Hacker News new | past | comments | ask | show | jobs | submit login
Physicists Have Identified a Metal That Conducts Electricity But Not Heat (2017) (sciencealert.com)
143 points by prostoalex 13 days ago | hide | past | web | favorite | 72 comments

That article is weirdly written and it is not clear how the title relates to the content.

However if that metal has a good electricity conductivity and a low heat one, wouldn't that make it an ideal candidate for seebeck-effect generators? My understanding was that the efficency of these cells is mostly limited by the heat conducted at the junction. If one of the metals at the junction is a heat insulator, it should work wonders, no?

The way they work is by having a bimetallic junction where a temperature difference between that junction and the other side of the wire pair is connected causes a voltage difference.

If you were able to prevent heat from going from the hot side to the cold side through the electrical wiring, yeah, I think the efficiency would go up. I worked on a battery project once where the internals were over 400C, and we needed like 000 gauge wiring which in copper meant our major thermal losses were through our conductors.

But the article is badly written and has a bad title. The actual article is here:


and indeed they specifically mention applications to thermoelectrics. The big takeaway was the 10x lower thermal conductivity than expected for the given electrical conductivity.

Kind of embarrassing that the linked article on sciencealert just copy-pastes random portions into a new article and doesn't even bother with the credits at the bottom.

Apparently it's transparent.

I'll have to remember this the next time I need to build a whale tank

> But a team in the US showed this isn't the case for metallic vanadium dioxide (VO2) - a material that's already well known for its strange ability to switch from a see-through insulator to a conductive metal at the temperature of 67 degrees Celsius (152 degrees Fahrenheit).

The next time you need to do whale sous-vide, i think.

The transition temperature can be tuned with dopants[0]. And the switching happens from transparent at low temperatures to reflective at higher temperatures.

That said, its highest transmittance is in the infrared, it's not that great in the visible spectrum.

[0] https://dr.ntu.edu.sg/bitstream/handle/10220/25845/for-your-...

Is that a Star Trek movie reference?

Yes. Original series when they went back in time. Scotty showed an earth engineer how to make transparent aluminum.

> how to make transparent aluminum

Which is just alumna (aluminium oxide, AKA ruby and other gems).

Homelab torch process can be seen on YouTube

A keyboard? How quaint!

proceeds to type at 400 wpm...


Using mostly his index fingers.

Suck it vim users, your editor has nothing on MOLECULAR DESIGN for System 3.0

Also makes you wonder how he was going to input that into the computer using voice commands.

It's not as surprising of you can imagine all the stuff probably going on in the background when you say "<assistant of choice>, play the next episode of Star Trek on my TV"

Lots of stuff going on, but based on concepts known to the developers. He had to teach the computer how to synthesize a hitherto unknown substance. Can't get around the information entropy of the stuff he'd've had to input into the computer.

How hard is it to say "take $known_aluminum_alloy and add $dopant using $heating_process"?

Obviously by using the corded speakerphone "Hello computer".

Aye laddie. Indeed.

Now if we only had computers we could talk to. (Without trying to sell us something in the process.)

I never understood why the whales needed to see outside their tank.

I never understood why the tank needed to be transparent.

It didn't need to be, it was transparent because it was glass.

They traded the formula for the glass. The company didn't manufacture transparent aluminum on the spot and give it to them.

My question is more, why not I’m just use steel or aluminum?

Because then you wouldn't get to see the whales.

It's just a movie. :-)

> The metal, found in 2017, contradicts something called the Wiedemann-Franz Law, which basically states that good conductors of electricity will also be proportionally good conductors of heat, which is why things like motors and appliances get so hot when you use them regularly.

Erm, no. That's not how it works.

Things become hot due to electrical and mechanical losses; if you run them long enough, the heat moves out no matter how good or bad of a heat conductor you use.

In fact, you usually want a not-too-bad heat conductor, to avoid overheating the active components.

That said, there are still good use cases for such a material, for example when conducting electricity through a heat shield (furnace or freezer).

Do I want to continue reading, after the second paragraph contains such confusing statements?

Came here to post this. Motors and appliances heat up because of friction and because they are less-than-perfect conductors of electricity. Their good heat conductivity is why they don't get even hotter.

There's more:

"The electrons were moving in unison with each other, much like a fluid, instead of as individual particles like in normal metals,"

Which describes a superconductor, but that word never appears in the writeup.

Subject is interesting but this article is clearly written by someone not very familiar with the science.

Is VO_{2}, a metal oxide, really a metal?

This actually annoys me a bit. Without knowings anything about it: it's a ceramic, almost certainly a semiconductor, and since its a semiconductor it can behave electronically like a metal under certain conditions.

Some semiconductors have poor phonon conduction, others quite good (ability to conduct heat).

Some semiconductors have good oxygen vacancy conductivity and poor electronic and hole conductivity. That gives you a transparent conductor.

And most ceramics are brittle. Is VO_{2} brittle?

In short, is this VO_{2} really a metal, or "just" a cool semiconductor?

The formal definition of a metal is just that it has overlapping conduction and valence bands.


So it doesn't matter whether it's brittle (superconductors are brittle) or whether it's an oxide or whether it's a ceramic as well (superconductors), or a glass as well (metglass). What matters is that at your temperature of interest the bands overlap so that you have free moving electrons.

But reading some other literature suggests that "metallic" VO2 is indeed a semiconductor (not actual insulator) at lower temperatures and thermal energy makes the electrons energetic enough to jump to the conduction band and become a metal. So I think I'd say it's a "hot" semiconductor.

Ruthenium Oxide, Tungsten Oxide, and Chromium Oxide (at least one of them) are also metallic, with 10-20x worse conductivity than copper. Just a weird fact I had to figure out for a project. Actual metallic, not temperature dependent just phase dependent (CrO2 vs Cr2O3 for instance).

On the other hand, alumina is a great thermal conductor (about 10x worse than copper) but has almost no electrical conductivity, so I'm not so sure about this law relating the two in the first place. I've never heard of it being a law, I've always just treated that as a rule of thumb...

Electrical conductivity is electron motion, and if you have electrons to carry charge, you also get thermal conductivity. So metals (in the general sense) are good electrical and thermal conductors.

If you have very hard materials, with high phonon conduction, you can get thermal transport via phonons. But, these are almost all poor electrical conductors, as you need very tight binding to get hard materials with high phonon velocity.

The prototypical example of the second case is diamond.

Oh yeah, fair. Good explanation. The relationship is one directional because there are other ways to carry heat.

Looks like his stuff has been known about for some time. What’s kept it from being used before now?

$14/kg for unprocessed V2O5, presumably VO2 is brittle.

Copper, despite all the news stories and it shooting up in 2011 (to almost $10/kg) is today only $5.70/kg pure.

I like the idea of tuning it to act as a passive way to heat and cool homes but compared to fiberglass insulation?

I buy thermoelectric and nonvolatile memory applications though.

One application is a thin window coating, you wouldn't need much of it. The manufacturing process is probably more expensive than the raw material.

This would likely not work well in desert areas over an appreciable period of time, since most vanadium compounds are pretty soft, and desert winds tend to contain very hard particles of silica.

I agree that the window application probably won't work (it's not like it's less thermally conductive than glass anyway), but why wouldn't they just coat the inside of the pane if they're worried about abrasive damage?

Either way, if I were to pick a technology to improve desert glass insulation it'd be with a coating that allows in light but reflects infrared. Windows are mostly radiative heat transport after all, you can always double pane them for more direct conductive insulation.

It's not about thermal conductivity, it's about IR transmissivity. Above a certain temperature it would reduce the IR transmitted to the inside, thus keeping things cooler in the summer while still allowing the inside to be warmed by the sun during winter.

Even in a desert the temperature-dependent behavior may be beneficial to cool during the night. But if it is not then yes, a different, harder coating that always reflects IR would be preferable.

It conducts heat, just much less than expected

Is it possible to have a material that doesn't conduct heat?

No, that would essentially mean that it doesn’t interact with neighboring matter (eg air bouncing off of it) since heat is really an emergent property of all the particles in a material bouncing/vibrating amongst each other. There are things that don’t conduct heat but they aren’t regular matter.

No, but you can get surprisingly close with aerogels:


The lack of a material, e.g. a perfect vacuum, as long as radiation doesn't get involved.

Radiation becomes the dominant heat transfer mode in that case though, usually gets really small at day to day temperatures though.

It's not that small. Blackbody radiation at room temperature is 450 watts per square meter. And it only takes an extra 15C to get 100 net watts transferred per square meter.

Radiation is always involved, even at very low temperatures.

Superconductors come pretty close.

From a quick google, Vanadium also exhibits interesting properties for superconducting. I wonder if studying these higher temperature electron effects could impact that field..

> which basically states that good conductors of electricity will also be proportionally good conductors of heat, which is why things like motors and appliances get so hot when you use them regularly.

NOOOOO!!!!!...... How can science reporting be so bad???

Because reporters aren’t scientists.

The article links directly to a page on the law so maybe a little grace could be given?


It is why they feel hot to human touch though :)

Educate us please?

Motors and appliances get hot because inefficiencies eventually become waste heat. That is the source of the heat. It propagates from tbe heat source to a surface you can feel through conduction and other mechanisms.


Motors become hot because they generate heat through friction and other forms of loss. Many other appliances generate heat this way (and some are designed to generate heat).

Conductive materials are likely to be good heat conductors because the electrons themselves can conduct heat, and because the crystal structures of conventional metals (e.g. copper) tend to be good for both.

These two effects are entirely unrelated.

Thanks, that makes sense.

The relationship is generally true for metals (https://en.wikipedia.org/wiki/Wiedemann–Franz_law) but an appliance being warm to the touch because you use it is not really related.

Making relatable analogies is hard.

but, if anything, the opposite is true. If I keep dumping a constant amount of heat onto a material it will get hot enough to cool itself.

So a piece of aluminum will be cooler than a piece of iron. The Al might transfer more heat to my fingers than the steel... or it might not (that depends on many more factors than just the conductivity).

That's not an analogy, it isn't explaining something in terms of something else.

Analogy : a comparison of one thing to another for the purpose of explanation.


That's exactly what your parent said.

Parent: "That's not an analogie, it isn't explaining something in terms of something else."

You: "Analogy : a comparison of one thing to another for the purpose of explanation."

Well, the parent says this isn't an analogy, as it's not "comparing one thing to another for the purpose of explanation".

You have to be patient and read past the inaccuracies and hyperbole in these physics reports. This is the first I've heard of VO2 and this surprising behavior and I think it's an amazing advance.

Quote: "The metal, found in 2017..."

Erm, say what again?

Here's wiki for vanadium, you might wanna check that year :


They’re talking about vanadium dioxide’s transition, but it’s still a poorly worded phrase

Form wiki:

The elemental metal is rarely found in nature, but once isolated artificially, the formation of an oxide layer (passivation) somewhat stabilizes the free metal against further oxidation.

That's literally 2nd phrase there, right at beginning. Tell me, does that strike you as something to be discovered only in 2017? Or the guys who made the discover(y)/(ies) already seen VO2?

Also wiki is wrong since Vanadium is as rare as Copper. Does Copper strikes you as something rare?

‘Elemental metal’ means the element not bonded to other elements (ie mostly pure vanadium). There’s no reason an element can’t be common but the pure element be rare.

Also the oxide the wiki article is talking about is V2O5, not VO2 (though VO2 was known beforehand - it’s easy to find articles about it from the 50s).

Maybe he meant that "elemental metal" is redundant?

In other words: can we call "metal" anything not in the periodic table?

Alloys of various elemental metals are also metals, but aren't elemental.

Ford was using Vanadium steel in the chassis of his model T in the early 1900s

In 1867 Henry Enfield Roscoe obtained the pure element.

The interior of a cathode ray tube?

From 2017

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact