New Evidence Shows Water Separates into Two Different Liquids at Low Temperatures

Peer-Reviewed Publication


Fresh evidence that water can change from one form of liquid into another, denser liquid, has been uncovered by researchers at the University of Birmingham and Sapienza Università di Roma.

This ‘phase transition’ in water was first proposed 30 years ago in a study by researchers from Boston University. Because the transition has been predicted to occur at supercooled conditions, however, confirming its existence has been a challenge. That’s because at these low temperatures, water really does not want to be a liquid, instead it wants to rapidly become ice. Because of its hidden status, much is still unknown about this liquid-liquid phase transition, unlike about everyday examples of phase transitions in water between a solid or vapour phase and a liquid phase.

This new evidence, published in Nature Physics, represents a significant step forward in confirming the idea of a liquid-liquid phase transition first proposed in 1992. Francesco Sciortino, now a professor at Sapienza Università di Roma, was a member of the original research team at Boston University and is also a co-author of this paper.

The team has used computer simulations to help explain what features distinguish the two liquids at the microscopic level. They found that the water molecules in the high-density liquid form arrangements that are considered to be “topologically complex”, such as a trefoil knot (think of the molecules arranged in such a way that they resemble a pretzel) or a Hopf link (think of two links in a steel chain). The molecules in the high-density liquid are thus said to be entangled

In contrast, the molecules in the low-density liquid mostly form simple rings, and hence the molecules in the low-density liquid are unentangled.

Andreas Neophytou, a PhD student at the University of Birmingham with Dr Dwaipayan Chakrabarti, is lead author on the paper. He says: “This insight has provided us with a completely fresh take on what is now a 30-year old research problem, and will hopefully be just the beginning.”

The researchers used a colloidal model of water in their simulation, and then two widely used molecular models of water. Colloids are particles that can be a thousand times larger than a single water molecule. By virtue of their relatively bigger size, and hence slower movements, colloids are used to observe and understand physical phenomena that also occur at the much smaller atomic and molecular length scales.

Dr Chakrabarti, a co-author, says: “This colloidal model of water provides a magnifying glass into molecular water, and enables us to unravel the secrets of water concerning the tale of two liquids.”

Professor Sciortino says: “In this work, we propose, for the first time, a view of the liquid-liquid phase transition based on network entanglement ideas. I am sure this work will inspire novel theoretical modelling based on topological concepts.”

The team expect that the model they have devised will pave the way for new experiments that will validate the theory and extend the concept of ‘entangled’ liquids to other liquids such as silicon.

Pablo Debenedetti, a professor of chemical and biological engineering at Princeton University in the US and a world-leading expert in this area of research, remarks: “This beautiful computational work uncovers the topological basis underlying the existence of different liquid phases in the same network-forming substance.” He adds: “In so doing, it substantially enriches and deepens our understanding of a phenomenon that abundant experimental and computational evidence increasingly suggests is central to the physics of that most important of liquids: water.”

Christian Micheletti, a professor at International School for Advanced Studies in Trieste, Italy, whose current research interest lies in understanding the impact of entanglement, especially knots and links, on the static, kinetics and functionality of biopolymers, remarks: “With this single paper, Neophytou et al. made several breakthroughs that will be consequential across diverse scientific areas. First, their elegant and experimentally amenable colloidal model for water opens entirely new perspectives for large-scale studies of liquids. Beyond this, they give very strong evidence that phase transitions that may be elusive to traditional analysis of the local structure of liquids are instead readily picked up by tracking the knots and links in the bond network of the liquid. The idea of searching for such intricacies in the somewhat abstract space of pathways running along transient molecular bonds is a very powerful one, and I expect it will be widely adopted to study complex molecular systems.”

Sciortino adds: “Water, one after the other, reveals its secrets! Dream how beautiful it would be if we could look inside the liquid and observe the dancing of the water molecules, the way they flicker, and the way they exchange partners, restructuring the hydrogen bond network.  The realisation of the colloidal model for water we propose can make this dream come true.”

The research was supported by the Royal Society via International Exchanges Award, which enabled the international collaboration between the researchers in the UK and Italy, the EPSRC Centre for Doctoral Training in Topological Design and the Institute of Advanced Studies at the University of Birmingham, and the Italian Ministero Istruzione Università Ricerca – Progetti di Rilevante Interesse Nazionale.


Nature Physics




Computational simulation/modeling


Not applicable



From EurekAlert!

4.1 9 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
August 19, 2022 2:36 am

“The team has used computer simulations to help explain what features distinguish the two liquids at the microscopic level.:

My reading stopped after that sentence.

Here’s my “water analysis”
I put a can of warm diet pop in the freezer for a quick cool down
I forget about the can
It later explodes inside the freezer
The wife goes berserk
Experiment repeated about once a year.
Same results.
No computer simulations required,

Last edited 1 month ago by Richard Greene
Reply to  Richard Greene
August 19, 2022 3:39 am

i usually do that with beer, same result, is there a message here

william Johnston
Reply to  Ian
August 19, 2022 4:00 am

They are making containers more fragile/ In my youth, we would put unconsumed cans of barley pop near certain fence post corners. Retrieved in spring, they were still consumable. Of course, back then, our standards were a bit lower!

william Johnston
Reply to  william Johnston
August 19, 2022 6:09 am

I neglected to mention, this was done during winter in northern Wisconsin!

Reply to  william Johnston
August 19, 2022 7:07 am

True. Only minimal aluminum is used in a can now. The real strength of the can is in the interior & exterior coating. The aluminum serves only as the substrate for the coating.

Bryan A
Reply to  Ian
August 19, 2022 5:15 am

Yep, the home freezer temperature is sufficient to stop the vibrational motion of many consumable liquids, even those containing CO2. However, the liquid cannot retain the dissolved CO2 after freezing which then builds up pressure in the containment vessel until catastrophic failure

Gunga Din
Reply to  Bryan A
August 19, 2022 7:36 am

Well, actually that’s not why the container might rupture.
Water can dissolve and hold more of a gas (like CO2) the colder it gets.
The container ruptures because when water becomes a solid (ice), unlike most solids, it becomes less dense and expands.
In other words, the same weight of solid water occupies a greater volume than the same weight of liquid water.

Pat Frank
Reply to  Gunga Din
August 19, 2022 5:51 pm

But in freezing, the water zone-refines out the CO2, which becomes gas-phase. The pressure increase is probably large.

Jim G.
Reply to  Pat Frank
August 20, 2022 8:44 am

Gasses compress easily.
Solids…Not so much.

Gunga Din
Reply to  Pat Frank
August 20, 2022 2:04 pm

But the water turning to ice would make for less pressure as the volume increases.
Wouldn’t it?
(I’d treated water for 38 years to make it safe to drink, the first 6 years included wastewater treatment making it safe to discharge into the receiving waters. But I didn’t treat ice.)

Pat Frank
Reply to  Gunga Din
August 20, 2022 10:56 pm

if the can is brim-full (zero head space), freezing the drink will cause the water to increase in volume by about 10%. The pressure of expansion will be exerted against the inner surface of the can.

The CO2 that is forced out of solution will cause an increase in pressure as well.

Apparently a 12 oz can of coke contains about 2.2 gm of dissolved CO2. Forced out of solution, that amount of CO2 will occupy about 1.12 liters at standard temp and pressure (STP).

A 12 oz coke can coke can is about 355 mL, so the excluded CO2 gas will exert about 3 atmospheres of pressure over and above what the frozen water exerts.

Altogether it may be enough to burst an aluminum can — almost certainly a plastic bottle.

Reply to  Pat Frank
August 22, 2022 5:24 pm

the solubility of CO2 in water increases as temperatures decreases, once the water is frozen it’s trapped.
if the can ruptures the pressure decreases to atmosphere and the solidus of the contents might be different, this may cause melting and spraying.

Reply to  Bryan A
August 19, 2022 5:12 pm

That type of failure used to be “catastrophic”. It is now “apocalyptic”.

Reply to  Ian
August 19, 2022 5:29 am



Does your wife go berserk too, or are you single?

I try to clean out the freezer before the wife finds out,
but that is a difficult job — the exploded pop is frozen,
and she always catches me. Every time. I claim to be
looking for ice cubes, but can’t explain why I have
removed every item from the freezer to find them.
This has been a running joke since 1977.
She pays me back by backing her car into a post,
or something similar.

william Johnston
Reply to  Richard Greene
August 19, 2022 6:11 am

Never a dull moment, is there!

Reply to  Richard Greene
August 19, 2022 7:18 am

Well, Richard, if Ian did that often enough, he’s probably single now.

Reply to  Richard Greene
August 19, 2022 7:52 am

Just because climate models are bad and misused, is not evidence that all models are worthless.

Reply to  MarkW
August 19, 2022 9:39 am

For climate computer games to be useful:

Very detailed knowledge of every climate change variable is required
That knowledge does not exist today.

The climate must be predictable. Predictions may not be possible.

The people programming the climate computer games must want to make accurate predictions. They currently want to make scary predictions, not accurate predictions

With the current knowledge of climate change, any model that appears to make an accurate climate prediction is likely to be a lucky guess.

