Tải bản đầy đủ - 0trang
II. ICP-AES and ICP-MS Instrumentation
PARVIZ N. SOLTANPOUR ETAL.
Figure 1 Magnetic fields (H) and eddy currents (shaded) generated by high-frequency currents (I)
flowing through coil. (Adapted from Fassel and Kniseley, 1974.)
gas at about 15-17 liters min-’ to the plasma to sustain the plasma and to isolate
the quartz tube from high temperatures. The innermost channel is for introducing
sample into the plasma. The middle channel conducts the auxiliary Ar gas at about
1 liter min-l and is used in ICP-AES only when starting the plasma or for organic samples; it is routinely used for all types of samples for ICP-MS (Fig. 3). The
ICP has an annular, or donut, shape when it is viewed from above. The hole has a
lower temperature than the donut body and offers less resistance to the sample injection. The sample is injected into the plasma by using Ar carrier gas at a rate of
about 1 liter min-’ for ICP-AES work. For ICP-MS work, the aerosol flow is approximately 1.5 liter min- l .
The ICP generater has unique physical properties that make it an excellent
source for vaporization, atomization, ionization, and excitation of elements. For
ICP-AES, the aerosol droplets containing the analyte are desolvated, the analyte
salts-oxides are vaporized, and the analyte is atomized at the high temperature region of the plasma in the vicinity of the Cu coil (Fig. 2). An initial radiation zone
(IRZ) has been defined by Koirtyohann et al. (1980) as the zone that begins in the
sample aerosol channel inside the load coil for ICP-AES (Fig. 4). The IRZ extends
upward to 1-2 mm above the load coil, taking on the appearance of an amber “bullet” during nebulization of many sample types related to agriculture. This is due
ADVANCES IN ICP EMISSION AND ICP MASS SPECTROMETRY 3 3
Figure 2 Temperatures in the plasma as measured by the spectroscopic slope method. (Adapted
from Fassel, 1977.)
to emission from CaO molecules on the surface of the “bullet,” with the color
changing to a deep blue or purple further downstream as emission from calcium
atoms and ions dominates. The blue-purple region is termed the normal analytical zone (NAZ) and is the region in which the analyte emission is observed by the
spectrometer (Fig. 4). Color photographs illustrating the appearance of the IRZ
and NAZ while nebulizing an elevated concentration of Y into an ICP have recently been published for ICP-AES (Winge ef al., 1988) and more clearly define
these critical regions. The NAZ is 15-20 mm above the coil, or about 14-19 mm
above the tip of the IRZ, in an environment relatively low in background emission.
The background consists of Ar lines and some weak band emission from OH, NO,
and CN molecules present in the plasma (Ward, 1978a). By the time the decomposition products of the sample reach the NAZ, they have had a residence time of
about 2 msec at spectroscopically measured temperatures ranging from about 8000
to 5000 K (Fassel, 1977). The residence time and temperature experienced by samples introduced into the plasmas are about twice as large as those in the hottest
flames, e.g., N,O-C,H,. The combination of high temperature and residence time,
at the sample aerosol flow rates typically used in ICP-AES, leads to complete sample vaporization and atomization in contrast to flames that require releasing agents
for refractory compounds (Larson et al., 1975). Once the free compounds, atoms,
and ions are formed in ICP-AES, they are in a chemically inert environment in
contrast to environments with highly reactive combustion flames. Ionization interferences are generally negligible in an ICP-AES experiment. Self-absorption (a
phenomenon responsible for the flattening of the standard curve at high analyte
PARVIZ N , SOLTANPOUR ET AL.
Aerosol ' Carrier
Figure 3 Typical quartz torch and inductively coupled plasma configuration. Flow A is auxiliary
flow used for organic samples. (Adapted from Fassel and Kniseley, 1974.)
concentrations) is practically absent, which leads to a wide linear dynamic analytical range of 3-5 decades. No sampling or skimmer cones and lense stack or
quadrupole rods are used in the ICP-AES, and, therefore, contamination from ablative processes off of them, e.g., secondary ion sputtering, is absent.
For ICP-MS, the vaporization and atomization begin at approximately the same
location relative to the load coil as do these processes in the ICP-AES, in a relatively hot region of the plasma in the vicinity of the Cu coil (Fig. 2). However, the
flow rates of sample and/or auxiliary argon are increased for ICP-MS to obtain an
analytically useful population of ions (Winge et al., 199 I), while keeping the sampling cone a safe distance from the load Cu coil to prevent arcing between the cone
and the load Cu coil. The IRZ extends well beyond the downstream side of the load
Cu coil (Fig. 4).The water droplets produced in a conventional concentric nebulizer, although apparently extremely few in number compared to the total number
of aerosol droplets produced, can survive the rigorous desolvation-atomization
conditions generated by the ICP (Winge et al., 1991).Although the downstream
side of the load coil-to-IRZ tip distance varies from one lab to another, it is generally between 10 and 20 mm for ICP-MS. Unlike ICP-AES, with ICP-MS this
leaves much of the analyte vaporization and atomization to be done in regions be-
ADVANCES IN ICP EMISSION AND ICP MASS SPECTROMETRY 3 5
Ion-atom emission =
Figure 4 Spatial nomenclature for ICP. In agronomic work, the initial radiation zone (IRZ) will
likely appear red to pink due to emission of CaO and CaOH molecules. The location of the normal analytical zone (NAZ) depends on whether analytical measurements are being performed by (A) AES or
(B) MS.Pressures: p, = 760 torr; pa = 1.2 torr; p3 = lo-’ to
torr. Cones: S, = sampling cone;
S, = skimmer cone. (Adapted from Koirtyohann et al.. 1980.)
yond the hottest parts of the ICP. The sampling cone orifice defines the NAZ in the
ICP-MS and is another 2-10 mm downstream from the tip of the IRZ (Fig. 4). In
the DANR Analytical Lab, the IRZ extends approximately 19 mm downstream
from the spectrometer side of the load coil, and the sampler cone orifice is positioned another 3 mm downstream from the IRZ tip; which results in placement of
the NAZ a total of 22 mm from the nearest surface of the load coil. Most of the
particle beam is sucked through the sampling cone into the intermediate vacuum
region of a differentially pumped aperture approximately 2-3 mm from the tip of
the bullet. The tip of a second cone, called the skimmer, is immersed in what is
termed a barrel shock (Gray, 1989) that results from supersonic expansion of the
plasma gas as it passes from atmospheric pressure through the sampling cone orifice into a vacuum of about 1 torr. The kinetic temperature of the gaseous particles
at the tip of the skimmer cone is 2200 K (Lim et al., 1989; Winge et al., 1991). Although the position of the sampler with respect to the extended IRZ of the ICP results in a maximum rate of ions per second at the detector, it is also sampling
aerosol that has undergone solute vaporization and atomization reactions outside
PARVIZ N. SOLTANPOUR ETAL.
the hottest regions of the ICP. This is thought to contribute to the appearance of
more molecular ions in the mass spectra and higher susceptibility to nonspectroscopic matrix effects than if the aerosol flow rate and/or auxiliary argon flow rate
could be slowed down enough to put the IRZ back to within 1-2 mm of the downstream side of the load coil. However, this is not possible because of the arcing
that occurs between the load coil and the metallic sampling cone in instances in
which the cone is placed too close spatially to the load coil. We have unsuccessfully located descriptions of ICP-MS experiments designed to reduce molecular
ion formation in the mass spectrum using a sampler constructed of a samplingcone
that does not conduct electricity. Among the possibilities for nonconducting materials are high-tech ceramics that can withstand prolonged exposure to the highest temperature regions of the ICP. These include AIN, Sic, A1,0,, or zirconia ceramics.I The sampler could be placed so that the NAZ is in a region closer to local
thermodynamic equilibrium (LTE) with respect to maximized ion populations
while the analyte solute vaporization and atomization is allowed to proceed in the
hottest parts of the plasma (Fig. 4).
In general, the NAZ is much closer to the tip of the IRZ in ICP-MS (2-10 mm)
than the NAZ is to the tip of the IRZ in ICP-AES (14-19 mm). The closer proximity used for the ICP-MS measurements increases the concentration of ions to an
analytically useful level (Winge et al., 1991). Ideally, ions should be extracted
from a region that approximates local LTE. Apparently, ion temperatures are sufficient to support high ion populations this close to the IRZ tip. Undoubtedly, the
requirement for high ion density at a distance well downstream from maximum
gas and excitation temperatures promotes formation of metal oxide ions and nonspectroscopic concomitant suppression effects that are observed in the ICP-MS. A
number of modifications mentioned later, most involving the usual sample introduction techniques, have significantly reduced these problems.
