Partial charges in a covalent bond

Question

My professor stated the following statement:

In a covalent bond, there is a presence of a partial charge in the atoms that combine to form a compound. For example in $\ce{H2O}$, since $\ce{O}$ is more electronegative, therefore there is a presence of a partial negative charge around $\ce{O}$. Similarly, as $\ce{H}$ is less electronegative, it has a presence of a partial positive charge around it. These partial charges are very feeble.

I found this statement to be little incomplete. So, I decided to find the reason of this partial charge. One of the video that talked about partial charges stated this:

Since $\ce{O}$ is more electronegative than $\ce{H}$, therefore the shared electrons tend to spend more time around $\ce{O}$ than $\ce{H}$ and this is the reason of the partial charges.

Now, my questions are:

1. If the shared electrons spend more time around $\ce{O}$ then it means that for some time $\ce{O}$ will be negatively charged and for some time $\ce{H}$. If this holds true, then why do we say that there is presence of partial charges around the atoms instead of saying that there is a presence of complete charge around the atoms but the charges keep on changing?
2. What does my professor mean by 'feeble' charges? Is the charge reduced or something else?

If the question is unclear, ask me to edit it in comments.

Answer 1

We describe an object, which does not exist, but is very real, called and 'orbital'. It is a region of space where an electron is 'likely' to be. It is the result of charge field interactions between positively charged nuclei. Normally we only discuss orbitals in the context of 1-2 Angstroms distance between atomic nuclei. Some orbitals have a character which we call 'bonding'. A bonding orbital, occupied by one or two electrons might be thought of as a 'bond'. However, bonds also do not exist, even if they are very real.

It is somewhat nonsensical to describe where in an orbital the electron is. If you must, the electron may be described as having a location of the entire orbital simultaneously, with vary weights to all of those locations, which are averaged. This needs to be thought of in the concept of 'wavefunction', which is a mathematical function describing the distributions of these weights. In the case of a polar covalent bond, the wavefunction more heavily favors locations near the more electron negative nucleus, and so the orbital 'concentrates' more of its charge near one end of the 'bond' and results in what we call a partial charge.

The reason we get squeamish about describing 'when' and 'where' is that the Heisenberg uncertainty principle renders the uncertainties energy and momentum of the system to be absurdly large when dealing with systems that are roughly the size of a chemical 'bond'.

At this point, nothing should be cleared up with regard to the original question.

The reality of the situation is that using when and where in this context is very likely to lead to predictions that will lead one astray when predicting physical phenomena. As bizarre as quantum mechanics is, it has been shown to be highly predictive of the bizarre world of atoms and molecules, and so you have to try to figure out how to not ask questions from a classical physics framework.

Answer 2

The study of chemistry requires one to build up every more complicated models to explain the behavior of the atoms. Dividing bonding into "pure" ionic bonds and "pure" covalent bonds is one such model that is wrong. No bonds are purely ionic or purely covalent. The bonds are always something in between.

The problem with the bonds in $\ce{H2O}$ being purely covalent is that water has a dipole moment. The way to explain the dipole is to note that the bonds are not purely covalent but have some ionic character too. Thus the oxygen has a partial negative charge but not a full -1 charge. The two hydrogen atoms must then each have a small positive charge. Now with partial charges the water molecule can have a dipole moment.

So the gist is that ionic bonds and covalent bonds are labels that we put on molecules to predict behavior. Such models are useful in limited circumstances. But if we try to overextend the model it leads to wrongful conclusions.