ArticlePDF Available

Ni-(In,Ga)As Alloy Formation Investigated by Hard-X-Ray Photoelectron Spectroscopy and X-Ray Absorption Spectroscopy


Abstract and Figures

The electrical, chemical, and structural interactions between Ni films and In0.53Ga0.47 As for source-drain applications in transistor structures have been investigated. It was found that for thick (> 10 nm) Ni films, a steady decrease in sheet resistance occurs with increasing anneal temperatures, however, this trend reverses at 450 degrees C for 5 nm thick Ni layers, primarily due to the agglomeration or phase separation of the Ni-(In,Ga) As layer. A combined hard-x-ray photoelectron spectroscopy (HAXPES) and x-ray absorption spectroscopy (XAS) analysis of the chemical structure of the Ni-(In,Ga)As alloy system shows: (1) that Ni readily interacts with In0.53Ga0.47 As upon deposition at room temperature resulting in significant interdiffusion and the formation of NiIn, NiGa, and NiAs alloys, and (2) the steady diffusion of Ga through the Ni layer with annealing, resulting in the formation of a Ga2O3 film at the surface. The need for the combined application of HAXPES and XAS measurements to fully determine chemical speciation and sample structure is highlighted and this approach is used to develop a structural and chemical compositional model of the Ni-(In,Ga)As system as it evolves over a thermal annealing range of 250 to 500 degrees C.
Content may be subject to copyright.
Investigation of Ni-InGaAs alloy formation by hard x-ray photoelectron spectroscopy
and x-ray absorption spectroscopy
Lee A. Walshand Greg Hughes
Dept. of Physical Sciences, Dublin City University, Glasnevin, Dublin 9, Ireland
Conan Weiland and Joseph C. Woicik
National Institute of Standards and Technology, Gaithesburg, Maryland 20899, USA
Rinus T.P. Lee, Wei-Yip Loh, Pat Lysaght, and Chris Hobbs
SEMATECH, 257 Fuller Road, Suite 2200, Albany, New York 12203, USA
(Dated: November 20, 2014)
The electrical, chemical, and structural interactions between Ni films and InGaAs for source/drain
applications in transistor structures have been investigated. It was found that for thick (>10 nm)
Ni films a steady decrease in sheet resistance occurs with increasing anneal temperatures, however,
this trend reverses at 450 C for 5 nm thick Ni layers, primarily due to the agglomeration/phase
separation of the Ni-InGaAs layer. A combined hard x-ray photoelectron spectroscopy (HAXPES)
and x-ray absorption spectroscopy (XAS) analysis of the chemical structure of the Ni-InGaAs alloy
system shows: (1) that Ni readily interacts with InGaAs upon deposition at room temperature
resulting in significant inter-diffusion and the formation of NiIn, NiGa, and NiAs alloys, and (2) the
steady diffusion of Ga through the Ni layer with annealing, resulting in the formation of a Ga2O3
film at the surface. The need for the combined application of HAXPES and XAS measurements to
fully determine chemical speciation and sample structure is highlighted and this approach is used to
develop a structural and chemical compositional model of the Ni-InGaAs system as it evolves over
a thermal annealing range of 250 to 500 C.
The III-V materials, such as GaAs and InGaAs, show
promise as a Si replacement as the channel material
in n-MOSFETs due to their higher injection veloci-
ties and electron mobilities. Research has recently fo-
cused on InGaAs due to promising improvements in
the InGaAs/high-κinterface1–5, however, the issue of
source/drain (S/D) contacts in InGaAs based MOS-
FETs remains. A possible solution is to find a self-
aligned silicide-like material (salicide), to act as the S/D
contacts6. The optimum material would ideally display
an abrupt ordered interface with InGaAs, low sheet re-
sistance (Rsh), as well as achieving the thermal stability
to withstand the temperatures involved in current MOS-
FET fabrication processes. The search for this material
has recently focussed on the Ni-InGaAs system, due to
its promisingly low Rsh, and its apparent sharp interface
with InGaAs6–8. In addition, investigations have been
performed on the ability to incorporate this material sys-
tem into standard device processing procedures7,9 . There
is a large body of work detailing the chemical interaction
between Ni and GaAs, and Ni and InP, yet there is sig-
nificantly less published data on the Ni-InGaAs system
which is known to form a mixed metallic alloy phase10–13.
Previous studies have shown a trend of decreasing Rsh
with increasing post deposition anneals, although there
is a reversal in this trend between 450 to 500 C7,14.
While it has been suggested that this change could be at-
tributed to the thermal desorption of the III-V elements
at this anneal temperature, more detailed work is needed
to understand this behaviour. The aim of this study is to
address this issue by exploring the details of the chemical
bond formation resulting both from the initial Ni depo-
sition as well as the structural changes which occur after
a range of thermal anneals up to 500 C.
X-ray photoelectron spectroscopy (XPS) is a widely
used technique for the identification of chemical changes
at metal/semiconductor interfaces, and can readily de-
tect the oxidation state of the chemical species present.
Additionally, XPS can be used to quantitatively de-
termine non-destructive depth-dependent composition
due to the well characterized photoelectron attenuation
lengths15. Hard XPS (HAXPES) using x-ray photons
of higher energy than conventional XPS can increase
the effective sampling depth of the photoemission mea-
surement to greater than 20 nm into the material1,16,17.
Chemical speciation in XPS, however, is complicated by
the dependence of peak energies and shapes on elec-
trostatic interactions such as interfacial dipoles or fixed
charges; that is, chemical differences cannot always be
judged by apparent binding energy (BE) shifts alone.
X-ray absorption spectroscopy (XAS) measures the ab-
sorption edges of the individual elements in a mate-
rial to deduce information about their local bonding
environment18. Unlike XPS, fluorescence yield hard x-
ray XAS is unaffected by sample charging, reducing the
uncertainty in chemical speciation. In the analysis of
XAS measurements the spectra can be compared with
those of reference materials thereby assisting in the iden-
tification of the chemical species present. This paper
demonstrates the advantage in combining these two tech-
niques, to obtain a clear chemical and structural model
of a complex material system.
In this study, nickel films of different thicknesses were
sputter deposited on undoped InGaAs substrates and
thermally annealed at a range of temperatures between
250 and 500 C in 50 C steps. These samples were
then investigated using sheet resistance, XAS, scanning
electron microscopy (SEM), transmission electron mi-
croscopy (TEM) and HAXPES measurements in order
to obtain information on the electrical resistivity, the
chemical composition and the physical structure of the
Ni/InGaAs interfacial region. Grazing incidence XAS
measurements were used to identify the chemical phases
present in the Ni/InGaAs interfacial region and com-
pared to more bulk sensitive measurements.
The undoped InGaAs samples consisted of 30 nm thick
In0.53Ga0.47 As layers grown by molecular beam epitaxy
(MBE) on InAlAs epi-ready layers on InP substrates.
The samples were cleaned for 60 seconds in dilute hy-
drofluoric acid prior to being loaded into the metal de-
position chamber. The samples for electrical and TEM
measurements were produced with 5, 15, and 25 nm thick
sputter-deposited Ni layers, and separate samples were
prepared by annealing in 50 C steps between 250 and
500 C for 60 seconds. For HAXPES analysis one sample
was left bare to act as a clean InGaAs reference, while all
other samples had 5 nm of Ni sputter-deposited. One of
these Ni capped samples was left unannealed, while the
remaining samples were annealed in-situ in the deposition
chamber in 50 C steps between 250 and 500 C. In or-
der to study the changes in the relative intensities of the
peaks as a function of thermal annealing, which can as-
sist in determining the extent of inter-diffusion and hence
the location of the chemical species, an internal reference
peak which remains at constant intensity is necessary.
This was achieved by capping all of the HAXPES sam-
ples with a sputter deposited 3 nm SiN capping layer
in a separate chamber, after the Ni deposition and an-
neal, to act both as a barrier to post processing oxida-
tion and as an internal reference. As such the Si 1speak
was acquired at the same time as each core level, with a
constant ratio between the number of scans of the Si 1s
and the relevant core level of 1:4. This was subsequently
used to correct for any photon energy drift during the
measurements, and to normalise the intensity of the core
level spectra across all samples. Using the Si 1sreference
spectra for each core level, the change in peak intensity
of a given chemical species throughout the anneal study
can be determined, and thus the diffusion behaviour of
the different elements can be investigated19 . An equiva-
lent sample set, was prepared using identical conditions
for the XAS measurements. However, the SiN layer was
not present, as a reference layer is not required in XAS.
Electrical sheet resistance (Rsh) measurements were
performed on the annealed samples using a four-point
probe apparatus. The samples for TEM analysis were
prepared using a dual beam FIB equipped with an in-
situ nano-manipulator. The samples were first protected
by applying layers of electron beam deposited C and Pt,
and then a second layer of Pt was deposited using the
ion beam. The samples were thinned to electron trans-
parency using 30 kV and 7 kV Ga ion beams followed
by a 2 kV clean up step. The samples were imaged us-
ing a STEM apparatus operated at 300 kV and the im-
ages were recorded using a beam convergence angle of
8.1 mRad. TEM images were filtered to produce zero-
loss images with a 10 eV energy window.
