Nottingham university boffins have devised a depleted uranium molecule that keeps a constant magnetic state and is many times smaller than bits on hard disk drives, promising 1,000-fold increases in hard drive capacity - if it can be turned into a product before solid state storage takes over. The Notts team, led by Dr Steven …
unlikely to succeed
The general public will scream OMG URANIUM and run. When properly informed that the drive is not actually radioactive, and neither does it emit poisonous particles when heated, they'll politely thank for the clarification, say that now it makes a very attractive proposal, and buy something else. Privately, they'll tell friends that "you never know", someone will say that the manufacturer is probably lying, and everyone will nod sagely, proud of their intelligence.
The "It holds its magnetic state stably if kept at a low temperature, around two degrees above absolute zero" is a bit of an issue to overcome first.
Re: irrelevant really
So, like most hard disks, if it doesn't work you bung it in the freezer overnight?
No need for blue LEDs
I predict a drop-off in the sales of those wanker-modified PC cases.
This is pretty much the opposite of the article earlier in the week about HP and its progress with memristors. None of the details which would allow the tech to be commoditized are nailed down. It would require a steady supply of depleted uranium, may not scale well to room temperature, might not even be able to flip its charge, and turns all puppies that approach your computer into Cujo (you have to fall asleep while reading the technical paper to learn about that last side-effect).
The potassium in my banana will reach half-life before this unvention comes to market.
Good idea. But?
Are these drives going to be a problem if disposed of improperly?
"Just how long will it take to backup a 3 petabyte drive with one read head per platter surface?"
Who says the platters have to be standard size? 2tb miniature hard drives for mobile devices would be more feasible.
Changing the Magnetism
Changing the magnetism of the molecule is done pretty much the old-fashioned way, I would suspect: by causing the molecule to physically change its orientation.
Another FUD source
You missed the great advantage of this technology. The technophobic would love it !. Another scare campaign about toxic waste containing the great satan of chemicals being forced upon a helpless world by corporate greed ! Best thing they have heard since the IPCC. Ignore the probability that it will never go into production. Many storage technologies work at 2K, but nothing higher.
Spinning in a death spiral
Which part of MOVING-PARTS MEMORY IS BAD is the industry not getting?
_Anything_ that has to microstep a bunch of heads into the right position, then twiddle its electronic thumbs waiting for the right bit of a spinning platter to rock up under the heads, whilst the OS has to find something else useful to do in the tens or hundreds of millions of cycles until the hard disk Telexes back its reply, is still living in the computational Stone Age of drum memory, thermionic valves and punched cards.
Motorised memory lives in the milliseconds world. Processors live in the nanoseconds world. It's like harnessing Dobbin to the Porsche and taking it on the M25.
Nurse, where are my dried frog pills?
I'll tell you which part the industry isn't getting
The part where data stays uncorrupted on your drive year after year. Flash/SS Drives are notoriously unreliable for long-term data storage. My own experience with an SSD which, while blindingly fast, fucked itself after 8 months, is less than optimal so far. When we see that the technology can store data reliably for longer than the lifespan of a butterfly more people might start taking it seriously. Until then, I'll keep my decade-plus-old MP3 collection on spinning media where I know it'll still be in 10 years time, thanks.
Hmmm, what's the spin-up power surge going to be like on a uranium platter then?
There is also the anti-nuke hurdle to cross
I can already picture the word URANIUM writ large in the headlines to whip the usual suspects into a frenzy
If the backup takes too long you make physically smaller drives. Say a few terabytes tucked in a spare corner in your watch.
No, if the backup takes too long you speed up the process. Like formatting. It already takes 10+ hours to full-format a 2 TB hard drive. A 2 PB hard drive at that rate would take more than a year to full-format. That needs to be fixed. Now. And the same goes for backing up.
10 hours to format a drive?
You need a different file system.
From tiny acorns
You pour scorn, but isnt this the nature of scientific research and innovation? someone somewhere finds a new way of doing things that int particularly useful. they file a paper on it, and then another team, maybe years removed from now or maybe just weeks, either in the same country or on another continent goes "holy shit guys, we've been looking at this all wrong, we could use this instead"
I understand your point, but you sound like the guy laughing at the helicopter going "but you'd need a guy who can turn that crank 1000 times a minute with the strength of a thousand men! what a fucking stupid plan!"
The idea is fine
But overzealous predictions of success are not. It seems that any such breakthough gets spun for all it is worth, ignoring the other 99% of technology that has to appear to make usable products. No wonder all the conspiracy theorists end up thinking the government/Big Oil/THEM supress technologies.
