Page 294 - Inorganic Mass Spectrometry - Fundamentals and Applications

of
sources). In a number implementations of this geometry, authors use the phrase
~to the comon CD-AES source. Although the
~
-
~
G ~ ~ geometry by analogy ~ e
Grim-type discharge does indeed employ an obstructed electrode arrangement,
most GD-MS sources are not truly of this geometry as auxiliary pumping between
the cathode and anode is not employed as in the AES case.
of
In concluding the discussion of rf GD-MS source geometries, the number
works involving the use of magnetic plasma enhancement methods cannot be ig-
nored [63-661. These “magnetron” arrangements have their roots firmly grounded
in the plasma deposition science and engineering literature [57]. ~agnetron GD
sources generally employ concentric permanent magnets (>1OO-G field strength)
located behind the cathode/target and thus not exposed to the plasma. Magnetic
fields permeate the cathode and trap plasma electrons in helical orbits close to the
surface, decreasing the atom-electron mean free paths and thus increasing the over-
GD
all plasma ionization efficiencies. As a result, magnetron enhanced sources can
operate at two to three orders of magnitude lower pressures than standard rf GD
of
ion sources (single vs. hundreds millitorrs). This greatly reduces the entire sys-
hold
to
tem vacuum pumping requirements. Lower operating pressures would seem
the promise of lower amounts of molecular ion species in the spectra and perhaps
higher ion signal fluxes as larger differential pumping apertures could be employed.
Interestingly, the groups describing the use magnetron rf GD-MS sources
of
each used different mass analyzer systems: Hecq and coworkers [63,64], a quadru-
pole mass filter; Saprykin et al. 1651, a double-focusing instrument; and Hieftje and
coworkers [66], a time-of-flight (TOF) spectrometer. Each group indeed found that
lower source pressures could be utilized than in “normal” powering, though only
rf
Saprykin [65] reported obviously lower levels of polyatornic ion contributions.
Hieftje E661 compared molecular and discharge gas ion ratios to those of analyte
elements, with and without magnetic coupling. Lower amounts of molecular
in ion sampling position, with the
species were eventually attributed to differences
magnetron producing AdCu ratios that were 15-20 times higher than in the %or-
mal” rf source operation. Although magnetron use was proposed as a means to re-
duce the deleterious effects of nonconductive sample thickness by virtue of en-
hanced plasma density, differences in magnetic field strength permeating through
to the sample surface (even for conductive samples) add yet another sample-de-
pendent variable to complicate quantification,
In addition to the differences in magnetic coupling described, two other lim-
itations hinder the use of magnetron geometries: (1) the sample must be in a disk
form and (2) the sputtered crater is a circular “track” rather than a flat surface.
Saprykin E651 presents profilometer tracings of “normal” and magnetron craters
that illustrate that although the sputtering rate is higher with magnetic coupling,
the ability to perform depth profiling is greatly compromised. Another aspect of
the circular sputtering track was indirectly addressed by Hecq and coworkers [64],
who sampled ions in the normal axial (end-on) and radial (side-on) directions. Very