Hacker News new | past | comments | ask | show | jobs | submit login
YInMn Blue (wikipedia.org)
212 points by zeristor 30 days ago | hide | past | web | favorite | 97 comments

> Under this project, he was particularly interested in synthesizing multiferroics based on manganese oxides. He directed Andrew E. Smith (Subramanian's graduate student) to synthesize an oxide solid solution between YInO 3 (a ferroelectric material) and YMnO 3 (an antiferromagnetic material) at 1,093 °C (2,000 °F). The resulting compound was not an effective multiferroic; it was instead a vibrant blue material.

The quintessential chemistry story: someone sets out to make something and ends up creating something useful but completely different by accident.

The discovery of artificial sweeteners is rife with this,

Saccharin: Chemist working on coal tar derivatives noticed a sweet taste on his hand from contamination.

Aspartame: Chemist working on anti-ulcer drugs discovers his finger's taste sweet when he licked them to turn a page.

Sucralose: Chemist working on novel uses of sucrose and synthetic derivatives asked a coworker to 'test' a compound. Coworker thought he said, 'Taste'.

Is the mortality rate for Chemists higher than average?

Tasting unknown chemicals can't always end so well can it?

It's a good question.

Reminded me of these entertaining columns: https://blogs.sciencemag.org/pipeline/archives/2011/11/11/th...

"Say what? I’d call for all the chemists who’ve ever worked with a hexanitro compound to raise their hands, but that might be assuming too much about the limb-to-chemist ratio"

They're all educational and very entertaining, even to a non-chemist like myself. Recommended for a chuckle, there are some really awful compounds he covers with wit.

Edit 1: original link de-borked.

Edit 2: full list of his 'things I won't work with' columns https://blogs.sciencemag.org/pipeline/archives/category/thin...

Just to give you an idea of why that series of articles was so darn entertaining:

a method to make a more stable form of it, by mixing it with TNT. Yes, this is an example of something that becomes less explosive as a one-to-one cocrystal with TNT. Although, as the authors point out, if you heat those crystals up the two components separate out, and you’re left with crystals of pure CL-20 soaking in liquid TNT, a situation that will heighten your awareness of the fleeting nature of life.

In a similar vein, https://blogs.sciencemag.org/pipeline/archives/2010/02/23/th...

"At seven hundred freaking degrees, fluorine starts to dissociate into monoatomic radicals, thereby losing its gentle and forgiving nature."

further on

"The paper goes on to react FOOF [dioxygen difluoride] with everything else you wouldn’t react it with: ammonia (“vigorous”, this at 100K), water ice (explosion, natch), chlorine (“violent explosion”, so he added it more slowly the second time), red phosphorus (not good), bromine fluoride, chlorine trifluoride (say what?), perchloryl fluoride (!), tetrafluorohydrazine (how on Earth...), and on, and on. If the paper weren’t laid out in complete grammatical sentences and published in JACS, you’d swear it was the work of a violent lunatic."

And who could forget:

"But I have to admit, I’d never thought much about the next analog of hydrogen peroxide. Instead of having two oxygens in there, why not three: HOOOH? Indeed, why not? This is a general principle that can be extended to many other similar situations. Instead of being locked in a self-storage unit with two rabid wolverines, why not three?" [1]

[1] https://blogs.sciencemag.org/pipeline/archives/2014/10/10/th...

A lot of the articles make me wonder if the labs that do highly reactive work are staffed by people who thought EOD jobs were too quiet.

If I remember correctly, chemists back in the day didn’t live too long. Chemists used to use copious amounts of chemicals that today are considered carcinogenic and tetragenic. College chem labs today tend to use far less toxic chemicals then they used too, in addition to using much smaller amounts.

Even Newton was half-poisoned by the time Principia was published.

How about the dangers of electronics repair?


I mean...there are SO many sweet-tasting chemicals. But most of them are toxic.

