Excimer laser excited atomic fluorescence detection of arsenic, and the design and characterization of a flowing electro...

MISSING IMAGE

Material Information

Title:
Excimer laser excited atomic fluorescence detection of arsenic, and the design and characterization of a flowing electrolytic hydride generator for arsenic and selenium analysis
Physical Description:
x, 206 leaves : ill. ; 29 cm.
Language:
English
Creator:
Hueber, Dennis Michael, 1965-
Publication Date:

Subjects

Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1994.
Bibliography:
Includes bibliographical references (leaves 200-205).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Dennis Michael Hueber.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001973107
notis - AKE9899
oclc - 31743890
System ID:
AA00003214:00001

Full Text








EXCIMER LASER EXCITED ATOMIC FLUORESCENCE
DETECTION OF ARSENIC, AND THE DESIGN AND CHARACTERIZATION
OF A FLOWING ELECTROLYTIC HYDRIDE GENERATOR FOR ARSENIC
AND SELENIUM ANALYSIS











By

DENNIS MICHAEL HUEBER


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

1994












ACKNOWLEDGEMENTS


I would like to thank my parents.


As if giving me life were not enough,


also gave me their love,


guidance, support and patience.


Thanks also go to my sisters,


for everything they have done for me, and everything they didn't do to me.


Professionally,

get me hooked on che.


James Winefordner.


the list of acknowledgements must start with Dr. Leo Bares who

mistry. I would like to acknowledge my research advisor, Dr.

)r. Winefordner has always found a way to give me the guidance


I needed while allowing me to make the mistakes it takes to really learn.


I will be


eternally grateful for the chance he has given me to pursue my own ideas and for his

patience and humor.


My thanks go to Ben Smith for all his assistance with my research.


always available to point me in the right direction.


research group.


Ben was


I would also like to thank the whole


I could not have completed this work if I did not enjoy coming to work


each morning with this great group of people.


I can't mention all their names here, but


I can't go without listing some of them.


My special thanks go to:


Tye Barber, Mike


Wensing,


Guisseppe


Petrucci


Sean


Madden,


Martin


Angebrant


Wellington


Massamba.


If I never work with a better group of people,


I'll consider myself lucky.














TABLE OF CONTENTS


PAGE


ACKNOWLEDGEMENTS

KEY TO ACRONYMS USED


* 3 3 3 3 3

3 . 3 3 3 9 3 .


. 1

. vi


ABSTRACT


CHAPTERS


.. IX


Si


INTRODUCTION
General Introduction .
The Scope of This Dissertati
Suggested General Texts on


Notes Concerning Atomic
Solution Nebulization for
in Atomic Spectrometry
Hydride Generation
Introduction .


Chemistry of Hydri
Methods Used to G
Method of Detectio.
Interference .
Electrolytic Hydride
Light Sources for Atomic i
Incoherent Line Sot
Dye Laser Sources
The Nature of ArF Laser I


INSTRUMENTATION USEE
FAAS, ICP-AES AND AFS
Introduction ..........
Atomizers, Emission Sources
The FAA Spectrometei
The ICP-AE Spectrom
The ICP-LEAF Spectre
Electronics Systems .


a,


on .............. 2

Atomic Spectroscopy .. .. 4
oectrometric Terminologv .. 5


Sample Introduction
3 6 S S 3 5 3 5 5 5 S 3 3 3 3 3 3 5


* 5 5 . S 3 3 3 . S 9
* 3 S 3 3 3 . .9
de Generation 10
generate Hydrides . . 12
n for Hydride Generation ...... 14
...... ................ 14
e Generation 17
Fluorescence Spectrometry ... 20
Jrces * * 20
. . 22
radiation . . . 24


...27
....................27
and Optics . . . 28
r .................. 28
eter 31
meter 35
. . ........ 39
*-F-


1
t
(







Electronics for ICP-AES
Electronics for LEAFS
Performance and Comparison to
Commercial Instrumentation
FAA Spectrometers .
ICP-AE Spectrometry .


.* . . .. .. . 43
* . . . . 43

* C . . . 57
. .. .. . ..... 57
. ............... 59


DESIGN OF THE CONTINUOUS
ELECTROLYTIC HYDRIDE GENERATOR


Introduction . . . . .
A Detailed Walk Through and Around thi
Voltage Regulation . . . .
The Cell Voltage and Overpotential .
Contributions to the Cell Potential
Resistance Terms .
Overpotential . . .


. . U.61


....... ..... 61
e Generator . .62
. . . . 74
.... 74
. . 745

.......... 75
........... 76


CHOICE OF ELECTRODE MATERIALS


AND ELECTROLYTES
Introduction . . .
Experimental
Cathodes . .
Reagents and Stock Solu
Preparation of the FAA


Results


Preparation of
Determination
Carrier Gas
Procedure for
and Discussic


the Gener


...................83
. ..................83
...................85
................. 85
.tions 85
Spectrometer . . 85
Sator 86


of Optimal
Flow Rate . . .
Cathode and Electrolyte Tests


In


. S S .


S87
S87
* . 88


ANALYTICAL PERFORMANCE OF CEHG
WITH FAAS AND ICP-AES DETECTION
Introduction . . . . .
Experimental. . . . .
Reagents and Stock Solutions .
General Procedure . . .
Procedure used for the FAAS
Calibration Curves
Procedure Used for the ICP-AES
Calibration Curves .
Sample Preparation and Procedure
for the Analysis of SRMs .
Procedure for the Estimation of


S. 101
* . 101
S102
S102
. 103


. 104


* *. S S S 105

* 106


--








Results and Discussion


Analytical Figures of Merit
Memory Effects
Concomitant Effects .


Results of the Analysis of the SRMs
Estimation of Efficiency .


. . . C110


* a a a 110
* a a a .. 115


. a a a a a a 115


* a . 126
. a a S S 126


Effect of Flow Rates with ICP-AES Detection


A Simple Mathematical Model
Conclusions .. ... .. .


. 130


S. 133


. 135


ARGON FLUORIDE LASER EXCITED ATOMIC


FLUORESCENCE OF


ARSENIC


.. .... ...... 137


Introduction


.. .. . . . 137


Measurement of the ArF Laser Spectral Profile


Experimental . .
Results and Discussion


U SS S 1410


* . C S . 140
a a a a a a a a 141


Comparison of two Atomizers for Arsenic LEAFS


Introduction
Experimental


* .*S S U S U a a a a a a .. . a a a a 144
a a 144


Results and Discussion


Conclusions


a a S 55 S S S C SC 5 14~7


. S a a a a a a a a S a a a a 5 170


Sample Introduction Using the
Electrolytic Hydride Generator


Introduction
Experimental


Sa a a a a a a a a 170


* a a a a a a a a a a a a a a a 170
* a a a a a a a a a A 170


Results and Discussion


Conclusion


* a a a a a a a a a .1'72
S. a a a 182


CONCLUSIONS AND POSSIBLE DIRECTIONS


FOR FUTURE RESEARCH
A Summation of the LODs
Observations on ArF Fluorescer
Observations on CEHG
Two Suggested Future Noveltie


APPENDIX.


187


REFERENCES


a 200


BIOGRAPHICAL SKETCH


a206


1447


.................. 183
.................. 183
nce 183
.................. 185
s ................185











KEY TO ACRONYMS USED


The following is a list of acronyms used in this dissertation.


Occasionally a


lower case "s" is added to an acronym in the text to pluralize the acronym (i.


is used for hollow cathode lamps).


HCLs


In contrast, a capital "S" added to the end of an


acronym


creates


a new


acronym;


usually


stands


spectrometry.


example,


stands


for atomic absorption,


AAS


stands


for atomic absorption


spectrometry.


In addition two or more of these acronyms may be combined with a


hyphen to create a new acronym.


For example, ICP-AES, ICP-LEAFS or HCL-AF.


Acronyms ending in


for spectrometry are included in this list.


The hyphenated


acronyms are not.


Atomic absorption.


AAS: Atomic absorption spectrometry.


AE: Atomic emission.

AES: Atomic emission spectrometry.

AF: Atomic fluorescence.

AFS: Atomic fluorescence spectrometry.

CEHG: Continuous electrolytic hydride generations.

DCP: Direct coupled plasma atomic.

DIN: Direct injection nebulizer.







EDL:

GF:

FAA:

FAAM

FAES

FIA:

FWH

HCL:

HPL(

ICP:

LEAF

LEAF

LEI:

LDR:

LIF:

LOD:

MIP:

MS:

ppm:

ppb:


I








I
4
E]



.,


Electrodeless discharge lamp.

Graphite furnace.

Flame atomic absorption.

S: Flame atomic absorption spectrometry.

i: Flame atomic emission spectrometry.

Flow injection analysis.

M: Full width at half maximum.

Hollow cathode lamp.

: High performance liquid chromatography.

Inductively coupled plasma.

: Laser excited atomic fluorescence.

S: Laser excited atomic fluorescence spectrometry.


Laser enhanced ionization.

Linear dynamic range.

Laser induced fluorescence.

Limit of detection.

Microwave induced plasma.

Mass spectrometry.

Parts per million.

Parts per billion.


OES: Optical emission spectrometry.







RF:

RSD:


Radio frequency.

Relative standard deviation.













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

EXCIMER LASER EXCITED ATOMIC FLUORESCENCE


DETECTION OF


OF


ARSENIC, AND THE DESIGN AND CHARACTERIZATION


A FLOWING ELECTROLYTIC HYDRIDE GENERATOR FOR ARSENIC


AND SELENIUM ANALYSIS

By

Dennis M. Hueber


April 1994


Chairperson:


James D.


Winefordner


Major Department:


Chemistry


Currently there is great interest in the detection of arsenic and selenium at trace


ultratrace concentrations


in environmental


biological


samples.


Two


new


methods have been assessed in


this work.


The first is a novel method of hydride


generation for sample introduction into an atomic spectrometer, and the second is a

method for the detection of arsenic.

Hydride generation is a popular alternative to solution nebulization in analytical

atomic spectrometry for elements that easily form volatile hydrides (i.e., As, Se, Sb,


because


it increases


sensitivity


analytical


atomic


spectrometric detection.


Presently, the common methods involve the reduction of these


elements by sodium tetraborohydride.


A continuous electrolytic hydride generator was


4~~r C C


1 1 1


r r





1 *







solutions containing trace aqueous concentrations of arsenic,


selenium or antimony.


This new hydride generator can be used for discrete or continuous sample introduction.

This method has the advantage of requiring no chemical reducing agent, and thus the

potential for sample contamination is reduced.

Laser excited atomic fluorescence spectrometry (LEAFS) can be an extremely


sensitive


method.


However,


some


elements,


such


as arsenic,


only


have


strong


absorption lines below 200 nm.


Dye laser systems capable of the resonance excitation


of arsenic are complicated and excessively expensive.


By coincidence, a simple fixed


frequency argon fluoride (ArF) excimer laser produces a broad band output centered at


193.0-193.2 m that overlaps with the arsenic absorption line at 193.7 nm.


The use of


an ArF excimer laser as a source for the LEAFS detection of arsenic in a flames and


an inductively


coupled


plasma


was evaluated.


This


method of


detection


was also


evaluated when coupled with sample introduction using the flowing electrolytic hydride

generator.













CHAPTER 1
INTRODUCTION

General Introduction


Arsenic and selenium have been known poisons since ancient times.


arsenic has long been a poison of choice, both in fiction and reality.


Indeed,


So for obvious


reasons, long before trace methods of analysis were developed for most elements they


were available for arsenic.


The classic Marsh test dates back to 1836, and the Gutzeit


test dates to


1870.


While selenium is poisonous in large amounts, it is a vital nutrient


in small concentrations.


biochemists and


Arsenic and selenium have been


environmental chemists.


great interest to both


In fact, this research was funded in part by


the United States Environmental Protection Agency (EPA


- CR 8188 48-01-0).


EPA has set maximum contaminant levels for arsenic and selenium in drinking water.

The maximum contaminant level for As is 50 pg/mL, and it is 10 pg/mL for Se.


Both


arsenic


selenium


present


special


difficulties


optical


atomic


spectrometric methods.


These elements have ground state absorption lines only below


200 nm,


region


often


called


vacuum


ultraviolet.


Flames


have


increased


background


absorption


levels


at these


wavelengths,


emission


lines


are weak


compared to lower energy transitions.


Hollow cathode lamps (HCL) are also weak for


these elements, and few laser systems can produce light below 200 nm.


Furthermore,












for use at longer wavelengths.


Even inductively coupled plasma mass spectrometry


(ICP-MS), which has become the premier atomic technique, has higher detection limits


for these elements than for most other elements.


This is likely due to a combination of


their high ionization energies and isotopic interferences1


Fortunately,


these elements are amenable to a


particularly


efficient form


sample introduction,


hydride generation.


Using hydride generation, these elements may


be detected at levels below many other elements.


course,


there is a drawback.


Hydride generation is far more complicated than solution nebulization, and more risk

of sample contamination exists because of the necessary addition of several reagents,


including


reducing


agent,


sodium


borohydride


(NaBH4).


With


improved


sensitivity provided by hydride generation, it is important that the chance of sample


contamination be reduced.


With the electrolytic approach to hydride generation,


reagent used in the largest quantity, the electrolyte, can be purified of hydride forming

elements by electrolysis.

In this dissertation an apparatus and a method for continuous electrochemical

hydride generation (CEHG) will be described and characterized for sample introduction


in atomic spectrometry.


In addition, the use of an ArF laser is examined as a potential


excitation


source


for the detection


arsenic by


laser excited atomic fluorescence


spectrometry (LEAFS).

The Scope of This Dissertation

The purpose of this dissertation is to describe original research and the results








perspective on these results.


It is not the intention of the author, nor is it feasible, that


this dissertation contain a complete introduction to the methodology and instrumentation

of atomic spectrometry nor any of the other broad topics that are peripherally touched


upon by the research described here.


Some recommended texts are suggested below.


In this first chapter a general introduction to solution nebulization and hydride


generation for sample introduction in analytical atomic spectrometry is given.


A brief


description of the most common sources that have been used for atomic fluorescence


spectrometry (AFS) will follow.


Emphasis will be on sources that have been used for


arsenic fluoresc

Chapter


ence.


Finally, the nature of excimer laser radiation will be described.


2 describes three atomic spectrometers constructed during this research,


and compares their performance with commercially available spectrometers.


description of the hydride generator is given in Chapter 3.


A detailed


In Chapter 4, experiments


designed to compare a variety of electrolytes and cathode materials are described.


results of these experiments are discussed.


Chapter


focuses on the analytical use of the hydride generator in the continuous


sample introduction mode of operation.

was used to detect arsenic, selenium


Flame atomic absorption spectrometry (FAAS)


and antimony introduced into the flame by the


generation of arsine (AsH3), hydrogen selenide (H2Se) and stibine (SbH3).


Inductively


coupled plasma atomic emission spectroscopy (ICP-AES) was used to detect arsenic and

selenium.


In Chapter 6,


the use of an ArF laser for the laser excited atomic fluorescence








described.


Besides


solution


nebulization


aqueous


samples,


CEHG


was used


produce


arsine


sample


introduction.


summation


results


given


Chapter 7.


Also, suggestions for future research are given.


The order of the chapters presented here does not represent the order of the


conception


or the


performance of the


research.


The idea of CEHG


preceded


invention of the actual device by over a year.

Suggested General Texts on Atomic Spectroscopy

As mentioned above, three types of optical atomic spectrometry were used during

the course of the research presented in this dissertation, FAAS, ICP-AES and LEAFS.

A complete description of any of these atomic spectrometric methods is beyond the


scope of this text.


Several general texts are recommended.


Ingle and Crouch have written an excellent text on the subject of spectrochemical


analysis including atomic methods.


They give a bibliography for each major area at


the end of


each chapter.


Among


the classic texts on flame spectroscopy is that of


Mavrodineanu and Boiteux.3


Detailed information about flame methods can be found


classic texts by


Alkemade and


coauthors.4,5


excellent review


of inductively


coupled


plasma


(ICP)


spectrometry


(atomic


emission,


fluorescence


mass


spectrometry) can be found in Montaser'


book.


Also, Boumans has edited a text on


ICP-AES.6


Among the general sources of information on atomic fluorescence (AF)


spectroscopy are those of Winefordner and Omenetto.7'









Notes Concerning Atomic Spectrometric Terminology


language


atomic


spectrometry


is rich,


can be


confusing


ambiguous.


For example, Mavrodineanu and Boiteux refer to atomization as the process


of breaking a solution into small


droplets,3 while most current sources classify this


process as nebulization and reserve the term atomization for the process of converting


the sample into free atoms.1,4'5'9.10


This dissertation uses the later standard,


and the


device that atomizes the sample is called the atomizer.


To avoid


unwieldy


phrase


"atomizer,


atomic emission


source or atomic


ionization source" the term "atomic source" is used to refer to all three devices.


Another


point


potential


confusion


use of


acronyms.


Hundreds


specialized acronyms have been used to describe analytical methods and instrumentation.

Ingle and Crouch list almost 100 acronyms which concern spectrochemical analysis.1


In this dissertation, the use of acronyms will be more limited, but the uninitiated reader

may become confused by such terms as AF, AAS, LIF and LEAFS. Each acronym is


given in parenthesis following its meaning the first time it appears in this text.


author has invented one new acronym, CEHG, which stands for continuous electrolytic


hydride generation.


However, he resisted the temptation of inventing more acronyms!


Solution Nebulization for Sample Introduction in Atomic Spectrometry

According to Twyman, the first true examples of spectrochemical analysis were


done by Kirchoff and Bunsen in the middle of the nineteenth century."


They introduced


salts into flames by placing the salt, or a drop of a salt solution, on a wire and then








supply their flames with sample3, an idea first suggested by Morton.1'2


many new sample introduction methods have been developed.


Since then,


Browner and Boomrn3


have published a review on sample introduction in atomic spectroscopy; also, several


of the texts suggested above include sections on sample introduction.1-


Although there


are several sample introduction methods for use with small liquid or solid samples, this


discussion


is concerned


only


with


methods


used


solution


introduction


continuous atomic sources.

For the introduction of solutions in continuous atomic sources, nebulization is the


most popular method.


In FAAS and flame atomic emission spectrometry (FAES), ICP-


AES and inductively coupled plasma mass spectrometry (ICP-MS), the most common


nebulizers are pneumatic.


Ultrasonic nebulizers are also common for use with an ICP.


Pneumatic nebulizers use a jet of compressed gas to aspirate and nebulize the


solution.


In cross flow nebulizers, a flow of compressed gas passes over the tip of a


sample capillary, and the resulting Bernoulli effect creates a local pressure drop and


draws the solution through the capillary.


Similarly, in the concentric nebulizer a flow


gas passes over the


tip of a sample capillary


located inside the compressed gas


capillary. In both cases,


aerosol.


the sample solution is broken into tiny droplets forming an


In some applications, the solution is pumped into the nebulizer to control the


solution flow rate.


Ultrasonic


nebulization


second


most


common


method


sample


introduction for solutions.


In the ultrasonic nebulizer, a flow of solution is pumped over






7

A spray, or cloud, chamber usually surrounds the nebulizer and prevents the


largest


droplets


from


entering


atomizer.


Perkin-Elmer


flame


concentric


pneumatic nebulizer (the type used for the work described in this dissertation) uses a set


of propeller shaped blades, or spoilers, which help break up large droplets and prevent

those that do not break up from leaving the nebulizer. Other spray chambers use impact


beads for the same purpose.


The Scott type spray chamber, used for the ICP, has a


cylindrical baffle to force the larger droplets to fall from the aerosol before they can


leave the spray chamber.


For some applications, a condenser is employed between the


spray chamber and the atomizer to reduce the amount of solvent vapor that enters the

atomizer.

The efficiency of a sample introduction system can be broken into several steps.


For nebulizers, the solution is first broken into droplets. The percentage of the sample

solution that forms an aerosol is the "nebulizer efficiency." Next the aerosol must be


transported to the atomizer.

"transport efficiency." Of cc


emission or fluorescence can occur.


The total efficiency of the first two steps is called the


)urse, several other processes must occur before absorption,


The aerosol must be dried (desolvation efficiency);


the dry aerosol must be volatilized volatilizationn efficiency); and finally the chemical


bonds (if any) must be broken.


The fraction of the sample (in the atom source) present


as free atoms is termed the free atom fraction (P).


In emission spectrometry, the atoms


must also be excited, and in mass spectrometry, they must be ionized.

While it is useful to think of each of these processes and efficiencies separately,








atomizer, act as a team.


While one nebulizer may have a large transport efficiency, it


may also produce droplets too large to be desolvated.


In other words,


the various


efficiencies (nebulization, transport, desolvation, etc.) are not independent.


Also,


while the ( is often the result of a local thermodynamic equilibrium, it is


possible to cool the atomizer with too much sample vapor and thus affect the 3.


also possible to greatly affect 3 in certain atomizers that lack the energy to thermally


decompose


analyte


molecules,


introducing


sample


an easily


decomposed form.


For example, when hydride generation is used to introduce samples


into a heated quartz tube atomizer, a large free atom fraction is achieved despite the low

temperature of the atomizer.

The typical pneumatic nebulizer used for FAAS has a transport efficiency of 4-


The typical Meinhard (concentric glass) type pneumatic nebulizer used with


ICP-AES has a transport efficiency of 1-3 %.*,2'13 Transport efficiencies of 20-30 % have

been reported for ultrasonic nebulization.113,15

One hundred percent transport efficiency can be realized in flame spectrometry


using a total consumption burner.2'3


In the total consumption burner (or direct injection


burner), the concentric nebulizer is incorporated into the design of the burner so that


100% of the sample is nebulized directly into a surface mixed flame.


However, the


largest


droplets


not have


a long


enough


residence


time


ensure


complete


desolvation and vaporization.


Similarly, direct injection nebulizers (DIN) have been


developed for interfacing ICP-AES and ICP-MS with flow injection analyses (FIA) and








flow rates are used with these nebulizers.


Also, the large amount of solvent can "load"


plasma


reduce


atomic


emission


signal


compared


to standard


solution


nebulization.


Houk and coworkers'7 have demonstrated a DIN using a microcapillary


(10-25 pm diameter),


which produces very uniform droplets.


However, this type of


nebulizer clogs easily.


Hydride Generation


Introduction


In many cases, analytes in solution can be chemically transformed to volatile


derivatives.


These


volatile


derivatives


introduced


atomic


spectrometric source as gases.


Unlike solution nebulization,


where over 90% of the


sample is usually discarded, gas generation techniques often approach 100% transport

efficiency.


Among


types of


gaseous


sample introduction


methods


which have been


applied to atomic spectrometry are, the generation of molecular halides (12,


19 Br2 and


Cl220'21),


generation


of OsO4,


generation


from


carbonate and


generation of volatile metal-chelates.24


The two most widely used techniques are the


generation of atomic mercury and hydrides of As, Sb, Bi, Ge, Sn, Pb, Se and Te.


is fortuitous that these elements form volatile hydrides, since As, Sb, Se and Te only


have resonance lines below 220 nm.


This complicates the optical atomic spectrometric


detection


these elements.


Flames


have a


background


absorption


at short


wavelengths, and a high excitation energy is required to produce atomic emission.






10

a result, solution nebulization detection limits are poorer for these elements than for

many others.

The use of hydride generation in analytical chemistry predates its use as a sample


introduction technique for atomic spectrometry.


Arsine generation has long been used


for the separation and preconcentration of arsenic from complex matrixes.Y Hydride

generation is also the basis of the Gutzeit and Marsh tests.

Holak first demonstrated the use of hydride generation for sample introduction


in spectrometric


analysis


1969.


Over


papers have


been


published


since,


including several recent reviews.27'28,29,30,31


D&dina


review is primarily concerned


with the use of hydride generation with atomic absorption (AA) detection in flame-in-


tube atomizers, externally heated quart tube atomizers and


electrothermal


atomizers.


A good review of the mutual interelement effects of the hydride forming elements is


given.


Nakahara


review


concentrates


hydride


generation


with


ICP-AES.


Cambell's29


review


concerns both


emission and absorption detection.


Hershey and


Keliher30


concentrate on the generation of


AsH3 and H2Se.


Finally,


the review by


Heitkemper31


and coauthors is a thorough


review of hydride generation for sample


introduction into an ICP with almost 150 references.

Chemistry of Hydride Generation


Holak26 originally used zinc metal as the reducing agent.


However,


sodium


borohydride has become the reducing agent of choice.


The general reaction is,


\TaBI.


+ 31-J +


HCL


--H,BO,


+ NaC1 + 8fH


E"'


Bit


+ IL(excess)







11

where E is the specific element, m is the formal charge on the element and n is the


number of hydrogens in the product hydride.


Table 1-1 lists several hydrides that may


be formed and their names.


All of these compounds are poisonous.


Less than 0.5 ppm


(by volume) of arsine can cause death.32


Table 1-1:
solution.


Volatile hydrides of main group elements that may be formed in aqueous


FORMULA NAME(S) BOILING pK,
POINT (C ) VALUES

AsH3 arsine, arsenic hydride -61"

SbH3 stibine, antimony hydride, -17"

BiHl3 bismuthane, bismuth hydride 16.8"

H2Se hydrogen selenide, selenium hydride -43.3" 3.89
HTe hydrogen telluride -2" 2.64

GeH4 germane, germanium hydride -90b

SnH4 stanane, tin hydride -52"

PbH4 plumbane, lead tetrahydride -18.4c


" Value from reference 32.
b Value from reference 28.
Values from reference 33.


In practice,


an acidified solution


of the analyte is treated


with an


excess of


NaBIH4


(usually in solution).


The rate of hydride production depends on the chemical


form of the analyte, the concentration and type of acid used, the concentration of the


N~RFT.


uerd and the de ion nf the opPnratnr 27,28,32








Arsenic(III) reacts almost instantaneously in 1 M hydrochloric acid (HC1),


arsenic(V) reacts only slowly except in 4 M HC1.34


Since the higher oxidation states of


Se and


(As(V),


Sb(V)


Se(VI))


react


more


slowly than


lower


oxidation states (As(III),


hydride generation.


Sb(III) and Se(IV)),


The iodide ion (KI,


these species often are prereduced before


Nal) is often used for the prereduction of Sb


and As


while Se is either heated with HC1 or acidified bromide.'29


Methods Used to Generate Hydrides


Two approaches, continuous and batch,


are used in hydride generation methods.


Generally, in the batch method, an aliquot of the solution to be analyzed is acidified and

transferred to a reaction vessel; the NaBH4 (usually a solution) is then added to the


reaction


vessel.


formed and


The resulting hydride is swept out of the vessel by hydrogen gas


, usually, an additional flow of argon or nitrogen.


The generator itself can


be an Erlenmeyer flask, a test tube or some specially designed vessel.


In the continuous process,


the reagents are continuously introduced into a mixing


chamber (or in a length of tubing) where the reaction takes place.


Pumps are required


to pump the solutions into the system.


The mixture is


swept into a gas-liquid separator


where the liquid passes to a drain and the gaseous products pass to the atomizer or


excitation source.


Several groups have developed continuous systems in which


sample solution is mixed with the NaBH4 solution as an aerosol.35'36

the advantage of being easily incorporated into a FIA system,37 oa


This approach has


r it may be coupled


an HPLC.31,38


Also,


continuous


methods


are less


affected


chemical








interference.


disadvantages


are the


increased


complexity


cost,


increased dependence on the kinetics of the hydride generation reaction.

The efficiency of hydride generation can be defined as the percentage of analyte


leaving the generator as the hydride.


The efficiency of hydride generation depends on


the use of optimized chemical and physical parameters,28 and many published studies do


not include any estimation of the efficiency.


In general, efficiencies of nearly 100% can


reached


batch


continuous


methods,31 but


a compromise


between


efficiency and analysis time may be preferred.

Once formed, the hydride may be swept directly into the atomic source, or it


may be trapped.


Liquid nitrogen,


or argon,


cooled U-tube traps have been used to


condense AsH3, SbH3 and H2Se.31'39'40

generator, the trap is warmed, and thi

into the atomic source. Also, the hyd


vapor which are also produced by the generator.

limit because of the preconcentration. However,


Once all of the hydride has been released by the


e hydride is introduced as a preconcentrated plug


ride can be separated from the H2, CO2 and water


There is a decrease in the detection


there is a great increase in the time of


analysis and in the complexity of the apparatus and procedure.

Several companies market commercial hydride generators, including most of the

major atomic spectrometric instrument manufacturers (i.e. Baird Corporation, Hitachi

Instruments, Leeman Laboratories, Perkin-Elmer Corporation, Thermo Jarrell-Ash and


Varian).31


The earliest commercial generators were of the batch type, but the recent


trend is toward fully automated FIA systems.


All of these generators are designed to








formed simultaneously.


In fact,


Ward and coworkers were able to generate the hydrides


Sn and Te simultaneously using a continuous hydride


generator.41

Method of Detection for Hydride Generation


Several


types


atomic


emission


sources


have


been


used


with


hydride


generation.


AES),


ICP-AES, microwave induced plasma atomic emission spectrometry (MIP-


capacitively coupled microwave plasma atomic emission spectrometry (CMP-


AES) and direct current plasma atomic emission spectrometry (DCP-AES).31'42'43


these emission sources, the ICP has been the most often utilized.

For AAS detection, the flame is rarely used, since the typical acetylene/air flame


a background


absorption


40%


typical


hydrogen/air


flame


background absorption of 15%.44 The most common atomizers are the externally heated


quartz tube and the flame-in-tube atomizer.27


In both cases, the hydride is heated in the


presence of hydrogen and decomposes forming free atoms.


These atomizers have also


been


used


detection.29,43,45,46


Graphite


furnace


atomizers


have


been


used.27,47'48 The hydride is purged through a heated graphite tube (300-600 C) where

it is retained. The trapped hydride is subsequently atomized at temperatures higher than

2000 C. Table 1-2 lists some detection limits for comparison.


Interferences


While hydride generation can eliminate spectral interference, a variety of species


can interfere with the production, or the release of the hydride from the solution.


There






15

have been many publications concerning interference effects in hydride generation; for


examples see references 50-54.


categories,


The chemical interference can be broken into two


those which act in the solution phase and those which act after the hydride


is released from solution.


Table 1


Detection limits (ng/mL) by solution nebulization and hydride generation.


ELEMENT SAMPLE INTRODUCTION ICP- AAS Flame- ICP-
METHOD AES" AFS MSd
As Pneumatic nebulization 41 100b 100b 0.4
Continuous hydride generation 0.8 1 0.8c 0.0015
LN2 trap preconcentraion 0.02 0.02b 0. 1
Se Pneumatic nebulization 75 70b 40b 1
Cont. hydride generation 0.8 P1 0.5 0.03
LN2 trap preconcentraion 0.1 0.02b 0.06b
Sb Pneumatic nebulization 32 30b 50b 0.02
Cont. hydride generation 1 0.0003

LN2 trap preconcentraion 0.1 0.5b 0.1b


a Values From reference 28.
b Values From reference 43.
SValues From reference 49.
d Values From reference 31.


Dissolved noble metal


(Ag,


, Pt and Pd) can reduce the hydride generation


efficiencies of As and Se at concentrations less than 1 tg/mL.


Transition metals


such


as copper, cobalt and nickel,


suppress the production of hydrides at levels that depend


greatly on the concentration of NaBH4 used, the concentration of acid (generally HC1)








transition metals


such as Fe


Mn and Zn


, have little effect on hydride generation.


mechanisms of transition metal interference are not well


understood.


Smith56 has


suggested that easily reduced metals may be preferentially reduced.


In turn, the metal


precipitate may coprecipitate the analyte, the metal may absorb the hydride, the metal


catalytically


decompose the hydride,


or the metal may retard its release


from


solution.


Bye57 has suggested that the formation of highly reactive metal borides may


be responsible for the decomposition of the hydride before they reach the detection


device.


Also


, Aggett and Hayashi58 have studied the effect of transition metal ions on


arsine production.

specific chemical


They concluded that the interference is due to "the formation of a

species between arsenic and the interferant in a lower than normal


oxidation state," reference 58,


page 282.


The findings of Bax and coauthors59 suggest


that the effect of transition metal ions on hydrogen selenide production is due to the

decomposition of the hydride by metal borides formed by the action of NaBH4 on the


metal ion.


Whatever the case,


the interference of transition metals is less important


when the NaBH4 concentration is kept low and when the hydride is separated quickly

from the solution.

In addition to the transition metals, there are interelement interference between


hydride forming elements that may depend upc

after they have been separated from solution.


a mechanism affecting the hydrides


Interferences between hydride forming


elements are generally more serious in heated quartz tube and flame-in-tube atomizers,


than in


more energetic


atomizers.


D&iina54


studied


the effect of


arsenic and







17

and SbH3 can reduce the concentration of hydrogen free radicals (believed to be involved


in the mechanism of hydride decomposition)


atoms.


and accelerate the decay of free selenium


(Also, it is suggested that modification of the quartz surface is involved).


Electrolytic Hydride Generation

It is possible to form arsine, stibine, stanane, germane, hydrogen selenide and


hydrogen


reactions


telluride


their


electrolysis.


standard


Table


potentials,


some


Table


relevant


electrochemical


some of the


half wave


potentials observed for arsenic using dropping mercury electrode polarography.

Several groups have used electrolytic hydride generation as an alternative to the

chemical reduction of arsine in variations of the Gutzeit or Marsh tests, for example see


Rogers and Heron.61


Rigin62 was the first to use electrolytic hydride generation for


sample introduction in atomic spectrometry.


Rigin62 used a stirred electrolytic cell to produce arsine.


platinum wire was the catho<

NaOH was the electrolyte.


A flat spiral of smooth


A current density of 250 mA/cm2 was used, and


The hydride, and hydrogen,


were directly transferred to a


heated quartz tube atomizer.

to detect the arsenic. Rigin


Hollow cathode lamp excited atomic fluorescence was used


i reported that most of the arsine was generated within 45-


Using the peak area as the analytical signal, he reported a detection limit of


15 pg, or 3 pg/mL in a 5 mL sample, and his relative standard deviation was

The linear dynamic range was 0.2-200 ng (only 40 ng/mL).


Rigin62 tested 20 ions as possible interferants.


The results were impressive.










Table 1-3:


Standard potentials for several reactions.


HALF CELL REACTION STANDARD FORMAL
POTENTIAL (E'), V


H3AsO4

HAsO


+ 2H+

, + 3H-


As +


HAsO2

As +

HAsO-2

As +

SbO+


+ 6H+

3H110 -


+ 2e- =

S+ 3e

+ 3H+

+ 6e- =

- 3e" =


+ 2H20 +


3H20 +

+ 2H+


Sb +


HAsO2

= As+ +


+ 2H20

2H120


= AsH3


AsH3

AsH3 -


+ 2H20

+ 30H-


=As+


= AsH3


+ 3e-


3H+ -


+4 2H20 +


3H20 +

+ 4H+


SbOz-

Sb +

H2SeO3


Se +


Se +

Ge + 4H+


Pb +


Sn + 4H+


30H-


+ 30H-


= Sb + HzO

= SbH3

= Sb + 40H-


= SbH3


= Se


+ 4e

+ 2e


+ 30H-

+ 3H20


= H2Se


= Se2-


+ 4e-

+ 2e-

+ 4e

+ 2e-


Te +


= GeH4

= PbH2

= SnH4

= H2Te


0.560

0.240

-0.225

0.01o

-1.37

-0.68

-1.37

0.20

-0.51

0.64

-1.33

-0.74

-0.115

-0.670

-0.29

-1.507

-1.07

-0.740


All values from reference 60 unless noted.
a Calculated.








appreciable effect.


Co(II), Cu(II)and Ni(II) had no effect when 50 mg were added to


the electrolytic cell.


Unfortunately,


the volume of the cell


was not given,


so the


corresponding concentration cannot be calculated.


If 500 mL is assumed, 50 ug/mL of


Au, Ag, Hg, Sb and Se did not interfere with arsine generation!


Rigin has also reported


electrolytic stanane generation from 10% NaOH using a smooth lead cathode with AA

detection.63


Table 1-4:


Reported half wave potential for dropping mercury electrode polarography.


Polarizer SUPPORTING REACTION HALF WAVE
ELECTROLYTE (CHANGE IN POTENTIAL
OXIDATION STATE) (E)A,V
As(IV) 11.5 M HCL 5 0 not measured
0 -- -3 -0.26
As(III) 1 M H2SO4 3 0 -0.455
0 -3 -0.755

As(III) 0.5 M KOH 0 --3 -0.575


All values from reference 60.
a Reported versus the standard hydrogen electrode.
b Half wave potential for As(V) is less than 0 V.


Despite these excellent results by Rigin, no further work has been reported in the

literature concerning electrolytic hydride generation for sample introduction in atomic


spectrometry.


Because of the success of NaBH4


reduction,


the rush to improve the


associated methodologies and the desire to understand matrix effects, work toward other

methndz nf rerdntii n hn hben cere






20

Light Sources for Atomic Fluorescence Spectrometry

The light sources that have been used for AFS can be categorized as incoherent


and coherent (laser) sources.


Incoherent sources may be further categorized as line and


continuum, in reference to their spectral distributions.


Laser sources can be spectrally


narrow or spectrally broad in comparison to the spectral profile of the atomic transition.


Line


sources


are preferred


AFS


because


have


high


spectral


radiance and


produce less scatter than continuum sources.

The ideal light source for analytical AFS would have a spectral profile smaller

than the spectral profile of the atomic absorption line of interest (under the conditions


existing in the atom reservoir).


In addition, no light would be emitted by the perfect


source beyond the limits of the atomic absorption line's profile, thus maximizing the


selective excitation of the element of interest, and minimizing the laser scatter.


course, the center wavelength of the ideal source spectral profile would be tunable to


match any atomic absorption line wavelength.


The perfect source would have a high


spectral radiance, and there would be no fluctuations in that radiance.


more fanciful,


To become even


the ideal source could produce a spectrum of narrow lines matching


several atomic transitions (choice of the analyst),


for multielement analysis.


Incoherent Line Sources


Butcher"


coworkers


have


recently


published


an extensive


review


"conventional" sources for AFS. The most widely used sources for AFS are the HCL

and the electrodeless discharge lamp (EDL). These lamps are atomic line sources that







21

a low pressure direct current discharge or a radio-frequency or microwave excited

plasma.


In their basic form,


HCLs are direct current glow discharges utilizing a cup


shaped cathode made of, or coated with, the element of interest.


Multielement HCLs


are also


available.


Operated


typical


current


10-50


linewidths


approximately 0.002 nm2'1'9


which


times narrower than


typical


FAA


profile.6

arsenic


4,65


The typical HCL lamp has a radiance of 103 to


lamps have a radiance of approximately 5 x 10i5 W/cm2sr.


x 106 W/cm2 sr, and

66 Since fluorescence


intensity is proportional to excitation intensity, HCL lamps are often operated at higher


currents for AFS than those used for AAS.


However, if the current is increased too far,


larger


density


sputtered


atoms


cause


emission


profile


to become


broadened and self-reversed.


HCL operating lifetimes are also greatly reduced at high


operating currents.

Several methods have been developed to increase the excitation efficiency within


HCLs without increasing the rate of sputtering from the cathode.


The simplest method


involves using short (15-40 /s), high current (200


1000 mA)


pulses at frequencies of


200-400 Hz.67


Walsh68 and, later, Lowe69 developed "boosted" HCLs through the use


of an additional electrode.


In these lamps, a low current is used to provide atomization,


a higher current


(200-400


mA)


is passed


through a


third


electrode


to provide


increased excitation.


Lamps operating on this principle are available from Photron Pty,


Ltd., and produce atomic lines with up to ten times the intensity of ordinary lamps.









Several other methods have been developed to increase the intensity of HCLs.


They are


based on the coupling of the discharge with a microwave or radio frequency field.64


Electrodeless discharge lamps are usually more intense than common HCLs.


an EDL, a sealed tube containing the element, or a compound of the element, is placed


in an intense radio frequency or microwave field.


For certain elements (As, Se, Sb, Te


and Hg), the improved performance of EDLs make them preferable to HCLs.


EDLs


can be over an order of magnitude more intense than hollow cathode lamps."

Since both EDLs and HCLs are line sources, they can be used for AFS with or


without a dispersive light collection system (i. e., a monochromator).


A simple filter


can be used to select the fluorescence wavelength, or a solar blind photomultiplier tube


(PMT) can be used.64"1


The excellent light throughput makes this type of detection


preferred


in atomizers that do not produce much scatter and have


little background


emission.

Atmospheric plasmas such as the DCP and ICP have also been used as line


sources for AFS.


However, these sources have been rarely used and have never been


used for the detection of arsenic.70


Table


1-5 lists some AFS limits of detection for


arsenic and, for comparison, a more typical element, cadmium.

Dye Laser Sources

For many elements, dye lasers approach the ideal source for atomic fluorescence.


They can have an extremely high radiance and are spectrally narrow.


Pulsed dye lasers


can produce peak powers of megawatts or more and have spectral linewidths that can









system does not extend below 200 nm.


Therefore, there have been almost no reports


of LEAFS for arsenic detection.


Table


atomizers.


Detection limits for As and Cd using several light sources with several
Pneumatic solution nebulization was used.


METHOD


HCL-Flame-AFS

EDL-Flame-AFS


HCL


-ICP-AFS


DCP-flame-AFS


As DETECTION
LIMITS (ng/mL)


6000
320

200
22000


Cd DETECTION
LIMITS
(ng/mL)


4

0.6

0.05
4


Data from reference 60.


Only Leong, D'Silva and Fassel71

arsenic using a dye laser system. In their


have reported an analytical detection limit for


system, the output of a frequency doubled dye


laser was used to generate anti-Stokes stimulated Raman scatter in a liquid nitrogen


cooled Raman cell.


Anti-Stokes stimulated Raman scatter is a nonlinear effect used to


shift coherent light to shorter wavelengths.


They measured arsenic and several other


elements


using an ICP atomizer with


solution nebulization.


The limit of


detection


(LOD) for arsenic was 20 ng/mL,


which is not significantly better then ICP-AES.


improvement of only two times was seen when the pneumatic nebulizer was replaced


by an ultrasonic nebulizer.


However, the same authors observed an improvement of


over 50 times for mercury and phosphorus by LEAFS when ultrasonic nebulization was







24

and undesolvated droplets were the limiting noise for mercury and phosphorus but that


some other source of noise was limiting for arsenic.


Unfortunately, the authors did not


discuss limiting noise sources.


All dye laser systems are complex and difficult to operate and maintain;


system described by Fassel and coworkers would be more so.

for LEAF detection of arsenic is the ArF excimer laser. Selw


Another possible source


'yn7 has demonstrated the


use of a simple untuned Lumonics ArF

fluorescence in a low pressure RF sputl


excimer laser (al


tearing chamber.


t 193 nm) to excite arsenic

The ArF laser pumped the


193.7 nm line of arsenic, and fluorescence was observed at 245.6 nm.


The analytical


performance of the ArF laser was never evaluated because of "the extremely low ArF

power available at the RF sputtering cell."n

The Nature of ArF Laser Radiation


In an excimer laser, a mixture of a rare gas and a halogen are contained between


two parallel conduction plates.


For an ArF laser, these gases are argon and fluorine.


A high power discharge is formed between the plates in fast pulses.


excimers) are created within the discharge.


Excited dimers (or


A population inversion is automatic, since


these excimers are not stable in


ground energy


state.


The excited


state of an


excimer is bound, but the ground state is dissociative (or very weakly bound).


common excimer lasers types are ArF, KrF, XeC1, XeF and KrC1.


The most


A typical single


laser may be used as an ArF, KrF, XeCl or XeF laser by changing the fill gas mixture.


However, some time is required to


"passivate" the laser to each halogen.


Also, the









spectral wavelength peak is approximately


193 nm.


exact wavelength depends


slightly on the composition of the gas fill.73


typical


excimer


laser


beam


(including


lasers)


spectrally


broad,


typically bandwidth


= 1 nm.74'75


The spectral profile typically shows a complex fine


structure


associated


partial


interference


radiation


field


within


chamber.76


Since the typical excimer laser chamber is about 1 m long,


light requires


6-7 ns to make one round trip between the rear mirror and the output coupler.


laser pulses last only 10-20 ns, so it is not possible for the laser radiation to make more


than one full pass between the output coupler and the rear mirror.


coupler is not a reflector, but a MgCI2 or quartz plate.


In fact, the output


Therefore, the resonance cavity,


typical


many


lasers,


never


established.


However,


some


destructive


constructive interference do occur, and some fine spectral structure is superimposed on

a broad peak.

Since the ArF laser is so broad, and only the wing of the band reaches out to


193.7 nm,


very little of its power is available to excite arsenic fluorescence.


Most of


the unabsorbed energy passes through the atomizer with some being scattered by small

molecules and particles.


Argon


fluoride


lasers


are capable


exciting


fluorescence


several


small


molecules present in air, flames and plasmas.


Among these are O,, NO, CO and H2.


Laufer and coworkers" have studied the radiative processes in air excited by


laser.


an ArF


They found several strong oxygen fluorescence lines between 200 and 260 nm.









247.8 nm. Presumably, CO2 is photodecomposed to CO, which is further decomposed

to C.74'7 Scatter (fluorescence and Raman) from such small molecules is an important


disadvantage in a LEAFS source,


since this scatter will be a source of noise.


important to remember that far


more energy


is available in


output to be


scattered by small molecules than is available to excite arsenic.


Narrow


tunable


lasers,


such


as the


Lambda


Physik


model


EMG


available, but they are far more expensive and complex than the untuned excimer laser.


Like most tunable excimer lasers,


the Lambda Physik uses two discharge devices.74


the first discharge,


the tuned laser radiation is created,


and in


the second,


it is


amplified.


The ArF laser used in these studies,


the Questec Model 2100,


had a spectral


profile of


0.47 + 0.02 nm full width at half maximum (FWHM) and a peak center of


193.25 nm.


The pulse duration was 10-12 ns FWHM.


(See Chapters 2 and 6 for the


experimental


details.)


manufacturer


gives


following


specifications:


maximum pulse energy is 150 mJ/pulse;


although, only 80 mJ/pulse was sustainable at


high repetition rates;


the maximum repetition rate is 50 Hz;


the beam dimensions are typically 6-14 mm by


the pulse to pulse stability

20-23 mm (a rectangular


beam); and the beam divergence is 1.0 mr in the vertical direction and 3.0 mr in the

horizontal direction.













CHAPTER


INSTRUMENTATION USED FOR FAAS, ICP-AES AND AFS

Introduction


purpose


instrumentation


chapter

atomic s


used


describe


pectrometric


atomic


measurements


spectrometric


presented


dissertation. Minor modifications made for particular measurements will be described

in later chapters. The goals of the experiments concerning the electrolytic hydride


generator were to develop and characterize this new method of sample introduction.

Because the analytical figures of merit measured depended on both the generator and the

spectrometers, it is important to understand the limitations of the spectrometers used.

Three instruments were constructed in the laboratory chiefly of commercially

available components: a flame atomic absorption spectrometer (FAA), an inductively


coupled plasma atomic emission (ICP-AE) spectrometer,


fluorescence spectrometer.


and a laser excited atomic


The emission instrument and the fluorescence instrument


resided on the same experimental setup and shared much of the same equipment.


the sake of clarity, they will be discussed below as if they were completely separate

instruments.

The atomic sources (the flames or the ICP) and optical systems will be described


first,


measurement electronics


will be described


in a separate section









be compared to commercially available instruments.


Because the spectrometers


were used at wavelengths near 200 nm, this comparison will focus on the use of these


instruments at short wavelengths.


Since there has never been a successful commercial


LEAF spectrometer, no comparison can be made between the laboratory constructed

LEAF spectrometer and a commercial counterpart.

Atomizers. Emission Sources and Optics

The FAA Spectrometer


The FAA spectrometer was a simple single beam spectrometer.


Figure 2-1 is


a schematic diagram of the instrument, and Table 2-1 lists some of the equipment used.


All the lenses were made of fused quartz (silicon glass). The light from a HCL was

focused, modulated by a mechanical chopper, and then collimated. The collimated light


was focused by a 15 cm focal length lens into the center of the flame and then refocused


by another 15 cm focal length lens.


An iris was used to block the outside portion of the


beam to match the f-number of the monochromator.


Next, the beam passed through a


filter (when used) and fell onto the entrance slit of the monochromator.


image of the lamp emission was formed on the entrance slit.


A one-to-one


The entire optical train


was carefully aligned to ensure that the system had a maximum throughput.


The burner was a premix type burner with a 5 cm slot burner head.


line was attached to the fuel inlet.


A hydrogen


The auxiliary air inlet of the mixing chamber was


connected


to the


outlet of


hydride


generator,


and all


the combustion


was


introduced through the nebulizer-air port.


The nebulizer was a common pneumatic type,




































a,

U




Cd,




Cd



S






C3)
I-


Cd

Uc
S

4-i
Ca
C
.9-
C)
0r











































































































4-


S



* S
* S
* S
* S
* S


.
*
S
.
.tEflh)
S.
-S
.5
S.
S.
S.
S.

S
"5- -
5*


S.
#5
*
S
*
*
S
*
*
*
S
.


S
S
A
S
*
.
.
.
S
.
A
S
S
S
A
A
A
A
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
(I) S.
S
S
S
C .5


100


* S
* S
* S
* A
* S
*
* S



* A
* S
* S
* S
* S
- S


Lc



D.o




a0


0-c


i


i


---~






31

and the sample uptake tube of the nebulizer was blocked when the nebulizer was not in


use.


Both


hydrogen and air


lines were


regulated by two stage regulators and


metered by needle valves in the base of two separate rotameters.


Table 2-1:


Equipment used for FAAS measurements.


EQUIPMENT DESCRIPTION / MANUFACTURER AND MODEL
Atomic line source Boosted hollow cathode lamps, Photron Pty. Ltd.,
Superlamp
Chopper and control Stanford Research Systems, model SR540
Premix burner Perkin Elmer, part no. 0040-0146

Nebulizer Perkin Elmer, part no. 0303-0352

Interference filter

For As and Se Acton Research Corp., 193-N-1D (195.0 peak
measurements wavelength) 20 nm bandwidth

For Sb measurements none

Monochromator Thermo Jarrell Ash (0.5 M focal length, 1200 groves
mm' grating, F-number 8.6), model 82020
High voltage supply Bertran Associates Inc., Model 301 (-1000 V)

Photomultiplier Hamamatsu, model R 1547 (1.3 x 106 gain, 1 nA dark
current


The ICP-AE SDectrometer


Figure


a schematic


equipment


used


ICP-AE


spectrometer, and Table


lists some of the equipment used.


The emitted light


from the 40 MHz argon ICP was collimated by a fused quartz planoconvex lens,






































C.)

E

(I-
C*
C



0






C.)






E
0
Cd3

Cu











0T)
C



r-
H











Ur)

0)


Cu
Cr
a.
0


I I H
-3 2a


4.
9,
S


S


-4


0
I
C

O a- e0

OJ)Q)
VC a
as
C 0)
"Oc-
^-00 X

*^ 8C
'4-,U


* a
* S


III II.
0
S


I
Sb


C (4
CU)
0L.


E
E
-oE
ro


:3 ***** .***.OleO********U**Se


S.





EL


11~111111111111111
I
;;IiI








of the monochromator.


A second planoconvex lens focused the light onto the entrance


of the monochromator where it was focused into a one-to-one image of the ICP.


Optionally, one or more filters could be placed in front of the entrance slit to filter the


ICP emission.


The plasma was viewed through the observation window in the front


access door of the ICP housing rather than


the window in


the side of the housing


normally used for this purpose.


(This geometry allowed the laser beam to enter at a 90


angle to the collection optics during the fluorescence experiments.)


The lenses were


mounted on to a kinematic base so that they could be easily removed and replaced when

the plasma access door was opened.


Table


Equipment used for ICP-AES measurements.


EQUIPMENT DESCRIPTION / MANUFACTURER AND MODEL

Torch Precision Glass Blowing of Colorado (standard)

Nebulizer Precision Glass Blowing of Colorado (concentric)

Monochromator Spex Industries Inc. (0.5 M focal length, 1200
grooves mma' grating, F-number 6.9), model 1870

Photomultiplier Hamamatsu, R 166

High voltage supply Bertran Associates Inc., Model 301 (-1000 V)

Radio frequency Baird Corp., 300-D
Generator (and ICP
housing etc.)


The argon gas sources,


coolant and plasma/nebulizer,


were regulated by two


stage


regulators.


The gas


meters and


control


valves


were supplied by


the plasma


rp nI t rn- ft-. a ,,-an-,1nnl ,-,n 1an TC'TD f^.-^l 'fo'.. ^,.n,,ar-fA 4f a l,,, a ,,- C tM70


rurnnlrEnnlrl*ne








generator, depending on the experiment being performed.


The ICP torch and nebulizer


were manufactured by Precision Glass Blowing of Colorado, and both were of the type


most


commonly


used


ICP-AES78


The


torch


was


developed


Greenfield


, which supports an annular plasma.


The nebulizer was a concentric glass


nebulizer similar to the type developed by Meinhard.so


Solutions were not pumped into


the nebulizer, but were drawn in by the vacuum created by the Bernoulli effect in the

nebulizer.

The ICP-LEAF Soectrometer


The system


used for atomic


fluorescence measurements,


in the


ICP and the


flames, is shown in Figure


and some of the equipment used is listed in Table 2-3.


For the flame measurements, a flame was in the position occupied by the ICP in Figure


Monochromator #1 was used for the early experiments and was later replaced by,


the higher resolution, monochromator #2.


Monochromator #1


was placed on its side,


so that the entrance slit was horizontal, parallel to the ground.


This allowed a larger


area of the fluorescence "stripe," formed as the laser passed through the atomizer, to


be imaged along the monochromator entrance slit.


In contrast, monochromator #2 was


used in the conventional vertical entrance slit attitude.

A 10 mm2 area of the edge of a quartz microscope slide was placed in the path


one corner


laser


beam


angle


to the


laser


beam.


Approximately


of the edge of the beam was thus reflected toward the photodiode


trigger.


The laser beam was then focused through a 50 cm focal length, fused


silica,














































4-.


ci














a)




S







C)

V1
4)


















































. .,,
* -


*m
(0
c
0
a--
a.

o
* 0




A A


Q. -
C0-
Or
OEU

C E


Sc~


o c
LO aI


LOr U)
rc~C
03)0.)
Vc -


COC:
cao0
14-0


--. **
********* "*** "* ** -***, .. *
... - -v ""**1" ".- .,


0)
E

(U
'4













Table 2-3:


Equipment used for LEAFS measurements.


EQUIPMENT DESCRIPTION / MANUFACTURER AND
MODEL
Excimer laser Questec (50 Hz. maximum pulse rate, 150 mJ
maximum pulse energy in ArF mode) model
2100
Monochromator #1 Spex Industries Inc. (0.33M focal length, 1200
grooves mm-1 grating, F-number 5.9), model
1681
Monochromator #2 Spex Industries Inc. (0.5 M focal length, 1200
grooves mm-' grating, F-number 6.9), model
1870
Photodiode trigger EG&G Photon Devices (<1 ns risetime) FND-
100Q
Arsenic electrodeless Imaging and Sensing Technology model WL-
discharge lamp 40011
Photomultiplier Hamamatsu, R 166
High voltage supply Bertran Associates Inc., Model 301 (-1000 V)
ICP torch Precision Glass Blowing of Colorado (standard)
ICP nebulizer Precision Glass Blowing of Colorado (concentric)

Radio frequency generator Baird Corp., 3000-D
(and ICP housing etc.)
Premix burner Perkin Elmer, part no. 0040-016
Nebulizer Perkin Elmer, part no. 0303-0352









lens.


distance


between


atomizer


was


adjusted


during


experiments to achieve various laser beam cross sectional areas in the atomizer.


fluorescence was collected at a right angle to the laser beam by a pair of 15 cm focal


length,


fused


quartz,


lenses.


was


used


to match


F-number


spectrometer, and the collected fluorescence was imaged on the spectrometer entrance


slit to form a one-to-one image.


Optionally, one or more filters could be placed in front


of the monochromator entrance slits.

The arsenic electrodeless discharge lamp was used to calibrate the wavelength


of the spectrometer; it was left off during most of the experiments.


splitter was used only when the EDL was used.


The 90/10 beam


It allowed 90 % of the fluorescence, or


laser scatter,


to enter the spectrometer at the same time as


10% of the EDL optical


emission.

The ICP used for the fluorescence experiments was the same as that used for the


emission experiments.


The burner and nebulizer used for the LEAF spectrometer were


of an identical type to those used in the FAA spectrometer, except that a 3 cm slot


burner head was used instead of a 5 cm slot burner head.


The burner head was placed


so that the laser beam passed through the long axis of the flame.

Electronics Systems


Electronics for FAAS


The electronic system used for the FAA measurements is shown in Figure 2-4,


and Table 2-4 lists some of the components used.


The HCL was modulated to reduce





























Figure 2-4:


The electronic system used for the atomic absorption measurements.


















Current to
voltage amplifier


Reference
timing signal


Optica
control


modulator
(chopper)


Lock-in


amp


ADC


Chart recorder


Computer








pass,


filter


was


used


to amplify


signal


from


PMT


before phase


sensitive


rectification by a lock-in amplifier.


The optical modulator,


or chopper,


provided a


reference signal to the lock-in.


with a


The current-to-voltage amplifier was generally operated


2 kHz electronic band pass, the chopper modulated the HCL beam at 200 Hz,


and the time constant of the lock-in amplifier's low pass filter was set to 10% of total


time used for each measurement.


The output of the lock-in was sent to an analog-to-


digital converter (ADC) and recorded by a computer and a strip chart recorder.


Table 2-4: Electronic equipment used for FAA measurements.


EQUIPMENT DESCRIPTION / MANUFACTURER AND MODEL
Current amplifier Keithley, model 427
Lock in amplifier EG&G Princeton Applied Research, model 5101
Analog to digital Stanford Research Systems, model SR245
converter

Computer Northgate 286


When a single beam spectrometer is used to measure absorption, the ratio of the

HCL atomic line intensity to the intensity of the line transmitted through the flame


cannot be directly measured.

intensity of the HCL. Eaci


Therefore, it is necessary to decrease the variation in the


i lamp used in these experiments was always allowed to


warm up for 30-60 min, or until the average intensity of lamp emission changed by less


than 1% in 10 minutes.


The lamps were powered by a current regulated high voltage









Electronics for ICP-AES


electronic


system


used


ICP-AE


measurements


shown


Figure


All of the equipment listed in Table 2-4 was also used for the ICP-AE


spectrometer, except for the lock-in amplifier.


The current from the PMT (a solar blind


model) was amplified by a current-to-voltage amplifier with a built in, low pass, filter.

A laboratory constructed voltage amplifier was used to adjust the output range of the

current amplifier (-1 to 1 V) so that it more closely matched the input range of the ADC


(-12 to 12 V).


The output of the current-to-voltage amplifier could be monitored either


a strip


chart


recorder


or a


computer using


an ADC.


Neutral


density


filters,


calibrated at 193.7 nm and 228.8 nm, were used to extend the region of linear response

of the instrument.

The time constant of the detection electronics was adjusted to approximately 10%


of the measurement time used for recording steady state signals.


For transient signal


measurement, a time constant of no more than 1/7t of the peak width of the transient,


or pulse, is required to accurately reproduce the pulse."8 Therefc

of the electronics was set for 1/10" of the fastest transient signal'


)re, the time constant


peak width.


Electronics for LEAFS


Figure 2-6 is a schematic of


spectrometer


's electronic measurement


system.


Much


equipment


used


was


identical


to that


listed


Table


However, a current signal from the PMT was converted to a voltage by a laboratory


designed transimpedance amplifier, rather than the Keithly current amplifier.


(A circuit





























Figure 2-5:


The electronic system used for the atomic emission measurements.




















Current to
voltage amplifier




Optional precision
voltage amp. gain = 10


ADC


Chart recorder


Computer





























Figure 2-6:


The electronic system used for the atomic fluorescence measurements.
corner inset is a circuit diagram for the transimpedance amplifier.















Optional voltage Amp (5 or 10X)

X 10 Transimpedance Amp.

***

PMT





Low Pass Filter ->


S70 V Battery


I<-- Trigger Photodiode


Boxcar


Avg.


15 pF


10okn


OP-07 (50 MHz)


*


ADC


Computer


Lock-in
amplifier


EDL


Power


Recorder


Function
Generator








diagram of this amplifier is shown in the inset in figure


average (Stanford Research Systems,


In addition, a boxcar


model 255) was used.


A fast voltage preamplifier was sometimes used to further amplify the signal


before its measurement by the boxcar average.


The fast voltage preamplifier was a


Comlinear Corporation (model CLC100) DC-500 MHz amplifier with a gain of 20dB


(10 times).


However, since the output resistance of the voltage amplifier and the input


resistance of the boxcar were each 50 (0


, a voltage divider was created,


and the effective


gain was only


10dB.


Whether the voltage preamplifier was used or not, a 50 (f series resistor was


used


to match the input impedance of the boxcar average to the impedance of the


coaxial cable that connected the various components.


provide by a photodiode.


The boxcar's trigger signal was


The signal from the boxcar was recorded on both a strip chart


recorder and a computer via an ADC.


Again,


calibrated neutral density filters were


used to extend the linear response range of the instrument.


The electronic


system


included


means


to record a


wavelength


calibration


spectrum simultaneously with a laser specular scatter or fluorescence spectrum.


atomic emission lines from the EDL were used.


Arsenic


This was particularly necessary during


experiments designed to measure the laser spectral profile (see Chapter 6.)


The EDL


emission spectrum was recorded simultaneously with the laser scatter by the computer


and strip chart recorder.


Since the EDL was modulated at a frequency of 500-1000 Hz,


the high pass filter (10 kHz cutoff frequency) built into the boxcar average prevented






49

emission was measured by a lock-in amplifier connected to the transimpedance amplifier


via a low pass filter (1 kHz


cutoff frequency) which prevented the laser induced signal


from reaching the lock-in device.

The excimer laser not only produced ultraviolet light but also radio frequency


(RF) pulses and X-rays with each laser pulse.

the lead lining of the laser housing. Howeve


The X-rays are effectively contained by


r, the pulses of radio frequency (MHz to


GHz) electromagnetic radiation presented a considerable difficulty in the measurement


of small fluorescence signals.


picked up by the PMT.


Figure


The PMT


shows a typical radio frequency noise pulse


was in a shielded housing and biased with -1000 V


A shielded


, 50 0,


coaxial


cable


used


to connect


PMT


to an


oscilloscope


(Hewlett


Packard model 54503A).


The oscilloscope's


internal 50 0 input resistance served as a


load resistor.


(The band width of the oscilloscope prevented the measurement of higher


frequencies.)

Figure 2-8 shows the average discrete Fourier transform of four, excimer laser


generated, RF pulses as detected at the PMT.


The frequency content of the RF pulses


generally increased with increasing frequency from


100 kHz to


1 GHz.


Among the


strong


bands


frequencies


several


near


100 MHz


(period


10 ns).


Unfortunately the atomic fluorescence pulses had a FWHM of approximately 12-15 ns,

and the RF pulses began before the fluorescence signal started and ended long after the


fluorescence signal had decayed.


Therefore, it was difficult to effectively eliminate this





























$I
*









I-l



C)

I-J









.2




-o

I-










Ur,

S!







t*



























































0
0
0








o


S






































a)

I-
'4
0,

* -i


I-

0)t
4-
0

I-







E4





4-




a)






so











































(sl!un AieJiqJe) epn)


due esIoN








filter (such as a boxcar).


Since the exact phase and intensity of each RF pulse was


random, the boxcar output showed random fluctuations (noise) that interfered with the

measurement of small fluorescence signals.


Several


steps


were


taken


to reduce


effect


noise


on the


measurement system.


The laser was encased with a copper mesh.


Several shielded


PMT housings were compared, and a model PR1402RF (Products for Research, Inc.)


housing


was chosen


for the experiments described in


this dissertation.


PMT


housing was electrically isolated from the monochromator's mounting flange by a rubber


gasket.


Every


cable


kept as


short as


possible,


a special


transimpedance


amplifier was built in a shielded box and was connected directly (no cable) to the PMT


output.


As a result, the system'


response to the RF noise was reduced to a level below


that produced by a single photoelectron.

Although the fluorescence signal consisted of pulses of light lasting not much

longer than the laser pulses (12-15 ns FWHM), the transimpedance amplifier stretched


these pulses to a FWHM of approximately 250 ns.


Figure 2-9 shows the shape of the


average of eight fluorescence signal pulses recorded with a simple 50 0 load resistor and


with the transimpedance amplifier.


The speed of the transimpedance amplifier had to


be compromised to prevent ringing due to


"gain peaking."


Gain peaking can occur


when a fast transimpedance amplifier is connected to a current source that has a high


output capacitance.82


An unstable resonator results if the product of output capacitance


output resistance of the current source


PMT)


exceeds


the product


of the










U










CO






"1*



U

01








C)
-C
-v

'-4
0






S-r

C)

CU
0


C









I-
















































[L.
U-


C
0 E
CJc
CM E


(pw) pn!ldIuwe


(ALw) apnl!ldwue leu6!s


'~-v'


SI


3 I








amplifier.


A feedback capacitance of 15 nF was experimentally determined to be the


minimum capacitance that prevented ringing.


In theory, it is always better not to stretch


a transient signal in the presence of white noise, since the signal intensity will decrease.

Also, if the boxcar gate width is increased to capture the same area of each peak, the


larger gate width will include more noise.


However, it was empirically confirmed that


the larger gate width did not degrade the signal to noise ratio, likely due to the special


nature of the RF noise.


The larger gate width reduced the bandwidth of the boxcar,


effectively filtering the higher frequencies contained in the noise spectrum as much as,

or more than, the signal.

Performance and Comoarison to Commercial Instrumentation


FAA Spectrometers


Generally,


FAA


spectrometer


manufacturer's publish


detection


limits and/or


characteristic concentrations for a variety of elements.


The characteristic concentration


is the concentration of the element that gives an absorbance of 0.0044.

can be calculated as 0.0044 divided by the characteristic concentration.


direct measure of the instrument'


the sensitivity times the detection


The sensitivity

Perhaps a more


capability is the minimum detectable absorbance, or

limit. This criterion is still a function of the flame


type, the HCL lamp type, and the wavelength, but it is less affected by the nebulizer


efficiency and atomization efficiency.


Another important performance criterion for FAA


spectrometers is the minimum transmittance for linear response, or the minimum optical


transmittance that produces a linear response from the instrument.


From this value and








sensitivity,


upper


limit of


a linear region


the calibration


curve can


estimated, if the spectrometer response is the limiting factor.

The minimum transmittance for linear response was estimated by using calibrated


neutral density filters.


An arsenic HCL was used, and the arsenic


193.7 nm line was


isolated by the monochromator (spectral band pass 0.1 nm).


The flame was not used.


The instrument responded linearly down to a transmittance of 0.78% (or an absorbance


of 2.1).


Stray radiation was the most likely cause of the nonlinear response at high


absorbance


values.


Typically,


commercial


FAA


spectrometer


can measure


transmittance down to 1% (an absorbance of 2).2

The minimum detectable absorbance depends greatly on the relative strength of


several noise sources that may be dominant at low absorbance values.


below 210


At wavelengths


- 220 nm, flame absorption flicker noise is usually the limiting noise source.


The instrument used in this work had minimum measurable absorbance of approximately


0.003 for both the 193.7 nm arsenic line and the 196.8 nm selenium line.


Both values


were measured using pneumatic nebulization of an arsenic or a selenium solution into


a H2/air


flame.


Ingle


Crouch2


typical


characteristic


concentrations


concentration that gives an absorbance of 0.0044) and detection limits for FAAS.


From


these values, a minimum detectable absorbance of approximately 0.001 can be calculated


elements


with


absorption


lines


below


nm.


From


Varian


commercial


literature83


, minimum detectable absorbance of 0.0025 (for arsenic at 193.7 nm) and


0.002 (for selenium at 196.8 nm) can be calculated.


At longer wavelengths, both the









laboratory


constructed


commercial


instruments


measure


smaller


absorbance.

The simple single beam instrument described here lacked two important feature


shared


most


modem


commercial instruments.


Modem


commercial


instruments


generally employ some type of background correction technique.


Background correction


reduces the effect of background absorption and allows the measurement of a lower


absorbance when the background absorption is high.8485


Flame background absorption


is significant at wavelengths shorter than 210-220 nm, even in a H2/air flame.


Another


important feature of most commercial FAA spectrometers is the incorporation of a


reference light beam.


In such an instrument (a spectrophotometer), the ratio of the two


beams


measured.


Therefore,


changes


HCL


lamp


intensity


not affect the


magnitude of the signal.


The absence of a reference beam to account for changes in the


lamp intensity decreased the precision of the measurements and necessitated constant

recalibration of the 100% T level.

ICP-AE Spectrometry


The comparison


emission


spectrometers


is more difficult than absorption


spectrometers since the only common figures of merit given for commercial instruments


are the limits of detection for various elements.


A major difference between the ICP-


spectrometer


used


work


a state


sequential,


ICP-AE


spectrometer, such as the Perkin-Elmer Plasma 200086 or the Varian Liberty


11087


commercial


spectrometers


benefit


from


larger,


higher


throughput






60

What is more important is that these commercial instruments take a number of


steps to stabilize the plasma.


For example,


the spray chamber of the Plasma 2000 is


cooled


, the gas flow rates are more accurately controlled by mass flow meters, and a


solid state RF supply with automatic impedance matching is used.


These improvements


allow


commercial


instruments


to detect


most


elements


(including


arsenic


selenium) at lower concentrations than were possible using the laboratory constructed


spectrometer.


Using pneumatic solution nebulization and the laboratory constructed


spectrometer,


the arsenic detection limit was approximately


70-90 ng/mL.


l0 s


integration time and three standard deviations (3 a) were used to calculate this LOD.


The Plasma 200's arsenic detection limit is 20 ng/mL.


, and the Liberty 110'


arsenic


detection limit is 12 ng/mL,


based on 10


s integration time and the same, 3 a, definition


of LOD.83













CHAPTER 3
DESIGN OF THE CONTINUOUS ELECTROLYTIC HYDRIDE GENERATOR

Introduction


A canon of all design is that the optimal compromise between performance,


convenience and cost defines the best design.


industry in the 1980's).


(An obvious exception is the defense


In this case, the budget, mechanical abilities of the author, time


and initial ignorance were the limiting factors.


much was learned.


This design is a prototype, from which


That there could be many improvements is obvious, and several will


be suggested in the seventh and final chapter of this dissertation.

The electrolytic hydride generator was designed and constructed at the University


of Florida.


The design was intended for ease of construction.


Much of the actual


fabrication was done in the "student machine shop," and the more difficult machining


was carried out by the UF Chemistry Department'


machine shop staff.


Appendix A


contains shop drawings for the various parts that must be machined.


Rigid acrylic


plastic (polymethylmethacrylate) was chosen for several parts, primarily for its inertness


in strong acids and alkali, excellent machineability and transparency.


Extruded PTFE


polytetrafluoroethylenee) plastic was used for all of the other parts that came in contact


with the solutions.


Chemical resistance was the principal reason for choosing PTFE.


generator


consisted


a cathode


flow


inside


a cylindrical


anode






62

electrical conduction between the cathode in the flow cell and the anode in the anode


compartment.


The solution to be analyzed was pumped through the cathode cell where


it contacted a coil of wire functioning as the cathode, and the analyte (As, Se or Sb) was


converted to the hydride. A large potential difference (over 20 V) was maintained

between the electrodes of the cell. Obviously, the principal reaction was the electrolytic


decomposition of water into oxygen and hydrogen gas.


The solution and the gases


(hydrogen and hydrides) produced at the cathode flowed up through the cathode cell and

into a "sparge" (phase separator) chamber where they were separated from the aqueous


phase with the aiding flow of a "carrier gas" (argon).

the solution stream below the sparge chamber. The


The carrier gas was mixed with


vaste liquid was drained out from


the side of the sparge chamber, and the gases were directed through a length of tubing


to either the burner or torch.


The oxygen gas generated at the anode was regulated to


maintain a higher pressure within the anode chamber than that within the cathode cell.

This created a net flow of solution through the glass frit from the anode compartment

and into the cathode cell, greatly reducing the possibility of the anode solution becoming

contaminated with the contents of the cathode cell.

A Detailed Walk Through and Around the Generator

Referring to Figure 3-1, the cathode flow cell was a cylindrical channel cut into


a solid block of extruded PTFE plastic.


A slot was cut into one side of the channel (A).


The chamber was shaped with a round flange (B) which was used to attach it to the


linking flange described below.


A cylindrical extension (C) above the flange had six






























Figure 3-1:


The cathode flow cell (not to scale).





























































































Front view






*
9
9
9
*
.
*
9


9
*



I 9


* 9
* 9
* 9
**

* 9

* 9


* S
*t
* .
* .
* .
* 9
* 9
* 9
* 9

* *


* 9


N
S

If
0**



a
S
a
a
a
S
a
a
a
S
S
S
a
a
S

SS
be.
S
S
S
a
S


43.


Side cross section


N






65

The solution to be analyzed was pumped into the bottom of the flow cell through a

fitting (E) that allowed the chamber to be connected to 1/8" PTFE tubing via a 1/4" to


1/8" PTFE compression union.


(All the tubing fittings and tubing unions were made


of PTFE unless otherwise note

(SwageLok Co., Solon, Ohio).


SwageLok


brand compression fittings were used


The entire chamber was -


10.5 cm long, while the slot


cut in the channel wall was t 4.3 cm long and 0.4 cm wide.


The channel was


- 0.5


cm in diameter and was lined with a coil of wire forming the cathode.


The cathode coil


(F) extended above and below the slot in the channel wall and was = 6 cm long.


the coil,


After


the wire continued up and out of the cathode flow cell and into the sparge


chamber which is described below.

An indentation was cut into the cathode flow cell to accept a porous fitted glass


plate (G).


This plate was cut from a fitted glass disc (supplied by Kontes,


Vineland,


nominal pore size 10-15 gm).


A thin layer of silicon sealant was spread along the


edge of the plate which pressed against the cathode flow cell to form a gasket.


plate was held in


place with a


TFE clamping plate (H)


which


was attached to the


cathode flow cell by four nylon screws (not shown).


The platinum wire mesh anode (I)


was held in place between the clamping plate and the porous plate.


The linking flange is illustrated in Figure 3-2.


This flange provided a means of


forming gas tight seals between the other parts of the reactor.


An o-ring,


which sat in


groove A, provided a gas tight seal between the linking flange and the flange on the


cathode flow cell.


The flow cell was held in place by six PTFE screws.


The sparge






























Figure 3-2:


The PTFE linking flange (not to scale).











Cross section


Bottom view


,I








a gas tight seal.


Two sets of six threaded holes (C and D) accepted the twelve nylon


screws which held the cathode flow cell and sparge chamber tightly in place.


linking flange was attached to the anode chamber by six additional screws.


Again, an


o-ring,


in groove E,


provided a gas tight seal between


the anode chamber and the


linking flange.


Sparge gas entered through a compression connection in port F in the


linking flange and was forced into the small holes in the cylindrical extension of the

cathode flow cell.


Figure 3-3 illustrates the assembled generator.

sparge chamber (B) were made of acrylic. The solution


The anode chamber (A) and


1 to be analyzed flowed through


the wall of the anode chamber in 1/8" PTFE plastic tubing (C) which was connected to


the bottom of the cathode flow cell (D) by a reducing compression union (E).


anode mesh (F) was electrically connected to a platinum wire that passed through a


compression fitting (G) in the anode chamber wall.


The oxygen produced at the anode


left the chamber through a port (H) near the top of the anode chamber.


Fresh anode


electrolyte solution could be pumped into the anode chamber through a connector (I).

The solution in the anode chamber was stirred during operation by the magnetic stirring


bar (J).


The anode chamber was 17.8 cm long, the outside diameter was 8.9 cm, and


the inside diameter was 6.35 cm.


The sparge chamber was 8.9 cm long and 3.6 cm in outside diameter.


inside diameter was 2.3 cm.


The compression connection that forms the drain of the


sparge chamber (K) was about 1 cm from the base of the sparge chamber.


A length of






























Figure 3-3:


The assembled generator (not to scale).




























































Si
4
S
.~ a.
ma a a St,0

a
a
a
a
a
n
La a


a
a
a
0% a
,..*.; *S*.eS. 0a
a
S S
*0 *
*ma
.0
S
a
0 5
a
a
*. a
0.
a. *


a.
a. %
*0*
*a. ............................ ~!
OS. 0*05.... *.* abates


a. S
.0.
S
S
a
a
S
S
S
a
*
S
S
S
S
S
a
S
a
a
S
a
a
a
a
a
S
a
S
S
0
S
a
S
a
a
S
-S
at .*
S
,st.
S St
III 't.
at a..
etarn
a.' ~ S.
* A .
a S


*
*
a
a
a 5
55* 5*


\ "w













**a
AS
*
S
S
a
S
S
S


S
S
S
S
S




*S





*S
* a

..****..
a#



*

:. *,
a *




**...** ".
S


a...
S m
*I S a
a~m
*l Sl S




gl 11j ga.*tg L 4pOb
*~p


.a*Saaa..Sb
.~ .00*000 .... .asSeeet0eeaSasaa.~~ -
*~ .................. 0 S ...S~ Ste...


-.***** "****'*" **.,


a

S
S
S
S
S
S
S
S
S
S
S
S
S
a
S
S
S
S
S
S
a
a
a
a
a
a.


~
a.







71

1/8" plastic tubing (L) formed a pressure equalizing arm between the sparge chamber


and the drain tubing (M).

the sparge chamber. Th


A stainless steel compression T-union (N) was attached to


e hydrogen and hydride gases passed out of the generator


through 1/4" TFE plastic tubing (0).


The cathode coil lead passed out of the sparge


chamber through a compression connector (P), and the sparge gas entered through a

compression connection (Q) in the linking flange (R).


Figure


illustrates


assembled


generator


attached


to the


various


additional devices and parts that comprised the complete apparatus.


The oxygen formed


at the anode was forced through a length of plastic tubing (B) and under several inches


of liquid in a cylinder (C).


The pressure in


the anode chamber was regulated by


adjusting the depth of the tubing in this cylinder (the vent of this cylinder was attached


to the laboratory exhaust system).


A d.c.


power supply


(D) was connected to the


electrode leads by two wires (E and F).


The solution being analyzed, or blank electrolyte solution,


the generator by a peristaltic tubing pump (G),


was pumped through


while the second pump (H) was used to


keep the anode chamber filled with clean electrolyte solution.

(I) was used to impel the stirring bar in the anode chamber.


A magnetic stirring motor

A loop of tubing (J) on the


liquid drain line prevented the gases from escaping out the drain port.


The drain bottle


(not shown) was carefully vented to the laboratory exhaust system to prevent the release


of toxic hydride gases into the laboratory.


The carter gas flow rate was controlled by


a needle valve in the flow meter (K).





















































































'-4
0









0.)
































0
--0


L(


c~








Voltage Regulation


The d.c.


power supply used was a Hewlett Packard Model 6268A, operated in


the voltage regulated mode.


Since this electrolytic cell contained no reference electrode,


only cell potential could be regulated and not the actual cathode potential.


When the


current regulation


mode of the power supply was used at these high


voltages,


regulation circuit fluctuated greatly.


It is assumed that the hydrogen generated within


the confined cathode flow cell erratically blanketed part of the cathode surface from the


solution.


This may be aggravated by the use of PTFE for the cathode flow cell since


it has a hydrophobic surface and gas bubbles tend to stick to the walls of the cell.


a result


, the actual working surface area of the cathode changed rapidly, and the current


regulation circuit could not respond. In any

production (as observed by FAAS and ICP-

potential is regulated rather than the current.


' event, the rate of hydride and hydrogen

AES) varied less with time when the cell

The total cell potential could be monitored


at one of the ADC channels simultaneously with the atomic spectrometric data.

current could also be monitored.

The Cell Voltage and Overpotential

Contributions to the Cell Potential


There are several types of voltage drops across an electrolytic cell.


The total cell


potential for water electrolysis may be given as,


=Ed


+ 71c


it,


+IRs


+ IRb


+ IRr~~


(3-1)






75

the conditions used (temperature, acid or base concentration, etc.) used, wr and fla are


the overpotentials of the cathode and anode respectively,


I is the current,


R, is the


electrical resistance of the solution, Rb is a correction factor (with units of Ohms) which

accounts for the effective loss of electrode surface area because of gas bubbles, and the

increase in solution resistance when small gas bubbles are dispersed into the solution.


The resistance of the metal parts of the circuit is denoted by Ra.


The value of


can be estimated using the Nernst equation for the hydrogen half cell reaction, or maybe

found in tables,88 and is about 1.15 V.

Resistance Terms


The IRm1e term can be ignored in most cases.


The resistance of the solution, R,,


can be estimated as,


1

K: (3-2)


where K is the conductivity of the solution, and C is the cell constant, or the distance


between the anode and


electrodes.


the cathode divided by the average surface area of the two


Solution conductivities can be found in tables for most common electrolytes.


However


in this


it is unlikely


equation


would


give


an adequate


approximation of the true R,.


(Because the electrodes are in very close proximity, but


are separated by a porous glass frit,


the ions


must travel further than the apparent


distance and must migrate through narrow channels in the barrier.)


The bubble effect


factor, Rb ,is difficult to estimate theoretically without specific experiments to measure









significance of these resistance terms is that they create heat within the cell (power


P2R).


therefore


, important


that they do not


become too large.


To reduce the


resistance of the cell, the

electrodes should be small


. electrode areas should be large,


the distance between the


, the barrier should be very porous and thin,


the void fraction


(fraction of space occupied by gas bubbles) within the cell should be small, and the


conductivity


of the


solution


should


high.


Strong


acids and


bases are


electrolytes to increase conductivity because hydroxide and hydronium ions have very

large ion mobilities.

Overpotential

The overpotential of a half cell is the difference between the equilibrium half cell


potential,


given by the Nernst equation, and the actual cell potential necessary to pass


a Faradaic current.


In this cell, ;h,


like the resistive losses


, is largely a nuisance,


the potential loss at the anode requires the use of a larger cell voltage.


On the other


hand


, a large ar is the goal.


If a reaction


, such as the hydride generation reactions,


which has an


more negative than


the hydrogen half


cell reaction for hydrogen


generation (


=1H2,


is to be forced to occur in aqueous solution,


a large cathode overpotential is essential.

There are two primary causes of overpotential (or polarization): mass transfer


overpotential,


or concentration


polarization,


electron


transfer


overpotential,


activation polarization.


Concentration polarization is important when the rate limiting


step is the diffusion of reactants toward the electrode.


This is usually important when









the reactants are dilute or when very high currents are used.


The activation polarization


is important when an electron exchange step is the rate limiting factor.

A large concentration polarization should be avoided since this leads to a voltage

drop across a thin solution boundary layer (or double layer) adjacent to the electrode and


does not increase the reducing power of the electrode;

activation energy to the hydride generation reaction. (T


this potential cannot provide


he rate of diffusion across this


double layer is the basis for the electroanalytical method of polarography.)


One way


to reduce concentration polarization is to provide a high concentration of reactants.


Another method is to reduce the effective depth of the double layer.


In practice this is


accomplished by stirring the solution or, in a tubular reactor, but using forced solution

flow.

The activation polarization for a reaction that is controlled by a single electron

transfer can be approximated by the famous Butler-Volmer equation,90


(3-3)

The symbol, f, denotes F/RT or the ratio of the Faraday constant to the product of gas


constant and the temperature.


The symbol, a, is the transfer coefficient (about 0.5).90


The term, i, denotes the current density (current per unit area of the electrode).


exchange


current,


is a


measure


current


density


flows


when


overpotential is present.


(A zero net current does not imply a zero io).


If a reaction has


a large io, it'


rate limiting step has a low activation energy; if io is low, the activation


_,(ld)nf'7]









energy is high.


At high negative overpotentials,


the second bracketed term is negligible,


and the current is proportion to the first term.


Solving for ',


or n,


we have,


RT
= n (In (i),
anF


n(i))


(3-4)


= a + blog(i)


(3-5)


where,


a =2


.3 RT log(io)
acnF


(3-6)


RT
.3 RT
anF (3-7)


This relation was derived from empirical data by


Tafel in 1905.90


For well studied reactions

in tables for a large variety of


, such as water decomposition,


conditions.


io and b can be found


One of the largest factors affecting the


exchange current is the nature of the electrode surface.


Some materials are active,or


catalytic,


to the hydrogen evolution reaction.


Other materials are less active


, such as Hg and


Active electrodes include Pt and Ag.

Pb. Table 3-1 lists some kinetic data


for hydrogen evolution on various materials.


Using the Tafel equation and the data in tables,


such as those in Table 3-1, it is









Unfortunately,


the cell current varied rapidly (as is noted above) and so only


anaverage current can be measured.


Figure 3-5 shows the current versus cell potential


curves


for the CEHG


using


M H2S04


electrolyte and


platinum


cathode.


exposed surface area of the cathode was roughly 50 cm


. The two curves were acquired


on two separate days,


the glass frit was replaced between experiments.


averaged for 1 mm, and the error bars represent a 90


The current was


confidence interval.


From the


data in Table 3-1 and the Tafel expression,


is roughly -0.55 V


the activation overpotential at 110 mA/cm2


. If the same current density were used with a lead cathode,


activation overpotential is about 1.6 V


. One thing should be stressed:


just because an


electrode can


produce a large overpotential


does


not necessarily


prove


that it will


produce hydrides quickly.


A surface which is not catalytic for hydrogen generation may


not necessarily be catalytic for hydride generation,


and it may interact with the analyte


in some way that is unfavorable for hydride generation.








Table 3-1:


Some kinetic data for hydrogen evolution.


ELECTRODE SOLUTION TAFEL SLOPE -LOG i,
b (mV) (A/cm2)
Aga 1MHC1 130 3.7
1 M HSO4 110 5.7
1 M NaOH 120 6.5
Cd" 0.5 M HzSO4 135 10.77
Fe" 0.5 M H2SO4 118 5.18
Hgb 1.0 M H2S04 124 11.8
Pbb 1 M H2SO4 120 12.7
0.5 M NaOH 129 6.5
Pd" 2.0 M H2SO4 120 3.2
Ptb (smooth)0 1 M H2SO4 120 3.6
0.5 M NaOH 117 4.1


a Values from reference 90.
b Values from reference 91.
c Platinum roughened by platinization (fast platinum electrodeposition), has a
larger exchange current.






























't
0







ca3







E






CU

* -


Lt~



U





ci)

U



I)

0






U


.










































o o O0 0!


pewnsse eeJe eoe ns epo jeo


wo09


. I a S'S II\ rl lnnllHA %SE *fI ~ HA













CHAPTER 4
CHOICE OF ELECTRODE MATERIALS AND ELECTROLYTES

Introduction


The choice of materials for the anode is limited to conductive materials that will


not corrode or contaminate the anode solution


platinum mesh was chosen.


under a large positive potential.


This mesh has an apparent surface area of 36 cm2 while the


platinum lead wire had a surface area of 31.7 cm2.


Similarly,


the cathode must be made of a conductive material


that will not


corrode or contaminate the solution under a large negative potential.


Also,


for the


present electrolytic hydride generator design, the cathode material must be easily formed


into a coil.


A large hydrogen overpotential might also be advantageous.


The electrolyte must be highly conductive. Any strong, or nearly strong, acid or


a base meets this criterion.


It must also not be reduced at the highly negative cathode.


In addition, if the electrolyte forms an oxidation product at the highly positive anode,


the product must not be very soluble in water. An oxidation product may be drawn

through the glass frit and into the cathode flow cell. Once in the cathode flow cell, the


reduction


oxidation


product


might


compete


with


hydrogen


generation


reaction.


Thus, the cathode activation overpotential would be reduced.


Hydrochloric


acid and HBr were not used as electrolytes to avoid chlorine or bromine generation at
JiLt ..-- -A






84

To choose the electrolyte and cathode material, several comparative tests were


performed.


CEHG was used


with four different cathodes and three electrolytes to


generate AsH3, SbH3 and H2Se.


and palladium.


The four cathodes were made of platinum, silver, lead


The three electrolytes were 1 M sodium hydroxide, 1 M sulfuric acid,


and 1 M phosphoric acid.


Silver was chosen on


laboratory.


the basis of the availability of pure silver wire in the


It was also included in order to test a relatively inexpensive and harmless


material.


Lead was chosen for its high hydrogen overpotential.


Lead is very toxic, and


pure lead wire is soft.


Care should be taken not to touch the wire with bare hands.


Palladium and platinum were chosen for their chemical resistance and their resistance


to corrosion.


Also,


platinum can be cleaned in concentrated nitric acid.


This was


particularly convenient during the concomitant tests to be described in Chapter 5.

The output of the hydride generator was directed into the burner of the FAA


spectrometer


described


Chapter


The atomic


absorption


signal,


produced


standard solutions of As(III), Sb(III) or Se(IV),


potential.


was monitored as a function of cell


Most of the possible combinations of the three analytes, four cathodes and


three electrolytes were tried.


Sb or Se


experiment.


For comparison, a solution having an equal concentration


was also nebulized at the conclusion of each hydride generation


The ratio of the analytical signal from hydride generation to the analytical


signal from solution nebulization was thus calculated for each test.








Experimental


Cathodes


The four cathodes were 1) a coil of silver wire, 50 cm long (1.0 mm diameter


99.9% pure), 2) a coil of platinum wire, 85 cm long (0.5 mm diameter


3) a coil of lead wire, 50 cm long (1.0 mm diameter

palladium wire, 50 cm long (1.0 mm diameter 99.


- 99.95% pure),


- 99.9985% pure) and 4) a coil of


997% pure).


Reagents and Stock Solutions


The stock arsenic, selenium and antimony solutions used throughout this study


are listed in Table 4-1.


The electrolyte being tested was used to make dilute standard


solutions from these stock solutions daily.


Table 4-1:


Stock solutions.


SOLUTION SOLUTE CONCENTRATION SOLVENT
As(III) As2O3 2000 xg/mL 0.05 M NaOH
Sb(III) Sb2O3 1000 pg/mL 0.1 M HCL
Se(IV) Na2SeO3 1000 pg/mL 0.1 M HC1


Preparation of the FAA SDectrometer


The FAA spectrometer was fitted with the appropriate HCL and aligned using


solution nebulization to produce an AA signal for alignment.


As stated in Chapter


each HCL lamp was "warmed up" for at least 30 minutes.








Preparation of the Generator


To prepare the generator for operation,


the electrode to be tested was placed


inside the cathode flow cell, and the generator was assembled. The anode compartment

was rinsed with distilled water and then clean electrolyte solution. The generator could


be quickly filled through the oxygen exhaust port using a large tubing pump, while the


displaced


air was


vented


through


the anode solution


refill


port.


Next,


the anode


chamber was filled with electrolytic solution to a point just under the oxygen exhaust


port.


The generator was then attached to all the equipment shown in Figure 3-4 and was


connected to the burner chamber by teflon tubing as described in Chapter


The flow


of blank electrolyte through the cathode cell was started, and the DC power supply was


turned on to begin electrolysis at 15-20 V

before any samples were introduced to


. The electrolysis was continued for 2-3 min


allow the anode compartment to pressurize.


During this time, the generator was carefully inspected for leaks to prevent the release


of hydride gases into the laboratory.


The anode solution level was always maintained


to within 1 cm of the oxygen vent port by adjusting the flow rate of the pump used to


refill


anode chamber.


anode


solution


stirred


constantly


during


operation.


Next, the flame was ignited, and a solution containing 0.1-1 pg/mL of the


analyte being tested was introduced at


- 6 mL/min for 2 or 3 minutes to pretreat the


electrode. (The production of AsH3, H2Se and SbH3 generally increased slowly when a


clean electrode was used.


After the initial few minutes of generation using a As(III),


Se(IV) or Sb(III), the electrodes did not need to be retreated for at least four hours of








the atomic absorption signal reached a steady state,


the blank electrolyte was again


introduced.


some


combinations


cathode,


electrolyte


analyte,


no signal


observed.


In these cases, a more concentrated analyte solution, up to 200 pg/mL,


used.


no signal


was observed


with


larger


concentration,


abandoned.

Determination of Optimal Carrier Gas Flow Rate


Arsine was generated using a solution of 10 pg/mL As(III) in 1 M H2SO4 and


a silver electrode.


The solution flow rate was 6.0 mL/min.


The arsenic absorption


signal was recorded as a function of argon carrier, or sparge, gas flow rate.


The signal


increased rapidly with increasing argon


flow until


the argon flow rate reached 200


mL/min.


The arsenic absorbance reached a plateau after 400-500 mL/min and did not


significantly increase even when 1000 mL/min of argon was used.


An argon sparge gas


flow


rate of 500-700


mL/min


was chosen


for use with


these


tests.


It was


confirmed that the carrier gas flow rate had a similar affect on the Sb and Se signals.

Procedure for Cathode and Electrolyte Tests


The optimal flame gas flow rates and flame height,


which gave the maximum


absorbance signal, were determined before each test using solution nebulization.


It was


confirmed that the optimal flame parameters were not different for sample introduction


using CEHG for arsenic and selenium.


However, the optimal air flow rate for antimony


measurement, using CEHG, was so low that the nebulizer could not be used.


Since the







88

purpose of these texts was to compare the hydride generation sensitivity to the solution

nebulization sensitivity, the optimal solution nebulization conditions were used.

While a blank was being introduced to the hydride generator, the power supply


was set to a high voltage setting (around 25 V).


recorded.


The blank response, or 0%


Next, several standard solutions of various concentration were introduced.


A solution that produced an easily measured absorbance,


linear response range,


but was clearly within the


was chosen for use in the remainder of the test.


Finally, the FAAS response to a blank solution and to the analyte solution was


recorded at a number of cell potentials.


The signal was averaged for 10-30 seconds.


After each test the absorbance due to a standard solution of equal concentration was


measured using solution nebulization.


Table 4-2 lists the conditions used.


All of the


solutions used were in a 2-50 pg/mL range.

Results and Discussion


In Figure 4-1


the atomic absorption signal shows the recorded relative arsenic


absorbance signal as a function of cell potential using 1 M H2SO4 for each cathode.


signal-axis,


Y-axi


scaled


to show


ratio


measured


absorbance of arsenic introduced by hydride generation to the absorbance recorded for


a solution of identical arsenic concentration introduced using the nebulizer.

bars in the figure represent one standard deviation from 4-5 measurements,


represent a best fit,


The error

the curves


third order polynomial equation.


Figure 4-2 shows similar data recorded for H2Se generation in 1 M sulfuric acid,






89

obtained for AsH3 generation using 1 M H3PO4 and 1 M NaOH, respectively. Finally,

Table 4-3 lists some results for a variety of less successful combinations of the various

cathode materials and electrolytes.


Table 4-2:


Experimental conditions for FAAS measurements.


Spectral bandpass (interference filter used) 0.32 nm
Wavelength

As 193.7 nm
Sb 217.6 nm
Se 196.8 nm
Hydrogen (flame gas) flow rate

As and Se 6.5 L/min
Sb 7.5 L/min
Air (flame gas) flow rate 4.2 L/min
Cathodic (sample) solution flow rate 5 mL/min
Nebulizer sample flow rate (when used) 5 mL/min
Carrier gas (argon) flow rate 600 mL/min
HCL current, boost current

As 25 mA, 60 mA
Sb 18 mA, 40 mA
Se 18 mA, 40 mA


In general, sulfuric acid was the best electrolyte for the generation of all three


hydrides.


There was no significant difference between the effectiveness of Pt, Ag and


Pd cathodes for AsH3 generation. Palladium gave the best results for H2Se generation,








be easily cleaned.


The lead cathode gave the largest signal for SbH3, but it corroded


quickly in sulfuric acid at cell potentials larger than 18 V.


Table 4-3:


Hydride generation results with several cathode materials and electrolytes.


ELEMENT ELECTROLYTE CATHODE MAX. SIGNAL3 (CELL
SOLUTION POTENTIAL)b
Se HzS04 Ag 2.25 0.22 (18 V)
Se NaOH Pt no signal0
Se NaOH Pb no signal
Se NaOH Ag no signal
Se H3PO4 Ag no signal
Sb H2SO4 Ag 0.43 .08 (25 V)
Sb NaOH Pt 0.57 0.09 (20 V)
Sb NaOH Pb no signal
Sb H3PO4 Pt 1.59 + 0.15 (18 V)
Sb H3PO4 Pb 1.64 0..17 (18 V)


a The ratio of the atomic absorbance signal for hydride generation to the atomic
absorbance signal for solution nebulization.
b The potential which gave the maximum signal, or the maximum potential used.
c No signal was observed for solutions of up to 200 ug/mL.