Skip to main content

Full text of "Characterization of atmospheric pressure ionization/tandem mass spectrometry for direct atmospheric analysis"

See other formats


THE CHARACTERIZATION OF ATMOSPHERIC PRESSURE lONIZATION/TANDEM 
MASS SPECTROMETRY FOR DIRECT ATMOSPHERIC ANALYSIS 



by 

KENNETH PAUL MATUSZAK 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL 

OF THE UNIVERSir/ OF FLORIDA 

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE 

DEGREE OF DOCTOR OF PHILOSOPHY 



UNIVERSITY OF FLORIDA 
1988 



To my parents with love. 



ACKNOWLEDGEMENTS 

I wish to express my sincere appreciation to Dr. Richard A. Yost 
for his guidance and direction during the completion of this work. I 
would also like to acknowledge and thank the members of my graduate 
committee, Dr. John Eyler, Dr. Gar Hoflund, Dr. John Dorsey, and Dr. 
Cliff Johnston. 

I acknowledge the U. S. Air Force Engineering and Services Center, 
at Tyndall, AFB , which provided the entire support for this work. 

Many thanks go to the departmental machine, electronics, and glass 
shops, including Russ Pierce of the electronics shop who bailed me out 
several times with his insights, knowledge, and ability to explain 
electronics on my level; Steve Miles of the electronics shop who 
refined the design and oversaw construction of the corona discharge 
current-regulated power supply; Chester Eastman, Vern Cook, and Dailey 
Burch of the machine shop who constructed almost the entire source and 
took time out of their schedules for last minute changes during crunch- 
time; and Dr. Sam Colgate for his advice on vacuum system design. 

I thank my fellow researchers, especially Mark Hail who helped 
extensively early on in this project, Dr. Jodie Johnson for his mass 
spectrometry advice, Mark Bolgar who assisted on the project, my fellow 
group members, especially Mike Lee, my roommates Bryan Hearn and Mike 
Cockman for putting up with me this past year, my friends here at UF, 

iii 



especially Paul and Anne McCaslin, Regina Cheong, and my little bro' 
Jeff Keaffaber. 

Sincere thanks go to my family, especially Mora and Dad, who have 
been so supportive, my friends back home in Chicago, and the members of 
St. Augustine Catholic Church and Inter-Varsity Christian Fellowship 
for uplifting me in their prayers. 

My most special thanks go to the person who held me up here at the 
end, who wouldn't let rae quit, who took care of me when I needed it, 
put up with me even when I was a jerk, made my days bright even when 
they were dark, and never ceased to shower me with love, my very soon- 
to-be wife, Laura Elizabeth Boudreaux (I love that name). 

Last but not least, I thank the Lord Jesus Christ my Savior, 
because it was only by his grace that I had the ability and strength to 
finish. 



IV 



TABLE OF CONTENTS 

PAGE 

ACKNOWLEDGEMENTS iii 

ABSTRACT vii 

CHAPTER 

1 INTRODUCTION 1 

Analytical Scenario 1 

Overview of API/MS 2 

Overview of Thesis Organization 13 

2 DESIGN OF THE API SOURCE 15 

Overall System Design 15 

Initial Source Design 18 

API Source Configuration for Finnigan 4500 25 

API Source Configuration for Finnigan TSQ 70 27 

3 EXPERIMENTAL RESULTS AND SOURCE DEVELOPMENT 34 

Results with API Source Configuration for 

Finnigan 4500 Mass Spectrometer 34 

Adaptation of the API Source to the 

Finnigan TSQ 70 Mass Spectrometer 42 

Experimental Physical Parameters of API Source 

and Mass Spectrometer 58 

Clustering/Declustering 74 

Experimental Methods for Direct Atmospheric Monitoring. . . 80 

Sampling Methodology 81 

4 ANALYTICAL PERFORMANCE 82 

Characterization of the API/MS /MS Instr\iment 82 

Analytical Potential 94 

5 CONCLUSIONS AND FUTURE WORK 104 

Summary of Results 104 

Evaluation as Direct Atmospheric Analyzer 106 

Suggestion for Future Work 107 



LITERATURE CITED ;L14 

BIOGRAPHICAL SKETCH nj 



VI 



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 



The Characterization of Atmospheric Pressure lonization/Tandem 
Mass Spectrometry for Direct Atmospheric Analysis 



by 

Kenneth Paul Matuszak 
April, 1988 

Chairman: Richard A. Yost 
Major Department: Chemistry 

Atmospheric pressure ionization (API) has been shown to be very 

useful when combined with tandem mass spectrometry in performing direct 

atmospheric monitoring of trace compounds. The design and development 

of a new API source that has been developed to be compatible with a 

commercial, turbomolecularly-pumped triple stage quadrupole (TSQ) 

tandem mass spectrometer (MS/MS) is presented. This API/MS/MS 

instrument has been used to study (1) direct atmospheric sampling by 

mass spectrometry, (2) the production of ions in the atmosphere, (3) 

the effects of supersonic expansion on the clustering of these ions 

with neutral species, (4) the effectiveness of declustering of these 

ions by collisional means, and (5) the ability to qualitatively 

identify atmospheric gases. These studies will seirve to lay the 

vii 



groundwork for the development of an instrument capable of performing 
trace analysis while performing direct atmospheric monitoring. 

This API source utilizes a corona discharge for the primary 
ionization of the atmospheric molecules and an orifice system to 
selectively leak the ions into the mass analysis region. It has been 
constructed so that it is interchangeable with a standard electron 
ionization (EI)/chemical ionization (CI) source on a Finnigan-MAT TSQ 
70 triple quadrupole tandem mass spectrometer. In developing this 
source, a program to model the ion optics (SIMION) has been used to aid 
in the design of the post-orifice lens configuration. 

The design of this source is such that memory effects have been 
essentially eliminated and the clustering of ions with molecules in the 
post-orifice region has been significantly reduced. Samples that have 
been studied include a variety of laboratory solvents . These solvents 
have been injected into streams of pure nitrogen carrier gas and have 
been analyzed under direct atmospheric conditions by bringing the caps 
of the solvent bottles near the sampling region of the API source. 

Recommendations are made for overcoming the limitations of the 
present API source design and lens system. These include the 
consideration of supersonic jet expansion theory to redesign the post- 
orifice lens system and additional modifications to the discharge 
needle svstem. 



Vlll 



CHAPTER 1 
INTRODUCTION 

The purpose of this work is twofold: first to design, construct, 
and develop, a new atmospheric pressure ionization (API) source for a 
turbomolecularly-pumped triple stage quadrupole (TSQ) tandem mass 
spectrometer (MS/MS) , and second, to characterize and apply this new 
source to study (1) direct atmospheric sampling by mass spectrometry, 
(2) the production of ions in the atmosphere, (3) the effects of 
supersonic expansion on the clustering of these ions with neutral 
species, (4) the effectiveness of declustering of these ions by 
collisional means, and (5) the ability to qualitatively identify 
atmospheric gases. These studies will serve to lay the groundwork for 
the development of an instrument capable of performing trace analysis 
while performing direct atmospheric monitoring. Such an instrument 
could be used as an airport "sniffer" for drugs and explosives, or as 
an on-line vapor detector in the microchip, incineration, or chemical 
production industries. The major use of such an instrument, however, 
is anticipated as an environmental monitor for chemical waste spills 
and chemical dump sites or other forms of environmental contamination. 

Analytical Scenario 

Because this API source is interchangeable with the standard 
electron ionization (EI)/chemical ionization (CI) source for a 



2 

conunercial tandem mass spectrometer, one such use would be as a 
preliminary pollution detection device, that is, to use this instrument 
to perform direct atmospheric analysis, sampling at various locations, 
and determining possible contamination of the air at each area of 
analysis. Thus, the instrument in that configuration would be able to 
perform preliminary identification of the pollutants present, as well 
as quantitation to determine the extent of the contamination. Once a 
site of contamination was found, the API source could be interchanged 
with the standard EI/CI source and the instrximent in this configuration 
could be used to analyze soil and water samples in the surrounding area 
with ionization techniques which are already established and 
standardized for these two forms of sample matter. In this way, the 
extent of contamination could be mapped out in the full environmental 
scheme of air, water, and soil. 

Overview of API /MS 

Background and History of API /MS 

To perform the analysis of atmospheric gases by mass spectrometry 
there are two basic approaches. Because a mass spectrometer operates 
under high vacuiom, in the case of a quadrupole instrument at 10'^ to 
10"' torr, the pressure of the atmospheric sample must be vastly 
reduced. The first approach (Figure 1.1) is to bleed the atmospheric 
sample through an orifice or membrane and use high-vacuum pumps to 
reduce the pressure to that typically used in the ionization region for 
normal mass spectrometry. Unfortunately, while doing this, most of the 



Vacuum To Atmospnsrlc 
Interface 



8 B 
B„ B 



B B 

S B B 
B " „8 R B - 
B 8 B B^ B 

B B 8 8 t^ g 

B fl B B B 

B B 8 B B 8 B 

B ■ B 8 B^fl e 

B 8^ SB 8 
B ^ 



B B^ B^ 

Pj ^ R fl 

b" 

Ionize 



Pu 



m p 



Analyze 



Figure 1.1 

An atmospheric sampling mechodology in which gas is reduced -o r.orTsal 
mass speccromecric pressures and Chen ionized. 



analyte is pumped away with the excess gas. For trace analysis this is 
obviously undesirable. 

The second method (Figure 1.2) is to somehow ionize the analyte at 
atmospheric pressure and focus the ions through a small orifice into 
the mass analysis region by means of electrostatic fields, while 
pumping away the excess gas. 

API is a unique ionization technique for mass spectrometry. 
Ionization occurs outside of the mass analysis (high vacuum) region of 
the mass spectrometer and the ions are permitted to enter the high 
vacuum region, usually via a sub-millimeter orifice in a thin 
diaphragm. 

Molecules in an atmospheric pressure gas can be ionized by passing 
high energy electrons through the gas allowing, the gas molecules to 
collide with these electrons. This ionization process can produce both 
positive and negative ions. Positive ions are formed when an electron 
(e') collides with a gaseous molecule (G) and causes a second electron 
to be ejected, as in Equation 1.1. 

G + e" - G"^ + 2e- (1.1) 

Negative ions are generated when a low energy (near thermal) 
electron is absorbed by a molecule of the gas, as in Equation 1.2. 
This is called electron capture ionization. 

G + eth" - G- (1.2) 



Vacuum To ATmoaphdric 
InTerfac© 



B 



B B B B . B B 

B B%8 „.B*B B 
B a%B^ B B/b 

^ B B* ^ S B . ^ . + 

B B B B^ 2 ^ ^ 
A B B a' B B B 

B B ^ B 
B B 



B^ 
B Q " -*°B ^ 



B 



Ionize 



B 



B 



B 



B 



B 



B 



B .a"" ^ 

+ A A^ A 

A " .+ A + _+ A 



B -^ 

+ A 



B 



B 



B 



B. 



B 



A B 



B 



pump 



Ana I y z a 



Figure 1.2 

An atmospheric sampling raechodology in which sample is ionized, sample 
ions are focused by electroscacic fields into mass analyzer, and 
gas is pumped away. 



6 

These ions can further react with background molecular species to 
form background reagent ions. In the case of direct atmospheric 
analysis, positive ions are formed when N2 and H2O molecules in the air 
react with the high energy electrons (Equations 1.3 and 1.4). 

N2 + e- -* N2'^ + 2e- (1.3) 

H2O + e* -» H20'^ + 2e" (1.4) 

These ions can further react by charge exchange and proton transfer 
reactions through collisions with other background molecules (Equations 
1.5 and 1.6) to form reagent ions. 

N2'^ + H2O ^ N2 + H2O+ (1.5) 

H20'*" + H2O - H30"^ + HO (1.6) 

Sample molecules (M) can react with these reagent ions by proton 
transfer or charge exchange to form molecular (M^) or pseudo-molecular 
(M+H''") ions (Equations 1.7 and 1.8). 

M + H20'^ ^ M"*" + H2O (1.7) 

M + H3O+ - M+H+ + H2O (1.8) 

Ion clusters can also be formed when ions associate with neutral 
species before the orifice or in the supersonic expansion in the post- 
orifice region. Low molecular weight molecules readily form clusters 
with ionized species. Thus sample and hydronium ions may cluster with 
either sample or water molecules (Equations 1.9-1.12). 



7 
M+ + M - M2"^ (1.9) 

M"*" + H2O - M+H2O+ (1.10) 

M+H"^ + H2O - M+H3O+ (1.11) 

H3O+ + H2O - H(H20)2'^ (1.12) 

In the preliminary studies with a commercially developed [1] but 
never marketed API source in an attempt to perform direct atmospheric 
monitoring, as many as 20 water molecules have been observed to cluster 
with a hydronium ion (Equation 1.13) to form a distribution of water 
cluster ions (Figure 1.3). 

H30"^ + nH20 - H(H20)n+i"*' where n = 1-20 (1.13) 

At atmospheric pressure, molecules will undergo multiple collisions 
and those with the highest proton affinity (for positive ions) or 
highest electron affinity (for negative ions) will quickly become 
ionized by means of charge or proton transfer with the ions from the 
bulk gas. Because this is a collisional energy transfer, API is a low 
energy process and therefore little fragmentation of the molecular 
analyte ion will occur. Also, due to the multiple collisions, chemical 
and thermal equilibria will be established in the gas and thus there 
should be nearly a 100% ionization efficiency for analyte molecules 
that possess these characteristics. These factors should lead to a 
very high sensitivity. High selectivity is also obtained for certain 
classes of compounds because of this non-democratic process. 
Sampling of Ions Formed at Atmospheric Pressure 

Some of the earliest cases of performing mass spectrometric 
analyses upon ions formed at atmospheric pressure were performed by 



o 

o 



cs 



u 

c 



o 



0^ 



ir. 
c 

E 






vO 



o 
o 



OS 



vn 



M 






(S 



4# 
O 
O 



o 





CI 




CO 


Q 




U 




s 


+ 


cs 


^m 


2 


c 




o 


w 


fN 


cs 


CS 


Ul 




H 


+ 


W 


a 


s 




d 






M 


a: 
u 


^ 


H 




< 


>4 


» 


a 



VA 



m 



n 

s 






O 

o 



f^ — 



£ia 



a. 




o 












•c 







<u 


Urf 


o 


^ 


u 




3 


w 


O 


c 


« 


c 




•H 


1— 1 




cu 


x^ 


< 


0) 




IJ 


C_i 


0) 


< 






^ 




f-^ 


1 


y 


»— 




C5 


^ 


M 


o 


•M 


>-i 


2 


n: 


(— 


3 


—4 




U 


oj 




o 


^ 




_i 


s 


—4 


3 


3 


W 




;j 


=a 


u 


^ 


« 


•«^ 




U 


Ul 


o 




;-) 


If) 


• ■H 


u: 


^ 


c 






aoirepanqv SATacT^H 



9 

Knewstubb and Sudgen [2-5]. They used a 50 ^m sampling orifice in a 
thin platinum foil attached to a water-cooled block, sampling ions 
produced in an atmospheric flame, to study the combustion chemistry of 
flames. Their system utilized three stages of differential pumping to 
reduce the pressure from atmospheric down to less than 10"° torr in the 
analyzer region of their magnetic deflection mass spectrometer. 

Kebarle and co-workers [6-8] used a-particle radiation to study the 
chemistry of the ionization processes for "spectroscopically pure" 
gases near atmospheric pressure. Their system utilized a 75 nm orifice 
to sample ions from these pure gases into a differentially pumped 
magnetic deflection mass spectrometer at pressures less than 10'° torr, 
a similar instrument to that used by Knewstubb and Sudgen. They found 
that the spectra from the gases were dominated by ions from trace 
impurities in the gases, especially from water. Kebarle [9] used these 
findings to study the kinetics of the ionization process and the 
production of water cluster ions, of the form (H20)^H''', formed in the 
post-orifice adiabatic expansion from the atmospheric pressure source 
to the high vacuum of the mass spectrometer. 

At about the same time, Shahin [10] developed a system for the mass 
spectrometric study of corona discharges in air at atmospheric 
pressures. Ion products were sampled using a 15 ;xm aperture into a 
differentially pumped region. This analyzer region (a quadrupole mass 
analyzer) was maintained at about 1 x 10"^ torr. The major positive 
ions that he observed were (H20)nH"'' and (H20)nN0'*" clusters in wet 
nitrogen. He also found that (H20)nH''" could be dissociated to H3O"'" by 
collisional means in the post-orifice region by increasing the pressure 
and accelerating the ions through that region. 



10 

Analytical API/MS 

Two excellent articles [11-12] review in detail the history of 
APT/MS and the currently available API/MS technology. Because API is a 
"soft" ionization technique, the major positive ions formed and 
detected with mass spectrometry are the molecular ((M)"*"), pseudo- 
molecular ((M-i-H)+), and cluster ((Mx)"^, (Mx+H) + , or [M(H20)x+H]"^) ions. 
In the API source chemical and thermal equilibria are established owing 
to the large number of collisions occurring at atmospheric pressure. 
This leads to a nearly 100% ionization efficiency for compounds with 
either high proton or electron affinities [11]. The absolute 
ionization efficiency for a normal EI source is approximately 1 ion for 
every 10^ molecules and a CI source may be 10 or 100 times greater than 
this [11]. Therefore an API source has a great potential as a very 
sensitive ionization source. 

Many researchers have taken advantage of this and used API sources 
in conjunction with gas chromatography (GC) , liquid chromatography 
(LC) , and supercritical fluid chromatography (SFC) . In a series of 
articles [13-15], Horning, et al., reported the development of an 
API/MS system for analysis of GC effluents. In 1973 they demonstrated 
the detection of 20 ng of nicotine, 20 ng of cocaine, 30 ng of 
methadone, and 500 ng of caffeine in 1 fiL of chloroform by GC/API/MS 
using negative ions. Also analyzed were barbiturates, extracted from 
urine, in the 3-5 Mg/niL concentration range. Mitchum et al . [16-17] 
used an Extranuclear API/MS instrument to detect 60 pg of 2,3,7,8- 
tetrachlorodibenzo-p-dioxin (TCDD) and 2 ,4 , 5- trichlorophenoxyacetic 
acid (20 ppm in 3 ^L of whole blood and 30 ppb in urine and feces 
samples) with capillary column GC as the sample introduction method. 



11 

Tsuchiya and Taira [18], Carroll et al. [19], Scott et al . [20], and 
Arpino et al. [21] have all developed LC/API/MS systems and applied 
them to a variety of sample types. 

Because of the design of the sources described above, cleanliness 
of the source and instrument was extremely important. Source regions 
for these instr\iinents were characteristically very small (<1 cm-^) and 
this lead to contamination and memory effects. These sources also 
suffered from orifice clogging and clustering problems and these two 
problems would have been even more evident if these sources had been 
used as direct atmospheric monitors. 

While performing direct atmospheric analyses, the low energy API 
method is likely to produce ions of the same mass- to -charge (m/z) 
values from components of the same molecular weight. In order to 
distinguish two such compounds it is therefore necessary to look at 
fragments of these compounds. However, if fragmentation is caused 
without first isolating the ion of interest, it would be impossible to 
tell which ions were fragments or merely ions formed from other 
compounds. Thus, to perform direct atmospheric analysis with an API 
source, a separation step is needed before mass analysis. 
API /Tandem Mass Spectrometry 

Much work has been performed showing that a mass spectrometer can 
be used as a mass separation technique followed by a second mass 
spectrometer for sample analysis [22]. Presently, only three examples 
have been reported of API sources being installed on tandem mass 
spectrometers, all triple stage quadrupole instruments. Caldecourt, 
Zackett, and Tou [23] constructed their own mass spectrometer and API 
source to analyze vapors from manufacturing processes. Extrel 



12 

Corporation [24] has reported installing its commercially available API 
source on a tandem mass spectrometer but as of this writing that system 
is not commercially available. 

Sciex Inc. [25] has developed a API/MS/MS system with liquid helium 
cryogenic pumping. This API system has seen the most analytical usage 
of any system to date, although conventional mass spectrometric 
ionization techniques cannot be used with this mass spectrometer. This 
instr\iment has been shown by Bruins et al . to be applicable to LC [26- 
27] by means of an "ion spray" interface for the analysis of sulfonated 
azo dyes and to SFC [28] for the analysis of anabolic steroids at the 
20-30 ppb concentration level. Snyder et al . have used the Sciex 
API/MS /MS instrument with a pyrolysis probe to analyze various 
pharmaceuticals in commercial polymer matrices [29]. The mobile Sciex 
TAGA 6000 has been applied to dioxin analysis [30] and for the 
detection of explosives and drugs in airports [31]. 

While the Sciex instrument has seen a wide spectrum of use, most of 
the analyses performed with it can be more easily performed with 
conventional, standardized methods utilizing EI/CI ionization. Its 
true advantage is its API source. However, because of the liquid 
helium cryopumping vacuum system, this instrument cannot be operated 
near any appreciable levels of hydrogen or helium gas , as the vacuum 
system has essentially no pumping speed for those gases. 

One other tandem mass spectrometer system for direct atmospheric 
analysis that needs to be addressed is that built by Glish et al . at 
Oak Ridge National Laboratories [32] . This instrument uses an 
atmospheric sampling ionization (ASI) source in which glow discharge 
ionization occurs at sub -atmospheric pressures (0.1 to 1 torr) , and a 



13 
tandem quadrupole/time-of -flight mass spectrometer for analysis. This 
source has been used to analyze the head space vapor of drinking water 
with 150 ppm trichloroethylene , and the head space vapor over 
trinitrotoluene (saturated head space vapor concentration for 
trinitrotoluene is approximately 6 ppb) . 

Overview of Thesis Organization 

This thesis is divided into 5 chapters. This introductory chapter 
provides background material on atmospheric pressure ionization (API) 
and its marriage with mass spectrometry (MS) to help the reader 
understand the research described in succeeding chapters. 

Chapter 2 describes the design of the API source which has been 
constructed for this work and its interfacing to two different mass 
spectrometers . Problems and characteristics of other early API sources 
are described to -emphasize some of the features of this source. 

The third chapter describes experimental results obtained during 
the development of the API/MS/MS instrument and the use of these 
results to guide the development of the API source to its final form. 
Included in this chapter are descriptions of computerized ion optic 
modeling which was used to help in source development and as an aid in 
gaining an understanding into some of the results that were obtained. 

A description of the analytical performance of the API/MS/MS 
instrximent as a direct atmospheric monitor can be found in Chapter 4. 

The final chapter includes conclusions drawn from this work as to 
the potential for this API/MS/MS instrument to analytically perform 



14 
direct atmospheric analyses and proposes future work that should help 
attain that goal. 



CHAPTER 2 
DESIGN OF THE API SOURCE 



In designing a new ion source of any type for a mass spectrometer, 
the design of the mass spectrometer and its component parts must be 
taken into consideration. The API source for this work was developed 
and modified to work with two different mass spectrometers. 

Overall System Design 

Even though the API source for this work was eventually modified to 
be compatible with a Finnigan-MAT TSQ 70 triple -stage quadrupole tandem 
mass spectrometer, at the inception of this work that instrument was 
itself, still in the design stages. Initial source design work was 
therefore performed on a Finnigan-MAT 4500 single quadrupole mass 
spectrometer with the goal of obtaining a functional API source that 
could later be adapted to the specifications of the mass spectrometer 
at hand. 

Early attempts at developing API sources for performing direct 
atmospheric analyses suffered from many problems, not the least of 
which were severe memory effects from previously analyzed samples , 
clogging of the orifice, and clustering of the sample ions with neutral 
molecules of the bulk gas [11]. Stringent cleaning procedures were 

15 



16 

required to clean these sources, including electropolishing of 
stainless steel surfaces, boiling in deionized water, washing with 
acetone or methanol, and baking at temperatures greater than 350 C. 
Indeed, an API source developed (but never marketed) by Finnigan [1] 
that generated the water cluster spectrum in Chapter 1 (Figure 1.3), 
demonstrated many of these problems. Figure 2.1 is a schematic drawing 
of this source, which is similar in design to many of the early API 
sources. The characteristics of this source include a small ion source 
region (approximately 0.25 cm-^ in volume), a discharge needle as the 
high energy electron supply, adjustable needle-to-orifice distance, 
sample inlet and outlet made out of 1/16" stainless steel tubing, and a 
single orifice interface between the atmospheric and vacuum regions. 
During operation Che source was heated to ca. 200 C, in order to reduce 
memory effects. As can be surmised from viewing Figure 1.3, this 
source had little use as a direct atmospheric monitor because of the 
severe clustering problems. To be fair, this source was developed as a 
GC-detector and not an atmospheric monitor, but even as a GC-detector 
it suffered from memory effects (because of the small source region) . 

The goals of the development aspect of this work were to develop an 
API source that would be able to perform real-time atmospheric 
monitoring, minimize memory and interference effects, minimize the 
amount of clustering, and achieve a high sensitivity for compounds of 
interest. 

Both the Finnigan-MAT [1] 4500 quadrupole and TSQ 70 triple 
quadrupole mass spectrometers utilize differential pumping (two 330 L/s 
turbomolecular pumps) to evacuate the separate ion source and analyzer 



17 




o 

C 

CO 
BO 



A 




_^ 








o 




a. 




o 




f-^ 




(U 




> 




a 




•o 




« 




u 




u 




^ 3 




^ 




-. ^ 








CJ A 




b < 


'_; 


=0 ^ 




rT' Ct! 


0) 


"^ U 


1= 


^ 





lU 


'.u 


s 


U 


•^ 


u 


u^ 


<u 


o 


^ 


^ 


al 


X 




Oi 


w 




J5 


u-> 


n: 






■* 


•" 


- 


a 


c 


^^ 







— > 


a. 


^ 


3 


u 


u 


r" 


■0 


Q 


eg 


W 


2 


0) 


w 


'^ 




a.o 


u 


o 


l-r 


lD 




~3- 


'J 




.^^ 


C 


U 


re 


15 


60 


E 


•^ 


a 


j^ 










o 


•<-t 


v: 


^ 



18 
regions. Because these instruments are pumped differentially, a much 
lower vacuum pumping speed is required (to maintain the analyzer 
regions at a sufficiently low pressure so that the performance of the 
analyzer is not degraded) than for a mass spectrometer which is pumped 
by normal means. These pumping speeds are significantly lower than 
those used by previously developed API/MS/MS instruments. Sciex has 
reported extremely high pumping speeds (as high as 50,000 L/s) [25] for 
its cryogenically pumped instrument. These pumping speeds allow for a 
much larger orifice diameter (>100 /im) than previously developed 
instruments (<20 urn) [33]. Sciex also claims that the large pumping 
speeds allow for a very quick, large drop in the pressure after the 
orifice to avoid the gas dynamics problem of shock waves at the end of 
the supersonic expansion of the gas in this region [34]. However, the 
differential pumping of the Finnigan instruments also allows for larger 
orifice diameters, on the order of 70 fim for this design. And since 
each of the two Finnigan instruments has easily interchangeable ion 
sources and lens assemblies, the potential for developing an 
interchangeable API source for these instruments was good. 

Initial Source Design 

Initial source design efforts began by developing an API source 
that was compatible with a Finnigan-MAT 4500 single quadrupole .nass 
spectrometer while waiting for the new TSQ 70 tandem mass spectrometer. 
In order to make the API source compatible and interchangeable with the 
Finnigan EI/CI ion source assembly, it was built into a standard 6" 



19 

stainless steel conflat flange. The Finnigan EI/CI ion source assembly- 
was mounted onto a similar flange, with associated electrical and gas 
feedthroughs also mounted in this flange. 

In order to reduce the interference and sample memory effects, it 
was necessary to start with a much larger internal volume for the 
ionization region (Figure 2.2(a)) than used by the Finnigan API source. 
To minimize the number of lenses needed for focussing the ions from the 
orifice to the quadrupole entrance, it was desirable to have the 
orifice as close as possible to the pre-quadrupole focussing lenses. 
Because the distance between the face of the flange on the front of the 
instrument and the entrance to the quadrupoles was greater than 6", the 
API source canister was designed to be inserted through this flange, 
held in place by an o-ring. This limited the outside diameter of the 
flange to less than 3.7", which defined the inside diameter to be 
approximately 2.5", a significant increase over the 0.5" diameter 
source region of the old Finnigan API source. 

The next step in designing a source for direct atmospheric 
monitoring was to develop a system to efficiently draw sample air into 
the source, past the discharge region, and then out of the source 
again. This was accomplished by drawing sample air through concentric 
glass tubes (Figure 2.2(b)), past the discharge region, and out a side 
port attached to the source canister, by means of a common laboratory 
blower fan that was modified to draw in sample air (=100 mL/min) 
instead of blowing it out. The inner of the two concentric glass tubes 
can be removed for cleaning when necessary. 



u u 

cs CI 



C U) 





'"i^' 


C5 






/— 


C 






;j 


!-< 






•w 









3 


c 


« 




CO 


.'■^ 









y 


M 




•^ 


>— ' 


f— 




)-i 




3 




c 


-. 


r-i 




iJ 


en 






iH 


Oi 


y 




>3 


r* 


•H 







3 


U 




s 


U 






c 


W 


w 




—4 


w 







i-l 


c: 


Ud 




Cj 


i—i 


^ 


CN 


f" 


aO 


y 




a 




lU 


CM 


w 


o 


1—1 




c 


•^ 


(U 


G 




^ 




1^ 


ij 


iJ 


■a 


3 


CC 


c 


Qi 


QO 




(U 


u 


•*^ 


u 


u 


tc 


U. 


y 




•w 




4) 


c 


U 




v-j 


CJ 


o 




— < 




Ul 



c? 



Uj 


Oi 


c 




:j 


« 


q; 


tn 




u 


:^ 


s 


i-j 


M 


c 


D 




iJ 


C 


jj 


w 


■J) 


q; 


>, 




f— 4 


■<n 


l-H 


;; 




04 


•w 


0) 


< 




y 



C ^ '-M 



c z 



- o 



c 


QO 


^w* 


a 


« 




e 


^ 


•. 


c 




^J 





a; 


0) 


1—1 


y 


N 


Q 


^ 


•^ 


> 


3 


^ 


O 





o 


Q 


w 


•H 



21 




^^y 



\ 




^ ■' . H II I! v^ N I 

k \ ' . -.- 1 T :.> .n\\^ 



// 




L \ 'x \ '\ N, K^-^-^ 



K^ 



^ 






.:-.\\ \\\\\N 



// 




\ 





'/ /a 






t '. 



22 

For a source of high energy electrons , the two most common methods 
in API have been a °^Ni foil /3' emitter and a corona discharge. The 
radioactive foil has the advantages of using no external power and of 
continually emitting a steady flux of electrons. However, the 
electrons it emits have a wide energy distribution (the most probable 
energy being ca. 20 keV) , and this energy distribution has a 
significant tail at high energies [11]. As a result, the spatial 
distribution of reactant ions is not well defined and the path length 
of the electrons and ions vary over a wide range [11]. 

The alternative source of high energy electrons in API is a corona 
discharge from the tip of a metal needle (often a common sewing needle) 
to a planar surface (usually the plate containing the diaphragm) . This 
discharge provides a more well defined spatial distribution of 
electrons as well as a much higher electron flux. One other advantage 
is the ability to select the needle-to-orifice distance, which may be 
used to alter the relative intensity and distribution of ions produced 
in the source. However, it should be pointed out that, in order to 
obtain a true corona discharge, a current -regulated, high voltage (up 
to 10 kV) , current limited (<20 fiA) power supply is required. Later in 
this chapter, the design of such a power supply is discussed (there are 
presently no such supplies commercially available); however, all 
preliminary work was performed using a Bertan 205A-05R voltage- 
regulated power supply which provided to ±5000 V [35]. Because this 
was a voltage-regulated supply, the discharge was unstable, and this 
contributed to not only clogging of the orifice and short needle 



23 

lifetimes (because of sputtering of the needle tip) , but an ion signal 
which was extremely noisy. 

To provide a supply of electrons for this source, a discharge 
needle was mounted in a teflon ring which threaded into the outer glass 
tube (Figure 2.2(c)). By attaching the needle to the tube in this way, 
rough control of the needle position and needle-to-orifice distance 
could be accomplished by rotating the glass tube or by sliding it in or 
out of the back plate of the source canister. 

Finally, Figure 2.2(d) shows a proposed two-orifice and lens system 
to focus the ions to the quadrupoles . Because clustering was such a 
problem in the early sources, and because it is not clear what portion 
of the clustering was occurring in the thermodynamic cooling of the 
post-orifice supersonic jet expansion, it is therefore desirable to 
prevent water and sample molecules from entering into this region. 
Sciex uses a similar system [33,34} and pressurizes the region between 
the two apertures with a relatively inert gas such as carbon dioxide or 
nitrogen to act as an "ion window" which is transparent to ions 
(because of potential fields which pull the ions through this gas) , but 
is restrictive to non- ionized species of the sample gas which might 
clog the orifice or cluster with the ions in the post orifice region. 
In ion mobility /mass spectrometry, it is common to have a backstream of 
a neutral nonreactive gas such as nitrogen or helium to prevent 
particulate matter from reaching the orifice [36], so this idea is not 
entirely unique to the Sciex instrximent. Thus, it was felt that such a 
stream of dry nitrogen gas might serve both purposes for this 
instrument, that is, to keep particulate matter away from the orifice 



24 

and to keep all neutrals, except nitrogen molecules and small amounts 
of impurities from the nitrogen supply, from entering the post-orifice 
region. 

Therefore, to provide this backstream of nitrogen gas, a two- 
orifice system was designed that could be pressurized with nitrogen 
gas. The gas would flow out into the atmospheric region of the source 
and also into the vacuum region of the mass spectrometer. The first 
orifice was a 500 nm aperture drilled into a stainless steel plate. 
This plate served to contain the nitrogen gas and possibly provide some 
focussing for the ions by electrically "floating" it relative to the 
ground. The major interface between the atmospheric pressure source 
region and the vacuxim region of the mass analyzer was a thin, 
replaceable, stainless steel diaphragm with a laser-drilled orifice 
(20-100 fim diameter) obtained from Precision Aperture [37]. 

Two skimmer cone lenses were designed to focus ions and to divert 
excess gas away from the axis of the mass analysis system so that the 
gas could be more easily pumped by the turbomolecular pump in the 
normal ion source region. The region after the orifice and before the 
first conical lens (hereafter referred to as the post-orifice region) 
is at relatively high pressure. Shahin in 1965 [10] noted the ability 
to perform some declustering of the background water clusters by 
accelerating the ions through this region. Fite in 1971 [38] and Levy 
in 1984 [39] noted the effect of multiple collisions in a supersonic 
jet on thermodynamic cooling of molecules. Sciex applied similar 
knowledge to develop an ion lens and gas skimming system to focus ions 
in a supersonic jet expansion with little increase in their kinetic 



25 

energy spread [34]. This can be done by creating a potential 
difference between the diaphragm and the first conical lens. If this 
potential difference is sufficiently large, it can add enough kinetic 
energy to the ion clusters that have formed to cause collisional 
declustering with neutral molecules in the supersonic expansion. This 
ability should help reduce clustering in the mass spectra. 

Potentials for additional lenses not in the Finnigan system were 
provided by a power supply designed and constructed by Mark Hail, a 
fellow research group member. The voltages for these lenses are 
adjustable by means of ten- turn potentiometers from to ±225 V and are 
monitored by a 3-1/2 digit LED display. 

API Source Configuration for Finnigan 4500 

Unfortunately, discharging to the first orifice was not possible 
with that lens floated because the lens power supply described above 
did not provide a low output impedance. When a discharge is created 
between the needle and another surface, the current will seek to follow 
the lowest path of resistance to ground. Thus the power supply for 
this floated orifice would need a very low output impedance. If the 
power supply does not have a low output impedance, a destructive 
situation can occur in which the current passes through the electronic 
circuitry of the mass spectrometer seeking a path to ground, and this 
may result in the damage of circuit components. Since such a power 
supply was not immediately available, the source design was simplified 
(Figure 2.3) to include just one electrically grounded orifice 



26 




0) 

u 

3 
00 



<u 

4J 
0/ 
E 
o 
u 
u 

V 

di 
a. 

en 

s: 

s 



a 

u 



o 
o 

LP 



00 



c 
c 



u 
o 



ID 



Si 
> 

13 



eg 

3 





u 

u 
u 

3 

c 

10 



04 

< 



27 

(circumventing the need for the additional power supply) . In early- 
studies, a cage electrode was inserted in place of the second conical 
lens to mimic the lens system of the old Finnigan API source. To solve 
the problem of orifice clogging and cluster formation a flow of 
nitrogen gas was directed in a stream from a tube placed near the 
orifice parallel to the plane of the diaphragm. This design 
significantly reduced the frequency of orifice clogging when performing 
direct atmospheric monitoring. 

API Source Configuration for Finnigan TSQ 70 

To interface this source to the TSQ 70 system, several 
modifications were made. Figure 2.4 shows a schematic of the current 
API source design. This configuration again incorporates the less 
complicated one-orifice system. However, in this design, the seat for 
the diaphragm has been machined directly into a standard iso-K flange 
which is a standard vacuum flange for the instrument. This flange is 
divided into two sections with the inner section (which contains the 
diaphragm seat) having electrical and vacuum isolation from the outer 
ring. The outer ring clamps directly onto the TSQ 70 vacuum manifold 
and makes interchanging sources relatively easy. The first conical 
lens (CLl in Figure 2.4) is mounted on the flange with the diaphragm 
seat. The second conical lens (CL2) , a lens directly behind it (the 
back lens), a cylindrical lens, and three pre-quadrupole focussing 
lenses are all attached to a grounded mounting plate which replaces the 



28 




CSI 

00 



O 

o 

1) 



31 
t8 



O 
C 



00 



I— I 

> 

G 



U 



O 
»] 

< 



29 

normal EI/CI source mounting plate and associated lenses. These lens 
assemblies can be switched by loosening two alien screws . 

The API source canister and the associated lens system have been 
designed to be totally interchangeable with the related Finnigan 
components. Indeed, the instrximent can be changed over from its normal 
EI/CI operating configuration to an API/MS/MS system, evacuated, and 
running in about 30 minutes. This increases the potential for the 
system to be later developed for possible field work. 

The most recent addition to the source configuration was the 
development of a current -regulated, high voltage, low current, corona 
discharge power supply. This supply is capable of producing "true" 
(stable and invisible) corona discharges up to ±5000 V and with 
currents in the range of 0.5 to 5 /iA. Figure 2.5 is a schematic for 
this power supply with the appropriate components labeled. The system 
is centered around an EMCO [40] 2 to 6 kV adjustable high voltage dc-dc 
converter. The current from the discharge is monitored and used in a 
feedback loop to regulate the input voltage into the dc-dc converter. 
Current and voltage are adjusted with two ten- turn potentiometers. The 
required voltage to generate a particular current depends upon 
different environmental factors such as the composition of the gas near 
the discharge needle, the flow rate of gas past the needle, and the 
needle-to-orifice distance. If the environment changes, the voltage 
changes so as to keep the current constant. However, if the dc-dc 
converter reaches its voltage limit, the system will become 
unregulated. In this case, the circuitry has been designed so that a 



a 

t5 



O 
> 








aa 






c 




u-i 


u 




CNl 


u 




Qi 


s 




Id 


t-l 










^ 






CkO 


2 


. 


•^ 


u 


>^ 


[^ 




1-^ 




3 


c 




0) 


■^ 




c 


M 




<4^ 






c 


0- 




'J 


3 




•^ 


c 




«^ 


CI. 




(7 






= 


a; 










w 


















'J 


r: 




^ 


•7^ 




-^ 


V. 




c^ 


.M 




'J 


" 




•ri^ 






'— 1 


« 




w 






c 


c 




c 


Ui 




f— 


C 




'^ 


U 



31 



<i>- 



V 



10 



3 

a n 



-N^V'S^ 



7/-. 



J S ^ 




72J?^-^2 



O 

a > 

u. a 

o • 

c 



= \ 



T 

A 



V^^' 



-i-f 



-J 



J' 



Ik. 



/ 





X^ 



^■' 



(4>- 



•- L 

C O 

* *- 

L — 

L C 

3 O 



c u 

C ^ 




s$ 



s£ 




Hi' 



32 

large drop in current will occur to prevent an erratic and possibly- 
dangerous discharge from ensuing. 

Figure 2.6 is a schematic of the power supply for electrically 
floating the orifice from to ±500 V based on an EMCO E05 dc-dc 
converter. This is a low output impedance power supply that is 
adjustable by means of a ten- turn potentiometer. 

All parts of this source (except the normal Finnigan pre-quadrupole 
focussing lenses) including the power supplies for the corona discharge 
and the orifice potential were designed in this laboratory and 
constructed in the University of Florida Chemistry Department machine, 
electronics, and glass shops. 



33 



in g 
+1 



(S>- 



4> 
0) 
O +- 

- o 

o 



(S 



c 

Z CO 

c u 

Z 3 

C -i 

CB & 



— Ill 



--J-^h- 




)-l 
3 
00 



a 
a. 

3 
CO 

u 

0) 

o 

o. 

a> 

u 

•M 
•M 
•r* 
U 

o 

> 
o 
o 
in 

-H 

(U 

o 
C 
to 
•a 
<u 
a 



3 
O 



E 

(U 

o 
u 



CO 



CJ 



Cd 



3 

n 



CHAPTER 3 
EXPERIMENTAL RESULTS AND SOURCE DEVELOPMENT 



Results with API Source Conf ig-uration for 
Finnjgan 4500 Mass Spectrometer 



As was stated in Chapter 2, the API source canister was designed to 
slide through a modified 6" conflat flange and into the vacuum region 
of the Finnigan 4500 single quadrupole mass spectrometer. While this 
design allowed the API source to be compatible with the Finnigan 
instrument, it caused several problems. The API source was designed to 
be modular in construction so that design changes could be made to 
selected parts of the assembly without the need to re -machine the 
entire API source for each modification made. Unfortunately, at each 
junction between individual parts of the source canister, there existed 
the potential for gas leaks from the atmospheric pressure source 
region, through the junctions in the assembly, into the vacutom region 
of the mass spectrometer. The source was designed so that the orifice 
plate was electrically isolated from the source canister (in the 
initial design, a second orifice flange was to be inserted between, and 
electrically isolated from, these two flanges) . This concept was to be 
used later in electrically floating this flange. Electrical isolation 
was accomplished by means of placing am anodized aluminum, ring-shaped 
spacer between these two components . Vacuum integrity was maintained 
by two o-rings, one on either side of the aluminum spacer. The 

34 



35 

assembly was to be held together with nylon screws. Because of the 
large force of atmospheric pressure pushing on the orifice flange, the 
nylon screws could not induce a sufficient counter pressure to maintain 
vacuum integrity. The nylon screws were replaced by metal screws to 
obtain vacuum integrity to some degree (although the metal screws did 
not allow for electrical isolation) , and a pressure in the analyzer 
region of 1.2 x 10 '-^ torr was eventually obtained for a 20 nm diameter 
orifice. To calculate the expected pump-down pressure for an aperture 
of this size, fluid dynamics can be applied to treat each stage of 
differential pumping separately. 

The conductance (C^isc) of a 20 /xm diameter orifice at atmospheric 
pressure can be obtained from equation (1) below [41] . 



Cvisc - 76.6 X 50-712 ^ (1-50-258)0.5 ^ ^ ^^ ^^-l (^^ 

1-5 



where 5 =- (P2/P1) < 1 (where ?i and ?2 are the respective pressures on 
either side of the orifice) and A is the surface area of the orifice 
(in cm^) . If P2 « p^, then C^isc approaches the limiting value for 



air in equation 2, 



Cvisc - 20 X A, in Ls'l. (2) 



A 20 fim diameter orifice has a surface area of 



nr 



2 - TT X (1.0 X 10-3 cj„)2 _ 3 1 X 10-6 ^^2 (3) 



36 

Thus, for this system, 



Cvisc = 20 X (3.1 X 10-^ cm2) = 6.3 x lO'^ Ls'^. (4) 



The gas throughput, qpv. can be defined as in Equation (5) 



qpv = Cvisc X (Pi - P2) . in torr-Ls"^ (5) 



Substituting for C^isc. ?! " 760 torr, and making the approximation 
that P]_ - P2 =■ ^1 ' ^he gas throughput is found to be 4.8 x 10 "^ 
torr-Ls'-'-. With q-^^ as defined above and s as the pumping speed of the 
first turbomolecular pump (330 Ls'^), P, the expected pump-down 
pressure for the first region of differential pumping, can be found 
from equation 6. 



./s = (4.8 X 10-2 torr-Ls-^)/330 Ls"^ - 1.5 x 10'^ torr. (6) 



qpv. 



Applying this calculation again, now with ?i =• 1.5 x 10"^ torr, and 
with an aperture radius (for the aperture between the first and second 
vacuum chambers), r, of approximately 0.15 cm, the expected pressure in 
the second region of differential pumping is 6.0 x 10" torr. 
Although, in the calculations above, the nominal pumping speeds for the 
turbomolecular pumps were used and the effective pumping speeds are 
somewhat less (because of restrictions) , the failure to get the 
analyzer pressure below 10*^ torr seems to indicate that the o-ring 
design was inefficient in its ability to maintain vacuum integrity. 



37 

Nevertheless, even though the vacuum integrity of the source was 
poor, these pressures were sufficient to perform mass spectrometry. The 
first ions generated with this source, in a one-orifice (at ground 
potential) system with no nitrogen jet, were cluster ions of the form 
H(H20)n'^ (n = 5 (m/z 91) to n = 21 (m/z 109)), (CH3)2COH(H20)n"^ [n = 3 
(m/z 113) to n = 8 (m/z 203)], [ (CH3) 2COH] 2(H20)n"^ [n = 2 (ra/z 153) to 
n = 4 (ra/z 189)], generated by inducing a flow of laboratory air past 
the discharge region (Figure 3.1). The acetone ions are from residual 
solvent left after source cleaning, which vaporizes off the source 
walls and diffuses to the discharge region because of the lack of a 
nitrogen jet. These ions were of very low intensity, and the lack of 
cluster ions with higher m/z values probably results from a lens tuning 
effect. 

Without the nitrogen gas or an electrically floated orifice, this 
source configuration had only limited abilities for declustering. 
Figure 3.2 shows the resulting mass spectra (all normalized to the same 
intensity) when the potential difference between the grounded orifice 
and the first conical lens was changed from 2 to 8 V. While the effect 
is not dramatic, the slight decrease in the cluster ions with higher 
m/z values relative to those with lower m/z values is apparent. 

However, even when analyzing vapor samples injected into a nitrogen 
stream (no direct atmospheric monitoring) , clustering was still a 
problem. Figure 3.3 (a) - (c) shows mass spectra obtained for 
representative compounds which undergo charge exchange ionization in 
the API source, carbon disulfide (MW 76), benzene (MW 78), and toluene 
(MW 92), respectively. Carbon disulfide forms almost exclusively the 



38 



0) 

o 
u 

O 
« 



o 
r-i 



i 



< 

a 

ID 

4J 



i3^ 

15^ 



CO 



13 
<0 
u 

t3 
U 
0) 

c 

0) 

w 

c 
o 



.= J 



c- 



— es 



< 



to 






O 



u 

(A 



o 



A^TSUa^UJ 3A.T^BXSa 



MI 



39 



(a) 2 V Potential Difference 

271 

253 
217 



199 



91 



127 



163 



T- i L^J 1 



289 



325 



343 



361 



397 



(b) 4 V Potential Difference 



253 



289 325 



4£2. 



115 181 



217 



+■ 



I*' ■* 



361 






CO 

c 



(c) 6 V Potential Difference 

235 271 



199 



163 



127 



91 



■T— 



^JmmmJt^H^mjJaam^^ 



307 



325 



343 



361 



397 



n/z 



(d) 8 V Potential Difference 



235 



199 



181 



127 



91 



163 



i t 



1C3 



■ r ■ 
263 



271 



307 



3S 



361 



1 ,1 397 



3ca 



4ca 



503 



Figure 3.2 

Declustering in post-orifice region with grounded orifice, no declust- 
ering gas, and with the potential of the first conical lens at (a) -2 
V, (b) -4 V, (c) -6 V, and (d) -8 V. 



Figure 3 . 3 

API/MS of compounds that undergo charge exchange ionization in a pure 
nitrogen stream (a) carbon disulfide, (b) benzene, and (c) toluene. 



41 



(a) 

'"•°" Carbon Disulfide 
\ MW = 76 



53.8- 



152 



I ' I ' ] ' I ' I ' 



I ■ 1 







(b) 




ICO.l;-; 

1 


Benzene 




■ 


MW = 78 


CO 


• 




c 






o 






c 


50.0- 




03 






i. 


■ 










<i^ 






CO 






© 






q: 







73 



155 



I 'I ' I 



(c) 

'"'•°^ Toluene 
MW = 92 



50.0 



I ' I ' I 



n/z 



53 



92 



K*- 



J ,1 .|ll, , 1,1 



tea 



153 



«2* 



lC-1 



23 



-I 1— — t 

S3 



42 

142"^ cluster ion. Benzene and toluene form both the M"^ molecular and 
the M2"'' cluster ions. Figure 3.4 (a) - (c) shows mass spectra obtained 
for representative compounds which undergo proton transfer ionization 
in the API source, acetone (MW 58), ethyl acetate (MW 88), and methanol 
(MW 32) , respectively. Acetone and ethyl acetate form both the (M+H)"^ 
pseudo -molecular and the (M2+H)''' cluster ions. The ions formed from 
methanol display severe clustering and are of the form [M+H(H20)n]'^ (n 
= 2 - 5), [M2+H(H20)n]'^ (n = - 2), [M3+H(H20)n]'^ (n = and 1), and 
(M4+H)+. 



Adaptation of the API Source to the 
Finnigan TSQ 70 Mass Spectrometer 



TSO 70 Characteristics 

This mass spectrometer has some unique characteristics, most of 
which have proven beneficial to performing source development on it. 
This instrument is almost totally under microprocessor control. All 
valves, pumps, heaters, and most importantly lenses and quadrupoles can 
be controlled by "firmware" (program code stored on erasable, 
programable read-only memory (EPROM) chips) which accesses 5 single 
board minicomputers located on the instrument. This allows 
sophisticated tuning and optimization to be performed with relative 
ease. The instrument is differentially pumped by two 330 Ls"-'- 
turbomolecular pumps, which can reach a vacuum suitable for mass 
spectrometry in less than 15 minutes (with no additional gas loads on 
the system). The vacuum manifold is shaped as a rectangular box, with 
a thick, glass lid, which presses down on two o-rings (one each for the 



Figure 3.4 

API/MS of compounds that undergo proton transfer ionization in a pure 
nitrogen stream (a) acetone, (b) ethyl acetate, and (c) methanol. 



100.9-1 



50.0- 



(a) 

Acetone 
MW = 58 



59 



44 



1 7 (M2+H) 



I ' I ' I 



CO 

C 



0) 



03 

a: 



(b) 

'•■''•'^ Ef/7// Acefafe 
MW = 88 



50.0- 



(M+H) * 



89 



177 (M2+H) + 



I ' I 



icT.: 



Methanol 
' MW = 32 



5O.0- 



■i7 



n/z 






rM+H(H20)n]* n - 2, 3, 4, 5 © 
[M2+H(H20)n]* n = 0, 1, 2 _^ 

[M3+H(H20)nj* n = 0, 1 <;n> 



69 



■44 — J. 



87 



1 , 115 7.,« 



(M4+H)+ 



129 



I I ■■ I I I I 'III 



53 



ISO 



za 



45 
differentially pumped ion source and analyzer sections) to maintain a 
suitable vacuum. Vacuum flanges on the instrument are standard 6" 
diameter iso-K flanges which clamp onto the manifold using an o-ring 
between the flange and the manifold to maintain vacuum integrity. All 
lenses and quadrupoles are mounted on an optical rail for easy 
maintenance (and in this case easy modification) and are each held in 
place by only one or two screws. The normal EI/CI ion source and lens 
assembly mounts to this optical rail in the first differentially pumped 
chamber of the mass spectrometer, and is removable by loosening two 
alien screws. 
Source Modifications and Ion Optical Modeling 

In order to interface the source canister to the TSQ 70 vacuum 
manifold, a new adapter flange was needed to replace the conflat flange 
compatible with the Finnigan 4500. Since all the vacuvim ports on the 
TSQ 70 vacuum manifold used iso-K type flanges, it was necessary to 
base the adapter flange on this standard. With the freedom to design a 
new flange, it was also decided to alter the design of the source to 
reduce the problems with vacuum integrity. This modified source design 
moves all the junctions between source parts outside of the vacuum 
manifold of the mass spectrometer by machining the diaphragm seat 
directly into the iso-K flange. Thus, the only possible locations for 
a gas leak are the knife-edge seal on the orifice and the o-ring 
between the iso-K flange and the vacuum manifold (which is in the 
normal instrument configuration) . The rest of the canister attaches to 
the adapter flange by four bolts and can be easily removed without 
breaking vacuum. When the orifice gets clogged, a jet of high pressure 



46 

gas (such as that from a can of common laboratory freon, or from 
compressed nitrogen, or air) can be used to clear the blockage. Later, 
the flange was re-designed so that in the final configuration it was 
divided into two concentric sections, the inner section (containing the 
diaphragm seat) being electrically isolated from the outer section. 

To test the vacuum integrity of this source, a diaphragm containing 
a 70 /xm diameter orifice was installed. Using the equations found 
earlier in this chapter, the theoretical pump-down pressures, for the 
first and second differentially pumped- regions of the mass spectrometer 
are calculated to be 1.8 x 10--^ torr and 7.2 x 10"° torr, respectively. 
The actual pump -down pressures were found to be 2.0 x 10"-^ and 3.0 x 
10"^. Therefore it appears that this interface serves to maintain 
vacuum integrity. 

An ion optical modeling program, SIMION [42] , that has been 
modified to run on a PC/AT microcomputer [43], was used to model the 
lens system for the API/MS/MS instrument configuration. This approach 
has proven advantageous in the design and development of the lens 
system for interfacing the API source to the mass spectrometer. Two 
examples of the use of this program are given below, although it should 
be pointed out that this program does not take into consideration the 
multiple inelastic collisions that the ions incur in the relatively 
high pressure post-orifice region. Therefore, the actual lens voltages 
used for maximum ion transmission can be very different from the values 
that give the best transmission in the collision- force model used by 
the SIMION program. Despite this, the model is very useful for the 



47 

physical design of the lenses themselves, rather than predicting the 
actual lens potentials. 

To simplify the system and facilitate testing, the source has been 
used in the one-orifice configuration. One characteristic of the TSQ 
70 mass spectrometer which turned out to be a disadvantage to the 
adaptation of the API source to the mass spectrometer was the 
electrically grounded source mounting plate. Since the API source has 
been installed on the front of the mass spectrometer, this grounded 
plate resides between the API source and the mass analyzer. This plate 
holds the normal EI/CI ion source assembly, including the pre- 
quadrupole lenses. Since the positioning of these pre-quadrupole 
lenses is critical, it was desired to use this same mounting plate with 
a modified lens block which held only the pre-quadrupole lenses and had 
the excess metal removed. Unfortunately, there is only a 3/4" aperture 
in this mounting plate through which the API generated ions are to be 
focussed. 

Figure 3.5 (a) shows the SIMION representation of the lens system 
for the API source interfaced to the TSQ 70 mass spectrometer. Figure 
3.5 (b) shows the equipotential lines on each of the lenses and the 
lens mounting block. Because the TSQ 70 lens mounting plate is 
electrically grounded, relatively low energy positive ions which enter 
even slightly off- axis are repulsed by the associated field. Figure 
3.5 (c) shows the ion trajectories for an ion with a m/z value of 181, 
an initial energy of 0.2 eV, and with initial angles of 0, 10, and 20 
degrees off the center axis. Figure 3.6 (a) shows the addition of a 
cylindrical lens to penetrate the field produced by the grounded source 



Figure 3 . 5 

Ion optic modeling of lens configuration for API source on TSQ 70 mass 
spectrometer showing (a) the representation of this lens system with 
the SIMION program, (b) with potential contours (at values 10% more 
positive than lens voltages), and (c) with representative ion 
trajectories. 



49 



Orifice 

Plate 



(a) 



CL1 
I 



,.11" 

t 

•■I!„ 



Lens 

Mounting 

Plate 

1 



Pre-QuadruDole 
iens Lenses 

BIOCK J 

I 



ib) 




-160 




GRND 



GRND 



(c) 




Figure 3 . 6 

Ion optic modeling of lens configuration for API source on TSQ 70 mass 
spectrometer with the addition of a cylindrical lens showing (a) the 
representation of this lens system with the SIMION program, (b) with 
equipotential contours (for potentials of 10% more positive than the 
lens potential), (c) with representative ion trajectories for -5 V, and 
(d) with representative ion trajectories for -30V. 




51 
Cylindrical . „„^ 
^ Lens ^^"^ 

; Mountir.g 

Plats 



(c) 



IIIIIIIIMMIMHIMIt 



■•••llllllltllllllllll 



-5 
i 



^ 



GRND 



GRND 



id) 



Pre-Ouadrupole 
Lens Lenses 

Block / 




GRND 




52 

mounting plate. Figure 3.6 (b) shows the equipotential contours for a 
cylindrical lens potential of -5 V. This improves the transmission of 
ions with m/z 181, initial ion energy of 0.2 eV , and initial angles of 
0, 10, and 20 degrees off the center axis. With this potential, ions 
<10 degrees off the center axis are focussed through the lens system. 
In Figure 3.6 (d) the potential on the cylindrical has been increased 
to -30 V. By plotting the trajectories for ions with the same initial 
conditions as above, it can be seen that ions with initial angles of 
<20 degrees off the center axis are focussed through the lens system to 
the quadrupoles. 

To aid in the focussing of ions and the pumping of the excess gas, 
a second conical lens (CL2) was added to the design. In order to 
implement this lens as a focussing lens element, a plate lens with a 
cylindrical hole is required directly behind the 2nd conical lens. The 
SIMION program was used to model the effect of the aperture size of a 
lens (the back lens), immediately behind CL2 , on the focussing of ions 
(now with a higher energy (« 5 - 10 eV) after being accelerated through 
the first conical lens) through CL2 and into the cylindrical lens. 
Figure 3.7 (a) is the SIMION representation of a portion of the lens 
assembly consisting of CL2 , and the back, and cylindrical lenses with 
potentials of -5 V, -50, and -150 V, respectively. Figure 3.7 (b)- 
(d) shows the effect of small (0.25" diameter), medium (0.50" 
diameter), and large (0.75" diameter) apertures on ion trajectories for 
ions with initial angles of 25, 20, 15, 10, and 5 degrees off the 
center axis. For a small aperture (b) , ions with initial angles <10 
degrees off axis are focussed through the cylindrical lens. For medium 



n 


s 


c <u 


3 


w 


•H 


•^ 


"O 


j: 


CU 


C W 


s 


o <u 




14-1 X 


CO 


0) -u 




c 


y^ 


<u c c 


o 


r-l CO 


s*^ 


•w £ 




^ U iJ 


. 


O CO 


0) 


CO iJ 0) 


u 


^ c > 


3 


i) -^ 


u 


0) en u 


S-i 


X 0) -H 


lU 


W !-i w 


a. 


Q. 


to 


n-i (u a 




1-1 


1— 1 


<u 


1—4 


V Z !-l 


CO 


N O 


E 


1-1 M E 


CO 


w S 




H-1 3^ 


CO 


0) C/3 O 




u ^ 


/'^v 


r^ 3 (U ^ 


J2 


4J ^ 




[^ i-l iJ w 




0) !-i 


JZ 


OJ D./-V 3 


JJ 


>-< « CO O 


•»^ 


3 w 4J 


3 


60 0) C 




ri £ 60 o 


01 


Ci. -u C O 


0) 


1^ 


•r^ 


14-1 S r-l 


!-l . 


O O CO 


o <u 


^ — H 


-u u 


U 05 J-l 


CJ 3 


CJ C 


3} XJ 


<U Ul 0) 


•—l (-1 


U-l C 4-1 


CO « 


U-l 0) o 


i- a 


<u r^ a. 


4J CO 


0) -^ TJ 


C 0) 


X CO (V 


O 60 


4J y 4J 


iH U 


— H CO 


CO 


•4-1 C -^ 


T3 i-i 


CJ 


c 


o o 


CO CO 


60 ui 




C CO en 


» f^ 


•w CO 


•^ T) 


r-l jr 


01 -w 


0) 60 a) 


c 


•O 3 £ 


0) T3 


iJ 


r-l C 


s ;- 


CO 


X "O 


•o 


O 4J C 


Oi - 


•W to 


u 0) 


U 60 


CO U 


a c 0) 


•H 3 


-H (U 


o -u 


01 en 


O ^-1 


CMC 


01 0) 


CO 0) 


01 a 


M D.rH 


CO CO 



54 









cs 


^ 




« 


C3 




** 


O 








« 




T: 


L^ 




c 


O 






~J 


___ 


:^ 






O 






JC 


(/I 


o 


u 




lO 



03^ 




O 




CO 


































:'" .'1. 




_ ' 


3 




•1 


Ij 1!;' 
|i il','' " 










r 






5 -"::: 


^ — 




____—- 


-:r' — - — 


3 '"'^ 

"•5 


^- 


/i: li 
1 ! 1 

■- i! 'i : ! 




III 






~- li 


I 


- 








'=1 


/ 


-~ 










1(1 


^'^ 












"Vi' ] 


I'l 






1j 


— — ____\!;h ill- 








i 







55 

(c) and large (d) apertures this value improves to 20 and 25 degrees, 
respectively. However, if this back lens is removed completely (not 
shown) , the potential field of the cylindrical lens is not capable to 
cause sufficient convergence of the ions to allow transmission through 
the pre-quadrupole lenses. 

The re -design of the source with a new adapter flange and the use 
of the SIMION program to model the lens system proved advantageous in 
developing an operable API source. By moving all junctions between 
source canister parts outside of the vacuum system, it became practical 
to use a 70 fim diameter orifice. Ion optical modeling saved much 
effort and time developing two new lenses for the system, a cylindrical 
lens to penetrate the potential field generated by the grounded source 
mounting plate and the optimum aperture size for the lens immediately 
following CL2. 
Gas Flows 

Using the equations presented earlier in this chapter, the 
conductance (C^xsc) °^ ^ ^'^ A*ni diameter orifice is calculated to be 
7.7x10'^ Ls"-'- or 46.2 mL-min"-'-. To measure the rate of gas flow 
through the orifice a common laboratory bubble meter was hooked up to 
the exhaust port of the mechanical backing pump for the turbomolecular 
pumps. Since one mechanical pxomp serves as the backing pump for both 
turbomolecular pumps, any gas that enters the system must exit through 
this port. With the orifice (70 fxm diameter) plugged, the flow rate, 
Fq, was « 1 mL-min' ■'•. When the orifice was unplugged, this flow rate 
increased to 45.3 mL-min"-'-. This would indicate a flow rate through 
the orifice of 45.3 mL-min"-'-, which agrees within the experimental 



56 

error (the actual size of the orifice is 70 urn ±10%) for the calculated 
conductance . 

For the measurement of the flow rate through the glass tube used 
for direct atmospheric sampling, the bubble meter was modified such 
that air was drawn through the bubble meter by attaching, with flexible 
tubing, the top of the bubble meter to the glass sampling tube. Both 
the orifice and the sampling fan served to draw air through the 
sampling tube. Because the flow rates through the orifice and through 
the sampling fan are fixed, any additional gas loads would reduce the 
flow through the sampling tube. With no flow of gas for the nitrogen 
jet, the sampling flow rate was found to be 77.4 mL-min"-'-. With the 
nitrogen gas jet on, the flow was found to be 74.8 mL-min' •'■. 
Therefore, the flow rate of the nitrogen jet (flowing out of a 1/16" 
tube) was 2.6 mL-min' •'-. This is the common configuration for the 
instriiment while performing direct atmospheric analysis. 
• Discharge Power Suppl-y 

As was stated in the second chapter, the discharge power supply 
which was used for most of this work was a Bertan 205A-05R voltage- 
regulated power supply. The discharge from which ions were detected 
(with this power supply) , was visible (violet in color) , unstable to 
varying degrees, had a current >20 /xA, and produced an ion signal which 
was extremely noisy. At times this discharge would become very 
unstable and erratic and would change to a bright white color. Since 
the power supply regulated the voltage, any change in resistance in the 
path of the discharge (such as those caused by changes in the gas flow 
rate or composition) would cause a change in the current and thus a 



57 
change in the ion intensity. Tuning was performed by scanning at a 
much slower rate (>1 s/scan) than is normally required for quadrupole 
mass spectrometry and still was, at best, difficult. 

A true corona discharge is invisible, stable, and of much lower 
current than the discharge obtained above [34]. The second chapter 
discusses the power supply that was designed and built to generate such 
a discharge. The corona discharge obtained with this current -regulated 
power supply meets the aforementioned requirements, with a stable, 
selectable current of 0.5 to 4.5 fiA.. The voltage required to produce 
the corona discharge is dependent on the current desired, the flow rate 
and composition of gases in the discharge region, and the needle-to- 
orifice distance, but usually was on the order of 3 to 4.8 kV. With 
this stable discharge, the ion signal was significantly less noisy, and 
therefore, tuning could be performed with automated tuning procedures 
and at a much faster rate. 
Orifice Potential 

With the present design it is possible to have the orifice 
electrically grounded or floated by a ±500 V, low- impedance power 
supply (see Chapter 2) . Initial experiments were performed with the 
orifice placed at instrxoment ground by connecting its electrical lead 
directly to the instrument manifold. Later, to improve sensitivity, 
the orifice was electrically floated at up to -1-150 V. 



58 

Experimental Physical Parameters of API Source and Mass Spectrometer 

Initial Ion Energy and Energy Spread 

An experiment was set up such that the intensity of the ions 
generated by the API source was monitored as the voltage on the second 
conical lens (CL2) was varied, while keeping the orifice at ground 
potential (0.0 V) and the first conical lens (CLl) at -4.4 V relative 
to ground. 

Figure 3.8, a "stopping-potential curve", is a plot of the ion 
intensity (scale on left-hand y axis) vs. the potential of lens CL2 
(with a potential of CLl = -4.4 V), together with the first derivative 
of this curve (intensity scale on right-hand y axis) . The first 
derivative provides (a) the potential that will stop an ion with the 
average kinetic energy (i.e. the peak of the first derivative curve) 
and (b) the kinetic energy spread of the ions (i.e. the peak width at 
half height) . A simple numerical method was employed to find the first 
derivative, in which the difference of consecutive y axis values 
relative to the difference of consecutive x axis values (Ay/Ax) was 
plotted vs. the average of their corresponding pair of x axis values. 
The difference curve that was generated yielded values of -3.8 V and 
0.8 eV for the potential required to stop an ion with the average ion 
kinetic energy and the ion kinetic energy spread, respectively. 

Since the orifice is at 0.0 V (ground potential), the first conical 
lens is at -4.4 V, and positive ions were analyzed, the ions should be 
accelerated by 4.4 eV between those two components. If this were true, 
it should take a potential of 0.0 V on the second conical lens to 







CO 


"w 


r-l 


r* 


<i; 


C 


o 


^c 


•^ 


»f-l 


c 


4-1 




^ 


j^ 


0) 


o 


a; 


•^ 


1-^ 


^ 


u 


M 






C tu 


^ 


c 


1—i 


0) 










r^ 




»^ 


(Z 


^ 


t; 


u 




P 


•^ 


2? 




u 




'^ 


^ 


•-^ 


M 


c 


3 




OJ 


C 


>, 


-- 


_^ 


00 


OJ 


;/) 





a 


q; 






> 


3 


"^ 


U4 




iJ 


-^ 


0) 


— ^ 







3 


1—1 


•^ 


tn 


£1 


"w 


c 


•T-t 


J^ 


(U 


JJ 


t5 


00 


3 


?^ 


f^ ""^ 


u 


qO 


CD 


(D 


^ 


2 •" 


tj 


S 


(M 




<^ <M 


»^ 


3 








.!f S 


■^ 


0) 


r 


■a 


00 


■^ -3 


<u 


c 


^ 


^ 


1-1 


CM 


15 


<u 




• r4 


> 


OJ 





to 


J^ 


O 




;^ 


ai 


0/ 




w 


^ 


!-j 


t3 


^ 









<M 


"^ 


'^ 




o; 


^ 


a 


« 


.■^ 


> 


c 


3 


u 






mJ 


~ 


^ 


O 


•i^ 







Ui 


•r^ 


r-^ 


a; 




ct 


S* 


^ 


• iH 


w 


c; 


iJ 


Ul 




c 




c 


a 


<u 





a , 


j: 


iJ 


o 


u 


n 


a 








•3 





to 


^ 


^ 


c 


<3 




*^ 




T3 


c. 


0) 


« 


o. 


U 


i-l 


o • 


•w < 


-^ 


u 1 


•u 


2 


w ■ 


•r-^ 


c" 




Vj 


<a 


< 





u 



60 



(A/s;uno3) |Di;U3J9|;!a 



o 


o 


o 


o 


o 


o 


o 


o 


ro 


00 


CM 

1 1 1 1 ! 


L.I_U.1-L 




o 
o 
o 
o 

CM 



' ' I 



_o 



CO 



o 
•iri 



c 

a; 

■^-^ 

Lo o 

T 



I I I I I 1 I I I I I I I M I 



I I I I I M I I I I I I I I I I I I I I I I I I I I M I I I I I 

o o o o 

o o o 

o o o 

o o o 

mom 



-O 



o 


o 


o 


o 


o 


o 


o 


o 


»n 


o 


w 


CM 



(s^unoo) /(;isus;u| 



61 
"stop" the ions (i.e. decelerate the ions by 4.4 eV, to 0.0 eV of 
energy). A simplistic model for this system is displayed in Figure 3.9 
(a). The ion, transversing the potential field, can be thought of as a 
ball rolling down a slope. If the ball is released down the slope from 
a resting position, it will be accelerated (i.e. gain in kinetic 
energy) due to gravity by falling some height, h, and therefore it will 
take a height equal to h to decelerate the ball (reduce its kinetic 
energy) until it stops again. The experimental results seem to 
indicate a situation more like Figure 3.9 (b) , in which the ball is 
stopped by a slope that is only a fraction of the initial height; that 
is, it takes a smaller decelerating energy than expected to stop the 
ball. Barring other factors, if the ball were released from the same 
initial height as in Figure 3.9 (a) it should not be stopped on the 
opposite slope but should continue on out of the system (Figure 3.9 
(c)). Likewise, barring other factors, the ion would be expected to 
pass through the decelerating field provided by CL2 instead of being 
stopped. The experimentally determined voltage needed to stop the ions 
was found to be -3.8 V (a deceleration of 0.6 eV from the CLl to CL2 
lenses); therefore, the ions must have been accelerated by only 0.6 eV 
on their way to the first conical lens. This could occur if the ions 
were generated at a potential field 0.6 V more positive than that of 
the first conical lens (i.e., at -3.8 V with respect to ground). This 
is analogous to the ball starting at a very low height. However, it is 
known that the ion is formed on the atmospheric side of the orifice 
(i.e. the ball is starting all the way up the slope). A second 
possible reason is that the ion's ability to increase its kinetic 




eo u 

U Qi 







c 



a = VI 

u- w n 

- c 

- !/! -^ 

-- >^ iJ 

::: ai ^ 

— "O 

X 71 O 

O y y 

c :: 0) 



"C — " -J 

c ^ — < 

= i« 3 

•J c y 

— in 

v) ■^ O 

I— 1 --N en 



(0 




— o 

^ J 

I- a 

1" s. 

o X 

c ^ 

C U 

c 

to ^^ 

u a; u 

M a; >-' 



en 



a> 



63 

energy was being retarded. If the ball was rolling down a surface 
where the friction was large, the ball's acceleration would be slowed 
by the friction; that is, the friction would retard the ball from 
increasing its kinetic energy. In the post-orifice region, a 
supersonic jet expansion occurs as the ions and molecules go from the 
atmospheric side of the orifice to the vacuum side of the orifice. 
Since near the orifice the ions and molecules are still at a relatively 
high pressure (several mtorr) , the bulk gas molecules undergo a 
Ithermodynamic cooling of their translational energy through multiple 
collisions with each other in the supersonic jet expansion [39]. As 
the ions exit the orifice, they are translationally cooled along with 
the bulk gas molecules. The inelastic collisions the ions have with 
the translationally cool molecules also allows the rotational and 
vibrational motion of the ions to be transferred to the translationally 
cool molecules. Thus, the ions do not experience the full accelerating 
potential field because of collisional energy exchange with the more 
abundant neutral molecules which remain unaffected by the potential 
field. Further downstream in the jet, the number of collisions between 
the ions and the molecules decreases and this collisional cooling no 
longer occurs . 

Figure 3.10 is stopping-potential curve (plotted on the left-hand y 
axis and its associated differential (plotted on the right-hand y axis) 
determined for CL2 with an electrically floated orifice with a 
potential of +120 V and with a potential for CLl of +81.9 V. The 
potential of CL2 required to decelerate the average ion to rest orifice 
is approximately 89 V and the energy spread of the ions is 



6^ 



(A/s;unoo) |Di:^usj3^^ia 



o o 

o o 

o o 

o o 

in "^ 



o 
o 
o 
o 



O 
O 
O 
O 
CM 



O 
O 

o 
o 



''""■■"'"""■"'"""■"'"'"""I"'"""'' 



CO 




I I 1 11 I I I I I I I I I I I I M I I I I I I I I I I I I I I I I I I I I I I I I I I I 1 I I I I I I I I I I I I I I ^ 



o 
o 
o 
o 
o 
n 



o 
o 
o 
o 
in 

CM 



o 
o 
o 
o 
o 

CM 



o 
o 
o 
o 
in 



o 
o 

o 
o 
o 



o 
o 
o 
o 
m 



(000l/s:^uno3) X:^isua^u| 



0) > 



O O" 

O.CSJ a) 

^ u w 

00 - c 

c ^ — o 

-4 n .-< 

Cl Qi •- 

C y u C) 

o — c j: 

u c iw 

< o a o 



65 

approximately 10 eV. Because it required only 7.1 V to decelerate the 
average ion to rest, the ion therefore could only have been accelerated 
by 7.1 eV from the orifice to CLl despite the 38.1 V potential 
difference between these components. At these higher voltages, 
however, there is a much greater spread in the ion energies. 
Needle-to-orifice Positioning 

For a discharge provided by a voltage -regulated power supply, the 
position of the needle was critical to the production of even a semi- 
stable discharge and to the number of ions which were detected by the 
mass spectrometer. Figure 3.11 (a) (all dimensions in inches) is a 
schematic representation of the needle position relative to the orifice 
and the orifice plate for which the greatest intensity of ions were 
obtained. Characteristically, the discharge was formed across the face 
of the orifice to the opposing sharp edge of the orifice plate. When 
the needle was moved across the face of the orifice towards the center 
axis of the orifice (Figure 3.11 (b)), the ion intensity decreased 
rapidly to zero. A semi -stable discharge was also obtained by moving 
the needle in towards the orifice (Figure 3.11 (c)); however, no ions 
were detected by the mass spectrometer for this needle-to-orifice 
position. Since the needle is at a large positive potential (= 4000 
V) , positive ions will be repelled from it and accelerated towards 
whatever surface is serving as the counter electrode. When the needle 
is in the position in Figure 3.11 (a), ions that are created in the 
region between the needle tip and the center axis of the orifice can be 
directed through the orifice by the potential field. However, when the 
tip of the needle is towards the center of the orifice, ions will be 



Figure 3.11 

Schematic representation (not to scale, 1/8" =» 1") of needle position 
relative to orifice. Position which generates (a) a discharge across 
face of orifice and from which ions are detected, (b) a discharge 
similar to (a) but shorter in length and for which no ions are 
detected, and (c) a discharge to the orifice. 



67 



(a) y 




(b) / 



8.83 




8.83 



ic) y 




.38 



68 

repelled into the diaphragm or into the source mounting plate and will 
not be directed through the orifice. Likewise, if the needle is moved 
in towards the orifice and discharges to the orifice, the potential 
field of the lens system will not be strong enough to direct ions 
through the orifice because the field generated by the discharge needle 
will be very strong in this region due to the close proximity of the 
needle to the grounded surface. For an electrically floated orifice, 
the position of the needle and the needle-to-orifice position are not 
as critical as in the case above for the production of a stable 
discharge and for the detection of ions. Factors that affect the 
discharge voltage required to produce a selected current include the 
gas flow rate and composition in the discharge region, the potential on 
the orifice, and the needle-to-orifice positions. Because of the 
limited output voltage (5 kV) of the dc-dc converter, the needle-to- 
orifice distance required to maintain a discharge is limited to only a 
few millimeters. However, within this range, this discharge was easily 
obtainable and always stable (unless the needle was shorted to the 
surface of the orifice plate) . 
Effect of Various Parameters on the Ethyl Acetate Ion Family 

While the potential field from electrically floating the orifice to 
a positive potential causes a greater transmission of ions through the 
orifice, it also increases the energy spread of the ions. The ions 
with a very small amount of kinetic energy (<2 eV) can cluster in the 
post-orifice region. Ions with a large amount of kinetic energy (>7- 
10 eV) can undergo fragmentation through collision- induced dissociation 
(CID) with the bulk gas. Ions with an energy between these values can 



69 

be declustered to form molecular (M"*") or pseudo-molecular (M+H"^) ions 
[34]. 

When ethyl acetate (M = CH3CO2CH2CH3) is ionized in the API source, 
ions with m/z values of 177 (M2+H"*') , 125 [M+H(H20)2"*'] , 107 (M+H3O+) , 89 
(M+H"*") , 61 (M+H"*" - C2H4) , and 29 (C2H5"^) are detected. Because of the 
kinetic energy spread, both cluster ions (m/z 177, 125, and 107) and 
fragment ions (m/z 61 and 29) are present with the pseudo -molecular ion 
(m/z 89) . It is desirable to limit the production of both cluster and 
fragment ions to increase the amount of the pseudo-molecular ion that 
is detected. Therefore, the orifice potential and the potential of the 
first conical lens were varied and the intensities of members of the 
"ion family" of ethyl acetate (m/z 177, 107, 89, 61, and 29) were 
plotted vs. these values. Figure 3.12 shows the intensities of these 
ions plotted (m/z 125 and 61 not shown) vs. the orifice potential, 
while holding the potential of CLl at +40 V. At approximately 100 V (a 
60 V potential difference in the supersonic jet expansion region) the 
greatest intensity of m/z 89 relative to the intensities of the other 
ions is achieved. At lower values of the orifice potential (lower 
potential difference between the orifice and CLl) clustering increases, 
and at higher values (higher potential difference) fragmentation 
increases . 

These same ions are plotted vs. the potential of the first conical 
lens, while holding the orifice at 100 V in Figure 3.13. The greatest 
intensity of m/z 89 relative to the intensities of the other ions 
occurred at ca. 36 V (approximately 2 to 1). Below this value the 
intensity drops off rapidly for all the ions. Above this value, all 



70 




(s^unoo) X^isu9^uj 



0) 



3 
=0 



c 


^ 


s 


u 


=0 


u 


c; 


G 


u 


C 




^, 




v^ 










V) 


^ 


i-4 




^ 


M 










■X. 


_, 






^ 


3 


o 


c 


"■ 


u 









<M 






rr, 








a 





CJ 


a; 



= o 

— o. 



a 


VI 




"^ 


> 




•J 






(^ 


— 




o 


- 




V) 






























-^ 


^^ 












'J 

0) 




u. 







** 


= 






c 

■3 




u 


3 


. 


rn 


71 


> 




u. 


C 


c 




^ 


u 


(U 


•*- 


c 


^ 






iJ 


D 


OJ 


-3 


^^ 














:— 


C3 


^w- 



71 



Gi 

00 •» 

N 

£ 



I 1 n I 1 M 

00 
o 
o 

+ 

UJ 

in 



-rrr 



co 

o 
o 
+ 

UJ 

o 




"'"M 1 ' 1 I I M ) 1 I I 1 1 I I r 

CO 

o 
o 
+ 
iij 



oo 

o 
o 

+ 

UJ 

o 



o 
o 

+ 

UJ 

o 



(s^unoo) X;isu9:^u 



to 



= I— I 






i! 3 



f3 

'J u 

O C3 



C != 
G 

f—i o 





« 


0) 


> 


f- 








u 


^v 










<4-l 








•^ 









(w 


o 




O 




> 


U) 







o 


■^ 


o 


-H 


^ 


fs; 


u 


0) 


1— 1 


^^ 


M 


+ 


<n 


a. 




r- 

















u 






— ^ 


•-H 


4-1 






(~l 


S) 


^ 


.^^ 


^ 


c 


'•^ 


r" 


o 






72 
the ions increase steadily in intensity until a potential of about 70 V 
for the conical lens, at which time the intensity of the fragment ion, 
ra/z 29, drops off. The other ions increase in intensity until 
approximately 85 V when the intensities for all the ions begins to 
decrease. It is apparent that, with the floated orifice, both clusters 
and fragments will be observed because of the kinetic energy spread; 
however, this seems to be a worthwhile tradeoff due to the increase in 
intensity that the floated orifice provides. 

The intensities of members of the ion family of ethyl acetate (m/z 
177, 107, 89, 29) were plotted relative to the discharge current 
(Figure 3.14) to determine what effect the current had upon the two 
processes of clustering and fragmentation, as well as on the 
intensities of all the ions. From these data, it appears that the 
highest intensities of the ions and the highest ratio of the pseudo- 
molecular ion intensity to that of the cluster and fragment ions occur 
in two regions, from 0.5 to 1.2 M and from 2 to 3 /jA- The lowest 
discharge current that can be set is 0.5 /xA, because below this value 
it becomes very difficult to regulate the current. Between 1.2 and 2 
mA the intensity of the protonated dimer increases significantly, while 
the intensities of the other ions decrease. Above about 3 ;:iA the 
intensities of all the ions decrease rapidly. Currently, the effect of 
current on the relative intensities of these ions is not yet 
understood. 



73 



=0 
!3 



fl 1 1 1 1 

00 


n 1 1 1 1 
CO 


o 


o 


O 


o 


+ 


+ 


UJ 


UJ 


o 


o 


o 


o 


CO 


in 




oo 
o 

O 
+ 
UJ 

o 
o 

• 



CO 

o 
o 
+ 

UJ 

o 
o 



00 

o 
o 
+ 

UJ 

o 
q 



00 

o 
o 

+ 

UJ 

o 
o 



(s^unoo) /<;isu3;ui 



o 
o 
o 

+ 

UJ 

o 
o 

d 



« 
1-1 
— ' 
CO 



■J) 

u 
a 

Cfl 






=0 

—I -C 

—I u 

S W 



0) ■ 

-> > 

c c 



s ^- 
o ~ 
£ u 
o 
iw .-> 

o o 

c 
Cfl 7 

0) o 



— . 0) 

to to 

- a. 
c 

c j: 



oj "2 



74 
Clustering/Declustering; 

The clustering of ions with molecules in the supersonic expansion 
of the post-orifice region has been one of the major problems in the 
development of an API source for direct atmospheric monitoring. 
Indeed, the only source that successfully overcomes this problem has 
been built by Sciex [25]. This problem can be minimized in one of 
three ways. The first is to start with a sample that contains no water 
vapor, such as the effluent from a GC . Unfortunately, in direct 
atmospheric monitoring removing the water vapor is nearly impossible. 

A second method to reduce clustering is to prevent water molecules 
and other neutral impurities from entering the post-orifice region. 
Sciex accomplishes this by drawing the ions that are formed through a 
pre-orifice region (enclosed by an aperture plate, to which the 
discharge is struck, and the orifice plate) which has been pressurized 
(slightly > 1 atm) with a pure, relatively inert, cryopumpable gas such 
as carbon dioxide (for liquid nitrogen or liquid helium cryogenic 
pumping) or nitrogen (for liquid helium cryogenic pumping). Sciex has 
referred to this region as a "gas curtain" or "ion window" [33]. In 
this instrument, with a less complicated one-orifice system, a gas 
"jet" has been produced by introducing a gas through the aperture of a 
1/16" tube positioned near the orifice. With an electrically grounded 
orifice, this gas jet has an appreciable effect on the amount of 
clustering that occurs. Figure 3.15 shows mass spectra of the water 
cluster ions obtained (with a grounded orifice) for direct atmospheric 
monitoring with (a) the nitrogen on and (b) the nitrogen off. With the 



(a) 



to 

C 



(b) 



r 



75 

(H20)nH^ 
n = 4 - 11, 13 



127 

1 
189 I 



n iii3 



n: 163 181 I '5^ 

r I I ■ =8' 23.:37 1 271.,, 



isa see 2:> 3et 3:> <ea 



T (H20)nH+ 
n = 4 - 22 



3t7 

r 



191 



1(3 

I 



32; 

f 



31 lei i:t 2S( ::i 3ai 3:1 



Figure 3 . 15 

Mass spectra of background water cluster ions generated from direct 
atmospheric monitoring of laboratory air with nitrogen jet (a) on and 
(b) off. 



76 
nitrogen on, the intensities of the large cluster ions are reduced, 
while the intensities of the small cluster ions are increased by ~2 
orders of magnitude relative to their intensities with the nitrogen jet 
off. 

The third method to reduce clustering is to perform collision- 
induced declustering in the post-orifice region. Sciex has performed 
this by applying a potential field in this region to impart enough 
kinetic energy to allow CID of the cluster ions to occur [34]. In 
theinstrument developed in this lab, this same effect can be observed. 
Figure 3.16 shows mass spectra obtained (orifice potential = +100 V and 
a dry compressed air jet flow rate of approximately 1 mL-min"-'-) for (a) 
a potential difference between the orifice and the first conical lens 
of 17 V and (b) a potential difference of 66 V (both mass spectra 
plotted on the same intensity scale) . While the cluster ions with 
large m/z values are reduced, the cluster ion with small m/z values are 
increased (by a factor of =2 for m/z 37 and =7 for m/z 19) . 
Furthermore, with a floated orifice (at a potential difference high 
enough to cause CID declustering), the gas jet appears to have no 
effect. Figure 3.17 (a) and (b) show the mass spectra attained with a 
floated orifice (potential difference =• 50 V) and a nitrogen jet on and 
off, respectively. This same effect is observed for sample cluster 
ions. Figure 3.18 shows that having the gas jet (a) on or (b) off has 
little or no effect on the spectra obtained when ethyl acetate is 
analyzed by direct atmospheric monitoring by this source with a floated 
orifice (potential difference - 50 V) . While with a floated orifice 
the gas jet has little effect on clustering, it may still be worthwhile 



77 



vo 












4 
f 


o 












, 


H 














X 














^ 


1 

II 

•*. 
Q 

•2 

c 

<B 

1 






U1. 




c 

in 

m < 




1 
































.. — r 












C4 


■ 








o 












A 




































IS. 












































" 






















« 


1 


1 










1 



o 



X 

«» 









CO 

1 






«3 
«3 




1 


II 






3 






5 

c 
o 




1 

: 
fv— - 

*»- — 

c- 

in 




m 




U1 










■■-: 






/■ 


■ 






A 






1 1 1 


1 


1 i 





u 


f-M 




>«^ 


C 




;_4 


»4 




o 


u 




^ 


C 




£i. 


u 




tn 


u 







o 




— 


c 




•J 












'w 


C3 




u 


•o 




o 


r^ 




OJ 


3 




u 






•w 


> 






o 




E= 


c 







t-^ 




'*« 


— 




u^ 


^> 




T3 


C3 o 




a 


^ 




u 


,_, 




B 


c;: -~- 




'n 


H-^ ^ 




0) 


jj ^^ 
















a 


« TS 




tiO 


iJ c 
o t^ 




tfl 


■** > 


vO 


c 




^^ 





^ r. 




•iH 


U ,-1 


m 




■*- ^ 


(U 


a) 


t: '^ 






o 


ao 


en 


(M 




u 






iJ c 






o 






^ *^ 







•n ^ 




ij 


>, « 




n 


3 '-' 












a) 



u o 



CO 



oa 


o 


1 


^ 


4J 


u 


u 


c 


U) 


(3 


1-1 


o 


— 





Q. 


(u 


(B 


O 





t-H 


•^ 


Ml 


O 


c 


Li 






f ) 


ao 


3) 


I") 


C 


U 


c 


•»H 


C 


U1 


M 


o 




o 


Ul 




u 


a> 


W 


•w 


H-l 


Ul 


:; 


<u 



o -^ 



Aiisuewi 



— ' O 



VO 










-r 


o 










H 










X 










^ 


O 

1 






I* 

rv 




~5 












C 








■ 














o 










k« 
















^ 
















2 




«^ 














^ ^ 










« 








.. 


r4 




, 
















rt 










«t 














, 














»»* 


























— - 










/^ 










■• 










"^ 




















«l 










ot 








?_2 



VO 






- 




o 










H 










X 










^ 






IS 


•1 




O 

1 




m 


• 




ca 
o 

2 




rs. 
n 

S 


-• 




A 












• 






rs 














^ 








fn 


"« 


\ 





























"te 



c 


c 


/^ 





2a 


•Vi^ 


« 


BO 


o 


<u 


B 


u 


=3 


o 




y 






iJ 


U-l 


U 


■^ 


a 


u 




o 


"3 


u 




w 


^ 


G 


^ 


O. 



U^ 


o 


o 


y 










a 


4) 


60 


u 




01 



ro 




r-t 




U 

0) 

U5 


to uj 


4) 


-.4 IM 


3 


- ° 


aO 


o ^. 




— 


iJ J 


r. , 





— • 

p. 




W 


ca c 




K 


'«< 




^ 


^^ 




a 


w c 




3 


-1 o 

3 




"C 


U :3 




c 


■^ >»^ 




3 


(0 







JJ 




'^ 


>, a 




eo 


u — ( 




^ 







u 


ij trt 




« 


a X 




^ 


u 00 
o 




U-l 


^ 0) 
<5 X 




c 




r-> ^ 




a 


'•^ -3 




'•^ 


-* 




u 


•^ 






00 3 

C 



w 



(0 



A;jsuQ]ui 



u c 
o « 

W -J > 

•Ji z 

coo 
S 5 m 



79 



^ 










[■^ 


o 










. 


H 


















ff> 












Ti 




X 








00 




^ 








fv -—J 


00 




K 








" 




fs. 












" 












S 






rt 






o 

1 








lA 
ri 




1 












a 












-5 






m 






c 








© 




0} 








■•1 




05 












o 












i. 












4m 












ig 




lA 


















M 












« 




ri 




fs 












C 












ri 






"to 












Ti 


•»' 






















-«» 












OB 










ot 












fN 






• fl^BM 












vi 








"« 




t* 
















*t'~" 












C* 


«• — 


■to 


1 


1 1 


1 




1 1 















r«- 


^ 










o 


o 












H 












X 








^L= 




^ 








cr. 










rt 






rs 












rv 












n 










u. 






«-» 


"o 




u, 
O 

1 






u> 


ri 




1 












"5 






~ 






"-5 






* 






C 








ri 




0) 












CD 












o 












^ 




























u> 








^ 


fv 


























■to 










" ■ 












oo 












o 


■to 

00 










r^ 












<A 


















I* 















tfl*""" 












« 






Pto 


1 


1 1 


1 


"T 


1 





tn 

to 





u 




u 




13 




u 




V 




o 




(0 




r-l 




>^ 




f» 








U 




Qi 




^ 










00 




rmmi 


qO 










m 


•^ 




'-I 


OJ 


O 


u 


u 


^ 


•^ 


oO 


c 


•M 


c 


Uw 


£ 



c. 

O 



CO 



o 




a 




u 




•■4 




"0 




u 









U.4 


t— 1 




1^ 


« 


o 


;. 




u 


y-^ 


u 


^ 


aj 


^.M'' 


c 




w 


•c 




f— 






w 


^ 


Ul 




t^ 


^ 


s 






Alisuewi 



80 
to employ the gas jet to prevent orifice clogging and to reduce memory 
effects . 

Experimental Methods for Direct Atmospheric Monitoring 

Mass Calibration of API/MS/MS Instrument 

In order to employ any mass spectrometer for mass analysis, the 
instrument must have a way to accurately assign m/z values to the 
signals it is obtaining. This process is called mass calibration. The 
tuning of a mass spectrometer usually refers to the adjustment of lens 
and quadrupole offset voltages so that the maximum ion intensity is 
obtained. In the case of a quadrupole mass spectrometer, a calibration 
compound which produces ions of known m/z values is analyzed (for EI/CI 
this is typically perflurotributylamine (FC43)). A calibration table 
is then generated which relates the scanning voltage of the quadrupole 
to the correct m/z value. These values can be compared to the voltages 
needed to pass ions, of otherwise unknown mass, from other samples to 
calculate their m/z values. Although mass calibration for this 
instrument does not change while switching between EI and CI , the 
installation of this source requires a new lens system which results in 
the need to re-calibrate and re- tune for API. Since background ions 
are abundant in API/MS , water cluster ions (no API ions were detected 
for FC43) are used to calibrate the instriunent and perform lens tuning, 
after which sample gases may be analyzed. As can be ascertained from 
the spectra shown in this thesis, this method for tuning and mass 



81 

calibration proved efficient at the unit mass resolution of the 
quadrupole mass analyzers . 

Sampling Methodology 

All vapors analyzed by direct atmospheric monitoring were generated 
by drawing air across the caps from bottles of liquid compounds. For 
studies which required a steady concentration of sample for long 
periods of time (such as calibration and tuning) , a few drops of 
compound were placed in the bottle cap and the cap was placed in front 
of the 1" diameter glass sampling tube of the API source. Although the 
vapor pressure of the liquid samples can be estimated, the amount of 
sample introduced is not accurately known, and absolute determination 
of the API/MS/MS instrximental sensitivity or LOD is therefore not 
possible. 



CHAPTER 4 
ANALYTICAL PERFORMANCE 

To evaluate the performance of an analytical instrijinent , several 
factors may be considered; however, not all factors are necessarily 
important for all types instruments. For an instrument which performs 
direct atmospheric analysis, these factors include the ability of the 
instrument to identify the sample it is analyzing, the ability of the 
instrument to detect trace amounts of the sample, the ability of the 
instrument to distinguish between compounds in the same sample , memory 
and interference effects in the analysis, ease of use of the 
instrximent, and the time required for sample analysis (including 
equilibration time between samples). 

Characterization of an API/MS /MS Instrument 

Analyses of Pure Compounds 

With the source configuration for the Finnigan 4500 quadrupole mass 
spectrometer, clustering proved such a large problem that direct 
atmospheric monitoring was nearly impossible. Indeed, even solvent 
vapors injected into a pure nitrogen stream displayed cluster ions as 
the predominant ions detected (Figures 3.3 and 3.4). In the source 
design for the TSQ 70 with a nitrogen jet, the instrument can be tuned 
to significantly minimize the clustering, even while performing direct 
atmospheric analyses. Figure 4.1 (a) - (c) shows mass spectra obtained 

82 



Figure 4 . 1 

Direct atmospheric monitoring/API/MS of compounds that undergo charge 
exchange ionization (a) carbon disulfide, (b) benzene, and (c) toluene. 



84 



■■^(a) 



76 



M^ 



^2 



151 



152 



Carbon Disulfide 



-Nf/b; 



00 
03 



<D 

^ ♦« -1 






78 



M^ 



Benzene 





(c) 


9 


'm^ 




t« - 








Toluene 


et - 










4e — 










2» — 


, , 1 









85 
(with a grounded orifice and a voltage -regulated power supply) while 
performing direct atmospheric monitoring of representative compounds 
which undergo charge exchange ionization in the API source, carbon 
disulfide (MW 76), benzene (MW 78), and toluene (MW 92), respectively. 
These compounds each form almost exclusively the M^ molecular ion, with 
carbon disulfide forming a small amount of the ^2"*" dimer ion. Figure 
4.2 (a) - (c) shows mass spectra obtained (with a grounded orifice) 
while performing direct atmospheric monitoring of representative 
compounds which undergo proton transfer ionization in the API source, 
acetone (MW 58), ethyl acetate (MW 88), and methanol (MW 32), 
respectively. Acetone and ethyl acetate both form almost exclusively 
the (M+H)"*" pseudo-molecular ion. The most predominant ion for methanol 
is the (M2+H)"'' protonated dimer ion. While the signal obtained under 
these conditions for these samples was not sufficient to perform tandem 
mass spectrometry (MS/MS) , these results demonstrate that this API 
source can reduce clustering significantly. 

Figure 4.3 shows the reconstructed ion current (RIC) traces for the 
two most predominant background ions (when the instrument is tuned for 
large amounts of declustering and analysis is performed by bringing 
caps of solvent bottles near sampling tube of API source) (a) (H20)2H''' 
(m/z 37) and (b) (H20)3H"*" (m/z 55), two representative sample ions (c) 
CgHg"'' and (d) (CH3C02C2H5)ir'", and (e) all the ions. In normal EI mass 
spectrometry, sample ionization occurs by the bombardment of the sample 
molecules with electrons from a filament. In positive CI mass 
spectrometry, ionization occurs through charge exchange or proton 
transfer with ions created from a reagent gas. Usually, however, these 



Figure 4.2 

Direct atmospheric monitoring/ API/MS of compounds that undergo proton 
transfer ionization (a) acetone, (b) ethyl acetate, and (c) methanol. 



87 



■■^(a) 



59 



MH^ 



Acetone 



is 

'to 

a: 



'^(b) 



MH^ 



89 



Ethyl Acetate 



I I 



■■-.(c) 



65 



m^h" 



Methanol 



+ 


to 










ro 


o 


^^ 


•^ 


O 




CM 


0) 


s 


X 


"*• 


u 


--N 


r^ 


^ 


1-^ 


s./ 


to 


T3 


.^ 


C 


lU 


to 


N--* 


+ 


T3 




r« 






c^ 


s 


y^ 




o 


. 


o^ 



CO 


Csl 


*^ 


CJ 




CN 


tn 


o 


c 


CJ 


o 


n 


•w 


^ 




u 


PO 73 


^-• 


C 




-J 3 


y-V 





T3 


lU u 


■s^* 


U 00 




3 ^ 


T3 


tiO u 


C 


1-1 to 


to 


* JD 






+ 


u 


vO 


c 


X 


to 


vO 


c 


U 


•^ 




s 


y^ 


o 





•o 


N-^ 


(U 




lu 


« 


a 


c 







u 


f^ 


to 







(U 


E 


r-^ 




Q. 





s 


3 


5 


iJ 


tn 


Oi 


tu 


^ 


> 


jj 


•^ 




u 


i-i 


to 


o 


jj 


iw 


c 




<u 


0) 


w 


<u 


0) 


a 


h 


to 


a. 


u 


0) 


4J 


)-i 


o 


o 


1— ( 


3 


Oi 


4J 



89 



•C_ 



c: 

CD 
O 
CO 



N 

C 

O 

>~ 

o 

CO 

03 



m 

N 

s 

O 

o 

CO 

CQ 



00 
N 



c 

N 

c 

0) 
CQ 



OS 

N 

i 

o 
"5: 



5 



CO 



(U 



AllSUQlUl 



90 
reagent ions are low in m/z value and the mass spectrometer can be set 
up to pass only ions of a greater mass . Ionization in API mass 
spectrometry occurs in a similar fashion to CI mass spectrometry (see 
Chapter 1). In direct atmospheric monitoring mass spectrometry, air 
ser-ves as the reagent gas, generating N2"'', H2O''", and H3O"'" ions 
(equations 1.2 - 1.5). Sample molecules (M) can react with the reagent 
ions through charge exchange or proton transfer reactions to form M"*" 
molecular or M+H"*" pseudo-molecular ions. The kinetics of the 
production of ions in discharges at atmospheric pressures has been 
studied by Shahin [10] and Kebarle [44]. Shahin stated that for direct 
atmospheric sampling (or for any samples with even a small quantity of 
water vapor) the production of H2O"'" (by charge exchange with the 
initial N2'*" ions) and 1130"*" ions would be the dominant reactions. In 
the case of either charge or proton transfer ionization reactions, the 
intensity of the reagent ions is reduced as the sample is ionized. 
Samples that undergo proton transfer ionization may totally deplete the 
population of the background ions, while samples that undergo charge 
transfer ionization reduce but do not deplete that population. Figure 
4.4 shows the relationship between the intensity of the reagent ions 
with respect to the time allowed for reaction. Equations 4.1 - 4.3 
[44] give the rate constants for the production of H20"*', H30"'', and 
H(H20)n'^. 

N2"^ + H2O - N2 + H20"*" k - 1.9 X 10^^ cm^ -molecule^^ • s"^ (4.1) 
H2O+ + H2O - H3O+ + HO k = 1.8 X 10^^ cm^-molecule^^-s^^ (4.2) 
H3O+ + nH20 - H(H20)n+i'*' k = lO'^^ cm^ -molecule'^ • s"^ (4.3) 



9> 

U 

3 
00 



o 

E 
t-i 

u 



c 
o 



0) OJ 

i-l iJ 

3 m 

w 5 

w 

(U u 

U Oi ^ 

ax i-H 

•u 3 

u o 

1-1 M-i 0) 
)>4 O i-i 

a. u 

o u 

£ 0) 
CO --^ O 

CQ 0) 

m iJ u 

(U (0 

&0 bO )-i 

^ C (U 

« -H C 

j= N a; 

O -w 00 

W 1— I 

i-( .w o 
"O iJ jj 
3 

•O 
E lU 
eg u 

bO E 

O -H 

s o. 
•^ >, 

M h-l 3 
C CO 4J 
O < OT 



c 
o 
u 
o 
u 



O m 

E • 

C O <}■ 

o u 



3 « • 

"D p^ <r 

O 3 
)-l U U) 

a 0} 
u 

u 

(4-1 

o j: u 

« -w c 

U 3 -^ 



C 

o 

u 
to 

3 
D" 



4j -a 

0) 0) 

C iJ 
-^ to 

c 



0) 



(U 0) 



H o 1-1 y 



92 




cn 



o.i 



fTTTT- 










o 
q 


o 
oo 


o o 

CO "^ 


o 


o 
o 


^ 


d 


6 6 


d 


d 



93 

sample molecules with H20'*" and H30''' are on the order of 3 x 10'^ 
cm^ •molecule'^ • s'-'- [45]. Proton transfer reactions with sample 
molecules will thus consume H3O'*" before it can further react to form 
(H20)2H'^ (m/z 37) and (H20)3H''" (m/z 55). Charge transfer reactions 
with sample molecules cannot consume all the H2O before some of it can 
be converted to H30''' and therefore, the population of (H20)2H"'" (m/z 37) 
and (H20)3H'*' (m/z 55) will be reduced but not depleted. 
Memory effects 

The RIC traces for the benzene molecular and ethyl acetate pseudo- 
molecular ions (Figure 4-. 4) show that, as the sample is brought to the 
sampling tube, the signal immediately increases and as it is removed, 
the signal decreases rapidly to zero. This would indicate that memory 
effects have been minimized with this source; however, it should be 
pointed out that this source has yet to be applied to trace analysis, 
and this application would give a much better indication of memory 
effects . 
Orifice Clogging 

Orifice clogging proved to be a severe problem with both the 
Finnigan developed API source and the one developed in this work for 
the Finnigan 4500 quadrupole mass spectrometer. The major reasons for 
this problem was that there was no way to prevent particulate matter in 
the sample gas stream from reaching the orifice. With the addition of 
the gas jet near the orifice, this problem has been almost eliminated. 
When the orifice does become clogged (maybe once or twice a day) , a gas 
jet of relative high pressure (= 20 - 40 psi) may be directed at the 
orifice to dislodge any matter. Occasionally, when all else has 



94 
failed, the instrument has needed to be vented to atmospheric pressure 
and the orifice removed and cleaned with an appropriate solvent 
(methylene chloride) . The diaphragms are surprisingly rugged and a 
single diaphragm has been used for several months of source development 
and sample analysis while the API source has been installed on the TSQ 
70. 

Analytical Potential 

Compound Identification 

An instrument must be able to distinguish between different 
compounds to be of analytical use. This may be done by separating 
compounds in time and using a universal detector, by analyzing all 
compounds at the same time with a detector that provides a different 
response for each compound, or by some combination of both of these 
techniques. A single stage of mass analysis can be used with low 
energy ionization techniques (such as CI, fast atom bombardment, or 
API) to determine the molecular weight of a compound, or with high 
energy ionization techniques (such as EI) to provide fragment ions 
which can be used for structure elucidation. Therefore, mass 
spectrometers are commonly used with some type of pre -separation step 
before mass analysis, such as GC , LC, or MS. When API/mass 
spectrometry is used for direct atmospheric monitoring, by definition 
there is no opportunity for any pre- ionization separation. 

Methyl benzoate (MW 136) and methoxy benzaldehyde (MW 136) are 
examples of compounds with the same empirical formula and the same 



95 

molecular weight that form ions with the same m/z value. Both of these 
compounds produce an intense M+H"^ pseudo-molecular ion (m/z 137) , as 
well as some less intense fragmentation ions when ionized by this API 
source with the potential difference of the supersonic jet expansion 
region set at = 50 V (Figure 4.5 (a) and (b)). In a normal API/MS 
system, it would be difficult to determine the presence of one or both 
of these compounds and distinguish between them (or other compounds 
that have MW 136 and form ions through proton transfer reactions). 
However, because the TSQ 70 is an MS/MS instrument, the pseudo- 
molecular ion can be fragmented in the center quadrupole to obtain a 
daughter spectr\im (spectrxim of fragment ions from a parent ion) of the 
m/z 137 ion. Figure 4.6 (a) and (b) show the daughter spectra (each 
under the same conditions of collision energy - 15 eV and pressure of 
nitrogen collision gas — 2 torr) for methyl benzoate and methoxy 
benzaldehyde , respectively. Major daughter ions observed for the API- 
generated M+H"*" ion from methyl benzoate are m/z 59 [ (M+H) -CgHg ]"*", 77 
[(M+H)-CH3C00H]"^, 91 [ (M+H) -HCOOH]"^, 93 [ (M+H) -CH3CH0]"^, 105 [ (M+H) - 
CH30H]'^, and 109 [ (M+H) -CO]"^. Major ions observed for the M+H"*" ion 
from methoxy benzaldehyde are m/z 77 [ (M+H) -CO-CH3OH]''", 94 [(M+H)-CO- 
^3]"*", and 109 [ (M+H) -CO]"'". There is a clear difference in these 
daughter spectra, and thus the compounds can be distinguished. 

A concern about API has been its ability to analyze mixtures of 
samples. Figure 4.7 shows a spectrum of a solvent mixture composed of 
benzene (MW 78), ethyl acetate (MW 88), benzaldehyde (MW 106), and 
methoxy benzaldehyde (MW 136) , obtained when the API source was used 
with a floated orifice (potential difference of supersonic jet 



c 
c 



0) 

u 
c 
c 

N 



4-1 

0) 



0) 
U 
3 



1.1 
O 
a. 

on 



< 

60 



c 

E 



w 


_« 





-> 


s 


^ 


4J 




CO 


"g 










4-1 


1— 1 


o 


c 


Q 


N 


U 


^ 


— H 


o 


c 


^ 



97 



a: 
5 " 



IS 

o 

N 
C 
<D 

"S, 

"5 



CO 
N 

C 

o 

QQ 

as 

X 

O 
« 



I r 



Ajisuejui BAijBiey 



o 



_:: CM 

■a u 
= 1-1 

Q O 
u 
o 

iJ CM 

c 

O II 
N 

c u 

a! 

t-i M 
>^ <U 

0) 

£ c 
o 

C5 05 








>? 


u 




tS 


a; 


!-i rs 


u 


U C 


3 


U t3 


CO 


a; 


•^ 


a- > 


u. 


W lU 




m o 




S ^ 




\ 




w II 




s 




\ >, 




r .'^ 








< a 




■\ = 




M u 




c 




H-t S 




>- O 




O •-( 




U Bl 




•»H •(-) 




Z ^H 




C — i 




= 









y 












u —1 




o •-* 




-= 3 




a 




w o 




-3 




= :^ 




U J^ 




c o 




•o 




4-1 1— 1 




O C3 




a N 




w c 




—1 S) 




G .2 



99 



5 






to 

o 

N 

c 
o 
00 

o 
CO 



























3: 








+ rs 






" 






3 










^ 








04 








rt 


















c* 








rt 






1 










"■"■ 








♦ 








A 


















(\ — 








rs. 










" 


• 
10 


(D 








-o 








X 




^ 




j: 








o 








2 
















"S 








N 

c 






« 


o 






■♦ 


QQ 




^ 




^ 








X 








o 








c 
















"5 








^ 






'o 


.ii^ 






■X 


•Q 








"*" 




^ 




1 


1 1 


r 11 





Ajjsuewi SAijsiay 



C X 

_ o 

o ~ 

« S) 

C 5 



o 





X ^— 




.^ 




5 2 












w N_^ 




1— 








o 




> -: 




^ >^ 












M O 












re --I 


r-- 


ra 




re N 


<■ 


ij ^ 




iJ c 


0) 


u ^ 


;-i 


o 


2 


a. - 


oO 


U) -^ 


■*-» 


03 


U. 


C/5 00 




S 




\ 3 




^ S 




C- w 




< 




\ o 




M ^ 




- c 




— ^ *^. 




Ij CJ 




c c 




— re 




>«N 




C 1— ' 




C >^ 

c: r- 



■^ 


cc 


">-• 


M 


r**,* 




;;: 




_« 




-> 






-.^ 




2 


^ 


>> 


w 


N.-' 


^ 


CC 




o 




<D 


"3 


■u 


r- 


,-^ 


c 


a; 


re 


a 


^3 


M 


u 


^ 


C 


•w 


0) 


Oi 


Q 


^ 


/^ 



101 



s 



s 
• If) 

CM 



i. Pll 





as 




■o 


iN 


iK 


X 


-C 


o 


<t> 


■c 


T) 






<D 


05 


S 


N 



-Q 






in 
CM 



CD h§ 



■t- O 00 ■ ^ — «- 

CO 0} 

N 

c 

qj 

•Q 



as ■ 



tr) 






I \ 1 1 r 

^3 C0 ^^ ^^ 1^ 

IS CO '.O -V CM 



Aifsuaiui dAiiBidy 



102 
expansion region = 50 V) and tuned for a high degree of declustering. 
The pseudo -molecular ion (m/z 89) from eChyl acetate clearly dominates 
the spectrum and a small quantity of the protonated dimer (m/z 177) is 
present. Fragment ions of ethyl acetate are also observed at m/z 29 
(C2H5'*') and m/z 61 (CH3COOH2''") . However, ions from benzene (m/z 78) 
and methoxy benzaldehyde (m/z 137) are clearly present. There is some 
confusion as to the identities of m/z 107 and m/z 125. Because of the 
large intensity of ethyl acetate it is possible that these ions are 
cluster ions of ethyl acetate (M+(H20)H'*' and M+(H20)2''") or that they 
are related to benzaldehyde (M+H"*" and M+(H20)H"'") . Unfortunately, the 
low intensity of these ions made it difficult to obtain their daughter 
spectra, and therefore, this uncertainty was left unresolved. 
Sensitivity 

The sensitivity of this API/MS/MS instrument can be roughly 
estimated from the daughter spectrum of methoxy benzaldehyde, which has 
a vapor pressure at room temperature and atmospheric pressure 
conditions of = 0.2 torr [46]. The concentration (mol/mol) of methoxy 
benzaldehyde in air above a bottle cap can be approximated to be , 

0.2 torr/760 torr = 0.0003. (4.3) 

Neglecting the slightly reduced pressure near the entrance to the glass 
tube (this occurs because air is being drawn through the tube) and the 
dilution of the concentration with the air flowing through the system, 
the API source has detected and positively identified by a full -scan 
daughter spectrum methoxy benzaldehyde at a concentration of 300 parts 



103 
per million (ppm) in room air. Given an air sampling rate of 100 
mL/min and an MS/MS acquisition time for the daughter spectrum of 5 s, 
this corresponds to 15 ng of methoxy benzaldehyde . Clearly, the 
resultant daughter spectrum has a high enough signal-to-noise ratio 
that one or two orders of magnitude less of the compound could still be 
identified. Furthermore, selected reaction monitoring of one or more 
parent ion-daughter ion pairs, would provide even lower detection 
limits . 



CHAPTER 5 
CONCLUSIONS AND FUTURE WORK 

Summary of Results 



The original purpose of this work was to design and develop a new 
atmospheric pressure ionization (API) source for a turbomolecularly- 
piomped triple stage quadrupole (TSQ) tandem mass spectrometer (MS/MS). 
This API source could be used to study API and direct atmospheric 
monitoring by mass spectrometry. To accomplish this, the work 
initially evaluated a commercially developed [1], but never marketed 
API source, which was developed as a GC-detector. This API source was 
characterized on a Finnigan 4500 single quadrupole mass spectrometer. 
Even as a GC-detector, the Finnigan- developed source suffered from 
memory effects (because of a small source region) and severe clustering 
of ions with molecules in the supersonic jet expansion of the post- 
orifice region. Because of these effects, direct atmospheric 
monitoring with this source was essentially impossible. 

Therefore, a new API source was designed, which would be compatible 
with a commercial turbomolecularly- pumped mass spectrometer, and which 
would reduce or eliminate the problems of clustering and memory 
effects. This source was at first developed to be compatible and 
interchangeable with the Finnigan 4500 mass spectrometer used to 

104 



105 
evaluate the Finnigan-developed API source; however, because of its 
design, this API source suffered from vacuum integrity problems. Even 
with these problems, ions were obtained with reduced, but not 
eliminated, clustering. Because of this clustering effect, this source 
was not viable for use as a direct atmospheric monitor. 

This API source was redesigned to be compatible and interchangeable 
with a state-of-the-art triple stage quadrupole (TSQ) tandem mass 
spectrometer (MS/MS), the Finnigan TSQ 70. The modifications included 
moving all junctions between parts of the source canister outside of 
the vacuum chamber of the mass spectrometer, thus reducing some of the 
vacuum integrity problems. A new interfacing flange was designed to 
mate the source with the mass spectrometer. This flange contained the 
diaphragm seat, which now became the only interface between the 
atmospheric pressure source region and the vacuum of the mass 
spectrometer, and allowed the orifice to be electrically floated. 
Replaceable diaphragms were obtained with laser-drilled orifices, and a 
diaphragm with a 70 nm diameter orifice was installed in the interface 
place. This interface flange maintained vacuum integrity to near 
theoretical levels calculated from fluid- flow theory. 

Unfortunately, moving all source components outside of the vacuum 
chamber resulted in a greater distance between the orifice and the 
first quadrupole. An ion optical modeling program (SIMION) was 
employed to model the region between the orifice and the first 
quadrupole. Although the model, which the program utilized, did not 
take into consideration the inelastic collisions of ions in the 
supersonic jet expansion in the post-orifice region, it proved 



106 
invaluable in the actual design of two lenses. These lenses increased 
the fraction of ions which were transmitted to the quadrupoles. 

This new version of the API source employed a small tube near the 
orifice to provide a gas jet in the discharge region. This gas jet 
served to keep particulate matter away from the orifice (to prevent 
orifice clogging) and reduced the amount of water vapor which entered 
the post-orifice region (to reduce clustering). 

A new current -regulated power supply, which was adjustable between 
0.5 and 4.5 /lA with a discharge voltage between 3 and 4.8 kV, was 
developed that generated a stable corona discharge. Along with this 
power supply, a ±500 V low output impedance power supply was developed 
to electrically float the orifice. The ability to electrically float 
the orifice allowed for an increased sensitivity and a strong 
declustering capability by creating a potential difference between the 
orifice and the first conical lens. The combination of these effects 
provided the ability to minimize clustering with or without the gas 
jet. The gas jet still served, however, to prevent orifice clogging. 

Direct atmospheric monitoring was accomplished for various common 
laboratory solvents by sampling vapors from the caps of the solvent 
bottles. The orifice and lens potentials could be tuned such that the 
molecular or pseudo -molecular ion was the predominant ion in the mass 
spectrtun for each solvent (methanol being the only exception) . 

Evaluation as Direct Atmospheric Analyzer 

The API/MS/MS instrument was applied to differentiating a pair of 



107 

solvents with the same molecular weight which generated pseudo- 
molecular ions of the same m/z value. MS /MS was used to fragment each 
of the pseudo -molecular ions and generate daughter spectra which were 
easily distinguishable. 

Electrical floating of the orifice caused a larger spread in the 
kinetic energy of the ions generated. Clustering and fragmentation 
occurred simultaneously, which in true mixture analysis would make the 
interpretation of mass spectra more difficult. While the API source 
was shown to detect compounds directly in the atmosphere at the ppm 
concentration range (=15 ^g of methoxy benzaldehyde) , the potential of 
the sensitivity of the API source remains to be evaluated. 

Suggestions for Future Work 

Consideration of Supersonic Expansion Theory 

The large body of work which has been performed to study the 
effects of a supersonic jet expansion of gases for the production of 
molecular beams has been reviewed by Bossel [47]. Most of the 
associated theory has been derived for systems with an "effusion 
source". An effusion source is one in which the diameter of the 
orifice is less than or equal to one mean- free -path of the species on 
the high pressure side of the orifice. This assures that no more than 
one molecule passes through the or.rfice at a given time. The mean- 
free-path for nitrogen (the major constituent of air) is 6.56 x 10 "2 ^m 
[48] or 656 nm. Clearly, an orifice of this size would reduce 
sensitivity and suffer from frequent clogging. However, this theory 



108 
can be used to calculate the velocity which the ions and molecules 
obtain in the supersonic jet expansion when they reach the plane of the 
aperture of the first conical lens. Equation 5.1 [38] can be used to 
calculate Mach number (the velocity of the molecule or ion relative to 
the speed of sound) at that point in the supersonic jet expansion. 



M = 



r+1 
r-1 



(r+l)/4 



z 
d 



(r-1) 



(5.1) 



For nitrogen r (Cp/C.^) - 1.4 [38], and for this system the diameter of 
the orifice, d, is 0.00276" (70 ^m) and the orifice- to-first conical 
lens (CLl) distance, z, is 0.075" (1900 /im) . Using these values, M can 
be calculated to be 11.0. The ions and molecules are therefore moving 
at 8360 mph (3740 m/s) . For a molecule of mass 100 g/mole , accelerated 
to 8360 mph, with a charge of 1, the kinetic energy can be calculated 
from equation 5.2. 



K.E. 



0.5-m-v2 



(5.2) 



Applying the values above (m - 1.66 x 10'^^ kg/molecule and v =- 3 . 40 x 
10 ■'^ ni/s), the kinetic energy can be calculated to be 7.2 eV. As the 
gas expands from the orifice, fluid dynamics predicts that the 
thickness of the beam of gas will be equal to one crifice diameter for 
a distance of approximately one orifice diameter [38]. To efficiently 
sample the beam of gas, CLl should then have the location of its 
projected apex within one orifice diameter of the orifice. Figure 5.1 



109 
shows the relationship between the Mach number M, the orifice- to-CLl 
distance 2, and the required diameter of the aperture on CLl D (see 
below). Sciex has stated [34] that if z is less than 50D then a 
potential field may be applied between the lens and the orifice to 
focus the ions while increasing the kinetic energy spread of the ions 
by less than 2 eV (for an electrically grounded orifice). Figure 3.8 
shows a kinetic energy spread of less than 1 eV with this system for z 
= 27D. However, for a floated orifice, the kinetic energy spread is 
much greater (Figure 3.10 shows a value of = 10 eV) . Therefore, by 
applying the knowledge of supersonic jet expansion theory, it may be 
possible to reduce this energy spread by reducing the size of the 
conical lens aperture and reducing z. 

Bossel also reviews the need for finely machined conical lenses 
[47] . The front of the cone should come to as sharp an edge as 
possible. This reduces the production of a shock wave at the plane of 
cone aperture, termed a "mach disk", which can reduce the transmission 
of ions through the cone. The conical lenses used in this work lack 
this precise machining. Therefore, new conical lenses should be 
obtained. Several companies specialize in the production of molecular 
beam skimmers for supersonic expansion, two of which are Beam Dynamics 
[49] and Custom Service Technologies [50]. 
Requirements for Future Work 

For the future development of this source, several componants 
should be modified or replaced. Sciex has shown [34] that varying the 
needle-to-orifice distance can affect the relative intensities of the 
ions produced. This may be useful for providing additional selectivity 





ID 


(U 




>-i 


y 




3 


•^ 


2 


;j 


M 




U 


•-H 


4-1 


<U 


u 




r> 


. o 


-3 


3" 




2 




!-i 


C3 


<u 


<u 




^^ 


u 


(U 


u 


CU 


a 






••-1 


(J-l 


5 


U.I 





•^ 


•^ 




■0 


u 


u 




o 


<u 


= 




u 


:i. 


2 


OJ 








O 


^ 


■ J 


'■^ 


n 


•3 


fl 


a 






<u 


2 


ap 


3 










u 


u 


•«-4 


OJ 




U 


JZ 


-3 


•W 




r- 


X 


0) 


5 


« 


'J 






/— 


• 


V5 


r-l 3 


M 


c 


iJ 




o 


J-1 71 


» 


•-4 


•w 


<n 




dj -0 


^ 


(U 


u 


<u 


^ 


3 tu 


r-^ 


4J 


qO j:; 






'W 4-1 


r-^ 


IM 


^ 


fl 


o 


- 


u 




2 


■M 


00 












w 


- 





—4 


u 


u 


i-H 


4) 




a. 


"^ 


u 


s 


S 


en 


3 


3 


u 


71 


c 


•^4 






U-l 


iJ 


r- 




c 


7; 


(U 


u 


a 


^ 


»^ 


s 


u 


o 

■f-t 


GJ 


U-l 


•J-t 


^ ^ 


c 


'J-l 


iJ 


CU 


0) 


c 


u 


3 


OJ 


3 


Q 


a 


u 


f-H 


3 


'-J 


f— 1 


u 


o 


ca 


a 


a. 




-^ 


03 


o 

4J 


w 


« 




"^ 


^ 


-0 


^ 


4J 


(U 


en 




u 


^ 


U.I 


•-H 


o 


o 


3 


•w 




O" 


u 


!U 


(U 


trj 


r; 


u 


<— 1 


ttJ 




(U 


--t 


» 


OS 


CQ 



Ill 




jaqtunN qooj^ 



112 
for the ionization process. Although the new corona discharge power 
supply has served to generate a stable corona discharge, the needle-to- 
orifice separation has been limited. For a given current, a larger 
needle-to-orifice distance requires a larger voltage from the power 
supply. The current power supply is limited to 5 kV. This has limited 
the needle-to-orifice separation to a few millimeters, and therefore it 
is proposed that a new discharge power supply be constructed based on a 
10 kV power supply. 

The vacuum system has proved inefficient in handling the large gas 
loads provided by the API source. When a normal gas load is put on the 
system, the forepump pressure (the pressure at the entrance to the 
mechanical pumps) is a few mtorr. With the API source installed, this 
pressure increases to approximately 400 mtorr. On the TSQ 70, one 
mechanical pump serves as the backing-pump for both turbomolecular 
vacuum pumps. For future work, the turbomolecular pump for the 
analyzer region should be provided with an independent backing-pump. 
This may allow for pressures closer to the theoretical value for the 
analyzer region and may allow the size of the orifice to be increased, 
increasing the sensitivity of the instrument. 

The API source canister for this work was designed to allow 
flexibility in its configuration. Since the one-orifice system has 
been shown to perform adequately, a new, second- generation source 
canister should be built that has a larger diameter source region, is 
shorter, and provides micrometer control over the needle position. The 
walls of the original canister are approximately 0.5" thick stainless 
steel. This makes the source heavy and cumbersome. Therefore, the 



113 
canister should be built out of a lighter material such as aluminum, or 
the wall thickness should be reduced where possible. 

To achieve total computer control of the API/MS/MS instrument, 
digital -to -analog (DAC) converters should be implemented to control the 
corona discharge current, the ±500 V orifice potential, and the 
potentials for the additional lenses that the API lens system has added 
to the mass spectrometer. 

To sample gases, the sample must currently be brought near the 
entrance to the glass sampling tube. A larger capacity gas sampling 
fan should provide the ability to sample from greater distances. 

Applications 

Finally, the applications of API/MS /MS with this instrioment should 
be extended. While this instrument has been applied to simple 
mixtures, it has not be applied to any "real" samples. With 
improvements listed above, the sensitivity of the instrument should be 
significantly increased. Also, this instrument has yet to be applied 
to samples which are more amenable to negative ionization. 

The research presented in this thesis has demonstrated the ability 
to develop a functional API source for a commercial, turbomolecularly- 
pumped tandem mass spectrometer. The future research proposed here 
should prove this combination to be a viable method for the direct 
atmospheric monitoring of trace compounds. 



LITERATURE CITED 



1. Finnigan-MAT Corporation, 355 River Oaks Parkway San Jose CA 
95134-1991. 

2. Knewstubb, P. F. ; Sugden, T.M. Nature 1958, 181, 474. 

3. Knewstubb, P. F. ; Sugden, T.M. Ibid. 1958, 181, 1251. 

4. Knewstubb, P. F. ; Sugden, T.M. 7th International Symposium on 
Combustion (Oxfo rd. igSS') Butterworth, London, 1959, 247. 

5. Knewstubb, P. F. ; Sugden, T.M. Proc . R. Soc. London I960 A255 
1920. ~ ' 

6. Kebarle, P.; Hogg, A. M. J. Chem. Phv.g 1965, 42, 688. 

7. Hogg, A. M. ; Kebarle, P. Ibid. 1965, 43, 449. 

8. Hogg, A. M.; Haynes , R. N. ; Kebarle, P. J. Am. Chem. Snr 1966 
88, 28. 

9. Kebarle, P.; Searles , S. K. ; Zolla, A.; Scarborough, J • 
Arshadi, M. Ibid. 1967, 89, 6393. 

10. Shahin, M. M. J. Chem. Phys 1965, 45, 2600. 

11. Carroll, D. I.; Dzidic, I.; Homing, E. C; Stillwell, R. N 
AppI. Spectrosc. Rpv 1981, 17(3) . 337-406. 

12. Mitchum, R. K. ; Korfmacher, W. A. Anal. Chem 1983 55 
1485A-1499A. ' — ' 

13. Horning, E. C; Homing, M. G.; Carroll; D. I., Dzidic, I • 
Stillwell, R. N. Anal. Chem. 1973, 45, 936-943. 

14. Dzidic, I.; Carroll, D. I.; Homing, E. C. Anal. Chem. 1976, 48, 
1763 - 1768 . 

15. Carroll, D. I.; Dzidic, I.; Homing, M. G.; Horning, E. C. Anal 
Chem. 1974, 46, 706-710. " 

16. Mitchum, R. K.; Althaus, J. R. ; Korfmacher, W. A.; Moler, G F 
Advances in Mas.q Soectrom 1980, 8B, 415-421. 



114 



115 

17. Mitchum, R. K. ; Althaus, J. R. ; Korfmacher, W. A.; Rowland, K. 
L. ; Nam, K. ; Young, J. F. Biomed. Mass Spectrom. 1981, 8, 
539-545. 

18. Tsuchiya, M. ; Taira, T. Int. J. Mass Spectrom. Ion Phvs . 1980, 
34, 351-359. 

19. Carroll, D. I.; Dzidic, I.; Stillwell, R. N. ; Haegele, K. D.; 
Horning, E. C. Anal. Chem. 1975, 47, 2369-2273. 

20. Scott, R. P.; Scott, C. G.; Munroe , M. ; Hess, J. J. Chromatogr. 
1974, 99, 395. 

21. Arpino, P. J.; Dawkins , B. G.; McLafferty, F. W. J. Chromatogr. 
Sci. 1974, 12, 574. 

22. Yost, R. A.; Fetterolf, D. D. Mass Spectrom. Rev. 1983, 2, 1-45. 

23. Caldecourt, V. J.; Zackett, D.; Tou, J. C. Int. J. Mass . 
Spectom. Ion Phvs. 1983, 49, 233-251. 

24. Extrel Corporation, P.O. Box 11512, Pittsburgh, PA 15238. 

25. Sciex Inc., 55 Glen Cameron Rd, Thornhill, Ontario, L3T 1P2 
Canada . 

26. Bruins, A. P.; Covey, T. R. ; Henion, J. D. Anal . Chem . 1987, 
59, 2642-2646. 

27. Bruins, A. P.; Weidolf, L. 0.; Henion, J. D. Anal. Chem. 1987, 
59, 2647-2652. 

28. Henion, J.D.; Weidolf, L. 0.; Lee, E. D.; Covey, R. C. 35th 
Annual Conference on Mass Spectrometry and Allied Topics 1987, 
15-16. 

29. Snyder, A. P.; Kremer, J. H.; Liebman, S. A.; Yost, R. A. 35th 
Annual Conference on Mass Spectrometry and Allied Topics 1987 , 
181-182. 

30. Davidson, W. R. ; Shusan, B. I.; Fulford, J. E. ; Thomson; B. A.; 
Tanner, S. D.; Ngo, A. 32nd Annual Conference on Mass 
Spectrometry and Allied Topics 1984, 7. 

31. Thomson, B. A.; Davidson, W. R. ; Lovett, A. M. Environ. Health 
Perspectives 1980, 36, 77. 

32. Glish, G. L. ; McLuckey, S. A. 35th Annual Conference on Mass 
Spectrometry and Allied Topics 1987, 290-291. 

33. French, J. B.; Reid, N. M. ; Buckley, J. A. U. S. Patent No. 
4,023,398, May 17, 1977. 



116 

34. French, J. B.; Reid, N. M. ; Buckley, J. A. U. S. Patent No. 
4,121,099, October 17, 1978. 

35. Bertan Associates, 121 New South Road, Hicksville, NY 11801. 

36. Karasek, F. W. Anal. Chem. 1974, 46, 710A-720A. 

37. Precision Aperture, P.O. Box 10863, Fort Wayne, IN 46803. 

38. Fite, W. L. Research note #1, Extranuclear Laboratories, Inc. 
January, 1971. 

39. Levy, D. L. Sci. Amer. 1984, 250, 96-109. 

40. Emco High Voltage Co., 11126 Ridge Rd. , Sutter Creek, CA 95685. 

41. Product and Vacuum Technology Handbook: Leybold-Heraeus , 
Inc.: San Jose, CA 1986. 

42. McGilvery, D. D. Latrobe University, Department of Physics and 
Chemistry, Bundoora Victoria, Australia, 1977. 

43. Dahl, D. A.; Delmore, J. E. Idaho National Engineering 
Laboratory, EG&G Idaho, Inc., P.O. Box 1625, Idaho Falls, ID 
83415. 

44. Good, A.; Durden, D. A.; Kebarle , P. J. Chem. Phv. 1970, 52, 
212-221. 

45. Harrison, A. G. Chemical Ionization Mass Spectrometry : CRC Press 
Inc.: Boca Raton, FL, 1982. 

46. Jordan, T. E. Vapor Pressure of Organic Compounds : Interscience 
Publishers, Inc.: New York, NY, 1954. 

47. Bossel, U. "Investigation of Skimmer Interaction Influences on 
the Production of Aerodynamically Intensified Molecular Beams", 
thesis. University of California-Berkeley, 1968. 

48. Barrow, G. M. Physical Chemistry : McGraw-Hill Book Company: New 
York, NY, 1979. 

49. Beam Dynamics, Inc., 708 E. 56th St., Minneapolis, MN 55417. 

50. Custom Service Technologies, 4300 N. W. 23rd Ave, Suite 289, 
Gainesville, FL 32602. 



BIOGRAPHICAL SKETCH 

Kenneth Paul Matuszak was born on March 21, 1961, at Cook County 
Hospital, in Chicago, Illinois. He attended St. Robert Bellarmine 
Grammar School for grades 1-8 and graduated from Albert G. Lane 
Technical High School in Chicago, IL, while earning athletic letters in 
three sports and was one of ten students selected by the faculty to 
receive the Wabeno award (the school's most prestigious award, for 
scholarship, citizenship, and leadership). 

After high school he attended Coe College in Cedar Rapids, Iowa, 
earning a Bachelor of Arts degree in 1983 for completing the 
requirements for both the chemistry and mathematics degrees in 4 years. 
At Coe, he also earned 5 varsity athletic letters in track and cross- 
country. He completed a semester of senior research at Argonne 
National Laboratories, in Argonne, IL, where he studied electron 
transfer reactions in frozen glass matrices. 

In August, 1983, he entered the University of Florida to begin work 
on a Ph.D. in chemistry. Following graduation. Dr. Matuszak will be 
working as an applications chemist at the Finnigan-MAT Corporation in 
San Jose, CA, a corporation which specializes in the research, 
development, and production of mass spectrometers. 
He is newly married. 



117 



I certify that I have read this study and that in ray opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. 



Richard A. Yost, Cha/rman 
Associate Professor of Chemistrv 



I certify that I have read this study and that in ray opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. 



John R. Eyler 
Professor of Chemistry 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. 



ar B. Hoflund 1/ 



Gar B. Hoflund 

Professor of Chemical Engineering 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. 




JoHn G.\Dorsey 

Assd^ia^e Professor of Chemis\try 




I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. 



lA 




Cliffor^'7. Johnston 
Assistant' Professor ol 




Soil Science 



This dissertation was submitted to the Graduate Faculty of the 
Department of Chemistry in the College of Liberal Arts and Sciences and 
to the Graduate School and was accepted as partial fulfillment of the 
requirements for the degree of Doctor of Philosophy. 



April, 1988 



Dean, Graduate School 



UNIVERSITY OF FLORIDA 



3 1262 08556 7617