ANALYTICAL AND DIAGNOSTIC STUDIES OF AN
ICP-EXCITED ICP FLUORESCENCE SPECTROMETER
ROBERT JOSEPH KRUPA
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
To my parents, whose loving support made all
I am grateful to Chester Eastman, Dailey Burch, and Vernon Cook of
the Chemistry Department machine shop for their construction of much of
the instrumentation used in this work.
I also wish to thank Dr. Benjamin Smith, Dr. Edward Voigtman, and
Lori Davis whose stimulating discussions and comments have helped me
enormously during my graduate career. Special thanks are extended to
Dr. Nicolo Omenetto for his inspiring enthusiasm and many helpful
suggestions concerning my research.
Steven Barber of R.F. Plasma Products deserves much of the credit
for demounting the ICP load coils. His assistance in maintaining the
ICP systems in our lab and in the initial set up of the present system
is gratefully appreciated.
I owe a special debt of gratitude to Dr. James D. Winefordner whom
I have had the privilege of working with during the past four years.
His concern for his research group, his many innovative suggestions, and
his encouragement during the course of this work have made my stay at
the University of Florida a pleasurable and invaluable learning experi-
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . . .
LIST OF TABLES . . . . .
LIST OF FIGURES . . . .
ABSTRACT . . . . . .
1. INTRODUCTION . . . . .
2. MATERIALS AND METHODS . . .
ICP-ICP-AFS System . . . .
Laser-Excited AFS System . .
Standard Operating Conditions .
Sample and Standard Solutions .
3. DIAGNOSTICS WITH THE EXTENDED SLE
Influence of R.F. Power . . .
Quantum Efficiency of the ICP .
Conclusions . . . . . .
4. FIGURES OF MERIT FOR ICP-ICP-AFS
Detection Limits for ICP-ICP-AFS
ICP-ICP Resonance Monochromator
Spectral Interferences . . .
Interelement Effects . . .
SRM Analysis . . . . .
Conclusions . . . . . .
5. FINAL COMMENTS AND SUMMARY . .
A. GLOSSARY OF ACRONYMS AND ABBREVIATIONS
B. QUANTUM EFFICIENCY MEASUREMENTS . .
C. COMPARATIVE DETECTION LIMITS . . .
REFERENCES . . .. . . . . . .. . . .. . 73
BIOGRAPHICAL SKETCH . .. .. .. ... .. .. ...... . ... 77
LIST OF TABLES
ICP Operating Conditions . . . . . . .
Ionization Potentials and Molecular Dissociation
Energies . . . . . . . . . . .
Wavelength of Transitions Used in This Study . ..
Comparative Detection Limits in ng/L (S/N=2) . ..
Resonance Monochromator Detection Limits (mg/L) and
Comparative Upper Linear Concentrations (mg/L) . .
Spectral Interferences in ICP-ICP-AFS . . . .
Comparison of Atomic Spectrometric Methods . . .
LIST OF FIGURES
1. Block Diagram of ICP-ICP-AFS . . . . . . . 5
2. Non-fractory AFS vs Power . . . . . . . .. 17
3. Refractory AFS vs Power . . . . . . . .. 18
4. Alkaline Earth AFS vs Power . . . . . . .. 19
5. Ar(m) AAS vs Power . . . . . . . . . 21
6. Ca AAS vs Power . . . . . . . . . .. 22
7. Ca AFS vs Power . . . . . . . . . . 23
8. CaOH/CaO AES vs Power . . . . . . . ... .25
9. Ca AES vs Power . . . . . . . . . .. 26
10. Li Lifetime vs Propane Flow . . . . . . ... .30
11. Electron Number Density vs Propane Flow . . . .. 31
12. Ca AFS vs Propane Flow . . . . . . . ... 32
13. Ca Atom/Ion AFS vs Propane Flow . . . . . ... 33
14. Hydroxyl AES vs Propane Flow . . . . . . .. 35
15. Ca AFS and RM Curves of Growth . . . . . .. . 41
16. Ca Spectral Interference on Al . . . . . . .. 44
17. Al Spectral Interference on Zn . . . . . . .. 45
18. Effect of Na on Ca AFS . . . . . . . . .. 51
19. Schematic Diagram of a 2-Level Atom . . . . .. 62
20. Fluorescence Decay Curve for Na . . . . . .. 65
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
ANALYTICAL AND DIAGNOSTIC STUDIES OF AN
ICP-EXCITED ICP FLUORESCENCE SPECTROMETER
Robert Joseph Krupa
Chairman: James D. Winefordner
Major Department: Chemistry
A 20 g/L solution of the element of interest is aspirated into a
1500 W inductively coupled plasma (ICP) and the resulting emission is
used to excite atomic and ionic fluorescence in a second ICP which is
used as the atomization cell. The fluorescence vs r.f. power curves
exhibit three trends which describe the atomization efficiency of the
elements investigated at different power levels. The non-refractory and
refractory elements exhibit trends with a decrease and increase in
fluorescence sensitivity vs power, respectively, while the alkaline
earth elements' behavior is quite different, possibly because of a
complex interaction of these elements with the decomposition products of
water. Laser-excited atomic fluorescence (LEAFS) measurements of
quantum efficiencies demonstrate that the ICP is an excellent atom and
ion reservoir for fluorescence measurements with quantum efficiencies
approaching unity. However, when propane is added to the plasma to
increase the atomization efficiency of the refractory elements, the
quantum efficiency is reduced and an increase in the atom/ion ratio is
observed. ICP-excited ICP atomic fluorescence spetrometry (ICP-ICP-AFS)
detection limits are comparable to ICP-AES LODs; however, the combina-
tion of the fluorescence and resonance monochromator techniques yields
linear dynamic ranges (LDRs) superior to atomic emission spectrometry
(AES), approaching eight orders of magnitude in some cases. Spectral
interference are minimal with this system because of the selectivity of
the fluorescence technique, making background correction unnecessary in
most cases. Interelement effects are negligible in many cases because
of the long sample residence time in the plasma; however, ionization
interference are slightly pronounced because of the relatively low
temperature of the plasma at the observation height employed. The
presentsystem should be considered a viable alternative to conventional
emission spectrometry when it is necessary to alleviate spectral
interference which may occur in complex sample matrices.
In the early 1960s, T.B. Reed developed the first inductively
coupled plasma which employed argon tangentially flowing in a cylindri-
cal torch . The torch design was later modified and investigations
into the analytical potential of the ICP were performed [2,3]. It was
not until the 1970s that the ICP became a commercially available analyt-
ical instrument; since its introduction, ICP-AES has rivalled the
detection limits of furnace AAS without the many matrix interference
encounted with electrothermal atomizers. Recently, the ICP has been
established as an excellent line source for exciting fluorescence in
flames [4-6] and a second ICP [7-9].
The efficient atomization and excitation properties of the ICP have
yielded emission detection limits in the ng/mL range for most elements.
The high intensity and stable emission from an analyte introduced into
an emission ICP makes it more versatile than any other conventional
source such as HCLs, EDLs, or Xe arc lamps. Several atomic line sources
are lacking in intensity and stability, such as HCLs and EDLs for the
non-metals, the lanthanides and actinides, and many refractory elements.
In addition to the higher spectral irradiance of the ICP, the plasma is
remarkably free of self-absorption at the observation heights commonly
employed [4, 10]. This allows the aspiration of high concentrations, up
to several percent, of the element of interest into the source ICP.
The ICP as an emission source for exciting fluorescence in a second
plasma also has the advantage of emitting many ionic lines. This is an
important consideration when the analyte element is highly ionized in
the atomization cell. In fact, two-thirds of the strong transitions
observed in the ICP are ionic transitions . Low temperature sources
of excitation, such as HCLs and EDLs, emit few ion lines and are pre-
dominantly atomic line sources; therefore, in many cases, these lamps do
not permit the optimum fluorescence transitions to be probed. It is
because of this limitation that the only commercially available atomic
fluorescence instrument, manufactured by the Baird Corp., utilizes low
r.f. powers and the addition of propane to the plasma . At higher
r.f. powers, the ionization of many elements is substantial, but the
HCLs cannot efficiently excite these ionic species, and so low powers
and low temperatures must be employed.
Because the emission of an element introduced into the ICP can be
considered a narrow line source, many of the spectral interference
which commonly plague ICP-AES can be avoided. This spectral selectivity
allows the analysis of even complex samples with the use of only a wide
bandpass monochromator or interference filter, without the need for
By employing a second ICP as the atomization/ionization cell, a
relatively high degree of freedom from matrix or interelement effects
can be achieved. This is especially apparent when the ICP is compared
to lower temperature cells such as flames, graphite furnaces, and
microwave plasmas [13,14]. The higher temperature plasma is also a much
more efficient atomizer, especially for the refractory metals which form
very stable metal-oxide molecules.
Another unique advantage of employing two ICPs is the ability to
determine trace species by the more sensitive fluorescence technique
while analyzing the major sample constituents by the resonance mono-
chromator technique [4-6]. When the fluorescence curve of growth
becomes nonlinear because of self-absorption at high concentrations, the
sample is used as the excitation source for exciting the fluorescence of
a standard aspirated into the sample cell ICP. This combination of AFS
and RM methods leads to linear dynamic ranges of up to eight orders of
A description of the ICP-ICP-AFS instrument and an evaluation of
the system's analytical figures of merit are now presented. Also, some
diagnostic studies are presented which attempt to add to the existing
knowledge we have of the ill-understood physical nature and excitation
processes in the inductively coupled plasma.
MATERIAL AND METHODS
A block diagram of the ICP-ICP-AFS instrument is presented in
Figure 1. Both ICPs were operated at 27.12 MHz with automatic matching
networks which tuned the capacitance of the generator to the plasma,
reducing the reflected power. These units were both Plasma-Therm
(Cherry Hill, N.J.) generators and matching networks, the source being a
2.5 kW unit (Model 2500K), with the atomization cell ICP being 1.5 kW
In the past, translation of the ICP torch and load coil assembly
was accomplished by moving the entire matching box/load coil assembly in
the desired direction. This usually involved the use of cumbersome x-y
milling tables and capstans which had poor spatial resolution and
reproducibility. To facilitate the translation of the torches, the load
coil assembly of each ICP was removed from the matching box. The load
coil was mounted on a 6 x 6 x 0.5 in Teflon sheet. The plasma gases and
cooling water were brought to the torch from the back side of the Teflon
sheet through nylon bulkhead feedthroughs. The r.f. power was transmit-
ted to the load coil from the matching box via two 12 x 0.75 x 0.01 in
silverplated brass straps. The procedure described by Carr et al. 
which employed tinned copper braid as the r.f. conductor proved unsuc-
cessful. We were unable to ignite the plasma because of high reflected
power which could not be tuned out with the capacitors in the matching
network. This tuning mismatch may have been due to the added inductance
introduced into the matching network by the copper braid. On a sugges-
tion from Koirtyohann , who experienced similar tuning problems, we
found brass straps to work without any increase in the reflected power.
This system has been previously employed in several ICP spectrometers
manufactured by the Baird Corporation (Bedford, MA) but did not reach
the literature until recently . With this system, the load coil
assembly could be separated from the matching box up to a distance of
18 in without any matching problems. A conductor length of 12 in was
chosen for this application which allowed vertical translation of the
load coil up to 6 in and horizontal travel up to 4 in. To position the
plasma, the Teflon mounting sheet was mounted on two micrometer-
controlled translational stages (Newport Research Corporation, Fountain
Valley, CA, model 440-2), connected at right angles to each other by an
aluminum L-bracket. The two load coil assemblies, chopper, and optical
components were mounted on a laboratory-constructed optical rail which
was enclosed in a 28 x 13 x 23 in (length x width x height) aluminum
box. The box served as a Faraday cage to contain the r.f. radiation
which may add to the electronic noise in the detection electronics, and
also reduced the stray light due to reflections of the source radiation
which may increase the spectral background detected by the monochromator
and photomultiplier tube. The interior of the box, which was painted
flat black to reduce the reflectance of source radiation, was divided
into two compartments by a grounded aluminum plate (with a window along
the optical axis) placed between the two ICP load coils. This plate
reduced the amount of scattered source radiation which may reach the
monochromator and also served to reduce the coupling of the r.f. from
the two ICPs which produced a high frequency tone at the beat frequency
of the two generators.
The emission from a 20 g/L solution of the element of interest,
which was aspirated into the source plasma operating at 1500 W, was used
as the excitation line source for exciting the fluorescence. The source
light was collected with a 2 in diameter, f 1, front-surface concave
mirror and a 2 in diameter, f 1, fused silica lens. This collimated
light was modulated at 168 Hz by a mechanical chopper (Princeton Applied
Research, Princeton, NJ, model 125A) and focused onto a second ICP,
acting as the atomization cell for the sample solution, by a 2 in
diameter, f 1.5, fused silica lens. An extended sleeve torch (Baird
Corporation, Bedford, MA) was employed for the atomization cell in order
to reduce the amount of air entrainment into the plasma and an observa-
tion height of 55 mm above the load coil was employed in order to
increase the residence time of the sample in the plasma. The fluores-
cence radiation was collected at 900 by a single lens which formed a
one-to-one image of the atomization cell plasma on the entrance slit of
a low resolution (2.2 nm FWHM bandpass) monochromator (GCA-McPherson
Corporation, Acton, MA, model EU-700). The light detected by the
photomultiplier tube (Hamamatsu, Inc., Middlesex, NJ, model 928) was
amplified by a current-to-voltage amplifier (PAR, model 181) and demod-
ulated and filtered by a lock-in amplifier (PAR, model 186A) with an
output time constant of 1 s. The lock-in output was displayed on a
strip chart recorder (Fisher Scientific Co., Pittsburgh, PA, model 5000)
and filtered by a laboratory constructed 10 s integrator which was
interfaced to a digital multimeter (Keithley Instruments, Inc., Cleve-
land, Ohio, model 175).
Absorption measurements employed the same detection system as in
AFS; however, the collimated source radiation was folded by three front
surface, plane mirrors before passing through the atomization cell ICP
and into the monochromator (0.1 nm FWHM bandpass). The emission mea-
surements from the atomization cell ICP were measured with the same
collection optics as used for AFS and AAS; however, the lock-in
amplifier was eliminated from the detection electronics.
Laser-Excited AFS System
A pulsed nitrogen laser (Laser Science, Inc., Cambridge, MA, model
VSL-337) with a pulse width of 3 ns was employed to pump a dye laser
(LSI, model DLM) and the emitted light, with a FWHM of 0.4 nm, was used
to excite fluorescence in the extended sleeve torch atomization cell in
order to determine the quantum efficiency of the ICP as a function of
r.f. power and propane flow rate. In this study, the torch aerosol
injection tube was positioned between the first and second turns of the
load coil in order to produce the "pencil plasma" as employed by the
Baird Plasma/AFS system (Baird Corp., Bedford, MA). The fluorescence
was observed at an observation height of 80 mm above the top of the
aerosol injection tube, or approximately 1 cm above the top of the
torch. The previously described collection optics and monochromator
were employed to obtain the fluorescence light which was detected by a
photomultiplier tube (Hamamatsu, Inc., Middlesex, NJ, model 928)
operated at -1000 V, its base modified for fast response  by adding
capacitors to the resistor network on the dynode chain. The
photomultiplier tube output was connected to a 400 MHz storage oscillo-
scope (Tektronix, Inc., Beaverton, Oregon, model 7834) by a 1 m RG58U
cable. The oscilloscope was triggered by a fast photodiode terminated
into 50 ohms. The convolved laser pulse/photomultiplier tube response
exhibited a 4 ns FWHM. The fluorescence lifetime measurement system was
also used to determine the quantum efficiency of Na in the extended
sleeve torch vs r.f. power with the aerosol injection tube positioned
5 mm below the load coil, as it is normally positioned.
Standard Operating Conditions
The operating conditions for the plasmas are listed in Table 1.
These were the parameters employed for all measurements with the excep-
tion of the studies involving the effect of propane on the "pencil
plasma." The pencil plasma employed the same extended sleeve torch as
the atomization cell ICP as well as the same concentric Meinhard
nebulizer. The cooling Ar was supplied at a rate of 10 L/min and there
was no auxiliary or plasma Ar used. The nebulizer was operated at
The source ICP nebulizer was fed with a peristaltic pump (Rainin
Instrument Co., Boston, MA, model rabbit) which reduced the source
flicker noise by approximately 10%. The major advantage of the pump,
however, was to eliminate the salt encrustation of the aerosol injection
tube. Previously the injection tube would clog after nebulizing the
20 g/L source solutions for 45 to 60 min, depending on the element.
With the peristaltic pump, the nebulizer was operated at a lower sample
uptake rate than pneumatic nebulization. This may have caused a
Table 1. ICP Operating Conditions.
(above the load coil)
Sample Uptake Rate
500 1500 W
aModel TN-1, R.F. Plasma Products, Cherry Hill, NJ.
Model T-230-A1, J.E. Meinhard Associates, Santa Ana, CA.
smaller aerosol droplet size which reduced the salt encrustation in the
Sample and Standard Solutions
Stock solutions of 20 g/L (2%) were prepared by dissolving the pure
metals or reagent grade salts in the minimum amount of acid and then
diluting with deionized water. One exception to this procedure was the
20 g/L B solution which was prepared by dissolving boric acid in dilute
ammonia because of this salt's limited solubility in water and dilute
acids. Volumetric dilutions of these stock solutions were made in order
to obtain the desired standard concentrations.
The National Bureau of Standards Standard Reference Material high
carbon steel sample (SRM 364) was prepared by refluxing 3.0515 g of the
sample for 10 h in 10 mL water and 20 mL concentrated, sub-boiling
distilled nitric acid. After this time, the sample was transferred to a
Teflon beaker and heated to dryness in the presence of 20 mL concen-
trated hydrofluoric acid. The dried sample was redissolved in 10 mL
concentrated, sub-boiling distilled nitric acid and diluted to 250 mL
with deionized water. A 10 mL aliquot of this solution was diluted to
100 mL with water and analyzed. The final sample concentration was
1.2206 g/L. A blank was also prepared using the same reagents and
procedure. In all other cases, only a deionized water blank was
DIAGNOSTICS WITH THE EXTENDED SLEEVE TORCH
Influence of R.F. Power
The most important factor influencing the fluorescence signal in
ICP-ICP-AFS was the r.f. power to the atomization cell ICP. Argon flow
rates and observation height play an important role in the signal and
background intensities, but to a much smaller degree [7-9,19]. It has
become apparent, especially in ICP-AES, that there exists a strong
dependence of the signal-to-background ratio on the operating param-
eters, leading several researchers to employ simplex optimization
schemes [20-22] to determine the optimum combination of operating
parameters for a particular analysis.
