MATERIALS: Don’t Guess Your Way through Root Cause Analysis: Part 2

If the power of the analysis laboratory to illuminate the reasons for product problems is so great, why are these resources utilized so seldom and so reluctantly? In many cases there is simply a lack of awareness of how these techniques and tools can help. But in other instances the people involved in the problem-solving process have experience with laboratory analysis, and the experience is not good.

I frequently receive reports from clients who have gone to an analytical laboratory for tests on problem products. The lab reports may contain more than 100 pages of charts and graphs and a summary that attempts to explain the work that was performed. The documents are typically signed by multiple people with advanced degrees. I recently read one report where the analyst’s resumé was part of the introduction—an attempt, I suppose, to convince the reader of the quality of the work and veracity of the conclusions that were drawn. But the report provided no clear explanation of the cause of the problem being experienced by the client and no guidance on a solution. This occurred despite the fact that the company performing the tests professed to be expert at providing just that type of assistance.

This happens because the expertise of the analytical people does not translate readily to the practical world of manufacturing where the problems arise. The people who work in the lab are expert chemists and are capable of performing detailed deformulations of polymers. But the connection between what they find and how it relates to the problem the client is experiencing is often not clear to them.

Plastic parts don’t simply represent a “polymer problem.” They are the result of converting raw material to a part through a process such as injection molding, extrusion, blow molding, or thermoforming. These processes use some combination of heat and pressure to form the final part.

Some of these processes are more aggressive than others in these respects. In the case of injection molding, the temperatures and pressures may be relatively high and there may be a considerable degree of shear applied as well. Drying the material may be required prior to processing, and this step may or may not be performed properly. The molding process moves the material at high velocities, resulting in orientation of the polymer chains as well as any fillers. The rate of cooling can affect the degree of orientation that is retained in the part, and it can also influence the development of crystallinity and internal stress. These factors all play a part in determining the properties of the material.

Considerations such as the geometry of the flow path in the mold, the gate location, and the process conditions are all important in determining the performance of the molded product.

This is a language that is just as foreign to the analysts as their report is to the manufacturer. Few manufacturing companies employ anyone who can truly understand the analysis report, and very few analytical firms have anyone on their staff with a manufacturing background.

This gulf between disciplines results in some unfortunate outcomes. Expensive tests are often performed that have little chance of helping to diagnose the problem. And when a test does produce a significant finding, its importance is sometimes overlooked while other less significant data points are highlighted as critical to understanding the problem. Often the volume of data that is generated can make it difficult to distinguish between the important and the trivial.

In one study performed on a part molded in a flame-retardant nylon, the end user was experiencing brittle part behavior and discoloration. Several tests were run that had little chance of addressing the problem. However, three tests were performed that actually produced results that were sufficient to diagnose the problem. A measurement of the molecular weight of the raw material and the molded parts showed a decline of 20% during processing. 

This represents a significant change that can be expected to reduce the toughness of the material. But it was not recognized by the analyst as problematic. If it had been identified as a key part of the picture, it is possible that poor process control by the molder would have been blamed. However, the results of other tests that were performed would have been understood as clues to the cause of the polymer degradation that were not necessarily the responsibility of the processor.

Thermogravimetric analysis (TGA) tests showed that the flame retardant used in the nylon began to decompose at a temperature only 20° C (36° F) above the crystalline melting point of the polymer, leaving the molder with an extremely narrow processing window. But the TGA result was not plotted or analyzed in a manner that enabled anyone viewing the final results to realize the importance of this finding.

Differential scanning calorimetry (DSC) tests that were also run had the potential to confirm and clarify the TGA results. But these were not performed at the right conditions, and therefore another opportunity to illustrate an inherent problem with the stability of the raw material was missed. 

Perhaps of even greater importance, the test data could have potentially been used by the molder to alter the molding conditions in order to determine whether it was possible to exercise enough control over the process to rectify the problem or whether the problem with the raw-material formulation was unmanageable.

Ultimately, the analyst focused on the moisture content of the nylon in the molded parts as the root cause, neglecting the fact that while this could account for a lack of ductility, it did nothing to explain the discoloration. The problem went unsolved until someone with both processing and analytical experience reviewed the test report, repeated the DSC and TGA tests, and explained to the processor what the results meant in a format that they could use to adjust their process conditions.

Unfortunately, the first outcome is very common. The result is that a lot of money is spent on testing and the final report is either inconclusive or draws an incorrect conclusion. This sends the problem-solving team off in the wrong direction, spending more funds on “corrective actions” that do not address the problem. And when this becomes apparent, the team becomes disenchanted with the analytical process and will be very reluctant to go back to an analytical lab again to solve a future problem.

INTERDISCIPLINARY APPROACH NEEDED 
How can this situation be improved? It begins with clear communication between the analysis laboratory and the people bringing the problem to the lab. The background information provided to the analytical people needs to be accurate. If good and bad product is being submitted, these designations need to be correct. Failed product is easy to identify, but “Good” may really mean “Not Failed Yet.” This needs to be clearly understood.

The analytical people need to be skilled in eliciting this information from the client. In order to do this effectively, they need to have people in the organization who have a working knowledge of the commercial side of the industry and of processing. 

Without this interdisciplinary approach, clients will continue to spend resources ineffectively and analytical labs will continue to turn out large and expensive reports that no one understands and that ultimately do not achieve the end goal—solving the client’s problem.