超高真空系統架設與鐵在低溫下成長於金與矽基板之研究

全文

(1)國立台灣師範大學物理研究所碩士論文. 超高真空系統架設與鐵在低溫下成長於金與矽基板之研究 Multifunctional UHV System Setup and Low-temperature Growth of Magnetic Nanostructures: Fe/Au(111), Fe/Si(111). 指導教授：林文欽博士研究生：林彥穎撰中華民國九十八年八月.

(2)

(3) Dedicated to my parents.

(4) Acknowledgements. I still can not believe that I can finish my thesis in time, It is very hurry for me to complete the jobs in such a short time, actually, without some people, I would not achieve, that I very appreciate. First I have to thank my parents for their supporting me to continue my school work, they are really patient with my late study, I know it is not easy for every parents to wait for their son to finish a primary study in school with such a long time. I will keep their hope in mind and doing my best to let them proud for me again. Thank prof. Cheng Hsiu-Fung and prof. Lin I-Nan, my early advisor in the first graduate year, they led me into the life of study and reaserch, I will never forget the time there, especially Horng-Yi and Chuan-Chic, my fist mates in NTNU, we shared many interestings. Thank Meng-Hsun, he brought me into prof. Lin Wen-Chin’s laberatory, here, and thank prof. Lin, he is very kind and taught me a lot about how to be a professional reasercher, although I still have to improve myself in many ways, he gave a gaol for me to achieve. In lab. C207, Chuang-Han really did great helps to me and he is also a exaplary person to me; Wen-Tin and Hung-Hsiang helped me to deal with the business in laboratory. Jie-Jhen in NSRRC provided a lot of experience and equipment information to me; mates in IAMS, C.J and Yi-Chen are also good partners in school and laboratory works. Finally, I’d like to give thanks to my lovely lady, Cynthia, she encourages me to be strong to overcome everything I suffered, I appreciate her being with me for the times..

(5) Multifunctional UHV System Setup and Low-temperature Growth of Magnetic Nanostructures: Fe/Au(111), Fe/Si(111) Yen-Ying Lin. Abstract The UHV system in C207 in the Department of physics, NTNU was established with the best base pressure lower than 3.0 × 10−10 torr. The image of STM can be corrected by Si(111) 7 × 7 and HOPG. The morphology of the growth of Fe on Au(111) herringbone surface at low temperature (LT) has been investigated. When depositing Fe on Au(111) herringbone surface at 180 K, doublets are uniformly located on fcc sites, that is different from the result of room temperature (RT) growth and exposure xenon at 90 K. After a repeat of 180 K deposition and RT annealing with total thickness of 0.45 ML, Fe clusters become random, and the 300 K annealing effect is not obvious. On the Au(111) surface, there is a region presenting a vicinal-like surface stucture, on which the Fe growth is totally different from the published results on Au(788) vicinal surface, the location of some small Fe clusters. Finally, the MOKE was set up with max field' 4000 Gauss. And we successfully obsreved the hyteresis loop of Fe/Si(111) in in-plane magnetization ..

(6) Table of Contents Abstract 1 Introduction 1.1 Silicon(111)7x7 ………………………………………………………….…...… 1.2 Highly ordered pyrolytic graphite (HOPG) ……………………….…………. 1.3 Reconstruction of Au(111) …………………………….……………………….. 1 1 5 7. 2 Basic Concepts 2.1 Growth of thin _lm and islands ……………………………………………… 2.2 Magnetic hysteresis loop …………………………………………………….. 2.3 Quantum well effect ………………………………………………………….. 11 11 13 15. 3 Experimental Apparatus 3.1 Multi-functional UHV systems ……………………………………………… 3.2 Home-made Evaporation Gun ………………………………………………. 3.3 LEED and I/V-LEED ……………………………………………………….. 3.4 Scanning Tunneling Microscopy (STM) ……………………………………. 3.5 Magneto-Optical Kerr Effect (MOKE) …………………………………….... 17 17 20 21 24 28. 4 Experiment and results 4.1 Si (111)7x7 observation …………………………………………………….. 4.2 Highly ordered pyrolytic graphite (HOPG) ………………………………… 4.3 Two step growth of Fe/Au(111) ……………………………………………. 4.3.1 Fe nanostructures on Au(111) herringbone surface ………………….. 4.3.2 Fe/Au vicinal growth ………………………………………...….……. 4.4 LT growth of Fe/Si(111 )7x7 and MOKE ………………………………….. 4.5 MOKE calibration …………………………………………………………... 33 33 37 38 38 40 41 46. 5 Discussion and conclusion 5.1 Si(111)7x7, HOPG calibration ……………………………………………… 5.2 Two step growth of Fe/Au(111) …………………………………………….. 5.2.1 Fe nanostructures on Au(111) herringbone surface …………………... 5.2.2 Fe/Au vicinal-like growth …………………………………………….. 5.3 LT growth of Fe/Si(111 )7x7 and MOKE ……………………………………. 49 49 50 50 53 55. 6 Summary. 57. Bibliography. 58. 2.

(7) Chapter 1 Introduction. Nano science arises in the recent year, being a field in the new generation, it is posible for human to manipulate atoms; further more, to apply quantum effects into practice. It was not untill the improvement in ultra high vacuum (UHV) and the invention of scanning probe microscope (SPM) that nano science comes so closely to our life. In such a tiny scale, a single molecule can even be a pollutant that we cannot easily clarify which one on substrate leads to the appearance. An UHV chamber provides a clean envoronment to give less unexpected factor, a SPM machine helps the investigation in a small area, the later becomes a standard equipment for surface science, a branch of condensed physics. With the development of spintronics, ferromagnetic semiconductors have been considered as the material for next micro-electronic devices owning to ferromagnetic semiconductor’s higher curie temperature, the control of spinor’s translation become possible for application in human life.. 1.1. Silicon(111)7 × 7. Semiconductors, especially silicon, being the basic material for electronic devices, really change our life. In the recent years, crystalline samples can be manufactured in ways-thanks to silicon-based computers, as a matter of fact-that their surface has 1.

(8) 1.1. Silicon(111)7 × 7. 2. a high degree of perfection [1]. Moreover, surface observation with atomic resolution in tunneling microscopy is ever easier with semiconductors than with most metals. The crystal structure of silicon is diamond cubic, in this structure,due to the hybridization of s and p electrons, stems from the fact that each atom ’likes’ prefers to have four neibours arranged in regular tetrahedron. Cleaving a highly covalent body in two parts to create a surface, one breaks bonds which are now ’dangling’. They recombine to lower their energy resulting in a surface reconstruction, a typical phenomenon for all semiconductors. The (001) and (111) surface orientations are especially used for growth of microelectronic components. In(001) surface, the diamond lattice has two atoms per unit cell, thus it is not a Bravais lattice1 In practice, (001) surfaces are most often used in MBE growth, and some steps are always present. The (111) face is characterized by a very complicated reconstruction. To understand its nature, consider first a 1 × 1 non-reconstruction strucutre. In a 1 × 1 struture, each surface atom has a dangling bond. Compared to (001), it is not much, but as it happens, one can easily reduce this number by placing one adatom on the head of each group of three atoms. This adatom possess one unsatisfied bond, but the number of dangling bonds per unit surface is reduced bt a factor three. However, such adatoms deeply perturb the other atoms. On silicon, this perturbationfavours the appearence of dimer bonds, stacking fualts, and even surface vacancies. The outcome is the celebrated DAS (dimer-adatom-stacking-fualt)7 × 7 model, which is sketched in figur 1.1. It includes 12 adatoms, 9 dimer bonds and a stacking fault in each unit cell. The order of magnitude for surface energues is the electron volt. But energy difference may be much smaller. For instance, take the cleavage (111)plane of silicon. This surface reconstructs after cleavage in metastable 2 × 1 structure, which transforms into the 7 × 7 upon annealing. AB − initio calculations showed that the energy difference between the two strucures is 60 mV per 1 × 1 cell [2, 3]. The charactgeristics of growing Fe on Si surface have been widely studied [4]. The 1. The crystal looks the same when viewed from any of the lattice points.. August 14, 2009.

(9) 1.1. Silicon(111)7 × 7. 3. Figure 1.1: Reconstructed 7 × 7 structure of the silicon (111) face.. electric properties [5, 6], magnetic properties [7–9] and gas reactions [13, 14] of various Fe silicides on Si also have been investigated. However, it is difficult to control the crystallinity and morphology of thin epitaxial phases depending strongly on the growth conditions, which seriously affect device properties. Moreover, the two step growth of Fe on Si(111)7×7 at low temperature, and its magnetic properties has not been well studied yet, since the various iron silicide are still well interested. A similar. Figure 1.2: Schematic of various silicide phase diagram in different coverage and annealing temperature. August 14, 2009.

(10) 1.1. Silicon(111)7 × 7. 4. system of ferromagnetic metal on Si(111)7 × 7 is Co/Si(111)7 × 7, which has investigated from room temperature to low temperature [10–12]. In room temperature growth of Co/Si(111)7 × 7 system, no specially preferable site has been identified for the bonding of Co on the rhombic unit cells of Si(111)7 ×7 substrate [11]. And there is still alloy between coblat and silicon atom [10] under lower temperature, that the appearance of doubled-spot defects shows that the chemical bonds are of the type in which one cobalt atom interacts with two substrate atoms. For Co deposited on a CoSi interface, the films possess an in-plane easy axis of magnetization. The different orientations of easy axis are attributed to the different magnetocrystalline anisotropies of the Co/CoSi and Co/Si interfaces.. Figure 1.3: (a) STM image (50 nmx50 nm) of 0.13 ML Co/Si(111). Doubled spot defects and some schematic insets are shown in the right panel.(b) Magnetic hysteresis loops for 2, 6, and 10 ML Co/Si(111) [10]. From figure 1.2 we see it is easily for iron grown on silicon surface to form iron silicides, we try to use MBE method to grow iron on silicon surface under lower August 14, 2009.

(11) 1.2. Highly ordered pyrolytic graphite (HOPG). 5. Figure 1.4: Magnetic hysteresis loops for 6 and 10 ML Co/Si(111) at different temperatures. Hysteresis behavior occurs in both the longitudinal and polar configurations [10]. temperature to reduce the formation of ”alloy”, therefore we may possibliy to study the reaction of real iron epitaxial growth on silicon surface.. 1.2. Highly ordered pyrolytic graphite (HOPG). HOPG is a semimetal. The basal or (0001) plane is reflective and electrically conductive. In the direction perpendicular to this plane, the electrical conductivity is much lower. The basal planes are held together weakly making it easy to expose clean atomically smooth sections of material by mechancially peeling off a group of planes closer to the surface of the material [15]. A single layer HOPG is called graphene, a pseudo-semiconductor, which has many special characteristics, and is well studied in its high mobility, high concentrations of n and p doping, and other features. In the obsevation of the structure of HOPG, the fact that in most STM images of graphite one observes with positive contrast only three of the six carbon atoms [16– 19]. This is because that there are two non-equivalent types of corbon atom sites: A-types (◦) have neighbors directly below in the second layer, and B-types are located above an hexagon (•) . The dashed lines show the bulk unit cell (also see August 14, 2009.

(12) 1.2. Highly ordered pyrolytic graphite (HOPG). 6. Figure 1.5: Schematic representation of the structure of the bulk hexagonal graphite crystal (ABAB stacking) [16].. Fig. 1.3). The in-layer nearest carboncarbon distance is 1.42 ˚ A and carbon layers are separated by 3.35˚ A . Figure1.3 illustrates the number of carbon atoms in the graphite structure which can be depressed by a tip consisting of an atom on its top part ( (•)) at different positions. When the tip is located above A- or B- sites, three additional carbon atoms can be depressed by, for instance, a metal tip atom (e.g. W, Pt, or Ir). Above a hollow-site all of the six carbon atoms of a hexagon are ”seen” by the tip, due to its comparable size (e.g. Pt, W: 1.39 ˚ A ) with the radius of the hexagon (1.42 ˚ A ). The size of the solid circle (2.79 ˚ A diameter, corresponds to the diameter of a Pt atom) has been drawn to the same scale as for the hexagons. In some practice of STM scanning,every other atom is enhanced.”Zig-zag” and ”arm-chair-configuration” of the edges of a graphite (0001) face are also seen, figure 1.4 shows the chematic representation. Generally,due to the variation in the local hardness between A-, B- and hole-sites (a hole is harder than an A- and an A-site is harder than a B-site). The STM thus ” sees ” the hitherto neglected hollow sites in the graphite structure and not the atom positions. That in order to visualize the true position of the carbon atoms in a hexagon, to invert the measured STM image was suggested [16]. August 14, 2009.

(13) 1.3. Reconstruction of Au(111). 7. Figure 1.6: The graphite structure which can be depressed by a tip consisting of an atom on its top part (•) at different positions. [16].. Figure 1.7: The enhanced ”Zig-zag” and ”arm-chair-configuration” [16].. 1.3. Reconstruction of Au(111). Nucleation control is one of the crucial issues for nanostructure synthesis since it August 14, 2009.

(14) 1.3. Reconstruction of Au(111). 8. is the dominant factor for how the atoms/molecules form self-assembled nanostructures on the surface [20]. Naturally formed nanostuctured-templates, such as square islands of N/Cu(100) [21] and dislocation network of Ag/Pt(111) [22,23], are extensively studied to provide the well-organized nucleation patterns. These techniques and knowledge are promising due to their capabilities to assemble the deposited materials into various nanostuctures which are of the importance for applications, especially in catalysis and nanomagetism [24–27]. However, each nanostructured template is mostly limited to only one nucleation pattern, which restricts the further application to more complicated catalytic process or to tune the magnetic coupling. So far the study for creating the variety of nucleation patterns on the same template is crucial, but still lacking. Au(111) is the lowest energy surface of gold, as reflected in the tendency of thin-film growth to propagate in the [111] direction [28, 29]. In recent technological applications, AuCnX are more likely to rely on thin film Au, rather than single-crystal Au.. Figure 1.8: (A)with the process as depositing decanethiol monolayer at 300 K and annealed to 325 K for 5 min in vacuum, the resulting surface exhibits c(4 × 2) domains, domain boundary network, and vacancy islands. (B) Monolayer as in A, annealed to 375 K for 10 min invacuum, the Surface exhibits large c(4 × 2) domains √ and large domains of the lower density (p× 3) phase. (C) Monolayer as in B, anneal Au to 575 K for 10 min in vacuum, the urface exhibits herringbone reconstruction characteristic of clean Au-(111) [28, 41].. Au(111) herringbone reconstruction surface is a high-lighted template for fabricaAugust 14, 2009.

(15) 1.3. Reconstruction of Au(111). 9. tion of Fe, Co, Ni, and Mn well-ordered nanoisland arrays [30–39]. The herringbone surface is constituted of zigzag domains of hexagonal close-packed (hcp) and face center cubic (fcc) crystalline structures. Both of the fcc and hcp kinks, are favored nucleation sites for the 3dtransition metals. Up to now, no clear differences between fcc and hcp kinks are observed. The Au(111) herringbone template is always limited in only using the regularly spaced kinks as preferred nucleation sites. Actually, by tuning the mobility or the diffusion behaviors of deposit atoms, one can very possibly control the nucleation patterns for the fabrication of various nanostructures on the same template [40]. The other interesting case is the Fe on Au(788) vicinal surface, with different annealing temperature, one can obtain morphologies from nanodots to step flow [47]. Figure 1.9: (a),(b)and(c) are smaller islands under 250K;note that in (a)44K and (b)58 K the islands locate randomly, (c)at 70 K one can se that there doublets appear.From(d)200 K and (e)300 K,doublets start merging at someposition and step flow growth begins. [47]. The STM image with a shematic picture of structure in figure 1.10 shows a stepflow gwoth; the linprofile in the left shows the size of terrace and three types of the Fe growth on Au(788) vicinal surface. The nucleation on Au(111) herringbone kinks are usually observed on larger latAugust 14, 2009.

(16) 1.3. Reconstruction of Au(111). 10. Figure 1.10: The structure of Au(788)vincinal surface and step growth begin at fcc site when annealing temperature increasing near 300 K [47]. tice misfit systems, such as Fe, Co, Ni on Au(111) but not intensitive on small lattice misfit systems, such as Al, Ag, Au on Au(111)systems. [42–45]. The topic for us to study of iron on gold(111) and its vicinal surface is that would the nucleation lead to special, not only epitaxial growth but also give some phenomenon in magnetism?. August 14, 2009.

(17) Chapter 2 Basic Concepts 2.1. Growth of thin film and islands. The forces rearrange the growing surface, a dominant role is played by the free energy γ on the surface and interface, which determing the growth modes in thermal equilibrium.. Figure 2.1: Surface energies γ for magnetic and non-magnetic materials [46]. From Figure 2.1 we can see that magnetic material exhibit a relatively high surface energy. In equilibrium growth, there are usually three growth modes: layer by layer growth, island growth and Stranski-Krastanov growth. The morphology depends on the balance between free surface of substrate, over-. 11.

(18) 2.1. Growth of thin film and islands. 12. Figure 2.2: Schematic display of three growth modes. layer and interface. The relations are : γsubstrate < γoverlayer + γinterf ace. (2.1). γsubstrate > γoverlayer + γinterf ace. (2.2). In case (1), the first atomic layer wants to coat the whole surface to provide optimum energy reduction, which is called layer-by-layer growth; in case (2), the over-layer has a tendancy to nucleate three-dimensional island, and leaves the lowenergy substrate exposed, this is island growth. The situation usually occurs when growing and magnetic materials on top of an inert substrate, such as noble metal or an oxide. By the ”rules” (1) and (2), if material A grows in material B layer by layer, then B on A will grow in islands, since the surface energies are reversed, this is Murphy’s law of epitaxy growth when A and B are with substrate surface energy difference. To overcome this, one would try non-equilibrium growth. The thermal-dynamics laws are very restrictive for growing desired nano-structure, especially in equilibrium growth. Therefore non-equilibrium growth, such as at lowtemperature or high deposition rate are applied. Figure 2b shows the general result in non-equilibrium growth. At low-temperature, or higher rate or lower step densities, the arriving atoms do nat have enough energy (or diffusion time) to find the nearest step edge and to nucleate into islands. Additionally, if the microscopic kinetic processes taking place at the surface is taken into account, the substrate August 14, 2009.

(19) 2.2. Magnetic hysteresis loop. 13. topology can be found to have much influence on the growth mode. The possible kinetic processes taking place at the surface is shown in. Figure 2.3: Schematic display of the ideal growth mode. (a) Equilibrium growth and (b) non-quilibrium growth Many factors might be responsible for the layer-plus-island growth, such as the lattice mismatch between the substrate and the deposited film and alternatively, the symmetry or orientation of the overlayers with respect to the substrate. Additionally, if the microscopic kinetic processes taking place at the surface is taken into account, the substrate topology can be found to have much influence on the growth mode. The possible kinetic processes taking place at the surface is shown in Fig. 2.4.. Figure 2.4: The possible kinetic processes taking place at the surface.. 2.2. Magnetic hysteresis loop August 14, 2009.

(20) 2.2. Magnetic hysteresis loop. 14. [!ht]. Figure 2.5: An example of hysteresis loop. The magnetization at zero external field is the remenance. The external field at zero net magnetization is called the coercive field and coercivity.. Hysteresis loop is one of the most distinctive experimental facts of ferromagnetism. An example is shown in Fig. 2.5. Loops like this are obtained by applying to the sample a cyclic magnetic field H and by recording the ensuing change of the magnetization M along the field, where M is defined as the average magnetic moment per unit volume. Hysteresis loops may be of many different shapes, thus it’s important to get some parameters to characterize the loop properties. Two quantities of particular importance are the remanent magnetization or remanence, Mr, and the coercive field or coercivity, Hc . As indicated in Fig. 2.5, remanence represents the magnetization obtained by applying a magnetic field and then removing it, and coercivity is the field needed to bring the remanence to zero. Unlike remanence, the coercive field spans an very wide interval, from less than 1 A/m to more than 10 6 A/m, in different specimens. The shape of a hysteresis loop strongly depends on not only the intrinsic properties of the specimen but also the external factors such as the measurement method, the specimen geometry etc.. Therefore, we have to make sure that the loops are measured in the same process, specimen geometry etc before comparing their results. The variety of the hysteresis loop shape is the direct consequence of the variety of possible magnetic domain structure. The primary mechanisms of the hysteresis loop are magnetization rotation and wall motion. In general, wall motion can occur in low field, and magnetization rotation needs strong applied field to overcome the energy barrier. August 14, 2009.

(21) 2.3. Quantun well effect. 15. Figure 2.6: Examples of hysteresis loops measured in (a) easy and (b) hard axis.. 2.3. Quantun well effect. Quantum well effect was clearly discussed in systems of thin film [48–55]. The ”well” confines electron to layer with lower inner potential and quantizes the moentumand energy perpendicular to the layers. The coresponding underyling is spin dependent owing to the spin dependence of the inner potential in ferromagnets, Namely, these quantized states become spin polarizes. As the system dimensions reduce, there is an increased locationlizationand overlap of electronic wave functions, and consequently, an increase in electron correlation. Thus, low-diemesional systems offer a convenient platform for experimenting with many body effects. To exame the descret states from confined electrons in space by a potential well, an elementary example can be found in general quantun mechenics textbooks. Here comes a simple model. Assume there is an electron confined in an 1-D box, the allowed wave vector k for stationary states , or quantum well states, are determined by the fitting in geometry, that standing waves exsist with k=. nπ d. where n is an interger quantum number and d is the width of box or the film. August 14, 2009.

(22) 2.3. Quantun well effect. 16. thickness. The energy levels are given by E=. ~2 nπ 2 ~2 k 2 = ( ) 2m 2m d. where m is the mass of a free electron, and the wave functions of the electron in different states are ψ(z) ∝ sin(. nπz ) d. Although the basic phisics is similar, energy levels in atoms or molecules are generally not reffer to as quantum well state. When the quantum well effects electronic and magnetic properties were discussed, The most important dorminant is the quantum well state at the Fermi-level. The density of states at the Fermi-level triggers electronic phase transitions, such as superconductivity,charge density waves, ferromagnetism, antiferomagnetism and electrical and thermal transport.. August 14, 2009.

(23) Chapter 3 Experimental Apparatus 3.1. Multi-functional UHV systems. The thickness of the deposited layer we made is about only few monolayer, that any nano-scale unexpected deposition, or, pollution will contribute unknown effect to our measurement, or kill the weak micro-phenomenon we want to observe. For example, Lieberman discovered the pollution will result in magnetic dead layer in 1969 [56,57], and Hope found that only 0.08 mono layer CO falls on Co/Cu(110) will also change the easy axis [58]. For the purpose to avoid the sample surface being polluted, an ultra high vacuum (< 10−10 torr) is basically recommend. The other way, we can also sketch the pollution rate with respect to the base pressure by the dynamics of gas. We know the frequency of the gas atoms collide the sample surface is r=. nva 4. where the root mean square velocity of a gas molecular of mass m under Kelvin degree T is 2 vrms =. 3kB T m. 17.

