPTURE SURFACES
7.1 Introduction
To measure the properties of single molecules and clusters to be used in single- molecule devices, we must be able to capture and to place these molecules and clusters on surfaces in specific locations. This chapter details our work towards creating surface that will be used to capture superatom clusters. These surfaces are being developed both to identify superatom clusters as having element-like properties and to immobilize clusters to a substrate for measurement with scanning probe microscopy.
Ionic species such as cyanide and thiocyanate have strong internal bonds, and in many cases they react as if they were a single halogen atom; thus, they are commonly referred to as pseudohaolgens. Similarly, clusters have been assembled that react as if they were single atoms [55,281-284]. These “magic” clusters have closed electronic
- shells leading to high stability and enhanced abundance. For example, an Al13 cluster has
40 valence electrons and using the Jellium model, forms a closed 2p shell [282-284]. It is
- expected that the Al13 clusters would act as halogens if reacted with electropositive species such as K, forming salt-like ionically-bound superatom complexes; however, binding with the electropositive species changes the electronic structure of the core 157 - influencing measurements with mass spectrometry [281,285,286]. The Al13 clusters have
- been reacted with HI, forming ionic Al13I polyhalides [281]. This led to the
- characterization of the Al13 core as having halogen properties [281]. In addition to halogen-like clusters, clusters of Al14 have been characterized as having alkaline-earth- metal-like properties [55]. Since these clusters have element-like properties, we are developing cluster capture surfaces that will test if the clusters react as do elements of specific groups of the periodic table.
The key test of the existence of a new element has always been to verify its chemistry. In the 1940’s, Seaborg and coworkers were comparing the chemistry of the transuranium element to that of other elements of the periodic table [287-291]. They used precipitation reactions to study the chemistry of each new element. For example, when using microscopic amounts of radioactive isotopes, the radioactive species were considered cationic if they could be “carried down” by forming a macroscopic precipitate with an excess of an anionic species [291]. If the radioelement was not “carried down”, i.e., no precipitate was formed, the compound was assumed to be relatively soluble.
Through precipitation reactions, the reactivities of the transuranium elements could be compared by the compounds they formed with the elements of other periods [287-291].
In analogy to this, we are developing the capability to measure the chemistry of superatoms and to compare these to their predicted (element-like) properties [55,281].
Toward this end, we are developing and applying patterned SAM substrates of isolated tether groups inserted into unreactive background matrices as reactive probes for superatoms and cluster assemblies. Previously, we have employed 1,10-decanedithiol tether molecules, inserted from solution into the defects of insulating host SAM matrices, 158 to isolate and to stabilize undecagold clusters on substrates and to measure the cluster assemblies’ electronic properties using an ultrastable extreme-high vacuum low temperature scanning tunneling microscope operating at cryogenic temperatures (4 K)
[121]. As a first step for capturing and measuring superatom clusters, we have created
SAMs containing isolated tether molecules that should be able to capture Ba2+ ions through electrostatic binding capture to test the reactivity of the surface to a specific ion.
Patterned electrostatic capture surfaces have been created using μCIP discussed in
Chapter 1 [26,292]. The advantages of using μCIP include precise control of the isolation and dilution of the inserted molecules, prevention of lateral surface diffusion of the inserted molecules, and the ability to place molecules selectively in patterned areas without solutions that are damaging to the preexisting SAM. The advantage of inserting single, isolated capture groups is two-fold. First, by isolating molecules within a host
SAM, we selectively and controllably define the physical and chemical environment for the captured clusters [121]; second, we are able to isolate single captured clusters from each other, thereby enabling measurements of single clusters.
7.2 Experimental Procedure
7.2.1 Sample Preparation
Our sample preparation is shown schematically in Figure 7.1205H . Inserted SAMs
were fabricated by μCIP (Chapter 1) on Au{111} substrates (Figure 7.1A206H ) that were annealed using a hydrogen flame just prior to deposition. All ethanolic solutions were
159
A
Exposure to SAM solution B
Tether molecule μCIP
C
Exposure to Ba solution
D
Figure 7.1: Schematic of the capture surface assembly. A. A Au{111} substrate is exposed to an C8 SAM solution. B. The C8 SAM with inherent defect sites is inserted with MUDA or sUPA molecules using μCIP (MUPA shown here). C. The substrate with 2+ the SAM and tether molecules is exposed to a BaTO or Ba(OH)2 solution. D. Ba ions are captured to the surface via the tether molecules. 160 sparged with nitrogen for 20 min prior to use. Octanethiolate SAMs were assembled onto
Au{111} substrates from 1 mM octanethiol ethanolic solutions for 1 min, rinsed with
neat ethanol and blown dry with nitrogen (Figure 7.1B207H ). To ink the PDMS stamps, 1 mM ethanolic solution of MUDA or 11-mercaptoundecylphosphoric acid (MUPA) were pipetted onto the stamps, allowed to adsorb for 1 min, and dried with under a stream of nitrogen. The molecular-inked PDMS stamps were brought into conformal contact with the Au{111} substrates containing the preformed C8 SAMs for 30 min; the MUDA (or
MUPA) molecules transferred from the stamp, inserting into defect sites in the C8 SAM
only in regions where the stamp and substrate were in contact (Figure 7.1C208H ). A cylindrical metal weight of ~55 g was placed on the back side of the stamp to ensure conformal contact. After insertion, the Au substrates were rinsed with neat ethanol and blown dry with nitrogen. The Au{111} substrates with the patterned MUDA (or MUPA) and C8 SAM were exposed to an ethanolic solution of 50 mM barium tetrahydrofurfuryl oxide (BaTO, Sigma Aldrich) or a saturated aqueous solution of barium hydroxide
(Ba(OH)2, Sigma Aldrich) for 5 min, rinsed with neat ethanol and blown dry with
nitrogen (Figure 7.1D209H ). As a control, we also exposed C8 SAMs without tether molecules to solutions of the BaTO and Ba(OH)2 so that we could characterize the nonspecific adsorption. All samples were characterized within 24 h by X-ray photoelectron spectroscopy (XPS) or time-of-flight secondary ion mass spectrometry (ToF-SIMS). 161 7.2.2 Time-of-Flight Secondary Ion Mass Spectrometry
The ToF-SIMS data were acquired using a Kratos Prism ToF-SIMS spectrometer.
A ToF-SIMS image is generated using primary ions from an ion source, which are rastered across a surface. The secondary ions produced from this bombardment are accelerated from the surface to a time-of-flight mass spectrometer where they are detected by a microchannel plate detector at each pixel [293]. A single-ion image is created by plotting the intensity of a selected mass/charge ratio from the total ion mass spectrum at each pixel. The ToF-SIMS used is equipped with a 15 kV indium liquid metal ion source (FEI, Beaverton, OR) oriented 45° to the sample. The ion source was focused to a beam size of ~500 nm and delivered current of 1 nA to the sample with a
50 ns pulse width. The sample was biased at +2.5 kV and an extraction lens biased at
-4.7 kV collected the secondary ions, which traveled along a 4.5 m flight path before being detected at a microchannel plate detector (Galileo Co., Sturbridge, MA). Imaging was performed by electrostatically rastering the primary ion beam across the sample and collecting a mass spectrum at every pixel in the 128 pixel x 128 pixel image.
7.2.3 X-Ray Photoelectron Spectroscopy
X-ray photoelectron spectroscopy uses a monochromatic X-ray source to eject electrons from the core levels of atoms into vacuum [294,295]. Core electrons are the inner quantum electron shells of an atom, and do not participate in chemical bonding. For each element, there is a characteristic binding energy (Eb) associated with each inner shell electron. If the energy of the incoming X-rays (hv) is larger than Eb, the excess energy is 162 converted to an emitted photoelectron with a specific kinetic energy (Ek). This kinetic energy can be measured with respect to the Fermi level of a metal standard (Φ is the work function of this metal standard), and thus the Eb can be determined by:
Eb = hv – Ek – Φ, 7.1
By measuring the Ek distribution, the elemental composition of a surface can be determined. The measured Ek (and Eb) is dependent on the bonding environment of the atom. Charge transfer can leave atoms with partial positive charges, leading to a lower Ek shift in the core levels due to the increased Coulombic attraction between the nucleus and the core electrons. Only core electrons, originating within tens of Ångstroms of the surface will be ejected into vacuum and measured by XPS. By increasing the incoming angle of the X-rays with respect to the surface normal, the surface sensitivity of XPS can be increased.
X-ray photoelectron spectra for the initial charge capture surfaces were obtained using a Kratos Axis Ultra photoelectron spectrometer equipped with a monochromatic
Al Kα source (1486.6 eV), and a base pressure of 1 × 10-9 torr, with a spot size of
300 μm × 700 μm. This instrument is located at the Penn State Nanofabrication Facility.
Survey spectra were acquired at a pass energy of 80 eV and high-resolution spectra were collected at a pass energy of 20 eV.
X-ray photoelectron spectra for the capture surfaces were also obtained using a
Scienta ESCA-300 photoelectron spectrometer equipped with a monochromatic Al Kα source (1486.6 eV). The x-ray spot is a line >5 mm long and >1 mm wide. This instrument is located at Lehigh University and was made available through collaboration 163 with Drs. Bruce Koel and Alfred Miller. All spectra were obtained at a 10° tilt angle with respect to the surface plane. Survey spectra were acquired at a pass energy of 300 eV and high-resolution spectra were collected at a pass energy of 75 eV.
The peaks were referenced to the Au 4f7/2 peak at Eb = 83.98 eV. All the peaks from the spectra were fit using GL line shapes (CasaXPS analysis software [296]). The
Au 4f peaks were fit with a fixed area ratio of 3:4, the S 2p peaks were fit with a fixed area ratio of 2:1 and with a fixed energy separation of 1.18 eV, and the P 2p peaks were fit with a fixed area ratio of 1:2 and with a fixed energy separation of 0.84 eV. The errors associated with all peak fits were 0.1 eV.
7.3 Results and Discussion
7.3.1 Electrostatic Capture Surfaces
The Al14 clusters, studied by Castleman and coworkers, were found to behave as alkaline-earth-like-metals [55]. Alkaline-earth metals ionize to form dicationic species, which form salts with anionic species. To emulate the experiments performed by Seaborg and coworkers detailed above [287-291], we have created electrostatic capture surfaces, which present carboxylate and phosphate pendant functionalities for electrostatic binding.
As a model for the superatom cluster capture, we have captured Ba2+ ions to these surfaces.
For our initial electrostatic capture surfaces, we chose MUDA as our capture tether molecules because it is a system with which we are familiar through the molecular rulers process where we use alternating layers of mercaptoalkanoic acid molecules and 164 Cu ions, building multilayers on a substrate [204-207]. The surfaces for the electrostatic capture were exposed to BaTO solutions. This capture scheme allows us to isolate the clusters from the substrate with the SAM, but also tether the molecules in place for imaging and measurement with STM, so that the clusters are not swept from the imaged area [109].
Using a PDMS stamp with 25 μm × 25 μm posts, we inserted MUDA molecules into an C8 SAM and imaged the patterned SAM using ToF-SIMS. The mass spectra from the ToF-SIMS displayed representative mass peaks for the Au substrate and the monolayer [297,298], as well as the Ba peak with a mass to charge ratio (m/z) of 138 corresponding to the 138Ba+ ion (and surrounding peaks with an isotropic distribution observed in previous experiments) [299]. The ToF-SIMS image mass selected for Ba
(Figure 7.2A210H ), displays features consistent with the size and shape of the PDMS stamp posts used to create the patterned MUDA regions in the host SAM. The high intensity
(red) regions correspond to area with more 138Ba+ counts. While the image is consistent with our pattern, the high intensity regions are larger than the size of the PDMS posts, indicating nonspecific adsorption occurred, yet the Ba ions were more localized in the regions where MUDA molecules were inserted. We believe this larger feature size is due to one or more of: an increase in the non-specific adsorption of Ba2+ ions surrounding the patterned features, lateral dissolution of Ba2+ ions because they are only electrostatically
bound, and stamp deformation during patterning. Figure 7.2B211H shows a line scan (blue
138 + line from Figure 7.2A212H ) where the intensity of the Ba signal was plotted as a function of position. Nonspecific adsorption of Ba2+ was observed across the substrate, which was consistent with the XPS data. 165
A B 2400
2100
1800
1500
1200
900
600 Signal Intensity (counts) 0 20 40 60 80 100 120 25 μm Distance (μm) Figure 7.2: A. Time-of-flight SIMS image mass selected for 138Ba+ (m/z = 138) of an C8 SAM with inserted MUDA stamped by μCIP for 30 min, then exposed to a BaTO ethanolic solution for 5 min. The PDMS stamp used in the μCIP had 25 μm × 25 μm posts. The higher intensity regions (red) correspond to higher ion counts, and the lower intensity regions (black) correspond to lower ion counts. The line across the image (blue) corresponds to the linescan in B. B. Signal intensity (counts) vs. position plot for the 138Ba+ ions in A. The 15 keV In+ dose was 3.7 × 1012 ions/cm2 with 72 ion pulses per pixel.
166
ABKinetic Energy (eV) Kinetic Energy (eV) 200 400 600 800 1000 1200 1400 700 710 720
Survey Ba 3d5/2 Au 4f Au C8 + MUDA + Ba 5/2 Intensity Intensity 3/2 3/2 Au 4d Au 1/2 Au 4d Au Au 4p Au
1/2 C8 + Ba Au 4p Au Au 4s Au O 1s C 1s S 2p Au 5p Au 1200 1000 800 600 400 200 0 790 780 770 Binding Energy (eV) Binding Energy (eV) Figure 7.3: A. X-ray photoelectron survey spectrum of the electrostatic capture surface using MUDA inserted into an C8 SAM exposed to BaTO. B. High-resolution spectra of the Ba 3d5/2 region with the MUDA inserted C8 SAM exposed to BaTO (upper) and an C8 SAM (no insertion) exposed to BaTO (lower). The spectra are offset for clarity.
167
Figure 7.3A213H displays an XPS survey spectrum for the MUDA-inserted C8 SAM exposed to BaTO. The major peaks displayed within the spectrum have been labeled.
Note that the strongest signals we observed were from the Au{111} substrate. In
Figure 7.3B214H , the high resolution Ba 3d5/2 region of the XPS spectra are shown for the
MUDA-inserted C8 SAM exposed to BaTO (upper) and an C8 SAM (no insertion) exposed to BaTO (lower). These spectra show that Ba has adsorbed to the surface, even when no tether molecules were present. This is consistent with the ToF-SIMS data, suggesting that nonspecific adsorption occurs.
Table 7.1 displays the atomic concentrations and atomic ratios calculated from the
Ba 3d, C 1s, S 2p, O 1s, and Au 4f high resolution spectra acquired for both the control
C8 SAM exposed to BaTO and the MUDA-inserted C8 SAM exposed to BaTO. The concentration of the Ba increased for the MUDA-inserted SAM; however, we again observe that the Ba2+ ions were nonspecifically adsorbed on the SAM with no tether molecules. We also observed small amounts of O present in both samples. As discussed below, we attribute this to solvent molecules trapped within the monolayer. The O concentration increased for the MUDA-inserted SAM. This was due to the O within the
MUDA molecules as well as the hydrophilicity of the MUDA molecules.