HAXPES measurements were carried out on the Na-
tional Institute of Standards and Technology (NIST)
beamline (X24A) at the National Synchrotron Light
Source (NSLS) at Brookhaven National laboratory
(BNL). A double Si (111) crystal monochromator allowed
for photon energy selection in the range of 2.1 to 5.0 keV.
An electron energy analyser was operated at a pass en-
ergy of 200 eV giving an overall experimental broadening
of 0.29 eV at the chosen photon energy of 2200 eV, 0.43
eV at 3000 eV, and 0.52 eV at 4050 eV. The total sam-
pling depth of the HAXPES measurement using a photon
energy of 2200 eV is estimated to be 13 nm20 which en-
sures the detection of photoemitted electrons from the
3 nm SiN and 5 nm Ni layers, as well as approximately
6 nm into the InGaAs, which is obtained from the in-
elastic mean free path of the As 2p3/2, Ga 2p3/2, and
In 3d5/2photoemitted electrons at this photon energy21 .
The sampling depth is estimated to be 18 nm and 21
nm for 3000 eV and 4050 eV, respectively. The XPS
core level spectra were curve fitted using Voigt profiles
composed of Gaussian and Lorentzian line shapes with a
Shirley-type background22.
Fluorescence yield hard x-ray XAS measurements were
performed on the As, Ga, and Ni K-edges at room tem-
perature using the NIST beamline (X23A2) at the NSLS.
A double Si (311) crystal monochromator allowed for
photon energy selection in the range of 4.9 to 30 keV.
XAS measurements were performed with the angle be-
tween the incident beam and sample surface at the criti-
cal angle (approximately 0.15) to produce surface sensi-
tive data and at both 0.5(Ga and As K-edges) and 1.2
(Ni K-edge) to sample either the entire Ni film and/or
a more bulk sensitive InGaAs signal. The grazing inci-
dence angle chosen was at a critical angle of 0.15where
the x-ray photon penetration depth into the sample is 2.5
to 5 nm. Below the critical angle total external reflection
occurs. The XAS signal is entirely bulk sensitive at an
incidence angle of 123, where an incidence angle of 90
is normal to the sample surface.
In order for Ni-InGaAs to be incorporated into the pro-
cessing of III-V MOSFETs, it is critical to know the effect
of both anneal temperature and initial Ni thickness (tN i)
on the electrical resistance. The sheet resistance of the
Ni-InGaAs samples over a range of annealing tempera-
tures was determined by four point probe measurements.
Figures 1 (a) and (b), show the plots of Rsh for a 5 nm
and a 15 nm thick Ni layer, respectively, annealed in 50
C steps between 250 and 500 C. A consistent drop in
Rsh with increasing anneal temperature up to 400 C is
observed for the 5 nm deposited layer, with the trend
reversing for higher temperatures. This reversal is not
observed for the 15 nm film which is attributed to the
decreasing influence of Ni-InGaAs interface effects as the
thickness of the deposited layer is increased. The even
lower value of sheet resistance for the 25 nm Ni film fol-
lowing a 400 C anneal evident from Figure 1 (c) confirms
this trend in agreement with previous studies of the Ni-
InGaAs system7–9.
Cross-sectional TEM measurements and top down
SEM images of the 5 nm deposited samples following
the 250, 400, and 500 C anneals were taken in order to
determine the morphological changes in the Ni-InGaAs
layer occurring over the course of the anneal study, and
in particular if any significant difference could be ob-
served following the 500 C anneal. In the TEM image
in Figure 2 (a) for the 250 C anneal sample, a uniform
thickness Ni-InGaAs layer can be clearly distinguished.
When the anneal temperature is increased to 400 C the
Ni-InGaAs layer thickness increases suggesting increased
inter-diffusion. After the 500 C anneal, there is evidence
of significant disruption of the ordered layer structure ap-
parent at lower temperatures and agglomeration/phase
separation is seen in the Ni-InGaAs layer. This effect is
also seen in SEM images, displayed in Figure 2, which
show a substantial increase in surface roughness and ev-
idence of islanding in the 500 C sample as compared to
250 C and 400 C samples. This is a possible reason
for the increase in sheet resistivity seen in Figure 1 (a)
following a 500 C anneal, as the sample surface is no
longer fully covered by a uniform Ni-InGaAs layer due to
the disruption of the layered structure.
In order to further understand the decrease and sub-
sequent increase in Rsh with increasing anneal temper-
ature for the 5 nm deposited Ni film, it is necessary to
determine the changes in the sample chemistry which oc-
cur in the Ni-InGaAs interfacial region probed by XAS
and HAXPES measurements. The normalised HAXPES
spectra of the In 3d5/2, As 2p3/2, and Ga 2p3/2core lev-
els for the SiN capped InGaAs reference sample with no
nickel layer are shown in Figure 3. The spectra are fit-
ted according to the parameters specified by Brennan et
al24 , who compiled a wide variety of reported oxide po-
sitions for the InGaAs core level peaks. It is clear from
the presence of higher BE components in the individ-
ual elemental spectra that the InGaAs surface was par-
tially oxidised due to air exposure prior to SiN deposition.
The corresponding As and Ga absorption K-edges from
the XAS measurements of this sample for both surface
and bulk sensitive modes, shown in Figure 3 (b), reflect
the homogenous composition within the XAS sampling
depth and match reference spectra of GaAs and InAs
samples25,26. The surface sensitive As K-edge spectra
show evidence of oxidation, seen as an increase in the
white line peak intensity (at 11868 eV) compared to the
bulk sensitive spectra, which is consistent with As oxidis-
ing more readily than Ga at GaAs and InGaAs surfaces27.
To understand the change in the chemical structure
of the sample throughout the study, it is necessary to
study both the chemical species present at each anneal
stage and the location and relative concentration of these
chemical species within the sample. As stated earlier the
use of an internal reference peak, such as the Si 1speak
used here, allows the location of each chemical species to
be determined. The Si 1score level spectra acquired from
the SiN capping layers on the bare InGaAs sample and
the Ni-InGaAs sample as-deposited and following a 500
C anneal, as shown in Figure 4, display a very similar
profile, apart from an increase in the higher BE shoulder
at +1.5 eV. The SiN surface would be expected to oxi-
dise after removal from the deposition chamber, forming
a thin layer of Si2N2O or SiO228, however, this would
be expected to be similar for all samples. Therefore, the
increased level of Si oxidation apparent for the samples
with the Ni interlayer would suggest that this oxidation
has occurred at the SiN/Ni interface which will be dis-
cussed later.
Figure 5 (a) shows the normalised (to the Si signal in
the SiN capping layer) and curve fitted Ni 2p3/2HAX-
PES spectra acquired at 2200 eV photon energy for the
SiN-Ni-InGaAs samples as-deposited (20 C) and after
300 C and 400 C anneals. The spectra are fitted with
two peaks, the metal Ni peak at 0 eV, and a broad plas-
mon loss feature at +2.4 eV29. The metallic peak is fitted
using an asymmetric Voigt function with Lorentzian val-
ues of 0.46±0.05 eV, and Gaussian values of 0.55±0.05
eV. The decrease in the intensity of the Ni 2p3/2peak
with successively higher anneals suggests either the dif-
fusion of the Ni into the InGaAs, and/or the diffusion
of the InGaAs substrate elements through the Ni. How-
ever, there is no obvious change in the lineshape of the
Ni 2p3/2throughout the annealing study which indicates
that no strong chemical interaction between the Ni and
InGaAs can be detected by HAXPES measurements of
the Ni 2p.
Figure 5 (b) shows the Fourier transformed Ni K-edge
XAS spectrum of a reference polycrystalline metal Ni
foil and the Ni-InGaAs sample prior to any anneal. It is
clear that the deposited Ni is not in a metallic Ni chem-
ical environment, suggesting that significant intermixing
with the InGaAs has occurred upon deposition. The Ni
K-edge spectra for the as-deposited layer shown in Fig-
ure 5 (c), acquired in both bulk and surface sensitive
modes, match the XAS spectrum of NiGa30 . While the
two stepped structure (with step features at 8334 and
8342 eV) is weaker in the as-deposited film, possibly due
to a mixed NiGa and metallic Ni phase in the reacted
layer, this structure becomes more prominent in both the
bulk and surface spectra after 250 C anneal (not shown)
FIG. 1. Sheet resistance measurements (a) on 5 nm and (b) 15 nm thick Ni layers as a function of post deposition anneal
temperature and (c) as a function of Ni thickness after a 400 C anneal.
FIG. 2. SEM (top) and TEM (bottom) images of 5 nm Ni capped samples after (a) 250 C, (b) 400 C, and (c) 500 C post
deposition anneals showing the increasing thickness of the Ni-InGaAs following the 400 oC anneal and the agglomeration/phase
separation which occurs following the 500 C anneal.
indicating the continued formation of a NiGa phase. The
NiGa formation also indicates that Ga atoms have dif-
fused into the Ni layer at room temperature. Increasing
the anneal temperature to 400 C, as seen in Figure 5 (c)
results in the surface spectrum still resembling a NiGa
alloy, but the bulk spectrum begins to show a change
consistent with a chemically mixed phase of NiGa and
NiAs, or NiAs2, as suggested by the appearance of peaks
at 8344 eV and 8353 eV in the bulk sensitive spectra31.
After a 500 C anneal (not shown) the peaks at 8342 and
8352 eV in the bulk sensitive spectrum which are assigned
to NiAs become more prominent, while the surface sen-
sitive signal also displays the same NiAs features. Even
for the highest anneal temperature, both surface and bulk
sensitive spectra display a mix of NiGa and NiAs bonding
configurations, once again confirming the diffusion of Ga
through the Ni layer, and continued formation of NiAs.
There is no evidence of Ni oxidation in any of the samples.
It is important to note that from the HAXPES measure-
ments no chemical change in the Ni spectra is detected,
while the XAS identifies a large Ni-InGaAs interaction
resulting in a variety of Ni compounds, which evolve as
the anneal temperature increases. This indicates that the
chemical reaction between the Ni and InGaAs initiates
FIG. 3. (a) Normalised and fitted In 3dd5/2, As 2pd3/2, and Ga 2pd3/2HAXPES spectra acquired at a photon energy of 2200
eV for a SiN-InGaAs sample showing surface oxidation of the InGaAs layer. The spectra are plotted relative to the binding
energy of the substrate InGaAs peaks, As-Ga for the As 2pd 3/2,Ga-As for the Ga 2pd 3/2and In-As for the In 3d d 5/2.
(b) XAS spectra of As and Ga K-edges showing the similarity in surface and bulk sensitive spectra of the InGaAs material
consistent with a uniform chemical composition within the XAS sampling depth.
FIG. 4. Normalised Si 1sspectra for the SiN capping layer
with (20 C) and without Ni and after 500 C anneals. The
spectra are plotted relative to the binding energy of the Si-N
upon Ni deposition and the only change following anneal
is the continued reaction and interdiffusion between the
Ni and InGaAs. As the XAS identifies the bonds present
in the surface and bulk of the reacted layer, a difference is
seen following anneal due to the diffusion of Ga through
the sample, leading to the NiGa being located closer to
the sample surface than the NiAs.
Figure 6 (a) shows the HAXPES spectra for the nor-
malised and curve fitted Ga 2p3/2core level acquired at
2200 eV photon energy for the SiN-InGaAs sample and
the SiN-Ni-InGaAs sample, as-deposited (20 C) and af-
ter a 400 C anneal. The spectrum from the SiN-InGaAs
sample is fitted with two peaks, one representing the Ga-
As peak present in the InGaAs bulk, and one which is
attributed to Ga2O due to surface oxidation. The spec-
trum of the as-deposited SiN-Ni-InGaAs sample shows
the growth of a broad peak at +1 eV BE, consistent with
an oxidised Ga state, although it is difficult to be defini-
tive as to the exact stoichiometry of the oxide as the at-
tenuation of the substrate Ga-As peak removes the refer-
ence by which the oxidation state is normally identified32.
In addition to the oxide growth, the lower BE peak at -
0.5 eV is attributed to a Ga-Ni bonding interaction at
the surface based on the respective electronegativities of
both Ga (1.81) and Ni (1.91)33 . The reaction between
Ga and Ni has been studied previously on GaAs and the
observed strong interaction between Ni and Ga12 was at-
tributed to the large enthalpy of formation for NiGa34.
As the anneal temperature increases to 400 C a large
growth in the Ga oxide peak, at +1 eV, is seen, and a
decrease in the Ga-Ni intensity. This assignment is con-
sistent with inter-diffusion between the InGaAs and Ni
layer as has been previously reported for the interaction
between Ni and Ga containing semiconductors34,35. The
subsequent oxidation of the up-diffused Ga, prior to SiN
deposition, results in the formation of a surface oxidised
overlayer which acts to suppress the Ga-Ni signal. The
Ga 2p3/2spectra following the 500 C anneal (not shown)
is similarly dominated by the Ga oxide peak, indicating
the formation of a thick Ga oxide layer prior to the de-
position of the SiN cap.
The bulk sensitive Ga K-edge spectra shown in Fig-
ure 6 (b) for InGaAs samples with and without Ni are
indicative of the gallium atoms being in either a GaAs
or InGaAs bonding environment with the characteristic
two peak structure at 10378 and 10384 eV25. However,
with Ni deposition, the surface sensitive spectrum shown
shows significant changes, and the appearance of a broad
feature at 10380 eV, which could suggest the presence
FIG. 5. (a) Normalised and fitted Ni 2p3/2HAXPES spectra for SiN-Ni-InGaAs samples as-deposited (20 C) and after 300
C and 400 C anneals. The spectra are plotted relative to the binding energy of the main Ni peak. (b) Fourier transformed
XAS spectra of the Ni K-edge, for a reference Ni foil and the Ni signal from the as-deposited (20 C) Ni-InGaAs sample and
(c) XAS spectra from the as-deposited (20 C) Ni-InGaAs sample, after 400 C anneal.
of NiGa or oxidised Ga at the surface of the Ni-InGaAs
film36,37. This additional feature in the spectrum is less
evident in the bulk sensitive spectra shown in Figure 6
(b) again suggesting that this species is localised at the
top of the Ni layer. Figure 6 also shows the spectra for a
Ni-InGaAs sample following a 400 C anneal where the
appearance of a structure (double peak feature at 10379
and 10382 eV) matching Ga2O3is seen, in agreement
with previous studies36–38. Direct evidence for the for-
mation of NiGa bonds cannot be determined from these
spectra due to the fact that the primary feature of NiGa
appears at 10378 eV30 , and thus would be difficult to dis-
tinguish from the Ga2O3signal, although NiGa was pre-
viously detected in the Ni K-edge spectra. These results
suggest that upon Ni deposition some Ga atoms diffuse
through the Ni layer to the surface where they oxidise,
which is assumed to have occurred when the Ni-InGaAs
samples were removed from the vacuum system. The 500
C spectra (not shown) display very similar results to the
400 C spectra. The Ga K-edge data is thus consistent
with the picture provided by the Ni K-edge analysis.
The complementary chemical information derived from
the HAXPES and XAS spectra is therefore particularly
beneficial in the analysis of the Ga 2p3/2. In relation to
the Ga 2p3/2peak profile after Ni deposition the only ev-
idence in the HAXPES spectra that this peak represents
a different chemical state of the Ga is the large increase
in the FWHM (1.38 eV in the oxide, as opposed to 0.78
eV in the Ga-As). The XAS spectra for the same sample
acquired in a surface sensitive mode identifies the Ga sig-
nal to be in an oxidized state which is further confirmed
by the HAXPES and XAS spectra following the 400 C
anneal, where the Ga oxide is identified as Ga2O3. It
is likely that oxidised Ga seen in the 20 C spectra is
composed of a number of suboxide states, and following
anneal at 400 C the greater concentration of Ga at the
Ni-InGaAs surface forms Ga2O3, which is the most stable
of the Ga oxides39.
The formation of Ga2O3at the surface of the Ni-
InGaAs reacted layer can be used to explain the changes
in SiN oxidation seen in Figure 4. The increase in the
SiN-O signal in the as-deposited SiN-Ni-InGaAs sample
is likely due to oxygen gettering by the SiN layer from the
Ga2O3at the SiN/Ni-InGaAs interface. The Gibbs free
energies of SiN oxidation products, either SiO2(G = -802
kJ/mole) or Si2N2O (G = -1063 kJ/mole)28,40, compared
to Ga2O3(G = -998.3 kJ/mole)41 suggest that Si2N2O
could form at the SiN/Ni-InGaAs interface. The fact
that there is no change in the extent of oxidation after a
500 C anneal, where the Ga2O3thickness is known to
increase, indicates that this gettering behaviour is self-
Figure 7 displays the normalised and curve fitted In
3d5/2spectra acquired at 2200 eV for the SiN-InGaAs
sample and the SiN-Ni-InGaAs sample as-deposited (20
C) and after a 400 C anneal. For the reference SiN-
InGaAs sample, the presence of a higher BE component
shifted by +0.46 eV BE with respect to the In-As peak
is indicative of an oxidised surface. A substantial change
occurs upon deposition of the Ni layer resulting in the
appearance of a very broad spectral feature which can
be curve fitted with four component peaks. The two
peaks on the lower BE side of the bulk In-As peak, are
attributed to In-In bonds, at -1 eV and In-Ni bonds at
-0.6 eV, consistent with the electronegativity values of In
(1.78) and Ni33. Previous studies on the interaction of
a deposited Ni layer with the InP surface, have reported
the dissociation of the InP with the appearance of In-
In bonds10–12 . The In-Ni peak at -0.3 eV continues to
grow as the anneal temperature increases. The strong
increase in intensity of the In-Ni peak over the anneal
range is attributed to the diffusion of In into the Ni layer,
forming more In-Ni bonds in the layer below the Ga2O3,
while both the In-In and In-As signals are suppressed
with increasing anneal temperature as both bonds are
localised at the InGaAs interface.
