RBS is a very powerful technique, albeit not so common. Nowadays many people prefer to go for a (locally) destructive XPS or MS to get analogous information.

I know just three RBS labs and one is quite close to here (and the colleagues usually bring our Master students in visit there).

The technique is definitely useful if you want to have very accurate composition profiles (both in quantity and in depth resolution. Depending on what you analyze you can get down to the ppm and the nm depth resolution).

So yes, RBS is quite an accurate tool for the characterization of a diffusion or an oxidation profile. By oxidation I mean quantifying oxygen vs metals present.

I had some depth profiling done on a thin complex metal nitride film and I could see not just the nitrogen and metal distribution, but also the presence of some residual oxygen present in the sputtering system

I agree with Mattoeo Leoni and would like to add something more.

In my experience, RBS data can be sometimes tricky to analyze specially when you want precision. Therefore, if you add other ion beam techniques like NRA (nuclear resonance analysis) and ERDA (elastic recoil detection analysis), you can complement your RBS results and be sure about your results.

As stated above, RBS is good for depth profiling of the composition. Therefore, I think this technique would be very useful for measure diffusion/oxidation in a coating. in the past, I have used RBS to measure the stoichiometry of thin films deposited using a solution-based technique, and for this application, RBS works very well. See our paper, "Dielectric Response of Variable Thickness Ba0.6Sr0.4TiO3 Films for Property-Specific Device Applications," which can be found here:

http://www.ims.uconn.edu/~alpay/Publications/publications.htm

as the first paper in the "2008" section

Thanks everyone for responding. RBS sounds like a useful and powerful technique.. but based on the papers I read, when measure coating composition/line profile, I see most people do cross-sectional analysis (include sectioning, mounting, and polishing the sample) with SEM/EDS/WDS which is much more time consuming and destructive methods. Is there any reason RBS is not used that common when measure thin film composition? Is there some significant limitation to what kind of samples it can work with? Or the equipment is very expensive..? Hard to maintain? Thanks

I guess the main reason is that an ion source like the one employed in RBS is not a common or easy-to-maintain equipment. Plus as already said, the results are not that easy to interpret.

I would put on the table a few limits of RBS:

- it needs an accelerator ($$ to buy and $$ to maintain)

- it needs He ($$$)

- results are not always easy to interpret

- the maximum depth of analysis is limited to the micron range (ok for several applications but not for everything)

- it cannot directly profile e.g. hydrogen (well, there are tricks there)

In the end any lab has a SEM and most labs have an EDS/WDS detector so the analysis of polished cross sections, albeit destructive, come "for free". I would hardly believe on ppm quantification done via fluorescence in a microscope, especially on light elements.

I think you use RBS when you can access that technique, when you cannot destroy your specimen or when you need a high accuracy (EDS/WDS cannot definitely give you a 1nm/10 ppm depth profile accuracy)

That makes perfect sense Matteo, does RBS run in vacuum like SEM setup? does ion source usually come from a FIB type system? Or does it require significant setup like a synchrotron level?

RBS is done in vacuum of course and the major limitation is the availability/access to the system. In considering only RBS, one more limitation is the elements present in the samples. RBS is very good for heavier elements but has relatively lower accuracy for lower element.

I recently did quantify Zr impurity in a HfO2 film of 35nm and found the content of Zr up to 0.06%. But if you have similar amount of e.g. oxygen in some thin film it is almost impossible to detect.

Continuing about the discussion on RBS, i fully agree with Matteo Leoni on the the reasons for not so common application of Ion Beam techniques for materials characterisation.

The RBS technique is powerful for the study of thin films and the inter-diffusion across the interfaces.In fact we routinely use this technique for the analysis of mult-ilayered films. This techniques is regularly used for estimation and depth profile of oxygen by using a resonance scattering (called Elastic Backscattering Spectrometry EBS) at 3.05 MeV energy for alpha particles as projectiles. At this energy the the scattering cross section increases by 22 times the regular Rutherford cross section thereby increases the sensitivity of the oxygen determination. Further, by progressively increasing the alpha energy beyond 3.05 MeV one can depth profile the oxygen.

And for depth profiling of other low Z elements such as Hydrogen one uses other IBA techniques called Nuclear Resonance Reaction Analysis using different projectiles. All these IBA techniques are normally carried out in Vacuum hence the sample should be vacuum compatible. However there are some IBA techniques which are carried out in non-vacuum conditions also.

This discussion is becoming more and more interesting for learning point of view.

I had been doing IBA from last 3-4 years but always in vacuum, since my samples were always single crystal pieces of wafers. I have also used carbon identification using 4.26 MeV alpha. Depth profiling is quite interesting using resonances.

Can you please give some examples/details about non-vacuum IBA?

The non-vacuum ( or in air) IBA i am referring to is the Proton (to be precise Particle) Induced X-Ray Emission (PIXE) technique for the analysis of paintings and antiques which cannot be done under vacuum for various reasons.

Ahhh.... OK. Thanks for the clarification. I thought there is something new that I am not aware of until now.

Of course for PIXE, for some samples vacuum environment is an issue.

RBS is a very interesting technique when...you have access to an accelerator!..which use an helium beam with an anergy around 2MeV. So it is not always easy to characterize a material with this technique. But when you use RBS to analyze a thin film, from the spectra (Number of count/energy) you can have very rapidly a good idea of what your sample is made with at the first glance. It is not the case for numerous spectrometry techniques. Quantitatively it is an other story because you needs to fit your spectra with the old code RUMP(1985) or a more recent one SIMNRA. If you needs to analyze light elements the NRA technique is better. Concerning the thickness it is difficult to have goods result after a thickness of 5000 Ang. The accuracy is around 50 Ang at 2MeV depending on the atomic mass of the element. To get a more better accuracy (10 Ang) you can use The Medium Energy Ion Spectrometry (MEIS) which use an helium beam with a rather low energy. But this technique uses an electrostatic spectrometer and the analysis is rather long in time. An other interesting possibility of RBS is the channeling experiment which can determine the crystalline quality of your sample and where are the impurity layer in a thin film.

A famous example of analysis is the RBS spectra of a thin film of YBaCuO where you can see clearly the four peaks in the spectra of the four elements and therefore measures quite easily the stoichiometry of the sample.

remark: measurement of the concentration of oxygen from an RBS spectra is not a good idea. Use nucleus resonance with RBS at higher energy.

A very good source of information is the article in Wikipedia:

http://en.wikipedia.org/wiki/Rutherford_backscattering_spectrometry

We in ST did not have an RBS, but we extensively used this technique cooperating with people having acces to it, in a number of commom research work. As RBS I mean all the possible technique available with an energetic heavy ion beam, like the ones alredy reported in this discussion. I would like only remember that you can obtain a depth profile if you know the element to be profiled; or know the elements present if you know their in-depth position. But this is usually the case when using RBS.

The limitation are alredy reported: low sensitivity - but high mass resolution - for light elements, and high sensitivity but low mass resolution for heavy elements. Finally, quantitation is not bad and effectively used.

The problem with RBS is that the spectra is a number of counts versus backscattered energy. The sensitivity spectra depends on the quality, the solid angle (omega) of the detector and where is this detector. With both amorphous and crystalline sample the only accurate result concerns the element AT the surface. In depth the spectra is more difficult to interpret with accuracy because of the straggling (loss and dispersion of the energy) due to the sample AND the detector. And also because don't forget that the spectra is not number of count/depth but is versus backscattered energy. You need to know with accuracy the atomic density if you want to calculate the depth which corresponds to your signal. It is easy when you have a well defined compound but when it is not the case it is much more difficult too get an accurate profile of your sample. As mentioned by Giuseppe it is OK when you know what you sample is made with. If it is not the case don't start a characterization of your sample by this technique.

(see the attached file)

RBS use typically 1,8-2MeV protons and He nucleus(alpha particles) and is exallent tool for measure heavi elements on light element samples. In many cases if light element is in sample surface layer this techniques is also applicable. For analyzing of oxigen is better to use PIXE technique-this means to use energy of back scattered projectilles equal to energy of scattering anomaly of oxigen nucleus due to nuclear reaction. In caseof organic layer investigations, you mast take account for changes of polimer during ion irradiation

The Gaboriaud answer enlight the fact that a single analytica technique is never sufficient to characterize an unknown sample. We used a number of technique (OPTICAL AND ELECTRON MICROSCOPY, AES, XPS, XRD, TEM, RBS, AFM, ...) in the first step of a research work, and then only the most effective for the body of the work.

To do that, we used the techniques we had in our laboratory, and those available in external labs, preferably as a common research activity.

I only want to remember tha RBS was (and is) used in all the Mars landing missions to have a characterizatin of the planet surface. Not using of course an accelerator, but a radioactive element to produce a beam of alpha particles.

I have been working on ion beam analysis for over 25 years. RBS (or in general ion beam analysis) is by far the most powerful technique for thin film analysis. Why is it not mentioned much these days? Several reasons: 1) it is an old and very well developed technique, 2} it may not be widely available, 3) people do not understand it well enough, 3) not much communication between ion beam professionals with current material fields,etc. There are a lot of misunderstandings (some I read here):

1) RBS equipment id expensive to get and maintain. True or not true. A state of the art fully functional accelerator with end station capable of all ion beam techniques will cost much less than US 1million (typically on the order of 750k$),less than a low end SIMS machine. It requires some maintenance, mainly in the ion source but typically simple work.

2) People always just talked about RBS, but in fact when you have an accelerator you can do several ion beam analysis, sometimes simultaneously (e.g. RBS, PIXE NRA together in channeling orientation). That is much more powerful than just RBS.

3) RBS spectra are complicated and difficult to interpret. True. they are difficult because they have a lot of information. Unlike most techniques that give you a specific information, a RBS spectrum has many different things in it, e.g. thickness, composition, uniformity, roughness, diffusion, etc. Luckily there are also a lot of very good analysis package on RBS (and other ion beam techniques as well) available these days and thery are very good and easy to use.

For more information, visit my site:

https://sites.google.com/a/lbl.gov/rbs-lab/home

RBS is a number one technique for thin film analysis. Great advantages of the technique is really quantitative analysis. Among others RBS is used for implanted dose examination. Of course, surface reaction like oxidation can be studied by using RBS. Alternative methods: XPS, AES, SIMS. All mentioned techniques have advantages as well as disadvantages.

I honestly think that RBS used to be a common tool in materials science. There exist today alternative methods (which have now become common tools) that are more powerful & can generate a lot more data than RBS.