TECHNET Archives

August 2002

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:
Reply To:
TechNet E-Mail Forum.
Date:
Tue, 13 Aug 2002 08:31:52 -0500
Content-Type:
text/plain
Parts/Attachments:
text/plain (207 lines)
Doug Wise-Guy-Person

**Who-doggies.  This ones gonna be a 3 Dew response.  Pardon me whilst I go
tank up.......Ahhhh, much better.

Very many thanks for the time you've taken to share such a full and
valuable response to my questions. If I may, I have some supplementaries:

**My pleasure.  I get a lot out of this forum and am happy to give back
when I can. In your questions, you essentially ask for an entire semester
of analytical or instrumental chemistry.  If you or other 'netters want to
really get into residue analysis, go to the nearest major university and
talk with some of the professors who teach intrumental analysis.  I have
found most are willing to talk about the ins and outs of analysis,
especially if you buy them lunch.

1. I have a theory from what you've said that may stand some debate. If the
two main elements of this problem are (a) identifying residues on a board
and (b) carrying out temperature/humidity tests to failure then analysing
the failure modes, then I think it is best to try and identify the residues
in situ. By identifying them in situ, one can them build a 'map' of where
they're located on the board and this can be used for correlation to the
failure results. What is your opinion on this?

**Is in-situ measurement preferrable to extraction testing?  Yes, but
dramatically harder, at least with the methods that I am familiar with.
Take two common methods, for example, surface reflectance FT-IR and
SEM-EDX.  In the FT-IR case, an infrared beam is bounced off a surface into
a detector and the absorption of radiation measured as a function of
frequency.  If the residue is thin, you will often get effects from the
polymeric solder mask or laminate under the residue, complicating the
analysis.  I have seen some FT-IR operators who can operate their equipment
like Yo Yo Ma does a cello, but they are rare.  For SEM-EDX, you can shoot
the X-ray through the sample and get a reading of what chemical elements
are present, but it is difficult to tell if the chemicals are part of the
surface residue, the bulk material, or the back side of the sample.  It is
possible to do ion milling and ion mapping with depth, but damn, it is very
expensive and very slow.  I have seen failure analysis reports in the past
where SEM-EDX on an FR-4 board showed high levels of bromide as the cause
for corrosion, but bromide was a part of the epoxy resin itself, not a
surface residue free to cause corrosion.  Another shortcoming of these
kinds of test is that most deal with a very focused spot on a substrate,
either an IR beam or an X-ray beam, neither are visible and often times
your residues are not visible either.  How confident are you that you have
managed to hit the residue, and the right residue, with this invisible
microscopic spot?  It can be done, if you know what you are doing.  So, the
answer to your question is yes, you can map residues in-situ, but you need
VERY experienced operators of the analytical equipment who know what to
look for and what to avoid, and you need some VERY deep pockets.

What are the main drawbacks to measuring in-situ and measuring extracted
residues? The main drawback to measuring extracted residues are twofold, as
far as I can see - (i) you don't know where on the board they came from and
(ii) you don't know local concentrations. Can one tell concentrations by
measuring in-situ, though?

**The main advantage of an extraction based test is that you can analyze
the extract with a wider array of instruments, and at a lower cost,
compared to the in-situ tests.  The main question to ask in any extraction
operation is:  is the soil/residue of interest soluble in my extract media,
at the temperature of extraction, in the time of extraction?  Lets look at
the two most common extractions:  SEC testing and ion chromatography
testing.  In ROSE/SEC testing, you put an assembly into 75% isopropanol /
25% DI water, for 10 minutes, at 110F.  If your residue is not soluble in
that solution, you won't see a result.  If the soil won't dissolve in 10
minutes, you won't see a result.  If the soil is not soluble at 110F, you
won't see a result.  If the soil is not electrically conductive, you won't
see a result.  Ion chromatography extraction is more rigorous, using 60
minutes at 80C, but the same reasoning applies.  In addition, as Susan
Mansilla pointed out recently, you get ionic residue not only from the
surface of the substrate but often from the bulk of the substrate as well.
Bromide fire retardant is one such material.

**In your two-part question, the answer is yes and no, or maybe and maybe.
If you do an extraction on the whole board or assembly, then you get an
averaging effect of the residue normalized to the extracted area.  This is
fine when you look for a general assessment of a board/assembly, but
unacceptable when you are doing a failure analysis of a particular area.
There are ways in which you can localize an extraction to a focussed area
of interest, such as a QFP or LCC, if you know what you are doing and use
the right chemistries.  As an example, I have extracted the central
cavities of ceramic integrated circuits.  I did not want to extract the
whole chip, because that would give me plating residues from the lead
plating/forming operation.  I wanted residues specific to the central
cavity because I suspected residues from the gold plating operation.  Using
precise air-displacement pipettes, I dispensed 80C isopropanol/water
solution into the well, waited 60 seconds, then collected the solution.
Doing this repeatedly got me plenty of solution to analyze and I was able
to make my determination. Localized extractions can be done if you can
localize the extraction media to the area of interest.  Again, you can do
it if you know what you are doing.


Do you (or anyone else) know if there is a software package that can map
contaminant locations and concentrations/levels on a board? What measuring
equipment can the computer be linked to to provide this map?

**Not that I know of.  SEM/EDX or Auger analysis can give you elemental
maps and most of the current software packages can do this for you, but it
is the only application I am aware of.  I am told that an experienced SEM
operator can give you concentrations from the strengths of the responses,
but I have yet to see this done reliably.

2. Residue problems seem to fall largely into two camps - corrosion and
dendrite growth. Is this fair to say? Do you need to be much of a chemist
to determine if/when residues are likely to be detrimental to adjacent
materials that are part of the board, or is there a good layman's
guide/chart to indicate reactivity?

**If they are in two camps, they are co-located.  They are both aspects of
the same problem.  An ionic material combines with moisture to form an
acidic electrolyte.  This can corrode metal, moving it into solution, where
an electrical potential creates a small reverse plating cell.  Even if
there is no corrosion, the electrolyte can cause electrical leakage.  This
is frequently the cause of intermittent failures in high impedance
circuitry.  Do you have to be a good chemist to understand this?  Gods I
hope not or I am out of a job.  If you want a top notch chemist, go talk to
Karen Tellefsen or Al Schneider at Alpha Metals.  In terms of reactivity,
you can look at the electronegativity scales of elements found in most CRC
Handbooks.  You would find the fluoride is more electronegative than
chloride which is more than bromide, more than sulfate, etc.  This is why
the fluxes have been moving away from halide activators.  The chloride and
bromide residues are very electronegative and so represent more of a risk
for electrochemical failures.  The dicarboxylic acids used as the
activators in low residue fluxes are not very electronegative, which makes
them a much smaller risk for these types of failures, provided they are
fully reacted.

3. You talked generally about the expected levels of various ionic
contaminants that might be found on boards of a certain material, with a
certain finish, etc - are you talking about acceptable levels after
cleaning or levels that might be detected prior to cleaning? It's what
post-cleaning levels need to be for reliability that we're interesting in
determining. They seem to be pre-cleaning levels, so we're still stuck with
finding reliable post-cleaning levels.

**I am talking about wherever you look for them.  It can either be on the
finished board or prior to solder mask application.  The guidelines I gave
were for finished boards, or post-cleaned boards to use your terms.  You
need to be even cleaner than that prior to solder mask/resist application
since the solder mask will lock in any residues after it is applied.

4. Can you comment on whether or not IPC-TM-650 covers all known test
methods, or are there other good ones outside of that spec? How might I
reliably find out about them, given that this process is only a small part
of my working life and I work largely alone (not many/any people here seem
particularly interested in the minutiae of processes, so I don't know how
they determine reliability and causes of failure without this
understanding)

**Does the TM-650 cover them all.  No.  Are there other good ones?  Yes.
Keep in mind that the TM-650 is put together by volunteers who want to get
standardized testing into the industry.  Anyone can introduce a test method
for consideration.  Susan Mansilla, Robisan Labs, chairs the test methods
committee and is usually looking for new people to help us find new better
more repeatable methods.  There are new techniques being developed all the
time.

5. What I was trying to get at with the PCB vs. PCA thang is should the
acceptable cleanliness level for a PCB be different (cleaner) from the ACL
for an assembly and if so, why. How clean do I tell my fab house to make
the bare boards?

**Yes it should be. Why?  Because I said so.  So There!  As I said before,
fabrication residues are very hard to adequately remove once the board has
been exposed to a reflow process.  So if the boards come in dirty, even if
you have a good cleaning process as part of assembly, there is no guarantee
that you will get those residues.  If you have a no-clean assembly
operation, you have NO opportunity to remove such residues.  In the latter
case, the quality of your produced hardware is held hostage to the incoming
cleanliness of parts and components.  So, yes, you need every bit of
cleanliness margin you can get or can afford.  How clean do you tell your
fabricator to be?  It depends.  See yesterdays response.

6. Following right along, if the bare board residues are so hard to remove
come assembly time, what contaminant testing/measuring methods are
effective? - surely not the extraction type, or I would use the measuring
equipment to clean the board.

**Why not?  Any residue detection method can be used, provided you know
what you are doing and have a firm grasp of the advantages and limitations
of the method used.  If you have a residue that defies extraction, change
the extraction.
You will wander into uncharted territory, but that is where the excitement
is.  Also, I did not say that the residues couldn't be removed, it is just
harder to do, and I was talking more about conventional cleaning processes
(like in-line spray under immersion).  Extraction protocols are usually a
little more rigorous and can get more material into solution.

Enough for now (more than, probably)

**Yah, same here.  Am wired now as the third Dew has fully kicked in.

Doug Pauls
Rockwell Collins

---------------------------------------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/html/forum.htm for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315
---------------------------------------------------------------------------------

ATOM RSS1 RSS2