5 “3%.

Ink

. :5.
«$0..

13.? .
53:. A. .
. r1 .. t u:

l. HMHNM. nukmafiouflwflcfu

...W
{otOs\u.

if. "a a.»

. l. . its}. Ix 1
$3 Iv
”lift. km. 1L. :JIM
. a: Ink
. . iofiamuu‘... Wm.“

344.1%...

.. I991

ill-l.»
. «I.
3|!

.i. ’4

.3 .15?an .
mmfiyxfia

. “6mg?
J... manna“

:mh.

vanishg “ .5
“my... (fly-Hg orbn
, a..u...‘.......u...i .531...
p f '..I
.32. I
2 Law
g
#:31 2 I 0 .fi
if, J T...
. anlia . All

.94?

s.
.flhfi‘fia
.iiL‘n

.6!
35?...

billiflu

 

 

Michigan State
_University

 

 

 

This is to certify that the
dissertation entitled

ANALYTICAL APPLICATIONS
OF FLUORESCENCE QUENCHING

presented by

MELISSA SUE MEANEY

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Chemistry

 

 

’U LLI'CJILIU “76}st v

Major Professor‘s Signature

IL! DJ 17,004-

 

Date

MSU is an affirmative-action, equal—opportunity employer

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K./Proi/Acc&Pres/CIRC/Date0ue indd

ANALYTICAL APPLICATIONS OF FLUORESCENCE QUENCHING
By

Melissa Sue Meaney

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Chemistry

2008

ABSTRACT
ANALYTICAL APPLICATIONS OF FLUORESCENCE QUENCHING
By
Melissa Sue Meaney

Fluorescence quenching has many forms, from trivial mechanisms
involving primary and secondary absorption to more selective formation of
ground- and excited-state complexes. Much work has been done to eliminate
quenching interferences from traditional solution-phase fluorescence
measurements, but little effort has focused on tuning the fluorophore chemistry
for selective detection of quencher molecules. Pyrene has demonstrated a
sensitive and selective quenching interaction with nitrated explosives, allowing a
40-fold improvement in signal-to-noise ratio compared with traditional UV
absorbance. Our Stem-Volmer data indicate that phenol red, malachite green,
and brilliant purpurin R are suitable, less toxic replacements for pyrene for the
detection of nitrated materials. In comparison to pyrene, these fluorophores
demonstrate comparable sensitivity and greater selectivity for model nitrated
compounds than for potential interfering species. When compared to Rehm-
Weller theory, the data suggest an electron-transfer mechanism is responsible
for the quenching signal.

For utility .in a field device, the fluorescence of malachite green is
investigated in the solid state. Because the greatest solution-phase fluorescence
is observed in dioxane, matrices with similar polarity to dioxane are considered.

Fluorescence of malachite green is not observed when adsorbed onto filter paper,

 

but is observed in a poly(ethylene glycol) thin film. Poly(ethylene glycol) films
containing malachite green are coated onto filter paper and silica particles and
their fluorescence investigated for application as a wipe material and air sampling
tube for explosives detection, respectively.

In addition to the application for nitrated explosives detection, an indirect
fluorescence method has been developed for the detection of acids. The pH
sensitivity of fluorescein demonstrates a novel method for selective identification
of acids in complex samples. The relationship between the logarithm of acid
concentration and fluorescence power of fluorescein is sigmoidal and dependent
on the fiuorescein concentration as well. The method is demonstrated for any
acid with pKa s 5.5 such that proton transfer to fluorescein is possible. Most
importantly, the approach provides a single calibration curve for acids of varying
strength and number of proton-donating sites. This method has been applied to
characterization of acids in fruit-derived beverages and food products such as
juice, wine, and vinegar following liquid chromatography.

A forensic extension of this method has been made to the determination of
y—hydroxybutyric acid (GHB) in adulterated beverages, including beer, wine, rum,
cola, and coffee. GHB is a common drug of abuse and is often found in alleged
sexual assault cases. The limit of detection of this method is 3 mM GHB in
various beverages, which corresponds to 90 mg GHB in an 8-oz beverage. This
detection limit is an order of magnitude below the therapeutic dose of GHB, and
20 — 40 times below the typical recreational dose. The sensitivity of the method

indicates that GHB may be detectable in beverage residues at a crime scene.

 

To My Friends and Family, and To Andrew, with Love

ACKNOWLEDGEMENTS

This road through graduate school has been challenging, enjoyable,
frustrating, and enlightening, and often all in the same day. As an undergraduate,
I never imagined being on this path, but I am grateful to those who directed me to
a future that involved higher education. l have been very blessed over the past
five and a half years at Michigan State University in meeting many amazing
friends and scientists and having numerous noteworthy experiences. The skills
and knowledge base that l have gained during graduate school will no doubt
guide me through the rest of my scientific career.

For the content of this dissertation, as well as the guidance only a true
mentor can give, I must thank my thesis advisor, Dr. Victoria McGuffin. Without
her continued support and encouragement, despite my own doubts and
frustrations, I would not have completed the research contained within this
dissertation. I am grateful for her genuine interest in the projects and areas of
interest to me, and for working diligently to keep the focus of my many projects
on high-quality, analytical forensic science. Vicki’s practical expertise and vast
knowledge of spectroscopy and separation science are directly reflected in the
content and structure of this work. In addition to her scientific advising, I have
also been blessed with an amazing woman role model who is incredibly
successful and more than willing to share her knowledge and experience.
Endless insight into the processes of peer-reviewed publication, grant
preparation and submission, education, and ethics has prepared me for any

career, and especially for a future in academia.

I would also like to thank the members of my thesis committee, Dr.
Merlin Bruening, Dr. John McCracken, Dr. Jay Siegel, and Dr. Gary Blanchard. I
greatly appreciate Merlin’s feedback on my dissertation and the time spent truly
thinking about the concepts I presented. I was also lucky to have spent several
semesters as a teaching assistant under Merlin’s guidance, and his trust in my
instructional ability only deepened my interest in teaching. I was also fortunate to
have the enjoyable experience of teaching for John in my final semester. Jay
was also very influential in my graduate studies, by accepting me into the
forensic science program and recommending that I pursue a Ph.D. in chemistry.
I must also thank Gary for always supporting me and stepping in to join my
committee in the final weeks. On a related note, I want to thank Lisa Dillingham,
who was always a friend and confidant to me, and who made many difficult
situations much more bearable.

l was fortunate enough to work with Heidi Bonta, an amazing
undergraduate student, on this work. Heidi worked diligently for three semesters
to design and properly characterize a field device for explosives detection. As a
BS. Chemistry major, she managed to fit undergraduate research into her
already overloaded schedule. Heidi’s creativity and hard work helped us
overcome a number of hurdles during this research project, and her enthusiasm
and dedication are what made this project so enjoyable. I am greatly
appreciative of her hard work and positive attitude, and truly I am very happy to
consider her a colleague and a friend.

Lastly, I must thank my friends and family who have supported me

vi

through my time in graduate school. Thanks to Sam and Carl for teaching me so
much about what it means to be a Ph.D. student. Thanks to John Goodpaster for
numerous emails and phone calls that saved me days of work. Thanks to Amber
for sharing in every aspect of being a McGuffin group member, and for becoming
one of my best friends. To the other members of the group: Xiaoping, Rashad,
Kahsay, Heidi, Lucas, John, and Melissa, for making the everyday tasks more
enjoyable. Thanks to Elizabeth, the one person with whom I shared nearly every
moment of graduate school and life — I couldn't have done it without you. Thanks
to Audrey for being a great scientist, roommate, and friend, and to Anne and
Jason for helping me to realize that there’s more to life than work. Thanks to Alli,
Lucas, Dan, and Lisa for being amazing friends, and knowing when it’s about
work and when it’s not. Thanks to Erin, Sarah, Kristin, and Lisa for keeping me
grounded. Thanks to my softball team for keeping me active and sharing lots of
memories on and off the field. Thanks to Mom, Dad, Melinda, Jerrod, Matt,
Stephanie, Gerri, Don, Geoff, Donna, and Kevin, as well as many other family
members, for their willingness to accept my absence at many family events and
for their support throughout my graduate career. A special thank you to my
parents for believing in me and encouraging me to pursue my goals and dreams
from as far back as I can remember. Thanks to my nieces and nephews who just
love their Aunt Mel. Lastly, thanks to Andrew, for being by my side
unconditionally through this process and for loving me so truly. We have been
on an amazing journey through these past few years, and I can’t wait for it to

continue as I spend the rest of my life with you.

vii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................... xi
LIST OF FIGURES ............................................................................................ xiii
CHAPTER 1. INTRODUCTION AND BACKGROUND ....................................... 1
1.1 Fluorescence and Fluorescence Quenching ........................................ 1
1.2 Luminescence-Based Methods in Forensic Science ............................ 6
1 .3 Explosives ............................................................................................ 7
1.4 Luminescence-Based Methods for Explosives Detection ................... 14
1.4.1 Direct Detection Methods ................................................................... 15
1.4.1. High-Energy Techniques .................................................................... 15
1.4.1. Reaction-Based Methods ................................................................... 17
1.4.2 Indirect Detection Methods ................................................................. 22
1.4.2. Quenching of Solution-Phase Fluorophores ....................................... 22
1.4.2. Quenching of Immobilized Fluorophores ............................................ 24
1.4.2.3 Quenching by Solid-State Fluorophores ............................................. 31
1.4.2.4 Indirect Laser-Induced Fluorescence ................................................. 38
1.4.2.5 Fluorescence Immunoassay ............................................................... 40
1.4.2.6 Reactions Producing Fluorescent Species ......................................... 46
1.5 Research Objectives .......................................................................... 48
1 .6 References ......................................................................................... 50
CHAPTER 2. SOLUTION-PHASE INVESTIGATIONS OF COMMON

FLUOROPHORES FOR THE DETECTION OF NITRATED EXPLOSIVES ...... 62
2.1 Introduction and Background .............................................................. 63
2.1.1 Analysis of Nitrated Explosives .......................................................... 63
2.1.2 Solution-Phase Fluorescence Quenching .......................................... 63
2.2 Experimental Methods ........................................................................ 67
2.2.1 Reagents ............................................................................................ 67
2.2.2 Determination of Quenching Constants .............................................. 67
2.2.3 Determination of Fluorescence Lifetimes ........................................... 70
2.2.4 Determination of Redox Potentials ..................................................... 71
2.2.5 Determination of Singlet Excitation Energies ...................................... 72
2.3 Results and Discussion ...................................................................... 72
2.3.1 Fluorophore Characterization ............................................................. 72
2.3.2 Quenching of Pyrene by Nitrated Explosives ..................................... 73
2.3.3 Preliminary Investigation of Selected Fluorophores ........................... 78
2.3.4 Quenching of Purpurin by Nitrated Explosives ................................... 78
2.3.5 Quenching of Malachite Green by Nitrated Explosives ....................... 81
2.3.6 Quenching of Phenol Red by Nitrated Explosives .............................. 85
2.3.7 Proposed Quenching Mechanism ...................................................... 89
2.3.7.1 Fluorescence Lifetimes ....................................................................... 90

viii

2.3.7.2 Redox Potentials ................................................................................ 94

2.3.7.3 Singlet Excitation Energies ................................................................. 94
2.3.7.4 Comparison to Rehm-Weller Theory .................................................. 96
2.4 Conclusions ...................................................................................... 101
2.5 References ....................................................................................... 103

CHAPTER 3. INVESTIGATION OF POTENTIAL FIELD-READY DEVICES FOR
THE DETECTION OF NITRATED EXPLOSIVES VIA FLUORESCENCE

QUENCHING .................................................................................................... 104
3.1 Experimental Methods ...................................................................... 105
3.1 .1 Reagents .......................................................................................... 105
3.1.2 Fluorescence Measurements ........................................................... 105
3.1.3 Preparation of Solid Substrates ........................................................ 106
3.2 Results and Discussion .................................................................... 108
3.2.1 Solution-Phase Fluorescence and Quenching ................................. 108
3.2.2 Solid-Phase Fluorescence ................................................................ 1 15
3.2.3 Poly(ethylene glycol)-Coated Substrates .......................................... 123
3.3 Conclusions and Future Work .......................................................... 130
3.4 References ....................................................................................... 1 31

CHAPTER 4. pH-ENHANCED FLUORESCENCE DETECTION OF ACIDS.. 132

4.1 Introduction and Background ............................................................ 132
4.2 Experimental Methods ...................................................................... 138
4.2.1 Reagents .......................................................................................... 138
4.2.2 Chromatographic System ................................................................. 139
4.2.3 Spectroscopic System ...................................................................... 141
4.3 Results and Discussion .................................................................... 142
4.3.1 Method Development and Optimization ............................................ 142
4.3.1.1 Analysis of Test Acids ...................................................................... 143
4.3.1.2 Chromatographic Optimization ......................................................... 148
4.3.1 .3 Detector Optimization ....................................................................... 151
4.3.2 Analysis of Real Samples ................................................................. 159
4.3.2.1 Analysis of Grape Juice .................................................................... 159
4.3.2.2 Analysis of Wines ............................................................................. 159
4.3.2.3 Analysis of Vinegars ......................................................................... 161
4.4 Conclusions ...................................................................................... 168
4.5 References ....................................................................................... 173
CHAPTER 5. pH-ENHANCED FLUORESCENCE DETECTION OF 7-

HYDROXYBUTYRIC ACID IN BEVERAGES .................................................. 176
5.1 Introduction and Background ............................................................ 177
5.1.1 GHB .................................................................................................. 177
5.1.2 Illicit GHB and Forensic Implications ................................................ 177

 

5.1.3 Determination of GHB ...................................................................... 180
5.2 Experimental Methods ...................................................................... 182
5.2.1 Sample Preparation .......................................................................... 182
5.2.2 Experimental System ........................................................................ 184
5.3 Results and Discussion .................................................................... 184
5.3.1 GHB Calibration Curve ..................................................................... 184
5.3.2 Beverage Analysis ............................................................................ 185
5.3.2.1 GHB in Alcoholic Beverages ............................................................ 188
5.3.2.2 GHB in Non-Alcoholic Beverages ..................................................... 188
5.3.2.2 Effect of Ethanol on GHB Retention ................................................. 194
5.3.3 lnterconversion of GHB to GBL ........................................................ 196
5.4 Conclusions ...................................................................................... 196
5.5 References ....................................................................................... 199
CHAPTER 6. CONCLUSIONS AND FUTURE WORK ................................... 201
APPENDIX. INVESTIGATION OF QUENCHING MECHANISM ..................... 207
A.1 Determination of Fluorescence Lifetimes ......................................... 207
A2 Determination of Redox Potentials ................................................... 207
A.3 Determination of Singlet Excited-State Energies .............................. 211

Table 1.1
Table 1.2
Table 2.1

Table 2.2

Table 2.3

Table 2.4

Table 2.5

Table 2.6

Table 2.7

Table 2.8

Table 2.9

Table 2.10

Table 2.11

Table 3.1

LIST OF TABLES

Common nomenclature for explosives and related compounds ...... 9
Twelve common military explosives and formulations ................... 12
Spectral properties of fluorophores ............................................... 69

Quenching constants (Kd) for pyrene (1) with target and interfering
quenchers in methanol at room temperature ................................ 74

Quenching constants (Kd) for fluorophores with nitrobenzene and 4-
nitrotoluene in methanol at room temperature .............................. 77

Quenching constants (Kd) for purpurin (5) with target and interfering
quenchers in methanol at room temperature ................................ 79

Quenching constants (Kd) for malachite green (8) with target and
interfering quenchers in methanol at room temperature ................ 82

Quenching constants (Kd) for phenol red (9) with target and
interfering quenchers in methanol at room temperature ................ 86

Fluorescence lifetimes (no) determined for each of the fluorophores
at the wavelengths listed in Table 2.1 (n = 6). All values
determined in methanol at room temperature ............................... 91

Quenching rate constants (kd) calculated for each of the
fluorophores in methanol at room temperature ............................. 92

Electrochemical data for fluorophores and quenchers in methanol
at room temperature ...................................................................... 93

Spectroscopic data for fluorophores in methanol at room
temperature. .................................................................................. 95

Free energy for electron-transfer (AGet, eV) from excited-state
fluorophore to ground-state quencher (Mechanism I) and electron-
transfer from ground-state quencher to excited-state fluorophore
(Mechanism ll). Boldface type indicates the more
thermodynamically favorable mechanism ..................................... 97

Optimized fluorescence excitation and emission parameters ..... 113

xi

Table 4.1

Table 4.2

Table 4.3

Table 4.4

Table 4.5

Table 4.6

Table 5.1

Table A.1

Correlation coefficient (R2) and parameters of fit (a-d) for test acids
to sigmoidal function (Equation 4.1) (top) and the corresponding
correlation to the sigmoid transform function (Equation 4.2)
(bottom) ....................................................................................... 146

Summary of retention factors (k) and pKa values for standard acids
in water ....................................................................................... 150

Optimization of the flow rate of a 10 mM sodium hydroxide
solution. Mobile phase: 0.5 mL/min, pH = 3. Fluorescein: 104 M,
50 LIL/min .................................................................................... 153

Optimization of the PMT voltage. Mobile phase: 0.5 mL/min, pH =
3. Fluorescein: 10“ M, 100 pL/min. Sodium hydroxide: 10'2 M, 75
uL/min. Integration time: 0.2 s. Slit widths: 0.2 mm. ................ 155

Optimization of the PMT integration time. Mobile phase: 0.5 mL/
min, pH = 3. Fluorescein: 10“ M, 100 pL/min. Sodium hydroxide:
10'2 M, 75 pL/min. PMT: 700 V. Slit widths: 0.1 mm ............... 157

Optimization of the entrance and exit slit widths. Mobile phase:

0.5 mL/min, pH = 3. Fluorescein: 10"4 M, 100 pL/min. Sodium
hydroxide: 10'2 M, 75 pL/min. PMT: 700 V. Integration time:

0.2 s ............................................................................................ 158

Summary of alcoholic and non-alcoholic beverages used in this
study ........................................................................................... 183

Parameters used for determination of fluorescence lifetimes ...... 208

xii

Figure 1.1

Figure 1.2

Figure 1.3

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

LIST OF FIGURES

Jablonski diagram depicting absorption (A), vibrational relaxation
(VR), fluorescence (F), and static (Ks) and dynamic (Kd)
fluorescence quenching. Solid lines represent radiative transitions,
whereas broken lines represent nonradiative transitions. The
electronic states shown are for a fluorophore in its singlet ground
(1Fo) and excited ('F1) states and for a fluorophore-quencher
complex in its ground ((FQ)o) and excited states for static ((FQ)1)
and dynamic ( FiQ) quenching ....................................................... 3

Structures of common nitroaliphatic and nitroaromatic explosives
and degradation products. NM (A), DMNB (B), NB (C), o-DNB (D),
TNB (E), 2-NT (F), 2,6-DNT (G), 2-am-DNT (H), TNT (I), Picric acid
(J), and Tetryl (K) .......................................................................... 10

Structures of common nitramine, nitrate ester, peroxide, and
inorganic explosives. RDX (L), HMX (M), NGU (N), NG (O), PETN
(P), EGDN (Q), TATP (R), HMTD (S), and AN (T) ........................ 11

Structure of fluorophores under investigation. Pyrene (1); 4-
Hydroxycoumarin (2); 4-Bromomethyl-7—methoxycoumarin (3); 7-
Diethylamino-4-methylcoumarin (4); Purpurin (5), Acridine orange
(6); Methylene blue (7); Malachite green (8); Phenol red (9);
Rhodamine 6G (10); Fluorescein (1 1) ........................................... 65

Fluorescence spectra of 5x10-5 M pyrene in methanol at room
temperature in the presence of increasing concentrations of 2,6-
dinitrotoluene (top) and the resulting Stern-Volmer plot (bottom).
The Stem-Volmer constant for this data set is 450 M-1 with a
correlation coefficient of 0.9988 .................................................... 75

Fluorescence spectra of 5x104 M purpurin in methanol at room
temperature in the presence of increasing concentrations of 2,6-
dinitrotoluene (top) and the resulting Stern-Volmer plot (bottom).
The Stem-Volmer constant for this data set is 129 M'1 with a
correlation coefficient of 0.9907 .................................................... 80

Fluorescence spectra of 2x104 M malachite green in methanol at
room temperature in the presence of increasing concentrations of
2,6-dinitrotoluene (top) and the resulting Stern-Volmer plot (bottom).
The Stem-Volmer constant for this data set is 56 M'1 with a
correlation coefficient of 0.9958 .................................................... 83

Fluorescence spectra of 6x10“1 M phenol red in methanol at room
temperature in the presence of increasing concentrations of 2,6-

xiii

Figure 2.6

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

Figure 3.8

Figure 3.9

dinitrotoluene (top) and the resulting Stern-Volmer plot (bottom).
The Stem-Volmer constant for this data set is 98 M" with a
correlation coefficient of 0.9979 .................................................... 87

Comparison of experimental quenching rate constants for pyrene
(0), purpurin (o), malachite green (A), and phenol red (:1) with target
quenchers (filled) and interfering quenchers (unfilled) in methanol
with Rehm-Weller theory (-) ........................................................ 100

Fluorescence spectra of 10‘1 M purpurin in water (—), methanol (---),

dioxane (—), and hexane (--) at room temperature. Excitation:
325 nm ........................................................................................ 109

Fluorescence spectra of 10'3 M malachite green in water (—),

methanol (--), dioxane (—), and hexane (-—) at room temperature.
Excitation: 325 nm ...................................................................... 110

Fluorescence spectra of 10" M phenol red in water (—), methanol

(--), dioxane (—), and hexane (--) at room temperature.
Excitation: 310 nm ...................................................................... 111

Structures of malachite green. Left: Neutral form. Right: Ionic
form. ............................................................................................ 1 14

Fluorescence spectra of 2x10" M malachite green in dioxane at
room temperature in the presence of increasing concentrations of
nitrobenzene (top) and the resulting Stem-Volmer plot (bottom).
The Stem-Volmer constant for this data set is 120 M" with a
correlation coefficient of 0.9997 .................................................. 116

Fluorescence spectra of a KBr pellet (—) and a KBr pellet
containing malachite green (—). Excitation: 325 nm ................ 117

Fluorescence spectra of a Kimwipe (—-), filter paper (--), a Grab-it

cloth (—), and fiberglass filter paper (—) at room temperature.
Excitation: 325 nm. Slit width: 1 mm (top), 5 mm (bottom) ....... 120

Fluorescence spectra of a Kimwipe (--), filter paper (-),and
fiberglass filter paper (—) coated with 10'3 M malachite green
solution at room temperature. Excitation: 325 nm. Slit width: 5
mm .............................................................................................. 121

Fluorescence spectra of filter paper coated with various

concentrations of malachite green solutions at room temperature.
Excitation: 325 nm. Slit width: 5 mm ........................................ 122

xiv

Figure 3.10

Figure 3.11

Figure 3.12

Figure 4.1

Figure 4.2

Figure 4.3

Figure 4.4

Figure 4.5

Fluorescence spectra of a glass slide coated with 1 mg malachite

green in 2 9 PEG 1500 (—) and PEG 2000 (—). Films prepared by
using the melting technique. Spectra acquired at room temperature.
Excitation: 325 nm. Slit width: 5 mm ........................................ 125

Fluorescence spectra of filter paper coated with 0.17 mg malachite
green and 0.8% (—), 1.6% (—), 4.1% (—), 5.4% (—), 8.1% (--),
16.3% (--), and 40.6% (---) PEG 2000. Percentages refer to the
ratio of polymer mass to filter paper mass. Spectra collected at
room temperature. Excitation: 325 nm. Slit width: 5 mm. PMT:
950 V (top), 700 V (bottom) ......................................................... 127

Fluorescence spectra of 0.1 g silica particles coated with 30% (--)
PEG 2000 (w/w), and 0.17 mg malachite green and 10% (---), 20%
(—), and 30% (—) PEG 2000 (w/w). Spectra acquired at room
temperature. Excitation: 325 nm. Slit width: 0.1 mm. Integration
time: 0.1 s .................................................................................. 128

Structures of common low-molecular-weight organic acids. Fonnic
acid (1 ), Acetic acid (2), Lactic acid (3), Oxalic acid (4), Succinic
acid (5), Malic acid (6), Tartaric acid (7), Citric acid (8) ............... 133

Acid-base equilibria of fluorescein in aqueous solution at room
temperature. Molar absorptivity (e) and quantum yield ((1)) are given
for each form at the wavelength of maximum absorbance .......... 137

Schematic diagram of the experimental system used for
determination of LMW organic acids in beverages by liquid
chromatography with pH-dependent fluorescence detection. Pump
A: Delivery of sample. Pump B: Delivery of 10 mM sodium
hydroxide. Pump C: Delivery of 5 mM fluorescein. I = injection
valve, L = lens, F = filter, PMT = photomultiplier tube, COD =
charge-coupled device. Dark lines indicate 1/16” stainless steel
tubing. Light lines indicate 340-pm o.d. capillary tubing ............. 140

Sigmoidal calibration curve for four test acids in water, determined
from the fluorescence power of fluorescein at 510 nm. Test acids:
hydrochloric acid (0), benzoic acid (0), maleic acid (I), citric acid
(Cl). Non-linear correlation coefficient (R2) = 0.9921 ................... 144

Linear calibration curve following sigmoid transform for four test
acids in water, determined from the fluorescence power of
fluorescein at 510 nm. Test acids: hydrochloric acid (0), benzoic
acid (0), maleic acid (I), citric acid (El). Linear correlation
coefficient (R2) = 0.9729 .............................................................. 147

XV

Figure 4.6

Figure 4.7

Figure 4.8

Figure 4.9

Figure 4.10

Figure 4.11

Figure 4.12

Figure 4.13

Chromatogram of common LMW organic acids obtained by UV-
absorbance detection (A) and laser-induced fluorescence detection
(B). Column: Hypersil ODS, 5-um particles, 250 mm x 4.6 mm.
Mobile phase: 99:1 water:methanol, pH 3. Flow rate: 1.0 mL/min.
Laser-induced fluorescence detection: 325-hm excitation, 510-nm
emission, 10-nm bandpass. UV-absorbance detection: 215 nm.
Solutes: oxalic acid (1 ), tartaric acid (2), formic acid (3), malic

acid (4), lactic acid (5), acetic acid (6), citric acid (7), succinic acid
(8) ................................................................................................ 149

Chromatogram of grape juice obtained by UV-absorbance detection
(A) and laser-induced fluorescence detection (B). Chromatographic
conditions as described in Figure 4.6. Solutes: oxalic acid (1 ),

tartaric acid (2), unknown (3), malic acid (4), citric acid (5). ........ 160

Chromatogram of red wine obtained by UV-absorbance detection
(A) and laser-induced fluorescence detection (B). Chromatographic
conditions as described in Figure 4.6. Solutes: Solutes: oxalic
acid (1 ), tartaric acid (2), malic acid (3), unknown (4), lactic acid (5),
succinic acid (6) .......................................................................... 162

Chromatogram of white wine obtained by UV-absorbance detection
(A) and laser-induced fluorescence detection (B). Chromatographic
conditions as described in Figure 4.6. Solutes: oxalic acid (1 ),

tartaric acid (2), malic acid (3) ..................................................... 163

Chromatogram of red wine vinegar obtained by UV-absorbance
detection (A) and laser-induced fluorescence detection (B).
Chromatographic conditions as described in Figure 4.6. Solutes:
oxalic acid (1 ), tartaric acid (2), malonic acid (3), acetic acid (4),
succinic acid (5) .......................................................................... 165

Chromatogram of white wine vinegar obtained by UV-absorbance
detection (A) and laser-induced fluorescence detection (B).
Chromatographic conditions as described in Figure 4.6. oxalic

acid (1 ), tartaric acid (2), unknown (3), acetic acid (4), succinic acid
(5) ................................................................................................ 166

Chromatogram of balsamic vinegar obtained by UV-absorbance
detection (A) and laser-induced fluorescence detection (B).
Chromatographic conditions as described in Figure 4.6. oxalic acid
(1 ), tartaric acid (2), formic acid (3), malonic acid (4), unknown (5),
acetic acid (6), fumaric acid (7) ................................................... 167

Chromatogram of white vinegar obtained by UV-absorbance
detection (A) and laser-induced fluorescence detection (B).

xvi

Figure 4.14

Figure 4.15

Figure 5.1

Figure 5.2

Figure 5.3

Figure 5.4

Figure 5.5

Chromatographic conditions as described in Figure 4.6. oxalic acid
(1 ), acetic acid (2) ....................................................................... 169

Chromatogram of malt vinegar obtained by UV-absorbance
detection (A) and laser-induced fluorescence detection (B).
Chromatographic conditions as described in Figure 4.6. oxalic acid
(1 ), unknown (2), unknown (3), acetic acid (4) ............................ 170

Chromatogram of rice vinegar obtained by UV-absorbance
detection (A) and laser-induced fluorescence detection (B).
Chromatographic conditions as described in Figure 4.6. oxalic acid
(1 ), tartaric acid (2), formic acid (3), lactic acid (4), acetic acid (5),
succinic acid (6) .......................................................................... 171

Structures of GHB and related compounds. y-hydroxybutyric acid,
GHB (1), y-aminobutyric acid, GABA (2), y-butyrolactone, GBL (3),
1,4-butanediol, 1,4-BD (4), y-valerolactone, GVL (5) .................. 178

Calibration curve for GHB measured in water, calculated by using
peak height (0) and peak area (0). Linear regression for peak
height: y = -1.0121x - 0.0165 (R = 0.9961 ). Linear regression for
peak area: y = -1.0470x — 0.0678 (R2 = 0.9946). Error bars
represent standard deviation of three measurements ................. 186

Chromatogram of common organic acids and GHB obtained by UV-
absorbance detection (A) and laser-induced fluorescence detection
(B). Column: Hypersil ODS, 5-pm particles, 250 mm x 4.6 mm.
Mobile phase: 99:1 water:methanol, pH 3. Flow rate: 1.0 mL/min.
Laser-induced fluorescence detection: 325-hm excitation, 510-nm
emission, 10-nm bandpass. UV-absorbance detection: 215 nm.
Solutes: oxalic acid (1 ), tartaric acid (2), formic acid (3), malic acid
(4), lactic acid (5), acetic acid (6), citric acid (7), GHB (8), succinic
acid (9) ........................................................................................ 187

Chromatogram of a 12-oz (360-mL) serving of beer spiked with
0.75 g GHB obtained by UV-absorbance detection (A) and laser-
induced fluorescence detection (B). Chromatographic conditions
as described in Figure 5.3. Solutes: oxalic acid (1), phosphoric
acid (2), GHB (3) ......................................................................... 189

Chromatogram of a 6-02 (180—mL) serving of red wine spiked with
0.75 g GHB obtained by UV-absorbance detection (A) and laser-
induced fluorescence detection (B). Chromatographic conditions

as described in Figure 5.3. Solutes: oxalic acid (1 ), tartaric acid (2),
malic acid (3), lactic acid (4), unknown (5), GHB (6). .................. 190

xvii

Figure 5.6

Figure 5.7

Figure 5.8

Figure 5.9

Figure A.1

Figure A.2

Figure A.3

Figure A.4

Figure A.5

Chromatogram of a 1.5-oz (45-mL) serving of rum spiked with 0.25
g GHB obtained by UV-absorbance detection (A) and laser-induced
fluorescence detection (B). Chromatographic conditions as
described in Figure 5.3. Solutes: oxalic acid (1), lactic acid (2),
GHB (3) ....................................................................................... 1 91

Chromatogram of a 12-02 (360-mL) serving of cola spiked with 0.75
g GHB obtained by UV-absorbance detection (A) and laser-induced
fluorescence detection (B). Chromatographic conditions as
described in Figure 5.3. Solutes: phosphoric acid (1), GHB
(2) ................................................................................................ 192

Chromatogram of an 8-oz (240-mL) serving of coffee spiked with
0.75 g GHB obtained by UV-absorbance detection (A) and laser-
induced fluorescence detection (B). Chromatographic conditions
as described in Figure 5.3. Solutes: oxalic acid (1 ), formic acid (2),
unknown (3), GHB (4) ................................................................. 193

Chromatogram of 50 mM each GHB and GBL in water obtained by
UV—absorbance detection (A) and laser-induced fluorescence
detection (B). Chromatographic conditions as described in Figure
5.3 ............................................................................................... 197

Fluorescence decay curves for pyrene (top) and purpurin (bottom).
For pyrene (top), F0 = 4477, C = 89.88, r = 15.3 ns, R2 = 0.9975.
For purpurin (bottom), F0 = 1607, C = 146.8, 1 = 4.00 ns, R2 =
0.9933 ......................................................................................... 209

Fluorescence decay curves for malachite green (top) and phenol
red (bottom). For malachite green (top), F0 = 2157, C = 38.30, T =
1.05 ns, R2 = 0.9887. For phenol red (bottom), F0 = 1554, C =
48.82, 1: = 1.97 ns, R2 = 0.9936 ................................................... 21o

Cyclic voltammograms for pyrene in 0.1 M NaN03 in methanol, with
potential scanned cathodically (top) and anodically (bottom) ...... 212

Cyclic voltammograms for purpurin in 0.1 M NaN03 in methanol,
with potential scanned cathodically (top) and anodically
(bottom) ....................................................................................... 213

Cyclic voltammograms for malachite green in 0.1 M NaNO3 in

methanol, with potential scanned cathodically (top) and anodically
(bottom) ....................................................................................... 214

xviii

Figure A.6

Figure A]

Figure A.8

Figure A.9

Figure A.10

Figure A.11

Figure A.12

Figure A.13

Figure A.14

Figure A.15

Cyclic voltammograms for phenol red in 0.1 M NaNOa in methanol,
with potential scanned cathodically (top) and anodically
(bottom) ....................................................................................... 215

Cyclic voltammograms for nitromethane in 0.1 M NaN03 in
methanol, with potential scanned cathodically (top) and anodically
(bottom) ....................................................................................... 216

Cyclic voltammograms for nitrobenzene in 0.1 M NaN03 in
methanol, with potential scanned cathodically (top) and anodically
(bottom) ....................................................................................... 217

Cyclic voltammograms for 2-nitrotoluene in 0.1 M NaNO3 in
methanol, with potential scanned cathodically (top) and anodically
(bottom) ....................................................................................... 218

Cyclic voltammograms for 2,6-dinitrotoluene in 0.1 M NaNOa in
methanol, with potential scanned cathodically (top) and anodically
(bottom) ....................................................................................... 219

Cyclic voltammograms for aniline in 0.1 M NaNOa in methanol, with
potential scanned cathodically (top) and anodically (bottom) ...... 220

Cyclic voltammograms for benzoic acid in 0.1 M NaN03 in
methanol, with potential scanned cathodically (top) and anodically
(bottom) ....................................................................................... 221

Cyclic voltammograms for phenol in 0.1 M NaNOa in methanol, with
potential scanned cathodically (top) and anodically (bottom) ...... 222

Normalized absorbance (—) and emission (—) spectra for pyrene
(top) and purpurin (bottom). Emission spectra are collected by
using the wavelength of maximum absorbance as the excitation
wavelength. Singlet excitation energies are determined from the
crossing point of the two spectra ................................................. 223

Normalized absorbance (—) and emission (——) spectra for
malachite green (top) and phenol red (bottom). Emission spectra
are collected by using the wavelength of maximum absorbance as
the excitation wavelength. Singlet excitation energies are
determined from the crossing point of the two spectra ................ 224

xix

CHAPTER 1

INTRODUCTION AND BACKGROUND

Fluorescence is a sensitive and selective analytical tool that is widely
applied in the analytical chemistry community, but is somewhat underutilized
among chemists in the forensic science community. Luminescence-based
methods have been used in a limited capacity in the determination of drugs of
abuse and in trace evidence analysis. One area in which luminescence-based
methods are widely used is in the detection of explosives and their degradation
products in complex matrices. Direct detection methods utilize the inherent
fluorescence of explosive molecules or the luminescence generated from
chemical reactions. These methods include high-energy excitation techniques
such as gamma-ray and x-ray fluorescence, detection of decomposition products
by fluorescence or chemiluminescence, and detection following reduction to
amines or other reactions that produce fluorescent products from the explosive.
Indirect detection methods utilize the interference with traditional processes of
fluorescence caused by the presence of explosive compounds. Indirect
detection methods include quenching of solution-phase, immobilized, and solid-
state fluorophores, fluorophore displacement, fluorescence immunoassay, and

reactions that produce fluorescent products other than the explosive.
1.1 Fluorescence and Fluorescence Quenching

Fluorescence can be described as the excitation and radiative relaxation
of an electron between orbitals of different energies and is traditionally depicted

by the Jablonski diagram (Figure 1.1, center). The first step in a fluorescence

event is the absorption of an incident photon of sufficient energy to promote an
electron to a higher energy orbital. The molecule releases the excess energy
first through vibrational relaxation and then by emission of a photon, leaving the
molecule in its ground electronic state. At low concentrations, where the
absorbance (8bC) of the fluorophore is less than 0.05, the fluorescence power (F)
can be described by
F = 2.3K8DCP0 (1 .1)
The constant K is related to the quantum efficiency of the fluorescence process
and the efficiency of optical irradiation and collection, 8 is the molar absorptivity
of the fluorophore, C is the concentration of the fluorophore, and P0 is the source
power.2 When the source power is constant, Equation 1.1 is simplified to
F = K'C (1.2)
by defining a new constant K'
K' = 2.3K sbPo (1.3)
When the absorbance is greater than 0.05, the assumptions made in Equations
1.1 and 1.2 are no longer valid and the fluorescence power must be represented
by an exponential function.2

Fluorescence- or luminescence-based methods offer many benefits over
other commonly used techniques. Generally, fluorescence is measured against
a low or zero background, as native fluorescence is not common to most

chemical species. Fluorescence also demonstrates enhanced sensitivity with

.mchcozc AGE;

288% new 20“.: ozfim Lo.— mo§m umgoxo new 20“.: 2505 m: E xanoo Lozoconcéozaeoac m .2 new 8.8
$5 8:96 new Gui 959m 865m m: E 90:88:: a .2 So $505 835 o_co.=oo_o of. .8868: 626698:
E8659 8:: c965 8983 .8865: 62568 E8659 8:: Sow .mchcoaa 8:88.53: 3: 2E8?

new av: ozflm ucm .Cv 8:88.03: .EZ :23me .mcoszS 6.3 5:988 @5838 E9926 28056.. ._.._. 959“.

can:

11ll
on: lr

L7
L7
.. I

 

 

 

L7

 

 

 

L
V

 

 

r-------------------+--”i

 

 

 

 

 

 

 

 

 

 

 

 

 

-————-——————_—____———__-.__..’

