Download Etching of Antimonide-Based Semiconductor Materials

Etching of Antimonide-Based Semiconductor Materials
Zhaoran He 1104111
Supervisor: Marc Sorel
Introduction
Light emitting diodes (LEDs) fabricated on narrow bandgap
semiconductors such as AlInSb and GaInSb offer a cheap and
compact solution for developing gas sensors in the 3-5 mm
wavelength range [1]. Fabrication of these devices requires a
precisely controllable etching method to build mesa structures.
This project aims at optimising the dry etching of AlInSb and
GaInSb materials.
Experiments
The wafers with designed structures
were grown on semi-insulating GaAs
substrate by molecular beam epitaxy
(MBE). Dry
etch
process
of
Al0.05In0.95Sb
p-i-n
diode
and
AlSb/Ga0.9In0.1Sb distributed Bragg
reflector (DBR) structures was
investigated. One major limitation is
that wafer temperature cannot exceed
120ºC during the fabrication process
Fig. 1 Wafer layer structure
to avoid performance degradation.
All dry etch tests carried out in this project were
processed in Oxford ICP-RIE 180 machine. The
following parameters were considered:
Methane/Hydrogen Based Etch
Results on the CH4/H2 based recipe show that etch rate is less than
100nm/min but gives low damaged sidewalls and smooth surfaces.
ICP power is used to generate and control
(a)
plasma density. RF bias power accelerates
the plasma dynamic energy to physically
bombard material. The etch rate therefore
RF bias power=100W
increases with increasing of ICP power due to
the larger quantity of plasma introduced into
the chamber. However, high ICP power leads
to rough surfaces because the substantial
(b)
polymer deposition cannot be completely
removed by the slow physical reaction. High
RF bias power gives a high etching rate as
well as smooth surface morphology, while,
ICP power=400W
excessively high bias power causes rough
surface by physical bombardment and it also
reduces selectivity through erosion of the Fig. 4 Etch rate vs ICP power (a)
masks.
and RF bias power (b),
(a)
Table temp=100ºC
•Gas component
•Gas flow and ratio
•ICP power & RF power
•Chamber pressure
•Table temperature
Fig. 2 Schematic
diagram of inductively
coupled plasma
enhanced reactive ion
etching (ICP-RIE)
machine [2]
(b)
Chamber Pressure=20mTorr
Chlorine (Cl) based and methane (CH4)/hydrogen(H2) based
recipes were tested on both diode and DBR structure.
Chlorine Based Etch
Due to the local accumulation of
the by-product InClx polymer on
the surfaces, chlorine based
etchings showed rough etched
surfaces although it provided
high etch rate. Moreover, the
temperature dots also indicated
that
the
actual
wafer
temperature was over 200ºC
which is beyond the maximum
temperature tolerated by this
layer structure.
a
b
c
d
Fig. 3 Cl based etch results (a)(b) diode and DBR
etched with recipe: Cl2/CH4/H2=7/8/5.5sccm,
ICP/bias power=2000/200W, Pressure=4mT, Table
temp=60ºC (c)(d) diode and DBR etched with recipe:
BCl3/Cl2/Ar=7.5/2.5/5sccm, ICP/bias power
=1000/100W, Pressure=4mT, Table temp=20ºC
CH4/H2/O2=6/50/0.2, chamber
pressure=20mTorr
Fig. 5 Etch rate vs chamber
pressure (a) and table
temperature (b),
CH4/H2/O2=6/50/0.2,
ICP/RF=350/150W
20 ℃
60 ℃
100 ℃
120 ℃
O2 is another factor influencing
sidewall profile. Experiments on the
diode structure indicate that O2
generates a passivation layer
protecting the sidewalls from being
etched. However, in the DBR
structure, a better sidewall profile
and higher etch rate was measured
without O2 because the material
contains aluminium-rich layers that
rapidly oxidise and tend to slow
down the etch process.
The chamber pressure and
table temperature also affect
the etch rate and mesa
sidewall profiles. A low
chamber pressure results in
sidewalls with positive slope
but etch rate is decreased
significantly.
The
table
temperature influences the
volatility of the polymer and
therefore
affects
both
sidewall and surface profiles.
As the table temperature is
increased the etch rate
increases and the sidewall
slope changes from positive
to vertical and even shows a
slightly undercut at 120ºC.
Without O2
Without O2
With O2
With O2
(a)
(b)
Fig. 6 Effect of O2 on diode sample (a) and
DBR sample (b)
Conclusion
In conclusion, it was found that chlorine based recipes are not suitable to etch antimonide material under low temperature conditions. The optimum
recipe is a CH4/H2/O2 chemistry with a ratio of 6/50/0.2 sccm, ICP/bias power=350/150W, Pressure=20mTorr, Table temperature=110ºC that
provides vertical sidewalls and smooth surfaces on the diode structure with an etch rate of 50nm/min. It was also found that reducing the chamber
pressure causes a sidewall with a positive slope but decreases the etch rate. The same recipe without O2 and the power increase to 400/200 W
appears to be promising for achieving vertical profiles in the DBR structure.
References: [1] Brain R. Bennett, et al. Antimonide-based compound semiconductors for electronic devices: A review. Solid-State Electronics 49 (2005) pp.1875–1895
[2] Zhang G., et al. Inductively coupled plasma-reactive ion etching of InSb using CH4/H2/Ar plasma. J. Vac. Sci. Technol. A, Vol. 27, No. 4 (2009)
University of Glasgow, charity number SC004401