With a theoretical storage capacity more than 10 times higher than today’s best lithium-ion batteries, it’s no wonder lithium-air batteries are being touted as one of the types of batteries that could make electric cars truly mainstream.
Now, as part of a US Department of Energy program to provide large amounts of supercomputer time to advance cutting edge, real world research, IBM scientists are partnering with government scientists from both Argonne and Oak Ridge National Laboratories to model and develop the materials needed to make lithium-air batteries a reality.
The award provides up to 24 million hours of computer time on two of the world’s fastest supercomputers–which, you guessed it, are housed at the two national labs. To give you some perspective, this only represents about 1.5% of the total time available on both computers, but considering how many other scientists use these computers, the fact that one project got 1.5% of the total time is pretty amazing.
IBM has been researching lithium-air batteries for some time. In the past they’ve indicated they’re not interested in building their own batteries, but want to partner with other players. The major stumbling blocks to building a consumer grade lithium-air battery have thus far been related to safety and recharging. Lithium can be extremely flammable and using air as a reactant instead of forms of cobalt or iron greatly increases the danger for the battery to catch fire.
The researchers hope that with the huge amount of computing time made available to them, they’ll be able to develop materials and methods to deal with some of these issues.
Given that recent reports clearly show lithium-ion batteries are coming up far too short in certain areas, the world can’t get next generation batteries soon enough.
Source: EurekAlert!
Image Credit: Argonne National Laboratory
PDF 
I wonder if they considered using a distributed computing model to do this work, instead? Because, I run both Folding@Home and SETI@Home of my home machines — but if I could contribute some spare CPU cycles on a battery project, I would!
Sincerely, Neil
I wonder if they considered using a distributed computing model to do this work, instead? Because, I run both Folding@Home and SETI@Home of my home machines — but if I could contribute some spare CPU cycles on a battery project, I would!
Sincerely, Neil
I’m an engineer and this post makes absolutely no sense to me. It’s either wrong or very poorly written. 24 million hours represents 2840 years if this is 1.5% of the available time these are seriously durable machines. Do the buildings housing “super computers” generate some sort of time warp or what? Need I go on. This article requires major rework. I’m not a computer geek nor should I need to be for it too make sense.
I’m an engineer and this post makes absolutely no sense to me. It’s either wrong or very poorly written. 24 million hours represents 2840 years if this is 1.5% of the available time these are seriously durable machines. Do the buildings housing “super computers” generate some sort of time warp or what? Need I go on. This article requires major rework. I’m not a computer geek nor should I need to be for it too make sense.
IBM is making a smart play here, as it appears their approach is to develop advanced technology, obtain the patents, then strike licensing deals with the actual manufacturers in China. We need many more technology engines like this to fuel the next generation of clean energy.
IBM is making a smart play here, as it appears their approach is to develop advanced technology, obtain the patents, then strike licensing deals with the actual manufacturers in China. We need many more technology engines like this to fuel the next generation of clean energy.
Simon,
It is neither wrong nor very poorly written. If you read the press release and other associated content that is linked throughout the post you will indeed see that these are the statistics given by the researchers and institutions themselves. “24 million hours out of an available 1.6 billion.” I’m not sure why it makes no sense to you. With my limited knowledge of how they calculate computing time on a supercomputer as imbued to me by my short perusal of wikipedia, it is done by calculating processor hours. You typically have one processor in your home computer. A supercomputer has many multiple thousands. As an example: an experiment running for 24 million hours on a 5,000 processor machine would take 200 days. That same experiment running on your home computer would take about 2,700 years.
Nick, Good explanation of the time involved. It would have been better understood if the story had reported as 24 million CPU hours, rather than as computer hours, although many, if not most home computers now sold have at least dual CPU processors.
Nick, Good explanation of the time involved. It would have been better understood if the story had reported as 24 million CPU hours, rather than as computer hours, although many, if not most home computers now sold have at least dual CPU processors.