Document 261692

Richard F. Browner
School of Chemistry
Georgia institute of Technology
Atlanta. Ga.30332
Andrew W. Boorn
55 Glen Cameron Road, X202
Thornhill, Ontario L3T 1P2. Canada
Sample lntmduction
Tihniques fior
Selection of the best sample introduction procedure for an analysis requires consideration of a number of
points. These include:
the type of sample (e.g., solid, liquid, gas),
the levels, and also the range of levels for the elements to be determined,
the accuracy required,
the precision required,
the amount of material available,
the number of determinations required per hour, and
special requirements, such as
whether speciation information is
The measurement techniques available, whether flame atomic absorption
spectracopy (FAAS), inductively coupled plasma (ICP) atomic emission
spectroscopy, or dc plasma (DCP)
atomic emission spectroscopy,will
also have a major effect on the choice
of the procedure selected.
Although this paper will concentrate specifically on sample introduction techniques, in any real analysis
sample introduction is an extension of
sample preparation. As a consequence,
the selection of a suitable sample introduction technique can depend
heavily on available and effective sample preparation procedures. Generally,
though, sample preparation will not be
discussed here, except where sample
introduction and sample preparation
are intimately linked.
0 I984 American Chemical Society
. . . . .
T o understand the limitations of
practical sample introduction systems
it is necessary to reverse the normal
train of thought, which tends to flow
in the direction of sample solutionnebulizer-spray chamber-atomizer.
and consider the sequence from the
opposite direction. Looking a t sample
introduction from the viewpoint of the
atomizer, the choice of procedure will
hinge on what the atomizer can usefully accept. Bearing in mind that every
atomizer has certain reasonably well
defined, but different, properties of
temperature, chemical composition
etc., an introduction procedure must
he selected that will result in rapid
breakdown of species in the atomizer,
irrespective of the sample matrix.
To ensure efficient free atom production, the following parameters
must be known for each analyte-matrix-atomizer combination:
maximum acceptable drop size,
optimum solvent loading, both aerosol and vapor,
maximum acceptable analyte mass
appropriate gas flow patterns for effective plasma penetration (for the
ICP), and
suitable observation height.
This last parameter should be selected in conjunction with the gas flow
pattern of sample introduction such
that adequate residence time is provided for the introduced material to
desolvate, vaporize, and atomize. In
certain cases, for instance, when organic solvents are introduced to an
ICP. it also may be necessary to adjust
the atomizer operating characteristics
to account for the change in plasma
properties induced by the solvent.
Here, an increase in forward power to
the plasma (e&, from 1.25 to 1.75 kW)
is necessary to aid the decomposition
of organic species.
Overall, then, it is the properties of
the atomizer that dictate the design
and operation of the sample introduction system. This is particularly true
for liquid sample introduction with
pneumatic nebulization.
Present Underbtanding of Sample
introduction Processes
There is a great deal of intuitive,
but relatively little experimentally
based, knowledge in this field. Clearly,
there is some upper limit to the size of
drop that can be vaporized in the typically 1-2 ms available in the atomizer.
Yet there are no tables available that
specify the upper limit of drop size
suitable for each matrix and atomizer.
Such tables would be of great help to
875 A
practicing analytical chemists. Of
course, the production of these tables
is not a trivial matter, and would require rather involved experimental
procedures at the state of the art in
particle generation and characterization. Nevertheless, when these data do
become available, which undoubtedly
they will in due time, this should allow
researchers to steer around the majority of interference problems other
than those of spectral overlaps, for
which tabular data are already available (I,2 ) .
ICP Systems. The data available
on solvent-loading limitations for organic solvents with the ICP have been
characterized recently ( 3 ) .Some typical limiting aspiration rates are shown
in Table I, together with evaporation
factors, E, of the solvents. These data
are useful as a guide for organic solvent introduction to the ICP. The
evaporation factor is a measure of the
rate of mass loss from an evaporating
drop, and is given by:
E = 48 D , u P , M ~ ( ~ R T ) - ~ (1)
where D , is the diffusion coefficient of
the solvent vapor, u is the surface tension, P, is the saturated vapor pressure, M is the molecular weight of the
solvent, 6 is the density, R is the gas
constant, and T i s the absolute temperature. In general, the ICP has decreasing tolerance to solvents as their
evaporation factors increase, and
there is an inverse correlation between
evaporation factor and limiting aspiration rate. However, the alcohols
have a much greater quenching effect
on the plasma than their evaporation
factors would indicate, and they may
readily extinguish the plasma under
normal operating conditions.
It is always possible to remove at
least part of the solvent vapor by condensation. Two groups of workers
have attempted this and shown that
the tolerance of the ICP to organic solvents is greatly improved when a large
fraction of the solvent vapor is removed from the gas stream passing to
the plasma ( 3 , 4 ) .This is an indication
of how the sample introduction process can be modified to produce analyte closer to the optimum for the atomizer. No published data are available on limitations of aqueous sample
introduction to the ICP, although
clearly water loading in the plasma
has a direct influence on plasma properties. In fact, it has been shown for
certain ionic lines that doubling the
water loading entering the plasma can
cause a 100-fold reduction in analytical signal (5).
From a practical standpoint, three
important conclusions can be reached.
First, it is necessary to introduce sample to the atomizer with drops no larger than a certain maximum size
Advertising removed
from this page
(dmx).Second, the solvent introduction rate must fall within a certain
permissible hand of values. Third, to
maintain good system reproducibility,
it is essential that all these parameters
be controlled carefully over the long
term. Any significant change in drop
size or solvent loading reaching the atomizer could have an adverse effect,
both on system accuracy and system
reproducibility. From this standpoint,
the need to maintain a constant temperature in the plasma box in an ICP
system becomes clear, as a means to
reduce the baseline drift caused by
variable solvent vapor loading (5).
Atomic Absorption Systems.
Flames are generally far less susceptible to variations in solvent loading
than ICPs are, although the introduction of organic solvents to an air-acetylene flame can lead to a significant
temperature drop. This in turn could
cause the onset of interferences due to
sample matrix problems. For flame
AAS systems, the design of nebulizers
and spray chambers appears to have
been empirically optimized to provide
the best aerosol drop size in the flame
for interference-free analyte vaporization. Solvent loading appears to be a
secondary factor.
In the past 20 years AA nebulizers
and spray chambers have undergone a
steady progression. They have
changed from devices producing very
coarse aerosols, with a corresponding
high incidence of vaporization interferences, to devices producing much
finer aerosols, which are largely free
from this type of interference. In fact,
recent data indicate that it is possible
to virtually eliminate all matrix-induced vaporization interferences in
AAS (e.g., calcium-phosphate, siliconaluminum, silicon-manganese, etc.)
(67).This is accomplished by shift-
Table 1. Limiting Organlc
Aspiration Rates for ICP a
Diethyl ether
* Ar ICP operated at 1.75XW rl power.
Defined as maxlmum whnt uptake POBslble fa stable opemtlon lw 1 h.
ing the aerosol distribution reaching
the flame to smaller values, through
modification of the spray chamber design.
In many respects it is surprising
that this process has taken so long, as
early work, particularly that of Stupar
and Dawson, gave a clear indication of
the importance of aerosol drop size in
minimizing interference (8).Since
publication of this paper, there appears to have been very little work
carried out on the systematic study of
aerosol properties and interference effects. Fortunately, it appears now that
such improvements can he accom.
plished relatively simply.
As a counterbalance to any complacence that this statement might imply,
it should also be noted that nebulizers
that produce improved detection limits for many volatile elements have
been marketed recently. These devices
operate by letting a higher proportion
of large-diameter droplets reach the
flame. While this can, in certain in.
Figure 1. Drop size distribution for AA nebulize1
Advertising removed
from this page
Capillary Adjustment
. . ,' . .... ..._.
. . ......
Solution Uptake
Figure 3. Crossflow nebulizer
Figure 2. AA nebulizer
stances, result in improved detection
limits, it can also cause a devastating
worsening of matrix-induced interferences (7). This is one instance where,
unless due care is exercised, we may
recycle to one of the most troublesome
aspects of early AAS.
Nebulizers: Pneumatic, Ullraronlc,
and Other
Liquid sample introduction, with
pneumatic nebulization, is the approach used in the vast majority of
atomic spectroscopydeterminations.
This situation appears unlikely to
change in the foreseeable future. The
precise microscopic processes by
which pneumatic nebulizers operate
are not well understood, though some
general principles are known. In simple terms, a liquid jet is shattered by
interaction with a high-velocity gas
jet. The best description of this process is probably some type of surfacestripping mechanism, such that successive thin surface films of liquid are
removed by the gas flow, and then
spontaneously collapse under surface
tension forces to produce the aerosol
droplets. Whatever the precise mecha-
Figure 4. MAK nebulizer
nisms, the result is an aerosol that
generally has a very wide drop size
distribution (see Figure 1 for a typical
AA pneumatic nebulizer distribution).
The construction of pneumatic nebulizers for atomic spectroscopy is a demanding engineering challenge, as tolerances must be very precisely held on
annular spaces. These may he as small
as 10-20 pm for ICP nebulizers, compared to 150-250 pm for AA nebulizers. The adjustable concentric nebulizer allows Substantial control over
the gas-liquid interaction by varying
the position of the liquid uptake tube
in a conically or parabolically converging gas tube (Figure 2). Crossflow nehulizers avoid this need, but require
very precise and rigid positioning of
the gas and liquid tubes (Figure 3) (9).
The main practical requirements
for pneumatic nebulizers are the following: A high-velocity (sonic to supersonic) gas stream, a reasonable
pressure drop at the liquid injection
capillary for venturi-effect natural aspiration (optional), maximum interaction between the gas and liquid
streams for fine aerosol production,
and freedom from blockage resulting
from either particles suspended in the
solution or from salt buildup at the
nebulizer tip. Matrix salt tolerance
will be determined by both the concentration of the salt and its solubility
characteristics (e.g., 1046 NaCl may
cause no blockage problem, whereas
1046 Na2SO. may rapidly block the
nebulizer). A more serious problem
with the ICP may he whether the plasma torch quartz injection tube can
withstand the high salt level without
clogging, and ultimately devitrifying.
Systems for dc plasma sample introduction generally have a substantial
advantage over ICP systems in their
tolerance to both suspended particles
and dissolved solids. Unfortunately,
the excitation characteristics of
dc plasmas can also he significantly
changed by high concentrations of
easily ionized elements (e.g., Na, Ca,
Atomic absorption nehulizers and
spray chambers are without doubt the
most robust sample introduction devices available for atomic spectroscopy. They are mechanically stable, corrosion resistant (with PtiIr nebulizer
capillaries and inert polymer chambers), difficult to block, and easy to
clean. At the other end of the spectrum are the rather delicate, all-glass
concentric nebulizers often used with
the ICP. Particle blockage with these
devices can be an irreversible process.
Crossflow nebulizers, which are much
less fragile than concentricnebulizers,
can be constructed for the ICP. They
can also be fabricated from acid-resistant materials such as PTFE or
Ryton. In recent years several proprietary pneumatic nebulizer designs
have been introduced that are claimed
to overcome some of the noise and stability problems common to ICP pneumatic nebulizers. For instance, the
MAK nebulizer (Figure 4), named
after Meddings, Anderson, and Kaiser, is a crossflow device that operates
at a very high back pressure (200 psi)
and is made of glass (10).
