Dear Anish Bhattacharya ,

The vacant oxygen is NOT an "O-specie" with an oxidation state of (-1), the vacant oxygen is a vacant position in the lattice created due to the absence of an oxygen ion, thus, there is NO XPS-signal for the vacant oxygen, that is the core of the present question.

Please note that the oxygen ion with an oxidation state of (-1), present in peroxides, for example, have O 1s in the range (531.1 - 532.0) which is out of the range of the O 1s in the article attached by Helen Par .

For more details regarding the XPS of oxygen in its different states, please refer to:

DOI: 10.1039/a908800h

Dear Helen Par ,

You are right, the absent atom can NOT have a signal in the XPS spectra. However, an unusual signal for an existing atom can be used as an indirect measure for the percent of the absent atom.

It is reported that O 1s XPS peaks are fitted by three curves:

1) Low energy curve: which is associated with the oxygen atoms bonded to cations which are in the stoichiometric-oxidation state, i.e. the cations have an oxidation state similar to that in the chemical formula of the material.

2) Medium energy curve: which is attributed to (O-2) ions bonded to non-stoichiometric oxidized cations, i.e. cations have oxidation state less than that of the chemical formula. Such cations are responsible, to maintain the charge balance in the structure, for the presence of the oxygen vacancies.

3) High energy curve: which is attributed to the presence of loosely bound oxygen on the surface such as in H2O and OH groups.

Thus, the medium energy curve, corresponds to oxygen ions in oxygen-deficient regions, can be used as an indirect measure for the oxygen vacancies.

Helen Par There are various species of oxygen, namely, chemisorbed oxygen, lattice oxygen and vacant oxygen present in the matrix of an oxide material. The deconvolution of O1s spectra is shows the amount of each species present in the sample. The vacant oxygen is a O- specie, rather than absence of oxygen, which has a separate binding energy than the other oxygen species. Hence, binding energy for vacant oxygen can be observed in XPS spectra of an oxide material.

Please refer to the following papers for a better explanation:

1. dx.doi.org/10.1039/C5CE00279F

2. dx.doi.org/10.1021/acssensors.5b00123

Dear Anish Bhattacharya ,

The vacant oxygen is NOT an "O-specie" with an oxidation state of (-1), the vacant oxygen is a vacant position in the lattice created due to the absence of an oxygen ion, thus, there is NO XPS-signal for the vacant oxygen, that is the core of the present question.

Please note that the oxygen ion with an oxidation state of (-1), present in peroxides, for example, have O 1s in the range (531.1 - 532.0) which is out of the range of the O 1s in the article attached by Helen Par .

For more details regarding the XPS of oxygen in its different states, please refer to:

DOI: 10.1039/a908800h

Imad Disher Please refer to the attached excerpt from the article you shared. I think it refers to the vacant oxygen sites in an oxide matrix through incoherent distribution of electron affinity of Oxygen atoms.

Dear Helen, why not analyzing the particular bonding partner of Oxygen? Its about Silicon, observed by Si2p (99.4 ... 103.8 eV) for instance, shows much more sensitivity on Oxygen vacancies the the O1s itself.

Please see the references below, you will find lots of interesting and pertinent information:

1- Saulius Kaciulis, G. Mattogno, A. Galdikas, Influence of surface oxygen on chemoresistance of tin oxide film; 1996, Journal of Vacuum Science & Technology A Vacuum Surfaces and Films 14(6):3164-3168. DOI: • 10.1116/1.580207.

2- SnO2 by XPS; 1993, Surface Science Spectra 2(1):50-54. DOI: 10.1116/1.1247724.

3- Dae-Kyoung Kim, Kwang-Sik Jeong, Yu-Seon Kang, Controlling the defects and transition layer in SiO2 films grown on 4H-SiC via direct plasma-assisted oxidation. Scientific Reports volume 6, Article number: 34945 (2016). https://doi.org/10.1038/srep34945.

Best regards

As above mentioned a missing oxygen cannot have an O1s signal, you have to look with high resolution into e.g. Si2por Ti 2p region. Also in the valence band there are sometimes emiisions which are belonging to suboxides e.g. TiO2. Sometimes you have as in Cu and CuO due to relaxation effects the same binding energy for + state and 0 state. in that case you might use the Auger parameter to obtain more information. But fitting Augers it might be cumbersome.

As an example you may look here:

Morphology of the Si–ZnO interface

U Meier, C Pettenkofer

Applied surface science 252 (4), 1139-1146

In principle, the nearest O atoms can experience some shifts due to a missing neighboring oxygen. For example, in O-Si-O, the first O does not experience the exact chemical environment such as in O-Si-[V]O ([V ] denotes a vacancy).

The binding energy would be slightly shifted.

In practice, I'm not sure whether you can resolve this shift (which would depend on the electronegativity of the cation, among others).

As pointed out earlier, a better candidate for looking into vacancies would be the 'real' neighbor, namely, the cation, which should have components with a lower oxidation states.

A nice example for that was recently published by Skaja et al. regarding TaOx (Fig. 1)

https://www.nature.com/articles/s41598-018-28992-9.pdf

To Lior: As you wrote "in principle" but in the spectra it would only contribute as a contribution to an inhomogenious broadening which analytically is hard to unravel.

Dear Helen Par,

This is possible due to fact that vacant oxygen ions will create more charge on the remaining of the lattice if the structure stay stable. However, the deconvolution analysis should be made on the data measured synchrotron source X-ray radiation due to precise and intense nature of it.

Thanks

Engin

The identification and quantification of chemical species by XPS must be done carefully. Before claiming the existence of any chemical species, one should check for the absence of surface charging effects on the sample. If these effects are discarded, then peak decomposition can be made. For decomposing peaks, one must correlate the relative concentrations of the elements of the species that one is proposing. This is done by calculating the stoichiometric ratios between the peaks ascribed for these elements. For example, if I claim that a certain oxygen peak corresponds to titanium dioxide, the ratio (O)-Ti/(Ti)-O must approximately be 2.0. The same applies for other oxides. Also, don't forget that adventitious carbon contains oxygenated species that must also be identified in the O 1s peak.

Please take a look at a procedure (see the attached picture) that we recently applied for analyzing TiO2.