7.3.2 Minimizing Nonspecific Adsorption
The BaTO depositions detailed above were performed from ethanolic solutions.
These solutions were able to wet the entire SAM surfaces, and thus lead to the
168
Atomic Concentration % Atomic Ratios C 1s C 1s C 1s C 1s C 1s S 2p O 1s Au 4f Ba 3d 5/2 S 2p Au 4f O 1s Ba 3d5/2 C8/Ba 24.8 2.9 0.4 71.8 0.08 8.7 0.35 62 310 C8/MUDA/Ba 24.7 2.6 1.2 71.3 0.14 9.6 0.35 23 176
Table 7.1: Atomic concentrations and ratios measured from XPS for an C8 SAM exposed to BaTO and a MUDA-inserted C8 SAM exposed to BaTO.
169 nonspecific adsorption observed. To minimize the nonspecific adsorption of the Ba2+ ions across the surface, we used aqueous solutions of Ba(OH)2 instead of the ethanolic solutions of BaTO. Since the SAM molecules contain a hydrophobic methyl pendant functionality, we expected the monolayer to repel the aqueous solution except in areas where capture molecules were inserted. We also used a second type of electrostatic capture molecule (MUPA), which contains a phosphoric acid pendant functionality that has two possible deprotonation sites. We expect the Ba2+ ions from the strong base
Ba(OH)2 to react with the weak phosphoric acid. We also exposed the MUDA-inserted
SAMs, discussed above, to Ba(OH)2 solutions for comparison, and we exposed C8 SAMs to Ba(OH)2 solutions, as a control, to understand if the strong base would disrupt the
SAMs. As an experimental observation, we noted that the C8 SAM without inserted molecules exposed to the aqueous Ba(OH)2 solutions floated when placed in the solution, and thus we had to hold these samples in the solution to expose the uninserted SAMs to the Ba(OH)2.
As with the initial MUPA/BaTO capture surfaces, we first used ToF-SIMS to characterize our surfaces to ensure that our capture scheme was functioning. We performed μCIP using a PDMS stamp with 25 μm × 25 μm posts to insert MUPA into
C8 SAMs and exposed these to saturated Ba(OH)2 aqueous solutions. Figure 7.4A215H displays the 138Ba+ mass-selected ToF-SIMS image. The high intensity regions (red) correspond to more ion counts compared to the low intensity (black) regions. We observe higher intensity regions corresponding to our PDMS stamp pattern, showing that Ba2+ had a greater preference for adsorbing to regions with the inserted MUPA. The areas
170
A B 450 400 350 300 250 200 150 100
Signal Intensity (counts) 50 0 20 40 60 80 100 120 25 μm Distance (μm) Figure 7.4: A. Time-of-flight SIMS image mass selected for 138Ba+ (m/z = 138) of an C8 SAM with inserted MUPA stamped by μCIP for 30 min, then exposed to a Ba(OH)2 aqueous solution for 5 min. The PDMS stamp used in the μCIP had 25 μm × 25 μm posts. The higher intensity regions (red) correspond to higher ion counts, and the lower intensity regions (black) correspond to lower ion counts. The line across the image (blue) corresponds to the linescan in B. B. Signal intensity (counts) vs. position for the 138Ba+ ions in A. Note the low intensity in the unpatterned regions as compared to + 12 2 Figure 7.2A216H . The 15 keV In dose was 3.7 × 10 ions/cm with 72 ion pulses per pixel.
171 with Ba2+ exposure were again larger than the stamped regions, as shown in the signal
intensity vs. position data (Figure 7.4B217H ). These enlarged feature sizes could again be due to nonspecific adsorption of the Ba2+ ions, lateral dissolution of Ba2+ ions, or stamp deformation during patterning.
We performed XPS analyses on four different SAMs to understand the interactions that are occurring in this binding system. We used an C8 SAM that was assembled from solution for 1 min as a reference, an C8 SAM exposed to aqueous
Ba(OH)2 to understand nonspecific adsorption, a MUDA-inserted C8 SAM exposed to aqueous Ba(OH)2, and a MUPA-inserted C8 SAM exposed to aqueous Ba(OH)2 to understand our binding scheme.
The XPS analyses detailed below were performed at Lehigh University using a state-of-the-art Scienta ESCA 300 (VG Scienta Inc., Newburyport, MA). Here, the samples were measured at a 10° tilt with respect to the surface plane, resulting in higher
sensitivity to the SAM species, as observed in the survey spectrum of Figure 7.5A218H as
compared to Figure 7.3A219H . In Figure 7.5B220H , the high-resolution P 2p region of the XPS spectra is shown for the MUPA-inserted C8 SAM exposed to Ba(OH)2. This peak is at
Eb = 133.8 eV, which indicates a metaphosphate (~134 eV) [295]. A metaphosphate is salt or an ester of metaphosphoric acid, HPO3, consistent with MUPA, which is an ester, or MUPA in a salt form with the Ba. The binding energy is ~1 eV higher than that measured for BaHPO3 (132.9 eV) [300]. The P 2p1/2 and 2p3/2 peaks were fit with GL line shapes with a 1:2 ratio and 0.84 eV separation [300,301].
172
ABKinetic Energy (eV) Kinetic Energy (eV) 400 600 800 1000 1200 1400 1348 13521356 1360 Survey P 2p Au 4f Au 3/2 5/2 5/2 3/2 3/2 Au 4d Au C 1s Ba 3d 4d Au O 1s Au 4p Au Ba 3d 3/2 Intensity Intensity S 2p Ba MNN O KLL Ba 3p 1/2 Au 5p Au
1000 800 600 400 200 0 140 138 136 134 132 130 128 126 Binding Energy (eV) Binding Energy (eV) Figure 7.5: X-ray photoelectron survey spectra of the electrostatic capture surface using MUPA inserted into an C8 SAM exposed to Ba(OH)2. A. Survey spectrum taken with a 10° tilt with respect to the surface plane, resulting in higher sensitivity to the SAM components. B. High-resolution spectra of the P 2p region with a peak at Eb = 133.8 eV. The P 2p1/2 and 2p3/2 peaks were fit with GL line shapes with a 1:2 ratio and 0.84 eV separation [300,301].
173
Figure 7.6221H shows high-resolution XPS spectra for the Ba 3d5/2 and O 1s regions.
The upper spectra correspond to an C8 SAM exposed to Ba(OH)2, the middle spectra correspond to a MUDA-inserted C8 SAM exposed to Ba(OH)2, and the lower spectra correspond to a MUPA-inserted C8 SAM exposed to Ba(OH)2. For the high-resolution
spectra of the Ba 3d region (Figure 7.6A222H ), the SAM exposed to aqueous Ba(OH)2 shows very little signal for nonspecifically adsorbed Ba, which is consistent with our ToF-SIMS images. The MUDA-inserted C8 SAM exposed to aqueous Ba(OH)2 has a peak at
Eb = 781.0 eV (red), and a MUPA-inserted C8 SAM exposed to aqueous Ba(OH)2
(lower) with a peak at Eb = 780.7 eV (red). Few literature values were found for Ba XPS spectra, and thus, we are using the Ba signal to determine the presence of on the substrates and in what relative concentrations. The literature value for the Ba 3d5/2 peak is at Eb = 780.6 eV [295], and both our measured values are close to this value. The higher energy peaks in the barium spectra (blue) are satellite peaks resulting from empty d states lowered below the Fermi level during the ionization process [302].
For the high-resolution spectra of the O 1s region (Figure 7.6B223H ), we observe a weak signal from the C8 SAM exposed to the Ba(OH)2, which we believe to be due to trapped solvent. The higher energy O 1s peaks in the MUDA- (535.1 eV) and MUPA-
(536.1 eV) inserted SAMs most likely arises from water that remained on the surface after the barium deposition [303]. The peak at Eb = 532.5 eV for the MUDA-inserted
SAM is attributed to the pendant carboxylate [304], while the peak at
Eb = 532.8 eV for the MUPA-inserted SAM is attributed to the phosphate pendant functionality [295]. The third peak for the MUPA-inserted SAM at Eb = 531.3 eV we
174
ABKinetic Energy (eV) Kinetic Energy (eV) 698 700 702 704 706 708 710 712 714 716 950 952 954 956 958
Ba 3d5/2 C8 + Ba O 1s C8 + Ba
Ba 3d C8 + MUDA Intensity Intensity + Ba
C8 + MUPA + Ba C8 + MUPA + Ba
790 788 786 784 782 780 778 776 774 772 770 538 536 534 532 530 528 Binding Energy (eV) Binding Energy (eV)
Figure 7.6: A. High-resolution spectra of the Ba 3d5/2 region. The spectra correspond to an C8 SAM exposed to aqueous Ba(OH)2 (upper), a MUDA-inserted C8 SAM exposed to aqueous Ba(OH)2 (middle) with a peak at Eb = 781.0 eV (red), and a MUPA-inserted C8 SAM exposed to aqueous Ba(OH)2 (lower) with a peak at Eb = 780.7 eV (red). The higher energy peaks in the barium spectra (blue) are satellite peaks [302]. B. High- resolution spectra of the O 1s region. The upper, middle and lower spectra correspond to the same samples as in A. The MUDA-inserted SAM was fit with two peaks with at Eb = 532.5 eV (red) and Eb = 535.1 eV (blue) and the MUPA-inserted SAM was fit to three peaks with at Eb = 532.8 eV (red), Eb = 536.1 eV (blue) and Eb = 531.3 eV (green). All fits were made with GL line shapes.
175 attribute to a barium-phosphate species since it is consistent with other barium compounds measured with XPS [295,305]. While this peak lies in the binding energy range for Ba(OH)2 [306], we infer that since we did not observe a Ba(OH)2 peak for the control substrate or the MUDA substrate, that we did not capture Ba(OH)2 on this substrate.
The final high-resolution area we analyzed for the electrostatic capture surfaces is the C 1s region (Figure 7.7). Here, we formed an C8 SAM from an ethanolic solution for
1 min as a reference and compared our Ba(OH)2-exposed C8 SAM, our Ba(OH)2- exposed MUDA-inserted SAM and our Ba(OH)2-exposed MUPA-inserted SAM to this standard. The GL line shape fits in red with peaks located at Eb = 284.9 eV for each of the spectra correspond to the C8 SAM matrix [22,307,308]. The small shoulder at
Eb = 287.4 eV appearing in each spectrum is attributed to solvent molecules trapped within the SAM matrix [309]. These are the only two C 1s peaks appearing in the spectra for the C8 SAM and the C8 SAM exposed to aqueous Ba(OH)2; thus, the other peak fits for the MUDA- and MUPA-inserted SAMs are attributed to the insertion and to the
Ba(OH)2 exposure. For the MUDA-inserted SAM, the peak centered at Eb = 285.6 eV is attributed to the carboxylate pendant functionality, similar to those measured for a polymer containing a carboxylic acid (poly(methacrylic acid)), and in phthalic acid
[310,311]. Similarly, for the MUPA-inserted SAM, we attribute the peak centered at
Eb = 285.7 eV to the phosphate pendant functionality. The final peaks observed in the high-resolution C 1s spectra for MUDA-inserted (Eb = 288.7 eV) and MUPA-inserted
2+ (Eb = 289.4 eV) SAM, are attributed to interactions of the Ba ions with the tether
176
Kinetic Energy (eV) 1192 119411961198 1200 1202 1204 1206
C8
C8 + Ba Intensity
C8 + MUDA + Ba
C8 + MUPA + Ba
294 292290288 286 284 282 280 Binding Energy (eV) Figure 7.7: High-resolution spectra of the C 1s region from top to bottom, C8 SAM, C8 SAM exposed to Ba(OH)2, MUDA-inserted C8 SAM exposed to Ba(OH)2, and MUPA- inserted C8 SAM exposed to Ba(OH)2. All fits were made with GL line shapes, and the peak positions are discussed within the text. 177
Atomic Concentration % Atomic Ratios C 1s C 1s C 1s C 1s C 1s S 2p O 1s Au 4f Ba 3d P 2p 5/2 S 2p Au 4f O 1s Ba 3d5/2 C8 76.2 3.5 0.4 19.9 0.0 N/A 21.8 3.8 190.5 N/A C8/Ba 72.6 3.7 0.7 21.6 1.4 N/A 19.6 3.4 103.7 51.9 C8/MUDA/Ba 63.6 2.0 9.3 12.2 12.9 N/A 31.8 5.2 6.8 4.9 C8/MUPA/Ba 48.3 1.6 14.2 12.5 22.3 1.1 30.2 3.9 3.4 2.2
Table 7.2: Atomic concentrations and ratios measured from XPS for an C8 SAM, an C8 SAM exposed to aqueous Ba(OH)2, a MUDA-inserted C8 SAM exposed to aqueous Ba(OH)2, and a MUPA-inserted C8 SAM exposed to aqueous Ba(OH)2.
178 molecules [305,312]. This could also be attributed to the carboxylates in MUDA [295]; however, since we observe this peak for the MUPA, we conclude this peak is from the
Ba interactions.
The atomic concentrations from the high-resolution XPS spectra are summarized in Table 7.2, which displays the atomic concentrations and atomic ratios calculated for the C 1s, S 2p, O 1s, Au 4f, Ba 3d5/2 and P 2p high-resolution spectra calculated for the control C8 SAM, the C8 SAM exposed to Ba(OH)2, the MUDA-inserted C8 SAM exposed to Ba(OH)2 and the MUPA-inserted C8 SAM exposed to Ba(OH)2. A small concentration of Ba was observed for the C8 SAM exposed to the Ba(OH)2 solution due to nonspecific adsorption; however, the concentration of the Ba displayed large increases for both the MUDA- and MUPA- inserted SAMs. Similarly, O was observed in each sample, but as with the Ba concentration for the MUDA- and MUPA- inserted SAMs, we observed a large increase in the O concentration when the tether groups were inserted.
7.4 Conclusions and Future Directions
Here, we have shown the first steps towards creating captures surfaces for superatom clusters. We have used MUDA and MUPA molecules inserted into C8 SAMs to act as tethers for the capture of Ba2+ ions. The electrostatic capture surfaces using an ethanolic BaTO solution were analyzed using ToF-SIMS and XPS, which indicate that nonspecific adsorption can occur on these surfaces. To circumvent this issue, we employed aqueous Ba(OH)2 solutions less likely to interact with the hydrophobic methyl termini of the alkanethiolate SAMs. 179 This work will be continued to tune the capture surfaces for specific ions, atoms, and clusters. This will include multiplexed surfaces, with several different tether molecules present to create modern-day analogues to the reactions performed to characterize transuranium elements. Finally, these capture surfaces will be used to tether superatom clusters to surfaces to be analyzed using STM.
