Friday, May 13, 2011

Re: PhD Oral Examination: Shankar Swaminathan (Thursday, May 27th, 9:30AM)

li

----- Original Message -----
From: "Shankar Swaminathan" <shanswam@stanford.edu>
To: mse-students@lists.stanford.edu, mse-faculty@lists.stanford.edu,
glamstuds@lists.stanford.edu,
labmembers@snf.stanford.edu
Sent: Friday, May 14, 2010 12:44:11 AM
Subject: PhD Oral Examination: Shankar Swaminathan (Thursday, May 27th,
9:30AM)


Nanoscale atomic-layer-deposited high- k gate oxides on Germanium:

Interface engineering for highly-scaled MOS devices

Shankar Swaminathan

Department of Materials Science and Engineering


Advisor: Dr. Paul C. McIntyre

Location: Bldg. 200 Rm. 002 (History Corner 200-002)

Date: Thursday, May 27, 2010


Time: 9:30 AM (refreshments served at 9:15 AM)


Germanium (Ge) has emerged as a promising candidate for surface channels
in highly-scaled field-effect-transistors (FETs), as performance and
reliability issues are likely to limit the use of conventional Si-based
complementary-metal-oxide-semiconductor transistors. Lack of a high
quality and stable thermal oxide of germanium has prompted interest in
the use of high- k (high dielectric-constant) gate dielectrics on Ge
channels. An interface layer (IL) between the high- k film and the Ge
substrate appears to be necessary to avoid large defect densities
characteristic of atomically-abrupt high-k (ZrO 2 or HfO 2 )/Ge
interfaces. Atomic layer deposition (ALD) is a useful high- k metal
oxide film growth technique due to the precise nature of thickness
control and uniformity of thickness for very thin films. The use of
ALD to synthesize deposited ILs interposed between the Ge channel and an
overlying high- k layer has not been studied extensively. In this
context, we present three highlights from our work:

First, we show that a pre-ALD surface functionalization by oxidant
dosing improves the electrical characteristics of Pt/4nm ALD-Al 2 O 3
/p-Ge devices, as opposed to a conventional precursor-first ALD process.
In situ x-ray photoelectron spectroscopy in the ALD ambient reveals the
influence of hydroxyl (-OH) termination of the Ge surface in passivating
dangling bonds that lead to fast trapping. The evolution of Ge-O bonding
states during this pre-pulsing process is correlated with the observed
improvements in hysteresis, frequency dispersion of the gate
capacitance, and the passivation of fast (band-edge) and slow (midgap)
interface states.


Second, we present findings on the effects of scaling the physical
thickness of the ALD-Al 2 O 3 (down to 1 nm) on important electrical
parameters such as interface state density (D it ), capacitance density,
fixed charge and leakage current density. The ultrathin-ALD-Al 2 O 3 /
Ge interface shows no apparent interfacial suboxide (GeO x ) while
retaining a low D it of 2x10 11 cm -2 eV -1 , indicating the potential
of ALD-Al 2 O 3 as an interface passivation layer.


Aggressive gate capacitance scaling necessitates the use of so-called
"higher-k" dielectrics such as TiO 2 (k > 25). However, the conduction
band offset (CBO) of the TiO 2 /Ge interface is very low (~ 0.2eV),
resulting in high gate leakage. We demonstrate that ultrathin (~ 1 nm)
Al 2 O 3 ILs, owing to their large bandgap (~ 6.2 eV), enhance the CBO
at the TiO 2 /Ge interface, and reduce the gate leakage by 4 to 6 orders
of magnitude at flatband. The Pt-gated bilayer devices exhibit excellent
C-V characteristics down to a capacitance-derived EOT of 1.2 nm. The
permittivity of the amorphous/nanocrystalline ALD-TiO 2 films was in the
range of 32-35. Forming gas annealing is beneficial in lowering the
interface state density via formation of monolayer-thickness interfacial
GeO 2 . Post-annealed devices exhibited a minimum D it ~ 3x10 11 cm -2
eV -1 near midgap. Finally, we show that thermally activated electron
transport into shallow defect states in the TiO 2 (~0.25eV below the CB
edge) near the TiO 2 /Al 2 O 3 interface appears to be responsible for a
temperature-dependent dispersion of the accumulation capacitance
density. The passivation of these defects is important to fully realize
the potential of novel bilayer high- k gate stacks on germanium channels
.

-- Shankar Swaminathan
Stanford Graduate Fellow '05
McIntyre Research Group
203P, McCullough (G-LAM)
Stanford University, CA
Ph: 650 796 6929

--++**==--++**==--++**==--++**==--++**==--++**==--++**== mse-students
mailing list
mse-students@lists.stanford.edu
https://mailman.stanford.edu/mailman/listinfo/mse-students

No comments: