U.S. patent application number 14/161745 was filed with the patent office on 2015-07-23 for method and structure to enhance gate induced strain effect in multigate device.
This patent application is currently assigned to International Business Machines Corporation. The applicant listed for this patent is International Business Machines Corporation. Invention is credited to Veeraraghavan S. Basker, Pranita Kerber, Junli Wang, Tanko Yamashita, Chun-chen Yeh.
Application Number | 20150206953 14/161745 |
Document ID | / |
Family ID | 53545542 |
Filed Date | 2015-07-23 |
United States Patent
Application |
20150206953 |
Kind Code |
A1 |
Basker; Veeraraghavan S. ;
et al. |
July 23, 2015 |
METHOD AND STRUCTURE TO ENHANCE GATE INDUCED STRAIN EFFECT IN
MULTIGATE DEVICE
Abstract
A FinFet formed by depositing a thin layer of polycrystalline
silicon followed by depositing a stress containing material,
including a high Ge percentage silicon germanium film and/or a high
stress W film on top of a polycrystalline silicon film. Freeing
space between fins allows stressor films to be deposited closer to
the transistor channel, improving the proximity of the stress
containing material to the transistor channel and enhancing the
stress coupling efficiency by defining a ratio between stress level
in the stressor film and stress transferred to the channel for a
mobility enhancement. The stress level is enhanced by patterning by
removing the n-type workfunction metal from the p-FinFET. After
stripping off the soft or hard mask, the p-type workfunction metal
is deposited in the n- and p-FinFET regions. The freed space
specifically for p-FinFet between the fins achieves an even higher
stressor coupling to further boost the carrier mobility.
Inventors: |
Basker; Veeraraghavan S.;
(Schenectady, NY) ; Kerber; Pranita; (Mount Kisco,
NY) ; Wang; Junli; (Slingerlands, NY) ;
Yamashita; Tanko; (Schenectady, NY) ; Yeh;
Chun-chen; (Clifton Park, NY) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
International Business Machines Corporation |
Albany |
NY |
US |
|
|
Assignee: |
International Business Machines
Corporation
Albany
NY
|
Family ID: |
53545542 |
Appl. No.: |
14/161745 |
Filed: |
January 23, 2014 |
Current U.S.
Class: |
257/192 ;
438/283 |
Current CPC
Class: |
H01L 29/7845 20130101;
H01L 27/0924 20130101; H01L 29/7846 20130101; H01L 21/845 20130101;
H01L 21/823821 20130101; H01L 29/785 20130101; H01L 29/66795
20130101 |
International
Class: |
H01L 29/66 20060101
H01L029/66; H01L 29/78 20060101 H01L029/78; H01L 29/165 20060101
H01L029/165 |
Claims
1. A method of forming a semiconductor device structure comprising:
forming fins on a substrate; forming a gate stack over said fins,
leaving a source region and a drain region exposed; forming a thin
layer of polycrystalline silicon over said fins; epitaxially
growing a high GE percentage silicon germanium film on top of said
polycrystalline silicon; and forming a transistor channel by
wrapping said gate stack around said fins.
2. The method as recited in claim 1, further comprising: patterning
a workfunction metal forming sequentially an n-type and p-type
transistor, and depositing a stress Tungsten (W) film in a space
freed by patterning between said fins.
3. The method as recited in claim 1, further comprising depositing
a high-k dielectric material in said gate stack.
4. The method of claim 3, wherein said high-k dielectric material
is made of HfO.sub.2, Al.sub.2O.sub.3, or La.sub.2O.sub.3, and said
deposition is followed by performing a post deposition anneal.
5. The method of claim 1 wherein said workfunction metal provides a
threshold voltage of said transistor producing a drive current for
switching a circuit.
6. The method of claim 1 further comprising: depositing a thin
layer of polycrystalline silicon; depositing a stress film made of
high Ge percentage silicon germanium film or a highly stress
Tungsten film on top of said polycrystalline silicon layer; freeing
a space between said fins to position said stress films deposited
closer to said transistor channel to improve a proximity of said
stress containing material to said transistor channel; and
enhancing a stress coupling efficiency by way of a ratio between a
stress level in said stress film, and said stress transferred to
said transistor channel for mobility enhancement.
7. The method of claim 6 wherein said stress is enhanced by
patterning to remove an n-type workfunction metal from a p-type
transistor.
8. The method of claim 7, further comprising stripping off a soft
or a hard mask of said deposited p-type workfunction metal.
9. The method of claim 7 wherein freeing said space for said p-type
transistor between said fins achieves a higher stressor and better
coupling from said stressor to said channel.
10. The method of claim 1, wherein growing said high Ge percentage
silicon germanium film on top of said polycrystalline silicon is
doped in-situ or ex-situ, said dopant lowering a gate
resistance.
11. The method of claim 10, wherein said in-situ doped film is
achieved by incorporating said dopants during said silicon
germanium epitaxial growth, and wherein said ex-situ doped film is
achieved by an ion-implantation followed by said epitaxial growing
of said silicon germanium film.
12. The method of claim 1, wherein a stress coupling efficiency is
determined by a ratio between a stress level in said stressor film
and a stress transferred to said channel for mobility
enhancement.
13. The method of claim 8 further comprising stripping a mask of
said p-type workfunction metal deposited in both n-type and p-type
MOSFET regions, and wherein freeing said space for said p-type
transistor between said fins provides a higher stressor coupling,
and boosts a carrier mobility.
13. A FinFET device structure comprising: a substrate provided with
fins; a gate stack on top of said fins and an exposed source region
and an exposed drain region; a thin layer of polycrystalline
silicon on top of said fins; a high Ge percentage silicon germanium
film on top of said polycrystalline silicon; and a transistor
channel with said gate stack wrapped around said fins.
14. The FinFet device structure of claim 13 further comprising: a
high-k dielectric material on a patterned fin structure with a
recess shallow trench isolation (STI); and a workfunction metal and
a thin layer of a high percentage Ge polycrystalline silicon on top
of said high-k dielectric material.
15. The FinFet device structure of claim 13, wherein a space
between said fins and stressor films are in a proximity to said
channel.
16. The FinFet device structure of claim 14 further comprising a
stripped mask of said p-type workfunction metal in both n-type and
p-type transistor regions, wherein a freed space for said p-type
transistor between said fins provides a higher stressor coupling,
for boosting carrier mobility.
Description
FIELD OF THE INVENTION
[0001] The present invention relates to the fabrication of
semiconductor devices, and more particularly, to the formation of a
gate stack of an n-type FET and a p-type FET to enhance the strain
level in the channel to generate carrier mobility and drive current
performance benefits in multigates or trigates.
BACKGROUND AND RELATED ART
[0002] Current techniques of forming gate stack in the FinFET
device structure starts with depositing high-k metal dielectric
material such as HfO.sub.2, Al.sub.2O.sub.3, or La.sub.2O.sub.3 by
ALD process. The process can be done either following the fin
formation, referenced as the "gate first" scheme (also referenced
as the metal inserted polysilicon), or it can be done after the
formation of the source and drain in a process referred to as "gate
last", (also referenced as "replacement metal gate" (RMG) scheme).
The gate first and gate last refer to whether a metal electrode is
deposited before or after the high temperature activation anneals
of the flow. It is then followed by a workfunction metal deposition
to set the transistor threshold voltage. Typical materials that are
employed include TiN, TiC, or TiAl.
[0003] Since the requirement for a threshold voltage between an
n-type MOSFET transistor and p-type MOSFET differs, the gate stack
formation can involve a certain patterning process, e.g., to
deposit either one type of workfunction metal first on both n-type
and p-type FETs, followed by removing it from one side, and then
depositing the desired workfunction metal to set the right
threshold voltage for both types of transistors. The next step
consists in depositing the gate contact material to lower the gate
resistance. In the gate first scheme, the process is followed by an
offset space deposition, and an epitaxial process to form the
source and drain of the transistors. While in the gate last (RMG)
scheme, the process is followed by the middle-of-the-line (MOL)
source/drain contact formation such as by a contact etch, TiN
barrier layer and W film deposition.