(24) 3.1. Multi-functional UHV systems. 18. where and the mean velocity of gas is r va = vrms. r =. 8 3π. 8kB T mπ. Combine these into idea gas formulation P = nkB T. We can get r=√. P 2πkB mT. Substituting, making this formula more readily useful by expressing P in torr and merging all convetion factors into a constant. The result is P (torr) r = 3.52 × 1022 p /cm2 s m(a.m.u)T (K). Using nitrogen of mass 28, room temperature is 300K and the nitrogen pressure is 1×10−6 torr to demonstrate the frequency of gas collide on the sample surface, we obtain r = 3.84 × 1014 bombardments /cm2 s. Assume that 100% nitrogen molecules will stick on the sample surface, after 0.1 seconds, well have about 2 ML 1 nitrogen atoms on or sample surface of 0.5 cm2 . To reach UHV condition, the pumping system and process play important roles. Usually we use a mechanical pump to fore-pumping down to the pressure about 10−2 to 10−3 torr, then a turbo pump is capably on to help further pumping to 10 −7 torr while the main chamber is also baked by about 120 o C for 24 hours and reach about 10−8 to 10−10 torr after baking stopped and cool down to room temperature. 1. 2 × 1015 atoms /cm2 for 1 ML. August 14, 2009.

(25) 3.1. Multi-functional UHV systems. 19. When the base pressure is near 10−7 torr, we on and off the ion pump to out gas, the ion pump will continuously on until it is clean enough. During the process with the base pressure is under 10−6 torr, TSP (titanium sublimation pump) is applied to help pumping and keep UHV getting better.. Figure 3.1: Illustation of the UHV system in lab. C207 in NTNU.. Figure 3.2: The translation of sample in the UHV system in C207 in NTNU.. August 14, 2009.

(26) 3.2. Home-made Evaporation Gun. 20. Figure 3.3: The real set up of the UHV system in lab. C207 in NTNU.. 3.2. Home-made Evaporation Gun. With the same concept of EFM, a tungsten filament is heated by DC current about 1.5-2 A, an additional high voltage(1 KV) difference between filament and source is also applied, therefore hot electrons escape from tungsten filament to the source in center, the hot electron current can be read from the current meter connected to the source, with the quantity is about 10 mA when the source evaporates in our expeiriment. Since all the parameters are kept fixed, the source evaporant is stable. The working power is usually about 10 W (10 mA × 1000 V). Figure 3.5 shows the structure of our gun, the gun body is made of thin Ta plate. A shutter is additionaly applied close to e-gun to ”stop” the evaporance while we wanting to. Thus the hole deposition system takes two ports on our chamber.. August 14, 2009.

(27) 3.3. LEED and I/V-LEED. 21. Figure 3.4: Another view of the multi-functional UHV system. Figure 3.5: a)Home made e-gun with Ta shell, and b) a view without shell. 3.3. LEED and I/V-LEED. Low Energy Electron Diffraction (LEED) is a technique to determine the surface structure of crystalline materials, which applied two fundamental physics, the de August 14, 2009.

(28) 3.3. LEED and I/V-LEED. 22. Broglie matter-wave and Bragg diffraction. For the development of LEED, Louis de Broglie introduced matter-wave in 1924, Clinton Davis and Lester Germer discover the electron in 1927, but LEED was popular used until the improvement of ultra high vacuum technique and detect method in the late 1960s. There are two applications for a LEED system, to determine surface crystalline structure and to accurate the atomic position.. Figure 3.6: Schematic display of a LEED structure The apparatus is also the same as LEED with the only difference that the sample is rotated by θ = 5o , thus the LEED (00) beams can be observed clearly. The I/V-LEED curve, can be obtained by continuously changing the incident electron beam energy and by recording the ensuing change of the (00) beam intensity. From theBragg condition and the de Brogile relation, we have. 2dcosθ = nλ = n. h h = np p 2m(Ek − V ). (3.1). with d the vertical interlayer distance, Ek the kinetic energy of the incident electron beam, and V the potential cost for electrons to escape from the atoms. Therefore Eq. 3.2 is got as August 14, 2009.

(29) 3.3. LEED and I/V-LEED. E k = n2 ·. 23. h2 +V 8m·d2 cos2 θ. (3.2). August 14, 2009.

(30) 3.4. Scanning Tunneling Microscopy (STM). 3.4. 24. Scanning Tunneling Microscopy (STM). Scanning tunneling microscope (STM) is a helpful instument to view surface of conductors in atomic level, the idea to build STM is based on the concept of quantum tunneling. in 1981,Gerd Binning and Heinrich Hohrer developed STM then they earned the Nobel Prize in Physics in 1986. It is very flexible to use STM, not only in ultra high vacuum, but also in air and in liquid or gas enviroments, and at temperatures ranging from near zero kelvin to about few hundred kelvin. In the real operation, STM probes the density of states with lateral resolution of 0.1nm (1˚ A) and depth resolution of 0.01 nm (0.1˚ A). For the sample is applied low voltages, the tunneling current will be the function of the local density of states (LDOS) at the Fermi level that we also interest in. Since STM is such a micro-probing machine, a clean surface,a sharp tip and a stable ambient are highly required. The basic working theory of STM is as fallow: We start form the Schrödinger equation, the fundamental equation in quantum physics. Inderivation, E > V is true for wave function inside the tip or inside the sample. p 2m(E − V ) k= (3.1) ~ E<V is for inside a barrier as between tip and sample, the wave function is decaying. p 2m(V − E) k= (3.2) ~ In scanning tunneling microscopy, a small bias voltage V is applied so that due to the electric field the tunneling of electrons results in a tunneling current I. The height of the barrier can roughly be approximated by the average work function of sample and 1 Φ = (sample +tip ) 2. (3.3). tip. If the voltage (eV) is much smaller than the work function (), Eq. 00 can be simplified to. √ k'. 2m ~. (3.4) August 14, 2009.

(31) 3.4. Scanning Tunneling Microscopy (STM). 25. The tunneling current is from the electrons near the Fermi level, and is proportional to the probability of the decaying wave function. I∝. EF X. | Ψn (0)2 |e−2kd. (3.5). En =EF −eV. Ψ(d) = Ψ(0)e−k·d. (3.6). The probability of finding an electron behind the barrier of the width d is W (d) =| Ψ(d)2 |=| Ψ(0)2 |e−2kd. (3.7). Figure 3.7: Schematic diagram of electron tunneling. It should be also mentioned that the number of empty states will affect the magnitude of tunneling current. See figure 3.8.. Figure 3.8: The number of empty states effests the tunneling current. There are three ways that STM images:. August 14, 2009.

(32) 3.4. Scanning Tunneling Microscopy (STM). 26. Figure 3.9: The loop of a working STM Constant current mode: By using a feedback loop, the tip is vertically adjusted in such a way that the current always stays constant. As the current isproportional to the local density of states, the tip follows acontour of a constant density of states during scanning. Recording the vertical position of the tip generates a kind of a topographicimage of the surface.. Figure 3.10: STM constant current mode.. Constant height mode: August 14, 2009.

(33) 3.4. Scanning Tunneling Microscopy (STM). 27. In this mode the vertical position of the tip is not changed, equivalent to a slow or disabled feedback. The current as a function of lateral position represents the surface image. This mode is only appropriate for atomically flat surfaces as otherwise a tip crash would be inevitable. One of its advantages is that it can be used at high scanning frequencies (up to 10 KHz). In comparison, the scanning frequency in the constant current mode is about 1 image per second or even per several minutes.. Figure 3.11: STM constant height mode.. Current imaging tunneling spectroscopy, CITS Equipment structure in NTNU The tip is made from Pt and the shock proof is consists of two level of springs.. Figure 3.12: Schematic display of STM stage in NTNU. August 14, 2009.

(34) 3.5. Magneto-Optical Kerr Effect (MOKE). 3.5. 28. Magneto-Optical Kerr Effect (MOKE). Magneto-optics effect is a phenomenon similar to optic activity. When a linear polarized electro-magnetic wave (EM wave, light) pass through a material which is with self-magnetism of induced magnetism, The propagation speed of right and left circular palorized light will be different, that a linear polarized EM wave which consists of right and left circular palorized EM wave will change its polarization angle, this is called Faraday effect, discovered by Faraday in 1845. In 1877, Kerr found a result similar to Faraday’s. If a linear polarized light is incident into a ferromagnetic sample, since of the different reflection coefficients of right and left circular polarization components, the reflection beam will become elliptical polarized. This phenomenon is called magneto-optical Kerr effect. The angle between the primary axis of the elliptical polarization and the linear polarization is called Kerr rotation, and the ellipticity of the elliptical polarization is called Kerr elliptical. Let r+ eiθ+ and r− eiθ− stand for the reflection coefficients of right and left circular polarization, respectively. The Kerr rotation and Kerr ellipticity can be illustrated + as ϕk = − θ+ −θ and εk = 2. b a. =. r+ −r− r+ +r−. , respectively. Both of them are proven to be. proportional to the magnetization of sample. Thus by measuring ϕk and εk with cyclic applied magnetic field, we can get thehysteresis loop. In general, there are three types of MOKEmeasurement. Each of them has different geometry of the magnetization and the light path, In the polar Kerr effect, the magnetization lies in the plane of incidenceand is perpendicular to the surface. In the longitudinal Kerr effect, the magnetization lies in the plane of incidence and is parallel to the surface. In the transverse geometry, the magnetization is perpendicular to the plane of incidence and on the surface. The angle between the primary axis of the elliptical polarization and the linear polarization is called Kerr rotation, and the ellipticity of the elliptical polarization is called Kerr elliptical, as shown in Fig. 3.13. In magnetic ultrathin films, the Kerr signal is so small that the noise may result August 14, 2009.

(35) 3.5. Magneto-Optical Kerr Effect (MOKE). 29. Figure 3.13: Schematic illustration of magneto optical Kerr effect. After reflected from the ferromagnetic sample, the linear polarized laser beam becomes elliptical polarized.. Figure 3.14: Different geometry for MOKE measurement.. August 14, 2009.

(36) 3.5. Magneto-Optical Kerr Effect (MOKE). 30. Figure 3.15: Schematic illustration of AC MOKE.. in significant effect. Therefore, Practically in experiment here, a modulator is added on the laser sourcer and modulated signal can be taken by lock-in technique with a larger ratio of signal to noise. The schematic illustration is shown in Fig. 3.16.. Figure 3.16: Schematic display of a DC-MOKE loop in C207, NTNU August 14, 2009.

(37) 3.5. Magneto-Optical Kerr Effect (MOKE). 31. Figure 3.17: The switch of controlling the monopolar DC-power supply for inverse current. Figure 3.18: Schematic display of various magnetic field direction switch. The production of magnetic field are consist of four eletrical magnets, with the max field of 1500 Oe each and 4300 Oe combinatively.. August 14, 2009.

(38) 3.5. Magneto-Optical Kerr Effect (MOKE). 32. Figure 3.19: In-plane measurement with associated perpendicular field by the MOKE in C207, Department of physics, NTNU. The magnetic field intensity of in-plane file is stable; but due to the sample or the light spotposition, the intensity of perpendicular field in in-plane measurement has a range of±70%. The max perpendicular field in in-plane peasurement is about10%.. August 14, 2009.

(39) Chapter 4 Experiment and results. All the experiments are under UHV environment, usually less than 8.0 × 10−10 torr. Before preparing sample, we fill the cold trap with liquid nitrogen, that would help TSP working better. Sometimes, we isolate main chamber to get the base pressure reach 3.0 × 10−10 torr to preserve sample from pollution in a time.. 4.1. Si (111)7 × 7 obseavation. The silicon(111) slab as a substrate we used is n-type high doped, which can easily be heated by low voltage due to its small resistance about 4 Ω. When the silicon substrate was sent into the UHV chamber, it must be heated at 600 o C for more than 6 hours to completely get rid of the oxidation. In the earlier time, our silicon slab were from Unisoku company, with it’s size of 2×12 × 0.5mm3 and resistance about4.6 Ω at room temperature. Later, we use the silicon slab from prof. Kuo Chien-Cheng’s laboratory in National Sun Yat-Sen university, with its size of 2×9 × 0.5mm3 and resistance about12 Ω at room temperature. Although both these silicon slab are all high doped, but we cannot apply the same current to degas or flash, since the resistance is different, after overheating and breaking few silicon slabs, we can catch the temperature by observing the color. 33.

(40) 4.1. Si (111)7 × 7 obseavation. 34. Figure 4.1: The process of preparation Si(111)7 × 7 of (a)Si(111)from Unisoku company and (b)from prof. Kuo Chien-Cheng’s laboratory in National Sun Yat-Sen university. The last flash process is the very key to obtain large Si(111)7 × 7 terrace or narrow terrace. If the current finally reduced rapidly from 1000 o C, or said, fast cooling down, there will be narrow terrace on the Silicon surface; on the contrary, a larger terraces is possibly available as the current is reduced very slowly. The pressure is higher than 1.0 × 10−8 torr only during the early degassing process, the maximum pressure of other flash process is under 2.0 ×10−9 torr. No matter which process we choosed, there is always 7 × 7 structure on silicon(111) surface. The morphology analysis in main chamber is equipped with a STM(RT-STM, Unisoku).. August 14, 2009.

(41) 4.1. Si (111)7 × 7 obseavation. 35. Figure 4.2: The STM image first obtained in our laboratory.. Figure 4.3: A comparison with reference to calibrate [59] In figure 4.2, we can see a terrace with the width about 30nm and its reconstruction of Si (111)7 × 7. There are few defects, this is becuase the cooling process August 14, 2009.

(42) 4.1. Si (111)7 × 7 obseavation. 36. is still not slow enough. In figure 4.3, we drag a line and have a lineprofile to see the high between two terraces, and compare with a reference, the high defference in these two detail illustraton is about 5 %. Then in figure 4.4 we can calibrate the size of our STM image with a theoretical model of Si (111)7 × 7.. Figure 4.4: A scale comparison for calibration. (a) Our STM image, (b) a Si (111)7× 7 model. And we can see the faulted half and un-faulted half unit cell of Si (111)7 × 7 in larger scale. Another special structure of Si (111)7 × 7, the fualted and unfualted half unit cell can also be distinguished in our STM image showed in figure 4.5. Usually, to see the ualted and unfualted half unit cell image by STM, a 100nm × 100nm scale (or larger, 200nm × 200nm) is better for us in our system.. Figure 4.5: (a)The triangles indicate the fualted half and un-faulted half unit cell and (b)the shcemic picture from Omicro company.. August 14, 2009.

(43) 4.2. Highly ordered pyrolytic graphite (HOPG). 4.2. 37. Highly ordered pyrolytic graphite (HOPG). We use a commercial graphite to calibrate our STM image. Figure 4.6 shows two STM HOPG images we took with the model comparision at the right side, the discussion of varios STM images of HOPG was mentioned in the previous section in chapter 1. Figure 4.7 is the geometrical, or the long-width scale, comparison between our STM image and reference model, taking a rectangle block is helpful to obsreve, the width of our STM image is shorter than that of reference’s, the reason is our sample is slightly on a slant, not perfectly paralell to the STM scanning surface, thus we have this result. And figure 4.8 shows a lineprofile and a reference comparison.. Figure 4.6: (a)and (c) are STM images of the same graphite with different scanning condition. (b)and (d)are their schematic illustraton respectively [60].. August 14, 2009.

(44) 4.2. Highly ordered pyrolytic graphite (HOPG). 38. Figure 4.7: A magnitude check [60].. Figure 4.8: The heigh calibration [60]. (a) our STM image, (b) from reference and (c) the lineprofile in our STM image from (a).. August 14, 2009.

(45) 4.3. Two step growth of Fe/Au(111). 39. Figure 4.9: (a)The STM image with positive contrast,(d)the STM image with inverse contrast; (b)and (c)are HRTEM image with positive and inverse contrast respectively and (e)the lineprofile of (b)and (c). [60].. 4.3 4.3.1. Two step growth of Fe/Au(111) Fe nanostructures on Au(111) herringbone surface. To produce herringbone structure on Au(111) surface, first we use Ar+ with energy of 1000 − 700 eV to sputter the Au(111) surface for about 2 hours. This time setting is becuase that we did not calibrate the sputter focus, and the sputter gun usually shuts down from time to time. After finish cleaning the surface, or, while sputtering Au, we fill liquid nitrogen into manipulator to cool sample stage before sent Au into main chamber, this is becuase the cooling rate is not efficient(see figure 4.11), When annealing Au, the base pressure is near 1.0 × 10−9 torr. The current and duration are 4.2 A and 12 minutes. After annealing is finished, we deposit Fe on Au(111) 2.5 hours later for LT= 190 K, or 1.5 hours later for LT= 200 K, with the deposition rate of 0.15 ML/min, then use STM to investigate the morphology.. August 14, 2009.

