Validity of the Pauling Scale of Electronegativity?

[cosmichorizon]

I propose the Pauling Scale is not valid in all circumstances, and other scales of electronegativity would do better to predict the behavior of electron density in molecules:

Electronegativity seems to be poorly defined for such a widely used chemical property: a 'tendency' of an atom or functional group to attract electrons. Although the Pauling Scale works well in most cases to describe the relative stability of ionic bonds, it does not always correctly predict which atom or functional group will 'more strongly' attract electrons. As an example:

According to the Pauling Scale:
Ru = 2.2
Si = 1.9

With these values in mind, we will compare the electron density in the compounds Ruthenium Dioxide and Silicon Dioxide:

Since Oxygen is more electronegative (on any electronegativity scale, not just Pauling) than either Ru or Si it is safe to assume that there will be a transfer of electron density from the Ru or Si cations to the Oxygen anions. This much we are sure of. If we were chemists we might be satisfied with this as both Ru and Si will be in a 4+ oxidation state. But as a physicist we know that an exact charge transfer does not occur in either case. Instead the electron probability density will reside closer to Oxygen, but again the exact charge transfer does not occur, so how much charge transfer exactly?

Using the Pauling Scale one might assume that the ruthenium will be relatively less likely to transfer its electron density to oxygen than silicon. Going along with this we will find that there will be a greater negative screening charge around the oxygen atoms in SiO2 than those in RuO2... But this is not the case.

Experimental data has shown that the binding energy of the Oxygen 1s electrons in RuO2 is lower than that of that of the Oxygen 1s electrons in SiO2. This shows the ruthenium transferred more electron density to the oxygen atoms in RuO2 than silicon to oxygen in SiO2. (Please comment if you do not accept this, and I will point you in the direction of XPS evidence.)

For those of you unfamiliar with the relation of core electron binding energies and the surrounding electron density:

The more reduced (greater excess electron density) an atom is the lower the binding energy of core electrons.

I find that these results agree with other electronegativity scales such as the Absolute (Mulliken) Scale and the Allen Scale, however the Pauling scale, horrendously fails in this situation.

Would the readers please comment on why the Pauling Scale is still used over others (besides historical reasons) specifically if it has any advantages over the Allen or Mulliken Scales.

Further, if the reader has any knowledge of the reason the Mulliken Scale is used over the Pauling Scale in XPS analysis, I would be grateful if you would show me why.

 

[Borek]

I propose the Pauling Scale is not valid in all circumstances

It is valid in all circumstances for which it is valid ;)

Electronegativity (of any kind) never gets better than just a "rule of thumb" kind indicator of what to expect. If you hope for more, be ready for the disappointment.

 

[DrDu]

Experimental data has shown that the binding energy of the Oxygen 1s electrons in RuO2 is lower than that of that of the Oxygen 1s electrons in SiO2. This shows the ruthenium transferred more electron density to the oxygen atoms in RuO2 than silicon to oxygen in SiO2. (Please comment if you do not accept this, and I will point you in the direction of XPS evidence.)

Of the 1s electrons? The 1 s binding energy forms the basis of Moseleys law and increases to a very good approximation linearly with nuclear charge.
Anyhow the absolute value of the 1s binding energy gives no information about the bonding, as the bonding involves the valence electrons and not the core electrons. However, the valence electrons lead to small shifts in the 1s binding energy which can be detected e.g. by Moesbauer spectroscopy and can be correlated to the oxidation state of the atom.

I have to second Borek. Nobody beyond introductory chemistry classes will claim to be able to predict charge distributions in complex B group compounds from electronegativity scales, and even less so for higher B-group elements. From a more theoretical point, it is even difficult to decide to which atom parts of the charge distribution are to be assigned. One of the few reasonable approaches seems to be the Bader atoms in molecules procedure.

 

[cosmichorizon]

It is valid in all circumstances for which it is valid ;)

Electronegativity (of any kind) never gets better than just a "rule of thumb" kind indicator of what to expect. If you hope for more, be ready for the disappointment.

Ok this is not really very clear. I asked for examples of why the Pauling Scale is a better 'rule of thumb' than other scales of Electronegativity. This is redundant.