Nebulizers are devices used for the injection of the sample into the plasmas.
There are three general types of nebulizers: pneumatic nebulizers, Babington style
nebulizers, and ultrasonic nebulizers (USNs) (Thompson and Walsh, 1983).Pneumatic nebulizers use the Venturi effect to draw sample solutions into the spray
chamber. Babington-style nebulizers require a pump to deliver the solution to a
pinhole orifice from which argon gas is emerging at high velocity. The USN also
'Coors Ceramics, 600 Ninth St., Golden, CO 80401.
ADVANCES IN ICP EMISSION AND ICP MASS SPECTROMETRY 3 7
requires a pump to deliver the solution, this time to a vibrating plate.
There are two common types of pneumatic nebulizers: cross-flow and concentric. For the cross-flow, as the solution emerges from the rigid capillary tube carrying the sample solution, another tube positioned at a right angle blasts argon past
it to shear off fine aerosol particles. Cross-flow nebulizers are often made of highly corrosion-resistant capillary metal tubes, e.g., Pt-Ir alloy. One capillary carries
Ar at approximately 1 liter min-l and the other carries the sample solution. The
orientation of the tips is fixed by the manufacturer and may include a sapphire edge
at the tip of the solution tube to produce a fine, uniform mist out of approximately 10% of the solution drawn in. The cross-flow systems in the authors' laboratories have held up to the most demanding applications for 2-3 years with no sign
of degradation.The concentric-flow (Meinhard-type)glass nebulizers are routinely used at the DANR Analytical Lab for both ICP-AES and ICP-MS work. These
are made entirely of glass in a T-type configuration. The main barrel of the nebulizer consists of a fine glass tube tapered to capillary size. The capillary portion
carries the sample solution, is approximately 1 in. long, and is surrounded by a
larger diameter tube carrying Ar. The Ar enters through a tube joined in a T shape
to this barrel. The Ar pressure is 241.5-345 KPa (35-50 psi) and flowing at about
0.75-1.5 liters min-'. The open ends of the Ar tube and the capillary tube meet at
a taper, and a fine mist is produced as the Ar flowing concentrically around the capillary shears off small fragments of water droplets at the capillary tip. These nebulizers are very steady and produce aerosol from about 10% of the solution going
through the tip.
The cross-flow and concentric nebulizers clog with high salt solutions. Soltanpour et al. (1979a) treated 1M NH,HCO,-O.OOSM DTPA (diethylenetriaminepentaacetic acid) soil extracts with 0.5 N HNO, to overcome clogging. However,
the Colorado State University Soil Testing Laboratory (CSUSTL) currently uses
a Legere2 Teflon nebulizer (Babington type) attached to a peristaltic pump that
eliminates the need for acid pretreatment. Wolcott and Butler (1979) designed a
pneumatic nebulizer that could aspirate solutions containing up to 36% suspended solids. To overcome differences in surface tension, density, and viscosity, the
analyst can use a peristaltic pump to introduce sample solutions into the nebulizer (Beasecker and Williams, 1978). For concentric nebulizers, care must be taken
to eliminate small insoluble particles from test solutions that would otherwise clog
the capillary. If a particle becomes lodged in the capillary or between the capillary
and the tapered tip, great care must be exercised while removing the blockage to
avoid breaking the fragile glass tubing. One method is to carefully remove the nebulizer from the Ar and sample delivery tubes and squirt acetone from the nebuliz-
2Distributed by Burtec Instrument Corporation, P.O. Box 235, Delmar, NY 12054.
PARVIZ N. SOLTANPOUR ET AL.
er tip into the barrel while tapping with a finger, then force Ar through the tip backwards while continuing to tap.3 This has been used successfully by one of the authors to remove a microscopic piece of glass fiber that became lodged between the
tip and the inner concentric capillary.
In Babington nebulizers (Suddendorf and Boyer, I978), aerosol is produced
when the solution is pumped into a V groove and is ruptured by gas coming from
a small hole in the groove. Glass frit nebulizers and the like (e.g., the Hildebrand
use the same principle as in the Babington except with many orifices.
These nebulizers can be used for high-salt solutions. Since no constricting orifices
are needed to produce aerosol, they are relatively free of clogs. For pneumatic and
Babington nebulizers, larger droplets settle out in the spray chamber and drain off,
leaving the finer aerosol droplets suspended in the flow stream of Ar that is transported to the plasma.
In USNs, transducers are used to produce the sample aerosol. Compared with
pneumatic nebulizers, USNs improve the detection limit of ICP spectrometers by
one to two orders of magnitude (Olson et al., 1977). A three to four order-of-magnitude improvement in ICP-MS detection limits has been noted using the USN
with a high-resolution, double-focusing ICP-MS instrument (Tsumura and Yamasaki, 1991). The USNs are operated with a sample aerosol desolvation system
that follows aerosol production by the transducer. The aerosol desolvation system
is a heating assembly followed by a condenser column. Thus, factors involved in
improved analytical performance of the ICP-MS with use of the USN observed in
Tsumura and Yamasaki (1991) are ( 1 ) improved sample transport to the plasma,
(2) reduced water vapor present in the aerosol introduced to the plasma (Hutton
and Eaton, 1987), (3) reduced oxygen and hydroxide present as reactive species in
the differentially pumped interface (Gregoire, 1989; Lim et al., 1989; Veillon and
Marghoshes, 1968), and (4) reduced background as a result of reduced oxygen and
hydroxide levels in the spectrum (Gregoire, 1989). Coupled with the high-resolution of the double-focusing mass spectrometer, detection limits achieved by
Tsumura and Yamasaki (199 1) are in the low parts-per-quadrillion range.
Ultrasonic, pneumatic, and Babington nebulizers can all be used with ICP-MS
instrumentation. In fact, any nebulization system used for ICP-AES can be used
for ICP-MS. Because of the severity of nonspectroscopic concomitant effects on
analyte ion arrival rate at the detector-per-unit analyte concentration-i.e., analytical response (Houk and Thompson, 1988; Gregoire, 1987a,b; Beauchemin et al.,
1987; Olivares and Houk, 1986; Douglas and Ken; 1988) encountered in routine
aqueous nebulization ICP-MS-variations on the usual aqueous sample aerosol
generation and introduction systems are more common in the ICP-MS area. Some
'Ken, Petrie. Precision Glassblowing of Colorado. 14775 East Hinsdale Ave., Englewood, CO
801 12. pers. comm.
4Leeman Labs Inc., Wentworth Dr., Hudson, NH 03051.
ADVANCES IN ICP EMISSION AND ICP MASS SPECTROMETRY 39
of the alternate methods of sample aerosol production and/or sample injection are
1. Hydride generators (Workman and Soltanpour, 1980; Thompson er al.,
1978a,b; Ek et al., 1991).
2. Laser ablation (Denoyer, 1991; Hager, 1989; Abell, 1991; Denoyer et al.,
1991b; Pearce et al., 1992).
3 . High-performance liquid chromatography (HPLC, including ion chromatography) (Braverman, 1992).
4. Liquid-liquid solvent extraction (Plantz, 1989; Serfass et al., 1986).
5. Flow-injection (FI) analysis (Thompson and Houk, 1986; Dean et al., 1988,
Denoyer et al., 1991a; Denoyer and Stroh, 1992).
6. ETV (Gregoire, 1989).
7. Aerosol-desolvation apparatus (Veillon and Margoshes, 1968).
8. Direct-injection nebulizers (DIN).5
9. Direct-insertion devices (Gervais and Salin, 1991).
10. USN systems (Olson et al., 1977).
Hydride generators (Workman and Soltanpour, I980), laser-ablation systems,
DIN, and flow-injection principles are discussed in the following sections and the
other systems have been described in the above references.
2. Hydride-Mercury Vapor Generator
Certain elements, when reduced by NaBH,, form gases that can be directly introduced into the plasma. Arsenic (As), Sb, Bi, Se, and Te are thus reduced to form
hydrides, and Hg is reduced to Hg vapor. This method of sample introduction
(Fig. 5) greatly improves the detection limits of these elements compared with
pneumatic nebulization because of an improvement in sample delivery and a decrease in matrix effect. Thompson et al. (1978a,b) simultaneously determined As,
Sb, Bi, Se, and Te by use of ICP-AES and a hydride generator. Studies at CSUSTL
indicate that by reducing As and Se to their hydrides and Hg to its vapor form and
introducing these gases into the ICP, they can be quantitatively detected at 1.O, 0.5,
and 0.5 pg liter-' of these elements (Workman and Soltanpour, 1980). Recently,
Ek et al. (199 1 ) have used an analogous system with ICP-MS instrumentation to
improve Se detection limits to 0.05 pg liter- I .
3. Laser Sampling of Solids
Many solid samples are difficult or time-consuming to put into solution, e.g.,
soils and ceramics. Sometimes the elemental composition of grain features and
'Transgenomic/CETAC Technologies, Inc., 5600 South 42nd St., Omaha, NE 68107.
PARVIZ N. SOLTANPOUR ETAL.
small inclusions in the solid are of greater interest than the overall composition,
e.g., minerals. To save time in sample pretreatment and to permit feature analysis,
surface-sampling methods using a laser have been developed (Denoyer, 1991;
Hager, 1989;Abell, 1991; Denoyer et al., 1991b; Pearce ef al., 1992). Laser ablation can be used in conjunction with ICP-AES, but most frequently the ablated
aerosol is injected into an ICP, and the ions produced are subsequently detected using a mass spectrometer. Two of the manufacturers of ICP-MS instrumentation
market a laser-ablation accessory.6 The accessory is equipped with an NdYAlG
(neodymium-yttrium-aluminum-garnet) laser and an ablation stand. The ablation
stand has an X-Y-Z translational specimen stage that is moved under computer
control. Vendor software supports time-resolved data acquisition and semiquantitative analytical reports. Laser repetition rates are adjustable from a single shot to
hundreds of bursts per second. Beams can be used in a defocused mode to cover
approximately 1 mm of surface area or sharply focused to less than 0.02 mm
(Pearce et al., 1992). The time durations and number of repeating shots are operator selectable. The amount of energy per pulse is variable. A threshold energy is
required to fire the laser. The upper limit on repetition rate and energy per pulse is
set either by the limitations of the laser output or by the window-material-degradation threshold.A typical pulse can be as short as a few nanoseconds (Q-switched)
and delivers approximately 0.1 J of energy.
In operation, the sample Ar flow to the ICP is momentarily interrupted while the
ablation stage cover is removed, the sample specimen placed on the Teflon ablation stage, and the ablation cover replaced. The sample Ar flow is then resumed,
and the portion of the sample to be ablated is located within the ocular of a light
microscope. The specimen is focused using the X-Y-Z movement of the sample
stage. The computer is notified of impending analysis, and the laser is fired. Preablation times, Laser repetition rates, and laser-power per pulse are important variables. The ablation stage is disk-shaped, with the circular top surface used to support the sample. A metal tube protrudes through the disk and serves to supply an
Ar flow into the sample area. A groove in the side of the disk is used to seat an 0
ring. The ablation cover makes a gas-tight seal with the 0 ring. The cover, resembling an upside-down glass beaker, is approximately 5 cm in diameter and height.
A glass sample aerosol exit tube protrudes from the side, toward the top, of the
sample cover. The cylindrical side of the sample cover and sample aerosol exit tube
are constructed of heavy gauge glass, and the top surface of the sample cover is
made of a glasslike material that is transparent to laser light. A relatively low-power light microscope is used for viewing the specimens, requiring a high-intensity
lamp inside the ablation stand next to the sample cover to illuminate the specimen.
%G Elemental Inc., 27 Forge Parkway, Franklin, MA02038; Perkin-Elmer Sciex, 761 Main Ave.,
Norwalk, CT 06859-0012.
ADVANCES IN ICP EMISSION AND ICP MASS SPECTROMETRY 41
A video camera is sometimes used to project the ablation process onto a television-type screen. The laser, light microscope, and video camera can all be made to
focus on the same point in space. In many applications, however, the light microscope and video camera focuses are set to coincide with some point on the sample
surface, and the laser focus is set 1-5 mm deeper.
