Two-Inch Diamond Wafers Could Store a Billion Blu-Ray's Worth of Data (newatlas.com) 81
Researchers in Japan have developed a new method for making 5-cm (2-in) wafers of diamond that could be used for quantum memory. The ultra-high purity of the diamond allows it to store a staggering amount of data -- the equivalent of one billion Blu-Ray discs. New Atlas reports: [R]esearchers at Saga University and Adamant Namiki Precision Jewelery Co. in Japan have developed a new method for manufacturing ultra-high purity diamond wafers that are big enough for practical use. With this technique, the team says the resulting diamond wafers measure 5 cm across, and have such immense data density that they can theoretically store the equivalent of a billion Blu-Ray discs. One Blu-Ray can store up to 25 GB (assuming it's single-layered), which would mean this diamond wafer should be able to store a whopping 25 exabytes (EB) of data. The company calls these wafers Kenzan Diamond. The key is that these diamonds have a nitrogen concentration of under three parts per billion (ppb), making them incredibly pure. The researchers say that these are the largest wafers with that level of purity -- most others only get to 4 mm2 (0.006 in2) at most.
Achieving this requires a new manufacturing technique. Diamond wafers are made by growing the crystals on a substrate material, and that material is usually a flat surface. The problem is, the diamond can crack under the strain, degrading the quality. In the new process, the team made a relatively simple change -- the substrate surface was shaped like steps, which spreads the strain horizontally and prevents cracking. This allows them to make larger diamond wafers with higher purity. The team hopes to commercialize these diamond wafers in 2023, and in the meantime are already working towards doubling the diameter to 10 cm (4 in).
Achieving this requires a new manufacturing technique. Diamond wafers are made by growing the crystals on a substrate material, and that material is usually a flat surface. The problem is, the diamond can crack under the strain, degrading the quality. In the new process, the team made a relatively simple change -- the substrate surface was shaped like steps, which spreads the strain horizontally and prevents cracking. This allows them to make larger diamond wafers with higher purity. The team hopes to commercialize these diamond wafers in 2023, and in the meantime are already working towards doubling the diameter to 10 cm (4 in).
Needs to be a bit bigger (Score:5, Informative)
If they can get the wafer up to 4 inches in diameter, they could adapt it to normal semiconductor processes. That would be a game changer for computer chip fab, as diamond has a higher bandgap and can handle much higher temperatures safely.
As is, 2 inches is still good, but would mostly have applications for optical lenses, rather than chip fab, Yields would be too low.
Re: (Score:1)
Re:Needs to be a bit bigger (Score:4, Informative)
Synthetic diamond is relatively cheap, and can be made perfectly clear.
Re:Needs to be a bit bigger (Score:4, Informative)
40mm round diamond PCB runs me almost FIVE THOUSAND DOLLARS each, before metallization layers.
Re: (Score:2)
what is this mm talk speak carats please.
also that is kinda awesome, what is the use case.
Re: (Score:2)
Use case is putting as much LED or LASER power on one PCB as possible with minimal thermal interfacing for maximum heat transfer.
Re: (Score:2)
Given the inherent difficulty in producing a diamond, or rather to get the crystal lattice to form perfectly, that is quite cheap. And if you needed it because of its properties, you're more than willing to do it.
It's a single crystal of carbon you got there, the price merely reflects the difficulty in growing it. $5000 at the DeBeers markup gets you a carat, maybe two, as a jewel.
Re: (Score:3)
It's a LOW-quality diamond, only about 700w/mK. If I wanted optical grade, pushing nearly 1200w/mK, I'd be looking at about $15,000 per.
Re: (Score:2)
I've paid almost that much for two traditional FR-4 PCBs on QuickTurn, they were very large but still it's just some fiberglass, metal, and epoxy. The rush order is what makes it expensive, but for prototyping and evaluation that cost doesn't really matter.
Re:Needs to be a bit bigger (Score:5, Informative)
Re:Needs to be a bit bigger (Score:4, Interesting)
Industrial diamonds are not subject to the DeBeers markup. A 4" wafer is expensive because it is a niche item that is probably difficult and/or slow to grow at the specified purity. If you just want a an equivalent amount of small industrial diamonds of no particular purity or shape, that is (relatively) cheap.
Diamonds for jewelry won't necessarily come down in price just because it becomes easier to grow artificial diamonds. DeBeers has been able to command a high price because they've been more or less effectively marketing "natural" diamonds as superior and more desirable. You can already get a reasonably priced artificial diamond gemstone today for much less than a mined diamond, but many people still demand a "real" one. Hell, before artificial diamonds were a thing, cubic zirconia was available as a virtually indistinguishable substitute to the casual consumer, but it didn't replace diamonds in jewelry due mainly to marketing.
Petabyte (Score:1)
One million gigabyte is a petabyte (e.g. look here: http://codycrossanswers.com/un... [codycrossanswers.com]). So 25 GB x 1,000,000 is 25 Petabytes, 25 Exabytes
Re: (Score:2)
Billion. With a B.
Transfer rate and technology? (Score:5, Insightful)
The potential storage density is impressive, of course, but what is involved in reading and writing data to media like this? It is not just going to be a matter of tweaking existing Blu-ray drives. Also, even if data can be written and read, if filling the wafers would take months or years, there would be few practical applications. This seems like fascinating research for the moment, rather than anything of short to medium term importance.
Re: (Score:2)
It sounds like data could probably be written to it in parallel using an array of read/write heads, like maybe 1 million of them.
Re: (Score:2)
Diamond permits a wider spectrum of light, so presumably, one could use a UV laser with it, and have accuracy to much smaller structures.
I doubt you could get the multi-petabyte densities the pundits are claiming, but this would just be an iteration on traditional optical media.
Somebody needs to tell them that optical media is basically dead though, eclipsed by much more portable, robust (in terms of use case), and reusable media, like flash.
I stand behind my earlier post-- if they can get this fab process
Re: Transfer rate and technology? (Score:4, Insightful)
https://m.dpreview.com/news/69... [dpreview.com]
Flash can't survive billions of years, can't handle thousand degree temperatures and is a great deal more fragile than a lump of quartz, so optical isn't dead.
Re: (Score:1)
this density outclasses flash by several orders of magnitude. Imagine having all of Wikipedia on a single disk -- even if it's read-only, data storage with that density would be super useful
Re: (Score:2)
Imagine having all of Wikipedia on a single disk
Wikipedia is 21GB of text and 23 TB of video and images.
So you could store a billion copies of the text or a million copies of the complete encyclopedia.
Storing fewer languages saves almost no space because the videos and images are shared, and they make up 99.9% of the content.
Re: (Score:1)
I had no idea it was so small! very cool, thanks for the info
Re: (Score:2)
Somebody needs to tell them that optical media is basically dead though, eclipsed by much more portable, robust (in terms of use case), and reusable media, like flash.
You are mistaken. A flash drive dies in a flash of a moment.
Optical media is used for archiving huge amounts of data, and only for the fact of a new stone age and losing the ability to retrieve that data: it lasts for millennia (and is actually used everywhere where you have to archive indeed extremely huge amounts of data).
Re: (Score:2)
It really depends how durable they are. If they are going to last a very long time them there are plenty of archival applications where write speed isn't all that important.
Re: Transfer rate and technology? (Score:4, Interesting)
https://en.m.wikipedia.org/wik... [wikipedia.org]
https://m.dpreview.com/news/69... [dpreview.com]
These folk, using a different crystal, claim billions of years but only 1% of the storage density.
https://www.science.org/doi/10... [science.org]
Diamond is nowhere near as good on longevity.
Re: (Score:2)
Diamond is nowhere near as good on longevity.
If only because diamond is flammable. Diamonds burn in air with an ignition temperature of 900 degrees C. Quartz is already oxidized, being silicon dioxide. It does not burn in air at any temperature.
Re: (Score:3)
Diamond is nowhere near as good on longevity.
Especially these diamonds. Quantum storage in diamond uses the electron spin of dislocations in the crystal where a nitrogen or silicon atom replaces a carbon atom in the lattice.
Maintaining spin state requires cryogenic temperatures.
On the plus side, it is easy to erase your data: Just warm it up to room temperature. Done.
Don't expect to see these disks available on Amazon anytime soon.
