Development of laser desorption ionization on a quadrupole ion trap mass spectrometer

MISSING IMAGE

Material Information

Title:
Development of laser desorption ionization on a quadrupole ion trap mass spectrometer
Physical Description:
xv, 170 leaves : ill. ; 29 cm.
Language:
English
Creator:
Vargas, Rafael Roberto, 1966-
Publication Date:

Subjects

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

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1993.
Bibliography:
Includes bibliographical references (leaves 166-169).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Rafael Roberto Vargas.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001950908
notis - AKC7450
oclc - 31200861
System ID:
AA00003262:00001

Full Text









DEVELOPMENT OF LASER DESORPTION IONIZATION ON A
QUADRUPOLE ION TRAP MASS SPECTROMETER

















By

RAFAEL ROBERTO VARGAS


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


UNIVERSITY OF FLORIDA

1 002




























To my family












ACKNOWLEDGEMENTS


I give my sincerest thanks to Professor Richard A.


learned in and out of the laboratory.


Yost for all the things I


He set an example of excellence in research


communication


which


always


remember when


pursue


goals.


especially would like to thank Rick for the opportunity to learn and teach in the


electronics


lab--a


valuable


experience


for all


future


analytical chemists.


Also,


appreciate the opportunities and encouragement he gave me to participate in many

scientific meetings.

Thanks go to Dr. Jodie Johnson for many helpful hints and for seeing my best


results not only analytically,


but with excitement as well.


reminded me that


research should be fun during times when I temporarily had forgotten.


Thanks go


to my committee members, Drs. James


Winefordner,


Vaneica


Young, John Eyler,


and Buna


Wilder, for performing their duties with the utmost attention.


Thanks go


to Mike


Rosenberg


participating


in my


research


with


helpful


suggestions and providing the very important funding from Bristol-Myers Squibb.

On the subject of funding, I also extend my appreciation for the NASA joint project

that got me started on quadrupole ion traps and gave me the opportunity to learn.


Thanks


to the


support


staff whose


skills


used


greatly


in my years


at the






department.


particular


would


to mention Jeanne


Karably for


quickly


preparing my


TRs no matter I how much I managed to complicate them.


Special


thanks


to Nate


Yates


whose


programming


electronics


knowledge helped me to interface the laser to the trap, and to Matt Booth for not

only setting up ion injection, but also making it work well.

All the friend and family support I have received over the past several years

and the assurance that there is another world out there beyond the ivory research

lab were given to me by some very good friends, Tony Annacchino, Uli Bernier, Jon


Jones, Henry Lin, and Donna Robie.


To my very special family and their unabated


support, Nancy, Janice, and Steve, I thank them all.


Finally, I wish to acknowledge


my wonderful parents, whose love and support are beyond measure.


Someone once


said,


"If you want to be successful, then choose your parents carefully."


I believe I


chose the best.













TABLE OF CONTENTS


Page


ACKNOWLEDGEMENTS


. iii


LIST OF FIGURES


ABSTRACT


CHAPTER


INTRODUCTION


Project and Goals


Previous


Work


Purpose of Using the Quadrupole Ion


Quadrupole Ion
History.


2


Trap


Trap Mass Spectrometry


How the Ion


Trap


Works


Performing MS/MS on the Ion
LDI and MALDI Combined with MS


Trap


* t1


Matrix-assisted LDI.
How MALDI Works


* .


* 0


* 0 *
* 0* ft


Comparison to Other Desorption Methods


Recent Advances


in LDI/Quadrupole Ion


Trap MS


Overview of Dissertation


* 0 0 5 *
* 0 0 0 0


INSTRUMENTAL AND EXPERIMENTAL DESCRIPTION


Instrumental Description . .
Ion Trap Mass Spectrometer


Trap


Theory


Nitrogen Laser and Interface
Experimental Setup ..........
Scan functions and Timing .


n... : i... .... ... n -..- i -... ----'--- / -a


7
7
8


. 22
. 22
*... 23
. 26


. . . 4







Sample Preparation ....... . . . .


FUNDAMENTAL STUDIES OF
IONIZATION/QUADRUPOLE
SPECTROMETRY USING
CONFIGURATION..............


LASER
ION
AN


DESORPTION
TRAP MASS
INTERNAL


Effect of Radiofrequency Phase Angle on Trapping Efficiency
Effect of Radiofrequency Amplitude on Trapping Efficiency 4
Effect of Buffer Gas Pressure on Trapping Efficiency ......
C conclusions . . . . . . ...


LASER DESORPTION IONIZATION/QUADRUPOLE ION TRAP
MASS SPECTROMETRY USING AN INTERNAL SOURCE
CONFIGURATION


Laser Desorption Ionization . . . . .
MS/MS of Trimethylphenylammonium Bromide
Positive and Negative Ion Spectra .. . .


* 9 9 4
* 4 4 9 4


Matrix Compounds .
Matrix-Assisted Laser Desorption
MALDI of Spiperone .
MALDI Signal Lifetimes .
Laser Irradiance .......
Matrix Preparation ..
Lowest Sample Amount


MALDI/MS/MS of Spiperone
MALDI of Leucine Enkephalin
Spiperone in Matrigel ..
LDI of Spiperone in Matrigel
MALDI of Spiperone in Matrige
Conclusions .....


S. Ionization
Ionization .


* 9 4
* 9 4
* 9 4 4


* 4 S S S


* S 4 4 9
* S 4 4 9
* S 5 9 4
* 4 5 5 .4


* 4 4 *
* a .


* St 4 4 4
* S 4 4 4


* 5 122


LASER DESORPTION IONIZATION/QUADRUPOLE ION TRAP
MASS SPECTROMETRY USING AN EXTERNAL SOURCE
CONFIGURATION .... .......... ..............


Laser Beam/Ion Axis Placement


Instru
Laser


mental Description .
Desorption Ionization
Low Mass Cutoff .
Laser Irradiance
MS3 of Trimethylphel


. .
. .


nylammonium Bromide


r A a V 4r 5







MALDI of Spiperone ....
MALDI of Spiperone in Matrigel
Conclusions .


* .
* .


* 0


CONCLUSIONS


Summary . . . .
Suggestions for Future Studies


* 0 0
* 0 0 *


APPENDICES


ELECTRONIC CIRCUIT CALIBRATION DATA FOR THE
TRAPPING RADIOFREQUENCY PHASE ANGLE
EXPERIM ENT ..... .. ......... ...... ..


FORTH KEY SEQUENCE
ASSISTED RECORDING
ANGLE DATA AND R1
DATA .. .. .


PROGRAMS FOR COMPUTER
OF RADIOFREQUENCY PHASE
ADIOFREQUENCY AMPLITUDE


REFERENCES


BIOGRAPHICAL SKETCH











LIST OF FIGURES


Figure


Schematic of quadrupole ion trap mass spectrometer showing which
electrodes the various voltages are applied and the configuration of the


filament and detector


. . . . 9


Illustrative comparison of LDI and MALDI sample composition and
ionization products . . . . . . . . .


Plot of stability diagram defined by iso-f3 lines. Ions which have az, qz
coordinates within the bound region have stable trajectories ....


Schematic


of laser


desorption


ionization/quadrupole


ion trap


mass


spectrometer


Schematic of laser/quadrupole ion trap showing the introduction of the
laser beam and sample probe tip . . . . . .


Drawing of rf scan function showing sequence of important events


Diagram of the rf phase synchronization circuit.


The variable resistor


(Rv) is adjusted for different rf phase angle settings during the laser
pu lse . . . . . . . .


Timing plot for rf phase synchronization experiment.
trace descriptions .. . . . .


text for


Plot of rf phase angle versus variable resister position settings used for
rf phase angle study . . . . . . ..


Structures and molecular weights of UV absorbing matrix compounds


used


MALDI


experiments:


sinapinic


acid,


dihydroxybenzoic acid, c) 2,4-dinitroaniline, d) coumarin-120, and e)
nicotinic acid . . .. . . . . * *


Page






absorption spectrum


matrix


compound


smlnaplmic


acid.


Molar absorptivity at 337 nm is 14391 M1 cm'


Laser desorption ionization mass spectrum of graphite.
scans (1 microscan each scan) . . .


C3 intensity and trapping rf amplitude versus


Average of 10


rf phase angle; a) Run


Run


Error


bars


designate


standard


error


mean.
. .


Average of 10 scans (1 microscan each scan)


intensity and trapping rf amplitude versus rf phase angle; a) Run


Run


Error


bars


designate


standard


error


mean.


Average of 10 scans (1 microscan each scan)


5+ intensity and trapping rf amplitude versus rf phase angle; a) Run


1, b)


Run


Error


bars


designate


standard


error


mean.


Average of 10 scans (1 microscan each scan)


intensity versus low mass cutoff.


microscans


scan each data


each scan)


point (10
. . .


56Fe +


intensity versus


scan


each


data


microscans each scan)


point
. .


Copper isotope ion intensities versus qz:


a) 63 Cu


intensity


vs. q, for


m/z 63, b) 65Cu+ intensity v,
microscans each data scan)


qz for m/z


intensity versus He buffer gas pressure.


standard error of the mean.


1 scan each data point (5


Error bars designate


50 scans each data point (1


microscan


each scan)


56Fe+


intensity versus Ar buffer gas pressure.


Error bars designate


standard error of the mean.


50 scans each data point (1 microscan


each scan)


Copper isotope ion intensities versus helium buffer gas pressure: a)


63C +


intensity


He pressure, b)


Cu+ intensities


He pressure.


Error bars designate standard error of the mean.
point (1 microscan each scan) .


50 scans each data


3-11.


63u +


Cu + isotope ratio versus


He buffer gas pressure.


50 scans each


data point (1 microscan each scan)


56Fe+


56Fe+






3-12.


Copper isotope
63Cu+ intensity


ion intensities versus argon buffer gas pressure:


vs. Ar pressure,


b) 65Cu+


intensities vs. Ar pressure.


Error bars designate standard error of the mean.
point (1 microscan each scan) . . . .


3-13.


50 scans each data


63Cu+'SCu+ isotope ratio versus Ar buffer gas pressure


Laser desorption ionization mass spectra of trimethylphenylammonium


bromide,


full mass scan,


single


ion isolation


of the


fragment ion


LDI/MS/MS of 121+ ion of trimethylphenylammonium bromide


LDI/MS of sodium ions, a) positive (average of 5 scans, 20 microscans


each scan), b) negative ion spectra.


Several peaks are shifted to the


right due to space charging (a) or calibration error (b)


LDI/MS of 2,5-dihydroxybenzoic acid

LDI/MS of 2,4-dinitroaniline .....


Mass chromatograms for LDI/MS of 2,4-dinitroaniline .. . .

LDI mass spectrum of sinapinic acid, a) full mass scan, b) isolation of
molecular ion at m/z 224 . . . . . .


Daughter mass spectrum of sinapinic acid molecular ions produced by
laser desorption ionization . . . .


4-9.

4-10.


4-11.


Laser desorption ionization mass spectrum of spiperone (MW 395)

Matrix-assisted laser desorption ionization mass spectrum of spiperone
using nicotinic acid matrix (0.5% TFA) in methanol . .


(M+H)+ signal of spiperone versus scan for MALDI using an acidified
nicotinic acid matrix and LDI .


4-12.


(M+H)


and 165' fragment ion of spiperone versus scan


4-13.


MALDI mass spectrum of spiperone using 2,5-dihydroxybenzoic acid


with


0.1%


TFA,


laser


irradiance


=6.4


x 106


W/cm2,


laser


irradiance =


x 106 W/cm2






4-14.


4-15.


4-16.


MALDI mass spectrum of spiperone using 2,5-dihydroxybenzoic acid
matrix in 70:30 acetonitrile:water and 0.1% trifluoracetic acid (low
m ass cutoff = 150 u) . . . . . . . . .


MALDI/MS of 74 ng of spiperone (4 pL of 47 CpM methanol solution)
mixed with 4 pL of 48 mM DHB matrix solution ....