(46) 4.3. Two step growth of Fe/Au(111). 40. Figure 4.10: The process of LT growth Fe on Au(111). Figure 4.11: The cooling curvature in C207 in NTNU. 4.3.2. Fe/Au vicinal growth. August 14, 2009.

(47) 4.3. Two step growth of Fe/Au(111). 41. Figure 4.12: (a)Reconstruction on Au(111) [61], (b) schematic structure (c) RT growth Fe/AU(111) [61] and (d) STM image in C207,NTNU. Figure 4.13: LT growth STM image (a) (b) (c) of Fe/Au(111) and (d) trend illustration. The growth process is showed in figure Once we want to scan the Fe/Au(111) surface by STM, found that in a part of the Au(111) crystal showed different morphology. August 14, 2009.

(48) 4.3. Two step growth of Fe/Au(111). 42. Figure 4.14: The comparison of 250 K and 180 K growth. (a) 250 K,0.18 ML [62] and (b) its schematic of sitting. (c) 180 K,0.15 ML and (d) its schematic of sitting. We can see the herringbone structure under this contrast.. Figure 4.15: The preferential nucleation of RT and 300 K. August 14, 2009.

(49) 4.4. LT growth of Fe/Si(111 )7 × 7 and MOKE. 43. Figure 4.16: (a) STM image for 180 K depositon, thickness= 0.45 ML, with 300 K anealing. (b) Shows its line profile. (c) is the process.. Figure 4.17: STM imageof the Fe thickness is 0.3ML on Au vicinal surface, deposited at 180K, RT annealing. (a) 50 × 50nm2 , (b) 80×80nm2 , (c) 100 × 100nm2 .. 4.4. LT growth of Fe/Si(111 )7 × 7 and MOKE. After Si(111 )7 × 7 substrate is prepared, we sent this Si substrate into MOKE chamber for low temperature deposition. The MOKE is cooled down by liquid nitrogen and has the lowest temperature about 98 K. The process we design is to deposit 0.5 ML on Si(111 )7 × 7 surface at low temperature 98 K, this is because we want to reduce the formation of iron silicide at Fe-Si interface. After the first LT deposition, we deposit Fe on it at RT to enhance its manetism, then measure the August 14, 2009.

(50) 4.4. LT growth of Fe/Si(111 )7 × 7 and MOKE. 44. magnetic characteristic in MOKE.. Figure 4.18: The shematic of two-step growth of Fe/Si(111). In the later process to return the sample temperature by taking off the sample from sample stage in MOKE. Before deposition, we measured MOKE of clean Si(111 )7 × 7 substrate, and we can see the result in figure 4.21.. Figure 4.19: A trying to see the effect from shielding. Without shielding, the Kerr intensity become larger, the shielding reduce the signal intensity. From figure 4.21 we see apparent noise background, once we try to take off the shielding on our magnets and measures again, the result is also put in figure 4.21.. In figure 4.19, it seems the Kerr intensity is stonger without shielding that we donot use is in the later experiments. In the early design, in order to prevent from Faraday’s effect in glass to break kerr signal, we add shieldings on view poles in optic’s path, showed by figure 4.20. August 14, 2009.

(51) 4.4. LT growth of Fe/Si(111 )7 × 7 and MOKE. 45. Figure 4.20: The Shielding on MOKE magnets to prevent from Faraday’s effect.. Figure 4.21: MOKE measurement of clean Si(111), in-plane, without shielding. The first try, figure 4.22 shows the process, with the results in figure 4.23, 4.24 and 4.25. While we do this experiment, STM was perturbed by unknown noise, we cannot by the time calibrate the deposition rate of the evaporation gun in MOKE, we deposit Fe on Silicon substrate in main chamber, the evaporation gun in main chamber was stable in its deposition rate1 , that was confirmed in earlier experiments.. 1. 1 KV, emissiom= 10 mV, the rate is 0.15 ML/min.. August 14, 2009.

(52) 4.4. LT growth of Fe/Si(111 )7 × 7 and MOKE. 46. Figure 4.22: Schematic of experiment process of the first try.. Figure 4.23: (a) and (b) are respectively the 1st and the 2nd in-plane measurement of thick Fe/Si(111). Figure 4.24: A zoom in view in max field of 300 gauss.. August 14, 2009.

(53) 4.5. MOKE calibration. 47. Figure 4.25: Perpendicular measurement of thick Fe/Si(111), compare with in-plane signal.. 4.5. MOKE calibration. In the bigining of MOKE testing, there was obviously a stong background. To clarify the problem, we insert a clean Si(111) slab and a copper plate into MOKE respectively, at first we donot want to vent UHV chamber, only test the effect from shielding. Later, we found that the evaporation gun in MOKE has a problem, then we vent MOKE chamber to fix it and try a complete test process by taking off the glasses, shielding and fore-polarizer to take MOKE signal. Since the shielding effect was confirmed in the previous experiments(see figure 4.19), we check the effect of the glass in view pole in MOKE. Than we check the glass effect.. Figure 4.26: Using copper plate as reflector, the comparison of with glass or not, in-plane.. August 14, 2009.

(54) 4.5. MOKE calibration. 48. Figure 4.27: Using copper plate as reflector, the comparison of with glass or not, perpendicular.. 8 Figure 4.28: Using copper plate as reflector, a combination view of with glass or not,max field is 3100 Oe, in-plane.. Figure 4.29: Using copper plate as reflector, a combination view of with glass or not,max field is 3100 Oe, perpendicular. August 14, 2009.

(55) Chapter 5 Discussion and conclusion 5.1. Si(111)7 × 7, HOPG calibration. From figure 4.7 and 4.8, we can check the magnitude reliability of our STM. In figure 4.3 we can find the difference of high is about 5%; figure 4.4 shows the longitudinal error is very small. It seems no problem in our STM measurement with Si as calibrator. When the sample replaced by HOPG, there is a small mismatch in some longitudinal scale, see figure 4.8. In order to check easily,we add a rectangle block in figure to help watch, we can see that if we set the length (the longer side) fixed, we must prolong the width to match the model. Compare with figure 4.7 and 4.9, we can conclud intuitively that the sample was tilt with a angle related to STM tip. This conjecture is resonable, becuase HOPG is layered graphene, the thicknes is not easily kept the same, especially everytime when we are going to sent a HOPG into UHV chamber, we just use a gummed tape to tear off the outer layer, thus the surface is possibly not parralel to the scaning path. The scale of our STM imgae is not probaly a problem, the oscilation noise is. Sometimes we suffered oscilation noise with frequency of hundreds Hz, 1.5KHz, and other stong oscillation noise when somewhere is under construction near our laboratory. That hurts the working efficiency and scanning quality, we are now trying to solve it. 49.

(56) 5.2. Two step growth of Fe/Au(111). 5.2 5.2.1. 50. Two step growth of Fe/Au(111) Fe nanostructures on Au(111) herringbone surface. Figure 5.1: Growth of Ag on Au(111) with different coverage. [42]. Figure 5.2: Growth of Al on Au(111) of 0.1 ML at different temperature. [43] The Au(111) reconstruction is one of the popular substrate for self-organization study. At lower coverage (less than 0.3ML) , transition metals with larger lattice August 14, 2009.