Re: (Score:2)
Re: (Score:2)
The could sequence a lot more of your DNA with storage like this. What is it, like 4GB per human genome? 350million people in the USA alone. Thats 1.4 million Terabytes right there. What about data thats constantly discarded and has to be sampled in realtime, such as say all the internet traffic. Imagine what the NSA could do if they could actually record everything in a day and go back and analyze everything. Seems a bit scary. And before someone t
Re: (Score:2)
Now to accelerate the process for ripping 4 billion DVDs.
Is this "holographic" too? (Score:1)
Somehow all those things that "could" store lots of data never manage to get in customers' hands.
Re: (Score:2)
That's because you always demand your two-inch, high-purity diamond for $10.
Re: Is this "holographic" too? (Score:3)
https://m.dpreview.com/news/69... [dpreview.com]
I'd go for the quartz storage. Less fragile.
Why are we measuring in Blu-Rays? (Score:2)
I don't think I even used one of those things. The cost per byte never seemed reasonable.
Is there a significance to stating that this new toy "could be used for quantum memory"? Is it somehow unsuitable for use as ordinary memory? Does it function more like RAM or as slow long-term storage, or only as some kind of storage that is both dead and alive until observed?
Re: (Score:2)
I also puzzled over that terminology. I eventually concluded that this was probably a clumsy way of trying to imply a quantum increase in memory density, but I could be wrong.
Re: Why are we measuring in Blu-Rays? (Score:1)
Re: (Score:2)
Information storage should measured in Libraries of Congress.
Does that mean that my existing 32 LC storage media continue to grow in capacity as books are added to the Library of Congress catalogue? Cool!
Re: (Score:3)
Well,
the word quantum as in "quantum leap" is abused greatly in our days anyway.
A "quantum leap" is the smallest thinkable leap. Aka: not a big one.
Re: (Score:2)
Diamonds can be used to store information as the electron spin state of dislocations in the crystal caused by either nitrogen or silicon replacing a carbon atom.
Since TFA says this wafer has extremely low nitrogen, I presume they plan to use silicon doping for storage.
Disclaimer: IANAQP* and have no idea how one would read or write such a device.
*IANAQP = I am not a quantum physicist.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
I think it means it can be used to store quantum states without having to "measure" them, which isn't useful for conventional tasks like storing movies but may have applications in quantum computing algorithms.
That's a LOT of pr0n! (Score:2)
Re: That's a LOT of pr0n! (Score:2)
The retention for diamond is extremely poor, apparently, although there are crystalline storage systems with an estimated lifetime in the billions of years.
Re: (Score:2)
... there are crystalline storage systems with an estimated lifetime in the billions of years.
And the availability of devices to retrieve the data on those billion-year media is probably measured in hundreds of years at best. At some point there's a Catch-22 situation: "We don't have the tech to recover this data, but we can replicate it. We have the engineering drawings, schematics, and protocols stored on the... oh, wait!
Re: Hardware availability (Score:2)
We have schemes to deal with this already, assuming there is a way to mark the medium which is visible to the naked eye. You start with small but recognizable printed instructions around the outside of the disc explaining how to construct something like a microfiche reader, which is then used to read the next part consisting of microscopic print explaining lasers, binary encoding, and basic optical drives, etc. The process continues in a spiral or concentric circles toward the center of the disc, each secti
Re: That's a LOT of pr0n! (Score:2)
Why Don't we see 200 Peta byte storage already? (Score:2)
So TFA touts the ability to make 50cm circular wafers which can hold 25 exabytes. But it also says that there is an existing ability to create 4mm chips of pure diamond already. If I do the math their large wafers are about 125 times larger than the 4mm chips which implies to me that they should be able to create storage that are 125 times smaller.
If I divide 25 exabytes by 125 I get 200 petabytes. That is a pretty big flash drive, and a much more impressive story if someone could actually make one. Me
Re: Why Don't we see 200 Peta byte storage already (Score:2)
Well, a 5N-pure Si-28 wafer scale flash chip is certainly doable today. It wouldn't have that kind of capacity, but it would still be impressive. But not cheap.
Re: (Score:2)
Quibble;
So TFA touts the ability to make 50cm circular wafers ...
I think that should be 50mm not 50cm. And 50mm is 150 times the area of 4mm, not 125 (assuming I did the math right).
But yeah, if I could pack over 12 petabytes per square mm, I wouldn't feel the need to make the wafer bigger than 4mm (or even 0.1mm -- 100 terabytes would still be an awesome amount of storage.)
Re: (Score:2)
Strictly speaking every kind of storage in sub milimeter range is "quantum".
Unless you want to nitpick that a few of them need quite a bunch of atoms.
However we already have spintronics and other stuff where a single electron is holding the information etc.
"a Billion Blu-Ray's Worth of Data" (Score:2)
Re: (Score:2)
C'mon, we're adults here, you can just use terabytes
Blu ray isn't a single capacity. Regular blurays store 25GB per layer, so a dual layer disc holds up to 50GB. 4K blu ray discs are available with 50GB, 66GB, and 100GB capacities. Given the range of 25-100GB per blu ray, I think we need a concrete measurement.
How many blu-ray's to a football field? (Score:3)
Least they could do is provide a universally accepted standard measure.
Re: How many blu-ray's to a football field? (Score:2)
Elebenty Libraries of Congress.
Re: (Score:2)
Re: (Score:2)
Let's see...
A BR disc measures 120 mm diameter x 1.2 mm thick, for a volume of 13571.68 cubic mm (we'll pretend the spindle hole isn't present)...
An standard Olympic-size swimming pool has a volume of about 2500 cubic meters...
Assuming you stack them optimally i.e. in a triangle pattern, and fit discs in sideways along the walls to fill space where they can't be stacked normally, you could probably get down to around 15% air inside the volume of the pool, so about 156.58 million BR discs should fit..
The max
Re: (Score:2)
What kind of football field?
NFL https://www.dimensions.com/element/american-football-field [dimensions.com]
CFL https://cflnewshub.com/cfl-news/cfl-vs-nfl-rules-what-are-the-differences/ [cflnewshub.com]
FIFA https://sportblurb.com/fifa-soccer-field-dimension/ [sportblurb.com]
It works with emeralds? (Score:2)
The Storage "Singularity"? (Score:2)
I can't be arsed to do the maths, but if you could store every possible permutation of (say) 4KB blocks on this thing (or an array of them), then you could achieve some pretty impressive data compression. To store a file, you no longer have to store 4K blocks of actual file content, instead you just store a list of pointers to the 4K blocks you've already stored (and those blocks will be read-only). You can store the pointer lists on your regular SSD just like normal.
As cool as this could be, I suspect the
Re: (Score:2)
Go back and study your number theory. The size of each pointer in your scenario would be .... 4k. No compression is possible without loss of information or access to information in some part of the system. For most schemes, that means either (a) a certain segment of the potentially representable data is just not representable, or (b) there is some sort of many-to-one mapping from indexes to data. The two stipulations are not fully independent.
If you have a list of all possible 4k blocks, that means you a
Re: (Score:2)
Yeah, okay, fair enough - good example, btw.
We can improve on your idea though (Score:2)
We can, however, take your idea and improve on it a bit.
Suppose that we just PRETEND to have stored every possible 4k block. We just decide ahead of time:
Block 0: 0000 ... 0000 0000 ... 0000 0001 ... 0000 0010
Block 1: 0000
Block 2: 0000
No need to actually write that to disk. Just if the "compressed" copy says to use block 2, we compute block 2 right there. That will significantly reduce the amount of storage space needed.
So to use the 250th block, store the index into the imaginary array, 250. You can compu
Re: (Score:2)
Re: (Score:2)
Actually finding something "popular" to compare it to?
Makes a lot of sense to ordinary people.
With the same argument as yours: I have no quick way to reverse calculate a 10E26 number into an amount od blue ray disks.
But knowing that a billion blue ray disks is more than fits in the entire building where I'm living in: is a no brainer.
Re: (Score:2)
Re: (Score:2)
We are nerds. We understand what a terrabyte is. A Blu-ray says nothing.
If blue ray says nothing to you: you are not a nerd.
It is btw. roughly a terra byte.
for real? (Score:2)
"i'm going to have to buy the white album again"
https://www.youtube.com/watch?... [youtube.com]
Check back in again later (Score:2)
After you’ve got a product working beyond prototype stage
What does the drive cost though? (Score:2)
How many football fields worth .. (Score:1)
Libraries of Congress (Score:2)
But the real question is: how many Libraries of Congress can the diamond wafers hold? That's the only unit of information people can truly understand.