On Whiskers and Dendrites
Many early electronic components had a short lifetime. The cause was identified and fixed, but then failures returned when the use of lead in electronics was banned under the EU RoHS directive, threatening the reliability of all electronic systems. What happened to the forecast apocalypse?
Image Credit: ESA, Wikipedia
Metallic hair growth
As long ago as 1951 a research paper was written describing a rather weird phenomenon: the tendency for certain smooth metallic surfaces to develop a growth of… hair. These hairs or whiskers are a fraction the width of a human hair but can grow to many millimetres in length. The whiskers are not a corrosion product, they are pure metal with a sharp point capable of piercing surface coatings aimed at stopping their development. To this day nobody is really sure why they appear on metals like Tin and Lithium, but there does seem to be a connection with mechanical and/or thermal stresses on the surface at the time of manufacture. It may have remained just an odd but inconsequential effect were it not for the extensive use of Tin plating on electrical circuit boards to prevent copper corrosion. Connector pins are similarly coated for the same reason, although gold is frequently used as a better, but more expensive alternative. It’s not difficult to see the harm caused by these electrically conductive filaments when, after an interval from months to years they eventually form a short circuit. Here is a video showing whisker growth around the pins of a DIL packaged chip:
The next example on an SMT capacitor reveals the intricate structure of a single filament:
Finally, a standard commercial D-Sub connector. Notice that the whiskers are so fine they are invisible to the human eye unless lit at precisely the correct angle:
The last video shows that even an unused component will develop whiskers. Clearly a flow of electric current is not a factor.
A temporary cure
Many of the early transistor devices of the 1950s had lamentably short lives, particularly those sealed in a tin-plated metal screening can with its own ‘grounding’ lead. The device failed with an internal short circuit, but with no external sign of what had gone wrong. The short appeared to be between the screen and one of the other three pins. A ‘cure’ was to cut the screen wire, but that might have caused increased noise in a sensitive circuit. When the device was opened up, whiskers were found sprouting from the inside of the can. Once one had grown long enough to touch the ‘chip’, the device ceased to function.
It had been known for some time that by adding a tiny portion of Lead to the Tin used for the plating (and solder), whisker growth could be halted or at least dramatically reduced. This was the widely adopted solution and remained so for the next 40-odd years. But it was a temporary fix because Lead was eventually declared a hazardous chemical and banned from use in electronics. Cue prophecies of doom and disaster like those concerning the effect of the ‘Millennium Bug’ on all computer real-time clocks in the year 2000. Nothing happened and the argument still continues as to whether the ‘bug’ was real and fixed by prompt action, or never existed in the first place. Lead-free components seemed set to cause a similar disaster, this time with hardware. Today, opinion seems divided on the impact of whiskers in the lead-free age.
Lithium-metal batteries have a whisker problem: like Tin, Lithium metal is prone to developing growths on its surface. In this case, though, tree-like structures called dendrites form in the electrolyte of the battery. The sharp tips can breach separators eventually causing short-circuits resulting in overheating and probably a fire. Researchers have recently found that the precise composition of the electrolyte has a direct impact on dendrite growth. Lithium-metal has a much higher energy density than Lithium-Ion and if its ‘whisker problem’ can be solved, a new generation of high-capacity batteries will become available.
A hidden problem?
Since electronics went lead-free, the forecast fall in reliability doesn’t seem to have transpired – at least as far as the ordinary consumer is concerned. Does this mean that an effective alternative to Lead was found and a whisker crisis avoided? Admittedly I have not carried out an exhaustive search of the literature, but I’ve not come across any reference to a new Tin formulation that works as well as Sn-Pb. Research is still continuing to find a whisker-resistant conformal coating. The list of RoHS-exempt product categories makes for interesting reading though. The list is extensive and it’s easier to describe what is not exempted: consumer electronics. Anything where reliability is a top priority, medical, space, military, industrial, safety systems, is exempt. Mobile phones, computers, TVs, domestic appliances, etc. must comply with RoHS restrictions. Probably a coincidence, but audio equipment I purchased in 1975 still works perfectly. Old computers from the 1980’s such as a ZX81 and a Jupiter Ace both still function. Two of the very first pocket calculators from the early 1970s, yep, they work too. On the other hand, a succession of recent smartphones and a tablet have developed faults after only a few years. My four-year-old digital SLR camera now seems to be completely dead. Maybe the more densely integrated circuits succumbed to bombardment by cosmic particles. Who knows? Modern consumers may accept the short life of their ‘tech’, but this attitude might change when the waste of global resources becomes apparent. You can see why NASA and the military don’t want to take chances by looking at this 2002 catalogue of expensive failures thought to be due to Tin whisker growth.
Here’s one for hackers of vintage equipment: ‘repairing’ an old AF117 transistor that may have failed due to an internal tin whisker. I should say the following video belongs in the ‘don't-try-this-at-home’ category, but anyone with a pressing need to rescue an AF117 knows what they’re doing. Probably.
Barbara Horváth “Examination of tin whisker growth in electronic assemblies”, 2012 Ph.D. thesis. An in-depth study of metallic whisker growth.
If you're stuck for something to do, follow my posts on Twitter. I link to interesting articles on new electronics and related technologies, retweeting posts I spot about robots, space exploration, and other issues.
CommentsAdd a comment
Nice article, and sure to get others‘ attention (maybe not like a charged capacitor, lol).
While the eraser-trick was successful on gold-plated connectors from the middle of the 80s onward, I frequently used air and toothbrush-scrubbed all kinds of discrete- and ASIC-/FPGA-electronics boards (and onboard mounted connectors) looking for "bridges" during maintenance and fault-finding. While most of the solder-joints looked good, and heat and cold pointed the way to some faulty areas, some re-soldering proved beneficial. Lastly, in my opinion, levitating hot solder-tips very near to virtually all components rejuvenated the boards (at least got rid of the 'immovable dust'). I often wondered where the ever-so-tiny elongated silvery "wires“ infrequently sticking out from underneath ASICs/FPGAs and recessed near pins of all kinds of components came from.
I am thinking that vast amounts of electronic-boards tested on Marconis suffered from whiskers and dendrites.