鈀/鐵,鈷,鎳/藍寶石基板(0001)系統晶格結構與磁性研究

全文

(1)國立台灣師範大學物理研究所碩士論文. 鈀/鐵,鈷,鎳/藍寶石基板(0001)系統晶格結構與磁性研究 Crystalline Structure and Magnetism of Pd/Fe,Co,Ni bilayers on Al2O3(0001). 指導教授: 林文欽研究生: 紀喬崧中華民國. 博士撰. 一百零一年六月.

(2)

(3) 致謝碩士班的這兩年,真的學到很多東西,也謝謝很多人的幫忙,我才能順利的完成學業。首先要很感謝我的父母親,他們都很支持我讀碩士班,有了他們的支持, 可以專心的做研究無後顧之憂,以及能體諒、替我加油的女朋友怡人。當然,非常感謝林文欽老師這兩年的指導,當初我考進師大,是備取生、沒做過專題、也不是本來師大的學生,想當然找教授也碰了不少釘子。但和老師的面談中,不是問考試排名,而問了關心學生的問題和詳細說明研究的領域,這大概也是我後來決定找老師的原因吧! 還要特別謝謝駱芳鈺老師,去年暑假學校防震工程,借我們一些空間讓我們的實驗可以繼續進行。還有中山的郭建成老師和吳世鈺學長 ,不管是在聯合 meeting 上,還有對問題的討論、合作的研究都讓我有很大的收穫,以及淡江王柏堯學長替我們於同步輻射中心的量測給予很大的幫助。最後謝謝蔡志申老師及魏德新老師擔任口試委員。除了各位教授之外,很感謝王志雄、蔡承叡、李寶生學長,黃雅筠學姐,同期的好隊友何宗穎,碩一薛昆仁、莊英奇。特別謝謝志雄學長、雅筠學姐教我們如何修理、操作儀器、真空技術等等,以及宗穎、英奇在我實驗分身乏術的時候替我買晚餐。最後勉勵之後的學弟妹:實驗室是一個團隊,期望未來能看到你們更多的研究成果。.

(4) 中文摘要: 此篇論文分為兩大主題,第一部分是透過斜角度之方式鍍鐵於藍寶石基板 (0001)上,最後覆蓋上鈀約 30 ML,目的是保護基板上的鐵磁層避免氧化。所有的樣品皆在超高真空腔(UHV:10-9torr)之下用熱蒸鍍原理製程,隨後破真空於大氣下量測,包含磁光柯爾效應(MOKE)、X 光繞射(XRD)、邊緣 X 射線吸收系微結構 (EXAFS)、掃描穿隧式電子顯微鏡(STM) 等等。藉由改變鍍膜鐵的角度和鐵的厚度,可以控制鐵薄膜的磁性行為;當鍍膜的傾斜角度越大的時候,會使得表面粗糙度和原子排列的亂度越大。第二部分是鈀吸附氫氣的實驗,於 UHV 系統下製備 n ML Pd/Fe,接著用 MOKE 量測曝氫後的磁性行為。發現曝氫於一大氣壓後,會產生消光角的位移以及 p 方向電場平方. 的改變。隨著鈀厚度的增加,消光角位移程度越大,. 的變化. 量也越大,在特定檢偏鏡角度下,發現 MOKE 訊號會隨著 Pd 的厚度(30 ML~60ML) 而增加 10%~40%。之後我們固定 Pd 的厚度在 60 ML,改變中間的鐵磁層,比較 Pd/Fe、Pd/Co、Pd/Ni 曝氫的行為,發現 MOKE 訊號的增加量分別增加約 40%、35%、和 60%。同時也量測吸附和脫附氫氣所需的時間以及飽和所需之氫氣壓力,三個樣品的氫氣飽和壓力都在 100 mbar 以下,當氫氣壓力於 1 atm 所需飽和時間皆小於 10 分鐘,但氫氣脫附的時間, Pd/Fe、Pd/Co 約需要 8 小時, Pd/Ni 只需不到 2 小時,而且其反應都是可逆的。關鍵字: 斜向鍍膜、磁光柯爾效應. 1.

(5) Abstract This study is separated into two parts: (1) Using oblique deposition to change the surface morphology. (STM measurement) We investigated the corresponding magnetic behavior (MOKE measurement), and the crystalline structure (XRD, EXAFS measurement) in Pd/Fe/Al2O3(0001) system. (2) H2 effect in MOKE, we absorption H2 from 10-3 mbar to 1013 mbar and measured MOKE in Pd/Fe,Co,Ni/Al2O3(0001) systems. In many studies, surface morphology plays an important role in magnetic behavior. They are many methods to change it, like ion-sputtering, different stepped substrate…etc. One of the feasible methods is changing the surface morphology by different deposit angle. In Pd/Fe/Al2O3(0001) system. In normal deposition, Fe atoms were follow the sapphire (0001) growth, and the corresponding magnetic hysteresis loop was almost square, it was isotropy. In oblique deposition, Fe grains were connected to form a 1-dimenstorn ripple structure, and habited interesting magnetic behavior. The easy axis is parallel to the ripples, and the hard axis is perpendicular to the ripples. In easy axis, the magnetic hysteresis loop was also square but the coercivity was much bigger than normal deposition. In hard axis, we could see the obvious uniaxial magnetic anisotropy. Another ideal in this experiment was H2 absorption effect on MOKE. We capped Pd films on the top surface after Fe deposition. There are two reasons for the Pd capping layer, one is to protect Fe films oxidation, the other is Pd can adsorption H2 and the optical of MOKE is changed. Key words: oblique deposition、magnetic optical Kerr effect. 2.

(6) Outline Abstract……………………………………………………………………….1 1. Introduction 1.1 Uniaxial Magnetic Anisotropy in Pd/Fe/Al 2 O 3 (0001) by oblique deposition. …………………………………………………………………...5 1.2 H2 absorption effect in room temperature by MOKE measurement in Pd/Fe,Co,Ni/Al2O3 (0001) system………………………………………….7. 2. Basic Concepts 2.1 Magnetic material…………………….……………….…………………….8 2.2 Hysteresis loop………………………….…………………….…………….8 2.3 Kerr Rotation and Intensity……………………………………………….10 2.4 Shadowing effect by oblique deposition…………………………………11 2.5 Magnetic Anisotropy Energy (MAE)……………………………………...12 2.6 H2 absorption effect with Pd thin films……………………………………15. 3. Experimental instrument 3.1 Ultra High Vacuum (UHV) system………………………………………..17 3.2 Low Energy Electron Diffraction (LEED)…………………………………20 3.3 Auger Electron Spectroscopy (AES)……………………………………..25 3.4 Magneto Optical Kerr Effect (MOKE)…………………………………….28 3.5 Scanning Tunneling Microscopy (STM)………………………………….33 3.6 X-Ray Diffraction (XRD)…………………………………………………...35 3.7 Extended X-ray Absorption Fine Structure (EXAFS)…………………...37. 4. Experiment and Result (Part 1.) Uniaxial Magnetic Anisotropy in Pd/Fe/Al2O3(0001) by oblique deposition. 4.1 Preparing Pd/Fe(0,45,65-deg)/Al2O3(0001)………………………………39 4.2 MOKE-magnetic behavior………………………………………………….40 4.3 XRD- crystalline structure……………………………………………….....44 4.4 EXAFS- crystalline structure……………………………………………….47 4.5 STM-morphology…………………………………………………………....49 4.6 MOKE-magnetic behavior in Pd/Fe(25ML)-0,45,65˚/Al2O3(0001) ……..51 4.7 Hc and Mr ration V.S. Azimuthal angle……………………………………53 3.

(7) 4.8 Hc in easy axis V.S. Fe thickness, deposition angle 0, 45, 65˚…………55 4.9 Ku in Hard axis V.S. Fe thickness, deposition angle 45, 65˚……………57. 5. Experiment and Result (Part 2.) H2 absorption effect in room temperature by MOKE measurement in Pd/Fe,Co,Ni/Al2O3(0001) system 5.1 Preparing Pd/Fe,Co,Ni/Al 2O3(0001)……………………………………....60 5.2 To proofread the Precise Rotator………………………………………….60 5.3 How to measure Kerr rotation curve and correspond intensity…………61 5.4 n ML Pd/Fe/Al2O3(0001) + H2………………………………………………63 5.5 60 ML Pd/Fe,Co,Ni/Al2O3(0001) + H2………………………………….….70. 6. Discussion and Questions….………………………………...74 7. Summary……………………………………………………………….…..77 8. Bibiography……………………………………………………………….79. 4.

(8) Chapter1 Introduction 1.1 Uniaxial magnetic anisotropy in Pd/Fe/Al2O3 (0001) induced by oblique deposition. In ferromagnetic (FM) materials, the magnetic anisotropy energy (MAE) is defined as the energy difference when the magnetic moment is aligned in different orientations. MAE not only determines the magnetization direction, the easy axis, and the stability of the ferromagnetic ordering, but also is correlated to the magnetization processes of FM materials [1,2]. Based on the various physical origins of MAE, the manipulation of MAE has been demonstrated to be important and feasible, especially in nano-sized materials. It was applications in magnetic data storages and nano-devices. One of the methods is using MAE by tuning the shape of the magnetic materials. Oblique deposition of magnetic atoms on suitable substrates can lead to highly elongated grains or ripples with nanometer scale width. It is because of the shadowing effect [3]. These elongated nanograins or ripples were induced uniaxial MAE.. Fig 1.1: STM image and hysteresis curves of the Co film after nanostructuring with an ion dose of about 12 ML. R. Moroni, D. Sekiba, F. Buatier de Mongeot et al, Phy. Rev. Lett. Vol. 91, Num. 16 (2003) [4]. Most of studies focus on the oblique-incidence metal (001) epitaxial growth, for example Co/Cu(001) [4] and Fe/Ag(001) [5]. Recent studies expand to the heterostructures of metal/semiconductor and metal/insulator, such as Fe/MgO(001) [6] and Co/Si(111) [7]. For cubic (001) substrates, the oblique-deposition induced uniaxial MAE usually mixes with a four-fold symmetric MAE, which is induced by the substrate crystalline structure. For the 5.

(9) six-fold symmetric case, the substrate induced MAE is relativity weak, and the grazing incidence growth effect on MAE is more obviously to examine. Up to now, the oblique-deposition induced uniaxial MAE is limited to a few substrates. Some important issues are still unclear. For example, will a protective capping layer destroy the self-organized nanostructure, as well as the uniaxial MAE? The capping layer effect is unavoidable in application, and it is seldom discussed. Besides, how is the evolution of crystalline structure with the increasing oblique angle? In our study, Fe films were grown on the single crystalline sapphire(0001) by oblique-incidence deposition. The transparency of sapphire substrate provides this system a suitable candidate for magneto-optical plamonic [8] applications. The changes of surface morphology, crystalline structure, and magnetic behavior, caused by oblique-deposition, were investigated.. 6.

(10) 1.2 H2 absorption effect in room temperature by MOKE measurement in Pd/Fe,Co,Ni/Al2O3(0001) system Hydrogen is absorbed by many metals and occupies interstitial sites in the host lattice. One of the popular metals is palladium (Pd), bulk Pd can absorb H2 and Pd lattice constant will expand about 3% [9]. In the initial stage H atoms occupy the interstitial sites of Pd crystalline lattice ( phase). With increasing hydrogen adsorption, the Pd lattice is expanded up to 2-3% ( phase) [10,11]. The physical properties of Pd film, such as electronic structures [12], optical properties [13,14] and mass [11], can be drastically changed by the hydrogen occupation in Pd thin film. In the experiment of H2 absorption, temperature plays an important role. In some surface studies, they investigate this effect in a UHV system, by cooling down the sample to 50K, Pd can absorption H2 spontaneously. But in our case, we study it at room temperature and the vacuum pressure is about 10-3 mbar. Study H2 absorption effect in RT has two requirements, one is the enough Pd thickness (>4nm), the other is the hydrogen pressure (>100mbar) [9]. In our studies, we are interested in the magnetic properties, so we deposit the double layer systems, like Pd/Fe,Co,Ni and expose them to 1 atm H2. After H2 exposure, the extinction angle of MOKE was shifted and correspondingly Kerr signal was significantly changed. When properly choosing the analyzer polarization angle, we can obtain large enhancement of 40%, 35% and 70% for Pd/Fe, Pd/Co and Pd/Ni bilayers, respectively. The enhancement was observed within a few minutes (min) after H 2 exposure. However after H2 was pumped out, the recovery of MOKE signal to 80% of original value took 30 min for Pd/Ni, and even longer (~175 min) for Pd/Co, and Pd/Fe. The reversibility of this H2 adsorption effect on MOKE was also demonstrated. These observations reveal the universal responses when combing hydrogenated Pd and magnetic thin films, and could be applied in various magnetic materials.. 7.

(11) Chapter2 Basic Concepts 2.1 Magnetic material Magnetic materials all can be magnetized with an external magnetic field. The magnetic moment in a unit volume is called magnetization, used . Because different materials have different atomic structure and particular arrangement of magnetic moment, they have really different magnetic behaviors. Generally speaking, we can used magnetic susceptibility for discussion, where. is the ratio of magnetization and magnetic field.. According to the magnitude of , magnetic materials can be separate into 4 kinds. Paramagnetism, Ferromagnetism.. Diamagnetism,. Anti-ferromagnetism,. and. 2.2 Hysteresis Loop. Fig 2.1: Hysteresis loop, Coercivity, and Retentivity [15] Hysteresis loop shows the relationship between the induced magnetic flux density (B) and the magnetizing force (H). The loop is generated by measuring the magnetic flux of a ferromagnetic 8.

(12) material while the magnetized or has been demagnetized will follow the dashed line as H is increased. As the line demonstrates, the greater the amount of current applied (H+), the stronger the magnetic field in the component (B+). At point "a" almost all of the magnetic domains are aligned and an additional increase in the magnetizing force will produce very little increase in magnetic flux. The material has reached the point of magnetic saturation. When H is reduced to zero, the curve will move from point "a" to point "b." At this point, it means that some magnetic flux remains in the material even though the magnetizing force is zero. This is referred in this point, some of the magnetic domains remain aligned but some have lost their alignment. As the magnetizing force is reversed, the curve moves to point "c", where the flux has been reduced to zero. This is called the point of coercivity on the curve. The force required to remove the residual magnetism from the material is called coercivity. As the magnetizing force is increased in the negative direction, the material will again become magnetically saturated but in the opposite direction (point "d"). Reducing H to zero brings the curve to point "e." It will have a level of residual magnetism equal to that achieved in the other direction. Increasing H back in the positive direction will return B to zero. Notice that the curve did not return to the origin of the graph because some force is required to remove the residual magnetism. The curve will take a different path from point "f" back to the saturation point where it with complete the loop. From the hysteresis loop, a number of primary magnetic properties of a material can be determined [15].. 9.

(13) 2.3 Kerr Ratation Curve and Correspnding Intensity. Fig 2.2: Kerr rotation angle and Kerr rotation curve Kerr rotation angle is the angle of original linear polarized light and the long axis of elliptically polarized light. By measuring Kerr rotation curve, we needed to define the magnetic field. Customarily, we define right-handed is the positive direction (clockwise) and left-handed is negative (counterclockwise). So in the process of rotating the polarization (Accuracy is about 0.0163˚) with an external magnetic field for opposite direction. We can get two Kerr rotation curves. And the corresponding intensity is the subtraction of two opposite magnetic field lock-in signal at the same polarization angle. In hydrogen absorption experiment, the curve and intensity are the two important quantities. By absorption H2, the curve will shift, and in specific polarization, we can see the intensity increase about 35~70% in 60 ML Pd/Fe,Co,Ni/Al2O3(0001) substrate.. Fig 2.3: Analysis angle corresponding Kerr intensity. 10.

(14) 2.4 Shadowing effect by oblique deposition We used oblique deposition, the first particles are grown following the substrate, like Fig 2.4(a). When the particles are large enough to become grains, There will be a shadow on the other side of grains, like Fig 2.4 (b). And ion particles are more difficult to growth in this part. By this reason, the surface morphology was changed by oblique deposition. For example, the 1D ripple structure in STM images, like Fe/Si(111) [16]. The orientation of grain columns, the porosity, the crystallographic texture, and grain size are sensitive to the deposition angle. The origin of this effect is associated with shadowing. In order to isolate the effects of shadowing from other physical effects (such as surface diffusion, deposition species size, flux divergence, etc.), we have constructed a simulation where all of these effects are completely removed. These simulations demonstrate that many of the observed structural properties of obliquely deposited films are controlled by shadowing [17].. Fig 2.4: Shadowing effect by oblique deposition 11.

(15) 2.5 Magnetic anisotropy Magnetic anisotropy is the energy difference of substances when the magnetization is changed to different directions. When a fixed external magnetic field is applied along different direction of the magnetic material, the magnetization processes in each direction is different. In all the directions, the easiest saturated direction is called easy axis, the most difficult direction to get saturated is called hard axis. This kind of internal energy is called magnetic anisotropy energy (MAE). We usually used MAE to discuss the direction of easy and hard axis. MAE is means the work needed about transfer self direction of magnetization to external magnetic field direction. And the easy axis is the direction which has minimum E. Generally the total energy expressed as:. : The constant of MAE, and the unit can be :. or. or. The angle of magnetization direction and normal direction. From equation (1), when , means easy axis prefer the direction of perpendicular to the sample surface. When , means easy axis prefer the direction of parallel to the sample surface.. 12.

(16) Structure and Magnetic properties in block Fe、Co and Ni. Fig 2.5: Block Fe,Co,Ni, anisotropy energy and prefer direction (Easy axis). Three kinds of Magnetic anisotropy:. Fig 2.6: Shape anisotropy and Magneto-crystalline anisotropy 1. Shape Anisotropy: When spin arrangement has to be affected by magneto crystalline, the sample surface will become magnetic dipole. Those magnetic dipoles have interaction by itself, and the energy of diploe-dipole interaction results in a net magnetic moment. Without external magnetic field, if all the spin arrangement in the same direction, it has the maximum net magnetic energy. In reality, spin 13.

(17) will prefer magnetic domain and decrease net magnetic energy. If the sample is a long and narrow stripe, the long axis will become an easy axis and it has bigger MAE. Therefore, the interaction of magnetic dipole moment produces magnetic anisotropy. It has affected by shape, and called shape anisotropy. 2. Magnetocrystalline anisotropy: In the magneto-crystalline structure, magnetic anisotropy energy and self magnetization are related to the direction of crystal axis. The symmetry of the materials and crystalline structure is called magneto-crystalline anisotropy. It comes from spin orbital coupling. Spin orbital interaction will let the materials self magnetized along the crystal axis, so the easy magnetization axis of magneto-crystalline anisotropy is the crystal axis of the crystal. If the crystal only has single magnetization axis, it is called uniaxial anisotropy. We often used magneto-crystalline anisotropy to describe the anisotropy of thin films. By uniaxial magnetic anisotropy sample, it can represent: : The constant of magneto-crystalline anisotropy, and : The angle between magnetization direction and easy magnetization axis. 3. Stress anisotropy: Put a magnetic material into a magnetic field, it can be magnetized by external magnetic field. The coupling effect between atoms are changed by the external field, create a small deformation. The size of the deformation is also anisotropic, so the symmetry of the magneto-crystalline axis is destroyed. The direction of self magnetization is changed. This phenomenon is called magnetoelastic anisotropy or magnetostriction anisotropy, or stress anisotropy.. 14.

(18) 2.6 H2 absorption effect with Pd thin films It is known that H2 is chemisorbed on the Pd surface, where the H2 molecule is dissociated into two H atoms, and the atomic H then diffuses to the interior. In the initial stages the phase is formed, in which the H atoms occupy interstitial sites in the Pd lattice. In this phase the H wave function is extended and the proton vibrates with large amplitude. Further absorption in bulk Pd causes significant volume expansion of the lattice due to the hydride phase. In these films the phase diagram appears to be thickness dependent. This is not well understood because the substrate plays a significant role by clamping the in-plane lattice expansion.. Fig 2.7: Cubic Pd structure and after absorption hydrogen, Pd lattice were expand. 15.

(19) Fig 2.8: The steps of H2 diffusion in Pd films Renaud Delmelle and Joris Proost, Phys. Chem. Chem. Phy.,2011,13,11412-11421[18] The reaction between hydrogen and Pd or Pd alloys is interest. It is known that H2 spontaneously adsorbs to Pd as atomic H and diffuses into the lattice to form PdHx. The initial -phase Pd becomes -phase PdHx through an phase transition. The Pd lattice spacing changes throughout these phase changes [18], depending on the H2 concentration in the surrounding atmosphere. The phase transitions and changes in lattice spacing lead to measurable changes in the optical properties, resistance, and mass of the Pd. In our studies, we used MOKE to do the measurement, we deposit Pd/magnetic layer, and want see difference after hydrogen absorption.. 16.

(20) Chapter3 3.1 Ultra High Vacuum (UHV) system In surface science study, ultra high vacuum chamber is necessary. In our Lab, we have three different chambers, each chamber has different purpose. First is the purely chamber, in the Fig 3.1. We only use this chamber to preparing samples, it doesn’t have any measurement device in this chamber. This chamber is disposed with a mechanical pump and a turbo pump, and the pressure can achieve about torr. We generally use this chamber to deposit ferromagnetic materials (Fe Co Ni) with a capping layer (Au Pd Cu) to protect under layer films. After a sample is prepared, we take it out off vacuum and measure MOKE, XRD, EXAFS …extra. Second is STM chamber, it is set up on the optic table. To avoid external vibrate, the optic table can float by the four pneumatic pillars. By using load-leg we can transfer our sample to STM. Third is the complicated and the biggest chamber. It disposes for a mechanical pump, a turbo pump, an ion pump, and a titanic sublimation pump (TSP). With all pumps, the bass pressure can achieve 10 -10 torr. Furthermore we have AES, LEED, MOKE in this chamber. By gas kinetic theory, at a constant temperature and pressure we can estimate the frequency for gas atoms to collide the sample. The frequency is. Consider a gas molecular of mass m, at temperature T, the root mean square velocity is. Where is Boltzmann’s constant the average velocity is. and ideal gas formulation Compare all the eq. we can get. 17.

(21) P in mbar, T in Kelvin (K), m in molecular mass(M). For example, N2 molecular mass is 28 g, at room temperature (300K), with pressure of 1 mbar, the frequency of gas atoms colliding with the sample surface is. , consider a. sample. 1ML have. atoms, assume all the atoms will stay on the surface, under 1 mbar, about second will become one atom layer ; under 10-6 mbar , about after 5s; under 10-9 mbar, about after 1hr. It’s means that, if we want the sample is clean enough, almost no pollution, the pressure were below 10-10 mbar.. Fig 3.1: Sample preparing chamber, the base pressure is. Fig 3.2: STM chamber is on the float optical table 18. torr.

(22) Fig 3.3: Main chamber, it equip with AES, LEED, and MOKE 19.

(23) 3.2.1 Low Energy Electron Diffraction (LEED) In surface science study, the popular device to observe the atoms arrangement and the structure of thin films is LEED. By LEED images, we can know whether the atoms arrange are regular or not. Generally used 20~1000eV for the incidence electronic energy. From de Broglie equation:. E: Incidence electronic energy λ :Corresponding the incidence electronic wave length When the incidence electronic energy is 20eV~1000eV, the corresponding wave length is 2.74~0.388Å . The materials lattice constant is close to the wave length, so the reflection will produce diffraction pattern.. For 1D space: a: The distance of two atoms, is also the materials lattice constant. θ: The angle of incidence electron beam and reflex electron beam. The distance of two reflex electron beam is , if two reflex electron beam become to complete constructive interference, then the distance has to be integer multiple of the wavelength of the electron. λ. λ and a are known, from the angle of different incidence electron beam and reflected electron beam, it can correspond the first maximum diffraction pattern. Incidence electron wave vector 20.

(24) λ to make up. is a constant,. is an incidence electron momentum, wave vector. is. relate to From. cancel. , we can get. is the direction which is parallel to the materials. Like this picture.. From eq. vector.. , with different. , we can get the difference of the parallel wave. Used 1D diffraction theory to extend to 2D. By the same reason:. : In 2D space, the angle of incidence and reflected electron beam. 21.

(25) : The lattice constant in 2D space Actuality we measure LEED, it is 2D space, if we needed diffraction pattern, then eq.(8),(9) are both satisfied. Used 2D reciprocal lattice vector independent 1D space.. , it to regard as the combination of two. Δ. Real space and K-space (reciprocal lattice space) transform type.(1D space). is the crystal lattice in real space,. is the crystal lattice in K-space.. From Δ Δ. ′. : Wave vector of incidence electron beam ′. : Wave vector of relax electron beam. The diffraction pattern needed Elastic Collision electrons, so the wave vector of incidence is equal to reflected. ′. The structure of Ewald sphere: Ewald sphere is using the value of the incidence electron wave vector for radius. The atoms in the sample surface are a 2D real space, corresponding to K-space is reciprocal lattice column. Like the Fig 3.4.. Fig 3.4: Ewald Sphere 22.

(26) When reciprocal lattice column and Ewald sphere intersect a point, this point occur diffraction pattern. In this point, the electron is perfect elastic collision, so incidence electron wave vector are equal to reflex.. Device. Fig 3.5: LEED device When the incidence electron beam is perpendicular to the sample surface, electrons are collide with the surface atoms, and reflex everywhere. In all reflected atoms, some atoms are elastic collision, some are not. By used LEED measurement, only elastic collision atoms are useful. So the device has to filter out unwanted electron signal, and keep the signal of elastic collision atoms. E-gun energy is from 0eV~750eV, when we measure LEED, the sample must to connect with ground. It can avoid sample surface charge by huge electrons. The reflex atoms will go through 3 network structure, and arrive to screen. The device structure: Grid 1 and Grid 3 are grounding; Grid 2 connects to retarding voltage, which is kept in negative voltage . Retarding voltage is used to filter unnecessary electrons. To ensure elastic collision electrons can arrive to the screen, we give a positive high voltage (5keV) to screen. Let electrons to accelerate and collides screen to arouse fluorescence. To analyze the spot position, we can know the structure of the sample. 23.

(27) 3.2.2 I/V-LEED By used LEED, we only get a plane lattice constant, if we want to know vertical plane’s interlayer distance. It can obtain by changing the incident electron energy. First, we rotate sample for a small angle ˚ ,in order to observe (00)-beam. We change the Incident electron energy by linear increase, and read the intensity of the brightness, so we can get the curve, x-axis is beam energy (eV) and y-axis is (00)-beam intensity. By Bragg's law:. : The vertical plane of the interlayer distance : Incident electron energy : The electron’s ionization energy. Take the peak of energy intensity curve,. for x-axis,. for y-axis, after. calculate the slope, we can get the vertical plane interlayer distance.. Fig 3.6: I/V LEED measurement. 24.

(28) 3.3 Auger Electron Spectroscopy (AES) Auger electron spectroscopy is a popular analytical technique in the study of surface science. It’s because that different atoms has different energy of an Auger electron. The Auger effect is an electronic process, at the heart of AES resulting from the inter and intrastate transitions of electrons in an excited atom. When an atom is probed by an external mechanism, such as a photon or a beam of electrons with high energies about 3KeV, a core state electron can be removed leaving behind a hole. As this is an unstable state, the core hole can be filled by an outer shell electron, by the electron moving to the lower energy level loses an amount of energy equal to the difference in orbital energies. The transition energy can be coupled to a second outer shell electron which will be emitted from the atom if the transferred energy is larger than the orbital binding energy. An emitted electron will have a kinetic energy of:. EK, EL1, EL2.3 are respectively the core level, first outer shell, and second outer shell electron energies.. Fig 3.7: Auger Electron 25.

(29) Fig 3.8: Connected Line when measure AES In the Fig 3.8, it is our AES circuit diagram at NTNU. Blue lines are output signal, and green lines are input signal. First, we used PC control NI card to output a negative electric potential to CMA analyzer controller. This signal was transfer to a high-voltage module and amplified by a high electric potential. In the same times, Lock-in sends out a sine wave with an oscillation 5V, frequency 10KHz to controller. Inside the controller, it has an isolation transformer, it can combine AC and DC signal and transfer to OC. Simultaneously analyzer controller gave a positive high-voltage to Collector port, in order to let collector receive electron much easier. We also want to get AC signal, so we make a high pass filter after Collector port. High pass filter can filter DC high-voltage and let small AC signal pass through, and then protect Pre-amplifier, Lock-in …etc.. 26.

(30) Small AC signal goes through a pre-amplifier, Lock-in amplifier, then to NI-card and at last we using VB to have Auger spectrum.. Fig 3.9: DPCMA. DPCMA is a device to select Auger electron. In AES system, E-gun ejects an electron beam to the sample. Lots of electrons are stimulated and a part of electrons fly into the concentric column. This two concentric columns can select specific energy electrons. The principle is very simple, when an electron goes through a parallel electroplate, the electron’s orbit is a parabolic. We fix the electric field and the incidence angle, than we can fix the range. By the principle, if we change the OC and IC electric field, we can choose the specific energy electron. Because of the collector has higher voltage than OC, so the specific energy electrons go through the focus and channeltron (like amplifier) to the collector. 27.

(31) 3.4 Magneto-Optical Kerr Effect (MOKE). Fig 3.10: MOKE We used a linear polarization laser beam, for MOKE measurement. When it in a ferromagnetic material with an external magnetic field. Because the left-handed and right-handed light has different reflection coefficients, the reflected light induced a phase difference. It will become elliptical polarized. This is called Magneto-Optical Kerr Effect. Surface Magneto-Optical Kerr Effect (SMOKE) and Magneto-Optical Kerr Effect (MOKE) have same principle. The angle of original linear polarized light and the long axis of elliptically polarized light is called Kerr Rotation angle , the ellipticity of the elliptically polarized light is called Kerr elliptically . and are all small (<<1˚), and proportional to magnetization (M). MOKE measurements can be classified into three types: Longitudinal (In-Plane), Polar (Perpendicular) and Transverse. Which are depending on the direction of the magnetization vector (M) with respect to the plane of incidence and the sample surface. Longitudinal (In-Plane) MOKE The magnetization vector is parallel to both the reflection surface and the plane of incidence. Polar (Perpendicular) MOKE The magnetization vector is perpendicular to the reflection surface and parallel to the plane of incidence. 28.

(32) Transverse MOKE The magnetization is perpendicular to the plane of incidence and parallel to the surface. Fig 3.11: Longitudinal (In-Plane), Polar (Perpendicular) and Transverse MOKE Linear polarized light is composed of Left-circularly polarized light and Right-circularly polarized light. Let the dielectric of the magnetic material produce non-diagonal number. When different direction polarized light Incident to ferromagnetic materials, it has different refraction coefficients (nr is the Index of refraction of right-handed polarized light, nl is the Index of refraction of left-handed polarized light). Because of the refraction coefficient are different, the two kinds polarized light have different velocity, and produce phase difference. The absorption coefficient is different, so the amplitude is also different. By the two reasons, it produces elliptical polarization. By mathematical formulas:. let. ,. is wave number, and discussion the relationship of. 29. and.

(33) By the above equation. By Fresnel eq. can know the reflection coefficient. In the eq.. and. :. are all complex, so we can know that L-circularly. polarized light and R-circularly polarized light combined with elliptically polarized. And last, we can get the Kerr Rotation angle and Kerr Elliptically :. Generally, dielectric. , so take dielectric replace refractive index:. Because is a linear function with M, so magnetization (M).. 30. and. also direct ratio with.

(34) Fig 3.12: Connected Line when measure MOKE. Fig 3.13 Air MOKE system 31.

(35) Fig 3.14: Absorption MOKE Chamber and The direction of external magnetic field. 32.

(36) 3.5 Scanning Tunneling Microscopy (STM) Scanning tunneling microscope (STM) is an instrument for imaging surfaces at the atomic level. It’s based on the quantum tunneling current to investigate the local morphology of material surface. Using a sharp tip to scan the surface and obtains two-dimensional (2D) images. Tunneling effect is arises from quantum mechanics. In classical physics, an object hitting an impenetrable barrier will not pass through. In contrast, objects with a very small mass, such as the electron, have wavelike characteristics which permit such an event, referred to as tunneling. Electrons behave as beams of energy, and in the presence of a potential U(z), assuming 1-dimensional case, the energy levels ψn(z) of the electrons are given by solutions to Schrödinger’s equation:. ħ is the reduced Planck’s constant, z is the position, and m is the mass of an electron. If an electron of energy E is incident upon an energy barrier of height U(z), the electron wave function is a traveling wave solution,. , if E > U(z), which is true for a wave function inside the tip or inside the sample. Inside a barrier, E < U(z) so the wave functions which satisfy this are decaying waves,. , Knowing the wave function allows one to calculate the probability density for that electron to be found at some location. In the case of tunneling, the tip and sample wave functions overlap such that when under a bias, there is some finite probability to find the electron in the barrier region and even on the other side of the barrier. Let us assume the bias is V and the barrier width is W. This probability, P, that an electron at z =0 (left edge of barrier) can be found at z =W (right edge of barrier) is proportional to the wave function squared,. If the bias is small, let U − E ≈ φ, φ is the work function, gives the minimum energy needed to bring an electron from an occupied level, the highest of 33.

(37) which is at the Fermi level. When a small bias V is applied to the system, only electronic states near the Fermi level are excited. These excited electrons can tunnel across the barrier. In other words, tunneling effect occurs when the electrons energies are near the Fermi level. When V is positive, electrons in the tip tunnel into empty states in the sample; for a negative bias, electrons tunnel out of occupied states in the sample into the tip. Tunneling current is given by. Fig 3.15: STM. 34.

