Two things here cry out for an investigation:
I don't recall when I last heard about a fire on a fab. Safety conditions in China fans may be worth looking into
Also now that the industry is building mega fabs that handle things like "ten percent of the world's DRAM supply" an industry/govt group should explore safeguards for this class of dependency
Not sure if the semiconductor fabs would spend money for installing so called "Safety Systems" for protection against such hazards, but at the minimum it should have had a basic fire protection system...isn't? Anyways, why the price has to go up...do the other companies cannot make up for this gap in Hynix production? Or they just might want to grab this opportunity to hike the price for DRAM which was expecting a downturn several months back?
Semiconductor manufacturing requires the use of some of the most toxic, flammable, corrosive, and simply dangerous gasses and chemicals known. Semiconductor fabs are so hazardous, they are classified all by themselves in International Building Code...and they are the only facility type to have this distinction.
Google Silane, Disilane, DCS, ClF3, WF6, BCl3, etc. for some properties of the more dangerous ones.
Manufacturing requires strict design and engineering controls including many sophisticated safety systems. Cutting corners with safety systems simply don't make sense because they are well known, widespread, inexpensive, and the risk of not having them is too great (life, property, output, etc). I have been in many many fabs in US and Asia, and in my experience safety systems are one universal constant among them.
The weakness in any fab will always be people. Even with training, people make mistakes and do things with significant consequences. I would bet human error is root cause of this fire, and key prevention measure is training/controlling the people.
@m00nshine: Thank you for sharing the knowledge. I have not visited a single fab yet and my comment was based on the assumptions that I did not think that the probability of occurrences of hazards in a semiconductor fab might be as severe as the same in the industries such as Oil refineries/off-shore platforms, where installing SIL2/3 rated safety systems is a regulatory requirement. You are correct, cost of lives (and properties) in case of an accident is much much more than cost of installing "costly" safety systems. I am interested to know which are the typical applications for the safety systems installed in fabs? for protection against toxic gas leakage & fire? Are there any standards to follow?
>> The weakness in any fab will always be people. Even with training, people make mistakes and do things with significant consequences
Fire cause problem in any situation and human beings can be blamed. Few years ago, a U.S. marine caused fire in a submarine that they need nearly a $1B to fix. The guy goes to jail while the Navy loses a warship. The problem with humans is that no one treats its well until it shows that it is the most important element in any business. You build a fab of $2B, you pay the worker min wage!
Capacity was already strained in DRAM. Apple said in their last CC that they expected DRAM prices to rise in the second half of the year. Both mobile DRAM and NAND are growing nicely, PC DRAM capacity has been diverted to those two segments of memory chips.
Unlike in the past, there is not any slack capacity in memory these days. Too many companies have gone out of business or have shrunk to niche players, and it costs way too much and takes too much technological expertise at the smaller process nodes that state of the art production requires to just throw up new fabs.
It looks fabs are more susceprtible to fire hazard. It has happened in India and few other places. It will interesting to know exact cause of fire at this fab. If we co-relate it to fire at other fabs, we can take preventive action in design of system. It is sad loss and hope it is resolved soon to begin production.
>> Safety conditions in China fans may be worth looking into
Should not be a problem. If they operate on market forces, they will go bankrupt. The problem is that they get free money and that is why we need to "look into" their operations. If you can afford $1B to build a fab, it makes sense you can at least mitigate fire risk
i purchased ten trays of 8 Gbyte, n 16 Gbyte ECC memory for servers. "inJune of this year" , well, to my surprise, the close of this week, i am getting offerings from server builders, a 25 percent premium on my offerings. Needless to say, what a quick reaction to a fire, this industry capitulates to.
Silane Gas fires/furnace explosions were so common that Japan banned the gas for many years. Hydrogen gas was not explosive usually but it did burn, often on purpose to make water vapor in oxidation tubes.
From 1960 to 2000 I can remember four small fires personally, and one HCl gas leak at National Semiconductor in new fab, that destroyed most of the control systems in all the process tools, as a regulator blew and an entire cylinder of HCl gas filled a new closed clean room. Those that saw it first, said it was like a white cloud along the floor (heavy, but trapped in the room overnight, eating away the systems). Anyone from National remember that story??
One TEOS fire in furnace was due to a technician's installation of the TEOS bubbler with the tubing hooked up backwards, so instead of argon bubbling through the liquid TEOS sending small vapor into the hot tube to grow oxide on wafers, the LIQUID was suddently pushed into the tube, blowing off the end-cap, and melting tubing, with the flames shooting from a waving tube while a tech tried to put out the fire (instead of waiting a few more seconds for the TEOS to run out) with an ancient fire extinguisher.....which dumped sodium all over the diffusion area (1962 as I remember, at Westinghouse fab in Elkridge MD.) Managers watched in horror through the clean room windows.
Many other fires were common in early epitaxial deposition systems where hydrogen was used. Famous diffusion gas story from transistor days in Connecticut: Arsenic diffusion furnace was outdoors, under tin roof. At end of furnace, manager notice a pile of dead flies on ground. The rest of the story has many versions.
From the days when we built most of our own equipment, at least the gas systems added to various furnaces, and the early days of plasma etching with nasty gases and vacuum pump corrosion issues, and crazy attemps at trapping the nasty gas byproducts, many maintenance techs were hospitalized. Bell Labs finally put out a paper on the very dangerous byproducts of plasma reactions. Very long story also about photoresist chemicals impacting pregnant women, and I personally remember washing my hands in tricholoethylene to get the negative resist off...before it was banned. We were a dangerous bunch for a long time.
What are the engineering and design challenges in creating successful IoT devices? These devices are usually small, resource-constrained electronics designed to sense, collect, send, and/or interpret data. Some of the devices need to be smart enough to act upon data in real time, 24/7. Are the design challenges the same as with embedded systems, but with a little developer- and IT-skills added in? What do engineers need to know? Rick Merritt talks with two experts about the tools and best options for designing IoT devices in 2016. Specifically the guests will discuss sensors, security, and lessons from IoT deployments.