Advanced mass spectrometric methods of jet fuel analysis

MISSING IMAGE

Material Information

Title:
Advanced mass spectrometric methods of jet fuel analysis
Physical Description:
viii, 160 leaves : ill. ; 28 cm.
Language:
English
Creator:
Gehron, Michael Joe, 1956-
Publication Date:

Subjects

Subjects / Keywords:
Jet planes -- Fuel -- Analysis   ( lcsh )
Mass spectrometry   ( lcsh )
Spectrometer   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1988.
Bibliography:
Includes bibliographical references.
Statement of Responsibility:
by Michael Joe Gehron.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001108621
oclc - 19761666
notis - AFK5067
sobekcm - AA00004801_00001
System ID:
AA00004801:00001

Full Text












ADVANCED MASS SPECTROMETRIC METHODS
OF JET FUEL ANALYSIS
















BY

MICHAEL JOE GEHRON


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



UNIVERSITY OF FLORIDA


1988






























To the Glory of GOD
whose grace through Jesus Christ
has brought me here and is taking me there.


To O. Dean Martin
Minister of Trinity United Method Church of Gainesville
whose teachings and courage has been an inspiration in my life.


To the "Original Little Leaguer"
Harold "Major" Gehron
who passed away on November 15, 1987.


To Mom and Dad


















ACKNOWLEDGEMENTS


I would like to thank my research director and friend Richard A.

Yost for the great discussions, time and effort he put into the

preparation of this research and dissertation. I would also like to

thank him for making my graduate career at the University of Florida

quite enjoyable.

I sincerely thank Gerhard M. Schmid and James D. Winefordner for

seeing me through till the end. Special thanks go to Joseph J. Delfino

and Kirk S. Schanze for also serving on my committee and reviewing this

work.

I would like to thank the entire group for making me feel at home

after the big switch. Special thanks go to Jodie Johnson, Mike Lee,

Ken Matuzsak, Todd Gillespie, Randy Pedder, and Mark Hail for all their

help; I mean all their help.

Special thanks go to a special friend, Bill Davis.

Thanks go to Dave White who taught me to do it rather than sit

around and theorize on the outcome.

I love my wife Cheryl, my little Laura, and the Bun in the oven.

They are what its all about.


















TABLE OF CONTENTS


Page

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

ABSTRACT ......................................................... vii

CHAPTERS

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

Objectives ................ .................... .......... 1
Background ................................................ 2
Type Analysis ........................................ 2
Total Fuel Analysis .................................. 5
History and Reviews of Hydrocarbon Type Analysis ........ 5
Specifications of Jet Fuels ............................. 9
Instrumental Methods of Analysis ........................ 14
Sample Introduction ................................ 14
Short Column GC/MS and Simulated Distillation ...... 15
Long Column GC/MS for TFA .......................... 16
Ancillary Jet Fuel Analysis Techniques ................. 17
Alternative Ionization Methods .................... 17
Mass Spectrometry/Mass Spectrometry ................ 18
Outline of Thesis ...................................... 19

II EXPERIMENTAL ............................................ 21

Materials ............................................... 21
Apparatus ............................................... 22
Mass Spectrometry .................................. 22
Gas Chromatography Conditions ...................... 22
Other Gas Chromatography Studies .................. 23
Data Handling Techniques ....................... ......... 24
Clarification of ASTM D-2789 ....................... 24
Computer Hardware and Software .................... 28
Ion Summation Chromatography ...................... 28
Simulated Distillation for Determining
Numbers .......................................... 30
Occupational Hazards .................................... 30











III GAS CHROMATOGRAPHIC/MASS SPECTROMETRIC HYDROCARBON TYPE
ANALYSIS .............................................. 32

GC Sample Introduction ....................... .......... 32
Ancillary Short-Column Experiments ..................... 40
Analytical Figures of Merit .............................. 42
Characterization of the GC/MS/HTA Method ................ 48
Comparison of ASTM and GC/MS/HTA Results ............. 48
Simple Mixture Analysis .......... .................. 52
Inverse Calibration Matrix Variation ............... 60
Evaluation of Olefin and Cycloparaffin Isomers ..... 69
Evaluation of Methods for Average Carbon Number
Determinations ................................... 78
GC/MS Simulated Distillation ...................... 79
Comparison Between GC/MS/SD and GC/MS/HTA .......... 83
Comparison Between GC/MS/HTA and ASTM .............. 83
Effect of Chromatographic Resolution on Carbon
Number Calculations .............................. 86
Summary of Results ..................................... 87

IV PRELIMINARY INVESTIGATIONS OF ALTERNATIVE IONIZATION
TECHNIQUES FOR JET FUEL ANALYSIS ..................... 90

Background ............................................... 90
Low Energy Electron Ionization .................... 90
Methane Chemical Ionization ....................... 91
Experimental ............................................. 92
Electron Energy Variation ......................... 92
Methane CI ............................................ 92
Results and Discussions ............................... 93
Low Energy Electron Ionization .................... 93
Methane Chemical Ionization ....................... 97
Summary of Results ..................................... 101


V PRELIMINARY INVESTIGATIONS OF GC/MS/MS TECHNIQUES
FOR JET FUEL ANALYSIS .................................. 104

Background ..................... ........................ 104
Mass Spectrometry/Mass Spectrometry ............... 104
Application of GC/MS/MS to Jet Fuel Analysis ...... 105
Experimental ........................................... 107
Results and Discussion ................................. 108
MS/MS for Hydrocarbon Type Analysis .............. 108
Other GC/MS/MS Modes of Analysis .................. 115
Differentiation of Isomeric Hydrocarbon
Types ........................................... 117
Summary of Results .................................... 124











VI CONCLUSIONS AND FUTURE WORK ............................

Summary of Results .....................................
Hydrocarbon Type Analysis .........................
Calculation of Average Carbon Number ..............
Ancillary Mass Spectrometric Methods ................


APPENDICES


A MATRIX INVERSION METHOD ................................


B EXAMPLE GC/MS/HTA CALCULATION ..........................


C GC/MS/HTA TURBO BASIC PROGRAM CODE .....................


D PRELIMINARY TEST METHOD FOR HYDROCARBON TYPE ANALYSIS
OF JET FUEL BY GAS CHROMATOGRAPHY/MASS SPECTROMETRY ..


BIBLIOGRAPHY ....................................................


BIOGRAPHICAL SKETCH .............................................


130


131


136



146


155


160

















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

ADVANCED MASS SPECTROMETRIC METHODS
OF JET FUEL ANALYSIS

BY

MICHAEL JOE GEHRON

April, 1988

Chairman: Richard A. Yost
Major Department: Chemistry


Hydrocarbon type analysis (HTA) is widely used to evaluate the

chemical composition of petroleum and shale-derived products, such as

jet fuels. The HTA methods provide a knowledge of the macro-

composition of the jet fuels which is important in engine design and

performance. Presently the methods used for HTA of jet fuels are

limited by the boiling point range of the analysis and the

instrumentation requirements. The HTA method for gasolines (used for

jet fuels), American Society of Testing and Materials standard test

method D-2789, was developed in the mid-1950s on magnetic sector mass

spectrometers which are no longer commercially available.

A gas chromatography/mass spectrometry (GC/MS) method has been

developed for the hydrocarbon type analysis (HTA) of jet fuels. The

method employs a short-column capillary (3 meters) which is interfaced

directly into a Finnigan TSQ45 tandem mass spectrometer. Advantages









of this method include employment of present day GC/MS instrumentation

and simple GC sample introduction. Additionally, a simulated

distillation profile may be obtained based on specific hydrocarbon

types which allows for the determination of more accurate average

carbon numbers of each hydrocarbon type. The results of the HTA on a

series of fuels and synthetic fuel mixtures will be reported along with

a complete characterization of the GC/MS/HTA method and results. The

results and methodology of the GC/MS simulated distillation method for

average carbon numbers will be reported.

A preliminary investigation of alternative MS techniques for jet

fuel analysis has been performed for the development of methods for

total fuel analysis (TFA) (determination of 95% of all components

present at 0.05% weight or greater). These techniques include

alternative ionization techniques such as low energy electron

ionization and methane chemical ionization; and tandem mass

spectrometry techniques such as daughter scan MS and parent scan MS.

The application and results of these techniques will be presented in

regards to the HTA and TFA of jet fuels.


viii


















CHAPTER I

INTRODUCTION


The need to understand the chemical composition of aviation fuels

has never been more important than in the last decade. This need stems

from two important aspects: the recent world-wide shortages of

petroleum crude reserves which have resulted in a transition to

alternative crude sources; and the desire to design and operate high

performance jet engines with the knowledge of fuel composition as a

controlling criterion. This has led to a resurgence in the development

of new analytical techniques for petroleum analyses and more

specifically aviation fuel analyses. In this laboratory investigations

are presently underway for the application of gas chromatography/mass

spectrometry (GC/MS) and GC/MS/MS for analysis of aviation fuels (in

particular turbine jet fuels). There are two modes of analysis of

interest, hydrocarbon type analysis (HTA) and total fuel analysis

(TFA). The concept of each application will be discussed below.



Objectives

This research project is organized around two goals. The first is

to develop, implement, and characterize a gas chromatography/mass

spectrometry (GC/MS) hydrocarbon type analysis (HTA) for jet fuels.









2

These fuels represent a specific boiling point fraction or range of

components of crude oil. The second goal is to begin preliminary

investigations into the development of a GC/MS method for total fuel

analysis (TFA) (determination of 95% of all components present at 0.05%

weight or more) of these jet fuels. Vast differences in

chromatographic separation will be required in these two tasks. The

ability to integrate the two methods with one instrumental

configuration (GC/MS) is a goal of significant importance.



Background

This section will discuss the idea of type analysis or more

specifically hydrocarbon type analysis of fossil fuels and the need and

significance of new methods. Total fuel analysis will also be

discussed illustrating the need for specific detection systems.

Type Analysis

The concept of "Type Analysis" is not widely known among analytical

chemists. One may ask, "what type (of) analysis did you perform on the

samples?" And one may reply, "a type analysis was performed on the

samples." Thus the confusion begins. Type analysis in general does

not provide information on specific individual components in a mixture

but rather supplies information on the relative amounts of certain

classes of compounds. In many applications the quantitation of

specific compounds is not required and may provide much more

information then is actually desired. This is particularly true in

cases where the analyst is only interested in the overall macro-

composition of a very complex mixture. The type analysis of









3

hydrocarbon mixtures, such as gasoline and heavier oils, represents

such an application of type analysis. Hydrocarbon type analysis is a

widely used procedure for the evaluation of petroleum feedstocks and

other complex mixtures. In 1951 Ralph Brown of the Atlantic Richfield

Corporation developed the first mass spectrometric technique for

Hydrocarbon Type Analysis (HTA) (1). The Brown method operates on the

premise that certain classes of compounds or homologous series, for

example paraffins, form a specific series of fragmentation ions in the

electron ionization source of the mass spectrometer. In this example

the paraffins form the series of mass fragments 43, 57, 71, 85, and 91.

Not all species of this series form all or only these ions. Similarly

the paraffin series is not the only class or homologous series of

compounds which contain any of these ions. Therefore an extensive

calibration scheme is required to normalize for this mixing of

fragmentation series between hydrocarbon types. Details of the present

method will be discussed in Chapter II, Clarification of ASTM D-2789.

Since the Brown method, there have been numerous improvements on his

method due to improved calibration schemes (2,3), and improved

instrumentation (4,5). In 1962 standard Hydrocarbon Type Analysis

methods for gasoline (6) and diesel fuels (7) were adopted into the

American Society of Testing and Materials (ASTM) standards. There have

been several revisions of these methods over the years, but the basis

of the methods remains the same. The methods, ASTM D-2789 and D-2425,

perform the analysis on magnetic sector instruments which are no longer

in commercial production. Because of this limitation, there is a much

greater demand for samples to be analyzed than there are laboratories









4

equipped to perform them. Additionally, the means of sample

introduction into the mass spectrometer, which is via a heated glass

inlet volume or microburette, is a time-consuming procedure and

facilitates component discrimination (8). The technique also relies on

manual measurements of peak height intensity which eliminates the

ability to computerize the calculations. Consideration of these

limitations illustrates the need for a modern HTA method performed on a

widely available modern instrument.

At the 1987 meeting of The American Society of Mass Spectrometry

in Denver, the ASTM committee on petroleum analyses (9) expressed

interest in the development of new and/or improved methods of HTA in

the fuels area. Presently several methods are used for hydrocarbon

type analysis such as the fluorescent indicator adsorption (FIA)

procedure which covers the determination of saturates, non-aromatic

olefins, and aromatics (10). Limitations of this method include long

analysis time (3-4 hours per sample), poor precision, and lack of

automation capabilities. Limitations of these methods led to the

development of several liquid chromatographic methods (11), but

limitations of these methods include the inability to separate and

quantitate enough classes. With the development of a GC/MS/HTA several

advantages would be realized. First, convenience in instrumentation,

since most modern analytical laboratories are equipped with a GC/MS,

many laboratories will be able to perform the analysis, and the sample

load will be decreased. Second, GC introduction will provide improved

sample introduction techniques which are more convenient and less prone

to component discrimination. Additionally, the use of chromatography









5

will provide information such as distribution of components as a

function of boiling point and make possible the extension of the method

to total fuel analysis by increasing the chromatographic resolution.

Total Fuel Analysis

The concept of total fuel analysis (TFA) is one to which most

analytical chemists may relate. The objectives are to quantitatively

determine specific components or classes of components of jet fuel.

One must be knowledgeable of the complexity of jet fuels to appreciate

the formidable task at hand. A typical jet fuel may contain from 400

up to 4000 individual components. Davis and Giddings (12) and Martin

and co-workers (13) have shown theoretically that, in order to

chromatographically resolve 90% of the components in a 100 component

mixture, 20,000,000 theoretical plates would be required. Considering

a standard capillary GC column of 0.25 mm i.d. and 0.25 pm thickness, a

column length of 5 km or 3 miles would be required for such an

analysis. This theoretical calculation merely illustrates the need for

more selective detection methods which could utilize normal high

resolution chromatography.



History and Reviews of Hydrocarbon Type Analysis

Discussions on the history of HTA of petroleum-derived mixtures

such as fuels will quickly lead one to a general discussion of the

early history and development of mass spectrometry. The impetus for

early mass spectrometry research was generally fueled by the petroleum

industry and the war-time demands on the industry to supply aviation

fuels of superior quality with good quality control and assurance.









6

Recently there have been several review articles which deal

specifically with this history (14,15). The Consolidated Engineering

Corporation (CEC) enjoyed the first commercially successful mass

spectrometer (CEC Model 21-101) which became the prototype instrument

which was be utilized by Ralph Brown (1) in the first mass

spectrometric HTA of gasoline mentioned above.

There have been many methods developed in the past for the

determination of hydrocarbon types in gasolines and fuels. Some of

these methods relied on rather laborious and time-consuming

separations, chemical reactions, and physical determinations. One

procedure developed by Kurtz et al. (16) required 24 man-hours for

fractionation, chemical treatment and determination by refractive

indices. Rampton (17) employed distillation, silica gel percolation,

hydrogenation, and measurements of refractive indices for the

determination of paraffins, cycloparaffins, aromatics, olefins, and

cycloolefins. The first applications of mass spectrometry to

hydrocarbon fractions were to the analysis of gaseous mixtures (18).

Later, with the advent of improved mass spectrometers and inlet

systems, the analysis of the lower-boiling fraction of gasolines was

performed (19-21). In 1951 Ralph Brown introduced the HTA method for

gasolines which represented a new concept and application of mass

spectrometry. The application of MS HTA soon expanded to the analysis

of heavier fractions such as kerosene and heating oils (2,22). At this

time O'Neal and Wier (4) and Brown et al. (23) independently described

an improved all-glass heated inlet system which would not only expand

the boiling point range of MS analysis even greater, allowing for the









7

analysis of waxes and other low vapor pressure compounds, but also

improve the precision and accuracy of the method as a whole. Of

particular interest at that time was more in-depth analysis of the

aromatic fractions of these complex heavier mixtures (24-27). Previous

to these MS techniques, little was known of the chemical make-up of

such aromatic fractions. From the middle 1950s, which may be regarded

as the end of the rapid growth period of mass spectrometry research in

the applications of petroleum-derived mixtures, there have been very

few publications appearing in the literature. However, with the advent

of high resolution mass spectrometers in the early 1960s, there was an

additional surge of interest in MS techniques for the analysis of

petroleum-derived complex mixtures (28-31). Traditionally, the major

use of high resolution MS was for the identification and structural

determination of pure compounds or major components of a simple

mixture. Since heavier petroleum fractions contain such a large

variety of compounds compared to lower boiling fractions such as

gasoline, a single low resolution mass spectrum cannot be interpreted

in terms of hydrocarbon types. The high resolution MS techniques could

single out each hydrocarbon type based on the exact mass of the

fragment ions and/or molecular ions. The classification scheme became

known as the "Z number" classification based on the formula CnH2n+z

where z is a number which represents the degree of unsaturation and/or

ring closure and signifies the hydrocarbon type. A more extensive

discussion on "Z number" classification may be found in the paper by

Teeter (29) which introduces an improved high resolution MS hydrocarbon

type analysis for 22 hydrocarbon types. This method determines the









8

relative amounts of eight types of saturates, ten types of aromatics,

and four types of sulfur-containing aromatics. In addition to high

resolution techniques a number of low resolution techniques began to

appear in the literature. The Robinson and Cook method (32) followed

by the Amoco method (26) took advantage of the appearance of saturate

fragments at lower masses and aromatic fragments at higher masses.

This enabled the spectra to be subdivided into a saturate and aromatic

region allowing mathematical treatment and normalization. The

elemental analysis, carbon number distribution, average molecular

weight, and the general aromaticity of coal-derived distillates were

determined by the combination of low-voltage, high resolution MS (33).

Other analytical techniques apart from MS techniques have been

utilized in the pursuit of a suitable HTA of fuels and other complex

petroleum mixtures. Based on the classic silica gel separation of

saturates, olefins, and aromatics, new high performance liquid

chromatographic methods have been developed which utilize silver-coated

silica gel particles and n-pentane mobile phases (34). Separation of

saturates, single ring aromatics, double ring aromatics, and

polynuclear aromatics was obtained for olefin free kerosenes and diesel

fuels using bonded aminosilane stationary phases (35). An excellent

review is available on the application of HPLC and GC to hydrocarbon

group analysis (36-38). Most recently several workers have evaluated

supercritical fluid chromatography (SFC) for the separation of

hydrocarbon types (39). Separation of paraffins, olefins, and

aromatics was obtained in two reports; using SF6 mobile phase, pressure









9

programming, and FID detection (40); and using silver nitrate

impregnated silica gel stationary phases with CO2 mobile phase and

refractive index detection (41). Combination of SFC with MS shows

great promise for future applications to fuels analysis.



Specifications of Jet Fuels

With the world-wide shortage of conventional high quality crude

oils, there has been a dramatic decrease in the availability of jet

fuels. As a result of this shortage, the industry has experienced a

transition from petroleum sources towards utilization of alternative

sources such as oil shale, tar sands, coal liquid, and heavy oils for

the production of jet fuels. Additionally fuels are being synthesized

by blending various refinery streams and mixing of pure compounds.

This shift has created increases in processing cost as well as many

uncertainties in market stability. More importantly here, new unknowns

have been generated in the physical and chemical characteristics of

these new source fuels. Recently it has become more important in the

design and operation of jet engines to be able to correlate the

physical parameters of jet fuels with their chemical composition. With

the addition of these new source fuels, there is a need for new and

improved physical and chemical methods of jet fuel analysis.

Traditionally the determination of the physical parameters of jet fuels

has been the focus of jet fuel analysis. Table 1.1 lists some of the

more common physical parameters determined for jet fuels (42). The

importance of a few of these parameters is as follows: thermal


















Table 1.1

List of Important Physical Parameters Determined for
Aviation Fuels and the Corresponding ASTM Methods Used


Physical Parameter


Test


ASTM Method


Volatility


Fluidity


Distillation
Specific Gravity
Vapor Pressure

Freezing Point


D-86
D-287
D-328, D-2551


D-2386


Combustion


Corrosion

Stability


Contaminants


Additives


Heat of Combustion
Aniline-Gravity Product
Knock Rating: Lean Fuel
Rich Fuel

Copper Strip Test

Potential Gum
Precipitate

Existent Gum
Water Reaction:
Interface Rating
Volume Change

Tetraethyllead Content
Dye Content


D-240, D-2382
D-611, D-287
D-2700
D-909


D-130

D-873
D-873

D-381

D-1094
D-1094


D-3341, D-2599, D-2547
D-2392









11

stability is important since jet fuels are exposed to extremely high

temperatures and minimal decomposition is desired; viscosity of fuels

as a function of temperature is important since the fuels must remain

fluid in flights where the fuels are exposed to extremely low

temperatures in storage; the density of fuels (or the volumetric

energy) determines the ultimate range in miles of the aircraft. For

more information on any of the specifications listed in Table 1.1,

refer to the appropriate reference listed. Most methods used in the

determination of the above mentioned physical parameters are ASTM

standard methods and represent important contributions to fuel and jet

engine research.

As mentioned above, the specifications of jet fuels are based

mainly on the physical parameters or usage requirements. The single

composition requirement common to all aviation fuels is they must

consist almost completely of hydrocarbon compounds. There are specific

limitations on particular hydrocarbon types such as aromatics and

olefins due to performance criteria. Additionally, low level additives

may be present as well as low level contaminants of various sorts.

Jet fuels or more specifically turbine jet fuels are made up of

hundreds of hydrocarbons that may be divided into four subgroups or

hydrocarbon types: paraffins and isoparaffins (branched paraffins);

cycloparaffins or naphthens; aromatics; and olefins. The former two

types make up the saturates and the latter two types the unsaturates.

Figure 1.1 illustrates some example compounds of each hydrocarbon type.

Paraffins and cycloparaffins are generally the major type component

of jet fuels. The paraffins are the most stable components and do not









Paraffin CH3
CH3(CH 2)CHCH3


CH3
Monocycloparaffin CH3CH CH3
isopropyl -cyclohexane


Dicycloparaffin


decalin


Alkylbenzene


ethylbenzene


Indane Tetralin


Naphthalene
2-methylnaphthalene


Figure 1.1.


SCH


Structures of some example compounds representative of
the six hydrocarbon types. Some specific examples are
also shown.


C):









13

readily react with materials in which they come in contact with such as

elastomers, paints, and various metals. They are clean-burning

compared to other hydrocarbons and have a very high heat release per

unit weight. Both of these advantages are due to the high hydrogen-to-

carbon ratio. Cycloparaffins have a lower hydrogen-to-carbon ratio

which lowers their heat release per unit weight, but increases their

density or volumetric energy. Cycloparaffins are also clean-burning

and stable and have a lower freezing point compared to paraffins of

equal carbon number.

Aromatics are fully unsaturated ring structures which may also

incorporate saturated structures. Because of the very low hydrogen-to-

carbon ratio, they have a much greater volumetric energy but a lower

heat release per unit weight than the paraffins. The aromatics are

reactive and tend to cause swelling of rubber seals. They produce

large amounts of smoke while burning, and a high-luminosity flame.

This particular aspect is important when considering secret

reconnaissance missions of the Armed Forces. Because of these factors

the maximum concentration of aromatics permitted in jet fuels is 20-25%

Virgin fuels generally contain no more than 10-20% aromatics.

Olefins are the most reactive hydrocarbon types and are capable of

reacting with many materials. They react with air to form varnishes or

rubber-like materials. They can dimerize or polymerize forming high

molecular weight contaminants. Because of their high reactivity,

olefins are generally not found in natural or virgin crudes. They may,

however, be formed in the refinery cracking processes. Specifications









14

limit the content of olefins to less than 5 percent in order to reduce

varnish and gum formation.

Other non-hydrocarbon types which may be found in aviation fuels

include heteroatomic species such as sulfur-, oxygen-, and nitrogen-

containing compounds. These include mercaptans, sulfides, thiophenes,

and free sulfur as example sulfur types; alcohols and naphthenic acids

as example oxygen types; and aza and amino polynuclear aromatics as

example nitrogen types. Combinations of these contaminants tend to

form high molecular weight gums and deposits when exposed to air during

storage. This creates many problems such as filter clogging, sticking

of valves, and plugging of orifices. Additives present which help to

eliminate some of these problems include antioxidants, metal

deactivators, fuel icing inhibitors, and corrosion inhibitors.



Instrumental Methods of Analysis

Sample Introduction

The utilization of GC/MS techniques can provide more up-to-date

methods of sample introduction into the mass spectrometer. This is a

direct advantage over the traditional ASTM methods, in which a heated

glass inlet or microburette is employed, and the sample is introduced

by effusion through a small orifice. A small fraction of the total

sample is analyzed with the microburette technique, in order to

minimize component discrimination which increases as the effusion of

the sample into the mass spectrometer proceeds. The method is time-

consuming and requires the maintenance of many instrumental parameters.

The GC/MS technique will vaporize the entire sample and analyze the









15

entire volume which has been loaded onto the column. Because of the

possibility of component discrimination during sample loading onto the

GC column, component discrimination must be evaluated as a function of

injection port temperature and injection techniques.

Short Column GC/MS and Simulated Distillation

An important concept of this HTA method is that chromatographic

resolution of the jet fuel components is not a matter of pursuit.

Trehy, Yost, and Dorsey (43) have discussed some of the important

instrumental considerations of normal GC/MS using short capillary

columns. In the GC/MS/HTA application of short capillary columns, the

GC is mainly used as a convenient device for sample introduction into

the mass spectrometer. The short columns, along with providing short

analysis times, provide the ability to observe the distribution profile

of the eluting jet fuel as function of column temperature. This

concept is similar to that of simulated distillation (44,45) which has

been utilized by the petroleum industry for many years to provide a

breakdown of the boiling point distribution of the hydrocarbon mixture.

The area percent of the eluting fuel is determined at various points

along the profile or chromatogram. Each area percent corresponds to a

specific retention time or boiling point. These boiling points are

determined by a parallel eluting sample of n-paraffin homologous series

in which the boiling points are known. A non-polar column must be

employed to assure the correlation of retention to boiling point. The

temperature corresponding to the 50 percent area point corresponds to

the average boiling point of the fuel. This is a useful physical

parameter, although the extraction of an average carbon number from the









16

results is not accurate. Application of GC/MS, however, can provide

not only average boiling points but also average carbon number. This

is provided by the ability of the mass spectrometer to distinguish

between hydrocarbon types within the fuel mixture. Corresponding

homologous series may be compared to the corresponding ion summation

chromatogram of the particular hydrocarbon type and an average carbon

number may be obtained. These experiments represent an important

advancement in the development and understanding of hydrocarbon type

analysis.

Long Column GC/MS for TFA

The quest for a GC/MS technique for TFA will require high

resolution (long-column) capillary gas chromatography. Since only a

select number of chromatographic peaks will be pure components,

additional selectivity in addition to chromatographic resolution will

be required. Many researchers have reported work in the fuels area

with chromatography alone (46,47). Here the application of MS, which

includes various modes of ionization, and MS/MS will be evaluated for

means of chromatographic peak deconvolution. There have been numerous

reports in the literature which have applied GC/MS techniques to fuel

analysis. Gallegos (48) has described a high resolution GC/MS method

for the analysis of a shale derived oil. Aczel has determined compound

type distribution in refinery streams and synfuels using GC/MS (49).

The compound distribution of a jet fuel has been determined by

recombination of separately determined components using GC/MS (50).











Ancillary Jet Fuel Analysis Techniques

The development of GC/MS methods for total fuel analysis (TFA) will

require the utilization of ancillary MS techniques such as low-energy

electron ionization, chemical ionization, and MS/MS techniques. These

techniques, when combined with chromatographic separation, will provide

specific detectors for the deconvolution of multicomponent

chromatographic peaks discussed earlier. The following is a review of

the application of such MS techniques.

Alternative Ionization Methods

Chemical ionization (CI) (51) is a softer ionization technique

which is capable of producing an increased abundance of molecular ion

and minimal fragmentation compared to conventional electron ionization

(El). Harrison discusses the use of CIMS with various reagent

compounds for the characterization of hydrocarbons (52). Although

there have been significant applications of CI to the characterization

of single classes of compounds in fuels and other complex mixtures,

there have been no known publications on the application of CI to an

overall hydrocarbon type analysis of jet fuels. Sieck characterized

gasolines for their aromatic content using low-energy photoionization

cyclohexane CI (53). Sieck, Burke, and Jennings (54) used N20 CI for

screening of aviation fuels. The OH- ion was employed for negative ion

CI, which eliminated ionization of aliphatic compounds. Other

applications of CIMS have been devoted to the determination of

polynuclear aromatic hydrocarbons (PAH) in fuels (55), as well as aza

and amino PAH in coal derived liquids (56). Negative ion ammonia CI

GC/MS has been used for the quantitation of sulfur-containing compounds









18

in gasoline (57). Bauer, Schubert and Enke (58) used methanol CI for

the characterization of heterospecies in the presence of hydrocarbons.

The essential aspect of CI GC/MS techniques in regard to total fuel

analysis is the ability to deconvolute co-eluting components of the

fuel. In the current research only methane chemical ionization is

evaluated.

Electron ionization (El) is considered the work-horse of mass

spectrometry. One aspect of EI/MS which is rarely explored in

petroleum analysis is the utilization of low-energy El. Since various

compound types have different ionization potentials, these differences

may be utilized as a source of compound-type discrimination. One of

the earliest applications of low-energy ionization MS (referred to as

low-voltage MS) was the determination of the unsaturated hydrocarbon

fraction in a petroleum naphtha (59). In this method only molecular

ions of the unsaturated species are formed, thus eliminating saturate

interference. Aczel and Johnson utilized low-energy electron

ionization in conjunction with high resolution MS for the analysis of

aromatic fractions of complex petroleum mixtures (31). As mentioned

above, low energy photochemical ionization was used for the analysis of

aromatics in gasoline (53). The use of low-energy electron ionization

GC/MS is evaluated here for the characterization of separate

hydrocarbon classes. This technique and its usefulness will be

discussed.

Mass Spectrometrv/Mass Spectrometry

Mass spectrometry/mass spectrometry (MS/MS) is quite useful for the

direct analysis of complex mixtures such as coal derived liquids.









19

Recently there have appeared in the literature many applications of

MS/MS to the analysis of complex petroleum mixtures (56,58,60-71).

There exist two MS/MS instrumental techniques which are generally used,

mass-analyzed ion kinetic energy spectrometry (MIKES) (72,73) and

triple quadrupole mass spectrometry (TQMS) (74). To date, no known

application of MS/MS to HTA of fuels has appeared in the literature;

however, as with CI applications, most MS/MS applications have been

dedicated to the determination of a single class of compounds or a

single component. In the current research, several aspects of TQMS

have been evaluated for possible HTA methods or TFA methods of jet

fuels. These results will be discussed along with the possibility of

combining MS/MS with alternative ionization techniques.



Outline of Thesis

The remainder of the thesis is divided into five chapters. Chapter

II, Experimental, will describe all experimental, instrumental, and

computational techniques. Chapter III, Gas Chromatography/Mass

Spectrometry/Hydrocarbon Type Analysis, will contain experimental

results and discussions based solely on the development,

implementation, and characterization of the GC/MS/HTA method. Chapter

IV, Alternative Ionization Techniques, will contain results and

discussions on the use of these techniques in the application of TFA

and HTA of jet fuels. Chapter V, Gas Chromatography/Mass

Spectrometry/Mass Spectrometry, will contain results and discussions on

the use of GC/MS/MS in the application of TFA and HTA of jet fuels.









20

Finally Chapter VI, Conclusions and Future Work, will summarize the

total effort and describe suggested future work. As in any research

effort each question answered brings about several new questions. This

project is no exception.


















CHAPTER II

EXPERIMENTAL


The following chapter will discuss the experimental parameters

which are involved in the development, characterization, and

implementation of the hydrocarbon type analysis of jet fuels. The

experimental parameters which deal specifically with studies in total

fuel analysis, such as chemical ionization and MS/MS will be discussed

in those specific chapters. The general experimental parameters, such

as materials, jet fuel preparation, and instrumentation will also be

discussed here.



Materials

All chemicals were ACS grade unless otherwise specified and were

used as received without further purification. Standards were

purchased from various vendors and were used as received without

further purification. All jet fuel samples were obtained from Pratt &

Whitney-Fuels Division, West Palm Beach, Florida. The fuels were

shipped at ambient temperatures without refrigeration in vials with

screw-cap or crimped cap teflon-lined replaceable septa. After every

use in the laboratory, the septa were replaced with new ones and

samples were refrigerated at 5 C.











Apparatus

Mass Spectrometer

A Finnigan MAT TSQ45 gas chromatograph/triple-stage quadrupole mass

spectrometer/data system was used in all GC/MS studies. An electron

energy of 70 eV and an electron current of 300 pA were used. The mass

spectrometer was mass calibrated with FC43 (perfluorotributylamine) in

accordance with the instrument's specifications. The mass spectrometer

was tuned, also with FC43, to obtain a specific ion intensity ratio of

m/z 69 to m/z 219, and m/z 502 to m/z 220. These ratios were monitored

to insure reproducible relative abundances of the fragmentation pattern

in regards to instrumental tuning alone. The low mass ratio is of

greater importance in these studies since this is the general mass

range in question. The mass spectrometer was scanned from 35 to 650

amu at a total scan time of 1.0 s. The electron multiplier was

operated between 950 and 1000 V, with the conversion dynodes at 3000

V. The preamplifier gain was set at 108 V/A with the fine adjustment

(zeroing) set for minimal acquisition of baseline noise. The mass

spectrometer was interfaced to a Finnigan Model 9610 gas chromatograph.

All GC operations were under computer control. The capillary GC

columns were inserted directly into the ion source of the mass

spectrometer through a heated interface. The ion source pressure was

0.3-0.6 Torr depending on the column head pressure. The ion source

temperature was 190 C.

Gas Chromatography Conditions

A 3 m DuraBond fused silica bonded phase open tubular (capillary)

column (J&W, Rancho Nuevo, CA) with an inner diameter of 0.25 mm and a









23

DB5 bonded phase (SE-54 equivalent or 5% phenyl-methyl silicone) of 1.0

pm thickness was operated at a head pressure of 8 pounds per square

inch of helium carrier gas. Split-type injection was used with a split

flow of 50 mL/min and a septum sweep of 2 mL/min. The injection port

temperature was 250 C and the transfer line was 280 C. The injection

port teflon lined septum was replaced after 30 40 injections and the

injection port glass insert was checked frequently and cleaned whenever

needed. The column temperature was programmed from 50 C to 250 C at 15

C/min after a hold time of 1 minute. Usually the column would not

require the full temperature to elute the entire jet fuel sample.

Generally 0.5-1.0 pL injections of neat jet fuel samples were made.

The injection technique used was as follows in order to reduce

injection error: the GC injection syringe (Hamilton) was thoroughly

rinsed between samples with hexane, followed by the sample jet fuel of

interest; the sample volume was pulled up from the needle volume into

the syringe volume; the needle was inserted into the GC injector; a

count of three allowed the needle to come to equilibrium temperature;

the sample was injected at a rate of approximately 1 pL/s; this

position was held for a count of three and the needle was removed.

This method decreases injection discrimination due to uneven heating of

the needle and components contained therein (75,76).

Other Gas Chromatographic Studies

Besides GC/MS, studies involved the application of GC with flame

ionization detection. A Varian 3300 gas chromatograph was used with a

flame ionization detector (FID). The same chromatographic conditions

were used as stated above including the column head pressure. However,









24

although the column head pressures were equivalent for the two systems,

different flow rates would result due to the different exit pressures.

This instrument was used for the studies of component discrimination as

a function of injection port temperatures and injection modes (split,

splitless, and on-column). Acquisition of peak area, peak height, and

retention time was obtained with an IBM 9000 minicomputer and the IBM

CAPMC4 Chromatography Applications Program. Additionally, the FID

system provided two advantages: a continuous mode of detection which

resulted in more highly resolved chromatograms, and increased

sensitivity (compared to full scan MS) which when combined with high

resolution gas chromatography (longer columns, 30 meters) would

demonstrate the extreme complexity of the jet fuels.



Data Handling Techniques

This section covers aspects of the data handling and reduction

which proved to be a major effort in obtaining the stated goals.

Considering the hydrocarbon type analysis alone, the mathematical

computations were quite abstract and had to be figured out conceptually

before application could be made. Below, these computations will be

discussed, along with the hardware used and the software generated to

deal with them.

Clarification of ASTM D-2789

The purpose of this section is to clarify the computational aspects

of ASTM D-2789 standard test method (6), which is the computational

basis of the GC/MS/HTA method developed here. The ASTM method, as it

is referred to, determines the volume percent of six separate









25

hydrocarbon types in low olefinic gasolines: paraffins,

monocycloparaffins, dicycloparaffins, alkylbenzenes, indanes and

tetralins, and naphthalenes. Each hydrocarbon type is determined based

on a characteristic set of ions as shown in Table 2.1. The method also

calculates the average aromatic and average paraffinic carbon numbers

based on mass spectral data. Calculations are made based on the

inverse calibration tables shown in Table 1 of the ASTM D-2789 standard

method (6) and are dependent upon the average carbon numbers. It is

these aspects of the ASTM method, development and application of the

inverse calibration matrices, which require clarification. The

calculations are similar to six simultaneous equations with six

unknowns; however, the unknowns are made up of a complex array of

variables and the set of equations used are chosen based on two other

variables as well.

The inverse calibration matrices were prepared by determining the

values Z43/T, E41/T, etc. for each hydrocarbon class mixture at each

carbon number, where T=Z(Z43+Z41+Z67+S77+Z103+1Z28), the sum of all the

characteristic hydrocarbon sums (Z). The compositions of the

hydrocarbon class mixtures at each carbon number (up to carbon number

nine) used in these determinations are given in Table 3 of the ASTM

standard (6). Consider this example calculation of the inverse

calibration matrix for carbon number six: the Z43/T, 441/T, Z67/T, and

277/T are determined for each of the three hydrocarbon type mixtures as

shown in the ASTM Table 3 under C6 blends. The results are given in

Table 4 of the ASTM standard (6) under the appropriate hydrocarbon type

and C6 listing along with the results for larger carbon numbers. For a

























The Characteristic
in


HYDROCARBON TYPE


PARAFFINS


MONOCYCLOPARAFFINS


DICYCLOPARAFFINS


ALKYLBENZENES


INDANES/TETRALINS


NAPHTHALENES


Table 2.1

Ion Summations Used For Each Hydrocarbon Type
the Hydrocarbon Type Analysis


MASSES OF EACH CHARACTERISTIC ION SUM (E) (m/z)


E43 = 43+57+71+85+99


S41 = 41+55+69+83+97


E67 = 67+68+81+82+95+96


E77 = 77+78+79+91+92+105+106+119+120+133+134
147+148+161+162

Z103= 103+104+117+118+131+132+145+146+159+160


Z128= 128+141+142+155









27

given hydrocarbon type, all values (Z/T) are divided by the largest

value of that set. The numbers are arranged in a matrix in proper

order. All elements of the array are then multiplied by an appropriate

pressure sensitivity factor. Table 5 of the ASTM standard (6) gives

the pressure sensitivity factors for each hydrocarbon class and carbon

number. These values represent the instrumental response as a function

of microburette pressure. The means of measuring the amount of sample

introduced into the mass spectrometer was via this monitoring

procedure. The C6 matrix is then inverted after multiplication by the

pressure sensitivity factors. The methods used for inverting a 3x3 and

3x4 matrix (77) are shown in Appendix A; this procedure is best handled

by computer programs. The elements of the C6 matrix are then

multiplied by the corresponding liquid volume factors, also given in

Table 5 of the ASTM standard, which are an indication of the liquid

volume per unit pressure. The matrix values are then divided by 100

and are in units of volume/counts.

Examination of the elements of the inverse calibration matrices

reveals much information on the use of the table. The carbon number 6

matrix illustrates this quite well. A relatively large positive entry,

such as +0.009016 for Z43/T/paraffins, is a good similarity

coefficient, as expected since paraffins response will contain a large

abundance of Z43 ions. Similarly, the -0.000003 for Z77/T/paraffins

indicates a low similarity coefficient, which is also expected since

very little E77 ions will be present for paraffin compounds. The

inverse relationship of these numbers represent the normalization

function of the inverse calibration matrices described above. A









28

thorough evaluation of the application of the inverse calibration

matrices will not be made since the method is quite common in

multicomponent analysis problems. The method involves the

multiplication of a 1x6 matrix (a paraffinic 1x3 and a aromatic 1x3) to

a 6x6 matrix and normalizing the results to 100%. However, as

mentioned above, the matrix used is dependent upon the average carbon

number of the sample in question. Since an integral average carbon

number is rarely obtained, two inverses should be applied with weighted

results. An example calculation of the GC/MS/HTA method is given in

Appendix B.

Computer Hardware and Software

The GC/MS/HTA method required extensive computations once the mass

spectral data had been obtained. This is evident from the discussion

above which described the methods and complexities of the calculations

used. A series of TURBO BASIC programs was written to completely

automate these computations. Additionally, a terminal emulator program

(Persoft, Smarterm 240) was utilized for the direct transfer of data

from the MS data system to an IBM PC. The program used for calculating

the volume % results and the average carbon numbers is shown in

Appendix C. Other programs included an inverse calibration development

program for carbon number 6; and a program which translates the

emulator-captured data to usable format for the HTA program.

Ion Summation Chromatograms

The characteristic ion sums listed in Table 2.1 were plotted as

separate mass chromatograms. These mass chromatograms, as shown in

Figure 2.1, are referred to as ion summation chromatograms and








Paraffin 29
16016


Monocycloparaffin
44096


Dicycloparaffin
27040


Alkylbenzene
34368




Indane/Tetralin
9776




Naphthalene
1312


A~'iI 1''iAe h


3:20 6:40 10:00


13:20 16:40


Retention Time (min)


Figure 2.1.


The ion summation chromatograms of each hydrocarbon type
with the relative areas of each for jet fuel JP-8X 2414.


! Nh.











i


'Ct-~jlY\;k"i~,~









30

represent the chromatogram of each ion sum listed in Table 2.1. The

area of each ion summation chromatogram was determined and normalized

to 100%. The area percent were then calibrated with the ASTM inverse

calibration matrices after determination of the average aromatic and

average paraffinic carbon numbers. All these calculations were done

automatically, as mentioned above.

Simulated Distillation for Determining Carbon Numbers

The application of ion summation chromatograms provided the

opportunity to investigate the average carbon numbers of the jet fuels.

Figure 2.2 illustrates how the average aromatic carbon number was

obtained using this method. The ion summation chromatogram represents

the aromatic ion series. After determination of the area centroid of

this chromatogram the line may be projected down to the chromatogram of

the corresponding homologous series. The value indicates the aromatic

average carbon number of the fuel. Variations were made using this

technique, such as combinations of ion summation chromatograms, as will

be discussed in the next chapter.



Occupational Hazards

As a final note, the occupational hazards of working with jet fuels

are quite evident considering the nature of these complex mixtures. It

should go without saying to avoid exposure to these materials by using

a well ventilated hood, wearing gloves, and keeping the samples in

properly sealed vials. The major concern of working with these fuels

is the effect of long term exposure. Many of the components in these

mixtures are fat-soluble and therefore bio-accumulation is likely (78).





















































KIL
















Figure 2.2.


alkylbenzenes


e50 SCAN
8:20 TIME


cia
I,


alkylbenzene homologous series



C14 C16 C18


tia 280 388 400 ?0 E'c
1:48 3:20 5:08 E:? :

The aromatic ion summation chromatogram of a jet fuel
JP-8X 2414 on a 3 meter DB5 column. The centroid of the
total area, indicated by the arrows, is referenced to the
alkylbenzene homologous series. The carbon number is
calculated to be 10.95.


















CHAPTER III

GAS CHROMATOGRAPHIC/MASS SPECTROMETRIC
HYDROCARBON TYPE ANALYSIS


Characterization of the hydrocarbon type analysis (HTA) method was

performed by evaluation of simple and complex synthetic mixtures,

perturbation of jet fuels, and comparison of results to existing ASTM

results. The experiments performed were to determine the effects of

various instrumental and method parameters on the results of the HTA of

jet fuels. The results of these experiments, as well as the simulated

distillation method for the determination of average carbon numbers,

will be discussed.



GC Sample Introduction

With the application of gas chromatography (GC), a new method of

sample introduction into the mass spectrometer was used for HTA. It

was therefore important to evaluate the specifics of sample

discrimination with respect to this sample introduction technique.

There are several injection techniques used in capillary gas

chromatography. These are split and splitless, which depend on the

flash vaporization of the sample within a heated glass-lined injection

volume, and on-column injection which allows the sample to be loaded

directly onto the capillary column at ambient temperatures. Specifics









33

of these capillary GC injection techniques are discussed elsewhere

(75,76). An aromatic mixture containing 012, 014, C16, and 018

straight chain alkylbenzene in equal concentrations was used for the

evaluation of component discrimination. The effect of injection port

temperature and injection technique was evaluated. Figure 3.1 and

Figure 3.2 illustrate the effect of injection port temperature on

component discrimination for splitless and split injections,

respectively, with flame ionization detection. In both cases the

ratios of peak areas were normalized to the C14/C12 ratio at the 280 C

injection port temperature to help illustrate the matter of

discrimination. Both cases indicate component discrimination taking

place at lower temperatures by the ratios being lower. This was due to

less efficient vaporization of the heavier components at lower

temperatures, as was expected. As the temperature was raised, the

ratios increased and leveled off, indicating a lack of discrimination

and the absence of component decomposition at higher injection port

temperatures. A larger degree of discrimination takes place for the

higher carbon number alkylbenzenes, as indicated by the larger ratio

spread between the low and high injection port temperatures. The

splitless technique more clearly illustrates these trends. A minimum

injection port temperature of 250 C should therefore be maintained to

avoid component discrimination. These results also demonstrate the

improvement in reproducibility obtained using splitless injections over

that obtained with split injection technique. A similar evaluation of

component discrimination was performed on the GC/MS instrument. Figure

3.3 illustrates the results which are similar to those from the GC/FID

























ToU

N O
co
CO



0
SC\2


0







a


D


I 1111 1 I 11111 II I I 1 I I I 1 I I I I I t I I I I T I I I I I I III

0 0 0 0 0 CC
C0 CO CO N C
-- l 0 0 0 0 C

* oV1^?i Tih~iH ^e~d PQIt~uvnj~oN


0 t o-
0) <



O O


w 0
a 44-1









O W
o J .a
;0 -4





-4 C r
L U *

LA



00







'-4
a 0







0 4J
0 U0

o (N 4.)
P4 U -4 '0a


-O
-0


N





-O





















- CO
: -
- 0
-- -




-4





C



mm


n
)4



.,-4
r4




















0 C







oo




U),
o

.- c








--
SCO


a-
.Ia 0 C l'r




)4-
"- o











S o -u
cJ--








C\ 0c 0 U

\Q -400 0
C' \)







C..)










, o.leEI qf.H :iuacd pozTLUIJOTO
o 0
























N 0

0 C







o
O


ed t(
CO U^


i l 1 111 il l I i l 151
0 0 0 0
O O0
-< '- 0

* OTI1 QmiIH


SI 5 I Ir 151 II 1115 I I I I Ir r~rrr


0
co
0
0~


0 0 0
0> 0 t0
d o d
pazi000ijo^


4-4 W
o
0 C
0




*4 U
4t 4-'












o
lB 4 (-)









co
0) 4
0-l -

P W









ca
W0 0








4-
SC,





04u 0








e,4
5 m r
c





*s
N.
0 U M

c-(f

r- r t
&4U

ho
(1r


A









37

technique. However, discrimination effects are less with GC/MS, as

indicated by the smaller spread of area ratios compared to GC/FID. The

vacuum outlet conditions result in a large pressure drop across the

short capillary column, giving higher flow rates and decreased

discrimination as a function of injection port temperatures.

On-column injection was also evaluated in terms of component

discrimination. It has already been demonstrated that on-column

techniques provide the best precision when compared to the other modes

of injection (79). Figure 3.4 illustrates component discrimination as

a function of initial column temperature. The initial column

temperature in on-column injection corresponds roughly to the injection

port temperature in split/splitless injections. For this reason,

evaluations were not possible at initial column temperatures above 120

C, since the early eluting C12 alkylbenzene component was lost in the

solvent front. The results indicated no discrimination, as would be

expected for on-column injections. However, there appears to be a

small enhancement effect for the less volatile components at lower

initial column temperatures. This cannot be explained at the present

time. Table 3.1 compares splitless and on-column injection

techniques for accuracy and precision. A mixture containing known

weight ratios was prepared and analyzed with GC/FID. The FID response

is proportional to the mass flux of analyte, and therefore direct

correlations may be made between weight ratios and area ratios.

Results for both splitless and on-column injections were high compared

to the actual weight ratio, which may be due to some systematic error.






















0





0




0_
0C


1.35 -


1.30


1.25


1.20


1.15


1.10


1.05


S/


I .UU


Figure 3.4.


C16/C12










C18/C12




C14/C1I


I I I I I i I I I I I 1 i I I I II I I I I I I I I II I I I I I I II I I 7 7 i
20 40 60 80 100 120
Initial Column Temperature (C)


Plot of initial column temperature versus area ratios for
C12, C14, C16, and C18 alkylbenzenes for on-column
injections on a 3 meter DB5 column with FID detection.


















Table 3.1

Comparisons of Accuracy and Precision of Area Ratios of Various
Components Between Splitless and On-Column Injection Techniques


Butylbenzene
Decane


Hexamethylbenzene
Decane


t-Stilbene
Decane


Actual Wt.
Ratio


Splitless

Area Ratio

%RSD


On-Column

Area Ratio

%RSD


0.785


0.280


0.894(.005)

0.54


0.905(.004)

0.44


0.30(.01)


4.32


0.330(.009)


2.82


Note:
number of replicates n=3
number in parenthesis indicates standard deviation


0.374


0.41(.03)


7.24


0.49(.01)


2.02









40

The splitless results agreed more closely with the weight ratios, while

the precision of the on-column results was somewhat better.

These results indicate that, as long as the minimum injection port

temperature is maintained, component discrimination is greatly reduced

if not eliminated. Also, better reproducibility is obtained using

splitless injections.



Ancillary Short-Column Experiments

The use of short (3 m) capillary columns in these studies has led

to the discovery of some interesting chromatographic aspects. The

injection technique previously described in the Experimental chapter,

in which the bulk of the sample volume was removed from the syringe

needle volume in order to let the needle equilibrate to the injection

port temperature, was used to eliminate component discrimination and to

obtain more reproducible injections. This technique when used with

short columns, however, leads to the splitting of the early eluting

chromatographic peaks (Figure 3.5). This was determined to be a

function of injection port temperature. At high injection port

temperatures, the residual amount of sample which remains in the

syringe needle is vaporized prior to the injection and vaporization of

the sample contained in the syringe itself. This results in a "pre-

peak" at high injection port temperatures which is absent at the lower

injection port temperature.

It was also observed that splitless injections provide better

chromatographic resolution than on-column techniques for short

capillary columns. This is presumably due to the large amount of band































































Figure 3.5.


109 158 :0, :-








2


3 4







I II





180 tSO 200 2* c ?CM|
SI1:40 2:38 3:20 4:1. TIflE




Chromatograms of nonane (1), cyclooctane (2),
propylbenzene (3), and decalin (4) using split injections
on a 3 meter column, a) injection port 280 C, b)
injection port temperature 130 C.









42

spreading incurred for on-column injections relative to the column

length. Figure 3.6 and Figure 3.7 illustrate the comparison of

injection techniques for chromatographic resolution for a narrow

boiling point range fuel and a wide boiling point range fuel,

respectively. The improved chromatographic resolution for the

splitless case is indicated by the sharper peaks, which are most

noticeable for the late eluting peaks.



Analytical Figures of Merit

The limits of detection which are determined here do not reflect

the limits that would actually be obtained if individual components

were to be determined. Figure 3.8 illustrates the reconstructed ion

current (RIC) full-scan mass spectra of a four-component mixture at 0.5

volume percent each, which corresponds to approximately 0.4 ng injected

on the column. The signal-to-noise in the RIC was approximately 3,

which represents the limit of detection for the HTA method. The

sensitivities of the four components are given by the slope of the

calibration curves illustrated in Figure 3.9. These sensitivities were

determined in order to compare the results to the actual ASTM

sensitivities in Table 5 of the ASTM D-2789 standard (6). Table 3.2

compares the sensitivities (slopes of the response curves in Figure

3.9) to the sensitivities of Table 5 of the ASTM standard (6), after

normalization to the C9 paraffin (nonane). The results indicate that

similar trends in sensitivities for the four components were observed

for the two methods, although the propylbenzene sensitivity was

significantly greater in the ASTM results.











































Retention


Figure 3.6.


00 8.C 0 4.00
Time (min) Retention Time


Chromatograms of JP-7 jet fuel on a 3 meter column
comparing on-column and splitless injections.


(min)




















On-Column


I I I II I I I I I
0 4.00
Retention Time


I TI I 7 I 7
8.0
(min)


Retention Time


Figure 3.7.


Chromatograms of JP-4 jet fuel on a 3 meter column
comparing on-column and splitless injections.


8.C
(min)


Splitless














































-- 3--


--__-



S----..


- u


"`


Jr, CC.




CO
0

C



-E-

















LWC




0
C U
r:
j Q














S3







.5




-.



^ ^ a

c, 3-




















DECALIN


000 n\r \fl


CYCLOOCTANE

NONANE


oUUU-








40000








20000








1 _-


Figure 3.9.


Response curve obtained by linear regression for nonane,
cyclooctane, propylbenzene, and decalin at various volume
percentages.


i1 i- i i- i l l l l I I I I i I I I l ] r I i 1 i [ i | i i i I I i [ i i |
2 4 6 8
VOLUME PERCENT


I






















Table 3.2

Comparison of Sensitivities of ASTM and GC/MS/HTA
Results for a Four-Component Mixture
Normalized to the ASTM Sensitivity of Nonane
(Response per Volume %)


ASTM
SENSITIVITY


GC/MS/HTA
SENSITIVITY


1.99


2.52


1.72


2.15


2.06


1.74


1.72 1.72


COMPONENT


DECALIN


PROPYL-
BENZENE

CYCLO-
OCTANE

NONANE









48

The precision of the HTA method was determined by replicate

analyses of several jet fuel samples. The precision was dependent upon

the jet fuel analyzed or the composition of the fuels. The precision

of the HTA method is summarized in Table 3.3, in which the average

percent relative standard deviation is listed by hydrocarbon type for

all fuels analyzed in replicate. The high average relative standard

deviations were a result of that particular volume percent being very

low. The dashed lines indicate a consistent result of zero volume

percent.

Characterization of the GC/MS/HTA Method

Comparisons of ASTM and GC/MS/HTA Results

One of the major tasks in this research effort was to analyze a

number of real jet fuel samples with the newly developed GC/MS/HTA

method and to compare the results to those from the long-established

ASTM D-2789 method. These comparisons would provide information on the

validity of the ASTM inverse calibration matrices with GC sample

introduction and quadrupole mass analysis. Table 3.4 shows these

comparisons for five jet fuels. Several methods were used to evaluate

trends in the comparison versus composition, but none could be found.

The accuracy of the ASTM method has not been evaluated; therefore, the

comparison does not necessarily reflect the accuracy of the GC/MS/HTA

method.

Table 3.5 demonstrates the comparisons of several synthetically

prepared jet fuels. The "recipe" results were calculated from the

known volumes of the components added. Again these comparisons do not

necessarily represent accuracy of the GC/MS/HTA method, since the


















Table 3.3

The Average Relative Standard Deviation for the GC/MS/HTA of
all Replicate Analyses and Separate Jet Fuels


HYDROCARBON
TYPE


AVERAGE
OF ALL FUELS


PARAFFIN


MONOCYCLO-
PARAFFIN

DICYCLO-
PARAFFIN

ALKYLBENZENE


INDANE/
TETRALIN

NAPHTHALENE


AVERAGE CARBON
NUMBER

AROMATIC


PARAFFIN 0.3


JP-4
2455


JP-7


JP-8X
2414


JP-8X
2429


1.2


0.6


1.2


18.8


2.5


18.1


3.8


10.7


5.3


1.8


5.4


2.4


6.7


0.5


0.9


0.5


0.2


0.7 0.3


0.1 0.0



















cw

ri



u4


41


ca
0
4:
<,


00








0
o







\O O'

o ,0

0' 0


5-'


0 -4


o -4
-) 4-4
















o w



r- 4-4

o c






0u

1-4m

044








'4-4
Cc
Cu
o0 -
(u
*r4
Q o.


PI


-


0 -z1





C-


5-



-l





C-










Co M






1;
U 00


C-'




CM


'. v


14 C(N








C( 0





co






0 0











-4


1-4
0
o



o\ o\
































CI4 -4
0-- --





















0C -\
C-













-4 C4


1\












5-^
o \o


















0 -4


a-' m aN E-
0 <\ 00 .
ccC co -C

: r- 4: p P4 Er

( -1 I cn C14 En CM C U C

\ 4 'E-4 \ E- \ H X

o 4 *< 4 I

o P ol0


C C%









cn 00





oo




o
C) C


(N


in -4









I- o%





0'




Un Un


00



a'
CV)


E -
\ E <


C


0
o0




a





-4
L1








Co
00





















0' r
-4 .
rno
4.1

0









aN
Sr
N 4J


ON
09





Co





i-4
20

















Co
0
0 E
N 0








o
i-l-





C-
































U,

u, -.
-4


- 0

0 0


0
in o
-d" fi


0 Co
CM CM


o

-4 C


a)









Q.
cJ






NJ
U2








0

3V
b<,-4







ca
oC






Ou









0
0




u
0f
,,,.


0


.4 0
.-1 -


0

O
m CM


co

O 0o
.4 C.,


44


I l 04
a u
a 1-
Q\ u
ZU W
< U
Ud


co
0
CO


u








S .)
*1-]
io

u,
<)


O
-4 Ln
-4





I- )
-c. t


Ln

Cs ~ c"
>-i i-


1 (a)
-14






-'4-
MC
C e









0







o -

C.) '4.,
0cu a


rw
0cu




0
4-
44

tfl


'1,

I~-


1-e

0

-i



o
0


n E-


0\


u









52

synthetic mixtures were formulated from complex petroleum fractions

(Exxon Corporation) such as Isopar C, G, and M (isoparaffinic and

paraffinic mixtures), and Aromatic 100 and 150 (aromatic mixtures).

These mixtures were obtained by simple fractionation and bulk

separation techniques, and were not intended for analytical standards.

Two of these mixtures were analyzed by GC/MS for specific group

components. The Isopar M mixture contained significant amounts of

cycloparaffinic components, with trace aromatics. The Aromatic 150

contained indanes and naphthalenes; the paraffin content was not

determined. The monocycloparaffin recipe results of the olefin-

containing fuels (Grand Mix and All Groups) agree better when the

olefin content is added to the monocycloparaffin content. These two

hydrocarbon types produce similar fragmentation patterns which tend to

equate the two types under the HTA method. More details of the

olefin/cycloalkane effects on the GC/MS/HTA method will be covered

later in this chapter and in Chapter V.

Simple Mixture Analysis

Simple mixtures were used to evaluate the accuracy of the GC/MS/HTA

method. Teeter states in his paper on high resolution mass

spectrometry for 22 hydrocarbon type analysis (29) that no sample can

be prepared which is complex enough to be a valid test for any HTA

method. This "Teeter Rule," as it has become known in our laboratory,

poses a real paradox for validation of these methods: any mixture

which is complex enough to be a valid test of the HTA matrix

calculation is too complex to be prepared. Certainly such a paradox is

difficult for any analytical chemist to accept. These experiments were









53

designed to test this rule as well as to evaluate the accuracy of the

GC/MS/HTA method. Table 3.6 shows the GC/MS/HTA results of a simple C6

mixture compared to the actual calculated volume percentages. The

mixtures were prepared solely of C6 components. The results clearly

indicate the inaccuracy in the analysis of such a simple sample,

although the zero percent components are correctly identified and the

aromatic average carbon numbers agree quite well. Table 3.7 shows the

GC/MS/HTA results of a simple C10 mixture compared to the actual

calculated volume percentages. In this case, better agreement was

obtained for those types present as well as the zero percent types.

The aromatic average carbon was well off the mark, yet the paraffin

average carbon number was very close.

The next step was to evaluate a mixture which contained a spread in

carbon numbers. Table 3.8 shows the GC/MS/HTA results of a simple C8-

C9 mixture compared to the actual calculated volume percent. The

results again indicated the inaccuracy of the method for simple

mixtures, although the precision of the analysis was quite good, as

indicated by the average percent relative standard deviation of 2.3%

for replicate injections. Again, the average aromatic carbon numbers

do not agree as well as the average paraffinic carbon numbers. Table

3.9 shows the GC/MS/HTA results of a simple C8-C10 mixture compared to

the actual calculated volume percent. The monocycloparaffin volume

percent was quite low, which resulted in a larger volume percent for

the dicycloparaffin and the two other hydrocarbon types present. The

average aromatic carbon number was low while the average paraffinic

carbon number was high compared to the calculated results. This seems


















Table 3.6

Comparison of HTA on a Simple C6 Mixture
With Actual Volume Calculated Results


HYDROCARBON
TYPE

PARAFFIN


MONOCYCLO-
PARAFFIN

DICYCLO-
PARAFFIN

ALKYLBENZENES


INDANE/
TETRALIN

NAPHTHALENE


AVERAGE CARBON
NUMBER

AROMATIC

PARAFFIN


PREPARED
BY VOLUME %


70.6


17.6


0.0


11.8


0.0


0.0


6.0

6.0


Simple C6 Mixture Make-up:


Paraffin:


Monocyclo-
paraffin:


300 pL hexane
200 pL 3-methylpentane
100 pL 2,2-dimethylbutane


100 pL cyclohexane
50 AL methylcyclopentane


100 iL benzene


HTA
VOLUME %


52.3


23.7


0.0


24.0


0.0


0.0


6.00

6.75


Alkylbenzene:


















Comparison of
With Actual


HYDROCARBON
TYPE

PARAFFIN


MONOCYCLO-
PARAFFIN

DICYCLO-
PARAFFIN

ALKYLBENZENE


INDANE/
TETRALIN

NAPHTHALENE


AVERAGE CARBON
NUMBER

AROMATIC

PARAFFIN


Table 3.7

HTA on a Simple C10 Mixture
Volume Calculated Results


PREPARED
BY VOLUME %

33.3


0.0


33.3


33.3


0.0


0.0






10.0

10.0


HTA
VOLUME %

29.8


0.0


35.8


34.3


0.0


0.0






8.44

9.96
















Table 3.8

Comparison of HTA on a Simple C8-C9 Mixture
With Actual Volume Calculated Results


HYDROCARBON
TYPE


PARAFFIN


MONOCYCLO-
PARAFFIN

DICYCLO-
PARAFFIN

ALKYLBENZENE


INDANE/
TETRALIN

NAPHTHALENE


AVERAGE CARBON
NUMBER


AROMATIC

PARAFFIN


PREPARED
BY VOLUME %


44.5


33.4


0.0


13.9


8.3


0.0


9.00

8.43


HTA
VOLUME %

39.2 (.5)


21.3 (.2)


6.9 (.07)


17.1 (.4)


15.4 (.9)


0.0


8.68

8.44


C8-C9 Mixture Composition (pL)


C8: octane 50
2,5-dimethylhexane 50
2,3,4-trimethylpentane 20
octane 50
ethylcyclohexane 20

C9: nonane 50
2,2,5-trimethylhexane 20
isopropylcyclohexane 50
indane 30
propylbenzene 50
















Table 3.9

Comparison of HTA on a Four-Component Mixture
With Actual Volume Calculated Results
(in Volume Percent)


HYDROCARBON
TYPE


CALCULATED
BY VOLUME


HTA
EXPERIMENTAL


PARAFFIN


MONOCYCLO-
PARAFFIN

DICYCLO-
PARAFFIN

ALKYLBENZENE


INDANE/
TETRALIN

NAPHTHALENE


AVERAGE CARBON
NUMBER

AROMATIC


9.0 9.59


27.0


9.8


33.7


29.2


0.0


0.3


8.68


PARAFFIN









58

to be the normal trend for these simple mixtures. Table 3.10 shows

the analysis of the same mixture as in Table 3.9 with the hexane

solvent included in the HTA calculation. The changing HTA volume

percent results generally agree with those expected, with the exception

of the monocycloparaffin results, which increase rather than decrease.

This would be expected since the mass spectrum of hexane contains a

large relative abundance of m/z 41, which is an ion of the

monocycloparaffin E-series. The average aromatic carbon number for the

GC/MS/HTA results remains constant when hexane is included in the

calculation, as it should. The average paraffinic carbon number

decreases as expected and shows excellent agreement within the

predicted value.

To summarize, the analysis of simple mixtures by GC/MS/HTA does not

result in correct volume percent, which may simply be a verification of

the Teeter Rule. The paraffin and alkylbenzene volume percent showed

the best and most consistent agreement. The average paraffinic carbon

numbers consistently showed the best agreement except for the simple C6

mixture. The average aromatic carbon number was consistently low

compared to the actual results. The C10 mixture analysis was quite

accurate, unlike the other two simple mixtures. This may be an

erroneous result or may have some significance. Perhaps the sample

make-up of the C10 mixture more closely matches that of the original

C10 mixture used in the calibration matrix, or the matrix calculation

works better for a more closely distributed (single carbon number)

mixture distribution.
















Table 3.10

Comparison of HTA on a Four-Component Mixture
With Actual Volume Calculated Results
With the Hexane Solvent Included in the Calculation
(in Volume Percent)


HYDROCARBON
TYPE


CALCULATED
BY VOLUME


HTA
EXPERIMENTAL


PARAFFIN


MONOCYCLO-
PARAFFIN

DICYCLO-
PARAFFIN

ALKYLBENZENE


INDANE/
TETRALIN

NAPHTHALENE


AVERAGE CARBON
NUMBER

AROMATIC

PARAFFIN


Sample Mixture in 1000 iL hexane:


nonane
cyclooctane
n-propylbenzene
decalin


70.5


16.2


5.5


7.7


0.0


0.0


8.72

6.70











Inverse Calibration Matrix Variation

An initial goal of this project was to validate the use of the

existing (ASTM D-2789) inverse calibration matrix with quadrupole mass

analysis. Since the matrix was prepared with one specific set of

calibration mixtures, it was desired to determine the effect of

calibration sample make-up on the inverse calibration matrix. A series

of C6 components was used to prepare an inverse calibration matrix for

carbon number 6 only. The hydrocarbon types were limited to paraffins,

monocycloparaffins, and aromatics because there are no C6 members of

the other types. Table 3.11 indicates the sample make-up used for the

determination of each hydrocarbon type similarity coefficient in the

inverse calibration matrix. The similarity coefficients are

normalization factors for the mixing of mass spectral ions of the

characteristic ion series. Various combinations were used for the

inverse calibration matrix. The calculational method used is similar

to that outlined in the experimental chapter. A computer program was

written to calculate the C6 inverse calibration matrix based on the

calculational method shown in Appendix A. A simple C6 mixture was

prepared and analyzed using the various inverse calibration matrices.

The results are compared to the actual volume calculated results in

Table 3.12. The results show a significant amount of fluctuation,

indicating the dependence of the calibration sample make-up on the

matrix results. The closest agreement comes with the P4C2B1

combination, whose monocycloparaffin make-up is similar to the test

mixture make-up but whose paraffin make-up is not. The worst agreement

comes with the P4CIB1 combination, which is expected since the paraffin






















Table 3.11

Component Codes and Make-Up for Preparation of Samples
Used in the Various Carbon Number 6 Inverse Calibration Matrices


Component Code


Sample Make-Up (in AL)


hexane 3-methylpentane


500
250
100
1000


250
100
500
250


2,2-dimethylbutane


100
500
250
100


Monocyclo-
paraffins


cyclohexane


500
250


methylcyclopentane


250
500


Alkylbenzenes


Paraffins


benzene


1000




















Table 3.12

Comparison of Volume Percent Results on a Simple Carbon Number 6
Mixture Using the Various Combinations of Calibration Mixtures as Listed
in Table 3.11 Previously Shown


Calibration Standard
Composition
by Component Codes


Paraffin


Monocyclo-
paraffin


Alkylbenzene
(benzene)


P1C1B1

P1C2B1

P2C1B1

P2C2B1

P3C1B1

P3C2B1

P4C1B1

P4C2B1


ACTUAL VOLUME CALCULATED


Simple C6 Mixture Make-up:


70.6


Paraffin:


Monocyclo-
paraffin:


17.6


300 pL
200 AL
100 uL


11.8


hexane
3-methylpentane
2,2-dimethylbutane


100 AL cyclohexane
50 AL methylcyclopentane


100 AL benzene


Alkylbenzene:









63

and monocycloparaffin make-up is quite different from the test mixture

make-up. The aromatic results, however, are in the closest agreement

with this combination. This series of experiments illustrates the

importance of the make-up of each single type component in

determination of the similarity coefficients and the inverse

calibration matrices. This C6 example represents the simplest case in

this analysis. The complications faced with increasingly higher carbon

number mixtures and matrices are unknown. The complexities of these

mixtures may tend to eliminate these problems or enhance them.

Perturbation Jet Fuel Analysis

The GC/MS/HTA validation experiments described so far indicate that

there still remains a need for validation of the GC/MS/HTA method since

the above validation experiments contained limitations such as the

Teeter Rule in the simple mixture analysis, and the unknown composition

of the petroleum fractions in the synthetic fuel comparisons. To

maintain the required complexity of the mixtures to be analyzed (to

avoid complications due to the Teeter Rule), the jet fuels were

perturbed by addition of a pure component of known volume. Table 3.13

shows the GC/MS/HTA results for a JP8X jet fuel before and after

spiking the fuel with different amounts of various pure components.

The paraffin volume percent of the first three spiked fuels should be

increased to a calculated 28.6%. The fourth spiked fuel, which

contains no paraffin spike, should agree with the "fuel only" results.

These were in fact the results observed. The fuel containing the

nonane and undecane spike showed the closest agreement. The volume
















Table 3.13

Comparison of GC/MS/HTA of 86-POSF-2429 JP8X FUEL
After Spiking with Various Hydrocarbon Types
in Volume %


HYDROCARBON FUEL
TYPE


SPIKE 1*


SPIKE 2


SPIKE 3


PARAFFIN


MONOCYCLO-
PARAFFIN

DICYCLO-
PARAFFIN


0.0


6.6


72.5


ALKYLBENZENE 6.2


INDANE/
TETRALIN


14.4


NAPHTHALENE 0.3


AVERAGE CARBON
NUMBER

AROMATIC 7.50

PARAFFIN 9.02


29.6 (28.6)


1.2 (4.7)


53.8 (57.8)


4.6 (4.4)


10.6 (10.3)


0.2 (0.2)


7.32

9.35


24.6 (28.6)


6.7 (4.7)


52.0 (57.8)


4.9 (4.4)


11.6 (10.3)


0.2 (0.2)


7.53

8.55


32.0 (28.6) 0.2 (0.0)


0.2 (4.7) 15.1 (19.0)


52.2 (57.8) 52.0 (57.8)


4.7 (4.4) 22.9 (18.7)


10.4 (10.3)


0.4 (0.2)


7.48

9.51


9.3 (10.3)


0.5 (0.2)


8.53

9.01


SPIKE 1: 500 pL
SPIKE 2:
SPIKE 3:
SPIKE 4:


FUEL +
+
+
+


100 pL n-nonane + 100 pL undecane
200 pl n-octane
200 pl n-decane
100 pL butylbenzene + 100 pL cyclooctane


Note: numbers in parenthesis indicate the expected results due to
dilution or addition of that specific type.


SPIKE 4









65

percent of the remaining hydrocarbon types of the first three spiked

fuels (paraffinic spikes) should decrease due to dilution and remain

relatively constant. The calculated values are monocycloparaffin,

4.7%; dicycloparaffin, 51.8%; alkylbezenes, 4.4%; indane/tetralin,

10.3%; and naphthalene, 0.2%. On examination of Table 3.13, excellent

agreement is observed for all types for the first three spikes, with

the exceptions of the monocycloparaffin results. Attempts were made to

explain this behavior by analysis of the mass spectra of each

particular component, but no distinctions could be made. Most

paraffins contain a large relative abundance of m/z 41 (30-50%) and

m/z 55 (5-15%) which are ions indicative of the monocycloparaffin type.

The fourth spiked fuel also demonstrates excellent agreement with the

calculated perturbations for the hydrocarbon types not present in the

spike. The monocycloparaffin results were low compared to the

calculated (19.0%), and the alkylbenzene was high compared to the

calculated (18.7%). Again no explanation could be drawn by examination

of the component spectra; however, the sum of the two types

(monocycloparaffin and alkylbenzene) did equal the sum of the

calculated results. The average carbon numbers demonstrated good

correlation with the perturbation of the fuel. The average aromatic

carbon number remained relatively constant for paraffin perturbation

but increased with the butylbenzene (C10) addition. The average

paraffinic carbon number also correlated with the paraffin

perturbation: 9.35 for nonane and undecane (C10 average) addition;

8.55 for octane addition; 9.51 for decane addition; and no change with

the addition of spike number four.









66

This perturbation approach appears to- be a much more successful

approach for validation of the GC/MS/HTA method than the previous two

methods (simple mixture analysis and ASTM comparisons). The much

better agreement seems to further support the "Teeter Rule" and

indicates that additional experiments should be performed.

A second set of fuel perturbation experiments was performed using

the standard addition method. Figure 3.10 illustrates the GC/MS/HTA

results for standard addition of n-propylbenzene to JP-4 2455 jet fuel.

The HTA aromatic type response was compared to the theoretical aromatic

type response of the volume percent added. The more positive slope of

the experimental line indicates that upon increased additions of n-

propylbenzene to the jet fuel an above-average GC/MS/HTA aromatic

response will result. However, the negative x-intercept (-13.2

aromatic volume percent), which indicates the volume percent of the

original fuel, agree to within experimental error with the original HTA

result (12.8). To test the hypothesis stated above in regard to the

above average GC/MS/HTA response of the n-propylbenzene alkylbenzene,

the same fuel was spiked with the Aromatic 150 mixture. Figure 3.11

illustrates the results of this standard addition experiment. The

"real line" intersects with the theoretical line and adheres more

closely to it in the region of the data points. This confirms the

above average response hypothesis of the n-propylbenzene. This may

indicate that the Aromatic 150 better represents the composition of the

original calibration solution used in the preparation of the inverse

calibration matrix. The larger negative x-intercept of the Aromatic

150 indicates that a larger volume percent must be added to equal the
















50.0



40.0



30.0



20.0



10.0



0.0 -
-20.0


n-propylbenzene


/ Real
A Theoretical

12/ /
/-
BA

/A

/&


/A
Ay
A/
A/
A/
A/
A/
Ay


-10.0 0.0
Volume


10.0
Percent


20.0 :30.0
Added


Figure 3.10.


Plot of volume percent standard addition of
n-propylbenzene to jet fuel JP-4 2455. The real line
indicates the experimental results.


0



C



C-















Aromatic


150


STheotetical
/ Real


, r ,i i I I II
-10.0 0.0
Volume


10.0
Percent


20.0 30.0
Added


Figure 3.11.


Plot of volume percent standard addition of Aromatic 150
to jet fuel JP-4 2455. The real line indicates the
experimental results.


50.0 -


a 40.0
Cn



0


30.0



20.0



10.0



0.0 -
-20.C


I
3r









69

theoretical volume percent. This result is consistent with the

composition of the Aromatic 150 mix which is known to contain

components other than alkylbenzenes such as indanes, tetralins, and

naphthalenes. Figure 3.12 compares the mass spectrum of n-

propylbenzene and the average mass spectra of the Aromatic 150 mixture.

The presence of non-alkylbenzene aromatics in the Aromatic 150 is

evident by the presence of m/z 117 and m/z 118, indanes and tetralins;

m/z 128 and m/z 142, naphthalenes.

The above analysis of the standard addition plots was formulated

based on the linear regression of the experimental standard additions.

This seems appropriate when considering the linear theoretical plots;

however, in both cases there was a clear curvature to the data points.

This effect is illustrated in Figure 3.13 in which the experimental n-

propylbenzene data have been connected by a spline fit. The decreasing

slope with increasing volume percent added could possibly be explained

by the partial molal volume of mixing theory, but the degree of change

is much too large. The more negative x-intercept which was obtained

in both of the previous results is not consistent with the types of

additions made and otherwise has no explainable significance. These

results confirmed the original linearization assumptions used above.

Evaluation of Olefin and Cycloparaffin Isomers

A major concern of the ASTM and GC/MS/HTA methods is the presence

of olefins and how they effect the results of the monocycloparaffin

volume percentages. In the ASTM method it is assumed that there is no

difference between the fragmentation pattern of any two olefinic or

cycloparaffin isomers of equal carbon number. This results in an
















































CD4



('4






=,


U)3
!2cm
L)-


to
In
U,.
10
10:


0 h,


, -



c3s---




0)


to
In


U-,


eouepunqv eAJBelIe %


. .. 1 1 I I


II I II I II
















50.0



a 40.0
&4


30.0



20.0



10.0



0.0 --r
-20.0


n-propylbenzene


Real
A Theoretical




A



/ A
/ A
/ /
/ A
/ A


-10.0 0.0
Volume


10.0
Percent


20.0 30.0
Added


Figure 3.13.


Spline fit to experimental volume percent standard
addition of n-propylbenzene to jet fuel JP-4 2455
compared to theoretical addition.









72

uncertainty in the HTA analysis if the presence of olefins is uncertain

or if their concentrations are unknown. The ASTM procedure requires

the subtraction of the olefin volume percent from the monocycloparaffin

volume percent when the olefin content is known. The olefin content

may be determined using the ASTM D-875 standard method. We were

interested in evaluating the effects of the olefin content on the

GC/MS/HTA results and how they would compare to the isomeric

monocycloparaffin content. Table 3.14 shows the differences in the

GC/MS/HTA paraffin, monocycloparaffin, and dicycloparaffin results for

the JP-4 2455 jet fuel after the addition of C7, C9, and C10 olefin and

monocycloparaffin isomers. It should be noted that the C10 case was

analyzed on a separate day which resulted in a slight variation in the

"FUEL ONLY" results. The table also compares these results to the

theoretical content expected. The results of the unsaturated

hydrocarbon types are not shown since changes were incurred only

through dilution. In all cases, the monocycloparaffin demonstrated a

larger increase in the monocycloparaffin content than the olefin.

Examination of the paraffin and dicycloparaffin content after

perturbation of the fuels demonstrates some interesting results which

can be evaluated in terms of the differences in the mass spectra of

each and reference to the characteristic ion series for each

hydrocarbon type (Table 2.1). Figure 3.14 shows the mass spectra of

the C7 isomers cycloheptane and 1-heptene. A larger relative abundance

of molecular ion (m/z 98) and [M-15]+ (m/z 83) was obtained for the

cycloheptane but produced little differences between the GC/MS/HTA

results. Figure 3.15 shows the mass spectra of the C9 isomers














Table 3.14


Comparison of
After Spiking with C7, C9,
(in


HTA of JP-4 2455 Jet Fuel
and C10 Cycloalkane and Olefin Isomers
Volume Percent)


HYDROCARBON
TYPE


FUEL ONLY
500 uL


Fuel+50 yL
cycloheptane


Fuel+50 iL
1-heptene


PARAFFIN 61.5 57.9 59.9 55.9


MONOCYCLO- 18.5 24.8 22.2 25.9
PARAFFIN

DICYCLO- 3.1 1.1 1.8 2.8
PARAFFIN


HYDROCARBON FUEL ONLY Fuel+50 tL Fuel+50 yL Expected
TYPE 500 pL propylcyclohexane 1-nonene Result


PARAFFIN 61.5 55.1 61.4 55.9


MONOCYCLO- 18.5 24.5 22.3 25.9
PARAFFIN

DICYCLO- 3.1 6.4 1.4 2.8
PARAFFIN


HYDROCARBON FUEL ONLY Fuel+50 iL iso- Fuel+50 pL Expected
TYPE 500 pL butylcyclohexane 1-decene Result


PARAFFIN


MONOCYCLO-
PARAFFIN

DICYCLO-
PARAFFIN


58.8


17.8


52.2


27.4


58.3


21.4


53.4


25.3


3.6


4.0


Expected
Result





























Co


CL


0
-4




0






(.-
0
U







04
U,
o4
0











r.a
o


















0
o-









M(


0



U


oupunq eA!el %c

eouspunqV SA!JBIGHb %















































w -
CM --


C"T






1)




n ~~.
rto


n-
w
(n_
cr,


cnc
m




te-=
ta


In

u,


C.)

o cn
m,
a -,,


eouepunqvy eA!le /1 %


. i | I


. i I


I1S

I9
Mo









76

isopropylcyclohexane and l-nonene. Again a larger relative abundance

of molecular ion (m/z 126) was obtained for the cycloparaffin, but also

the fragmentation pattern was quite different than that of the olefin.

The large relative abundances of the m/z 82 fragment (96%) and the m/z

67 fragment (42%) for the cycloparaffin isomer, which were very low for

the olefin (9%), caused the dicycloparaffin volume percent of the

cycloparaffin-spiked fuel determined by HTA to be high. Similarly the

large relative abundances of paraffinic fragment ions (m/z 43, 57, and

71) for the olefin isomer caused the paraffin volume percent of the

olefin-spiked fuel to be well above that of the cycloparaffin-spiked

fuel. Figure 3.16 shows the mass spectra of the C10 isomers

isobutylcyclohexane and 1-decene. Again a larger relative abundance of

molecular ion (m/z 140) was obtained for the cycloparaffin. The

fragmentation patterns, although different, both contributed response

to the dicycloparaffin volume percent of the spiked fuels. The

slightly higher volume percent calculated for the cycloparaffin-spiked

fuel was due to the larger relative abundance of m/z 82 fragment. The

paraffinic volume percent was again larger for the olefin spiked fuel

because of the large relative abundances of the paraffinic fragment

ions, especially m/z 43.

These studies validate the need for separate determinations of

cycloparaffin and olefin content in the HTA of jet fuels. New

complications have been introduced as to the effects on the HTA

calculated paraffin and dicycloparaffin content. Examination of the

differences in the mass spectra of the isomeric monocycloalkanes and







































































N


U,____________ __
co-








LO



=


I.





co


......
r -




i_-


eouepunqv eaeIJleB %


0
so



u

0 c





Eu
o X
So


. I I I I


I- '- i I a I









78

olefins of the three carbon numbers studied suggest the possibility of

developing techniques which could identify and quantitate each isomer

separately in the presence of the other.

Evaluation of Methods for Average Carbon Number Determinations

The GC/MS/HTA method relies heavily on the accuracy of the average

carbon number calculations, since the inverse calibration matrices

chosen are dependent upon these results. As in the ASTM methods, these

average carbon numbers (aromatic and paraffinic) are determined based

on the intensities of the paraffin and alkylbenzene molecular and

fragment ion series such as [M-15]+. This has raised a certain degree

of speculation concerning the accuracy of this method. First, as

mentioned, the method is based on the intensities of molecular and

fragment ions, which are generally of very low relative abundance.

This is particularly true for paraffins, whose molecular ion relative

abundances are on the order of 1-5%. Alkylbenzenes, however have

relatively abundant molecular ions, on the order of 20-50%. Therefore

less error would be expected for the aromatic results. Second, the

volume percent results of the remaining hydrocarbon types are dependent

upon the average carbon number results of the paraffins and

alkylbenzenes. For example, the dicycloparaffin volume percent is

calculated based on the inverse calibration matrix determined by the

average paraffin carbon number. This would obviously cause even

greater error in the results for those samples which contain small

amounts of paraffins relative to the dicycloparaffins. To clarify some

of these uncertainties, a new method of calculating the average carbon

numbers of a jet fuel was developed.









79

As was -shown in Figure 2.2, the GC/MS simulated distillation

(GC/MS/SD) method determines the centroid of the total area of a

chromatographic profile. The GC/MS facilitates the ability to profile

an ion series characteristic of a certain hydrocarbon type with an ion

summation chromatogram, which may then be compared to a similar

chromatogram of a homologous series of that hydrocarbon type. Figure

3.17 again illustrates this method, only in this case showing the

paraffin series (243) of the JP-8X fuel with the corresponding n-

paraffin homologous series in parallel. Several variations in this

technique were used for the determination of the average carbon

numbers. These variations were obtained by variations in the

particular set of ions used in the ion summation chromatograms. Table

3.15 shows the results of these carbon number determinations for

several jet fuels. From this table, three comparisons can be made for

both aromatic and paraffinic average carbon numbers: comparison between

the two GC/MS/SD methods (which will be explained below); comparison

between the two GC/MS/SD methods and the GC/MS/HTA method; and

comparison between the GC/MS/HTA and the ASTM method.

GC/MS Simulated Distillation

Table 3.15 provides data for the comparison of the two GC/MS/SD

methods, SIM DIST #1 and SIM DIST #2 for both aromatic average carbon

numbers and paraffinic average carbon numbers. The comparisons

demonstrate the variations which were due to the characteristic set of

ions used in the calculations as discussed above. The methods used for

the two GC/MS/SD average carbon numbers determinations were as follows:

the SIM DIST #1 method calculates the paraffinic carbon number based








































paraffin homologous series


PIc j


0 CD


~C3
Cl~l (::1. i


ieo 28 308 4e 5'00 SCAN
1:40 3:29 5:80 :49 9:;0 TIME


Figure 3.17.


The paraffinic (Z43) ion summation chromatogram of jet
fuel JP-8X 2414 on a 3 meter DB5 column. The centroid of
the total area, indicated by the arrows, is referenced to
the n-paraffin homologous series. The average paraffinic
carbon number is calculated to be 12.89.



































































































o3

u z o
Cd Z :: 0

0

I-. Ld
-) on AM P-)
, ,

0.4Sd
I~q p~


w cn

( 0
e-4
zC

3 -4
SE-




u co
0 0



0





CB)
w u




-4Eo


-4










o44

CO b
0 -4

(0
SL3





E0

U


4

-r -4


1-4



00 00

CO
o EE
C 0

Uo

CC
0 0
-4 -4




-.i Cc c

-4 00











o C
C a














C W
c 3





0 0 -
u Cu


































0m 0 0
m c
a 01 0
C w>











a)
-4
o 04-J








ca 3.0






CUda1
i- -i C1
<; < *^






a; r sa *
S1 rL 1 UU
M h Cfl 4
(!< -)
(B C)Q Z-









82

solely on the ion summation chromatogram of the E43 ions (paraffin),

and the aromatic carbon number based on the ion summation chromatogram

of the 277 ions (alkylbenzene); the SIM DIST #2 method calculates the

paraffinic carbon numbers based on the ion summation chromatogram of

all the saturated ion sets, Z43, Z41, and 267, (paraffin,

monocycloparaffin, and dicycloparaffin), and the aromatic carbon

numbers based on the ion summation chromatograms of all the unsaturated

ion sets, Z77, 2103, Z128, (alkylbenzene, indane/tetralin, and

naphthalene). To make a proper evaluation of these results, reference

must be made to the composition of these jet fuels, as shown in Table

3.4. For those fuels whose saturated and unsaturated compositions were

made up mainly of paraffins and alkylbenzenes, respectively, there were

only small differences between the average carbon numbers obtained by

two techniques. For example, the JP-4 2455 jet fuel has a saturated

fraction composed mainly of paraffin, which results in little

difference between the average paraffinic carbon numbers from the two

techniques, 12.1 and 12.0. The JP-8X 2429 jet fuel, on the contrary,

has a saturated fraction composed mainly of dicycloparaffin which has

resulted in a large difference between the average paraffinic carbon

numbers of the two techniques, 14 and 11.1. For this fuel the same

trend was observed for the average aromatic carbon numbers, 12.1 and

10.7, since the unsaturated fraction was composed mainly of

indane/tetralin rather than alkylbenzenes. In summary these two

techniques of average carbon numbers determinations demonstrate a large

dependency upon the composition of the jet fuel, which demonstrates the

failures of the present ASTM method and the need for more specific









83

methods for carbon number determinations of each particular hydrocarbon

type. These conclusions are further supported by the second set of

comparisons between the GC/MS/SD and GC/MS/HTA results.

Comparison Between GC/MS/SD and GC/MS/HTA

Table 3.15 provides data for the comparison of the GC/MS/SD and

GC/MS/HTA methods of average carbon number determination. The original

purpose of the development of the GC/MS/SD method was to demonstrate

the dependence of the average carbon number results on the sample

concentration, and the limitations of calculating the paraffinic

(saturated) and aromatic (unsaturated) average carbon number based on

the paraffin and alkylbenzene composition only (both discussed above).

The large differences between the GC/MS/SD and GC/MS/HTA results were

not expected. In all cases a larger average carbon number (sometimes

larger by 3 or 4 carbons) was obtained for the GC/MS/SD results than

for the GC/MS/HTA results. This raised uncertainty in the results of

the GC/MS/SD technique, although the simulated distillation methods

seemed scientifically sound and much more tangible than the mass

spectral technique of the GC/MS/HTA method discussed above.

Consultation with the engineers at the Pratt and Whitney Fuels

Division, who have analyzed many jet fuels for average boiling point by

the ASTM simulated distillation method (44,45), suggested a consistent

underestimation of the average carbon numbers by the ASTM D-2789

method.

Comparison Between GC/MS/HTA and ASTM

The final set of comparisons made in Table 3.15 were between the

GC/MS/HTA and ASTM D-2789 results. These results were determined based









84

on the same computational technique as discussed at the beginning of

this section. Large deviations between the results of the two

techniques were apparent for most fuels, with the GC/MS/HTA values

generally larger. The results of jet fuel JP-8X 2429 showed the best

paraffinic and aromatic average carbon number agreement, while jet fuel

JP-4 2455 showed the worst. The measure of paraffinic carbon number

agreement may be correlated to the saturate composition of the fuels

(see Table 3.4), since the JP-8X fuels (such as 2429 and 2414) were

composed mainly of dicycloparaffins, and the JP-4 fuel was composed

mainly of paraffins. These correlations are of course dependent upon

the accuracy of the volume percent results, which are themselves

dependent upon the average carbon number results. No correlation

between the composition of the jet fuels and the aromatic carbon

numbers could be determined. However, agreement between the two

methods was generally better for the aromatic carbon numbers than for

the paraffinic carbon numbers. This may be due to the increased

accuracy of the technique as reflected in the larger relative abundance

of the alkylbenzene molecular ions, as discussed above.

The presence of large amounts of naphthalenes in jet fuels can

cause a large amount of error in the paraffinic average carbon number

and ultimately in the final volume percent results (6), since the m/z

128, m/z 142, and m/z 156 ions used in the ASTM average paraffinic

carbon number determination correspond to abundant molecular ions of

naphthalene, methylnaphthalene, and dimethylnaphthalene, respectively.

Investigations of the effect of this perturbation were made by spiking

the JP-4 2455 jet fuel with dimethylnaphthalene. Table 3.16


















Table 3.16

Comparison of GC/MS/HTA of JP4 Fuel
After Spiking with Dimethylnaphthalene
in Volume Percent


HYDROCARBON
TYPE


FUEL ONLY
500 AL


Fuel+100 iL
Dimethylnaphthalene


PARAFFIN


MONOCYCLO-
PARAFFIN

DICYCLO-
PARAFFIN

ALKYLBENZENE


INDANE/
TETRALIN

NAPHTHALENE


AVERAGE CARBON
NUMBER

AROMATIC


9.50 10.58


61.5


18.5


Expected
Result


43.9


18.0


3.1


51.3


15.4


14.8


11.4


2.0


0.0


12.3


24.5


16.7


9.06


8.91


9.55


9.50


PARAFFIN









86

illustrates the results of this perturbation of the jet fuel. The

drastic increase in the paraffinic average carbon number from 9.50 to

10.58 reflects the effect of the addition of the C12

dimethylnaphthalene. An increase in the aromatic average carbon number

was expected since dimethylnaphthalene is a C12 aromatic; however the

results actually decreased. This is due in fact to the average

aromatic carbon number calculational method which is based solely on

ions of the alkylbenzene series. The mass spectra of

dimethylnaphthalene contains a small relative abundance of m/z 91,

which is an ion of the alkylbenzene series, which explains the decrease

in aromatic average carbon number. The saturate volume percent results

of the spiked fuel did not agree with changes expected due simply to

dilution. This may be attributed to the incorrect average paraffinic

carbon number on which the volume percent calculation is now based. It

should be noted that the naphthalene volume percent increased to 24.5%

for the spiked fuel which was close to the expected 16.7% theoretical

value. These results illustrate the significance of the naphthalene

content on the hydrocarbon type analysis results, not only on

paraffinic carbon numbers, but on volume percent results as well.

Effect of Chromatographic Resolution on Carbon Number Calculations

The evaluations and comparisons above were made under the

assumption that the carbon number calculational techniques produced

valid average carbon number results. Since chromatographic separation

played such an important role in these determinations, especially for

the GC/MS/SD techniques, it was desired to determine the effect of

increased chromatographic resolution on the average carbon number









87

results. Table 3.17 shows the effect of chromatographic separation on

the average carbon number results for the four calculational techniques

and two separate jet fuels, JP-4 2455 and JP-8X 2429. The

chromatographic resolution was increased by changing from a 3 meter

column to a 30 meter column. Both fuels demonstrated a dependence on

chromatographic separation for all carbon number calculational

techniques. In all cases the determination of carbon numbers with

increased chromatographic separation, with the exception of the

aromatic average carbon number for JP-8X using the SIM DIST #2 method,

better agreement was obtained between the GC/MS/DS and the GC/MS/HTA

results, and between the GC/MS/HTA and ASTM results. Increased

chromatographic separation reduced all average carbon numbers for the

JP-4 fuel. Increased chromatographic separation decreased all average

carbon numbers for the JP-8X fuel with the exception of the paraffinic

average carbon number using the SIM DIST #1 method. These results

indicate that the average carbon number results are very much dependent

upon the degree of chromatographic separation employed, and that

increased chromatographic resolution results in more accurate average

carbon number determinations.



Summary of Results

An extensive amount of effort was dedicated to characterization of

the GC/MS/HTA method. The injection methods were evaluated in terms of

component discrimination which resulted in the development of a non-

discriminating injection technique. The method was evaluated in terms

of sensitivity and precision. Similar sensitivities were determined to












88



o -4 ,



a 0 0 0'-
,mv v v




SE n 0o Co


a cr o 0 0
o





wo 0 0



0 .w 0
SE- 0





-4~
0 B r v r 0 0
i-4 v 0) a\ CM 00 (5





N co -oo








0 a 04
4 CO E

,0 0 0 u
0N


c :<3 E, c a
n ?N N cu Mo *
cl E Mrr r- O cN


4 00 0
4) .1 CU





ni3 0, --C
4 a0 >-, c ,o
c 5 o1 4) cw -4













0i c0
4 4. L" w 41


So o v cb U ca
0 co r_ u^^




o 4 U a o

, (4











0 4.JC
pr -, 0









89

those calculated in the ASTM technique. The precision was dependent on

hydrocarbon type and volume percent.

To characterize and validate the GC/MS/HTA method, a series of

experiments were performed. Comparisons of GC/MS/HTA results with the

ASTM results were inconsistent and followed no trends. Evaluations of

GC/MS/HTA results of synthetic mixtures showed good agreement in some

cases but were still inconsistent. The analysis of simple mixtures

seemed to validate the Teeter Rule, although excellent agreement was

often present. The perturbation fuel analysis introduced a new method

for the characterization of HTA methods. The results obtained here

were consistent with the expected results and represented a significant

contribution towards validation of the method.

A new method for carbon number determination was developed based on

GC/MS/simulated distillation. A critical evaluation of this method,

the GC/MS/HTA method, and the ASTM method was performed. The results

indicated that the GC/MS/HTA and ASTM method underestimate the average

carbon numbers which have a significant effect on the volume percent

results of HTA methods. Chromatographic resolution affected the

GC/MS/SD determination of average carbon numbers. The results

indicated better agreement with mass spectral methods upon increased

chromatographic resolution.

Finally, as an intricate part of this project, the developed

GC/MS/HTA method was documented in the standard ASTM format which was

to be utilized by the Pratt and Whitney engineers of the Fuels

Division. A duplication of the ASTM standardized GC/MS/HTA method is

shown in Appendix D.


















CHAPTER IV

PRELIMINARY INVESTIGATIONS OF ALTERNATIVE IONIZATION
TECHNIQUES FOR JET FUEL ANALYSIS


The primary goal of this area of study was to investigate

alternative ionization techniques in mass spectrometry for the analysis

of jet fuels. These studies will cover aspects of both hydrocarbon

type analysis and total fuel analysis. The combination of high

resolution gas chromatography and alternative ionization techniques

such as low-energy electron ionization and methane chemical ionization,

will be a powerful tool for the determination of specific components in

jet fuels.



Background

Low Energy Electron Ionization

The main effort in the application of low energy electron

ionization GC/MS is to evaluate a technique for the specific detection

of aromatic type compounds in jet fuels. Studies have been reported

(31,53,59) on the determination of specific aromatic components in

fuels and coal-derived liquids by such methods. Aromatic compounds

have a lower ionization potential than aliphatic molecules which allows

for the selective ionization of aromatics in the presence of aliphatics

(53).









91

In the present study the utilization of GC separation, combined

with the specific detection of aromatics, is demonstrated as a

potential method for the determination of specific aromatic components

in jet fuels. A method such as this would contribute to the

development of a total fuel analysis method which was previously

discussed.

Methane Chemical Ionization

The main effort in the application of methane chemical ionization

(CI) GC/MS to the analysis of jet fuels was to demonstrate the

potential of the technique as a new method for hydrocarbon type

analysis. Much work has been reported on the analysis of fuels and

other coal derived liquids using CI/MS (53-58) for the determination of

specific fuel components and specific classes of components. With

chemical ionization, a large array of specific ionization schemes are

available by choosing an appropriate CI reagent gas. Harrison (52) in

his book on chemical ionization mass spectrometry, discusses in detail

the fundamentals, instrumentation, and applications of CI/MS. In

general terms, the ionization specificity in CI/MS is obtained by the

specific reactions and energetic encountered by the ionized reagent

compound and the analyte. For example, in positive methane CI/MS

(which is used in this study), a methane molecule is first ionized to

CH4+ (and CH3+, CH2+, etc.) by electron impact ionization. Collisions

of these ions with another methane molecule leads to proton transfer to

form CH5+. This process continues to form higher m/z ions such as

C2H5+. Only those analytes with a higher proton affinity than CH4 and

C2H4 (the conjugate bases of CH5+ and the C2H5+ ions) will be ionized.









92

The application of methane CI/MS will be demonstrated as a

potential new method of hydrocarbon type analysis of jet fuel. The

results are only preliminary and demonstrate the need for continued

experimentation. An additional potential for CI/MS is combined gas

chromatography CI/MS. This would enable the determination of specific

components of specific classes which would contribute to the total fuel

analysis effort.



Experimental

All instrumental parameters were unchanged unless otherwise

specified in this section.

Electron Energy Variation

The electron energy variation was obtained by manual adjustment of

the electron energy control knob. The electron current was maintained

at 300 iA which limited the lowest electron energy to 11.7 eV. The

high end of the electron energy was 100 eV.

The sample analyzed was a 50/50 mixture of the Aromatic 150 and

Isopar M mixtures. Analysis of the sample at each electron energy was

performed separately after adjustment to the proper energy.

Methane CI

Methane CI was performed at an electron energy of 100 eV and a

methane source pressure between 0.2 and 0.3 Torr. The mass

spectrometer was scanned from 65 to 650 amu with a total scan time of

1.0 s. The samples were analyzed by separate injections of pure

components diluted in hexane.