Scanning Tunneling Microscopy and Related Methods

In these two lectures, I give a short introduction to local probe methods, pointing out certain aspects which very often do not receive proper attention, and present some of the interesting methods not represented in the main lectures. The various problems addressed should in no way discourage anybody from solving them, and doing it right.

In this set of two lectures the emphasis is on the tunneling of electrons between two semi-infinite electrodes with parallel, planar surfaces. Although directly relevant to scanning tunneling microscopy only in the “blunt tip” regime, it is hoped that they will provide a useful background for the other lectures. The topics include the stationary-state scattering and transfer Hamiltonian approaches to elastic and inelastic tunneling, resonant tunneling in semiconductor heterostructures, electron tunneling spectroscopy of single-particle electronic excitations and electron-phonon interactions in (conventional) superconductors and of the electron-phonon interaction in normal metals, inelastic electron tunneling spectroscopy of molecular vibrations, and single-electron charging effects. A third lecture is devoted to tunneling times.

The desire to minimize the response time of tunneling devices has led to a renewed interest in an unresolved fundamental question: “How long does an incident electron of energy E take, on average, to tunnel through a classically inaccessible region?” Since there is considerable controversy and confusion surrounding this problem, several recent approaches that illustrate the main points of dispute are discussed in some detail.

In the few years which have passed since its invention¹ by Binnig, Rohrer, and coworkers, scanning tunneling microscopy (STM) has established itself as a remarkably powerful and versatile tool for studying surfaces. This paper reviews the present theoretical understanding of STM, with emphasis on the interpretation of atomic-resolution STM images. Some familiarity with the basic principles of STM is assumed here; the basic ideas and instrumentation have been described in detail elsewhere.¹

A transition metal tip consisting of a single tungsten atom adsorbed on a W(110) surface is modelled. Chemisorpion theory is applied to treat the interaction of the tip atom with the flat W(110)-surface. The basis set on the tip atom includes 6s-, 6p- and 5d-orbitals. The importance of the d-electrons for the tunneling to an Al(111)-surface is investigated. Inclusion of the d- orbitals affects strongly the local electronic structure and modifies the current. For short distances a significant part of the total tunnel current flows directly via the tip d-orbitals. The theory is used to investigate the ‘barrier height’ of the clean surface. A corrugation of the Al(111)-surface is obtained in STM-theory, if a elastic deformation of the tip is taken into account.

In the earlier studies of Scanning Tunneling Microscopy the tip-sample distance was generally assumed to be large enough to disregard the tip-surface interaction effects, and the electrodes were considered to be independent. However, modifications of the electronic structure depending upon the tip-surface separation have led to the identification of different regimes (i.e. independent electrode, electronic contact and point contact regimes) in the operation of STM. As the tip approaches the sample surface, the tip-sample interaction gradually increases and the potential barrier is lowered. The charge density is rearranged and the ions of the tip and sample are displaced to attain the minimum of the total energy at the preset tip-sample distance. Owing to the overlap of the tip and sample states site-specific localized states appear and provide a net binding interaction, whereby the tip and sample are connected electronically. Upon further approach of the tip to the sample surface the tunneling barrier is perforated even before the tip enters in the strongly attractive force region. Eventually, the point contact regime is initiated, and new channels of conduction are opened through the electronic states localized in the gap. The conductance undergoes a qualitative change and its commonly accepted proportionality to the local density of states of the unperturbed sample surface is invalidated. In this lecture note, a microscopic analysis of the tip-surface interaction is presented and recent STM studies are examined within the framework of this analysis.

The electron elastic conductance of small contacts on constrictions between 2-D reservoirs is investigated. Experimental results in two-dimensional electron gas (2 DEG) GaAs/GaAIAs structures show that this conductance, to a good approximation, is quantized. We present calculations for general geometries of the contacts that describe the experimental data reasonably well. Our calculations also show how the resonant scattering structure superimposed on the quantized conductance for some geometries tend to be washed out when the contacts are made smooth. Finally we discuss the interpretation of those experiments in terms of ballistic transport. Calculations using a first iterative selfconsistent procedure show that it may not be correct to use ballistic transport for all the modes in the constriction. The importance of the electronic mean free path concept in the constriction when compared with that of the reservoirs is also discussed. We conclude that for the higher mode in the constriction the transport may be sequential.

The effect of geometry on the quantized steps and resonance structure is investigated for a quasi 1D constriction within the assumption of ballistic transport and by using the “mixed basis” boundary matching technique.

This paper is intended as a brief tutorial on work functions. It summarises some basic facts about the concept of “work function” as conventionally used in surface science. It then discusses various other parameters that get called “work function” in the contexts of field emission and scanning tunnelling microscopy.

