Relativistic particles are heavier, and this actually shrinks the atomic radius of the heavier elements

(written by lawrence krubner, however indented passages are often quotes). You can contact lawrence at: lawrence@krubner.com, or follow me on Twitter.

Interesting:

Some of the properties of mercury (and of copernicium) are due to “lanthanide shielding“, and that is at least understandable in a classical mental-picture way. The lanthanides (and higher elements beyond them) have atoms with smaller radii than you’d predict from just following the trends earlier in the periodic table. But that’s because those atom sizes have to do with the attraction of the outermost electrons to the nucleus (negative and positive charges), and that attraction is partly “shielded” by the inner electrons in the way. That effect diminishes as you lay on more electrons, though: the d and f electron orbitals are progressively less effective at shielding, and the lanthanide elements generally get smaller as you go up. This same effect is responsible, among other things, for making hafnium a lot more like zirconium, one row up, than you’d figure it would be, and separating out really pure hafnium took quite a bit of work.

But the bigger effect is relativistic. That’s actually a notable example of Paul Dirac being completely wrong about something in physics – he had stated back in 1929 (PDF here if you’re up for it!) that relativistic corrections to quantum mechanics were of “no importance” because they would apply only to very high-speed particles (that is, those moving at an appreciable fraction of the speed of light). But as it turns out, the inner electrons of the heavier elements are moving at such speeds (they get faster as the positively charged nucleus gets bigger and more charged), and this has effects out to the chemically important outer electrons as well. For one thing, relativistic particles are heavier, and this actually shrinks the atomic radius of the heavier elements still more and has complex effects on the various orbitals.

In the 1960s and 1970s these effects began to be more appreciated. Mercury’s outermost electrons were believed to be much more involved than they would normally be in interactions with the nucleus, and thus much less involved in attraction with other mercury atoms. But it wasn’t until 2013 that Peter Schwerdtfeger (Massy Univ., New Zealand) and colleagues at other centers were able to nail down that the exact contribution to mercury’s melting point. (I may have mentioned this before, but I’ve long thought that a book titled “Quantum Mechanics: A Hand-Waving Approach” would sell quite well in the textbook market). Without relativistic corrections, mercury’s melting point is predicted to be 82C, rather than -39. (These calculations, direct quantum-mechanical influence on bulk melting point, are extremely painful, which is why it took until the 21st century for hardware and algorithms to be up to the task).

Source