Sample Test and Analysis

The standard diameter is 4 mm corresponding to optimum compromise between crater shape and quantity of light collected. 2 mm, 7 mm and 1 mm anodes are also available. We have also done some 8 mm, 10 mm and 6 mm anodes on request. As long as the sample size is larger than the o-ring of the ceramic, we can directly analyze these samples.

Samples smaller than the o-ring can also be measured using an accessory – the small sample holder shown below. With this accessory, the vacuum seal is assured by the holder itself; the sample just needs to cover the anode diameter (5 mm for a 4 mm anode, 3 mm for a 2 mm anode etc). This accessory can also be used for rough surfaces.

Hygroscopic or inflammable samples can also be done with care. A dedicated accessory (“Li bell”) was designed for the measurement of Li ion electrodes specifically. It allows transporting the samples to the instrument under an inert atmosphere and running the analysis without contact to air.

Fragile samples and soft thin films can be measured by Pulsed RF GDOES.

GD operation requires the use of a primary vacuum. Thin films or soft samples can be deformed due to the mechanical stress induced when vacuum to close the chamber, is applied on the samples. The analysis is then affected or even made impossible if the sample comes in contact with the anode, possibly creating a short circuit. So when analyzing this kind of sample, we need to glue the sample on a rigid substrate.

Appropriate gluing of such samples on a rigid support is the best way to maintain the quality of the analysis (control of the crater shape and preservation of the depth resolution).

Pulsed operation is most often needed for such fragile samples.

The domain of polymer films is vast, applications are many and one could wonder why an elemental technique as GD would be of interest, thinking that Raman would be more appropriate.

Actually GD provides complementary information to other techniques that are useful for process development and failure analysis. Pulsed RF GDOES has for instance, measured paints on car bodies, plastic films to protect mobile phones, plastic DVD, or organic solar cells. Several papers have been published on these applications.

Pulsed operation is often needed to avoid thermal damage to the samples and preparation strategies may be required if thin films are to be measured.

The patented UFS might also be of interest in many cases as, for example, the analysis of protective films for mobile phones. In this example, 110 microns are sputtered in 11 minutes using the UFS, the presence of tracer elements permits identification of the layers and their function: Na, for instance is there for the capacitive contact – assuring that when typing on your film, information is sent to the phone glass window below.

A very new patented development from HORIBA Scientific allows rapid (1-10 mn) sputtering of polymer layers, providing an optimum flat crater bottom when previously hours were needed. This is ideal to look at buried interfaces for painted car bodies, plastic DVDs, encapsulated solar cells, etc.

This new development keeps enlarging the range of applications for which the Pulsed RF GDOES benefits of speed and ease of use can be employed with great profit.

This applies to bulk analysis only where the purpose is to measure the core of the sample. The idea is to use preburning. Pre-burning the sample, also called pre-integration, simply means igniting the discharge, and waiting to reach a steady state before integrating the spectral intensities. Igniting the discharge also helps remove vapors, such as water, from the discharge cavity by destroying their molecular structure.

The result is an improved relative standard deviation (RSD) for calibration and bulk analysis measurements. Typical preintegration times are between 30s and 90s. Double preintegration and the use of different conditions could help to minimize this time or make it more efficient.

For obvious reasons, this pre-integration is not applied for surface and coating analysis, but plasma cleaning strategies (described later) can be applied.

Even with clean Ar plasma gas, residual gas species (H, N, O, CO₂ or H₂O) are present from the sample surface (that can be porous, hydrated, etc.) or from leaks. They have a nonnegligible effect on the light intensities as they contribute to changes in the plasma. They should be minimized as much as possible.

“Plasma Cleaning” strategies offer elegant, practical ways to minimize such effects. They are based upon the following considerations: anode and sample surfaces, and gas atmosphere are sources of contamination (one more than the others, depending on the configuration) requiring the setting of a dedicated optimum strategy:

• Use of higher purity Ar and longer flush time prior to analysis.
• A silicon or alumina sample can be sputtered just prior to the analysis of a clean sample (without lamp cleaning in between). This deposits a small protective film on the tip of the anode that contributes to clean anode walls.
• A first measurement can be done and the sample then moved on the side without the lamp being in contact with air. The second measurement will be cleaner.
• Finally, and this is unique to pulsed RF operation, it is possible to operate the RF source at ultra low power in pulsed mode without sputtering the material. This creates a soft plasma which has too low an energy for material sputtering, but is sufficient in combination with the gas flow to clean the surfaces prior to analysis. After a few minutes of such “surface cleaning,” the normal analysis can start.

The Plasma Cleaning has been described in a paper published in JAAS.

Calibration principles are simple, and based on the unique feature that in the GD plasma, the erosion and the excitation are physically, spatially separated and can be treated in first approximation as independent features.

Step by step procedures are available from HORIBA Scientific and presented during our user trainings.

The sputtering efficiency (SR or q depending on notations) is taken into account - this is expressed as the mass removed per unit of time. It is dependent on the sample (or the layer) and of the operating conditions. It is a material property, not an element property.

The intensity of light for an element is proportional to the concentration of this element in the plasma, and this relates to the concentration in the sample by the simple q factor.

c_iq_M = k_iI_i

c_i composition
q_M sputtering rate
I_i intensity
k_i constant

If the SR is taken into account, one can perform calibration mixing samples from different families.

During a measurement, at any time, all intensities are recorded so all (cc*q) values are calculated, and by summing these numbers (knowing that sum of cc is always 100% at any depth) both the SR and the cc are known at each depth.

So quantification simultaneously gives the concentrations and the mass removed at each time, hence the depth (by calculating densities). In this standard approach, depths are calculated, such as erosion rates. Of course with DiP, these values are measured and therefore the quantification becomes more accurate, as no approximation is made at any step.

A CRM is by definition a Reference Material, accompanied by a certificate, one or more of whose property values are certified by a procedure which establishes its traceability to an accurate realization of the unit in which the property values are expressed, and for which each certified value is accompanied by an uncertainty at a stated level of confidence. CRMs are most often bulk samples, though some layered samples can be found.

For most GD applications, CRMs do not exist and so reference materials – often provided by the customers themselves – are to be used.

For such reference materials, homogeneity can be checked by GD itself, while composition is to be determined by other techniques.

As with all comparative techniques, Pulsed RF GDOES requires calibration to provide concentrations vs. depth (named “QDP” Quantitative Depth Profile or “CDP” Compositional Depth Profile) from the measured Intensities vs Time (Qualitative Depth Profile).

With pulsed RF, all types of samples can be used for calibrations, and not only bulk metallic samples as at the origin of GD. Samples can now be bulk (conductive or non) or layered samples – conductive or non. Oxides can be mixed within the calibrations to provide data points for O, and polymer layers can be added when organic layers are of interest.

For some advanced materials and new materials, it is very hard to get adequate reference materials to do normal quantitative analysis. But HORIBA GDOES instruments are designed to perform quantification with only one sample representative of the material investigated. This is called the layer mode.

The layer mode is a very elegant and the most simple way to calibrate the instrument for a certain application. It only requires knowing one type sample, representative of the material of interest. In the case of PV cells for instance, the availability of one cell, well characterized in thickness and composition, is enough to calibrate the instrument for the application.

The measurement is done by interferometry. A unique laser beam (from a laser diode) is split. The reference beam goes to the intact sample surface near the crater when the sensing beam is directed in the middle of the crater. The reflecting signals are collected and measured in real time providing erosion rates and depth.

Fig 47: DiP logo

The Time Plus function allows increasing the measurement time during the analysis without stopping the source. It is especially useful in depth profiling when unknown samples are run and one realizes that the present timing is not sufficient, and that the source will stop before the last layer is reached. Time Plus is a standard feature of the HORIBA GD instruments.

Each time the sample is removed from the source, the source is exposed to air. The discharge cavity will therefore contain some air every time the sample is changed or moved to a new position, even if the source is flushed with argon during the positioning process, though a constant argon stream should reduce the ingress of air. To obtain a pure argon atmosphere inside the source, after placing the sample, the closed source cavity is first evacuated and then flushed with argon. An effect of flushing the cavity prior to analysis is to reduce the time for the plasma to become stable during the analysis.

Typical questions arise when people look at a GD profile for the first time, notably the qualitative analysis.

The following example is shown during our trainings. Sample is a fake 1 Euro coin.

The central part which should be pure Ni in a real Euro is here a coated material: Ni/Cu/Fe.

Different lines are measured with different detectors and have different sensitivity and after quantification, this issue disappears. GD is a comparative technique.

This is why you cannot easily compare elements from their relative intensities in a qualitative profile. After quantification, this issue disappears.

The second question is how can it be assessed that there is diffusion within a layer? Diffusion and roughness will both lead to enlarged interfaces (also the crater shape, but this is easier to assess with a profilometer, and to correct using adequate conditions).

A single profile does not easily permit differentiation between the two phenomena.

But if one looks at a sample before and after annealing for instance, and overlay the two results, the diffusion can easily be derived.

For this question also comparing samples is the key to proper interpretation of a GD profile.