UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The growth of novel thin film oxides : calibrations, preliminary growths and quantitative characterizations Wicks, Ryan Christopher 2007

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


ubc_2007-0640.pdf [ 7.43MB ]
JSON: 1.0084919.json
JSON-LD: 1.0084919+ld.json
RDF/XML (Pretty): 1.0084919.xml
RDF/JSON: 1.0084919+rdf.json
Turtle: 1.0084919+rdf-turtle.txt
N-Triples: 1.0084919+rdf-ntriples.txt
Original Record: 1.0084919 +original-record.json
Full Text

Full Text

The Growth of Novel Thin Film Oxides Calibrations, Preliminary Growths and Quantitative Characterizations by Ryan Christopher Wicks B.Sc, Queen's University, 2005 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF Master of Science in The Faculty of Graduate Studies (Physics and Astronomy) The University Of British Columbia October, 2007 © Ryan Christopher Wicks 2007 Abstract In this thesis, preliminary growths of SrCuO^ epitaxial films were under taken. Despite none of the growths yielding the desired stoichiometry or oxidation, the difficulties that were encountered, such as properly oxidiz ing the film, are demonstrated to be surmountable. The structure of the films was probed with RHEED, and found to display growth patterns that are consistent with epitaxial growth. The desire to perform future ARPES studies on this material imposes strict sample quality requirements and the need for the most accurate measurements of stoichiometry possible. This warranted an investigation of the properties of the SPECS PHOIBOS 150 hemispherical electron analyzer used to perform XPS measurements of films chemistry. The degree of non-linearity of the electron detection system was determined and a method introduced that could correct for it. Correcting for the transmission function of the analyzer was found to be a necessary step in quantitative chemical analysis. Therefore, transmission functions were mea sured using two techniques and were found to be in good agreement with calculated transmission functions provided by SPECS. While correcting for detector non-linearity and measuring the actual transmission function did not drastically alter the stoichiometry measurements for scans taken using reasonable analyzer settings for XPS ( 1% for the films grown in this thesis, mostly from detector non-linearity), it is possible to drive the analyzer into regimes where the corrections will be necessary. The methods used to first correct for the intensity response of the analyzer, and then the transmission function would be more applicable in these instances. The work performed in this thesis will ensure that future growths of SrCuOi films will yield samples of adequate quality to perform an ARPES study. ii Table of Contents Abstract : ii Table of Contents iiList of Tables v List of Figures • • vi List of Abbreviations yiiAcknowledgements ix Dedication x 1 Introduction 1 1.1 Motivation . . .1.2 Scope of Thesis 5 2 Calibration of an Electron Analyzer 7 2.1 Introduction . 7 • 2.1.1 The Electron Analyzer .,2.2 Experiment and Results 14 2.2.1 Intensity Calibration2.2.2 The Transmission Function . . • 20 2.3 Conclusions : 25 3 Growth of SrCu02 31 3.1 Introduction3.1.1 Molecular Beam Epitaxy 31 3.2 Experimental 3 3.2.1 Reflection High Energy Electron Diffraction 33 3.2.2 X-Ray Photoelectron Spectroscopy- .......... 37 3.3 Results • • • • 41 iii Table of Contents 3.4 Conclusions • 44 4 Creation of Atomic Oxygen 46 4.1 Introduction 44.1.1 Creation of Atomic Oxygen . . . : 46 4.2 Experimental . . . 47 4.3 Results • 48 4.4 Conclusions 51 5 Conclusions 2 Bibliography 54 Appendices ^ A Intensity Calibration Programs 58 A.l Calibration Script 5A.2 Data Correction . . 9 List of Tables 2.1 The Various Lens Modes of an. Electron Analyzer [14] .... 9 3.1 Summary of growth parameters for SrCu02 ' • • • 42 i v List of Figures 1 A graduate student in thesis writing mode x 1.1 A pictorial demonstration of epitaxy. 3 1.2 The Polar Catastrophe and Electronic Reconstruction 4 2.1 Components of an Electron Analyzer (Adapted From [13]) . . 8 2.2 XPS scans of the same SrCu02 sample in FAT and FRR modes 11 2.3 Intensity response of the SPECS PHOIBOS detection system. 13 2.4 A portion of a Cu spectrum, normalized for X-ray power, but not intensity corrected : . . 17 2.5 A portion of a Cu spectrum.that has been normalized for X-Ray power and intensity corrected. . 18 2.6 Measured intensity response of the SPECS PHOIBOS detec tion system 19 2.7 Fit to data for Method 1 22 2.8 Measured and Expected Product of (PE)(RR) for two Differ ent Ratios 23 2.9 In versus In ^ for an electron gun 26 2.10 Offset corrected and fitted data from the electron gun 27 2.11 Comparison of the measured and calculated transmission func tion 28 2.12 Measured FR.R transmission function 29 3.1 The Oxide MBE in the Quantum Materials Laboratory at UBC (from [22]) 32 3.2 A RHEED diffraction pattern of a Mica sample 34 3.3 An explanation of RHEED oscillations 35 3.4 Au (111) surface reciprocal lattice reconstructed from RHEED data. 36 3.5 Pictorial description of the photo-emission process 37 3.6 An example XPS scan of a SrCuOi sample . 38 vi List of Figures 3.7 Comparison of CuO. and Cu20 2p structure 42 3.8 RHEED time series for SrCu02 growth #1 3 3.9 RHEED time series for SrCu02 growth #3 44 4.1 Flux of AO as a Function of Power at a Constant 02 Pressure of 6.2 x 1CT6 Torr (1.1 seem Throughput) 49 4.2 Flux of AO as a Function of Throughput at a Constant Power of 250 W : . 50 List of Abbreviations AO - Atomic Oxygen ARPES - Angle Resolved Photo-emission Spectroscopy CAE - Constant Analyzer Energy (also FAT) CCD - Charge Coupled Detector CRR - Constant Retard Ratio (also FRR) FAT - Fixed Analyzer Transmission FRR - Fixed Retard Ratio LEED - Low Energy Electron Diffraction MBE - Molecular Beam Epitaxy MCP - Multi-Channel Plate PLD - Pulsed Laser Deposition r PVD - Physical Vapour Deposition QCM - Quartz Crystal Monitor RF - Radio Frequency RHEED - Reflection High Energy Electron Diffraction SCCM - Standard Cubic Centimeters per Minute STM - Scanning Tunneling Microscopy TMO - Transition Metal Oxide XPS - X-Ray Photoelectron Spectroscopy Acknowledgements I would like to acknowledge the support of both my supervisors, Dr. Andrea Damascelli and Dr. Nicholas Ingle. They have both spent considerable time and effort to take me from an ignorant student fresh out of undergrad, to a full-fledged graduate student, just as ignorant, but better able to answer his own questions. Despite each of them having a different perspective and approach to doing science, I know I am a better scientist for having the opportunity to work with both of them. I would also like to thank the other students and post-docs who worked in the lab with me. You are a valuable source of information and assistance, but, more importantly, you can appreciate a dirty joke, and have made coming to the lab each day interesting. ix For my parents, who got me here, and for Marja, who still loves me despite me being me (See figure 1). Removed Pending Copyright Permission Figure 1: A graduate student in thesis writing mode. x Chapter 1 Introduction 1.1 Motivation The transition metal oxides (TMOs) are a strange class of materials. They exhibit a broad spectrum of unique and inexplicable properties, such as colossal magneto-resistance and high temperature superconductivity. The failure of conventional condensed matter physics to explain even the most basic properties of the TMOs' is due to the localized and strongly correlated nature of their outermost electrons. The strong localization of electrons lies at the root of the TMOs more exotic properties. As science begins to obtain a better understanding of the nature of electron correlations in condensed matter systems, new vistas in device fabrication will be revealed, ushering in an era where the striking and bizarre properties of strongly correlated systems, like the TMOs, can be tailored to suit the needs of human ingenuity. TMOs are characterized by strongly localized electronic orbitals. In a noble metal or semi-conductor, the outermost electrons are of primarily s and p character. The large spatial extent of these orbitals means that the electron density is spread roughly evenly throughout the solid. The inter actions between electrons in this approximately homogeneous electron gas are well described by a non-interacting, single particle picture (specifically through the local density approximation of density functional theory). The TMOs, on the other hand, have outermost electrons that consist of d orbitals which are strongly localized. The electron density is no longer homogeneous in these systems, and the electronic correlations cannot be described by the local density approximation. In principle, electrons on one site can tunnel to adjacent sites, but the large electronic density localized at the new site invokes a strong Coulomb repulsion (called U) that prevents this. Systems with half filled orbitals which would normally be metals can behave as in sulators with gaps as large as 5-10 eV (Mott-Hubbard insulators, in this case). Strong localization also makes possible various exchange interactions between sites, leading-to magnetism in these systems. The host of exciting properties that are displayed by these strongly localized, highly correlated systems make them interesting both from a scientific and technological point 1 Chapter 1. Introduction of view. In an attempt to understand these types of materials, theoretical ap proaches have been developed that go beyond the local density approxi mation picture. Tools developed in the field of particle physics to deal with many body interactions have also been adapted for use in condensed matter systems with some success. Overall, however, our understanding of electronic correlations is still incomplete. Fortunately, there are a large number of experimental probes that can be used further our understanding of electronic correlations in solids, and TMOs in particular; for example, Angle Resolved Photo-Emission Spectroscopy (ARPES) or Scanning Tun neling Microscopy (STM). Unfortunately, surface sensitive techniques, such as these, require the use of atomically flat and clean samples to extract meaningful data. This limits the number of systems available for study to those with a natural cleavage plane (a 2-D layer of the crystal that has a weaker bond than adjacent layers, which will break preferentially when subject to mechanical stress). A heuristic approach to studying these materials would be to modify a parameter of the system and then measure how its properties change. The most common example is chemical doping. Other options could include changing the lattice parameter of the system, or the bond angles between atoms. In a de-localized system, these changes would only have small ef fects on the overall properties because the electronic densities remain almost unchanged. In a system with localized electrons however, changing the over lap between adjacent orbitals could have enormous effects. Unfortunately, changing lattice parameters and bond angles independent of chemical doping is a difficult challenge in conventional crystal growth, where thermodynam ics puts tight constraints on how the material grows. An alternate approach to modifying a system's parameters uses an alter nate method of crystal growth called epitaxy. The goal of epitaxial growth is to form a thin film of single crystalline material through some deposition process on a suitable single crystal substrate. There are two main advantages to epitaxial growth. One advantage of an epitaxial growth technique is that a flat surface, suitable for surface sensitive experiments, is left after growth. The second advantage is that changing the seed substrate changes the po tential that the film feels as it forms, creating thermodynamically accessible meta-stable phases that would be otherwise untenable. Epitaxial films are characterized by a shared unit cell at the interface [1, 2]: the substrate and film are "lattice matched" at their common interface. In homo-epitaxy (where the film and substrate are made of the same material) this is not surprising. This behavior in hetero-epitaxy (epitaxy between two different 2 Chapter 1. Introduction materials), on the other hand, offers a variety of interesting possibilities for crystal growth. For example, materials normally grown at high pressures by conventional means can be grown epitaxially in a vacuum environment; the potential of the seed substrate provides the needed strain [3]. Hetero-epitaxy can also cause a film to grow with a modified structure, such as altered lattice parameters or bond angles [4]. Furthermore, this structural modification can be achieved without resorting to chemical doping. Figure 1.1: A pictorial demonstration of epitaxy (courtesy of Nicholas J.C. Ingle). The growth of lattice matched films does present its own set of draw backs. Growth of a strained film promotes the formation of defects in an attempt to relax the film to its thermodynamically stable state. It has been demonstrated, however, that strained TMO films can be grown well past the point where conventional energy and strain considerations predict that they would form defects and relax [3, 5]. Hetero-epitaxy also offers other exciting prospects for engineering the properties of TMOs. If the chosen substrate has a polar unit cell, as many TMO substrates do, the need to balance this polarity leads to many novel possibilities for altering the film's behaviour. Semi-conductors will normally compensate for this "polar catastrophe" by allowing defects and changes in stoichiometry and/or structure at the interface (structural or chemical 3 Chapter 1. Introduction reconstruction)^]. In some systems, however, it is energetically favourable to change the oxidation level of the film's constituents at the interface to compensate for the polar nature of the film rather than to adjust its structure or chemistry [7]. This is called an "electronic reconstruction". Figure 1.2: The Polar Catastrophe and Electronic Reconstruction. This figure shows the different possible arrangements at the interface between films of SrTiOz and LaAIOz (which has a polar unit cell). Parts a and b demonstrate the two ways in which these films can stack, and the resultant divergence of potential with film thickness (the so called "polar catastro phe"). Parts c and d show how this divergence can be avoided by allowing oxidation state of half of the Tz4+ to change to Ti3+ (an overall change to Tz35+). In this case, the polar catastrophe is avoided, and the half of an electron per unit cell that is moved to the interface results in a metallic layer between the two insulators. From [8] This aspect of oxide thin film research has already begun to realize its promise, with the first exciting results of Hwang and coworkers discovering a metallic layer between two insulating oxides (SrTiOz and LaTiOs) [9] spurring a flurry of related activity. Figure 1.2 provides an explanation of how the interface between insulators can form a metallic layer through an electronic reconstruction. Superconductivity at this interface has also been demonstrated [10]. In the case of cuprate based films, this could be an interesting mechanism for doping the CuO planes deemed essential for high temperature superconductivity without changing the film's stoichiometry (the cuprates are 2-D materials that contain corner sharing CuO planes, and 4 Chapter 1. Introduction are the only materials that exhibit high temperature superconductivity). The. potential of using epitaxial growth techniques on TMOs is enor mous. There is the obvious advantage of being able to prepare materials with an atomically flat surface, making a wider variety of materials avail able for surface sensitive measurements. It must be remembered, however, that epitaxy is not only a method for preparing bulk samples suitable for surface science. Ultra-thin films can be distinct materials with properties that are totally unique from their bulk counterparts [11]. The ability to con trol the structure through the interaction of the substrate on the film could have enormous impact on the electronic properties of the material due to the TMOs strongly localized nature. Electronic reconstructions at the interface between substrate and film also present exciting possibilities for tuning the TMOs properties. 1.2 Scope of Thesis In this thesis, growth of SrCu02 is performed on SrTiOz substrates using Molecular Beam Epitaxy (MBE). SrCuOz has the simplest structure that still contains the CuO planes considered necessary < for high temperature superconductivity. This makes it an interesting candidate for an ARPES study, but this has not happened due to the lack of a cleavage plane in the material. Thin film growth makes this sort of study possible. To perform a meaningful ARPES study, however, a high sample quality is paramount. Scientific advances in ARPES studies of the TMOs have been more a result of increased sample quality than instrumental resolution. This requires that a careful quantitative chemical analysis of our XPS data is undertaken. As a result, an examination and quantification of the properties of a SPECS PHOBIOS 150 hemispherical electron analyzer is undertaken. The principles of operation and properties of this device will be presented, with an emphasis on the qualities of the analyzer that will be investigated more closely, specifically the transmission function of the analyzer, and the linearity of the detector response with electron intensity. Several methods for determining the transmission function and intensity response of the analyzer will be introduced, and the results of these measurements are discussed. The results of the chemical analysis reveal that the oxidation level of the SrCu02 films achieved with the use of molecular Oxygen is insufficient to properly oxidize the SrCuOi films. As a result, the viability of an alternate Oxygen delivery system is investigated. The effectiveness of a Mantis Deposition MATS 30 RF Atom Source for creating Atomic Oxygen 5 Chapter'!. Introduction (AO) is determined and compared to other previously characterized AO sources used to grow similar materials epitaxially. The preliminary growths performed in this thesis are the first steps to wards performing a systematic ARPES study of SrCuOi thin films which before now was impossible due to the inability to cleave the material. This is not only exciting because it offers the chance to study the simplest of the cuprate materials with a surface sensitive measurement technique, but also because epitaxial growth offers a variety of novel mechanisms for altering the electronic structure of the material; through substrate lattice matching to change the lattice parameters and bond angles, or through a substrate/-film interaction, such as an electronic reconstruction at the substrate/film interface. The lessons learned through these preliminary growths, and the subsequent analyzer and AO source calibrations ensure that future studies will have the highest chances for success. 6 Chapter 2 Calibration of an Electron Analyzer 2.1 Introduction The requirement of samples of high quality in performing future ARPES measurements on SrCuOz films requires that the stoichiometry and oxida tion of the films be known to the highest possible precision. To extract this information from a thin film, the best tool is photo-emission spectroscopy. The details of the technique will be discussed more thoroughly in section 3.2, but at its heart, the measurement involves collecting electrons being excited out of the sample into the vacuum, discriminating them as a function of their kinetic energy, and then recording this spectrum. The tool used to do this is an electron analyzer, and this chapter looks at different techniques for correcting non-linearities in its intensity response and transmission. 2.1.1 The Electron Analyzer The function of an electron analyzer is to measure the kinetic energy dis tribution of collected charged particles. This ability has made them a very popular and versatile tool for both research and industry. There are several companies offering commercial analyzers. There have been many different designs, but the most common design is the electrostatic hemispherical an alyzer. For a description of other analyzer designs, please see reference [12]. It should be noted that, in principle, any charged particle can be studied; however, in practice, these analyzers are only used in electron spectroscopies. This is primarily due to' the low count rates found in experiments that study other types of particles. Most electron analyzers are designed to operate at higher count rates than are found in these types of studies. Modern hemispherical electron analyzers have two active components (figure 2.1). The first stage is the accelerating/focusing stage. Electrons enter the analyzer through an electron lens system. This stage has the dual task of focusing electrons on the entrance of the hemisphere, as well as 7 Chapter 2. Calibration of an Electron Analyzer Removed Pending Copyright Permission Figure 2.1: Components of an Electron Analyzer (Adapted From [13]) accelerating them in preparation for the next stage. Earlier electron ana lyzers could only accelerate the impinging electrons, preparing them for the second stage energy discrimination. Modern analyzers, with their electron optics, make possible a variety of different types of focusing (which usu ally require the use of a 2-D detector). In this thesis, a SPECS PHOIBOS 150 hemispherical electron analyzer is used. Some of the modes available on the this analyzer include transmission modes which boost low-signals, magnification modes that magnify a part of a sample, and angle resolved modes useful in Angle Resolved Photo-Emission experiments. These modes are detailed in table 2.1. The second component of the electron analyzer is the hemispherical capacitor. The electrons entering between two concentric hemispheres at different voltages follow a curved path from the entrance slit to the detection system due to a Lorenz force acting on the electrons. The detection system can be either a Charge Coupled Detector (CCD) recording a phosphour screen detecting the electrons that have been amplified by a Multi-Channel Plate (MCP) in newer systems, or a set of channeltrons in older or custom designed systems. The range and median kinetic energy of the electrons that are incident on the screen, as well as the resolution of the analyzer are all defined by the "pass energy" of the analyzer. The concept of a pass energy needs to be defined because the term oc-8 Chapter 2. Calibration of an Electron Analyzer Mode Characteristics Use Magnification -spatially resolved studies with non-monochromated x-ray sources -acceptance area defined by lens, not iris Small Area XPS Transmission -spatially resolved studies requiring strong signal -acceptance area defined by iris, not lens < ' High Energy or Low In tensity XPS Angular Dis persion -Fourrier electron optics give angular dependent information -increased angular range limits energy range Angle Resolved Photo-emission Spectroscopy (ARPES) Table 2.1: The Various Lens Modes of an Electron Analyzer [14] curs often in discussions of the operation of an electron analyzer. Electrons that enter the center of the hemispherical analyzer with a kinetic energy that equals the " pass energy" of the analyzer will arrive at the center of the detection system. Electrons that enter the system with energies not at the pass energy are deflected by an amount proportional to the difference be tween their energy and the pass energy. Hence, the hemispherical capacitor acts as an energy discriminator. The window of kinetic energies exposed to the detector is proportional to the pass energy, as is the resolution of the analyzer. Modes of Operation The narrow energy window exposed to the detection system at a fixed pass energy is rarely sufficient to perform a proper experiment. Data is normally acquired by scanning the analyzer through a range of kinetic energies. In practice, this is done by changing the voltages on different parts of the electron analyzer so that only electrons with the desired kinetic energy reach the detection system. The energy window exposed to the detection system at different kinetic energies is then averaged to produce the final spectrum. There are two ways of doing this. The mode used historically in Auger Electron Spectroscopy is the Fixed Retard Ratio mode (FRR mode) or Constant Retard Ratio (CRR mode). In this mode, the accelerating potential in the electron lens system is held constant (the retarding ratio is constant). To measure a range of kinetic 9 Chapter 2. Calibration of an Electron Analyzer energies, the pass energy of the hemisphere is scanned. Scanning the pass energy of the hemisphere has the additional effect of changing the range of kinetic energies that are exposed to the detection system. The main ad vantage of this mode is that one has high intensity at large kinetic energies due to a wider energy window in this range; data collected at high kinetic energy has a larger energy window exposed to the detector, and is sampled for longer, resulting in more intense photoelectron peaks. This will be dis cussed later, in the context of a transmission function in FRR mode that is approximately linear in kinetic energy. The disadvantage of this mode is that the increased energy window that increases the signal also changes the energy resolution (the resolution worsens as kinetic energy increases). FRR mode is the ideal choice for studies that are interested in identifying the po sition of weak photoelectron peaks at high kinetic energy where resolution is a secondary concern, In XPS, Fixed Analyzer Transmission mode (FAT mode), or Constant Analyzer Energy mode (CAE mode), is used. In this mode; the hemispheres pass energy is held constant (fixing the analyzers transmission), and the elec trons are accelerated to the pass energy by the lens system. Now, the energy window exposed to the detector remains a constant width as the kinetic en ergy is scanned. The advantage of this approach is that the instrumental resolution, which is proportional to the pass energy, remains constant over the entire kinetic energy range. The primary disadvantage is a lowering of intensity at high kinetic energies relative to a spectra obtained in FRR mode (see figure 2.2). FAT mode is used when a quantitative analysis of peak shape and size is required. The Transmission Function In addition to the two primary modes of operation, there are many other settings that can be changed on modern spectrometers. In the case of the SPECS PHOBIOS analyzer, an iris, as well as entrance and exit slits change the acceptance area or the acceptance angle of the analyzer. Different com binations of voltages in the lens also provide different types of focusing of the electrons entering the analyzer (these are the lens modes, see table 2.1). Changing any of these settings also changes the intensity of the final spec trum. The function that describes these changes is called the transmission function. The transmission function describes the ability of the electron analyzer to allow electrons to pass through it. Ideally, the analyzer would have no impact on the signal it was measuring. In reality, the analyzer has a significant effect 10 Chapter 2. Calibration of an Electron Analyzer Cu LMM Kinetic Energy (eV) Figure 2.2: XPS scans of the same SrCuOz sample in FAT and FRR modes. Notice that in FRR mode the intensity of the signal increases as the kinetic energy increases. This highlights the importance of understanding the be haviour of the transmission function in different analyzer modes when per forming quantitative peak analysis. Also notice how the width of the peaks in FRR mode increase as the kinetic energy increases (relative to the FAT mode peaks). 1 1 Chapter 2. Calibration of an Electron Analyzer on the measured signal (see figure 2.2). The transmission function describes how the intensity depends on the kinetic energy being measured; if one wants to make a quantitative measurement of intensity, it has to be corrected for. The transmission function of an electron analyzer is rarely taken into account properly when performing experiments. Approximate theoretical transmission function corrections are applied to the data, or tabulated, ana lyzer specific structure factors are used (see section 3.2.2). This is primarily due to the fact that measuring a transmission function exactly is a difficult task. In this thesis, two separate methods for determining the transmission function are investigated. Linearity of Detector Response The transmission function describes the kinetic energy dependent intensity response of the analyzer. There is another effect that alters the measured signal: the intensity response of the detection system. Ideally, the flux measured by the detection system is proportional to the flux incident on the detection system. In reality, at large count rates, the system saturates and no longer responds linearly with applied flux (see figure 2.3). This high intensity behaviour is undesirable, but expected. Most measurement systems have a dead time, the time after a measurement during which the detector cannot respond to new events. At low flux, the period between events is much longer than the dead time of the detector, and all of the events are recorded. At high flux, however, the period between events approaches the dead time, creating the possibility that a measurement event occurs during the dead time and is not recorded. As the incident flux continues to increase, the detector approaches the saturation limit, where the measured rate is equal to the inverse of the dead time. In this limit the detector is useless. A common but limiting method for avoiding this problem is to simply ensure that the detector is always operated in the low count, linear range, A better way would be to measure, and then correct- for, the intensity response of the detection system, making available a much larger range of detection. Not only does this increase the dynamic range of the detection system, but it has been previously demonstrated that the response of the detection system commonly used in new electronic analyzers (MCP/phosphour screen/CCD detection system) has- a low intensity quadratic response to applied flux [15]. This surprising result indicates that limiting the range of intensities is an inadequate method of dealing with the linearity of the detector, and that measuring the intensity response is a necessary calibration step for performing quantitative photoelectron spectroscopy. 12 Chapter 2. Calibration of an Electron Analyzer Sample Current (nA) Figure 2.3: Intensity response of the SPECS Phobios detection system (a MCP/phosphour screen/CCD based system). The system responds linearly over a large region. The saturation of the detector can be seen at high intensities. A small quadratic response is evident at low intensities (inset) (From [16]). L Chapter 2. Calibration of an Electron Analyzer 2.2 Experiment and Results 2.2.1 Intensity Calibration Correcting for the detection systems non-linearities must be done before calculating the transmission function. The method used in this thesis to correct for non-linearities in the detection system is based on the one used in reference [15].. This method involves taking several scans of the same sample at different X-ray powers. Scans of the same sample taken at N different x-ray powers yields broad range of intensities at all sampled kinetic energies. The premise of the method is that the real intensity can be represented as a linear combination of the measured intensity raised to different exponents (up to a maximum exponent of P, chosen as the lowest exponent that gives a reasonable fit) (equation 2.1). r(nj,Ek) = ' ^cii x (m(nj, £fc))* (2.1) =o r(rij,Ek) is the real spectrum, corrected for detector non-linearities, as a function of kinetic energy Ek and x-ray power rij, while m(nj, Ek) is the uncorrected, measured spectrum. The spectra each contain Q equally spaced kinetic energy values. The initial co-effiecient, ao is the dark count value, or.the value that all the scans approach as the x-ray power goes to 0. As all of the spectra are of the same sample in the same position, and have the same analyzer settings, the real spectra normalized by current should all be equivalent. Substituting equation 2.1 into equation 2.2 yields equation 2.3. r(nj,Ek) r(ni,Ek) rij n\ p , p (2.2) 1 y 1 -— 22aix(m(nj>Ek))% = — yjoi x Mni.Efc))* (2-3) ni i=o ni i=o j labels scans between 2 and N. Scan 1 is chosen to be in the linear range of the detector, as all of the other scans will be corrected relative to it. The choice of an appropriate scan 1 is the most subjective part of the process as it is based entirely on the users experience with the analyzer, and can also introduce an unavoidable systematic error because of the wide range of intensities in a single spectra. • . 14 Chapter 2. Calibration of an Electron Analyzer Arbitrarily setting a\ = 1 (this is an algebraic convenience, and allowed because real and measured count rates are allowed to different by an arbi trary factor) and rearranging equation 2.3 gives: m(rij,Ek) m(ni,Ek) ( 1 — Q0 X n-j ni 1 + 5>i ni rij '{min^E'k))1 (m(nj,Ek)y i=2 V 711 UJ ) This can be re-written as a matrix equation via the following definitions: B as ap m2k! ~ milt, m2k2 ~ mut2 m2kQ ~ mlkQ m3kl - mlkl m3k2 - mifc2 ™3fcc mike mNkl - mlkl mNki - mlk2 "rriNkr (2.5) (2.6) C = _1_ ; 1_ (jnlkl)2 K*! )2 (mlfel)3 m n.2 "i 1_ (™lfc2) (w2fc2)2 ni ri2 m U2 m J_ 1_ (™ifcQ)2 (mNkQ)2 (mifcQ)3 n\ m UN ni Q2fc2)3 n2 (mNkQ)3 nN . Kt/ _ (rn2fc1)J m • H2 ("l2fc,)P n,2 (.mlkQ)p _ (mNkQ)p nN 15 Chapter 2. Calibration of an Electron Analyzer B-CA = 0 • (2.8) A is a column vector of length P, B is a column vector of length. (iV — 1) x Q and C is a (N — 1) x Q by P matrix, is the j'th kinetic energy point of the i-'th dataset. j goes from 1 to Q, the length of each spectrum, and i goes from 2 to N, the number of spectra. Equation 2.8 is equivalent to equation 2.4. Simply solving for A will not yield the coefficients because of statistical errors found in real data. The best that can be done is to find the most likely values by minimizing |B — CA|2 (solving the normal equation). Va|B-CA|2 = 0 (2.9) 2CRCA - 2CTB = 0 (2.10) : A = (.CTC)-1 x CTB (2.11) . Solving equation 2.11 will yield the coefficients needed to express the true intensity in terms of the measured intensity. The values can be substituted into equation 2.1 and applied to any other scan as long as the intensity values are within the range that was used in the calibration. The algorithm used for finding the coefficients and correcting the data are provided in Appendix A. This algorithm was successfully applied to a set of seven Copper scans taken in Medium Magnification mode. A section of the data around the Cu 2p peaks after normalization by x-ray power but before intensity correction is shown in figure 2.4. While the different datasets are similar, there is a definite difference in the various intensities. The inset shows the normalized intensity of the 2p3 peak. Ideally, the data would be a flat line. This data does display a trend that suggests that the detection system is being operated in both the saturation and low intensity regimes, as well still having a multiplicative scaling factor missing. To claim this with certainty, however, the true current of electrons being photo-emitted from the sample should be measured directly, rather than inferred through an indirect measurement of X-ray power. The algorithm was applied to correct the data to various powers: A plot '•> of the data corrected with a P — 6 exponent is shown in figure 2.5. Higher order than P = 6 does not significantly improve the quality of the correction. A mode or measurement yielding higher intensities and hence stronger non-linearities would be expected to require higher order corrections. The data in figure 2.5 now appear to be properly normalized because the different datasets now lie on top of each other, and the intensity of the 2p§_ peak as a 16 Chapter 2. Calibration of an Electron Analyzer Binding Energy (eV) Figure 2.4: A portion of a Cu spectrum, normalized for X-ray power, but not intensity corrected. The inset shows the peak area intensity of the Cu 2p3 peak. The intensity appears to display non-linearity, saturating at high X-ray powers, and also appears to display quadratic behaviour at low intensities. There is also a multiplicative constant that was not accounted for in the normalization. 17 Chapter 2. Calibration of an Electron Analyzer function of X-ray power is flat. This indicates that the non-linearities have been corrected. 1200 1100 1000 900 Binding Energy (eV) 800 700 Figure 2.5: A portion of a Cu spectrum.that has been normalized for X-Ray power and intensity corrected. This is a considerable improvement over the uncorrected data shown in figure 2.4. The non-linear behaviours appears to have been corrected by the algorithm with a maximum exponent of P = 6. The range of intensities of the inset is same as the inset of figure 2.4. In figure 2.6, the measured intensity is plotted as a function of the actual (corrected) intensity. This plot bears many similarities to the data provided by SPECS regarding the intensity response of the detection system (figure 2.3). The detector is beginning to saturate at high intensities, and displays quadratic behaviour at low intensities. The saturation at high intensities is by far the most pronounced, but only the most intense peaks were affected. However, the settings used to collect this data were probably conservative in terms of maximum intensities for an average XPS measurement. Therefore, correcting these non-linearities is an important step in performing any quan-18 Chapter 2. Calibration of an Electron Analyzer titative analysis of data collected by this electron analyzer. This includes calculating transmission functions, as well as performing quantitative chem ical analysis. In the case of the SrCu02 films grown in this thesis (discussed in the next chapter), correcting for the intensity accounted for an approxi mately 1% difference in the measurement of stoichiometry. Actual Intensity Figure 2.6: Measured intensity response of the SPECS PHOIBOS detection system. The measured intensity is plotted as a function of the actual (cor rected) intensity. The blue line represents linear behavior. At high count rate, the detector begins to saturate, and at low count rates, the detector displays a quadratic response. This agrees qualitatively with the expected analyzer response, as well as the information provided by the manufacturer (see figure 2.3). The data shown in figures 2.4 and 2.5 correspond to the high count rate region of this graph. 1!) Chapter 2. Calibration of an Electron Analyzer 2.2.2 The Transmission Function Once the intensity response has been corrected, it is possible to measure the analyzer's transmission function. If the transmission function for one combination. of settings of an' analyzer is known, it is trivial to find the transmission function for any other combination of settings experimentally. By measuring a spectra of any sample with the analyzer set to a mode with a known transmission function, and then re-measuring the same sample with only the analyzer settings changed, the unknown transmission function can be extracted. mri ^-Unknown rp jp fry i cs\ "unknown — j ^known . v.^-"/ 1 known There have been many different techniques suggested to measure the transmission function of an electron analyzer [17, 18]. Here, two techniques have been used. The first technique, the FAT/FRR ratio, is very simple, but assumes' a known transmission function for one mode and a limited functional form for the other. The second technique, the transmission func tion using peak intensities, is more robust than the first method because it assumes a more flexible functional form for the transmission function with out assuming a known transmission function for any mode. Each technique yields a function that is proportional to the transmission function, but since all quantitative XPS measurements yield relative quantities, this is not a problem. Both methods take into account and correct for the additional kinetic energy dependance of the photoelectron cross section, cr, and the mean free path of the photoelectrons, F(Ek). ( Method I: FAT/FRR Ratio The premise of this method is to assume a known form for the FRR mode transmission function (a linear transmission function) and use this to find the transmission function in FAT mode by assuming a functional form for it [19, 20]. The ideal transmission functions for the two modes based on electron optics calculations are: P F'2 K TFAT(Ek) = -g— (2.13) TFRR(Ek) = |£ (2.14) 20 Chapter 2. Calibration of an Electron Analyzer Ek is the kinetic energy, PE is the pass energy of the hemisphere (both in units of eV), RR is the retard ratio of the lens and K is an analyzer dependent constant (both dimensionless). If one takes a scan in both FRR and FAT mode of the same sample with the same photon flux, the ratio for an ideal analyzer will have the form m).K™i! (,15) k By taking the ratio of measured spectra, all of the terms from equation 3.4 that do not depend on the analyzer mode cancel out. If the analyzer is not ideal, the ratio is modified by assuming that FRR mode retains its linear form, but that FAT mode becomes an inverse function of kinetic energy raised to a variable power, resulting in the following form [20]: R(Et) = • (2.16) Where a: is a constant with a value close to 1 (a value of 1 would indicate ideal behaviour), x is found by fitting the measured R(Ek) to equation 2.16. The transmission function in FAT mode is then given by: ''VAT. DC ^ . (2.17) The primary advantage of this method is that it is very easy to imple ment. Figure 2.7 shows R(Ek) for settings of medium magnification, iris 23, slits set to 1.6X20 and 2:open, pass energy of 60 eV and a retard ratio of 15. As seen in figure 2.7, it is not possible to fit the entire range of data with the simple function given by equation 2.16. At high KE, the exponent x is closer to the theoretical value of 2 and the fitted constant is closer to the expected product of the retard ratio and pass energy raised to the exponent l+x. At low KE, the function does fit the data well, but the values obtained are not what one would expect. This suggests that the model presented in equation 2.16 is not valid at low kinetic energies. This is supported by the data in figure 2.8. Figure 2.8 is a plot of the expected and measured product of the pass energy and retard ratio. Fits were done to each 100 eV section of R(Ek) 21 Chapter 2. Calibration of an Electron Analyzer Kinetic Energy (eV) Figure 2.7: Fit to data for Method I. The fit to the full dataset does not match the data well, and the fitted parameters are not what one would expect based on the model in equation 2.16 (the constant does not equal ((PE)(RR))1+X and the exponent does not equal 2). At low kinetic energy the fitted values deviate even more strongly from the expected behaviour. At high kinetic energies, however, the fitted values agree well with the expected fit values. 22 Chapter 2. Calibration of an Electron Analyzer 1400 > 1200 9 « 1000 T3 c 1 800 — rx X 600 CD LU (0 10 CD n. 400 60PE/20RR Expected 60 PE / 20 RR 60PE/15RR Expected 60PE/15RR 200 i i i i i i I I I I i i i i 200 400 600 800 1000 Kinetic Energy (eV) 1200 1400 Figure 2.8: Measured and Expected Product of (PE)(RR) for two Different Ratios. The values fit at high kinetic energies agree well with the model function, equation 2.16. At low kinetic energies, however, the fit values begin to deviate from the expected behavior. 23 Chapter 2. Calibration of an Electron Analyzer for scans with the same settings as figure 2.7. The measured values were obtained from the fitted constant, C, and fitted exponent, (1 + x). Figure 2.8 show that at low kinetic energy, the measured constant devi ates more strongly from the expected value. This suggests that this method is not effective at low kinetic energies. It is shown in the next section that the reason for this behavior is that the assumption of a linear transmission function in FRR mode is incorrect: Method II: Transmission Function From Peak Intensities This method of determining the transmission function assumes a functional form for the transmission function, yet allows enough freedom in the pa rameters of that model to fit any function [17, 18, 21]. It does not require assuming a known transmission function for any analyzer mode. The func tional form of the transmission function chosen for FAT mode is given by equation 2.20 Where n is a function of This is the same functional form as was postulated in method I, but now, instead of a constant exponent x, we have a function n (^§f Inserting equation 2.20 into equation 3.4, placing the various constants into the factors A and B(Ek), and taking the natural logarithm of both sides yields: C = l(PE)(RR)} {PE)(RR) = C-(1+x) (2.18) (2.19) (2.20) A = xfcf>yA B(Ek) = a(Ek)F(Ek) (2.21) (2.22) I (2.23) In I RE = ]nA + ]nB(Ek)+n PE ) In PE (2.24) 24 Chapter 2. Calibration of an Electron Analyzer One approach to collecting data to find the transmission function is to •measure peak areas from XPS scans of the same sample at different pass energies. Alternatively, one can use an electron gun at different kinetic energies. In this thesis, an electron gun was used instead of a XPS scan because it provided more peaks at evenly spaced kinetic energy levels. "(f)—f-('"f)2c '<"•> The data is shown in figure 2.9 was taken with the same analyzer settings used for correcting the detector non-linearity (Medium Magnification mode, slits set to 1.6:20, 2:open and with an iris of 15), with In plotted as a function of In The logarithm of the data must be taken so that when it is plotted, multiplicative factors become constant additive constants (In A and InB(Ek) from equation 2.20). For each kinetic energy value, the constants can be subtracted off so that data from different kinetic energies can be compared to each other. Finally, the function n (chosen to be a quadratic in this case) can be fitted to the data. See figure 2.10 for the offset corrected and fitted data. Higher order than quadratic did not significantly improve the fit of n. The resultant FAT mode transmission function agreed well with the calcu lated function provided by SPECS as shown in figure 2.11. A small deviation is present, but it is still within the spread of the SPECS data. The measured difference in stoichiometry between the measured and calculated transmis sion function of the SrCu02 films grown in the next chapter was much less than 1% (not correcting for the transmission function at all, however, introduced a difference of about 2 — 3%). The calculated and experimental transmission functions were used to cal culate the actual transmission function in FRR mode using equation 2.12. The experimentally determined FRR transmission function is shown in fig ure 2.12. The data does not exhibit the expected linear behaviour. This is most likely the reason for the poor results obtained using Method I. The FRR transmission function found using the calculated data from SPECS deviated from linear behaviour in a similar way. 2.3 Conclusions A method for correcting the intensity response of the detection system of a SPECS PHOIBOS 150 Hemispherical Analyzer has been investigated. 25 Chapter 2. Calibration of an Electron Analyzer 7.6 7.4 7.2 111 0-•— £ 7.0 6.8 6.6 6.4 X • + + + 300eVKE 500eVKE 700 eV KE a900eVKE • • • HOOeV KE X X ><1300eVKE I I I I I I I I I I I ll I I I I I I I I I I I _L_L -4.0 -3.5 -3.0 -2.5 In(PBKE) -2.0 -1.5 Figure 2.9: In -p^. versus In for an electron gun. Each point represents the area under the main peak from an electron gun at a fixed current. For each kinetic energy, a scan was taken at different analyzer pass energies. By plotting ln-^g versus hi^, unknown multiplicative constants can be subtracted off. 26 Chapter 2. Calibration of an Electron Analyzer ln(PE/KE) Figure 2.10: Offset corrected and fitted data from the electron gun. This data is the same as shown in figure 2.9, but with the unknown offsets cor rected. The fitted n is the unknown transmission function exponent from equation 2.20. It has the form n f^j = 0.717 + 0.196 x In ffjl1| +0.038 x In \EkJJ 27 Chapter 2. Calibration of an Electron Analyzer Figure 2.11: Comparison of the measured and calculated transmission func tion for FAT mode, medium magnification lens setting, pass energy 40 eV and Iris 15. The measured transmission function agreed well with the cal culated one.The calculation was performed by SPECS. 28 Chapter 2. Calibration of an Electron Analyzer 200 400 600 800 1000 1200 1400 Kinetic Energy (eV) Figure 2.12: Measured FRR transmission function for Medium Magnifica tion, Retard Ratio 20 and Iris 15. The measured spectra deviates form the expected linear behaviour. This is the most probably reason for the failure of Method I, as the method is based on this assumption. 29 Chapter 2. Calibration of an Electron Analyzer Sets of data taken with conservative powers and analyzer settings for XPS measurements have shown non-linearities in their intensity response. The method presented was successful in removing these non-linearities from the data. This procedure should be used before performing any quantitative peak analysis, including measuring the transmission function of the ana lyzer. This correction created a 1% improvement in measurements of relative chemical abundance of the SrCuOi films. Two methods were used to measure the transmission function of the an alyzer. The first method, the FAT/FRR method, was been found to rely on an incorrect assumption of a linear transmission function in FRR mode. The second method, the transmission function from peak intensities, has been successfully used to determine the analyzer's transmission function. The measured transmission function was found to agree well with the calculated transmission function provided by SPECS (making almost no difference in the measured stoichiometry of the SrCuC-2 films). Using this measured transmission function, it was also demonstrated that the FRR mode of this analyzer does not have a linear transmission function, and that this was the most likely cause of the failure of FAT/FRR method. Correcting the non-linearities of the detector is a necessary step in per forming any quantitative peak analysis with this analyzer. Measuring the transmission function may not be a necessary step, however, as long as a calculated transmission function is used. Not correcting for the transmission function at all introduces significant errors, on the order of 2 — 3%. These types of corrections will be necessary in future to ensure the quality of the SrCuO-z films is high enough to perform a meaningful ARPES study. 30 Chapter 3 Growth of SrCu02 3.1 Introduction Growing SrCuO,2 films epitaxially offers some advantages over conventional bulk growth. For instance, epitaxial growth results in a very thin sam ple composed primarily of surface atoms and a flat surface suitable for an ARPES or STM study. This growth technique makes it possible to perform-these surface sensitive measurements on wider variety of materials (ones that do not cleave to expose a-flat surface). Unfortunately, the structural and chemical probes available for thin film samples are severely limited. This can be a problem due to the high sample quality required to extract meaningful information from the materials. The thinness of the film makes normal bulk sensitive probes much less accurate or impossible. However, there are several options for examining the structure and chemistry of films. The majority of these techniques rely on the short mean free path of elec trons in solids to probe only the surface of the material. Techniques such as electron diffraction and X-Ray Photoelectron Spectroscopy (XPS) can probe the structure and and chemistry of a thin film without being influenced by the substrate. In this chapter, the growth of SrCuOi films is discussed, followed by a discussion of the results of Reflection High Energy Electron Diffraction (RHEED) and XPS characterization measurements. 3.1.1 Molecular Beam Epitaxy The goal of thin film crystal growth is to form a thin film of single crys talline material through some deposition process on a suitable single crystal substrate. To have epitaxial growth, the film and substrate must have their lattice constants and crystal symmetry match at the interface between them (as opposed to thin film crystal growth, where there does not need to be any lattice matching). In this thesis, the technique of Molecular Beam Epitaxy (MBE) is used to prepare epitaxial SrCuO-i thin films. MBE is a Physical Vapour Deposition (PVD) technique. The MBE technique involves the evaporation of elemental sources in an ultra-high 31 Chapter 3. Growth of SrCuC-2 vacuum environment. Sources are heated, whether with an electron beam, a laser or directly (using a Knudsen cell), until their vapour pressure is high enough to allow a reasonable amount of material to traverse the distance between the source and the substrate. The low pressure of the growth chamber means that the evaporated material has a very long mean free path (many times larger than the sample/substrate distance). The evaporated material is collimated, producing a "molecular beam" that impinges the sample. The rate of evaporation can be precisely controlled to an accuracy of less than a mono-layer of deposited atoms and the resulting films are extremely clean, due to the ultra high vacuum environment. Removed Pending Copyright Permission Figure 3.1: The Oxide MBE in the Quantum Materials Laboratory at UBC (from [22]) The Oxide MBE in the Quantum Materials Laboratory is a system cus tom designed with simplicity, modularity and connectivity in mind. The sample holding system is compatible with the lab's ARPES chamber and transfer chamber, allowing films to be transported from the growth cham ber to the analysis chamber without breaking the vacuum. Sources include several Knudsen cells with various temperature capabilities, a Mantis De position RF Atom Source, a sputter gun and several gas leak valves. Rates are monitored with a Quartz Crystal Monitor (QCM) on a custom designed stage that flips to expose the sample to the molecular beams. Until re-32 Chapter 3. Growth of SrCu02 cently, substrate heating was accomplished with filaments incorporated into the sample holder.. Substrates are now -heated by a custom external C02 iaser-system and temperatures are measured with a pyrometer. A high en ergy electron gun and detection system are used to measure Reflection High Energy Electron Diffraction (RHEED) patterns (explained in section 3.2.1). Molecular Beam Epitaxy Vs. Pulsed Laser Deposition The majority of the work done on the growth of oxide thin films has utilized the technique of Pulsed Laser Deposition (PLD) to grow samples, another type of PVD. In this process, a high power laser (normally a KrF excimer -laser) is fired repeatedly into a sintered target made of the film material. The target material is vaporized and travels through the vacuum to form a film on the substrate. The rate of deposition is controlled by changing the lasers firing rate. The difference between PLD and laser MBE is that the PLD target is a composite material with the same stoichiometry as the desired film, while the laser MBE target is an elemental source that will react with other materials at the substrate surface to form the desired film. The primary advantage of the PLD process is its simplicity. The disadvantage of the PLD technique, and the reason it was not used for this work, is that the laser/target interaction and its effect on the va porized material is very complicated and poorly understood. The process creates an ionized plasma plume at .the laser impact site, and the nature of the ablated units is unknown. This presents possible problems for perform ing APRES measurements on the samples in the future. The sample quality requirements are very stringent in APRES, and in an attempt to avoid pos sible complications arising from the unknown nature of the ablated units, the MBE technique is used. 3.2 Experimental 3.2.1 Reflection High Energy Electron Diffraction Reflection High Energy Electron Diffraction (RHEED) is an electron diffrac tion technique that can be used during film growth. High energy electrons are fired at the sample at a glancing angle, and a phosphour screen on the other side of the chamber is used to collect the reflected diffraction pattern. The diffraction pattern produced by RHEED satisfies the normal condi tions for diffraction (the Bragg condition, or the Ewald construction) except that only the in plane lattice parameters are probed. This is due to the 33 Chapter 3. Growth of SrCu02 short mean free path of electrons in the crystal perpendicular to the surface. While the electrons energy is high, and one would expect a deeper penetra tion, the electrons only hit the surface at a glancing angle therefore do not penetrate very far. This implies that the reciprocal lattice that is sampled is 2-D. The diffraction pattern exhibits "stripes" as a result (see Figure 3.3). This can be explained qualitatively by realizing that, in the plane, the elec trons sample a large area, with many diffraction planes, resulting in discrete points when switched to reciprocal space. Out of the plane, however, the electrons sample only a short distance, resulting in an extended rod when switched to a reciprocal space representation. Figure 3.2: A RHEED diffraction pattern of a Mica sample. The 2-D nature of the diffraction pattern is evident in the extended, rod-like nature of the diffraction spots. Traditionally, RHEED is used to monitor growth rate by measuring os cillations in the intensity of the diffracted pattern. A perfectly flat surface will produce the maximum diffracted intensity. As material is deposited on that surface, the flat surface begins to roughen and the intensity of the beam lessens as the scattering becomes increasingly diffuse. As a full mono-layer of material is deposited, the surface becomes flat again (assuming 2-D growth), and the intensity reaches another maximum. The period of oscillation of the RHEED pattern corresponds to the rate of mono-layer growth. 34 Chapter 3. Growth of SrCu02 Removed Pending Copyright Permission Figure 3.3: An explanation of RHEED oscillations. As the atomically flat surface begins to roughen, the amount of diffuse scattering of the electron beam decreases to a minimum corresponding to maximum surface roughness. As the amount of material deposited begins to approach-that required for a full mono-layer, diffuse scattering decreases and the RHEED intensity increases to a maximum corresponding to a full mono-layer deposited. From [23] 35 Chapter 3. Growth of SrCuOz RHEED can also be used to extract crystallographic information. Nor mally, only one direction of the crystal is probed by the beam. The change in one lattice parameter relative to its corresponding substrate parameter can be monitored as a function of time (see figures 3.8 and 3.9). It is pos sible, however, to take RHEED images at various angles and reconstruct the 2-D reciprocal lattice of the surface. This provides information similar to Low Energy Electron Diffraction (LEED), except that the data can be taken during crystal growth. This is particularly useful for observing changes in crystal symmetry or changes in lattice parameter at different stages of growth [22] (an example is given in figure 3.4). While this technique is not used in this thesis, it is mentioned because it promises to be an invaluable tool in later growths of SrCuOz films. Figure 3.4: Au (111) surface reciprocal lattice reconstructed from RHEED data. The pattern has the expected symmetry for a (111) Au surface. 36 Chapter 3. Growth of SrCuOi 3.2.2 X-Ray Photoelectron Spectroscopy XPS measurements involve exposing the sample under study to X-ray ra diation from an X-ray source or synchrotron light source and collecting the photo-emitted electrons. An example spectra is shown in figure 3.6. The kinetic energy of the photo-emitted electrons is given in the simplest ap proximation by equation 3.1. Ek = hco-<t>-EB (3.1) Removed Pending Copyright Permission Figure 3.5: Pictorial description of the photo-emission process. An incoming photon of energy %LO excites electrons out of their bound levels in the solid and promotes them to empty vacuum levels (from [12]). Ek is the kinetic energy of the photo-emitted electrons, Ti is Planck's constant, UJ is the photon frequency (hto is the energy of the photon), and 37 Chapter 3. Growth of SrCu02 Cu 2p3/2 Cu 2p1/2 200 400 600 800 1000 1200 1400 Kinetic Energy (eV) Figure 3.6: An example XPS scan of a SrCu02 Sample. This spectra can be used to extract many different quantiites, including the oxidation level of the Cu and the relative concentrations of Sr and Cu. 38 Chapter 3. Growth of SrCuO^ EB is the binding energy of the electron (measured relative to Ef, the Fermi energy). </> is the work function of the material (the difference between the vacuum level and Ef) and is included because of the convention of measuring binding energies from the Fermi level and not the vacuum level. This is process is, demonstrated pictorially in figure 3.5. A more accurate way of representing the photo-emission process is the following: Ek = hu - <f> - (Ef'1 - Ef) (3.2) Now, instead of the binding energy, the kinetic energy depends on the difference between the energies of the final state of the system with one electron removed (E^~l) and the initial N electron state (Ef)(N is on the order of Avagadro's Number). Fortunately, for XPS data where core level information is the most important, one can usually use the relation in equation 3.3. EB = (Ef^-E?) (3.3) This gives us back equation 3.1 when substituted into equation 3.2. In systems where electronic correlations are important, however, one must take into account the possibility that the core hole left behind by the photo-emission' process can interact with other particles in the system. Then, equation 3.2 must then be used. This description of the photo-emission process is suited to a scientist interested in understanding the details of the electronic structure of the material. For a film grower who is more interested in characterizing their samples, an experimental description of the photo-emission process is more appropriate. The intensity of a photoelectron peak is given by equation 3.4 [21, 24]. . I = xfa{Ek)<t>yAT{Ek)F{Ek) (3.4) < Where x is the atomic concentration, / is the x-ray flux, Ek is the electron kinetic energy, cr{Ek) is the photoelectron cross section, <j> is the instrumental geometry factor (a function of the angle between the photon beam and the detector, normally constant), y is the efficiency factor for the production of a photoelectron that leaves the system in the N-l system 39 Chapter 3. Growth of SrCuO^ ground state, A is the area from which the photoelectrons are emitted, F(Ek) is the mean free path of the photoelectrons in the sample, and T(Ek) is the transmission function of the analyzer. All of the physics based initial and final state energy considerations are lumped into a(Ek) and y, leaving the experimentalist free to deal with the other terms in equation 3.4. Extracting Chemical Information From XPS When one wants to extract chemical information, the relevant quantity is the atomic concentration x from equation 3.4. This is normally extracted from the peak area or from intensity through tabulated atomic sensitivity factors S = a(Ek)4>yAT(Ek)F(Ek) (3.5) / = xfS (3.6Each peak intensity is divided by the appropriate atomic sensitivity fac tor, and then all the peaks for each element are added together. Only relative concentrations can be determined in this manner, as the unknown x-ray flux / must be divided out. The disadvantage of using tabulated atomic sensitivity factors is that, at best, they are only valid for a single set of analyzer settings. The transmis sion function (T(Ek)) is very sensitive to the settings used by the analyzer, and can vary between analyzers from the same manufacturer. It also changes as a function of time [25, 26]. A better method would be to deal with each term in equation 3.4 individually. A and <j) are constants at fixed photon energy and are independent of the chemical species being analyzed. If all of the satellites around the main peak are included in the intensity of the main peak, then y goes to 1, One can use calculated a values for each pho toelectron peak [24]. F(Ek) has an approximate functional form (cx y/E^ m the XPS limit [19]). The transmission function, T(Ek), must be measured (see Section 2.2.2). Determining Oxidation Level of Cu with XPS As is discussed in Section 4.1.1, proper oxidation of the grown films is the most difficult part of growing SrCuOi epitaxially. XPS is used in this thesis to measure the oxidation of the Cu (Sr has only one oxidation level and is trivial to oxidize properly relative to the Cu). It is possible to determine the oxidation level of Cu by measuring small shifts in the binding energy 40 Chapter 3. Growth of SrCu02 of various photo-electron peaks, but in the case of Cu, a more accurate technique is possible. The correlated nature of CuO and the effect that it has on its core level spectrum is used to differentiate it from Cu20, which is well described by a single electron, non-interacting approach (this is an instance where the assumption used in equation 3.1 fails, and equation 3.2 must be used). Cu metal has a valence configuration of 3d10As1. At its most accessi- . ble oxidation, 1+, it forms Cu20, and the valence configuration of the Cu becomes 3d10. In terms of photoemission, this means that there is only one peak that will be measured when an electron is photo-excited (ignoring the possibility of multiplet splitting). For the higher oxidation level of Cu, 2+ (forming CuO), there are two possible initial state levels. The obvious possibility is 3d9, but there is also the option of a charge transfer from the ligand Oxygen, creating a 3d10L state (L represents a ligand hole). In this instance, the linear combination of two initial states creates 2 possible peaks in the spectrum (See Figure 3.7), a main line, and a shake-up satellite. This presents several ways of determining the oxidation level of the Cu. One method uses the ratio.of cross sections of ^c^id) in the valence band. While this works well at low photon energies (He II and He I), the cross section ratio for AlKa X-ray radiation is very small. (0.02)(predominantly d character) and hence not suited for XPS measurements of oxidation. For XPS, the best way is to measure the ratio of peak area intensities of the Cu 2pz peak and its shake-up satellite. This ratio of /3at is 0.55, and the 2 •'•main. intensity of both peaks is very strong and easy to measure [12, 28]. 3.3 Results In total, three growths were completed. Several more were attempted, but were unsuccessful due to equipment failures, which in some cases required several weeks to fix. The details of the growths are outlined in table 3.1 RHEED time series for each of these samples all showed similar be haviour. At the start of growth, the films all began to display 3-D style growth, characterized by discrete spots in the RHEED diffraction pattern as opposed to extended rods. The behaviour quickly disappeared and was replaced with a 2-D style growth. For growth #1, the lattice parameter changed suddenly from 3.905 A measured along the (100) direction of the SrTi05 substrate to 1.760 A (see figure 3.8). Growths #2 and #3 did not exhibit this change in lattice parameter, indicating possible epitaxial growth (data for all lattice parameters is needed to say this conclusively). 41 Chapter 3. Growth of SrCuO^ 2.51 1 1 i : ;—i 1 n 1 C~i r 930 935 940 945 950 955 960. • _965 970 975 Binding Energy (eV) Figure 3.7: Comparison of CuO and Cu-^O 2p structure. The two peaks in the CuO 2p^ shake-up lines are due to multiplet splitting of the higher binding energy 3d9 level (from [27]). Growth Substrate Deposition Oxygen .Br Cu Oxidation . # Temperature Rate Pressure Ratio Level (T***-) (°C) (+) \ mm J (Torr) (0.55 = 100%) 1 • 620 10 1.2 x-10-e 0.82 0.05 2 608 5 1.0 x'lO"8 1.14 0.06 3 545. 6 5.9 x 10-Y 0.85 0,05 Table 3.1: Summary of growth parameters for SrCuOi 42 Chapter 3. Growth of SrCuQ2 The largest difference between growth #1 and growths #2 and #3 is the deposition rate. A lower deposition rate appears to be more conducive to growth with a matched lattice. There were no visible RHEED oscillations during any of the growths. This data suggests that the substrate surface had some atomic scale roughness to it, and that the film initially grew in a 3-D fashion to compensate for this. A new substrate preparation technique has been implemented which leaves the SrTiO-s substrates with an atomically fiat surface; this should alleviate this problem. Time Figure 3.8: RHEED time series for a SrCuC>2 growth #1. The vertical axis is the diffraction pattern and the horizontal axis is time. The start of the growth is indicated by the shaded area. The signal before the start of growth is that of SrTiO^, and the signal after the growth is that of an unknown phase. The change in reciprocal lattice spacing corresponds to a lattice parameter change from 3.905 A to 1.760 A. XPS data from all of the samples indicate that none of the films were properly oxidized (predominantly Cu20 with a small amount of CuO). The peak height ratio is so small that increasing the 02 pressure to the maximum possible operating pressure of the oxide MBE chamber will not properly oxidize these films. A different substrate temperature, along with a more reactive species of Oxygen will be necessary to properly oxidize the films. As a result, in Chapter 4, the effectiveness of an atomic Oxygen source is determined. XPS data also indicates that the ratio of ^ was not correct in any of the growths. For equal deposition rates of each element, small changes in other growth parameters caused the stoichiometry to vary wildly. Fortunately, the rate at which material is deposited is much easier to control than any other parameter in the growth, and can be tuned easily after all of the other variables (such as oxidation) are correct. 