Page 291 - Inorganic Mass Spectrometry - Fundamentals and Applications

Sample Types
Analysis of Nonco~ductive 277
ized to some extent, until the potential swings to the positive portion of the wave-
form. At this point, negatively charged species (i-e., electrons) are accelerated to
the cathode, effecting an opposing charge reduction, Because of their much lower
to
mass, electrons cause the applied potential approach 0 volt at a much higher rate
than do ions in the negative potential half-cycles. This difference charge mobil-
in
ity acts to displace the applied potential toward an average negative value over a
number of voltage cycles. Thus the potential on the cathode surface alternates about
a dc bias potential wherein there is a negative potential on the sample for SO% of
the rf cycle and a positive potential for a relatively short period. The sample is
therefore bombarded (sputtered) by positive ions for a majority of the time, and
charge compensation by electrons occurs for the remainder of the cycle. In this
way, all of the desired functions of an analytical CD source can be achieved for in-
sulating samples and conductors.
Although rf CD-MS in spectrochemical analysis may have been underuti-
lized to date, there is a substantial body of work regarding the fundamental traits
of these plasmas that supports their use in analytical spectroscopy [58-601. Many
of these studies have dealt more generally with the use of rf sputter deposition
plasmas but are important to our understanding for better source designs and ap-
plications in direct solids analysis. Because these plasmas operate on a different
basic principle than the traditional dc sources, a few of the design and plasma char-
acteristics are presented here. The establishment of the negative dc-bias potential
is key to the ability to sputter the sample cathode and establish an effectively field-
free region about the ion sampling orifice. Coburn and coworkers [58] have stud-
ied the role of the respective electrode sizes within the plasma in isolating sput-
tering to the target (i.e., the analytical sample). Very simply, the maximum dc bias
is obtained when the area anode-to-cathode r is maximized. Therefore, it is ad-
vantageous to power a small sample housed in a relatively large-volume, grounded
anode. Unlike in the case of dc GD-MS sources, the ionization processes in plas-
rf
mas actually maximize at very different pressure8 than the atomization rates. In
general, optimal analyte ion signals are obtained at pressures of an order of mag-
nitude less than in the dc case (hundreds of milliTorrs vs. Torrs) [ 1 1,28--321. The
lower pressure leads to far greater spatial segregation of the plasma processes and
is
resultant analyte ions. Therefore, ion sampling position a key aspect in system
optimization as discussed in the following sections. To a first appro~imation, op-
eration of GD plasmas at lower pressures results in much higher cathode poten-.
tials, but this is not the case of the dc bias in the rf discharges. These values tend
to be less than for the dc, implying that a greater fraction of the applied power is
directed into the gas-phase excitation and ionization processes. Detailed Langmuir
probe measurements in both process and analytical GD plasmas confirm that this
is indeed the case as electron energies and temperatures are much higher in rf
the
discharges 1[59,61]. This is also true for the all-important metastable gas species
that Coburn identified early on as crucial to analyte ionization in deposition plas-
mas [60].