I prefer my models to be tall, female, and strutting on runways.

Last edited 1 month ago by Richard Greene
Reply to  Richard Greene
August 19, 2022 11:08 am

“With the current knowledge of climate change, any model that appears to make an accurate climate prediction is likely to be a lucky guess.”

I thought the Russian model, a major outlier to the consensus models, was pretty much close to matching the “temperature anomaly”.

So is there actually one model that is “tuned” correctly, if not actually truly “modeling” climate change?

AGW is Not Science
Reply to  Drake
August 19, 2022 2:22 pm

The reason the Russian model is the most realistic is because it is the one model that assigns little effect to…


Reply to  AGW is Not Science
August 19, 2022 4:30 pm

We can’t have that.
Sanctions on the Russian model !

Reply to  Drake
August 19, 2022 4:30 pm

The Russian INM model is less inaccurate than all others. It gets almost no attention. That tells you a lot.
Accurate predictions are not a goal. If they were, the Russian model would get 99% of the attention.

The Russian model can also appear somewhat accurate by chance. There is not enough knowledge of every cause of climate change for it to be accurate only because of its design. At least it is in the ballpark of reality. But one could do that by extrapolating +0.1 degree C. global warming per decade on the back of an envelope.

Or using my 100% accurate “model” from 1997:
“The climate will get warmer, unless it gets colder”.

Reply to  Richard Greene
August 19, 2022 11:21 am

It also helps to know the initial conditions across the entire globe, at the same time, to a high level of accuracy.

Reply to  Retired_Engineer_Jim
August 19, 2022 4:32 pm

Knowing initial conditions and having a very precise understanding of all the causes of climate change … and then it is still possible that climate is unpredictable. Caused by random variations of a complex system.

Reply to  MarkW
August 19, 2022 11:19 am

“All models are wrong. Some models are useful.”

Can’t remember who said that.

I was employed building, and using, models.

Reply to  Retired_Engineer_Jim
August 19, 2022 4:34 pm

“All models are wrong, but some are useful” is a famous quote often attributed to the British statistician George E. P. Box. Thus, the idea of this quote is that every single model will be wrong, meaning that it will never represent the exact real behaviour. …

Pat Frank
Reply to  Retired_Engineer_Jim
August 19, 2022 5:56 pm

And your models were parameterized to accurately reproduce observables within the specified calibration bounds.

Extrapolation outside those bounds was done with caution, even at small scale. Large scale extrapolation, and no engineer would trust the result.

Climate models are engineering models. And yet, large scale extrapolation is what climate modelers do with confidence. They’re not scientists and certainly not engineers.

Gary Pearse
Reply to  Richard Greene
August 19, 2022 10:00 am

Years ago, I went to the family summer cottage in winter and found a 24 of beer with several bottles in it, all liquid. When opened one the contents solidified in about 10 second. I put the rest to thaw by the fireplace before opening.

Reply to  Gary Pearse
August 19, 2022 11:12 am

I have found that super cooled beer or soda in the freezer, or water left in the car overnight in cold weather, turns solid immediately upon the release of pressure by opening the bottle.

10 seconds to solidify must have been close to the freezing point.

Reply to  Richard Greene
August 19, 2022 4:05 pm

I managed to do that with a bottle of champagne . The frozen champagne had the consistency of SNOW. I was longing to sample it, but since it also contained glass shards I thought better. Champagne snow could make for a unique dessert.

I tried to reproduce the results by pouring champagne into a plastic pump sprayer, pumping up the pressure, and freezing it, then quickly releasing the pressure. Didn’t work. I think once you pop the cork on the bottle that’s it. You can’t repressurize it to the same degree.

Reply to  Richard Greene
August 19, 2022 5:51 pm

So you don’t drive a car, cross a bridge, or go near a large building? Or use modern technology?

Models are fundamental to all engineering.

The critical points you miss are to what level the model is understood, checked and verified.

Reply to  Dean
August 20, 2022 2:31 pm

Big difference between using FEA (finite element analysis) to model a bridge. You know the inputs are valid. Almost nothing in climate models is known.

August 19, 2022 2:41 am

Hmm. You have to read that carefully because—not by accident—it gives you a wrong idea about what they actually achieved.

It is a press-release. It is intended to excite.

(Fair enough; everyone wants clicks and grants. Just be a careful reader.)

michael hart
Reply to  Quelgeek
August 19, 2022 7:05 am

Yup. Just add it to the 11 (or is it 12? I forget) phase changes of pure water.

Be careful twice, reader. I said pure water. Once it starts interacting with other molecules it becomes exponentially more complex. That’s water for you.