Such the yummy lead white paint that kids peel from the walls

See also, Polytetrafluoroethylene (Teflon): https://en.wikipedia.org/wiki/Polytetrafluoroethylene

See also also, lysergic acid diethylamide (LSD): https://en.wikipedia.org/wiki/History_of_lysergic_acid_dieth...

See also also also, mauveine (Mauve): https://en.wikipedia.org/wiki/Mauveine

See also also also also Scotchgard

See also also also also also Saccharin

See also also also also also also Viagra

See also also also also also also also Post-It notes (specifically, the adhesive they use)

Given that the entire field of chemistry appeared in this fashion from alchemy, I suspect the tradition is very well ingrained by this point.

Why does this happen so much? I mean, this is science: shouldn’t chemicals react as they do in theory? It doesn’t seem like it’d leave a lot to chance.

Computational chemistry involves solving multiple NP-hard problems. There is an increasing reliance on computational models and simulation, but chemists realize that the models are approximations (since it's computationally infeasible to have perfect fidelity simulations). The only surefire way to observe a compound's properties is to synthesize it and measure.

Surely there can't be any problems in computational chemistry that have actually been proven to be NP-hard? Simulating an experiment atom-for-atom is in BQP, which is not known to contain NP (and is suspected not to contain NP).

EDIT: To put it another way, we can't solve NP problems by doing chemistry experiments.

Layman understanding ahead, see scott aaronson's NP-complete Problems and Physical Reality[0] if you want more rigorous arguments.

BQP gets you collapsing measurements.

The Schrödinger equation (NP-hard) gives you all probabilities at once, solving it is equivalent to performing non-collapsing measurements.

[0] https://www.scottaaronson.com/papers/npcomplete.pdf

Surely you would need to be modelling at a subatomic level to accurately simulate chemical interactions?

Sure. It's not an essential distinction though, because subatomic physics is also simulable in BQP.


Here’s one example. Though it’s probably different from what you had in mind


Chemistry never had good theoretical classical theory like physics had. Theories have been just rules of thumb or heuristics based on experiments.

Theoretically all chemistry is quantum physics with nucleons, electrons and photons (quantum chemistry). You need to numerically simulate what happens in reactions and it's very expensive.

Predicting reactions is hard. But why is predicting the physical—rather than chemical—properties of simple inorganic molecules like this one hard? Presumably, trying to figure out how light interacts with a molecule is an easier problem than, say, protein folding, no?

Light interaction is physical reaction of photons with electrons and it depends on how the electron cloud is distributed. All the vibrations and complex geometries make it really hard to calculate exact solutions.

Even that is not always enough. Gold has appears yellow because 5d orbital's distance from nucleus increases and 6s orbital's distance decreases due to relativistic effects. You need relativistic quantum chemistry to figure that out.

Sure, it’s challenging math (if you’re doing pure analysis rather than a numerical computation), but that doesn’t imply that it’s NP-complex math, does it?

Getting all the details in such a model right seems to be the same kind of problem as getting a weather-prediction model right—in that, as you add factors, all your predictions asymptotically approach correctness, with smaller and smaller preturbations.

Those preturbations might cause one or two elements “local” to some part of the configuration-space to be misclassified; but remember, here, that the point is to get a predictive screening test, one that hopefully gives more false positives than false negatives. Any time it says “this molecule might look cool if you made it”, we can just make it and see; and then, if it looks different than we were expecting, we can refine the model with the additional details that make that novel molecule also appear correctly in it.

The current upper limit is molecules with 40 electrons or so with 1-2 kJ/mol errors. After that the accuracy decreases dramatically.

YInMn molecule has way more electrons that can be solved numerically. Indium atom alone has more than 40.

No, because there’s a bunch of quantum goop that gets really complicated and your eyes not going to be able to replicate your theoretical 100% pure result anyways.

What does properties mean, to you? For the sweetness test, the property is the way the compound (or its breakdown constituents, or the compounds whose formation it catalyzes) reacts to our bodies.

So you're trying to figure out how our tastebuds work, in response to a novel stimulus.

The fact of whether or not a chemical reaction takes place in a certain context is a chemical property of a substance. Any fact about a substance on its own, not interacting with other molecules (or at least interacting with them a way that causes no chemical reactions to take place, e.g. being cooled by convection of Neon gas, or scratching in a Moh’s hardness test, or piezoelectric effects), is a physical property.

Sweetness is, as you’ve defined it, a chemical property, so it’s not one of the things I was talking about being “easy” to predict.

But pigmentation is a physical property, as it is a reaction between a molecule and inbound photons†. Likewise, the output of a “pigmentation calculation”—if we know of one—should just be more photons. Wavefunction in, wavefunction out.

We know how photons work. We know how our eyes perceive them. As long as we have a framework for figuring out how a material physically manipulates the photons absorbed by it, we should be able to calculate the pigmentation of every simple molecule by brute force.


† The pigmented molecule can be in the chemical form or physical state that gives it that pigment very ephemerally, due to e.g. a clock reaction. But at any given instant in time, we can say “the molecule is in this state, so what function does it apply to inbound photons now.”

Note also that I’m defining pigmentation here, not total appearance, so this ignores any fluorescence, phosphorescence, triboluminescence, Cherenkov radiation, etc. that the molecule also emits and which we perceive as light. All those purely-physical processes are able to be admitted into this model without actually making it untenable, but it becomes harder to understand, since it becomes a Feynman diagram thing, rather than simple transformation over wavefunctions.

It's often impossible to theoretically predict how two chemicals will react. The interaction of the molecules is governed by quantum mechanics, which can't be simulated quickly by classical computers.

In the words of the Great Yogi (Berra);

'In theory there is no difference between theory and practice. In practice there is.'

For a small set of other color-related videos, have a look at two from Tom Scott:

• "The Library of Rare Colors": https://www.youtube.com/watch?v=rApTzWboLrA

• "I Can't Show You How Pink This Pink Is": https://www.youtube.com/watch?v=_NzVmtbPOrM

The latter video makes me wonder: Is there a good way of finding out if a monitor, or even an entire color space, can contain this color? This may be a pigment where you really can't tell exactly what it looks like on-screen, because either the screen cannot display the color or the color space being used can not encode the color (or because the camera sensor can't properly capture the color).

I actually ordered some of this Pinker than Pink! As well as Black 2.0 and a few of his other pigments.

I believe what he's doing is mixing in a pigment that fluoresces from sunlight, at the same frequency as the primary visible-spectrum pigment. Thus, more of that frequency of light is coming off the page than the visible light landing on the page.

It's pretty crazy to look at in sunlight. Indoor, it's not quite as vibrant.

Like the optical brighteners that are used in detergents: https://en.m.wikipedia.org/wiki/Optical_brightener

Not quite as vibrant, but still startlingly pink.

>Is there a good way of finding out if a monitor, or even an entire color space, can contain this color?

You are describing the gamut. You can generally look up the gamut for any decent screen. Also, image editors have gamut warnings for going between colour-spaces.

> YInMn Blue is chemically stable, does not fade, and is non-toxic.

I'm calling a big old [citation needed] on a compound that's only existed for a decade. It might not be acutely toxic, but recently we seem to be finding all sorts of subtle effects from compounds that've been in common use for decades.

How does one establish non-toxicity?

You test it with animals.

It's inorganic, has simple structure and contains only non-toxic compounds, so there is no reason to be worried.

If you want to worry, direct your attention towards Vantablack and other carbon nanotubes. It has specific target organ toxicity from single exposure and while it does not appear to be carcinogenic it needs more testing.

yttrium and indium are both suspected to be toxic, manganese can certainly be toxic. no idea about the relative doses, I suspect that the chemists certainly thought about this etc, so I'm not trying to say YInMn is toxic, but "inorganic, has simple structure" doesn't mean anything, as it also describes Cobalt Blue, which is definitely toxic.

> It's inorganic, has simple structure and contains only non-toxic compounds, so there is no reason to be worried.