The role and application of scanning tunneling microscopy (STM) for the investigation of clean and adsorbate covered metal surfaces are discussed. STM studies on the periodic structure and structural defects on these surfaces are reviewed. Experimental results related to atomic resolution imaging of close-packed metal surfaces are presented. The contribution of electronic effects to STM imaging of adsorbates is discussed, electronic modifications of the surrounding metal substrate atoms illustrate the effect of the chemical bond between adsorbate and substrate atoms. Time resolved observations of local adsorbate and substrate structures are shown to gain detailed information on various surface processes, in particular they allow direct access to mechanistic details of surface reactions.

The application of the scanning tunneling microscope (STM) to the study of clean and adsorbate covered semiconductor surfaces is discussed. Various imaging and spectroscopic methods are described, and the methods are illustrated with examples of experimental results. Emphasis is placed on the relationship between the electronic and structural properties of a surface, and the role that this relationship plays in the interpretation of STM images.

These notes describe two spectroscopic methods which use conduction electrons as spectroscopic probes. Mainly low temperature experiments will be discussed.

Local photon emission properties in the visible range of granular and island silver films have been characterized using STM. The experimental data are discussed in the light of model calculations. The observed fluorescence spectra and their dependence on tunnel voltage are interpreted theoretically in terms of radiative decay of local surface Plasmons excited dominantly via inelastic tunneling. Spatial mapping of the resulting integrated light intensity (photon maps) reveals a close relation to topographic features. The sharp spatial contrast, lack of significant polarity dependence, and the detection of fieid emission resonances in isochromat photon spectra favor inelastic tunneling as the dominant process.

The Tunneling and Force Microscopes are superb instruments for imaging atomic and molecular structures. From the beginning it has been clear that they can be used to modify surfaces in various ways. In this article we review the conventional methods and compare these with potential for surface modification with the STM and AFM. We reach the conclusion that the new instruments can be used for microfabrication of structures on solid substrates with a resolution that is improved by “a factor of ten beyond the present capabilities.”

Four scanning probe microscopes, the scanning tunneling microscope (STM),¹ the atomic force microscope (AFM),² the scanning electrochemical microscope (SEM)³ and the scanning ion-conductance microscope (SICM)⁴ have all been used to image liquid-solid interfaces. Images in this report illustrate the variety of systems that can be studied: from iron corroding in salt water to selenium atoms in liquid nitrogen and from graphite covered with vacuum grease to proteins covered with oil. The technological and biological importance of liquid-solid interfaces has driven and will ensure rapid growth in this field.

We discuss principal aspects and experimental concepts of STM at potential-controlled electrodes in electrolytic environment and illustrate them with typical electrochemical applications.

I review the state of the art of scanning tunneling microscopy in biology, and present STM images of coated and uncoated biological macromolecules. I further discuss the electron-transfer mechanisms which allow STM imaging of mesoscopic organic objects.

We have investigated some aspects of the interface between metallic biomedical implants and living tissue. To characterize this interface we have developed an STM coupled with a high resolution reflective optical microscope. The STM is used in connection with other microscopic techniques: TEM, SEM, and Auger spectroscopy. Initial biological measurements show that fibronectin (Fn), a glycoprotein which plays an important role in cell attachment and bacterial promotion, is attached on Ti and V substrates, but is not biologically active on V, a non-biocompatible substance. We present STM images of the Ti oxide layer and of single and multiple Fn molecules shadowed with a conductive layer.

Uncoated DNA molecules marked with an activated Tris 1-Aziridinyl Phosphine Oxide (TAPO) solution were deposited on gold substrates and imaged in air with high resolution Scanning Tunneling Microscope (STM). Constant current and barrier height STM images show a clear evidence of the helicity of the DNA structure: pitch periodicity ranges between 25 and 34 Å while the average diameter is 20 Å. Molecular structure within a single helix turn is also observed.

The results of scanning tunneling microscopy (STM) at ambient conditions of several representatives of inorganic layered materials (α-RUCl₃, VS₂), conductive charge transfer complexes (tetracyanoquinodimethane with different donors: tetrathiofulvalene, 4-ethyl pyridine and triethylammonium) and conductive polymers (polypyrrole, polythiophene) are presented. In the case of (α-RUCl₃ the threedimensional distortion of surface unit cell of chlorine atoms has been discovered. The STM imaging of atomic scale features of VS₂ surface had indicated the existence of charge density waves at room temperature. The STM images of crystal surfaces of the charge transfer complexes give detailed information concerning the surface charge distribution which has been analyzed in comparison with X-ray data. The arrangement of polymer chains in polypyrrole and polythiophene is discussed on the basis of the STM results.

