Mechanical Testing of Aluminum Microbeams Using an Atomic Force Microscope

1
Mechanical Testing of Aluminum Microbeams Using
an Atomic Force Microscope

• Adam Falk
• May 10th 2002
• Advisor Dr. Gleixner

2
Presentation Overview

• Project Objective
• Experimental Procedure
• Problems Encountered
• Results of Experiment
• Further Work
• Conclusion

3
Project Objective

• Overall Goal
• Calculate Youngs Modulus of aluminum on a micro
  scale
• To Accomplish this
• Fabricate aluminum microbeams
• Develop testing protocol using AFM
• Test microbeams
• Calculate Youngs Modulus from data received

4
Reason for Experiment

• Mechanical properties are important to designers
  of any device micro or macro
• Little is known about the properties of materials
  on such a small scale
• Youngs modulus describes the ability of a
  material to undergo recoverable deformation
• Aluminum used in contacts or MEMS devices are
  under stress and it is important to know the
  limits of the strain the material can withstand
  

5
Why Use the AFM?

• AFM has the ability to apply a known force to a
  small surface
• AFM can measure the deflection caused by that
  force
• AFM is relatively easy to use and they are widely
  available

6
Beam dimensions

• Beams will be varying dimensions
• Height and thickness will be controlled by
  experimental procedure
• Length and width will be controlled by the
  dimensions of the mask

7
Mask Parameters

• 16 beams per group with varying dimensions X 8
  groups per block X 25 blocks per wafer 3,200
  possible beams

8
Procedure for fabrication of Microbeams
Oxidation and 1st Photolithography Step
The thickness of the oxide determines the height
of the beam and the 1st mask defines the
dimensions of the base
9
Metallization
Thickness of the aluminum layer defines the
thickness of the beam
10
Results of Thickness Measurements
11
Second Photolithography Step and Final Etch

• Dimensions of the mask define the length of the
  beams

12
Problems with etching the Oxide from below the
beams

• An anisotropic etch is necessary to etch from
  underneath the beams
• It must attack oxide but not attack the aluminum
• Two different etching techniques were used to
  etch the oxide from beneath the beams

13
Etching method 1 Wet Etch

• Wafers were placed in buffered oxide etch (BOE)
  for 40 min
• BOE etches oxide at 400A per sec
• BOE etches aluminum considerably slower
• Experiment was designed to see if all of the
  oxide could be etched with minimal damage to the
  aluminum

14
Results of BOE Etch

• BOE overetched the majority of the beams
• Aluminum surface showed high degree of overetch

15
Aluminum Microstructure

• BOE attacked the grain boundries
• Photo resist was not sufficient to protect the
  top of the beams

16
Was all of the Oxide Etched From Below the Beams?

• Energy Dispersive Spectroscopy (EDS) was
  preformed on randomly selected beams
• This was done with an Scanning Electron
  Microscope (SEM) equipped with an IXRF detector
• EDS can accurately detect what elements are
  present in a given location and can determine the
  relative atomic percentages present
• If the oxide was etched from underneath the beams
  there will be little or no oxygen detected

17
Results of EDS Analysis
18
EDS Quantitative Analysis

• Oxygen content is over 20 which indicates the
  presence of oxide beneath the beams

19
Etching Method 2 Dry Etch

• Wafers were placed in the Applied Materials
  plasma etcher for 5 hours
• The Applied Materials etcher does not attack
  aluminum at all
• To make the etching less isotropic the bias was
  set to a lower voltage
• Zero volts was attempted but the etcher was
  unable to form plasma

20
Results of Plasma Etch

• All beams were high quality. Resolution was high
  and all of the lines were clean

21
Did the Plasma Etch the Oxide From Beneath the
Beams?

• No, Oxide was still detected

22
Solution to the Etching Problem

• BOE was not a good etching media
• The plasma etch had good results.
• Research needs to be done to see if there is a
  way to set the bias to zero
• Photoresist can be stripped anisotropically using
  the March plasma etcher

23
Change in Procedure

• Instead of using oxide as the sacrificial layer
  use photoresist

24
Results of New Procedure

• Most successful run yet
• Beam height can be seen by SEM
• Some problems with overetching of aluminum during
  second photolithography step

25
(No Transcript)
26
(No Transcript)
27
Did the New Procedure Etch Beneath the Beams?

• Photoresist would leave behind organic traces.
  EDS analysis shows no contaminants.

28
Further Proof of Complete Etching

• A fast map was performed on a randomly selected
  beam
• A fast map searches for selected elements and
  displays the exact location that each element is
  found at

29
Fast Map
30
AFM Testing Mechanism

• AFM testing conducted using contact mode

31
AFM Testing Procedure

• Cantilever arm will press down on the top of the
  beams with a known force
• AFM will measure and return the deflection of the
  cantilever beams
• The spring constant of the cantilever arm must be
  calculated
• The deflection data can be mathematically
  manipulated to calculate Youngs modulus

32
Future Work

• AFM testing has not began due to difficulty with
  scheduling and difficulty with etching the beams
• One more run of beams will be completed this
  semester
• The completed beams will be tested this summer

33
Conclusion

• Fabricate microbeams using IC processing
  techniques
• BOE etch is too aggressive to etch the oxide and
  not destroy the aluminum
• Plasma etching and using photoresist as the
  sacrificial layer have yielded reasonable
  results
• AFM testing will be conducted this summer

34
Thank You!!