35
Ferromagnetic integrated inductor/noise suppressor Masahiro YAMAGUCHI Department of Electrical and Communication Engineering Tohoku University OUTLINE 1. Thin film permeameter for material evaluation 2. Slit works on thin film inductor 3. Thin film electromagnetic noise suppressor 4. Side channel attack protection 5. Summary PwrSoc’08, Cork, Ireland, September 24th, 2008

PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

  • Upload
    others

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

Page 1: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Ferromagnetic integrated inductor/noise suppressor

Masahiro YAMAGUCHIDepartment of Electrical and Communication Engineering

Tohoku UniversityOUTLINE1. Thin film permeameter for material evaluation2. Slit works on thin film inductor3. Thin film electromagnetic noise suppressor4. Side channel attack protection5. Summary

PwrSoc’08, Cork, Ireland, September 24th, 2008

Page 2: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Permeability vs Ferromagnetic Resonance Freq.

1 10 100 1000 10000-1000

0

1000

2000

Frequency (MHz)

Real

part

of re

lative

perm

eabil

ity

ρ=500μΩcmMs=1.2T

Hk=80 A/m (10 Oe)

160 A/m (20 Oe)400 A/m (50 Oe)

800 A/m (100 Oe)1600 A/m (200 Oe)

(a) Real part (b) Imaginary part

H

M Ms

Hk

1 10 100 1000 100000

500

1000

1500

2000

2500

Frequency (MHz)

Imag

inary

part

of re

lative

perm

eabil

ity

ρ=500μΩcmMs=1.2T

Hk=80 A/m (10 Oe)160 A/m (20 Oe)400 A/m (50 Oe)

800 A/m (100 Oe)1600 A/m (200 Oe)

Inductor Noise Suppressor

0/2

μπγ

ksr HMf =

ks HM /=μconstfr =⋅ 2μ

Page 3: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Shielded loop coil method

Strip line

GND line

Eac

V+

Strip line

GND line

Eac

V+

Development of thin-film permeameter

M. Yamaguchi et al.T. Mag. Soc. Jpn., 3 (2003) 137-140

AlsoRyowa Electronics Co.Model PMM-9G1

Coplanar line method(coming)

Magnetic filmInput

Reflection, S11 Transmission, S21

I/2I/2

I

Magnetic film

Insulator

Magnetic filmInput

Reflection, S11 Transmission, S21

I/2I/2

I

Magnetic film

Magnetic filmInput

Reflection, S11 Transmission, S21

I/2I/2

I

Magnetic film

Insulator

3GHz 6GHz

9GHz

20GHz

RF power supplied by a network analyzer

Signal detection by a network analyzer

Shielded-loop type planar pickup coil

50Ω terminator

Test sampleSide-open TEM cellas RF field generator

Page 4: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

9GHz measurement jig

Centered pick-up coil arrangementStrip line structure (Side-open TEM cell)

Yielding TRAVELING EM fieldLoaded end:

Improved EM uniformity at the coil window

Improved EM uniformity throughout the jig

Page 5: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Planar Shielded-loop coil

Strip line

GND line

50 Ω signal line

Hac

1 turn coilwith gap

V

Page 6: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Voltage transfer along the strip line, I

Strip line

GND line

50 Ω signal line

Hac

1 turn coilwith gap

V

Page 7: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Voltage transfer along the strip line, II

Strip line

GND line

50 Ω signal line

1 turn coilwith gap

New calibration point

Page 8: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

E-filed Voltage Suppression

Strip line

GND line

Eac

V+

Page 9: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

E-voltage suppression

Comparison of E- and M- voltages extracted by simulation & experiments.

-100

-80

-60

-40

-20

0

0 2 4 6 8 10

Meas. (0 deg.)Meas. (90 deg.)HFSS (0 deg.)HFSS (90 deg.)

Gai

n[dB

]

Frequency[GHz]

30.3dB(3GHz)

35.9dB(6GHz)

Simulated by ANSOFT HFSS Ver. 8.0.21

Page 10: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

TDR Profiles

30

35

40

45

50

55

60

65

70

43400 43500 43600 43700 43800 43900 44000 44100 44200

Position [psec]

Impe

dance [

Ω]

lower

middle

upper

Point

Page 11: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

-1000

-500

0

500

1000

1500

2000

100 1000 10000

Experimental data - Calculated

Perm

eabi

lity

Frequency (MHz)

FeCoBN: thickness 978 nm, 4πMs~22kG,Hc~5Oe, Hk~60 Oe, resistivity ~ 73.8 μΩ⋅cm

By Prof. J-R Kim, Hanyang Univ., Korea

Validity check

1. Comparison with L.L.G. equation 2. Int’l cross measurementsTohoku U, Japan(Yamaguchi)

Frequency domainTEM cell/Pick-up coil

NIST, USA (Silva)Time domainCoplanar line

CEA, France (Acher)Frequency domainMicrostrip line

FMR frequencies agreedwithin the error of 15%

Page 12: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Frequency profile of the cross-measured permeability on a CEA-made 0.14μm thick CoNbZr film.