(38) 3.6 X-Ray Diffraction (XRD) X-ray scattering technique is a non-destructive analytical technique to study the crystal structure, chemical composition, and physical properties of materials and thin films. These techniques are based on observing the scattered intensity of an X-ray beam hitting a sample as a function of incident and scattered angle, polarization, and wavelength or energy. One of the X-ray scattering techniques is called XRD. X-ray powder diffraction (XRD), is an instrumental technique that is used to identify minerals, as well as other crystalline materials.. Fig 3.16: X-Ray Diffraction Incident X-Rays are diffracted by the layers of atoms in a crystalline material.. If an X-ray beam incidence to a crystal lattice, general scattering occurs. Although most scattering interferes is eliminated, diffraction occurs when scattering in a certain direction is in phase with scattered rays from other atomic planes. Under this condition the reflections combine to form new enhanced wave fronts that mutually reinforce each other. The relation by which 35.

(39) diffraction occurs is known as the Bragg law or equation. Because each crystalline material has a characteristic atomic structure, it will diffract X-rays in a unique characteristic pattern.. Basic concept of an X-Ray Diffractometer. Fig 3.17: X-Ray Diffractometer. 36.

(40) 3.7 Extended X-ray Absorption Fine Structure (EXAFS) During X-ray absorption, X-ray photon incident with energy E, it can stimulate the inner atomic orbital to unfilled high energy states orbital. When the kinetic energy of the photoelectron (EXAFS range is 40eV~1000eV) are much bigger than the atomic interaction (3eV), photoelectron used absorption atomic for center and spread out with spherical wave. The wavelength is , from de-Broglie relation and law of conservation of energy, we can know that:. The needs energy for the inner atomic orbital to unfilled high energy states orbital. The quality of electron When the kinetic energy of the photoelectron is much bigger than the atomic interaction, the absorption atoms are scattering by neighbor atoms. It produce backscattering wave, resulting in positive interference and negative interference to each other. An EXAFS function χ(k) can be defined:. : The absorption coefficient of isolated atoms, it also called background absorption. When analyze EXAFS, we used some relationship and coefficient like this:. : path multiplicity : nominal path length. : change in half path length 37.

(41) : effective scattering amplitude : effective scattering phase shift : passive electron reduction : mean free path : Debye-Waller factor Debye-Waller factor plays an important role in EXAFS, generally speaking, it compose of Static Disorder and Thermal Vibration .. Fig 3.18: EXAFS. 38.

(42) Chapter4 Experiment and Result Part1. Uniaxial Magnetic Anisotropy in Pd/Fe/Al2O3(0001) by oblique deposition. 4.1 Preparing Pd/Fe(0,45,65-deg)/Al2O3(0001) All the samples were prepared in a UHV chamber with a base pressure of torr, by using e-beam evaporation method. After the samples were well prepared, we took it out and measured MOKE, XRD and EXAFS. The magnetic behavior could survive in the air more than a week. Sapphire substrates (3*5*0.3 mm) with (0001) orientation was prepared by vibration washes (used ethyl alcohol 95%) and baking at 400K for 12 hours in vacuum. The Pd and Fe thin films were deposited at room temperature (RT) by e-beam heated thermal evaporation. We deposited Fe on Al 2O3 (0001) than capping Pd films on the top surface to protect the Fe films. We fixed the Pd thickness and only used normal deposition. We changed two parameters in our system, one was Fe thickness (10 to 150 ML) and the other was deposition angle (0, 45, 65-degree). The deposition rate was calibrated from the epitaxial growth on Cu (100) using Auger electron spectroscopy. Thus, 1 ML is equivalent to the nominal surface atom density of. .. Fig 4.1: Sample prepared by using oblique deposition. 39.

(43) 4.2 Magnetic optical Kerr effect (MOKE)-Magnetic behavior Magnetic hysteresis loops are measured by MOKE with azimuthal rotation of 360˚. In this study, the magnetic anisotropy was obvious when obliquely depositing Fe films. We measured MOKE angle dependent. So we can clearly know that the normalized remanence and magnetic coercivity corresponding to the azimuthal angle.. Fig 4.2: Azimuth rotation (360˚) measure MOKE In the next three pages, we show three deposition angle (0,45,65˚) with various Fe thickness. Each sample we measured from 0˚~170˚ with azimuthal angle Φ in a step of 10˚. In normal deposition, the results are separated into two parts for discussion, Fe thickness up from 25 ML and below to 25 ML. Even though by normal deposition, we still see magnetic anisotropy in 10 and 20 ML. But when we increase Fe thickness up to 25 ML, it turns to isotropy. By oblique deposition for 45 and 65-deg, we measured clear anisotropy magnetic behavior even after capping Pd films.. 40.

(44) Fig 4.3: Pd/Fe-0˚/Al2O3(0001). 41.

(45) Fig 4.4: Pd/Fe-45˚/Al2O3(0001). 42.

(46) Fig 4.5: Pd/Fe-65˚/Al2O3(0001). 43.

(47) 4.3 X-ray Diffraction (XRD) – Crystalline structure. Fig 4.6: XRD data for Pd/ Fe-0˚、45˚、65˚/ Al2O3(0001) For the characterization of crystalline structure, an X-ray diffractometer with a Cu target was used to obtain the Bragg reflections from the samples. Fig 4.6 shows the X-ray diffraction patterns with three different deposit angles. The Fe films were deposited with the oblique angle of 0˚, 45˚and 65˚. XRD show the large main peak of Al 2O3 (0006) at 2θ≈41.6˚, which confirms the highly oriented crystalline structure of the sapphire substrate. As indicated by the orange dashed line of the square box, the peaks of Fe(110) and Fe(200) are observed at 2θ≈ 44.6˚ and 2θ≈ 64.4˚. For 0-deg, the Fe(110) peak is especially pronounced and much larger than the Fe(200) peak, indicating that body-centered-cubic (110) is the preferred orientation for normal-deposited Fe on Al2O3(0001). For 45-deg, the peaks of Fe(110) and Fe(200) are still observable and of similar intensity, suggesting the coexistence of both orientations. For 65-deg, only the peak of Fe(110) is observed and the Fe(200) peak disappears. This also shows the preference for bcc(110) orientation in 65˚-Fe. The above observations indicate two facts. One is that the (110) orientation is preferred in the Fe films grown on Al 2O3(0001), especially for 0˚ 44.

(48) and 65˚-Fe. The other one is that the off-normal 65˚grazing incidence growth will leads to the disappearance of the (100) texture of Fe. Roughly speaking, the grazing-incidence growth leads to less pronounced XRD peaks of Fe, more disordering in crystalline structure. Besides, the peak of Pd(111) appear at 2θ ≈ 20.0˚, showing the preference for Pd(111) texture in the Pd polycrystalline layer. The combination of metals with oxide substrates is of crucial importance for the understanding of various phenomena related to composite materials, and heterogeneous catalysts. Recently, a substantial amount of experimental and theoretical interest has focused on the Al 2O3(0001) surface. The literatures agree on that the most stable (0001) surface of Al 2O3 is terminated by a single Al layer. This surface undergoes large relaxations, where the surface aluminum moves into the bulk, ending up almost coplanar with the oxygen.. Fig 4.7: (a) Al2O3(0001) structure, Fe(110) structure (b) Fe(110) on Al2O3(0001). 45.

(49) Fig. 4.7 (b) exhibits the crystalline structures of Al 2O3(0001) and bcc Fe(110) from the top view. The hollow circles and gray circles indicate the first layer of aluminum and oxygen atoms, respectively. Due to the surface relaxations, the oxygen atoms do not form a perfect hexagonal structure. Instead, the surface atomic structure of oxygen is composed of a rectangular and a parallelogram lattice cell, as indicated in the upper panel of Fig.4.7 (b).The Fe(110) surface crystalline is also composed of a rectangular lattice cell. Although the lattice mismatch along the long edge direction is much larger, ≈ 17%, the ratio of the two lattice lengths is close to the ratio of two integers 6/5. This means that 5 cells of Al2O3(0001) can match 6 cells of Fe(110) well. One can see from Fig.4.7 (b) that the crystalline structures of Fe(110) do not perfectly match Al2O3(0001). But it is reasonable to expect some short range ordered Fe(110) structure on Al2O3(0001), due to the similar lattice structure. Actually, many studies of epitaxial Al 2O3(0001) layers on Fe(110) also supports this suggestion. The lower panel shows the three possible orientations of Fe(110) grown on Al2O3(0001) surface, which is consistent with the 3-fold symmetry observed in our STM picture.. Fig 4.8: Corresponding to the STM images. 46.