While optimizing the operating conditions for ICP-ICP-AFS, it
became obvious that the elements investigated could be divided into
three categories: refractory elements, non-refractory elements and the
alkaline earth elements, depending on the shape of the fluorescence vs
r.f. power plots. In order to investigate the cause of the behavior of
the group II elements which were expected to follow the trends exhibited
by the refractory or non-refractory elements, ICP-excited ICP-AFS, AAS,
and ICP-AES were employed. The emission and fluorescence techniques
have been well established as analytical and diagnostic methods;
however, we believe this is the first reporting of the use of an ICP as
a source for atomic/ionic absorption employing a second ICP as an
atom/ion reservoir, although there have been several attempts to use
conventional sources with an ICP as an atom reservoir for AAS [23-25].
The advantage of using an ICP as an excitation source for AAS and AFS is
that it is not limited to only atomic emission lines as are the low
temperature hollow cathode lamps and electrodeless discharge lamps.
This flexibility has allowed the study of not only atomic transitions
but also ionic absorbance and fluorescence transitions as well.
For AAS, the relative source shot noise decreased with increasing
source concentration aspirated into the source ICP because the signal-
to-noise ratio (S/N) of I increases as the square root of the intensity
(S/N a 4To). However, collisional and self-absorption broadening
increases the source linewidth as the source concentration is increased
. This leads to a decrease in the absorbance signal and a deter-
ioration of the sensitivity. The source concentration which yielded the
highest analyte absorbance S/N was typically 10 mg/L for the elements
investigated. Absorbance curves of growth exhibited a linear relation-
ship between absorbance and concentration from the detection limit up to
approximately 500 mg/L. The detection limits for the alkaline earth
elements and Cu (S/N=3) were between 1-10 mg/L. The relatively poor
LODs are in part due to the short absorption pathlength through the
absorption cell ICP, approximately 1 cm, and the flicker noise on the
excitation source ICP.
The analyte solutions for AFS and AES were 1 mg/L, with the excep-
tion of P which was 10 mg/L. For AAS, 100 mg/L solutions were used
because of the poor sensitivity. Curves of growth demonstrated that
these concentrations were all within the linear dynamic ranges of the
three techniques. During this investigation, only one parameter, the
r.f. power to the atomization cell ICP, was varied while all other
operating parameters were held constant at the values listed in Table 1.
As a general rule, the elements which are easily atomized and which
are determined in the air/acetylene or air/hydrogen flames are termed
non-refractory. In contrast, those elements which form stable oxide
molecules with dissociation energies (D ) in excess of 6 eV and which
require a nitrous oxide/acetylene flame for atomization are labelled
refractory. The D values of several representative elements along with
their ionization potentials are listed in Table 2, while Table 3 lists
the wavelength of the transitions employed.
Most elements which were investigated by AFS fall neatly into these
two categories. The non-refractory elements, exemplified in Figure 2 by
Cu, Na, and Zn, show a decrease in fluorescence intensity with
increasing r.f. power. The decrease is due to a decline in the ground
state population because of an increase in the excited state and ionic
populations at higher powers, thereby decreasing the ground state atom
number density which can be probed by resonance fluorescence. The
refractory elements, exemplified in Figure 3 by B, P, and Zr, show an
increase in fluorescence intensity with increasing r.f. power because
the refractory oxide molecules are more efficiently dissociated at
higher powers, and hence, higher temperatures where there is more energy
available to dissociate the strong metal oxide bond.
Figure 4, however, demonstrates the fluorescence behavior of the
alkaline earth elements. This behavior was not typical of either the
refractory or non-refractory trends, and appeared to be a combination of
the behaviors portrayed in Figures 2 and 3. To determine whether this
Table 2. Ionization Potentials and Molecular Dissociation Energies.a
Element I.P. (eV) MO MO M
MO MOH M(OH)2
B 8.298 8.1 -- --
Ba 5.212 5.8 4.9 9.7
Ca 6.113 4.8 4.3 9.4
Cu 7.726 4.9 -- -
Mg 7.646 3.9 2.4
Na 5.139 2.8 3.3
P 10.486 6.1 -- --
Sr 5.695 4.8 4.2 9.3
Zn 9.394 4 -- -
Zr 6.84 7.8
Table 3. Wavelength of Transitions Used in This Study.
Specie Wavelength (nm)a
aThe summation sign indicates that more than one transition falls within
the monochromator spectral bandpass of 2.2 nm (AFS).
bWavelengths from reference 28, all other from reference 29.
05 0.7 0.9 UI L3 15
R.F. POWER (KW)
Figure 2. Non-refractory AFS vs power.
Figure 3. Refractory AFS vs power.
05 0.7 0.9 LI 1.3
R.F. POWER (KW)
0 \ \ \ M(ll)
0- 'Fi' i I I
05 0.7 0.9 LI 13 L5
R.F. POWER (KW)
Figure 4. Alkaline earth AFS vs power.
behavior was due to a change in the physical nature of the plasma as the
r.f. power was varied, such as abrupt changes in Ar metastable number
density, quantum efficiency, etc., the Ar(m) absorbance was measured
while varying the r.f. power and nebulizer conditions. Figure 5 shows
the absorbance vs power for the conventional short torch and the
extended sleeve torch used in this work. The Ar(m) absorbance (and
therefore its population) increased smoothly with increasing power and
was reduced when the nebulizer gas is turned on and when water is intro-
duced into the plasma. However, the trends in the Ar(m) absorbances did
not exhibit the dips or peaks in intensity which might account for the
behavior of the alkaline earth elements. This was expected since the
curves for the non-refractory and refractory elements follow the
behavior expected according to their various dissociation energies and
do not appear to be influenced by the changing nature of the plasma as
the power is increased (other than the increase in plasma temperature).
Calcium was chosen as the test element for further investigation
because of this element's high transition probabilities, relatively
simple atomic and ionic spectra , and its well-studied behavior in
both flames and plasmas [8,28,30-34].
Figures 6 and 7 show the behavior of Ca AAS and AFS intensities vs
r.f. power. The fluorescence signals follow the absorbance signals
quite closely indicating that these trends were due to the changing atom
and ion populations in the plasma and were not due to selective quench-
ing processes involving the excited state. If the quenching rate was
changing as the r.f. power was increased, the fluorescence and
absorbance curves would not exhibit the same trends because the
Ar(m) 811.531 nm
o SHORT TORCH
* LONG TORCH
R.F. POWER (KW)
Ar(m) AAS vs power. a)no nebulizer Ar, b)nebulizer Ar
flowing at 1 L/min, c)nebulizer Ar flowing at 1 L/min
with water aspirated at 1 mL/min. Excitation ICP power
at 800 W. Spectrometer bandpass is 0.1 nm.
Figure 6. Ca AAS vs power.
0.5 0.7 0.9 1.1 L3 1.5
R.F. POWER (KW)
Figure 7. Ca AFS vs power.
0.5 0.7 0.9 1.1 1.3
R. F. POWER (KW)
quenching would only serve to decrease the fluorescence signal while
leaving the absorbance signal unchanged. This is because absorption
involves only the absorption of a photon, and this process does not
involve the population of the excited state; it only depends upon the
population of atoms or ions in the lower level from which absorption
takes place. Fluorescence, on the other hand, is a two-step process
which involves absorption as a first step; the second step, the emission
or fluorescence of a photon, depends upon the quenching environment of
the excited atom or ion. Further evidence that the quenching envi-
ronment was not changing as the power was increased has been shown by
Uchida et al.  who measured the lifetime of Na(I) vs r.f. power and
found no change in the excited state lifetime. This was repeated using
the present system by measuring the lifetimes of Na(I) and Li(I) using
laser-excited fluorescence. It was demonstrated that the lifetimes of
these two elements remained constant at 16 ns for Na and 24 ns for Li
over a range of r.f. powers from 400 to 1500 W. Because the quantum
efficiency is directly proportional to the observed lifetime, this
indicates that the quantum efficiency of the atomization cell ICP does
not change with changing r.f. power. A more detailed description of the
quantum efficiency measurements is presented in Appendix B.
The emission signals vs r.f. power also showed a strong dependence
on power as shown in Figures 8 and 9. The emission signals, however,
were not necessarily correlated with the absorbance and fluorescence
signals because the emission is related not only to the total population
but also to the thermally excited population, whereas the absorbance and
fluorescence signals are related to the ground state number density.
The molecular species of Ca, as shown in Figure 8, exhibited the
05 0.7 0.9 L3
Figure 8. CaOH/CaO AES vs power.
Figure 9. Ca AES vs power.
0.5 0.7 0.9 UI L3
R. F. POWER (KW)
unexpected behavior of an increase in emission intensity at r.f. powers
above 900 W. Although not strictly related to the ground state popula-
tion, it was expected that the oxide and hydroxide populations in the
plasma would decrease at the higher powers because of their relatively
low D values. These emission plots seem to indicate that the Ca
molecules become less abundant as the power increases from 500 to 900 W
where the largest free atom population should occur. At higher powers,
however, the increase in emission intensity may indicate a recombination
of the free atoms/ions with the dissociation products of water. Similar
trends have been observed for the absorption, emission, and fluorescence
vs power plots of Mg(I), Mg(II), Sr(II), and Ba(II) using an ICP as an
excitation source, and also with laser-excited AFS .
One possible explanation for this behavior is a complex "equi-
librium" which may exist between the akaline earth atoms/molecules and
the dissociation products of water. Although the plasma is not in local
thermodynamic equilibrium, that is all radiative and nonradiative
processes are not in equilibrium, certain equilibria still exist in the
ICP. Examples of these equilibra, referred to as partial thermodynamic
equilibria, which have been observed are Boltzman equilibrium of high-
lying energy levels of Fe and between the ground state and metastable
states of Ar .
