QUAKER CHEMICAL LABORATORY RESULTS

Parallel studies were ongoing by McCrone Associates, Inc. on the surface analysis of the defect cans, and by the Quaker Chemical research laboratory on the analysis of the lubricant/coolant samples obtained during the cratering problem.  Fortunately, this plant was on a schedule of sending a sample for routine analysis to the Quaker Chemical laboratory every Monday.  This turned out to be most helpful since the problem started one day after the Monday sample was sent.  During the midst of the problem, a five-gallon sample of the in-use lubricant/coolant had been captured.

Extensive analysis of the lubricant/coolant samples was performed, and the most relevant data are shown in Figure 17.

click image to enlarge (71K) Figure 17

Elemental composition of the “before” sample versus the “during the problem” sample showed some significant differences:  Note the absence of iron, nickel and zinc in the “before” sample and the presence of these elements “during the problem” frame.  Something else very important to note is the ratio between the magnesium and the calcium for the “before” problem time frame.  This ratio is about 1 part calcium to 3.7 parts magnesium.  In the “problem” time frame, a significant change occurred to 1 part calcium to 1.2 parts magnesium.  This was a dramatic change in the fluid chemistry, especially in a nine-month-old lubricant/coolant that had reached a quasi-equilibrium state.

With the surprising conclusion of a corrosive mechanism from McCrone Associates, Inc. plus the contamination findings from the Quaker Chemical laboratories, a meeting was held with the production plant management to share our information.  This lasted for several hours and the plant personnel shared some information that they had discovered from their own internal investigation on changes that had occurred in their processes.  Key comments from the plant personnel included:

1. Initially, there seemed to be a pattern to the high ME (metal exposure)readings.

    

2. From the review of the plant chemical process log, it was determined that excessive make-ups on the closed loop of the cooling tower water system had occurred.

    

3. The bodymaker operators had noticed sporadic color changes in the lubricant/coolant, similar to the dye used in the cooling tower water system.

    

4. The lubricant/coolant concentration was dropping during the “problem” time frame, indicating that some excess water was getting into the system.

    

5. The heat exchanger was replaced as the lubricant/coolant was dumped and recharged.  In this plant, the lubricant/coolant flow traveled from the central lubricant/coolant system, through the heat exchanger, directly to the bodymakers.

The main conclusion from this meeting was that substantial amounts of cooling tower water were clearly contaminating the bodymaker lubricant/coolant system during the “problem” time frame.

No one knew the impact of this, but the obvious next step was to take some samples of the cooling tower water back to the laboratory and analyze its components.  Figure 18 is similar to Figure 17 with an additional column for the cooling tower water.  Note the presence of iron, zinc and nickel in the cooling tower water; also note the magnesium to calcium ratio of 1 to 6.  We also speculated that a borate corrosion inhibitor was used in the water, which would explain the presence of boron detected during the SIMS analysis.  Our conclusion was that the cooling tower water contamination was the assignable cause of this ratio change during the crater problem.

click image to enlarge (76K) Figure 18

Based on these findings, we hypothesized that the cooling tower water contamination might be the cause of the crater defect during the can making process.  Additional samples of the cooling tower water were submitted to the Quaker Chemical laboratory to see if this defect could be reproduced under laboratory conditions.