Fundamental studies of thermospray ionization processes

MISSING IMAGE

Material Information

Title:
Fundamental studies of thermospray ionization processes
Uncontrolled:
Thermospray ionization processes
Physical Description:
xiii, 236 leaves : ill. ; 28 cm.
Language:
English
Creator:
Lee, Mike S., 1960-
Publication Date:

Subjects

Subjects / Keywords:
Ionization   ( lcsh )
Liquid chromatography   ( lcsh )
Mass spectrometry   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1987.
Bibliography:
Includes bibliographical references (leaves 232-235).
Statement of Responsibility:
by Mike S. Lee.
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 - 001076288
notis - AFG1047
oclc - 19045275
System ID:
AA00003373:00001

Full Text



















FUNDAMENTAL STUDIES OF THERMOSPRAY IONIZATION PROCESSES


By

MIKE S. LEE























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






UNIVERSITY OF FLORIDA


1987











ACKNOWLEDGEMENTS


I wish to express my sincere gratitude to Dr. Richard A. Yost for

his guidance, direction, and friendship during the completion of this

work. I wish to acknowledge the members of my committee, Dr. John

Eyler, Dr. James Winefordner, Dr. John Dorsey, and Dr. Joseph Delfino

for their various contributions to my thesis work and education at the

University of Florida.

I acknowledge the U.S. Army Chemical Research, Development, and

Engineering Center for their support of this work.

Many thanks go to the guys in the machine shop, Chester, Vern, and

Daly, for their help in the construction of the thermospray vacuum

chamber.

I thank special friends Mike Gehron, Ken Matuszak, and Kevin

McKenna who provided and shared the laughter, joy, and nervous energies

of a gradual student lifetime.

Mostly, I thank my number one family for their endless support and

enthusiasm in everything I do.
































A guitar, blanket, burrito, and you.













TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS................................................... ii

LIST OF TABLES.................................................... vi

LIST OF FIGURES ................................................... vii

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

CHAPTERS

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

Developments in Ionization Techniques for
Nonvolatile Molecules.................................. 2
Thermospray LC/MS Background............................. 12
Present Understanding of the Thermospray
Ionization Mechanism................................... 15
Overview of Thesis Organization.......................... 21

II EXPERIMENTAL............................................. 23

Vestec Thermospray LC/MS Interface....................... 23
HPLC Pumps ............................................... 27
Mass Spectrometer ........................................ 28
Vacuum Chamber Measurements.............................. 28

III PRELIMINARY CHARACTERIZATION OF THE VESTEC
THERMOSPRAY LC/MS INTERFACE............................ 30

Vaporization Studies..................................... 30
Characterization of thermospray
plume types....................................... 31
Variable flow with constant heat.................... 35
Constant flow with controlled heat.................. 38
Thermospray Mass Spectra................................. 42
Ammonium acetate .................................... 43
Ribavirin........................................... 48
Reproducibility of thermospray spectra
of ribavirin ...................................... 53
Purine and xanthine-type compounds.................. 56

IV THERMOSPRAY PLUME MEASUREMENTS IN VACUUM.................. 71

Plume Pressure Measurements .............................. 71
Plume Temperature Measurements ........................... 71
Plume Current Measurements............................... 82















CHAPTER PAGE

V INSTRUMENTAL PARAMETERS................................. 93

Tip Temperature......................................... 101
Source Temperature ...................................... 114
Flow Rate ............................................... 126
Source Pressure......................................... 137
Probe Position .......................................... 148
Summary of the Effects of Instrumental Parameters....... 158

VI SOLUTION CHARACTERISTICS ................................ 162

Mobile Phase and Sample Solvent......................... 162
Sample Concentration .................................... 174
Buffer Concentration .................................... 184
pH Study ................................................ 202
Thermospray Ionization Behavior in a Mixture............ 220
Summary of the Effects of Solution Characteristics...... 225

VII CONCLUSIONS AND FUTURE WORK............................. 228



LITERATURE CITED ................................................. 232

BIOGRAPHICAL SKETCH .............................................. 236













LIST OF TABLES


TABLE PAGE

I Summary of the major techniques used for ionizing
nonvolatile molecules..................................... 6

II Summary of the positive and negative ions observed
in the mass spectra of the purine and xanthine-type
compounds................................................. 62

III The effects of tip temperature on source temperature
and HPLC pump pressure .................................... 102

IV The effects of source temperature on source pressure...... 115

V The effects of flow rate on source pressure obtained
at tip temperatures corresponding to nearly dry plumes.... 127

VI Summary of the mobile phase-solvent studies............... 163

VII Ratios of various positive and negative ion intensities
of ribavirin .............................................. 185

VIII Positive and negative ion intensities of pure
ammonium acetate versus 100 ppm ribavirin................. 199

IX Forms of the ionizing groups of ammonium acetate
at various pH's........................................... 206

X Forms of the ionizing groups of histidine
at various pH's........................................... 207

XI Positive and negative RIC values of
0.1 M ammonium acetate .................................... 208

XII Positive and negative RIC values of 100 ppm histidine
in 0.1 M ammonium acetate................................. 214

XIII Summary of the two-component mixture study for
allantoin and 2-hydroxypurine............................. 222

XIV Summary of the two-component mixture study for
allantoin and 2,6-diamino-8-purinol....................... 223

XV Summary of the two-component mixture study for
2-hydroxypurine and 2,6-diamino-8-purinol................. 224

XVI Summary of the three-component mixture study for
allantoin, 2-hydroxypurine, and 2,6-diamino-8-purinol..... 226

vi













LIST OF FIGURES


FIGURE PAGE

1.1 A simplified thermospray vaporization scheme
illustrating the four major processes which occur
during thermospray ionization ............................ 17

1.2 Direct ion evaporation and chemical ionization
mechanisms of thermospray ................................ 19

2.1 A schematic representation of the Vestec
thermospray LC/MS interface .............................. 25

3.1 A plot of tip temperature versus control
temperature for water at 1.0 mL/min...................... 33

3.2 A plot of probe temperature versus distance
along the probe for water at various flow rates.......... 37

3.3 A plot of probe temperature versus distance
along the probe for wet, nearly dry, and dry
water plumes............................................. 41

3.4 Positive and negative ion full scan thermospray
mass spectra of ammonium acetate......................... 45

3.5 MS/MS daughter spectra obtained for selected
positive and negative ions of ammonium acetate........... 47

3.6 Positive ion full scan mass spectrum of 1 jag ribavirin... 50

3.7 Negative ion full scan mass spectrum of 1 jAg ribavirin... 52

3.8 Positive ion full scan thermospray mass spectra
of ribavirin acquired several weeks apart................ 55

3.9 Negative ion full scan thermospray mass spectra
of ribavirin acquired several weeks apart................ 58

3.10 Model purine and xanthine-type compounds used for
determining the effects of various substituents
on thermospray ionization................................ 60

3.11 Positive ion intensity summation profiles of the
purine and xanthine compounds, ammonium acetate,
and RIC .................................................. 66








3.12 Negative ion intensity summation profiles of the
purine and xanthine compounds, ammonium acetate,
and RIC .................................................. 68

4.1 Plume temperatures measured along wet, nearly dry,
and dry plumes at 2.0 mL/min............................. 74

4.2 Plume temperatures measured along a nearly dry plume
at 2.0 mL/min for three manifold temperatures............ 77

4.3 Plume temperatures measured along nearly plumes
produced at various flow rates........................... 79

4.4 Plume temperatures measured along a nearly dry plume
produced at 2.0 mL/min at two manifold pressures......... 81

4.5 Plume currents measured along wet, nearly dry, and dry
plumes at 2.0 mL/min..................................... 84

4.6 Plume currents measured along a nearly dry plume at
2.0 mL/min for three manifold temperatures............... 87

4.7 Plume currents measured along nearly dry plumes
produced at various flow rates........................... 89

4.8 Plume currents measured along a nearly dry plume
produced at 2.0 mL/min at two different manifold
pressures ................................................ 91

5.1 A model showing the effects of temperature on
hypothetical thermospray plumes .......................... 95

5.2 A model showing the effects of pressure on
hypothetical thermospray plumes.......................... 98

5.3 A model showing the effects of probe position on
hypothetical thermospray plumes.......................... 100

5.4 A plot showing the effects of tip temperature on
the ion intensities of pure ammonium acetate............. 105

5.5 A plot showing the effects of tip temperature on
the ion intensities of ammonium acetate with 100 ppm
ribavirin present ........................................ 107

5.6 A plot showing the effects of tip temperature on
the positive ion intensities of 100 ppm ribavirin........ 109

5.7 A plot showing the effects of tip temperature on
the negative ion intensities of 100 ppm ribavirin........ ill

5.8 A plot showing the effects of source temperature
on the ion intensities of ammonium acetate............... 118


viii








5.9 A plot showing the effects of source temperature
on the ion intensities of ammonium acetate with
100 ppm ribavirin present ................................ 120

5.10 A plot showing the effects of source temperature
on the positive ion intensities of 100 ppm ribavirin..... 122

5.11 A plot showing the effects of source temperature
on the negative ion intensities of 100 ppm ribavirin..... 124

5.12 A plot showing the effects of flow rate on the
ion intensities of pure ammonium acetate................. 130

5.13 A plot showing the effects of flow rate on the
ion intensities of ammonium acetate with 100 ppm
ribavirin present ........................................ 132

5.14 A plot showing the effects of flow rate on the
positive ion intensities of 100 ppm ribavirin............ 134

5.15 A plot showing the effects of flow rate on the
negative ion intensities of 100 ppm ribavirin............ 136

5.16 A plot showing the effects of source pressure
on the ion intensities of pure ammonium acetate.......... 140

5.17 A plot showing the effects of source temperature
on the ion intensities of ammonium acetate with
100 ppm ribavirin present ................................ 142

5.18 A plot showing the effects of source pressure
on the positive ion intensities of 100 ppm ribavirin..... 144

5.19 A plot showing the effects of source pressure
on the negative ion intensities of 100 ppm ribavirin..... 147

5.20 A plot showing the effects of probe position on the
ion intensities of pure ammonium acetate................. 151

5.21 A plot showing the effects of probe position on the
ion intensities of ammonium acetate with 100 ppm
ribavirin present ........................................ 153

5.22 A plot showing the effects of probe position on the
positive ion intensities of 100 ppm ribavirin............ 155

5.23 A plot showing the effects of probe position on the
negative ion intensities of 100 ppm ribavirin............ 157

6.1 Positive ion thermospray mass spectra of ribavirin
with 0.1 M ammonium acetate buffer as the sample
solvent and mobile phases of methanol, water, and
ammonium acetate......................................... 166









6.2 Negative ion thermospray mass spectra of ribavirin
with 0.1 M ammonium acetate buffer as the sample
sample solvent and mobile phases of methanol, water, and
ammonium acetate......................................... 168

6.3 Positive ion thermospray mass spectra of ribavirin
with 0.1 M ammonium acetate buffer as the mobile phase
and sample solvents of methanol, water, and
ammonium acetate......................................... 171

6.4 Negative ion thermospray mass spectra of ribavirin
with 0.1 M ammonium acetate buffer as the mobile phase
and sample solvents of methanol, water, and
ammonium acetate......................................... 173

6.5 Positive ion thermospray mass spectra of various
amounts of ribavirin..................................... 176

6.6 Negative ion thermospray mass spectra of various
amounts of ribavirin..................................... 178

6.7 Calibration plots for the major positive ions of
ribavirin and the RIC .................................... 181

6.8 Calibration plots for the major negative ions of
ribavirin and the RIC .................................... 183

6.9 Positive ion intensities of the major ions of
ammonium acetate at various concentrations............... 187

6.10 Positive ion intensities of the major ions of
ammonium acetate obtained from 100 ppm ribavirin
standards prepared in various concentrations of
ammonium acetate......................................... 190

6.11 Positive ion intensities of 100 ppm ribavirin
prepared in various concentrations of ammonium acetate... 192

6.12 Negative ion intensities of the major ions of
ammonium acetate at various concentrations............... 194

6.13 Negative ion intensities of the major ions of
ammonium acetate obtained from 100 ppm ribavirin
standards prepared in various concentrations of
ammonium acetate......................................... 196

6.14 Negative ion intensities of 100 ppm ribavirin
prepared in various concentrations of ammonium acetate... 198

6.15 The structures of ammonium acetate and histidine
at pH 6 .0 ................................................ 204









6.16 Positive ion thermospray mass spectra of 0.1 M
ammonium acetate at various pH values.................... 210

6.17 Negative ion thermospray mass spectra of 0.1 M
ammonium acetate at various pH values.................... 212

6.18 Positive ion thermospray mass spectra of 100 ppm
histidine in 0.1 M ammonium acetate at various
pH values ................................................ 216

6.19 Negative ion thermospray mass spectra of 100 ppm
histidine in 0.1 M ammonium acetate at various
pH values ................................................ 218













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




FUNDAMENTAL STUDIES OF THERMOSPRAY IONIZATION PROCESSES


by

Mike S. Lee

December, 1987


Chairman: Richard A. Yost
Major Department: Chemistry

The effects of various parameters on thermospray ionization were

determined on a thermospray liquid chromatography/mass spectrometry

(LC/MS) interface. Particular changes in these parameters affect

specific processes of thermospray [vaporization, droplet evaporation,

droplet charging, and direct ion evaporation/chemical ionization (CI)]

and result in a change in the overall sensitivity and the relative

abundances in the mass spectra. The instrumental parameters such as

tip temperature, source temperature, flow rate, and source pressure

most directly affect the vaporization, droplet evaporation, and direct

ion evaporation/CI processes, while solution characteristics such as

mobile phase, sample solvent, sample concentration, buffer

concentration, pH, and mixture components most directly affect the

droplet charging and direct ion evaporation/CI processes. Each

parameter was varied systematically and the effects on a model sample

compound, ribavirin, was determined.


xii








Conditions of vaporization and droplet evaporation which produce

droplet-rich plumes favor the formation of fragment ions while

conditions which produce vapor-rich plumes favor the production of

pseudo-molecular ions. The role of a volatile electrolyte in

thermospray ionization was demonstrated with mobile phase-sample

solvent combinations of methanol, water, and ammonium acetate. The

distinction between the direct ion evaporation and CI mechanisms was

difficult to make. Although inconclusive, these results do indicate

that CI most likely plays a major role in sample molecule ionization by

thermospray. Recommendations for future work in thermospray

fundamentals are made and focus on methods which will help to

distinguish between the direct ion evaporation and CI mechanisms.


xiii














CHAPTER I
INTRODUCTION

This thesis describes fundamental studies of thermospray

ionization, a relatively new ionization technique used in liquid

chromatography/mass spectrometry (LC/MS). Although the thermospray

technique has experienced rapid growth and has proven to be a

successful method for the analysis of polar and labile compounds, the

actual mechanism of thermospray ionization is still unclear. The

objectives of our fundamental studies were to gain further insight into

the thermospray mechanism and to improve the reproducibility of the

thermospray ionization technique. This work was undertaken in hopes

that a better understanding of the various parameters associated with

the processes involved with thermospray would aid in the investigation

of much more complex and detailed conditions of thermospray ionization.

A better understanding of these processes should ultimately lead to a

more routine applicability of the thermospray LC/MS interface, improved

day-to-day reproducibility, and perhaps some potential instrumental

improvements.

In this introductory chapter, developments in ionization techniques

used for nonvolatile molecules are discussed. Background to the

thermospray technique along with the present understanding of the

thermospray ionization mechanism is presented, followed by an overview

of the thesis organization.











Developments in Ionization Techniques For Nonvolatile Molecules

Through the years mass spectrometry has undergone dramatic changes.

Exploration into new and novel ionization techniques has contributed

significantly to this change, resulting in the application of mass

spectrometry to an increasingly wide range of organic compounds.

Described below are the various ionization techniques used in mass

spectrometry and their capabilities and limitations.

In the late 1960s combined gas chromatography/mass spectrometry

(GC/MS) instruments with computer data systems became commercially

available so that routine quantitative analysis was possible. On these

systems electron ionization (EI) was used in essentially all analytical

applications. A "standard" electron energy of 70 eV was used for both

qualitative and quantitative analyses resulting in the production of an

intense, stable ion beam, and a highly reproducible fragmentation

pattern (eqn. 1.1).



M + e" M+ + 2e- (1.1)



Although the El mode remains the accepted preliminary ionization

mode for identification purposes, two factors exist which restrict El

to a limited range of organic compounds. First, the El mode requires

that the sample be vaporized prior to ionization which may lead to

decomposition of thermally unstable compounds. Therefore, useful

spectra of very polar and thermally labile compounds cannot be obtained

with EI. Second, the use of electron energies of 70 eV usually results

in extensive fragmentation and sometimes undetectable molecular ions.








In many cases chemical ionization (CI) techniques are capable of

providing complementary information to El spectra. CI achieves

ionization without transferring excessive energy to the sample

molecules resulting in the formation of adduct ions (often a protonated

molecule) which contain the intact molecular species of the analyte.

CI differs from El in that the sample molecules are ionized by

interaction with ions of a reagent gas rather than with electrons. The

site of protonation is most likely to occur on the heteroatom of

greatest proton affinity. CI is the subject of a book (1) and several

reviews in the literature (2,3) where processes of ion formation are

discussed in detail. Below is a brief description of ion formation by

CI.

Production of reagent ions is initiated by El of the reagent gas.

Even though the analyte and the reagent gas are present in the

ionization source, the reagent gas is present in great excess

(typically, the partial pressure of an analyte will be 10-5 torr,

whereas the reagent gas will be 1 torr); therefore, only reagent ions

result from El. The principle electron/molecule and ion/molecule

reactions that occur in CI for the case involving methane as reagent

gas are illustrated in eqns. 1.2-1.6. The methane molecules are

converted to molecular ions by EI, some of which decompose to fragment

ions of methane (CH3+, CH2+, CH+, etc.). The ions of methane react

with a methane molecule to produce the CH5+, C2H5+, and C3H5+ ions,

which do not react further with methane.



CH4 + e- & CH4+, CH3+, CH2+, (1.2)

CH+, H2+, H+ + 2e"








CH4 + + CH4 CH5+ + *CH3 (1.3)

CH3+ + CH4 C2H5+ + H2 (1.4)

SC2H4+ + H2
CH2+ + CH4 C2H3 + H2 + (1.5)
IL- C2H3+ + H2 + *H

C2H3 + CH4 C3H5+ + H2 (1.6)

There are many more reactions which occur between the methane ions

and the methane molecules; however, those shown above are the most

important because they lead to reagent ions which are unreactive with

molecular methane resulting in their accumulation in the ionization

source. Most other organic molecules have a sufficiently high proton

affinity to abstract a proton from these reagent ions resulting in a

reaction as shown in eqn. 1.7. An analyte molecule, M, will interact

with a reagent ion to produce a protonated molecule of the analyte,

(M+H)+. Similar reactions with C2H5+ and C3H5+ may occur resulting in

(M+H)+ as well as (M+C2H5)+ and (M+C3H5)+ adduct ions.



M + CH5+ (M+H)+ + CH4 (1.7)



Typical CI spectra contain abundant protonated molecular ion

species, adduct ion species (associated with the reagent gas ions) and

less fragmentation than observed in El spectra. Thus, molecular weight

information is usually more easily and reliably obtained. However, as

in the case with El, sample vaporization occurs prior to ionization and

thus, the sample must be thermally stable.

Recent research aimed at overcoming the thermal stability or

volatility barrier has resulted in a wide variety of ionization

techniques in which the direct production of pseudo-molecular ions from








nonvolatile molecules have been reported. These techniques produce

relatively stable even-electron species rather than the odd-electron

species produced initially by gas-phase El and can provide a

potentially attractive approach to applying mass spectrometry to large,

nonvolatile, or thermally labile compounds not amenable by conventional

ionization techniques such as El and CI. These techniques differ

primarily in the means employed for disrupting the solid or liquid

surface of the sample. However, a common feature of all these

techniques appears to be the "direct" production of molecular ions [M+,

(M+H)+, (M+Alkali)+] from a condensed phase without the formation of a

neutral gas-phase molecule as an intermediate.

There are several techniques which have been developed for

desorbing molecular ions from a condensed phase. All of the these

techniques have been reviewed in detail (4-6) or have been described

elsewhere in various publications (7-15). These techniques are listed

in Table I according to the method of energy application and the

physical state of the sample prior to energy deposition. Samples may

be either solid or liquid on the surface, depending on the technique

used, and an electrical field may be of primary importance to the ion

production mechanism or may only serve to focus the ions into the mass

spectrometer. Below is a brief description of these ionization

techniques.

Several of these desorption ionization techniques employ electrical

energy via a high electrical field and/or ohmic heating of the sample.

They generally involve ohmic heating of a metallic substrate in contact

with the sample or the production of charged droplets in the presence

or absence of an electrical field. These techniques include field















O
4.4




4 4
0
.-4
00
.9-4
U4-1

4.,

ti1


0
u44.





0
.-4




'r-I
~*m1
rj -


U



4-1
-4


H -H
No
.or- .0


OI





Be




i a
0r4

'Io


0

ra


A3~(



4a4)
0 '4
4.4

(4.,

I.H
in41







"-4
9-4



0
|.9-





0

.9-4












4.4
a




Ir<
Ln
C%


M4-4

44i


4S-S


IJI


co



-44 N
s-0
-i
4., "0




9-i







oit
0 S4


0

4


ii

4

rz4


*H

c44
*H 0

'O00
9-4


9-4
y)

4-1
r-I
4.4


r-1




L




-0
M(d
+JN








desorption, electrohydrodynamic ionization, electrospray, and

thermospray.

With the field desorption (FD) technique, the sample is deposited

from solution onto the surface of a field anode, usually a 10 pm

tungsten wire covered with field-enhancing microneedles. Ionization

and desorption of molecules is achieved by applying a high voltage of

about 10 kV between the wire and a 2-3 mm distant counter-electrode

while resistively heating the emitter wire. The mass spectra produced

by FD exhibit molecular ions such as M+ and (M+H)+, and adduct ions

such as (M+Na)+ and (M+Cl)" depending on the sample solution. Three

distinctly different mechanisms contribute to the production of ions by

the FD technique. The ionization of molecules can occur by an electron

tunneling process, which takes place on the tips of the field enhancing

microneedles and requires a high threshold field strength (>109 V/m).

Ions can also be formed in condensed sample layers and extracted via a

field-induced desolvation mechanism. This mechanism requires lower

threshold field strengths for the production of gaseous ions. The

third mechanism is a thermal mechanism which is principally independent

of external fields. In general this mechanism does not contribute to

the emission of ions except in the cases of metal ions at emitter

temperatures > 800 K. Recent studies by Giessmann and Rollgen (16)

show that field emission of molecular ions depends on the temperature

of the emitter, the field strength, and the thickness and physical

state of the sample surface and that the "best anode temperature"

corresponds to melting of the solid sample on the field emitter.

The electrohydrodynamic (EHD) ionization technique developed

primarily by Evans and co-workers (17-19) is similar to FD in that a









high electrical field (6 kV) is applied to the sample. However, in

this case the sample is present in a liquid meniscus at the end of a

capillary tube. When liquids contained in a hypodermic syringe needle

or similar capillary are introduced into a vacuum, the interaction of

the electrical field and the liquid surface of the capillary causes a

distortion of the tip of the meniscus into a sharp cone. The liquid

surface is distorted further by the EHD process and charged droplets

are produced and accelerated away from the tip. This technique has

been shown to successfully generate pseudo-molecular ions

characteristic of any sample material which can be dissolved, together

with an electrolyte, in a nonvolatile organic solvent which can also

undergo appreciable interaction with cations and anions in solution.

In the electrospray (ES) technique developed by Dole and co-workers

(20,21) solutions are flowed through a hypodermic needle (5-20 pL/min)

biased at ca. 10 kV into a spray chamber at room temperature and

atmospheric pressure. Charged droplets are formed and rapidly

evaporate until the increase in surface charge density and the decrease

in radius of curvature result in electric fields strong enough to

desorb sample ions. The resulting dispersion of ions is passed through

a small orifice into an evacuated region to form a supersonic free jet.

Ions are sampled from the core of the jet and are introduced into the

mass spectrometer using a standard nozzle and skimmer.

The thermospray ionization technique produces ions from liquid

solutions without requiring strong electrical fields or an external

source of ionization. The technique developed by Vestal and co-workers

(22-25) produces ions directly from a flow of vaporized liquid (1-2

mL/min) containing a volatile electrolyte such as ammonium acetate.








When this vaporized liquid is introduced into the ion source of the

mass spectrometer, a supersonic jet of vapor containing a mist of fine

particles and droplets results. Molecular ions desorb or evaporate

directly from the evaporating droplets typically forming (M+H)+ and

adducts of ammonium acetate such as (M+NH4)+ and (M+CH3COO)>. A

detailed description of the basic principles and apparent ionization

mechanism of the thermospray ionization technique will be presented in

the next section.

Another group of ionization techniques employs particle impact on a

surface of the sample or a matrix containing the sample. In these

techniques the source of sample energization is a beam of ions or

neutrals with kinetic energies ranging from a few keV to more than 100

MeV. The compound of interest is desorbed and ionized directly from a

thin solid or liquid layer deposited on a metal substrate. These

techniques include secondary ion mass spectrometry, fast atom

bombardment, and 252Cf fission fragment desorption.

The secondary ion mass spectrometry (SIMS) technique involves the

bombardment of a sample, either solid (26) or dissolved in solvent

(27), with a beam of primary ions and analyzing the secondary ions that

emanate from the sample surface. In most cases an ion gun which

generates Xe+ or Ar+ ions is used to produce the primary bombarding ion

beam. The primary-ion beam is usually oriented perpendicular to the

mass analyzer with the sample mounted on a probe or target which can be

adjusted to provide for an angle of incidence ranging from 0 to 90.

The ionization mechanism of the SIMS technique appears to be identical

to the fast atom bombardment (FAB) technique (28,29).








In the FAB technique, a beam of atoms of Xe or Ar are directed

toward a sample contained in a polar, viscous, low-volatility solvent

(usually glycerol). Essentially the same orientation of atom beam,

sample target, and mass analyzer is used as in the SIMS technique. The

resulting energetic bombarding neutrals collide with the sample surface

producing many particles, both ions and neutrals. The bombardment of

the sample surface causes intense "hot spots" with the energy

deposition dependent on the angle on incidence. Angles of incidence

which approach 90 (grazing angle) result in the majority of the energy

deposition limited to the surface layers of the solution, resulting

primarily in the desorption of molecular ion species (30). Intact

molecular ions (M+H)+ and adduct ions (M+Na)+ which are desorbed are

characteristic of both SIMS and FAB spectra. Angles of incidence which

approach 0* result in a more intense mode of energy deposition which

causes intramolecular vibrations deeper within the matrix (31). This

orientation results in the degradation of the analyte and matrix and

consequently a low abundance of molecular ions. The presence or

absence of charge on the impacting primary particle has little effect

on the desorption process; however, the use of the neutral beam is

somewhat more convenient with mass spectrometers in which the ion

source is at high potential.

The 252Cf fission fragment desorption technique employs a beam of

fission fragments resulting from the radioactive decay of 252Cf which

penetrates the sample coated on a thin (<10 am) nickel foil (32,33).

The fission fragments pass through a 10 nm diameter region of the foil

depositing considerable amounts of energy and inducing electronic

excitation. Atoms and molecules within the foil perceive the swiftly








moving ions passing through the matrix as equivalent to a short burst

of photons (32). The electromagnetic radiation induces electronic

excitation, which quickly dissipates into a complex spectrum of atomic,

molecular, and matrix excitation, resulting in the desorption of intact

molecular ions, (M+H)+, and adduct ions, (M+Na)+, from the sample.

Finally, techniques which involve lasers with power densities

ranging from ca. 103 to 1010 W/cm2 and wavelengths from the UV to the

far-IR have been successfully employed to produce desorption of ions

from solid samples (34-36) or samples contained in a liquid matrix such

as glycerol (37). The laser (essentially a photon gun) is used to

bombard a specific area of the sample with photons in a manner

analogous to the SIMS and FAB techniques. Protonated molecules or

adduct ions with cations such as Na+ or K+ are typically produced, with

the extent of their fragmentation influenced by the power of the laser.

From a review of the recent literature on ionization of nonvolatile

molecules, it is evident that a wide variety of dissimilar techniques

produce mass spectra with similar features, even-electron pseudo-

molecular ions from nonvolatile molecules. The primary processes

involved in the first-step (energy deposition) are apparently quite

dependent on the kind of excitation used; furthermore, some of the

properties of the intermediate must depend on the nature of the primary

processes. It is not yet clear, however, whether the final step

leading to the production of a pseudo-molecular ion is dependent or

independent of the primary processes occurring in the condensed phase

or whether neutral gas-phase molecules play an integral role in pseudo-

molecular ion formation via CI processes.








Thermosprav LC/MS Background

For over a decade, several research groups have worked on the

development of a practical LC/MS interface. Because of the basic

incompatibilities between high performance liquid chromatography (HPLC)

and MS, the successful on-line coupling of these instrumental

techniques has been difficult to achieve (38-41). This is because the

normal operating conditions of a mass spectrometer (high vacuum, high

temperatures, gas-phase operation, and low flow rates) are opposed to

those used in HPLC (high pressures, low temperatures, liquid-phase

operation, and high flow rates).

The moving belt interface (42,43) and direct liquid introduction

(44-47) are two of the principal methods which have been used for LC/MS

interfacing other than thermospray. These two techniques are capable

of generating El (moving belt) and/or CI (either technique) spectra of

volatile and nonpolar type compounds which are also amenable to GC/MS

techniques. The moving belt interface provides one of the more

versatile approaches to coupling LC to MS in that it permits the

operation of the mass spectrometer in both the El and CI modes. In

this method the LC effluent is deposited on a polyimide or Kapton belt

where it is desolvated; it is then transported into the ion source

region of the mass spectrometer where the remaining sample residue is

flash vaporized and ionized. Direct introduction of the effluent from

a LC into the ion source was first described by McLafferty and co-

workers (48). In this technique, a fraction of the eluent (sample and

mobile phase) is passed through a capillary tube at flow rates of 10-60

pL/min. Microdroplets are formed through a pinhole (ca. 4 Am diam.)

and directed into the ions source where the neutral sample molecules








undergo CI processes with the mobile phase serving as the CI reagent

gas.

While it may be desirable and perhaps more convenient in certain

situations to have a LC/MS technique which is capable of providing El

and CI information identical to GC/MS, the moving belt and direct

liquid introduction techniques do not provide for a complementary

method; namely, the ability to provide useful mass spectral information

of chromatographically separated polar and thermally labile compounds

which are not amenable to GC/MS techniques. Although other recent

LC/MS approaches such as atmospheric pressure ionization (49), liquid

ion evaporation (50), electrospray (51,52), and the monodisperse

aerosol generator for introduction of liquid chromatographic effluents

(53) have attempted to provide for this capability, thermospray has by

far been the most successful, and as a result, has been the most widely

used. All of the LC/MS interfacing techniques listed above differ in

their principle of operation; however, all are capable of transport,

desolvation, and ionization of a liquid sample. Several reviews (54-

58) have appeared in the literature which describe in more detail the

principles of operation of these techniques. Thermospray is unique of

the LC/MS approaches in that it is the only technique which does not

require an external form of ionization or external electrical fields.