180
Chapter 8
CONCLUSIONS AND FUTURE PROSPECTS
8.1 Summary
This dissertation focused on single-molecule electronic measurements and strategies for capturing molecules and clusters to surfaces to measure them. In Chapter 1, we introduced our SAM assemblies, STM instrumentation and how we have combined microwave frequencies in our STM tunneling junction. In Chapter 2, we gave an overview of the current progress towards molecular devices including typical molecules used, the testbeds in which they are measured and possible nanoscale junctions in which single molecules could be assembled. In Chapters 3, 4 and 5, we looked at the conductance switching exhibited by OPE molecules to understand the mechanism by which they switch, the motions that these molecules exhibit at substrate step edges and their real-time switching and motion measured on the millisecond time scale. In
Chapter 6, we discussed our use of ACSTM to measure the relative polarizability of
SAMs and the nitro-functionalized OPE switches; we used the polarizability signal to predict which molecules would be active as switches or by exhibiting motion. In
Chapter 7, we developed methods to create capture surfaces for alkaline-earth metals.
These capture surfaces are being designed for use as superatom cluster capture surfaces. 181
The background for this dissertation resulted in the following reviews written on molecular devices and STM:
Molecular Devices, A. M. Moore, D. L. Allara, and P. S. Weiss, in NNIN Nanotechnology Open Textbook (2007) pp. 11.1-11.29. Available at: http://www.nano.umn.edu/nnin_opentext/chapter_11-nnin_open_textbook.pdf.
Functional and Spectroscopic Measurements with Scanning Tunneling Microscopy, A. M. Moore and P. S. Weiss, in Annual Review of Analytical Chemistry 1, 852-882 (2008).
8.2 Single-Molecule Switches
The work in this dissertation represents a contribution towards the understanding of molecular interactions on the single-molecule level for OPE switch molecules. We have analyzed how our measurements fit in with the broader field of molecular devices, as detailed in Chapter 2. The work on OPE switch molecules has been performed in collaboration with the Tour group at Rice University, who synthesized the OPE molecules and derivatives. We have determined the mechanism by which conductance switching in OPE molecules occurs. The experiments detailed in Chapter 3 discuss how we have used molecular engineering, where components of OPE molecules were designed and synthesized to test each of the theoretically and experimentally hypothesized mechanisms for the observed conductance switching. The hypothesized mechanisms included reduction of OPE functional groups, ring rotation of the OPE phenyl backbone, bond fluctuations between the OPE and the substrate and a change in hybridization between the Au{111} substrate and the OPE molecules [109]. The only mechanism consistent with our and others’ data was that of a change in hybridization 182 occurring at the molecule-substrate interface via a molecular tilt of the OPE molecules, and possibly concomitant rearrangement of the substrate atoms [24,25,106-110,224,225].
We have characterized motions exhibited by the OPE molecules in addition to conductance switching, as detailed in Chapter 4. In some of the OPE inserted SAMs, our analyses for the occurrences vs. apparent height found three apparent heights for the OPE switches. To determine the nature of this ‘third’ apparent height, we designed a new extraction method for determining apparent heights. This was done by using an area of the SAM containing no defect sites adjacent to each inserted OPE molecule as our background height. Using this new background extraction method, we were able to determine that molecules located at substrate step edges were able to place-exchange up and down the substrate step edge in addition to exhibiting conductance switching.
Molecules isolated in the SAMs at domain boundaries with no substrate steps in proximity only exhibited two apparent heights. We were able to show that motion at the step edges was not necessary for conductance switching to occur, however switching and motion could occur simultaneously and independently.
Finally, the switching and motion events that had been analyzed previously were on an image by image basis. However, the question remains of how fast switching and motion events could occur, and thus, in Chapter 5, we used apparent height vs. time measurements to analyze these events on a millisecond time scale rather than the second to minute time scale of imaging. Using these measurements, we determined that switching and motion both occur in real time (10 kHz and faster) and we also discovered that further motions (~0.5 Å) take place, which we attributed to substrate atom rearrangements, consistent with recent X-ray diffraction measurements [268]. 183
The work described in Chapters 3, 4 and 5 has resulted in the following publications:
Cross-Step Place-Exchange of Oligo(phenylene-ethynylene) Molecules, A. M. Moore, B. A. Mantooth, A. A. Dameron, Z. J. Donhauser, F. Maya, D. W. Price Jr., Y. Yao, J. M. Tour and P. S. Weiss, Nano Letters 5, 2292-2297 (2005).
Molecular Engineering and Measurements to Test Hypothesized Mechanisms in Single- Molecule Conductance Switching, A. M. Moore, A. A. Dameron, B. A. Mantooth, R. K. Smith, D. J. Fuchs, J. W. Ciszek, F. Maya, Y. Yao, J. M. Tour and P. S. Weiss, Journal of the American Chemical Society 128, 1959-1967 (2006).
Real-Time Measurements of Conductance Switching and Motion of Single Oligo(phenylene-ethynylene) Molecules, A. M. Moore, B. A. Mantooth, Z. J. Donhauser, Y. Yao, J. M. Tour and P. S. Weiss, Journal of the American Chemical Society 129, 10352-10353 (2007).
Measurements and Mechanisms of Single-Molecule Conductance Switching, A. M. Moore, B. A. Mantooth, A. A. Dameron, Z. J. Donhauser, P. A. Lewis, R. K. Smith, D. J. Fuchs and P. S. Weiss, in press for Nano- and Micromaterials, eds. K. Ohno, M. Tanaka, J. Takeda and Y. Kawazoe, Advances in Materials Research 10, 29-47 (2008).
Further experiments that should be performed for the OPE switch system include analyses of the S-Au interaction. Since we have determined the importance of this bond for the observed conductance switching for the OPE molecules, it will be important to characterize the changes that occur at the molecule-substrate interface. This could be done by changing the metal used for the substrate to which our molecules are assembled.
For example, Au-molecule complexes have been shown to have mobility upon adsorption of the SAM molecules [16,18-20,255,256] and after the SAM has assembled [27,28,224].
If a substrate that were less susceptible to motion could be used to create these assemblies, would the same switching observed on Au{111} substrates occur? Could substrates such as semiconductor surfaces with SAMs assembled [313-318] be used, since the use of Si would be more compatible with current processor fabrication 184 techniques? The question of the influence of substrate atom rearrangements in the observed conductance switching could also be addressed by changing the substrate.
In addition to changes to the substrate, the S linkage could be replaced with other functionalities such as amines, isonitriles or selenolates. These varied contacts could address the question of are there any functionalities that will not exhibit conductance switching on Au{111} substrates when isolated within SAM matrices? Also, how much does the conductivity change through the use of different functional groups
[105,319,320]? Could different conductivities be used for device fabrication by having multiple switches with different contacts that determine different conductance states?
The S-Au interaction could be further analyzed with the barrier height measurements developed by S. U. Nanayakkara to image and to identify both the pendant and bonding groups of alkanethiolates [321]. If this method were applied to the OPEs, we could explore the changes that occur at this bond when the molecule switches conductance states, including rearrangements that occur, and we could determine the tilt angle of the molecule as it changes upon conductance switching.
The charge transport of the OPE molecules could also be characterized using
STM to record inelastic tunneling spectroscopy (IETS) [48,195,322-326]. Using IETS, the interactions between the electron transport and the molecular vibrations could be understood similar to the work performed by D. L. Allara, J. G. Kushmerick, M. A. Reed and T. S. Mayer and coworkers, who have used IETS to trace the electronic pathways in molecular transport junctions [195,322-326].
Finally, experiments should be performed to place the OPE molecules within the nanoscale junctions described in Chapter 2. It is important for device fabrication to 185 understand if the switching phenomenon observed with STM is a function of the testbed in which it is measured. It would be interesting to observe if isolated OPE molecules within a junction could exhibit conductance switching, and it would aid in the assembly of future devices by placing molecules in defined junctions and not having the necessity of being addressed by STM.
8.3 Expanding the Capabilities of Scanning Tunneling Microscopy
In Chapter 1 we detailed the instrumental setup for coupling microwave frequencies into the STM tunneling junction by applying two microwave frequencies offset by a difference frequency. The signals are heterodyned at the lower difference frequency set between the two microwave sources, enabling measurement using a LIA.
This technique was applied to measure the relative polarizabilities of the SAMs and the
OPE switches, as detailed in Chapter 6. The microwave frequencies distort the electron cloud around the molecules within the STM junction, where the degree of distortion possible for any molecule is that molecule’s polarizability. We measured different species within the STM junction and compared the relative values, recorded through the MDF magnitude signal, to calculated polarizability values. We used this system further to analyze the nitro-functionalized OPE switch molecules within the ACSTM tunneling junction. Here, we determined that a stable molecule had a constant profile and MDF magnitude. However, for molecules that exhibited switching or motion, we found that the
MDF signal displayed an irregular profile and large changes in the MDF magnitude, thus 186 giving us predictability of which molecules would switch or exhibit motion; this was not predicted by topography alone.
The work towards coupling microwave frequencies into the STM tunneling junction will result in the following manuscript:
Imaging Single-Molecule Polarizability and Buried Interface Dynamics, A. M. Moore, S. Yeganeh, Y. Yao, J. M. Tour, M. A. Ratner and P. S. Weiss, in preparation.
Currently, it would be difficult to distinguish between two molecules that were imaged with the same apparent height using STM. Scanning tunneling spectroscopy is currently used in STM to aid in molecular identification. However, spectroscopic tools such as IETS requires ultrahigh vacuum and cryogenic temperatures to reduce the thermal spread of electronic energies, and the IETS selection rules have not been determined. If the relative polarizabilities of molecules deposited on the surface were determined theoretically, their relative polarizabilities could be measured via ACSTM further enabling molecular identification. In addition to coupling microwave frequencies into the STM junction, other frequencies including visible, UV and IR could also be coupled leading to further spectroscopic characterization of molecules within the STM tunneling junction.
8.4 Cluster-Capture Surfaces
The final work detailed in this dissertation discussed our initial steps towards creating functional surfaces to capture superatom clusters. This work has been performed 187 as part of a collaboration with the Sen and Castleman groups at The Pennsylvania State
University, who have been synthesizing precise nanoscale clusters. Our goals for creating capture surfaces have been to identify the clusters as having specific element-like properties, to tether them to surfaces for analysis with STM, and to tether the clusters at specific locations on the substrate. We have used ToF-SIMS in collaboration with M. E.
2+ Kruczy and N. Winograd to show that Ba ions bind preferentially in patterned MUDA and MUPA areas, and we used XPS in collaboration with both the Penn State
Nanofabrication Facility and Drs. B. E. Koel and A. C. Miller at Lehigh University.
Chapter 7 discusses our initial tether surfaces, which were prepared to create an electrostatic capture for superatom clusters for analysis with STM. The electrostatic capture tethers have been developed, using Ba2+ ions as a model, in order to keep the clusters stationary so they are not swept away when the STM tip is rastered over them or when they are measured using scanning tunneling spectroscopy. Similar investigations of tethered clusters were performed for Au11 clusters by R. K. Smith et al. to determine the electron transfer properties of these clusters [121]. The electrostatic tether will allow the clusters to be held in place, isolating them from the substrate for measurement. The initial samples here used MUDA inserted SAMs to capture Ba2+ ions from ethanolic BaTO solutions.
We are further striving to create a modern-day analogue to the work done by
Seaborg and cowokers [287-291], who looked at defining the properties of transuranium elements by reacting them with various species and determining if they formed precipitates. Similarly, we are working towards creating capture surfaces to capture specific ‘elements’ that will be applied to the clusters, such as those that that act as 188 alkaline-earth metals to define their chemical and electronic properties. The initial system described above used MUDA. We tested a second capture scheme using MUPA inserted
2+ in host C8 SAMs to capture Ba ions from Ba(OH)2 solutions. The Ba(OH)2 was dissolved in aqueous solutions. Since the methyl termini of the host C8 SAM molecules are hydrophobic, we have decreased the nonspecific adsorption that was present in the initial electrostatic capture surfaces.
The work described in Chapter 7 on the capture surfaces will result in the following manuscript:
Alkaline-Earth Metal Capture Surfaces, A. M. Moore, T. J. Mullen, M. E. Kurczy, N. Winograd, A. C. Miller, B. E. Koel and P. S.Weiss, in preparation.
The work on cluster-capture surfaces will be continued in the creation of surfaces with a variety of tethers to test element properties across the periodic table. The use of
μCIP facilitates the ease of inserting molecules in specific locations. Seaborg and coworkers tested transuranium elements for their insolubility with many different anionic species. Similarly, we would like to be able to put a variety of isolated tether molecules in specifically patterned areas across the surface to determine which specific clusters bind to them. This multiplexing will facilitate the determination of the specific atomic properties of the clusters.
The clusters will also need to be tethered to the surfaces and studied by STM, similar to the work done by R. K. Smith et al. for Au11 clusters [121]. This will require 189 further collaboration with the Castleman group to develop strategies to bring the clusters they have synthesized to our tether surfaces.
8.5 Final Thoughts
The research themes presented in this dissertation demonstrate the assembly of molecules on Au{111} substrates for the characterization of their electronic properties.
We have described studies of the conductance switching and motion of OPE molecules, have measured the polarizability of SAMs, and have used the polarizability to characterize switch molecules further. Finally, we have taken the first steps to create capture surfaces to tether molecules with specific element-like properties to surfaces.
Understanding the ways in which single molecules interact with a surface and within their local environment will continue to gain importance as we continue to create ever smaller devices, and as devices reach their limit at the single-molecule level.
Visionaries Gordon Moore and Richard Feynman were leaders with ideas unheard of in their time, pointing towards this field of molecular devices that we are striving to realize.
190 REFERENCES
[1] A. M. Moore, D. L. Allara, and P. S. Weiss, Molecular devices, in NNIN Nanotechnology Open Textbook, 2007, pp. 11.1-11.29.
[2] B. A. Mantooth and P. S. Weiss, Fabrication, assembly, and characterization of molecular electronic components, Proceedings of the IEEE 91, 1785-1802 (2003).
[3] G. E. Moore, Cramming more components onto integrated circuits, Electronics 38, 114-117 (1965).
[4] P. Ebert, N. D. Jäger, K. Urban, and E. R. Weber, Nanoscale fluctuations in the distribution of dopant atoms: Dopant-induced dots and roughness of electronic interfaces, Journal of Vacuum Science and Technology B 22, 2018-2025 (2004).
[5] M. R. Castell, D. A. Muller, and P. M. Voyles, Dopant mapping for the nanotechnology age, Nature Materials 2, 129-131 (2003).
[6] F. G. Pikus and K. K. Likharev, Nanoscale field-effect transistors: An ultimate size analysis, Applied Physics Letters 71, 3661-3663 (1997).
[7] R. P. Feynman, There's plenty of room at the bottom, Engineering and Science 23, 22–36 (1960).
[8] A. Aviram and M. A. Ratner, Molecular rectifiers, Chemical Physics Letters 29, 277-283 (1974).
[9] R. H. Friend, R. W. Gymer, A. B. Holmes, J. H. Burroughes, R. N. Marks, C. Taliani, D. D. C. Bradley, D. A. Dos Santos, J. L. Brédas, M. Lögdlund, and W. R. Salaneck, Electroluminescence in conjugated polymers, Nature 397, 121-128 (1999).