Reply to  Quelgeek
August 19, 2022 8:56 am

Astronomers say that there are huge massive “blobs” of water in some regions of space – what form(s) of water?

Pat Frank
Reply to  Antigriff
August 19, 2022 5:58 pm

They’re very large clouds of very rarified gaseous water molecules.

August 19, 2022 2:44 am

With the correct model I change the liquid milk into the liquid wine.

Reply to  Krishna Gans
August 19, 2022 4:23 am

That might require a novel statistical approach.

Reply to  MarkH
August 19, 2022 11:14 am

So just call Mann, he will make one up for you, BUT it will be statistically invalid, like his Hookey stick.

Reply to  Krishna Gans
August 19, 2022 5:31 am

The only good models.
are those tall female models

Paul C
Reply to  Richard Greene
August 20, 2022 11:18 am

Short female models can be good too,
but the most interesting models are the bad female models 🙂

David E
Reply to  Paul C
August 20, 2022 1:46 pm

Hahaha – I concur, but think we need to conduct more experiments to map the boundaries 🤣

John Endicott
Reply to  David E
August 22, 2022 3:43 am

cuation: Mapping the boundaries may lead to getting slapped in the face 😉

AGW is Not Science
Reply to  Krishna Gans
August 19, 2022 2:31 pm

Reminds me of a sign at a local ice cream stand – “Patience. In time, the grass becomes ice cream.”

Jay Willis
August 19, 2022 2:47 am

I really find it annoying that the title misrepresents the work. There is no new evidence presented here. It’s speculation. That’s not a bad thing, its interesting and beautiful, but it’s still speculation. New evidence will only come from observation of reality. Otherwise, great stuff, thanks for posting.

Reply to  Jay Willis
August 19, 2022 8:35 am

Speculation is the basis of thought.
I might speculate that “entangled” water has a higher viscosity than “unentangled,” and thus might exhibit different microfiltration rates.
Another might speculate a way to use osmotic pressure or nano-filtration to separate the two liquids, thus concentrating two levels of entropy from each other.
Even more, perhaps others may figure out how to make money from this.
Speculate away!

August 19, 2022 2:49 am

Very interesting post/article! May be somewhat related to something I have pondered for a few years with regards to the make-up and formation of the cloud base as a result of the condensation of the H2O molecules to their formation/bonding into a visible cluster of ‘fog’. There is a series of entanglement and unentanglement as the molecules coalesce and then get hit by upwelling LWIR causing a partial re-evaporation of at least some of the H2O molecules from this cluster. Is this process that is occurring at the critical stage of the phase change something that would be considered an entanglement and unentanglement process? Inquiring minds need to know!

Michael ElliottMichael Elliott
Reply to  eyesonu
August 19, 2022 3:06 am

This line of research will be good for many years of government grants.

It’s one way to make a living, bur what for ?

Yes water is a interesting substance, but let’s deal with it at normal temperatures.

Michael VK5ELL

Reply to  eyesonu
August 19, 2022 4:42 am

Fly into a cloud that is “supercooled’ and watch the ice immediately form on your windscreen and who knows what else. Maybe this has something to do with that.

Bob B.
August 19, 2022 3:02 am

I find that Dense water is very hard to swallow.

Last edited 1 month ago by Bob B.
Reply to  Bob B.
August 19, 2022 5:31 am

I thought they might be able to turn water into whine, but it’s still all wet.

August 19, 2022 3:06 am

With no valid experimental evidence that the ‘entangled’ denser state actually exists, this paper has no merit. It is yet another computer model based on the flawed assumptions about molecular bonding, which is poorly understood.

Reply to  Tenuc
August 19, 2022 3:30 am

It is best to think of this as an hypothesis to be tested later. It is theoretical, and the title should have said something like “Hypothesis for dense water phase structure proposed.”

Like many things in science these days, the paper is being overhyped. Theoretical work does sometimes lead to some interesting discoveries, but many times lead nowhere. I think it is because the people in the university press office know absolutely nothing about science – especially if they have degrees in “science communications”.

Reply to  OweninGA
August 19, 2022 7:27 am

Right you are, Owen.

From the article:

The team expect that the model they have devised will pave the way for new experiments that will validate the theory and extend the concept of ‘entangled’ liquids to other liquids such as silicon.

They appear to be using the model to be validated by experiment at some later point, so kudos to them. That’s the way to do it.

At least I am hoping that is what they really meant. If it paves the way for “new experiments” with yet more models, then I withdraw my kudos.

Rich Davis
Reply to  H.R.
August 19, 2022 4:19 pm

With a melting point of 1410C, I’m sure that once again the political science interns who write up the YouReekAlot! press releases did not mean silicon. They have once again confused silicon with silicone.

Tim Gorman
Reply to  OweninGA
August 19, 2022 11:23 am

In working with Bell Labs folks many years back it was very interesting how many “theoretical” proposals turned out to be meaningless. You had to be able to demonstrate how things worked to get continued financing, no such things as computer models.

Smart Rock
Reply to  Tenuc
August 19, 2022 11:00 am

The article is a typically overstated description of the paper by a so-called journalist who doesn’t understand the subject. Putting “evidence” in the title is totally misleading and violates a thing called “journalistic ethics”, which used to exist back in middle part of the 20th century (some of you may remember it fondly).

The actual paper, entitled “Topological nature of the liquid–liquid phase transition in tetrahedral liquids” proposes a new hypothesis about the structure of water, which is a complex subject that’s been kicked around for over a century. Water is a critical component of our earthly environment, crucially essential for life, and water’s behaviour is a fit subject for study.

Their hypothesis is capable of being tested experimentally, which it probably will be.

I’m baffled by why these news-release writers, who know their work is going to be picked up by on-line rags like EurekAlert, aren’t required to have their output edited by the actual authors.

Simon Derricutt
August 19, 2022 3:29 am

A nicely-chosen illustration, referring to “Cat’s Cradle” and thus ice-9 too. Nice one, Charles the Moderator!

However, the existence of this new form of water is only in computer models. All sorts of things can be proposed using a model that may or may not represent reality close-enough to exist in real life. A model is after all a hypothesis, and always needs verification.

Maybe more worthwhile looking at Pollack’s work for real measurements, which are weird enough anyway.

August 19, 2022 3:31 am

“I am sure this work will inspire novel theoretical modelling”

We’ve had quite enough of that, thank you.

Derek Wood
August 19, 2022 3:52 am

Great Header! “Cat’s Cradle” Kurt Vonnegut Jr, a terrific SF novel from a great author, proposing the existence of ice nine.

Captain climate
August 19, 2022 3:53 am

“Scientists form hypothesis with computer model.”

Loren C. Wilson
August 19, 2022 4:54 am

We have a new model, completely unverified, that models what they are looking for. No experimental data to support. Does this sound familiar to anyone? I am being a bit harsh here since the modelled behavior of water may actually exist. Now someone needs to propose and conduct some measurements to see if this really happens.

Ed Zuiderwijk
August 19, 2022 5:00 am

And the actual observations are published where?

Richard Page
Reply to  Ed Zuiderwijk
August 19, 2022 5:39 am

What actual observations? All they have is what they refer to as ‘computational evidence’.

Reply to  Ed Zuiderwijk
August 19, 2022 7:37 am

Ed the article said (my bold):

“The team expect that the model they have devised will pave the way for new experiments that will validate the theory and extend the concept of ‘entangled’ liquids to other liquids such as silicon.”

Dare we hope for some real science?

The article didn’t say anything about the experiments. They just may be more models, in which case phllbbbbt!!!

. But I’m cautiously optimistic and in ‘wait and see’ mode. Perhaps we’ll get an update in a year or so.

Maybe science isn’t completely dead and is still twitching a little. I dunno.

Richard Page
Reply to  H.R.
August 19, 2022 8:24 am

Now that’s optimism on a massive scale!

Mark BLR
Reply to  Ed Zuiderwijk
August 19, 2022 8:29 am

the actual observations

The paper includes :

Recent numerical studies [8,9], as well as experiments [10–12] (despite the difficulties induced by the rapid formation of ice around the predicted critical temperature and pressure conditions), have provided strong support to this fascinating hypothesis.

The most recent of those references is :
“10. Kim, K. H. et al. Experimental observation of the liquid-liquid transition in bulk supercooled water under pressure. Science 370, 978–982 (2020).”

The Abstract of the Kim et al (2020) paper [ plus a direct link to the Science website for people with “access rights” … ] :

We prepared bulk samples of supercooled liquid water under pressure by isochoric heating of high-density amorphous ice to temperatures of 205 ± 10 kelvin, using an infrared femtosecond laser. Because the sample density is preserved during the ultrafast heating, we could estimate an initial internal pressure of 2.5 to 3.5 kilobar in the high-density liquid phase. After heating, the sample expanded rapidly, and we captured the resulting decompression process with femtosecond x-ray laser pulses at different pump-probe delay times. A discontinuous structural change occurred in which low-density liquid domains appeared and grew on time scales between 20 nanoseconds to 3 microseconds, whereas crystallization occurs on time scales of 3 to 50 microseconds. The dynamics of the two processes being separated by more than one order of magnitude provides support for a liquid-liquid transition in bulk supercooled water.

