Citation
Characteristics and application of a laser ionization/evaporation source for tandem mass spectrometry

Material Information

Title:
Characteristics and application of a laser ionization/evaporation source for tandem mass spectrometry
Creator:
Perchalski, Robert John, 1948- ( Dissertant )
Yost, Richard A. ( Thesis advisor )
Winefordner, James D. ( Reviewer )
Eyler, John R. ( Reviewer )
Dorsey, John G. ( Reviewer )
Wilder, B. Joe ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1985
Language:
English
Physical Description:
xiii, 195 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Desorption ( jstor )
Ion currents ( jstor )
Ionization ( jstor )
Ions ( jstor )
Lasers ( jstor )
Mass spectrometers ( jstor )
Mass spectroscopy ( jstor )
Nickel ( jstor )
Quadrupoles ( jstor )
Signals ( jstor )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Lasers in biochemistry ( lcsh )
Lasers in chemistry ( lcsh )
Mass spectrometry ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
An instrument system is proposed that is based on tandem mass spectrometry (MS/MS) and a focused laser for vaporizing selected areas of samples within the ion source so that the molecular components of the sample can be identified and quantitated. This instrument is similar in principle to a commercially available laser mass spectrometer (LAMMA), but additional selectivity is possible with the proposed system because it has two stages of mass analysis, making molecular analysis feasible. The instrument used for this study consists of a triple quadrupole mass spectrometer and a single-shot, coaxial flash lamp-pumped dye laser, interfaced through a simple system of mirrors and a lens,mounted near the ion source, which serves as a vacuum seal and focuses the laser radiation on the sample. The system is evaluated by observation of short-lived (less than 1 ms) and long-lived (greater than 1 ms) events after the laser is fired. Ion signals lasting several seconds are observed when samples are irradiated in a heated chemical ionization (CI) source. These signals are a combination of a desorption event and a vaporization event, and desorption CI is used as a model to characterize the laser-initiated processes. The latter are called Laser-Induced or Laser-Enhanced Desorption CI. The current system proved unsatisfactory for quantitative analysis because the time limitations of the instrument combined with a concentration-dependent ion signal lifetime produced calibration curves with low correlation coefficients. Qualitative confirmation of the presence of a drug in rat liver was obtained, however, after laser vaporization of a small sample of intact tissue within the ion source. Recommendations are made for overcoming the time limitations of the present system so that a microprobe capable of molecular analysis of heterogeneous and compartmentalized biological samples might be developed.
Thesis:
Thesis (Ph. D.)--University of Florida, 1985.
Bibliography:
Bibliography: leaves 185-193.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Robert John Perchalski.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
029845411 ( AlephBibNum )
AEK6764 ( NOTIS )
015477434 ( OCLC )

Downloads

This item has the following downloads:

PDF ( 6 MBs ) ( .pdf )

characteristicsa00perc_Page_059.txt

characteristicsa00perc_Page_123.txt

characteristicsa00perc_Page_185.txt

characteristicsa00perc_Page_008.txt

characteristicsa00perc_Page_046.txt

characteristicsa00perc_Page_096.txt

characteristicsa00perc_Page_022.txt

characteristicsa00perc_Page_122.txt

characteristicsa00perc_Page_014.txt

characteristicsa00perc_Page_093.txt

characteristicsa00perc_Page_049.txt

characteristicsa00perc_Page_072.txt

characteristicsa00perc_Page_160.txt

characteristicsa00perc_Page_134.txt

characteristicsa00perc_Page_039.txt

characteristicsa00perc_Page_021.txt

characteristicsa00perc_Page_076.txt

characteristicsa00perc_Page_066.txt

characteristicsa00perc_Page_193.txt

characteristicsa00perc_Page_035.txt

characteristicsa00perc_Page_119.txt

characteristicsa00perc_Page_136.txt

characteristicsa00perc_Page_044.txt

characteristicsa00perc_Page_012.txt

characteristicsa00perc_Page_156.txt

characteristicsa00perc_Page_150.txt

characteristicsa00perc_Page_169.txt

characteristicsa00perc_Page_205.txt

characteristicsa00perc_Page_104.txt

characteristicsa00perc_Page_081.txt

characteristicsa00perc_Page_034.txt

EEVWPEPGW_9GR0F5_xml.txt

characteristicsa00perc_Page_028.txt

characteristicsa00perc_Page_145.txt

characteristicsa00perc_Page_200.txt

characteristicsa00perc_Page_051.txt

characteristicsa00perc_Page_184.txt

characteristicsa00perc_Page_188.txt

characteristicsa00perc_Page_074.txt

characteristicsa00perc_Page_033.txt

characteristicsa00perc_Page_202.txt

characteristicsa00perc_Page_088.txt

characteristicsa00perc_Page_178.txt

characteristicsa00perc_Page_025.txt

characteristicsa00perc_Page_114.txt

characteristicsa00perc_Page_186.txt

characteristicsa00perc_Page_011.txt

characteristicsa00perc_Page_111.txt

characteristicsa00perc_Page_139.txt

characteristicsa00perc_Page_155.txt

characteristicsa00perc_Page_174.txt

characteristicsa00perc_Page_189.txt

characteristicsa00perc_Page_061.txt

characteristicsa00perc_Page_069.txt

characteristicsa00perc_Page_058.txt

characteristicsa00perc_Page_201.txt

characteristicsa00perc_Page_112.txt

characteristicsa00perc_Page_107.txt

characteristicsa00perc_Page_029.txt

characteristicsa00perc_Page_128.txt

characteristicsa00perc_Page_010.txt

characteristicsa00perc_Page_078.txt

characteristicsa00perc_Page_162.txt

characteristicsa00perc_Page_042.txt

characteristicsa00perc_Page_183.txt

characteristicsa00perc_Page_179.txt

characteristicsa00perc_Page_143.txt

characteristicsa00perc_Page_176.txt

characteristicsa00perc_Page_146.txt

characteristicsa00perc_Page_037.txt

characteristicsa00perc_Page_197.txt

characteristicsa00perc_Page_095.txt

characteristicsa00perc_Page_167.txt

characteristicsa00perc_Page_192.txt

characteristicsa00perc_Page_040.txt

characteristicsa00perc_Page_198.txt

characteristicsa00perc_Page_057.txt

characteristicsa00perc_Page_109.txt

characteristicsa00perc_Page_013.txt

characteristicsa00perc_Page_041.txt

characteristicsa00perc_Page_064.txt

characteristicsa00perc_Page_173.txt

characteristicsa00perc_Page_102.txt

characteristicsa00perc_pdf.txt

characteristicsa00perc_Page_130.txt

characteristicsa00perc_Page_177.txt

characteristicsa00perc_Page_161.txt

characteristicsa00perc_Page_048.txt

characteristicsa00perc_Page_182.txt

characteristicsa00perc_Page_172.txt

characteristicsa00perc_Page_181.txt

characteristicsa00perc_Page_159.txt

characteristicsa00perc_Page_084.txt

characteristicsa00perc_Page_056.txt

characteristicsa00perc_Page_092.txt

characteristicsa00perc_Page_098.txt

characteristicsa00perc_Page_003.txt

characteristicsa00perc_Page_090.txt

characteristicsa00perc_Page_117.txt

characteristicsa00perc_Page_151.txt

characteristicsa00perc_Page_101.txt

characteristicsa00perc_Page_166.txt

characteristicsa00perc_Page_100.txt

characteristicsa00perc_Page_082.txt

characteristicsa00perc_Page_191.txt

characteristicsa00perc_Page_016.txt

characteristicsa00perc_Page_180.txt

characteristicsa00perc_Page_062.txt

characteristicsa00perc_Page_138.txt

characteristicsa00perc_Page_070.txt

characteristicsa00perc_Page_004.txt

characteristicsa00perc_Page_208.txt

characteristicsa00perc_Page_132.txt

characteristicsa00perc_Page_071.txt

characteristicsa00perc_Page_099.txt

characteristicsa00perc_Page_148.txt

characteristicsa00perc_Page_210.txt

characteristicsa00perc_Page_199.txt

characteristicsa00perc_Page_097.txt

characteristicsa00perc_Page_052.txt

characteristicsa00perc_Page_110.txt

characteristicsa00perc_Page_118.txt

characteristicsa00perc_Page_103.txt

characteristicsa00perc_Page_054.txt

characteristicsa00perc_Page_030.txt

characteristicsa00perc_Page_091.txt

characteristicsa00perc_Page_142.txt

characteristicsa00perc_Page_190.txt

characteristicsa00perc_Page_120.txt

characteristicsa00perc_Page_024.txt

characteristicsa00perc_Page_203.txt

characteristicsa00perc_Page_206.txt

characteristicsa00perc_Page_196.txt

characteristicsa00perc_Page_075.txt

characteristicsa00perc_Page_027.txt

characteristicsa00perc_Page_154.txt

characteristicsa00perc_Page_063.txt

characteristicsa00perc_Page_053.txt

characteristicsa00perc_Page_168.txt

characteristicsa00perc_Page_164.txt

characteristicsa00perc_Page_187.txt

characteristicsa00perc_Page_009.txt

characteristicsa00perc_Page_144.txt

characteristicsa00perc_Page_170.txt

characteristicsa00perc_Page_019.txt

characteristicsa00perc_Page_126.txt

characteristicsa00perc_Page_113.txt

characteristicsa00perc_Page_157.txt

characteristicsa00perc_Page_047.txt

characteristicsa00perc_Page_137.txt

characteristicsa00perc_Page_055.txt

characteristicsa00perc_Page_165.txt

characteristicsa00perc_Page_006.txt

characteristicsa00perc_Page_038.txt

characteristicsa00perc_Page_015.txt

characteristicsa00perc_Page_094.txt

characteristicsa00perc_Page_153.txt

characteristicsa00perc_Page_079.txt

characteristicsa00perc_Page_195.txt

characteristicsa00perc_Page_073.txt

characteristicsa00perc_Page_129.txt

characteristicsa00perc_Page_043.txt

characteristicsa00perc_Page_083.txt

characteristicsa00perc_Page_209.txt

characteristicsa00perc_Page_020.txt

characteristicsa00perc_Page_152.txt

characteristicsa00perc_Page_060.txt

characteristicsa00perc_Page_089.txt

characteristicsa00perc_Page_007.txt

characteristicsa00perc_Page_175.txt

characteristicsa00perc_Page_194.txt

characteristicsa00perc_Page_135.txt

characteristicsa00perc_Page_133.txt

characteristicsa00perc_Page_001.txt

characteristicsa00perc_Page_031.txt

characteristicsa00perc_Page_106.txt

characteristicsa00perc_Page_140.txt

characteristicsa00perc_Page_087.txt

characteristicsa00perc_Page_127.txt

characteristicsa00perc_Page_158.txt

characteristicsa00perc_Page_065.txt

characteristicsa00perc_Page_036.txt

characteristicsa00perc_Page_147.txt

characteristicsa00perc_Page_080.txt

characteristicsa00perc_Page_026.txt

characteristicsa00perc_Page_068.txt

characteristicsa00perc_Page_018.txt

characteristicsa00perc_Page_017.txt

characteristicsa00perc_Page_124.txt

characteristicsa00perc_Page_149.txt

characteristicsa00perc_Page_116.txt

characteristicsa00perc_Page_086.txt

characteristicsa00perc_Page_121.txt

characteristicsa00perc_Page_002.txt

characteristicsa00perc_Page_045.txt

characteristicsa00perc_Page_204.txt

characteristicsa00perc_Page_032.txt

characteristicsa00perc_Page_207.txt

characteristicsa00perc_Page_005.txt

characteristicsa00perc_Page_023.txt


Full Text




















CHARACTERISTICS AND APPLICATION OF A LASER
IONIZATION/EVAPORATION SOURCE FOR TANDEM MASS SPECTROMETRY




by

ROBERT JOHN PERCHALSKI


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


1985







































To my parents with love.
















ACKNOWLEDGMENTS


I wish to express my sincere appreciation to Dr. Richard A. Yost

for his direction and faith in this project. I am also grateful to the

members of my committee: Dr. James D. Winefordner whose initial advice

should have been given more serious consideration and whose instrumenta-

tion helped to make this project more successful; Dr. B.J. Wilder whose

continuing personal and professional support over the last 13 years has

been invaluable in the attainment of this goal; Dr. John G. Dorsey who

offered advice and encouragement; Dr. L. James Willmore who offered

counsel and expertise; and Dr. John Eyler who took over where

Dr. Willmore left off. I also appreciate the generosity of

Dr. Martin Vala who loaned us the laser.

Thanks are also due to Drs. Neal Brotherton and Dean Fetterolf who

got me started, and to the members of the research group who freely gave

advice and assistance. Special thanks are due to Dr. Jodie V. Johnson

who kept the instrument running, and to Dr. Edward Voigtman for help

with the laser and the storage oscilloscope experiments.

The LAMMA experiments were completed through the kindness of

Dr. David M. Hercules of the Department of Chemistry, University of

Pittsburgh. The assistance of Dr. Edward J. Hammond in the animal

experiments and Dr. William Ballinger in the preparation of samples for

LAMMA analysis is appreciated.












This project could not have been completed without the help and

cooperation of my children, Matthew, Abby, Ben, and David.

Finally, for her support and patience, I thank my wife, Lynn.

















TABLE OF CONTENTS


CHAPTER PAGE

ACKNOWLEDGEMENTS................................................ iii

LIST OF TABLES............................................. vii

LIST OF FIGURES............................................ viii

ABSTRACT.................................................... xii

I INTRODUCTION .................................................. 1

Overview of Current Technology............................. 1
Overview of Present work................................... 13

II CHARACTERISTICS OF LASER-INDUCED PROCESSES
IN MASS SPECTROMETRY ......... ............................ 15

Effects of Absorption of Laser Radiation
at Solid Surfaces......................................... 16
Laser-Induced Vaporization and Particle Emission........... 18
Analytical Usefulness of Laser Desorption
Mass Spectrometry........................................ 25

III EXPERIMENTAL ................................................. 27

Laser Microprobe Study..................................... 28
Preliminary Triple Quadrupole Mass Spectrometry
Study: Regional Distribution of Phenobarbital
and Phenytoin in Rat Brain................................ 30
Laser-Triple Stage Quadrupole.............................. 35

IV PRELIMINARY STUDIES AND INSTRUMENT CHARACTERISTICS.......... 47

LAMMA Experiments.......................................... 47
Regional Distributions of Phenobarbital
and Phenytoin in Rat Brain................................ 55
Laser-Triple Stage Quadrupole Characteristics.............. 62












CHAPTER PAGE

V SHORT-LIVED PROCESSES......... ................... .... ..... 68

Triple Stage Quadrupole Control............................ 69
Comparison of Storage Oscilloscope and Data System.......... 75

VI LONG-LIVED PROCESSES.................... ..... ........... ... 99

Long-Lived Signal Profiles.................................. 100
Quantitative Analysis and Application of
Laser Induced Desorption Chemical Ionization............. 147

VII CONCLUSIONS AND FUTURE WORK ............................... 170

Conclusions Based on the Present Work...................... 170
Future Work.... ............................................ 178

APPENDIX

CHARACTERISTICS OF DIRECT PROBE, DESORPTION
CHEMICAL IONIZATION, AND LASER-INDUCED
DESORPTION CHEMICAL IONIZATION BEHAVIOR OF
TEST COMPOUNDS............................................ 183

LITERATURE CITED ............................................ 185

BIOGRAPHICAL SKETCH .................... .................... 194
















LIST OF TABLES


TABLE PAGE

I Common Operational Modes of the Triple-Stage
Quadrupole Mass Spectrometer............................... 33

II Masses and Flight Times for Nickel, Molybdenum
and Lead Isotopes in LAMMA Calibrant........................ 49

III Effects of Programmed RF Function and Rod Polarity
Switch Settings During Various TSQ Operational Modes......... 73

IV Weight of Phenytoin Removed from Copper Grid
by a Single Laser Shot...................................... 126

V (M+H)+ Ion Intensity Ratios for LIDCI of PHT and
and PHT* Deposited with Parlodion on Copper Grids.
Comparison of 2 Shots at Different Sites on
Each of 5 Grids ............................................. 130

VI Graded Melting Point Standard Solution Containing
Compounds Having Melting Points from 76'C to
Above 300C ................................................. 131

VII Comparison of Maximum and Summed Ion Currents
for DCI of Graded Melting Point Stock Dilution
from Copper, Nickel, and Silver EM Grids.................... 135

VIII Calibrant Solutions Made from Graded Melting Point
Standard for DCI/CAD from Nickel and Teflon
Substrates ........................... ............ .......... 152

IX Phenytoin Calibrant Solutions for LIDCI
from Nickel Grids ........................................... 158

X Coefficients of Determination for DCI calibration
with Nickel Grids and Teflon Paddle, and for LIDCI
Calibration with Nickel Grids................................ 161

XI Phenytoin Concentrations in Rat Brain and
Liver Samples and in Fixative Solutions..................... 163
















LIST OF FIGURES


FIGURE PAGE

3.1 Analyzer section of Finnigan TSQ mass spectrometer.......... 32

3.2 Diagrams for machining blank side flange of TSQ
for fused silica lens, and spacers for adjusting
focal point of lens within ion source........................ 37

3.3 Diagrams for machining 900 and 45 stainless steel
solids probe extensions..................................... .40

3.4 Layout of lasers and beam directing optics in
relation to TSQ ion source.................................. 42

4.1 Plot of flight time vs. the square root of ion
mass of LAMMA calibrant .. .................. ................ 52

4.2 LAMMA spectra for nickel-molybdenum-lead
calibrant, section of rat brain from cobalt lesion
site, and section of rat brain from in control
animal...................................................... 54

4.3 Ion current profiles for direct solids probe
vaporization and and selected reaction monitoring of
phenobarbital and phenytoin in rat brain extract............. 58

4.4 Concentrations of phenytoin and phenobarbital in
various areas of rat brain (ug/g) and in plasma
(pg/mL)................................................... 60

4.5 Variation of focal spot diameter with distance in
front of and in back of focal point......................... 64

4.6 Laser pulse shapes from photodiode signal as
monitored by storage oscilloscope at various
driver voltages and laser cavity diameters.................. 67

5.1 Stability diagram for quadrupole mass filter................ 72
+
5.2 Plots of total ion current vs. time for Ni
(m/z 58) showing storage scope data and profile
data for nickel EM grid target............................... 77











FIGURE


5.3 Plots of total ion current vs. time for cationized
phenytoin (m/z 275) showing storage scope data
and profile data for phenytoin over Na2CO3 on
copper EM grid.............................................. 83

5.4 Plots of total ion current vs. time showing the
efficiency of Q1 and Q3 mass filtering (profile
data) for phenytoin over Na2CO3 on copper EM grid ........... 85

5.5 Plots of total ion current vs. time showing
efficiency of Q1 mass filtering and effects of
programmed RF function in Q1................................ 89

5.6 Plot of total ion current vs. time for 4-vinyl-4-
aminobutyric acid (mol wt 129) and reserpine (mol
wt 608) over Na2CO3, showing time resolution of
respective cationized molecules with m/z 152 and
m/z 631 ..................................................... 92

5.7 Plots of total ion current vs. time for phenytoin
over Na2CO3 showing effects of reagent gas on
signal for cationized molecule.............................. 95

5.8 Plot of total ion current vs. time for phenytoin
in parlodion film (no Na2CO3) with auxiliary
CI (m/z 253) ................................................ 97

6.1 Plot of reconstructed ion current vs. time
for laser evaporation of carbamazepine epoxide
(mol wt 252) in CI source.................................... 102

6.2 Plots of ion current vs. time for laser
vaporization of tetrabutylammonium
perchlorate, menthol glucuronic acid,
adenosine, and acetazolamide................................ 105

6.3 Plots of RIC vs. time and solids probe temperature
vs. time for desorption chemical ionization of
phenytoin from copper grid floating in CI source
at the electron entrance aperture........................... 109

6.4 Plots of RIC vs. time and solids probe temperature vs.
time for direct probe evaporation from a glass solids
probe vial, and desorption chemical ionization from a
400 mesh nickel EM grid of hydergine........................ 112











FIGURE PAGE

6.5 Plot of RIC with respect to position of sample in CI
ion source for phenytoin on copper grid...................... 115

6.6 Plots of RIC vs. time for DCI of sodium phenytoin
with filament turned on at various times.................... 118

6.7 Plots of RIC vs. time and solids probe temperature
vs. time showing DCI and LEDCI of phenobarbital............. 122

6.8 Plots of RIC vs. time for a mixture of phenytoin
and 1,3-15N-2- 13C-phenytoin with parlodion
evaporated on copper grids.................................. 128

6.9 Plots of ion current vs. time for DCI of graded
melting point stock dilution from copper, nickel
and silver grids............................................ 134

6.10 Plots of ion current vs. time for DCI of graded
melting point standard (60-76 ng of each drug)
from nickel EM grid, and Teflon probe extension............. 138

6.11 Plot of negative ion RIC vs. time for DCI of
sodium phenytoin at various source temperatures.............. 142

6.12 Plot of positive ion RIC vs. time for DCI of
sodium phenytoin at various source pressures................. 144

6.13 Plot of ion current vs. time for LIDCI of
components of graded melting point standard................. 146

6.14 Plots of RIC vs. time for DCI/CAD of 0.558 ng and
5.42 ng of hydergine on nickel grid showing
dependence of signal lifetime on sample size................ 150

6.15 Plot of response ratio (analyte/internal standard)
vs. amount of analyte on grid for DCI/CAD of 1 pL
of graded melting point calibrant solutions on
nickel grids................................................ 154

6.16 Plot of response ratio (analyte/internal standard)
amount of analyte on grid for DCI/CAD of 1 ul of
graded melting point calibrant solutions on Teflon
paddle..................................................... 156











FIGURE


6.17 Plot of response ratio (PHT/PHT*) vs. amount of PHT
on nickel grid for LIDCI/CAD of PHT calibrant
solutions................................................... 160

6.18 Plots of ion current vs. time for phenytoin parent
ion, major daughter ions and ion recorded at laser
firing for LIDCI/CAD of rat liver tissue section............ 166

6.19 Daughter spectra of phenytoin (m/z 253) from laser
shot for rat liver tissue section, and sample of
pure PHT on nickel grid..................................... 169

















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






CHARACTERISTICS AND APPLICATION OF A LASER
IONIZATION/EVAPORATION SOURCE FOR TANDEM MASS SPECTROMETRY


By

Robert John Perchalski

May, 1985


Chairman: Richard A. Yost
Major Department: Chemistry

An instrument system is proposed that is based on tandem mass

spectrometry (MS/MS) and a focused laser for vaporizing selected

areas of samples within the ion source so that the molecular components

of the sample can be identified and quantitated. This instrument is

similar in principle to a commercially available laser mass spectrometer

(LAMMA), but additional selectivity is possible with the proposed system

because it has two stages of mass analysis, making molecular analysis

feasible. The instrument used for this study consists of a triple

quadrupole mass spectrometer and a single-shot, coaxial flashlamp-pumped

dye laser, interfaced through a simple system of mirrors and a lens,

mounted near the ion source, which serves as a vacuum seal and focuses

the laser radiation on the sample.


xii











The system is evaluated by observation of short-lived (< 1 ms) and

long-lived (> 1 ms) events after the laser is fired. Ion signals

lasting several seconds are observed when samples are irradiated in a

heated chemical ionization (CI) source. These signals are a combination

of a desorption event and a vaporization event, and desorption CI is

used as a model to characterize the laser-initiated processes. The

latter are called Laser-Induced or Laser-Enhanced Desorption CI. The

current system proved unsatisfactory for quantitative analysis because

the time limitations of the instrument combined with a concentration-

dependent ion signal lifetime produced calibration curves with low

correlation coefficients. Qualitative confirmation of the presence of

a drug in rat liver was obtained, however, after laser vaporization of a

small sample of intact tissue within the ion source.

Recommendations are made for overcoming the time limitations of the

present system so that a microprobe capable of molecular analysis of

heterogeneous and compartmentalized biological samples might be

developed.


xiii
















CHAPTER I

INTRODUCTION




The purpose of this work was twofold: first, to design, construct,

and evaluate a laser evaporation/ionization source for a triple

quadrupole mass spectrometer, and second, to determine the feasibility

of developing a milli- or microprobe, based on a microscopically-

directed sampling laser interfaced to a tandem mass spectrometer that is

capable of molecular analysis of heterogeneous samples. Such an

instrument could be used for chemo-morphological mapping of geological,

metallurgical, environmental and astronomical samples, and of intact

plant and animal tissues by direct, in situ elemental and molecular

analysis. The major use of this instrument, however, is anticipated in

the biomedical sciences, in which samples and their chemical

constituents are highly compartmentalized from the organism to the

ultrastructural level.



Overview of Current Technology




Much of the work of medical, agricultural, biochemical, and

pharmacological research involves experiments designed to provide

insight into the chemical and physical processes occurring at the

cellular and subcellular levels. Elucidation of these processes is












sought to define, in molecular terms, the normal chemical condition of

organisms of interest from microscopic flora and fauna to man,

abnormalities in this chemical condition that are evident as disease or

deficiency, and effects and mechanisms of action of the various curative

measures employed to produce and maintain an optimal state of being.

The experiments are conducted in five general areas: morphology,

physiology, pharmacology, histochemistry, and chemistry. Morphological

studies relate shape and structure with physical function.

Physiological and pharmacological studies define organism and

substructure reaction to endogenous and exogenous (both therapeutic and

anti-therapeutic) chemical species, respectively. Histochemical studies

relate, usually indirectly and qualitatively, chemical species, and

anatomical structure. Chemical analysis also relates chemical species

with structure but has greater potential for quantitation through direct

determination of the compounds of interest. Each of these

methodologies, when used in this context, is site-specific at various

levels of spatial resolution, and depends wholly or in part upon

microscopy (1, 2).

Methodologies for Physical and Chemical Analysis

Morphology is generally the initial subject of interest and is

pursued at three levels, delineated by the instruments used.

Macrostructural surveys include all work done with the unaided eye and

therefore are limited to gross anatomical correlations in large

organisms and organ systems with dimensions measured in meters to

millimeters. Microstructural studies are conducted with the light












microscope in its various forms and can delve deeply into organ systems

in large species or give information of more general nature for one- or

few-celled organisms. Dimensions are measured in millimeters to

micrometers. Ultrastructural investigations are conducted with the

electron microscope and give information at the cellular and subcellular

levels with dimensions normally measured in micrometers to nanometers.

Physiology and pharmacology are usually restricted to micro- to

macrostructural studies, since the probes used to measure or observe the

chemical or physiological response are incapable of spatial resolution

below the cellular level (5 to 10 pm) in a living organism, or in a

normally functioning tissue or cellular preparation. These studies

include clinical evaluations of drug effects on animals and humans

(pharmacological), and nutritional studies designed to evaluate effects

of and requirements for biomolecules such as vitamins, amino acids,

hormones, and transmitter substances necessary for proper nervous system

function (physiological). In many cases, subject response coupled with

chemical analysis of easily obtained body fluids (blood or urine) can

give the information sought, since all foreign substances and most

biomolecules present in an organism are in dynamic equilibrium with the

circulatory system. At the lower limits of spatial resolution for these

studies, test chemicals are usually applied by microiontophoresis (3)

and responses are monitored by potentiometric (4) or voltammetric (5)

microelectrodes.

The gross chemical analyses which necessarily precede physiological

studies and coincide with pharmacological studies can be carried out by












usual methods (chromatography, ultraviolet, visible and infrared

spectrophotometry, nuclear magnetic resonance and mass spectrometry,

atomic absorption, emission and fluorescence spectroscopy) applied to

physiological fluids and tissue homogenates. As the work progresses,

however, to the stage at which a more exact localization of active

substances is desired, more spatially selective techniques must be used

which usually place more stringent requirements on the analytical

techniques, particularly with regard to sensitivity. The techniques

that are most used for localization studies are based on three general

methodologies, histochemistry, autoradiography, and direct chemical

analysis of microdissected tissue samples or subcellular fractions

obtained by density gradient centrifugation.

The histochemical methods rely upon chemical, enzymatic or

immunological reactions which involve the molecule of interest, an

enzyme specific for the molecule of interest, an inhibitor of that

enzyme, or an antibody created to bind specifically to the molecule of

interest, as reagents to give, either directly or indirectly, a product

visible with a light or electron microscope (6). The method is devised

so that all reactions occur rapidly in succession, and the final visible

product is synthesized at and does not diffuse from the location of the

reactant of interest. Potential for qualitative analysis is predicated

on the specificity of the chosen reagent, enzyme or antibody for the

molecule of interest. Insuring this specificity presents the most

challenging problem during method development. Quantitative analysis is

possible by densitometry if normal stains are used, or by fluorometry if












the final product is naturally or can be made fluorescent with a

fluorogenic reagent or fluorescent-labeled enzyme or antibody (7-9).

The introduction of raster beam microscopes, coupled with computerized

data acquisition and imaging techniques, have contributed greatly to

improving the accuracy of these methods (10).

Autoradiography is a direct method of localizing radioactively-

labeled molecules of interest at the light and electron microscope

levels (11). The system under investigation is sacrificed and sectioned

at an appropriate time after the labeled compound is introduced. The

sections can be treated as usual for the microscopic method of choice

(staining) and are then coated with a thin layer of silver halide

emulsion (gelatin). Resolution is determined by the homogeneity of the

emulsion, the size of the particles, and the thicknesses of tissue and

emulsion layers. After exposure of the emulsion for a suitable period

of time, which depends upon the half-life of the nuclide (normally

several weeks to a few months), the emulsion is developed and the sample

is subjected to microscopy.

Direct chemical analysis of anatomical structures isolated by

microdissection or density gradient centrifugation can sometimes be

performed using scaled-down versions of normal analytical methods.

Typical methods for microdissection of brain, however, allow

differentiation of more than 220 cell types and areas of a rat brain

(whole brain wet weight: 1.5 g) with samples measuring 200 500 um

square x 100 pm thick, a volume of 4 25 nL (12), and methods for

isolation and analysis of single brain cells (20 50 pL) have been












reported (13). As sample size is decreased to effect a more exact

localization, enzymatic cycling methods must be used (14) which take

advantage of the chemical amplification (104-108) afforded by the

continuous recycling of enzymes in a reaction. This catalytic effect

increases the concentration of the substance chosen as an indicator to

the point where it can be accurately measured. These methods have

generally been applied to determination of endogenous substances,

biomolecules and enzymes, rather than exogenous substances.

Microbeam and Probe-Type Instruments

Another methodology that has been developed is the use of

microbeams of electrons, ions or light, not only to visualize

microstructures, but also to obtain a simultaneous chemical analysis of

the surface by analysis of emissions of electromagnetic radiation,

electrons and ions resulting from the interaction of the incident beam

with the sample. The first such instrument, patented in 1947 and

commercially available in the late 1950s, was the electron microprobe

(15) which uses a focused (0.1 to 3 pm) beam of electrons to produce X-

rays characteristic of the elements in the sample. Other probe-type

instruments that followed include those based on Auger electron

spectroscopy, ion backscattering spectrometry, electron spectroscopy for

chemical analysis (ESCA) and secondary ion mass spectrometry (SIMS) (16,

17). The SIMS instruments gave rise to the ion microprobe and ion

microscope, which differ in the method of obtaining an image of the

sample. The microprobe uses a focused beam of ions to scan in raster

fashion over the sample surface. The microscope uses a flood beam about












250 to 300 im in diameter to image an area of the surface. Final images

of the secondary ions emitted maintain a direct relationship to the

point of origin at the sample surface (18).

Laser-based instruments. Shortly after the introduction of the

laser in 1960 (19), it was incorporated into a microprobe instrument

based on emission spectrography (20), and emission of electrons, ions

and neutral atoms was observed from conductors, semiconductors and

insulators irradiated in a mass spectrometer (21). Several reviews

discuss the early work in both of these fields, concentrating primarily

on emission spectroscopic instruments, laser characteristics, and

interactions of laser radiation with solid surfaces (22-24). The most

recent spectroscopy-based instrument to be introduced commercially was

the laser Raman molecular microprobe (MOLE) (25). Point or global

illumination was possible to allow recording of a complete Raman

spectrum of a particular portion of a sample or mapping of a larger area

by selecting a Raman line characteristic of an analyte of interest.

Although this instrument was capable of identifying some molecular

species, the complexity of the identifiable molecules and of their

environment was limited. Most applications to date have shown

identifications of inorganics, or simple organic deposited in

biological systems in concretions of essentially pure crystals. This

system is no longer commercially available, but work is still continuing

in this area (26). Another microprobe instrument based on nitrogen-

pumped dye laser-excited atomic fluorescence of analytes after site-












specific sample evaporation by a second microscopically-directed laser

(ruby) has also been reported (27).

Laser-based mass spectrometers. Lasers have been applied

extensively in mass spectrometry (MS) over the past 20 years as a means

of rapid vaporization and ionization of samples in the ion source.

Conzemius and Capellan (28) have reviewed much of the work conducted

prior to 1980, including instrument system characteristics (lasers and

mass spectrometers) and applications. This review has been updated to

include all work published through 1982 (29). In 1975, Hillenkamp et

al. (30) reported the use of a LAser Microprobe Mass Analyzer (LAMMA)

which was designed primarily for in situ determination of

physiologically significant elements in intact biological samples. The

instrument, introduced commercially in 1977 (LAMMA 500), is based on a

time-of-flight mass spectrometer with a Q-switched, frequency-multiplied

Nd-YAG sampling laser and a He-Ne spotting laser, both directed and

focused through a high quality microscope. The system has been

described (31), and two reviews have covered sample requirements,

performance characteristics (32), and applications to structural

analysis (33). Proceedings of a symposium on analysis with LAMMA have

been published (34) and include articles on characteristics of laser

spectra of various molecular classes (35, 36), preparation of biological

specimens (37-39), and techniques for quantitative standardization (40,

41). A second generation instrument (LAMMA 1000) has been introduced

(42-44) which allows for analysis of bulk samples rather than restrict-

ing the analyst to the use of thin sections, as did the LAMMA 500.












Molecular analysis with current probes. The characteristic that

links all of the above microprobe instruments is their almost exclusive

application to elemental analysis. This is not to say that they have

not been used to obtain spectra of molecules; however, there is an

almost complete lack of applications for which the microprobe capability

has been used in real, heterogeneous samples for qualitative or

quantitative determination of molecules. In fact, in a search of the

literature, only four such applications were found (two for the MOLE and

two for the LAMMA). In three of these (25, 39, 45), the analysis

succeeded because the spatial resolution of the laser beam allowed

sampling of microcrystals or pure concretions of analyte, and not

because of the resolution or chemical specificity of the spectrometric

method. The only application that succeeded in identifying a molecular

species in an intact, heterogeneous sample involved detection of a

fungal metabolite, glyceollin, in plant tissue with LAMMA (46). In this

study, two ion masses indicative of the metabolite were observed as the

infection boundary was crossed. Part of the problem is that these

instruments were never really designed to do molecular analysis. It is

not, however, unrealistic to expect that the considerable selectivity

necessary for molecular analysis is obtainable with appropriate

combinations of currently available methodologies. In fact, several

review articles on mass spectrometry and tandem mass spectrometry that

have come out of the laboratory of R.G. Cooks in the past four years

have included some reference to a "molecular microscope" (47-51),









10


although no specific instrument has been described. The current project

grew out of a desire to begin working toward this ideal.

Projections for a molecular probe. Of the three instruments

described above that are capable of localized molecular analysis, the

laser mass spectrometer was thought to be the most promising. Due to

the non-exclusive nature of the electromagnetic spectra of molecules

(i.e., all organic molecules contain essentially the same types of bonds

which give rise to signals contained within a limited portion of the

electromagnetic spectrum for any given method), the selectivity and

specificity of a microprobe instrument based upon optical (Raman)

spectroscopy would be seriously limited. The molecular reordering that

occurs during surface bombardment with energetic ion beams (50) would

ultimately limit the resolution obtainable with an instrument based upon

SIMS. The primary deficiency that precludes molecular analysis of

heterogeneous samples with the LAMMA is the lack of confirmation of the

identity of an ion corresponding to a specific mass peak. Elemental

analysis is possible because molecular interference can be eliminated

by using a high power density (greater than 109 W cm-2) and confirmation

can be based on relative isotopic abundances for the analyte. This

deficiency might be corrected by substituting a tandem mass spectrometer

for the current single stage mass spectrometer to allow online verifi-

cation of the identity of selected ions by sequential fragmentation and

mass analysis steps.

Tandem mass spectrometry. Modern tandem mass spectrometry (MS/MS)

was developed in the laboratories of Cooks (52) and of McLafferty (53)












using reversed-geometry, double-focusing instruments for Mass-analyzed

Ion Kinetic Energy Spectrometry (MIKES). The introduction of the triple

quadrupole mass spectrometer by Yost and Enke (54) simplified the

automation of tandem mass analysis, increased the allowable speed of

data acquisition, and brought the technique within reach of laboratories

other than those specializing in design and construction of unique MS

systems (similar instruments are now available from several

manufacturers). A recently published volume (55) summarizes the

developments and applications of this relatively new field. A tandem

mass spectrometer consists of an ionization-fragmentation region

followed in space (tandem sector and quadrupole instruments) or time

(Fourier transform ion cyclotron resonance instruments (56)) by a mass

selector-analyzer region. Therefore, a molecular ion or fragment

produced in the ion source can be selected by the first mass analyzer,

fragmented, usually by collisionally activated dissociation (CAD), and a

complete spectrum of the fragments can be obtained by scanning the

second analyzer. Carrying this process to its limit, a single pure

compound can be totally characterized by obtaining CAD spectra for the

molecular ion and each fragment produced in the ion source; or a mixture

may be characterized by using a soft ionization technique producing

primarily molecular ions with little fragmentation, and obtaining a CAD

spectrum for each component. Used in this way, the instrument is

analogous to a chromatograph/mass spectrometer system; however, the

relatively lengthy chromatographic separation is replaced by a nearly

instantaneous mass separation. This decreases analysis and method












development time, but more importantly, allows additional modes of

operation that require simultaneous presence of signals from all

analytes of interest. An example of full use and an explanation of

these special MS/MS operating modes are given in a report describing a

rapid, general method for detection and structural elucidation of drug

metabolites (57).

The recording of mass spectra is a statistical process in that the

reproducibility of relative ion intensities is dependent upon the length

of time the ion signals are present, and the length of time available

for monitoring each mass of interest. Therefore, the primary problem

associated with interfacing a laser with a mass spectrometer is a

temporal one, since the ion signal resulting from interaction of a

pulsed laser (pulsewidth = 10-8 to 10-6 s) normally has a lifetime of

less than a millisecond. Zackett et al. (58) used a laser pulsed at

10 Hz to obtain MS and MS/MS spectra of sucrose deposited on a silver

foil. They reported that a 10 ns pulse gave ions for 300 us in a

chemical ionization (CI) source maintained at 0.1 torr with isobutane

and for 1.5 ms at a source pressure of 0.5 torr. The same group

identified a naturally occurring quaternary amine in cacti using laser

desorption MS/MS after extracting the plant material and evaporating the

extract on a silver or platinum foil (59). They used a CI source

maintained at 0.5 torr with isobutane, and concluded that the CI

conditions contributed only to increasing the ion signal lifetime by

causing more efficient extraction of ions from the source, and not to

the actual ion formation process. Cotter (60), however, has used an












auxiliary chemical ionization source to take advantage of the abundant

and relatively long-lived neutral species desorbed by the laser as

discussed by Ready (61). A molecular ion for guanosine was recorded

with a lifetime of 4 ms after a 40 ns laser pulse.



Overview of Present Work



The scope of the present work will include all aspects of

interfacing a single shot-laser with a tandem quadrupole mass

spectrometer, and prolonging the signal obtained upon interaction of the

laser radiation with a solid sample. Chapter II will cover the work

that has been done to develop an understanding of the interaction of

laser radiation with matter and the processes leading to ion formation

in laser desorption mass spectrometry. Chapter III will describe the

instrumental systems and procedures used in this study. The

capabilities of a commercially available laser microprobe (LAMMA

500) will be evaluated along with those of the triple quadrupole mass

spectrometer for localization of species of interest (Chapter IV). The

characteristics of the new system will be studied with respect to short-

lived (Chapter V) and long-lived (Chapter VI) processes, so that

recommendations can be made for future research efforts (Chapter VII) in

the development of a microprobe capable of site-specific molecular

analysis of heterogeneous and compartmentalized samples. Finally, an

attempt will be made to answer a question which may make the development









14


of a molecular microprobe-microscope for analysis of biological samples

somewhat academic; that is,

Although a molecule may be present in tissue, is it in a

form that is recoverable by direct vaporization or is it

incorporated into membranes, proteins and anatomical

structures in such a way as to make it unrecognizable?

This question may not have a single answer covering all species of

interest; however, it is important to show that discrete molecular

structures can be identified by vaporization from intact tissue.

















CHAPTER II

CHARACTERISTICS OF LASER-INDUCED PROCESSES
IN MASS SPECTROMETRY



This chapter is primarily a literature review covering the most

notable work directed toward developing an understanding of the laser-

induced processes leading to ion formation. The initial discussion of

absorption of laser radiation at solid surfaces covers the physical

basis for prediction of the extent of temperature increase at the target

surface upon laser impact. This treatment assumes that no vaporization

or melting of the target material occurs, and is applicable to

experiments conducted on bulk samples or substrates at power densities

at which little visible surface evidence exists of laser impact. The

second section deals with the investigation of the laser vaporization-

ionization processes. The primary questions that have been asked

include (1) what is the energy spread of particles of mass spectral

significance emitted (ions, atoms, neutral and ionized molecules), and

does the system exist in a state of thermal equilibrium? (2) how does

the nature of the laser-induced surface emission change with time? and

(3) what are the mechanisms of laser-induced ionization? The chapter

ends with a discussion of the techniques used to enhance the analytical

usefulness of laser-initiated processes.











Effects of Absorption of Laser Radiation at Solid Surfaces



The unique properties of laser light include a high degree of

brightness, directionality and monochromaticity, these features

resulting from the coherence of the emitted radiation. The laser is,

therefore, a versatile source of spatially and temporally controllable

spectral and thermal power. In laser desorption mass spectrometry

(LDMS), it is the thermal power that is of primary interest because of

the need for sample evaporation and ionization. The advantages of

laser-induced vaporization over more common methods involving resistive

heating lie in the high rate of heating which can approach 1010 C/s and

the extreme localization of the effects to areas as small as 10-8 cm2

(62).

Experimental systems used for LDMS are of two basic types: (1)

those in which samples are mounted on thin substrates such as electron

microscope grids and sample and substrate are completely vaporized by

the laser (MS instruments based primarily on time-of-flight); and (2)

those in which samples are mounted on relatively thick substrates (thin

foils to bulk metal stages) and are partially or totally vaporized with

no apparent effect on the substrate (studies employing time-of-flight,

quadrupole and sector instruments). Because the latter process is

viewed as an effect on the substrate which is transferred to the sample,

much effort has been expended in determining the temperature rise

induced in the sample substrate by the laser radiation, and the time-

course of laser-induced thermal effects.











Ready (63) has developed equations for predicting temperature

variations in metallic substrates for lasers having pulse widths greater

than 10-8 s. An assumption of local equilibrium at the site of impact

is based on the premise that the energy is absorbed by electrons whose

collision frequency is 1013 s-1 (the Debye frequency) so that

distribution of the energy to other electrons and the lattice is

instantaneous with respect to the duration of the laser pulse. A

generalized dimensionlesss) equation was presented by Ready to allow

calculation of curves applicable to any material. This treatment was

used by van der Peyl et al. (64) to calculate the time-dependent

temperature variation for a stainless steel surface irradiated by a

pulsed (200 ns) C02 laser, whose pulse profile was experimentally

determined. A comparison of stainless steel (m.p. 1808C) and quartz

(m.p. 1883C) indicated that a lower energy was required to bring quartz

to its melting point (2.4 mJ) than was required for stainless steel

(5.6 mJ). Lincoln and Covington (65) used a specially designed time-of-

flight mass spectrometer (66) which allowed variation of specimen

distance from the ion source (30 or 100 cm) and monitoring of ions or

neutrals. They calculated surface temperatures for two different types

of graphite having different thermal conductivities by measuring the

expansion velocity of the vapor plume and calculating the vaporization

temperature from gas dynamic models for the vapor expansion process.

Both of these studies indicated a more rapid increase in surface

temperature for materials with lower thermal conductivities, the

rationale being that materials of low thermal conductivity do not allow











for rapid redistribution of the absorbed energy to the bulk substrate.

Knox (67) has reviewed some of the earlier work in which surface

temperatures were determined experimentally. Results for normal pulsed

lasers with power densities of 106 to 108 W cm-2 were 2200'C to 9700C

for metals, and were relatively reproducible. Q-switched lasers with

power densities of 108 to 1011 W cm-2, however, gave widely variant

results for similar materials.



Laser-Induced Vaporization and Particle Emission



This discussion will focus on the vaporization processes that occur

when laser radiation strikes a target. The initial section will cover

vaporization from the point of view of the substrate, primarily with

respect to spatial and temporal characteristics. The second section

will cover vaporization from the point of view of the vapor plume, and

will summarize most of the work that has been done to elucidate the

mechanisms of laser-induced ion formation.

Laser-Induced Vaporization

Ready (68) again appears to be the jumping-off place for discussion

of the vaporizing actions of normal-pulsed and Q-switched lasers. Data

have been given for depths of holes produced in a number of thick metal

targets. For a normal-pulsed laser (700 ps pulsewidth), the depth was

dependent on thermal conductivity of the substrate at low laser power

density, but became independent of this property at higher power density

and dependent on latent heat of vaporization of the metal. Therefore,











at low power, the material having the greater thermal conductivity

redistributes the energy to the bulk phase rapidly, experiencing

relatively little vaporization, whereas at higher power, energy is

absorbed too fast for redistribution, regardless of the thermal

conductivity, and deeper holes are produced in the material having the

lower heat of vaporization. The power density, F (W cm-2), at the point

where heat of vaporization becomes dominant is given approximately by

Eq. 2.1:



F > 2LpK t- (2.1)



where L is the latent heat of vaporization (J g-1), p is the density

(g cm-3), Kis the thermal diffusivity (cm2 s-1), and t is the laser

pulse length (s). This view is consistent with the types of laser

interactions outlined by Kovalev et al. (69) and Furstenau (70), who

described interactions for metals, semiconductors and dielectrics based

on a model in which the extent of the interaction was determined by the

concentration of free charge carriers (electrons) in the substrate and

the difference between the valence band-conduction band energy gap and

the photon energy of the laser radiation.

For Q-switched lasers (F > 109 W cm-2), vaporization is not a

simple matter of ablation of material at its boiling point at a front

that advances into the substrate. Other processes occur, including

exertion of pressure on the target by the vapor plume, allowing

superheating of the surface and leading to explosive ejection of











material, and absorption of incoming laser radiation by the plume. In

general, much less material is removed by short (ns), very high power

density pulses than by longer pulses of lower power density.

Scott and Strasheim (71) have studied the spatial and temporal

characteristics of vapor plume generation by normal-pulsed (850 us),

semi Q-switched (30 us) and Q-switched (50 ns) lasers. Metal substrates

were used and impressive photographic evidence was presented in each

case. In the normal mode, vapor emission began at about 30 us after

initiation of laser action and ended at about 400 ps, even though lasing

continued. This was thought to be due to scattering of incoming light

by particles emitted from the surface. The vapor expanded at

8.3 x 103 cm s-1 with the solid angle of ejection decreasing as the

crater became deeper. The semi Q-switched laser provided radiation at a

higher power density and emission began after about 20 us. No particle

ejection was observed and the plume expanded at a rate of

3.2 x 105 cm s-1. For the Q-switched pulse, a small spherical plasma

was visible for about 20 us. The expansion rate was greater than

106 cm s-1 and no particle ejection was seen. In all cases, the vapor

plume expanded in a direction normal to the target surface regardless of

the angle of incidence of the laser beam (33, 72).

Laser-Induced Particle Emission

The primary mathematical relationship found in articles dealing

with LDMS is the Langmuir-Saha equation (Eq. 2.2):


+/i0 = (g/g0)exp[(O I)/kT]


(2.2)











where i+ and i0 are the positive ion and neutral molecule fluxes

leaving the surface at temperature T. The quantities g+ and go are the

statistical weights of the ionized and neutral states, k is the

Boltzmann constant, is the work function of the surface (eV), and I

is the ionization potential (eV). The equation gives the temperature

dependence of the degree of ionization of surface emissions (73). For

substances for which I is greater than D, evaporation occurs at a

temperature less than that required for efficient ionization (74). Datz

and Taylor (75) reported that the alkali metals, the species most

commonly implicated in laser-induced ionization processes, follow this

equation when adsorbed on platinum and tungsten. From a review of the

literature, Conzemius and Capellan (28) divided ionization efficiency

into two general regions based on power densities of lasers employed.

Below 108 W cm-2, ionization efficiency was low, about 10-5. Above

5 x 109 W cm-2, emissions were almost completely ionized.

Energy- and Time-Resolved Studies

Measurement of the energies and time-course of emission of desorbed

ions and neutrals is useful for elucidation of the mechanisms of laser

desorption-ionization, and also for providing a firm basis for increasing

the analytical potential of the technique. Cotter (76, 77) has recently

reviewed the area of energy- and time-resolved studies, some of the most

notable work having been done in his own laboratory. Numerous reports

of energy spreads of elemental and molecular ions have offered

conflicting evidence. The review by Ready (78) cited ion energies up to

5 keV and typical values of up to 400 to 500 eV. Bernal et al. (79),










using a Q-switched CO2 laser (775 ns pulsewidth; 7 x 106 W cm-2),

reported time-resolved energy spreads during the 1.3 us ion pulse as 0

to 60 eV initially and rising to about 200 eV. Ion energies of up to

20 eV have been reported for the LAMMA 500 for thin non-metallic

samples, and up to several hundred eV for bulk metallic samples (31).

Hardin and Vestal (80), using a quadrupole instrument with samples being

introduced continuously on a moving belt, reported energies of 6 to

25 eV for potassium iodide clusters and protonated and cationized guanosine

ions and fragments. They reported that the observed time-of-flight

distributions were dependent on laser power density, and the relative

concentrations of sample and alkali salt on the belt. All work was

performed at a power density of about 107 W cm-2. Energies as low as

0.26 eV were reported by van der Peyl et al. (81) for Na+ in a double-

focusing instrument.

Cotter and coworkers have published a series of papers in which

they report results obtained on a versatile time-of-flight instrument

designed to allow observation of ions produced directly by the laser, or

laser-produced neutrals ionized by a pulsed electron beam (82-85). An

ion-drawout pulse was provided to sample the contents of the ion source

at any time up to 500 us after the laser was fired. They have used this

system to resolve the convoluted profiles of laser-desorbed ions into

separate contributions from prolonged vaporization envelopes and finite

energy spreads. The energy spread for K+ was 15.3 eV at 10 ps after the

laser pulse, and decreased to 0.31 eV at about 44 ps. This is probably









23


the most accurate work that has been done to date, and while it does not

negate the value of the previous work, it certainly offers a much

improved method for solving the problems.

Mechanistic Studies

Early work in laser-assisted mass spectrometry was concerned mostly

with elemental analysis or pattern recognition after pyrolysis of large

molecules or molecular aggregates. Vastola et al. (86, 87) introduced

the concept of using laser radiation for volatilization and ionization

of molecules that could not be vaporized intact by the usual methods of

ohmic heating. Until 1978, laser desorption continued to take a back

seat to field desorption (FD), secondary ion mass spectrometry (SIMS)

and the newly introduced (88) plasma desorption (PD) in the field of

molecular analysis. The commercial introduction of the LAMMA 500 and

the report by Posthumus et al. (89) in particular provided impetus to

define the optimum conditions for laser desorption and the mechanisms of

ion formation. In this work, laser desorption spectra were presented

for important biomolecules which were generally considered intractable,

including oligosaccharides, cardiac glycosides, steroid conjugates,

antibiotics, nucleotides, nucleosides, amino acids, peptides, and

chlorophyll. All gave abundant molecular ions as alkali metal ion

adducts (M+Na+, M+K+), indicative of a process called "cationization".

This study, along with others (88, 90, 91), indicated that a rapid

heating rate was important for molecular ion production in preference to

fragmentation and decomposition. Although most researchers have used

laser power densities of 105 to 108 W cm-2, Stoll and Rollgen (92)












observed the cationization process and desorption of quaternary ammonium

ions (preformed ions) using a CW CO2 laser at a power density of about

20 W cm-2; however, after scrupulously cleaning the copper substrate and

purifying a sucrose sample to eliminate traces of alkali metals, no

molecular ions or protonated or cationized molecules were detected, even

though the sucrose did evaporate.

In an effort to determine the mechanism of the cationization

reaction which is predominant in laser desorption, van der Peyl et al.

(93-95) performed experiments using several different instrumental

configurations in which thermionically produced alkali metal ions

originated in a different area of the ion source than the test compound

(sucrose). Temperatures of both thermionic emitter and sample holder

were monitored during resistive heating and laser heating cycles and

plotted along with ion intensity curves. They proposed a model in which

the laser serves to produce a steep temperature gradient at the surface

with alkali metal ions thermionically emitted from the hot central zone

of the laser spot (greater than 427C) and neutral sample molecules

vaporized from the relatively cooler periphery. Ion formation occurs in

the gas phase as an ion-molecule reaction producing a stable, even-

electron species with a positive charge on the alkali metal atom. Stoll

and Rollgen (96, 97) arrived at a similar conclusion using a system in

which the thermionic emitter was not completely isolated from the sample.

Heinen (98) and Hercules et al. (33) have reviewed the types of

ions formed from molecules containing various functional groups.

Hillenkamp (99, 100) has summarized the work reported by the most












prominent LDMS research groups, and listed some features of LD mass

spectra. These features include (1) in contrast to the normal

situation with resistive heating, polarity and ionic character promote

desorption and ionization; (2) ions are generally even-electron species;

(3) cationization by alkali and other singly-charged metal ions is

common; (4) anionization has been observed but not frequently reported;

(5) positive and negative ions are produced in approximately equal

amounts, with less fragmentation of negative ions; (6) cluster ions are

frequently seen; (7) matrix effects are observed; and (8) extent of

fragmentation is dependent on laser power density. Cooks and Busch

(101) have compared the various desorption ionization techniques (SIMS,

FD, PD, LD) and have developed a general model for the ionization-

fragmentation events that occur. The model is based on the fact that

these techniques, which differ in the types of energy put into the sample

result in the production of similar ion types.



Analytical Usefulness of Laser Desorption Mass Spectrometry



The utility of a laser desorption/evaporation source in mass

spectrometry derives from the characteristics of the laser beam, i.e.,

it can be focused to sub-micrometer diameter spot sizes, contains a high

thermal power, can desorb or vaporize large, polar molecules intact,

and can ionize them directly or present them to the instrumental system

in a state suitable for ionization by other means. The primary tasks

associated with coupling a laser with a mass spectrometer are (1)












prolonging the laser-induced events so that they persist for the time

required for the mass spectrometer to acquire statistically relevant

data; or (2) designing the mass spectrometer so that it can acquire such

data during the lifetime of laser-produced ions. Ion lifetimes have

been extended by operating the ion source in the chemical ionization

mode (102). This not only takes advantage of the neutral species that

are desorbed, but also increases the lifetime of ions produced directly

by the laser. Using a renewable surface with a pulsed or CW laser

provides a continuous signal, but this is not ideal for microprobe work,

because the intrinsic nature of the investigation precludes the

assumption of a renewable surface (i.e., why use a microprobe if the

sample is homogeneous enough for the surface to be considered

renewable). A variation of this approach involves pulsing the laser in

synchronization with mass scanning so that only one mass is monitored

for each laser pulse (repetitive laser desorption (103)). McCrery et

al. (104) used ion storage in a Fourier Transform mass spectrometer and

were able to digitize and average twelve decay transients for each laser

pulse. Adapting the mass spectrometer to the laser duty cycle has been

done in the LAMMA instruments with time-of-flight mass analysis. Other

workers have used sector instruments with photoplate or array detectors

to simultaneously detect ions over a certain mass range. Since the

objective of this work is to evaluate a tandem mass spectrometer system

for application to laser microprobe mass spectrometry, emphasis was be

placed on obtaining as much information as possible from a single laser

shot.
















CHAPTER III

EXPERIMENTAL




This chapter describes the instrumentation and procedures used in

this three-part study. The first experiments, performed in the

laboratories of D. M. Hercules at the University of Pittsburgh, were

done to gain familiarity with a commercially available laser microprobe

and to illustrate the problems inherent in using a single stage mass

analyzer for analysis of heterogeneous samples. The second study was

done to evaluate the capabilities of the Triple-Stage Quadrupole Mass

Spectrometer (TSQ) for quantitative analysis of drugs in small (2-10 mg)

samples of rat brain in preparation for anticipated laser/MS/MS analysis

at some later date. The objectives of the final series of experiments

were (1) to design and construct a laser interface for the triple

quadrupole mass spectrometer; (2) to evaluate the instrument system with

respect to short-lived and long-lived events produced directly or

indirectly by the laser; and (3) to determine the feasibility of and

make recommendations for construction of a laser/tandem mass

spectrometer-based milli- or micro- probe capable of molecular analysis

of heterogeneous samples.










Laser Microprobe Study



Apparatus

Laser microprobe mass analyzer (LAMMA 500). The LAMMA 500

(Leybold-Heraeus GmbH, Federal Republic of Germany) has been described

in detail (31). The instrument consists of a time-of-flight mass

spectrometer with an open ion source, coupled with a Q-switched,

frequency-multiplied Nd-YAG laser through a high quality microscope. A

He-Ne laser aligned coaxially with the Nd-YAG laser is used for site

selection and aiming. The laser light is incident on the sample at 90

to the sample rear surface and ions are extracted into the mass

spectrometer at 90* to the sample front surface (the instrument operates

with a 90/-90 configuration, according to the convention established

by Conzemius and Capellan (28)). For this reason, samples are

restricted to thin films, or if granular forms are used, the laser must

be aimed so as to strike a grain edge. An ion reflector is employed to

minimize the effect of the initial ion energy spread on mass resolution

and to increase resolution by allowing a longer flight path without

greatly increasing instrument size.

Procedure

Preparation of animals. An epileptic focus was created in an adult

male Sprague-Dawley rat by sub-dural deposition of powdered cobalt

(105). After 2 weeks, the lesioned animal along with a control (non-

lesioned) animal were sacrificed. The brains were removed (106) and

fixed in phosphate-buffered glutaraldehyde for 48 hours. The lesioned












area of the test animal and a corresponding area from the brain of the

control animal were dissected out and embedded for sectioning. Thin

sections (1 pm) from each sample were mounted on electron microscope

grids for use in the LAMMA, and serial sections were mounted on glass

microscope slides for pre-test selection of analysis sites by optical

microscopy.

Analysis. A calibration mixture of Ni, Mo, and Pb was made by

grinding together NiS04, MoS2, and PbSO4, and dusting the fine particles

onto an EM grid. Spectra were acquired with a 0.020 ps sampling

interval and a 1100 channel delay time. Delay time is the time in

sampling intervals that the transient recorder waits after the laser is

fired before beginning to accumulate data and store it in the 2048

channels available. Sampling interval is the time between each sample,

or channel width in ps, and determines the upper limit of the mass

range. The delay time multiplied by the sampling interval determines

the low mass recorded. The delay time plus 2048 multiplied by the

sampling interval is the total flight time that is covered in a run.

Laser energy before filtering (energy incident upon the sample is

selected very roughly by neutral density filters) is read in mJ after

each shot from a digital panel meter which receives a signal from a

photodiode. Energy from shot to shot is variable and uncontrolled.

Energies varied from 9.7 mJ to 32.5 mJ, but were usually between 10 and

20 mJ. Mass calibration was obtained from the Ni-Mo-Pb calibrant

spectrum by identification of mass peaks from the isotopic patterns and

correlation of flight times with chart distances. Numerous spectra were












recorded from the lesion sample and the control sample with delay times

of 1000 channels and a sampling interval of 0.020 js.Specimens were

observed in the LAMMA with the 32X objective (total magnification 320X).

The mass range covered was m/z 20-219.



Preliminary Triple Quadrupole Mass Spectrometry Study:
Regional Distribution of Phenobarbital and
Phenytoin in Rat Brain



Apparatus

Finnigan triple stage quadrupole (TSQ) mass spectrometer. The TSQ

(Finnigan MAT, San Jose, Calif.) (Fig. 3.1), consisting of an electron

impact (EI)/chemical ionization (CI) ion source followed in series by a

mass selector-analyzer quadrupole (Ql), an ion-focusing fragmentation

quadrupole (Q2), a second mass selector-analyzer quadrupole (Q3), and a

continuous dynode electron multiplier, has been described (107, 108).

Finnigan INCOS software (TSQ revision C) was used for instrument

control, and data acquisition, storage, and manipulation. The center

quadrupole (Q2) functions as a second fragmentation region when

pressurized with an inert gas (N2, Ar), operating by collisionally

activated dissociation (CAD) as described by Yost et al. (119). Table I

shows the primary operating modes of the instrument along with the

corresponding configurations of the mass analyzer and collision

quadrupoles.




















Figure 3.1

Analyzer section of Finnigan TSQ mass spectrometer.



















ION QUADRUPOLE QUADRUPOLE QUADRUPLE ELECTRON
SOURCE ONE TWO THREE MULTIPLIER


1st STAGE 2nd STAGE
MASS COLLISION, MASS
IONIZATION- SEPARATION -> FOCUSING SEPARATION -> DETECTION






SAMPLE -J L l
INLET i, nc -:i Hi $--



















Table I. Common Operational Modes of the Triple-Stage Quadrupole Mass Spectrometer


mode


(1) single stage mass
spectrometer



(2) parent ion experiment




(3) daughter ion
experiment






(4) neutral loss experiment
(linked scan)


quad 1


a) pass all m/z
b) scan m/z


scan m/z




select m/z







scan m/z


quad 2


no collision gas;
pass all m/z



collision gas.a
pass all m/z



collision gas,a
pass all m/z






collision gas,a
pass all m/z


quad 3


a) scan m/z
b) pass all m/z


select m/z


(a) scan m/z or
(b) select daughter
ions for each ion
passed through
quad Ib


scan m/z at same
rate as quad 1 but
with a constant
mass difference


spectrum


normal mass spectrum
(CI, El, positive
and/or negative ion)


spectrum of all ions
that fragment to give
selected m/z


spectrum of all ions that
arise from fragmentation
of selected m/z




spectrum of all daughter
ions that result from
loss of selected
mass difference


a) If no collision gas is present, the unimolecular dissociation products (metastable ions) will be observed.
b) Selected reaction monitoring. Only daughter ions or ions characteristic of parent ion fragmentation are monitored.











Procedure

Preparation of animals. Rats were given phenobarbital or phenytoin

(20 mg kg-l twice daily) for 5 days. Animals were sacrificed, the

brains were removed, and various structures cerebellarr gray and white

matter, sensory motor cortex, basal ganglia, hippocampus, corpus

callosum, occipital cortex, pontine white matter) were separated by

vacuuming into weighed glass pipets plugged with glass wool.

Analysis. The pipets containing the brain samples were reweighed

to obtain sample weight, and the section of pipet containing the sample

was crushed in a screw cap culture tube (Teflon-lined screw cap) for

analysis. Whole blood was also collected and plasma was analyzed.

Methylated analogues of the drugs [5-(p-methylphenyl)-5-ethyl barbituric

acid, 5-(p-methylphenyl)-5-phenylhydantoin] were used as internal

standards. Samples were homogenized by sonication in 300 uL of 1 M

phosphoric acid. After addition of 300 )jL of 1 M phosphoric acid

containing the internal standard, samples were extracted with 3 mL of

chloroform. The chloroform residue was dissolved in 100 uL chloroform

and 1 to 2 uL was placed in a probe vial for MS. The mass spectrometer

was operated in the selected reaction monitoring (SRM) mode by

sequentially passing the drug and internal standard molecular ions

resulting from methane CI through Q1, fragmenting those by CAD in Q2

(pressurized with nitrogen at 2.3 mtorr; collision energy of 18 eV) and

selecting the major daughter ions for each in Q3. Thus, phenobarbital

(mol. wt. 232) was monitored by selecting the m/z 162 daughter ion in Q3

while passing m/z 233 through Ql. Quantitation was based on the ion












intensity ratio of m/z 162 to m/z 191, the daughter ion of the

phenobarbital internal standard parent ion m/z 247. Phenytoin (mol. wt.

252) concentration was calculated from the ratio of the m/z 182 daughter

ion of the phenytoin molecular ion, m/z 253, to the m/z 196 daughter ion

of the internal standard (parent ion, m/z 267).



Laser-Triple Stage Quadrupole Studies



Apparatus

Laser. The laser was a coaxial flashlamp-pumped dye laser (Model

SLL-625, Candela Corporation, Natick, Mass.) operated primarily with

Rhodamine 6G (4 x 10-5 M in absolute ethanol) with driver voltages

between 16 kV and 20 kV. The output coupler was a sapphire window

without reflective coatings. Pulsewidth was 1.5 to 2 ps. The laser

could be operated only in the single shot mode with a maximum firing

rate of 3 shots min-1. An iris diaphragm, calibrated in 2 mm increments

between 2 mm and 18 mm, was inserted into the laser cavity between the

dye cell and the output coupler to reduce the diameter of the laser beam

and focal spot. Pulse energy, measured with a volume absorbing energy

meter (Gen-tec, Ste-Foy, Quebec, Canada), was 0.65 J, 1.49 J, and 1.64 J

for driver voltages of 16 kV, 18 kV and 20 kV, respectively.

Laser-mass spectrometer interface. The side flange of the TSQ was

machined according to the specifications of Fig. 3.2.a to accommodate a

1.5" diameter fused silica lens (focal length: 12.5 cm). Three spacers

(Fig. 3.2.b) were also machined to allow adjustment of the point of




































Figure 3.2

Diagrams for machining (a) blank side flange of TSQ to hold 1.5"
diameter fused silica lens, and (b)spacers for adjusting focal
point of lens within ion source. Measurements marked are:
(A) 4.5 in.; (B) 1/4 in.; (C) 7/8 in.; (D) 5/8 in.; (E) 1 in.;
(F) 1/4 in. (G) 1/8 in., 3/16 in., and 1/4 in. (3 spacers).












focus over a 9/16" range in 1/16" steps. As it happened, the focal

point was about 1/16" from the reagent gas inlet side of the ion volume

with no spacers. Vacuum was maintained with O-ring seals. To allow

entrance of laser light into the ion source, a 1/8" hole was drilled

into the ion volume (El and CI) directly opposite the reagent gas entry

hole. Extensions to the solids probe were machined out of stainless

steel according to the specifications of Fig. 3.3 to allow insertion of

samples on electron microscope (EM) grids and angles of laser incidence

of 90 and 45. The 45 tip was later modified to allow grids to float

in the ion volume with no backing. This tip was used for the desorption

CI experiments. The dye laser was aligned (Fig. 3.4) by directing a He-

Ne laser (LI) through the dye cell (L2) with the retroreflector (MO)

removed, through the output coupler (WI) to a plane mirror (Ml) which

reflected the beam to another plane mirror (M2) mounted on the TSQ, then

to plane mirror (M3) mounted on a gimbal mount on a line axial with the

focusing lens (W2). Final adjustments in laser aiming could be made by

slight adjustments of M3 while determining the point of impact on an

electron microscope grid inserted into the ion source.

Detection systems. Short-lived events were monitored with one of

two MS detection systems. In the first, a storage oscilloscope (Model

549, Tektronix, Inc., Beaverton, Ore.) was connected directly to the

electron multiplier of the TSQ through a Type D high gain differential

amplifier plug-in, and was triggered by a TTL pulse from a photodiode-

to-TTL interface connected to the output of a fast photodiode (FND100,

EG&G Electro-Optics, Salem, Mass.). The laser light (about 8% of total


































Figure 3.3

Diagrams for machining (a) 900 and (b) 450 stainless steel solids
probe extensions. The 450 extension was later modified for
desorption chemical ionizaton (DCI) by grinding the tip until it
was flat (dotted line). Measurements shown are: (A) 9/32 in.;
(B) 1/4 in.; (C) 0.0765 0.0005 in.; (D) 1/16 in.;
(E) 0.215 0.001 in.; (F) 1/64 in.; (G) 0.133 0.001 in.; (H)
13/32 in.. Depression is 7/64 in. in diameter and between 1/64
in. to 1/32 in. deep. A small leaf spring tightened by a screw
is used to hold the tabbed EM grid on the 900 tip, and a cylinder
made to just fit over the end of the 45 tip is used to secure
the grid on that probe extension.

















a) 900 SOLIDS


PROBE EXTENSION


SIDE VIEW


I B


C
i ;


SF
E DEPRESSION
DEPRESSION


END VIEW


b) 450 SOLIDS PROBE EXTENSION WITH DCI MODIFICATION

END OF EXTENSION AFTER
S A ---G MODIFICATION FOR DCI


FI !DEPRESSION


' D T
I I
*- H -t,


TOP VIEW





















Figure 3.4

Layout of lasers and beam directing optics in relation to TSQ ion
source. Components are: (LI) He-Ne alignment laser; (L2) dye
cell of dye laser; (MO) 100% reflecting plane mirror for laser
back reflector; (Ml, M2, M3) plane mirrors for beam steering with
M3 on gimbal mount; (Sl) iris diaphragm calibrated in 2 mm
increments from 2 mm to 18 mm; (WI) sapphire window for laser
output coupler; and (W2) fused silica lens with 12.5 cm focal
length for focusing laser light inside ion source.















D; ----a--

N N~
N~ C~


Sl MO



t^ M3

I\

I \
i\


I M2

IONS ---- SAMPLE
ION
SOURCE


ATMOSPHERIC PRESSURE

HIGH VACUUM


I I I











output) was reflected to the photodiode from a microscope slide placed

in the light path. Comparison of the TSQ instrument oscilloscope with

the storage scope indicated a sensitivity of 1.6 x 10-7 A V-1 for the

latter. The storage scope could not be run from the amplified

instrument signal, because the output of the amplifier contained the

instrument scope trigger signal and mass marker signals which disrupted

the storage scope signal and triggering. In the second detection

system, the TSQ data system was used for monitoring fast signals by

acquiring data in profile mode. In this mode, time resolution of 25 us

was obtained because the instrument acquires data at the maximum analog-

to-digital conversion (ADC) rate of 40 kHz. A single long scan (20 s)

of profile data was acquired during which time the laser was fired. By

setting the mass of interest with the unscanned quadrupole and turning

off the DC voltage to the scanned quadrupole, a signal of ion current

vs. time was obtained. This operation will be described in more detail

in Chapter V. The beginning of the scan was monitored by connecting an

oscilloscope (Model 2213, Tektronix, Inc.) to the positive Y-axis input

(Y+) at the rear of the TSQ scope and observing the scan signal (TSQ

scope set for multiple mass marker display) with a fast time base. Long

lived events were monitored with the TSQ data system operated normally,

but with the centroid sampling interval set at less than the normal 200

us. The centroid sampling interval is the number of ADC readings (1 to

8) summed to produce a single datum point. Therefore, with an ADC rate












of 40 kHz, the centroid sampling interval can be varied, in 25 ps

increments, from 25 ps to 200 ps with the latter being the default

setting.

High performance liquid chromatograph. A high performance liquid

chromatograph was used for quantitation of test drugs in animal samples

and to determine the quantities of sample vaporized in the ion source by

the laser. A high pressure syringe pump (Model 4100, Varian Associates,

Palo Alto, Calif.), 6-port rotary injection valve (Valco Instrument Co.,

Houston, Tex.), and a fixed-wavelength (214 nm) ultraviolet absorbance

detector (Model 1203, Laboratory Data Control, Riviera Beach, Fla.) were

used with a 4.6 mm i.d. x 250 mm stainless steel column, packed with a

5-pm particle diameter octadecyl silica gel stationary phase (Biosphere-

ODS, Bioanalytical Systems, Inc., West Lafayette, Ind.) and maintained

at 52C with a circulating water bath (Neslab Instruments, Inc.,

Portsmouth, N.H.). The mobile phase was pumped at 80 mL/h, and

consisted of acetonitrile:tetrahydrofuran:water (28.5:0.5:71) with

potassium dihydrogen phosphate added at a concentration of 2.7 g L-1

Procedure

Laser system evaluation. The laser system was evaluated with

Coumarin 2 and Rhodamine 6G laser dyes (Eastman Kodak, Rochester, N.Y.).

The coumarin lased at about 450 nm, however, the intensity was low and

the useful life of the dye was only a few shots. Rhodamine 6G lased at

about 590 nm with high pulse energy and a single dye solution allowed












operation for several hundred shots. Pulse shapes were measured for

various laser cavity diameters at 16 kV, 18 kV and 20 kV by monitoring

the signal directly from the fast photodiode.


Laser-TSQ evaluation. The laser-mass spectrometer system was

evaluated with respect to short-lived and long-lived laser-initiated

processes. The short-lived events (ps to ms) were observed with an open

electron impact ion source at high vacuum, while long-lived processes

(ms to s) occurred with a closed chemical ionization source with reagent

gas present at greater than 0.35 torr. Phenytoin, a relatively

nonvolatile (m.p. 295 298C) drug, which was also used in the animal

experiments, was used in most of the system evaluations, since it was

readily available as the free acid and the sodium salt, and proved to be

a suitable test substance. A series of compounds (Appendix) having

various functional groups and molecular weights was also tested under

desorption chemical ionization conditions (110) in which only the source

block was heated in most cases, under normal direct probe conditions,

and under laser vaporization conditions to determine the versatility and

utility of the present system. In all cases, samples were inserted into

the source as solids packed into a depression in the stainless steel

probe tip or as residues from solutions evaporated on copper, nickel or

silver electron microscope grids (Ted Pella, Inc., Tustin, Calif.).

Preparation, collection and analysis of animal samples. To provide

animal tissue samples suitable for laser analysis, two adult male rats

were given a loading dose (150 mg/kg) of phenytoin intraperitoneally











75 min before decapitation. At the time of sacrifice, the animals were

ataxic and lethargic, indicating substantial drug absorption. Brains

and livers were removed and washed in isotonic phosphate buffer. One

set was frozen at -20C. The other set was fixed for 72 h in phosphate-

buffered glutaraldehyde (KC1, 0.0834 M; KH2PO4, 0.0127 M; K2HP04,

0.0846 M; 25% aqueous glutaraldehyde; 925 mL water). Samples from fixed

and frozen brains and livers were analyzed by high performance liquid

chromatography (111). Aliquots of homogenized weighed tissue samples

were deproteinized with acetonitrile. The aqueous acetonitrile phase

was washed with iso-octane and then poured into a tube containing

granular KC1. Vortexing caused the acetonitrile to be salted out of the

water, at the same time extracting the phenytoin and internal standard.

The acetonitrile layer was transferred to a clean tube, evaporated to

dryness, and the residue was dissolved in mobile phase for injection

into the HPLC system. Sections of the fixed tissues were frozen in

isotonic phosphate for later MS analysis.















CHAPTER IV

PRELIMINARY STUDIES AND INSTRUMENT CHARACTERISTICS



This chapter includes discussions of (1) the LAMMA experiments with

thin slices of rat brain tissue taken from the cobalt lesion site in one

animal and a corresponding site from a non-lesioned animal, (2) the

determination of the regional distributions of phenytoin and

phenobarbital in rat brains by triple quadrupole MS/MS, and (3) the

basic characteristics of the laser-TSQ system, including laser pulse

shape and variation of diameter of laser spot size with target position

and cavity diameter.



LAMMA Experiments



The LAMMA instruments, initially the Model 500 and, more recently,

the Model 1000, were designed ideally to monitor atomic mass spectra

during the brief lifetime of laser-induced emissions from thin film and

bulk samples. The time-of-flight (TOF) mass spectrometer operates

within the time restrictions of Q-switched, single shot lasers and the

lifetimes of the ions which they produce, so that a complete mass

spectrum covering a significant mass range can be acquired in less than

100 us. After the laser fires, kinetic energy is imparted to the ions

in the source area by a high voltage accelerating pulse. This is

47











translated into an ion motion according to eqn. 4.1,


(1/2)mv2 = zeV


(4.1)


where m and v are the ion mass and velocity, respectively, z is the ion

charge, e is the electron charge, and V is the accelerating voltage.

The time of ion flight is determined by the length of the flight tube

(L) and the ion velocity (eqn. 4.2).


t = Lml/2/(2zeV)1/2


(4.2)


Since m is equal to the molecular weight divided by Avogadro's number

(6.023 x 1023) and the voltage in electron volts (eV) is the

accelerating voltage times the charge per molecule (ze), for singly

charged molecules with time measured in seconds, eqn. 4.2 becomes,


t = 7.199 x 10-7(L/V1/2)m1/2


t = cmI/2


(4.3)


(4.4)


where c is the instrumental constant, incorporating unit conversion

factors, length of the flight path and instrument accelerating voltage.

Table II gives the masses, their square roots and abundances for the













Table II. Masses and Flight Times for Nickel, Molybdenum and Lead
Isotopes in LAMMA Calibrant


Isotope Ion Mass (u) Ion Massl/2 % Abundancea dpb


58Ni
60Ni
92Mo
94Mo
95Mo
96Mo
97Mo
98Mo
1OOMo
206pb
207pb
208pb


57.933
59.931
91.907
93.905
94.906
95.905
96.906
97.905
99.907
205.974
206.976
207.977


7.6115
7.7415
9.5868
9.6905
9.7420
9.7931
9.8441
9.8947
9.9953
14.352
14.387
14.421


68.3
26.1
14.8
9.3
15.9
16.7
9.6
24.1
9.6
24.1
22.1
52.4


101.5
106.4
176.4
180.2
182.1
184.1
186.1
188.0
191.9
356.8
358.0
359.0


tc dtd


32.579
33.089
40.385
40.781
40.979
41.188
41.396
41.594
42.001
59.187
59.312
59.416


a) % Abundance of naturally occurring isotopes.
b) dp = Chart distance from recorder start to peak in mm.
c) t = Total flight time in us.
d) dt = Total chart distance (dp + delay distance) in mm.


312.6
317.5
387.5
391.3
393.2
395.2
397.2
399.1
403.0
567.9
569.1
570.1











calibrant isotopes (Ni, Mo, Pb), and the distance and time data derived

from the calibration spectrum. The flight times (t) were calculated as

the sum of chart time (tc) plus the delay time (td) with



tc = nts(dp/D) (4.5)



and



td = ndts (4.6)



where n and nd are the total number of chart channels (2048) and the

number of delay channels, respectively, ts is the sampling interval, and

dp and D are the chart distance to the peak and total chart distance

covered by 2048 channels (393.0 mm at 0.020 ps sampling interval),

respectively. Regression analysis of flight time (ps) vs. the square

root of the ion mass indicated a 2.563 us lag (Y-intercept) between the

start of recorder timing and the ion acceleration pulse. The slope of

this plot (Fig. 4.1) gave the instrument constant, 3.944 ps u-1/2

Using this value with eqns. 4.3 and 4.4 gave a flight path length of

300.2 cm.

Figure 4.2 shows the calibration spectrum (A) with spectra acquired

from the lesion site (B) and the control site (C). Although the peak at

m/z 59 (*) is higher in the lesion site than in the control sample, it

is impossible to say if this signal is actually due to cobalt. The

presence of large peaks at almost all masses throughout this range


























Figure 4.1

Plot of flight time vs. the square root of ion mass for LAMMA
calibrant (data from Table II). Data points cover all isotopes
for each metal (Ni: 58, 60; Mo: 92, 94, 95, 96, 97, 98, 100;
Pb: 206, 207, 208).

















































6.0 9.0
SQUARE ROOT OF ION MASS


TIME (vs)


50.0





40.0-





30.0





20.0-





10.0.


3.0


12.0
























Figure 4.2

LAMMA spectra for (A) nickel-molybdenum-lead calibrant, (B)
section of rat brain from cobalt lesion site, and (C) section of
rat brain from corresponding site in control (unlesioned) animal
(m/z 59 designated by *).












precludes any accurate assessment of elemental or molecular composition.

These results point out the advantage of having two stages of mass

analysis (filtering and selection) for microprobe analysis. By

increasing the laser power density, the chemical noise from molecular

fragments could be decreased, but it would still be impossible to

definitely identify the ion at m/z 59 as cobalt since there is only one

naturally occurring isotope. Uncertainty of ion identification becomes

greater as mass increases because the number of possible elemental

compositions also increases. This deficiency is made obvious in a recent

report on LAMMA analysis of coal and shale samples (112). The laser was

fired at grazing incidence at 1/8" diameter chips, and ions produced

directly by the laser were recorded. Although the authors gave

explanations of the spectra and made mass assignments, primarily on the

basis of natural isotopic abundances and past literature reports of

laser pyrolysis gas chromatography and mass spectrometric studies of

homopolymers, a second stage of mass analysis would have added greatly

to the certainty of some of the assignments.



Regional Distributions of Phenobarbital and
Phenytoin in Rat Brain



At the time this project was started, very little quantitative work

on real biological samples had been done by MS/MS. A quantitative study

of typical samples amenable to analysis by other methods (i.e., gas-

liquid or high performance liquid chromatography) was considered to be a












necessary starting point, to define instrumental conditions,

particularly with respect to the CAD behavior of analytes, which would

give accurate, reproducible results. Since one of the primary

advantages of MS/MS is elimination of the need for pre-MS

(chromatographic) separation of mixture components, demonstration of

reliable analysis by direct vaporization of samples (extracts) from the

solids probe was also considered important.

Rapid vaporization of the entire sample from the solids probe vial

with as little temporal separation of analyte and internal standard

evaporation profiles as possible was found to be of primary importance.

Separation of the analyte and internal standard ion signals (because of

differences in their vaporization temperatures) caused the response

ratio to be dependent on the number of scans over which the signal was

summed. To minimize this effect, a heating rate of about 100C s-1 was

employed. This generally resulted in complete sample vaporization

within a 30 s period with a near-Gaussian ion current profile. Analyte-

internal standard pairs vaporized simultaneously within about 20 s (Fig.

4.3).

Response ratios for drug and internal standard ion currents were

also dependent on instrument tuning, so that long-term (day to day)

reproducibility of calibration curve slopes was poor; however, linearity

was good, with correlation coefficients generally between 0.996 and

0.999. Figure 4.4 shows a histogram of the results for phenobarbital

and phenytoin in plasma and various brain regions. The numbers in

parentheses beside each bar give the number of samples (before slash)

























Figure 4.3

Plots of ion current vs. time for direct solids probe
vaporization and selected reaction monitoring of rat brain
extract from an animal that received phenobarbital (m/z 233 to
m/z 162) and phenytoin (m/z 253 to m/z 182). Other signals are
from internal standards, p-methyl phenobarbital (m/z 247 to m/z
191) and p-methyl phenytoin (m/z 267 to m/z 196). RIC is
recombined ion current.









ION CURRENT (CNTS)
607 1 233' 162+


23201 253+ -182+


241281 247+ 191


30400 267+ 196


27601 233+ 233


249281 247+ 247+



845] 253+ 253+


51681 267+ 267


490241 RIC -


10.0 20.0 30.0 40.0
TIME (s)


















Figure 4.4

Histogram showing concentrations of phenytoin and phenobarbital
in various areas of rat brain (ug/g) and in plasma (ug/mL).
Number before slash indicates the number of samples taken and the
number after slash shows the number of animals from which those
samples were obtained. See caption Fig 4.3 for reactions
monitored during analysis.













(4/
PLASMA
(2/2)

CEREBELLUM (10/4)
(GRAY MATTER) (6/2)


CEREBELLUM (9/4)
(WHITE MATTER) (5/2)


SENSORIMOTOR (6/3)
CORTEX (4/2)

(6/3)
BASAL GANGLIA
(4/2)

(6/3)
HIPPOCAMPUS
(4/2)


CORPUS (6/3)
CALLOSUM (2/2)


PONTINE (5/4)
(WHITE MATTER) (3/2)


OCCIPITAL (2/1)
CORTEX (2/1)
I I I I I I I


PHENOBARBITAL

PHENYTOIN


L I


10 20 30 40 50
CONCENTRATION (pg/g OF TISSUE)
(pg/mL OF PLASMA)


4)


60 70












taken from the number of animals (after slash) sacrificed. Regional

distribution of phenobarbital is relatively even. Although phenytoin

appears to be slightly more concentrated in white matter and less

concentrated in basal ganglia and hippocampus, it also is fairly evenly

distributed. Concentrations in structures that are possibly linked with

epileptic activity or drug effects (hippocampus, cerebellum,

sensorimotor cortex) are not greatly different from those in areas that

are not linked with seizures (pontine white matter, occipital cortex).

The averages of the histogram values for all structures are

38 + 2 ig g-1 and 11 + 3 jg g-l for phenobarbital and phenytoin,

respectively; brain to plasma ratios for these averages are 0.568 and

1.73, respectively.

The low relative standard deviations for these samples, 5.6% for 50

phenobarbital determinations, and 20% for 30 phenytoin samples, indicate

that monitoring analyte daughter ion intensities for selected reactions

with respect to those of suitable internal standards is a valid

technique for direct analysis of plasma and brain extracts by triple

quadrupole mass spectrometry. The response ratios are made more

reproducible by rapid, complete vaporization. This was assured by rapid

heating of the solids probe in this experiment. In the instrument

under consideration, rapid heating and sample vaporization would be

affected by using a pulsed laser as a sampling device.










Laser-Triple Stage Quadrupole Characteristics



Variation of Focal Spot Size with Distance from Focal Point and

Laser Cavity Diameter.

The laser beam, which is about 20 mm in diameter when emanating

from a totally open cavity, was reflected from a flat mirror 24 inches

in front of the output coupler to the 1.5 inch diameter fused silica

plano-convex lens used as the laser window in the TSQ. The target was a

black piece of cardboard that was placed from 12 mm behind the focal

point to 18 mm in front of the focal point. Spot size varied from

3.2 mm at the former to 5.0 mm at the latter and was 1.3 mm at the focal

point (Fig. 4.5). Spot size initially increased more slowly with

distance in front of the focal point. The cavity diameter was varied by

placing an iris diaphragm between the dye cell and the output coupler.

The target was located 4 mm in front of the focal point of the lens.

Spot size was 1.7 mm, 1.0 mm and 0.5 mm with iris openings of 18 mm,

12 mm and 6 mm, respectively. Below 6 mm the spot size was too small to

be measured accurately, but a flash could still be seen at the target

with the cavity diameter set to 2 mm. Normal cavity diameters used in

this study were 6 mm to 10 mm, depending on the age of the flashlamp and

laser dye solution. Focal spot sizes for these settings were 0.5 mm to

1 mm.

























Figure 4.5

Variation of focal spot diameter with distance in front of (+,
toward) and in back of (-, away from) focal point for 1.5 in.
diameter, 12.5 cm focal length fused silica lens used for
focusing laser in ion source of TSQ.











Laser Pulse Shape and Duration.

Laser pulse shape and duration were measured using the output from

the fast photodiode to drive the storage oscilloscope directly. Figure

4.6 shows the tracings produced by varying the cavity diameter (2 to

12 mm) at 16 kV, 18 kV and 20 kV. In all cases, the light pulse was

reflected from two microscope slides to the photodiode so that all

pulses remained on the oscilloscope scale (5 V cm-1). Therefore, only

about 0.6% of the laser light was incident on the photodiode (8%

reflected from each slide). The time base of the scope was set at

2 ps cm-1, indicating a pulse width (full width at half maximum) varying

between 1 ujs and 3 us and depending on cavity diameter and driver

voltage. At 18 kV and cavity diameters of 6 mm to 10 mm (normal

settings), pulse width was 1.5 ps to 2 ps.

































Figure 4.6

Laser pulse shapes from photodiode signal as monitored by storage
oscilloscope at driver voltage of (A) 16 kV, (B) 18 kV, and (C)
20 kV. Cavity diameter was varied from 2 mm to 12 mm in 2 mm
increments giving curves running from low intensity to high
intensity in each figure (drawn from photographs of storage
oscilloscope screen).








67








V

A

20



o10






2 4 6 8
TIME (us)



B

20



10



0

2 4 6 8
TIME (Is)






20



10



0

2 4 6 8
TIME (ps)
















CHAPTER V

SHORT-LIVED PROCESSES



Although the quadrupole mass filter scans the fastest of all the

scanning mass analyzers, normally covering 1000 to 2000 mass units each

second, it is still relatively slow when compared to the ion lifetime of

pulsed laser-generated ions. In spite of this incompatibility, an

attempt to observe the events immediately following the laser pulse was

thought to be desirable for this study for two reasons. First, the

laser used in this study was a single shot device (1.5 to 2 ius

pulsewidth) and prolonged signals were not expected, and second,

observation and verification of normal laser-induced processes such as

cationization required mass discrimination and time resolution in the us

range. This chapter will present results of experiments conducted using

an electron impact ion volume, modified to transmit the laser light, and

operated without reagent gas. Under these conditions, ion signals

lasted less than 1 ms and were monitored either with a storage

oscilloscope or with the TSQ data system. Limited mass discrimination

was maintained by selecting the mass of interest with Q1 and, with no

collision gas in Q2, selecting a range of masses which may or may not

include the mass of interest with Q3 operated in RF-only mode.











Triple Stage Quadrupole Control



The goal in these experiments was to have the TSQ operate as a

time-of-flight mass spectrometer for the selected ion, allowing data to

be obtained at a single m/z during the course of an experiment. This

had been accomplished in a single stage quadrupole instrument by Hardin

and Vestal (80), who used a pulsed laser with single ion counting.

Setting the analyzer to pass the selected ion mass, and increasing the

instrument resolution until only one event was recorded for each laser

pulse, they collected data until a statistically significant

distribution of ion flight times was acquired for each mass of interest.

The TSQ is electronically configured to operate with at least one

analyzer quadrupole scanning a range of masses (although that range may

be 1 u wide) in all modes. A system having both quadrupoles set to pass

a single mass had not been defined in programmable read only memory

(PROM). Since the laser used in this experiment was a single shot

device, the instrument had to be set up to provide a continuous signal

for the ion selected, with time resolution of a few microseconds. This

could have been done either by determining and adding the appropriate

instruction or instructions to the PROM to allow each analyzer to be set

to pass a single mass without scanning, or by causing the scanned

quadrupole to operate in the RF-only mode to allow a range of masses to

pass simultaneously rather than sequentially. Since modifying the PROM

could have had far-reaching and unforeseen effects on other parts of the

TSQ system, elimination of the DC voltage from the scanned quadrupole











was selected as the best way to solve the problem. This was

accomplished by changing the rod polarity switch located at the rear of

the quadrupole electronics module (QEM) from "+" to "off". This is a

three position switch ("+", "off", "-") that allows the operator to

change the DC polarity on the quadrupole rods. Under normal conditions,

there is no difference in response if the switch is set to "+" or "-"

The "off" setting removes the DC component from the rods. The range of

masses passed under these conditions can be determined from the

stability diagram for the quadrupole. The stability diagram (Figure

5.1) is a plot of the ratio of DC to RF voltages imposed on the

quadrupole rods vs. a function including the RF frequency, a dimensional

constant for the quadrupole (one-half the diagonal distance between the

rods in cm), the RF voltage, and the inverse of the mass-to-charge ratio

of the ion. Under normal conditions, the voltage ratio is set so that

only the mass (+0.5 u) under the apex has a stable trajectory through

the quadrupole. All other ions have unstable trajectories and are lost

by collisions with the rods. If the DC voltage component is removed,

all ions that fall under the curve have stable trajectories and will

pass. The ratio of the ordinate value under the apex (1/ml) to that at

the low-mass abscissa zero (1/m2) is about 7/9. Therefore, if an RF-

only quadrupole is set to pass ml, it will actually pass all masses

greater than m2 which equals 7/9 of mi.

Table III gives the effects on response for the most pertinent

combinations of operation mode (parent and daughter ion experiment),

programmed RF function (on/off) and rod polarity switch (+/off). Other



































Figure 5.1

Stability diagram for quadrupole mass filter. Plot of DC to RF
voltage ratio vs. a function proportional to (m/z)-l. Shaded
region shows range of ion masses passed through quadrupole set to
pass mi with both RF and DC voltages imposed (normally, mi 0.5
u). The RF-only quadrupole will pass all masses higher than m2
where m2 = (7/9)m1. Parameters in the ordinate function are: RF
voltage, Vrf; quadrupole rod spacing (diagonal), ro; electron
charge, e; angular frequency of Vrf, w ; mass-to-charge ratio,
m/z.
















Table III. Effects of Programmed RF
During Various TSQ Operational Modes


Condition


Function and Rod Polarity Switch Settings


Operation Programmed Rod Polarity
Mode RF Switch
Q1 Q3


Effects


1 Daughter
Spectrum


2 Daughter
Spectrum

3 Daughter
Spectrum


Parent
Spectrum


Parent
Spectrum


on off


6 Parent
Spectrum


+ Daughter spectrum. Increasing Ql first
mass removes low m/z (all above 7/9 Q3
first mass passed).

off High broad signal. Low m/z and signal
intensity controlled by Q1 first mass.

off Broad peak displayed only when Q1 first
mass set on peak.

+ Parent spectrum. Increasing Q3 first mass
removes low m/z (all above 7/9 Ql first
mass passed).

+ Broad signal. Signal intensity 75% of
that with Ql rod polarity +.

+ Broad peak displayed only when Q3 first
mass set on peak. Signal intensity less
than that in daughter mode.











combinations were not listed because they were unremarkable or had no

bearing on the present discussion. The programmed RF function removes

the DC component from the unscanned quadrupole. The effects were

observed with the instrument under manual control with perfluoro-

tributylamine (FC43) being introduced continuously into the El source.

Effects that include a measurement of relative signal intensity for the

rod polarity switch on "+" or "off" were made by observing m/z 69 on the

instrument oscilloscope. Loss of mass resolution is expected when the

scanned quadrupole is operated in RF-only mode. Condition 3 was selected

for most of the work on short-lived processes because Q1 provides mass

discrimination and Q3 can be set up to pass or not pass the ion selected

by Q1. The advantage to this condition is that a long scan (20 s) can

be acquired to eliminate the delay time of a few milliseconds normally

required between scans. Further, since Q3 is not resolving masses, this

mode will provide an ion intensity vs. time response for the selected

ion rather than an ion intensity vs. mass response. In addition, the

beginning of the scan can be detected by an oscilloscope attached to the

"Y+" input of the TSQ oscilloscope. This is important for monitoring

fast processes with the data system because data acquisition begins at

the first regular scan after the experiment is initiated at the computer

terminal. Data acquisition does not begin immediately because the scan

sequence is not interrupted and restarted. Condition 6 is analogous to

condition 3 but Q1 is the scanned quadrupole (parent mode). Signal

intensities were generally lower for this condition, however, the major












drawback was that the scan signal was not visible at the "Y+" input as

it was for the daughter mode.



Comparison of Storage Oscilloscope and Data System



Signal for Ni+ from Nickel EM Grid

With the TSQ operating in daughter mode with Q3 rod polarity switch

off (condition 3), Q1 passing m/z 58 (Ni+) and Q3 passing all ions above

m/z 56 (scanning from m/z 71 to 72 in 20 s with rod polarity switch

off), the laser was fired at a clean nickel electron microscope grid

inserted into the ion source on the 90' probe tip. The storage scope

was attached directly to the electron multiplier through a high-gain

differential amplifier module (Type D, Tektronix, Inc.) operated at

1 my cm-1 with a time base of 50 jis cm-1. In Figure 5.2.A, traces 1 and

2 show response with the laser light blocked from the ion source and the

light admitted to the source, respectively. Trace 3 was obtained by

switching Ql to pass m/z 90 and Q3 to pass all masses above m/z 140

(scanning m/z 180 to 181 in 20 s), so that Ql and Q3 would be mutually

exclusive and neither quadrupole should pass any ion produced by

vaporization of the nickel target. In figure 5.2.B, traces 2 and 3 are

analogous to traces 2 and 3, respectively, in part A, except that the

response was monitored with the TSQ data system by acquiring in profile

mode. Also, Q3 was scanned from m/z 60 to 70 (passing all m/z above 47

to 54) for trace 2 and from m/z 170 to 180 (passing all m/z above 132 to

140) for trace 3. These data show that this method of recording





























Figure 5.2

Plots of total ion current vs. time showing (A) storage scope
data and (B) profile data for nickel EM grid target. In A, trace
1 is the response with laser light blocked from ion source;
trace 2 in A and B is the response with Ql passing m/z 58 and Q3
passing all m/z above 56 (A) and all m/z above 47 to 54 (B).
Trace 3 in A and B is the response with Q1 passing m/z 90 and Q3
passing all m/z above 140 (A) and all m/z above 132 to 140 (B).
See text for explanation of differing Q3 scans for storage scope
and data system operation. Ion currents are given as counts
(CNTS).




























1


I I
0.1 0.2
TIME (ms)


I3
0.3 0.4


TOTAL
ION CURRENT (CNTS)


4000













2000


0.2
TIME (ms)


0.4













intensity vs. time data for fast events may be valid, that quadrupole 1

provides some mass discrimination, even for low mass elemental ions, and

that the storage scope and the data system give equivalent data. There

is always a minimum signal (rapidly decaying spike) observed by the data

system even when no sample is in the ion source, the laser light is

blocked before it reaches the source, and the electron multiplier is

off. This response is probably due to electrical interference from

discharge of the laser storage capacitor when the flashlamp is fired.

Instrumental Acquisition and Display Characteristics

It might be helpful at this point to describe the TSQ scan

parameter (SYSTem, System Status, Multiple Ion Detection), calibration

(CALIrate), acquisition (ACQUire), and display (PROFile, mass

CHROmatogram, and SPECtrum) programs. The above programs will,

henceforth, be referred to by the abbreviations used to access them

(underlined). Data can be acquired in two modes, centroid and profile.

Both are calibrated using the CALI program which establishes a linear

relationship between mass and time elapsed since start of a scan.

Therefore, even under normal circumstances, the mass of an ion is

determined by the appearance time after the scan starts, and the

spectrum is displayed as ion intensity vs. mass. Acquisition in profile

mode causes storage of all points detected at the maximum sampling rate

(25 js). Storage requirements are maximized, but a relatively

continuous plot of ion intensity vs. time is produced, allowing

observation of peak shapes. It is this mode that has been used in these

experiments on short-lived processes (Chapter V) to obtain a plot of









79


signal intensity vs. time. A threshold intensity can be chosen (in ADC

units), which allows immediate rejection of points having intensities

less than the selected threshold. Thresholding is the only decision

made by the data system at the full 40 kHz ADC rate. A setting of 1

means all data is recorded (no thresholding). For these experiments,

the threshold was routinely set at 2 to eliminate spurious electronic

noise. Therefore, a scan 20 s long, during which a signal is produced

for a brief instant by a laser shot, would result in a profile lasting

only from the time the signal exceeds threshold until it goes below

threshold. PROF is the program used to display profile data. Although

the program converts time to mass units, it plots essentially a time-

dependent, ion intensity curve, and calculates the time the signal is

above threshold. The plots of profile data shown in the figures were

calculated from these curves. Since time scales were usually different

for each profile curve, signal intensities were replotted on a common

time scale for comparison. The maximum intensity in a profile plot is

4096 counts, as determined by the 12-bit ADC. Q3 scan parameters were

most easily set using the MID program. Q1 mass was set using the SS

program. In this way, experiments could be run rapidly by writing the

required MID descriptors, and calling them as needed from SS. Q1 mass

was changed as required, also from SS.

Acquisition in centroid mode (used for experiments on long-lived

processes in Chapter VI) causes calculation and storage of the mass

centroid of a peak, minimizing storage requirements and making data

manipulation faster. In this mode, the centroid sampling interval is












chosen (using MID) in 25 ps increments between 25 js and 200 ps to give

a reasonable number of data points per mass peak. Therefore, data is

acquired at 40 kHz, and 1 to 8 data points are summed to give a single

datum point to be used in calculating the peak centroid. If a scan is

set up in SS, the highest possible centroid sampling interval (200 ps)

will be used. In SYST, a selection is available for centroid samples

per peak, however, the number actually used will be determined by the

mass range covered in the scan time. The only way to insure use of a

particular centroid sampling interval is to set that interval in MID.

The CHRO program is used to view the recombined ion current (RIC), or

the intensity profiles for each ion with respect to scan number (time).

This is analogous to the use of the PROF program for viewing short-lived

processes; however, an additional dimension is available (in the SPEC

program), which allows display of the mass centroids making up the RIC

(mass spectrum). Scan parameters for acquisition in centroid mode

(Chapter VI) were again set up using the MID program, both for normal

mass spectra and, as required, for parent and daughter ion spectra.

The centroid sampling interval was chosen so that scan time could be

minimized, and there would be at least 5 points/mass peak for

calculation of the peak centroid.

Filtering Characteristics of Quadrupoles 1 and 3

Although the quadrupole is supposed to effectively filter out all

ions of unselected mass regardless of ion energy, it is possible for

ions of inappropriate mass to pass through the quadrupole field without

deflection if their kinetic energy is high enough (109, 113, 114).












Therefore, a series of experiments was performed to determine the

filtering capabilities of the system under the conditions used for time

profile monitoring.

Figure 5.3.A and B show a comparison of storage scope and profile

data, respectively, for phenytoin. Ql was set to pass m/z 275, (M+Na)+

for phenytoin, and Q3 was scanned from m/z 343 to 353 in 18 s with the

rod polarity switch off (passing all m/z above 266 to 274). The

phenytoin (4 pL of a solution of 4.35 mg/mL) was layered over a Na2C03

residue (2 uL of a solution of 25 mg/mL) on a 400-mesh copper electron

microscope (EM) grid which was inserted into the ion source on the 90

probe tip. Trace 1 shows the first shot. Trace 2 is a second shot at

the same site. Although this figure shows an intense signal for

cationized phenytoin and significant reduction for a repeat shot (the

grid was totally disintegrated at the point of impact of the first shot

creating a hole about 0.5 mm in diameter), the signal is not necessarily

due to cationized phenytoin alone. Figure 5.4 shows responses for a

similar sample for which Ql was set to pass m/z 150 and Q3 was set to

pass all m/z above 202 to 210 (scanning m/z 260 to 270 in 18 s with rod

polarity off). In this case Q1 and Q3 are mutually exclusive and Q1 is

not passing an ion that should be produced by the sample; however, Q3 is

covering a mass range that includes ions that should be present. Trace

1 shows a shot at the center of the grid where the phenytoin

concentration, due to uneven evaporation of the methanol, was lowest.

Solids were usually concentrated in a ring describing the edge of the

sample droplet. Traces 2 through 4 show successive shots at the same



































Figure 5.3

Plots of total ion current vs. time showing (A) storage scope
data and (B) profile data for phenytoin over Na2CO3 on copper EM
grid. In A and B Q1 is passing m/z 275 and Q3 is passing all m/z
above 267 to 275 (scanning m/z 343 to 353 in 18s with rod
polarity switch off). In both A and B, trace 2 is the response
from a second laser shot at the same site that gave trace 1.






























0.2 0.4 0.6 0.8
TIME (ms)


TOTAL
ION CURRENT (CNTS)
B

4000












2
2000 -


0.4 0.8
TIME (ms)
































Figure 5.4

Plots of total ion current vs. time showing the efficiency of Ql
and Q3 mass filtering (profile data) for phenytoin over Na2CO3 on
copper EM grid. In traces 1-4, Q1 is set to filter out all m/z
produced by the sample (passing m/z 150) and Q3 is set to pass
ions that should be present (passing all m/z above 202 to 210.
In trace 5, neither quadrupole is passing ions produced by the
sample and they are set to be mutually exclusive (QL passing m/z
700 and Q3 scanning normally with rod polarity "+" from m/z 300
to 310. Trace 1 is at grid center where phenytoin concentration
is low. Trace 2 is at edge of droplet where concentration is
highest. Traces 3 and 4 are successive shots at site 2. Trace 5
is from an identically prepared grid at the edge of sample
droplet.



















TOTAL
ION CURRENT (CNTS)













400











200


0.4


TIME (ms)












site on the edge of this ring where concentration was highest. These

traces indicate that Q1 is not totally efficient at filtering ionic

species. Trace 5 was recorded with Ql set to pass m/z 700 and Q3

scanning from m/z 300 to 310 with rod polarity "+" (DC on) so that both

quadrupoles were mutually exclusive and neither was passing ions that

were produced by the sample. A minimum signal, generally present under

any conditions, was produced. This indicated that Ql and Q3 effectively

filtered out sample ions when set to pass masses that differed greatly

from the sample ion masses. Another experiment was conducted to

determine whether the filtering capabilities of Ql were lessened when

the selected mass was close to the mass of an ion produced by the sample

(i.e., is unit mass resolution maintained under the conditions produced

by laser desorption in the present ion source). Ql was set to pass m/z

275 and then, with a new grid, m/z 270 (test point arbitrarily chosen

5 u away from the ion mass), with Q3 passing all m/z above 202 to 210.

The signal intensity ratio for the former with respect to the latter was

about 5, indicating that unit mass resolution in Ql is not maintained

under these conditions. The particles that are causing these unwanted

responses may be fast ions or neutrals which do not "see" the low

voltage fields of the lenses and quadrupoles. Discrimination of a

selected ion mass from its nearest neighbors is less efficient than from

ions that have masses much greater or much less than the selected mass.

This problem was approached from another angle to show the total

ion signal with all quadrupoles operated in RF-only mode, and to

contrast that with the signal obtained when QI was operated as a mass













filter. Q1 was set to pass m/z 275 and Q3 to pass all m/z above 266 to

274 with the sample again phenytoin over Na2CO3. Successive shots were

fired at the same site (Figure 5.5.A, traces 1-3). Ql was then set to

pass m/z 350 and the programmed RF was turned on so that Q1 would

actually pass all m/z above 272. Two shots were fired at the same site

(new grid; Figure 5.5.B, traces 1 and 2). Programmed RF was then turned

off so that Ql would then only pass m/z 350 (an ion not produced by the

sample) and a single shot was fired at a new grid (Figure 5.5.B, trace

3). These figures show that Q1 does actually filter out a large portion

of the total signal that is possible under these conditions. Therefore,

the combination of quadrupoles 1 and 3 can be used to obtain limited

mass resolution in this rapid acquisition mode to allow monitoring of

the fast signal profiles.

