Engineers use heat to create 100-million-cycle flash memory

By on December 3, 2012, 6:00 PM

Engineers at Macronix, one of the world's largest flash memory makers, have reportedly developed a new technology that can extend the life of NAND cells by more than a thousand times by simply applying heat.

Most of today's flash modules fail after cells have been programmed and erased about 10,000 times -- or even less, in the case of many budget-minded storage products such as the 3,000 p/e Kingston HyperX 3K.

Although that type of longevity is sufficient for consumer devices that are replaced regularly anyway, it's hardly ideal for enterprise situations where solid-state drives may see heavy usage around the clock for many years. 

While companies have used various software schemes such as wear-leveling to help improve the rate of a flash cell's deterioration, Macronix's engineers have produced a compelling hardware-based solution.

When a cell dies, it's generally due to the degradation of oxide layers around its floating gate (see the image below). Years ago, engineers found that they could restore this oxide layer by applying heat -- specifically, by baking flash chips in an oven at 250C (482F). Obviously, this wouldn't be practical for end users, so researchers had to create some form of internal mechanism, and that's just what Macronix has done.

The company has produced chips with integrated heat plates that "heal" worn cells by zapping them at 800C (1,472F) for a few milliseconds. With that occasional jolt of heat, researchers say that flash cells can handle more than 100 million write/erase cycles -- and by more than 100 million, they haven't seen any signs of death at that point, noting that a billion cycles could be possible, but it'll take several months to test.

In addition to boosting the longevity of flash cells, the heat also enables faster erasing, which Macronix believes may bring "thermally assisted" drives that tout better performance. It's worth noting that although the heat results in greater power consumption, it shouldn't reduce battery life because it can be configured to run when a device is plugged in. There's no word on when the technology may hit consumer electronics.




User Comments: 14

Got something to say? Post a comment
1 person liked this | ikesmasher said:

This is a huge breakthrough. The wearing out of SSDs is often exagerated, But this brings it to the point where the consumer will never eeven have to think about wearing out a SSD. This + Price decrease=future.

1 person liked this | p51d007 said:

I can't see "the industry" really getting behind this from the standpoint if the flash cells last longer and don't "wear out" or stop working, then they won't sell as many?

ikesmasher said:

I can't see "the industry" really getting behind this from the standpoint if the flash cells last longer and don't "wear out" or stop working, then they won't sell as many?

from another standpoint, OEMs might become interested in putting these in the everyday cheap desktop PC, then the industry would hop behind it with big dollar signs in their eyes, wouldnt you think?

1 person liked this |
Staff
Matthew Matthew, TechSpot Staff, said:

@p51d007 In my opinion, that assumes a large portion of current SSD sales stem from people replacing drives with exhausted p/e cycles, and I don't think that's the case, so what is there to lose? There'll always be some new innovation to attract buyers, be it speed, capacity or just some random feature like the one in the article (imagine how many new enterprise customers this tech would draw in). Besides, I bet if flash companies were *that* concerned about this tech reducing unit shipments, they'd just offset the loss with a sizable premium on the drives that have a zillion p/e cycles. Also, if even one company adopted the tech, I assume the rest would have to follow suit to remain relevant, especially as it became more affordable and mainstream.

Guest said:

So I guess in the end they will be hotter than HDDs, will consume more power and - hell - you'll probably fry egs on them which is not that bad. I already fry egs on my computer ... I will be able to fry more in the same time. Great!

avoidz avoidz said:

If only the same could be done for batteries.

VitalyT VitalyT said:

Hallo, tech-support? I tried to increase SSD longevity by heating it up in a microwave, and it suddenly stopped working... I think I want my money back...

Guest said:

"Hallo, tech-support? I tried to increase SSD longevity by heating it up in a microwave, and it suddenly stopped working... I think I want my money back..."

Answer from tech-support: I'm sorry, sir. You must heating up your SSD in an oven or stove, not in a microwave. And, I'm truly sorry you can't have your money back. Please buy another SSD and good luck with the new one.. :D *LoL

Scshadow said:

So I guess in the end they will be hotter than HDDs, will consume more power and - hell - you'll probably fry egs on them which is not that bad. I already fry egs on my computer ... I will be able to fry more in the same time. Great!

Uh... thats a bit of a stretch. It only heats for a few milliseconds and only has to happen on occasion to revive the dead cells. And the oxide layer is so small, I doubt you'd even be able to feel whether or not your drive is in a healing cycle even if you were touching the drive.

9Nails, TechSpot Paladin, said:

"Years ago, engineers found that they could restore this oxide layer by applying heat -- specifically, by baking flash chips in an oven at 250C (482F)."

That's well in the range of my oven. But also within the temp range of melting some solder joints. I haven't had any SDD drives fail, but is this a potential solution to repair a failed drive? What would be required to prepare a hard drive for a baking? Other than removing the plastic parts, anything else?

What frequency defines the "occasional jolt of heat" to keep it healthy? This is sounding like the new Defrag of the SDD world. =0

Staff
Per Hansson Per Hansson, TS Server Guru, said:

"Years ago, engineers found that they could restore this oxide layer by applying heat -- specifically, by baking flash chips in an oven at 250C (482F)."

That's well in the range of my oven. But also within the temp range of melting some solder joints. I haven't had any SDD drives fail, but is this a potential solution to repair a failed drive? What would be required to prepare a hard drive for a baking? Other than removing the plastic parts, anything else?

What frequency defines the "occasional jolt of heat" to keep it healthy? This is sounding like the new Defrag of the SDD world. =0

Yea, I have the same thoughts but also keep in mind that for this to be viable to test we need to wear out an SSD.

And I'm not even close to wearing out any of mine, perhaps in a decades time or so, which also helps to put this "problem" into perspective I think

hellokitty[hk] hellokitty[hk], I'm a TechSpot Evangelist, said:

So, who's going to be the first to take off the casing of their ssd and give it a pass through the oven?

If I had a dead SSD on hand I'd go for it, though I think I'd use a hair dryer.

misor misor said:

So iJesus is already here?

Guest said:

Somewhat obscure news to those of you don't follow computer history. BUT, this moves SSD into living as long as regular hard drives. It will transform the normal user computing experience. Since it's flash, this makes possible to drag your complete system (in a phone size device) along with you to any hardware standard PC. Or to put it another way, your phone size device can plug into standardized full size systems and you can continue working with all the full size device display capability because BOOT can be close to instant.

Load all comments...

Add New Comment

TechSpot Members
Login or sign up for free,
it takes about 30 seconds.
You may also...
Get complete access to the TechSpot community. Join thousands of technology enthusiasts that contribute and share knowledge in our forum. Get a private inbox, upload your own photo gallery and more.