RF Excitation of the STM Eudean Sun, UC Berkeley, EECS Joonhee Lee, Xiuwen Tu Dr. Wilson Ho IM-SURE Fellow: Graduate Students: Faculty Mentor:

RF Excitation of the STM

Eudean Sun, UC Berkeley, EECSJoonhee Lee, Xiuwen TuDr. Wilson Ho

IM-SURE Fellow:Graduate Students:

Faculty Mentor:

Scanning Tunneling Microscope Angstrom Resolution Microscope Tunneling Current

Quantum Tunneling

JT – tunneling current

VT – DC bias

s – tip-sample distance

Order of magnitude change in JT for every angstrom change in s.

sAVJ TT2/1exp

Scanning Feedback loop

Piezoelectric drives to position sample Ceramic materials that distort with voltage for high-precision

positioning

Maintain constant tunneling current by adjusting tip-sample distance

RF Excitation via a Coil 800MHz – 2.0GHz, 100MHz steps Resonance

Experimental Data

The Model SolidWorks, AutoCAD, Ansoft’s HFSS – High

Frequency Structure Simulator Finite Element Method, like FEMLAB

The Model cont’d Parts:

Inner radiation shield Sample holder Sample RF coil Crosspiece Tip

Excitation 2mA current

Solution Frequency sweep Fields along polylines

Preliminary Results Resonance due to radiation shield

Making a Better Model Added parts:

Outer shield

Coaxial cable

Rails / Grabber

Making a Better Model cont’d New excitation

Wave port Finer mesh

Making a Better Model cont’d Rebuilt all parts in HFSS

Transferring between SolidWorks/AutoCAD and HFSS was inconsistent

Refined solution setupAdded parametric analysis to plot E field

across sample for four different tip-sample distances: 1e-6, 1e-5, 1e-4, 1e-3 in.

Increased data points across polylines to 10,000 for plotting fields.

The New Model…in Color

Results Frequency Sweep

Results cont’d Parametric Analysis

Results cont’d

Results cont’d

Results cont’d Field plots

Results cont’d

Problems Resonance in simulation at 1.2GHz,

1.4GHz, 2.0GHz primarilyResonance in experiment at 800MHz,

1.2GHz, 1.3GHz, 2.0GHz primarily Parametric analysis shows large E field

differences between 1e-3 in, 1e-4 in, and 1e-5 in, but not a big jump between 1e-5 in and 1e-6 in.

“Out of memory”

Potential Fixes Better geometry Finer meshes

Limitations Model complexity

Can’t include everything, but what parts will affect resonance most?

Computer speed 2.66GHz Pentium 4 – 20 hours to complete one

analysis

RAM 1GB physical + 4GB virtual memory “Out of memory”, literally, when finer meshes applied

Acknowledgements Dr. Wilson Ho Joonhee Lee Xiuwen Tu IMSURE Program