(57) 5.2. Two step growth of Fe/Au(111). 51. misfit (Fe, Co, Ni) on Au(111) are usually found nucleating at the kink of herringbone stucture, but in the contrast, this is not seem in small lattice misfit system (Al, Ag, Au/Au(111)). Although in the view of surface energy, Fe (2.9Jm−2 ), Co (2.7Jm−2 ), Ni(2.5Jm−2 ), Au (1.6Jm−2 ), these metal tend to form layer growth on Au’s surface, but the dislocation and different structures on the reconstruction surface of Au(111) lead to local difference. From figure 5.3, figure 4.13, 4.14 and 4.15 we can see that Fe doublets locate on fcc site when deposited at 180 K and by 300 K annealed, very different from RT growth and exposure xenon at 90 K. Next we compare figure 4.16, a repeat of 180 K deposition and RT annealing, from STM image we see no cluster merging and the distribution seems random, even the surface energy of Fe(2.9 Jm −2 )is greater than Au(1.6 Jm−2 ), iron clusters remain to amass high and are more disorder than be in exposured xenon at 90 K. To explain the doublets on herringbone fcc kinks, take figure 5.4 into consideration, in a 33 K growth; 80 K annealing and STM scanning experiment(figure5.4), except the nucleation at herringbone kinks, the preferential site of metal with large lattice misfit deposited on Au(111), there are Fe particles elsewhere. Although the random distribution might be due to the limited thermo energy of Fe particles, we can still see Fe depositions with a shape of long narrow piece on fcc elbow site, not cross discommensuration line but along fcc area, extending from the center of fcc elbow. In the Co/Au(111) system, see figure 5.5, the cobalt cluster on fcc site is larger than that on hcp site and cobalt has more tendancy to form bilayer growth. The Fe/Au(111) system seems not entirely similar to the Co/Au(111)’s, the height not directly means bilayer growth, but it gives ideas to explain. The harringbone kink is the primary nucleation site and Fe cluster on fcc kink is more flat than that on hcp’s, Fe grows stacking along the narrow channel of fcc site at herringbine kink. In another case of deposition at 180K; 300 K annealing and scanning (figure5.4(b)), we see doublets, singlets on the kink of fcc site and hcp site respectively, and we no longer see random distribution but clear nucleation on herringbone kinks. At 180 K the Fe particles have enough diffusion energy to site at the kinks, on the August 14, 2009.

(58) 5.2. Two step growth of Fe/Au(111). 52. Figure 5.3: (a) 0.18 ML Fe grown at 250 K. (c) 0.15 ML Fe grown at 180 K. (e) 0.27 ML Fe grown at 90 K on 0.09 ML RT-prepared seeds. (f) 0.27 ML Fe /8 L Xe on RTgrown 0.09 ML regular seeds. (b), (d), (f) and (h) are magnified images, revealing the herringbone background. Blue andred circles indicate the islands located on hcp and fcc kinks, respectively. [63] other hand, as the number of atoms in cluster increases, the narrow fcc channel walled by discomensuration lines then provides weaker stress relatively, the Fe-Fe stress overcome the Fe-Au’s at fcc site, thus the stacking clusters seperate into two, that we see doulets there. This evolution is not observed on hcp kinks, perhaps that the Fe-Au(111, hcp)stress does not change so dramatically. In the Co/Au(111) system, we see doublets both on fcc and hcp kinks only at 120K (not appear again at 70 K or 150 K), in this case, monolayer and bilayer growth is the spot point. Although August 14, 2009.

(59) 5.2. Two step growth of Fe/Au(111). 53. Figure 5.4: The comparision of deposition at (a)33 K and (b)180 K and their schematic nucleation below. [62]. Figure 5.5: (a)and (c) are the thickness as functions of temperature of cobalt on Au(111), and (b) the phase diagram. [44] the surface free energy between iron and cobalt is small( 0.2Jm −2 ), their cluster behavior on gold (111) surface is differnet, in order to find the most dominative factor, more clues should be found.. 5.2.2. Fe/Au vicinal-like growth. From figure 4.17 we can see straight wires, which seems a set of narrow terrace with width about 5 − 8nm. On the edge, clusters connect and merge into August 14, 2009.

(60) 5.2. Two step growth of Fe/Au(111). 54. Figure 5.6: Idea of the formation of doublets on fcc kinks. straight lines along the terrace edge. The most differences between this case and Fe/Au(788)’s are that there are few smaller cluster locate on the terrace belly and both these experiments are with 300 K annealing, the Fe/Au(788) obviously goes in step-flow growth, but we see Fe cluster amassing on the terrace edge and merging into line here. By observing the cluster density, we can draw a picture to illustrate the growth process of Fe on Au vicinal at 180 K.. Figure 5.7: A chematic illustrstion of the growth of Fe on Au vicinal at 180 K. In figure 5.3 (d)and (e),the Au(788) case, the terrace width are about 4−5nm, the clusters almost occupy the site on edge and be doublets. The terrace width is broader in our result, an explain is that the process includes a room temperature annealing, probably as like the Fe/Au(788) system with 300 K annealing, the step-flow effect may have been happened, if so, it implies the initial width of this terrace(5 − 8 nm, in STM image) would be close to that of Au(788)’s( 4 nm). August 14, 2009.

(61) 5.3. LT growth of Fe/Si(111 )7 × 7 and MOKE. 55. Figure 5.8: A full comparison, (a),(b)and (c)are our STM image of Fe/Au vicinal; (d),(e)are Fe/Au(788) at 200 K and 165 K respectively; (f)shows the annealing effect at 300K and (g)shows the lineprofile of (d),(e),(f). There are questions that cannot easily to gain a clear idea with our scanty information: What the structure the edge is? It seems there will be doublets growth on fcc site at lowtemperature whether in Au(111)herring bone surface or Au(788) surface. The reference shows that in Au(788) vincinal it goes step-flow growth [47], but what is the behavior in the Au(111) herringbone of Fe cluster with higher coverage? These quetion are quite interesting now.. 5.3. LT growth of Fe/Si(111 )7 × 7 and MOKE. The motivation we want to growth of Fe/Si(111 )7 × 7 at low temperature is to prevent the formation of iron silicide at the interface and expect we can see the August 14, 2009.

(62) 5.3. LT growth of Fe/Si(111 )7 × 7 and MOKE. 56. quantum well effect on it. The background is really a problem in measuring MOKE, but we still found a in-plane hysterisr loop in this system (figure 4.23, 4.24). From figure 4.26 to 4.29, we can make sure the glass effect, and hope these experience can help us to clarify the MOKE measurement in the later work. In the later work, we are going to fix the bug in program and repeat this experiment again.. August 14, 2009.

(63) Chapter 6 Summary • A multifunctional UHV system with the best base pressure of 3.0 × 10−10 torr is now working. • The sample preparation process has been confirmed. And the scale of STM image therefore can be corrected. • At 180K, the growth of Fe on Au(111) herringbone surface has a defferent behavior from RT growth and exposuring xenon at 90 K, we found Fe doublets on fcc site. • At 180 K, the growth of Fe on Au vicinal surface behavors different from Fe/Au(788), there is no obvously step-flow growth but amass and merging on terrace adge. • LT growth of Fe on Si(111)7 × 7 is a practicable idea, since we get a in-plane magnetic signal from MOKE.. 57.

(64) Bibliography [1] A. Pimpinelli,and J. Villain in Physics of crystal growth,(Cambridge University Press, United Kingdom, 1998), p. 144. [2] Brommer, J. Appl. Phys. 89, 7139 (2001). [3] Stich, Phys. Rev. B 55, 5886 (1997). [4] K. Kataoka, K. Hattori, Y. Miyatake, and H. Daimon,Phys. Rev. B 74, 155406 (2006). [5] K. Radermacher, A. Schppen, and S. Mantl, Solid-State Electron. 37, 443 (1994). [6] N. G. Galkin, D. L. Goroshko, S. Ts. Krivoshchapov, and E. S. Zakharova, Appl. Surf. Sci. 175-176, 230 2001. [7] D. Berling, G. Gewinner, M. C. Hanf, K. Hricovini, Loegel, A. Mehdaoui, C. Pirri, M. H. Tuilier, Magn. Magn. Mater. 191, 331 (1999). [8] Y. Takagi, A. Nishimura, A. Nagashima, and J. Yoshino, Surf. Sci. 514, 167 (2002). [9] A. N. Hattori, K. Hattori, K. Kataoka, and H. Daimon unpublished. [10] J.S. Tsay , T.Y. Fu, M.H. Lin, C.S. Yang, and Y.D. Yao, Appl. Phys. Lett. 88 , 102506 (2006). [11] T.Y. Fu, J.S. Tsay , M.H. Lin, and Y.D. Yao, J. Magn. Magn. Mater. 304 , E128 (2006). 58.