At low r.f. powers, the sample aerosol is desolvated and the
resulting molecular species were efficiently dissociated yielding a high
atom or ion ground state population to be probed by fluorescence. As
the power and temperature were increased, the decomposition of water
into hydroxyl, oxygen, and hydrogen radicals may have increased which
could have shifted the metal-metal oxide/hydroxide/dihydroxide
equilibrium toward the molecular species side, thereby decreasing the
atom/ion population and the AFS signal. However, the higher tempera-
tures also caused an increase in the excited state population relative
to the ground state atom/ion and molecular populations. These processes
at the higher powers may have resulted in the sudden increase in the
fluorescence signals near 1200 W where the efficient molecular dissocia-
tion processes and the low hydroxyl and oxygen radical populations
produced a high atom/ion population.
Quantum Efficiency of the ICP
As was discussed previously, lifetime measurements by LEAFS have
demonstrated the high quantum efficiency of the Ar ICP when used as an
atomization cell for fluorescence measurements (see Appendix B). This
is because the inert Ar environment has a very low quenching cross-
section as compared to combustion flames which contain large concentra-
tions of very efficient quenchers such as N2, CH, and C2 . However,
it has been demonstrated by several researchers [8,12] that the addition
of organic compounds enhanced the fluorescence signals of many elements.
Typically, propane was added to the Ar carrier gas at a rate of
10-30 mL/min. For the refractory elements, the addition of propane
appeared to scavenge the oxygen in the plasma forming either CO or CO2.
This reduced the free oxygen in the plasma which could react with the
metal atoms which form stable refractory oxide molecules. However, even
for several non-refractory elements, such as Li and Cu, which do not
form stable oxide molecules in the ICP, there existed an enhancement in
the fluorescence signals upon the addition of a small amount of propane
to the plasma. The addition of organic molecules to the ICP was
expected to reduce the fluorescence signals of the non-refractory
elements because of the decrease in quantum efficiency. In order to
investigate the mechanism of this fluorescence enhancement, the quantum
efficiency, electron number density, atom/ion populations, and OH
signals were studied.
As was expected, the quantum efficiency (and the lifetime) of atoms
the ICP was reduced when propane was added to the plasma as is
demonstrated in Figure 10. A decrease in the excited-state lifetime of
Li(I) with increasing propane flow was measured using LEAFS. In itself,
this appeared to indicate that the fluorescence intensity should
decrease as the propane flow increased, but just the opposite trend
occurred. The reason for this increase in AFS signal seems to be linked
to the increase in the electron number density (n ) when propane was
introduced. As is shown in Figure 11, the n increased with increasing
propane flow which in turn reduced the population of Li(II) while
increasing the Li(I) population. This is also demonstrated in
Figures 12 and 13 which show the decrease in Ca (II) AFS signal while
the Ca (I) signal increased with increasing propane. Because only the
atom lines are observed in HCL-ICP-AFS, which is the technique which
utilizes propane as a reductant, the increase in atom population offset
the decrease in quantum efficiency, thereby producing a larger AFS
signal for the non-refractory elements.
The reducing nature of the plasma when an organic gas is added
produced an enhancement in the refractory elements' fluorescence as
well. This was primarily due to the reduction of oxygen and hydroxyl
radical concentrations in the plasma caused by the combustion of
propane. When propane is added, the plasma emits the green light
0 10 20 30 40
Figure 10. Li lifetime vs power.
1.0- I I
0 10 20 30 40
Figure 11. Electron number density vs propane flow. The
relative number density was determined by the
Starke broadening of the H(a) line.
Ca AFS vs propane flow. Both the atomic and ionic
fluorescence signals were normalized to a signal of
unity at no propane flow.
0 10 20 30 40
Figure 13. Ca atom/ion AFS vs propane flow.
characteristic of CH and C2 emission. When the water to the nebulizer
was stopped and only dry Ar was introduced into the plasma, this green
emission increased dramatically, demonstrating qualitatively that there
may have been a combustion process occurring between propane and water.
This was quantitatively supported by the decrease in OH emission, as
shown in Figure 14, as the propane flow increased.
Although the addition of propane to the ICP decreased the fluor-
escence quantum efficiency, the subsequent increase in electron number
density increased the fluorescence signals for the non-refractory
elements by suppressing ionization in the plasma. For the refractory
elements, the propane decreased the amount of refractory oxide species
by reacting with the oxygen and hydroxyl radicals, producing an increase
in the free metal atom population which can be excited and produce
Although most of the elements investigated fall neatly into the
categories of refractory or non-refractory depending upon their molecu-
lar dissociation energies and behavior with changing r.f. power, the
alkaline earth elements' behavior appeared to be a combination of these
two behaviors. This may have been due to a complex partial thermo-
dynamic equilibrium between these atoms and the decomposition products
of water. Although these results are inconclusive, they seem to indi-
cate that water and its decomposition products play a major role in the
atomization/excitation process in the extended sleeve torch ICP .
This is especially apparent for the Ar(m) population. Work is presently
in progress to determine spatially resolved populations of hydroxyl,
0 10 20 30
Hydroxyl AES vs propane flow. The OH emission
was observed at 306.3 nm while aspirating water
into the ICP.
hydrogen, and oxygen species in the ICP. These measurements may provide
an explanation of these unexpected curves, and contribute to a better
understanding of the extended sleeve torch atom reservoir.
By measuring the lifetime of Na and Li, it was demonstrated that
the low quenching cross-section Ar ICP is an excellent atom/ion
reservoir for fluorescence measurements, the quantum efficiency for
Na(I) between 400-1500 W being unity. Even when the quantum efficiency
was lowered by introducing propane into the plasma, the ICP retained its
integrity as an atomization cell. This was because the increase in n
and the reduction of oxygen in the plasma increased the atom/ion ratio
and diminished the formation of refractory oxide species. This is
especially important to atomic fluorescence measurements with hollow
cathode lamp excitation; however, with a second ICP as the excitation
source, ionic transitions can be used and the reduction of refractory
molecules can be accomplished by increasing the r.f. power.
FIGURES OF MERIT FOR ICP-ICP-AFS
Detection Limits for ICP-ICP-AFS
As was demonstrated in Figures 2 and 3, the non-refractory elements
exhibit their maximum fluorescence signal at low r.f. powers while the
more difficult to atomize refractory elements exhibit their maximum
sensitivity at approximately 1200 W. For this reason, the LODs of the
non-refractory elements were determined at 700 W while those of the
refractory elements were measured at 1200 W. The detection limits are
defined as the concentration which gives a signal equal to twice the
standard deviation of sixteen consecutive blank readings. The LODs
obtained by ICP-ICP-AFS for twenty-two elements are listed in Table 4
along with those obtained by HCL-ICP-AFS and ICP-AES. A comparison of
detection limits for several spectometric methods is presented in more
detail in Appendix C.
With the exception of Na and K, the detection limits for the
non-refractory elements are similar for all three methods. For Na, the
fluorescence techniques were superior to emission because of the high
degree of ionization of this element low in the AES plasma. By using an
observation height higher in the plasma tail, the temperature and degree
of ionization decreased , yielding a larger atom population to be
probed by AFS. Potassium also suffered from a high degree of ionization
which resulted in a poor AES LOD. In ICP-ICP-AFS, the high degree of
Table 4. Comparative Detection Limits in ng/mL (S/N=2).
Element AFS,nm ICP-ICP-AFS HCL-ICP-AFSb ICP-AESc
Ca 393.4 0.4 0.4 0.1
Cr 357.9 10 5 4
Cu 324.7 0.4 1 4
K 766.5 100 0.8 300
Mg E279.1-280.3 0.2 0.3 0.1
Na E589.0-589.6 1 0.3 20
Pt 214.4 30 75e 20
Sr 407.7 0.2 2 0.3
Zn 213.9 2 0.4 1
Al E394.4-396.2 10 15 20
B E249.7-249.8 10 2000 3
Ba 455.4 0.9 50 0.9
Hf E263.9-264.1 30 -- 10
Ho E345.3-345.6 10 -- 4
P E253.4-255.5 80 -- 50
Si Z251.4-252.9 7 300 8
Sm Z359.3-360.9 20 -- 30
Th Z283.2-284.3 100 -- 40
V Z309.3-310.2 40 300 3
Y E360.1-361.1 20 500 2
Yb 369.4 10 -- 1
Zr 339.2 10 -- 5
This work only. The summation sign indicates that more than one
fluorescence transition falls within the monochromator spectral
ionization in the source ICP resulted in a poor atomic source irradiance
as compared to the HCL emission. Even when 1% Cs was added to the 2% K
excitation solution, there was no increase in the fluorescence sensi-
The LODs for the refractory elements for ICP-ICP-AFS and ICP-AES
were quite similar, with the exception of V. The most intense V fluor-
escence signal was at the 309-310 nm lines which were in the midst of
the OH band. The increased plasma emission in this region increased the
background shot noise and produced a higher LOD.
The HCL-ICP-AFS LODs for the refractory elements differ signifi-
cantly from the other two methods. This is primarily due to the poor
atomization efficiency of these elements and the use of HCLs as exci-
tation sources. At low r.f. powers, the refractory oxide molecules are
not efficiently dissociated, while at higher powers, there is a signifi-
cant degree of ionization for many of these elements. The HCL's output,
however, is primarily atomic resonance radiation and these ionic tran-
sitions cannot be probed. For elements such as aluminum, this is not a
significant drawback because of this element's high ionization poten-
tial. For elements like Ba and Sr, the HCL's inability to excite
efficiently these ionic transitions results in comparatively higher LODs
because the atomic transitions must be used. In general, HCL-excited
ICP-AFS will be comparable to ICP-AES and ICP-ICP-AFS when the elements
of interest form molecular oxides with low dissociation energies com-
pared to their ionization potentials. Another major drawback of
HCL-ICP-AFS is the lack of availability of high irradiance HCLs for many
elements. This is a contributing factor in the poor detection limits
obtained by HCL-ICP-AFS for B and Si, and the cause for the lack of
information for many elements such as the lanthanides.
In many applications, there is a need to determine both trace
elements as well as the major sample constituents. Even with the large
LDR of AES and AFS, it is often necessary to prepare several sample
dilutions or to use less sensitive analytical lines in order to accom-
plish this. Sample dilution is time consuming and may introduce
dilution errors and contaminate the sample, while choosing less sensi-
tive emission or fluorescence lines may increase the influence of matrix
effects and spectral interference because of the lower signal-to-noise
ratio. An alternative is to employ the resonance monochromator (RM)
technique to increase the LDR. Figure 15 illustrates typical curves of
growth for Ca AFS and RM. When the AFS curve begins to bend toward the
concentration axis because of self-absorption , the sample solution
can be aspirated into the source ICP and its emission used to excite the
fluorescence of a 100 mg/L standard solution which is aspirated into the
atomization cell ICP. By doing this, the LDR can be extended by one to
two orders of magnitude. Typical detection limits are listed in
Table 5, along with comparative upper linear concentrations.
Because the emission from the excitation ICP has a similar spectral
bandwidth as the absorbing atoms in the atomization cell ICP, the entire
spectral line is integrated , which leads to upper linear concentra-
tions as high as 10,000 mg/L. Atomic emission spectrometers generally
employ a very narrow spectral bandpass in order to reduce spectral
interference. As the concentration of the analyte increases and the
( m 0 -Y 0 -
^t- \. H
^ X.E ^ -^ -
S \ ^
0t \ cuO^
Table 5. Resonance Monochromator Detection Limits (mg/L) and
Comparative Upper Linear Concentrations (mg/L).
Element ICP-ICP-RM Upper Linear Concentration
LOD ICP-ICP-RM HCL-ICP-AFSa ICP-AES
Ba 1 5,000 -- 100
Ca 1 3,000 90 20
Cr 2 5,000 500 150
Cu 3 5,000 75 150
Mg 3 10,000 20 50
Na 20 2,000 10 200
Zn 4 5,000 20 150
spectral line begins to broaden, the emission spectrometer only views
the line center. When self-absorption begins to occur, it affects the
line center while the wings of the broadened line continue to increase
in intensity, but this is not seen by the detector. A similar spectral
bandpass limits the LDR of HCL-ICP-AFS; however, in this instance, the
narrow source profiles of the HCLs are the limiting factor. The HCLs,
because they are at low pressure and at a significantly lower trans-
lational temperature than the atoms in the ICP, have a much narrower
emission profile than the absorption profile of the atoms in the ICP.
Therefore, even though a wide spectrometer (filter) bandpass can be
employed without increasing the effect of spectral interference, only
the line center is being measured as in AES. Another limiting factor is
the use of large bandpass filters in HCL-ICP-AFS which can have a
bandpass of 2-12 nm. This tends to pass a large fraction of the plasma
background and analyte emission signal which can saturate the photo-
multiplier tube at high analyte concentrations, even though the AFS
signal may still be linear.
Elaborate background correction techniques are often necessary in
ICP-AES in order to reduce the effect of spectral interference. This
usually requires a computer-assisted scan of the spectral region near
the line of interest and a deconvolution of the interfering and analyte
emission lines. Fluorescence spectra, when a line source of excitation
is used, are inherently simple, containing only the fluorescence lines
of interest in most cases. Comparisons of ICP-AES and ICP-ICP-AFS
backgrounds are shown in Figures 16 and 17. In scan (a) of Figure 16,
Figure 16. Ca spectral interference on Al. a) 1 mg/L Ca and
50 mg/L Al AES, b) 1000 mg/L Ca and 50 mg/L Al AES,
c) 1000 mg/L and 50 mg/L Al AFS with 2% Al excitation.
Figure 17. Al spectral interference on Zn. a) AES from 10 mg/L
Zn, b) AES from 10 mg/L Zn and 1000 mg/L Al, c) AFS
from 10 mg/L Zn and 1000 mg/L Al with 2% Zn excitation.
the emission spectrum of a solution containing a low concentration of Ca
and Al exhibits the doublets of these two elements as well as an Ar
doublet. Scan (b) shows the effect of increasing the Ca concentration
on the Al peaks. The Al doublet now lies in the valley between the
two collisionally broadened Ca peaks  making background correction a
necessity. The fluorescence spectrum (c) of the same solution, using
2% Al in the excitation source ICP, demonstrates that there was no
increase in the background around the Al peaks, although the background
shot noise at the Ca wavelengths was increased. Also absent in the AFS
spectrum were the Ar emission doublet and the plasma recombination con-
This type of spectral interference is well documented in the
literature [11,46,47] and can be corrected by appropriate background
correction techniques developed for AES. Many spectral interference,
however, are not listed in the prominent line tables and must be
evaluated on a sample by sample basis. This is illustrated in Figure 17
by the Al interference on Zn AES. Scan (a) shows the emission spectrum
of 10 mg/L Zn demonstrating the relative simplicity of the plasma
background near 200 nm. However, when 1000 mg/L of Al was added to this
Zn solution, the background recombination continuum increased and
several Al transitions appeared around the Zn(I) line. The wavelength
separation of the Al and Zn peaks is sufficiently large that no direct
spectral overlap occurs, but the increase in background must be sub-
tracted from the Zn AES signal. Although no spectral interference
occurred, this points out the potential for interference to occur from
interfering lines which are not considered "prominent lines" and are not
commonly tabulated in the literature. The AFS spectrum of the Zn and Al
solution, shown in scan (c), once again demonstrates the simplicity of
fluorescence spectra which enables one to use AFS to analyze real
samples without the need for background correction in most cases.
There are several types of background sources in the ICP which may
cause spectral interference in AFS. The plasma and other ion recombi-
nation continue  are emission phenomena and do not produce fluores-
cence signals. Molecular fluorescence of species such as metal oxides
or OH radicals are not efficiently excited by the source ICP to produce
any noticeable fluorescence signal. For example, no fluorescence signal
was observed for A10, even at an Al concentration of 10 g/L. Atomic or
ionic emission lines within the spectrometer bandpass only add to the
background shot noise and do not add to the modulated fluorescence
signal. Their presence can be reduced by increasing the integration
time on the electronic measurement system, but their presence will
degrade detection limits depending upon the interferent concentration in
the matrix by increasing the blank noise. Direct spectral overlap of
two elements does present a small problem in ICP-ICP-AFS. Several con-
trived cases of direct spectral overlap are presented in Table 6. The
majority of the interferent wavelengths do not appear in Boumans' "Line
Coincidence Tables"  or in the "Prominent Lines Table" of Winge et
al. . Many of the reported spectral interference encountered in
ICP-AES  were evaluated using ICP-ICP-AFS, but only a small fraction
of these interference were detectable at concentrations below 10 g/L.
For a spectral interference to occur in AES, the interfering
transition must fall within the spectral bandpass of the monochromator.
In AFS, the criteria for spectral interference are much more stringent
and depend upon a) the interfering specie's absorption cross section
0 0000 0
0 0 0
\. l10 0MnO 0
mo \%m- rfonW
^o\o ~Qr-~-O\O UI t> m
O mc C C r rN
cf) T~ T3 T3 Mn Mn Mn T ~-Z CI "M CM4 cM CM4 cM CM
-r ( wj cd 3
<4 pq U U Uz
-T 3' M c4 - CM r- O M 0 M- O
r- 00 r-. CT
c0 CM 10 -4
-i ~3 cn c
(71 C (01 Cl)
r- CIA 1. I
O 4-0 0 W W ) 00 U E E
z N21 Nu =u 4 u-. z 4 ;4 4 zz > >> >m
M 0 0m
Ufl C,4 CM4
C 0 0
0 00 00
00 00 00
over the emission source's spectral profile; b) the population of the
interfering species in the proper energy level; and c) the fluorescence
quantum efficiency of the interfering species within the detector's
spectral bandpass. Therefore, substantially fewer spectral interfer-
ences occur in AFS as compared to AES.
The inert, high temperature environment of the ICP makes it an
excellent atomization cell for atomic spectrometry. A relatively high
degree of freedom from interelement (matrix) effects is a major advan-
tage of the ICP over flame, furnace, and microwave discharge atom
The solute vaporization interference of sulfate and phosphate 
on Ca were investigated by measuring the fluorescence signal at 393.4 nm
of a 1 mg/L Ca solution while varying the interferent concentration.
Solutions containing up to 50,000 mg/L SO or PO produced no depres-
sion or enhancement of the fluorescence signal. This is due to the high
temperature of the plasma and the long residence time of the sample in
the plasma plume.
Figure 18 illustrates the interference of Na on Ca which resulted
from the shifting of the ion/atom equilibrium . There was a minimal
effect upon the Ca fluorescence signal at Na concentrations below
10 mg/L. However, at higher concentrations, the effects became pro-
nounced. The onset of the Na interference occurs at a Na number density
of approximately 1 x 10 cm which is comparable to the n trends
reported in reference 49 at a temperature of 4000 K (Fe Boltzman plot).
S10 100 1,000 10,000
No CONCENTRATION (Lg/mL
Figure 18. Effect of Na on Ca AFS. 0 Ca(I), 0 Ca(II).
The ionization interference occurs because of the introduction of
an element which is easily ionized in the plasma, and which produces a
shift in the ion/atom equilibrium. This equilibrium is described by the
Saha equation, which states that the ratio of ions to atoms in a given
system is constant at a specified temperature and n :
+log e -5040IP 5 6.1818
log T + log T + log 6.1818
where n+ and n are the number densities of the ions and atoms, respec-
tively, IP is the ionization potential of the atom (in eV), T is the
temperature (in K), and Q and Qo are the ionic and atomic partition
functions, respectively. When Na is introduced into the plasma, n
increases due to the ionization of Na atoms, which have a low I.P. This
reduces the ion/atom ratio for Ca, producing a larger Ca(I) population
in the plasma. Because of the lower temperature and n under the
present operating conditions as compared to typical emission parameters,
the ionization interference were slightly worse in ICP-ICP-AFS compared
to ICP-AES . The enhancement of the Ca(I) signal was much larger
than the depression of the Ca(II) fluorescence which was due to the
higher population of Ca(II) and its relatively larger fluorescence
sensitivity. The small depression in the large ion population caused by
Na caused a large change in the relatively small atom population. This
also illustrates the increased interelement effects which may be encoun-
tered if one were to choose a less sensitive emission or fluorescence
line for analysis in order to increase the LDR or to avoid spectral
Several elements were determined in NBS SRM 364 (high carbon steel)
with the atomization cell ICP operating at 1200 W. The weight percent
solid concentrations ranged from 96.7% Fe to 0.0005% Zn. Only one
sample solution (1.2206 g/L) was necessary to determine these elements
by AFS (Cr, Cu, Zn) and RM (Fe), with the average relative error and
relative standard deviation both being 1%. However, we did experience a
problem with the analysis of Co. The measured concentration was
1.5 times the certified value. The same high value was measured at two
different Co lines, which indicated that a spectral interference was
probably not the cause. Zeeman AA yielded the correct result, which
indicated that sample contamination did not cause the discrepency. This
may have been due to a vaporization interference, but this had not been
confirmed. No matrix matching of standards nor matrix blank was used.
Only a solvent (acid) blank was employed. Even though the sample
contained a large Fe concentration, no spectral interference were
encountered, despite the hundreds of Fe emission lines.
A log-log plot of S/N vs integration time displays a slope of 0.5,
indicating that the present system is at the shot noise limit. There-
fore, detection limits could be improved by increasing the signal
integration time, but this improvement in LODs would only increase as
the square root of the integration time. Further improvements in LODs
may come about by using more efficient source collection optics, but
this may be limited to a factor of two or three over the present system
which employs f 1 optics. The major instrumental modification necessary
for this system to become a viable analytical instrument would be to
combine the two r.f. generators and run the plasmas off a single 3 kW
generator. The matching units would have to be slightly modified, but
two 100 ohm units could be run in parallel off the single generator
The versatility of ICP-ICP-AFS and RM is evident from its large
LDR, up to 108, and LODs which are comparable to ICP-AES. The system
should be considered an alternative to emission spectrometry in order to
alleviate spectral interference which may occur in complex sample
matrices, without the need for an expensive, high resolution mono-
FINAL COMMENTS AND SUMMARY
The effect of r.f. power on the shapes of the fluorescence curves
has demonstrated that, in the extended sleeve torch, there are three
groups of elements: the refractory, non-refractory, and alkaline earth
elements. The behavior of the alkaline earth fluorescence signal vs
power seems to indicate a complex relationship between the metal atom
and the decomposition products of water. These elements form stable
oxide, hydroxide, and dihydroxide molecules in flames and in plasmas
which may account for this behavior. This work is inconclusive at this
time because the measurements were all line of sight measurements, which
do not afford the spatial resolution necessary to state conclusively the
mechanism of this behavior. It has demonstrated, however, that care
must be taken in the choice of an element which will be used to model
the ICP, for example, calcium in many past studies. Also, the influence
of water on the mechanism of the ICP must be considered, as was demon-
strated by the dramatic effect water can have on the Ar(m) population.
The effect of propane on the atomization mechanism in the plasma
has been investigated using fluorescence and emission spectrometry. The
inert, monoatomic Ar atmosphere of the ICP has been shown to be an
excellent atom and ion atomization cell for atomic fluorescence spec-
trometry because of its high atomization efficiency and quantum
efficiency. Propane, however, decreases the quantum efficiency of the
plasma, although it does aid in the atomization of refractory elements
and shifts the Saha equilibrium to produce a larger atom population by
suppressing ionization in the ICP. Although this may enhance the
fluorescence signal when only atomic transitions are considered, as in
HCL-ICP-AFS, it does not improve the S/N of ICP-ICP-AFS .
The figures of merit for ICP-ICP-AFS are equal or superior to many
atomic methods of analysis. The detection limits are comparable to
ICP-AES for both the refractory and non-refractory elements, and the
linear dynamic range is superior to all spectroscopic methods. This
large LDR, up to eight orders of magnitude, is accomplished by combining
the fluorescence and resonance monochromator curves of growth. The
selectivity of the fluorescence technique when a narrow line source of
excitation is used, as the ICP, results in very few spectral interfer-
ences. Even in the instances when these interference occur, the
interferent detection limit is usually three or more orders of magnitude
higher than the LOD for the element of interest. This allows the use of
a low resolution spectrometer for fluorescence detection without the
need for background correction, as compared to ICP-AES which requires a
high resolution monochromator and elaborate background correction
The use of the extended sleeve torch in fluorescence spectrometry
also provides a long residence time for the sample in the hot plasma
tail before it reaches the observation zone. This yields a high atomi-
zation efficiency and reduces the effect of solute vaporization inter-
ferences which can plague ICP-AES and flame spectrometry. By viewing
the fluorescence signal at a height of 55 mm above the load coil, the
electronic temperature of the atomic vapor is relatively low producing a
large free atom or ion ground state population to be probed while
reducing the plasma background emission which can contribute to the
background shot noise. However, because the plasma temperature is low,
the electron number density is quite low which leads to a more pro-
nounced ionization interference when large quantities of easily
ionizable elements are present in the sample matrix.
The ICP-ICP-AFS system should be considered a viable alternative to
ICP-AES when a large LDR is necessary and where spectral interference
must be minimized. However, this system is not expected to become
commercially available because of several drawbacks, but could be
assembled from existing laboratory equipment when the need for its
special attributes is warranted. The drawbacks which limit the fea-
sibility of this instrument are the need for two generators which may
make the initial cost of such an instrument comparable to an ICP emis-
sion system with a high resolution polychromator. The operating costs
are expected to be double that of an ICP-AES instrument because of the
increase in power and argon consumption. Two minitorches, which are
becoming popular in ICP-AES instruments, could be employed, thereby
reducing the Ar consumption rate. Also, a single r.f. generator could
be configured to run both plasmas. The instrument also has the drawback
of being a sequential system, capable of analyzing only a single element
at a time whereas several ICP emission systems permit simultaneous
detection of up to fifty elements. The excitation solutions could be
supplied to the excitation ICP by means of an auto-sampler, and the
monochromator could be controlled so as to slew scan to each wavelength.
This would automate the system to the point where it could be
competitive with sequential ICP-AES systems; however, the multi-element
capabilities of ICP emission spectrometers with a direct reader could
never be matched. Lastly, aspiration of a 2% excitation solution for
long periods of time may become expensive, as in the case of the
precious metals, and may be dangerous as in the case of many toxic
metals such as Be, Hg, Pb, etc. The possible health hazard of aspirat-
ing concentrated solutions of toxic metals into the source ICP could be
minimized or eliminated if a good hood/ventilation system was an
integral part of the instrument. However, aspirating 2% solutions of
the precious metals for extended lengths of time will be prohibitively
Nonetheless, employing an existing ICP emission system as an
excitation source for exciting fluorescence in a second ICP may provide
valuable information on a sample which could not be obtained by
other atomic techniques.
ICP emission spectrometers have been slowly replacing atomic
absorption spectrometers in the laboratory for routine analysis, and
this trend seems likely to continue in the future. The main reasons for
this are the larger linear dynamic range, excellent precision, reduced
matrix effects, and multi-element capabilities of ICP-AES. Atomic
fluorescence, on the other hand, possesses many of the attractive
advantages of ICP-AES when an ICP is employed as an atomization cell;
however, fluorescence systems are lacking in easy to use, high intensity
excitation sources which may make them competitive with AES. The ideal
AFS system of the future, in my opinion, will be a combination of ICP
emission and fluorescence spectrometers. With a high resolution
polychromator and the system operating in the emission mode, routine
multi-element analyses can be performed on a simultaneous basis. When
spectral interference occur which cannot be corrected by background
correction techniques, the fluorescence mode can be employed. In this
mode, a high intensity pulsed laser-dye laser combination can be
employed to excite fluorescence of the atoms or ions of interest in the
emission ICP. Because of the narrow bandwidth of the laser and the
capability of doing non-resonance fluorescence, spectral interference
can be eliminated. Also, the fluorescence system can be employed when
lower detection limits than AES must be achieved.
The complexity of present laser systems, and the high cost, at
present make this system infeasible; however, the rapid developments in
the field of electro-optics may produce a user friendly laser capable of
satisfying the needs of an analytical instrument in the near future. I
project that within 10 years, this type of ICP emission/fluorescence
instrument will be commercially available to satisfy most analytical
needs of the atomic spectroscopist. The knowledge we have gained by
investigating atomic and ionic fluorescence in an ICP will play an
integral part in development of this optimal atomic spectrometer.
GLOSSARY OF ACRONYMS AND ABBREVIATIONS
Einstein coefficient of spontaneous emission
atomic absorption spectrometry
atomic emission spectrometry
atomic fluorescence spectrometry
Einstein coefficient of stimulated absorption (emission)
electrodeless discharge lamp
electro-thermal atomizer (graphite furnace)
full width at half-maximum
hollow cathode lamp
inductively coupled plasma
ICP-excited ICP atomic/ionic fluorescence spectrometry
inductivity coupled plasma-mass spectrometry
collisional rate constant
linear dymanic range
laser-excited atomic fluorescence spectrometry
laser enhanced ionization
limit of detection
local thermodynamic equilibrium
national bureau of standards
parts per billion
partial thermodynamic equilibrium
radio frequency radiation
standard reference material
QUANTUM EFFICIENCY MEASUREMENTS
The quantum efficiency (Y) of a transition is the fraction of the
absorbed energy which is converted into a fluorescence signal. The
quantum efficiency is a guage of the quenching environment of an atom
reservoir, and has a direct bearing on the intensity of the fluorescence
signal. In analytical atomic fluorescence spectrometry, a high quantum
efficiency atomization cell is desirable. A description of a simple
2-level system (Figure 19) is now presented and used to illustrate the
significance and measurement of the quantum efficiency. Several excel-
lent treatments of the intensity of transitions in spectrometry are
available in the literature [30, 51-54], and the reader is referred to
these for a more extensive description of these processes.
If an external light source impinges on an atom reservoir, two
induced processes may occur. An atom in the ground state may absorb a
photon with the probability per unit time of this stimulated absorption
given by the Einstein probability of induced absorption (B12) multiplied
by the spectral energy density of the source (P ). The reverse process,
stimulated emission, is governed by a similar expression, and has a rate
of B2 P. Atoms in an excited state may emit a photon at an emission
rate which is defined by the Einstein coefficient of spontaneous emis-
sion (A21). These three processes are governed by the Einstein
-- -- 2
Figure 19. Schematic diagram of a 2-level atom. The ground
-state, level 1, has a population nI while the excited
state, level 2, has a population n2. The energy
difference between the two levels, AE, is hvo.
probability coefficients which are intrinsic properties of an atom, and
do not depend on the atom's environment.
There are two other processes to be considered in the description
of the 2-level system which are governed by the properties of the atom
reservoir. Two collisional rate constants, k12 and k21, define the
nonradiative excitation and de-excitation processes which an atom may
experience. In a given atom reservoir, an atom may be excited by a
collision with another specie at a rate of kl2. Analogously, an atom in
an excited state may decay to the ground state upon collision with
another specie without the emission of a photon. This radiationless
deactivation rate is expressed as k21.
Because of the low source spectral irradiance of the laser employed
in this study, the rate of stimulated emission (B21P ) can be neglected,
because a linear interaction between the source radiation and the atomic
system is maintained. In other words, B21P is negligible if saturation
is not approached.
Lifetime measurements of a 2-level atom can be used to determine
the quantum efficiency. The spontaneous radiative lifetime of a transi-
tion is defined as
while a purely nonradiative lifetime can analogously be defined as
The fundamental parameter that defines the influence of collisions which
depopulate the excited state, namely the quantum efficiency of the
transition (Y21), can be described by combining the above two expres-
sions. The quantum efficiency is the probability that the excited atom
will lose its energy by the emission of a photon, and is defined as
21 A 21 + k21
in terms of the rate coefficients, or in terms of the lifetime of the
where Tobs is the observed or measured lifetime of the transition.
Therefore, by measuring the excited state lifetime and knowing the
spontaneous lifetime (or A21), a value-for the quantum efficiency of the
transition can be obtained. Reports of lifetime measurements in the
literature are very scarce because most atomic T are on the order of a
few ns. Also, many atoms reservoirs, such as flames, have very low
quantum efficiencies which shortens the observed lifetimes to values
below 1 ns, which are very difficult to measure.
When a temporally narrow excitation light source impinges upon an
atom reservoir containing Na atoms at time t = 0, the fluorescence decay
curve depicted in Figure 20 is produced. As can be seen, simply by
measuring the time it takes for the fluorescence intensity to decay by
z Na(I) 589.Onm
S \ rp= 16ns, Y= 1.0
Figure 20. Fluorescence decay curve for Na.
50% (or some other convenient value), one can determine the radiative
lifetime from the exponential decay expression
F(t) = F e
where F(t) is the fluorescence signal at time t, F is the maximum
fluorescence signal, and T is the observed lifetime of the transition.
From the measurement of the time it takes for the fluorescence signal to
decay to 50% of its maximum value (t = t50 = 0.5), the above
equation reduces to
This was the procedure used in this study to determine the quantum
efficiencies of Na and Li in the ICP. A high value for Y demonstrates
that the atom reservoir is an excellent choice for atomic fluorescence
measurements, as long as the atomization efficiency is also high. For
the ICP, both the quantum efficiency and atomization efficiency are very
high, making it an attractive atomization cell and atom reservoir for
COMPARATIVE DETECTION LIMITS
The ultimate goal of analytical atomic spectrometry is to be able
to detect single atoms. Single atom detection has been achieved in
certain special cases, and schemes have been proposed to determine
single atoms by a variety of methods [55-56]. However, most analytical
methods possess detection limits in the parts per billion or parts per
trillion range, clearly, many order of magnitude above single atom
detection. But what is the significance of these detection limits?
Surely if one technique possesses lower LODs, it would be the method of
choice for a particular analysis. Unfortunately, this is the view of
many analytical chemists, and can lead to tragedy if all the figures of
merit for a particular technique are not viewed together.
In 1978, the International Union of Pure and Applied Chemists
(IUPAC) adopted a model for the calculations of LODs , a model which
was also approved by the American Chemical Society (ACS) Subcommittee on
Environmental Analytical Chemistry in 1980 . Even though a standard
method for determining LODs has been presented, many spectroscopists,
and chemists in general, refuse to employ this method and continue to
calculate LODs by their own methods. This can lead to a great diversity
in detection limits reported for the same method, which leads to LODs
that can easily vary by an order of magnitude or more through the use of
different statistical methods . For the purpose of this work, the
IUPAC definition was used to calculate the LODs. This definition is:
LOD = ,
where k is a scale factor which determines the confidence level, Sb is
the standard deviation of the blank readings, and m is the slope of the
calibration curve, often referred to as the sensitivity. Typically, a k
value of 2 or 3 is chosen in the calculation of the LOD: however, k
values vary depending on the researcher. The standard deviation of the
blank, Sb, also depends upon the method used to calculate and measure
this quantity. Because the standard deviation is only valid for an
infinitely large number of measurements, Sb is only an approximation of
the actual standard deviation and depends on the number of blank reading
used to calculate this parameter. When reported in the literature, most
standard deviations are calculated from 10 to 25 measurements, although
reports of standard deviations for only two measurements have made their
way into publication.
It is clearly evident that there exists a diversity of methods used
to determine LODs , and extreme care must be taken when comparing
detection limits from different sources. LODs should only be used as
approximations of the minimum concentration which can be detected under
the optimum conditions (detection limits are typically determined under
the optimal experimental conditions using the simplest matrix available,
usually, high purity salts dissolved in the purest water available).
Actual quantitative determinations usually require an analyte
concentration between 5 and 10 times the calculated LOD, while some
samples with complex matrices require the analyte to be present at 50 to
100 times the LOD before meaningful results can be obtained.
Bearing all these factors in mind, Table 7 presents a comparison of
typical detection limits and linear dynamic ranges for several atomic
spectrometric methods. A more complete comparison of LODs and the
historical progression of detection limits over the past two decades is
given in reference 60.
At this point, a few comments on Table 7 are necessary to put the
tabulated values in perspective. Methods employing electrothermal
atomizers, commonly referred to as graphite furnaces, possess some of
the best detection limits obtained. However, furnaces suffer from many
vaporization interference which usually require the use of matrix
modifiers to reduce these effects . In many cases, a different
matrix modifier is required for each element, or group of elements,
which is to be determined in a complex sample matrix. This tends to
make sample preparation and methods development very time consuming, and
does not insure the elimination of matrix effects.
Methods which employ lasers as excitation sources, AFS and laser
enhanced ionization (LEI), also exhibit very low LODs. However, lasers
are notoriously noisy, having shot-to-shot instabilities ranging from a
few percent up to 20%. These instabilities, or pulse-to-pulse shot
noise, can be eliminated if the laser has enough power to saturate the
transition being studied ; however, most laser systems available
today cannot optically saturate transitions below approximately 300 nm.
This is a significant drawback, since most of the intense atomic reso-
nance transitions lie between 180-300 nm. Another significant
Table 7. Comparison of Atomic Spectrometric Methods.
METHOD SOURCE CELL LOD(ppb)a LDR(decades)b
AFS Laser Flame 1-1,000 3-6
ETA 0.001-1,000 2-5
ICP 0.1-1,000 3-6
HCL ICP 0.1-1,000 3-5
Xe Arc Flame 1-1,000 2-4
ICP ICP 0.1-100 4-8
Flame 1-1,000 4-7
LEI Laser Flame 0.001-100 3-6
AAS HCL Flame 1-1,000 1-3
ETA 0.001-100 1-3
AES ICP 0.1-100 2-5
Flame 0.1-10,000 2-5
ICP-MS ICP 0.01-10 2-4
aRange of LODs, in parts per billion, commonly found in the literature.
bLinear working range of the calibration curve in decades (orders of
drawback to employing laser systems is that they are expensive, in the
initial cost of the system and in the maintenance costs. These systems
have also not become "user friendly," as have many computer systems over
the past few years, making them difficult to operate on a daily basis.
Because of their complexity, it is often necessary to spend several
hours to tune a laser to a particular atomic transition, making them
ineffective as multi-element sources.
