CV Scanning Tunneling Microscopy Studies of the Ge(111) Surfacereu/REU07/rosen.pdf ·...
Transcript of CV Scanning Tunneling Microscopy Studies of the Ge(111) Surfacereu/REU07/rosen.pdf ·...
Scanning Tunneling Microscopy Studies of the Ge(111) Surface
Anna RosenUniversity of California, Berkeley
Advisor: Dr. Shirley ChiangUniversity of California, Davis
August 24, 2007
Abstract: This research project used a Scanning Tunneling Microscope (STM) to study the Ge(111) surface. The goals of this project were to obtain an atomically clean surface, dose it withAg, and use the STM to obtain images of the phase transitions of Ag on Ge(111). However aclean surface was not obtained. The images revealed the surface was covered with contaminantsand atomic resolution was not observed. Due to experimental difficulties the sample was scannedonly four times. When a sample is maintained within the Ultra High Vacuum (UHV) system longenough to achieve an atomically clean surface the sample can then be dosed with Ag and scannedwith the STM to observe the desired phenomenon.
Introduction: The Scanning TunnelingMicroscope (STM) reveals a topographicimage of a conducting or semiconductingsurface. It is capable of resolving individualatoms. The STM uses a very sharp tungstenwire (tip) that is brought within severalÅngstroms of the sample. When a biasvoltage is applied to the sample a quantummechanical tunneling current is measured.1
The tunneling current allows the tip to sensecorrugations in the electron density of thesurface that arise from the positions of thesurface atoms. The STM is used to probe the electronicproperties of surfaces. It probes the chargedensity at the Fermi level rather than thetotal charge density. If a positive biasvoltage is applied to the sample thenelectrons tunnel from the tip to the sample,
producing an image of the charge density ofthe lowest unoccupied states. If the sampleis negatively biased, electrons tunnel fromthe sample to the tip thus producing animage of the charge density of theenergetically highest occupied electronstates on the surface.2
Figure 1. Schematic view of an STM. The tip isbrought within a few Ångstroms of the sample.Tunneling junction is blown up to atomic scale.
Sample Piezoelectric tube
VC
VC
Tip
ElectronFlow
The lateral resolution of an STM imagedepends upon the sharpness of the tip. Thebest resolution occurs when the tip isatomically sharp. Ideally the resolutionobtained is 2 Å laterally and 0.01 Åvertically. The STM can operate in one oftwo modes: constant current mode orconstant height mode. Figure 2 shows aschematic representation of each scanningmode.
(a) (b)Figure 2. Schematic views of scanning a STM tip in(a) constant current mode and (b) constant heightmode. An image consists of multiple scan lines.
In constant current mode the tip isscanned across the sample while thepiezoelectric transducer, a ceramic materialthat expands or contracts in response to achange in the applied voltage (~100 Å/V),changes the height of the tip to maintain aconstant current. The change in the appliedvoltage is controlled by a feedback loop.The distance between the tip and sample iskept nearly constant because the currentvaries exponentially with distance.1 The zposition of the tip above the surface givesthe sample topography. In constant height mode, the tip isscanned across the sample at a nearlyconstant height and the current varies. In thismode the applied voltage across thepiezoelectric transducer is held constant.This allows for a much faster scan rate of 1kHz compared to the constant current mode
scan rate of ~10 Hz.1 This method is usedfor atomically flat surfaces. The currentvariation is proportional to the sampletopography.
Figure 3. STM image of clean Ge(111) surface.Image taken by Cory Mullet. Size: 500 Å square.White spots are defects and contaminants and darkspots are vacancies. In this image you are able to seethe individual Ge atoms. Brightness of image isproportional to height of tip above sample andtherefore proportional to image topography.
Experiment: The research project involvedpreparing and scanning a Ge(111) sample.The goal was to obtain a clean sample anddose it with Ag to study the phasetransitions of Ag on Ge(111). The samplewas placed in an Ultra High Vacuum (UHV)system that contained an STM, XrayPhotoemission Spectrometer (XPS), andLow Energy Electron Microscope (LEEM)as shown in Figure 4. The cleaning of thesample took place in the XPS chamber andthe scanning took place in the STMchamber. It was necessary to place thesample in a UHV system with a pressure of~1010 – 1011 Torr to keep the sample free ofcontaminants while conducting anexperiment.
Figure 4. The UHV system in which the Ge(111)sample was prepared and scanned.
