A high-quality (that is, low surface state density) oxide does not exist for
GaAs.  This has been known 'forever', and is why Si MOSFETs are the dominant
electronic device by far.  (As a benchmark, the surface state density for Si-
SiO2 is equal to or less than 1E10 cm^-2.)

On the other hand, there have been some papers which show possible approaches.
One is:

Nitride-based passivation of GaAs for reduced surface state density
Hariu, T.  Suzuki, N.  Matsushita, K.  Shibata, Y.
IEEE International Electron Devices Meeting, 1978
Page(s): 598 - 599

They use gallium oxynitride deposited on GaAs, with a nitrogen anneal.

Al2O3 insulator is used in:

Microelectronics Reliability
Volume 47, Issue 12, December 2007, Pages 2082-2087
Electrical measurements of voltage stressed Al2O3/GaAs MOSFET
Z. Tanga, P.D. Yeb, D. Leea and C.R. Wiea

Frankly, the VT shifts and gate current look awful, but no mention is made of SS
density.

ALD Al2O3 is discussed in the next reference, which suggest good characteristics
are at least possible:

Processing and Characterization of III-V Compound Semiconductor MOSFETs Using
Atomic Layer Deposited Gate Dielectrics
Book Series     Advanced Microelectronics
ISSN    1437-0387
Volume  Volume 27
Book    Advanced Gate Stacks for High-Mobility Semiconductors
Publisher       Springer Berlin Heidelberg
DOI     10.1007/978-3-540-71491-0
Copyright       2007
ISBN    978-3-540-71490-3 (Print) 978-3-540-71491-0 (Online)
DOI     10.1007/978-3-540-71491-0_16
Pages   341-361
Subject Collection      Engineering
SpringerLink Date       Tuesday, January 01, 2008

Finally, Ga2O3(Gd2O3) ALD oxide insulator is used in the last ref, again with
purported good effect (but for me the gate leakage is very, very high):

Jpn. J. Appl. Phys. 46 (2007) pp. 3167-3180
III-V Metal-Oxide-Semiconductor Field-Effect Transistors with High κ Dielectrics
Minghwei Hong, J. Raynien Kwo1, Pei-chun Tsai, Yaochung Chang, Mao-Lin Huang,
Chih-ping Chen, and Tsung-da Lin

Al Henning
ahenning@nanoink.net

-----Original Message-----
From: mikas remeika [mailto:remeika@physics.ucsd.edu]
Sent: Tuesday, August 03, 2010 3:05 PM
To: General MEMS discussion
Subject: [mems-talk] Surface states between GaAs and SiO2

Hello, everyone,

I am working with a device that involves a layer of SiO2 sputtered on
top of epitaxially grown GaAs. It appears that there is a large number
of surface states that form at the interface (GaAs-SiO2) that
effectively screen any electric fields that are applied perpendicular
to the surface.

Could anyone point me to some literature on eliminating these surface
states? Also, is there another insulating NIR-transparent material
that would not form such surface states at the interface with GaAs?

To give a bit mode detail, I'm trying to fabricate a capacitor-like
device where the half the dielectric is SiO2 and half is GaAs. It is
very important for the potential to drop linearly in the device so
that the field inside a material would only depend on the dielectric
constant of the material.  What I am getting right now is that all
field is screened by the interface between SiO2 and GaAs.

thank you,
-mikas