So is asbestos...

I knew that someone would reply with asbestos when I wrote that. "simple structure" is there to exclude asbestos.

Asbestos is dangerous because it's a tiny spike that sticks into cell. Just like nanotube of similar length.

Why do people get to patent the products of federal grants?

Why should you deprive an inventor of the rights to his invention merely because he's funded by the government? Do note that, in these situations, you have to notify the government of your intent to file a patent (and you lose the patent if you fail to do so!), and the government gets an automatic license to the patent.

> Why should you deprive an inventor of the rights to his invention

"Rights" to inventions aren't innate, they're explicitly granted. Why should we grant an exclusive monopoly on publicly funded research?

To add a profit motive, which in effect accelerates the fruits of the research being made available to the public and not hidden on the shelf of a lab. At least, that's the theory...

Tax money doesn't come from the government, any more than a politicians power comes from the government. It comes from the people. It ought to serve the people, not the recipients.

Everywhere I've ever worked claimed patent rights over anything I invented.

>Why should you deprive an inventor of the rights to his invention merely because he's funded by the government?

Same reason the company i work for owns any code I write while I'm on the clock.

I’m funded by my employer and my parents are assigned to and owned by them. Why isn’t this analogous?

I just hope this is a typo... ;)

Sounds like Worryfree from 'Sorry to Bother you', strange to think that film is utopian in that unskilled labour isn't replaced by robots

Because it was in the public’s interest to fund them and the public should reap the benefits. It should be “open”. I know this is largely a philosophical difference of view, but let me ask you: why should the public fund anything if it doesn’t do anything to benefit the public?

Patent rights exist because they benefit the public. A patent is a grant of a monopoly of an invention for a limited amount of time. After the patent expires, then the public is free to use the invention. I get what you're saying, but states typically operate on a different time-scale. So, long-term the public definitely benefits. They also benefit in the short term by having access to an invention -- in this case a pigment -- that didn't exist before.

The government gets an automatic license to that patent, so there is also that benefit that the government doesn't end up paying for the invention multiple times.

The counter argument to your question is: why should I work to invent something if I can't benefit from it?

I believe that the concept of the patent is a decent compromise between the two points of view. I don't think that the question of who does the initial funding is necessarily as important. And the Bayh-Dole act was specifically written as an effort to encourage more federally funded research to be made commercially available to the public. Before Bayh-Dole, this compound may have been academically interesting, been researched, published, but then left on the shelf for many years before someone found a commercial use for it. Now, trying to find a commercial use is actively encouraged (or required depending on which tech-transfer office you're talking to). The thinking behind this is that an invention that is available in the market (even patent-protected) is better than an invention that is sitting on the shelf in a lab. Eventually the patent protection runs out and the public gets an even better benefit.

But you're right that it's a trade-off.

>Patent rights exist because they benefit the public.

I don't see how your following statements support this claim. Specifically in the short term, you say that the patent provides access to an invention (through the company or whoever will sell the invention). Is this necessarily the case?

>[...] but then left on the shelf for many years before someone found a commercial use for it

It doesn't seem clear to me how the ability for something to be privately patented suddenly makes it accessible to the public.

If government funded research discovered this pigment, then DuPont can still benefit from the process-- they'd just have to pay royalties to the government (i.e. the taxpayers) whose money gave rise to the invention in the first place.

"[...]state influences on innovation and technological developments within the private sector using Apple as an example, for the way they popularized the government created technologies of GPS navigation, touch screen technology, and voice recognition into the modern smartphone. She also gives the example of how the US National Science Foundation funded the algorithm which helped create Google's search engine. Mazzucato argues that the private sector makes up the last and least risky part of technological innovation and entrepreneurship." [0]

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

> Specifically in the short term, you say that the patent provides access to an invention (through the company or whoever will sell the invention). Is this necessarily the case?

It is not necessarily the case that a patent will lead to a commercially available product. But, a patent is a public document. At the end of the protected time period, the invention described by the patent is available for the public to use.

If the patent didn't exist, the secrets behind an invention would be hidden from the public. But, this comes with a risk for the company in question... if someone else figured out your secret (independently), then you would have no protections and your secret could then be used by this new competitor. A patent is a defense against this. An inventor agrees to make their secrets public, in exchange for a time-limited monopoly to their invention.

Both parties gain something. The public gets to know how something works -- and the rights to use this knowledge for free in the future. The inventor gets a short-term monopoly on this IP and immediate protection from competitors to make money.

> It doesn't seem clear to me how the ability for something to be privately patented suddenly makes it accessible to the public.

Because the institution (normally a University) now has a motive to market this invention. Before Bayh-Dole, something like this pigment would have been noted in a lab notebook and maybe the bright blue color would have been mentioned in a journal article. But, because the University can now use this IP to license the pigment commercially, there is a motive to pull this IP off the shelf at the lab and into the market.

> If government funded research discovered this pigment, then DuPont can still benefit from the process-- they'd just have to pay royalties to the government (i.e. the taxpayers) whose money gave rise to the invention in the first place.

Which would you rather handle the licensing of the IP for this new pigment? A large federal bureaucracy, which is now on the hook for managing the IP for all government grants, and really doesn't have any strong motivation to market the IP? Or the University which stands to make a lot of money to help fund their research and academic missions? One of those entities is more motivated to get this invention licensed and (hopefully) available to the public.

Again, Bayh-Dole was designed specifically for this purpose -- to move the gatekeepers of government funded research away from the federal government and out to the institutions that had more motivation to push these discoveries to the public faster.

I haven't read the The Entrepreneurial State, but the opening synopsis on Wikipedia supports this logic.

> book written by Mariana Mazzucato which argues that the United States' economic success is a result of public and state funded investments in innovation and technology, rather than a result of the small state, free market doctrine that often receives credit for the country's strong economy. (from the Wikipedia page)

A key driver behind the US economy has been public and state funded research. The mechanism that this research makes it to the broader market (faster) is by moving the licensing away from the federal government and towards the research institutions.

Does the US government get royalties from product sales on an invention that came solely du to its funding sources?

If the answer is no then I'm with the other commenter, all the rest of the points are a sidetrack. You're saying everything about pushing inventions into market faster. Great. However the argument was around "why aren't we getting the money back from direct profit from these inventions?" The taxpayers are the ones that front the money. The taxpayers can be likened to the investors. However it is, in every case, that the tax payers don't get anything out of this. Benefit to society, you might argue. But I'm not feeling as the beneficiary when I pay a shit ton for medicine or a specific item because it's patented, even though NSF or NIH funded the research.

And the other side effect you mention, of protection from competition, I'd argue that is a negative. You already get the headstart on research and the money to do it in the first place. Why do you get so much time to have your product untouchable after the fact? I wouldn't doubt it if this protection is why we see abuse of the medical or automobile industry's prices for things.

If you are a private inventor, using your own money, sure. Give the protections. If you use government grants, I'd say that the government owes it to us to document any invention from it and give it open competition access or get royalties from its sales.

And yes, I'd rather the government handle IP rather than DuPonts research team. Every single time.

Why should I work to pay taxes if I don't benefit from it?

> Why do people get to patent the products of federal grants?

A few reasons:

* The core hypothesis behind publicly-funded research is that simply advancing knowledge is a good thing. Patents (versus public domain, versus assigning a patent to the government) can change who benefits to what degree, but the raw public good is still there.

* Government grants often don't fully fund a research project, or interesting research happens as a byproduct of another 'core' project. That happened here -- undoubtedly the original grant said nothing about finding new dyes.

* Research grants don't fund all the steps to commercialization, they just fund "basic" research. Paradoxically, denying IP rights to funded research might stifle commercialization by making it not worth the effort to turn a lab development into a commercially-viable product. This is similar to the idea that old drugs and traditional remedies are under-studied in part because pharmaceutical companies cannot patent (and thus profit from) associated discoveries.