The detection limits of metals in the solid are usually less than 1 pg g-' with
the laser setup, and the elemental coverage is superior. The dry-sample aerosol produced by the laser bursts is free of many of the recombination polyatomic ions that
would ordinarily accompany a major element (M) in a nebulized sample (e.g.,
MO+, MH+, MOH+; see Date et al., 1987). Argide polyatomic ion species, e.g.,
MAr+, may persist, however. The sample analysis rate can be rapid, but it depends
on the analytical objectives and the variability between samples. The accuracy of
the analyses is highly dependent on the availability of certified materials of composition similar to the sample. At ultra low concentrations, memory effects must
be considered. For example, assume that a gold nugget is to be ablated to determine approximate elemental composition. On the next sample an elemental assay
for gold content is requested on a metallic inclusion in a piece of quartz. To reduce
the gold background between the two samples, the entire ICP-MS system should
be shut down to permit thorough cleaning of the sample and skimmer cones, the
ICP torch, the aerosol carrier line from the Laser stand, and the interior of the glass
sample stage cover. Cleaning the glass sample stage cover is probably the most
critical process, because the interior of the laser-ablation window becomes coated
with a metallic film of elemental composition generally representative of the ablated sample, and re-ablation of the film can occur during ablation on subsequent
samples. Thus, for the analysis problem at hand, the total analysis time can be a
few minutes or a few hours, depending on whether the quartz piece can be run
ahead of the gold nugget and, more generally, on the detection limit and accuracy
4. Direct Injection Nebulizers
Direct injection nebulizers (DIN) provide for the direct injection of microvolume aqueous liquid samples into the base of torch plasma using fused silica capillary tube and a high-pressure HPLC type pump. A DIN can be used on either ICPAES or ICP-MS instrumentation. It is uniquely suited for determining nebulizer
memory-prone mass isotopes of B, Hg, I, C, S, and Br or where sample volume is
limited. Smith et al. (1991) have shown it capable of detecting as little as 1 ng/g
of B in biological materials using ICP-MS. Powell and Boomer (1995) have shown
the technique under optimal conditions accurately capable of detecting Cr at the
30 ngfliter concentration range for multiple Cr"' and Cr"' species. In addition, the
technique provides for fast sample washout and high sample throughput.
PARVIZ N. SOLTANPOUR ETAL.
A. ICP-ATOMIC EMISSION
Atoms of elements in a sample when excited emit light of characteristic wavelengths with an intensity directly proportional to the element concentration. The
light is focused on the entrance slit of the spectrometer to illuminate the diffraction grating. The diffraction grating separates light into its component wavelengths
of lines (spectrum). The spectral line of an analyte passes through the aperture of
an exit slit and strikes a photomultiplier tube. Photomultiplier tubes produce signals directly proportional to the intensity of the spectral line. The signal is fed to
the readout system, which displays intensities, concentrations, or both. Readout
systems are computer controlled. The computer stores the intensities of standards
and uses these data to calculate the concentrations of unknowns. Systems are available that check calibration-curve accuracy periodically, so that if the quality control (QC) limits are exceeded, the system automatically updates the calibration.'
If the system is equipped with tandem nebulizers,8 one nebulizer could be shut
down by the computer if clogging or another irrecoverable error has occurred,
leaving the second nebulization system to finish running the samples. If the sensitivity is degraded beyond prescribed limits, or if the run is finished, there are commercially available systems that automatically shut down the ICP generator, Ar
flow, and other system functions.
Two types of spectrometers are commonly used (Slavin, 1971): (1) direct-reading polychromators (direct readers) and (2) scanning monochromators. Some systems are equipped with both spectrometers.
Direct readers are designed to reduce the possibility of unwanted light reaching
the photomultiplier tubes. The refractor plates used for fine alignment of the spectral lines are also filters that exclude stray light. The exit-slit assemblies of the photomultiplier are protected by a light shield, and the internal surfaces of the spectrometer are blackened to reduce reflections.
Scanning monochromators use a variety of techniques to make a wide range of
useful analytical wavelengths accessible. Fixed or movable gratings, single or
multiple detectors, and movable entrance and exit slits are a few of the options
available from a variety of manufacturers. The scanning is computer-controlled,
fast, and accurate. In a recent demonstration for the DANR Analytical Lab, one
manufacturer was able to produce 150 elemental concentrations per hour on a set
7Thom Zalinski, Thermo Jarrell Ash/Baird Corporation, 27 Forge Parkway, Franklin, MA 02038,
8Jim O'Dell, Leeman Labs, Inc., 6 Wentworth Dr., Hudson, NH 03051, pers. comm.