It is true that dewetting can be time-dependent. I know of two potential 
explanations:
1. Back to abrasive cleaning: when brushing with fibrous gritty brushes, 
such as Sc****br**e, or with pumice, abrasive particles get implanted 
into the copper that is then swaged over the embedded particles with a 
<1 µm thick layer of metal. Subsequent power rinsing removes the surface 
particles but not the swaged over ones. When soldering, the copper wets 
perfectly initially but, as the thin copper over the particles is 
"eaten" away, forming the IMCs, dewetting can subsequently start. It 
takes only a second or so to pass from good wetting to dewetting.

2. Most PCB laminates are made with electrodeposited copper foil which, 
for reasons of speed of manufacture and thus economy, are deposited as 
rapidly as possible. This produces a coarse crystal structure in the 
copper, growing from the drum side (outer surface of PCB) to a rough 
(subsequently treated) side which promotes adhesion to the laminate. 
This copper is horrendously porous, some microsectioned pores being 
visible under an optical microscope i.e. >0.4 µm across, most being 
smaller. Before pressing the laminate, the glass cloth is impregnated in 
an epoxy resin solution in e.g. a ketone solvent which is then dried in 
a heated tunnel or similar. This hardly cross-links the epoxy prepolymer 
which may not be 100% stoichiometric (i.e., it may have a small excess 
of monomer). This B-stage pre-preg is assembled with the copper foil and 
pressed for a given time at very high pressure and an appropriate 
temperature profile to ensure cross-linking of the epoxy. The crunch is 
that the B-stage still contains, say, 1% solvent. The heating to temps 
of 2-3 x the boiling point causes this to evaporate but the vapours have 
little place to go, except into the pore structure of the copper. In 
doing so, they may take with them any residual monomers or even some 
prepolymer. The pores therefore become lined with a soup of organic 
compounds of almost indeterminate composition. The larger pores may even 
have cross-linked polymer filling them (bleed-through). During board 
manufacture, chemical cleaning processes may remove these organics on 
the surface, providing a solderable surface, but the contaminants may 
lie untouched a µm or so under the surface, causing dewetting after a 
second or so of IMC formation.

A lesser-known addendum to this is that subsequent copper plating is 
often epitaxial to the original copper structure, so that the pores may 
extend through to the new surface and any volatile monomers may reach 
the new surface during soldering, still causing dewetting. This is 
sometimes also ascribed to organic additives (levellers, brighteners, 
throwing power additives etc.) co-deposited in the copper, but I'm sure 
that polymer residues can reach the surface in some cases, especially 
with sulfate baths, as opposed to pyrophosphates.

I researched these phenomena in quite good depth in the second half of 
the 1960s and published them in a paper in 1969. I used a home-built 
automated (including fluxing in 25% WW rosin, drying and solder-bath 
dross removal) edge-dip test with 0 dwell time in the range of 0-10 s in 
a carefully temp controlled pot for determining the point at which 
dewetting would start and, more importantly, how to push it up the time 
scale so that dewetting would not occur during wave soldering. One 
practical discovery was that with ultrasolderable boards from processing 
to reduce dewetting, we could solder ultra-fine line boards with no 
bridging. One such board, I remember, had 4 conductors passing between 
adjacent pads of a dual-in-line package and this always soldered without 
bridging (orientation in the wave soldering machine was important, 
though). Of course, this was in the days before solder masks were 
current. We manufactured the demo boards for three major soldering 
machine manufacturers at that time, one of which, I remember had 35 µm 
spacings, quite a challenge in the late 1960s!

Brian

Louis Hart wrote:
> Thanks, Brian, this explanation makes sense. Mike Sewell sent a link to an article claiming excess time and temp will lead to an intermetallic depletion and attendant poor solder joint. If true, that is something of which I was not aware. Louis
> 
> -----Original Message-----
> From: Brian Ellis [mailto:[log in to unmask]]
> Sent: Thursday, February 19, 2009 12:02 PM
> To: TechNet E-Mail Forum; Louis Hart
> Subject: Re: [TN] dewetting
> 
> The most usual cause of dewetting is myriad tiny unwettable areas in an
> otherwise wettable metal. A very common cause is abrasive cleaning,
> where abrasive particles are implanted into the copper surface, but
> there are other causes. What happens is that IMCs are formed where the
> e.g. copper is solderable and this bridges over the contaminants. When
> the solder flows away, the bridging breaks down, leaving areas with
> thickish solder and areas with just the residual IMC, which is very thin.
> 
> Brian
> 
> 

---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 15.0
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/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100 ext.2815
-----------------------------------------------------