(65) Bibliography. 59. [12] H.W. Chang, J.S. Tsay , Y.C. Hung, F.T. Yuan, W.Y. Chan, W.B. Su, and Y.D. Yao,J. Appl. Phys. 101(9) , 09D124 ( 2007). [13] K. Hattori, T. Nishimura, K. Kataoka, Y. Shimamoto, and H. Daimon, Thin Solid Films 464-465, 5 (2004). [14] P. Rygus, P.-R. Steiner, and H. Froitzheim, Surf. Sci. 352-354, 523 (1996). [15] http://en.wikipedia.org/wiki/HOPG [16] F. Atamny, O. Spolleck and R. Schlögl, Phys. Chem. Chem. Phys.,1, 41134118 (1999) [17] G. Binnig, H. Fuchs, C. Gerber, H. Rohrer, E. Stoll and E. Tosatti, Europhys. L ett., 1, 31 (1986,). [18] H. J. Mamin, E. Ganz, D. W. Abraham, R. E. Thomson and J. Clarke, Phys. Rev., B, 34, 9015 (1986). [19] R. Wiesendanger and D. Anselmetti, in Scanning Tunneling Microscopy I, ed. H.-J. Gu ntherodt and R. Wiesendanger,(Springer, Berlin, 1992), vol. 1, p. 131 [20] S. Rousset, B. Croset, Y. Girard, G. Prevot, V. Repain, S. Rohart, C. R. Physique 6, 33-46 (2005). [21] H. Ellmer, V. Repain, M. Sotto, and S. Rousset, Surf. Sci. 5,11 183 (2002). [22] K. Bromann, M. Giovannini, H. Brune, and K. Kern, Eur. Phys. J. D 9, 25 (1999). [23] H. Brune, M. Giovannini, K. Bromann, and K. Kern, Nature 394, 451 (1998). [24] W. C. Lin, C. C. Kuo, M. F. Luo, K. J. Song, M.-T. Lin, Appl. Phys. Lett. 8,6 043105 (2005). [25] M. F. Luo, C. I. Chiang, H. W. Shiu, S. D. Sartale and C. C. Kuo, Nanotech. 17, 360 (2005). August 14, 2009.

(66) Bibliography. 60. [26] W. C. Lin, P. C. Huang, K. J. Song, M.-T. Lin, Appl. Phys. Lett. 88, 153117 (2006). [27] Z. Gai, B. Wu, J. P. Pierce, G. A. Farnan, D. Shu, M. Wang, Z. Zhang, and J. Shen, Phys. Rev. Lett. 89, 235502 (2002). [28] G.E.Poirier, Chem. Rev. 97, 1117-1127 (1997). [29] Chidsey, C. E. D.; Loiacono, D. N.; Sleator, T.; Nakahara, S. Surf. Sci. 200, 45 (1988). [30] S. Rohart, Y. Girard, Y. Nahas, V. Repain, G. Rodary, A. Tejeda, S. Rousset, Surf. Sci. 602, 28 (2008). [31] J. Meyer, I. Baikie, E. Kopatzki, R. Behm, Surf. Sci. 365, L647 (1996). [32] Joseph A. Stroscio, D. T. Pierce, R. A. Dragoset, and P. N. First, J. Vac. Sci. Technol. A10, 1981 (1992). [33] Neetha A. Khan, Christopher Matranga, Surf. Sci. 602, 932 (2008). [34] D. T. Dekadjevi, B. J. Hickey, and S. Brown, Phys. Rev. B 71, 054108 (2005). [35] P. Ohresser and N. B. Brookes, S. Padovani, F. Scheurer and H. Bulou, Phys. Rev. B 64, 104429 (2001). [36] H. Bulou, F. Scheurer, P. Ohresser, A. Barbier, S. Stanescu, and C. Quiros, Phys. Rev. B 69, 155413 (2004). [37] S. Padovani, F. Scheurer, J. Bucher, Europhys. Lett. 45, 327 (1999). [38] D.D. Chambliss, R. J. Wilson, and S. Chiang, Phys. Rev. Lett. 66, 1721 (1991). [39] M. Fonin, Yu.S. Dedkov, U. Rudiger, G. Guntherodt, Surf. Sci. 529, L275L280 (2003). [40] W. C. Lin, S. S. Wong, P. C. Huang, C. B. Wu, B. R. Xu, C. T. Chiang, H. Y. Yen, M.-T. Lin, Appl. Phys. Lett. 89, 153111 (2006). August 14, 2009.

(67) Bibliography. 61. [41] N.Camillone; et al. J. Chem. Phys. 101, 11031(1994). [42] M. M. Dovek, C. A. Lang, J. Nogami, and C. F. Quate, Phys. Rev. B 40, 11973 (1989). [43] B. Fischer, H. Brune, Johannes V. Barth, A. Fricke, and K. Kern, Phys. Lett. 82, 1732 (1999). [44] I. Chado, C. Goyhenex, H. Bulou,* and J. P. Bucher, Phys. Rev. 69, 085413 (2004). [45] S. Boukari, and E. Beaurepaire, Phys. Rev. B 55, 13932 (1997). [46] F. J. Himpsel, J. E. Ortega, G. J. Mankey and R. F. Willis, Advances in Physics, 47, No. 4, 511- 597 (1998). [47] S. Rohart, Y. Girard , Y. Nahas, V. Repain, G. Rodary, A. Tejeda, S. Rousset, Surf. Sci. 602, 28 (2008). [48] T.Chiang, Surf. Sci., 39, 7-8, 181-235 (2000). [49] S.˚ A. Lindgren, L. Walldén, Phys. Rev. Lett. 59, 3003 (1987). [50] T. Miller, A. Samsavar, G.E. Franklin, T.-C. Chiang, Phys. Rev. Lett. 61, 1404(1988). [51] S.˚ A. Lindgren, L. Walldén,Phys. Rev. Lett. 61, 2894 (1988). [52] M.A. Mueller, A. Samsavar, T. Miller, T.-C. Chiang, Phys. Rev. B 40, 5845(1989). [53] S.˚ A. Lindgren, L. Walldén,J. Phys.: Condens. Matter 1, 2151(1989). [54] M.A. Mueller, T. Miller, T.-C. Chiang, Phys. Rev. B 41, 5214 (1990) . [55] T. Miller, A. Samsavar, M. Mueller, G. Franklin, T.-C. Chiang, Physica Scripta T 31, 35 (1990) . [56] L.N. Leiberman, D.R. Fredkin, H.B. Shore, Phys. Rev. Lett. bf2, 539 (1969.) August 14, 2009.

(68) Bibliography. 62. [57] L.N. Leiberman, J. Clinton, D.M. Edwards, J. Mathon, Phys. Rev. Lett. 25, 232 (1970). [58] S. Hope, E.Gu, B. Choi and J. K. Howard,J.Appl. Phys.62,2989 (1987). [59] Becker R. S., Golovchenko J. A., McRae E. G. and Swartzentruber B. S., phys. Rev. Lett., 55 2028 (1985) [60] F. Atamny, O. Spillecke and R. Schlo. gl. Phys. Chem., 1, 4113-4118 (1999). [61] Neetha A.Khan,Christopher Matranga, Surf.Sci. 602, 932942 (2008). [62] Provided by prof. Kuo Chien-Cheng’s laboratory in National Sun Yat-Sen university. [63] W.C. Lin, H.Y. Chang, Y.C. Hu, Y.Y. Lin, C.H. Hsu, and C.C. Kuo, Appl. Phys. Lett. in press (2009). August 14, 2009.

(69)