Like most planets, Uranus has a very cold outer atmosphere and a very hot core. What we see is a very thick primary atmosphere with plenty of hydrogen. Deeper in, we might suppose that water condenses in clouds at maybe 200 atm of pressure. (Above that it would be ice crystals)

By "might suppose", I've used the dodgiest source conceivable to plot the atmospheric conditions on a phase diagram of water, shamelessly thefted from Duke here, whereby I mean I took a Wikipedia contributor's dotted blue line, "an extrapolation of the 1987 Lindal et al. data (solid blue) using a constant lapse rate of 0.82 K/km."

enter image description here

Given this impeccable data (or other data you might contribute!), it appears that the extrapolated atmospheric conditions of Uranus take a shot straight down the fairway toward a layer of liquid water. IF it's there. The usual treatment of the topic says that there is an atmosphere over a layer of ices ... without many specifics, since after all nobody can see it.

But do we have any indication at all? Is there a reason to say probably yes or probably no to the question of whether a probe dropped into Uranus would reach 374 °C (the critical temperature of water) before it would reach a liquid (probably, in that case) surface of water?

(Note: I recall seeing some simulation of the overall temperature issue before, but with a range of outcomes, and this question is prompted by this report that MgO dissolves extensively into H2O in Uranus, potentially altering the internal heat distribution.

  • 2
    $\begingroup$ Interesting question, welcome to Astronomy SE! I've adjusted your title to better match your question since some users just read titles and then start answering. Feel free to edit further. $\endgroup$
    – uhoh
    May 19, 2021 at 21:47
  • $\begingroup$ You might want to check 1.) whether chemical equilibrium at those temperatures allows water to exist (oxygen might be soaked up by other molecules) and 2.) whether water is soluble in H/He mixtures at those temperatures ( see ui.adsabs.harvard.edu/abs/2012ApJ...745...54W/abstract). $\endgroup$ May 20, 2021 at 19:54
  • $\begingroup$ A well-posed question! I tried to see whether there is some literature on the topic, but I only found a 2010 called Interior Models of Uranus and Neptune going somewhat in the correct direction, maybe that's helpful for you. $\endgroup$
    – B--rian
    May 20, 2021 at 21:17
  • $\begingroup$ @AtmosphericPrisonEscape Concerning your 2nd point about ice-mixtures of different substances: More than a decade ago, I attended a physics collquium talk by some specialist on strange ice phases (in the shown phase diagram, the red numbers are refering to different water ice phases). Sadly, I do not recall the name of the speaker, but I remember that he mentioned the application of his research for ice planets. Maybe I can figure out more about the colloquium... $\endgroup$
    – B--rian
    May 20, 2021 at 21:22
  • 1
    $\begingroup$ @B--rian: The works by Helled, Vazan and others on those topics continues until today, so you should easily find updates. With that 2010 articles you've posted when comparing their central densities and temperatures with those $\rho$ and T required for liquid water in the Wilson&Militzer article, we see that the central values barely scratch that area of phase space. So the answer to OP's question is 'probably no, even if Oxygen would remain in $H_2 O$' $\endgroup$ May 20, 2021 at 22:11

2 Answers 2


Likely your conclusion is true.

It's hard to look into the interiors and one only has data from some form of seismic, thermal and rotational measurements as well as lab data from dynamic measurments on the phase behaviour of the materials under high pressure and temperature as found in the interior (which is a quasi-static state).

However simulations like this from Gao et al or Naumova et al (and references therein) seem to suggest that highly silica- or generally ion-enriched water is found in Uranus' and Neptune's interior.


I would think that the solubility of the various gases present would be critical to any phase diagram. To what extent turbulence will allow any phase separation may be pertinent. Often, actually getting ones hands dirty in a lab gives insights.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .