This certainly has been in the making for longer than the "everything we do must be for AI" bubble. In fact s390 has its own on-die inference engines and they have access to the same caching mechanisms as the main processor (which are quite insane).
It's about refining theoreticals models that are used to predict nucleosynthesis of heavier elements. The researchers used indium because we can obtain the required neutron-heavy isotopes for indium but not for heavier elements such as gold or platinum. But improving the model with data from indium, they say, makes it more accurate for gold as well?
Why then gold in the title? Probably just because it's shiny.
Platinum is also a peak of element abundance, together with its neighbor elements.
So any model of how the elements have been produced must explain why the probability of making platinum and its neighbor elements, osmium, iridium and gold was higher than the probability of making other elements.
The existence of other abundance peaks is easier to understand, e.g. the peaks at tin and at lead happened because these 2 metals have "magic" numbers of protons, i.e. 50 and 82, which correspond to complete nucleon layers.
The peak at platinum is higher to understand, so to explain it you need accurate models.
On Earth it is not obvious that the heavy platinum-group metals and gold are located on an abundance peak, because all these precious metals have gone deep inside the Earth, into its iron core, so the crust of the Earth is depleted in them, which has made them rare and precious.
There are asteroids where the iron cores are easily accessible and they contain great amounts of platinum and related metals. However, the idea that mining that would be easy is extremely naive.
On Earth, mining gold and platinum is easy, because they do not mix with silicate rocks so they can be found as native metals or sulfides/arsenides/tellurides that can be easily separated from silicate rocks and then the metals are easy to extract.
On the other hand, in asteroids platinum and the other precious metals are dissolved in iron uniformly, so they are extremely diluted, in proportions of less than 1 part per million. Therefore, even if the total amount of platinum and gold is huge, concentrating one gram of platinum from one ton of iron would be tremendously difficult, requiring a huge amount of energy.
Mining asteroids for the purpose of bringing something back to Earth will certainly not happen before solving much easier problems, e.g. growing back an amputated leg or any other part of the body. The fact that at least a startup exists that claims to work to achieve such mining is just a certain scam with no other goal than mine money from naive investors.
How much less? I believe most gold produced in the US is from ore with under a half ppm gold (E.g open pit mines in Nevada).
Maybe the point there is that we already have practically endless supplies of quarter ppm ore ready for the taking on the surface of the earth. Gold is rare only in so far that the current price reflects the breakeven point of these most abundant sources. Adding more supply with similar or worst production costs wouldn't change anything.
All the other precious metals are less than 1 ppm compared to iron, but platinum is more abundant, and by weight it is about 6 ppm in iron.
The advantage of an asteroid is that its entire metal core has 6 ppm of platinum and a fraction of a ppm of gold, while on Earth the quantities of ore containing such amounts of precious metals like a half ppm or a quarter ppm of gold are much smaller.
There certainly exists no "endless supply" of gold ore with a quarter ppm gold, because the average concentration of gold in the crust of the Earth is a few parts per billion, so the few places where the concentration is as high as a fraction of a ppm are compensated by vast areas where the gold concentration is much less than one part per billion.
While an asteroid may have a lot of iron containing 6 ppm of platinum and a little less than 1 ppm of gold, that is not comparable at all with a terrestrial ore with 1 ppm or a few ppm of precious metals.
The precious metals are the easiest to separate from rocks, which is why one can exploit on Earth ores with a so low content of metal. On the other hand, precious metals are very hard to separate from iron, which is the very reason why in any planet or asteroid these metals end up being dissolved in the iron core.
So the extraction of platinum or gold in so small quantities from iron would be extremely expensive on Earth and much more so on an asteroid, where it is impossible to produce most of the chemicals used on Earth, like acids or cyanides.
Those are the average abundances in the iron that forms the asteroid metallic cores, which are exposed in a few asteroids, presumably because of ancient collisions.
The asteroids where such cores are exposed, instead of being buried under huge amounts of rocks, like in the planets, are those that are targeted for mining.
The iron meteorites are pieces detached from such asteroid cores, so they provide samples of their composition.
Some meteorites, the so-called chondrites, come from small bodies that have never aggregated into bigger asteroids or planets since the formation of the Solar System, so they have a chemical composition close to the average composition of the Solar System.
Other meteorites have been detached from big bodies, like asteroids, planets (e.g. from Mars) or from the Moon.
These meteorites are either made of rocks, when they have been detached from the surface of such bodies, or made of an alloy of iron, nickel, cobalt, germanium, some times also silicon, together with other metals that are present in much smaller quantities, when they have been detached from exposed asteroid cores.
>concentrating one gram of platinum from one ton of iron would be tremendously difficult, requiring a huge amount of energy.
melting one ton of iron requires 500KWh, 12 gallons of gasoline, less than $100 on Earth. Or 5 Tesla car batteries fully charged by say 30x30 m solar array in 2.5 hours - cost nothing in space once you got the hardware there. This is why mining in space is going to be a pretty big thing once/if we get cheap launch capability.
Since you didn't show your math, I did a quick calculation. .45J/g/C specific heat of iron means .45MJ/tonne. 1811K to melt iron means 815MJ/tonne. 3.6kWh/MJ, so 226.4 kWh should melt 1t of iron.
Yes, but melting is just the beginning of the process. Even your computation is incomplete, because it is not enough to heat iron until the melting temperature, you must also provide the additional latent heat of melting. Similarly for boiling iron, after heating to the boiling temperature there is an additional latent heat of vaporization.
There is still no easy way to separate platinum-group metals from liquid iron, so you must vaporize the iron, to exploit the fact that platinum-group metals have higher boiling temperatures. It is true however that at the low pressures easily achievable in vacuum, vaporization is easier than on Earth.
Otherwise than by vaporization, you could dissolve iron with an acid, but on such asteroids you do not have with what to make an acid, so you would have to transport it from some other asteroid, or more likely from a satellite of Jupiter. You would also need a chemical plant to make the acid and also to recycle the iron salts into regenerated acid. This is so much more complicated, that vaporization of the iron might be simpler.
Finally, you must account for the fact that the energy required to vaporize one ton of iron produces less than a gram of platinum and of each other platinum-group metals. It is unlikely that you could build there a solar array big enough to provide energy for vaporizing a million of tons of iron, to make a ton of platinum, so you would need a nuclear reactor.
While platinum-group metals might be obtained as a minuscule residue after vaporizing the iron, gold has about the same vapor pressure as the much more abundant iron, nickel, cobalt and germanium, so it would be impossible to extract it from iron by vaporization. It could be extracted only with a chemical method, e.g. with an acid or with oxygen, which need to be brought from elsewhere.
Taking all these into account, it seems that there is no chance of being able to mine precious metals at a cost less than on Earth any time soon, e.g. in the current century. Extraordinary reductions in the cost of interplanetary transport would be needed and in the cost of building a metallurgic plant on an asteroid.
Mining asteroids would make sense only if some people would decide to live in huge space stations with artificial gravity, instead of on Earth, and then some asteroids would be mined for making steel and other construction materials, to be used in the interplanetary space, not on Earth.
> gold has about the same vapor pressure as the much more abundant iron, nickel, cobalt and germanium, so it would be impossible to extract it from iron by vaporization
>energy required to vaporize one ton of iron produces .... so you would need a nuclear reactor.
it is less than 2500KWh - under $250 of nuclear generated power on Earth. The best - fastest and efficient - way to travel outside planet's LEO that is available today is solar or nuclear powering ion thruster, with only nuclear really beyond Mars. So anyway you come into the asteroid belt with a reactor. A submarine or icebreaker like reactor - 70MW - would power vaporizing of almost 30 tons/hour of iron. Note, that nuclear reactor in space is tremendously cheaper than on Earth as all the regulation, safety, etc. costs either disappear completely or reduced a lot.
If you produce a few grams of a precious metal, that cannot justify the trip until there.
To produce something of the order of one ton, which still seems too low to cover the expenses, you need to process something of the order of one million tons of iron.
With your estimation that could take several years.
In reality the energy consumption would be much greater, because one must cut chunks of iron and transport them to the vaporization installation, then also transport elsewhere the condensed iron.
So you would need a decent number of submarine like reactors in order to achieve an acceptable productivity.
There is no doubt that it would be feasible, but the problem is that at the current prices there would be no way to recover the expenses.
Your calculation assumes the heat must be considered wasted, but what prevents a counter-current heat exchange configuration from attaining ridiculously higher efficiencies? not to speak of just using saner approaches like chemical separation (gold and iron are very different chemically)
Heat exchangers for metal vapors at temperatures of a few thousand kelvin would be a significant technical challenge.
A heat exchanger needs fluids between which heat can be exchanged. Besides the fact that it would be very difficult to have pipes for fluids at such temperatures, it would not be so easy to efficiently heat the fluid more than it was heated by the recovered heat and then control somehow a fluid jet to transfer efficiently heat to the iron that must be vaporized.
Even if some heat would be recovered from the vapors, the losses due to imperfect heat transfer from fluid to iron might be greater than the recovered heat. Moreover, it is not clear what could be used as the working fluid, because those asteroids are depleted in volatile elements, so any fluid must be brought from elsewhere and any fluid losses would be irreplaceable.
Probably the easiest and most efficient way to heat iron until vaporization would be with an electron beam, but it would not be easy to ensure that the iron vapors do not destroy the installation and they condense in a safe place, from which the iron can be somehow evacuated.
working fluid? the same hot iron vapour is used to heat the incoming molten iron, no heat exchanger is perfect so the preheat would inevitably be a few percent short of the target temperature, the remainder is just the energy you supply to negate any heat lost through insulation (space is large, so one could use a ridiculously large insulation)
not that any of this matters, since chemical methods would be much more efficient
Chemical methods would be much more efficient, but they would need huge amounts of chemicals that do not exist on asteroids, so they need to be brought from elsewhere.
It would be impossible to bring millions of tons of acid and of water, so if an acid would be used it would have to be regenerated, e.g. by the electrolysis of the iron-nickel-cobalt salts, which would also need a lot of energy.
Designing a process that could regenerate and purify the acid in a closed cycle, with no losses of any fluids, due to the difficulty of replacing them, would be a very difficult task. Nothing remotely similar has ever been achieved. On Earth, any such methods use at least vast amounts of water and air that are not recycled.
Also, any chemical methods would need to be performed inside a perfectly sealed installation.
Vaporizing iron and the other more volatile metals with an electron beam could be made in a partially open vessel, in the vacuum from the surface of the asteroid. The main difficulty would be to ensure that the metal vapor goes in a certain direction and not omnidirectionally, to avoid its condensation all over the installation.
When vaporizing metals in vacuum with an electron beam, you do not pass through a liquid phase, but the metal is vaporized directly from solid pellets. This method ensures a high efficiency of conversion between electrical energy and heat that is actually used for vaporizing the iron, instead of being lost in the environment.
Thus there would be no molten iron to be preheated, even supposing that there would be materials suitable for a heat exchanger working at such temperatures.
Moreover, even if one would first melt the iron in a closed vessel, heat exchangers transfer heat well only between dense fluids, i.e. liquids, supercritical fluids or at least gases at high pressures. The liquid iron qualifies, but not the iron vapor, from which transferring the heat would be bad. Better heat transfer could be achieved if the iron vapor would condense inside the heat exchanger, but for that a means to ensure a high enough pressure for the vapor would be needed, but that may be difficult to ensure without preventing its advance in the pipes. Liquid iron can be pumped with magnetohydrodynamic pumps, but for pumping iron vapor there is no easy method. Perhaps one could ionize the vapor, to be able to move it with electric fields.
A heat exchanger working at a temperature so high would tend to have a very high heat loss, due to radiation. It may be difficult to ensure that you recover more heat than the extra heat that is lost.
In any case both the attempt to use chemical methods or the attempt to make a heat exchanger for iron vapor would be engineering challenges that require solutions far beyond everything that has ever been done on Earth.
By contrast, vaporizing metals in vacuum with an electron beam is a routine technology on Earth. The only big challenge is that in normal vaporization installations the vapors go in all directions. On Earth this is not a problem, because everything around is covered with some thin metallic foil, on which the vapors condense. After that the metal foil is dumped if the evaporated metal is cheap, or it is sent to metal extraction by chemical methods if the metal is precious and it must be recycled.
On an asteroid, in order to avoid the deterioration of the installation, one needs either a method to move the useless vapor in a certain directon, e.g. by ionization and then moving with an electric field, or perhaps by making iron foil and covering the installation, and then working in batches, where one vaporizes an iron pellet, so-that the platinum-group metals remain in the pellet holder and the volatile metals are deposited on the iron foil, which is then dumped and the cycle repeats.
This is the only method that could be done with existing technologies, with minimal improvements over them.
