Constant energy synchronous luminescence spectroscopy

MISSING IMAGE

Material Information

Title:
Constant energy synchronous luminescence spectroscopy theory and applications
Physical Description:
xii, 106 leaves : ill. ; 28 cm.
Language:
English
Creator:
Files, Leigh Ann, 1960-
Publication Date:

Subjects

Subjects / Keywords:
Luminescence spectroscopy   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1986.
Bibliography:
Includes bibliographical references (leaves 102-105).
Statement of Responsibility:
by Leigh Ann Files.
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 - 000989390
notis - AEW6252
oclc - 17689598
System ID:
AA00003386:00001

Full Text















CONSTANT ENERGY SYNCHRONOUS LUMINESCENCE SPECTROSCOPY:
THEORY AND APPLICATIONS










By

LEIGH ANN FILES


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


1986























dedicated to my parents
with all my love















ACKNOWLEDGEMENTS

First I would like to thank my parents, brothers, and grandparents

for all their support and love. Their faith in me allowed me to have

enough confidence to complete a task which I never could have

accomplished otherwise.

I would also like to express my appreciation for the wonderful

instructors; Dr. J. E. Bennett, Mrs. M. Cooper, Dr. N. L. Trautwein,

Mr. L. Battles, and Dr. L. D. Kispert, who encouraged me to pursue

research in chemistry.

I am deeply grateful to all the people I have worked with during

my stay at the University of Florida. I would especially like to thank

Jonell Kerkhoff, Ben Smith, Ed Voigtman, and Linda Hirshy for their

input and guidance in helping to establish my research system. I would

also like to acknowledge Brad Jones and Monica Moore for their

assistance with the PAH studies. Best wishes go to Mike Mignardi in

his future work with CESLS, and many thanks for his help with the final

project involving phosphorescence.

The acceptance and guidance as well as the invaluable friendship

of the people I have worked with at the University of Florida have made

my stay not only successful, but very enjoyable. Among these people I

must mention Kathy, Jonell, Doug, Jose, Brad, Moi, and Keith. I can

never express how much these people have meant to me.

Finally, I would like to say a very special thank you to Dr.

Winefordner. His dedication and enthusiasm for research as well as his

sincere interest in his students and colleagues are an example which I

will always admire and strive to emulate.
iii
















TABLE OF CONTENTS


Page

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

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

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

ABSTRACT.............................................................xi

CHAPTERS

1 MIXTURE ANALYSIS

Introduction..................................................1
Goals.... ..................oo.o- ..... ....................... .. 8

2 THEORETICAL OPTIMIZATION OF PARAMETER SELECTION
IN CONSTANT ENERGY SYNCHRONOUS LUMINESCENCE
SPECTROMETRY

Introduction.................. ................... .......... 11..
Theory................... .......... ...... ..... ....... ........ 12
Discussion...................................................20
Conclusions ..... ............................................ 39

3 ANALYSIS OF ENVIRONMENTAL SAMPLES CONTAINING
POLYCYCLIC AROMATIC HYDROCARBONS

Introduction............................... ................. 41
Experimental............. ....................................42
PAH Mixture Analysis......................................... 45
Gasoline Engine Exhaust Analysis............................46
Gasoline and Crude Oil Fingerprinting.........................55
Conclusions ......................................... ......... 68

4 FEASIBILITY STUDY OF CONSTANT ENERGY SYNCHRONOUS
LUMINESCENCE SPECTROMETRY FOR PESTICIDE DETERMINATION

Introduction................................................. 69
Experimental............. ...................... ................70
Results and Discussion.......................................71












CHAPTERS

5 TIME RESOLVED PHOSPHORIMETRY

Introduction ................................................. 85
Experimental ................................................. 88
Results and Discussion........................................ 90

6 CONCLUSION

Summary...................................................... 96
Future Work .................................................. 99

GLOSSARY................................. ........................... 100

REFERENCES .......................................................... 102

BIOGRAPHICAL SKETCH ................................................. 106

















List of Tables


Table Page

I Spectral Data for Three Model Compounds...................... 22

II Limiting Cases............................................... 26

III Spectral Data for Two Model Compounds........................ 32

IV Experimental Components for CESLS............................44

V Calculated Energy Transitions (in cm 1) for Carbary.........74

V Calculated Energy Transitions (in cm ) for Naphthol......... 75

VII Limits of Detection for Pesticides with Variable
Parameters ...................................................79

















List of Figures


Figure Page

1 Energy level diagram illustrating the luminescence
processes for an idealized molecule............................ 3

2 CESLS spectra of anthracence. (A) AV = 1 vibrational
quantum, (B) Av = 3 vibrational quanta, (C) Jablonski
diagram illustrating Av = 1 vibrational quantum, (D)
Jablonski diagram illustrating Av = 3 vibrational
quanta................................................ ......... 9

3 Geometric model for CESLS. The ellipse defines the
contour equivalent to half the maximum peak intensity.
The dashed line defines a CESLS scan with Av =v io-v .......18

4 Graphical representation of various values of Av (- -)
and the line (-----) defining the position of the CESLS
maxima for this scan.......................................... 19

5 Graphical representation of three special cases for
a) a. = b) a, << a c) o >> a ..........................25

6 Comparison_ f peaks generated by using a variety of Av
values (cm ) for compound A. See Table II.
Av = 15,000 (----), 14,000 (-----), 13,000 (----),
16,000 (-.-), and 17,000 (----)......................... 27

7 Excitation, A (- -) and emission, A (-- ) peaks
for compound X. oa = a The CESLS spectra, A ,
(------) is shown for tAe scan path Av = v v ............29
o 0 o
8 Excitation, B (----) and emission, B (---- ) peaks
for compound a. o. = a./2. The CESLS spectrum, B ,
(----) is shown for tie scan path Av = v v. ............30
o 0 o
9 Excitation, C (-- -) and emission, C (-- ) peaks
for compound a. o = 2a.. The CESLS spectrum, C ,
(----) is shown for thd scan path Av = v. v. ........... 31
1 0
o o
10 Excitation (- --) and emission ( ) peaks for
model compounds D (D ,D ) and E (E ,E ). See Table III.
CESLS spectrum (-- --x) with Av = 'v m for both
compounds (D and E ), arbitrarily p oated in terms of
excitation energy...........................................33












11 Peaks described in Figure 10 with x-axis in units of
wavelength (nm).............................................34

12 Excitation (----) and emission ( ) peaks for
compounds D and E as described in Table III. CWSLS
spectrum with AX = AX for compound D (85.227 nm)
(-- -- _)...............ma...................................... 35

13 Excitation (- --) and emission ( ) peaks for
compounds D and E. CWSLS spectrum with AX = AX
max
for compound E (121.457 nm) (-- --).........................36

14 Scans defined by A = 12,000 cm-1 (C ), AX 85.227
n (----.), and AX = 121.457 nm (----). For compounds
D and E as described in Table III.............................37

15 Schematic diagram of experimental system for obtaining
CESLS spectra............................................... 43

16 CESLS scan of 16 component PAH mixture AV = 1400 cm .........47


17 CESLS scan of 16 component PAH mixture Av = 4800 cm ..........4

18 CESLS scan of leaded gasoline exhaust extracts obtained
using multiple sample collection system ( AV 1400 cm ).
Compounds identified include benzo(a)pyrene (BAP),
benzo(k)fluoranthene (BKF), anthracene (A), coronene
(COR), and perylene (PER)..................... ............... 50

19 CESLS scan of leaded gasoline exhaust extracts obtained
using multiple sample collection system ( Av 2800 cm ).
Compounds identified include phenanthrene (PHE), chrysene
(CHR), pyrene (PYR), anthracene (A), benzo(a)pyrene (BAP),
benzo(k)fluoranthene (BKF), and anthanthrene (ANT)............ 51

20 CESLS scan of exhaust from different gasoline samples
obtained using a U-tube collection system at 1400 cm .
Samples: (a) Brand A Super Unleaded, (b) Brand A
Unleaded, (c) Brand B Super Unleaded with Ethanol,
(d) Brand B Unleaded with Ethanol, and (e) Brand C
Regular Leaded...............................................53

21 CESLS scan of exhaust from different gasoline samples1
obtained using a U-tube collection system at 4800 cm .
Samples: (a) Brand A Super Unleaded, (c) Brand B Super
Unleaded with Ethanol, and (e) Brand C Regular Leaded.........54

22 CESLS scans at 77 K and 1400 cm- for a) Brand D
gasoline (1:10 dilution), b) mixture containing benzo(a)
pyrene (BAP) and anthracence (A), and c) mixture
containing pyrene (PYR) and fluoranthene (FLU)................57


viii











-l
23 CESLS scans at 77 K and 4800 cm for samples identified
in Figure 22................................................... 58

24 CESLS scans at 77 K and 1400 cm-1 for 1:50 dilutions of
a) Brand E Super Unleaded Gasoline, b) Brand E Regular
Gasoline, and c) Brand F Regular Gasoline..................... 59

25 CESLS scans at 298 K and 1400 cm-1 for samples
identified in Figure 24 (no dilutions required)...............60

26 CESLS scans at 77 K and 1400 cm-1 for 1:50 dilutions of
a) Brand G Diesel and b) Brand G Regular Gasoline............. 62

27 CESLS scans at 298 K and 1400 cm-1 for samples
identified in Figure 26 (no dilutions required)...............63

28 CESLS scans at 77 K and 1400 cm-1 for 1:1000 dilutions
of crude oil number one distillation fractions a) below
35000C, b) 160 to 240 C, and c) 240 to 350C................... 64

29 CESLS scans at 298 K and 1400 cm-1 for 1:10 dilutions of
crude oil number one fractions described in Figure 28.........65

30 CESLS scans at 77 K and 1400 cm-I for 1:1000 dilutions
of crude oil number two fractions a) below 350 C,
b) 160 to 2400C, and c) 240 to 3500C ..........................66

31 CESLS scans at 298 K and 1400 cm-1 for 1:10 dilutions of
crude oil number two fractions described in Figure 30.........67

32 Excitation and emission scans of carbaryl. Prominent
peaks are listed in Table V................................... 72

33 Excitation and emission scans of naphthol. Prominent
peaks are listed in Table VI.................................. 73
-1
34 CESLS scan of naphthol with Av = 1400 cm Excitation
wavelength range 250-350 nm. Excitation and emission
bandpasses 1.5 unm.............................................76

35 CESLS scans of a mixture of carbaryl, naphthol, and
carbofuran with Av = 1400 and 2650 cm Excitation
and emission bandpasses 1.5 nm................................ 78

36 Constant energy scans of naphthol with Av = 2650 cm-1
Demonstrating the comparison between spectra obtained with
bandpasses of 1.5 nm on both monochromators versus a) main-
taining a 1.5 nm emission bandpass and opening the excitation
bandpass to 4 nm, and b) maintaining a 1.5 nm excitation
bandpass and opening the emission bandpass to 4 nm............ 81

37 Constant energy scans of carbaryl withAv= 2650 cm .
Showing the results obtained with the same variation in
parameters described in Figure 36............................. 82












38 Constant energy scans of a mixture of carbaryl, naphthol,
and carbofuran with a) bandpasses of 1.5 nm on both
monochromators, b) excitation bandpass of 2.5 nm and
emission bandpass of 1.5 nm, and c) excitation bandpass
of 1.5 nm and emission bandpass of 2.5 nm.....................84

39 Illustration demonstrating the ability to obtain more
selective spectra through the use of time resolution.
FL--flash lamp pulse, BC-boxcar, f--fluorescence,
p--phosphorescence, t --delay time, t --gate time.
Subscripts used to denote compounds with different
excited state lifetimes....................................... 86

40 Schematic diagram of experimental system modified to
allow one to obtain time-resolved CESLS spectra............... 89
-1
41 Constant energy scans with A = 12,000 cm for benzo(e)
pyrene obtained using (a) a continuous source and measuring
emission directly, and (b) a flashlamp as the source and
measuring emission with gated detection....................... 91


42 Constant energy scans with Av = 12,000 cm- for a mixture
of benzo(e)pyrene and phenanthrene obtained using (a) a
continuous source and measuring emission directly, and
(b) a flashlamp as the source and measuring emission with
gated detection............................................... 92
-1
43 Constant energy scans with Av = 12,000 cm for a mixture
of benzo(a)pyrene and carbazole obtained using (a) a
continuous source and measuring emission directly, and
(b) a flashlamp as the source and measuring emission with
gated detection .............................................. 93

















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

CONSTANT ENERGY SYNCHRONOUS LUMINESCENCE SPECTROSCOPY:
THEORY AND APPLICATIONS

By

Leigh Ann Files

August 1986

Chairman: James D. Winefordner
Major Department: Chemistry

Constant energy synchronous luminescence spectroscopy (CESLS) is a

simple, inexpensive method which demonstrates excellent selectivity for

mixture analysis. The spectral parameters of CESLS have been evaluated

for optimization of experimental variables. Mathematical relationships

have been derived for peak wavelength maximum, intensity maximum, and

peak bandwidth, defining their dependence upon the luminescence

excitation and emission spectral characteristics and the selected scan

path.

CESLS at low temperature (77 K) has been applied to environmental

analysis involving fingerprinting gasoline exhaust samples containing

polyaromatic hydrocarbons (PAHs). Combining this technique with a

system which allowed a crude sample separation based on temperature

provided increased selectivity and sample information.

CESLS was also applied to analysis of gasoline and crude oil

samples containing PAHs. Sample identification at low temperature

(77 K) with quartz tubes and at room temperature on filter paper was

xi












compared. The ability to identify individual PAHs in samples as well

as fingerprint different samples based on total PAH content was

demonstrated.

The results obtained theoretically involving optimization of

experimental parameters have been applied to determination of

pesticides. Limits of detection, analytically useful ranges, and

results obtained using a variety of scan parameters have been

determined for carbaryl, naphthol, and carbofuran. Results showed

promise and indicated that CESLS has great potential for a wide variety

of applications involving complex mixture analysis.

An experimental system has been developed to allow time-resolved

measurements in combination with constant energy scanning. Time-

resolution further enhances the selectivity of constant energy

synchronous luminescence spectra. The system has been designed for

maximum versatility. Room and low temperature measurements as well as

fluorescence and phosphorescence spectra can be obtained with minimal

instrumentation rearrangement required.

















CHAPTER I
MIXTURE ANALYSIS

Introduction

Luminescence techniques have often been used for detection of

polycyclic aromatic hydrocarbons (1-6), pesticides (7-12), and

pharmaceuticals (13-16). These methods are very sensitive due to the

high quantum yields of many compounds in these classes. Conventional

luminescence methods can be used to identify and quantitate species in

such mixtures, however, overlap in spectra often becomes a problem,

especially with complex mixtures or a mixture containing compounds with

similar structures. Mixture analysis is often done by first separating

the sample into its individual components through the use of separation

techniques such as HPLC, GC, or TLC, and then detecting the analytes

with another technique such as fluorescence, phosphorescence, infrared,

or mass spectrometry. Combining techniques in this manner provides the

desired selectivity at the expense of increased time and cost required

for each sample. Other approaches to mixture analysis focus on

increasing the selectivity of the detection method, and eliminating the

need for a separation technique. To appreciate the methods which can

be implemented to increase the selectivity of a particular technique,

it is necessary to understand the fundamentals of the process being

monitored. Therefore, a brief explanation of molecular luminescence

will be presented. For more detailed descriptions the reader is

referred to books by Schulman (17), Turro (18), Parker (19), and

Winefordner, O'Haver, and Schulman (20).

1












Figure 1 presents a Jablonski diagram illustrating the

luminescence processes for an organic molecule. Transitions labelled A

refer to the absorption process. An absorption spectrum corresponds to

the decrease in the intensity of the light passing through a sample due

to its absorption by the sample. The wavelength of light being

absorbed corresponds to the amount of energy required to excite the

molecule from the ground state to an excited state.

Once the molecule is in the excited state many processes are

available. One possibility is intersystem crossing, IS, to the triplet

state where it may then undergo internal conversion, IC, (nonradiative

decay), or it may undergo radiative decay in a process termed

phosphorescence, P. Phosphorescence occurs at wavelengths longer than

those of excitation and fluorescence--longer wavelengths correspond to

lower energy.

Another possibility for the molecule in the excited state is for

it to undergo vibrational relaxation, VR, to the lowest excited state--

at this point if the molecule decays radiatively to one of the ground

state levels this transition is described as fluorescence, F.

Molecular luminescence spectroscopy is the measurement of emission

accompanying the deexcitation of a molecule from an excited electronic

state, i.e. fluorescence and phosphorescence. Luminescence signals are

measured against essentially a zero background, whereas absorption

measurements involve small differences in two large signals.

Therefore, luminescence spectroscopy often provides an enhancement in

detection power. The low limits of detection generally obtained

produce large linear dynamic ranges--usually in the range of five

orders of magnitude. Also, detection systems involving fluorescence



























c
-o0
C C *
0 00
=x



.0

<:( I




c
0



S>






0
C-


0

C
02 0
o 0o



U. i 13.
0.....E 0






i 0r 0
L, O.
Li* a.


to C4-


0



Cu
5-'


CO
CO

0
w
(U


0

u
0



(1)
CO,
z


4-i
C


co
*--I







4.1

ca
-4
-4
co



C-u
" U





-4
- 0
S(U1


a) a)
-40
N
>5--
00-4 n




- r4

60

i-l


CO
0f)












and phosphorescence are inherently selective in that they determine

only those compounds which are excited by absorbing light and then

decay radiatively to the ground state. Two degrees of selectivity are

possible in such detection schemes due to the ability to select the

energy of the exciting light and the energy of the emitted light.

Still another degree of selectivity involves the use of delayed and

gated detection to monitor the lifetimes of the excited states of the

analytes. This aspect will be dealt with later.

The differences in the excitation and emission wavelengths for

different compounds are based on the relative positions of the ground

and excited states and the energy spacings between the levels in these

states. Many different techniques are used to exploit these

differences, and the choice of techniques is based on considerations

such as speed, resolution, and the number of components of interest.

One of the most widely used techniques for increasing selectivity

is the use of measurement at low temperature (( 77 K). Examples of

such methods include matrix isolation, supersonic jet expansion, and

the use of glassy and Shopolskii solvents. Lowering the temperature of

a sample reduces the internal degrees of freedom, molecular rotation,

and vibration. The lower the temperature, the narrower the energy

distribution profile of molecules within vibrational energy levels.

This narrowing effect reduces overlap between neighboring levels.

Increased sample information can then be obtained by selectively

monitoring transitions between specific energy levels, provided the

instrumentation is capable of accurately interrogating the exciting

energy and emitted radiation. When one increases the selectivity of

detection to the point at which it is possible to excite only those








5



molecules which inhabit very similar matrix sites, the term site

selection is applied. Lasers are generally used as sources of

excitation in site selection due to the intense, narrow bands which

they provide. However, xenon arc lamps are also used as sources at low

temperature, due to their low cost, reliability, and broad band

emission. Through the use of a monochromator, a xenon arc lamp can be

used to interrogate a wide range of excitation transitions. Xenon

lamps are especially useful in conjunction with Shopolskii

spectroscopy where the analyte and solvent are matched to produce a

reduction in the total number of possible orientations for a molecule.

Wehry and Mamantov (21) give more complete explanations of the theory

and advantages of the use of low temperature for enhanced spectral

resolution, as well as references demonstrating the use of such

techniques.

An added advantage of analysis at low temperatures is the increase

in quantum yields due to a decrease in nonradiative collisional

deactivation of analytes in the excited state. For a detailed

discussion of the radiative and nonradiative electronic decay

processes, the reader is referred to Jaffe and Orchin (22).

Still another method for improving the selectivity of a detection

system involves the options available for monitoring excitation and

emission energies. The most simple method of luminescence detection

involves fixed excitation and emission wavelengths, and monitoring the

luminescence signal intensity (typically through the use of a

photomultiplier tube). This method requires prior knowledge of system

response to individual analytes, and it must be verified that the

response is proportional to analyte concentration and independent of












sample matrix and other variables. It is also essential in such

measurements that a true, consistent blank be determined. These

requirements generally limit measurements of this type to determination

of one known analyte in a mixture, or require that the sample be

separated into relatively pure components prior to detection.