MALDI/MS/MS daughter spectrum of protonated spiperone, m/z 396
(average of 10 scans) . . . . . .


4-17.


MALDI/MS/MS daughter spectrum
spiperone (average of 10 scans) .


fragment


4-18.


4-19.


4-20.








4-21.


4-22.


Electron ionization of spiperone off the solids probe on the ion trap
mass spectrometer (10 microscans). Low mass cutoff during trapping
was 40 u and ionization time was 2 ms . . . . . .


Methane chemical ionization of spiperone off the solids probe on the
ion trap mass spectrometer (10 microscans). Reaction time was 25
l S S S S S S S S S S S S S S S S S S S S S S * S S S S S S S S S S S S


Intensity versus scan plot of protonated spiperone (m/z 396) using
methane chemical ionization of spiperone off the solids probe (10
microscans). Reaction time was 25 ms. Manifold temperature was 100
degrees celsius. Initial solids probe temperature was 70 degrees celsius
and the temperature ramp was 120 degrees/minute, but the maximum
ion signal (at scan #132 and retention time 4:57) was obtained after
the reaching the final probe temperature . . . . .

MALDI mass spectrum leucine enkephalin. an example of a compound
which cannot be thermally desorbed. . . . . . .


MALDI mass spectrum of leucine enkephalin showing the appearance
of adduct ions...........


LDI/MS of 5 IpL
matrix compound
during trapping
microscan .


of spiperone in Matrigel mixture (100 ppm). No UV
ds or TFA were added. The low mass cutoff is 40 u
and storage. The mass spectrum is a single


4-24.


MALDI/MS of spiperone in Matrigel (average of 20 scans)


4-23.





4-25.


MALDI/MS of spiperone in Matrigel with isolation of (M+H)' of
spiperone using DHB matrix solution (0.1% TFA) and low mass cutoff


= 150 u.


Mass


spectrum shows effects of delayed ejection of ions or


neutrals into the ion trap analyzer volume


Drawing


laser


beam


sample


probe


placement


external source. .


Schematic


external


source


configuration


laser


desorption


ionization/quadrupole ion trap mass spectrometer


Schematic of laser desorption external ionization ion trap showing the
configuration of components for delayed laser triggering and resonant


excitation


LDI/MS mass spectrum of trimethylphenylammonium bromide.


mass cutoff is 10 u.


Low


Average of scans #5 through #20


Plot of trimethylphenylammonium bromide ion intensities versus low
mass cutoff for trapping and storage. Each data point is an average
of the ion intensity from scan #5 through #20 . . . .


of trimethylphenylammonium bromide


120/121


intensity


ratio versus low mass cutoff for trapping and storage.


Each data point


is an average of the ion intensity from scan #5 through #20


Plot of trimethylphenylammonium bromide ion intensities versus laser


energy


transmission.


Each


data


point is an


average of the


intensity from scan #5 through #20


LDI mass spectra of trimethylphenylammonium bromide: a) MS/MS
of molecular ion at m/z 136, b) MS/MS/MS of molecular ion (136+ --


121f


-fragments)


MALDI


mass


Average of


spectrum


10 microscans.


spiperone
1 pL of 4'


using external
7 piM spiperon<


configuration.
e in methanol


sample mixed with 1


IpL of


20 mM DHB matrix solution


5-10.


intensity


protonated


spiperone,


from


MALDI/MS


versus scan for external source configuration. . . .






5-11.


5-12.


5-13.


MALDI/MS/MS daughter spectrum of protonated spiperone, m/z 396,
using external source configuration. Average of 10 microscans. 1 pL
of 47 pM spiperone in methanol sample mixed with 1 iL of 520 mM
DHB matrix solution. Resonant excitation frequency is 113 kHz and
amplitude is 6 V (p-p). . . . . . .


MALDI/MS of spiperone in Matrigel membrane. Average of 10
microscans. 2 pL of 0.5 mg/mL of spiperone in Matrigel sample mixed
with 2 pL of 520 mM DHB matrix solution . . . .


Ion intensity of protonated spiperone from MALDI/MS of spiperone
in Matrigel membrane versus scan for external source configuration.
2 p.L of 0.5 mg/mL of spiperone in Matrigel sample mixed with 2 ipL
of 520 mM DHB matrix solution. . .












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

DEVELOPMENT AND APPLICATION OF LASER DESORPTION
IONIZATION ON A QUADRUPOLE ION TRAP MASS SPECTROMETER

By


Rafael Roberto Vargas

December 1993


Chairperson: Richard A.


Yost


Major Department: Chemistry

The purpose of this research is to develop and characterize laser desorption


ionization


(LDI)


on a quadrupole


ion trap


mass


spectrometer


to evaluate


potential for development of the first true molecular microprobe.


was used to desorb and ionize solid samples,


mass spectrometry (MS/MS).


A nitrogen laser


with the ions analyzed using tandem


The ultimate goal of this method is to determine drugs


in biological tissues and to obtain spatial information about drugs in


matrix.


the cellular


This method would not only be able to detect molecular ions produced by


laser desorption ionization, but also to determine the structure of the molecular ion


of interest by performing MS/MS.


quadrupole


ion trap


mass spectrometer is


utilized because of its sensitivity and selectivity, and its capability to obtain complete

MS and MS/MS spectra from a pulsed source of ions.






In an internal ionization configuration, the laser beam is directed into the ion


to focus


upon


sample


surface


positioned


flush


with


ring electrode


surface.


In an external configuration, the laser beam is directed into an external


ionization source with injection of the ions into the trap.


LDI on the ion trap has


been achieved and tandem mass spectrometry has been performed on laser-desorbed


ions.


Matrix-assisted laser desorption ionization (MALDI) has been evaluated to


improve


determination


spiperone


other


thermally


labile,


nonvolatile


compounds.


Spiperone


drug used


this research as a


model


of the


drug


compounds


to which


method


could


applied.


Spiperone


Matrigel


membrane,


an extracellular


matrix


used


as a tissue


model,


been


detected.


Results were obtained which provide a better understanding of the laser-desorbed


ion trapping process to increase the sensitivity of the method.


For example,


trapping efficiency was found to be dependent on the radiofrequency phase angle


during


laser


pulse.


effects of buffer


pressure


radiofrequency


amplitude


trapping


efficiency


were


discussed.


results


obtained


demonstrated


potential


LDI/quadrupole


ion trap


mass


spectrometry


achieving a highly sensitive and selective imaging technique for drugs in biological

matrices.











CHAPTER 1
INTRODUCTION


Project and Goals


The ability to analyze specific drug compounds located in biological tissues


provides


pharmacological


researchers


with


very


useful


information


about


physiological


microprobe


metabolic


method,


action


those


specific


compounds.


reactivity


or drug


With


use of


location


could


determined


various


regions


cellular


matrix


(e.g.,


intracellular


extracellular).


Microprobe


methods


currently


available,


such


laser


microprobe/time-of-flight mass spectrometry, however, do not provide the selectivity


to determine a


wide


range of


drug compounds.


Lack


of selectivity can


lead to


erroneous


identification


compounds,


particularly


a complex


matrix.


purpose of this research is to develop and characterize laser desorption ionization

on a quadrupole ion trap mass spectrometer in order to evaluate the potential for


development of the first true molecular microprobe.


This method would not only


detect molecular ions produced by laser desorption ionization, but also determine the


structure of the


molecular ion of interest by performing multiple stages of mass


spectrometry.


quadrupole


ion trap


mass spectrometer


is used


because


of its








sensitivity and selectivity,


its capability to obtain


complete


MS/MS


spectra from a pulsed source of ions.

In this research, laser desorption ionization has been achieved on the ion trap


and tandem mass spectrometry of laser-desorbed ions has been performed.


Matrix-


assisted laser desorption ionization has been


utilized on


the ion


trap to improve


detection and determination of spiperone and other thermally labile,


compounds.


nonvolatile


Spiperone is an antipsychotic drug that is used in this research as an


example of the drug compounds to which this method can be applied.


Spiperone in


Matrigel membrane, an extracellular matrix used in this work as a tissue model, has


been detected.


Experiments to better understand the process of trapping laser-


desorbed ions have been performed so that the sensitivity of the method may be


increased.


example,


a radiofrequency phase


angle


effect


on the


trapping


efficiency


laser-desorbed


ions


was


studied,


an experiment


required


development of phase angle/laser trigger synchronization electronics.


The effects of


buffer gas pressure and radiofrequency amplitude on the trapping efficiency of laser-

desorbed ions have also been shown.


Previous Work


This research arose from the successful results previously obtained by Bob

Perchalski performing laser desorption on a triple quadrupole mass spectrometer.1'2

The researchers attempted to take advantage of long lived neutral production from








quadrupole


could


mass


scanned


across


resultant


ion pulse.


triple


quadrupole was used because it of its ability to perform MS/MS and to perform


rapid mass scans.


In that work, a pulsed dye laser was interfaced with a Finnigan


TSQ45 triple quadrupole mass spectrometer.


Samples were mounted on electron


microscope grids attached to stainless steel solids probe extensions.


angles were 450


The surface


or 900 for laser incidence and the ion pulse generation angles were


(along


quadrupole


axis)


or 900


(perpendicular


to quadrupole


axis),


respectively.

laser event.


The method involved a mass scan over several seconds after a single

With the long-lived ion production from CI of laser-desorbed neutrals,


the quadrupole mass analyzer could be scanned across a selected mass range.


The results included MS/MS of phenytoin for successive laser shots.


Peaks


in the ion current versus time represented each laser pulse and the mass spectra


obtained were


found to be similar to


those


obtained from


direct probe CI.


another study, the instrument was employed to monitor parent ion/daughter ion pairs


for each of nine drugs in an antiepilectic drug mixture.


The most impressive results


were the detection of phenytoin in mouse liver tissue at low levels of concentration


by obtaining complete daughter spectra of the laser-desorbed (M+H)


ions.


research,


it was apparent that obtaining complete


or MS/MS


spectra of ions produced by the laser event, and therefore with lifetimes of less than


1 ms,


was beyond the ability of the instrument due to the length of the mass scan.


However,


the beam type instrument could analyze the ions produced by chemical








laser pulse.


The research showed the potential of MS/MS for multicomponent, site-


specific, molecular analysis by laser desorption MS/MS, but clearly demonstrated the


value of


an MS/MS


instrument capable


of analyzing the


narrow pulse


of laser-


desorbed ions.


Purpose of Using the Quadrupole Ion Trap


One of the most sensitive and selective mass spectrometers is the quadrupole


ion trap mass spectrometer.3


Originally utilized to study volatile organic species, the


has been


used recently to analyze various samples with


the addition of


various


desorption


and spray


techniques.


Nearly


recently


developed


desorption


and/or


ionization


methods


have


been


combined


with


ion trap4


because of the widely recognized potential of the instrument.


Jodie Johnson in our


group


demonstrated


ion trap


higher


collision-induced


dissociation


efficiencies


(typically


80-90%)


MS/MS


than


triple


quadrupole


mass


spectrometer (typically 50-60%).3


In the same report, quantitative studies revealed


that the ion trap could obtain complete daughter spectra on low picogram amounts


of analyte,


hundred times


lower than the amounts required for comparable


spectra on the triple quadrupole.

While several researchers have demonstrated tandem mass spectrometry using

dual time-of-flight stages of mass analysis,5'6 this capability is not widely available


as yet and has not been applied to laser sampling methods.


Time-of-flight mass








the 1970s, laser microprobes were developed with highly focused beams (<

size) which could be aimed at a selected location on a sample surface.7


1 pm spot

Initially,


these instruments were used only for elemental analysis and mapping surfaces for

elemental concentration. Several years later, the first methods to produce molecular


from


laser


desorption


appeared


were


applied


laser


microprobe


instruments.


Materials that were studied with


this technique included polymers,


metals, biological materials, and semiconductors.8


rabbit liver tissue,


For biological materials such as


the localization of aluminum in cells was performed using this


method.9


These laser microprobes,


however,


incorporates only a


single stage of


time-of-flightmass spectrometry and thus cannot perform tandem mass spectrometry.


Laser


microprobe


Fourier-transform


mass


spectrometry


(FTMS)


instrumentation is available commercially (e.g., EXTREL Laser ProbeTM MS).


instruments provide accurate mass measurement and high resolution.


These


Molecular


weight and structural information on biomolecules and polymers have been obtained


using these instruments.'0


Multiple stages of analysis can be performed with dual


cells


stored


waveform


inverse


Fourier-transform


excitation,


picomole


detection limits have been obtained for biomolecules.1'


Laser microscopy for trace


level analyses, however, has not been utilized with this method.

Multiple stages of mass analysis (MS", where n 13) have been demonstrated


on the quadrupole ion trap.4


The ion trap is also much simpler and more compact


than


other


tandem


mass


spectrometry


instruments


such


sector,


hybrid






6

as well as less stringent vacuum requirements, is also characteristic of the ion trap


not shared by the Fourier-transform instruments.


In terms of figures of merit, the


best results obtained for limits of detection for the ion trap are in the attomole (10'18


mole) region.4 The best linear dynamic range reported equaled over six orders of

magnitude.12 As are these figures of merit, analytical utility is an important quality

of an instrument. With the addition of very different ionization methods, the ion


trap has been used to study a wide range of samples from volatile organic molecules


to involatile


inorganic and


biological


molecules.


particular,


when


analyzing


biological


molecules,


ionization


methods


including


electrospray


ionization


atmospheric glow discharge ionization have been performed successfully on the ion


can provide


analytical


information


complementary


to the


information


obtained using laser desorption ionization.


The future of the quadrupole ion trap holds much promise as well.


Recent


research has demonstrated dramatic increases in the mass range; ions of masses up


to 70,000 u have been detected.14'15


High mass resolution has been performed by


others as well.16'17


These advances in


technology have greatly increased


the ion trap's ability to analyze large molecules.18.19

Ongoing research also suggests that the future of the quadrupole ion trap will

include a wider range of analytical applications, improved resolution, combination


of the


trap with other separation methods,


the automation


of the acquisition of


MS/MS data, and the study of different trap geometries as well as nondestructive







has also been studied


for use


as analyzer for an


expert system


to monitor


enclosed environments autonomously; suggesting an instrument completely free of

operator input and control.


Quadrupole Ion


Trap Mass Spectrometry


History


In 1956,


Wolfgang Paul first described a device using only electric fields for


containing gaseous ions.20


By the time Paul and Hans Dehmelt were awarded half


the Nobel Prize in physics in


1989,


the quadrupole ion trap had successfully made


a place for itself in the analytical sciences.

was in 1983 as the Finnigan MAT Ion T1


The commercial introduction of the trap


rap Detector; over the past decade it has


become a standard instrument for mass spectrometry.


Certain advances in ion trap technology were required for the instrument to


become an analytically useful tool in chemistry.


These advances include the addition


of a light buffer gas,21 alternative scan functions such as parent scans to increase


analytical


tandem


usefulness,22 positive


mass


negative


spectrometry.20


chemical


ability


ionization,:3


isolation


to combine


with


chromatography has proven to be the most useful of all applications, and currently

the ion trap is most widely used in the form of a benchtop gas chromatography/MS

system.








How the Ion


Trap Works


quadrupole


ion trap


is made


three


cylindrically


symmetric


electrodes, two endcap electrodes with a ring electrode between them.


surfaces of these electrodes are hyperbolic.


two endcaps are grounded while


The inner


In the common method used here, the


trapping radiofrequency (rf) potential,


known as the drive potential, is applied to the ring electrode, as shown in Figure 1-1.


This radiofrequency is typically around 1 MHz.


The Finnigan MAT Ion


Trap Mass


Spectrometer


(ITMS),


used


internal


configuration,


operates


a drive


frequency of 1.1


MHz.


The trap is controlled by a computer processor to allow


complete adjustment of scan functions and data collection using only a


computer.


personal


Scan functions represent the temporal variation of the rf potential applied


to the ring electrode and are used to display the relationship between the trapping

field and other events that occur during a mass analysis experiment (e.g., ionization,


cooling, isolation).


MS or


Using a personal computer, the operator only needs to write an


MS/MS scan function into the software; no other hardware or electronics


modification is necessary.

After ions are created within the trap by injection of electrons through an


endcap


electrode,


or after


are injected


trap,


a range


mass/charge values may be held in stable orbits inside the trap depending on the


amplitude of the rf potential applied to the ring electrode.


A stable orbit has the


fnrm nC n thra _.. mnin niLL l T oocliic Tfu'llrch inA c lf npA\\rd l fha r2Ai!11l anti navinl













Ion


Trap


Mass


Spectrometer


(ITMSTM)


Filamen


Electron Multiplier Detector


To Preamplifier
(Ion Signal)


Amplifier and
RF Generator,


upple


mentary


RF Voltage


Figure


Schematic of quadrupole ion trap mass spectrometer showing which
electrodes the various voltages are applied and the configuration of the


filament and detector.


Amplifier and
RF Generator
Fundamental
RF Voltage


Scan Acquisition


Processor
(Computer4






10

As the amplitude of the rf is increased, ions of increasing mass/charge value


approach


low mass


cutoff,


become


unstable


the z-direction


(towards


endcaps), and exit the ion trap through holes in one of the endcaps and strike a


detector.


The detector positioned behind one of the endcaps is typically an electron


multiplier that produces the


ion current signal.


Most often, several mass scans,


called "microscans,"


are summed


and the averaged mass spectrum


is sent to the


computer for display and recording.


Performing MS/MS on the Ion Trap


As stated previously,


performing MS/MS on


the quadrupole


ion trap only


requires editing the software in order to carry out the added steps required in a scan


function.


Several methods exist for performing MS/MS.


Apex isolation is the most


common method of ion


isolation


ion trap;


as do


most other methods,


requires a dc potential applied to ring electrode.


The proper setting of rf and dc


potentials, which will be discussed in detail in Chapter 2, will create an electric field


inside


the ion


trap where only ions of a particular mass/charge value will have a


stable trajectory and all other ions of masses above and below that value will have


unstable trajectories.


These other ions will be ejected from the trap or strike an


electrode, and they will no longer be stored and detected in


the subsequent mass


scan.


The purpose of isolating ions of a single mass/charge is so


that they (the








be interfered with by the other ions.


These daughter ions, will then be products only


of the parent ion and the information obtained in the daughter ion spectra can be


utilized


as such.


Several


methods


exist


fragmenting


parent


Throughout


work


presented


dissertation,


method


resonantly


enhanced excitation of the parent ion for collisionally induced dissociation (CID) was


used.


Application


a resonant excitation


voltage


(applied


across


endcap


electrodes)


induces the parent ions to achieve more energetic orbits in


the axial


direction.


These energetic orbits cause the parent ions to collide with the required


buffer gas inside


trap;


these collisions can eventually


lead to dissociation as


presented in the reaction


(1-1)


where P+


is the parent ion, D


represents the daughter ions, and N represents the


neutral fragments. The daughter ions that are trapped can then be scanned to obtain

a daughter mass spectrum. As stated previously, tandem mass spectrometry provides


a second dimension of selectivity and valuable information for ion identification.


LDI and MALDI Combined with MS


Laser desorption is a method by which atoms and molecules in the condensed

phase are transferred to the gas phase through the absorption of sufficient energy


from laser radiation.


The process produces both neutral and ionic species.


Laser


- Jr -r 1 4 -. n


- D


L,,,,,:~. ~..~ln:rnlrr


-1. d1-.... A. --


1








production of ionic species.


The history of lasers combined with mass spectrometry


began about thirty years ago with the use of high laser energies to perform elemental


analyses.6


With


eventual


addition


molecular


analyses


to the


laser


desorption


method,6 the


technique showed great potential as an


analytical


tool.


Adding to this potential is the ability to tightly focus laser beams all the way down


to their diffraction limits making microscopic imaging by LDI possible.


When tight


laser focussing was not necessary, laser radiation can easily be introduced into the


vacuum


chambers


ionization


sources


using


fiber-optics.24.25


LDI


mass


spectrometry has been used primarily as a qualitative technique, but there have been

many applications of LDI/MS to quantitative studies as well.26

Several other uses of lasers with mass spectrometry have made a widespread


appearance


performed


over


with


several


mass


decades.


spectrometry."7


Laser


many


photodissociation


different


been


photodissociation


methods include two-pulse experiments, where the first pulse is used to desorb ions


and the second pulse is used to photodissociate those ions.


The pulses can both


come


from


a single laser


or individually from


two different lasers,


the second of


which (the photodissociation laser) is tuned to the desired frequency for absorption


by the ions.


Combination of laser desorption with chemical ionization, to perform


CI on laser-desorbed neutrals, has also been achieved.2


Laser desorption is a


very versatile analytical tool because it can produce


positively and negatively charged species.


" The ionic species may be either singly








molecular and quasi-molecular particles.


Much research has been performed over


two decades


to study


the controlling factors in


determining the species


produced and to manipulate the laser desorption conditions to obtained the desired

products.6

Most of the theories presented for the production process of laser-desorbed

particles are divided into two categories based upon the general conditions of laser


desorption.

laser irradia


The first condition involves the slow thermal heating of the sample upon

... S, 8 rlZ~ rr't:. -I -n -n 1 -. -1-


LIUIn.


This process typically creates a long la s


from several


milliseconds


to several


seconds


length.


Often,


this thermal


process is performed using an infrared laser.


The second condition involves the


quick ablation


sample


upon


laser


irradiation.


This


fast process


usually


produces an equally fast ion pulse that lasts from several nanoseconds up to several


microseconds in length and is often performed with an


ultraviolet laser.


very


important characteristic of this method is the lack of thermal energy transfer from


the sampling spot to the surrounding sample.30


In electron micrographs of sample


craters produced from this nonthermal process, the hole is usually very clean and has


the same shape as the laser beam profile.


surrounding sample.


Also, there is no visible charring of the


The methods of laser desorption ionization and matrix-assisted


laser desorption ionization discussed in the following chapters are both part of the

second condition of laser desorption.

Mass spectra of the ions produced from laser desorption typically show ions








electron removal or addition.30


frequently.


Cationization by alkali metal ions also occurs very


This tendency has been utilized to improve the production of ions by


increasing the appearance of molecular adduct ions (also called


quasi-molecular


ions).


ions


produced


routinely


have


a widespread


kinetic


energy


range.31


Average kinetic energy of laser-desorbed ions also varies over the time immediately


after the


laser


pulse.29


The more


polar a sample


more


easily


it can


desorbed as an ion.


This relationship is opposite to what is normally considered for


electron ionization or chemical ionization, in which the required volatility is generally


favored by lower polarity.


It is also common practice to use preformed ions for


calibration and


testing purposes in LDI/MS.


important parameters in laser


desorption are photon wavelength, laser irradiance, laser pulse width, absorption

spectrum of the sample, the orientation of the impinging laser beam to the sample,

sample thickness and smoothness, sample homogeneity, and substrate composition.


These


parameters


given


each


various


experiments


presented


throughout this dissertation and periodically discussed for their effect on obtaining

good mass spectra.


Matrix-assisted LDI


As stated before, when the use of lasers with mass spectrometry first began,


high


laser


irradiances


available


were


applied


to the


rapid


vaporization


materials for elemental analysis and pyrolysis.


The method of matrix-assisted laser







1987


makes


possible


routine


analysis


intact


biological


molecules.30'32


molecular weights of biomolecules,


therefore, can be obtained.


These molecules


include peptides, proteins, glycoproteins, glycosides, nucleosides, nucleic acid, and


oligosaccharides.


Molecular masses of up to approximately


300,000 u have been


obtained using MALDI and studies to improve sensitivity and resolution continue to


reported.33


Although


MALDI


been


primarily


used


to study


biological


molecules with molecular masses of 1000 u or greater, the mass range of analytes is


much


lower in


research


discussed


in later chapters.