To interpret tunneling current through a molecule, it is important to distinguish between through bond and through space tunneling processes. A definition of these processes is proposed. For each one, the methods to calculate the tunneling current are briefly presented. A tight-binding example is analytically solved to illustrate them. Recent results on the STM imaging of Alkanes are discussed by considering through bond processes.

We develop a model for calculating the conductance through disordered linear chains. The behaviour of the conductance for disordered systems shows a strong dependence with the length of the system, number of chains and energy of the incoming electrons. We suggest a link of these results and STM measurements on biologicals. Some of the features of those experiments could be explained in terms of properties of disordered systems.

This paper deals with point sources for charged particles. The importance of a well-characterized ensemble of particles in the framework of the quantum mechanical measuring process is illustrated. The engineering of these sources for electrons and noble gas ions and their emission properties are reviewed. Two practical microscopy applications, the resolution of which depends on the “quality” of the particle beam, are presented.

The influence of quantum size effects and atomic geometry on the field emission characteristics (energy distribution, intensity-voltage, angular spread of the emitted beam, resolution) of atomic-size microtips is analyzed. Theoretical models are propose to analyze the influence of atomic-size protrusions on the properties of the emitted beam. It is shown that a tip with a small protrusion can be used as a source of very collimated electron beams. The conditions under which is it possible to obtain atomic resolution are discussed. Simple formulas relating the angular spread and the resolution with the experimental parameters are presented. Field emission experiments on “build up” and “teton” tips are presented together with the basic principles of the fabrication technique.

The pressure sensitivity of commercially available electret microphones is typically about 1 mV/µbar. A pressure of 1 µbar corresponds to a force of 10^-6 N acting on the diaphragma. We show that, with suitable electronics, it is possible to detect — at the center of the diaphragm — even forces in the range of 10^-9 N.

This sensitivity is in an interesting range, because the interaction forces occuring in standard tunneling microscopy between the tip and sample are of the same order of magnitude.

We present data of STM experiments under ambient air conditions on samples of graphite (H0P6) and of gold mounted at the center of the diaphragm. During the tunnel process the height of the tip was modulated (z-direction) with an amplitude of about 1 nm at a frequency of 2 kHz. We were able to measure the vibrations of the diaphragm caused by the periodic interaction forces between tip and sample.

In summary, we present a new method of simultaneous tunneling and force-microscopy.

In scanning near field optical microscopy (SNOM), an antenna of subwavelength size is scanned accross the object in close proximity. The modulated light emission due to near field interactions with the object is recorded for image formation. With 50 — 100 nm apertures in a metal film or protrusions serving as antennas a resolution of 30 — 100 nm — limited by the antenna size — is demonstrated. Phase contrast depending on electric excitation and contrast enhancement depending on local plasmon excitation of the antenna are demonstrated in two types of reflection mode SNOM, one of which — scanning optical tunneling microscopy — is closely analogous to STM.

For overpassing the classical limit of resolution in optical microscopy, it is necessary to detect the light diffracted from small objects in the near field and not in the far field as in classical microscopy. A particular case is the detection of the evanescent field lying on the surface of a guiding structure. These surface waves interact with the object details and then can be used for determining the topography of the object. The main problem is the detection because the light beam is confined on the object surface. A solution consists of frustrating the evanescent field by means of a dielectric probe. The conversion of the inhomogeneous waves into homogeneous ones is fundamentally similar to the electronic tunneling effect. Subwavelength resolution can be obtained by placing a suitable optical stylus connected to an optical fibre near the surface. A xyz piezo-electric micropositioning system allows then to scan the object surface under test.

Scanning Nearfield Acoustic Microscopy (SNAM) is a new method for imaging the topography of non-conducting surfaces using a vibrating tip of high Q as distance sensor. A 32 kHz quartz tuning fork is used as an oscillator driven at resonance by a feedback loop. The feedback signal is used both to excite and to detect the oscillation. An edge of one leg serves as a tip for imaging. If the tip is approached to the surface, the resonance frequency and the oscillation amplitude both decrease. The dependence of this decrease on the gas pressure shows that hydrodynamic forces in the gas and adsorbed vapor films are responsible for the interaction. For imaging, the sensor is piezoelectrically scanned accross the surface. The distance is feedback controlled such that the vibration amplitude stays constant. Using the control signal for imaging, a resolution of 1.5 µm laterally — limited by the tip geometry — and of 10 nm vertically are demonstrated.