Another approach to luminescence measurement involves using either

a fixed excitation or emission wavelength and scanning to determine the

corresponding response. Spectra obtained in this manner provide

information concerning the vibrational energy spacings in the excited

electronic state, excitation/ absorption spectra, and the ground

electronic state, emission spectra. Often spectra obtained in this

manner are used for identification as well as quantitation. Through

the use of more advanced electronic design and computer interfaces,

spectra corrected for system response have been obtained for a wide

variety of compounds (23,24). It is expected that in the future

libraries containing luminescence spectra similar to Sadtler indices

for IR, NMR, and UV absorption will be available.

More recent attempts at selectivity enhancement for multicomponent

analysis have included such methods as total luminescence spectroscopy,

TLS, and synchronous luminescence spectroscopy, SLS. TLS involves the

use of multichannel analyzers and/or computers to obtain emission

spectra of a sample over a range of excitation wavelengths. One of the

major advantages of TLS is that it requires minimal prior knowledge

concerning the luminescence characteristics of an unknown.

Disadvantages of TLS include the cost of equipment required to obtain

and store the extensive data corresponding to total luminescence

spectra. Also, the amount of information in a TLS spectrum can often












be of such a magnitude that it is difficult to interpret. For a more

complete review and explanation of the theory, advantages, and

applications of total luminescence spectroscopy the reader is referred

to Inman (25) and included references.

Synchronous luminescence spectroscopy is a technique which

involves measuring luminescence as the excitation and emission

monochromators are scanned simultaneously while maintaining a

well-defined relationship between the wavelengths of the monochromators

(26). Conventionally this relationship is a constant wavelength

difference, Alternately, variable-angle synchronous luminescence

spectroscopy implements different scan speeds between the

monochromators, thus providing a variable wavelength difference during

the scan (27,28). Recently, constant energy synchronous luminescence

spectroscopy (CESLS) has been described where a constant energy

difference is maintained between the monochromators (29). In this way,

the natural energy relationships fundamental to spectroscopic

evaluation of fluorescent compounds are instrumentally exploited.

These energy relationships are represented in the Jablonski diagram

presented in Figure 1. The energy separation maintained between the

monochromators in CESLS is chosen to correspond to the overall

vibrational energy loss of an absorption-fluorescence transition.

The energy difference for a particular transition can be

experimentally determined by solving the equation


A= (1/ X 1/ X ) 107
ex em

where = excitation peak maximum in nm, em = emission peak
ex em transition in
-a
maximum in nm, and Av = energy difference for the transition in cm












The ability to selectively monitor transitions with CESLS is

illustrated in Figure 2 (30). Here constant energy scans of anthracene

are shown with their respective Jablonski diagrams depicting the
-l
absorption-fluorescence transitions being observed. The 1400 cm

scan yields two peaks, which are attributed to transitions resulting
-i
in an overall vibrational energy loss of one quantum. The 4800 cm

scan yields four peaks. These peaks are assigned to transitions which

involve a loss of three vibrational quanta.

The advantages of synchronous techniques have been discussed, with

examples demonstrating qualitative spectral improvements (31). These

advantages include a reduction in spectral complexity, a reduction in

peak bandwidths, and a decrease in interference from Rayleigh scatter

(32) and even Raman scatter (33).


Goals

The initial goal of this work is to provide a theoretical and

practical evaluation of constant energy synchronous luminescence

spectrometry (CESLS). The spectral parameters of CESLS will be

evaluated for optimization of experimental variables. Mathematical

relationships will be given for determining location of synchronous

peak maximum, maximum intensity, and peak bandwidth; defining their

dependence upon the luminescence excitation and emission spectral

characteristics and the selected scan path. Graphical representations

of CESLS scans illustrate these relationships, with hypothetical

compounds used to demonstrate their application.

The feasibility of fingerprinting and identifying PAHs in

environmental samples and analyzing synthetic mixtures of pesticides

and pharmaceuticals using CESLS will be demonstrated. Special emphasis




















(A)
z.C 1400cm-'


anrnracene
iO0ppb


$-


500
538
wavelength


....c 0


2C0O
206
(nm)


3
3
2





SO i


So 0
A' = I vibrational units


(B)

AT = 4800cm-1
anthracene
100 ppb


a ZOO
em 221


I I I


500
658
wavelength


(nm)


S-3







Sr i
So 0
av = 3 vibrational units


CESLS spectra of anthracene. (A) Av = 1 vibrational
quantum, (B) Av = 3 vibrational quanta, (C) Jablonski
diagram illustrating Av = 1 vibrational quantum,
(D) Jablonski diagram illustrating Av = 3 vibrational
quanta. Taken from M. J. Kerkhoff, reference (30).


"() 356,377
"( 375, 399


ex 200
em 206


(C)


Figure 2.







10



is placed on combining techniques and implementing novel sampling

systems for acquiring maximum sample information.

The final goal of this work is to develop and evaluate a system

capable of constant energy synchronous phosphorescence spectrometry.

















CHAPTER 2
THEORETICAL OPTIMIZATION OF PARAMETER SELECTION IN
CONSTANT ENERGY SYNCHRONOUS LUMINESCENCE SPECTROSCOPY

Introduction

The advantages of synchronous techniques include a reduction in

spectral complexity; a reduction in peak bandwidths; and a reduction in

interference by Rayleigh (32) and Raman (33) scatter. Full realization

of these advantages has been limited due to the lack of a thorough

theoretical evaluation and subsequent parameter optimization to aid in

the routine application of these techniques. The characteristics of

the luminescence excitation and emission spectra and their effect on

the selection of scan parameters, and the resultant synchronous

spectral features are considered here.

The prediction of peak wavelength maxima, intensity maxima, and

peak bandwidths in conventional constant wavelength synchronous

luminescence (CWSLS) spectra have been described (34,35). However,

these have included approximations due to the complexity of the

mathematical manipulations necessary when assuming excitation and

emission peaks to be Gaussian in frequency and determining their

response in terms of constant wavelength scanning (i.e. the reciprocal

relationship between wavelength and frequency). The calculations are

less formidable for CESLS and exact relationships can be obtained.

Thus, the spectral advantages of CESLS can be quantitated and utilized

for scan selection, providing for systematic parameter optimization.

These will be presented, along with a discussion of the experimental

consequences of these relationships.

11












Theory

The fundamental spectral parameters of interest in CESLS include

the location of the synchronous peak maximum, its relative intensity

and its bandwidth as functions of luminescence excitation band

characteristics, emission band characteristics, and scan parameters.

Lloyd and Evett (34) obtained the following relationship for the

synchronous peak wavelength maximum, s in CWSLS:
o
-1
Xo = 2XXo (X + Ao- AX) (1)


where X. is the wavelength of maximum luminescence intensity in the
o
excitation spectrum, X. is the wavelength of maximum intensity in the
Jo
luminescence excitation spectrum, and AX is the wavelength difference

in the conventional synchronous (CWSLS) scan mode. An assumption is

made in their treatment that the bandwidths of the excitation and

emission peaks are equivalent when expressed in energy terms.

Satisfactory agreement was observed with experimental data for a number

of compounds over a range of AA values. From the experimental data, an

empirical relationship was derived for the synchronous bandwidth


6Xs = 0.364 (61i + 61X) 3.309 (2)


where 6Xs is the synchronous bandwidth, 6Xi is the luminescence

excitation bandwidth, and 6A. is the luminescence emission bandwidth.

Andre et al. (36) approximated the bandwidths of synchronous peaks as


6XS = 6Xi + 6Xj 6Xi2 + 6j 1/2 (3)


While no derivation was provided for this expression, it seemed to

agree with observed experimental data. It was noted that 6Xs was











always less than 65 confirming the bandwidth reduction observed for

synchronous techniques. Thus, experimental data have suggested the

direction for parameter optimization for synchronous techniques.

For the discussion to follow, luminescence excitation and emission

peaks are represented as single Gaussian functions of frequency

(wavenumbers) or reciprocal wavelength. A range of energy differences

(Avalues) is considered where the scan path does not cross the

excitation-emission peak maxima. The peak location and the bandwidth

of the CESLS peak, vs. and s' respectively, are derived in the

discussion to follow. Appropriate modification is necessary when the

derived expressions are used with other SLS scanning techniques.

Peak Maximum Location and Relative Intensity

To obtain equations which will define the location, vs and
o
intensity, M ijso, of peak maxima in constant energy scans, it is

necessary to establish functions which describe the intensity of these

scans. To determine these functions in terms of the excitation and

emission spectra assuming peaks which are Gaussian in frequency, the

following definitions and substitutions were required.

The terms used throughout this discussion were selected in an

attempt to maintain consistency throughout the literature and are

defined in a Glossary. The general equation defining a Gaussian

distribution can be written as


y = -- 1 exp[-(x j) /2 2] (4)
G (27r)/2

where y is the normalized peak intensity, a is the standard deviation

of the peak, x is an independent variable, and p is the average value











of x over the peak. If the luminescence excitation and emission

spectral band shapes are Gaussian functions, then


xi = x exp[-(7i )2/2, 2 (5)
i Vi ()
o o

and


y = y exp[-(. 7 )2/2a 2] (6)
J o J o J

The relative luminescence measured for a single component sample can be

written as


Mij = a xi j (7)


where a is the product of wavelength independent terms, xi is the

product of excitation wavelength dependent terms, and y. is the product

of emission wavelength dependent terms (37). The conversion from

wavelength (nm) to frequency (cm- ) is written as


V = 10 /I (8)


Substitution of Equations 5 and 6 into 7 yields

2 2 2 2
M. ="ax y exp[-(v. v. )2/2o2 ( 2V ) /2 2(9)
0 0 1 1 Jo

Equation 9 defines the intensity at any excitation and emission

wavelength relative to the maximum peak wavelengths. The wavenumber

difference between the excitation and emission monochromators in a

CESLS scan is defined as:


Av = vi v. (10)


where Av is the constant frequency difference maintained between the











excitation and emission monochromators during the CESLS spectral

measurements. Placing this restriction on Equation 9 gives


Mjs =axi y exp[-(v + Av- )2/2G2 (V- )2/20 2] (11)
00 0 0

This equation defines the synchronous spectrum as a function of only

the emission wavenumber, v., at a selected energy difference, A .

To determine the location of the intensity maximum of the CESLS

scan, M.. the CESLS intensity function is differentiated with
o
respect to emission frequency and set equal to zero. Therefore,


dM.ij. /dj = 0
1Js j


This results in a peak wavenumber maximum, v. of
3so

Vjs = j (0i2/j2) + Vi 0 A)/(l + a2/a2)


Because 6V by definition is related to a by


(13)





(14)


6v = (8 In 2)1/20


Equation 13 can be rewritten as


v. =(v 6 (v/6 2) + v -Av )/(1 + /6 )
0 0 0

Equation 15 defines the location of the peak maximum in a

constant energy scan in terms of the emission wavenumber.

the intensity of the CESLS maximum in a specific constant

is given from Equation 11 by


(15)


given

Therefore,

energy scan


2 2 2 2
Mijs =ax y exp[-(v.s + Av- )2/20 ( ) /20 ] (16)
o0 0 0 0 0 0











CESLS Bandwidths

To calculate the bandwidth of a constant energy peak, 6- one

can determine the location of the peak maximum at js. and the location
o
at which the intensity is half the maximum value. The difference in

these two locations is equal to half the CESLS bandwidth. This goal is

accomplished through the use of Equation 16 and an analogous equation

for the relative intensity at half the maximum intensity

2 2 ( 2 2
M.. = ax y0 exp[-(v + Av i ) /22 (V ) /2G ] (17)
h o o h o i jSh o

Under these conditions


M.. /M.. = 2 (18)


Substitution and simplification results in

2 2 2 2 2 2- 2-
0 = (V v s. )( + + (v. v )(2a.Av 2j vi 20 v. )
jsh 3So i 3 ) ( Jsh jso 3 o i3

2 2
(2i j 2 In 2) (19)


The bandwidth, 6v is introduced and defined as


6s = 2 vjsh Vjs (20)


Solving Equation 19 for 6s gives the CESLS bandwidth

2 2 1/2
6vS = 8 In 2 [a(io/(o + ) 1/2 (21)


and so


6Vs = 6 6vj/(6vi + 6V2)1/2 (22)












Geometric Representation

The total luminescence spectrum provides a clear picture of the

CESLS scanning procedure. This spectrum includes the relative

luminescence intensity contours measured as a function of excitation

and emission wavelengths (38). For this discussion, this spectrum can

be plotted in frequency units, maintaining spectral symmetry (29). For

simplicity, Figure 3 illustrates the total luminescence spectrum of a

single peak, where the ellipse defines the contour equivalent to half

the maximum peak intensity. The following terms are used for this

model:


x = Vi (23)
o
y = v. v. (24)
Jo
2a = 6Sv (25)

2b = 6v (26)


The dashed line represents the CESLS scan path for the special case


Av = v V. (27)


The ellipse can be expressed as

2 2 2 2
x2/a2 y2/b = 1 (28)


or

2 2 2 2 2 2
x b + y a = ab (29)


As Av is varied, a number of scan paths are defined as shown in Figure

4, each with a slope of 1, or:


dy/dx = 1


(30)
























dy/dx = 1


Geometric model for CESLS. The ellipse defines the
contour equivalent to half the maximum peak intensity.
The dashed line defines a CESLS scan with
A = v
0 -0


Figure 3.





























//7/
////
/ / /


\


dy/dx = -b /a
\l
'\

\I
\


Figure 4. Graphical representation of various values of
Av (--) and the line (---) defining the
position of the CESLS maxima for this scan.


/ / dy/dx =

/











These definitions can be used to determine the intensity maximum for

each CESLS spectrum. The maxima fall on a line that runs through the

origin in Figure 4. The equation for this line can be established

using two points, one of these being the origin and the other being

determined by finding one of the two points on the ellipse where the

scan path intersects the ellipse at a single point. While this may not

seem intuitively obvious, the addition of a number of contour lines

will verify this conclusion. Differentiating Equation 29 yields


y = -(b2/a2) x (31)


Equation 31 represents the line defining the intensity maxima for CESLS

scans. Substituting the appropriate CESLS terms into Equation 31 and

solving for v. gives
Jso

2 -2 -2 -2
v. = [v. (v. /6v. ) + v Av]/(l + 6v /v.2) (32)
is 0 1 J 1 0i

Equation 32 represents the location of the CESLS peak maximum in terms

of the emission wavenumber, v. Note that this result agrees with

Equation 15.

Discussion

From these expressions, a number of observations can be made

concerning CESLS spectral characteristics. For illustration, a number

of hypothetical compounds were evaluated. These observations will

provide guidelines for experimental design of CESLS applications.












Peak Wavelength and Intensity Maxima

The shape and location of each CESLS peak is a function of the

intensity maxima of the excitation and emission bands and the selected

Av Thus, for a specific compound, the synchronous spectrum intensity

maximum and location varies only with the scan parameter, Av. The peak

characteristics of three hypothetical compounds are shown in Table I.

For each compound, Av is varied to compare the peak maximum locations,

relative intensities, and peak bandwidths. The values of Av were

selected to coincide with vi V and reasonable variation around
o o
this value. Equation 15 suggests that if AK is chosen to coincide with

v v then v s0 is equivalent to J. That is, if the scan path
o o o o
passes through the peak maximum, the synchronous and emission peak

maxima are identical. If Av is less than vi v, then vs is
S0 0
greater than v (thus shorter in wavelength terms), and the scan
Jo
crosses the x-axis in Figure 4 to the left of the origin. If Av is

greater than v v, then v. is less than v. and the scan
i 0 s 30
o 3o So
crosses the x-axis in Figure 4 to the right of the origin. If Equation

13 is differentiated, then


dv. /dAv = -1/(1 + 2 /a ) (33)
so J

This defines the slope of the line of CESLS maxima. This line is also

depicted in Figure 4.

The three compounds in Table I represent three cases for relative

values of ai and a.. These are


1. a. = a.

2. O. << .

3. oi >> a.
1 j












Table I. Spectral Data for Three Model Compounds

Compound A


- -1
v (cm )
Jso


28,500

28,000

27,500

27,000

26,500


- -1
6v (cm )




1,061

1,061

1,061

1,061
1,061


Compound B


-js (cm )
o __ _


29,100

28,300

27,500

26,700

25,900


- -1
6vs (cm )
S


Compound C


,js (cm )


- 1
6v (cm )


27,900

27,700

27,500

27,300

27,100


Av(cm )



13,000

14,000

15,000

16,000

17,000


M js/M
o


0.641

0.895

1.000

0.895

0.641


- -1
Av (cm )



13,000

14,000

15,000

16,000

17,000


Mijs /M
0


0.491

0.837

1.000

0.837

0.491


AV (-1
Av (cm )



13,000

14,000

15,000

16,000

17,000


Mijs /Mo



0.491

0.837

1.000

0.837

0.491












Table I. Continued
*


- -1



v (cm )
jo

(C-1
0 (cm-1)
i

0 (cm )
j


Compound A


42,500


27,500


1,500


1,500


Compound B


42,500


27,500


750


1,500


Compound C


42,500


27,500


1,500


750











The total luminescence spectra for these cases are illustrated in

Figure 5. Case 1 is representative of the solution fluorescence

spectra of the majority of common fluorescent compounds. If = .,

then Equation 13 reduces to


v = (Jo + v-i Av)/2 (34)


From Equation 33, dv. /dAv reduces to -1/2. The data for compound A
3so
in Table I confirms these relationships. For Case 2, v. approaches
J so
v v.. That is, the synchronous maximum is proportional to Av ,
o 3o
representing the conditions for maximum change in .js as Av is varied;

also from Equation 33, dv. /dAv approaches -1. For Case 3, v.
Jso so
approaches v That is, the synchronous maximum does not deviate
1o
significantly from the emission maximum as Av is varied. This is

readily apparent from Figure 5C, and is further proven by Equation

33, since dv /dAv approaches 0. These three cases are summarized in
Js0
Table II.

CESLS Bandwidths

Several observations can be made about CESLS bandwidths. Note

that from Equation 22, 6vs is independent of Av This is also

confirmed by the data in Table I. This is graphically illustrated in

Figure 6 where five CESLS spectra are shown with varying values of Av

The bandwidth at half maximum is identical for each peak. In

conventional synchronous (CWSLS) techniques, it has been observed that

6xs is dependent upon AX (36). The three cases previously described

are also important in this discussion. For Case 1, ci = G1 and 6X s

is equivalent to 6 ./2. This effect is clearly seen in the example














































0


S-4



.-IA



A
*f4 *




u .J.I
-U 0
0) A








0


4J
P4 V
v








C,.
0
--4



Ia 0











LrX
L











Table II.


Case I


(oai = a j)


Limiting Cases


Case 2


(i << j)


-1/2


(- + Av)/2
O0 O


Vi A
o
0


o / /2



6I/ /2
J


dv. i/dA
^o


Case 3


(ai >> j)
---------


Sjs0
O o


vJo


6v

























/* / \\\\

c... / / \ \ \,\,



20000 25000 30000 35000
Energy (cm-1)




Figure 6. Comparison of peaks generated by using a variety of
Av values (cm ) for compound A. See Table II.
Av = 15,000 (-----), 14,000 (- -), 13,000 (---),
16,000 (- *-), and 17,000 (-- -).











shown in Figure 7. For Case 2, a << and 6v approaches6v as this
i j S
difference increases. An example of this is shown in Figure 8 where

a. = 0./2. Correspondingly for Case 3, a >> and 6v approaches
i J 3 s
6V. An example of this is shown in Figure 9 where ao = 2o These

two cases have significant experimental consequences. This suggests

that the CESLS bandwidth corresponds to the narrower of the excitation

and emission bandwidths as one becomes much larger than the other.

Therefore, 6Vs is theoretically always less than or equal to 6v .

CESLS vs CWSLS

For comparison of CESLS and CWSLS two additional hypothetical

compounds are evaluated. Relevant spectral characteristics for these

compounds, D and E, are given in Table III. For demonstration

purposes, peak locations (v values) were chosen based on molecular

theory, which shows that compounds within a class (such as PAHs)

generally exhibit similar vibrational energy separations, and

consequently comparable Av values. Figure 10 compares the
max

excitation and emission spectra for D and E to the synchronous spectra

obtained using a constant energy difference equal to Avmax for both

compounds. Figure 11 demonstrates the change in appearance of spectra

presented in Figure 10 when they are plotted in terms of signal vs

wavelength rather than energy.

Figure 12 shows the result obtained if a constant wavelength

difference is scanned with AA = 85.227 nm (this is A max for compound

D). Figure 13 shows the result obtained if a constant wavelength

difference is scanned with AA = 121.457 nm (this is A max for compound

E). Figure 14 shows on an expanded scale comparison of plots defined

























CI




i I
I I

/ A / \


/ \ /

20000 30000 40000 50000
Energy (cm-1)




Figure 7. Excitation, A (-- -) and emission, A (- )
peaks for compound A. o. = o.. The CESLS spectra,
A (--- -) is shown for the scan path
s
Av = v. .
o 0



























1I I






I s / I




20000 30000 40000 50000
Energy (cm-1)





Figure 8. Excitation, B (- -) and emission, B (- )
x m
peaks for compound B. a = a./2. The CESLS spectrum,
Bs, (-- -) is shown for the scan path
Av = -. .
0 Jo



























i-"I
I I t
C0


I I





/ \

c \ / \
| s :


20000 30000 40000 50000
Energy (cm-1)





Figure 9. Excitation, C (- -) and emission, C (- )
x x
peaks for compound C. o. = 2 o.. The CESLS spectrum,
Cs, (- -) is shown for the scan path
S= v. .
o 0












Table III. Spectral Data for Two Model Compounds


Compound D Compound E


-- -1
Vex (cm ) 44,000 38,000
o

7em (cm ) 32,000 26,000
o
-1
Avmax (cm ) 12,000 12,000


ex (nm) 227.273 263.158
o

Xem (nm) 312.5 384.6
o

Amax (nm) 85.227 121.457


0ex (cm ) 1,000 1,000

Gem (cm-1 1000 1,00
o (cm ) 1,000 1,000
em


















































Figure 10.


30000 40000 50000
Energy (cm-1)





Excitation (- -) and emission (- ) peaks for
model compounds D (D ,D ) and E (E ,E ). See Table
x m x m
III. CESLS spectra (-- -) with Av = Amv for
max
both compounds (D and E ), arbitrarily plotted in
terms of excitation energy.
terms of excitation energy.









































200


Figure 11.


300 400 500
Wavelength (nm)





Peaks described in Figure 10, with x-axis plotted
in units of wavelength (nm).























C
c H
I






I X
is


fDx \H i

200






Figure 12.


300 400 500
Wavelength (nm)





Excitation (- -) and emission (- ) peaks
for compounds D and E as described in Table III.
CWSLS spectrum with AX = AX for compound D
(85.227 m) (----max
(85.227 um) (----).





























C1
0 1












200I





Figure 13.


300 400 500
Wavelength (nm)




Excitation (- -) and emission (- ) peaks for
compounds D and E. CWSLS spectrum with AX = AX
max
for compound E (121.457 nm) (--- -).































-q
0














200








Figure 14.


300


Wavelength (nm)


Scans defined by Av = 12,000 cm ( ),
AX = 85.227 nm (----), and AX = 121.457 nm
(- -). For compounds D and E as described
in Table III.











-- -l
by Av = 12,000 cm AA = 85.227 nm, and AX = 121.457 nm, for

hypothetical compounds D and E.

Figures 10 through 14 demonstrate the superior ability of CESLS

over CWSLS to exploit the natural energy relationships in luminescence

spectrometry, i.e. peak shapes Gaussian in energy and equivalent

vibrational energy spacings exhibited by compounds within a given

class.

Experimental Design

Room temperature fluorescence and phosphorescence spectra

generally fall within the Case 1 category where oi = o For these

spectra, monochromator bandpasses are selected to limit instrumental

broadening of the luminescence bands. A reduction of the spectral

bandwidth by a factor of 2 can be advantageous. Cases 2 and 3

describe those systems where significant band-narrowing is observed in

the luminescence excitation and emission spectra, such as fluorescence

observed at reduced temperatures (39,40). The combination of

low temperature spectral band narrowing techniques with synchronous

scanning provides extensive application of the selectivity advantages

described for synchronous techniques. Experimental parameter selection

for these combinations requires a careful evaluation of the fundamental

relationships as they have been presented above.

The major consideration for synchronous scanning of narrow peaks

include insuring that the scan path crosses the peak of interest;

providing sufficient light throughput for detection; and maintaining

the spectral features provided by the band-narrowing technique. All

three objectives are met by maintaining the bandpass of one

monochromator at a level where the bandwidth is not instrumentally











broadened and by increasing the bandpass of the other monochromator to

a significantly larger value. The narrow bandpass is maintained on the

monochromator where narrower bands are observed in the excitation and

emission spectra. While a thorough evaluation of experimental

parameters was not available at the time, reference 13 demonstrated the

effectiveness of this approach. If the excitation and emission spectra

are comparable in bandwidth, increasing the excitation monochromator is

recommended, forcing the system into the Case 3 category. Case 3

offers several advantages for this application. The CESLS maximum

approaches the emission peak maximum, providing a means of peak

comparison for identification. Note that the CESLS maximum is

independent of Av Also 6v is then approximately equal to 6v By
s J
increasing the excitation monochromator bandpass, sensitivity is

increased for adequate detection. Finally, the luminescence intensity

of the synchronous peak relative to the maximum peak intensity is

maintained over a wider range of scan values for Av using these

experimental conditions. Figure 5 demonstrates that a Case 3 peak has

a greater probability of being included in a synchronous scan than a

Case 1 peak without instrumental broadening.

Conclusions

The CESLS parameters of peak wavelength maximum, intensity

maximum, and peak bandwidth have been evaluated, with mathematical

relationships derived for each parameter as it is affected by the

luminescence excitation and emission spectral characteristics and the

scan path. These relationships are solved without approximation based

on the initial assumption of single Gaussian peakshapes. Graphical

representations of CESLS scans confirm these relationships, with












hypothetical compounds used to demonstrate their application.

Calculations are significantly influenced by the ratio of the

excitation and emission bandwidths, resulting in three special cases

used for detailed discussion. Observations from three cases were

translated into experimental design considerations for CESLS

applications. This provides a parameter optimization strategy based on

fundamental spectral characteristics rather than empiracal

determinations. As a result, the combination of synchronous scanning

with other luminescence techniques should implement the selectivity

advantages.

















CHAPTER 3
ANALYSIS OF ENVIRONMENTAL SAMPLES CONTAINING
POLYCYCLIC AROMATIC HYDROCARBONS

Introduction

Polycyclic aromatic hydrocarbons are formed during incomplete

combustion of fuel in an engine. The concentration of chemical species

in vehicle exhaust is dependent upon factors such as engine type, oil

and fuel consumption, and operating conditions for the engine. PAHs

have also been found to occur naturally in crude oil and gasoline

samples (41). The number and quantity of PAHs in such materials

depends on a number of variables, including: degree of maturity,

temperature fractionation, sampling, and processing parameters. Fuels

and exhausts are merely examples of sources of PAHs. The reader is

referred to Grimmer (41), Cooke and Dennis (42), and Jacob et al.

(43,44) for comprehensive reviews concerning the occurrence and toxicit

of PAHs found in the environment.

Due to the documented carcinogenic and mutagenic activity of many

PAHs, there is an obvious need for a simple, sensitive, and reliable

method for determining these compounds. Also, a simple method capable

of fingerprinting a sample based on the PAH content could be used to

identify the source of an environmental hazard, (i.e. an oil spill,

gasoline leak, etc.). Such a method could also be used to study the

metabolism of oil products by marine species, to aid in forensic

analysis, and to evaluate systems proposed for reducing PAH production

and emission by engines.












Widely employed techniques for analysis of samples containing PAHs

include MS/MS (45), GC/MS (46), and luminescence techniques (1-6).

Conventional luminescence measurements are generally very sensitive due

to the high quantum yields exhibited by PAHs; however, they offer

limited application to analysis of complex mixtures containing PAHs due

to the broad, overlapping spectral characteristics exhibited by many of

these compounds. Methods often used for enhancing the selectivity of

luminescence techniques for mixture analysis were outlined in Chapter

1. Constant Energy Synchronous Luminescence Spectrometry (CESLS) is

among the more successful techniques used for selectivity enhancement.

CESLS has been shown to be useful for identifying and quantitating PAHs

(29,39) in mixtures. To further enhance selectivity CESLS has been

carried out in a rapid scan mode for HPLC detection (47). This system

showed promise and would be especially useful for extremely complex

samples. Based on the success achieved for PAH analysis in the past,

CESLS is proposed for fingerprinting gasoline and oil samples as well

as gasoline engine exhaust particles.

Experimental

Instrumentation

The experimental setup used for this study is shown in Figure 15.

Components and manufacturers are listed in Table IV. The scan rate was

approximately 50 nm/min allowing a constant energy scan with an

excitation range from 200 to 500 nm to be collected in less than 6 min.

The excitation monochromator was pulsed at a constant rate while the

emission monochromator was pulsed at a variable and faster rate to

maintain the desired constant energy difference. Scan rates were

controlled by an Apple II plus microcomputer.














































4)
U, 0.
CE

4) 0
U,


43

















4J
U

























0..
Ci


U,
U4





















C-1
0
0
41
0

4-W
4.4
0)
Co


4i


1-4
0)
0.






















-Il


00-












Experimental Components for CESLS.



Model Manufacturer


Xenon Arc Lamp, 300W


Illuminator Power Supply



Excitation and Emission
Monochromators f/3.5
holographic grating 1200
grooves/mm

Monochromator Scan Controls


Photomultiplier


High Voltage Power Supply
operated at 800V

Current Voltage Converter/
Amplifier TC 300msec



Recorder


Microcomputer


VIX-300UV


PS300-1



EU700-56




EU700-51


IP 28


EU701-30


427




Omniscribe


Apple II plus


EIMAC, Division of
Varian, San Carlos
CA 94070

EIMAC, Division of
Varian, San Carlos


GCA McPherson
Instruments, Acton
MA


Heath Co., Benton
Harbor, MI 49022

Hamamatsu, Waltha
MA 02154

Heath Co., Benton
Harbor, MI 49022

Keithley Instru-
ment Inc. 28775
Cleveland, OH
44139

Houston Instrument
Austin, TX 78753


Apple Computer
Cuportino, CA
95014


Table IV.


Equipment












Sample Analysis

Measurements at low temperature (77 K) were obtained using quartz

tubes immersed in a liquid nitrogen dewar. Room temperature

measurements were made on filter paper (Fisher Scientific Company,

Pittsburgh, PA) placed at a 45 angle with respect to the excitation and

emission monochromators. The filter paper was held in position using a

device similar to one described previously (48). The holder resembles

an aluminum cuvette with the center cut at a 45 angle and has a metal

plate with a hole in the center to hold the filter paper in place. The

metal cuvette was attached to a cover designed to fit over the liquid

nitrogen dewar holder. This design allowed us to obtain spectra at

room temperature and low temperature with a minimum of alteration in

the system. Also, both methods required less than 0.5 mL of sample

volume for analysis.

PAH Mixture Analysis

To determine the suitability of CESLS at low temperature (77 K)

for PAH mixture analysis, a 16 component mixture (Chem Service,

Westchester, PA) was used. All of these compounds are on the EPA

priority pollutant list.

The mixture was first scanned at a variety of constant energy

differences ( Av values). The Av values were chosen based on

preliminary work determining optimum Av values for each compound

calculated from excitation and emission spectra (49). The values used
-- -1
for this study included Av = 1400, 2800, and 4800 cm The small

number of values necessary for multi-component identification agrees

with molecular luminescence theory which shows that compounds within a

class (such as PAHs) generally exhibit similar vibrational energy level











spacings and consequently comparable Av values. After the mixture was

scanned, a library of scans was established containing each of the 16

compounds (obtained from Foxboro Analabs, North Haven, CT, and the EPA

Repository, Cincinnati, OH) at the different values. This library was

used to identify the peaks in the scans of the mixture. Standard

additions were also used when identification was difficult. Results of

two of the scans obtained after dilution to a concentration of 200 ppb

in each component (with peak identification) are shown in Figures 16

and 17. Thirteen of the compounds in the 16 component mixture are

identified in the two scans used for demonstration purposes. The other

three compounds were dibenz(a,h)anthracence (which was identifiable in
-I
the 2800 cm scan), acenaphthylene, and indeno(l,2,3-cd)pyrene. The

last two were not identifiable at the sample concentration chosen for

the demonstration scans. These scans were later used for comparison to

unknown mixtures for preliminary identification of components.

Gasoline Engine Exhaust Analysis

Sampling Methodology

The samples in this study were collected from the exhaust of a

small 4-cycle 1975 model #21 Toro lawn mower. One of the first

considerations in identifying and quantitating the components of the

exhaust of an engine is the choice of the sampling system. The number

and quantity of PAHs measured at a sampling site shows a strong

dependence on the sampling methodology (50-54). Two different sampling

systems were used in this study. In the first system (outlined in

Figure 18), the exhaust was routed through a short (20 cm) flexible

metal tube into an open glass tube 120 cm long for 30 min. To

determine the effect of sampling location on PAH content, sections 3 cm














-J L7= 1400

0
a


0 -
-I



I a
0

Cu



a
1 (0

a "
01
U 3%
A N
a
0: 1
AS N
0 & a
0 0
0 a
0 *


00 EXCIT ATION


Figure 16.


WAVELENGTH


CESLS scan of 16 component PAH mixture

AV = 1400 cm-
Avl= 400 cm


soo















SAV= 4800


4)
a
0
1.42

9
M SI







A c
O
0. |
z0 I


0
C:
0


JC
A

cc







M
-4


4O
0
14
0
a


I
m
J
o


Fll

U-l


EXCIT ATION


WAVELENGTH


Figure 17. CESLS scan of 16 component PAH mixture

Av = 4800 cm .


200


500











in length were cut from the tube. Three sections were taken; one from

the end connected to the metal tube (a), one from the center (b), and

one from the open end (c). Although the temperature was not calibrated

along the length of the tube, it was expected that this sampling

procedure would be equivalent to a crude chromatographic separation

based on the volatility of the compounds. The 3 cm sections labelled

a, b, and c were each extracted with 25 mL of hexane (obtained from

Burdick and Jackson). The extraction process involved immersing the

3 cm sections in hexane contained in amber glass vials and placing them

in an ultrasonic bath at room temperature for 30 min. Constant energy

scans were then obtained for these samples. The second system for

sample collection was designed for qualitative comparison of PAHs

produced by different samples. The system involved routing the exhaust

into an open 120 cm glass U-tube (i.d. 8 mm o.d. 12 mm) placed in a

dewar containing liquid nitrogen. Samples were collected for 6 min

and subsequently extracted into 25 mL of hexane by repeated rinsing.

The samples included exhaust collected from two different brands of

super and regular unleaded and one brand of regular leaded gasoline.

The two brands of super and regular unleaded were chosen because

although they have comparable octane ratings only one of them contains

ethanol.

Results and Discussion

Figures 18 and 19 show results obtained for exhaust samples

collected using the first system described. The A- values chosen for
-I
demonstration purposes were 1400 and 2800 cm The exhaust samples

shown here were from regular leaded gasoline and the separation results

were typical of other gasoline samples which were measured with the





















































500


EXCITATION


Figure 18.


WAVELENGTH (nm)


CESLS scan of leaded gasoline exhaust extracts
obtained using multiple sample collection
system (Av = 1400 cm- ). Compounds identified
include benzo(a)pyrene (BAP), benzo(k)fluoranthene
(BKF), anthracene (A), coronene (COR), and
perylene (PER).





















































EXCITATION WAVELENGTH (nm)


Figure 19.


CESLS scan of leaded gasoline exhaust extracts
obtained using multiple sample collection system
(Av = 2800 cm-1). Compounds identified include
phenanthrene (PHE), chrysene (CHR), pyrene (PYR),
anthracene (A), benzo(a)pyrene (BAP), benzo(k)
fluoranthene (BKF), and anthanthrene (ANT).


200


500












same sampling system. Results indicate that section a which would be

the hottest since it is closest to the engine is too hot to allow

significant condensation of any of the PAHs. (Spectra of section a are

typical of the background obtained with hexane alone.) The next

sections indicate condensation of PAHs in a trend which agrees with

their boiling points (41). Identification of PAHs was made through

comparison to standard spectra at all AV values and confirmed by

standard additions. This experiment confirms previous studies which

indicate that the results obtained from a particular sampling system

may not be indicative of the total PAH content of an exhaust. It also

demonstrates the advantage of combining CESLS with a preliminary

separation, when one wishes to analyze an especially complex mixture to

obtain further information about a sample. Of course, a more elegant

separation technique would enhance the information content at the

expense of time. Another possibility for information enhancement would

be the use of mathematical techniques such as factor analysis.

Figures 20 and 21 show scans of the extracts from the U-tube

samples collected for the exhaust of the five different gasolines

analyzed in the second system. The Av values chosen for demonstration
-i
purposes were 1400 and 4800 cm No attempt was made in this study to

label peaks in Figures 20 and 21 because of the confusion which would

arise with the need for multiple labeling of individual peaks.

However, this system demonstrates the sensitivity of CESLS for PAH

analysis, especially for fingerprinting and screening. The results of

the study indicate comparable PAH production from different gasoline

samples with slight differences in relative concentrations of

particular species, such as pyrene, perylene, and acenapthene.




















































200

EXCITATION


500


WAVELENGTH (nm)


Figure 20.


CESLS scan of exhaust from different gasoline
samples obtained using a U-tube collection system
(Av = 1400 cm- ) Samples: (a) Brand A Super
Unleaded, (b) Brand A Unleaded, (c) Brand B
Super Unleaded with Ethanol, and (d) Brand B
Unleaded with Ethanol, and (e) Brand C Regular.













-1
Av= 4800 cm














a


200
EXCITATION


500
WAVELENGTH (nm)


Figure 21.


CESLS scan of exhaust from different gasoline
samples obtained using a U-tube collection
system (Av = 4800 cm- ). Samples: (a) Brand A
Super Unleaded, (c) Brand B Super Unleaded with
Ethanol, and (e) Brand C Regular.












Tests to determine the efficacy of the extraction processes and

sampling systems were not performed because the intent of this study

was to demonstrate the sensitivity and selectivity of CESLS and not

to propose a new sampling method. Also, the engine used in this study

could not be optimized and reproducibly run to obtain controls and

therefore prohibited the possibility of obtaining data necessary for a

meaningful efficacy analysis.

The difference in the appearances of spectra obtained with the two

different methods (Figures 18 and 19 compared to 20 and 21) can be

attributed to the sampling procedures. In the second system the

extracts contain all PAHs collected over the 120 cm length of tube

while in the first system each extract contains only the PAHs condensed

in a 3 cm interval of the tube.

Possible applications of this method include preliminary screening

to evaluate systems proposed for reducing PAH emission and

fingerprinting of samples to identify their origin.

Gasoline and Crude Oil Fingerprinting

Sampling Methodology

Gasoline samples were obtained from various service stations in

the Gainesville, FL, area. They were stored at 4C in amber glass

vials prior to analysis. Crude oil samples were obtained from the

Stazione Sperimentale dei Combustible (Milano, Italy) through the

Joint Research Centre of European Communities (Ispra, Italy). To

achieve analyte concentrations within the linear dynamic range of the

method, samples were diluted with hexane. For the gasolines and light

crude oil fractions, a direct dilution was performed. For the heavier,

more viscous crude oil fractions, samples were weighed and then












extracted by 25 mL of hexane in an ultrasonic bath. These samples were

then centrifuged and spectra were obtained for the supernatant. When

necessary, these samples were also further diluted with hexane.

Dilution requirements were determined by first running the samples

directly and then diluting and scanning until no significant spectral

changes were observed. Once dilution requirements were established

for the most complex, concentrated samples with the different methods

all fingerprints were run at this concentration. However, dilutions

for gasolines and crude oils were determined separately.

Results and Discussion

Figures 22 and 23 demonstrate the ability of CESLS to identify

several PAHs within gasoline samples. These scans were obtained at
-i
low temperature with energy differences of 1400 and 4800 cm ,

respectively. The synthetic mixtures were matched by comparing the

unknown to a library of scans obtained for pure PAHs. Figures 22 and

23 also indicate the advantage of combining scans at different
-I
values. For the 1400 cm scan, anthracene (A) and benzo(a)pyrene
-I
(BAP) are easily identified, and the 4800 cm scan allows

identification of pyrene (P) and fluoranthene (FLU).
-i
Figures 24 through 27 contain scans obtained with Av = 1400 cm -

They demonstrate the fingerprinting capabilities of CESLS as well as

contrast room temperature versus low temperature measurements. Figures

24 and 25 show scans obtained for three different gasolines at 77 K and

298 K, respectively. Although brand names will not be given, it is

especially interesting to note major spectral differences for Brand E

super unleaded and regular gasoline (scans a and b).
















































EXCITATION


Figure 22.


WAVELENGTH


-1
CESLS scans at 77 K and 1400 cm for a) Brand D gasoline
(1:10 dilution), b) mixture containing benzo(a)pyrene
(BAP) and anthracene (A), and c) mixture containing
pyrene (PYR) and fluoranthene (FLU).











AV = 4800


a- 1:10 gasoline

b- BAP,A
c- PYR,FLU


Figure 23.


EXCITATION WAVELENGTH
-1
CESLS scans at 77 K and 4800 cm for samples identified
in Figure 22.


















































Figure 24.


\b


EXCITATION WAVELENGTH
CESLS scans at 77 K and 1400 cm-I for
1:50 dilutions of a) Brand E Super
Unleaded, b) Brand E Regular, and
c) Brand F Regular.





















































EXCITATION WAVELENGTH


Figure 25.


CESLS scans at 298 K and 1400 cm-1 for
samples identified in Figure 24 (no dilutions
required).











Scans of Brand G regular and diesel obtained at 77 K and 298 K are

shown in Figures 26 and 27, respectively. Once again the sensitivity

and selectivity of CESLS are demonstrated. Also, these figures

illustrate the feasibility of using CESLS to identify the source of an

environmental hazard, for example a leak in a tank at a service

station.

Figures 28 and 29 compare scans of selected fractions of crude

oil number one at 77 K and 298 K, respectively. Scans are for

fractions collected by distillation a) below 350C, b) between 160

and 240C, and c) between 240 and 350 C. Scans of the same

temperature fractions of crude oil number two are presented at 77 K,

Figure 30, and 298 K, Figure 31.

The spectra presented here are representative of results obtained

for many different gasoline and crude oil samples which were analyzed.

The different dilution requirements for the sample types (i.e.,

gasolines and crude oil fractions) and the methods (i.e., low

temperature and room temperature) can be explained by variations in

a) concentration of compounds contributing to the fluorescence

fingerprint, b) sample matrices, and c) increase in quantum yields of

fluorescing compounds at low temperature.

The room temperature scans on filter paper often yielded enough

information for conclusive sample identification and are simpler to

obtain than low temperature measurements, which require liquid nitrogen

and dewars with quartz windows. These scans would be especially

suited for distinguishing between known samples. The room temperature

scans could also be used for preliminary sample screening and to

provide complimentary information to that obtained at low temperature.






















































Figure 26.


b





EXCITATION WAVELENGTH
-1
CESLS scans at 77 K and 1400 cm for
1:50 dilutions of a) Brand G Diesel and
b) Brand G Regular Gasoline.























































Figure 27.


EXCITATION WAVELENGTH

-1
CESLS scans at 298 K and 1400 cm- for
samples identified in Figure 26 (no
dilutions required).




















































EXCITATION WAVELENGTH


Figure 28.


-i
CESLS scans at 77 K and 1400 cm for
1:1000 dilutions of crude oil number one
distillation fractions a) below 3500C,
b) 160 to 2400C, and c) 240 to 350C.



















































EXCITATION WAVELENGTH


Figure 29.


-I
CESLS scans at 298 K and 1400 cm for
1:10 dilutions of crude oil number one
fractions described in Figure 28.
























































Figure 30.


EXCITATION WAVELENGTH

CESLS scans at 77 K and 1400 cm-I for 1:1000 dilutions
of crude oil number two fractions a) below 3500C,
b) 160 to 2400C, and c) 240 to 3500C.




















































EXCITATION WAVELENGTH


Figure 31.


-i
CESLS scans at 298 K and 1400 cm-1 for 1:10
dilutions of crude oil number two fractions
described in Figure 30.











Measurements made at low temperature provide greater spectral

resolution, and it has been demonstrated that this enhancement in

resolution can allow identification of specific compounds and provide

information necessary for distinguishing between samples which are very

similar.

Conclusions

The sensitivity and selectivity of CESLS for fingerprinting and

identifying PAHs in environmental samples has been demonstrated.

Reproducibility for all samples and methods was found to be excellent.

CESLS is an inexpensive, simple, and reliable method, which is expected

to have a wide range of applications for complex mixture analysis in

the future.

















CHAPTER 4
FEASIBILITY STUDY OF
CONSTANT ENERGY SYNCHRONOUS LUMINESCENCE SPECTROMETRY
FOR PESTICIDE DETERMINATION

Introduction

Carbamate pesticides are widely used to protect plants from

insects (55). Due to the toxicity of these compounds and the

possibility of residual presence in the environment and crops, there is

an obvious need for a sensitive and reliable method for determining

them. A method for distinguishing pesticides from their hydrolysis

products and/or metabolites could also have a wide variety of

applications for studies involving optimizing application

concentrations and times, as well as metabolism of pesticides by

insects, animals, and humans.

Pesticide determination has been done by a wide variety of

methods ranging from liquid chromatography followed by mass

spectrometry (56) to room temperature phosphorescence (7), UV detection

(8), and a variety of fluorescence techniques (9-12). Fluorescence

techniques are among the most sensitive because of the high

fluorescence quantum yields of many pesticides (12). Generally

luminescence detection follows separation methods such as HPLC (9) or

TLC (11). Separation is usually required due to overlap in

conventional luminescence spectra of the compounds of interest.

The many advantages of CESLS over conventional luminescence

measurements were evaluated theoretically in Chapter 2. CESLS was

applied to determination of PAHs in Chapter 3 with very good results.

69











Based on previous success with mixture analysis, it was decided to

extend the application of CESLS to the determination of pesticides.

The enhancement in selectivity achieved with CESLS at low

temperature over conventional luminescence measurements will be

demonstrated in this article. It will be shown that often CESLS allows

analysis of compounds in a mixture without prior separation. This is a

very attractive possibility since separation techniques generally

increase the cost and time required for each analysis.

For this study carbaryl, naphthol (a hydrolysis product of

carbaryl), and carbofuran were evaluated with respect to

identification, limits of detection, and linear dynamic ranges.

Bandpasses and constant energy differences were varied to evaluate

their effect upon spectra obtained.


Experimental Section

Instrumentation

The experimental setup used for this study was described in

Chapter 2 and shown in Figure 15. For this study spectral bandpasses

were varied from 1.5 to 4 nm. Constant energy scans were obtained with

an excitation range from 250 to 350 nm in approximately 2 min. Spectra

were taken at low temperature (77 K) through the use of a liquid

nitrogen dewar system, and round quartz sample tubes (3 mm i.d. and

5 mm o.d.).

Sample Preparation

Stock solutions of 100 ppm carbaryl, naphthol, and carbofuran

(obtained from EPA, Research Triangle Park, NC) were made in reagent

grade ethanol (obtained from Aaper Alcohol & Chemical Co., Shelbyville,

KY). These stock solutions were subsequently diluted in ethanol to











form samples for calibration curves and to obtain spectra for

demonstration purposes.


Results and Discussion

First, excitation and emission scans were obtained for the three

compounds. Carbofuran luminescence scans yielded excitation peaks at

234 and 284 nm, and one broad emission peak centered at 299 nm. The

width of the emission peak allowed determination at a wide range of

energy differences, while the position of these peaks prevented overlap

with constant energy scans of naphthol and carbaryl. Figures 32 and 33

show spectra for carbaryl and naphthol, respectively. Spectra were

taken at low temperature and were not corrected for instrumental

response. These spectra will be used here to demonstrate the results

one can expect with constant energy scanning and variations in

bandpasses.

From conventional luminescence excitation and emission spectra of

the three compounds, tables of possible energy differences between each

excitation and emission pair were calculated using the expression

previously given for AV.

Tables V and VI contain results of these calculations for carbaryl

and napthol respectively. Based on these calculations, constant energy

scans were obtained at a variety of energy differences. For example,

naphthol has three transitions with an overall vibrational energy loss

close to 1400 cm-1 (see Table VI). Figure 34 shows the spectrum of

napthol obtained by scanning with Av= 1400 cm If one then compares

both tables for naphthol and carbaryl (Table V), it can be seen that an

energy difference of 2650 cm-1 is a good compromise for identification
energy difference of 2650 cm is a good compromise for identification



















Z
zLU V


0



C_)













WAVELENGTH (nm)

Figure 32. Excitation and emission scans of carbaryl.
Prominent peaks are listed in Table V.












































WAVELENGTH (nm)


Figure 33.


Excitation and emission scans of naphthol.
Prominent peaks are listed in Table VI.














-I
Table V. Calculated Energy Transitions (in cm ) for Carbaryl.





Emission
peaks (nm) 315.7 320.7 330.5 335.5


Excitation
peaks (nm)

273.2 4926 5413 6347 6792


283.1 3650 4137 5070 5515


294.3 2310 2797 3730 4176













-i
Table VI. Calculated Energy Transitions (in cm ) for Naphthol.





Emission
peaks (nm) 326.4 331.6 341.7 347.5 358.2


Excitation
peaks (nm)

304.6 2193 2673 3565 4053 4913


311.8 1435 1915 2806 3295 4154


318.3 780 1260 2152 2640 3500


325.7 66 547 1438 1927 2786




















IU
U





0


F-
LL

LQ



o-










__.__


WAVELENGTH (nm)

Figure 34. Constant energy scan of naphthol with
Av = 1400 cm Excitation wavelength
range 250-350 nm. Excitation and
emission bandpasses 1.5 nm.










of both compounds. As mentioned earlier, carbofuran could be

determined at a wide range of constant energy differences. This fact

is illustrated in Figure 35 which includes scans of a mixture of 4 ppm

carbaryl, 4 ppm carbofuran, and 2 ppm naphthol at Av = 1400 and 2650
-1
cm Also, this figure demonstrates the advantage of using a

combination of constant energy scans for confirmation of

identification. Peaks used for identification and limit of detection

measurements are labeled. These scans were obtained with 1.5 nm

bandpasses on the excitation and emission monochromators.

Concentrations were calculated based on peak heights and comparison to

standard calibration curves. The mixture was run in triplicate and

peak heights were averaged to minimize variation in signal intensity

due to effects such as solvent cracking and changes in sample tube

alignment. Values determined in this manner gave results within 13% of

the true value for all three compounds. In conventional luminescence,

the accuracy and precision of such measurements is dependent on the

relative concentrations of the compounds being quantitated, due to

overlap in spectral characteristics. The excellent resolution obtained

with CESLS minimizes this problem.

Theory presented in Chapter 2 indicated that it should be possible

to maintain good resolution while increasing signals by opening either

the excitation or emission monochromator and keeping the other narrow.

Therefore, it was decided to evaluate these same compounds under

varying conditions. Table VII gives limits of detection obtained for

the three compounds at a variety of bandpasses and energy differences.

Analytically useful ranges were approximately 3.5 orders of magnitude

for the three compounds under all conditions, and included the limits



















Ai = 1400 = 2650




0 34 -
-7 a
0. .0 ^
0o- -
z '- Q.
I 0 0
1 b,
z


WAVELENGTH (nm)


Figure 35.


Constant energy scans of a mixture of carbaryl,

naphthol, and carbofuran with Av = 1400 and
-1
2650 cm Excitation and emission bandpasses

1.5 nm.













Table VII.


Compound


Carbaryl







Naphthol










Carbofuran


Limits of Detection for Pesticides with Variable
Parameters.


-1
(cm )


2650

2650

2650



1400

2650

2650

2650



1400

2650

2650

2650


bandpasses


LOD (ppb)


1.5,1.5

4.0,1.5

1.5,4.0


1.5,1.5

1.5,1.5

4.0,1.5

1.5,4.0


1.5,1.5

1.5,1.5

4.0,1.5

1.5,4.0


a Bandpasses for excitation, emission monochromators in nm.



Note: Analytically useful ranges were approximately 3.5 orders of
magnitude extending from the limit of detections for all
conditions and compounds.











of detection. Figure 36 shows the effect on the constant energy

spectra of naphthol when a) opening the excitation monochromator to a

4 nm bandpass and maintaining a 1.5 nm bandpass for the emission

monochromator, and b) opening the emission monochromator to a 4 nm

bandpass and maintaining a 1.5 nm bandpass for the excitation

monochromator. The lower scans in Figure 36 were obtained with the

same sensitivity scale and 1.5 nm bandpasses on both monochromators.

Figure 37 shows the effect of the same variation in parameters on

constant energy luminescence spectra of carbaryl. The effect of CESLS

can be visualized in a simplified way as moving simultaneously across

the excitation and emission spectra. (This would be an exact

representation if the spectra were plotted in units of wavenumber and

the separation between the points moving across these spectra was equal

to the constant energy difference being maintained.) For naphthol,

opening the excitation slit causes a loss in resolution in the constant

energy peaks at higher wavelength. This loss can be correlated to the

excitation and emission scans (see Figure 33). There is a broad

emission peak at higher wavelength, and therefore the use of a narrow

emission slit merely reduces light throughput. However, resolution is

lost by opening the excitation slit because the higher wavelength

excitation peaks are quite narrow. Resolution for naphthol is

maintained very well by opening the emission slit and keeping the

excitation slit narrow.
-i
For carbaryl, in a 2650 cm scan, resolution is maintained by

opening the excitation slit and keeping the emission slit narrow.

Again this effect can be related back to the excitation (broad peaks)

and emission (narrow peaks) scans for this compounds (see Figure 32).







































Figure 36.


a















_^L


WAVELENGTH (nm)
-1
Constant energy scans of naphthol with Av = 2650 cm
Demonstrating the comparison between spectra obtained
with bandpasses of 1.5 nm on both monochromators
versus a) maintaining a 1.5 nm emission bandpass and
opening the excitation bandpass to 4 nm, and b)
maintaining a 1.5 nm excitation bandpass and opening
the emission bandpass to 4 nm.


J L1















































WAVELENGTH


Figure 37. Constant energy scans of carbaryl
Showing the results obtained with
in parameters described in Figure


-1
with Av = 2650 cm
the same variation
36.


(nm)











For carbofuran, the constant energy spectrum is superimposed on a
-1
solvent background peak at Av = 2650 cm To minimize the solvent

background interference and to optimize throughput and resolution based
-i
on excitation and emission scans, a 1400 cm constant energy

difference is used for detection of carbofuran. However, it should be

noted that even in the presence of the solvent peak, carbofuran could

be readily detected at concentrations only slightly higher than the

limit of detection when a constant energy difference of 1400 cm-1 is

used (see Table VII and Figure 35).

Figure 38 shows the effect of varying bandpasses on the excitation

and emission monochromators for the three component mixture. For scan

a) both monochromator bandpasses are 1.5 nm; for b) the excitation

bandpass is 2.5 nm and the emission bandpass is 1.5 nm, and for scan c)

the excitation bandpass is 1.5 nm and the emission bandpass is 2.5 nm.

It can be seen that for these variations in bandpasses, signal

intensities are increased while resolution is maintained for

identification and quantitation purposes. Optimum parameter selection

will be determined by the complexity of the mixture being studied, the

specific components being determined, and the concentration of the

analytes. Depending on the application, one may wish to trade

resolution for increased sensitivity or vice versa.

In conclusion, we feel that we have demonstrated the sensitivity

and selectivity of CESLS, an inexpensive and reliable method, for

analysis of pesticides and hope that in the future this technique will

find wide applicability to studies involving pesticides as well as

other complex mixtures where physical separations may be avoided.

















































WAVELENGTH (nm)


Figure 38.


Constant energy scans of a mixture of carbaryl, naphthol,
and carbofuran with a) bandpasses of 1.5 nm on both
monochromators, b) excitation bandpass of 2.5 nm and
emission bandpass 1.5 nm, and c) excitation bandpass
1.5 nm and emission bandpass 2.5 nm.
















CHAPTER 5
TIME RESOLVED PHOSPHORIMETRY

Introduction

If a molecule absorbs light and is excited into a singlet excited

state, it may then undergo intersystem crossing into the excited

triplet state. If this molecule then decays radiatively to the ground

state this transition is termed phosphorescence. This process is

outlined in the Jablonski diagram presented in Figure 1, Chapter 1.

The reader is referred to Vo-Dinh (57) and previous references (17-20)

for a thorough explanation of parameters affecting phosphorescence, as

well as descriptions of methods currently employed for measuring

phosphorescence signals.

Phosphorimetry has a unique advantage over many analytical

measurement methods. This advantage is the possibility of conducting

time-resolved measurements to distinguish between compounds whose

phosphorescence spectra overlap but which have different lifetimes.

This possibility is graphically illustrated in Figure 39.

Time resolution measurements can be made by using many different

types of devices. Mechanical choppers are the simplest and most common

devices employed. The excitation radiation is periodically interrupted

and the emission is observed after a time delay following the

excitation cycle. The phosphorescence signal that persists after the

excitation has ceased can therefore be detected without interference

from scattered light and fluorescence emission. The time delay

between the excitation cycle and emission measurement can be varied to

85












PHOSPHORESCENCE

Time Resolution


(A)


I I I FL


Thh h h BC
I t I
td tg


f p
I e


(B)


.. (C)


FL




I I I
td, tg
I I I


FL


BC

td tg


Figure 39. Illustration demonstrating the ability to obtain more
selective spectra through the use of time-resolution.
FL flash lamp pulse, BC boxcar, f fluorescence,
p phosphorescence, td delay time, t gate time,
Subscripts used to denote compounds with different
excited state lifetimes.


f

IA

I











distinguish between short and long lived phosphors. Mechanical devices

used for chopping the excitation light include Becquerel discs (58),

rotating cylinders (59), and rotating mirrors (60).

A method superior to those using mechanical choppers employs

pulsed excitation and electronically gated detection. O'Haver and

Winefordner, 1966, and Fisher and Winefordner, 1972, demonstrated the

efficiency of pulsed-source time-resolved detection for analyzing

mixtures of fast-decaying phosphors (59,61). Later Winefordner and

coworkers improved the pulsed-source system by replacing the CW xenon

arc lamp and a mechanical chopper with a pulsed xenon flash tube for

excitation (62-64). A signal average or a boxcar integrator was used

for detection.

Systems using pulsed nitrogen lasers and flash-tube pumped dye

lasers with electronically gated detection have been used to measure

phosphorescence lifetimes in the submicrosecond range, record total

luminescence, and obtain well resolved phosphorescence spectra (65-68).

Other devices which have been used for detection in time-resolved

luminescence measurements are gated PM tubes (69-71), and silicon-

intensified target (SIT) vidicons (72).

O'Donnell and Winefordner used time-resolution extensively for

analysis of phosphors at low-temperature (77 K) (73). Low temperature

is used to minimize non-radiative de-excitation processes of the first

excited triplet state such as collisions with solute or solvent

molecules. One disadvantage of working at low temperature is the

requirement for liquid nitrogen and a dewar with quartz windows to

allow passage of ultra violet radiation. A further disadvantage of the

dewar system is the loss of energy at each interface in the system











(i.e. air to quartz, quartz to vacuum, vacuum to quartz, quartz to

liquid nitrogen, etc.). Also, as the sample freezes it may form a

clear glass, a cracked glass, or a snow. Many different variables can

affect the appearance of the frozen sample matrix and reproducibility

is often a problem. Spinning the sample tube as is done in NMR helps

to increase reproducibility, but is not 100% effective.

Vo-Dinh (57) describes room temperature measurements and the many

methods such as internal and external heavy atom effects and the use

of a variety of substrates to enhance phosphorescence, thereby allowing

measurement at temperatures greater than 77 K.

A new system is presented which combines sensitivity, selectivity,

and resolution of spectra obtained using constant energy synchronous

luminescence spectroscopy with the added selectivity of time-resolved

phosphorimentry measurements.


Experimental


Instrumentation

The experimental setup was designed to allow maximum versatility.

A schematic of the system is presented in Figure 40. Room temperature

and low temperature measurements can be made with minimal rearrangement

through the use of a room temperature cell attached to a cover which

fits over the dewar holder. A pulsed flashlamp (EG&G Model FX-239E)

is employed for excitation in phosphorescence measurements and a gated

integrator is used for detection. A mirror on a kinematic mount

allows one to change over to a continuous xenon arc lamp for excitation

in fluorescence measurements, again the emphasis is on versatility.