(Or How Can Anything So Simple Be So Easy To Screw Up?)
By Rudy Sedlak
The first of three articles on the Develop-Etch-Strip process line.
Most PCB Fabricators today are using negative working photoresist to produce innerlayers, usually a dryfilm type. The photoresist is typically developed using a dilute Sodium or Potassium Carbonate solution. This mildly alkaline solution reacts with the unexposed portion of the “acidic” photoresist, neutralizing its acidity, and making it water soluble. The exposed portion is cross-linked by the UV light during exposure, and rendered essentially insoluble to the developer.
Developing is one of the simpler chemical processes in the making of a PCB, yet one of the more crucial, as it defines what the circuitry will look like when the PCB is finished. Because developing is apparently so simple, it is easy to overlook its significance to the production of high quality PCB’s, and to overlook the details necessary to produce high quality developing.
The standard developer today is a 1% solution of either Sodium Carbonate Monohydrate (Na2CO3.H2O), or Potassium Carbonate (K2CO3). There is not much to choose from between the two, although the use of Potassium Carbonate has been reported to give slightly better quality developing, and have a wider process window. Potassium Carbonate can be purchased in concentrated liquid form, allowing convenient feed and bleed application, without large feed tanks.
In the days when a 10 mil trace was a “fine line”, using developer solutions to the point where the developing slowed dramatically, then dumping them, was an acceptable scenario. This was because the quality of the developing that occurred during the last half of the developer bath was not a problem. Today the “resist foot” that is left behind in such a use scheme is utterly unacceptable.
Today automated pH controlled feed and bleed developing is the norm, and the best use mode for developers. Using pH control on a developer is the only way to control the feed and bleed system that makes sense. Other control systems are insensitive to the varying amounts of photoresist on a given job, and will either waste chemistry, or give sporadic quality in developing.
When using pH control, the pH that the developer is controlled to is recommended by the photoresist vendor, and is usually around 10.5. This number is chosen only partially because it is a “good” pH at which to develop the photoresist, the other reason why a (usually higher) pH is chosen, is that the higher pH promotes greater solution turnover, which insures that the level of dissolved photoresist in the developer sump is kept low. It turns out that the level of dissolved photoresist in the developer sump profoundly affects the quality of developing.
Using a pH of 10.5 as a set point, assuming a 50% exposed area, and two-sided photoresist application, the typical 1.3 ml thick dryfilm photoresist will consume about one gallon of a 1% solution of Potassium or Sodium Carbonate made-up developer solution per square foot of laminate, This means that there is only 1.3 ml-square feet of photoresist per gallon of developer solution in the sump.
The issue of temperature and line speed need mention in passing. The photoresist vendor will specify the optimum temperature for developing, but it is usually around 85-90 F. The line speed should be set such that the breakpoint (the point where the resist apparently is just gone, exposing the Copper) is at, or less than 50% of the way through the developer chamber.
On the subject of equipment configuration, it is crucial that the spray nozzles are set up so that the sprays barely overlap. Sprays that meet too far away from the board lose energy, and will not develop as well. If the sprays do not overlap at all, those areas that they do not touch will not get the spray pressure required for good developing. Note that this spray pattern issue is also very important in the first rinse, to assist in removing the resist foot.
There has been much discussion about whether cone or fan nozzles are superior for use in developing, and further discussion will change no ones mind. There are proponents on both sides. One area that is obvious is that fan nozzles have more degrees of freedom in spray pattern adjustment (and thus more ways to be fouled up?).
What is clear is that in most PCB fabrication shops the choice of which nozzle type to use becomes a minor issue about a month after the developer module is installed, and the develper module becomes subject to the abuses that are so common during routine maintenance and cleaning. One of the best way to insure high quality developing is to make sure that the people who are responsible for dismantling the developer equipment, for any reason, clearly understand the critical importance of replacing the nozzles in the chosen configuration.
Another issue that has a powerful impact on the quality of developing is the “mundane” issue of equipment maintenance. Simply keeping the developer process equipment clean is apparently so simple, that it is too often overlooked, or given a standard cleaning using generic chemistry (Caustic Soda and/or acid). Cleaning with caustic soda, and/or acid, can knock off the big, easily removed deposits, but ignores the long term, tightly bound residues, which build up and ultimately clog nozzles, and manifolds.
There is no way that the typical generic chemistry (dilute Caustic Soda and/or acid) solutions can clean a developer process chamber thoroughly. The residues in these chambers are the reaction product of water hardness and photoresist, perhaps combined with certain defoamers. The only way that this sort of residue can be removed is by chelating the metal out of this residue. When the metal is removed from these residues by chelation, the alkalinity remaining in the cleaner easily dissolves the rest of the deposit. I have personally seen developers that had glass “observation” windows which had been opaque for years, even though the developer was cleaned out weekly with caustic soda. These windows become completely clear in less than a minute when cleaned with a chelated cleaner.
Equally worthy of discussion is the first rinse after the developer. Cleaning this chamber, and its nozzles, can be as difficult, or worse, a cleaning problem than cleaning the developer itself. Although the carbonate/photoresist levels are much lower in the first rinse than in the process sump, the water hardness is higher. Because of the higher water hardness any carbonate, or photoresist residue rinsed off the panel has lots of water hardness with which to react, which means that the deposits in the first rinse chamber are all very tenacious, and require a good chelation system to remove.
It is real important when using these type of cleaners to realize that they only need pH adjustment to be dumped, (the required legal disclaimer here) based on the laws in most areas (please check with your local authorities). The cleaners only need pH adjust because they will have no controlled heavy metals. Given this scenario, it is critical that these spent chelated cleaners be dumped directly, after pH adjust, to the waste streams going out of the plant. Never, never, ever, allow the spent cleaner, or rinses of this chamber, to go through a clarifier. (This warning is issued based on some very painful, costly, personal experience.)
The above article is meant to trigger investigation into the issues mentioned, and discussions with vendors, but is entirely too brief to be considered a thorough going review of all the variables that control the developing process. Next month, forward to the etcher!