Time Resolution of Low-Mass and High-Mass Ions

With these limitations in mind, an attempt was made to observe the

resolution of two compounds by time alone by operating the instrument in

a time-of-flight mode without mass discrimination in either analyzer

quadrupole. Q1 was set to pass m/z 160 with programmed RF on (actually

passing all m/z above 124), Q3 was set to pass all m/z above 117 to 124

(scanning m/z 150 to 160 in 18 s with rod polarity off). Data were

acquired with the TSQ data system in profile mode. The sample was a

mixture of 4-vinyl-4-aminobutyric acid (vinyl gaba, mol. wt. 129) and

reserpine (mol. wt. 608) over Na2CO3. Expected cationized species were

m/z 152 and m/z 631, respectively. All quadrupole offset voltages and

lens voltages were set at -20 V. The flight path length from the ion




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EEVWPEPGW_9GR0F5 INGEST_TIME 2017-07-20T21:25:47Z PACKAGE UF00099340_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

CHARACTERISTICS AND APPLICATION OF A LASER IONIZATION /EVAPORATION SOURCE FOR TANDEM MASS SPECTROMETRY by ROBERT JOHN PERCHALSKI 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 1985

PAGE 2

To my parents with love.

PAGE 3

ACKNOWLEDGMENTS I wish to express my sincere appreciation to Dr. Richard A. Yost for his direction and faith in this project. I am also grateful to the members of my committee: Dr. James D. Winefordner whose initial advice should have been given more serious consideration and whose instrumentation helped to make this project more successful; Dr. B.J. Wilder whose continuing personal and professional support over the last 13 years has been invaluable in the attainment of this goal; Dr. John G. Dorsey who offered advice and encouragement; Dr. L. James Willmore who offered counsel and expertise; and Dr. John Eyler who took over where Dr. Willmore left off. I also appreciate the generosity of Dr. Martin Vala who loaned us the laser. Thanks are also due to Drs. Neal Brotherton and Dean Fetterolf who got me started, and to the members of the research group who freely gave advice and assistance. Special thanks are due to Dr. Jodie V. Johnson who kept the instrument running, and to Dr. Edward Voigtman for help with the laser and the storage oscilloscope experiments. The LAMMA experiments were completed through the kindness of Dr. David M. Hercules of the Department of Chemistry, University of Pittsburgh. The assistance of Dr. Edward J. Hammond in the animal experiments and Dr. William Ballinger in the preparation of samples for LAMMA analysis is appreciated. iii

PAGE 4

This project could not have been completed without the help and cooperation of my children, Matthew, Abby, Ben, and David. Finally, for her support and patience, I thank my wife, Lynn.

PAGE 5

TABLE OF CONTENTS CHAPTER PAGE ACKNOWLEDGEMENTS iii LIST OF TABLES vii LIST OF FIGURES viii ABSTRACT xii I INTRODUCTION 1 Overview of Current Technology 1 Overview of Present work 13 II CHARACTERISTICS OF LASER-INDUCED PROCESSES IN MASS SPECTROMETRY 15 Effects of Absorption of Laser Radiation at Solid Surfaces 16 Laser-Induced Vaporization and Particle Emission 18 Analytical Usefulness of Laser Desorption Mass Spectrometry 25 III EXPERIMENTAL 27 Laser Microprobe Study 28 Preliminary Triple Quadrupole Mass Spectrometry Study: Regional Distribution of Phenobarbital and Phenytoin in Rat Brain 30 Laser-Triple Stage Quadrupole 35 IV PRELIMINARY STUDIES AND INSTRUMENT CHARACTERISTICS 47 LAMMA Experiments ^7 Regional Distributions of Phenobarbital and Phenytoin in Rat Brain 55 Laser-Triple Stage Quadrupole Characteristics 62

PAGE 6

CHAPTER PAGE V SHORT-LIVED PROCESSES 68 Triple Stage Quadrupole Control 69 Comparison of Storage Oscilloscope and Data System 75 VI LONG-LIVED PROCESSES 99 Long-Lived Signal Profiles 100 Quantitative Analysis and Application of Laser Induced Desorption Chemical Ionization 147 VII CONCLUSIONS AND FUTURE WORK 170 Conclusions Based on the Present Work 170 Future Work 178 APPENDIX CHARACTERISTICS OF DIRECT PROBE, DESORPTION CHEMICAL IONIZATION, AND LASER-INDUCED DESORPTION CHEMICAL IONIZATION BEHAVIOR OF TEST COMPOUNDS 183 LITERATURE CITED 185 BIOGRAPHICAL SKETCH 194

PAGE 7

LIST OF TABLES TABLE PAGE I Common Operational Modes of the Triple-Stage Quadrupole Mass Spectrometer 33 II Masses anU Flight Times for Nickel, Molybdenum and Lead Isotopes in LAMMA Calibrant 49 III Effects of Programmed RF Function and Rod Polarity Switch Settings During Various TSQ Operational Modes 73 IV Weight of Phenytoin Removed from Copper Grid by a Single Laser Shot 126 V (M+H)+ Ion Intensity Ratios for LIDCI of PHT and and PHT* Deposited with Parlodion on Copper Grids. Comparison of 2 Shots at Different Sites on Each of 5 Grids 130 VI Graded Melting Point Standard Solution Containing Compounds Having Melting Points from 76°C to Above 300°C 131 VII Comparison of Maximum and Summed Ion Currents for DCI of Graded Melting Point Stock Dilution from Copper, Nickel, and Silver EM Grids 135 VIII Calibrant Solutions Made from Graded Melting Point Standard for DCI /CAD from Nickel and Teflon Substrates 152 IX Phenytoin Calibrant Solutions for LIDCI from Nickel Grids 158 X Coefficients of Determination for DCI calibration with Nickel Grids and Teflon Paddle, and for LIDCI Calibration with Nickel Grids 161 XI Phenytoin Concentrations in Rat Brain and Liver Samples and in Fixative Solutions 163

PAGE 8

LIST OF FIGURES FIGURE PAGE 3.1 Analyzer section of Finnlgan TSQ mass spectrometer 32 3.2 Diagrams for machining blank side flange of TSQ for fused silica lens, and spacers for adjusting focal point of lens within ion source 37 3.3 Diagrams for machining 90° and 45° stainless steel solids probe extensions 40 3.4 Layout of lasers and beam directing optics in relation to TSQ ion source 42 4.1 Plot of flight time vs. the square root of ion mass of LAMMA calibrant 52 4.2 LAMMA spectra for nickel-molybdenum-lead calibrant, section of rat brain from cobalt lesion site, and section of rat brain from in control animal 54 4.3 Ion current profiles for direct solids probe vaporization and and selected reaction monitoring of phenobarbital and phenytoin in rat brain extract 58 4.4 Concentrations of phenytoin and phenobarbital in various areas of rat brain (>ig/g) and in plasma (>ig/mL) 60 4.5 Variation of focal spot diameter with distance in front of and in back of focal point 64 4.6 Laser pulse shapes from photodiode signal as monitored by storage oscilloscope at various driver voltages and laser cavity diameters 67 5.1 Stability diagram for quadrupole mass filter 72 + 5.2 Plots of total ion current vs. time for Ni (m/z 58) showing storage scope data and profile data for nickel EM grid target 77

PAGE 9

FIGURE PAGE 5.3 Plots of total ion current vs. time for cationized phenytoin (m/z 275) showing storage scope data and profile data for phenytoin over Na2C03 on copper EM grid 83 5.4 Plots of total ion current vs. time showing the efficiency of Ql and Q3 mass filtering (profile data) for phenytoin over Na2C03 on copper EM grid 85 5.5 Plots of total ion current vs. time showing efficiency of Ql mass filtering and effects of programmed RF function in Ql 89 5.6 Plot of total ion current vs. time for 4-vinyl-4aminobutyric acid (mol wt 129) and reserpine (mol wt 608) over Na2C03, showing time resolution of respective cationized molecules with m/z 152 and m/z 631 92 5.7 Plots of total ion current vs. time for phenytoin over Na2C03 showing effects of reagent gas on signal for cationized molecule 95 5.8 Plot of total ion current vs. time for phenytoin in parlodion film (no Na2C03) with auxiliary CI (m/z 253) 97 6.1 Plot of reconstructed ion current vs. time for laser evaporation of carbamazepine epoxide (mol wt 252) in CI source 102 6.2 Plots of ion current vs. time for laser vaporization of tetrabutylammonium perchlorate, menthol glucuronic acid, adenosine , and acetazolamide 105 6.3 Plots of RIC vs. time and solids probe temperature vs. time for desorption chemical ionization of phenytoin from copper grid floating in CI source at the electron entrance aperture 109 6.4 Plots of RIC vs. time and solids probe temperature vs. time for direct probe evaporation from a glass solids probe vial, and desorption chemical ionization from a 400 mesh nickel EM grid of hydergine 112

PAGE 10

FIGURE PAGE 6.5 Plot of RIG with respect to position of sample in CI ion source for phenytoin on copper grid 115 6.6 Plots of RIG vs. time for DCI of sodium phenytoin with filament turned on at various times 118 6.7 Plots of RIG vs. time and solids probe temperature vs. time showing DCI and LEDGI of phenobarbital 122 6.8 Plots of RIG vs. time for a mixture of phenytoin and l,3~l%-2^^G-phenytoin with parlodion evaporated on copper grids 128 6.9 Plots of ion current vs. time for DCI of graded melting point stock dilution from copper, nickel and silver grids 134 6.10 Plots of ion current vs. time for DCI of graded melting point standard (60-76 ng of each drug) from nickel EM grid, and Teflon probe extension 138 6.11 Plot of negative ion RIG vs. time for DGI of sodium phenytoin at various source temperatures 142 6.12 Plot of positive ion RIG vs. time for DCI of sodium phenytoin at various source pressures 144 6.13 Plot of ion current vs. time for LIDGI of components of graded melting point standard 146 6.14 Plots of RIG vs. time for DCI/GAD of 0.558 ng and 5.42 ng of hydergine on nickel grid showing dependence of signal lifetime on sample size 150 6.15 Plot of response ratio (analyte/internal standard) vs. amount of analyte on grid for DCI/GAD of 1 ;iL of graded melting point calibrant solutions on nickel grids 154 6.16 Plot of response ratio (analyte/internal standard) amount of analyte on grid for DGI/CAD of 1 }il of graded melting point calibrant solutions on Teflon paddle 156

PAGE 11

FIGURE PAGE 6.17 Plot of response ratio (PHT/PHT*) vs. amount of PHT on nickel grid for LIDCI/CAD of PHT calibrant solutions 160 6.18 Plots of ion current vs. time for phenytoin parent ion, major daughter ions and ion recorded at laser firing for LIDCI/CAD of rat liver tissue section 166 6.19 Daughter spectra of phenytoin (m/z 253) from laser shot for rat liver tissue section, and sample of pure PHT on nickel grid 169

PAGE 12

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 CHARACTERISTICS AND APPLICATION OF A LASER lONIZATION/EVAPORATION SOURCE FOR TANDEM MASS SPECTROMETRY By Robert John Perchalski May, 1985 Chairman: Richard A. Yost Major Department: Chemistry An instrument system is proposed that is based on tandem mass spectrometry (MS/MS) and a focused laser for vaporizing selected areas of samples within the ion source so that the molecular components of the sample can be identified and quantitated. This instrument is similar in principle to a commercially available laser mass spectrometer (LAMMA), but additional selectivity is possible with the proposed system because it has two stages of mass analysis, making molecular analysis feasible. The instrument used for this study consists of a triple quadrupole mass spectrometer and a single-shot, coaxial f lashlamp-pumped dye laser, interfaced through a simple system of mirrors and a lens, mounted near the ion source, which serves as a vacuum seal and focuses the laser radiation on the sample.

PAGE 13

The system is evaluated by observation of short-lived (< 1 ms) and long-lived (>1 ms) events after the laser is fired. Ion signals lasting several seconds are observed when samples are irradiated in a heated chemical ionization (CI) source. These signals are a combination of a desorption event and a vaporization event, and desorption CI is used as a model to characterize the laser-initiated processes. The latter are called Laser-Induced or Laser-Enhanced Desorption CI . The current system proved unsatisfactory for quantitative analysis because the time limitations of the instrument combined with a concentrationdependent ion signal lifetime produced calibration curves with low correlation coefficients. Qualitative confirmation of the presence of a drug in rat liver was obtained, however, after laser vaporization of a small sample of intact tissue within the ion source. Recommendations are made for overcoming the time limitations of the present system so that a microprobe capable of molecular analysis of heterogeneous and compartmentalized biological samples might be developed.

PAGE 14

CHAPTER I INTRODUCTION The purpose of this work was twofold: first, to design, construct, and evaluate a laser evaporation/ionization source for a triple quadrupole mass spectrometer, and second, to determine the feasibility of developing a millior microprobe, based on a microscopicallydirected sampling laser interfaced to a tandem mass spectrometer that is capable of molecular analysis of heterogeneous samples. Such an instrument could be used for chemo-morphological mapping of geological, metallurgical, environmental and astronomical samples, and of intact plant and animal tissues by direct, in situ elemental and molecular analysis. The major use of this instrument, however, is anticipated in the biomedical sciences, in which samples and their chemical constituents are highly compartmentalized from the organism to the ultrastructural level. Overview of Current Technology Much of the work of medical, agricultural, biochemical, and pharmacological research involves experiments designed to provide insight into the chemical and physical processes occurring at the cellular and subcellular levels. Elucidation of these processes is 1

PAGE 15

sought to define, in molecular terms, the normal chemical condition of organisms of interest from microscopic flora and fauna to man, abnormalities in this chemical condition that are evident as disease or deficiency, and effects and mechanisms of action of the various curative measures employed to produce and maintain an optimal state of being. The experiments are conducted in five general areas: morphology, physiology, pharmacology, histochemistry, and chemistry. Morphological studies relate shape and structure with physical function. Physiological and pharmacological studies define organism and substructure reaction to endogenous and exogenous (both therapeutic and anti-therapeutic) chemical species, respectively. Histochemical studies relate, usually indirectly and qualitatively, chemical species, and anatomical structure. Chemical analysis also relates chemical species with structure but has greater potential for quantitation through direct determination of the compounds of interest. Each of these methodologies, when used in this context, is site-specific at various levels of spatial resolution, and depends wholly or in part upon microscopy ( 1 , 2) . Methodologies for Physical and Chemical Analysis Morphology is generally the initial subject of interest and is pursued at three levels, delineated by the instruments used. Macrostructural surveys include all work done with the unaided eye and therefore are limited to gross anatomical correlations in large organisms and organ systems with dimensions measured in meters to millimeters. Microstructural studies are conducted with the light

PAGE 16

microscope in its various forms and can delve deeply into organ systems in large species or give information of more general nature for oneor few-celled organisms. Dimensions are measured in millimeters to micrometers. Ultrastructural investigations are conducted with the electron microscope and give information at the cellular and subcellular levels with dimensions normally measured in micrometers to nanometers. Physiology and pharmacology are usually restricted to microto macrostructural studies, since the probes used to measure or observe the chemical or physiological response are incapable of spatial resolution below the cellular level (5 to 10 ;im) in a living organism, or in a normally functioning tissue or cellular preparation. These studies include clinical evaluations of drug effects on animals and humans (pharmacological), and nutritional studies designed to evaluate effects of and requirements for biomolecules such as vitamins, amino acids, hormones, and transmitter substances necessary for proper nervous system function (physiological). In many cases, subject response coupled with chemical analysis of easily obtained body fluids (blood or urine) can give the information sought, since all foreign substances and most biomolecules present in an organism are in dynamic equilibrium with the circulatory system. At the lower limits of spatial resolution for these studies, test chemicals are usually applied by microiontophoresis (3) and responses are monitored by potentiometric (A) or voltammetric (5) microelectrodes . The gross chemical analyses which necessarily precede physiological studies and coincide with pharmacological studies can be carried out by

PAGE 17

usual methods (chromatography, ultraviolet, visible and infrared spectrophotometry, nuclear magnetic resonance and mass spectrometry, atomic absorption, emission and fluorescence spectroscopy) applied to physiological fluids and tissue homogenates. As the work progresses, however, to the stage at which a more exact localization of active substances is desired, more spatially selective techniques must be used which usually place more stringent requirements on the analytical techniques, particularly with regard to sensitivity. The techniques that are most used for localization studies are based on three general methodologies, histochemistry, autoradiography, and direct chemical analysis of microdissected tissue samples or subcellular fractions obtained by density gradient centrifugation. The histocheraical methods rely upon chemical, enzymatic or immunological reactions which involve the molecule of interest, an enzyme specific for the molecule of interest, an inhibitor of that enzyme, or an antibody created to bind specifically to the molecule of interest, as reagents to give, either directly or indirectly, a product visible with a light or electron microscope (6). The method is devised so that all reactions occur rapidly in succession, and the final visible product is synthesized at and does not diffuse from the location of the reactant of interest. Potential for qualitative analysis is predicated on the specificity of the chosen reagent, enzyme or antibody for the molecule of interest. Insuring this specificity presents the most challenging problem during method development. Quantitative analysis is possible by densitometry if normal stains are used, or by fluorometry if

PAGE 18

the final product is naturally or can be made fluorescent with a fluorogenic reagent or fluorescent-labeled enzyme or antibody (7-9). The introduction of raster beam microscopes, coupled with computerized data acquisition and imaging techniques, have contributed greatly to improving the accuracy of these methods (10). Autoradiography is a direct method of localizing radioactivelylabeled molecules of interest at the light and electron microscope levels (11). The system under investigation is sacrificed and sectioned at an appropriate time after the labeled compound is introduced. The sections can be treated as usual for the microscopic method of choice (staining) and are then coated with a thin layer of silver halide emulsion (gelatin). Resolution is determined by the homogeneity of the emulsion, the size of the particles, and the thicknesses of tissue and emulsion layers. After exposure of the emulsion for a suitable period of time, which depends upon the half-life of the nuclide (normally several weeks to a few months), the emulsion is developed and the sample is subjected to microscopy. Direct chemical analysis of anatomical structures isolated by microdissection or density gradient centrif ugation can sometimes be performed using scaled-down versions of normal analytical methods. Typical methods for microdissection of brain, however, allow differentiation of more than 220 cell types and areas of a rat brain (whole brain wet weight: 1.5 g) with samples measuring 200 500 ^m square x 100 ;am thick, a volume of 4 25 nL (12), and methods for isolation and analysis of single brain cells (20 50 pL) have been

PAGE 19

reported (13). As sample size is decreased to effect a more exact localization, enzymatic cycling methods must be used (14) which take advantage of the chemical amplification (10^-10^) afforded by the continuous recycling of enzymes in a reaction. This catalytic effect increases the concentration of the substance chosen as an indicator to the point where it can be accurately measured. These methods have generally been applied to determination of endogenous substances, biomolecules and enzymes, rather than exogenous substances. Microbeam and Probe-Type Instruments Another methodology that has been developed is the use of microbeams of electrons, ions or light, not only to visualize microstructures, but also to obtain a simultaneous chemical analysis of the surface by analysis of emissions of electromagnetic radiation, electrons and ions resulting from the interaction of the incident beam with the sample. The first such instrument, patented in 1947 and commercially available in the late 1950s, was the electron microprobe (15) which uses a focused (0.1 to 3 jam) beam of electrons to produce Xrays characteristic of the elements in the sample. Other probe-type instruments that followed include those based on Auger electron spectroscopy, ion backscattering spectrometry, electron spectroscopy for chemical analysis (ESCA) and secondary ion mass spectrometry (SIMS) (16, 17). The SIMS instruments gave rise to the ion microprobe and ion microscope, which differ in the method of obtaining an image of the sample. The microprobe uses a focused beam of ions to scan in raster fashion over the sample surface. The microscope uses a flood beam about

PAGE 20

7 250 to 300 jam in diameter to image an area of the surface. Final images of the secondary ions emitted maintain a direct relationship to the point of origin at the sample surface (18). Laser-based instruments. Shortly after the introduction of the laser in 1960 (19), it was incorporated into a microprobe instrument based on emission spectrography (20), and emission of electrons, ions and neutral atoms was observed from conductors, semiconductors and insulators irradiated in a mass spectrometer (21). Several reviews discuss the early work in both of these fields, concentrating primarily on emission spectroscopic instruments, laser characteristics, and interactions of laser radiation with solid surfaces (22-24). The most recent spectroscopy-based instrument to be introduced commercially was the laser Raman molecular microprobe (MOLE) (25). Point or global illumination was possible to allow recording of a complete Raman spectrum of a particular portion of a sample or mapping of a larger area by selecting a Raman line characteristic of an analyte of interest. Although this instrument was capable of identifying some molecular species, the complexity of the identifiable molecules and of their environment was limited. Most applications to date have shown identifications of inorganics, or simple organics deposited in biological systems in concretions of essentially pure crystals. This system is no longer commercially available, but work is still continuing in this area (26). Another microprobe instrument based on nitrogenpumped dye laser-excited atomic fluorescence of analytes after site-

PAGE 21

specific sample evaporation by a second microscopically-directed laser (ruby) has also been reported (27). Laser-based mass spectrometers . Lasers have been applied extensively in mass spectrometry (MS) over the past 20 years as a means of rapid vaporization and ionization of samples in the ion source. Conzemius and Capellan (28) have reviewed much of the work conducted prior to 1980, including instrument system characteristics (lasers and mass spectrometers) and applications. This review has been updated to include all work published through 1982 (29). In 1975, Hillenkamp et al. (30) reported the use of a LAser Microprobe Mass Analyzer (LAMMA) which was designed primarily for in situ determination of physiologically significant elements in intact biological samples. The instrument, introduced commercially in 1977 (LAMMA 500), is based on a time-of-flight mass spectrometer with a Q-switched, frequency-multiplied Nd-YAG sampling laser and a He-Ne spotting laser, both directed and focused through a high quality microscope. The system has been described (31), and two reviews have covered sample requirements, performance characteristics (32), and applications to structural analysis (33). Proceedings of a symposium on analysis with LAMMA have been published (34) and include articles on characteristics of laser spectra of various molecular classes (35, 36), preparation of biological specimens (37-39), and techniques for quantitative standardization (40, 41). A second generation instrument (LAMMA 1000) has been introduced (42-44) which allows for analysis of bulk samples rather than restricting the analyst to the use of thin sections, as did the LAMMA 500.

PAGE 22

Molecular analysis with current probes . The characteristic that links all of the above microprobe instruments is their almost exclusive application to elemental analysis. This is not to say that they have not been used to obtain spectra of molecules; however, there is an almost complete lack of applications for which the microprobe capability has been used in real, heterogeneous samples for qualitative or quantitative determination of molecules. In fact, in a search of the literature, only four such applications were found (two for the MOLE and two for the LAMMA). In three of these (25, 39, 45), the analysis succeeded because the spatial resolution of the laser beam allowed sampling of microcrystals or pure concretions of analyte, and not because of the resolution or chemical specificity of the spectrometric method. The only application that succeeded in identifying a molecular species in an intact, heterogeneous sample involved detection of a fungal metabolite, glyceollin, in plant tissue with LAMMA (46). In this study, two ion masses indicative of the metabolite were observed as the infection boundary was crossed. Part of the problem is that these instruments were never really designed to do molecular analysis. It is not, however, unrealistic to expect that the considerable selectivity necessary for molecular analysis is obtainable with appropriate combinations of currently available methodologies. In fact, several review articles on mass spectrometry and tandem mass spectrometry that have come out of the laboratory of R.G. Cooks in the past four years have included some reference to a "molecular microscope" (47-51),

PAGE 23

10 although no specific instrument has been described. The current project grew out of a desire to begin working toward this ideal. Projections for a molecular probe . Of the three instruments described above that are capable of localized molecular analysis, the laser mass spectrometer was thought to be the most promising. Due to the non-exclusive nature of the electromagnetic spectra of molecules (i.e., all organic molecules contain essentially the same types of bonds which give rise to signals contained within a limited portion of the electromagnetic spectrum for any given method), the selectivity and specificity of a microprobe instrument based upon optical (Raman) spectroscopy would be seriously limited. The molecular reordering that occurs during surface bombardment with energetic ion beams (50) would ultimately limit the resolution obtainable with an instrument based upon SIMS. The primary deficiency that precludes molecular analysis of heterogeneous samples with the LAMMA is the lack of confirmation of the identity of an ion corresponding to a specific mass peak. Elemental analysis is possible because molecular interferences can be eliminated by using a high power density (greater than 10^ W cm~2) and confirmation can be based on relative isotopic abundances for the analyte. This deficiency might be corrected by substituting a tandem mass spectrometer for the current single stage mass spectrometer to allow online verification of the identity of selected ions by sequential fragmentation and mass analysis steps. Tandem mass spectrometry . Modern tandem mass spectrometry (MS /MS) was developed in the laboratories of Cooks (52) and of McLafferty (53)

PAGE 24

11 using reversed-geometry, double-focusing instruments for Mass-analyzed Ion Kinetic Energy Spectrometry (MIKES). The introduction of the triple quadrupole mass spectrometer by Yost and Enke (54) simplified the automation of tandem mass analysis, increased the allowable speed of data acquisition, and brought the technique within reach of laboratories other than those specializing in design and construction of unique MS systems (similar instruments are now available from several manufacturers). A recently published volume (55) summarizes the developments and applications of this relatively new field. A tandem mass spectrometer consists of an ionization-f ragmentation region followed in space (tandem sector and quadrupole instruments) or time (Fourier transform ion cyclotron resonance instruments (56)) by a mass selector-analyzer region. Therefore, a molecular ion or fragment produced in the ion source can be selected by the first mass analyzer, fragmented, usually by collisionally activated dissociation (CAD), and a complete spectrum of the fragments can be obtained by scanning the second analyzer. Carrying this process to its limit, a single pure compound can be totally characterized by obtaining CAD spectra for the molecular ion and each fragment produced in the ion source; or a mixture may be characterized by using a soft ionization technique producing primarily molecular ions with little fragmentation, and obtaining a CAD spectrum for each component. Used in this way, the instrument is analogous to a chromatograph/mass spectrometer system; however, the relatively lengthy chromatographic separation is replaced by a nearly instantaneous mass separation. This decreases analysis and method

PAGE 25

12 development time, but more importantly, allows additional modes of operation that require simultaneous presence of signals from all analytes of interest. An example of full use and an explanation of these special MS/MS operating modes are given in a report describing a rapid, general method for detection and structural elucidation of drug metabolites (57). The recording of mass spectra is a statistical process in that the reproducibility of relative ion intensities is dependent upon the length of time the ion signals are present, and the length of time available for monitoring each mass of interest. Therefore, the primary problem associated with interfacing a laser with a mass spectrometer is a temporal one, since the ion signal resulting from interaction of a pulsed laser (pulsewidth = 10"^ to 10"^ s) normally has a lifetime of less than a millisecond. Zackett et al. (58) used a laser pulsed at 10 Hz to obtain MS and MS/MS spectra of sucrose deposited on a silver foil. They reported that a 10 ns pulse gave ions for 300 p.s in a chemical ionization (CI) source maintained at 0.1 torr with isobutane and for 1.5 ms at a source pressure of 0.5 torr. The same group identified a naturally occurring quaternary amine in cacti using laser desorption MS/MS after extracting the plant material and evaporating the extract on a silver or platinum foil (59). They used a CI source maintained at 0.5 torr with isobutane, and concluded that the CI conditions contributed only to increasing the ion signal lifetime by causing more efficient extraction of ions from the source, and not to the actual ion formation process. Cotter (60), however, has used an

PAGE 26

13 auxiliary chemical ionization source to take advantage of the abundant and relatively long-lived neutral species desorbed by the laser as discussed by Ready (61). A molecular ion for guanosine was recorded with a lifetime of 4 ms after a 40 ns laser pulse. Overview of Present Work The scope of the present work will include all aspects of interfacing a single shot-laser with a tandem quadrupole mass spectrometer, and prolonging the signal obtained upon interaction of the laser radiation with a solid sample. Chapter II will cover the work that has been done to develop an understanding of the interaction of laser radiation with matter and the processes leading to ion formation in laser desorption mass spectrometry. Chapter III will describe the instrumental systems and procedures used in this study. The capabilities of a commercially available laser microprobe (LAMMA 500) will be evaluated along with those of the triple quadrupole mass spectrometer for localization of species of interest (Chapter IV). The characteristics of the new system will be studied with respect to shortlived (Chapter V) and long-lived (Chapter VI) processes, so that recommendations can be made for future research efforts (Chapter VII) in the development of a microprobe capable of site-specific molecular analysis of heterogeneous and compartmentalized samples. Finally, an attempt will be made to answer a question which may make the development

PAGE 27

14 of a molecular microprobe-microscope for analysis of biological samples somewhat academic; that is, Although a molecule may be present in tissue, is it in a form that is recoverable by direct vaporization or is it incorporated into membranes, proteins and anatomical structures in such a way as to make it unrecognizable? This question may not have a single answer covering all species of interest; however, it is important to show that discrete molecular structures can be identified by vaporization from intact tissue.

PAGE 28

CHAPTER II CHARACTERISTICS OF LASER-INDUCED PROCESSES IN MASS SPECTROMETRY This chapter is primarily a literature review covering the most notable work directed toward developing an understanding of the laserinduced processes leading to ion formation. The initial discussion of absorption of laser radiation at solid surfaces covers the physical basis for prediction of the extent of temperature increase at the target surface upon laser impact. This treatment assumes that no vaporization or melting of the target material occurs, and is applicable to experiments conducted on bulk samples or substrates at power densities at which little visible surface evidence exists of laser impact. The second section deals with the investigation of the laser vaporizationionization processes. The primary questions that have been asked include (1) what is the energy spread of particles of mass spectral significance emitted (ions, atoms, neutral and ionized molecules), and does the system exist in a state of thermal equilibrium? (2) how does the nature of the laser-induced surface emission change with time? and (3) what are the mechanisms of laser-induced ionization? The chapter ends with a discussion of the techniques used to enhance the analytical usefulness of laser-initiated processes. 15

PAGE 29

16 Effects of Absorption of Laser Radiation at Solid Surfaces The unique properties of laser light include a high degree of brightness, directionality and raonochromaticity, these features resulting from the coherence of the emitted radiation. The laser is, therefore, a versatile source of spatially and temporally controllable spectral and thermal power. In laser desorption mass spectrometry (LDMS), it is the thermal power that is of primary interest because of the need for sample evaporation and ionization. The advantages of laser-induced vaporization over more common methods involving resistive heating lie in the high rate of heating which can approach 10^^ °C/s and the extreme localization of the effects to areas as small as 10~° cm^ (62). Experimental systems used for LDMS are of two basic types: (1) those in which samples are mounted on thin substrates such as electron microscope grids and sample and substrate are completely vaporized by the laser (MS instruments based primarily on time-of-f light ) ; and (2) those in which samples are mounted on relatively thick substrates (thin foils to bulk metal stages) and are partially or totally vaporized with no apparent effect on the substrate (studies employing time-of-f light , quadrupole and sector instruments). Because the latter process is viewed as an effect on the substrate which is transferred to the sample, much effort has been expended in determining the temperature rise induced in the sample substrate by the laser radiation, and the timecourse of laser-induced thermal effects.

PAGE 30

17 Ready (63) has developed equations for predicting temperature variations in metallic substrates for lasers having pulse widths greater than 10~^ s. An assumption of local equilibrium at the site of impact is based on the premise that the energy is absorbed by electrons whose collision frequency is 10^^ s~^ (the Debye frequency) so that distribution of the energy to other electrons and the lattice is instantaneous with respect to the duration of the laser pulse. A generalized (dimensionless) equation was presented by Ready to allow calculation of curves applicable to any material. This treatment was used by van der Peyl et al. (64) to calculate the time-dependent temperature variation for a stainless steel surface irradiated by a pulsed (200 ns) CO2 laser, whose pulse profile was experimentally determined. A comparison of stainless steel (m.p. 1808°C) and quartz (m.p. 1883°C) indicated that a lower energy was required to bring quartz to its melting point (2.4 raJ ) than was required for stainless steel (5.6 mJ). Lincoln and Covington (65) used a specially designed time-offlight mass spectrometer (66) which allowed variation of specimen distance from the ion source (30 or 100 cm) and monitoring of ions or neutrals. They calculated surface temperatures for two different types of graphite having different thermal conductivities by measuring the expansion velocity of the vapor plume and calculating the vaporization temperature from gas dynamic models for the vapor expansion process. Both of these studies indicated a more rapid increase in surface temperature for materials with lower thermal conductivities, the rationale being that materials of low thermal conductivity do not allow

PAGE 31

18 for rapid redistribution of the absorbed energy to the bulk substrate. Knox (67) has reviewed some of the earlier work in which surface temperatures were determined experimentally. Results for normal pulsed lasers with power densities of 10^ to 10^ W cm~2 were 2200°C to 9700°C for metals, and were relatively reproducible. Q-switched lasers with power densities of 10^ to 10^^ W cm~2, however, gave widely variant results for similar materials. Laser-Induced Vaporization and Particle Emission This discussion will focus on the vaporization processes that occur when laser radiation strikes a target. The initial section will cover vaporization from the point of view of the substrate, primarily with respect to spatial and temporal characteristics. The second section will cover vaporization from the point of view of the vapor plume, and will summarize most of the work that has been done to elucidate the mechanisms of laser-induced ion formation. Laser-Induced Vaporization Ready (68) again appears to be the jumping-off place for discussion of the vaporizing actions of normal-pulsed and Q-switched lasers. Data have been given for depths of holes produced in a number of thick metal targets. For a normal-pulsed laser (700 )is pulsewidth) , the depth was dependent on thermal conductivity of the substrate at low laser power density, but became independent of this property at higher power density and dependent on latent heat of vaporization of the metal. Therefore,

PAGE 32

19 at low power, the material having the greater thermal conductivity redistributes the energy to the bulk phase rapidly, experiencing relatively little vaporization, whereas at higher power, energy is absorbed too fast for redistribution, regardless of the thermal conductivity, and deeper holes are produced in the material having the lower heat of vaporization. The power density, F (W cm~2), at the point where heat of vaporization becomes dominant is given approximately by Eq. 2.1: F > 2LpkV^ (2.1) where L is the latent heat of vaporization (J g"-'-), pis the density (g cm~3), K is the thermal diffusivity (cm^ s~M , and t is the laser pulse length (s). This view is consistent with the types of laser interactions outlined by Kovalev et al. (69) and Furstenau (70), who described interactions for metals, semiconductors and dielectrics based on a model in which the extent of the interaction was determined by the concentration of free charge carriers (electrons) in the substrate and the difference between the valence band-conduction band energy gap and the photon energy of the laser radiation. For Q-switched lasers (F > 10^ W cm"^), vaporization is not a simple matter of ablation of material at its boiling point at a front that advances into the substrate. Other processes occur, including exertion of pressure on the target by the vapor plume, allowing superheating of the surface and leading to explosive ejection of

PAGE 33

20 material, and absorption of incoming laser radiation by the plume. In general, much less material is removed by short (ns), very high power density pulses than by longer pulses of lower power density. Scott and Strasheim (71) have studied the spatial and temporal characteristics of vapor plume generation by uormal-pulsed (850 jus), semi Q-switched (30 ;is) and Q-switched (50 ns) lasers. Metal substrates were used and impressive photographic evidence was presented in each case. In the normal mode, vapor emission began at about 30 jus after initiation of laser action and ended at about 400 jas, even though lasing continued. This was thought to be due to scattering of incoming light by particles emitted from the surface. The vapor expanded at 8.3 X 10 cm s"l with the solid angle of ejection decreasing as the crater became deeper. The semi Q-switched laser provided radiation at a higher power density and emission began after about 20 jus. No particle ejection was observed and the plume expanded at a rate of 3.2 X 105 cm s"^. For the Q-switched pulse, a small spherical plasma was visible for about 20 jus. The expansion rate was greater than 10 cm s~l and no particle ejection was seen. In all cases, the vapor plume expanded in a direction normal to the target surface regardless of the angle of incidence of the laser beam (33, 72). Laser-Induced Particle Emission The primary mathematical relationship found in articles dealing with LDMS is the Langmuir-Saha equation (Eq. 2.2): V^O " (S+/SQ)e^v[(^ D/kT] (2.2)

PAGE 34

21 where i+ and Iq are the positive ion and neutral molecule fluxes leaving the surface at temperature T. The quantities g+ and gg are the statistical weights of the ionized and neutral states, k is the Boltzmann constant, $ is the work function of the surface (eV), and I is the ionization potential (eV). The equation gives the temperature dependence of the degree of ionization of surface emissions (73). For substances for which I is greater than $, evaporation occurs at a temperature less than that required for efficient ionization (74). Datz and Taylor (75) reported that the alkali metals, the species most commonly implicated in laser-induced ionization processes, follow this equation when adsorbed on platinum and tungsten. From a review of the literature, Conzemius and Capellan (28) divided ionization efficiency into two general regions based on power densities of lasers employed. Below 10° W cm~2, ionization efficiency was low, about 10~^. Above 5 X 10^ W cm~2, emissions were almost completely ionized. Energyand Time-Resolved Studies Measurement of the energies and time-course of emission of desorbed ions and neutrals is useful for elucidation of the mechanisms of laser desorption-ionization, and also for providing a firm basis for increasing the analytical potential of the technique. Cotter (76, 77) has recently reviewed the area of energyand time-resolved studies, some of the most notable work having been done in his own laboratory. Numerous reports of energy spreads of elemental and molecular ions have offered conflicting evidence. The review by Ready (78) cited ion energies up to 5 keV and typical values of up to 400 to 500 eV. Bernal et al. (79),

PAGE 35

22 using a Q-switched CO2 laser (775 ns pulsewidth; 7 x 10^ W cm"^), reported time-resolved energy spreads during the 1.3 us ion pulse as to 60 eV initially and rising to about 200 eV. Ion energies of up to 20 eV have been reported for the LAMMA 500 for thin non-metallic samples, and up to several hundred eV for bulk metallic samples (31). Hardin and Vestal (80), using a quadrupole instrument with samples being introduced continuously on a moving belt, reported energies of 6 to 25 eV for potassium iodide clusters and protonated and cationized guanosine ions and fragments. They reported that the observed time-of-f light distributions were dependent on laser power density, and the relative concentrations of sample and alkali salt on the belt. All work was performed at a power density of about 10^ W cm"^. Energies as low as 0.26 eV were reported by van der Peyl et al. (81) for Na"^ in a doublefocusing instrument. Cotter and coworkers have published a series of papers in which they report results obtained on a versatile time-of-f light instrument designed to allow observation of ions produced directly by the laser, or laser-produced neutrals ionized by a pulsed electron beam (82-85). An ion-drawout pulse was provided to sample the contents of the ion source at any time up to 500 us after the laser was fired. They have used this system to resolve the convoluted profiles of laser-desorbed ions into separate contributions from prolonged vaporization envelopes and finite energy spreads. The energy spread for K"^ was 15.3 eV at 10 jus after the laser pulse, and decreased to 0.31 eV at about 44 jjs. This is probably

PAGE 36

23 the most accurate work that has been done to date, and while it does not negate the value of the previous work, it certainly offers a much improved method for solving the problems. Mechanistic Studies Early work in laser-assisted mass spectrometry was concerned mostly with elemental analysis or pattern recognition after pyrolysis of large molecules or molecular aggregates. Vastola et al. (86, 87) introduced the concept of using laser radiation for volatilization and ionization of molecules that could not be vaporized intact by the usual methods of ohmic heating. Until 1978, laser desorption continued to take a back seat to field desorption (FD), secondary ion mass spectrometry (SIMS) and the newly introduced (88) plasma desorption (PD) in the field of molecular analysis. The commercial introduction of the LAMMA 500 and the report by Posthumus et al. (89) in particular provided impetus to define the optimum conditions for laser desorption and the mechanisms of ion formation. In this work, laser desorption spectra were presented for important biomolecules which were generally considered intractable, including oligosaccharides, cardiac glycosides, steroid conjugates, antibiotics, nucleotides, nucleosides, amino acids, peptides, and chlorophyll. All gave abundant molecular ions as alkali metal ion adducts (M+Na"^, M+K"^), indicative of a process called "cationization" . This study, along with others (88, 90, 91), indicated that a rapid heating rate was important for molecular ion production in preference to fragmentation and decomposition. Although most researchers have used laser power densities of 10^ to 10^ W cm"2, Stoll and Rollgen (92)

PAGE 37

24 observed the cationization process and desorption of quaternary ammonium ions (preformed ions) using a CW CO2 laser at a power density of about 20 W cm~2; however, after scrupulously cleaning the copper substrate and purifying a sucrose sample to eliminate traces of alkali metals, no molecular ions or protonated or cationized molecules were detected, even though the sucrose did evaporate. In an effort to determine the mechanism of the cationization reaction which is predominant in laser desorption, van der Peyl et al. (93-95) performed experiments using several different instrumental configurations in which thermionically produced alkali metal ions originated in a different area of the ion source than the test compound (sucrose). Temperatures of both thermionic emitter and sample holder were monitored during resistive heating and laser heating cycles and plotted along with ion intensity curves. They proposed a model in which the laser serves to produce a steep temperature gradient at the surface with alkali metal ions thermionically emitted from the hot central zone of the laser spot (greater than 427°C) and neutral sample molecules vaporized from the relatively cooler periphery. Ion formation occurs in the gas phase as an ion-molecule reaction producing a stable, evenelectron species with a positive charge on the alkali metal atom. Stoll and Rollgen (96, 97) arrived at a similar conclusion using a system in which the thermionic emitter was not completely isolated from the sample. Heinen (98) and Hercules et al. (33) have reviewed the types of ions formed from molecules containing various functional groups. Hillenkamp (99, 100) has summarized the work reported by the most

PAGE 38

25 prominent LDMS research groups, and listed some features of LD mass spectra. These features include (1) in contrast to the normal situation with resistive heating, polarity and ionic character promote desorption and ionization; (2) ions are generally even-electron species; (3) cationization by alkali and other singly-charged metal ions is common; (A) anionization has been observed but not frequently reported; (5) positive and negative ions are produced in approximately equal amounts, with less fragmentation of negative ions; (6) cluster ions are frequently seen; (7) matrix effects are observed; and (8) extent of fragmentation is dependent on laser power density. Cooks and Busch (101) have compared the various desorption ionization techniques (SIMS, FD, PD, LD) and have developed a general model for the ionizationfragmentation events that occur. The model is based on the fact that these techniques, which differ in the types of energy put into the sample result in the production of similar ion types. Analytical Usefulness of Laser Desorption Mass Spectrometry The utility of a laser desorption/evaporation source in mass spectrometry derives from the characteristics of the laser beam, i.e., it can be focused to sub-micrometer diameter spot sizes, contains a high thermal power, can desorb or vaporize large, polar molecules intact, and can ionize them directly or present them to the instrumental system in a state suitable for ionization by other means. The primary tasks associated with coupling a laser with a mass spectrometer are (1)

PAGE 39

26 prolonging the laser-induced events so that they persist for the time required for the mass spectrometer to acquire statistically relevant data; or (2) designing the mass spectrometer so that it can acquire such data during the lifetime of laser-produced ions. Ion lifetimes have been extended by operating the ion source in the chemical ionization mode (102). This not only takes advantage of the neutral species that are desorbed, but also increases the lifetime of ions produced directly by the laser. Using a renewable surface with a pulsed or CW laser provides a continuous signal, but this is not ideal for microprobe work, because the intrinsic nature of the investigation precludes the assumption of a renewable surface (i.e., why use a microprobe if the sample is homogeneous enough for the surface to be considered renewable). A variation of this approach involves pulsing the laser in synchronization with mass scanning so that only one mass is monitored for each laser pulse (repetitive laser desorption (103)). McCrery et al. (104) used ion storage in a Fourier Transform mass spectrometer and were able to digitize and average twelve decay transients for each laser pulse. Adapting the mass spectrometer to the laser duty cycle has been done in the LAMMA instruments with time-of-f light mass analysis. Other workers have used sector instruments with photoplate or array detectors to simultaneously detect ions over a certain mass range. Since the objective of this work is to evaluate a tandem mass spectrometer system for application to laser microprobe mass spectrometry, emphasis was be placed on obtaining as much information as possible from a single laser shot .

PAGE 40

CHAPTER III EXPERIMENTAL This chapter describes the instrumentation and procedures used in this three-part study. The first experiments, performed in the laboratories of D. M, Hercules at the University of Pittsburgh, were done to gain familiarity with a commercially available laser microprobe and to illustrate the problems inherent in using a single stage mass analyzer for analysis of heterogeneous samples. The second study was done to evaluate the capabilities of the Triple-Stage Quadrupole Mass Spectrometer (TSQ) for quantitative analysis of drugs in small (2-10 mg) samples of rat brain in preparation for anticipated laser/MS/MS analysis at some later date. The objectives of the final series of experiments were (1) to design and construct a laser interface for the triple quadrupole mass spectrometer; (2) to evaluate the instrument system with respect to short-lived and long-lived events produced directly or indirectly by the laser; and (3) to determine the feasibility of and make recommendations for construction of a laser/tandem mass spectrometer-based raillior microprobe capable of molecular analysis of heterogeneous samples. 27

PAGE 41

28 Laser Mlcroprobe Study Apparatus Laser mlcroprobe mass analyzer (LAMMA 500) . The LAMMA 500 (Leybold-Heraeus GmbH, Federal Republic of Germany) has been described in detail (31). The instrument consists of a time-of-f light mass spectrometer with an open ion source, coupled with a Q-switched, frequency-multiplied Nd-YAG laser through a high quality microscope. A He-Ne laser aligned coaxially with the Nd-YAG laser is used for site selection and aiming. The laser light is incident on the sample at 90° to the sample rear surface and ions are extracted into the mass spectrometer at 90° to the sample front surface (the instrument operates with a 90°/-90° configuration, according t9 the convention established by Conzemius and Capellan (28)). For this reason, samples are restricted to thin films, or if granular forms are used, the laser must be aimed so as to strike a grain edge. An ion reflector is employed to minimize the effect of the initial ion energy spread on mass resolution and to increase resolution by allowing a longer flight path without greatly increasing instrument size. Procedure Preparation of animals . An epileptic focus was created in an adult male Sprague-Dawley rat by sub-dural deposition of powdered cobalt (105). After 2 weeks, the lesioned animal along with a control (nonlesioned) animal were sacrificed. The brains were removed (106) and fixed in phosphate-buffered glutaraldehyde for 48 hours. The lesioned

PAGE 42

29 area of the test animal and a corresponding area from the brain of the control animal were dissected out and embedded for sectioning. Thin sections (1 jam) from each sample were mounted on electron microscope grids for use in the LAMMA, and serial sections were mounted on glass microscope slides for pre-test selection of analysis sites by optical microscopy. Analysis . A calibration mixture of Ni, Mo, and Pb was made by grinding together NiSO^, M0S2, and PbS04, and dusting the fine particles onto an EM grid. Spectra were acquired with a 0.020 )is sampling interval and a 1100 channel delay time. Delay time is the time in sampling intervals that the transient recorder waits after the laser is fired before beginning to accumulate data and store it in the 2048 channels available. Sampling interval is the time between each sample, or channel width in ;is, and determines the upper limit of the mass range. The delay time multiplied by the sampling interval determines the low mass recorded. The delay time plus 2048 multiplied by the sampling interval is the total flight time that is covered in a run. Laser energy before filtering (energy incident upon the sample is selected very roughly by neutral density filters) is read in mJ after each shot from a digital panel meter which receives a signal from a photodiode. Energy from shot to shot is variable and uncontrolled. Energies varied from 9.7 mJ to 32.5 mJ , but were usually between 10 and 20 mJ. Mass calibration was obtained from the Ni-Mo-Pb calibrant spectrum by identification of mass peaks from the isotopic patterns and correlation of flight times with chart distances. Numerous spectra were

PAGE 43

30 recorded from the lesion sample and the control sample with delay times of 1000 channels and a sampling interval of 0.020 ^s. Specimens were observed in the LAMMA with the 32X objective (total magnification 320X) , The mass range covered was m/z 20-219. Preliminary Triple Quadrupole Mass Spectrometry Study: Regional Distribution of Phenobarbital and Phenytoin in Rat Brain Apparatus Finnigan triple stage quadrupole (TSQ) mass spectrometer . The TSQ (Finnigan MAT, San Jose, Calif.) (Fig. 3.1), consisting of an electron impact (EI )/chemical ionization (CI) ion source followed in series by a mass selector-analyzer quadrupole (Ql), an ion-focusing fragmentation quadrupole (Q2), a second mass selector-analyzer quadrupole (Q3), and a continuous dynode electron multiplier, has been described (107, 108). Finnigan INCOS software (TSQ revision C) was used for instrument control, and data acquisition, storage, and manipulation. The center quadrupole (Q2) functions as a second fragmentation region when pressurized with an inert gas (N2, Ar), operating by collisionally activated dissociation (CAD) as described by Yost et al. (119). Table I shows the primary operating modes of the instrument along with the corresponding configurations of the mass analyzer and collision quadrupoles.

PAGE 45

32 o

PAGE 46

33 a o -H -1 (I) 0) CO 3 — 1 ^ 6 CD *Q. O O (D O X-H O 0) -*-D ^ ~ -^ C (/) (0 (0 -Q O (D 3 E (0 ID T3 -t^ (D

PAGE 47

34 Procedure Preparation of animals . Rats were given phenobarbital or phenytoin (20 mg kg"-'twice daily) for 5 days. Animals were sacrificed, the brains were removed, and various structures (cerebellar gray and white matter, sensory motor cortex, basal ganglia, hippocampus, corpus callosum, occipital cortex, pontine white matter) were separated by vacuuming into weighed glass pipets plugged with glass wool. Analysis . The pipets containing the brain samples were reweighed to obtain sample weight, and the section of pipet containing the sample was crushed in a screw cap culture tube (Teflon-lined screw cap) for analysis. Whole blood was also collected and plasma was analyzed. Methylated analogues of the drugs [5-(p-methylphenyl)-5-ethyl barbituric acid, 5-(p-methylphenyl)-5-phenylhydantoin] were used as internal standards. Samples were homogenized by sonication in 300 juL of 1 M phosphoric acid. After addition of 300 juL of 1 M phosphoric acid containing the internal standard, samples were extracted with 3 mL of chloroform. The chloroform residue was dissolved in 100 juL chloroform and 1 to 2 jaL was placed in a probe vial for MS. The mass spectrometer was operated in the selected reaction monitoring (SRM) mode by sequentially passing the drug and internal standard molecular ions resulting from methane CI through Ql, fragmenting those by CAD in Q2 (pressurized with nitrogen at 2.3 mtorr; collision energy of 18 eV) and selecting the major daughter ions for each in Q3 . Thus, phenobarbital (mol. wt. 232) was monitored by selecting the m/z 162 daughter ion in Q3 while passing m/z 233 through Ql. Quantitation was based on the ion

PAGE 48

35 intensity ratio of m/z 162 to m/z 191, the daughter ion of the phenobarbital internal standard parent ion m/z 247. Phenytoin (mol. wt . 252) concentration was calculated from the ratio of the m/z 182 daughter ion of the phenytoin molecular ion, m/z 253, to the m/z 196 daughter ion of the internal standard (parent ion, m/z 267). Laser-Triple Stage Quadrupole Studies Apparatus Laser . The laser was a coaxial f lashlamp-pumped dye laser (Model SLL-625, Candela Corporation, Natick, Mass.) operated primarily with Rhodamine 6G (A x 10~^ M in absolute ethanol) with driver voltages between 16 kV and 20 kV. The output coupler was a sapphire window without reflective coatings. Pulsewidth was 1.5 to 2 ;is. The laser could be operated only in the single shot mode with a maximum firing rate of 3 shots min~^. An iris diaphragm, calibrated in 2 mm increments between 2 mm and 18 mm, was inserted into the laser cavity between the dye cell and the output coupler to reduce the diameter of the laser beam and focal spot. Pulse energy, measured with a volume absorbing energy meter (Gen-tec, Ste-Foy, Quebec, Canada), was 0.65 J, 1.49 J, and 1.64 J for driver voltages of 16 kV, 18 kV and 20 kV, respectively. Laser-mass spectrometer interface . The side flange of the TSQ was machined according to the specifications of Fig. 3. 2. a to accommodate a 1.5" diameter fused silica lens (focal length: 12.5 cm). Three spacers (Fig. 3.2.b) were also machined to allow adjustment of the point of

PAGE 49

Figure 3.2 Diagrams for machining (a) blank side flange of TSQ to hold 1.5" diameter fused silica lens, and (b)spacers for adjusting focal point of lens within ion source. Measurements marked are: (A) 4.5 in.; (B) 1/4 in.; (C) 7/8 in.; (D) 5/8 in.; (E) 1 in.; (F) 1/4 in. (G) 1/8 in., 3/16 in., and 1/4 in. (3 spacers).

PAGE 50

a) FLANGE 37 A ^ b) SPACER 0-RING GROOVE

PAGE 51

38 focus over a 9/16" range in 1/16" steps. As it happened, the focal point was about 1/16" from the reagent gas inlet side of the ion volume with no spacers. Vacuum was maintained with 0-ring seals. To allow entrance of laser light into the ion source, a 1/8" hole was drilled into the ion volume (EI and CI) directly opposite the reagent gas entry hole. Extensions to the solids probe were machined out of stainless steel according to the specifications of Fig. 3.3 to allow insertion of samples on electron microscope (EM) grids and angles of laser incidence of 90° and 45°. The 45° tip was later modified to allow grids to float in the ion volume with no backing. This tip was used for the desorption CI experiments. The dye laser was aligned (Fig. 3.4) by directing a HeNe laser (LI) through the dye cell (L2) with the retroref lector (MO) removed, through the output coupler (Wl) to a plane mirror (Ml) which reflected the beam to another plane mirror (M2) mounted on the TSQ, then to plane mirror (M3) mounted on a gimbal mount on a line axial with the focusing lens (W2). Final adjustments in laser aiming could be made by slight adjustments of M3 while determining the point of impact on an electron microscope grid inserted into the ion source. Detection systems . Short-lived events were monitored with one of two MS detection systems. In the first, a storage oscilloscope (Model 549, Tektronix, Inc., Beaverton, Ore.) was connected directly to the electron multiplier of the TSQ through a Type D high gain differential amplifier plug-in, and was triggered by a TTL pulse from a photodiodeto-TTL interface connected to the output of a fast photodiode (FNDIOO, EG&G Electro-Optics, Salem, Mass.). The laser light (about 8% of total

PAGE 52

Figure 3.3 Diagrams for machining (a) 90° and (b) 45° stainless steel solids probe extensions. The 45° extension was later modified for desorption chemical ionizaton (DCI) by grinding the tip until it was flat (dotted line). Measurements shown are: (A) 9/32 in.; (B) 1/4 in.; (C) 0.0765 ± 0.0005 in.; (D) 1/16 in.; (E) 0.215 ± 0.001 in.; (F) 1/64 in.; (G) 0.133 ± 0.001 in.; (H) 13/32 in.. Depression is 7/64 in. in diameter and between 1/64 in. to 1/32 in. deep. A small leaf spring tightened by a screw is used to hold the tabbed EM grid on the 90° tip, and a cylinder made to just fit over the end of the 45° tip is used to secure the grid on that probe extension.

PAGE 53

40 a) 90° SOLIDS PROBE EXTENSION SIDE VIEW TOP VIEW END VIEW b) 45° SOLIDS PROBE EXTENSION WITH DCI MODIFICATION \f\ lii END OF EXTENSION AFTER MODIFICATION FOR DCI DEPRESSION -> •*D 1 l«—

PAGE 54

O ^-' O t-l O U C •^1 4-1 0) 5 "^ O l-i O CO 4-1 o • w

PAGE 55

42 2 < ' ^ 1 » :; (==2 ^ / ^

PAGE 56

43 output) was reflected to the photodiode from a microscope slide placed in the light path. Comparison of the TSQ instrument oscilloscope with the storage scope indicated a sensitivity of 1.6 x 10~' A V"-'for the latter. The storage scope could not be run from the amplified instrument signal, because the output of the amplifier contained the instrument scope trigger signal and mass marker signals which disrupted the storage scope signal and triggering. In the second detection system, the TSQ data system was used for monitoring fast signals by acquiring data in profile mode. In this mode, time resolution of 25 }is was obtained because the instrument acquires data at the maximum analogto-digital conversion (ADC) rate of 40 kHz. A single long scan (20 s) of profile data was acquired during which time the laser was fired. By setting the mass of interest with the unscanned quadrupole and turning off the DC voltage to the scanned quadrupole, a signal of ion current vs. time was obtained. This operation will be described in more detail in Chapter V. The beginning of the scan was monitored by connecting an oscilloscope (Model 2213, Tektronix, Inc.) to the positive Y-axis input (Y+) at the rear of the TSQ scope and observing the scan signal (TSQ scope set for multiple mass marker display) with a fast time base. Long lived events were monitored with the TSQ data system operated normally, but with the centroid sampling interval set at less than the normal 200 JUS. The centroid sampling interval is the number of ADC readings ( 1 to 8) summed to produce a single datum point. Therefore, with an ADC rate

PAGE 57

44 of 40 kHz, the centroid sampling interval can be varied, in 25 jjs increments, from 25 )is to 200 ;is with the latter being the default setting. High performance liquid chromatograph . A high performance liquid chromatograph was used for quantitation of test drugs in animal samples and to determine the quantities of sample vaporized in the ion source by the laser. A high pressure syringe pump (Model 4100, Varian Associates, Palo Alto, Calif.), 6-port rotary injection valve (Valco Instrument Co., Houston, Tex.), and a fixed-wavelength (214 nm) ultraviolet absorbance detector (Model 1203, Laboratory Data Control, Riviera Beach, Fla.) were used with a 4.6 mm i.d. x 250 mm stainless steel column, packed with a 5-jam particle diameter octadecyl silica gel stationary phase (BiosphereODS, Bioanalytical Systems, Inc., West Lafayette, Ind.) and maintained at 52°C with a circulating water bath (Neslab Instruments, Inc., Portsmouth, N.H.). The mobile phase was pumped at 80 mL/h, and consisted of acetonitrile: tetrahydrof uran:water (28.5:0.5:71) with potassium dihydrogen phosphate added at a concentration of 2.7 g L~l . Procedure Laser system evaluation . The laser system was evaluated with Coumarin 2 and Rhodamine 60 laser dyes (Eastman Kodak, Rochester, N.Y.). The coumarin lased at about 450 nm, however, the intensity was low and the useful life of the dye was only a few shots. Rhodamine 60 lased at about 590 nm with high pulse energy and a single dye solution allowed

PAGE 58

45 operation for several hundred shots. Pulse shapes were measured for various laser cavity diameters at 16 kV, 18 kV and 20 kV by monitoring the signal directly from the fast photodiode. Laser-TSQ evaluation . The laser-mass spectrometer system was evaluated with respect to short-lived and long-lived laser-initiated processes. The short-lived events (}is to ms) were observed with an open electron impact ion source at high vacuum, while long-lived processes (ms to s) occurred with a closed chemical ionization source with reagent gas present at greater than 0.35 torr. Phenytoin, a relatively nonvolatile (m.p. 295 298°C) drug, which was also used in the animal experiments, was used in most of the system evaluations, since it was readily available as the free acid and the sodium salt, and proved to be a suitable test substance. A series of compounds (Appendix) having various functional groups and molecular weights was also tested under desorption chemical ionization conditions (110) in which only the source block was heated in most cases, under normal direct probe conditions, and under laser vaporization conditions to determine the versatility and utility of the present system. In all cases, samples were inserted into the source as solids packed into a depression in the stainless steel probe tip or as residues from solutions evaporated on copper, nickel or silver electron microscope grids (Ted Pella, Inc., Tustin, Calif.). Preparation, collection and analysis of animal samples . To provide animal tissue samples suitable for laser analysis, two adult male rats were given a loading dose (150 mg/kg) of phenytoin intraperitoneally

PAGE 59

46 75 min before decapitation. At the time of sacrifice, the animals were ataxic and lethargic, indicating substantial drug absorption. Brains and livers were removed and washed in isotonic phosphate buffer. One set was frozen at -20°C. The other set was fixed for 72 h in phosphatebuffered glutaraldehyde (KCl, 0.0834 M; KH2PO4, 0.0127 M; K2HPO4, 0.0846 M; 25% aqueous glutaraldehyde; 925 mL water). Samples from fixed and frozen brains and livers were analyzed by high performance liquid chromatography (111). Aliquots of homogenized weighed tissue samples were deproteinized with acetonitrile. The aqueous acetonitrile phase was washed with iso-octane and then poured into a tube containing granular KCl. Vortexing caused the acetonitrile to be salted out of the water, at the same time extracting the phenytoin and internal standard. The acetonitrile layer was transfered to a clean tube, evaporated to dryness, and the residue was dissolved in mobile phase for injection into the HPLC system. Sections of the fixed tissues were frozen in isotonic phosphate for later MS analysis.

PAGE 60

CHAPTER IV PRELIMINARY STUDIES AND INSTRUMENT CHARACTERISTICS This chapter includes discussions of (1) the LAMMA experiments with thin slices of rat brain tissue taken from the cobalt lesion site in one animal and a corresponding site from a non-lesioned animal, (2) the determination of the regional distributions of phenytoin and phenobarbital in rat brains by triple quadrupole MS /MS, and (3) the basic characteristics of the laser-TSQ system, including laser pulse shape and variation of diameter of laser spot size with target position and cavity diameter. LAMMA Experiments The LAMMA instruments, initially the Model 500 and, more recently, the Model 1000, were designed ideally to monitor atomic mass spectra during the brief lifetime of laser-induced emissions from thin film and bulk samples. The time-of-f light (TOF) mass spectrometer operates within the time restrictions of Q-switched, single shot lasers and the lifetimes of the ions which they produce, so that a complete mass spectrum covering a significant mass range can be acquired in less than 100 us. After the laser fires, kinetic energy is imparted to the ions in the source area by a high voltage accelerating pulse. This is 47

PAGE 61

48 translated into an ion motion according to eqn. 4.1, (l/2)mv2 = zeV (4.1) where m and v are the ion mass and velocity, respectively, z is the ion charge, e is the electron charge, and V is the accelerating voltage. The time of ion flight is determined by the length of the flight tube (L) and the ion velocity (eqn. 4.2). t = Lml/2/(2zeV)l/2 (4.2) Since m is equal to the molecular weight divided by Avogadro's number (6.023 X 10^^) and the voltage in electron volts (eV) is the accelerating voltage times the charge per molecule (ze), for singly charged molecules with time measured in seconds, eqn. 4.2 becomes, t = 7.199 X 10-7(L/vl/2)ml/2 (4.3) t = cml/2 (4.4) where c is the instrumental constant, incorporating unit conversion factors, length of the flight path and instrument accelerating voltage. Table II gives the masses, their square roots and abundances for the

PAGE 62

49 Table II. Masses and Flight Times for Nickel, Molybdenum and Lead Isotopes in LAMMA Calibrant sotope

PAGE 63

50 calibrant isotopes (Ni, Mo, Pb) , and the distance and time data derived from the calibration spectrum. The flight times (t) were calculated as the sum of chart time (t^-) plus the delay time (t^j) with tc = nts(dp/D) (4.5) and td = "d^s (^-6) where n and n^j are the total number of chart channels (2048) and the number of delay channels, respectively, tg is the sampling interval, and dp and D are the chart distance to the peak and total chart distance covered by 2048 channels (393.0 mm at 0.020 ;is sampling interval), respectively. Regression analysis of flight time (^is) vs. the square root of the ion mass indicated a 2.563 ps lag (Y-intercept ) between the start of recorder timing and the ion acceleration pulse. The slope of this plot (Fig. 4.1) gave the instrument constant, 3.944 jas u"-'-'^. Using this value with eqns. 4.3 and 4.4 gave a flight path length of 300.2 cm. Figure 4.2 shows the calibration spectrum (A) with spectra acquired from the lesion site (B) and the control site (C), Although the peak at m/z 59 (*) is higher in the lesion site than in the control sample, it is impossible to say if this signal is actually due to cobalt. The presence of large peaks at almost all masses throughout this range

PAGE 64

^-t 0) 4-) 0^ rH T3 !-> r^ O t\3 O vo CU O M-l Oj

PAGE 65

52

PAGE 66

c

PAGE 67

54

PAGE 68

55 precludes any accurate assessment of elemental or molecular composition. These results point out the advantage of having two stages of mass analysis (filtering and selection) for microprobe analysis. By increasing the laser power density, the chemical noise from molecular fragments could be decreased, but it would still be impossible to definitely identify the ion at m/z 59 as cobalt since there is only one naturally occurring isotope. Uncertainty of ion identification becomes greater as mass increases because the number of possible elemental compositions also increases. This deficiency is made obvious in a recent report on LAMMA analysis of coal and shale samples (112). The laser was fired at grazing incidence at 1/8" diameter chips, and ions produced directly by the laser were recorded. Although the authors gave explanations of the spectra and made mass assignments, primarily on the basis of natural isotopic abundances and past literature reports of laser pyrolysis gas chromatography and mass spectrometric studies of homopolymers, a second stage of mass analysis would have added greatly to the certainty of some of the assignments. Regional Distributions of Phenobarbital and Phenytoin in Rat Brain At the time this project was started, very little quantitative work on real biological samples had been done by MS /MS. A quantitative study of typical samples amenable to analysis by other methods (i.e., gasliquid or high performance liquid chromatography) was considered to be a

PAGE 69

56 necessary starting point, to define instrumental conditions, particularly with respect to the CAD behavior of analytes, which would give accurate, reproducible results. Since one of the primary advantages of MS/MS is elimination of the need for pre-MS (chromatographic) separation of mixture components, demonstration of reliable analysis by direct vaporization of samples (extracts) from the solids probe was also considered important. Rapid vaporization of the entire sample from the solids probe vial with as little temporal separation of analyte and internal standard evaporation profiles as possible was found to be of primary importance. Separation of the analyte and internal standard ion signals (because of differences in their vaporization temperatures) caused the response ratio to be dependent on the number of scans over which the signal was summed. To minimize this effect, a heating rate of about 100°C s~^ was employed. This generally resulted in complete sample vaporization within a 30 s period with a near-Gaussian ion current profile. Analyteinternal standard pairs vaporized simultaneously within about 20 s (Fig. 4.3). Response ratios for drug and internal standard ion currents were also dependent on instrument tuning, so that long-term (day to day) reproducibility of calibration curve slopes was poor; however, linearity was good, with correlation coefficients generally between 0.996 and 0.999. Figure A. 4 shows a histogram of the results for phenobarbital and phenytoin in plasma and various brain regions. The numbers in parentheses beside each bar give the number of samples (before slash)

PAGE 70

01

PAGE 71

58

PAGE 72

iH • 0) 0) W e « ;3 •V a. o u c 6 w ~aft M ,Q O O 4-1 60 3 S-i 4-1 O P C
PAGE 73

60

PAGE 74

61 taken from the number of animals (after slash) sacrificed. Regional distribution of phenobarbital is relatively even. Although phenytoin appears to be slightly more concentrated in white matter and less concentrated in basal ganglia and hippocampus, it also is fairly evenly distributed. Concentrations in structures that are possibly linked with epileptic activity or drug effects (hippocampus, cerebellum, sensorimotor cortex) are not greatly different from those in areas that are not linked with seizures (pontine white matter, occipital cortex). The averages of the histogram values for all structures are 38 ± 2 jjg g~^ and 11 ± 3 Mg g"-*for phenobarbital and phenytoin, respectively; brain to plasma ratios for these averages are 0.568 and 1.73, respectively. The low relative standard deviations for these samples, 5.6% for 50 phenobarbital determinations, and 20% for 30 phenytoin samples, indicate that monitoring analyte daughter ion intensities for selected reactions with respect to those of suitable internal standards is a valid technique for direct analysis of plasma and brain extracts by triple quadrupole mass spectrometry. The response ratios are made more reproducible by rapid, complete vaporization. This was assured by rapid heating of the solids probe in this experiment. In the instrument under consideration, rapid heating and sample vaporization would be affected by using a pulsed laser as a sampling device.

PAGE 75

62 Laser-Triple Stage Quadrupole Characteristics Variation of Focal Spot Size with Distance from Focal Point and Laser Cavity Diameter . The laser beam, which is about 20 mm in diameter when emanating from a totally open cavity, was reflected from a flat mirror 24 inches in front of the output coupler to the 1.5 inch diameter fused silica plano-convex lens used as the laser window in the TSQ. The target was a black piece of cardboard that was placed from 12 mm behind the focal point to 18 mm in front of the focal point. Spot size varied from 3.2 mm at the former to 5.0 mm at the latter and was 1.3 mm at the focal point (Fig. 4.5). Spot size initially increased more slowly with distance in front of the focal point. The cavity diameter was varied by placing an iris diaphragm between the dye cell and the output coupler. The target was located 4 mm in front of the focal point of the lens. Spot size was 1.7 mm, 1.0 mm and 0.5 mm with iris openings of 18 mm, 12 mm and 6 mm, respectively. Below 6 mm the spot size was too small to be measured accurately, but a flash could still be seen at the target with the cavity diameter set to 2 mm. Normal cavity diameters used in this study were 6 mm to 10 mm, depending on the age of the flashlamp and laser dye solution. Focal spot sizes for these settings were 0.5 mm to 1 mm.

PAGE 76

a) ;s c u-i CM 0) •a 'H CO C CO l-l

PAGE 77

64 + W H en

PAGE 78

65 Laser Pulse Shape and Duration . Laser pulse shape and duration were measured using the output from the fast photodiode to drive the storage oscilloscope directly. Figure 4,6 shows the tracings produced by varying the cavity diameter (2 to 12 mm) at 16 kV, 18 kV and 20 kV. In all cases, the light pulse was reflected from two microscope slides to the photodiode so that all pulses remained on the oscilloscope scale (5 V cm~^). Therefore, only about 0.6% of the laser light was incident on the photodiode (8% reflected from each slide). The time base of the scope was set at 2 ;is cm~^, indicating a pulse width (full width at half maximum) varying between 1 yis and 3 }is and depending on cavity diameter and driver voltage. At 18 kV and cavity diameters of 6 mm to 10 mm (normal settings), pulse width was 1.5 }is to 2 ;as.

PAGE 79

Figure 4.6 Laser pulse shapes from photodiode signal as monitored by storage oscilloscope at driver voltage of (A) 16 kV, (B) 18 kV, and (C) 20 kV. Cavity diameter was varied from 2 mm to 12 mm in 2 mm increments giving curves running from low intensity to high intensity in each figure (drawn from photographs of storage oscilloscope screen).

PAGE 80

67 20 _ 10 _ — I r 4 6 TIME (ys) 20 _ 10 _ — I r 4 6 TIME (|js) 20 10 _ I I 4 6 TIME (ps)

PAGE 81

CHAPTER V SHORT-LIVED PROCESSES Although the quadrupole mass filter scans the fastest of all the scanning mass analyzers, normally covering 1000 to 2000 mass units each second, it is still relatively slow when compared to the ion lifetime of pulsed laser-generated ions. In spite of this incompatibility, an attempt to observe the events immediately following the laser pulse was thought to be desirable for this study for two reasons. First, the laser used in this study was a single shot device (1.5 to 2 jus pulsewidth) and prolonged signals were not expected, and second, observation and verification of normal laser-induced processes such as cationization required mass discrimination and time resolution in the us range. This chapter will present results of experiments conducted using an electron impact ion volume, modified to transmit the laser light, and operated without reagent gas. Under these conditions, ion signals lasted less than 1 ms and were monitored either with a storage oscilloscope or with the TSQ data system. Limited mass discrimination was maintained by selecting the mass of interest with Ql and, with no collision gas in Q2, selecting a range of masses which may or may not include the mass of interest with Q3 operated in RF-only mode. 68

PAGE 82

69 Triple Stage Quadrupole Control The goal in these experiments was to have the TSQ operate as a time-of-f light mass spectrometer for the selected ion, allowing data to be obtained at a single m/z during the course of an experiment. This had been accomplished in a single stage quadrupole instrument by Hardin and Vestal (80), who used a pulsed laser with single ion counting. Setting the analyzer to pass the selected ion mass, and increasing the instrument resolution until only one event was recorded for each laser pulse, they collected data until a statistically significant distribution of ion flight times was acquired for each mass of interest. The TSQ is electronically configured to operate with at least one analyzer quadrupole scanning a: range of masses (although that range may be <1 u wide) in all modes. A system having both quadrupoles set to pass a single mass had not been defined in programmable read only memory (PROM). Since the laser used in this experiment was a single shot device, the instrument had to be set up to provide a continuous signal for the ion selected, with time resolution of a few microseconds. This could have been done either by determining and adding the appropriate instruction or instructions to the PROM to allow each analyzer to be set to pass a single mass without scanning, or by causing the scanned quadrupole to operate in the RF-only mode to allow a range of masses to pass simultaneously rather than sequentially. Since modifying the PROM could have had far-reaching and unforeseen effects on other parts of the TSQ system, elimination of the DC voltage from the scanned quadrupole

PAGE 83

70 was selected as the best way to solve the problem. This was accomplished by changing the rod polarity switch located at the rear of the quadrupole electronics module (QEM) from "+" to "off". This is a three position switch ("+", "off", "-") that allows the operator to change the DC polarity on the quadrupole rods. Under normal conditions, there is no difference in response if the switch is set to "+" or "-". The "off" setting removes the DC component from the rods. The range of masses passed under these conditions can be determined from the stability diagram for the quadrupole. The stability diagram (Figure 5.1) is a plot of the ratio of DC to RF voltages imposed on the quadrupole rods vs. a function including the RF frequency, a dimensional constant for the quadrupole (one-half the diagonal distance between the rods in cm), the RF voltage, and the inverse of the mass-to-charge ratio of the ion. Under normal conditions, the voltage ratio is set so that only the mass (+0.5 u) under the apex has a stable trajectory through the quadrupole. All other ions have unstable trajectories and are lost by collisions with the rods. If the DC voltage component is removed, all ions that fall under the curve have stable trajectories and will pass. The ratio of the ordinate value under the apex (l/mj^) to that at the low-mass abscissa zero (l/m2) is about 7/9. Therefore, if an RFonly quadrupole is set to pass m^ , it will actually pass all masses greater than m2 which equals 7/9 of mj^ . Table III gives the effects on response for the most pertinent combinations of operation mode (parent and daughter ion experiment), programmed RF function (on/off) and rod polarity switch (+/off). Other

PAGE 84

Figure 5 . 1 Stability diagram for quadrupole mass filter. Plot of DC to RF voltage ratio vs. a function proportional to (m/z)~^. Shaded region shows range of ion masses passed through quadrupole set to pass m^ with both RF and DC voltages imposed (normally, mj^ ± 0.5 u). The RF-only quadrupole will pass all masses higher than m2 where m2 = (7/9)m]^. Parameters in the ordinate function are: RF voltage, Vj-f; quadrupole rod spacing (diagonal), rQ; electron charge, e; angular frequency of Vj-f, oj ; mass-to-charge ratio, m/z.

PAGE 85

72 dc rf NORMAL ION STABILITY REGION (SHADED)

PAGE 86

M

PAGE 87

74 combinations were not listed because they were unremarkable or had no bearing on the present discussion. The programmed RF function removes the DC component from the unscanned quadrupole. The effects were observed with the instrument under manual control with perfluorotributylamine (FC43) being introduced continuously into the EI source. Effects that include a measurement of relative signal intensity for the rod polarity switch on "+" or "off" were made by observing m/z 69 on the instrument oscilloscope. Loss of mass resolution is expected when the scanned quadrupole is operated in KF-only mode. Condition 3 was selected for most of the work on short-lived processes because Ql provides mass discrimination and Q3 can be set up to pass or not pass the ion selected by Ql. The advantage to this condition is that a long scan (20 s) can be acquired to eliminate the delay time of a few milliseconds normally required between scans. Further, since Q3 is not resolving masses, this mode will provide an ion intensity vs. time response for the selected ion rather than an ion intensity vs. mass response. In addition, the beginning of the scan can be detected by an oscilloscope attached to the "Y+" input of the TSQ oscilloscope. This is important for monitoring fast processes with the data system because data acquisition begins at the first regular scan after the experiment is initiated at the computer terminal. Data acquisition does not begin immediately because the scan sequence is not interrupted and restarted. Condition 6 is analogous to condition 3 but Ql is the scanned quadrupole (parent mode). Signal intensities were generally lower for this condition, however, the major

PAGE 88

75 drawback was that the scan signal was not visible at the "Y+" input as it was for the daughter mode. Comparison of Storage Oscilloscope and Data System Signal for Ni"^ from Nickel EM Grid With the TSQ operating in daughter mode with Q3 rod polarity switch off (condition 3), Ql passing m/z 58 (Ni"^) and Q3 passing all ions above m/z 56 (scanning from m/z 71 to 72 in 20 s with rod polarity switch off), the laser was fired at a clean nickel electron microscope grid inserted into the ion source on the 90° probe tip. The storage scope was attached directly to the electron multiplier through a high-gain differential amplifier module (Type D, Tektronix, Inc.) operated at 1 mv cra"l with a time base of 50 jus cm"l. In Figure 5. 2. A, traces 1 and 2 show response with the laser light blocked from the ion source and the light admitted to the source, respectively. Trace 3 was obtained by switching Ql to pass m/z 90 and Q3 to pass all masses above m/z 140 (scanning m/z 180 to 181 in 20 s), so that Ql and Q3 would be mutually exclusive and neither quadrupole should pass any ion produced by vaporization of the nickel target. In figure 5.2.B, traces 2 and 3 are analogous to traces 2 and 3, respectively, in part A, except that the response was monitored with the TSQ data system by acquiring in profile mode. Also, Q3 was scanned from m/z 60 to 70 (passing all m/z above 47 to 54) for trace 2 and from m/z 170 to 180 (passing all m/z above 132 to 140) for trace 3. These data show that this method of recording

PAGE 89

Figure 5.2 Plots of total ion current vs. time showing (A) storage scope data and (B) profile data for nickel EM grid target. In A, trace 1 is the response with laser light blocked from ion source; trace 2 in A and B is the response with Ql passing m/z 58 and Q3 passing all m/z above 56 (A) and all m/z above 47 to 54 (B). Trace 3 in A and B is the response with Ql passing m/z 90 and Q3 passing all m/z above 140 (A) and all m/z above 132 to 140 (B). See text for explanation of differing Q3 scans for storage scope and data system operation. Ion currents are given as counts (CNTS).

PAGE 90

mV 77 4 _ 2 TOTAL ION CURRENT (CNTS) 4000 2000 0.1 0.2 TIME (ms) 0.3

PAGE 91

78 intensity vs. time data for fast events may be valid, that quadrupole 1 provides some mass discrimination, even for low mass elemental ions, and that the storage scope and the data system give equivalent data. There is always a minimum signal (rapidly decaying spike) observed by the data system even when no sample is in the ion source, the laser light is blocked before it reaches the source, and the electron multiplier is off. This response is probably due to electrical interference from discharge of the laser storage capacitor when the flashlamp is fired. Instrumental Acquisition and Display Characteristics It might be helpful at this point to describe the TSQ scan parameter ( SYST em, System Status, Multiple Ion Detection), calibration ( CALI rate), acquisition (ACQUire), and display (PROFile, mass CHRO matogram, and SPEC trum) programs. The above programs will, henceforth, be referred to by the abbreviations used to access them (underlined). Data can be acquired in two modes, centroid and profile. Both are calibrated using the CALI program which establishes a linear relationship between mass and time elapsed since start of a scan. Therefore, even under normal circumstances, the mass of an ion is determined by the appearance time after the scan starts, and the spectrum is displayed as ion intensity vs. mass. Acquisition in profile mode causes storage of all points detected at the maximum sampling rate (25 jas). Storage requirements are maximized, but a relatively continuous plot of ion intensity vs. time is produced, allowing observation of peak shapes. It is this mode that has been used in these experiments on short-lived processes (Chapter V) to obtain a plot of

PAGE 92

79 signal intensity vs. time. A threshold intensity can be chosen (in ADC units), which allows immediate rejection of points having intensities less than the selected threshold. Thresholding is the only decision made by the data system at the full 40 kHz ADC rate. A setting of 1 means all data is recorded (no thresholding). For these experiments, the threshold was routinely set at 2 to eliminate spurious electronic noise. Therefore, a scan 20 s long, during which a signal is produced for a brief instant by a laser shot, would result in a profile lasting only from the time the signal exceeds threshold until it goes below threshold. PROF is the program used to display profile data. Although the program converts time to mass units, it plots essentially a timedependent, ion intensity curve, and calculates the time the signal is above threshold. The plots of profile data shown in the figures were calculated from these curves. Since time scales were usually different for each profile curve, signal intensities were replotted on a common time scale for comparison. The maximum intensity in a profile plot is 4096 counts, as determined by the 12-bit ADC. Q3 scan parameters were most easily set using the MID program. Ql mass was set using the SS program. In this way, experiments could be run rapidly by writing the required MID descriptors, and calling them as needed from SS. Ql mass was changed as required, also from SS. Acquisition in centroid mode (used for experiments on long-lived processes in Chapter VI) causes calculation and storage of the mass centroid of a peak, minimizing storage requirements and making data manipulation faster. In this mode, the centroid sampling interval is

PAGE 93

80 chosen (using MID) in 25 jus increments between 25 )ss and 200 ;us to give a reasonable number of data points per mass peak. Therefore, data is acquired at 40 kHz, and 1 to 8 data points are summed to give a single datum point to be used in calculating the peak centroid. If a scan is set up in SS, the highest possible centroid sampling interval (200 }is) will be used. In SYST, a selection is available for centroid samples per peak, however, the number actually used will be determined by the mass range covered in the scan time. The only way to insure use of a particular centroid sampling interval is to set that interval in MID. The CHRO program is used to view the recombined ion current (RIG), or the intensity profiles for each ion with respect to scan number (time). This is analogous to the use of the PROF program for viewing short-lived processes; however, an additional dimension is available (in the SPEC program), which allows display of the mass centroids making up the RIG (mass spectrum). Scan parameters for acquisition in centroid mode (Chapter VI) were again set up using the MID program, both for normal mass spectra and, as required, for parent and daughter ion spectra. The centroid sampling interval was chosen so that scan time could be minimized, and there would be at least 5 points/mass peak for calculation of the peak centroid. Filtering Characteristics of Quadrupoles I and 3 Although the quadrupole is supposed to effectively filter out all ions of unselected mass regardless of ion energy, it is possible for ions of inappropriate mass to pass through the quadrupole field without deflection if their kinetic energy is high enough (109, 113, 114).

PAGE 94

81 Therefore, a series of experiments was performed to determine the filtering capabilities of the system under the conditions used for time profile monitoring. Figure 5. 3. A and B show a comparison of storage scope and profile data, respectively, for phenytoin. Ql was set to pass m/z 275, (M+Na)"^ for phenytoin, and Q3 was scanned from m/z 343 to 353 in 18 s with the rod polarity switch off (passing all m/z above 266 to 274). The phenytoin (4 ^L of a solution of 4.35 mg/mL) was layered over a Na2C03 residue (2 ;aL of a solution of 25 mg/mL) on a 400-mesh copper electron microscope (EM) grid which was inserted into the ion source on the 90° probe tip. Trace 1 shows the first shot. Trace 2 is a second shot at the same site. Although this figure shows an intense signal for cationized phenytoin and significant reduction for a repeat shot (the grid was totally disintegrated at the point of impact of the first shot creating a hole about 0.5 mm in diameter), the signal is not necessarily due to cationized phenytoin alone. Figure 5.4 shows responses for a similar sample for which Ql was set to pass m/z 150 and Q3 was set to pass all m/z above 202 to 210 (scanning m/z 260 to 270 in 18 s with rod polarity off). In this case Ql and Q3 are mutually exclusive and Ql is not passing an ion that should be produced by the sample; however, Q3 is covering a mass range that includes ions that should be present. Trace 1 shows a shot at the center of the grid where the phenytoin concentration, due to uneven evaporation of the methanol, was lowest. Solids were usually concentrated in a ring describing the edge of the sample, droplet. Traces 2 through 4 show successive shots at the same

PAGE 95

Figure 5.3 Plots of total ion current vs. time showing (A) storage scope data and (B) profile data for phenytoin over Na2C03 on copper EM grid. In A and B Ql is passing m/z 275 and Q3 is passing all m/z above 267 to 275 (scanning m/z 343 to 353 in 18s with rod polarity switch off). In both A and B, trace 2 is the response from a second laser shot at the same site that gave trace 1.

PAGE 96

mV 83 20 10 0.2 0.4 0.6 TIME (ms) TOTAL ION CURRENT (CNTS) 4000 2000 0.4 1 1 — 0.8 1.2 TIME (ms) 1.6

PAGE 97

Figure 5.4 Plots of total ion current vs. time showing the efficiency of Ql and Q3 mass filtering (profile data) for phenytoin over Na2C03 on copper EM grid. In traces 1-4, Ql is set to filter out all m/z produced by the sample (passing m/z 150) and Q3 is set to pass ions that should be present (passing all m/z above 202 to 210. In trace 5, neither quadrupole is passing ions produced by the sample and they are set to be mutually exclusive (Ql passing m/z 700 and Q3 scanning normally with rod polarity "+" from m/z 300 to 310. Trace 1 is at grid center where phenytoin concentration is low. Trace 2 is at edge of droplet where concentration is highest. Traces 3 and 4 are successive shots at site 2. Trace 5 is from an identically prepared grid at the edge of sample droplet .

PAGE 98

TOTAL ION CURRENT (CNTS) 85 400 200 -

PAGE 99

86 site on the edge of this ring where concentration was highest. These traces indicate that Ql is not totally efficient at filtering ionic species. Trace 5 was recorded with Ql set to pass m/z 700 and Q3 scanning from m/z 300 to 310 with rod polarity "+" (DC on) so that both quadrupoles were mutually exclusive and neither was passing ions that were produced by the sample. A minimum signal, generally present under any conditions, was produced. This indicated that Ql and Q3 effectively filtered out sample ions when set to pass masses that differed greatly from the sample ion masses. Another experiment was conducted to determine whether the filtering capabilities of Ql were lessened when the selected mass was close to the mass of an ion produced by the sample (i.e., is unit mass resolution maintained under the conditions produced by laser desorption in the present ion source). Ql was set to pass m/z 275 and then, with a new grid, m/z 270 (test point arbitrarily chosen 5 u away from the ion mass), with Q3 passing all m/z above 202 to 210. The signal intensity ratio for the former with respect to the latter was about 5, indicating that unit mass resolution in Ql is not maintained under these conditions. The particles that are causing these unwanted responses may be fast ions or neutrals which do not "see" the low voltage fields of the lenses and quadrupoles. Discrimination of a selected ion mass from its nearest neighbors is less efficient than from ions that have masses much greater or much less than the selected mass. This problem was approached from another angle to show the total ion signal with all quadrupoles operated in RF-only mode, and to contrast that with the signal obtained when Ql was operated as a mass

PAGE 100

87 filter. Ql was set to pass m/z 275 and Q3 to pass all m/z above 266 to 274 with the sample again phenytoin over Na2G03. Successive shots were fired at the same site (Figure 5. 5. A, traces 1-3). Ql was then set to pass m/z 350 and the programmed RF was turned on so that Ql would actually pass all m/z above 272. Two shots were fired at the same site (new grid; Figure 5.5.B, traces 1 and 2). Programmed RF was then turned off so that Ql would then only pass m/z 350 (an ion not produced by the sample) and a single shot was fired at a new grid (Figure 5.5.B, trace 3). These figures show that Ql does actually filter out a large portion of the total signal that is possible under these conditions. Therefore, the combination of quadrupoles 1 and 3 can be used to obtain limited mass resolution in this rapid acquisition mode to allow monitoring of the fast signal profiles. Time Resolution of Low-Mass and High-Mass Ions With these limitations in mind, an attempt was made to observe the resolution of two compounds by time alone by operating the instrument in a time-of-flight mode without mass discrimination in either analyzer quadrupole. Ql was set to pass m/z 160 with programmed RF on (actually passing all m/z above 124), Q3 was set to pass all m/z above 117 to 124 (scanning m/z 150 to 160 in 18 s with rod polarity off). Data were acquired with the TSQ data system in profile mode. The sample was a mixture of 4-vinyl-4-aminobutyric acid (vinyl gaba, mol. wt . 129) and reserpine (mol. wt . 608) over Na2C03. Expected cationized species were m/z 152 and m/z 631, respectively. All quadrupole offset voltages and lens voltages were set at -20 V. The flight path length from the ion

PAGE 101

Figure 5.5 Plots of total ion current vs. time showing efficiency of Ql mass filtering and effects of programmed RF function in Ql. Fart A shows successive laser shots (1-3) at a single site on a copper EM grid with phenytoin over Na2C03 and Ql passing m/z 275 and Q3 passing all m/z above 267 to 275. Part B shows responses from an identical sample with Ql passing all m/z above 272 (Ql set at m/z 350 with programmed RF on); traces 1 and 2 are first and second shots at same site. Trace 3 is from a third identical sample under the latter conditions but with programmed RF off (Ql passing m/z 350).

PAGE 102

89 TOTAL ION CURRENT (CNTS) 2000 -I 1500 1000 500 4000 3000 2000 1000 0.4 0.6 TIME (ms)

PAGE 103

90 source to the detector was measured at 50 cm, making the total expected flight path about 75 cm since the oscillations of ions in the quadrupole fields make the effective length about 1.5 times the actual length (109). Figure 5.6 shows the ion signal from this experiment. The flight times calculated from the peak maxima were 247 ;us for vinyl gaba and 372 MS for reserpine. A plot of chart time vs. the square root of ion mass for these two points had a slope of 9.77, and an intercept of 127 JUS. This intercept indicates that there is a 127 ^s lag between the laser impact and the time when the data system detects the first nonzero signal. The instrumental constant (Eq. 4.4) is 9.77 ms u~^/2, and the apparent accelerating voltage seen by the ions is, from Eq. 4.3, 30.5 V. This is not grossly out of line, considering that the drift space is not field free, that the acceleration field imposed by the conversion dynodes at the end of the last quadrupole was not considered, and that no true accelerating pulse was provided at the ion source. There is also a possibility that this doublet does not truly represent time resolution of these two ions. Additional attempts to reproduce this pattern resulted only in peaks with shoulders; however, at no other time during these experiments was a doublet obtained when a single species was vaporized. Contribution of Reagent Gas and Sample Matrix to Ion Signal It appeared at this point that monitoring fast processes with the TSQ was interesting from an academic standpoint, but, because of limited mass and time resolution, was not very practical. Further, it did not have the potential for producing accurate numerical data on ion flight

PAGE 104

o J-" S -H -H -H -H CO to JJ vo 4-) 4-1 3 O S iH N H <

PAGE 105

92

PAGE 106

93 times or energy spreads. Time, it was decided, could be better spent observing processes that occurred over longer times, so that the capabilities of the TSQ could be put to better use. Figure 5.7 shows the effects of reagent gas on yield and longevity of signal for cationized phenytoin. Two identical samples of phenytoin over Na2C03 on copper EM grids were vaporized within the electron impact ion volume without reagent gas (trace 1) and with 0.58 torr of methane (trace 2). The filament was off. Ql was set to pass m/z 275; Q3 all m/z abov 244 to 252. This figure shows that reagent gas does not prolong the signal lifetime of the cationized molecule, and may not contribute to efficient extraction of this ion into the analyzer. A similar experiment was performed without Na2C03, but with parlodion, a nitrocellulose film material used in electron microscopy, added to the methanolic phenytoin solution (20 mg parlodion/mL) . This matrix appeared to cause the phenytoin to be more evenly distributed over the EM grid mesh after droplet evaporation (parlodion alone formed a clear film, but parlodion with solute was white). Q3 was scanning from m/z 170 to 180 with rod polarity off (passing all m/z above 132 to 140) while Ql was passing m/z 253. The experiment was performed with 0.64 torr of methane in the EI ion volume. Although the maximum ion current was less than that observed for the cationized molecule in Figure 5.7 (most likely because a different ion was being monitored), the lifetime of the signal was significantly increased (Figure 5.8). This is reasonable since the cationization process is directly dependent on thermal effects induced by the laser. The protonation reaction is

PAGE 107

Figure 5.7 Plots of total ion current vs. time for phenytoin over Na2C03 showing effects of reagent gas on signal for cationized molecule with Na2C03 present in sample. Traces show signal without reagent gas (1) and with 0.58 torr of methane (2) with filament off, Ql passing m/z 275, and Q3 passing all m/z above 244 to 252. All experiments done in EI ion volume.

PAGE 108

TOTAL ION CURRENT (CNTS) 95 1800 1200 600 _ I I 0.2 O.A TIME (ms) 0.6

PAGE 109

O 3 0) iH n « o ta o ^ 4-1 O -U 4-1 O
PAGE 110

97 m g

PAGE 111

98 dependent on conditions existing within the ion source (reagent gas, source pressure, filament on or off, open or closed source). This relatively long-lived signal is indicative of a reaction mechanism that may not be directly linked to the laser pulse. This last result gave renewed hope that signal lifetime could be increased by combining chemical ionization in a closed source with laser vaporization to produce a workable tandem mass spectrometer-based probe-type instrument.

PAGE 112

CHAPTER VI LONG-LIVED PROCESSES Initial experiments conducted to gain an impression of the possible success of laser microprobe MS/MS produced very encouraging results (115). In contrast to the reports of other workers using single shot lasers, signal lifetimes of seconds, rather than microseconds or, at most, milliseconds, were observed from a number of compounds, irradiated in a CI source pressurized with 0.9 torr of methane. No signals were observed that corresponded to cationized molecules or metal ions, species usually linked directly to the laser event; however, a mixture of nine anticonvulsant drugs was analyzed and each component was detected as a daughter ion which resulted from collisional dissociation of the corresponding molecular ion. This greatly increased the selectivity of the method and indicated that the system had possibilities for probe-type determinations of single components in mixtures. This chapter presents a more in-depth investigation of these results, with the purpose of explaining the apparent disparity between the preliminary results outlined above and the work of other investigators. A model for the mechanism of ion production is developed and used to show the dependence of the process on source 99

PAGE 113

100 temperature and pressure, and possible dependence on substrate material. Finally, the laser-induced process and the model are evaluated for their analytical potential. Long-Lived Signal Profiles Extending the lifetimes of laser-induced signals is of primary • importance if they are to be useful in quantitative analysis. Since the quadrupole mass spectrometer has the capability of scanning at 1 to 5 u ms"-'-, depending on the centroid sampling interval selected, this requirement can be satisfied if signals can be made to last for one or more seconds. To meet this requirement, the closed CI ion source was substituted for the open EI ion source used for the work reported in Chapter V. Figure 6.1 shows the signals produced by laser evaporation of carbamazepine epoxide (mol wt 252) in parlodion film inserted on a copper EM grid into the CI source heated to 160°C and pressurized to 0.90 torr with methane. The centroid sampling interval was set to 50 ^s for this experiment, and Q3 was scanned from m/z 240 to 310 in 29 ms to give a normal mass spectrum of the molecular ion region (2 ms delay between scans). Traces 1-3 in Figure 6.1. A show successive shots at the same site. Trace 4 is from a new site on the same grid. The diameter of the hole produced by the laser was about 0.5 mm for 18 kV laser driver voltage and 8 mm laser cavity diameter. Figure 6.1.B shows trace 4 on an expanded time scale, demonstrating a decay back to zero

PAGE 114

0) (0 ro -aCO O CO 4-1 Oj u-1 nj s 4-1 w 4-) W l-i Cu C oi B H u « S-l X C -U 1-1 4J i-i O 03 -H so C C aj E --I « 3 0) O 03 o o 0) W (-1 (\3
PAGE 115

102 J

PAGE 116

103 intensity after about 1.7 s. This recombined ion current (RIC) consisted of the protonated molecular ion at m/z 253 and normal methane CI adducts, (M+29)"'" and (M+Al)"*", all of which persisted throughout the signal lifetime. Ion signals having this longevity allow scanning over a reasonable mass range during the time the signal is present, and allow use of CAD conditions to provide the selectivity necessary for molecular analysis of mixtures. Figure 6.2 shows laser-induced ion signals from (A, B) tetrabutylammonium perchlorate, (C) menthol glucuronic acid, (D) adenosine, and (E) acetazolamide , respectively, acquired under the same conditions used for carbamazepine epoxide (normal mass spectrum). The most intense ions are the protonated molecular ions for adenosine (m/z 268) and acetazolamide (m/z 223), and a thermal degradation product for menthol glucuronic acid (m/z 139, M+H-glucuronic acid). These ions were present throughout the signal lifetime. The most intense ions from tetrabutylammonium perchlorate initially were the degradation products, m/z 184, 185, 186 (M-butyl fragments). These made up the initial spike, however, there was a considerable contribution (A) from the tetrabutylammonium ion (m/z 242) to signal B, which rapidly became the predominant ion, indicating that degradative processes were not totally responsible for the signal. All laser-induced RIC signals had similar profiles showing an intense initial spike that lasted for one to several scans before decreasing rapidly to a more slowly decaying signal, which sometimes began with an obvious secondary wave (D and E). Specific ion

PAGE 117

Figure 6.2 Plots of ion current vs. time for laser vaporization of (B) tetrabutylammonium perchlorate, (C) menthol glucuronic acid, (D) adenosine, and (E) acetazolamide with Q3 scanned over the appropriate 70 u mass range in 29 ms with a 2 ms delay between scans. Base peaks are (B) m/z 184, (C) m/z 139, (D) m/z 268, and (E) 223. The upper trace (A) shows the profile of the tetrabutylammonium ion from tetrabutylammonium perchlorate. B-D are reconstructed ion currents. Profile A is single ion current.

PAGE 118

105 RIC (CNTS) 25216 AvKWhV^ '^^\V^, '^VHyv.^ 'Wa^. ''Vv*^ 354816 -, 472576 ''"^'^Ww.i^. 1382390 91586 3.0 6.0 9.0 TIME (s) 12.0 15.0

PAGE 119

106 signals, primarily the protonated molecular ion and typical adduct ions (e.g., (M+C2H5)"^ and (M+C3H5)"^ for methane reagent gas) generally, but not always, persisted (within the limits of sensitivity) throughout both spike and wave sections. The scan being acquired at the time the laser was fired always contained an intense ion signal, due to electronic interference, that corresponded to the mass being passed at that instant. All scans after the first contained only ion signals resulting from ionization (protonation) , adduct formation, and fragmentation of the sample. The secondary wave (more noticeable in D and E than in B and C) indicates that a second mechanism, which may be laser-induced but continues to operate after the thermal effects of the laser should have subsided, is responsible for part of the signal. The appearance of an obvious second wave is not linked to specific compounds or functional groups, but occurs randomly with all compounds that exhibit intense, long-lived ion signals. Characteristics of laser induced ion signals for other test compounds are given in the appendix and show signal lifetimes (time for signal to reach 10% of maximum) which varied from about 0.1 s to 70 s. Desorption Chemical Ionization As is often the case in research, several serendipitous occurrences provided clues to the origin of the signals. The first of these occurrences was the replacement of an ailing electron multiplier in the TSQ, and the second was the modification of the 45° probe tip to allow samples to float in the ion source without the stainless steel backing that the 90° probe tip provided. The new electron multiplier was

PAGE 120

107 extremely sensitive, and could be operated at 30-40% of the voltage required with the old multiplier. It was only after this multiplier was installed that spontaneous ion signals were obtained prior to firing the laser when an EM grid with sample was inserted into the CI ion source on the 90° probe tip. Modification of the 45° probe tip (henceforth called the DCI probe tip) enhanced this process (apparently by more efficiently exposing the sample to the reagent gas plasma) so that it was observed whenever a sample was introduced and not just occasionally as with the 90° probe tip. This phenomenon occurred with all compounds that could be vaporized from the solids probe by direct heating, and required no stimulus other than being placed in a heated source with reagent gas (filament on). This process is not new and was reported by Baldwin and McLafferty (116) in 1973 and called Direct Chemical Ionization. Since then it has been investigated by a number of other workers (91, 110, 117-120) and is now commonly referred to as Desorption Chemical Ionization (DCI) . The process (not laser-induced) is shown in Figure 6.3 for phenytoin, a relatively nonvolatile compound with a melting point of 294-298°C. The RIC is shown along with the solids probe temperature for CAD of m/z 253 with 2 mtorr of nitrogen in Q2 and a collision energy of 18 eV. Q3 was scanned from m/z 100 to 300 in 2 s. Source temperature was 16G°C and the methane reagent pressure was 0.90 torr. The probe was cooled to 10°C using freon and allowed to reheat by conduction from the source block. After the ion signal had risen to a plateau, the probe was cooled to -7°C. For this experiment, about 9 ug

PAGE 121

Q>

PAGE 122

109 2: u W o

PAGE 123

110 of phenytoin was evaporated onto a copper grid from a methanol solution containing parlodion. The grid was placed in the source directly in front of the electron entrance aperture. All daughter ions followed this profile. Figure 6.4 shows negative ion CI/CAD summed ion current profiles and solids probe temperature profiles for direct probe evaporation (evaporation from a glass solids probe vial by resistive heating of the solids probe) and DCI of 7.2 ng of Hydergine (ergotoxine methanesulfonate) . The reagent gas was methane :nitrous oxide (100:11) at a combined pressure of 1.11 torr. A collision energy of 25 eV was used with nitrogen collision gas at a pressure of 2 ratorr. Source temperature was 160°C. The sample was evaporated from methanol solution in the tip of a glass solids probe vial for the direct probe CI experiment, and on a 400 mesh nickel EM grid for the DCI experiment. Ergotoxine is a 1:1:1 mixture of three hydrogenated ergot alkaloids, dihydroergocristine (mol wt 611), dihydroergocryptine (mol wt 577), and dihydroergocornine (mol wt 563), each made up of dihydrolysergic acid (mol wt 270) and a cyclized tripeptide. The lysergic acid portion (as dihydrolysergamide, mol wt 269) is lost in the CI ion source to give, almost exclusively, ions characteristic of the polypeptide portion of each molecule. These ions undergo neutral loss of mass 99, again almost exclusively, to give representative daughter ions for each alkaloid (122). Therefore, Ql was set to pass the (M-269)~ ions for each component, m/z 294, 308, and 342. Q3 was scanned from m/z 180 to 350 to detect the selected parent ions as well as the respective daughter ions

PAGE 124

Figure 6.4 Plots of RIC vs. time and solids probe temperature vs. time for (A) direct probe evaporation from a glass solids probe vial, and (B) desorption chemical ionization from a 400 mesh nickel EM grid of hydergine (dihydroergotoxine methane-sulf onate) . Negative ion CI/CAD with methane: nitrous oxide (100:11) with Ql passing successively ra/z 294, 308, and 342 and Q3 scanning from m/z 180 to 350 in 0.145 s. Nitrogen collision gas pressure: 2 mtorr; collision energy: 25 eV.

PAGE 125

112 11328 RIC (CNTS) 373 TEMP (°C) 44 Jl ...amAIi 1 B 22848 1 RIC (CNTS) 332 TEMP (°C) i i ' ^iii^.>Lui^i v. 1. .-..j.. J .. JJl 43 — f— 15 30 60 75

PAGE 126

113 arising from loss of 99, m/z 195, 209, and 243. All ions gave similar ion current profiles. As can be seen from the probe heating cycle (B) initiated after the DCI signal had decayed to baseline, essentially all of the sample vaporized while the probe remained at ambient temperature (43°C). In contrast, significant vaporization from the probe vial did not begin until the heating cycle (A) was almost completed (247°C). This is consistent with the findings of other investigators (116, 120) who reported vaporization of compounds by DCI at temperatures about 150°C lower than those required for direct probe evaporation. Effect of sample position in source . Figure 6.5 shows the effect of the position of the EM grid in the source on the ion current during DCI (normal mass spectrum, Q3 scanning from m/z 100 to 300 in 2 s). The inset shows the relative positions within the ion source. The sample (phenytoin in parlodion film on copper grid) was inserted into the source, suspended from the end of the DCI probe tip, so that the grid would initially be in front of the electron entrance (A) . The solids probe was then rotated to position the grid successively in front of the reagent gas inlet (B), the collector (C), the laser aperture (D), the electron entrance (A), the reagent gas entrance (B), and the collector (C). Relative ion intensities were similar throughout the experiment. The data show that the sample was most readily desorbed or vaporized when it was directly in front of the reagent gas inlet. If the grid were inserted on the 90° probe extension which prevented the reagent gas from directly hitting the grid, DCI was delayed or prevented, and, if it occurred, was less intense.

PAGE 127

(U ^ T3 '+-< 4-i C C U 4-1 0) O O -H O l-i -H -i-J -H 3 S to -O O u u U C C w
PAGE 128

115

PAGE 129

116 The above experiment, and the fact that sample was not vaporized from a grid placed in the CI source under the same conditions but without reagent gas, pointed to the action of the reagent gas molecules on the sample surface as the primary mechanism of sample vaporization. Since the optimum sample position was directly in front of the reagent gas inlet, this mechanism may even involve a sample bombardment by reagent gas entering the source. This was further proved by turning the filament on at different times after inserting a series of identical samples (sodium phenytoin) into the CI source with reagent gas present (helium at 0.90 torr). Helium was selected as the reagent gas for this experiment because it caused desorption of molecules, but allowed ionization and fragmentation primarily by electron impact (121). The positive He-CI spectra of phenytoin free acid and sodium phenytoin were similar and showed extensive fragmentation of the molecular ion. The major ions were m/z 165 (100%) and m/z 180 (80%). The molecular ion for both forms of phenytoin was m/z 252 (M"^, 8%). The sodium salt of phenytoin was used in this experiment for several reasons. First, no signal could be obtained when 800 ng of the salt was placed in a probe vial and heated to 353°C. Second, while no signal was obtained for 800 ng of salt placed on a copper grid and subjected to DCI conditions, a substantial signal was obtained when a nickel grid was used. Third, the presence of a preformed ion (PHT~) in the sample might provide a signal if ablation occurred when the filament was off. Figure 6.6 shows the results of this experiment (normal positive mass spectrum), in which the sample was inserted into the source (160°C) on the DCI probe extension

PAGE 130

Figure 6.6 Plots of RIC vs. time for DCI of sodium phenytoin inserted into source at 160°C with 0,90 torr of methane. Filament turned on (A) before insertion, (B) 12 s after insertion, and (C) 34 s after insertion. Points 1 and 2 show the points at which filament came on in B and C, respectively.

PAGE 131

118 RIC (CNTS X 10 ^) 30 20 _ 10 TIME (s)

PAGE 132

119 (A) with the filament on, (B) with the filament off initially but turned on after 12 s, and (C) with the filament off initially but turned on after 34 s. Apparently, ablation occurs as long as reagent gas is impinging on the sample, whether or not the filament is on. Signals B and C begin when the filament is turned on, but appear to pick up the profile of signal A at that point that corresponds to the time the sample has been in the source. There is no signal, however, when the filament is off, so the molecules that vaporize or desorb during that time, do so as neutral species and are not observed. The molecular ion was evident in A only for the first 5 to 6 s. Thereafter, the m/z 165 and 180 fragments made up the total ion current in A, B, and C. The intensities of these two ions summed over the time of each curve were 5.49 X 10^, 1.32 X 10^, and 1.32 x 10^ counts for A, B, and C, respectively. These values appear to indicate that there is a relatively loosely bound (possibly by inter-molecular forces) group of molecules, which can be easily vaporized. There is, however, another group of more tightly bound molecules (possibly by molecule-surface binding forces), which requires input of more energy to be vaporized. Otherwise, the summed ion intensity for curve C would be less than that for curve B, since the sample was in the source about 3 times longer before the filament was turned on. Therefore, the ionized reagent gas must impart more energy to the sample surface than the neutral gas molecules.

PAGE 133

120 Laser-Induced (Enhanced) Desorption Chemical Ionization Since the process of DCI could be stopped by cooling the grid indirectly through the solids probe, the question remained as to the effect of cooling on the laser-induced process. As it turned out, cooling the grid to -20°C had no effect on the response elicited when the laser was fired. Figure 6.7 shows this for phenobarbital deposited on a copper grid with parlodion (normal positive methane CI spectrum, Q3 scanned from m/z 160 to 300 in 33 ms, 75 ;as centroid sampling interval). The plot of RIC vs. time and the corresponding plot of solids probe temperature vs. time show the presence of the DCI signal (1) and the decrease in DCI with probe cooling. The effect of probe cooling is not so immediate as with phenytoin because the melting point (174-178°C) of phenobarbital and, presumably, the heat of vaporization are much lower than those for phenytoin. The laser was fired (2) when the ion current had decreased to a low level, and the signal that resulted contained the initial high spike and long tail, typical of the process. This process can therefore be called Laser-Induced or Laser-Enhanced Desorption Chemical Ionization (LIDCI or LEDCI), depending on whether the DCI signal is absent or present when the laser is fired. Ions present during the DCI phase and after the laser was fired included the protonated molecular ion, (M+29)"^, and (M+41)+. Point 3 on the RIC curve indicates the point at which the laser was fired again at the same point on the sample. The ion current dropped rapidly to zero, evidently because the input of additional energy hastened the clearing of the ion source, and no additional sample was available at the point of focus.

PAGE 134

•H C ^-^ -H J-i -H M -H C i-j O
PAGE 135

122

PAGE 136

123 Cotter (60) reported an effect on reagent gas peak intensities, (transient 2to 10-fold increase in m/z 56, 57, and 58 from isobutane) which occurred over about 5 ms after the laser was fired. Therefore, it is not unreasonable to think that species already vaporized in the source could be affected by acquiring an additional kinetic energy, which either decreased the efficiency of ion extraction, or caused extensive fragmentation to species outside the mass scan range. Model for Laser-Induced Desorption Chemical Ionization There is some question as to whether the term, desorption , is applicable in this context, since the process appears to be a thermallyinduced vaporization, which involves input of energy sufficient to overcome intermolecular interactions. There is still some question as to whether the process involves the input of energy sufficient to overcome a molecule-to-surface binding energy (desorption) . Actually, both processes appear to be at work, the latter predominating when concentrations of analyte are low (a low surface density of sample). The following discussion gives a physical description of the laser effects, and also gives evidence showing that the substrate contributes to sample ablation. Dependence of response on ion source temperature and pressure is determined using DCI as a model for LIDCI. Pattern of sample removal by a single laser shot . Incidence of focused laser radiation on the EM grid containing a solute residue in a parlodion film results in a specific pattern of sample removal. Each 400 mesh grid (copper, nickel, or silver) is 3 ram in diameter, about

PAGE 137

124 13 }im thick, having a regular square mesh with holes 37 urn square defined by 25 ^um thick "wires" (the grid is acually of one-piece construction). The grid is totally vaporized at the point of focus of the laser. In that area of the target, therefore, it would be difficult to propose a mechanism based on transfer of energy from the substrate to the sample. There is, however, a circular area around the hole, having about 2 times the hole diameter, that appears to be cleared of sample with the grid still intact. Examination of the grids under a microscope (lOOX) revealed that the tips of the wires showed signs of melting (they were bent or of smaller diameter than the rest of the wire. This is to be expected and indicates a gradation of heat emanating from the point of laser focus. Therefore, it is reasonable to propose more than one mechanism of sample ablation to correspond to the different temperature zones present. Amount of sample removed by a single laser shot . Ion current profiles obtained for identical samples were variable with respect to RIC and, therefore, the amount of sample removed from a grid by a single laser shot was determined by high performance liquid chromatography. To eliminate effects from DCI, the reagent gas was turned off and 5 of 10 identically prepared copper grids (2 ;jL of a mixture of 0.40 mL of methanol plus 0.20 mL of parlodion solution containing 20 mg/mL of methanol : ether (1:3)) were placed sequentially in a 160° CI source on the 90° probe extension. Each grid contained 1.11 jug of phenytoin in a residue about 1.7 mm in diameter (48 Mg phenytoin cm"^). xhe laser was fired once at each grid, and the residue left on each grid (as well as

PAGE 138

125 the residue on each of the other 5 control grids) was extracted and analyzed by high performance reversed-phase liquid chromatography with p-methylphenobarbital (MPB) as the internal standard. The results of the analysis including visual (microscopy at lOOX magnification) determination of the hole areas and areas from which the film was vaporized around the hole are shown in Table IV. From the average amounts of PHT recovered from the test (0.92+O.OA ;jg) and control ( 1 .05+0.03 >ig) grids, it appears that an average of 13% of the PHT was removed by the laser shot. The average area of the residue noticeably affected (hole plus de-filmed area) is 18%. These values indicate that the distribution of solute in the film may not be as even as it appears, with solute still being concentrated near the edge of the residue (laser was focused in the center of the grid). Even so, the residue is noticeably more evenly distributed than it is without the parlodion film. Reproducibility of ion current profiles and relative ion intensities . Figure 6.8 shows plots of RIC vs. time for LIDCI (normal Q3 mass spectrum) of a mixture of PHT and 1 , 3-^5N-2-13c_phenytoin (PHT*) (2 jjL of a mixture of 0.40 mL of a solution of 9.60 mg/mL of PHT and 8.94 mg/mL of PHT* in methanol plus 0.40 mL of a solution of 20 rag/raL of parlodion in methanol :ether (1:3)). This solution was evaporated on 5 copper grids, and each grid was inserted into the ion source on the 90° probe tip. Two shots were fired at each grid on opposite edges (but within the boundaries) of the residue. Plot A shows all first shots, and plot B shows all second shots. All traces decayed to baseline

PAGE 139

126 LO O tN 00 lO I I < < -d
PAGE 140

(U 4-1 60 CQ •H ro o o t-i O 0) c •H ro
PAGE 141

128

PAGE 142

129 within 4 s. Table V gives the ratios of the integrated (M+H)''" ion intensities for PHT (ra/z 253) to those of PHT* (m/z 256), calculated by summing the scans from the first scan after the laser shot to the scan whose RIC was 25% of that of the first scan. The scan being acquired at the time the laser was fired was always eliminated from calculations, because this scan contained an intense artifact peak which appeared at the mass being passed at the instant of laser firing. Although the RIC curves were variable in intensity, relative ion intensities were reproducible for the protonated molecular ions. These ions accounted for between 82% and 99% of the total RIC. The only other ions present were the normal methane CI adduct ions. Maximum RIC at the scan after the laser shot varied between 8448 and 172288 counts. This variability in RIC is not unexpected, because laser ablation and laser desorption mass spectrometry have been plagued from the outset by a pronounced irreproducibility of relative and absolute ion intensities (32). The relatively stable ion intensity ratios, however, provide hope that, if ion lifetimes can be extended to allow signal averaging, the precision required for quantitation may be obtained. Effect of substrate on PCI signals . The uncertainty over whether there is a desorption process occurring in addition to a vaporization process was resolved by comparing signals obtained with different sample substrates. An analyte stock solution (graded melting point standard) was made up which contained a series of compounds having melting points spanning the range from 76°C to greater than 300°C (Table VI). A working solution was made by adding 1 ;jL of stock to 100 ^L of methanol,

PAGE 143

Table V. (M+H)+ Ion Intensity Ratios for LIDCI of PHT and PHT* Deposited with Parlodion on Copper Grids. Comparison of 2 Shots at Different Sites on Each of 5 Grids 130 Ion Intensity Ratio^ Grid 1 2 3 4 5 ave. ± SD ave ± SD for all shots: irst

PAGE 144

131 vo o> ro r~00 o\ m ^H CO so to vo ^ og rsi fM CM CN n CM oo LO — I CO vD LO \£) ^H CN CM OJ CN fsl • c CO

PAGE 145

132 and 1 jjL was spotted on 400 mesh copper, nickel, and silver EM grids. Each was inserted into the CI source on the DCI probe tip. The source was maintained at 180°C and was pressurized with 0.90 torr of methane. Q3 was scanned in the normal mass spectrum mode from m/z 100 to 320 in 0.1 s. Figure 6.9 shows the ion current vs. time curves for the protonated molecular ions for the copper grid (A), the nickel grid (B), and the silver grid (C). The copper grid does not give a continuous signal for the highest-melting compound, HPHT . Vaporization (desorption) envelopes begin at later times as compound melting point increases. This is more easily seen for the copper grid since compounds appear to be more strongly held to that material than to the nickel or silver. It is possible that there is some complex formation between the sample components and the copper (123, 12A), rather than a simple physisorption of the sample on the less reactive metals, nickel and silver (124). It is doubtful that these effects can be correlated with thermal conductivities of the metals, since those of copper and silver are about equal (3.98 and 4.27 W cm~^ °C~^, respectively), and that of nickel is much lower (0.899 W cm"^ °C~M. The only species for which the maximum ion current for copper is higher than that for nickel and silver is MPS, the compound having the lowest melting point (Table VII). Summed ion currents are greater in all cases for nickel than for silver, but greater for the lower melting point components on copper than on nickel. Also, while maximum and summed ion currents decrease as melting points increase for sample components on copper, these values appear less dependent on component melting point for nickel and silver. These

PAGE 146

Figure 6 . 9 Plots of ion current vs. time for protonated molecular ions from graded melting point stock 1:100 dilution (Table 6.3), DCI at 180°C from (A) copper, (B) nickel, and (C) silver 400 mesh EM grids with 0.90 torr of methane. Ions are for MPS (156), MES (219), PB (233), PHT (253), and HPHT (269).

PAGE 147

134 ION CURRENT (CNTS x 10 ) m/z 156 147 504 651 601 30 m/z 156 m/z 219 m/z 233 m/z 253 m/z 269 123 -

PAGE 148

135 Table VII. Comparison of Maximum and Summed Ion Currents for DCI of Graded Melting Point Stock 1:100 Dilution^ from Copper, Nickel and Silver EM Grids Ion Current (Counts) Grid MPS MES PB PHT HPPH Cu

PAGE 149

136 results indicate that there is a substrate effect, and that, of the three metals used here, nickel and silver are preferable to copper because there appears to be less of an interaction between the substrate and the sorbed species. This, in turn, gives an evaporation profile that is less dependent on melting point or heat of vaporization of the analyte. The substrate effect is an indication that true desorption (molecules in contact with surface) may be occurring in addition to vaporization (molecules in contact with other molecules). Investigators who have studied DCI have shown an enhancement of the process when nonreactive substrates such as Teflon (118) and Vespel (111) are used. Therefore, desorption profiles for DCI of the graded melting point standard on nickel and Teflon were examined by evaporating identical 1 ^uL samples onto nickel grids and a Teflon "paddle" made by flattening the end of a short piece of 1.5 mm o.d. capillary tubing. The length of the Teflon probe extension was adjusted so that the paddle would be directly in front of the reagent gas inlet when the probe was inserted into the source (sample side facing inlet). Dilutions of the graded melting point standard (Table VI) were made to give concentrations ranging from about 30 to 300 ng/;aL of each drug. The mass spectrometer was set up to acquire in parent spectrum mode sequentially monitoring the daughter ions from the corresponding parent ions as given in Table VI. Figure 6.10 shows the difference in the desorption profiles for a sample containing about 60 to 76 ng of each drug evaporated on nickel and Teflon. Vaporization according to melting point is more noticeable for the sample on Teflon; each ion profile is

PAGE 150

Figure 6.10 Plots of ion current vs. time for desorption of graded melting point standard (60-76 ng of each drug) from (A) nickel EM grid, and (B) Teflon probe extension. Selected reaction monitoring parent spectrum. See Table VI for standard composition and parent ion-daughter ion designations. Discontinuous profiles are due to scan characteristics of TSQ operating in MS/MS mode.

PAGE 151

138 ION CURRENT (CNTS) 79616 , il 156^ -^ se"*" llllhin 176128 -, 219^ ^ 134^ HfiUAAiiiii... 90240 46336 233^ ^ 162^ HuUtAum....... 1.^ !!!!!!!! nil 253^ ^ 182"'' f.lll.ll..l..»..'..UJ.(.U i>iii.ii.ii I.. 1.. 31712 -, 267^ ^ 196"^ Mh lHhUitiiuitiiiini>i...i...^...iA... 3.0 6.0 3460 -1 6464 2496 2184 1 2536 9.0 156"^ ^ 86'*'

PAGE 152

139 nearly Gaussian, and lasts 7 to 10 times longer than that for the same sample on nickel. The maximum ion current for the sample on nickel, however, is 13 to 36 times greater, and the area under the curve is 2.5 to 5.6 times greater (depending on the compound, but not correlating with melting point) for the sample on nickel. The profile more closely resembles a Poisson distribution. These results are indicative of a primarily thermal process occurring on Teflon, which is impeded by the low thermal conductivity of the substrate. The sample seems to be efficiently desorbed from the nickel, and the metallic surface does not hinder, but enhances the process, because its higher thermal conductivity increases the rate of heat transfer to the sample. The nickel may, in fact, be better for this reason than the inert but less conductive materials. In contrast to copper and stainless steel, nickel has been shown to be nonreactive with respect to labile compounds when used as a column material in gas chromatography (123). Another reason for the faster desorption from the nickel grid could be simply that the area available for contact with the reagent gas is greater with the 400 mesh grid than with the solid Teflon paddle. Effect of source temperature and pressure . The influence of source temperature and pressure on DCI and LIDCI were examined in view of the results given above which indicate a thermal process, dependent upon close contact between the sample and the reagent gas. Sodium phenytoin (411 ng) was evaporated from methanol solution on nickel EM grids and inserted on the DCI probe tip into the ion source. The mass spectrometer was set up to acquire normal positive ion/negative ion mass

PAGE 153

140 spectra, scanning from m/z 160 to 300. For the temperature experiments, the source pressure (methane) was maintained at 0.70 torr, and the source temperature was varied from 90°C to 180°C. For the pressure experiments, the source temperature was maintained at 160°C, and the pressure was varied from 0.25 torr to 0.98 torr. Figures 6.11 and 6.12 show the RIG vs. time profiles for the temperature and pressure experiments, respectively. As would be expected, ion signal decreases in intensity and longevity as temperature or pressure is decreased. Extent of fragmentation for both negative and positive molecular ions decreased with decreasing temperature and pressure. Negative ions predominated over positive ions as temperature and pressure decreased; however, both negative and positive ion spectra showed the same trends. Figure 6.11 shows the negative ion RIG because there was no positive ion current at the lowest source temperature. Figure 6.12 shows the positive ion RIG. Temperature and pressure conditions are also interdependent, in that a higher temperature will compensate somewhat for a lower pressure. In another experiment with phenytoin free acid, the threshold pressure for desorption was 0.18 torr at a source temperature of I60°G and 0.35 torr at 130°G. These temperature and pressure requirements were also observed for LIDGI; however, it was possible to obtain signals upon firing the laser when temperature and pressure conditions were not sufficient for DGI. Figure 6.13 shows the results of an experiment (normal mass spectrum with selected ion monitoring) in which the graded melting point standard on a nickel grid was inserted on the 90° probe tip into a 160°G source pressurized with

PAGE 154

Figure 6.11 Plot of negative ion RIC vs. time for DCI of sodium phenytoin at various source temperatures. Source temperatures are (1) 180°C, (2) 140°C, (3) 120°C, (4) 100°C, (5) 90°C. Source pressure: 0.70 torr .

PAGE 155

142 RIC (CNTS X 10 ) 400 200 TIME (s)

PAGE 156

Figure 6.12 Plot of positive ion RIC vs. time for DCI of sodium phenytoin at various source pressures. Source pressures are (1) 0.98 torr, (2) 0.71 torr, (3) 0.52 torr, (4) 0.34 torr, (5) 0.25 torr. Source temperature 160°C.

PAGE 157

144 RIC (CNTS X 10 ) 340 255 170 85 TIME (s)

PAGE 158

TD a> c • • •
PAGE 159

146

PAGE 160

147 0.90 torr of methane. The component having the lowest melting point (MPS, It-ll^Q.) began to vaporize before the laser was fired (1), and DCI resumed after the ion source was cleared by a second laser shot (2) at the same target site. This is an example of LEDCI. The signal for the second most volatile component, PB, terminated at the second laser shot, again due to a rapid clearing of the source by the input of additional energy, but did not resume because conditions were not conducive to DCI. None of the other components desorbed prior to the laser shot and they exhibited LIDCI signals having lifetimes and intensities which were dependent upon the melting point (heat of vaporization). This is a definite disadvantage and portends difficulties in use of LIDCI for quantitative analysis for two reasons. First, the versatility of the method will be limited because compounds having high melting points will give short-lived signals. Second, sensitivity will decrease with increasing melting point of the analyte resulting in selective detection of low melting species and discrimination against high melting species. Quantitative Analysis and Application of LIDCI In view of the above findings, numerous attempts were made to obtain analytical curves for mixtures of antiepileptic drugs using DCI and LIDCI. The LIDCI experiment in which reproducible ratios of PHT:PHT* were obtained (Figure 6.8, Table V) were very promising; however, this experiment was run with the TSQ in the normal Q3 mass mode. Since the forte of tandem mass spectrometry is the analysis of

PAGE 161

148 mixtures without prior chromatographic separation, but with a mass separation followed by CAD and mass analysis of component fragments, calibration experiments were performed with the instrument operated as a true tandem mass spectrometer. Lower Limit of Detection for PCI The lower limit of detection was checked for hydergine by DCI from a nickel grid in a 160°C source pressurized with methane:nitrous oxide (1.00:0.11) at 1.11 torr to give a mixture of acidic (C2H5"'") and basic (0H~) reagent ions (125). Figure 6.1A shows the RIG plots for the negative parent ions obtained in a selected reaction monitoring experiment with Q3 successively selecting m/z 195, 209, and 243, and Ql passing each corresponding parent ion, m/z 294, 308, and 342, respectively. Ion current profiles for all three compounds were similar. Plot A shows the desorption envelope for 0.558 ng of hydergine, and plot B, for 5.42 ng. A blank run under the same conditions had a peak signal level of 294 counts, as compared with signal maxima of 1160 counts (3 times blank level) for the low standard and 4640 counts (15 times blank level) for the high standard. If the lower limit of detection is taken to be the amount of sample giving a signal 5 times the blank level, the limit of detection is 1.42 ng . A lower limit was not determined for LIDCl because of experimental difficulties in working with small sample weights in the present instrument system, however, the DCI limit gives an idea of the possible LIDCI limit for an optimized system. Another observation that can be made from this figure is that the signal life-time is dependent on

PAGE 162

v^

PAGE 163

150

PAGE 164

151 sample size. Signals for the 0.558 ng sample and the 5.42 ng sample decayed to 25% of maximum in 1.4 s and 3.7 s, respectively. This could be a problem and must be considered if any further work is done in this area. When detection is time-limited, the signal lifetime must be independent of concentration if the optimum limit of detection is to be realized. Calibration Curves for PCI and LIDCI PCI calibration for graded melting point standard on nickel grids and Teflon paddle . These calibrations were done using a dilution of the graded melting point standard which provided standard solutions containing from 30 to 380 ng/pL of each analyte. Table VllI gives the make-up of the calibrant solutions. Figures 6.15 and 6.16, respectively, show the curves of analyte-to-internal standard response ratio (parent ion experiment, selected reaction monitoring according to Table VI) obtained for each drug upon PCI from a nickel grid and a Teflon paddle. Lines shown for 2-methyl-2-propylsuccinimide (MPS), and phenobarbital (PB) were calculated by using all but the highest calibrant. The line for mephenytoin (MES) was calculated using only the data for the three low calibrants. PHT gave the best curves with coefficients of determination (r^) of 0.955 for nickel and 0.988 for Teflon (all calibrants included). The other three drugs showed a change in slope to negative values at high calibrant concentration. These results indicate the importance of choosing the correct internal standard for an analysis. The dependence of signal lifetime on sample concentration and signal intensity on sample heat of vaporization causes

PAGE 165

1-1

PAGE 166

Figure 6,15 Plot of response ratio (analyte/internal standard) vs. amount of analyte on grid for DCI/CAD of 1 juL of graded melting point calibrant solutions (Table VIII) on nickel grids. Parent ion experiment with selected reaction monitoring (Table VI). Symbols: # , MPS; , PB; A , PHT; ^ , MES; # , MPS+PB+PHT (3 points overlap) .

PAGE 167

154 RESPONSE RATIO 6.0 4.0 2.0 0.1 x 0.2 0.3 Ug ON NICKEL GRID

PAGE 168

Figure 6.16 Plot of response ratio (analyte/internal standard) vs. amount of analyte on grid for DCI/CAD of 1 ^iL of graded melting point calibrant solutions (Table VIII) on Teflon paddle. Parent ion experiment with selected reaction monitoring (Table VI). Symbols: # , MPS; , PB; A , PHT: ^ , MES; ^ , MPS+PB+PHT (3 points overlap).

PAGE 169

156 RESPONSE RATIO 6.0 4.0 2.0 0.1 0.2 0.3 yg ON TEFLON PADDLE 0.4

PAGE 170

157 the DCI technique to be less tolerant of differences between the analyte and internal standard than direct solids probe evaporation (glass probe vial) with rapid heating. LIDCI calibration for PHT on nickel grid . The difficulty experienced in obtaining an adequate calibration for a multicomponent sample in DCI prompted the decision to try calibration for a single analyte by LIDCI. Table IX gives the contents of each calibrant for PHT with PHT* used as the internal standard. A 1 >iL sample was spotted on each grid and 2-3 shots were made at different sites on a grid. Data were acquired in the parent ion mode with Q3 sequentially selecting m/z 182 (PHT) and m/z 183 (PHT*) and Ql scanning from m/z 250 to 260 in 9 ms with a 2 ms delay between scans to detect the respective parent ions, m/z 253 and m/z 256. Figure 6.17 shows a plot of ratio of selected reaction response for PHT to that for PHT* vs. amount of PHT on the grid (line 1). Again, the coefficient of determination was low (r2=0.910) indicating poor correlation between response ratios and weight of drug. Elimination of the high results and the high and low results for one of the calibrants which gave 3 widely spaced ratios (marked by *) left 10 results spanning the range from 0.098 ;jg to 1.37 ng PHT on a grid. The best straight line calculated for this group of points is line 2 in Figure 6.17. Table X gives the coefficients of determination for the three sets of calibrations (DCI-nickel, DCI-Teflon, LIDCI) and also the coefficients obtained after eliminating obviously deviant data points, which improved correlation significantly in almost all cases. The only data sets for which data did not have to be eliminated (no points

PAGE 171

o o o o o o o 158 e c ^ — 1
PAGE 172

Figure 6.17 Plot of response ratio (PHT/PHT*) vs. amount of PHT on nickel grid for LIDCI/CAD of PHT calibrant solutions (Table IX). Parent ion experiment with selected reaction monitoring of m/z 253 going to m/z 182 (PHT) and m/z 256 going to m/z 183 (PHT*). All points used for least squares calculation (1); Indicated points (*) excluded for least squares calculation (2). Symbols: ^ , single datum point; ifi , two data points.

PAGE 173

160 RESPONSE RATIO 2.0 , 1.0 1.0 yg PHT ON NICKEL GRID

PAGE 174

161 Table X. Coefficients of Determination (r^) for DCI Calibration^ with Nickel Grids and Teflon Paddle, and for LIDCI Calibration^ with Nickel Grids (See Figures 6.15, 6.16, 6.17). Data MPS^ MESC Pb^ PHT' DCI-Ni-All 0.524 DCI-Ni-High Calibrant Omitted 0.909 DCI -Teflon-All 0.868 DCI-Tef Ion-high calibrant omitted 0.945 LIDCI-Ni-all _ _ _ LIDCI-Ni-high calibrant and widely variant results (*) omitted

PAGE 175

162 obviously out of line) were those for PHT DCI. Of these, the samples evaporated from Teflon gave the best results, indicating that the process which occurred in that case was more controlled, probably because of decreased or non-existent substrate effect. It is also probable that methylphenytoin is a more suitable internal standard for phenytoin than for the other analytes. The end result of these attempts at calibration is that the process, with this instrumentation, is too uncontrolled at this point to be used for quantitative analysis; however, it can be used for qualitative analysis. Qualitative Analysis of Rat Brain by LIDCI This experiment was performed to attempt to answer the question posed at the end of Chapter I. Again, that is: Although a molecule may be present in tissue, is it in a form that is recoverable by direct vaporization or is it incorporated into membranes, proteins and anatomical structures in such a way as to make it unrecognizable? The procedure used to prepare samples for this part of the study was described at the end of Chapter III. Table XI gives the sample descriptions and phenytoin concentrations found in the frozen and fixed brain and liver samples taken from the rats that had been loaded with phenytoin (150 mg/kg). The total amounts of phenytoin found in the fixative solutions are also given. The two frozen liver samples (1 and 2) were homogenized by sonication. Since liver is relatively difficult to homogenize, it is possible that the concentration difference in these two samples is due to incomplete homogenization. There are also

PAGE 176

163 Table XI . Phenytoin Concentrations in Rat Brain and Liver Samples and in Fixative Solutions Sample Brain (1) Brain (2) Brain (3) Liver (1) Liver (2) Liver (3) fixative-Brain fixative-Liver Description frozen frozen Fixed 72 hr frozen frozen fixed 72 hr (0.50 mL analyzed out of 10 mL) Concentration ^7.4 ;ig/g 46.6 jug/g 25.7 ;ig/g 211 >ig/g 130 ;ig/g 82.4 ;ag/g 2.66 jag/10 mL 8.32 jug/lO mL

PAGE 177

164 discrepancies between the concentrations in frozen and fixed brain and in frozen and fixed liver. The amount of phenytoin found in the fixative is not enough to account for the differences. The fixed samples, however, were homogenized in a ground glass grinder with a ground glass pestle, which is very efficient. Therefore, the reason for the differences between the frozen and fixed samples cannot be explained by the present data. Since the phenytoin deficits in the fixed tissues are unaccounted for, it is possible that the fixed tissues used for LIDCI MS may have concentrations as high as those in the frozen samples, or as low as those in BRAIN 3 and LIVER 3. Therefore, the brain sample used for LIDCI probably had a phenytoin concentration between 26 and 47 ;ig/g wet weight; the liver sample may have contained from 82 to 211 ;ug/g wet weight. Sections of fixed brain and liver were trimmed to be about 0.5 mm thick and 2 mm square with a small tab on one corner. The TSQ was set up to acquire in daughter ion mode with Ql passing m/z 253 and Q3 scanning from m/z 170 to 260 in 19 ms with a 1 ms delay between scans. Centroid sampling interval was set at 75 jus. Source pressure (methane) was 0.90 torr and the source temperature was 160°C. Collision energy was 18 eV and collision gas pressure (nitrogen) was 2.6 ratorr. Fresh rhodamine 6G dye was put into the laser and the driver voltage was set at 18 kV. Cavity diameter was 10 mm. Figure 6.18 shows the ion current profiles for the major PHT ions (parent and daughter) which resulted from a single laser shot fired at the liver section. The hole produced

PAGE 178

u

PAGE 179

166 > t t > > > > >

PAGE 180

167 in the tissue sample was about 0.7 mm in diameter (2 x 10"^ cm-^ vaporized.).. The large spike at m/z 213 is from the laser .firing. Several other shots at different sites on the tissue produced ions representative of PHT, but this one gave the best profile. PHT was not detected in the brain sample. Figure 6.19 shows spectra from the liver sample (12 scans after the laser shot summed), and from a sample of pure PHT placed on a grid and vaporized with the laser immediately after the tissue samples were run. These results, although inconclusive because they were not repeated, are promising in that PHT, introduced into a sample in vivo, was detected by direct vaporization of the tissue in the ion source. It is also possible that there is "free" PHT which is accessible and "bound" PHT which is not. There is some hope that, if a microprobe tandem mass spectrometer can be developed by overcoming the time limitations of the present system, it will be possible to detect molecular species directly in intact tissue samples.

PAGE 181

Figure 6.19 Daughter spectra of m/z 253 from summation of 12 scans after laser shot for (A) rat liver tissue section from Figure 6.18, and sample of pure PHT on nickel grid. Pure sample analyzed after liver sample.

PAGE 182

ION CURRENT (CNTS) 38848 A 182 175 210 I I ] I, .— aJ 225 169 253 1087480 B 182 175 210 225 253 m/z 1! 200 220 240 260

PAGE 183

CHAPTER VII CONCLUSIONS AND FUTURE WORK The original purpose of this work was to introduce the concept of using tandem mass spectrometry as a means of obtaining the selectivity required for molecular analysis under probe-type sampling conditions. To achieve this end, a single-shot laser was interfaced to a triple quadrupole mass spectrometer, and the system was evaluated with respect to short-lived and long-lived processes which were induced by the incidence of the laser radiation on molecular samples. This chapter is a summary of the conclusions based upon this work, and a prospective view of future research which may contribute to the development of a microprobe capable of providing a molecular image of intact, heterogeneous and compartmentalized samples. Conclusions Based on the Present Work Direct Analysis of Mixtures by MS/MS Initially, tandem mass spectrometry, implemented with a triple quadrupole analyzer, was shown to be a valid technique for quantitation of drugs in small sections of rat brain. The internal standard technique was used and crude extracts were vaporized into the ion source from a rapidly heated solids probe while the molecular ions of interest 170

PAGE 184

171 were selected with Ql, fragmented in Q2 , and the corresponding daughter ions were observed after selection by Q3 (selected reaction monitoring). The same samples, analyzed with the mass spectrometer operating in the normal mass spectrum mode, gave little or no evidence of the presence of the analytes or internal standards. In addition, the chemical background was intense with significant signal at all masses in the range of interest. This condition was similar to that obtained with a currently available laser mass spectrometer microprobe (LAMMA 500), which proved unsuitable for detection of cobalt in sections of a cobaltinduced lesion in rat brain. These studies pointed out the deficiency of the single stage mass analyzer in direct analysis of mixtures, both in the bulk analysis and in the microprobe modes, and showed the advantage that could be gained by introduction of a second stage of mass analysis. Laser Studies Short-lived processes . Although the triple quadrupole mass spectrometer is not specifically designed for monitoring brief signals, a system having limited mass resolution for ions generated by the laser was developed in which the ion of interest was selected by Ql, and Q3 was configured to allow a range of masses to pass simultaneously. In this way, a time-dependent signal of the selected ion intensity was produced. Time resolution was limited to 25 >is by the ADC rate. All signals monitored in this mode had lifetimes of less than 1 ms. Mass filtering was not totally efficient (i.e., less than unit mass resolution), and some signal was observed if a mass close to that of an ion present in the source were selected by Ql. If Ql and Q3 were configured to pass

PAGE 185

172 ions widely separated in mass, a minimal signal, which was always present when the data system was used, was obtained. This signal was due to electrical interference which resulted from discharge of the storage capacitor used to fire the laser flashlamp. When the storage oscilloscope, triggered by the output of a fast photodiode which received a portion of the laser light, was used to monitor the signal directly from the electron multiplier under the same conditions, a baseline signal was obtained. This indicated that the storage scope was less sensitive to electrical interference than the data system, and that the interference was probably due to an antenna effect of the mass spectrometer cables. When both quadrupoles were operated in RF-only mode (passing all ions having masses greater than 7/9 of the selected mass), a two partially resolved peaks were obtained when a sample containing two compounds differing in mass by about 500 u was vaporized by the laser. From this time-of-f light experiment, an instrumental constant of 9.77 jds u"-*-'^ was calculated. The intercept of the time vs. square root of mass curve for these two data points indicated a delay of 127 jas between the time the ions were formed, ideally at the time the laser was fired, and the first detection of a signal above threshold. This set of experiments gave some information about how the quadrupoles functioned as filters for laser generated ions. The available mass and time resolution of the instrument was insufficient to allow positive identification of ions detected, or to allow operation of the instrument as a true tandem mass spectrometer. Experiments with reagent gas

PAGE 186

173 present in the ion source indicated that long-lived signals could be obtained, which would allow more effective use of the mass selection capabilities of the TSQ. Long-lived processes . When reagent gas was present in a closed chemical ionization source, ion signals which lasted for hundreds of milliseconds to several seconds could be obtained for a number of compounds (Appendix). Initially, it appeared that these signals were initiated by the laser and that their longevity was due to conditions in the source which prolonged the lifetime of the ions vaporized by the laser. Subsequent experiments showed that the signal was actually a convolution of two signals; one the result of ions produced by chemical ionization of the laser produced molecules, and the second longer lasting signal the result of desorption chemical ionization (DCI) of sample vaporized by the action of reagent gas molecules on the sample surface. Use of the laser as a means of introducing thermal energy into the sample initiated these processes when conditions unfavorable to DCI existed (when the sample was blocked from direct contact with reagent gas entering the source) to give a phenomenon called Laser-Induced Desorption Chemical Ionization (LIDCI). If conditions were favorable for DCI, the laser was used to enhance the signal in a process called Laser-Enhanced Desorption Chemical Ionization (LEDCI). Compounds that could be vaporized from the solids probe by normal resistive heating could be vaporized without the addition of probe heat, but with the heat provided only by the ion source. In one case, a compound that could not be vaporized from a glass probe vial gave a signal under DCI conditions. The DCI signal was influenced by sample substrate, and exhibited

PAGE 187

174 threshold pressure and temperature requirements. Threshold levels of ion source pressure and temperature were also required for production of laser-initiated signals; and signal lifetime was dependent on compound melting point and sample size. Analytical application of laser-initiated processes . The dependence of the laser-initiated processes on compound melting point and sample concentration limits the analytical potential of these techniques. Although limits of detection for DCI of a higher molecular weight alkaloid were in the very low nanogram range, attempts at calibration with a mixture of commonly used antiepileptic drugs were relatively unsuccessful, and gave low coefficients of determination (r^). For 3 out of 4 of the compounds studied, suppression of the signal at the highest concentration was observed. The compound that did not show this effect was the most similar to the internal standard. When identical samples containing a mixture of phenytoin (PHT) and a stable isotope-labeled analog (PHT ) were subjected to LIDCI with the instrument operated in the normal mass spectrum mode, ion response ratios for the protonated molecular ions were reproducible with a relative standard deviation of 5%. When samples having equal weights of PHT* (internal standard) and weights of PHT varying from 0.098 jug to 1.95 jjg were subjected to LIDCI (selected reaction monitoring), most analyte/internal standard response ratios showed a linear increase with concentration. Elimination of 4 outlying points out of 14 points gave a reasonable coefficient of determination of 0.988. These results indicate that, with proper internal standardization (126, 127), it may be possible to use this technique for quantitation. Practical lower

PAGE 188

175 limits of quantitation are high, however, compared to results obtained by more conventional methods (128). In addition, versatility is limited because of the dependence of signal lifetime on analyte melting point (heat of vaporization). This will introduce an undesired selectivity because mixtures containing components having a range of vaporization temperatures will exhibit signals favoring the components having the lower heats of vaporization. Detection of phenytoin in rat liver . The most promising result obtained in this study was the detection of phenytoin, which was introduced in vivo, in a section of rat liver by direct laser vaporization of the tissue inside the CI ion source. The protonated molecular ion was selected in Ql, fragmented in Q2 , and daughter ions were detected for about 400 ms after the laser shot by scanning Q3 over a 90 unit mass range. The spectrum obtained in this way from the liver sample matched that of a sample of pure phenytoin vaporized under the same conditions. This result indicates that molecular species can be recognized by direct vaporization from tissue, and that their identity may not be obscured by binding to macromolecules and tissue membranes. This is not so important when elemental analysis is performed by a microprobe method because the elements will be recognizable by their isotopic patterns even if all surrounding tissue is destroyed. A primary concern, however, at the start of this project, was that the lack of a working microprobe for molecular analysis many years after the development of ion microprobes and microscopes was due to the difficulties of in situ detection of molecules after they had been incorporated into biological systems. This result gives hope that

PAGE 189

176 molecular analysis by direct vaporization of tissue is possible, and gives, in this preliminary stage of investigation, a positive answer to the question posed at the end of Chapter I. Previous work with laser desorption and laser ablation MS, which has been cited earlier in this report, indicated that auxiliary ionization techniques coupled with laser sampling could give extended lifetimes to ion signals. The current work has extended the lifetimes of laser-initiated signals from a few milliseconds to several seconds for a single 1 to 2 ;js laser pulse. This prolonging of signal is important because of the statistical nature of the detection process. To produce an analytically useful response, several conditions must be satisfied. These include: (1) the sampling and response times of the detection system must be significantly shorter than the potential lifetime of the signal; (2) the potential lifetime of the signal must not be dependent upon analyte concentration; (3) the mechanism of signal generation must not discriminate between analyte response and internal standard response; and (4) analyte and internal standard responses must be distinguishable within the constraints of signal lifetime and detector response time. In addition, several conditions should be inherent in the design of a versatile molecular microprobe, including: (1) it should be possible to maintain the sample at or below room temperature to maintain the integrity of fragile samples such as biological tissues; (2) the system should be capable of generating ions from molecules that are normally amenable to MS and also

PAGE 190

177 those that are thermally unstable or nonvolatile; and (3) the system should be able to take advantage of processes such as cationization and anionization. The present system, while showing that combination of a spatially controllable desorption ionization technique (laser) with tandem mass spectrometry may be a valid approach to building a molecular microscope, lacks the versatility that would be demanded of such an instrument. The triple quadrupole mass spectrometer is not fast enough to adequately characterize molecules ionized by cationization or anionization after desorption by a single shot laser. The laser-induced desorption CI process, demonstrated on this instrument, does not satisfy the need for a vaporization/ionization technique applicable to a wide variety of molecules (polar, nonvolatile, thermally unstable). Lastly, the discrimination among molecules according to heat of vaporization, as evidenced by widely different sensitivities for compounds having different melting points, makes quantitation difficult, and places more stringent requirements on internal standard selection. A fast-pulsed laser system could have contributed greatly to the characterization of the present instrument system and the processes observed by allowing continuous monitoring of ion signals while instrumental and sample conditions were varied.

PAGE 191

178 Future Work Proposed Instrument Designs The primary considerations at this time are the choices of sampling device and mass analyzer. Although this report has dealt with the laser as a sampling device, an ion beam should also be considered. The mass analyzer or combination of analyzers must provide mass selection, secondary fragmentation, and mass analysis of the resulting fragments, within the lifetimes of the primary molecular ions desorbed or vaporized from the sample. This system must have a high transmission, and the ion source and fragmentation stage must be very efficient to optimize detection limits. Sampling devices . Both laser (34) and ion beams (18) have been used with success for sampling on a microscopic scale for elemental analysis. As has been mentioned previously, speculative comments have been made during the past 5 years (47-51) concerning an instrument for molecular microscopy. More recently, mass spectrometric systems have been described which are capable of molecular imaging. Stoll et al. (129) have recently described an instrument, based on SIMS (secondary ion mass spectrometry), with a high resolution, double-focusing mass spectrometer. The primary cesium ion beam was focused to a 0.3 >im diameter spot with a current density of 1 A cm~2 and a power density of 10^ W cm~2 . Molecular maps of a low mass (m/z 131) and a high mass (m/z 1281) ion were produced from a perf luoropolyether , deposited on a 500 mesh EM grid. This instrument provided more selectivity for molecular analysis than one having nominal mass resolution; however, an instrument

PAGE 192

179 having a second, confirmatory stage of fragmentation and mass analysis is more desirable. Kelner et al. (130) have recently introduced a system which they call a "molecular microprobe". They considered all of the requirements listed above, and designed an instrument, based on SIMS, which incorporates a high transmission, triple quadrupole analyzer. The system had not been tested at the time of the report. Sampling with a primary ion beam has several advantages, including: (1) fineness of focus (20 nm for a cesium ion gun); (2) ease of beam translation by electronic rastering under computer control; and (3) accurate control over and measurement of beam current and primary ion kinetic energy. Sensitivity and signal lifetime, however, are highly dependent upon the sample matrix (50). Since a microprobe is primarily designed to detect species in their natural state, this is a definite disadvantage. In fact, much of the success of molecular SIMS is due to pretreatment of the sample, or dispersing it in a nonvolatile liquid matrix, to form polar or pre-ionized species, and to provide a continuously renewable surface (131, 132). The liquid matrix also allows use of a high current density primary ion beam, which increases sensitivity. The wide-ranging molecular reordering which occurs through collision cascades resulting from primary particle impact, limits the beam current that can be used on static surfaces (133). Another disadvantage of basing a microprobe system on SIMS is the small sampling depth (molecular monolayer). This factor limits sensitivity, increases analysis time, and limits the versatility of the technique, particularly with biological samples. In this case, an instrument with an instantaneous sampling depth of at least 0.5 )im, and capability for

PAGE 193

180 penetration to 0.5 mm would be more desirable, since cell dimensions uspally span tens of micrometers. The primary disadvantages of the laser as a sampling device are its expense, lack of reproducibility of beam characteristics (energy/pulse), and poor functional reliability. These drawbacks, however, are diminishing with advances in laser technology (134). The advantage of the laser lies in its capability of sampling at depths which are more reasonable for biological systems. In addition, the sampling depth can be determined by the power density of the laser emission at the sample surface. The power density also determines the extent of fragmentation of molecular ions or ion adducts (32, 135). Therefore, this parameter can be adjusted for optimum sensitivity (lower power density produces only molecular ions) or for maximum structural information (higher power density produces fragmentation). Actually, since the cost of the sampling device would be a relatively small fraction of the total cost of an instrument having the proposed capabilities, and since laser and ion beam sampling are complementary, it would be most reasonable to incorporate both sampling systems. Mass analyzers . At this time, multi-stage mass spectrometry has seemingly been performed with every available type and combination of mass analyzer. From the original magnetic-electric sector (B-E) reversed geometry MS/MS instrument (136), progress has seen the introduction of triple quadrupole (QQQ) (54), multiple sector (137, 138), and hybrid multiple sector-multiple quadrupole instruments (139). In addition, MS/MS experiments have been performed on Fourier transform (FT) mass spectrometers (56, 140), and by monitoring ion flight times

PAGE 194

181 through a magnetic sector after CAD (141, 142), Recently, Glish and Goeringer (143) have described a tandem quadrupole time-of-f light (Q-TOF) instrument, which could be operated as a Q-TOF, or as a TOF-TOF MS/MS. It may be a design similar to this that provides the key to linking desorption ionization with tandem mass spectrometry to produce a workable molecular microprobe. The dependence of signal lifetime upon sample size and heat of vaporization seriously limits the applicability of the long-lived LIDCI and LEDCI processes to the development of a molecular microprobe. Therefore, the performance of the triple quadrupole mass spectrometer in this application is also limited, because the required scan time is still too long to take advantage of true laser desorption processes. These processes produce intense, brief pulses of ions and neutral molecules having energy spreads from tens of electron volts, initially, to a fraction of an electron volt after several microseconds (83). Decay of the energy spread is beneficial because increased resolution can be obtained in a TOF mass spectrometer by delaying the ion acceleration pulse (77). Sensitivity, however, is decreased because the particle abundance is also decreasing rapidly. Therefore, to take full analytical advantage of the ions and neutral molecules generated by the laser pulse, the first stage of the instrument should be an ion source and storage device, in which laser-desorbed ions and ions created from neutral molecules could be contained. This storage device could be an ion cyclotron resonance (ICR) cell, such as is used in FTMS. Laser desorption has been used successfully with FTMS (104,144). The only drawback to such a system is the need for a high field magnet, which

PAGE 195

182 surrounds the ICR cell. This would complicate and perhaps preclude the design of the specimen viewing system for the microprobe. One possible solution to this problem might be to use a quadrupole to inject the ions into the ICR cell (145); however, this does not accomplish the initial goal of fully using both ions and neutrals desorbed by a laser. A 3diraensional quadrupole ion trap (146, 147) could provide an alternative solution, although the mass range of current designs is limited to about 650 u. The sample could be located within or outside of the trap, which would be operated in the total storage mode. Neutrals could be ionized by electron impact. After a variable delay, a high voltage ion drawout pulse could sample the contents of the source-trap, accelerating the ions through a short field-free region into a deceleration region, and, finally, into a Q-TOF system. After the TOF analysis of daughter ions from the selected ion, the source could be pulsed again, and the same or a different parent ion could be selected. The system could be operated as a single stage TOF, as a Q-TOF, or as a TOF-TOF mass spectrometer, as described by Glish and Goeringer (143). It is clear from the literature that a number of investigators are now convinced that it is possible to construct an instrument having the capabilities described in this work. There is even a hint that the instrument will combine desorption ionization and tandem mass spectrometry (148). At this point, it may only require the correct combination of technologies that currently exist.

PAGE 196

183 Q

PAGE 197

184 4-1 4J D UJ U C -P 0) W -H C M 3 4-1 -H HOC T3 ^ O 0) C S -H (TJ O + rH ^1 M tNI O O O tN ^ Q J Q Z 2 Z CO ^ 4-1 u

PAGE 198

LITERATURE CITED 1. Rochow, T.G.; Rochow, E.G. "An Introduction to Microscopy by Means of Light, Electrons, X-Rays or Ultrasound"; Plenum Press: New York, 1978. 2. Siegel, B.M.; Beaman, D.R. "Physical Aspects of Electron Microscopy and Microbeam Analysis"; John Wiley and Sons, Inc.: New York, 1975. 3. Kelley, J.S.; Simmonds, M.A. ; Straughan, D.W. In "Methods in Brain Research"; Bradley, P.B., Ed.; John Wiley and Sons, Inc.: London, 1975. Chapter 7. A. Berman, H.J.; Hebert, N.C., Eds.; "Ion Selective Microelectrodes" ; Adv. Exp. Med. Biol. 50; Plenum Press: New York, 197A. 5. Adams, R.N. Anal. Chem (1976) 48, 1126A-1137A. 6. Costa, E.; Iversen, L.L.; Paoletti, R. , Eds.; "Studies of Neurotransmitters at the Synaptic Level"; Adv. Biochem. Psychopharm. 6; Raven Press: New York, 1972. 7. Nairn, R.C., Ed.; "Fluorescent Protein Tracing", 4th ed. ; Churchill Livingstone: Edinburgh, 1976. 8. Lojda, Z.; Gossran, R. ; Schiebler, T.H. "Enzyme Histochemistry A Laboratory Manual"; Springer-Verlag : Berlin, 1979. 9. Bullock, G.R.; Petrusz, P., Eds.; "Techniques in Immunohistochemistry" ; Academic Press: London, 1982 (Vol. 1), 1983 (Vol. 2). 10. Meek, G.A.; Elder, H.Y. , Eds.; "Analytical and Quantitative Methods in Microscopy"; Cambridge University Press: Cambridge, 1977. 11. Williams, M.A. "Autoradiography and Immunocytochemistry" ; NorthHolland Publishing Co.: Amsterdam, 1977. 12. Cuello, A.C., Ed.; "Brain Microdissection Techniques"; John Wiley and Sons, Inc.: New York, 1983. 13. Otsuka, M. ; Miyata, Y. In "Studies of Neurotransmitters at the Synaptic Level"; Costa, E.; Iversen, L.L.; Paoletti, R. , Eds.; Adv. Biochem Psychopharm. 6; Raven Press: New York, 1972. Pp. 61-74. 185

PAGE 199

186 14. Lowry, O.H.; Passoneau, J.V. "A Flexible System of Enzymatic Analysis"; Academic Press: New York, 1972. 15. Birks, L.S. "Electron Probe Micronanalysis" ; Interscience Publishers: New York, 1963. 16. Evans, C.A. Anal. Chem. (1975) 47, 818A-829A. 17. Evans, C.A. Anal. Chem. (1975) 47, 855A-866A. 18. Morrison, G.H.; Slodzian, G. Anal. Chem. (1975) 47, 932A-943A. 19. O'Shea, D.C.; Callen, W.R. ; Rhodes, W.T. "Introduction to Lasers and Their Applications"; Addison Wesley Publishing Co.: Menlo Park, 1978. 20. Brech, F.; Cross, L. Appl. Spectrosc. (1962) 16^, 59. 21. Honig, R.E.; Woolston, J.R. Appl. Phys. Lett. (1963) 2, 138-139. 22. Harding-Barlow, I.; Snetsinger, K.G.; Keil, K. In "Microprobe Analysis"; Anderson, C.A., Ed.; John Wiley and Sons, Inc.: New York, 1973. Pp. 423-455. 23. Harding-Barlow, I. Laser Appl. Med. Biol. (1974) 2, 133-149. 24. Laqua, K. In "Analytical Laser Spectroscopy"; Omenetto, N., Ed.; John Wiley and Sons, Inc.: New York, 1979. Pp. 47-118. 25. Dhamelincourt, P.; Wallert, F. ; Leclerg, M. ; N'Guyen, A.T.; Landon, D.O. Anal. Chem. (1979) M, 414A-421A. 26. Delhaye, M. ; Bridoux, M. ; Dhamelincourt, P.; Barbillat, J.; DaSilva, E. ; Roussel, B. Microbeam Anal. (1982) 17, 275-278. 27. Kwong, H.S.; Measures, R.M. Anal. Chem. (1979) M, 428-432. 28. Conzemius, R.J.; Capellan, J.M. Int. J. Mass Spectrom. Ion Phys. (1980) 34, 197-271. 29. Conzemius, R.J.; Simons, D.S.; Shankai, Z.; Byrd, CD. In "Microbeam Analysis — 1983"; Gooley, R. , Ed.; San Francisco Press: San Francisco, 1983. Pp. 301-328. 30. Hillenkamp, F.; Unsold, E. ; Kaufmann, R. ; Nitsche, R. Nature (1975) 256, 119-120. 31. Vogt, H.; Heinen, H.J.; Meier, S.; Wechsung, R. Fresenius' Z. Anal. Chem. (1981) 308, 195-200.

PAGE 200

187 32, Denoyer, R.J.; VanGrieken, R. ; Adams, F.; Natusch, D.F.S. Anal. Chem (1982) 5A, 26A-41A. 33. Hercules, D.M. ; Day, R.J.; Balasanmugam, K. ; Dang, T.A.; Li, C.P. Anal. Chem (1982) 54, 280A-305A. 34. Sessions of LAMMA Symposium, Dusseldorf, West Germany, October 810, 1980. Fresenius' Z. Anal. Chem. (1981) 308, 193-320. 35. Schiller, C; Kupka, K.D.; Hillenkamp, F. Fresenius' Z. Anal. Chem. (1981) 308, 304-308. 36. Muller, J.F.; Berthe, C; Magar, J.M. Fresenius' Z. Anal. Chem. (1981) 308, 312-320. 37. Kupka, K.D.; Schropp, W.W. ; Schiller, C; Hillenkamp, F. Fresenius' Z. Anal. Chem (1981) 308, 229-233. 38. Hirche, H. ; Heinrichs, J.; Schaefer, H.E.; Schramm, M. Fresenius' Z. Anal. Chem. (1981) 308, 224-228. 39. Mathey, A. Fresenius' Z. Anal. Chem. (1981) 308, 249-252. 40. Schroder, W.H. Fresenius' Z. Anal. Chem. (1981) 308, 202-217. 41. Seydel, U.; Lindner, B. Fresenius' Z. Anal. Chem. (1981) 308, 253257. 42. Heinen, H.J.; Meier, S.; Vogt, H.; Wechsung, R, Int. J. Mass Spectrom. Ion Phys. (1983) 47, 19-22. 43. Heinen, H.J.; Meier, S.; Vogt, H. Spr. Ser. Chem. Phys. (1983) 25, 229-234. 44. Feigle, P.; Schueler, B.; Hillenkamp, F. Int. J. Mass Spectrom. Ion Phys. (1983) 47, 15-18. 45. Dhamelincourt , P. In "Microbeam Analysis"; Newbury, D.E., Ed.; San Francisco Press: San Francisco, 1979. Pp. 155-164. 46. Seydel, U.; Lindner, B. Spr. Ser. Chem. Phys. (1983) 25 240-244. 47. Day, R.J.; Unger, S.E.; Cooks, R.G. Anal. Chem. (1980) 52, 557A572A. 48, Unger, S.E.; Vincze, A.; Cooks, R.G.; Chrisman, R. ; Rothman, L.D. Anal. Chem. (1981) 53, 976-981. 49. Cooks, R.G.; Busch, K.L. J. Chem. Educ . (1982) 59, 926-933.

PAGE 201

188 50. Busch, K.L.; Hsu, B.H.; Xie, Y. ; Cooks, R.G. Anal. Chem (1983) 55, 1157-1160. 51. Cooks, R.G.; Busch, K.L.; Glish, G.L. Science (1983) 222^, 273-291. 52. Kondrat, R.W. ; Cooks, R.G. Anal. Chem. (1978) 50, 81A-92A. 53. McLafferty, F.W.; Bockoff, F.M. Anal. Chem. (1978) 50, 69-76. 54. Yost, R.A.; Enke, C.G. Anal. Chem. (1979) 5J^, 1251A-1264A. 55. McLafferty, F.W., Ed.; "Tandem Mass Spectrometry"; John Wiley and Sons, Inc.: New York, 1983. 56. White, R.L.; Wilkens, C.L. Anal. Chem. (1982) 54, 2211-2215. 57. Perchalski, R.J.; Yost, R.A. ; Wilder, B.J. Anal, Chem. (1982) 54, 1466-1471. 58. Zackett, D.; Schoen, A.E.; Cooks, R.G.; Hemberger, P.H. J. Am. Chem. Soc. (1981) 103, 1295-1297. 59. Davis, D.V. ; Cooks, R.G.; Meyer, B.N.; McLaughlin, J.L. Anal. Chem. (1983) 55, 1302-1305. 60. Cotter, R.J. Anal. Chem. (1980) 52, 1767-1770. 61. Ready, J.F. "Effects of High Power Laser Radiation"; Academic Press: New York, 1971. Pp. 157-161. 62. Ready, J.F. "Effects of High Power Laser Radiation"; Academic Press: New York, 1971. P. 82. ' 63. Ready, J.F. "Effects of High Power Laser Radiation"; Academic Press: New York, 1971. Pp. 67-125. 64. van der Peyl, G.J.Q.; Haverkamp, J.Q.; Kistemaker, P.G. Int. J. Mass Spectrom. Ion Phys. (1982) 42, 125-141. 65. Lincoln, K.A. ; Covington, M.A. Int. J. Mass Spectrom. Ion Phys. (1975) 16, 191-208. 66. Lincoln, K.A. Int. J. Mass Spectrom. Ion Phys. (1974) 13, 45-53. 67. Knox, B.E. Dyn. Mass Spectrom. (1971) 2, 61-96. 68. Ready, J.F. "Effects of High Power Laser Radiation"; Academic Press: New York, 1971. Pp. 95-114.

PAGE 202

189 69. Kovalev, I.D.; Maksimov, G.A.; Suchkov, A.I.; Larin, N.V. Int. J. Mass Spectrom. Ion Phys. (1978) 27, 101-137. 70. Furstenau, N. Fresenius' Z. Anal. Chem. (1981) 308, 201-205. 71. Scott, R.H.; Strasheim, A. Spectrochim. Acta (1970) 25b, 311-332. 72. Ready, J.F. "Effects of High Power Laser Radiation"; Academic Press: New York, 1971. P. 154. 73. Ready, J.F. "Effects of High Power Laser Radiation"; Academic Press: New York, 1971. Pp. 140-161. 74. Inghram, M.G.; Chupka, W.A. Rev. Sci. Instrum. (1953) 24, 518-520. 75. Datz, S.; Taylor, E.H. J. Phys. Chem. (1956) 25, 389-394. 76. Cotter, R.J. Anal. Chem. (1984) 56, 485A-504A. 77. Cotter, R.J.; Tabet, J.-C. Am. Biotechnol. Lab. (1984) 2, 10-20. 78. Ready, J.F. "Effects of High Power Laser Radiation"; Academic Press: New York, 1971, P. 147. 79. Bernal, E.; Ready, J.F.; Allen, F.J. J. Appl. Phys. (1974) 45, 29802986. 80. Hardin, E.D.; Vestal, M.L. Anal. Chem. (1981) 53, 1492-1497. 81. van der Peyl, G.J.Q.; van der Zande, W.J.; Bederski, K. ; Boerboom, A.J.H.; Kistemaker, P.G. Int. J. Mass Spectrom. Ion Phys. (1983) 47, 7-10. 82. Van Breeman, R.B.; Snow, M. ; Cotter, R.J. Int. J. Mass Spectrom. Ion Phys. (1983) 49, 35-50. 83. Tabet, J.-C; Cotter, R.J. Int. J. Mass Spectrom. Ion Phys. (1983) 54, 151-158. 84. Cotter, R.J.; Tabet, J.-C. Int. J. Mass Spectrom. Ion Phys. (1983) 53, 151-166. 85. Cotter, R.J.; Snow, M. ; Colvin, M, Spr. Ser. Chem. Phys. (1983) 25, 206-210. 86. Vastola, F.J.; Pirone, A.J. Adv. Mass Spectrom. (1968) 4, 107-111. 87. Vastola, F.J.; Mumma, R.D,; Pirone, A.J. Org. Mass Spectrom. (1970) 3, 101-104,

PAGE 203

190 89, 90. 91 MacFarlane, R.D.; Torgerson, D.F. Science (1976) j^, 920-925. Posthumus, M.A.; Kistemaker, P.G.; Meuzelaar, H.L.C.; Ten Noever de Brauw, M.C. Anal. Chem (1978) 50, 985-991. Hunt, D.F.; Shabanowitz, J.; Botz, F.K.; Brent, D.A. Anal. Chem. (1977) 49, 1160-1163. Cotter, R.J.; Fenselau, C. Biomed. Mass Spectrom. (1979) 6, 287293. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102, 103, Stoll, R.; Rollgen, F.W. Org. Mass Spectrom. (1979) 14, 642-645. van der Peyl, G.J.Q.; Isa, K. ; Haverkamp, J,; Kistemaker, P.G. Org. Mass Spectrom. (1981) J^, 416-420. van der Peyl, G.J.Q.; Isa, K. ; Haverkamp, J.; Kistemaker, P.G. Nucl. Inst. Meth. Phys . Res. (1982) j^, 125-130. van der Peyl, G.J.Q.; Isa, K. ; Haverkamp, J.; Kistemaker. P.G. Int. J. Mass Spectrom. Ion Phys. (1983) 47, 11-14. Stoll, R, ; Rollgen, F.W. Org. Mass Spectrom. (1981) 16, 72-75. Stoll, R. ; Rollgen, F.W. Z. Naturforsch. (1982) 37A, 9-14. Heinen, H.J. Int. J. Mass Spectrom. Ion Phys. (1981) 38, 309-322. Hillenkamp, F. Int. J. Mass Spectrom. Ion Phys. (1982) 4^, 305-313. Hillenkamp, F. Spr. Ser. Chem. Phys. (1983) 2^, 190-205. Cooks, R.G.; Busch, K.L. Int. J. Mass Spectrom. Ion Phys. (1983) 53, 111-124. Cotter, R.J. Anal. Chem. (1981) 53, 719-720. Heresch, F. ; Schmid, E.R.; Huber, J.F.K. Anal. Chem. (1980) 52, 1803-1807. 104. McCrery, D.A.; Ledford, E.B.; Gross, M.L. Anal. Chem. (1982) 54, 1435-1437. 105. Dow, R.S.; Fernandez-Guardiola, A.; Manni , E. Electroenceph. Clin. Neurophysiol. (1962) ]A, 399-407. 106. Palkovits, M. ; Brownstein, M.J. In "Brain Microdissection Techniques"; Cuello, A.C., Ed.; John Wiley and Sons, Inc.: New York, 1983. Pp. 1-36.

PAGE 204

191 107. Slayback, J.R.B.; Story, M.S. Ind. Res. Dev. (1981) 23(2), 128-134. 108. Kirby, H. ; Sokolow, S.; Steiner, U. In "Tandem Mass Spectrometry"; McLafferty, F.W., Ed.; John Wiley and Sons, Inc.: New York, 1983. Pp. 337-352. 109. Yost, R.A.; Enke, C.G.; McGilvery, D.C.; Smith, D. ; Morrison, J.D. Int. J. Mass Spectrom. Ion Phys. (1979) 30, 127-136. 110. Cotter, R.J. Anal. Chem. (1979) 5j,, 317-318. HI. Perchalski, R.J.; Bruni, J.: Wilder, B.J.; Willmore, L.J. J. Chromatogr. (1979) j[63, 187-193. 112. Vanderborgh, N.E.; RolandJones, C.E. Anal. Chem. (1983) 55, 527532. 113. Dawson, P.H.; Fulford, J.E. Int. J. Mass Spectrom. Ion Phys. (1982) 42, 194-211. 114. Dawson, P.H. In "Quadrupole Mass Spectrometry and Its Applications"; P.H. Dawson, Ed.; Elsevier: Amsterdam, 1976. Pp. 23-24. 115, Perchalski, R.J.; Yost, R.A. ; Wilder, B.J. Anal. Chem. (1983) 55, 2002-2006. 116. B aldwin, M.A. ; McLafferty, F.W. Org. Mass Spectrom. (1973) 7, 13531356. 117. Hansen, G.; Munson, B. Anal. Chem. (1978) 50, 1130-1134. 118. Rapp, U.; Dielraann, G.; Games, D.E.; Gower, J.L.; Lewis, E. Adv. Mass Spectrom. (1980) 88, 1660-1668. 119. Reinhold, V.N.; Carr, S.A. Anal. Chem. (1982) 54, 499-503. 120. Cotter, R.J. Anal. Chem. (1980) 52^, 1589A-1606A. 121. Fentiman, A.F. Jr.; Foltz, R.L.; Kinzer, G.W. Anal. Chem. (1973) 45, 580-583. 122. Plattner, R.D. ; Yates, S.H.; Porter, J.K. J. Agric. Food Chem. (1983) 31, 785-789. 123, Fenimore, D.C.; Whiteford, J.H.; Davis, CM.; Zlatkis, A. J, Chromatogr. (1977) 140 9. 124. Benninghoven, A. Int. J. Mass Spectrom. Ion Phys. (1983) 53, 85-99.

PAGE 205

192 125. Oehme, M. Anal. Chem. (1982) 55, 2290-2295. 126. Balasanmugam, K. ; Hercules, D.M. Anal. Chem. (1983) 55, 145-1A6. 127. Verbueken, A.H.; VanGrieken, R.E.; Paulus, G.J.; Bruijn, W.C. Anal. Chem. (1984) 56, 1362-1370. 128. Yost, R.A,: Perchalski, R.J.: Brotherton, H.O.; Johnson, J.V.; Budd, M.B. Talanta (1984) 31, 929-935. 129. Stoll, R.G.; Harvan, D.J.; Cole, R.B.; Haas, J.R. (1984) Paper presented at the 32nd Conference on Mass Spectrometry and Allied Topics, San Antonio, May 27-June 1. Pp. 838-839. 130. Kelner, L.; Fales, H.M.; Markey, S.P. (1984) Paper presented at the 32nd Conference on Mass Spectrometry and Allied Topics, San Antonio, May 27-June 1. Pp. 850-851. 131. Busch, K.L.; Cooks, R.J. Science (1982) 218, 247-254. 132. Busch, K.L.; Unger, S.E.; Vincze, A.; Cooks, R.G.; Keough, T. J. Am. Chem. Soc. (1982) 104, 1507-1511. 133. Magee, C.VJ. Int. J. Mass Spectrom. Ion Phys. (1983) 49, 211-221. 134. Laser Focus Buyer's Guide (1984), 19th Ed., PennWell Publishing Co., Littleton, MA. 135. Dutta, P.K.; Talmi, Y. Anal. Chim. Acta (1981) 1^, 111-118. 136. Benyon, J.H.; Cooks, R.G.; Amy, J.W.; Baitinger, W.E. ; Ridley, T.Y. Anal. Chem. (1973) 45, 1023A-1031A. 137. Gross, M.L.; Russell, D.H. In "Tandem Mass Spectrometry"; F.W. McLafferty, Ed.; John Wiley and Sons, Inc.; New York, 1983. Chapter 12. 138. Todd, P.J.; McGilvery, D.C., Baldwin, M.A. ; McLafferty, F.W. In "Tandem Mass Spectrometry"; F.W. McLafferty, Ed., John Wiley and Sons, Inc.: New York, 1983. Chapter 13. 139. Ciupek, J.D.; O'Lear, J.R.; Cooks, R.G.; Dobberstein, P.; Schoen, A.E. Paper presented at the 32nd Conference on Mass Spectrometry and Allied Topics, San Antonio, May 27-June 1, 1984. Pp. 378-379. 140. B ricker, D.L.; Adams, T.A., Jr.; Russell, D.H. Anal. Chem. (1983) 55, 2417-2418.

PAGE 206

193 141. Stults, J.T.; Enke. C.G.; Holland, J.F. Anal. Chem. (1983) 55, 1323-1330. 142. Holland, J.F.; Enke, C.G.; Allison, J.; Stults, J.T.; Pinkston, J.D.; Newcome, B.; Watson, J.T. Anal. Chem. (1983) 55, 997A-1012A. 143. Glish, G.L.; Goeringer, D.E. Anal. Chem. (1984) 56, 2291-2294. 144. Hein, R.E.; Cody, R.B. Paper presented at the 31st Conference on Mass Spectrometry and Allied Topics, Boston, May 9-13, 1983. P. 800. 145. Mclver, R.T.; Hunter, R.L. ; Story, M. ; Syka, J.; Labunsky, M. Paper presented at the 31st Conference on Mass Spectrometry and Allied Topics, Boston, May 9-13, 1983. Pp. 44-45. 146. Armitage, M.A. ; Fulford, J.E.; Hughes, R.J.; March, R.E. ; Mather, R.E.; Waldren, R.M. ; Todd, J.F.J. Adv. Mass Spectrom. (1980) 8B, 1754-1763. 147. Stafford, G.C.; Kelley, P.E.; Reynolds, U.E.; Syka, J.E. Paper presented at the 31st Conference on Mass Spectrometry and Allied Topics, Boston, May 9-13, 1983. Pp. 48-49. 148. Glish, G.L. Paper presented at the 31st Conference on Mass Spectrometry and Allied Topics, Boston, May 9-13, 1983. Pp. 191192.

PAGE 207

BIOGRAPHICAL SKETCH Robert John Perchalski was born in Collingdale, Pennsylvania on February 8, 1948. He graduated from Central Catholic High School in Melbourne, Florida, in June, 1965, and entered the University of Florida in September. He received the Bachelor of Science degree with a major in chemistry in June, 1969. In the fall of 1969, he entered the University of Florida graduate school and completed the course work for the Ph.D. in analytical chemistry in June, 1971. At that time, he was commissioned as a 2nd Lieutenant in the U.S. Army and served three months on active duty at Ft. Benning, Georgia. Upon release from the Army, he accepted a position as a laboratory technologist at the University of Florida, College of Pharmacy. In August, 1972, he accepted the position of Research Chemist in the Medical Research Service at the Gainesville, Veterans Administration Medical Center. Until December, 1984, he directed the analytical research program for the Neurology Service for the investigation of the pharmacology and clinical effectiveness of new and established antiepileptic drugs. A large part of this effort involved the development of reliable methods of determining concentrations of these drugs, their metabolites, and neurologically important biomolecules in biological samples. In March, 1975, he received the Master of Science degree in analytical chemistry from the University of Florida. In September, 194

PAGE 208

195 1980, he returned to the University of Florida to complete work on the Ph.D. in analytical chemistry. Dr. Perchalski is now working as Head of Analytical Research and Development for Pharmatec, Inc., Alachua, Florida, a research firm specializing in the development of soft drugs and drug delivery systems, with particular emphasis on targeting drugs to specific organs. He is also Adjunct Assistant Professor in the Department of Medicinal Chemistry, College of Pharmacy, University of Florida. He has authored and co-authored a number of articles and chapters dealing with the determination of drugs in body fluids and tissues, and the pharmacology of antiepileptic drugs. He is married and has four children.

PAGE 209

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Richard A. Yost, Chairman Associate Professo\r of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. f Jy/)iU L InX^ lames D. Winefordi 'Graduate Research'' Prof essor of Chemistry I certify that I have read this study and that in ray opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. g . Cfj^ Eyler f Profess of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. r A^ John G.Dorsey \ Assistant Professor of Chemistry,

PAGE 210

I certify that I have read this study and that in ray opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. \=^ Joe Wilder Professor of Neuroscience L. This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Liberal Arts and Sciences and to the Graduate School, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May, 1985 Donald R. Price Dean for Graduate Studies and Research

PAGE 211

UNIVERSITY OF FLORIDA 3 1262 08553 3767