Development and application of particle beam LC/MS on the quadrupole ion trap mass spectrometer

MISSING IMAGE

Material Information

Title:
Development and application of particle beam LC/MS on the quadrupole ion trap mass spectrometer
Physical Description:
viii, 181 leaves : ill. ; 29 cm.
Language:
English
Creator:
Kleintop, Brent Lamar, 1966-
Publication Date:

Subjects

Subjects / Keywords:
Particle beams   ( lcsh )
Liquid chromatography   ( lcsh )
Mass spectrometry   ( lcsh )
Quadrupoles   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1993.
Bibliography:
Includes bibliographical references (leaves 172-180).
Statement of Responsibility:
by Brent Lamar Kleintop.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001890692
oclc - 29642207
notis - AJW5916
System ID:
AA00003244:00001

Full Text







DEVELOPMENT


AND APPLICATION OF PARTICLE BEAM LC/MS


ON THE QUADRUPOLE


ON TRAP MASS SPECTROMETER


BRENT LAMAR KLEINTOP













A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL


OF THE UNIVERSITY OF FLORIDA


OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


N PARTIAL FULFILLMENT





























To Mom and Dad in appreciation for al


they've done.











ACKNOWLEDGEMENTS


wish to express my sincere gratitude to my advisor


Richard A.


Yost


for his


night and guidance throughout my graduate studies.


to acknowledge and thank the members of my


would also


graduate committee,


. Jim


Winefordner, Dr. Vanecia Young,


wish to acknowledge Dr


Dr. Jim Boncella, and Dr


Jodie Johnson


Joseph Delfino.


a valuable member of the


also


Yost group,


for many fruitful discussions,


both scientific and personal.


also wish to acknowledge those who provided funding for my research


salary which


prevented


me from


seeking


employment at


McDonald's


support myself.


This


includes the NASA Amtech Joint Agreement,


the Sandoz


Research


institute,


U.S.


Environmental


Protection


Agency/EMSL


Cincinnati


An extra thank you goes to Dr.


Thomas Behymer of the U


EPA/EMSL for his willingness to assist at any time.


have great difficulty finding the words to express the gratitude


have for


the many special friendships


have shared while at UF


. These include fellow Yost


group


members


Don


"The


Gecko


Eades,


Stacy


Rossi-Barshick,


Brad


"Smoothie" Coopersmith,


and Nate "Air Quaker" Yates


, as well as fellow graduate


students Kevin Kinter, Keith Palmer


and Larry "Moti Oti"


Villanueva.


would like


to make an extra acknowledgement of Don Eades, whom


thoroughly enjoyed






having as my coworker on this particle beam project, and Nate Yates, for always


knowing how to change the


trap software to suit my needs.


wish them and


everyone else too numerous to mention the best of


uck in their careers.


memories


we have shared will never be forgotten.


These were truly the


salad


days.


Finally and most


importantly,


express my sincerest


ove and thanks to my


family.


It is not possible to express enough thanks to my mother and father, who


have worked their entire life to provide me with whatever was necessary to help


me achieve my goals.


've never felt closer to my brother Gary whom


have


watched change careers,


attending Florida.


get married,


thank Gary,


and start (and expand) a family while


Nadine, Ryan,


was


and Jarrett for not letting the family


become boring wh


was away.


Although hundreds of miles have separated me


from my family during recent years,


have never felt closer to all of them.













TABLE OF CONTENTS


ACKNOWLEDGEMENTS


S S a S S S S ii


ABSTRACT


VII


CHAPTERS


INTRODUCTION


S. . 1


The Quadrupole Io
Brief History


General Theory of


Trap Mass Spectrometer


. 2


SS 5 2


on Motion in the QITMS


Principles of Operation . .
LC/MS using the QITMS .


The Particle Beam LC/MS


nterface.


Overview of Organization of Dissertation


CHARACTERIZATION OF


IMPORTANT PARTICLE BEAM


OPERATING PARAMETERS


Introduction
Experimental
Characterization Studies


* S S
* S S S


Nebulizing Helium Flow Rate


Desolvation Chamber


. S S S
* S S S S


Temperature


Desolvation Chamber Pressure


Target Temperature
Conclusions


EVALUATION OF STRATEGIES TO MINIMIZE ADVERSE


EFFECTS OF RESIDUAL SOLVENT


introduction


Experimental . . . . .
Problems Caused by Residual Solvent


6
8







Strategies to Minimize Adverse Solvent Effects
Additional Stage of Momentum Separation
Elevation of rf Level . . . .
Resonant Ejection of Single Ion Species


Compariso
Application to
Mixtures
Conclusions


in of Solvent 14
LC/PB/ITMSm


on Ejection Methods


Analyses
*. a .a .a a .a
* a a a a a .a


Simple


* a .a .a a a a
* a a a a a a a a a


Pesticide


* a a
* a a a


* a a
* a a a


APPLICATION OF LC/PB/ITMS TO ANALYSIS OF PESTICIDE
MIXTURES AND COMPARISON TO OTHER LC/PB/MS


SYSTEMS


introduction


Experimental .
Results and Discussion


El Mass Spectra . . .
Instrument Detection Limits
Instrument Calibration Curves
Precision of Peak Areas .
Alternating EI/CI Acquisitions
Conclusions . . . .


IMPROVED QUANTITATION USING COELUTING


LABELED


SOTOPICALLY


INTERNAL STANDARDS


introduction


Experimental
Gas Chromatography/Mass


pectrometry


Benefits of Axial Modulation


Alternating M


ass-


Selective Storage Scans


Liquid Chromatography/Mass Spectrometry
Conclusions .


CONCLUSIONS


Conclusions
Suggestions


AND FUTURE WORK


for Future Work


REFERENCES


BIOGRAPHICAL SKETCH











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 PARTICLE BEAM-LC/MS ON THE


QUADRUPOLE


ON TRAP MASS


SPECTROMETER


Brent LaMar Kleintop


May


Chairperson: Richard A.


Yost


1993


Major Department: Chemistry


This dissertation presents results from


investigations of coupling a particle


beam (PB) interface for liquid chromatography/mass spectrometery (LC/MS) with


a quadrupole


on trap m


spectrometer (QITMS).


Unlike most other LC/MS


nterfaces


, PB interfaces are capable of producing


structural


information from


electron


onization (El) mass spectra which may allow identification of unknown


compounds.


Also,


the acknowledged sensitivity


and versatility


of the QITMS


might provide lower detection limits and provide the impetus for development of

future benchtop LC/MS/MS instrumentation.


Initially,


residual solvent introduced into the QITMS by the interface caused


space charging and a large degree of solvent-chemical ionization (CI).


The use


of an


additional


stage of momentum


separation


in the interface reduced the






amount of solvent which reached the interface; however


it was also necessary to


eject solvent ion


from within the trap prior to m


analysis.


Two modes of ion


trap


operation


implemented


with


standard


instrument


control


software


were


demonstrated to efficiently eject solvent ions.


El m


This allowed generation of classical


spectra of several pesticides which compared favorably to both library


spectra and spectra acquired from solid


probe/QITMS


analyses (i.e. a "solvent-


free" method) of pure compounds.


The ability to perform


both isocratic and gradient elution


LC/PB/QITMS


analyses of pesticide mixtures has been demonstrated.


The results obtained from


gradient


elution


LC/PB/QITMS


analyses


were


also


compared


to results


obtained


LC/PB/quadrupol


e-MS


system.


Both


system


provided


comparable


mits of detection and precision; however


n the relative intensities of fragment ions


of self-C


were observed


at large analyte concentrations was observed


some minor differences

i. Also, a larger degree

on the ion trap system.


Quantitation using coeluting isotopically labeled internal standards (IS) has


also been investigated.


selective storage scan


A new method of rapidly performing alternating m


has been demonstrated to provide excellent


ass-


nearity and


precision over four orders of magnitude


for gas chromatography/QITMS analyses.


coeluting


IS also


provided


linear


calibrations


using


LC/PB/QITMS.


improvements were largely


attributed to the coeluting


IS acting


as a


"carrier"


which improved analyte transport through the interface.











CHAPTER 1
INTRODUCTION


Since becoming commercially available in 1983,


the quadrupole ion trap


mass spectrometer (QITMS) has seen widespread use in both fundamental and


applied


mass spectrometric studies.


Although


introduced


commercially as a


benchtop mass spectrometric detector for gas chromatography (GC),


the QITMS


currently


seeing


increased


applications


liquid


chromatography/mass


spectrometry (LC/MS) systems.


In this dissertation


, results from


investigations


focused on coupling a particle beam (PB)


nterface for LC/MS


with a QITMS wil


be presented.


initially


problems were encountered which were caused by the


introduction


of residual


solvent into the QITMS by the


nterface.


Several


strategies were investigated to minimized these adverse solvent effects and allow


acquisition


classical


electron


ionization


mass


spectra


a variety


environmentally significant compounds which compared favorably with


brary El


spectra


acquired


more


traditional


mass


spectrometers


(e.g.


sector


quadrupole instruments).


The capabilities of the LC/PB/QITMS system were then


evaluated by applying it to the LC/MS analysis of pesticide mixtures.


Since al


commercially


available


systems


employ


quadrupole


mass


analyzers,


analytical figures of merit obtained on the LC/PB/QITMS system were compared


ft -- I *- 1 t I I a -


a CIA .-


*-l i








operational modes were


investigated to


improve quantitation using isotopically


labeled internal standards.


This


introductory chapter provides a historical perspective, as well as a


presentation of the operating principles of both the QITMS and the PB interface


for LC/MS


and also provides an overview of the organization of this dissertation.


The Quadrueole Ion


Trao M


Spectrometer


Brief History


A summary of the history of the development of the


QITMS


is provided


Table 1


The quadrupole


on trap was


initially described


n 1953 by Pau


coworkers


at the


University


Bonn


(Paul


Steinwedel


, 1953),


was


detailed


same


patent


described


operating


princip


quadrupole mass filter (Paul and Steinwedel,


1960).


The same


deas were also


proposed that same


year by Post and Heinrich (1953) and by Good (1953) at the


University of California,


mass


Berkeley.


spectrometer by Fischer (19


The quadrupole ion trap was first used as a

)59), who used mass-selective ion detection


using resonant absorption to produce a m


spectrum of krypton.


Detection via


resonant absorption involves the application of an auxi


ary alternating current (ac)


voltage of a particular frequency across the endcap electrodes of the ion trap.

An ion undergoes resonant absorption when its fundamental frequency of motion


!FI *lk a earvAn ae *ha #ranm aanr"ri *ha


IFIAAIIIAC Crnhr Ihnhu\


su. *r Irr~ ^ ** ___

















Table 1


Summary of the stages of development of the QITMS.


column


also depicts the time periods of each operational


ie third
mode


used during these stages of development.


1953


1959
1962


First disclosure (Paul and Steinwedel)
Use as a mass spectrometer (Fischer)


Storage


Mass-selective detection


of ions for rf spectroscopy


(Dehmelt and Major)


1968


Ejection of


ons to an external detector


(Dawson and Whetten)


1972


1980


Use of QUISTOR as ion source for
quadrupole mass filter (Todd et al.)


First use


Mass-selective storage


as GC detector (Armitage


March)


1984


1985


Disclosure of ion trap detector (ITDTM)
(Stafford et al.)
Ion trap mass spectrometer (ITMSTM)
(Kelley et al.)


ass-se


lective instability


1987


MS/MS, C
extension


ion injection


, mass-range


Th







applied potential


When this occurred


, ion signal was detected by measuring the


power absorbed by the ion from the resonant excitation circuit.