However, it would not be worthwhile, as the cost of extracting thus platinum-group metals from an asteroid would be many times greater than on Earth.
when semiconductor boules are czochralsky grown, afterwards impurities are swept away by heating a zone, and moving the heated zone (the impurities dissolve better in the hot solid compared to cold solid), could one similarily move the gold by such transport?
instead of a heating device, a large mirror could focus sunlight on the asteroid, so that one doesn't need to do induction joule heating powered with solar panels
I'm wondering if simpler solvents for gold (like mercury) could work
Or perhaps faradayic electrodeposition of iron? like how conducting current through 2 copper electrodes in a copper sulfate bath can transport copper from one electrode to the other.
Obviously any proposals would have to be tested on Earth before porting to space...
A lot of processes commonly used on earth are not necessarily the most efficient ones, certain aspects like environmental regulations or poisonous or dangerous-for-human substances can preclude their commercial utilization, but that doesn't mean an automated refinery in space should avoid it too. Thus we can't just point at the properties of on-earth-commercial methods and assume space-based refineries would have to inherit the same issues.
You cannot use a centrifuge to separate solid iron.
Using a centrifuge with liquid iron would create a gradient of concentration of the heavier elements dissolved in it, but that would not be enough to separate them.
All that could be done with a centrifuge with liquid iron would be to obtain an iron alloy enriched in heavy elements. However, I doubt that it would be possible to make a centrifuge for liquid iron that would have a lifetime sufficient to process quantities of the order of one million tons of iron. I do not think that until now anyone has ever tried to make a centrifuge that could work with a liquid metal at such a temperature. Most materials lose their strength at such temperatures, so the risk of breakage for the centrifuge would be extremely high, a risk that is increased by how heavy iron is.
It is also not clear if such an enrichment of the heavy elements would bring a sufficient simplification to further processing steps to make it worthwhile.
Iron and platinum have different melting points. If you melt the alloy, then spin it to concentrate the platinum, couldn't you coax the platinum to separate out as solid clumps by adjusting the temperature?
Alternatively, there are differences in magnetic properties that could be exploited...
This isn't my field, so I'm just spitballing. I bet if you can get the cost of launch and interplanetary transit to be low enough for people to really start tinkering with asteroid mining though, someone will crack the metallurgy issues...
Different melting points are easy to exploit only when metals do not mix in liquid state.
Even when metals do not mix in solid state, but they mix in liquid state, that usually cannot be used for separation, because the liquid solution will become solid at a temperature different from the melting temperatures of the components and lower than them, and the solid alloy will consist of the component metals intimately mixed at the level of microscopic crystals, so you cannot separate them (this is called an eutectic alloy, like the lead-tin alloy used for soldering, where by solidifying it you do not obtain separate lead and tin, but just a non-separable alloy, and by remelting the solid alloy you obtain a liquid solution, where again, the metals cannot be separated).
If the metals also mix when solid, the solid metal is a solid solution that does not melt at any of the melting temperatures of its components, but at an intermediate temperature, and the metals cannot be separated regardless whether the alloy is solid or liquid.
Here, in asteroid cores, the precious metals are present in a very small proportion, so they form either a liquid solution when molten or a solid solution when solidified.
The melting temperatures of platinum et al. do not matter, the melting temperature of the alloy is slightly lower than that of iron, corresponding to that of an iron-nickel alloy. The other alloying elements are in quantities small enough that they have negligible influence on the melting temperature.
In conclusion, differences in melting points can only very seldom be exploited for metal separation and they cannot be used for the iron alloys of planetary or asteroid cores.
You can exploit only either the difference in boiling points or the differences in chemical reactivity with acids or oxidizing agents.
True, but even with that, the amount of siderophile elements like platinum and gold in the crust is much less than in the core of the Earth ("siderophile" means that at the contact between molten iron and molten silicate rock such elements go into the molten iron).
Without that impact, it is assumed that almost no platinum-group metals and gold would have remained in the crust.
> Without that impact, it is assumed that almost no platinum-group metals and gold would have remained in the crust.
Wow, its wild to think of a counterfactual world without gold. Would those metals have emerged to the crust from volcanism or is that material not sourced deeply enough?
Volcanism at most brings material from the upper mantle, but usually such material becomes mixed with material from the crust, while ascending.
The mantle has slightly bigger concentrations of precious metals than the crust, but the concentrations remain many times smaller than in the core.
The reason is that both the mantle and the crust are made mostly of silicate rocks. The mantle is made of heavy silicate rocks and the crust is made of light silicate rocks, which float on the denser mantle.
The metals that are resistant to oxidation do not mix well with silicates, so they tend to segregate from them, and then, being heavier than rocks, they tend to descend towards the core. If they reach the core, then they dissolve into the melted iron.
When lava is expelled by volcanism, the precious metals contained in it usually separate from the silicates together with metal sulfides and arsenides, which makes them easier to find than if they were dispersed uniformly in the rocks. Other elements that ere much more abundant, for instance germanium and gallium, are harder to mine than the precious metals because they are not concentrated in distinct minerals but they are uniformly distributed in many rocks.
Stackful coroutines also can't be used to "send" a coroutine to a worker thread, because the compiler might save the address of a thread local variable across the thread switch (happened in QEMU).
> all the inputs and outputs share the same ground, it's not just the values for that pair of wires?
No, it depends on the converter. There are converters that leave 160V on the DC power rail for a 110V AC input, and 155V on the DC "ground" rail.
They are economic and you could find then when galvanic isolation is at least in theory not important, but they're terribly unsafe when used on PCBs that people might muck with.
If you have some "normal" converters and some of this kind, sharing the ground would be quite dangerous.
I have done some projects that needed some generic dc-dc converters from aliexpress (eg stepping down 12v to 5 or 3.3) I alway treated the output of each step down as a pair of wires that share no ground. It sounds like that would be overkill if they were reputable but it's probably best to not try tying the grounds together.
I figured any happenstance from the multimeter that the grounds match was transitory and not to be trusted.
> meant to have good ratios and good (de)compression speeds as compared to other tools
That does not mean it's Pareto optimal; Pareto-optimality forms a curve and while zstd, LZMA, LZ4, ZPAQ all want to be as close as possible to the curve, they focus on different parts of it. In particular zstd tries to stay on the middle part of the curve, while LZMA and LZ4 focus on opposite sides
Also, the Pareto curve is not necessarily known in advance. All you can do is add more and more algorithms or tweaks to understand what it looks like. For example, this blog post [https://insanity.industries/post/pareto-optimal-compression/] shows that prior to zstd, bzip2 and gzip2 were both pretty much Pareto optimal in the same area for ratio vs. compression speed. LZMA at low settings was a bit better but much slower. There was a huge gap between LZMA and LZ4, and bzip2/gzip filled it as best as they could.
The same blog post shows that zstd is an absolute speed demon at decompression; while not all zstd settings are Pareto optimal when looking at size vs compression speed (in particular LZMA wins at higher compression ratios, and even considering zstd only there's hardly a reason to use levels 11-15), zstd is pretty much Pareto optimal at all settings when looking at size vs. decompression speed. On the other hand at intermediate settings zstd is faster and produces smaller files than gzip, which therefore is not Pareto optimal (anymore).
Interesting approach. The fastest of the 4 presented compressors ("LSTM (small)") is 24 times slower than xz, and their best compressor ("LSTM (large1)") is 429 times slower than xz. Let alone gzip or, presumably, zstandard (not shown in paper). They also ran the models on different CPUs (a Core i5 and a Xeon E5) so the results are not even comparable within the same paper. A linked webpage lists the author's decompression times, which are even worse: xz decompresses twelve thousand times faster (50MB/s vs. 4kB/s) when nncp has an Nvidia RTX 3090 and 24GB RAM available to it, which apparently speeds it up by 3x compared to the original CPU implementation.
At half the size of xz's output, there can be applications for this, but you need to:
- not care about compression time
- not be constrained on hardware requirements (7.6GB RAM, ideally let it run on a GPU)
- not care about decompression time either
- and the data must be text (I can't find benchmarks other than from English Wikipedia text, but various sources emphasize it's a text compressor so presumably this is no good on e.g. a spacecraft needing to transmit sensor/research data over a weak connection, even if the power budget trade-off of running a GPU instead of pumping power into the antenna were the optimal thing to do)
Wait, this is a modern-ish ISA with a software-managed TLB, I didn’t realize that! The manual seems a bit unhappy about that part though:
> In the current version of this architecture specification, TLB refill and consistent maintenance between TLB and page tables are still [sic] all led by software.
reply