OES

Is Spatial Resolution really necessary when measuring plasma emission?

This is not a difficult question to answer since all the microscopic properties of the plasma are space and energy dependent. In addition, the time dependence (during the rf cycle) of many parameters is often important.

However, most of the times even the space dependence of the time-averaged properties (like emission intensity) is ignored. In fact almost all of the existing literature is based on optically-averaged or single point emission measurements.

Most of the times the optics of the experiments are not even sufficiently described. Sometimes the trends of emission intensities of various species, as a function of a macroscopic parameter (power, pressure, etc), or even worst as a function of a characteristic of the material deposited or treated in the process, is used for establishing relationships or for drawing conclusions concerning the mechanism of the process!!

Why is this wrong?

This is better explained by using an example:

The first figure shows the spatial distribution profile of SiH* emission from highly diluted silane in hydrogen rf discharges at 0.5 Torr, recorded with a resolution better than 0.5 mm.

One can easily observe that the various features of the profile behave differently when increasing the RF voltage, depicting the various effective electron heating mechanisms.

The second figure shows the trend of emission intensity at several different points in the discharge space along with the total emission intensity, calculated by integrating each curve in space. (Note that Total Intensity resulting from the integration of the spatial distribution is much more accurate than an optical average)

It is obvious that there are many different trends that could be used to fit any parameter (say the deposition rate).
The question is which one would you choose?

One could add here many more examples demonstrating that this malpractice is a big source of errors. However, even more important are....
 

The advantages of spatially distributed emission measurements:

Among many things, one can:

- visualize electron heating

- measure sheath lengths

- observe plasma-surface interaction

- calculate accurately fluxes of species to the surfaces   

- correlate with the effective electron density for the specific process
   

In many cases they can be used as ionization or radical generation profiles, permitting thus to avoid laborious calculations