Many early applications of the ion trap


involved its use with a wide variety


of gas-phase rf


pectroscopic experiments


(Dehmelt,


1967


1969


Major


Dehmelt


the late


,1968).


The evolution of the ion trap as a mass spectrometer began in


1960s with the development of mass-selective storage


(Dawson


Whetten


, 1968a,b,


1969),


which is


milar to the operation of the quadrupole mass


filter.


n this mode, appropriate rf and direct current (dc) potentials were applied


to the electrodes such that only a very narrow mass range of ion


remained


trapped


small


Ions were detected by ejecting them from the


holes


trap through a series of


n one of the endcap electrodes to an external detector such as an


electron multiplier.


Although this method was usefu


for analysis over a fairly


narrow mass range,


it was cumbersome for acquiring a full m


spectrum,


which


required repeating the process of mass-selective storage for each mass over the

range of interest.


a result


of the


limitations


mass-selective


storage


mode


operation,


vast


majority


initial


experimental


applications


mass


spectrometric


studies


used


as a mass-selective


source


for the


quadrupole m


filter. For th


applications,


Todd and coworkers coined the


term


"_yadrupole ion store" or QUISTOR for the ion trap


(Lawson and Todd,


1971).


The QUISTOR/quadrupole combination was successfully used throughout








theoretical


(Todd


, 1980a)


aspects


trap,


develop


new


instrumentation (Kishore and Ghosh,


1979) and operational modes (Fulford and


March,


1978; Fulford et al.,


1980), and to study a variety of ionic processes such


as ion/molecule reactions


(Bonner et al.,


1973,


1974) and ion energetic


(Todd


et al.


,1980b).


capabilities


traps


have


dramatically


increased


with


development of the mass-selective instability mode of operation (Stafford et al.,


1984


Kelley et al


1985).


n this mode, ions covering a broad range of mass-to-


charge ratios (m/z) are trapped by application of an appropriate rf voltage to the


ring electrode.


A mass spectrum


is obtained by


nearly increasing the amplitude


of the rf voltage, which causes


ons of increasing m/z to be ejected through holes


in one of the endcap electrodes and detected by an external electron multiplier.

This new scan method provides a much simpler and more rapid approach for


acquisition


of mass spectra.


From


a single ionization


event


, an entire mass


spectrum can be obtained.


The development of the mass-selective instability scan was the impetu


ed to the commercial introduction of the ion trap detector (ITDTM) by Finnigan


MAT


n 1983 as a benchtop mass spectrometric detector for GC.


and sensitivity of the


analyses.


TDT resulted


1985 Finnigan MAT


The low cost


n widespread applications for routine GC


introduced the more versatile ion trap mass


spectrometer (ITM


STM) which allows both routine MS and more advanced MS








S/MS/MS, etc.).


Interest in ion trap technology increased dramatically after the


introduction


of the


TDM


TMS


a result,


instrumentation


operation of the QITMS have been the subject of several reviews (Cooks and


Kaiser


1990


Nourse and Cooks,


1990; Cooks et al.,


1991


March


1991


Todd,


1991


Todd and Penman


, 1991


March


, 1992).


A recent book also reviews the


fundamental


and applications of the quadrupole ion trap (March and Hughes,


1989).


General Theory of Ion Motion


n the QITMS


The QITMS


is operated by app


cation of an rf voltage to the ring electrode


which creates a three-dimensional electric field within the


electrode


surfaces.


Ions


are either trapped or ejected from the trap depending upon the m/z of the ion and


the applied electric fields.


The motion of an ion in a quadrupole electric field can


be described by equation


which is also known as the Mathieu


equation


(Dawson,


1976):


- 2 q cos 2) u


where u represents either the radial (r) or axial (z) axis,


voltage,


Jis the


and a. and q, are the Mathieu stability parameters.


phase of the rf


These parameters


are related to the


zero-to-peak


rf amplitude


amplitude


S .'


angular


S


+ (au


df









8eU


(1-2)


m ro Q2



4eV
mr02 2


(1-3)


Since ro and


are constant


, the


Mathieu


parameters can


be simplified


expressed as:


= U(m)
e


x const.


(1-4)


= V ()
e


x const.


(1-5)


From th


simplified equations,


it becomes


evident that the Mathieu parameter


q, is related to the rf potential applied to the ring electrode and that for a given


rf potential,


each m/z


has a unique value of q,.


Similarly


, au is related to the


applied dc potential and for a given dc potential,


each m/z


has a unique a. value.


values


determine


whether


solutions


Mathieu


equation


are stable,


through zero


meaning the displacement of an


(the center of the trap),


ion periodically passes


or whether the displacement


increases


= az








which yield stable solutions to the Mathieu equation,


it will be trapped and wi


process


within


trap


particular


secular


frequency.


Graphical


representations of stable solutions in a and q space are called stability diagrams.


The stability region which is relevant to the QITMS is shown in Figure 1


Ions


with values of a and q which


ie within the stab


ity diagram wi


be effectively


trapped.


The stability region is bounded by so called


so-pn


for .,


=0 or 1


a complex function of a. and q, and determines


the secular frequencies


of the


trapped ions.


The trajectories of ion


with the same l,,


value have the same


secular frequency


however


, the amplitude of the trajectories is governed by the


a and q values of each m/z.


The trajectory of an ion in the rz plane is illustrated


n Figure 1


It has the form of a


Lissajous figure composed of two secular


frequency components with a superimposed high frequency ripple resulting from


higher order harmon


of the ion's


secular frequ


Principles of Operation


quadrupole


trap


is the


three-dimensional


analog


of the more


iar quadrupole m


filter.


A schematic of the Finnigan


STM which was


used for most of these studies is


hown


in Figure 1-3.


It con


ists of a hyperbolic


ring electrode and two hyperbolic endcap electrodes,


radius of the ring electrode and


where ro represents the


is the half distance between the two endcap














0.20




0.00




-0.20


-0.40




-0.60




-0.80


iso-a 1.0
lines na o


0.0 0.2


0.00


0.0 q, eject = 0.908
0.2


0.4 rf-only
0.4
Yiso-p,
lines

0.6



0.8


0.40 0.80 1.20 1


Figure 1-1:


Stab


ity diagram for the quadrupole


on trap plotted


n a,q


pace.


with a,q parameters


ocated within the region of stability wil


ssess


stab


trajectories and be effectively trapped.








































Figure


plane.


Lissajous form of the trajectory of a trapped ion in the rz
'he figure is from a photograph of an illuminated charged


aluminum
trajectory


particle


spend


composed


in a quadrupole


lower


secular


uperimposed high frequency ripple (Todd,


ion trap.


frequency


with


1991)





















Filament


End Cap


Ring Electrode


Electron Multiplier Detector


End Cap


To Preamplifier
(Ion Signal)


Figure 1


Schematic diagram of the Finnigan ITMSTM used throughout these
studies.


I


J


Amplifier and
RF Generator,
Fundamental
RF Voltage


-


Scan Acquisition
Processor
(Computer)


Amplifier and
RF Generator,
Supplementary
RF Voltage








(Knight,


1983).


It has recently been disclosed however, that al


commercially sold


ion traps are actually "stretched" in the axial,


z, direction such that r


= 1.63zo2


(Louris et al.,


1992) where r,


1.000 cm and zo


= 0.783 cm.


The quadrupolar trapping field of the QITMS is generated by application


of an rf voltage to the ring electrode.


For the


TMST


, the frequency of the applied


rf signal is 1.1 MHz.


The generated field is actually not an ideal quadrupolar field,


but has contributions from higher order fields as we


Octapolar and dodecapolar


field contributions result from the "stretched" geometry of the electrode spacing.

Imperfections in the electrode surfaces and holes for injection of ionizing electrons


and ejection of ions


also


introduce hexapolar contributions to the trapping field.


Studying the effects of these nonlinear resonances has been the subject of many


recent


investigations (Guidugli et al.


1992; Eades and Yost,


1992)


Electron


onization


most


commonly


used


method


production of positive ions


n mass spectrometry (Duckworth et al


1986).


In the


ITMST


, ionizing electrons are created from a heated rhenium filament located


outside one of the endcap electrodes.


Electrons enter the trap by gating an


entrance


ens at


180 V for a period of time equivalent to the ionization time


(typically 1 ms) and ionize the neutral


present within the trap.


Ions are effectively


trapped in the


TMST when the values of their az and q, parameters fall within the


stability diagram (Figure 1


When operated


in the rf-only mode (az


= 0),


ions


with q, values less than q,


= 0.908 wil


be effectively trapped and stored.








pressure of a light buffer gas,


typically


mtorr of helium.


Collisions between


ions and helium atoms serve to kinetically "cool" the ions, which causes the ions


to migrate toward the center of the trap.


This results


n dramatic increases


sensitivity and resolution (Stafford et al.,


1984).


ITMSTM


acquires


a mass


spectrum


using


mass-selective


instability scan method (Stafford et aL


1984).


During the mass-selective instability


scan there is no dc potential


= 0)


applied to the trap


such that all ions


possess values of q, which lie along the qz axis of the stability diagram.


Since q,


inversely proportional to m/z


(Equation


1-5),


ions of lower m/z are located


towards the right of the q, axis.


Ions are "moved" along the q, axis by


nearly


ramping the rf potential,


*, (qz = Vfrom Equation 1-5) until they become unstable


at the stability boundary


where q,


= 0.908.


This results


n sequential ejection of


ions of


increasing


m/z from the trap where they are detected


by an external


electron multiplier


of 180 psec/u.


. Mass spectra are recorded on the


Typically during


TMSM using a scan rate


mass analysis, an additional axial modulation


signal


is applied across the endcap


electrodes.


This supplemental


signal


applied at slightly less than half the rf drive signal at approximately 6 Vp.


Just


prior to ejection,


the ion


come into resonance with the axial modulation signal,


which increases the amplitude of the


on's trajectory in the


z direction.


This


results in the ions being more tightly bunched upon ejection,


resolution and detection efficiency (Weber-Grabau et al.


yielding increased


1988).







simultaneous dc, rf,


and ac voltages.


For example, isolation of a single m/z


or a


narrow range of m/z's


n the trap can be accomplished


by application of an


appropriate combination of rf and dc voltages such that the m/z of interest is near


the apex of the stability diagram (Figure


-1) at q,


= 0.78, az


= 0.15 (Weber-


Grabau


, 1987).


Alternatively


, a "two-step"


isolation


method


been


described where higher mass ions are ejected at a working point along the $z


0 stability edge and lower mass ion


are ejected at a working point along the $z


stability edge (Gronowska et al.,


1990;


Yates et a/.


, 1991).


Tandem mass spectrometry (MS/MS) experiments are performed by first


mass-selecting a particular precursor ion.


supplemental resonant excitation


potential


is then


applied


across the endcaps at the secular frequency


of the


precursor


Resonance


excitation


kinetically


excites


causing


increase


n the amplitude of its trajectory


in the z-direction.


The amplitude of the


resonant excitation voltage,


however


, is kept low enough to avoid losses due to


resonant


ejection.


The kinetically excited precursor ion


undergoes


energetic


collisions


with


helium


buffer


which


can


result


collisionally


activated


dissociation (CAD).


