Loadunload curves of aaaxis and b caxis ZnO loaded to 10 and 50 mN' Insets show loadunload curves fo

1
Mechanical behaviour of a- and c- axis epitaxial
ZnO grown on sapphire
V. A. Coleman1, J. E. Bradby1, C. Jagadish1, M.
R. Phillips2, M. V. Swain3 and P. Munroe4 1
Department of Electronic Materials Engineering,
Research School of Physical Sciences and
Engineering, The Australian National University,
Canberra, ACT 2 Microstructural Analysis Unit,
University of Technology Sydney, Broadway, NSW 3
Department of Oral Sciences, School of Dentistry,
The University of Otago, Dunedin, New Zealand 4
Electron Microscope Unit, The University of New
South Wales, Randwick, NSW
WHY ZINC OXIDE?
DEFORMATION MECHANISM
LOAD-UNLOAD CURVES

• Zinc Oxide (ZnO) is a future material with
  great promise for blue and ultra violet
  optoelectronic devices
• Wide band gap (3.4 eV) and large exciton
  binding energy (60 meV at room temperature)
• Radiation hardness
• Possibility for wet chemical etching
• Low power threshold for optical pumping
• Can grow high quality single crystals and
  epitaxial layers
• However.
• ZnO is a very soft material highly susceptible
  to mechanical (contact induced) damage1
• Device processing involves extensive surface
  contact
• Need to know if device processing will affect
  the active material and reduce the
  attractiveness of ZnO for device purposes

BF XTEM of an indent in a-axis epi ZnO. Note
there is no evidence of delamination of the ZnO
film. The area under the indent shows heavy
damage, however no slip lines are evident.
Dislocations and substrate inhibit slip making
the epi-layers harder than bulk.
ORIENTATION EFFECTS
Conversely to what is seen in bulk ZnO, a-axis
epilayers are harder than c-axis layers. This
arises due to the orientation of the basal
planes. In c-axis epi the planes are oriented
parallel to the substrate, where-as for a-axis
epi, they are perpendicular to the substrate.
During indentation, deformation along the basal
plane is inhibited by the underlying sapphire
substrate for a-axis layers, but not for c-axis
layers.
EXPERIMENTAL

• a- and c- axis epitaxial layers were indented
  with an Ultra-Micro Indentation System 2000
  (UMIS) up to a maximum load of 50 mN, using
  continuous load and partial load-unload
  nanoindentation schemes (4.3 mm spherical
  diamond indenter, ambient conditions)
• Compared bulk a- and c- axis single crystals
• Field and Swain2 analysis method was used to
  extract hardness and elastic modulus values
• Cross sectional electron microscopy (XTEM), and
  monochromatic cathodoluminescence (CL) imaging
  were used to characterize the damage created by
  indentation
• XTEM samples were prepared using a focused-ion
  beam system with 30keV Ga ions. Prior to
  milling, a Pt layer was deposited to protect the
  surface

EFFECT ON LUMINESCENCE
a-axis bulk
c-axis bulk
Load-unload curves of (a)a-axis and (b) c-axis
ZnO loaded to 10 and 50 mN. Insets show
load-unload curves for single crystal bulk ZnO of
corresponding orientation. No pop-in events are
seen for the epitaxial ZnO, whereas the bulk ZnO
has either a single pop-in event (a-axis) or
multiple pop-in events (c-axis).
CL imaging (380 nm at lN2) shows that the exciton
luminescence is quenched at the site of
indentation as well as along slip lines.
CONCLUSIONS
THE ROLE OF DISLOCATIONS
Epitaxial ZnO is harder than bulk ZnO, making it
more favourable for device fabrication. a-axis
epilayers are also harder than c-axis epilayers,
indicating that they will be the most robust
choice for ZnO devices. This work has significant
implications for the design of ZnO-based
optoelectronics.
BF XTEM image of a ZnO epilayer taken away from
the indent region showing many threading
dislocations present throughout the material.
These are responsible for the spread in data
values of the epilayers, and inhibit slip making
the epilayers harder.
RESULTS HARDNESS AND MODULUS
REFERENCES

• S. O. Kucheyev, J. E. Bradby, J. S. Williams, C.
  Jagadish, M. V. Swain, P. Munroe and M. R.
  Phillips, Appl. Phys. Lett. 80, 956 (2002)
• J. S. Field and M. V. Swain, J. Mater. Res. 8,
  297 (1993)