Failure Analysis of Ball Shear Test Failures

EESemi.com Forum Archives

Aside from the Wire Pull Test, the Ball Shear Test provides another measure of how well the wirebonding process was performed on an assembly lot. Lots exhibiting ball shear test failures must be properly investigated to ensure that no lots are being shipped with a wirebonding problem that can escalate into a field reliability problem later on. The archived forum thread below examines one such case. Fortunately, the thread starter was able to resolve the issue and identify the root cause of the ball shear test failures to be the dirty oven used for die attach curing.

Posted by Tom Wang: Fri Jul 07, 2006 12:55 pm Post subject: ball shear failed

Hi:
Recently, my company suffered from a trouble. Because of a mass of ball shear low, a majority of of wb machine down. We did EDS and cratering and intermetallics, but no any abnormity was found. And we adjust the wb parameter and wp/bs/ds parameter, again and again. But the trouble is continued still. how difficult for us to get the cause root!!
Any good suggestion, please let me know
thanks a lot
best regards
yours
farel

Posted by FARel Engr: Fri Jul 07, 2006 2:26 pm Post subject:

Hi Tom,

How about your bond pull readings - are they low too? Are all bonds affected - or just localized to certain bonds? Anyway, here's a little suggestion: cross-section bonds from several units. The objective is to look for anomalies such as voids at the following: a) bond pad metal layer; b) Al-Au intermetallic layer; and c) the ball bond itself. You need to know where and how the failure is occurring before you can correct it.

FARel Engr

Posted by Jefriz: Mon Jul 10, 2006 12:31 pm Post subject: Deteriorating BST

Hello Tom,

Were there any changes in the materials, parameter specs, environmental conditions or clamp/insert design? Any changes on the method of doing BST? Do you have SPC monitoring on your bonding temperature? Was the product running already a long time (over 6 months) when the deterioration happened? Have you also consider doing FTIR on the fresh and prior bonding bond pads? Do this on the bondpads after exposing to heat. Do you have periodic cleaning of inserts (workholder) and spark rods?

I am sorry if I cannot give you direct solution on the problem but rather the areas where you can also consider.

Best Regards, Jefriz

Posted by Tom Wang: Mon Jul 10, 2006 2:26 pm Post subject:

On Jul.1'06, we found this issue, but till now, no any one give a convincing answer. trouble is just like a hill. No BST is changed. oh, No, we have not SPC monotoring on bonding temp. we only do ball shear/wire pull for w/b.
what is FTIR?
below is the bst data:
47.5 50.6 34.9 32.3 32.0 34.8 27.5 29.1 37.9 43.9
20.3 21.7 23.9 28.6 23.8 23.3 23.2 23.3 29.8 28.7
58.1 23.6 21.4 12.2 30.9 29.3 21.7 21.8 23.1 17.3
the LCL OF spc is 35
any good suggetion, please help me
rga
farel

Posted by Jefriz: Mon Jul 10, 2006 6:25 pm Post subject: Probability of smashed ball - hill in the ballbond

Hello Tom,

It seems not clear if the product is running a long time. You mean that the ball have an inner depth just like a hill? (a photo will help) BST methodology had no change, which means that your shear tester controls the 3 to 5 um tool height. For FTIR, please check ------/FTIR.htm If your EDS did not show Carbon, skip this suggestion.

General comment:
1) Your parameters still needs DOE
2) SPC on peak temp would later be necessary to stabilize and predict the process.
3) Are you using wire size > 1.0 mil diameter? Your LCL seems a little too high (anyway, I assume it came from SPC study). FYI, LSL for 1 mil diameter based from MIL-STD883F is only 20 grams.
4) Your BST data had a high standard deviation (10) which means large variation and data are skewed right, which means most of the data are near the LCL.

Recommended areas to consider:
1) Maintain ball thickness atleast 0.5x wire diameter.
2) Need to monitor your bonding temperature is stable within your target value. Overheating may result to smashed ball and underheating can cause poor intermetallics and weak bonding.
3) Adjust bondforce only as necessary. The ball could have been smashed so much.
4) You may need to evaluate a new tip angle for your capillary. This is if the hill on the ball is still present.
5) Verify that no blowing air is flowing towards the bonding stage.

These are just few things that might help, considering details are limited. Let me know the response.

Best Regards, Jefriz

Posted by Tom Wang: Mon Jul 17, 2006 12:06 pm Post subject:

Hello friend:

Thanks a lot for your kindly help. Now question was resolved. issue is the dirty oven at D/A PMC. We cleaned it well, the ball shear pass. But I did not inspect contamination by SEM. Why? yes. I know X-RAY from sample s the one from underside of sample surface.
But how can I inspect contamination composition?

Farel Wang

Posted by Jefriz: Mon Jul 17, 2006 6:14 pm Post subject: Oven contaminants

Hello Tom,

Thank you for informing us on the rootcause. Sorry if the oven was not mentioned as a suspect because your EDS result you mention was normal (I assume peaks of Aluminum only). Can you check again your EDS results if you found peaks of Carbon and Oxygen. SEM (high magnified view) would unlikely show the contaminants. Common contaminants are Carbon and Oxygen base which the signature can be characterize by FTIR. Several other composition analysis tools can also be used (pls look on FA tools on this website).

You may also look into your die attach epoxy. Outgassing potentially can cause similar issue even you have clean oven.

As prevention, include a periodic cleaning & PM of your oven (also include SPC monitoring). Don't forget the wirebond stage temp SPC.

To other readers,
I invite others who have experience on characterizing contaminants to comment or add more suggestion. I believe this subject matter is quite interesting.

Best regards, Jefriz

Back to the 'Best of Forums' Archives

HOME