a study on formation of high resistivity phases of …...resistivity [mw・cm] c54:13~20, c49:80~100...
TRANSCRIPT
-
1
A Study on Formation of High Resistivity Phases of Nickel Silicide at Small Area and its
Solution for Scaled CMOS Devices
Doctor Thesis
Ryuji Tomita
Supervisor : Prof. Iwai
-
2
Outline
Back ground
Purpose of this work
Configuration of this thesis
Chapter 3~6
Conclusion
-
3
MOS Scaling and SALICIDE
Si Substrate
Well
STI
Si Substrate
Well
STI
Scaling :k (
-
4
Titanium SALICIDE
Ti
Si
a-TiSix
Si
C49-TiSi2 80-100 mW cm
C54-TiSi2 13-20 mW cm
Si Si
>700℃ 500 – 600℃
400 -
500℃
*J.A. Kittl , W.T. Shiau, Q.Z. Hong, D. Miles
Microelectronic Engineering 50 (2000) 87–101
Limitation of TiSi2 as scaling
1. Difficulty in transforming C49 to C54 TiSi2. 2. Low thermal stability as thinning.
3. Si is dominant diffusion specie.
The first SALICIDE material used in fabrication.
-
5
Cobalt SALICIDE
Limitation of CoSi2 as scaling
1. Large Si consumption. 2. Rough interface by nucleation control.
3. High formation temperature of CoSi2.
4. Si is dominant diffusion specie in CoSi.
Co
Si
Co2Si 110 mW cm
Si
CoSi 147 mW cm
CoSi2 15-20 mW cm
Si
Si
700℃ 375℃ 350℃
*H. Iwai, T. Ohguro, and S.-i. Ohmi, NiSi salicide technology for
scaled CMOS, Microelectronic Engineering 60, 157 (2002).
Cobalt SALICIDE has better scalability than Titanium SALICIDE.
-
6
Nickel SALICIDE: Superiority
Ni
Si
Ni2Si 110 mW cm
Si
NISi 147 mW cm
NiSi2 15-20 mW cm
Si
Si
700 -
800℃ 300
-400℃ 250
-300℃
Superiority of Nickel SALICIDE
1. low Si consumption
2. low formation temperature
3. Ni diffusion kinetics
TiSi2 CoSi2 NiSi
Resistivity [mW・cm] C54:13~20, C49:80~100 14~20 10.5~18
Formation Temperature 600~700℃ 600~700℃ 300~400℃
Dominant moving species Si Co(Si) Ni
Si consumption ratio 2.27 3.64 1.83 *参考 Silicide Technology for Integrated Circuits L.J. Chen
-
7
Nickel SALICIDE: Issues for practical use Critical Issues of Nickel SALICIDE
1. Controlling phase formation
2. Controlling and limiting diffusion of Ni in Si
3. Poor thermal stability of NiSi
4. Oxidation of NiSi
5. Reducing the specific contact resistivity
Phase Resistivity
(mW・cm)
Ni 7-10
Ni3Si 80-90
Ni31Si12 90-150
Ni2Si 24-30
Ni3Si2 60-70
NiSi 10.5-18
NiSi2 34-50
Silicide Phase and Resistivity
*参考 Silicide Technology for Integrated Circuits, L.J. Chen
Series Resistance v.s. Gate Length
*J-G Yun et al. Japan Journal App. Phy.Vo43, pp.6998 2004
*M Tsuchiaki et. al.
Japan Journal App. Phy.
Vo44, pp.1673 2005
*S-D Kim et al. TED2002
-
8
Purpose of this work
Following issues of Nickel silicide
are investigated.
1. Phase control of nickel silicide at
small areas.
2. Thermal stability of NiSi.
3. Schottky barrier height lowering at
NiSi and silicon interface.
-
9
Configuration of this thesis
Chapter 1 Introduction
Chapter 4
Improvement on
Uniformity
of Sheet Resistance
Chapter 7 Summary and
Conclusions
Chapter 6
Schottky Barrier Height
between NiSi and Si
Interface
Chapter 5
Thermal Stability
of NiSi
Chapter 2 Fabrication and Characterization Method
Chapter 3
High Resistivity of
Nickel Silicide
at Small Area
-
Chapter 2 CMOS Silicide Process Flow
and Sheet Measurement Pattern
Sectional view of test structure Top View of test structure
L=0.5~100mm W=0.10 ~ 1.0mm
STI or Side Wall
silicide
V I
silicide
contact
Substrate or Poly-Si
STI or
Side
Wall
STI, Gate, Junction, Wet Clean
Ni/TiN sputter:Ni→TiN in-situ
1st Anneal:270℃ 150sec heating up by hot plate
Selective Wet Etch:SPM(H2SO4+H2O2)
2nd Anneal:395℃ 30sec hold
Contact, Cu Interconnects
-
11
Chapter 3 High Resistivity Phase of
Nickel Silicide at Small Area
Purpose of this chapter
Understanding the formation of high resistivity
phase of Nickel silicide at small area.
Identified nickel silicide phase
Compared to CoSi2
Reaction Mechanism and Thermodynamics of nucleation
-
12
Width dependence of Sheet Resistance
Sheet resistance increase in only narrow line.
n+ active area (L=100mm) p+ active area (L=100mm)
0
5
10
15
20
0.1 1 10
Sh
eet
Res
ista
nce
[ W
/sq
]
Width [ mm ]
Mid
Max
Min 0
5
10
15
20
0.1 1 10 S
hee
t R
esis
tan
ce [
W/s
q ]
Width [ mm ]
Mid
Max
Min
-
13
Sheet Resistance on n+ Active Areas
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
L=0.5um
L=5.0um
L=10um
L=100um
Sheet Resistance[ohm/sq.]
Cu
mu
lati
ve p
rob
abil
ity
[%]
W=1.0mm
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
L=0.5mm
L=5.0mm
L=10mm
L=100mm
Sheet Resistance[ohm/sq.]
Cu
mu
lati
ve p
rob
abil
ity
[%]
W=0.10mm
TEM
&NBD
(a)
(b) (a) L/W=0.5/0.10mm. 6.4W/sq.
(b) L/W=0.5/0.10mm. 19.5W/sq.
Si 200
Si 111
NiSi
102
NiSi
010
NiSi
112
(a) NBD
Thickness:21.8nm, Resistivity:14.0mm・cm
Thickness:19.8nm, Resistivity:38.7mm・cm
Ni3Si2
311
Ni3Si2
311
Ni3Si2
020
(b) NBD
*NBD: Nano Beam Diffraction *Ni3Si2: JCPDS 17-0881
High resistance sample has NiSi and Ni3Si2. Long line has NiSi only. No increase in resistance.
-
14
W=1.0mm
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
L=0.5um
L=5.0um
L=10um
L=100um
Sheet Resistance[ohm/sq.]
Cum
ula
tive p
rob
abil
ity[%
]
W=0.10mm
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
L=0.5mm
L=5.0mm
L=10mm
L=100mm
Sheet Resistance[ohm/sq.]
Cum
ula
tive p
rob
abil
ity[%
]
TEM
&NBD
(a)
(b) (a) L/W=0.5/0.10mm. 7.2W/sq.
Thickness:19.8nm, Resistivity:14.3mm・cm
(b) L/W=0.5/0.10mm. 27.0W/sq.
Thickness:19.4~59.9nm
NiSi
211
NiSi
102
NiSi
113
(a) NBD Si 200
Si 111
NiSi2
111
NiSi2
200
NiSi2
111
(b) NBD
Sheet Resistance on p+ Active Areas
High resistance sample has NiSi and NiSi2. Long line has NiSi only. No increase in resistance.
-
15
NiSi on poly also has degradation in uniformity.
⇒Not crystal states of Si but size of area is dominant.
Sheet Resistance on Poly-Si n+ poly (W=0.11mm) and
n+ active area (W=0.10mm)
p+ poly (W=0.11mm) and
p+ active area (W=0.10mm)
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
Npoly 1.0mm
Npoly 10mm
Npoly 100mm
Ndiff 1.0mm
Ndiff 10mm
Ndiff 100mm
Sheet Resistance[ohm/sq.]
Cum
ula
tive
pro
bab
ilit
y[%
]
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
Ppoly 1.0mm
Ppoly 10mm
Ppoly 100mm
Pdiff 1.0mm
Pdiff 10mm
Pdiff 100mm
Sheet Resistance[ohm/sq.]
Cum
ula
tive
pro
bab
ilit
y[%
]
-
16
Comparison to CoSi2
n+ active area (W=0.10mm) p+ active area (W=0.10mm)
CoSi2 has no degradation in uniformity.
⇒Complex phase diagram of Ni-Si system is main cause in the formation of high resistivity phase.
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
L=0.5mm
L=5.0mm
L=10mm
L=100mm
Sheet Resistance [ W/sq ]
Cum
ula
tive
pro
bab
ilit
y[%
]
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
L=0.5mm
L=5.0mm
L=10mm
L=100mm
Sheet Resistance [ W/sq ]
Cum
ula
tive
pro
bab
ilit
y[%
]
-
Presumed Ni-Si Reaction Path on n+ Area
17
Step Chemical
equation
ΔH
(kJ/mol of
Ni atoms)
Total Volume
Change
[(post-pre)/pre]
Volume Change
of Silicide
[(post-pre)/pre]
Volume
Expansion
[(Silicide-Si)/Si]
1 2Ni+Si→Ni2Si -66 -23.3% - 0.0%→60.5%
2 3Ni2Si+Si→2Ni3Si2 -9 -1.3% +19% 60.5%→43.5%
3 Ni3Si2→Ni2Si+NiSi +2.7 -2.1% -2.1% 43.5%→40.5%
4 Ni2Si+Si→2NiSi -19 -7.5% +50% 60.5%→20.5%
Reaction Step Model
1. Formation of Ni2Si. Increasing
compressive stress.
2. Formation of Ni3Si2. Relaxing stress.
Reaction is noticeable at fine patterns.
