In our last newsletter, we described a few of the interesting projects we have done for the microelectronics industries in the Pacific Northwest.  In this issue, we continue with a few more examples.

In this project, a high-tech client called with a problem; one of his stainless steel, water-cooled, vacuum furnace tanks used in silicon crystal manufacture had developed a leak.  Because he had dozens of similar units, which were essential to his process, he was very interested in knowing what had caused the leak and, more importantly, whether similar problems could be expected with his other units.

Extent Of Problem

Inspection of the tank quickly revealed that in addition to the known leak, the tank was riddled with cracks.  Furthermore, a radiographic inspection of several other tanks revealed cracks in them also.

The decision was quickly made to forgo any repair attempts on the leaking furnace tank, and instead, sacrifice it for a detailed study to establish the root cause of the problem.

Stress Corrosion Cracking

By removing several coupons and examining them under the microscope, we soon learned that the cracks were initiating on the internal cooling waterside surface.  The cracks were all associated with welds, and in many cases, these were just small tack welds rather than the major structural welds.

Metallographic examination quickly identified the cracking as transgranular stress-corrosion.  (See figures below.)  This cracking mechanism occurs in susceptible materials as a result of the combined effects of tensile stresses and an aggressive chemical environment.

Cause of Problem

In this case, the tensile stresses were the result of residual stresses from welding.  But the unit was only exposed to water, so what could be the source of the “aggressive chemical environment”?

Well, it turns out that the tank was made of a 300-Series stainless steel, and while this material does an excellent job of resisting general corrosion, it is quite susceptible to stress corrosion cracking in the presence of chlorides.  In fact, under just the right conditions, cracking can be initiated from chloride concentrations as low as a few parts per million (ppm).

Cooling Water

The client reported that at one time, the cooling water for the furnace had consisted of untreated tap water.  Could this be the source of the chlorides?  Yes—the natural concentration of chlorides in untreated water in this part of the country is on the order of 6 ppm.

Even so, the attack was somewhat of a surprise, as this type of stainless steel is oftentimes used in similar water environments, usually without a problem.  However, in this instance there were just enough other contributory factors to push the material beyond its range of acceptable service.

Contributory Factors

First of all, the untreated water contained, in addition to the chlorides, a minor amount of silt.  The silt collected at the tack weld sites, which acted as small crevices; this led to anaerobic corrosion under the deposits, which created local sites for the initiation of the stress corrosion cracking.

Although the operating stress was quite low, the residual stress from welding was just enough to push the material beyond the threshold for stress corrosion cracking.  Finally, this whole process was accelerated by an elevated operating temperature, which served to increase the rate of attack.

The interesting thing about this whole project was that had any one of the contributory factors not been present, the unit probably would have run indefinitely without a problem.

This next project also involved a corrosion-related problem with a vacuum furnace, but in this case, the problem was caused by an apparently inadvertent substitution of materials, in combination with the use of unusually pure cooling water.

The project began with an emergency call to one of our engineers at home late one Friday evening.  One of our high-tech clients had just discovered that the cooling water system for their silicon crystal growing furnaces was contaminated with a brown sediment.  Although the cooling water was not in direct contact with their end product, the presence of an unidentified contaminant anywhere in the system was cause for serious alarm, given the extreme sensitivity of the process to foreign mater.

Could we be on site at 7:00 the next morning to help them find the source of contamination and eliminate it before it affected other parts of their system?

Like all our projects, the place to start was with a review of what the client already knew about the problem:  What materials were present in the cooling system?  What was the source of cooling water? What type of water treatment was used?  How did they find the problem?  Was it new?  Had they changed anything in the system or process recently?

Source of Contamination

In this case, the problem turned out to be quite simple, but it had an interesting twist.  After taking a series of water and deposit samples from the system, we came back to the lab and analyzed them; they were primarily iron oxides and iron hydroxides, with trace amounts of other impurities.  The composition and morphology of the contaminants indicated they were corrosion products of steel (see figure at left), but from what component?

All the internal components of the system were supposed to be composed of 300-Series stainless steel, which should have been nearly immune to corrosion.  So what could be corroding? 

Through inspection and sampling, we found out that not quite everything was 300-Series stainless steel.  In this case, the coils in the cooling tower were carbon steel tubing.  Sure, the tubing was galvanized, undoubtedly to protect it from corrosion, but the galvanized coating was on the outside surface for external corrosion protection, and the inside surface, which was in contact with the cooling water, was unprotected.  The tubing was clearly the source of the corrosion products in the system (see figure at right).

An Interesting Twist

 A review of the client's cooling water system for the furnace revealed an interesting contributory factor in the corrosion.  Because of concerns about potential cross contamination between the cooling water and other components of the system, the designer had specified the use of ultra pure water, with no corrosion inhibitors.  While this would have been acceptable if all the components had been stainless steel, it actually accelerated the corrosion when the cooling tower coils somehow ended up as carbon steel.  How so?

Well, the corrosion rate of carbon steel in water can vary by several orders of magnitude, depending on what's in the water.  And while everyone knows that salt water is much more corrosive than fresh water, what about the fresh water itself—how does it vary?

It turns out that our nice, clean, sparkling-clear water from the Bull Run watershed has a higher corrosion rate than that horrid-tasting, mineral-laden water that is found in some parts of the country.  Why, you ask?  All those minerals in the well water produce quite an off-taste, but they form a scale that protects against corrosion.

And as our Bull Run water is purified even further by removing the trace amounts of minerals, its corrosivity goes up (that is, unless the dissolved oxygen is also removed).

So, instead of protecting the system from contamination, the high purity water actually accelerated the corrosion, and helped cause the contamination.

 The solution was quite simple: Replace the carbon steel tubing with stainless steel.

One of our clients had been experiencing a high rejection rate during electrical tests on completed circuit boards.  Prior to calling us, the client had already determined that neither the boards nor the components were faulty; instead, they suspected that the boards were not being inserted far enough into the automated test machine to consistently make full electrical contact.

The most obvious solution would have been to simply readjust the stroke of the test machine so the boards would be inserted further.  However, before resorting to this, the client needed a better understanding of exactly what was happening.

The first thing we did was to observe the actual board insertion process deep inside the test machine using a small borescope.  Sure enough, the terminals were just barely engaging; sometimes they made electrical contact, and sometimes they did not.  But it wasn't because the stroke was too short per se; instead,  it was because the boards themselves were deflecting.

Again, the obvious solution would have been to simply increase the stroke to compensate for the board deflection.  However, doing so would cause the boards to deflect even further, and the client was concerned that the resulting increase in bending stresses might damage some of the boards' sensitive components and connections.

To evaluate this possibility, we installed electrical resistance strain gages at several critical locations on a circuit board, then ran a series of insertion tests at different stroke settings, all the while recording the board strain and deflection.  With the results of these tests, the client was able to adjust the stroke of the test machine just enough to allow full engagement of the contacts without excessively straining the board.

Another type of project where we have assisted local electronics firms and/or their suppliers is with independent, third-party verifications of design criteria.  In one project, we instrumented a clean room with several dozen thermocouples.

We then ran a 24-hour test, during which we monitored the room temperatures at 15-minute intervals with a computerized data logger to determine if the temperatures were being controlled within specification.

In another project, we ran a series of load deflection tests on modular ceiling panels to determine if they complied with the strength and deflection criteria specified by the end user.

High-Tech Photo Gallery Surface of solder-tinned circuit board trace Overheating of solder resulted in the formation of tin-rich mounds where lead was lost preferentially.

MEI-C and General Chemical of Anacortes, Washington are sponsoring David Blindheim, a junior/senior at Lincoln High School, in a cooperative research venture this summer. Dr. Andrew Held at MEI-C and Michael Sternberg at General Chemical will be directing the project, and David will be conducting the research.

The project will be administered through Saturday Academy's Apprenticeships in Science and Engineering (ASE) Program, which coordinates summer symposia and other networking and educational activities for over a hundred students.

The team will be investigating the properties of cesium vanadium sulfosilicate catalysts.  The catalysts are used to produce sulfuric acid from sulfur dioxide.

Ross Sundberg, engineering technician, passed the Fundamentals of Engineering exam, or Engineer-in-Training (EIT), as it is more commonly called.  Ross has been taking engineering classes at Portland State University while working full time at MEI-C.

Dr. Andrew Held, chief chemist, has completed 20 years of service at MEI-C.  Don Cushing, chemist, has completed 10 years.