Improving Yields of Innerlayers and Flexible Composites with One Step Cleaning Without a Microetch

James M. Taylor
Duratech Industries, Inc.
Orange, CA

Maintenance of dimensional tolerances on thin core material is maximized when process steps are minimized and surface roughening by microetching or other means is eliminated.

There are any number of ways to clean copper. With the trend towards thinner core materials and ever denser circuitry, the problem with cleaning large sheets of thin core material is dimensional distortion. Because of this, the circuit industry is moving toward total chemical processing with a concomitant reduction in process steps.Each additional step beyond what is absolutely necessary brings with it the threat of diminished yields.

Regardless of how many steps are employed there are certain requirements that must be met before laminating dry film to an innerlayer or flexible circuit panel, and these requirements have not changed in well over thirty years. First, the panel must be cleaned of organic contaminates, fingerprints and oxides, and the chromate conversion coating removed. Second, the surface is etched and, finally, deoxidized and passivated in sulfuric acid. The need to reduce the number of process steps is fueling the increased use of drum-side-treated-foil (DSTF) and so-called "double-treat", or reverse-side-treated-foil, in order to eliminate the microetch surface structuring otherwise necessary to achieve good resist adhesion.

The problem with conventional chemical cleaning of DSTF is that without a microetch to aid the acid cleaner in attacking and removing the chromate layer, there simply is no way of being sure it is gone. In spite of improvements in the coating of chromates over the years, the time it takes for complete removal can and does vary widely. I still run across occasional lots of innerlayer material in my laboratory that exhibit substantial differences in chromate removal rates, not only side to side, but end to end on the same side of a single panel. In a production setting there simply is no way to tell if you have an entire lot of such material. Depending on your dry film it might not matter. Then again, it could be a cause of poor resist adhesion with resulting edge lift, or too much adhesion causing developing, etching, and stripping problems.

Discrepancies in chromate removal rates can probably be related to coating thickness or to formula differences from different fabricators. Whatever the reason, the effect is the same, and the need for a simple means to identify these discrepancies when they occur is an important one when no microetch is going to be used.

Without exception, all the physical and chemical processes used to prepare substrate materials for dry film lamination are based on a single concept: that a clean board is a water-wet board that has no water-break. While this concept can tell you about oils and other non-water soluble contaminates, it can tell you nothing about the chromates because the chromate layer will wet-out just as well as the underlying copper.

A completely new and different cleaning technology addresses this need by combining in a single processing chemistry the ability to clean panels of organic surface contaminates and oxides, remove chromate conversion coatings, and apply an adhesion promoting layer that eliminates the need for a microetch or pre-roughened (DSTF) foil to achieve the degree of dry film adhesion necessary for fine line work. This cleaning technology identifies un-removed chromates by reversing the normal cleaning process and replacing/displacing contaminates with an extremely hydrophobic (water-repelling) layer that bonds only to clean copper. The formation of the adhesion promoting, water repelling layer will not take place on chromates, only on the underlying copper. Thus panels cleaned in the bath will show immediately the status of any remaining chromates on the panel surface by the presence of water clinging to and wetting those areas after rinsing. Water will be repelled from the rest of the panel. In such cases, panels can be re-cleaned and adjustments made in the process to assure complete removal on subsequent panels.

Identifying un-removed chromates is only one facet of this cleaning methodology. Another is the ability to process dry films on smooth, non-surface roughened copper. No microetch is used or needed. And even though strong acids are used to attack and remove the chromate layer, the underlying copper is inhibited from attack by those same acids due to the hydrophobic layer which forms after the chromates are dissolved away. It is that layer which functions to retard oxidation and bond dry film and KaptonŽ coverlays once panels are rinsed and dried.

This technology has given rise to some very interesting phenomena. That there is a substantial increase in the adhesion of most dry films is without question. But other unusual things are happening as well. Dry films laminated to smooth, un-etched copper processed through this cleaning system exhibit a degree of flexibility they do not normally possess. A number of different dry films tested on the hydrophobic surface produced by this cleaner were laminated on thin core flexible substrates in our laboratory and exposed to 100 millijoules of UV radiation. The laminate was then folded over and creased as one would a piece of paper and the dry film did not crack or otherwise lose adhesion. Films were also cross-hatched with an Xacto knife, the laminate again folded and creased, and the same results achieved. Tests such as these were repeated many times during the development of this product.

The question naturally arises, if the adhesion is that good how do you strip the resist? The answer is that the stripping process is unchanged with the only difference being cleaner stripping. Surprising as it may seem, the dry film comes off without leaving the usual thin resist residue. The reason appears to be that the resist does not come in contact with the copper.

Chemically speaking, copper is a very catalytic metal. When a dry film resist is laminated to bare copper that has been cleaned in the conventional manner and exposed to actinic radiation, a free radical mechanism is set in motion that causes the acrylic monomers in the resist formula to polymerize. When the light source shuts off the polymerization process stops (for all practical purposes), except at the resist/copper interface. Here the catalytic effect of the copper causes a slow hardening of the polymer that is a cause of the film becoming brittle and to also leave a thin, almost monomolecular residue on the copper that can be difficult to remove if panels are allowed to sit too long before stripping. On the other hand, aqueous dry films laminated over the hydrophobic surface produced by the cleaning system that is the subject of this paper, simply do not exhibit this effect. Laboratory testing could determine no difference in stripped substrate cleanliness between day-old laminated parts and those of several months.

It is for this reason that flex circuit manufacturers are able to process from imaging to coverlay without a microetch. The panels are cleaned in a single step, imaged, developed, etched, the dry film stripped, and the panels sent back through the same one-step cleaner to prepare for laminating coverlays. KaptonŽ coverlay materials for encapsulating flexible circuits bond to this surface with such tenacity that once soldered it cannot be removed. Prior to solder, the peel strength is nearly double what it is on bare copper cleaned in the conventional manner.

The concept of cleaning and displacing contaminates, rather than dissolving them for later re-deposition, is applicable to other metals and alloys. This is the reason that chemical milling materials such as Alloy 42 (iron w/42% nickel), also known as Kovar, can be cleaned in a single cleaning step in less than a minute with this technology. Because of the hydrophobic layer, the surface structure of the metal substrate, whether roughened or as smooth as glass, makes absolutely no difference with respect to the adhesion of the dry film. Successful testing has been carried out on sputtered copper so smooth it looked like a mirror. Other metals include brass, bronze, aluminum, nickel, iron, and nickel/iron alloys. The acidic solution removes oxides from aluminum and then prevents oxide re-formation for a considerable length of time. It removes protective oils from Kovar and the hydrophobic layer prevents rust formation during production delays. The cleaner also works with stainless steel, even though it does not leave a water-repelling surface on this material.

To sum up, I believe that changing the fundamental concept of how we clean metal substrates is going to allow extreme abridgement of the multiple processes used by circuit manufacturers and the chemical milling industry. These processes have served industry well for many years, and no one will argue that fact. But now it is time for a change; for new technology to overtake the old.