The thermospray technique, developed by Blakley and Vestal (22-25),

emerged from efforts to develop an LC/MS interface suitable for

efficiently analyzing samples dissolved in aqueous mobile phases at

typical HPLC flow rates (1-2 mL/min). The original approach involved

the production of a molecular beam of the vaporized effluent which

could be directed into an El or CI source until it was accidently









discovered that, under certain conditions of flow rate, temperature,

and mobile phase composition, ions were produced without the use of a

filament (22). Early attempts to vaporize the liquid eluent employed a

CO02 laser which was focused on droplets exiting the probe. However,

when the laser was accidently focused on the end of the vaporizer tube

instead, it was discovered that gas-phase ions could be produced

without the use of an external ionization source (23). The laser was

later replaced by an array of oxy-hydrogen torches (24,25). In the

present version, the power required to vaporize the liquid is supplied

by passing an electrical current through the capillary tube itself.

Thermospray LC/MS interfaces have been commercially available for

several years with a wide variety of applications reported in the

literature. The analysis of peptides and amino acids (59,60),

glucuronides (61), pesticides (62), dyes (63), and metabolites in

biological fluids (64) have been described recently. In general the

mass spectra obtained from these nonvolatile compounds contain

protonated molecular ions as the major peaks with few structurally

significant fragments. This lack of fragmentation is a significant

limitation in attempting to use thermospray LC/MS to determine the

structure of eluted compounds.

An instrumental development was recently reported which employs the

use of a repeller electrode in which increased fragmentation was

observed in thermospray mass spectra of caffeine, D 4030, D 2439,

bambuterol, and 16-hydroxyprednisolone (65). In most examples repeller

voltages of 120-180 V have been used to produce significant

fragmentation. The thermospray source which was used in our studies

was a relatively early model and was not equipped with a repeller








electrode; however, we have at our disposal collision-induced

dissociation and MS/MS to provide structural information.



Present Understanding of the Thermospray Ionization Mechanism

Thermospray begins with the spraying of aqueous mobile phases

containing a volatile electrolyte such as ammonium acetate through an

electrically heated stainless steel capillary into the source region of

the mass spectrometer at flow rates of 1 to 2 mL/min. A simplified

thermospray vaporization scheme is shown in Figure 1.1 and illustrates

the four major processes which occur during thermospray ionization

(vaporization, droplet evaporation, droplet charging, and direct ion

evaporation/CI). The vaporization of liquids in vacuum results in the

production of a supersonic jet of vapor containing a mist of fine

droplets and particles with estimated velocities of 105 cm/s (22). A

portion of the droplets produced are electrically charged, with an

equal distribution of positively and negatively charged droplets

(66,67). The actual size of the droplets is dependent on the vaporizer

tip diameter and the conditions affecting vaporization (temperature and

flow rate). As the droplets travel through the heated source region

they continue to vaporize because of the heat input from the

surrounding hot vapor. As the charged droplets shrink to a certain

diameter (1 pm) the electric field increases until it is strong enough

(107 V/m) to eject ions (direct ion evaporation) from the liquid

surface (68,69). Ions are then sampled into the mass analyzer region

through a conical exit aperture.

Two distinct stages are believed to be involved in the formation of

thermospray ions (Figure 1.2). The first stage is the formation of


































0 Ew
0

, o









M4



O -4

0






n N
o o,

1 0












-4 Q-
- -4

> N















4-i c














5o
z z
o co


0 0





































0

0





.4)

0 4

0

41

-44




0 -

0

'41




~0

44




0 )-4


U)

















c
+ .

N
U-
0
++-*+ +

C103 .imS
0

U-
E

C)








0
0




a: "5
0.
CO
w

LU


0
C)
1_
Q









gaseous ions out of the microdroplets produced by a direct ion

evaporation mechanism. These ions will be referred to as primary

thermospray ions. Primary ions produced in the thermospray process are

apparently identical to those present in solution. For example,

ammonium acetate produces (NH4)+ and (CH3COO)- ions with clusters of

these ions with water, ammonia, and acetic acid. The second stage

involves the primary thermospray ions in gas-phase ion/molecule

reactions with neutral gas phase sample molecules. These CI reactions

may remove the primary thermospray ions and produce new product ions.

This process has been recognized as of possible importance in the

thermospray ionization mechanism but has received much less attention

than the primary direct ion evaporation processes.

The difficulty in elucidating further details of the thermospray

mechanism lies in the complexity of the thermospray plume itself. The

thermospray plume consists of a dynamic collection of evaporating

droplets and particles moving at supersonic speeds, producing ions

without an external source of ionizing electrons. Studies performed by

Iribarne and Thomson (68,69) using an atomizer to generate charged

droplets have shown that small ions can separate or "evaporate" from

evaporating droplets carrying electrical charges. This model of ion

evaporation from liquid droplets appears to account satisfactorily for

the initial ionization (primary ions) observed in the thermospray

technique; however, it may not account for ionization produced in the

later stages of evaporation (secondary ions). In our studies we have

chosen to retain the integrity of the thermospray plume and perform

studies on actual thermospray plumes under actual thermospray








conditions of temperature and pressure in order to better understand

the processes occurring in these later stages of ion evaporation.

Our experiments were based on the fact that a particular

instrumental parameter or solution characteristic can greatly affect a

particular thermospray process. For example, source temperature would

be expected to have a major role in affecting the droplet evaporation

process. The approach used in our studies was to first obtain direct

physical measurements of actual thermospray plumes in vacuum and then

determine the effects of instrumental parameters and solution

characteristics on the ionization of a pure buffer (ammonium acetate)

and of a model compound (ribavirin) prepared in ammonium acetate.

Ribavirin was a particularly attractive choice as a model compound

because of its thermal liability which renders classical mass

spectrometric ionization techniques ineffective; thus, to date little

mass spectral data has been obtained on the compound. Also, ribavirin

is currently under investigation as a broad-based anti-viral compound

for which there is no sensitive analytical method currently available;

therefore, the characterization of its thermospray mass spectral

behavior would be desirable.



Overview of Thesis Organization

This work represents a systematic study of the effects of

instrumental parameters and solution characteristics on thermospray

ionization. The remaining chapters which follow this introductory

chapter will provide a detailed description of the experimental set-up,

and will discuss the results of our fundamental studies of thermospray

ionization. Chapter II will describe the various instruments used in








these studies which include a thermospray LC/MS interface, HPLC pumps,

mass spectrometer, and a specially designed vacuum chamber.

Preliminary studies dealing with the characterization of the Vestec

thermospray LC/MS interface are discussed in Chapter III. In Chapter

IV the results of direct physical measurements of actual thermospray

plumes in vacuum will be discussed. The effects of instrumental

parameters (Chapter V) and solution characteristics (Chapter VI) will

be discussed with respect to pure ammonium acetate and a ribavirin

standard prepared in ammonium acetate. The thesis ends with

conclusions and future work (Chapter VII), which will state our

conclusions from these fundamental studies and provide recommendations

for future work in this area of research based upon our results.











CHAPTER II
EXPERIMENTAL

Samples and Reagents

Analytical-grade ammonium acetate was obtained from Fisher

Scientific (Fair Lawn, New Jersey), and Mallinckrodt (St. Louis,

Missouri). Ribavirin was received as a gift from Dr. B.J. Gabrielson

of the U.S. Army Medical Research Institute of Infectious Diseases

(Fort Detrick, Frederick, Maryland). Purine, 2-aminopurine, adenine,

2-hydroxypurine, 2,6-diamino-8-purinol, 6-thioguanine, 6-

mercaptopurine, xanthine, hypoxanthine, 6-thioxanthine, uric acid,

allantoin, alloxan monohydrate, and histidine were obtained from Sigma

Chemical Co. (St. Louis, Missouri). All reagents were analytical

grade.



Vestec Thermospray LC/MS Interface

A Vestec thermospray LC/MS interface (Vestec Corp., Houston, TX)

was used. A schematic representation of the Vestec thermospray

interface is shown in Figure 2.1. The interface consists of a

vaporizer probe, thermospray ion source, equipped with a thoriated

iridium filament, discharge electrode, and pump out line (1 cm i.d.) to

a 300 L/min mechanical pump. Below is a description of the interface

components and a brief description of their primary functions.

The vaporizer probe consists of a stainless steel capillary tube

(0.015 mm i.d. X 1.5 mm o.d.) located inside a 6.5 mm o.d. probe which

is inserted into the mass spectrometer via a vacuum lock. Two

thermocouples are spotwelded to the capillary, one located near the

inlet and the other near the tip of the probe. The thermocouple





































co 4)c

c 4J
0)
U 0
a''
ca

c)n C
0JJ(





E- 0

4-





01 4.j


U
(D~ .















































ww


i
z



WW
t m
cr. 0
0 x-,,,
a. CL
5 M









positioned at the inlet measures the temperature at which no

vaporization is occurring and is commonly referred to as T1 or the

control temperature. The thermocouple located at the tip measures T2

or the vaporizer tip temperature. The power required to vaporize the

liquid is supplied by passing an electrical current through the

capillary tube itself. A feedback controller is used to control the

power input to the vaporizer so that T1 is maintained constant. A

digital panel meter displays either Tj or T2.

The Vestec thermospray ion source is specially designed to

accommodate the excess vapor flows being delivered from the vaporizer

probe. Flow rates of up to 2.0 mL/min could be handled by the source

without shutting down the mass spectrometer. The source consists of a

4.11 cm X 1.91 cm X 1.27 cm block of stainless steel with openings for

the vaporizer probe (0.95 cm diam.), source heater (0.64 cm diam.),

pump out line (0.95 cm diam.), and an ion sampling cone (1.27 cm

diam.). A thermocouple is located downstream past the ion sampling

cone to measure the vapor temperature. The power input to a cartridge

heater is controlled by a temperature controller so that either the

vapor temperature or source block temperature is held constant. The

ion sampling cone measures 0.64 cm in height and the aperture is 0.050

mm in diameter. With the probe fully inserted into the source the

distance between the probe tip and the ion sampling cone aperture was

0.6 cm. We modified the source to accommodate a thermogauge so that

source pressures could be measured. (This line for source pressure

measurements was not a standard feature on the ion source.) Typical

indicated pressures inside the source ranged from 1.8 to 2.6 Torr. The

source pressure could be increased to a maximum of 8 Torr by partially








closing a valve on the pump out line. The source pressure line (1 mm

i.d.) was made to be interchangeable with the vapor thermocouple; thus,

only one reading (vapor temperature or source pressure) could be

obtained at any one time. In these studies the source was operated to

measure source pressure exclusively. Although the source was equipped

with external sources of ionization, a thoriated iridium filament and a

discharge electrode, they were not used in these studies.



