Sabine
patreon
Science News May 3
🕑 Added 2023-05-03 15:00:04 +0000 UTC[This is a transcript with references.]
Welcome everyone to this week’s science news. Today we talk about wave-like dark matter, wires that learn, the core of planet mars, why quasars ignite, a superconducting highway, fabric that changes shape and colour, chemistry on a quantum computer, the Brazil nut effect, and of course, the telephone will ring.
A lot of you have asked me to comment on a paper that made headlines last week with the claim of having found evidence that dark matter is wave-like and it’s got something to do with space-time crinkles.
Dark matter is supposedly stuff that fills most of the universe and if it exists, which it may not, then it’s even more abundant than normal matter. Averaged over the entire universe it would have to make up about 85 percent of all matter or so, but in galaxies it’s usually even more, 90 percent and up. Yes, that’s also the case in the Milky Way – its mass is mostly dark matter which forms a big spherical halo like a huge invisible cloud around the visible disk and if it exists dark matter is passing through us as we speak.
The currently most popular hypothesis is that dark matter is some kind of particle. Astrophysical observations unfortunately tell us very little about exactly what kind of particle, they just tell us where the stuff is and how much of it there is. This is why physicists have put forward hundreds if not thousands of different hypotheses, a completely fruitless strategy that I’ve bitched about in an earlier video.
Now, all particles are also waves because quantum mechanics and all, so the headlines make little sense to me, even leaving aside the question of what a space-time crinkle is supposed to be. Is this like a French fries crinkle cut? Do you want ketchup with it?
What they say in the paper is that they test ultralight particles as the constituents of dark matter. One type of those particles are invisible axions. This is an option which has become increasingly popular among physicists in the last decade or so, partly for good reasons and partly for not-so-good reasons.
The thing is now that the lighter the particle, the larger the wavelength that’s associated with the particle. The dark matter particles could be so light indeed that the wavelength becomes comparable to the size of a galaxy. And stuff that has wave-like properties can interfere and that can create specific patterns.
The authors of the paper looked at a particular gravitational lens. In a gravitational lens, one galaxy curves spacetime, and the light from a second galaxy then bends around the first. For this particular galaxy the result of the gravitational lensing is somewhat confusing with multiple dots and has been difficult to explain. This is because in the usual dark matter models, the dark matter halo is very smooth and therefore the lensed image should also be smooth. Though personally I don’t find this particularly surprising because there’s a lot of stuff going on out there that we don’t know about, so you get a lot of weird observations.
In any case, what they now do is add random perturbations on the smooth dark matter halo, choose the parameters suitably, and show that this fits the observations better. So, no waves, just random fluctuations. That’s what they mean by “crinkles”. And no ketchup either.
A few things worth mentioning. First, they don’t actually use any “wave-like properties” of the dark matter, . That’s just an inspiration for adding random bumps on a smooth halo. Second, massive particles could also create halos with inhomogeneities or crinkles, if you wish, depends on their self-interaction about which we know very little. Third, it’s unsurprising that if you make a model more complex you can fit the data better. Fourth, one generally shouldn’t jump to conclusions from a sample size of one.
In summary, I’d say the results do not support the conclusion. Albert also thinks this is a little sketchy.
An international team of researchers has developed wires that learn like the human brain.
They took a bunch of tiny silver nanowires and splattered them on glass. This created a network with about 700 nodes. They sent currents through the network, which they used to ask questions and extract answers. Then they train the nanowire network by providing feedback signals that either increased or decreased the conductance of the junctions, depending on whether the network's output is correct or incorrect.
This is very much like learning works in the human brain, in which connections between neurons strengthen or weaken. The wire network could also remember things at the same time as they were processing new information, also very similar to the human brain.
The artificially intelligent systems that we use at the moment, in contrast, do not learn on a physical level. The physical level remains the same, they learn on the software level.
The researchers also ran the wire networks through what’s known as the “n-back test,” a common working memory task for humans. It’s a bit like the game Concentration, when you turn over pairs of playing cards to find two that match. You have to remember which card is which from earlier turns. Typical humans can recall about 7 steps back. So could this nanowire network.
Who knows, that box with old cables you have in the basement might soon enrol for college.
An international group of researchers has found out that Mars probably has a liquid core and it’s smaller than previously thought. They arrived at that conclusion by analysing seismic data from two Mars-quakes in 2021.
One quake was likely from a rock fracture caused by heat or stress. The other was triggered by the impact of a meteoroid. Being hit by a meteoroid is not great if you’re Mars, but it is great if you’re a seismologist studying Mars, because in that case you can very sharply pinpoint the source of the initial waves.
The data came from NASA’s InSight Mars lander, which retired from duty in December. On the left one of the last images it took, the seismometer on the surface of the planet.
