Actually you can get die discoloration by a badly controlled back grinding process - just a thought
		

		Sent from my iPad Pro
	





On Tue, Sep 12, 2017 at 11:45 AM -0700, "Yuan-chia Joyce Koo" <[log in to unmask]> wrote:










thermal runaway is drastic process... once you hit above threshold,  
it all go downhill from there until something busted,  like your  
glass passivation ... your vendor should be able tell you that...  
best luck.  check your vendor's datasheet, if it got restricted  
reflow profile, if you go above that, you might damaged chip to start  
with (ramp  rate included).  my 1.6 cents.
jk
On Sep 12, 2017, at 11:12 AM, Ioan Tempea wrote:

> Thanks Joyce,
>
> Good points. What temperature is necessary to create this kind of  
> damage?
> Can the die hit this temperature if the heat evacuation is not  
> perfect?
>
> Also, the discoloration really goes along all metallized features  
> and is not localized. Something affected the whole surface to the  
> same extent.
>
> Dean,
> EDX did not reveal any contaminant so I wonder if there could be  
> anything else.
>
> Best regards,
>
> Ioan
>
> -----Message d'origine-----
> De : Yuan-chia Joyce Koo [mailto:[log in to unmask]]
> Envoyé : September 12, 2017 10:46 AM
> À : TechNet E-Mail Forum; Ioan Tempea
> Objet : Re: [TN] Die discoloration
>
> over heating?  what is your die attach?  any voiding % more than  
> 25% in any of quadroon?  only one unit or many?  does the part  
> datasheet required heat sink at backplane with multiple stack  
> vias?  any of those vias are good large enough? (datasheet spec for  
> thermal plus 20-30% if your PWB FAB house via recessed to prevent  
> contact... any via outgas resulted die attach problem?)... my guess.
> On Sep 12, 2017, at 9:25 AM, Ioan Tempea wrote:
>
>> Dear Technos,
>>
>> I know many of you, not only Steve Creswick and Wayne Thayer, have
>> experience in microelectronics, so I dare poking you with this
>> situation I'm having.
>>
>> There's this Ka band MMIC die that was properly working in the
>> beginning, only to see its performance degrading over time. Visual
>> inspection revealed a strange discoloration running along all the
>> metallized features and the glassivation is cracked here and there.
>>
>> Please see image graciously placed by Stevie G on his website
>> http://stevezeva.homestead.com/MMIC_Die.jpg
>>
>> What during the assembly process could cause this kind of visual
>> aspect?
>>
>> Can this be a cause of the degrading performance?
>>
>> Is it possible the die was like this from manufacturing (improper
>> pre-glassivation cleaning or something)?
>>
>> Any other impressions?
>>
>> Thanks,
>>
>> Ioan Tempea