41-}
f-WN

 

 

 

 

 

 

 

 

 

 

”0+

 

Or“:

 

respect to absorbance methods because the fluorescence power, which is
dependent on concentration, is also dependent on the source power.2
Absorbance methods, in contrast, relate concentration to a ratio of the source
power before and after sample interaction. Therefore, the sensitivity of a
fluorescence technique can be increased by use of a higher power source, such
as a laser. For these reasons, fluorescence methods are typically one to three
orders of magnitude more sensitive with expanded linear ranges relative to
absorbance-based methods.

Fluorescence quenching is a related technique that allows detection of
non-fluorescent species while taking advantage of many of the benefits
discussed above. The term “fluorescence quenching” describes any event in
which the observed fluorescence signal is decreased.3 When the decrease in
fluorescence results from a change in the optical properties of the sample, the
quenching is considered trivial and nonselective.4 Trivial quenching typically
arises from refractive index effects or from primary and secondary absorption,
which can be reduced by using a small optical path length of the sample.5 More
selective quenching arises from specific interactions between a non-fluorescent
compound, such as an explosive, and a carefully chosen fluorophore. Dynamic
quenching involves the transient collisional interaction between an excited-state
fluorophore and a ground-state quencher (Figure 1.1, right). In contrast, static
quenching involves the formation of a ground-state fluorophore-quencher
complex (Figure 1.1, left). Because complex formation between the fluorophore

and quencher is required, both dynamic and static quenching are more selective

and, thus, more analytically useful than the trivial mechanisms. By careful choice
of fluorophore, the quenching interaction can be tuned to provide selectivity for a
variety of analytes and to reduce potential interferences.

Dynamic quenching is described mathematically by the Stem-Volmer
equation6

0 0
%—z-i—=1+kdtgcq=1+Kqu (14)
f

where the quantum efficiency of the fluorophore in the absence (CD?) and
presence (<1)f ) of a quencher is related to the Stem-Volmer constant (Kq) and the
concentration of the quencher (Cq). The measured fluorescence power ratio
(FOIF) is directly proportional to the quantum efficiency ratio ((1)9/ (1),) when the

source power, efficiency of optical irradiation and collection, and fluorophore
absorbance remain constant.4 A graph of the fluorescence power ratio versus
the quencher concentration is linear, with a slope equal to the Stem-Volmer
quenching constant and an intercept of unity. A similar relationship is observed
for static quenching, but the quenching constant is the equilibrium constant for
ground-state complex formation between the fluorophore and quencher. The
Stern-Volmer constant (Kq) can also be expressed as the product of the second-

order rate constant for quenching (kq) and the lifetime of the fluorophore in the
absence of quencher (1?). Because dynamic quenching is a collisional process

that occurs during the excited-state lifetime of the fluorophore, the fluorescence

lifetime changes in the presence of quencher molecules. Hence, measurement

of the fluorescence lifetime in the presence and absence of a quencher can be

used to elucidate whether the mechanism is static or dynamic.
1.2 Luminescence-Based Methods in Forensic Science

Luminescence techniques are widely used by forensic biologists, but
very few of these methods have been adopted by forensic chemists. One well-
known luminescence technique is the detection of blood and body fluids by the
oxidation of luminol to produce chemiluminescence.7 A solution of luminol is
sprayed in areas where bloodstains are suspected to occur, and the reaction of
iron from heme with luminol is observed at 425 nm in the absence of ambient
light. Another common use of fluorescence in forensic science is the
visualization of latent fingerprints. When sprayed with a ninhydrin solution, the
amino acids contained in the fingerprint residue react to form a ninhydrin dimer
that is purple in color.8 In the presence of a metal salt such as ZI'ICI2 or CaClz,
the dimer forms a metal complex that is fluorescent under visible light (500 - 530
nm).8 Other reagents for fluorescence detection of latent fingerprints have been
developed and include magnetic powders,9 CdSe—ZnS quantum dots,10 and 1,2-
indanedione.11 The visualization mechanism has also been adapted for
determination of latent lip prints using a solution of Nile Red.12

In addition to these commonly used tools, a limited amount of work has
investigated the utility of fluorescence-based techniques for analysis of other

forensic samples. A number of methods have been identified for determination

1 3-20 21 ,22 23-26

of drugs of abuse such as amphetamines, opiates, and hallucinogens

17-23.26-29

using native fluorescence,14‘17'2“'25 derivatization, and fluorescence

quenching.13 Although these methods have been reported in the scientific
literature, traditional gas chromatography-mass spectrometry and infrared
spectroscopy methods are more commonly used in routine forensic analysis of
these drugs of abuse. Several luminescence-based methods have also been
introduced for trace evidence analysis. Some other unique applications include
fluorescence imaging of trace evidence (paints, tapes, adhesives, inks, firearm
propellants)?o age determination from fluorescence of dentin in teeth,31 and
fluorescence imaging of accelerants32 and gunshot residue.33 Despite the wide
use of luminescence-based methods by analytical chemists, the forensic science
community has yet to fully utilize these techniques within state and local
laboratories. The analysis of explosives and related compounds, however, is
one area in which a variety of luminescence-based techniques have been

explored and applied with great success.
1.3 Explosives

The analysis of nitrated explosives is of great importance in a variety of
fields.1 In forensic science, detection of these explosives and their degradation
products can be used to identify persons in recent contact with an explosive
device.“ In addition, in environmental chemistry, explosives present in
undetonated landmines can leach into and persist in the soil and groundwater,

35'“ In industrial

posing a threat to the environment and any living inhabitants.
quality control, manufacturers of explosive materials must assure consumers that
their products are safe, effective, and free from contamination by monitoring the

composition throughout the manufacturing process. Because the requirements

in each of these fields are diverse, a versatile method is needed for the analysis
of nitrated explosives.

Of these, the most widely studied are nitrated explosives, including
nitroaromatics such as TNT, nitramines such as RDX, and nitrate esters such as
PETN (Table 1.1, Figures 1.2-1.3). Military explosives are often composite
materials consisting of several of these explosives together with fuels and other
excipients (Table 1.2).1 Extensive characterization of these explosives has led to
extremely low detection limits in a laboratory setting. In the field, however, where
untrained personnel must identify explosive compounds in the presence of
abundant contamination, a simple and effective system is needed to detect these
compounds at trace levels. Frequently, detection of hidden explosive devices is
difficult owing to the low vapor pressures of many explosives.

Rising interest in defense and homeland security has led to improved
technologies for military explosives detection. These improvements have, in turn,

led to increased use of non-nitrated explosive materials.37

Most recently, liquid
or peroxide-based explosives such as TATP and l-IMTD have been involved in
cases of terrorism or drug-related crime (Table 1.1, Figure 1.3).37"9 Although
these compounds are easily synthesized from readily available materials and are
comparable in power to TNT, their instability and rapid sublimation make them of
little use in military applications. Because these materials are new and

significantly different from other classes of explosives, the best methods for their

determination have yet to be identified.

 

z<
:FEI
oe<e
2000
zemo
oz
:02
x2:
xom
seer
28 2.8
FZP
EQEo
ezo
ez
02»
02:
02
022:
22
5.6.558

“80985
8_x0.0n_
8_x0.0n_
.88 982
.80 982
.880 982
8_E0.=z
8_E0.=2
0560.32
8_:.0.=2\0_.0E0.00.u_2
88.0.0052
0=0E0.8.:z
88.0.0052
88.0.0052
88.0.0022
88.0.0052
0_“0E0.00.:z
0385.00.32
2.29.882
208.822
80.0 0.30.5.5

AC 90.x: E:_:0EE<
Amy 8_E0_0 820.09: 80350E0x01
Ev 820.83 8980....
.9 98% .820 8850
E 98:89 0.254389.
6: 5.82082
22v 8258932
:3 802002058082-...m.m. To.::0..0._.-n.m.m€
:v oeoxo;o_o>ooNooE-m.m. 72.2.30. F
o: 8:co_>EoE-z-o..Eo..ohz.o.ed
3 _oco;aoe_cE-o.e.N
e ocoaeoeathoed
Ev ocoa_o.o.._:_e-o.N-oc_E<-e 852985.90.#oc_E<-N
.9 8822850.“ 8822235040
E 86292.2-.. 8529232-..” 85292.20
E oeoNcoaoEeEeg
ADV 88:82:50-0 .80~:80.=:_Q.E .0:0~:80.._:_Q.0
A0: 88:80.:2
Ev 8939.53.«4.58580
As 805882
8:09:00

 

08:00:80 090:... 0:0 0028506 .0. 0.30.885: coEEoo .5. 030...

~~o 8 0
t i
*ji 7*

Figure 1.2. Structures of common nitroaliphatic and nitroaromatic explosives
and degradation products. NM (A), DMNB (B), NB (C), o-DNB (D), TNB (E), 2-
NT (F), 2,6-DNT (G), 2-am—DNT (H), TNT (I), Picric acid (J), and Tetryl (K).

10

N
02N\ /\ /N02 \N/NOZ NH
N N M
L k U /N J N 02N\
N OZN \ N m NH2
I \
N02 N02
T02 02N\ /N02
0 O O
02N\ /\/O\
o 0 N02
0 N
2 \O P Q
0
\
N02 /0 0\
02N No2
/ O
80 ° 7 o
o\ >< QNN. + _

Figure 1.3. Structures of common nitramine, nitrate ester, peroxide, and
inorganic explosives. RDX (L), HMX (M), NGU (N), NC (0), PETN (P), EGDN
(Q), TATP (R), HMTD (S), and AN (T).

11

 

3.8.. + ._.z._. .890...

 

25: + 52 + x9. 82...
:28 + 222 + x8 EC:
22» + 25: 898:
Eh + x2: 80
82260.: + 2.50 20800800
52 + x8 .9296
__o 8:. + 982828....58828 + 22338.28 + x8 .48 #0 8:888
82208.2: + 5.8 + #2 + 220+ eze + x8 28v no 8.2.0.858
oe_EeE.o._§oE_e + 82208.2: + 82 + 220+ Eze + x0: a2 8 8:88:50
x8. + ._.z.. + xom Am anOV m :28an00
98:: EECOEEG + HZ... _oumE<
:o=0_:E.on_ 0E0: :oEEoo

 

.8852:th 0:0 820208 302:... coEEoo 0203:. .N... 030...

12

In the laboratory, nitrated explosives have been separated by thin-layer

1 .40 1 ,40-45

chromatography, gas chromatography, supercritical fluid

1,5,35,40,47-50 1,51

chromatography,”46 liquid chromatography, ion chromatography,

34.52

capillary electrophoresis, capillary electrochromatography,53 and micellar

electrokinetic chromatography.‘°’6'54 Detection and characterization of nitrated

explosives have been accomplished by colorimetric reactions,1"“"55

0 1,35,40,47,48.50 40.56.57

densitometry,”4 UV absorbance, infrared spectroscopy,

0 1.40.50 1 ,40,58-60

nuclear magnetic resonance,"4 electrochemistry, immunoassay,

1.61.62 1,43,61,62

mass spectrometry, ion mobility spectrometry,

4345.58.63 3653.54.64

chemiluminescence, indirect laser-induced fluorescence, and
fluorescence quenching.5'6574 Unfortunately, many of these methods are difficult
to adapt for field use owing to the size of laboratory-scale instrumentation and
the need for other accessories that limit portability.

In view of these limitations, explosives are most frequently detected on-
site by trained canines75 or by ion mobility spectrometry.61 Canines require
extensive training and continuous care. and IMS is a destructive technique.
Because forensic analyses require confirmation by a second independent
technique, sample preservation is critical in cases involving trace quantities of
explosive residue. In addition, IMS is plagued by a limited linear dynamic range,
interferences from complex matrixes, and poor resolution. Further advances in
non-destructive detection techniques for explosives screening will overcome the

disadvantages of many of these methods, and can only serve to improve current

technology.

13

1.4 Luminescence-Based Methods for Explosives Detection

As discussed previously, fluorescence- or luminescence-based methods
for explosives detection offer many benefits over other common techniques. For
example, fluorescence methods are typically one to three orders of magnitude
more sensitive than absorbance-based methods. Although mass-spectral
techniques are also common for explosives detection, instruments with the
required sensitivity have not been developed as field-ready and user-friendly.
Fluorescence methods require only a source and detector, both of which can be
incorporated easily into a handheld device for field detection of explosives.

Unfortunately, many explosive compounds are not inherently fluorescent.
Explosives such as nitramines, nitrate esters, and peroxides have non-
conjugated structures that allow efficient vibrational relaxation. Nitroaromatic
explosives, despite having aromatic structure, do not fluoresce as a result of the
strong electron-withdrawing nature of the nitro substituents. This lack of native
fluorescence notwithstanding, many techniques have been proposed in recent
years for explosives detection using luminescence-based methods. Several
direct techniques have been investigated, such as high-energy excitation
methods including gamma resonance technology and x-ray fluorescence. Other
direct methods involve reactions to form fluorescent and chemiluminescent
products. Indirect methods have also been used in which the fluorescence of a
secondary species is quenched by the explosive. Quenching methods have
been utilized with solution-phase, immobilized, and solid-state fluorophores.

Other indirect methods rely on changes in concentration of a secondary

14

fluorescent species, including indirect laser—induced fluorescence and
fluorescence immunoassay. These methods, as summarized below, have

contributed significantly to the current status of field detection of explosives.
1.4.1 Direct Detection Methods

As discussed previously, explosive compounds are not inherently
fluorescent upon excitation by UV or visible light. Despite this limitation, many
direct methods have been explored for the determination of explosives. Some
techniques utilize high-energy excitation sources, such as x-rays or gamma rays
to generate luminescence. Other direct methods involve a chemical reaction
resulting in decomposition or the formation of fluorescent products.
Decomposition products are detected by fluorescence emission following
excitation at deep-UV wavelengths or by using a chemiluminescence reaction
with ozone. In all of these methods, the explosive compound or its products are

detected as the source of the fluorescence signal.
1.4.1.1 High-Energy Techniques

X-ray fluorescence (XRF) is a high-energy excitation method that has
been investigated for determination of nitrated explosives from a distance without

direct contact (i.e.. “standoff detection”).76

In this method, a tungsten x-ray
source excites the sample and the return fluorescence signature is captured by a
germanium crystal detector. Fluorescence profiles are generated by nitro
functionalities, where the XRF results from the ejection of an outer-shell electron

of a nitrogen atom and replacement by an electron from a higher energy shell (E

z 0.4 keV). The XRF is increased above the background level of soil or other

15

components in the surrounding environment in the presence of excess nitrogen
atoms from explosives or their residues. XRF spectra have been used for
identification of TNT, RDX, PETN. tetryl, C4, Comp B, black powder, smokeless
powder, flash powder, ammonium nitrate-fuel oil, detonator cord, and several
landmine types. Blair and coworkers report 100% accuracy in field testing for
identification of 10 newly buried landmines, and the method is not affected by
environmental factors such as wind or presence of vegetation.76 Explosives
residues are also detectable up to six months after removal of the landmine. The
current technology allows detection of these explosives from a distance of 2 m.
Gamma resonance technology (GRT) is another high-energy excitation
method for detection of nitrated explosives that involves nuclear resonance
fluorescence.”79 ln GRT, monoenergetic gamma rays that match excited levels
of 14N (E z 9 MeV) are produced by in-flight annihilation of fast positrons. When
focused on a suspect object, gamma rays are absorbed and reemitted as nuclear
fluorescence in areas containing high nitrogen concentrations, such as near the
location of a buried landmine. This technology is most useful for scanning large
cargo by generation of a real-time, three—dimensional map of nitrogen content as
cargo is passed between the source and detector. Although nuclear resonance
fluorescence is a potential method for explosives detection, a similar
absorbance-based technique has dominated field studies as a result of the
higher resonance cross section.77 Overall, the GRT method is limited by the size
of the required instrumentation, namely a particle accelerator and nuclear

reaction chambers. Handheld XRF instrumentation is commercially available

16

(e.g., Therrno Fisher Scientific, Bruker Biosciences, Innov-X Systems, and
Oxford Instruments), making XRF the most promising tool of the high-energy

excitation methods for explosives detection.
1.4.1.2 Reaction-Based Methods

Another approach to direct luminescence detection is to decompose or
fragment the explosive prior to detection. Many of these methods involve
decomposition using UV photolysis (220 — 272nm) or pyrolysis (675 — 1000 K) of
the nitrated explosive to NO, followed by measurement of laser-induced
fluorescence. The (A22+(v'=0) «— X2H3/2(v"=2)) transition of NO is monitored
from 245 - 248 nm with excitation at 226 nm by using a frequency-tripled

Nd:YAG laser in conjunction with a frequency-doubled dye laser.8°‘86

Using
photolytic decomposition, reported vapor-phase detection limits are 25 ppb NB,84
13 ppb DNB,82 3.7 ppb DNT,82 and 8 ppb TNT.85 A laboratory detection limit of
40 ppb TNT81 and a field detection limit of 4 ppm TNT80 are reported for spiked
soil samples. Swayambunathan and coworkers report similar detection limits for
TNT, but no signal for RDX and PETN.86 The authors speculate that the lack of
signal is a result of preferential decomposition of RDX and PETN to N02 instead
of NO. By using low-temperature pyrolysis (300 — 473 K) instead of photolysis to
generate NO radicals, low ppm detection limits are reported for RDX and
PETN.“6

Chemiluminescence is an alternative method for NO determination

following decomposition of an explosive. Chemiluminescence detection of

explosives is most commonly performed with thermal energy analysis (TEA).

17

The TEA method involves high-temperature pyrolysis (675 — 1000 K) of
explosives vapors to produce an NO radical that is oxidized by ozone to form
electronically excited nitrogen dioxide (N02*).87 Upon relaxation, the N02* emits
at near-infrared wavelengths (600 — 800 nm). Portable explosives detectors
based on this technology have been developed at Therrno Electron Corporation
and are commercially available.87""’2 The utility of TEA-based field detectors has

been reviewed recently,“93

and initial studies report ppb and sub-ppb detection
limits for nitrated compounds and a linear range extending five orders of
magnitude.“"’7‘89’94'96 Detection of nitrated explosives in post-blast debris, air
samples, solvent extracts of handswabs, and spiked blood has also been
successful.“5"9°'91'97 Miller and coworkers have also used this technique to map
contamination of work areas after sampling with Teflon wipes.43

The utility of TEA for detection of explosives following chromatographic
separation has also been investigated. Following gas chromatography (GC), a
detection limit of 2.6 ppb DNT in a supercritical fluid extract of soil is reported by
Francis and coworkers.98 Crowson and coworkers report low ng detection limits
for NG, PETN. TNT, and RDX in solvent extracts of handswabs analyzed by GC-
TEA.99 To further improve sensitivity and selectivity following GC, the blue-green
chemiluminescent emission observed upon heating of explosive vapors, rather

than the near-infrared emission, has also been investigated.100

Signals are
enhanced in the presence of helium or nitrogen carrier gases. Although no

signal is observed for TNT, detection of PETN, RDX, and NG, as well as a

18

number of related nitrated compounds such as isopropyl nitrate, n-propyl nitrate,
and ethylnitrate is reported.100

TEA has also been utilized following chromatographic separations using
condensed mobile phases. Bowerbank and coworkers report the coupling of
TEA with solvating GC, utilizing a packed column with a carbon dioxide mobile
phase for determination of DNT, TNT, NG, and PETN.42 Using TEA, NG
detection limits of NG are 10-fold less than those of the common flame-ionization
detector, with a linear range from 0.1 — 0.5 ppb NG. Coupling of TEA with
capillary supercritical fluid chromatography (SFC) has been investigated by
Francis and coworkers101 and Douse.46 Following SFC, TEA detection limits are
reported in the pg range for TNB, DNT, TNT, tetryl, HMX. RDX, PETN, and NG,
with linearity from 1 - 10,000 ppm.”101 Explosives detection using TEA has also
been adapted for use following liquid chromatography (LC) by Lafleur and
Morriseau102 and Selavka and coworkers.103 Detection limits of 4 ng RDX and 6
ng PETN are reported, with a linear range extending three orders of

magnitude.102

The major limitation of LC-TEA is that the high temperatures
required for pyrolysis of explosives result in increased background noise and
decreased sensitivity and selectivity.103 Selavka and coworkers demonstrate that
NO radicals generated by using UV photolysis, rather than pyrolysis, have
reduced baseline noise and increased sensitivity for nitroaromatic compounds.103
Using UV photolysis with LC-TEA, a detection limit of 50 pg TNT is reported.
Detection of nitrite ions (NOz‘) has also been utilized for determination of

104.105

nitrated explosives. In a method described by Nguyen and coworkers,

19

nitrite ions are generated by pyrolysis of RDX at NH2 0

temperatures above 673 K. Gaseous nitrite ions oxidize

Z—Z
I I

luminol (1) to aminophthalic acid in the presence of sodium

sulfite to produce chemiluminescence at 425 nm.104 The 1 0
system is currently designed to signal an alarm in the presence of TNT, RDX,
PETN, NG, EGDN, DMNB, and AN. The pyrolysis of TATP also produces
oxidizing agents, resulting in the same chemiluminescent reaction and signal
generation. In another approach, Lapat and coworkers generate nitrite ions by
decomposition of RDX in basic solution at elevated temperature (333 K). In one

method, aqueous nitrite ions are reacted with 6-amino-4-chloro-1-phenol-2-

sulfonic acid to form a diazonium salt, which reacts with resorcinol and Ga3+ to

form a hi hi fluorescent lum allion- HO O
Ga3+ complex (2) (Max = 495 nm, kEM = Ill
/
N
608 nm).105 Using the fluorescence of H
2
this complex, RDX is detectable from CI

10 ppb to 3 ppm.105 In a second method, aqueous nitrite ions are reacted with 4-
aminofluorescein to form a diazonium salt, which is converted to a highly
fluorescent diazotate (1.5x = 492 nm, 7.5M = 518 nm) in the presence of base.
Using this derivatization, RDX is detectable from 5 ppb to 3 ppm.105

Another direct method involves reduction of electron-withdrawing nitro
substituents on explosives to electron-donating amine substituents, without
further decomposition or fragmentation. Replacement with amine groups

increases native fluorescence by increasing the electron density of the aromatic

20

system. Eastwood and coworkers use carbon-nanotube supported Pd particles
for reduction of nitroaromatic compounds followed by laser-induced fluorescence
(2.5x z 287 nm, XEM z 342 nm).1°‘3'107
Alternatively, the amines can be derivatized

with fluorescamine (3) for more sensitive

 

fluorescence detection (lax z 394 nm, kEM ==
487 rtm)."’6'107 Limits of detection as low as 1 pM DNT are achieved by using
fluorescamine derivatization.106

Bruno and coworkers utilize electrochemiluminescence after
complexation to Au” or Cu2+ ions for determination of diaminotoluenes (DATs).108
When a current is passed through a solution containing Au-DAT or Cu-DAT,
emission from the complex is observed. Detection limits reported for this method
are less than 1 ppb for various DATs related to TNT.108 Mohammadzai and
coworkers have developed an extraction and detection method for
aminonitroaromatic compounds based on reverse micelle-mediated

chemiluminescence (RMM-CL).109

In this method, the pH of the sample is
increased to protonate the amine groups, and chloroaurate ions, AuCl4‘, are
added as ion pairing agents. The ion pairs are extracted into dichloromethane
and introduced into a micelle/luminol mixture. Oxidation of luminol (1) to
aminophthalic acid by AuCl4‘ generates a chemiluminescence signal at 445 nm
that is proportional to both the AuCl4‘ and aminonitrotoluene concentrations. The

detection limits for aminonitrotoluene and dinitroaniline are 0.1 pM with a linear

range of 100 pM - 10 mM.

21

1.4.2 Indirect Detection Methods

An alternative approach to luminescence detection of non-fluorescent
explosive compounds is measurement of an indirect effect on a secondary
fluorescent species. In these indirect methods, the explosive is not involved in
the fluorescence process, but affects the fluorescence of another compound.
One type of indirect detection method is fluorescence quenching, in which the
explosive reduces the fluorescence power of a compound through a collisional
interaction or formation of a ground-state complex. Fluorescence quenching has
been investigated in both the solution phase and solid phase. Solid-phase
materials include fluorescent molecules immobilized on a non-fluorescent
support as well as inherently fluorescent solids such as polymers and quantum
dots. Other indirect methods have also been used for detection of explosives,
including simple spatial or charge-based displacement of a fluorophore, as well
as competitive antigen-antibody displacement. Finally, the explosive may
undergo reaction with another compound that subsequently becomes fluorescent.
In each of these methods, the explosive is not the source of the fluorescence

emission, but instead alters the fluorescence of another species.
1.4.2.1 Quenching of Solution-Phase Fluorophores

A variety of inorganic and organic fluorophores have been studied in the
solution phase for fluorescence quenching detection of nitrated explosives. Book
and coworkers have investigated the quenching of ruthenium tris-bipyridine
(Ru(bpy)32+) fluorescence (1.5M = 620 nm) by nitroaromatic compounds in

acetonitrile using a standard fluorimeter."°’112 Quenching is greatest for m-DNB

22

and p-DNB, with rate constants of 109 - 1010 M" s", but is also observed for NB
and o-DNB, with rate constants of 103 - 105 M" 5".”0 The static quenching

2* to the nitroaromatic

mechanism involves electron transfer from Ru(bpy)3
quencher. Although the focus of this work is on the measurement of excited-
state redox potentials of Ru(bpy)32+, the potential of this inorganic fluorophore for
the detection of nitrated explosives has clearly been demonstrated.

Glazier and coworkers have expanded on the solution-phase work of
Book and coworkers"°"12 by incorporating Ru(bpy)32* into poly(amidoamine)
(PAMAM) dendrimers.113 The fluorescence of a solution of Ru(bpy)32+-modified
PAMAM in ethanol (Max = 450 nm, XEM = 575 — 640 nm) is quenched by TNT,
with a quenching constant of 2500 M". Relative to TNT, the quenching constant
for DNT is 3 to 4 times less at 700 M", and no quenching is observed for NT.
Although this approach may be useful for TNT detection. the inability to detect
other related compounds is limiting for field applications.

The solution-phase quenching of organic fluorophores has also been
studied.5'65 Goodpaster and McGuffin report that nitrated
explosives are effective quenchers of pyrene (4)
fluorescence (15x = 325 nm. 1.5M = 375 nm). 4
Nitroaromatics are the most efficient quenchers, with rate
constants of 3x1010 M" s", as a result of the stable charge-transfer complex that
is formed with pyrene.5 Nitramine and aliphatic explosives also quench the

fluorescence of pyrene, but to a lesser extent with rate constants from 5x109 —

1.5x1010 M" s". Quenching of pyrene is used as a detection method following

23

LC with detection limits of 44 ng explosive. Focsaneanu and Scaiano also
studied fluorescence quenching of excited-state pyrene by nitrated compounds.65
These authors compared the quenching of monomer fluorescence (lax = 355 nm,
REM < 400 nm) and excimer fluorescence (lax = 355 nm, XEM = 470 nm) by
nitroaromatic compounds. Similar to the results of Goodpaster and McGuffin,
nitroaromatic explosives demonstrate greater quenching than aliphatic
explosives. Another solution-phase organic fluorophore with promise for
detection of explosives is meso-tetra(4-sulfonatophenyl)porphyrin (TPPS).114
The fluorescence of TPPS (Max = 413 nm, XEM = 645 nm) in pH 7 buffer
decreases in the presence of TNT, with a quenching constant on the order of 103
M" and a detection limit of 200 ppb TNT. The proposed static quenching
mechanism involves binding of TNT to TPPS in a 1:1 ratio. Although pyrene has
been identified as the most sensitive and selective fluorophore for quenching
detection of nitrated explosives, the less toxic alternatives are more favorable for

routine use.
1.4.2.2 Quenching of Immobilized Fluorophores

For greater utility in field detection of explosives by fluorescence
quenching, it is beneficial for the fluorophore to be in the solid 5 SO3H
state. The use of solid-state materials avoids possible leakage
and evaporation that can occur when working with solutions.

Fang and coworkers have investigated the quenching behavior
/N\
of dimethylaminonaphthalenesulfonic acid (dansyl) and pyrene

immobilized on glass substrates coated with a self-assembled monolayer for

24

detection of nitrated explosives.”115 The fluorescence of a dansyl (5) film (kgx =
340 nm, AEM = 505 nm) is quenched by aqueous solutions of NB, NT, DNB, and
TNT, but not by NM or aromatic compounds such as benzene.115 The response
of this sensor is fast and reversible, with a detection limit of 1 pM TNT in aqueous
solution. The fluorescence of a pyrene (3) film (1.5x = 355 nm, 7.5M = 500 nm) is
quenched by gaseous TNT, but more significantly by gaseous NB.69 Again, no
interference is observed for the vapors of common organic solvents. In both
cases. a static mechanism is proposed based on electron transfer from dansyl or
pyrene to the nitrated explosive.

Other pyrene-related compounds have also been utilized for explosives
detection by incorporation into membranes or thin films. Jian and Seitz utilized a
cellulose acetate polymer membrane containing (CH2)3COOH
pyrenebutyric acid (PBA, 6) for fluorescence quenching
detection of explosives.116 The fluorescence of PBA (2.5x =
354 nm, XEM = 412 nm) is quenched in the presence of 6
aqueous TNT, DNT, and RDX. Despite a slight sensitivity
to dissolved oxygen, these sensors are reusable and have detection limits of 2
ppm TNT and DNT and 10 ppm RDX. Because the membrane acts as an
extraction device, additional selectivity is gained in the analysis of complex
natural water samples. The response time is quite long, however, with a 40—min
sample exposure required for constant signal. Pyrenebutyric acid is photostable
for 8 hours of continuous exposure to the light source. but some leaching into

water is observed after 72 hours. Kumar and coworkers have incorporated

25

pyrene-based fluorophores into electrospun polymer films for detection of
nitroaromatic compoundsm'118 For PBA (1.5x = 275 nm, kEM = 395 nm)
immobilized in poly(4-hydroxybenzyl alcohol), a solution of DNT in
dimethylsulfoxide reduces the fluorescence with a CHZOH
quenching constant of approximately 104 M".118 For 1-
pyrenemethanol (7, lax = 350 nm, 7.5M = 450 nm)
copolymerized with poly(acrylic acid) in an electrospun thin
film on a glass slide, the quenching constant for aqueous
DNT is even larger at 106 M".117 For both systems, the quenching is proposed to
occur through electron transfer from the pyrene compound to DNT. Of all
membrane-bound organic fluorophores, the PBA membrane developed by Jian
and Seitz reports the lowest detection limits and is the most useful for aqueous
explosive detection.

Quenching of benzo[k]fluoranthene (8).
an organic fluorophore similar in structure to 8 0.0C
pyrene (4), is used for explosives detection by Q
Patra and Mishra.119 Poly(vinyl alcohol) films
containing benzo[klfluoranthene (kEx = 308 nm, 7.5M = 435 nm) supported on a
glass slide have been developed. These sensors give a reproducible and
reversible response for nitrated explosives, with strong quenching from 0.1 — 1
mM solutions of DNB, NT, nitrobenzoic acid, nitroaniline, nitrobromobenzene,
and nitrophenol in methanol. The sensors are most sensitive to p-nitrophenol,

with a detection limit of 10 uM in methanol. The greatest interferences are from

26

acetone (51.6 % quenching relative to 60 mM NB), benzylalcohol (15.6 %), and
acrylic acid (11.0 °/o). The sensors have a response time of 2 — 10 s and some
photobleaching of benzo[kjfluoranthene is observed after 1 h of constant source
exposure.

Other organic fluorophores have also been investigated in an
immobilized form for the determination of nitrated explosives. Yu and coworkers
have incorporated o o

curcumin (9) into a \ /

poly(vinyl chloride)

HO OH
membrane for the

O

0
detection of \ /

nitrophenols such as picric acid.120 The fluorescence of curcumin (kEx = 426 nm,
AEM = 512 nm) is quenched most effectively by aqueous o-nitrophenol because
ground-state complex formation is more sterically hindered by other nitrophenols.
The detection limit for this sensor is 0.08 mM o-nitrophenol with a linear range of
0.15 — 10 mM. Other nitrophenols are also detected to a lesser degree, including
m- and p-nitrophenol, 2,4-dinitrophenol, and

picric acid. Preparation of the sensing film can NH2
be simplified by incorporation of a fluorophore

directly into the polymer structure. This has /T 10

been demonstrated by incorporating 3—amino-9—

ethylcarbazole (AEC, 10) (1.5x = 275 nm, 1.5M = 384 nm) into the monomer 3-(N—

methacryloyl) amino-9-ethylcarbazole (MEAC).121 A thin film is prepared on a

27

glass slide by copolymerization of MEAC with 2-hydroxypropyl methacrylate.
The fluorescence of the film is quenched in the presence of aqueous picric acid
through formation of a 1:1 ground-state complex between picric acid and AEC.
The quenching response is linear from 10 pM — 10 mM picric acid, and no signal
is observed for a number of common interferences. In addition, the sensor is
stable for more than two months, and leaching of the dye into water is not
observed.

Another approach for fluorophore immobilization is by incorporation into
a silica monolith. Li and coworkers have developed optical sensors for TNT and
related compounds by using this methodm":Z4 Using polymer beads as a
template, a monolith with a bimodal pore structure is formed on a glass slide.
During monolith growth, a porphyrin dye is incorporated into the silica gel matrix,
yielding a fluorescent thin film. Upon exposure to saturated TNT vapor (10 ppb),
the fluorescence from the porphyrin dye (hex = 420 nm, XEM = 660 nm) is
decreased by 50% after 10 s, and is decreased by 97% within 2 min.124
Quenching is also observed upon exposure to DNT and NB vapors.123'124 The
quenching mechanism is believed to involve hydrogen bonding between
porphyrin and the nitro group on the explosives. Incorporation of metallic
character into the porphyrin-doped films improves the quenching response of
TNT by increasing the binding constant with the film. Films containing Cd are
superior for TNT sensing, with 56% quenching within 10 3.124 In comparison, Zn-
containing films demonstrate 43% quenching and undoped films demonstrate

33% quenching in the same amount of time. In all cases, the quenching is

28

reversible and the fluorescence can be regenerated by purging with nitrogen to
remove explosive vapors.

Polymer and silica beads have also been used as a solid support for
metal—doped organic fluorophores. Lu and Zhang have developed an optical

sensor for aqueous picric acid based on fluorescence

11
quenching of thenoyltrifluoroacetone (TTA, 11) complexed \ S

O

with Eu3+ immobilized on a chelating resin, Chelex 100.125

By coordination of Eu3+ to the carboxylic acid groups on O CF3
the poly(styrene—divinylbenzene) resin. water ligands are removed and the
luminescence of the film (2.5x = 352 nm, 9.5M = 612 nm) is increased. Chelation of
the hydroxyl group on picric acid to Eu-TTA results in static quenching and
fluorescence is decreased. A detection limit of 2 pM picric acid is reported.
Sensors utilizing Eu-‘l‘l’A fluorescence quenching have a lifetime of more than
one month and are insensitive to a wide range of interferents including organic
acids and inorganic salts.

Sensors containing immobilized fluorophores for analysis of volatile
organic compounds‘zr"129 have been adapted for vapor-phase explosives

detection by Walt and coworkers.

These sensors incorporate Nile Red

0 O N\/
(12) (XEX = 500 nm, XEM = 580 nm),130- ’/
\ 12
‘33 indodicarbocyanine Cy5 (13) (AEX N

= 577 nm, item = 670 nm),134 and

Texas red cadaverine (14) (kgx = 577 nm, 1.5M = 610 nm)134 into polymer

29

microspheresmmz'135 and porous silica beads.‘3°'1‘°’1'13’3'135 The sensor particles
are fixed onto a glass
- 130133135 13
slide ' ' or onto the
end of an optical fi/ / / / N

I
fiber,'3°'131'133'134 which is G// K

exposed to the explosive vapor. The response of the sensor is related to the
diffusion of the vapor into the particles and the resulting decrease in fluorescence
with time, generating a unique

profile for each analyte.133'134

N O N
/C+)
Pattern recognition via statistical O
/
9

techniques such as principal

14

66,136 or a 803

components analysis
generalized Whitney-Mann-
Wilcoxen classifier135 is used to

discriminate between nitrated O=S=O

HN\/\/\/NH2
and non-nitrated vapors in

complex samples. Using an array of particles allows signal averaging that

133 In

dramatically decreases the detection limits for explosive compounds.
addition, an array of particles containing several different fluorophores in an
electronic nose configuration increases specificity and allows simultaneous

66"“ These sensors detect nitroaromatic

detection of multiple explosives.
compounds in spiked soil and groundwater samples with detection limits of 50 —

80 ppb DNT.66 For buried landmine detection, these sensors give no false

30

positives and have a detection limit of 120 ppb DNT.”132 The sensor arrays are

also very stable, with a shelf life greater than 10 months.133

1.4.2.3 Quenching by Solid-State Fluorophores

Detection of explosives by quenching of inorganic and organic
fluorescent materials is another method under investigation. Quenching of these
intrinsically fluorescent materials is similar to quenching of immobilized
fluorophores. One example of this method is the visualization of explosives on
thin-layer chromatography (TLC) plates containing a fluorescent stationary phase.
Fisco describes a portable explosives identification kit based on such
visualization.137 Many commercially-available TLC plates contain a manganese-
activated zinc silicate fluorophore (1.5x = 254 nm, 1.5M = 525 nm). Upon
application of explosives to the plate, the explosive absorbs UV light and appears
as a dark spot on the fluorescent green background. This mechanism is
consistent with trivial quenching, and although it is less selective than other
mechanisms. it has been widely utilized in the forensic community for qualitative
analysis. Recent work by Meaney and McGuffin describes a quantitative TLC
detection method using hand-held UV lamps and digital camera imaging.138'139
This method yields low microgram detection limits of 12 nitrated explosives and
degradation products.

Quenching of photoluminescence from Si nanocrystals, another
inorganic material, by nitroaromatic compounds has also been reported.140

Gerrnanenko and coworkers measure quenching of Si nanocrystals (lax = 355

nm, 2.5M = 590 — 720 nm) in a methanol suspension by adding an aliquot of 0.042

31

M quencher solution. Quenching is observed for DNB and DNT with rate
constants on the order of 106 - 107 M" s", but no quenching is observed for NB
or NT. A static quenching mechanism is proposed in which an electron is
transferred from the conduction band of the Si nanocrystals to the vacant orbitals
of the quencher molecules. As a laboratory detection method for nitrated
explosives, quenching of Si nanocrystals is very promising. However, because
these measurements are made in solution and require a frequency-tripled
Nd:YAG laser, the utility of this technique for field detection and sensing of
explosives is limited.

Quantum dots are another type of inorganic fluorescent material that has
been investigated for detection of explosives. Nieto and coworkers investigated
the fluorescence quenching of CdSe quantum dots (Max = 400 nm, item = 534 nm)
in toluene by TNT.141 The quenching constant for TNT is estimated to be 107 M",
and the detection limit using this method is approximately 5 ng TNT. In addition
to a decrease in fluorescence, a change in the maximum emission wavelength of
the quantum dots is also described. Based on these observations, an energy-
transfer mechanism of static quenching is proposed. Similar to the limitations of
Si nanocrystals methods, quantum dot quenching measurements are made in
solution and require a femtosecond laser, therefore decreasing the utility for field
detection and sensing of explosives.

Functionalized carbon nanotubes aree another example of nanoparticles
that can be used for fluorescence quenching detection of nitrated explosives.

Kose and coworkers functionalize well-dispersed carbon nanotubes with a

32

poly(propionylethylenimine—ethylenimine) copolymer and report fluorescence
(Max = 400 nm, item = 510 nm) as a result of excited-state energy trapping by
passivated surface defects.68 A decrease in the photoluminescence of these
nanotubes in methanol is observed in the presence of nitroaromatic compounds
such as NB, NT, and DNT. Quenching constants increase from NB to NT to DNT,
but the magnitude of quenching constants is small, from 67 — 132 M".
Measurements of fluorescence lifetime indicate a significant contribution from
static quenching.

An organometallic material that has been utilized for the fluorescence
quenching detection of explosives is a polymetallole or polymeric
metallocyclopentadiene.142 The incorporation of a heavy element from Group 14
into the polymer backbone reduces the band gap and increases fluorescence
power. Trogler and coworkers have investigated the potential of porous silicon143
as well as polymetalloles containing silicon and gerrnanium70'142'144'146 for the
detection of nitrated explosives such as TNT. The fluorescence of porous silicon
thin films (9.5x = 480 nm, XEM = 650 nm) is decreased in the presence of
explosive vapors in a flowing air stream.143 Vapor detection limits of 500 ppb NB,
2 ppb DNT, and 1 ppb TNT are reported based on 5 min exposure. The
quenching mechanism is believed to be electron transfer from the silicon to the
explosive. which is most effective for TNT and DNT as a result of their more
positive reduction potentials relative to NB. Because this electron transfer occurs
upon adsorption to the silicon surface, high concentrations of NB result in

quenching despite the less favorable reduction potential. The authors speculate

33

that any compound in sufficient concentration is a potential interference and
sought improved selectivity for nitrated compounds. Additional specificity, but
decreased sensitivity, is observed by incorporation of a PtOz or PdO catalyst in
the vapor stream to decompose explosives to N02 prior to detection.

Thin films of polymetalloles are more sensitive to quenching than porous
silicon films because the conjugated, delocalized structure of the polymer is

142,144-146 For thin films Of

interrupted by the explosive vapors.
poly(tetraphenylsilole) (XEX = 340 nm, 9.5M = 513 nm), Trogler and coworkers
report quenching constants on the order of 103 M" with NB. DNT, and TNT in
seawater.144 Limited interference is reported from organic solvents or inorganic
salts. Picric acid demonstrates greater quenching than other nitroaromatics, but
the quenching is nonlinear with concentration. Nonetheless, the detection limit
for picric acid is 6 ppb in seawater. Modification of the polymetallole to contain a
germanium backbone or an alternating Si-Ge backbone leads to comparable or
decreased sensitivity relative to the polysilole described above.145 Fluorescence
lifetime studies of these systems indicate that a static mechanism is most
probable, and no interferences are observed from organic solvents or inorganic
salts. Other polymetallole configurations have also been investigated by Trogler
and coworkers for the determination of nitrated explosives.7°'146 By organizing
oligomers of polymetalloles into colloidal nanoparticles, quenching efficiency by
aqueous TNT is improved, with a detection limit of 100 ppb.146 As with other

polymetalloles. the mechanism for nanoparticle quenching is also static and

occurs because TNT disrupts colloidal aggregation. Polymetalloles have also

been sprayed onto non-porous substrates containing explosive contamination
and, upon illumination, quenching is observed.70 These devices are highly
portable and field-ready with limited toxicity once prepared. They offer detection
limits of 10 ng TNT, 50 ng DNT, and 50 ng picric acid.

Polysiloles have also been modified to contain extensive fluorination,
thereby increasing the positive charge on the Si center and increasing attraction

to N02 groups on explosive molecules.”147

Saxena and coworkers report
quenching constants for fluorinated polysiloles (1.5x = 280 nm, XEM = 335 nm) on
the order of 102 — 103 M" for picric acid, DNT, TNB, and DNB in tetrahydrofuran
solution. For fluorinated polysiloles in a thin film, quenching constants increase
to 104 M'1 for picric acid in water.”147 Trends in quenching constants for both
solutions and thin films of fluorinated polysiloles are consistent with those
observed by Trogler and coworkers.7°'1“2'14“"46 The quenching is reversible, and
no interference is observed from organic solvents or inorganic acids.

The most widely studied fluorescent materials for detection of nitrated
explosives are conjugated fluorescent polymers, first developed by Swager and
coworkers.71'72"‘“"162 The performance of fluorescent polymers is a result of the
“molecular wire” effect, in which the conjugated backbone of the polymer
promotes delocalization of electron density and migration of excited-state
electrons over large distances.“"3'149 The result is an amplified sensor response,
where a single molecule of the target analyte can result in quenching of

numerous polymer units. The amplified response was verified initially using

paraquat, an herbicide and well-known electron-transfer quenching agent.148

35

Many of the original fluorescent polymer materials are iptycene-based and are
utilized in a thin film construction. The quenching response of these polymer
films to TNT and DNT vapors is dependent on the vapor pressure of the

71,150,151

analyte as well as the charge distribution of both the polymer and

analyte.”155

Changes to the polymer backbone and side chains cause
corresponding changes in the fluorescence properties as well as the quenching
response of nitrated explosives. Incorporation of a pentiptycene backbone into
the structure of poly(p-phenylene ethynylene) (PPE) (kg = 450 nm, 7.5M = 460
nm) and poly(p-phenylene butadiynylene) (15x = 400 nm, 1.5M = 430 nm)
polymers further amplifies the response, increasing sensitivity and selectivity for
TNT.71'152"6"162 Poly(phenylene) (9.5x = 405 nm, 1.5M = 413 nm) polymers have
also been investigated with some success for the detection of the taggant 2,3-
dimethyI-2,3-dinitrobutane (DMNB), which is a required additive in all legally
manufactured plastic explosives.160 Sensitivity is also increased by pumping the
polymer films with a nitrogen laser (337 nm) to generate stimulated emission at
535 nm.72'159 The stimulated emission is disrupted in the presence of TNT vapor,
and sub-ppb detection limits are reported.7‘°"159 In all cases, the quenching of
these films is reversible and demonstrates little interference from other aromatic
compounds."'""*152 Quenching occurs through an electron-transfer mechanism
and is dominated by a static interaction between the analyte and the polymer
film.”72 These polymers have been incorporated into a field-deployable sniffer

153-159

device (FIDO) through collaboration with Nomadics, Inc The sniffer

responds to TNT and degradation products in approximately 5 s, with a detection

36

limit reported as low as 1 fg TNT in air and soil.“""158 The sniffer also
demonstrates similar selectivity to a trained canine, accurately identifying the
presence or absence of a buried mine in 90% of field tests.15“'158 The success of
further adaptation of the FIDO device for minefield edge detection is unknown, as

(1.157

appropriate field testing is yet to be complete Overall, conjugated

fluorescent polymers are the state-of-the-art for field detection of explosives and

3

are currently being used by the US. military in Iraq.16 Other conjugated

fluorescent polymers have been utilized for detection of explosives with similar
reports of sensitivity and detection limits.57'7""“""166

Other developments in the area of conjugated fluorescent polymers have
improved the range of applications of these systems for explosives sensing.
Simonson and coworkers have developed fluorescent particle sensors for
standoff detection of explosives in minefields.167 Poly(styrene—divinylbenzene)
capolymer particles are embedded with a patented PPE polymer as well as p-
bis(o—methylstyryl)-benzene (MSB) as an internal standard. The fluorescence of
the PPE component (2.5x = 355 nm, AEM = 460 nm) is generated by using a
frequency-tripled Nd:YAG laser and is quenched in the presence of explosives
vapors. These fluorescent particles are designed for distribution over soil in
areas where contamination is suspected. Quenching of the fluorescence is
observed using light detection and ranging (LIDAR) technology, in which
decreases in PPE polymer fluorescence (normalized to MSB) indicate the

presence of nitroaromatic compounds. The limit of detection using these

fluorescent particles is 1 ppm TNT in soil from a distance of 0.5 km, but fresh

37

particles must be distributed frequently as photodegradation is observed in direct
sunlight.

Chen and coworkers found that addition of a cationic surfactant to
poly(2,5-methoxypropyloxysulfonate phenylene vinylene), an anionic conjugated
polymer, enhances quenching by nitroaromatics while decreasing interference
from cationic molecules.168 Irreversible quenching by DNT is observed down to 1
nM for polymer thin films containing the surfactant, but quenching is more
reversible when the polymer and surfactant are arranged in a bilayer structure.
These investigations of conjugated polymer modifications broaden the range of

polymers that can be used for explosives sensing devices.
1.4.2.4 Indirect Laser-Induced Fluorescence

Several authors have used indirect laser-induced fluorescence (ILIF) for
the determination of various explosives. In this approach, the fluorophore is
displaced by the explosive based on either space
or charge considerations, leading to a local 15
decrease in fluorescence. Kennedy and coworkers have separated NG, PETN,
NGU, EGDN, RDX, HMX, TNT and tetryl using micellar electrokinetic
chromatography (MEKC) with detection by
ILIF.54 Four organic fluorophores are
investigated for utility as the background

COOH
fluorophore, including anthracene (15),

\
fluorescein (16), eosin (17) and rhodamine B O
O O

(18). The fluorophores are added to the H0

38

background electrolyte, producing a

constant fluorescence upon excitation with 1.,
. , COOH
an Ar laser at 488 nm and detection at 520
Br \ Br
nm. When the fluorophore is contained C
wrthin the sodium dodecyl sulfate (SDS) HO O O
micelle, radiative relaxation processes are Br Br

favored and lead to increased fluorescence. In aqueous solution, explosives
preferentially reside within the hydrophobic micelle environment as well. Thus,
the presence of the explosives will perturb the equilibrium between the

fluorophore and the micelle,

resulting in a change in 13

COOH
fluorescence. The authors
suggest that some decrease in \O
fluorescence may be the result /\N 0 \fi/\
of quenching by the explosives J g 1\

in the aqueous phase. The decrease in fluorescence is proportional to the
concentration of explosive. To determine the most favorable fluorophore, the
quantum yield upon excitation at 488 nm and the effect of the SDS solution on
the fluorescence signal is first considered. Anthracene (15) is unsuitable for this
application because no fluorescence is observed at 520 nm upon excitation at
488 nm. The fluorescence of eosin (17) decreases in the presence of SDS, while
that of fluorescein (16) and rhodamine B (18) increase in the presence of SDS.

The authors conclude that because rhodamine B exhibits a more stable baseline,

39

it is a more promising background fluorophore than fluorescein for this
application. Detection limits for HMX, RDX, PETN, tetryl, TNT, and NG are in the
range of 1 - 10 pM, but the method lacks selectivity because the presence of any
solute will cause the same disturbance.54

Bailey and Wallenborg have also developed an ILlF method for detection
of explosives following MEKC and capillary electrochromatography (CEC) using
cyanine dyes Cy5
(13) and CW (19)
to generate the

fluorescence

 

background?“53 Nitroaromatic explosives yield a large signal, while nitramines
are not detectable. Because a pure displacement mechanism would yield
comparable results for any explosive, this observation implies a large contribution
from electron-transfer fluorescence quenching. For a mixture of explosive
standards, detection limits of 1 — 5 ppm following separation by CEC and 1 - 10
ppm following separation by MEKC are reported.53 For soil extracts containing
explosive materials separated by using the MEKC method, detection limits of 1 —

5 ppm are reported.36
1.4.2.5 Fluorescence Immunoassay

Fluorescence immunoassay has been widely studied for detection of
nitrated explosives, including TNT, RDX, and PETN. All of the sensors involve a
competitive displacement immunoassay with fluorescence detection. In these

systems. an explosive-specific antibody is initially saturated with a fluorescently

40

labeled analog of the explosive molecule (antigen). The fluorescence signal
changes when explosives in the sample compete for the antibody binding site,
thereby displacing the labeled antigen.

Detection of TNT and degradation products has been explored by
Bromberg and Mathies by using a solution-phase fluorescence immunoassay in
conjunction with capillary electrophoresis (CE) on a microchip or microchannel

”'59 In these studies, anti-TNT antibodies and fluorescein-labeled TNB

plate.
antigens (16) (Max = 488 nm, 1.5M = 520 nm) are added to each sample at
constant concentration. For TNT-containing samples, unlabeled TNT binds
selectively with the antibody, displacing the labeled antigen. Following
separation by CE, the relative intensities of the peaks corresponding to the
labeled antigen and the labeled antigen-antibody complex are compared and the
amount of displaced antigen is determined. The amount of displaced antigen is
directly proportional to the concentration of TNT in the sample, with a detection
limit of 1 ppb TNT and a linear range of 1 — 300 ppb TNT.

A similar competitive immunoassay has been utilized for explosive
detection by immobilization of the antibody on a solid substrate. Work at the US.
Naval Research Laboratory (NRL) by Ligler and coworkers has focused on the
development of specific and sensitive sensors for TNT and RDX,169 with
detection limits as low as 15 ppt TNT170 and 10 ppt RDX.171 Antigens used in
these studies are labeled with fluorescein-coupled cadaverine (16) (1.5x = 490 nm,

1.5M = 520 nm), Texas Red cadaverine (14) (Max = 583 nm, 9.5M = 607 nm), or

cyanine-5-ethylenediamine (20) (kg = 640 nm, AEM = 665 nm). These

41

immunoassay systems have been incorporated into a number of field devices. In

addition to traditional
plate—based O 20 O

immunosensors,172 flow (3/ / / / N

I NH;
immunosensors have H2N\// e \\/

been developed in which the antibody is immobilized on a bead and then packed

169.173-181 60.170.171.181-184

in a cylindrical tube or bound to the inner wall of a capillary

or microchip.185

Fiber optic probes have also been developed in which the
antibody is bound to the surface of an optical fiber, allowing the fluorescence
signal to be detected in real time.""‘*192 These devices have been tested on real

183.186

samples including river, harbor. and bilge water. tap water,183

172,178-181,192-195 60.175,176,181,195 172.182.190.196

groundwater, seawater, and soil extracts
at low cost and with minimal interference. Improvement in background stability
and response of the biosensors is achieved through sample pretreatment by
solid-phase extraction.195 Two devices have been designed and marketed
commercially by Research International, Inc. for TNT and RDX detection,
including a membrane-based device (FAST 2000) for water analysis‘m"93"9‘“196
and a fiber optic sensor (Analyte 2000).192 The US. Environmental Protection
Agency (EPA) has approved Method 4655 for detection of TNT and RDX using
the continuous flow immunosensor developed at the NRL.197 This work has been
extended to the detection of PETN utilizing an anti—PETN antibody, with detection

limits of 1 ppb.17‘“77'183 All flow-based immunosensors have response times less

than 1 min. Despite the fast response time and low detection limits,

42

improvements in antibody and antigen stability at elevated temperatures and
over long periods of time are necessary to increase the utility of immunosensors.

More recently. work at the NRL by Medintz and coworkers has exploited
the strong binding between biotin and avidin for improvements in competitive

‘98 In these

TNT immunoassay.
devices, the anti-TNT antibody
and TNB antigen are both
contained in a DNA arm,
biotinylated, and bound to an

avidin surface. In addition, the

antibody fragment is labeled with

 

an AlexaFluor 532 donor dye
(21) (Max = 520 nm, 1.5M = 555
nm) while the antigen fragment is labeled with a carboxytetramethylmodamine
(TAMRA) acceptor dye (22) (9.5x = 565 nm, ABM = 590 nm). The fluorescence
emission of the system is monitored with excitation at 510 nm and emission at
600 nm. In the absence of TNT, the COOH

baseline fluorescence is high due to

fluorescence resonance energy 22

6
transfer (FRET) between the two dye COO

molecules. In the presence of TNT, O \
however, FRET is interru ted and a G)
P \N O \N/

decrease in the fluorescence signal is

43

observed. Using this method, TNT degradation products such as TNB, am-DNT,
and DNT are also detectable due to their structural similarity to TNT and
consequent affinity for the anti-TNT antibody. Limits of detection of 1 ppm TNT
and am-DNT, 5 ppb RDX, and 7.5 ppb DNT are reported, with a linear range of 1
— 25 ppm TNT.

A further extension of this work at the NRL by Goldman and coworkers
involves the use of CdSe—ZnS quantum dots as the substrate to which the

antibody is bound for competitive immunoassay.199'201

In these sensors,
quantum dots (715x = 400 nm, 9.5M = 530 nm) are derivatized with anti-TNT or anti-
RDX antibodies, and the TNB antigen is labeled with a dye such as the Black
Hole Quencher 10 acceptor (23) (has = 530 nm).201 The antibody sites are

SO3Na

23
N N
NaO3S N N N

COOH

saturated with the labeled antigen prior to introduction of the sample. In the
absence of TNT, FRET occurs between the quantum dot and the labeled antigen
and no fluorescence is observed. In the presence of TNT, however, the labeled
antigen is displaced. resulting in an increase in fluorescence from the quantum

dot. A small increase in fluorescence is observed for am-DNT and tetryl with the

anti-TNT antibody, although no response is observed for DNT. The sensors are
operational in both simulated seawater and soil extracts, with detection limits of 1
ppb Rl:>x200 and 20 pptTNT.199'201

An electrochemiluminescence enzyme immunoassay for TNT and PETN
detection has also been developed.58 In this method, antibodies corresponding
to each explosive are labeled with horseradish peroxidase (HRP) and bound to
high-affinity paramagnetic beads through an avidin-biotin linkage. The beads are
mixed with a sample containing luminol (1) and are concentrated magnetically
onto the gold-coated iron working electrode. Hydrogen peroxide is generated
electrochemically by HRP from dissolved oxygen in the solution. As hydrogen
peroxide is generated, luminol is oxidized to aminophthalic acid, producing
chemiluminescence at 445 nm. When TNT or PETN are present in the sample
and bind to their respective antibodies, the activity of HRP is decreased and
peroxide generation is suppressed, resulting in a decrease in luminol
chemiluminescence. Detection limits for this method are 0.11 ppb TNT and 19.8
ppb PETN.

Another unique approach to a fluorescence immunoassay for the
detection of TNT is based on a comparison of the photosynthetic activity of two
groups of algae.202 In this method developed by Altamirano and coworkers, the
control algae strain exhibits reduced photosynthetic activity in the presence of
TNT, as determined by the fluorescence of chlorophyll. A second strain has
been made resistant to TNT and maintains a constant level of chlorophyll

production in the presence or absence of TNT. The relative fluorescence of the

45

two strains can be related to the concentration of TNT with a detection limit of
500 ppb. Although the response time of 3 min is comparable to other
immunoassays, this method is less sensitive and less durable than conventional
immunoassays.

Another potential immunoassay for determination of nitrated explosives
is based on a yeast strain engineered by Radhika and coworkers?”204
Preliminary reports indicate that this yeast strain couples olfactory receptor
signaling to green fluorescent protein production for detection of DNT.203 In the
presence of DNT. a rat olfactory protein is triggered and increased levels of cyclic
adenosine monophosphate (cAMP) are produced. The yeast strain has been
engineered to recognize the increased concentration of cAMP and respond
through production of a green fluorescent protein (lax = 485 nm, 7.5M = 535 nm).

This pioneering work is promising for development of artificial nose technology

and field detection of explosives.
1.4.2.6 Reactions Producing Fluorescent Species

An alternative method for detection of explosives involves the redox
reaction of a secondary species by nitrated explosives to generate a fluorescent
product. Woltman and coworkers have developed a method in which nitrated
explosives are electrochemically reduced to amines following LC separation.64
These amines are combined post-column with a solution of Ru(bpy)33*, which is
reduced to the highly fluorescent Ru(bpy)32+ (9.5x = 488 nm. 7.5M = 600 nm). This

method is demonstrated using both isocratic and gradient mobile phases with Ar+

46

laser-induced fluorescence, yielding sub-nM TNT detection limits and a linear
range from 1 nM to 20 uM TNT.

Another related mechanism for detecting explosives involves the use of
an analogue of nicotinamide adenine dinucleotide (NADH). Andrew and Swager
have designed a sensor material for detection of RDX and PETN based on the
fluorescence activation of a Zn analogue of 10—methyl-9,10-dihydroacridine
(Acer).2°5'206 The fluorescence of Zn-Acer increases 80-fold in the presence of
RDX and 25-fold in the presence of
PETN upon hydride abstraction to form
Zn-AcrH‘ (24) (Max = 313 nm, 7.5M = 480 24
nm). The proposed mechanism involves

photoreduction of RDX and PETN by

 

excited-state Zn-Acer to form Zn-AcrH“, which releases a photon upon

relaxation.205

The quantum yield increases linearly with increasing [Acer],
indicating that a single Zn-complexed AcrH" is involved in the photoreduction.
The reduced forms of RDX and PETN formed during the photoreaction have not

(I.205 The sensor shows no

been identified, but N-nitroso forms are propose
response to nitroaromatic explosives such as TNT, and has no interference from
other aromatic organic compounds.

Fluorescence detection of peroxide-based explosives TATP and HMTD
has been investigated through the formation of catalytic hydrogen peroxides”39
Schulte-Ladbeck and coworkers use irradiation at 254 nm to decompose

aqueous explosive standards or extracts of exploded samples to hydrogen

47

peroxide. The generated hydrogen peroxide OH
catalyzes the dimerization of p- 25

hydroxyphenylacetic acid (25) in the presence HO

of horseradish peroxidase. This dimer is highly fluorescent at 405 nm upon

excitation with 320 — 325 nm light.”39 Detection limits for this method are 0.8
pM TATP and HMTD with a linear dynamic range from 3 — 50 pM TATP and 3 —
100 uM HMTD.38 Using this method, no false positives or false negatives are

observed among 30 real samples containing potential contaminants such as soil,

dust, and cleaning agents.
1.5 Research Objectives

Luminescence-based methods are of great importance in the search for
a sensitive and selective detector for explosives and other forensically important
compounds. A more detailed understanding of the mechanism of interaction
between explosives and sensor materials is necessary. This information will aid
in the development of new fluorescent materials with an improved ability to detect
explosives in complex field settings. The focus of this research has been the
identification of new fluorescence quenching interactions that can be used to
advance the field of forensic science. The fluorescence quenching response of
several solution—phase fluorophores in the presence of nitrated explosives is
presented in Chapter 2. For those fluorophores with a sensitive response for
nitrated explosives, the quenching mechanism is explored. The most promising
fluorophores are incorporated into solid-state materials and their fluorescence

properties explored further in Chapter 3. In a second application, the pH-

48

sensitivity of fluorescein is exploited for the post-column detection of acids in
beverages and food products in Chapter 4. This method is expanded to include
the forensic detection of GHB in adultered beverages for alleged sexual assault
cases in Chapter 5. This fluorescence quenching work will introduce more
luminescence-based applications into the forensic science community and

improve current technologies for determination of explosives and drugs of abuse.

49

1.6

10.

11.

12.

13.

14.

15.
16.

References

Yinon, J.; Zitrin, S. Modern Methods and Applications in Analysis of
Explosives; Wiley: Chichester, UK, 1993.

Skoog, D. A.; Holler, F. J.; Nieman, T. A. Principles of Instrumental
Analysis, 5th ed.; Saunders College Publishing: Philadelphia, PA, 1998.

Lakowicz, J. Principles of Fluorescence Spectroscopy, Plenum Press:
New York, NY, 1983.

McGuffin, V. L.; Goodpaster, J. V. In Encyclopedia of Environmental
Analysis and Remediation; Meyers, R., Ed.; Wiley: New York, NY, 1998,
pp 3815-3831.

Goodpaster, J. V.; McGuffin, V. L. Anal. Chem. 2001, 73. 2004-2011.

Goodpaster, J. V.; Harrison, J. F.; McGuffin, V. L. J. Phys. Chem. A 2002,
106, 10645-10654.

Lee, H. C. In Forensic Science Handbook. Saferstein, R., Ed.; Prentice—
Hall, lnc.: Englewood Cliffs, NJ, 1982; Vol. 2.

Almog, J.; Levinton-Shamuilov. G.; Cohen, Y.; Azoury. M. J. Forensic Sci.
2007, 52, 330-334.

Seah, L. K.; Dinish. U. S.; Phang, W. F.; Chao, Z. X.; Murukeshan, V. M.
Forensic Sci. Int. 2005, 152, 249-257.

Sametband, M.; Shweky, l.; Banin, U.; Mandler, D.; Almog, J. Chem.
Comm. 2007, 1142-1144.

Alaoui, I. M.; Menzel, E. R.; Farag, M.; Cheng, K. H.; Murdock, R. H.
Forensic Sci. Int. 2005. 152, 215-219.

Castello, A.; Alvarez-Segui, M.; Verdu, F. Forensic Sci. Int. 2005, 155,
185-187. .

Kochana, J.; Wilamowski, J.; Parczewski, A. Chromatographia 2004, 60,
481-484.

Concheiro, M.; de Castro, A.; Quintela, O.; Lopez-Rivadulla, M.; sz, A.
Forensic Sci. Int. 2005, 150. 221-226.

Caslavska, J.; Thormann, W. Electrophoresis 2004, 25, 1623-1631 .

Nakamura, S.; Tomita. M.; Wada, M.; Chung, H.; Kuroda, N.; Nakashima,
K. Biomed. Chromatogr. 2006, 20, 622-627.

50

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.
28.
29.

30.

31.

32.
33.

Alnajjar, A.; Butcher, J. A.; McCord, B. Electrophoresis 2004, 25, 1592-
1 600.

Wada, M.; Nakamura, S.; Tomita. M.; Nakashima, M. N.; Nakashima, K.
Luminescence 2005, 20, 210-215.

Nakamoto, A.; Namera, A.; Nishida, M.; Yashiki, M.; Kuramoto, T.; Takei,
Y.; Furuno, M.; Minakuchi, H.; Nakanishi, K.; Kimura, K. Forensic Toxicol.
2006, 24, 75-79.

Tomita. M.; Nakashima, M. N.; Wada, M.; Nakashima, K. Biomed.
Chromatogr. 2006, 20, 1380-1385.

Ghazi-Khansari, M.; Zendehdel, R.; Pirali-Hamedani, M.; Amini, M. Clin.
Chim. Acta 2006, 364, 235-238.

Pulli, T.; Hoyhtya, M.; Soderlund, H.; Takkinen, K. Anal. Chem. 2005, 77,
2637-2642.

Kato, N.; Kojima, T.; Yoshiyagawa, S.; Ohta, H.; Toriba, A.; Nishimura, H.;
Hayakawa, K. J. Chromatogr. A 2007, 1145, 229-233.

Huhn, C.; Piltz, M.; Martin, N.; Dahlenburg, R.; Pyell, U. Electrophoresis
2005, 26, 2391 -2401.

Pietsch, J.; Schulz, K.; Kdrner, B.; Trauer, H.; DreBler, J.; Gey, M.
Chromatographia 2004, 60, 89-92.

Saito, K.; Toyo'oka, T.; Kato, M.; Fukushima, T.; Shirota, O.; Goda, Y.
Talanta 2005. 66, 562-568.

Yang, X.; Wang, X.; Zhang, X. Anal. Chim. Acta 2005, 549, 81-87.
Zhou, J.; Xu, X.; Wang, Y. J. Chromatogr. B 2007, 848, 226-231.

Matsuda, K.; Asakawa, N.; lwanaga, M.; Gohda, A.; Fukushima, S.; Ishii,
Y.; Yamada, H. Forensic Toxicol. 2006, 24, 41-47.

Payne, G.; Wallace, C.; Reedy, B.; Lennard, C.; Schuler, R.; Exline, D.;
Roux, C. Talanta 2005, 67. 334-344.

Yekkala, R.; Meers, C.; Van Schepdael, A.; Hoogmartens, J.; Lambrichts,
l.; Willems, G. Forensic Sci. Int. 2006, 159, $89-$94. '

Saitoh, M; Takeuchi, S. Forensic Sci. Int. 2006, 163, 38-50.

Bird, D. K.; Agg, K. M.; Barnett, N. W.; Smith, T. A. J. Microsc. 2007. 226,
18-25.

51

34.

35.

36.
37.

38.
39.
40.

41.

42.

43.

44.
45.

46.
47.

48.

49.
50.
51.

52.

Casamento, S.; Kwok, B.; Roux, C.; Dawson, M.; Doble, P. J. Forensic
Sci. 2003, 48, 1075-1083.

Gholamian, F.; Chaloosi, M.; Husain, S. W. Prop. Expl. Pyrotech. 2002,
27, 31-33.

Wallenborg, S. R.; Bailey, C. G. Anal. Chem. 2000, 72, 1872-1878.

Schulte-Ladbeck, R.; Vogel, M.; Karst, U. Anal. Bioanal. Chem. 2006, 386,
559-565.

Schulte—Ladbeck, R.; Kolla, P.; Karst, U. Analyst 2002, 127, 1152-1154.
Schulte-Ladbeck, R.; Kolla, P.; Karst, U. Anal. Chem. 2003, 75. 731-735.

Yinon, J.; Zitrin, S. The Analysis of Explosives; Pergamon: Elmsford, NY,
1 981 .

Oxley, J.; Smith, J.; Resende, E.; Pearce, E. J. Forensic Sci. 2003, 48,
742-753.

Bowerbank, C. R.; Smith, P. A.; Fetterolf, D. D.; Lee, M. L. J. Chromatogr.
A 2000, 902, 413-419.

Miller, C. J.; Glenn, D. F.; Hartenstein, S. D. Proc. SPIE—lnt. Soc. Opt.
Eng. 1998, 3575, 335-341.

Douse, J. M. F. J. Chromatogr. A 1987, 410, 181-189.

Batlle, R.; Carlsson, H.; Tollback. P.; Colmsjo, A.; Crescenzi, C. Anal.
Chem. 2003, 75, 31 37-3144.

Douse, J. M. F. J. Chromatogr. A 1988, 445, 244-250.

US Environmental Protection Agency 1994, Method 8330: Nitroaromatics
and Nitramines by High-Performance Liquid Chromatography (HPLC).

Gates, P.; Furlong, E.; Dorsey, T.; Burkhardt. M. Trends Anal. Chem.
1996, 15, 319-325.

Doali, J.; Juhasz, A. J. Chromatogr. Sci. 1974, 12, 51-56.
Spiegel, K.; Welsch, T. Fresenius J. Anal. Chem. 1997, 357, 333-337.

Reutter, D.; Buechele, R.; Rudolph, T. Anal. Chem. 1983, 55, 1468A-
1472A.

Bromberg, A.; Mathies, R. A. Anal. Chem. 2003, 75, 1188-1195.

52

53.

55.
56.

57.
58.

59.
60.

61.
62.
63.

65.

66.

67.

68.

69.

70.

71.

Bailey, C. G.; Wallenborg, S. R. Electrophoresis 2000, 21 , 3081-3087.

Kennedy, S.; Caddy, B.; Douse, J. M. F. J. Chromatogr. A 1996, 726, 211-
222.

Stevanovic, S.; Mitrovic, M. Int. J. Environ. Anal. Chem. 1990, 40, 69-76.

Pristera, F.; Halik, M.; Castelli, A.; Fredericks, W. Anal. Chem. 1960, 32,
495-508.

Conduit, C. P. J. Chem. Soc. 1959, 3273-3277.

Wilson, R.; Clavering, C.; Hutchinson, A. Anal. Chem. 2003, 75, 4244-
4249.

Bromberg, A.; Mathies, R. A. Electrophoresis 2004, 25, 1895-1900.

Charles, P. T.; Rangasammy, J. G.; Anderson, G. P.; Romanoski, T. C.;
Kusterbeck, A. W. Anal. Chim. Acta 2004, 525, 199-204.

Eiceman, G. A.; Stone, J. A. Anal. Chem. 2004, 76. 390A-397A.
Buryakov, l. A. J. Chromatogr. B 2004, 800, 75-82.
Jimenez, A. M.; Navas, M. J. J. Hazard. Mater. 2004, 106, 1-8.

Woltman, S. J.; Even, W. R.; Sahlin. E.; Weber, S. G. Anal. Chem. 2000,
72, 4928-4933.

Focsaneanu, K.-S.; Scaiano, J. C. Photochem. Photobiol. Sci. 2005. 4,
81 7-821 .

Albert, K. J.; Myrick, M. L.; Brown, S. B.; James. D. L.; Milanovich, F. P.;
Walt, D. R. Environ. Sci. Technol. 2001, 35, 3193-3200.

Kim, T. H.; Kim, H. J.; Kwak, C. G.; Park, W. H.; Lee. T. S. J. Polym. Sci,
Part A: Polym. Chem. 2006, 44, 2059-2068.

Kose, M. E.; Harruff, B. A.; Lin, Y.; Veca, L. M.; Lu, F.; Sun, Y. P. J. Phys.
Chem. B 2006, 110, 14032-14034.

Zhang, 8.; Lu, F.; Gao, L.; Ding, L.; Fang, Y. Langmuir 2007, 23, 1584—
1590.

Toal, S. J.; Sanchez, J. C.; Dugan, R. E.; Trogler, W. C. J. Forensic Sci.
2007, 52, 79-83.

Yang, J. S.; Swager, T. M. J. Am. Chem. Soc. 1998, 120, 11864-11873.

53

72.

73.

74.

75.
76.

77.

78.

79.

80.

81.

82.
83.
84.
85.
86.

87.
88.
89.

90.

Rose, A.; Zhu, Z.; Madigan, C. F.; Swager, T. M.; Bulovic, V. Nature 2005,
434, 876-879.

Saxena, A.; Fujiki, M.; Rai, R.; Kwak, G. Chem. Mater. 2005, 17, 2181-
2185.

Naddo, T.; Che, Y.; Zhang, W.; Balakrishnan, K.; Yang, X.; Yen, M.; Zhao,
J.; Moore, J. S.; Zang, L. J. Am. Chem. Soc. 2007, 129, 6978-6979.

Furton, K. G.; Myers, L. J. Talanta 2001, 54, 487-500.

Blair, H. M.; Poteet. W. M. Proc. SPlE-lnt. Soc. Opt. Eng. 2000, 4129,
494-502.

Guardala, N. A.; Farrell, J. P.; Dudnikov, V.; Merkel, G. AIP Conf. Proc.
2003, 680, 905-908.

Brondo, J.; Wielopolski, L.; Thieberger, P.; Alessi, J.; Vartsky, D.;
Sredniawski, J. AIP Conf. Proc. 2003, 680, 931-934.

Wielopolski, L.; Thieberger, P.; Alessi, J.; Brondo, J.; Vartsky, D.;
Sredniawski, J. AIP Conf. Proc. 2003, 680, 919-923.

Boudreaux, G. M.; Miller, T. S.; Kunefl<e, A. J.; Singh, J. P.; Yueh, F.-Y.;
Monts, D. L. Appl. Opt. 1999, 38, 1411-1417.

Wu, D.; Singh, J. P.; Yueh, F. Y.; Monts, D. L. Appl. Opt. 1996, 35, 3998-
4003.

Shu, J.; Bar, l.; Rosenwaks, S. Appl. Phys. B 2000, 70, 621-625.

Shu, J.; Bar, l.; Rosenwaks, S. Appl. Phys. B 2000, 71 , 665-672.

Shu, J.; Bar, l.; Rosenwaks, 8. Appl. Opt. 1999, 38, 4705-4710.
Arusi-Parpar, T.; Hefiinger, D.; Lavi, R. Appl. Opt. 2002, 40, 6677-6681.

Swayambunathan, V.; Singh, G.; Sausa, R. C. Appl. Opt. 1999, 38, 6447-
6454.

Fine, D. H.; Lieb, D.; Rufeh, F. J. Chromatogr. 1975, 107, 351-357.
Fine, D. H.; Rounbehler, D. P. J. Chromatogr. 1975. 109, 271-279.

Fine, D. H.; Rufeh, F.; Lieb, D.; Rounbehler, D. P. Anal. Chem. 1975, 47,
1188-1191.

Fine, D. H.; Wendel, G. J. Proc. SPlE-Int. Soc. Opt. Eng. 1993, 2092, 131-
136.

91.

92.

93.
94.
95.
96.
97.
98.

99.

100.

101.

102.
103.

104.

105.

106.

107.
108.

Fine, D. H.; Yu, W. C.; Goff, E. U.; Bender, E. C.; Reutter, D. J. J.
Forensic Sci. 1984, 29, 732-746.

Conrad, F. J.; Kenna, B. T.; Hannum, D. W. Nucl. Mat. Man. 1990, 19,
902-905.

Moore, D. 8. Rev. Sci. Instrum. 2004, 75, 2499-2512.
Lafleur, A. L.; Mills, K. M. Anal. Chem. 1981, 53, 1202-1205.
Douse, J. M. F. J. Chromatogr. A 1983, 256, 359-362.
Collins, D. A. J. Chromatogr. A 1989, 483, 379-383.

Douse, J. M. F. J. Chromatogr. A 1985, 328, 155-165.

Francis, E. S.; Wu, M.; Famsworth, P. B.; Lee, M. L. J. Microcol. Sep.
1995, 7, 23-28.

Crowson, C. A.; Cullum. H.; Hiley, R. W.; Lowe, A. J. Forensic Sci. 1996,
41 , 980-989.

Crowson, A.; Hiley, R. W.; lngham, T.; McCreedy, T.; Pilgrim, A. J.;
Townshend, A. Anal. Comm. 1997, 34, 213-216.

Francis, E. S.; Eatough, D. J.; Lee, M. L. J. Microcol. Sep. 1994, 6, 395-
401.

Lafleur, A. L.; Morriseau, B. D. Anal. Chem. 1980, 52, 1313-1318.

Selavka, C. M.; Tontarski Jr., R. E.; Strobel, R. A. J. Forensic Sci. 1987,
32, 941-952.

Nguyen, D. H.; Locquiao, S.; Huynh. P.; Zhong, Q.; He, W.; Christensen,
D.; Zhang, L.; Bilkhu, B. In Electronic Noses and Sensors for the Detection

of Explosives; Gardner, J. W., Yinon, J., Eds.; Kluwer Academic
Publishers: NonNell, MA, 2004, pp 71-80.

Lapat, A.; Székelyhidi, L.; Homyak, I. Biomed. Chromatogr. 1997, 11 . 102-
104.

Eastwood, D.; Fernandez, C.; Yoon, B. Y.; Sheaff, C. N.; Wai, C. M. Appl.
Spectrosc. 2006, 60, 958-963.

Sheaff, C. N.; Eastwood, D.; Wai, C. M. Appl. Spectrosc. 2007, 61, 68-73.
Bruno, J. G.; Comette, J. C. Microchem. J. 1997, 56, 305—314.

55

109.

110.

111.

112.

113.

114.
115.

116.
117.

118.

119.
120.

121.
122.
123.
124.
125.
126.
127.
128.

Mohammadzai, l. U.; Ashiuchi, T.; Tsukahara, S.; Okamoto, Y.; Fujiwara,
T. J. Chin. Chem. Soc. 2005, 52, 1037-1042.

Bock, C. R.; Connor, J. A.; Gutierrez, A. R.; Meyer. T. J.; Whitten, D. G.;
Sullivan, B. P.; Nagle, J. K. J. Am. Chem. Soc. 1979, 101, 4815-4824.

Bock, C. R.; Connor, J. A.; Gutierrez, A. R.; Meyer, T. J.; Whitten, D. G.;
Sullivan, B. P.; Nagle, J. K. Chem. Phys. Lett. 1979, 61, 522-525.

Bock, C. R.; Meyer, T. J.; Whitten, D. G. J. Am. Chem. Soc. 1975, 97,
2909-2911.

Glazier, 8.; Barron, J. A.; Morales, N.; Ruschak, A. M.; Houston, P. L.;
Abruna, H. D. Macromolecules 2003, 36, 1272-1278.

Rahman, M.; Harmon, H. J. Spectrochim. Acta A 2006, 65, 901-906.

Kang, J.; Ding, L.; Lu, F.; Zhang, S.; Fang, Y. J. Phys. D: Appl. Phys.
2006, 39, 5097-5102.

Jian, C.; Seitz, W. R. Anal. Chim. Acta 1990, 237. 265-271.

Wang, X.; Lee, S.-H.; Drew, 0.; Senecal, K. J.; Kumar, J.; Samuelson, L.
A. Polymer Preprints 2002, 43, 130-131 .

Kumar, V.; Dhawan, A.; Wang, X.; Parmar, V. S.; Samuelson, L. A.;
Kumar, J.; Cholli, A. Polymer Preprints 2003, 44, 1224-1225.

Patra, D.; Mishra, A. K. Sensor Actual. B - Chem 2001, 80, 278-282.

Wang. Y.; Wang, K.-M.; Shen, G.-l.; Yu, R.-Q. Talanta 1997, 44. 1319-
1327.

Hu, X.; Jiao, K.; Sun, W.; You, J.-Y. Electroanalysis 2006, 18, 613—620.

Tao, S.; Li, G.; Zhu, H. J. Mater. Chem. 2006, 16, 4521-4528.

Tao. S.; Li, G. Colloid Polym. Sci. 2007, 285, 721-728.

Tao, S.; Shi, 2.; Li, G.; Li, P. Angew. Chem. Int. Ed. 2006, 7. 1902-1905.
Lu, J.; Zhang, 2. Anal. Chim. Acta 1996. 318, 175-179.

Barnard, S. M.; Walt, D. R. Environ. Sci. Technol. 1991, 25, 1301-1304.
Barnard, S. M.; Walt, D. R. Nature 1991, 353, 338-340.

White, J.; Kauer, J. S.; Dickinson, T. A.; Walt, D. R. Anal. Chem. 1996, 68,
2191-2202.

56

129.

130.

131.

132.

133.
134.

135.

136.

137.
138.
139.
140.

141.

142.
143.
144.

145.

146.

Johnson, S. R.; Sutter, J. M.; Engelhardt, H. L.; Jurs, P. C.; White, J.;
Kauer, J. S.; Dickinson, T. A.; Walt, D. R. Anal. Chem. 1997, 69, 4641-
4648.

Albert, K. J.; Dickinson, T. A.; Walt, D. R. Proc. SPIE-lnt. Soc. Opt. Eng.
1998, 3392, 426-431.

Dickinson, T. A.; Michael, K. L.; Kauer, J. S.; Walt, D. R. Anal. Chem.
1999, 71. 2192-2198.

Albert, K. J.; Myrick, M. L.; Brown, S. B.; Milanovich, F. P.; Walt, D. R.
Proc. SPlE-Int. Soc. Opt. Eng. 1999, 3710, 308-314.

Albert, K. J.; Walt, D. R. Anal. Chem. 2000, 72, 1947-1955.

Michael, K. L.; Taylor, L. C.; Schultz, S. L.; Walt, D. R. Anal. Chem. 1998,
70, 1242-1248.

Stitzel, S. E.; Cowen, L. J.; Albert, K. J.; Walt, D. R. Anal. Chem. 2001, 73,
5266-5271.

Bakken, G. A.; Kauffman, G. W.; Jurs, P. C.; Albert, K. J.; Stitzel, S. S.
Sensor Actuat. B - Chem 2001, 79, 1-10.

Fisco, W. J. Forensic Sci. 1975, 20, 141-148.
Meaney, M. S.; McGuffin, V. L. J. Forensic Sci. 2007 (submitted).
Meaney, M. S., Michigan State University, East Lansing, MI, 2007.

Gerrnanenko, I. N.; Li, S.; EI-Shall, M. S. J. Phys. Chem. B 2001, 105, 59-
66.

Nieto, S.; Santana, A.; Hernandez, S. P.; Lareau, R.; Chamberlain, R. T.;
Castro. M. E. Proc. SPlE-Int. Soc. Opt. Eng. 2004, 5403, 256-260.

Toal, S. J.; Trogler, W. C. J. Mater. Chem. 2006, 16, 2871-2883.
Content, 8.; Trogler, W. C.; Sailor, M. J. Chem. Eur. J. 2000, 6, 2205-2213.

Sohn, H.; Calhoun, R. M.; Sailor, M. J.; Trogler, W. C. Angew. Chem. Int.
Ed. 2001, 40, 2104-2105.

Sohn, H.; Sailor, M. J.; Magde, D.; Trogler, W. C. J. Am. Chem. Soc.
2003, 125, 3821-3830.

Toal, S. J.; Magde, D.; Trogler, W. C. Chem. Comm. 2005, 5465-5467.

57

147.

148.
149.
150.
151.

152.

153.

154.

155.

156.

157.

158.

159.

160.

161.
162.
163.
164.

Saxena, A.; Rai, R.; Kim, S.-Y.; Fujiki, M.; Naito, M.; Okoshi, K.; Kwak, G.
J. Polym. Sci, Part A: Polym. Chem. 2006, 44, 5060-5075.

Zhou, Q.; Swager, T. M. J. Am. Chem. Soc. 1995, 117, 12593-12602.
Swager, T. M. Acc. Chem. Res. 1998, 31, 201-207.
Yang. J. S.; Swager, T. M. J. Am. Chem. Soc. 1998, 120, 5321-5322.

Williams. V. E.; Yang, J. S.; Lugmair, C. G.; Miao. Y. J.; Swager, T. M.
Proc. SPIE-lnt. Soc. Opt. Eng. 1999, 3710, 402-408.

Rose, A.; Lugmair, C. G.; Miao, Y. J.; Kim, J.; Levitsky, l. A.; Williams, V.;
Swager, T. M. Proc. SPlE-lnt. Soc. Opt. Eng. 2000, 4038, 512-518.

la Grone, M. J.; Cumming, C. J.; Fisher, M.; Reust, D.; Taylor, L. C. Proc.
SPlE-Int. Soc. Opt. Eng. 1999, 3710, 409-420.

la Grone, M. J.; Cumming, C. J.; Fisher. M.; Fox, M.; Jacob, S.; Reust, D.;
Rockley, M. G.; Towers, E. Proc. SPlE-Int. Soc. Opt. Eng. 2000, 4038,
553-562.

Cumming, C. J.; Aker, 0.; Fisher, M.; Fok, M.; la Grone, M. J.; Reust, D.;
Rockley, M. G.; Swager, T. M.; Towers, E.; Williams, V. IEEE T. Geosci.
Remote 2001 , 39, 1119-1128.

Swager, T. M.; Wosnick, J. H. MRS Bulletin 2002, 27, 446-450.

Fisher, M.; Sikes, J. Proc. SPIE-Int. Soc. Opt. Eng. 2003, 5089, 1078-
1087.

Cumming, C. J .; Fisher, M.; Sikes, J. In Electronic Noses & Sensors for
the Detection of Explosives; Gardner, J. W., Yinon, J., Eds.; Kluwer
Academic Publishers: Norvvell, MA, 2004, pp 53-69.

Henry. C. Chem. Eng. News 2005, 83(16), 11.

Thomas I", S. W.; Amara, J. P.; Bjork, R. E.; Swager. T. M. Chem. Comm.
2005. 4572-4574. '

Zhao, D.; Swager, T. M. Polymer Preprints 2005, 46, 1 194-1 195.
Zhao, D.; Swager, T. M. Macromolecules 2005, 38, 9377-9384.
Wang, L. Chem. Eng. News 2007, 85(15), 11.

Liu, Y.; Mills, R. C.; Boncella, J. M.; Schanze, K. S. Langmuir 2001, 17.
7452-7455.

58

165.

166.

167.

168.

169.

170.
171.
172.

173.

174.

175.

176.

177.

178.

179.

180.

Chang. C.-P.; Chao, C.-Y.; Huang, J. H.; Li, A.-K.; Hsu, C.-S.; Lin, M.-S.;
Hsieh, B. R.; Su, A.-C. Synth. Met. 2004, 144, 297-301.

Levitsky, l. A.; Euler, W. B.; Tokranova, N.; Rose, A. Appl. Phys. Lett.
2007, 90, 041904-041903.

Simonson, R. J.; Hance, B. G.; Schmitt, R. L.; Johnson, M. S.; Hargis Jr.,
P. J. Proc. SPlE-Int. Soc. Opt. Eng. 2001, 4394, 879-889.

Chen, L.; McBranch. D.; Wang, R.; Whitten, D. Chem. Phys. Lett. 2000,
330, 27-33.

Whelan, J. P.; Kusterbeck, A. W.; Wemhoff, G. A.; Bredehorst, R.; Ligler,
F. S. Anal. Chem. 1993, 65, 3561-3565.

Narang, U.; Gauger, P. R.; Ligler, F. S. Anal. Chem. 1997, 69, 1961-1964.
Charles, P. T.; Kusterbeck, A. W. Biosens. Bioelect. 1999, 14, 387-396.

Goldman, E. R.; Cohill, T. J.; Patterson, C. H.; Anderson, G. P.;
Kusterbeck, A. W.; Mauro, J. M. Environ. Sci. Technol. 2003, 37, 4733-
4736.

Judd, L. L.; Kusterbeck, A. W.; Charles, P. T.; Ligler. F. S.; Whelan, J. P.;
Foster, K. E. The Flow Immunosensor Used to Detect the Small Molecular
Weight Molecule TNT; Air and Waste Management Association:
Pittsburgh, PA, 1993.

Kusterbeck, A. W.; Gauger, P. R; Charles, P. T. Proc. SPIE-lnt. Soc. Opt.
Eng. 1997, 2937, 191-196.

Green, T. M.; Charles, P. T.; Anderson, G. P. Anal. Biochem. 2002, 310,
3641.

Goldman, E. R.; Pazirandeh, M. P.; Charles, P. T.; Balighian, E. D.;
Anderson, G. P. Anal. Chim. Acta 2002, 457, 13-19.

Kusterbeck, A. W.; Judd, L. L.; Yu, H.; Myles, J.; Liger, F. S. Proc. SPIE-
lnt. Soc. Opt. Eng. 1993, 2092. 218-226.

Charles, P. T.; Bart, J. C.; Judd, L. L.; Gauger, P. R.; Ligler, F. S.;
Kusterbeck, A. W. Proc. SPIE-Int. Soc. Opt. Eng. 1997, 3105, 80-87.

Bart, J. C.; Judd, L. L.; Kusterbeck, A. W. Sensor Actuat. B - Chem 1997,
39, 411-418.

Bart, J. C.; Judd, L. L.; Hoffman, K. E.; Wilkins, A. M.; Kusterbeck, A. W.
Environ. Sci. Technol. 1997, 31 , 1505-1511.

59

181.

182.
183.

184.

185.

186.

187.

188.

189.

190.

191.

192.

193.

194.

195.

196.

197.

Shriver-Lake, L.; Charles, P.; Kusterbeck, A. Anal. Bioanal. Chem. 2003,
377, 550-555.

Narang, U.; Gauger, P. R.; Ligler, F. S. Anal. Chem. 1997, 69, 2779-2785.

Judd. L. L.; Kusterbeck, A. W.; Conrad, D. W.; Yu, H.; Jr., H. L. M.; Ligler,
F. S. Proc. SPlE-lnt. Soc. Opt. Eng. 1995, 2388, 198-204.

Narang, U.; Gauger, P. R.; Kusterbeck, A. W.; Ligler, F. S. Anal. Biochem.
1998, 255, 13-19.

Holt, D. B.; Gauger, P. R.; Kusterbeck, A. W.; Ligler, F. S. Biosens.
Bioelect. 2002, 1 7, 95-103.

Shriver-Lake, L. C.; Breslin, K. A.; Charles, P. T.; Conrad, D. W.; Golden,
J. P.; Ligler, F. S. Anal. Chem. 1995, 67, 2431-2435.

Bakaltcheva, I. B.; Shriver-Lake, L. C.; Ligler, F. S. Sensors and Actuators
B: Chemical 1998, 51, 46-51.

Bakaltcheva, l. B.; Ligler, F. 8.; Patterson, C. H.; Shriver-Lake, L. C. Anal.
Chim. Acta 1999, 399, 13-20.

Shriver-Lake, L. C.; Donner, B. L.; Ligler, F. S. Environ. Sci. Technol.
1997, 31, 837-841.

Shriver-Lake, L. C.; Patterson, C. H.; van Bergen, S. K. Field Anal. Chem.
Tech. 2000. 4, 239-245.

Shriver-Lake, L. C.; Breslin, K. A.; Golden, J. P.; Judd, L. L.; Choi, J.;
Ligler, F. S. Proc. SPlE-Int. Soc. Opt. Eng. 1994, 2367, 52-58.

van Bergen, S. K.; Bakaltcheva, l. B.; Lundgren, J. S.; Shriver-Lake, L. C.
Environ. Sci. Technol. 2000, 34, 704-708.

Kusterbeck, A. W.; Charles, P. T. Field Anal. Chem. Tech. 1998, 2, 341-
350.

Charles, P. T.; Gauger, P. R.; Patterson, C. H.; Kusterbeck, A. W. Environ.
Sci. Technol. 2000, 34, 4641-4650.

Charles, P. T.; Dingle, B. M.; Van Bergen, S.; Gauger, P. R.; Patterson Jr.,
C. H.; Kusterbeck, A. W. Field Anal. Chem. Tech. 2001, 5, 272-280.

Gauger, P. R.; Holt. D. B.; Patterson Jr., C. H.; Charles, P. T.; Shriver-
Lake, L.; Kusterbeck, A. W. J. Hazard. Mater. 2001, 83, 51-63.

Lesnik, B. Environ. Test. Anal. 1999, 8, 8.

60

198.

199.

200.

201.

202.

203.

204.
205.
206.

Medintz, I. L.; Goldman, E. R.; Lassman, M. E.; Hayhurst, A.; Kusterbeck,
A. W.; Deschamps, J. R. Anal. Chem. 2005, 77, 365-372.

Medintz, l. L.; Goldman, E. R.; Clapp, A. R.; Uyeda, H. T.; Lassman, M.
E.; Hayhurst, A.; Mattoussi, H. Proc. SPlE-lnt. Soc. Opt. Eng. 2005, 5705,
166-174.

Tran, P. T.; Goldman, E. R.; Anderson, G. P.; Mauro, J. M.; Mattoussi, H.
Proc. SPIE-lnt. Soc. Opt. Eng. 2002, 4636, 23-30.

Goldman, E. R.; Medintz, I. L.; Whitley, J. L.; Hayhurst, A.; Clapp, A. R.;
Uyeda, H. T.; Deschamps, J. R.; Lassman, M. E.; Mattoussi, H. J. Am.
Chem. Soc. 2005, 127, 6744-6751.

Altamirano, M.; Garcia-Villada, L.; Agrelo, M.; Sanchez-Martin, L.; Martin-
Otero, L.; Flores-Moya, A.; Rico, M.; Lopez-Rodas, V.; Costas. E. Biosens.
Bioelect. 2004, 19, 1319-1323.

Radhika, V.; Proikas-Cezanne, T.; Jayaraman, M.; Onesime, 0.; Ha, J. H.;
Dhanasekaran, D. N. Nat. Chem. Biol. 2007, 3, 325-330.

Everts, S. Chem. Eng. News 2007, 85(20). 10.
Andrew, T. L.; Swager, T. M. J. Am. Chem. Soc. 2007, 129, 7254-7255.
Science & Technology Concentrates. Chem. Eng. News 2007, 85(23), 32.

61

CHAPTER 2
SOLUTION-PHASE INVESTIGATIONS OF COMMON FLUOROPHORES
FOR THE DETECTION OF NITRATED EXPLOSIVES

Previous studies indicated that nitrated explosives may be detected by
fluorescence quenching of pyrene and related compounds. The use of pyrene,
however, invokes numerous health and waste disposal hazards. In this chapter,
ten safer fluorophores are identified for quenching detection of target nitrated
compounds. Initially, Stern-Volmer constants are measured for each fluorophore
with nitrobenzene and 4-nitrotoluene to determine the sensitivity of the quenching
interaction. For fluorophores with quenching constants greater than 50 M",
sensitivity and selectivity are investigated further using an extended set of target
quenchers. Nitromethane, nitrobenzene, 4-nitrotoluene, and 2.6-dinitrotoluene
are chosen to represent nitrated explosives and their degradation products;
aniline, benzoic acid, and phenol are chosen to represent potential interfering
compounds. Among the fluorophores investigated, purpurin. malachite green,
and phenol red demonstrate the greatest sensitivity and selectivity for nitrated
compounds. Correlation of the quenching rate constants for these fluorophores
to Rehm-Weller theory suggests an electron-transfer quenching mechanism. As
a result of the large quenching constants, purpurin, malachite green, and phenol
red are the most promising for future detection of nitrated explosives via

fluorescence quenching.

62

2.1 Introduction and Background
2.1.1 Analysis of Nitrated Explosives

The analysis of nitrated explosives is of great importance in a variety of
fields.1 In forensic science, as discussed in Chapter 1, identification of these
explosives and their degradation products can be used to identify persons in
recent contact with an explosive device.2 In addition. in environmental chemistry,
explosives present in undetonated landmines can leach into and persist in the
soil and groundwater, posing a threat to the environment and any living
inhabitants?"4 For industrial quality control, manufacturers of explosive materials
must assure consumers that their products are safe, effective, and free from
contamination by monitoring the composition throughout the manufacturing
process. Because the requirements in each of these fields are so diverse, a

versatile method is needed for the analysis of nitrated explosives.
2.1.2 Solution-Phase Fluorescence Quenching

As discussed briefly in Chapter 1, the fluorescence quenching of pyrene
and related compounds by nitrated explosives and degradation products has
been explored.5’8 Goodpaster and McGuffin report that nitroaromatic compounds
are more efficient quenchers than their aliphatic counterparts, with quenching
constants up to three times greater.7 In addition, measured quenching constants
increase with increasing nitration of the molecules. These experimental results
support the ab initio calculations of Goodpaster and coworkers, which suggest
that the fluorescence quenching mechanism involves the formation of an excited-

state ion pair.7 Upon collision of excited-state pyrene and ground-state

63

nitromethane, an electron is transferred from pyrene to nitromethane. As the
molecules separate, the electron is transferred back to pyrene and the excitation
energy is dissipated by vibrational relaxation of nitromethane. A similar
mechanism can be postulated for the nitrated explosives.7 Aliphatic explosives
may accept the transferred electron with similar efficiency to nitromethane.
However, aromatic explosives have a large number of resonance structures that
can stabilize the resulting ion and enhance the efficiency of electron transfer.
Because of the higher efficiency, Stem-Volmer quenching constants are
expected to be greater for aromatic than aliphatic explosives, as confirmed
experimentally.7 Increasing the number of electron-withdrawing nitro groups
further enhances the ability of the aromatic explosive to accept an electron from
pyrene. Consequently, the Stern-Volmer constants are expected to be greater
with increasing degree of nitration, as confirmed experimentally.7 In direct
comparison to UV absorbance detection, the fluorescence quenching method of
Goodpaster and McGuffin demonstrates a 40-fold improvement in signal-to-noise
ratio.7 Although pyrene offers excellent sensitivity and selectivity for nitrated
explosives, the hazards associated with pyrene discourage its routine use.

In this study, several alternative fluorophores (Figure 2.1) have been
selected based on the proposed quenching mechanism and other criteria
essential to their incorporation into a new or existing field-ready device. The
selected fluorophores have lower toxicity than pyrene to reduce health and waste
disposal concerns that may arise from their routine use. These molecules are

selected for testing because they have conjugated, electron-rich structures with

.3: 58852". é: co oseoooem ”é so. scone ”E code 25%:
gt 63.2 3233.2 x8 3ch ocfito< it 5595 x$ c:mE3oo_>£oE-¢.ocE9355-“ x8 atmano>xo£oE
-n-_>£oEoEo._m-v xuv ctmEzoo>xoLu>I-v ”E ocoin. .cozmmzmgs Loan: 88390:: Co 6.202% .—..N 239”.

o O o o: Bazofozz o
/ O/
2

zxfofo

   

3
I0 .: ofofo

 

O a
“2.10”:on 2.3110319
\z
2. o :0
6.311912 /m «3.9 o:
N3..on z 2:319 m
\
o o :o
/

.36

e \
\ Q on \
o 0 com:
o o 2.93...on

65

significant electron-donating capabilities to facilitate ion-pair formation during the
quenching process.7'9 Another relevant selection criterion is a high quantum
efficiency, which allows high sensitivity to be achieved with a low fluorophore
concentration. The excitation and emission wavelength of the fluorophore are
also important when considering potential field applications. Because handheld
UV lamps (254 and 366 nm) are readily available in the forensic community,
fluorophores with excitation wavelengths in the UV region are beneficial. In a
field device, detection may be achieved visually or through the use of a digital
camera or similar imaging device. In the latter approach, where glass optics may
be involved, fluorophores with emission wavelengths in the visible region are of
greatest utility. Finally, good physical, chemical, and photolytic stability of the
fluorophore are desirable features that will maximize durability and longevity of a
field-ready device.

The goal of this study is to evaluate the selected fluorophores with
regard to their potential utility for on-site explosives detection. Stem-Volmer
constants for each fluorophore are measured with target nitrated compounds as
well as potential interfering compounds to determine sensitivity and selectivity.
Of the target quenchers, nitromethane and nitrobenzene are models of aliphatic
and aromatic nitrated explosives, respectively, while 4-nitrotoluene and 2,6-
dinitrotoluene are models of degradation products. Behavior of specific
explosives and their degradation products is not investigated in this preliminary
work, but will likely be consistent with the chosen models. Of the interfering

quenchers, aniline is representative of bases, while benzoic acid and phenol are

66

representative of acids. These functional groups and aromatic molecules, in
general, are found ubiquitously in the environment. Through the Stem-Volmer
constants, the sensitivity and selectivity of the selected fluorophores for the
compounds of interest are evaluated and compared directly to those for pyrene.
Promising fluorophores identified from this study may be useful in a variety of

field-ready devices, including sensors, wipes, and air sampling tubes.
2.2 Experimental Methods
2.2.1 Reagents

Reagent-grade fluorophores are used as received and include 4-
hydroxycoumarin (Aldrich). 4-bromomethyl-7-methoxycoumarin (Aldrich), 7-
diethylamino-4-methylcoumarin (Aldrich), acridine orange (Eastman Kodak),
fluorescein (Eastman Kodak), phenol red (Aldrich), purpurin (National Aniline
Division), malachite green (Sigma), methylene blue (Coleman and Bell), pyrene
(MCB), and rhodamine 6G (MCB). Target quenchers, selected to model
nitroaromatic explosives and their degradation products, are used as received
and include nitromethane (Aldrich), nitrobenzene (Mallinckrodt), 4-nitrotoluene
(Aldrich), and 2.6-dinitrotoluene (Aldrich). Interfering quenchers are used as
received and include aniline (Fisher). benzoic acid (Spectrum), and phenol

(Mallinckrodt).
2.2.2 Determination of Quenching Constants

Fluorescence power measurements are performed by using a flow-

injection laser-induced fluorescence spectrometer. The sample is introduced by

67

a helium gas-displacement pump (30 psi) into a 50-pm i.d. fused-silica capillary
flow cell. The excitation source, a continuous-wave He-Cd laser (Melles Griot,
model 3074-40M, 325 nm, 38 mW), is filtered and focused onto a section of the
capillary where the polyimide coating has been removed to allow UV
transmission. The fluorescence emission is collimated, filtered to remove
scattered laser light, and refocused onto the entrance slit of a 0.34-m Czerny-
Tumer monochromator (Instruments SA, model 340E, 300 groove/mm grating, 2-
nm band-pass). The emission is collected by a thermoelectrically-cooled charge-
coupled device (CCD) detector (Instruments SA, model (A)TECCD-2000x800—7)
operated at 240 K. A commercially available electronic interface (Instruments
SA, models Datascan 2 and CCD 2000) and associated software (Instruments
SA, Spectramax for Windows, version 3.1) are used for instrument control and
data acquisition.

Stock solutions of the fluorophores (10'3 M) and quenchers (10'2 M) are
prepared in high-purity spectroscopic-grade methanol (Honeywell Burdick 8
Jackson). For preliminary studies, dilute solutions of each fluorophore are
prepared at concentrations ranging from 10'3 to 10"5 M. Fluorescence spectra
are measured and the peak heights at the maximum emission wavelength (Table
2.1) are used to determine the linearity of the fluorescence response. Linearity is
necessary to ensure accuracy of the measured quenching constant, as the
addition of quencher is equivalent to a reduction in the active fluorophore
concentration. For quenching studies, stock solutions are combined and diluted

in methanol to prepare working solutions of constant fluorophore and varied

68

Table 2.1. Spectral properties of fluorophores.

 

 

Km“ (EM) Relative
Fluorophore (nm)‘I Fluorescence Power”
Pyrene (1) 397.2 0.345
4-Hydroxycoumarin (2) 381.0 0.00151
4-BromomethyI-7-methoxycoumarin (3) 408.1 0.206
7-Diethylamino-4-methylcoumarin (4) 448.4 0.764
Purpurin (5) 412.6 0.0802
Acn'dine Orange (6) 556.0 0.0757
Methylene Blue (7) 716.9 0.136
Malachite Green (8) 379.1 0.0225
Phenol Red (9) 368.6 0.00840
Rhodamine 6G (10) 585.8 1.00
Fluorescein (1 1) 559.1 0.0283

 

3 Maximum emission wavelengths (kmax (Em) are given in methanol with

excitation at 325 nm.

b Relative fluorescence power is normalized with respect to concentration,
integration time of. the CCD detector, and to the fluorescence of Rhodamine 6G.

69

quencher concentrations. Fluorophore concentrations are selected to obtain
approximately equal fluorescence power for each fluorophore while maintaining
linearity, and are generally on the order of 10" to 10'5 M. Quencher
concentrations are varied over the range of 10'2 to 10" M.

For each solution, five replicates of the fluorescence spectrum are
acquired over a 220-nm range. From the average peak height at the maximum
emission wavelength, the fluorescence power is determined for the fluorophore
solution in the absence of quencher (F0) and in the presence of quencher (F).
The fluorescence power ratio (Po/F) is then graphed against the quencher
concentration to generate a Stem-Volmer plot, as described in Equation 1.4.
Using a spreadsheet program (Microsoft Excel, version 2003), the slope (Stem-
Volmer constant) and intercept are determined by linear regression. The Stem-
Volmer constant for each fluorophore with a given quencher is compared to that
for pyrene, the reference fluorophore. This ratio of Stem-Volmer constants
allows a direct comparison of the relative sensitivity of the fluorophore with
respect to that of pyrene. The Stem-Volmer constant for each fluorophore with a
given quencher is also compared to that for nitrobenzene, the reference
quencher. This ratio of Stem-Volmer constants allows a direct comparison of the

selectivity of the fluorophore for the target and interfering compounds.
2.2.3 Determination of Fluorescence Lifetimes

Fluorescence lifetimes are determined in collaboration with Dr. Gary
Blanchard of Michigan State University by using a time-correlated single-photon-

counting spectrometer.10 In this system, the second harmonic of a continuous-

70

wave mode-locked Nd:YAG laser (Coherent, model Antares 76-S) is used to
excite a cavity-dumped, synchronously pumped dye laser (Coherent, model 702-
2, 325 nm, 1 mW). Fluorescence emission is collected and focused on a
subtractive double monochromator of Czemy—Tumer design (American
Holographic, model DB10-S, 10-nm bandpass). The emission is detected with a
cooled two-stage microchannel plate photomultiplier (Hamamatsu, model
R3809U-51) operated at 273 K. Single-photon counting is performed with
commercially available electronic instrumentation (Tennelec, models TC454,
TC864, TC412A, TC525, and PCA—II) and associated software (National
Instruments, Labview, version 7.1). Six replicate measurements of the
fluorescence time decay are acquired by single-photon counting. The data are fit
by nonlinear regression to a single-exponential function using a peak-fitting
program (Table Curve 2D, Jandel Scientific, version 2.02). Because the
fluorescence lifetimes range from 1 to 15 ns, deconvolution of the instrument
response function (~35 ps full width at half-maximum) from the experimental data
is not necessary. A more detailed description of the data analysis is given in the

Appendix.
2.2.4 Determination of Redox Potentials

To investigate the quenching mechanism, oxidation (on) and reduction
(Ered) potentials are determined for each fluorophore and quencher by using
cyclic voltammetry in collaboration with Dr. Elizabeth McGaw of Michigan State
University. Solutions are prepared containing 10'3 M fluorophore or quencher

and 0.1 M NaN03 as the supporting electrolyte in methanol. All solutions are

71

degassed by nitrogen purge for 5 minutes prior to measurement. The
electrochemical measurements are performed using a glassy carbon working
electrode (0.07 cm2 area), a Pt counter electrode, and a nonaqueous “no leak”
AglAgCl reference electrode (Cypress Systems, model EE009, 3 M KCI fill
solution). A potentiostat (CH Instruments, model 650A) is used to scan the
potential from 0.0 to +1.5 V at 0.5 V/s for determination of oxidation potentials
and from 0.0 to -1.5 V at 0.5 V/s for determination of reduction potentials. For
reversible systems, on and Ered are calculated as the average of the anodic and
cathodic peak potentials. For irreversible systems, Eox is reported as the anodic
peak potential and Ered is reported as the cathodic peak potential. A more
detailed description of the data analysis is given in the Appendix.
2.2.5 Determination of Singlet Excitation Energies

Also needed to investigate the quenching mechanism are the singlet
excitation energies (Eon) for each fluorophore. UV-visible absorbance and
fluorescence emission spectra are acquired by using commercially available
spectrophotometers (Hitachi, models U-4001 and F-4500, respectively).
Because the absorbance and emission spectra for each fluorophore are mirror
images, the energy of the singlet excited state is determined from the wavelength
at which the normalized absorption and emission spectra intersect.9 A more

detailed description of the data analysis is given in the Appendix.
2.3 Results and Discussion
2.3.1 Fluorophore Characterization

Table 2.1 summarizes the emission wavelengths and relative

72

fluorescence powers of the fluorophores as determined on the flow-injection
laser-induced fluorescence spectrometer. All fluorophores have excitation
wavelengths in the ultraviolet region, which is desirable for the development of a
field detection device. These fluorophores demonstrate a wide range of
maximum emission wavelengths, from phenol red at 369 nm to methylene blue at
717 nm. Most fluorophores have sufficient emission in the visible region to be
useful in a field detection device. The relative fluorescence power with 325-nm
excitation is also variable, with rhodamine 6G having the highest fluorescence
and with phenol red and 4-hydroxycoumarin having the lowest, approximately

100- and 1000-fold less than that of rhodamine 6G, respectively.
2.3.2 Quenching of Pyrene by Nitrated Explosives

Because the quenching of pyrene by nitrated compounds is well
documented,7 pyrene is used as a reference fluorophore throughout these
studies. Quenching constants determined for pyrene with the target and
interfering quenchers, as well as correlation data for Stem-Volmer plots, are
given in Table 2.2. An example spectral data set and Stem-Volmer plot are
shown in Figure 2.2. As noted previously by Goodpaster and McGuffin,7 the
Stem-Volmer constant for the target aliphatic quencher, nitromethane, is less
than those for its aromatic counterparts, such as nitrobenzene (Table 2.2).
Moreover, the addition of an electron-donating methyl group slightly decreases
the quenching constant for 4-nitrotoluene compared to nitrobenzene.
Conversely, the addition of an electron-withdrawing nitro group significantly

increases the quenching constant for 2,6-dinitrotoluene compared to

73

....m.w..oo .m..._o>-c.o.w 93.39. m E9. no.m_:o_mo on 3:58 o:_m> .o_nm__m>m .o: .1. <2 o
.ocoNcmno.._.. o. o>_.m_o. 093328 9m wo_._>_.oo_om .onocoac o
.Nm .coBEooo co_.m_o..oo ..._>> .vé .8353 E2. no:_....o.oo s

 

 

 

o<Z momd 5.0 H mad m H m- .0..ch
8.0 on... «8... a 8.... om n em 29.. 285.. Whumuwuh
mmd chad mod H mad m H V5 9.55.. .
mvé vmmd Nd H md cm H omv mcw3_o.o.._:_0-©.m
vmd 0mm.o 00.0 H vmd m H NmN mama—9.0.321? Mums—095:0
Do... mmmd 5.0 H voé m H m Fm mcmNcmnoEZ «090...
mmd mmmd mod H #0.? n H nor 0.559.532

ag>=o£ow «m 33.2... Areal.” 3305.5

3305.5

 

s.o.3.m.og...o. Eoo. .m _o..m...oE
5 9620.530 octoto..._ ccm Ema. 5.2. E meet... .0. 3.: 9.8.2.8 9.32.9.0 .N.~ min...

74

 

 
  
 
    

 

 

 

 

 

60
— o M
35 5o — 5x10'4 M
g ----- 2x10-3 M
n. 40 - — 3x10'3 M
3 4x10-3 M
“21 3o _ 5X10'3 M
U
(0
Lu
1:
O
D
_I
IL
370 390 410 430 450 470
WAVELENGTH (nm)
3.5
3.0 -
2.5 -
LI.
1
ll.
2.0 -
1.5 —
1.0 . .

 

 

 

0.000 0.002 0.004 0.006
2.6-DNT CONCENTRATION (M)

Figure 2.2. Fluorescence spectra of 5x10'5 M pyrene in methanol at room
temperature in the presence of increasing concentrations of 2.6-dinitrotoluene

(top) and the resulting Stern-Volmer plot (bottom). The Stern-Volmer constant
for this data set is 450 M'1 with a correlation coefficient of 0.9988.

75

4-nitrotoluene.

The Stem-Volmer constants determined for the interfering aromatic
compounds are significantly less than those for the target aromatic compounds.
Benzoic acid and phenol have quenching constants that are statistically
indistinguishable from zero. The correlation coefficients for these regressions
deviate significantly from unity as a direct result of the small quenching constant
and, therefore, there is greater uncertainty in the data. The quenching constant
for aniline, however, is comparable to that measured for nitromethane.
Consequently, the presence of amines in the environment could cause
interference in the detection of nitromethane or other aliphatic nitrated
explosives. At higher concentrations, such amines may even cause interference
in the detection of nitroaromatic explosives. However, the ability to detect
selected amines may be beneficial, as the degradation of nitrated explosives can
result in reduction of the nitro groups to amine groups (e.g., 2,4,6-trinitrotoluene
to 2-amino-4,6-dinitrotoluene or 4-amino-2,6-dinitrotoluene).

The selectivity of pyrene for each quencher is calculated with respect to
nitrobenzene, as summarized in Table 2.2. Ideally, the selectivity for nitrated
compounds is near or greater than unity and the selectivity for interfering
compounds is much less than unity. All of the target quenchers demonstrate a
greater selectivity with pyrene than all of the interfering quenchers. The
selectivities for the nitroaromatic compounds are near or greater than unity,
although the value for nitromethane is significantly less and more similar to that

for aniline.

