Enhancing the chemiluminescence determination of biologically
and forensically important molecules
by
Jessica Michele Terry
B.FSc. (Hons)
Submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
Deakin University
May, 2012
DEAKIN UNIVERSITY
ACCESS TO THESIS A
I am the author of the thesis entitled:
‘Enhancing the chemiluminescence determination of biologically and
forensically important molecules’
submitted for the degree of Doctor of Philosophy
This thesis may be made available for consultation, loan and limited copying in
accordance with the Copyright Act 1968.
'I certify that I am the student named below and that the information provided in
the form is correct'
Full Name....................................................................................…….......................
(Please Print)
Signed...................................................................................……………..............….
Date......................................................................................……………..............….
i
DEAKIN UNIVERSITY
CANDIDATE DECLARATION
I certify that the thesis entitled:
‘Enhancing the chemiluminescence determination of biologically and
forensically important molecules’
submitted for the degree of Doctor of Philosophy is the result of my own work and
that where reference is made to the work of others, due acknowledgment is given.
I also certify that any material in the thesis which has been accepted for a degree or
diploma by any university or institution is identified in the text.
I certify that I am the student named below and that the information provided in
the form is correct
Full Name....................................................................................…….......................
(Please Print)
Signed...................................................................................……………..............….
Date...................................................................................……………..............….
ii
"Do not go where the path may lead, go instead where
there is no path and leave a trail."
Ralph Waldo Emerson
iii
Table of Contents
Table of Contents....................................................................................................... iv
Abstract ...................................................................................................................... ix
Acknowledgements ................................................................................................... xii
Publications ..............................................................................................................xiii
Chapter One
Introduction ................................................................................................................ 1
1. Chemiluminescence ........................................................................................... 2
1.1
General principles .......................................................................................... 2
1.2
Acidic potassium permanganate .................................................................... 3
1.3
Tris(2,2ƍ-bipyridine)ruthenium(III) chemiluminescence................................ 4
1.4
Manganese(IV)............................................................................................... 6
1.5
Instrumentation .............................................................................................. 8
1.5.1
Flow injection analysis ...................................................................... 8
1.5.2
Sequential injection analysis ............................................................. 9
1.5.3
High performance liquid chromatography ...................................... 10
1.5.4
Stopped flow analysis ..................................................................... 11
2. Chemiluminescence flow-cells ........................................................................ 12
3. Biologically important molecules ................................................................... 14
3.1
Low-molecular mass thiols and disulfides ................................................... 14
3.1.1
3.2
GSH and oxidative stress ................................................................ 16
Determination of biologically important thiols and disulfides ..................... 17
4. Forensically important molecules .................................................................. 20
4.1
Opiate alkaloids............................................................................................ 20
4.2
Determination of opiate alkaloids ................................................................ 22
iv
5. Project aims ...................................................................................................... 26
Chapter Two
Solution mixing in flow-cells for chemiluminescence detection ........................... 27
1. Introduction ..................................................................................................... 28
2. Experimental .................................................................................................... 30
2.1
Chemicals and reagents ................................................................................ 30
2.2
Flow injection analysis................................................................................. 31
2.3
Sequential injection analysis ........................................................................ 32
2.4
Flow-cells ..................................................................................................... 34
3. Results and Discussion .................................................................................... 37
3.1
Preliminary investigations into novel flow-cell configurations ................... 37
3.2
Further investigations into channel design ................................................... 41
3.3
Flow-cell material ........................................................................................ 43
3.4
Confluence point and mixing chamber ........................................................ 44
4. Conclusions....................................................................................................... 50
Chapter Three
Enhanced permanganate chemiluminescence ....................................................... 51
1. Introduction ..................................................................................................... 52
2. Experimental .................................................................................................... 54
2.1
Chemicals and reagents ................................................................................ 54
2.2
Flow injection analysis................................................................................. 55
2.3
Flow-cells ..................................................................................................... 56
2.4
Stopped flow ................................................................................................ 56
2.5
Sequential injection analysis ........................................................................ 57
2.6
UV-Visible spectrometry ............................................................................. 58
3. Results and Discussion .................................................................................... 59
3.1
Manganese(II) .............................................................................................. 59
3.1.1
Preliminary chemiluminescence investigations .............................. 59
v
3.2
3.1.2
Optimisation of reagent conditions ................................................. 61
3.1.3
Analyte screening test ..................................................................... 63
3.1.4
Stopped flow analysis ..................................................................... 65
3.1.5
Flow-cell comparisons .................................................................... 68
Sodium thiosulfate ....................................................................................... 71
3.2.1
Preliminary experiments ................................................................. 71
3.2.2
Stopped flow analysis ..................................................................... 73
3.3.3
Stability studies ............................................................................... 75
3.2.4
Reagent comparison and screening test .......................................... 77
3.2.5
Detection limits ............................................................................... 80
4. Conclusions....................................................................................................... 81
Chapter Four
Direct detection of biologically significant
thiols and disulfides with post-
column chemiluminescence ..................................................................................... 82
1. Introduction ..................................................................................................... 83
2. Experimental .................................................................................................... 86
2.1
Chemicals and reagents ................................................................................ 86
2.2
Flow injection analysis................................................................................. 89
2.2.1
Acidic potassium permanganate...................................................... 89
2.2.2
Tris(2,2‫މ‬-bipyridine)ruthenium(III) ................................................. 90
2.2.3
Soluble manganese(IV) ................................................................... 91
2.3
Chemiluminescence spectra ......................................................................... 92
2.4
High performance liquid chromatography ................................................... 92
2.5
Sample collection and analysis .................................................................... 94
3. Results and Discussion .................................................................................... 95
3.1
Acidic potassium permanganate chemiluminescence detection ................ 95
3.2
Tris(2,2‫މ‬-bipyridine)ruthenium(III) chemiluminescence detection .............. 97
3.3
Manganese(IV) chemiluminescence detection........................................... 101
3.3.1
Separation conditions and analytical figures of merit ................... 106
vi
3.3.2
Determination of GSH and GSSG in whole blood........................ 109
4. Conclusions..................................................................................................... 114
Chapter Five
Precision milled chemiluminescence flow-cells for flow injection analysis and
high performance liquid chromatography ........................................................... 115
1. Introduction ................................................................................................... 116
2. Experimental .................................................................................................. 118
2.1
Chemicals and reagents .............................................................................. 118
2.2
Flow-cells ................................................................................................... 119
2.3
Flow injection analysis............................................................................... 124
2.4
High performance liquid chromatography ................................................. 124
3. Results and discussion ................................................................................... 126
3.1
Channel size ............................................................................................... 126
3.2
Chip material .............................................................................................. 128
3.3
Channel design ........................................................................................... 130
3.4
Flow-cell comparison ................................................................................. 133
3.4.1
Flow injection analysis comparison .............................................. 134
3.4.2
High performance liquid chromatography .................................... 138
4. Conclusions..................................................................................................... 146
Chapter Six
Determination of Papaver somniferum alkaloids in drug seizure samples using
high
performance
liquid
chromatography
with
UV-absorbance
and
chemiluminescence detection ................................................................................. 148
1. Introduction ................................................................................................... 149
2. Experimental .................................................................................................. 153
2.1
Chemicals and reagents .............................................................................. 153
2.2
Flow injection analysis............................................................................... 154
2.3
High performance liquid chromatography ................................................. 155
vii
2.3.1
Conventional high performance liquid chromatography with
post-column chemiluminescence detection................................................ 156
2.3.2
High performance liquid chromatography with multiplexed
detection ..................................................................................................... 157
3. Results and discussion ................................................................................... 159
3.1
Initial experiments...................................................................................... 159
3.1.1
3.2
Application to real samples ........................................................... 160
Conventional high performance liquid chromatography with post-column
UV-absorbance detection .................................................................................... 162
3.3
3.2.1
Optimisation of separation conditions........................................... 163
3.2.2
Application to real samples ........................................................... 165
Conventional high performance liquid chromatography with post-column
tris(2,2ƍ-bipyrine)ruthenium(III) chemiluminescence detection.......................... 168
3.3.1
3.4
Application to real samples ........................................................... 170
Conventional high performance liquid chromatography with post-column
enhanced permanganate chemiluminescence detection ...................................... 172
3.4.1
3.5
Application to real samples ........................................................... 174
High performance liquid chromatography with multiplexed UV-absorbance
and chemiluminescence detection ....................................................................... 176
3.5.1
Application to real samples ........................................................... 179
4. Conclusions..................................................................................................... 183
References ............................................................................................................... 184
viii
ABSTRACT
Chemiluminescence, the production of light as the result of a chemical reaction, is
an attractive method of detection for a variety of liquid-phase techniques, due to the
high selectivity and sensitivity. This thesis encompasses a series of investigations
into (i) the design, evaluation and optimisation of novel flow-cells, (ii) improving the
sensitivity of several analytically important chemiluminescence reactions via
optimisation of reagent chemistry, and (iii) the development of methodologies that
exploit these innovations for the determination of biologically significant thiols and
disulfides as well as forensically important opiate alkaloids.
Constructing flow-cells for chemiluminescence detection by machining channels
into polymer disks has enabled the exploration of new configurations and materials
that can improve signal intensity beyond that attainable with the traditional
coiled-tubing design. The optimal position of the confluence point was found to be
external to the detection zone for several relatively fast chemiluminescence reactions.
However, for those with extremely fast kinetics, merging solutions directly in front
of the photomultiplier tube provided greater signals. Modifying this design to
incorporate central mixing zones with larger widths than the channel reduced the
chemiluminescence response. Although the spirally propagating serpentine channel
promotes efficient mixing and greater chemiluminescence intensities than a spiral
channel, increasing the sharpness of the turns created areas of poor solution flow and
decreased the chemiluminescence response. Disks constructed from Teflon
impregnated with glass microspheres increased the quantity of light reaching the
photodetector, highlighting the importance of surface reflectance.
The chemiluminescence oxidation of organic compounds with acidic potassium
permanganate can be catalysed by manganese(II) sulfate or sodium thiosulfate. This
ix
provides greater control of reaction rates, enhanced emission intensities and
improved stability when compared to the standard permanganate reagent. This
‘enhanced’ permanganate was used to detect a wide variety of compounds including
biomolecules, antioxidants, neurotransmitters, opiate alkaloids and pharmaceuticals.
The increased sensitivity arising from this approach resulted in detection limits
superior to all previously reported values with acidic potassium permanganate for
four model analytes (morphine, synephrine, fenoterol and tyrosine).
The quantification of low-molecular mass thiols and disulfides involved in cellular
redox processes is hindered by oxidation or degradation of analytes during
conventional
sample
preparation
steps
(including
deproteinisation
and
derivatisation). A manganese(IV) colloid was found to provide a simple and sensitive
approach for the direct detection of biologically significant thiols and disulfides
using both FIA and HPLC methodologies. A mixture of seven thiols and disulfides
(cysteine, cystine, homocysteine, homocystine, glutathione (GSH), glutathione
disulfide (GSSG) and N-acetylcysteine) in their native forms were separated within
twenty minutes. Detection limits for these analytes ranged from 5 × 10 -8 M to
1 × 10-7 M. This procedure was successfully applied to the determination of two key
biomarkers of oxidative stress, GSH and GSSG in whole blood taken from twelve
human subjects. Samples were simply deproteinised, centrifuged and diluted prior to
analysis, using a procedure that was shown to avoid significant auto-oxidation of
GSH.
Novel flow-cells with integrated confluence points and reaction channels were
evaluated to refine several key design attributes identified throughout the proceeding
investigations. Integrating the confluence point into the detection zone facilitated
rapid presentation of the reacting solutions to the photodetector. Previously reported
x
chemiluminescence detectors constructed by machining channels into polymers have
almost exclusively been prepared using transparent materials. However, far greater
emission intensities were obtained using an opaque white chip with a thin transparent
seal, which directed more of the emitted light to the photodetector. This approach
also enabled the exploration of channel designs that could not be incorporated into
traditional coiled-tubing flow-cells. The combination of these flow-cells (that more
efficient mix fast chemiluminescence reactions) with the enhanced permanganate
reagent has improved superior limits of detection for a range of analytes.
A rapid HPLC based post-column chemiluminescence screening test for Papaver
somniferum opiate alkaloids and opiate derivatives in drug seizure samples was
developed using the complimentary selectivity of the enhanced acidic potassium
permanganate and a stabilised tris(2,2ƍ-bipyridine)ruthenium(III) reagent. The
combination of these two stabilised chemiluminescence reagents proved highly
compatible with the extended analysis times inherent to HPLC procedures and
generated limits of detection comparable to or superior than those previously
reported for the P. somniferum opiate alkaloids. Furthermore, the addition of an
annular frit to the end of the chromatographic column enabled segmentation of the
column eluate between three unique modes of detection (UV-absorbance and two
chemiluminescence reagents) with only minor losses of sensitivity (less than an order
of magnitude) whilst significantly reducing the sample and solvent consumption,
waste generation and total analysis time.
xi
ACKNOWLEDGEMENTS
I would like to extend my gratitude to a number of people without whom this
thesis would not have come to fruition.
Firstly I would like to thank my supervisors Doctor Paul Francis and Professor
Neil Barnett for their encouragement, guidance, patience, unwavering support and
friendship over the last few years. It has been an absolute privilege to work alongside
you.
I would also like to thank the following people who contributed to this work:
Doctor Xavier Conlan for his willingness to troubleshoot HPLC based problems,
Donna Squire for her endless patience when required to spend hours in the dark
taking the amazing photos featured in this thesis, Doctor Stephan Mohr for
enthusiastically designing and constructing the flow-cells featured in chapter five,
Doctor Jacqui Adcock for developing the SIA software, Gregory Van Egmond for
the intravenous blood collection, Michelle Camenzuli for her assistance with the
novel column technology featured in the final chapter, the Australian Government
for providing the funding for this research and all the honours and PhD students that
I have shared a lab with over the last couple of years.
I am deeply grateful to Danielle Bassanese, David Donaldson, Doctor Geoffrey
McDermott, Kelly Morris, Teo Slezak, Zoe Smith, Rebecca Waite and Elizabeth
Zammit for their invaluable advice, assistance and friendship. I quite simply would
not have been able to do this without you!
Finally, I would like to thank my family, Mum, Dad and Kylie. The realisation of
this dream is due in no small part to the love, support and understanding you
continue to offer me and I am eternally grateful.
xii
PUBLICATIONS
Listed below are the references to publications that have resulted from the work
presented in this thesis.
1.
Stephan
Mohr,
Jessica
M.
Terry,
Jacqui L. Adcock, Peter R. Fielden, Nick J. Goddard,
Neil W. Barnett, Duane K. Wolcott and Paul S. Francis,
‘Precision milled flow-cells for chemiluminescence
detection,’ Analyst, 2009, 134 (11), 2233-2238.
An image from this work was featured on the front
cover of the November 2009 issue of the Analyst.
2.
Teo Slezak, Jessica M. Terry, Paul S. Francis,
Christopher
M.
Hindson,
Don
C.
Olson,
Duane K. Wolcott and Neil W. Barnett, ‘Autocatalytic
nature of permanganate oxidations exploited for highly
sensitive chemiluminescence detection,’ Analytical
Chemistry, 2010, 82 (6), 2580–2584.
3.
Paul S. Francis, Christopher M. Hindson,
Jessica M. Terry, Zoe M. Smith, Teo Slezak,
Jacqui L. Adcock, Bronwyn L. Fox and Neil W.
Barnett, ‘Enhanced permanganate chemiluminescence,’
Analyst, 2011, 136 (1), 64-66.
An image from this work was featured on the inside
front cover of the January 2010 issue of the Analyst.
xiii
4.
Jacqui
L.
Adcock,
Jessica
M.
Terry,
Colin J. Barrow, Neil W. Barnett, Don C. Olsen and
Paul S Francis, ‘Chemiluminescence detectors for
liquid chromatography,’ Drug Testing and Analysis,
2011, 3 (3), 139-144.
An image from this work was featured on the front
cover of the March 2011 issue of Drug Testing and
Analysis.
5.
Jessica M. Terry, Elizabeth M. Zammit,
Teo Slezak, Neil W. Barnett, Don C. Olson,
Duane K. Wolcott, Donna L. Edwards and Paul S.
Francis, ‘Solution mixing and the emission of light in
flow-cells for chemiluminescence detection,’ Analyst,
2011, 136 (5), 913-919.
An image from this work was featured on the inside
front cover of the March 2011 issue of the Analyst.
6.
Geoffrey P. McDermott, Jessica M. Terry,
Xavier A. Conlan, Neil W. Barnett and Paul S. Francis,
‘Direct detection of biologically significant thiols and
disulfides with manganese(IV) chemiluminescence,’
Analytical Chemistry, 2011, 83 (15), 6034-6039.
An image from this work was featured on front cover of
the August 2011 issue of Analytical Chemistry.
xiv
7.
Jessica
M.
Terry,
Stephan
Mohr,
Peter R. Fielden, Nick J. Goddard, Neil W. Barnett,
Don C. Olson, Duane K. Wolcott and Paul S. Francis,
‘Chemiluminescence
injection
analysis
detection
and
flow-cells
high
for
performance
flow
liquid
chromatography’, Analytical and Bioanalytical Chemistry,
accepted February 2012.
8.
Jessica
Geoffrey
Neil
Jarrad
P.
M.
Terry,
McDermott,
W.
Barnett,
M.
Altimari
Zoe
Rebecca
Luke
and
M.
Smith,
J.
Waite,
C.
Paul
Henderson,
S.
Francis,
‘Chemiluminescence detection of amino acids and related
compounds
using
manganese(IV)
or
acidic
potassium
permanganate,
tris(2,2‫މ‬-bipyridine)ruthenium(III)’,
submitted to Talanta, January 2012.
xv
CHAPTER ONE
Introduction
x Chemiluminescence
x Chemiluminescence flow-cells
x Biologically important molecules
x Forensically important molecules
x Project aims
1
1. Chemiluminescence
1.1
General principles
Chemiluminescence is defined as the production of light (ultra-violet, visible or
infra-red) as the result of a chemical reaction [1-5]. This reaction typically yields an
electronically excited species (C*) which may either emit a photon (direct
chemiluminescence; Scheme 1) or donate its energy to a suitable compound (F)
which then luminesces (indirect or sensitised chemiluminescence; Scheme 2) [1-4].
This emission process, radiative relaxation of an excited state species to its ground
state, is indistinguishable from other forms of luminescence (i.e. fluorescence) [1-4,
6].
A + Bĺ C* + Products
C* ĺ C + Light
Scheme 1. Direct chemiluminescence
A + Bĺ C* + Products
C* + F ĺ F* + C
F* ĺ F + Light
Scheme 2. Indirect chemiluminescence
The rate of light generation from these reactions is highly dependent on physical
and chemical processes such as solution mixing and reaction kinetics [1-4, 7]. The
intensity of the emitted light is proportional to the concentrations of the chemical
species involved, and as such, can be used for quantitative analysis [1-4, 7, 8].
There are several inherent advantages of chemiluminescence which have resulted
in it becoming an attractive mode of detection for a variety of liquid-phase
techniques (such as flow injection analysis, stopped flow analysis, sequential
2
injection analysis and high performance liquid chromatography) [1-3, 5-8]. Only a
relatively low number of chemical reactions are capable of producing
chemiluminescence, which results in higher selectivity than other luminescence
based techniques [1-4, 7, 8]. Furthermore, the absence of an excitation source results
in improved signal-to-noise ratios and greater sensitivity which can lead to superior
limits of detection [1-3, 7, 8]. These features, combined with the use of relatively
simple instrumentation, have resulted in chemiluminescence detection becoming a
widely applied analytical technique [1-3, 5, 7-9].
1.2
Acidic potassium permanganate
In 1917, E.N. Harvey reported the use of acidic potassium permanganate as a
chemiluminescence reagent after observing the production of light during the
oxidation of a weak solution of pyrogallol [10]. However, the earliest analytical
application was not reported until 1975, by Stauff and Jaeschke for the determination
of sulphur dioxide [11]. Since Harvey’s initial report, hundreds of papers have been
published in scientific literature detailing its analytical applications, including
comprehensive reviews by Hindson et al. [12] and Adcock et al. [13].
Despite being used extensively for more than three decades, the light-producing
pathway of acidic potassium permanganate has only recently been elucidated [14].
Whilst many researchers erroneously attributed the characteristic red emission to
singlet oxygen [13, 15], it has been proven to emanate from a manganese(II) species
upon relaxation from an excited state (4T1 ĺ 6A1 transition) [13, 16, 17]. Hindson
et al. [14] recently demonstrated that the oxidation of an analyte by acidic potassium
permanganate [manganese(VII)] generates a radical intermediate and manganese(III)
(Figure 1). Subsequent reaction of these species produces an excited state
manganese(II) emitter (wavelength of maximum emission 734 ± 5 nm) [14]. The
3
presence of sodium polyphosphates increases the emission intensity (up to 50-fold)
[12, 13] by forming ‘‘cage-like’’ structures around the manganese(II) emitter that
prevents deactivation by non-radiative pathways and shifts the wavelength of
maximum emission to 689 ± 5 nm [13, 14].
Manganese(VII) + Analyte
Oxidation
Reduction
Manganese(III) + Analyte radical
Oxidation
Reduction
Manganese(II)* + Other products
hv Ȝmax 689 nm
Manganese(II)
Figure 1. Postulated mechanism for permanganate chemiluminescence (adapted
from Hindson et al.[14]).
The determination of numerous organic analytes with acidic potassium
permanganate has enabled some general trends regarding analyte structure and
chemiluminescence intensity to be established. In general, molecules containing
phenolic or amine moieties elicit the greatest chemiluminescence responses [12, 13].
Accordingly, several important classes of analytes (antioxidants, biomolecules, illicit
drugs and pharmaceuticals) which contain these functional groups have received the
most attention to-date [12, 13, 18-27].
1.3
Tris(2,2ƍ-bipyridine)ruthenium(III) chemiluminescence
Tris(2,2ƍ-bipyridine)ruthenium(II) [Ru(bipy)3]2+ was initially synthesised by
F.H. Burstall in 1936 [28], but it was not until three decades later that Hercules and
Lytle [29] published the first account of chemiluminescence from this complex. They
4
had observed the production of orange luminescence when a solution of
tris(2,2ƍ-bipyridine)ruthenium(III) was reacted with concentrated sodium hydroxide
[29]. Comprehensive reviews by Gerardi et al. [30] and Gorman et al. [31] illustrate
the numerous analytical applications of this reagent.
Whilst a definitive reaction mechanism for the light-producing pathway has yet to
be elucidated, the general detection chemistry is well-established (Figure 2).
Common
to
all
analytical
tris(2,2ƍ-bipyridine)ruthenium(III)
applications
reagent
is
the
[Ru(bipy)3]3+
generation
via
of
chemical
the
or
electrochemical oxidation of [Ru(bipy)3]2+ [5, 30-33]. Subsequent reaction with a
suitable reducing agent (the analyte or analyte oxidation product) results in the
formation of an electronically excited state species [Ru(bipy)3]2+*, which then emits
its characteristic orange luminescence (wavelength of maximum emission 610 nm)
[5, 30-33].
hv Ȝmax 610 nm
Ru(bipy)32+
[Ru(bipy)3
2+]*
Oxidation (chemical or electrochemical)
Ru(bipy)33+
Reduction (analyte)
Figure 2. Generalised
chemiluminescence.
mechanism
for
tris(2,2ƍ-bipyridine)ruthenium
The ability to offer a constant, stable supply of the [Ru(bipy)3]3+ species has seen
in situ electrochemistry become the prevalent method of oxidation [30, 31].
However, this technique suffers from downfalls associated with equipment
complexity and electrode fouling [30, 31]. An alternative is to employ chemical
oxidation using cerium(IV) sulfate or lead dioxide [30, 31]. This approach is rapid
5
and simple, but does exhibit limited temporal stability, which has to-date
significantly impeded the analytical utility of this method of oxidation [1, 5, 30, 31,
33-36]. To overcome this, McDermott et al. [33] recently synthesised
[Ru(bipy)3](ClO4)2, which (when dissolved in acetonitrile containing 0.05 M
perchloric acid and oxidised with lead dioxide) proved to be exceedingly stable over
extended analysis periods (up to 48 hours) and removed the need for constant
re-calibration or preparation of fresh reagent solutions.
Tris(2,2‫މ‬-bipyridine)ruthenium(III)
has
been
used
extensively
for
the
chemiluminescence and electrochemiluminescence determination of various analytes
containing an amine moiety (in particular amino acids, antioxidants and illicit drugs)
[5, 23, 25, 26, 30, 31, 34, 37, 38]. In general, tertiary substituted amines generally
elicit the largest response, followed by secondary and then primary [30, 31, 38, 39].
The
presence
of
phenol
substituents
appears
to
drastically
affect
the
chemiluminescence signal, with several studies reporting inhibition or quenching of
the emission [31, 40, 41].
1.4
Manganese(IV)
Manganese(IV) in its most commonly encountered form - the dark brown
manganese dioxide - has been used extensively as an oxidant in many fields of
chemistry as detailed in the comprehensive review by Pastor and Pastor [42].
However, the exceedingly poor solubility of manganese(IV) has limited its analytical
utility [43].
In 1984, and again in 1993 Jáky et al. reported a method for preparing a ‘soluble’
form of manganese(IV) [44, 45]. This involved the reduction of potassium
permanganate with excess sodium formate to yield the characteristic dark brown
6
manganese dioxide precipitate which was then dissolved into concentrated
orthophosphoric acid [44, 45]. Experiments conducted by Jáky et al. using
transmission electron microscopy demonstrated that the solution contained small
particles less than 2 nm in diameter [44, 45]. Subsequent research conducted by
Brown et al. [46] using both scanning and transmission electron microscopies
revealed the presence of small uniform nanoparticles (20 nm), which indicated that
this form of soluble manganese(IV) was actually a colloidal suspension rather than a
solution.
The first report of chemiluminescence from soluble manganese(IV) occurred
nearly a decade later when B.J. Hindson [47, 48] used it to determine 25 organic and
inorganic analytes. Corrected chemiluminescence spectra revealed that the light
emitted from reactions with manganese(IV) had an emission maxima centred at
734 ± 5 nm, which is characteristic of many reactions with manganese-based
oxidants [13, 16, 43, 47, 48]. Furthermore, it is suggestive that the red emission
emanates from an excited state manganese(II) species [16, 17]. Incorporation of
formaldehyde into the reaction system (as an enhancer) significantly improves
emission intensity and limits of detection [46-48].
Only a small number of papers concerning the use of manganese(IV) as a
chemiluminescence reagent have been published [43]. From these, no general trends
with respect to molecular structure and intensity have emerged, but it is apparent that
the reagent exhibits markedly different selectivity to that of acidic potassium
permanganate [43].
7
1.5
Instrumentation
1.5.1
Flow injection analysis
Flow injection analysis (FIA) was first introduced in 1975 by Ruzicka and Hansen
[49] and Stewart et al. [50] as a rapid method of sample handling for laboratory
analysis. This simple dispersion technique allows reactions to be performed by
injecting a precise aliquot of sample into a continuously flowing carrier stream,
which subsequently merges with a reagent [2, 4, 7, 8, 49-53]. Three decades ago,
Rule and Seitz published the first account of FIA with chemiluminescence detection,
in which copper was used to catalyse the luminol-hydrogen peroxide reaction [54].
Many chemiluminescence reactions occur very rapidly, and due to the transient
nature of the emission, a significant advantage of using FIA methodology is the
ability to reproducibly merge two reacting solutions as close as possible to the
detector [2, 4, 7, 8, 52, 53]. To ensure that the maximum amount of radiation is
captured by the detection system (usually a photomultiplier tube) factors such as the
reaction kinetics, volume between the confluence and observation points, the
geometry of the flow-cell and the flow rate all need to be carefully optimised [2, 7, 8,
52, 53].
Since Rule and Seitz’s initial report [54], hundreds of papers have been published
in scientific literature examining the use of FIA to: (i) explore the fundamental
chemistry behind chemiluminescence reactions, (ii) optimise post-column reaction
conditions for liquid chromatography and (iii) quantify analytes in simple matrices as
detailed in the comprehensive review by Fletcher et al. [8].
There are however, some limitations associated with using chemiluminescence as a
detection method for FIA [2, 7, 52]. The addition of reagents and modifications to
8
the instrumental setup increases the overall complexity of the manifold [2, 7, 52].
This can lead to lower reproducibility, particularly if different flow rates are used.
Furthermore, due to the simple instrumentation and the lack of a separative
component in traditional manifolds, FIA does not have the selectivity to determine
multiple analytes within complex biological matrices [2, 7, 52].
1.5.2
Sequential injection analysis
Sequential injection analysis (SIA) was developed in 1990 by Ruzicka and
Marshall to address some of the aforementioned limitations of FIA and the need for a
more powerful and versatile flow-based sample-handling methodology [55]. Unlike
FIA where a sample is injected into a continuously flowing carrier stream and a
reagent is subsequently merged on-route to the flow-cell [2, 4, 7, 8, 49-53], SIA
utilises one flowing stream under the control of a bidirectional pump [55-59]. This
pump initially aspirates the carrier solution followed by defined aliquots of sample
and reagent; each introduced via the action of a computer controlled multi-position
valve [55-59]. This sequential aspiration results in a defined series of sample and
reagent zones which are directed towards a holding coil [55-59]. The direction of
flow is then reversed and the sample and reagent zones are propelled towards the
flow-cell creating product zones from the overlapping regions of the sample and
reagent(s) [55-59]. It is these zones that generate the radiation detected by the
photomultiplier tube [56-58] (Figure 3).
9
(i)
(ii)
R
R
Valve
S
Valve
(iii)
Valve
R
P
S
Figure 3. Schematic diagram illustrating: (i and ii) the pump operating in the reverse
mode to aspirate and sequentially stack plugs of reagent (R) and sample (S) and (iii)
reversal of flow resulting in the formation of a zone of detectable product (P).
Arrows indicate the direction of flow (adapted from [56]).
Owing to the use of robust syringe or Milligat pumps and computer control, the
advantages of SIA over the more traditional FIA are; (i) the capacity to stack, in any
order, precisely defined zones of sample and reagent that are mixed to yield
detectable products, (ii) lower reagent consumption, and (iii) production of far less
waste [2, 55-58]. A disadvantage of SIA is the lower sample throughput, as the
indispensable step of stacking the sample and reagents in the holding coil prior to
reversing the direction of flow results in longer run times than FIA [2, 55-58].
1.5.3
High performance liquid chromatography
Reversed-phase high performance liquid chromatography (HPLC) is a powerful
method of identifying, quantifying and purifying specific compounds in complex
matrices (i.e. process samples, pharmaceuticals and physiological fluids) [60-62].
The separation of each component in a given sample is attributable to hydrophobic
interactions and repulsive forces occurring between a moderately polar mobile phase
(typically water containing an organic modifier), the sample, and a nonpolar
stationary phase (consisting of a hydrophobic ligand chemically bound to a
10
particulate or solid support) [60-62]. As the components within a sample are
separated, they elute off the column and into a detector, typically UV-absorption or
fluorescence, where their individual retention times are recorded in the form of a
chromatogram [60-62].
Coupling HPLC with chemiluminescence detection involves merging a flowing
stream of the chemiluminescence reagent(s) (via the action of a peristaltic or syringe
pump) with the column eluate at a confluence point located prior to the point of
detection [5, 6, 9, 52]. Post-column chemiluminescence introduces selectivity that is
normally unattainable with FIA, and offers improved sensitivity when compared to
the aforementioned detection techniques [5, 6, 9, 52]. Compromises between the
chemical parameters required for an efficient chromatographic separation and
sensitive post-column detection are usually required, as chemiluminescence
emissions are extremely susceptible to changes in pH, flow-rate and the amount and
type of organic modifier in the mobile phase [5, 6, 9, 52].
1.5.4
Stopped flow analysis
In flow-based methodologies such as FIA, SIA or HPLC, the analyte, carrier and
reagent(s) are propelled in streams and subsequently merged to initiate the
chemiluminescence reaction close to or within the flow-cell [2, 4, 7, 8, 49-53]. This
results in only a small portion of the total emission being captured by the
photodetector [63, 64] (Figure 4i). In stopped flow analysis, discrete quantities of the
reagent(s), carrier and analyte are propelled by a syringe pump into a detection
chamber or flow-cell, where they are held for a set period of time [65-68]. This
enables the entire chemiluminescence intensity versus time profile to be monitored
[68] (Figure 4ii). Generally, the temporal distribution of the signal is proportional to
the analyte concentration and can provide information about the reaction kinetics
11
which may be exploited to aid in the optimisation of reaction and manifold
conditions for flow-based methodologies [68]. Furthermore, in cases where the
chemiluminescence kinetics are distinct for two or more compounds, the signals
resulting from mixtures of analytes can be mathematically derived to provide
(ii)
Chemiluminescence intensity (mV)
(i)
Chemiluminescence intensity (mV)
enhanced selectivity without a separation step [67-69].
Time (s)
Time (s)
Figure 4. Schematic of a typical chemiluminescence peak where (i) depicts the
portion of the emission captured by the detector using FIA methodology and (ii)
depicts the portion of the emission captured by the detector with stopped flow
analysis.
2. Chemiluminescence flow-cells
Whilst each of the aforementioned analytical techniques is unique with regard to
the manipulation, mixing and transportation of analytes and reagents to the point of
detection, inherent to each is the presence of a flow-cell which retains the reacting
solution in view of a photomultiplier tube [2, 4, 6, 7, 52, 70]. The design and
geometry of the flow-cell is crucial; it needs to be transparent, have appropriate
channel dimensions to ensure efficient mixing and maximise sensitivity, and be inert
to the chemical reaction being employed [2, 4, 6, 7, 52, 70].
Coiled-tubing flow-cells were first introduced for chemiluminescence detection in
FIA by Rule and Seitz in 1979 [54]. The cells were simple to construct, compatible
12
with photodetector geometry, and enabled collection of the emitted light during the
first few seconds of the reaction [6, 53, 54, 71, 72]. More than three decades later the
most commonly reported flow-cell configuration for flow analysis based techniques
still consists of a coil of glass or polymer tubing (normally 0.5-1.0 mm i.d.) mounted
against a photomultiplier tube within a light-tight container [2, 4, 6, 7, 52, 53, 70].
Solutions merge at a T- or Y-shaped junction shortly before entering the coil
(Figure 5). However this simple design has several serious limitations; (i) the walls
of the tubing are curved, and therefore most of the surface is not flat against the
photodetector window, (ii) polymer tubing is often translucent rather than totally
transparent, (iii) the geometry of the flow-cell is hindered by the limited flexibility of
polymer or glass tubing, (iv) mixing is initiated before the reacting mixture enters the
coil and (v) the internal diameter of the tubing is limited by availability [7, 73].
Confluence point
Reagent line
Reaction coil
Photomultiplier tube
Carrier line
High voltage input
and signal output
Waste line
Figure 5. Schematic of a traditional chemiluminescence flow-cell used in
various flow analysis based techniques [53].
A variety of alternative chemiluminescence detection cells have been described,
yet to-date none of these designs have been adopted for routine analysis [74-83]. In
the fountain flow-cell [83], the reacting mixture enters the centre of an open, shallow
cylindrical space and drains into a ring-shaped well around the edge, which contains
the outlet hole. This configuration allows a greater volume of solution to be in
contact with a flat surface facing the photodetector. Another design, referred to as the
13
bundle cell [81], consists of a bundle of Teflon tubing packed into a plastic cuvette.
This configuration was found to be 50% more efficient than a coil composed of the
same tubing [81]. However, the first turn of the coil used in that particular study had
a diameter of 1 cm, leaving a relatively large empty space in the centre of the coil
compared to similar flow-cells constructed by other researchers. These studies help
to highlight the importance of optimising flow-cell design and geometry to maximise
mixing efficiency and detection of the emitted light.
3. Biologically important molecules
3.1
Low-molecular mass thiols and disulfides
Low-molecular mass thiols such as homocysteine, cysteine and glutathione, are
critical physiological components endogenous to all biological tissues and fluids [8488]. These thiols are mainly present in their reduced forms which can be converted to
the corresponding disulfide (cystine, homocystine and glutathione disulfide) under
conditions of oxidative or nitrosative stress [84-88].
Homocysteine is an endogenous thiol-containing amino acid synthesised in the
liver as an intermediate of the methionine cycle (Figure 6) [84, 86, 89, 90]. Under
normal conditions, intracellular concentrations of homocysteine are kept relatively
low as the result of (i) remethylation back to methionine using the
vitamin B12-dependent methionine synthetase and/or (ii) conversion to cysteine via
the trans-sulfuration pathway [84, 86, 89-92]. Elevated concentrations of
homocysteine in physiological fluids are considered to be an important risk factor or
marker for several diseases, and as such interest in the determination of this
biologically significant thiol has rapidly grown during the last decade [84, 89, 90, 9395].