[0004] As the device continues to the nanometer scaling, the
requirement for the transistor drive current performance becomes
increasingly more difficult to meet. One difficulty that
encountered resides in the conventional stress effect, such as an
embedded silicon germanium, embedded silicon carbon source drain,
and dual stress silicon nitride liner set to boost the carrier
mobility that diminishes significantly with the scaling of the gate
contact pitch.
[0005] The process typically starts with the recess of a silicon
source and the drain region, followed by a SiGe or SiC film
epitaxial growth in the trench region. Typically, it consists of
three different layers, the first layer referenced as the buffer
layer which is usually lightly doped to provide the junction
gradient control and device short channel electrostatic benefit.
Then, the process continues with the main layer heavily doped to
lower the source and drain sheet resistance. A typical dopant is
boron for a p-type MOSFET, and phosphorus or arsenic for the n-type
transistor. The incorporation of the germanium for a p-type MOSFET
and carbon for the n-type MOSFET is by introducing the strain from
the source and drain region to the channel region. The fact that
germanium atom has a larger lattice constant than silicon atom
produces a tensile strain in the SiGe film itself when it creates
bonding with the silicon atom. The stress is then transferred to
the channel region to generate a compressive strain that is
favorable for the transport of holes and thus enhances the p-type
MOSFET drive current. The effect of the carbon atom is just the
opposite. The fact that carbon atom has a smaller lattice constant
than silicon atom produces a compressive strain in the SiC film
itself when it creates bonding with the silicon atom. The stress is
then transferred to the channel region to generate the tensile
strain that is favorable for the transport of electrons and thus
enhances the n-type MOSFET drive current. The stress effect from
the embedded source drain either is made of SiGe or SiC that drops
significantly when the device gate pitch continues to scale. The
pitch scaling is necessary because of the chip area reduction set
to improve the cost structure of the semiconductor. However,
because of the pitch scaling, the available space for the source
and drain stressor formation becoming smaller and smaller, this
leading to the decrease of the stressor volume and a greatly
reduction of the stress effect from the source and the drain
region. The stress effect has been estimated to be about only
20-30% left when the technology feature size shrinks from the 22 nm
node technology to the 14 nm node technology.
[0006] Another factor that limits the usage of source and drain
extrinsic stressor is the transition from the planar device
structure to the FinFFT type of the transistor structure. The fact
that FinFET device is provided of a thin slab of silicon as the
channel (typical dimensions are 10 nm width and 30 nm tall) which
signifies that the area for stress coupling is much reduced
compared to the planar device structure which makes the stress
coupling efficiency, defined as the ratio between stress level in
the source and drain region to the stress level in the device
channel region that becomes significantly less compared to the
planar device. The reduction of the aforementioned two effects
(stressor volume and stress coupling) makes the stress engineering
of the FinFET device structure ever more challenging compared to
that of previous generations.
[0007] The challenge for the gate induced strain, however, is how
to implement it through the selection of gate stack materials to
meet not only the channel strain requirement for carrier mobility
boost, but also a workfunction setting to meet the transistor
threshold voltage requirement. It further has to have minimized
gate stack inversion thickness impact so that the transistor drive
current benefit from gate induced strain will not be compromised.
Gate stack materials such as TiN, TaN, and TiC have a certain level
of strain itself, preferably in the range of 2 GPa to 3 GPa which
is quite significant in terms of producing channel strain when the
coupling ratio is sufficiently high. The contact material such as W
is also known for having a tensile strain favorable for the
transistor mobility enhancement.
[0008] The challenge is, as previously mentioned, how to design the
gate stack thickness and structure such that the stress benefit can
be maximized, but at the same time, without affecting other device
parametrics, such as the transistor threshold voltage and the
inversion gate stack thickness.
