The laser light emitted using a Nd:YAG laser is generally at 1064 nm in the infra-red range. This wavelength couples easily with samples containing significant quantities of the transition elements. Longerich et al. [1993] incorporated a harmonic generator into the laser apparatus which allowed shorter wavelength (532 nm and 266 nm) laser radiation to be generated. Jenner et al. [1994] determined crystal-matrix partition coefficients for a variety of trace elements using 266 nm wavelength laser radiation and reported a fourfold decrease in the diameter of the ablation pit from that produced at 1064 nm on this particular LA-ICP-MS system. This is important for controlled ablation of transition-element-poor materials (e.g., the minerals calcite and feldspar). However, Abell [1990] noted that materials which are transparent to laser light could be ablated using the 1064 nm wavelength if the laser pulse has sufficient energy. Feng [1994] used this modus operandi to undertake controlled ablation and analysis of carbonates using 1064 nm laser radiation.
The laser may be operated in two modes: (a) ``Q-Switched'' where
a short laser pulse (
10 ns) contains practically all of
the energy; and (b) ``Fixed-Q'' or ``Free-Running'' where the laser
pulse is much longer (120-150 
sec) and the power delivered is
considerably less [see Denoyer et al., 1991, for detailed
descriptions]. The resulting ablation characteristics are very
different and produce very different ablation pits, thus affecting
the size of the sample analyzed. In Q-switched mode, the laser
energy is higher (relative to the free-running mode), and much of
the ablation occurs through total vaporization and mechanical
ablation. Calculated Relative Sensitivity Factors (RSFs) are
relatively uniform across the mass range [e.g., Denoyer et
al., 1991]. In Fixed-Q or Free-Running mode, the power of the
laser is lower, the laser interacts with the sample for a longer
period of time and is conducted more deeply into the sample. This
produces a deeper crater of smaller diameter relative to Q-switched
mode, but the elements are ablated selectively on the basis of
their vaporization energies [e.g., Thompson et al., 1990].
This fractionation produces variable RSFs across the mass range
relative to those produced in Q-switched mode. Generally, the
laser is operated in Q-switched mode.
By its very nature, the signal induced by the laser pulse is a
transient one, thus making tuning difficult even in Q-switched
mode. Hollocher [1993] reported a technique involving the
by-pass of the argon carrier from the sample chamber over a crystal
of iodine held in a glass tube. Iodine is evaporated at room
temperature, is monoisotopic having an atomic weight of 127 which
is in the middle of the mass range, and is relatively resistant to
forming polyatomic species (i.e., ArI
). While the memory of
iodine may be long in the system, if this element does not need to
be quantified and is only used for tuning, such a set up would seem
ideal for LA-ICP-MS.
Detection limits are intimately related to the signal intensity, counting time per element for the ablation mass, and on the sample cell design which affects the size and configuration of the ablation pit and, thus, on the amount of material ablated. The precision of LA-ICP-MS is dependent on signal fluctuations as a result of pulse-to-pulse variations in the amount ablated and hence the amount reaching the plasma [ van de Weijer et al., 1992]. A quantitative analysis of both major and trace elements in geological samples can be obtained by normalizing the intensities of the observed peaks to either the weight of the sample removed or a true internal standard [e.g., Imai, 1990; Denoyer et al., 1991]. Determining the accurate weight of sample removed is an extremely involved process, especially as not all of the material ablated reaches the plasma or collector [e.g., Remond et al., 1990]. Internal standardization removes the need of knowing an accurate volume of material ablated and amount transported to the ICP torch. Also, normalizing signals from the unknown sample to an internal standard concentration removes any change in response with time between analyses [e.g., Pearce et al., 1992 a,b]. However, this requires a knowledge of matrix composition and if it has an isotopic abundance which is less than 1% of the total matrix [ van de Weijer et al., 1992]. Choice of an internal standard is critical in that its behavior during ablation must be representative of the unknown elements being quantified [c.f., Jarvis and Williams, 1993]. If knowledge of the matrix is known, then such data can be used as internal standards. This is of particular significance for geological applications, where major and minor elements are usually determined via other methods (i.e., electron microprobe for minerals and XRF or INA for bulk samples).
The requirement of careful matrix matching in order to obtain quantitative analyses of small samples via LA-ICP-MS is well documented in the recent literature [e.g., Denoyer et al., 1991; Jarvis and Williams, 1993; Williams and Jarvis, 1993]. In a study of pressed powder standard reference materials, Williams and Jarvis [1993] concluded that geological standards for LA-ICP-MS should not only be matched in chemistry, but more importantly in mineralogy. This is a particularly critical observation for the analysis of small geological samples which will tend to be individual minerals. However, it has been demonstrated that if the laser pulse has sufficient energy to ablate the sample via plasma plume expansion and not from absorption of the laser beam with resulting thermal vaporization (and matrix-dependent element fractionation), then nonmatrix matched standards may be used [e.g., Abell, 1990; Jackson et al., 1992; Jenner et al., 1994; Feng, 1994]. Note that all procedures using nonmatrix matched standards are conducted in Q-switched mode which produces a more intense but shorter duration laser pulse (see above).