Fritting: Experiences with Non-Ohmic Contact Resistance while ...
Fritting: Experiences with Non-Ohmic Contact Resistance while ...
Fritting: Experiences with Non-Ohmic Contact Resistance while ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>Fritting</strong><br />
<strong>Experiences</strong> <strong>with</strong> non-ohmic contact<br />
resistance C RES <strong>while</strong> wafer test probing<br />
Jan Martens<br />
Philips Semiconductors Hamburg<br />
Awarded “Best Overall Presentation at SWTW-2006”
Overview<br />
• <strong>Fritting</strong> – What‘s that?<br />
• <strong>Fritting</strong> – How to measure?<br />
C RES measurements <strong>with</strong>in production<br />
environment<br />
• Definition of the <strong>Fritting</strong> ratio Φ<br />
• Why measuring <strong>Fritting</strong> and not only C RES?<br />
• Measurement results<br />
• Summary and Outlook<br />
06/12/2006 SWTW2006 – Jan Martens 2
<strong>Fritting</strong> – What‘s that?<br />
• The vertical Probe tip<br />
touches the contact pad.<br />
• Depending on the contact<br />
pressure the oxide film is<br />
broken partly and<br />
electrical bridges arise.<br />
• The number and size of<br />
the bridges is equivalent<br />
to the C RES quality<br />
Probe Tip<br />
<strong>Contact</strong> Pad<br />
06/12/2006 SWTW2006 – Jan Martens 3
<strong>Fritting</strong> – What‘s that?<br />
Probe tip<br />
Oxide film<br />
<strong>Contact</strong> pad<br />
• What happens, if bridges<br />
are only few and small?<br />
Small bridge through oxide film.<br />
Before high current flow.<br />
06/12/2006 SWTW2006 – Jan Martens 4
<strong>Fritting</strong> – What‘s that?<br />
• Current must flow through small bridge.<br />
• Bridge and neighbourhood are heated up<br />
• <strong>Contact</strong> Pad material migrates to the bridge.<br />
Probe tip<br />
Oxide film<br />
<strong>Contact</strong> pad<br />
High current flow situation:<br />
Black Lines of current flow.<br />
White Lines of equipotential surface.<br />
06/12/2006 SWTW2006 – Jan Martens 5
<strong>Fritting</strong> – What‘s that?<br />
• Bridge is widened C RES decreased<br />
• <strong>Contact</strong> pad material migrated to the bridge and tip<br />
surface<br />
Probe tip<br />
Oxide film<br />
<strong>Contact</strong> pad<br />
Wide bridge through oxide film.<br />
After high current flow.<br />
Tip surface is contaminated.<br />
06/12/2006 SWTW2006 – Jan Martens 6
<strong>Fritting</strong> – What‘s that?<br />
Probe Tip<br />
<strong>Contact</strong> Pad<br />
• <strong>Fritting</strong> is a kind of electrical breakdown at the<br />
contact surface between the probe tip and the<br />
contact pad of the IC.<br />
• It improves the electrical contact by building or<br />
stabilizing bridges through the oxide film, if the<br />
film was not mechanically broken completely.<br />
• After <strong>Fritting</strong> the probe tip is welded <strong>with</strong> the<br />
contact pad. After removing the contact<br />
residuals of the welding remain at the probe<br />
tip and will oxidize.<br />
06/12/2006 SWTW2006 – Jan Martens 7
<strong>Fritting</strong> – How to measure?<br />
Needs:<br />
• A C RES measurement is needed before and after fritting.<br />
• The change of C RES is the indicator of occurence.<br />
Implementation:<br />
• C RES Monitor is installed twice <strong>with</strong>in the production test<br />
program flow of a smart card measuring C RES on the<br />