3. ΔH is small. Nucleation reaction like.
4. Formation of NiSi.
• Step3 is limiting step. Ni3Si2 tends to remain. • 2 reaction paths complicate phase sequence.
1:Ni2Si→NiSi, 2:Ni2Si→Ni3Si2→NiSi.
*Rivero et. al. Appl. Phys. Lett. 87, 041904 (2005)
-
18
Effect of Compressive Stress on silicidation *C. Rivero et. al, App. Phy. Lett. (2005)
147Å
92Å
175Å
122Å
220Å
183Å
Ni2Si Ni3Si2 NiSi
Silicon
ΔH=
-26.0(KJ/mol)
ΔH=
-10.3(KJ/mol)
Silicide /
Consumed
Si
1.61
(=147/92)
1.44
(=175/122)
1.20
(=220/183)
Large impurities; As, P, and Silicide stress.
⇒Compressive stress retards the reaction. *L.W. Cheng et al. / Materials Science and Engineering A 409 (2005) 217–222
⇒Difficulty in transition from Ni3Si2 to NiSi.
-
Presumed Ni-Si Reaction Path on p+ Area
19
Step Chemical
equation
ΔH
(kJ/mol of
Ni atoms)
Total Volume
Change
[(post-pre)/pre]
Volume Change
of Silicide
[(post-pre)/pre]
Volume
Expansion
[(Silicide-Si)/Si]
1 Ni+2Si→NiSi2 -87 -22.5% - 0.0%→-1.4%
2 3Ni+NiSi2→2Ni2Si -10.5 -11.1% +62.8% -1.4%→60.5%
3 NiSi2+Ni2Si→3NiSi -12 +1.0% +1.0% 19.3%→20.5%
4 NiSi2→NiSi+Si +2 -11.8% -38.9% -1.4%→20.5%
Reaction Step Model
1. Formation of NiSi2. Small stress. Thickness of
NiSi2 depends on implanted dose spices and
amounts. 2. Unreacted Ni and NiSi2 form Ni2Si. Large
stress is formed.
3. NiSi2 and Ni2Si form NiSi.
4. Without Ni, NiSi2 hardly transforms to NiSi.
Large amount B makes thicker NiSi2 layer. Thicker
NiSi2 tends to remain because of lack of Ni. *T. Isshiki et. Al. RTP 2006 14th IEEE International Conf. on RTP
-
20
Effect of Tensile Stress on silicidation
Implanted Boron and STI makes tensile stress.
⇒Tensile stress promotes the reaction. *L.W. Cheng et al. / Materials Science and Engineering A 409 (2005) 217–222
⇒NiSi2 is formed in early stage of silicidation.
Interfacial Energy increase by implanted B (?)
*W. J. Chen and L. J. Chen, J. App. Phys. (1991)
Si Substrate
Well
-Si-B-Si-B-Si STI
Small atom, B and STI
makes tensile stress.
-
21
Thermodynamics of Nucleation
Fre
e E
ne
rgy C
han
ge
Surface Energy S1+S2
Volume
Energy DG* t
DG*
t tco=
(S1+S2) / DG
(a) The free energy of a nucleus
as a function of its radius
Surface Energy
S1+S2+SB
tc1= (S1+S2+SB) / DG
(b) The free energy of a nucleus
When uniform planer film
Boundary surface increases free energy change
→ nucleation reaction is promoted at small areas.
*S1:Surface Energy
S2:Interface Energy
SB:Interface Energy of boundary
*F.M. d’Heurele J. Mater. Res. 3(1), 1988, pp.167-195
Fre
e-E
nerg
y C
hange [
DG
]
r
Surface
bsr2
Volume
aDGr3
r*
DG*
r
S2
SB
S1
t
-
22
Comparison between TiSi2 and NiSi cases
Ni2Si Ni3Si2
Reaction C49-TiSi2→C54TiSi2 Ni2Si→Ni3Si2→NiSi
Figure
Reaction
Mechanism
TiSi→C49-TiSi2→C54TiSi2 Sequential reaction.
1. Ni2Si→NiSi
2. Ni2Si→Ni3Si2→NiSi
2paths to form NiSi.
Volume Change
by reaction
-2%
C49-TiSi2→C54TiSi2
+50%
Ni2Si→Ni3Si2→NiSi
Effect of Scaling on
Phase Transition
Transition from C49 to C54-
TiSi2 is difficult for decreasing
nucleation sites.
Ni2Si→Ni3Si2 reaction is
promoted at small area for
stress reduction.
*T. Isshiki et. Al. RTP 2006 14th
IEEE International Conf. on RTP
*J. A. Kittl et. Al. Microelectronics Eng.
Vol.50. pp.87-101(2000)
Formation of NiSi is much complex than TiSi2 case in terms of stress formation and reaction mechanism.
-
23
Conclusions of Chapter 3 Firstly, formation of Ni3Si2 on small n
+ or NiSi2 on small p
+ active and poly-Si areas is confirmed.
The formation of high resistivity phases at small areas is characteristic features of nickel silicide which has complex phase diagram and reaction mechanism.