MALDI


is used


research to increase the production of molecular ions to increase sensitivity, since the

analytes are thermally labile compounds and easily fragmented by laser desorption

ionization.


How MALDI Works


In order to avoid fragmentation, a matrix compound rather than the analyte


absorbs the laser pulse energy.


Further, analyte molecules are separated within the


matrix to prevent clustering, as shown in Figure


The matrix molecules.


which


typically


outnumber


analyte


molecules


1000-to-1,


provide


a source


hydrogen for protonation reactions.


In effect, the matrix absorbs the laser energy


which volatilizes the analyte molecules, and soft chemical ionization of analyte occurs

in the selvedge region just above the sample surface.30 The matrix compounds share

the following characteristics: low sublimation temperature, high absorptivity at the
U4 4 *4 t* -* d* 4* 9*- -- f


























FRAGMENT IONS


UV LASER PULSE






MOLECULAR I
MATRIX IC


PURE ANALYTE


ANALYTE IN MATRIX


Figure


Illustrative comparison of LDI and MALDI sample composition and
ionization products.







Thus


MALDI


ionization


mechanism


not been


completely


understood.


What is known is that MALDI provides an effective and controlled


transfer of the laser energy to the condensed phase by predominantly one-photon


absorption.


This absorption


then


followed


an expansion


analyte


carried away by the ejection of matrix molecules from the sample surface.


MALDI


therefore involves the ablation of several layers of molecules per laser shot and is


considered a bulk analysis method, as opposed to a surface method.


The irradiance


threshold for optimum molecular ion production is uniform and around 106 W cm2

Pan and Cotter, using a delayed ion extraction linear time-of-flight MS method, have


shown the average velocity of ions produced by MALDI to be approximately


1000


a separate


study,


Beavis


Chait


showed


average


velocity


approximately


independent


mass.35


These


velocity


characteristics,


along with


other


characteristics,


demonstrate


that ion


production


depends primarily on


the matrix molecules,


not on


the analyte


molecules.


This


property is seen in the uniformity of the sample preparation methods and the quality

of the mass spectra obtained from MALDI.


Typical ultraviolet energy sources used are pulsed Nd:YAG,


excimer, and


nitrogen


lasers.


matrix


compounds


used


are sinapinic


acid,


gentisic


acid,


nicotinic acid, nitrobenzyl alcohol, and derivatives of these compounds.







Comparison to Other Desorption Methods


As mentioned before, MALDI is a bulk method, whereas secondary ion mass

spectrometry (SIMS) and fast atom bombardment (FAB) MS are surface methods.


MALDI


involves


analysis of many molecular


layers,


where


SIMS


FAB


analyze a single layer or a few layers.36


For the purpose of microprobe analysis,


FAB is typically not used since good spatial resolution or control of fragmentation


cannot be achieved as with laser desorpi

used as an ion microprobe technique.


SIMS, on the other hand, is commonly


Originally, used for elemental mapping like


laser microprobe mass spectrometry, SIMS is now also being used for the analysis

of organic species as well.


the SIMS


method,


a focused


primary ion


beam


is rastered


across


sample


to produce


secondary ions


representative


sample


surface.


primary beam may be typically Cs'


ions with several keV


of energy.


The images


produced give the intensity of mass-selected secondary ion emitted as a function of


position of the primary ion beam.


SIMS has been combined with triple quadrupole


mass spectrometers, with the second stage of mass analysis improving the quality of


chemical


image.37


nonconducting


samples


such


as biological


tissues,


however, sample surface charging can occur from the influx of the primary ions, and

decrease the efficiency of the desorption process.







Recent Advances in LDI/Quadrupole Ion


Trap MS


To introduce


following chapters on


the dissertation research


on laser


desorption ionization/quadrupole ion trap


reported in this area.


MS, a short summary is given of work


Details of these various studies by other research groups will


not be given here as they will be provided in the appropriate chapters for comparison

to the results presented in this dissertation.

Laser desorption on the quadrupole ion trap was first reported by Heller et


in 1989.


A CO2 laser was used to produce molecular and fragment ions from


organic and


biomolecular


compounds.


These compounds


included


sucrose


leucine enkephalin doped with potassium chloride to promote cationization.


configuration allowed the instrument to be used in its normal electron ionization


mode as well.


Electron ionization was also performed during the laser desorption


ionization experiments, resulting in additional fragmentation in the mass spectra.

Tandem mass spectrometry on laser-desorbed ions in the quadruple ion trap was

first reported by Glish et al. and used to demonstrate the production of daughter ions


and granddaughter ions from laser-desorbed parent ions.39


The researchers also


applied this new technology for mixture analysis.40

Laser ablation sampling inside the ion trap volume has been used to obtain


direct


atomic


mass


spectrometnric


analysis


solid


metal


samples.41


pulsed


Nd:YAG laser beam was used to ablate sample pins inserted radially through the


ring electrode and a mass spectrum of silver solder was obtained.


The preliminary







combine the attributes and advantages of both methods to achieve an instrument

unique in analytical capability, sensitivity, and selectivity.


Overview of Dissertation


At the beginning of this research project,


only preliminary results by two


research


groups


been


reported


laser


desorption


inside


ion trap.38'39


Therefore, little was known about either the fundamentals of trapping laser-desorbed


ions,


or the ability to detect thermally labile


molecules with


this method.


interface of hardware, electronics, and software to achieve laser desorption on the


ion trap is described in Chapter


of this dissertation.


With the instrumentation in


place


functioning,


research


was


to determine


characterize the most important factors in trapping ions formed at the ring electrode

surface, particularly when it was possible that laser-desorbed ions were too energetic


to be trapped efficiently.


These factors include buffer gas pressure and low mass


cutoff settings.


It was also believed


that the radiofrequency phase angle at the


instant of laser desorption could affect trapping efficiency if the ion pulse was shorter


in time than the trapping radiofrequency cycle.


All of these fundamental studies are


described in Chapter 3 of this dissertation.

The second step in this research was to develop the capability to desorb and


ionize molecules of interest intact with little or no fragmentation.


Matrix-assisted


laser desorption ionization had not been performed on the ion trap for this purpose.


F-.- .1 ,' .1 A W IW*I


rrP1 n







irradiance for trapping, the length of the ion signal, and the sample preparation with


UV-absorbing matrices were


studied.


results of these


studies


are given in


Chapter 4.


In Chapter 5, the external configuration for laser desorption is presented


and compared to the internal method.


Both methods are considered for their effect


on ion


signal


lifetime,


sensitivity,


ability


to perform


MS/MS.


Chapter


summarizes all of the conclusions drawn from the previous chapters and includes

suggestions for future studies.











CHAPTER


INSTRUMENTAL AND EXPERIMENTAL DESCRIPTION


Instrumental Description


Ion Trap Mass Spectrometer


A Finnigan MAT Ion


Trap Mass Spectrometer (ITMSTM) was used for this


research.


It has MS" capability with the use of a Selective


Mass Storage box (dc


voltage supply) and an auxiliary frequency synthesizer for resonance excitation (tickle


voltage) and axial modulation.

inside the heated manifold by


A base pressure around 8


turbomolecular pump (170


10-8 torr is maintained

L/s pumping speed).


Helium is introduced into the manifold through a Granville-Phillips leak valve to act

as a buffer gas and help trap and collisionally cool ions to the center of the trap.

Gas pressures were measured on the ITMS by a Bayard-Alpert type ionization gauge


(Granville-Phillips)


mounted


on the


vacuum


chamber.


Pressure


measurement


corrections were made for helium using factors determined in our laboratory.42


mass-selective


instability


scan,


or ramping


radiofrequency


voltage


described in Chapter 1,


was used for ion detection.21


Ion currents were measured


outside the ion trap during rf voltage ramping by an electrically shielded electron


multiplier.


In several experiments, a conversion dynode was also used to increase


sensitivity.








Trap


Theory


Several


parameters


to describe


ion motion


are used


later


chapters to


discuss the laser desorption results.


The derivation of these parameters from ion


trap theory is given here; equations were obtained from March and Hughes.3

The potential ( ) inside the ion trap is described by the following equation


+ VcosfRt)


- 2z2


+ VcosQt


(2-1)


where U represents the maximum dc potential and V the maximum rf potential (0-

peak) applied between the ring and endcap electrodes, is the angular frequency

of the rf drive potential, t is time, x, y, and z are the directions inside the ion trap.

The x,y plane leads out from the center of the trap to the nearest ring surface (e.g.,


radius)


z axis


leads out from


the center


of the


to the


endcap


electrodes.


parameter


internal


radius


electrode.


Conventional ion traps use a geometry in which r0-


- 2zoz,


where


is the closest


distance


between


endcaps,


although


commercial


traps


actually


stretched in the z-axis from the conventional geometry.44


The oscillation of the rf


potential creates a dynamic electronic field inside the trap where ions are forced in

one direction and then forced in the opposite direction in both the radial and axial


planes.


The force acting upon an ion of mass m and charge


is given by


1
2









= -ze *


(2-2)


where


V represents (d/dx +


d/dy


+ d/dz) and A


is the acceleration vector.


This


relationship leads to representations of the forces acting upon the ion in the x, y, and

z and given by


m
m x+(U
e


-r y+(U
e


+ VcosQt)-- = 0
ro


(2-3)


+ Vcos Qt)- = 0
ro


(2-4)


m
- z
Se


-(u


2z
+ Vcos tt)
ro


(2-5)


None of these expressions contain cross-terms between, y, and z, so the ion motion


may be resolved into each of the respective perpendicular coordinates.


The x and


y components are identical and may be treated independently if angular momentum


is ignored about the z axis.


Due to the cylindrical symmetry, the x and y components


are combined to give the radial component r using x2


The z component of motion is out of phase by half a cycle with respect to the

x and y motions (therefore the minus sign); the factor of two arises because of the

asymmetry of the device caused by the need to observe the Laplace condition V2h


= 0 when applied to equation 2-1.


These equations are all examples of the Mathieu


v4 =


+ y2 = r2






25

equation (the mathematics of ion stability within the trap follow the Mathieu second-

order differential equation), which has the generalized form


+ (au,
du2


+ 2q, cos2O)u


(2-6)


where


= X, y,


z (2-7)


= Qt/2


(2-8)


8eU
mro 2


(2-9)


= -2q


= -2qy


4eV
mr22
mro 2


(2-10)


and where ( is a dimensionless quantity.


Thus the equations 2-7 to 2-10 relate the


Mathieu parameters az and qz to the experimental variables and also to the time


variables il and t.


The a and q parameters are very important to the operation of


= -2a


= -2a






26

diagram shown in Figure 2-1 defines the primary region within which the axial and


radial components of motion are stable,


the region of overlap indicating the az, qz


coordinates corresponding to those ions that are held in the trap.


The plot is called


the stability diagram.

Lines drawn across the stability region in Figure 2-1 are called iso-/f lines, and

describe the detailed trajectories of the ions at that point; the boundaries of the


diagram correspond to fr,,Iz


= 0 and 3,,zp


= 1, with the boundary 3z,


= 1 being that


at which mass-selective instability is normally achieved during a mass scan. /3 values,


as well as az, q2 coordinates and ion secular frequencies,


will be discussed in later


chapters in reference to CID conditions, axial modulation, and black holes.


Nitrogen Laser and Interface


diagram


laser


desorption


ionization


quadruple


mass


spectrometer is shown in Figure


The beam from a pulsed nitrogen laser (Laser


Science 337ND) is directed through a UV-grade focussing lens and a low distortion


quartz window mounted in a vacuum flange.


The focussing lens (Melles-Griot) has


a focal length of 25.4 cm.


Neutral density filters were


used to control the laser


irradiance.


Average beam spot size, as measured by the hole created by the laser on


the sample surface, was approximately 0.04 mm2


nitrogen laser


operates at a


wavelength of 337.1


nm with


a spectral


bandwidth of 0.2 nm.


It has a 3 ns pulse width (FWHM) with an average jitter of















0.20


0.00


-0.20


-0.40


-0.60


-0.80 ---
0.00


Pr 0o


-Pz-1


=0


- Pr


0.25 0.50 0.75 1.00 1.25


Figure 2-1.


Plot of stability diagram defined by iso-a3 lines. Ions which have az, qz
coordinates within the bound region have stable trajectories.


a
Z


4z

















hv


Quartz
window


Ion
tra


I


p


Sa
----*-- ---


:tron multiplier


nple probe


Vacuum manifold

Probe lock


Figure 2-2.


Schematic of laser desorption ionization/quadrupole ion trap mass


spectrometer.


Nitrogen
Laser


II


Fl pr





29

peak power of 85 kW, more than sufficient to produce large numbers of ions from


all types of samples.


mrad.


The beam area is 40 mm2 with a beam divergence of


Pulse-to-pulse stability is given to be t4% at a 10 Hz repetition rate.


<0.3

With


the laser focused to a spot size of around 0.13 mm2, the irradiance is calculated to


be approximately 6 x 10


7 W/cm2 (this value does not include any constant attenuating


factors such as the flange window).


typical applications, however,


the laser is


attenuated with neutral density filters to transmit only 5-10% of the energy, thus


giving calculated irradiances around 3


x 106


106 W/cm2.


These values are near


the optimum laser irradiances (1


106 W/cm2) for the ion production threshold for


MALDI.

To enter the ion trap, the laser beam passes through a modified ring electrode


via a 3.18 mm diameter hole as shown in Figure 2-3.


The endcap electrodes were


not modified.


A sample is introduced into the vacuum manifold on


the tip of a


stainless steel probe through a flange-mounted vacuum probe lock.


The tip of the


sample probe is then inserted into the ion trap through a 3.18 mm diameter hole

opposite the laser beam entry hole and positioned along the ring electrode surface.


The instrument can be used in its conventional El


and CI modes with


laser


desorption interface hardware in place.

Various materials have been used for the probe tip including stainless steel,


copper, aluminum, glass, graphite, and Macor.


Probe tips have 0.10 inch diameters


and approximately 1 inch lengths.


They are held in


Teflon holders that are press fit
















































I







/%/
I


6/;
/:-


-I


- ---%l. .






-+ -









: :: : 1


j-



:::t:::~::" A-
'* '<; ":: "ih .? .


% i ^ +- ^ -


ZZ \
vr :











: ~ t _:
.-
I,






31

provide the electrical insulation between the rf ring and the operator-handled probe.

The distance between the ring inner surface and the outside of the ring hole is 0.433

inches, and a mark indicating this distance on the probe tip is used as a guide to

achieve a proper and constant insertion depth of the probe tip.

In order to aim the laser for sampling, fluorescence from the Macor probe tip


surface is used to see the ultraviolet beam spot.

mirror and adjustable mirror mount. Typically,


The laser beam is aimed using a


the laser spot is placed to one side


of the probe tip surface and the tip is rotated so that several sampling spots can be

obtained from a single sample residue.


Experimental Setup


Scan functions and Timing


As stated in Chapter 1 for ion trap operation, scan functions represent the rf

voltage amplitude over time for comparison to various events that occur during a


single analysis step.


The various events include, in sequential order,


1) electronics


initialization trigger, 2) rf voltage warmup period, 3) laser trigger and ion trapping,

4) ion cooling and storage period, 5) multiplier warmup period, 6) rf voltage ramp


and data


acquisition


scan,


voltages off period


emptying


of ions


between scans.


A scan function showing the important events is given in Figure 2-4.


This scan function shows that a single laser pulse occurs during the trap and store


I. n T. r~ C w ,n rli rn I. .4. a ar- 1rr car n.. ln in en+ i-c. 'ii I i + ;r


*l^ s-i tc Fjkn i torI T^


r" x-^4 inrt ^ IA r


c: mn























RF
amplitude


Laser
trigger


time
- pre-ionization
- ion formation & trapping
- ion storage
- ion detection


Figure 2-4.


Drawing of rf scan function showing sequence of important events.





33

The proper timing of the laser pulse with the scan function is important to


insure trapping of ions.


Repetition rate of the scan function must not exceed the


maximum pulse rate of the nitrogen laser (20 Hz).


Typically, the scan function has


a length of 100 ms that corresponds to the laser's optimum operating repetition rate


of 10 Hz.


The trigger that is sent to the laser comes from the ITMS electronics.


Computer control of the laser trigger was established through software written in our

laboratory.


Trapping Radiofrequency Phase Synchronization Circuit


order to synchronize the


laser pulse with


rf phase angle, a


trigger


circuit was developed which utilized the clock pulse of the rf generator circuit and


a TITL pulse controlled by the software.


uis to insure laser triggering. As shown in Figure


The scan table in the software is set to 100


the 1.1 MHz 15 V square wave


(rf clock signal) of the rf generator and the asynchronous laser trigger (TTL pulse)

were passed through a buffer and sent to the inputs of a dual edge-triggered flip flop


(Motorola MC14013).


The 1.1 MHz signal is sent to the clock input of the flip-flop


and the


TTL pulse is sent to the data input.


The output signal of the flip flop circuit


(Q1) is similar to the laser trigger signal except that the rising edge (Lo Hi) of


the output pulse is synchronized with the rising edge of the rf clock signal.


be used to trigger the laser synchronously with the rf phase.


Q1 could


Instead, Q1 is used to


trigger a pair of monostables so that the rf phase angle during the laser pulse may


- t)







34




C]C
cn C 1
a U2^ ^
o -

S0 "0C.

I
ct ,


l~(d
^^ v




in 0
+ d -
/C.

d-)
_ C) H

C ) a I


v-^ Q) Id .^ ^
I----------^--- & ^ .--



0 C

) k
c ,,




0 r


r s
^~~C '3^
&J) L^^ P "
*r^ ^ I
r"F\v






35

Q1 and the complementary output (Q2), or low pulse, of the signal is then sent to


the second monostable.


The Q2 signal is varied in length using a calibrated 10-turn


variable


10 klZ precision resistor (see Appendix A for calibration data).


The rising


edge of Q2 then triggers the second monostable that creates a high pulse (Q3) which

is synchronized with the rf phase and variable over 1/2 rf cycles.

A comparison is made between the laser pulse and the rf voltage on a digital


oscilloscope


using the amplified response from a


silicon


photodiode (TL032)


which is impinged upon by part of the laser pulse and the rf voltage (measured at


the vacuum manifold feedthrough).


The various signals described above are shown


in Figure 2-6 where trace (a) is the initial software produced trigger, (b) is the rf

clock, (c) is the rf synchronous trigger, (d) is the delay pulse with variable width v,

(e) is the rf synchronous laser trigger, (f) is the photodiode response to the laser


pulse, and (g) is the actual trapping rf field.


It is important to note that at all times


the low mass cutoff was maintained on the ITMS so that the rf voltage was measured

at a safe level.

The turn settings of the precision resistor were calibrated to the pulse width


(v) of the first monostable and rf phase angle.


Values obtained from this calibration


were used to increment the laser pulse over several rf phase angles.


The relationship


between the pot number setting of the variable resistor and the rf phase angle is


shown in Figure





















Asynchronous
trigger


clock


rf synchronous
trigger


Monostable


Monostable


Photodiode


trapping


500 1000


1500


me


Figure 2-6.


Timing dplot for rf phase synchronization experiment.
trace descriptions.


See text for


-~ur" '~-r'F


_f















600


500


400


300


200


0 1 2 3 4 5 6 7


pot


Figure


number


Plot of rf phase angle versus variable resistor position settings used for
rf phase angle study.








Sample Preparation


Compounds of


interest were


dissolved


distilled


water


or methanol


various concentrations from a


few micromolar to a


few millimolar.


matrix


solutions for MALDI were always dissolved in methanol with 0.1% trifluoroacetic


acid.


matrix


compounds


used


were


sinapinic


(MW


224),


dihydroxybenzoic acid (MW 154), 2,4-dinitroaniline (MW 183), coumarin-120 (MW


175),


nicotinic acid


(MW


123).


Structures and


molecular weights of these


matrices are given in Figure


Note that the molecular weights are all lower than


225 u and therefore do not interfere in the typically mass range of interest (>350 u)


As described in Chapter


1, the matrix compounds should absorb efficiently at the


operating


wavelength.


absorption


spectrum


sinapinic acid


(3,5-


dimethoxy-4-hydroxycinnaminic acid) is given in Figure 2-9.


The molar absorptivity


at 337 nm was found to be


14391 M'cm1


The molar absorptivities at 337 nm for


the other matrices are similar (coumarin- 120, 14519 M-'cm ; 2,4-dinitroaniline, 11971


M-'cm-1)


except for nicotinic acid


which


is much


lower


(29.0


M-'cm-1).


For the


analyte compound studied


following MALDI


experiments,


spiperone,


molar absorptivity at 337 nm is


74.0 M'cm'


For laser desorption-ionization,


2 to 4 pL of sample solution is deposited on


a clean probe tip and air dried before placing in vacuum.


For MALDI experiments,


pL of the


analyte solution is deposited on a clean probe tip followed by


2 pL of


matrix soliutionn


The samole-matrix mixture is also air dried before olacinQ in the


Ill L1 .A tLU ,J /1


U ELA *











CH30


O
II
CH = CH-C-OH


CH30


a) MW 224


NO2


N02


b) MW 154


C H3


H2N


c) MW 183


d) MW 175


O
C-OH



N

e) MW 123


Figure 2-8.


Structures and molecular weights of UV absorbing matrix compounds
used for MALDI experiments: a) sinapinic acid, b) 2.5-




































































f g g 9 a 9 a
nvt*


09E









Oa



O









OBZ
0a













022


3JMVRHIOSv






41
homogeneous mixture of analyte and matrix on the probe tip, such that the analyte


molecules are individually surrounded by matrix molecules.


The equal volumes and


1:1000


concentrations


analyte


to matrix ratio


that works


best.


When


obtaining the first few mass spectra from a new sample, the laser transmission usually

must be adjusted with the neutral density filters.












CHAPTER 3
FUNDAMENTAL STUDIES OF LASER DESORPTION
IONIZATION/QUADRUPOLE ION TRAP MASS SPECTROMETRY USING
AN INTERNAL CONFIGURATION


The laser desorption/ion trap system was modified and important instrumental


parameters were studied in order to better understand the trapping process.


In the


instrumental configuration presented here, the trapping process for laser-desorbed

ions resembles that of ion injection systems which have been previously demonstrated


on the


ion trap.45'46


contrast to


the conventional ionization methods such


electron ionization which takes place within the ion trap volume, the laser-desorbed


these


experiments are


formed


at the


surface


of the


electrode,


therefore at both the physical and electrical stability boundaries for the ion trap.

Ions formed at the position of the ring surface are then "injected" into the ion trap

with the kinetic energy imparted by the laser desorption ionization process.

In theory, ions injected into the trap will eventually be ejected from the trap


boundary because of the


initial kinetic energy which allowed


them


to enter the


radiofrequency trapping field.4


The millitorr of helium buffer gas commonly used


in the trap to improve sensitivity and mass resolution can also serve to collisionally


cool injected ions,21 as well as ions formed at the ring electrode.


Trapping would be


most favorable for ions of low kinetic energy and divergent from the symmetrical


Sr Ainfrnnu10nn ri xi onr*? iln f nlnn n thblh raiallir rnnn^rtc tho rinn alc'rtrno ciirfar'nc 2


T/-\no






43

formed upon laser ionization at a solid surface would no doubt contain a wide range

of injection angles to satisfy the latter property, but the average kinetic energy of


laser-desorbed ions typically would be too high for efficient trapping.


Therefore


instrumental parameters during ionization which affect trapping would need to be


studied


their


effect


on trapping


efficiency.


more


important


these


parameters, radiofrequency phase angle,


radiofrequency amplitude,


and buffer gas


pressure, are the focus of this chapter.

To perform these experiments, samples were needed which would provide an


abundant and steady source of ions over hundreds of laser pulses.


Samples which


could be shaped in the form of the probe tip can be ablated by the laser with little


effect on the ion signal for very many shots.


Rods of graphite, stainless steel, and


copper were used in these studies without any other preparation.


Effect of Radiofrequency Phase Angle on


Trapping Efficiency


While rf amplitude and buffer gas pressure have been studied previously for


ion injection experiments,


the phase of the sinusoidal trapping voltage during ion


injection has not been investigated; however, the effect of the rf phase at the instant

of laser desorption was evaluated experimentally in our laboratory because of the


possibility


have


a significant effect


on trapping


efficiency for


produced at the ring electrode surface.


Previously,


Kishore


Ghosh47


Ghosh


et al.48


calculated


phase








were


injected


asymptotically


optimum


initial


phase


(occurring


after


specified delay before the rf field was turned on) was approximately


1800.


They


calculated the optimum initial phase as that which the ions were stable for hundreds


of rf cycles.


O and Schuessler also showed a theoretical dependency of injected ions


on the rf phase.49


Their calculations demonstrated that the trapping of externally


produced ions (without buffer gas) was possible for a very narrow range of injection

phases and that the optimum rf phase angle was dependent on the velocity of the


In an analogous study,


Pedder


et al.


used an ion


trap simulation program,


Hyperion,


to study


rf phase


angle effect on


kinetic


energy


electrons


entering the trap through the endcap electrode and thus on electron ionization.50

In preliminary studies in our laboratory (and in reports by others), the laser


trigger was not synchronized with the rf phase.

varied greatly from pulse to pulse. Due to the s


It was observed that the ion signal


hort laser desorption event (the laser


pulse width is 3 ns in our system compared to the 909 ns period of the 1.1 MHz rf),


the ion trapping efficiency might be expected to be dependent on the rf phase.


rf phase synchronous triggering circuit described in Chapter 2 can vary the delay time


between the rf trigger and the laser trigger from 200 to

study of laser pulse triggering over an entire rf period.


1200 ns, thus enabling the

In order to perform this


experiment, a constant source of sample ions is needed for a large number of laser


pulses.


Toward this end, a graphite rod was used as the probe tip and ions were


produced from its surface.


A mass spectrum of graphite from an average of the first






45

set of 10 laser pulses in the rf phase angle data discussed below is given in Figure

3-1.


During the experiments,


ten laser desorption ionization mass spectra were


taken at each


of 29 different points over approximately


1/2 rf cycles,


each point


accounting for a phase delay increment of 200 (50 ns).


Each set of 10 spectra was


averaged to obtain the plots of the C3+ ion intensity with respect to rf phase angle,


as seen in Figure 3-2a and 3-2b for two separate runs.


These plots show patterns of


the C3' intensity (displayed as averaged data points and error bars) compared to the


pattern of the rf voltage sine wave (displayed by the solid lines).


The ion current


patterns are seen to decay over the entire period of 1/z rf cycles which is attributable

to a small change in focal length as the sample is ablated from the surface, or to


change in surface composition.


Continued desorption of ions from


graphite


surface, however, showed no overall change in the ion compositions seen in the mass


spectra.


However, cyclical patterns can be seen which are comparable to the sine


wave except that they have slower rises and quicker falls in intensity, and are shifted


in phase from the rf field.


For the plot shown in Figure 3-2a, the first maximum is


seen to occur at an rf phase angle of 123, shifted from the rf sine wave maximum


at 900.


The first minimum is seen to occur at an rf phase angle of 2380, shifted from


the sine wave minimum at 2700.

The C3+ ion is more intense than the other graphite ions detected using laser


desorption ionization (as seen in the mass spectrum in Figure 3-1).


However, it is













































+
in








+-
(9


-In


I I !1


I- 1r I I


suauU


'I I I I


I II I T









300000





200000


100000


-50 0 50 100 150 200
rf phase ai


300000





>"200000

0)
I'
C

100000





0


n


250 300 350 400 450 500
gle (degrees)


~II


~II


100
pha


angle


I I I I I I I I I I I I I I I 1 I I I I I
00 350 400 450 500 550
(degrees)


Figure 3-2.


C3+ intensity and trapping rf amplitude vs. rf phase angle: a) Run
b) Run 2. Error bars designate standard error of the mean. Avera


of 10 scans (1 microscan each scan).


ige


E= E


E


1








a similar shape to that of the C3+


, although each is shifted in phase angle by different


amounts, as shown in Figures 3-3 and 3-4, respectively, for two runs.


of these ions are also more sporadic due


The patterns


to their lower signal-to-noise ratio.


comparison of the ions at different masses brings up the question of whether or not

this rf phase dependence of the ion trapping efficiency is also mass dependent, thus

complicating the matter of determining the optimum point of laser desorption in the

rf phase of the ion trap for any given sample.


Effect of Radiofreauencv Amplitude on


Trapping Efficiency


Ions stored in the ion trap are subject to a low mass cutoff; ions of m/z lower


than this value have unstable trajectories,


trajectories in the trap.


while ions of higher m/z will have stable


The low mass cutoff level is directly proportional to the rf


voltage applied to the ring electrode.


The effect of low mass cutoff on ion trapping


efficiency during trapping of laser-desorbed ions was studied.


The intensity of laser-


desorbed 56Fe+ ions versus the low mass cutoff during trapping and storage is shown


in Figure


As seen in the plot, the ion intensity varies with the low mass cutoff


level, with a maximum occurring at 39.4 u and complete loss of ion storage above 56

u. The maximum ion intensity at the low mass cutoff of 39.4 u is nearly two times


greater than that measured at a low mass cutoff of 10 u.

minimum which occurs at a low mass cutoff of 43.2 u.


Also seen in the plot is a

When the data is plotted


versus the qz for m/z 56, we can see that this minimum occurs at a q, equal to 0.69,










40000




30000




^ 20000
C
C


10000




0




60000


50000


40000


* -
0 30000
C<
-4-J
C
- 20000


10000


C'


U


- EF


S I I I i i i i i i i I t i i J i 1 I t 1 I I T r I I I I J | i t r | I I I I | i T i | I I I I I
-50 0 50 100 150 200 250 300 350 400 450 500 550
rf phase angle (degrees)




-I


Si i i i i i i i i i i i i i i i i i i i IT I I I i I i i i i i i I i [
-50 0 50 100 150 200 250 300 350 400 450 500 550
rf phase angle (degrees)


Figure 3-3.


C4+ intensity and trapping rf amplitude vs. rf phase angle: a) Run 1,
b) Run 2. Error bars designate standard error of the mean. Average
of 10 scans (1 microscan each scan).


itfr E3=


XIi'


"E35q:5EE










30000






20000


10000


IT -I


rr TT ; T i 1 i i i i i i I I I I I I. i i 1 1 i I I 1
0 50 100 150 200 250 300 350 400 450 500 550
rf phase angle (degrees)


40000




30000




20000




10000




0


III-


TE F T--


-50 0 50
rf


100 150
phase


IT-T V T 1 1 I I
200 250
angle


661 S S 466 I i I T
00 350 400 450 500 550
(degrees)


Figure 3-4.


Cs5 intensity and trapping rf amplitude vs. rf phase angle: a) Run 1,
b) Run 2. Error bars designate standard error of the mean. Average
of 10 scans (1 microscan each scan).


E3-


^E-fasasE


&

















40000





30000


000


0000


20 30 40 50


ow


ma


utoff


Figure 3-5.


intensity vs.


low mass cutoff.


1 scan each data point (10


microscans each scan).


56Fe+


















40000





30000


20000


10000





0-
0.1


0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9


qz,56+


Figure 3-6.


30Fet intensity vs. qz 1
microscans each scan).


For m/z 56.


scan


each


data


point (10


v'









from


octapolar


distortions


quadrupolar


trapping


field


stability


coordinate (az,


= 0,


0.69) corresponding to a fz


= 1/2;


it has been described


theoretically by


Wang and Franzen,5152 has been seen in chemical ionization and


CID experiments as well, and is currently the focus of


research in our laboratory


others.


53.54.55


Also


shown for


comparison


Figures 3-7a


and 3-7b


are the


intensity versus qz plots for copper-63 and copper-65 ions, respectively.


All three


plots show that the ions follow a similar pattern for trapping efficiency for their


respective q,


values during the


trapping and storage periods.


The maximum ion


intensity for all three ions was found to occur at a setting of approximately q,


= 0.65.


In Figures 3-7a and 3-7b for the copper isotope ions, a second minimum is seen to


occur


which


was


not clearly


evident


data.


This


second


minimum


corresponds to a black hole arising from a hexapolar field distortion at /z


= 2/3.


In the data presented here, the m/z of the ions are quite close, so the effects


of intensity versus qz are similar and predictable.


The effect of rf amplitude on


trapping efficiency has been found to be different for laser-desorbed ions of higher


This effect, however, is possibly not a storage effect but a trapping effect.


trapping (or "acceptance") volume at high rf voltages is very small for ions that are


injected,


such as ions formed at the ring surface.


A similar reason was given to


support the data


presented


a report by


Louris et al.45


Higher mass ions will


undergo fewer collisions than lower mass ions because of their lower velocities.


m/z ions, however, need more collisions to remove


High


kinetic energy because of the
























250




0-
0.1


0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0


0 5 0 I
0.5 0.6


qz,65


Figure 3-7.


Copper isotope ion intensities versus q2: a) 63Cu +
m/z 63, b) 65Cu+ intensity vs. qz for m/z 65. 1 scan
microscans each data scan).


intensity vs.


q, for


each data point (5


qz,63'






55

injection, the higher mass ions will be increasingly unlikely to reach the center of the


trap and become stabilized.


The combination of these effects probably leads to the


observed decrease in trapping efficiency.


In fact,


very few ions are detected when


low mass cutoff levels are above


150 to 200 u in our experiments for detecting ions


between 350


u and


Although laser-desorbed


ions of m/z up


to several


thousands, have been detected in the ion trap,


the low mass cutoff levels were set


low and the ions were detected using axial modulation for mass range extension.57


Effect of Buffer Gas Pressure on Trapping Efficiency


The experimental parameter with most significant effect on trapping efficiency


is buffer gas pressure.


Buffer gas collisionally cools the ions so that they may be


trapped once they enter the ion trap volume.


After ion trapping has occurred, buffer


gas continues to play an important role during ion storage by relaxing the ion motion


toward the center of the trap.


This process is used to achieve better resolution and


sensitivity, as was first discussed by Stafford et al.z2


In the case for helium buffer gas,


Glish et al. showed increased intensity for quaternary ammonium salt ions produced


by laser desorption at the ring surface with an increase in pressure."


Figure 3-8


shows the effect helium pressure has on the intensity of laser-desorbed iron ions of


In this plot, the 56Fe+


intensity increases rapidly from 0.0025 mtorr to


mtorr.


No ions were detected at the base pressure (8


x 10s torr).


To improve the


cooling ability of the buffer gas, argon was used instead; the relationship between

















800





600


O


400


200


0.0001


1 I I I 1 1 1I


0.001


pre


I I-I I lllI


ure


( m


I I I I 1111I
1
itorr)


I I I 1 1 11n


Figure 3-8.


56Fe+ intensity vs. He buffer gas pressure.


Error bars designate


standard


error


mean.


scans


each


data


point


microscan each scan).


I Ir I TTI Ill
















200


000



800


600


400



200


I I I 1 1 1 11


1 I I I II IT


T 1 I I III I


I I I I I I


I I I I I Ill


0.000


0.001


pre


0.01


(m


rr)


Figure 3-9.


intensity


Ar buffer gas pressure.


Error bars designate


standard


error


mean.


scans


each


data


point


microscan each scan).


r3


56Fe+






58

intensity decreases from 0.0005 mtorr to 0.5 mtorr with a rapid decline beginning at


pressure


greater than 0.01


mtorr.