[10] J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay, R. H. Friend, P. L. Burns, and A. B. Holmes, Light-emitting diodes based on conjugated polymers, Nature 347, 539-541 (1990).
[11] G. Gustafsson, Y. Cao, G. M. Treacy, F. Klavetter, N. Colaneri, and A. J. Heeger, Flexible light-emitting diodes made from soluble conducting polymers, Nature 357, 477-479 (1992).
[12] R. K. Smith, P. A. Lewis, and P. S. Weiss, Patterned self-assembled monolayers, Progress in Surface Science 75, 1-68 (2004).
[13] A. Ulman, Formation and structure of self-assembled monolayers, Chemical Reviews 96, 1533-1554 (1996). 191 [14] R. G. Nuzzo and D. L. Allara, Adsorption of bifunctional organic disulfides on gold surfaces, Journal of the American Chemical Society 105, 4481-4483 (1983).
[15] R. G. Nuzzo, F. A. Fusco, and D. L. Allara, Spontaneously organized molecular assemblies. 3. Preparation and properties of solution adsorbed monolayers of organic disulfides on gold surfaces, Journal of the American Chemical Society 109, 2358-2368 (1987).
[16] G. E. Poirier and E. D. Pylant, The self-assembly mechanism of alkanethiols on Au(111), Science 272, 1145-1148 (1996).
[17] E. Barrena, E. Palacios-Lidón, C. Munuera, X. Torrelles, S. Ferrer, J. Ulrich, M. Salmeron, and C. Ocal, The role of intermolecular and molecule-substrate interactions in the stability of alkanethiol nonsaturated phases on Au(111), Journal of the American Chemical Society 126, 385-395 (2004).
[18] G. E. Poirier, Mechanism of formation of Au vacancy islands in alkanethiol monolayers on Au(111), Langmuir 13, 2019-2026 (1997).
[19] G. E. Poirier, Characterization of organosulfur molecular monolayers on Au(111) using scanning tunneling microscopy, Chemical Reviews 97, 1117-1127 (1997).
[20] G. E. Poirier and M. J. Tarlov, The c(4×2) superlattice of n-alkanethiol monolayers self-assembled on Au(111), Langmuir 10, 2853-2856 (1994).
[21] G. E. Poirier, M. J. Tarlov, and H. E. Rushmeier, Two-dimensional liquid phase and the p x √3 phase of alkanethiol self-assembled monolayers on Au(111), Langmuir 10, 3383-3386 (1994).
[22] R. G. Nuzzo, L. H. Dubois, and D. L. Allara, Fundamental studies of microscopic wetting on organic surfaces. 1. Formation and structural characterization of a self-consistent series of polyfunctional organic monolayers, Journal of the American Chemical Society 112, 558-569 (1990).
[23] L. A. Bumm, J. J. Arnold, M. T. Cygan, T. D. Dunbar, T. P. Burgin, L. Jones II, D. L. Allara, J. M. Tour, and P. S. Weiss, Are single molecular wires conducting? Science 271, 1705-1707 (1996).
[24] Z. J. Donhauser, B. A. Mantooth, K. F. Kelly, L. A. Bumm, J. D. Monnell, J. J. Stapleton, D. W. Price, A. M. Rawlett, D. L. Allara, J. M. Tour, and P. S. Weiss, Conductance switching in single molecules through conformational changes, Science 292, 2303-2307 (2001).
192 [25] Z. J. Donhauser, B. A. Mantooth, T. P. Pearl, K. F. Kelly, S. U. Nanayakkara, and P. S. Weiss, Matrix mediated control of stochastic single molecule conductance switching, Japanese Journal of Applied Physics 41, 4871-4877 (2002).
[26] T. J. Mullen, C. Srinivasan, J. N. Hohman, S. D. Gillmor, M. J. Shuster, M. W. Horn, A. M. Andrews, and P. S. Weiss, Microcontact insertion printing, Applied Physics Letters 90, 063114-1-3 (2007).
[27] P. A. Lewis, Z. J. Donhauser, B. A. Mantooth, R. K. Smith, L. A. Bumm, K. F. Kelly, and P. S. Weiss, Control and placement of molecules via self-assembly, Nanotechnology 12, 231-237 (2001).
[28] S. J. Stranick, A. N. Parikh, Y. T. Tao, D. L. Allara, and P. S. Weiss, Phase separation of mixed composition self-assembled monolayers into nanometer-scale molecular domains, Journal of Physical Chemistry 98, 7636-7646 (1994).
[29] Z. J. Donhauser, P. S. Weiss, D. W. Price, and J. M. Tour, Control of alkanethiolate monolayer structure using vapor-phase annealing, Journal of the American Chemical Society 125, 11462-11463 (2003).
[30] T. J. Mullen, A. A. Dameron, and P. S. Weiss, Directed assembly and separation of self-assembled monolayers via electrochemical processing, Journal of Physical Chemistry B 110, 14410-14417 (2006).
[31] T. Kakiuchi, K. Sato, M. Iida, D. Hobara, S. Imabayashi, and K. Niki, Phase separation of alkanethiol self-assembled monolayers during the replacement of adsorbed thiolates on Au{111} with thiols in solution, Langmuir 16, 7238-7244 (2000).
[32] P. A. Lewis, R. K. Smith, K. F. Kelly, L. A. Bumm, S. M. Reed, R. S. Clegg, J. D. Gunderson, J. E. Hutchison, and P. S. Weiss, The role of buried hydrogen bonds in self-assembled mixed composition thiols on Au{111}, Journal of Physical Chemistry B 105, 10630-10636 (2001).
[33] R. K. Smith, S. M. Reed, P. A. Lewis, J. D. Monnell, R. S. Clegg, K. F. Kelly, L. A. Bumm, J. E. Hutchison, and P. S. Weiss, Phase separation within a binary self-assembled monolayer on Au{111} driven by an amide-containing alkanethiol, Journal of Physical Chemistry B 105, 1119-1122 (2001).
[34] A. A. Dameron, J. R. Hampton, R. K. Smith, T. J. Mullen, S. D. Gillmor, and P. S. Weiss, Microdisplacement printing, Nano Letters 5, 1834-1837 (2005).
[35] G. Binnig, H. Rohrer, C. Gerber, and E. Weibel, Surface studies by scanning tunneling microscopy, Physical Review Letters 49, 57-61 (1982).
193 [36] D. A. Bonnell, Scanning probe microscopy and spectroscopy: Theory, techniques, and applications, second edition (Wiley-VCH, Inc., New York, 2001).
[37] C. J. Chen, Introduction to scanning tunneling microscopy (Oxford University Press, New York, 1993).
[38] J. A. Stroscio and W. J. Kaiser, Scanning tunneling microscopy (Academic Press, San Diego, 1993).
[39] G. Binnig and H. Rohrer, Scanning tunneling microscopy, IBM Journal of Research and Development 30, 355-369 (1986).
[40] G. Binnig and H. Rohrer, Scanning tunneling microscopy - from birth to adolescence, Reviews of Modern Physics 59, 615-625 (1987).
[41] J. E. Griffith and G. P. Kochanski, Scanning tunneling microscopy, Annual Review of Materials Science 20, 219-244 (1990).
[42] J. H. Coombs, M. E. Welland, and J. B. Pethica, Experimental barrier heights and the image potential in scanning tunneling microscopy, Surface Science 198, L353-L358 (1988).
[43] J. Tersoff and D. R. Hamann, Theory and application for the scanning tunneling microscope, Physical Review Letters 50, 1998-2001 (1983).
[44] N. Garcia, Theory of scanning tunneling microscopy and spectroscopy: Resolution, image and field states, and thin oxide layers, IBM Journal of Research and Development 30, 533-542 (1986).
[45] D. M. Eigler, P. S. Weiss, and E. K. Schweizer, Imaging Xe with a low- temperature scanning tunneling microscope, Physical Review Letters 66, 1189- 1191 (1991).
[46] J. Tersoff and D. R. Hamann, Theory of the scanning tunneling microscope, Physical Review B 31, 805-813 (1985).
[47] A. M. Moore and P. S. Weiss, Functional and spectroscopic measurements with scanning tunneling microscopy, Annual Review of Analytical Chemistry 1, 857- 882 (2008).
[48] B. C. Stipe, M. A. Rezaei, and W. Ho, Single-molecule vibrational spectroscopy and microscopy, Science 280, 1732-1735 (1998).
[49] M. F. Crommie, C. P. Lutz, and D. M. Eigler, Spectroscopy of a single adsorbed atom, Physical Review B 48, 2851-2854 (1993). 194 [50] R. M. Feenstra, Electronic states of metal atoms on the GaAs(110) surface studied by scanning tunneling microscopy, Physical Review Letters 63, 1412-1415 (1989).
[51] L. A. Bumm and P. S. Weiss, Small cavity nonresonant tunable microwave frequency alternating current scanning tunneling microscope, Review of Scientific Instruments 66, 4140-4145 (1995).
[52] S. J. Stranick and P. S. Weiss, A versatile microwave frequency compatible scanning tunneling microscope, Review of Scientific Instruments 64, 1232-1234 (1993).
[53] S. J. Stranick and P. S. Weiss, Alternating current scanning tunneling microscopy and nonlinear spectroscopy, Journal of Physical Chemistry 98, 1762-1764 (1994).
[54] S. J. Stranick and P. S. Weiss, A tunable microwave frequency alternating-current scanning tunneling microscope, Review of Scientific Instruments 65, 918-921 (1994).
[55] D. E. Bergeron, P. J. Roach, A. W. J. Castleman, N. O. Jones, and S. N. Khanna, Al cluster superatoms as haolgens in polyhalides and as alkaline earths in iodide salts, Science 307, 231-235 (2005).
[56] D. Kuciauskas, P. A. Liddell, S. Lin, T. E. Johnson, S. J. Weghorn, J. S. Lindsey, A. L. Moore, T. A. Moore, and D. Gust, An artificial photosynthetic antenna- reaction center complex, Journal of the American Chemical Society 121, 8604- 8614 (1999).
[57] J. P. Sumida, P. A. Liddell, S. Lin, A. N. Macpherson, G. R. Seely, A. L. Moore, T. A. Moore, and D. Gust, Contrasting photoinduced electron transfer properties of two closely related, rigidly linked porphyrin-quinone dyads, Journal of Physical Chemistry A 102, 5512-5519 (1998).
[58] J. S. Hwang, K. J. Kong, D. Ahn, G. S. Lee, D. J. Ahn, and S. W. Hwang, Electrical transport through 60 base pairs of poly(dG)-poly(dC) DNA molecules, Applied Physics Letters 81, 1134-1136 (2002).
[59] D. Porath, A. Bezryadin, S. de Vries, and C. Dekker, Direct measurement of electrical transport through DNA molecules, Nature 403, 635-638 (2000).
[60] B. Xu, P. Zhang, X. Li, and N. J. Tao, Direct conductance measurement of single DNA molecules in aqueous solution, Nano Letters 4, 1105-1108 (2004).
[61] N. R. Armstrong, Phthalocyanines and porphyrins as materials, Journal of Porphyrins and Phthalocyanines 4, 414-417 (2000). 195 [62] X. Lu and K. W. Hipps, Scanning tunneling microscopy of metal phthalocyanines: d6 and d8 cases, Journal of Physical Chemistry B 101, 5391- 5396 (1997).
[63] N. Bampos, C. N. Woodburn, M. E. Welland, and J. K. M. Sanders, STM images of individual porphyrin molecules on Cu(100) and Cu(111) surfaces, Angewandte Chemie, International Edition in English 38, 2780-2783 (1999).
[64] T. A. Jung, R. R. Schlittler, and J. K. Gimzewski, Conformational identification of individual adsorbed molecules with the STM, Nature 386, 696-698 (1997).
[65] A. M. Kuznetov and J. Ulstrup, Mechanisms of in situ scanning tunneling microscopy of organized redox molecular assemblies, Journal of Physical Chemistry A 104, 11531-11540 (2000).
[66] R. A. Bissell, E. Cordova, A. E. Kaifer, and J. F. Stoddart, A chemically and electrochemically switchable molecular shuttle, Nature 369, 133-137 (1994).
[67] D. I. Gittins, D. Bethell, D. J. Schiffrin, and R. N. Nichols, A nanometre-scale electronic switch consisting of a metal cluster and redox-addressable groups, Nature 408, 67-69 (2000).
[68] J. M. Tour, L. Jones II, D. L. Pearson, J. J. S. Lamba, T. P. Burgin, G. M. Whitesides, D. L. Allara, A. N. Parikh, and S. V. Atre, Self-assembled monolayers and multilayers of conjugated thiols, α,ω-dithiols, and thioacetyl-containing adsorbates: Understanding attachments between potential molecular wires and gold surfaces, Journal of the American Chemical Society 117, 9529-9534 (1995).
[69] M. Akai-Kasaya, K. Shimizu, Y. Watanabe, A. Saito, M. Aono, and Y. Kuwahara, Electronic structure of a polydiacetylene nanowire fabricated on highly ordered pyrolytic graphite, Physical Review Letters 91, 255501-1-4 (2003).
[70] B. L. Feringa, In control of motion: From molecular switches to molecular motors, Accounts of Chemical Research 34, 504-513 (2001).
[71] P. R. Hania, R. Telesca, L. N. Lucas, A. Pugzlys, J. van Esch, B. L. Feringa, J. G. Snijders, and K. Duppen, An optical and theoretical investigation of the ultrafast dynamics of a bisthienylethene-based photochromic switch, Journal of Physical Chemistry A 106, 8498-8507 (2002).
[72] A. S. Lukas and M. R. Wasielewski, Approaches to a molecular switch using photoinduced electron and energy transfer, in Molecular Switches, B. L. Feringa, Ed. (Weinheim: Wiley-VCH, 2001), pp. 1-35.
196 [73] M. Elbing, R. Ochs, M. Koentopp, M. Fischer, C. von Hanisch, F. Weigend, F. Evers, H. B. Weber, and M. Mayor, A single-molecule diode, Proceedings of the National Academy of Sciences of the United States of America 102, 8815-8820 (2005).
[74] N. J. Geddes, J. R. Sambles, D. J. Jarvis, W. G. Parker, and D. J. Sandman, Fabrication and investigation of asymmetric current-voltage characteristics of a metal Langmuir-Blodgett monolayer metal structure, Applied Physics Letters 56, 1916-1918 (1990).
[75] M. A. Reed, C. Zhou, M. R. Deshpande, C. J. Muller, T. P. Burgin, L. Jones II, and J. M. Tour, The electrical measurement of molecular junctions, in Molecular Electronics: Science and Technology, vol. 852, Annals of the New York Academy of Sciences, (New York: New York Academy of Sciences, 1998), pp. 133-144.
[76] C. Joachim and J. K. Gimzewski, An electromechanical amplifier using a single molecule, Chemical Physics Letters 265, 353-357 (1997).
[77] J. P. Wittke, Molecular amplification and generation of microwaves, Proceedings of the IEEE 85, 1144-1168 (1997).