The fundamental limitation of all
pneumatic nebulizers of conventional
design is that they produce aerosols
with a wide drop size range. This
means that high transport efficiency
can be achieved only at the expense of
allowing large drops to reach the atomizer. One device that produces a
much finer aerosol is the fritted disk
nebulizer (11). This device has been
claimed to generate aerosols with a
mean primary drop size of <1 pm,
and to result in a transport efficiency
(c)of 9046 (12). In our laboratories
we have been able to obtain en values
of 3046. To obtain this performance,
very low sample flow rates of 0.1 mL1
min are necessary, which means that
the net analyte mass transport rate
( WtOt)is not as high as the value
Advertising removed
from this page
Advertising removed
from this page
Flgure 5. Ultrasonic nebulizer
might initially indicate. Nevertheless,
the increase in WtOtcompared to conventional pneumatic nebulizers could
he important in chromatographic and
flow injection interfacing. Some negative aspects of the fritted disk nebulizer are its tendency to froth and block
when concentrated solutions (e.&
>loo0 pglmL) are nebulized for more
than a minute or so, its similar frothing problem with some organic solvents, and its extremely long wash-out
time of several minutes. An external
wash cycle can reduce the clean-out
time to 4 5 4 0 s (12).
The ultrasonic nebulizer has been
suggested as a replacement for the
pneumatic nebulizer since 1964. With
this device, the principle of aerosol
production is significantly different
from pneumatic nebulization. In the
ultrasonic nebulizer, instead of drops
being stripped from a liquid cylinder
hy a high velocity gas jet, surface instability is generated in a pool of liquid by a focused or unfocused ultrasonic beam. The beam is generated by
a piezoelectric transducer. These devices produce aerosols with mean drop
diameters that appear to be a function
of exciting frequency.
At low frequencies (i.e., 50 kHz or
leas),cavitation is the main mode. At
the high frequencies commonly used
in modem ultrasonic nebulizers, typically 1MHz or greater, the mechanism
of aerosol production changes from
cavitation to geyser formation. With
the geyser formation mechanism,
power density in the liquid surface,
rather than the operating frequency,
becomes of major importance.
Experiments in our laboratory with
an ultrasonic transducer of the type
described in Olson et al. (Figure 5)
(13)have shown power-dependent
particle size distributions. However,
ultrasonic nebulizers generally give
much more efficient fine aerosol pro-
duction than pneumatic nebulizers.
Up to 30%efficient production of
droplets in the size range from 1.5 to
2.5 pm has been found a t a 0.3-mL/
min solution flow rate.
Desolvation is essential for ultrasonic nebulization if worthwhile signal
gains are to he made (14). With surface water (15) and seawater (26) samples, improvements in analytical
working range hy factors of 1.1-12
have been claimed, the improvement
varying with element. Cross-contamination problems often encountered
with desolvation systems have been
much reduced by using a concentric
gas sheath to prevent deposits on tube
walls (I 7). Nevertheless, many unan-
swered questions remain about the
general reliability and freedom from
interference of ultrasonic nebulizers
(18);these must be addressed before
they will achieve widespread use.
Nebulizers and spray chambers operate interactively, and must be optimized as a unit rather than individually. There are, however, certain requirements relating specifically to the
spray chamber:
the effective removal of aerosol
droplets larger than the cutoff diameter (d,) found to be necessary
for interference-free measurement,
rapid wash-out characteristics, both
to increase the possible rate of analysis and to avoid cross-contamination problems, and
smooth drainage of waste aerosol
from the chamber, to avoid pressure
pulses in the atomizer.
Slow wash-out is a particular prohlem of ICP spray chambers, caused
because both gas and liquid flows are
substantially lower (typically 1 L/min
and 1 mL/min, respectively) than for
AA systems (typically 18 L/min and
6 8 mL/min). The wash-out times
necessary for a drop to 1%and 0.1%of
peak for a typical ICP spray chamber
are 25 s and 40 s, compared to AA values of 1 sand 3 s. Paradoxically, washout problems with the ICP are aggravated by its exceptionally good linear
working range of up to five orders of
magnitude. The possiblity always exists, with unknown samples of widely
varying composition, that a loo0
pg/mL solution will be followed by one
of 0.1 pg/mL. The consequencesof
Figure 6. Wash-oui and carry-over with ICP spray chamber
Advertising removed
from this page
Figure 7. (a) Babington nebulizer (b) V-groove nebulizer
such a sequence are shown in Figure 6.
