The Mystery Glue-Forming Plastic Case Failure
We’ve all heard of or (hopefully not) experienced irritating mystery device failures before: A brand new iPhone explodes and causes some minor burns; a week old Nexus screen suddenly develops a hairline crack that the phone company won’t replace; an SD card that was working perfectly fine one day and has never been ejected, is suddenly unreadable etc. These are maddening to a consumer that spent some hard earned cash looking forward to a new toy, but if you are an engineer tasked with figuring out the root cause of a rare, hard to reproduce failure, you may need to get on some blood pressure meds.
As frustrating as the examples above are, I believe the mother of all head scratchers was a uniform disintegration of the plastic enclosure assembly across 30 devices I was once tasked with investigating. Let’s start from the beginning. The device in question was an automated medical device with outstanding battery life that is suitable for transport use and mass casualty events. For this reason, a customer purchased 30 of the devices to store in their climate controlled medical storage facility. However, to secure the sale, the customer requested an ability to periodically charge the devices within their storage facility as well as have an auxiliary external battery. Now, on the face of it, this does not seem like an unreasonable or particularly complicated request, so the production team made some adjustments to accommodate these options, without carrying out a risk analysis. As you can imagine, this wasn’t a good decision. About a year later reports came in that all 30 devices had developed numerous cracks and fractures along the plastic case parts that form the enclosure assemblies of the devices. Even more mysterious, some strange adhesive substance had formed on the surfaces of these plastic enclosures. The customer insisted that they had never used the equipment, aside from the periodic charging they had requested. This was so puzzling because no other customer had experienced such issues, especially given the other customers’ storage and operating temperatures ranged from 120°F in the Middle East to -5°F in Alaska. The plastic enclosure assembly parts were injection molded from impact-resistant ABS/Polycarbonate blend resin with excellent mold flow qualities. The parts themselves had some intricate features, but the design process had taken great care to ensure standard material wall thickness strength and uniform cooling rates would be maintained during the manufacturing process. So this was not an obvious design issue. When I finally got to assess the damage, I was astounded at the level of the physical destruction. It looked like someone had relieved a temper tantrum by running through the devices with a ball pin hammer, then smeared some glue all over them for good measure. It was hard to imagine that the damage was not caused by some form of high velocity trauma. However, since all the devices had their plastic enclosures fail at exactly the same points, it was clear the answer to this mystery could be narrowed down to the packaging setup of the shipped devices, and whatever was done to them during storage. The device (which has an internal 97 W-h Li-Ion battery), its power supply brick, and an auxiliary Lead Acid battery are placed inside a hermetically sealed shipping case surrounded by closed cell polyurethane foam for protection. Take a look at the mock-up illustration to the right to get an idea of the packaging setup (Note that this illustration only shows the bottom half and that there is a top foam sheet that is not shown for clarity): The devices were being charged periodically in the
hermetically sealed case as shown. The Lead Acid battery pack was being charged as well. The adhesive-like substrate was only present where the foam and the ABS/PC plastic were in contact. The cracks and fractures also tended to be in areas where the foam had some contact with the plastic as well, though this was not as clear-cut as with the adhesive substance. So What Happened? Initial theories centered mainly around the idea that either or both the Lead Acid and Li Ion were outgassing something that was reacting with the packaging foam to form the glue-like stuff that was then denigrating the plastic. However, the lead acid battery pack is supposedly sealed, although if charged to capacity it could outgas hydrogen (which, by the way, isn’t a good idea in a hermetically sealed case, particularly one with a charging electrical device). Hydrogen gas does not react with polyurethane, and there was no evidence that anything in the case could react with hydrogen to produce some glue (unless there is a chemist out there who knows otherwise). As for the Li Ion battery pack inside the medical device, it is also sealed, and contains a current limiting protection circuit. It would be a statistical nightmare if all 30 protection circuits messed up, and if they did there would probably be 30 exploding devices. The first attempt at figuring out what caused this strange failure was to ask the ABS/PC resin manufacturer whether they’d ever experienced any such behavior from the polymer before. Of course occasionally cracks could occur in injection molded parts because of residual stresses, but a situation in which such extensive plastic failure occurred, coupled with the adhesive-like substrate, was unheard of to them. Next, I contacted the Polyurethane packaging foam supplier and asked them the same thing. They had never heard of it either. In fact, one of the main characteristics of any packaging foam material is that it has to be inert, and to their knowledge, the packaging foam had never exhibited any reactive properties while in use. Next, I attempted to reproduce the problem in an accelerated manner. Now, the 30 devices that had experienced the plastic disintegration had been stored (according to the customer, but who knows?) in a “cave” maintained at 58°F – 63°F throughout the year. Remember that they noticed the issue after about a year in storage. To try to accelerate the process, I placed a packaged, charging device in the same arrangement into a thermal chamber and cycled the temperature to just below the glass transition temperature of the polymer over a weekend. I was not able to reproduce the failure. Obviously it is daunting to recreate a failure that took a year to manifest itself in just a couple of days. I realized that the best explanation of this failure would have to be a reasoned theory. I put the packaged arrangement back into the thermal chamber. I set the thermal chamber to maintain a temperature of 63°F (17°C) to simulate the approximate condition the devices were stored in according to the customer. I allowed the device to charge but monitored the temperatures using thermocouples at various points inside the hermetically sealed case. What I discovered was that the temperatures at different points inside the sealed packaging case varied greatly, and never reached equilibrium through the duration of the study. Right around the power supply was the hottest, reaching almost 70°C, fully 50°C above ambient. In positions where the foam and plastic were pressing against each other, the temperatures were roughly around 20°C – 30°C above ambient, though this varied somewhat depending on the exact position and the LIB state of charge (the device itself emits heat as well when charging, depending on the charging current). Elsewhere, the air temperature inside the sealed packaging case did not get higher than 15°C above the set temperature of 17°C. Final Theory: Charging the devices in an airtight packaging setup traps a lot of heat. This is particularly true in areas where the polyurethane foam acts as a direct insulator. Over time, if the device is continually charged, a large heat gradient develops on the plastic in areas where it is tightly insulated compared to areas where it is not as well insulated. In addition, there are surface imperfections on the plastic (e.g. weld marks) that can propagate cracks and cause fracture from expansion caused by the localized heating. If charging takes place over a period of months, the heat buildup could cause the insulated areas to heat up to the glass transition temperature of the polymer, resulting in a rubbery feel that becomes sticky. This phenomenon can be readily observed in some internal car dashboard and door panels, which are mostly molded Polystyrene (or similar hard plastic) parts. These panels can become rubbery and sticky if the car has been out in the sun for a long time on a particularly hot summer day with all the windows up, thus trapping a lot of heat. So this is what I think happened. But I could be wrong. Would have loved to be able to have the time to set up a separate, independent, controlled study to test out this theory, but time waits for no man. Anyway, moral of the story is, this packaging setup was a terrible idea from the beginning, so consult an engineer before agreeing to honor your customer’s terrible idea.