Sounds … errrrrmmmm … “interesting”, but also “above my pay-grade” (intellectually speaking), as they say.

That Science article is paywalled (for me at least), so it looks like Kim et al saw evidence of some sort of “liquid-liquid phase change”, while the new paper is trying to come up with a theoretical model it can compare the Kim et al data to (as well as data from other, non-cited papers ???), while asking for some new experiments better “tuned” to checking out their particular model to be devised and run by other (experimental, solid-state) physicists … maybe …

Pete Bonk
August 19, 2022 5:35 am

What is this paper really suggesting? At the local level water a few water molecules might organize as knots and chains. Fine, but so what? At best these knots and chains- are they more or less closed, or can they form an extended 2D (ish) or 3D structure?- might be form minute imperfections in garden variety ice. Would this new (and improved?) ice have a higher melting point than 0C? The Greenland Ice Cap, made of “knotty”ice, melting even sooner than std ice due to global warming! Oh, the humanity!

Geoff Sherrington
August 19, 2022 5:54 am

“other liquids such as silicon”???
Maybe silicone???
Geoff S

It doesn't add up...
Reply to  Geoff Sherrington
August 19, 2022 6:44 am

Silicon does melt if you heat it enough

Joao Martins
August 19, 2022 6:19 am

No catastrophe: that kind of less liquid water can (sic, in the first sentence) only occur when supercooling; that will not happen because Climate Change.

Gunga Din
August 19, 2022 6:25 am

Sounds like frazil ice.
That occurs when the water is 32*F or below but turbulence hinders the hydrogen bonds from forming and/or the tiny ice crystals form but the turbulence keeps them from rising to the surface.
When the water contacts a cold object below the waterline it can begin to coat the object with ice. That can be problem at water intakes.

Gunga Din
Reply to  Gunga Din
August 19, 2022 7:45 am

PS Water is most dense at about 39*F (4*C). That’s the H2O molecules begin to “unpack” and line up with each other.

It doesn't add up...
August 19, 2022 7:01 am

So now the race is on to design apparatus that can create 2,000bar at 188K and allow observation of the molecular structure.

August 19, 2022 7:04 am

“It’s only a model.”🏰

dodgy geezer
August 19, 2022 7:14 am

At STP water molecules are free-flowing. At 0C they are arranged in a crystalline lattice.

At some point close to 0C the molecules will start to migrate towards their ordered structure. Inside this range you probably have connections between molecules developing and breaking as the system aligns itself.

This research seems to be examining the actual progress of this process. Interesting – but we are not really learning anything new. We knew that something like this must be happening.

There seems to be a lot of hyper-excitement and exaggeration in modern science papers…

Chris Hall
August 19, 2022 7:23 am

The anomalous behavior of water has been a subject of study for over a century. A truly amazing site that outlines all of the history and science connected with this is Martin Chaplin’s Structure of Water web site (it only has a few thousand references):

The idea that liquid water comes in two “flavors”, a structured low-density, ice-like form and a high density compressed form has been around for several decades. These two forms have been postulated to come in small clusters consisting of a few hundred molecules. What the study mentioned in this post does is use molecular dynamics modeling that bolsters the idea that these two types of clusters do indeed exist at room temperature. The ice-like clusters are favored from the point of view of enthalpy, but the high density clusters are favored because of entropy. The cluster hypothesis is especially successful at explaining the anomalous difference between the heat of fusion and vaporization, as well as the anomalous temperature dependence of liquid water density. You can toss in about another hundred or so anomalous features of water that has perplexed researchers for a very long time.

I recently published a paper that uses the cluster hypothesis to explain anomalous noble gas solubility in cloud water (aka fog). We’d previously measured the solubility of the stable noble gases in snow/ice, and the temperature dependence of their solubilities is well known. The big surprise was that fog behaves like liquid water that has a thin shell of ice-like water, and the effect is anything but subtle. We suggested that the cluster hypothesis combined with surface tension and Archimedes Principle (!!) can explain the results. The doi of the paper is : 10.1029/2020WR029306

It’s nice that MD simulations are getting to the point where the cluster idea can be tested at room temperatures.

Gunga Din
Reply to  Chris Hall
August 19, 2022 8:44 am

Chris, I clicked on the link at the beginning of your comment and got a 404 message, site not found.

Chris Hall
Reply to  Gunga Din
August 19, 2022 9:44 am

Sorry, it got moved. Here’s the new site:

It’s not https, so your browser might complain.

Reply to  Chris Hall
August 19, 2022 9:46 am


I searched the doi of the paper but it is paywalled. If you can offer free access, now is the time and place for exposure. There are more eyes here at WUWT than anywhere else.

Chris Hall
Reply to  eyesonu
August 19, 2022 10:44 am

WRR has the copyright, so I hesitate to upload the paper to researchgate or some other public site. However, I can answer requests for reprints. I’m at cmhall at umich dot edu.

August 19, 2022 8:00 am

Yeah, yeah; but what about Kurt Vonnegut’s ‘Ice Nine’?

Andy Pattullo
August 19, 2022 8:48 am

As many others have commented this is not an objective finding, it is an hypothesis imbedded in a computer model. The quality of the output depends more on the programming, the assumptions and the adjustments made in the model as it does on the laws of physics in the real world. Maybe they are right, maybe not, but they have no right to claim a new discover other than some interesting math and geometry in their model.

August 19, 2022 8:53 am

Got it H2O & OH2.

Gordon A. Dressler
August 19, 2022 9:20 am

Polywater is back . . . courtesy of computer simulations!

Reply to  Gordon A. Dressler
August 21, 2022 9:17 pm

Yeah, this is the first thing I thought of too. I remember back in the 1960s it was said that the polywater meme was started by a Russian scientist with the intent to sidetrack U.S. scientist into fruitless experimentation. Since then, polywater has been classified as an example of a pathological science – kind of like CAGW is.

Gary Pearse
August 19, 2022 9:29 am

I was told over 60yrs ago that water was liquid far below its freezing point in very fine clays. I thought it was because the fine clays impeded the joining of the molecular ends into a solid lattice. This was probably an engineering discovery that perhaps didn’t get published.

Gary Pearse
Reply to  Gary Pearse
August 19, 2022 9:52 am

Here is a medical research study by engineers that found a way to maintain liquid water well below freezing.

Methinks physicists today with little new science to add may be trolling the internet for stuff to “discover”. Google water below freezing in fine clays” for a whole bunch of these. There is even one about subsurface water on Mars. Seems the phenomenon is rather commonplace. Interesting to see if the paper has a big bibliography!

Matt G
August 19, 2022 11:00 am

Supercooled water that requires to be pure with no impurites has been well known for decades. It remains well below freezing point of water but requires a trigger to suddenly change it to ice.

Schrodinger's Cat
August 19, 2022 11:02 am

There is a phase separation process known as coacervation. Dilute solutions of oppositely charged colloids can separate into two phases, one colloid rich and the other colloid depleted. If you think of a colloid as mid way between a solution and a solid you can get the idea. This is not just a laboratory curiosity, it is the basis of a microencapsulation process used industrially to encapsulate oil emulsions in batches of several thousand gallons at a time.

I’m not saying this two phase water claim is true, just mentioning the colloid system. It is possible that there is a pre-frozen state when water molecules have aggregated into colloidal size particles. Aggregates of these aggregates could probably be centrifuged into a separate “denser” phase with the upper layer composed of normal free water molecules.

They are probably getting excited about what the rest of us call slush.

Walter Sobchak
August 19, 2022 1:46 pm

“METHOD OF RESEARCH: Computational simulation/modeling”

I.e. they played video games and here are the results.

Not science. Not close.

Walter Sobchak
August 19, 2022 1:50 pm

Ice Nine! We’re all going to die.

Classical reference explained:

“Cat’s Cradle” by Kurt Vonnegut

RTWT | Spoilers

Last edited 1 month ago by Walter Sobchak
Walter Sobchak
Reply to  Walter Sobchak
August 19, 2022 10:40 pm

I added the links. Herr Rotter was clever to add the picture above. But, the book was written in 1963, and I do not know if it is much read anymore.

Elliot W
August 19, 2022 2:59 pm

If they wanted funding to pursue this further, they should have know they needed to link it with global warming. No climate change, no $$$.

August 19, 2022 3:47 pm

Sounds like they are using the same computer models as the Global Warming XXXXX XXXXXXX Climate Change XXXXX XXXXXX Climate Disruption people.

Pat Frank
August 19, 2022 5:49 pm

I have a little familiarity with molecular dynamics. The models are heavily parameterized. Small changes in the parameters governing electrostatic interactions between water molecules (Lennard-Jones potentials) can strongly influence the simulated behavior of water molecules.

In 2016, my colleagues and I published the first experimental determination of the solvation structure of chloride, bromide, and iodide ions in water.

In doing so, we falsified an asymmetric molecular dynamics solvation model that had become very popular and cutting edge. It was wrong.

All-in-all, I take MD simulations with a grain of salt. At best, they may give one an idea of where to look. But experimental work tells the tale.

Such model work, absent experiment, is hardly more than philosophy. The computer graphics are very pretty and are made to look convincing; especially the videos. But the simulations are speculative, not indicative.

