HfO2InAs Ellipsometry Results

1
HfO2/InAsEllipsometry Results

• Dane Wheeler
• February 15, 2008

2
Setup

• Large or small beam can be used
• Large beam (4 mm diameter) ? better
  signal-to-noise ratio (SNR) but need more area
• Smaller beam (2 mm diameter) ? worse SNR but can
  probe smaller areas
• Large piece of bulk InAs was measured with big
  and small beams
• Next, the InAs reference sample from Lund was
  measured with the large beam (it has a small area
  of InAs not covered by metal)
• Lastly, an HfO2/InAs sample was measured with a
  large and small beam (it also has a small area
  not covered by metal)

3
Bulk InAs sample with small spot
Reasonable model fit but data is noisy The
problem is InAs is not very reflective due to the
narrow band gap
4
Bulk InAs sample with large spot
Excellent fit And the extracted refractive index
is a pretty good match to published data
5
Lund InAs sample with large spot
No signal Not a large enough area on sample to
get anything
6
HfO2/InAs sample with small spot
Sample 6b 300 C, HCl, TiAu 50 cycles (4.8 nm)
Some kind of signal present, but its noisy and
not sure if its even from HfO2/InAs Tried a
bunch of different models and fitting techniques
and could not get a good fit
7
HfO2/InAs sample with large spot
Sample 6b 300 C, HCl, TiAu 50 cycles (4.8 nm)
Large spot improves SNR Still could not get a
good model fit, however, could be because of my
inexperience with the tool Not sure if the
HfO2/InAs is being probed here data is much
different in character from bare InAs data (note
that the angle trend is reversed)
8
Conclusions

• Small and large beam data agree, but large data
  less noisy
• Refractive index data extracted for InAs seems
  reasonable, so I think the technique is okay
• However, the samples from Lund do not have large
  enough bare HfO2 areas to probe (maybe with the
  exception of the sample presented on slides 6
  7, sample 6b)
• Future depositions should include a sample the
  receives no contact metal for ellipsometry