Dear Lu,

I think it is normal when your metal oxide films are not very smooth. In my experience, you can optimize the concentration and spin coating speed to minimize some aggregation on the films. Did you anneal your films as well? I suggest to check XRD to see structural and crystal quality for your MO films.

Best!

Dear Sir,

to find out whether the black dots are formed by the transition metal (compound), I would proceed as follows:

1. Estimate from your AFM image the fraction alpha of the surface that is covered by the black dots.

2. Probe a specific area A of the sample by XPS in surface sensitive mode, i.e. record a particular core level intensity ITM from the transistion metal for large polar angle.

3. Prepare a reference sample from the transition metal (compound).

4. Repeat step 2 for the reference sample with the same (!) parameters and get intensity ITM,ref

5. In case that you find a relation close to ITM = alpha x ITM,ref , there is strong evidence that the black dots are formed by the transition metal (compound).

Sincerely, FM

In general I agree with the advice by Dr. Műller above, however I would be very careful while comparing XPS results from point (2) with those from (4). The reason is that metal particles dispersed in a non-conducting matrix most likely will give you core level signals shifted towards higher binding energy with respect to the reference sample. This is because there is no good electrical contact to the spectrometer in the former case.

Dear Dr. Müller

Thanks for your long and detailed explanation. I am no familiar with XPS analysis technique. Therefore I have a such question, if XPS analysis is used, would it be straightforward to have a composition analysis to detect the whether there is the existence of un-oxidized Mo particles.

Dear Sirs,

concerning the remark by G.G:

please note that intensities are compared in step 2 and 4. Therefore, a shift in binding energy due to charging is no problem (at least in first order approximation) as long as the energy window is much broader than the particular core level signal that is probed.

concerning the question by C.L.:

in general, the chemical state of an element can be determined via the so-called "chemical shift" that provides information about the bonding of this element. In case of metallic Mo, the binding energy of the Mo-3d doublet is a few eV smaller than in case of MoOx, see e.g. our paper.

C. R. Chimie 12 (2009) 1228e1247

In case of charging of the sample you can at least decide whether there are different Mo species or whether there is a single Mo species.

Regards, FM

Dear Frank,

The first concern is how much metal particles will charge. It can be so severe that signals are way up and no spectra can be recorded. Even if we assume this is not the case you still get the problem of peak intensities being affected by charging.

One can, of course, try charge compensation with a flood gun but then most likely different regions of the sample, metal vs oxide, will react differently. It can be difficult to interpret the spectra.

Apart from all of the above, there is an additional concern - how can one distinguish between oxide peaks from the matrix and those from the native oxide on the surface of metal particles?

Dear G.G.

it seems that our experiences with charging are different. Even when using highly insulating samples (such as hydroxyapatite, i.e. the material that forms our enamel) we observe a shift towards higher binding energies up to 15 eV with sharp spectral features.

Of course you are right that inhomogenous charging increases the peak width but the increase of peak width is at the cost of an decreasing peak amplitude, i.e. the peak intensity in terms of peak area should not be affected.

You are also right that it is difficult to distinguish the matrix from the "black dots" and therefore I thought of an approximation when suggesting the method mentioned above. This is the way I would proceed with my 35 years old VG ESCA MKII setup without spatial resolution. When working with setups of the latest generation

(e.g. ThermoScientific ESCALab or - even better - Omicron NanoEsca) it should be no problem to probe the "hundreds of nm to several µ black dots" seperately.

Regards, FM

@ ResearchGate: Quite annoying that sometimes I have to write answers twice because the first one disappears upon clicking the "Add Answer" button.

1 please specify the point ".. transitional metal with quite high conductivity aggregate..", what does it mean?

2. is you organic solvent able to disperse the metal nanopowder and its oxide? there is possibility that its the original powder (metal) simply is not dispersed,

3. the size of dots ( a few horehounds of nm to um) and the density (according to SEM) will allow you directly check XRD,

4. I am not XPS experienced so I can not comment colleagues approach. apply XPS as you see useful at any point of my suggestions.

5. "...then I deposited the solution on silicon using spin coating. The films.." what sort of film you expect? your solvent must be evaporated (check the T of boiling) and only particles (metal, oxide, crystaline or amorphous) must be around.

6. make sure your metal nanoparticle is not initially oxide covered or not oxidized in ambient condition as it is exposed to open air, this sort of process can deteriorate you XPS XRD and EDS/SEM data,

7. apply EDS/SEM

8. collect back the powder into the same (or inert to that metal/oxide) solvent and precipitate by centrifuging and apply XRD

9. check you initial solution for particle.

10. it can be NP but agregated during cpin coating due to surface incompatability.

11. lookc for surfase plasmonic phenomena in solution and solid film.

12. in personal email you mentioned about crystalline and amorphous state. XRD/TEX diffraction will help you

Do you know how to simulate the p-i-n junction solar cell in COMSOL?