One way to reduce wash-out time significantly is to use flow injection techniques, which lie discussed in more
detail later in the article.
High Solids Nebulizers: Babington and V-Groove. The problems of
nebulizer blockage inherent in conventional pneumatic nebulizers are ef-
fectively overcome with Bahington.
type nebulizers. The original Babington concept, developed for paint
spraying, involved a spherical surface
with an array of holes around a circumference. The gas supply came
from within the sphere, and as the liquid flowed over the outside of the
sphere and passed over the gas stream,
Figure 8. Electrothermal vaporizer for ICP
it was nebulized. Devices based on this
concept (Figure 7a), and suitable for
atomic spectroscopy, have been deScribed (19). However, a simpler design
(Figure 7b), in which a liquid stream is
passed down a V-groove, with a small
hole drilled in its center for the gas
stream, has also proved popular, and
is available commercially (20). Babinpton-type nebulizers can be made either entirely of glass, metal, or Teflon,
or by embedding a sapphire orifice in
a Teflon block.
All Babington-type nebulizers are
inherently blockage-free because of
their method of operation. As such,
they are ideal when solutions with suspended particles must be analyzed,
and when prior acid or other dissolution to dissolve the particles is not
convenient. The trap to be avoided
when using these nebulizers is to assume that simply because the sample
is capable of nebulization the analytical results are therefore necessarily
meaningful. The limitations regarding
particle vaporization, discussed earlier, are especially critical in this situation. Simply transporting the analyte
to the flame or plasma does not guarantee that a proportional supply of
atoms or ions will result. In certain situations, such as the direct analysis of
animal tissue after preparation of a
slurry of the finely divided sample, the
accuracy of the procedure has been
well documented (21).In other circumstances, caution should he observed until good agreement with
standard reference materials can be
Electrothermal Vaporization
In many respects electrothermal vaporization (ETV) is better for sample
introduction in ICP emission spectroscopy than it is for atomization in AAS.
Specifically, the limitations in normal
furnace AAS, including condensation
problems within the furnace, light
scattering due to particulates and molecular species, and lack of adequate
linear working range, disappear when
these devices are used for sample introduction rather than for atomization. It is possible to maintain the microsampling capabilities of electrothermal furnaces while producing de&tion limits generally comparable to, or
only a little poorer than those obtainable with furnace AAS. In addition,
the wide linear working range (up to
five orders of magnitude), freedom
from interference, and multielement
capabilities of the ICP are maintained.
It is possible to interface many commercially available electrothermal vaporizers to an ICP with relatively little
system modification. However, the
most suitable type of vaporizer is
probably the open graphite rod device
(22,23).A typical system is shown in
Advertising removed
from this page
Figure 8. As the signal observed will
he transient, the same electronic requirements will apply as for furnace
AAS. This means that direct-reading
spectrometer circuitry must be redesigned. Otherwise, plasma background
rather than analytical signal will be
observed during the extended integration period, and this will lead to degraded detection limits. Typical ETV/
ICP emission detection limits, compared to ETV/AAS performance, are
shown in Table 11.
Vapor Introduction
The fundamental advantages of
vapor introduction, compared to liquid sample introduction, are the following: it allows preconcentration of
the sample from a relatively large volume of solution into a relatively small
volume of vapor, sample transport can
he accomplished with an efficiency approaching 1W%,compared to the
1-1W typical of liquid sample introduction, and the procedures can readily be automated (24).The greater
transport efficiency of vapor generation can be critically important for
several elements. This is especially so
for arsenic, selenium, and tellurium in
ICP emission spectroscopy, where
conventional liquid sample introduction gives inadequate detection limits.
Hydride introduction is necessary for
the determination of these elements at
levels acceptable for environmental
monitoring applications. Typical hydride detection limits, both by ICP
emission and by AAS, are compared
with normal liquid sample introduction values in Table 111.
Table II. Selected Electrothermal Vaporizer Detection Limits
Reference 34
Relwmce 35
Reference 36
Table 111. Detection Limits for Hydride Generation (ng/mL)
0 02
0 03
0 05
All data taken born Reference 37.
Flow Injection and Liquid
Chromatography Introduction
Flow injection- (FI-) and liquid
chromatography- (LC-) coupled atomFigure 9. FllLC system
ic spectroscopy share many common
features, both theoretical and practical. A typical F I L C system is shown
in Figure 9. For LC, a low-pulse, highpressure pump is necessary, whereas
for FI a peristaltic pump may be used.
Flow injection has recently received
some careful study for both AA and
ICP applications (25,26).The primary advantages over conventional
sample introduction are the following.
First, only a relatively small volume of
sample is necessary to achieve a signal
comparable to continuous nebulization. For example, with the ICP and a
suitable LC-type injection system,
50% of the equilibrium signal can be
achieved with a 1W-pL injection. Second, because of the transient nature of
the signal, exponential decay in the
wash-out process starts much sooner
than with continuous sample introduction (Figure 10).Consequently, the
signal decays to baseline more rapidly
Figure 10. Flow injection peaks
than with normal continuous sample
introduction. It is therefore possible,
with the ICP, to inject samples at the
rate of approximately 4/min, as opposed to 1.5/min with conventional
sample introduction. Additionally, in
AAS, where the addition of ionization
buffers, lanthanum releasing agents,
etc., may be desirable, it is possible to
add the analyte as a spike into a flowing stream of the desired buffer, making for a relatively simple experimental system (26).
Other Techniques for Sample
The techniques considered so far
have achieved substantial practical
use; there are others which have more
specialized applications. Laser ablation, in which the power from a focused ruby laser is used to vaporize a
spot of material directly from a solid
surface, has considerable promise (27,
28). Another approach of great interest in metallurgy is the use of spark or
arc vaporization (29,30).Some interesting studies have been made in
which sample is introduced into the
ICP with a carbon rod and placed into
the torch in the region of the plasma
coils but below the plasma itself (31,
32). Direct inductive heating of the
carbon occurs, and the sample vaporizes directly into the plasma. With
this system, very efficient transport of
sample to the plasma is readily accomplished. However, vaporization is not
always very rapid, and the broad emission peaks that result can sometimes
lead to poor detection limits.
Other devices aimed at obtaining efficient sample transfer of solid and
liquid samples to the ICP have been
described recently, including a system
where the rf plasma is led into a chamber below the torch for sample vaporization (33).
Many of the devices presently proposed as alternatives to liquid sample
introduction offer great promise for
specific applications; however, to
achieve widespread use, they will have
to demonstrate the reliability, freedom from interference, and the ease of
use that liquid sample introduction
currently enjoys. Finally, there is always the possibility that some truly
new sample introduction technique,
with general applicability, will be developed. The need is certainly there.
This material is based on work supported by the National Science Foundation under Grant No. CHESO19947.
(1) Parsons, M. L.; Forster, A.; Anderson,
D. “An Atlas of Spectral Interferences in
ICP Spectroscopy”; Plenum: New York,
(2) Boumans, P.W.J.M. “Line Coincidence
Tables for Inductively Coupled Plasma
Emission Spectrometry”; Pergamon:
New York, 1981; Vols. I and 11.
(3) Boorn, A. W.; Browner, R. F. Anal.
Chem. 1982,54,1402.
(4) Hausler, D. W.; Taylor, L. T. Anal.