76

.Nm .:0_0E0o0 8000.50 :._2. 6.. 80005 E0: 00:_:..0.0n_ s

 

 

 

ems... No... a 8.. m h mm .00... me... a .o. o. a em :0 58098.“.
08... 8... n S... 0... u .0. 08... 8... h we... . u .0 3.. 0.... 8.582.”.
as... so... a .o. m. a 8. .8... 8... n .o.. N a e. a. .80 .89.:
80... 08.0 a .00.. o n mm 08.0 no... a 3... . a 0.. .0. 590 2.50.0.2
000... No... a No. .0 a 0.0 80... .0... a 8.. .s. a . E on... 22.3.22
30... .o... n 8... . s 0.. 000... No... a .o. . a on .0. 8.0.0 8.2.0....
000... .0... a .o. m a 9.. 08... 8... a a... o a m... .0. ages:
.00... .0... u 3... . u 0.. 30... me... a 00... m s 0... .3 5.0...soo.....oe-...oe_eo_..£o_o-.
.00... 8... n 8.. v u 0.. 8...... 8... a 8.. N a an E e.0...soo..xo...oE-.-_.....oEoEo.m-..
0.0... No... a 8.. 0.0 a 0.0 80.. me... n 0.... o... n .1. a. eeoEsoo..xo.o.f-..
08... 8... u .0... m a 30 80... .00 n 8.. 0 a 0.0 . 0 ocean.

Na 2.8.2... .1... a. a”. .0023... 5...... J. 29.8.8...

 

 

77

3020.05.24 2.05.0005:

 

e0.3.0.095.
E00. .0 6:050... :_ 0:0:_0.o.._:-v 0:0 0:0N:0..0...: 5.2. 00.9.0900: .0. A.1. 9:00:00 0:30:25 .m.~ 030..

2.3.3 Preliminary Investigation of Selected Fluorophores

Quenching constants are measured for the selected fluorophores with
nitrobenzene and 4-nitrotoluene (Table 2.3). The magnitude of the Stem-Volmer
constant is representative of the sensitivity of the quenching interaction, and the
values can be directly compared for a given quencher. The quenching constants
for all of the selected fluorophores with nitrobenzene and 4-nitrotoluene are
significantly less than those for pyrene. Purpurin demonstrates the greatest
quenching constant, with a value half that of pyrene. In addition, malachite green
and phenol red exhibit satisfactory quenching, with measured Stem-Volmer
constants of nearly 50 M'1 or greater. The other fluorophores show very little
sensitivity for nitrobenzene or 4-nitrotoluene, as illustrated by the small
quenching constants for 4-hydroxycoumarin, methylene blue, rhodamine 6G, and
fluorescein. Based on the values of the quenching constants, only purpurin,

malachite green, and phenol red are chosen for further investigation.
2.3.4 Quenching of Purpurin by Nitrated Explosives

The Stern-Volmer constants and rate constants measured for purpurin
with the extended set of target and interfering quenchers are shown in Table 2.4.
An example spectral data set and Stem-Volmer plot are shown in Figure 2.3. As
with pyrene, the quenching constant for the target aliphatic compound is less
than those for the target aromatic compounds. The quenching constants for
nitrobenzene, 4-nitrotoluene, and 2,6-dinitrotoluene are all approximately the
same value, 140 M", within experimental error. This similarity indicates that, in

the presence of a nitro group, the addition of other functional groups to the

78

....0.0:o0 .0..._o>-:.0.w 02.000: 0 E9. 00.0.30.00 0.. .0::00 03.0> 0.00.00 .0: u <2

 

 

 

0:0...0 0.02.0.0. 00.0.3200 0.0 0922.00.00 0.0..00.o3_n. M
.0:0~:0..0...: 0.02.0.0. 00.0.30.00 0.0 0022.00.00 .0:0:03O 9
.mm HCQOEQOO warm—Otoo 53> .Vé cozmscm E0: UOEE..0~00 0
<2 8... 08... No... a 8.. . h o .89.:
0.... m .... 000... 0...... a 5.... . h .0 00.. 2850 “.2830
5.0.5....
<2 .<2 80... .o... a 0.... . n m- 2.0..
mm... .0... 80... 8... a 0.... o. a om. «53.22.0500
.4... we... 80... .o... a .0. m a e... 88.92.24 «35530
0.... 0.... 80... we... a .0... 0 h 0... «5282.2 .83.
8... 0...... 08... .o... n 0.... . n 8 22.9.22
3.2.83... n2.2.8.8 «m 08.9... 2.2.1.4 35530

05:00.03... .0..0..03O

 

«0.3.0.0080. E00.
.0 .0:0...0... :. 0.020530 m:..0..0.:. 0:0 .090. 2...... Am. 5.30.30 .0. 3.. 3:00:00 0:30:03. Jaw 0.00...

79

 

FLUORESCENCE POWER

 

 

 

 

350 400 450 500 550
WAVELENGTH (nm)

 

1.6~
u.
E1.4~

1.2-

 

 

1.0 I .
0.000 0.002 0.004 0.006

2,6-DNT CONCENTRATION (M)

Figure 2.3. Fluorescence spectra of 5x10“1 M purpurin in methanol at room
temperature in the presence of increasing concentrations of 2,6-dinitrotoluene
(top) and the resulting Stem-Volmer plot (bottom). The Stern-Volmer constant
for this data set is 129 M’1 with a correlation coefficient of 0.9907.

 

80

aromatic ring, whether electron-donating or -withdrawing, does little to alter the
quenching of purpurin. These quenching constants for purpurin are
approximately three-fold less than those for pyrene (Table 2.2).

The Stem-Volmer constants for the interfering compounds are
significantly less than those for the target compounds. Aniline and phenol have
quenching constants near zero, while benzoic acid has a quenching constant that
is small but statistically indistinguishable from zero at the 95% confidence level.
Interestingly, the quenching constant for aniline with purpurin is significantly less
than that with pyrene and, hence, has less potential for interference.

The selectivity of purpurin for each quencher is also calculated with
respect to nitrobenzene, as summarized in Table 2.4. All of the target quenchers
demonstrate a greater selectivity with purpurin than all of the interfering
quenchers, with values near or greater than unity. In contrast to pyrene, purpurin
exhibits a selectivity for nitromethane that is more similar to the target quenchers
than the interfering quenchers. This suggests that purpurin may be more

effective for detection of nitrated explosives than pyrene.
2.3.5 Quenching of Malachite Green by Nitrated Explosives

The Stem-Volmer constants and rate constants measured for malachite
green with the extended set of target and interfering quenchers are shown in
Table 2.5. An example spectral data set and Stern-Volmer plot are shown in
Figure 2.4. As with pyrene, the quenching constant for the target aliphatic
compound is less than those of the target aromatic compounds. The quenching

constants increase only slightly upon the addition of electron-donating and

81

.8050... 01.0. 0:.E0_0..0...0_00 0.80 .2 00.0 5.; 00.003 000200
22.0.00 00.0..00::

0:0...0 0. 02.0.0. 00.0.0200 0.0 0022.00.00 90:00.02".
.0:0N:000.._: 0. 02.0.0. 00.0.0200 0.0 0022.00.00 .0:0:00O
.0”. .:0.0...000 5:20.80 :22. .0.. 80003. 0.0.. 00:.....0.0o

..:0.0:00 .0:._0>-:.0.w 02.000: 0 80.. 00.0.0200 00 .0::00 00_0> .0.00_.0>0 .0: u <2 .

0
0

0

 

 

 

<2 00.0 00.0 00.0 H 00.0 0 a 0 .8020
m0... m.20 m.200 m.000 a 3.. w. a 00
.000 20.0 .0000 v.0 « 0.. .00 u. 0.. 0_0< 0.00.00 Muumuwuh
<2 <2 .0000 m00.0 « .0.. mm a 00- .
<2 .<2 .0000 D000 0 00.. u. a 0.- 0...__..<
3.0 00.. 000.0 .00 u. .0.. 0 a .0 0000.92.20-00
0.0 0... 000.0 000.0 a .00.. 0 a 00 0000.99.20 0.0 2.00
m0.0 .00.. m.0000 m.000 . 0... 00 a 0.. 040.0. 0
.020 .00.. .0000 .000 u. .00 u. a 0.. 0:0N..000.._z
00.0 0.0 .000 .00 a .0.. 0 u 00 000500.022
0.0.2.0200 a.....>..00_00 .0 .0028... .02. ulx .05500
03:00.00... .0202000

0.0.0.0.00E0. Eoo. .0 .0:0:.0E
:_ 0.0:0:000 m:..0..0.:. 0:0 .090. :22. .0. :00... 02:00.2. .0. 3.. 0.:0.0:00 m:.:0:000 .m.~ 0.00..

82

 

FLUORESCENCE POWER

 

 

 

 

 

 

 

 

 

340 390 440 490 540
WAVELENGTH (nm)
1.4
1.3 0
u.
E 1.2 0
o
1.1 -
1.0 l I
0.000 0.002 0.004 0.006

2,6-DNT CONCENTRATION (M)

Figure 2.4. Fluorescence spectra of 2x10“1 M malachite green in methanol at
room temperature in the presence of increasing concentrations of 2,6-
dinitrotoluene (top) and the resulting Stem-Volmer plot (bottom). The Stem-
Volmer constant for this data set is 56 M'1 with a correlation coefficient of 0.9958.

83

-withdrawing groups. These quenching constants for malachite green are
approximately an order of magnitude less than those for pyrene (Table 2.2).

The Stem-Volmer constants for the interfering compounds with
malachite green, however, are less straight-forward. The quenching constant for
phenol is zero, for aniline is less than zero, and for benzoic acid is substantially
greater than any of the target quenchers. During preparation of the malachite
green solutions, the color is visibly changed upon addition of both aniline and
benzoic acid, and becomes more intense as the concentration of quencher
increases. The color change and unexpected quenching behavior is believed to
be a pH effect, resulting from the addition of the acidic and basic quenchers.
Malachite green is known to change color with pH,11 and there is some evidence
for a decrease in quantum yield in acidic solution.12 This decrease in
fluorescence power may be considered as a form of trivial quenching, which can
be minimized by proper buffering of the solutions.

To minimize the pH change, malachite green solutions are buffered
using acetic acid and triethylamine (2:1) in methanol. The quenching constants
are measured for benzoic acid and aniline, and for nitrobenzene as a control
(Table 2.5). The quenching constants for nitrobenzene and aniline are
statistically equivalent in the buffered and unbuffered solutions, whereas the
quenching constant for benzoic acid decreases significantly upon buffering. In
fact, the quenching constant for benzoic acid is statistically indistinguishable from
zero at the 95% confidence level. Clearly, these results indicate that the large

quenching constant observed for benzoic acid in unbuffered solution is the

consequence of trivial and not dynamic quenching.

The negative quenching constant determined for malachite green and
aniline, however, is not the consequence of trivial quenching, as the value does
not change upon buffering the solution. Based on this result, the interaction
between malachite green and aniline is not related to a simple proton transfer.
The negative value may arise from the formation of a ground-state complex
between malachite green and aniline that has a greater absorbance or quantum
yield than malachite green alone. The resulting increase in fluorescence power
relative to a solution in the absence of aniline yields a negative Stem-Volmer
constant.

The selectivity of malachite green for each quencher is also calculated
with respect to nitrobenzene, as summarized in Table 2.5. Selectivities for the
target compounds, including nitromethane, are all near or greater than unity, and
are greater than those measured for the interfering quenchers. The selectivity for
benzoic acid in unbuffered solution, however, is three times greater that for the
target quenchers, indicating a large potential for interference. When the
fluorophore solution is buffered, the selectivity for benzoic acid decreases
dramatically, to a value less than 0.5. This clearly indicates the importance of

buffering malachite green solutions to avoid interference from acidic species.
2.3.6 Quenching of Phenol Red by Nitrated Explosives

The Stern-Volmer constants and rate constants measured for phenol red
with the extended set of target and interfering quenchers are shown in Table 2.6.

An example spectral data set and Stem-Volmer plot are shown in Figure 2.5. As

85

..00.0000 .00..o>-0.0.w 02.0000 0 0.0... 00.0.0200 00 .00000 00_0> .0_00_.0>0 .00 n <2 .
.8050... 0. .00. 00.50.200.000 0..000 s. 000 .0.; 00.0.00 00.0.00

02.0.00 00.00000:

000.3 0.02.0.0. 00.0.0200 0.0 0022.00.00 0.000902“.

0000000820 0.02.0.0. 00.0.0200 0.0 00.22.0200 00000:...

.01 0.0.0.0000 02.0.0000 0...... .0.. 02.00am. E00 002000.00

U

0
a

 

 

 

<2 <2 0000 .00 0 00.. 0 0 0- .000...
00.0 0.0 0.00 .00 0 .0.. 0 0 .. 0.0< 0.00000 0.0000000
<2 <2 .0000 .000 0 00.. .0 0 0- 05.002...
<2 .<2 .0000 .000 0 .00 .00 0 0.0- 00._.c<
.00 00.. .000 000.0 0 0.0.. 0 n. 00 0000.00.05.00
.00 .0.. .000 .00 0 .0.. 0. 0 00. 0000.00.22-.. 0.0200000
.000 .00“. .0000 .000 0 000 .0 0 00 .00....
.00 0 .00 . .000 0 .00 0 0 .0 . .0 0 0. 0000000002
000 00.0 .000 .00 0 00.. 0 0 0.. 000500.002
".........00.00 «22.00.00 .0 0.00.0.0. .0.... .0. 0000000

0.2.0960... .0000000

 

«0.0.0.0080. E00. .0
6000.00. 0. 0.0000000 02.2.0.0. 000 .00.0. 0...... .0. 00. .00000 .0. 3.. 0.00.0000 00.000000 0.0 030...

86

35

 

 
  
 
   
  

 

 

 

 

 

 

 

— 0 M
E 30 ‘ mm3 M
-3
o 25 _ ----- 2x10 M
D- — 3x10'3 M
8 20 - --- 4x10'3 M
E ..... 5x10'3 M
0 15 -
(D
‘0’
0 10 ~
3
r.’ 5 -
O I I “A
335 385 435 485 535
WAVELENGTH (nm)
1.6
1.4 -
IL
1
II.
1.2 -
1.0 T ,
0.000 0.002 0.004 0.006

2,6-DNT CONCENTRATION (M)

Figure 2.5. Fluorescence spectra of 6x104 M phenol red in methanol at room
temperature in the presence of increasing concentrations of 2.6-dinitrotoluene
(top) and the resulting Stem-Volmer plot (bottom). The Stem-Volmer constant
for this data set is 98 M'1 with a correlation coefficient of 0.9979.

87

with pyrene, the quenching constant for the target aliphatic compound is less
than those of the target aromatic compounds. The quenching constants increase
only slightly upon the addition of an electron—donating methyl group but remain
constant for an additional electron-withdrawing nitro group. These quenching
constants for phenol red are three- to four-fold less than those for pyrene (Table
2.2).

Consistent with the trends of the other fluorophores, the Stem-Volmer
constants for the interfering compounds are significantly less than those for the
target compounds. In fact, these quenching constants with phenol red are
statistically indistinguishable from zero at the 95% confidence level. A negative
quenching constant is measured for phenol red and aniline, similar to that
observed for malachite green. Phenol red (pKa = 7.9) is a common pH indicator,
so the addition of aniline increases solution pH and causes a color change.
However, solutions of phenol red and aniline buffered with acetic acid and
triethylamine (2:1) give statistically equivalent quenching constants to those
measured in unbuffered solutions (Table 2.6). Hence, the negative quenching
constant is not the result of trivial quenching and may arise from the formation of
a ground-state complex between the fluorophore and quencher, as with
malachite green.

The selectivity of phenol red for each quencher is calculated with respect
to nitrobenzene, as summarized in Table 2.6. All of the target quenchers
demonstrate a larger selectivity with phenol red than all of the interfering

quenchers. The selectivities for the nitroaromatic compounds are all near or

88

greater than unity. The value for nitromethane is significantly less, but is still
greater than that for the interfering quenchers. Although selectivities could not
be calculated for aniline and phenol because the quenching constants were less
than zero, there is little potential interference from these or related compounds in

the detection of nitrated explosives.
2.3.7 Proposed Quenching Mechanism

As discussed previously, fluorescence quenching of pyrene is believed
to occur by electron transfer from the excited-state fluorophore to the ground-
state quencher.13 This mechanism can be examined in more detail by using the
empirical model of Rehm and Weller.14 The free energy change for an outer-
sphere electron-transfer process (A60) is given by

AGet = on _ Ered _ E0,0 _" C (2-1)

where on and Ered are the ground-state oxidation and reduction potentials of the
electron donor and acceptor, respectively, E00 is the energy of the singlet
excited-state fiuorophore, and C is a Coulombic term relating the energy of the
separated ions. In polar solvents such as methanol, the Coulombic term of

Equation 2.1 is less than 2% of AG... and can be neglected in the determination of

1360.15 The rate constant for fluorescence quenching. k0. is related to the free
energy for electron transfer by

k0

kd = 1+ 0.25[exp(AG;, /RT) + exp(AG6t /RT):|

 

(2.2)

where k0 represents the diffusion-limited rate constant (1.4 x 1010 M'1 s'1 in

methanol),16 R is the gas constant, T is the absolute temperature, and AG“. is

89

the activation energy for the electron-transfer process, which is given as a

monotonic function of AG...14

2 )é
AG;t=[[A:°‘] +(AGet(O))2:| +5239: (2.3)

 

ln Equation 2.3, AGet(O) is the activation free energy when AG... = 0 and has been
determined experimentally to be 0.104 eV.“ In general, agreement between
calculated and experimental values of kd within a factor of two is considered

evidence in favor of an outer-sphere electron-transfer quenching mechanism.
2.3.7.1 Fluorescence Lifetimes

To compare Stem-Volmer constants with Rehm-Weller theory,
calculation of quenching rate constants is necessary. Fluorescence lifetimes are
determined in methanol for pyrene, purpurin, malachite green, and phenol red, at
wavelengths listed in Table 2.1. Fluorescence lifetimes are determined from
nonlinear regression to a single-exponential function, and are given in Table 2.7.
Typical correlation coefficients (R2) for the nonlinear regression ranged from
0.9887 to 0.9975 (see Appendix). The Stem-Volmer constants (Tables 2.2 and
2.4-2.6), together with the fluorescence lifetimes, are used to calculate second-
order rate constants according to Equation 1.4, and are given in Table 2.8. The
rate constants for pyrene, purpurin, malachite green, and phenol with all
quenchers are 109 to 1010 M" s", approaching the diffusion-limited value of 1.4 x

1010 M" s'1 in methanol.16

90

Table 2.7. Fluorescence lifetimes (no) determined for each of the fluorophores at
the wavelengths listed in Table 2.1 (n = 6). All values determined in methanol at
room temperature.

 

 

Fluorophore 0° (ns)
Pyrene 15.32 :l: 0.06
Purpurin 3.93 d: 0.05
Malachite Green 1.050 :1: 0.007
Phenol Red 1.99 :l: 0.02

 

91

.6559: 5 cs 35552.58 38¢ .2 one 5;, .8.th 8:28

 

 

 

.8328 88:32:: W
.Efimcoo .mE_o>-E$m m>=mmoc m E0: 85:28 on “cacao m:_m> .o_om__m>m Ho: n <2 2
cote co cozmmmaoa 3 85.3.8 359.58: 5.3 vé cozmacm So: 3585qu a
<2 md H od «.0 H Nd n_<z _0cm;n_
No H Nd
. . o . . . . 220530
c o o m
mOHoo omHoF NOHmo FOHro E< _~:m actors...
<2 <2
<2 <2 <2 mod H 33 oc_._c<
to H 3. md H 2m md H On to H mN mcmaouozEBéd
m.o H o.m od H to mod H 85 mod H em... mcmaoaougzé «.8 use:
.3 H 3 fig H 5. Mask 0
u_..o H ed 16 H m6 Nd H ed mod H mod 83:80.22
2o H Wm md H Wm 3.0 H 9N mod H nod ocméoEoEZ
tom 3:25 .820 3202.22 5593.. 2.23 350.330

 

kwmvs. so 5 fiacfimcoo ouamdczocasa

 

.EBEmaEE Eco. um .0559: E 88390:: or: ho zoom .2 36.8.8 3: $23ch 29 9:20:30 .mN 03a...

92

Table 2.9. Electrochemical data for fluorophores and quenchers in methanol at
room temperature.

 

 

 

on i!) End 1!)

Pyrene 0.41 0%

Flu oroph ore Purpurin 0.89 -0.99
Phenol Red 1.22 -0.35

Malachite Green 0.75 -0.51

Nitromethane NA3 -1 .03

Nitrobenzene 1 .23 -0.50

4-Nitrotoluene 1 .24 -0.53

Quencher 2,6—Dinitrotoluene 1.20 -o.37
Aniline 0.41 -1.04

Benzoic Acid 0.41 -1.01

Phenol 0.63 -1 .03

 

8 NA = not available. No oxidation potential is observed for nitromethane.

93

2.3.7.2 Redox Potentials

Oxidation and reduction potentials for each fluorophore and quencher
are also needed for comparison of quenching data to Rehm-Weller theory
(Equation 2.1). Redox potentials for each fluorophore and quencher are
determined in methanol from cyclic voltammograms (see Appendix) and are
summarized in Table 2.9. Based on the redox potentials, general trends can be
hypothesized about the electron-donating or -accepting ability of the fluorophores
and quenchers. Pyrene and phenol red are more likely to be reduced, or accept
an electron. Conversely, purpurin and malachite green are more likely to be
oxidized, or donate an electron. Of the quenchers, no oxidation potential is
measurable for nitromethane over the range investigated, indicating that
nitromethane is more likely to be reduced. The target quenchers are all more
likely to be reduced, with reduction potentials that are closer to zero. The
opposite is true for the interfering quenchers, in that all are more likely to be

oxidized.
2.3.7.3 Singlet Excitation Energies

Singlet excitation energies for each fluorophore are also needed for
comparison of quenching data to Rehm-Weller theory (Equation 2.1). Singlet
excitation energies (Eon) are determined in methanol from absorbance and
fluorescence spectra (see Appendix) and are summarized in Table 2.10. Phenol
red has the largest singlet excitation energy (3.70 eV), while purpurin has the
smallest (3.22 eV). Because the emission of all of the fluorophores occurs within

the same spectral region, the range of singlet excitation energies is narrow, less

94

Table 2.10. Spectroscopic data for fluorophores in methanol at room
temperature.

 

 

Fluorophore Eoo (e!)
Pyrene 3.44
Purpurin 3.22
Phenol Red 3.70
Malachite Green 3.63

 

95

than 0.5 eV.
2.3.7.4 Comparison to Rehm-Weller Theory

In order to examine the quenching mechanism for each fluorophore,
measured rate constants are compared with Rehm-Weller theory. The relevant
electrochemical (on, Ered) and spectroscopic (Egg) parameters for the
fluorophores and quenchers are used to evaluate Equation 2.1 for each
fluorophore-quencher pair. The AGet values calculated for each fluorophore-
quencher pair are compared in Table 2.11 for each of the possible electron-
transfer mechanisms. Mechanism I involves the transfer of an electron from the
excited-state fluorophore (F*) to the ground-state quencher (Q), denoted as F* +
Q —> (F+Q‘)*. Mechanism ll involves the transfer of an electron from the ground-
state quencher to the excited-state fluorophore, denoted as F* + Q —> (F’Q+)*.
Of the two, the mechanism that yields the more negative value of AGet is more
thermodynamically favorable.

The AGet values for pyrene with each of the target quenchers (Table
2.11) suggest that the most probable quenching mechanism is via transfer of an
electron from excited-state pyrene to the ground-state quencher (Mechanism I).
This mechanism is consistent with that previously demonstrated by ab initio
calculations with nitromethane.13 This mechanism is also consistent with the
trends in electron-donating and -accepting ability of the quencher molecules.
During ion-pair formation with pyrene, the aromatic structure of nitrobenzene
allows greater stabilization of the incurred negative charge than nitromethane, as

reflected in their reduction potentials (Table 2.9). The result is a greater rate

96

0052.000 00 82:00 00:_0> so< om .0:0£0E0:.: .0: 0020000 0. .0::060 5:00.08 02 .0_00__0>0 .0: u <2 m

 

 

 

97

.0..- 8... Q... me. .- 8..- om..- mm..- 80- 6:9:
.0... .0. .- 2.....- .4. .- «0... mm. .- .3.- 80- o_o< 0.88m.
.3- 5..- 3H. 3...- 3..- 8... SN- 8..- 2.22
mom- .3- EN- EN- no..- 8..- mm..- 8....- 0co=_o.o:_c_00.~
mom- and- ...N. mm..- 9.0- 8..- E... and- 05222.24
3. .- unw- N..~- 8..- 8..- 8..- mm..- «.3- 08229.2
<2 8..- <2 mu..- <2 2....- m<2 80- 22059.2
.. _ __ _ __ _ __ _ 3282.0
200.0 02:00.02 00:... .0202... 5.595. 2.2%...

 

 

 

 

 

.EmEmco0E 0.00.02: >..00_E0:>00E._0£ 0.0E 0.: 00.00.05 00>. 000:0.0m
.2. E25505: 05:00.02: 0:90-00:00 0. 0202030 0.06-0:006 E0: .0:mcm:-c0:00_0 0:0 A. Em_c0_..00s_v
050:0:0 0:90-02:06 0. 0.050200: 0.06-00:00.0 E0: A>0 80$ .0:mc0:-c0:00_0 .0: 3.000 00.“. ._._..~ 030...

constant for quenching by nitrobenzene compared to nitromethane (Table 2.2).
Similarly, the electron-donating methyl group of nitrotoluene increases the
electron density of the aromatic ring compared to nitrobenzene. The increase in
electron density results in less stabilization of the negatively charged ion, as
reflected in the reduction potential, and a smaller rate constant. Finally, the
additional nitro group of 2.6-dinitrotoluene significantly decreases the electron
density on the aromatic ring, resulting in a substantially greater rate constant
compared to 4-nitrotoluene. Because the electron-withdrawing capacity of the
nitro group is much greater than the electron-donating capacity of the methyl
group, the effect on the magnitude of the rate constant is not equal and opposite.
Each of these trends is consistent with the measured Stem-Volmer quenching
constants, rate constants, and the predicted AGet values for pyrene with the
target quenchers.

The calculated AGO, values (Table 2.11) indicate that Mechanism I is
most likely for purpurin and malachite green with the target quenchers. It is
noteworthy that the mechanism is similar to that for pyrene, despite the presence
of functional groups on these fluorophores that may alter the electronic structure
and interactions with the quenchers. In the case of phenol red, however, the
alternative mechanism involving electron-transfer from the ground-state
quencher to the excited-state fluorophore (Mechanism II) is more likely for the
target quenchers. Regardless of the direction of electron transfer, all of the
quenching rate constants are approximately equal to the diffusion limit of 1.4 x

1010 M'1 s'1 in methanol.16 It is interesting that Mechanism II is also most

98

probable for all of the fluorophores with the interfering quenchers (Table 2.11).
The values of AGe. are relatively large and comparable to those for the target
quenchers. However, the rate constants are substantially less than those for the
target quenchers (Table 2.8).

Quenching rate constants for each fluorophore-quencher pair are
compared to the AGe- values predicted by Rehm-Weller theory in Figure 2.6. As
a result of their shorter fluorescence lifetimes, the rate constants for all
fluorophores with the target quenchers are comparable. In general, the
magnitude of the rate constants is directly related to the magnitude of the AG,-
values according to Rehm-Weller theory. For the target quenchers, many of the
measured rate constants are within a factor of two, and all are within a factor of
five, of the predicted values.

In contrast, the rate constants for the fluorophores with the interfering
quenchers are notably less than the values predicted by Rehm-Weller theory. A
possible explanation for this discrepancy is that Rehm-Weller theory assumes
the kinetic behavior is simply related to the thermodynamic behavior (i.e., AG*--- is
a monotonic function of AGet, Equation 2.3). This may not be true if steric or
statistical considerations are important in the collisional interaction between the
fluorophore and quencher. Another possibility is that Rehm-Weller theory
assumes the redox potential of the excited-state fluorophore is simply related to
that of the ground state (i.e., E*ox = B», — Eop or E*-ed = Ered + EQO). This may not
be true if the excited and ground states differ substantially in electron distribution

and density. Finally, it is possible that the model of simple outer-sphere electron

99

 

._x
O

 

 

 

 

'- A
z A I
on

:5 A.Q I.
‘2’ 1~
8.: '3 3 O
m.“ o
g 201 ~ 0

“c
g 2
:I:
0 0.01 4
z
I-IJ
:
O

0.001 -

-3 -2 -1 0

FREE ENERGY FOR ELECTRON TRANSFER (eV)

Figure 2.6. Comparison of experimental quenching rate constants for pyrene
(O), purpurin (0), malachite green (A), and phenol red (:1) with target quenchers
(filled) and interfering quenchers (unfilled) in methanol with Rehm-Weller theory

H-

100

transfer does not accurately describe the interaction between the fluorophore and

the interfering quenchers.
2.4 Conclusions

Of the extended group of fluorophores investigated in this study, pyrene,
purpun'n, malachite green, and phenol red are shown to have sensitive and
selective quenching interactions with a target group of nitrated quenchers. The
remaining fluorophores under investigation demonstrate little or no response to
nitrated compounds. Because each fluorophore may interact with the quenchers
through a different mechanism, various degrees of sensitivity and selectivity are
observed. Of the fluorophores studied, pyrene demonstrates the greatest
sensitivity for nitrated molecules, but its routine use presents health and waste
disposal hazards. Purpurin, malachite green, and phenol red demonstrate less
sensitivity but comparable or greater selectivity for nitrated molecules. Both
malachite green and phenol red exhibit pH-dependent fluorescence, but only
malachite green shows quenching behavior that is a function of pH. With
buffering, the interference of acidic species on the quenching of malachite green
fluorescence can be reduced. In general, pH-sensitive fluorophores are not ideal
for application in the detection of explosives on-site, as various matrices may be
encountered that could lead to erroneous results. Thus, among these potential
fluorophores, purpurin shows the most promise for detection of nitrated
explosives by fluorescence quenching.

Based on the results reported herein, purpurin, malachite green, or

phenol red are suitable for a variety of applications. For example, they can be

101

used in the laboratory detection of nitrated explosives separated by using liquid
chromatography. Following LC, the fluorophore can be added post-column and
quenching monitored with a standard fluorescence detector.7 In addition, field-
ready devices can be developed by incorporation of these fluorophores into
optical fiber sensors, wipes, air and water sampling tubes, etc. The use of hand-
held UV lamps and digital cameras is appropriate for these applications. The
detection limits of the explosives must be determined and optimized for each of
these applications.

In this work, only a limited number of potential fluorophores have been
selected based on their structural similarities to pyrene. From these results,
however, it is evident that mechanistic similarities do not follow directly from
structural similarities. Consequently, a more broad and diverse selection of
fluorophores may prove essential for the development of a fluorescence

quenching method for the detection. of nitrated explosives.

102

2.5

10.

11.
12.
13.

14.
15.
16.

References

Yinon, J.; Zitrin, 8. Modern Methods and Applications in Analysis of
Explosives; Wiley: Chichester, UK, 1993.

Casamento, 8.; Kwok, B.; Roux, C.; Dawson, M.; Doble, P. J. Forensic
Sci. 2003, 48, 1075-1083.

Gholamian, F.; Chaloosi, M.; Husain, S. W. Prop. Expl. Pyrotech. 2002,
27, 31-33.

Wallenborg, S. R.; Bailey, C. G. Anal. Chem. 2000, 72, 1872-1878.
Jian, C.; Seitz, W. R. Anal. Chim. Acta 1990, 237, 265-271.

Zhang, 8.; Lu, F.; Gao, L.; Ding, L.; Fang, Y. Langmuir 2007, 23, 1584-
1590.

Goodpaster, J. V.; McGuffin, V. L. Anal. Chem. 2001, 73, 2004-2011.

Focsaneanu, K.-S.; Scaiano, J. C. Photochem. Photobiol. Sci. 2005, 4,
81 7-821 .

Goodpaster, J. V.; McGuffin, V. L. Anal. Chem. 2000, 72, 1072-1077.

DeWitt, L.; Blanchard, G. J.; LeGoff, E.; Benz, M. E.; Liao, J. H.;
Kanatzidis, M. G. J. Am. Chem. Soc. 1993, 115, 12158-12164.

Chemyak, V.; Reisfeld, R. Chem. Phys. Lett. 1991, 181, 39—44.
Hu, X.; Jiao, K.; Sun, W.; You, J.-Y. Electroanalysis 2006, 18, 613-620.

Goodpaster, J. V.; Harrison, J. F.; McGuffin, V. L. J. Phys. Chem. A
2002, 106, 10645-10654.

Rehm, D.; Weller, A. lsr. J. Chem. 1970, 8, 259-&.
Tachiya, M; Murata, S. J. Phys. Chem. 1992, 96, 8441-8444.

Zanini, G. P.; Montejano, H. A.; Previtali, C. M. J. Chem. Soc., Faraday
Trans. 1995, 91, 1197—1202.

103

CHAPTER 3
INVESTIGATION OF POTENTIAL FIELD-READY DEVICES
FOR THE DETECTION OF NITRATED EXPLOSIVES
VIA FLUORESCENCE QUENCHING

Field-ready devices for explosives detection based on fluorescence
quenching are currently being used in Iraq, and the improvement of such devices
will inevitably increase the security of soldiers and other law enforcement agents.
Because solid-state devices are most desirable for such applications, this
chapter examines immobilization of the fluorophores identified in the solution-
phase fluorescence quenching studies described in Chapter 2. The goal of
fluorophore incorporation is to determine the viability of solid-state devices such
as wipes and air sampling tubes for explosives detection. To identify the most
appropriate matrix polarity, fluorescence of the dyes is first compared in a variety
of solvents, including water, methanol, dioxane, and hexane. The most
promising fluorescence is observed for malachite green and phenol red in
dioxane, and for purpurin in methanol. Fluorescence of immobilized dyes is
compared in matrices including filter paper and poly(ethylene glycol) of various
molecular weights. For malachite green, incorporation into a poly(ethylene
glycol) matrix yields the greatest fluorescence. Poly(ethylene glycol) containing
malachite green is coated on filter paper, glass slides, and silica particles and the
resulting devices compared with regard to their fluorescence and fluorescence
quenching behavior. Ideally, stable fluorescence with UV excitation and visible
emission would be observed in a matrix that could be easily incorporated into

field devices such as wipes or air sampling tubes.

104

3.1 Experimental Methods

The experiments related to the incorporation of fluorophores into solid-
state devices are conducted with the assistance of Ms. Heidi Bonta, an

undergraduate research assistant.
3.1.1 Reagents

Reagent—grade fluorophores are used as received and include purpurin
(Fisher Scientific), phenol red (Aldrich), and malachite green (Sigma). Stock
solutions of each fluorophore (10'3 to 10“ M) are prepared in dioxane (Sigma-
Aldrich), hexane (Mallinckrodt), methanol (Sigma-Aldrich), methylene chloride
(Mallinckrodt), and distilled, deionized water (Corning Glass Works, model MP-
3A). Stock solutions are diluted to give working solutions of concentrations
varying from 10'3 to 10'7 M. Poly(ethylene glycol) (Sigma-Aldrich) is obtained in
molecular weights of 750, 1000, 1500, and 2000 g/mol. Malachite green is
added into the melted polymer or into a solution of the polymer in dioxane or

methylene chloride.
3.1.2 Fluorescence Measurements

Fluorescence of dye-containing substrates is visually observed by using
a 366-hm handheld UV lamp (Mineralight Lamp, Multiband UV-254/366nm,
Model UVGL-58) in the absence of ambient light. A conventional fluorimeter
(Hitachi F-4500) is utilized for measurement of excitation and emission spectra.
Quartz cuvettes with a 1-cm path length (Stama Cells) are used for solution-

phase measurements, while a reflectance attachment is utilized for solid-phase

105

measurements (Hitachi Solid Sample Holder #650-0161). To use this reflectance
attachment, the substrates must be attached to a support such as a microscope
slide. Initially, fluorescence spectra are collected by using an excitation
wavelength of 325 nm based on previous results (Chapter 2). Emission
wavelengths are scanned at 240 nm/min from 330 - 500 nm. Excitation and
emission wavelengths are optimized through iterative comparison of maxima in
the emission and excitation spectra, respectively. Entrance and exit slit widths of
5 mm and a photomultiplier tube voltage of 950 V are employed for most
measurements. Background fluorescence of each solvent or substrate is also
measured for comparison. Analysis of the resulting data is completed using a
spreadsheet program (Microsoft Excel, version 2003).

For quenching studies, solutions are prepared by combining stock
solutions of malachite green and quencher in dioxane. The concentration of
malachite green in each solution is 2x10'4 M, and the concentrations of
quenchers nitromethane and nitrobenzene vary from 2x10'3 - 1x10'2 M. The
fluorescence emission of these solutions is measured by using the spectroscopic
system described previously (Section 2.2.2). Stem-Volmer constants for
malachite green in dioxane are compared to those determined previously in

methanol (Section 2.3.5).
3.1.3 Preparation of Solid Substrates

To verify the fluorescence in the solid state, solid malachite green is
mixed with potassium bromide (Spectrum, IR grade), ground, pressed into a

pellet, and analyzed in the fluorimeter by using the reflectance attachment (vide

106

infra). Fluorophores are also examined on substrates including filter paper
(Whatman), fiberglass filter paper (Hurlbut Paper Co.), Kimwipe sheets
(Kimberly-Clark), and Grab-it dust cloths (S.C. Johnson & Son, Inc.). The
fluorescence of the substrates before and after methanol pretreatment is
compared. Pretreatment includes soaking in 3 mL of methanol and drying at 100
°C in an oven for 10 minutes and is designed to remove adsorbed compounds
from the substrate. Fluorophores are adsorbed to paper-based substrates by
soaking in 10 mL of fluorophore solution in an open Petri dish. The Petri dishes
are placed in a fume hood the solvent is allowed to evaporate. Other materials
including blue cotton cloth, unbleached paper towel, unbleached cardboard, and
a sterile bandage pad are also investigated as substrates for malachite green.
Fluorophore solution is applied dropwise onto the untreated substrate. Dry
substrates are cut to fit onto a support (vide infra) and the fluorescence of the
mounted substrates is measured by using the reflectance attachment.

Various techniques for mounting of substrates are evaluated, including
different supports and adhesives. A glass microscope slide (Fisher Scientific)
and black cardboard are investigated as potential support materials. Double-
sided tape (3M Scotch Brand Tape), green laboratory tape (Shamrock), and
white laboratory tape (Shamrock) are investigated as potential adhesives.
Double-sided tape is applied between the substrate and support, whereas single-
sided tape is applied on the front edges of the substrate and also with the
substrate folded around the support and secured on the opposite side. Tapes

and supports are compared to minimize the background fluorescence and

107

potential interference. For polymer substrates, mixtures of malachite green and
poly(ethylene glycol) are poured onto the support. If solvent is used to dissolve
the polymer and malachite green, the solvent is allowed to evaporate prior to
fluorescence measurement.

Poly(ethylene glycol) is also supported on silica particles (Johns-Manville
Products, Chromosorb, hexamethyldisilazane (HMDS) treated, 30/60 mesh).
Particles are coated with a solution of poly(ethylene glycol) and 5x104 M
malachite green in methylene chloride. The mixture is placed in a rotary
evaporator and the solvent is allowed to evaporate slowly under vacuum.
Adjusting the ratio of polymer mass to particle mass from 10 — 30% allows
adjustment of the film thickness. The coated particles are packed into a quartz
pyrolysis tube (CDS Analytical, Inc., part 10A1-3008) and secured on both ends
with quartz wool (Alltech). Fluorescence of the coated particles is measured by

using the spectroscopic system described in Chapter 2.
3.2 Results and Discussion
3.2.1 Solution-Phase Fluorescence and Quenching

For solution-phase measurements, solvent polarity is known to directly
affect fluorescence emission.1 Determination of optimum solvent polarity for
each fluorophore provides insight into the most appropriate matrix for a solid-
state device. For example, if the greatest fluorescence is observed in aqueous
solution, polar matrices are most favorable, and a substrate such as cellulose
may be ideal. The fluorescence of purpurin, malachite green, and phenol red are

measured in water, methanol, dioxane, and hexane, as shown in Figures 3.1-3.3.

108

.Ec mum E0..0..0xm. .0.:.0.00E0.
E00. .0 T-v 020x02 0:0 All. 020.8... 2.... .0:0£0E Al. .903 2. 5.20.30 5. so. .0 0:00am 080000.02”. é.» 0.59...

.E... 2.020.032.
com. 0.... 84 m? 84 0... 8m mmm

 

..:.l- I. - o
r oom H
.s. n
... o
- 8.. w...
._.. - 80 N
.... a
. - com d
. O
- 82 a

 

 

 

CON .-

109

.E2 mmm .20_.0..0xm 0.0.0.00E0. E00. .0
To 020x02 020 2'. 020x06 xiv .0202.0E .Alv 0.0.5 2. 200.6 0..2o0_0E .2 so. :0 0.8000 002080.02“. .N...” 0.5m...

.5... 2.025052.
com 0.4 ems mm. 004 m.m omm 00m

 

   

. r...) </ H... .. s t i ‘
- 80 m
n.
U
a.
- 8.. o
.4.
N
3
..oom .d
d
m
-oow :.
ua

 

 

 

 

000.

110

.E2 o5 ”20:0..0xm. 0.2.0.00E0. E00.
.0 .---V 020x02 020 2'. 020x22 ..---. .0202.0E Al. 0.0.5 2. 00. .0202... .2 so. :0 0:00am 002080.02“. .m.n 0.39”.

.55 2.025033
8.. m... 8.. ms. 8.. 0... 8m 0%

 

......_. ...... ...... . ........ . .2. _ . o
...... .. .. i {5% <\o\.w\av-I-v—...

altnl 0| 5

...). N n 1

- .. . n

’........-J J. .- m

x . i DON—...—

S

3

3.—

N

3

a

- oovd

0

M

3..

a

 

 

 

ooo

111

The fluorescence power of malachite green and phenol red are greatest in
dioxane, while the fluorescence power of purpurin is greatest in methanol and
water. This result indicates that the fluorescent forms of malachite green and
phenol red require less polar environments, while that of purpurin prefers polar
environments. The optimum solvent conditions and maximum emission
wavelengths are summarized in Table 3.1. The fluorescence power of both
malachite green and purpurin is greater than that of phenol red. In addition, the
fluorescence of purpurin is somewhat inconsistent, and spectra could not be
measured reproducibly for solutions of similar concentration. Because the
fluorescence spectra of malachite green are more reproducible than those of
purpurin during the initial studies, malachite green is selected for use in further
exploratory studies.

In addition to the fluorescence power, the appearance of malachite
green in water, methanol, dioxane, and hexane is used to identify the most ideal
environmental polarity. Malachite green is known to exist in two forms in
solution, as shown in Figure 3.4. The ionic form, which predominates in polar
solution, is blue-green and absorbs strongly at 618 nm. The color observed for
solutions of malachite green in water and methanol is characteristic of the ionic
form. The neutral form, however, predominates in non-polar solutions and is
nearly colorless. The color observed for solutions of malachite green in dioxane
and hexane is characteristic of the neutral form. Based on the color of malachite
green solutions and the fluorescence spectra in Figure 3.1, the colorless form is

the most fluorescent and is desirable for subsequent measurements.

112

Table 3.1. Optimized fluorescence excitation and emission parameters.

 

 

Excitation Emission
Solvent Wavelength (nm) Wavelength (nm)
Malachite Green Dioxane 325 375
Purpurin Methanol 325 405
Water 325 418
Phenol Red Dioxane 310 360

 

113

IO

>m\

.E.0. 0.20. ”.292 22029502 20.. .2005 0._200.0E .0 005.02% .1.» 0.20.“.

 

1)

O...

D Q. r

114

To determine the utility of an explosives detection device based on
malachite green, the quenching behavior with model compounds is measured in .
dioxane. Quenching data for malachite green with nitrobenzene is given in
Figure 3.5. The quenching constant for nitrobenzene in dioxane is 120 M",
which is significantly greater than the value measured previously in methanol (50
M", Chapter 2). The quenching constant for nitromethane in dioxane is 5 M",
which is significantly less than the value measured previously in methanol (36
M"). Despite the decrease in quenching constant for nitromethane, the increase
observed for nitrobenzene is promising and indicates that a device based on
malachite green in a non-polar matrix could be useful for nitroaromatic explosives

detection.
3.2.2 Solid-Phase Fluorescence

For nitrated explosives detection, the use of a solid-state fluorophore is
more advantageous than a solution-phase fluorophore to avoid concerns of
leakage or evaporation of the solvent. First, to verify that malachite green
fluoresces in a solid form, a KBr pellet is prepared containing solid malachite
green. The fluorescence of this mixture is measured in the fluorimeter, and the
spectra are given in Figure 3.6. The increase in fluorescence that is observed
upon addition of malachite green to the KBr indicates that the solid form of
malachite green is sufficiently fluorescent. The emission spectrum of malachite
green in KBr is similar to that in water, indicating that the polar KBr matrix is not

optimal for this application.

115

A
O

 

""" 0 M
"‘ 2X10'3 M
— 5x10'3 M

00
O
l

FLUORESCENNCE POWER
0

 

 

 

 

335 360 385 410 435 460 485 510
WAVELENGTH (nm)

 

2.2-

2.0-
u.
$1.8-

1.6—

1.4-

 

 

 

1.2 l l I
0.004 0.006 0.008 0.010 0.012
NB CONCENTRATION (M)
Figure 3.5. Fluorescence spectra of 2x10" M malachite green in dioxane at
room temperature in the presence of increasing concentrations of nitrobenzene

(top) and the resulting Stern-Volmer plot (bottom). The Stern-Volmer constant
for this data set is 120 M'1 with a correlation coefficient of 0.9997.

116

.E2
mmm 20:0..0xm. ..l. 200... 0..200_0E 9.2.0.200 .0..00 5v. 0 020 Al. 0:00 .9. 0 .0 0.6000 002080.02“. .0..” 050.“.

.55 Eozmd><>>

 

8.. 8.. 8... omm
_ _ L O

J

- 08 m

o

H

...

- 82 o

....—

N

3

- com. a

d

o

- ooom w.

H

 

 

 

 

oomN

117

The fluorescence power of malachite green is greatest in dioxane,
indicating that a non-polar matrix is most desirable. The structure of
poly(ethylene glycol) is similar in polarity to dioxane, as the polymer contains
ether groups and only terminal hydroxyl groups. For design of an inexpensive
field device, however, a poly(ethylene glycol) polymer will require the use of an
additional support material for durability.

The effect of substrate pretreatment is investigated in which the
fluorescence from substrates rinsed in methanol is compared to those that are
unrinsed. Contaminants such as dust and other particulates that could be
adsorbed to the surface may affect the observed fluorescence or the interaction
of the fluorophore with the substrate. Overall, pretreatment of substrates has
little effect on the fluorescence of the substrate or the fluorophore-coated
substrate. Unrinsed substrates are utilized for all subsequent measurements.

The fluorescence of the support materials is also investigated. Glass
slides are found to have lower backgrOund fluorescence in the range of 385 —
400 nm compared to black cardboard. The low background is important to
ensure that the contribution to the overall signal of the sample is minimal.
Mounting of the paper substrates using double-sided tape also minimizes
background fluorescence compared to using white or green tape on the back of
the support. The orientation of the mounted substrate is also unimportant, as
filter paper taped to a glass slide in a variety of possible orientations gives
consistent fluorescence. Sufficient reproducibility is also observed, as the

fluorescence power is consistent upon removing and reinserting mounted

118

substrates. Based on these initial observations, paper substrates are mounted
onto glass slides using double-sided tape for all subsequent measurements.

The fluorescence of malachite green is also studied on various
substrates. Of the substrates investigated, fiberglass filter paper has the lowest
background fluorescence (Figure 3.7, bottom). The fluorescence of the Grab-it
cloth is significantly greater than that of the other substrates (Figure 3.7, top),
indicating that interference with malachite green fluorescence may occur. Other
substrates such as unbleached cardboard, unbleached paper towel, blue cloth,
and sterile bandage pads have greater background fluorescence that is more
irregular, also limiting their use in this application. In comparison to fiberglass
filter paper and a Kimwipe, coated filter paper demonstrates the greatest
fluorescence upon application of 10'3 M malachite green in methanol (Figure
3.8). Based on these results, filter paper coated with malachite green is used for
subsequent measurements.

The concentration dependence of malachite green coated on filter paper
is also of interest for future development of a field device. Fluorescence spectra
for filter paper coated by using 10 mL of various concentrations of malachite
green from 2x10'5 - 1x10'3 M are given in Figure 3.9. The filter paper coated by
using the most concentrated solution displays the greatest fluorescence, and
yields the only quantifiable peak observed for the different malachite green
concentrations. The fluorescence of the filter paper coated with 10‘3 M malachite
green, however, is not reproducible. Repeated studies using 10'3 M and greater

concentrations on filter paper do not yield measurable fluorescence. In addition,

- 119

200

 

160 —

 

FLUORESCENCE POWER

 

 

 

330 380 430 480

 

4000 4

2000 -

FLUORESCENCE POWER

 

 

 

 

0 i . -

330 380 430 480
WAVELENGTH (nm)

Figure 3.7. Fluorescence spectra of a Kimwipe (--), filter paper (-—), a Grab-it
cloth (—), and fiberglass filter paper (—) at room temperature. Excitation: 325
nm. Slit width: 1 mm (top), 5 mm (bottom).

120

.EE m “2.0.3 ...m E: mum .20..0..0xm. .0.:.0.00E0. E00. .0 20.3.00 200.0 0._200_0E
s. no. 2...... 00.000 Al. .0000 .0.... 000.900: 020...... .0000 .0:: ..---. 00.3E.v. 0 .0 0.6000 002080.02... .0.: 0.50...

.E... 2.020.052.
om. om. own 2mm

_ b _ O

 

- OOON

H ooov

 

1 ooom

T ooow

HHMOd EONHOSSHOITH

 

O
.
.
-.-----.

 

 

 

oooo r

121

.EEm ”2.23:5 .2.2 mwm ”20:0:0xm 0.30.0020“

E00. 00 0202200 200.0 0._200_0E .0 020005200200 m:020> 525 00.000 .0000 .0:... .0 0.00000 0020000820 .06 0.090

85 25204052,

 

8.. 8.. 80 o8

_ _ _ O
. .. . . I:
. . ... . germ... .. - 88 m
..w. nu
.. “a
_ ...
- 88 0
3.—
N
. 3
2.00.0... . -88 a
_>_ 90va ..... . W
s. .800 | . M
Evotsl -88 m

_>_ 00:; I

 

 

 

0000—.

122

the fluorescence emission of malachite green on filter paper is more similar to the
spectra observed in polar solvents such as water, indicating that greater
fluorescence can be obtained in a less polar matrix.

Overall, although malachite green fluoresces in the solid state, the
fluorescence cannot be measured reliably when coated on filter paper. Filter
paper, like other cellulose-based paper products, contains many hydroxyl groups
that contribute to the high polarity of the substrate. The result observed for
malachite green on filter paper is consistent with that observed for polar solvents
such as water and methanol, indicating that a less polar matrix is required for this
application. The low background fluorescence of filter paper before and after
coating with malachite green, however, indicates that this material is suitable as

a support for poly(ethylene glycol) films containing malachite green.
3.2.3 Poly(ethylene glycol)-C0ated Substrates

As discussed previously, dioxane and poly(ethylene glycol) (PEG) are
comparable in polarity. Various molecular weights of PEG are investigated for
utility in a field device for explosives detection, including 750, 1000, 1500, and
2000 g/mol. PEG 750 is a viscous liquid, PEG 1000 is a waxy solid, and PEG
1500 and 2000 are solids. Because a solid-state device is desired, the utility of
PEG 750 is limited in this application. In addition, the melting point of PEG 1000
is 37 - 40 °C, indicating that at slightly elevated temperatures, such as in a warm
climate, this material would also be liquefied. Based on this, the PEG 1500 and
2000 are utilized for measurements that follow. For incorporation of malachite

green into the solid-state polymer, two methods are employed. One method

123

involves heating of the polymer beyond the melting point followed by addition of
solid malachite green. Another method involves combination of solutions of the
polymer and malachite green prepared in a suitable solvent. The solubility of
PEG in various solvents is investigated, and methylene chloride, a common
solvent for stationary phase preparation in gas chromatography, is determined to
adequately dissolve both polymers as well as malachite green. The fluorescence
of poly(ethylene glycol)-malachite green mixtures is investigated when supported
on a glass slide, filter paper, and silica particles.

The fluorescence power of glass slides coated with PEG 1500 and 2000
containing malachite green is measured. The spectra in Figure 3.10 correspond
to polymer mixtures prepared by heating the polymer above the melting point
followed by addition of solid malachite green. Malachite green completely
dissolves in the polymer matrix at slightly elevated temperature. The
fluorescence power of malachite green in PEG 2000 is greater than that in PEG
1500 when coated on a glass slide. In addition, the fluorescence emission of
malachite green in the polymer matrix is similar to the spectrum observed in
dioxane, indicating that this matrix is sufficiently non-polar to allow maximum
fluorescence emission. Based on this result, PEG 2000 is utilized for all
subsequent measurements.

Malachite green and polymer solutions are coated onto filter paper in
various mass ratios to determine the ratio that yields the greatest fluorescence
power. Initial studies are performed by applying the polymer-malachite green

solution in methylene chloride dropwise onto the filter paper in a Petri dish.

124

.EE 0 H2.00.3.5
.2.2 0N0 ”20.00.63 0.30.0020. E00. .0 00.5000 00000.0 000.2200. 020.02. 0... 02.0: >0 00.0000 05.... All. 0000
Gun. 020 Al. com? Own. 0 N 2. 200.0 0:50.02. 0E .. 5.3 00.000 02.0 000.0 0 .0 000000 0020000820 .0...» 0.50.0

 

 

.5... 152040225
owe owe 000 omm
t t...) l... s - _ _ o
4 . 2
J.—
m
- 88 o
H
3
S
0
- 88 3N.
O
.4.
m
- 88 M
.4.
H

 

 

 

ooow

 

125

These studies indicate that this method is not reliable or reproducible because
some mass of polymer may be lost to the Petri dish. The amount of polymer lost
is inversely related to the amount of polymer present in the solution, as the
increased viscosity in solutions of high polymer concentration prevents complete
coverage of the filter paper. In addition, the capillary action that causes the
polymer mixture to spread to the edges of the filter paper results in preferential
migration of the malachite green with the solvent front. Once the solvent has
evaporated, regions of higher malachite green concentration are visible by eye.

To overcome this problem, an application system based on capillary
pressure is developed in which the solution is painted onto the filter paper using
a fused-silica capillary. Solutions containing PEG 2000 and malachite green are
prepared in methylene chloride and coated on the filter paper by applying 15 psi
air pressure to the solution. The applied pressure causes displacement of the
polymer solution through the capillary and onto the filter paper. As shown in
Figure 3.11, as the proportion of polymer is increased in the solution, the
fluorescence power of malachite green also increases.

The malachite green polymer is also coated onto silica particles as a
method for explosives detection. The effect of polymer concentration or density
is compared for silica particles with polymer-to-particle mass ratios of 10%, 20%,
and 30% with incorporation of 0.17 mg malachite green. The fluorescence power
for each of these coated particles is given in Figure 3.12. The fluorescence from
these coated particles is also visible by eye under irradiation from a handheld UV

lamp. As observed for the coated filter paper, the greatest fluorescence power is

126

 

1 0000

8000 -

6000 -

4000 -

2000 -

FLUORESCENCE POWER

 

 

 

O r l I

380 430 480
WAVELENGTH (nm)

2000

 

1500 -‘

 

1000 .

500 -

 

FLUORESCENCE POWER

 

 

0 . . .

330 380 430 480
WAVELENGTH (nm)

 

Figure 3.11. Fluorescence spectra of filter paper coated with 0.17 mg malachite
green and 0.8% (—), 1.6% (—), 4.1% (—), 5.4% (—), 8.1% (--), 16.3% (--), and
40.6% (---) PEG 2000. Percentages refer to the ratio of polymer mass to filter
paper mass. Spectra collected at room temperature. Excitation: 325 nm. Slit
width: 5 mm. PMT: 950 V (top), 700 V (bottom).

127

.0 to ”0E: 2000.092. .EE v.0 ”2.0.3 ...m .E2 000 H20:0..0xm
0.0.0.0020. E00. .0 00.5000 0..000w A305. ooom Own. .1. $00 020 ..l. #8 ..---. $3 020 200.0 0..200.0E
0E ... to 020 05.2,. 080 Own. .1. $00 5.2. 00.000 00.00.00 0050 0 v.0 ..0 0.00000 0020000820 .N_..n 0.00.0

E... 152000205

 

omm owe owe 000 800
_ o

- . - 88 a
- 82: m
o
- 88. w.
x
- 880 a
N
- 880 o
3..
- 880 d
m
- 880 w

- 88..

 

 

 

00000

128

observed for greater mass ratios or film thicknesses of the PEG polymer. Some
discoloration of the particles is observed upon exposure to 325-nm laser
irradiation, but when removed from the laser beam path the original color returns.
An exposure study in which the tube is continuously irradiated with a laser source
indicates that the fluorescence decreases rapidly (within 3 min), and only the
background fluorescence of the polymer remains. At this point, it is unclear if
fluorophore degradation would occur to the same degree if a lower power source

is used.
3.3 Conclusions and Future Work

A general protocol for incorporation of a solution-phase fluorophore into
a solid-state device has been developed. Initial studies with various substrates
indicate that a polar paper-based matrix alone is not sufficient to support the
fluorescence of malachite green. In contrast, a less polar PEG solution
containing malachite green fluoresces in the solid state when coated on glass
slides, filter paper, or silica particles. The paper-based substrate coated with
polymer can be developed as a wipe for detection of explosives on surfaces such
as luggage or cargo. An air sampling device could be developed based on
coated silica particles enclosed in a quartz tube. Malachite green may be a
useful fluorophore for these devices if a lower power source is utilized, such as a
handheld lamp. Similar procedures could be investigated for devices containing
phenol red and purpurin. For any device, the fluorescence must be stable upon
source irradiation to prevent false positives from fluorophore degradation. In

addition, significant testing with explosive particles and vapors is necessary to

129

verify that the quenching behavior observed in the solution phase is also

observed in the polymer matrices.

130

3.4 References

1. Lakowicz, J. Principles of Fluorescence Spectroscopy, Plenum Press:
New York, NY, 1983.

131

CHAPTER 4

pH-ENHANCED FLUORESCENCE DETECTION OF ACIDS

A novel method for determination of acidic species based on the pH-
sensitive fluorescence of fluorescein is presented. Fluorescein is a commonly
used fluorophore whose quantum yield changes dramatically as a function of
protonation state and, therefore, of solution pH. For a compound more acidic
than fluorescein (PK. 5 5.5), proton transfer will lead to a decrease in
fluorescence that can be employed in quantitative or qualitative analysis. The
fluorescence response of fluorescein is investigated using a series of organic and
inorganic acids, and is shown to be sigmoidal with the logarithm of proton
concentration. The linear range extends two orders of magnitude and is
centered about the fluorescein concentration in solution. The linear range can be
extended to four orders of magnitude using a sigmoidal curve fit, and can be
shifted to higher or lower concentrations by adjustment of the fluorescein
concentration. This method is applied to the analysis of grape juice and a variety
of wine and vinegar samples following liquid chromatography. Improvements in
sensitivity and selectivity are observed for acids found in these beverages and

foods upon direct comparison to UV-absorbance detection.
4.1 Introduction and Background

The determination of low-molecular—weight (LMW) organic acids (Figure
4.1) is important in a variety of fields. These compounds are present in
physiological fluids as intermediates or final products in many metabolic

pathways. The role of these acids in metabolism of fats, amino acids, and

132

OH OH
Dr

03

O
O
3 4
OH OH
HO HO
O O
O O O OH
5 6 7
OH OH OH
HO HO HO
O OH O OH O
0 0H
8 O O
HO OH

OH

Figure 4.1. Structures of common low-molecular-weight organic acids. Formic
acid (1 ), Acetic acid (2), Lactic acid (3), Oxalic acid (4), Succinic acid (5), Malic
acid (6), Tartaric acid (7), Citric acid (8).

133

carbohydrates makes their determination an important indicator of organ
function.“3 In the environment, LMW organic acids may originate from
combustion of fossil fuels, burning of biomass in incinerators, or from
photooxidation of organic compounds.4 High concentrations of these acids are
found in soils and plants, especially at the soil-root interface where nutrient
uptake occurs. The presence of these acids affects plant uptake of nutrients and
toxins by enhancing the availability of anions to plants.5'6 In addition, the
solubility of toxins such as heavy metals and phosphorous in soil can be affected
by LMW organic acids through complexation and acid-base reactions.7

Another field in which LMW organic acids are important is food and
beverage analysis.1'8'9 One of the most common applications is in quality control
of wines, as the LMW organic acid content is strongly correlated with the color,
odor, and flavor of the wine."'9 Grapes, the source of many traditional wines,
contain numerous acids such as tartaric, citric, and malic acids. The properties
of the wine are related to the distribution of these acids, which are often added to
wines after fermentation to decrease the pH.8 In addition, malic acid can
undergo bacterial fermentation to form lactic acid during wine production, which
results in a pH decrease and decreased wine quality.8 The complexity of the
fermentation process and the dependence of wine quality on pH motivate the
analysis of acids in wines and related products in the food and beverage industry

In addition to wine, the quality of several related food and beverage
products can also be affected by acid content. For example, grape juice contains

acids such as tartaric, citric, and malic acids. As with wine, these acids are often

134

added to grape juice for pH adjustment, and their relative concentrations are
important in the overall quality of the product. Upon fermentation, carbohydrates
or sugars in the juice are converted to alcohols to produce wine. Further
fermentation of wine or grape products involves oxidation of alcohols to acids,
producing Vinegars. Thus, many types of vinegar also contain a number of acids
that are related to the origin of the vinegar as well as the degree of oxidation. To
decrease cost, many types of vinegar are produced from acetic acid solutions, in
which sugars and colors are added to reproduce the flavor of natural aging and
fermentation. Analysis of the acid content of such Vinegars can reveal the
natural products present only in high quality, aged vinegar.

The most common methods for analysis of LMW organic acids in foods

and beverages are liquid chromatography (LC)9'22

or capillary electrophoresis
(CE)8'""23"33 with UV-absorbancemo'18'32 or conductimetric detection.9'1“‘1"""“'33
Both LC and CE provide sufficient separation of LMW organic acids in the neutral
and ionic forms, respectively. The sensitivity of UV-absorbance and
conductimetric detection methods are also sufficient for the primary LMW organic
acid components in food and beverage samples. The greatest drawback of UV-
absorbance and conductimetric detection, however, lies in the lack of selectivity
of these methods. Conductimetric detection responds to all ionic species by
measuring the change in current that results from the presence of charged
molecules. Similarly, UV-absorbance detection responds to all species with

sufficient molar absorptivity at the monitored wavelength. Although UV-

absorbance detection is often considered more selective than conductimetric

135

detection, this is not the case for LMW organic acids. The main UV-absorbing
functionality on LMW organic acids is the carboxylic acid group, for which the
wavelength of maximum absorbance is from 180 — 220 nm.‘"”"3'9'w'30'31 The
potential for interference is great, as many, if not most, other functional groups
also absorb in this region. To overcome this limitation, indirect UV-absorbance
detection is used with CE and provides greater selectivity than direct detection.
The selectivity and sensitivity of indirect UV-absorbance detection arise from the
use of a secondary compound that absorbs at a longer wavelength, thereby
reducing interferences.6

This work describes a new alternative to indirect UV-absorbance
detection to improve selectivity in the detection of LMW organic acids. A novel
detection method based on pH-dependent fluorescence of fluorescein is
presented. Fluorescein is a commonly used fluorophore with desirable
properties such as a high molar absorptivity, high fluorescence quantum yield,

“'35 ln solution, fluorescein occurs in cationic, neutral.

and good photostability.
anionic, and dianionic forms, depending on the pH (Figure 4.2). The
fluorescence properties, including the molar absorptivity (a) and quantum yield
(CD), change as a function of the degree of protonation.34 These quantities are
greatest for the dianionic form and decrease with increasing protonation to the
anionic, neutral, and cationic forms.34 The pH-sensitivity of fluorescein results in
a decrease in fluorescence in the presence of acidic species.

In this study, an experimental system has been developed for separation

of neutral LMW organic acids by reversed-phase LC using an unbuffered mobile

136

COOH 2. = 2.08 COOH

pK
\ O 4 v \
(+3
0 OH HO O O

 

 

H0
843, = 53,000 M-1 cm" 2,3,, = 11,000 M" cm-1
0437 = 0.18 //V 043, = 0.30
pk: = 4.31
e e
—->
0
HO 0 o o o o
8472 = 29,000 M’1 cm-1 8490 = 76,900 M‘1 cm'1

0,7, = 0.37 0,9,, = 0.93

Figure 4.2. Acid-base equilibria of fluorescein in aqueous solution at room

temperature. Molar absorptivity (s) and quantum yield ((1)) are given for each
form at the wavelength of maximum absorbance.

137

phase. After the separation, the pH is adjusted to a more neutral value and
fluorescein is introduced post-column. When no acids are present, a constant
elevated background of fluorescein is detected by laser-induced fluorescence. In
the presence of LMW organic acids, however, the fluorescence is decreased
below the background level to yield a negative signal. The decrease in
fluorescence arises because the acids transfer protons to the fluorescein
molecules, causing a decrease in the fluorescence quantum yield. This method
facilitates the determination of LMW organic acids in grape-based juice, wines,
and Vinegars by simplifying the chromatograms of the complex sample matrices

by detecting only acidic species.
4.2 Experimental Methods
4.2.1 Reagents

Reagent-grade acetic acid (Aldrich), benzoic acid (Spectrum), citric acid
(MCB Chemical), formic acid (JT Baker), hydrochloric acid (JT Baker), lactic acid
(JT Baker), maleic acid (JT Baker), malic acid (Sigma), oxalic acid (JT Baker),
succinic acid (Sigma), and tartaric acid (JT Baker) are used to prepare aqueous
stock solutions at 25 mM concentration. Aqueous fluorescein solutions are
prepared at 5 mM concentration by using the disodium salt (Sigma). High-purity,
spectroscopic grade methanol (Honeywell Burdick and Jackson) and distilled
deionized water (Corning Glass Works, model MP-3A) are used for solution
preparation and to prepare mobile phases for liquid chromatography. Fresh
solutions are prepared daily, and stock solutions are stored in a refrigerator to

prevent degradation. Grape juice, wines, and Vinegars are purchased from local

138

supermarkets and are filtered prior to analysis. Sodium nitrate (JT Baker) is used
as a void marker for UV-absorbance detection and is added to samples in a final

concentration of 0.1 mM.
4.2.2 Chromatographic System

A system has been designed for the analysis of acids in beverages and
other food samples (Figure 4.3). In this system, a reciprocating piston pump
delivers the mobile phase at a nominal flow rate of 1.0 mL/min. The mobile
phase for separation of acids in beverages consists of 1% methanol in distilled
deionized water, adjusted to pH 3.0 using hydrochloric acid. Samples are
introduced by using a 20—pL injection valve (Valco Instruments, model EC6W)
onto the reversed-phase chromatographic column (Supelco Hypersil ODS, 5 pm
particles, 250 mm x 4.6-mm id). The terminus of the column is connected to a
stainless steel tee (Valco Instruments, model ZT 1), where the effluent is mixed
with an aqueous solution of 10 mM sodium hydroxide (Aldrich) to increase the pH
to approximately 8.0. The sodium hydroxide solution is delivered to this tee at
0.15 mL/min by using a syringe pump (Applied Biosystems, model 140) and a
10-mL injection valve (Rheodyne, model 7125). The pH-adjusted effluent is then
directed to a second tee where an aqueous solution of 5 mM fluorescein is
introduced at 0.30 mL/min by using another syringe pump and 10-mL injection
valve. After mixing, the effluent is directed to a fluorescence detector (vide infra).
For comparison, the column effluent is also directly connected to a UV-visible
absorbance detector (Jasco, model UVIDEC-100-V, 215 nm, 1 cm pathlengh).

The UV-absorbance data are collected by using a commercially available

139

02.00. E02000 .0.0 E1000

0.00.02. 002.. .00... 02.00. .00.0 000.2.0.0 .0:. 0.00.02. 002.. 0:00 .00.>00 00.008-00.000 u 000 .000. .0._0...0E0.0...0
n 0.5.0 00.... u 0 .020. u .. .0>_0> 20000.2. n . .2.0000.00.. .2E m .0 302.00 .0 0E00 00.08.00. E0.000 SE 0. .0
E0200 ”m. 0E0n. .0_0E00 .0 E0200 u< 0E0n. .20..00.00 002800.00: .20020000-...0 0...... >000.00.0E0.00 0.00.. >0
0000.0>00 2. 00.00 2200.0 >>_>.._ .0 22.02.220.00 .0. 0000 E0.0>0 .0.20E..00x0 02. .0 E0.00.0 0..0E000w .06 0.00.0

_ m...w<>>_
mmwdj ._ - u. ._
COOI 1milllllriflaIDOIIlllllllllrllllllhlll“lrllllllWHlmlmlululultw
E ..n...,,..,..,..,..,..,..,M

 

 

 

 

 

 

 

 

 

 

 

140

interface (National Instruments, model BNC 2120) and associated software
(National Instruments, Labview, version 5.1). Retention times are consistent for
known solutes with each detector, indicating that no correction is needed in order
to align chromatograms obtained using UV-absorbance and fluorescence

detection.
4.2.3 Spectroscopic System

Chromatographic detection is achieved by using a laser-induced
fluorescence spectrometer as shown in Figure 4.3 and described in detail
elsewhere.36 In this system, a helium-cadmium laser (Melles Griot, model 3074-
40M, 325 nm, 20 mW) is focused onto a fused-silica capillary flow cell (Polymicro
Technologies, 180-um i.d., 340-pm o.d.). Fluorescence emission is collimated,
filtered to remove scattered laser light, and refocused onto the entrance slit of a
0.34-m Czemy-Turner monochromator (Instruments SA, model 340E, 300
groove/mm grating). Emission is collected by using a therrnoelectrically cooled,
charge-coupled device (CCD) detector (Instruments SA, model (A)TECCD-
2000x800-7) operated at 240 K, which allows collection of the entire fluorescence
spectrum over a 220—nm range. The spectrometer is also equipped with a
photomultiplier tube (Hamamatsu, model R-106, 10-nm bandpass) used for
chromatographic studies. Instrument control and data acquisition are provided
by a commercially available electronic interface (Instruments SA, model
Datascan 2) and associated software (Instruments SA, Spectramax for Windows,

version 3.1).

141

For initial steady-state studies with test acids, the CCD detector is used to
collect the fluorescein emission spectrum as a function of acid concentration.
The peak height at the maximum emission wavelength (510 nm) is used for the
construction of a static calibration curve for each test acid. For chromatographic
studies, the photomultiplier tube is used to measure the fluorescence emission at
510 nm as a function of time. Peak heights and areas are calculated from the
chromatographic peaks and are used for the construction of calibration curves.
Peak-fitting software is used for data analysis (Microsoft Excel, version 2003)

and non-linear regression (Table Curve 20, Jandel Scientific, version 2.02).
4.3 Results and Discussion
4.3.1 Method Development and Optimization

As discussed previously, the decrease in fluorescence is related to the
number of protons transferred from the acid to fluorescein. The number of
protons available is dependent on the relative strength of the acid compared to
the strength of fluorescein. For acids with pKa s 5.5, approximately 1 pH unit
below the pKa of the fluorescein dianion, protons are completely transferred to
fluorescein and measured as a decrease in fluorescence. For monoprotic acids,
the concentration of protons available for donation to fluorescein is equal to the
concentration of acid in the sample. For diprotic and triprotic acids with all pKa
values 3 5.5, however, the concentration of protons available is two and three
times the concentration of acid in the sample, respectively. In all cases, the
behavior of fluorescein and an acid is inherently similar to the titration of a weak

base. In a traditional titration of this type, the pH is monitored as a function of the

142

volume of acid added, while in this method the fluorescence is monitored as a
function of pH (or —log [acid]). Thus, the calibration curve is inherently sigmoidal

in shape, as demonstrated in Figure 4.4.
4.3.1.1 Analysis of Test Acids

The response of fluorescein is first investigated by measurement of the
solution fluorescence in the presence of various acid concentrations. Four test
acids are selected for steady-state studies, including a strong acid (hydrochloric
acid), a monoprotic weak acid (benzoic acid, pKa = 4.21), a diprotic weak acid
(maleic acid, pKa = 1.97, 6.07), and a triprotic weak acid (citric acid, pK.I = 3.15,
4.77, 6.40). The fluorescence of a 10‘5 M fluorescein solution is measured in the
presence of 10'3 to 10'7 M concentration of each acid (Figure 4.4). As mentioned
previously, the number of available protons is dependent on the pKa values for
the acids relative to that of disodium fluorescein. Although maleic acid is a
diprotic acid, it behaves as a monoprotic acid with fluorescein because the
second pKa value is greater than ~5.5. Similarly, citric acid behaves as a diprotic
acid with fluorescein because the third pKa value is greater than ~5.5. For this
reason, the concentration of available protons in the solution is twice the
prepared concentration of citric acid.

After correction of the proton concentration, Figure 4.4 demonstrates
that fluorescein has the same sigmoidal response to all four acids, regardless of
acid strength for pKa values 5 5.5. The resulting advantage is that a single
calibration curve can be constructed using a test acid and applied for the

quantitation of any acid in an unknown sample. It is also noteworthy that the

143

.080 u .00. .8058". 8.0.28
005.202 ..0. 0.00 0....0 .... 0.00 0.0.0E .8. 0.00 0.0N200 .... 0.00 000200.02. ”00.00 .00... .E2 05 .0 2.000082.
.0 .0260 002000060: 02. E0... 002.E.0.00 20.02. 2. 00.00 .00. .00. .0. 0>.00 20..0.0..00 .00.0E0.w 6.0 0.00.0

.2. 20.22.200200 0_0<

 

0-0.. v.2. 0-0.. 0.0?
_ _ _ — _ o
....
-o.m
O
a
......
1 ON 3
a
N
O
1 om 3..
d
0
- 83M:
8

 

 

 

 

0
L0

144

inflection point of the sigmoidal curve is centered about 10'5 M acid concentration
and the curve is linear for two orders of magnitude. At the inflection point, the
concentrations of fluorescein and acid are equivalent at 10'5 M. Hence, another
advantage of this method is that the linear range can be translated to higher or
lower concentrations for a variety of applications by adjustment of the fluorescein
concentration.

To linearize the calibration curve, the fluorescence power of fluorescein
as a function of acid concentration is fit by nonlinear regression to a sigmoidal
function.

b

1+ expLELi)
d

In this function, y is the fluorescence power, x is the logarithm of the acid

y=a+

 

(4.1)

concentration, and a, b, c, and d are fit parameters determined by least-squares
regression (Table 4.1, top). For the test acids, correlation coefficients (R2) for the
sigmoidal function are greater than 0.994, indicating a strong correlation. The fit
parameters a, b, c, and d are use to predict the fluorescence power resulting
from a given concentration by graphing the logarithm of the acid concentration (x)

as a function of the sigmoid transform function (2).

z=—dln[—b——1]+c (4.2)
y—a

The remaining variables in Equation 4.2 are as defined in Equation 4.1. The

resulting calibration curve is linear over four orders of magnitude (Figure 4.5,

Table 4.1, bottom).

145

Table 4.1. Correlation coefficient (R2) and parameters of fit (a-d) for test acids to
sigmoidal function (Equation 4.1) (top) and the corresponding correlation to the
sigmoid transform function (Equation 4.2) (bottom).