(50) 4.4 Extended X-ray Absorption Fine Structure (EXAFS) - Crystalline structure. Fig 4.9: (a) Original data (b) After Fourier Transfer -EXAFS. Table1: Nearest neighbor distance and Debye-Waller factor In order to know the clear crystalline structure, we also performed extended X-ray absorption fine structure (EXAFS) measurement, for the local atomic geometry of obliquely-deposited Fe. Fig.4.9 (a) shows the Fe K-edge EXAFS oscillation k2-weighted extracted EXAFS data. The EXAFS spectra were recorded at room temperature in normal incidence with the polarization parallel to the 1-dimensional Fe ripples. Fig.4.9 (b) represents the Fe K-edge Fouries transform (FT) amplitudes of the EXAFS k2χ data for the Fe films. For the oblique angles of 0˚, 45˚ and 65˚, the first peaks in FT spectra appear to have roughly the same location, though they have different heights and full widths at the half maximum. Besides of the 47.

(51) normal incidence geometry with polarization parallel to Fe ripples, we also performed EXAFS measurement with other geometry: "gracing incidence + polarization perpendicular to surface normal "and "normal incidence + polarization perpendicular to Fe ripples". These measurements reveal similar results to Fig. 4.9(b). The first FT peak positions are invariant, but the height of peak significantly decreases with the increasing of oblique deposition angle. This also suggests that the local crystalline structure of Fe becomes more disordered, when the Fe layer is deposited with a larger oblique angle. For a quantitative comparison, fitting of theory to the EXAFS data shown in Fig. 4.9(b) was carried out. The coordinating number of Fe was fixed at 8 in fitting a model compound to the experimental EXAFS. From the fitting, the nearest neighbor distance around Fe (R), and the mean square relative displacement (σ2) for each NN bound are deduced, as listed in Table 1. For the different oblique angels, the nearest neighbor distance around Fe (R) is nearly the same: ≈ 2.46±0.01Å, which is consistent with the body-centered-cubic Fe structure. On the other hand, the σ2 significantly increases with the oblique growth angle, from 8.9±0.5 (*10-3Å 2) for 0˚ to 14.5±0.5 (*10-3Å 2) for 65˚. Apparently the large oblique deposition angle not only induces the 1-dimensional Fe surface nanostructures, but also leads to the disordering of crystalline structures. The stacking of grazing-incident atoms are getting more random with increasing oblique angle, leading to the larger variation of the mean square relative displacement (σ2) for each NN bound.. Fig 4.10: bcc structure (Fe) Nearest neighbor distance. 48.

(52) 4.5 Scanning Tunneling Microscopy (STM) – Surface Morphology. Fig 4.11: STM images 40 ML Pd/100 ML Fe-0˚、70 ML Fe-45˚、65˚/ Al2O3(0001) In previous studies, the shadowing effect induced by oblique deposition will change the surface morphology. This picture shows the STM images of Pd/Fe bilayers grown on Al 2O3 (0001) with different oblique deposition angle of Fe: 0˚, 45˚ and 65˚. The STM images reveal the surface morphology of Pd capping layer on the oblique-deposited Fe films. Because the Pd layer was deposited in the surface normal direction, it is interesting to know if the Fe oblique deposition induced 1D nanostructures are still observable even after capping of Pd layer. In Fig 4.11 (a), the morphology of Pd on 0˚-Fe is composed of nano stripes. The width of the stripes is 5±2 nm, and the length ranges several tens nanometer. The surface corrugation is within ±3 nm. In the magnified STM image, the orientations of the stripes clearly reveal the three-fold symmetry, which might originate from the hexagonal crystalline structure of the substrate Al2O3(0001). In picture (b) and (c), the morphology of 40 ML Pd on 70 ML 45˚ and 65˚-deposited Fe is composed of randomly distributed clusters. 49.

(53) The cluster size ranges 10±3 nm, and the surface roughness is within ±5nm. From the comparison of Fig 4.11 (a)-(c), we can conclude that the 40 ML Pd capping layer fully covered the 1-dimensional Fe nanostructures for the cases of 45˚and 65˚-Fe. However, the Pd capping layer can still preserve the 3-fold symmetry from the 0˚-Fe/Al2O3 (0001) films. Fig 4.11(d) shows another example of Pd/45˚-Fe bilayer, in which the Pd capping layer is much thinner than those shown in Fig. 4.11(a)-(c). This only 5 ML thick Pd overlayer can still fully cover the 1D nanostructure of the Fe under layer, because the uniaxial shape is unobservable in the morphology. The surface morphology is composed of nano-sized clusters, like Fig. 4.11(b),but with a much smaller average size (≈5±2 nm) and a smoother surface corrugation (≤ ±2 nm). The comparison between Fig. 4.11(b) and (d) suggests that the thicker the Pd coverage is, the larger the average size of Pd surface nanoclusters becomes.. 50.

(54) 4.6 MOKE-magnetic behavior in Pd/Fe(25ml)-0,45,65/Al2O3 (0001). MOKE hysteresis loops of Pd/0˚、45°、65°-deg Fe/Al2O3(0001), measured with variation of azimuthal angle Φ in a step of 10°. WhenΦ=0°, Hc(0-deg)=60 Oe, Hc(45-deg)=250 Oe, Hc(65-deg)=400 Oe. Fig 4.12: Pd/25 ML Fe-0˚、45˚、65˚/ Al2O3(0001) with Φ=0˚~90˚ This page shows the same Fe thickness but with different deposition angle. The azimuthal angle Φ-dependent on MOKE hysteresis loops of Pd/0˚, 45˚, 65˚-deposited Fe/Al2O3(0001). By normal deposition, the hysteresis loops are of square shape and almost invariant with variation of Φ. This invariance indicates the absence of magnetic anisotropy as a function of Φ in the surface plane for 0˚-deposited Fe. it’s isotropy and the coercivity are all about 60 Oe. By oblique deposition, for 45˚-deposition, the sample shows the square loop when Φ is close to 0˚. When Φ approaches 90˚, which means that the magnetic field (H) is perpendicular to the 1-dimensional ripples of Fe, the hysteresis loops gradually become titled and the remanence decreases to almost zero. The Φ-dependent evolution of MOKE hysteresis loops show a uniaxial magnetic anisotropy induced by the oblique deposition. The magnetic easy 51.

(55) axis is along the Φ=0˚ direction, which is parallel to the 1D ripples of Fe, while the magnetic hard axis is along the Φ=90˚ direction, which is perpendicular to the Fe ripples. For 65˚-deposition, Pd/65˚-deposited Fe/Al2O3(0001) reveals the similar uniaxial anisotropy like Pd/45˚-Fe, but with a larger coercivity field (HC) in the easy axis measurement and a larger saturation field (H S) in the hard axis measurement. In the more complete study, for each oblique deposition angle (0˚, 45˚ and 65˚), a series of Pd/Fe bilayers was prepared with the variation of Fe thickness.. 52.

(56) 4.7. Hc and Mr ration V.S. Azimuthal angle. Fig 4.13: Hc and Mr ration V.S. Azimuthal angle Picture in Fig 4.13 (a) (b) (c) is Mr/Ms. Fig 4.13 (d) (e) (f) is Hc Before to analyze all the picture, we isolate the 10ML and 20ML Fe-0˚. We discuss it first, in (a) and (d) Fe-10ML and 20ML by normal deposition still have anisotropy. It was not surprised, and it comes from substrate effect. If the Fe thickness is too thin, the growth patterns of the Fe atoms are affected by the substrate. Our basis substrate is Al 2O3(0001), Fe atoms prefer some directions. But when the Fe is thick enough (more than 25 ML), the anisotropy is not such obvious. Series of Pd/Fe bilayers was prepared with the variation of Fe thickness. This picture summarized the azimuthal angle Φ-dependent MOKE measurement. The ratio of remanence (MR) to saturation (MS), and coercivity field (HC) are plotted as a function of Φ. In (a) and (d), 25-60 ML 0˚-Fe shows the isotropic magnetic behavior. The coercivity (H C) ranges 80-100 Oe and is invariant with azimuthal Φ. The ratio of MR/MS is always close to 1, expressing the square of hysteresis loops. When the oblique deposition angle is increased to 45˚, as shown in Fig 4.13 (b), the HC measured along the easy axis gradually increases from ~100 Oe to ~380 Oe with Fe film thickness increased 53.

(57) from 10 ML to 68 ML. Similar to Fig 4.12, for each sample of 45˚-Fe, the hysteresis loop gradually become tilted with the increasing azimuthal Φ. The HC does not change seriously, until Φ=70˚. Because of the tilting shape of hysteresis loop the H C drops to almost zero in hard axis (Φ=90˚). Unlike the insignificant change of HC during Φ=0˚-70˚, the ratio of MR/MS starts to decrease right after the deviating from easy axis (Φ=0˚). In the hard axis (Φ=90˚), Mr drops to nearly zero, indicating that no magnetic moment can be stabilized in the direction. For Pd/65˚-Fe/Al2O3(0001), similar to the case of 45˚-Fe, the drop of HC and MR/MS ratio to the minimum (Φ=90˚) clearly indicates the presence of an unaxial anisotropy. As shown in Fig. 4.13 (c) and (f), both HC and MR/MS ratio of Pd/20 ML 65˚-Fe descends more quickly with Φ rotation than that of the other samples do. This implies the uniqueness of 65˚-Fe around the thickness region of 20 ML.. 54.

(58) 4.8 Hc in easy axis V.S. Fe thickness. Fig 4.14: Hc in easy axis V.S. Fe thickness The summarized phase diagram in Fig 4.14 also exhibits the distinctive evolution of HC displayed by 65˚-Fe. The coercivity (HC) of 65˚-Fe gradually increases with Fe film thickness and reaches the maximum HC of ~450 Oe at 20-40 ML. Afterward, the HC gradually decreases to ~300 Oe and becomes nearly invariant. Similar tendency of thickness-dependent HC evolution have been observed in other systems. The increase and then decrease of HC as a function of Fe film thickness might be related to the formation of the 1-dimensional ripples. Before 20 ML, the ripples become more and more steep with increasing Fe thickness. With the higher coverage of more than 40 ML, the Fe ripples start to merge with nearest ones, which thus cause the decrease and saturates of HC. On the other hand, the coercivity of 45˚-Fe monotonously increases with Fe film thickness and gradually saturates at HC ~ 350 Oe after 60 ML. In the case of 0˚-Fe, the HC ranges 70±10 Oe and is nearly invariant with the change of Fe film thickness. Besides of Pd/0˚, 45˚, 65˚-Fe bilayers, the magnetic properties of Pd/45˚-Fe/Pd trilayers are investigated for comparison. For the trilayers, both the bottom and top layer of Pd are deposited in the surface normal direction. The thickness-dependent evolution of HC is shown in Fig 4.14, as indicated by open squares. Interestingly, the inserting of a Pd buffer layer between 45-deposited Fe layer and Al 2O3 (0001) destroys the 55.

(59) uniaxial MAE, leading to a relative small value of H C ≈ 70±10 Oe, which is similar to the HC of Pd/0˚-Fe bilayers. This implies that the oblique deposition of Fe can induce a uniaxial MAE only on a smooth and clean Al 2O3(0001) surface. The presence of a Pd buffer layer may already generate a relative rough surface, which can destroy the shadowing effect during the grazing-incidence growth of Fe, as well as the induced uniaxial MAE.. Fig 4.15: Pd/Fe-65˚ V.S. Fe thickness and discussion. J.L.Bubendorff et al, Europhys. Lett., 75(1),pp.119(2006) and Surface Science 603 (2009) 373-379. 56.

(60) 4.9 Ku in Hard axis V.S. Fe thickness. Fig 4.16: Magnetic Anisotropy Energy in hard axis V.S. Fe thickness In Fig 4.16, we show the summarized phase diagram of saturation field (H S) measured in hard axis (left axis), and the corresponding uniaxial MAE (right axis). By neglecting the crystalline MAE in Neel's model of surface anisotropy, the uniaxial MAE (Ku) is deduced from the HS:. ,. this eq. come from:. │. │. M is the magnetic moment of Fe (2.5 µ B/atom). Usually in previous studies, both the crystalline MAE and shape-induced uniaxial MAE should be taken into consideration. Nevertheless the absence of minor loops in Fig 4.12 signifies the insignificance of crystalline MAE in this Pd/Fe/Al 2O3(0001) system. Although the crystalline MAE of bulk Fe(3.4 µeV/atom) is of the same order of magnitude as the uniaxial MAE shown in Fig 4.16, the disordering of crystalline 57.

(61) structure in the grazing-incidence grown Fe films randomizes the orientation of crystalline MAE and thus reduces its total contribution. Eventually the shape induced uniaxial MAE dominates the magnetic behavior of the Pd/Fe bilayers.. Fig 4.17 (a) The path of hysteresis loop (b) Oblique deposition produces shape anisotropy in our system. In order to discuss the shape anisotropy, we define the magnetostatic energy first. Assume a material is magnetized in an external magnetic field, like A point in Fig 4.17 (a), than remove the external magnetic field, the magnetization of the material will decrease to point C. In the same time, it will effect by demagnetize field Hd,. means the line of the demagnetize field, and the slpoe is. ,. is demagnetization factor, the yellow triangle Δ OCD is the stored energy of the material. This area, generally called magnetostatic energy, the energy equation is:. M is the magnetization in C point, transfer to vector representation:. Direction of demagnetization field is opposite to M.. Now we take a long and flat shape for example, c is the length of half of the long axis, b is half of the short axis. The angle of magnetization and c axis is θ, then. and. are demagnetization factor along c axis and a axis. 58.

(62) is uniaxial crystalline anisotropy energy, and also be the easy axis. is shape anisotropy, like this eq.. By this eq. if long axis equal to short axis, shape anisotropy will disappear.. 59.

(63) Chapter 5 Experiment and Result Part 2. H2 absorption effect in room temperature by MOKE measurement in Pd/Fe,Co,Ni/Al2O3(0001) system. 5.1 Preparing Pd/Fe,Co,Ni/Al2O3(0001). Fig 5.1: H2 absorption effect in n ML Pd/Fe and 60 ML Pd/Fe、Co、Ni All the samples are prepared in UHV chamber by thermal evaporation method. First we want to figure out the relation of Kerr rotation and Pd thickness. Therefore, were deposited n ML Pd on the top surface. In the experimental data, the curve of measure Kerr rotation will start to shift in Pd (10 ML). And Pd (60 ML) is the suitable thickness for the observation of H2 adsorption effect.. 5.2 To proofread the Precise Rotator. Fig 5.2: The analyzer corrected In the experiment of measuring the Kerr Rotation curve, A Precise Rotator is needed. Like Fig 5.2, the resolution of the X-axis is 0.01mm, and the center to the apex of the rotator is 35mm. 60.

(64) By using trigonometric function, we can get the corresponding rotation. It means, when we rotate X for 0.01mm, it equal rotate 0.0163˚.. 5.3 How to measure Kerr rotation curve and corresponding intensity?. In MOKE measurement, a linear polarization laser beam injects in a ferromagnetic material with an external magnetic field. By changing the external magnetic field with opposite direction, we can get two parabolic curves, as shown Fig 5.3.. Fig 5.3: Kerr rotation curve with positive and negative magnetic field Corresponding the two curves, the difference of two curves minimum is Kerr Rotation. When we fixed the analyzer angle, we can get the corresponding intensity (Δ y), like Fig 5.4. 61.

(65) Fig 5.4: Analyzer angle and corresponding Kerr intensity The minimum of the curves is called extinction angle. The physics of extinction angle comes from the magnetic optical Kerr effect (MOKE). A linear polarization light to ferromagnetic materials with an external magnetic field, the reflected light is ellipse polarization, like Fig 5.5. When we rotate the analyzer angle, we can get the corresponding intensity. When the analyzer angle is perpendicular to the long axis of the ellipse, it has the weakest optical signal and it is also the minimum intensity, we called it extinction angle.. Fig 5.5: Defined Extinction angle 62.

(66) 5.4 n ML Pd/Fe/Al2O3(0001) + H2. Fig 5.6: n ML Pd/Fe First we focus on Pd/Fe system. In many studies, Pd thickness and H2 pressure are the most important parameters. If we want to see Pd lattice expanding at RT, Pd thickness needs to be thicker than 4 nm, and the H2 pressure needs to be higher than 20 mbar. In our experiment, H2 pressure is always in 1013 mbar. Pd 3 ML and 5 ML, the Kerr rotation curve is similar after absorption H2. When we increase Pd thickness to 10 ML, two curves start to shift. 60 ML Pd /Fe has the obvious phenomenon of curve shift, so we take this sample do some analysis. The blue circle is measured in a vacuum of 10-3 mbar. The red triangle is measured after exposure to 1 atm H2, and the green square is in air. Take vacuum curve for basis, air curve has less than 0.02˚ shift, and H2 curve has 0.6˚ shift. It all reasonable, it means H2 changing something by 63.

(67) absorption H2. Even in the air, it also has a small H2, it still possibly to absorption in Pd thin films. In the series of n ML Pd/Fe curve, we need to pay attention to two observable features. One is the extinction angle shift, the other is the curvature change after absorption H2.. Fig 5.7: Absorption H2 cause extinction angle shift and change. Fig 5.8: Physical meaning about MOKE intensity Z. Q. Qiu and S. D. Bader, Rev. Sci. Instrum., Vol. 71, No.3, March 2000 [19] The physical meaning of the parabolic and extinction angle comes from magnetic optical Kerr effect. [19] Consider linear p-polarized light reflected from a sample surface. If the sample is nonmagnetic, the reflected light is purely p polarized. If the sample is ferromagnetic then the reflection beam should consist of an s component in addition to the dominant p component , with being the Kerr rotation. Therefore, measuring this s component will be the goal of the experimental setup. Experimentally, the measurement of the s component could be realized by 64.