Studies have demonstrated up to 10059


CAD efficiency and


overall


MS/MS


efficiencies approximately 14 times


higher than triple quadrupole


instruments which suggests that 'the ion trap may be


the most sensitive MS/MS


instrument


ever"


(Johnson


, 1990,


2172).


Multiple


stages


mass


spectrometry


(MS",


n>2)


experiments can be performed


by mass-selecting a








experiments up to MS12 have been successfully demonstrated


on the


TMSTM


(Louris et al.


1990).


LC/MS usina the QITMS


The combination of high-performance liquid chromatography (HPLC) with


mass


spectrometry


been


widely


recognized


as possessing


enormous


potential for analyzing a wide variety of compounds. Although HPLC does not

provide the chromatographic efficiency of capillary GC, analytes remain in the


condensed


phase


ambient


temperature


during


separation,


thermal


decomposition is not a problem.


This makes HPLC-based methods an appealing


choice for the analysis of a wide variety of thermally labile, nonvolatile, and polar

compounds which are either not amenable to GC or require derivatization for GC


analysis.


Indeed


, it has been estimated that only about 20% of al


known organic


compounds are amenable to GC and GC/MS analyses and that HPLC-based

methods should be capable of analyzing the vast majority of the remaining 80%


(Budde et al.


1990).


The major challenge involved with using a mass


detector is the design of a suitable


spectrometer as an HPLC


nterface which can convert the high pressure


mobile


phase


flow


phase,


maintain


typical


operating


conditions.


Vaporizing this quantity of liquid generates a volume of gas that must


be substantially reduced to maintain the high vacuum (typically 1


x 10.5 torr) of








direct liquid


introduction


(Baldwin


McLafferty,


1973),


interfaces


represented the first successful LC/MS interfaces.


However, the moving belt was


mechanically complex and subject to memory effects, while DLI required splitting


a large portion of the LC eluent which limited the sensitivity of the technique.


1980s


, the development of a "new generation" of improved aerosol-based


LC/MS


interfaces


including


thermospray


(Blakley et al.,


1980),


particle beam


(Willoughby and Browner,


1984),


and electrospray (Fenn etal.


1989) dramatically


broadened the applicability of LC/M


Although


majority


LC/MS


research


been


performed


using


quadrupole


instruments


, ion traps are seeing


increased use


n LC/MS


systems.


initial attempts to


nterface the widely used thermospray (TSP)


nterface with an


TMSTM


(Kaiser


et al.


, 1991)


exhibited


good sensitivity;


however,


high


solvent


background


pressures


caused


significant


peak


broadening


poor


mass


resolution).


performance


TSP/ITMS


system


was


significantly


improved using a differentially pumped ion source (Bier etal.,


1991).


Electrospray


(ESP) interfaces have seen the greatest number of LC/ITMST applications


n the


literature


(Van


Berkel et aL


, 1990,


McLuckey


et al.


, 1991)


. Since ESP yields


multiply charged ion


, even high molecular weight biological compounds are well


within the working mass range of the


TMSTM (m/zm,


= 650 u).


Ion trap systems


coupled with


TSP and ESP require the use of an ion injection system because


ions are generated within the interfaces,


external to the trap.








A number of features of the


TMST make it particularly attractive as a mass


analyzer for


TSP and ESP generated ions.


These include high sensitivity


ability to perform MS" experiments with high CAD efficiency,


perform kinetic measurements.


and the ability to


Since TSP and ESP are known as "soft" ionization


techniques which produce predominantly molecular ion


the capability to obtain


structural information from MS" experiments is particularly appealing.


The Particle Beam LC/MS


Interface


interface


represents


another


of the


new


generation


LC/MS


nterfaces based upon aerosol formation.


aerosol-based LC/MS


The primary advantage of using an


nterface is that solvent can evaporate far more rapidly from


droplets than from bulk solutions because of the


much higher surface area-to-


volume ratio (Browner et a.,


1986).


The original design was disclosed in 1984


and was named MAGIC for monodisperse aerosol generation interface for
---- 0 vww ^^^w


iquid


chromatography (Willoughby and Browner,


1984).


As variations based on MAGIC


became commercially


available


(Apffel


Nordman


1987


Willoughby


Poeppel,


1987),


they became simply known as particle beam


nterfaces.


Although differences exist between various commercial PB interfaces, all


consist of a nebulizer and some type of desolvation chamber


momentum


coupled with a


separator which removes solvent vapor from the beam of largely


desolvated particles. A schematic of the prototype


interface used


n these studies










~.2 o
a-

( C
I w


0) 18
LLS -
u.c

'I,


a>
Sr-
4-'


"V
/K







(d
ci3



3
(IS

()
0)


~IL


--E


-> O


ci1
^-


_I


J






19
employs an additional third stage of pumping in the momentum separation region

to further reduce the amount of solvent which reaches the ion trap.


The interface works by first converting the LC


effluent into an


aerosol


having a narrow range of droplet diameters.

the liquid with a high concentric flow of heliu


The nebulizer in this interface shears

m. The droplets then travel through


a heated desolvation chamber where volatile solvent is vaporized resulting in a


mixture


of solvent vapor


, helium


, and


largely


desolvated


particles


containing


analyte.


This mixture then passes through a beam collimator which form


particle


beam


which


then


travels


through


momentum


separator


momentum separator


n this


nterface consists of three stages of pumping which


are separated by a series of skimmers.


Solvent and helium are pumped away


the momentum separator while the more massive and higher momentum analyte


particles


are transported


through


momentum


separator


region.


These


particles are transported through a transfer probe


nto the mass spectrometer


source region.


The transfer probe was


inserted


nto the


TMST via a


/2" o.d.


transfer line probe which allowed insertion and removal without disturbing the


high vacuum of the ITMSTM


The beam of particles is "flash" vaporized when it


collides with the hyperbolic surface of the filament endcap upon


entering the


TMST analyzer cavity.


The resulting vapor then travel


toward the center of the


trap where it is ionized


alternatively


by chemical


ionization


ion/molecule reaction


with reagent ions produced from a reagent gas.


Typical






20
operating parameters and optimization of the interface will be described in greater


detail


n Chapter


Particle beam interfaces have generated recent widespread interest largely


because of their ability to produce classical El m


spectra for a wide variety of


relatively


involatile


thermally


labile


compounds


environmental


biological interest (Behymer et al.


1990,


Voyksner et al.,


1990).


The generation


of El spectra is advantageou


because the fragmentation patterns found


n these


spectra


reveal


structural


information


which


affords


ability to


identify


unknown compounds,


a typical goal of many LC/MS


analyses.


El spectra also


can be compared to existing library spectra for the identification and confirmation


of unknown compounds.


TSP and ESP interfaces produce spectra unique to the


nterface which yield predominantly molecular weight


information.


This typically


mits th


techniques


targeted


compound


analysis,


although


use of


tandem mass spectrometry for obtaining structural


information has been reported


(Smith et al.


1990


Van Berkel et al.


, 1990).


Coupling a PB interface with a QITMS would appear to be an appealing


choice for many applications employing LC/MS.


The ability of PB to generate El


spectra affords the


identification of unknown compounds,


a typical goal of many


types of analyses, while the


sensitivity of the


QITMS


should also provide


ow limits


of detection.


mass


Currently,


all commercially available PB systems employ quadrupole


analyzers due to their ruggedness and low cost.


Although the coupling of








ions from an external ion source (Bier et al.,


1991) has been demonstrated,


rigorous studies have been performed which thoroughly evaluate the analytical

capabilities of a PB/QITMS system.


Overview of Oraanization of Dissertation


In this dissertation, results from investigations of coupling LC with a QITMS

via a PB interface are presented. A historical perspective and a review of the

basic operating principles of the QITMS and the PB interface are provided in this


introductory chapter.


The optimization of important PB operating parameters and


instrumental modifications are discussed in Chapter


Chapter 3 presents results


of several methods investigated to minimize adverse effects of residual solvent.


Application


of the PB/ITMSTM


system to the analyses of pesticide mixtures


presented


Chapter


Chapter


4 also


compares


performance


of the


PB/ITMSTM system with a PB system which employs a quadruple mass analyzer.


investigations of several ion trap operational modes to


improve quantitation using


isotopically labeled


internal


analyses are discussed


standards for both GC/ITMS


n Chapter 5.


Finally,


LC/PB/ITMS


conclusions and perspectives for


future work are presented in Chapter 6.











CHAPTER 2


CHARACTERIZATION OF


IMPORTANT PARTICLE BEAM


OPERATING PARAMETERS



Introduction


As discussed


n Chapter 1


particle beam (PB)


nterfaces have attracted


widespread


interest


recently


because


they


allow


generation


classical


electron


ionization


mass


spectra


a wide


variety


compounds


environmental (Brown et al.


1990) and biological interest (Voyksner et al.,


1990).


Generation


spectra


is advantageous


because they typically


exhibit rich


fragmentation patterns which can reveal structural information.


patterns of th


The fragmentation


spectra can then be compared to reference spectra using readily


available computer library spectral searches for the identification and confirmation

of unknown compounds, a typical goal of many LC/MS analyses.


Although the important PB operating parameters are well known (Voyksner


et al.


, 1990),


variation


between commercially available PB


nterfaces and MS


system


necessitate the need to characterize th


parameters for each system


n order to intelligently develop effective LC/MS methods.


chapter describes


and characterizes the important PB operating parameters for the LC/PB/ITMS


system


used


throughout


these


studies.


order


improve


performance,








modifications were made to both the interface and the analyzer.


and the results of these modifications wil


A description


also be discussed.


Experimental


instrumentation


Two HPLC systems were used for these characterization studies.


The first


was


ISCO


(Lincoln,


LC-2600


syringe


pump


which


used


a Valco


(Springfield,


NJ) 6-port manual injection valve with a 10 pL sample loop.


second


HPLC


system


was


a Hewlett-Packard


(Palo


, CA)


1090L


liquid


chromatograph fitted with a Rheodyne (Cotati,


CA) manual


injection valve with a


5 pL sample loop.


Solvent compositions and flow rates used are indicated in the


text.


characterization studies


were performed using flow injection analysis


(FIA) of analytes.


The LC/MS


interface was a prototype Finnigan MAT (San Jose, CA) particle


beam


interface


(Figure


1-4).


This


prototype


differs


from


commercially


available interface in that it employs three stages of pumping in the momentum

separator region to further reduce the amount of solvent which reaches the ion


trap.


Supplemental helium was added to the third stage of momentum separation


to prevent backstreaming of pump oil from the mechanical pumps caused by low


pressures.


Typical pressures in the three stages of the momentum separator


were


5-10


mtorr,


mtorr


(adjusted


addition





24
TMSTM via a V1/ probe lock which allowed insertion and removal of the interface


without disturbing the high vacuum of the mass spectrometer.


The probe tip was


positioned


- 14" away from the


on trap entrance to prevent high pressures inside


the trap.


The Rulone fitting normally


inserted


n the entrance hole to electrically


isolate a probe from the endcap was removed so a larger entrance


nto the ion


trap would be present.

The mass spectrometer used in these experiments was a Finnigan MAT


(San Jose, CA)


TMSTM


The vacuum chamber was maintained at 100C for al


experiments.


El was performed within the ion trap; detection was accomplished


with an electron multiplier set to yield


gain


and with


a single conversion


dynode at 0V.


No dynode voltage was used because the dynodes caused high


noise problems on the


STM at the time th


experiments were performed.


