Accelerating the next technology revolution

Copyright ©2009SEMATECH, Inc. SEMATECH, and the SEMATECH logo are registered servicemarks of SEMATECH, Inc. International SEMATECH Manufacturing Initiative, ISMI, Advanced Materials Research Center and AMRC are servicemarks of SEMATECH, Inc. All other servicemarks and trademarks are the property of their respective owners.

Understanding Chemical Interactions Between Thin Buried Layers of Microelectronic Materials in Advanced CMOS Structures

International Workshop for New Opportunities in Hard X-ray Photoelectron Spectroscopy: HAXPES 2009

Patrick S. Lysaght1, Joseph C. Woicik2, Daniel A. Fischer2, M. Alper Sahiner3

1SEMATECH, Austin TX

2National Institute of Standards and Technology, Gaithersburg, MD

3Seton Hall University, Physics Department, South Orange, NJ

2228 August 2009 2

Outline

• Intro – About SEMATECH

• High-k/Metal Gate Scaling Trends

• More than Moore - Novel Structures & New Materials

• HAXPES Characterization – Practical Examples• High-k Integration - VKE-XPS: Si/SiOx /HfO2

• Dielectric Cap Layers: Higher-k & EWF Tuning

• Disruptive Technology – Toward Zero IF

• Flash Memory Retention - Impact of Intermixing

3

About SEMATECH…Global Consortium

SEMATECH Front End Processes (FEP) DivisionAdvanced Dielectrics & Electrodes for Logic & Memory

Advanced CMOS Scaling – Planar and Non-Planar

Electrical & Physical Characterization /Reliability & Metrology

Novel Emerging Technologies

4

SEMATECH Extensive R&D Network

• University Network: 80+ schools• Examples: MIT, Stanford, Berkeley,

UIUC, CMU, GATech, UT Austin• Students completed (or completing)

their PhD research: 27• Co-authored papers = 180+• Jointly arranged conferences,

workshops = 9

Albany, NY:SEMATECH North Funded by State of NYHome of Lithography, 3D Interconnect, Metrology and FEP

300,000 sq. ft. class 1 clean roomState of the art tools for 300 mm wafers

National Lab Network:• BNL Synchrotron XPS, XAS• LANL Neutron SANS, SPEAR• ORNL HR TEM• ANL Ferroelectric, piezoelectric• Sandia MEMS• NIST FTIR,1/f Noise, XRD

Austin, TX:SEMATECH corporate HQ Home of FEP, Emerging Technologies and ISMI

Premier user of 68,000 sq. ft. class 1 clean room for 200/300mm wafers

R&D of advanced materials/structuresHighly Productive Collaborations w/NIST Scientists,

Joe Woicik & Dan Fischer at NSLS/BNL

5

XG = LG /45

When LG scaled to 90nm, XG =>2nm – 6-7 atomic layers SiO2 -Ig

MOSFET control depends on the capacitance given by:

where, k is the dielectric constant, A is the area and t is the thickness

What’s up with High-k?

kSiO2 = 3.9, kHfO2 = 18-20HfO2

film 5x SiO2

will provide same EOT with better leakage EOT = (kSiO2 /kHfO2

)tHfO2

equivalent oxide thickness (EOT) :

6

OutlineIn the beginning there was Si/SiO2 /Poly-Si

Position (nm)1050

nitrogen K EELS oxygen K EELS titanium L2,3 EELS

Cou

nts

(arb

. uni

ts)

hafnium La EDXS

Position (nm)1050

nitrogen K EELS oxygen K EELS titanium L2,3 EELS

Cou

nts

(arb

. uni

ts)

hafnium La EDXShafnium La EDXS

Position (nm)

Cou

nts

(arb

. uni

ts)

151050

nitrogen K EELS oxygen K EELS

hafnium La EDXS

Position (nm)

Cou

nts

(arb

. uni

ts)

151050

nitrogen K EELS oxygen K EELS

hafnium La EDXS

151050

nitrogen K EELS oxygen K EELS nitrogen K EELS oxygen K EELS

hafnium La EDXShafnium La EDXS

Si Poly

Si TiN

Hf

Hf

5 nm

Si TiN

Si PolyHfO2

HfO2

P o s i t i o n ( n m )

Cou

nts

2 01 51 050

3 5 0 0 0

3 0 0 0 0

2 5 0 0 0

2 0 0 0 0

1 5 0 0 0

1 0 0 0 0

5 nm

HF-last/NH3/HfO2/N2/600°C/TiN

1.5nm 2.6nm 9.1nm

Si TiNSi HfO2

HfO2

polypoly

TiN

Density Gradient

TiN electrode facilitated scaling

Poly-Si produced thicker BIF and formed an additional low-k TIF

Hi-k deposition tools and clever processing tricks have yielded robust, stable films

7

50 nm50 nm~32nm

50 nm50 nm~32nm

Thermal stability: S/D activation to ~ 1000 ºC

high ų

materials

selective Epi• SiGe, Ge, III/V• Burried/ Surface Channels/QW• Selective Dep; Defects; strain;

Junction Extensions• USJ: Flash/Spike anneals• Infusion, Plasma

Contact Resistance

Schottky Barrier of Silicide / Germanides Doped Ni Silicide/Germanide Schottky FETs

Si

HfO2SiOx

1.2 nm<0.5 nm

Planar CMOS Scaling:

Critical Modules

8

CMOS Scaling Enabled by New Materials and/or Architecture

9

Img: 'E0318 071221011_7071107_10_GJ2_rot_s.dm3' MAG: 49kX

50 nm50 nm~32nm

Img: 'E0318 071221011_7071107_10_GJ2_rot_s.dm3' MAG: 49kX

50 nm50 nm~32nm

Materials & Device Roadmap - Logic

2009

High-k / Metal Gate

FinFET

2011 2013 2015 2017 2019

Bulk SiBulk Si

SiGeBulk SiIII-V

Hi Mobility Channel

Si

Bulk Si

(b)

r ans i s tor _200_T aN _Z r O_ InGaAs .Ed_r ot_s .dm3' M AG:

Metal

High-k

InGaAs

Characterization Issues:• SiGe-HK Interface• Damage at p-n in SiGe• Strain in channel

SiO2Si

Nanowires III-V

Many options!

Characterization Issues:• Sidewall S/D doping• Sidewall silicidation• Sidewall etch

Characterization Issues:• strain, Eg for III-V on Si• III-V – HK interface• Doping

10

Materials & Device Roadmap -Memory

2009 2011 2013 2015 2017 2019

SiN Trap Flash

SMSG 2008 IEDM

50nm

300A SiN

2nm

Metal Nanocrystal Flash

Many options:• Resistance Change• Molecular• Spin Torque

1T1C Issues:• HK/MG interface• HK/MG crystal

1T Issues:•Charge storage•Etch damage•Junction quality

-

SiO2

SiNx

TaN

Al2 O31C

1T

DRAM

Nanosys, Inc.

Characterization Issues:• Conduction mechanism• Oxygen Vacancies• NiO, TiO, WO, TaO

Characterization Issues:• traps, defects in Al2O3• Storage node: SiN -> metal NC• High work function metal

M

IM

RRAMR. Waser et al. Nature 2007

11 1128 August 2009 11

Outline

• Intro Intro Intro ––– About SEMATECHAbout SEMATECHAbout SEMATECH

••• HighHighHigh---k/Metal Gate Scaling Trends k/Metal Gate Scaling Trends k/Metal Gate Scaling Trends

••• More than Moore More than Moore More than Moore --- Novel Structures & New Materials Novel Structures & New Materials Novel Structures & New Materials

• HAXPES Characterization – Practical Examples• High-k Integration - VKE-XPS: Si/SiOx /HfO2

• Dielectric Cap Layers: Higher-k & EWF Tuning

• Disruptive Technology – Toward Zero IF

• Flash Memory Retention - Impact of Intermixing

12

Influence of NHInfluence of NH33

on HfOon HfO22

/SiO/SiO22

/Si(100)/Si(100)

Si 2pNormalized to Si0

A

B

C

Kinetic Energy (eV)2103

Inte

nsity

(a.u

.)