The predictions we see here mirror those for high temperature superconductors around 1990. Low loss cables across the country (ignoring for a bit the trivial detail of building a 1000 mile refrigerated tunnel and the power required to keep it cool). Over 20 years on and high temp superconductors are still limited to a few very niche applications. Same too for a whole raft of "breakthrough" technologies from optical computing to EV batteries.
fission for facts
In the "no understanding whatever of physics" category, I wonder how long it will take for someone to complain that if you get too many of these drives together in a rack you risk creating an uncontrolled nuclear reaction? After all, everyone knows uranium always creates nuclear reactions. Admittedly, I can see someone in senior management actually ordering that this be done, because of course nuclear reactions = free electricity.
"Let's say it is 1,000 times denser areally speaking than current hard drives; what does that mean? The I/O density will be abysmally atrocious. Just how long will it take to backup a 3 petabyte drive with one read head per platter surface? We should envisage a time scale in days unless the spin speed and/or the number of read/write heads is increased. As for RAID re-builds, forget it; we're talking weeks."
How about physically smaller drives? These drives would open the door for much smaller drives to be released and adopted. You could have a 1.8" drive (or smaller) and have more capacity than todays largest 3.5" drive. That would make adoption and sure thing, they could even make a 1.8" drive 15k or faster.
Would it solved RAID issues, no, but as storage is increasing faster than performance. In a heavy IO state, larger drives are not what you are looking for.
They could make drives small enough where and still have more capacity than todays drives where one could have a mirrored RAID 5, 10 or 6 config, so if one drive fails, the second array is used and that would allow the other array time to rebuild and still be updated by the other array. Rebuild times would drop as you would have the issue of rebuilding while data is trying to be accessed.
That "boffin with a Back To The Future hair-style" would be Martyn Poliakoff. He supervised my mate's postdoctoral research and is well-respected both in the UK and internationally. Give the man some credit.
But I'll concede that he does have Back To The Future style hair.
"It holds its magnetic state stably if kept at a low temperature, around two degrees above absolute zero."
Zero Kelvin will hold any state perminently - even my last thought ....
That many icons on a desktop would make anyone's hair stand on end.
Apart from that: such cool stuff, science rocks! qB^D
element versus isotope
"created a depleted uranium molecule, one with its radio-active element removed"
An element refers to an atoms with equal number of protons. For example, Uranium atoms all have 92 protons. The 92 protons are, in fact, what makes Uranium Uranium and not, for example, Lead. SO if we remove all the radio-active element, we would end up with no Uranium at all. Some might say well we are only taking about the radio-active element, the Uranium that has 235 protons and neutrons, but that is imprecise as well. Why the majority of Uranium decays a factor of 1000 slower than the U-235, it is still radio active. In fact all elements heavier than lead tend to eventually decay to lead.
What we are talking about here is isotopes. Isotopes have the same number of protons, but different number of neutrons, that is they are the same elements with slight different properties. Depleted uranium has the the majority of the hightly radioactive isotope removed leaving the less radioactive isotopes, not elements.
"... the El Reg storage desk thinks this is bonkers..."
OK, so you're probably not going to be able to create a working hard disk by painting a platter with this stuff...
However, if a way can be found to embed an SMM in a magneto-optical substrate, where the alignment of the molecule can be changed by heating the substrate with a UV laser, applying an external magnetic field to change the alignment of the molecule as it "floats" in the melted substrate, then turn off the laser to let the substrate cool and lock the molecule into its new orientation, then you'd probably have a pretty good shot at making an ultra-high-density archival medium. (The short wavelength of UV light would probably be necessary to achieve the desired storage density; it is likely that not even blue lasers would be fine enough to manipulate domains this small.)
The key to all of this is finding -- as the article indicates -- an SMM that can maintain its magnetic moment at room temperature. Since the uranium-based molecule mentioned in the article needs to be kept at 2 Kelvin to make this process work, it seems highly unlikely that it would be capable of being used in a phase-change-based storage system, much less a GMR-based hard disk drive. On the other hand, there are a number of other materials, such as certain samarium compounds, that may work for this kind of storage. (Some samarium compounds can maintain their magnetic moments at temperatures approaching 980 Kelvin / 707 degrees Celsius.)
Of course, when it comes down to it, if you're trying to pack that many bits into that small of a space, why not find a way to shrink a nanometer-scale lithography machine down to the size of a hard drive? Who needs magnetic or optical storage, when you can insert a cartridge of raw silicon, hafnium, and germanium, and carve your own chips? :-)
and since SSDs have definite cell life then a tiny lithography that simply "carves over" when you delete, it will have a cell life too.
Wow, I mean wow...
"However, if a way can be found to embed an SMM in a magneto-optical substrate, where the alignment of the molecule can be changed by heating the substrate with a UV laser, applying an external magnetic field to change the alignment of the molecule as it "floats" in the melted substrate, "
Well, that's more or less HAMR when you've got the laser and you operate above the curie temp. If you really want to stick with MO, then you operate above the comp temp and require some other element to make it a ferrimagnet. Oh, and the substrate *never* melts in MO.
"it seems highly unlikely that it would be capable of being used in a phase-change-based storage system, much less a GMR-based hard disk drive. "
Well that's a safe assumption given that drives haven't used GMR for about 5 years, and why do you care about the reader anyway? The writer is more of the limiting factor especially at such small scales.
"On the other hand, there are a number of other materials, such as certain samarium compounds, that may work for this kind of storage. "
No, you'd use FePt with its ~50kOe Hk.
I concede that the word "melt" may be a misnomer, the way I used it above, but I was speaking in generalities.
That said, the basic premise still holds: Embed an SMM in a phase-change substrate that limits the mobility and fixes the orientation of the SMM when the area surrounding the it IS NOT being illuminated with the laser, and that allows freedom of (re-)orientation when it IS illuminated.
Thank you, also, for bringing HAMR to my attention. That one somehow slipped by me, but I've been focusing on Spintronics-based memory technologies...
I don't generally go ultra-pedant on a single paragraph fragment, but this is one on those times where it is necessary.
"The Notts team, led by Dr Steven Liddle, created a depleted uranium molecule, one with its radio-active element removed... "
This is utterly confusing writing. Firstly, people know about Uranium-238 atoms (Uranium with 238 nucleons in its nuclei), and Uranium-235 atoms (235 atoms in its nuclei). Depleted uranium is just uranium with almost all of the 235 atoms removed. I've never heard of a "depleted uranium molecule". Is that the same as a Uranium-238 atom? Then why don't you say so?
The second thing: how the hell is the "radioactive element removed"? All forms of Uranium are radioactive - 238 just has a longer half-life of 4 * 10^9 years. Unless you are talking about someone who has worked out to make the half-life time infinite... In that case, that achievement would be far, far more newsworthy than just using in a hard rive. That's Nobel Prize territory.
I'm disappointed, El Reg. All this regurgitated knowledge is _high school physics_. It might not be so easy for journos with Arts backgrounds to get their head around it. I understand. But for the love of god, can you get a second opinion somewhere before printing this confusing morass of embarrassment?
I love you, El Reg. But sometimes you need to lift your game.
Now that they can read and write DNA, all they need to do is speed it up a bit.
Of course, the entertainment industry will want to apply a levy on PCR (DNA replicators) equipment.
Your assumption is that you would make a disk 1000x bigger in capacity. Why not 1000x smaller in (physical) size? Imagine 100 CF card sized TB drives in a 1U array.
"Hot" hard disk
GIves new meaning to that term, innit?
Also, would the densities be sufficient to achieve critical mass?
Maybe they could be self powered.
Ok, I'll leave now.
Don't mean to nitpick, but...
"The Notts team, led by Dr Steven Liddle, created a depleted uranium molecule, one with its radio-active element removed..."
Considering that there is no stable isotope of uranium, depleted uranium is that in which the most active radionuclides have been removed - leaving behind uranium-238, which has a half-life of 4.5 billion or so years. So it's not *very* radioactive, but it certainly still counts as radioactive.
Still, a nice article. :)
All Uranium is radioactive
All uranium is radioactive. Depleted uranium has had U-235 removed leaving mainly U-238 which is still radioactive, but not fissile i.e. it can't be used in a bomb.
Other transuranic elements have special magnetic properties. Plutonium would make a great magnet - much stronger than NdFeB - but for some reason the technology has never caught on. Mmmmm.
>>it can't be used in a bomb
It can't be used to make simple fission bomb but you can use it in a multi-stage bomb...
Non radioactive uranium?
"depleted uranium molecule, one with its radio-active element removed"! Depleted uranium (U238) is radioactive too with a half life of ~4 billion years.
"a depleted uranium molecule, one with its radio-active element removed"
U-238 still decays with a half-life of 4 x 10⁹ y, which is around 10 times longer than U-235. Still, not a lot.
Removing radioactive element
The article originally said the radio-active element of the depleted uranium molecule had been removed, and this was based on the U of Nottingham release about Liddle's SMM, which said this: "Although it may have somewhat negative PR it seems depleted Uranium — a by-product from uranium enrichment and of no use in nuclear applications because the radioactive component has been removed ..."
I didn't check this - FAIL!
It's been pointed out that this is hogwash. A commentard said: "U-238 still decays with a half-life of 4 x 10⁹ y, which is around 10 times longer than U-235. Still, not a lot."
But still, definitely still radio-active. So the article has been corrected and I'm grateful for the education.
"created a depleted uranium molecule, one with its radio-active element removed, built from two uranium atoms with a bridging atom of toluene"
This is one of those sentences where it is hard to start listing all the mistakes. Could someone with some vague scientific background proofread these pieces sometime? All well and good to criticise "boffins obviously know diddly squat", if only the writer had a clue.