Partial pressures of residual solvent were typically


- 1 X 106 torr (uncorrected) as


measured by a Bayard-Alpert ionization gauge (Granville-Phillips,


Boulder,


CO)


mounted on the vacuum chamber.


Helium was added into the


STM vacuum


chamber to produce typical operating pressures of 1


Five microscans were averaged on the


X 104 torr (uncorrected).


Sm and returned to the data system


as a single scan.


Axial modulation (Weber-Grabau et al.


1987) was employed at


530 kHz and 6Vpp during the mass-selective instability scan to improve mass

resolution and sensitivity.








Samples and Reaaents


The structures of the standards used throughout these characterization


studies are shown in Figure 2-1.


Also shown are molecular weights and the


fragmentation which yields the base peak of each compound's


El mass spectrum.


Carbaryl


diuron


were


obtained


from


U.S.


Environmental


Protection


Agency (Pesticide Chemical Repository,


Research


Triangle Park,


NC); caffeine


and rotenone were purchased from Sigma (St. Louis,


MO).


All standards were


used without further purification. HPLC g

obtained from Fisher Scientific (Pittsburgh,


;rade methanol and acetonitrile were

PA) and reagent water was obtained


from a Milli-Q water purification system.


Characterization Studies


main


operating


parameters


interface


which


affect


performance


nebulizing


helium


flow


rate,


desolvation


chamber


temperature, desolvation chamber pressure, and source target temperature.


of these parameters were


investigated to determine what effects they have on


sensitivity


performance


system.


These


parameters


were


investigated with various commonly used LC mobile phases, including methanol,


acetonitrile


, and acetonitrile/reagent water mixtures and different flow rates.


assistance of Mr. Donald Eades during these characterization studies is gratefully























Z-


z


Z -0


0=0
Jvwvv~


i 0
< (C
U)

U)-
wa
v) Q-


00


I
Z-I--


5-o
O "n
C
cn c
fl vU


JVV\NVV\/\//\//\//v\^
+ \ _








Nebulizina Helium Flow Rate


The nebulizer used in the PB interface was a glass Meinhard (Santa Ana,


CA) nebulizer which consists of a pair of concentric tubes.


While the position of


the inner capillary tube is adjustable on other commercial interfaces (Apffel and


Nordman


,1987),


its position is fixed in the Meinhard nebulizer


. The nebulizer was


installed onto the desolvation chamber through a Cajon fitting which ensured

a vacuum-tight seal.

The HPLC eluant entered the nebulizer by passing through a length of 0.1


deactivated fused silica capillary column,


which was


inserted into the


inner tube of the nebulizer


. A high flow (


-0.6 L/min.)


of He was


introduced


through a gas port and flowed through the outer tube.


the inner capillary,


As the


iquid flow exited


the concentric flow of He dispersed the liquid flow


nto an


aerosol consisting of a distribution of micron-sized droplets (Browner et al.


1986).


The aerosol was then swept into the desolvation chamber by the flow of He.


The flow of He into the nebulizer was controlled


pressure of the regulator on the He supply.


by varying the


The plots in Figure


head


show how


varying the head


pressure


of the


He (nebulizing


flow)


affected the signal


different mobile phase compositions.


The data points represent the average peak


areas obtained from triplicate injections of 20 ng diuron.


The peak areas were


calculated from the extracted ion profile of the base peak (m/z 72) of the El mass


spectrum of diuron.


The curves are asymmetric and show that Door sianal


were


,













500000


400000

+

300000




200000


100000




0


- 100
,aa-&a 75/25
*^-- 50/50


CH3CN
CH3CN/H20
CH3CN/H20O


I
I
/
ti


- . . U tEl UptUttEl I * III..
iiu..i.i U~I~I' III I
I U U


1 1 1 120 I
20


He


Regu


ator


Pre


60


ig)


Figure


Plots of peak area vs.


nebulizing gas


mobile phase compositions.


triplicate


injections


standard deviation.


head pressure for different


Data points


diuron.


represent


Error


bars


averages
indicate


I IIIIIIII


) 80


I


IIIIII lilllllIl IlllIlllIl ii








primarily composed of low velocity drops with


arge mean diameters which are


difficult to desolvate.


high


pressures,


smaller


, high


velocity


droplets


formed.


Significant transport losses result from surface impact, turbulence, and


evaporation when the aerosol droplets are either too large or too small


et a.


(Browner


, 1982).


The plots also show that, as the percentage of water in the mobile phase

increases, a higher pressure (flow) of helium is required to produce optimum


performance.


This trend


related to the surface tension and viscosity of the


mobile phase.


Aerosol


are generated


because the


gas jet causes frictional


interactions with the liquid surface.


The main forces opposing aerosol generation


are the surface tension and the viscosity of the liquid phase.


much higher surface tension and is more viscou


Since water has a


than organic solvents, more


energy (hence higher gas flow) is required to efficiently nebulize aqueous solvents


(Browner et aL


1986).


It also appears that sensitivity decreases as the percentage of water in the


mobile phase is


increased.


Even at optimum conditions for each mobile phase


composition,


phase.


the peak areas were greatest when using a pure organic mobile


This observation can be explained by considering both the nebulization


desolvation


processes.


Similar studies


involving


characterization


of the


desolvation


chamber


temperature


showed


sensitivity


decreased


as the


percentage of water


n the mobile phase increased.


Since water has a much








phase


(Browner


, 1986).


Desolvation


aerosols


comprised


larger


aqueous droplets is more difficult because the aerosol has less surface area and

water also has a higher heat capacity and AH- than other common LC solvents


(Weast,


1982).


Indeed


other studies have demonstrated that the observed order


of sensitivity for different mobile phases is inversely related to the energy to heat


and vaporize the solvent (Voyksner et al.


1990).


It does not appear for this PB interface that the flow rate of the mobile

phase affects the optimum nebulizing gas flow as much as the mobile phase


composition.


Figure 2-3


shows nebulizing gas optimization curves for different


flow rates of 100% acetonitrile.


All curves indicate optimum conditions at 30-35


psig


head


pressure.


Although


optimum


conditions


vary


significantly


the best sensitivity was observed at lower mobile phase flow rates.


It was also


observed


that the


optimum


nebulizing


flow rate was


largely


compound-independent.


Desolvation Chamber


Temperature


Once


aerosol


generated,


droplets


are swept


a heated


desolvation chamber.


The main function of the desolvation chamber is to remove


the volatile solvent molecules from the


aerosol droplets.


Solvent evaporation is


achieved by the transfer of heat from the chamber walls to the


drops.


Any analyte


molecules which are present are typically less volatile and remain as particles.













1000000-


800000


600000


400000


200000




0


Figure 2-3:


A*a6a 0.3


I \
,
I
I
I


I
/
/
^I
.-)C


11111120 1
20


He


mL/min.
mL/min.
mL/min.


$ \ I


30
Regu


Plots of peak area vs.
mobile phase flow rates.


ator


Pre


nebulizing gas


head


70


80


ig)


pressure for different


Data points represent averages of triplicate


injections of 20 ng diuron.


iiiiiiniTTlililllnillllllltii








The desolvation chamber in this interface was an 8" long,


hollow


nickel-


coated aluminum


block


(Figure


1-4).


It was heated


by two 300 W


cartridge


heaters inserted


nto the chamber block.


The temperature of the desolvation


chamber was


controlled


an extemal


Ogden


(model


ETR


9080,


Arlington


Heights,


L) feedback temperature controller which used a thermocouple to sense


the chamber temperature.


Figure 2-4 illustrates the


effects of desolvation chamber temperature and


mobile phase composition on the signal obtained for 20 ng of diuron.


The data


points represent the average areas obtained from triplicate FIA injections.


plots reveal that the optimum temperature of the desolvation chamber depends


on the composition of the mobile phase. Thb

the percentage of water in the mobile phase


heat capacity and AH, than organic solvents,


e optimum temperature


increased.


increased as


Since water has a larger


more energy is con


umed during


vaporization.


Thus, higher temperatures are needed to replace the heat lost from


adiabatic expansion due to solvent evaporation to maintain efficient desolvation.


The plots also show that signal


decreases as the percentage of water in the


mobile phase increased.


As mentioned previously,


this trend is probably a result


of both the nebulization and desolvation processes.


appears


from


plots


Figures


optimizing


desolvation chamber temperature was not as critical as optimizing the flow of


nebulizing gas.


While the signal obtained at non-optimum desolvation chamber












900000


800000


700000


600000


< 500000


0 400000
C0

300000


200000


100


CH3CN


__ -___
I--r 75/25 CH3CN/HO0
/


50/50


I I I I I1 II


Deso


30 40 50


Chamber


Temp


CH3CN/H20


60
(c)


Figure 2-4:


Plots illustrating the effect of desolvation chamber temperature on
peak areas obtained for FIA of 20 ng diuron at various mobile phase
compositions.









nebulizing g

(Figure 2-2).


las flow rates resulted


n as much as an 80% decrease in signal


However, this was not true for all the test compounds used.


Figure


2-5 shows desolvation temperature optimization curves for diuron and carbaryl.

Each curve was normalized with respect to its largest data point to better illustrate


the loss of signal at non-optimum conditions.


Although the peak areas for diuron


not vary by more than


20%,


peak areas for


carbaryl


at higher,


non-


optimum desolvation chamber temperatures decreased by as much as 80% with


respect to the maximum peak area.


Since carbaryl is more volatile than diuron,


at higher temperatures carbaryl was probably being vaporized in the desolvation

chamber and pumped away in the momentum separator region of the interface.


Although of a limited test set, Figure


chamber temperature is largely compound


also illustrates that optimum desolvation


independent as both curves exhibit


maximum signal at the same temperature.


Desolvation Chamber Pressure


The pressure in the desolvation chamber is largely determined by the flow


nebulizing


mobile


phase


flow


rate.


optimum


nebulization conditions


, the pressure in the desolvation chamber is typically 200


torr.


pressure


regime,


imiting


process


influencing


solvent


evaporation is the


rate of heat transport from the


surrounding gas to the


surface


of the aerosol droo


(Rrnwner


19R86


Helium


is used


WIiW *W I V* I I Y -


as the nebulizina


2-M













1.0



( 0.8



0
0Y


I
20


Desolv.


Diuron
(m/z 72)


Carbaryl
(m/z 144)


I I II I I I I I I I I 1 1 II
30 40 50


Chamber


Temp.


I I I I I III
60
(c)


Figure 2-5:


Normalized desolvation chamber temperature optimization curves
for diuron (m/z 72) and carbaryl (m/z 144) using a 100% acetonitrile
solvent at a flow rate of 0.3 mL/min.








aerosol


drops.


order


provide


a better


thermally


conductive


environment, supplemental He was added to the desolvation chamber to enhance


solvent


vaporization.


Helium


was


introduced


a Negretti


(Southampton,


England) fine metering valve through a Swagelok (Jax Valve, Jacksonville,


installed on the desolvation chamber


. A pressure gauge (Omega, Stamford,


CT) was also added to monitor the pressure of the desolvation chamber (Figure

1-4).


Figure 2-6 shows the effect of adding


supplemental He


to the desolvation


chamber.


The mobile phase for th


studies was 100% acetonitrile; the data


points represent the average peak areas obtained from either 20 ng caffeine or


100 ng carbaryl.


Optimization of the nebulization process accomplished prior to


these studies resulted in a desolvation chamber pressure of


210 torr.


The plots


indicate that a two-to-threefold increase in signal was obtained by


pressure in the desolvation chamber by approximately 120 torr


plateau was observed for both compounds.


increasing the


at which point a


These results are consistent with


those observed by other laboratories (Browner,


1986).


interesting to note that


commercially available


supplemental He


nterfaces of similar design do not utilize the addition of


n the desolvation chamber.


Target Temperature


After travelling through the momentum separator


, the analyte-containing



























Figure 2-6:


Plots illustrating affect of adding supplemental He to the desolvation
chamber for FIA of (a) 20 ng caffeine and (b) 100 ng carbaryl.










Desolv.


Chamber


Pressure


(torr)


80000

+
'*70000


V60000


50000


40000


30000


Amount He Added (torr)


Desolv.


Chamber


Pressure


(torr)


210
100000 J


90000


80000

70000

60000

50000

40000

30000


r-


210
900004^








particles


enter the ion trap where they strike the hyperbolic surface of one of the


endcap electrodes which served as the PB target.


The collis


on with the endcap


causes the particles to undergo a rapid flash vaporization.


This results in the


production of predominantly


ntact gas-phase analyte molecules.


The analyte


molecules are then ionized by E


, or alternately,


byC


with the


introduction of a


reagent gas.


The temperature of the PB target (endcap) plays an


important role


n the


signal obtained for PB analyses.


The flash vaporization process must be efficient


to prevent peak tailing,


which requires sufficiently high temperatures.


However,


thermal degradation of some compounds can occur if too high temperatures are


utilized.


Typical target temperatures used for PB analyses are between 250 and


3000C.


initially


the temperature of target endcap was maintained by radiatively


heating the entire


SM vacuum chamber


, including the ion trap and electron


multiplier, with the standard quartz heaters mounted


In order to prevent damage to the electron multiple


inside the vacuum chamber.

ier, temperatures were not


raised above 140C.


This resulted


n bad sensitivity and a large degree of peak


tailing even for FIA because these


ow temperatures yielded an


efficient flash


vaporization process.


n order to permit the use of typical PB target temperatures,


modifications


were made to the filament endcap electrode which served as the target.


Two


large holes were drilled into the body of the endcap to insert two 10W cartridge









to sense the endcap temperature.


The temperature of the endcap was controlled


by an external feedback Omega (model CN-2012,


Stamford


, CT) temperature


controller.


These modifications allowed the endcap to be heated to typical PB


operating temperatures of up to 290C while maintaining the electron multiplier
mounted within the vacuum chamber at normal operating temperatures of 1000C.


improvements these modifications brought


about are illustrated


Figure 2-7


. Shown


are mass chromatograms


of m/z


from triplicate FIA


injection of 200 ng carbaryl at endcap temperatures of 125 and 2500C.


temperatures,


the peaks exhibited a significant degree of peak tailing caused by


inefficient flash vaporization of the particle beam.


At higher temperatures,


peak


shapes were dramatically


observed.


The signal intensities


improved and little evidence of peak tailing was


hown in the figure also show that there was an


approximate three-fold


increase in peak height at the higher temperature.


Studies


characterizing


target


temperature


also


indicated


optimum temperature was largely compound-dependent.


Optimization curves for


the FIA of 100 ng carbaryl and 75 ng rotenone are shown


n Figure 2-8.


The plots


are normalized with respect to the maximum peak area in each data set for clarity


Data


points


represent


average


ion peak


areas


obtained


from


triplicate


injections.


While


maximum


response


carbaryl


occurs


at temperature


200oC


290C.


rotenone exhibits a maximum


at much higher target temperatures of


Rotenone is much less volatile than carbaryl,


so it is not surprising that






























L..

1L


0)
C




(U
W4-


0
0

.0

U,




o g
N
n

04-'
10 $



0 0
Og


(00L.



00 (
4-
Ef


Sc~a

SQ)
0)0
.-5

^-a
(0 (0

WO
Sa
[d L

(0-
eQQ





























cl--
0
Cl)


Ai!sueiul

















0
( 0.8
L.



( 0.6


-o
N 0.4


180


Carbaryl
(m/z 144)


Rotenone


(m/z


200


192)


'''l220ll
220


End cap


Il Ill' II


240


II II II I


26(


I I
0


Temperature


I 111111 II
280
(c)


300


Figure 2-8:


Target temperature optimization curves for carbaryl (m/z


144) and


rotenone


(m/z


192).


different


maxima


curves


exhibit


indicate that the optimum target temperature is largely compound-
dependent.


^*-






44

The carbaryl optimization curve illustrates that higher temperatures are not


optimum for all compounds.


This is because thermal decomposition of some


compounds


can


occur


high


target temperatures.


carbaryl,


best


response


of the 1 44 fragment ion (base peak) occurred at 200 C and decreased


by approximately 50% at target temperatures of 290C (Figure


2-8).


However, it


was observed that the total ion current signal


increased by approximately 20%


when the target temperature was increased from 200 C to 290 OC.


This indicated


carbaryl


likely underwent


some


degree of thermal


decomposition


which


limited the sensitivity of the m/z


144E


fragment ion at high target temperatures.


Conclusion


studies demonstrate the need for


optimization


of important


operating


parameters to


intelligently


develop effective


LC/MS


methods.


nebulizing gas flow rate, desolvation chamber temperature,


desolvation chamber


pressure,


and target temperature have al


been shown to affect the


sensitivity of


the LC/PB/ITMSTM system.


Most of th


interface parameters were found to be


largely mobile phase dependent; however,


some compound dependance was


observed


, most notably for the source target temperature.


mobile phase


dependance wil


make it difficult to optimize interface parameters for gradient


elution analyses because the mobile phase composition will change during the


analysis


Modifications to the ion trap endcap electrode were necessary to allow











CHAPTER 3
EVALUATION OF STRATEGIES TO MINIMIZE ADVERSE
EFFECTS OF RESIDUAL SOLVENT


introduction


Although ion traps are noted for their sensitivity


versatility


and potential


cost,


benchtop


currently


instrumentation


LC/MS


(including


, most


those


mass


using


spectrometric


interfaces)


systems


employ


quadrupole mass analyzers.


primarily


Ion traps have seen limited use


because of the large amount of solvent introduced


n LC/MS systems

by most LC/MS


nterfaces.


Excess solvent ions and neutrals typically cause space charging and


undesired


on/molecule reactions,


resulting in


poor mass resolution and poor


overall spectral quality.


n this chapter are reported several strategies which were evaluated,


aimed


at minimizing adverse effects caused by introduction of residual solvent from a PB


nterface.


These strategies


included one method which reduced the


amount of


solvent which reached the trap.


The versatility of the


ST also allowed the


creation of customized scan functions which ejected unwanted solvent ion


from


within the ion trap prior to mass analysis, while efficiently storing analyte ions of


interest.


Although the coupling of PB with an ion trap using both direct coupling








1991


, 1992),


currently


no information


appears


in the


literature


comparing


operational modes of the ion trap which minimize space charging by ejecting


residual LC solvent ions prior to mass analysis.


determine the best way to operate the


The goal of this work was to


STM to produce quality El spectra from


analytes


introduced via LC.


El spectra obtained from isocratic LC/PB/ITMSTM


analyses of simple pesticide mixtures (


_3 components) are compared to library


El spectra and


El spectra obtained from solids probe/ITMSTM


"solvent-free" method) of pure compounds.


Also


analyses


instrument calibration curves


and limits of detection (LOD) obtained from these isocratic LC/MS analyses on

this system are reported.


Experimental


instrumentation


Liauid


Chromatoaraphv.


Two


LC pumping


system


were employed


studies.


Solvent ejection studies utilized an ISCO (Lincoln,


NE) LC-2600


syringe pump.


Chromatographic separations were performed using a Hewlett-


Packard


(Palo


Alto


, CA)


1090L


high-performance


liquid


chromatograph.


Separations were performed on a 100


x 4.6 mm Hewlett-Packard octadecasily


column


, packed with 5 pm particles.


Separations were accomplished isocratically


with a mobile-phase composition of 80/20 acetonitrile/reagent water at a flow rate


of 0.3 mL/min.


The analytical column was thoroughly conditioned by pumping






47

a 50/50 acetonitrile/reagent water mixture through the column overnight to remove

residual impurities and column bleed.


nterface.


The LC/MS interface was a prototype Finnigan MAT


(San


Jose,


three-stage


particle


beam


interface


(Figure


1-4).


operating


princip


optimization


of the interface are described


greater


detail


Chapter 2.


important PB operating parameters including desolvation chamber


temperature and nebulizing


sensitivity.


He flow rate were optimized to provide maximum


The desolvation chamber pressure was optimized by addition of He


increase


pressure


approximately


400 torr


probe


was


positioned about /4" away from the


on trap entrance to prevent high pressures


inside the trap.


Spectrometer


. The mass spectrometer used


experiments


was a Finnigan MAT (San Jose, CA) ITMS


Fundamentals of ion trap operation,


theory


and ion motion are described in greater detail in Chapter 1


. The


TMS


vacuum chamber was maintained at 100C for all experiments.


Chapter


As described in


, the filament endcap of the ion trap was modified to allow insertion of


10W heater cartridges which were controlled by


an external temperature


controller (Omega Engineering


Model CN 2012,


Stamford


, CT).


This endcap


served as the particle beam target and was operated at temperatures of 250-


290 C.


Electron


ionization was employed within the ion trap;


detection was


.. a








typically


x 10 torr (uncorrected) as measured by a Bayard-Alpert ionization


gauge (Granville-Phillips, Boulder, CO) mounted on the vacuum chamber. Helium

was added into the vacuum chamber to produce typical operating pressures of


x 10" torr (uncorrected).


TMST was repeatedly scanned from 55 to 500


amu to check for formation of dimers and ion/molecule reaction products


Axial


modulation


(Weber-Grabau et al.,


1987)


at 530 kHz and 6V


was employed


during the analytical scan in al


studies to reduce the effects of space charge and


improve sensitivity.


Samples


and Reaaents


Methomyl and


nuron were obtained from AccuStandard (New Haven, CT),


caffeine from Sigma (St. Louis,


MO); al


other pesticide standards were obtained


from the U


Environmental Protection Agency (Pesticide Chemical Repository,


Research Triangle Park, NC)


All standards were used without further purification.


The structures of the standards used in these studies are shown


in Figure 3-1.


Also shown are molecular weights and the El fragmentation which yields the


base


peak of each compound's


mass spectrum.


Structures


, molecular weights and El


fragmentation for carbaryl,


rotenone, and diuron were shown previously in Figure


HPLC grade methanol and acetonitrile were obtained from Fisher


ientific


(Pittsburgh,


PA);


reagent water was obtained from


a Milli-Q water purification


system.












I
Z-I


sVwwvw^


WVWVV W


m-.1-


O r
(ho

< s


nc
(O~
--
**
Q.


.- U)
* o

I
("1
0) ^
t0U)


I
Z-I
I
o= -
'O -
gvwvvwuw
GD 0


.c *
E C
05 -


Z-- I


swCvVvw^
+


o 0=0

Z-I


CO
Da3
0)


I

II
0 I
If
0-0


JVw/VVwV


0=c =








Problems Caused by Residual Solvent


The major challenge encountered when interfacing LC directly with the ion

trap was minimizing adverse effects caused by the introduction of residual solvent


from the interface.