Tensile or compressive stress affects the formation of NiSi2 and Ni3Si2 respectively.
-
24
Chapter 4 Improvement on
Uniformity of Sheet Resistance
Purpose of this chapter
Modify silicidation conditions in order to
improve the uniformity of sheet resistance
at small area.
-
25
Effect of First Anneal
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
270oC
280oC
290oC
320oC
350oC
Sheet Resistance[ohm/sq.]
Cum
ula
tive p
rob
abil
ity[%
]
n+ active area W/L=0.10/0.5mm
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
270oC
280oC
290oC
320oC
350oC
Sheet Resistance[ohm/sq.]
Cum
ula
tive p
rob
abil
ity[%
]
p+ active area W/L=0.10/0.5mm
Only p+ active area is improved.
*Process condition 2nd Anneal 395oC
-
26
200
200
111
111 111
111
NiSi2 1 304
304
600
304
600
304
NiSi 2
Si 220
Si 220 Si 220
Si 220
Si 4 204
212 412
412
212
204
NiSi 3
1 2
4
3 L/W=0.5/0.10mm. 11.7W/sq at 575℃
Effect of second anneal on n+ active area
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
395oC
500oC
550oC
575oC
Sheet Resistance[ohm/sq.]
Cu
mu
lati
ve p
rob
abil
ity
[%]
N+diff. W/L=0.10/0.5mm
TEM
&NBD
Not Ni3Si2 but NiSi2 is formed.
*Process condition 1st Anneal 270oC
-
27
NiSi2
220
220
220
220
020
200 020
200
1
220
220
220
220
020
200 020
200
NiSi2 2
Si 220
Si 220 Si 220
Si 220
Si 4 NiSi2
220
220
220
020
200 020
200
220
3
Effect of second anneal on p+ active area
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
395oC
500oC
550oC
575oC
Sheet Resistance[ohm/sq.]
Cu
mu
lati
ve p
rob
abil
ity
[%]
P+diff. W/L=0.10/0.5mm
TEM
&NBD
1 2
4
3
L/W=0.5/0.10mm. 14.8W/sq. at 575℃
Epi-NiSi2 is formed.
*Process condition 1st Anneal 270oC
-
28
Effect of Nickel Thickness n+ active area W/L=0.10/0.5mm p+ active area W/L=0.10/0.5mm
thicker Ni increases volume energy.
*Process condition 1st Anneal 270oC /2nd Anneal 395oC
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
80A60A100A
Sheet Resistance [ W/sq ]
Cu
mu
lati
ve
pro
bab
ilit
y [
%]
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
80A60A100A
Sheet Resistance [ W/sq ]
Cu
mu
lati
ve
pro
bab
ilit
y [
%]
-
29
n+ Active Area after 1st Anneal Sample1 L/W=0.5/0.10mm
002
320
322
Ni2Si
301 320
021
Ni2Si
242 551
313
Ni3Si2
Both Ni3Si2 and Ni2Si are identified.
⇒Process before 1st Anneal should be improved.
*Process condition 1st Anneal 270oC
-
30
High Temp. Ni PVD on n+ active area N+diff. W/L=0.10/0.5mm
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
200oC SPT
300oC SPT
RT SPT
Sheet Resistance[ohm/sq.]
Cu
mu
lati
ve
pro
bab
ilit
y[%
]
TEM
&NBD
414 123
511
NiSi2
111
113 224
NiSi2
422
531 113
NiSi2
220 400
Si
High resistance sample, 18.2W/sq. of Ni PVD at 200℃
Not Ni3Si2 but NiSi2 is formed.
A B C
D
A B
D C
*Process condition 1st Anneal 270oC /2nd Anneal 500oC
-
31
High Temp. Ni PVD on p+ active area P+diff. W/L=0.10/0.5mm
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
200oC SPT
300oC SPTRT SPT
Sheet Resistance[ohm/sq.]
Cu
mu
lati
ve
pro
bab
ilit
y[%
]
TEM
&NBD
High resistance sample, 22.6W/sq. of Ni PVD at 200℃
200
020 200
020
NiSi2
200
020 200
020
NiSi2
200
020 200
020
NiSi2 220 400 Si
Epi-NiSi2 is formed.
A B C
D
A B
D C
*Process condition 1st Anneal 270oC /2nd Anneal 500oC
-
32
Effects of Anneal Time
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
270C REF.270C 2 times long
Sheet Resistance [ W/sq ]
Cu
mula
tiv
e p
rob
abil
ity
[%
]
N+ Active Area
L/W=0.5/0.10mm
P+ Active Area
L/W=0.5/0.10mm
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
270C REF
270C 2 times long
Sheet Resistance [ W/sq ]
Cum
ula
tive
pro
bab
ilit
y [
%]
Longer anneal improves the uniformity.
*Process condition 1st Anneal 270oC /2nd Anneal 500oC
-
33
Effect of Furnace Anneal n+ active area (L/W=0.5/0.10mm) p+ active area (L/W=0.5/0.10mm)
*Process condition 1st Anneal 270oC /2nd Anneal 500oC
Furnace anneal improves the uniformity dramatically.