power supply pin for every IC.<br />
• First monitor at test program start, last monitor at the end.<br />
• Hardware used is Teradyne J750, Accretech (TSK)<br />
prober UF200A and 32-multisite vertical probe card<br />
Viprobe from Feinmetall.<br />
• Data of first and last C RES per touchdown is collected and<br />
analyzed.<br />
06/12/2006 SWTW2006 – Jan Martens 8
<strong>Fritting</strong> – How to measure?<br />
Last CRES CRES / Ω<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
First C RES = Last C RES<br />
Stable contact!<br />
CRES CRES Scatter Plot<br />
First C RES > Last C RES<br />
<strong>Fritting</strong> contact!<br />
0 2 4 6 8 10<br />
First CRES CRES / Ω<br />
06/12/2006 SWTW2006 – Jan Martens 9
<strong>Fritting</strong> – First data review<br />
Number of TDs<br />
5000 5000<br />
4000 4000<br />
3000<br />
2000<br />
1000<br />
0<br />
CRES CRES Distribution<br />
Distribution<br />
C RES Improvement<br />
0 22 44 66 8 10 10<br />
CRES CRES / / Ω<br />
First Cres<br />
First Cres<br />
Last Cres<br />
06/12/2006 SWTW2006 – Jan Martens 10
Definition of <strong>Fritting</strong> ratio Φ<br />
Number of TDs<br />
5000<br />
4000<br />
3000<br />
2000<br />
1000<br />
0<br />
CRES CRES CRES Distribution<br />
<strong>Fritting</strong> Ratio Φ<br />
=<br />
0 2 4 666 8 10<br />
CRES CRES CRES / Ω<br />
NFrit<br />
+ N<br />
First Cres<br />
First Cres<br />
First First Cres Cres / Stable<br />
First Cres / Stable<br />
First Cres / <strong>Fritting</strong><br />
06/12/2006 SWTW2006 – Jan Martens 11<br />
N<br />
Frit<br />
Stab
<strong>Fritting</strong> ratio Φ. Typical examples<br />
<strong>Fritting</strong> ratio Φ<br />
100%<br />
90%<br />
80%<br />
70%<br />
60%<br />
50%<br />
40%<br />
30%<br />
20%<br />
10%<br />
0%<br />
Change of <strong>Fritting</strong> ratio Φ per time<br />
Within 2k TDs<br />
Φ80%<br />
example 1<br />
example 2<br />
0 500 1000 1500 2000 2500<br />
Touchdowns<br />
06/12/2006 SWTW2006 – Jan Martens 12
Why measuring <strong>Fritting</strong> and not only C RES ?<br />
• C RES measurement only indicates a problem in<br />
the signal path but doesn‘t answer where.<br />
<strong>Fritting</strong> measurement verifies a probe tip<br />
contamination.<br />
• In cases where an absolut C RES measurement is<br />
impossible, a relative C RES decrease can still be<br />
measured.<br />
• The fritting analysis is more convenient and<br />
simpler to analyze.<br />
06/12/2006 SWTW2006 – Jan Martens 13
How influencing the fritting ratio?<br />
• Changing overdrive<br />
• Probing on gold bumps instead of<br />
aluminum<br />
• Using refurbishment methods<br />
– Probe Polish (online cleaning)<br />
– T.I.P.S. (offline cleaning)<br />
06/12/2006 SWTW2006 – Jan Martens 14
Changing overdrive<br />
<strong>Fritting</strong> ratio Φ<br />
100%<br />
90%<br />
80%<br />
70%<br />
60%<br />
50%<br />
40%<br />
30%<br />
20%<br />
10%<br />
0%<br />
<strong>Fritting</strong> ratio Φ<br />
Overdrive improves Φ<br />
(nonlinear probing force only small improvement)<br />
2 Mil 3 Mil 4 Mil 5 Mil<br />
Overdrive<br />
06/12/2006 SWTW2006 – Jan Martens 15
Probing on gold bumps (instead of aluminum)<br />
• Searched in many production lots<br />
• <strong>Fritting</strong> ratio always Φ=0%<br />
Gold is not oxidized.<br />
No <strong>Fritting</strong>!<br />
06/12/2006 SWTW2006 – Jan Martens 16
Probe tip refurbishment – How does it work?<br />
Probe Tip<br />
Highly cross-linked polymeric<br />