Solvent molecules (S) in the trap are ionized by El:


-* S+'


resulting


a large


population


solvent


ions


which


caused


space


charge


conditions.


Space charge results when the density of ion


n the trap becomes


large enough that ion-ion interactions become significant.


This can distort the


trajectories


of trapped


ions


resulting


spectral


quality with


poor


mass


resolution and


correct mass assignments.


In the ion trap,


ions and neutrals are present within a confined space for


a period of time prior to ejection/detection.


As a result,


a large population of


solvent ion


and molecules can also cause undesired ion/molecule reactions to


occur


Most notably


a large degree of solvent-C


of analyte neutral


can occur


as illustrated below:


(S + H)+


(S + H)+


-H)


(M + H)+


These equations show that a solvent radical ion


(S+) can react with a neutral


a a a. S a ..1 -1 .t-a


^ -* -1


1 rl ~ 1 .








proton transfer will occur and a protonated molecule, (M+H)+


forms.


This is


typically


a "soft"'


ionization


process,


meaning


that the


(M+H)+


ions


typically


undergo little fragmentation as compared to the M'*"


ions initially formed by El.


Since the primary advantage of using PB is the ability to acquire El mass spectra,

these adverse effects of residual solvent must be minimized.

These adverse effects are illustrated in the profile mass spectrum shown


n Figure 3-2.


The mass spectrum was acquired from FIA of


-100 ng of carbary


n 100% methanol mobile phase.


The spectrum shown is an expansion of the m/z


25-150 mass range to better illustrate the resolution between adjacent masses;


the abundance of the M'

abundance of methanol


' ion at m/z 201 was

solvent ions centered


<5% relative to


around


145*


m/z 33 caus


The large

ed space


charging which resulted in poor mass resolution between adjacent low masses.


Notice that resolution


improved at higher masses.


nset


Figure 3-2 is


provided to better illustrate the mass resolution between


144


145


. As


discussed


Chapter


TMSM


generates


mass


spectra by


sequentially


ejecting trapped ion


in solvent ions being


from low to high mass (Stafford et al.


ejected/detected


1984).


This results


before higher mass ions during


mass


analysis,


reducing space charging when higher mass


are detected.


144*/145+


ratio in this spectrum


also indicated


a large degree of


solvent-C


occurred.


The base peaks of the El and Cl mass spectra of carbaryl


are 144+


(M-57)*


and 145* (M+H-57)*


, respectively.


Although some 145" from































0
t 0)
a o
(0 C
0 0


U-
O 0)

< Q)
U- Sc


a)






O
C





O
O)
0






O
15





or
c-
a+
0
0








6 -




O c
I -
+'
05 (


0w
a -
---r ^


*o

(O


i -

Wo E













































I -7

_ .-


I--b


-IS
-4^









-04


-ai
a^


:CO
-r ~


_I


r


! I


(Qsi)


Aiisueiul


CI
=r


-----


.--_-_


&


I~


W








carbaryl (11 % relative to 144"),


the figure shows the intensity of 145+ being about


three times greater than


that of 144k


The abundant methanol


solvent ion


evidently


caused


a large amount


of solvent-C


to occur


ndeed


the use of


residual solvent ions to perform solvent-CI analyses of pesticides with the same


LC/PB/ITM


Sm system has been demonstrated (Eades et aL,


1992).


However,


here the large degree of solvent-CI was undesirable since the goal of this work


was to produce quality El spectra.


prior to mass analysis to minimize


This figure illustrates the


need to eiect solvent


space charging and solvent-C


to obtain


good quality El m


Stratea


spectra.


to Minimize Adverse Solvent Effects


different types


adverse effects of residual


strategy


which


solvent are illustrated


used


in the


block


to minimize


diagram of the


LC/PB/ITM


STM system


analyte molecu


in Figure 3-3.


(S and A


The block diagram shows that solvent and


, respectively) elute from the


HPLC and pass through


the PB


nterface


into the ion trap where solvent and analyte ions (S+


and A+) are


formed by E


The number of solvent molecules which reach the trap could be


reduced


by using


various


nstrumentation-based


approaches


before the trap.


Alternately


the trap could be operated such that solvent ions are ejected from


within the trap prior to m


analysis.


By employing these strategies,


ideally only


analyte ions of interest eventually reach the detector




















+
O)














0
(4


(Ac
ID>
>6
Vy)
Cu
I.-


ci
'U


E N .0-,
oE .
N I-
o a
E3
o
.2 2


S-


a )
*1 s
CO C
w rc
o .w


a)

30
)1:
(U"
WI


Eut
-0~



O-
s c
0-
Ama


O



c,,








As mentioned previously, the


nterface used in these studies employs an


additional


third


stage


momentum


separation.


additional


stage


momentum


separation


is an


example of one strategy


aimed


at reducing the


number of solvent molecules which reach the trap.


The amount of solvent which


reaches the trap could also be reduced by employing a membrane drier


n place


of the desolvation chamber.


In the membrane


volatile solvent diffu


through


membrane


pumped


away


while


involatile


analyte


particles


transferred to the momentum separator.


The commercially available Universal


nterface from Vestec is based on this operating principle (Vestal et al.,


1990).


Solvent


molecules


also


prevented


from


reaching


trap


using


injection of ions created


n an external ion source.


been demonstrated using both on-axis (Louri


et al.


Previously


injection has


,1989) and off-axi


(Pedder


et al.


, 1989) ion sources.


Different


TMSTM


operational


modes can


also


implemented to


eject


solvent ions


from within the trap.


Previous


research


our labs


here at UF


investigated method


for mass-selective ionization in the quadrupole ion trap to


eliminate undesired ion


of a range of mass/charge (m/z) values


(Eades et al.,


1991).


These


included the use of resonant


excitation


employing


waveform


comprised of multiple frequencies.


However


, implementation of th


methods


required modifications to the


ion trap operation


STM source code.


implemented with standard


In this chapter, two modes of


TMST software are evaluated for






57

analysis: 1) elevation of rf voltages to eject ions below a chosen m/z ratio (i.e.

impose a low mass cutoff) and 2) application of a supplemental rf signal to eject


a single solvent ion species.


Both methods were evaluated when applied either


during ionization or during an ion storage period following ionization.


A summary


of strategies which may minimize adverse solvent effects is shown below in Table


Table 3-1


Summary of Strategies to Minimize Adverse Solvent Effects.


Remove Solvent Molecules Before the


Trap


. Membrane Drier prior to Momentum Separator
. Additional Third Stage of Momentum Separation in


External Source and Ion


nterface


injection


On-Axis
Off-Axis


Eject Solvent Ions within the Ion


Trap


Elevate rf Voltage (Impose Low Mass Cutoff)


During


onization Step


b. During Ion Storage
Resonant Ejection of Single I
a. During Ionization Step
b. During Ion Storage


Species (Notch Filter)








Additional Staae of Momentum Separation


The same PB interface was used to determine the benefits of using the

additional third stage of momentum separation to reduce the amount of solvent


which reached the trap.


The interface was converted to a two-stage interface by


removing the third skimmer and third stage roughing pump from the interface.

Pressures in the momentum separator region were approximately 5-10 torr and


400-500 mtorr is stage one and stage two,


respectively.


pressures are not


significantly


different


than


those


observed


when


employing


three


stages


momentum separation.

Figure 3-5 illustrates that the additional stage of pumping does reduce the


amount


solvent


which


reaches


trap.


Shown


are the


ntensities


protonated acetonitr


e (m/z 42) at different flow rates of 100% acetonitrile mobile


phase.


figure


demonstrates


solvent


intensity


increased


with


increased flow rate and that about four times


less solvent ions were formed when


using


three stages of pumping.


Table 3-2


ists the ion


gauge pressures for


various


solvent compositions for both configurations.


obtained at a flow rate of 0.3 mL/min.


and with no


pressures were


supplemental He added to the


desolvation chamber


STM vacuum chamber, or third stage of the momentum


separator


pressures


were


read


directly


from


gauge,


with


correction factors


used.


The table indicates that the partial


pressures


of the














1400

1200


1000

800

600

400

200

0


0.1 0.3 0.5


Flow Rate (mL/min)


Figure 3-5:


Comparison


various


flow


of intensities


rates


using


of acetonitrile


both


(M+H)+


a two-stage


solvent ions
three-stage


interface.


I 3-Stage 2-Stage









Table 3-2:


Comparison


gauge


pressures


various


solvent


compositions using two and three stages


of momentum separation


Ion Gauge Pressure (torr)


100%
CHOCN


2-stage

3-stage


1.9 X1 0i

5.8 X 10


75/25
CHCCN/H,_O
19X 10


1.9 X 104

5.7 X 106


50/50
CHCN/H1O


2.8 X 104

5.7 X 104


25/75
ChLCN/HO


3.2 X 104

3.5 X 10-5s


All pressures obtained at solvent flow rate of 0.3 mL/min. and no


He added to desolvation chamber


supplemental


. Pressures were read directly from the ion


gauge meter with no correction factor used.


The performance of both configurations was evaluated by comparing the

El spectra obtained for FIA of 20 ng caffeine in 100% acetonitrile mobile phase.


The spectra are presented


n Figure 3-6.


spectrum obtained from a computer


Also


brary


shown is a library reference El


included with the


TMST software.


spectrum


obtained


using


three


stage


nterface


shows


excellent


comparison with the


library spectrum with respect to both fragmentation pattern


and relative


ntensities of the m


peaks.


The spectrum obtained using two


stages,


however, does not compare favorably with the library spectrum.


. .. N


-8 S C a 8


f 8


- -


- -. .Y -LL~ .;- -r 1 r -. a a a -l a ar a~ a* *l aC a a a sa ;- ~ au C fl S














I- -0
V
mC
ap
-on

CO
00


Q Q-

'=CO
a)
L 0
So
t

*" .


S0)
E


'CO
10
*6^


st
O5 *
C
(uIJ, *-

) WO
0- i
wn n


0 0) LI.




E -

* m
w0<
Oma 0.

CO
Sa>


cm


cv,
0)
























_^Y
~cmn
'-4


N



SD--


*\T I 1
'-4

(p


O -


N-
03-


(D --


SI I I I I


-Co





-N



-03


+


+a)
S^-


r"


-. "


-CM
-NC\
-1


a,
O -


-to
-0
- 7"


-0
CO


cN
0--


_ I
CD


I 1 1 1 1


-LO
-N
-(N

L0
-0
-CN


-Lo
-0L




-CD
- r



- ft


10
-CN
-1 cU


NC)

tLO
-0






cx,



1CD




it )
4






63

activated dissociation (CAD) to occur from collisions with background solvent


molecules.


Also


, the most abundant molecular ion peak was the (M+H)+


peak


at m/z 195 which


indicated a


arge degree of solvent-C


occurred.


These figures


illustrate


that


additional


third


stage


momentum


separation


removed


significant amounts of solvent and improved the performance of the LC/PB/ITMST

system.


Elevation of rf Level


One method


of solvent ion ejection studied


involved elevation of the rf


voltage applied to ring electrode to eject ions below a chosen m/


to impose


a low mass cutoff.


In this mode, the ion trap was operated


n the


rf-only mode


with


Mathieu


parameter


=0).


Since


Mathieu


parameter


proportional to the rf voltage (Equation 1-5),


elevation of the rf voltage increased


the q, values


of trapped ions.


In the rf-only mode, ions with values of q, greater


than 0.908 wi


(z) direction.


follow unstable trajectories


and be ejected from the trap


Since q, is inversely proportional to m/z


(Equation 1-5),


n the axial

ions from


low molecular weight LC solvents (e.g.


CH3OH,


CH3CN,


H20) were ejected at a


properly


chosen


low rf potential while analyte ions of higher m/z


remained


stored.


In this mode, the ion trap was basically operating as a high-pass mass


filter.


The effect of elevating the rf voltage upon both solvent and analyte ions is












oa cm
OtC T



O-S -




CL.q
oo, c
cDCS 0 I

*n~ y) J





t o.2
O r


maga
Eo
N-En C-

4~- 4-'


e-
- tf **o -5







Ew D t
S- -*









-ces
O -
|^%S
C*-






a $ iO



^.^ w m~
1(8. O S *-


-, cl
N-. CU


E w o
N3 0'CV.



.. o c ca
C O0 0)L.
E-4-'-



#ce

o- a), 4
* ~o Ci
-Ct



cowoc
-._ S' wE'






>5 (0c-ja
+*g 0- T0
^ mo -s
u *B > *B

CO O O>4a

---CS
* U)- 0
0) 0)1 aStfl





-
4-' 0 -
Rests' s
4- ...

+* O *


0
L.

















0
0

%- 0


00


Cn

C_



O-
O

o


6D 1jOA Op


aO




0+

OL








solvent (m/z


42) and the most abundant fragment ion of diuron (m/z


Note


here that the stability diagrams are plotted in terms of actual rf and dc voltage


applied to the electrodes, rather than a,q space.


rf-only mode, the rf voltage determines the


Since the trap is operated


ow mass cutoff of the trap,


n the


which is


also shown.


At an rf level corresponding to point A in the figure, both the solvent


and analyte ion will be trapped because this point is located within the stability


diagram of both ions.


Point B is located outside the solvent


but is still within the stability diagram of the analyte.


on stability diagram


At this rf level


the solvent


on wil


be ejected while the analyte ion will remain effectively trapped.


This mode of operation was investigated when applied both during the


ionization event and afterwards


, during ion storage.


The effect of raising the rf


level (low mass cutoff) upon the ion intensity of some commonly used LC solvents


is shown


n the plots of Figure 3-8.


As expected,


as the instrument's


low mass


cutoff


approached


each


solvent


species,


solvent


ions


were


efficiently (100%) ejected from the ion trap.


These plots were created by raising


the rf level after ionization


, during ion storage.


Although not shown,


similar plots


were obtained by elevating the rf during ionization.


Figure 3-9


Ilustrates how the


intensity of the carbaryl analyte ion (144+)


was affected by elevating the rf


evel (q,) both during


onization and during ion


storage.


These plots were obtained by varying either the ionization or storage


values of q, with a constant amount of carbary


(2 ng/p4L yielding 10 ng/s


at 0.3

















19+
H30O


33+


CH3OH2


42+
CH3CNH+


0 10 20 30 40 50


Low


Ma


Cutoff


Figure 3-8:


Plots of normalized solvent


on intensities


vs. rf


evel (expressed as


low mass cutoff) during ion storage illustrating efficient ejection of


common LC solvent


ons.


m/












0.00
' 1


0.20


I I. I I I I I


0.40


III III IIII


0.60


I iiIIIII


0.80


I lJ III II


IaI III III


+
4 0.8


Elevate rf
\ During Ion
vI Storage

1i


Elevate


During


onization


S.-


I I I I I I I I I


Low


I I I I I I I I 1 I
bco


Ma


Cutoff


I I I I I I I I I 50
150


(m/


Figure 3-9:


Plots


intensity


144+


fragment


carbaryl


rf level


(expressed
ion storage.


as low mass cutoff) both during ionization and during








signal


was


relatively


unaffected


storage


until


stability


limit was


exceeded


, signal decreased significantly with increasing ionization q,.


As the


value


q, increases,


so does the


initial


kinetic


energy


of the


ions formed,


resulting in large initial velocities and ion losses due to quasi-unstable trajectories


(Dawson,


1976).


These losses occur when the magnitude of an ion's


oscillation


exceeds


ntemal


dimensions


of the


trap


even


though


mathematical


conditions for stability still exist.


Although raising an ion's


q, value during ion


storage also increases its kinetic energy,


the ions are already effectively trapped


so ion


losses


are minimal


until


ions


become


unstable


when


qz=0.908.


Elevating the


onization value of q,


of an ion also decreases the


initial ionization


volume which limits the region


n which ions can be created and remain stable


(Dawson,


1976).


The observed decreases


n intensity could also be the result of


decreased ionization cross-sections due to high electron energies at elevated rf


levels.


Preliminary studies investigating the effects of ionization rf level (q2) of n-


butylbenzene ions also showed decreased ion


(Pedder et al.


ntensities at higher values of q,


1992)


There were also some differences


n the extent


of solvent-C


between


spectra obtained by ejecting solvent during


ionization and during ion storage.


Figure 3-10


shows how the ratio of 144+ (El fragment) to 145


fragment plus


'"C isotope of 144+)


is affected by elevating the rf


during


ionization and


storage for carbaryl


n 100% methanol.


The rf level where the methanol solvent












Eject


+


0
--+- 2.0
CD


Solvent
Ions:


Elevate rf during
ionization




Elevate rf during
Ion Storage


III IIII I III IIIII II


SI I II II I I I I I I I I I I I I I I I I I I
) 40 50 60


'I l70l
) 70


Low


Ma


Cutoff


Figure 3-10:


Plots of ratio of 144'


(expressed as


(El fragment) to 145+


ow mass


cutoff) both during


ion storage for carbaryl in 100% methanol.


fragment) vs. rf level
onization and during


Dashed


indicates rf


level where methanol solvent ions are efficiently ejected from the
trap.


m/








in the figure.


In a "pure" El spectrum,


144+/145+


intensity ratio would be


expected to be 9.0,


based on the natural abundance of 'C isotopes of 144'


.The


low ratios obtained at low rf levels resulted from solvent ions causing a


arge


degree of solvent-CI.


Ejection


of solvent ions


decreased


the amount


of CI


resulting in increased 144C/145* ratios; the inability to achieve the expected ratio


for a pure El spectrum indicated that some solvent-CI still


occurred in the time


before solvent ions were ejected.


Also


lower ratios were observed when the rf


was elevated during ion storage.


the ionization time,


This is because solvent ions were stored during


which allowed more time for solvent-CI to occur before the


solvent ions were ejected.


Resonant Election of Sinale Ion Species


The other method of solvent ion ejection studied involved resonant ejection


of a single m/z


solvent ion species by applying a notch filter (Kelley et al.


1985).


This was accompli


hed by


applying


a supplemental


gnal


(resonant


excitation voltage) across the endcap electrodes to kinetically excite a particular


solvent ion m/z species


such that it was ejected from the trap.


The trajectories


of trapped ions of a particular m/z are generally characterized by their unique

secular frequencies, wz, which are a complex function of the Mathieu parameters


az and q,, and 8z,


which relates to the frequency of the oscillation of the ions.


Application of the resonant excitation voltage at the wz of a particular solvent ion


6V p








oscillations.


The excited ion can then undergo CAD with residual buffer gas


molecules.


Alternatively,


and of relevance to this discussion,


the excited ion


species can also be ejected from the trap if sufficient power is absorbed such that


the amplitude of the oscillations become larger than the trap's


dimensions.


use of resonant ejection was investigated because it permitted ionization to be


performed


lower


values


which


may


result


better


sensitivity


compounds whose El mass spectra have low m/z base peaks.


As with elevating the rf amplitude, the use of a notch filter was


investigated


when


applied


either


during


ionization


or ion storage.


Figure 3-11


illustrates


resonant ejection of the m/z 42 ion of acetonitrile solvent both during ionization


and afterwards


, during ion storage.


The figure shows that m/z 42 was efficiently


100%) ejected when resonant ejection was applied during ion storage.


Since


ions of different m/z p

signal were observed.


'ossess different secular frequencies,


no losses in analyte


This process was not nearly as efficient, however


when


applied during


the ionization


pulse.


In this case only about one-third


of the


solvent ions were ejected.


This is because the ionization and ejection of ions


were acting


as competing


processes.


Ions were constantly being


produced


during the ionization event; those created towards the end of the ionization period


were only resonated for a very short period of time,


and thus were not efficiently


ejected.


Furthermore


onization


times


were


typically


only


This


represents the maximum amount of time an ion was resonated during ionization.













(M+H)


1277










C-)


35 40 45 50 55


Resonant Excitation Voltage
OFF


Resonant Excitation Voltage
ON







-M -h. sAJL v-tt


1' "41!I
i 4G


* I V I j


I TTTI I II I"
0 708


80


mass/charge


Figure 3-11:


a) Profile mass spectra
(M+H) + solvent ions (n


illustrating resonant ejection of acetonitrile
i/z 42) during ion storage.













(M +H)+


Resonant Excitation Voltage
OFF


35 40 45 50 55 60 65 70 75 8B


iViA


Resonant Excitation Voltage
ON





A -A. 4'


35 40 45 50 55 60 65 70 75 80


mass/charge


Figure 3-11:


(continued) b) Profi
acetonitrile (M+ H)+


mass


spectra


solvent ions


ustrating resonant ejection of
42) during ionization.


13B1







ComParison of Solvent Ion Election Methods


The analyte signal obtained for methods which efficiently eject solvent ions


are compared in Figure 3-12.


These plots were obtained by varying the ionization


time with a constant amount of diuron (1 ng/pL yielding 5 ng/s at 0.3 mL/min.) in


100% methanol flowing through the


nterface.


Illustrated


n Figure


base peak of the El spectrum of diuron occurs at m/z 72; its


intensity is plotted


in Figure 3-1


evels were elevated such that al


below m/z 45 were


ejected for methods involving elevated rf leve


For the notch filter mode, the


resonant excitation signal was set to eject the (M+H) ion of methanol solvent


(m/z 33) for 10 ms during ion storage. As expected,


increasing the ionization time


resulted in


inear increases in signal for al


methods unti


space charge develops


due to too many analyte ions being trapped, at which point the plots begin to


level off. The plots also reveal that methods which involved ejection of solvent

ions during ion storage produce about two times greater signal than ejection


during ionization.


As discussed previously


we believe this to be because of ion


osses due to quasi-unstable trajectories


and smaller ionization volumes


at higher


values of qz during


ionization


(Dawson,


1976).


It should


be noted that these


effects might not be as significant for compounds whose El mass spectral base


peaks are at higher m/z values.


Similar studies


monitoring m/z


144 (qz


=0.28) of


carbaryl show less than two times difference


n signal between these operational


modes.


This


was not further


investigated


because the


mass spectra of the













2500


2000


Elevate


Storage


during


I


0 1500


1000


Resonant
during lo


Elevate r
Ioniozation


Ejection
n Storage





luring


500


0-,r
0.0


I I i I I I I I I


on


1.0
on


1.5
me


(m


I I I I I I
2.5


Figure 3-1


Plots


diuron


comparing


analyte


methods


ion signal


described


vs. i


the text which


onization


efficiently


time
eject


solvent ions.


I I I I II II I I I I I I I I I I I I I I I








To obtain maximum sensitivity,


all subsequent studies ejected solvent ions


by elevating the rf level after ionization, during ion storage. Although resonant

ejection during ion storage yielded essentially the same sensitivity, elevation of the


rf amplitude ejected all solvent ion species,


as well as background


ions


water


, column bleed, etc.),


below the chosen m/z.


Also,


with standard


TMSTM


software, ejection of more than one m/z solvent ion species by resonant excitation


would require multiple resonant ejection events.


Although not all solvent ion


species require


10 ms for efficient ejection,


multiple resonant ejection events


would allow more time for undesired ion-molecule reactions to occur.


Multiple


resonant ejection events would be necessary for LC/PB/ITMSTM analyses which

would employ gradient elution or isocratic solvent mixtures.


ADDlication to LC/PB/ITMSTM


Analyses of Simple Pesticide Mixtures


The capabilities of the LC/PB/ITMSTM


system were


initially evaluated by


performing


isocratic


LC/MS


analyses


simple


pesticide


mixtures


components).


The El


mass spectral quality,


estimated LODs


and instrument


calibration


curves


were


used


evaluate


performance


under


realistic


conditions.


Ionization was performed at an rf level corresponding to a low mass


cutoff


of m/z


20 for


ms; solvent ions


were ejected


during


storage


elevating the rf level to eject all ions below m/z 55.


LC/PB/ITMS


analysis


a mixture


of three


pesticides


using




























Figure 3-13:


Total ion chromatogram (TIC), mass chromatogram


and amounts


analyzed


from


LC/PB/ITMS


analysis


a three-pesticide


mixture using LC conditions described in the text.











18Bx





TIC


TIC


250 ng
m/z 88 Methomyl


m/z 72


100 ng
Diuron


m/z 61


250 ng
Linuron


1B8 288 388


A J0


4- A1


m _


wl


I


10~








figure displays the total ion chromatogram


(TIC),


mass chromatograms of the


base peaks of the El mass spectra of each component,


and amounts of each


component injected.


Structures and E


fragmentation of each component were


shown previously in Figure 2-1


and 3-1


These chromatogram


illustrate better


than baseline separation of the compounds was achieved


ess than 6 mm.


The background-subtracted mass spectrum of 80 ng of diuron (m.w. =232


amu)


obtained


from


these


analyses


compared


with


both


a reference


spectrum


(Hites,


1985)


mass


spectrum


obtained


from


solids


probe/ITMSTM analysis


n Figure 3-14.


The spectra are quite similar with respect


to both fragmentation patterns and


relative fragment ion


ntensities.


The low


abundance of the (M+1)+ ion


n the LC/PB/ITMSTM spectra indicated that only


a small


amount of Cl occurred


spectral quality.


, and did not significantly degrade the overall


The spectra of the other compounds


n the test mixture also


compared favorably with reference and solids probe/ITM


instrument calibration curve for diuron


STM El spectra.


is shown in Figure 3-15.


calibration


curves were constructed


by plotting the


integrated


ion abundance


(peak


area)


of the


quantitation


as a function


amount


injected.


quantitation ion for each compound was the base peak of its fu


spectrum.


-scan El mass


It appears from the figure that a linear calibration was obtained from


diuron.


which better


The inset of the figure is an expansion of the region from


ustrates the linearity at the low end.


10-100 ng


The value of the correlation






























Figure 3-14:


El m.
solid


pr


spectra of diuron obtained from (a) library (Hites, 1985), (b)
robe/ITMSTM analysis, and (c) LC/PB/ITMS analysis of 80


ng diuron (background-subtracted).


w
















100


M+*
232


187


60 80 100 120 140 160 180 200 220 240


10wx


63
- n il-


124133


159 187 215


60 88 180 120 140 160 180 200 220 240


100%


81597
159


-^- *^k e n- rr'ia


80 100 120 140 160 180 200 220 240


M +

















- c
c Cc




0a
Om


C a

Im


irn



La




































SI 1 1 1 1 1 1 i I I I II i i I | | i i i I I
3 0 0 0
3 0 0 0
S0 0 0
- r0
'5y) +DZ aJV


O
.o


O


C




: *D
O
: o

-CM





--O
0


CO
cof



Q0


E

0<
0






0
OI





O


0 0 0


L 1








analyzed.


The non-linearities at the ends of the calibration curve are presumably


due to particle beam transport problems at the low end (Bellar et al.


1990) and


the onset of space charging caused by too many analyte ions at the high end.

Although diuron here shows linear behavior, nonlinear calibration curves, like that


obtained for linuron in Figure 3-16,


are far more prevalent


n LC/PB/MS.


nonlinear behavior of PB has been attributed to transport losses through the


momentum separator region of the


nterface.


This has been observed by several


other laboratories as a limitation of PB/LC/MS


for quantitative analyses (Doerge


Miles,


1991


, Brown


Draper


, 1991).


The use of isotopically labelled


ntemal standards has been suggested as a possible solution to this problem and


has been the topic of several recent reports (Ho et al.,


1992, Doerge et al.


1992).


Quantitation using PB and the carrier effect will be discussed


n greater detai


Chapter


instrument


detection


limits


were estimated


selecting


a sample


concentration which reliably yielded a 3:1 signal/noise ratio for each compound's


quantitation


full-scan


mass


spectrum.


Figure


3-17


illustrates


estimated detection limits routinely obtained on this system for several carbamate


and urea-based pesticides.


The LOD's


are typically in the low ng range and are


comparable to LOD's


reported for PB systems which employ quadrupole mass


analyzers (Behymer et al.


1990).


The LOD reported here for


inuron is 10 times


lower than previously reported (Behymer et aL,


1990) because the base peak of













400000





300000


0
a 200000
<


100000


III1111111I
200


I II IIII


Amt.


111111111 ll


400


111111 1


600


I I I I I II ll


800
(ng)


1000


SlIll200
1200


Figure 3-16


Linuron
Dashed


instrument calibration curve from PB/LC/ITMS


analysis.


ne represents region where the onset of space charging


occurred.


injected

























(d
0
0)
cl-
Q


--O
oI


C
0
Ow
L.
*S
0


Figure 3-17:


Estimated
Pesticides


detection
routinely


mits for several carbamate and


obtained


PB/ITMSTM


urea-based


isocratic


anal


described


n the experimental


section.


'4-


I.-
L.
0.
0


m n1








report to eliminate interference from using


ammonium acetate in the mobile


phase,


which


necessitated


a less


intense


(M+', m/z 248)


quantitation.

These LODs were obtained from the calibration runs and are not intended


to be


presented


as the


lowest


LODs


possible


each


compound


on the


LC/PB/ITMST system.


Note that there are a number of options offered by the


LC/PB/ITM


ionization


STM system that could further


times


improve th


, higher multiplier voltages,


LOD'


of a high


, including longer


voltage conversion


dynode, and addition of mobile phase additives such as ammonium acetate to


improve


analyte


transport


through


interface.


Automatic


variation


of the


ionization time using automatic gain control,


AGC (Stafford et al.


1987


Yost et


, 1987),


might lower detection limits and extend the linear dynamic range as


well.


Conclusions


successful


coupling


a PB


nterface


with


an I


TMSTM


been


demonstrated


n this chapter


When directly coupling


a PB


nterface with an


TMST


, it was necessary to eject solvent ion


to minimize space charging and


prevent a significant degree of solvent-C


from occurring.


The additional third


stage of momentum separation has been demonstrated to reduce the amount of


solvent which reached the trap.


Ejection of solvent ion


prior to m


analysis


:








spectra and El spectra obtained by solids probe/ITM


STM analyses. Although linear


instrument


calibration


curves


were


observed


, nonlinear


behavior


was


more


prevalent for LC/PB/ITMSTM determinations of several pesticides.


for several pesticides have been estimated to be


Typical LODs


n the low ng range.











CHAPTER 4
APPLICATION OF LC/PB/ITMS TO ANALYSIS OF PESTICIDE MIXTURES


AND COMPARISON TO OTHER LC/PB/MS


SYSTEMS


Introduction


The analytical utility


of PB relies on the ability to generate classical El


spectra


wide


variety


environmentally


biologically


significant


compounds introduced via LC.


The vast majority of PB/LC/MS system


currently


n use employ quadrupole mass analyzers due to their widespread availability, low


cost, and tolerance of relatively high operating pressures.


mited use


Ion traps have seen


n LC/MS systems primarily because of the large amount of solvent


introduced by most LC/MS


interfaces.


n the previous chapter


several strategies


were investigated which minimized adverse solvent effects and allowed acquisition


mass


spectra


from


isocratic


LC/PB/ITMS


analyses


which


compared


favorably to reference library spectra.

Although isocratic separations are well suited for individual applications


which require simple separation strategies,


they are limited for the analysis of


complex mixtures


with


components


having


a wide


range


retention


times.


Isocratic separations of multicomponent mixtures typically exhibit poor resolution


of early eluting components,


bad peak shape of late eluting components,








complex mixtures


require the use of gradient elution,


which


analogous to


temperature programming for GC.


n gradient elution,


the composition of the


mobile phase is changed throughout the analysis to


increase its eluent strength.


An optimized gradient affords the ability to efficiently analyze a


arge number of


components in a minimum amount of time.

This chapter will report the application of the LC/PB/ITMST system for the


analysis of a ten-component pesticide mixture.


The pesticide mixture required the


use of gradient elution to achieve adequate separation of the components.


performance of this system for gradient elution analyses wil


be compared to that


an LC/PB/MS


system


regards to detection


limits,


which


employed


calibration


a quadrupole


curves


mass


, precision,


analyzer with


overall


spectral


quality.


Finally, the capability of the ion trap to acquire both El and Cl spectra in


alternate scans over the course of an LC peak will be demonstrated.


ExDerimental


Liauid ChromatoaraDhv


Chromatographic separations were performed using


a Hewlett-Packard


(Palo


Alto,


1090L


high-performance


iquid


chromatograph


fitted


with


Rheodyne


(Cotati,


manual


injection


valve


with


a 5 upL


sample


loop.


Separations were performed on a Waters (Milford,


a I -


MA) Nova-Pak 150 mm


-I a a' .,,,,,, ....... ..:L Xt.. S-1 -1- -1- ..... ...- S


* J 11 ii __K


- L






92

through the column overnight to remove residual impurities and column bleed.

The mobile phase used for these studies was an acetonitrile/0.01M ammonium


acetate in


reagent water mixture at a flow rate of 0.3


mL/min.


Post-column


addition


was


used


increase


percent


organic


phase


through


interface.


SCO


(Lincoln,


LC-2600 syringe pump was used for post-


column addition of 100% acetonitrile at a flow rate of 0.1


mL/min,


resulting in a


total mobile phase flow rate of 0.4 mL/min.


through the interface.


nterface


The LC/MS interface was a prototype Finnigan MAT (San Jose, CA) three-


stage particle beam interface (Figure 1-4).


nterface is described


n greater


detai


Chapters


Important


parameters


including


desolvation


chamber temperature, pressure, and nebulizing He flow rate were optimized by

making several injections of the test mixture under different conditions to provide


maximum s

Optimization


sensitivity


more


a majority


difficult


gradient


compounds in thl

elution analyses


mixture.


because


composition of the mobile phase changes throughout the course of the analysis.

Chapter 2 revealed that most PB parameters are largely mobile phase dependant.


Typical pressures


in the three stages of the momentum separator were 5-10 torr,


300 mtorr, and 100 mtorr (adjusted by addition of He),


respectively.


nterface


was inserted into the mass spectrometers via a /" o.d. transfer line probe through