parametric study of contact fritting for improved … · parametric study of contact fritting for...

Parametric Study of Contact Fritting for Improved CRes Stability

06/12/2006 – SWTW, San Diego

Dr. Christian Degen, Oliver Nagler, Michael Horn, Dr. Florian Kaesen

Infineon Technologies AG, GermanyCommunication Solutions – Test Techonolgy (COM TT)

C. Degen: Contact Fritting (SWTW 2006)

??? Fritting ???Electrical conditioning of contacts by ‘overvoltage’ / ‘overcurrent’

••HolmHolm‘‘ss theory of electrictheory of electric contactscontacts

CanCan frittingfritting bebe utilizedutilized forfor contactcontact conditioning of probe needles?conditioning of probe needles?

••OverviewOverview overover evaluationevaluation setupsetup

PRPRobeobeFOFOrcerceIInvestigationnvestigationTTool (PROFIT)ool (PROFIT)

••Parameters underParameters under study:study:

••AA--fritting and Bfritting and B--frittingfritting

••CurrentCurrent level and polaritylevel and polarity

••propro‘‘s and cons and con‘‘s, conclusionss, conclusions


Holm‘s theory of electric contact:(Holm: Electric Contacts, Springer 1967)

•two conductors in mechanical contact, separated bythin insulating layer (e.g., probe needle on Al-padseparated by layer of native oxide)

•the insulating layer can be broken mechanically

•the insulating layer can brake-down locally due to electrical voltage (A-fritting), creating a channel forinitial current flow (A-spot)

•the A-spot can be broadened by transport effectsdue to electric current (B-fritting), reducing CRes

??? Fritting in theory ???What happens during formation of an electrical contact?


How to study fritting of probe needles? universal tool @ IFX probing lab: PROFIT = PRobeForceInvestigationTool(SWTW2005: Nagler et al / ‘An Advanced Probe Characterization Tool for Online Contact Basics Measurements‘)

•motorized x-y-z stages, x-z force sensors, top-view microscope•standard PCBs for mounting various single probes•substrates with standard pad material (Si-wafer pieces, coated with Al, Au, …)•LabView based custom-made software for automatic multi-TD investigations•high precision SMU for parametric tests •capabilities: probe force, current/voltage dc, RF, leakage, S-parameter, …


How stable is a probe contact?TungstenRhenium cantilever 20µm tip on Aluminum, OD=20µm, 1000TDs

024

68

101214

161820

CRes @ I=1mA

CRes @ I=10mA

0 200 400 600 800 1000024

68

101214

161820

CRes @ I=30mACR

es in

Ohm

s

number TDs

0 200 400 600 800 1000

CRes @ I=100mA

•CRes degrades after several 100 TDs without cleaning•probes accumulate Al-oxide -> isolating layer between tip and pad


Why are electric contacts bad? -> Conductors separated by isolating layer!(here: high current relay contact, from literature)

from: Diagnostic routines for automotive relays in the field, IC- and EOL-tests, application note by Tyco Electronics

I-U characteristics during fritting cycle•A-fritting is voltage induced formation of initial conductive channel•B-fritting is current induced broadening of existing channel

(for details see Holm: Electric Contacts, Springer 1967)


0 -20m -40m -60m -80m -100m0.00

-0.05

-0.10

-0.15

-0.20

-0.25

-0.30

-0.35

-0.40

2.8Ohms

18Ohms

sens

e vo

ltage

in V

olts

drive current in Amps

0.0 -20.0m -40.0m -60.0m -80.0m -100.0m0.0

-0.1

-0.2

-0.3

-0.4

-0.5

contact overdrive

volta

ge in

Vol

ts

current in Amps

A-fritting

B-fritting

Formation of electric contact: cantilever TD on Aluminum pad(high precision quasistatic measurement / full trace ~1sec)

•observation of A-fritting and B-fritting for probe needles on pad material


0

10

20

30

40

50

60

70

0.0E+00 1.0E-04 2.0E-04 3.0E-04 4.0E-04 5.0E-04time in seconds

CR

es in

Ohm

s

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

curr

ent i

n A

mps

CRes during frittingCRes after frittingcurrent ramp

What are the typical timescales for contact fritting?-> investigation of voltage drop across contact during current step

•current step <50µs•full breakdown of insulating layer after ~150µs (typ. <<1msec)•instantaneously low CRes for 2nd cycle (after fritting)


Characteristic curve of contact fritting:voltage drop across contact during current step

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.02 0.04 0.06 0.08 0.1 0.12

current in Amps

volta

ge d

rop

in V

olts

Cres ~20Ohms(t=25µs)

Cres ~ 9Ohms(t=45µs)

Cres ~ 4Ohms(t=145µs)

•typical timescale <<1msec


1

10

100

1000

10000

volta

ge

curre

nt

volta

ge

curre

nt

volta

ge+c

urren

t

fritting method (200TDs each)

CR

es in

Ohm

s (lo

g sc

ale!

