Photoemission Spectroscopy

Transcription

1 FY13 Experimental Physics - Auger Electron Spectroscopy Photoemission Spectroscopy Supervisor: Per Morgen <per@fysik.sdu.dk> SDU, Institute of Physics Campusvej 55 DK Odense S Ulrik Robenhagen, Irvin Wadzanayi Mangwiza, Jakob Kjelstrup-Hansen, May 23,

2 Contents 1 Photoemission Spectroscopy 4 2 Equipment X-ray Source Metal Evaporator Concentric Hemispherical Analyser Experimental Procedure XPS System Hardware Set-Up Acquisition of XPS Spectra Acquisition of Bremsstrahlung Spectra Metal Deposition on Silicon Measurements 12 5 Conclusion 15 Appendix 16 2

3 List of Figures 1.1 Photoemision of a K-shell electron Photoemission process in Ni with adsorbed atomic oxygen Mean free path of electrons in a solid Concentric Hemispherical Analyser Spectrum of Si(111) using an Mg anode Spectrum of Si(111) using an Mg anode - measured with two different systems Spectrum of Si(111) using an Mg anode (kinetic energy) Spectrum of Si(111) using an Al anode (kinetic energy) Spectrum of Si(111) using an Mg anode (binding energy) Spectrum of Si(111) using an Al anode (binding energy) Spectrum of Si(111) using an bremsstrahlung Spectrum of Si(111) (Mg anode) Spectrum of Si(111) and Ag (Mg anode)

4 Chapter 1 Photoemission Spectroscopy Photoemission spectroscopy is an experimental technique, which can be used to determine the composition of a given sample as well as to investigate the energy band structure of a solid. The underlying physical principle is the emission of photoelectrons from a solid, when it is bombarded with photons. An incoming photon can then transfer its energy to an electron in the solid, which then can be emitted as illustrated in fig. 1.1 for a K-shell electron. Figure 1.1: Photoemision of a K-shell electron Photoemission spectroscopy (PS) can be performed with X-ray photons, which is known as XPS, or with ultraviolet photons, in which case it is called UPS. In XPS the X-rays are generated using an electrode, which most often is made of either magnesium or aluminium. These sources emit photons at energies of 1253, 6 ev and 1486, 6 ev respectively. (From: w_xps/ag_xps_senior.htm.) The energy of a photon is transferred to a single electron, and the energy conservation principle then states (as shown in fig. 1.1): hν = E Kinetic + E Binding + φ (1.1) 4

5 , where hν is the photon energy, E Kinetic is the kinetic energy of the electron, when it is leaving the sample, E Binding is the energy necessary to lift the electron up to the fermi-level and φ is the work function. The emitted electrons are collected by an electron analyser, which can show either the kinetic or the binding energy. The electrons, which are due to the photoemission process are called the photoelectrons. However some of the peaks in the spectrum might be due to Auger electrons. The way to distinguish between the photoelectrons and the Auger electrons is by making two spectra at two different values of photon energy. If the spectrum being observed is showing the kinetic energy, then an increase in photon energy will lead to an increased kinetic energy of the photoelectrons. This is not the case for the Auger electrons, since their kinetic energy only depend on the internal energy levels of the sample atoms. The peaks, which have moved, will then be the photoelectron peaks. If the spectrum is showing binding energies instead, then the exact opposite is the case (the Auger peaks will move). When photons of a distinct energy are incident on the sample they will raise the energy of the electrons in the different bands inside the solid with a specific value, whereby the energy spectrum of the output electrons will reveal the structure of the bands. This is illustrated in fig. 1.2, which shows an example of a phtoemission process in nickel on which oxygen has been adsorbed. The shadowed area shows the energy levels, which are occupied (up to the Fermi level). Figure 1.2: Photoemission process in Ni with adsorbed atomic oxygen The mean free path of electrons is heavily dependent on their kinetic energy as shown in fig Because of this, an energy in the region around 100 ev will show how the states very close to the surface are organised. By taking advantage of this property, photoelectron spectroscopy can be used as a very surface sensitive characterization method. Some of the photoelctrons will be scattered on their way to the surface and 5

6 Figure 1.3: Mean free path of electrons in a solid will therefore be detected at a lower energy. These will then form a continous background (called secondary electrons). 6

7 Chapter 2 Equipment The system consists of a UHV chamber on which the instruments utilized are mounted. For this exercise an x-ray source, a metal evaporator and a concentric hemispherical analyser (CHA) were used. 2.1 X-ray Source The x-ray source has a Mg and an Al anode, these gives spectral lines at ev and ev respectively. To produce x-rays, electrons are accelerated from a filament at the cathode and into the anode. When the electrons hit the anode they excite the atoms which produces X-rays upon relaxation. The electrons impinging on the anode cause it to be heated. Therefore it is neccesary to cool it using water. Electrons that are not immediately stopped in the anode gives rise to bremsstrahlung, which is the background X-ray radiation and can also be used for XPS. The x-rays enters the UHV chamber through an aluminum window. 2.2 Metal Evaporator A metal evaporator consists of a small crucible containing the metal, that has to be evaporated. When the crucible is heated the evaporated metal enters the UHV chamber. To monitor the rate at which the metal deposits on the sample, it is controlled by a quartz-crystal which oscillates near 5 M Hz. The frequency at which the crystal oscillates depends on the weight of the metal deposited on its surface. It has been found that a monolayer of metal reduces the oscillation by 13 Hz. 2.3 Concentric Hemispherical Analyser The CHA is used to analyse the electrons emitted from the sample. As seen on fig. 2.1 the CHA consists of two hemispherical metal plates. The outer plate is negatively charged to repel the electrons and the inner plate is positively charged to attract the electrons, thus making the path of the electrons curve. 7

8 Due to the charge on the plates only electrons with a certain energy will be able to complete the curve, electrons with a lower energy will hit the inner plate and electrons with a higher energy will hit the outer plate. The energy needed for the electrons to make it through the hemisphere is called the pass energy. The lenses placed before the CHA enables two operating modes: Constant Retard Ration (CRR) or Constant Analysis Energy (CAE). In CRR mode the energy of the electrons is reduced by a constant ratio. This means that if electrons with an energy of 1000 ev are to be detected and the retard ratio is 10, the energy of the electrons is reduced to 100 ev and the pass energy is set to 100 ev. The CAE mode utilizes a constant pass energy. This means that if electrons, with an energy of 1000 ev, are to be detected using a pass energy of 50 ev, the energy of the electrons has to be reduced by 950 ev. Using the CRR mode constant resolving power is achieved and using the CAE mode constant energy resolution is achieved. Slits are placed on the CHA to limit the area from which the electrons are received. In order to obtain enough electrons to make a spectrum an electron multiplier is used. 8

9 Figure 2.1: Concentric Hemispherical Analyser 9

10 Chapter 3 Experimental Procedure The experiment is divided into two parts. The objective of the first part is to introduce the participants to the UHV system, X-ray source and photoelectron spectroscopy. Here XPS and bremsstrahlung spectra are made on a Si(111) surface and the results analyzed. The second part of the experiment is carried out in a separate vacuum system and it involves deposition of a metal film on silicon surfaces and then checking the result using XPS. 3.1 XPS System Hardware Set-Up The sample had already been transferred into the UHV chamber, so it was just moved to the correct position with respect to the X-Ray source and detector. The UHV was checked in the same way described in report 2 on Auger Electron Electron Spectroscopy. A light was used to illuminate the inside of the UHV system through a glass window. The spectrometer and X-Ray source was switched on. The X-Ray source was water cooled and flow could be boosted to about 4.3 L/min by use of a pump. Care was taken not to recycle the water to void making it conductive. There are two available anodes, Al and Mg. The X-Ray source was first stabilized with the Mg anode at 10 kv and 10 ma, with a filament current of 4 A. It is import to ground the sample, otherwise an offset voltage develops due to the emission of electrons. Sometimes the sample can be deliberately floated at some voltage level depending on the purpose. 3.2 Acquisition of XPS Spectra After the hardware set-up was complete, a computer program was used to control the spectrometer and create the spectra in the EDX mode. The software was set to Mg anode and to scan the kinetic energies between 100 ev and 1300 ev. The digital voltmeter (DVM) is set and the system was calibrated. After all the necessary settings are made according to the software s guideline, both kinetic energy and binding energy spectra are created. The same procedure is repeated for Al. The program SHOWXPS is used to manipulate the spectra. The peaks were identified and the spectra analyzed and compared. 10

11 3.3 Acquisition of Bremsstrahlung Spectra Acquisition of Bremsstrahlung Spectra is useful because KLL Auger lines for Si are at a higher energy that either Al or Mg X-Ray lines, and these Auger lines can not be excited by X-Ray line spectra. The Mg anode is used and the spectrometer scans between 1400 ev and 1700 ev. 3.4 Metal Deposition on Silicon The sample was transferred into the UHV chamber (by use of a magnetic coupling) and the pressure stabilized at To do this, the transfer chamber was filled with nitrogen. First the Mg source at an anode voltage of 10 kv and current of 20 ma was used together with a filament current of 5A. After calibration of the spectrometer and before deposition, XPS spectra of the sample are taken in order to make a before and after comparison. It was important to record and monitor the position of the sample so the the same point was measured. In this exercise silver is deposited on a silicon surface. The thickness of the layer to be deposited is determined by use of a 5 MHz Quartz crystal, whose frequency decreases with weight. The density of the metal, change in crystal frequency per unit weight, and the area of the deposition area are used to calibrating the crystal. This way one calculates the deposition thickness. In this system every 13 Hz corresponded to 1 monolayer (i.e 2.5 Å). It took thirty seconds to deposit one monolayer. This means 10 minutes are required to deposit the target of 50 Å (i.e 50/2.5 monolayers at 30 seconds per monolayer). The system contains four crucibles with Ag, Au, Zr and Cu, but only Ag was used for this exercise. The deposition settings used are shown in table 3.1. Parameter Setting Value Units Filament Current 3.93 A Emission Current 17 ma Flux ma Energy(as a voltage) 1.99 kv Table 3.1: Deposition Settings The deposition was carried out and this was followed by a second round of XPS spectra using the Mg anode. Correct positioning can not be overemphasized. The spectra are then analyzed and compared. 11

12 Chapter 4 Measurements The first measurement was made on a Si(111) surface using a magnesium anode (photon energy = ev ). The spectrum on fig. 4.1 shows the number of particles vs. kinetic energy. Figure 4.1: Spectrum of Si(111) using an Mg anode In the appendix the same spectrum can be seen, where four distinct peaks have been labeled. Also shown in the appendix is the spectrum obtained with an aluminium anode. This shows, that some of the peaks have shifted, while others remain at the same energy. The kinetic energies of these four peaks are compared in table 4.1 As can be seen, only the first peak remains at the same kinetic energy, when the photon energy is changed. This means, that this line corresponds to an Auger transistion in the silicon, while the other three lines are due to photoelectrons. It is also possible to do the same comparison using the binding 12

13 Energy [ev] Peak no. Mg anode Al anode 1 496,0 496, ,0 940, Table 4.1: Kinetic energies energies. The binding energy spectra are also included in the appendix. From these it can be seen, that three distinct binding energies of Si are: 115 ev, 166 ev and 546 ev. As the final activity of the first part, a spectrum was obtained using bremsstrahlung (also shown in the appendix). By using bremsstrahlung it is possible to obtain spectra at a higher X-ray energy. The KLL Auger transistion in Si is supposed to give of an electron of 1616 ev. A line can be seen in the spectrum at an energy a little lower than 1600 ev, which must be this transistion since there can be a slight offset. In the second part of the experiment, silver was deposited on a Si surface. An XPS spectrum of the sample before the deposition is shown in the appendix (Mg10kVSi111). The spectrum after deposition of a silver layer of 50 Å is shown in the appendix. Later it was discovered that the sample unfortunately was not grounded. This means, that the peaks will be shifted due to possible charging of the sample. Also the spectrometer used in this part was not calibrated. By comparing the spectrum obtained before the deposition of silver (Mg10kVSi111) with the corresponding spectrum from the first part (Si111MgSource in the appendix), it can be seen that the peaks are shifted approx. 100 ev. These two spectra are also shown in fig. 4.2 and this is used calibrate the second system. When comparing the spectra before and after the silver deposition (Mg10kvSi111 and Mg10kvSi111-Ag), it is seen, that no new peaks are present after the deposition. This is unexpected, since the spectrum from silver should be observable. An explanation to this could be, that the sample holder already had silver deposited on it prior to the first measurement, meaning that the silver peaks are already present in the initial spectrum. This of course makes the identification of the silver peaks more difficult. However the large peak around 800 ev (which actually is close to 700 ev, since the system was not calibrated) is seen to vanish when the sample is covered, indicating that this is not an Ag-peak. 13

14 Figure 4.2: Spectrum of Si(111) using an Mg anode - measured with two different systems 14

15 Chapter 5 Conclusion In this exercise X-ray photoelectron spectroscopy was performed on a Si111 surface. At first a clean Si-surface was investigated with two different photon energies (accomplished using two different anodes in the X-ray gun). This allows for the determination of which peaks are due to an Auger transition and which are true photoelectrons. An attempt was made to use bremsstrahlung to obtain a spectrum, which turned out to be a useful technique for making measurements at higher X-ray energies. In the last part, silver was deposited on a Si-surface, and XPS-spectra were taken both before and after. Unfortunately this experiment was not too succesful, since it was discovered afterwards, that the sample had not been grounded and that the sample holder had already been contaminated with silver before the measurements. Therefore these measurements did not produce any good results. 15

16 Appendix Figure 1: Spectrum of Si(111) using an Mg anode (kinetic energy) Figure 2: Spectrum of Si(111) using an Al anode (kinetic energy) 16

17 Figure 3: Spectrum of Si(111) using an Mg anode (binding energy) Figure 4: Spectrum of Si(111) using an Al anode (binding energy) 17

18 Figure 5: Spectrum of Si(111) using an bremsstrahlung Figure 6: Spectrum of Si(111) (Mg anode) 18

19 Figure 7: Spectrum of Si(111) and Ag (Mg anode) 19

20 Bibliography [1] Lüth Surfaces and Interfaces of Solid Materials 20