SUMMARY
[0009] In one aspect, embodiments of the invention provide a method
and a FinFET device structure that form a gate stack stress of a
finFET to increase the mobility and drive current.
[0010] In an embodiment, a method is described of forming a FinFET
that includes depositing high-k dielectric on a patterned fin
structure with recess shallow trench isolation. A workfunction
metal and a very thin layer of poly crystalline silicon are then
deposited on top of the high-k materials. A stress containing
material such as high Ge percentage silicon germanium film and/or
highly stress W film is then deposited on top of the poly
crystalline silicon film. In the case of a high Ge percentage
silicon film, the film can be formed either in-situ or ex-situ
doped with dopant to lower the gate resistance. The in-situ doped
film can be achieved by incorporating dopants during the silicon
germanium epitaxial growth, whereas the ex-situ doped film can be
achieved by ion-implantation following a silicon germanium
epitaxial growth. Preferably, the gate stack has thick
poly-crystalline silicon on top of the workfunction metal that
occupies the space between the fins, leaving no room for strain
producing material to impart a stress in the transistor channel for
mobility enhancement.
[0011] In an embodiment, a method and structure are described
freeing up the space between the fins to allow the stressor films
being deposited closer to the channel, thus improving the proximity
of the stress containing material to the transistor channel,
thereby enhancing the stress coupling efficiency defined as the
ratio between the stress level in the stressor film and the stress
transferred to the channel for mobility enhancement.
[0012] In an embodiment, a method is described forming a
semiconductor device that includes depositing high-k dielectric on
the patterned fin structure with recess shallow trench isolation. A
workfunction metal for n-type MOSFET is deposited on top of high-k
materials. Next, a patterning process can remove the n-type
workfunction metal from the p-type MOSFET. After stripping off the
soft or the hard mask, a p-type workfunction metal is deposited in
both the n-type and p-type MOSFET regions. The stress containing
material can be a high percentage Ge silicon germanium film and/or
a highly stress W film is then deposited on top of the
poly-crystalline silicon film.
[0013] In an embodiment, the structure is provided with fins on a
substrate and includes workfunction metal and a highly stressed
Tungsten (W), wherein the workfunction metal and the highly
stressed W create a strain in the channel, with the workfunction
metal wrapping around the fins. Notably, a highly stress W would
fail to achieve the same results.
BRIEF DESCRIPTION OF THE DRAWINGS
[0014] The foregoing and other objects, features and advantages
described herein will be apparent from the following more
particular descriptions of example embodiments as illustrated in
the accompanying drawings wherein like reference numbers generally
represent like parts of exemplary embodiments.
[0015] FIG. 1 is a cross-sectional view of a semiconductor device
having a substrate and fin structures formed thereon;
[0016] FIG. 2 shows a cross-sectional view of a semiconductor
device having a substrate and fin structures formed thereon, with a
shallow trench isolation oxide deposited between the fins;
[0017] FIG. 3 is a cross-sectional view of a semiconductor device
structure on a substrate where fin structures are formed with a
shallow trench isolation oxide recessed between the fins;
[0018] FIG. 4 illustrates a cross-sectional view of a semiconductor
device having a substrate and fin structures formed and covered
with a high-k dielectric;
[0019] FIG. 5 is a cross-sectional view of a semiconductor device
having a substrate and fin structures formed and covered with
high-k dielectric and a workfunction metal;
[0020] FIG. 6 shows a cross-sectional view of a semiconductor
device having a substrate and fin structures formed and covered
with high-k dielectric, a workfunction metal, to which is added a
very thin layer of poly-crystalline silicon film, according to an
embodiment of the invention;
[0021] FIG. 7 is a cross-sectional view of a semiconductor device
having a substrate and a plurality of fin structures formed and
covered with a high-k dielectric, a workfunction metal, a very thin
layer of poly-crystalline silicon film, and a high Ge percentage
silicon germanium film or highly stress W film, according to one
embodiment of the invention;
[0022] FIG. 8 illustrates a cross-sectional view of a semiconductor
