TECHNET Archives

December 2013

TechNet@IPC.ORG

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Wayne Thayer <[log in to unmask]>
Reply To:
TechNet E-Mail Forum <[log in to unmask]>, Wayne Thayer <[log in to unmask]>
Date:
Mon, 2 Dec 2013 18:30:20 +0000
Content-Type:
text/plain
Parts/Attachments:
text/plain (1 lines)
Hi Dean-

I'm not in sync with your observations.  

First, to be clear, we're talking about the paper George gave Steve to post entitled "Brittle_ENIG_Failures.pdf.

Figure 2 shows a completely undisturbed spherical profile for the solder ball at the pcb interface, suggesting no mask interference.  There appears to be some bluish white gunk packed into the gap between the solder mask and the pad, but it didn't affect ball formation.  Could be debris from the polishing, or could be flux residue.

IF mask interferes with ball formation, I expect to see in a cross section the distorted ball and a small gap between the ball and the mask.  This gap is because the mask has a higher TCE than the ball, so during cooldown it contracts a bit more.  This logic opposes your suggestion that you could induce the mask to push up on the ball during thermal cycling.

If the idea of solder mask pushing on solder balls at below solidus temperatures is well-documented, could you please provide a reference?  The TCE numbers just don't add up on the back of my envelope!

While mask defined pads aren't a good solution to all situations, they still have their place:  Most BGA packages I've looked at use them, particularly at finer pitches.  Sometimes the "weakest link" is at the pad/substrate interface, so increasing the pad size without increasing the solder sphere size can make sense.  Also, mask defined pads have better repeatability of wetting since all exposed solderable apertures are identical (yes, you can reach this condition with non-mask defined pads if you do all via-in-pad).  Finally, mask apertures are inherently more repeatable and better defined than etched copper.

I've seen another phenomenon which you've suggested  can be responsible for the observed brittle cracks (improper reflow profile) which could also explain the observed symptoms:  Too hot a profile.   Note that Andy indicated the problem was observed when they ran the BGA alone as opposed to a fully loaded board.  Perhaps the temperature at the part in question is different in these two conditions.

Wayne Thayer




-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Stadem, Richard D.
Sent: Monday, December 02, 2013 11:55 AM
To: [log in to unmask]
Subject: Re: [TN] Separation between the BGA sodler ball and pad on the PCB

My interpretation of the black line is that it is a crack. 

If you look closely in the pictures, in Fig 3 and 4 and especially magnified in Fig. 8, you can see the left end of the "black line" extending as a crack through the solder on the left end of the pad. The origination of the pressure that causes the crack is the soldermask expansion during thermal cycling. On each end of the pad, where the solderball extends down the sides of the pad to the PWB surface, the soldermask is tapered down, but makes full contact with the solderball where they meet at the PWB surface. There is even a small indentation near the left side of Fig. 4 from the soldermask.

During thermal cycling, the soldermask expands in every direction. When the soldermask makes full contact with the solderball completely around the circumference of the pad, the pressure forces the solderball up in the Z-axis, thus causing the crack. In figure 3, it is easy to see the crack started at the left end of the pad where the solder meets the soldermask, and then propagated through the solder joint along the IMF. Then, very near the opposite (right) side of the pad, it extended up into the solderball itself, taking advantage of the porosity of the solder.

This phenomenon is well-documented. It is the reason BGA pads should be completely non-soldermask defined. While the intention of this board layout appears to be non-soldermask defined, there is actually not enough clearance between the edge of the soldermask and the edge of the BGA pads, such that the soldermask comes in direct contact with the circumference of the solderball around the circumference of the pad.

The black line caused by diffusion of nickel that Vladimir describes IS a true observation in other microsections, but is usually only 1 or 2 uinches thick (like as shown in Fig. 6) and is so thin that it is not always seen in microsectional pictures. That is not what we are seeing in Fig 4. What we have there is a crack.

The root cause of the crack may not be just the soldermask contact but also an improper reflow profile. The IMF is usually shown as a 2-4 uinch layer between the nickel and the solder. In Fig. 4 the IMF layer is so thin that it is almost non-existent, especially on the right end of the pad. This is because nickel's dissolution into solder happens at a much slower rate than copper. The reflow profile must be slightly hotter and for a little longer time above solidus temperature (TAST) in order to develop a good IMF with the nickel. It does not appear that this is consistently happening, according to what I see in the various figures.

My 5 bucks worth.
dean


-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Vladimir Igoshev. PhD
Sent: Thursday, November 28, 2013 1:43 PM
To: [log in to unmask]
Subject: Re: [TN] Separation between the BGA sodler ball and pad on the PCB

 Hi George!

 With all due respect I have to say that none of  the failures in the paper were caused by Black Pad.nd you of cause know that any solder joints formed on an ENIG finished pad will have that 'Black Ni Line"
which is simply a region depleted of Ni as it diffused out to form intermetallics.

Best regards,

Vladimir Igoshev. PhD                          mailto:[log in to unmask]

SENTEC Testing Laboratory Inc.
11 Canadian Road, Unit 7.
Scarborough, ON M1R 5G1
Tel: (416) 899-1882
Fax: (905) 882-8812
www.sentec.ca

Thanks! Got it posted, here 'tis:

http://stevezeva.homestead.com/Brittle_ENIG_Failures.pdf

This is why I love the Technet, what may be old to you may be gold to others...I'd venture there are those who have never seen this paper, myself included.

Once again thanks!!

Steve 

-----Original Message-----
From: Wenger, George M. [Contractor] [mailto:[log in to unmask]]
Sent: Wednesday, November 27, 2013 1:39 PM
To: TechNet E-Mail Forum; Steve Gregory
Subject: RE: [TN] Separation between the BGA sodler ball and pad on the PCB

Hi Steve,

Attached is a really old PDF file of ENIG brittle interface fractures that show how you can get a lot of useful information from a well polished and micro-etched cross section.  Although the failures in this file are really old tin-lead failures before switching to Pb-Free, if you have room to post the file it might be helpful to Andy and others.
 

Regards,
George
George M. Wenger
Failure Signature & Characterization Lab LLC
609 Cokesbury Road, High Bridge, NJ 08829
(908) 638-8771 Home (732) 309-8964 Mobile
E-mail: [log in to unmask]
[log in to unmask]


-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Steve Gregory
Sent: Wednesday, November 27, 2013 12:46 PM
To: [log in to unmask]
Subject: Re: [TN] Separation between the BGA sodler ball and pad on the PCB

Hi Andy,

Got all of your pictures and attachments. All the cross-section pictures are here:

http://stevezeva.homestead.com/A1.jpg
http://stevezeva.homestead.com/A16_1.jpg
http://stevezeva.homestead.com/A16_2.jpg
http://stevezeva.homestead.com/T16_1.jpg
http://stevezeva.homestead.com/T16_2.jpg

Your dye & pry is here:

http://stevezeva.homestead.com/256BGA-mod_for_TechNet.pdf

Lastly, here is the reflow profile:

http://stevezeva.homestead.com/Andy_s_Reflow_Profile.jpg


Steve 

-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of andi1978
Sent: Wednesday, November 27, 2013 6:56 AM
To: [log in to unmask]
Subject: Re: [TN] Separation between the BGA sodler ball and pad on the PCB

George,
 First apologize for not timely replying to this topic, I am living within Europe time zone.
Here more facts about this condition and as well to clarify some questions already posted.
I did re-sent pictures again and in addition will send 2 pages of our dye &amp; pry report document we did in March this year (from the same subcontractor).
As George W. already indicated we can eliminate lab issue and poor sample preparation doing the dye &amp; pry test. Fortunately I am in possession of such a report from March 2013 before our BGA cross-section from September 2013. The short message is that there are evidences of dye ingression at the PCB pad level and only existing in the corners of the BGA package. To clarify George's question on completed inter-metallic fusion. What I can see from pictures I sent to Steve, the solder paste perfectly melted with the BGA solder balls and the sample board was fully cooled. The cross-section was done after the initial reflow (single thermal processing). 
The whole board is double sided assembly but because of the need for cross-section, only the BGA component was populated (no other parts were assembled).
There was a request to send the reflow profile to Steve, have done that now. That's all I know and far I can go myself.
I kind of excluded PCB warpage due to small size of the sample board (65mm/70mm/1.5mm) and that this PCB was held in the thermal panel secured with 4 clamps during reflow (single pass).
Regards,
Andy
27 November 2013 3:02 "Wenger, George M. [Contractor]" &lt;[log in to unmask]&gt; wrote:
Andy,
Once Steve has your photomicrographs of the cross section posted it will be much easier to provide meaningful advice. Without a photograph and additional information one can only guess. You indicated separation occurred after the reflow process was completed because there is evidence of completed inter-metallic fusion. I assume when you say "after the reflow process you most likely mean after the solder melted and then cooled below liquidus. Is this a single sided reflow assembly or a double sided reflow assembly? Where the separations discovered after the first reflow or the second reflow? 
Regards,
George
George M. Wenger
Failure Signature &amp; Characterization Lab LLC
609 Cokesbury Road, High Bridge, NJ 08829
(908) 638-8771 Home (732) 309-8964 Mobile
E-mail: [log in to unmask]
[log in to unmask]
-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Andrzej Zielinski
Sent: Tuesday, November 26, 2013 10:33 AM
To: [log in to unmask]
Subject: [TN] Separation between the BGA sodler ball and pad on the PCB Dear Colleagues ..
I am facing an issue with the BGA part where in three locations we found separation between the BGA solder balls and pads on the PCB (ENIG finish).
This condition was discovered during the cross-section that we have requested from our local lab.
I can share few pictures I have taken from our report (can't attach the whole document as this report is confidential to our business).
Basically this BGA is the plastic ACTEL package 896 pins, SAC305 finish, soldered with leaded process using 62Sn/36Pb/2Ag solder paste. Temperature was slightly elevated to accommodate SAC305 solder finish from the BGA package.
The separation condition we have observed was only found in location A1, A16 &amp; T16 which are the corners of the BGA package. As well all 3 separations occurred after the reflow process was completed because there is an evidence of completed inter-metallic fusion. 
In location A16 &amp; T16 you can see the copper is protruding of the pad but I have been told this is not an issue and cause of the separation.
I was considering package or PCB warpage as the cause of this separation or something related to cooling the board after reflow.
I asked our lab to measure the height of the solder balls across row A &amp; T and they all deem to be ok.
I have asked to check for the phosphorus content as the pad level to see if this is related to black-pad symptom and the phosphorus was measured as 10.3%.
So far the board assembly house is adamant their reflow profile is correct.
PCB manufacturer is confident with their PCB and pointing to the assembly house.
There is a thought about poor sample preparation where the separation issue would be caused by the lab (during cross-section).
We are attempting to perform another cross-section on different board but here I am not really sure if this is the right way to go (extra cost).
Because I have never seen such a defect, I am not sure what should be the next step to take.
Any advice guys ?

--
Best regards,

Vladimir Igoshev. PhD                          mailto:[log in to unmask]

SENTEC Testing Laboratory Inc.
11 Canadian Road, Unit 7.
Scarborough, ON M1R 5G1
Tel: (416) 899-1882
Fax: (905) 882-8812
www.sentec.ca

______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask] ______________________________________________________________________

______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask] ______________________________________________________________________

______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask] 
______________________________________________________________________

ATOM RSS1 RSS2