The problem is that the pertinent interactions are below the energy resolution limits of the theoretical mechanics. The physical theory (quantum mechanics) will need serious serious and detailed improvement before it can command belief at the level of solution interactions.

Chris Hall
Reply to  Pat Frank
August 20, 2022 6:35 am

Looking at your abstract (the link needs to be edited to work), you may have stumbled onto an effect of cluster theory. Solubility may be controlled by spaces between clusters and it’s not surprising that this might be missed in an MD experiment, as a single cluster might consist of dozens to hundreds of water molecules. In our noble gas paper (definitely non polar and symmetrical), we found that fog had ice-like solubility, with enhanced He solubility (like ice), Ne solubility nearly the same as liquid water (like ice) and depleted concentrations of Ar, Kr and Xe (like ice, with Xe the most depleted). He and Ne can be accommodated within ice structures, but the three heavies cannot. Now a “crushed” or “entangled” high density cluster should not be able to handle having any noble gas atoms within the cluster. So most solubility must occur in spaces between clusters and for Ar, Kr and Xe, ALL solubility must occur between clusters. It’s just a guess at this point, but it’s a 10 to 20 sigma result that can’t be explained by a traditional theory of liquids.

Pat Frank
Reply to  Chris Hall
August 20, 2022 7:42 am

Thanks for the comment, Chris. One could do an XAS experiment — at the Ar edge, for example — and use MXAN to derive an empirical model of the solvation environment.

Maurizio Benfatto is the MXAN expert and may be interested in the project.

If you’re interested, feel free to contact him and mention my name. Maurizio is a great guy and a brilliant condensed matter physicist.

All three halides had discrete local solvation structures. Chloride was unique and was particularly challenging.

Figure 11 shows the test of the MD-predicted solvation structure of chloride.

I’ll have to look at your paper to understand your description. 🙂

Chris Hall
Reply to  Pat Frank
August 20, 2022 10:09 am

I’m retired now, so unfortunately I can’t do any experimental work. I was hoping to learn how to do some MD simulations, but the crypto-currency guys made high powered graphics cards too expensive!

Pat Frank
Reply to  Chris Hall
August 20, 2022 7:46 am

Found it: Anomalous Noble Gas Solubility in Liquid Cloud Water: Possible Implications for Noble Gas Temperatures and Cloud Physics

les online
August 19, 2022 8:31 pm

Only this morning i read of the book “How Superstition Won and Science Lost. Popularising science & health in the United States.” by John Burnham… Has anyone any thoughts or advice, as it’s quite pricey for a pensioner ?

John Hultquist
August 19, 2022 9:06 pm

 The “Higgs boson” or Higgs particle is an elementary particle in the Standard Model.
Oh, did I just use the term “model”? OK, don’t think of that.

Proposed in 1964, after a 40 year search, a subatomic particle with the expected properties was discovered in 2012by experiments at the Large Hadron Collider (LHC) at CERN near Geneva, Switzerland.  

August 19, 2022 9:37 pm

So a number of interactions can take place by attraction between electrons in outer orbitals of this v-shaped molecule as it changes from water to ice. So minor we’ve hardly noticed even though water is the most studied fluid ever. Nobody is going to care about the pretty knot pics in this waste-of-research funds grant application.

Chris Hall
Reply to  DMacKenzie
August 20, 2022 6:21 am

It’s interesting that you’re not interested in the answer to the question of why ice floats (very anomalous), or that the bottom of the ocean and most northern lakes are all at nearly the same temperature, or details of steam thermodynamics, or details of water surface tension, etc., etc. The list of anomalous features of the most important and probably least understood liquid on the planet are nearly endless.

August 20, 2022 6:31 am

I missed something. The article never says at what temperature this unusual type of water happens.

How do use a computer model to show effects of cooling unless you program it to do those effects?

Chris Hall
Reply to  mkelly
August 20, 2022 10:04 am

It happens at all temperatures when water is in liquid form. It seems that liquid water always has this “Janus” personality.

Reply to  Chris Hall
August 20, 2022 2:58 pm

The article says “… occur at supercooled conditions…”. It just isn’t specific.

Chris Hall
Reply to  mkelly
August 21, 2022 2:13 pm

Ahh, but the hypothesis that they are addressing needs these two types of liquid water structures also must occur at normal liquid water temperatures. However, most attempts to verify this via methods like X-ray diffraction have used super-cooled water because the effects are more easy to document at those temperatures..

Allen Stoner
August 20, 2022 7:46 pm

I was looking for something solid to have them hang their hats on, and there is nothing in this article worth reporting at a high quality place such as this.

%d bloggers like this: