Fluorescence dip spectroscopy of copper and silver as a diagnostic tool in several atomization reservoirs

MISSING IMAGE

Material Information

Title:
Fluorescence dip spectroscopy of copper and silver as a diagnostic tool in several atomization reservoirs
Physical Description:
vi, 257 leaves : ill. ; 29 cm.
Language:
English
Creator:
Robie, Donna Jean, 1966-
Publication Date:

Subjects

Subjects / Keywords:
Atomic emission spectroscopy   ( lcsh )
Fluorescence spectroscopy   ( lcsh )
Excited state chemistry   ( lcsh )
Copper   ( lcsh )
Silver   ( lcsh )
Chemistry thesis Ph. D   ( lcsh )
Dissertations, Academic -- Chemistry -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1993.
Bibliography:
Includes bibliographical references (leaves 249-256).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Donna Jean Robie.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001950903
oclc - 31200866
notis - AKC7445
System ID:
AA00003255:00001

Full Text







FLUORESCENCE DIP SPECTROSCOPY
OF COPPER AND SILVER
AS A DIAGNOSTIC TOOL
IN SEVERAL ATOMIZATION RESERVOIRS











By

DONNA JEAN ROBIE


DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA











ACKNOWLEDGEMENTS


I want to express my gratitude to many people at the University of Florida for

their insight and suggestions regarding the work presented in this dissertation especially


Ben Smith and Giuseppe Petrucci.


The entire Winefordner group deserves recognition,


since, at some time or another, each person has given me their support in some way.

I am especially grateful to Jim Winefordner for giving me the opportunity to learn from

the best.

In addition to professional support, I was lucky enough to have an emotional


support system including my parents and my sister.


I want to give a much deserved


thank you to Stefanie Pagano for always being there to tell me what I needed to hear,


whether I wanted to hear it or not. I also want to acknowledge Mike Naughton for

several years of support through graduate school. And I want to say thank you to Rafael


Vargas,


without whom I never would have made it through my


last months at the


University of Florida.


especially want to acknowledge


my brother,


Daniel Robie,


without whose


support and love I would not be receiving this degree.


Through his constant optimism


and patience, he has walked me through some difficult times, and there is no way I can


completely express my gratitude.


He has been a constant inspiration to me, and I hope


he will continue to share his strength with me throughout my life.













TABLE OF CONTENTS


ACKNOWLEDGEMENTS


ABSTRACT


S * . ii1


S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S a S S S S SV


CHAPTER 1


CHAPTER


INTRODUCTION


. 1


ATOMIC FLUORESCENCE SPECTROSCOPY


Conventional Source Atomic Fluorescence Spectroscopy


History
Theory


* S . 4
* S S C 4


* S . .
* . .


Types of atomic fluorescence


. 6


Theoretical


treatment


atomic


fluorescence


spectroscopy
Applicatons of Conventional Source AFS
Advantages and Disadvantages .. .


. .. ... S CSa39


. S 43


Laser Excited Atomic Fluorescence Spectroscopy (LEAFS)
History of Flame LEAFS . . .


History of Furnace LEAFS
History of ICP-LEAFS


.* S . . 52
.* . . 5 54


History of LEAFS in Other Atomization Reservoirs


History of Two Photon LEAFS .
History of Two-Color Excitation Fluorescence
Advantages and Disadvantages of LEAFS


CHAPTER 3


S* S 57


DIAGNOSTIC APPLICATIONS OF LEAFS


Diagnostic Characterization of Atom Sources . . . . .
Investigation of Atomic Parameters Using LEAFS . ... ..


Investigation of Atomic Parameters Using Two-Color LEAFS
Advantages and Disadvantages of LEAFS as a Diagnostic Tool


CHAPTER 4


* S S S
* . .


INSTRUMENTATION USED IN LEAFS


Excitation Sources


72


Principles of Lasers
Types of Lasers


.A







Flames
Plasmas


.. ....... 88


...... 92


Electrothermal Atomizers


a a a a a a a a S 96


Other Atomization Reservoirs used in LEAFS


Wavelength Selection Devices
Detection Methods used in LEAFS


. . . . 98


. 101


CHAPTER


RATE


EQUATIONS


THEORETICAL


TREATMENT


. . . . . .. . 105


Two Level Atomic System ....
Three Level Atomic System . .
Fluorescence Dip Spectroscopy ......
Three Level Atom System ... .
Five Level Atom System .
Negative/Inverse Fluorescence Dip
Time Dependent Fluorescence Dips


* . . S .. 105
.* . *. . . 115
. .. .. .. 122
. . . . . 122
. .. . . 129


. . . . . 148


CHAPTER 6


EXPERIMENTAL


Instrumentation


. . . . . . . . 167
. . . . . . . . 167


Analyte Solutions . . . . . . . 174
Experimental Procedure . . . . . . . 175


CHAPTER 7


RESULTS AND DISCUSSION


S... ... 179


Experiments Performed
Theoretical Modeling


Excitation at X.3 and X- .
Excitation at X1.2 and X
Excitation at X..2 and X .-
Excitation at Xz. and X2-.4
Saturation Through Laser Excitation


* S . S S C . . 5 179
* S S S S S S S C C 205


. . .. . . .. 207
a a a a a a a S C S S S S S S 207

. . . . . . 212
* C S S . . . . 214
. .. .. .. . . . 215


CHAPTER 8


CONCLUSIONS


S. . 246


REFERENCE LIST


. .. . ..249


BIOGRAPHICAL SKETCH .


S.. .. . 250


LEAFS












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

FLUORESCENCE DIP SPECTROSCOPY
OF COPPER AND SILVER
AS A DIAGNOSTIC TOOL
IN SEVERAL ATOMIZATION RESERVOIRS

By

Donna Jean Robie


December 1993


Chairperson:


James D.


Winefordner


Major Department: Department of Chemistry


application


two-step


laser


excited


atomic


fluorescence


spectroscopy


(LEAFS) to the field of analytical chemistry has proven to be a method demonstrating

unprecedented sensitivity and selectivity as well as being quite useful as a diagnostic tool

in the examination of atom reservoirs and processes involved in the atom systems being


studied.


One variation on two-step


LEAFS is fluorescence dip spectroscopy which


involves the monitoring of the fluorescence signal,

an excited level populated through thermal means,;


from a directly excited level or from

as the result of one-step excitation and


two step excitation with the second step tuned to deplete the population of the monitored


level and


therefore


, the fluorescence signal.


This kind of study can be used to examine relationships between excited atomic







the determination of physical parameters for transitions involving the ground state.


work performed for this dissertation was designed to examine the collisional relationship

between two intermediate, excited state atomic levels and the effect this relationship has


on the fluorescence dip measurement.


Through one-step excitation,


fluorescence is


monitored from both the directly populated level and the collisionally coupled level.


second laser is then added while monitoring the fluorescence from the same level and


calculating the difference.


This decrease in signal is affected by several experimental


parameters including the electronic spacing between these intermediate levels as well as


the atom reservoir implemented in the experiment.


Different atoms were examined in


different


collisional


environments


results


compared


to theoretical


results


following a rate equations approximation, which is discussed and outlined in detail in this

dissertation.










CHAPTER 1
INTRODUCTION


Atomic fluorescence was first studied by Wood in 19021 and then in 1924 by


Nichols and Howes


,2 both of these studies dealing with possible diagnostic applications


technique.


1963


Alkemade3


suggested


application


resonance


fluorescence


flames as an analytical


technique.


Winefordner and


Vickers' then


explored the application possibility and in doing so obtained sensitivities on the order of


1 gg/ml for several elements causing them to comment that "


... it [atomic fluorescence


spectrometry] may prove superior to either atomic emission or atomic absorption flame


spectrometry.


.." (161) as an analytical tool.


Since these preliminary studies,


the field


of atomic fluorescence spectroscopy has quickly developed,


the main goal being to obtain


better detection limits, greater linear ranges, etc.


than previous studies.


In contrast, the


growth in


field of analytical fluorescence spectroscopy


as applied


to diagnostic


techniques,


resulting in physical information about both the atom being studied as well


as the atom cell of interest in the investigation, has fallen behind despite the fact that the


results


such


studies


are crucial


complete


understanding


analytical


spectroscopic techniques.

Many diagnostic studies have been performed in both the inductively coupled


plasma (ICP) and an air/acetylene flame,


the atom reservoirs of interest in the studies


presented in this dissertation, as well as other atom cells, yet it is without argument that







2

fact, improvement upon previous analytical studies requires the understanding of the atom

reservoir that is only possible through diagnostic studies to determine such parameters


as the spatial distribution and temperature profile of the atom source.


The availability


such information can facilitate the choices a chemist has to make as pertains to


experimental conditions in a subsequent experiment employing a similar atomization

environment.

In addition to information about the atom reservoir being employed in a study,

it is also necessary for chemists to know as much as possible about the atom they are

performing the studies on in order to select the best possible experimental conditions and


procedure to yield the desired results.


The need for fundamental atomic reference data


was the study of a questionnaire distributed by P.W.J.M. Boumans and A.


Scheeline.


Several well recognized analytical chemists responded to the questionnaires8,9,1' with

the consensus being that more reference data was needed about atomic systems since


...any information regarding the basic mechanism of excitation and deexcitation of


selectively


excited


states


be of


general


usefulness


to all


techniques."


Seliskar's response to the questionnaire,1


he mentions that while fundamental physical


information is incomplete for one-photon absorption studies, the same information for

two-photon and multi-photon studies is practically nonexistent.


With


advent


these


multi-photon


excitation


fluorescence


studies


subsequent


growth


field


since


results


above


questionnaire


were


published has resulted in substantial routine use of such techniques, the need for physical









one.


This is the void in fundamental atomic information that the work presented in this


dissertation will serve a small role in filling.


Two atoms (Ag and Cu),


approximated as


five level systems,


were studied using a two-photon excitation fluorescence technique


resulting in information about the collisional relationships between excited state levels.


These


studies


were


performed


three


different atomization reservoirs,


namely


air/acetylene


flame,


inductively


coupled


plasma


(ICP),


a reduced


pressure


inductively coupled plasma to determine the effect that the different environments would


have on the fluorescence measurements and


, therefore, on the relationship of the excited


state levels themselves.












CHAPTER 2
ATOMIC FLUORESCENCE SPECTROSCOPY


Conventional Source Atomic Fluorescence Spectroscovpy


History


Atomic fluorescence spectroscopy involves the absorption of a photon at a certain

wavelength by an atom resulting in an excited state atom and the subsequent de-excitation


of this atom.


This de-excitation results in the release of a photon of a certain wavelength


usually


equal


or lesser


energy


originally


absorbed


photon.


Both


wavelengths of the absorbed and emitted radiation are characteristic of the atom species

present.


The fluorescence of atomic vapors has been studied by


19th and 20th century


physicists.


These


investigations


have


been


discussed


detail


Mitchell


Zemansky'2 and Pringsheim13 and will not be addressed in this dissertation.


In 1905,


Wood14


observed


phenomenon


...non-luminous


sodium


vapor


radiating


brilliant yellow light when illuminated by the light from a very intense sodium flame."


(513).


Wood1S further studied sodium in a sealed glass tube with the excitation of the


sodium


D-doublet emission


of a


flame naming the resulting fluorescence resonance


radiation.


Atomic fluorescence was then explored by other spectroscopists in a similar









Howesr,'


suggested


implementation


a flame


as the


atom


in atomic


fluorescence experiments.


In doing so,


they were able to detect strontium,


lithium,


sodium


calcium


barium


and thallium.


The next substantial advance in the field of atomic fluorescence spectroscopy


came in


1956 with the suggestion from Boers et al." that these techniques could be


used to study the analyte atom,


not simply detect its presence.


This suggestion was


expanded upon by Alkemade3 at the Tenth International Spectroscopy Colloquium in 1963


where he described methods resulting in excited state atoms in flames.


At this time


Alkemade also suggested the use of atomic fluorescence spectroscopy as an analytical

tool.


In 1964


, Winefordner and Vickers4 investigated the possibility of the application


of atomic fluorescence flame spectroscopy as discussed by


metal


Alkemade.


vapor discharge tubes as excitation sources for mercury,


They utilized


, cadmium, and


thallium


acetylene/oxygen


flame


eventually


detecting


these


elements


concentrations


as low


gig/ml.


This


technique


further


explored"s26,n"28a


with the goal of these investigations being the development of a more sensitive technique


through the optimization of optics, instrumentation, and flame conditions.


In fact, these


studies resulted in detection limits that established atomic fluorescence spectroscopy as

a technique rivalling atomic emission and atomic absorption techniques in sensitivity.

Further improving upon atomic fluorescence spectroscopy was the suggestion of

the usefulness of a continuum source for the excitation of the analyte species in the atom









element.30


This


introduction


made


it possible


to evaluate atoms


not determinable


through previous excitation methods, and the field of atomic fluorescence spectroscopy

virtually exploded with these previously undetectable atoms being examined at the same

time as improvements were constantly being made on the experimental procedure to

obtain the best sensitivity possible.


Theory


Types of atomic fluorescence

Atomic fluorescence spectroscopy is a technique in which the emission collected


measured


direct


result


radiational,


rather


thermal,


excitation.


Fluorescence


techniques


have


been


defined


based


on their


different excitation


fluorescence schemes,3132,33


this discussion


of the types of atomic fluorescence


will be based on these previous designations.


There are two basic types of atomic


fluorescence


spectroscopy,


resonance


fluorescence


non-resonance


fluorescence.


Resonance fluorescence (Figure 2-1) involves the same two levels in the excitation


de-excitation


processes,


meaning


wavelength


absorbed


photon


responsible for the excitation of the atom species is identical to that of the emitted photon


involved in the atom's relaxation.


This type of fluorescence has been found to be the


most useful as pertains to analytical applications since the transition probabilities for


these


transitions


are usually


greater


those


other


transitions.


Resonance


fluorescence techniques do not necessarily involve the ground state in the excitation and








excited state


, absorbs a photon at a given wavelength, and relaxes back to its original


ground state fluorescing at that same wavelength,


fluorescence process,


this is still considered a resonance


but is designated as excited state resonance fluorescence (Figure


2-2) to distinguish it from conventional resonance fluorescence.


term


non-resonance


fluorescence


is applied


to all


transitions


where


wavelength of the photon absorbed is different than that of the emitted photon with the


energy of the photon involved in the fluorescence process usually, but not always,


lower than the energy corresponding to the excitation process.


being


This type of fluorescence


can be


divided


categories


direct-line


fluorescence


stepwise


fluorescence.

Direct-line fluorescence results when the excitation and de-excitation processes


have a common upper level.


Again,


this process does not necessarily involve the ground


state in either the absorption or emission processes,


Stokes and anti-Stokes categories.


and has, therefore been divided into


Stokes processes involve the emission of a wavelength


at a lower energy than the absorbed photon (Figure 2-3) while anti-Stokes fluorescence

pertains to a process where the emitted photon is of a wavelength corresponding to an


energy greater than that used for the excitation process (Figure 2-4).


If these processes


occur between two excited states, the terms excited state Stokes and excited state anti-


Stokes are applied (Figures


& 2-6).


Stepwise-line fluorescence is said to occur when the upper levels involved in the


excitation and emission steps are different.


This can occur in several variations involving








thermally-assisted


excitation


a higher


excited


level


than


reached


through


absorption as a result of the excitation step prior to the fluorescence takes place.


many


types


stepwise-line


fluorescence


include


Stokes


stepwise-line


fluorescence


(Figure 2-7),


excited state Stokes stepwise-line fluorescence (Figure 2-8),


anti-Stokes


stepwise-line


fluorescence


(Figure


2-9),


excited


state


anti-Stokes


stepwise-line


fluorescence


(Figure


thermally


assisted


Stokes


anti-Stokes


stepwise-line


fluorescence (Figure 2-11),


and excited state thermally assisted Stokes and anti-Stokes


stepwise line fluorescence (Figure 2-12).


In addition


to the types of fluorescence


techniques just discussed,


sensitized


fluorescence can also occur in analytical studied

an excited state atom or molecule, the "donor"


This type of fluorescence occurs when


, transfers its excitation energy to another


atom species present, the "acceptor"


, in the atom cell through collisions.


The acceptor


then undergoes radiative de-excitation resulting in atomic fluorescence (Figure


This type of fluorescence requires a high population of "donor" atoms,


however, and this


is not normally the case especially in flame atom cells where energy transfer fluorescence


mechanisms are of virtually no use because of the collisional


deactivation


of these


"donors" resulting in an inadequate concentration to fulfill its intended purpose.

Another type of fluorescence is multi-photon excitation fluorescence in which two

photons of the same wavelength are absorbed by the analyte atom through a virtual level


(Figure


This type of fluorescence is not of much analytical use because the


improbability


the transitions involved


makes it an


unlikely choice


for analytical






































Figure 2-1 : Resonance fluorescence (either process)






















citation


Fluorescence


Excitation


2






1



Fluorescence


0






































Figure


: Excited state resonance fluorescence



















2



ce


--1







-o






































Figure 2-3


Stokes direct line fluorescence






14











S2



Fluorescence



1


XExcitati




I___________0 _







































Figure 2-4


Anti-Stokes direct line fluorecence






16















Excitation AFuorescence










__0






































Figure


Excited state Stokes direct line fluorescence


















Xluoe


citation






































Figure 2-6


Excited state anti-Stokes direct line fluorescence






















Excitation


Fluorcence






































Figure


Stokes stepwise line fluorescence



























Excitation


2












--1



uorescence

0


__


__





































Figure 2-8


Excited state Stokes stepwise line fluorescence






24












3




2


I~citation fluorescence



- --. _- 1




--____0





































Figure 2-9


Anti-Stokes stepwise line fluorescenece






26











3

2

^ IL


Excitation___ ____ 1
A1
Fluorescence





_____________ o





































Figure 2-10:


Excited state anti-Stokes stepwise line fluorescence
















4

3



citation 2

Fluorescence


1
0






































Figure 2-11: Thermally assisted Stokes or anti-Stokes fluorescence
(Depending upon whether the fluorescence occurs at a longer or
shorter wavelength, respectively, than the absorbed radiation)





30










3







II A
I Fluorescence
2




Excitation
1



0





































Figure 2-12 : Excited state thermally assisted Stokes or anti-Stokes fluorescence
(Depending upon whether the fluorescence occurs at a longer or
shorter wavelength, respectively, than the absorbed radiation)






























Fluorescence


citation


_ _~ _____





































Figure 2-13 : Sensitized fluorescence (D = donor; D'
A = acceptor; A' = excited acceptor; hvp
and hvF = fluorescence radiation)


= excited state donor;
= excitation radiation;

















D+hvE


A*


+ hv


D


D*


+D






































Figure 2-14


Multi-photon excitation fluorescence





F





































Figure 2-15


Two-color atomic fluorescence spectroscopy





















Excitation #2


Excitation #1


Fluorescence





1




0








specifically,


two-color


excitation


techniques


where


photons


responsible


excitation of the analyte atom are of different wavelengths corresponding to transitions


between


levels


(Figure


2-15),


resulting


name


two-color


excitation


fluorescence.


These


transitions


have


much


greater


probabilities


can be


combination


of the aforementioned


resonance and


nonresonance


techniques.


These


techniques were not implemented or recognized for their analytical potential until the


application


of laser


excitation


to atomic


fluorescence


spectroscopy,


which


discussed later.

Theoretical treatment of atomic fluorescence soectroscoDv


Winefordner


Vickers4


have


outlined


theoretical


basis


atomic


fluorescence as an analytical technique, and their development will be closely followed

in this discussion.

The intensity of a resonance fluorescence signal per unit time is proportional to

the intensity of the absorbed excitation radiation per unit time through the expression


(2-1)


P,=Pabs


where


Pp is the fluorescence power, P,, represents the amount of incident radiation


absorbed,


is a


proportionality


constant


representing


number


atoms


undergoing the


fluorescence process per unit time divided


by the number of atoms


leaving the ground state as a result of the incident radiation, involved in this absorption

process.

The relationship between the incident radiation and auantitv of this vower that is









Pab,


-e-k L) A


(2-2)


where P,O is the spectral intensity of the incident radiation (W/sec),


absorption of center of the absorption line (cmr'),


ko is the atomic


L is the average absorption path length


(cm),


and Av is the half-width of the absorption profile (sec').


is approximated,


in this case, as a triangle,


This absorption profile


which is a good estimate since the absorption


to follow


a Gaussian


distribution


flame


spectroscopy.


With


approximation,


this half-width can be expressed as


Av=


2(1n2)1/2


G(2-3)


where Avo is the half intensity spectral linewidth of the Gaussian curve (sec').

constant ko is defined as


(In2)1


(2-4)


An v


where AvD is the Doppler half-width of the absorption line (sec"),


statistical weights of the lower and higher states,


being studied,


go and g, represent the


respectively, involved in the transition


no is the the ground state population density (cm3) of the analyte atom


species,


X is the wavelength of the center of the absorbtion line (cm),


A, is the Einstein


coefficient of spontaneous emission from level 1 to level 0 (sec'"),


and 6 is defined as (In


times the ratio of the sum of the Lorentz, Holtsmark, and natural half-widths to the


Doppler half-width (sec')12


2gn
-~no e6


A








Combining equations (2-1) and (2-2), and adding a self-absorption factor,


(koL/2),


determined


Kolb


Streed,"3


following


expression


approximates the energy emitted per unit time as fluorescent radiation:


PF=PPAv (l-e


-koL/2)


-koL/2cosh(koL/2)


(2-5)


This equation can be rewritten in terms of fluorescence intensity, Ip (W/cm2* steradian),


by dividing Pp by the area,


of the fluorescent cell from


which the fluorescence


energy is emitted, and by 4wr steradians resulting in


PoAv


(1-e


-koL)


-koL/2cosh (koL


(2-6)


mentioned


beginning


theoretical


treatment,


resonance


fluorescence is assumed, and for this case equation 2-6 is an exact expression upon which

expansion is required in order to compensate for other types of fluorescence, or a case

where several absorption lines are responsible for the fluorescence intensity.


intensity


of the fluorescence is proportional to


the number of atoms


absorbing


energy.


As n


increases,


(1-ek-) approaches


unity.


Therefore,


Ip goes


through a maximum as n increases.


For small values of n, i.e. low concentrations, the


fluorescence intensity can be expressed as


*4PAv ko L


(2-7)


Substituting for ko, defined earlier, the following expression is obtained:


e -tLcosh


4x A,









(1n2) 1/24 vLg129A6
16 5/2AA v g2


PoNo


(2-8)


This equation for fluorescence intensity is an exact expression for any experimental


arrangement and any spectral line.


This relationship then can be simplified to


I,=CPONo ( W/ cm2


s ter)


(2-9)


which indicates that a linear relationship exists between the intensity of the fluorescence


signal and a proportionality constant, C,


of the analyte species,


providing the theoretical


basis for atomic fluorescence spectroscopy.


ADplicatons of Conventional Source AFS


During the introduction of atomic fluorescence spectroscopy and its early growth,


there was very little use for this technique for the analysis of "real" samples.


There were


already other established methods involving either atomic emission or atomic absorption

spectroscopy that could be used, and the technique of atomic fluorescence spectroscopy


was confined to mainly fundamental research.


method


The sensitivity and selectivity of this


, however, quickly demonstrated its usefullness, and AFS began to be applied to


samples previously only studied by other spectroscopic techniques.

Atomic fluorescence spectroscopy has been applied to biological samples such as


blood35'36y'


urine38'39


determination


mercury,


gold,


zinc,


cadmium,


and other metallic elements of biological interest detectable in quantities as low as 5


ne/ml.


The main tonics of discussion surrounding these arolications of AFS were in








would


result in


the lowest limit of


detection.


This was


the same point of interest


discussed

petroleum


pertaining

products


to the

and


early analytical

fuels.42'43 and


applications


environmental


AFS


to metallurgy,4041


analyses"4445


trace


determination of metals.


Advantages and Disadvantages


Conventional source atomic fluorescence spectroscopy has been demonstrated to

achieve lower detection limits and larger linear dynamic ranges than conventional atomic


absorption or atomic emission techniques for the same atom species.


Since these are the


main goal


to be achieved in an analytical technique,


it would seem beneficial


examine


atomic


fluorescence


spectroscopy


instrumentation


not achieved


appreciable commercial attention.

It is true that atomic fluorescence techniques enjoy unprecedented sensitivity and


selectivity


along


linear


dynamic


range


usually


seven


orders


Add to these advantages that the relationship between the incident radiation


intensity and the fluorescence signal is a linear one at low concentrations,


the case in atomic absorption spectroscopy,

especially for trace analysis. Atomic fluid

simultaneous multielement analysis, suffers


which is not


and AFS would appear to be a powerful tool,


orescence spectroscopy has the potential for

from minimal chemical interference because


of the excellent selectivity of the technique, and the instrumentation involved is rather


simple and relatively inexpensive.


Despite all of these benefits of atomic fluorescence


magnitude.








This


is mainly


to the


many


disadvantages


conventional


source


AFS


techniques including self-absorption at higher concentrations resulting in a non-linear


response between incident intensity and fluorescence intensity.


These methods also have


high background signals due to scatter of the excitation radiation that is difficult to


correct for,


decreasing


the sensitivity of the measurements.


Matrix problems have


prevented the application of AFS to real samples, and sample preparation can result in


blank matrix matching problems.


The quantum efficiency of the fluorescence signal,


well as other physical parameters, is dependent upon several atomization source variables

which must be monitored closely to maintain reproducible results.


main


reasons


AFS


instrumentation


manufactured


commercially


optical


emission


mass


spectrometric


techniques


result


comparable detection limits are a multi-element techniques virtually free from matrix


effects.


These instruments are much more costly than a commercial AFS instrument,


, nonetheless,


chemists are using


either mass


spectrometry


or established atomic


absorption or atomic emission methods for their trace analysis.


This does not take away


from the need for information supplied through diagnostic applications of AFS.


Laser Excited Atomic Fluorescence Spectroscopy (LEAFS)


One major area of research in the field of AFS has been the study of a variety of

radiation sources meeting the criteria for an ideal excitation source including stability,


long life,


low cost,


high versatility,


, most importantly,


high intensity since the






45

studies have resulted in the use of excitation sources such as the high-intensity hollow-


cathode lamp,


demountable hollow-cathode lamps,


and spectral vapor discharge lamps.47


While these sources did fulfill most of the aforementioned criteria, they lacked versatility


in that only a limited number of elements could be studied using them.

sources including high-pressure xenon arcs, overcame this disadvantage, 1


Continuum


but added the


disadvantage of increased


background


noise and lack


of intensity,


especially


ultraviolet region.


It became obvious that in order for the vast potential of atomic


fluorescence spectroscopy techniques to be realized, some excitation source had to be

found which satisfied most of the above criteria.


History of Flame LEAFS


1971


, Denton and Malmstadt48 and Fraser and Winefordner9 independently


reported using a tunable dye laser as an excitation source for the analytical study of


atomic species in a flame.


These are the first documented analytical applications of


lasers to the field of atomic fluorescence spectroscopy

for the use of dye lasers in AFS was demonstrated.


and in these studies the potential

Denton and Malmstadt4A used a


frequency-doubled ruby laser to pump a dye laser and observed fairly sensitive detection


of barium atoms in a flame.


They also observed a linear response over three orders of


magnitude once the high scatter signal resulting from the laser radiation being reflected

off of water droplets present in the flame was compensated for.

Fraser and Winefordner9 used a nitrogen pumped dye laser system continuously








output of


laser


system


used in


study


would


be considered


low by today's


standards for laser systems, the authors were able to show that if a stable, high powered,

pulsed light source with a narrow spectral bandwidth and small duty cycle was used, the


signal to noise ratio could be greatly improved over conventional source AFS.


limiting noises in AFS were usually random noises including flame background and dark

current, and any reduction in the "on time" of the detector would reduce the contribution


of these noises to the total noise considerably.


Therefore, the short pulses afforded by


a dye laser system made them a nearly ideal source for AFS.


Through their studies, the


authors observed a decrease in the flame background noise coupled with an increase in


total noise due to laser scatter from the atom source.


This had also been a problem for


Denton and Malmstadt48 simply because increasing the intensity of the excitation source

through the implementation of a dye laser resulted in an increase in scatter as a result of


that source.


The sensitivities reported in this study for their determination of aluminum,


calcium


, chromium, iron,


gallium,


indium


, manganese,


strontium, and titanium in either


a hydrogen/air or an acetylene/air flame compared favorably to those obtained through


the use of conventional source AFS


, but were not superior.


Also in this article,


authors discussed the limitations of this technique based on the wavelength range not


encompassing transitions below 360 nm, the range where many ground state,


transitions are found.


resonance


It was suggested at this time that this limitation could be overcome


through the use of doubling crystals resulting in laser radiation in the ultraviolet region.


the early


1970'


Kuhl


, Marowsky,


Torge0,s"


reported


use of a








sodium in an absorption cell.


This report was significant because it represented the first


mention of the use of a dye laser whose output was narrowed to a linewidth,


reported at


5x10" nm, that was comparable to true elemental line source linewidths resulting in an


absolute detection limit of


According to the authors, the main advantage of using


laser


as the


excitation


source


its tunability


which


allowed


background


interference to be corrected for through small detunings of the laser wavelength.

Almost from the beginning of the applications of lasers as excitation sources in


AFS it was observed that


, at high intensities of laser radiation,


there was a nonlinear


relationship between the laser power and the resulting fluorescence signal.


In 1972,


Piepmeier


discussed


advantages


phemonenon,


known


saturation


fluorescence, could have in the field of AFS.


As mentioned earlier, one of the main


advantages of conventional source AFS over atomic absorption spectroscopy (AAS) was

the linear relationship between the source intensity and the monitored fluorescence signal


resulting in


AFS


demonstrating improved


sensitivity


over


AAS.


With


pulsed laser


excitation


, the peak source


power was


now


enough


to saturate


the absorption


resulting in a now nonlinear relationship between source intensity and fluorescence signal


observed.


According to Piepmeier, laser excitation resulted in an optically saturated


dilute population in a flame causing


fluorescence several


orders of magnitude more


intense


population


excited


a conventional


source at lower power.


This


saturated population demonstrated a much lower dependence on source variations than


an unsaturated population.


The author used the rate equations approach to approximate








the saturated atomic population,


which assumed steady state conditions during the laser


pulse.


main


conclusions


of his"


theoretical


treatment were


excitation


spectral irradiances beyond the irradiance required to reach saturation conditions,


atom population responsible for the fluorescence signal was only weakly dependent upon


factors that had a much greater effect when saturation was not achieved,


to pulse variations and quenching within the atomization medium.


such as pulse


The result of this was


a fluorescence signal which reached a maximum at high laser irradiances and suffered


little fluctuation with variations in the source.


While the fluorescence signal remained


constant for laser irradiances greater than the saturation irradiance, laser scatter in the


atom


continued


to increase


linearly


with


increasing


laser power resulting in


decreased fluorescence to scatter ratio at high laser powers.


It was suggested that there


was an optimum laser power resulting in saturation conditions while minimizing the


signal due to scatter.


This theoretical treatment will be discussed in more detail in


another section of this dissertation.


Also in 1972


, Fraser and Winefordner53 reported results for laser excited atomic


fluorescence


spectroscopy


(LEAFS)


a flame


atom


These


results


included


detection limits for thirteen elements and some general conclusions based on the results


of their experimentation.


For resonance fluorescence measurements


, the signal to noise


ratio was high due to the laser scatter in the atom source as well as the shot noise of the


detector


pulse


variations.


This


noise


was


scatter


limited


at low


analyte








required


a decrease in


scatter noise.


This


brought the authors


to suggest and


implement the first nonresonance laser excited atomic fluorescence scheme.


The result


was a reduced scatter background since the wavelength of the laser radiation and the

wavelength being monitored were different, and detection limits as low as 2x10(3 pg/ml


were reported.


This publication also reported the first use of multiphoton excited AFS


for the detection of cadmium and zinc.


As mentioned earlier,


two photon excitation


techniques are only possible for analytical applications through the implementation of

lasers as excitation sources.

In 1973, Omenetto et al.54S reported the analysis of several transition elements


using LEAFS in a nitrous oxide/acetylene flame.


The significance of these studies was


the fact that these elements had never before been analyzed using AFS because of the


lack of suitability or availability of line sources. This again demonstrated the usefulness

of the application of dye lasers to the field of AFS. Also in these studies, nonresonance


fluorescence


was


monitored


again


demonstrating


decrease


in scatter


noise and


subsequent improved detection linits.

Also in 1973, Omenetto et al.56 published a paper similar to the theoretical paper

by Piepmeier52 on saturated AFS with the difference between the theoretical treatments


being


the assumption


a line


(monochromatic)


source


Piepmeier and a quasi-


continuum source by Omenetto.


At the time of this publication, the quasi-continuum


source assumption was a better one for the application to dye lasers with bandwidths on


the order of 0.1 to 1 nm,


which is much greater than most atomic linewidths.







50

The authors also presented experimental results that demonstrated the effects of


optical saturation as a result of laser excitation on the observed AFS signals.


This data


caused them to conclude that, as expected, for low excitation source irradiances,


fluorescence signal enjoyed a linear relationship with the source intensity, but that at high

source irradiances, this fluorescence signal would become increasingly independent upon


the source intensity and eventually reach a maximum value.


This maximum fluorescence


signal was shown to be a property of the atom system itself and was described using the

theoretical treatment presented in this paper.


Also


the article


Omenetto et


' the authors


derived


an expression


describing the relationship between

resulting fluorescence radiance, assi


the source flux for a continuum source and the


jming a two level system at steady state conditions.


This relationship is valid at both high and low source intensities.


The authors discussed


the possibility of mapping the population of the atom source using saturated LEAFS since

such measurements would be independent of the quantum efficiency (i.e. independent of


collision processes).


This would be important analytically because of the high quenching


environment in most flame reservoirs.

This paper confirmed the conclusions of Pipemeier'2 that saturated fluorescence


signals are independent of source fluctuations.


It also confirmed that at laser irradiances


greater than the saturation irradiance, the fluorescence signal would eventually reach a


maximum


signal


while


scatter noise would


continue


increase.


The authors


suggested the best way to circumvent this disadvantage was to optimize the fluorescence








signal to noise


ratio rather than optimizing conditions resulting in simply the greatest


signal.


1978


, in an article by


Boutilier


et al.


steady state atomic


fluorescence


radiance expressions were given for two and three level atomic systems excited by a


continuum source for both saturation and non-saturation conditions.


Two types of three


level atomic systems were considered in their treatment for the fluorescence radiance


expressions that were derived.


potassium,


The first case was for alkali-like atoms (i.e. sodium,


etc.) where the two excited state levels were assumed to be very close in


energy,


with radiative processes between these two levels being forbidden.


The second


case considered elements where all three levels were assumed to be well-separated in


energy,


with radiative transitions being allowed between all levels except the two lowest


levels


the intermediate level


is a metastable


level).


This theoretical


treatment


included derivations of all possible radiative and nonradiative transition combinations

resulting in several conclusions to be made by the authors about the fluorescence radiance


expressions.


First,


for dilute atom


vapors


(low optical


densities),


fluorescence


radiance is linearly related to the total atom population for all source intensities.


level atomic systems,


For two


the fluorescence signal resulting from excitation from a source


under saturation conditions is independent of collisional deactivation processes.


atom not accurately depicted as a two level system,


For any


however, some knowledge of these


nonradiational processes would be required to gain information about the relationship

between the fluorescence radiance and the atom population.








Also in 1978


, Weeks, Haraguchi, and Winefordneri8 published the results of a


comprehensive study of 24 elements using flame LEAFS.


For comparison purposes,


several


of the elements


nonresonance


excitation


were investigated,


fluorescence


implementing


schemes.


both resonance as


authors


well as


determined


resonance


fluorescence


was


limited


laser


scatter


noise


under


otherwise


ideal


conditions


while nonresonance fluorescence schemes and the freedom from laser scatter


interference enjoyed by these techniques were limited by either noises due to the dark


current,


amplifier,


flame


background


emission


molecular


background


fluorescence, or any combination of these noise sources.

This study resulted in detection limits demonstrating improvement over previous

studies by up to two orders of magnitude, yet the authors still felt that there was much


room for improvement of flame LEAFS as an analytical technique.


Many of these


suggested improvements revolved around the need for a higher output power dye laser


since the laser used in this study was sometimes close to,


but not always reaching,


saturation spectral energy density for many of the ultraviolet transitions.


This limitation


prevented the authors from being able to take full advantage of saturated conditions in


these schemes.


The use of an alternative atomizer was also suggested to reduce noise


and improve the sensitivity of this method.


History of Furnace LEAFS


While flame LEAFS


was enjoying some success,


the search


was still on for









results obtained in a flame cell.


As atomizers,


flames suffer from


many problems


preventing the analytical application of LEAFS from achieving its full potential in terms


sensitivity


selectivity.


Flames


are susceptible


to chemical


ionization


interference as well as low quantum efficiencies of fluorescence in most flame gas


mixtures.


result


these


limitations,


field


of LEAFS


was


redirected


applications in graphite furnaces, ICPs,


and other atomization reservoirs.


The first mention of LEAFS employing a graphite furnace as an atom source was

back in 1974 when Neumann and Kriese59 reported the use of a flashlamp pumped dye


laser system employed in


the analysis of lead.


A comparison of the laser to other


excitation sources, such as the hollow cathode lamp and the electrodeless discharge lamp,


was made at this time.


The graphite furnace was proven superior to flame atomizers for


several


reasons.


furnace


an exceptionally


high


atom


density


during


atomization cycle and the analysis is performed in an inert environment helping to reduce


interference.


The authors reported that LEAFS in a graphite furnace provided superior


detection


limits as


as a


greater


linear


range


response


previously


mentioned conventional source AFS technique.


Since this first report of graphite furnace LEAFS,


this technique has been used


to determine many elements at very low levels.


In fact


, the use of graphite furnace


LEAFS


has resulted in


the lowest detection limits of any spectrochemical analytical


technique to date.S3,60,61'62


Graphite


furnace


LEAFS


has been demonstrated


to be a


technique with selectivity and sensitivity coupled with the ability to handle small samples








promising,


could


an eventual


achievement


intrinsic


limit


detection.63

While this technique would seem ideal for continued application to the field of


LEAFS


it is not without its limitations.


These include the limited number of elements


that can be analyzed through this technique as well as the difficulty with interference


despite the inert atmosphere in which the measurements are taken.


Successful application


of this technique to real samples would require vacuum atomization which entails a

considerable increase in experimental complexity.


History of ICP-LEAFS


The ICP was first implemented as an atomizer for AFS by Montaser and Fassel"


1976.


The excitation was provided by EDLs and Osram lamps which the authors


considered the best existing excitation source since they were the most intense sources


for transitions in the ultraviolet range.


Detection limits were reported for cadmium,


zinc,


and mercury comparable to previously reported detection limits for atomic emission ICP


studies.


In addition to the good detection limits, this method demonstrated excellent


selectivity with no adverse effects on detection limits for complex mixtures of up to 17


concomitant elements.


Since this initial ICP-AFS study, conventional source AFS has


also been performed very successfully using hollow cathode lamp (HCL) excitation."

HCL excitation is effective for those elements that provide sufficient radiant intensity of


a resonance transition in a hollow cathode discharge.


Because of this limitation, there







55

The first report of ICP-LEAFS appeared when Pollard et al.6 discussed the first


application of dye laser excitation to fluorescence excitation in the ICP


. Although the


results were not very favorable using a continuous wave (CW) laser, the authors reported

excellent linearity and suggested that pulsed laser excitation could result in improved


limits of detection of this method.


This technique was also proposed as a useful method


to perform diagnostic studies in the ICP.

Quickly following this study in 1980 was a report by Epstein et al.67 using both

a flashlamp pumped dye laser and a nitrogen pumped dye laser as excitation sources for


ICP-LEAFS.


The detection limits of this method were not an improvement over ICP


emission techniques or ICP-AAS.


Nonetheless, the authors were optimistic about the


future of this technique stating that these were simply preliminary studies,


was much room for improvement.


and that there


It was again mentioned at this time that ICP-LEAFS


could be a very powerful tool in the field of plasma diagnostics.


In 1984


, Omenetto et al.8 published an article discussing the analytical potential


of ICP-LEAFS beginning with the admission that the results presented to that point were


disappointing as compared to emission ICP methods.


In their study, they were unable


to reproduce the poor and erratic detection limits previously reported,


, in fact, their


experimentation resulted in LODs superior to other fluorescence results and,


cases, to emission methods as well.


in most


Both resonance and nonresonance schemes were


investigated


resulting


better


signal


to noise


to decreased


scatter


nonresonance cases,


as expected.


The authors also discussed


the ease


with


which








this atom cell as compared to a flame.


The forward power was also mentioned to be


about 1kW to increase the atomic population available for probing by the laser radiation.


In 1985


, Huang,


Lanauze


, and Winefordner69 utilized ICP-LEAFS to study some


precious metals and refractory elements resulting in detection limits in the range of 1.3


to 58 ng/ml and linear ranges of over four orders of magnitude in most cases.


These


limits of detection were found to be superior or comparable to those obtained through the


use of flame AAS and similar to those obtained by ICP emission.


Also investigated in


this study was the effect of rf power on the fluorescence signal with the intensity of the


atomic line monotonically decreasing with increased rf power.


the greater signal,


The lower the rf power,


with the plasma becoming unstable at powers less than 600 W


History of LEAFS in Other Atomization Reservoirs


While the majority of previous LEAFS studies have been performed in a flame

cell and the most recently, and probably future studies, have been carried out in an ICP,

other atomization reservoirs have been implemented in the field of laser excited atomic


fluorescence spectroscopy.


These include the glow discharge,


an atom source primarily


utilized in emission studies for trace element analysis in metal samples.


This technique


has the advantage of solid sampling with little sample preparation required.70


analytical potential of GD-LEAFS has not been recognized,


The full


but the studies that have been


performed7071 reported detection limits on the order of 10 ng in aqueous solution and

8 gg/g in a solid for indium and 20 pg in aqueous solution and 0.1 psg/g in a solid for









In 1983


Kosinski


, Uchida, and Winefordner


reported the use of a modified


ICP torch for ICP-LEAFS resulting in detection limits for


those obtained with HCL excitation.


12 elements comparable to


The comparison of this method with ICP-OES,


however, showed the LEAFS technique lacking in sensitivity.


History of Two Photon LEAFS


As mentioned earlier, two-photon excitation fluorescence spectroscopy was carried


out in a flame atom cell for the analysis of cadmium and zinc.53


This study entailed the


absorption of two photons from the laser excitation radiation at the same wavelength

corresponding to an energy of half that of the resonance wavelength through a virtual


level.


As was also previously mentioned, this technique has no analytical usefulness with


any other excitation source other than a laser,


because of the low probabilities of the


transitions


involving the virtual level.


Even


with


laser


excitation


, it is not always


possible to obtain saturation conditions in these cases.


In 1978


, two-photon excitation was implemented in a study of the excited states


of sodium7 to investigate the collisional coupling between higher excited levels.


authors were able to attain these higher states through the absorption by the sodium atom


of two excitation photons of equal wavelength through a virtual level.


The wavelengths


were chosen based on the desired scheme to be studied


either 3S


-* 3D, 3S -' 4D, or


--5S.


Using a flashlamp pumped dye laser, the authors reported achieving saturation


conditions providing an estimate of the value of the saturation parameter which can then







58

Miziolek and Willis74 then used two photon excitation for the analytical analysis

of lead using two different wavelengths and exciting through a real excited state level to


populate the higher lying excited state.


The authors called this process double-resonance


emission spectroscopy, and utilized this method to reduce scatter noise by monitoring the

fluorescence signal at a wavelength originating from the uppermost excited level at a

different wavelength than either of the excitation steps.

Another study employing two-color excitation for analytical determination was


reported


Rogers


et al.


for the determination of mercury with


graphite furnace


atomization. The authors reported detection limits on the order of 10' atoms/cm3.

In 1988, Leong et al.76 used what they called double resonance spectroscopy for


the analytical determination of lead in a graphite furnace.


The fluorescence signal was


monitored as the result of excitation at both 283.306 nm and 600.193 nm from a Nd-


YAG dual dye laser system from both of the excited state levels taking part in


excitation steps.

being monitored,


With the direct line fluorescence signal as a result of two step excitation

the authors reported detection limits in the picogram range and direct


line fluorescence as the result of one step excitation resulted in an LOD of 3 fg.


addition


analytical


applications,


two-photon


excitation


fluorescence


spectroscopy has proven to be a very useful diagnostic tool.


These types of techniques


have been used to a great extent in the evaluation of combustion species in flames such


as oxygen,


7 atomic nitrogen,78 and hydrogen.79.s80,18s2









History of Two-Color Excitation Fluorescence


Two photon excitation schemes involving a real level as an intermediate excited

state would facilitate saturation conditions as a result of greater transition probabilities

involved in the two excitation steps, but would then require two different wavelengths


of laser radiation for the two respective excitation steps.


This is typically carried out


utilizing a dual dye laser system with both lasers pumped by the same laser providing the

advantage of tunability as well as intensity in order to study the transitions involving the

high lying excited states.


There


are many


possible


variations


on the


theme


two-color


excitation


spectroscopy including monitoring the fluorescence from the intermediate excited state


level instead of the highest excited state.


These experiments usually entail the monitoring


of a fluorescence signal as the result of one step excitation and the subsequent addition


a second


excitation


tuned


to a


transition allowing


it to directly perturb


population of the monitored level.


In most cases,


the fluorescence signal from the


intermediate


level


decrease


upon


introduction


second


excitation


responsible for depleting the population of the monitored level.


This technique has taken


on many


names


in the


literature


but the


methodology


techniques


essentially the same.

These types of techniques are based on the monitoring of an excited level as the

population of that level is being depleted through the introduction of a second excitation


T S


S4 Sl 4 S 4 4 ** f -


a[






60

excitation radiation and this decrease in signal has caused these techniques to more

generally be termed fluorescence dip spectroscopy.

Omenetto et al.'3 used fluorescence dip spectroscopy for the determination of


calcium


, barium, strontium,


and magnesium.


The authors reported this to be a very


sensitive


technique


demonstrating


unprecedented


selectivity.


excitation


fluorescence schemes, the signal was monitored at a wavelength different than that of


either excitation wavelengths thereby eliminating scatter noise.


Also in this article, the


authors outlined a theoretical approach for modeling such excitation and fluorescence


schemes based on the rate equations approach.


A similar model will be presented later


in this dissertation.


Advantages and Disadvantagees of LEAFS


While laser excited atomic flame spectroscopy has been shown to be a technique


with excellent sensitivity,


superior to atomic emission or atomic absorption methods, and,


with


saturated


step capabilities,


unprecedented


selectivity


perhaps


approaching


elemental specificity, it is not without its limitations.

systems is the cost involved in the instrumentation.


The main disadvantage of these

In order to take advantage of the


tunability of a dye laser, there needs to be some means of pumping this laser, and while

that was earlier performed using a flashlamp, it is now almost exclusively performed by


another laser.


Add to that the two color excitaton schemes improving the sensitivity and


selectivity,


and another laser is required.


With all other optics and atomization sources







61

industrial lab can afford to spend on a fluorescence system, which is why there are many

ES and AAS instruments yet only a handful of fluorescence instruments commercially


available


, and the ones that exist do employ HCL excitation, even though laser excitation


has many advantages over HCL excitation, simply because of the cost involved.


With the exception of thi


large disadvantage,


LEAFS has many advantages over


OES


AAS


conventional


source


AFS


techniques,


which


have


been


outlined


throughout the previous discussion.


For most routine analysis,


however, the existing,


less expensive methods are sufficient,


industry.


and LEAFS has not made it to the forefront in


This does not detract from the need for fundamental physical information about


atom systems as well as temperature profiling and other physical information about the

atomization reservoirs that are used in conjunction with these other detection methods


everyday.


The best way to carry out the studies resulting in this type of information is


through the use of LEAFS.












CHAPTER 3
DIAGNOSTIC APPLICATIONS OF LEAFS


Diagnostic Characterization of Atom Sources


It is difficult to distinguish between the diagnostic and analytical applications of


LEAFS since they complement each other so well.


It is undisputed that a knowledge of


the underlying physical principles of analytical as well as physical information about the


analyte species are necessary to improve upon existing analytical techniques.


Initial


studies regarding the state of atom sources typically involved the insertion of some sort

of probe into the flame, plasma, or other atom reservoir, and sampling the condition of


probe


over


a small


volume."


many


disadvantages


involved


with


such


measurements


resulting


include the effect that the probe itself has on


effect this would have on


the atom


the measurement being taken.


cell and


It was therefore


necessary to find a method to determine physical information about these atom sources,


as well


as the


species


being


investigated


them,


without


disturbing


source.


Spectroscopic methods provide the best means for obtaining spatially and temporally


resolved


measurements


several


physical


parameters


these atom


cells


without


perturbing the processes occurring in them.


First,


emission


spectroscopic


techniques


were


used


investigate


horizontal85'"


vertical7"*8


spatial


distribution


excited


atoms.


Also explored









were


determination


various


kinds


temperature,89,90


electron


number


densities.9'92


Atomic absorption spectroscopy was implemented in the investigation


of number densities of analyte atoms using hollow cathode lamps" as well as other


excitation


sources.


Argon


number


densities


were


also examined


using


AAS


microwave induced argon plasma.


These


spectroscopic


methods provide


"line-of-


sight" measurements, however, not allowing for temporal or spatial resolution.


In order


to gain spatial information, the desired quantity must be derived by somewhat elaborate

procedures, usually the Abel inversion method."9

Omenetto, Benetti, and Rossi" reported using atomic fluorescence spectroscopy


measurement


flame


temperatures.


authors


proposed


different


fluorescence schemes and compared the resulting temperature measurements to each other

as well as with those determined using the other spectroscopic methods mentioned above.

They found that the ability to investigate a smaller volume than with other spectroscopic

methods allowed for more reliable temperature profiling resulting in more consistent

values.


Subsequent applications of


AFS


to atom


source diagnostics


have


resulted


physical


information about these sources


such as


flow velocities,


electron and ion


densities.97"98


flame


temperatures


or plasma


within


species." 9,1


atom


AFS


cell,54


been


applied


collisional


cross


sections


to measurement


some


constants of atomic studies including damping constants of atoms in flames,10' lifetime

of excited atoms,12 atomic cross sections,3 and diffusion studies.






64

AFS found many uses as a diagnostic tool in a variety of atomization reservoirs


with the eventual application of lasers as excitation sources to these techniques.


Laser


excitation combines the advantage of small volume probing with increased intensity over


conventional sources thereby maximizing the fluorescence signal.


In addition, lasers


more often cause saturation conditions to occur simplifying the investigations of atom

species in these atom cells.


In 1968


, Measures" introduced what he called selective excitation spectroscopy


which


was


, basically,


atomic fluorescence using a pulsed laser as excitation.


This


method was proposed and utilized as a means for determining the local values of electron


temperature and electron density in low-temperature plasmas.


that because a pulsed laser was implemented,


The author commented


both spatial and temporal resolution in


these measurements were possible, demonstrating the versatility of this technique.

While it would seem that only so much information can be gained through such


studies


, resulting in the stagnation of this field,


these properties are very dependent upon


atomizer conditions.


These conditions of the atom cells used in AFS are ever changing


to meet the needs of the proposed study, resulting in discrepancies between values for


the same type of atomizer.


laboratory


Even if the same conditions are reported from laboratory to


there are inconsistencies in reported information.


Investigation of Atomic Parameters Using LEAFS


While it is not necessary to use laser excitation AFS for the determination of









theoretical modeling for calculation of this information under saturation conditions.


most commonly


used


theory


to model


atomic transitions


follows the


equations


approach,


and this approximation will be discussed in detail in a later chapter.


Through the implementation of LEAFS and the just mentioned theory,


several


physical


parameters


have


been


evaluated


including


transition


lifetimes.' 0l'04 '1S


transition


transition


oscillator


strengths. 07


With


temperature


information and the knowledge of one or more physical parameters, it is sometimes


possible


to utilize


rate equations


theoretical


modeling to calculate other atomic


parameters such as quantum efficiencies."

LEAFS was employed in a reduced pressure plasma" as a diagnostic tool in


this atom


source


to gain a better understanding


ICP as


sampled


a mass


spectrometer.


The sodium D lines were observed because of their high fluorescence


quantum efficiencies, and the resulting information was used to profile the temperature


of the expanded plasma.


A similar study was carried out for a reduced pressure flame


atomizer,'10


again as


a diagnostic tool


to investigate the


temperature profile


flame expansion.


Investigation of Atomic Parameters Using Two-Color LEAFS


While the need for fundamental atomic information for transitions involving the

ground state is not yet completely met, the immediate need in this field is information


about transitions involving two atomic excited levels.


Two-color excitation spectroscopy


probabilities, 06









is a fast growing field,


and yet the fundamental information about these transitions being


utilized in these investigations is practically nonexistent."

Two-color excitation fluorescence spectroscopy has proven to be a very useful


diagnostic tool.


These types of techniques have been used to a great extent in the


evaluation of combustion species in flames such as oxygen,110 atomic nitrogen,111 and

hydrogen."*2,113'14'115


There


are many possible variations on


the technique of two-color excitation


spectroscopy including monitoring the fluorescence from the intermediate excited state


level instead of the highest excited state.


These experiments usually entail the monitoring


of a fluorescence signal as the result of one step excitation and the subsequent addition


of a


second


excitation


step tuned


to a


transition


allowing it to directly perturb


population of the monitored level.


In most cases,


the fluorescence signal from


intermediate


level


decrease


upon


introduction


second


excitation


responsible for depleting the population of the monitored level.


This technique has taken


on many names


literature


methodology


techniques


essentially the same.

This type of excitation and fluorescence scheme was implemented in 1983 for the

determination of the energies of the Rydberg levels of nitric oxide in a supersonic


jet.'"6


Further


work


same


research


group117",


same


area


showed


that it was possible to determine absorption cross sections for excited state transitions

involved in the excitation and fluorescence scheme.







67

Extinction spectroscopy was discussed by Pedrotti"9 for the study of lithium in


determination of radiative lifetimes and autoionization rates.


The author used "high-


powered narrowband dye laser radiation to extinguish the emission of VUV


from lines emitted by a hot plasma."


radiation


This was accomplished through the introduction


of the radiation from this dye laser tuned to a transition connecting the upper level of a


strongly emitting


VUV line to another excited state.


The author suggested


that this


technique was useful because it allowed the observation of levels not seen in emission


spectroscopy


to rapid


upper


level


radiation


seen


in absorption


spectroscopy because these transitions often times do not involve the ground state.

Another variation on this theme is photionization controlled-loss spectroscopy used


Salmon


Laurendeau'2o,12'


investigation


atomic


hydrogen


premixed hydrogen/oxygen/nitrogen flame.


The main obstacle in the prior studies for


this purpose was that quenching was dominant as a loss mechanism difficult to control


or monitor.


In the case of photoionization controlled-loss spectroscopy, however, the


addition of a second excitation step selectively ionized atoms initially excited through a


first excitation step making this selective ionization the primary loss mechanism.


This


was a process that could be controlled and compensated for with the fluorescence signal

from the originally excited level being monitored and decreasing upon introduction of the


second excitation


step due to this


second laser'


depletion of the population of this


monitored level.


Using this technique,


the authors were able to profile the number


density


atomic


hydrogen


in several


different


ratios


demonstrating








Fluorescence


reduction


spectroscopy


used


Bonin


determination of the absolute photoionization cross section of the cesium


'D, level.


They


used


lasers


to first


excite


an atom


an excited


state


through


absorption of two photons of the same wavelength via a virtual level and then subsequent

photoionization as a result of the introduction of another dye laser while monitoring


direct line fluorescence from the initially populated level.


This study also demonstrated


the utility of these methods as diagnostic tools in providing much needed physical atomic

information.

These types of techniques are based on the monitoring of an excited level as the

population of that level is being depleted through the introduction of a second excitation


step.


Intuitively,


the fluorescence signal will decrease upon introduction of this depleting


excitation radiation and this decrease in signal has caused


these techniques to more


generally be termed fluorescence dip spectroscopy.


Pago and


Gudeman'23 used fluorescence dip spectroscopy to study transition-


metal atoms produced in a rf glow-discharge sputtering machine.


Rydberg spectra of


titanium


vanadium


, cobalt, and nickel were analyzed to obtain ionization potentials


agreeing with other reported results to 1 cmn


. Also obtained were quantum defects for


nd and ns series not showing detectable s-d mixing which would lead to asymmetric line

shapes characteristic of autoionization.


Axner,


Norberg,


Rubinsztein-Dunlop14


investigated


physical


properties of laser enhanced ionization (LEI) and LIF (laser induced fluorescence) signals


et al./








obtained theoretically through a rate equations approximation.


The authors also studied


the effect of laser intensity on the resulting fluorescence signal as well as the fluorescence

dip reinforcing what Peipmeieri2 discussed about the importance of saturation conditions.

Another study presented was the relationship between LEI enhancement and the LIF dip.


As was expected,


the closer to the ionization continuum the atom was promoted,


greater the resulting LEI signal due to thermal ionization from the excited level.


This


phenomenon was studied while varying both first and second step intensities, and the


results were presented and fully explained


through theory.


It was suggested in this


article that fluorescence dip spectroscopy had tremendous potential as a diagnostic tool.


1992,


et al.125


used


fluorescence


spectroscopy


to investigate


barium atom and, specifically, the 6s6pT'P to 6s5d'D2 transition.


The authors reported


that the nonspontaneous radiative (amplified spontaneous emission or superradiation)

relaxation of barium 6s6p1P1 to 6s5d'D2 from the laser excited upper level was enhanced

by laser light that was resonant with the 6s5d'D2 to the 5d6p'F3 transition to deplete the

population of the lower level.


Simeonsson,


Winefordner126


used


fluorescence


spectroscopy


measure the quantum efficiencies of fluorescence for excited state transitions of silver,


copper, iridium, and lead in an ICP environment.


The results were shown to be in good


agreement with theory, again based on a rate equations approach assuming a three level


system.


authors


commented


in most


cases


three


level


model


oversimplification and that some modifications needed to be made to compensate for the








Taking into consideration the fact that, as mentioned,


a three level approximation


was not a


valid one in


most cases,


Simeonsson et al127


discussed a rate equations


approach for an approximated five level atom system and the possibility of a negative


fluorescence dip.


In other words


, the authors predicted that there was the possibility of


an increase in the fluorescence signal upon introduction of a second excitation step in a


fairly


collision


atomization


medium.


More


possibility


experimental data to support it will be discussed later.


Advantages and Disadvantages of LEAFS as a Diagnostic Tool


actual


In the previous discussion,


it has been demonstrated that LEAFS is an invaluable


better


understand


atom


reservoirs


which


analytical


chemists


perform


investigations as well as the atom systems they perform these investigations on.


This has


been


shown


to be


one of


pnmary


needs


field


analytical


chemistry.


Unfortunately, while LEAFS is the best method through which to obtain this information,

it carries with it the limitation of the expense and complexity of the required system.

This has been discussed earlier and is the greatest disadvantage as applied to LEAFS.


In the case of analytical studies,


this disadvantage was overcome through the


implementation of conventional excitation sources for AFS or the use of OES or AAS


as alternative techniques.


The application of these methods for diagnostic studies as


alternative


techniques


not yield


desired


information.


This


most easily


performed through the use of lasers since the theoretical treatment applied requires






71

saturation conditions which only the use of an excitation source with the power of a laser

can attain.












CHAPTER 4
INSTRUMENTATION USED IN LEAFS


The instrumental requirements


LEAFS


are similar to those


for emission


spectroscopy with the exception of the need for an external excitation source, in this


case


, a laser.


block diagram


of instrumentation required


to perform a LEAFS


experiment is shown in Figure 4-1.


source


The experimental setup consists of an excitation


, an atomizer in which the analyte is investigated, a wavelength selection device,


a photodetector, and an electronic amplifier/readout system.


The excitation source is


positioned perpendicular to the collection axis to minimize the detection of radiation from

the excitation source.


Excitation Sources


The ideal source for AFS experiments would be stable and emit radiation at the


desired wavelength at a very high radiance.


Lasers make excellent sources for AFS with


radiant power emitted from them being several orders of magnitude greater than that for


conventional excitation sources.


Both continuous wave (CW) and pulsed dye lasers have


been used to provide tunable radiation as the excitation source, but pulsed dye lasers have


seen the most use in LEAFS.


The very high peak irradiances achieved through the use


of pulsed dye lasers easily result in saturation conditions for the transition of interest.


**** 1- A -_ _.__-- -- -- ---- _* A. L a 1


ak .































e
10



E

u,
e
C^


* -
C3

*<

O
0
E
S
s


* -

a
d-

8
60

I
II-




















tl
* C,








reducing the noise contribution of the laser scatter in the atom cell.


While the use of


laser excitation is the closest to an ideal source for


AFS,


the high cost involved in


obtaining appropriate instrumentation has kept the full potential of this technique from

being realized.

While dye laser excitation is typically used as the means for excitation in LEAFS,


dye lasers require excitation themselves in order for a lasing action to occur.


Flashlamp


pumping has been used, but the resulting pulse from the dye laser has a low average


power because of


long


pulse duration,


which in


many


cases


will not result in


saturation conditions.


Laser pumping results in shorter pulses from the dye laser with


total per pulse power sometimes lower than that for flashlamp pumped, but with peak

intensities that are much greater usually effecting saturation conditions.


Principles of Lasers


The laser consists of three components


1) an active medium which amplifies an


incident electromagnetic wave, 2) an energy source which selectively pumps energy into

the active medium resulting in the population of selected levels and eventually resulting


in a population inversion, and 3) an optical resonator, composed of two mirrors,


stores part of the emission and allows some to escape as the lasing.'28


which


A population


inversion occurs when the upper level population can be made to exceed that of the lower


level, and when this occurs, the system behaves as an amplifier.

atomic or molecular system is said to be an active medium.129


When this occurs, the

Such an amplifier is








The optical resonator


(Figure 4-2) causes selective


feedback of the radiation


emitted from the active medium.128


This active medium is located between two mirrors


in the resonator, o

cavity length as d'


me of the mirrors being partially transparent.


, the optical pathlength as d, and


Defining the resonator


the length of the active medium as


b, d is given as


(4-1)


where q is the refractive index of the active medium and j'


atmosphere in the rest of the resonator.


is the refractive index of the


The electromagnetic field is directed back and


forth through this resonator cavity and is amplified with each pass through the active


medium.


Lasing is produced when some of this amplified radiation is allowed to pass


through the partially transparent mirror.

oscillator above a certain pump threshold.


This laser amplifier can be converted into an

Oscillation begins when the gain of the active


medium becomes equal to the losses in the system. 29


There are many ways through which a population inversion can be attained.


is not possible to reach a population inversion in a two level system since the absorption

and emission processes must be equal and the maximum population that the upper level


can attain would be one half of the total population. Therefore, systems resulting in

lasing action must consist of three or more levels (Figure 4-3). The required excitation


process,


called


pumping,


can be achieved


through a number


methods including


electrical methods emDlovine an electrical discharge. by chemical reactions.


or by rapid


d=-l b+' /(d d







































*
I-N
0

Cu




a

"I

* -


-4




'4.





78












Fr
06
-----1





| i





[e m
- ---r--- I- 11111




1 I

--


*
I I- --








































Figure 4-3 : L


asing schemes. (A) In the three level scheme, atoms are
pumped to level 2 and this population undergoes rapid relaxation
to level 1. If the lifetime of level 1 is significantly greater
than that of level 2, the population of level 2 will grow and
eventually become greater than the population of the ground state.
(B) In the four level scheme, pumping occurs from level 1 to level
3 resulting in a population inversion between levels 2 and 1. It is
easier to attain a oooulation inversion in the four level case.


. ,I .





80



2
Fast Decay
I1


Lasing


Fast Decay


Pump


I








sometimes with efficiencies of 20% or lower, but the resulting radiation has properties


unique to lasers alone and,


specifically


different than conventional excitation sources.


Types of Lasers


One laser that has been used as a pump for a dye laser in LEAFS experiments is


the Nd:YAG.


This laser consists of Nd"3 in a yttrium-aluminum-garnet host.


The active


medium in the case of thi

or pulsed radiation. Th


laser is a rod which is optically pumped to result in either CW


ie pulsed Nd:YAG has laser pulse durations on the order of


nanoseconds and enjoys very high output powers.


at 1.06 prm,


The Nd:YAG laser emission occurs


and this wavelength is frequency doubled into the visible range for use as


a pump source for a dye laser.


The most common


and numerous


, of all lasers are gas lasers.


All of the noble


gases have been made to lase,


and these lasers


, along with the helium-neon laser, are


some of the most reliable lasers.

as pump sources for dye lasers.


Unfortunately, they are not really appropriate choices

One gas laser commonly used as a pump laser for a dye


laser is the nitrogen laser which uses a vibronic transition of the nitrogen molecule at


337.1 nm.129


The nitrogen laser emits a short pulse with peak powers up to 100 kW


Excimer lasers are among the newest of the gas lasers and are used to a large


extent as pump lasers for LEAFS experiments today.


mixture (e.g.


In this type of laser, a gaseous


, and Ne) is subjected to an electrical discharge resulting in the


formation of an excimer (e.g.


XeCI),


hence the name excimer.


Excimers are molecules





































Figure 4-4


Schematic potential diagram of an excimer molecule.
















































- -


A+B








(Figure 4-4).128


The excimers of different gas species are ideal candidates for the active


medium of lasers since a population inversion is constantly maintained.


ultraviolet wavelengths are available with different gas mixtures (Table 4-1).


riety of

These


lasers are tunable over a small


wavelength range dependent upon


"the slope of the


repulsive potential and on the internuclear positions r, and r2 of the turning points in the


excited


vibrational


levels"


(354). 28


(Figure


4-4).


Because


relatively


long


wavelength (308 nm) and long lifetime of the gas mixture,


the XeCI excimer laser has


proven to be the most useful of the excimer lasers for pumping dye lasers.


The dye laser is the main source of excitation for LEAFS experiments,


mentioned earlier, it requires pumping from another excitation source.


but, as


The lasers just


discussed as well as a flashlamp and CW laser not included in this discussion have been


used for this purpose,


and the results of these studies are shown in Table 4-2.


The dye


laser cavity (Figure 4-5) consists of a dispersive device allowing for wavelength selection


and optics directing the excitation radiation through a dye cell.


This cell is filled with


a solution of dye which absorbs the incident radiation resulting in oscillations within the

dye cell and subsequent lasing emitted from the cell.


Atomization Reservoirs used in LEAFS


Several atomization cells have been evaluated for implementation with LEAFS.

Most LEAFS work has been performed in flames, but more recently other atomizers are


being


routinely


used


LEAFS


including


plasmas,


glow


discharge


devices,


V2








Table 4-1

Typical Excimer Laser Characteristics'29


Laser Wavelength Pulse Energy Average
Medium (nm) (J) Power (W)

ArF 193 0.2 0.3 10
KrC1 222 0.03 1
KrF 248 0.3 0.4 18
XeF1 308 0.08 0.2 8
XeF 351 0.08- 0.15 7









Table 4-2

Typical Pulsed Dye Laser Characteristics29

Pump Tunability Peak Pulse Repetition
Source Range (nm) Power (kW) Duration (ns) Rate (Hz)

Flashlamp 220 960 100 500 250 750 1 10
N2 400-970 1-100 1-8 1-100
Nd:YAG 195 500 100 10,000 5 -10 1 -30
Excimer 217 970 100 1000 10 -20 1 -500
CW' 400 -1000 0.1 5 0.015 103 4x103


Synchronously pumped, mode-locked,


cavity dumped.







































Figure 4-5


: Dye laser.















Grating







Telescope



Oscillator Cell


Amplifier Cell


Dye Laser


Pump
Laser


Output


1








ideal atomizer would have a large and stable nebulization-atomization efficiency.


That


it is desirable


to maximize


the percentage of nebulized


analyte


that is actually


atomized and present for analysis in the atom cell.


particular,


AFS


methods,


atomizer


should


have


a high


enough


temperature and long enough residence time to insure complete vaporization of solution


particles,


thereby reducing the scatter noise which, in some cases, is the limiting noise


in a fluorescence measurement.


If an excited state fluorescence scheme is being studied,


or if thermal excitation is necessary from the excited level to the level from which the


fluorescence is being monitored,


these processes should easily take place in the chosen


atomizer.


Flames


The flames used in spectroscopic analyses are hot, chemical flames resulting from


reaction between a fuel and an oxidant. Chei

based on the fuel/oxidant combination used.


mical flames have characteristic structures

Flame types can be separated into the


categories of turbulent flames and laminar flames.


Turbulent flames


were


flame type applied


to LEAFS


because


total


consumption burners result in a higher concentration of atomic vapor in this atom cell


available for excitation.


Total consumption burners are unpremixed burners which,


the name implies,


introduce the entire aspirated


sample into


flame cell.


This,


combined with the fact that these burners are operated at lower temperatures than laminar








causes scatter noise from the laser radiation reflecting off of these droplets.


Because of


this large disadvantage, turbulent flow burners have enjoyed limited use in LEAFS.130

The implementation of laminar, premixed flames in LEAFS results in reduced

droplet size and, therefore, a decrease in the laser scatter noise, but at the same time

introduces the disadvantage of the consequent lower atom concentrations available for


analysis.


addition


to this


reduction


noise,


premixed


flames


are sometimes


surrounded by a sheath of inert gas (i.e. argon or nitrogen) to minimize the interaction


of flame gases


with air.3"'


These flames have


less flicker noise problems than the


turbulent flow burners simply because of the smooth, steady flow produced by this type


of burner.


Most laminar


flames


use a


pneumatic,


concentric


nebulizer


spray


chamber (Figure 4-6).


In most cases, capillary burner heads are used to make it possible


to change the shape and pathlength of the flame.'2

The type of flame to be used is typically selected on the basis of compromises

among atomization efficiency, background emission, and quenching characteristics29


Qualitatively, quenching properties increase in the order Ar


< H2


< H20


< N,


< CO


< 02


< CO2.


Early on, hydrogen-based flames were employed because of the low


background


improved


noise


detection


quenching.


limits over those obtained


mixture


oxygen


in hydrogen


based


argon


flames,


provides


but these


mixtures result in lower temperature flames which do not efficiently vaporize all of the

introduced sample.

Most LEAFS experiments have been carried out in air/acetylene or a nitrous





























E
=-

a)
.0




irt

3
I








Id-
U








.0;
Cu
ca
F3
U


04




* -




"0
gi-i
tc





91



















t 1




,i iI l








interference and scattering.


There is a higher quenching efficiency in such sources, but


many times this limitation can


be offset through the use of lasers as


the source of


excitation.


Plasmas


A plasma is a hot, ionized gas.


Electrically generated plasmas have been used


as an atom source for optical emission spectroscopy,


AAS


, more recently,


AFS.


These


flame-like


plasmas


operate


at significantly


higher


temperatures


flame


atomizers and are chemically inert environments compared to flames.


The sample is


introduced into the plasma through the use of a pneumatic nebulizer where it is readily

vaporized and atomized in this high temperature atom cell.

The most commonly used plasma in the field of AFS is the inductively coupled


plasma


(ICP).


The plasma is


typically


formed in a quartz tube surrounded


by an


induction coil which is connected to a high frequency generator operating at frequencies


from


to 50 MHz (Figure 4-7).


. An inert gas,


usually argon,


The resulting output powers are typically between 1 to


acts as the support gas for the plasma as well as a


the coolant gas for the quartz tube."12


The plasma is formed when a Tesla coil produces


"seed"


electrons


and ions in


this quartz tube.


plasma


will form


, provided the


magnetic


field


strength


is strong


enough


streams


follow


a particular,


rotationally symmetrical pattern. 32


For the implementation of LEAFS in an ICP


, the forward power must be kept as





































Figure 4-7:


Operation of the ICP. (A) An induction coil connected
to an RF generator surrounds a quartz tube through which Ar is
flowing. A plasma then forms when the Ar becomes conductive.
(B) The magnetic fields (H) and eddy currents (I). The high
frequency alternating currents in the coil generate magnetic
fields. The ions and electrons are accelerated in an eddy current.
(C) Complete ICP torch.













PlMaax


RF Power
seaPS


S Quarz Tube


Argom Tagetial
SCoolant Flow


Ar and
Sample


Optimal Ar


Indnmeimn


H/T