 

Sigmoidal Function

 

 

 

 

Acid a b c d R2
Hydrochloric 2912 41420 5.012 0.5587 0.9946
Benzoic 8289 35560 5.027 0.3732 0.9962
Maleic 5153 39290 5.145 0.5384 0.9944
Citric 5389 35380 5.282 0.4646 0.9944
Linear Sigmoidal Transform
Acid m b R2
Hydrochloric 0.9854 0.0718 0.9943
Benzoic 0.9432 0.2707 0.9320
Maleic 0.9998 -0.0193 0.9898
Citric 0.9949 0.0213 0.9874

 

146

.830 u ...... .8088 880:8 .08... ..u. 28
0....0 .... 0.00 0.0.0E ..o. 0.00 0.0N200 .3 0.00 0.8.2008? ”00.00 .00... .E2 05 .0 2.000082. .0 .0300 002000080...
02. E0.. 002.E.0.00 20.0.5 2. 00.00 .00. .00. .0. 220.020.. 0.0E0.0 02250.8. 0>.00 20:08:00 .002... .00 0.00.0

.2. 29220200200 08...

 

 

 

 

$0? Vow 80F @OF

P — F L _ N
.1 me
9
1VW
D
[mu
m
uoN
ma.
1&0
a
.. W

. 1w

..

 

 

 

 

147

4.3.1.2 Chromatographic Optimization

This pH-sensitive detection method can be practically applied to
separations by LC or CE. In order to separate a standard mixture of LMW
organic acids by reversed-phase liquid chromatography, the acids must be
uncharged. Thus, the pH of the mobile phase is adjusted to pH 3.0 using
hydrochloric acid. A buffer cannot be used for acidification as it will interfere with
the pH-sensitive detection. A small amount of methanol is included in the mobile
phase to solubilize the hydrocarbon stationary phase and prevent collapse.37
Using a mobile phase of 99:1 waterzmethanol (pH = 3.0), near baseline resolution
of several standard acids is achieved in less than 7 min (Figure 4.6). Retention

factors are calculated by

(4.3)

where tR is the retention time of the standard acid and to is the retention time of
sodium nitrate (a non-retained marker). The retention factor for each acid,
summarized in Table 4.2, is compared with those in the beverage samples in
order to identify the components.

The retention order of these LMW organic acids follows directly from a
comparison of their structures (Figure 4.1). The retention factors are precise
within 0.5% for all of the acids under these conditions. For a reversed-phase
separation, retention depends on the number of carbonsin the backbone of the
acid, as well as the polarity of the molecule. Polarity is related to the number of

acid groups and hydroxyl groups present on the molecule. For compounds with

148

.0. 28 0.58:0 .... 0.8 8.0 ..0. .28 2.80 ..0. .28 2.8. .0.. 28 gas .8.

0.00 0.E.0. AN. 0.00 02020. .3. 0.00 0:06 ”00.0.00 .E2 0.0 ”20.80.00 00200.0000->D .00000200 E20. .20.00.E0
E2-0..0 .20..0..0x0 E2000 ”20.80.00 0020000800 000002...000._ .2.E...E 0.. ”0.0. 2.0.0 .0 ...0 ..0202.0EH.0.0>>

:00 ”00020 0:00.). .EE 0.0 x EE 000 008.200 E10 .000 ..0.00>I .2E0.0O .8. 20.80.00 0020000800
000002...000. 020 2. 20.80.00 00200.0000->3 >0 002.0.00 00.00 0.2008 >>.>... 20EE00 .0 E0.00.0E0.2O 0.0 0.00.0

 

2.2.0.2:
. 0 0 e 0 N F
0 . 0 0 v m N .
F
m

 

 

 

 

HSNOdSHH OSZI'IVWHON

 

 

 

 

149

Table 4.2. Summary of retention factors (k) and pKa values for standard acids in
water.

 

 

Acid k‘ pK2
Oxalic acid15 -0.006 1 0.003 1.23, 4.19
0.206 :1: 0.004
Tartaric acid 0.5263 1 0.0004 2.98, 4.34
Formic acid 0.646 1: 0.002 3.75
Malic acid 0.855 :1: 0.001 3.40, 5.11
Lactic acid 1.160 :I: 0.006 3.85
Acetic acid 1.386 :I: 0.007 4.76
Citric acid 1.739 :I: 0.009 3.15, 4.77, 5.40
Succinic acid 2.52 i 0.01 4.16, 5.61

 

a Retention factors determined by using sodium nitrate as a void marker. Errors
are representative of the standard deviation of replicate measurements (n = 3).

b Two peaks are often observed for oxalic acid. The first peak may be a
decomposition product.

150

the same polarity, retention increases as the number of carbons increases. For
monocarboxylic acids, retention increases from formic to acetic acid, and for
dicarboxylic acids, retention increases from oxalic to succinic acid. Increasing
the number of acid groups increases the polarity of the molecule, resulting in
decreased retention. For acids with the same number of carbons, retention
decreases with increasing number of acid groups from acetic to oxalic acid.
Addition of hydroxyl groups to the molecule also increases the polarity and
decreases the retention. For acids with the same number of carbons and acid
groups, retention decreases with increasing number of hydroxyl groups from
succinic to malic to tartaric acid. Overall, this column and mobile phase are
behaving as expected and the precision of the retention factors allows

identification of unknown compounds with confidence.
4.3.1.3 Detector Optimization

After the separation, the pH is adjusted to ~80 and the fluorescein
solution is introduced. To avoid corrosion and salt precipitation in
chromatographic pumps, the sodium hydroxide and fluorescein solutions are
introduced by using large volume (10 mL) injection loops. The solutions are
introduced separately to allow independent control of pH and fluorescein
concentration. After optimization, introduction of a solution containing fluorescein
at high pH could be used to accomplish pH adjustment and fiuorescein addition
simultaneously.

For this new detection system, several parameters must be optimized.

The first consideration is dilution of the sample as a result of the post-column

151

addition of fluorescein and sodium hydroxide solutions. To this end, high
concentrations at low flow rates of each are desirable. Based on the corrosive
nature of high concentrations of sodium hydroxide, a high concentration limit is
set at 10 mM (pH z 12). The optimum flow rate of the sodium hydroxide solution
will introduce enough hydroxide ions to neutralize the excess protons contained
in the pH 3 mobile phase. The flow rate is varied from 10 — 30% of the mobile
phase flow rate, or 50 — 150 uL/min with a constant mobile phase flow rate (0.5
mL/min) and fluorescein flow rate (50 pL/min). As shown in Table 4.3, the
background fluorescence increases dramatically for sodium hydroxide flow rates
from 10 — 15% of the mobile phase flow rate, but is relatively constant at flow
rates greater than 15%. Therefore, when the flow rate of sodium hydroxide is
greater than ~15% of the mobile phase flow rate, the pH has been increased to a
level at which fluorescein is completely deprotonated. Above this level, the
fluorescence background no longer increases with increasing pH. This result is
consistent with combining a mobile phase pH =1 3 ([H”] = 10'3 M) with a basic
solution of pH z 12 ([OH‘] = 10'2 M) in a 10:1 ratio to give a neutral pH.

The optimum fluorescein concentration is directly related to the
concentration of acidic species in the sample. At the detector, the concentration
of fluorescein should be equal to or greater than the acidic solute concentration.
In the region in which the fluorescein concentration is up to an order of
magnitude greater than the acid concentration, the change in fluorescence with
acid concentration is measurable and quantitative (Figure 4.4). To determine the

concentration of the sample at the detector (C), Equation 4.4 is used with a

152

Table 4.3. Optimization of the flow rate of a 10 mM sodium hydroxide solution.
Mobile phase: 0.5 mL/min, pH = 3. Fluorescein: 10'4 M, 50 uL/min.

 

 

Relative Background
NaOH Flow NaOH Flow Fluorescence Background]
Rate (pL/min) Rate (%)' (9A) Noise (HA) Noise
50 10 0.0258 0.0037 6.913
55 11 0.0664 0.0168 3.962
60 12 0.4017 0.0099 40.39
75 15 0.4092 0.0098 41 .64
100 20 0.4435 0.0099 45.04
125 25 0.4588 0.0099 46.19
150 30 0.5103 0.0111 46.08

 

a Relative sodium hydroxide flow rate is with respect to the mobile phase flow

rate of 0.5 mUmin.

153

sample Chromatogram.

C — (30‘4”sz 4 4
‘ V 1/2 ( ' )
0(1+k)(2n)

 

In this equation, Co is the initial or injected concentration of the solute, Vin; is the
injected volume, k is the retention factor of the solute, N is the plate number, and
V0 is the void volume of the column. The plate number N is determined from the

retention time (tR) and the width of the chromatographic peak (w) by

t 2
4.3N =16(—£—] (4.5)

The void volume of the column is determined as the product of the flow rate and
the void time from injection of a non-retained solute. Using these equations, a
25-fold dilution factor of the column in this configuration is determined. The
fluorescein concentration and flow rate are selected by using the information
about dilution of the sample as well as the expected concentrations of acids in
real samples to be analyzed. As a result of fluorescein solubility in water, a high
concentration limit of 5 mM is established. Using this concentration and a
column flow rate of 1.0 mL/min, a flow rate of 300 uL/min fluorescein is selected
to yield a concentration of ~1 mM fluorescein at the detector.

Detector-related parameters are also optimized, including PMT voltage,
integration time, and slit width. The PMT voltage is varied from 500 — 800 V at
constant integration time (0.2 s) and slit width (0.2 mm), and the background and
noise are given in Table 4.4. A greater PMT voltage provides a greater

amplification of the photocurrent, but may also amplify the noise simultaneously.

154

Table 4.4. Optimization of the PMT voltage. Mobile phase: 0.5 mL/min, pH = 3.

Fluorescein: 10‘4 M, 100 uL/min. Sodium hydroxide: 10'2 M, 75 pL/min.
Integration time: 0.2 s. Slit widths: 0.2mm.

 

 

Background Background!
PMT Voltage Current (0A) Noise (0A) Noise
500 0.0652 0.0030 21.620
600 0.2302 0.0039 59.649
700 0.8083 0.0120 67.476
800 2.1355 0.0319 66.940

 

155

Although the background fluorescence increases with increasing PMT voltage,
the noise also increases. Beyond 700 V, no further increase in background/noise
ratio is observed. A PMT voltage of 700 V is utilized in all subsequent
measurements.

Integration time is another parameter that affects the output signal for
analysis of acids by directly affecting the number of data points collected per unit
time. The number of data points collected does not affect the background
fluorescence level, but does affect the noise (Table 4.5). Therefore, the most
beneficial integration time is 1.0 3. However, for the best representation of
chromatographic peaks, a sufficient number of data points must be collected
during elution. As seen in Figure 4.6, many of the chromatographic peaks are 10
— 15 s in width, and an integration time of 1.0 5 may cause loss in information.
For this reason, an integration time of 0.2 s is utilized in all subsequent
measurements.

The entrance and exit slit widths are also optimized to provide the best
enhancement in the background/noise ratio. The slit widths are varied from 0.1 —
1.0 mm at constant PMT voltage (700 V) and integration time (0.2 s), and the
background and noise are given in Table 4.6. Increasing the slit widths allows
more light to be collected by the photomultiplier tube, increasing the background
fluorescence. As with PMT voltage, however, the increased fluorescence is
accompanied by increased noise. As the slit widths increase, the background
and noise increase as expected. The background/noise ratio also increases over

this range. At slit widths of 1.0 mm, the photomultiplier signal is out of range and

156

Table 4.5. Optimization of the PMT integration time. Mobile phase: 0.5 mL/min,

pH = 3. Fluorescein: 104 M, 100 uL/min. Sodium hydroxide: 10'2 M, 75 pL/min.
PMT: 700 V. Slit widths: 0.1mm.

 

 

Integration Background Background]
Time (3) Current (0A) Noise (0A) Noise
0.1 0.2355 0.0084 26.665
0.2 0.2359 0.0063 35.372
0.25 0.2365 0.0056 40.055
0.5 0.2374 0.0039 57.145
1.0 0.2367 0.0031 71.392

 

157

Table 4.6. Optimization of the entrance and exit slit widths. Mobile phase: 0.5
mL/min, pH = 3. Fluorescein: 10'4 M, 100 pL/min. Sodium hydroxide: 10'2 M,
75 uUmin. PMT: 700 V. Integration time: 0.2 s.

 

 

Slit Widths Background Background]
(mm) Current (pA) Noise (0A) Noise
0.1 0.1665 0.0054 31.058
0.2 0.8083 0.0120 67.476
0.5 8.1670 0.0804 101.534
0.8 29.1043 0.2045 142.286

 

158

could not be measured. Based on this data, slit widths of 0.8 mm are most
beneficial. However, in order to prevent damage to the PMT from continuous
exposure to such large background currents, slit widths of 0.2 mm are selected to

maintain the current at 1 - 2 pA for all subsequent measurements.
4.3.2 Analysis of Real Samples
4.3.2.1 Analysis of Grape Juice

As discussed previously, grapes and related products contain a number
of LMW organic acids. The acids determined in a filtered red grape juice sample
are shown in Figure 4.7. By comparison to the standard acid chromatogram
(Figure 4.6), the grape juice sample contains oxalic, tartaric, malic, and citric
acids. A comparison of the chromatograms in Figure 4.7 indicates that some of
the native UV-absorbing compounds in the red grape juice sample are not acidic
in nature. As a result, the chromatogram collected by using fluorescence
detection is somewhat simplified in comparison to that using UV-absorbance
detection. The utility of the proposed method lies in the simplicity of the resulting
chromatograms by eliminating potential interferences and improving the ability to

correctly identify acidic species in a variety of samples.
4.3.2.2 Analysis of Wines

Red and white wine are alcoholic beverages prepared by the natural
fermentation of grapes. Yeast is added to crushed grapes to aid the fermentation
process, converting sugars into alcohols. The type of wine produced depends

both on the type of grape and yeast used for fermentation, and the quality of the

159

..m. 000 8.0 .3 000 205 .8. 565.5
AN. 0.00 02020. .. ... 0.00 2.006 ”00.0.8 .00 050.“. 2. 00000000 00 0200.028 0.2002090E900 .8. 20:00.00
0020000200.. 0000020000. 020 2. 20:00.00 00200.0000->: >0 002.0000 00.0.. 000.0 .0 E0.00.0E200 .50 050.“.

.22... ms:
0 0 m .V m N F

_ _ _ _

r

 

 

 

SSNOdSEH CEZI'IVINHON

 

 

 

 

 

160

wine is related to the acid content. Because they are prepared directly from
grapes, the acid content of wine is expected to similar to that of grape juice. As
shown in Figure 4.8, red wine contains many of the same acidic species as
previously identified in red grape juice. By comparison to the standard acid
chromatogram (Figure 4.6), the red wine sample contains oxalic, tartaric, malic,
lactic, and succinic acids. A comparison of the chromatograms in Figure 4.8 also
indicates that some weakly absorbing acids such as succinic acid would be non-
detectable using UV-absorbance detection. The fluorescence method may
identify acids important to wine quality that could be undetected using UV-
absorbance detection. The chromatograms obtained by using UV-absorbance
and fluorescence detection for determination of acids in a white wine sample are
shown in Figure 4.9. By comparison to the standard acid chromatogram (Figure
4.6), the white wine sample contains oxalic, tartaric, and malic acids. As
expected, the acid content of both wine samples is consistent with that of the

grape juice sample.
4.3.2.3 Analysis of Vinegars

Vinegars are the product of further fermentation or oxidation of wines,
resulting in the conversion of ethanol into acetic acid. Just as the quality of wine
depends on the acid content, the quality and flavor of vinegar is related to the
quality of the original grape product. The most flavorful Vinegars are allowed to
ferment for many months or years and contain a number of acids in addition to
acetic acid, the primary component. Because they are prepared directly from

wines, the acid content of red and white wine Vinegars is expected to be similar

161

.6. 000 2585 ..m. 00.... 200 ...V. :25ch .8. god 2.05 ..N.
0.00 00000. .A 3 0.00 2.008 ”00.0.0m ”00.2om 0.0 050.”. 2. 00000000 00 0.8.0.0200 0.000.020E200 .8. 2000000
0020000200: 0000052000. 020 2. 20:00.00 0020020000->D >0 002.0000 02.3 00. .o E0590Eo220 .06 0.50.".

.22... 0.2:
N 0 m 4 m N F

_ _ b _ _

 

 

 

 

ESNOdSEH GSZI'IVWHON

 

 

 

162

..m. .000
0..0E ..N. 0.00 0.2000. A 5 0.00 0..0xo “00.0.0m 0.0 050.“. 2. 00000000 00 0200.028 0.200.090E0200 .8. 280000
0020000200... 0000021000. 020 .00 20.00000 00.200.800-20 >0 002.050 02.3 00.22, .0 E0.00.0EoEO .06 0.50....

.22... 055
n 0 0 e m N F

. _ . . _ L .
_
.
.

 

 

 

 

ESNOdsaa daznkuon

 

163

to that of grape juice and wines. As shown in Figure 4.10, red wine vinegar
contains many of the same acidic species as previously identified. By
comparison to the standard acid chromatogram (Figure 4.6), the red wine vinegar
sample contains oxalic, tartaric, malonic, acetic, and succinic acids. As shown in
Figure 4.11, the white wine vinegar sample contains oxalic, tartaric, acetic, and
succinic acids.

Balsamic vinegar is another type of vinegar prepared from the
concentrated juice or must of grapes. Balsamic vinegar is allowed to age for
several years and, through the fermentation process, becomes a more viscous,
sweet, and flavorful vinegar. The resulting chromatograms from a sample of
balsamic vinegar are given in Figure 4.12. By comparison to the standard acid
chromatogram (Figure 4.6), the balsamic vinegar sample contains oxalic, tartaric,
malonic, acetic, and fumaric acids. As discussed previously for the wine analysis,
the succinic and fumaric acids that are detected in the Vinegars by the
fluorescence method would be undetectable by UV-absorbance detection. For
all of the grape-derived foods and beverages, the fluorescence technique is a
novel and more selective alternative to traditional UV—absorbance detection.

Many other Vinegars are prepared in ways other than from the
fermentation of wines or other grape products, as discussed previously. White
vinegar, for example, can be prepared by oxidizing a distilled alcohol or from
fermentation of grains such as maize. The most common forms of inexpensive
white vinegar are prepared directly from acetic acid as a 5% aqueous solution.

An analysis of a white vinegar sample reveals only one component in addition to

164

..0. 200 228:0 ..v. 200 2.000 ..0. .200 2020... ..N. 200
0.2020. A S 0.00 0..0xo ”00.0.0.0 0.0 050.“. 2. 00000000 00 0200.028 0.000.00.0Eo.0o Am. 20000.00 0020000200..
000002.000. 020 .3 20000.00 00200.0000->: >0 002.0.00 .0002.> 02.3 00.. .0 E0.00.0E0220 .3... 0.50.".

.22... ms:
N 0 0 0 0 N .

_ _ F _ _

0

 

 

 

 

 

ESNOdSEH CIBZI'IVINHON

 

 

 

 

165

.8. 0.00 0.2.0000 .30. 0.00 00000 .8. 23020.20
AN. 0.00 00020. A r. 0.00 0..0xo .0... 050.". 2. 00000000 00 0200.028 0.200.00.0E0.00 .3. 20000.00 002800.000
000005.000. 020 .0.. 20000.00 8200.0000->D >0 002.0.00 .0002; 02.2. 0.0.2, .0 E0.00.0E022O ......0 0.30.".

.25. ms:
N . 0 0 0 0 N .

_ _ _ p _

0

 

 

<
‘1
SSNOdSHH CIHZI'IVWHON

 

 

 

 

166

.... 200 2.025. ..0. 200 2.80 ..0. 0265.5 ..0. 200 2:20.: ..0. 200 255.
..N. 0.00 0.2000. .. F. 0.00 2.0.6 .0... 0.00.“. 2. 00000000 00 0200.028 0.200.00.0E0.20 .3. 20000.00 0020000600
0000020000. 020 9... 20000.00 00200.0000->3 >0 002.0.00 .0002.> 0.E00_00 .0 E0.00.0E0.2O NP... 0.00....

.22... ms:
N 0 0 0 0 N .

_ _ _ P .

O VmN

 

 

 

 

HSNOdSEH CBZI'IVWHON

 

 

 

167

acetic acid (Figure 4.13). A low concentration of oxalic acid is determined, near
the detection limit of 3 mM. The simple LWM organic acid profile indicates that
this sample is most likely prepared directly from an acetic acid solution.

Malt vinegar is another type of vinegar that is traditionally prepared from
malting barley. During this process, the starches in the barley are converted to
maltose, which is fermented and aged. As with white vinegar, 3 more cost-
effective malt vinegar can be prepared from an acetic acid solution with added
caramel color to reproduce the effects of natural aging. An analysis of a malt
vinegar sample reveals numerous acidic components, including oxalic and acetic
acids (Figure 4.14). The remaining components could not be identified using the
standard set of LMW organic acids. The complexity of the profile, however,
indicates that this malt vinegar sample is prepared through the traditional process
of malting barley.

Another common type of vinegar is rice vinegar, prepared from the
fermentation of the starches in rice or rice wine. Similar to balsamic or other
wine Vinegars, the acid profile of rice vinegar is complex (Figure 4.15). By
comparison to the standard acid chromatogram (Figure 4.6), oxalic, tartaric,
formic, lactic, acetic, and succinic acids are identified in the rice vinegar sample.
For all vinegar samples, this method provides a clear profile of the acid content,

and allows conclusions to be drawn regarding the origin of the sample.
4.4 Conclusions

A novel method for determination of acids based on the pH-sensitive

fluorescence of fluorescein has been developed. The native acid-base

168

..N. 0.00 00000 .A F. 0.00 0..0x0 .06 0.00.0 2. 00000000 00 0200.028 0.200.00.0E0.2O .8. 20000.00
820000.000 000005.000. 020 2. 20000.00 00200.0000->: >0 002.0.00 .0002.> 0.0.2. .0 E0.00.0Eo.20 .26 0.00.0

.z__.._. 0.2:
. 0 0 0 m N .

_ _ _ _ _

N

 

 

 

 

HSNOdSEH CIBZI'IVINHON

 

 

 

 

169

.0. 200
00000 .8. 22.05.20 ..N. 22.05.20 .A F. 0.00 2.0.6 .06 0.00.0 2. 00000000 00 0200.028 0.200.00.0E0.2O .8. 20000.00
002800.000 0000023000. 020 2. 20000.00 00200000053 >0 002.0.00 .0002.> ..0E .0 E0.00.0Eo.20 .36 0.00.0

.22... 0.2:
. 0 0 0 0 N .

_ _ — b _

6 ..

 

 

 

 

 

 

ESNOdSEU OSZI'IVINHON

 

 

 

170

..0. 200 22800 ..0. 200 2.80 .0.. 200 2.00_
.8. 0.00 0.E.0. .8. 0.00 02020. .A F. 0.00 0..0x0 .06 0.00.0 2. 00000000 00 0200.0200 0.200.00.0E0.2o .8. 20000.00
002800.000 0000023000. 020 2. 20000.00 00200.0000->D >0 002.0.00 .0002.> 00.. .0 E0.00.0Eo.2o .0 F6 0.00.0

.22... ms:
N 0 0 0 m N .

_ _ _ p —

m

 

 

 

ESNOdSHH CIBZI'IVINEION

 

 

 

 

171

properties of fluorescein and the related changes in fluorescence quantum yield
are exploited for detection of acidic species in foods and beverages. This
method is useful for any compound that is more acidic than fluorescein, allowing
sufficient proton transfer in both aqueous and polar organic solvents. The
behavior of fluorescein fluorescence is sigmoidal with the logarithm of proton
concentration. The linear range extends two orders of magnitude and is
centered about the fluorescein concentration in solution. The linear range can be
extended to four orders of magnitude using a sigmoidal transform function, and
can be shifted to higher or lower concentrations by adjustment of the fluorescein
concentration. This method is validated by using a series of test acids, including
inorganic and organic acids, strong and weak acids, and acids of varying
degrees of protonation. The application of this method to the analysis of grape
juice and derivatives, including a variety of wine and vinegar samples, following
liquid chromatography is demonstrated. Improvements in sensitivity and
selectivity are observed for acids found in these samples upon direct comparison
to UV-absorbance detection. This fluorescence method provides a new tool for
analysis of acidic species with potential use in a variety of disciplines including

forensic, environmental, and clinical analyses.

172

4.5

10.

11.
12.

13.

14.

15.

16.
17.

References

Galli, V.; Garcia, A.; Saavedra, L.; Barbas, C. Electrophoresis 2003, 24,
1951 -1 981 .

Presto Elgstoen, K. B.; Zhao, J. Y.; Anacleto, J. F.; Jellum, E. J.
Chromatogr. A 2001, 914, 265-275.

Verhaeghe, B. J.; Lefevere, M. F.; De Leenheer, A. P. Clin. Chem. 1988,
34, 1077-1083.

Tedetti, M.; Kawamura, K.; Charriere, B.; Chevalier, N.; Sempere, R.
Anal. Chem. 2006, 78, 6012-6018.

Naidu, R.; Chen, Z. Chromatographia 2001, 54, 495-500.

Xu, J.; Chen, 2.; Yu, J. C.; Tang, C. J. Chromatogr. A 2002, 942, 289-
294.

Chen, Z.; Tang, C.; Yu, J. C. J. High Resolut. Chromatogr. 1999, 22,
379-385.

Mato, l.; Huidobro, J. F.; Simal-Lozano, J.; Sancho, M. T. Anal. Chim.
Acta 2006, 565, 190-197.

Hayenga, I. In Analytix: Advances in Analytical Chemistry, 2006.

Dopico-Garcia, M. S.; Valentao, P.; Guerra, L.; Andrade, P. B.; Seabra,
R. M. Anal. Chim. Acta 2007, 583, 15-22.

Grayeski, M. L.; DeVasto, J. K. Anal. Chem. 1987, 59, 1203-1206.

Paredes, E.; Maestre, S. E.; Prats, S.; Todoli, J. L. Anal. Chem. 2006,
78, 6774-6782.

Edelmann, A.; Diewok, J.; Baena, J. R.; Lendl, B. Anal. Bioanal. Chem.
2003, 376, 92-97.

Kotani, A.; Miyaguchi, Y.; Tomita. E.; Takamura, K.; Kusu, F. J. Agric.
Food Chem. 2004, 52, 1440-1444.

Helaleh, M. l. H.; Tanaka, K.; Taoda, H.; Hu, W.; Hasebe, K.; Haddad, P.
R. J. Chromatogr. A 2002, 956, 201-208.

Palma, M.; Barroso, C. G. Anal. Chim. Acta 2002, 458, 119-130.

Poels, l.; Nagels, L. J.; Verreyt, G.; Geise, H. J. Anal. Chim. Acta 1998,
370, 105-113.

173

18.

19.
20.

21.

22.

23.
24.

25.
26.

27.
28.

29.

30.

31.

32.
33.

34.

35.

Castro, R.; Moreno, M.; Natera, R.; Garcia-Rowe, F.; Hernandez, M.;
Barroso, C. Chromatographia 2002, 56, 57-61.

Zotou, A.; Loukou, Z.; Karava, O. Chromatographia 2004, 60, 39-44.

Perez-Ruiz, T.; Martinez-Lozano, C.; Tomas, V.; Martin, J. J.
Chromatogr. A 2004, 1026, 57-64.

Kerem, Z.; Bravdo, B.-a.; Shoseyov, O.; Tugendhaft, Y. J. Chromatogr.
A 2004, 1052, 211-215.

de Villiers, A.; Lynen, F.; Crouch, A.; Sandra, P. Chromatographia 2004,
59, 403-409.

Soga, T.; Ross, G. A. J. Chromatogr. A 1999, 837, 231-239.

Klampfl, C. W.; Katzmayr, M. U.; Buchberger, W. Electrophoresis 1998,
19, 2459-2464.

Xiong, X.; Li, S. F. Y. J. Chromatogr. A 1998, 822, 125-136.

Kaniansky, D.; Masar, M.; Madajova, V.; Marak, J. J. Chromatogr. A
1994, 677, 179-185.

Kenney, B. F. J. Chromatogr. 1991, 546, 423-430.

Arellano, M.; Andrianary, J.; Dedieu, F.; Couderc, F.; Puig, P. J.
Chromatogr. A 1997, 765, 321-328.

O'Flaherty, B.; Yang, W.-P.; Sengupta, S.; Cholli, A. L. Food Chem.
2001, 74,111-118.

Castineira, A.; Pena, R. M.; Herrero, C.; Garcia-Martin, S. J. High
Resolut. Chromatogr. 2000, 23, 647-652.

Castineira, A.; Pena, R. M.; Herrero, C.; Garcia-Martin, S. J. Food
Compos. Anal. 2002, 15, 319-331.

Nutku, M. S.; Erim, F. B. J. Microcolumn Sep. 1999, 11, 541-543.

Huang, X.; Luckey, J. A.; Gordon, M. J.; Zare, R. N. Anal. Chem. 1989,
61, 766-770.

Sjoback, R.; Nygren, J.; Kubista, M. Spectrochim. Acta, PartA 1995, 51,
L7-L21.

Margulies, D.; Melman, G.; Shanzer, A. Nat. Mater. 2005, 4, 768-771.

174

36.

37.

Goodpaster, J. V.; McGuffin, V. L. Appl. Spectrosc. 1999, 53, 1000-
1008.

Bidlingmeyer, B. A.; Broske, A. D. J. Chromatogr. Sci. 2004, 42, 100-
106.

175

CHAPTER 5

pH-ENHANCED FLUORESCENCE DETECTION OF
y-HYDROXYBUTYRIC ACID IN BEVERAGES

Because of the short half-life of y-hydroxybutyric acid (GHB) in
physiological fluids, an abandoned beverage may be the best source of forensic
evidence for the identification of this drug in a sexual assault case. This study
presents a novel method for detection of GHB in beverages. The proposed
method utilizes separation by liquid chromatography with detection based on the
pH-sensitive fluorescence of fluorescein, as described in Chapter 4. This method
is sensitive only to acids and, therefore, is a vast improvement over UV-
absorbance methods in highly colored, complex beverage matrices. Several
alcoholic and non-alcoholic beverages are spiked with GHB and analyzed using
the proposed method. Under the conditions used in this study, the limit of
detection is 3 mM GHB in beer, red wine, rum, cola, and coffee with a linear
range of 1 — 100 mM GHB. This detection limit corresponds to a dose of 90 mg
of GHB in an 8-oz beverage, which is an order of magnitude below the
therapeutic dose and 20 — 40 times below the typical recreational dose. A dose
of GHB administered with malicious intent may further exceed these levels.
Variation in GHB retention time is observed for beverages with higher alcohol
content (e.g., rum), but identification can be readily confirmed by coinjection of a
GHB standard. Using the proposed method, no selective interconversion of GHB
and y-butyrolactone (GBL) is observed, and in samples containing both

compounds, GBL is not detected.

176

5.1 Introduction and Background
5.1.1 GHB

y-Hydroxybutyric acid (GHB, Figure 5.1) is a minor metabolite of y-
aminobutyric acid (GABA), a neurotransmitter that inhibits uptake of dopamine.
At endogenous levels, GHB has only a mild affinity for GABA receptors, but at
therapeutic and recreational doses, binding increases dramatically. As a central
nervous system depressant, GHB can produce stimulatory effects at low doses
(1 — 2 g) and sedative effects at higher doses.M Clinically, GHB is used to treat
narcolepsy, for weight control, as an anesthetic, to induce amnesia, and has
been considered for treatment of alcohol and opioid addiction?57 The use of
GHB has also been promoted in the body-building community as a tool for weight
loss and muscle gain. In 1990, GHB was banned by the US. Food and Drug
Administration due to serious side effects.8 Some of these side effects include
headache, dizziness, disorientation, seizures, nausea, vomiting, respiratory
depression, and coma.”10 In 2000, GHB was added to the list of Schedule I
controlled substances,11 and the main GHB precursor, y-butyrolactone or GBL,
became a List I chemical. Currently, the only legal form of GHB is a prescription

drug (Xyrem®) used to treat narcolepsy.12
5.1.2 Illicit GHB and Forensic Implications

Illicit GHB is available under street names of Liquid Ecstasy, Liquid X,
Grievous Bodily Harm, Georgia Home Boy, G, and Scoop,7'13 and abuse as a

recreational drug has increased in recent years. Doses of 3 — 4 g GHB can

177

O O
/\/\/u\ [%0
HZN OH
2 3

/\/\/OH O
HO
4 5

Figure 5.1. Structures of GHB and related compounds. y—hydroxybutyric acid,
GHB (1), y-aminobutyric acid, GABA (2), y—butyrolactone, GBL (3), 1,4-butanediol,
1,4-BD (4), y-valerolactone, GVL (5).

178

cause euphoria, reduced inhibitions, and rapid loss of consciousness. Because
GHB is odorless and tasteless, increased use of GHB in sexual assault cases
has been reported.2'3"°""'16 In addition, the onset of effects is extremely rapid,
resulting in loss of consciousness within 10 - 30 min.7 Analogs of GHB (Figure
5.1) such as GBL (Renewtrient, Revivarant, Blue Nitro), y—hydroxyvaleric acid
(GVL), and 1,4-butanediol (1 ,4-BD, FX, Thunder Nectar, Zen) are also commonly
abused, as in vivo conversion to GHB produces similar effects to direct
consumption of GHB.7"7'18 Typically, illicit GHB is manufactured clandestinely
from GBL, which can be purchased legally as an industrial solvent or illegally as
a health food product.13 Sodium or potassium hydroxide is then used to
hydrolyze GBL to GHB in aqueous solution.19 The GHB product may be isolated
as a powder, partially dried to a paste or wet mass, concentrated, or left in
solution. Prior to consumption, the GHB product is dissolved or further diluted in
aqueous media such as beverages.

After consumption, the half-life of GHB in physiological fluids is
extremely short.20 Peak plasma concentrations are observed within 20 — 40 min,
and peak urine concentrations are observed within 4 h."12 Of the original GHB
dose, only about 5% is excreted in urine and is undetectable after 12 hours.12
Because GHB can induce amnesia and disorientation, many victims of sexual
assault may not be aware of the urgency required for determination of GHB in
bodily fluids. As a result of these short physiological timeframes, the ability to
detect GHB after an extended time in beverages will improve evidentiary

procedures related to sexual assault cases.

179

The potential for interconversion of GHB and GBL between manufacture
and consumption has also been a legal concern. Although the overall effect of
consumption is the same, a legal distinction exists between possession and
consumption of GHB (a Schedule I controlled substance) and GBL (a List I
chemical). As a result, any proposed detection methods must distinguish
between GHB and GBL, although no method can overcome the problem of
interconversion prior to analysis. The stability of aqueous GHB and GBL
solutions has been studied as a function of pH and temperature, and a natural
equilibrium of GBLzGHB (2:1) has been reported in pure water.19 Hydrolysis of
GBL to GHB is slowest in pure water and in slightly acidic solution (pH 4.0 — 6.5).
The rate of hydrolysis increases with decreasing pH and increasing temperature.
In basic solution, however, GBL is completely converted to GHB within 15 min.19
Conversely, GHB is stable for over 200 days in slightly acidic solution, and
undergoes esterification to GBL only at extremely low pH. Commercial GBL
products have been found to contain ~30% GHB, consistent with the natural

equilibrium. In beverages spiked with GBL, GHB is detectable after 1 — 3 days.19
5.1.3 Determination of GHB

Preliminary methods for GHB screening are based on colorimetric

21-23 24,25

tests, microcrystal tests, and enzyme-based assays.26 Because the
beverage matrices in which GHB is administered are complex and often colored,
determination based on simple color tests may be difficult or impossible.25 Fillers
and dyes in adulterated beverages as well as the presence of alcohol may

interfere with the determination of drugs such as GHB.17 Isolation of GHB from

180

aqueous matrices has been accomplished by using liquid-liquid extraction25'27

and solid-phase microextraction.28 In forensic laboratories, the most common
method for GHB determination is via precipitation of the silver salt from the
beverage or solution. This precipitate is then isolated by filtration and identified
by infrared (IR) spectroscopy.192752330 Although this method is widely available
and effective for identification of GHB, quantitation by IR spectroscopy is difficult.
Quantitation is important as the amount of GHB present in a beverage may
discriminate between recreational ingestion and malicious administration of GHB.

In order to quantitate the amount of GHB present in a sample, a
chromatographic method is often necessary to separate GHB from other
compounds. Separation methods for GHB analysis in aqueous beverage
matrices include gas chromatography,1927283"33 liquid chromatography,1929'34
and micellar electrokinetic chromatography.”35 Detection of GHB has been

17,19,27,29

achieved by using UV-visible absorbance, indirect UV absorbance,35

19,27,28,33,34

mass spectrometry, ion mobility spectrometry,” nuclear magnetic

”'37 and flame ionization detection.33 Some of these methods rely on

resonance,
conversion of GHB to GBL prior to analysis, which may complicate the legal
interpretation of the result.”38

Of these methods, only a limited number involve separation of GHB from
beverages matrices.17'19'33 In the current work, a liquid chromatographic method
for separation of GHB in beverages is presented together with a novel detection

method based on pH-dependent fluorescence. Beverage components and GHB

are separated by reversed-phase liquid chromatography using an unbuffered

181

acidic mobile phase, as described in Chapter 4. After separation, the pH is
adjusted to a more neutral value and fluorescein is introduced post-column.
When no acids are present, a constant fluorescence background of fluorescein is
observed. In the presence of GHB or other acids, however, the fluorescence is
decreased below the background level to yield a negative signal. This method
facilitates the determination of GHB in beverages by simplifying the

chromatograms of complex beverage matrices by detecting only acidic species.
5.2 Experimental Methods
5.2.1 Sample Preparation

Alcoholic and non-alcoholic beverage samples are purchased from local
supermarkets and convenience stores (Table 5.1). Ground coffee is purchased
from a local supermarket and brewed according to manufacturer’s specifications.
Carbonated beverages are ultrasonicated for 10 min to remove dissolved carbon
dioxide, and all beverages are gravity filtered using filter paper (Whatman) to
remove particulates. Sodium nitrate (JT Baker) is used as a void marker for UV-
absorbance detection and is added to all solutions in a final concentration of 0.1
mM. Beverages are spiked with GHB (Aldrich) at concentrations determined by
using typical serving sizes together with a typical dose of 0.75 9 (Table 5.1). For
construction of the calibration curves, solutions containing 1 — 100 mM GHB are
prepared in distilled deionized water (Corning Glass Works, model MP-3A). For
GBL conversion studies, solutions containing 50 mM each of GHB and GBL
(Aldrich) are prepared in distilled deionized water. Reagent-grade acetic acid

(Aldrich), citric acid (MCB Chemical), formic acid (JT Baker), lactic acid (JT

182

Table 5.1. Summary of alcoholic and non-alcoholic beverages used in this study.

 

 

Alcohol Serving size [GHB] in 1-3 9
Beverage content (%) oz mL dose (mM)
Beer 4.2 12 360 27 — 82
Red wine 9 6 180 54 — 162
Rum 35 1.5 45 217 — 650
Cola 0 12 360 27 - 82
Coffee 0 8 240 41 - 122

 

183

Baker), malic acid (JT Baker), oxalic acid (JT Baker), succinic acid (Sigma), and
tartaric acid (JT Baker) are used to prepare standard solutions at 25 mM
concentration in distilled deionized water. Fresh solutions are prepared daily,

and stock solutions are stored in the refrigerator to prevent degradation.
5.2.2 Experimental System

A system has been designed for the analysis of acidic species in
beverages, and has been described in the previous chapter (Section 4.2.2,
Figure 4.3). The mobile phase for separation of GHB in beverages consists of
1% methanol (Honeywell Burdick and Jackson) in distilled deionized water,
adjusted to pH 3 by using hydrochloric acid. Post-column, the pH is adjusted to
8.0 by addition of 10 mM sodium hydroxide and fluorescein is introduced from a
5 mM solution. Chromatographic detection is achieved by using a laser-induced
fluorescence spectrometer (kg = 325 nm, AEM = 510 nm) and UV-absorbance
detector (M33 = 215 nm). All other experimental parameters are as described in

Chapter 4.
5.3 Results and Discussion
5.3.1 GHB Calibration Curve

As discussed in Chapter 4, the decrease in fluorescence is related to the
number of protons transferred from GHB to fluorescein. Because GHB is a
monoprotic acid, the concentration of GHB in the sample is equal to the
concentration of protons available for donation to fluorescein. The behavior of

fluorescein and GHB is inherently similar to the titration of a weak base with a

184

weak acid. In this study, calibration curves are generated by injection of GHB
solutions of varying concentration from 1 — 100 mM. The chromatographic peak
height or area measurements are fit by nonlinear regression to a sigmoidal
function (Table Curve 2D, Jandel Scientific, version 2.02). A graph is then
prepared of the sigmoidal function of peak height or area versus concentration
(Equation 4.1) to yield a linear calibration curve, as shown in Figure 5.2. The
correlation coefficient for this curve is related to the quality of sigmoidal fit and is
greater than 0.994 for both peak height and area. This indicates that the
concentration of GHB in an unknown sample can be determined with good
accuracy and precision.

From the calibration curve, the limit of detection of GHB is determined to
be 3 mM in aqueous solution (S/N = 3) and the linear dynamic range is two
orders of magnitude. The detection limit and linear range are sufficient for
determination of GHB in adulterated beverages, as demonstrated by the typical

concentration ranges in Table 5.1.
5.3.2 Beverage Analysis

For analysis of GHB in beverages, determination of potential
interferences is critical. The most probable interferences are low-molecular-
weight organic acids such as formic, acetic, lactic, and citric acids. Using the
method described in Chapter 4, eight common organic acids are successfully
resolved from GHB by using a mobile-phase composition of 99:1 water:methanol
adjusted to pH 3 (Figure 5.3). The retention factor for each acid, summarized in

Table 4.2, is compared with those in the beverage samples in order to identify

185

.0.20E0.0000E 00.2. .0 20..0.>00 0.0020.0 2000.00. 0.00 .020 8600.0 u N0. 00000 I .6060. F- u >
”00.0 2000 .0. 20.000.00. .002... A F0000 u N0. 0030 I xFNF0. F- u > 020.0... 2000 .0. 20.000.00. .002...
..o. 00.0 2000 020 .0. 0.0.0.. .000 02.00 >0 00.0.0200 ..0.02. 2. 00.0000E 0:0 .0. 0200 2000.0..00 .00 0.00.0

...... 2925200200 020

 

Tow Wow flow YOF
_ _ moo
i. S
OFS
W
nlu
Tméa
m
-0.NN
S
....
w
imNW

 

 

 

0.0

186

..0. 0.00 0.2.0000 .8. 00.0 .8. 0.00 0....0 .8. 0.00 00000 .8. 0.00 0000. ..6. 0.00 0._0E .8.

0.00 0.220. .8. 0.00 02020. ..F. 0.00 0..0x0 ”00.0.00 .E2 0FN ”20000.00 00200.0000->3 .00000200 E2-0F .20.00.E0
E2-0F0 .20..0..0x0 E2000 ”20000.00 82800.02. 080020.800 .2.E.._E 0.F ”0.0. 2.0.0 .0 ...0 ..0202.0E..0.02.
F00 ”00020 0:00.). .EE 06 x EE 000 00.00.00 E00 .000 ..0.00>... .2E0.00 .8. 20000.00 82800.00... 000002.
-.000. 020 2. 20000.00 00200.0000->: >0 002.0.00 010 020 00.00 0.2000 20EE00 .0 E0.00.0E0.2o .00 0.00.0

.22... 0.2:
N 0 0 0 0 N .

_ _ _ _ _

 

 

 

 

 

 

 

ESNOdSEH CISZI'IVINHON

 

 

 

 

187

the components.
5.3.2.1 GHB in Alcoholic Beverages

Alcoholic beverages such as beer, red wine, and rum are spiked with
GHB as described in Table 5.1, and the resulting chromatograms are shown in
Figures 5.4-5.6. For all of the alcoholic beverages, the chromatograms with
fluorescence detection are greatly simplified with respect to those with UV-
absorbance detection. The UV-absorbance method responds to any compound
that absorbs 215-hm light, yielding a complex chromatogram with many peaks
that can interfere with GHB determination. The fluorescence method, however,
responds only to acids. For selective determination of GHB in beverages, the
fluorescence method eliminates potential interferences from non-acidic absorbing
species such as fermentation products. By comparison to the chromatogram of
standard acids (Figure 5.3), oxalic and phosphoric acids are identified in the beer
sample (Figure 5.4). Oxalic, tartaric, malic, and lactic acids are identified in the
red wine sample (Figure 5.5), whereas oxalic and lactic acids are identified in the
rum sample (Figure 5.6). In all of the alcoholic beverages, the detection limit of
GHB is determined to be 3 mM (S/N = 3). As mentioned previously, this
detection limit is sufficient for determination of GHB in adulterated beverages

(Table 5.1).
5.3.2.2 GHB in Non-Alcoholic Beverages

Non-alcoholic beverages such as cola and coffee are also spiked with

GHB and analyzed by using the current method (Figures 5.7 and 5.8). As

188

 

..0. 020 ..N. 200 2.2.82.0 .. 0 200 2.0.8 00.000
.00 0.00.0 2. 00020000 00 0200.028 0.200.00.0E0.20 .8. 20000.00 82800.00: 080020.000. 020 2. 20000.00
00200.0000->: >0 002.0.00 0:0 0 00.0 ....2. 002.00 .000 .0 02.200 ...E-000. No.NF 0 .0 E0.00.0E0.2O 6.... 0.00.0

.22... 0.2:
N 0 0 0 0 N .

_ _ _ _ _

 

ESNOdSHH OEZI'IVINHON

 

 

 

 

 

189

.8. 0:0 .6. 2205.2: .E 28 2.8. .8. 28 2.2.. .3. 28 2.28. .2 c 28 2.98 8.200
.00 0.00.". 2_ 00220000 00 020_._0200 038.00.080.20 .3. 22.00.00 002800.02. 080020.002 020 2. 22.00.00
00202502053 >2 0050.20 9.0 a mud 2.25 09:00 052, 00. .0 023.00 32.62: No.0 0 .0 E0.00.0—20.20 dd 0.59“.

22.5.. 05.:
2 0 0 v 0 m .

_ — — _ —

 

 

<1:
asuoasaa aaznkuon

 

 

 

 

 

190

.8. 0:0 .20. 28 2.8. .. 0 28 2.98 00.200
.04... 0.00.... 2. 00220000 00 0202.028 0.200.020.2220 Am. 20200.00 002800.00: 080022.000. 020 2. 20200.00
00202.0020->D >2 002.0.20 mIO 0 mwd 2...... 002.00 2.0. .0 02.200 32..va No.0... 0 .0 60.00.028.20 6.0 0.00.“.

.220 ms...
2 0 w 0 N .

p r _ _ —

LO

 

 

<
asuoasau aaznkuou

 

 

 

 

 

 

 

191

0.010 .E 28 82002.0 00.200
.00 0.00.... 2. 00220000 00 0202.0200 0.200.00.02.0.20 .3. 20200.00 002800.002 08002...000_ 020 2. 20200.00
8202.0020->3 >2 002.0.20 9.20 0 02.0 2.25 002.00 000 .0 02.200 32.000. N0.N_. 0 .0 20.00.025.20 ....m 0.00.“.

22.5.. 0.2:
2 0 0 0 0 N .

 

 

 

<E
SSNOdSEH GEZI'IVWHON

 

 

 

 

 

 

192

.00 0:0 .8. 23022.... ..N. 28 25.0. .. 2 28 2.0.6 00.200
0.0 0.00.“. 2. 00220000 00 0202.0200 0.200.00.0E0.20 .8. 20200.00 82800.00: 08002...000_ 020 .0.. 20200.00
00202.0020->3 >2 002.0.20 $5 0 02.0 2.25 002.00 00200 .0 02.200 3.2-023 No.2 20 .0 20.00.025.20 .0.... 0.00.“.

.22... 05.:
2 0 0 0 0 N .

_ _ _ _ _

 

<
EISNOdSSH aaznvme

 

 

 

 

 

193

observed previously for the alcoholic beverages, the chromatograms of the non-
alcoholic beverages with fluorescence detection are greatly simplified compared
to those with UV-absorbance detection. Simplification of the cola chromatogram
is most impressive at retention times less than 3 min, where a number of non-
acidic species are observed with UV-absorbance detection but not with
fluorescence (Figure 5.7). These species may include artificial sweeteners,
preservatives, and coloring agents. In the coffee chromatogram, the
fluorescence method eliminates potential overlap between an unknown
absorbing compound and GHB (Figure 5.8). Because this absorbing species is
not acidic, no interference is observed with fluorescence detection. The cola and
coffee samples, both dark brown in color, have the greatest potential for
interferences when using UV-absorbance detection. The ability to determine
GHB in these matrices by using the fluorescence method will be a great utility to
forensic scientists. By comparison to the chromatogram of standard acids
(Figure 5.3), phosphoric acid is identified in the cola sample (Figure 5.7) and
oxalic and formic acids are identified in the coffee sample (Figure 5.8). Under
these experimental conditions, the detection limit of GHB in non-alcoholic
beverages is also determined to be 3 mM in the non-alcoholic beverages (S/N =
3). As mentioned previously, this detection limit is sufficient for determination of

GHB in adulterated beverages (Table 5.1).
5.3.2.3 Effect of Ethanol on GHB Retention

For qualitative identification of GHB, the retention factor is compared to a

known value. The precision of the retention factor, therefore, will determine the

194

reliability of the method for identification of GHB in various unknown samples.
The retention factor for GHB is constant at 2.1 in water, beer, wine, cola, and
coffee, while the retention factor for GHB is 1.9 in rum. Because
chromatographic retention is highly dependent on the relative polarity of the
mobile and stationary phases, the decrease is most likely caused by the
increased alcohol content of rum (35%) compared to the other beverages (<10%).
In the rum sample, the decreased polarity causes GHB to spend less time in the
stationary phase, relative to the other samples. Because less time is spent in the
stationary phase, GHB elutes more rapidly. To confirm this theory, a 50 mM
solution of GHB is prepared in 35:65 ethanolzwater as a simulant of an alcoholic
beverage. Upon injection of the simulant, a decreased retention factor (1.9) is
observed for GHB. To avoid this effect, it is often recommended to dissolve
samples in the mobile phase prior to injection in liquid chromatography. For the
direct analysis of beverages, however, this approach is undesirable.

Another approach to confirm the presence of GHB is by coinjection of a
standard. If an alcoholic beverage is suspected to contain GHB but the retention
time is not statistically consistent, a small volume of a standard GHB solution
may be added and the beverage reanalyzed. The presence of GHB is confirmed
if a single peak is observed, where the retention factor remains constant and the
peak height and area increase. For analysis of unknown beverages, the
coinjection method is advantageous because a priori knowledge of the alcohol

content is not necessary.

195

5.3.3 lnterconversion of GHB to GBL

Because GHB is more strongly regulated, the quantitation of GHB in a
suspect beverage, without interconversion to GBL, is the most important
determination to be made in a forensic case. As a result of these legal
implications, it is necessary to validate the current method to ensure that
selective interconversion of GHB and GBL is not promoted during sample
analysis. A solution of 50 mM each GHB and GBL is analyzed using both UV-
absorbance detection and the pH-dependent fluorescence detection method
(Figure 5.9). Injections of individual standards are used to identify the first peak
as GHB and the second peak as GBL. No GBL peak is observed for injection of
a GHB solution, indicating that the method does not cause esterification of GHB
to GBL. Also, no GHB peak is observed for injection of a GBL solution, indicating
that the method does not cause hydrolysis of GBL to GHB. Because the indirect

fluorescence method only detects acids, no peak is observed for GBL.
5.4 Conclusions

The complex matrices of alcoholic and non-alcoholic beverages can
obscure the forensic analysis of beverages adulterated with GHB. The proposed
method eliminates interferences and allows quantitation of GHB in adulterated
beverages with a detection limit of 3 mM and a linear range of 1 - 100 mM. This
detection limit corresponds to a dose of 0.09 g of GHB in an 8-oz beverage,
which is an order of magnitude below the therapeutic dose and 20 - 40 times
below the typical recreational dose. A dose of GHB administered with malicious

intent may further exceed these levels. The sensitivity of the method indicates

196

 

.00 0.00.“. 2. 00220000 00 0202.0200 0.200.00.0E0.20 .3. 20200.00 002800.002 000002.
-.000_ 020 0.2. 20200.00 00202502053 >2 002.050 0.0.... 2. ..m0 020 9.20 2000 EE cm .0 E0.00.0E0.20 .0.... 0.50.".

 

 

.z__.._. 0...:
o. 0 0 0 N
N
o
H
w
0 w
.4.
3
it 0
< m...
S
m
.50 N
S
3
0:0

 

 

 

197

that GHB may be detectable in trace amounts of a beverage left behind.
lnterconversion to GBL does not occur during sample analysis, and GBL is not
detected using the proposed method. The ability to detect GHB in abandoned
beverages may provide important forensic evidence in sexual assault cases in

which biological evidence of GHB consumption is difficult to obtain.

198

5.5

10.
11.

12.

13.

14.
15.
16.
17.

18.

References

Craig, K.; Gomez, H. F.; McManus, J. L.; Bania, T. C. J. Emerg. Med.
2000. 18, 65-70.

Galloway, G. P.; Frederick, S. L.; Staggers Jr, F. E.; Gonzales, M.;
Stalcup, S. A.; Smith, D. E. Addiction 1997, 92, 89-96.

Tunnicliff, G. Journal of Toxicology: Clinical Toxicology 1997, 35, 581-
590.

Ropero-Miller, J. D.; Goldberger, B. A. Clin. Lab. Med. 1998, 18, 727-
746.

Gallimberti, L.; Cibin, M.; Pagnin, P.; Sabbion, R.; Pani, P. P.; Piratsu,
R.; Ferrara, S. D.; Gessa, G. L. Neuropsychopharmacol. 1993, 9, 77-81.

Caputo, F.; Addolorato, G.; Trevisani, F.; Bemardi, M. Lancet 2005, 366,
981-982.

Couper, F. J.; Marinetti, L. J. Forensic Sci. Rev. 2002, 14, 101-121.
Centers for Disease Control J. Am. Med. Assoc. 1991, 256, 447-448.

Chin, M.-Y.; Kreutzer, R. A.; Dyer, J. E. West. J. Med. 1992, 156, 380-
384.

Kam, P. C. A.; Yoong, F. F. Y. Anaesthesia 1998, 53, 1195-1198.

Drug Enforcement Administration 2000, Addition of gamma-
hydroxybutyrate (GHB) to Schedule I. 65 (49), 13235-13238.

Haller, C.; Thai, 0.; Jacob, P.; Dyer, J. E. J. Anal. Toxicol. 2006, 30, 360-
364.

Del Signore, A.; McGregor, M.; Cho, B. P. J. Forensic Sci. 2005, 50, 81-
86.

Couper, F. J. ; Logan, B. K. J. Forensic Sci. 2001, 46, 919-923.
ElSohly, M. A.; Salamone, S. J. J. Anal. Toxicol. 1999, 23, 141-146.
Wells, D. Sci. Justice 2001, 41, 197-199.

Bishop, S. C.; Lerch, M.; McCord, B. R. Forensic Sci. Int. 2004, 141, 7-
15.

Zvosec, D. L.; Smith, S. W.; McCutcheon, J. R.; Spillane, J.; Hall, B. J.;
Peacock, E. A. New Engl. J. Med. 2001, 344, 87-94.

199

19.

20.

21.
22.
23.
24.

25.

26.

27.

28.

29.

30.
31.

32.
33.
34.
35.
36.

37.
38.

Ciolino, L. A.; Mesmer, M. Z.; Satzger, R. D.; Machal, A. C.; McCauley,
H. A.; Mohrhaus, A. S. J. Forensic Sci. 2001, 46, 1315-1323.

Brenneisen, R.; ElSohly, M. A.; Murphy, T. P.; Passarelli, J.; Russmann,
S.; Salamone, S. J.; Watson, D. E. J. Anal. Toxicol. 2004, 28, 625-630.

Alston, W. C.; Ng, K. Forensic Sci. Int. 2002, 126, 114-117.
Badcock, N. R.; Zotti, R. Ther. Drug Mon. 1999, 21, 376.
Koppenhaver, D. J. Microgram1997, 30, 130.

Andera, K. M.; Evans, H. K.; Wojcik, C. M. J. Forensic Sci. 1999, 45,
665-668.

Bell, S. C.; Oldfield, L. S.; Shakleya, D. M.; Petersen, J. L.; Mercer, J. W.
J. Forensic Sci. 2006, 51, 808-811.

Bravo, D. T.; Harris, D. 0.; Parsons, S. M. J. Forensic Sci. 2004, 49,
379-387.

Chappell, J. S.; Meyn, A. W.; Ngim, K. K. J. Forensic Sci. 2004, 49, 52-
59. .

Blair, 8.; Song, M.; Hall, 3.; Brodbelt, J. J. Forensic Sci. 2001, 46, 688-
693.

DeFrancesco, J. V.; Witkowski, M. R.; Ciolino, L. A. J. Forensic Sci.
2006, 51 , 321-329.

Couper, F. J.; Logan, B. K. J. Anal. Toxicol. 2000, 24, 1-7.

Ferrara, S. D.; Tedeschi, L.; Frison, G.; Rossi, A. J. Forensic Sci. 1995,
40, 501-504.

Elian, A. A. Forensic Sci. Int. 2000, 109, 183-187.

Elliott, S.; Burgess, V. Forensic Sci. Int. 2005, 151, 289-292.
Mesmer, M. Z.; Satzger, R. D. J. Forensic Sci. 1998, 43, 489-492.
Dahlen, J.; Vriesman, T. Forensic Sci. Int. 2002, 125, 113-119.

Patchett, M. L.; Minoshima, Y.; Harrington, P. B. Spectroscopy 2002, 17,
16-21 .

Chew, S. L.; Meyers, J. A. J. Forensic Sci. 2003, 48, 292-298.
Lettieri, J. T.; Fung, H. L. Biochem. Med. 1978, 20, 70-80.

200

CHAPTER 6

CONCLUSIONS AND FUTURE WORK

Luminescence-based methods are of great importance in the search for
a sensitive and selective detector for explosives and other forensically important
compounds. By gaining a more detailed understanding of the mechanism of
interaction between explosives and sensor materials, new fluorescent materials
can be developed with an improved ability to detect explosives in complex field
settings. This insight may also lead to identification of fluorescent materials that
could be useful in determination of other forensically relevant compounds. The
focus of this research has been the identification of new fluorescence quenching
interactions that can be used to advance the field of forensic science.

The fluorescence quenching response of several solution-phase
fluorophores in the presence of nitrated explosives has been investigated. Of the
extended group of fluorophores investigated in this study, pyrene, purpurin,
malachite green, and phenol red are most promising for explosives detection.
These fluorophores demonstrate sensitive and selective quenching interactions
with a target group of nitrated quenchers, with quenching constants greater than
50 M". The remaining fluorophores under investigation demonstrate little or no
response to nitrated compounds, with quenching constants less than 50 M".
Because each fluorophore may interact with the quenchers through a different
mechanism, various degrees of sensitivity and selectivity are observed for
pyrene, purpurin, malachite green, and phenol red. Of these, pyrene

demonstrates the greatest sensitivity for nitrated molecules with quenching

201

constants on the order of 102 M". The routine use of pyrene for explosives
detection is undesirable, however, because of numerous health and waste
disposal hazards. Purpurin, malachite green, and phenol red demonstrate less
sensitivity than pyrene but comparable or greater selectivity for nitrated
molecules. Both malachite green and phenol red exhibit pH-dependent
fluorescence, but with buffering, the interference of acidic species on the
quenching of malachite green fluorescence can be reduced. In general, pH-
sensitive fluorophores are not ideal for application in the detection of explosives
on-site, as various matrices may be encountered that could lead to erroneous
results. Thus, among these potential fluorophores, purpurin shows the most
promise for detection of nitrated explosives by fluorescence quenching.

Based on these solution-phase quenching results, the fluorescence of
purpurin, malachite green, and phenol red is studied in more detail. Initial studies
with a variety of solvents indicate that the fluorescence power of malachite green
and phenol red is greatest in dioxane, while that of purpurin is greatest in more
polar solvents such as methanol or water. The most promising substrate for
incorporation of malachite green while maintaining fluorescence is poly(ethylene
glycol). A mixture containing malachite green and poly(ethylene glycol)
fluoresces in the solid state when coated on glass slides, filter paper, or silica
particles. The use of a high-power source such as a laser causes degradation of
malachite green fluorescence, however, indicating that a low-power source, such
as a handheld UV lamp, is better for field applications. Upon further testing, the

polymer containing malachite green can be coated on a paper-based substrate

202

for use as a wipe for detection of explosives on surfaces such as luggage or
cargo. In addition, an air-sampling device could be developed based on coated
silica particles enclosed in a quartz tube. For both of these applications, the use
of available equipment such as handheld UV lamps and digital cameras is
appropriate.

Future work in the development of sensors for explosives detection
involves investigations of other solution-phase fluorophores to identify those that
exhibit sensitive and selective quenching by nitrated compounds. In this work,
only a limited number of potential fluorophores have been selected on the basis
of their structural similarities to pyrene. From the results of this study, however, it
is evident that mechanistic similarities do not follow directly from structural
similarities. Consequently, a more broad and diverse selection of fluorophores
may prove essential for the development of a fluorescence quenching method for
the detection of nitrated explosives. For phenol red and purpurin, fluorophores
identified in the original study, parallel studies to those conducted for malachite
green are necessary to determine appropriate matrix polarity and verify
fluorescence in the solid state. In addition, significant testing of any solid-state
device with explosive particles and vapors is necessary to verify that the
quenching behavior observed in the solution phase is also observed in the
polymer or other matrices. Use of malachite green in the solution phase is still
viable, and investigations into a post-column liquid chromatography system could

lead to a selective explosives detection method. For all methods, the detection

203

limits of the explosives must be determined and optimized before the true utility
can be known.

In a second application of luminescence-based methods for forensic
sample determination, a novel method for detection of acidic species based on
the pH-sensitive fluorescence of fluorescein has been developed. The native
acid-base properties of fluorescein and the corresponding changes in
fluorescence quantum yield are exploited for detection of acidic species in foods
and beverages. Any compound that is more acidic than fluorescein allows
sufficient proton transfer in aqueous and polar organic solvents and can thereby
be detected. The fluorescence of fluorescein is sigmoidal with the logarithm of
proton concentration. The linear range extends two orders of magnitude and is
centered about the fluorescein concentration in solution. The linear range can be
extended to four orders of magnitude using a sigmoidal transform function, and
can be shifted to higher or lower concentrations by adjustment of the fluorescein
concentration. This new method is verified by using a series of test acids, and
applied to the analysis of grape juice and derivatives, including a variety of wine
and vinegar samples following liquid chromatography. Improvements in
detection sensitivity and selectivity compared to UV-absorbance detection are
observed for acids found in these samples. This fluorescence method provides a
new tool for analysis of acidic species with potential use in a variety of disciplines
including forensic, environmental, and clinical analyses.

A forensic application of this acid detection method is in the

determination of y-hydroxybutryic acid (GHB), a drug commonly used in sexual

204

assaults. GHB is often administered in complex alcoholic and non-alcoholic
beverages, complicating the forensic analysis of adulterated beverages. The
proposed method eliminates interferences from other beverage components and
allows quantitation of GHB with a detection limit of 3 mM and a linear range of 1
— 100 mM. This detection limit corresponds to a dose of 90 mg of GHB in an 8-
oz beverage, which is an order of magnitude below the therapeutic dose and 20
— 40 times below the typical recreational dose. A dose of GHB administered with
malicious intent may further exceed these levels. The sensitivity of the method
indicates that GHB may be detectable in beverage residues remaining at the
scene of a crime. In contrast to other GHB detection methods, interconversion to
y—butyrolactone (GBL) does not occur during sample analysis. In addition, GBL is
not detected using the proposed method. The ability to detect GHB in
abandoned beverages may provide important forensic evidence in sexual assault
cases in which biological evidence of GHB consumption is difficult to obtain.

For applications of this method in other areas, the instrument parameters
must be optimized for operation at lower fluorescein concentrations. This
adjustment will allow detection of acids at lower concentrations, as present in
environmental or clinical samples. The most important variables to control are
the relative flow rates of fluorescein and sodium hydroxide. At low fluorescein
concentrations, accurate flow control is critical to maintain a constant
fluorescence background. Because this flow control can be difficult, adaptation
of this method to capillary electrophoresis (CE) may be beneficial. In a CE

method, the acidic species are separated as ions, and fluorescein could be

205

added into the non-buffered separation electrolyte, maintained at pH 8. As with
the liquid chromatography method, acids present in the sample transfer a proton
to fluorescein and the corresponding decrease in fluorescence is detected.
Using the CE method, lower concentrations of fluorescein could be utilized, and
lower acid concentrations could thereby be detected.

Overall, this work demonstrates a variety of analytical and forensic
applications of fluorescence quenching. Each of these techniques demonstrates
the use of luminescence methods in a way that has not been utilized in the
forensic community. In addition, many of these applications can be easily
implemented in a forensic laboratory by utilizing common equipment such as a
handheld lamp, a digital camera, or a conventional fluorescence detector. In all,
these methods provide a number of new tools for the forensic science and

analytical chemistry communities.

206

APPENDIX

INVESTIGATION OF QUENCHING MECHANISM

A.1. Determination of Fluorescence Lifetimes

The fluorescence lifetime of each fluorophore is determined from single-
photon counting data as described in Section 2.2.3. The fluorescence lifetime (1:)

is exponentially related to the photons counted (F) at time t.
F(t)=Foe'%+C (A.1)

The initial number of photons counted (F0), the fluorescence lifetime, and the
offset constant (C) are determined from nonlinear regression of the data. Six
replicate measurements are acquired for each fluorophore in methanol under the
conditions given in Table A.1. The concentrations for each fluorophore are
chosen to give a minimum integration time for single-photon counting.
Wavelengths are determined from the emission spectra determined with
excitation at 325 nm. Fluorescence decay curves for pyrene and purpurin are

given in Figure A.1, and for malachite green and phenol red in Figure A2.
A2. Determination of Redox Potentials

Redox potentials for each fluorophore and quencher are determined
from cyclic voltammetry as described in Section 2.2.4. Potentials are scanned
from 0.0 to +1.5 V at 0.5 V/s for determination of oxidation potentials and from
0.0 to -1.5 V at 0.5 V/s for determination of reduction potentials. For reversible

systems, E», and Ered are calculated as the average of the anodic and cathodic

207

Table A.1. Parameters used for determination of fluorescence lifetimes.

 

 

Fluorophore Concentration Emission Wavelength
Pyrene 10‘4 M 390 nm
Purpurin 10'3 M 413 nm
Malachite green 10'4 M 379 nm
Phenol red 10'3 M 369 nm

 

208

 

COUNTS

 

 

O I I I I

0 10 20 30 4O 50
TIME (ns)

 

1400

 

1200 —
1000 -

1’3 800 -

z

3

0 600 -
400 -

200 -

 

 

O I I I I

0 2 4 6 8 1 0
TIME (ns)
Figure A.1. Fluorescence decay curves for pyrene (top) and purpurin (bottom).

For pyrene (top), F0 = 4477, c = 89.88, 0 = 15.3 ns, R2 = 0.9975. For purpurin
(bottom), F0 = 1007, c = 146.8, 0 = 4.00 ns, R2 = 0.9933.

 

209

 

 

 

 

TIME (ns)

 

1200

1000 -

COUNTS
0 oo
o o
o o

b

O

O
l

200 -

 

 

 

O I I l f

0 2 4 6
TIME (ns)
Figure A.2. Fluorescence decay curves for malachite green (top) and phenol

red (bottom). For malachite green (top), F0 = 2157, C = 38.30, 1' = 1.05 ns, R2 =
0.9887. For phenol red (bottom), F0 = 1554, C = 48.82, 1: = 1.97 ns, R2 = 0.9936.

00

10

210

peak potentials. For irreversible systems, on is reported as the anodic peak
potential and EM is reported as the cathodic peak potential. Cyclic
voltammograms for pyrene, purpurin, malachite green, and phenol red are given
in Figures A.3-A.6. Cyclic voltammograms for nitromethane, nitrobenzene, 4-
nitrotoluene, 2,6—dinitrotoluene, aniline, benzoic acid, and phenol red are given in

Figures A.7-A.13.
A.3. Determination of Singlet Excitation Energies

The singlet excitation energy for each fluorophore is determined from
UV-visible absorbance and fluorescence emission spectra as described in
Section 2.2.5. Spectra are acquired by using traditional spectrophotometers with
1-cm pathlength. The energy of the singlet excited state can be approximated as
the wavelength at which the normalized absorption and emission spectra
intersect. Spectra are normalized to the maximum absorbance or emission
wavelength and overlaid to determine the intersection point. This wavelength is
then converted to energy units of eV for calculation of the free energy of electron
transfer (Equation 2.1). Normalized spectra for pyrene and purpurin are given in

Figure A.14, and for malachite green and phenol red in Figure A15.

211

N
O

 

_x
O
1

CURRENT (pA)

0 r'0 2
O O O O
l l l l

 

 

.40 I r I I

-1.6 -1.2 -0.8 -0.4 o
POTENTIAL (V)
200

 

 

150 -

.0
O
0

CURRENT (0A)
01
O

 

 

 

 

'50 I I I I I

-0.1 0.2 0.5 0.8 1.1 1.4
POTENTIAL (V)

Figure A.3. Cyclic voltammograms for pyrene in 0.1 M NaN03 in methanol, with
potential scanned cathodically (top) and anodically (bottom).

 

212

N
O

 

.0
O O
1 I

PURRPNT (,0)
a

204

-30 -

 

 

'40 I I j I

-1.6 -1.2 -0.8 -0.4
POTENTIAL (V)

 

O

 

00
O
I

CURRENT (0A)
a B

L.
o o
I l

 

 

'20 I I I I

-0.1 0.3 0.7 1.1 1.5
POTENTIAL (V)

Figure A.4. Cyclic voltammograms for purpurin in 0.1 M NaN03 in methanol,
with potential scanned cathodically (top) and anodically (bottom).

 

213

N
O

 

PA)
3

O
1

CURRENT(
8

I'C
O
1

 

 

 

 

-30 I I I T
-1.6 -1.2 -0.4

O

-O.8
POTENTIAL (V)

 

 

CURRENT (“A

 

 

..__ \/—'
-20 -

'40 I I I I

 

 

 

-0.1 0.3 0.7 1.1 1.5
POTENTIAL (V)

Figure A.5. Cyclic voltammograms for malachite green in 0.1 M NaNOa in
methanol, with potential scanned cathodically (top) and anodically (bottom).

214

 

9URRIENT(,,IA)
B 3 o 3

00
O
l

 

 

 

O

-1.6 -1.2 -0.8
POTENTIAL (V)

 

CURRENT (pA)

 

 

 

‘20 I I I I
-o.1 0.3 0.7 1.1 1.5
POTENTIAL (V)

Figure A.6. Cyclic voltammograms for phenol red in 0.1 M NaNOa in methanol,
with potential scanned cathodically (top) and anodically (bottom).

215

01
O

 

 

(IA)
'8 o

-100 -

CURRENT

-150 -

-200 -

 

-250 T I I

 

 

-1.6 -1.1 -O.6 -o.1
POTENTIAL (V)

 

30

—¥ N
O O
1 1

CURRENT (pA)
O

 

.L
O
I

 

 

 

'20 I I I

-o.1 0.4 0.9 1.4
POTENTIAL (V)

Figure A.7. Cyclic voltammograms for nitromethane in 0.1 M NaNOa in
methanol, with potential scanned cathodically (top) and anodically (bottom).

216

 

O)
O

 

 

 

 

 

 

 

 

30 .
2; ° ‘
7: -3o —
2
E:
m -60 -
D
0 -90 -
-120 -
'150 I I I
-1.6 -1.1 -0.6 -o.1
POTENTIAL (V)
80
60 -
:40 .
Z
|.|.|
n:
a: 20 -
D
O
O -
-20 I I I
-o.1 0.4 0.9 1.4
POTENTIAL (V)

Figure A.8. Cyclic voltammograms for nitrobenzene in 0.1 M NaN03 in methanol,
with potential scanned cathodically (top) and anodically (bottom).

217

 

 

200

100)

O
l

-100 -

CURRENT (pA)

-200 -

-300 -

 

 

 

-400 I l I

-1.6 -1.1 -0.6 -o.1
POTENTIAL (V)
40

 

 

30-

20-

CURRENT (pA)
ES

 

_'L
O O
I 1

 

'20 I I I

 

 

-O.1 0.4 0.9 1.4
POTENTIAL (V)

Figure A.9. Cyclic voltammograms for 2-nitrotoluene in 0.1 M NaN03 in
methanol, with potential scanned cathodically (top) and anodically (bottom).

218

150

 

 

 

 

 

 

 

 

50 ..
V -50 -
'—
z
32
m -150 -
a
O
-250 -
-350 . , f
-1.6 -1.1 -0.6 -0.1
POTENTIAL (V)
40
3O -
i 20 -
E
m 10 -
a:
"a‘
o 0 I
-10 ..
-20 I I T
-0.1 0.4 0.9 1.4
POTENTIAL (V)

Figure A.10. Cyclic voltammograms for 2,6-dinitrotoluene in 0.1 M NaN03 in
methanol, with potential scanned cathodically (top) and anodically (bottom).

219

A
O

 

N N
O O O
1 I I

CURRENT (IIA)
<5: 1:.
O 0

do
0
I

-100 -
-120 -

”140 I I I

-1.6 -1.1 -0.6 -o.1
POTENTIAL (V)

 

 

 

 

(D
0

CURRENT (,IA)
I'\> N -b o:
o o o o o

 

 

 

'40 I I I

61 0.4 0.9 1.4
POTENTIAL (V)

Figure A.11. Cyclic voltammograms for aniline in 0.1 M NaNOa in methanol, with
potential scanned cathodically (top) and anodically (bottom).

220

 

CURRENT (,IA)

 

 

 

-140 I I I

-1.6 -1.1 -O.6 -0.1
POTENTIAL (V)

 

40

 

CURRENT (IIA)
L; —x N 00
O O O O O

 

 

 

-20 I I I

-o.1 0.4 0.9 1 .4
POTENTIAL (V)

Figure A.12. Cyclic voltammograms for benzoic acid in 0.1 M NaNO3 in
methanol, with potential scanned cathodically (top) and anodically (bottom).

221

 

N b
O O O
I 1

CURRENT (IIA)
do c'n L lb
0 O O O

-100 -

 

 

 

-120 I I I
-1.6 -1.1 -O.6 -o.1
100 POTENTIAL (V)

 

80-

-§ 03
O O
1 I

CURRENT (IA)
8

 

I'O

O O
l l
l

 

'40 I I I

 

 

-o.1 o.(9 1.4
POTENTIAL V)

Figure A.13. Cyclic voltammograms for phenol in 0.1 M NaN03 in methanol,
with potential scanned cathodically (top) and anodically (bottom).

222

 

0.9 -

0.6 -

0.3 -

NORMALIZED RESPONSE

 

 

0.0 I I
300

 

50 400 450
WAVELENGTH (nm)

 

1.2

0.9 -

0.6 d

0.3 -

NORMALIZED RESPONSE

 

 

0.0 I | I

300 350 400 450 500
WAVELENGTH (nm)

Figure A.14. Normalized absorbance (—) and emission (—) spectra for pyrene
(top) and purpurin (bottom). Emission spectra are collected by using the
wavelength of maximum absorbance as the excitation wavelength. Singlet
excitation energies are determined from the crossing point of the two spectra.

 

223

1.2

 

0.9 1

0.6 -

0.3 -

NORMALIZED RESPONSE

 

 

 

0.0 I I I I

310 330 350 370 390
WAVELENGTH (nm)

1.2

 

NORMALIZED RESPONSE
o o
'O) o:

 

 

0.0 I I

250 300 350 400
WAVELENGTH (nm)

Figure A.15. Normalized absorbance (—) and emission (—) spectra for
malachite green (top) and phenol red (bottom). Emission spectra are collected
by using the wavelength of maximum absorbance as the excitation wavelength.
Singlet excitation energies are determined from the crossing point of the two
spectra.

 

224

 

 

 

llllllllllllllllllllllllllllllllllllllllllllll‘llllllllllil
3 1293 02956 5565