21012099209720952093

Si4+

Si 2pNormalized to Si0

A

B

C

Kinetic Energy (eV)2103

Inte

nsity

(a.u

.)

21012099209720952093

Si4+

hv = 2200 eV

NH3 post-deposition anneal

NH3 pre-deposition anneal

NH3 pre- & post-deposition anneal

NH3 Pre-DA strongly influences interface Si coordination (samples B & C)

NH3 PDA strongly influences Hf coordination (samples A & C)

@ NIST Beamline X24A (NSLS/BNL)@ NIST Beamline X24A (NSLS/BNL)

A

21872181 2183 2185Kinetic Energy (eV)

Hf 4fHfO2

Inte

nsity

(a.u

.)In

tens

ity (a

.u.)

Inte

nsity

(a.u

.)

NH3 pre-deposition anneal

NH3 post-deposition anneal

B

C

NH3 pre- & post-deposition anneal

HfO2

HfO2

Hf-N

Hf-N

hv = 2200 eVHigh-k Deposition

A ISSG 2 nm None HfO2 NH3 / PDA

B HF - last NH3 / Pre-DA HfO2 None

C HF - last NH3 / Pre-DA HfO2 NH3 / PDA

Sample Interface Pre Deposition

Anneal

Post Deposition

Anneal

High-k Deposition

A ISSG 2 nm None HfO2 NH3 / PDA

B HF - last NH3 / Pre-DA HfO2 None

C HF - last NH3 / Pre-DA HfO2 NH3 / PDA

Sample Interface Pre Deposition

Anneal

Post Deposition

Anneal

1799 1781 1783 1805 1807

N 1sHf-N

1809

Hf-N

A

B

C

Si-N

Kinetic Energy (eV)

Inte

nsity

(a.u

.)

hv = 2200 eV

NH3 pre-deposition anneal

NH3 pre- & post-deposition anneal

NH3 post-deposition anneal

13

HAXPES: VKE-XPS Depth ProfilingVariable Kinetic Energy-XPS: Si 1s BE as f(depth) through the IF

Hf 4f B.E. constant surface & interface, Oxygen deficient Si, at the SiO2 /HfO2 IF

Inte

nsity

(a.u

.)

Sample A

Nor

mal

ized

Inte

nsity

( arb

. uni

ts)

Kinetic Energy (E-Ec)0 2 4-2-4-6-8

Si 1s

moreoxidized

less oxidized

moreoxidized

less oxidized

“Interface effect”hv=2100 eV

hv=2500 eV

hv=2600 eV

hv=3000 eV

hv=3500 eV

3 nm HfO2

Si

2nm SiO2

P. S. Lysaght, J. C. Woicik, D. A. Fischer, G. I. Bersuker, H-H. Tseng, and R. Jammy, J. Appl. Phys. 101, 024105 (2007).

hv ~ 2170 eV

Sample A

80º - surface sensitive

Glancing -interface sensitive

2179 2181 2183 2185 2187

Hf 4 f

Kinetic Energy (eV)

NH3 Post Deposition Anneal only

hv = 2200 eV

@ NIST Beamline X24A (NSLS/BNL)

Variable Kinetic Energy –1)Depth profile2)Avoid Auger interference3)Capture higher BE surface sensitive core lines to compare with bulk sensitive

14

R (Å)

0 1 2 3 4 5 60.0

0.4

0.8

1.2

1.6

2.0

2.4

2.8

3.2

1.4 nm PDA

1.8 nm PDA

4.0 nm PDA

R (Å)

0 1 2 3 4 5 60.0

0.4

0.8

1.2

1.6

2.0

2.4

2.8

3.2

1.4 nm PDA+RTA

4.0 nm PDA+RTA

1.8 nm PDA+RTA

Monoclinic Tetragonal Orthorhombic 1.4 nm PDA 0.00 0.96 0.04 1.8 nm PDA 0.00 0.99 0.01 4.0 nm PDA 0.67 0.33 0.00 1.4 nm PDA+RTA 0.57 0.43 0.00 1.8 nm PDA+RTA 0.60 0.40 0.00 4.0 nm PDA+RTA 0.96 0.04 0.00