43 Chapter 3. Growth of SrCu02 Time Figure 3.9: RHEED time series for a SrCu02 growth #3. The vertical axis is the diffraction pattern and the horizontal axis is time. The start of the growth is indicated by the shaded area. The signal before the start of growth is that of SrTiO^, and the signal after growth is that of a material with the same lattice parameter. The RHEED time series for sample #2 was similar. 3.4 Conclusions Despite several attempts, SrCuOz epitaxial films were not grown success fully on SrTiOz substrates. The RHEED patterns taken during growth all show the same initial 3-D growth, and then switch to a 2-D growth pattern. A possible reason for this behaviour is that the SrTiOs substrates have a poor surface quality, and hence the films initially grow in a 3-D manner. This will be dealt with in future growths with a new substrate preparation procedure. Low evaporation rates (~ 5 I -4jj J) produced potentially epi taxial films, while high deposition rates cause a new phase to be formed. The stoichiometry of the films varied between each growth despite equal deposition rates of the Sr and Cu. This is not a major issue due to the ease of altering the rate of deposition after all of the other growth parameters have been established. In all cases, the level of oxidation of the film was in adequate. Chapter 4 deals with the calibration of an atomic Oxygen source, and its applicability for oxidizing SrCuO^ films. A new sample heating mechanism in the Oxide MBE chamber will make it possible to increase the temperature of the film to higher values in future growths. This will also help to increase the oxidation level of the films. What has been learned in these preliminary growths will ensure that future films will be grown with the correct structure and chemistry. All 44 Chapter 3. Growth of SrCuO^ of the difficulties encountered here are surmountable. The work that has been done to improve the utility of RHEED data [22], the quality of the substrate's surface, the range of the sample heating mechanism and the accuracy of XPS data will ensure that subsequent growths of this material will yield samples of high enough quality to perform meaningful ARPES measurements on. 45 Chapter 4 Creation of Atomic Oxygen 4.1 Introduction In the previous chapter, it was found that molecular Oxygen (02) was not reactive enough to properly oxidize the SrCu02 films. To rectify this, a more reactive form of Oxygen is needed. This chapter deals with the 'calibration of a Mantis Deposition MATS30 RF Atom Source for creating Atomic Oxygen (AO), in the hope that it can be used to drive the oxidation of the SrCu02 films to the required levels. 4.1.1 Creation of Atomic Oxygen The conventional bulk growth of SrCu02 requires both high pressure and a high oxidation level to grow successfully. The pressure cells used in the growth provide enough heat and pressure to satisfy both of these require ments. In epitaxial growth, the substrate lattice matching provides the required strain/pressure, but reaching the proper oxidation level is difficult in a low pressure MBE system. Properly oxidizing the films is, in fact, the most difficult part of growing SrCu02 . SrCu02 can be regarded for the purposes of oxidation as a mixture of SrO and.CuO. Formation of SrO is trivial, as. Sr readily reacts with Oxygen, and has only one available oxidation level of 2+. The formation of CuO is much more difficult. The most -accessible oxidation level for Cu is which forms Cu20 when exposed to oxygen. Reaching the 2+ oxidation level for Cu requires high pressures of 02 (which is impossible in a vacuum environment), high temperatures for the Cu, and/or a more reactive species of Oxygen. It is possible to raise the temperature of the sample to increase the oxidation level of the material, but in thin film growth, the vacuum envi ronment and presence of the substrate can complicate matters (for example, the deposited film can- re-evaporate off of the substrate, or the substrate can undergo a change of phase). An element where we have more control is the reactivity of the Oxygen. Ozone (O3) and Nitrous Oxide (N02) have 46 Chapter 4. Creation of Atomic Oxygen both been used previously to increase the reactivity of the Oxygen to the point where CuO films can be grown, but both gases are hazardous and require special handling and pumping procedures [29, 30]. Additionally, the possibility that Nitrogen could be included in the film when Nitrous Oxide is used makes this approach even less desirable. An alternative is to use a source of Atomic Oxygen (AO). Molecular Oxygen is fed into a device that breaks apart O2 and ideally produces a stream of AO, free of any ions. There are two methods of achieving this. The first involves thermal crack ing of O2 over a hot metal surface. This is a common practice for cracking non-reactive gases, but becomes more difficult due to-the reactive nature of the Oxygen (some sources designed specifically for cracking Oxygen do this effectively with a hot Platinum surface). The other method involves running O2 through a Radio Frequency (RF) or microwave cavity. The frequency is tuned to produce a standing wave in the cavity that tears apart the O2 and creates a plasma of AO ions. These AO ions quickly take an extra electron from any available source and are allowed to leak out of the cavity into the vacuum chamber. The resultant beam of AO is directed at the sample. Any free ions that escape can be removed from the beam by deflecting them with an electrostatic field. The method of creating an Oxygen plasma with a RF source is investi gated in this thesis. As the success of growing SrCu02 epitaxially depends on properly oxidizing the films, the cracking efficiency of the source must be determined and compared with devices that have been used previously to grow SrCuOz films. 4.2 Experimental The method used in this thesis to measure the amount of AO produced by a Mantis Deposition MATS 30 RF Atom Source (using a quartz tube with 6 holes escape holes as the plasma cavity) involved measuring the rate of increase of mass of a silver film deposited on a QCM exposed to AO. Molecular O2 does not diffuse into a Ag film, but AO does. The rate of AO atoms impinging on the substrate can be can be measured with a QCM by measuring the rate of thickness/mass increase of the film as a function of time [27, 31]. The rate of AO diffusion into the film increases the measured QCM thickness as an inverse exponential function of time. T(t) = A0-B0e-Xt (4.1) 47 Chapter 4. Creation of Atomic Oxygen Where T(t) is thickness as a function of time, A0 is the final film thickness when the Ag film is saturated with AO, B0 is the difference between the initial and final thickness of the film, and A dictates the rate at which the film increases in thickness. Differentiating 4.2 gives us the rate of thickness increase: dt = B0\e~xt (4.2) The rate of diffusion into the film is a maximum at t=0, giving an initial rate of B0X, with units of tht^gSS• It must be remembered that this mea sured thickness does not actually represent the thickness of the film. The QCM measures thickness by measuring the change in resonant frequency of an oscillating quartz crystal, which changes as more mass is accumulated on the film. What is needed is a conversion from thickness to number of AO molecules. This is given by equation 4.4. AO Rate = -^_(£0A) (4.3) mAO AO Rate = 3.948 x 1015(EoA) (4.4) s cmz PAg is the density of Ag and TTIAO is the mass of AO. (B0\) are measured from fits of the QCM data. • It should be noted that the initial rate of absorbed AO may not be the actual rate of AO impinging on the sample; there is no way to tell if all of the AO is actually absorbed by the Ag film. This information is still useful, however, if the measured rates are compared to the rates of other sources of AO measured in the same way. Additionally, understanding the behaviour of the source under different conditions ensures that it is used to its greatest possible effect. 4.3 Results For each measurement, 200nm of Ag were evaporated onto an Au QCM. The films were exposed to molecular O2 to ensure that the molecular species would not affect the measurement. Closing the shutter on the AO source while in operation successfully stopped the Ag film from absorbing AO while the plasma was lit. QCM thickness measurements were taken as a function of time, sampling every 6 seconds. The thickness/time data was fit to a function of the form equation 4.2 to determine A0, BQ and A. 48 Chapter 4. Creation of Atomic Oxygen The first study looked at the oxygen flux as a function of power at constant pressure/flow rate. The results are presented in figure 4.1. Power (W) Figure 4.1: Flux of AO as a Function of Power at a Constant 02 Pressure of 6.2 x 10"6 Torr (1.1 seem Throughput). These results fit better to an inverse exponential function than to a linear function, implying that the AO rate is saturating as more power is introduced. While there is too little data to fit the function properly, a rough estimate of the maximum possible AO rate from the fit is 2.2 x 1014at(T6 at a flow rate of 1.1 scan. The second part of this study involved measuring the cracking efficiency of the atomic source as a function of increasing Oxygen pressure. The power of the source was kept at 250 W. The range of pressures chosen covered the span between where the plasma was just able to sustain itself (2.7 x 10~6 Torr or 0.5sccm) to close to the maximum pressure allowed by the Oxide MBE chamber interlock (4.1 x 10-5 Torr or 7.4 seem). The data in figure 4.2 suggests that the maximum rate of AO flux is 49 Chapter 4. Creation of Atomic Oxygen •k 2.0 X101' CO ? I 1.5 X1014 «s EC o < ., „ .-,-14 1.0x10' 5.0x10 ,13 J L Throughput (seem) J I L Figure 4.2: Flux of AO as a Function of Throughput at a Constant Power of 250 W 50 Chapter 4. Creation of Atomic Oxygen achieved at 250 W at a flow rate of between 2 and 6 seem. If the trend suggested by the rate versus power data at low Oxygen throughput, shown figure 4.1, holds for higher Oxygen throughputs, then it would be reasonable to expect that rates on the order of 1015 are possible with this source. A future extension of this study would be to investigate the AO flux versus power at Oxygen flow rates in the high cracking efficiency range seen in 4.2, as well as filling gaps in the data presented in figure 4.2. Another possible extension that would increase the utility of the AO source would be to use this information to calibrate a system that could be used to measure AO rates during crystal growth. For example,, a hot Platinum wire placed in the AO stream could be used to measure the rate of recombination of the AO as it strikes the Platinum surface. As the AO recombines on the surface, a certain amount of heat is generated, which changes the resistance of the wire. This change of resistance is proportional to the amount of AO flowing over the surface of the wire. 4.4 Conclusions The Oxygen cracking' efficiency of a Mantis Deposition RF Atom Source has been determined. AO rates in the range of 8 x 1013 to 2.4 x 1014 9^f1 ° era' s cm* s have been demonstrated. AO rates on the order of 1015 atmPs are deemed cm's possible, but the nature of the cracking efficiency of the plasma is poorly understood; further study is required to confirm this. These results are comparable to other sources whose rates were measured in the same way. For example, 100W at 10 seem yielded 3.5 x 1015 ^jpf for a microwave plasma source [27], and similar AO rates at 1.0 Torr of pressure for another microwave source [32]. It is difficult to make a direct comparison, however, due the fact that the Oxygen pressures involved were higher than achievable here in both cases. , Whether this source is appropriate for properly oxidizing SrCuOi films still needs to be determined. Rates of 5 x 1016 atm?s have been found nec-emr s essary to properly oxidize copper in a related material SrCu^Oz [27], but these values were measured using atomic absorption techniques and are not necessarily comparable to QCM based measurements. 