Portfolio Micro and nano-technology MEMS – Thin Film Lise Bilhaut, PhD

Lise Bilhaut

1

Content 1.

Nano-transfert Printing

2.

Microfluidic Systems

3.

Microstructed Radiator

4.

2-D Constriction Model

5.

Thin-film Resistivity

6.

VLSI Bistable Nano-Switch 1. 2. 3. 4. 5.

7.

Lise Bilhaut

Magnetic Thin-films Modeling Process flow on 8’ wafers Design and layout Failure Analysis

VLSI Nano-Resonator 1. 2. 3. 4. 5.

Modeling Design and layout Process flow : 4 and 8 ’ wafers XeF2 release process Motion measurement

2

Nano-Transfert Printing 1/2 • • • •

Lise Bilhaut

Single-step patterning of large areas Nonplanar surface (R2R manufacturing) High resolution (< 2nm!) Proved by myself Compatible with organic and bio materials

3

Nano-Transfert Printing 2/2 • Optimized the technique

Master (Si Wafer)

Regular printed Au-line

Stamp (PDMS)

Optimized printed Au-line

• Assessed its maximum resolution Master (Si Wafer)

Lise Bilhaut

Stamp

4

Microfluidic System 1/2

Lise Bilhaut

5

Microfluidic System 2/2 • Nozzle prototyping  2-level photolithography  DRIE  Laser drilling

Lise Bilhaut

6

Microstructured Radiator Layer of dielectric  increase e Metallic TF  decrease as

Relief  increase A Si Wafer

A(a S I solar  a S I solaralbedoFalbedo  eI EIR FEIR )  Qint  e AT 4

Lise Bilhaut

7

2-D Constriction Model x

Body 1

-i

l’(x) L

l2(x)

dx

V2 V1 l1(x)



V

w

b1 0

y

a

+i

i

100 µm

Body 2

Rconstriction 

 f  1  2 L tan( )  1  2 L tan(a )    ln    ln    2t    w a w    a

Van der Pauw pattern

f 

 ln 2

b

c

tR d

Lise Bilhaut

100 µm

8

Thin-film Resistivity 1/2 Crystalline grain Surface

• ρthin film ≠ ρbulk (FuchsSondheimer/Mayadas-Shatzkes) Surface loss



 f  bulk 1  

Grain boundaries losses

Grain boundaries

p

Electron

3e 1  p  7e R    8t 5d 1  R  

R

i

250

300

Substrate 100 90 80 70 d [nm]

• λe : electron mean free path • p : specular reflection parameter at film surface (0 ≤ p ≤ 1) • R : electron reflection coefficient at grain boundaries (0 ≤ R ≤ 1) • d : mean grain size, function of film thickness t

60 50 40 30 20 10 0 0

50

100

150

200

350

t [nm]

Au sample Lise Bilhaut

9

Thin-film Resistivity 2/2 • ρsample vs ρtheoretical (Au sample)  p = 0 (specular surface)  R = 0.27 (set to fit theoretical curve)  ρbulk = 2.3.10-8 Ω.m (monocrystalline)  λe = 38 nm

Au (λe = 38 nm)

Pt (λe = 19 nm)

Experimental ρ (50 nm) [Ω.m]

3.89.10-8

1.77.10-7

Formula ρ (50 nm) [Ω.m]

3.80.10-8

1.70.10-7

Bulk ρ [Ω.m]

2.3.10-8

1.04.10-7

• t = 35 nm  ballistic regime? 6,E-08 Sample Formula Bulk monocrystalline

Resistivity [Ω.m]

5,E-08 4,E-08 3,E-08 2,E-08 1,E-08 0,E+00 0

Lise Bilhaut

50

100

150 200 t [nm]

250

300

350

10

Young Modulus • Nanoindentation  EPt (100 nm) = 145 GPa  EPt (bulk) = 177 GPa 200 Si (100)

180

Module d(Young (GPa)

160 140 E10% du film = 145 ±8 GPa

120 100 Pt

80

Ti

Si

60 40 20 0 0

50

100

150

200

250

300

350

profondeur d'indentation (nm)

Lise Bilhaut

11

From MEMS to NEMS • NEMS : Nano-Electro-Mechanical Systems o Size < 10s µm, with at least one < µm o Surface < MEMS/100 o Mass < MEMS/1000

• Miniaturization Impact o o o o

Very low power level High resolution sensors Multi-sensors approach (More-than-Moore) Microelectronic convergence  Co-integration  Applications : memory, clock, switch o Cost reduction o Nano-science

• Issues linked to decreased dimension o Surface effect, proximity force, noise level o Technological reproducibility Lise Bilhaut

MEMS Accelerometer

NEMS Accelerometer

Hair

12

VLSI Bistable Nano-switch • Applications: interruptor, mechanical memory… • Bistability  Equilibrium Factuation – Fproximity o Demonstrated at the nanoscale in the lab o Industrialization doubtful  Magnetism o Magnetoconstriction o Coil + magnetic materials  Doesn’t exist at the nanoscale Ziegler et al, APL, 84, 2004

Nanoscale actuation system Ensure bistability  Approach VLSI VLSI ≡ Very Large Scale Integration

Lise Bilhaut

Chandler, Microwave Journal, 47, 2004

13

Inheritance of TA-MRAM •

Magnetic Memory

i

Ferromagnetic 1

i

Free layer Pinned layer

Spacer (AlO3, MgO…) Ferromagnetic 2

Low resistivity « 1 »

High resistivity « 0 » Free layer

Ferromagnetic 1

Pinned layer

Spacer Ferromagnetic 2 Antiferromagnetic

High-magnetization nano-magnet

400 nm 400 nm 50 nm FM AF T ~ 170°C TA-MRAM ≡ Thermally Assisted Magnetic Random Access Memory FM ≡ FerroMagnetic (FeNi, FeCo) AF ≡ AntiFerromagnetic (IrMn, NiMn, FeMn, PtMn)

Lise Bilhaut

B ~ 100 Oe

Reversible JFeNi = 1 T JFeCo = 2,4 T

 B ~ 10s mT at 50 nm (for FeCo)

14

Bistable Nano-Switch • Actuation principle  Dipolar interaction  Bistable  Commutation: electro-thermal system Top magnet

Cantilever

Top magnet S

N Cantilever

N

S

S

Substrate

Closed state

N Bottom magnet

N

 

Works at room temperature Manufactured in a co-integration process along with the mechanical structure Lise Bilhaut

S

Substrate

Open state

Bottom magnet 15

Scientific Approach Design /Modeling

• Concept • Design • Modeling tools

NEMS Measurement • Nano-indentation • U (i) • Parametric electrical testing •Optical Lise Bilhaut

Analysis

Technology • Process flow • Implementation (equiments)

16

Mechanical Modeling • Dipolar Interaction 1D Problem

Fz  M x

3D Problem

    F  (m.grad ) B        mBr F

Simplifications

Bz B B  My z  Mz z x y z

y1  mz Bx  mx Bz Nx

y 2   xi Fz ( xi ) i 1

F magnetic force Γ mechanical moment

Γy

m magnetic moment

Fz

B magnetic field

Fz 50 nm

z

y x

• Cantilever deflection: Euler equation + boundary conditions E Young modulus I 2nd moment of area

Lise Bilhaut

d 2z  EI 2    fléchissant ( x j ) dx  x x j

17

Electro-thermal System 1/2  Requirement: limit the number of photolithography masks  1 level for all metallic lines  Read with T line Bottom magnet

Thermistance

Reading current T Line

Heating current Magnetization current H Line

T line ≡ Heating H line ≡ Mag. Field µ0 100 Oe = 10 mT

Lise Bilhaut

T Line

FeMn

Substrate

FeCo FeCo FeMn

Operating Point (FeMn/FeCo) {TFeMn = 170°C ; HFeCo = 100 Oe} 18

Flux3D vs Analytical program (Matlab)

Lise Bilhaut

19

Electro-thermal System 2/2 Heating (T Line)

Magnetic Field (H Line)

T  2T mc p  Lwt th 2  P t t

Loi de Biot et Savart Magnetic Field in FeCo > 100 Oe

 FEM

wligne H = 5 µm 910

200°C Hy [Oe]

Write TFeMn > 170°C

Estimated consumption 230 to 540 mW

Read TFeMn 200 nm)  VLSI  Resolution ~ 200 nm  Very good device quality  Smaller gaps need improvement  Costly

VLSI ≡ Very Large Scale Integration

7.5 µm

(Intégration à très grande échelle)

Lise Bilhaut

35

XeF2 Release Process •

Dry chemical etching process No stiction Good selectivity

•

Usual sacrificiel layer : polySi (~ 2.3 µm/min)

 

Beam

•

In my case : Ti (~ 20 nm/min at 45°C)  OK for NEMS OK for magnetic materials Inhomogeneous etching rate Material deposition if etching time is too long

  

•

Determining factors  Exposure area  Sample size  Number of samples in the process chamber

Magnet

Encastrement Beam

XeF2 ≡ Xenon Difluorure

Lise Bilhaut

Titane (partially etched)

Magnet

36

100 mm wafer – Release Process  Pt beam release 

L = 6 µm ; w = 1.6 µm ; t = 50 nm



Gap ~ 150 nm

~ 150 nm

2 µm 500 nm

Lise Bilhaut

37

200 mm wafer – Release Process  Smaller NEMS 

L = 1 µm ; w = 200 nm ; t = 50 nm  f0 ~ 135 MHz



Gap ~ 200 nm

55 700 NEMS

Encastrement Poutre

Without space optimization

Aimant

Lise Bilhaut

38

Detection • Scanning vibrometer (doppler effect) Reference laser beam Laser beam

3 µm

Lise Bilhaut

39

Motion measurement (static) • Scanning vibrometer (doppler effect) Laser beam

V 30

Déplacement [pm]

25

Proof of concept

20

z(t )  FLaplace  Iexcitation  Vexcitation

15 10 Mesures Calcul

5 0 0

Lise Bilhaut

20

40 60 V_actionnement [mV]

80

100

40

Motion measurement (dynamic) •

Resonance peak  

Normalized at the displacement outside resonance Lorentzian peak fitting

 f0 ~ 6.9 MHz (model: f0 ~ 5.39 MHz) (L = 5 µm ; w = 1 µm ; t = 50 nm)  Q ~ 10 (atmospheric pressure) 0 2

A( ) 

(0   )  2

Lise Bilhaut

2 2

0 2 2 Q2

41

Comparison modeling/measure • f0 measured ~ 6.9 MHz ≠ f0 modeling ~ 5.39 MHz • Anchors’ over-etching effect:  FEM : over-etching  → f0 

• 2nd hypothesis: Pt residual stress f constraints  f 0

L2S 1 4EI

 Pt = 30 MPa  In agreement with known values

Lise Bilhaut

42

f0=1 GHz

detection

Comparison with the state of the art Integrated

f0=125 MHz [5]

[3] f0=1,5 MHz

? [4]

external

f0=0,485 MHz

[1] [1] D. W. Carr et al, APL, 77, 2000 [2] Sotiris et al, Science, 317, 2007 [3]Huang et al, New J. Physics, 7, 2005 [4] J. Arcamone et al, IEEE Trans. on circuits and systems, 54, 2007 [5] M. Li et al, Nature Nanotech, 2, 2007 Lise Bilhaut

actuation integrated

f0=8 MHz

[2] external

f0=6,9 MHz > 100 MHz

43

Lise Bilhaut

Back to presentation

44