Critical grain size (~ 4 nm) for t- to m- transformation, below which the t- phase is retained at temperatures sufficient to produce m-HfO2 in slightly thicker films with slightly larger grains.

t-HfO2 (k

= 28-29), m-HfO2 (k

= 16-18)

EXAFS: Characterizing HfO2 crystalline phase

15 15

Single channel dual cap or Dual Channel (SiGe pFET) Single Cap (LaOx) Integration

AlOx or SiGe for pFET and LaOx for nFET

Si SubstrateSTI

Si SubstrateSTI

Si Hard MaskTiN Dummy

High-kAlOx

Si Hard MaskTiN Dummy

High-kAlOx

SC1 cleanSC1 clean LaOx + PDA (Optional)LaOx + PDA (Optional)

Si SubstrateSTI

Si SubstrateSTI

Si Hard MaskTiN Dummy

High-kAlOx

Si Hard MaskTiN Dummy

High-kAlOx

SC1 cleanSC1 clean LaOx + PDA (Optional)LaOx + PDA (Optional)LaOx + PDA (Optional)LaOx + PDA (Optional)

STI

Si substrate

STI

SiGe pMOS High-k

Si Hard MaskTiN dummy

STI

Si Hard Mask

SC1 clean

STI

Si Hard Mask

LaOx + PDA

STI

Si substrate

STI

Si substrate

STI

SiGe pMOS High-k

Si Hard MaskTiN dummy

STI

SiGe pMOS High-k

Si Hard MaskTiN dummy

STI

Si Hard Mask

SC1 clean

STI

Si Hard Mask

SC1 clean

STI

Si Hard Mask

LaOx + PDA

STI

Si Hard Mask

LaOx + PDA

16

XPS & EXAFS: AlO Cap Layer Spectra Hf 4f & Hf L3 edge - N2 Interactions

AlO

Si

SiOHfO

N2 PDA forms stable m-HfO2 – wrt Al diffusion

2- 2nm HfO2/N2 PDA (reference)5- 2nm HfO2/N2 PDA/ 1nm Al2O3/N2 PDA11- 2nm HfO2/ 1nm Al2O3/ N2 PDA

m-HfO2 results when pre-annealed via N2 PDA

m-HfO2

EXAFS

: HfO2/N2 PDA

2- 2nm HfO2/N2 PDA (reference)5- 2nm HfO2/N2 PDA/ 1nm Al2O3/N2 PDA11- 2nm HfO2/ 1nm Al2O3/ N2 PDA

2152 2156 2160 21640

5000

10000

15000

20000

Kinetic Energy (eV)

Shift due to Al

No HfO2 pre-anneal

Hf4f

slight Al shift

Rel

ativ

e In

tens

ity (a

.u.)

hν

= 2170eV

2- 2nm HfO2 /N2 PDA (reference)5- 2nm HfO2 /N2 PDA/ 1nm Al2 O3 /N2 PDA11- 2nm HfO2 / 1nm Al2 O3 /N2 PDA

11not m-HfO2

171728 August 2009 17

K4: Si/2nm HfO2/ NH3 PDAK5: Si/2nm HfO2/1nm Al2O3 NH3 PDAK6: Si/2nm HfO2/ NH3 PDA/1nm Al2O3 /NH3 PDA

Onset of m-HfO2 retarded by N (& Al,Si) incorporation

Tetragonal & monoclinic polymorphs of various volume fraction

AlO

Si

SiOHfO

Al & N as diffusing species

K5: HfAlO – XPS Hf 4f

and EXAFS Hf L3 look like Hf silicate K6: More N than Al incorporation in HfO2

Hf 4f

Shift due to Al

Shift due to N

ref

K6: more N than Al affect on Hf

EXAFS

XPS

XPS & EXAFS: AlO Cap Layer Spectra Hf 4f & Hf L3 edge - NH3 Interactions

K52112 2116 2120

0

5000

10000

15000

20000

Inte

nsity

(cps

)