14
Cysteine can be derived in vivo from two independent metabolic pathways [84, 86,
89, 91, 92, 96]. In the trans-sulfuration pathway, homocysteine is converted to a
cystathionine intermediate which is then metabolised to form cysteine [84, 86, 89,
91, 92, 96] (Figure 6). In the Ȗ-glutamyl cycle, the breakdown of glutathione by
Ȗ-glutamyl transpeptidase produces a cysteinylglycine fragment which is
subsequently cleaved by membrane-bound dipeptidases to form cysteine and glycine
(Figure 6) [85, 86, 96]. A variety of important cellular functions, including protein
synthesis, detoxification, and metabolic processes are associated with this
endogenous thiol [84-86, 89, 91, 92, 96].
The tripeptide glutathione (L-Ȗ-glutamyl-L-cysteinyl-glycine) is the most abundant
non-protein thiol produced in mammalian organs, tissues and cells [84-89, 96].
Glutathione (GSH) is synthesised via two adenosine triphosphate (ATP) dependent
steps (Figure 6) [85, 86, 89, 96]. In the rate limiting step, the dipeptide
Ȗ-glutamylcysteine is synthesised via condensation of cysteine and glutamate by the
enzyme Ȗ-glutamylcysteine synthetase [85, 86, 89, 96]. Glycine is then added to the
C-terminal of Ȗ-glutamylcysteine via the action of glutathione synthetase to form
GSH [85, 86, 89, 96]. The presence of a Ȗ-glutamyl linkage and thiol moiety enables
GSH to participate (either directly or indirectly) in many biological processes [85,
86, 89, 96]. These include protein and DNA synthesis, regulation of enzyme activity,
detoxification of electrophilic xenobiotics, maintenance of sulfhydryl groups in
proteins, protection of cell membranes from oxidative damage and antioxidant
defence [84-86, 89, 96-99].
15
Ȗ-GLUTAMYL CYCLE
NADP+
NADPH
glutathione
reductase
amino acid
Ȗ-glutamyl
transpeptidase
Ȗ-glutamyl-amino acid
Ȗ-glutamyl
cyclotransferase
cysteinyl-glycine
cysteinyl-glycine
dipeptidase
glutathione
synthetase
glycine
glutathione
peroxidase
ROOH
R-OH
ADP + Pi
ATP
Ȗ-glutamyl-cysteine
amino acid
Ȗ-glutamylcysteine synthetase
cysteine
5-oxoproline
glutathione
disulfide
glutathione
ADP + Pi
ATP
oxoprolinase
glutamic acid
ATP
ADP + Pi
TRANS-SULFURATION
PATHWAY
cystathionine
cystathionine ȕ-synthase
homocysteine
methionine
synthetase
protein
- ADO
methionine
+ ADO
S-adenosylhomocysteine hydrolase
S-adenosylhomocysteine
methionine
adenosyltransferase
ATP
methylated acceptor
S-adenosylmethionine methyltransferase
ADP + Pi
S-adenosylmethionine
methylated product
METHIONINE CYCLE
Figure 6. Schematic representation of the methionine cycle, trans-sulfuration
pathway and Ȗ-glutamyl cycle (adapted from [85, 86, 89]).
3.1.1
GSH and oxidative stress
Reactive oxygen species (ROS), as well as reactive nitrogen species (RNS) are
produced as a result of normal cellular metabolic processes such as respiration and
photosynthesis [100-102]. These species play a dual role within biological systems,
having both deleterious and beneficial effects [88, 100-105]. At low/moderate
16
concentrations, ROS/RNS play crucial roles in various physiological processes
including: defence against infectious agents, enhancement of signal transduction
from various membranes, sensing of oxygen tension, induction of mitogenic
responses and regulation of redox homeostasis [88, 100-105]. The deleterious effects
occur when there is an over production of ROS/RNS, coupled with a deficiency of
enzymatic and non-enzymatic antioxidants [88, 100-105]. This causes damage to
and/or inhibits the normal functioning of cell structures, including lipids and
membranes, proteins, and DNA and is referred to as oxidative stress or nitrosative
stress [88, 100-105].
In protecting against oxidative stress and free-radical-mediated cell injury, GSH
can donate a reducing equivalent, resulting in the formation of glutathione disulfide
(GSSG) [85, 88, 97-100]. Regeneration of GSH from GSSG is mediated by the
widely distributed enzyme glutathione reductase in a nicotinamide adenine
dinucleotide phosphate (NADPH) dependant reaction [85, 88, 97-100] (Figure 6).
Typically, the molar ratio of GSH/GSSG in eukaryotic cells is between 10:1 and
1000:1; however, under conditions of oxidative stress, this ratio decreases [85, 88,
97-100]. As oxidation of the free thiol to the corresponding disulfide is one of the
earliest responses to an overproduction of reactive oxygen species, the accurate and
interference-free measurement of the GSH/GSSG ratio offers great potential for
diagnosing and monitoring pathological and physiological conditions related to
oxidative stress [84, 85, 87, 88, 97-100, 106-110].
3.2
Determination of biologically important thiols and disulfides
Since the isolation of GSH from yeast and animal tissues by Hopkins [111] in
1921, numerous analytical methodologies have been developed to quantify
low-molecular mass thiols and disulfides in biological fluids due to their importance
17
in diagnosing and monitoring pathophysiological conditions associated with
oxidative stress [85, 97, 99, 112, 113].
However, several significant analytical challenges associated with the accurate
measurement of thiols and disulfides have emerged [84, 85, 87, 107, 109, 110, 114,
115]. Due to the lack of strong chromophores or fluorophores, derivatisation of these
analytes
is
essential
for
techniques
employing
spectrophotometric
or
spectrofluorimetric detection [84, 85, 87, 107, 109, 110, 114, 115].
Thiols are usually derivatised with either a colorimetric (such as Ellman’s or
Sanger’s) or fluorogenic reagent(s) (such as bimanes, malimides, halides,
halogenosulfonylbenzofurazans, or ortho-phthaldehyde) prior to chromatographic
separation and detection [84, 85, 87, 107, 109, 110, 114, 115]. Disulfides then need
to be reduced to their corresponding thiols via the use of reagents such as
dithiothreitol, 2-mercaptoethanol or tris(2-carboxyethyl)phosphine prior to the
aforementioned derivatisation, separation and detection procedures [84, 85, 87, 107,
109, 110, 114, 115]. The disulfide concentration is then calculated from the
difference between the measurements obtained in the two separate analytical steps.
This approach is time consuming and vulnerable to several major sources of error
[87, 115], the most important of which is the facile auto-oxidation of thiols under the
alkaline conditions required for derivatisation and disulfide bond reduction [84, 85,
87, 106-109, 116, 117].
Physiological fluids, particularly whole blood, plasma and tissue samples contain
proteinaceous materials that need to be removed prior to analysis [84, 85, 87, 106,
107, 110]. Typically, acidification of the physiological fluid (with either
5-sulfosalicylic, metaphosphoric, perchloric, trichloroacetic or trifluoroacetic acids)
in combination with centrifugation provides a simple method of inducing protein
18
precipitation and yields a clear protein-free supernatant suitable for analysis [84, 85,
87, 106, 107, 110]. Disulfides are present in biological samples at much lower
concentrations than thiols, and even a small degree of auto-oxidation or incomplete
removal of proteins during these sample preparation procedures can lead to their
considerable overestimation [84, 85, 87, 106-109, 116, 117].
Therefore, the determination of these analytes should ideally combine minimal
sample handling (under conditions that prevent auto-oxidation) with direct detection
of all target analytes in a single chromatographic separation. Thiols and disulfides
have previously been detected without derivatisation using electrochemistry, mass
spectrometry or fluorescence quenching [84, 87, 107, 110, 116-119]. Whilst these
modes of detection offer advantages over more popular derivatisation techniques in
terms of procedural simplicity, their application is limited by equipment complexity,
cost, electrode stability and/or analysis time [108, 110, 120, 121].
Several
reviews
have
highlighted
the
selectivity
and
sensitivity
of
chemiluminescence detection when applied to a variety of liquid-phase techniques
such as FIA, SIA and HPLC [2, 4, 5, 8, 9, 12, 13, 30, 31, 122]. The inherent
selectivity of this technique arises as a result of the limited number of species
capable of emitting light [1]. This effectively eliminates interferences from nonreacting species present in complex matrices (such as biological fluids) that are
typically observed with other detection methods.
Our research group recently reported the determination of GSH in cultured skeletal
muscle cells with acidic potassium permanganate chemiluminescence detection
following a rapid chromatographic separation [18]. However, like other previously
reported chemiluminescence-based detection systems for thiols, this reagent was not
suitable to detect disulfides [123, 124]. Therefore quantification of GSSG required
19
thiol blocking and disulfide bond reduction steps prior to a second separative step
[18].
These issues point to the need for an approach that (i) prevents or minimises
auto-oxidation during sample preparation, (ii) removes the requirement of a timeconsuming derivatisation step and (iii) allows for the direct detection of biologically
significant thiols and disulfides with simple instrumentation following a single
chromatographic separation.
4. Forensically important molecules
4.1
Opiate alkaloids
Opium is a natural product obtained from the unripened seed pods of Papaver
somniferum, or the opium poppy [125-130]. Small incisions in the skin of an
unripened seed pod are used to obtain the milky opium exudate which consists of
proteins, sugars, lipids and approximately thirty alkaloids [125-130].
Despite opium and the opium poppy being used recreationally and medicinally for
centuries, the active ingredient responsible for physiological effects such as
analgesia, sedation and respiratory depression [127, 131] was not identified until the
early nineteenth century when F.H. Serturner reported the isolation of morphine
[126, 132]. Since this time, many more P. somniferum alkaloids have been identified,
and in general they fall into two distinct categories based upon their chemical
structure; phenanthrene alkaloids (such as morphine, codeine, thebaine and
oripavine) and isoquinoline alkaloids (such as papaverine and noscapine) [125, 126,
128]. The relative concentrations of these alkaloids vary substantially from poppy to
poppy and are highly dependent on factors such as climate, soil fertility, altitude,
available moisture, and the age of the plant [125, 126, 128].
20
The commercial sale of preparations containing morphine, as well as the invention
of the hypodermic syringe saw a rapid increase in the popularity of the drug, and by
the late nineteenth century, efforts were being made to find a non-addictive
alternative [133]. In 1874, whilst working at Saint Mary’s Hospital Medical School,
C.R. Alder Wright acetylated carbon 3 and 6 of morphine’s phenanthrene backbone
to form the semi-synthetic derivative 3,6-diacetylmorphine or heroin [125, 133-135].
Early reports suggested that when compared to morphine, heroin was more effective
at depressing the respiratory system and had a much weaker narcotic action [135,
136]. As a result of these studies, heroin was introduced into the commercial market
as a cough suppressant and treatment for tuberculosis [133, 135]. After addiction to
these medications rapidly increased, it was discovered that by simply acetylating the
phenolic and hydroxyl moieties of the phenanthrene backbone, Wright had created a
narcotic even more powerful and addictive than its parent compound [130, 133, 135].
Despite products containing heroin being quickly withdrawn from commercial sale
and the establishment of the International Convention on Narcotics in 1924, the
1950’s saw a significant rise in the illegal manufacture and misuse of the drug [133,
135].
The presence of acetyl groups on the heroin molecule significantly increases the
lipid solubility of the drug, allowing it to easily cross the blood-brain barrier [127,
131, 137]. Heroin is considered to be a prodrug; a drug that is rapidly converted into
the active therapeutic form after absorption [131, 138]. Once in the brain, it is rapidly
deacetylated
to
its
pharmacologically
active
metabolites
morphine
and
6-monoacetylmorphine [127, 131, 137]. These metabolites interact with numerous
—-opioid receptors located in the central nervous system to lessen or remove the
sensation of pain via disruption of nerve impulses [127, 131, 137]. Further
metabolism results in morphine being conjugated to a glucuronide to render it
21
soluble for excretion from the body [127, 137]. This glucuronide is a polar molecule
and a derivative of an acidic compound made from glucose [127, 137]. Once
conjugated to morphine it is detectable in urine for 2-4 days after administration of
the drug [127, 137].
Each year the pharmaceutical industry extracts and purifies hundreds of tonnes of
opiate alkaloids from P. somniferum for medicinal use [139]. Therefore, the accurate
measurement of these species is of great importance when monitoring
industrial-scale extractions of alkaloids from opium poppies and establishing the
alkaloidal content of these process extracts and pharmaceutical formulations [126,
139]. Furthermore, the on-going illicit production of heroin has resulted in it being
considered as one of the most significantly abused narcotics [128, 129]. This has
created the need to identify and/or quantify the drug (and its precursors or
metabolites) in suspected drug seizure samples, biological fluids, hair samples and
the larvae of insects found on or near the bodies of drug overdose victims [25, 26,
125, 130, 140].
4.2
Determination of opiate alkaloids
Numerous analytical methodologies exist for the identification and quantification
of opiate alkaloids [125, 130, 141-144]. Screening tests are often utilised to obtain
rapid preliminary identification of certain illicit substances or classes of drugs prior
to confirmatory testing with more specific instrumental methods [125, 128, 129, 145,
146].
The most commonly used screening method for opiates and opiate derivatives
involves mixing seizure samples with various reagents, such as Mandelin’s, Marquis
or Mecke’s, and monitoring the resultant colour change [125, 128, 129, 145, 146].
22
These spot tests are usually selective for a particular class of drug (i.e. opiates), but
are not selective for specific drugs (i.e. heroin), so a series of tests is usually
performed. Greater specificity can be obtained by performing microcrystalline
precipitation tests [125, 128, 129, 145]. However, the presence of cutting agents and
diluents can result in significant alterations to the crystal structure which can impede
identification [125, 128, 129, 145].
Characterisation and quantification of single or multiple analytes in complex
matrices (such as drug seizure samples, pharmaceutical process streams and bodily
fluids) is generally undertaken with either; (i) gas chromatography coupled to flame
ionisation or mass spectrometric detection, or (ii) HPLC and/or capillary
electrophoresis (CE) with UV-absorbance, fluorescence, electrochemical or mass
spectrometric detection [125, 130, 141-144].
As previously mentioned, chemiluminescence offers highly sensitive and selective
detection for species capable of emitting light, effectively eliminating interferences
from non-reacting species present in complex matrices that are typically observed
with other detection methods [1, 2]. As described in several reviews, the
chemiluminescence detection of opiate alkaloids and their derivatives has been
largely dominated by two reagents; acidic potassium permanganate and
tris(2,2‫މ‬-bipyridine)ruthenium(III) [12, 13, 25, 26, 30, 31].
The first report of chemiluminescence from the reaction of illicit drugs and
pharmaceuticals (antibiotics, monoamine oxidase inhibitors and opiate alkaloids)
with acidic potassium permanganate was published in 1986 by Abbott and
Townshend [147]. Whilst numerous analytes have been examined since this time,
very few can be detected at the exceedingly low concentrations reported for
morphine and other P. somniferum opiates [13]. A range of narcotic analgesics
23
(including opiate alkaloids) were subsequently screened with acidic potassium
permanganate by Abbott and co-workers [75] and Barnett and co-workers [148].
Analytes containing a phenolic moiety at carbon 3 on the phenanthrene backbone
and a furan bridge between carbons 4 and 5 were found to produce superior
chemiluminescence emissions when compared to other analytes investigated [75,
148]. Substitution of the functional group at carbon 3 had a dramatic influence on
chemiluminescence intensity, with opiates containing a methoxy moiety (codeine
and thebaine) generating responses 0.5-2% that of morphine [148].
Our research group has previously reported several applications of acidic
potassium permanganate chemiluminescence for the selective and sensitive
determination of phenolic morphinan alkaloids (morphine and oripavine) in process
samples using FIA, SIA, HPLC and CE [27, 148-154].
The first chemiluminescence determination of non-phenolic opiate alkaloids
(codeine,
heroin
and
dextromethorphan)
with
electrochemically
oxidised
tris(2,2‫މ‬-bipyridine)ruthenium(III) was reported by Greenway and co-workers [155]
in 1995. Subsequent research by Barnett and co-workers [156] employed chemically
oxidised tris(2,2‫މ‬-bipyridine)ruthenium(III) to determine non-phenolic morphinan
alkaloids in P. somniferum process extracts. A comprehensive examination of several
phenolic and non-phenolic opiate alkaloids and their derivatives revealed that
emission intensities were highly dependent on reaction pH and molecular structure,
with phenolic moieties and quaternary nitrogens quenching the response [41].
Conversion of the phenolic moiety to a methoxy (codeine, 6-methoxycodeine and
thebaine) or an acetyl group (heroin) evoked intense chemiluminescence emissions
with the tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent [41]. Our research group has
reported the selective determination of several non-phenolic morphinan alkaloids
24
using tris(2,2‫މ‬-bipyridine)ruthenium(III) chemiluminescence detection coupled to
FIA, SIA, HPLC and CE [27, 33, 152-154, 156-158].
The complimentary selectivity of these reagents for phenolic and non-phenolic
morphinan alkaloids served as the basis of a rapid screening test for heroin in drug
seizure samples developed by Agg and co-workers [158] (Figure 7). Heroin produces
an intense chemiluminescence response with tris(2,2‫މ‬-bipyridine)ruthenium(III) and a
relatively weak response with acidic potassium permanganate. When the seizure
sample is mixed with a small amount of a concentrated base, heroin is rapidly
converted into its metabolites, morphine and 6-monoacetylmorphine, which elicit an
intense response with acidic potassium permanganate and a negligible response with
tris(2,2‫މ‬-bipyridine)ruthenium(III) [158]. This methodology has been successfully
applied to several drug seizure samples using both FIA and SIA approaches [27,
158]. However the absence of a separative component in the two instrumental
approaches has, to-date prevented the identification of various opiates and/or
alkaloidal impurities within these seizure samples.
Hydrolysis
Heroin
+
6-Monoacetylmorphine
Tris(2,2‫މ‬-bipyridine)ruthenium(III)
Morphine
Acidic potassium
permanganate
hv Ȝmax 689 nm
hv Ȝmax 610 nm
Figure 7. Concept for the rapid chemiluminescence screening test for heroin
(adapted from Agg et al. [158]).
25
5. Project aims
This thesis describes a systematic approach to the design, evaluation and
optimisation of novel flow-cells for chemiluminescence detection in both flow
analysis and HPLC. Various approaches were employed for improving the sensitivity
of several analytically important reactions utilised for the determination of
biologically significant thiols and disulfides as well as forensically important opiate
alkaloids.
26
CHAPTER TWO
CHAPTER TWO
Solution mixing in flow-cells for
chemiluminescence detection
x
Introduction
x
Experimental
x
Results and discussion
x
Conclusion
27
CHAPTER TWO
1. Introduction
Chemiluminescence detection has been used extensively in procedures based on
flow injection analysis (FIA) and sequential injection analysis (SIA) methodology [2,
8, 9, 57, 159], due in part to the excellent sensitivity and wide calibration ranges that
have been obtained for diverse classes of analytes. Moreover, the instrumentation is
simple, essentially comprising a reaction vessel or conduit with a transparent surface,
mounted against a photodetector. The most commonly used configuration consists of
a coil of glass or polymer tubing (normally 0.5-1.0 mm i.d.) mounted against a
photomultiplier tube within a light-tight container. Solutions merge at a T- or
Y-shaped junction shortly before entering the coil.
The emission of photons from a chemiluminescence reaction is transient and
occurs at a rate that is dependent on both the kinetics of the chemical reaction and the
physical processes of solution mixing [8, 71]. For the greatest sensitivity, the
instrumental manifold and the flow-cell should be configured to maximise the
emission and detection of light when the reacting mixture passes through the cell [2].
For relatively fast chemiluminescence reactions, the analyte and reagent solutions
should merge at (or as close to) the point of detection. In addition, the dead volume
should be minimised to ensure rapid rinsing of the cell between analyses.
Etching or machining channels into glass/polymer materials and sealing the
channels with a transparent surface provides a convenient and reproducible
alternative that enables exploration of new materials or configurations to enhance
solution mixing, maximising the transfer of light to the photodetector [73, 160]. This
approach is also utilised to create chemiluminescence reaction zones within
microfluidic chips [161-164]. Previously reported configurations include linear
channels [165, 166], spirals (that mimic the traditional coil of tubing approach) [73,
28
CHAPTER TWO
160, 167-169], and ‘meandering’ or ‘serpentine’ designs comprising a linearly
propagating series of reversing turns [163, 170]. Spiral or meandering channels
consisting of linear segments between right-angled turns have also been reported
[161, 162, 164, 171, 172].
This chapter describes a novel chemiluminescence detector designed by
Global FIA to maximise both the generation and transmission of light from
chemiluminescence reactions. The key component of this new detector is a thin
Teflon disk with channels machined into one side. The Teflon disk is placed against
a sapphire window to form a flow channel with a flat transparent seal allowing
efficient transmission of light to the photomultiplier tube. The channel configuration
can be conveniently modified by replacing the Teflon disk, and as such, several
novel designs have been explored and compared with a wide range of analytically
useful chemiluminescence reactions, such as: morphine with acidic potassium
permanganate; vanilmandelic acid with luminol and hexacyanoferrate(III); codeine
with tris(2,2ƍ-bipyridine)ruthenium(III); and arginine, urea or ammonium with
hypobromite.
29
CHAPTER TWO
2. Experimental
2.1
Chemicals and reagents
Deionised water (Continental Water Systems, Victoria, Australia) and analytical
grade reagents were used unless otherwise stated. Chemicals were obtained from the
following sources:
L-arginine,
luminol, sodium polyphosphate (+80 mesh) and
vanilmandelic acid from Sigma-Aldrich (New South Wales, Australia); hydrogen
peroxide (3%) and potassium permanganate from Chem-Supply (South Australia,
Australia); bromine, lead dioxide, potassium ferricyanide, sodium hydroxide and
urea from Ajax Finechem (New South Wales, Australia); codeine and morphine from
GlaxoSmithKline (Victoria, Australia); sulfuric acid from Merck (Victoria,
Australia);
ammonium
chloride
from
BDH
(Poole,
England)
and
tris(2,2ƍ-bipyridine)ruthenium(II) dichloride hexahydrate from Strem Chemicals
(Minnesota, USA).
Stock solutions (1 × 10-3 M) of morphine and codeine were prepared in acidified
deionised water and sonicated to aid dissolution. Vanilmandelic acid (1 × 10-3 M)
was dissolved in aqueous sodium hydroxide (0.1 M), whilst ammonium chloride,
arginine and urea (1 × 10-3 M) were prepared daily in deionised water. When
required, the analytes were diluted into deionised water.
The acidic potassium permanganate reagent (1 × 10-3 M) was prepared by
dissolution of potassium permanganate in a 1% (m/v) sodium polyphosphate solution
and adjusted to pH 2.5 with sulfuric acid.
The luminol reagent (3.5 × 10-6 M) was prepared by dissolution of the solid in
deionised water, following the addition of a small volume of aqueous sodium
30
CHAPTER TWO
hydroxide (1.0 M). Potassium hexacyanoferrate(III) (potassium ferricyanide,
3 × 10-4 M) was prepared in sodium hydroxide (0.35 M).
To obtain intense luminol chemiluminescence suitable for visual examination, the
luminol reagent was prepared by dissolution of luminol (1.1 × 10-3 M) in sodium
hydroxide (0.1 M), whilst potassium hexacyanoferrate(III) (potassium ferricyanide,
3 × 10-2 M) was prepared in deionised water containing 4% (v/v) hydrogen peroxide
[173].
The tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent (1 × 10-3 M) was prepared by
dissolving tris(2,2‫މ‬-bipyridine)ruthenium(II) dichloride hexahydrate in dilute sulfuric
acid (0.05 M). The ruthenium(III) species was obtained by adding lead dioxide
(0.2 g per 100 mL) to the reagent solution and stirring for 5 minutes. Following the
characteristic change in colour from bright orange to emerald green, excess oxidant
was filtered (0.45 —m, Millipore MillexHN Nylon) from the solution prior to
analysis.
Hypobromite (0.08 M) was prepared daily by disproportionation of bromine in ice
cold sodium hydroxide (0.4 M). To prevent photodegradation of the reagent, the
Schott bottle was covered in aluminium foil.
2.2
Flow injection analysis
The FIA manifold was constructed from a Gilson Minipuls 3 peristaltic pump
(John Morris Scientific, Victoria, Australia) with bridged PVC pump tubing
(1.02 mm i.d.; DKSH, Queensland, Australia), black manifold tubing (0.76 mm i.d.,
Global FIA, Washington, USA) and six-port injection valve (Vici 04 W-0192L,
Valco Instruments, Houston, Texas, USA). The conventional and novel flow-cells
were mounted against the window of an extended range photomultiplier module
31
CHAPTER TWO
(Electron Tubes P30A-05, ETP, NSW, Australia) in light-tight housing. The output
from the photomultiplier captured by a chart recorder (YEW type 3066, Yokogawa
Hokushin Electric, Tokyo, Japan) or an ‘e-corder 410’ data acquisition system
(eDAQ, NSW, Australia).
For acidic potassium permanganate and hypobromite measurements, the analyte
standards were injected (70 —L) into a deionised water carrier stream that merged
with the chemiluminescence reagent shortly before, or within the flow-cell
(Figure 8). In the case of the ruthenium(III) complex, the reagent was injected into a
dilute sulfuric acid (0.05 M) carrier stream that merged with the analyte solution
shortly before or within the flow-cell. Solution flow rates were optimised over the
range 0.9 mL/min to 3.5 mL/min per line, for each chemiluminescence reaction.
Detector
Reagent
Flow-cell
Peristaltic
pump
PMT
Injection valve
Carrier
Waste
Analyte
Figure 8. Schematic of the FIA manifold
2.3
Sequential injection analysis
The SIA manifold was constructed from a milliGAT pump (model CP-DSM-GF,
Global FIA), ten-port multi-position valve (model C25Z, Valco) and black manifold
tubing (0.76 mm i.d., Global FIA). The conventional and novel flow-cells were
mounted against the window of an extended range photomultiplier module (Electron
Tubes P30A-05, ETP, NSW, Australia) in light-tight housing. A desktop computer
equipped with data acquisition board (LabJack U12, National Instruments, Victoria,
32
CHAPTER TWO
Australia) and LabView Software (Version 8.0, National Instruments) was used to
automate the instrument and acquire experimental data.
2.3.1
Conventional SIA mode
This automated system was programmed to aspirate and sequentially stack the
carrier (line 1, 2050 ȝL, 167 ȝL/s), analyte (line 2, 500 ȝL, 20 ȝL/s), and reagent
(line 3, 150 ȝL, 20 ȝL/s) into a holding coil (PTFE tubing, DKSH, 2.7 m × 0.8 mm
i.d.). The direction of flow was then reversed to propel the reacting mixture
(167 ȝL/s) through the detector (Figure 9).
Reagent
Analyte
Detector
Carrier
Flow-cell
Bidirectional
pump
PMT
Holding coil
Multi-position
valve
Waste
Figure 9. Schematic of the conventional SIA manifold
2.3.2
Direct aspiration mode
The detector was placed in-line between the pump and multi-position valve, so that
the solutions were sequentially aspirated directly through the flow-cell without the
reversal of solution flow (Figure 10). This results in minimal mixing of solutions
prior to entry into the chemiluminescence detector. This automated system was
programmed to repeatedly aspirate 150 PL of the reagent (valve position 1;
167 PL/s), 1500 PL of the analyte standard (position 2; 167 PL/s), and 1000 PL of
the carrier solution (position 3; 100 PL/s).
33
CHAPTER TWO
Reagent
Analyte
Detector
Carrier
Flow-cell
PMT
Multi-position
valve
Bidirectional
pump
Waste
Figure 10. Schematic of the SIA manifold in ‘direct aspiration mode.’
2.4
Flow-cells
Flow-cell A was constructed by mounting a tight coil (28 mm diameter) of
transparent PTFE tubing (0.8 mm i.d.; DKSH) onto a thin metal sheet (35 × 60 mm).
The tubing at the centre of the coil passed through a slit in the sheet and was
connected to a barbed plastic T-piece by slipping silicone tubing (1.02 mm i.d.;
DKSH) over both the tubing and the fitting. The distance from the confluence point
to the beginning of the coil was 1 cm (equivalent to approximately 5 PL).
A variety of flow-cell designs were explored by encasing different Teflon disks in
a ‘GloCel’ chemiluminescence detector (Global FIA; Figure 11). The path of
solution flow was set by channels (width 0.76 mm × depth 0.89 mm) machined into
each disk. Solution inlet and outlet ports in the back plate of the detector were
aligned with holes drilled though the disks. A sapphire window served as the
transparent top surface of the flow-cell. After an appropriate back plate was fastened,
an extended range photomultiplier module (Electron Tubes model P30A-05, ETP)
was secured in the holding chamber with a nut and ferrule. When the disks contained
a single solution inlet, the solutions were merged at a Y-piece outside the detector
and the length between the confluence point and detection was approximately 4 cm
34
CHAPTER TWO
(equivalent to approximately 20 PL). In cases where the disk contained two solution
inlets, the confluence point was located directly in front of the photomultiplier tube.
iii
ii
i
vi vii viii
iv v
Figure 11. GloCel chemiluminescence detector (Global FIA) consisting of (i) flowcell, (ii) PMT chamber, and (iii) nut and ferrule to secure the PMT and light-seal the
chamber. Flow-cell components: (iv) flat gasket, (v) sapphire window, (vi) Teflon
disk, (vii) ‘O’ ring and (viii) back plate (single-inlet design shown).
Flow-cell B was a Teflon disk with machined single-inlet spiral channel (0.76 mm
width × 0.89 mm depth, Figure 12i). The volume of the flow-cell was 275 —L.
Flow-cell C was a Teflon disk with machined single-inlet serpentine channel
(0.76 mm width × 0.89 mm depth) containing 114 reversing turns (Figure 12ii). The
volume of the flow-cell was 245 —L.
Flow-cell D was a modified version of flow-cell C, containing two solution inlets
at the centre of the Teflon disk (Figure 12iii) and an appropriately adjusted back
plate. The volume of the flow-cell was 235 —L.
Flow-cell E was constructed by replacing the Teflon disk with a spacer ring,
forming a shallow open cylindrical space (Figure 12iv). The back plate was equipped
with a custom designed Y-piece.
35
CHAPTER TWO
Flow-cell F was a modified version of flow-cell C, in which the rounded edges of
the single-inlet serpentine channel were replaced with approximately right-angled
turns.
Flow-cells G and H were identical to flow-cells C and D, except that the disks
were constructed from Teflon impregnated with 25% glass microspheres to improve
the transmission of light to the photomultiplier tube.
Flow-cell I, J and K were modified versions of flow-cell D containing mixing
zones located shortly after (I: 1.0 mm diameter, J: 2.0 mm diameter) or at the
confluence point (K: 4 mm diameter).
Flow-cell L was identical to flow-cell C, except that a 3 mm diameter well was
removed from the centre of the Teflon disk and equipped with an appropriately
adjusted back plate.
(i)
(ii)
(iv)
(iii)
Figure 12. Teflon disks with machined channels and inlet/outlet holes: flow-cells
(i) B, (ii) C, (iii) D and (iv) E.
With each change of flow-cell, the housing was re-sealed and the instrument was
left for 40 minutes to avoid the temporary increase in baseline signal that was
observed under the most sensitive settings.
36
CHAPTER TWO
3. Results and Discussion
3.1
Preliminary investigations into novel flow-cell configurations
The reaction of morphine with acidic potassium permanganate (in the presence of
sodium polyphosphates) was selected for the comparison of different flow-cell
configurations because it has been used extensively as a method of detection for flow
analysis and HPLC [13, 25, 174, 175]. It is also a representative example of the
relatively fast chemiluminescence reactions between strong oxidising agents and
organic analytes [13]. The emission from this particular reaction occurs as a result of
an electronically excited manganese(II) species and reaches maximum intensity
within a few seconds under stopped flow conditions [174].
With the use of FIA methodology, morphine standards (1 × 10-8 M to 1 × 10-5 M)
were injected into a deionised water carrier stream, which merged with the
permanganate reagent. The chemiluminescence intensities obtained using flow-cells
B, C and D were similar to those obtained using the conventional coiled-tubing
approach (flow-cell A). Flow-cell C gave the greatest signals, generally 6í11%
superior to flow-cells A and B (Figure 13.). Flow-cell E was found to be significantly
inferior in terms of chemiluminescence intensity (87% poorer than flow-cell A) and
precision; the relative standard deviation (R.S.D) for 10 replicate injections of
1 × 10-7 M morphine (9.6%) was much greater than that of the other configurations
(all < 1.5%).
37
CHAPTER TWO
8
Chemiluminescence intensity (mV)
7
6
5
4
3
2
1
0
A
B
C
D
E
Flow-cell
Figure 13. The effect of flow-cell configuration on chemiluminescence intensity
(peak height) during the oxidation of morphine (1 × 10-7 M) with acidic potassium
permanganate.
In contrast to the single-inlet cells (A, B, C and E, where mixing started at a T- or
Y-piece located within close proximity to the inlet), flow-cell D allowed the reaction
to be initiated directly in front of the photomultiplier tube. However, this design was
unable to significantly improve the chemiluminescence response. This was further
explored using the reactions between (i) tris(2,2ƍ-bipyridine)ruthenium(III) and
sodium
hydroxide
and
(ii)
luminol,
hydrogen
peroxide
and
potassium
hexacyanoferrate(III), which both produce intense emissions of light suitable for
visual observation (Figures 14 and 15). An examination of the chemiluminescence
emission within flow-cell C revealed that some light may have been emitted prior to
entry into the cell as demonstrated by the bright spot visible in the inlet hole
(Figures 14i and 15i). Despite this, the most intense emission occurred within the
first reversing turns of the serpentine design. In the case of flow-cell D, the two
38
CHAPTER TWO
solutions merged at the centrally located confluence point, delaying the onset of the
chemiluminescence reaction and as a result, the maximum emission occurred several
reversing turns away from this point (Figures 14ii and 15ii).
(ii)
(i)
Figure 14. Photographic evidence of solution mixing efficiency in flow-cells (i) C
and (ii) D, based on the distribution of light when solutions of
tris(2,2‫މ‬-bipyridine)ruthenium(III) and sodium hydroxide are continuously merged.
An exposure time of 20 s was used for both photographs.
(ii)
(i)
Figure 15. Photographic evidence of solution mixing efficiency in flow-cells (i) C
and (ii) D, based on the distribution of light when solutions of luminol, hydrogen
peroxide and potassium hexacyanoferrate(III) are continuously merged. An exposure
time of 0.3 s was used for both photographs.
Flow-cells B and C were also compared to the conventional coiled-tubing
approach (flow-cell A) using SIA methodology and the reaction of morphine
(1×10-7 M) with acidic potassium permanganate (5×10-4 M). Deionised water (line 1,
39
CHAPTER TWO
2050 ȝL, 167 ȝL/s), morphine (line 2, 500 ȝL, 20 ȝL/s), and the permanganate
reagent (line 3, 150 ȝL, 20 ȝL/s) were aspirated into the holding coil. The flow was
then reversed to propel the reacting mixture (167 ȝL/s) through the detector. A
comparison of the chemiluminescence signals obtained with flow-cells A, B and C
(Figure 16, blue columns) revealed that the greatest signals were obtained from
flow-cell A (15-35% superior to flow-cells B and C).
The manifold was then reconfigured (Figure 10) so that the solutions were
aspirated directly through the detector, without reversing the flow of solution. This
configuration is well-suited to chemiluminescence detection that involves relatively
fast reactions, as mixing of solutions prior to entry into the flow-cell is minimised,
enabling a greater proportion of the emitted light to be detected. The permanganate
reagent (5 × 10-4 M; line 1, 150 ȝL, 167 ȝL/s) was aspirated prior to morphine
(line 2, 1500 ȝL, 167 ȝL/s) and deionised water (line 3, 1000 ȝL, 100ȝL/s). Nine
standard solutions (between 1 × 10-9 and 1 × 10-5 M) were tested using the three
different flow-cells (A, B and C, Figure 16, pink columns). At equivalent
concentrations of morphine, the signals obtained with flow-cell C were greater than
those from flow-cells A and B by 7í23% and 75í104% respectively.
40
CHAPTER TWO
Chemiluminescence signal (peak area/arbitrary units)
50
40
30
20
10
0
A
B
C
Flow-cell
Figure 16. The effect of channel design on chemiluminescence signal (peak area)
using traditional SIA (blue columns) and direct aspiration SIA methodologies (pink
columns) during the oxidation of morphine (1 × 10-7 M) with acidic potassium
permanganate.
3.2
Further investigations into channel design
Several research groups have utilised chemiluminescence reaction zones consisting
of meandering or spiral channels with right-angled turns (in flow-cells or
microfluidic devices) [161, 162, 164, 171, 172]. A flow-cell comprising a spirally
propagating serpentine channel with approximately 90° turns (hereafter referred to as
flow-cell F), was compared with three previously described single-inlet flow-cell
configurations, A, B (Figure 12i) and C (Figure 12ii).