Chem. 1981,53,1223.
(5) Kull, R., Jr.; Browner, R. F. Spectrochim. Acta B 1983,38,51.
(6) Smith, D. D. PhD Thesis, Georgia Institute of Technology, Atlanta, Ga., 1983.
(7) Smith, D. D.; Browner, R. F., submitted for publication in Anal. Chem.
(8) Stupar, J.; Dawson, J. B. Appl. Opt.
1968, 7,1351.
(9) Novak, J. W.; Lillie, D. E.; Boorn,
A. W.; Browner, R. F. Anal. Chem. 1980,
(10) Anderson, H.; Kaiser, H.; Meddings,
B. In “Developments in Atomic Plasma
Spectrochemical Analysis”; Barnes,
R. M., Ed.; Heyden and Son: London,
U.K., 1981, p. 251.
(11) Apel, C. T.; Duchane, D. V. Abstracts
of Papers, Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Cleveland, Ohio, 1979.
(12) Layman, L. R.; Lichte, F. E. Anal.
Chem. 1982,54,638.
(13) Olson, K. W.; Haas, W. J., Jr.; Fassel,
V. A. Anal. Chem. 1977,49,632.
(14) Boumans, P.W.J.M.; de Boer, F. J.
Spectrochim. Acta B 1975,30,309.
(15) Taylor, C. E.; Floyd, T. L. Appl.
Spectrosc. 1981,35,408.
(16) Berman, S.S.; McLaren, J. W.; Willie,
S. N. Anal. Chem. 1980.52.488.
(17) Mermet, J. M.; Trassy, C. In “Develo ments in Atomic Plasma SpectrocEemica1 Analysis”; Barnes, R. M., Ed.:
Heyden and Son: London, U.K., 1981,
p. 245.
(18) Boumans, P.W.J.M.; de Boer, F. J.
Spectrochim. Acta B 1976,31,355.
(19) Garbarino, J. R.; Taylor, H. E. Appl.
Spectrosc. 1980,34,584.
(20) Suddendorf, R. F.; Boyer, K. W. Anal.
Chem. 1978.50.1769.
(21) Mohamed, N.; Brown, R. M., Jr.; Fry,
R. C. Appl. Spectrosc. 1981,35,153.
(22) Kirkbright, G. F.; Gunn, A. M.; Millard, D. L. Analyst (London) 1978,203,
(23) Barnes, R. M.; Fodor, P. Spectrochim.
Acta B 1983,38,1191.
(24) Godden, R. G.; Thomerson, D. R. Analyst (London) 1980,105,1137.
(25) Greenfield, S. Spectrochim. Acta B
(26) Tyson, J.; Idris, A. B. Analyst (London) 1981,106,1125.
(27) Carr, J. W.; Horlick, G. Spectrochim.
Acta B 1982,37,1.
(28) Thompson, M.; Goulter, J. E.; Sieper,
F. Analyst (London) 1981,106,32.
(29) Human. H.G.C.: Scott. R. H.: Oakes.’
A. R.; West, C. D. Analyst (London)
(30) Marks, J. Y.; Fornwalt, D. E.; Yungk,
R. E. Spectrochim. Acta B 1983,38,107.
(31) Salin, E. D.: Horlick, G. Anal. Chem.
(32) Kirkbright, G. F.; Walton, S. J. Analyst (London) 1982,207,241.
(33) Farnsworth, P. B.; Hieftje, G. M.
Anal. Chem. 1984,55,1414.
(34) Long, S. E.; Snook, R. D.; Browner,
R. F., submitted for publication in Spectrochim. Acta B.
(35) Fassel, V. A,; Kniseley, R. N. Anal.
Chem. 1974,45,1110 A..
(36) Knoller, B. N.; Bloom, H.; Arnold,
A. P. Prog. Anal. At. Spectrosc. 1981,4,
(37) Nakahara, T. Prog. Anal. At. Spectrosc. 1983,6,163.