Discovery: Metals can heal themselves | Hacker News Reader

Discovery: Metals can heal themselves

newsreleases.sandia.gov

90 pointsbobvanluijt28 comments

Certainly a discovery, but a new one?

Ok, this is based on an article from Nature and I'm just a random dude from the Internet, so take this with a grain of salt:

What they claim is what I've learned in materials science class 20 years ago.

Metals fatigue under cyclic loading because of micro-cracks. This effect can be quantified by the Wöhler curve, which very simply put tells you how often you can bend a piece of metal until it breaks.

From what I remember, it is well known that this fatigue reverses by itself slowly and depending on the temperature and type of metal.

It is true that the effect is negligible for steel at room temperature, which admittedly is the most interesting case.

But for other metals this self healing effect can actually be exploited at room temperature. The example we learned in class was the leaden rulers everyone of us owned back then. These are rulers you can bend to draw nice curves in technical drawings. They are made out of pure lead. If you bend them very very often in one sitting they can eventually break. But if you stop before and you leave them alone over night they will be like new.

At least that is how I remember the story.

Statisticians borrowed the idea of “simulated annealing” and used it as a regularization technique for some time now.

Annealing metal requires starting at a high temperature and gradually lowering it. Simulated annealing does the same thing with some appropriate analog of temperature.

There's no temperature lowering happening in TFA.

You owned a lead ruler 20 years ago? I get the idea but I had never heard of such a thing. I know about physical splines, but I thought those were flexible wood, or maybe plastic later on.

Oh yes, and apparently you can still buy them[1]. Mine was blue and probably a different brand (Faber-Castell, maybe) but otherwise looked the same.

I didn't use it much for drawing, though, because I liked the French Curves (or Burmester Schablonen, as they are called in German) much better. The lead ruler, on the other hand, was fun to play with during boring lessons, I like to think of it as my generations fidget toy.

[1] https://www.xn--praehistorika-archologenbedarf-dtc.de/produk...

Sounds super useful for tracing curves onto a material you need to cut.

Is there still lead in it? Any health issues? Is that why I can't find them in North America?

EDIT: I do find hits for Flexible Curve Ruler on Amazon but several reviews complain they don't hold their shape well. Are they made of a substitute that doesn't perform as well as lead?

Just looked them up and found some on Amazon from $17-40ish depending on brand, length, and probably other factors.

I had no idea these were a thing!

https://en.wikipedia.org/wiki/Lesbian_rule

Off topic but that lesbian rule would have been a great punchline to this sketch: https://m.youtube.com/watch?v=QIBwSHZAOU0

From anothet perspective, i was taught that metals in a non-oxidizing environment, if touching, will reform crystal bonds. So a crack deep inside a metal, if given time, can heal across a previous break. But this wound heeling will not realign complex structure, so the effect is more noticeable in non-alloy, such as a lead spline.

this is called cold welding

https://en.wikipedia.org/wiki/Cold_welding

Funny story about lead rulers:

An alternative name for them (and the name that I originally learned for them) is a "Lesbian curve" because the lead to make them originally came from Lesbos. I once went up to a woman working at an art store and asked if they had lesbian curves. Immediately after the words left my mouth I noticed that she was wearing a lesbian pride pin and based on her reaction probably misinterpreted what I was saying. I awkwardly made some excuse and left the store in embarrassment as quickly as possible. -_-

[Recycling a comment from an earlier submission yesterday]

Isn't this a bit like cold welding? [0] Unless something is in-between them or some force is pulling them apart, there's no big reason for two exactly-fitted pieces of the same crystalline metal to _not_ join together. It's not like the atoms "know" which side they belong to.

Actually... [tap tap tap] It looks like that's explicitly in the original Nature title: "Autonomous healing of fatigue cracks via cold welding." [1]

[0] https://en.wikipedia.org/wiki/Cold_welding

[1] https://www.nature.com/article

Cold welding can really mess you up when you deal with precision ground stuff.

Article in nature: https://www.nature.com/articles/s41586-023-06223-0

Metals have been known to have a kind of memory. If you bend a cable out of shape, it will bend itself right back. I have to think that this self-healing is an instance of the same principle.

Yeah, there are formulations of wire you can buy that are specifically designed to have this property to increase fatigue resistance and bend life.

Nitinol is neat

if you bend a thick wire of nitinol (nickel titanium alloy, discovered on accident, aka shape memory alloy) quickly, it heats at the site of the bend and that is cool, but paper clips do the same thing.

once you've let the thick wire settle back to room temperature---here's the cool part---quickly force it straight again and it gets suddenly very cold at the site where the bend was.

nitinol is neat

No preprint although government labs have direction to make their research openly available asap. Probably a paper written just for other researchers to cite and not for any generic interest.

It seems the paper isn’t in sandia’s publications database yet, but it’s only been a few days since the journal version was published. I’m not sure how it works at sandia, but at NIST one of the authors has to get around to uploading the manuscript and there’s no way to get an automated reminder. Then our publications database only updates once a day I think.

I’m sure if you send a polite request for the manuscript they would be happy to share. Maybe even responsive to a polite request that they post a preprint next time, but in MatSci that’s much more common with modeling than experimental work.

https://www.sandia.gov/research/publications/search/?pub_s=A...

Prof. Demkowic was a very good instructor at MIT.

Once I learned about the amazing world of dislocation growth/glide in materials, it helped me see physical objects in a whole new way.

https://youtu.be/EXbiEopDJ_g

The math describing these defect strings was also interesting and challenging :) I wondered if they could be used for one-way computation.

I thought reforming the bonds of microfractures was already well known as a differentiator between metalloids and true metals? So does this mean that metalloids just shouldn't be a group anymore, or that they instead have to be expanded to include more materials? And certain substitutional alloys can be argued to be self healing. In bronze the copper will replace any displaced tin and zinc bonds without heat, for example.

It's cool that they observed this, but I'm a little confused why this is seen as surprising. Unless there's oxidation at the fissure, metals seem like they should be able to glue back together pretty easily...

...unless the metal at the crack location reconfigures itself to a lower energy crystal structure due to the boundary condition. Which apparently is not happening here.

Isn't this the same effect as vacuum welding? Eg NASA doors in space station has to have some kind of isolation painting?

Reply on news.ycombinator.com