After the sample was placed in the UHVsystem the next step was to obtain anatomically clean surface. It was firstdegassed (heated) to remove adsorbedmolecules. The degassing of the sample wasdone in the XPS chamber, where the samplewas heated in situ for several hours.Degassing of the sample was only necessarywhen the sample was first introduced to theUHV system. Next, the sample underwentseveral cleaning cycles. Each cycle consistsof sputtering the sample with Ar+ ions for 15minutes and then annealing the sample for10 minutes. Ion bombardment with Ar+ ions removesmore stubborn contaminants such as carbon,oxygen, and sulfur; but it can leave thesurface disordered. The sample wassputtered with 250 eV Ar+ ions. Annealingthe sample by electron beam bombardmentreorders the surface but also causescontaminants within the bulk of the sampleto diffuse to the surface. Multiple cycles ofsputtering and annealing were required tocreate a depletion region near the surfacethat is free of contaminants. Annealing thesample reorders the surface to a c(2 x 8)crystal structure. After the sample was properly prepared itwas ready to scan. The sample was mounted
onto the STM stage and the tip was placedover the sample. The STM was operated inconstant current mode. The bias voltageapplied to the sample was –2.0 V. Thepiezoelectric transducers changes the heightof the tip with respect to the sample until atunneling current of ~0.500 nA is reached.
Results/Discussion: The sample wasscanned multiple times before attempting todose it with Ag. However the imagesobtained from the STM showed that thesample was still contaminated. Since thesample appeared dirty it was not dosed withAg because it would be difficult to identifywhich structures in the image were Ag andwhich structures were contaminants. The image resolution obtained was poor.Atomic resolution was not seen in any of theimages. Morphological features such asmonatomic steps and terraces were not seenin the majority of the images. However,large areas of contaminants were observed.
Figure 5. STM image of clean Ge(111) surface. Size: 1000 Å square. Monatomic steps are observed.White spots are contaminants.
Figure 6. STM image of clean Ge(111) surface. Size: 400 Å x 200 Å . Rows seen in image appear tobe rows of Ge atoms. Individual atoms are notresolved.
Several experimental difficulties wereencountered during this project. The Ge(111) sample is very thin (~0.5 mm thick)and cracks very easily. Throughout thisproject six samples cracked in the UHVsystem. Of these samples: two cracked whentransferring the sample from the load lock tothe STM transfer bar and three samplescracked when the sample was transferredfrom the STM transfer bar to the STM stage.To help prevent the sample from crackingwhen transferring it from the load lock to theSTM transfer bar, stainless steel nuts wereadded to the sample holder screws tostabilize the top plate relative to the baseplate as seen in figure 7.
Hidden text Figure 7. Sample holder, Stainless steel nuts wereadded to screws to prevent the sample from cracking.
Difficulty also arose with the sampleholder parts. After the last recorded scan one
of the sample holder’s feet became bent andwas replaced. The sample cracked when itwas placed in the STM stage and a newsample was put in. The sample was scanned last on July 26,2007. The images were lost when the STMsoftware closed unexpectedly. However,good largescale images were obtained. Theimages showed monatomic steps andterraces but individual atoms were notresolved. No other scans were taken because thesamples continued to crack and had to bereplaced and prepared.
Conclusions/Future Work: Difficultiesarose in maintaining a sample long enoughin the UHV system that it could be cleanedsufficiently to achieve an atomically cleansurface. Difficulties included crackedsamples and damaged sample holder parts.When a sample does appear to be clean itcan be dosed with Ag and then rescanned tostudy the phase transitions of Ag on Ge(111). This experiment will continue with aclean and undamaged sample. In order tomaintain a sample long enough in the UHVsystem more solutions are needed to preventthe sample from cracking.
Acknowledgements: I would like to thankDr. Shirley Chiang for overseeing thisresearch project, graduate student CoryMullet for supervising and guiding thelaboratory work, graduate student JasonGiacomo and undergraduate student DanielMembreno for their assistance, Dr. RenaZieve for coordinating the ResearchExperience for Undergraduates (REU)program at UC Davis, the PhysicsDepartment at UC Davis, and the NationalScience Foundation for funding the REUsummer program.
1 P. Hansma, J. Tersoff, J. Appl. Phys. 61, R1 R23 (1986).2 M. Herz, F. J. Giessibl,, and J. Mannhart, Phys. Rev. B 68, 045301 (2003)