Inductively coupled plasma-mass spectrometry (ICP-MS) is the most
recently introduced technique listed in Table 7. On the surface, this
technique has excellent detection limits and has the unique capability
of being able to determine isotope ratios of the elements contained in
the sample [61-62]. However, ICP-MS is plagued with spectral interfer-
ences and many matrix effects. The spectral interference arise from
the many isotopes of both atomic and molecular species which are present
in the ICP. For many elements, especially the refractory elements, a
large percentage of the total element is in the form of molecular
species, typically oxides and hydroxides. If these molecular peaks
coincide with the analyte element's peak, a spectral interference, not
unlike spectral interference in ICP-AES, will occur. Also, many matrix
effects exist in ICP-MS. Although the instruments manufacturers refer
to these matrix effects as "apparent matrix effect," they are very real,
and may eventually cause this technique to be abandoned within the next
few years if these problems are not resolved .
In summary, LODs are useful in comparing one spectroscopic tech-
nique to another, if the same method is used for their determination.
Beyond only a superficial comparison, though, LODs should be used in
conjunction with all the figures of merit for a given method of
analysis. Linear dynamic range, spectral and matrix interference, ease
of operation of the instrument, sample preparation requirements, and
instrument cost should all be incorporated into the evaluation of a
particular technique. Also, many applications do not require detection
limits in the parts per trillion range. In these instances, a particu-
lar technique should be chosen on the basis of whether or not it can
satisfy the analysis requirements, and not on the basis of which tech-
nique possesses the best LODs. In short, LODs are a guide to analytical
performance, and not the only figure of merit which describes the
utility of the technique or method.
 T.B. Reed, J. Appl. Phys. 32, 821 (1961).
 S. Greenfield, I.L. Jones, and C.T. Berry, Analyst 89, 713 (1964).
 R.H. Wendt and V.A. Fassel, Anal. Chem. 37, 920 (1965).
 N. Omenetto, S. Nikdel, J.D. Bradshaw, M.S. Epstein, R.D. Reeves,
and J.D. Winefordner, Anal. Chem. 51, 1521 (1979).
 M.S. Epstein, N. Omenetto, S. Nikdel, J. Bradshaw, and J.D.
Winefordner, Anal. Chem. 52, 284 (1980).
 N. Omenetto, P. Cavalli, and G. Rossi, Rev. Anal. Chem. 5, 185
 M.A. Kosinski, H. Uchida, and J.D. Winefordner, Anal. Chem. 55,
 G. L. Long and J.D. Winefordner, Appl. Spectrosc. 38, 563 (1984).
 R.J. Krupa, G.L. Long, and J.D. Winefordner, Spectrochim. Acta
40B, 1485 (1985).
 H.G.C. Human and R.H. Scott, Spectrochim. Acta 31B, 459 (1976).
 R.K. Winge, V.J. Peterson, and V.A. Fassel, Appl. Spectrosc. 33,
 D.R. Demers, Spectrochim. Acta 40B, 93 (1985).
 G.F. Larson and V.A. Fassel, Anal. Chem. 48, 1161 (1976).
 G.F. Kirkbright and M. Sargent, Atomic Absorption and Fluorescence
Spectroscopy, Academic Press, London (1974).
 J.W. Carr, M.W. Baldes, and G.M. Hieftje, Appl. Spectrosc. 36, 689
 S.R. Koirtyohann, Personal communication, University of Missouri
 C.A. Monnig and S.R. Koirtyahann, Appl. Spectrsc. 39, 884 (1985).
 L.M. Fraser and J.D. Winefordner, Anal. Chem. 44, 1444 (1972).
 D.R. Demers and C.D. Allemand, Anal. Chem. 53, 1915 (1981).
 J.J. Leary, A.E. Brooks, A.F. Dorrzapf, Jr., and D.W. Golightly,
Appl. Spectrosc. 36, 37 (1982).
 M.W. Routh, P.A. Wartz, and M.B. Denton, Anal. Chem. 49, 1422
 L.R. Porker, Jr., N.H. Tioh, and R.M. Barnes, Appl. Spectroc. 39,
 S. Greenfield, P.B. Smith, A.E. Breeze, and N.M.D. Chilton, Anal.
Chim. Acta 41, 385 (1968).
 B. Magyar and F. Aeschbach, Spectrochim. Acta 35B, 839 (1980).
 R.H. Wendt and V.A. Fassel, Anal. Chem. 38, 337 (1966).
 J.D. Windfordner and T.J. Vickers, Anal. Chem. 36, 1947 (1964).
 M.L. Parsons, B.W. Smith, and G.E. Bentley, Handbook of Flame
Spectroscopy, Plenum Press, New York (1975).
 R. Mavrodineanu and H. Boiteux, Flame Spectroscopy, John Wiley and
Sons, Inc., New York (1965).
 C.H. Corliss and W.R. Bozman, Experimental Transition
Probabilities for Spectral Lines of Seventy Elements, NBS
Monograph 53, Washington, DC (1962).
 C. Th. J. Alkemade, T. Hollander, W. Snellman, and P.J.Th.
Zeegers, Metal Vapours in Flames, Pergamon Press, New York (1982).
 R.N. Savage and G.M. Hieftje, Anal. Chem. 52, 1267 (1980).
 M.A. Kosinski, H. Ulchida, and J.D. Winefordner, Talanta 30, 339
 M.W. Blades and G. Horlick, Spectrochim. Acta 36B, 881 (1981).
 P.J.Th. Zeegers, W.P. Townsend, and J.D. Winefordner, Spectrochim
Acta 24B, 243 (1969).
 H. Uchida, M.A. Kosinski, N. Omenetto, and J.D. Winefordner,
Spectrochim. Acta 38B, 529 (1983).
 S. Huang, D. Mo, K.S. Yeah, and J.D. Winefordner, in press, Canad.
J. Spectrosc. (1986).
 R.S. Houk, V.A. Fassel, and B.R. LaFreniere, Appl. Spectrosc. 40,
 D.R. Jenkins, Spectrochim. Acta 25B, 47 (1970).
 G.F. Kirkbright, Chapter in Developments in Atomic Plasma
Specrochemical Analysis, R.M. Barnes (ed), Heyden Press, London
 W.H. Gunter, K. Visser, and P.B. Zeeman, Spectrochim. Acta 37B,
 Plasma/AFS Comparative Detection Limits, Baird Corporation,
Bedford, MA (1985).
 Plasma/AFS Precious Metal Analysis Bulletin, Baird Corporation,
Bedford, MA (1985).
 Periodic Table of Detection Limits and Upper Linear Range for
ICP-AES, Allied Analytical Systems, Andover, MA (1985).
 P.J.T. Zeegers, R. Smith, and J.D. Winefordner, Anal. Chem. 40,
 D. Demers, Personal communication, Baird Corporation, Bedford, MA
 G.F. Larson and V.A. Fassel, Appl. Spectrosc. 33, 592 (1979).
 P.W.J.M. Boumans, Line Coincidence Tables for Inductively Coupled
Plasma Atomic Emission Spectrometry, Vols. 1 and 2, Pergamon
Press, New York (1980).
 G.R. Harrison, Massachusetts Institute of Technology Wavelength
Tables, M.I.T. Press, Cambridge, MA (1969).
 B.L. Coughlin and M.W. Blades, Spectrochim. Acta 39B, 1583 (1984).
 S. Barber, Personal communication, R.F. Plasma Products, Inc.,
Cherry Hill, N.J. (1985).
 A.P. Thorne, Spectrophysics, Science Paperbacks, London (1974).
 N. Omenetto (ed), Analytical Laser Spectrometry, John Wiley, New
 A.C.G. Mitchell and M.W. Zemansky, Resonance Radiation and Excited
Atoms, University Press, Cambridge, England (1961).
 N. Omenetto and J.D. Winefordner, Prog. Analyt. Atom. Spectrosc.
2, 1 (1979).
 G.S. Hurst, M.G. Payne, S.D. Kramer, and C.H. Chen, Phys. Today
33, 24 (1980).
 C. Th. J. Alkemade, Appl. Spectrosc. 35, 1 (1981).
 International Union of Pure and Applied Chemists, Spectrochim.
Acta 33B, 242 (1978).
 American Chemical Society, Anal. Chem. 52, 2242 (1980).
 G.L. Long and J.D. Winefordner, Anal. Chem. 55, 712A (1983).
 M.L. Parons, S. Major, and A.R. Forster, Appl. Spectrosc. 37, 411
 V. Fassel, paper no. 1, Pittsburgh Conference on Analytical
Chemistry and Applied Spectroscopy (1986).
 R.S. Houk, paper no. 2, Pittsburgh Conference on Analytical
Chemistry and Applied Spectroscopy (1986).
 R.C. Hutton, J.E. Cantle, and D.E. Kelly, paper no. 229,
Pittsburgh Conference on Analytical Chemistry and Applied
 G. Horlick, S.H. Tan, M.A. Vaughn, Y. Shao, and J. Lam, paper
no. 226, Pittsburgh Conference on Analytical Chemistry and Applied
Robert Joseph Krupa was born in Holyoke, Massachusetts, on
July 12, 1960. He attended Georgetown Preparatory School in Rockville,
Maryland, and graduated from Marquette University in Milwaukee,
Wisconsin, with a B.S. in chemistry in 1982. Since that time he has
attended the University of Florida where he received his Ph.D. in
chemistry in 1986 under the direction of Dr. James D. Winefordner.
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Jnes D. Winefordner
Graduate Research Professor of
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Jo i G. Dorsey
As ci te Professor of Chemis ry
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Eric R. Allen
Professor of Environmental Engineering
This dissertation was submitted to the Graduate Faculty of the
Department of Chemistry in the College of Liberal Arts and Sciences and
to the Graduate School and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.
Dean, Graduate School