Water molecules are described as an oxygen atom covalently (ignoring for the monent protonization/deprotonization) bonded to two hydrogen atoms with an angle of ~104° between the hydrogen atoms (like Mickey Mouse). I can understand this with multiple water molecules (the stacking would - for want of better descriptive powers - surround the oxygen's relative negative charge.)

Would a single water molecule retain this configuration? It would seem that in space, it should look like CO2?

  • 6
    $\begingroup$ It is probably useful to google for "why does water have a bond angle of 104.5". Once you understand why the 104.5 angle occurs you will also be able to answer your own question $\endgroup$
    – Michiel
    Apr 29 '14 at 5:36
  • 4
    $\begingroup$ The linked Google search mainly finds "Why is the bond angle of water 104.5$^\circ$ and not $109.5^\circ$. If using the VSEPR model, the nonbonding electron domains on water (the lone pairs) produce some "repulsion" contributing to the overall shape. The repulsion is minimize is tetrahedral arrangements. If using quantum mechanics, the shape of water arises from the minimum energy solution to the multielectron wavefunction describing the bonding in water. $\endgroup$
    – Ben Norris
    Apr 29 '14 at 10:39

The HOH bond angle for an isolated water molecule is accurately known to be 104.5° (104.52°±0.05° from Rotation‐Vibration Spectra of Deuterated Water Vapor J. Chem. Phys. 24, 1139).

In the liquid phase, values of 105.5° (calculated) and 106° (experimental) are reported in Structural, electronic, and bonding properties of liquid water from first principles J. Chem. Phys. 111, 3572.

The gas phase water dimer (two otherwise isolated water molecules hydrogen bonded to each other) has also been considered and an angle of 104.7° is reported in the above reference.

For ice Ih (common form of ice), values ranging from 106° to 108° have been reported.

Attempting to explain the trend, the intermolecular hydrogen bonding in the liquid and solid phases weakens the intramolecular covalent O-H bonds, and from a VSPER approach causes the electrons of the O-H bond to be more like the lone pair electrons.


As shown by Uncle Al there is some (old) experimental data on the bond angle of water in this article: Shibata and Bartell J. Chem. Phys. 42, 1147 (1965).

For water in vacuum they find:

For $\ce{H2O}$: r_g (OH) = 0.976±0.0030 Å, mean HOH angle α_g = 107.2°±3°

This means that there is no statistically significant deviation from the value at atmospheric pressure.

  • 1
    $\begingroup$ (-1) for "liquid water in vacuum". Clearly, the reference is about gas phase water, and the reference recognizes that the angle is much more accurately known from spectroscopy. The gas phase water bond angle is 104.52°±0.05° from J. Chem. Phys. 24, 1139 (1956). $\endgroup$
    – DavePhD
    Dec 8 '14 at 18:37
  • $\begingroup$ @DavePhD You are right. Changed it. Anyway, that was besides the point from my perspective (although that's no excuse for the careless error). I only wanted to point out that $107.2\pm3$ is not a statistically significant deviation from $104.5$ $\endgroup$
    – Michiel
    Dec 11 '14 at 10:28
  • $\begingroup$ I agree it's not statisitically significant and commented on Uncle Al's answer too. But does the reference you cite really say what pressure the experiment was done at? $\endgroup$
    – DavePhD
    Dec 11 '14 at 14:46

You want the electron diffraction structure of water in vacuum. DOI:10.1038/248405a0 Liquid water

For $\ce{H2O}$: r_g (OH) = 0.976±0.0030 Å, mean HOH angle α_g = 107.2°±3°
For $\ce{D2O}$: r_g (OD) = 0.970±0.0025 Å, mean DOD angle α_g = 104.2°±3°

It looks like the HOH angle slightly opens in vacuum as opposed to condensed phase. That is 1965. See if you can find a more recent study.

  • $\begingroup$ $107.2\pm3$ includes $104.5$ so I don't see a reason to assume that the HOH angle opens up in vacuum. $\endgroup$
    – Michiel
    Apr 29 '14 at 17:17
  • $\begingroup$ Thank you very much! That answers my question, and paired with the reading, helps me to better understand the significance of it. $\endgroup$ Apr 29 '14 at 18:39
  • $\begingroup$ @Michiel If you were conversant with statistics, the angle opens in vacuo. Find a more recent electron diffraction, ground state microwave, or vibronic spectroscopy structure for tighter angle constraints. To criticize is to volunteer. $\endgroup$
    – Uncle Al
    Apr 30 '14 at 15:26
  • $\begingroup$ @UncleAl I am certainly aware of statistics and you are wrong on this one - if you give $107.2 \pm 3$ that means that you trust the measurement enough to say that it can be any value within the range $104.2-110.2$. It can very well be the case that a newer experiment with a smaller error margin would result in e.g. $105.5\pm 1.5$ without contradicting the current measurement. $\endgroup$
    – Michiel
    Apr 30 '14 at 18:10
  • $\begingroup$ But, instead of arguing, let's have the other users decide. I've written an answer with the same information, but different conclusion $\endgroup$
    – Michiel
    Apr 30 '14 at 18:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.