Hint, there is no such thing as a depleted uranium molecule, or a depleted uranium atom even. "one with its radio-active element removed" is almost meaningless in context. Nor is there any such thing as a toluene atom. One suspects that some of the blame might come down to ham fisted attempts by the researchers to defuse any fear about the use of a radioactive element. But it would be nice if el Reg could QC this stuff before spouting off about things they don't understand. The video was actually remarkably well done for a bunch of academics trying to explain something, and vastly better than the article.
"Unless the paper's costly full text in Nature Chemistry discusses it" so we assume that el Reg's budget doesn't run to the the $32 needed to actually read the article and find out the answer. Or maybe just email one of the researchers. Maybe this sets some sort of metric to the value placed on finding out information for publication here.
Where we're going, we don't need hard drives!
I think the "scientist" icon should be mostly pedantic grammar nazi icons.
So U238 has a half life of of 4 *billion* years
So in 4 billion years time time 1/2 of the U238 atoms will have decayed to Thorium 234.
Would any one like to calculate how many atoms of a *whole* drive would be UJ238 to begin with.?
Hint. I would expect it to be *substantially* below 1g. In fact I would doubt the mass of the layer on *all* platters would be <1g given that in 2206 mag layer thickness was 14-20nm (MRS bulletin) and Nickel's density is 8800 Kg/m^3
Note People have been *claiming* spinning disk drives were obsolete since the mid 1970's, when Magnetic Bubble Memories were 2x the density of DRAM at the time, had 0 static power and a fabrication process *significantly* simpler (which means cheaper in this highly cost sensitive area) than DRAM's of the time (a feature that if anything has *improved* given the *very* complex manufacturing cycles now employed in silicon fab processes).
It didn't happen then.
Note that as a *near* term storage technology this is < v 0.1. Either studies predict *huge* benefits if they can make it stable at room temperature (in reality up to say 80c inside the case) or this is just a lab curiosity and a reminder that transuranic chemistry has uses other than blowing stuff up, building nuclear reactors or nuclear waste processing.
Well I don't see a problem ...
After all anyone with a domestic smoke detector has a (very) small amount of Americium in their home which is quite radioactive indeed. No one makes a huge fuss about them being in a domestic setting.
In contrast 'depleted' U238 is only quite mildly active and in minute quantities embedded in epoxy resin would not even be detectable since its mainly an alpha emitter.
i gather that "ordinary" depleted uranium has about one-third the radioactive decay of regular uranium - they don't get all the U-235 out of it.
This device probably is just the same and lasts only long enough to run the experiment before radioactive decay annihilates the special molecule and/or zaps the whole computer.
As Scotty would no doubt point out
you can't have a Diuranium drive. You need Dilithium.
Why not Diuranium? Well "The engines cannae take it", natch.
Too Cold To Be Practical
Liquid Nitrogen at 77K is cheap, available in large volume, and only somewhat inconvenient - you won't use it at home, but might use it in a data center. Unfortunately, most useful superconductors can't run at this high a temperature, a mere -I96 Celsius.
Liquid Helium at 4K is expensive, and helium's in short supply, and you'd mainly use it in a laboratory or very special application.
This storage runs at 2K, which means that merely cooling it with liquid helium isn't good enough - you have to actively cool it after that. It's really really difficult. It's extremely valuable to have this kind of discovery in a lab, because it points out directions for interesting research, and maybe some of that will lead to something practical.
Never Mind The Hysteria
The government, at least in the U.S., allows people to own old lanterns with Welsbach mantles, but since Thorium can be converted to fissionable material (just add neutrons) it's tightly regulated. As the same is true of U-238, I suspect that there will be severe regulatory hurdles.
What, they can scrape the coating off their old hard drives, and make a NUCLEAR BOMB with it?
So public hysteria about the drives being unsafe may be the least of their worries...
bis(bis(N-trimethylsilyliminodiphenylphosphorano)methane uranium dodo)toluenediide
Post in reply to this comment the number of goes it took you to pronounce the chemical name of that molecule without stumbling over it.
My count is 6 tries before I was able to say it out loud without tripping up. Anyone do any better?
but then I am a chemist.
More important, where do they get the dodos from in this day and age???
While stocks "last"
But isn't DU still unstable? Ok, very very long half life, but still, statistically, someone somewhere will have an uranium atom splitting. Ok, so maybe not more likely than a head crash, but still something to think about!
Also, if your RCD trips when you're on holiday, does that mean you lost both your pizza and data for the same reason?
for a giant ROM, if you ignore the giant fridge keeping it cool, and the difficulty of setting the bits in the first place and...
What's that Lassie? You think this is a stupid bit of grant-grabbing nonsense with no practical applications?
You know - you could be right. Have a bone.