51 Chapter 5 Conclusions In this thesis, preliminary growths of SrCuQi epitaxial films were under taken. Despite none of the growths yielding the desired stoichiometry or oxidation, the difficulties that were encountered, such as properly oxidizing the film, are demonstrated to be surmountable. The structure of the films was probed with RHEED, and found to display growth patterns that are consistent with epitaxial growth. The lessons learned from these attempts will make it possible to grow an epitaxial film of SrCuOi for an ARPES study in the future. The strict sample quality requirements imposed by a future ARPES study, and hence the need for the most accurate measurements of stoichiom etry possible, warranted an investigation of the properties of the SPECS PHOIBOS 150 hemispherical electron analyzer used to perform XPS mea surements of films chemistry. The degree of non-linearity of the electron detection system was determined and a method introduced that could cor rect for it. Correcting the data for the transmission function of the analyzer was found to be a necessary step in performing quantitative chemical anal ysis (2 — 3% differences were. observed in stoichiometry before and after applying the transmission function). Therefore, the transmission function of the analyzer was measured using two techniques. The measured trans mission function agreed well with calculated transmission functions provided by SPECS (the difference in SrCuOi film stoichiometry being less than 1% between the 2 transmission functions). While correcting for detector non-linearity, and measuring the actual transmission function did not drastically alter the stoichiometry measurements for scans taken using reasonable ana lyzer settings for XPS (about 1%, mostly from detector non-linearity), it is possible to drive the analyzer into regimes where the corrections would be necessary. The methods used to first correct for the intensity response of the analyzer, and then the transmission function would be more applicable in these instances. This work constitutes the first -steps towards an ARPES study of the electronic properties of SrCuOi films. The possibilities offered by epitaxial growth, such as structural tuning through lattice matching, or electronic 52 Chapter 5. Conclusions reconstruction through substrate/film interactions, will be exploited to al ter the films properties. This will make it possible to not only to study the properties of SrCuOi using a surface sensitive measurement technique (which before now was not possible), but also to investigate novel methods of tuning the properties of the TMOs. This study aims to shed light on the nature of electronic correlations in these materials, and investigate new ways to control their properties. 53 Bibliography [1] KJ Franke, P. Gille, KH Rieder, and W. Theis. Achieving Epitaxy be tween Incommensurate Materials by Quasicrystalline Interlayers. Phys ical .Review Letters, 99(3):36103, 2007. [2] A. Zangwill. Physics at Surfaces. Cambridge University Press, 1988. [3] NJC Ingle, RH Hammond, and MR Beasley. Molecular beam epitaxial growth of SrCuzOz: Metastable structures and the role of epitaxy. Journal Of Applied Physics, 91(10), 2002. [4] A. Vailionis, A. Brazdeikis,. and AS Flodstrom. Observation of local oxygen displacements in Cu02 planes induced by a misfit strain Jn heteroepitaxially grown infinite-layer-structure Ca\-xSrxCu02 films. Physical Review B, 55(10):6152-6155, 1997. [5] MA James and T. Hibma. Thickness-dependent relaxation of NiO (001) overlayers on MgO (001) studied by X-ray diffraction. Surface science, 433:7181722, 1999. [6] WA Harrison, EA Kraut, JR Waldrop, and RW Grant. Polar hetero-junction interfaces. Physical Review B, 18(8):4402-4410, 1978. [7] R. Hesper, LH Tjeng, A. Heeres, and GA Sawatzky. Photo'emission evidence of electronic stabilization of polar surfaces in K$C§Q. Physical Review B, 62(23): 16046-16055, 2000. [8] N. Nakagawa, H.Y. Hwang, and DA. Muller. Why some interfaces • cannot be sharp. Nature Materials, 5(3)-.204-209, 2006. [9] A. Ohtomo, DA Muller, JL Grazul, and HY Hwang. Artificial charge-modulationin atomic-scale perovskite titanate superlattices. Nature, 419(6905):378-380, 2002. [10] N. Reyren, S. Thiel, AD Caviglia, , L. Kourkoutis, G. Hammerl, C. Richter, CW Schneider, T. Kopp, A.S. Ruetschi, D. Jaccard, 54 Bibliography et al. Superconducting Interfaces Between Insulating Oxides. Science, 317(5842):1196, 2007. [11] C. Freysoldt, P. Rinke, and M. Scheffler. Ultrathin oxides: bulk-oxide-like model surfaces or unique films? Physical Review Letters, • 99(8):086101, 2007. [12] Stefan Hiifner. Photoelectron spectroscopy : principles and applications, volume 82 of Springer series in solid-state sciences. Springer-Verlag, Berlin ; New York, 2nd edition, 1996. [13] A. Damascelli, Z. Hussain, and Zhi-Xun Shen. Angle-resolved photoe-mission studies of the cuprate superconductors. Reviews of Modern Physics, 75(2) :473-541, 04 2003. [14] SPECS GmbH. PHOBIOS Hemispherical Energy Analyzer Series, 2.08 edition. [15] N. Mannella, S. Marchesini, AW Kay, A. Nambu, T. Gresch, SH Yang, BS Mun, JM Bussat, A. Rosenhahn, and CS Fadley. Correction of non-linearity effects in detectors for electron spectroscopy. Journal of Electron Spectroscopy and Related Phenomena, 141(l):45-59, 2004. [16] SPECS GmbH. PHOBIOS CCD Imaging Detector, 1.13 edition. [17] L. T. Weng, G. Vereecke, M. J. Genet, P, Bertrand, and W. E. E. Stone. Quantitative xps. part i. experimental determination of the rel ative analyser transmission function of two different spectrometers - a critical assessment of various methods, parameters involved and errors introduced. Surface and Interface Analysis, 20(3): 179-192, 1993. [18] L. T. Weng, G. Vereecke, M. J. Genet, P. G. Rouxhet, J. H. Stone-Masui, P. Bertrand, and W. E. E. Stone. Quantitative xps. part ii. comparison between different quantitative approaches for.two differ ent spectrometers - determination of the contamination-reduced thick ness, application of the determined transmission functions and accuracy achieved. Surface and Interface Analysis,. 20(3): 193-205, 1993. [19] P. RufHeux, P. Schwaller, O. Groning, L. Schlapbach, P. Groning, Q. C. Herd, D. Funnemann, and J. Westermann. Experimental.determina tion of the transmission factor for the omicron eal25 electron analyzer. Review of Scientific Instruments, 71(10):3634, 2000/10/. 55 Bibliography [20] Y. M. Cross and J. E. Castle. The relationship between transmission efficiencies in the frr and fat modes of an electron spectrometer. Journal of Electron Spectroscopy and Related Phenomena, 22(l):53-60, 02 1981. [21] Carol S: Hemminger, Terry A. Land, Alex Christie, and John C. Hemminger. Empirical electron spectrometer transmission function for applications in quantitative xps. Surface and Interface Analysis, 15(5):323-327, 1990. [22] Sebastian Leung. Visualization of Reciprocal Lattice Structures Using RHEED Data, 2007. [23] K. Ploog. Microscopical Structuring of Solids by Molecular Beam Epi-taxySpatially Resolved Materials Synthesis. Angewandte Chemie In ternational Edition in English, 27(5):593-621, 1988. [24] C. D. Wagner, L. E. Davis, M. V. Zeller, J. A. Taylor, R. H. Raymond, and L. H. Gale. Empirical atomic sensitivity factors for quantitative analysis by electron spectroscopy for chemical analysis. Surface and Interface Analysis, 3(5):211-25, 10 1981. [25] M. P. Seah and G. C. Smith. Aes. accurate intensity calibration of electron spectrometers, results of a bcr interlaboratory comparison co-sponsored by the vamas sea twp. Surface and Interface Analysis, 17(12):855-874, 1991. [26] M. P. Seah. Xps reference procedure for the accurate intensity cal ibration of electron spectrometers - results of a bcr intercomparison co-sponsored by the vamas sea twa. Surface and Interface Analysis, 20(3):243-266, 1993. [27] Nicholas J:C. Ingle. Molecular Beam Epitaxial Growth of Complex Novel Oxides. PhD.thesis, Stanford Universtiy, 2001. [28] J. Ghijsen, LH Tjeng, J. van Elp, H. Eskes, J. Westerink, GA Sawatzky, and MT Czyzyk. Electronic structure of Cu^O and CuO. Physical Review B, 38(16):11322-11330, 1988. [29] A. Brazdeikis and A. S. Flodstrom. Surface morphology and nucleation of infinite-layer (Sr,Ca)Cu02 films on atomically flat SrTiO^ (100) substrates studied by atomic force microscopy. Applied Surface Science, 107:95-101, 1996. 56 Bibliography [30] S. Gonda, H. Nagata, M. Kawasaki, M. Yoshimoto, and H. Koinuma. In situ surface analysis of SrCu02 heteroepitaxy on SrTiO^ substrate using X-ray photoelectron spectroscopy. Physica C: Superconductivity, 216(1):160-164, 1993. [31] LAA Pettersson and PG Snyder. Preparation and characterization of oxidized silver thin films. Thin Solid Films, 270(1-2):69-72, 1995. [32] V. Matijasevic. Atomic oxygen detection by a silver-coated quartz de position monitor. Review of Scientific Instruments, 61(6):1747, 1990. 57 Appendix A Intensity Calibration Programs All of the following scripts were written for Interpretive Data Language (IDL) version 6.3. A.l Calibration Script This script finds the column vector A, as per the algorithm described in section 2.2.1. It presupposes the existence of arrays of XPS intensity data named for the X-ray power used in Watts (scan400, scan350, scan300, scan250, scan200, scanl50, scanlOO). p = 6 B= [ -scan400/400.0+scanl00/100.0,-scan350/350.0 + scanlOO/100.0 , -scan300/300.0+ scan 100 / 100.0 , -scan250/250.0+scanl00/100.0 , -scan200/200.0 . +scanl00/100.0 , -scanl50/150.0+ scanlOO /100.0] C = DBLarr ( size (B, /dimensions), p) FOR i = 2, p+lDOC[*, i-2] = [ double (scan400 )*( i )/400.0 -double(scanl00)~(i)/100.0, double(scan350)~(i)/350.0 double ( scanlOO ) ~ (i) /100.0 , double (scan300 ) ( i )/300.0 -double ( scanlOO )" (i)/ 100.0 , double (scan250 )"( i )/250.0-double(scaril00)"(i)/100.0 , double (scan200 )'(i )/200.0 double (scanlOO)" (i)/100.0 , double (scanl50 )" ( i)/150.0 -double(scanlOO)" (i )/100.0] A= Invert ( Matrix_Multiply (double (C) ,double( C) , / Atranspose )) # Matrix_Multiply (double (C) , double(B), /ATranspose) 58 Appendix A. Intensity Calibration Programs Print , A . . A.2 Data Correction This function accepts the data to be corrected and the column vector A as its inputs, and returns the corrected spectrum. function makegood, scan, A fixed = dblarr ( size (scan , /dimensions)) temp = size (A, /dimensions) fixed = scan for i = 2, temp[0] + l DO fixed += double( A [ i-2] * scan " ( return , fixed END 59 


Citation Scheme:


Usage Statistics

Country Views Downloads
China 7 26
United States 4 0
France 2 0
City Views Downloads
Beijing 7 0
Unknown 3 0
Ashburn 2 0
Dayton 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items