Using FIA methodology, six morphine standards (between 1 × 10-10 M and
1 × 10-5 M) were injected into a deionised water carrier stream that merged with the
permanganate reagent (1 × 10-3 M) at an external confluence point located within
close proximity (4 cm, approximately 20 —L) to the inlet of the flow-cells. The
41
CHAPTER TWO
chemiluminescence intensities obtained using flow-cell C were again greater than the
other configurations, although the difference across all of the cells was generally less
than 10% (Figure 17, blue columns). Using SIA (direct aspiration of 150 —L of the
permanganate reagent, 1500 —L of the morphine standard, and then 1000 —L of the
deionised water carrier), the same trend was observed, but the differences in intensity
were much greater (Figure 17, pink columns). Over five orders of magnitude
(1 × 10-9 M to 1 × 10-5 M), the signals obtained with flow-cell C were significantly
14
70
12
60
10
50
8
40
6
30
4
20
2
10
0
SIA chemiluminescence signal (arbitrary units)
FIA chemiluminescence intensity (mV)
greater than those from A (4-56%), B (51-76%) or F (11-103%).
0
A
B
C
F
Flow-cell
Figure 17. The effect of channel design on chemiluminescence signals obtained
with (i) FIA (peak height, blue columns) and (ii) direct aspiration SIA
methodologies (peak area, pink columns) during the oxidation of morphine
(1 × 10-7 M) with acidic potassium permanganate. R.S.D. was less than 1.4% for
each flow-cell.
It appears that although the repeated approximately 90º turns of flow-cell F
promote more efficient mixing than the steady curve of the spiral configuration
(flow-cell B), the relatively sharp corners also create areas of poorer solution flow,
resulting in inferior emission intensities compared to the serpentine design. A visual
42
CHAPTER TWO
examination of the chemiluminescence generated in flow-cell F revealed that air
bubbles in the solution tended to remain in the corners of the channel, a problem also
encountered in flow-through Z- or U-cells for photometric detection [176]. This
problem
became
more
apparent
when
using
the
relatively
long-lived
chemiluminescence reaction between luminol, hydrogen peroxide and potassium
hexacyanoferrate(III), which produces nitrogen gas [173] (Figure 18).
Figure 18. Photographic evidence of bubble formation within flow-cell F when
solutions of luminol, hydrogen peroxide and potassium hexacyanoferrate(III) are
continuously merged.
3.3
Flow-cell material
Although chemiluminescence flow-cells and microfluidic reactors have often been
constructed by machining channels into glass [165, 170, 171] or transparent polymer
materials [163, 164, 166-168, 172], an investigation was undertaken to determine if
the material used to construct these novel flow-cells had any impact on the
proportion of light that can be captured by the photomultiplier tube. Using FIA
methodology, a disk constructed from Teflon impregnated with 25% glass
microspheres (flow-cell G) was compared to one of identical design constructed from
standard Teflon (flow-cell C) and the conventional coiled-tubing (flow-cell A). The
chemiluminescence reactions selected to examine these flow-cells were: (i) the
43
CHAPTER TWO
oxidation of morphine with acidic potassium permanganate [13] and (ii) the
enhancement of luminol and hexacyanoferrate(III) by vanilmandelic acid [177]. The
R.S.D for ten replicate injections of the 1 × 10-7 M morphine standard solution was
below 0.9% for all flow-cells. Across a wide range of analyte concentrations,
13-17% more light was captured using the Teflon disk impregnated with glass
microspheres (Figure 19, blue columns). A similar trend was observed for the
reaction between vanilmandelic acid and luminol, with signals from flow-cell G on
average 14-27% superior to the other flow-cells (Figure 19, pink columns). This is
attributed to the greater reflection of stray light towards the photodetector.
Permanganate chemiluminescence intensity (mV)
12
500
10
400
8
300
6
200
4
100
2
0
Luminol chemiluminescence intensity (mV)
600
14
0
A
C
G
Flow-cell
Figure 19. The effect of the material used in flow-cell fabrication on
chemiluminescence intensity (peak height) for the reactions between (i) morphine
(1 × 10-7 M) and acidic potassium permanganate (blue columns) and (ii)
vanilmandelic acid (1 × 10-5 M) and luminol (pink columns).
3.4
Confluence point and mixing chamber
The position of the confluence point is crucial for chemiluminescence detection in
FIA or HPLC methodologies that involve fast light-producing reactions. It is thought
44
CHAPTER TWO
that ideally, the solutions should merge as close as possible to the point of detection
[53, 147, 178-180]. Despite the design of flow-cell D enabling solutions to be
merged directly in front of the photodetector window, this design was unable to
improve the transmission of light to the detector, even for the relatively fast
oxidation of morphine with permanganate (Figure 13). This design was further
explored by; (i) introducing mixing zones into the dual-inlet design, (ii) creating a
larger mixing well in the centre of the detection zone by modifying both the disk and
backing plate, and (iii) examining both the previously constructed and new flow-cell
inserts with a wider range of rapid chemiluminescence reactions.
The modified dual-inlet disks had mixing zones of located shortly after the
confluence point (1.0 mm diameter; flow-cell I, 2.0 mm diameter; flow-cell J), or at
the confluence point (4.0 mm diameter; flow-cell K). The flow-cells were compared
to the standard single-inlet (C) and dual-inlet (D) configurations using FIA
methodology with the analytically useful reactions of: morphine with permanganate
[25, 73]; codeine with tris(2,2ƍ-bipyridine)ruthenium(III) [25, 181]; and arginine,
ammonium chloride and urea with sodium hypobromite [168, 182-185]. The
production of the emitting species in these reactions is rapid; with the maximum
chemiluminescence intensity being obtained within a few microseconds to several
seconds, under stopped flow conditions [41, 168, 174, 186]. For the reaction of
morphine and permanganate, the greatest intensity was obtained with flow-cell C
(Figure 20, blue columns). Interestingly, the chemiluminescence signal decreased as
the size of the mixing zone was increased (a 13% to 16% reduction in light,
compared to flow-cell C).
To overcome the poor temporal stability associated with using chemically oxidised
tris(2,2ƍ-bipyridine)ruthenium(III), the reagent was oxidised immediately prior to
45
CHAPTER TWO
each flow-cell test. Furthermore, to account for variations in signal intensity due to
changes in the tris(2,2‫މ‬-bipyridine)ruthenium(III) and hypobromite reagent
concentrations over the day, each experiment was also simultaneously performed
with a second FIA instrument (containing a coiled-tubing chemiluminescence
detector) to obtain a relative response. A similar trend was observed across the six
flow-cell configurations for the reaction between codeine and tris(2,2ƍbipyridine)ruthenium(III). Flow-cell C again resulted in the greatest signal intensities
and proved to be 10-19% superior to the modified dual-inlet flow-cells (I, J and K,
Figure 20, pink columns).
3.5
3.0
8
2.5
6
2.0
1.5
4
1.0
2
0.5
Relative chemiluminescence intensity for
tris(2,2'-bipyridine)ruthenium(III) (mV)
Permanganate chemiluminescence intensity (mV)
10
0.0
0
A
C
D
I
J
K
Flow-cell
Figure 20. The effect of mixing zones on chemiluminescence intensity (peak height)
for the chemiluminescence reactions of (i) morphine with acidic potassium
permanganate (blue columns) and (ii) codeine with tris(2,2c-bipyridine)ruthenium(III)
(pink columns).
The reactions of ammonium, arginine and urea with hypobromite (Figure 21)
provided responses analogous to those obtained with the permanganate and
tris(2,2‫މ‬-bipyridine)ruthenium(III)
reaction
46
systems.
However,
the
CHAPTER TWO
chemiluminescence intensities from flow-cells C and D were very similar and in the
case of the reaction of ammonium with hypobromite, the greatest response was
obtained with the dual-inlet configuration. This suggests that flow-cell D may prove
to be advantageous for chemiluminescence reactions with extremely fast kinetics.
Relative chemiluminescence intensity (mV)
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
A
C
D
I
J
K
Flow-cell
Figure 21. The effect of mixing zones on chemiluminescence intensity (peak height)
for the reaction of ammonium chloride (blue columns), arginine (pink columns) and
urea (yellow columns) with hypobromite.
A visual examination of the light emitted from flow-cells D, J and K using the
reaction
between
tris(2,2ƍ-bipyridine)ruthenium(III)
and
sodium
hydroxide
(Figure 22) revealed that the addition of mixing chambers delayed the onset of
maximum emission (Figures 22v and vi). Furthermore, it is clear from the
photographs that the solutions rapidly move through the path of least resistance
within the mixing chambers, and only poor solution mixing occurred in surrounding
areas.
47
CHAPTER TWO
(i)
(iv)
(ii)
(v)
(iii)
(vi)
Figure 22. Teflon disks (for GloCel detector) containing dual-inlet serpentine
configuration channels with (i) no mixing zone (flow-cell D), (ii) 2.0 mm mixing
zone after confluence point (flow-cell J), and (iii) 4.0 mm mixing zone at confluence
point (flow-cell K). (iv-vi) Photographic evidence of solution mixing efficiency,
based on the distribution of light when solutions of sodium hydroxide and
tris(2,2‫މ‬-bipyridine)ruthenium(III) are continuously merged. The exposure times for
the three photographs were 0.8, 0.8 and 2.5 s, respectively.
48
CHAPTER TWO
To enhance solution mixing within a central reaction chamber, a 3 mm diameter
well was removed from the centre of the serpentine flow-cell (flow-cell L) and two
new back plates with integrated confluence points were created. The first contained
angled inlets that met in a shallow well (4 mm diameter, 2.5 mm depth including
both disk and back plate). The second contained tangential inlets located on opposing
sides of the well with grooves to encourage vortex flow. This approach was similar
to the mixing device employed by Kobayashi and Imai [187] to merge the reactant
solutions of a peroxyoxalate chemiluminescence system prior to detection. However,
under the stated conditions, these designs did not improve the chemiluminescence
response (13% and 9% poorer than flow-cell C for the reaction of morphine with
permanganate), presumably due to similar issues as those described above.
49
CHAPTER TWO
4. Conclusions
The exchange of disks in the GloCel detector provides a convenient way to
optimise the flow-cell configuration for any particular chemiluminescence reaction.
In general, the spirally propagating serpentine channel configuration (flow-cell C)
was superior to the spiral (flow-cell B) and the square-serpentine designs
(flow-cell F) in terms of generating the most intense emission. Even for relatively
fast chemiluminescence reactions, the optimum position of the confluence point is
often external to the detection zone. However, for some reactions, a dual-inlet
configuration (flow-cell D) that enabled the solutions to be merged directly in front
of the photodetector provided greater emission intensities. The introduction of larger
chambers in the centre of the detection zone did not enhance the signal for any of the
reactions investigated in this study. Disks constructed from Teflon impregnated with
glass microspheres increased the quantity of light reaching the photodetector,
highlighting the importance of flow-cell materials in minimising the loss of stray
light.
The development of flow-cells that efficiently mix solutions and capture a greater
proportion of the emitted light from rapid reactions will be particularly useful for
chemiluminescence detection in HPLC. This will allow for the maximum emission to
be obtained within a very small volume of solution, minimising the contribution of
band broadening. The findings of this chapter have resulted in a recent publication
[188].
50
CHAPTER THREE
CHAPTER THREE
Enhanced permanganate chemiluminescence
x
Introduction
x
Experimental
x
Results and discussion
x
Conclusion
51
CHAPTER THREE
1. Introduction
The emission of light from chemical reactions can be utilised for sensitive
detection with relatively simple instrumentation in various flow-based analytical
techniques [5, 13, 31, 43, 189-191]. However, its application is limited by the
relatively small number of reactions with sufficient chemiluminescence quantum
yields to be used under the conditions required for routine analysis. The intensity of
the emission is determined not only by the concentration of the molecules and the
overall quantum yield, but also by the reaction kinetics [68]. This aspect is crucial for
chemiluminescence detection as analytes are merged with a reagent to initiate the
chemiluminescence reaction, and only a small proportion of this solution is exposed
to the photodetector at any particular moment [122, 192]. Ideally, the reaction should
be sufficiently fast so that the majority of the emitted light can be captured before the
analyte zone is flushed out of the detector.
One reagent that meets these requirements is acidic potassium permanganate,
which over the last few decades has become widely used for detection in flow
analysis based techniques [13]. The oxidation of various organic analytes with acidic
potassium permanganate is known to be autocatalytic [193, 194] – the rate of
reaction is influenced by the concentration of reaction intermediates or
products - which has been attributed to the formation of manganese(II) and colloidal
manganese(IV) [195-197].
Several researchers have described a change in permanganate chemiluminescence
intensity or reaction rate in the presence of manganese(II) salts [15, 198-201]. Agater
and co-workers added 200 mM manganese(II) to catalyse the slow oxidation of
mono- and di-saccharides. The lowest analyte concentration that they examined was
1 ×10-4 M with precipitation of manganese dioxide occurring at concentrations of
52
CHAPTER THREE
manganese(II) higher than 200 mM [198]. Zhu and co-workers determined
manganese(II) based on its influence on the rate of the reaction between
permanganate and 2,3-butanedione [199]. Townshend and co-workers added 1 mM
manganese(II) to the carrier stream of their flow injection analysis system, which
they found increased signal intensity by 16% for one analyte, but decreased the
signal by 51% and 80% for two other analytes [15, 200, 201]. However, at that time,
the emitting species in these reactions had not yet been confirmed [17, 202] (the
luminescence was often erroneously attributed to the production of singlet oxygen
[13, 15]) and the full implications of these observations were not apparent.
This chapter describes a systematic investigation into the relationship between the
addition of reducing agents to freshly prepared acidic potassium permanganate
solutions and the resulting impact on the intensity and rate of chemiluminescence
reactions. These reagents were evaluated in terms of their stability and capacity to
elicit analytically useful chemiluminescence in comparison with a conventional
acidic potassium permanganate solution using a number of flow analysis techniques.
53
CHAPTER THREE
2. Experimental
2.1
Chemicals and reagents
Deionised water (Continental Water Systems, Victoria, Australia) and analytical
grade reagents were used unless otherwise stated. Chemicals were obtained from the
following
sources:
3-aminophenol,
amoxicillin,
caffeic
acid,
dopamine
hydrochloride, fenoterol hydrobromide, gallic acid, homovanillic acid, 4-nitrophenol,
octopamine, quercetin, resveratrol, serotonin hydrochloride, sodium polyphosphate
(+80 mesh), sodium thiosulfate, synephrine, tyramine, L-tyrosine and vanilmandelic
acid from Sigma-Aldrich (New South Wales, Australia); ascorbic acid,
manganese(II) sulfate and vanillin from BDH (Poole, England); potassium
permanganate from Chem-Supply (South Australia, Australia); codeine, morphine,
oripavine and pseudomorphine from GlaxoSmithKline (Victoria, Australia) and
sulfuric acid from Merck (Victoria, Australia).
The ‘standard’ permanganate reagent was prepared by dissolution of potassium
permanganate (1 × 10-3 M) in 1% (m/v) sodium polyphosphate and adjusting to
pH 2.5 with sulfuric acid.
The ‘enhanced’ permanganate reagent was prepared by dissolution of potassium
permanganate (1.0 × 10-3 M) in 1% (m/v) sodium polyphosphate and adjusting to
pH 2.5 with sulfuric acid and then adding either solid manganese(II) sulfate or
sodium thiosulfate, using a small volume of a 0.1 M solution.
Stock solutions (1 × 10-3 M) of the analytes were prepared and diluted in deionised
water and sonicated to aid dissolution. Opiate alkaloid stocks (1 × 10-3 M) were
prepared in acidified deionised water.
54
CHAPTER THREE
2.2
Flow injection analysis
The FIA manifold was constructed from a Gilson Minipuls 3 peristaltic pump
(John Morris Scientific, Victoria, Australia) with bridged PVC pump tubing
(1.02 mm i.d.; DKSH, Queensland, Australia), black manifold tubing (0.76 mm i.d.,
Global FIA, Washington, USA), six-port injection valve (Vici 04 W-0192L, Valco
Instruments, Houston, Texas, USA) and GloCel chemiluminescence detector
(Global FIA, Washington, USA) equipped with either flow-cell C (single-inlet
serpentine) or D (dual-inlet serpentine) and mounted flush against the window of an
extended range photomultiplier module (Electron Tubes P30A-05, ETP, New South
Wales, Australia). All tubing entering and exiting the detector was black PTFE
(0.76 mm i.d., Global FIA). The output signal from the detector was recorded with
an ‘e-corder 410’ data acquisition system (eDAQ, New South Wales, Australia).
The analyte standards were injected (70 —L) into a deionised water carrier stream
that merged with the chemiluminescence reagent, shortly before or within the
flow-cell (Figure 23). Solution flow rates were optimised over the range 0.9 mL/min
to 3.5 mL/min per line.
Detector
Reagent
Flow-cell
Peristaltic
pump
PMT
Injection valve
Carrier
Waste
Analyte
Figure 23. Schematic of the FIA manifold.
55
CHAPTER THREE
2.3
Flow-cells
The single-inlet and dual-inlet serpentine flow-cells (C and D, Figures 12ii and
12iii) were utilised in this study. They were encased in a GloCel chemiluminescence
detector (as described in Chapter 2, section 2.4).
2.4
Stopped flow
Stopped flow analysis experiments were performed with an FIA manifold
constructed from a programmable dual-syringe pump (Model sp210iw, World
Precision Instruments, Victoria, Australia), Valco six-port injection valve, and
GloCel chemiluminescence detector (Global FIA) equipped with either flow-cell C
or flow-cell D (Figure 24). The carrier and reagent syringes (Terumo, 10 mL Luer
lock) were loaded with deionised water and the permanganate reagent. After the
injection loop was filled with the analyte solution, the pump was activated.
Equivalent, precise volumes of the carrier and reagent solutions were dispensed,
which propelled the analyte and reagent into the serpentine reaction channel, where it
was held for a set period of time to obtain the entire chemiluminescence intensity
versus time profile. The output signal from the photomultiplier module was recorded
with an ‘e-corder 410’ data acquisition system (eDAQ, NSW, Australia).
Longitudinal dispersion of the analyte zone was minimised using the shortest
possible length of tubing between the valve and detector. By inserting a small
volume of analyte solution into a carrier stream, rather than dispensing an analyte
stream from the syringe pump, convenient and thorough flushing of the detector was
achieved. Activating the pump for an extended period of time without filling the
injection loop) after each profile was collected, ensured that the manganese(II)
product of the reaction did not affect subsequent tests.
56
CHAPTER THREE
Detector
Flow-cell
Reagent
PMT
Injection valve
Carrier
Syringe pump
Waste
Analyte
Figure 24. Schematic of the stopped flow manifold
2.5
Sequential injection analysis
Analyte standards were reacted with the permanganate reagent by sequential
aspiration of solutions through a multi-position valve (model C25Z, Valco) and the
GloCel detector equipped with flow-cell C, using a milliGAT pump (Global FIA,
Figure 25). A desktop computer equipped with data acquisition board (LabJack U12,
National Instruments, Victoria, Australia) and LabView Software (Version 8.0,
National Instruments) was used to control the pump, valve, and record the output
from the photomultiplier module. This automated system was programmed to
repeatedly aspirate 150 PL of reagent (valve position 1; 167 PL/s), 1500 PL of the
analyte standard (position 2; 167 PL/s), and 1000 PL of deionised water (position 3;
100 PL/s), with a 60 min pause after every fifth cycle.
Reagent
Analyte
Detector
Carrier
Flow-cell
PMT
Multi-position
valve
Bidirectional
pump
Waste
Figure 25. Schematic of the SIA manifold.
57
CHAPTER THREE
2.6
UV-Visible spectrometry
UV-Visible absorbance spectra were collected using a Cary 300 Bio UV-visible
spectrophotometer (Varian, Victoria, Australia) with 10 mm path length, sealable
quartz cuvettes (Starna, New South Wales, Australia). A solution of acidic potassium
permanganate (1 × 10-3 M) was prepared, and immediately following the addition of
sodium thiosulfate (0.6 mM), the reagent was transferred into a cuvette within the
spectrophotometer, where absorption spectra (wavelengths between 300-700 nm)
were taken once every hour for 48 hours.
58
CHAPTER THREE
3. Results and Discussion
3.1
Manganese(II)
3.1.1
Preliminary chemiluminescence investigations
Acidic potassium permanganate solutions containing sodium polyphosphate have
limited stability due to both the degradation of the oxidant [203] and a pH dependent
hydrolysis of the polyphosphate [204], a process which is accelerated by exposure to
natural light. The stability of a ‘standard’ permanganate reagent (1.0 mM, 1% m/v
polyphosphate, pH 2.5) was assessed using automated SIA methodology. The
reagent, analyte (morphine, 1 × 10-6 M) and deionised water carrier were sequentially
aspirated through flow-cell C, with a 60 minute pause after every fifth cycle. This
approach minimises the mixing of solutions prior to entry into the flow-cell, enabling
detection of a greater proportion of the emitted light. Over a 48 hour period, a 26%
increase in chemiluminescence intensity was observed (R.S.D. of the 49 sets of 5
replicates was 5.1%).
Chemiluminescence signal (arbitrary units)
300
250
200
150
100
50
0
0
6
12
18
24
30
36
42
48
Time (h)
Figure 26. Stability of the standard permanganate reagent (1.0 mM) with morphine
(1× 10-6 M), over 48 hours with SIA methodology.
59
CHAPTER THREE
Slezak et al. [205] recently demonstrated that acidic potassium permanganate
solutions exposed to natural light for up to six months can improve the
chemiluminescence emissions for several simple phenols. In this case, significant
degradation of both the oxidant and enhancer (polyphosphates) was counteracted by
the manganese(II) catalysis of the light-producing pathway resulting in dramatic
increases in the reaction kinetics for these organic analytes [205].
The addition of manganese(II) sulfate to the standard reagent results in the partial
reduction of permanganate to form a concomitant mixture of both manganese(VII)
and manganese(III) [205]. This effectively replicates the reduction process outlined
by Slezak, without the lengthy time frame. The presence of polyphosphates in the
reagent is crucial in preventing the formation and subsequent flocculation of a murky
red manganese dioxide suspension which is characteristic of the Guyard reaction
[206].
The viability of this method of reagent preparation was assessed via repeated
reaction of morphine (1 × 10-6 M) with a reagent containing 1.0 mM permanganate
and 0.6 mM manganese(II) sulfate over 72 hours using SIA methodology. An
immediate 3-fold increase in chemiluminescence intensity was observed after the
first hour, which became even greater (8.3-fold) 24 hours after reagent preparation
(Figure 27). During the ensuing 48 hours, minimal variation in signal intensity was
observed (R.S.D. was 1.1%), and therefore all subsequent experiments were
performed using reagents left to equilibrate for 24 hours. The significant increase in
chemiluminescence intensity resulted in this reagent being denoted ‘enhanced’
permanganate.
Townshend and co-workers [15, 200, 201] have previously reported relatively
small levels of enhancement (or even inhibition) of the chemiluminescence response
60
CHAPTER THREE
for several analytes when manganese(II) was used in flowing systems. However, this
may have been attributable to merging solutions of manganese(II), permanganate
(without polyphosphates) and the analyte online (which allowed very little time for
the equilibration of manganese species).
Chemiluminescence signal (arbitrary units)
1800
1600
1400
1200
1000
800
600
400
200
0
0
8
16
24
32
40
48
56
64
72
Time (h)
Figure 27. Stability of the enhanced permanganate reagent (1.0 mM containing 0.6 mM
manganese(II) sulfate) with morphine (1× 10-6 M), over 48 hours with SIA
methodology.
3.1.2
Optimisation of reagent conditions
An evaluation of reagent conditions was then conducted using FIA methodology to
determine the optimum concentration of manganese(II) required to produce the
greatest enhancement in chemiluminescence response with two model analytes,
morphine and the adrenergic amine synephrine.
A set of six reagents were prepared, containing 1.0 mM permanganate, 1% (m/v)
sodium polyphosphate, adjusted to pH 2.5 using sulfuric acid. Concentrations of
manganese(II) ranged from 0.0 mM to 1.0 mM in each of the reagents. The optimal
concentration of manganese(II) appeared to be analyte dependent (Figure 28). A
61
CHAPTER THREE
moderate increase in chemiluminescence intensity was obtained for morphine
(7-fold), whilst a significant 80-fold increase was observed for synephrine.
Chemiluminescence intensity (mV)
2400
(i)
2000
1600
1200
800
400
0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Manganese(II) sulfate concentration (mM)
Chemiluminescence intensity (mV)
160
(ii)
140
120
100
80
60
40
20
0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Manganese(II) sulfate concentration (mM)
Figure 28. The effect of manganese(II) sulfate concentration on the
chemiluminescence response (peak height) obtained from the reaction of acidic
potassium permanganate (1.0 mM) with (i) morphine (1 × 10-5 M) and (ii)
synephrine (1 × 10-5 M) using FIA.
62
CHAPTER THREE
The greater chemiluminescence intensities obtained for both morphine and
synephrine with the enhanced reagent also translated into much lower limits of
detection for both analytes, 4.6 × 10-11 M and 1.2 × 10-9 M, respectively. This
represents over an order of magnitude greater sensitivity than that obtained using the
standard reagent (Tables 1 and 2) and is superior to all previously reported detection
limits obtained with acidic potassium permanganate for these analytes [13, 21, 22].
Table 1. Analytical figures of merit obtained for morphine with the standard and
enhanced permanganate reagents.
a
Reagent
Calibration function
(peak height, mV)
R2
LOD
(M)a
R.S.D.
%*
Standard
permanganate
y = 0.928x + 8.588
0.9962
5.4 × 10-10
1.3
Enhanced
permanganate
y = 0.847x + 8.512
0.9985
4.6 × 10-11
0.69
Calculated as 3× the standard deviation of the blank response.
*1 × 10-7 M (n = 10)
Table 2. Analytical figures of merit obtained for synephrine with the standard and
enhanced permanganate reagents.
a
Reagent
Calibration function
(peak height, mV)
R2
LOD
(M) a
R.S.D.
%*
Standard
permanganate
y = 0.962x + 7.538
0.9989
1.6 × 10-8
0.89
Enhanced
permanganate
y = 0.721x + 6.397
0.9993
1.2 × 10-9
0.58
Calculated as 3× the standard deviation of the blank response.
* 1 × 10-6 M (n = 10)
3.1.3
Analyte screening test
Sixteen compounds from four important classes of analytes (antioxidants [19, 20],
neurotransmitters [24, 207], adrenergic amines [21, 22] and opiate alkaloids [27,
153]) that have been previously detected with acidic potassium permanganate were
63
CHAPTER THREE
selected to examine the influence of the enhanced reagent on chemiluminescence
signals (Figure 29).
Figure 29. Structures of the sixteen compounds (antioxidants, neurotransmitters,
adrenergic amines and opiate alkaloids) used in the reagent screening test.
Moderate increases in chemiluminescence intensity (less than an order of
magnitude) were obtained for phenols containing ortho-hydroxy or -alkoxy groups
64
CHAPTER THREE
(such as morphine, dopamine, resveratrol, quercetin and caffeic acid) when using the
enhanced reagent (Figure 30 and Table 3). Significant enhancement (between
20- and 95-fold) was obtained for the simple phenol hordenine and three closely
related compounds (octopamine, synephrine and tyramine). The enhanced reagent
affords a significant improvement in sensitivity, which may allow for the detection of
compounds not traditionally examined with acidic potassium permanganate
chemiluminescence.
1100
Chemiluminescence intensity (mV)
1000
900
800
700
600
500
400
300
200
100
Re
sv
e
Q r ato
ue l
Ga rce
l ti
Ca lic n
f f e aci
d
V
a n D ic a
ilm op cid
H an am
om d
i
ov elic ne
a n ac
ill id
ic
S e aci
O roto d
ct n
ap i n
S y am
n e ine
p
T y hrin
ra e
H mi
or ne
de
n
Co i n e
d
M ein
or e
P s O phi
eu rip ne
do av
m in
or e
ph
in
e
0
Analyte
Figure 30. Chemiluminescence intensities (peak heights) for selected analytes with
standard permanganate (1.0 mM; blue columns) and enhanced permanganate
(1.0 mM containing 0.6 mM manganese(II) sulfate; pink columns) using FIA.
3.1.4
Stopped flow analysis
Stopped flow analysis was used to obtain the chemiluminescence intensity versus
time profile for synephrine (1 × 10-5 M) using the standard reagent (Figure 31).
65
CHAPTER THREE
Despite acidic potassium permanganate having previously been applied for the
post-column determination of synephrine [21, 22], the kinetics of this reaction are
not ideal for flow-through detection (Figure 31i). The maximum emission intensity
occurred between 10 and 35 seconds, and did not decay to half its maximum until
40-80 seconds. However, using an enhanced reagent containing 0.1 mM
manganese(II), the time taken to reach maximum intensity decreased to 2.4 seconds
(Figure 31ii). Increasing the manganese(II) concentration between 0.3 mM and
1.0 mM (Figure 31iii and Figure 31iv) resulted in small additional increases in
reaction rate, which were accompanied by improvements in maximum intensity
(peak height), but decreases in the total chemiluminescence emission (peak area)
(Figure 32).
400
iv
Chemiluminescence intensity (mV)
350
300
iii
250
200
ii
150
100
50
i
0
0
5
10
15
20
25
Time (s)
Figure 31. Chemiluminescence intensity versus time profiles for the reaction of
synephrine (1 × 10-5 M) with acidic potassium permanganate containing (i) 0.0 mM,
(ii) 0.1 mM, (iii) 0.3 mM, and (iv) 0.6 mM manganese(II) sulfate.
66
CHAPTER THREE
1400
ii
350
1200
300
1000
250
i
800
200
600
150
400
100
Total cehmiluminescence emission (mVs)
Maximum chemiluminescence intensity (mV)
400
200
50
0
0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Manganese(II) sulfate concentration (mM)
Figure 32. Effect of manganese(II) sulfate concentration on (i) maximum
chemiluminescence intensity (peak height), and (ii) total emission (peak area) for the
reaction of acidic potassium permanganate with synephrine (1 × 10-5 M).
The distribution of light within the single inlet serpentine reaction channel
(flow-cell C) was examined by continuously merging the reactant solutions at the
external confluence point and capturing the emitted light with a digital SLR camera
(Figure 33). The chemiluminescence emission from the reaction between the
standard reagent and synephrine (1 × 10-3 M) was spread throughout the flow-cell,
with a slightly more intense emission emanating from the outer coil of the serpentine
channel (Figure 33i). In agreement with the intensity versus time profile (Figure 31i),
it appears that the reaction mixture would continue to emit light for a significant
period of time after exiting the detection zone. In contrast, use of the enhanced
reagent resulted in an intense emission in the centre of the detection zone
(Figures 33ii and 33iii), which decreased considerably before the solution exited the
cell.
67
CHAPTER THREE
Although the reaction between morphine and the standard permanganate reagent is
already rapid (Figure 33iv), an increase in reaction rate was observed upon reaction
with the enhanced reagent, producing more intense chemiluminescence in the inner
coil of the serpentine channel (Figures 33v and 33vi).
(i)
(ii)
(iii)
(iv)
(v)
(vi)
Figure 33. Chemiluminescence emissions occurring within flow-cell C of the
GloCel detector. Photographs (i), (ii), and (iii) show the reaction between synephrine
(1 × 10-3 M) and acidic potassium permanganate, with 0.0 mM, 0.3 mM, and
0.6 mM manganese(II) sulfate, respectively, using an exposure time of 5 min.
Photographs (iv), (v), and (vi) show the reaction of morphine (1 × 10-3 M) under the
same three reagent conditions and an exposure time of 2.5 s.
3.1.5
Flow-cell comparisons
As discussed in chapter 2, the design of chemiluminescence detectors is crucial in
obtaining optimal mixing efficiency to maximise the chemiluminescence response
[73, 188]. Despite morphine being an exemplar of the many fast chemiluminescence
reactions between oxidising agents and organic analytes [13], a flow-cell designed to
allow the reaction to be initiated in front of the photomultiplier tube (flow-cell D)
failed to improve the chemiluminescence response. However, the reaction of
hypobromite and ammonium chloride has extremely rapid reaction kinetics and in
68
CHAPTER THREE
this instance, flow-cell D proved superior (by 3%) to an approach utilising an
external confluence point (flow-cell C). The results obtained for synephrine using
stopped flow analysis indicated that considerable increases in reaction rate and
chemiluminescence intensity were obtained using the enhanced reagent. To examine
the influence of increased reaction rates on the position of maximum
chemiluminescence emission, the aforementioned flow-cells (C and D) were
compared to the conventional coiled-tubing approach (flow-cell A) using FIA
methodology.
Unlike the standard permanganate reagent (Figure 34, blue columns), the reaction
of morphine with the enhanced reagent produced greater signals with flow-cell D
(reaction initiated in front of the detector) than the other configurations (A or C), by
27% and 14% respectively (Figure 34, pink columns).
28
Chemiluminescence intensity (mV)
24
20
16
12
8
4
0
A
C
D
Flow-cell
Figure 34. Comparison of chemiluminescence flow-cells for the reaction of
morphine (1 × 10-7 M) with (i) standard permanganate (blue columns) and (ii)
enhanced permanganate (pink columns) using FIA. The R.S.D. (n = 10) for each
configuration was less than 1.8%.
69
CHAPTER THREE
Under stopped flow conditions, the chemiluminescence intensity versus time
profile for the reaction of morphine (1 × 10-5 M) with the standard reagent reached a
maximum after 2 seconds with both flow-cells C and D. The overall signal (peak
area) obtained using flow-cell C with an external Y-piece was 5% greater than that of
flow-cell D (Figure 35i, blue trace). The enhanced reagent halved the time taken to
reach maximum emission. Using flow-cell D, an 18% increase in area and a 260%
increase in height was obtained (Figure 35ii, pink trace). This is attributable to the
dual-inlet design allowing initiation of reaction in front of the photomultiplier tube
and the increased reaction kinetics. The increases were smaller for flow-cell C: 1.3%
and 180%, respectively, indicating that in this configuration, a significant proportion
of the emission (approximately 15%) occurred before the reacting mixture entered
the detection zone (Figure 35ii, blue trace). It was concluded that flow-cell D was the
most suitable configuration for reactions with such rapid reaction kinetics and it was
therefore used for all future experiments with the enhanced reagent.
4000
Chemiluminescence intensity (mV)
3500
ii
3000
2500
2000
1500
i
1000
500
0
0
3
6
9
12
15
Time (s)
Figure 35. Chemiluminescence intensity versus time profiles for the reaction
between 1 u10-5 M morphine and acidic potassium permanganate with flow-cells (i)
C (blue) and D (pink dotted) and (ii) C and D after addition of 0.6 mM
manganese(II) sulfate.
70
CHAPTER THREE
3.2
Sodium thiosulfate
3.2.1
Preliminary experiments
In 1989, Perez-Benito et al. [208] demonstrated that the addition of sodium
thiosulfate to a solution of potassium permanganate (without polyphosphates)
produced a transparent brown solution of colloidal manganese(IV). This reagent has
been used as an oxidant in reactions with formic acid [209], oxalic acid [210], lactic
acid [211] and L-tryptophan [212]. Recently, Hindson et al. [213] discovered that
when 1% (m/v) polyphosphates are added to the acidic potassium permanganate
reagent solution prior to the addition of aqueous sodium thiosulfate, the result is a
near instantaneous partial reduction of permanganate to form a concomitant mixture
of both manganese(VII) and manganese(III).
An optimisation of the permanganate and thiosulfate concentrations required to
produce the greatest chemiluminescence intensities with morphine and synephrine
was undertaken using FIA. A series of twenty reagents were prepared containing
concentrations of potassium permanganate between 1.0 and 1.9 mM, thiosulfate
between 0.0 and 1.5 mM and 1% (m/v) sodium polyphosphates. The partial reduction
of permanganate was accompanied by a visual colour change; as the concentration of
thiosulfate added to the reagent increased, the characteristic purple colour
diminished.
Whilst the optimal concentrations of oxidant and reductant appeared to be analyte
dependant (Figures 36 and 37), in general greater concentrations of permanganate
required the addition of more thiosulfate to achieve similar levels of enhancement.
However, concentrations of thiosulfate greater than 1.0 mM generally resulted in
71
CHAPTER THREE
significant decreases in chemiluminescence intensity due to the complete reduction
of permanganate.
Moderate increases (5- to 9-fold) in signal intensity were achieved following the
addition of 0.3 mM thiosulfate (to the four permanganate reagents) and subsequent
reaction with a morphine standard (1 × 10-7 M). Further enhancement (between
11- and 18-fold) was obtained when the concentration of thiosulfate was increased to
0.6 mM. Interestingly, the addition of 1.0 mM thiosulfate resulted in quenching of
the chemiluminescence signal, except for the reagent containing 1.9 mM
permanganate, where a 17-fold increase in intensity was observed (Figure 36).
160
i
Chemiluminescence intensity (mV)
140
120
ii
100
iii
80
iv
60
40
20
0
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
Thiosulfate concentration (mM)
Figure 36. The effect of thiosulfate concentration on the chemiluminescence
response (peak height) obtained for the reaction of four acidic potassium
permanganate reagents (i) 1.0 mM, (ii) 1.3 mM, (iii) 1.6 mM and (iv) 1.9 mM with
morphine (1 × 10-7 M), using FIA methodology.
Superior levels of enhancement were achieved when the twenty permanganate
reagents were reacted with synephrine (1 × 10-5 M, Figure 37). As with morphine,
moderate increases in signal intensity were achieved with the reagents containing
72
CHAPTER THREE
0.3 mM thiosulfate (37- to 45-fold). Significant enhancement was achieved with
reagents containing 0.6 mM thiosulfate (63-fold from a 1.9 mM permanganate
solution, Figure 37iv).
800
Chemiluminescence intensity (mV)
700
600
iv
500
iii
400
ii
300
i
200
100
0
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
Thiosulfate concentration (mM)
Figure 37. The effect of thiosulfate concentration on the chemiluminescence
response (peak height) obtained for the reaction of four acidic potassium
permanganate reagents (i) 1.0 mM, (ii) 1.3 mM, (iii) 1.6 mM and (iv) 1.9 mM with
synephrine (1 × 10-5 M), using FIA methodology.
3.2.2
Stopped flow analysis
Stopped flow experiments were conducted to examine the change in reaction
kinetics with increasing thiosulfate concentration. Five permanganate reagents
(1.9 mM) were prepared with concentrations of thiosulfate between 0.0 and 1.5 mM
and were subsequently reacted with a synephrine standard (1 × 10-5 M ).
As with the previous mode of preparation (utilising the addition of manganese(II)),
the production of the enhanced reagent using thiosulfate significantly increased the
rate of the reaction between synephrine and permanganate. This was reflected in the
73
CHAPTER THREE
chemiluminescence versus time profiles (Figure 38), with a significant reduction in
the time required to reach maximum emission (from 60 seconds to 3 seconds) and
increases in both peak intensity (peak height) and overall light emitted (peak area,
Table 3). Concentrations of thiosulfate greater than 1.0 mM resulted in notable
decreases in both peak intensity and overall signal, until quenching of the signal was
obtained (1.5 mM thiosulfate) due to the complete consumption of permanganate.
The enhancement of chemiluminescence signals for compounds such as synephrine
(that react slowly with the standard reagent) was of greatest interest to this study and
therefore all subsequent experiments were conducted using an enhanced reagent
containing 1.9 mM permanganate and 0.6 mM thiosulfate.
550
iv
Chemiluminescence intensity (mV)
500
450
iii
400
350
300
250
200
ii
150
100
50
i
0
0
5
10
15
20
25
Time (s)
Figure 38. Chemiluminescence intensity versus time profiles for the reaction of
synephrine (1 × 10-5 M) with acidic potassium permanganate (1.9 × 10-3 M)
containing (i) 0 mM, (ii) 0.3 mM, (iii) 0.6 mM, and (iv) 1.0 mM sodium thiosulfate.
74
CHAPTER THREE
Table 3. Effect of sodium thiosulfate concentration on the chemiluminescence intensity
(peak height) and total chemiluminescence emission (peak area) from the reaction of
acidic potassium permanganate (1.9 × 10-3 M) with synephrine (1 × 10-5 M).
Thiosulfate
concentration (mM)
Chemiluminescence
intensity (mV)
Total chemiluminescence
emission (mV.s)
0.0
10
838
0.3
468
1686
0.6
498
1558
1.0
179
605
1.5
0
0
3.3.3
Stability studies
The stability of the enhanced reagent was assessed by monitoring the
chemiluminescence intensity from the reaction with a morphine standard
(1 × 10-6 M) using SIA. A set of five replicate injections was performed each hour
for 48 hours following the addition of thiosulfate to the regent solution (Figure 39).
Unlike the previous mode of reagent preparation (utilising the addition of
manganese(II)), the addition of thiosulfate produced an immediate increase in the
chemiluminescence signal (11-fold) that did not change significantly over 48 hours
(a change of < 1% based on the line of best fit). The R.S.D. for the 49 sets of five
signals was 0.8%. Similar results were obtained when monitoring the reaction with
synephrine for 18 hours (< 0.1% change; 1.3% R.S.D.).
75
CHAPTER THREE
Chemiluminescence intensity (arbitrary units)
2800
2400
2000
1600
1200
800
400
0
0
4
8
12
16
20
24
28
32
36
40
44
48
Time (h)
Figure 39. Stability of the enhanced permanganate reagent (1.9 mM containing
0.6 mM thiosulfate) with morphine (1× 10-6 M), over 48 hours with SIA
methodology. The first point (t = 0) was obtained using the permanganate reagent
prior to the addition of thiosulfate
The stability of the enhanced reagent was also assessed by monitoring its visible
absorbance over 48 hours (Figure 40). A minor decrease in the structured band at
approximately 525 nm (characteristic of permanganate) and no fluctuation at lower
wavelengths, (where a large absorbance is typical of manganese(III) complexes)
confirmed that the enhanced reagent is extremely stable over an extended period of
time.
76
CHAPTER THREE
3.0
2.5
Absorbance (a.u.)
2.0
1.5
1.0
0.5
0.0
300
350
400
450
500
550
600
650
700
Wavelength (nm)
Figure 40. Absorption spectra for the enhanced permanganate reagent (1.9 mM,
containing 0.6 mM thiosulfate) measured every hour for 48 hours (only every sixth
spectrum shown). The initial spectrum is shown in bold.
3.2.4
Reagent comparison and screening test
The previously observed improvement in chemiluminescence intensity obtained
using the manganese(II) enhanced reagent demonstrated that the greatest levels of
enhancement are achieved with simple phenols (such as synephrine), rather than
those containing ortho-hydroxy or -alkoxy groups (such as morphine). A comparison
was undertaken to examine the improvements in chemiluminescence intensities
obtained from the permanganate reagents (standard, manganese(II) enhanced, and
thiosulfate enhanced) upon reaction with several simple phenols (Figure 41) using
FIA methodology.
77
CHAPTER THREE
Figure 41. Structures of the simple phenols used in the reagent screening test.
Significant increases in chemiluminescence intensity were obtained for both of the
enhanced reagents, however, the thiosulfate method proved on average 16% superior
to the manganese(II) preparation. This analyte screen confirmed that; (i) thiosulfate
is a viable alternative to manganese(II) as an enhancer for permanganate
chemiluminescence, and (ii) increases of well over an order of magnitude can be
obtained for a wide range of phenolic compounds (Figure 42). The greatest
improvements in chemiluminescence intensity were obtained for the phenolic amino
acid tyrosine (164-fold) and the sympathomimetic ȕ-adrenergic agonist, fenoterol
(118-fold, Table 4).
78
CHAPTER THREE
700
400
Chemiluminescence intensity (mV)
500
6
5
4
3
2
1
ol
LA
sc
or
bi
cA
ci
d
en
ph
4-
N
itr
o
V
300
an
ill
in
0
Analyte
200
100
0
V
an
ill
4in
N
itr
op
he
Lno
A
sc
l
or
bi
cA
3ci
A
d
m
in
op
he
no
l
Ty
ro
s in
e
A
m
ox
ic
ill
in
Ty
ra
m
in
e
Sy
ne
ph
rin
e
Fe
no
te
ro
l
M
or
ph
in
e
Chemiluminescence intensity (mV)
600
7
Analyte
Figure 42. Chemiluminescence intensities (peak heights) for selected analytes with
the standard reagents (1.0 mM potassium permanganate, blue columns and 1.9 mM
potassium permanganate, pink columns) and the enhanced reagents (1.0 mM
potassium permanganate containing 0.6 mM manganese(II), yellow columns and
1.9 mM potassium permanganate containing 0.6mM thiosulfate, green columns)
using FIA. All analytes were tested at a concentration of 1 × 10-5 M, except for
morphine (1 × 10-6 M).
Table 4. Chemiluminescence intensities (peak heights) for reactions between
various analytes (1 × 10-5 M) and the standard reagent (1.0 mM permanganate) and
the enhanced reagent (containing 0.6 mM thiosulfate) using FIA.
Chemiluminescence intensity (mV)
Enhancement
factor
Analyte
Standard
permanganate
Enhanced
permanganate
Vanillin
0.1
4
40.0
4-Nitrophenol
0.1
0.5
5.0
Ascorbic acid
6
5
0.8
Amoxicillin
3
177
59.0
Tyrosine
1
164
164.0
Tyramine
11
485
44.0
3-Aminophenol
10
146
14.6
Synephrine
14
679
48.5
Fenoterol
3
355
118.3
84
698
8.3
Morphine
a
79
CHAPTER THREE
a
3.2.5
1 × 10-6 M
Detection limits
The greater chemiluminescence intensities obtained for both fenoterol and tyrosine
with the enhanced reagent also translated into much lower limits of detection for both
analytes, 1 × 10-9 M and 4 × 10-9 M, respectively. This represents a significant
increase in sensitivity compared to that obtained with the standard reagent (Tables 5
and 6). Furthermore, the detection limit for tyrosine is superior to all previously
reported values obtained with acidic potassium permanganate [23]. Fenoterol has not
previously been detected with permanganate chemiluminescence.
Table 5. Analytical figures of merit obtained for fenoterol with the standard and
enhanced permanganate reagents.
a
Reagent
Calibration function
(peak height, mV)
R2
LOD
(M)a
R.S.D.
%*
Standard
permanganate
y = 0.862x + 4.715
0.9969
4.8 × 10-8
0.95
Enhanced
permanganate
y = 0.839x + 6.542
0.9985
1 × 10-9
0.56
Calculated as 3× the standard deviation of the blank response.
*1 × 10-5 M (n = 10)
Table 6. Analytical figures of merit obtained for tyrosine with the standard and
enhanced permanganate reagents.
a
Reagent
Calibration function
(peak height, mV)
R2
LOD
(M)a
R.S.D.
%*
Standard
permanganate
y = 0.908x + 4.429
0.9916
2.3 × 10-7
1.05
Enhanced
permanganate
y = 0.835x + 6.154
0.9983
3.6 × 10-9
0.61
Calculated as 3× the standard deviation of the blank response.
*1 × 10-5 M (n = 10)
80
CHAPTER THREE
4. Conclusions
Acidic potassium permanganate chemiluminescence has been used to detect a wide
variety of compounds, including biomolecules, antioxidants, neurotransmitters,
opiate alkaloids and pharmaceuticals. The greater control of reaction rates and
enhanced emission intensities obtained using manganese(II) or sodium thiosulfate
catalysis will improve the detection compatibility of various flow analysis
methodologies. This approach not only provides the potential for increased
sensitivity
in
existing
applications
of
acidic
potassium
permanganate
chemiluminescence detection, but may also enable its extension to a multitude of
new analytes and analytical techniques. The findings of this chapter have resulted in
two recent publications [205, 213].
81
CHAPTER FOUR
CHAPTER FOUR
Direct detection of biologically significant
thiols and disulfides with post-column
chemiluminescence
x
Introduction
x
Experimental
x
Results and discussion
x
Conclusion
82
CHAPTER FOUR
1. Introduction
Over the past two decades there has been a growing interest in the quantification of
low-molecular mass thiols and disulfides as researchers gain new insight into their
role in significant cellular processes [85, 88, 97-99, 108, 112, 113, 116, 117, 121,
214, 215]. Particular attention has been paid to the redox couple, glutathione (GSH)
and glutathione disulfide (GSSG). Oxidation of the free thiol to the corresponding
disulfide is amongst the earliest responses to increases in reactive oxygen species in
eukaryotic cells, and therefore assessment of the GSH/GSSG ratio offers great
potential for diagnosing and monitoring pathological and physiological conditions
related to oxidative stress [85, 97, 99, 112, 113].
From the numerous methods previously developed for the detection of thiols and
disulfides, several key analytical challenges have emerged [84, 85, 87, 107, 109, 110,
114, 115]. The analytes typically lack strong chromophores or fluorophores, which is
most
commonly overcome
by
derivatisation
of
the
free
thiols
before
chromatographic/electrophoretic analysis, and then repeating the procedure on a
second sample aliquot after reduction of the disulfides to their respective thiols [84,
85, 87, 107, 109, 110, 114, 115]. The disulfide concentrations are then calculated
from the difference between the measurements obtained in the two separate
analytical steps. This approach is not only time consuming, but also vulnerable to
several major sources of error [87, 115], the most important of which is the oxidation
of thiols during either the initial sample deproteinisation, or under the alkaline
conditions often employed for derivatisation and reduction steps. Disulfides are
present in biological samples at much lower concentrations than thiols, and even a
small degree of auto-oxidation during these extensive sample preparation procedures
can lead to their considerable overestimation [84, 85, 87, 106-109, 116, 117].
83
CHAPTER FOUR
Therefore, the determination of biologically significant thiols and disulfides should
ideally combine minimal sample handling (under conditions that prevent autooxidation) with direct detection of all target analytes in a single chromatographic
separation. Thiols and disulfides have previously been detected without
derivatisation using electrochemistry, mass spectrometry and fluorescence quenching
[84, 87, 107, 110, 116-119]. Whilst these modes of detection offer advantages over
the more popular derivatisation techniques in terms of procedural simplicity, their
application is limited by equipment complexity, cost, electrode stability and/or
analysis time [108, 110, 120, 121].
Several procedures coupling a separation with chemiluminescence detection have
been developed for the direct determination of thiols [18, 123, 124]. The post-column
chemiluminescence oxidation of GSH and cysteine with cerium(IV) (sensitised by
rhodamine B) was reported by Li et al. [123], however this methodology resulted in
unsatisfactory resolution of the two analyte peaks. Zhao et al. [124] described the use
of microchip electrophoresis and luminol chemiluminescence to determine GSH,
cysteine and haemoglobin in red blood cells. McDermott et al. [18] reported a
procedure to determine GSH in cultured muscle cells, based on HPLC with direct
acidic potassium permanganate chemiluminescence detection. However, like other
previously reported chemiluminescence-based detection systems for thiols [123,
124], the reagent was not suitable to detect disulfides in biological systems, and
therefore the quantification of GSSG required thiol blocking and disulfide bond
reduction steps prior to a second chromatographic run [18].
In recent years, there have been several major advances in two of the most
prolifically
used
chemiluminescence
reagent
systems
(acidic
potassium
permanganate and tris(2,2c-bipyridine)ruthenium(III)), in terms of their stability [33,
84
CHAPTER FOUR
205, 213] and sensitivity [205, 213], and the understanding of their light producing
pathways [14, 216, 217]. Furthermore, the development of a colloidal
manganese(IV) chemiluminescence reagent [46, 48] which has been found to
produce light with a wider range of compounds than other manganese based oxidants
offers the potential for new detection strategies of native (underivatised) analytes.
This chapter presents a systematic evaluation of the relative chemiluminescence
responses from several biologically significant thiols and disulfides (Figure 43) with
conventional
and
new
adaptations
of
both
the
permanganate
and
tris(2,2c-bipyridine)ruthenium(III) reagents, in addition to the manganese(IV) colloid.
A chromatographic separation of these analytes was subsequently developed,
employing chemiluminescence detection to assess the redox state of whole blood
from twelve healthy volunteers.
Figure 43. Structures of biologically significant thiols and disulfides.
85
CHAPTER FOUR
2. Experimental
2.1
Chemicals and reagents
Deionised water (Continental Water Systems, Victoria, Australia) and analytical
grade reagents were used unless otherwise stated. Chemicals were obtained from the
following sources: N-acetylcysteine (•99%), L-cysteine (•98.5%), L-cystine (•99%),
L-cysteinylglycine
(•85%),
disodium
phosphate,
N-ethylmaleimide
(NEM),
Ȗ-glutamyl-cysteine (•80%), L-glutathione (•98%), L-glutathione disulfide (•99%),
homocysteine (•95%), homocystine, L-methionine (•99%), monosodium phosphate,
sodium polyphosphate (+80 mesh), sodium tetraborate, trichloroacetic acid,
trifluoroacetic acid and tris(2-carboxyethyl)phosphine hydrochloride (TCEP) from
Sigma-Aldrich (New South Wales, Australia); formaldehyde (37%), orthophosphoric
acid (85%), potassium permanganate and sodium chloride from Chem-Supply (South
Australia,
Australia);
sodium
formate
from
BDH
(Poole,
England);
ethylenediaminetetraacetic acid disodium salt (EDTA), methanol, sulfuric acid
(98%) and tris(hydroxymethyl)methylamine from Merck (Victoria, Australia);
formic acid from Hopkin and Williams (Essex, England); perchloric acid (70% w/v)
from Univar (New South Wales, Australia); aqueous soluble starch, lead dioxide and
sodium perchlorate from Ajax Finechem (New South Wales, Australia); potassium
iodide from Fisons Scientific Equipment (Loughborough, England); acetonitrile from
Burdick & Jackson (Michigan, USA) and tris(2,2ƍ-bipyridine)ruthenium(II)
dichloride hexahydrate from Strem Chemicals (Minnesota, USA).
Stock solutions (1 × 10-3
M) of
N-acetylcysteine, cysteine,
cystine,
cysteinylglycine, Ȗ-glutamyl-cysteine, GSH, GSSG, homocysteine, homocystine and
methionine were prepared daily in deionised water that had been adjusted to pH 2.57
86
CHAPTER FOUR
with trifluoroacetic acid, and diluted into the mobile phase (98% Solvent A:
deionised water adjusted to pH 2.57 with trifluoroacetic acid; 2% Solvent B:
methanol) as required.
The thiol selective acidic potassium permanganate reagent was prepared by
dissolution of potassium permanganate (2.5 × 10-4 M) in 1% (m/v) sodium
polyphosphate and adjusting to pH 3 with sulfuric acid.
The ‘standard’ acidic potassium permanganate reagent was prepared by dissolution
of potassium permanganate (1.0 × 10-3 M) in 1% (m/v) sodium polyphosphate and
adjusting to pH 2.5 with sulfuric acid.
The ‘enhanced’ acidic potassium permanganate reagent was prepared by
dissolution of potassium permanganate (1.9 × 10-3 M) in 1% (m/v) sodium
polyphosphate and adjusting to pH 2.5 with sulfuric acid and then adding sodium
thiosulfate (0.6 mM), using a small volume of a 0.1 M solution.
The ‘conventional’ tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent was prepared by
oxidising tris(2,2‫މ‬-bipyridine)ruthenium(II) (5 × 10-5 M) with lead dioxide
(0.2 g/100 mL) in 0.02 M sulfuric acid. Following the change in colour from orange
to emerald green (approximately 5 minutes with stirring), the excess oxidant was
filtered from the solution using a 0.45 Pm syringe-tip filter prior to injection into the
FIA manifold.
The ‘stabilised’ ruthenium perchlorate reagent was prepared as previously
described [33]. This involved treating [Ru(bipy)3]Cl2 with sodium perchlorate in
aqueous solution to yield the characteristic bright orange [Ru(bipy)3](ClO4)2
precipitate, which was collected by vacuum filtration, washed twice with ice water,
and dried over phosphorus pentoxide for 24 hours. The [Ru(bipy)3](ClO4)2 crystals
87
CHAPTER FOUR
(5 × 10-5 M) were oxidised with lead dioxide (0.2 g/100 mL) in acetonitrile
containing 0.05 M perchloric acid. Following the change in colour from orange to
blue-green (approximately 1 minute with stirring), the excess oxidant was filtered
from the solution using a 0.45 Pm syringe-tip filter prior to injection into the FIA
manifold.
For use as a flowing stream of reagent, the [Ru(bipy)3](ClO4)2 crystals (1 × 10-3 M)
were oxidised with 0.2 g/100 mL lead dioxide in acetonitrile containing 0.05 M
perchloric acid. Following 1 minute of vigorous mixing and the characteristic change
in colour from orange to blue-green, the resultant suspension was allowed to settle
prior to filtration through an in-line filter (consisting of a small Pasteur pipette
packed tightly with glass wool).
Soluble manganese(IV) was prepared as previously described [46], based on the
method of Jáky and Zrinyi [45]. This involved the reduction of potassium
permanganate using excess sodium formate to yield the characteristic dark brown
manganese dioxide precipitate which was collected by vacuum filtration
(GF/A Whatman, England). After rinsing with deionised water, the freshly
precipitated wet manganese dioxide (1.2 g) was added to 1.0 L of orthophosphoric
acid (3.0 M) and ultrasonicated for 30 min. The colloid was then heated for 1 hour
(80 °C) with stirring, and then cooled to room temperature. The concentration was
determined by iodometric titration as described by Vogel [218]. The stock
manganese(IV) reagent was diluted daily to the required concentration (5 × 10-4 M)
using orthophosphoric acid (3.0 M). Formaldehyde was filtered and diluted to the
required concentration with deionised water.
88
CHAPTER FOUR
2.2
Flow injection analysis
The FIA manifold was constructed from a Gilson Minipuls 3 peristaltic pump
(John Morris Scientific, Victoria, Australia) with bridged PVC pump tubing
(1.02 mm i.d.; DKSH, Queensland, Australia), black manifold tubing (0.76 mm i.d.,
Global FIA, WA, USA) and six-port injection valve (Vici 04 W-0192L, Valco
Instruments, Texas, USA) equipped with a 70 —L sample loop. A GloCel
chemiluminescence detector (Global FIA) with flow-cell D (dual-inlet serpentine)
and extended range photomultiplier module (Electron Tubes model P30A-05; ETP,
New South Wales, Australia) was used for the permanganate and manganese(IV)
experiments, whilst an in-house fabricated detector containing an aluminium-backed
coiled-tubing flow-cell and extended-range photomultiplier module within a
light-tight housing was used for the tris(2,2‫މ‬-bipyridine)ruthenium(III) experiments.
All tubing entering and exiting the detectors was black PTFE (0.76 mm i.d). The
output signal from the photomultiplier module was recorded using ‘e-corder 410’
data acquisition software (eDAQ, New South Wales, Australia).
2.2.1
Acidic potassium permanganate
The analytes (1 × 10-5 M) were injected (70 —L) into a 100% aqueous carrier
stream (2.5 mL/min or 3.5 mL/min) that merged with the acidic potassium
permanganate reagent (2.5 mL/min or 3.5 mL/min) within the reaction channel of the
flow-cell (Figure 44).
89
CHAPTER FOUR
Detector
Reagent
Flow-cell
Peristaltic
pump
PMT
Injection valve
Carrier
Waste
Analyte
Figure 44. Schematic of the FIA manifold used for acidic potassium permanganate
measurements.
2.2.2
Tris(2,2ү-bipyridine)ruthenium(III)
Manifold 1: A continuously flowing analyte stream (5 × 10-6 M; 3.5 mL/min) was
merged with a sodium tetraborate buffer (0.1 M, pH 10; 3.5 mL/min) at a T-piece
located 15 cm prior to the flow-cell inlet. The ruthenium reagents were injected
(70 —L) into a 100% aqueous carrier stream (deionised water; 3.5 mL/min) that
merged with the combined analyte-buffer solution just prior to entry into the
flow-cell (Figure 45).
Analyte
Buffer
Detector
Flow-cell
Peristaltic
pump
PMT
Injection valve
Carrier
Waste
Reagent
Figure 45. Schematic of the FIA
tris(2,2‫މ‬-bipyridine)ruthenium(III) reagents.
manifold
used
to
inject
the
Manifold 2: The analytes (1 × 10-5 M) were each injected (70 —L) into a 100%
aqueous carrier stream (formic acid pH 2.8; 1 mL/min) that merged with a phosphate
buffer (0.04 M; pH 8.0; 1 mL/min) at a T-piece located 15 cm prior to the entrance of
the flow-cell. A continuously flowing stream of the ruthenium reagent
90
CHAPTER FOUR
(1 × 10-3 M; 2 mL/min) then merged with the combined analyte-buffer solution just
prior to entry into the flow-cell (Figure 46).
Detector
Reagent
Buffer
Flow-cell
PMT
Peristaltic
pump
Injection valve
Carrier
Waste
Analyte
Figure
46.
Schematic
of
the
FIA
manifold
tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent as a flowing stream.
2.2.3
used
with
the
Soluble manganese(IV)
The analytes (1 × 10-5 M) were each injected (70 —L) into a 100% aqueous carrier
stream (deionised water; 3.5 mL/min) that merged with a flowing stream of 2.0 M
formaldehyde (used as an enhancer to significantly improve the emission intensity;
3.5 mL/min) at a T-piece located 22 cm prior to the flow-through detector inlet. A
continuously flowing stream of the manganese(IV) reagent (3.5 mL/min) merged
with the combined analyte-formaldehyde solution in the reaction channel of the
flow-cell (Figure 47).
Detector
Manganese(IV)
Formaldehyde
Flow-cell
PMT
Peristaltic
pump
Injection valve
Deionised water
Waste
Analyte
Figure 47. Schematic of the FIA manifold used for manganese(IV) chemiluminescence
measurements.
91
CHAPTER FOUR
2.3
Chemiluminescence spectra
A Cary Eclipse fluorescence spectrophotometer (Varian, Mulgrave, Victoria,
Australia) with R928 photomultiplier tube (Hamamatsu, Japan) was operated in
Bio/Chemiluminescence mode. A three-line continuous flow manifold was used to
merge the GSH or GSSG solution (1 × 10-4 M; 1.0 mL/min) with formaldehyde
(2.0 M; 1.0 mL/min), and then with the manganese(IV) reagent (5 × 10-4 M in 3.0 M
orthophosphoric acid; 1.0 mL/min), and propel the reacting mixture through a coiled
PTFE flow-cell (200 ȝL, 0.8 mm i.d.) that was mounted against the emission window
of the spectrophotometer. Final spectra were an average of 20 scans (1000 ms gate
time, 1 nm data interval, 20 nm band pass, PMT: 800 V) and corrected as previously
described [219].
2.4
High performance liquid chromatography
Chromatographic analysis was carried out on an Agilent Technologies 1200 series
liquid chromatography system, equipped with a quaternary pump, solvent degasser
system and autosampler (Agilent Technologies, Victoria, Australia), using an Alltech
Alltima C18 column (250 mm × 4.6 mm i.d., 5 —m) at room temperature, with an
injection volume of 100 —L and a flow rate of 1 mL/min. Isocratic elution was
performed with 98% Solvent A: deionised water adjusted to pH 2.57 with
trifluoroacetic acid and 2% Solvent B: methanol. A Hewlett-Packard analogue to
digital interface box (Agilent Technologies) was used to convert the signal from the
chemiluminescence detector. Before use in the HPLC system, all sample solutions
and solvents were filtered through a 0.45 ȝm nylon membrane.
For manganese(IV) chemiluminescence measurements, the column eluate
(1.0 mL/min) and a formaldehyde carrier (2.0 M; 1.0 mL/min) were merged at a
92
CHAPTER FOUR
T-piece located 22 cm from the entrance of the detector. This stream was then
combined with the manganese(IV) reagent (5 × 10-4 M in 3.0 M orthophosphoric
acid; 1.0 mL/min) in the detector employed for FIA as described above. A peristaltic
pump was used to deliver the reagents (Figure 48).
HPLC pump
Column
Detector
Flow-cell
Injector
PMT
Formaldehyde
Manganese(IV)
Peristaltic
pump
Waste
Figure 48. Schematic of the HPLC manifold employed for manganese(IV)
chemiluminescence measurements.
For acidic potassium permanganate chemiluminescence measurements, the column
eluate (1.0 mL/min) and reagent (2.5 mL/min) merged at a T-piece and the light
emitted from the reacting mixture was detected with a custom built flow-through
chemiluminometer, which consisted of a coiled flow-cell comprising of 0.8 mm i.d.
PTFE tubing (DKSH), mounted flush against the window of Electron Tubes
photomultiplier tube (model 9878SB, ETP) set at a constant voltage of 900 V from a
stable power supply (PM20D, ETP) via a voltage divider (C611, ETP). The
flow-cell, photomultiplier tube and voltage divider were encased in light-tight
housing (Figure 49).
93
CHAPTER FOUR
HPLC pump
Detector
Column
Flow-cell
Injector
PMT
Reagent
Peristaltic
pump
Waste
Figure 49. Schematic of the HPLC manifold employed for acidic potassium
permanganate chemiluminescence measurements.
2.5
Sample collection and analysis
The method of Stempak et al. [120] was followed when collecting and processing
blood samples. Venous blood samples (5 mL) were drawn from twelve healthy, nonfasting adult volunteers into ethylenediaminetetraacetic acid disodium salt (EDTA)
lined blood tubes (Becton and Dixon, UK) and immediately placed on ice. Aliquots
of ice-cold perchloric acid (500 —L, 15% (v/v) containing 2.0 mM EDTA) were
pipetted into 1.5 mL Eppendorf tubes, followed by whole blood (500 —L). The
samples were then vortexed and incubated on ice for 10 minutes to completely
precipitate blood proteins. The acidic suspensions were then centrifuged at 13000
rpm for 15 minutes at 4 °C, and immediately analysed. For analysis, 10 —L of the
acidic supernatant was mixed with 990 —L of mobile phase (98% Solvent A:
deionised water adjusted to pH 2.57 with trifluoroacetic acid; 2% Solvent B:
methanol), vortexed and filtered with a 0.2 —m cellulose acetate filter, prior to
injection onto the HPLC column.
94
CHAPTER FOUR
3. Results and Discussion
3.1
Acidic potassium permanganate chemiluminescence detection
Li and co-workers previously reported the determination of thiols such as cysteine,
GSH, N-acetylcysteine and captopril based on the sensitised chemiluminescence
reaction of acidic potassium permanganate and quinine [123]. This system contains
two light producing pathways: (i) the reduction of permanganate to form an excited
state manganese(II) species, and (ii) the oxidation of the thiol compounds to generate
an excited intermediate capable of transferring energy to the quinine fluorophore [13,
202].
This approach was improved by McDermott and co-workers via the substitution of
quinine with sodium polyphosphate to promote the manganese(II) pathway [18].
Removing quinine eliminated the background emission from the reaction between
the sensitiser and oxidant, enabled post-column chemiluminescence detection of
GSH using a single reagent solution and provided greater sensitivity (the limit of
detection was 5 × 10-7 M, compared to 6.5 × 10-6 reported by Li et al.[123]).
The selectivity of this permanganate reagent for a wider range of biologically
significant thiols and disulfides (cysteine, cystine, GSH, GSSG, homocysteine and
homocystine) was evaluated using FIA methodology. Analyte standard solutions
(1 × 10-5 M) were injected into a stream of aqueous formic acid (pH 2.8;
2.5 mL/min) which merged with the permanganate reagent (2.5 × 10-4 M in 1% (m/v)
sodium polyphosphate solution, adjusted to pH 3 with sulfuric acid; 2.5 mL/min)
within the reaction channel of the flow-cell. Significant chemiluminescence signals
were obtained for the three thiols, a negligible emission for cystine (which may have
resulted from a minor breakdown of the disulfide to its corresponding thiol ; less than
95
CHAPTER FOUR
0.6% relative to cysteine), and no response for either GSSG or homocystine (Figure
50).
20
Chemiluminescence intensity (mV)
18
16
14
12
10
8
6
4
2
SS
CY
H
H
CY
S
G
SS
G
G
SH
SS
CY
CY
S
0
Analyte
Figure 50. Relative chemiluminescence intensities (peak heights) obtained for the
oxidation of biologically significant thiols and disulfides (1 × 10-5 M) with a
permanganate reagent optimised for the detection of thiols. Columns: cysteine
(CYS), cystine (CYSS), glutathione (GSH), glutathione disulfide (GSSG),
homocysteine (HCYS) and homocystine (HCYSS).
As demonstrated in chapter three, considerable increases in reaction rate and
chemiluminescence intensity can be obtained by a preliminary partial reduction of
permanganate to create a stable, relatively high concentration of manganese(III) in
permanganate reagents [205, 213]. Therefore the thiols and disulfides were also
screened with a standard permanganate reagent and the enhanced reagent.
Each of the analytes generated a response with both of the permanganate reagents.
On average, the signals obtained from the thiols were 17-fold more intense than their
corresponding disulfide using the standard permanganate reagent (Figure 51, blue
columns). In agreement with the results presented in chapter three, the enhanced
reagent provided increases in chemiluminescence intensity for the six analytes
96
CHAPTER FOUR
(1-2 fold for the thiols and 2.3-4.2 fold for the disulfides, Figure 51 pink columns).
Although there are very few methodologies capable of directly detecting thiols and
disulfides, the sensitivity achieved using the standard or enhanced permanganate
reagents was insufficient for transfer of these approaches to post-column
chemiluminescence detection.
1.8
Chemiluminescence intensity (mV)
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
SS
CY
H
H
CY
S
G
SS
G
G
SH
SS
CY
CY
S
0.0
Analyte
Figure 51. Relative chemiluminescence intensities (peak heights) obtained for the
oxidation of biologically significant thiols and disulfides (1 × 10-5 M) with (i)
standard permanganate (blue columns) and (ii) the enhanced permanganate reagent
(pink columns). Columns: cysteine (CYS), cystine (CYSS), glutathione (GSH),
glutathione disulfide (GSSG), homocysteine (HCYS) and homocystine (HCYSS).
3.2
Tris(2,2‫މ‬-bipyridine)ruthenium(III) chemiluminescence detection
As previously discussed, tris(2,2‫މ‬-bipyridine)ruthenium(III) has been used
extensively for the determination of various analytes containing an amine moiety,
with tertiary substituted amines generally eliciting the largest responses, followed by
secondary and then primary [30, 31, 38, 39]. Despite this, several studies have shown
that an alkaline reaction environment can promote analytically useful signals from
analytes containing secondary amines [23, 38, 220, 221]. The aforementioned thiols
97
CHAPTER FOUR
and disulfides possess both primary and secondary amine functionalities, and as
such, tris(2,2‫މ‬-bipyridine)ruthenium(III) chemiluminescence holds the potential for
direct detection of these biologically important analytes. Initial experiments were
conducted with a conventional tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent as
described by Costin and co-workers [23]. Using a three-line FIA manifold
(Figure 45), a continuously flowing analyte stream (5 × 10-6 M; 3.5 mL/min) was
merged with a sodium tetraborate buffer (0.1 M, pH 10; 3.5 mL/min) and then
combined with a deionised water carrier (3.5 mL/min) immediately prior to entering
the flow-cell. The reagent (5 × 10-6 M) was injected into the carrier stream. The
relatively large signal obtained for homocystine (Figure 52) initially suggested that
this reagent had the potential to determine biologically significant disulfides.
However the response for GSSG was less than 10% relative to homocystine, and the
other analytes produced emissions lower than the blank.
1800
Chemiluminescence intensity (mV)
1500
1200
900
600
300
0
SS
CY
H
H
CY
S
G
SS
G
G
SH
SS
CY
CY
S
-300
Analyte
Figure 52. Relative chemiluminescence intensities (peak heights) obtained for the
oxidation of biologically significant thiols and disulfides (5 × 10-6 M) with the
conventional tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent (5 × 10-5 M). Columns:
cysteine (CYS), cystine (CYSS), glutathione (GSH), glutathione disulfide (GSSG),
homocysteine (HCYS) and homocystine (HCYSS).
98
CHAPTER FOUR
A
longstanding
weakness
of
tris(2,2‫މ‬-bipyridine)ruthenium(III)
as
a
chemiluminescence reagent has been its limited stability in aqueous solution,
particularly when employing off-line methods of preparation and/or alkaline
conditions [1, 5, 30, 31, 33-36]. To address this issue, an exceedingly stable reagent
developed by McDermott and co-workers [33] was prepared by oxidising
[Ru(bipy)3](ClO4)2 in acetonitrile containing 0.05 M perchloric acid. This reagent
was injected into the same three-line FIA manifold as described above. The stabilised
reagent was found to exhibit similar selectivity to that of the conventional reagent,
however in this instance, the response for cystine was marginally larger than the
blank (despite being only 1% relative to that of homocystine, Figure 53).
1200
Chemiluminescence intensity (mV)
1000
800
600
400
200
0
SS
CY
H
H
CY
S
G
SS
G
G
SH
SS
CY
CY
S
-200
Analyte
Figure 53. Relative chemiluminescence intensities (peak heights) obtained for the
oxidation of biologically significant thiols and disulfides (5 × 10-6 M) with the
stabilised tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent (5 × 10-5 M). Columns: cysteine
(CYS), cystine (CYSS), glutathione (GSH), glutathione disulfide (GSSG),
homocysteine (HCYS) and homocystine (HCYSS).
99
CHAPTER FOUR
Previous
work
undertaken
by
McDermott
to
optimise
the
stabilised
tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent to derive the greatest response for thiols
and disulfides focused solely on the determination of GSH and GSSG [222]. The
selectivity of these conditions for a wider range of thiols and disulfides was
evaluated by injecting analyte standard solutions (1 × 10-5 M) into a stream of
aqueous formic acid (pH 2.8; 1.0 mL/min) which merged with a phosphate buffer
(0.04 M; pH 8.0; 1.0 mL/min) at a T-piece located 15 cm prior to the entrance of the
flow-cell (Figure 46). A continuously flowing stream of the reagent (1 × 10-3 M;
2.0 mL/min) then merged with the combined analyte-buffer solution just prior to
entry into the flow-cell. Under these conditions, each of the thiols produced
emissions bigger than the blank (Figure 54). Homocystine produced the dominant
response of the three disulfides, and no light was detected from the oxidation of
cystine.
Chemiluminescence intensity (mV)
2500
2000
1500
1000
500
SS
CY
H
H
CY
S
G
SS
G
G
SH
SS
CY
CY
S
0
Analyte
Figure 54. Relative chemiluminescence intensities (peak heights) obtained for the
oxidation of biologically significant thiols and disulfides (1 × 10-5 M) with the stabilised
tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent (1 × 10-3 M). Columns: cysteine (CYS),
cystine (CYSS), glutathione (GSH), glutathione disulfide (GSSG), homocysteine
(HCYS) and homocystine (HCYSS).
100
CHAPTER FOUR
3.3
Manganese(IV) chemiluminescence detection
As outlined in the review by Brown et al. [43] only a small number of analytical
applications of manganese(IV) chemiluminescence have been reported to-date. Of
these, only one manuscript describes the use of manganese(IV) for post-column
chemiluminescence detection, where a mixture of six opiate alkaloid standards were
determined following HPLC separation with a monolithic column [46]. Owing to the
limited number of studies, little is known regarding the relationship between analyte
structure and the chemiluminescence response with this reagent [43]. Moreover,
there have been no published reports of manganese(IV) chemiluminescence emission
upon reaction with thiols or disulfides.
Previous work by Brown et al. found that the emission intensities for various
analytes were significantly enhanced by the addition of formaldehyde [46].
Therefore, the ability of the thiols and disulfides to provide analytically useful
signals upon reaction with the manganese(IV) reagent was examined using a
three-line FIA manifold. The analytes (1 × 10-5 M) were each injected (70 —L) into a
100% aqueous carrier stream (1 mL/min) that merged with a flowing stream of 2.0 M
formaldehyde (1.0 mL/min) at a T-piece located 22 cm prior to the flow-through
detector inlet. The manganese(IV) reagent (1.0 mL/min) was pumped into the second
inlet and merged with the combined analyte-formaldehyde solution within the
serpentine reaction channel of the flow-cell. Strong signals were recorded for each of
the analytes (Figure 55).
101
CHAPTER FOUR
Chemiluminescence intensity (mV)
125
100
75
50
25
SS
CY
H
H
CY
S
G
SS
G
G
SH
SS
CY
CY
S
0
Analyte
Figure 55. Relative chemiluminescence intensities (peak heights) obtained for the
oxidation of biologically significant thiols and disulfides (1 × 10-5 M) with
manganese(IV) (5 × 10-4 M). Columns: cysteine (CYS), cystine (CYSS), glutathione
(GSH), glutathione disulfide (GSSG), homocysteine (HCYS) and homocystine
(HCYSS).
Using the aforementioned analytes, a series of univariate searches were performed
on flow rate and the concentration of formaldehyde. A 2.0 M formaldehyde solution
and flow rate of 1.0 mL/min per line were found to afford the greatest
chemiluminescence response for each of the analytes and were used for all further
experiments. The percentage of organic modifier (methanol or acetonitrile) added to
the aqueous carrier stream (representing the HPLC mobile phase) was found to have
a critical effect on manganese(IV) chemiluminescence. For example, the addition of
only 1% acetonitrile resulted in an average 70% decrease in signal intensity for GSH
(Figure 56, blue columns) and GSSG (Figure 56, pink columns). In contrast, adding
methanol (in 1% increments to 5%) did not significantly alter the response
102
CHAPTER FOUR
(Figure 56). Therefore, use of acetonitrile in combination with the manganese(IV)
reagent is not recommended.
Chemiluminescence intensity (mV)
60
50
40
30
20
10
5%
M
eO
H
M
eO
H
3%
4%
M
eO
H
M
eO
H
2%
M
eO
H
1%
1%
A
CN
0
Organic modifier
Figure 56. Effect of the percentage of organic modifier on the relative
chemiluminescence intensities (peak heights) obtained for the oxidation of response
of GSH (blue columns) and GSSG (pink columns, 1 × 10-5 M) with manganese(IV)
(5 × 10-4 M).
Using the optimised parameters, a wider range of biologically important thiols and
disulfides
(cystine,
cysteine,
homocystine,
homocysteine,
GSH,
GSSG,
N-acetylcysteine, cysteinylglycine, Ȗ-glutamyl-cysteine, and methionine) were
screened with the manganese(IV) reagent. All of the analytes elicited a response,
demonstrating the potential of manganese(IV) chemiluminescence for the detection
of both thiols and disulfides (Figure 57).
103
CHAPTER FOUR
55
Chemiluminescence intensity (mV)
50
45
40
35
30
25
20
15
10
5
N
A
CY
S
ET
M
H
CY
S
H
CY
SS
LU
-C
Y
S
G
SS
G
J-G
G
SH
SS
CY
S
CY
CY
SG
LY
0
Analyte
Figure 57. Relative chemiluminescence intensities (peak heights) obtained for the
oxidation of biologically significant thiols and disulfides (1 × 10 -5 M) with
manganese(IV) (5 × 10-4 M). Columns: cysteinylglycine (CYS-GLY), cysteine
(CYS), cystine (CYSS), glutathione (GSH), glutathione disulfide (GSSG),
Ȗ-glutamylcysteine (Ȗ-GLU-CYS), homocysteine (HCYS), homocystine (HCYSS),
methionine (MET) and N-acetylcysteine (NACYS).
The chemiluminescence spectra (corrected for the detector response and
monochromator transmission [202]) for the oxidation of a thiol (GSH) and a
disulfide (GSSG) both contained a single broad band with a maximum at
734 r 5 nm, which matched the characteristic emission observed from other reactions
with manganese-based reagents [13, 16, 43, 48], attributed to an electronically
excited manganese(II) species [17] (Figure 58).
104
CHAPTER FOUR
14
(i)
GSH
Chemiluminescence signal (arbitrary units)
12
10
8
GSSG
6
4
2
0
550
600
650
700
750
800
850
Wavelength (nm)
140
(ii)
Chemiluminescence signal (arbitrary units)
120
GSH
100
80
60
GSSG
40
20
0
550
600
650
700
750
800
Wavelength (nm)
Figure 58. (i) Uncorrected and (ii) corrected chemiluminescence spectra for the
reactions between manganese(IV) (5 × 10-4 M) and glutathione (GSH, 1 × 10-4 M) or
glutathione disulfide (GSSG, 1 × 10-4 M), using formaldehyde (2.0 M) as an enhancer.
105
CHAPTER FOUR
3.3.1
Separation conditions and analytical figures of merit
As the vast majority of techniques for the detection of biological thiols and
disulfides employ pre-column derivatisation, very few involve separation of these
molecules in their native forms [84, 85, 87, 106, 108-110]. However, from these
accounts, satisfactory separation has been achieved using reverse phase HPLC with
an almost 100% aqueous isocratic solvent system [118-120]. A mobile phase
consisting of 98% deionised water adjusted to pH 2.57 with trifluoroacetic acid
(similar to that used by Khan et al. [118] and Pelletier et al. [119]) was initially
examined for the separation of a mixture containing seven routinely measured thiols
and disulfides. All of the analytes were separated within 35 minutes (Figure 59i) with
unnecessarily high resolution between N-acetylcysteine and GSSG. The separation
was tunable by simply adjusting the amount of methanol present in the mobile phase
by 1% increments, with the largest effects on retention time observed for
N-acetylcysteine and GSSG (co-elution of these peaks occurred with 3% methanol;
Figure 59iv). The optimal mobile phase composition consisted of 98% deionised
water adjusted to pH 2.57 with trifluoroacetic acid and 2% methanol. These
conditions were found to provide a resolution of 1.3 between cysteine and cystine
and baseline separation for all other analytes within 20 minutes (Figure 59iii).
106
Chemiluminescence intensity (a.u.)
CHAPTER FOUR
160
CYS
(i)
140
HCYS
CYSS
120
HCYSS
100
GSH
80
NACYS
60
40
GSH
20
0
0
5
10
15
20
25
30
35
40
Chemiluminescence intensity (a. u.)
Time (min)
180
(ii) CYS
160
140
HCYS
HCYSS
CYSS
120
GSH
NACYS
100
80
60
GSSG
40
20
0
0
5
10
15
20
25
30
Chemiluminescence intensity (a. u.)
Time (min)
180
(iii)
CYS
HCYS
160
HCYSS
CYSS
140
GSH
120
NACYS
100
80
60
GSSG
40
20
0
0
5
10
15
20
25
Chemiluminescence intensity (a. u.)
Time (min)
180
(iv)
160
CYS
HCYS
HCYSS
140
CYSS
GSH
120
NACYS
100
80
60
GSSG
40
20
0
0
5
10
15
20
Time (min)
Figure 59. Separation of a mixture containing thiols and disulfides (1 × 10 -5 M)
using various percentages of methanol in the mobile phase (deionised water adjusted
to pH 2.57 with trifluoroacetic acid); (i) 0%, (ii) 1%, (iii) 2% and (iv) 3%. Peaks:
cystine (CYSS), cysteine (CYS), homocysteine (HCYS), homocystine (HCYSS),
glutathione (GSH), N-acetylcysteine (NACYS), and glutathione disulfide (GSSG).
107
CHAPTER FOUR
Furthermore, to highlight the advantage of this procedure in comparison to the
previously reported chemiluminescence approach [18], the same mixture was
subjected to an identical separation but detected using acidic potassium
permanganate (Figure 60ii). Whilst both reagents can be used to detect thiols, the
manganese(IV) reagent also enabled the direct detection of the disulfides
(Figure 60i).
280
(i)
Chemiluminescence intensity (arbitrary units)
CYS
240
HCYS
CYSS
200
HCYSS
GSH
160
NACYS
120
80
GSSG
40
0
0
2
4
6
8
10
12
14
16
18
20
16
18
20
Time (min)
5
(ii)
Chemiluminescence intensity (arbitrary units)
CYS
4
3
GSH
HCYS
2
NACYS
1
0
0
2
4
6
8
10
12
14
Time (min)
Figure 60. Separation of a mixture containing thiols and disulfides (1 × 10 -5 M)
using either (i) manganese(IV) (5 × 10-4 M) or (ii) acidic potassium permanganate
(2.5 × 10-4 M) chemiluminescence detection.
108
CHAPTER FOUR
The procedure was evaluated in terms of linearity, sensitivity and precision
(Table 7). Calibration curves were constructed for each analyte using twelve
standards prepared over the range of 3 × 10-8 M to 5 × 10-5 M. As can be seen from
the correlation coefficients (R2, Table 7), all thiols and disulfides exhibited a highly
linear relationship between signal (peak area) and concentration over their respective
dynamic range. The limit of detection (LOD, defined as a signal-to-noise ratio of 3),
for all analytes was in the range of 5 × 10-8 M to 1 × 10-7 M, which is comparable to
other HPLC techniques for the detection of thiols and disulfides without
derivatisation [87, 106, 110, 117, 118]. The precision of repeated injections was
excellent for all analytes; relative standard deviations were less than 0.3% for
retention times and 2% for peak areas.
Table 7. Analytical figures of merit obtained for the thiols and disulfides using
manganese(IV) (5 × 10-8 M) chemiluminescence detection.
Analyte
Retention
R2
Liner range
(—M)
R.S.D.
(%)*
LOD
(M)
Time
(min)
R.S.D.
(%)*
cystine
2.6
0.1
0.9881
0.05 - 10
1.3
5 × 10-8
cysteine
3.2
0.1
0.9958
0.05 - 10
1.1
5 × 10-8
homocysteine
4.3
0.1
0.9985
0.05 - 7
0.8
5 × 10-8
homocystine
5.5
0.2
0.9978
0.05 - 7
1.4
5 × 10-8
GSH
7.0
0.1
0.9982
0.07 - 10
1.4
7 × 10-8
N-acetylcysteine
14.1
0.1
0.9990
0.07 - 10
1.1
7 × 10-8
GSSG
18.2
0.3
0.9982
0.1 - 10
2.0
1 × 10-7
*n = 10
3.3.2
Determination of GSH and GSSG in whole blood
As previously noted, quantifying the molar ratio of GSH to its dimer, GSSG, is one
of the most important indicators of cellular health [85, 97-99, 112]. Although these
109
CHAPTER FOUR
biomarkers are ubiquitous in all human organs, it has been suggested that their
concentrations in blood may reflect the redox status of other less accessible tissues
and hence provide a useful indicator of a whole subject’s oxidative status [88, 114,
115, 223, 224]. However, the sample processing required in most procedures for the
determination of GSH and GSSG in blood (i.e. thiol blocking and/or derivatisation,
disulfide bond reduction, sample pH neutralisation, etc.) can lead to erroneous
measurements [87, 114, 115, 224]. Therefore, the use of HPLC with manganese(IV)
chemiluminescence detection was explored for the determination of GSH and GSSG
in whole blood, where samples were simply deproteinised, centrifuged and diluted
prior to analysis.
The deproteinisation of physiological fluids prior to GSH and GSSG determination
is an indispensable step [84, 85, 87, 106, 107, 110]. Stempak et al. [120] examined a
range of commonly employed acids for the deproteinisation of blood and found that
combining samples (1:1) with perchloric acid (15% (v/v) containing 2.0 mM ETDA)
prior to storage at -80 ° C afforded the greatest GSH and GSSG stability. However,
Rossi et al. [114] noted that deproteinisation with perchloric acid or trichloroacetic
acid significantly altered the GSH/GSSG ratio in blood. The methods of sample
preparation described by both Stempak et al. [120] and Rossi et al. [114] were
examined using two blood samples, prepared in triplicate, with and without the initial
introduction of N-ethylmaleimide (NEM), which reacts with the free thiol of GSH to
prevent unwanted oxidation.
Based on the deproteinisation procedure described by Rossi et al. [114], a whole
blood sample was divided into two aliquots. The first aliquot (370 PL) was combined
with a solution of 1.5 g/L EDTA and 250 mM sodium chloride (40 PL) and then
trichloroacetic acid (600 g/L; 77 PL), centrifuged and diluted 38-fold into the mobile
110
CHAPTER FOUR
phase,
immediately prior
to
analysis
using
HPLC
with
manganese(IV)
chemiluminescence detection (Figure 61i; black trace). The second aliquot was
mixed with a solution of 1.5 g/L EDTA and 0.5 M NEM (40 PL) and trichloroacetic
acid (600 g/L; 77 PL) [114], centrifuged, diluted 38-fold into the mobile phase, and
analysed (Figure 61i; pink trace). The disappearance of the peak at 7.0 minutes
confirmed the complete reaction of GSH with NEM. A difference (55-62%) in the
GSSG peak (18.2 min) was observed for samples with and without NEM, suggesting
that significant oxidation of GSH had occurred during the deproteinisation
procedure. However, the change was much lower than that reported by Rossi et al.
[114], possibly arising from additional error introduced during their subsequent
liquid-liquid extraction, analyte derivatisation (pH 10.8, room temperature, 2 hr) and
re-acidification, prior to analysis.
To replicate the sample preparation conditions of Stempak et al. [120], a whole
blood sample was divided into two aliquots, one of which was combined (1:1) with
perchloric acid (15% (v/v) containing 2.0 mM EDTA), centrifuged and diluted
25-fold into the mobile phase, immediately prior to analysis (Figure 61ii; black
trace). The second aliquot was mixed with NEM (3.6 mM) and a Tris-HCl buffer
(0.1 M; pH 8.0) in a 2:1:1 ratio. It has previously been shown that under these
conditions, NEM reacts immediately and completely with GSH to form a thioether
derivative [18]. This solution was then acidified by combining with an equal volume
of perchloric acid (15% (v/v) containing 2.0 mM EDTA), centrifuged, and diluted
12.5-fold into the mobile phase and analysed (Figure 61ii; pink trace). The
disappearance of the GSH peak again confirmed complete reaction with NEM. In
contrast to the results shown in Figure 61i, only a small change (less than 6%) in the
111
CHAPTER FOUR
peak area for GSSG was observed, indicating that this deproteinisation procedure
does not result in significant oxidation of GSH to GSSG in whole blood samples.
180
without NEM
with NEM
GSH
(i)
Chemiluminescence intensity (arbitrary units)
160
140
120
100
80
60
40
GSSG
20
0
0
2
4
6
8
10
12
14
16
18
20
Time (min)
220
without NEM
with NEM
GSH
(ii)
Chemiluminescence intensity (arbitrary units)
200
180
160
140
120
100
80
60
40
GSSG
20
0
0
2
4
6
8
10
12
14
16
18
20
Time (min)
Figure 61. Typical manganese(IV) chemiluminescence chromatograms obtained for
the analysis of whole blood, using the deproteinisation procedures of (i) Rossi and
co-workers and (ii) Stempak and co-workers. Pink and black and traces show sample
preparation with and without the addition of NEM.
112
CHAPTER FOUR
The method was then used to determine GSH and GSSG in whole blood from
twelve healthy volunteers (six men and six women, 21-31 years old). The results of
the study are presented in Table 8, with values obtained for GSH (723–1260 —M)
and GSSG (120–191 —M) within the range previously reported by numerous other
researchers for healthy subjects of similar age [223-227].
Table 8. Concentrations of GSH and GSSG in whole blood.
Sex
Age (years)
GSH (—M/L)
GSSG (—M/L)
Mean
R.S.D. (%)*
Mean
R.S.D. (%)*
GSH/GSSG
M
25
914
5.6
167
1.0
5.5
M
25
990
3.2
158
2.0
6.3
F
24
723
2.8
134
4.7
5.4
F
24
1249
1.2
180
3.8
6.9
F
22
851
2.9
182
4.7
4.7
M
31
1260
1.8
163
1.0
7.7
M
30
1115
4.7
129
1.8
8.6
F
24
1074
5.4
191
2.5
5.6
F
24
843
3.1
120
7.1
7.0
F
24
880
4.8
134
3.9
6.6
M
21
1176
2.1
137
3.5
8.6
M
23
878
3.2
146
3.2
6.0
*n = 3
113
CHAPTER FOUR
4. Conclusions
This chapter presented a systematic evaluation of the relative chemiluminescence
responses from several biologically significant thiols and disulfides with
conventional
and
new
adaptations
of
permanganate
and
tris(2,2c-bipyridine)ruthenium(III) reagents, in addition to a manganese(IV) colloid.
This colloid provided a simple and sensitive approach for the direct detection of
thiols and disulfides using both FIA and HPLC methodologies. A mixture of seven
thiols and disulfides (cysteine, cystine, homocysteine, homocystine, glutathione
(GSH), glutathione disulfide (GSSG) and N-acetylcysteine) in their native forms
were separated within twenty minutes. Detection limits for these analytes ranged
from 5 × 10-8 M to 1 × 10-7 M, and the precision for retention times and peak areas
was excellent, with relative standard deviations of less than 0.3% and 2%,
respectively. This procedure was successfully applied to the determination of two
key biomarkers of oxidative stress, GSH and GSSG in whole blood taken from
twelve human subjects. Preliminary experiments showed that preparation of samples
via centrifugation, deproteinisation (using perchloric acid (15% (v/v) containing
2.0 mM EDTA)) and dilution was simple and did not result in significant GSH
oxidation. Concentrations of GSH and GSSG in the blood samples were comparable
to other reports. The findings of this chapter have resulted in two publications [228,
229].
114
CHAPTER FIVE
CHAPTER FIVE
Precision milled chemiluminescence flow-cells
for flow injection analysis and high
performance liquid chromatography
x
Introduction
x
Experimental
x
Results and discussion
x
Conclusion
115
CHAPTER FIVE
1. Introduction
The analytical utility of liquid-phase chemiluminescence detection is dependent on
the reproducibility of solution mixing, reaction kinetics and the proportion of the
transient emission exposed to a suitable photodetector. As described in the preceding
chapters, greater sensitivity can be obtained by combining exceedingly stable
chemiluminescence reagents that enable greater control of reaction rates with
machined flow-cells that more efficiently mix the reacting solutions. This work has
further demonstrated the need to refine several key design attributes of
chemiluminescence flow-cells, such as the proximity of the confluence point to the
detection zone in relation to the rate of the light-producing reaction [73, 77], the use
of reflective surfaces to direct more of the emitted light towards the photodetector
[234, 235], and increased mixing efficiency via greater deviation in the flow path
[73, 81].
To address some of the aforementioned issues, flow-cells have previously been
constructed by engraving or machining channels into Perspex [167, 168], Acetal
[169] or Teflon [73] and sealing the channels with transparent films or plates.
Similarly, chemiluminescence detection has been incorporated into microfluidic
devices fabricated from silicon, glass or polymer materials [230]. In these systems,
solutions are often merged in T- or Y-shaped channels with a linear detection zone
[165, 166, 231], which exposes only a very small volume of solution to the relatively
large photodetector window. Microscope optics have been used to partially alleviate
this limitation [232]. Alternatively, spiral [161] or meandering [163, 170, 233]
channels have been used to improve mixing and increase the volume of solution
within the detection zone. In spite of these developments, there has been very little
116
CHAPTER FIVE
optimisation of chemiluminescence flow-cell or microfluidic-reactor configuration,
which has often been limited by fabrication capabilities.
Other flow-cell designs have also been explored [73, 77, 81, 83, 234-238], such as
the fountain [83], sandwich [77], liquid core waveguide [234], bundle [81], vortex
[236], and droplet [237] flow-cells, but as yet none have attained widespread
acceptance, generally due to the complexity of fabrication, current lack of evidence
for significant advantages over the coiled-tubing approach, and/or limited
applicability.
This chapter details the use of computer-aided design and high-precision
machining technology to create novel flow-cells with integrated confluence points
and reaction zones
for highly sensitive
detection using relatively fast
chemiluminescence reactions. These new designs were investigated in direct
comparison with the conventional coiled-tubing approach and several previously
described flow-cells to determine their effect on the generation and transmission of
light to the photodetector.
117
CHAPTER FIVE
2. Experimental
2.1
Chemicals and reagents
Deionised water (Continental Water Systems, Victoria, Australia) and analytical
grade reagents were used unless otherwise stated. Chemicals were obtained from the
following sources: amoxicillin, epinephrine hydrochloride, fenoterol hydrobromide,
isoprenaline hydrochloride, luminol, metaproterenol hemisulfate salt, octopamine,
sodium polyphosphate (+80 mesh), sodium thiosulfate, synephrine, trifluoroacetic
acid, tyramine and vanilmandelic acid from Sigma-Aldrich (NSW, Australia);
N-methyltyramine from CSIRO Animal Health Laboratories (Victoria, Australia);
salbutamol
hemisulfate
from
Fluka
(NSW,
Australia);
morphine
from
GlaxoSmithKline (Victoria, Australia); potassium permanganate from Chem-Supply
(SA, Australia); methanol and sulfuric acid from Merck (Victoria, Australia)
potassium ferricyanide and sodium hydroxide from Ajax (NSW, Australia).
Hordenine was synthesised as previously described [22].
Morphine (1 × 10-3 M) was prepared in acidified deionised water and sonicated to
aid dissolution. Vanilmandelic acid (1 × 10-3 M) was prepared by dissolving the solid
in sodium hydroxide (0.1 M) and diluting in deionised water as required. The Citrus
aurantium compounds (hordenine, N-methyltyramine, octopamine, synephrine and
tyramine, 1 × 10-3 M) were prepared in deionised water and diluted into the HPLC
mobile phase (98% Solvent A: deionised water adjusted to pH 2.15 with
trifluoroacetic acid; 2% Solvent B: methanol) as required. The ȕ-adrenergic agonists
(epinephrine, fenoterol, isoprenaline, metaproterenol and salbutamol, 1 × 10-3 M)
were prepared and diluted in deionised water adjusted to pH 2.5 with trifluoroacetic
acid.
118
CHAPTER FIVE
The ‘standard’ acidic potassium permanganate reagent was prepared by dissolution
of potassium permanganate (1 × 10-3 M) in 1% (m/v) sodium polyphosphate and
adjusting to pH 2.5 with sulfuric acid.
The ‘enhanced’ acidic potassium permanganate reagent was prepared by
dissolution of potassium permanganate (1.9 × 10-3 M) in 1% (m/v) sodium
polyphosphate and adjusting to pH 2.5 with sulfuric acid and then adding sodium
thiosulfate (0.6 or 1.0 mM), using a small volume of a 0.1 M solution.
The luminol reagent (3.5 × 10-6 M) was prepared by dissolution of the solid in
deionised water, following the addition of a small volume of aqueous sodium
hydroxide (1.0 M), whilst potassium hexacyanoferrate(III) was prepared by
dissolution of potassium ferricyanide (3 × 10-4 M) in sodium hydroxide (0.35 M).
To obtain intense luminol chemiluminescence suitable for visual examination, the
luminol reagent was prepared by dissolution of luminol (1.1 × 10-3 M) in sodium
hydroxide (0.1 M), whilst potassium hexacyanoferrate(III) (potassium ferricyanide,
3 × 10-2 M) was prepared in deionised water containing 4% (v/v) hydrogen peroxide
[173].
2.2
Flow-cells
Flow-cell A was constructed by mounting a tight coil (28 mm diameter) of
transparent PTFE tubing (0.8 mm i.d.; DKSH) onto a thin metal sheet (35 × 60 mm).
The tubing at the centre of the coil passed through a slit in the sheet and was
connected to a barbed plastic T-piece by slipping silicone tubing (1.02 mm i.d.;
DKSH) over both the tubing and the fitting. The distance from the confluence point
to the beginning of the coil was 1 cm (approximately 5 —L).
119
CHAPTER FIVE
Flow-cell C was a Teflon disk with a machined serpentine channel (0.76 mm width
× 0.89 mm depth), contained within a commercially available GloCel
chemiluminescence detector (as described in chapter 2, section 2.4).
Flow-cell D was a modified version of flow-cell C, containing two solution inlets
at the centre of the Teflon disk and an appropriately adjusted back plate.
Flow-cells G and H were identical to flow-cells E and F, except that the disks were
constructed from Teflon impregnated with 25% glass microspheres to improve the
transmission of light to the photomultiplier tube.
Several novel flow-cells with integrated confluence points and reaction channels
designed for efficient mixing of fast chemiluminescence systems were constructed
by machining opposing sides of a polymer chip and sealing the channels with
transparent epoxy-acetate films. Prior to fabrication, 3D models of the flow-cells
were drawn using Autodesk Inventor (Autodesk Inc., San Rafael, US), with the
exception that the sinusoidal reaction zone was calculated using Visual Basic
(Microsoft) and imported into AutoCAD (AutoDesk). Each flow-cell had dimensions
of 46 u 57 u 6 mm and contained a 24-28 mm diameter reaction zone – such as the
spiral-configuration (Figure 62i) – and a confluence point located below the central
inlet of the reaction zone (Figure 62ii). EdgeCAM software (EdgeCam UK, Reading,
UK) was used to convert the 3D models into machine code for the CNC precision
milling machine (Datron CAT3D M6, Datron Technology Ltd, Milton Keynes, UK).
The channels were machined into 6 mm thick sheets of polycarbonate or white acetal
(both RS, Corby, UK) and sealed using a transparent epoxy-acetate film [230], which
also served as the detection window on the reaction face of the chip. Solution lines
were connected to the flow-cells using 062 Minstac screw-in fittings (Aviaquip,
Cheltenham, Victoria, Australia).
120
CHAPTER FIVE
(i)
(ii)
Figure 62. 3D model of a spiral-configuration flow-cell: (i) top face (solid view),
(ii) bottom face (transparent view).
Flow-cell L was constructed by machining a spiral channel (0.44 × 0.44 mm) into
the front face of polycarbonate chip (46 × 57 × 6 mm) and sealing the channels with
a transparent epoxy-acetate film. A central entrance located above the confluence
point was machined onto the back face.
Flow-cell M was identical to flow-cell L except a mirror was attached to the back
face of the flow-cell to enhance the transmission of light to the photodetector.
Flow-cell N was a modification of flow-cell L, containing a 0.7 × 0.7 mm spiral
channel machined into the front face of polycarbonate chip.
Flow-cell O was identical to flow-cell N except a mirror was attached to the back
face of the flow-cell to enhance the transmission of light to the photodetector.
Flow-cell P was a modification of flow-cell N containing a 0.8 × 0.8 mm spiral
channel machined into the front face of polycarbonate chip.
Flow-cell Q was identical to flow-cell P except a mirror was attached to the back
face of the flow-cell to enhance the transmission of light to the photodetector.
Flow-cell R was a modification of flow-cell N, containing a spiral channel
(0.7 × 0.7 mm), machined into an opaque Acetal chip and sealed with a transparent
epoxy-acetate film. The total volume of the detection zone was 144 —L.
121
CHAPTER FIVE
Flow-cell S was a modification of flow-cell R, containing a sinusoidal channel
configuration (0.7 × 0.7 mm), machined into an opaque Acetal chip. The total
volume of the detection zone was 133 —L.
Flow-cell T was a modification of flow-cell R, containing a serpentine channel
configuration (0.7 × 0.7 mm), machined into an opaque Acetal chip (Figure 63). The
total volume of the detection zone was 198 —L.
Figure 63. Flow-cell T: an opaque Acetal chip with serpentine reaction channel.
Flow-cell U was identical to flow-cell T, containing a serpentine channel
configuration (0.7 × 0.7 mm), machined into polycarbonate (Figure 64) and a mirror
attached to the back face of the flow-cell to enhance the transmission of light to the
photodetector.
Figure 64. Flow-cell U: a polycarbonate chip with serpentine reaction channel.
Flow-cell W was a novel variation to the conventional coiled-tubing design. The
tubing was glued into a spiral channel machined into a polished aluminium plate
122
CHAPTER FIVE
(Figure 65). Acting as a reflector, the machined channel had a semi-spherical profile
matching the radius of the tubing.
(ii)
(i)
Figure 65. Construction of flow-cell W: (i) a spiral channel machined into an
aluminium plate, (ii) PTFE-PFA tubing (0.8 mm i.d.) glued into the channel.
Flow-cell X was a commercially available borosilicate chip (48 × 57 × 11 mm)
containing a spiral channel (1 mm i.d.) sealed by thermal bonding (Hellma,
Singapore). The confluence point was located below the central entrance to the spiral
(Figure 66). Inlet and outlet ports had ¼Ǝ-28 thread for conventional fittings. The
back of the cell was coated with a mirror surface.
Figure. 66 Flow-cell X: a borosilicate chip with spiral reaction channel.
Flow-cell Y was identical to flow-cell X, except that all surfaces not exposed to the
PMT were covered with aluminium foil.
123
CHAPTER FIVE
2.3
Flow injection analysis
The FIA manifold was constructed from a Gilson Minipuls 3 peristaltic pump
(John Morris Scientific, Balwyn, Victoria, Australia) with bridged PVC pump tubing
(1.02 mm i.d.; DKSH), black manifold tubing (0.76 mm i.d., Global FIA, WA,
USA), and six-port injection valve (Vici 04W-0192L, Valco Instruments, Houston,
Texas, USA) with 70 ȝL injection loop. The output signal from the photomultiplier
module was captured on a chart recorder (3066 Pen Recorder, Yokogawa Electric
Works, Tokyo, Japan). Peak heights were measured manually.
The analytes (1 × 10-5 M) were injected (70 —L) into a deionised water carrier
stream that merged with the acidic potassium permanganate reagent within the
reaction channel of the flow-cell (Figure 67). Solution flow rates were optimised
over the range 0.9 mL/min to 3.5 mL/min per line, for each reaction within each
flow-cell.
Detector
Reagent
Flow-cell
Peristaltic
pump
PMT
Injection valve
Carrier
Waste
Analyte
Figure 67. Schematic of the FIA manifold used for acidic potassium permanganate
chemiluminescence measurements.
2.4
High performance liquid chromatography
Analyses were carried out on an Agilent Technologies 1200 series liquid
chromatography system, equipped with a quaternary pump, solvent degasser system
and autosampler (Agilent Technologies, Victoria, Australia), using a Synergi
Hydro-RP 80A (250 mm × 4.6 mm i.d., 4 —m) (for the C. aurantium compounds) or
124
CHAPTER FIVE
an Alltech Alltima C18 column (250 mm × 4.6 mm i.d., 5 —m) (for the ȕ-adrenergic
agonists) with an injection volume of 20 —L and a flow rate of 1 mL/min. An
analogue to digital interface box (Agilent Technologies) was used to convert the
signal from the chemiluminescence detector. Before use in the HPLC system, all
sample solutions and solvents were filtered through a 0.45 ȝm nylon membrane.
For the separation of C. aurantium compounds, isocratic elution was performed
with 98% solvent A: deionised water adjusted to pH 2.15 with trifluoroacetic acid
and 2% solvent B: methanol. The column eluate (1.0 mL/min) and acidic potassium
permanganate reagent (1.0 mL/min) merged at a confluence point located
immediately prior or within the detection zone of the flow-cells (Figure 68).
For the separation of ȕ-adrenergic agonists, gradient elution was performed with
deionised water adjusted to pH 2.5 with trifluoroacetic acid (solvent A) and methanol
(solvent B) as follows: 0-7 min: 10% B, 7-9 min: 10-50% B, 9-10 min 50% B,
10-11 min: 50-40% B, 11-12 min: 40-10% B, 12-13 min: 10% B.
HPLC pump
Detector
Column
Flow-cell
Injector
PMT
Reagent
Peristaltic
pump
Waste
Figure 68. Schematic of the HPLC manifold employed for acidic potassium
permanganate chemiluminescence measurements.
125
CHAPTER FIVE
3. Results and discussion
3.1
Channel size
Three spiral configuration flow-cells were prepared by machining channels with a
width and depth of 0.44 u 0.44 mm, 0.7 u 0.7 mm and 0.8 u 0.8 mm into transparent
polycarbonate chips (Figure 62) and sealing with an epoxy-acetate film. The spirals
comprised eight turns of the 0.7 mm or 0.8 mm channels or fifteen turns of the
0.44 mm channel. The channels have cross sectional areas similar to tubing with
internal diameters of 0.5, 0.8 and 0.9 mm (Table 9). The glass or polymer tubing
used in conventional coiled flow-cell chemiluminescence detectors typically has an
internal diameter between 0.5 and 1 mm.
Table 9. Comparison between cross sectional area of tubing and square channels.
Tubing
Square channels
2
Diameter
Area (mm )
Width
Area (mm2)
0.5
0.196
0.44
0.194
0.8
0.503
0.7
0.490
0.9
0.636
0.8
0.640
These flow-cells were compared using the rapid chemiluminescence reaction of
morphine with acidic potassium permanganate. Using FIA methodology, six
morphine standards (between 1 u10-10 M and 1 u 10-5 M) were injected into a
deionised water carrier stream that merged with the permanganate reagent at the
confluence point within the chips (Figure 62ii). Increasing the channel dimensions
from 0.44 u 0.44 mm (flow-cell M) to 0.7 u 0.7 mm (flow-cell O) improved the
chemiluminescence signal by an average of 30% (Figure 69, blue columns). A
further 27% (average) improvement was obtained using channel dimensions of
126
CHAPTER FIVE
0.8 u 0.8 mm (flow-cell Q). In both cases, the percentage differences were greater at
higher morphine concentrations.
The transmission of light to the photomultiplier tube was enhanced by 36-99%
(0.44 mm channel), 25-45% (0.7 mm) and 18-30% (0.8 mm) by placing a mirror
against the back face of the chip to reflect light back to the detector (flow-cells N, P
6
12
5
10
4
8
3
6
2
4
1
2
0
Signal/cross sectional area
Chemiluminescence intensity (mV)
and R respectively, Figure 69, pink columns).
0
0.44 × 0.44
0.7 × 0.7
0.8 × 0.8
Channel width and depth (mm)
Figure 69. Chemiluminescence intensities (peak heights) obtained for the
polycarbonate chips (blue columns) and with the addition of a backing mirror (pink
columns), for the reaction of morphine (1 × 10-7 M) with permanganate. Black
points: chemiluminescence signal divided by the cross-sectional area of the channel
(from data obtained without a backing mirror).
Whilst the absolute signals were lower for the flow-cells M and N (0.44 mm
channels) than those with flow-cells O and P (0.7 mm channels) or Q and R (0.8 mm
channels), the ratio of signal intensity to channel cross sectional area was superior
with flow-cells M and N (Figure 69, black points).
127
CHAPTER FIVE
3.2
Chip material
Even with a backing mirror to enhance the transmission of light to the
photomultiplier tube, the chemiluminescence signal from the polycarbonate
flow-cells was lower than that obtained using the conventional coiled-tubing
approach (flow-cell A). A visual inspection of the light emanating from the relatively
long-lived and intense chemiluminescence reaction of luminol with hydrogen
peroxide and potassium hexacyanoferrate(III) in alkaline solution within the
polycarbonate flow-cells revealed that a significant portion of the light was
transferred out of the sides of the chips (Figure 70).
Figure 70. Chemiluminescence from the reaction between luminol, potassium
hexacyanoferrate(III) and hydrogen peroxide within a polycarbonate chip with a
backing mirror (exposure time 1.3 s).
Flow-cell O (0.7 u 0.7 mm channels) was then compared to a flow-cell with
identical configuration that had been machined into white acetal (flow-cell S; both
were sealed with a transparent epoxy-acetate film), and flow-cell A (0.8 mm i.d.),
using the reactions of morphine with permanganate, and luminol with potassium
hexacyanoferrate(III) enhanced by vanilmandelic acid. Compared to the luminol
system used for visual examination (Figure 70), the reaction without the presence of
hydrogen peroxide produces a relatively weak emission. Vanilmandelic acid can be
128
CHAPTER FIVE
determined based on its enhancing effect [177], which is most prominent in the first
few seconds after the solutions are merged.
More light was transferred to the photomultiplier tube from flow-cell S than
flow-cells O or A, for both chemiluminescence reactions (Figure 71). The difference
could be observed visually, using the intense chemiluminescence reaction of luminol
with hydrogen peroxide and potassium hexacyanoferrate(III) (Figure 72). To ensure
identical chemical conditions, the reactant solutions were simultaneously pumped
20
1000
15
750
10
500
5
250
0
Luminol chemiluminescence intensity (mV)
Permanganate chemiluminescence intensity (mV)
through both chips.
0
A
O
S
Flow-cell
Figure 71. Chemiluminescence intensities (peak heights) obtained from the
transparent and opaque flow-cells with identical spiral channel design
(0.7 × 0.7 mm) and a coil of tubing (0.8 mm i.d.), for the reaction of morphine
(1 u 10-7 M) with permanganate (blue columns) and the reaction of luminol with
hexacyanoferrate(III) enhanced by vanilmandelic acid (1 u 10-6 M, pink columns).
129
CHAPTER FIVE
(ii)
(i)
Figure 72. Visual comparison of the emission intensity from (i) flow-cell S and (ii)
flow-cell O with an identical channel design (exposure time 0.5 s).
3.3
Channel design
As demonstrated in chapter 2, the emission of light from spiral reactors can be
enhanced by introducing a series of reversing turns, which has been referred to as a
‘serpentine’ configuration. A comparison was therefore undertaken between white
acetal chips that incorporate spiral (Figure 62i) and serpentine (Figure 73) reaction
zones, and a third design, in which a sinusoidal pattern with a wavelength of
2.35 mm and amplitude of 0.55 mm was imposed on the spiral channel (flow-cells S,
T and U respectively). Each flow-cell had an integrated confluence point (Figure 62i)
and was sealed on both sides with a transparent epoxy-acetate film.
Figure 73. The serpentine channel configuration incorporated into flow-cell T.
130
CHAPTER FIVE
These novel flow-cells were then compared to flow-cell A (conventional coiledtubing approach). In general, the signals from the opaque chips were two-fold greater
than those from flow-cell A (Figure 74). The response from flow-cells T and U were
on average 7.8% and 2.8% greater than flow-cell S. The difference was greater at
lower concentrations; at a morphine concentration of 1 × 10-9 M, the signals from
flow-cells T and U were 20.2% and 8.1% greater, respectively.
20
Chemiluminescence intensity (mV)
18
16
14
12
10
8
6
4
2
0
A
S
T
U
Flow-cell
Figure 74. The effect of channel design on chemiluminescence intensity (peak
height). Data for the reaction of 1 u 10-7 M morphine with permanganate shown.
A visual examination of the light emitted from each flow-cell using the reaction of
luminol with hydrogen peroxide and potassium hexacyanoferrate(III) revealed
significant differences in solution mixing. In each case, this relatively long-lived
reaction did not reach its maximum intensity until the solution had travelled a short
distance in the reaction channel (Figure 75). In flow-cell S (spiral), the increase in
intensity during the initial mixing period was most obvious near the walls of the
channel (Figure 75i). This was attributed to the slower movement of solution in those
131
CHAPTER FIVE
regions (compared to the centre of the channel), which allowed greater reaction time
over the same distance. The configuration of flow-cell U appeared to disrupt the
gradual onset of light from the sides of the channel and reduced the distance required
to reach the maximum intensity (Figure 75iii). However, the distribution of light
during the initiation period suggested a slower movement of solution in the crests
and troughs of the sinusoidal design (Figure 75iv). A more rapid onset of the
emission was also observed in flow-cell T (serpentine), but the emission was more
evenly distributed across the channel (Figure 75ii).
(i)
(ii)
(iii)
(iv)
Figure 75. Chemiluminescence from the reaction between luminol, potassium
hexacyanoferrate(III) and hydrogen peroxide in the opaque (Acetal) flow-cells (i) S,
(ii) T, (iii) U and (iv) zoomed in image of the initial turns of flow-cell U (contrast
adjusted and converted to greyscale (0.3 s exposure time).
The sudden changes in the direction of flow within flow-cell U, and to a greater
extent within flow-cell T, create disturbances in a similar manner to knitted open
tubular reactors [51] used for on-line derivatisation and photochemistry. This results
in increased radial mixing and decreased axial dispersion, which in this case leads to
greater chemiluminescence intensity. The effect of the channel configuration on
132
CHAPTER FIVE
mixing efficiency was evident in photographs of the chemiluminescence from the
reaction between the permanganate reagent and morphine, using an extended
exposure time of 25 seconds (Figure 76).
In flow-cell S, the most intense emission occurred within the detection zone, but a
significant amount of light was emitted as the solution passed along the exit channel
(Figure 76i). Whilst in flow-cell U, the emission in the detection zone was more
intense, and only a faint emission was observed in the exit line (Figure 76iii). In
flow-cell T, the exit line was barely visible, even when using a longer exposure time
of 90 seconds, confirming that under these conditions the chemiluminescence
reaction was essentially complete before the solution exited the detection zone
(Figure 76ii).
(i)
(ii)
(iii)
Figure 76. Chemiluminescence from the reaction between the standard
permanganate reagent and morphine (1 × 10-3 M) in the opaque (Acetal) flow-cells
(i) S, (ii) T and (iii) U (25 s exposure time).
3.4
Flow-cell comparison
As detailed in chapter 2, studies by our research group and others have shown:
(i) Repeated changes in the direction of solution flow (such as reversing turns)
enhances the mixing efficiency and therefore the emission intensity [73, 81, 160,
231]. (ii) Relatively large chambers or wells in the detection zone generally delay the
onset of the maximum emission as the solution rapidly moves through the path of
least resistance, leaving surrounding areas of poor solution mixing [73, 188].
133
CHAPTER FIVE
(iii) Channels machined or etched into an opaque white polymer and sealed with a
transparent seal minimise the loss of light through surfaces not exposed to the
photodetector [160, 169, 170]. (iv) A mixing tube/channel prior to the entrance of the
detection zone is normally required for the reacting mixture to reach maximum
emission intensity in front of the photodetector [71, 160, 231], but in the case of very
rapid chemiluminescence systems, merging the solutions within the detection zone
may be superior [188].
A variety of flow-cells that incorporate one or more of these advances were
compared to the best opaque flow-cell (T) and the conventional coiled-tubing
approach (A) to determine the effectiveness of each on the generation and
transmission of light to the photomultiplier tube.
3.4.1
Flow injection analysis comparison
Ten flow-cell configurations were initially compared using the chemiluminescence
reactions of morphine and amoxicillin with two acidic potassium permanganate
reagents. As previously mentioned, the reaction of morphine with permanganate has
been used extensively in both flow analysis and HPLC applications [25, 26].
Amoxicillin is a ȕ-lactam antibiotic that has been detected with various
permanganate
chemiluminescence
reagent
systems
[232-234].
Previous
investigations involving these and other analytes (detailed in chapter 3) have shown
that considerable increases in reaction rate and chemiluminescence intensity can be
obtained by a preliminary partial reduction of permanganate to create a stable,
relatively high concentration of manganese(III) in the reagent solution [205, 213].
The flow-cells were therefore also compared with the enhanced permanganate
reactions as examples of extremely rapid chemiluminescence systems.
134
CHAPTER FIVE
Using FIA methodology, the analytes were injected (10 replicates) into a water
carrier stream, which merged with the permanganate reagent at a flow rate of
3.5 mL/min. In agreement with previous work [213], the enhanced reagent provided
a considerable increase in chemiluminescence intensity (4-5 fold for morphine and
32-38 fold for amoxicillin for most flow-cells; Table 10). Greater enhancement
(7-fold for morphine and 50-58 fold for amoxicillin) was observed for flow-cells F
and H (dual-inlet serpentine machined into Teflon and Teflon impregnated with glass
microspheres, respectively), where the merging of reactants within the detection zone
enabled a greater proportion of the light emitted from these reactions to be captured
(Figures 77 and 78).
Table 10. Improvement of the chemiluminescence response obtained from the
enhanced permanganate reagent compared to the standard reagent with morphine
(1 × 10-7 M) and amoxicillin (1 × 10-5 M), using FIA with the ten different
flow-cells.
Enhancement factor
Flow-cell
Morphine
Amoxicillin
A
4.8
34.9
C
4.9
37.4
D
6.5
50.2
G
5.1
37.7
H
7.1
52.7
T
4.5
32.4
V
4.1
33.0
W
5.2
37.6
X
4.2
37.8
Y
4.6
35.6
Several flow-cells provided greater chemiluminescence signals (Figure 77) than
the traditional coiled-tubing approach (flow-cell A). Flow-cell W enabled the tubing
to be tightly coiled on a reflective backing plate (Figure 79), the advantage of which
135
CHAPTER FIVE
was most evident using the enhanced permanganate reactions. The GloCel detector
(flow-cells C, D, G and H) provided a relatively inexpensive commercial option with
a purpose-built housing. Machining the reaction channels into disks constructed from
Teflon impregnated with 25% glass microspheres (flow-cells G and H) increased the
signal intensities by up to 23%.
Although flow-cell T was also constructed by machining channels into a white
polymer material, it produced significantly larger chemiluminescence signals
(between 36% and 210% greater than the signals for flow-cells C, D, G and H;
Figures 77 and 78). These differences can be attributed to the extremely thin
transparent seal and the ability to place the photomultiplier tube flush against the
20
90
18
80
16
70
14
60
12
50
10
40
8
30
6
20
4
10
2
0
0
A
C
D
G
H
T
V
W
X
Y
Enhanced permanganate chemiluminescence intensity (mV)
Permanganate chemiluminescence intensity (mV)
chip, which enables a greater transfer of light to the photodetector.
Flow-cell
Figure 77. Comparison of chemiluminescence detection flow-cells for the oxidation
of morphine (1 × 10-7 M) with the standard permanganate reagent (blue columns)
and the enhanced permanganate reagent (pink columns).
136
12
360
10
300
8
240
6
180
4
120
2
60
0
0
A
C
D
G
H
T
V
W
X
Y
Enhanced permanganate chemiluminescence intensity (mV)
Permanganate chemiluminescence intensity (mV)
CHAPTER FIVE
Flow-cell
Figure 78. Comparison of chemiluminescence detection flow-cells for the oxidation
of amoxicillin (1 × 10-5 M) with the standard permanganate reagent (blue columns)
and the enhanced permanganate reagent (pink columns).
(i)
(ii)
(iii)
(iv)
Figure 79. Chemiluminescence from the reaction between morphine (1 × 10-3 M)
and the enhanced permanganate reagent in flow-cells (i) A, (ii) H, (iii) T and (iv) W.
A 20 s exposure time was used for each photograph.
137
CHAPTER FIVE
In agreement with work presented earlier in this chapter [160], flow-cells U and X,
which were both constructed from transparent materials and backed with a mirror,
produced much lower signals due to the loss of light through surfaces not exposed to
the photodetector (Figure 80). The addition of a layer of aluminium foil around the
surfaces of flow-cell Y enhanced the transmission of light to the detector by 25-44%.
However, the absolute signals were still lower than most of the other flow-cells.
(i)
(ii)
Figure 80. Chemiluminescence from the reaction between morphine (1 × 10-3 M)
and the enhanced permanganate reagent in flow-cells (i) U and (ii) X. A 30 s
exposure time was used for each photograph.
The greater chemiluminescence intensity obtained using flow-cell T also translated
to a slightly lower limit of detection for amoxicillin (9 × 10-8 M) compared to the
conventional coiled-tubing approach (2 × 10-7 M). Using the enhanced reagent, the
detection limit was further extended to 3 × 10-9 M, which is superior to all previously
reported values for this analyte with permanganate chemiluminescence [232, 233,
235].
3.4.2
High performance liquid chromatography
The flow-cells that provided the greatest chemiluminescence intensities with the
enhanced permanganate system (H and T) were evaluated for post-column HPLC
detection, in comparison with the conventional coiled-tubing approach (A). An initial
comparison of these flow-cells was undertaken for the detection of a mixture of
138
CHAPTER FIVE
Į-adrenergic agonists from C. aurantium (Figure 81), due to the recent interest in
their potentially adverse health effects as ingredients in weight-loss supplements
[236]. Work previously conducted by Slezak et al. [21] and Percy et al. [22] has
demonstrated the rapid determination of these analytes with monolithic column
chromatography and acidic permanganate chemiluminescence detection. More
recently, it has been shown that the emission intensities with these analytes can be
improved by up to two-orders of magnitude with enhanced permanganate reagents,
largely due to increased reaction rates that were more compatible with HPLC-based
detection [205].
Figure 81. Structures of the Į-adrenergic agonists from C. aurantium.
In this study a reverse-phase particle-packed column designed for highly aqueous
mobile phases (98% Solvent A: deionised water adjusted to pH 2.15 with
trifluoroacetic acid; 2% Solvent B: methanol) was employed. Separation times with
this column were much longer than those in previous reports, providing greater peak
resolution [21, 22].
Using flow-cell A, the enhanced reagent provided 15- to 49-fold larger signals than
the standard reagent (Figure 82). The use of flow-cells H and T, which both
contained channels machined into white polymer materials, further increased signal
intensities by 10% and 55% respectively.
139
CHAPTER FIVE
480
OCT
SYN
Chemiluminescence intensity (arbitrary units)
420
360
TYR
300
N-MET
HOR
240
180
120
60
0
0
4
8
12
16
20
24
28
Time (min)
Figure 82. HPLC separation of C. aurantium compounds. Peaks: octopamine
(OCT), synephrine (SYN), tyramine (TYR), N-methyltyramine (N-MET) and
hordenine (HOR). Detection conditions: flow-cell A using the standard
permanganate reagent (black trace); and flow-cells A (blue trace), H (pink trace) and
T (yellow trace) using the enhanced permanganate reagent.
The flow-cells were also examined for the separation and detection of ȕ-adrenergic
agonist pharmaceuticals (Figure 83), which required higher proportions of organic
modifier in the chromatographic mobile phase (up to 50% solvent B: methanol). Poor
compatibility between the epoxy resin (used to attach the transparent seal to
flow-cell T) and the HPLC mobile phase meant that it was not suitable for use in this
analysis.
140
CHAPTER FIVE
Figure 83. Structures of the ȕ-adrenergic agonist pharmaceuticals.
As only fenoterol had previously been detected with the enhanced reagent [213],
the concentrations of permanganate and thiosulfate required to generate the greatest
chemiluminescence intensities with each analyte were optimised using FIA
(Figure 84), prior to implementing this detection system on the HPLC. Whilst the
degree of enhancement was somewhat analyte dependant, the greatest intensities
were generally obtained using a reagent containing 1.9 mM permanganate and
1.0 mM thiosulfate.
141
CHAPTER FIVE
70
(i)
Chemiluminescence intensity (mV)
Chemiluminescence intensity (mV)
60
50
40
30
20
10
0
0.0
0.3
0.6
0.9
1.2
(ii)
60
50
40
30
20
10
0
1.5
0.0
Sodium thiosulfate concentration (mM)
400
Chemiluminescence intensity (mV)
50
40
30
20
10
0
0.0
0.3
0.6
0.9
1.2
0.9
1.2
(iv)
300
250
200
150
100
50
0
1.5
0.0
0.3
0.6
0.9
1.2
350
300
250
200
150
100
50
0
0.3
1.5
Sodium thiosulfate concentration (mM)
(v)
0.0
1.5
350
Sodium thiosulfate concentration (mM)
400
0.6
Sodium thiosulfate concentration (mM)
(iii)
Chemiluminescence intensity (mV)
Chemiluminescence intensity (mV)
60
0.3
0.6
0.9
1.2
1.5
Sodium thiosulfate concentration (mM)
Figure 84. Optimisation of thiosulfate concentrations in reagents containing
1.0 mM permanganate (blue points) or 1.9 mM permanganate (pink points), using
FIA for (i) epinephrine, (ii) isoprenaline, (iii) metaproterenol, (iv) salbutamol and
(v) fenoterol.
142
CHAPTER FIVE
The use of the enhanced reagent resulted in significant increases in
chemiluminescence intensity for both salbutamol and fenoterol (Table 11). Moderate
enhancement was obtained for metaproterenol (11-fold), and as a consequence, the
extremely low response obtained with the standard reagent became comparable with
that of epinephrine and isoprenaline. Similar trends were observed in the HPLC
experiments (Figure 85 and Table 13), with discrepancies between the two
experimental approaches attributable to differences in flow rates and the proportion
of organic modifier in the carrier/eluent stream.
Table 11. Relative chemiluminescence intensities (peak heights) for the
E-adrenergic agonists with the two permanganate reagents, using FIA methodology.
Chemiluminescence signal (mV)
Analyte
Permanganate
Enhanced
Permanganate
Enhancement
Factor
Epinephrine
16.5
55.2
3.3
Isoprenaline
22.0
62.5
2.8
Metaproterenol
5.0
56.5
11.3
Salbutamol
16.6
326.6
20.8
Fenoterol
12.2
394.2
30.5
The limits of detection for the five analytes using FIA under these conditions (with
flow-cell H) ranged from 1 × 10-10 M to 5 × 10-9 M (compared with 3 × 10-8 M to
1 × 10-7 M using the standard permanganate reagent, Table 12), which were better
than all previously reported values for epinephrine, isoprenaline, salbutamol and
fenoterol using permanganate chemiluminescence systems [213, 237-241].
Metaproterenol had not previously been detected with this reagent.
143
CHAPTER FIVE
Table 12. Limits of detection for the E -adrenergic agonists using FIA methodology.
Limit of Detection (M)a
a
Analyte
Standard
permanganate
Enhanced
permanganate
Epinephrine
4.7 × 10-8
4.1 × 10-9
Isoprenaline
4.4 × 10-8
5.5 × 10-9
Metaproterenol
1.5 × 10-7
5.0 × 10-9
Salbutamol
3.2 × 10-8
3.4 × 10-10
Fenoterol
2.6 × 10-8
1.3 × 10-10
Calculated as 3× the standard deviation of the blank response.
The use of flow-cell F (dual-inlet serpentine channel machined into a Teflon disk)
for post-column chemiluminescence detection provided an increase in signal of
34-69% across the five analytes, compared to flow-cell A with the same reagent
(Figure 85). Flow-cell H, in which the Teflon disk was impregnated with glass
microspheres, was on average 45% and 129% superior to flow-cells F and A,
respectively.
Table 13. Relative chemiluminescence responses (peak areas) for the E-adrenergic
agonists with the two permanganate reagents, using HPLC.
Chemiluminescence signal (mV)
Analyte
Standard
permanganate
Enhanced
permanganate
Enhancement
Factor
Epinephrine
16.5
55.2
3.3
Isoprenaline
22.0
62.5
2.8
Metaproterenol
5.0
56.5
11.3
Salbutamol
16.6
326.6
20.8
Fenoterol
12.2
394.2
30.5
144
CHAPTER FIVE
450
FEN
Chemiluminescence intensity (arbitrary units)
400
350
SAL
300
250
200
150
EPI
100
ISO
MET
8
10
50
0
0
2
4
6
12
14
16
Time (min)
Figure 85. HPLC separation of ȕ-adrenergic agonist pharmaceuticals. Peaks:
epinephrine (EPI), isoprenaline (ISO), metaproterenol (MET), salbutamol (SAL) and
fenoterol (FEN). Detection conditions: flow-cell A using the standard permanganate
reagent (black trace); and flow-cells A (blue trace), D (pink trace) and H (yellow
trace) using the enhanced permanganate reagent.
145
CHAPTER FIVE
4. Conclusions
The construction of chemiluminescence detection flow-cells based on a coil of
polymer tubing is an effective approach that has been used in both FIA and HPLC
over decades of research. In flow-cell W, the signals from this simple design have
been improved by the use of a reflective backing plate with grooves to ensure a tight
coil of the transparent PTFE-PFA tubing.
Advances in computer-aided design and precision milling technology driven by
microfluidics research can be exploited for the manufacture of flow-through cells for
highly sensitive chemiluminescence detection in flow analysis. This approach allows
exploration of channel configurations not attainable using conventional coiled-tubing
flow-cells, to create superior designs that are more suitable for commercial
production or scaling down for incorporation in microfluidic devices. Previously
reported chemiluminescence reactors constructed by machining channels in polymers
(for flow analysis or microfluidic applications) have almost exclusively been
prepared using transparent materials, but far greater emission intensities can be
obtained using opaque polymers with a transparent seal. Further increases in the
emission intensity from these relatively fast chemiluminescence reactions can be
achieved by incorporating channels that produce sudden changes in the direction of
flow to enhance mixing efficiency.
Comparison of several novel flow-cells revealed the importance of surface
reflectance. The introduction of glass microspheres in the Teflon disks of the GloCel
detector (flow-cells G and H) improved the transmission of light to the
photodetector. Even greater signals were observed from an Acetal chip with a similar
channel configuration (flow-cell D), which may in part arise due to the shorter
distance from the reaction channel to the photomultiplier window. Despite
146
CHAPTER FIVE
incorporating highly reflective back surfaces, the two flow-cells constructed from
transparent materials (polycarbonate and glass) produced inferior signals due to the
loss of light from the sides of the chips.
A further advantage of machining into polymer materials is the inclusion of
threaded inlet/outlet ports for standard fittings (which overcomes problems
associated with pressure at high flow rates) and custom designs for incorporation in
purpose-built light-tight housings. The combination of these flow-cells, which
provide efficient mixing of fast chemiluminescence reactions close to or at the point
of detection, with the enhanced permanganate reagent has improved limits of
detection for a range of organic analytes. The findings of this chapter have resulted in
two recent publications [160, 242].
147
CHAPTER SIX
CHAPTER SIX
Determination of Papaver somniferum alkaloids
in drug seizure samples using high performance
liquid chromatography with UV-absorbance
and chemiluminescence detection
x
Introduction
x
Experimental
x
Results and discussion
x
Conclusion
148
CHAPTER SIX
1. Introduction
Heroin (3,6-diacetylmorphine) is a semi-synthetic compound derived from the
acetylation of the Papaver somniferum opiate alkaloid morphine [125, 133-135].
Originally synthesised as a non-addictive morphine substitute, heroin was introduced
into the commercial market as a cough suppressant and treatment for tuberculosis
[125, 133-135]. However, as addiction to these medications rapidly increased, it was
discovered that by simply acetylating the phenolic and hydroxyl moieties (on carbons
3 and 6 respectively of the phenanthrene backbone), a narcotic even more powerful
and addictive than its parent compound had been created [125, 133-135]. The ongoing illicit production and trafficking of heroin has resulted in it being considered as
one of the most significantly abused narcotics [128, 129]. This has created the need
to identify and/or quantify the drug (and its precursors or metabolites) in suspected
drug seizure samples.
Screening tests are often utilised to obtain rapid preliminary identification of
certain drugs or classes of drugs prior to confirmatory testing with more specific
instrumental methods [125, 128, 129, 145, 146]. The most commonly used screening
method for opiates and opiate derivatives involves mixing seizure samples with
reagents such as Mandelin’s, Marquis or Mecke’s and monitoring the resultant
colour change [125, 128, 129, 145, 146]. These spot tests are usually selective for a
particular class of drug (i.e. opiates), but are not selective for specific drugs (i.e.
heroin), so a series of tests is usually performed [125, 128, 129, 145, 146].
Microcrystalline tests have been used to provide greater specificity, however, the
presence of cutting agents and diluents can result in significant alterations to the
crystal structure which can impede identification [125, 128, 129, 145].
149
CHAPTER SIX
As previously mentioned, chemiluminescence is a highly selective and sensitive
method of detection that has been applied to various flow based techniques [1-3, 68]. Our research group has demonstrated the analytical utility of chemiluminescence
reactions,
in
particular
those
of
acidic
potassium
permanganate
and
tris(2,2‫މ‬-bipyridine)ruthenium(III) for the determination of various opiate alkaloids in
process samples using FIA, SIA, HPLC and CE [148-151, 153, 154, 156, 243]. The
complimentary selectivity of these reagents for phenolic and non-phenolic opiates
served as the basis of a rapid screening test for heroin developed by Agg and
co-workers [27, 158].
In stage one of the screening test, a suspected heroin seizure sample is reacted with
the tris(2,2ƍ-bipyridine)ruthenium(III) reagent [27, 158]. The production of a strong
chemiluminescence response is suggestive that the sample may contain heroin [27,
158]. Preliminary confirmation occurs in stage two of the test which involves mixing
the seizure sample with concentrated base prior to reaction with acidic potassium
permanganate [27, 158]. If the sample contains heroin it will be rapidly hydrolysed to
morphine and 6-monoacetylmorphine, which produce intense chemiluminescence
responses with the second reagent [27, 158].
Some tertiary amines (such as codeine, strychnine and chloroquine) have been
found to elicit false positives with the tris(2,2ƍ-bipyridine)ruthenium(III) reagent [27,
158]. However, following hydrolysis, these species fail to elicit an emission with
acidic potassium permanganate to rival that of morphine and 6-monoacetylmorphine
[27, 158]. This rapid screening test has been successfully employed for the
determination of heroin in seizure samples using both FIA and SIA methodologies
[27, 158]. However, an inherent disadvantage of performing this test with these
techniques is the lack of a separative step, which has made it almost impossible to
150
CHAPTER SIX
identify the presence of multiple opiates and/or opiate impurities within these
complex matrices.
The use of HPLC coupled with chemiluminescence detection has previously been
shown to be an effective means of monitoring opiate alkaloids in process samples
[25, 26, 153, 154, 243, 244]. However, this methodology has not yet been utilised for
the rapid screening test for heroin.
Camenzuli and co-workers [245, 246] have recently designed a chromatographic
column equipped with an annular frit to enable processing of flow streams in a
manner that emulates conventional post-column detection [245, 246]. The frit is
contained in an outlet fitting equipped with multiple exit ports: a central port to
capture flow eluting from the central section of the column bed, and three exit ports
that enable the peripheral portion of flow to be captured [245, 246]. This column
technology
has
the
potential
to
allow
multiple
post-column
detectors
(i.e. UV-absorbance, chemiluminescence, fluorescence, mass spectrometry etc) to be
coupled for simultaneous analysis in a way that until now, has not been feasible
using conventional approaches.
In recent years, there have been several major advances in two of the most
prolifically
used
chemiluminescence
reagent
systems
(acidic
potassium
permanganate and tris(2,2c-bipyridine)ruthenium(III)), in terms of their stability [33,
205, 213] and sensitivity [205, 213], and the understanding of their light producing
pathways [14, 216, 217]. As such, this chapter details the implementation of a rapid
HPLC based post-column chemiluminescence screening test for opiate alkaloids and
opiate derivatives in drug seizure samples using the more stable and sensitive forms
of the acidic potassium permanganate and tris(2,2‫މ‬-bipyridine)ruthenium(III)
151
CHAPTER SIX
reagents. The novel column technology described above was utilised to
simultaneously determine these opiates with three unique modes of detection.
152
CHAPTER SIX
2. Experimental
2.1
Chemicals and reagents
Deionised water (Continental Water Systems, Victoria, Australia) and analytical
grade reagents were used unless otherwise stated. Chemicals were obtained from the
following sources: sodium polyphosphate (+80 mesh) and trifluoroacetic acid from
Sigma-Aldrich (New South Wales, Australia); sodium thiosulfate from Fluka, (New
South Wales, Australia); potassium permanganate from Chem-Supply (South
Australia, Australia); lead dioxide, sodium hydroxide and sodium perchlorate from
Ajax Finechem (New South Wales, Australia); methanol and sulfuric acid from
Merck (Victoria, Australia); glacial acetic acid and perchloric acid (70% w/v) from
Univar (New South Wales, Australia); acetonitrile from Burdick & Jackson
(Michigan, USA) and tris(2,2ƍ-bipyridine)ruthenium(II) dichloride hexahydrate from
Strem Chemicals (Minnesota, USA). Codeine, heroin (•85%), morphine, oripavine,
papaverine and thebaine were provided by GlaxoSmithKline (Victoria, Australia).
Drug seizure samples were provided by the Australian Federal Police.
The ‘enhanced’ permanganate reagent was prepared by dissolution of potassium
permanganate (1.9 × 10-3 M) in 1% (m/v) sodium polyphosphate and adjusting to
pH 2.5 with sulfuric acid and then adding sodium thiosulfate (0.6 mM), using a small
volume of a 0.1 M solution.
The ‘stabilised’ tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent was prepared as
previously described [33]. Briefly, this involved treating [Ru(bipy)3]Cl2 with sodium
perchlorate in aqueous solution to yield the characteristic bright orange
[Ru(bipy)3](ClO4)2 precipitate, which was collected by vacuum filtration, washed
twice with ice water, and dried over phosphorus pentoxide for 24 hours. The reagent
153
CHAPTER SIX
solution was prepared by oxidising the [Ru(bipy)3](ClO4)2 crystals (1 × 10-3 M) with
lead dioxide (0.2 g/100 mL) in acetonitrile containing 0.05 M perchloric acid.
Following 1 minute of vigorous mixing and the characteristic change in colour from
orange to blue-green, the resultant suspension was allowed to settle prior to filtration
through an in-line filter (consisting of a small Pasteur pipette packed tightly with
glass wool).
Stock solutions of the opiate alkaloids (morphine, oripavine, codeine, thebaine and
papaverine, 1 × 10-3 M) were prepared by dissolution in acidified deionised water.
Heroin (1 × 10-3 M) was prepared in 0.1% (v/v) acetic acid and diluted in
0.05% (v/v) acetic acid.
Stock solutions of the drug seizure samples were prepared by dissolving
approximately 15 mg of the solid material into 50 mL of 0.1% (v/v) acetic acid.
The ‘non-hydrolysed’ samples were prepared by taking 1.0 mL of stock solution
(heroin or the drug seizure sample) and diluting to either 100 mL (HPLC with
multiplexed detection) or 200 mL (FIA/ HPLC with conventional end column
detection) with a 0.05% (v/v) acetic acid solution.
The ‘hydrolysed’ samples were prepared by mixing 1.0 mL of stock solution with
100 —L of sodium hydroxide (1.0 M) and then diluting the mixture to 100 mL or
200 mL with the 0.05% (v/v) acetic acid solution.
2.2
Flow injection analysis
The FIA manifold was constructed from a Gilson Minipuls 3 peristaltic pump
(John Morris Scientific, Victoria, Australia) with bridged PVC pump tubing
(1.02 mm i.d.; DKSH, Queensland, Australia), black manifold tubing (0.76 mm i.d.,
154
CHAPTER SIX
Global FIA, Washington, USA), six-port injection valve (Vici 04 W-0192L, Valco
Instruments, Texas, USA) and an in-house fabricated detector containing
flow-cell W (a tight coil of Teflon tubing attached to an aluminium back-plate)
mounted flush against the window of an extended range photomultiplier module
(Electron Tubes P30A-05, ETP, New South Wales, Australia) within light-tight
housing. All tubing entering and exiting the detector was black PTFE (0.76 mm i.d.,
Global FIA). The output signal from the detector was recorded with an
‘e-corder 410’ data acquisition system (eDAQ, New South Wales, Australia).
The analytes (5 × 10-6 M) were injected (70 —L) into a 100% aqueous carrier
stream (deionised water adjusted to pH 2.5 with trifluoroacetic acid, 3.5 mL/min)
that merged with a flowing stream of the chemiluminescence reagent(s)
(3.5 mL/min) at a confluence point located immediately prior to the inlet of the flowcell (Figure 86).
Detector
Reagent
Flow-cell
Peristaltic
pump
PMT
Injection valve
Carrier
Waste
Analyte
Figure 86. Schematic of the FIA manifold.
2.3
High performance liquid chromatography
Analyses were carried out on an Agilent Technologies 1200 series liquid
chromatography system, equipped with a quaternary pump, solvent degasser system
and autosampler (Agilent Technologies, Victoria, Australia), using a reversed phase
Hypersil Gold chromatography column (100 mm × 4.6 mm i.d., 5 —m, ThermoFisher
155
CHAPTER SIX
Scientific, Cheshire, United Kingdom), with an injection volume of 20 —L, flow rate
of 2.5 mL/min and a UV-absorbance detector operating at 280 nm (Agilent
Technologies). An analogue to digital interface box (Agilent Technologies) was used
to convert the signal from the chemiluminescence detector. Before use in the HPLC
system, all sample solutions and solvents were filtered through a 0.45 ȝm nylon
membrane.
Opiate standards were separated via gradient elution with deionised water adjusted
to pH 2.5 with trifluoroacetic acid (solvent A) and methanol (solvent B) as follows:
0-1 min: 5-10% B, 1-2 min: 10-25% B, 2-6 min: 25-35% B, 6-6.5 min: 35% B, 6.5-8
min: 35-5% B, 8-9 min: 5% B.
2.3.1
Conventional
high
performance
liquid
chromatography
with
post-column chemiluminescence detection
The column eluate (2.5 mL/min) and chemiluminescence reagent(s) (1.0 mL/min)
merged at a confluence point located immediately prior to or within the detection
zone of the flow-cells (Figure 87). A GloCel detector (Global FIA, Washington,
USA) equipped with flow-cell D and extended range photomultiplier module
(Electron Tubes, model P30A-05, ETP, New South Wales, Australia) was used for
the permanganate experiments, whilst an in-house fabricated detector consisting of
flow-cell W and an extended range photomultiplier module within light-tight housing
was used for the tris(2,2‫މ‬-bipyridine)ruthenium(III) experiments.
156
CHAPTER SIX
HPLC pump
Detector
Column
Flow-cell
Injector
PMT
Reagent
Peristaltic
pump
Waste
Figure 87. Schematic of the HPLC manifold employed for conventional postcolumn chemiluminescence detection.
2.3.2
High performance liquid chromatography with multiplexed detection
The column eluate (2.5 mL/min) was segmented through the annular frit in such a
way that (i) 27.4% (flow rate 0.68 mL/min) exited peripheral port 1 and entered the
UV-absorbance detector, (ii) 30.4% (flow rate 0.76 mL/min) exited peripheral port 2
and
merged
with
the
stabilised
tris(2,2‫މ‬-bipyridine)ruthenium(III)
reagent
(1.0 mL/min) at a confluence point located immediately prior to flow-cell W (as
described in section 2.3.1), (iii) 29.8% (flow rate 0.74 mL/min) exited peripheral port
3 and merged with the enhanced permanganate reagent (1.0 mL/min) at a confluence
point located within the GloCel detector (Global FIA, Washington, USA) equipped
with flow-cell D (as described in section 2.3.2) and (iv) 12.4% (flow rate
0.31 mL/min) exited the central port to waste (Figure 88).
157
CHAPTER SIX
UV-absorbance
detector 280nm
HPLC pump
(i)
(iv)
Column
Waste
Flow-cell
Injector
(iii)
(ii)
PMT
Peristaltic
pump
Flow-cell
Waste
PMT
Where:
(i) UV-absorbance detection
(ii) Stablised tris(2,2‫މ‬-bipyridine)ruthenium(III) chemiluminescence reagent
(iii) Enhanced permanganate chemiluminescence reagent
(iv) Waste
Figure 88. Schematic of the HPLC manifold employed for multiplexed
UV-absorbance and chemiluminescence detection.
158
CHAPTER SIX
3. Results and discussion
3.1
Initial experiments
Preliminary FIA experiments were conducted to examine the responses from a
pure heroin standard solution, prepared with and without the hydrolysis step, upon
reaction with the two chemiluminescence reagents. In accordance with the
observations of Agg et al. [158], a large response was obtained from the reaction of
heroin with the stabilised tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent and a negligible
response with the enhanced permanganate reagent (Figure 89). Following hydrolysis
of
the
sample,
a
modest
response
was
observed
with
the
stabilised
tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent and an intense response observed with
enhanced permanganate.
Relative chemiluminescence intensity
1.0
0.8
0.6
0.4
0.2
0.0
Stabilised
tris(2,2'-bipyridine)ruthenium(III)
Enhanced permanganate
Chemiluminescence reagent
Figure 89. Comparison of the relative chemiluminescence responses (peak heights)
for a non-hydrolysed (blue columns) and hydrolysed (pink columns) heroin standard
(5 × 10-6 M) with the stabilised tris(2,2‫މ‬-bipyridine)ruthenium(III) and enhanced
permanganate reagents. The chemiluminescence responses have been normalised for
each reagent.
159
CHAPTER SIX
3.1.1
Application to real samples
Six drug seizure samples were obtained from the Australian Federal Police and
examined using FIA methodology. The relative signal intensities (peak heights) with
the stabilised tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent (Figure 90) and the
enhanced permanganate reagent (Figure 91) for the non-hydrolysed and hydrolysed
samples are shown below.
The chemiluminescence signals for the hydrolysed samples upon reaction with the
tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent were between 66% and 81% of that
observed from the corresponding non-hydrolysed samples (Figure 90). This is
attributable to the conversion of heroin into two phenolic morphinan alkaloids,
6-monoacetylmorphine and morphine, which quench the response of the stabilised
tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent. The intense signals obtained from the
hydrolysed seizure samples may signify the presence of naturally occurring nonphenolic opiate impurities such as 6-acetylcodeine, which is often present in opium
extracts [128, 129].
When reacted with the enhanced permanganate reagent, the non-hydrolysed
samples produced moderate responses (Figure 91). However, following hydrolysis, a
significant increase in chemiluminescence intensity was observed. Whilst the
absolute chemiluminescence intensities varied, four of the non-hydrolysed seizure
samples produced responses between 3% and 11% of the corresponding hydrolysed
sample. Non-hydrolysed samples 1 and 6 produced responses that were 26% and
63% of their corresponding hydrolysed sample. This may be attributable to the
presence of significant quantities of 6-monoacetylmorphine formed via natural
degradation of the samples following synthesis.
160
CHAPTER SIX
600
Chemiluminescence intensity (mV)
500
400
300
200
100
0
1
2
3
4
5
6
Heroin seizure sample
Figure 90. Chemiluminescence responses (peak heights) for the non-hydrolysed
(blue columns) and hydrolysed (pink columns) heroin seizure samples with the
stabilised tris(2,2‫މ‬-bipyridine)ruthenium(III)reagent.
700
Chemiluminescence intensity (mV)
600
500
400
300
200
100
0
1
2
3
4
5
6
Heroin seizure sample
Figure 91. Chemiluminescence responses (peak heights) for the non-hydrolysed
(blue columns) and hydrolysed (pink columns) heroin seizure samples with the
enhanced permanganate reagent.
161
CHAPTER SIX
These results demonstrate that the combination of two chemiluminescence
reagents, coupled with a simple hydrolysis step, provide an effective method of
screening for heroin in drug seizure samples. However, an inherent disadvantage of
performing this test with FIA methodology is the lack of a separative step, which
has, to-date, prevented the identification of multiple opiates and/or opiate impurities
within these samples.
3.2
Conventional
high
performance
liquid
chromatography
with
post-column UV-absorbance detection
Camenzuli and co-workers [245, 246] have recently designed a chromatographic
column equipped with an annular frit that enabled processing of flow streams in a
manner that emulated conventional post-column detection. The capacity to
simultaneously employ multiple modes of detection should enable reductions in (i)
total analysis time, (ii) sample and/or solvent consumption and (iii) waste generation
for various analytical methodologies, including the rapid screening test for heroin.
The performance of this column was initially evaluated by sealing each of the
peripheral exit ports of the annular frit, allowing passage of the solvent through only
the central outlet (emulating conventional post-column detection, Figure 92).
Sealed
To UV-absorbance or
chemiluminescence
detector(s)
Sealed
Sealed
Figure 92. Illustration of the column outlet showing the three peripheral exit ports
and the single central exit port.
162
CHAPTER SIX
3.2.1
Optimisation of separation conditions
Initial separation experiments were conducted with UV-absorbance detection. An
optimisation of HPLC conditions (mobile phase composition, solvent gradient and
flow rate) was undertaken to rapidly separate heroin from several P. somniferum
opiate alkaloids (morphine, codeine, oripavine, thebaine, and papaverine, Figure 93).
HO
O
H
O
O
H
N
N
HO
HO
Morphine
Codeine
HO
O
O
O
N
N
O
O
Oripavine
Thebaine
O
O
O
O
H
O
O
O
N
N
O
O
Papaverine
Heroin
Figure 93. Structures of the P. somniferum opiate alkaloids.
The compatibility between the chemical parameters necessary for an efficient
chromatographic separation and sensitive chemiluminescence detection is extremely
important [5, 6, 9, 52]. Both acetonitrile and methanol have been used as an organic
modifier
in
HPLC
procedures
employing
post-column
tris(2,2‫މ‬-bipyridine)ruthenium(III) chemiluminescence detection [30, 31]. However,
several studies have shown that incorporating acetonitrile into the mobile phase
resulted in significant reductions in signal intensity or complete quenching of the
emission with acidic potassium permanganate reagents [13, 75, 153, 154, 243, 247].
In contrast, the addition of methanol causes negligible changes to the background
163
CHAPTER SIX
response and signal intensity [13, 75, 153, 154, 243, 247]. As the column eluate was
to be segmented between detectors employing tris(2,2‫މ‬-bipyridine)ruthenium(III) and
permanganate based chemiluminescence reagents, methanol was selected as the
organic modifier for this separation.
The solvent gradient and flow rate through the column were optimised to afford a
rapid separation that provided sufficient resolution between all analytes, whilst
maintaining satisfactory pressure within the column (< 200 bar). The gradient
conditions outlined by Agg and co-workers [158] were initially employed, however
poor resolution between thebaine, heroin and papaverine resulted in a modification
of the gradient to that described in the experimental. The flow rate was increased
from 1.0 mL/min to 2.5 mL/min with no loss of resolution between the six opiate
alkaloids, and a significant reduction in run time (from 11.5 minutes to 6.5 minutes;
Figure 94).
12
ORI
Signal (arbitrary units)
10
HER
8
THE
6
PAP
4
MOR
COD
2
0
0
1
2
3
4
5
6
7
8
9
10
Time (min)
Figure 94. HPLC separation of a mixture of P. somniferum opiate alkaloids
(5 × 10-6 M) with UV-absorbance detection at 280 nm. Peaks: morphine (MOR),
codeine (COD), oripavine (ORI), thebaine (THE), heroin (HER) and papaverine
(PAP).
164
CHAPTER SIX
To establish the sensitivity of UV-absorbance detection, calibration curves were
prepared for each analyte using seventeen standard solutions between 1 × 10-9 M and
1 × 10-5 M. As demonstrated by the correlation co-efficients (R2, Table 14), each of
the opiates exhibited a highly linear relationship between absorbance (peak area) and
analyte concentration. The detection limits, defined as a signal to noise ratio of 3
were in the range of 5 × 10-8 M and 5 × 10-7 M. The precision of repeated injections
was adequate, with relative standard deviations between 2.96% and 4.37%.
Table 14. Analytical figures of merit obtained with UV-absorbance detection.
Analyte
Calibration function
(peak area, arbitrary units)
R2
LOD
(M)a
R.S.D.
%*
Morphine
y= 6.29 × 105x + 0.001
0.9986
2.5 × 10-7
2.96
Codeine
y= 6.89 × 105x – 0.161
0.9980
5.0 × 10-7
3.62
Oripavine
y= 3.35 × 106x – 0.404
0.9997
1.0 × 10-7
3.41
Thebaine
y= 3.10 × 106x – 0.040
0.9996
5.0 × 10-8
4.37
Heroin
y= 7.85 × 105x + 0.303
0.9992
5.0 × 10-8
3.13
Papaverine
y= 3.80 × 106x – 0.301
0.9997
1.0 × 10-7
3.21
a
The lowest response detected with a signal to noise ratio of 3.
*n = 5; 1 × 10-6 M
3.2.2
Application to real samples
The aforementioned separation conditions were utilised to determine the heroin
content of six drug seizure samples, each of which were within ranges previously
reported by other researchers [248-254] (Table 15).
165
CHAPTER SIX
Table 15. Concentration of heroin in drug seizure samples obtained with
UV-absorbance detection.
Sample
Heroin
(mass percent)
1
48
2
25
3
69
4
46
5
85
6
27
Whilst the absolute signal intensities varied, heroin (retention time 4.7 minutes)
was detected in each of the non-hydrolysed seizure samples (Figure 95). A response
from an unknown compound was observed in five of the non-hydrolysed and
hydrolysed samples (retention time at 3.0 minutes; Figure 95i, ii, iv, v and vi). This
may be attributable to the presence of naturally occurring opiate impurities in the
opium extract used to manufacture these samples. Elucidation of this species and its
precursor would require further investigation.
Following
hydrolysis,
morphine
(retention
time
1.7
minutes)
and/or
6-monoacetylmorphine (retention time 2.9 min, Figure 95) were detected in each of
the samples, with the peak at 4.7 minutes (heroin) noticeably absent.
166
CHAPTER SIX
1.6
non-hydrolysed
hydrolysed
(i)
MOR
1.4
1.2
1.0
0.8
0.6
HER
0.4
0.2
0.0
non-hydrolysed
hydrolysed
(ii)
2.8
Signal (arbitrary units)
Signal (arbitrary units)
3.2
2.4
6-MAM
2.0
MOR
1.6
1.2
0.8
6-MAM
0.4
0
2
4
6
8
0.0
10
HER
0
2
Time (min)
0.8
0.5
0.4
8
10
HER
0.3
0.2
non-hydrolysed
hydrolysed
(iv)
6-MAM
MOR
0.8
0.6
Signal (arbitrary units)
Signal (arbitrary units)
1.0
6-MAM
MOR
0.7
6
Time (min)
non-hydrolysed
hydrolysed
(iii)
4
0.6
0.4
HER
0.2
0.1
0.0
0
2
4
6
8
0.0
10
0
2
Time (min)
4.0
non-hydrolysed
hydrolysed
(v)
4
6
8
10
Time (min)
2.0
non-hydrolysed
hydrolysed
(vi)
1.6
6-MAM
3.0
Signal (arbitrary units)
Signal (arbitrary units)
3.5
MOR
2.5
2.0
1.5
1.0
MOR
1.2
0.8
0.4
HER
HER
0.5
0.0
6-MAM
0
2
4
6
8
0.0
10
Time (min)
0
2
4
6
8
Time (min)
Figure 95. Chromatograms obtained for the analysis of six drug seizure samples
with UV-absorbance detection at 280 nm. Pink and black traces show sample
preparation with and without hydrolysis. Peaks: morphine (MOR),
6-monoaceetylmorphine (6-MAM) and heroin (HER).
167
10
CHAPTER SIX
3.3
Conventional
high
performance
liquid
chromatography
with
post-column tris(2,2ƍ-bipyrine)ruthenium(III) chemiluminescence detection
Greater selectivity and sensitivity can be derived from coupling HPLC with
chemiluminescence detection [5, 6, 9, 52]. In recent years, the limited temporal
stability of chemically oxidised tris(2,2ƍ-bipyridine)ruthenium(III) in aqueous
solutions has significantly impeded its use as a post-column chemiluminescence
reagent [1, 5, 30, 31, 33-36]. To overcome this, McDermott and co-workers [33]
recently synthesised [Ru(bipy)3](ClO4)2, which (when dissolved in acetonitrile
containing 0.05 M perchloric acid and oxidised with lead dioxide) was successfully
applied as a post-column chemiluminescence reagent. This enabled analysis without
the need for constant re-calibration or preparation of fresh reagent solutions over
extended periods of time (48 hours) [33].
Using the optimised conditions (from section 3.2.1), a mixture of the six opiates
(5 × 10-6 M) was separated. Only codeine, thebaine and heroin (retention times of
2.6, 4.3 and 4.7 minutes respectively) elicited a response with the stabilised
tris(2,2ƍ-bipyridine)ruthenium(III)
reagent,
non-phenolic morphinan alkaloids (Figure 96).
168
demonstrating
its
selectivity
for
CHAPTER SIX
2400
Chemiluminescence intensity (arbitrary units)
COD
2000
1600
1200
800
THE
400
HER
0
0
2
4
6
8
10
Time (min)
Figure 96. HPLC separation of a mixture of P. somniferum opiate alkaloids
(5 × 10-6 M) with stabilised tris(2,2ƍ-bipyridine)ruthenium(III) chemiluminescence
detection. Peaks: codeine (COD), thebaine (THE) and heroin (HER).
The effect of reagent flow rate on emission intensity was examined between
1.0 mL/min and 3.5 mL/min. A flow rate of 1.0 mL/min was found to afford the
greatest chemiluminescence response for each of the three analytes and was selected
for use in all further experiments.
Analytical figures of merit were obtained for codeine, thebaine and heroin
(Table 16). Up to two-orders of magnitude greater sensitivity was achieved for the
three
non-phenolic
morphinan
alkaloids
using
the
stabilised
ruthenium
chemiluminescence reagent (when compared to UV-absorbance). These limits of
detection (between 1 × 10-9 M and 2.5 × 10-9 M) are comparable to, or better than,
previously
reported
limits
of
detection
for
these
analytes
with
tris(2,2‫މ‬-bipyridine)ruthenium(III) chemiluminescence [35, 153, 154, 156, 157, 165].
Of further note were the improvements in linearity (R2) and precision (% R.S.D.).
169
CHAPTER SIX
Table 16. Analytical figures of merit obtained with the stabilised ruthenium reagent.
Analyte
Calibration function
(peak area, arbitrary units)
R2
LOD
(M)
R.S.D.
%*
Codeine
y= 1.40 × 109x – 10.308
0.9998
1.0 × 10-9
0.57
Thebaine
y= 7.30 × 108x + 4.306
0.9997
2.5 × 10-9
0.89
Heroin
y= 2.40 × 108x + 8.181
0.9972
1.0 × 10-9
0.74
a
The lowest response detected with a signal to noise ratio of 3.
*n = 5; 1 × 10-6 M
3.3.1
The
Application to real samples
improved
selectivity
and
sensitivity
of
the
stabilised
tris(2,2ƍ-bipyridine)ruthenium(III) reagent afforded much simpler chromatograms
when applied to the post-column determination of the six drug seizure samples.
Heroin was detected in each, with strong responses from samples 1, 3 and 5
(Figure 97i, iii and v), consistent with results obtained during the preliminary
experiments conducted with FIA methodology. The chemiluminescence signals for
the hydrolysed samples (upon reaction with the tris(2,2ƍ-bipyridine)ruthenium(III)
reagent) were between 38% and 60% of that observed from the corresponding
non-hydrolysed samples (Figure 97). This is attributable to the conversion of heroin
into 6-monoacetylmorphine and morphine, which quench the emission from the
tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent [31, 40, 41]. A response from the
unknown compound initially detected using UV-absorbance was also observed in
each of the hydrolysed samples. This provides a clue with regard to elucidating its
molecular structure, as the tris(2,2‫މ‬-bipyridine)ruthenium(III) reagent is highly
selective for non-phenolic opiates [30, 31, 41, 156]. However, as previously
mentioned, further investigation is required to positively identify this species.
170
(i)
Chemiluminescence intensity (arbitrary units)
non-hydrolysed
hydrolysed
160
HER
140
120
100
80
HER
60
40
20
0
0
2
4
6
8
10
Time (min)
(iii)
Chemiluminescence intensity (arbitrary units)
non-hydrolysed
hydrolysed
210
HER
180
150
120
HER
90
60
30
0
0
2
4
6
8
10
Time (min)
275
(v)
250
225
200
175
150
125
100
75
50
25
0
0
non-hydrolysed
hydrolysed
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
CHAPTER SIX
HER
HER
2
4
6
8
10
Time (min)
110
(ii)
100
90
80
70
60
50
40
30
20
10
0
0
non-hydrolysed
hydrolysed
HER
HER
2
4
6
8
Time (min)
non-hydrolysed
hydrolysed
160
(iv)
140
HER
120
100
80
HER
60
40
20
0
0
2
4
6
8
10
Time (min)
110
(vi)
100
90
80
70
60
50
40
30
20
10
0
0
non-hydrolysed
hydrolysed
HER
HER
2
4
6
8
Time (min)
Figure 97. Chromatograms obtained for the analysis of six drug seizure samples
with stabilised tris(2,2ƍ-bipyridine)ruthenium(III) chemiluminescence detection.
Pink and black traces show sample preparation with and without hydrolysis. Peaks:
heroin (HER).
171
10
10
CHAPTER SIX
The heroin content of the six samples (Table 17) was found to be within the range
previously reported [248-254] and correlated well with the results obtained using
UV-absorbance detection. The difference between the two modes of detection ranged
between 1.45% and 10.42% across the six samples.
Table 17. Concentration of heroin in drug seizure samples obtained with stabilised
tris(2,2ƍ-bipyridine)ruthenium(III) chemiluminescence detection.
3.4
Conventional
Sample
Heroin (percent mass)
1
53
2
26
3
70
4
50
5
85
6
28
high
performance
liquid
chromatography
with
post-column enhanced permanganate chemiluminescence detection
As previously discussed, the preliminary partial reduction of manganese(VII) by
sodium thiosulfate creates a stable, relatively high concentration of manganese(III) in
permanganate reagents which provides two major advantages for chemiluminescence
detection; (i) improved temporal stability (a negligible change in peak area over
48 hours; < 1% based on the line of best fit) and (ii) dramatic increases in reaction
rates translating to enhanced limits of detection for various analytes [213]. These
features are highly compatible with the extended analysis periods utilised in many
chromatographic procedures and may provide superior analytical figures of merit for
the post-column detection of opiate alkaloids [213].
The enhanced permanganate reagent is highly selective for phenolic morphinan
alkaloids [213], and as such when a mixture of the six opiates (5 × 10-6 M) was
172
CHAPTER SIX
separated, only morphine and oripavine (retention times of 1.7 and 2.85 minutes
respectively) elicited a response (Figure 98). A small peak from an unknown
compound
was
observed
at
2.9
minutes.
This
was
believed
to
be
6-monoacetylmorphine, present in the opiate mix as a result of minor degradation of
the heroin standard. The identity of this peak was confirmed by analysing a
6-monoacetylmorphine standard (5 × 10-6 M) and obtaining an identical retention
time. A small degree of co-elution occurred between the oripavine and
6-monoacetylmorphine peaks. This was deemed to be insignificant and due to the
anticipated large increase in the 6-monoacetylmorphine peak following hydrolysis of
the seizure samples, the separation conditions were not altered.
2400
Chemiluminescence intensity (arbitrary units)
ORI
2000
1600
MOR
1200
800
6-MAM
400
0
0
2
4
6
8
10
Time (min)
Figure 98. HPLC separation of a mixture of P. somniferum opiate alkaloids
(5 × 10-6 M) with enhanced permanganate chemiluminescence detection. Peaks:
morphine (MOR), oripavine (ORI) and 6-monoacetylmorphine (6-MAM).
The use of the enhanced permanganate reagent provided two-and-a-half-orders of
magnitude greater sensitivity than UV-absorbance, with limits of detection for
173
CHAPTER SIX
morphine and oripavine of 7.5 × 10-10 M and 1 × 10-9 M respectively (Table 18).
These are comparable to or better than previously reported limits of detection for
these analytes with acidic potassium permanganate chemiluminescence [13, 75, 147,
148, 150, 153, 154, 243, 247].
Table 18. Analytical figures of merit obtained with enhanced permanganate
chemiluminescence detection.
Analyte
Calibration function
(peak area, arbitrary units)
R2
LOD
(M)a
R.S.D.
%*
Morphine
y= 8.54 × 108x + 15.827
0.9993
7.5 × 10-10
0.67
Oripavine
y= 1.13 × 109x + 2.657
0.9999
1.0 × 10-9
0.37
a
The lowest response detected with a signal to noise ratio of 3.
*n = 5; 1 × 10-6 M
3.4.1
Application to real samples
When the non-hydrolysed drug seizure samples were screened with the enhanced
permanganate reagent, 6-monoacetylmorphine and/or morphine were detected in
each (Figure 99). Samples 1, 2, 4 and 6 generated particularly intense
6-monoacetylmorphine peaks, which is consistent with results obtained using FIA
methodology. Following hydrolysis, a significant increase in chemiluminescence
intensity was observed for both morphine and 6-monoacetylmorphine across the six
seizure samples. Whilst the absolute chemiluminescence intensities varied,
non-hydrolysed samples 3 and 5 produced responses between 3.8% and 6% of their
corresponding hydrolysed samples. The remaining non-hydrolysed seizure samples
produced responses for 6-monacetylmorphine that were between 29% and 41% of
their corresponding hydrolysed samples. This may be attributable to the presence of
significant quantities of 6-monoacetylmorphine formed via natural degradation of the
samples following synthesis.
174
1200
non-hydrolysed
hydrolysed
(i)
6-MAM
1000
MOR
800
600
400
6-MAM
MOR
200
0
0
2
4
6
8
10
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
CHAPTER SIX
1000
6-MAM
MOR
800
600
6-MAM
400
200
MOR
0
0
2
4
(iii)
non-hydrolysed
hydrolysed
6-MAM
360
300
MOR
240
180
120
60
0
6-MAM
0
2
4
6
8
10
non-hydrolysed
hydrolysed
6-MAM
500
400
MOR
300
200
100
MOR 6-MAM
0
0
2
4
6
10
(iv)
600
MOR 6-MAM
540
480
420
360
300
240
180
6-MAM
120
MOR
60
0
0
2
4
6
8
10
Time (min)
8
10
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
(v)
8
non-hydrolysed
hydrolysed
660
Time (min)
600
6
Time (min)
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
Time (min)
420
non-hydrolysed
hydrolysed
(ii)
Time (min)
700 (vi)
non-hydrolysed
hydrolysed
6-MAM
600
500
MOR
400
300
200
6-MAM
100
MOR
0
0
2
4
6
8
Time (min)
Figure 99. Chromatograms obtained for the analysis of six drug seizure samples
with enhanced permanganate chemiluminescence detection. Pink and black traces
show sample preparation with and without hydrolysis. Peaks: morphine (MOR) and
6-monacetylmorphine (6-MAM).
175
10
CHAPTER SIX
3.5
High
performance
liquid
chromatography
with
multiplexed
UV-absorbance and chemiluminescence detection
As previously mentioned, the column was equipped with an annular frit that
enabled flow from the central and/or peripheral portions of the column bed to be
segmented through multiple detectors. Therefore each port of the annular frit was
attached to an individual detector or fed through to waste (Figure 100), enabling the
simultaneous detection of opiate alkaloids in suspected heroin seizure samples with
UV-absorbance and two chemiluminescence reagents.
To UV-absorbance detector
Waste
To chemiluminescence detector 1
(enhanced permanganate)
To chemiluminescence detector 2
(stabilised tris(2,2ƍ-bipyridine)ruthenium(III))
Figure 100. Illustration of the column outlet showing the three detectors attached to
the peripheral exit ports and the single central exit port leading to waste.
Using the separation conditions optimised for conventional post-column detection,
a mixture of the six opiates (5 × 10-6 M) was separated and simultaneously
determined by the three detectors. When the chromatograms from the two modes of
UV detection (conventional and multiplex) were compared, it was discovered that
the peaks were essentially coincident in time, but the multiplex approach appeared to
have poorer sensitivity (Figure 101). By equipping each port of the annular frit with a
detector (or line through to waste), the total flow from the column was segmented
between the central and peripheral ports, resulting in much lower flow rates passing
through the detectors than when using conventional post-column HPLC.
Consequently the flow rates through the three detectors were on average 3.4-fold
176
CHAPTER SIX
slower than when using the conventional post-column approach (0.73 mL/min
compared to 2.5 mL/min). A closer examination of these chromatograms revealed
that whilst the absolute intensities (peak heights) were considerably lower than that
of the conventional mode, the peaks were significantly broader (due to the longer
residence time in the UV-absorbance flow-cell), and consequently the peak areas
were comparable.
14
Conventional HPLC
Multiplex HPLC
Signal (arbitrary units)
12
ORI
10
THE HER
PAP
MOR COD
8
ORI
6
THE
4
HER
PAP
2
MOR
COD
0
0
1
2
3
4
5
6
7
8
9
10
Time (min)
Figure 101. HPLC separation of a mixture of P. somniferum opiate alkaloids
(2.5 × 10-6 M) with conventional post-column UV-absorbance detection (black
trace) and multiplexed UV-absorbance detection (pink trace). Peaks: morphine
(MOR), codeine (COD), oripavine (ORI), thebaine (THE), heroin (HER) and
papaverine (PAP).
Analytical figures of merit were obtained for the six opiate alkaloids in the three
multiplexed detectors (Tables 19, 20 and 21). In general, the limits of detection
obtained via the multiplexed approach were slightly inferior (less than an order of
magnitude) to those achieved with conventional post-column detection. This
reduction in sensitivity did not pose a problem for the rapid screening test, as the
concentration of heroin in suspected seizure samples is typically much higher than
177
CHAPTER SIX
the limits of detection obtained using the multiplexed approach. The precision of
repeated injections (% R.S.D.) was also somewhat poorer; however this is most
likely attributable to sampling different regions of the peripheral portion of solution
flow.
The ability to simultaneously run three modes of detection resulted in considerable
reductions in sample and solvent consumption, waste generation as well as total
analysis time (from 30 hours to 10 hours).
Table 19. Analytical figures of merit obtained with UV-absorbance detection.
Analyte
Calibration function
(peak area, arbitrary units)
R2
LOD
(M)a
R.S.D.
%*
Morphine
y= 6.66 × 105x – 0.102
0.9979
7.5 × 10-7
3.0
Codeine
y= 5.43 × 105x – 0.151
0.9998
7.5 × 10-7
3.3
Oripavine
y= 3.49 × 106x – 1.786
0.9984
5.0 × 10-7
3.9
Thebaine
y= 3.03 × 106x – 0.004
0.9992
2.5 × 10-7
2.8
Heroin
y= 2.30 × 106x + 2.074
0.9935
5.0 × 10-7
3.0
Papaverine
y= 2.39 × 106x – 0.187
0.9987
2.5 × 10-7
2.7
a
The lowest response detected with a signal to noise ratio of 3.
*n = 5; 1 × 10-6 M
Table 20. Analytical figures of merit obtained with stabilised ruthenium
chemiluminescence detection.
Analyte
Calibration function
(peak area, arbitrary units)
R2
LOD
(M)a
R.S.D.
%*
Codeine
y= 4.97 × 108x – 57.596
0.9989
1.0 × 10-8
1.3
Thebaine
y= 5.16 × 108x – 4.078
0.9996
1.0 × 10-8
1.1
Heroin
y= 9.49 × 107x + 79.766
0.9882
5.0 × 10-8
1.2
a
The lowest response detected with a signal to noise ratio of 3.
*n = 5; 1 × 10-6 M
178
CHAPTER SIX
Table 21. Analytical figures of merit obtained with enhanced permanganate
chemiluminescence detection.
Analyte
Calibration function
(peak area, arbitrary units)
R2
LOD
(M)a
R.S.D.
%*
Morphine
y= 8.54 × 108x + 15.827
0.9993
1.0 × 10-9
1.6
Oripavine
y= 1.13 × 109x + 2.657
0.9999
1.0 × 10-9
1.6
a
The lowest signal detected with a signal to noise ratio of 3.
*n = 5; 1 × 10-6 M
3.5.1
Application to real samples
The six drug seizure samples were then screened using the multiplexed approach.
The chromatograms from the three simultaneous modes of detection were almost
identical to those obtained using conventional post-column detection, with only
minor differences in sensitivity observed (Figures 102, 103, and 104).
As can be seen in Table 22, application of multiplexed UV-absorbance and
chemiluminescence detection generated results that compared favourably. The
difference between the two modes of detection ranged between 1.42% and 8.0%.
These results are in good agreement with those obtained using conventional
post-column detection and ranges previously reported by other researchers [248254].
Table 22. Concentration of heroin in drug seizure samples obtained with
multiplexed UV-absorbance and stabilised tris(2,2ƍ-bipyridine)ruthenium(III)
chemiluminescence detection.
Heroin (percent mass)
Sample
UV-absorbance
Chemiluminescence
1
48
50
2
25
27
3
71
70
4
46
49
5
87
83
6
25
27
179
CHAPTER SIX
1.6
3.2
MOR
1.4
1.2
1.0
0.8
0.6
HER
0.4
0.2
0.0
2.4
MOR
2.0
1.6
1.2
0.8
MOR
0.4
0
2
non-hydrolysed
hydrolysed
(ii)
2.8
Signal (arbitrary units)
Signal (arbitrary units)
non-hydrolysed
hydrolysed
(i)
4
6
8
0.0
10
0
2
Time (min)
1.0
0.9
0.7
0.6
0.5
HER
0.4
0.3
0.2
0
2
4
6
8
1.0
0.8
0.6
HER
0.4
0.0
10
MOR
0
2
4.5
1.6
Signal (arbitrary units)
Signal (arbitrary units)
6
8
10
non-hydrolysed
hydrolysed
(vi)
1.4
4.0
3.5
MOR
3.0
2.5
2.0
1.5
1.0
HER
0.5
0.0
4
Time (min)
non-hydrolysed
hydrolysed
(v)
10
non-hydrolysed
hydrolysed
Time (min)
5.0
8
MOR
0.2
0.1
0.0
6
(iv)
1.2
6-MAM
Signal (arbitrary units)
Signal (arbitrary units)
1.4
MOR
0.8
4
Time (min)
non-hydrolysed
hydrolysed
(iii)
HER
0
2
4
1.2
MOR
1.0
0.8
0.6
0.4
HER
0.2
6
8
0.0
10
Time (min)
0
2
4
6
8
Time (min)
Figure 102. Chromatograms obtained for the analysis of six drug seizure samples
with multiplexed UV-absorbance detection. Pink and black traces show sample
preparation with and without hydrolysis. Peaks: morphine (MOR),
6-monoacetylmorphine (6-MAM) and heroin (HER).
180
10
120
non-hydrolysed
hydrolysed
(i)
HER
100
80
60
HER
40
20
0
0
2
4
6
8
10
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
CHAPTER SIX
120
non-hydrolysed
hydrolysed
(ii)
HER
100
80
60
HER
40
20
0
0
2
150
non-hydrolysed
hydrolysed
(iii)
HER
125
100
75
HER
50
25
0
0
2
4
6
8
10
110
HER
150
125
100
HER
75
50
25
0
0
2
4
6
8
10
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
175
10
100
HER
90
80
70
60
HER
50
40
30
20
10
0
0
2
4
6
8
10
Time (min)
non-hydrolysed
hydrolysed
(v)
8
non-hydrolysed
hydrolysed
(iv)
Time (min)
200
6
Time (min)
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
Time (min)
4
Time (min)
100
non-hydrolysed
hydrolysed
(vi)
HER
80
60
40
HER
20
0
0
2
4
6
8
Time (min)
Figure 103. Chromatograms obtained for the analysis of six drug seizure samples
with multiplexed stabilised tris(2,2ƍ-bipyridine)ruthenium(III) chemiluminescence
detection. Pink and black traces show sample preparation with and without
hydrolysis. Peaks: heroin (HER).
181
10
420 (i)
MOR
360
non-hydrolysed
hydrolysed
6-MAM
300
240
180
120
6-MAM
60
0
MOR
0
2
4
6
8
10
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
CHAPTER SIX
350
non-hydrolysed
hydrolysed
(ii)
MOR
300
250
200
6-MAM
150
6-MAM
100
MOR
50
0
0
2
4
150 (iii)
non-hydrolysed
hydrolysed
MOR
6-MAM
120
90
60
30
0
MOR 6-MAM
0
2
4
6
8
10
180
(iv)
160
140
120
100
80
60
6-MAM
40
MOR
20
0
0
2
4
80
70
60
50
40
30
20
6-MAM
0
0
2
4
6
8
10
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
90
10
6
8
10
Time (min)
non-hydrolysed
hydrolysed
(v) MOR
100
6-MAM
10
non-hydrolysed
hydrolysed
MOR
6-MAM
Time (min)
110
8
Time (min)
Chemiluminescence intensity (arbitrary units)
Chemiluminescence intensity (arbitrary units)
Time (min)
6
Time (min)
125
(vi)
non-hydrolysed
hydrolysed
MOR
6-MAM
100
75
50
6-MAM
25
0
MOR
0
2
4
6
8
Time (min)
Figure 104. Chromatograms obtained for the analysis of six drug seizure samples
with enhanced permanganate chemiluminescence detection. Pink and black traces
show sample preparation with and without hydrolysis. Peaks: morphine (MOR) and
6-monoacetylmorphine (6-MAM).
182
10
CHAPTER SIX
4. Conclusions
The rapid separation of P. somniferum opiate alkaloids (morphine, oripavine,
codeine, thebaine, heroin and papaverine) was achieved in six minutes using
conventional HPLC with post-column UV-absorbance or chemiluminescence
detection. All of the alkaloids were detectable with UV-absorbance, whilst greater
selectivity was imparted via the use of the enhanced permanganate reagent
(morphine and oripavine) and the stabilised tris(2,2ƍ-bipyridine)ruthenium(III)
reagent (codeine, thebaine and heroin).
The combination of these two stabilised chemiluminescence reagents not only
provides the basis of a rapid screening test for heroin, but proved highly compatible
with the extended analysis times inherent to HPLC procedures and generated limits
of detection comparable to or superior than those previously reported for the
P. somniferum opiate alkaloids.
The addition of an annular frit to the end of the chromatographic column which
enabled segmentation of the column eluate between three unique modes of detection
(UV-absorbance and two chemiluminescence reagents) resulted in minor losses of
sensitivity (less than an order of magnitude) whilst significantly reducing the sample
and solvent consumption, waste generation and total analysis time. Furthermore,
application of the multiplexed UV-absorbance and chemiluminescence detectors for
the post-column determination of opiate alkaloids in drug seizure samples, resulted
in good agreement with values obtained using conventional post-column detection.
183
REFERENCES
REFERENCES
1.
N.W. Barnett and P.S. Francis,
Chemiluminescence: Overview
in
Encyclopedia of Analytical Science, Editors, P.J. Worsfold, A. Townshend,
and C.F. Poole, 2005, Elsevier, Oxford, 506-510.
2.
A.C. Calokerinos and L.P. Palilis, Chemiluminescence in Flow Injection
Analysis in Chemiluminescence in Analytical Chemistry, Editors, A.M.
García-Campaña and W.R.G. Baeyens, 2001, Marcel Dekker, New York,
321-348.
3.
A.M. García-Campaña, W.R.G. Baeyens and X. Zhang, Chemiluminescence
in Analytical Chemistry in Chemiluminescence in Analytical Chemistry,
Editors, A.M. García-Campaña and W.R.G. Baeyens, 2001, Marcel Dekker,
New York, 41-67.
4.
A. Roda, M. Guardigli, E. Michelini, M. Mirasoli and P. Pasini, Anal. Chem.,
2003, 75 (21), 462A-470A.
5.
L. L.Gamiz-Gracia, A.M. García-Campaña, J.F. Huertas-Perez and F.J. Lara,
Anal. Chim. Acta, 2009, 640 (1-2), 7-28.
6.
J.L. Adcock, J.M. Terry, C.J. Barrow, N.W. Barnett, D.C. Olson and
P.S. Francis, Drug Test. Analysis, 2011, 3 (3), 139-144.
7.
A. Roda, M. Mirasoli, B. Roda and P. Reschiglian, Flow-assisted analysis in
Chemiluminescence and Bioluminescence : Past, Present and Future, Editor
A. Roda, 2011, Royal Society of Chemistry, Cambridge, 191-227.
8.
P. Fletcher, K.N. Andrew, A.C. Calokerinos, S. Forbes and P.J. Worsfold,
Luminescence, 2001, 16 (1), 1-23.
9.
A.M. García-Campaña and F.J. Lara, Anal. Bioanal. Chem., 2007, 387 (1),
165-169.
10.
E.N. Harvey, J. Biol. Chem., 1917, 31 (2), 311-336.
11.
J. J. Stauff and W. Jaeschke, Atmos. Environ., 1975, 9 (11), 1038-1039.
12.
B.J. Hindson and N.W. Barnett, Anal. Chim. Acta, 2001, 445 (1), 1-19.
13.
J.L. Adcock, P.S. Francis and N.W. Barnett, Anal. Chim. Acta, 2007, 601 (1),
36-67.
14.
C.M. Hindson, P.S. Francis, G.R. Hanson, J.L. Adcock and N.W. Barnett,
Anal. Chem., 2010, 82, 4174-4180.
184
REFERENCES
15.
C. Thongpoon, B. Liawruangrath, S. Liawruangrath, R.A. Wheatley and
A. Townshend, J. Pharm. Biomed. Anal., 2006, 42 (2), 277-282.
16.
N.W. Barnett, B.J. Hindson, P. Jones and T.A. Smith, Anal. Chim. Acta,
2002, 451 (2), 181-188.
17.
J.L. Adcock, P.S. Francis, T.A. Smith and N.W. Barnett, Analyst, 2008, 133
(1), 49-51.
18.
G.P. McDermott, P.S. Francis, K.J. Holt, K.L. Scott, S.D. Martin, N. Stupka,
N.W. Barnett and X.A. Conlan, Analyst, 2011, 136 (12), 2578-2585.
19.
X.A. Conlan, N. Stupka, G.P. McDermott, N.W. Barnett and P.S. Francis,
Anal. Methods, 2010, 2 (2), 171-173.
