Chapter 2

Is there a correlation between ionic contamination testing and SIR. The
easy answer is "maybe". Let us take four hypothetical situations:
1) a board is contaminated with, say, salt (e.g. from sea spray) and
nothing else. There would be a close correlation between the two.
2) a board is contaminated with, say, sugar and nothing else. Ionic
contamination would say the board was perfectly clean but SIR would tell
you it's hopeless.
3) a board is contaminated with well-hardened, aged, halide-activated
rosin flux residues. Ionic contamination testing will dissolve just
sufficient to register severe contamination, while SIR testing may
(depending on the conditions) reveal no problems.
4) a board is contaminated with freshly soldered "no-clean" organic acid
activated flux residues. Ionic contamination testing may reveal
impossibly high (by cleaned standards) contamination, while SIR testing
may reveal it to be safe (because the chosen temperature was so high
that the residues sublimated).
So....! Pragmatically, one must say there is no sure correlation and
probably none at all with the usual porridge of contaminants encountered
in practice.

However, there is another enormous difference: ionic contamination
testing is a simple, low-cost (equipment and operating), rapid (15
minutes), workshop-floor, process control system. SIR is a high-cost,
laboratory, slow (days, weeks or months), qualification procedure. Ion
chromatography is a moderately high-cost, laboratory, fairly rapid
(hours), test very useful for identifying ionic contaminants.

They are therefore all complementary to each other and, if you wish to
know the whole story, at all stages, you must do them all, regularly. I
say "the whole story" but, with these three alone, you still will not
have it. For example, spray some mould release (silicone or
fluoropolymer type) or many oils on a circuit and none of these methods
will budge a hair, but the contamination will be there and cause
potential problems (e.g. no adhesion of conformal coatings). In fact, it
is conceivable that some of these contaminants could even improve ionic
contamination or SIR figures! A "pressure cooker test" (vesication test)
could reveal the presence of these contaminants.

_______


Other things that must be considered with SIR testing are:
- absorbed contaminants (epoxy substrates will absorb contaminants of
all types, especially over the Tg or with small molecules). These may
desorb on a time scale ranging from hours to years.
- adsorbed contaminants (hydrogen bond linking with the organic
substrate or chemisorption). This is often fairly permanent if the
adsorbed molecule is stable, but it may have a hydrophilic end on the
outer surface, creating an increase of hygroscopicity with electrical
consequences. Less stable molecules may break down by any one of a
hundred mechanisms (heat or thermolysis, IR, light, UV, up to cosmic
radiation or photolysis, reaction with hydroxyl radicals or hydrolysis,
bacteriolysis, plasmolysis, electrolysis, chemical reactions etc.) to
release harmful contaminants that could lower the SIR
- the substrate chemistry. For example, most epoxy resins are formed
(simplistically) by an initial prepolymerisation, usually of an epoxide
substance (typically epichlorohydrin which has an epoxy ring C-O-C and a
chlorine atom stuck on the end) and a complex phenolic molecule such as
bisphenol A (or tetrabromobisphenol A for flame-retarding). A hydroxyl
group is needed to complete the reaction and this is supplied by a third
component, sodium hydroxide, where the sodium atom has the supreme merit
of capturing the chlorine one, which would otherwise be a darned
nuisance. Yes, you've read it right, there is sodium chloride in the
epoxy prepolymer, although most, but not all, of it is removed in the
electrical grades. This prepolymer (tube A in your hardware store) is
then crosslinked using such horribles as amines or dicarboxylic acids
(contained in tube B). Under ideal conditions, you have exactly the
right number of molecules in tubes A and B that every reactive epoxy
group in the prepolymer clicks into a molecule of the crosslinker, with
no excess of either, and that the mixture is perfectly homogeneous to
allow this. However, this is utopian and there is always a local excess
of one component or the other, especially at the surface. If it is the
prepolymer, it means you have some highly reactive epoxy groups just
waiting to capture anything that comes along. If it is the crosslinker,
then these are usually nicely ionic or can break down into ionic
species. In either case, you can expect some salt molecules near the
surface, as well. All this has an influence on SIR measurements, whether
you are testing the quality of the substrate or trying to find the
effect of possible contaminants.
- moist air. Yes, moist air - and we are talking of 80-95% RH - is a
conductor of the same order as some of the other factors we are looking
at. A student at the Swiss Federal Institute of Technology who was
working on the electrical characteristics of epoxy resins for his
dissertation stretched two 1 mm diameter wires copper wires between PTFE
insulators, 1 mm apart, across a humidity chamber and obtained readings
of the order of 1E9 ohms (if my memory is good, the "test pattern" was
about 50 cm long). Control wires cut off at the insulators were abour
two orders of magnitude higher. This is therefore not negligible.
- history of the measurement. If you apply a voltage gradient to the
substance of a damp epoxy resin, the salt molecules (and any other
ionics) will dissociate and start to migrate. This migration requires
energy which is supplied by the applied voltage in the form of an
additional current (i.e. an apparent lowering of resistance). I don't
know the distances involved, probably in the order of angstroms, but I
do know it takes hours for equilibrium to be reached (i.e. the
resistance reaches maximum). If the gradient is removed, it will take
days for a new electrical equilibrium to be reached, because the
re-associative forces are much lower, but this can be accelerated by
reversing the field which causes an initial relatively high peak of
current as the ions are re-pushed towards one another. This is why the
power factor of most epoxy resins is never brilliant, even at low
frequencies. It is also why it is very important that the SIR
measurements should always be made with exactly the same time interval
between applying the test voltage and actually making the measurement,
especially important if a lower, no or reverse bias voltage is used.

In conclusion of Chapter 2: make sure you know exactly what you are
measuring.

Brian

Bev Christian wrote:
>
> OK, let's see what I can stir up today.  :)
>
> I am looking for a clear, succinct definition of surface insulation
> resistance.  If I look in IPC-9201 "Surface Insulation Resistance Handbook",
> I don't find what I am looking for. (Sorry, Joe, Doug et al).  What I find
> on page 2, section 2.1, Definitions is the following: "It represents the
> electrical resistance between two electrical conductors separated by some
> dielectric material(s).  This property is loosely based on the concept of
> sheet resistance, (see ASTM-D-263), but also contains elements of bulk
> conductivity, leakage through electrolytic contaminants, multiple dielectric
> and metallization materials and air."  Whew!
>
> Any takers?  Hmmm?
>
> regards,
> Bev Christian
> Research in Motion
>
> ---------------------------------------------------------------------------------
> Technet Mail List provided as a free service by IPC using LISTSERV 1.8d
> 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 delivery of Technet send the following message: SET Technet NOMAIL
> Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail 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
> ---------------------------------------------------------------------------------

---------------------------------------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8d
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 delivery of Technet send the following message: SET Technet NOMAIL
Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail 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
---------------------------------------------------------------------------------