[78] S. Das Sarma, J. Fabian, X. D. Hu, and I. Zutic, Spintronics: Electron spin coherence, entanglement, and transport, Superlattices and Microstructures 27, 289-295 (2000).
[79] H. A. Engel, P. Recher, and D. Loss, Electron spins in quantum dots for spintronics and quantum computation, Solid State Communications 119, 229-236 (2001).
[80] T. Gruber, M. Keim, R. Fiederling, G. Reuscher, W. Ossau, G. Schmidt, L. W. Molenkamp, and A. Waag, Electron spin manipulation using semimagnetic resonant tunneling diodes, Applied Physics Letters 78, 1101-1103 (2001).
[81] Z. Z. Gu, O. Sato, T. Iyoda, K. Hashimoto, and A. Fujishima, Spin switching effect in nickel nitroprusside: Design of a molecular spin device based on spin exchange interaction, Chemistry of Materials 9, 1092-1097 (1997).
[82] Y. H. Lee, Y. T. Jang, C. H. Choi, D. H. Kim, C. W. Lee, J. E. Lee, Y. S. Han, S. S. Yoon, J. K. Shin, S. T. Kim, E. K. Kim, and B. K. Ju, Direct nanowiring of carbon nanotubes for highly integrated electronic and spintronic devices, Advanced Materials 13, 1371-1373 (2001).
[83] C. Y. You and S. D. Bader, Voltage controlled spintronic devices for logic applications, Journal of Applied Physics 87, 5215-5217 (2000). 197 [84] C. P. Collier, E. W. Wong, M. Belohradsky, F. M. Raymo, J. F. Stoddart, P. J. Kuekes, R. S. Williams, and J. R. Heath, Electronically configurable molecular- based logic gates, Science 285, 391-394 (1999).
[85] A. P. de Silva, N. D. McClenaghan, and C. P. McCoy, Molecular logic systems, in Molecular Switches, B. L. Feringa, Ed. (Weinheim: Wiley-VCH, 2001), pp. 339- 361.
[86] O. Hod, R. Baer, and E. Rabani, A parallel electromagnetic molecular logic gate, Journal of the American Chemical Society 127, 1648-1649 (2005).
[87] A. Bachtold, P. Hadley, T. Nakanishi, and C. Dekker, Logic circuits with carbon nanotube transistors, Science 294, 1317-1320 (2001).
[88] P. S. Weiss, L. A. Bumm, T. D. Dunbar, T. P. Burgin, J. M. Tour, and D. L. Allara, Probing electronic properties of conjugated and saturated molecules in self-assembled monolayers, in Molecular Electronics: Science and Technology, vol. 852, Annals of the New York Academy of Sciences, (New York: New York Academy of Sciences, 1998), pp. 145-168.
[89] Y. Selzer and D. L. Allara, Single-molecule electrical junctions, Annual Review of Physical Chemistry 57, 20.1-20.31 (2006).
[90] E. G. Petrov, Y. V. Shevchenko, and V. I. Teslenko, Nonadiabatic donor- acceptor electron transfer mediated by a molecular bridge: A unified theoretical description of the superexchange and hopping mechanism, Journal of Chemical Physics 115, 7108-7122 (2001).
[91] L. A. Bumm, J. J. Arnold, T. D. Dunbar, D. L. Allara, and P. S. Weiss, Electron transfer through organic molecules, Journal of Physical Chemistry B 103, 8122- 8127 (1999).
[92] M. Magoga and C. Joachim, Conductance and transparence of long molecular wires, Physical Review B 56, 4722-4729 (1997).
[93] D. J. Wold, R. Haag, M. A. Rampi, and C. D. Frisbie, Distance dependence of electron tunneling through self-assembled monolayers measured by conducting probe atomic force microscopy: Unsaturated versus saturated molecular junctions, Journal of Physical Chemistry B 106, 2813-2816 (2002).
[94] Y. Selzer, A. Salomon, and D. Cahen, Effect of molecule-metal electronic coupling on through-bond hole tunneling across metal-organic monolayer- semiconductor junctions, Journal of the American Chemical Society 124, 2886- 2887 (2002).
198 [95] J. M. Seminario, C. E. de la Cruz, and P. A. Derosa, A theoretical analysis of metal-molecule contacts, Journal of the American Chemical Society 123, 5616- 5617 (2001).
[96] J. M. Seminario, A. G. Zacarias, and J. M. Tour, Theoretical study of a molecular resonant tunneling diode, Journal of the American Chemical Society 122, 3015- 3020 (2000).
[97] Y. Q. Xue, S. Datta, and M. A. Ratner, Charge transfer and 'band lineup' in molecular electronic devices: A chemical and numerical interpretation, Journal of Chemical Physics 115, 4292-4299 (2001).
[98] J. M. Seminario, A. G. Zacarias, and P. A. Derosa, Theoretical analysis of complementary molecular memory devices, Journal of Physical Chemistry A 105, 791-795 (2001).
[99] Y. Selzer, A. Salomon, and D. Cahen, The importance of chemical bonding to the contact for tunneling through alkyl chains, Journal of Physical Chemistry B 106, 10432-10439 (2002).
[100] Y. Selzer, A. Salomon, J. Ghabboun, and D. Cahen, Voltage-driven changes in molecular dipoles yield negative differential resistance at room temperature, Angewandte Chemie- International Edition 41, 827-830 (2002).
[101] S. Datta, W. Tian, S. Hong, R. Reifenberger, J. I. Henderson, and C. P. Kubiak, Current-voltage characteristics of self-assembled monolayers by scanning tunneling microscopy, Physical Review Letters 79, 2530-2533 (1997).
[102] N. D. Lang and P. Avouris, Electrical conductance of parallel atomic wires, Physical Review B 62, 7325-7329 (2000).
[103] J. Chen, W. Wang, J. Klemic, M. A. Reed, B. W. Axelrod, D. M. Kaschak, A. M. Rawlett, D. W. Price, S. M. Dirk, J. M. Tour, D. S. Grubisha, and D. W. Bennett, Molecular wires, switches, and memories, in Molecular Electronics II, vol. 960, Annals of the New York Academy of Sciences, (New York: New York Academy of Sciences, 2002), pp. 69-99.
[104] J. M. Beebe, V. B. Engelkes, L. L. Miller, and C. D. Frisbie, Contact resistance in metal-molecule-metal junctions based on aliphatic SAMs: Effects of surface linker and metal work function, Journal of the American Chemical Society 124, 11268- 11269 (2002).
199 [105] J. D. Monnell, J. J. Stapleton, J. J. Jackiw, T. D. Dunbar, W. A. Reinerth, S. M. Dirk, J. M. Tour, D. L. Allara, and P. S. Weiss, Ordered local domain structures of decaneselenolate and dodecaneselenolate monolayers on Au{111}, Journal of Physical Chemistry B 108, 9834-9841 (2004).
[106] P. A. Lewis, C. E. Inman, Y. Yao, J. M. Tour, J. E. Hutchison, and P. S. Weiss, Mediating stochastic switching of single molecules using chemical functionality, Journal of the American Chemical Society 126, 12214-12215 (2004).
[107] P. A. Lewis, C. E. Inman, F. Maya, J. M. Tour, J. E. Hutchison, and P. S. Weiss, Molecular engineering of the polarity and interactions of molecular electronic switches, Journal of the American Chemical Society 127, 17421-17428 (2005).
[108] A. A. Dameron, J. W. Ciszek, J. M. Tour, and P. S. Weiss, Effect of hindered internal rotation on packing and conductance of self-assembled monolayers, Journal of Physical Chemistry B 108, 16761-16767 (2004).
[109] A. M. Moore, A. A. Dameron, B. A. Mantooth, R. K. Smith, D. J. Fuchs, Y. Yao, J. W. Ciszek, F. Maya, J. M. Tour, and P. S. Weiss, Molecular engineering and measurements to test hypothesized mechanisms in single molecule conductance switching, Journal of the American Chemical Society 128, 1959-1967 (2006).
[110] G. K. Ramachandran, T. J. Hopson, A. M. Rawlett, L. A. Nagahara, A. Primak, and S. M. Lindsay, A bond-fluctuation mechanism for stochastic switching in wired molecules, Science 300, 1413-1416 (2003).
[111] R. A. Wassel, R. R. Fuierer, N. Kim, and C. B. Gorman, Stochastic variation in conductance on the nanometer scale: A general phenomenon, Nano Letters 3, 1617-1620 (2003).
[112] A. S. Blum, J. C. Yang, R. Shashidhar, and B. R. Ratna, Comparing the conductivity of molecular wires with the scanning tunneling microscope, Applied Physics Letters 82, 3322-3324 (2003).
[113] D. S. Seferos, A. S. Blum, J. G. Kushmerick, and G. C. Bazan, Single-molecule charge-transport measurements that reveal technique-dependent perturbations, Journal of the American Chemical Society 128, 11260-11267 (2006).
[114] B. A. Mantooth, Z. J. Donhauser, K. F. Kelly, and P. S. Weiss, Cross-correlation image tracking for drift correction and adsorbate analysis, Review of Scientific Instruments 73, 313-317 (2002).
[115] M. Di Ventra, S. G. Kim, S. T. Pantelides, and N. D. Lang, Temperature effects on the transport properties of molecules, Physical Review Letters 86, 288-291 (2001). 200
[116] J. Cornil, Y. Karzazi, and J. L. Brédas, Negative differential resistance in phenylene ethynylene oligomers, Journal of the American Chemical Society 124, 3516-3517 (2002).
[117] J. M. Seminario, A. G. Zacarias, and J. M. Tour, Theoretical interpretation of conductivity measurements of a thiotolane sandwich: A molecular scale electronic controller, Journal of the American Chemical Society 120, 3970-3974 (1998).
[118] H. Sellers, A. Ulman, Y. Shnidman, and J. E. Eilers, Structure and binding of alkanethiolates on gold and silver surfaces: Implications for self-assembled monolayers, Journal of the American Chemical Society 115, 9389-9401 (1993).
[119] Y.-T. Tao, C.-C. Wu, J.-Y. Eu, and W.-L. Lin, Structure evolution of aromatic- derivatized thiol monolayers on evaporated gold, Langmuir 13, 4018-4023 (1997).
[120] A. S. Kumar, T. Ye, T. Takami, B.-C. Yu, A. K. Flatt, J. M. Tour, and P. S. Weiss, Reversible photo-switching of single azobenzene molecules in controlled nanoscale environments, Nano Letters 8, 1644-1648 (2008).
[121] R. K. Smith, S. U. Nanayakkara, G. H. Woehrle, T. P. Pearl, M. M. Blake, J. E. Hutchison, and P. S. Weiss, Spectral diffusion in the tunneling spectra of ligand- stabilized undecagold clusters, Journal of the American Chemical Society 128, 9266-9267 (2006).
[122] R. P. Andres, T. Bein, M. Dorogi, S. Feng, J. I. Henderson, C. P. Kubiak, W. Mahoney, R. G. Osifchin, and R. Reifenberger, 'Coulomb staircase' at room temperature in a self-assembled molecular nanostructure, Science 272, 1323- 1325 (1996).
[123] T. P. Bigioni, L. E. Harrell, W. G. Cullen, D. E. Guthrie, R. L. Whetten, and P. N. First, Imaging and tunneling spectroscopy of gold nanocrystals and nanocrystal arrays, European Physical Journal D 6, 355-364 (1999).
[124] M. Dorogi, J. Gomez, R. Osifchin, R. P. Andres, and R. Reifenberger, Room- temperature Coulomb blockade from a self-assembled molecular nanostructure, Physics Review B 52, 9071-9077 (1995).
[125] C. Burda, X. Chen, R. Narayanan, and M. A. El-Sayed, Chemistry and properties of nanocrystals of different shapes, Chemical Reviews 105, 1025-1102 (2005).
201 [126] M.-C. Daniel and D. Astruc, Gold nanoparticles: Assembly, supramolecular chemistry, quantum-size-related properties, and applications toward biology, catalysis, and nanotechnology, Chemical Reviews 104, 293-346 (2004).
[127] C. B. Murray, C. R. Kagan, and M. G. Bawendi, Synthesis and characterization of monodisperse nanocrystals and close-packed nanocrystal assemblies, Annual Review of Materials Science 30, 545-610 (2000).
[128] R. P. Andres, S. Datta, M. Dorogi, J. Gomez, J. I. Henderson, D. B. Janes, V. R. Kolagunta, C. P. Kubiak, W. Mahoney, R. F. Osifchin, R. Reifenberger, M. P. Samanta, and W. Tian, Room temperature Coulomb blockade and Coulomb staircase from self-assembled nanostructures, Journal of Vacuum Science and Technology A 14, 1178-1183 (1996).
[129] D. J. Wold and C. D. Frisbie, Formation of metal-molecule-metal tunnel junctions: Microcontacts to alkanethiol monolayers with a conducting AFM tip, Journal of the American Chemical Society 122, 2970-2971 (2000).
[130] D. J. Wold and C. D. Frisbie, Fabrication and characterization of metal- molecule-metal junctions by conducting probe atomic force microscopy, Journal of the American Chemical Society 123, 5549-5556 (2001).
[131] H. Cohen, C. Nogues, R. Naaman, and D. Porath, Direct measurment of electrical transport through single DNA molecules of complex sequence, Proceedings of the National Academy of Sciences of the United States of America 102, 11589-11593 (2005).
[132] H. Cohen, C. Nogues, D. Ullien, S. Daube, R. Naaman, and D. Porath, Electrical characterization of self-assembled single- and double- stranded DNA monolayers using conductive AFM, Faraday Discussions 131, 367-376 (2006).
[133] X. D. Cui, A. Primak, X. Zarate, J. Tomfohr, O. F. Sankey, A. L. Moore, T. A. Moore, D. Gust, G. Harris, and S. M. Lindsay, Reproducible measurement of single molecule conductivity, Science 294, 571-574 (2001).
[134] X. D. Cui, A. Primak, X. Zarate, J. Tomfohr, O. F. Sankey, A. L. Moore, T. A. Moore, D. Gust, L. A. Nagahara, and S. M. Lindsay, Changes in the electronic properties of a molecule when it is wired into a circuit, Journal of Physical Chemistry B 106, 8609-8614 (2002).
[135] S. M. Lindsay, Single molecule electronics and tunneling in molecules, Japanese Journal of Applied Physics 41, 4867-4870 (2002).
202 [136] X. D. Cui, X. Zarate, J. Tomfohr, O. F. Sankey, A. Primak, A. L. Moore, T. A. Moore, D. Gust, G. Harris, and S. M. Lindsay, Making electrical contacts to molecular monolayers, Nanotechnology 13, 5-14 (2002).
[137] A. M. Rawlett, T. J. Hopson, L. A. Nagahara, R. K. Tsui, G. K. Ramachandran, and S. M. Lindsay, Electrical measurements of a dithiolated electronic molecule via conducting atomic force microscopy, Applied Physics Letters 81, 3043-3035 (2002).
[138] C. J. Muller, J. M. van Ruitenbeek, and L. J. de Jongh, Experimental observation of the transition from weak link to tunnel junction, Physica C 191, 485-504 (1992).