device structure having a substrate and fin structures formed with
shallow trench isolation oxide recessed in between the fins;
[0023] FIG. 9 shows another cross-sectional view of a semiconductor
device having a substrate and fin structures formed and covered
with high-k dielectric;
[0024] FIG. 10 is a cross-sectional view of a semiconductor device
having a substrate and fin structures formed and covered with
high-k dielectric and n-type MOSFET workfunction;
[0025] FIG. 11 depicts a cross-sectional view of a semiconductor
device having a substrate and fin structures formed and covered
with high-k dielectric and an n-type MOSFET workfunction, with an
n-type MOSFET region being covered by soft or hardmask for
pattering, according to an embodiment of the invention;
[0026] FIG. 12 shows a cross-sectional view of a semiconductor
device having a substrate and fin structures formed and covered
with high-k dielectric with n-type MOSFET workfunction etched away
in the p-type MOSFET region according to another embodiment of the
invention;
[0027] FIG. 13 is a cross-sectional view of a semiconductor device
having a substrate and fin structures formed thereon and covered
with a high-k dielectric, an n-type MOSFET workfunction and p-type
MOSFET according to a further embodiment of the invention;
[0028] FIG. 14 illustrates a cross-sectional view of a
semiconductor device having a substrate and fin structures formed
thereon covered with high-k dielectric, n-type MOSFET workfunction,
p-type MOSFET and a high Ge percentage silicon germanium film or a
highly stress Tungsten film, according to still another embodiment
of the invention; and
[0029] FIG. 15 depicts a cross-sectional view of multiple layers of
gate stack in both, an n-type and a p-type MOSFET region to explain
the structure provided with a potential mobility boost from the
gate stack.
DETAILED DESCRIPTION
[0030] It will be appreciated that for simplicity and clarity of
illustration, elements shown in the drawings are not necessarily
drawn to scale. For example, the dimensions of some of the elements
may be exaggerated relative to other elements for clarity. As
previously stated, the present disclosure relates to a
semiconductor structure including locally thinned semiconductor
fins, and a method for manufacturing the same. Aspects of the
present disclosure will now be described in detail with
accompanying figures. It is noted that like reference numerals
refer to like elements across different embodiments. As used
herein, ordinals such as "first" and "second" are employed merely
to distinguish similar elements, and different ordinals may be
employed to designate a same element in the specification and/or
claims.
[0031] In a first exemplary semiconductor structure according to a
first embodiment of the present disclosure can be formed by
providing a semiconductor substrate, which can be a bulk
semiconductor substrate or a semiconductor-on-insulator (SOI)
substrate. At least an upper portion of the semiconductor substrate
includes a semiconductor material, which can be selected from
elemental semiconductor materials (e.g., silicon, germanium,
carbon, or alloys thereof), III-V semiconductor materials, or II-VI
semiconductor materials. In one embodiment, the semiconductor
substrate can include a single crystalline semiconductor
material.
[0032] Generally, an upper portion of the semiconductor substrate
can be patterned by a combination of lithographic methods and an
anisotropic etch used to form a plurality of semiconductor fins.
The plurality of fins can include two outermost semiconductor fins
and at least one nested semiconductor fin located between the two
outermost semiconductor fins.
[0033] As used herein, the semiconductor fin refers to a
semiconductor material portion having a pair of parallel vertical
sidewalls laterally spaced by a uniform dimension. In one
embodiment, each semiconductor fin can have a rectangular
horizontal cross-sectional area such that the spacing between the
pair of parallel vertical sidewalls is the same as the length of
the shorter sides of the shape of the rectangular horizontal
cross-sectional area. As used herein, a fin field effect transistor
(finFet) refers to a field effect transistor in which at least a
channel region is located within a semiconductor fin. As used
herein, an "outermost semiconductor fin" refers to a semiconductor
fin within a plurality of semiconductor fins located at an
outermost location. Furthermore, a "nested semiconductor fin"
refers to a semiconductor fin within a plurality of semiconductor
fins located between the two outermost semiconductor fins of the
plurality of semiconductor fins.