20.
P.S. Francis, J.W. Costin, X.A. Conlan, S.A. Bellomarino, J.A. Barnett and
N.W. Barnett, Food Chem., 2010, 122 (3), 926-929.
21.
T. Slezak, P.S. Francis, N. Anastos and N.W. Barnett, Anal. Chim. Acta,
2007, 593 (1), 98-102.
22.
D.W. Percy, J.L. Adcock, X.A. Conlan, N.W. Barnett, M.E. Gange,
L.K. Noonan, L.C. Henderson and P.S. Francis, Talanta, 2010, 80 (5), 21912195.
23.
J.W. Costin, P.S. Francis and S.W. Lewis, Anal. Chim. Acta, 2003, 480 (1),
67-77.
24.
J.L. Adcock, N.W. Barnett, J.W. Costin, P.S. Francis and S.W. Lewis,
Talanta, 2005, 67 (3), 585-589.
25.
P.S. Francis, J.L. Adcock, J.W. Costin, S.D. Purcell, F.M. Pfeffer and
N.W. Barnett, J. Pharm. Biomed. Anal., 2008, 48 (3), 508-518.
26.
J.L. Adcock, C.J. Barrow, N.W. Barnett, X.A. Conlan, C.F. Hogan and
P.S. Francis, Drug Test. Analysis, 2011, 3 (3), 145-160.
27.
L.A. Hill, C.E. Lenehan, P.S. Francis, J.L. Adcock, M.E. Gange, F.M. Pfeffer
and N.W. Barnett, Talanta, 2008, 76 (3), 674-679.
28.
F.H. Burstall, J. Chem. Soc. A, 1936, 1 (1), 173-175.
29.
D.M. Hercules and F.E. Lytle, J. Am. Chem. Soc., 1966, 88 (20), 4745-4746.
30.
R.D. Gerardi, N.W. Barnett and S.W. Lewis, Anal. Chim. Acta, 1999, 378
(1-3), 1-41.
31.
B.A. Gorman, P.S. Francis and N.W. Barnett, Analyst, 2006, 131 (5),
616-639.
185
REFERENCES
32.
R.D. Gerardi, N.W. Barnett and P. Jones, Anal. Chim. Acta, 1999, 388 (1-2),
1-10.
33.
G.P. McDermott, P. Jones, N.W. Barnett, D.N. Donaldon and P.S. Francis,
Anal. Chem., 2011, 83 (13), 5453-5457.
34.
M.M. Richter, Chem. Rev., 2004, 104 (6), 3003-3036.
35.
N.W. Barnett, R. Bos, H. Brand, P. Jones, K.F. Lim, S.D. Purcell and
R.A. Russell, Analyst, 2002, 127 (4), 455-458.
36.
N.W. Barnett, B.J. Hindson, S.W. Lewis, S.D. Purcell and P. Jones,
Anal. Chim. Acta, 2000, 421 (1), 1-6.
37.
J.W. Costin, N.W. Barnett and S.W. Lewis, Talanta, 2004, 64 (4), 894-898.
38.
J.B. Noffsinger and N.D. Danielson, Anal. Chem., 1987, 59, 865-868.
39.
L. He, K.A. Cox and N.D. Danielson, Anal. Lett., 1990, 23 (2), 195-210.
40.
J. McCall and M.M. Richter, Analyst, 2000, 125 (3), 545-548.
41.
N.W. Barnett, R.D. Gerardi, D.L. Hampson and R.A. Russell, Anal.
Commun., 1996, 33 (8), 255-260.
42.
T.J. Pastor and F.T. Pastor, Talanta, 2000, 52 (6), 959-970.
43.
A.J. Brown, P.S. Francis, J.L. Adcock, K.F. Lim and N.W. Barnett, Anal.
Chim. Acta, 2008, 624 (2), 175-183.
44.
M. Jáky, L.I. Simándi and V.Y. Shafirovich, Inorg. Chim. Acta, 1984, 90 (3),
39-41.
45.
M. Jáky and M. Zrinyi, Polyhedron, 1993, 12 (11), 1271-1275.
46.
A.J. Brown, C.E. Lenehan, P.S. Francis, D.E. Dunstan and N.W. Barnett,
Talanta, 2007, 71 (5), 1951-1957.
47.
B.J. Hindson, Ph.D Thesis, 2000, Deakin University, Waurn Ponds.
48.
N.W. Barnett, B.J. Hindson, S.W. Lewis, P. Jones and P.J. Worsfold, Analyst,
2001, 126 (10), 1636-1639.
49.
J. Ruzicka and E.H. Hansen, Anal. Chim. Acta, 1975, 78 (1), 145-157.
50.
K.K. Stewart, G.R. Beecher and P.E. Hare, Anal. Biochem., 1976, 70 (1),
167-173.
51.
J. Ruzicka and E.H. Hansen, Flow Injection Analysis, Second edition, 1988,
New York, John Wiley & Sons.
52.
N.W.
Barnett
and
chemiluminescence in
P.S.
Francis,
Chemiluminescence:
Liquid-phase
Encyclopedia of Analytical Science, Editors,
186
REFERENCES
P.J. Worsfold, A. Townshend, and C.F. Poole, 2005, Elsevier, Oxford,
511-520.
53.
P.S. Francis and C.F. Hogan, Luminescence in Advances in Flow Injection
Analysis and Related Techniques, Comprehensive Analytical Chemistry
Series, Editors, I.D. McKelvie and S.D. Kolev, 2008, Elsevier, Oxford,
343-373.
54.
G. Rule and W.R. Seitz, Clin. Chem., 1979, 25 (9), 1635-1638.
55.
J. Ruzicka and G.D. Marshall, Anal. Chim. Acta, 1990, 237 (1), 329-343.
56.
C.E. Lenehan, Seqential Injection Analysis in Encyclopedia of Analytical
Science, Editors, P.J. Worsfold, A. Townshend, and C.F. Poole, 2005,
Elsevier, Oxford, 290-297.
57.
C.E. Lenehan, N.W. Barnett and S.W. Lewis, Analyst, 2002, 127 (8),
997-1020.
58.
N.W. Barnett, C.E. Lenehan and S.W. Lewis, Trend. Anal. Chem., 1999,
18 (5), 346-353.
59.
A.M. Pimenta, M.C.B.S.M. Montenegro, A.N. Arujo and J.M. Calatayud,
J. Pharm. Biomed. Anal., 2006, 40 (1), 16-34.
60.
R.A. Shalliker and S. Kayillo, Liquid Chromatography: Reversed Phase in
Encyclopedia of Analytical Science, Editors, P.J. Worsfold, A. Townshend,
and C.F. Poole, 2005, Elsevier, Oxford, 152-157.
61.
L.R. Snyder, J.J. Kirkland and J.W. Dolan, Introduction to modern liquid
chromatography, 2010, Hoboken, Wiley.
62.
L.R. Snyder, J.J. Kirkland and J.L. Glajch, Practical HPLC method
development, 1997, New York, Wiley.
63.
M.D. Luque de Castro, Flow Injection Analysis: Principles in Encyclopedia
of Analytical Science, Editors, P.J. Worsfold, A. Townshend, and C.F. Poole,
2005, Elsevier, Oxford, 31-39.
64.
M. Miró and W. Frenzel, Flow Injection Analysis: Detection Techniques in
Encyclopedia of Analytical Science, Editors, P. Worsfold, A. Townshend, and
C. Poole, 2005, Elsevier, Oxford, 48-56.
65.
J. Ruzicka and E.H. Hansen, Anal. Chim. Acta, 1979, 106 (2), 207-224.
66.
G.D. Christian and J. Ruzicka, Anal. Chim. Acta, 1992, 261 (1-2), 11-21.
187
REFERENCES
67.
M. Silva, Kinetic Methods: Noncatalytic Techniques in Encyclopedia of
Analytical Science, Editors, P.J. Worsfold, A. Townshend, and C.F. Poole,
2005, Elsevier, Oxford, 9-24.
68.
D. Pérez-Bendito and M. Silva, Kinetics in Chemiluminescence Analysis in
Chemiluminescence in Analytical Chemistry, Editors, A.M. García-Campaña
and W.R.G. Baeyens, 2001, Marcel Dekker, New York, 175-209.
69.
H.A. Mottola and D. Pérez-Bendito, Anal. Chem., 1992, 64 (12), 407R-428R.
70.
M. Miró and W. Frenzel, Flow Injection Analysis: Detection Techniques in
Encyclopedia of Analytical Science, Editors, P.J. Worsfold, A. Townshend,
and C.F. Poole, 2005, Elsevier, Oxford, 48-56.
71.
A. Townshend, Analyst, 1990, 115 (5), 495-500.
72.
W.R. Seitz and K.V. Dyke, Crit. Rev. Anal. Chem., 1981, 13 (1), 1-58.
73.
J.M. Terry, J.L. Adcock, D.C. Olsen, D.K. Wolcott, C. Schwanger, L.A. Hill,
N.W. Barnett and P.S. Francis, Anal. Chem., 2008, 80 (24), 9817-9821.
74.
J.L. Burguera, A. Townshend and S. Greenfield, Anal. Chim. Acta, 1980,
114, 209-214.
75.
R.W. Abbott, A. Townshend and R. Gill, Analyst, 1986, 111 (6), 635-640.
76.
R.L. Veazey and T.A. Nieman, J. Chromatogr. A, 1980, 200 (1), 153-162.
77.
J.L. Pérez Pavón, E.R. Gonzalo, G.D. Christian and J. Ruzicka, Anal. Chem.,
1992, 64 (8), 923-929.
78.
F. Preuschoff, U. Spohn, G. Blankenstein, K.-H. Mohr and M.-R. Kula,
Fresenius J. Anal. Chem., 1993, 346 (10-11), 924-929.
79.
D.J. Tucker, D. Toivola, C.H. Pollema, J. Ruzicka and G.D. Christian,
Analyst, 1994, 119 (5), 975-979.
80.
P.B. Martelli, B.F. Reis, A.N. Araújo, M. Conceição and B.S.M. Montenegro,
Talanta, 2001, 54 (5), 879-885.
81.
P. Campíns-Falcó, L.A. Tortajada-Genaro and F. Bosch-Reig, Talanta, 2001,
55 (2), 403-413.
82.
E.J. Llorent-Martínez, P. Ortega-Barrales and A. Molina-Díaz, Anal. Chim.
Acta, 2006, 580 (2), 149-154.
83.
K.M. Scudder, C.H. Pollema and J. Ruzicka, Anal. Chem., 1992, 64 (21),
2657-2660.
84.
T. Toyo'oka, J. Chromatogr. B, 2009, 877 (28), 3318-3330.
188
REFERENCES
85.
A. Pastore, G. Federici, E. Bertini and F. Piemonte, Clin. Chim. Acta, 2003,
333 (1), 19-39.
86.
A. Meister and M.E. Anderson, Ann. Rev. Biochem., 1983, 52 (1), 711-760.
87.
P. Monostori, G. Wittmann, E. Karg and S. Túri, J. Chromatogr. B, 2009,
877 (28), 3331-3346.
88.
D. Giustarini, I. Dalle-Donne, D. Tsikas and R. Rossi, Crit. Rev. Clin. Lab.
Sci., 2009, 46 (5-6), 241-281.
89.
E. Mosharov, M.R. Cranford and R. Banerjee, Biochemistry, 2000, 39 (42),
13005-13011.
90.
P.M. DiBello, S. Dayal, S. Kaveti, D. Zhang, M. Kinter, S.R. Lentz and
D.W. Jacobsen, Mol. Cell. Proteomics, 2010, 9 (3), 471-485.
91.
J.D. Finkelstein and S.H. Mudd, J. Biol. Chem., 1967, 242 (5), 873-880.
92.
S.H. Mudd, J.D. Finkelstein, F. Irreverre and L. Laster, J. Biol. Chem., 1965,
240 (11), 4382-4392.
93.
G. Chwatko and H. Jakubowski, Clin. Chem., 2005, 51 (2), 408-415.
94.
K. Rasmussen and J. Moeller, Ann. Clin. Biochem., 2000, 37 (5), 627-648.
95.
P.M. Ueland, A. Mansoor, A.B. Guttormsen, F. Muller, P. Aukrust,
H. Refsum and A.M. Svardal, The J. Nutr., 1996, 126 (4), 1281S-1284S.
96.
A. Meister, J. Biol. Chem., 1988, 263 (33), 17205-17208.
97.
J. Estrela, A. Ortega and E. Obrador, Crit. Rev. Clin. Lab. Sci., 2006, 43 (2),
143-181.
98.
C. Perricone, C. De Carolis and R. Perricone, Autoimmun. Rev., 2009, 8 (8),
697-701.
99.
D.M. Townsend, K.D. Tew and H. Tapiero, Biomed. Pharmacother., 2003,
57 (3-4), 145-155.
100.
M. Valko, D. Leibfritz, J. Moncol, M.T.D. Cronin, M. Mazur and J. Telser,
Int. J. Biochem. Cell Biol., 2007, 39 (1), 44-84.
101.
M. Valko, C.J. Rhodes, J. Moncol, M. Izakovic and M. Mazur, Chem-Biol.
Interact., 2006, 160 (1), 1-40.
102.
B. Halliwell and J.M.C. Gutteridge, Free radicals in biology and medicine,
1989, Oxford, Clarendon Press.
103.
B.P. Yu, Physiol. Rev., 1994, 74 (1), 139-162.
104.
E.H. Sarsour, M.G. Kumar, L. Chaudhuri, A.L. Kalen and P.C. Goswami,
Antioxid. Redox Signal., 2009, 11 (12), 2985-3011.
189
REFERENCES
105.
K. Apel and H. Hirt, Annu. Rev. Plant Biol., 2004, 55, 373-399.
106.
E. Camera and M. Picardo, J. Chromatogr. B, 2002, 781 (1-2), 181-206.
107.
F. Carlucci and A. Tabucchi, J. Chromatogr. B, 2009, 877 (28), 3347-3357.
108.
X. Chen, Y. Zhou, X. Peng and J. Yoon, Chem. Soc. Rev., 2010, 39 (6),
2120-2135.
109.
R.E. Hansen and J.R. Winther, Anal. Biochem., 2009, 394 (2), 147-158.
110.
Y. Iwasaki, Y. Saito, Y. Nakano, K. Mochizuki, O. Sakata, R. Ito, K. Saito
and H. Nakazawa, J. Chromatogr. B, 2009, 877 (28), 3309-3317.
111.
F.G. Hopkins, Biochem. J., 1921, 15 (2), 286-305.
112.
M. Kemp, Y.-M. Go and D.P. Jones, Free Radical Biol. Med., 2008, 44 (6),
921-937.
113.
N. Rouhier, S.D. Lemaire and J.-P. Jacquot, Ann. Rev. Plant Biol., 2008,
59 (1), 143-166.
114.
R. Rossi, I. Dalle-Donne, A. Milzani and D. Giustarini, Clin. Chem., 2006,
52 (7), 1406-1414.
115.
R. Rossi, A. Milzani, I. Dalle-Donne, D. Giustarini, L. Lusini, R. Colombo
and P. Di Simplicio, Clin. Chem., 2002, 48 (5), 742-753.
116.
J. Lock and J. Davis, TrAC, Trends Anal. Chem., 2002, 21 (12), 807-815.
117.
O. Nekrassova, N.S. Lawrence and R.G. Compton, Talanta, 2003, 60 (6),
1085-1095.
118.
A. Khan, M.I. Khan, Z. Iqbal, Y. Shah, L. Ahmad, S. Nazir, D.G. Watson,
J.A. Khan, F. Nasir, A. Khan and Ismail, Talanta, 2011, 84 (3), 789-801.
119.
S. Pelletier and C.A. Lucy, Analyst, 2004, 129 (8), 710-713.
120.
D. Stempak, S. Dallas, J. Klein, R. Bendayan, G. Koren and S. Baruchel,
Ther. Drug Monit., 2001, 23 (5), 542.
121.
V. Ducros, K. Demuth, M.-P. Sauvant, M. Quillard, E. Caussé, M. Candito,
M.-H. Read, J. Drai, I. Garcia and M.-F. Gerhardt, J. Chromatogr. B, 2002,
781 (1-2), 207-226.
122.
N. Kuroda, M. Kai and K. Nakashima, Chemiluminescence Detection in
Liquid Chromatography in Chemiluminescence in Analytical Chemistry,
Editors, A.M. García-Campaña and W.R.G. Baeyens, 2001, Marcel Dekker,
New York, 393-425.
123.
H.-N. Li, Y.-X. Ci and L. Huang, Anal. Sci., 1997, 13 (5), 821-824.
190
REFERENCES
124.
S. Zhao, Y. Huang, F. Ye, M. Shi and Y.-M. Liu, J. Chromatogr. A, 2010,
1217 (36), 5732-5736.
125.
N.N. Daeid, Illicit Drugs in Encyclopedia of Analytical Science, Editors,
P.J. Worsfold, A. Townshend, and C.F. Poole, 2005, Elsevier, Oxford,
446-453.
126.
L.D. Kapoor, Opium Poppy: Botany, Chemistry and Pharmacology, 1995,
New York, Haworth Press.
127.
O.H. Drummer and M. Odell, The Forensic Pharmacology of Drugs of
Abuse, 2001, London, Arnold.
128.
United Nations Illicit Drug Control Program, Recommended Methods for
Testing Opium, Morphine and Heroin, 1998, New York, United Nations
Office on Drugs and Crime
129.
United Nations Illicit Drug Control Program, Rapid Testing of Drugs of
Abuse, 1994, United Nations Office on Drugs and Crime
130.
M.L. Merves and B.A. Goldberger, Heroin in Encyclopedia of Analytical
Science, Editors, P.J. Worsfold, A. Townshend, and C.F. Poole, 2005,
Elsevier, Oxford, 260-266.
131.
S. Bullock, E. Manias and A. Galbraith, Fundementals of Pharmacology,
Fifth edition, 2007, Sydney, Peason-Prentice Hall.
132.
F.H. Serturner, J. der Pharm., 1805, 13, 229-234.
133.
T. Carnwath and I. Smith, Heroin Century, 2002, New York, Routledge.
134.
C.R.A. Wright, J. Chem. Soc., 1874, 27 (1), 1043-1044.
135.
W. Sneader, Lancet, 1998, 352 (21), 1697-1699.
136.
H. Dreser, Deut. Med. Wochenschr., 1898, 24 (1), 185.
137.
M.R. Hutchinson and A.A. Somogyi, Eur. J. Pharmacol., 2002, 453 (1),
27-32.
138.
J. Temsamani, J.M. Scherrmann, A.R. Rees and M. Kaczorek, Pharm. Sci.
Technol. To., 2000, 3 (5), 155-162.
139.
International Narcotics Control Board, Narcotic Drugs 2011 Report:
Estimated World Requirements for 2012—Statistics for 2010, 2010, New
York, United Nations
140.
S.W. Lewis, K.M. Agg and N.B. Anastos, N.W., Can. J. Pol. Sec. Serv.,
2005, 3 (1), 17-25.
191
REFERENCES
141.
M.E. Bosch, A.R. Sánchez, F.S. Rojas and C.B. Ojeda, J. Pharm. Biomed.
Anal., 2007, 43 (3), 799–815.
142.
J. Garrido, C. Delerue-Matos, F. Borges, T. Macedo and A.M. Oliveira-Brett,
Anal. Lett., 2004, 37 (5), 831-844.
143.
F. Tagliaro, D. Franchi, R. Dorizzi and M. Marigo, J. Chromatogr. B, 1989,
488 (1), 215-228.
144.
F. Tagliaro, S. Turrina, P. Pisi, F.P. Smith and M. Marigo, J. Chromatogr. B:
Biomed. Sci. App., 1998, 713 (1), 27-49.
145.
B. Levine, Presumptive Chemical Tests in Encyclopedia of Forensic
Sciences, Editors, J. Siegel, P. Saukko, and G. Knupfer, 2000, Academic
Press, San Diego.
146.
C.L. O'Neal, D.J. Crouch and A.A. Fatah, Forensic Sci. Int., 2000, 109 (3),
189-201.
147.
R.W. Abbott and A. Townshend, Anal. Proc., 1986, 23 (1), 25-26.
148.
N.W. Barnett, D.G. Rolfe, T.A. Bowser and T.W. Paton, Anal. Chim. Acta,
1993, 282, 551-557.
149.
N. W. Barnett, C. E. Lenehan, S. W. Lewis, D. J. Tucker and K. M. Essery,
Analyst, 1998, 123 (4), 601-605.
150.
N.W. Barnett, B.J. Hindson and S.W. Lewis, Analyst, 2000, 125 (1), 91-95.
151.
N.W. Barnett, S.W. Lewis and D.J. Tucker, Fresenius J. Anal. Chem., 1996,
355 (5-6), 591-595.
152.
J.L. Adcock, P.S. Francis, K.M. Agg, G.D. Marshall and N.W. Barnett, Anal.
Chim. Acta, 2007, 600 (1-2), 136-141.
153.
J.W. Costin, S.W. Lewis, S.D. Purcell, L.R. Waddell, P.S. Francis and
N.W. Barnett, Anal. Chim. Acta, 2007, 597 (1), 19-23.
154.
C.E. Lenehan, N.W. Barnett, S.W. Lewis and K.M. Essery, Aust. J. Chem,
2004, 57 (10), 1001-1004.
155.
G.M. Greenway, A.W. Knight and P.J. Knight, Analyst, 1995, 120 (10),
2549-2552.
156.
N.W. Barnett, T.A. Bowser, R.D. Gerardi and B. Smith, Anal. Chim. Acta,
1996, 318 (3), 309-317.
157.
N.W. Barnett, B.J. Hindson, S.W. Lewis and S.D. Purcell, Analyst, 1998,
35 (10), 321–324.
192
REFERENCES
158.
K.M. Agg, A.F. Craddock, R. Bos, P.S. Francis, S.W. Lewis and
N.W. Barnett, J. Forensic Sci., 2006, 51 (5), 1080-1084.
159.
K. Mervartová, M. Polášek and J. Martínez Calatayud, J. Pharm. Biomed.
Anal., 2007, 45 (3), 367-381.
160.
S. Mohr, J.M. Terry, J.L. Adcock, P.R. Fielden, N.J. Goddard, N.W. Barnett,
D.K. Wolcott and P.S. Francis, Analyst, 2009, 134 (11), 2233-2238.
161.
H. Nakamura, Y. Murakami, K. Yokoyama, E. Tamiya, I. Karube, M. Suda
and S. Uchiyama, Anal. Chem., 2001, 73 (2), 373-378.
162.
Z.-R. Xu and Z.-L. Fang, Anal. Chim. Acta, 2004, 507 (1), 129-135.
163.
É. Tyrrell, C. Gibson, B.D. MacCraith, D. Gray, P. Byrne, N. Kent, C. Burke
and B. Paull, Lab Chip, 2004, 4 (4), 384-390.
164.
Z. Zhang, D. He, W. Liu and Y. Lv, Luminescence, 2005, 20 (4-5), 377-381.
165.
G.M. Greenway, L.J. Nelstrop and S.N. Port, Anal. Chim. Acta, 2000, 405
(1-2), 43-50.
166.
X. Wang, O. Hofmann, R. Das, E.M. Barrett, A.J. de Mello, J.C. de Mello
and D.D.C. Bradley, Lab Chip, 2007, 7 (1), 58-63.
167.
A. Economou, A.K. Clark and P.R. Fielden, Anal. Commun., 1998, 35 (12),
389-390.
168.
S.W. Lewis, P.S. Francis, K.F. Lim and G.E. Jenkins, Anal. Chim. Acta,
2002, 461 (1), 131-139.
169.
L.M. Magalhães, M.A. Segundo, S. Reis, J.L.F.C. Lima, J.M. Estela and
V. Cerdà, Anal. Chem., 2007, 79 (10), 3933-3939.
170.
Y.-X. Guan, Z.-R. Xu, J. Dai and Z.-L. Fang, Talanta, 2006, 68 (4),
1384-1389.
171.
L. Marle and G.M. Greenway, Anal. Chim. Acta, 2005, 548 (1-2), 20-25.
172.
W. Liu, Z. Zhang and Z. Liu, Anal. Chim. Acta, 2007, 592 (2), 187-192.
173.
R. Carlson, S.W. Lewis and K.F. Lim, Chemeda: Aust. J. Chem. Ed., 1999,
51-53, 14-21.
174.
S.W. Lewis, P.S. Francis, K.F. Lim, G.E. Jenkins and X.D. Wang, Analyst,
2000, 125 (10), 1869-1874.
175.
P.S. Francis and J.L. Adcock, Chemiluminescence Detection in Liquid
Chromatography in Hyphenated and Alternative Methods of Detection in
Chromatography, Editor R.A. Shalliker, 2010, CRC Press, Boca Raton, In
press.
193
REFERENCES
176.
P.S. Ellis, B.S. Gentle, M.R. Grace and I.D. McKelvie, Talanta, 2009, 79 (3),
830-835.
177.
N.W. Barnett, P.S. Francis, S.W. Lewis and K.F. Lim, Anal. Commun., 1999,
36 (4), 131-134.
178.
A.T. Faizullah and A. Townshend, Anal. Proc., 1985, 22 (1), 15-16.
179.
T. Pérez-Ruiz, C. Martínez-Lozano and M.D. García, J. Chromatogr. A,
2007, 1169 (1-2), 151-157.
180.
L.L. Shultz and T.A. Nieman, J. Biolumin. Chemilumin., 1998, 13 (2), 85-90.
181.
M.M. Cooke, E.H. Doeven, C.F. Hogan, J.L. Adcock, G.P. McDermott,
X.A. Conlan, N.W. Barnett, F.M. Pfeffer and P.S. Francis, Anal. Chim. Acta,
2009, 635 (1), 94-101.
182.
X. Hu, N. Takenaka, M. Kitano, H. Bandow, Y. Maeda and M. Hattori,
Analyst, 1994, 119 (8), 1829-1833.
183.
X. Hu, N. Takenaka, S. Takasuna, M. Kitano, H. Bandow, Y. Maeda and
M. Hattori, Anal. Chem., 1993, 65, 3489-3492.
184.
P.S. Francis, N.W. Barnett, S.W. Lewis and K.F. Lim, Talanta, 2004, 64 (2),
283-289.
185.
P.S. Francis, J.L. Adcock, J.W. Costin and K.M. Agg, Anal. Biochem., 2005,
336 (1), 141-143.
186.
X. Hu, N. Takenaka, M. Kitano, H. Bandow and Y. Maeda, Bull. Chem. Soc.
Jpn., 1996, 69 (5), 1179-1185.
187.
S. Kobayashi and K. Imai, Anal. Chem., 1980, 52 (9), 1548-1549.
188.
J.M. Terry, E.M. Zammit, T. Slezak, N.W. Barnett, D.C. Olson,
D.K. Wolcott, D.L. Edwards and P.S. Francis, Analyst, 2011, 136 (5),
913-919.
189.
M. Tsunoda and K. Imai, Anal. Chim. Acta, 2005, 541 (1-2), 13-23.
190.
P.S. Francis, N.W. Barnett, S.W. Lewis and K.F. Lim, Luminescence, 2004,
19 (2), 94-115.
191.
C.A. Marquette and L.J. Blum, Anal. Bioanal. Chem., 2006, 385 (3),
546-554.
192.
W.R.G. Baeyens, S.G. Schulman, A.C. Calokerinos, Y. Zhao, A.M. GarcíaCampaña, K. Nakashima and D. De Keukeleire, J. Pharm. Biomed. Anal.,
1998, 17 (6-7), 941-953.
193.
H.F. Launer, J. Am. Chem. Soc., 1932, 54, 2597-2610.
194
REFERENCES
194.
H.F. Launer and D.M. Yost, J. Am. Chem. Soc., 1934, 56, 2571-7.
195.
J.F. Pérez-Benito, C. Arias and E. Brillas, Int. J. Chem. Kinet., 1990, 22 (3),
261-287.
196.
S.A. Farokhi and S.T. Nandibewoor, Can. J. Chem., 2004, 82 (9), 1372-1380.
197.
K.A. Kovács, P. Gróf, L. Burai and M. Riedel, J. Phys. Chem. A, 2004, 108
(50), 11026-11031.
198.
I.B. Agater, R.A. Jewsbury and K. Williams, Anal. Commun., 1996, 33 (10),
367-369.
199.
C. Zhu, L. Wang, Y. Li and F. Gao, Fenxi Huaxue, 1997, 25 (4), 387-390.
200.
A. Townshend, J.A.M. Pulgarin and M.T.A. Pardo, Anal. Bioanal. Chem.,
2005, 381 (4), 925-931.
201.
A. Townshend, R.A. Wheatley, A. Chisvert and A. Salvador, Anal. Chim.
Acta, 2002, 462 (2), 209-215.
202.
J.L. Adcock, P.S. Francis and N.W. Barnett, J. Fluoresc., 2009, 19 (5),
867-874.
203.
S. Dash, S. Patel and B.K. Mishra, Tetrahedron, 2009, 65 (4), 707-739.
204.
S.J. Omelon and M.D. Grynpas, Chem. Rev. (Washington, DC, U. S.), 2008,
108 (11), 4694-4715.
205.
T. Slezak, J.M. Terry, P.S. Francis, C.M. Hindson, D.C. Olson, D.K. Wolcott
and N.W. Barnett, Anal. Chem., 2010, 82 (6), 2580–2584.
206.
M. Orbán and I.R. Epstein, J. Am. Chem. Soc., 1989, 111 (22), 8543-8544.
207.
Y.-T. Lee and C.-W. Whang, J. Chromatogr. A, 1997, 771 (1-2), 379-384.
208.
J.F. Pérez-Benito, E. Brillas and R. Pouplana, Inorg. Chem., 1989, 28 (3),
390-392.
209.
M. Tunçay, N. Yüce, B. Arlkan and S. Göktürk, Colloids and Surfaces A:
Physicochem. Eng. Aspects, 1999, 149 (1-3), 279-284.
210.
J.F. Pérez-Benito, C. Arias and E. Amat, Journal of Colloid and Interface
Science, 1996, 177 (2), 288-297.
211.
Z. Khan, Raju, M. Akram and U.D.D. Kabir, Int. J. Chem. Kinet., 2004,
36 (6), 359-366.
212.
U.D.D. Kabir, M. Altaf and M. Akram, J. Disper. Sci. Technol., 2008, 29 (6),
809-816.
213.
P.S. Francis, C.M. Hindson, J.M. Terry, Z.M. Smith, T. Slezak, J.L. Adcock,
B.L. Fox and N.W. Barnett, Analyst, 2011, 136 (1), 64-66.
195
REFERENCES
214.
J.S. Stamler and A. Slivka, Nutr. Rev., 1996, 54 (1), 1-30.
215.
M.Á. Medina, J.L. Urdiales and M.I. Amores-Sánchez, Eur. J. Biochem.,
2001, 268 (14), 3871-3882.
216.
C.M. Hindson, G.R. Hanson, P.S. Francis, J.L. Adcock and N.W. Barnett,
Chem. Eur. J., 2011, 17 (29), 8018-8022.
217.
T. Slezak, Z.M. Smith, J.L. Adcock, C.M. Hindson, N.W. Barnett,
P.N. Nesterenko and P.S. Francis, Anal. Chim. Acta, 2011, 707 (1-2),
121-127.
218.
A.I. Vogel, A Textbook of Quantitative Inorganic Including Elementary
Instrumental Analysis, Second edition, 1957, London, Longmans, Green and
Co.
219.
P.S. Francis, J.L. Adcock and N.W. Barnett, Spectrochim. Acta, Part A, 2006,
65 (3-4), 708-710.
220.
W.A. Jackson and D.R. Bobbitt, Anal. Chim. Acta, 1994, 285 (3), 309-320.
221.
J.L. Adcock, N.W. Barnett, R.D. Gerardi, C.E. Lenehan and S.W. Lewis,
Talanta, 2004, 64 (2), 534-537.
222.
G.P. McDermott, Ph.D Thesis, 2012, Deakin University, Geelong.
223.
J.P. Richie, Jr, P. Abraham and Y. Leutzinger, Clin. Chem., 1996, 42 (7),
1100-1105.
224.
D. Giustarini, I. Dalle-Donne, R. Colombo, A. Milzani and R. Rossi, Free
Radic. Biol. Med. , 2003, 35 (11), 365-1372.
225.
V. Angeli, H. Chen, Z. Mester, Y. Rao, A. D'Ulivo and E. Bramanti, Talanta,
2010, 82 (2), 815-820.
226.
R. Kand'ár, P. Záková, H. Lotková, O. Kucera and Z. Cervinková, J. Pharm.
Biomed. Anal., 2007, 43 (4), 1382-1387.
227.
G. Žuniü and S. Spasiü, J. Chromatogr. B, 2008, 873 (1), 70-76.
228.
G.P. McDermott, J.M. Terry, X.A. Conlan, N.W. Barnett and P.S. Francis,
Anal. Chem., 2011, 83 (15), 6034-6039.
229.
J.M. Terry, Z.M. Smith, G.P. McDermott, R.J. Waite, N.W. Barnett,
L.C. Henderson, J.M. Altimari and P.S. Francis, 2012, Submitted for
publication.
230.
S. Mohr, Y.H. Zhang, A. Macaskill, P.J.R. Day, R.W. Barber, N.J. Goddard,
D.R. Emerson and P.R. Fielden, Microfluid. Nanofluid., 2007, 3 (5), 611-621.
196
REFERENCES
231.
E.A. Boyle, B. Handy and A. Van Geen, Anal. Chem., 1987, 59 (11),
1499-1503.
232.
J. Du, Y. Li and J. Lu, Anal. Lett., 2002, 35 (14), 2295-2304.
233.
W. Cao, J.-H. Yang, C.-X. Sun, Y.-J. Chen and Q.-F. Gao, Luminescence,
2005, 20 (1), 20-24.
234.
J. Shi, K.-L. Zhao, X.-Y. Gong and Q.-G. Song, Fenxi Kexue Xuebao, 2006,
22 (1), 43-45.
235.
Y.-D. Liang, J.-F. Song and S.-Y. Zhao, Xi'an Keji Xueyuan Xuebao, 2004,
24 (3), 336-338, 349.
236.
D.M. Marcus and A.P. Grollman, Science, 2003, 301 (5640), 1669-1670.
237.
H. Ikkai, T. Nakagama, M. Yamada and T. Hobo, Bull. Chem. Soc. Jpn.,
1989, 62 (5), 1660-1662.
238.
S.A. Al-Tamrah, A.A. Al-Warthan and A.S. Al-Amri, J. Saudi Chem. Soc.,
1997, 1 (1), 1-8.
239.
L.-Q. Li, B.-Y. Qin, M.-L. Feng and J.-R. Lu, Shandong Keji Daxue Xuebao,
Ziran Kexueban, 2001, 20 (4), 25-27.
240.
N.W. Barnett, B.J. Hindson and S.W. Lewis, Anal. Chim. Acta, 1999, 384 (2),
151-158.
241.
J. Huclová, D. Šatínský, H. Sklenárová and R. Karlícek, Anal. Bioanal.
Chem., 2003, 376 (4), 448-454.
242.
J.M. Terry, S. Mohr, P.R. Fielden, N.J. Goddard, N.W. Barnett, D.C. Olson,
D.K. Wolcott and P.S. Francis, 2011, Submitted for publication.
243.
C.E. Lenehan, Ph.D Thesis, 2002, Deakin Univeristy, Waurn Ponds.
244.
A.J. Brown, Ph.D Thesis, 2009, Deakin University, Waurn Ponds.
245.
M. Camenzuli, H.J. Ritchie, J.R. Ladine and R.A. Shalliker, Analyst, 2011,
136 (24), 5127-5130.
246.
M. Camenzuli, H.J. Ritchie, J.R. Ladine and R.A. Shalliker, J. Chromatogr.
A, 2011, DOI: 10.1016/j.chroma.2011.09.021.
247.
R.W. Abbott, Ph.D Thesis, 1987, University of Hull, Hull.
248.
M.D. Cole, The Analysis of Controlled Substances, 2003, New York, Wiley.
249.
T. Wen, X. Zhao, G. Luo, J. Wang, Y. Wang, B. Yao, J. Zhao, J. Zhu and
Z. Yu, Talanta, 2007, 71 (2), 854-860.
250.
I.S. Lurie, P.A. Hays, A.E. Garcia and S. Panicker, J. Chromatogr. A, 2004,
1034 (1–2), 227-235.
197
REFERENCES
251.
C. Barnfield, S. Burns, D.L. Byrom and A.V. Kemmenoe, Forensic Sci. Int.,
1988, 39 (2), 107-117.
252.
F. Besacier, H. Chaudron-Thozet, M. Rousseau-Tsangaris, J. Girard and A.
Lamotte, Forensic Sci. Int., 1997, 85 (2), 113-125.
253.
L.A. King, Forensic Sci. Int., 1997, 85 (2), 135-147.
254.
S.P. Sharma, B.C. Purkait and S.C. Lahiri, Forensic Sci. Int., 2005, 152 (2-3),
235-240.
198