Characterization of Ferroelectric Thin Films

1
Characterization of Ferroelectric Thin Films

• Evan Pickett
• Advisor Dr. John Blendell
• MSEL
• Thin Films Group, Ceramics Division

2
Overview

• Ferroelectric Thin Films
• Atomic Force Microscopy
• AFM and Ferroelectrics
• Research Results

3
Ferroelectrics

• Spontaneous electrical polarization

Ti
O
Ba
Images c.o. Joshua Hertz
4
Its not that simple
The crystal cells can be oriented in any
direction, not just up or down.
5
Ferroelectrics II

• Regions of similar polarization

1. Under unpoled conditions, domain distribution
is approximately 50/50 (lowest energy arrangement)
2. An applied electric field causes domain
switching
6
Why study ferroelectrics?

• Domains either up or down - 1 or 0
• Smaller domain size results in increased storage
  density
• Domains retain polarization until switched - can
  store data after voltage is removed

7
Ferroelectric Thin Films PZT

• PbxZr1-xTiO3

• Tetragonal crystal structure
• a0, b0 4.036 Å
• -c0 4.146 Å (2.7 longer)
• lt111gt lattice direction normal to film surface
  (gray triangle)

c0
b0
a0
8
Reference Frames
Co. John Blendell
9
Atomic Force Microscopy

• Measuring topology

10
Atomic Force Microscopy II

• Measuring Vertical Polarization Vectors

11
Atomic Force Microscopy III

• Measuring Lateral Polarization Vectors

12
AFM and Ferroelectrics
Vertical
E
E
Electric field is in phase with response
Electric field is out-of-phase with response
Z
Z
C-axis up
C-axis down
Electric field is in phase with response
Electric field is out-of-phase with response
C-axis right
C-axis left
Lateral
c.o. Joshua Hertz
13
Results I
Vertical polarization vectors
10 V DC applied
10 minutes
20 minutes
and removed
30 minutes
40 minutes
14
Results II
Vertical
(Z)
Deflection (Topology)
Polarization (Phase)
Phase and Amplitude
Lateral
(Y)
15
Results III

• Domain Switching

The image at right is of the same area. Vertical
polarization vectors are shown. The upper image
was subjected to a -6.0V DC bias, and the lower
to 6.0V. So switching really does occur.
16
Results IV

• Motion of the sample surface due to electric
  field 1/1000th to 1/100th of amount predicted by
  piezoelectric effect
• e3d33E3 and e1d13E3 E is the electric field, d
  is a constant, and e is the strain.
• Microscope was not calibrated incorrectly
• Is mis-orientation of c-axis responsible?

17
Results V (Piezoelectric Coefficient)
d33 d33cos3(x) (d13 - d15)cos(x)-cos3(x)

• Laboratory 3 axis aligned with crystal lt111gt
  axis
• Angle of 36 degrees between laboratory 3 and
  crystal 3 axes.

d33
X (degrees) angle between crystal axis 3 and
laboratory 3
18
Conclusions

• Domain boundaries (polarization changes) exist
  within grains
• Over time, switched domains relax back to unpoled
  state starting from pinned sites
• Observed movement of thin film surface is far
  less than expected for bulk sample of same
  material

19
Future Developments

• Show that 90o domain switching occurs in
  practice, not just in theory
• Determine source of error in predicted
  piezoelectric movement
• Use collected data to determine actual
  orientation of polarization vector

20
Acknowledgements

• NIST and the SURF Program
• Dr. Terrell Vanderah and Dr. Marc Desrosiers
• Dr. John Blendell
• Dr. Jay Wallace and Dr. Grady White
• Deb Brown, Dr. Clive Randall, and Dr. Richard
  Tressler