At the


lower pressures,


argon,


with


its much


heavier mass than helium, removes kinetic energy from the ions more efficiently per

collision than helium. A combination of scattering losses and ion peak reduction due

to peak spreading (loss of mass resolution) explains the signal decay.

The observations made of the buffer gas pressure effect on ions produced at

the ring surface are similar to those seen in a study of injection of gold ions through


the endcap electrode from an external source.45


Helium, neon, argon and xenon


buffer


gases


promoted


ion trapping.


Neon


was


found


similar


to helium


promoting trapping, but argon and xenon yielded signals only 10-20% as intense as


helium.


Xenon produced broad peaks in the mass spectrum whereas the other three


gases produced peaks of unit resolution.

The results obtained using a copper sample show the expected increase in


and 65Cu+ intensity with increasing helium pressure, as seen in Figures 3-10a


and 3-10b, respectively.


It is important to note, however, that the isotope ratio in


these


measurements


deviate


from


expected


natural


abundance


ratio


(63Cu/65Cu).


Figure 3-11


shows a


plot of 63Cu +5Cu +


intensity ratio.


ratio


increases from approximately 0.4 to 2.1 from 0.0025 mtorr to 2.5 mtorr helium.


appearance of the incorrect isotope ratio at the lower pressures is not completely


understood.


While ion abundance is maintained below the onset of space charge


conditions, the apparent loss of 63Cu


(or gain of 65Cu+) is unexpected.


No other


63 Cu +








200




150

0

x 100


0.0001
0.0001


0.001 0.01 0.1 1


pressure


(millitorr)


I I I I I00
0.001


i rr n| I i II
0.01 0.1
sure (millito


I I I TTT10
10


Figure 3-10.


Copper isotope ion intensities vs. h
intensity vs. He pressure, b) 65Cu +


helium buffer gas pressure: a) 63Cu+


intensities vs. He pressure.


Error


bars designate standard error of the mean.
(1 micronscan each scank


50 scans each data point


























O
(D


0.0001


1 I I 111 11


0.001


I I 1 111 1


pre


I I IT I I1


ure


0.
(m


I I I I E II i
1
itorr)


I I I I IIIh


Figure 3-11.


63Cu/65Cu +


isotope ratio vs.


He buffer gas pressure.


50 scans


each data point (1 microscan each scan).






61

And it is doubtful that the incorrect ratio is due to the sample or ionization process.


In the use of


a stainless steel probe tip to make iron ions, the calculated isotope


ratios for observed iron isotopes, chromium isotopes, and nickel isotopes were all

found to be correct.


use of


argon


buffer


was


studied


with


copper


isotopes.


Figures 3-12a and 3-12b show the intensity versus argon pressure plot for 63Cu +


The data are similar to those seen for 56Fe+


in Figure 3-9 in that there is a


decrease


ion intensity at the


high end


of the


pressure


range.


In the plots of


copper intensities versus argon pressure, however, a reduction in ion intensity is also

seen at the high end of the pressure range creating maximum intensities at 0.025

mtorr for both isotopes.

A plot of the copper isotope intensity ratio versus argon pressure gives the


results in Figure 3-13.


The copper isotope ratio varied 0.6 to 1.3 over the measured


argon pressure range from 0.0005 mtorr to 0.5 mtorr.


Although the isotope ratio


increased


with


increasing


argon


pressure,


ratio


never


reached


natural


abundance ratio of


Conclusions


Efficient trapping of ions produced by laser desorption ionization at the ring

electrode surface requires careful control of rf phase angle, rf amplitude, and buffer


gas pressure.


The trapping efficiency of laser-desorbed graphite ions was found to


65"Cu+































0 -
0.000


TVTfWTT7Tf


0


TI I 10.01
0.01


pressure


0.1
mil it


orr)


I I rTT
10


0001
0.0001


r r I I rrrTJ
0.001
Ar p


T i i i ii
0.01
pressure


-I T r I i T
0.1
(millit


1
orr)


I IT 1 1l
10


Figure 3-12.


Copper isotope
63Cu+ intensity


intensities versus argon


vs. Ar pressure,


65C +


buffer gas pressure:


intensities vs. Ar pressure.


Error bars designate standard error of the mean.
noint (1 microscan each scan\.


50 scans each data



























(9
0


/1
Co
C


c-I-
+
Li
'
(0

C


0.0 t-
0.0001


I I I ti II I
0.001


I I I I I II 1
0.01


pressure


III I I I I I
0.1
mil


I 1 I 1 I I 1 E


I I I II II


itorr)


Figure 3-13.


63Cu /65Cu+ isotope ratio vs. Ar buffer gas pressure.








pattern was seen in the intensity of C3+


which is comparable to the rf voltage sine


wave except that it has a slower rise and quicker fall in intensity, and is shifted in


phase from the rf field.


+ and Ci+


signals were also found to be dependent on rf


phase angle with different patterns over the measured rf phase angle range.


results from the different ions suggest that the rf phase dependence of the trapping

efficiency is also mass dependent, thus complicating the matter of determining the


optimum


point of laser desorption


the rf phase of the


for any given


sample.

The rf voltage was found to have a similar effect on the intensities of 56Fe+


and 65Cu


Also, minima in the ion intensities demonstrated the appearance


of black holes in the ion trapping efficiency when the data was plotted versus qz.

The maximum intensities measured were approximately twice as large as the lowest


measured low mass cutoff value and found to occur at coinciding qz


values of 0.65.


Buffer gas pressure was found to have the most significant effect on trapping


efficiency of all of the instrumental parameters studied.


No ions were detected at


the base pressure.


0.005 mtorr to 5 mtorr.


With helium buffer gas,


With argon buffer gas,


intensity increased rapidly from


56Fe+ intensity decreased from 0.0005


mtorr to 0.5 mtorr, most likely caused by a combination of scattering losses and loss


of resolution (peak spreading and height reduction).


Results obtained here were


similar to those in other reports using laser-desorbed ions and ion injection systems

on the quadrupole ion trap.


63 Cu+


56Fe+






65

Copper ion results were similar to those obtained using iron for helium and


argon buffer gas.


The isotope ratio of the copper ions, however, was found to vary


over the measured buffer gas ranges and deviate from the natural abundance ratio.

The observations made here of the effects of these instrumental parameters

are in general more similar to those that would be obtained performing ion injection

from an external source than ionization inside the ion volume (e.g. conventional El


and CI on the trap).


These observations suggests that in order to obtain the best


trapping


efficiency using


this configuration,


effect of


phase


angle


amplitude need to be determined on the sample of interest; the buffer gas pressure

in the case of helium should be set to its highest possible pressure (e.g. before the


onset of space charge or arcing).


At optimum rf phase angle and rf amplitude, the


intensities


this study


doubled


both


cases.


helium


buffer


pressure study,


the 56Fe+


intensity varied by a factor of 4 from the lowest helium


pressure to the highest pressure.


For 63CU+


and 65Cu+


, the intensities varied by a


factor


approximately


over


same


pressure


range


helium,


respectively.












CHAPTER 4
LASER DESORPTION IONIZATION/QUADRUPOLE ION TRAP MASS
SPECTROMETRY USING AN INTERNAL SOURCE CONFIGURATION


Laser Desorotion Ionization


Whereas the previous chapter dealt with the fundamentals of trapping laser-

desorbed ions formed at the ring electrode surface, this chapter presents research

performed to evaluate the combination of the capabilities of laser desorption with


the capabilities of the quadrupole ion trap mass spectrometer.


An abundant number


of ions over many laser pulses can be obtained with laser desorption ionization on


organic salts.


Therefore, to initially evaluate tandem mass spectrometry of laser-


desorbed


trap,


laser


desorption


ionization


was


performed


trimethylphenylammonium bromide.


MS/MS of Trimethylphenylammonium Bromide


Using


a five


microliter


syringe,


solution


trimethylphenylammonium bromide (TMPA) in methanol was placed on the stainless


steel probe tip and dried in air.


The probe tip was placed inside the ring electrode


and the sample surface positioned flush with the ring surface.


The results obtained


from laser desorption ionization (LDI) of TMPA are given in Figures 4-1 and 4-2.
T -". IP A f l A11----------- ...... .










25000


20000


>.15000

C
44 10000
C


5000


0


+ K+
Noa 39
25


(M-CH3)7
121


30 40 50 60 70 80 90 100 110 120 1.30
miz


1000


800


140 150


121


30 40 50 60 70 80 90
m/z


100 110 120 130 140 150


Laser desorption ionization mass spectra of trimethylphenylammonium
bromide, a) full mass scan, b) single ion isolation of the 121' fragment


Figure 4-1.


























E


O 0
0 0








from 20 to 640 u is shown.


In this spectrum, the displayed data from m/z 20 to 150


show two large and unresolved peaks at the low mass end due to highly abundant


sodium and potassium ions.


These ions are


typical LDI


products observed most


often in the mass spectra of unpurified organic compounds and biological samples.

We also see a set of unresolved peaks between 50 and 60 u due to the stainless steel


probe tip background.

chromium, and nickel.

fully resolved for study.


TMPA sample.


The peaks are comprised of the various isotopes of iron,

Under different conditions for LDI, each of these ions were

The interest here, however, is in any ions obtained from the


There is an extremely small peak at m/z 136 for the molecular ion,


but the ion used for the following analysis


the one found at m/z


121 which is a


fragment of TMPA.


This ion is formed by the


loss of a methyl group from the


trimethylphenylammonium cation (a preformed ion).


Using a two-step ion isolation


routine developed in our


lab by Nathan


Yates,


TMPA fragment ion


can be


isolated by ejecting all of the other ions from the trap.

Performing this isolation step and taking a mass scan gives the mass spectrum


shown in Figure 4-lb.


In this spectrum, the only major peak is that for the


TMPA


fragment ion


at m/z


few smaller


peaks at m/z


are due


incomplete ion isolation, but they are less than


of the intensity of m/z 121 ion.


The efficiency of the single ion isolation displayed in Figures la and lb is only about

33%, but it must be noted that isolation step involved ejecting more of m/z 121 ion


that was really necessary.


Typically, isolation is performed to maintain as much of






70

m/z 120 and 122) having a relative intensity much greater than in this experiment,


usually up to


10%.


collision-induced


dissociation


(CID)


was


then


added


to the


scan


function after the ion isolation step.


During the CID process,


the m/z


121 ion is


resonantly excited causing the ions to obtain larger and more energetic trajectories.

The additional collisions with helium buffer gas induce the ions to fragment and the


ions produced can then be stored if their trajectories are stable in the ion trap.


mass scan


of these


"daughter"


ions are shown in Figure 4-2.


This daughter ion


spectrum shows several peaks which can be identified by the amount of mass lost.

The peaks noted in the figure are m/z 118 which has the composition C(CH3)N-+,


m/z 103 which is CN-#+


, m/z 93 which is H2N- +


, and m/z 91 which is N-4+


It is important to note in this mass spectrum, and in all of the spectra shown

later, that the method of assigning peaks to the mass axis is different on the ITMS


from most other mass spectrometers.


With the ITMS, the peak for m/z 118 should


be found to the right of the hatch mark signifying 118 and to the left of the hatch


mark signifying 119.


The center of the peak, therefore, is placed in the center of the


space between these two hatch marks.


This convention is used for all of the data


obtained from the ITMS and ITS40 instruments and will be noticeable where the

hatch marks in the mass spectra represent a single mass unit.


combination


factors


determines


efficiency


with


which


CID


performed in the ion trap, an efficiency usually described as the conversion of parent






71

buffer gas pressure, resonant excitation frequency and voltage, and low mass cutoff),


most


important


determinant


compound


itself


ability


to be


fragmented with the obtainable collision energies in the trap.


The CID efficiency for


the mass spectrum shown in Figure 4-2 was not optimized, but yet the appearance


of several different fragment ions was the desired result of this experiment.


These


data show that tandem mass spectrometry can be performed on laser-desorbed ions

in the ion trap, and the results obtained can be used to identify ions which are not

the most abundant products of the laser desorption process.


Positive and Negative Ion Spectra


As stated in the introduction, laser desorption ionization can produce a wide


range of particles including both positive and negative ions.


The quadrupole ion trap


store


both


positive


negative


species;


method


detection


presumably allows only one polarity to be detected at a time.


Figures 4-3a and 4-3b


show the positive and negative ion spectra for sodium iodide, respectively.


In order


to obtain the positive ion spectrum, a negative voltage (-4 kV) was applied to the


conversion dynode to attract positive ion species toward it.


we see peaks at m/z


In this mass spectrum,


, shifted on the mass axis due to space charge), and


sodium iodide cluster peaks at m/z 173 (Na2Il


also shifted on the mass axis due to


space charge), m/z


(NaI2+), and m/z


323 (Na3)l2).


In order to obtain the mass spectrum shown in Figure 4-3b, a positive voltage


can










7000

6000

5000

4000

3000

2000

1000

0





1500


+ shifted
127




+






NaI shifted
173


Nal2
277 Na I2+
I 525


150 200 250 300 550 400
m/z


1000


150 175


200
m/z


225


250


Figure 4-3.


LDI/MS of sodium ions, a) positive (a
each scan), b) negative ion spectra.


Average of 5 scans, 20 microscans
Several peaks are shifted to the


right due to space charging (a) or calibration error (b).


275


300






73

Two peaks are seen in this mass spectrum at m/z 127 for the I ion and m/z 277 for


the NaI2,


In both experiments,


the ions hit the conversion dynode and any


emitted negative particles were attracted to the more positively charged electron


multiplier for detection.


The opposite charged species may appear as artifacts in the


mass spectra (e.g., a negative ion may hit the detector although the voltage settings

make it highly unfavorable), but trapping conditions may be altered to remove these

ions.8


Matrix Compounds


Of the different compounds studied using laser desorption ionization in the

ion trap mass spectrometer, the UV absorbing matrices were expected to give good

ion signals due to their strong energy absorbance characteristics at the nitrogen laser


wavelength as discussed


Chapter


LDI mass spectra


were


obtained for the


various matrix compounds used so that ions which are produced


MALDI


experiments could be attributed to either the analyte or the matrix compound.


matrices


used


were


methanol


solutions


sinapinic


acid,


nicotinic


acid,


dinitroaniline, coumarin-120, and 2,5-dihydroxybenzoic acid.


Of these compounds,


the ones which were used consistently were sinapinic acid and 2,5-dihydroxybenzoic


acid.


These


matrices were


observed


to give


the desired


MALDI


results in


laboratory.


Reports in


literature also showed these matrices to be


the most


popular due to their universal applicability and uniform MALDI spectra.








LDI mass spectrum


2,5-dihydroxybenzoic acid


(DHB)


is given in


Figure 4-4.


In this spectrum, the largest peaks observed are at m/z ratios of 136, 137


and 153 (the masses are off the axis assignment by about 0.5 u).


These fragment ion


peaks are attributable to a loss of a hydrogen (m/z 153), a loss of a hydroxyl group


(m/z


137), and a loss of water (m/z


136).


All of these fragment ions were easily


assigned and the mass spectrum is typical of those obtained using laser desorption


ionization.


No ions are seen for the molecule at m/z


The characteristics of the


LDI mass spectrum are similar to spectra obtained performing electron ionization


(El) on gaseous samples.


It is noted then for future reference when performing


matrix-assisted laser desorption ionization using a DHB matrix solution that these


ions may appear in the mass range below


154 u.


Another


compound


which


was


observed


work


preliminary


experiments performed in our laboratory was 2,4-dinitroaniline (2,4-DNA).


The LDI


mass spectrum


2,4-DNA


is given in


Figure 4-5.


this mass spectrum,


observe the protonated molecular ion peak, (M+H)


, for 2,4-DNA at m/z


184 that


may have been produced by self chemical ionization.


We also detect peaks at m/z


167 for [M-NHj2]


, m/z 154 for [M+H-NO]+


, m/z 136 for [M-HNO2]J


, m/z


123 for


[M-N202]+


, and m/z 89 for [C6H3N])


A peak is seen at m/z 39 for potassium and


at m/z 24 for sodium shifted to the right by 1 mass unit.


mass


spectrum


2,4-DNA


given


Figure


is the


33rd


scan


performed in this experiment.


The relationship of several of the ions in the mass



























(M-H20)+


(M-OH)+


(M-H)+


10 115 120 125 10 15 140 145 150 I 155 160
10 115 120 125 150 155 140 145 150 155 160


Figure 4-4. LDI/MS of 2,5-dihydroxybenzoic acid.


I Aa A A --


_ _~__


_L

























































m-


I K if


O
bo O

0
y2


0 E


m i II I1 I


Y-rr~Cc~-L~Y~Lyll --


d
--


Ai~snamni


















4000-


2000



0 5 10 15 20 25 30 35 40 45 50
1500-

1000 -
1000: (M-NH2)+

500


0 5 10 15 20 25 30 35 40 45 50


4000


2000


0-
C


(M+H)+


I i i i I ~ I i i i I I I i 5 I i I
) 5 10 15 20 25 30 35 40 45 50


Scan

Figure 4-6. Mass chromatograms from LDI/MS of 10 pg of 2,4-dinitroaniline
(MW 183). 1 microscan each scan.


I-






78

versus scan, the potassium ion intensity is seen to begin at a maximum intensity and


decrease over 50 scans. Each scan represents a single laser pulse followed by a mass

scan (e.g., no scan averaging). The intensities for 2,4-DNA ions at m/z 164 and m/z


184 are seen to increase from nearly from 10% of their maximum intensity to 100%


after the first two pulses.

the following 45 scans. E


The ion intensities then decrease from that maximum over


between scans 10 through 50, however, the 2,4-DNA ions are


greater in intensity than the potassium ions.


This relationship is typical of many of


the experiments performed using LDI in our laboratory.


The first few laser shots


may contain a large number of alkali


metal ions,


the sample


ions dominate the


spectra after several shots, and the overall sample ion intensities decay over multiple


laser shots.


The first few mass spectra also tend to be greatly space charged as the


initial layers of sample material are removed.


The space charge can occur because


either there is a higher abundance of preformed alkali metal ions at the sample top


layers


which


are efficiently


desorbed


laser


greatly


increase


number, or the combination of fresh sample material, alkali metal ions, and laser

irradiance setting produces a large number of sample and alkali ions producing space


charge conditions.


Therefore, to obtain the best mass spectra of resolved sample


situation,


a later mass


scan


can be


used


such


as scan


experimental series as was given in


Figure 4-5.


The occurrence of space charge


conditions causing the loss of mass resolution was also observed by Gill and Blades

for LDI/MS of both organic and inorganic samples under varying conditions and






79

The full scan mass spectrum for sinapinic acid, the other most commonly used


matrix compound, is given in Figure 4-7a.


In this figure is a peak at m/z 224 for the


molecular ion surrounded by many smaller peaks, particularly below m/z 210.


While


the molecular peak is easily seen, the smaller ion peaks are not easily discernable


and their assignment to sinapinic acid fragments or background is difficult.


This


result


is another


example


of the


need


for tandem


mass


spectrometry


to obtain


analytically useful information from laser desorption ionization. As shown before for

TMPA, Figure 4-7b gives the single ion isolation scan of molecular sinapinic acid.


this mass spectrum,


we also see a


few small


peaks at (M+1)


and (M-1),


isolation is essentially complete.


The ion intensity for the peak at m/z 224 is more


than ten times greater than the ion intensity in Figure 4-7a and probably due to a


change


sample


surface


conditions


improve


ion production


(similar


to the


process


seen


2,4-DNA


Figure


Figure


gives


daughter mass


spectrum of sinapinic acid.


While we still have molecular ions at m/z 224 due to


incomplete dissociation conditions, there are also peaks at m/z 209 for [M-CH3]


193 for [M-OCH3]+


, m/z


179 for [M-COOH]+


, and m/z 164 for [M-CzO2H4]+


. The


quality and characteristics of the LDI mass spectra shown above are similar to those


obtained by Heller et al. and Glish et al. for similar studies.


38.39


Matrix-Assisted Laser Desorption Ionization


purpose


of the


following


experiments were


to determine


whether a













M+








m..\ A-A A.VA


180 185 190 195 200


205 210 215 220 225 230 235 240
m/z


5000


4000



3000

C

2000


1000



0 I II i ir l '' i l i i t I il 1 i i i l i1 si iI i I t' r 1 1 i i i i 1r i I i
180 185 190 195 200 205 210 215 220 225 230 235 240
m/z


Figure 4-7. LDI mass spectrum of sinapinic acid, a) full mass scan, b) isolation of
molecular ion at m/z 224.




















150





100





50





0-
16


(M-CH3)+
209


(M-COOH)+
179

(M-C202H4)+ (M-OCH3)+
164 193


01
0


170 180 190 200 210 220
m/z


Figure 4-8.


Daughter mass
produced by LDI.


spectrum


sinapinic


molecular








the production


of molecular


or pseudo-molecular analyte


ions.


As discussed in


Chapter


1, matrix-assisted laser desorption ionization (MALDI) was developed to


produce molecular ions for thermally labile and involatile compounds. This method

is usually applied to biological compounds such as peptides and proteins. MALDI


was initially utilized in our laboratory for the enhancement of the production of


molecular


spiperone


Ions


from


solid


residue


of the


drug


compound.


Two


important aspects, however, needed to be considered for the application of MALDI


to spiperone.


First, suitability of the method to spiperone and ease in obtaining


molecular ions needed to be evaluated.


Second, the production of higher mass ions


formed at


ring surface


using


MALDI


may pose


a problem


for the


internal


ionization configuration used


on the ion


trap.


Initial experiments by Beavis and


Chait showed that peptide ions formed by MALDI carry initial kinetic energies that

increase linearly with mass, which could be a potential problem for efficient trapping


ions.


However,


later work


Cotter


as well


as Beavis


Chait,


demonstrated that initial kinetic energies begin to level off at higher mass.34'35


two research groups showed that MALDI


These


ions essentially have around the same


velocity (750


- 1000 m/s).


Therefore,


with nearly equal velocities,


the higher the


mass the higher the kinetic energy.


These results, along with the successful trapping


of peptide and proteins ions in a Fourier-transform mass spectrometer, suggested

that MALDI ions formed in ion trap could be stored for analysis as well.

Another consideration for the application of MALDI to the ion trap, although








charge unlike the multiply charged ions produced by electrospray ionization.


This


characteristic of MALDI limits the mass range that can be achieved utilizing ions


carrying


a high


number


charges.


While


limitation


is not a


problem


spiperone (MW 395) or other drug compounds which have molecular masses below

the conventional ion trap maximum mass (650 u), the utilization of MALDI for high

mass ions will require mass range extension methods to be applied to the ion trap


as well.


Results from


this approach have been


demonstrated by a


few research


groups for the analysis of peptides and proteins.19


MALDI of Spiperone


Spiperone,


spiroperidol


(8-[3-(p-fluorobenzoyl)propyl]- 1-phenyl- 1,3,8-


triazaspiro[4.5]decan-4-one), is an antipsychotic drug which was first suggested for


use in evaluating the ion trap by Mike Lee of Bristol-Myers Squibb.


He was trying


to detect spiperone that had been sprayed onto the surface of a tissue sample using


secondary ion/time-of-flight mass spectrometry.


our laboratory using LDI.


Initially, spiperone was analyzed in


The resultant mass spectrum is shown in Figure 4-9.


molecular ion peak which would be situated at m/z 395 is hardly discernable in this


spectrum.


All of the spiperone that is ionized is also fragmented.


The two major


ions obtained are the spiperone fragment ions at m/z 165 and m/z 98.


The low mass


cutoff during trapping and storage was 40 u and the helium pressure was 0.6 mtorr.

For similar conditions of low mass cutoff and helium pressure, matrix-assisted laser


























S


0 0























I CO
+ To
+(7)


-4


I I


I I


E


co


/1*
a O
Y '-
1-.
tu s