Kinetic Energy (eV)

K3: Si/2nm HfO2 /1nm Al2 O3 as-depK5: Si/2nm HfO2 /1nm Al2 O3 NH3 PDAK6: Si/2nm HfO2 / NH3 PDA/1nm Al2 O3 /NH3 PDA

Rel

ativ

e In

tens

ity (a

.u)

hν

= 2140eV

181828 August 2009 18

K4: Si/2nm HfO2 /1nm Al2 O3 as-depK5: Si/2nm HfO2 /1nm Al2 O3 NH3 PDAK6: Si/2nm HfO2 / NH3 PDA/1nm Al2 O3 /NH3 PDA

AlN

N 1s

1720 1740 17600

4000

8000

12000

Inte

nsity

(cps

)

Kinetic Energy (eV)

k4 k5 k6

Hf4p3/2

~380eV BE

N-O N-O3

AlO

Si

SiOHfO

XPS & XAS: AlO Cap Layer Spectra N 1s & N K1 edge - NH3 Interactions

390 400 410 420 430 440

0.0

0.2

0.4

0.6

0.8

1.0

K5 K6

Inte

nsity

(cps

)

Energy (eV)

Al-N

Hf-N

N K-edge absorption

Low E XASXPS

K6

Al cap more easily etched if HfO2 is pre-annealed

Rel

ativ

e In

tens

ity (a

.u)

hν

= 2140eV

1928 August 2009

K4: Si/2nm HfO2 /1nm Al2 O3 as-depK5: Si/2nm HfO2 /1nm Al2 O3 NH3 PDAK6: Si/2nm HfO2 / NH3 PDA/1nm Al2 O3 /NH3 PDA

2028 2032 2036 20400

2500

5000

7500

10000

Inte

nsity

(cps

)

Kinetic Energy (eV)

k4 k5 k6

Si 2p

Stoichiometry of BIF impacted by NH3 PDA of HfO2

No pre-anneal of HfO2 – Al incorporation alloys HfO2 => HfSiO-like very similar for N2 and NH3 PDA

XPS & EXAFS: AlO Cap Layer Spectra Si 2p & Hf L3 edge - NH3 Interactions

Impact of NH3 PDA on Si/SiO BIF

XPS

Impact of No HfO2 PDA => HfAlO

EXAFSSi/2nm HfSiO2/N2 PDASi/2nm HfO2/1nm Al2O3 N2 PDASi/2nm HfO2/1nm Al2O3/ NH3 PDA

Alloy with Al or Si results in very similar structure

No pre anneal of HfO2

Rel

ativ

e In

tens

ity (a

.u)

hν

= 2140eV

20

2060 2070 2080 2090 2100 2110 2120

0

2000

4000

6000

8000

10000

12000

14000In

tens

ity (a

.u.)

Kinetic Energy (eV)

asdep 700 950

2094 2096 2098 2100 2102 2104 21060

500

1000

1500

2000

2500

Inte

nsity

(a.u

.)Kinetic Energy (eV)

asdep 700 950

2070 2071 2072 2073 2074 2075400

600

800

1000

1200

1400

1600

1800

2000

Inte

nsity

(a.u

.)

Kinetic Energy (eV)

asdep 700 950

Shallow cores normalized to Si0, intensity & energy

Al2p

Si2p 4+

Si

SiOx

AlOx

AlOx/SiOx Intermixing1nm AlO/1nm chem ox/Si

Si0

21

1nm AlO/1nm chem ox/Si

320 325 330 335 340 34510000

20000

30000

40000

50000

60000

70000

80000

90000

Inte

nsity

(a.u

.)

Kinetic Energy

950C 700C asdep

Si1s

Si1s = 1839 eV Binding Energy

Si

SiOx

AlOx

SiSiO

Oxidation increase at 950ºC with Si-O at/near the surface

Spectra normalized to Si peak intensity

SXPS AdvantageR

elat

ive

Inte

nsity

(a.u

)

hν

= 2180eV

AlOx/SiOx Intermixing

22

What is the best strategy to EOT = 0.5nm? New material vs. SiOx scaling

• Higher-k

doped HfSiON show no scaling benefit vs. undoped HfSiON.• STRATEGY: Focus on scaling SiOx layer rather than a new material.

0.7 0.8 0.9 1.0 1.1 1.210-2

10-1

100

101

102

103

104

SiON-HfTiSiON SiON-HfOx/HfZrOx/HfSiON SiON-HfSrSiON SiON-HfBiSiON SiO2-HfSiON

SiO2/PolySi

Gat

e C

urre

nt (V

fb-1

) [-A

/cm

2 ]

EOT (nm)

1000x

5 nm5 nm

TaN

HfO2

SiOx

Zero Interface Risk:

Phonon, dipole, coulomb scatter [1,2,3]

Transient charging significant [2][1] M. Fischetti et al., J. Appl. Phys., 90, p. 4587 (2001)[2] B. H. Lee et al., IEDM Tech. Dig., (2004) p. 689.[3] H. Ota et al. IEDM Tech. Dig., (2007) p. 65.

23

Toward “Zero IF”Can’t use Si2p4+ w/La : La4d3/2 = 105eV La4d5/2 = 103eV

2035 2040 2045 2050 2055

6000

12000

18000

24000

Inte

nsity

(cps

)

Kinetic Energy (eV)

208Aw1 208Aw2 208Aw3 208Aw4

Si2s

1638 1640 1642 1644 1646 1648 1650 1652

5000

10000

15000

20000

25000

30000

35000

40000

45000

50000

X Axis Title

A1 A2 A3 A4

O1s

O1s BE:

HfO2 – 530.2 eV

SiO2 – 532.4 eV

KE = 2178 – ~ 3eV

2175 eV – 1645 = 530 eV

Hf-O like La-O like

Si-O like

Skip O

10s O

20s O

40s O

520 525 530 535 540 5450

6000

12000

18000

24000

Kinetic Energy (eV)

208Aw1 208Aw2 208Aw3 208Aw4

Hf3d5/2

1638 1640 1642 1644 1646 1648 1650 16520

5000

10000

15000

20000

25000

30000

35000

40000

Y A

xis

Title

X Axis Title

134 138 139 1310

O1s

Skip O

10s O

20s O

40s O

KE = 2178 – ~ 3eV

2175 eV – 1645 = 530 eV O1s BE:

HfO2 – 530.2 eV

SiO2 – 532.4 eV

Hf-O La-O

Si-O

520 525 530 535 540 545

0

10000

20000

30000

40000

50000

Kinetic Energy (eV)

3013w4 3013w8 3013w9 3013w10Hf3d5/2

2035 2040 2045 2050 2055

5000

10000

15000

20000

Inte

nsity

(cps

)

Kinetic Energy (eV)

3031w4 3013w8 3013w9 3013w10Si2s O1s

O1s

Si/Hf [LaOx]/TaN as-dep

Si/Hf [LaOx]/TaN {poly/spike/poly etch}

hν

= 2200eV

2424

Mid to Late-transition metals are meta-stable and can release oxygen

7.0%

7.5%

8.0%

8.5%

9.0%

9.5%

10.0%

10.5%

11.0%

% C

harg

e lo

st a

t 24h

rs-1

50C

from

4V

Vth

prog

ram

TaNMoON

Increasing O2 concentationIn the electrode stack

RuO2

MoO

N(a

)

MoO

N(b

)

MoO

N(c

)

7.0%

7.5%

8.0%

8.5%

9.0%

9.5%

10.0%

10.5%

11.0%

% C

harg

e lo

st a

t 24h

rs-1

50C

from

4V

Vth

prog

ram

TaNMoON

Increasing O2 concentationIn the electrode stack

RuO2

MoO

N(a

)

MoO

N(b

)

MoO

N(c

)

RETENTION: MANOS Flash device

Defects / Oxygen vacanciescreated near interface

Oxygen from electrodepassivates vacancies for improved retention

Al2 O3

W-O-N O

Al2 O3

Ta-NO

Gibbs Energy of formation: (kJ-mole-1 @298K)

IrO2 ~ -260MoO2 ~ -580TiO2 ~ -900

Late Transition Metal-oxides: Less stable

Early Transition Metal-oxides: More stable

•Retention improvement for oxygen-bearing electrode may be due available free oxygen to passivate defects/vacancies

2525

360 365 370 375 380 385 390 395-5000

0

5000

10000

15000

20000

25000

30000

35000

40000

45000

Inte

nsity

(cps

)

Kinetic Energy (eV)

W-O-N

W-W-N

W 3W 3dd5/25/2

900 ºC anneal

As-deposited

WON/Al2 O3 interactions as-dep vs. 900°C

W reduction => Retention Improvement

• Loss of O from WON upon anneal may contribute to less defective Al2 O3 bulk and/or interface

1660 1662 1664 1666 1668 1670

0

20000

40000

60000

80000

Inte

nsity

(a.u

.)

Kinetic Energy (eV)

905w5 905w8

O 1O 1ss

905_5: Al2O3/WON/900C

905_8: Al2O3/WON/as-dep

900 ºC anneal

As-deposited

W 3d5/2 core line indicates reduction of WON and narrow- more single component O 1s core line due to 900°C anneal

hν

= 2200eV

26

905_6: Al2O3/TaN/900C

905_9: Al2O3/TaN/as-dep

430 440 450 460 470

16000

20000

24000

28000

32000

36000

40000

Inte

nsity

(a.u

.)

Kinetic Energy (eV)

905w6 905w9

Ta 3d5/2 oxidized as deposited with only slight reduction due 900°C anneal

2164 2166 2168 2170 2172 2174 2176

0

1000

2000

3000

4000

5000

6000

7000

Inte

nsity

(cps

)

Kinetic Energy (eV)

Ta4+

Ta2+

Ta5+

Ta 4Ta 4ffAs-dep 900°C

Ta2 O5 Ta sub-oxides

Ta oxidation => Retention DegradationTa 3dTa 3d5/25/2

1660 1662 1664 1666 1668 167015000

30000

45000

60000

75000

Inte

nsity

(a.u

.)

Kinetic Energy (eV)

O 1sO 1s

Scavenging of O by Ta may contribute to defects in Al2 O3 bulk and/or interface

TaN/Al2 O3 interactions as-dep vs. 900°C

As-deposited

900 ºC anneal

hν

= 2200eV

272728 August 2009 27

Summary

• We’re in an extremely dynamic and exciting period in semiconductor research – no consensus on the next architecture, - many options will continue to be explored

• Complex new materials systems - physics of single atom thick lattice networks - will be integral to many future technologies

• A suite of measurement instruments & techniques are critically needed to address fundamental engineering problems

• HAXPES will play an increasing role in characterizing heterostructures & buried interfaces – “real” samples

• Increasing need for industry engineers to partner with spectroscopy experts for multiyear investigations

28

Example topics: High-k dielectrics/electrodes Physics of dielectric defects Modeling Reliability DRAM, flash, ReRAM, spin RF and Mixed Signal III-V, graphene channel Characterization Gate stack for M/NEMS

Confirmed Invited Speakers…more to comeEvgeni Gousev, QualcommKirklen Henson, IBMGeoffrey Yeap, QualcommSayeef Saluhuddin, UC-BerkeleyJung Dal Choi, Samsung Semiconductor Krishna Parat, Intel

International Symp on Adv. Gate Stack Technology (ISAGST) August 23-26, 2009 ~ San Francisco, California

Abstract Submission Deadline – May 22, 2009 www.sematech.org

29

Computed Electron Chemical Tomography

This movie is a sequence of views of a 3D computed model reconstructed from HAADF-STEM data

TiSi2Si-Fin

SiO2

Si-Gate

TiSi2TiN

HfSiO

• Image intensity in this representation of the 3D model relates position with higher material density • For this sample, contrast therefore dominated by the spatial distribution of heavy metal atoms such as titanium and hafnium

Collaboration w/Brendan Foran, The Aerospace Corp.