Compositionally Dependent Band Offsets in Aln/Al x Ga 1-x N Heterojunctions Measured by Using X-Ray Photoelectron Spectr
- PDF / 322,035 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 9 Downloads / 224 Views
775 Mat. Res. Soc. Symp. Proc. Vol. 482 01998 Materials Research Society
160000
"GLMM
140000 NIs
120000 ols
100000
42 00000
Ga3p
SGa3
GO
60000
>,
I
Pre Exposure
C40000
-E20000
Exposure
-20000Post -200002 700
500
500
400
300
200
100
0
Binding Energy (eV)
Figure 1:
XPS survey spectrum of GaN sample before and after exposure to nitrogen plasma showing
decrease in 01s peak. 800a
GaNI !P•600
b
6G
b
U
a
"AIN/aNI
|Ga3d
~2o
E0I~o.,J~%~~1 ":iz2°
3
03
25 20 1510
-5
0
Binding Enegy (eV)
-5
odd
AP (nlboenbonded)
300
04000 400. =~0
~~~e
350P c
G~d
___-AI
2J2p (oxMbnded)
200 •i15°
":i 0G:o
I,•.
.. 80
60
40
20
Binding energy [eV]
80
60
40
20
0
Binding Energy (eV)
Figure 2: X-ray Photoelectron Spectroscopy data showing a) Ga3d and valence band from GaN
sample,
b) Al2p and Ga3d core levels from AiN/GaN and c) A12p and valence band from A1N sample. The samples were then returned to the MBE chamber and exposed to nitrogen plasma to remove any oxygen contamination. Fig. 1 shows two XPS spectra illustrating the removal of oxygen by nitrogen plasma exposure. The change in adsorbed oxygen can be seen by comparing the Ols peak, located at -531 eV, in the pre exposure spectrum to the Ols peak in the post exposure spectrum. By comparing sensitivity adjusted electron flux from the Ols peak relative to Ga3d core level peak, the contribution to the net XPS spectrum from
the adsorbed oxygen on the surface was estimated at
",,
6% before exposure. After a fifteen
minute exposure to nitrogen plasma the Ols peak was undetectable(< 0.5%). A thin layer (,,25 A) of AlN was then grown on top of the A1GaN. The thickness of the AlN cover layer fulfilled two requirements. First, it was thick enough to relax strain due to the A1N/AlGaN lattice mismatch. Second, it was thin enough to allow significant electron flux from the underlying AlGaN layer to escape the cover layer without scattering. Measurements of the core level separations were then made. Fig. 2 a-c) show the XPS data required to measure the valence band offset in AIN/A1GaN hetreo junctions. ANALYSIS The valence band offset (AEv) is given in terms of the measured separations shown in
776
fit1
.
.
.
.
.
.
Fio FitI Ftt
•"45-•t
.
I _______ 1.0 --- Bloom brecht] . - .............. LambrctI°
___ql,
a). a
=1 358 "30 -0.
CO 0.66
"250 -1
•)
0.8
210.8
-
0.2-
0
S4
0.0
-10
Energy (eV)
-5
0
5
Energy (eV)
Figure 3: a) Broadened theoretical valence band density of states fit to the leading edge of AlGaN valence
band XPS data. 0 eV is the valence band maximum. b)Comparison of theoretical valence band densities of states. Fig. 2a-c by: =
AE, +
AE, =A~d+
(E AIGaN
•Ja3d
EAIGaN)
-
--
Ev
(•AIN (E-A12p
E-AIN)(
E--
,
1
Core level peak positions (EPN,E!) were determined by fitting a linear combination of Gaussian and Lorentzian peak shapes to core peak data. Valence band maximum positions (EAlGeN,EýlN)were
determined by convolving a theoretical valence band density of states
(TVBDOS) with a predet
Data Loading...