0 5 10 15 20 25 30 35.01
.1
1
5102030
50
70809095
99
99.9
99.99
Hot Plate 270oC
Hot Plate 290oC
Hot Plate 350oC
Furnace 300oC
Furnace 350oC
Sheet Resistance [ W/sq ]
Cum
ula
tive
pro
bab
ilit
y[%
]0 5 10 15 20 25 30 35
.01
.1
1
5102030
50
70809095
99
99.9
99.99
Hot Plate 270oC
Hot Plate 290oC
Hot Plate 350oC
Furnace 300oC
Furnace 350oC
Sheet Resistance [ W/sq ]
Cum
ula
tive
pro
bab
ilit
y[%
]
-
34
Width dependence of Sheet Resistance
on n+ active area by Furnace Anneal
No increase in sheet resistance at narrow line by furnace anneal.
Hot plate(Reference) (L=100mm) Furnace Anneal (L=100mm)
0
5
10
15
20
0.1 1 10
Sh
eet
Res
ista
nce
[ W
/sq
]
Width [ mm ]
Mid
Max
Min 0
5
10
15
20
0.1 1 10 S
hee
t R
esis
tan
ce [
W/s
q ]
Width [ mm ]
Med.
Max.
Min.
*Process condition
1st Anneal single 270oC
2nd Anneal 395oC
*Process condition
1st Anneal furnace 350oC
2nd Anneal 500oC
-
35
Width dependence of Sheet Resistance
on p+ active area by Furnace Anneal
No increase in Sheet resistance at narrow line by furnace anneal.
0
5
10
15
20
0.1 1 10
Sh
eet
Res
ista
nce
[ W
/sq
]
Width [ mm ]
Mid
Max
Min 0
5
10
15
20
0.1 1 10 S
hee
t R
esis
tan
ce [
W/s
q ]
Width [ mm ]
Me
d.
Hot plate(Reference) (L=100mm) Furnace Anneal (L=100mm)
*Process condition
1st Anneal single 270oC
2nd Anneal 395oC
*Process condition
1st Anneal furnace 350oC
2nd Anneal 500oC
-
36
Comparison among Annealing tools Anneal
Tool RTA Hot Plate Furnace
Structure
Atmosphere N2 Vacuum or N2 N2
Advantage Excellent temperature
control.
•Easy for process below 300 oC.
•Available for in-situ
treatment after PVD.
・Suitable for long-term treatment.
・Suitable for low temperature treatment.
Drawback
•Unsuitable for temperature
below 300 oC.
•Unsuitable for long-term
treatment.
•Hard to control the
temperature profile.
•Unsuitable for long-term
treatment.
・Impossible for short- term treatment.
Silicide
application
Commonly used in
silicidation. ・Hardly used in silicidation.
・Hardly used in silicidation.
Furnace suitable for long and low temperature Annealing.
-
37
Effect of Furnace Anneal on The
Formation of High Resistivity Phases High resistivity
Phase Ni3Si2 on n type NiSi2 on p type
Example
Reaction path
2 way paths
Path1:Ni2Si→NiSi
Path2:Ni2Si→Ni3Si2→NiSi
2 way paths
Path1:Ni2Si→NiSi
Path2:NiSi2→Ni2Si→NiSi
Effect of
furnace anneal
1. Promote Ni3Si2→NiSi
reaction by large heat
quantity by long anneal.
2. Grain growth Ni2Si phase.
1. Promote NiSi2→Ni2Si reaction by
large heat quantity by long anneal.
2. Initial NiSi2 layer is not thicken (?).
*T. Isshiki et. al. 14th IEEE
International Conf. on RTP, 2006 [4.4]
*T. Isshiki et. al.
14th IEEE International
Conf. on RTP, 2006
[4.4] Ni2Si Ni3Si2
Long anneal by furnace promotes the reaction to NiSi.
-
38
Conclusions of Chapter 4
Furnace anneal successfully
improves the uniformity of sheet
resistance at small areas.
Increasing temperature of 1st anneal or
thickness of Ni improves the uniformity of Rs
on p+ active area only.
Increasing temperature of Ni PVD stage or
2nd anneal promotes the formation of NiSi2
phase.
-
39
Chapter 5 Thermal Stability of NiSi
Purpose of this chapter
To Investigate thermal stability of NiSi.
Clarify thermal stability difference between n+ and
p+ active areas.
Clarify thermal stability difference between NiSi on
Si(100) and Si(110).
-
40
Crystal Orientation and Thermal stability of NiSi
Crystal Orientation of NiSi depends on Si.
⇒Orientation of NiSi or Si changes thermal stability ?