[0034] By way of example, a photoresist layer (not shown) can be
applied over the top surface of the semiconductor substrate and
lithographically patterned to mask portions of the semiconductor
substrate, in which a plurality of semiconductor fins is
subsequently formed. The pattern in the photoresist layer can be
transferred to an upper portion of the semiconductor substrate to
form the plurality of semiconductor fins. If the semiconductor
substrate is a bulk substrate, the remaining portion of the
semiconductor substrate underlying the plurality of semiconductor
fins is referred to as a semiconductor material layer. In this
case, the semiconductor material layer is a substrate on which the
semiconductor fins are formed. The semiconductor material layer
functions as a substrate mechanically supporting the plurality of
semiconductor fins. The plurality of semiconductor fins and the
semiconductor material layer collectively constitute a contiguous
semiconductor material portion. In one embodiment, the entirety of
the contiguous semiconductor material portion can be single
crystalline. Alternatively, if the semiconductor substrate is a
semiconductor-on-insulator (SOI) substrate, a vertical stack of a
buried insulator layer and a handle substrate layer can be present
underneath the plurality of semiconductor fins in lieu of the
semiconductor material layer. In this case, the vertical stack of
the buried insulator layer and the handle substrate layer is the
substrate on which the semiconductor fins are formed.
[0035] The height of the semiconductor fins can range from 5 nm to
1,000 nm, although lesser and greater heights can also be employed.
The plurality of semiconductor fins and the semiconductor material
layer can be doped with electrical dopants, i.e., p-type dopants or
n-type dopants, or be intrinsic.
[0036] Referring now to FIG. 1, a fin formation, preferably on a
bulk substrate (106) is illustrated. The substrate material can be
either silicon, silicon germanium, or III-V materials such as
InGaAs, InAs or GaN. The fin is preferably formed by a sidewall
image transfer (SIT) process to produce small dimensional features
beyond the capability of the current lithography tools. The silicon
substrate is first deposited with the amorphous silicon, then
patterned to form mandrels. Next, the SIT spacer is formed by first
depositing either a silicon oxide (Si.sub.2) or silicon nitride
Si.sub.3N.sub.4 layer, followed by etching the SIT spacer, leaving
material only on the side of the mandrels. The mandrels are then
pulled away by wet etch process, leaving the SIT spacer standing
alone on top of the silicon substrate, serving as a hardmask (102).
A silicon fin etch process is then performed to fabricate silicon
fins with the dimension of 10 nm wide and 30 nm deep (104).
[0037] FIG. 2 illustrates a shallow trench isolation material (STI)
deposition (108) that can be formed among the plurality of
semiconductor fins. The shallow trench isolation layer includes a
dielectric material such as silicon oxide or silicon nitride or a
combination thereof, referenced as a hybrid STI with good
conformality and gap fill capability to fill the space between the
fins. The shallow trench isolation layer can be formed by
depositing a dielectric material over the semiconductor fins and
the semiconductor material layer. The deposition of the dielectric
material can be performed, for example, by chemical vapor
deposition (CVD) or spin coating. Excess portions of the deposited
dielectric material can be removed from above the top surfaces of
the semiconductor in, e.g., by planarization preferably employing a
chemical mechanical planarization (CMP). The shallow trench
isolation layer laterally surrounds the plurality of semiconductor
fins. The top surface of the shallow trench isolation layer can be
coplanar with the top surfaces of the plurality of semiconductor
fins. The material lying above the fin cap (102) is then polished,
as previously described.
[0038] Referring to FIG. 3, an STI etch back follows. The etching
process can be achieved by a dry etch process with chemistry to
etch away either the silicon oxide or the silicon nitride or a
combination thereof (i.e., hybrid STI). The target etch depth is 30
nm, so that the same number fin (104) is exposed above the
remaining STI (108).
[0039] FIG. 4 shows a high-k material deposition (110). The
material serves as the gate dielectric to provide a high-k value to
achieve a better device drive current benefit, but without
increasing the physical thickness which is prone to cause
additional leakage current. Typical high-k materials include HfO2,
Al2O3, and La2O3. The process is followed by performing a
post-deposition anneal at 700.degree. C. in a furnace to densify
the film following the deposition. Usually because of the oxygen
effect, there is a 1 nm interfacial SiOx layer formed between the
high-k material and the silicon substrate. The deposition is
achieved by an atomic layer deposition (ALD) to achieve the
thickness control and good conformality from the top of the fin to
the bottom.
[0040] FIG. 5 is an illustration of the deposition of the
workfunction metal (112). A typical workfunction metal can consists
of TiN, TaN, TiAlN, and TiC. The deposition is preferably achieved
by the atomic layer deposition (ALD) to obtain thickness control
and good conformality from the top of the fin to the bottom. The
purpose of the workfunction metal is to provide the right threshold
voltage of the transistor so that the device can produce enough
drive current for switching the circuit.
[0041] FIG. 6 is an illustration of the deposition of a very thin
layer of polycrystalline silicon (114) with dimensions preferably
ranging between 5-10 nm. It is deposited first in the furnace as
amorphous silicon, and then annealed at a 1025.degree. C. spike
temperature to form the crystalline structure. The purpose of the
thin polycrystalline silicon layer is to provide a template of
silicon atoms to facilitate the ensuing silicon germanium epitaxial
process.
[0042] FIG. 7 is an illustration of an epitaxial growth of silicon
germanium film (116) atop of the polycrystalline silicon. The
epitaxial process can be preferably achieved in an epitaxial
chamber, starting with a pre-clean by HF or pre-bake to ensure an
oxide free surface, followed by silane and other ambient with
carrier gas to grow the high Ge concentration silicon germanium
films. The Ge percentage can advantageously rise to 52%. The
purpose of using SiGe as the gate contact material is because the
Ge has a larger lattice constant compared to silicon, indicative of
the SiGe film having a tensile strain in the film itself.
Simulation results shows that with 35% Ge silicon germanium film,
the intrinsic stress level is 2 GPa. Stress can be induced to the
channel to achieve 500 MPa along the fin length and a fin height
direction and 400 MPa compressive stress along the fin width
direction. When placed atop and on the sidewall of the fin channel,
it produces strain in the channel region which is favorable for the
n-type MOSFET electron carrier transport. The SiGe film also needs
to be heavily doped with either boron or phosphorus up to 5e20/cm3
to ensure that the gate resistance is not an issue during the logic
circuit or RF microwave transistor operation.
[0043] FIG. 8 illustrates an embodiment (200) starting from the STI
etch back process. The etching can be achieved by a dry etch
process with chemistry etching away either the silicon oxide or
silicon nitride or the combination thereof (i.e., a hybrid STI
scheme). The target etch depth can be 30 nm, such that the same
number of fins (204) is exposed above the remaining STI (208).
[0044] FIG. 9 illustrates the high-k material deposition (210). The
material serves as the gate dielectric to provide a higher-K value
for better device drive current benefit but without increasing the
physical thickness that may cause additional leakage current. The
typical high-k material used can include HfO2, Al2O3, and La2O3.
The process proceeds preferably with a post-deposition anneal at
700.degree. C. in the furnace to densify the film after deposition.
Usually, because of the oxygen effect, there is a 1 nm interfacial
SiOx layer formed between the high-k material and the silicon
substrate. Deposition is achieved by the atomic layer deposition
(ALD) to achieve thickness control and good conformality from the
top of the fin to the bottom.
[0045] FIG. 10 illustrates the deposition of workfunction metal
(212). A typical workfunction metal includes TiN, TaN, TiAlN, and
TiC. The deposition is achieved by an atomic layer deposition (ALD)
to achieve thickness control and good conformality from the top of
the fin to its bottom. The purpose of the workfunction metal is to
provide a right threshold voltage of the transistor to ensure that
the device can produce enough drive current for switching the
circuit.
[0046] FIG. 11 is an illustration of the gate stack patterning
process. The structure is first covered with either a soft or a
hard mask (216). It then proceeds going through a photo-lithography
process to initiate the patterning so that the n-FET region is
covered while the p-FET region opens up.
[0047] FIG. 12 is an illustration of the etch of n-type MOSFET
workfunction metal (212) from the p-type MOSFET region. Most of the
workfunction metal shows a certain level of strain but only
favorable for one type of transistor carrier transport. By way of
example, a TaN film has a tensile strain up to 2 GPa which is
beneficial to the electron transport but is detrimental to the hole
transport. Furthermore, experimentally, from several dark field
holography strain measurement results, one can observe the TiN,
either by ALD (25A) or PVD (140A) that induces a significant strain
in the silicon channel. The `Fin-only` structure shows no strain.
An amorphous silicon deposition added to annealing induces a strain
in the opposite direction to TiN. The amorphous silicon layer
itself induces no strain. Based on the aforementioned results, the
most significant contributors are the TiN deposition and the
annealing of the amorphous silicon. The purpose of removing the
n-type workfunction metal not only provides the right threshold
voltage in the p-type MOSFET region, but it also reduces the impact
of an unfavorable strain in the p-type MOSFET region. Another added
benefit is to free the space between the fins as only a single
layer workfunction metal instead of two to be deposited. This leads
to a potentially large volume of the ensuing W gate contact to
further enhance the strain effect using W as a film known for
strain generation.
[0048] FIG. 13 is an illustration of the deposition of p-type
MOSFET workfunction metal (214) in the p-type MOSFET region. As
previously mentioned, the patterning process enables individual
tuning of a strain level and a type of strain in both n-type and
p-type MOSFET.
[0049] FIG. 14 is an illustration of the deposition of W gate
contact (218) in both, the n-type and the p-type MOSFET regions. As
previously mentioned, W is a film known for generating a strain,
but the presence of multiple layers, e.g., such as a high-k
dielectric, interfacial layer, and workfunction metal layer(s)
underneath sandwiched between the W film and the silicon channel
leads to a poor stress coupling from the W gate contact to the
channel. Having the aforementioned patterning process frees the
space in the p-type MOSFET region, which can create a large volume
of W deposited and which potentially can lead to a better stress
coupling, is believed to improve the strain effect from the gate.
Experimentally, it has been observed that with the new scheme, the
hole mobility and the short channel resistance reduction leads to
an overall mobility enhancement because of the Tungsten applying a
compressive strain and the workfunction displaying a tensile
strain. By removing the first layer of workfunction metal in the
p-type MOSFET region, the compressive strain can be applied to the
channel in both the in-plane (220) and the off-plane (002)
direction that is favorable for a hole transport. There is also a
highly stress W film available if a further boost is required for
an improved performance. The stress level measured in the blanket
films can be enhanced from 2.3 GPa to 3.5 GPa, while in a contact
array, the improvement ranges from 1.6 GPa to 2.2 GPa.
[0050] FIG. 15 shows a cross-sectional view of multiple layers of
the gate stack in both the n-type and p-type MOSFET regions. The
n-FET first patterning freezes up the space for the W fill to
induce the channel strain from the gate, where the p-type FET
consists of only a very thin TiN metal that leaves sufficient room
for the W fill along the PC and the fin direction. Experimental
data suggests the proposed device structure improves long-channel
and short-channel mobility significantly in an advanced CMOS
technology. The high stress W film can also be incorporated to
further boost the performance.
[0051] While the present invention has been particularly described,
in conjunction with a specific preferred embodiment, it is evident
that many alternatives, modifications and variations will be
apparent to those skilled in the art in light of the foregoing
description. It is therefore contemplated that the appended claims
will embrace any such alternatives, modifications and variations as
falling within the true scope and spirit of the present
invention.
* * * * *