GE’s Molten Salt Battery Failure
Author’s Note: if you want to watch the video first, it is below
The idea for this video came via a viewer recommendation. Yes, I actually do the requests that people suggest to me. I do! It just takes an ungodly amount of time.
I was surprised that the GE battery was a direct descendant of the ZEBRA battery that I mentioned in my CATL sodium ion battery video. That was pretty cool, I have to say.
This video did terribly at the start and at the time I promised to swear off any more battery videos. But I am a terrible addict, and this video seems to have legged out pretty well. I will once more work on another battery video soon.
In 2011, then-President Barack Obama visited a General Electric or GE facility in the town of Schenectady, New York. There, he mostly discussed wind turbine exports. But he also briefly mentioned an "advanced battery" business with great promise.
Obama was referring to a molten salt stationary battery technology branded as Durathon. GE CEO Jeff Immelt believed that it will become a billion dollar business.
But Durathon fell far short. In 2015, the company closed its battery manufacturing factory in New York after investing nearly $200 million. Nearly a hundred people lost their jobs.
In this video, we are going to look at General Electric's failed molten salt battery business venture.
Beginnings
Molten salt batteries are an interesting technology. They only work at very high temperatures. Ford seriously explored the space back in the late 60s with their sodium sulfur batteries. The company invested considerable resources into the venture before abandoning it for safety issues.
In the late 1970s, South African scientists working for the mining giant Anglo American pioneered the ZEBRA battery. The ZEBRA used a different chemistry from Ford:
A sodium-metal halide battery, sometimes also referred to as sodium–nickel–chloride.
After first patenting the concept in 1978, Anglo American then founded a company called Beta R&D in 1982 in the United Kingdom to commercialize the concept.
ZEBRA batteries are made using nickel powder, plain salt, and a special beta-alumina solid electrolyte. Despite the name, Beta-alumina is a sodium polyaluminate - more a salt than actual aluminium.
ZEBRA in the EV
Beta R&D at first targeted the battery for use in electric vehicles. They partnered with automaker Daimler Benz to do this and made tangible progress.
For instance, the Mercedes buses used in the 1992 Barcelona games - the ones with the Dream Team - ran on ZEBRA batteries.
In the end, however, the technology hit a double dead end at least in the EV space. First, they were expensive. Plain salt is far more plentiful than lithium, but the need for nickel powder added a great deal of cost.
Second, a ZEBRA-powered car would not be all that competitive. The battery's energy density by the late 2000s was about 94-148 watt hours per kilogram.
In comparison, good lithium ion batteries today can offer 250 watt hours per kilogram. Other more specialized chemistries can do even higher, up to 300 and even 400.
In the end automakers saw the EV space as an unprofitable niche, lacking the electric infrastructure for widespread adoption. GM for instance ditched its billion dollar EV1 car in the mid 1990s - around the same time. That basically would be the case until Tesla got big.
The ZEBRA battery's casing and heat requirements made it unfeasible for much else. Would you want a molten salt Walkman? But Beta R&D continued studying and developing the technology.
NATO eventually chose to use it for their rescue submarine, which is cool because I like submarines. Then GE came along.
The Evolution Hybrid
In 2007, GE purchased Beta R&D and integrated it into their prototype diesel-electric train - named the Evolution Hybrid. GE Transportation, now a division of Wabtec, had strong market share in the North American freight train market.
New carbon emissions regulations asked GE to improve the carbon footprint of their locomotives. Rather than achieving these thresholds with infrastructure, they opted to design and build this brand new locomotive.
The company reviewed several battery technologies for this - including a nickel metal hyride technology and lithium ion technology. But in the end, they chose to use the Beta R&D molten salt technology.
The 4,000-plus horsepower train paired a four stroke diesel engine with several batteries that captured and stored the energy generated from recuperative braking. This would help the train be 20% more efficient than its peers.
I do not believe the hybrid locomotive ever developed into a viable business. In 2011, then-CEO Jeff Immelt cited the hybrid locomotive in an interview with the Financial Post as a positive example of GE being innovative by developing products before the market was ready for them.
In this interview, he said that they sold over 4,000 of these hybrid locomotives. But I can't find any documentation for this assertion being true.
It seems strange that the CEO of a publicly listed company would lie so blatantly. But I can't find any press releases of GE ever delivering vehicles in large volume to customers nor can I track down a significant number of these hybrid trains currently running around the world.
A Harvard Business School case study noted that in 2006 the battery - presumably the molten salt technology - was not ready for prime time:
> It became clear that the battery technology at the core of its design was not yet able to support the proposed customer benefits or to achieve them at a cost that would make the project economical
This makes a lot of sense. The ZEBRA battery this battery descends from falls far short of competing lithium ion batteries in terms of energy capacity.
And 16 years later, a former engineer for GE Transportation at the time confirms:
> From a practical standpoint, the battery wasn't up to the task for energy and reliability
Today, it is worthy to note that advances in lithium ion EV battery technology eventually changed the equation. And now the industry is starting to buy into the economic cost-benefits surrounding a hybrid locomotive.
GE Transportation is even getting back into the mix, using lithium ion batteries.
Anyway going back to 2011, the GE hybrid train never left the station. But GE management decided that the molten salt battery technology still had a warm future ahead and rebranded it as Durathon. A stationary energy storage technology capable of providing backup power to critical facilities.
Technology Overview
I think before we move forward it is worth pausing to briefly discuss how Durathon actually works. This reading is based on recent reviews of technical papers and grant applications.
The first thing about this sodium metal halide technology is that it works at very high temperatures - about 270 to 350 degrees Celsius. This is because salt is conductive for sodium ions only in a liquid format - which it achieves at 154 degrees Celsius.
This makes molten salt batteries ideal for extreme conditions, and I read proposals to use them on the surface of Venus. Which is hot.
Back on the surface of Earth, however, this requires some extra work to maintain the ideal temperature throughout usage. The Durathon system consists of the battery cell pack, enclosure, and a battery management interface.
The cell pack is insulated with a ceramic fiber thermal board and heated with an attached heating device. Mica insulators and blocks are also placed between and on top of the cells. This tends to add a lot of weight, part of the reason why they weren't feasible for EVs.
When the battery is in use, the battery management interfaces activates a heater attached to the system to heat the cells to 280 degrees Celsius (536 degrees Fahrenheit). It then monitors the cells as they charge or discharge, activating a blower to keep temperatures at under 320 degrees Celsius.
GE knew from the very start that Durathon's battery chemistry did not have the same power capacity as lithium ion. So they focused on safety and durability. The system has a benign failure mode - as in it doesn't explode - and lasts up to 20 years or 5 times as long as lithium ion.
Thusly, Durathon was positioned as being complementary to lithium. As Immelt said in his remarks:
> We really believe that lithium provides power, sodium provides storage and the combination of these two technologies will find great fits ... this gives GE a real portfolio advantage in technology and innovation. We like that.
Rebrand
When GE management was pitched the Durathon system idea, analysts said that this market presented a $6 billion opportunity by 2020. Their arguments moved Jeff Immelt to sign off on a new $100 million manufacturing facility to build these at scale.
In a written account, he purportedly said at the time:
Build it big ... it's like when I was in plastics. The orders will come.
I suspect that this is why the early messaging around the New York facility was so muddled.
In the 2009 announcement, GE signaled that the factory's first batteries would go into the aforementioned Evolution hybrid locomotive.
You know. The one that supposedly had sold 4,000 units.
After the locomotive market was fulfilled, the facility's batteries would service the mining industry, telecoms, utilities, heavy service vehicles, backup storage, and load leveling.
This eclectic mix feels more like second-day fried rice to me: Everything that happens to be in the fridge.
Which is delicious by the way, but also chaotic. And therein lies the problem.
The Stationary Energy Problem
Back in 2009, Immelt said that GE expected to generate about $500 million in battery sales by 2015. And it seemed like as if the effort got off to a good start.
In September 2012, GE announced that they had secured about $63 million in new orders for Durathon. This includes 10 new telecom customer orders in 7 countries.
They profiled one of these customers - Adrian Group in Nairobi, Kenya. The company maintains cell phone towers for Kenya's largest telecom - Safaricom.
Cell phone towers had been one of the big market opportunities for Durathon. But GE found the telecom market difficult to enter. Existing suppliers locked in their customers for very long periods. Even if customers wanted to switch to Durathon, they needed to wait until those prior contracts ended.
GE then discovered that the energy storage systems or ESS space is actually a very broad category. And very fragmented. Users in the residential, commercial/industrial, off-grid, or utility fields all have different needs and requirements for their energy storage systems.
For instance, off-grid systems want to have uninterrupted power and backups from their stationary energy storage systems.
Utilities on the other hand want to be able to integrate their growing but unpredictable renewables portfolio.
Defense projects on the other other hand want peak-shaving, reliability for energy security reasons, and cost reduction as well.
Residential owners on the other other other hand want cost reduction and also arbitrage - the ability to earn money by contributing energy back into the grid.
All of these situations and use cases require their own system integration. Even if the raw battery technology is capable of meeting all of these technical promises - not a given - working them into existing systems that were not originally designed for them is challenging.
You need to customize the existing system - which means lots of specialized knowledge as well as finding a great system integrator partner to be on the ground. Or building a new system from the ground up. Both options are expensive.
Failure
In the end, I wouldn't say that the ESS market actually was a single $6 billion market like as GE saw it.
Rather, it was a jumbled fried rice of several dozen markets. Each loosely connected by a common but vague need.
There is no scale here. It's the business equivalent of house-to-house fighting. You shouldn't be building a $100 million factory for that.
All the while, competing lithium ion battery technologies and their associated costs were rapidly improving. These improvements were driven by companies in China and South Korea responding to market demands in the massive consumer electronics and electric vehicle markets. Molten salt's technical upsides over lithium weren't so prominent anymore.
GE invested millions of dollars on the technical issues of scaling battery manufacture. But they did this before realizing that the market was not how they imagined. After four years in the market, GE called it and cut their losses.
Conclusion
It has long been known that the number one reason why startups fail is premature scaling. GE was a hundred billion dollar company. They should have known better.
But to me, the root cause of the GE Durathon debacle was the company's desire for big home runs. Each year, the company had to deliver revenue growth and it desperately grasped for ways to achieve it.
In 2009, GE generated $154 billion in revenue. Growing that even 5% means finding $7.7 billion in new revenue in a single year. That's the size of the world's 1,000th largest company.
This crave for growth led the company to lay down the track before they proved that the train can move. It serves as a warning for any company or country looking to scale big into advanced manufacturing. For GE, the company is now splitting into three. Hopefully this gives them more breathing room to choose their future projects more carefully.