*C. Lavoie et. al, Microelectronics Engineering 70, 144-153 (2003)
*Process condition
P type poly-Si,
Ni=15nm, 3 oC/s N2 ambient
Crystal structure of Nickel Silicide on Si(100)
*C.Detavernier et al. Nature 2003
NiSi(112) NiSi(202)/(211)
-
41
Thermal Stability of Sheet Resistance
0 10 20 30 40 50 60.01
.1
1
5102030
50
70809095
99
99.9
99.99
500 oC
600 oC
650 oC
700 oC
Sheet Resistance [ W/sq ]
Cu
mu
lati
ve
pro
bab
ilit
y [
%]
0 10 20 30 40 50 60.01
.1
1
5102030
50
70809095
99
99.9
99.99
500 oC
600 oC
650 oC
700 oC
Sheet Resistance [ W/sq ]
Cu
mu
lati
ve
pro
bab
ilit
y [
%]
n+ active area (L/W=100/10mm) p+ active area (L/W=100/10mm)
p+ active area has better thermal stability than n+ active area.
*Process condition
1st Anneal 270oC /2nd Anneal 500oC
-
42
Cross-sectional TEM Images of NiSi at 700oC
101 213
312
n+ active area
p+ active area
NiSi
Agglomeration of NiSi on n+ active area is confirmed.
-
43
20 30 40 50 60 700
100
200
300
400
2theta angle [degree]
In
ten
sit
y [
CP
S]
NiSi 200 Si 400
w/o impla.
n+-impla.
p+-impla.
20 30 40 50 60 700
100
200
300
400
2theta angle [degree]In
ten
sit
y [
CP
S]
w/o impla.
n+-impla.
p+-impla.
NiSi 020
NiSi 101
Si 220
(a) Parallel to Si (b) Parallel to Si
In-plane XRD patterns of NiSi films
NiSi (020) // Si (110) is observed on p+ implanted Si.
*Process condition
1st Anneal 270oC
2nd Anneal 500oC
-
44
Calculated From D. F. Wilson et al., Scripta Metal., 26, 85 (1992)
Thermal Expansion of NiSi films
R.T.
700˚C
3.00
3.50
4.00
4.50
5.00
5.50
6.00
0 100 200 300 400 500 600 700
Unit
cel
l dim
ensi
on [
nm
]
Temperature [ oC ]
a
b
c
NiSi(211)
Si(220)
NiSi(211)
Si(220) (a) Epitaxial grain
Si(220)
NiSi(020)
NiSi(202)
(b) Transrotational grain
The space of b axis decreases as temperature increases.
⇒ NiSi grains relax by arranging each grains.
*NiSi: Orthorombic axbxc=5.233x3.258x5.659
-
45
n+ active area (all)
n+ Ext-SD only
Deep-SD As only
Deep-SD P only
p+ active area (all)
p+ Ext-SD only
Deep-SD B only
None Dope
Annealed NiSi Images by optical microscope
NiSi on deep-SD B shows no agglomeration.
⇒Highly implanted B is the cause in high thermal stability.
*Process condition
1st Anneal 270oC
2nd Anneal 500oC
Add. Anneal 650oC
-
46
Effect of B dose on NiSi orientation
Si direction Si direction
NiSi (020) // Si (110) is formed at 5x1015 atoms/cm2 and above of B dose.
*H. Kimura and R. Tomita, Micro. Eng. (2010) *Process condition
1st Anneal 270oC
2nd Anneal 500oC
atoms/cm2
-
47
Thermal Stability Comparison
between NiSi/Si(100) and Si(110)
Process Flow
SPM 9:1
↓
BOE 20:1
↓
Ni-PVD 10nm
↓
RTA 225 to 700℃ 30sec
0
10
20
30
40
50
60
70
80
90
100
200 300 400 500 600 700Anneal Temp.[℃]
She
et R
esis
tanc
e[Ω
/sq.
]
Si(100)
Si(110)
Thermal stability: NiSi on Si(100) > Si(110).
-
48
In-plane XRD, f-rotation Analysis
0
500
1000
1500
2000
2500
3000
0 20 40 60 80 100 120 140 160 180
φ [℃] fromSi(110)
Inte
nsity[
coun
t]
NiSi (101)NiSi (011)NiSi (200)NiSi (111)NiSi (112)NiSi (202)/(211)NiSi (103)NiSi (301)NiSi (020)
NiSi/Si(110)
0
500
1000
1500
2000
2500
3000
0 20 40 60 80 100 120 140 160 180
φ [°] fromSi(110)
Inte
nsity[
coun
t]
NiSi (101)NiSi (011)NiSi (200)NiSi (111)NiSi (112)NiSi (202)/(211)NiSi (103)NiSi (301)NiSi (020)
NiSi/Si(100)
NiSi on both Si(100) and Si (110) align to Si substrates.
・NiSi on Si(100) is 4 fold symmetry. ・NiSi on Si(110) is 2 fold symmetry.
Strong NiSi(020)//Si(110) relation is observed from both
samples.
Si(2-20)
Si(-113) Si(1-13)
Process
Ni-PVD 10nm
RTA 500℃30sec
-
49
Thermal Stability of NiSi on Si (110)
NiSi on Si(110) cannot relax by contracting b-axis of NiSi because of 2 fold symmetry.
