This year we pushed the boundaries of science to new extremes. We saw new levels of supercomputer performance, peeled the plastic off the biggest telescope and solar plant on earth, did the first-ever spectroscopy on antimatter, and even made liquid light.
To start with, we put spinners on the world’s most powerful X-ray laser. The SLAC National Accelerator Laboratory is a linear electron accelerator at Stanford. The X-ray laser is created by forcing electrons through a gauntlet of carefully spaced magnets, which provokes the emission of a beam of X-rays so powerful that it can even image a single molecule, without needing to produce a crystal first. This means that it can be used to capture chemical reactions in progress.
The current XFEL laser, no slouch by any means, is still limited by its copper linac line. So, Stanford teamed up with Cornell, Berkeley, Fermilab, Jefferson Lab, and Argonne to install a niobium superconductor, because it’ll let the linac run ten thousand times faster and brighter than ever before. LCLS-II, as the new accelerator will be called, will be capable of producing bursts of electrons at rates of up to a million pulses per second.
ATLAS is definitely the most visually imposing of the LHC experiments.
Where LCLS-II is like a microscope, used to image specific tiny things, proton colliders like the LHC are closer to sledgehammers. CERN is no stranger to extremes, this year in particular. In 2016 the LHC produced as much data as all its other operating years combined, and dropped 300TB of it to the internet with a resounding thud. Why so much data now? The LHC is operating at the highest energy (13 TeV) and greatest luminosity of any particle accelerator, ever. And it’s only going to get bigger and brighter from here. Their planned HL-HLC upgrade is expected to increase collisions tenfold.
Particle physics isn’t the only thing to use a focused beam. Dubai’s water and power authority is taking final bids for developers on their massive concentrated solar plant. The project is the largest concentrated solar power installation in the world, slated to generate an outrageous 1,000 MW by 2020, and 5,000 MW — 5 gigawatts — once it’s running at full tilt in 2030. Concentrated solar plants use mirrors to corral photons from the sun into a beam so dense it’s capable of melting salt to store energy and run steam turbines even when the sun isn’t shining.
Ivanpah concentrated solar thermal power tower and mirrors, seen from above
That extreme intensity is also capable of char-broiling birds on the wing if they’re unlucky enough to fly into the death ray between the concentrating mirrors and the solar panels. The Ivanpah plant, in California, actually gave the poor dears a name: “streamers.” Fun math exercise: If Ivanpah has 176,000 heliostats spread evenly around three collecting towers, that means each collecting tower has (roughly) 58,000 heliostats, each with their two mirrors, trained on it. Assuming that each heliostat is covered in dust and dead bugs and so returns only 90% of the light it catches, if an average size bird intercepted just one angular degree’s worth of focused solar radiation at its point of greatest intensity, it would experience the effective output of about a hundred and fifty suns. SolarReserve told CleanTechnica that the solution was to slightly rearrange Ivanpah’s heliostats such that they focused more diffusely into a flat disc around the collecting towers. This way, even if a bird flew through the most dangerous zone, it still wouldn’t experience “more than 4 suns at a time.”
Beyond the big hot intense powerful stuff, though, this was also a year of developments in the world of the very small and very cold. Researchers shot a laser into a declivity in a slice of semiconductor, forcing photons from the laser to interfere with excitons from the semiconductor and forming an ephemeral thing called polaritons. At the cryogenic temperatures they were working with, the polaritons were willing to crowd in closely enough to form a Bose-Einstein polariton condensate: liquid light.
How does this work? Remember that photons and electrons have both wavelike and particle-like properties, and they both have the mass-related property of momentum. When they’re forced to interact by being mashed together close enough, even though their particle selves don’t want to touch because photons are chargeless, their waveform selves will engage in interference through the medium of the all-permeating quantum field. That’s how they can interfere with each other’s momentum. But the electrons sort of wear the pants in that relationship. The polaritons mostly act like electrons and obey the “right hand rule,” changing how they rotate in 3-space in response to electrical fields they’re exposed to. The researchers were able to toggle the spin of the polaritons, from clockwise to counterclockwise and back again, by toggling the electrical field inside the semiconductor, as with AC power. The result: spin polarized photons. And they did so at a cost of half a femtojoule per toggle.
Speaking of lasers, scientists also did the first laser spectroscopy on antimatter. Antihydrogen behaves exactly the same as hydrogen when you bounce a photon off it, which confirms the null hypothesis that antimatter and matter are exact inverses of one another. We still don’t know why there’s more matter than antimatter.
2016’s Nobel Prize in physics went to a trio who sorted out the mathematics of phase transition on the quantum scale. Using donuts.
Phase transitions in a material start out as vortices on the quantum scale. When it’s cold enough, they hover politely in place, or even stop entirely. When it warms up, they pelt away from each other through the material, and propagate the disorder that creates phase change. The trio of researchers used the mathematics of topology to describe this observation with equations. Topology is its whole own can of worms, but it deals with integers, whole numbers, not necessarily fractions or the whole set of rational numbers. You can have part of a whole, but you can’t have part of a hole. You also can’t have part of a vortex, or 3.7x of a vortex. You either have one, or you don’t. Describing the transition from not-vortex to vortex is what won these guys the Prize.
2016 was also a year of speed. For the seventh year in a row, China has the world’s fastest supercomputer. The Sunway TaihuLight is now the fastest system in the world, running at 93 petaflops: more than three times faster than the previous fastest supercomputer in the world, the Tianhe-2, also from China. What will they do with this much processing power? Climate and weather modeling, life science research, manufacturing, and data analytics. It’s becoming more and more difficult to make gains in how far we can miniaturize transistors. Moore’s Law isn’t a law like we call laws in physics laws. Right now, we’re scraping the absolute bottom of the barrel in terms of component size. To carry a given quantity of electrical current without melting, a copper wire has to have a large enough cross-section and be able to shed heat fast enough.
These problems are still significant, and we haven’t had any enormous breakthroughs, but 2016 delivered some significant advances nonetheless. One way to increase chip density without relying on smaller geometries is to build 3D chip stacks. Samsung moved from 32-layer 3D NAND to 48-layer 3D NAND this year, and has outlined plans to drive SSDs to hard drive-equivalent prices through the further advancement of this technology. Meanwhile, HBM2 memory hasn’t shipped on any consumer devices, but SK Hynix has announced that the next-generation GPU memory is now in volume production, with debuts scheduled on consumer cards in H1 2017. But even with 3D chip stacks, you have to be able to get all that heat out somehow.
No comments:
Post a Comment