Morphology of MOCVD grown aplane ZnO thin films:

1
Morphology of MOCVD grown a-plane ZnO thin
films film thickness, temperature and substrate
effects Olga Dulub, Erie Morales and Ulrike
Diebold
2
Surface preparation

• After insertion into vacuum chamber the following
  steps were employed
• Mild anneal (show weak LEED pattern
• Slight sputtering (1-2 keV Ar, 1 ?A, 5-10 min,
  removes ca. 1-4 monolayers)
• and anneal in UHV (500 - 650C) -
  sharper LEED pattern, samples get cleaner
• (could only judge from STM) but overall
  morphology does not change.
• Side Note the new samples we got (Box 1 - 2)
  were not as clean as the old ones. More cleaning
  was necessary to get good STM images. Where
  these handled differently?
• The STM experiments were carried out using an
  Omicron UHV VT-STM at
• Room temperature. All STM data were recorded in
  constant current mode at a positive sample
• bias of 1.5- 2.5 V

3
a-plane ZnO film on r-Al2O3
4
a-plane ZnO films with various thickness(Old
data. First the summary, then the details for
each thickness.)Main Point It seems that more
than 1 critical parameter was varied in this
previous data set.
5
SUMMARY (a-plane ZnO films with various thickness)

• STM -20nm -thick and large areas of 2000nm-thick
• films have smallest roughness.
• LEED - Relatively sharp spots, but streaky
• (small terrace size)

Questions

• What is the reason for the typical waves? Their
  spacing/height?
• What is the influence of growth parameters?
• What was the orientation/value of the miscut
  directions of the sapphire?
• At what temperature was each film grown?

STM size(a-c, e, f) 500 x 500 nm
6
20 nm - thick

• RESULTS
• flat morphology
• terraces are small
• terrace size independent of preparation
  procedure,
• Sputtering cleans surface (easier to image,
  therefore images look better)
• Smallest terraces have similar size to
  dislocation distance at interface observed
  previously with HRTEM
• QUESTIONS
• Why is this very thin film so smooth? Are all
  thin films so smooth? Are the dislocations
  developed yet? What was the growth
  temperature/substrate miscut?

7
100 nm - thick

• Results Typical wavelike features along c-axis.
  100 nm periodicity. 12 nm high. No step-bunching
  (waves symmetrical along c)
• Q Why is P950 different from P854? What was the
  growth temperature? (Or is this just a more dirty
  surface that was harder to images (the waves look
  similar for both). Substrate miscut??

8
100 nm - thick (cont.)

• R Wave-like surface morphology with
• needle-shaped domains running along
• the c-axis, periodicity 150 nm, no step
  bunching (terraces are symmetric)
• Q Why do the surfaces of the two samples look
  different? Was P854 just a dirty surface? Or
  different growth parameters? Temperature? Miscut?

9
450 nm - thick

• R P863 looks very different from P927. WHY?
• R P863 looks like OTHER PHASE (c-oriented!) is
  present) Growth temperature??
• R P927 shows the waves, but they are even more
  pronounced (18 nm high, 150 nm wide) than for 100
  nm thick samples. Waves paralle to c. MISCUT?
  GROWTH TEMP?

c
10
2000 nm - thick

• Very flat morphology, but c-oriented inclusions!
• Different growth conditions were used which
  ones?
• What is the reason for the flat morphology?
  Growth temp? Or are all thick films that smooth??

11
a-plane ZnO films on various sapphire substrates
(Box 2 0.2-0.3, 0.8, 2, 500 nm thick, 2
growth sets (520C, 512C))We first give the
summary, then the detailsMain point The miscut
matters a lot. But we need to know the miscut
orientation for these samples. And we need to
pay close attention to the miscut for all future
samples! For a thorough physical explanation, we
would need to have a fixed miscut angle (film
thickness, growth temperature, etc.) but
different miscut orientations.
12
SUMMARY a-plane ZnO films on various sapphire
substrates (all 500 nm thick, T520C)

• The miscut angle matters a lot!
• smallest miscut gives smooth films, but waviness
  is still there.
• Highest miscut shows indications of waviness, but
  has large terraces and also small rms roughness.
  
• Small and Large angle gives better results than
  0.8 WHY?

We need to know

• miscut directions of the sapphire?

13
0.2-0.3 (growth T520C)

• Flat surface, but waves are there (100 nm
  periodicity, but only 3 nm high)
• waves are parallel to c-axis. Small terrace
  size.

14
0.8 (growth T520C)

• Note that waves are not parallel to c
• different miscut orientation?
• Terrace size is bigger than for smaller miscut
  orientation.
• From atomically-resolved image in the inset we
  can tell the 0001bar direction - apparent
  difference in Zn- and O-terminated step edges.

15
2 (growth T520C)

• Here the miscut is very apparent (ie,
  perpendicular to c). But it would be good to
  confirm this orientation independently!
• There is still some indications for waves
  parallel to c (same periodicity).
• At a small scale, these waves manifest
  themselves as areas with narrower terraces.
• Despite (or because of??) the high miscut angle
  the quality of the surface is very good.

16
0.2 (growth T512C)
17
0.8 (growth T512C)
18
520C vs. 512C
ZnO1044 (grown at 520C )
The images do look different, but it might just
be an imaging artifact. Our impression is that
the growth at 512 and 520 on the 2 different
sapphire substrates gave reproducible results,
i.e., - For 0.8 degrees we have waves, but they
are not paralle to c - For 0.2 degrees miscut we
have waves along c but with a small
amplitude. It is important that we include such
checks in the future!
ZnO1032 (grown at 512C )
19
a-plane ZnO film growth as a functionof
temperature(new data, box 1 - again, first the
summary, then the details about each sample)Main
point The higher the growth temperature the
flatter surfaces! But We need to make sure there
is not also an influence of the miscut
angle/orientation. What was the miscut for these
samples?
20
SUMMARY a-plane ZnO film growth as a function of
temperature

• Films grown at T
• morphology (not shown here)
• Surface quality improves with temperature
• 580C has a similar morphology as 2000 nm thick
  film in previous data set, with grooves
  perpendicular to c-axis. No waviness parallel to
  c!
• 460 film shows waves with high corrugation, 200
  nm periodicity
• 400 film shows similar waves.
• Buffer layer film is very flat (not clear if b/c
  of buffer layer, b/c it was grown at very high
  temperature, which gives smooth film w/out buffer

• miscut directions of each substrate?

21
400C

• Films with typical waves, larger periodicity,
  high roughness.
• NOTE triangular holes in the images - maybe an
  indication for another (c-oriented?? phase)
• Very intersting in addition to waves parallel to
  c also lines perpendicular to c (similar to the
  very smooth film grown at 580C)
• Small-scale image 3 shows terraces not aligned
  with c-direction (we dont know why - curious)
• Miscut angle/orientation?

100 nm x 100 nm
22
460C

• Films with typical waves, large periodicity,
  high roughness.
• Small-scale image 4 again shows terraces not
  aligned with c-direction (??). C-direction
  deduced from LEED
• Miscut angle/orientation?

23
580C

• By far the flattest films!
• Again the black lines perpendicular the
  c-direction.
• This particular film was for some reason hard to
  clean (had to sputter/anneal a number of times.
  Did not change large-scale morphology, but
  terraces in smaller scale images look more
  reasonable.
• Miscut angle/orientation?

24
300C - 520C (buffer layer)

• Also pretty flat film. But other films at hi
  Temp are also flat - buffer layer seems not to
  make much sense?
• Miscut angle/orientation?