Si(220)
NiSi(020)
NiSi(202)
(a) NiSi on Si(100) Si(220)
Si(200)
NiSi(202)
NiSi(020)
(b) NiSi on Si(110) Si(220)
-
50
Conclusions of Chapter 5
・Thermal stability of NiSi depend on crystal orientation of NiSi to Si.
・Thermal stability: NiSi on p+ > n+ active area. Highly implanted B (> 5E15atms/cm2)
⇒NiSi(020) aligns to Si(110). ⇒b-axis of NiSi contacts by increasing Temp. .
・ Thermal stability: NiSi on Si(100) > Si(110) NiSi on Si(110) cannot relax by b-axis of NiSi
because of two fold symmetry of Si(110).
-
51
Chapter 6 Schottky Barrier Height of
NiSi and Si Interface
Purpose of this chapter
Lowering Schottky barrier height of
NiSi on Si by Al.
Comparing Schottky barrier height of
NiSi on Si (100) and (110).
-
52
Logic CMOS Series Resistance
Series Resistance
Specific Contact Resistivity
To decrease Specific Cotact Resistivity
・Doping Density N ↑ ・Barrier Height φB ↓
*S-D Kim et al. TED2002
Schottky Barrier Height need to be reduced in order to reduce rc.
-
53
Schottky Barrier Height Interface State
・Dangling Bond ・Metal Induced Gap State (MIGS)
Fermi Level of Metal/Si interface
・φbp=1/3Eg
Modulation of Barrier Height
・PtSi for PFET, ErSi For NFET ・Introducing impurities at the interface.
Interface state and Fermi Level Pining
*Y-C Yeo et al. JAP2002
Semiconductor
*M.Saitoh et al. IEDM2008
CMOS on Si(110) and NiSi
*C.Detavernier et al. Nature 2003
NiSi(112) NiSi(202)/(211) CMOS Performance
Si(110) > Si(100)
NiSi is affected by
Surface Orientation of Si.
Is Schottky Barrier Height on Si(110) different from Si(100) ?
-
54
Schottky Diode I-V Characteristic w or w/o Al
Reverse Current increases by Al.
Process Flow
1. SPM 9:1
2. SPM 4:1,HF50:1,HPM
3. Thermal SiO2 1000Å 4. Photolithography
5. SiO2 Etch, BOE 20:1
6. Resist Strip
7. Al-Implantation
1e14atm/cm2 5keV
8. HF Treatment
9. Ni-PVD 10nm
10.RTA 500℃ 30sec 11. Metal strip SPM4:1
12. Al-PVD 1000Å for back side contact
1.0E-02
1.0E-01
1.0E+00
1.0E+01
1.0E+02
-2 -1 0 1 2
I [A
/cm
2]
Voltage [V]
Si(100)
Si(100) with Al
Si(110)
Si(110) With Al
*Si(100):p-type, 10-20Ω・cm, 475-575mm *Si(110): p-type, 1-10Ω・cm, 500-550mm
-
55
Activation-Energy Measurements
For SBH Determination, without Al
NiSi/Si(110) NiSi/Si(100)
fBp=0.38 eV fBp=0.34 eV
fBp of Si(100) > fBp of Si(110)
⇒Charge neutral level : Si(100) > Si(110)
1.00E-14
1.00E-13
1.00E-12
1.00E-11
1.00E-10
1.00E-09
1.00E-08
3.00 3.50 4.00 4.50 5.00 5.50 6.00
I/T
2 (A
/K
2 )
Measurement Temperature 1000/T(K-1)
-0.02 V
-0.04 V
-0.06 V
-0.08 V
-0.10 V
fBp=0.38 eV
-
56
SIMS Profile after NiSi Formation
Al conc. of Si(110) ≈ Si(100) at NiSi/Si interface.
1E+00
1E+01
1E+02
1E+03
1E+04
1E+05
1E+06
1E+07
1E+14
1E+15
1E+16
1E+17
1E+18
1E+19
1E+20
1E+21
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Se
co
nd
ary
ion
inte
ns
ity (c
ou
nts
/se
c)
Co
nc
en
tra
tio
n (
Ato
ms/c
m3
)
Depth (nm)
Al in Si(100)
Al in Si(110)
Si in Si100(raw ion counts)
Ni in Si100(raw ion counts)
Si in Si(110)(raw ion counts)
Ni in Si(110)(raw ion counts)
Al in Si(110)
Si(raw ion counts)->
Ni(raw ion counts)->
Al in Si(100)
-
57
Activation-Energy Measurements
For SBH Determination, with Al NiSi/Si(110) with Al NiSi/Si(100) with Al
fBp=0.17 eV fBp=0.23 eV
fBpof Si(100) < fBpof Si(110)
NiSi/Si(100): DfBp=0.21eV, NiSi/Si(110): DfBp=0.11eV
⇒interface states: Si(100) < Si(110)
-
58
In-plane XRD
No orientation change observes by Al implantation.