) before frittingafter fritting

What is dominant for contact fritting of probe needles? voltage effects vs current effects

U=-20V(I<1mA)

I=-100mA(U<2.5V)

U=-20V(I<1mA)

I=-100mA(U<2.5V)

U=-20V &I=-100mA

•Is the effect always initiated @ U<2.5V ? -> YES•Or do higher voltages assist conditioning? -> current is crucial


summary part1 – contact fundamentals

contact formation and fritting:

•fritting cycle observed on probe-pad-contacts•fritting cycle typically <<1msec•A-spot prepared by A-fritting at voltages <2.5V (assisted bymechanical scrub)•current fritting (B-fritting) significantlylowers CRes

next questions:

•polarity of fritting current ?•level of fritting current ?•multi-TD test ?


Flow of multi-TD experiments:1. meas. CRes before – do fritting – meas. CRes after -> all the same TD2. no fritting, only meas. CRes -> reference without any fritting effects


Which polarity of fritting current gives better stability?(cantilever probe, OD=20µm, I=60mA changing polarity)

0

2

4

6

8

10

12

14

16

18

20

-60mA +60mA -60mAfritting current (500TD each)

CR

es in

Ohm

sbefore frittingafter fritting

•negative current = better CRes stability (pad=ground, probe=negative)


Which polarity of fritting current gives better stability?(cantilever probe, OD=20µm, I=100mA changing polarity)

0

2

4

6

8

10

12

14

16

18

20

-100mA +100mA -100mAfritting current (500TD each)

CR

es in

Ohm

sbefore frittingafter fritting

•negative current = better CRes stability (pad=ground, probe=negative)


How much current is necessary for CRes stabilisation by fritting?

0

2

4

6

8

10

12

14

16

18

20

0mA

-10mA

-20mA

-30mA

-40mA

-50mA

-60mA

-70mA

-80mA

-90mA

-100m

A-90

mA-80

mA-70

mA-60

mA-50

mA

fritting current (200TDs each)

CR

es in

Ohm

s

before frittingafter fritting

•significant improvement from I > -40..-60mA to I=-100mA


1

10

100

1000

10000

1 501 1001 1501 2001number TDs

CR

es in

Ohm

s

before frittingafter fritting

So fritting works – are there any drawbacks?•2000TDs, fritting I=-100mA, CRes @ U=1mV (i.e., no current load)

•Very stable CRes 2.0 .. 3.0 Ohms•High fritting current enhances probe pollution -> probe gets ‘addicted’ to fritting


Once more about drawbacks•1000TDs, fritting I=-100mA, CRes @ I=10mA (i.e., with current load)•less degradation (‘before fritting’) than CRes @ U=1mV

1

10

100

1000

10000

1 501number TDs

CR

es in

Ohm

s (@

I=10

mA

)

before frittingafter fritting •susceptibility for degradation

(contamination) depends on load, i.e., power dissipation or current across contact point

•CRes stability and fritting may vary between IO and power pins

•tests with low current values are more sensitive to poor contacts

•test applications need to be evaluated individually

•Again very stable CRes 2.0 .. 3.0 Ohms•Enhanced pollution is less significant if standard testing current is higher!


How significant is fritting enhanced contamination?-> Comparison of fritting OFF / ON / OFF


Fritting OFF-ON-OFF:contact conditioning and subsequent probe degradation

BC

Res

@I=

1mA

nofri

tting C

CR

es@

I=1m

Afri

tting

ON P

CR

es@

I=1m

Afri

tting

OFF D

CR

es@

I=10

mA

nofri

tting E

CR

es@

I=10

mA

fritti

ng O

N LC

Res

@I=

10m

Afri

tting

OFF F

CR

es@

I=30

mA

nofri

tting G

CR

es@

I=30

mA

fritti

ng O

N MC

Res

@I=

30m

Afri

tting

OFF H

CR

es@

I=10

0mA

nofri

tting

IC

Res

@I=

100m

Afri

tting

ON J

CR

es@

I=10

0mA

fritti

ng O

FF

0

10

20

30fritting: I=-100mA(200TD per box)

CRes@I=1mA CRes@I=10mA CRes@I=30mA CRes@I=100mA

CR

es in

Ohm

s

•fritting improves contact stability but makes probes ‘addicted’ to current•negatives effects become negligible for higher testing currents (power pins)


summary / conclusions

FRITTING WORKS!

•A-fritting starts at voltages <2.5V (Tungsten-Rhenium on Al)

•B-fritting significantly lowers CRes•currents >40mA required (with negative polarity)

•fritting makes ‘addicted‘: after turning fritting off, CRes ishigher than before (current assisted contamination)

•amount of CRes improvement depends on current levelduring testing (operation) mode

☺ fritting is suitable to reduce CRes for power pins where negative effects due to enhanced contamination are less significant

fritting very efficiently reduces CRes also for IO-pins – but probe cleaning / maintenance strategy must be adjusted accordingly in order to control enhanced contamination

fritting cannot substitute online cleaning – but it can stabilize CRes in between cleaning executions


perspectives

FRITTING WORKS!

currently used in IFX production•fritting used in several test programs but so far only limited productive experience•application in front-end and back-end test•observation of both, yield improvement and enhanced probe wear out

future tasks•adjust fritting parameters according to new findings•re-evaluate yield improvement vs probe wear out•unified recommendations for test development groups

Thank you for your attention !

parametric study of contact fritting for improved … · parametric study of contact fritting for...

Documents