Versão do seminário apresentado por Celia Olivero (Horiba) na seção UCS do Instituto Nacional de Engenharia de Superfícies no dia 28 de junho para um público de 18 estudantes, professores e profissionais de empresas.
We will today mainly present some results in different fields
Brief principle of the technique (identical to a Ne light tube)
View of the HJY lamp and of a crater. The 2 electrodes are totally disymetric. This is responsible of the creation on the sample surface (conductive or non) of a constant negative voltage, the DC bias (Vdc). This DC bias voltage accelerates the Ar ions which sputters the sample surface. The sputtering is very fast and the unique double pumping system permits to get flat and deep craters. The depth resolution is sample dependant but has been shown to be as good as 1-2 nm and the maximum depth attainable in one shot can be more than 150 microns.
The design principle of GD has also no much changed since its introduction by Dr Grimm in the 70’s. The sample is usually placed on the o’ring sealing the chamber and is one of the electrode. It faces a Cu tube that is the other electrode. The operation is not that different from a lamp used for lightning. A low pressure gas is flushed in the chamber. When the RF power is switched on an electrical plasma is set up. This plasma assures both the sputtering of the sample and the excitation of the sputtered species. In GD we have a spatial separation of the erosion at the sample surface and the excitation in the gas phase and so (in first approximation) a dissociation of the 2 phenomena that allows multimatrix calibrations and depth profile quantification.
Benefits of the RF source. Better for surface (often oxidized so partly non conductive) and allowing simpler calibrations
Pulse permits to do fragile samples and to get better depth resolution
The instrument is essentially a polychromator for the simultaneous measurement of all elements (in depth profile we need a simultaneous measurement).An additional monochromator can be added for flexibility.
Polyscan is limited. Not simultaneous, same resolution as poly, limited spectral range. Mono is better. The addition of a monochromator in the instruments provides a superior flexibility. The mono permits to measure any N+1 element simultaneously in a depth profile The mono has the highest resolution The mono also permits (if the sample is homogeneous) to run a full spectrum measurement (Image)
The GD instrument is first an optical spectrometer and that commands for the largest part the obtainable results due to the spectral range covered, the resolution you can achieve and as the GD source is not a very intense source of light, the grating quality that mainly commands light throughput of the instrument. We have 2 types of GD instruments, the GD-Profiler 2 shown here and the GD-Profiler HR. They differentiate by the different focal length of their optics.
This is to explain the differences between the bulk analysis and the Depth Profile Analysis (named QDP: Quantitative Depth Profile). Bulk analysis: Analysis of a metal piece…, sputtering to remove contamination, analysis time: several seconds Results: Chemical composition of the sample Depth profiling: Substrate with different layers or surface treatments, Sputtering through the different layers, no pre-integration time, very fast acquisition rate up to 2000 acquisition/second. A lot of points of measurements from the surface until the matrix Results: plot of chemical composition of the sample for all elements simultaneously versus the depth (from the surface to the substrate)
Quick summary
To summarize, these are the general questions GD can bring a piece of information
If we look now to the depth profile of a more industrial multi layers sample, here an alternance of TiN and TiAln deposited on a stainless steel, we can sse the degradation of the depth resolution with the depth due to large change of sputtering rate between the materias and the roughness effect (I will come back on this issue later).
Samples here are various steels, Ti, Ni and carbides. Coatings can also be used within multimatrices calibrations.
We will today mainly present some results in different fields
Help for industry and for research : an example in the paper from Prof. Alvarez
An other example showing a benefit of Rf GD-OES, the ability to measure O, N and Cl, H not shown on the slide. This allows to follow all sorts of surface treatments and indicates why the technique is precious for corrosion studies.
Classical applications where GD has a lot to bring. Control of treatments. Quick comparisons of samples. Very popular in Iran
This is an overlap of both profiles: Good and bad process showing clearly the different behavior of the N and C, and the influence of the process on the treatment.
Full range of applications
Now some examples coming from EU projects. GD can be used to follow ionic implantation
European project about C3N4 coating. Realisation of layers harder than diamond.
Let us now focus on thin layers This result has been presented in Spain at AIN by the customer. The sample has 107 layers deposited (CrN, TiN) on steel. The diagram shows the result of 2 quantified measurements performed on the same sample. The depth resolution and the repeatability are clearly revealed by this example. Of course the depth resolution degrades with the depth due to preferential sputtering, induced roughness etc but it remains pretty good.
Now back to flat samples and real thin layers Lot of work done on such thin films. Anodised Al Consider next another coated sample. This one is prepared by anodising a polished Al substrate in various solutions. The result is an Al substrate, then a 240 nm thick Al2O3 layer, followed by a 120 nm outer AlOx layer containing about 3% B, and a Cr spike about 2 nm thick at 40 nm from the outer surface. Just in front of the Cr spike is a region with bubbles.
Let us see an example illustrating the benefits of RF GD-OES. As we have seen the instrument is capable to sputter microns/minute (hundreds of nm per second) but as optical signals can be recorded fast this speed of analysis is not an handicap for the depth resolution. Here is shown a thin anodised layer. Anodisation was done in acid pH, the anodisation is porous and a delta layer (Cr) a few nm thick was deposited at the bottom of the pores (See the TEM view).
This sample has been used to compare RF GD and SIMS. The depth resolution was similar with the 2 techniques and of course the time of analysis is without comparison. Such an approach certainly changes the way people appreciate surface analysis. It is now possible to quickly run multiple samples assuring nearly immediate feedback.
Yes we can do surface !! Surface contamination can be important to check and monitor when it can be linked to the process. An example illustrating that point is given here. It comes from the electronic industry and shows the capability of RF GD for ultra thin films analysis. Customer is producing ultra pure Cu, base to electronic circuit boards. The first result shows the top surface (less than 1s of sputtering) of a good sample, the second one is from problematic sample: contamination is obvious and coming from the rinsing water used.
This result was presented in 2004 in JAAS. The sample is a monolayer of thiourea deposited on electropolished Cu. The layer is sputtered in 20ms but the distribution of species precisely follow the structure of the molecule.
To be sure of the previous result a second molecule parallel to the surface was deposited and immediately measured
An other important aspect is the control of the source. Our source can operate in continuous mode or in pulsed mode. Source can be pulsed at ms frequency (from 33Hz up to several kHz). This example is on an other type of coated glass and shows the potential benefits of the pulse mode. Without pulsing (left), even with a low power some crack may happen, with the pulse mode (with high power sent in each pulse – but a resulting power average equivalent to the first case) we obviously have less problems. The synchronised acquisition allows to record signals only during the time the source is on improving signal/background ratio.
With pulse glasses can be done easily without damage (here a thick glass and deep crater)
Use of pulsed in PV. For Na. Here an other glass with coatings : pulse : no diffusion of Na !!
Example of CIGS
Other benefit of this pulsed mode is the reduction of the sputtering rate and the minimization of surface damages while keeping interesting information. Reducing the fast SR of GD is often asked: usually the only way to do it is to reduce the applied power (hence limiting the sensitivity). The pulsed operation provides enough power for excitation while reducing the SR. It is also a useful approach to identify details on the surface and it minimizes sample heating and damage.
We will today mainly present some results in different fields
We have seen that RF GD has a very fast sputtering, it is also a very soft one – when compared to SIMS for instances. This is due to the plasma characteristics. The possibility to obtain a very fast erosion due to the high density of the plasma but a soft one due to the relatively low energy of the incident particles is at the origin of a new application: to use the RF GD as a cleaning system tool and a sample preparation technique for SEM.
The example here shows a stainless steel sample, mirror polished and a part of the GD spot inside. The preferential sputtering, usually a drawback, is here turned into an advantage and used to clearly reveal the structure of the sample beneath the surface.
Let us see now a couple of slides showing some SEM views after mechanical polishing and RF GD operation.
We will today mainly present some results in different fields
Now some figures of merits comparing rf GD to classical surface techniques as a possible conclusion.
The capability of GD to characterize quickly multilayers at the nm scale make it an interesting tool in many fields.
GD operates directly from the sample surface, SEM requires a cross section. Very often in a lab the 2 techniques are used complementarily: can we go further and have the 2 domains more closely together: may be !
Finally I would like to mention 2 recent books released on GD where valuable information can be found. Thanks for your attention.