Yes and no...

Yes, you will detect the presence of pi-bonds, which are needed for conduction of electrons, but that will mainly tell you that it is not an insulator (as with sp3 in diamond).

The electrical conductivity of a metal or semimetal is dominated by the crystallinity. So defects, or mechanical strain, that you don't see in XPS or NEXAFS will still influence and may well dominate the electrical conductivity.

Yes and no...

Yes, you will detect the presence of pi-bonds, which are needed for conduction of electrons, but that will mainly tell you that it is not an insulator (as with sp3 in diamond).

The electrical conductivity of a metal or semimetal is dominated by the crystallinity. So defects, or mechanical strain, that you don't see in XPS or NEXAFS will still influence and may well dominate the electrical conductivity.

Hi Mazdak,

from a more detailed analysis of the carbon edge you can get even more informations than the presence of Pi orbital, and as Kristen correctly said, this is no measure of electrical conductivity. For example CNT can be metallic or semiconducting depending on chirality, but are both characterized by the presence of Pi bonds.

For example, consider that the Pi orbital are always perpendicular to the C-C sigma bond. With an angle resolved XPS you can get the direction of a CNT material (macroscopic geometry and even the so called chirality if they are pure enough) and this will tell in which direction they conduct and possibly even if they are metallic or semiconducting CNT. See for example:  Chem. Mater., 2006, 18 (23), pp 5624–5629

yes, as pi-pi* bond increase the conductivity will also be increased.

I agree with Lolli and Krister. 

This can be only solved by very thorough study.

Hello Mazdak,

really this transition is evidence, but not absolute, as well Krister says there are other facts that contribute to conductivity and are not objectively measured by this technique. Therefore, make quantifications only from these data may be erroneous

Hi Krister,

Thanks for your answer. I assume that you have meant “yes, but just qualitatively and not quantitatively”.

One of the features of the XPS is its capability to be used for quantitative analysis. Actually, in practice, one will obtain an elemental analysis of the surface (through survey scans) and the concentration of the oxidation states of each element (through narrow scans). Therefore, any other property of the material evaluated via XPS, will be through an indirect relation between that property and the elemental/chemical properties of the surface. Even the very initial quantitative elemental analysis in XPS is established based on some deeper physical properties of elements such as their ionization energies and Relative Sensitivity Factors.

I mean, a potential feature is there and full efflorescence and utilization of that becomes possible after some people explore it thoroughly and establish the required relationships between the acquired data and the desired property.

As the first step, it is good to know that the π–π* transition has something to say about the conductivity of the sample. This is that implication that I was pointing to in my question in a very general sense.

The next step is to see if anybody has ever tried to establish a relationship between the contributions of this component and the gross conductivity of the sample (which in fact should be measured and checked through the appropriate method).

Thanks also for the examples you gave about the parameters such as crystallinity, defects and mechanical strain that can influence the conductivity. Interestingly, in case of carbon, the most important features related to crystallinity are dependent on the atomic bonding of carbon, namely sp2 or sp3 hybridization. This can fortunately be studied via XPS. Defects are also other things that can be explored with XPS since they will affect the chemical state of atoms surrounding them.

Anyways, you might be right about the point that XPS is not the best thing to study the conductivity of carbon. It can however, be also something more than just a qualitative sign.

Regards,

Mazdak

Hi Giulio and thanks for your comment.

I agree with you that this might not be the best way studying the conductivity of carbon. However, I would like to know just as a “side-possibility” if it is conceivable to extract some more info from the π–π* contribution.

Thanks also for the nice example about studies of CNTs using XANES.

I can also recommend you having a look at the following paper that is good example of XPS potential for studying different aspects and properties of carbon.

Okpalugo, T. I. T., P. Papakonstantinou, H. Murphy, J. McLaughlin, and N. M. D. Brown. "High resolution XPS characterization of chemical functionalised MWCNTs and SWCNTs." Carbon 43, no. 1 (2005): 153-161.

Article High resolution XPS characterization of chemical functionali...

Hi Firoz, thanks for your answer.

By π–π* I meant the transition revealed in the higher energy end of the C 1s spectra and not simply the π bonding. I assumed that these two should be correlated, but I don’t know how and to which extent and whether it can be used for any estimation about the conductivity.

Do you have some evidence or personal experience or reports in the literature about it?

Can you please let me know?

Mazdak

Hi Anna, thanks for your comment. I agree with you and actually I am looking for any attempt by people who may have tried to explore this a bit.

Dear Mazdak, as my simple suggestion, it seems that a relative comparison is possible. When you examine a change of at.% for pi-pi bonding throughout an electrochemical test of the carbon bounded with an ion so as to possibly react with each other, then, you can compare both impedance behavior just before and after the electrochemical test. In my experience, after long cycles(repetition) of the electrochemical(reaction) test the pi-pi bonding at.% was decreased as well as the conductivity, eventually.

Dear Jong-Huy,

Thanks for your answer and suggestion. It is good to know that you have some personal experience about the relation, at least qualitative and for rough comparison, of π–π* and the conductivity of carbon.

In fact, as you told about your experience, the carbon has been corroded (undergone some electrochemical attack) and has lost some of its conductive properties that has been revealed in π–π* transition as a decrease in the intensity of this components after testing.

My calculations suggest that the question of the electrical conductivity of CNTs depends essentially on the Coulomb interaction of pi-electrons at one site.

Pi-electron jumps from one site to create the conduction band, and the Coulomb interaction prevents pi-electrons jump from site to site. Such a situation can be correctly described in the Hubbard model.

This is dedicated to our article:

Journal of Experimental and Theoretical Physics 07/2014; 118(6):935-944.

Russian Physics Journal 12/2013; 56(7):791-800.

Physics of the Solid State 01/2013; 55(1).

Physics of the Solid State 09/2012; 54(9).

These article These articles are presented on my page: https://www.researchgate.net/profile/Arkadiy_Murzashev/contributions

Dear Arkadiy,

Thanks for your comment and for your papers. Actually, I am not familiar with the computational and theoretical solid state physics principles you have used in your papers. But, maybe I can at least conclude according to your brief comment that availability of Pi-electrons (that can exist in sp2 type of carbon and the π–π* transition of the XPS spectra can in some way indicate their presence), at least through some probabilistic relationships, can imply some facts about the electrical conductivity. Probably, you have meant that quantification of such relationship could be possible via some models such as Hubbard. But for the moment, just a rough qualitative relation could be imagined like: the stronger the π–π* transition component of the C 1s, the higher the amount of the Pi-electrons, and therefore the higher the population of the Pi-electrons who can pass the barrier of Coulomb interaction. So, a higher conductivity could be achieved.

Hi Roman, Thanks for your comment.

Dear Mazdak Hashempour 

The fact that at each node in the CNT is one pi-electron. And the conductivity is determined not by their number and width of the conduction band. A conduction band width depends on the parameter describing the electron hopping from site to site. This option increases the conduction band. A Coulomb repulsion prevents jumping. And the conduction band is divided into two subzones. The distance between them: HOMO-LUMO is the difference between the Coulomb repulsion U parameter at one site and the band width W.

The number of pi-electrons may be increased or decrease only by adding to the system of chemical elements having an excess or deficiency of electrons.

Dear Mazdak,

I am interested in the π–π* transition component of the C 1s. XPS spectrum I observed  that in graphene material, the shake up is well visible with ESCA Lab machine. However using photons of 350 eV in the synchrotron, the satellite is not visible on graphene. Is this explainable by differnt probabilities ?

Thnks

Hi Jean Michel,

Thanks for your interest in this discussion. Unfortunately, I have not used the synchrotron for XPS and am not familiar with tricks of this technique. However, as a rough idea, since the energy of the photons you have used in synchrotron (350 eV) are almost close to the binding energy of π–π* transition (≈ 291 eV) it might has not been energetic enough to excite a visible signal. What if you use a higher energy photon?

Hi Mazdak,

Thanks for answer.

If we use higher energies (100 eV for example), we see clearly the transition on the C1s spectrum at 292 eV. So this trasition depends on the energy of photoelectrons. Regards

Right, so it makes sense.

Hello Mazdak,

Have you found answer on this question? I'm curious because I have the same question right now.

Regards,

Dmitry

Hi Dmitry,

Thanks for your interest in this question. My general intuition after all discussions interchanged, has been that a precise and quantitative conclusion about the electrical conductivity of carbon, is not possible just based on the π–π* transition component of C 1s XPS spectra, or at least, based on the current knowledge and investigations. Maybe a dedicated work, could cast some more light on the quantitative capacities of this technique for this particular characteristic (electrical conductivity) of carbon. For the time being, we could at most, say that there is a qualitative relation, where more conductive forms of carbon (like CNTs based on my personal research) show a clear π–π* transition and less conductive forms of carbon (like activated carbon or graphene oxide) do not show that, or show a very weak transition.

Hi Mazdak,

Thank you!