Int’l Cross Measurements, I

A jig for the transmission line method.

9mm1.7mm

< 12.5mm (λ/4 at 6GHz)

Z0=50Ω9mm

1.7mm

< 12.5mm (λ/4 at 6GHz)< 12.5mm (λ/4 at 6GHz)

Z0=50ΩZ0=50Ω

40(lc)

1030

30(Wc)

5.5

SMAconnector Parallel plates

Semi-rigidcable

Pickup coil1.4mm×6.4mm

Unit:mmTaper partion

A jig for the pickup coil method.

M. Yamaguchi, O. Acher et al, Journal of Magnetism and Magnetic Materials 242-245 (2002), 970-972.

Page 13: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Printer

Network Analyzer

Display

PC

Power Source

Solenoid Coil

Key Board

Main jig

Completed System (PMM-9G1, by Ryowa)

M. Yamaguchi et al, IEEE Intermag2003, EA-03, Boston USA April 2003.

Page 14: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

CoZrOThickness 1.82μm, Resistivity 1800μΩcmS. Ohnuma et al, 25th Annual Conference on Magnetics in Japan, 25pC-1, (2001.9).

0 1 2 3 4 5 6-100

0

100

200

Meas. Cal.(L.L.G.)

Frequency (GHz)

Rel

ativ

e pe

rmea

bilit

y

μr'

μr"

10-1 100 101

-100

0

100

200

300

Frequency (GHz)

Rel

ativ

e P

erm

eabi

lity

CoFeAlO

CoPdSiO

CoFe-SiO2

10-1 100 101

-100

0

100

200

300

10-1 100 101

-100

0

100

200

300

Frequency (GHz)

Rel

ativ

e P

erm

eabi

lity

CoFeAlO

CoPdSiO

CoFe-SiO2

10-1 100 101

-100

0

100

200

300

10-1 100 101

-100

0

100

200

300

Frequency (GHz)

Rel

ativ

e P

erm

eabi

lity

CoFeAlOCoFe-SiO2

CoPdSiO

CoFeAlO: S. Ohnuma et al, J. Appl. Phys., 85, (1999) 4574.CoFe-SiO2: M. Munakata, et al, J. Magn. Soc. Japan, 28, (2004) 240.CoPdSiO: S. Ohnuma et al, J. Magn. Soc. Japan, 23, (1999) 240.

Granular films

Page 15: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

FeCoBSi(80 nm)/NiFe(20 nm) bilayerA. Gerber, et al, IEEE Trans. Magn., 43, 2624

(2007).

NiFe/CoFeN/NiFe(by Prof. Wang, Stanford U)

Metallic films (FeCo-base system)

Page 16: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

100 10000

50

100

150

3000

μ'B

ED

C

A

Rel

ativ

e pe

rmea

bilit

y

Frequency ( MHz )

A, D : NiZn-Ferrite films, B, C, E : NiZnCo ferrite films

E

C

DA

B

μ”

Spin-Spray ferrite films

Page 17: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

lm = 4mmwm = 20μmtm = 0.2μmdm = 2.3 μm,3.3 μm,4.0 μm

wm dm

tm

lm

Easy ax

is

0.01 0.05 0.1 0.5 10

500

1000

1500

Frequency (GHz)

Rea

l par

t of r

elat

ive

perm

eabi

lity

d m

calc. H k

no slitd m =2.3d m =3.3d m =4.0

800A/m(10Oe)2960A/m(37Oe)4000A/m(50Oe)5600A/m(70Oe)

sdkkeff MNHH +=

sdk

seff MNH

M+

0

2

2 μπγ sdks

rMNHMf +=

Hk : anisotropy fieldNd : demagnetizing factorMs : saturated magnetization

γ : gyromagnetic constant

45o

Intrinsiceasy axis

Slit works on magnetic film

Page 18: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Observed domain structures (1)

Test pattern of bi-directional micro wire array

4mm

44

425

25

EasyAxis

4

unit: mm

(a) 10μm wide micro wire array

Observed domain structure.0.1mm thick and 2mm long samples.

50μm

50μm

(b) 30μm wide micro wire array

Page 19: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Hysteresis curves

0.1mm thick10 μm wide 0.6 μm spacing 2mm long(a) Easy axis direction (b) Hard axis direction

[Oe] 100-100

10

-10

[kG]

0

[kG]

-100

10

100

-10

0 [Oe]

[kG]

1000-100

10

-10

[Oe]

[kG]

-10

[Oe] 100-100 0

10

Horizontally aligned micro wire array

(a) Easy axis direction (b) Hard axis direction

Vertically aligned micro wire array

Broken lines:before annealingSolid lines:after annealing

Page 20: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Frequency profile of complex permeability

Width: varied, 0.1μm thick, 0.6 μm spacing, 4mm longno heat treatment

0

500

1000

1500

2000

2500

1 10 100 1000

tm=0.1μm

Noslit

wm=20μm

wm=10μm

Calc.

Frequency (MHz)

0

500

1000

1500

1 10 100 1000

tm=0.1 μm

Noslit

wm=20μm

wm=10μm

Calc.

Frequency (MHz)

(a) Real part (b) Imaginary part

Page 21: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

■MiniaturizationMagnetic film enhances magnetic flux.

■Low coil lossCoil length reduces with miniaturization.

■Low eddy current loss in Si and SiGeMagnetic flux is mostly associated with magnetic film.

■Thin insulator between coil and Sibecause low eddy current loss in Si.

■High self-resonance frequency???Size is small but magnetic film can be a electrode of a capacitor...

Toward High efficiency

Possible advantages of magnetic thin-film inductors

Page 22: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Spiral inductor

Bi-directional micro wire array structure

Inducedeasyaxis

Intrinsiceasy axis

Advantages:■Enhace the air-core inductance

up to 100%■Post Si process compatible■by micro wire array structure:

- Enhance FMR frequency■by bi-directional wire structure

- Utilize 100% area of magnetic film

Si

SiO2

Spiral coilPolyimide

Magnetic filmUnderlayer

Coil Spiral:

Width:

Spacing:

Thickness:

Area:

4

11.0

11.0

2.6-3.0

337 x 337

turns

μm

μm

μm

μm2

Magnetic Thickness: 0.1, 0.3 μm

thin film Wire array

patterns:

None, Parallel orMicro wire array

Substrate Thickness: 600 μm

Page 23: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

0.1 0.2 0.4 1 24

6

8

10

12

Frequency (GHz)

Induc

tance

(nH)

air core0° type0° type(polished)45° type(polished)

0.1 0.2 0.4 1 20

5

10

15

20

Frequency (GHz)

Resis

tance

(Ω)

air core0° type0° type(polished)45° type(polished)

0.1 0.2 0.4 1 20

2

4

6

8

Frequency (GHz)

Quali

ty fac

tor

air core0° type0° type(polished)45° type(polished)

■L=10.1nH (+ 50%) but R=19.2Ω and Q=3.4

Non-patterned CoNbZr film inductor

Page 24: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

(i) Air-core

(ii) Plain

(iii)Aligned : The slits on magnetic film faces to the coil gap.

(iv)Shifted : The slits are shifted by a half coil pitch

(v)Closed : The top and bottom magnetic layers are terminatedat the edges

Magnetic film

Substrate

Coil

InsulatorMagnetic film

Substrate

Coil

Insulator100μm

Structure

(a) Cross sectional structure(b) Magnetic film

patterning(c) Outlook of the

completed inductor

(i) Air-core

(ii) Plain

(iii)Aligned : The slits on magnetic film faces to the coil gap.

(iv)Shifted : The slits are shifted by a half coil pitch

(v)Closed : The top and bottom magnetic layers are terminatedat the edges

Magnetic film

Substrate

Coil

InsulatorMagnetic film

Substrate

Coil

Insulator100μm

Structure

(a) Cross sectional structure(b) Magnetic film

patterning(c) Outlook of the

completed inductor

Sandwich structure thin film inductors

Page 25: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

L & Q

0.1 0.2 1 10 200

5

10

15

20 Air-core Plain Aligned Shifted Closed

2

L [n

H]

frequency [GHz]0.1 1 2 10 200

10

20

30

40 Air-core Plain Aligned Shifted Closed

frequency [GHz]0.2

Q

Page 26: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

High Freq. operation

High Density Packaging

Low Power drive

Signal and communicationfrequencies conflict.Wave length becomes similarto device and Equipment size.

Intra-EMI problems of modern IT equipments

For further miniaturizationFor new device installation(camera, et al)

Low S/N ratioradiating wave source(Common Mode Current)

Noise Suppression Sheet(NSS)

To reduce unexpected signal tonoise couplingsBy means of FMR loss generation in magnetic materials

Complimentary: NEC-Tokin Co.

Page 27: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Reflection, S11

AttenuationInput

Transmission, S21

I/2

I/2I

Insulator

magnetic layer

Ideal frequency profile Schematic diagram

Idea of thin film noise suppressor

0Frequency (GHz)

Atte

nuat

ion

in

trans

mis

sion

line

(dB

)

0

Signal frequencies

Noise frequencies

Insertion loss

Without magnetic film

With magnetic film

P

Ploss/Pin = 1-(|S21|2+|S11|2)

Page 28: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

0.1 1 10-60

-50

-40

-30

-20

-10

0

20Frequency (GHz)

S21

(dB

)

-60

-50

-40

-30

-20

-10

0

5

C

B

D

E

Thickness(magnetic film, insulator)

CoNbZr magnetic film15 mm(L) x 2 mm(W) x 0.5, 1, 2 μm(T)CoNbZr magnetic film15 mm(L) x 2 mm(W) x 0.5, 1, 2 μm(T)

A: CPW.

~ 27

dB

~ 15

dB

~ 25

dB

B: (1 μm, 7.5 μm)C: (0.5 μm, 2 μm)D: (1.0 μm, 2 μm)E: (2.0 μm, 2 μm)

Ki Hyeon Kim, S. Ohnuma, M. Yamaguchi, IEEE Trans. Mag.,40(4), (2004)

Gap length vs. transmission properties

A: CPW.

Page 29: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Sheet resistance , Rs (Ω/□)

Sheet resistance

b

t

a

Resistivityρ

Current

tRs ρ

=ρ: Resistivityt: Film thickness

[Ω/square]

100 101 102 103 104 105

1.0

1GHz

2GHz

6GHz

6GHz2GHz

1GHz2GHz6GHz

Al-O

Co-Al-OFMR

S. Ohnuma et al: Intermag2005

Size: 35x35mm2

Page 30: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

d = 50, 100, 200, 500μm

d = 500μm

6μm

2μm

Lead portion

Loop portion

Ground gap

A A’

C C’B’

B

Thin-film shielded-loop coil

Page 31: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Time domain analysis

BG-1

BV-1

BG-2

BV-2

BV-3

++

-150

-100

-50

0

50

100

150

120 130 140 150 160 170 180 190 200 210 220 230 240Time [ns]

Curr

ent [

mA]

BV-3

-150

-100

-50

0

50

100

150

120 130 140 150 160 170 180 190 200 210 220 230 240Time [ns]

Curr

ent [

mA]

BV-1

(BV-2)+(BV-3)

(BV-1)-(BV-2)-(BV-3)

-150

-100

-50

0

50

100

150

120 130 140 150 160 170 180 190 200 210 220 230 240Time [ns]

Cur

rent

[mA

]BV-1 BG-1

-150

-100

-50

0

50

100

150

120 130 140 150 160 170 180 190 200 210 220 230 240Time [ns]

Curr

ent [

mA]

BV-2 BG-2

Page 32: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

A variety of side channel attack methods

Computer virus

Illegal inputSide channel attack

Keyboard input

Nondestructive attack

Analysis by combining various I/O to the module

・Frequency operation・Voltage operation・Adding noise・Electric field, Magnetic field and Irradiation Processing

timeEM radiation current,voltage

Leakage information

Destructive attackLight,

EM radiation,and irradiation

Input in moduleOutput from inside of module

•Circuit pattern analysis•Wiring probe•Radiation observation

Side channel attack on cryptographic modules

Page 33: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Digital oscilloscope

LSI board

PC for wave analysis

Micro magneticfield probe

Magnetic field probe

Measurement point

Electromagnetic radiation analysis system that use micro magnetic field probe

Page 34: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Differential Electromagnetic Analysis (DEMA)

Outline of DEMAI) Acquire wave form corresponding

to various plaintexts (ciphertexts).II) Assume a part of the key.III) Assume 1 bit of internal variety

from the key assumed.IV) Classify to two groups according to

an internal variable is 0 or 1.V) Calculate difference between averages of

the two groups.VI) Repeat II)-V) changing assumption of the key.VII)Consider the one with the maximum peak to be a correct key.

Electromagnetic radiation from CPU depends changes of the data.…DEMA is attack method that uses change in EM wave radiated from CPU while encryption.

Assume the correct key

Assume the false key

Page 35: PwrSoc’08, Cork, Ireland, September 24th, 2008 ...pwrsocevents.com/wp-content/uploads/2008...PwrSoc’08, Cork, Ireland, September 24th, 2008. Permeability vs Ferromagnetic Resonance

Relation between sampling frequency and analytical accuracy

Number of wave form data

Num

ber o

f erro

r bits