 

[Borek]

I asked for examples of why the Pauling Scale is a better 'rule of thumb' than other scales of Electronegativity.

Problem is - it isn't. As all these scales are just proxies none of them is substantially better - they are all similar, sometimes one works better, sometimes the other, but they all can be used to make just some rough predictions.

 

[cgk]

OP, as others said, electronegativity is a somewhat vague concept, and predictions derived from it cannot be expected to hold under all circumstances. This is also the main reason why the Pauling scale is still prevalent: Just as you said, it has historical reasons, and since the concept is vague anyway, better scales do not get adopted easily. Note also that there are not only the Allen and Mulliken scales; infact, there are *more* electronegativity scales than researchers who worked on this. But, since you asked about those two specifically:

- Allen's electronegativity scale is one of the newest ones. In contrast to many other scales, it is directly based on physically observable data (ionization potentials). However, many people in the field do not like that it disregards the properties of the anions *completely*, and thus can only give a somewhat limited view of the electronic structure in the negative charge direction. I personally also use the Allen scale when I need electronegativities, but one has to recognize that it is limited.

- Mulliken's scale is probably the most flexible and powerful scale, but it is also somewhat empirical. It's main problem in practice it that is much harder to use, since electronegativities are assigned to *orbitals*, not to atoms. This makes it highly impractical for rule-of-thumb predictions, and offers lots of opportunities for using fudge factors because *you* decide which orbitals/hybridization/etc to use for an atom, and thus which data to use as input for the scale.

One final note: While I normally agree with DrDu, I do not think that Bader charges are a sensible way of assigning atomic partial charges, especially not when it comes to electronegativity. The charges are often highly counter intuitive in practice and do not agree with any expectations based on "chemical intuition". Now one could in principle say that this is a defect of the electronegativity scales, and real electrons (as predicted by Bader's AIM) simply do not work that way. However, there *are* purely ab initio based charge scales which are highly consistent with electronegativities. See, e.g.:
http://dx.doi.org/10.1002/jcc.10351 or http://pubs.acs.org/doi/abs/10.1021/ct400687b .

 

[cosmichorizon]

Of the 1s electrons? The 1 s binding energy forms the basis of Moseleys law and increases to a very good approximation linearly with nuclear charge.
Anyhow the absolute value of the 1s binding energy gives no information about the bonding, as the bonding involves the valence electrons and not the core electrons. However, the valence electrons lead to small shifts in the 1s binding energy which can be detected e.g. by Moesbauer spectroscopy and can be correlated to the oxidation state of the atom.

I have to second Borek. Nobody beyond introductory chemistry classes will claim to be able to predict charge distributions in complex B group compounds from electronegativity scales, and even less so for higher B-group elements. From a more theoretical point, it is even difficult to decide to which atom parts of the charge distribution are to be assigned. One of the few reasonable approaches seems to be the Bader atoms in molecules procedure.

Ok well I know it is possible to see these small shifts in the O1s with XPS, Moesbauer spec. is not needed also. My question was directed towards the use of electronegativity to predict relative shifts based on the electronegativity, and the pro's and con's of the different scales of electronegativity.

Problem is - it isn't. As all these scales are just proxies none of them is substantially better - they are all similar, sometimes one works better, sometimes the other, but they all can be used to make just some rough predictions.

Ok can you give me an example of when the Pauling scale does a better job than say Absolute Electronegativity? As I understand it the shift in binding energy in XPS is better predicted with the Absolute Scale, and I would like to know why.

 

[DrDu]

One final note: While I normally agree with DrDu, I do not think that Bader charges are a sensible way of assigning atomic partial charges, especially not when it comes to electronegativity. The charges are often highly counter intuitive in practice and do not agree with any expectations based on "chemical intuition". Now one could in principle say that this is a defect of the electronegativity scales, and real electrons (as predicted by Bader's AIM) simply do not work that way. However, there *are* purely ab initio based charge scales which are highly consistent with electronegativities. See, e.g.:
http://dx.doi.org/10.1002/jcc.10351 or http://pubs.acs.org/doi/abs/10.1021/ct400687b .

Thank you, this seem to be really interesting articles. I didn't know that Bader performs so bad in practice.