XAS spectra of the In K-edge were recorded, however,
due to the high energy (27000 eV) the peak was very
broad making it hard to obtain information on the In
bonding environment. The interpretation of the Ni K-
edge spectra previously discussed suggests that any Ni-
In formed is at a lower concentration than both NiAs
and NiGa, as both of these alloys dominate the Ni edge
spectra at different temperatures. However, due to the
lack of XAS data, specific In chemical assignments are
necessarily more speculative.
Figure 8 (a) shows the normalised and peak fitted As
2p3/2spectra for the SiN-InGaAs sample and the SiN-
Ni-InGaAs sample as-deposited and after 400 C anneal.
For the SiN-InGaAs sample, the As 2p3/2profile has a
higher BE component consistent with an oxidised sur-
face. Although the overall line-shape does not signifi-
cantly change in the as-deposited SiN-Ni-InGaAs sample,
the substantial shift of +0.9 eV BE relative to the As-Ga
peak, indicates a possible change in chemical state of the
As located at the surface of the InGaAs, consistent with
the formation of NiAs as also seen in the Ni K-edge data.
Throughout the anneal study, the As chemical bonding
environments do not change, however, there is a signifi-
cant reduction in the intensities of the overall As 2p3/2
peak as the anneal temperature is increased. As in the
case for the Ni 2p3/2peak in Figure 5, this is primarily at-
tributed to the formation of the Ga2O3layer above the
InGaAs interface. Additionally, the substantial reduc-
tion in the intensity of the total As profile as a function
of thermal anneal indicates that the As remains primarily
localised at the InGaAs surface throughout the study. As
K-edge spectra were acquired to confirm the sample com-
position profile, as shown in Figure 8 (b). The spectra
of bare InGaAs are characteristic of a clean InGaAs, or
a GaAs As K-edge spectrum, with a double peak struc-
ture at 11870 and 11875 eV25 and the bulk spectra of
the Ni-InGaAs sample show no change with anneal. The
surface sensitive spectra do show a change indicative of
a chemical interaction at the InGaAs surface, however,
it has not been possible to identify the precise chemical
species present, as the layer may consist of a number of
As species, such as As-As, AsNi, and AsOx. Upon an-
nealing at 400 C, there is a significant change in the
surface sensitive spectra with a strong peak emerging at
11868 eV corresponding to NiAs or NiAs2phases31,42 ,
while the bulk sensitive signal shows no change. Follow-
ing the 500 C anneal (not shown), the surface sensitive
spectrum becomes less well defined, while the bulk signal
still resembles InGaAs. This suggests the formation of
NiAs at the Ni-InGaAs interface, in agreement with the
HAXPES results.
The As 3d5/2peak can also be used to determine the
location of the As-Ni reacted phase with respect to the
InGaAs substrate. As such it was measured at higher
photon energies, thereby increasing the sampling depth
from 6 nm into the InGaAs layer for the As 2p3/2peak
acquired at 2200 eV photon energy, to 11, 15, and 20 nm
for the As 3d5/2taken at 2200, 3000 and 4050 eV respec-
tively, as shown in Figure 920. If the proposed structure
of an As-Ni overlayer on the InGaAs substrate is correct,
then a more bulk sensitive measurement should be able
to detect both As-Ni bonds and As-Ga bonds, attributed
to the bulk InGaAs. As can be seen in Figure 9, in the
SiN-Ni-InGaAs sample an As-Ga bonding component can
be identified at 0.48 eV lower BE than the As-Ni com-
ponent peak, with the As-Ga peak area increasing with
increasing photon energy, as the measurement becomes
increasingly bulk sensitive. This confirms the presence of
an As-Ni bonding interaction in a layer at the surface of
the InGaAs substrate.
In order to determine the relative positions of the NiIn
and In-In bonding interactions within the Ni-InGaAs
structure, HAXPES spectra were also acquired at 3000
eV photon energy which increases the sampling depth of
the In 3d5/2from 10 nm at 2200 eV to 13 nm at 3000
eV. From comparison of the normalised spectra following
a 300 C anneal, acquired at the two photon energies,
as shown in Figure 10, the change in relative intensity
between the In-In and In-Ni can be seen as the photon
energy is varied. The results indicate that the In-In is
located below the In-Ni, as in the more bulk sensitive
(3000 eV) measurement the In-In intensity increases rel-
ative to the In-Ni. A similar comparison for the Ga 2p3/2
spectra of the as-deposited SiN-Ni-InGaAs sample, where
the sampling depth is increased from 7 nm at 2200 eV
to 11 nm at 3000 eV, demonstrates the Ga-Ni intensity
increasing relative to the Ga oxide in the more bulk sensi-
tive measurement, indicating the Ga-Ni is located below
FIG. 6. (a) Normalised and fitted Ga 2p3/2HAXPES spectra, and (b) Ga K-edge XAS spectra, for a SiN-InGaAs sample and
SiN-Ni-InGaAs samples as-deposited (20 C), and after a 400 C anneal. The HAXPES spectra are plotted relative to the
binding energy of the substrate Ga-As peak.
FIG. 7. Normalised and fitted In 3d5/2spectra for sample
with (20 C) and without Ni, and after 400 C anneal. The
spectra are plotted relative to the binding energy of the sub-
strate In-As peak.
the Ga oxide in the interface structure.
In order to produce an accurate model of the sample
structure throughout the anneal study, it is important
to understand the location of each chemical species. To
this end the photoionisation cross section and inelastic
mean free path (IMFP) normalised HAXPES peak ar-
eas, referenced with respect to the 3 nm SiN capping
layer, are plotted as a function of anneal temperature for
each sample in Figure 11. The error bars in Figure 11
are taken from the discrepancy between the experimental
data and the theoretical fit acquired using the XPS fit-
ting software. The steady decrease in the Ni 2p3/2signal
intensity as a function of anneal temperature is consis-
tent with both the diffusion of Ni into the InGaAs layer
as well as the growth of an overlayer of Ga2O3. Con-
versely, the increase of the Ga signal over the course of
the anneal study agrees with the continual up-diffusion
of Ga through the Ni layer, forming a Ga2O3overlayer
between the SiN and Ni layers. The Ga area begins to
decrease after the 500 C anneal, which can be explained
by the agglomeration of the Ni-InGaAs layer at this tem-
perature, resulting in the unreacted substrate InGaAs
moving toward the sample surface around the agglomer-
ated region. A steady decrease in both the In 3d5/2and
As 2p3/2signals up to 400 C agrees with both of these
elements out-diffusing less than the Ga and remaining
closer to the InGaAs surface. The increases seen in both
peaks between 450 and 500 C are due to the Ni-InGaAs
agglomeration at these temperatures which significantly
disrupts the overall film structure. The segregation of
Ga to the top of the interacted layer has also been seen
in studies of Ni-GaAs, where Ni2GaAs is detected upon
deposition and this segregates into NiGa near the sam-
ple surface, and NiAs near the GaAs substrate, following
thermal anneal13,43. The much lower formation energy
of Ga2O3as opposed to Ni oxides is consistent with the
conversion of the surface localised NiGa into a Ga2O3
surface layer41.
Combining the results from both XAS and HAXPES
measurements, a model of the Ni-InGaAs system over the
course of the anneal study can be produced. All of the
chemical interactions observed appear to initiate upon
Ni deposition, contrary to previous results8, and only
the volume of the reacted layers changes as the anneal
temperature increases as Ni continues to diffuse into the
InGaAs layer. This expansion of the reacted Ni-InGaAs
layer results in the trend of decreasing sheet resistance as
a function of temperature seen in Figure 1. The HAX-
PES and XAS experimental data can be used to con-
struct a schematic model of the sample structure prior
to Ni deposition, upon Ni deposition, and after a 400
C anneal, as shown in Figure 12. The diverse range
FIG. 8. (a) Normalised and fitted As 2p3/2HAXPES spectra for a SiN-InGaAs sample and SiN-Ni-InGaAs samples as-
deposited (20 C) and after a 400 C anneal (b) As K-edge XAS spectra, for a SiN-InGaAs sample and SiN-Ni-InGaAs samples
as-deposited (20 C), after a 400 C anneal. The HAXPES spectra are plotted relative to the binding energy of the substrate
As-Ga peak.
FIG. 9. Normalised and fitted As 3d5/2spectra for an as-
deposited SiN-Ni-InGaAs sample taken at 2200, 3000, and
4050 eV photon energy. The spectra are plotted relative to
the binding energy of the substrate As-Ga peak.
of the chemical species formed coupled with their phys-
ical location in relation to the original Ni-InGaAs inter-
face makes this a challenging experimental study which
necessitated the enhanced sampling depth of HAXPES
and the definitive chemical species identification capa-
bilities of XAS. Previous studies have described a very
abrupt Ni-InGaAs/InGaAs interface, and constant com-
position throughout the Ni-InGaAs layer6–8. The results
in this study are contrary to these findings, as signifi-
cant diffusion of certain species throughout the anneal
study indicate a graded layered structure within the re-
acted region. These previous studies used secondary ion
mass spectroscopy (SIMS) to characterise the diffusion
profile of each element as well as the abruptness of the
FIG. 10. Normalised and fitted Ga 2p3/2spectra for an as-
deposited (20 C) SiN-Ni-InGaAs sample and In 3d5/2spectra
for a SiN-Ni-InGaAs sample after a 300 C anneal, acquired
at 2200 and 3000 eV photon energy.
interface. While SIMS is a powerful method in deter-
mining the diffusion profile for a given element, the mea-
surement in itself can cause intermixing and it does not
provide chemical speciation. The key point in this study
is that the combination of HAXPES measurements to
identify reactions between Ni and InGaAs, and the dif-
fusion profile of the elements throughout the Ni-InGaAs
layer, and XAS measurements to provide chemical speci-
ation allows the chemical compositional structure of this
complex material system to be investigated. As minimis-
ing the source/drain contact resistance is a key require-
ment for low power devices, the ability to correlate the
chemical and electrical measurements facilitates the de-
velopment of a more comprehensive understanding of the
FIG. 11. Plots of the normalised HAXPES peak areas as a
function of temperature, of (a) As 2p3/2, (b) Ga 2p3/2, (c)
In 3d5/2, and (d) Ni 2p3/2peaks. All areas have been nor-
malised by photo-ionisation cross-section, and inelastic mean
free path.
relationship between the chemical compositional profile
and the conduction properties.
The change in chemical composition, physical struc-
ture and the accompanying change in resistivity of a Ni-
InGaAs interface was studied as a function of post depo-
sition anneal temperature. It was found that while the
resistivity steadily decreases as the anneal temperature
increases for thick films, thinner films show a reversal
of this trend at 450 C due to the agglomeration/phase
separation of the Ni-InGaAs film. The decrease in sheet
resistance with anneal temperature is due to the increase
in thickness of the Ni-InGaAs interaction layer containing
Ni-In, Ni-As and Ni-Ga bonds. The analysis of both the
chemical and structural composition of Ni-InGaAs con-
tacts has been shown to significantly benefit from the use
of two complementary measurement techniques. While
XAS can readily identify the chemical species present,
HAXPES is required to determine the relative concen-
trations and diffusion trends throughout the annealing
study. The combination of XAS and HAXPES analysis is
necessary to fully describe the chemical composition and
structure of the Ni-InGaAs layer. The Ni-InGaAs model
derived from the experimental results reveals a highly
reactive and inter-diffused interface with substantive Ga
out-diffusion and the formation of Ni-In and Ni-As alloy
phases closer to the Ni-InGaAs interface. These studies
provide a detailed understanding of this material system
and have potential application for fabrication strategies
for S/D contacts in future InGaAs MOSFETs.
The authors from Dublin City University acknowl-
edge the financial support of SFI under Grant Num-
ber: SFI/09/IN.1/I2633. Access to the X24A HAXPES
beamline, and the X23A2 XAS beamline at Brookhaven
National Laboratory was obtained through a General
User Proposal. Use of the National Synchrotron Light
Source, Brookhaven National Laboratory, was supported
by the U.S. Department of Energy, Office of Science, Of-
fice of Basic Energy Sciences, under Contract No. DE-
1Patrick S. Lysaght, Joel Barnett, Gennadi I. Bersuker,
Joseph C. Woicik, Daniel A. Fischer, Brendan Foran,
Hsing-Huang Tseng, and Raj Jammy. Chemical analysis
of HfO2/Si (100) film systems exposed to NH3thermal
processing. Journal of Applied Physics, 101(2):024105–
024105–9, January 2007.
2Y. T. Chen, J. Huang, J. Price, P. Lysaght, D. Veksler,
C. Weiland, J.C. Woicik, G. Bersuker, R. Hill, J. Oh, P.D.
Kirsch, R. Jammy, and J.C. Lee. III-V gate stack interface
improvement to enable high mobility 11nm node CMOS. In
2012 International Symposium on VLSI Technology, Sys-
tems, and Applications (VLSI-TSA), pages 1–2, 2012.
3J. Huang, N. Goel, P. Lysaght, D. Veksler, P. Nagaiah,
S. Oktyabrsky, J. Price, H. Zhao, Y.T. Chen, J.C. Lee, J.C.
Woicik, P. Majhi, P.D. Kirsch, and R. Jammy. Detailed
high-κ/In0.53Ga0.47 As interface understanding to enable
improved In0.53Ga0.47As gate stack quality. In 2011 In-
ternational Symposium on VLSI Technology, Systems and
Applications (VLSI-TSA), pages 1–2, 2011.
4Han Zhao, Jeff Huang, Yen-Ting Chen, Jung Hwan Yum,
Yanzhen Wang, Fei Zhou, Fei Xue, and Jack C. Lee. Ef-
fects of gate-first and gate-last process on interface qual-
ity of In0.53Ga0.47As metal-oxide-semiconductor capacitors
using atomic-layer-deposited Al2O3and HfO2oxides. Ap-
plied Physics Letters, 95(25):253501–253501–3, December
5Jonathon B. Clemens, Sarah R. Bishop, Joon Sung
Lee, Andrew C. Kummel, and Ravi Droopad. Initia-
tion of a passivated interface between hafnium oxide and
In(Ga)As(001)(42). The Journal of Chemical Physics,
132(24):244701–244701–8, June 2010.
6Xingui Zhang, Huaxin Guo, Xiao Gong, Qian Zhou, You-
Ru Lin, Hau-Yu Lin, Chih-Hsin Ko, Clement H. Wann, and
Yee-Chia Yeo. In0.53 Ga0.47 As channel n-MOSFET with
self-aligned NiInGaAs source and drain. Electrochemical
and Solid-State Letters, 14(2):H60–H62, February 2011.
7L. Czornomaz, M. El Kazzi, M. Hopstaken, D. Caimi,
P. Mchler, C. Rossel, M. Bjoerk, C. Marchiori, H. Sieg-
wart, and J. Fompeyrine. CMOS compatible self-aligned
S/D regions for implant-free InGaAs MOSFETs. Solid-
State Electronics, 74:71–76, August 2012.
8Ivana, Yong Lim Foo, Xingui Zhang, Qian Zhou, Jisheng
Pan, Eugene Kong, Man Hon Samuel Owen, and Yee-
FIG. 12. Schematic diagram showing the deduced Ni-InGaAs sample structure and composition before and after (20 C) Ni
deposition and following 400 C anneal.
Chia Yeo. Crystal structure and epitaxial relationship of
Ni4InGaAs2films formed on InGaAs by annealing. Jour-
nal of Vacuum Science & Technology B: Microelectronics
and Nanometer Structures, 31(1):012202, 2013.
9Xingui Zhang, Ivana, Hua Xin Guo, Xiao Gong, Qian
Zhou, and Yee-Chia Yeo. A self-aligned Ni-InGaAs contact
technology for InGaAs channel n-MOSFETs. Journal of
The Electrochemical Society, 159(5):H511–H515, January
10 L. J. Brillson, C. F. Brucker, A. D. Katnani, N. G. Stoffel,
and G. Margaritondo. Atomic and electronic structure of
InPmetal interfaces: A prototypical IIIV compound semi-
conductor. Journal of Vacuum Science and Technology,
19(3):661–666, 1981.
11 T. Kendelewicz, M. D. Williams, W. G. Petro, I. Lindau,
and W. E. Spicer. Interfacial chemistry and schottky-
barrier formation of the Ni/InP(110) and Ni/GaAs(110)
interfaces. Physical Review B, 32(6):3758–3765, 1985.
12 R.H. Williams, A. McKinley, G.J. Hughes, V. Mont-
gomery, and I. T. McGovern. Metal-GaSe and metal-
InP interfaces: Schottky barrier formation and interfa-
cial reactions. Journal of Vacuum Science and Technology,
21(2):594–598, 1982.
13 M.D. Williams, W.G. Petro, T. Kendelewicz, S.H. Pan,
I. Lindau, and W.E. Spicer. An exploratory study of the
reactive Ni-GaAs (110) interface. Solid State Communica-
tions, 51(10):819–822, September 1984.
14 R.T.P. Lee, R.J.W. Hill, W.-Y. Loh, R.-H. Baek, S. De-
ora, K. Matthews, C. Huffman, K. Majumdar, T. Micha-
lak, C. Borst, P.Y. Hung, C.-H. Chen, J.-H. Yum, T.-W.
Kim, C.Y. Kang, Wei-E. Wang, D.-H. Kim, C. Hobbs,
and P.D. Kirsch. VLSI processed InGaAs on Si MOS-
FETs with thermally stable, self-aligned Ni-InGaAs con-
tacts achieving: Enhanced drive current and pathway to-
wards a unified contact module. In Electron Devices Meet-
ing (IEDM), 2013 IEEE International, pages 2.6.1–2.6.4,
December 2013.
15 S. Hofmann. Quantitative depth profiling in surface anal-
ysis: A review. Surface and Interface Analysis, 2(4):148–
160, August 1980.
16 Keisuke Kobayashi. High-resolution hard x-ray photoelec-
tron spectroscopy: Application of valence band and core-
level spectroscopy to materials science. Nuclear Instru-
ments and Methods in Physics Research Section A: Ac-
celerators, Spectrometers, Detectors and Associated Equip-
ment, 547(1):98–112, July 2005.
17 K. Kakushima, K. Okamoto, K. Tachi, J. Song, S. Sato,
T. Kawanago, K. Tsutsui, N. Sugii, P. Ahmet, T. Hattori,
and H. Iwai. Observation of band bending of metal/high-κ
Si capacitor with high energy x-ray photoemission spec-
troscopy and its application to interface dipole measure-
ment. Journal of Applied Physics, 104(10):104908–104908–
5, November 2008.
18 Dale E. Sayers, Edward A. Stern, and Farrel W. Ly-
tle. New technique for investigating noncrystalline struc-
tures: Fourier analysis of the extended x-ray absorption
fine structure. Physical Review Letters, 27(18):1204–1207,
November 1971.
19 C. Weiland, P. Lysaght, J. Price, J. Huang, and J. C.
Woicik. Hard x-ray photoelectron spectroscopy study of as
and ga out-diffusion in In0.53Ga0.47 As/Al2O3film systems.
Applied Physics Letters, 101(6):061602, August 2012.
20 S. Tanuma, C. J. Powell, and D. R. Penn. Calculations of
electron inelastic mean free paths. IX. data for 41 elemental
solids over the 50 eV to 30 keV range. Surface and Interface
Analysis, 43(3):689–713, 2011.
21 Jill Chastain and John F Moulder. Handbook of X-ray
photoelectron spectroscopy : a reference book of standard
spectra for identification and interpretation of XPS data.
Physical Electronics, Eden Prairie, Minn., 1995.
22 D. A. Shirley. High-resolution x-ray photoemission spec-
trum of the valence bands of gold. Physical Review B,
5(12):4709–4714, June 1972.
23 Patrick S. Lysaght, Joseph C. Woicik, Jeff Huang,
Jungwoo Oh, Byoung-Gi Min, and Paul D. Kirsch.
Process driven oxygen redistribution and control in
Si0.7Ge0.3/HfO2/TaN gate stack film systems. Journal of
Applied Physics, 110(8):084107, October 2011.
24 Barry Brennan. Surface and interface characterisation of
high-κdielectric materials on silicon and III-V semicon-
ductor substrates. doctoral, Dublin City University. School
of Physical Sciences, March 2010.
25 G. Dalba, D. Diop, P. Fornasini, A. Kuzmin, and F. Rocca.
EXAFS and XANES study of GaAs on Ga and As K-edges.
Journal of Physics: Condensed Matter, 5(11):1643, March
26 M. G. Proietti, S. Turchini, J. Garcia, G. Lamble,
F. Martelli, and T. Prosperi. Glancing-angle extended
x-ray absorption fine structure study of strained In-
GaAs/GaAs heterostructures. Journal of Applied Physics,
78(11):6574–6583, 1995.
27 P. Pianetta, I. Lindau, C. M. Garner, and W. E. Spicer.
Oxidation properties of GaAs (110) surfaces. Physical Re-
view Letters, 37(17):1166–1169, October 1976.
28 S. I. Raider, R. Flitsch, J. A. Aboaf, and W. A. Pliskin.
Surface oxidation of silicon nitride films. Journal of The
Electrochemical Society, 123(4):560–565, April 1976.
29 Andrew P. Grosvenor, Mark C. Biesinger, Roger St.C.
Smart, and N. Stewart McIntyre. New interpretations of
XPS spectra of nickel metal and oxides. Surface Science,
600(9):1771–1779, May 2006.
30 Li-Shing Hsu, Y.-K. Wang, and G. Y. Guo. Experimental
and theoretical study of the electronic structures of Ni3Al,
Ni3Ga, Ni3In, and NiGa. Journal of Applied Physics,
92(3):1419–1424, August 2002.
31 Anna Puig-Molina, Lars Pleth Nielsen, Alfons M. Molen-
broek, and Konrad Herbst. In situ EXAFS study on the
chemical state of arsenic deposited on a NiMoP/Al2O3hy-
drotreating catalyst. Catalysis Letters, 92(1-2):29–34, Jan-
uary 2004.
32 C. L. Hinkle, M. Milojevic, B. Brennan, A. M. Sonnet, F. S.
Aguirre-Tostado, G. J. Hughes, E. M. Vogel, and R. M.
Wallace. Detection of Ga suboxides and their impact on
III-V passivation and fermi-level pinning. Applied Physics
Letters, 94(16):162101, April 2009.
33 David R. Lide. Handbook of Chemistry and Physics: 2004-
2005. CRC PressI Llc, June 2004.
34 L. J. Brillson, C. F. Brucker, N. G. Stoffel, A. D. Katnani,
and G. Margaritondo. Abruptness of semiconductor-metal
interfaces. Physical Review Letters, 46(13):838–841, March
35 S. Huang, B. Shen, F. Lin, N. Ma, F. J. Xu, Z. L. Miao,
J. Song, L. Lu, F. Liu, Y. Wang, Z. X. Qin, Z. J. Yang,
and G. Y. Zhang. Ni diffusion and its influence on electri-
cal properties of AlxGa1xN/GaN heterostructures. Applied
Physics Letters, 93(17):172102–172102–3, October 2008.
36 Emiel J. M. Hensen, Mayela Garca-Snchez, Neelesh Rane,
Pieter C. M. M. Magusin, Pang-Hung Liu, Kuei-Jung
Chao, and R. A. van Santen. In situ Ga K-edge XANES
study of the activation of Ga/ZSM-5 prepared by chemi-
cal vapor deposition of trimethylgallium. Catalysis Letters,
101(1-2):79–85, May 2005.
37 Ken-ichi Shimizu, Atsushi Satsuma, and Tadashi Hattori.
Metal oxide catalysts for selective reduction of NOxby
hydrocarbons: Toward molecular basis for catalyst design.
Catalysis Surveys from Japan, 4(2):115–123, July 2001.
38 Seong-Kyun Cheong, Bruce A. Bunker, D. C. Hall, G. L.
Snider, and P. J. Barrios. XAFS and x-ray reflectivity
study of IIIV compound native oxide/GaAs interfaces.
Journal of Synchrotron Radiation, 8(2):824–826, March
39 D. J. Fu, Y. H. Kwon, T. W. Kang, C. J. Park, K. H.
Baek, H. Y. Cho, D. H. Shin, C. H. Lee, and K. S. Chung.
GaN metaloxidesemiconductor structures using Ga-oxide
dielectrics formed by photoelectrochemical oxidation. Ap-
plied Physics Letters, 80(3):446–448, January 2002.
40 F.L. Riley. Silicon nitride and related materials. Journal
of the American Ceramic Society, 83(2):245–265, 2000.
41 Donald D. Wagman, William H. Evans, Vivian B. Parker,
Richard H. Schumm, Iva Halow, Sylvia M. Bailey, Ken-
neth L. Churney, and Ralph L. Nuttall. The NBS Tables
of Chemical Thermodynamic Properties, volume 11. Na-
tional Bureau of Standards.
42 Shunsuke Sakai, Akira Yoshiasa, Hiroshi Arima, Maki
Okube, Chiya Numako, and Tsutomu Sato. XAFS study
of As in K-T boundary clays. AIP Conference Proceedings,
882(1):274–276, February 2007.
43 Masaki Ogawa. Alloying reaction in thin nickel films de-
posited on GaAs. Thin Solid Films, 70(1):181–189, July
... 10 Contact resistance values of 1.1 to 4.7 X lm 2 and the ability to incorporate Mo-InGaAs into current device processing techniques have been previously demonstrated. [10][11][12][13][14] However, unlike one of the other candidate material systems, Ni-InGaAs, 15 there has been no definitive study on the interface chemistry and stability as a function of anneal temperature. The aim of this preliminary work is to study the chemical and electrical stability of the Mo/InGaAs system at room temperature and following a range of post deposition thermal anneals between 250 and 500 C. ...
... These results are in contrast to the Ni-InGaAs system, which is an alternative candidate for source/drain contacts, where substantial fluctuation of R sh is seen as a function of anneal temperature. 15,[17][18][19] In order to explore the chemical stability of the Mo-InGaAs system, the interfacial region between the Mo and InGaAs was probed by HAXPES and XAS measurements. Characteristic HAXPES spectra are shown in Figure 2. ...
... These results are in contrast to a previous study of the Ni-InGaAs system which describes a large extent of inter-diffusion between the Ni and InGaAs as a function of thermal anneal, with the Ga diffusing through the Ni to the sample surface. 15 This comparison further demonstrates the improved thermal stability of Mo-InGaAs over Ni-InGaAs. ...
Full-text available
The electrical and chemical stability of Mo-InGaAs films for source-drain applications in transistor structures has been investigated. It was found that for 5 nm thick Mo films, the sheet resistance remains approximately constant with increasing anneal temperatures up to 500 °C. A combined hard x-ray photoelectron spectroscopy and x-ray absorption spectroscopy analysis of the chemical structure of the Mo-InGaAs alloy system as a function of annealing temperature showed that the interface is chemically abrupt with no evidence of inter-diffusion between the Mo and InGaAs layers. These results indicate the suitability of Mo as a thermally stable, low resistance source-drain contact metal for InGaAs-channel devices.
... Continuous engineering of contacts compatible with state-of-the-art semiconductor technology relies upon a detailed understanding of the critical relationships between processing conditions, interface chemistry and structure, and contact performance [1]. Silicides [2] and salicides [3,4] exhibit a broad spectrum of composition-dependent contact resistances (R c ) and have long been employed as standard, low resistance contacts in traditional (Si, Ge) and compound (e.g. InGaAs) semiconductorbased CMOS technologies. ...
Full-text available
Sc has been employed as an electron contact to a number of two-dimensional (2D) materials (e.g. MoS 2 , black phosphorous) and has enabled, at times, the lowest electron contact resistance. However, the extremely reactive nature of Sc leads to stringent processing requirements and metastable device performance with no true understanding of how to achieve consistent, high-performance Sc contacts. In this work, WSe 2 transistors with impressive subthreshold slope (109 mV dec ⁻¹ ) and I ON / I OFF (10 ⁶ ) are demonstrated without post-metallization processing by depositing Sc contacts in ultra-high vacuum (UHV) at room temperature (RT). The lowest electron Schottky barrier height (SBH) is achieved by mildly oxidizing the WSe 2 in situ before metallization, which minimizes subsequent reactions between Sc and WSe 2 . Post metallization anneals in reducing environments (UHV, forming gas) degrade the I ON / I OFF by ~10 ³ and increase the subthreshold slope by a factor of 10. X-ray photoelectron spectroscopy indicates the anneals increase the electron SBH by 0.4–0.5 eV and correspondingly convert 100% of the deposited Sc contacts to intermetallic or scandium oxide. Raman spectroscopy and scanning transmission electron microscopy highlight the highly exothermic reactions between Sc and WSe 2 , which consume at least one layer RT and at least three layers after the 400 °C anneals. The observed layer consumption necessitates multiple sacrificial WSe 2 layers during fabrication. Scanning tunneling microscopy/spectroscopy elucidate the enhanced local density of states below the WSe 2 Fermi level around individual Sc atoms in the WSe 2 lattice, which directly connects the scandium selenide intermetallic with the unexpectedly large electron SBH. The interface chemistry and structural properties are correlated with Sc–WSe 2 transistor and diode performance. The recommended combination of processing conditions and steps is provided to facilitate consistent Sc contacts to WSe 2 .
... 1 Transition metal silicides 2 in the case of Si, or salicides in the case of compound 3D semiconductors (e.g. InGaAs), 3,4 have long been the industry standard contacts in conventional CMOS technology. These materials exhibit phase and stoichiometry dependent electronic properties, which are tunable with carefully designed processing conditions. ...
Palladium has been widely employed as a hole contact to WSe2 and has enabled, at times, the highest WSe2 transistor performance. However, there are orders of magnitude variation across the literature in Pd–Se2 contact resistance and ION/IOFF ratios with no true understanding of how to consistently achieve high–performance contacts. In this work, WSe2 transistors with impressive ION/IOFF ratios of 10⁶ and Pd–WSe2 Schottky diodes with near–zero variability are demonstrated utilizing Ohmic–like Pd contacts through deliberate control of the interface chemistry. The increased concentration of a PdSex intermetallic is correlated with an Ohmic band alignment and concomitant defect passivation, which further reduces the contact resistance, variability, and barrier height inhomogeneity. The lowest contact resistance occurs when a 60 minute post metallization anneal at 400 °C in forming gas (FG) is performed. X-ray photoelectron spectroscopy indicates this FG anneal produces 3× the concentration of PdSex and an Ohmic band alignment, in contrast to that detected after annealing in ultra–high vacuum, during which a 0.2 eV hole Schottky barrier forms. Raman spectroscopy and scanning transmission electron microscopy highlight the necessity of the fabrication step to achieve high–performance contacts as no PdSex forms and WSe2 is unperturbed by room temperature Pd deposition. However, at least one WSe2 layer is consumed by the necessary interface reactions that form PdSex requiring strategic exploitation of a sacrificial WSe2 layer during device fabrication. The interface chemistry and structural properties are correlated with Pd–WSe2 diode and transistor performance and the recommended processing steps are provided to enable reliable high–performance contact formation.
... 3,54 Ni has been predicted by density functional theory to show significant reaction with other inert materials such as MoS2, 55 and similar interactions have previously been used to form alloyed source/drain contacts in III-V materials. [56][57] Figure 11a shows the Bi 5d and Se 3d spectra for a Bi2Se3 sample after Ni deposition, along with bulksensitive AR spectra. In the Bi 5d spectrum, a new peak appears at lower BE after Ni deposition. ...
Full-text available
The interface between the topological insulator Bi2Se3 and deposited metal films is investigated using x-ray photoelectron spectroscopy including conventional contact metals (Au, Pd, Cr, and Ir) and magnetic materials (Co, Fe, Ni, Co0.8Fe0.2, and Ni0.8Fe0.2). Au is the only metal to show little or no interaction with the Bi2Se3, with no interfacial layer between the metal and the surface of the TI. The other metals show a range of reaction behaviors with the relative strength of reaction (obtained from the amount of Bi2Se3 consumed during reaction) ordered as: Au < Pd < Ir < Co ≤ CoFe < Ni < Cr < NiFe < Fe, in approximate agreement with the behavior expected from the Gibbs free energies of formation for the alloys formed. Post metallization anneals at 300°C in vacuum were also performed for each interface. Several of the metal films were not stable upon anneal and desorbed from the surface (Au, Pd, Ni, and Ni0.8Fe0.2), while Cr, Fe, Co, and Co0.8Fe0.2 showed accelerated reactions with the underlying Bi2Se3, including inter-diffusion between the metal and Se. Ir was the only metal to remain stable following anneal, showing no significant increase in reaction with the Bi2Se3. This study reveals the nature of the metal-Bi2Se3 interface for a range of metals. The reactions observed must be considered when designing Bi2Se3 based devices.
... FETs 151,152,153 . However, N-type dopants in InGaAs come from either group IV or group VI elements. ...
The economic health of the semiconductor industry requires substantial scaling of chip power, performance, and area with every new technology node that is ramped into manufacturing in two year intervals. With no direct physical link to any particular design dimensions, industry wide the technology node names are chosen to reflect the roughly 70% scaling of linear dimensions necessary to enable the doubling of transistor density predicted by Moore’s law and typically progress as 22nm, 14nm, 10nm, 7nm, 5nm, 3nm etc. At the time of this writing, the most advanced technology node in volume manufacturing is the 14nm node with the 7nm node in advanced development and 5nm in early exploration. The technology challenges to reach thus far have not been trivial. This review addresses the past innovation in response to the device challenges and discusses in-depth the integration challenges associated with the sub-22nm non-planar finFET technologies that are either in advanced technology development or in manufacturing. It discusses the integration challenges in patterning for both the front-end-of-line and back-end-of-line elements in the CMOS transistor. In addition, this article also gives a brief review of integrating an alternate channel material into the finFET technology, as well as next generation device architectures such as nanowire and vertical FETs. Lastly, it also discusses challenges dictated by the need to interconnect the ever-increasing density of transistors.
... It has been demonstrated that all metals are pinned toward the conduction band of InGaAs. This leads to low electron barriers for N-type contacts in InGaAs FETs 150,151,152 . However, N-type dopants in InGaAs come from either group IV or group ...
The economic health of the semiconductor industry requires substantial scaling of chip power, performance, and area with every new technology node that is ramped into manufacturing in two year intervals. With no direct physical link to any particular design dimensions, industry wide the technology node names are chosen to reflect the roughly 70% scaling of linear dimensions necessary to enable the doubling of transistor density predicted by Moore’s law and typically progress as 22nm, 14nm, 10nm, 7nm, 5nm, 3nm etc. At the time of this writing, the most advanced technology node in volume manufacturing is the 14nm node with the 7nm node in advanced development and 5nm in early exploration. The technology challenges to reach thus far have not been trivial. This review addresses the past innovation in response to the device challenges and discusses in-depth the integration challenges associated with the sub-22nm non-planar finFET technologies that are either in advanced technology development or in manufacturing. It discusses the integration challenges in patterning for both the front-end-of-line and back-end-of-line elements in the CMOS transistor. In addition, this article also gives a brief review of integrating an alternate channel material into the finFET technology, as well as next generation device architectures such as nanowire and vertical FETs. Lastly, it also discusses challenges dictated by the need to interconnect the ever-increasing density of transistors. Read More:
... These applications are made possible by HAXPES due to the large photoelectron IMFPs that assure significant signal from the bulk of the sample itself. 31, 32 The enhanced bulk sensitivity also leads to the ability to study samples with intentional surface and/or overlayer modification, either for chemical protection (see, e.g., Ref. 33, where a passive SiN capping layer was used to prevent oxidation of Ni films deposited on InGaAs), or for in situ and in operando measurements. ...
