The development and evaluation of constant energy synchronous luminescence techniques for PAH mixture analysis

MISSING IMAGE

Material Information

Title:
The development and evaluation of constant energy synchronous luminescence techniques for PAH mixture analysis
Physical Description:
viii, 178 leaves : ill. ; 28 cm.
Language:
English
Creator:
Kerkhoff, M. Jonell, 1954-
Publication Date:

Subjects

Subjects / Keywords:
Hydrocarbons -- Spectra   ( lcsh )
Liquid chromatography   ( lcsh )
Spectrum analysis   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1984.
Bibliography:
Includes bibliographical references (leaves 172-177).
Statement of Responsibility:
by M. Jonell Kerkhoff.
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 - 000496909
notis - ACR6136
oclc - 12042141
System ID:
AA00003416:00001

Full Text













THE DEVELOPMENT AND EVALUATION OF CONSTANT
ENERGY SYNCHRONOUS LUMINESCENCE TECHNIQUES FOR
PAH MIXTURE ANALYSIS









By

M. JONELL KERKHOFF


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


1984
















ACKNOWLEDGEMENTS


I would like to express my deep appreciation and

gratitude to Dr. James D. Winefordner for his encouragement,

knowledgeable advice, and interest throughout my stay at the

University of Florida. This dissertation would not have

been possible without his patience and kindness. I would

like to thank Dr. Ben Smith, Dr. Ed Voigtman, and Dr. Lucas

Hart for their invaluable help and input. A special thanks

goes to Dr. Ed Voigtman who continues to serve as a role

model as he strives to set and attain higher scientific

goals. A special thanks goes to Dr. Lucas Hart for his

advice on the computer projects, and his support and friend-

ship.

My thanks go to Terrie Lee and Dr. Eric Allen, Depart-

ment of Environmental Engineering and Sciences, for their

interactions on our joint project. I would like to thank

Leigh Ann Files and Rik Faith for all of their help and time

spent with this work. I would like to thank Dr. John Dorsey

and Dr. Rick Yost for all of their advice and interest.

Thanks go to Jeanne Karably for her assistance throughout

the years. To Dr. Leslie Oliver for his concern and sup-

port, thank you.









To all my friends and colleagues, especially Barbara

Kirsch, Leigh Ann Files, Melanie Elder, Linda Hirschy, and

Jodie Johnson, thank you for filling the last five years

with good times and memories.

A very special thank you goes to my parents and sisters

for their total support and love, and for always being

there.

















TABLE OF CONTENTS


ACKNOWLEDGEMENTS . .

ABSTRACT . .

CHAPTER


INTRODUCTION . .

Polyaromatic Hydrocarbons .
Analysis of PAHs . .
Goals of Work . .

THEORETICAL CONSIDERATIONS .

Molecular Luminescence . .
Synchronous Luminescence . .
Constant Energy Synchronous Luminescence .

RAPID SCANNING CONSTANT ENERGY SYNCHRONOUS
FLUORESCENCE SPECTROSCOPY AT 20 nm/s .

Introduction . .
Experimental . .

Reagents ... .
Instrumentation ..
Methodology . .

Results and Discussion . .

Scan Mechanics .
Flow Cell Comparison .
Selection of Av . .
Conclusion . .


1


Page

S ii

. vii



1

1
S 6
S 8

S 11

S 11
21
S 26


S 34

S 34
S 36

S 36
S 36
S 41

S 42

S 45
S 48
S 51
S 63


. .









CHAPTER


4 RAPID SCANNING CONSTANT ENERGY SYNCHRONOUS
FLUORESCENCE SPECTROSCOPY AT 200 nm/s 64

Introduction . .. 64
Experimental ... 69

Instrumentation . 69
Methodology .. . .. 80

System Evaluation . 80

Stop Flow vs Continuous Flow .. 80
Analytical Figure of Merit 85
Evaluation of the Data Processing 86
Comparison of the PMT and the SIT for
Synchronous Scanning . 94

5 RAPID SCANNING CONSTANT ENERGY SYNCHRONOUS
FLUORESCENCE SPECTROSCOPY AT 200 nm/s AS
A LIQUID CHROMATOGRAPHY DETECTOR .. 103

Introduction . .. 103
Experimental . . 104

Apparatus . 104
Reagents . 104
Methodology . 104

Results and Discussion .. .. 105

Analysis by RSCESFS at 200 nm/s 105
RSCESFS at 200 nm/s vs UV Absorbance 108
RSCESFS at 200 nm/s vs Variable
Wavelength Fluorescence 115
Conclusion . 116

6 SPECTRAL FINGERPRINTING OF POLYAROMATIC
HYDROCARBONS IN HIGH-VOLUME AMBIENT
AIR COLLECTIONS BY CONSTANT ENERGY
SYNCHRONOUS LUMINESCENCE SPECTROSCOPY 118

Introduction . ... .118
Experimental .. .. 120

Instrumentation .. .120
Methodology .. .. ... 120
Sampling System and Methodology 123


Page









CHAPTER Page

Results and Discussion . 126

CESLS Spectral Fingerprinting 126
GFF Extracts . 128
PUF Extracts. . 133
Semiquantitative Analysis .. 136

7 ANALYSIS OF PAH MIXTURES BY LOW TEMPERATURE
CONSTANT ENERGY SYNCHRONOUS FLUORESCENCE
SPECTROSCOPY . 143

Introduction. . 143
Experimental . 145

Instrumentation . 145
Methodology .. .. 145

Results and Discussion .. .. 146

Enhancement of Spectral
Resolution by LTCESFS ... .146
Mixture Analysis by LTCESFS 148

8 CONCLUSION . . 166

Summary . 166
Future Work . .. 170

REFERENCES ...... ......... 172


BIOGRAPHICAL SKETCH ........ .


. 178
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE DEVELOPMENT AND EVALUATION OF CONSTANT
ENERGY SYNCHRONOUS LUMINESCENCE TECHNIQUES FOR
PAH MIXTURE ANALYSIS

By

M. Jonell Kerkhoff

December 1984

Chairman: James D. Winefordner
Major Department: Chemistry

Constant energy synchronous luminescence (CESLS)

scanning techniques are viable methods of analysis for

polyaromatic hydrocarbon (PAH) mixtures because of the

improvement in spectral resolution and reduction of Raman

and Rayleigh scatter obtained. CESLS was applied in two

ways: 1) a constant energy difference was selected to give

spectra for the PAHs separated by liquid chromatography

(HPLC) and 2) constant energy differences were chosen to

promote wavelength separation in PAH mixtures.

A rapid scanning fluorescence system at 200 nm/s based
-1
on constant energy synchronous scanning at 4800 cm was

built for continuous monitoring of chromatographic eluants.

A 200 nm spectral resolution wavelength range (250 nm to

450 nm) was covered in 1 s time period with data collection

in the forward and reverse directions. The area of each









RSCESFS spectrum was integrated. The integrated value was

output in real time to reconstruct the chromatogram. A

RSCESFS spectrum was saved for each chromatographic peak.

Chromatographic separation and RSCESFS analysis for 3 com-

ponent and 9 component mixtures were run. Detection limits

for 17 PAHs were determined. Combining HPLC with RSCESFS

resulted in the following advantages: 1) separation of PAH

mixtures by HPLC, 2) detection limits in the parts per

billion range, and 3) RSCESFS spectra used in conjunction

with retention times for identification.

CESLS techniques promoted spectral wavelength sepa-

ration and identification of PAHs in mixtures at room and

low temperatures. CESLS was adapted for screening and

assessing PAHs in ambient air collections received from the

Environmental Engineering Sciences Department. Spectral

fingerprints of particulate phase PAHs and gas phase PAHs

were obtained at Av values of 200 cm 1, 1400 cm and 4800
-1
cm Five major PAHs in the gas phase extracts were

identified and semiquantitated.

A significant enhancement in spectral resolution was

obtained with low temperature constant energy synchronous

fluorescence (LTCESFS). Further spectral separation of

peaks and reduction of peak halfwidths were achieved.

Isomeric PAH mixtures and alkyl homologue mixtures were

analyzed at Av values from 1400 cm to 2000 cm-. LTCESFS

provided a simple, selective, and inexpensive method for PAH

evaluation in mixtures.
















CHAPTER 1

INTRODUCTION



Polyaromatic Hydrocarbons

Polyaromatic hydrocarbons (PAHs) are environmental

pollutants. They are stable products of natural and anthro-

pogenic sources formed from their organic precursors upon

exposure to high temperatures. PAHs can also form over time

at lower temperatures, e.g. the PAHs found in crude oils.

As a result of biosynthesis, natural combustion, and long

term degradation, PAHs occur naturally in the environment.

The anthropogenic PAHs are the major contributors of

environmentally hazardous compounds. Examples of man-made

sources are the coking of coal, incomplete combustion and

emission of PAHs in furnaces and automobile exhaust, and

tobacco smoke. According to Blumer, a reasonable guess as

to the PAH source can be made since the higher the trans-

formation temperature of the organic precursor to aromatic

process, the more likely unsubstituted aromatics (without

alkyl substituent side chains) occur (1). The majority

of PAHs found in air pollution from the high temperature

processes of coal coking and engine combustion are unsub-

stituted PAHs while crude oil samplings contain more complex

mixtures of alkyl PAHs.









This work is directed towards those PAHs found in

environmental pollution. A spectral data base of 17 PAHs

was compiled in this study. These structures are shown in

Figure 1 with the corresponding molecular weights, chemical

formulas, and carcinogenic activity levels tabulated in

Table 1. Of the 17 PAHs in Figure 1, 15 are listed as EPA

priority pollutants for monitoring of water samples (2,3).

PAHs consist of two or more fused benzene rings. The

rings can be arranged in a linear, angular, or cluster

fashion. A cluster arrangement differs from an angular in

that one benzene ring in the cluster is surrounded on three

sides. The linear rings are the least stable and the

angular rings are the most stable. The cluster PAHs are

most commonly found in soil, tobacco smoke, and petroleum

and automobile emissions. The cluster PAHs show the most

carcinogenic activity.

The arrangement and number of benzene rings in each PAH

as well as the substitution position of the alkyl homologues

can greatly affect the carcinogenicity (1,4,5). Carcino-

genic activity is found generally in PAHs with 3 to 6 ben-

zene rings as shown in Table 1. PAHs demonstrating strong

carcinogenic activity are benzo(a)pyrene, benzo(b)fluoran-

thene, benzo(k)fluoranthene, and dibenz(a,h)anthracene.

Structural isomers of the PAHs in Table 1 occur at

molecular weights of 178 (C14H10), 202 (C18H10), 228

(C18H12), 252 (C20H12) and 276 (C22H12). A well-known

isomer pair is that of benzo(a)pyrene, which shows strong





























ra



a)




rl

41-
To
w
En



En







TO



0
En










4J
En









"-I
U)










.14
4

a








04
r4




0
u
w













412
w




















0'
tar
-1

FZ4























4 \z u

z z


z z














u /z
oz
I M O4




Iz z




















44 J--i
E-r 44








Table 1. Molecular Weights, Chemical Formulas, and
Carcinogenic Activity Levels for 17 PAHs


Molecular Chemical Carcinogenic
Compound Weight Formula Activity


naphthalene 128 C10H8 0

acenaphthene 154 C12H10 0

fluorene 166 C13H10 0

anthracene 178 C14H10 0

phenanthrene 178 C14H10 0

fluoranthene 202 C18H10 0

pyrene 202 C18H10 0

benz(a)anthracene 228 C18H12 +

chrysene 228 C18H12 +

benzo(a)pyrene 252 C20H12 ++

benzo(e)pyrene 252 C20H12 0/+

benzo(b)fluoranthene 252 C20H12 ++

benzo(k)fluoranthene 252 C20H12 ++

perylene 252 C20H12 0

benzo(g,h,i)perylene 276 C22H12 +

indeno(1,2,3-c,d)pyrene 276 C22H12 +

dibenz(a,h)anthracene 278 C22H14 ++


a. 17 PAHs, 15 of which are EPA priority pollutants in
water (2,3).

b. Carcinogenic activity determined by the ability of PAHs
to induce cancer in rats: (0) noncarcinogenic, (+)
weakly carcinogenic, and (++) strongly carcinogenic (4).









carcinogenic activity and benzo(e)pyrene, which shows little

carcinogenic activity. The structural isomers most diffi-

cult to resolve and separate are the compounds with molec-

ular weights of 228: benz(a)anthracene and chrysene; of

252: benzo(a)pyrene, benzo(e)pyrene, benzo(b)fluoranthene,

and benzo(k)fluoranthene; of 276: benzo(g,h,i)perylene and

indeno(1,2,3-c,d)pyrene.



Analysis of PAHs

Analytical techniques for multicomponent PAH analysis

must be capable of identifying and quantitating individual

components at the parts per million or parts per billion

level. Lee, Novotny, and Bartle (4) have recently reviewed

analytical methods of analysis for complex PAH mixtures.

For the most part, PAHs are measured by gas chromatography/

mass spectrometry, liquid chromatography with UV absorption

or fluorescence detection methods, and luminescence tech-

niques. The luminescence methods with and without liquid

chromatography for PAH analysis will be briefly reviewed.

Recent advances in multicomponent luminescence tech-

niques have been discussed in books edited by Wehry (6,7,8)

and Eastwood (9). Because PAHs exhibit strong absorption

and fluorescence transitions fluorescence analysis of PAHs

is widely used. Spectra obtained by conventional lumin-

escence techniques are generally broad, hindering the

identification of PAHs in mixtures. Quenchofluorometry has

been used to quench selectively PAH in mixtures. Through









quenching with oxygen individual components in benz(a)-

anthracene/chrysene/naphthacene mixtures were determined

(10). A solution 10% by volume of iodomethane was used to

promote phosphorescence of fluoranthene (11).

In order to reduce the complexity of spectra which are

normally measured at room temperature, low temperature

techniques such as laser excited Shpol'skii spectrometry

(12), fluorescence line narrowing (13), and matrix isolation

spectrometry (14,15) have been developed. Shpol'skii

fluorometry has been applied to coal and coal tar pitch

extracts (16) airplane engine emissions (17) and liquid

chromatography fractions of automobile exhaust (18). Matrix

isolation fluorometry has been used to analyze PAHs in steel

mill coking plant water (19). Fluorescence line narrowing

or site selection spectroscopy has been applied to synthetic

PAH mixtures (13). PAHs have been examined by supersonic

jet molecular spectroscopy (20).

PAH analysis of coal tar has been obtained with deriva-

tive and selective modulation fluorometry (21). Modulation

fluorometry has been used for the analysis of PAHs in air

particulates fractioned by liquid chromatography (22).

A number of authors have combined liquid chromatography

with fluorescence analysis for PAH detection (22-26).

Liquid chromatography effluents have been analyzed by

a three mode detection scheme measuring fluorescence,

photoionization, and photoacoustic signals (27-29). A

laser-induced fluorescence detector has been discussed (30)









for PAH analysis. Liquid chromatography fractions from

carbon black samples were analyzed by mass spectrometry and

fluorescence (31).

Total luminescence spectra of PAH mixtures have been

obtained on oil samples (32,33). Synchronous luminescence

has been applied to PAH analysis of automobile oils, air

water and coal liquid samples (34-37). Eastwood has

reviewed luminescence techniques for oil identification

(38).



Goals of Work

The goal of this work was to develop and evaluate

constant energy synchronous luminescence techniques for

multicomponent PAH analysis. Constant energy synchronous

luminescence spectroscopy (CESLS) developed by Inman and

Winefordner, is a simple, inexpensive, sensitive, and

selective luminescence technique (39). Luminescence is

measured while synchronously scanning both monochromators

and maintaining a constant energy difference between the

excitation and emission monochromators. CESLS offers a

selectivity improvement through the reduction of spectral

complexity and spectral interference as well as a gain in

spectral resolution while maintaining the sensitivity

advantage of conventional fluorescence scanning methods

(39,40). Thus, CESLS scanning results in a spectrum of

reduced spectral complexity with comparable sensitivity

found in conventional fluorescence measurements obtained









at the excitation and emission wavelengths of maximum

intensity for each PAH.

A major part of this research consisted of the devel-

opment of a rapid scanning constant energy synchronous

fluorescence spectrometer (RSCESFS) and its subsequent

application as a liquid chromatography (HPLC) detector for

PAH analysis.

A RSCESFS system with a 20 nm/s scan rate was developed

and is compared with a conventional fluorescence system in

Chapter 3. The choice of constant energy difference optimal

for analysis of seventeen PAHs was evaluated. RSCESFS at

20 nm/s at the optimal constant energy difference compared

favorably to conventional fluorescence measurements obtained

at the wavelengths of maximum intensity for each PAH.

RSCESFS limits of detection ranged from 700 fg for perylene

to 200 pg for phenanthrene. The majority of the PAHs were

in the low pg range. Furthermore, spectra obtained for the

seventeen PAHs at the optimal constant energy difference

resulted in reduction of spectral complexity over conven-

tional fluorescence scanning and provided a means to iden-

tify each PAH.

A faster RSCESFS system was necessary to provide

continuous, on-line monitoring for HPLC. A 200 nm/s system

was developed and evaluated for figures of merit in

Chapter 4. A discussion of the programming and hardware

requirements as well as the signal processing has been

given.









With the RSCESFS detection system set up, mixtures of

PAHs were separated by HPLC and analyzed with the constant

energy synchronous system. Analytical figures of merit for

15 PAHs are given in Chapter 5. Limits of detection ranged

from 200 pg for perylene to 70 ng for phenanthrene. The

majority of PAHs were in the pg to low ng range.

A high-volume sampling technique was combined with

CESLS to obtain spectral fingerprints of PAHs in ambient air

collections and is discussed in Chapter 6. Spectral finger-

prints were obtained on samplings of gasoline engine

exhaust, diesel engine exhaust, air from a heavily-travelled

interstate site, and air from a moderate to light urban

site. PAHs present in the gas and particulate phases of the

ambient air were collected. With various constant energy

differences employed, spectra were obtained for the four

samples in both phases. Five major PAHs in the gas phase

extracts were identified and estimated.

A further improvement in spectral resolution for

mixture analysis by combining CESFS with measurements at

low temperature (77K) is discussed in Chapter 7. Isomeric

mixtures were evaluated by low temperature CESFS. Constant

energy differences were optimized for individual components

in the mixtures. Individual components in the mixtures were

spectrally separated and resolved.
















CHAPTER 2

THEORETICAL CONSIDERATIONS



Molecular Luminescence

In order to discuss the concepts of constant energy

synchronous luminescence spectroscopy (CESLS), it is neces-

sary to review the basic principles of molecular lumines-

cence. After the absorption of light and the formation of

an electronic excited species, radiative and nonradiative

electronic decay processes occur. Luminescence is the term

describing the radiative electronic decay mechanisms.

Fluorescence and phosphorescence are two types of lumines-

cence which are schematically described by Figure 2a,b. In

both examples, the electronic states are labelled G, S, and

T. A singlet electronic state S contains electron pairs of

opposite spin, and a triplet electronic state T contains

electron pairs of the same spin. The ground electronic

state G is generally a singlet state.

The electronic states of molecules are comprised of

both vibrational and rotational levels. Energy differences

between the electronic levels at room temperature are

several electron volts, while vibrational and rotational

spacings are tenths of electron volts and hundredths of

electron volts, respectively. The energy differences of the





































cn



r-* C0


r .-4 0f
0 -r4


~-4H ul
v4C~


41 9
E)O r4







*H 0r rn
.4Jv0









~4 (d.
ra w Q
4)U













*f4
Ea 0Q
Cd 0
-H 0 E
:j 0r







(3)0






rX4















)- C0


o() 0D









rotational transitions are usually too small to distinguish

between at the room temperature and liquid phase conditions

used in the major portion of this work. The vibrational

states are the levels within the electronic states shown in

Figure 2a,b.

The fluorescence transition for an excited molecule is

shown by the potential energy diagram in Figure 2a. A Morse

curve for the ground electronic level G as well as a curve

for the first excited electronic level S are given. The

molecules absorb a certain amount of energy sufficient to

populate the electronic level S as well as the vibrational

levels within S. The excited molecules vibrationally relax

from the upper vibrational levels to the ground vibrational

level within S. Spontaneous emission of energy in the

excited molecule occurs from S G. Radiative emission of

energy from S + G is called fluorescence.

The potential energy diagram in Figure 2b illustrates

the mechanism of phosphorescence. With the presence of

another excited state of similar potential energy T, a

transfer from S + T of the excited species can occur through

intersystem crossing. The crossing takes place at the point

of intersection between the potential energy curves. Again,

the excited molecules undergo vibrational relaxation to

the ground vibrational level in T. Phosphorescence is the

radiative energy transfer from T G. The triplet state

will have lower energy than the corresponding singlet state.

Thus, the fluorescence spectrum will always occur at shorter









wavelengths than the phosphorescence spectrum for a given

compound.

Nonradiative decay processes occur for many compounds

undergoing electronic absorption transitions. Internal

conversion is the rapid non-radiative transfer of energy of

excited molecules into vibrational energy. Molecules that

lack rigidity and have some degree of flexibility tend to

undergo internal conversion. Since excited solute molecules

are constantly colliding with solvent molecules, energy can

be transferred to the solvent molecule through external

conversion. Predissociation is another nonradiative decay

mechanism. At some vibrational level in the excited S, the

excited molecule can cross over into another excited singlet

S' at the point where the potential energy of the two states

are equal. Dissociation will occur from S' if the energy of

the excited molecule is approximately equal to the disso-

ciation energy of S'. Jaffe and Orchin give an excellent

discussion on the radiative and nonradiative electronic

decay processes (41). A number of good reference books by

Parker (42), Guilbault (43), and Winefordner, O'Haver, and

Schulman (44) describe luminescence in greater detail.

There are two types of electrons that determine the

electronic spectral features of a molecule's valence elec-

trons and lone pair electrons (n-electrons) present in the

valence shell (44). Even though the n-electrons cannot be

involved in the molecular bonding, they contribute to the

electronic spectra observed for some molecules. The valence









electrons of the atoms that make up the molecules can form

two types of bonding orbitals: the a orbitals for single

bond formation and the 7 orbitals for double bond formation.

Strong bonds are formed with the a orbital electrons. A

large amount of energy is necessary to excite the a orbital

electrons involving transitions in the vacuum ultraviolet

region, < 200 nm. Compounds with electronic transitions in

the far UV have a greater tendency towards predissociation

than luminescence. For example, 200 nm corresponds to

approximately 150 kcal/mole of energy. This amount of

energy is enough to cause bond breakage in most molecules.

Electrons in double bonds require less energy for exci-

tation than in single bonds. Transitions for the electrons

would occur in the UV and visible region (200 nm 700 nm).

The i electrons move freely throughout the atoms in the

molecule and are said to be delocalized. Molecules such

as benzene and the polyaromatic hydrocarbons (PAHs) have

delocalization extending over the entire molecule. These

cyclic compounds contain conjugated double bonds which

confer a special stability to the molecule because the

hexagonal structure is ideal for the formation of bonds in a

relaxed and unrestrained manner. For PAHs, the transition
*
w + T is possible (w orbitals are antibonding orbitals).

PAHs undergoing i n transitions produce strong absorption

and luminescence spectra in the near UV to near IR region.

The electronic spectra of PAHs have been studied extensively









by Clar (45), Berlman (46), and Birks (47) and have been

used in analytical analysis for a number of years.

The absorption and fluorescence spectra of anthracene

and perylene in Figure 3a,b will be used to discuss the

spectral shapes and relationships. A mirror image relation-

ship generally exists between the absorption and fluores-

cences spectra as observed in Figure 3a,b. The absorption

spectrum shows the vibrational spacings of the excited

electronic states, and the fluorescence spectrum provides

information concerning the vibrational spacing of the ground

electronic state. The absorption spectrum can contain more

than one electronic band of vibrational spacings as transi-

tions can occur from the ground electronic level to the

first excited electronic or the second electronic states.

The fluorescence spectrum should contain only one band since

transitions back to the ground level usually originate in

the first electronic level.

Compounds with aromatic structure display three or

four rather distinct vibrational bands with spacings of
-i
S1400 cm due to the predominant -C=C- spacings. Weaker

=C-H stretching vibrations are present with spacing of
-I
- 3000 cm (47,48). The fluorescence spectra of anthracene

and perylene exhibit a slight shift to the red, a Stokes

shift, due to the interaction of the solvent molecules with

the ground state and excited state PAH molecules. For PAHs

in alkanes the Stokes shift is 200 cm-1
in alkanes the Stokes shift is = 200 cm.




















41




2


44
0

41





'U
0





:3
0
'U
























4.
0

















>1
r- 0
C.


a)
























0)
a)













a)















0-
'-I















u











4.
0



ON










4.)-
T4.
o


C',



a,

cZ





19










I0
CL

0.0
CD




0 a




I0




A4i.SUs4U! GA4OIQOJ













C0
oo

















00


CC
'o
O~OC


















i0


'ci: 4)
re)



















3

0
Cu



A~isualui GAI~D;8J









The fluorescence spectrum of perylene is red-shifted

with respect to the spectrum of anthracene because of the

increase in the number of benzene rings due in part to an

increase in the degree of conjugation (43). Alkyl substi-

tutions result in a shift to slightly longer wavelengths.

Linear PAHs also tend to fluoresce at longer wavelengths

than do nonlinear PAHs of equal ring size (46).

Molecular luminescence spectroscopy has its main

advantages over absorption spectroscopy through the improve-

ments obtained in selectivity and sensitivity. A specific

absorption-fluorescence process can be monitored through the

use of two wavelengths, one excitation and one emission,

increasing selectivity. Many compounds may absorb at a

particular wavelength but all of these will not undergo

fluorescence and phosphorescence at the same wavelength.

Because luminescence measurements are made at angles of 600

to 1200 (usually 900) from the absorption path length,

background noise and scatter from the excitation wavelength

can be reduced; resulting in an improvement in the sensi-

tivity. Analytical limits of detection have been published

for PAHs measured by both room temperature and low tempera-

ture fluorescence spectrometry at wavelength maxima for each

compound (49). The limits of detection ranged from 50 fg

for perylene to 3.5 pg for phenanthrene.

Problems with spectral interference exist in lumin-

escence techniques from Raman scattering and Rayleigh

scattering. Rayleigh scatter is scatter resulting from the









interaction of the excitation wavelength with the electrons

of the molecules. With proper scanning procedures, Rayleigh

scatter can be reduced. Raman inelastic scattering occurs

at some energy loss from the excitation wavelength due to

interactions of the solvent molecules with the wavelength of

light. In obtaining a luminescence spectrum of a compound,

the Raman energy band appears as a broad peak in the spec-

trum and can mask spectral features of the compound. The

luminescence spectrum of a compound is generally broad,

and appears in a wavelength range from 300 nm to 700 nm.

Subsequently, the luminescence spectrum of a mixture would

consist of overlapping peaks from each compound comprising

the mixture. Identification of individual components would

be essentially impossible. Synchronous luminescence spec-

troscopy was developed as a way to improve mixture spectral

resolution.



Synchronous Luminescence

Synchronous luminescence spectroscopy is a technique

of measuring luminescence as the excitation and emission

monochromators are scanned simultaneously while maintaining

a constant energy or a constant wavelength difference

between the two monochromators. For dilute concentrations

of luminescing species, synchronous luminescence follows the

law of conventional luminescence; the intensity observed at

a given excitation and emission wavelength, Mexem is given

by:









Mex,em = 2.3 Ioex f ex) b c f(em) em) (1)


where Io ( ex) = relative intensity of the incident

radiation

S= fluorescence quantum efficiency

E(X ) = molar absorption coefficient

Yf(Xem) = fraction of total luminescence emitted

<(Aem) = detection system relative response function





Lloyd was the first to use constant wavelength syn-

chronous luminescence (CWSLS) for the analysis of forensic

samples in the early 1970s (34,50,51). The constant wave-

length difference maintained during scanning was empirically

determined. Subsequently, Vo-Dinh has laid the theoretical

foundation of constant wavelength luminescence (52-54).

CWSLS has been applied to biological and environmental

problems. John and Soutar characterized crude oils from

various locations with CWSLS (55). Vo-Dinh has applied

CWSLS to numerous environmental problems monitoring waste

water and air pollutants and coal liquefaction products

(35-37). Other researchers have used CWSLS for the deter-

mination of pharmaceuticals and LSD (56,57). Synthetic PAH

mixtures were analyzed by Thompson and Pardue using a SIT

vidicon with synchronous luminescence (58). The application

of synchronous luminescence for the characterization of

automobile engine oils has been described (59).









The constant wavelength difference (AX) in nm is

obtained for a compound by the following equation:





AX = (em ) (2)



where em = fluorescence wavelength, nm

Aex = absorption wavelength, nm





Through the selection of AX, a CWSLS scan offers spectral

bandwidth reduction, spectral profile simplification, and

minimization of Rayleigh and Tyndall scatter in contrast to

a conventional luminescence scan. These advantages can be

shown in Figure 4. Figure 4 is a contour plot of anthracene

obtained from a 3-dimensional plot of excitation wavelength

vs. emission wavelength vs. intensity. A conventional

fluorescence spectrum is obtained by scanning the emission

monochromator at a constant excitation wavelength. These

scans are labelled in Figure 4 and can be shown spectrally

in Figure 3. Three broad peaks are present in the conven-

tional fluorescence scan. With a AA scan of 2 nm, one

spectral peak of reduced bandwidth is obtained as shown in

Figure 4. Rayleigh and Tyndall scatter can be minimized by

selecting the spectral bandpass to be smaller than the AX

value. Synchronous scanning results in a spectral pro-

file simplification.


















4J

(




-41
Cd


OW0



O
Q)H












H-4 E-
4c











Q0















>4W






r
0 0



-4
:3
C
















00W
rl ..









rzC
0 i-1

















0'
> H














U)












-1-4
** I
*P 0,


(0
e





































































D UL)


(WN) H13N313AHM NOI1UBIIX3


0
(n










en









In CWSLS, scanning with a AA results in a change in the

energy difference between the excitation and emission

monochromators throughout the spectral scan range. Two

compounds with the same constant energy difference at

different wavelength pairs would require two different AX.

Synchronous luminescence scanning with a constant energy

difference would make use of inherent spectral relationships

of the absorption and fluorescence spectra.



Constant Energy Synchronous Luminescence

Constant energy synchronous luminescence spectroscopy

(CESLS) was developed by Inman and Winefordner in 1982;

here, a constant energy difference is maintained between the

excitation and emission monochromators while synchronously

scanning (39). The constant energy difference (Av) in cm-1

is obtained by the following equation:




AT = (1/ex 1/A em) 107 = (3)




CESLS allows the synchronous scan to be optimized for a

specific absorption-fluorescence process by setting the Av

equal to the overall vibrational energy loss of the process.

In addition to the advantages of CWSLS, CESLS can also

decrease the Raman scatter interference by selecting a AV

greater than or less than the Raman energy band (39). Inman









has compared CESLS to CWSLS with room temperature and low

temperature analysis of PAHs (39,40).

The distinguishing feature of CESLS over CWSLS is that

a constant energy difference is maintained while scanning

instead of a constant wavelength difference. The advantage

of CESLS scanning becomes more apparent with the data shown

in Table 2 and Figure 5 of a 3-component mixture of anthra-

cene, perylene, and naphthacene. For a constant energy

difference of 200 cm-1 (the Stokes shift), a spectral peak

for anthracene at 375 nm, 377 nm and a peak for perylene at

435 nm, 438 nm would be obtained. This would correspond to

a constant wavelength difference of 2 to 3 nm. CESLS and

CWSLS show similar spectra. Larger variations result

between CESLS and CWSLS spectra as the energy difference is

increased to 1 vibrational quantum and to 3 vibrational

quanta. For a Av of 1 vibrational quantum, two AX values,

22 nm and 30 nm are necessary to monitor major wavelength

pairs of the two compounds. A AX scan at 22 nm would

provide spectral information at major wavelength pairs for

anthracene but not for perylene. For a Av of 3 vibrational

quanta, the AX values needed to provide equivalent spectral

information range from 53 nm to 92 nm.

By scanning with the Av to be equal to the overall

vibrational energy loss of an absorption-fluorescence

process, spectral information will be obtained for all PAHs

with transitions resulting in that energy loss.
















4 r::




0444~




rdE-4
U).

'PU)












()W
0

WN

Crd 4
OU)






41



0 JQ)

roa
wwu

. 0







0J 0 -q

0U) 4

Q4 4-)~c
4- U)
(d




o gJ.



o U)

r-4




Nc H

0U
U)c
Nc



































































0 0 0
U,) 0 N U


(WN) HIDN313AUM NOIiW.IDX3


0
- 0







0
or)
o















0
44

CU

3

ra)












0

0
1-4






a)






',
4l























0 r
0














UW-)
0
1a1
EU











-4
a



o(

'aH





a




Ca




*a
a)










'a


in
en


0

0





P4 co co
Z n




















Ln



'0 '0
W %







r "



-n

y3 1^
M in ^
W en *q
M ..
U o <
rijf^ (








n, r^


MI m oo >rl- (N



% n %a%%





X n n aCO o ao m
) e M 0o co a























a) rL C"On IVO 0N



-<
-H S m S
X in 'n w "o In
(n en m r~n m mMn











4 r-l






(> 1) 14-
S>a >ao








I 1> I \m
< < :









CESFS spectra of anthracene at two constant energy

differences are shown in Figure 6. From conventional

fluorescence data in Table 2 and Figure 3a, spectral peaks

can be calculated and predicted from major excitation,

emission wavelength pairs and equation 3 for Av of 1 vibra-

tional quantum (- 1400 cm- ) and 3 vibrational quanta. The

wavelength pairs and corresponding spectral peaks are shown

in Figure 6a,b for the two constant energy differences. The

Jablonski diagrams in Figure 6c,d further demonstrate the

overall vibrational energy loss and correlate the transi-

tions with the corresponding spectral peaks. A reduction

in spectral bandwidth of CESLS spectra can be seen when

compared to the conventional fluorescence spectrum in

Figure 3a.

CESLS will be applied in two different ways in this

work: 1) a constant energy difference will be selected to

give characteristic spectra for eluting PAHs in a liquid

chromatography detection system, and 2) a constant energy

difference will be selected to promote spectral separation

for spectral fingerprinting in both room and low temperature

work.





























Figure 6. CESF spectra of anthracene: (A) Av = 1 vibra-
tional quantum, (B) Av = 3 vibrational quanta,
(C) Jablonski diagram illustrating
Av = 1 vibrational quantum, and (D) Jablonski
diagram illustrating Av = 3 vibrational quanta.

















(A)

LS =. 1400cm1-


anthracene
100 ppb

)


I I I


200
206


(B)

AT = 4800 cm-
anthracene
100 ppb


ex 200
em 221


500 200
658 221


wavelength (nm)

0
3
2


4


0




3
2


SI

(D)




So


A, E I vibrational units


wavelength (nm)


2

C



3
-- 2
Ay:--3 --- 0i
30
A = 3 vibrational units


() 356,377
'0 375, 399


ex 200
em 206


S-
s,-


(C)


--- ,


I


I


dJ
















CHAPTER 3

RAPID SCANNING CONSTANT ENERGY SYNCHRONOUS
FLUORESCENCE SPECTROSCOPY AT 20 nm/s



Introduction

There is a great need to monitor trace levels of poly-

aromatic hydrocarbons (PAHs) in the environment. Complex

PAH mixtures can be determined by combining liquid chromato-

graphy (HPLC) with a fluorescence detection system. Fox and

Staley collected HPLC sample fractions and obtained conven-

tional absorption-fluorescence spectra of each fraction as

well as derivative and selective modulation fluorescence

spectra for peak identification (22). Das and Thomas used

a variable wavelength fluorescence detection system for

HPLC analysis of 9 PAHs with three sets of excitation,

emission wavelengths (23). Ogan, Katz, and Slavin success-

fully separated 16 PAHs listed as EPA priority pollutants

with three sets of wavelength pairs (24). Several environ-

mental water samples were also analyzed for the 16 PAHs.

Nielsen purified PAHs from automobile exhaust by thin layer

chromatography separated them by HPLC, and used a stop-flow

fluorescence detection system to obtain fluorescence spectra

for 7 PAHs (25). Allen, Hurtubise, and Silver used


, A









corrected excitation fluorescence spectroscopy as a detec-

tion system for PAH components (26).

Fluorescence has been shown to be a selective and

sensitive technique with improvements in detection limits

over absorption by at least one order of magnitude (49).

Problems with currently used fluorescence detection systems

include the inability to resolve spectral overlap of the

components by conventional fluorescence techniques and the

inability to optimize the absorption and fluorescence

wavelengths for each compound.

The most desirable fluorescence detection system would

use major absorption-fluorescence wavelength pairs for

monitoring eluting PAHs and would allow spectra to be

obtained for identification. The potential utility and

feasibility of a rapid scanning constant energy synchronous

fluorescence spectroscopy system (RSCESFS) at 20 nm/s was

evaluated for this purpose. Constant energy synchronous

luminescence has proven to be a selective and sensitive

scanning technique in which spectral bandwidth reduction,

spectral profile simplification, and minimization of

Rayleigh, Tyndall, and Raman scatter occur in contrast to

conventional fluorescence scanning (39,40).

RSCESFS encompasses synchronous fluorescence scanning

at a fast rate with a constant energy difference. This

technique has been evaluated for optimal vibrational energy

difference for a number of PAHs. For application of









synchronous luminescence to multicomponent analysis, the

energy difference is generally chosen to promote spectral

separation (39,40,60). For this work, a constant energy

difference is chosen to give characteristic spectra for all

seventeen PAHs. A discussion of scan mechanics, flow cell

designs, and selection of the constant energy difference is

included. RSCESFS is compared to conventional fluorescence

with measurements at excitation and emission wavelength

maxima.



Experimental

Reagents

The PAH solutions were made with spectroscopic grade

heptane obtained from Burdick and Jackson. Suppliers of the

PAHs were Aldrich and Analabs.



Instrumentation

The data and spectra discussed in this chapter were

obtained with the RSCESFS system shown in Figure 7. The

components and manufacturers have been listed in Table 3.

The scan speed of the monochromators motors was mechanically

limited to 20 nm/s. The motor resolution was one pulse per

angstrom. Simultaneous pulsing of the monochromator wave-

length motor drives was controlled by a KIM microcomputer.

The excitation monochromator was pulsed at a constant rate

while the emission monochromator was pulsed at a faster,

































4J
Q)
to



E
.f-4
U)

4)

x
u)


C>
0







C1~4

4.)


co
M
















C-)
U)




0
.Z4
rX4

























-ii-
CL0

Et
oo
wn E

















>
44
Co







m 0
>0





H-H
o






S(0
LO











*H 4
2 cn



























0
U,
i-I

H























0













U)
i-


44
04
rdo








2 0
>* H





*H
0O




C(
u0




















o

























4 U
M 4




P4 (
W


































C
0

















E
C



























CO

k







H-


0
Oo
00


cH)


u 1


0
H.
IQ




















-u4
0l
0 41

















I





m












o
U



0



ONJ

0-<
0,0
0


















"Jo
O0
M r
P Cn
















M


0
co
qo

.O
00
o

HC







04-i

H
0















4 0
U,
























I1



0



O
O

u
mc
- *1>



*k0


OC





S< -I
O





00




Om
-,-o
>


"-4
04


0

t 0
00


c4

U)




4-)
H



rl (0
u -


5-4
0
,.Q





41


CNQ
aN
0,0

0 Tr
U

U)


41
(C





E 0
'I-i



fl


0
r(
5:l
0)

4J0
.4-i






















-,
H(






















>i.U)
r

















cu
















0 -i

II






44
Cr4



U


>1
r-l


<0
o
0
5.0





. r-4
>)C
0





A (a
0


(0
CP'


I-i


0 0
> 0
wo
L n
0
U II
) u
aH
CPE-4
a(

0 -
> 4-4
-H
.14



S-4

U



















!-1
,a


U
(0



r4
3




















a,
0)
'0a















O
0
5




























*H
o1
Hr


0
C


















.1
0
f-
U




























4
I
O0





























r0
C
E
































0
u
O



0

r-(-



U






o;
HO
0


S4
10












































0
0
U

























O
E












H






























,


E
-p

Qi









variable rate. The excitation wavelength scan ranged from

200 nm to 500 nm with the emission wavelength scan from

206 nm to 538 nm for a constant energy difference of
-1
1400 cm- and from 221 nm to 658 nm for a constant energy
-i
difference of 4800 cm The KIM program was written in

6502 assembly language with a "lookup" table providing the

number of extra emission pulses needed to attain the con-

stant energy synchronous scan.

A laboratory-constructed nanoammeter with a 500 ms time

constant was used for conventional luminescence measure-

ments. For RSCESFS measurements at 20 nm/s, a Keithley

current amplifier with a 1 ms time constant was employed.

From the current to voltage converter, data was output to a

recorder where measurements were obtained from the recorder

tracings for the RSCESFS evaluation study. Spectra were not

corrected for instrument response. Spectral bandpasses were

8 nm for all measurements.



Methodology

Three micro flow cells were evaluated. A schematic

diagram as well as sectional view of each cell is shown in

Figure 8. The first flow cell, (A), was constructed of a

black Teflon tube holder, 1.2 cm x 1.2 cm x 4.5 cm with

1 mm x 8 mm slits and 2 mm o.d., 1 mm i.d. Suprasil quartz

tubing. The illuminated volume was 6 UL. The second cell,

(B), was a U-shaped quartz tube of the same dimensions and

illuminated volume. The U-shaped tube was fitted with a









Teflon cuvette cover and was immersed in a CC14-filled

Suprasil cuvette, 1.2 cm x 1.2 cm x 4.5 cm. A commercially

available flow cell (Hellma Cells Inc., Jamaica, NY 11424)

is shown in (C). The Hellma cell consisted of a rectangular

18 pL illuminated volume with 1.5 mm x 8 mm quartz windows.

The sample cavity was surrounded by a black absorber for

stray light and scatter reduction.

Stock solutions of 100 ppm with heptane were made

for the seventeen PAHs listed in Table 1 of Chapter 1. A

separatory funnel was attached to the micro flow cell and

acted as a sample reservoir. Varying the flow rate from

0 to 10 ml/min while scanning had no effect on the fluor-

escence measurements. Figures of merit from fluorescence

measurements obtained with excitation, emission wavelength

maxima were compared to figures of merit obtained with
-1
RSCESFS at constant energy differences of 1400 cm- and
-1
4800 cm Limits of detection were run on each PAH using

dilutions from 10 ppm to 1 ppb. A spectral data base and

calibration curves for the seventeen PAHs were generated in

this manner.



Results and Discussion

The RSCESFS system was evaluated for 1) scan mechanics,

2) optimal flow cell in both detection limits and stray

light reduction, 3) best constant energy difference for all

seventeen PAHs, and 4) comparison of conventional fluores-

cence to RSCESFS at 20 nm/s.



























Sa)


'-4 0
0 1
0


C0 >
Q) .4-
4lM




0
0 -4
HO








494


E
H *a
*H









O-,














* *
'm



























































I\0,
5'%1


C






0,>





45



Scan Mechanics

When developing a scanning system it is important to

evaluate the maximum scan speed that results in the smallest

amount of spectral distortion in the reproducibility of the

forward and reverse scans, and in the optimization of the

signal to noise ratio.

The maximum scan speed, R, in nm/s, has been defined

by O'Haver (61) as that speed providing minimal spectral

distortion. The maximum scan speed depends upon the half-

width, 6Ab, of the most narrow portion of the molecular

fluorescence band and the electronic bandwidth, 6f, of the

measurement electronics,





6&b
R < 6X 6f =
b 2 rT
c





where t is the time constant of the measurement elec-
c
tronics. The time constant, T used in the RSCESFS at

20 nm/s studies was from 1 ms to 100 ms and the fluorescence

bandwidth, 6Xb, was approximately 8 nm. The maximum scan

speed was 1300 nm/s. Since this RSCESFS system was mechan-

ically limited by the scanning motors to 20 nm/s no spectral

distortions were observed in any of the recorded synchronous

spectra.









Noise and background interferents such as stray light,

spurious fluorescence, scatter and dark current inherent

in the detector can be picked up by the detection system

decreasing the signal-to-noise ratio (62-64). It is impor-

tant to reduce the noise in the RSCESFS system as much as

possible; resulting in a system that is shot noise limited.

Signal-to-noise ratios, S/N, vary with the square root of

the measurement time/spectral interval if the system shot

noise prevails and is independent of the measurement time/

spectral interval if the system is flicker noise limited.

To evaluate the signal-to-noise ratio for the RSCESFS sys-

tem, the parameter, RSD or relative standard deviation is

used. The RSD is inversely proportional to the S/N. The

RSD will decrease by a factor of 2 if the measurement

time/spectral interval is increased by a factor of 4 for the

shot noise limited case and will be independent of the

measurement time/spectral interval for the flicker noise

limited case. The RSD values are listed for RSCESFS spectra

of 100 ug/ml anthracene for 3 scan rates and 3 spectrometer

slit arrangements in Table 4. Using the example of 1 mm

slits, as scan rate is decreased from 20 nm/s to 5 nm/s, the

time/spectral interval increased from 0.4 s to 1.6 s; a

factor of four increase. The RSD dropped by (4)1/2 from 1.9

to 0.9. This trend follows throughout the other cases.

Therefore, it is apparent that the RSCESFS system is shot

noise limited and is optimized for signal-to-noise as much

as possible.

























a

4.
* C

S a
a C
U a
'o E





0I


0
0
'-4


T



Sn
r-.4 L



Ca




.4

OCO




)











C1 0 .ma A .
i> *


O. dp
Ca

0)

0




CD

-.. OLOOLA



C(


0



II
01
4> a .
a oomo oin






-H *H
Iu UI











rn + w U4


a
-e)



0

4.

-P,















0
41
-r-
u

























09
0)
x







a4
Ln










4-)







54
0













o
4C
a)
a


(z












0)
O* *H











0 E!
u -)
oCa
3 C


St
H IH
'a H
(0 '-

H
Ca
-0







Q LA
I0 0
d o
dP


'-I (


a)
N

*il m

C a

0 to (U
o c


X 0 (
O C 04
*E 4


04 4I
0 a
C o l




) a
'-1 1 -
0
O

U C C








o o

(U C
N *H O
Ca lr
a)
0

+1 r- *

- C M
)a a
SO o
c ao



0C C*

H H C
(0 00

(a r- 41

4J 0 >
**H a)
3a, C



C4-) *4 1


C) H
0

i rAC Q)

SQ)* m "
e *H a






( Q U


a
*



C





u


Ca 0
4(0
S4-







*H r
*i-l
0.
C o


-U < O
41.
















-,-4
x r









w 0
l H
'a a)
C a
*H a)

0 10
*0
(N


04 M

0
a)
a a

x a
a o
< *H
4 J
O 'U


t o
C 0
*H
o T


a) '0
U C


0 0
0 4
a a


H *a
u
o a
k0 4


,I
To























Q)
aI
a
04


C











I-l






.0







a









Flow Cell Comparison

The optimal flow cell for the RSCESFS system was one

which provides maximum reduction in stray light and scatter

throughout the scan. According to Lyons and Faulkner (65),

four scattering/ luminescence mechanisms can contribute to

background interference: 1) scattering of the excitation

beam by refraction and reflection at both the inner and

outer flow cell surfaces and interfaces; 2) Rayleigh and

Tyndall scattering; 3) Raman scattering; and 4) luminescence

from interferents and contaminants.

Rayleigh and Tyndall scattering can be reduced in

RSCESFS by using spectral slits smaller than the constant

energy difference selected. The interference of Rayleigh

scatter is most prominent when the Stokes energy shift,

200 cm-, is chosen as the constant energy difference. The

-1
Raman band for heptane is approximately 3000 cm and Raman

interference can be eliminated by choosing a constant energy
-1
difference other than 3000 cm Elimination of the reflec-

tion and refraction from the flow cell walls while scanning

was the most crucial problem to overcome in the flow cell

selection.

The stray light levels can be seen for the three flow

cells in Figure 9. For each flow cell, blank spectra of
-1 -1
heptane were obtained at Av = 1400 cm, and 4800 cm

Signal levels have been given in nA for comparison. The

black Teflon cell has both circular inner and outer bound-

aries as shown in the sectional cell view in Figure 8.































-1
Figure 9. RSCESFS _pectra of heptane at Av = 1400 cm and
4800 cm in the black Teflon cell, CC4 cell,
and Hellma cell.










Black Teflon Cell
Ai7 = 1400cm"1


AV = 4800cm-!


em 206 538 206 em 221 658 221
wavelength (nm) wavelength (nm)


CCI4 Cell
Av = 1400cm"'


Av= 4800cm-"


ex 200 500 200 ex 200 500 200
em 206 538 206 em221 658 221
wavelength (nm) wavelength (nm)


= 4800 cm-


500
538
wavelength (nm)


ex 200 500 200
em221 658 221
wavelength (nm)


ex. 200
em 206









The inner and outer circular interfaces cause refraction and

reflection of the light giving the largest background

levels. Since CCl4 and quartz have the same refractive

index at 589 nm a decrease of two in the background was

obtained by immersing the quartz tubing in CC14. Scatter

due to the inner circular surface was still present. The

greatest background reduction was seen with the Hellma cell

which has square inner and outer boundaries. A decrease in

scatter of approximately 100 was found with the Hellma cell

versus the black Teflon cell.

A figure of merit comparison for the three flow cells

was run with six PAHs and the results are listed in Table 5.

Fluorescence measurements were obtained at the excitation,

emission wavelength maximas for each compound. The Hellma

cell gave the best figures of merit over the black Teflon

cell and the CC14 cell. Based upon these results, the

Heilma cell was chosen for all further studies.



Selection of Av

The optimal constant energy difference for RSCESFS

scanning is the one which produces synchronous luminescence

spectra for the 17 PAHs. Spectral complexity is of second-

ary importance in this application. From conventional

fluorescence spectra, energy differences were calculated

from various excitation, emission wavelength pairs of each

compound with the equation (2). In Table 6, wavelength

pairs for 17 PAHs are listed and grouped according to energy
















0

H *3
r4




r U
J 0






0 H
- rII-I o O r






H C 000
SO u u






t7> 4-1 i 00
a I gr r,- c0 o n
o 4 c4 HU4.ij



S0
C r





O H
C 0
0 0













.1 0 a -H0
4Ci 4-J .0
QA C





0 -i, u o









-0 0 4
C X 3 ii uo mc o co r o a

o 0


0 0 0









0 0 4
-I -l C





















4E C' )






-4-4 < *
C N N O 0

0 411



-4 0 0 0 r-

) 0) 01 a) H 01
E-4 it Q 4-4 C
^~~ (0 X) 3 < Q

































C 4N *-4 v


4 -i
0 4 -





-4-U )
U)3 4








0
a )


- H
s Q E



C



, -Hri
cM
W 0 04
Q:









r4
m4e

























0 r4 tIq
U, 4
.HO










4Ji
it: Q 0a







C I
S f
0 *
m -E 6





Q C
U -P -

r d -
-ue-
a) r <
a i-i


-a U V


C Q
A-4
CQ dP


-44
0

rc.
0

E- (0




.C
CO
om


Ha


r-i
aC



Q4 0U
wm

r-H

-H

C rd
0 Cr
tLn






4 -r

E 0)
a, -H
0
SH
4C
O

EU )













a, -H
-.-)













(n *4-i
0 4
0 0
4 -
n -c









w CD
G 0 m
(U0- f
m -1> (
3 U *-
(0 <> ^
(0 +


0
m N N 0


4-
O


-r
-,-
0

4-1

0

,1




4-J




0
a,




rOdP





oa
4



.O0





0
a,
0 C


a) to


.*0
0
4Q)
a ,
cM





C 0














$4
M u




















tm0
0 .,










rEU
4a,














-4 0
Q)C





*HO








*H
i- *











S m w Cn N O o oM

fn IV o ro0 o 0 0 C(oo


0g m rm mmmm m m m(m
I M
0 00 U000 000 0000

> 00 0000 000 0 0000
)r M e m n Mem en vm








C 4
o .-4L4 .0 ( '
04
-4 j r4 0) -ir n c' ON w o (n m Cr


*^ o gr~ o CMOvI oof



( e ( n r 00 C r C -Y n
H 4-(Uf X (Nmm) mmm mm qw vR T

oI C
il *4 00 000 000 000 000
--.) 00 000 000 000 000
Sd mX m mmnm m m (v mmm mm w

4*H

U 0

M N


02 .Q '-4 LA 'n O mo
U (q C co o0 om
U H U) m m m
O u(( M < S S S.
U >cE o CC 03 o00
0 0 CO A

4 0

En jQG1r- o oo o o oo
0r3 .. 0 00 0 0 00

UO ) n O O
>a > z (N Lo'O 0n CD C-
S< (N -I r-0






C: .
U) C a 0 0

4-1 -4-4 4

U) 0)-w 0 0


0 *4.)) "-
0 a 4 ( fo -
rH UC N N N

E 0 G fi aQ oa














Oda a)

0 41
O ,

)( 4-) (d x


- -
4J


r-H
























Ot
0 (







o-'



*H
c E




*Hl


Sr- T- C *IV



W 00 v m w







IV IV) in in Ln
00000
M4AII0


0in C1


mN N


00
00
r-4 CM
oo

GNao owr- CpA re rv
(aOr-Il4rm-Il NOLn


m r- S N 5 CN v
m mcMf'mmo mcom


00

c00
CNM n


oo
.. 0.
S0
S0


m cn M m
D CO CO a




0000










0)
c









1)
04
-I


00
00

M m


000
000
r co '
CN CNM


OCN



m m
coo
n m



00
00
m 0
r- cM


co r-

n en
a CD

m o




00
0 0
oo

o n


CM 0
o 00
m n

lo 00

M cn
mm


(NM r- (V



r- m qw (m
Nr- o n
mN;r q
mmwa
mmes



0000
0000
(N COr~ o
T T ^r in)


00
*co


-NWN



00
00
CM 0


V- m D
1ri o
d*sr n *<
f


000
000

CMmm


O


ca

c,

0

ONo

mm.
--
0-rol


n m



m r
oo
M rIn




o o
00
00
-4 (C4


00
00
'-",













.40 Coo .O aM n O L %D m
H 1-qN r Cen ON wf r co r-H


(1) 0 o Ln o co Co oa o 0 -I -4
P 4+ ZC X m m n c m -q m c m mf 4- U)4
4 m (a ) (0 C




S00000 0 000 0 00 I0

i4--.
S 0 a4- Q
S000 000 0 1 I 1 0
O) COl r.4
0 c0 u
) ~ 4-4 % -rj
S H r-

0 O0 O0 t.w 0)0
-- ,4 en a% n 4-0 0 "00 r.
04J c) : 0 A L 00wv I V



04 0 00 0
S00omo 0C 00 a o a UI
^+j (a rn fn n CQ Cn m co m r W







I o 4J -
OMo mm *M >4 W




S44
M tJ < O U0 U

,o r- 0 000 00 m 0
0 o coo cto i o > I tf
(D -cmm mm m '0
9: 4)42
4 '0C *4
O CaO 5
1 *H) 0
0 co MC
SP -i a V
S4-r
S4 m41v 0 (a


X mmm mm a v-410
4>1 00 00 00 4 () M 4-i
Q) 0) rl 0 0)


r4 000 0 0 0 wr )

0 % i m a) 0 0
0l < a) 4r Q) H w 0w w w




C4 4J' C0 =U r- g 4


0) (1 a)Wu )*H
4 000 00 00 4)




rq 000 0 00 0 4I -- 0)
< irN rr iN tO
t7 U0 C
I C) C *





C) CU. ^ U 004



C0 T! ( 0 0
*^ H a) > a, a) U)l





+* O UJ H 0 ) 0CU
H I U M U) 4
.0 04 r-l (U (Q1


S3 C C 0 0 a 0




( -C ( 04 4 >4 *









differences of 1 vibrational quantum, 2 vibrational quanta,

and three vibrational quanta. Very few of the 17 PAHs had

excitation, emission wavelength pairs corresponding to the
-I
Stokes energy shift, 200 cm1. These pairs were not

included in the table. From the compiled data in Table 6,

several PAHs had no spectral peaks with an energy difference

of 1 vibrational quantum and/or two vibrational quanta.
-1
Heptane has a Raman band of approximately 3000 cm- and so a

constant energy difference of 2 vibrational quanta was

eliminated as a scan choice. All PAHs in the data base had

excitation, emission wavelength pairs corresponding to

3 vibrational quanta. A compromise constant energy differ-

ence of 4800 cm- was chosen to approximate the vibrational

energy loss of 3 vibrational quanta. Four of the 17 PAHs,

benzo(b)fluoranthene, fluoranthene, naphthalene, and phenan-

threne had only one pair of wavelengths apiece with A
-1 -1
ranging from 4500 cm to 5300 cm .
-1
To further substantiate the choice of 4800 cm- as the

best compromise energy difference for RSCESFS, a comparison

of figures of merit for 6 PAHs with Av at 1400 cm- and Av
-i
at 4800 cm1 was carried out. In Table 7, the detection

limits ranged from 700 fg to 9 pg for the 6 PAHs at both

energy differences with the exception of benzo(b)fluoran-

thene which gave a detection limit 10 times better in the
-I
4800 cm1 scan. Excluding benzo(b)fluoranthene, the detec-

tion limits were within a factor of 4 for both Av values.









The linear dynamic range varied from 3.5 to 5.1 decades and

the blank %RSD values from 2.4 to 5.9% for both At values.

Analytical figures of merit for all 17 PAHs obtained at
-I
the 4800 cm energy difference are shown in Table 8. The

limit of detection range was from 700 fg for perylene to

500 pg for phenanthrene.

One final evaluation of the RSCESFS system at 20 nm/s

was necessary. The best case for each PAH in chromatography

would be detection and quantitation at the excitation, emis-

sion wavelength maxima for each compound. The data in
-1
Table 7 for the 6 PAHs at 4800 cm- energy difference with

the Hellma flow cell can be compared to the data obtained in

Table 5 for the 6 PAHs at the excitation, emission wavelength

maxima using the Hellma cell. The RSCESFS system at 20 nm/s

has comparable linear dynamic ranges, blank %RSD values, and

slightly worse limits of detection by 2 to 5 times.

In Figure 10, conventional excitation emission spectra

for benzo(k)fluoranthene as well as the RSCESFS spectra at

Av = 1400 cm- and 4800 cm- are given. The reproducibility

of the forward and reverse scans with both constant energy

differences is evident. The synchronous spectra in Fig-

ures 10b and c contain numbered peaks (1,2,...) correspond-

ing to one vibrational quantum or three vibrational quanta

energy losses in the absorption-fluorescence spectra. This

energy loss can be further seen in the Jablonski diagrams in

Figures d and e, showing the transitions corresponding to

the wavelength pairs.















1-1








II


I,


0
<









0



















O
0
II

I1


















O
r-i
u











0
0




12














Ur-




a)


0





P4 '-E
co












S-4


(a1
0




(al


OCN
O~ e~ -4 *


u u


04 =


<0
0
1-4

Q) r4l


r-1 Be-i


u u


C C


S ar -0


ro
0


0
r- (4-

4.)








co
4-I



C 4



U)M
S4. -
00


C -

40







O4-
n o


4->
(O 0
fcl


c tr





u

rn
(no


4-)


O (D
r0






l a )
2 -H C













o 0
U)O
W 0
















a) -H4
G 0
itn

4
U (U
K -r4




O
a( 0


















. n <


0



0

-J
0

.HJ

4-
0






4-4
0
*4-









(a

4-i



Or

a)

0a a
0a)



0a)


CdP


a





x
0 03







> .4
V 0Z
1) C: H
4) >i4-i
0 (U






0 O3



2 0

.Q 4
(U C U





e c
C; !-
*^ -












o^ o
cnJ















0 0
0)









S0
C4
u






w 0
C 0


iD ^









Table 8. Analytical Figures of Merit for 17 PAHs Using
-1
RSCESF at Av = 4800 cm and with the Hellma Flow
Cell



Detectiona Absoluteb Linear
Limit Detection Dynamic Blank
Compound ng/mL Limit, pg Range %RSD


acenaphthene 3. 54 3.4 4.8
anthracene 0.5 9 5.1 5.1
benz(a)anthracene 2 36 3.5 4.9
benzo(b)fluoranthene 0.3 5 4.8 3.7
benzo(k)fluoranthene 0.2 4 5.1 3.8
benzo(g,h,i)perylene 2 36 3.7 3.6
benzo(a)pyrene 0.08 1 5.1 2.5
benzo(e)pyrene 2 36 3.5 4.5
chrysene 4 72 3.2 3.8
dibenz(a,h)anthracene 2 36 3.6 5.4
fluoranthene 1 18 4.0 5.3
fluorene 0.3 5 4.1 5.9
indeno (1,2,3-c,d)pyrene 0.9 16 4.1 3.2
naphthalene 30 500 2.5 3.9
perylene 0.04 0.7 5.1 5.1
phenanthrene 10 180 2.7 3.4
pyrene 2 36 3.5 3.8



a. Measurements taken from RSCESF scans with 1 ms time
constant and 8 nm spectral bandpass. The limit of
detection is the concentration giving a signal 3 times
the standard deviation of 20 blank measurements.

b. The illuminated volume was 18 ul.

c. Linear dynamic range is expressed as the linear orders
of magnitude from the detection limit to the concentra-
tion where the slope falls by 5%.

d. Relative standard deviation is based on 20 blank
measurements.






















Figure 10. RSCESFS and conventional fluorescence spectra of
100 ppb (benzo(k)fluoranthene.

a. conventional fluorescence excitation-emission
spectra.

b. RSCESFS at Av = 1400 cm1 (one vibrational
quantum). Corresponding wavelength pairs
given for labelled peaks 1 and 2.
-1
c. RSCESFS at Av = 4800 cm- (three vibrational
quanta). Corresponding wavelength pairs
given for labelled peaks, 1,2,3, and 4.

d. Jablonski diagram correlating wavelength
pairs 1 and 2 to transitions with one vibra-
tional quantum difference.

e. Jablonski diagram correlating wavelength
pairs 1,2,3, and 4 to transitions with
3 vibrational quanta difference.












(A)
benzo (k) fluoranthene
100ppb
430
400



'3
457
360
334 0 4
479


250 360 450 550
wavelength (nm)


(B)

Av'= 1400cm-'
benzo (k) fluoranthene
0 ( 100ppb


' (D. 310,40;
* 400,431
c
S

0
a





ex 200
em 206


S,


(D)


so-


500 200
538 206
wavelength (nm)


3
2


0





2
0 2


8y = I vibrational units


(C)
Ap= 4800cm-'


ex zuu
em 221


658
wavelength (nm)


SI

(E)




So
A~ = 3 vibrational units


O
0









Conclusion

RSCESFS at 20 nm/s has potential as a HPLC detection

system because

1) the RSCESFS system has detection power and lin-

earity approaching conventional fluorescence measurements

taken at excitation, emission wavelength maxima.

2) the RSCESFS spectra are excellent for identifi-

cation; with the 4800 cm- scan, a reliable spectrum is

obtained for each of the 17 PAHs.

With RSCESFS system at 20 nm/s a stop flow liquid

chromatography system would be required. Faster stepper

motors added to the system (10 times faster) would allow

on-line synchronous fluorescence spectra to be measured.
















CHAPTER 4

RAPID SCANNING CONSTANT ENERGY SYNCHRONOUS
FLUORESCENCE SPECTROSCOPY AT 200 nm/s



Introduction

A rapid scanning fluorescence system was developed by

Warner, Callis, and Christian in the early 1970s in which

total luminescence spectra were obtained (66-69). A video-

fluorometer was used, combining two polychromators, a

silicon-intensified target vidicon (SIT), and a multichannel

analyzer.

The excitation polychromator was turned on its side so

that the excitation wavelengths were dispersed vertically up

and down the sample cuvette as shown in Figure 11a. The

polychromatic excitation beam spanned a 260 nm range from

the ultraviolet radiation at the lower part of the cuvette

to the near infrared radiation at the top of the cuvette.

The emission polychromator was set up in the normal manner

with the 260 nm wavelength span at the excitation poly-

chromator perpendicular to the polychromatic excitation

beam. The excitation image was observed along the vertical

axis and the emission image was observed along the horizon-

tal axis resulting in a 2 dimensional image (total lumines-

cence spectrum) focused onto the SIT. A plot of a






























Figure 11.


The excitation monochromator optical axis is
dispersed vertically up and down the cuvette on
(a). A plot of the excitation-emission image
appearing on the SIT can be seen in (b). A
3 dimensional plot of the excitation-emission
image is shown in (c). Taken from D.W. Johnson,
J.B. Callis, and G.D. Christian, Anal. Chem. 49,
747A (1977), reference (67).






















:0 a.

I *


400 500
wavelength,


A


Red-


Yellow-

Blue-


Violet-


600
nm


Ui


eoo
Xce


T |









2 dimensional image is shown in Figure llb. A total lumi-

nescence scan spanning a 260 nm range along both the excita-

tion and emission axes was obtained in 18 ms with 1 nm

spectral resolution (68). The total luminescence scan can

be envisioned better in Figure llc with a 3 dimensional

plot.

Contour plots of total luminescence spectra are shown

in Chapter 2. The spectral interference of Raman and

Rayleigh scatter can be seen in the total luminescence

contour plot of anthracene in Figure 4. The Raman scatter

ridge runs through the total luminescence spectrum of

anthracene. As a result, the Raman scatter masks a portion

of the anthracene spectrum. Figure 5 is a contour plot of a

3 component mixture of anthracene, naphthacene, and pery-

lene. The total luminescence spectrum of the mixture is

spectrally complex. Overlap between the anthracene and

perylene spectral images and between the perylene and

naphthacene images can be seen. The difficulty of identi-

fying individual PAHs in total luminescence spectra would

increase for mixtures with a larger number of PAH compon-

ents. Mixtures of PAH isomers such as benzo(a)pyrene,

benzo(e)pyrene, benzo(b)fluoranthene, benzo(k)fluoranthene,

and perylene would be even more of a problem to resolve and

identify by total luminescence spectroscopy because of the

overlap and spectral complexity of the scan.

A reduction in spectral complexity is the major advan-

tage of synchronous luminescence scanning (39,40,53,54). A









SIT vidicon detection system was combined with constant

wavelength synchronous fluorescence spectroscopy (CWSLS) by

Thompson and Pardue for measurement of several single and

two component PAH mixtures (70). A total luminescence

system, with 2 polychromators, and a SIT detection system,

was set up similar to the videofluorometer of Johnson,

Callis, and Christian (67). Spectra were collected from the

SIT system and stored on the computer disk for data pro-

cessing; a constant wavelength difference (AA) was calcu-

lated between the excitation and emission wavelength pair of

maximum spectral intensity for each PAH. For various

concentrations, the fluorescence measurements made at these

excitation, emission wavelength pairs were used to determine

the PAH detection limits. From the two dimensional total

luminescence data matrix, the fluorescence values of all of

the excitation, emission wavelength pairs corresponding to

particular AX values were obtained. CWSF scans were recon-

structed for the values. Thus, with the SIT detection

system, synchronous scans of various AX values were possi-

ble.

A new approach to rapid scanning fluorometry will be

discussed in this chapter with constant energy synchronous

fluorescence spectroscopy at a 200 nm/s scan rate. Mechan-

ically-scanned monochromators were used for rapid scanning

constant energy synchronous fluorescence spectroscopy

(RSCESFS) at 200 nm/s. The merits of a constant energy

synchronous fluorescence system at 20 nm/s scan rate were









given in the previous chapter. With a single constant

energy difference (Av) of 4800 cm1, 17 PAHs were identified

and detected. CESFS spectra at a Av of 4800 cm- provided

valid peak identification of 17 PAHs. With the 20 nm/s

system, 17 PAHs were detected at levels in the parts per

trillion to parts per billion range.

The instrumental development and software for the

RSCESFS system at 200 nm/s will be discussed. Since data

are analyzed during the forward and reverse scannings, the

reproducibility in peak position and intensity between the

forward and reverse scans will be evaluated. The RSCESFS

system at 200 nm/s was set up to measure PAH eluants contin-

uously for further application as a chromatography detection

system. Thus the performance of the data processing system

will be studied through the generation of a simulated

chromatographic separation.

The RSCESFS system at 200 nm/s uses a photomultiplier

(PMT) detection system. A comparison of the SIT detection

and the PMT detection system for synchronous scanning will

be given.



Experimental

Instrumentation

For the RSCESFS system at 200 nm/s shown in Figure 12,

faster stepping motors with controllers and signal pro-

cessing components were added to the 20 nm/s system dis-

cussed in Chapter 3. These components are listed in




































CN

4-)

uu
+rz



U)


4-4
0
E)





-4J







C4)










--4






71






CC-)
o (n 0 Ia. E E







w a- 0







u) Q. 0

c: E









Table 9. The motor resolution was 200 steps per turn

resulting in 0.5 nm per step.

An Apple II plus microcomputer performed 3 primary

functions: 1) synchronous scanning of the excitation and

emission monochromators in the forward and reverse direc-

tions; 2) data acquisition; and 3) data transmission to the

PDP 11 minicomputer.

To provide the constant energy synchronous scan at

200 nm/s, an initialization routine generated a "lookup"

table of the emission pulse rates. To attain a RSCESFS scan
-1
at 4800 cm the excitation monochromator was pulsed at a

constant rate of 400 Hz and the emission monochromator was

pulsed at a variable rate up to approximately 700 Hz. The

synchronous scan was controlled by an assembler program.

Acquisition, conversion, and memory storage time per

data point was 100 us. Data acquisition and transmission to

the PDP were synchronized with the excitation monochromator

pulse rate as shown in Figure 13. The width of the pulses

for motor stepping was 15 vs. The excitation and emission

monochromator pulses were interrupt-controlled; the excita-

tion interrupts occurred every 2.5 ms and the emission

interrupts at shorter intervals. Data were acquired on

alternate pulses to the excitation monochromator. Eight

analog to digital conversions were obtained for an excita-

tion monochromator pulse. The signals were averaged, and,

after the next excitation monochromator pulse, the averaged




















0

0 -
*H 0


a)
HE4

0 -



















Or














0
02

















0
O
Q



o















0










r.
0
2:

En











E


0
o
UO
0 O
Or-
4- 0

0

U
,-I.






















0
OHr
*O 0
M +
0)0




















E-4












4t
0
3



























4



E0







H
0




9I

H
4o-
'-l

ro
r:



4-


CN




S0


Q











U0






S
0
















4,

0

0
o



















4in










H
0)
-S
0)









C


C.C
H



(o
0 m
4



U)
H <
> >1
4Ju
C.

04
) rd

c m
Hm


H Ln
aCo
Sn04
Ik
4O



OC
o c
u -


04


0


Ro
4Jn

0

























rlz
0























.HP










-"-I

-,-







a


0
u
O

4J n


r04





O4
Eo

a














c.c
0J






































C.%
a










































-4
-I-


-47








O































Figure 13. Time sequence of pulsing for excitation mono-
chromator in (a) and emission monochromator in
(b). The data acquisition and transmission to
the PDP is shown in (c). The signal handling
processes are based on the excitation pulse
rate.























15 US
P--I


I I
ex pulsing ,
I I I
II I
I I I
I I I
I I I


I I I


11111 I I
I transmission
A/D transmission


(A)







(B)



(C)









signal was sent to the PDP 11/34 through the Apple II

asynchronous serial interface (California Computer Systems,

Model 7710) at 19.2 kBaud.

A versatile interface adapter (VIA, SY6522) was used

for communication between the Apple II microcomputer and the

RSCESFS system. An interface was constructed for use with

the stepper motor translation modules and the VIA; optional

manual operation of the motors was also included. A current

to log voltage converter was used, giving an output of 0 to

10 V corresponding to a five decade range of -0.1 nA to

-10 1A.

A PDP 11/34 minicomputer received data from the

Apple II for processing in real time. Data were transmitted

through the Digital DZ11 serial port at a rate of

19.2 kBaud. Serial data transmission was chosen instead of

parallel data transmission because the latter is more

sensitive to outside noise. The serial data transmission

used error checking making it easier to verify data in the

developmental programming stages. The DZ11 serial port has

a data buffer. Data were read in only when the buffer was

full.

The main program was written in Fortran with Fortran

and assembler subroutines. Data were read in from the

serial data buffer of the PDP under interrupt control by an

assembler subroutine. The assembler subroutine processed

16 ASCII characters at a time, converting the characters to

octal values. A total of 201 data points (201 nm) were









transmitted to the PDP for each RSCESFS scan. Figure 14a

gives a schematic of the RSCESFS scanning. Each 12 bit data

point was encoded by the Apple II program as 1 direction

character and 3 data characters every 2.5 ms with parity

checking. After the 201 data points per RSCESFS scan were

transmitted, the area of the RSCESFS spectrum was inte-

grated. The integrated log fluorescence value was output to

the recorder through the Digital Laboratory Peripheral

System (LPS) digital-to-analog converter with an output

voltage range of -5 V to 5 V for the 5 decade current range

of -0.1 nA to -10 uA. The integrated values output to the

recorder reconstructed the chromatogram in real time as

shown in Figure 14b for 14 scans. Each value corresponds to

log fluorescence intensity per second. The integrated value

was stored for further processing.

Signal processing included baseline, drift, and peak

monitoring and the storage and output of RSCESFS spectra.

Each integrated value was compared to a threshold value

above the noise level. A program counter was used to

distinguish a valid peak increase or decrease from drift

over time or a noise spike. If a peak was present, moni-

toring continued until the slope was zero or negative. The

RSCESFS spectrum was then saved. Times at the beginning and

end of each chromatographic peak were also stored.

At the end of the run, the integrated values were

printed out. The spectrum saved for each chromatographic

peak was output to the recorder for PAH identification. A






























Figure 14.


(a) The RSCESFS scanning scheme is shown for a
total of 14 s. Each scan is integrated to
determine the area under the spectrum. The
integrated values are output every 1 s to
reconstruct a chromatogram (b).













(A)


time, s


(B)


I I4 Isi l l l l I Ii I I I
0 5 10 15 20 25 30
time, s









direction character, the 201 data points, and the start and

stop times were printed for each spectrum stored.



Methodology

Stock solutions were made for 6 PAHs with heptane. A

separatory funnel was attached to the Hellma flow cell for

sample introduction. To determine figures of merit for the

RSCESFS system at 200 nm/s, data were collected in the

forward and reverse directions for 10 scans. Calculations

were performed by the Apple II microcomputer.

To evaluate the PDP data processing programs, a program

was written on the Apple to simulate a chromatographic

separation. Data values representing data points of RSCESFS

scans were generated by the Apple II and sent into the DZ11

serial port to test the data transmission, monitoring, peak

recognition and RSCESFS spectral storage.



System Evaluation

Stop Flow vs Continuous Flow

The RSCESFS system at 200 nm/s was developed for

continuous monitoring of the eluant in high performance

liquid chromatography. Unfortunately, a stop flow chroma-

tography system would be required for peak detection with

the CESFS system at 20 nm/s. With stop flow systems,

broadening of the chromatography peaks over time is a

disadvantage. With the 20 nm/s scan, a minimum stop flow

time of 11 s would be necessary for a 200 nm scan of each









eluting PAH. The 20 nm/s system would rely upon an addi-

tional peak detection system to detect the oncoming PAHs, to

synchronize the timing of the stop flow process with the PAH

entering the CESFS flow cell, and to start the 20 nm/s

scanning. There was also concern as to whether one scan at

the peak maximum would be a statistically valid measurement.

With a continuous analysis system (200 nm/s), peak

broadening would be reduced. Assuming a chromatographic

peak had at least a 15 s halfwidth, 15 scans would be

obtained. The integrated value of the area under the

chromatographic peak could then be determined for quantita-

tive measurements.

A disadvantage of the RSCESFS system at 200 nm/s is

that the data processing in real time at the faster scan

rate required more programming and hardware control. A loss

in the signal to noise was expected over the 20 nm/s system.

The evaluation of the RSCESFS system at 200 nm/s consisted

of three parts: 1) determination of the reproducibility

between the forward and reverse scans for both peak position

and intensity; 2) determination of analytical figures of

merit for 6 PAHs; the 200 nm/s system was compared to the

20 nm/s system for signal to noise, detection limits, and

linear dynamic ranges; and 3) microcomputer generation of a

simulated chromatography model to test the signal processing

capabilities of the PDP 11/34.

The reproducibility of the forward and reverse scans

was dependent on the "memory" effect of the motors and the









play in the mechanical linkage to the gratings as the

monochromators were scanned in both directions. As the scan

direction was changed in the assembly program from forward

to reverse or reverse to forward, the speed of the motors

was of sufficient magnitude to cause the motors to continue

in the same direction for a small number of pulses. For

example, if the excitation motor continued in the forward

direction for 4 pulses after a change of direction was made,

the wavelength readout at the end of the reverse scan would

be 2 nm less (4 pulses X 0.5 nm/pulse) than the correct

reading. For continuous forward and reverse scanning during

a chromatographic separation, a scan time of 25 minutes

would not be unrealistic. Gross discrepancies occurred in

the experimental values obtained versus the correct values

(250 nm to 450 nm) for continuous forward and reverse scan

times as short as 1 minute. The same errors were observed

with the emission monochromator.

The "memory" effects observed in these motors were

corrected by adding a 1 ms time delay after the direction

change, before the program continued with the scanning

pulses in the new direction. With the 1 ms time delay

between changes of direction, accurate wavelength readouts

from the start to the end of the spectral range were

obtained for continuous forward and reverse scanning times

of 25 minutes.

For wavelength scanning in optical systems, a sine-bar

mechanism is used to rotate the grating to provide a linear









wavelength readout. This mechanism consists of a sine bar

and lead screw. Constant linear motion is converted to a

sinusoidally-varying angular motion by this mechanism so

that the wavelength appearing at the exit slit of the

monochromator is proportional to the angle through which the

motor shaft has turned.

To evaluate the sine-bar mechanism for reproducibility

of spectral peak positions in the forward and reverse scans,

a Hg pen lamp was used as the source. Spectra were obtained

for each monochromator during forward and reverse scannings;

the emission monochromator was scanned in the forward

direction from 250 nm to 450 nm and in the reverse direction

from 450 nm to 250 nm. Slits of 0.1 mm (0.8 nm spectral

bandpass) were used. A similar procedure was employed for

the excitation monochromator with a 450 angle mirror placed

in the sample cell holder to reflect the Hg lines through

the emission monochromator. The excitation monochromator

slits were 0.1 mm and the emission monochromator slits were

1 mm. The monochromators were scanned synchronously in both

directions. Data were acquired by the Apple II microcom-

puter as described previously in the experimental section.

The positions of the spectral lines were compared between

the forward and reverse scans for both monochromators. For

the excitation monochromator, the positions of the Hg lines

in the reverse spectrum were correct (within 0.5 nm) but

were red-shifted by 1 nm in the forward spectrum. For the

emission monochromator, the position of the Hg lines on the









reverse spectrum were red-shifted by 2 nm and the forward

spectrum was correct.

Since data was collected in the forward and reverse

directions, it was important that the spectral intensities

measured were reproducible between the forward and reverse

scans, from forward scan to forward scan, and from reverse

scan to reverse scan. Ten scans in the forward and the

reverse directions from 250 nm to 450 nm were run at

Av = 4800 cm- with a 100 ppb solution of anthracene. The
-1
RSCESFS spectrum of 100 ppb anthracene at A7 = 4800 cm-1 was

previously given in Figure 6b. Four major spectral peaks

occur in the RSCESFS spectrum at excitation, emission

wavelengths of (1) 324,337 nm; (2) 338,399 nm;

(3) 356,422 nm; and (4) 375,446 nm. The means and standard

deviations of the spectral intensities at the 4 wavelength

pairs were calculated for the 10 forward scans. The same

calculations were made for the 10 reverse scans. The %RSD

was obtained from the mean and standard deviation values.

The average %RSD of the 4 spectral peak intensities in the

forward direction was 3.2% and in the reverse direction was

3.4%. The mean values of the 4 major spectral peak inten-

sities in the forward and reverse scans were compared and

differed by only 2%. No significant difference was observed

in the spectral intensities measured in the forward and

reverse scans. The small differences in peak positions

within the forward and reverse scans had no adverse effect









on the spectral intensities measured at the major spectral

peaks for anthracene.



Analytical Figures of Merit

The major disadvantage of the RSCESFS system at

200 nm/s over the 20 nm/s system was the reduction in the

signal to noise for an equivalent number of scans (note the

200 nm/s system obtains the signal in 1/10 the time of the

20 nm/s system). The RSCESFS system at 20 nm/s was mechan-

ically limited to a maximum scan rate of 20 nm/s corre-

sponding to a time/spectral interval of 400 ms as discussed

in Table 10. For the RSCESFS system at 200 nm/s, the

time/spectral interval was reduced by a factor of 10.

There was a significant difference in the way that the

data measurements were processed in the 200 nm/s system and

the 20 nm/s system. With the 20 nm/s system, data was taken

from the recorder tracings. The recorder has a 100 ms

response time. In the 200 nm/s system, data were digitally

averaged by the Apple II microcomputer. The time constant

of the current to log voltage converter varied from 400 us

for -0.1 nA to 4 ps for -10 pA. The signal to noise ratio

(S/N) is proportional to the square root of the time con-

stant (t ) as shown below,



1/2
S/N at 20 nm/s t of 20 nm/s
S/N at 200 nm/s (t of 200 nm/s)
c









By comparing the 400 us time constant in the 200 nm/s system

and the 100 ms time constant in the 20 nm/s system, a signal

to noise reduction by a factor of 16 would be expected.

Accounting for the increase in scan rate and the difference

in time constants, a loss in the signal to noise of approxi-

mately 160 was expected (a factor of 10 reduction for the

faster scan rate and a factor of 16 reduction by using the

smaller time constant) with the RSCESFS system at 200 nm/s.

Analytical figures of merit for the two systems are

given in Table 10 for 6 PAHs. The detection limits became

poorer by a factor of 25 to 50 with the 200 nm/s system

compared to the 20 nm/s system. The linear dynamic range

for each PAH decreased by at least one order of magnitude

with the faster scanning system. Detection levels for the

6 PAHs are in the parts per billion range for the RSCESFS

system at 200 nm/s. RSCESFS (200 nm/s) spectra for anthra-

cene and perylene are shown in Figure 15a and b.



Evaluation of the Data Processing

Analytical methods are needed to identify PAHs in

mixtures and detect PAHs at levels in the parts per billion

to parts per million range (4). Combining liquid chroma-

tography with RSCESFS at 200 nm/s detection system offers

the following advantages: 1) liquid chromatography systems

are capable of separating PAH mixtures into individual PAHs

or PAH fractions; 2) PAH levels are detected by RSCESFS at

200 nm/s in the parts per billion range; 3) RSCESFS scanning



















4




-D
U,






(N

4-
af



a



En

0
0

(N



-41






a)
CJrH















4J '
U

HS
#-oz





uca)
0.-







a)0


HIIl

ca)




bH

C)E
H

a)


a)

















oo
o
C)

CqqE










a

U



u~0
Sa)0






4-) -H I
0 -
f-A 0 0) -





a)







a) 10a r

0
r-I
rT4














.H
4-)


rz J r; -










a) .14



00
I~r

















00
c04
e(~r
rl Oat
u,~~


a)


dC


H D 00 W (4

















C C4 a;~h
mN mq (N (4














IV( HN 1H CN
























o 0 0 0a0)
(~ H
















(N







a) a)

a) a)




r. 4-i 4-4
Q) C)




o4 0 0 p) -
4) N N N >
0 ..Q .u a)
a) c- C) C

U C) ) C) H C)
a) -1a
k000k44


44





0 -i -
(N .-)
4-i
a)

CA,
r14

U,
cOU)
u:cu
0) 0)04
+t C)0
4J- -H


I-0 OC)-
Z 4J~


a) U 2'4-J
0 a
W 0J





U .0 )
3:
0
r-i

(z(

ri



V-1
Iq





coc






00

00
exu


r-4 N0














rorc
*r





















a)
0
C%~ Ln g

4-)
9:
0
U
















Q)cr
0 0







C)
a)


Z
44




fo
a)

C) H

C) -

C 0,
.m N
















(U )

C d

m c


u
301

c

QA










4-
U4-
-*H
ii
a-H


a o




4-1
0 a
l


.r-
4 r-l4 V 0


CN

0










C!




4-1
0
,-q



0
N
a


(0
4






o
it

0

EM
02.
U0


4-
04











0*



060
GI















o


a) >

00
0 0








4-)
H <






) CH -

Oa>








+l 0 C




o 0o
OO 0


Srl u

r-4r
- 0 e4
0>
SCOH










U> H
SOHcO
U +)4.)






a 3


















a
0)- U>
g> n


4 )
+1 (0

t0 (U
0H N

.r 4


Go
*-H)

HQ)
Oc


U)
WO
-P
04 r3




a z
a

r0



M 4
.0 U)
(0(0
M c
(0 4-)

Ha




rm
$H 4-
4-,






4-) 0
CO




-H r-4






>
4J -rH







'( 0



0a)
CO 0


aQ) 0
m u


.-)



a

4u-
0



0








0
*4-












eL
*Hl.















(4a
0,2

0
O

(1W
*
(
OCn

W








0

c0
co 04
a '



0




S(W



XM

m M
020






ao -
0









C1
rn 0)











t0 +


4-)



0
U
4)
tb


0 g
> -r

41
C(

S00

W
a)
H a
OE




4-)
W02
0r
S(0



4 .-I
CO









4:44
0)
04 4



0





e u
n 0
2: 44


0 1a 4





























Figure 15. RSCESFS spectra at 200 nm/s and with
A = 4800 cm for (a) anthracene and
(b) perylene.



















RSCESFS at 200 nm/s

(A) anthracene (B) perylene -





S- -
C



0 355 386
C 338 408




u* 324 375
0







250 350 450 250 350 45


excitation wavelength, nm


i0


II

HI









results in a spectrum obtained for each of the PAHs analyzed

while reducing the spectral complexity seen in total lumines-

cence and conventional fluorescence scans; and 4) RSCESFS

spectra can be used for PAH identification in conjunction

with retention times.

Before further application as a chromatography detec-

tion system, the PDP 11/34 data processing programs were

tested to insure correct data transmission, peak recogni-

tion, and proper storage and output of the integrated values

and RSCESFS spectra. The Apple II program outputs the data

values to simulate the 201 data points of each RSCESFS

spectrum. The data were sent out of the Apple II serial

port and into the DZ11 serial port. The capabilities of the

data transmission program to read and interpret the direc-

tion characters and data characters were verified at this

stage. The Apple II program generated RSCESFS spectral data

to simulate a 2 component chromatographic separation of

150 s in duration. The area of each RSCESFS spectra was

integrated and sent out to the recorder. The output is

shown in Figure 16a. At the end of the 150 s chromatography

simulation, two RSCESFS spectra (a spectrum stored in memory

for each chromatographic peak) were sent out to the recorder

as shown in Figures 16b and c. Proper working of the peak

monitoring programs were confirmed. Finally, the stored

integration values and RSCESFS values were printed out and

crosschecked correctly with the values sent out by the





























Figure 16. A two component chromatographic simulation by
the Apple microcomputer results in (A) the two
component chromatogram reconstructed by the PDP;
(B) the RSCESFS spectrum corresponding to peak
(1); and (C) the RSCESFS spectrum corresponding
to peak (2).