material <strong>with</strong> spatially distributed<br />
abrasive particles. (Probe Polish)<br />
Flattened tip<br />
Semiround tip<br />
06/12/2006 SWTW2006 – Jan Martens 17<br />
Pictures from T.I.P.S. Messtechnik GmbH
Online cleaning method: Probe Polish (ITS)<br />
<strong>Fritting</strong> ratio Φ<br />
90%<br />
80%<br />
70%<br />
60%<br />
50%<br />
40%<br />
30%<br />
20%<br />
10%<br />
0%<br />
<strong>Fritting</strong> ratio Φ<br />
Online using<br />
Probe Polish<br />
(1000 touchdowns)<br />
<br />
Probe tip<br />
Refurbishment<br />
<br />
<strong>Fritting</strong> prevention and<br />
Yield win 1%<br />
0 200 400 600 800 1000 1200<br />
Touchdowns<br />
06/12/2006 SWTW2006 – Jan Martens 18
<strong>Fritting</strong> measurement and Probe Polish<br />
Number of TDs<br />
3000<br />
2500<br />
2000<br />
1500<br />
1000<br />
500<br />
0<br />
CRES CRES Distribution<br />
First Cres / before cleaning <strong>with</strong> Probe Polish<br />
First Cres / before cleaning <strong>with</strong> Probe Polish<br />
Last Cres / before cleaning <strong>with</strong> Probe Polish<br />
First Cres / after cleaning <strong>with</strong> Probe Polish<br />
Last Cres / before cleaning <strong>with</strong> Probe Polish<br />
Last Cres / after cleaning <strong>with</strong> Probe Polish<br />
<strong>Fritting</strong> prevention<br />
<strong>with</strong> Probe Polish<br />
0 2 4 6 8 10 10<br />
CRES CRES / ΩΩ<br />
06/12/2006 SWTW2006 – Jan Martens 19
Offline cleaning method: T.I.P.S.<br />
<strong>Fritting</strong> ratio Φ<br />
100%<br />
90%<br />
80%<br />
70%<br />
60%<br />
50%<br />
40%<br />
30%<br />
20%<br />
10%<br />
0%<br />
<strong>Fritting</strong> ratio Φ<br />
Offline using<br />
T.I.P.S.<br />
(24k touchdowns)<br />
<br />
Probe tip<br />
Refurbishment<br />
<br />
<strong>Fritting</strong> prevention and<br />
Yield win 3%<br />
0 500 1000 1500 2000<br />
Touchdowns<br />
06/12/2006 SWTW2006 – Jan Martens 20
<strong>Fritting</strong> measurement <strong>with</strong> T.I.P.S.<br />
Number of TDs<br />
5000<br />
4000<br />
3000<br />
2000<br />
1000<br />
0<br />
CRES CRES Distribution<br />
0 2 4 6 8 10<br />
CRES CRES / Ω<br />
First Cres w/o T.I.P.S.<br />
First Cres w/o T.I.P.S.<br />
Last Cres w/o T.I.P.S.<br />
First Cres w/ T.I.P.S.<br />
Last Cres w/o T.I.P.S.<br />
Last Cres w/ T.I.P.S.<br />
<strong>Fritting</strong> prevention<br />
<strong>with</strong> T.I.P.S.<br />
06/12/2006 SWTW2006 – Jan Martens 21
Summary<br />
• <strong>Fritting</strong> was detected by comparing C RES values<br />
during one touchdown.<br />
• <strong>Fritting</strong> ratio Φ was defined and used as C RES<br />
quality indicator.<br />
• <strong>Fritting</strong> contaminates probe tips.<br />
• <strong>Fritting</strong> was prevented and yield was increased by<br />
using probe tip refurbishment techniques:<br />
Probe Polish www.inttest.net<br />
T.I.P.S. www.tips.co.at<br />
06/12/2006 SWTW2006 – Jan Martens 22
Outlook<br />
• <strong>Fritting</strong> ratio Φ can indicate preventive<br />
maintenance of probe cards.<br />
• <strong>Fritting</strong> ratio Φ can help to optimize<br />
cleaning setups.<br />
• <strong>Fritting</strong> can help to identify and engineer<br />
(too) high current tests.<br />
• <strong>Fritting</strong> could be an explanation for a fast<br />
abrasion of probe tip material.<br />
06/12/2006 SWTW2006 – Jan Martens 23
Thank You!<br />
Questions?<br />
06/12/2006 SWTW2006 – Jan Martens 24