(68) placing a linear polarizer in front of the photodetector to eliminate the p component. However, this measurement geometry has the following disadvantage. First, since the photodetector measures the light intensity , the measured quantity is proportional to the square of the magnetization. Second, it is difficult to quantify the absolute value of the Kerr rotation. This disadvantage can be circumvented by setting the polarizer at a small angle from the p axis. In this way, the intensity measured by the photodetector after the polarizer is. ,. [19]. We fitting the curve by eq.. Fig 5.9: Fitting parabolic, Pd thickness V.S. extinction angle shift Right axis is the extinction angle shift =. Left axis is the curvature change after absorption H2= 65.

(69) We take 60 ML Pd/Fe to analysis, some studies show the Kerr Intensity will increase by absorption H2. By measure Kerr Rotation Curve, we can know the intensity change not only in the specific angle but also in the all polarization angle in the range of Kerr rotation curve. By using this eq:. Fig 5.10: Kerr rotation curve analysis. 66.

(70) Fig 5.11: Discussion the intensity increase for Pd thickness and analyzer angle In Fig 5.11, When Pd is thicker enough, up to 30 ML. The enhancement of Kerr intensity and Pd thickness is a positive correlation. It increase about 10%、 20%、and 40% for Pd thickness 30 ML、45 ML、60 ML respectivity.. 67.

(71) Fig 5.12: Pd thickness V.S. Increase Intensity with different analyzer angle In order to discussion the relation Pd thickness, intensity and analyzer angle. We defined the extinction angle in vacuum is 0˚.. Fig 5.13: We defined the extinction angle in vacuum is 0˚ And take X axis for Pd thickness, Y axis for increase intensity after absorption H2, like Fig 5.13. If we want to see the intensity enhancement after absorption H2. Pd thickness needed thick enough (>30 ML), and the analyzer angle were positive shift about 2~3˚.. 68.

(72) Fig 5.14: (60 ML Pd/Fe) (a) H2 absorption and (b) desorption V.S. time Now we know that in Pd 60 ML/Fe system, with positive angle shift 3˚, under 1atm H2 absorption pressure, Kerr intensity can increase ~40%. But how long will it takes when H2 has been absorbed by Pd to achieve saturation? In Fig 5.14, X axis is the absorption and desorption time. Left Y axis is the intensity increase (H2/Vac.) after absorption H2, it corresponding to the blue circle and blue line. In H2 pressure about 1atm, it can saturate less than 20 min. Right axis is the intensity decrease after saturate sample pump down to vacuum, it corresponding to the red triangle and red line. From red line, we can know that it much difficult to desorption H 2 out, it need more than 500 min.. 69.

(73) 5.5 Pd/Fe,Co,Ni/Al2O3(0001) + H2. Fig 5.15: Corresponding 60ML Pd/Fe, Pd/Co, Pd/Ni In this experiment, we combine three kinds of magnetic materials, Fe, Co, and Ni and fix the top Pd films for 60 ML. And report on the reversible change of magneto optical Kerr effect (MOKE) in Pd covered magnetic thin films by controlling H2 absorption. After H2 exposure, the extinction angle of MOKE was shifted and correspondingly Kerr signal was significantly changed. When properly choosing the analyzer angle, we can obtain large enhancement of 40%, 35% and 60% for Pd/Fe, Pd/Co and Pd/Ni bilayers, respectively.. 70.

(74) Fig 5.16: H2 pressure V.S. Intensity with Pd/Fe,Co,Ni system As the before experiment, we know that absorption H 2 pressure for 1 atm, Kerr intensity can increase 35%~60% for different magnetic materials with capping Pd thin films. But it has something question, does less H2 pressure still produce the same intensity increase effect? In this picture, it shows the Kerr intensity as a function of H2 pressure, which is normalized by the initial value measured before hydrogen exposure. For Pd/Ni, the Kerr intensity drastically increased with H2 exposure and saturated at ~50 mbar. Pd/Co and Pd/Fe saturated at ~80 and ~110 mbar respectively. Afterward the Kerr intensity always keeps invariant. In previous studies, the transmittance, reflectance, and refractive index of Pd are drastically changed by H2 absorption. The variations also saturate with H2 pressure 100 mbar, implying the similar mechanism in the MO enhancement.. 71.

(75) Fig 5.17: H2 pressure V.S. absorption time in Pd/Co system In order to know the relationship of H2 pressure and saturation time, we do the two experiments. In Fig 5.17(a), H2 pressure is 1013 mbar and the saturation time is less than 10 min. In Fig 5.17(b), we absorption H2 for 20 mbar first, in Fig 5.17 (b), the saturation time is more than 60 min and the increase is only about 25%, by this we know that H2 pressure is an important factor, it has influence the accuracy for measure MOKE.. 72.

(76) Fig 5.18: (a) Pd/Fe, Pd/Co, Pd/Ni: H2 absorption V.S. time (b) Pd/Fe, Pd/Co, Pd/Ni: H2 desorption V.S. time Fig 5.18 (b) shows how the Pd covered magnetic layers recover the original value of Kerr intensity after fully hydrogenated. The sample was firstly saturated by 1 atm H2, and then the MOKE chamber was pump down to vacuum for the investigation of recovery rate. The arrows in (b) indicate the time constant when Kerr intensity recovery 80% of the H2-induced enhancement. The 80% recovery took 30 min for Pd/Ni, and even longer (~175 min) for Pd/Co, and Pd/Fe. Because Pd/Ni has the characteristic of quickly desorption H2, we do the repeated experiment for absorption and desorption H 2, than we can get the reversibility result.. Fig 5.19: Reversibility effect in Pd/Ni system. 73.

(77) Chapter 6 Discussion and Questions: Q1. How to define the thickness (ML) ? The deposition rate was calibrated from the epitaxial growth on Cu (100) using Auger electron spectroscopy. Thus, 1 ML is equivalent to the nominal surface atom density of . By using AES, we know the Fe-gun deposition rate on Cu (100). We use the same parameter to deposit Fe on Al 2O3 (0001), because the substrates are different, it has some error when define Fe thickness.. Q2. In page 41, why the hysteresis loop is oblique when 10 ML and 20 ML Fe by normal deposition in Φ=110°? Is it possibly to prefer some direction of Al2O3 (0001) ? We think it comes from the substrate effect. Al2O3 (0001) sample is only cleaned by vibration wash, maybe it had some glue on it. And by cutting of the C-Plane (0001) sapphire, it also has about 2˚ error. The two reasons above also have large influence when Fe film is too thin.. Q3. In page 44, Fig 4.6. Fe prefers (200) orientation in Fe-0˚、45˚, but why (100) peak can’t be seen in XRD pattern? The occurrence of diffraction not only has to be Bragg’s condition but also influence by crystal symmetry. The symmetry of atoms restricts some diffraction, called extinction condition. Like the table below:. 74. [20].

(78) Bravais Lattice. O (possible to produce diffraction). X(impossible to produce diffraction). bcc. h+k+l= 2n (even). h+k+l= 2n+1 (odd). fcc. h,k,l all even, or all odd. h,k,l mixed even and odd. hcp. h+k+l= 3n. h+k+l 3n. In our case, Fe is bcc structure, so (100) represent (h=1,k=0,l=0). h+k+l=1 (odd) is impossible to produce diffraction. Q4. What is the meaning of Kerr angle increase? Does it mean the increase of magnetism? By MOKE measurement, is a linear function with M, so and also direct ratio with magnetization (M).. Kerr angle increase has two possible reasons. By increasing the magnetism or decreasing n. In our study, we identified it comes from decrease n. By absorption H2, the reflectivity was decreased, and by MOKE measurement, only the intensity changed. It proves that, after absorption H 2, it only change the optical signal. Q5. Any evidences to prove the absorption of H2 by Pd? Does H2 even go through the Pd layer to Fe、Co、Ni films? From literature [21], we know that Pt can stop H2. So, we can prepare Pt/Pd/Fe and Pd/Pt/Fe sample and measure MOKE in the future.. 75.

(79) From another literature [22] , we know that Pd roughness and Pd particle size also influence H2 absorption effect. So another idea is used oblique deposition to change the Pd roughness and particle size. [23]. 40ML Pd/Fe-0°, 65°/Al2O3(0001) [23]. 76.