20 30 40 50 60 70 800
200
400
600
800
1000
2theta angle [degree]
In
ten
sit
y [
CP
S]
NiSi with Al-impla. on Si(100)
Si
22.5o from
Si
Si
20 30 40 50 60 70 800
200
400
600
800
1000
2theta angle[degree]
NiSi w/o Al impla. on Si(100)
Inte
nsity [C
PS
]
Si
22.5o from
Si Si
NiSi/Si(100) NiSi/Si(100) with Al
20 30 40 50 60 70 800
200
400
600
800
1000
2theta angle [degree]
In
ten
sit
y [
CP
S]
NiSi with Al-impla. on Si(110)
Si
Si
Si
NiSi/Si(110)
20 30 40 50 60 70 800
200
400
600
800
1000
2theta angle [degree]
NiSi w/o Al-impla. on Si(110)
Inte
nsity [C
PS
]
Si
Si
Si
NiSi(020)
NiSi/Si(110) with Al
-
59
Analysis on Schottky barrier lowering by Al
implant.
1E+00
1E+01
1E+02
1E+03
1E+04
1E+05
1E+06
1E+07
1E+14
1E+15
1E+16
1E+17
1E+18
1E+19
1E+20
1E+21
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Se
co
nd
ary
ion
inte
ns
ity (c
ou
nts
/se
c)
Co
nc
en
tra
tio
n (
Ato
ms/c
m3
)Depth (nm)
Al in Si(100)
Al in Si(110)
Si in Si100(raw ion counts)
Ni in Si100(raw ion counts)
Si in Si(110)(raw ion counts)
Ni in Si(110)(raw ion counts)
Al in Si(110)
Si(raw ion counts)->
Ni(raw ion counts)->
Al in Si(100)
Band structure of NiSi2/Si interface
by Y addition. *Li Geng et. Al. EDL Vol.29 pp.446(2008)
Solid solubility of Al in Si is below 7x1018atms/cm3
(600℃) ⇒Accepter like Al in Si is few.
Addition of Se, S, Y, Al to NiSi/Si interface has following effects.
①Charge Neutral Level is changed by interface state change. ②Work function is changed by dipole formation.
-
60
Conclusions of Chapter 6
Schottky Barrier Height is successfully
reduced by Al implantation for both
Si(100) and Si(110)
Barrier lowering by Al implantation is
significant for Si(100) than Si(110).
No crystal structure change of NiSi was
observed by Al implantation.
-
61
Conclusions of Chapter 7-1
Firstly the formation of high resistivity
phase at small area has been observed,
the issue has successfully been solved
by furnace anneal for 1st anneal.
This process has successfully been used
for a CMOS fabrication condition.
-
62
Conclusions of Chapter 7-2 Chapter 3
The formation of Ni3Si2 or NiSi2 at small areas is
confirmed.
⇒The formation of Ni3Si2 or NiSi2 is related to complex Ni-Si phase diagram and reaction mechanisms.
Chapter 4
Modifying silicide condition for improve the uniformity of
sheet resistance at small area.
⇒Furnace anneal for 1st anneal dramatically improves the uniformity.
-
63
Conclusions of Chapter 7-3 Chapter 5 Thermal stability: NiSi on p+ > n+ active area.
⇒NiSi(020) aligns to Si(110) on p+ active area. ⇒b-axis of NiSi contacts by increasing Temp. .
Thermal stability: NiSi on Si(100) > Si(110)
⇒ NiSi on Si(110) cannot relax by b-axis of NiSi. Because of Two folds symmetry of Si(110).
Chapter 6 Schottky Barrier Height is successfully reduced by Al
implantation for both Si(100) and Si(110)
Barrier lowering by Al is significant for Si(100) than Si(110).
-
Academic Paper & Presentation
タイトル 年月日 雑誌・学会名 著者
Formation of High Resistivity Phases
of Nickel Silicide at small Diffusion Region 2007年6月
International Workshop
on Junction Technology
R.Tomita, H. Kimura,
M. Yasuda, T. Nakayama,
K. Maeda, Y. Sugiyama,
Y. Kikuchi, M. Moritoki
Analysis and Improvement
on the Uniformity of
Sheet Resistance of Nickel Silicide
2008年5月 International Workshop
on Junction Technology
R.Tomita, M. Yasuda,
H. Kimura, K. Maeda,
S. Ueno, M. Moritoki
微小領域におけるニッケルシリサイド の層抵抗均一性とその改善方法
2008年6月
応用物理学会
シリコンテクノロジー分科会
第102回研究集会
R.Tomita, H. Kimura,
M. Yasuda, K. Maeda,
S. Ueno, M. Moritoki
Formation of High Resistivity Phases
of Nickel Silicide at Small Area accepted
Microelectronics
Reliability
R. Tomita, H. Kimura,
M. Yasuda, K. Maeda,
S. Ueno, T. Tomizawa,
Y. Kunimune, H. Nakamura,
M. Moritoki, H. Iwai
Improvement on Sheet Resistance
Uniformity of Nickel Silicide
by Optimization of Silicidation Conditions
accepted Microelectronics
Reliability
R. Tomita, H. Kimura,
M. Yasuda, K. Maeda,
S. Ueno, T. Tonegawa,
T. Fujimoto, M. Moritoki,
H. Iwai