Scientists can use seismic waves from earthquakes to figure out what is going on inside a planet because the waves move at different velocities depending on what they encounter. In the new paper they tried different possibilities for the size and composition of the core to see which one fit the data best.
They found that the best fit was a core radius of about one thousand 800 kilometres. For comparison, the core of earth has about twice that radius. Like Earth’s core, that of Mars is mostly molten iron, but on Mars it contains about twice as much light elements. Those light elements include a lot of sulfur, plus oxygen, carbon, and hydrogen, which together make up about a fifth of the mass.
Our sun and its planets all formed from the same protoplanetary disk, but they don’t have the same share of chemical elements. Scientists have models for how this happens but very little data. The composition of Mars is the obvious place to start.
It's also a message to all fitness gurus on Tiktok, some of us were born with a liquid core of low density, please leave us alone.
Astronomers at two British universities have figured out the main reason that quasars ignite.
Quasars, short for “quasi-stellar object,” are among the brightest objects in the universe. They sit at the centres of some galaxies, where huge amounts of gas get sucked into by black holes. The gas swirls around the black hole and heats up. That emits a lot of radiation, so much that we see it as brilliant light even from billions of light years away.
But not all supermassive black holes are quasars. For example, the one in the centre of our own galaxy isn’t. So why do some of them ignite and others not?
These British astronomers looked at the properties and shapes of 48 galaxies with quasars and compared them to 100 galaxies without quasars. The observational data came from the Isaac Newton Telescope on La Palma in the Canary Islands.
They found that nearly two-thirds of the quasars showed distortions in the outer regions of the galaxies, evidence of galactic collision. By contrast, only 22 percent of galaxies without quasars showed those distortions. They concluded that the dominant trigger for the formation of quasars is the collision of galaxies.
Moreover, in 61 percent of cases the quasars were visible before the nuclei of the colliding galaxies had merged. That means that rather than being formed at the peak of the merger, they can be formed before that.
I don’t want to be morbid, but you all know I want to. This finding is probably a prophecy for our own galaxy. In about 5 billion years, the Milky Way is projected to collide with Andromeda, and who knows, maybe we’ll make a quasar, too. So the future may be brighter than you thought.
A team from the U.S. and Germany has developed a vision for a superconducting highway. Yes, you heard that right.
It’s a riff on the old idea of magnetically levitated trains, or maglev trains, some of which are in use in Japan and China. Maglev trains work with superconducting alloys that repel magnetic fields very strongly, strongly enough to levitate a train carriage. This dramatically reduces friction and allows the trains to reach much higher speeds. Maglevs are rare because superconductors must be cooled to extremely low temperatures, which costs a lot of money.
This concept paper proposes to combine maglevs with hydrogen transport. Because, as you’ve probably heard, we’re supposed to get this wonderful hydrogen economy. The trouble with hydrogen is, well, many things , but for today’s purposes it’s that to transport it you need either high pressure or low temperature. So the authors of this paper want to run liquefied helium at about 20 Kelvin through highways, and that way convert the highway into one large superconductor.
Here’s the plan. See all that liquid hydrogen coming from the factory off to the side of the road? It’s cooling for the road, and at the same time storage plus helping transport the cars on the road. There’s a layer of liquified nitrogen and a vacuum layer on top of the hydrogen to keep the hydrogen cold enough.
And then, all you have to do is fit the undersides of vehicles with permanent magnets. That’ll make the vehicles levitate over the superconductor highway and they can zoom to their destination at 500 to 800 kilometres an hour. Sounds like fun. Just don’t step on the highway, you might freeze to the ground.
For the experimental part of the paper, they levitated a magnet above a superconductor. The rest of the technical details are still to come. Oh, ya, plus the financing. Maybe Elon will be interested.
No luck today.
A group of Canadian and Chinese researchers has created a fabric that can change both colour and shape when either heat or electricity is applied to it. When the fabric cools down or the voltage is removed, it goes back to its original shape and colour. Earlier smart fabrics have been capable of changing colour or shape but not both.
This new smart fabric is woven on a loom from two different types of plastic yarn. You can see it here in this video. One type of yarn, the blue one, is plastic mixed with a material that changes colour when it’s heated. The other is plastic plus fibres of stainless steel, which conduct electricity. The colour changes in 20 seconds when the heat goes from 20 to 60 degrees Celsius.
The researchers made a simple fabric model of a dragonfly – if you squint your eyes, you can see it here in this video – and heated it up. You can see its “wings” unfold as it turns from purple to blue. Once the heat is turned off, it goes back to its original shape. Somewhat. The recovery ratio was nearly 80 per cent after repeated transformations.
The changes can also be specific to just part of the fabric as you see here in this video. When the voltage is applied only to the left side of the fabric, the left changes. If it’s applied to the right, the right changes.
The researchers say that such fabrics could eventually be used as wearable biomedical devices or as environmental sensors, or maybe they could use them to make compression socks that I can put on without ruining my nails.
Guten Tag Olaf,
I’m not talking to you because last time I did no one had any idea who the German chancellor is.
A six-hour trip in a hot air balloon. For 130 thousand dollars. Well, I guess the hot air is true to character. But where does it go?
A round trip. That’s a lot of money just to go back to where you started. But then that’s what politics is all about, I’m sure people will understand that.
Tschuessi.
A group of Swedish researchers has done simple chemical calculations on a quantum computer.
They used an existing quantum computer at Chalmers University in Sweden and a conventional computer in addition. They fed the same simple reference problem into both computers – in this case, calculating the ground state energy of some small molecules: hydrogen, helium hydride, and lithium hydride.
Then they used the difference between the solutions to the reference problems to correct for noise on the quantum computer. This allowed them to improve the accuracy on the quantum computer by a factor of 100, as long as they combined it with an earlier correction algorithm.
So, they did a calculation on a quantum computer of which not only the result was known, but that could have been done faster on a conventional computer. Then they corrected the errors on the quantum computation with the conventional computer. Ok, that was a little underwhelming.
A group of European chemists and physicists have conducted an experiment showing the so-called Brazil nut effect without the need for external energy.
The Brazil nut effect is named after the phenomenon of shaking a bag of nuts only to find that the largest – the Brazil nuts – gather at the top. This happens because it’s more likely for the small pieces to fall down through the large ones than the other way round. But it only happens when you shake the bag.
Theoretical models suggested that the same Brazil nut phenomenon could happen spontaneously under the right conditions, and this is what they showed in this new paper. They used a mixture of different sizes of charged particles made of a hard plastic in some kind of oil. They don’t have to shake the mixture because the shaking is provided by Brownian motion, the random tiny movement of atoms as they bounce off each other. And indeed, the larger plastic particles – the green ones -- move upward more easily than smaller particles, which are coloured red in this video. It’s because they have a greater charge, so they repel each other more strongly.
It only works under certain conditions, though. You need particles with sufficiently large charges compared to their mass. When masses were too large or when the charge was too low, it didn’t work.
How matter arranges itself is an important ingredient for models of stars and planet formation. It also helps us understand how sediments are laid down in geology, archaeology, and children’s bedrooms.
Comments
Regarding the transport problem with hydrogen. We don't need to transport it, or at least transport it far. Hydrogen can be produced anywhere given the right mix of water supply and enough baseload power, such as with the integral fast reactor that I keep mentioning. Therefore, hydrogen can be produced locally everywhere and directly pumped into storage by high pressure pumps (https://www.haskel.com/en-us/industries/hydrogen) for distribution. The hydrogen problems are engineering problems with a bit of politics messing everything up. We could have commercial scale IFRs but for politics. Oh, and that Mars has so little magnetosphere is but one more reason that we will not settle Mars to anything but smaller Antarctic scale research facilities. We better reverse our destructive behavior here on the only planet that can support us and life as we know it.
Thank you Tracy, for the explanation. Didn't know anything about relicts of earlier 'quasar activity' in the milky way. Will go to Zooniverse this evening.
You're exactly right about quasars, Thomas. They are far more common in the early universe and it has long been thought that mergers (also far more common in the early universe) were the primary means of supplying the "food", but the merger hypothesis had never been adequately tested. A couple interesting things from this new quasar study: 1. they looked at relatively nearby quasars in order to get the spatial resolution needed to confirm that mergers are the dominant cause -- quasars in the local universe could be a biased sample, and 2. the correlation between quasar activity and mergers is not 100%, so there are other means by which plenty of "food" can infall to the central black hole of a galaxy to get quasar activity. Also interesting, our galaxy has what look like relic bubbles above and below the disk that are likely remnants of past quasar jet activity when the Milky Way was much younger. It makes me wonder if a past merger led to that quasar activity and what structural signatures could be looked for to support the hypothesis.
A lot interesting to learn here again. My knowledge about Quasars was, that they are all far away, so a phenomenon of an earlier phase of the universe, when there was much more 'food' for massive black holes. And I like the imagination, that this box in my cellar with countless old cables developes to a super-computer slowly
One of the neat things about the quasar study is that it used "citizen science" via Zooniverse (https://www.zooniverse.org/). Anyone can participate completely free in any of the various Zooniverse projects. Although it all started as astronomy/galaxy zoo projects over 10 years ago, Zooniverse has grown to many projects from many varied fields. I encourage everyone to go take a look and have a bit of fun.
Get Olaf to buy you all the snacks, including crinkle-cut fries with tomato sauce.
