P-type Doping of Silicon Nanowires

Download Report

Transcript P-type Doping of Silicon Nanowires

Principles and Applications of Molecular
Beam Epitaxy
Aaron Vallett
EE 518
April 5th, 2007
Instructor: Dr. J. Ruzyllo
Outline

Introduction

Review of epitaxial growth

MBE Process
 Chamber construction
 Beam sources
 Characterization

MBE Applications
 Devices
 R&D/Commercial

Summary
Introduction

Invented in late 1960s at Bell Labs by J. R. Arthur and
A. Y. Cho

An epitaxial growth process involving one or more
molecular beams of atoms or molecules physically
arranging themselves on a crystalline surface under
ultrahigh-vacuum conditions

Growth is tightly controlled – layer compositions and
thickness can be adjusted at an atomic scale
Epitaxy Review



Growth of thin, high quality, single-crystal layers on a
similar-type crystal substrate
Molecules are adsorbed on the surface
Diffuse across the surface until finding a suitable crystal
site
Image from http://www.phys.ubbcluj.ro/~rote/Zahn/Introduction.pdf
MBE Process Overview






Very similar to
thermal evaporation
with one big
difference - UHV (10-8
- 10-11 torr)
Solid source materials
are heated to melting
point in effusion cells
UHV gives source
molecules a large
mean free path,
forming a straight
beam
Beam impinges on heated substrate (600°C)
Incident molecules diffuse around the surface to the proper
crystal sites and form crystalline layers
Characterization tools allow growth to be monitored in-situ
Image modified from http://projects.ece.utexas.edu/ece/mrc/groups/street_mbe/mbechapter.html
MBE Chamber




Stainless steel chamber and seals reduce leaks
After servicing, chamber must be baked and outgassed
at ~200°C for 2-5 days
UHV achieved through use of cryo, Ti-sublimation and
ion pumps – no oil
Cryo-shroud promotes condensation of contaminants
and stray particles
Image from http://ocw.mit.edu/NR/rdonlyres/Electrical-Engineering-and-Computer-Science/6772Spring2003/B5D923F5-9B4C-4436-A1F1-0343B35E1928/0/lect8_part1.pdf
Sample Preparation and Loading

Starting substrate must be
ultra clean and flat
 Wafer usually comes “epiready” with a protective oxide





Substrate loaded in load-lock
and heated for outgassing for
several hours
Substrate may then move to a
buffer chamber and be
outgassed again
Growth substrate then loaded
onto holder in growth chamber
Protective oxide desorbed by
heating substrate on the chuck
in UHV
Goal is to keep the chamber and sample as pure as
possible
Image from http://www.uwo.ca/isw/images/Mbeiiism.gif
Effusion Sources and the Molecular Beam




Effusion: the process where individual molecules flow through a
hole without collisions
Source material is heated to vapor phase
Ultra-low pressure in UHV leads to molecules with mean free paths
of hundreds of meters
Opening in effusion cell is small – molecules travel straight out of
it with no collisions, forming a beam
Images from http://www.mbe-kompo.de/products/effusion/effusioncell_ome.html
and http://zumbuhllab.unibas.ch/060929GufeiMBE.pdf
Effusion Cell Construction



A typical MBE system may feature 8 effusion cells
Crucible is constructed of pyrolytic boron nitride (PBN) to withstand
temperatures up to 1400°C
Thermocouple must accurately measure crucible temperature
 Change in T of .5°C changes flux by 1%
 During the day flux variations of <1%, day-to-day <5%
 T must be controlled within a half-degree at 1000°C


Images from http://www.riber.com/en/public/solidcells.htm
and http://www.hlphys.jku.at/fkphys/epitaxy/mbe.html
Sources seated in a cooling shroud to
maintain flux and eliminate thermal crosstalk
between cells
Mechanical shutters in front of sources control
the beam
In-situ Characterization


Deposition in UHV allows unique in-situ measurements to be taken
RHEED – reflection-high-energy-electron-diffraction
 Electrons from a gun strike the growing surface at a shallow angle
 The crystal reflects electrons into a diffraction pattern
 Diffraction pattern and intensity can provide information on the state of the
surface

Mass spectrometry
 Used to measure surface and
chamber composition

Ionization gage
 Used to measure chamber
pressure or molecular beam flux
Images from http://www.elettra.trieste.it/experiments/beamlines/lilit/htdocs/people/luca/tesihtml/node25.html
and http://www.phys.ubbcluj.ro/~rote/Zahn/Introduction.pdf
MBE Abilities




Deposition rate is ~ 1 μm/hr or 1
monolayer/sec
Computer controlled shutter can
be opened or closed in 100 mS
Defect free, super abrupt, singleatom layers can be grown – only
MBE allows this precision
Multiple beams can impinge the
surface at once to create III-V
materials or dope a layer during
growth
AlGaAs/GaAs alternating layers
15 monolayers
Images from http://www.phys.ubbcluj.ro/~rote/Zahn/Introduction.pdf and
http://research.yale.edu/boulder/Boulder-2005/Lectures/Willett/boulder1.pdf
Device Applications

Traditionally used for very
specific, commonly compoundsemiconductor, applications





HBTs, MESFETs and HEMTs
Quantum wells
Semiconductor lasers
Silicon-on-sapphire growth
Also being considered for use
in commercial production of
SiGe MOSFETs
Images from http://www.micro.uiuc.edu/mbe/laserd.htm
and Thompson et. al. IEEE Trans. On Semicon. Manufacturing, Vol. 18, No.1, February 2005
MBE in Industry


By nature MBE has a very low throughput
If it is needed for future CMOS processing, manufacturers
will install clustered MBE chambers to increase throughput
Images from http://users.ece.gatech.edu/~alan/ECE6450/Lectures/ECE6450L13and14CVD%20and%20Epitaxy.pdf
Summary

MBE creates near-perfect crystalline layers

MBE is a purely physical process, so blocking the beam can
stop layer growth

Slow growth time allows atomically thin and super abrupt
layers to be grown

Mixing of beams permits growth of compound semiconductor
and doped layers

MBE is a costly and time consuming technique, but its high
level of precision may drive it into the commercial CMOS
world
References
Parker, E. “Technology and Physics of Molecular Beam Epitaxy” 1985
Chang, L. and K. Ploog “Molecular Beam Epitaxy and Heterostructures” 1985
Liu, W. “Fundamentals of III-V Devices” 1999
http://www.phys.ubbcluj.ro/~rote/Zahn/Introduction.pdf
http://projects.ece.utexas.edu/ece/mrc/groups/street_mbe/mbechapter.html
http://www.uwo.ca/isw/images/Mbeiiism.gif
http://www.mbe-kompo.de/products/effusion/effusioncell_ome.html
http://zumbuhllab.unibas.ch/060929GufeiMBE.pdf
http://www.riber.com/en/public/solidcells.htm
http://www.hlphys.jku.at/fkphys/epitaxy/mbe.html
http://www.elettra.trieste.it/experiments/beamlines/lilit/htdocs/people/luca/tesihtml/node25.html
http://www.phys.ubbcluj.ro/~rote/Zahn/Introduction.pdf
http://research.yale.edu/boulder/Boulder-2005/Lectures/Willett/boulder1.pdf
http://www.micro.uiuc.edu/mbe/laserd.htm
R M Sidek et al 2000 Semicond. Sci. Technol. 15 135-138
Thompson et. al. IEEE Trans. On Semicon. Manufacturing, Vol. 18, No.1, February 2005
http://users.ece.gatech.edu/~alan/ECE6450/Lectures/ECE6450L13and14-CVD%20and
%20Epitaxy.pdf