Recent applications of hard x-ray photoelectron spectroscopy (HAXPES) demonstrate its many capabilities in addition to several of its limitations. Examples are given, including measurement of buried interfaces and materials under in situ or in operando conditions, as well as measurements under x-ray standing-wave and resonant excitation. Physical considerations that differentiate HAXPES from photoemission measurements utilizing soft x-ray and ultraviolet photon sources are also presented.
... delaminate and the contact resistance increases rapidly. Similar contact resistance degradation at around this temperature has also been observed in the Ni-InGaAs contact system [19], [20]. In 6 sets of test structures consisting of a total of 72 CTLMs, the average ρc obtained after 350 °C anneal is 4.5•10 -8 Ω•cm 2 with a standard deviation of 3.1•10 -8 Ω•cm 2 . ...
Full-text available
We demonstrate ultralow ohmic contact resistance to antimonide-based, p-channel quantum-well field-effect transistor (QW-FET) structures using a new p±-InAs/InAsSb cap structure. The incorporation of a p±-InAsSb layer enables the use of a thicker cap with lower sheet resistance, resulting in an improved contact resistivity. Using a Pd-based ohmic scheme, the composite cap structure resulted in a 4x reduction in contact resistance compared with a standard p±-InAs cap. This translates into nearly 3x improvement in the gm of fabricated InGaSb p-channel QW-FETs. Furthermore, Ni contacts on the composite cap were fabricated and a contact resistance of 45 Ω · μm was obtained. An accurate contact resistivity extraction in this very low range is possible through nanotransmission line models with sub-100 nm contacts. In devices of this kind with Ni-based contacts, we derive an ultralow contact resistivity of 5.2 · 10-8 Ω · cm2.
Full-text available
We derive an analytical formula for the electron inelastic mean free path (IMFP) from its definition within the dielectric formalism. The parameters in this formula are determined solely by the optical energy-loss function of the material of interest. This formula is valid for electrons of energy larger than 500 eV, including relativistic electrons.
Conference Paper
Full-text available
Parasitic resistance (Rpara) is a grand challenge to successfully hetero-integrate III-V channels onto Si for CMOS application. Here, we report the first statistical IDsat comparison for non-self-aligned and self-aligned contacts of In0.53Ga0.47As MOSFETs fabricated on large scale Si substrates with VLSI toolsets. We compare non-self-aligned Mo and self-aligned Ni-InGaAs contacts. Devices with self-aligned contacts exhibit a 25% enhancement in IDsat over devices with non-self-aligned contacts largely due to the 27% reduction in Rpara. We have also extended the thermal stability of Ni-InGaAs to 500 °C (highest reported) enabling it to be compatible with BEOL processes. The impact of the Ni-InGaAs process module on tool contamination is discussed. These results represent significant progress towards establishing a path to a unified Ni-based S/D contact module for Si/SiGe/Ge/III-V co-integration on VLSI platforms.
Full-text available
The structural, compositional, and electrical properties of epitaxial Ni4InGaAs2 (denoted as Ni-InGaAs) film formed by annealing sputtered Ni film on InGaAs were investigated. It was found that Ni-InGaAs adopts a NiAs (B8) structure with lattice parameters of a = 0.396 ± 0.002 nm and c = 0.516 ± 0.002 nm, and exhibits an epitaxial relationship with InGaAs, with orientations given by Ni-InGaAs[110]//InGaAs[001] and Ni-InGaAs[110]//InGaAs[110]. The epitaxial Ni4InGaAs2 film has bulk electrical resistivity of ∼102 μΩ·cm, which increases as the film thickness scales below 10 nm. The results of this work would be useful for the development of contact metallization for high mobility InGaAs metal-oxide-semiconductor field-effect transistors.
Silicon nitride has been researched intensively, largely in response to the challenge to develop internal combustion engines with hot-zone components made entirely from ceramics. The ceramic engine programs have had only partial success, but this research effort has succeeded in generating a degree of understanding of silicon nitride and of its processing and properties, which in many respects is more advanced than of more widely used technical ceramics. This review examines from the historical standpoint the development of silicon nitride and of its processing into a range of high-grade ceramic materials. The development of understanding of microstructure-property relationships in the silicon nitride materials is also surveyed. Because silicon nitride has close relationships with the SiAlON group of materials, it is impossible to discuss the one without some reference to the other, and a brief mention of the development of the SiAlONs is included for completeness.
A salicide-like self-aligned Ni-InGaAs contact technology suitable for InGaAs metal-oxide-semiconductor field-effect transistors has been developed. It has been confirmed that Ni film sputtered onto single crystalline InGaAs substrate is uniformly converted into Ni-InGaAs by low temperature (250-400 degrees C) rapid thermal annealing. A comprehensive study of Ni-InGaAs contacts was performed by employing characterization of High Resolution Transmission Electron Microscopy, Energy-dispersive X-ray spectroscopy, X-Ray Diffraction and Secondary Ion Mass Spectroscopy. The electrical properties of the contacts were also characterized. Sheet resistance mapping of Ni-InGaAs thin film by microscopic 4-point probe shows a uniform sheet resistance of 21.3 Omega/square and low resistivity of similar to 96 mu Transfer Length Method test structure shows Ni-InGaAs has a contact resistance of 1.27 on n(+) InGaAs doped by Si+ implant. This self-aligned Ni-InGaAs contact technology was then used in the experimental fabrication of InGaAs channel n-MOSFETs. Devices with self-aligned metallic Ni-InGaAs S/D as well as Si-doped S/D with Ni-InGaAs contacts were realized and show good electrical characteristics with on-state/off-state drain current ratio of 10(3)similar to 10(5).
ESCA is used to characterize silicon nitride surface oxidation. Si 2p, N 1s, and O 1s binding energies and photoelectron line intensities of oxidized nitride films are compared with the corresponding lines from thick reference films of silicon, silicon nitride, silicon dioxide, and a series of oxynitrides. Rapid initial oxidation of silicon nitride surfaces occurs at room temperature on exposure of nitride films to air. A graded oxidized nitride film forms between the film surface and the nitride. Similarly, oxynitride films with gradations in composition are obtained upon oxidation of nitride films at high temperatures.
Recommended values are provided for chemical thermodynamic properties of inorganic substances and for organic substances usually containing only one or two carbon atoms. Where available, values are given for the enthalpy of formation, Gibbs energy of formation, entropy, and heat capacity at 298.15 K (25 C), the enthalpy difference between 298.15 and 0 K and the enthalpy of formation at 0 K. All values are given in SI units and are for a standard state pressure of 100 000 pascal. This volume is a new collective edition of 'Selected Values of Chemical Thermodynamic Properties,' which was issued serially as National Bureau of Standards Technical Notes 270-1 (1965) to 270-8 (1981). Values are given for properties of gaseous, liquid and crystalline substances, for solutions in water, and for mixed aqueous and organic solutions. Values are not given for alloys or other solid solutions, fused salts or for substances of undefined composition. Compounds of the transuranium elements are not included. (Author)
Hard x-ray photoelectron spectroscopy (HAXPES) was performed on In0.53Ga0.47As/Al2O3 gate stacks as deposited and annealed at 400 °C, 500 °C, and 700 °C to test for out-diffusion of substrate elements. Ga and As core-level intensities increase with increasing anneal temperature, while the In intensity decreases. HAXPES was performed at two different beam energies to vary the surface sensitivity; results demonstrate Ga and As out-diffuse into the Al2O3 film. Analysis suggests the presence of an interlayer containing Ga and As oxides, which thickens with increasing anneal temperature. Further diffusion, especially of Ga, into the Al2O3 film is also observed with increasing anneal temperature.
CMOS compatible self-aligned access regions for indium gallium arsenide (In0.53Ga0.47As) implant-free n-type metal–oxide–semiconductor field effect transistors (MOSFETs) are investigated. In situ doped n+ source/drain regions are selectively grown by metal-organic vapor phase epitaxy and self-aligned Nickel–InGaAs alloyed metal contacts are obtained using a self-aligned silicide-like process, where different process conditions are studied. Soft pre-epitaxy cleaning is followed by X-ray photoelectron spectroscopy, while the Ni–InGaAs/III–V interface is characterized by back-side SIMS profiling. Relevant contact and sheet resistances are measured and integration issues are highlighted. Gate-first implant-free self-aligned n-MOSFETs are produced to quantify the impact of Ni–InGaAs contacts on the device performance.
Soft X-ray photoemission spectroscopy measurements have been carried out on cleaved n-type GaAs (1 1 0) surfaces covered with Ni overlayers ranging in thickness from 0.05 to 53 Å. The results of these room temperature measurements show that we have band bending effects occurring in conjunction with strong interfacial chemical reactions. Deconvolution of the Ga 3d core line into substrate and metallic components shows dissolution of the substrate at the interface with Ga diffusing into the surface of the metal overlayer for the intermediate coverages (1–15 Å). Observation of the As 3d core level shows out-diffusion of As to the surface over the entire Ni coverage range. Using this deconvolution scheme we are able to follow the band bending of the Schottky barrier formed here up to the 8 Å coverage. The Schottky barrier height is 1.0 ± 0.1 eV for this overlayer thickness.