Page 295 - Inorganic Mass Spectrometry - Fundamentals and Applications

simply, the inhomogeneous magnetron plasma requires greater spatial selectivity
than the more diffuse plasma produced in the “normal” powering mode.

The major driving force for the development of rf-powered GD-MS sources is of
of
course the broad diversity possible analytical samples to which the devices may
be applied. It should be noted at the outset that a number of the cited works have
equal
shown that the performance characteristics of the sources are to or better than
that of dc GD-MS for metallic, conductive samples. In the discussion that follows,
the use of rf GD-MS is highlighted for the analysis of bulk insulators, oxide pow-
ders, and polymeric materials.
The analysis of bulk nonconductive materials by rf GD-MS has been un-
dertaken with greater emphasis on source development and characterization than
on pure quantitative analysis within specific analytical systems. For the most
part,
of
studies by Marcus and coworkers have demonstrated general analytical figures
merit that could be attained for given source geometries on commercial GD-MS
systems [28-30,331. As such, the results are something of benchmarks relative to
dc source operation, but for the case of nonconductive samples in rf powering.
be
Throughout these studies, a few general observations can made; they are for the
most part consistent throughout the GD-MS literature cited in this section. These
rf
to
traits include (l) lower operating pressures relative those of dc powered systems
(hundreds of milliTorrs vs. Torrs), (2) high degree of signal and spectral respon-
sivity to changes in discharge conditions and ion sampling position, (3) faster
plasma stabilization times for metals relative to dc discharges (single minutes ver-
sus tens of minutes), and (4) an inverse relationship between nonconductive Sam-
ple thickness and ion signal intensities. The latter factor would be expected to pro-
vide severe limitations to quantification. Although these limitations have been
successfully addressed in this laboratory [29], it is still true that absolute sensitiv-
ity is sacrificed as thicker samples are analyzed.
Many of the primary analytical characteristics of rf GD-MS analyses of flat
glass samples are demonstrated in the data presented in Table 7.3 [29]. These data
were obtained for NIST 610 Trace Elements in Glass samples that are in the form
of l-m-thick, 10-m-diameter disks analyzed on a VG GloQuad insfo ford,
cell.
Cheshire, IJK) spectrometer system equipped with a flat sample holder In gen-
eral, the ion signals for the analyte elements of these samples stabilize to better
l60+ signals take as much as min to reach
than 7% RSD in <20 min, whereas the 50
a steady state. The data presented in Table 7.3 illustrate the stability of the ele-
mental ion signals (ratioed to the 2sSi+ internal standard) over what would be con-
sidered a typical 45-min analysis time. Interestingly, only the 56Fe+ signal shows
a variation over 2.5 RSD. It is expected that this value is elevated as a result of the
presence of the isobaric ArO+ species. Si-referenced RSFs produce elemental con-
centration accuracies of less than 6% error, except in the case of Fe. The pooled