Imperial College London

Henry Rzepa

Why is this post orphaned from the
[previous](http://www.ch.imperial.ac.uk/rzepa/blog/?p=13458 "How many water molecules does it take to ionise HF and HBr?"){target="_blank"}?
In order to have the opportunity of noting that treating iodine
computationally can be a little different from the procedures used for
F, Cl and Br.

As the nuclear charge increases proceeding down the periodic table, the
inner electron shells start becoming
[relativistic](http://www.ch.imperial.ac.uk/rzepa/blog/?p=12715 "Why is mercury a liquid at room temperatures?"){target="_blank"}.
Iodine is the first halogen where this might really start to matter.^\*^
There are two ways in which one can compute molecules with I; the first
adopts the same procedure as for the earlier halogens, whereby **all**
the electrons are described by basis functions (called an all-electron
basis). This effect does not really include the effects of relativistic
contractions on the inner (1s) shell unless special relativistic
Hamiltonians are also used. The second replaces these inner cores with a
pseudopotential, and this does incorporate some of the relativistic
effects. To find out how much this might matter, I have included both
types:

|     |                   |                                                 |
|-----|-------------------|-------------------------------------------------|
|     | I                 |                                                 |
| n   | I-H               | H-O                                             |
| 1   | 1.637^†^/1.623^‡^ | 2.032/2.060\[cite\]10.14469/ch/190924\[/cite\]  |
| 2   | 1.657/1.641       | 1.863/1.889\[cite\]/10.14469/ch/190921\[/cite\] |
| 3   | 1.696/1.675       | 1.641/1.670\[cite\]10.14469/ch/190925\[/cite\]  |
| 4   |  2.316/2.304      | 1.014/1.015\[cite\]10.14469/ch/190927\[/cite\]  |

^†^Non-relativistic calculation with an all-electron 6-311G(d,p) basis
on I, 6-311++G(2d,2p) on O and H. ^‡^Def2-TZVPPD basis, with
pseudopotential just on I.

As with bromine, iodine shows a precipitous ionisation when the **[4th
water molecule]{style="color: #993366;"}** is added. In the previous
post, I compared this with pKa values, and a
[comment](http://www.ch.imperial.ac.uk/rzepa/blog/?p=13458&cpage=1#comment-162075 "How many water molecules does it take to ionise HF and HBr?"){target="_blank"} posted
there reminded us that a pKa is measured for macroscopic bulk water and
that all sorts of new effects due to free energy/entropy, continuum
solvation and much else will take hold. But qualitatively at least, the
ionisation of HI in a gas-phase cluster of water molecules seems to
match the bulk properties. Relativistic effects do not appear to play a
major role here.

------------------------------------------------------------------------

^\*^Whilst such effects can be prominent for I, arguably they actually
start at Cl *via* an effect called spin-orbit (SO) coupling. This
manifests in the calculation of chemical magnetic shieldings. If one
uses standard GIAO NMR theories, one can calculate shieldings for e.g. C
pretty accurately. But with Cl, the shieldings may be SO-perturbed by
about 3ppm, with Br it's about 12 ppm and with I it reaches 50
ppm\![cite\]10.1021/np0705918\[/cite\]

::: {.wp_orcid_field}
[https://orcid.org/0000-0002-8635-8390](https://orcid.org/0000-0002-8635-8390){target="_blank"
rel="author"}
:::


Why is this post orphaned from the previous? In order to have the opportunity of noting that treating iodine computationally can be a little different from the procedures used for F, Cl and Br.    As the nuclear charge increases proceeding down the periodic table, the inner electron shells start becoming relativistic.

Rogue Scholar Beiträge

How many water molecules does it take to ionise HI?