SoylentNews is people
SoylentNews
Researchers at Delft University of Technology have developed a sensor that is only 11 atoms in size. The sensor is capable of capturing magnetic waves and consists of an antenna, a readout capability, a reset button and a memory unit. The researchers hope to use their atomic sensor to learn more about the behaviour of magnetic waves, so that hopefully such waves can one day be used in green ICT applications.
In theory, engineers can make electronic data processing much more efficient by switching to spintronics. Instead of using electrical signals, this technology makes use of magnetic signals to transmit data. Unfortunately, magnetism tends to get incredibly complicated, especially at the tiny scale of computer chips. A magnetic wave can be viewed as millions of compass needles performing a complex collective dance. Not only do the waves propagate extremely quickly, causing them to vanish in mere nanoseconds, the tricky laws of quantum mechanics also allow them to travel in multiple directions at the same time. This makes them even more elusive.
Journal Reference:
Elbertse, R.J.G., Coffey, D., Gobeil, J., et al. Remote detection and recording of atomic-scale spin dynamics, (DOI: 10.5281/zenodo.3759448)
The sensor is intended to help make progress with spintronics.
(Score: 4, Informative) by Wobzter on Wednesday May 27 2020, @12:45PM (7 children)
Hi! One of the authors here.
What we are able to sense is a "magnon", also known as a "magnetic wave" that gets transported by atoms.
In particular we use magnetic atoms (Iron) on a surface where they are partially decoupled from the unlying surface, such that they behave mostly (though certainly not exactly!) as free atoms.
If you have any questions, let me know.
The full paper is here: https://www.nature.com/articles/s42005-020-0361-z [nature.com]
(Score: 0) by Anonymous Coward on Wednesday May 27 2020, @01:23PM (1 child)
So the sensor is just 11 (13, 15, 17) iron atoms appropriately spaced in a Cu2N lattice?
I mean, bravo at getting the number of atoms into the teens, but if you aren't counting the lattice (nor the system to keep the whole thing at 1.5K), it seems a bit... hyperbolic.
Still impressive, I'd love to understand it better.
(Score: 2, Informative) by Wobzter on Wednesday May 27 2020, @01:46PM
Correct. The sensor was 11 atoms, and we attached a chain of 3, 5 and 9 atoms to it. So the largest structure would be 20 atoms.
A different research group has managed to about a hundred of them on the same surface in a very closely packed fashion: https://science.sciencemag.org/content/335/6065/196 [sciencemag.org]
So it definitely possible to scale up. The goal of this experiment was moreso for the science: how do these magnetic waves move? That's why we built a sensor.
Once we (science) understands how these magnetic waves operate better, it will be that much easier to design structures that are much easier to scale up: well beyond a hundred atoms.
I understand where you're coming from with calling it a bit... hyperbolic. However, a mouse trap (which is a kind of sensor) requires more than just itself to operate as well: it needs a surface to stand on and it needs to the right temperature, gravity and other variables (fortunately for us "ambient pressure/temperature" usually suffices). With that in mind, if we can consider the mouse trap in isolation, one could argue in favour of considering these 11 atoms in isolation as well. But I definitely see your point. Unfortunately, funding agencies like to stories that are easy to understand for the public without too many "if"s and "but"s - so we are encouraged to present it as such to the public. But I do stand by the title of the article.
If you have any more questions, I'm happy to help!
(Score: 2) by Phoenix666 on Wednesday May 27 2020, @02:07PM (3 children)
Thanks for the link. The memory aspect of the sensor is meant to accommodate the travel time of the STM to provide both the excitation source at the input lead and then read the result at the output lead. Would the sensor serve in its current form to record excitation from a non-STM source, say, via discrete exposure to exogenous excitation?
Also, if the sensor is sensitive up to 9 atoms away, presumably you'd be able to set up an array of them. How many could you read with one STM in a cycle?
Washington DC delenda est.
(Score: 1) by Wobzter on Wednesday May 27 2020, @02:49PM (2 children)
Hi there Phoenix.
The sensor would serve to record a spin excitation from any source - provided this excitation travels through the input lead. To get such an excitation into the input lead, however, is quite difficult through means other than the STM tip. A laser pulse would be too large and most likely excite (i.e. trigger) the sensor as well*. We have also been looking into using the sensor as the input lead for the next sensor (i.e. spintronics). We have not gone further in this direction as of yet (there is only so much time in a day, and other experiments were waiting as well).
* I mentioned that the laser pulse would trigger the sensor as well. You could argue that this itself is a form of recording. However, it would trigger to a different kind of source (EM waves, rather than M waves). This would also be possible through the bits described in the science paper I quoted earlier. However, I again want to point out the issue of locality.
Regarding your last question: the STM tip can easily move 100 um in a short time (1-100 seconds, depending on how safe you want to travel at 300pm height). We wanted to set up a kind of array where we have a long line of, say, 100 atoms in a row. Every 10 atoms there would be a perpendicular line of atoms branching off to one of these sensors; so 10 sensors in total. It would be interesting to see how far a single pulse is able to trigger these sensors. Unfortunately, theory predicted (and we seem to corroborate that, though I'm hoping to get the chance to give experimental proof) that these waves on this particular surface would last only 10ps. With a travel speed of about 50m/s (surprisingly slow, actually! Who knew!), it would not able to go much beyond 9 atoms. So such an array would look great on paper... but in practice the results will probably disappoint.
(Score: 2) by Phoenix666 on Wednesday May 27 2020, @03:22PM (1 child)
Thank you, Wobzter, very interesting.
Is the placement of the sensors strictly limited by the substrate, ie. such that radial configurations aren't possible? Are you able to play with the substrates like, say, etching a circuit board?
Washington DC delenda est.
(Score: 2, Informative) by Wobzter on Wednesday May 27 2020, @03:41PM
Correct. This substrate (CuN on Cu3Au(100)) has a square lattice. We could've also done this on a hexagon lattice (Cu3Au(111)), though the coupling strengths between the different components (interchain and intrachain) would be very different, as well as the ease of building such structures.
In a world with infinite money it would be possible build an algorithm that builds these structures for us to scale things up. For now things are done manually at the atomic level - etching of circuit boards is still too big for that (1-2 orders of magnitude, according to the latest etching techniques). Even then, the exact spacing of the atoms is difficult to do with etching techniques, but very crucial.
I hope it helps!
Cheers, Wobzter (i.e. RJG Elbertse)
(Score: 2) by Bot on Thursday May 28 2020, @09:20AM
>What we are able to sense is a "magnon"
kinky!
Account abandoned.