* Patents expire, eventually, in a more reasonable timeframe than, say, copyrights.


This is the act that makes this all possible and actually encourages this practice.

Patents themselves are federal grants. They are the government giving you a monopoly and enforcing it.

If you don't think people should profit from the governments efforts to promote science, cats already out the bag and run round corner down the alley.

Because federal grants don't require exclusivity.

Presumably because it keeps the profits in the country that funded it?

No it doesn't - the patent can be read by anyone, and if someone in China wants to duplicate your invention they can. This just stops other people who paid the taxes that funded the work from getting it without paying more for a long time.

Reminds me of the story of how Lithium was initially thought help “brain gout” by dissolving uric crystals, but it still ended up working.

Read an interesting story of a physicist that got curious about it. https://www.quantamagazine.org/a-new-spin-on-the-quantum-bra...

NPR had a good article about the discovery: "A Chemist Accidentally Creates A New Blue. Then What?" [1]

[1] https://www.npr.org/2016/07/16/485696248/a-chemist-accidenta...

This blue pigment is used on the shrouds of AMD Radeon Pro graphics cards. https://www.amd.com/en/graphics/workstations

Here's a great article about the discovery of this color, and the pigment industry in general, which hit the front page of HN a year ago:

The Quest for the Next Billion-Dollar Color



YInMn is supposed to have high NIR reflectivity useful on a sunny roof. The Apollo 11 Lunar Module used gold foil to reflect heat, I suppose future vehicles could have a brilliant blue foil, apt for Blur Origin.

I liked how the lunar module used gold foil, and Apollo the Greek god was the Sun deity; gold often used to represent the Sun.

If all you need is IR reflection for a building then TiO2 based paint will do the job and is much cheaper. The advantage of YInMn is that it also provides color.

Under properties one an see the reflectivity and whilst far better than Cobalt blue, its a long way from Gold and silver:


Here's a nice interview with the discoverer: https://news.artnet.com/art-world/chemist-mas-subramanian-on...

Yttrium and Indium are two rare earths, Indium especially, used in displays. Mention has been made that the Indium required makes this pigment expensive.

Apparently it costs $1k/kg right now, if it were even available.

Indium looks to cost $5 a gram, twice the price of silver, although it’s supposed to be as ubundant

> https://www.shepherdcolor.com/products/yinmn/

Looks like it hasn't received approval for commercial sale yet. Presumably some health / safety entity?

That's cheap by pharmaceutical standards and expensive by pigment standards.

About right for exotic engineering applications though.

Is it still true that there are no real inorganic red dyes? Or is that ancient history?

It may depend on the definition of red, but that doesn't sound true at all. Cadmium red, lead tetroxide, cinnabar, all in use for a hundred+ to even thousands of years.

I think where that comes from is that there isn't really an inorganic red that is all of stable, durable, non-toxic, and only then even then good looking.

If I remember correctly, once Mas Subramanian sort of stumbled into "I guess I can try to make pigments with these compounds" he's tried making other colors, but sort of the big prize is in finding a good red with properties similar to YInMn.

>>stable, durable, non-toxic, and only then even then good looking.

Yes, I think that was the combination.

I was thinking in particular red car paint -- it was only organic compounds through the early 2000's, and therefore all of it fades.

> It may depend on the definition of red, [...] Cadmium red, lead tetroxide, cinnabar

For that matter, iron oxide (rust) technically qualifies.

Cinnabar is red...

The article makes no mention of how the pigment makes the blue colour.

It's due to the "trigonal bipyramidal" lattice structure. By changing atoms in the structure, they've created pigments with a range of colors. https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=921441

Excellent find!

This article also talks of the three other variants for different colours

Ooh yeah, that's very interesting. Thanks!

It reflects only blue photons

As mentioned in this YouTube video by Seeker, albeit in passing, but credit where it’s due:


Registration is open for Startup School 2019. Classes start July 22nd.

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