[139] C. J. Muller, J. M. Krans, T. N. Todorov, and M. A. Reed, Quantization effects in the conductance of metallic contacts at room temperature, Physical Review B 53, 1022-1025 (1996).
[140] C. J. Muller, B. J. Vleeming, M. A. Reed, J. J. S. Lamba, R. Hara, L. Jones II, and J. M. Tour, Atomic probes: A search for conduction through a single molecule, Nanotechnology 7, 409-411 (1996).
[141] J. Reichert, R. Ochs, D. Beckmann, H. B. Weber, M. Mayor, and H. von Löhneysen, Driving current through single organic molecules, Physical Review Letters 88, 176804-1-4 (2002).
[142] J. Reichert, H. B. Weber, M. Mayor, and H. von Lohneysen, Low-temperature conductance measurements on single molecules, Applied Physics Letters 82, 4137-4139 (2003).
[143] M. A. Reed, C. Zhou, C. J. Muller, T. P. Burgin, and J. M. Tour, Conductance of a molecular junction, Science 278, 252-254 (1997).
[144] E. G. Emberly and G. Kirczenow, Models of electron transport through organic molecular monolayers self-assembled on nanoscale metallic contacts, Physical Review B 64, 235412-1-8 (2001).
[145] D. Q. Andrews, R. P. Van Duyne, and M. A. Ratner, Stochastic modulation in molecular electronic transport junctions: Molecular dynamics coupled with charge transport calculations, Nano Letters 8, 1120-1126 (2008).
[146] J. Ulrich, D. Esrail, W. Pontius, L. Venkataraman, D. Millar, and L. H. Doerrer, Variability of conductance in molecular junctions, Journal of Physical Chemistry B 110, 2462-2466 (2006).
203 [147] L. Venkataraman, J. E. Klare, C. Nuckolls, M. S. Hybertsen, and M. L. Steigerwald, Dependence of single-molecule junction conductance on molecular conformation, Nature 442, 904-907 (2006).
[148] L. Venkataraman, J. E. Klare, I. W. Tam, C. Nuckolls, M. S. Hybertsen, and M. L. Steigerwald, Single-molecule circuits with well-defined molecular conductance, Nano Letters 6, 458-462 (2006).
[149] L. Venkataraman, Y. S. Park, A. C. Whalley, C. Nuckolls, M. S. Hybertsen, and M. L. Steigerwald, Electronics and chemistry: Varying single-molecule junction conductance using chemical substituents, Nano Letters 7, 502-506 (2007).
[150] B. Xu and N. J. Tao, Measurement of single-molecule resistance by repeated formation of molecular junctions, Science 301, 1221-1223 (2003).
[151] H. Park, A. K. L. Lim, A. P. Alivisatos, J. Park, and P. L. McEuen, Fabrication of metallic electrodes with nanometer separation by electromigration, Applied Physics Letters 75, 301-303 (1999).
[152] H. Park, J. Park, A. K. L. Lim, E. H. Anderson, A. P. Alivisatos, and P. L. McEuen, Nanomechanical oscillations in a single C60 transistor, Nature 407, 57- 60 (2000).
[153] J. Park, A. N. Pasupathy, J. I. Goldsmith, C. Chang, Y. Yaish, J. R. Petta, M. Rinkoski, J. P. Sethna, H. D. Abruña, P. L. McEuen, and D. C. Ralph, Coulomb blockade and the Kondo effect in single-atom transistors, Nature 417, 722-725 (2002).
[154] W. Liang, M. P. Shores, M. Bockrath, J. R. Long, and H. Park, Kondo resonance in a single-molecule transistor, Nature 417, 725-729 (2002).
[155] Y. Selzer, M. A. Cabassi, T. S. Mayer, and D. L. Allara, Thermally activated conduction in molecular junctions, Journal of the American Chemical Society 126, 4052-4053 (2004).
[156] Y. Selzer, M. A. Cabassi, T. S. Mayer, and D. L. Allara, Temperature effects on conduction through a molecular junction, Nanotechnology 15, S483-S488 (2004).
[157] O. M. Magnussen, B. M. Ocko, M. Deutsch, M. J. Regan, P. S. Pershan, D. Abernathy, G. Grubel, and J. F. Legrand, Self-assembly of organic films on a liquid metal, Nature 384, 250-252 (1996).
204 [158] R. E. Holmlin, R. Haag, M. L. Chabinyc, R. F. Ismagilov, A. E. Cohen, A. Terfort, M. A. Rampi, and G. M. Whitesides, Electron transport through thin organic films in metal- insulator-metal junctions based on self-assembled monolayers, Journal of the American Chemical Society 123, 5075-5085 (2001).
[159] M. A. Rampi, O. J. A. Schueller, and G. M. Whitesides, Alkanethiol self- assembled monolayers as the dielectric of capacitors with nanoscale thickness, Applied Physics Letters 72, 1781-1783 (1998).
[160] M. A. Rampi and G. M. Whitesides, A versatile experimental approach for understanding electron transport through organic materials, Chemical Physics 281, 373-391 (2002).
[161] K. Slowinski, H. K. Y. Fong, and M. Majda, Mercury-mercury tunneling junctions. 1. Electron tunneling across symmetric and asymmetric alkanethiolate bilayers, Journal of the American Chemical Society 121, 7257-7261 (1999).
[162] K. Slowinski and M. Majda, Mercury-mercury tunneling junctions - Part II. Structure and stability of symmetric alkanethiolate bilayers and their effect on the rate of electron tunneling, Journal of Electroanalytical Chemistry 491, 139-147 (2000).
[163] R. E. Holmlin, R. F. Ismagilov, R. Haag, V. Mujica, M. A. Ratner, M. A. Rampi, and G. M. Whitesides, Correlating electron transport and molecular structure in organic thin films, Angewandte Chemie- International Edition 40, 2316-2320 (2001).
[164] K. S. Ralis, R. A. Buhrman, and R. C. Tiberio, Fabrication of thin-film metal nanobridges, Applied Physics Letters 55, 2459-2461 (1989).
[165] D. C. Ralph, C. T. Black, and M. Tinkham, Spectroscopic measurement of discrete electronic states in single metal particles, Physical Review Letters 74, 3241-3244 (1995).
[166] D. C. Ralph, C. T. Black, and M. Tinkham, Studies of electron energy levels in single metal particles, Physica B 218, 258-261 (1996).
[167] C. Zhou, M. R. Deshpande, M. A. Reed, L. Jones II, and J. M. Tour, Nanoscale metal self-assembled monolayer metal heterostructures, Applied Physics Letters 71, 611-613 (1997).
[168] J. Chen, L. C. Calvet, M. A. Reed, D. W. Carr, D. S. Grubisha, and D. W. Bennett, Electronic transport through metal-1,4-phenylenediisocyanide-metal junctions, Chemical Physics Letters 313, 741-748 (1999).
205 [169] J. Chen, M. A. Reed, A. M. Rawlett, and J. M. Tour, Large on-off ratios and negative differential resistance in a molecular electronic device, Science 286, 1550-1552 (1999).
[170] J. Chen, W. Wang, M. A. Reed, A. M. Rawlett, D. W. Price, and J. M. Tour, Room-temperature negative differential resistance in nanoscale molecular junctions, Applied Physics Letters 77, 1224-1226 (2000).
[171] M. A. Reed, J. Chen, A. M. Rawlett, D. W. Price, and J. M. Tour, Molecular random access memory cell, Applied Physics Letters 78, 3735-3737 (2001).
[172] K. Konstadinidis, P. Zhang, R. L. Opila, and D. L. Allara, An in-situ X-ray photoelectron study of the interaction between vapor-deposited Ti atoms and functional-groups at the surfaces of self-assembled monolayers, Surface Science 338, 300-312 (1995).
[173] R. L. Opila and J. Eng, Thin films and interfaces in microelectronics: composition and chemistry as function of depth, Progress in Surface Science 69, 125-163 (2002).
[174] T. B. Tighe, T. A. Daniel, Z. H. Zhu, S. Uppili, N. Winograd, and D. L. Allara, Evolution of the interface and metal film morphology in the vapor deposition of Ti on hexadecanethiolate hydrocabon monolayers on Au, Journal of Physical Chemistry B 109, 21006-21014 (2005).
[175] A. V. Walker, T. B. Tighe, B. C. Haynie, S. Uppili, N. Winograd, and D. L. Allara, Chemical pathways in the interactions of reactive metal atoms with organic surfaces: Vapor deposition of Ca and Ti on a methoxy-terminated alkanetholate monolayer on Au, Journal of Physical Chemistry B 109, 11263- 11272 (2005).
[176] A. V. Walker, T. B. Tighe, J. J. Stapleton, B. C. Haynie, S. Uppili, D. L. Allara, and N. Winograd, The interaction of vapor-deposited Ti and Au with molecular wires, Applied Physics Letters 84, 4008-4010 (2004).
[177] Z. Zhu, T. A. Daniel, M. Maitani, O. M. Cabarcos, D. L. Allara, and N. Winograd, Controlling gold atom penetration through alkanethiolate self- assembled monolayers on Au{111} by adjusting terminal group intermolecular interactions, Journal of the American Chemical Society 128, 13710-13719 (2006).
[178] A. Bezryadin, C. Dekker, and G. Schmid, Electrostatic trapping of single conducting nanoparticles between nanoelectrodes, Applied Physics Letters 71, 1273-1275 (1997).
206 [179] I. Amlani, A. M. Rawlett, L. A. Nagahara, and R. K. Tsui, An approach to transport measurements of electronic molecules, Applied Physics Letters 80, 2761-2763 (2002).
[180] D. P. Long, C. H. Patterson, M. H. Moore, D. S. Seferos, G. C. Bazan, and J. G. Kushmerick, Magnetic directed assembly of molecular junctions, Applied Physics Letters 86, 153105-1-3 (2005).
[181] J. G. Kushmerick, D. B. Holt, S. K. Pollack, M. A. Ratner, J. C. Yang, T. L. Schull, J. Naciri, M. H. Moore, and R. Shashidhar, Effect of bond-length alternation in molecular wires, Journal of the American Chemical Society 124, 10654-10655 (2002).
[182] L. T. Cai, H. Skulason, J. G. Kushmerick, S. K. Pollack, J. Naciri, R. Shashidhar, D. L. Allara, T. E. Mallouk, and T. S. Mayer, Nanowire-based molecular monolayer junctions: Synthesis, assembly, and electrical characterization, Journal of Physical Chemistry B 108, 2827-2832 (2004).
[183] A. S. Blum, J. G. Kushmerick, D. P. Long, C. H. Patterson, J. C. Yang, J. C. Henderson, Y. X. Yao, J. M. Tour, R. Shashidhar, and B. R. Ratna, Molecularly inherent voltage-controlled conductance switching, Nature Materials 4, 167 (2005).
[184] J. G. Kushmerick, D. B. Holt, J. C. Yang, J. Naciri, M. H. Moore, and R. Shashidhar, Metal-molecule contacts and charge transport across monomolecular layers: Measurement and theory, Physical Review Letters 89, 086802-1-4 (2002).
[185] J. G. Kushmerick, J. Naciri, J. C. Yang, and R. Shashidhar, Conductance scaling of molecular wires in parallel, Nano Letters 3, 897-900 (2003).
[186] J. G. Kushmerick, S. K. Pollack, J. C. Yang, J. Naciri, D. B. Holt, M. A. Ratner, and R. Shashidhar, Understanding charge transport in molecular electronics, in Molecular Elecronics III, vol. 1006, Annals of the New York Academy of Sciences, (New York: New York Academy of Sciences, 2003), pp. 277-290.
[187] S. Gregory, Inelastic tunneling spectroscopy and single-electron tunneling in an adjustable microscopic tunnel junction, Physical Review Letters 64, 689-692 (1990).
[188] J. G. Kushmerick, C. M. Whitaker, S. K. Pollack, T. L. Schull, and R. Shashidhar, Tuning current rectification across molecular junctions, Nanotechnology 15, S489-S493 (2004).
[189] J. M. Beebe and J. G. Kushmerick, Nanoscale switch elements from self- assembled monolayers on silver, Applied Physics Letters 90, 083117-1-3 (2007). 207 [190] C. Liang, K. Terabe, T. Hasegawa, and M. Aono, Resistance switching of an individual Ag2S/Ag nanowire heterostructure, Nanotechnology 18, 485202-1-5 (2007).
[191] C. Schindler, K. Szot, S. Karthäuser, and R. Waser, Controlled local filament growth and dissolution in Ag-Ge-Se, Physica Status Solidi 2, 129-131 (2008).
[192] D. R. Stewart, D. A. A. Ohlberg, P. A. Beck, Y. Chen, R. S. Williams, J. O. Jeppesen, K. A. Nielsen, and J. F. Stoddart, Molecule-independent electrical switching in Pt/organic monolayer/Ti devices, Nano Letters 4, 133-136 (2004).
[193] R. Waser and M. Aono, Nanoionic-based resistive switching memories, Nature Materials 6, 833-840 (2007).
[194] J. G. Kushmerick, D. L. Allara, T. E. Mallouk, and T. S. Mayer, Electrical and spectroscopic characterization of molecular junctions, MRS Bulletin 29, 396-402 (2004).
[195] L. T. Cai, M. A. Cabassi, H. Yoon, O. M. Cabarcos, C. L. McGuiness, A. K. Flatt, D. L. Allara, J. M. Tour, and T. S. Mayer, Reversible bistable switching in nanoscale thiol-substituted oligoaniline molecular junctions, Nano Letters 5, 2365-2372 (2005).
[196] Y. Selzer, L. T. Cai, M. A. Cabassi, Y. Yao, J. M. Tour, T. S. Mayer, and D. L. Allara, Effect of local environment on molecular conduction: Isolated molecule versus self-assembled monolayer, Nano Letters 5, 61-65 (2005).
[197] I. Kratochvilova, M. Kocirik, A. Zambova, J. Mbindyo, T. E. Mallouk, and T. S. Mayer, Room temperature negative differential resistance in molecular nanowires, Journal of Materials Chemistry 12, 2927-2930 (2002).
[198] J. K. N. Mbindyo, T. E. Mallouk, J. B. Mattzela, I. Kratochvilova, B. Razavi, T. N. Jackson, and T. S. Mayer, Template synthesis of metal nanowires containing monolayer molecular junctions, Journal of the American Chemical Society 124, 4020-4026 (2002).
[199] B. R. Martin, D. J. Dermody, B. D. Reiss, M. M. Fang, L. A. Lyon, M. J. Natan, and T. E. Mallouk, Orthogonal self-assembly on colloidal gold-platinum nanorods, Advanced Materials 11, 1021-1025 (1999).
[200] P. A. Smith, C. D. Nordquist, T. N. Jackson, T. S. Mayer, B. R. Martin, J. Mbindyo, and T. E. Mallouk, Electric-field assisted assembly and alignment of metallic nanowires, Applied Physics Letters 77, 1399-1401 (2000).
208 [201] N. B. Zhitenev, H. Meng, and Z. Bao, Conductance of small molecular junctions, Physical Review Letters 88, 226801-1-4 (2002).
[202] T. Graves-Abe, Z. Bao, and J. C. Sturm, Self-aligned, insulating-layer structure for integrated fabrication of organic self-assembled multilayer electronic devices, Nano Letters 4, 2489-2492 (2004).
[203] T. Graves-Abe and J. C. Sturm, Programmable organic thin-film devices with extremely high current densities, Applied Physics Letters 87, 133502-1-3 (2005).
[204] M. E. Anderson, M. Mihok, H. Tanaka, L.-P. Tan, M. W. Horn, G. S. McCarty, and P. S. Weiss, Hybrid approaches to nanolithography: Photolithographic structures with precise controllable nanometer-scale spacings created by molecular rulers, Advanced Materials 18, 1020-1022 (2006).
[205] A. Hatzor and P. S. Weiss, Molecular rulers for scaling down nanostructures, Science 291, 1019-1020 (2001).
[206] M. E. Anderson, L.-P. Tan, H. Tanaka, M. Mihok, H. Lee, M. W. Horn, and P. S. Weiss, Advances in nanolithography using molecular rulers, Journal of Vacuum Science and Technology B 21, 3116-3119 (2003).
[207] M. E. Anderson, C. Srinivasan, R. Jayaraman, P. S. Weiss, and M. W. Horn, Utilizing self-assembled multilayers in lithographic processing for nanostructure fabrication: Initial evaluation of the electrical integrity of nanogaps, Microelectronic Engineering 78-79, 248-252 (2005).
[208] C. Srinivasan, M. E. Anderson, R. Jayaraman, P. S. Weiss, and M. W. Horn, Electrically isolated nanostructres fabricated using self-assembled multilayers and a novel negative-tone bi-layer resist stack, Microelectronic Engineering 83, 1517-1520 (2006).
[209] C. T. Black, Self-aligned self assembly of multi-nanowire silicon field effect transistors, Applied Physics Letters 87, 163116-1-3 (2005).
[210] C. T. Black and O. Bezencenet, Nanometer-scale pattern registration and alignment by directed diblock copolymer self-assembly, IEEE Transactions on Nanotechnology 3, 412-415 (2004).
[211] T. Bailey, B. J. Choi, M. Colburn, M. Meissl, S. Shaya, J. G. Ekerdt, S. V. Sreenivasan, and C. G. Willson, Step and flash imprint lithography: Template surface treatment and defect analysis, Journal of Vacuum Science and Technology B 18, 3572-3577 (2000).
209 [212] M. Colburn, A. Grot, M. Amistoso, B. J. Choi, T. Bailey, J. G. Ekerdt, S. V. Sreenivasan, J. Hollenhorst, and C. G. Willson, Step and flash imprint lithography for sub-100nm patterning, Proceedings of the SPIE 3997, 453-457 (2000).
[213] M. Colburn, S. Johnson, M. Stewart, S. Damle, T. Bailey, B. Choi, M. Wedlake, T. Michaelson, S. V. Sreenivasan, J. G. Ekerdt, and C. G. Willson, Step and flash imprint lithography: A new approach to high-resolution patterning, Proceedings of the SPIE 2676, 379-389 (1999).
[214] E. K. Kim, J. G. Ekerdt, and C. G. Willson, Importance of evaporation in the design of materials for step and flash imprint lithography, Journal of Vacuum Science and Technology B 23, 1515-1520 (2005).
[215] D. J. Resnick, W. J. Dauksher, D. Mancini, K. J. Nordquist, T. Bailey, S. Johnson, N. Stacey, J. G. Ekerdt, C. G. Willson, S. V. Sreenivasan, and N. Schumaker, Imprint lithography for integrated circuit fabrication, Journal of Vacuum Science and Technology B 21, 2624-2631 (2003).
[216] M. Austin and S. Y. Chou, Fabrication of nanocontacts for molecular devices using nanoimprint lithography, Journal of Vacuum Science and Technology B 20, 665-667 (2002).
[217] Y. Chen, D. A. A. Ohlberg, X. Li, D. R. Stewart, J. O. Jeppesen, K. A. Nielsen, J. F. Stoddart, D. L. Olynick, and E. Anderson, Nanoscale molecular-switch devices fabricated by imprint lithography, Applied Physics Letters 82, 1610-1612 (2003).
[218] S. Y. Chou, P. R. Krauss, and P. J. Renstrom, Imprint lithography with 25- nanometer resolution, Science 272, 85-87 (1996).
[219] L. Qin, J.-W. Jang, L. Huang, and C. A. Mirkin, Sub-5-nm gaps prepared by on- wire lithography: Correlating gap size with electrical transport, Small 3, 86-90 (2007).
[220] L. Qin, S. Park, L. Huang, and C. A. Mirkin, On-wire lithography, Science 309, 113-115 (2005).
[221] R. D. Piner, J. Zhu, F. Xu, S. Hong, and C. A. Mirkin, "Dip-pen" nanolithography, Science 283, 661-663 (1999).
[222] C. M. Fischer, M. Burghard, and S. Roth, Microelectronic and nanoelectronic aspects in conjugated systems, Materials Science Forum (New Materials: Conjugated Double Bond Systems) 191, 149-157 (1995).
[223] J. M. Seminario and P. A. Derosa, Molecular gain in a thiotolane system, Journal of the American Chemical Society 123, 12418-12419 (2001). 210 [224] A. M. Moore, B. A. Mantooth, Z. J. Donhauser, F. Maya, D. W. Price, Y. Yao, J. M. Tour, and P. S. Weiss, Cross-step place-exchange of oligo(phenylene- ethynylene) molecules, Nano Letters 5, 2292-2297 (2005).
[225] A. M. Moore, B. A. Mantooth, Z. J. Donhauser, Y. Yao, J. M. Tour, and P. S. Weiss, Real-time measurements of conductance switching and motion of single oligo(phenylene-ethynylene) molecules, Journal of the American Chemical Society 129, 10352-10353 (2007).
[226] H. Sakaguchi, J. Matsumura, and H. Gong, Electrochemical epitaxial polymerization of single-molecular wires, Nature Materials 3, 551-557 (2004).
[227] J. M. Tour, A. M. Rawlett, M. Kozaki, Y. X. Yao, R. C. Jagessar, S. M. Dirk, D. W. Price, M. A. Reed, C. W. Zhou, J. Chen, W. Y. Wang, and I. Campbell, Synthesis and preliminary testing of molecular wires and devices, Chemistry - A European Journal 7, 5118-5134 (2001).
[228] R. H. Mathews, J. P. Sage, T. C. L. Gerhard Sollner, S. D. Calawa, C.-L. Chen, L. J. Mahoney, P. A. Maki, and K. M. Molvar, A new RTD-FET logic family, Proceedings of the IEEE 87, 596-605 (1999).
[229] M. T. Cygan, T. D. Dunbar, J. J. Arnold, L. A. Bumm, N. F. Shedlock, T. P. Burgin, L. Jones II, D. L. Allara, J. M. Tour, and P. S. Weiss, Insertion, conductivity, and structures of conjugated organic oligomers in self-assembled alkanethiol monolayers on Au{111}, Journal of the American Chemical Society 120, 2721-2732 (1998).
[230] Z. J. Donhauser, Probing nanoscale electronics using scanning tunneling microscopy, Ph.D. Thesis, Department of Chemistry, The Pennsylvania State University, University Park, PA (2003).
[231] B. A. Mantooth, Nanoscopic analyses: Single molecule characterization in molecular electronics and surface science, Ph.D. Thesis, Department of Chemistry, The Pennsylvania State University, University Park, PA (2005).
[232] A. M. Rawlett, T. J. Hopson, I. Amlani, R. Zhang, J. Tresek, L. A. Nagahara, R. K. Tsui, and H. Goronkin, A molecular electronics toolbox, Nanotechnology 14, 377-384 (2003).
[233] C. W. J. Bauschlicher, A. Ricca, N. Mingo, and J. Lawson, On the current flow for benzene-1,4-dithiol between two Au contacts, Chemical Physics Letters 372, 723-727 (2003).
[234] J.-J. Hwang and J. M. Tour, Combinatorial synthesis of oligo(phenylene ethynylene)s, Tetrahedron 58, 10387-10405 (2002). 211 [235] J. W. Ciszek and J. M. Tour, Synthesis of ladder polyaromatics as new molecular device candidates, Tetrahedron Letters 45, 2801-2803 (2004).
[236] F. Maya, A. K. Flatt, M. P. Stewart, D. E. Shen, and J. M. Tour, Formation and analysis of self-assembled monolayers from u-shaped oligo(phenylene ethynylene)s as candidates for molecular electronics, Chemical Materials 16, 2987-2997 (2004).
[237] D. W. Price, S. M. Dirk, F. Maya, and J. M. Tour, Improved and new syntheses of potential molecular electronics devices, Tetrahedron 59, 2497-2518 (2003).
[238] Y. Yao and J. M. Tour, Facile convergent route to molecular caltrops, Journal of Organic Chemistry 64, 1968-1971 (1999).
[239] L. T. Cai, Y. X. Yao, J. P. Yang, D. W. Price, and J. M. Tour, Chemical and potential-assisted assembly of thiolacetyl- terminated oligo(phenylene ethynylene)s on gold surfaces, Chemistry of Materials 14, 2905-2909 (2002).
[240] L. A. Bumm, J. J. Arnold, L. F. Charles, T. D. Dunbar, D. L. Allara, and P. S. Weiss, Directed self-assembly to create molecular terraces with molecularly sharp boundaries in organic monolayers, Journal of the American Chemical Society 121, 8017-8021 (1999).
[241] J. M. Seminario, P. A. Derosa, and J. L. Bastos, Theoretical interpretation of switching in experiments with single molecules, Journal of the American Chemical Society 124, 10266-10267 (2002).
[242] Molecule naming schemes were obtained from Chemdraw (CambridgeSoft; 5.0; CambridgeSoft Corporation: Cambridge, MA, 1998) and Beilstein Commander (GmbH, M. I. S.; 4.0; MDL Information Systems GmbH: Frankfurt, Germany, 2000).
[243] R. G. Nuzzo, B. R. Zegarski, and L. H. Dubois, Fundamental studies of the chemisorption of organosulfur compounds on Au(111): Implications for molecular self-assembly on gold surfaces, Journal of the American Chemical Society 109, 733-740 (1987).
[244] M. J. Hostetler, J. E. Wingate, C.-J. Zhong, J. E. Harris, R. W. Vachet, C. M. D., J. D. Londono, S. J. Green, J. J. Stokes, G. D. Wignall, G. L. Glish, M. D. Porter, N. D. Evans, and R. W. Murray, Alkanethiolate gold cluster molecules with core diameters from 1.5 to 5.2 nm: Core and monolayer properties as a function of core size, Langmuir 14, 17-30 (1998).
[245] E. G. Emberly and G. Kirczenow, The smallest molecular switch, Physical Review Letters 91, 188301-1-4 (2003). 212 [246] P. E. Kornilovitch and A. M. Bratkovsky, Orientational dependence of current through molecular films, Physical Review B 64, 195413-1-4 (2001).
[247] P. Fenter, A. Eberhardt, and P. Eisenberger, Self-assembly of n-alkyl thiols as disulfides on Au(111), Science 266, 1216-1218 (1994).
[248] H. Basch, R. Cohen, and M. A. Ratner, Interface geometry and molecular junction conductance: Geometric fluctuation and stochastic switching, Nano Letters 5, 1668-1675 (2005).
[249] S. J. Stranick, A. N. Parikh, D. L. Allara, and P. S. Weiss, A new mechanism for surface diffusion: Motion of a substrate-adsorbate complex, Journal of Physical Chemistry 98, 11136-11142 (1994).
[250] D. J. Trevor and C. E. D. Chidsey, Room temperature surface diffusion mechanisms observed by scanning tunneling microscopy, Journal of Vacuum Science and Technology B 9, 964-795 (1991).
[251] M. Di Ventra, S. T. Pantelides, and N. D. Lang, First-principles calculation of transport properties of a molecular device, Physical Review Letters 84, 979-982 (2000).
[252] S. Wakamatsu, S. Fujii, U. Akiba, and M. Fujihira, Motions of single molecules inserted in a self-assembled monolayer matrix of a bicyclo 2.2.2 octane derivative on Au(111), Nanotechnology 14, 258-263 (2003).
[253] D. J. Fuchs and P. S. Weiss, Insertion of 1,10-decanedithiol in decanethiolate self-assembled monolayers on Au{111}, Nanotechnology 18, 044021-1-7 (2007).
[254] L. Häussling, B. Michel, H. Ringsdorf, and H. Rohrer, Direct observation of streptavidin specifically adsorbed on biotin-functionalized self-assembled monolayers with the scanning tunneling microscope, Angewandte Chemie, International Edition in English 30, 569-572 (1991).
[255] O. Chailapakul, L. Sun, C. J. Xu, and R. M. Crooks, Interactions between organized, surface-confined monolayers and vapor-phase probe molecules 7. Comparison of self-assembling n-alkanethiol monolayers deposited on gold from liquid and vapor-phases, Journal of the American Chemical Society 115, 12459- 12467 (1993).
[256] K. Edinger, A. Golzhauser, K. Demota, C. Woll, and M. Grunze, Formation of self-assembled monolayers of n-alkanethiols on gold - a scanning tunneling microscopy study on the modification of substrate morphology, Langmuir 9, 4-8 (1993).
213 [257] G. L. Kellogg and P. J. Feibelman, Surface self-diffusion on Pt(001) by an atomic exchange mechanism, Physical Review Letters 64, 3143-3146 (1990).
[258] P. J. Feibelman, Interlayer self-diffusion on stepped Pt(111), Physical Review Letters 81, 168-171 (1998).
[259] R. L. Schowoebel and E. J. Shipsey, Step motion on crystal surfaces, Journal of Applied Physics 37, 3682-3686 (1966).
[260] N. D. Lang and P. Avouris, Electrical conductance of individual molecules, Physical Review B 64, 5323-5328 (2001).
[261] G. E. Poirier and M. J. Tarlov, Molecular ordering and gold migration observed in butanethiol self-assembled monolayers using scanning tunneling microscopy, Journal of Physical Chemistry 99, 10966-10970 (1995).
[262] A. M. Moore, B. A. Mantooth, A. A. Dameron, Z. J. Donhauser, P. A. Lewis, R. K. Smith, D. J. Fuchs, and P. S. Weiss, Measurements and mechanisms of single- molecule conductance switching, Advances in Materials Research 10, 29-47 (2008).
[263] S. J. Stranick, M. M. Kamna, and P. S. Weiss, Atomic-scale dynamics of a two- dimensional gas-solid interface, Science 266, 99-102 (1994).
[264] J. A. Stroscio and R. J. Celotta, Controlling the dynamics of a single atom in lateral atom manipulation, Science 306, 242-247 (2004).
[265] K. Wang, C. Zhang, M. M. T. Loy, and X. Xiao, Time-dependent tunneling spectroscopy for studying surface diffusion confined in nanostructures, Physical Review Letters 94, 036103-1-4 (2005).
[266] H. Nyquist, Certain topics in telegraph transmission theory, Transactions of the A. I. E. E. 47, 617-644 (1928).
[267] C. E. Shannon, Communication in the presence of noise, Proceedings of the Institute of Radio Engineers 37, 10-21 (1949).
[268] X. Torrelles, E. Barrena, C. Munuera, J. Rius, S. Ferrer, and C. Ocal, New insights in the c(4x2) reconstruction of hexadecanethiol on Au(111) revealed by grazing incidence x-ray diffraction, Langmuir 20, 9396-9402 (2004).
[269] M. Yu, N. Bovet, C. J. Satterley, S. Bengio, K. R. J. Lovelock, P. K. Milligan, R. G. Jones, D. P. Woodruff, and V. Dhanak, True nature of an archetypal self- assembly system: Mobile Au-thiolate species on Au(111), Physical Review Letters 97, 166102-1-4 (2006). 214 [270] P. Maksymovych, D. C. Sorescu, and J. T. Yates, Gold-adatom-mediated bonding in self-assembled short-chain alkanethiolate species on the Au(111) surface, Physical Review Letters 97, 146103-1-4 (2006).
[271] S.-W. Tam-Chang, H. A. Biebuyck, G. M. Whitesides, N. Jeon, and R. G. Nuzzo, Self-assembled monolayers on gold generated from alkanethiols with the structure RNHCOCH2SH, Langmuir 11, 4371-7382 (1995).
[272] L. J. Richter, C. S.-C. Yang, P. T. Wilson, C. A. Hacker, R. D. van Zee, J. J. Stapleton, D. L. Allara, Y. Yao, and J. M. Tour, Optical characterization of oligo(phenylene-ethynylene) self-assembled monolayers on gold, Journal of Physical Chemistry B 108, 12547-12559 (2004).
[273] S. J. Stranick, P. S. Weiss, A. N. Parikh, and D. L. Allara, Alternating current scanning tunneling spectroscopy of self-assembled monolayers on gold, Journal of Vacuum Science and Technology A 11, 739-741 (1993).
[274] R. C. Miller, Optical second harmonic generation in piezoelectric crystals, Applied Physics Letters 5, 17-19 (1964).
[275] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. Montgomery, J. A., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Lui, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al- Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, and J. A. Pople, "Gaussian 03," Revision C.01, Wallingford: Gaussian, Inc., 2004.
[276] S. Yeganeh and M. A. Ratner, Polarizability calculations, personal correspondence (2008).
[277] G. H. Yang, Y. L. Qian, C. Engtrakul, L. R. Sita, and G.-Y. Liu, Arenethiols form ordered and incommensurate self-assembled monolayers on Au(111) surfaces, Journal of Physical Chemistry B 104, 9059-9062 (2000).
215 [278] F. P. Cometto, P. Paredes-Olivera, V. A. Macagno, and E. M. Patrito, Density functional theory study of the adsorption of alkanethiols on Cu(111), Ag(111), and Au(111) in the low and high coverage regimes, Journal of Physical Chemistry B 109, 21737-21748 (2005).
[279] L. H. Dubois and R. G. Nuzzo, Synthesis, structure, and properties of model organic surfaces, Annual Review of Physical Chemistry 43, 437-463 (1992).
[280] G. Yang and G.-Y. Liu, New insights for self-assembled monolayers of organothiols on Au(111) revealed by scanning tunneling microscopy, Journal of Physical Chemistry B 107, 8746-8759 (2003).
[281] D. E. Bergeron, A. W. J. Castleman, T. Morisato, and S. N. Khanna, Formation of - Al13I : Evidence for the superhalogen character of Al13, Science 304, 84-87 (2004).
[282] R. E. Leuchtner, A. C. Harms, and A. W. J. Castleman, Thermal metal cluster anion reactions: Behavior of aluminum clusters with oxygen, Journal of Chemical Physics 91, 2753-2754 (1989).
[283] A. C. Harms, R. E. Leuchtner, S. W. Sigsworth, and A. W. J. Castleman, Gas- phase reactivity of metal alloy clusters, Journal of the American Chemical Society 112, 5673-5674 (1990).
[284] S. N. Khanna and P. Jena, Atomic clusters: Building blocks for a class of solids, Physical Review B 51, 13705-13715 (1995).
[285] S. N. Khanna and P. Jena, Designing ionic solids from metallic clusters, Chemical Physics Letters 219, 479-483 (1999).
[286] S. N. Khanna, B. K. Rao, and P. Jena, Electronic signature of the magicity and ionic bonding in Al13X (X=Li-K), Physical Review B 65, 125105-1-5 (2002).
[287] E. McMillan and P. H. Abelson, Radioactive element 93, Physical Review 57, 1185-1186 (1940).
[288] G. T. Seaborg and A. C. Wahl, The chemical properties of elements 94 and 93, Journal of the American Chemical Society 70, 1128-1134 (1948).
[289] G. T. Seaborg, Transuranium elements: Past, present and future, Accounts of Chemical Research 28, 257-264 (1995).
[290] G. T. Seaborg, The chemical and radioactive properties of the heavy elements, Chemical and Engineering News 23, 2190-2193 (1945).
216 [291] O. Hahn, Applied radiochemistry (Cornell University Press, Ithica, 1936).
[292] C. Srinivasan, T. J. Mullen, J. N. Hohman, M. E. Anderson, A. A. Dameron, A. M. Andrews, E. C. Dickey, M. W. Horn, and P. S. Weiss, Scanning electron microscopy of nanoscale chemical patterns, ACS Nano 1, 191-201 (2007).
[293] J. C. Vickerman and D. Briggs, ToF-SIMS: Surface analysis by mass spectrometry (Manchester and IM Publications, Chichester, 2001).
[294] V. I. Nefedov, X-ray photoelectron spectroscopy of solid surfaces (VSP BV, Utrecht, Netherlands, 1988).
[295] J. F. Moulder, W. F. Stickle, P. E. Sobol, and K. D. Bomben, Handbook of X-ray photoelectron spectroscopy: A reference book of standard spectra for identification and interpretation of XPS data, J. Chastain editor (Perkin-Elmer Corporation, Physical Electronics Division, Eden Prairie, MN, 1999).
[296] N. Fairley, "CasaXPS," version 2.3.5, Freeware: http://www.casaxps.com/.
[297] D. J. Graham and B. D. Ratner, Multivariate analysis of ToF-SIMS spectra from dodecanethiol SAM assembly on gold: Spectral interpretation and ToF-SIMS fragmentation process, Langmuir 18, 5861-5868 (2002).
[298] M. Schroder, S. Sohn, and H. F. Arlinghaus, Investigation of secondary cluster ion emission from self-assembled monolayers of alkanethiols on gold with ToF- SIMS, Applied Surface Science 231, 164-168 (2004).
[299] T. Shahwan, H. N. Erten, B. L., and G. C. Allen, ToF-SIMS depth profiling analysis of the uptake of Ba2+ and Co2+ ions by natural kaolinite clay, Journal of Colloid and Interface Science 277, 23-28 (2004).
[300] M. Pelavin, D. N. Hendrickson, J. M. Hollander, and W. L. Jolly, Phosporus 2p electron binding energies. Correlation with extended Hückel charges, Journal of Physical Chemistry 74, 1116-1121 (1970).
[301] S. J. Splinter, R. Rofagha, N. S. McIntyre, and U. Erb, XPS characterization of the corrosion films formed on nanocrystalline Ni-P alloys in sulphuric acid, Surface and Interface Analysis 24, 181-186 (1996).
[302] J. A. Leiro and E. E. Minni, Satellite structure of the x-ray photoemission in Rb, Ca, and Ba metals, Physical Review B 31, 8248-8251 (1985).
[303] P. D. Schulze, S. L. Shaffer, R. L. Hance, and D. L. Utley, Adsorption of water on rhenium studied by XPS, Journal of Vacuum Science and Technology A 1, 97-99 (1983). 217 [304] A. Abdureyim, K. K. Okudaira, Y. Harada, S. Masuda, M. Aoki, K. Seki, E. Ito, and N. Ueno, Characterization of 4-mercaptohydrocynnamic acid self-assembled film on Au(111) by means of X-ray photoelectron spectroscopy, Journal of Electron Spectroscopy and Related Phenomena 114-116, 371-374 (2001).
[305] A. B. Christie, J. Lee, I. Sutherland, and J. M. Walls, An XPS study of ion-induced compositional changes with group-II and group-IV compounds, Applied Surface Science 15, 224-237 (1983).
[306] J. A. T. Verhoeven and H. Van Doveren, An XPS investigation of the interaction of CH4, C2H2, C2H4 and C2H6 with a barium surface, Surface Science 123, 369- 383 (1982).
[307] A. S. Duwez, Exploiting electron spectroscopies to probe the structure and organization of self-assembled monolayers: A review, Journal of Electron Spectroscopy and Related Phenomena 134, 97-138 (2004).
[308] P. E. Laibinis, C. D. Bain, and G. M. Whitesides, Attenuation of photoelectrons in monolayers of normal-alkanethiols adsorbed on copper, silver and gold, Journal of Physical Chemistry 95, 7017-7021 (1991).
[309] A. A. Dameron, T. J. Mullen, R. W. Hengstebeck, H. M. Saavedra, and P. S. Weiss, Origins of displacement in 1-adamantanethiolate self-assembled monolayers, Journal of Physical Chemistry C 111, 6747-6752 (2007).
[310] G. Beamson and D. Briggs, High resolution XPS of organic polymers: The Scienta ESCA 300 database (Wiley, New York, 1992).
[311] G. Barth, R. Linder, and C. Bryson, Advances in charge neutralization for XPS measurements of nonconducting materials, Surface and Interface Analysis 11, 307-311 (1988).
[312] M. Ayyoob and M. S. Hegde, Electron spectroscopic studies of formic acid adsorption and oxidation on Cu adn Ag dosed with barium, Journal of the Chemical Society, Faraday Transactions 82, 1651-1662 (1986).
[313] J. M. Buriak, Organometallic chemistry on silicon and germanium surfaces, Chemical Reviews 102, 1271-1308 (2002).
[314] S. F. Bent, Heads or tails: Which is more important in molecular self-assembly? ACS Nano 1, 10-12 (2007).
[315] T. He, J. Ding, N. Peor, M. Lu, D. A. Corley, B. Chen, Y. Ofir, Y. Gao, S. Yitzchaik, and J. M. Tour, Silicon/molecule interfacial electronic modifications, Journal of the American Chemical Society 130, 1699-1710 (2008). 218 [316] K. Demirkan, A. Mathew, C. Weiland, Y. Yao, A. M. Rawlett, J. M. Tour, and R. L. Opila, Energy level alignment at organic semiconductor/metal interfaces: Effect of polar self-assembled monolayers at the interface, Journal of Chemical Physics 128, 074705-1-6 (2008).
[317] B. Chen, M. Lu, A. K. Flatt, F. Maya, and J. M. Tour, Chemical reactions in monolayer aromatic films on silicon surfaces, Chemistry of Materials 20, 61-64 (2008).
[318] M. Lu, B. Chen, Y. Li, and J. M. Tour, Synthesis, grafting and film formation of porphyrins on silicon surfaces using triazenes, Chemistry of Materials 19, 4447- 4453 (2007).
[319] J. M. Seminario, A. G. Zacarias, and J. M. Tour, Molecular alligator clips for single molecule electronics. Studies of group 16 and isonitriles interfaces with Au contacts, Journal of the American Chemical Society 121, 411-416 (1999).
[320] X. Jiang, K. Ataka, and J. Heberle, Influence of the molecular structure of carboxyl-terminated self-assembled monolayer on the electron transfer of cytochrome c adsorbed on a Au electrode: In situ observation by surface- enhanced infrared absorption spectroscopy, Journal of Physical Chemistry C 112, 813-819 (2008).
[321] S. U. Nanayakkara, Precise and directed assembly and chemistry at the atomic scale, Ph.D. Thesis, Department of Chemistry, The Pennsylvania State University, University Park, PA (2006).
[322] J. G. Kushmerick, J. Lazorcik, C. H. Patterson, and R. Shashidhar, Vibronic contributions to charge transport across molecular junctions, Nano Letters 4, 639-642 (2004).
[323] J. M. Beebe, H. J. Moore, T. R. Lee, and J. G. Kushmerick, Vibronic coupling in semifluorinated alkanethiol junctions: Implications for selection rules in inelastic electron tunneling spectroscopy, Nano Letters 7, 1364-1368 (2007).
[324] A. Troisi, J. M. Beebe, L. B. Picraux, R. D. van Zee, D. R. Stewart, M. A. Ratner, and J. G. Kushmerick, Tracing electronic pathways in molecules by using inelastic tunneling spectroscopy, Proceedings of the National Academy of Science 104, 14255-14259 (2007).
[325] L. H. Yu, C. D. Zangmeister, and J. G. Kushmerick, Origins of discrepancies in inelastic electron tunneling spectra of molecular junctions, Physical Review Letters 98, 206803-1-4 (2007). 219 [326] W. Wang, T. Lee, I. Kretzschmar, and M. A. Reed, Inelastic electron tunneling spectroscopy of an alkanedithiol self-assembled monolayer, Nano Letters 4, 643- 646 (2004).
VITA
Amanda Michelle Moore
Education Ph.D. in Chemistry; The Pennsylvania State University University Park, Pennsylvania, August 2008, GPA: 3.9/4.0 Advisor: Professor Paul S. Weiss Thesis: Creating and Probing Molecular Assemblies for Single-Molecule Devices B.S. in Chemistry and Music (Summa Cum Laude); Pacific University Forest Grove, Oregon, May 2002, GPA: 3.9/4.0 Advisor: Professor Kevin E. Johnson Undergraduate Thesis: Degradation of Electrodes in Lithium-Ion Batteries
Professional Experience Graduate Research Assistant; The Pennsylvania State University Advisor: Prof. Paul S. Weiss (2002-2008) Teaching Assistant; The Pennsylvania State University General Chemistry Lecture, Honors Lecture, Recitation and Laboratory Summer Research Assistant; Pacific University Advisor: Prof. Kevin E. Johnson (2000-2002) Research Assistant; University of Arizona Advisor: Prof. Jeanne Pemberton (January 2002) Teaching Assistant; Pacific University (2000-2002) General Chemistry Laboratory and Grader
Honors and Awards 2007 Anna Louise Hoffman Award, Iota Sigma Pi 2007 Rohm and Haas Travel Award, The Pennsylvania State University 2007 AVS Dorothy M. and Earl S. Hoffman Travel Grant 2006 Apple and Geiger Fellowships, The Pennsylvania State University 2005 Department of Chemistry Travel Award, The Pennsylvania State University 2004 Department of Chemistry Travel Award, The Pennsylvania State University 2003 Department of Chemistry Fellowship, The Pennsylvania State University 2002 Cheney Scholar in the Natural Sciences, Pacific University 2001 Harold Zeh Scholarship, Portland Section of the American Chemical Society 2001 Promethium Chapter Iota Sigma Pi Scholarship
Professional Affiliations Sigma Xi American Vacuum Society American Chemical Society Iota Sigma Pi