Friday, February 9th 2024

Hafnia Material Breakthrough Paves Way for Ferroelectric Computer Memory

Scientists and engineers have been experimenting with hafnium oxide over the past decade—many believe that this "elusive ferroelectric material" is best leveraged in next generation computing memory (due to its non-volatile properties), although this requires a major scientific breakthrough to get working in a practical manner. Hafnia's natural state is inherently non-ferroelectric, so it takes some effort to get it into a suitable state—a SciTechDaily article explores past efforts: "Scientists could only get hafnia to its metastable ferroelectric state when straining it as a thin, two-dimensional film of nanometer thickness." Research teams at the University of Rochester, New York and University of Tennessee, Knoxville have presented evidence of an exciting landmark development. Sobhit Singh, assistant professor at UoR's Department of Mechanical Engineering, believes that the joint effort has created a lane for the creation of bulk ferroelectric and antiferroelectric hafnia.

His "Proceedings of the National Academy of Sciences" study proposes an alternative material path: "Hafnia is a very exciting material because of its practical applications in computer technology, especially for data storage. Currently, to store data we use magnetic forms of memory that are slow, require a lot of energy to operate, and are not very efficient. Ferroelectric forms of memory are robust, ultra-fast, cheaper to produce, and more energy-efficient." Professor Janice Musfeldt's team at the University of Tennessee have managed to produce a ferroelectric form of hafnia—through an experimental high pressure process, based on Singh's exact calculations. The material remained in a metastable phase post-experiment, even in a pressure-relieved state. Musfeldt commented on the pleasing results: "This is as an excellent example of experimental-theoretical collaboration." Memory manufacturers are likely keeping an eye on Hafnia's breakthrough potential, but material costs are dampening expectations—Tom's Hardware cites shortages (going back to early 2023): "Hafnium (the key component in Hafnia) has seen a nearly fivefold price increase due to increased demand since 2021, raising its cost from about $1,000 per kilogram to about $5,000. Even at $1000 a kilogram, though, hafnium is by far more expensive than silicon, which measures in the tens of dollars per kilogram."
Sources: Sci Tech Daily, Tom's Hardware, PNAS Org, Reuters
Add your own comment

9 Comments on Hafnia Material Breakthrough Paves Way for Ferroelectric Computer Memory

#1
FreedomEclipse
~Technological Technocrat~
Price means its obviously going to be pushed into the non-commercial market first. It might be used in data servers a long time before eventually trickling down to every day joes.
Posted on Reply
#2
user556
It's odd that he compares it against tapes and HDDs rather than Flash. That probably means their results also fall way short of Flash performance for the moment.
Posted on Reply
#3
Wirko
FreedomEclipsePrice means its obviously going to be pushed into the non-commercial market first. It might be used in data servers a long time before eventually trickling down to every day joes.
But one could make many exabytes of memory from one kilogram of nanometer-thick hafnia.
Posted on Reply
#4
FreedomEclipse
~Technological Technocrat~
WirkoBut one could make many exabytes of memory from one kilogram of nanometer-thick hafnia.
Thats not how business works...

Naturally they would want to keep production at an all time low so they can mark up the price by 999%
Posted on Reply
#5
bug
I understand price is a concern, it always is. But is price per kg a good metric here? Do we need as much Hf per memory cell as we do Si? Does it matter that Hf is something like 6x denser than Si?
Posted on Reply
#6
Frick
Fishfaced Nincompoop
FreedomEclipseThats not how business works...

Naturally they would want to keep production at an all time low so they can mark up the price by 999%
Volume is where the real money is.
Posted on Reply
#7
FreedomEclipse
~Technological Technocrat~
FrickVolume is where the real money is.
Not if you saturate the market it aint. why would chip makers need to bump up the price by 15-20% and slow down production despite having no problem pumping it out at the same speed they have for the last 2-3 years?
Posted on Reply
#8
Frick
Fishfaced Nincompoop
FreedomEclipseNot if you saturate the market it aint. why would chip makers need to bump up the price by 15-20% and slow down production despite having no problem pumping it out at the same speed they have for the last 2-3 years?
If it works out the market will be huge though.
Posted on Reply
#9
bug
FreedomEclipseNot if you saturate the market it aint. why would chip makers need to bump up the price by 15-20% and slow down production despite having no problem pumping it out at the same speed they have for the last 2-3 years?
Slow down what? Did you not get the memo about everybody fighting over fab capacity, despite fab capacity being at an all-time high?
Posted on Reply
May 21st, 2024 19:19 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts