bit-tech.net

Lasers prove key to terabit-per-second storage

Lasers prove key to terabit-per-second storage

Lasers could prove key to boosting the performance of the humble magnetic hard drive.

Researchers have published a paper detailing a way of using laser heating to significantly speed up the data transfer capabilities of the humble hard drive.

In a paper published in the Nature: Communications journal, a team of researchers led by Thomas Ostler from the University of York detail a method of dramatically increasing data storage speeds of magnetic hard drives using a laser-based heating system.

In the paper, a system is detailed whereby a laser briefly heats the magnetic medium to a temperature of around 800 degrees Celsius for a brief period using a sub-picosecond laser pulse. This heating, the team found, significantly speeds up the process of reversing the magnetic polarity of a particular bit compared to the traditional methods of using another magnet or electricity.

While much of the paper is indecipherable to the layman, its conclusions are clear: it's now proven possible to alter the magnetic polarity of a medium by blasting it with a laser.

The team's work is far from a commercial reality, of course. There are questions to be answered about how the technology can be implemented in a way that won't bankrupt anyone hoping to buy a laser-based hard drive, while the experimental results don't address the issue of medium longevity.

It's a promising start to a potentially disruptive new technology for the storage space, however. According to the team's findings, a laser-based hard drive would have two main advantages over a typical magnetic drive. The first is energy efficiency: while high power, the laser pulses for such a tiny amount of time the overall power draw is much less than a magnetic drive would require for the same operation.

It's the performance that is of the most interest, however. A laser-based storage system could write at speeds in the terabytes per second region, making the fastest solid-state drive available today seem positively pedestrian in comparison.

Sadly, Ostler and his team are silent on when - or, indeed, if - the technology will be making ti to the desktop.

Fancy strapping some frickin' lasers to your storage system, or is this just another pie-in-the-sky idea from the boffins? Share your thoughts over in the forums.

21 Comments

Discuss in the forums Reply
kosch 8th February 2012, 13:56 Quote
I'll add this to the list of other technologies that may or may not appear in the next 30 years :)
specofdust 8th February 2012, 14:23 Quote
This sort of research is really important, but I've seen lots of stuff like this over the last decade promising super massive disks, or super fast disks, and it rarely pans out.

Off the top of my head, some problems:

Hard disks have platters made from aluminium, or from glass. I don't know what kind of glass. Rapidly heating aluminium is going to cause you all sorts of metal fatigue problems in the long run. It also produces massive problems within the system with regards to heat dissipation. If you think that you've got maybe 4 platters, so 30 square micrometres of metal being heated to 800C during all read/writes (fairly constantly).

Now if you're looking at glass, can the structure handle that sort of heat? I've made glass at 750C, so if HDD's don't use a high temperature enough varient, then you're looking at melting. Even without melting, there'll be serious problems maintaining the amorphous structure I suspect.

Finally, and perhaps more solvable but I'm not sure it would be, is the legality of high power lasers. I don't know the required power for the lasers involved because I can't be bothered to access the journal article, but I suspect that anything which can heat up metal to 800C in a picosecond is either a class 3B (hazardous to eyes) or class 4 (hazardous to everything) laser. No way are governments going to be happy letting every joe public have four of those in his computer.

Still, cool piece of research, and really interesting that polarity switches can be sped up via the use of heat in this manner.

Thing is:
Quote:
Originally Posted by article
it's now proven possible to alter the magnetic polarity of a medium by blasting it with a laser.

From memory, this may well be due to an expansion of the structure of the metal or glass due to the increased heat, allowing for electron flow through the network in a manner similar to ferroelectric effects or something. Again though, expansion and contraction of a crystal or metallic structure are, in the long term, going to be very very bad for integrity.
Teelzebub 8th February 2012, 15:05 Quote
Quote:
Originally Posted by specofdust
I've made glass at 750C,

I was wondering if you could clarify the above, what type of glass did you make at 750c and how did you make it?
Bauul 8th February 2012, 15:16 Quote
Quote:
Originally Posted by specofdust
. It also produces massive problems within the system with regards to heat dissipation. If you think that you've got maybe 4 platters, so 30 square micrometres of metal being heated to 800C during all read/writes (fairly constantly).

Now if you're looking at glass, can the structure handle that sort of heat? I've made glass at 750C, so if HDD's don't use a high temperature enough varient, then you're looking at melting. Even without melting, there'll be serious problems maintaining the amorphous structure I suspect.

Simple, use blue lasers instead of red ones, as blue lasers are cooler. Simples.
xrain 8th February 2012, 15:32 Quote
Quote:
Originally Posted by specofdust
This sort of research is really important, but I've seen lots of stuff like this over the last decade promising super massive disks, or super fast disks, and it rarely pans out.

Off the top of my head, some problems:

Hard disks have platters made from aluminium, or from glass. I don't know what kind of glass. Rapidly heating aluminium is going to cause you all sorts of metal fatigue problems in the long run. It also produces massive problems within the system with regards to heat dissipation. If you think that you've got maybe 4 platters, so 30 square micrometres of metal being heated to 800C during all read/writes (fairly constantly).

Now if you're looking at glass, can the structure handle that sort of heat? I've made glass at 750C, so if HDD's don't use a high temperature enough varient, then you're looking at melting. Even without melting, there'll be serious problems maintaining the amorphous structure I suspect.

Finally, and perhaps more solvable but I'm not sure it would be, is the legality of high power lasers. I don't know the required power for the lasers involved because I can't be bothered to access the journal article, but I suspect that anything which can heat up metal to 800C in a picosecond is either a class 3B (hazardous to eyes) or class 4 (hazardous to everything) laser. No way are governments going to be happy letting every joe public have four of those in his computer.

Still, cool piece of research, and really interesting that polarity switches can be sped up via the use of heat in this manner.

Thing is:



From memory, this may well be due to an expansion of the structure of the metal or glass due to the increased heat, allowing for electron flow through the network in a manner similar to ferroelectric effects or something. Again though, expansion and contraction of a crystal or metallic structure are, in the long term, going to be very very bad for integrity.

I don't think thermo-cycling will be all that much of a problem, if the glass sub-strait is really unsuitable for the laser-hardrive I don't think it would be all that much of an issue to replace it with a different material, or dope the glass differently. I don't see that many reasons why they cant substitute the glass for a ceramic instead. The aluminum shouldn't be at that high risk of metal fatigue since generally the layer of aluminum on a platter is extremely thin, it should keep the conduction to the adjacent material to a minimum, with your only major risk is the chance of the aluminum becoming de-laminated from the substrate. It's really that picosecond heating time that is key to making this technology practical.

Even then the microscopic area that is going to be heated should mitigate most of the thermal issues. In the case of extreme disk usage I imagine that with proper material selection, and maybe some moderate redesigns for better thermal dissipation wouldn't be that major of a challenge.

The short duration and high focus of the laser required for this technique should keep it out of the Class IV laser classification. It shouldn't take more than a few mW at an extremely high focus to perform these tasks no problem. I'm not sure what kind of laser was used in the experiment, since they seem to want to charge me $40 to view the article.


There are certainly challenges facing this technology, but they all seem like pretty standard sized challenges you would face in implementing any new technology of this caliber.
specofdust 8th February 2012, 17:11 Quote
Quote:
Originally Posted by Teelzebub
I was wondering if you could clarify the above, what type of glass did you make at 750c and how did you make it?

Sodium phosphate, and a few different acetate glasses. Can't remember precisely which were in the furnace at 750C, since this was about 5 months ago now. How did we make? IIRC just took the raw materials, crushed them up as much as possible, added the dopants we were adding to the various samples, melted down using heat of around 400 or 500C, then once they were melty enough, stuck them in ceramic crucibles before sticking them into the furnaces at the various temperatures the furnaces were at. After various times (20 minutes to over an hour, variously) we took them out, stamped the glass into discs, and then experimented on them.

You'll have to excuse the lack of specificity, it's a good while since I did this and the report I wrote on it is fairly scant on details of actual production, and mostly just filled with theory. You can make glass from about 400C onward though, but as far as I'm aware there's rarely need to go much above 2300 K.
Quote:
Originally Posted by Bauul
Simple, use blue lasers instead of red ones, as blue lasers are cooler. Simples.

I'm not talking about the heat of the laser, I'm talking about the heat of the disc. No matter what colour laser you use, if you have to heat the disc to 800C, you have to heat it to 800C.
Quote:
Originally Posted by xrain
I don't see that many reasons why they cant substitute the glass for a ceramic instead.

I haven't researched it at all, but you'd need a ceramic capable of the distinctly magnetisable sectors in a tiny area. I imagine that's possible, although I can't think off the top of my head what materials you'd use.
Quote:
The aluminum shouldn't be at that high risk of metal fatigue since generally the layer of aluminum on a platter is extremely thin

Heat stress is surely going to inevitably cause gradual deformations in, what I assume must be, a precisely ordered metal structure though? You must have a good deal of expansion of the metal when rapidly heating from STP to 800C, and I'm thinking that the level of stress can't be good for the structure of the metal over time. Obviously that's just speculation though.
Quote:
Even then the microscopic area that is going to be heated should mitigate most of the thermal issues.

Hard disks are fairly constantly in use though. You must be looking at the entire platter being heated to 800C in the course of a matter of a few minutes.
Quote:
I'm not sure what kind of laser was used in the experiment, since they seem to want to charge me $40 to view the article.

I'll have a look, I should have journal access to nature.
Quote:
There are certainly challenges facing this technology, but they all seem like pretty standard sized challenges you would face in implementing any new technology of this caliber.

I'm inclined to agree, I just think that instead of, as the press normally does, making out like this is 90% of the way there, we should recognise that this is 10% of the way there, with 90% of the work still to be taken care of by chemists and engineers.

Out of interest, what field of STEM are you in?
borandi 8th February 2012, 17:32 Quote
Quote:
Originally Posted by xrain
I'm not sure what kind of laser was used in the experiment, since they seem to want to charge me $40 to view the article.

That's a standard journal fee if you're not at an institution to a general subscription. Not dictated at all by the researchers. You might even find a copy on their website for free if you're lucky.

But we're heating up a small amount of aluminium with a sub-picosecond pulse. So what's the specific heat capacity of aluminium, the volume that's being heated, then the time, and you can calculate the power.

SPH of Al = 900 J / Kg K-1
Volume of bit = 10nm x 10nm x 10nm (probably wrong) = 1e-24 m3
Density of Al = 2700 kg m-3
Mass of Al heated = 2.7e-21 kg
Energy Needed to heat that mass up 1K = 2.43e-18 J
Energy for 780 K = 1.8954e-15 J
If we say a 1 picosecond pulse = 1e-12 s
Laser has to be 0.0018954 Watts minimum (depends on transparency of air too)

But then we could go to http://en.wikipedia.org/wiki/Heat-assisted_magnetic_recording which describes it :

"Heat-assisted magnetic recording (HAMR) is a technology that magnetically records data on high-stability media using laser thermal assistance to first heat the material. HAMR takes advantage of high-stability magnetic compounds such as iron platinum alloy. These materials can store single bits in a much smaller area without being limited by the same superparamagnetic effect that limits the current technology used in hard disk storage. The only catch being that they must be heated to apply the changes in magnetic orientation. HAMR was developed by Fujitsu in 2006 so that it could achieve one terabit per square inch densities.[1]"

"Outlook
HAMR could increase the limit of magnetic recording by more than a factor of 100. This could result in storage capacities as great as 50 terabits per square inch.
Seagate believes it can produce 300 terabit (37.5 terabyte) Hard disk drives using HAMR technology.[5] Some news sites erroneously reported that Seagate would launch a 300 TB HDD by 2010. Seagate responded to this news stating that 50 terabit per-square-inch density is well past the 2010 timeframe and that this may also involve a combination of Bit Patterned Media. [6]"
Krikkit 8th February 2012, 17:35 Quote
I don't think bulk heating over a short period would be too much of a problem. If you think the storage density of the drive might be conservatively approaching 250Gb/sq.in, you would need to read 4 sq.in/s to achieve their theoretical best-speed. In a 3.5" drive with two double-sided platters that's about 28 sq in of available space at full capacity.

Randomise your data placement a little to avoid sequential reads/writes and you've got up to 4s to dissipate that mW of heat. Shouldn't be too hard.
The_Crapman 8th February 2012, 17:57 Quote
Quote:
Originally Posted by specofdust




I'm not talking about the heat of the laser, I'm talking about the heat of the disc. No matter what colour laser you use, if you have to heat the disc to 800C, you have to heat it to 800C.

I can tell you're clearly passionate about this, but serious funny missed their. :?
Quote:
Originally Posted by Bauul
Simple, use blue lasers instead of red ones, as blue lasers are cooler. Simples.
:):):) kEl5RvbGdik
lilgoth89 8th February 2012, 18:04 Quote
are hdd platters in a vacum ?? if they are then heat dissapation could be a real problem

if not...and they are filled with air...then the spin of the platters would create airflow inside the drive

promicing tech... but every few months a 'revolution' in hdds are promiced...
and it usually fizzles out
The_Crapman 8th February 2012, 18:08 Quote
dam double post.

also, can anyone advise on how to embed youtube videos please! have tried searching the forums but can't find a guide.
specofdust 8th February 2012, 18:13 Quote
Quote:
Originally Posted by The_Crapman
dam double post.

also, can anyone advise on how to embed youtube videos please! have tried searching the forums but can't find a guide.

Fixed it for you. Just put the youtube video code, in this case "kEl5RvbGdik", inside of the youtube and /youtube (in square brackets, obviously).
The_Crapman 8th February 2012, 18:19 Quote
Quote:
Originally Posted by specofdust
Fixed it for you. Just put the youtube video code, in this case "kEl5RvbGdik", inside of the youtube and /youtube (in square brackets, obviously).
cheers spec! i'd tried that and it didn't work:?, although may have been that you were editing it at the same time.

thanks again.;) i'll go sort out another dodgy one i put in. lol
Fizzban 8th February 2012, 19:14 Quote
To help with the possible heat issues could you not coat the disc in some form of carbon layer before adding the magnetic layer?
specofdust 8th February 2012, 19:23 Quote
Why would that help? The specific heat capacity of graphite is even lower than solid alu. I don't get what advantage that would bring :/
Fizzban 9th February 2012, 00:03 Quote
I don't know anything about specific heat capacitys and what not. I just wondered, as Expandable graphite is used in metallurgical industries as a covering material for hot molten metal. Thermal insulation basically.
l3v1ck 9th February 2012, 00:23 Quote
You've still got to wait to the disk to spin to the required location and for the head to move across the disk, so It has no change of beating SSD's, and that to me makes this a less interesting development.
specofdust 9th February 2012, 10:05 Quote
Quote:
Originally Posted by Fizzban
I don't know anything about specific heat capacitys and what not. I just wondered, as Expandable graphite is used in metallurgical industries as a covering material for hot molten metal. Thermal insulation basically.

The whole point of this seems to be heating the metal to 800C though, since the mechanism of action is, I'm fairly certain, EM waves.
mclean007 9th February 2012, 11:03 Quote
I agree material stress may be a problem but I am sure they are working on it. I agree with borandi that the heat production isn't going to be an issue on the macro scale. Regardless of the fact it is heating a tiny area very quickly to 800C, a 2 mW (or whatever) laser is only capable of generating 2 mW of energy to be dissipated as heat. That is orders of magnitude less than what an SSD controller uses, let alone a high performance conventional drive. Of course you will still need the motors, head actuators, magnetic coils, drive controllers etc. that are present in any other HDD, but even if you have 4 platters, each reading both sides simultaneously using a 5 mW laser, you're only talking about an additional 40 mW of heat in a drive which is probably already dissipating conservatively a hundred times that.

In other words, yes there are local hotspots of 800C, but they will be so tiny that the power dissipated by them will be miniscule, similar to how a thimble full of boiling water contains much less heat energy and would dissipate its heat much more quickly and with less warming of the surrounding room than a bath full of water at 30C.
Quote:
Originally Posted by l3v1ck
You've still got to wait to the disk to spin to the required location and for the head to move across the disk, so It has no change of beating SSD's, and that to me makes this a less interesting development.

Well the idea I think is that HAMR allows tighter bit density because the magnetic coil not only flips bits faster (so can perform more write ops per second) but also because the heat is necessary for the bit to be written, the bits can be placed much more closely together. So even with the same rotational speed, you can increase the bitrate significantly. Think of it like this - if you have 100 bits in a given area, and you increase the bit density by 4, so there are now 400 bits in the same area, then you have a 20x20 array instead of 10x10 - a head scanning across that area at a fixed speed therefore reads 20 bits in the same time it would have taken to read 10 previously, so the data rate is doubled while the speed remains the same. This is why a modern 7200 rpm 2TB disk will produce much higher sequential data rates than a first gen 18GB WD Raptor spinning at 10,000 rpm.

Of course any spinning platter technology will never reach the random access times of an SSD, so SSDs will have the advantage on latency, but it is an oversimplification to say that HDD data rates are constrained solely by platter speed.
chrisc1277 10th February 2012, 14:04 Quote
Yeah! thats how we get stuff done here in york! BOOM! lol
N17 dizzi 10th February 2012, 17:40 Quote
So no point breaking out the blowtorch and laser pointers? I already ruined one drive throwing salt on it :(
Log in

You are not logged in, please login with your forum account below. If you don't already have an account please register to start contributing.



Discuss in the forums