Kuo-Jen Chao

Formation of Atomically Flat Silver Films on GaAs with a "Silver Mean" Quasi Periodicity

A flat epitaxial silver film on a gallium arsenide [GaAs(110)] surface was synthesized in a two-step process. Deposition of a critical thickness of silver at low temperature led to the formation of a dense nanocluster film. Upon annealing, all atoms rearranged themselves into an atomically flat film. This silver film has a close-packed (111) structure modulated by a “silver mean” quasi-periodic sequence. The ability to grow such epitaxial overlayers of metals on semiconductors enables the testing of theoretical models and provides a connection between metal and semiconductor technologies.

Cross‐sectional scanning tunneling microscopy study of GaAs/AlAs short period superlattices: The influence of growth interrupt on the interfacial structure

We report studies of GaAs/AlAs short period superlattices using cross‐sectional scanning tunneling microscopy. In particular, we investigate the role of growth interrupt time on the resulting interfacial structure. Superlattices with repeated periods of four layers of GaAs and two layers of AlAs are resolved atom by atom. Superlattices grown using a 30 s growth interrupt time are observed while those grown with a 5 s growth interrupt time are not. We also discuss residual effects of the growth interrupt process on layers grown on top of the short‐period superlattice.

Scanning tunneling microscopy of doping and compositional III–V homo‐ and heterostructures

Scanning tunneling microscopy (STM) was used to study the (110) cross‐sectional surfaces of molecular‐beam epitaxially grown III–V homo‐ and heterostructures, which include GaAs multiple p–n junctions, (InGa)As/GaAs strained‐layer multiple quantum wells, and (AlGa)As/GaAs heterojunctions. Both doping and compositional effects can be resolved by the topographic contrasts of constant‐current STM images. The samples were prepared by either cleaving in ultrahigh vacuum or cleaving ex situ followed by sulfide [(NH4)2S] passivation. Sulfide passivated samples have been found to be advantageous for the measurements of scanning tunneling spectroscopy.

Factors influencing the interfacial roughness of InGaAs/GaAs heterostructures: A scanning tunneling microscopy study

Using cross-sectional scanning tunneling microscopy, we have investigated factors which influence interfacial roughness in InGaAs/GaAs heterostructures and have found that the roughness of the growth front and In segregation are two major factors influencing the interfacial roughness. In addition, we noticed no preferential clustering of indium atoms along the [001] growth direction as previously reported by others. Furthermore, a growth procedure which combines substrate temperature ramping with a growth interruption results in an atomically smooth interface.

Two-dimensional pn-junction delineation and individual dopant identification using scanning tunneling microscopy/spectroscopy

We have used scanning tunneling microscopy and spectroscopy (STM/S) to study multiple pn junctions on cross-sectional surfaces of both Si and GaAs devices. The spectroscopy results indicate that pn junctions can be resolved at the nanometer scale by using the two-dimensional STS technique. STM is also used to identify Zn dopants on GaAs(110) surfaces. A detail dopant location identification method is presented.

Variable low‐temperature scanning tunneling microscopy study of Si(001): Nature of the 2×1→c(2×4) phase transition

We have studied the Si(001) surface from 120 K to room temperature using a variable low‐temperature scanning tunneling microscope. Complementary investigations were carried out on two distinctly different types of surfaces: first, the normal 2×1 surface and second, the 2×n (4<n<12) surface. For the 2×1 surface, the defects are scattered randomly. By plotting out the fraction of buckled dimers as a function of temperature, we find a slow transition from predominantly c(2×4) at low temperature to mostly 2×1 at room temperature for a defect concentration of about 8.5%. For the 2×n surface, the much larger number of surface vacancies form long‐range ordered chains, dividing the surface into many short dimer segments. These dimer segments predominantly appear to be unbuckled at room temperature. Upon cooling to 190 K, we observe very little change in the amount of buckling. The implications of this result are discussed.

Temperature dependent compensation of Zn‐dopant atoms by vacancies in III–V semiconductor surfaces

Scanning tunneling microscopy is used to study the compensation of Zn‐dopant atoms by surface vacancies on InP and GaAs(110) surfaces as a function of the annealing time and temperature. An attractive interaction is observed between negatively charged Zn atoms and positively charged anion vacancies. Uncharged Zn‐vacancy defect complexes are formed and their structure is analyzed. The concentration of the complexes increases with the vacancy concentration. Simultaneously the concentration of charged Zn decreases. The total observable Zn concentration is found to be the sum of the concentration of isolated charged Zn atoms and of the concentration of the defect complexes. This sum is constant at low temperatures, but decreases at 480 K. The observations are explained by the formation of subsurface defect complexes at higher temperatures and surface complexes at low temperatures.

Scanning tunneling microscopy investigation of the dimer vacancy–dimer vacancy interaction on the Si(001) 2×n surface

Scanning tunneling microscopy has been used to investigate the formation of the 2×n vacancy line structure on Si(001). We find that quenching the surface from high temperatures results in the formation of vacancies. After further quenching, these vacancies nucleate into chains running perpendicular to the dimer rows. Finally, the vacancy chains connect and develop into vacancy lines which extend for many thousands of A’s. Each vacancy line consists of mainly two types of dimer vacancies: (1) a di‐vacancy and (2) a combination of a single vacancy and a di‐vacancy separated by an isolated dimer. All the vacancy lines together with the dimer rows form a 2×n structure with 6≤n≤12. Calculations using the Stillinger–Weber potential support the view that the formation of the vacancy line structure is due to the interaction between vacancies.

Cross‐sectional scanning tunneling microscopy of doped and undoped AlGaAs/GaAs heterostructures

Scanning tunneling microscopy (STM) was used to study the (NH4)2S‐passivated (110) cross‐sectional surfaces of both doped and undoped Al0.3Ga0.7As/GaAs heterostructures on n+‐substrates. The ex situ (NH4)2S treatment of the cross‐sectional surfaces of heterostructures was found to be very stable against oxidation. STM images showed no appreciable deterioration of surface quality in vacuum after more than 40 days. The spectroscopic results on the undoped epilayer showed diodelike behavior, confirming that an undoped large band gap region can be imaged by STM through carrier injection from the conductive regions.

Influence of various growth parameters on the interface abruptness of AlAs/GaAs short period superlattices

Cross‐sectional scanning tunneling microscopy has been used to investigate the effects of several key growth parameters on the resulting interfacial quality of AlAs/GaAs short period superlattices. For growth on top of AlGaAs layers, only superlattices grown with periodicity no smaller than 4 unit cells of GaAs and 2 unit cells of AlAs and grown with a minimum of 30 s of growth interrupt time are resolved. On the other hand, when grown on top of GaAs layers, superlattices as fine as 2 unit cells of GaAs and 1 unit cell of AlAs grown with only 5 s of growth interrupt time are resolved. This result suggests that the material on which the superlattice is grown is at least as important as the growth interrupt time. In particular, GaAs seems to provide a smoother starting surface than AlGaAs and hence aids in the formation of abrupt interfaces. We also compare our scanning tunneling microscopy data with some predictions based on simple atomic models of the interfacial regions.