HPLC Pumps

An ISCO LC-5000 precision syringe pump (ISCO, Inc., Lincoln, NE)

was used for all thermospray studies. The pump was capable of

producing a precise, pulse-free delivery of liquids (500 mL capacity)

at flow rates ranging from 1.0 pL/min to 6.7 mL/min with an output

pressure limit of 3,700 psi. The pump operates in a constant flow mode

and the pump pressure was digitally displayed in psi. The syringe

barrel and piston are made of 304 stainless steel and the pump seals

are graphite-filled Teflon. When changing solvents, the cylinder was

refilled and pumped three to four times with 50 mL volumes of the

desired solvent to ensure the thorough rinsing of the pump.

For experiments dealing with pump pulsing and plume measurements in

vacuum, an EM Science MACS 100 pump was used. This pump is a single

piston, constant stroke, reciprocating HPLC pump capable of producing

accurate flow rates ranging from 5 pL/min to 5.0 mL/min with a pressure

limit of 6000 psi. Along with flow rate, the refill speed (125 to 650

ms) could be controlled. Slow refill speeds were selected to yield a

pulsing flow while fast refill speeds were selected to yield a smoother

solvent flow.











Mass Spectrometer

A Finnigan triple stage quadrupole (TSQ45) mass spectrometer was

used. The El and CI spectra were obtained using a standard Finnigan

4500 EI/CI ion source with solids probe introduction of the samples.

Samples were heated and vaporized rapidly from 45 to 400*C in 3-4

minutes under temperature control. Electron energies of 70 and 100 eV

were used for El and CI, respectively. Source pressure for ammonia was

0.60 torr with the source temperature maintained at 140*C. Nitrogen

was used as the collision gas at pressures ranging from 1.2 to 2.5

mTorr. Collision energies of 25 to 30 eV were used. The electron

multiplier was typically operated at -1000 V for full scan mass spectra

and -1500 V to -1700 V for MS/MS spectra.

The mass spectrometer used was not dedicated to LC/MS resulting in

the frequent conversion back and forth between the standard EI/CI

source and the thermospray source. Thus, normal El tuning was

performed on a standard EI/CI source with the standard calibration

compound perfluorotributylamine (PCR Research Chemicals, Inc.,

Gainesville, FL) on the TSQ45 prior to installation of the thermospray

source. After installation of the thermospray source the lens voltages

were optimized with a 0.1 M ammonium acetate solution followed by a

solution of polyethylene glycol (PEG 200) (Sigma Chemical Co., St.

Louis, MO).



Vacuum Chamber Measurements

The vacuum chamber consisted of a modified Finnigan 1015 vacuum

console designed with entrance ports for the vaporizer probe,









temperature, pressure, and current probes, a six inch diameter

observation window, and a pump out line connected to a 300 L/min

mechanical pump. The pump-out line was situated 180* from the

vaporizer probe. Temperature and pressure measurements of the plumes

were obtained by placing the thermocouple (1 mm o.d.) and the pressure

sampling tube (2 mm i.d.) directly into the center of the plume,

perpendicular to the direction of the vapor and droplet flow from the

vaporizer probe. Current measurements were made using a Keithly 480

picoammeter (Cleveland, OH) obtained in the same manner as with the

temperature and pressure measurements with a 2" X 2" wire mesh screen

placed into the center of the plume. Sampling distances were varied by

adjusting the distance of the vaporizer probe from the sampling probes.












CHAPTER III
PRELIMINARY CHARACTERIZATION OF THE VESTEC THERMOSPRAY LC/MS INTERFACE

Preliminary characterization of the Vestec thermospray LC/MS

interface involved vaporization studies performed outside the mass

spectrometer where measurements of temperatures along the outside of

the vaporizer capillary were obtained. The resulting temperature

profiles were used to characterize the effects of tip temperature and

flow rate on vaporization and predict the state of the liquid inside

the probe under the various vaporization conditions typically

encountered with thermospray. These characterization studies also

involved the acquisition of thermospray mass spectra obtained under

"standard" thermospray conditions so that the features typical of

thermospray ionization mass spectra could be readily observed on our

LC/MS interface. These preliminary and general performance studies

were beneficial in providing results and information which were used

for the experiments discussed in Chapters IV and VI.



Vaporization Studies

The vaporization process of thermospray is responsible for the

production of a plume of droplets and vapor from a flow of liquid.

Flow rate and tip temperature are the necessary instrumental parameters

to achieve the partial or complete vaporization of a liquid mobile

phase. In these studies of vaporization, the effects of flow rate and

tip temperature on the temperature of the outside of the thermospray

vaporizer probe were determined. These studies involved experiments

which were performed outside the mass spectrometer on a vaporizer probe









with the outer (6.5 mm o.d.) metal sheath removed exposing the bare

(1.5 mm o.d.) capillary. From these studies information on the

processes occurring inside the probe during various conditions of

vaporization was obtained.



Characterization of Thermospray Plume Types

Two thermocouples are positioned along the capillary tube to allow

for temperature measurements at the point where electrical heating

starts (Tl), and at the tip of the probe (T2). These temperatures are

commonly referred to as the control temperature or T1, and the tip

temperature or T2. By measuring these temperatures on the vaporizer

probe as it is heated at a constant flow rate, a plot can be generated

which represents the resulting tip temperature (T2) versus the control

temperature (Tl), related to the power input to the vaporizer probe.

With these plots the various stages of vaporization can be illustrated.

We performed a "T2 versus TI" experiment with 1.0 mL/min of water to

illustrate these stages of vaporization and to help characterize the

thermospray plumes.

Figure 3.1 shows a typical T2 versus TI plot obtained for water

sprayed outside the mass spectrometer at a flow rate of 1.0 mL/min.

These plots are typically characterized by three slopes or regions

which represent the state of the liquid sample exiting the probe tip.

The first region (I) of the plot represents the steady heating of the

mobile phase. No vaporization occurs in this region as only liquid

exits the probe (tip temperature below the boiling point). The second

region (II) of the plot represents the partial vaporization of the

mobile phase. In this temperature region a fine mist of droplets and






























- 0 4J 0 0 ,.44 ,
,-4'0 0 0 r 4 4J 0C 0 .W

E4 d .- 4 0 1 o 3
S0 4.) 0 6 r- 4
M NN l>o n o


"4 c. 0. b4j 4)' to





0 0 --
1 0 0 C o .1



"o P w
b1 > > ca.

0. o O; 41 ) > o.







0-C ,4
4 0 0

0 *g -










0 .-4 -
00 0
0 ) 4 W 0 0)- w- $ 0 *







- 4 4-. .! 0 .^ 0 to 4J> 4-) .-4> 4J

























































l l ll i jlg l i ll lj i lli
0 0
O O 0
CN CN


LC)


II
0
0


I II I T I I I I
0 C


(0)ianjejedwoej dIjL


0
0






-o
--0
00
o
-(0







CN



0

o0
0



co






-N



-0


LM



CL

0




0



0


I








vapor are observed to exit the probe. As the tip temperature is

increased, a greater fraction of the mobile phase is transformed into

the vapor phase (wet to nearly dry plumes). The boundaries of this

second region are marked by two points, point A and point B. Point A

corresponds to the temperature at which vaporization begins (T2 -

boiling point of the liquid) while point B corresponds to the

temperature at which 100% vaporization of the mobile phase is achieved.

The third region (III) of the plot represents the complete vaporization

of the mobile phase. At these tip temperatures only vapor exits the

probe tip (dry plume) resulting in the rapid increase in tip

temperature.

From this study three distinct plume types can be defined with

respect to their apparent droplet content and physical characteristics

(see Figure 3.1). Wet plumes consist of relatively large droplets and

can be produced when operating at relatively low temperatures within

region II. These plumes appear as a dense cloud of droplets and vapor

accompanied by a "sizzling" sound as when water is sprayed on a hot

frying pan. When an object is placed within a wet plume at atmospheric

pressure, condensation occurs on the surface. Nearly dry plumes appear

to contain smaller droplets than wet plumes and can be produced when

operating at higher temperatures within region II. When an object is

placed within a nearly dry plume at atmospheric pressure, no

condensation occurs on the surface. Dry plumes consist of only vapor

(invisible plume) and can be produced when operating at temperatures

within region III. No condensation occurs on the surface of an object

when placed within a dry plume. These three plume types can be

reproduced for a given flow rate and will be referred to in Chapters








IV-VI. Further studies were performed which investigated the effects

of flow rate and temperature (power input) on vaporization and are

discussed below.



Variable Flow with Constant Heat

To determine the effects of flow rate on vaporization, temperature

was measured along the vaporizer capillary (1.5 mm o.d.) under

conditions of variable flow with fixed control temperature. A

thermocouple was placed directly on the side of the probe until a

stable temperature reading was attained (ca. 10 sec). A plot of probe

temperature versus distance along the probe across a 30 cm portion of

the probe at flow rates of 0.5 mL/min, 1.0 mL/min, and 4.0 mL/min is

shown in Figure 3.2. Increased flow rates result in higher

temperatures along the probe, higher pump back pressures, and a larger

plume of droplets and vapor (visually observed to be larger in diameter

and contain larger droplets).

We believe the increase in temperature along the probe with

increasing flow rate is due to compression of the vapor inside the

probe. As vaporization occurs at some point within the probe, the

volume increases enormously, resulting in the compression of this vapor

and an increase in pump back pressure. At higher flow rates the

compression is greater; therefore, the pressure increases. With the

increase in pressure, the average speed of the gas molecules increase

which results in an increase in temperature. This process is analogous

to the behavior of a gas contained in a cylinder with a moveable

piston. The pressure of the gas is the net effect of the collisions of

the moving molecules with the walls, tending to push the piston out.




































.0
0 -It




4-


bo


U0



C-4






4-

U41i




U)


U) 4
:s0
41i

co
4-

-4
















































o o
SCN


0 00 O


(0) eJnjejedwuejL


E
U

&)
.0
0



0


0
C

0)


I
) O








When the piston is moved downward into the cylinder, the gas is

compressed, resulting in the increase in temperature and the increase

in the average speed of the molecules.

At each flow rate a temperature maximum was observed at various

distances along the probe. For example, temperature maxima occur at

ca. 8 cm for the flow rates of 0.5 and 1.0 mL/min. Temperatures along

the probe decrease from this maximum towards the vaporizer tip. We

believe this cooling effect is due to the evaporation of the droplets

produced within the probe and the temperature maximum represents the

position within the probe corresponding to the onset of vaporization.

Temperature along the probe is maximum at ca. 17 cm for water at 4.0

mL/min. This shift in temperature maxima can be explained by the fact

that the same amount of power is input to the probe at 1.0 and 4.0

mL/min. With a greater flow rate we predict that an increase in the

power input would be required with higher flow rates to produce

temperature maxima at corresponding distances along the probe.

However, with a constant power input the temperature maximum shifts

towards the probe tip which indicates that the onset of vaporization

occurs at distances closer to the probe tip with increasing flow rates.



Constant Flow with Controlled Heating

To determine the effects of control temperature (related to power

input) on vaporization, temperature was measured along the vaporizer

probe under conditions of constant flow with controlled heating. Three

TI temperature settings were selected at a 1.0 mL/min flow rate of

water to produce the three thermospray plume types, wet, nearly dry,

and dry, which correspond to visual and physical observations made with








the probe spraying into an open room at atmospheric pressure. A plot

of probe temperature versus distance along the probe at conditions of

constant flow and controlled heating is shown in Figure 3.3. Three

control temperatures were selected which produced the three described

plume types.

The increase in power to the vaporizer probe at a constant flow

rate resulted in higher temperatures along the probe, higher pump back

pressures, and a more complete vaporization of the mobile phase.

Temperatures toward the end of the probe decreased for the temperature

settings corresponding to wet and nearly dry plumes. However, at the

temperature setting which resulted in the production of a dry plume,

the probe temperature increased. The temperature increases here

because only the limited heat capacity of the vapor (dry plume) is

available to absorb the input heat. The gradual decrease in

temperature along the vaporizer probe observed at various portions of

the plot is probably due the evaporation of the resulting droplets

within the probe (similar to the flow rate study discussed above). The

increase in HPLC pump backpressure with increasing power input is

probably the result of more vapor created within the vaporizer probe.

The excess vapor cannot escape as rapidly as when it was liquid,

resulting in increased backpressures.

From the discussion of the effects of flow rate on temperature

measurements along the probe, it would seem likely that increasing the

power input to the vaporizer probe at a constant flow rate would result

in increased vaporization efficiency while shifting the vaporization

onset position (temperature maximum) further away from the probe tip.

Although increased vaporization was observed with an increase in power






































4f -.4 -.





4 -



ca


0ca* 1-4
'-4 01)


Ca2




04J 0:

(n4 0

4j



0)

4j 0


0)


-'-4

00)



0 .



0 41 -

44 4






41







_L'
tro









-Ln
0




0 0 0 0 0
\ r -- -















(O C 0Q 00" 0l
VC) \ ^ .o



0)


o m q





aD -O
) OC


,,
00000000
CO WN OO LO 0









input to the probe, this shift in temperature maximum along the probe

was not clearly evident. A temperature maximum was observed at ca. 15

cm for a wet plume, while nearly dry and dry plumes each appear to

generate two temperature maxima which might be indicative of a sequence

of vaporization followed by condensation of the resulting vapor and

droplets followed by re-vaporization of the liquid. From these results

we cannot be certain whether this sequence is actually occurring or

whether this is just a reflection of the heating behavior of the probe.

A more detailed investigation into this heating process at various flow

rates might provide more conclusive results.



Thermospray Mass Spectra

The second portion of our preliminary characterization studies

involved the acquisition of positive and negative ion full scan

thermospray mass spectra of ammonium acetate, ribavirin, and several

purine and xanthine-type compounds. In cases where more fragmentation

was desired for structural information, daughter spectra were obtained

and the resulting fragmentations were interpreted. Spectral

characteristics of thermospray such as the observed ions of the

ammonium acetate buffer, formation of pseudo-molecular ions and

fragment ions are discussed. Experiments performed with ribavirin to

determine the reproducibility of positive and negative ion intensities

and relative abundances of the ions contained in the mass spectra are

described. Preliminary observations on the effect of various

functional groups on thermospray performance are discussed.









Ammonium Acetate

Positive ion and negative ion full scan thermospray mass spectra of

ammonium acetate obtained at 1.5 mL/min are shown in Figure 3.4.

Abundant positive ions at m/z 18, 36, 54, 60, 77, and 78 which

correspond to NH4+, NH4*-H20, NH4'-2H20, CH3CONH3+, CH3CONH2NH4+, and

CH3COOH*NH4+, respectively, were present in the positive ion mass

spectrum while abundant negative ions at m/z 59, 77, and 119 which

correspond to CH3COO-, CH3COO'H20, and CH3COOH-CH3COO", respectively,

were present in the negative ion mass spectrum. A negative ion was

observed at m/z 155 in the spectrum of ammonium acetate at a very low

relative abundance and corresponds to CH3COOH*CH3COO'-2H20.

Figure 3.5 shows the MS/MS daughter spectra obtained for selected

thermospray ions of ammonium acetate. The positive daughter spectrum

of m/z 36 shows the loss of 18 (H20) resulting in the formation of NH4+

at m/z 18. The positive daughter spectrum of m/z 60 shows the loss of

17 (NH3) resulting in the formation of the fragment ion at m/z 43 which

corresponds to CH3CO+. Here it is assumed that the positive ion at m/z

60 is not an acetic acid ion (CH3COOH)+ but rather (M+H)+ of an

impurity of acetamide, CH3CONH2. Likewise, the positive ion at m/z 77

is presumably not a M+ of ammonium acetate (CH3COO-'NH4+) but rather an

ammonia adduct ion formed from the acetamide impurity, CH3CONH2NH4+.

The positive daughter spectrum of m/z 77 shows two successive losses of

17 (NH3) to yield the fragment ions at m/z 60 amidee impurity) and m/z

43 (CH3CO)+. Another fragmentation observed was the loss of 59

(CH3CONH2) resulting in the formation of NH4+ at m/z 18. The negative

daughter spectrum of m/z 59 shows that this ion is very stable to

fragmentation. We presume this behavior to correspond the acetate ion



































cd


4 -

ra

U)



Uc,
















o
0 e
o-o
4-1


ca
U)

-4




-4
44








4103













Owi 0



























+
NX


0
0


0
a?


,_


0

+3c'


X
Z4


eouepunqv GA!We/IG %


1-


ej

0
0

0


0



0
0
0

0 -~


0

Ic'
0





s--


-w








- (N





N


. i I i I


eouepunqy eAulely1 %
























Figure 3.5

MS/MS daughter spectra obtained for selected positive and negative ions
of ammonium acetate. A collision energy of 25 eV was used with 1.2
mTorr collision gas pressure (N2).












NH4- H20



180.8





.0


120 140


12 149


I/E 28 40


- NH;


180.0














50.
T9.


NH

4 43
I


CH3CONH 2NH 4

77


S I I I I I
28 40 68 t o1o 120 148
















:O 40 E6 88 1Be 120 14








CH3COOH CH3COO

CH3COOH 119


20 408 88 1

CH3CONH;


W E .B


49 68 is


148








(CH3COO)-. The negative daughter spectrum of m/z 119 shows the loss of

60 (CH3COOH) to yield the acetate fragment ion at m/z 59.

All positive ions of ammonium acetate produced by thermospray at

these conditions contain an ammonium ion in their structure, while the

negative ions contain an acetate ion. Under these conditions the

positive and negative reconstructed ion currents (RIC) were essentially

equal. The origin of the acetamide impurity is not known. This

compound may be the result of a side reaction which occurs during the

commercial synthesis of ammonium acetate. Buffer solutions were

prepared from ammonium acetate obtained from two sources, Fisher

Scientific and Mallinkrodt. The resulting spectra were virtually

identical, each containing abundant positive ions at m/z 60 and 77

corresponding to the (M+H)+ and (M+NH4)+ of the amide impurity.



Ribavirin

The positive ion full scan thermospray mass spectrum of 1 pg of

ribavirin injected is shown in Figure 3.6. The spectrum contains ions

at m/z 113, 130, 245, 262, and 489 which correspond to the protonated

triazole-carboxamide ring of ribavirin, the NH4+ adduct of the

triazole-carboxamide molecule, (M+H)+, (M+NH4)+, and (2M+H)+,

respectively. The negative ion full scan thermospray mass spectrum of

ribavirin is shown in Figure 3.7 and contains ions at m/z 111, 171,

243, 303, and 355 which correspond to the loss of a proton from the

triazole-carboxamide ring, the acetate adduct of the triazole-

carboxamide molecule, (M-H)', (M+CH3COO)', and (M+lll)', respectively.

Although not observed in this particular mass spectrum, a (2M-H)" ion

at m/z 487 was frequently observed. In both the positive ion and







































444



a)



4- 4-





CO,



a 0)
>- '-
Pu(.4




Q)


'-4

co


Cicc
-4

44)


co

",4 CjEn


_-4 .,-
4>~-

.-4 (t 0)

o '-,4 4.J
PL4 I ~0


a)

'-4

0



u.

C0
L0


0)0




41) -4


0)


414




0
'.4


0
0



(-4
o0)


Cu$-

0C42
50)

0'41


















I'I










*o














Z j


+ _. "
-4 L
+-1c-----4





































bo 0-4 0~





(n





a) 0J~

0f C-4 a
0 044u



41
-4 W ca


0-r4 a) 41 040



W) Q- .44) En

bL0 0~~a
,C: r= (1) .4M
Zn 40 0~





















z
I





0
o=
1 z^ O
OP

O~


0
C)


N-


! I

t J'O

eouepunqV eA!J lel:G %









negative ion spectra the ammonium acetate ions are observed in the

background. Under the particular conditions at which these spectra

were obtained, the summed negative ion intensity of ribavirin was about

ten times greater than the summed positive ion intensity. The (M+H)+

and (M+CH3COO)" ions of ribavirin were present as the base peaks in

their respective spectra. In both ion modes abundant fragment ions

were formed. Especially interesting in the negative ion mode was the

addition of the triazole-carboxamide ring fragment (m/z 111) to a

molecule of ribavirin which suggests the occurance of gas-phase

ion/molecule reactions.



Reproducibility of Thermospray Spectra of Ribavirin

Because a mass spectrometer dedicated for thermospray was not

available in our laboratory, the standard EI/CI source was replaced

with the thermospray source prior to its use. The thermospray source

remained on the mass spectrometer for two to four days at a time and

yielded reproducible ion intensities and mass spectra during this time.

However, from week-to-week the relative abundances of ions observed in

the mass spectra would sometimes change as well as the relative

positive and negative ion intensities. These week-to-week fluctuations

are illustrated below.

The positive ion full scan thermospray mass spectrum of ribavirin

shown in Figure 3.8a was acquired three weeks prior to the spectrum

shown in Figure 3.8b. Although these spectra were obtained under

identical conditions of thermospray (instrumental parameters and

solution characteristics), the resulting spectra are significantly

different. The low relative abundances of the ions at m/z 113 and 489




































.-14 CO 4

CL 4-'4j




o o-
4-



0 4




U ,4 (-4
U ) 0

U) U)

ca


4








ca




a)~ 0




03 C.) 0


~ U ) En-


M U) En a)
tr) EE cduao
C~j 0 0
a) 0'
a) C"4








E -, aC




o~a Q)





0 o 40 1





a) a)



50) a4.)


0 4j~ ,



0) $4 Cd Q

_ 4 0) 0. t "C %



-4C.-4 r. L44 )










































C4

L



C4
<*------------, -------


tJ
('4
in
If- -- -


C.-
00.


eouepunqv eAIIeBIG %


I I l i l. .f l f l


eouepunqV GeA!IJl1 %









(observed in Figure 3.8a) are the major differences between the two

spectra. However, the relative abundances of the ions at m/z 130, 245,

and 262 remain relatively constant and the pseudo-molecular ion at m/z

245 remains the base peak. The negative ion full scan thermospray mass

spectra of ribavirin (Figure 3.9), also acquired three weeks apart,

showed similar differences as observed with the positive ions. The

decreased relative abundances of the ions at m/z 111, 243, 355, and 487

(observed in Figure 3.9a) are the major differences between the two

spectra while the pseudo-molecular ion at m/z 303 remains the base

peak.



Purine and Xanthine-type Compounds

As part of our preliminary characterization of the LC/MS interface

we investigated the effects of various substituents on thermospray

ionization. The compounds which were selected for this study have

structures similar to purine with most differing by only one or two

functional groups. These compounds can be divided into two groups:

one group consisting of substituted purines and the other consisting of

substituted xanthine or substructures of xanthine (Figure 3.10).

Purine along with pyrimidine are the parent compounds of the two

classes of nitrogenous bases found in nucleotides and consist of a

pyrimidine ring and an imidazole ring fused together. Xanthine is a

degradation product of purines and has a core structure similar to

purine.

Both positive ion and negative ion full scan thermospray mass

spectra were obtained from standard solutions (50-100 ppm) of each of

the purine and xanthine compounds. Most of the compounds showed



































4 WW0CN U 0
> 01 W (U ~ -0 -9



o o

4 ',4 .J4 0e
44 41 44 '


0, r-4 $
SW (1)

Ca 41J j"j
W E1--Wca cN i

(U ca J a) w
3: 0 a




cn 0W W'L
~ca cn U 04 c4 a)



ca W W.-
(D- 4) (UN Q.

En 41 U) 04.)


ca cd cz I -4 En(1
Zj (UU






























,w


In


In
0in


01


eOUBpunqV eAIlBelf %


| | i | | = = = |


I I I I i I I I I


-*- _._ i__mi_ i III-


lilli


eouepunqy eAII81e %

























Figure 3.10

Model compounds used for determining the effects of various
substituents on thermospray ionization. Note that most of the
compounds contain the purine core structure.







H


'N:-N
Purine NW 20
Purine (MW 120)


2-Aminopurine
(MW 135)


NH2 H
N NY',, ,


Adenine
(MW 135)


2-Hydroxypurine
(MW 136)


2,6- Diamino-8 -purinol
(MW 166)


6-Thioguanine
(MW 167)


6-Mercaptopurine
(MW 152)


0 H
HN N"


H
Xanthine (MW 152)


0 H
N



Hypoxanthine
(MW 136)


SH H



H
6-Thioxanthine
(MW 168)


0 H

HNN

H
Uric Acid
(MW 168)


0
II H
H2NCNH N0
HO
0


Alloxan Monohydrate
(MW 160)


Allantoin
(MW 158)









characteristics in the mass spectrum typical of thermospray ionization;

abundant pseudo-molecular ions with very little fragmentation.

Molecular ions appeared as the (M+H)+ and/or (M+NH4)+ in the positive

ion mass spectra and as the (M-H)- and/or (M+CH3COO)" in the negative

ion mass spectra. The instrumental parameters and the results of the

positive ion and negative ion thermospray mass spectra are summarized

in Table II. These ions were present in the mass spectrum with ion

intensities greater than three times the background ("chemical noise").

Thermospray of purine yielded the (M+H)+ ion in the positive ion

spectrum and (M-H)', (M+CH3COO)', and (2M-H)" ions in the negative ion

spectrum. All of the substituted purine compounds show the same

behavior as purine in the positive ion mode and yield a (M+H)+

molecular ion. Two of the substituted purines, 2-hydroxypurine and 6-

mercaptopurine, also produced (M+NH4)+ ions. 2-hydroxypurine was the

only purine-type compound which contained a hydroxy group in its

structure. 6-thioguanine has a structure very similar to 6-

mercaptopurine but did not form an ammonium adduct. The only

difference between these two structures is the presence of an amine at

the two-position on 6-thioguanine.

All of the substituted purine compounds yield (M-H)- ions in their

negative ion mass spectra. In addition to (M-H)" ions, some compounds

(2-aminopurine, adenine, and 2,6-diaminopurinol) formed (M+CH3COO)"

adduct ions. The structure of these compounds differ from purine by

the addition of an amine group (two amine groups in the case for 2,6-

diaminopurinol) at various positions on their structures. Purine,

hydroxypurine, 6-mercaptopurine, and 6-thioguanine do not show the

























I 0





0-U 0 -4 -0 0
0rt Y 014 <410 0

.-I 0 % I- F qw .-
r, f~n -I r.- 04 ,4V r .-4 .-M











4
+ + + 4




r-,
o 0 0 "
o 0 0 .


4 en (n *
M3 r- m


1"4









0
1*4












I


o


0















E|4


Cl)
-4


I I


I I _'1 I I

-. Il |








- I I -I-. 4
o' 0 -0 -. 00
0 0o 0 0 0-0 0


-4 14 o 'ON n ro 4 4
*- W- Ifl (M i- U*l l-l >%












t- +




0- 0 0 0
1 r-

Ln o, wo % r, to
unr- 'C '4 ri -n
1-1 r4 r-4 l- c4 r-


0 %0 r-4

0O r-
r In 1 1
1-1(' n -l .-l.











+



+' '0.



in r
14-1 4


IA t rI ko N4 %0 co co co 0
V) ( U %o 3o In n '3 '3m %0 %3
4 1 i4 -4 -4 14 r4 1 r-4 i-4 -


cI 'j I


-I C~J
Li -




63
















~ 4- 4 4
A0 0




'- r4 -





*4J
t) U -)

00

00 A



I :, 9.2 I

o2
,?Q 444I









addition of acetate in their negative ion spectra nor do they contain a

primary amine group on their structures.

Xanthine, 6-thioxanthine, uric acid, and alloxan did not produce

any positive thermospray ions. A common feature to these compounds is

that their structures contain more than one doubly-bonded

electronegative group such as oxygen or sulfur. Hypoxanthine, with a

similar in structure to xanthine with the doubly-bonded oxygen at the

2-position removed, yields (M+H)+ and (M+NH4)+ ions. Similarly,

allantoin (an oxidation product of uric acid) also yields abundant

(M+H)+ and (M+NH4)+ ions. Xanthine, hypoxanthine, 6-thioxanthine, and

allantoin yield (M-H)- negative ion species. Alloxan monohydrate

produced negative ions at m/z 113, 115, and 175. The ions at m/z 113

and 115 correspond to fragment ions for the loss of HCOOH and CO2,

respectively. The ion at m/z 175 corresponds to an acetate adduct of

the 115- fragment ion. Both hypoxanthine (which contains no primary

amine functionality in its structure) and allantoin yield (M-CH3COO)"

ions while uric acid did not produce any negative ions.

Figures 3.11 and 3.12 show the resulting ion intensity summation

profiles of the abundant positive and negative ions of these purine and

xanthine-type compounds and ammonium acetate. These chromatograms

represent a single 25 AL flow-injection of each compound. Several

sharp downward spikes are present both profiles. These spikes

correspond to the switching of the sample injection valve from load to

inject or vice-versa. These spikes occur since the switching of the

injection valve causes a momentary interruption of the liquid flow and

thus, the signal decreases abruptly. A rapid switching technique will

eliminate these spikes while a slower, more deliberate switching will





























) (U 0) Q0.) 0 -s 4-4 )
r -. 4" ) .4 U -Y r-4 (L) O' CM 0 .4
-r4 1 c4 Ci uJ C o 0 I
4 4-1 Z Q) r. -4-



0 u H p P O
41 ,c 0) 0Ca)- C )





-4 0 -4


0 0 0
o g a: -, o O.. o






o C 3 3
4-4 4. o 0


0 C 0 u C OO


0 .4 0










S. 0 4 0
Q) 0 0 CN '0' 4-1 Z
W 4-4










0 M E*
2 4 0)- 4 <








U) C )X ,)GJUC -H cn
O Ci C) w O S CV) -4
o < i 0 ^ 0c Co 0 10




i 4 0 4.1 U '-' C4 C1






























i
-3



co -







co o


C03


04


juaJJno uol


luejjno uol


Old



































4)o U ) P. ()
Z -4 414 .,-q C.
-r 4 CU 4-J -J

WNa 0
(j C cn








0 a)


*1 0 C14
o4 4. 0
~-0 4) a


-4 -4 9 :

.,4 4 41



w 0U

oz u, a)a
co E a

0 ~






o 1 0

P.- -4 E. 0 -

.,- V-4 Z

CU O4-)~

b C'-,4 C*


0 CU4iO )


W)


'10 a1) %D CC

C-4 .4 0
a -4

P)-'. ca
c') 0



0. 4) CU.
IL -4 E
CO
"-4 r=
f-C I bOCo





0 a)
-4~

r---4-


0 0


0
.,.j
41
"4
(a
0


.f0 "'4 CUC
0 Z # 0
W .) CN 4
CC CC




o _4

0.

.-



co 0 .'4-
5-4 P.0 0

(0 a) "~4 (J x
(4 w C- 0
a) W- CO-.-
P- 0 "'4
0. 0 '1 4.)


E: 4.4 a)
0 0,4

r4-4 44
-4 -,4 0
a) 4-4




S4.)


0 (L.) a.


4.J .

M1 -'-4 0

Qa)

0. 4.)


01 4J-,4






zn


-'a)



ca





ca)4.) o

a0

CU0 0
5.4 (L)U


0' r

E C14 CU w)




































C14 __ _









00



o


LO







.CV) ,-



CM


juaJjno uol


-o













LO


772


oDi


luaijno uoI









result in these spikes. These profiles illustrate the rapid nature of

thermospray with a flow injection technique where 13 samples were run

in approximately eight minutes, which corresponds to 100 samples per

hour. The RIC appears significantly elevated above the zero and does

not increase significantly for any of these compounds. This is because

the scanning interval of the mass spectrometer included the masses

corresponding to the ammonium acetate buffer ions which were present

throughout the acquisition.

It is interesting to note that the ion intensity profile of the

ammonium acetate ions decrease during the ion production of the purine

and xanthine-type compounds. Thus, it seems that the sample ions are

produced at the expense of the ammonium acetate ions. Whether this

ionization behavior is the result of direct competition between the

sample and ammonium acetate in the direct ion evaporation mechanism, or

a depletion process involving sample molecules and ammonium acetate

ions (reagent ions) in a CI mechanism, or a combination of both, cannot

be concluded here.

These preliminary investigations show the effects of various

substituent groups on the thermospray ionization of purine and

xanthine-type compounds. These results indicate that particular

functional groups present in the sample compound structure can

influence the thermospray mass spectra. Structures which contain at

least two thiol or doubly-bonded oxygen groups generally do not produce

positive ions by thermospray ionization. Allantoin is an exception to

this observation. We believe that the presence of a primary amine

group in the allantoin structure helps to overcome this effect of the

doubly-bonded oxygens and promoted the formation of both (M+H)+ and





70



(M+NH4)+ ions. The presence of amine groups combined with the absence

of thiol or doubly-bonded oxygen groups seem to promote the formation

of (M+CH3COO)" ions in the negative ion thermospray ionization mode.

Uric acid, which contains three doubly-bonded oxygens on its structure,

does not produce any positive or negative ions by thermospray

ionization.










CHAPTER IV
THERMOSPRAY PLUME MEASUREMENTS IN VACUUM

Direct physical measurements of thermospray plumes were made in

vacuum to obtain information on the droplet evaporation and droplet

charging processes of thermospray. The measurements were made in a

specially designed vacuum chamber which permitted visual observation of

the plume, as well as sampling of pressure, temperature, and current

within the plume. Because this chamber was considerably larger (8.5 cm

i.d.) than the Vestec thermospray source (1 cm i.d. with an ion

sampling cone and other structures protruding into the source), the

effect of the source geometry may change the thermospray plume

behavior. Thermospray plumes were generated with the Vestec vaporizer

probe as measurements were obtained at positions along the center of

the plume in the manner described in Chapter II. The results of these

plume measurements are discussed below.



Plume Pressure Measurements

Plume pressures remained constant for all distances measured within

the plume at constant flow rate. Pressures of 2.4, 2.6, and 2.8 Torr

were obtained with flow rates of 1.5, 1.75, and 2.0 mL/min water,

respectively. Other instrumental parameters had no effect on pressure

measurements.



Plume Temperature Measurements

Plume temperature measurements were obtained at various tip

temperatures, manifold temperatures, flow rates, and manifold

pressures. (The adjustment of manifold temperatures and manifold









pressures approximated the effects of source temperature and source

pressure.) Much condensation (in the form of ice) was observed to

occur on the thermocouple at distances close to the probe tip (<1 cm)

while no condensation was observed to occur on the thermocouple at

distances further from the probe tip. The formation of ice on the

thermocouple would likely result in low temperature readings until the

ice either falls off or melts away at which point the temperature

readings would be expected to increase to temperatures more

representative of the surrounding vapor. Therefore, low and perhaps

fluctuating temperature readings may be indicative of a droplet-rich

plume, while stable or constant temperature readings may be indicative

of a vapor-rich plume. For these measurements no specially designed

temperature sensing probes were used. Therefore, these measurements

would be expected to provide only an approximation of the actual plume

temperature. Essentially, these measurements provide a means of

monitoring the production of droplets within the plume.

A plot of plume temperatures measured at various distances along

the plume for the three plume types at a flow rate of 2.0 mL/min is

shown in Figure 4.1. All three plume types behave similarly with the

greatest temperature fluctuations occurring at distances less than 1.5

cm. Beyond this distance the temperature readings become more stable.

Wet plumes produce the lowest temperature readings while dry plumes

produce the highest temperature readings. Thus, the vapor-rich

portions of wet, nearly dry, and dry plumes are attained at virtually

the same distances away from the probe tip. Also, higher tip

temperatures result in higher temperatures within the vapor region of

the plumes.






























4-J


C/)




'-4





















r-4
co



a-)

ca




'4J

ca


CO 4



4j4




4-

























































0 0
CN c(


0
C"j





0-
6



0
CM


0
U-n


0
0


I lilt II jilt I I liii


0
U~)


(0) eJneJedwjei.


L6












lo
C?
-0n
0





Mq
-^










l0
- r












-N





-



o6


0

a.




0
I-.
0


CL
E
0
a-
U.





U)









Figure 4.2 shows the plume temperatures measured at various

distances along a nearly dry plume produced at 2.0 mL/min for three

manifold temperatures. With increasing distances from the probe tip

the resulting plume temperature readings surpass the manifold

temperature. It appears that higher manifold temperatures result in a

more rapid attainment of stable plume temperatures. Therefore,

increased manifold temperatures are likely to result in a more rapid

droplet evaporation process and the attainment of droplet-rich plumes

at distances closer to the probe tip.

Figure 4.3 shows a plot of plume temperatures measured at various

distances along the plume for nearly dry plume types produced at

various flow rates. Plumes produced at flow rates of 1.5 and 1.75

mL/min show the greatest fluctuation in temperature at distances less

than 3 cm from the probe tip. At distances beyond 3 cm the measured

temperatures were more stable. Plumes produced at 2.0 mL/min show the

greatest fluctuation in temperature at distances less than 1 cm from

the probe tip. At distances beyond 1 cm the measured temperatures were

more stable. Thus, it appears that higher flow rates may yield vapor-

rich plumes at closer distances to the probe tip than plumes produced

at lower flow rates. This faster evaporation of droplets with

increasing flow rates may be due to the increase in tip temperatures

with higher flow rates.

Figure 4.4 shows the plume temperatures measured at various

distances along a nearly dry plume produced at 2.0 mL/min for manifold

pressures of 2.8 and 760 Torr and a manifold temperature of 24C.

Plume temperatures increase with increasing distances from the probe at

2.8 Torr, however, at 760 Torr plume temperatures decrease with






























0


z

















-4



-4






bo~
Z-4
0.



-4)


c~o0
0 '4,4




r.


0Ei

Z- C

















































I I I I I l l I I I l I I I I
0 0 0
0 Lo 0


Q)0Q
to to
olc
04 ..V.


cyj


Ii I I I O


(C3) einlyajodwe.j


0


.0





lo
















O
-0


- C
lq


10%%
E
0





0.
0.
I.


E
0
I=
U.


c
C)

Cu
(0


I





I













Cl (0




N o 04 3 t
CLO % D C
CM.




-j-

E



0



(IL

0












0 0C 0 0 0 0
00 00(N 0 CC



(3 fimejedweL


































ca




10
Q.)

0


0)







U)















,4.









r.)

,-4 r.



44~
















N CC\ (D





co ot
~E


ca ::



0
CL

E

0

U..


0 ... .. .. I I... ..1 111VVTT rr1
0 0 0 0 0
00 LO I- (N


(C3) einjejedwoej









increasing distances along the probe. One reason which may account for

this behavior at high pressures may be the fact that at 760 Torr much

condensation occurs at all distances measured along the plume. This

may be the case since at 2.8 Torr, the probe temperature is enough to

maintain a "warm" spray with little condensation occurring. At 760

Torr, however, higher pressures combined with a manifold temperature of

only 24*C results in a "cool" spray with much condensation occurring

within the plume.



Plume Current Measurements

Plume current measurements were obtained at various tip

temperatures, manifold temperatures, flow rates, and manifold

pressures. As with the plume temperature measurements, much

condensation was observed to occur on the wire mesh used as an

electrode at distances close to the probe tip (<1 cm), while at

distances further from the probe tip no condensation was observed to

occur. Condensation on the wire sampling mesh would affect the current

measurements by changing the fraction of the droplets/vapor which are

intercepted by the mesh (nominally 80% transparent). We believe that

the plume currents measured here are most representative of the droplet

charging process and that it is unlikely that ion formation (if any) in

the gas phase contributes significantly to these current measurements

of water plumes.

Plume currents measured at various distances along the plume for

the three plume types at a flow rate of 2.0 mL/min of water are shown

in Figure 4.5. Wet and nearly dry plumes show similar plume current

behavior with distance sampled and produce negative currents of




























0
CN










ca)










r.



0

2-4






0)









:J0





84





0


0 0 0 "




.0






\ I-
o c :

\\ 0 0 < 10 0


0C 0
0














\ ~ -- 00
00

I I I -









approximately -500 nA at 0.5 cm. The magnitude of these plume currents

decrease with increasing distances from the probe tip (as droplets

evaporate to produce vapor) and gradually approach zero. Dry plumes

generate little current (<20 nA) and shows virtually no fluctuation

with distance sampled.

Figure 4.6 shows the plume currents measured at various distances

along a nearly dry plume produced at 2.0 mL/min for three manifold

temperatures. These plume currents generally show similar behavior

with distance sampled. The largest negative currents were observed at

0.5 cm from the probe tip and decrease to current values approaching

zero. Lower manifold temperatures have less effects on the plume

current readings. This is probably due to the increased condensation

which occurs at lower manifold temperatures.

Figure 4.7 shows a plot of plume currents measured at various

distances along the plume for nearly dry plume types produced at

various flow rates. Plume currents were greatest at distances close to

the probe tip. As the sampling distance was increased, the resulting

plume current gradually approached zero. Higher flow rates produce

more droplets and thus, yield higher plume currents.

Figure 4.8 shows the plume currents measured at various distances

along a nearly dry plume produced at 2.0 mL/min for manifold pressures

of 2.8 and 10 Torr. Plume currents are initially negative at distances

close to the probe tip and change to positive values at distances

further from the probe. At 10 Torr plume currents change from negative

to positive at ca. 1 cm while at 2.8 Torr plume currents change from

negative to positive at ca. 2 cm.

The results from these plume current studies suggest that water
































(4.







0












(0
0)












bo w
z04-

0 4


)r--4m




4) 4


~0

04 )


















0


C'







OO

0 0
o |


0 O
n


0






.0
-o








0









i
_o



0



-N








0





K
I


(VU) lueJJno


0
0
0
CL

0.
E
0





(0
U.

Q


n 0


0
0


fill, 11


I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1


I