Page 49 - Inorganic Mass Spectrometry - Fundamentals and Applications

Glow ~ischar~e Mass Spectrometry 39

Most sputter yield data, involve tightly focused, highly energetic primary ion
beams; this is the basis for secondary ion mass spectrometry (SIMS) (see Chapter
4). Because the physical processes are identical in the two techniques, sputter
yield data reported in the SMS literature are often used to explain the parameters
influencing glow discharge sputtering. One should recognize, however, that inci-
dent ion energies are much lower in the glow discharge than in SIMS, and that
there is a significant amount of sample redeposition in the GD due to the ambient
fill pressure. For a detailed discussion of the influence of these parameters the
reader is referred elsewhere [22,36,37]. In general, the sputter yield increases as
the ratio of the mass of the incident ion to the mass of the target atom ( ~ ~ / ~ ~ )
approaches unity, as the angle of incidence moves away from normal (up to 70°),
and as the incident ion energy increases. The dependence of sputtering on the
target material is a slightly more complicated relationship. The yield is seen to
increase going from left to right in any one row of the periodic table. Carter and
Colligon [38] and Wehner [36] have shown that, with minor exceptions, the
sputter yield closely follows the state of the electron concentrations in the atoms’
“d” shells (i.e., the greater d-shell filling, the great the sputter yield). This
phenomenon is believed to be related to the penetration depth of primary ions. In
targets with more open electronic structures, ions penetrate to such depths that the
transmission of energy back to the surface where sputtering occurs is less efficient.
In filled d-shell atoms, the penetration of ions is relatively small and energy is
more readily projected back toward the target surface, resulting in larger sputter
yields.
Besides the general sputtering theory that was developed to accompany
experimental ion beam data [39], several mathematical models have been devel-
oped [40-421. The success of these models to the physicist is measured by how
accurately they predict sputter yields; to the rest of the scientific community, the
success of these models is measured by how well they convey the sputtering
theory to their audiences. Sophisticated computer programs, often employing
Monte Car10 calculations and requiring large amounts of computer memory, have
generated data about primary ion penetration depth, sampling depth, and colli-
sional transfer of energy, as well as information about the nature of the ejected
species. To this author’s knowledge, however, no satisfactory model has been
developed to explain glow discharge sputter/ato~zation.

Once cathodic species are atomized, sufficient energy must be imparted to ionize
them. Collisions occurring throughout the discharge volume are central to main-
taining the stability of the discharge. However, only those collisions that occur
in the negative glow region (less than one mean free path from the ion exit orifice)
produce a significant enough population of ions to perrnit trace elemental analysis.
Figure 2.4 illustrates the three principal types of collisions in the negative glow