|
PULSED SOURCE TIE RESOLVED PHOSPHORIHETRY FOR QUANTITATIVE
ATD/OR QUALITATIVE ANALYSIS OF DRUGS
By
KAREN FRANCES HAR3AUGH
A DISSERTATION PRESENTTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1973
A MOiM bEiD Ur"HENT S
For the good things I have accomplished and learned
so far in my life, I owe a great debt of gratitude to a
number of people. The gifts of joy and wonder they have
given me are absolutely priceless. Though they are not
listed by name, they are remembered and loved.
Others have shown me that our world is less than
ideal, and these worthwhile lessons .I also acknowledge.
It is largely because of them that I sincerely hope I can
add to the world's balance on the side of goodness.
TABLE OF CONTENTS
Page
ACKNOWLEDG ETTS . . . . . . . . . ii
LIST OF TABLES. . . . . . . . . . v
LIST OF FIGURES . . . . . . . . . vi
ABSTRACT. . . . . . . . . . . .. viii
Chapter
I. INTRODUCTION. . . . . . . . . 1
II. PHOSPHORESCENCE THEORY AND 'EASUREMENT. . 6
Phobphorescence Theory . . . . . 6
Phosphorescence Lifetimes. . . . . 9
Pulsed Source Phosphorescence Measurement. 12
III. EXPERIMENTAL SYSTEM AND PROCEDURE . . . 18
Instrumentation . . . . . ... 18
Instrumental Procedure . . . . .. 28
Data Reduction . . .... . . . 32
Drugs and Reagents . . ... . . 39
IV. RESULTS AND DISCUSSION. . .. . . . 42
Noise and Precision. . . . . .. 42
Quantitative Analysis of Mixtures via Time
Resolved Phosphorimetry . . . . 51
Qualitative Analysis via Time Resolved
Phosphorimetry. . . . . . 62
iii
Chapter
V. SUDI~ILABY AMID FUTUYZE WORK . . ..
APPENDICES
Appendix I. . . . . . . . .... ..
Appendix II . . . . . . . . . ..
LIST OF REFERE:NCES. . ..............
BIOGRAPHICAL SKETCH ................ .
Page
. . 65
LIST OF TABLES
Table Page
1. Specifications and Operating Conditions for
the Flashtube Excitation Source . . . 22
2. Comparison of Curve Fit and Graphical Methods
for Data Reduction. . . . . . . 38
3. Illustration of the Precision Obtainable with
the Pulsed Source Phosphorimeter. . . 50
4. Phosphorescence Lifetimes (s) of Drugs in
Various Solvents. . . . . . . 54
5. A Comparison of the Relative Initial
Phospnorescence Intensities of Selected
Drugs in Various Solvents . . . . 56
6. Results of Quantitative Analysis of Drugs by
Pulsed Source, Time Resolved Phosphorimetry 59
7. Result of Quantitative Analysis of Drugs by
Pulsed Source, Time Resolved Phosphorimetry,
with Spectral Isolation . . . . . 61
LIST OF FIGURES
Figure Page
1. Modified Jablonski diagram showing the possible
transitions in the excitation and relaxation of
an organic molecule . . . . . . . 7
2. Plot of relative response (P /I ) versus
phosphorescence lifetime,r,p 0
according to Equation 13. . . . . . 17
3. Block diagram of the instrumental components of
the pulsed source phosphorimeter. . . . 21
4. Diagram of the Xenon flashlamp plexiglass
housing .. . . . . . . . . 23
5. Spectral irradiance of Xenon Novatron 599
Flashtube as a function of wavelength . . 25
6. Circuit diagram of the impedance matching
device between the signal average and the
flashlamp power supply. . . . . . . 29
7. Diagram of the events that occur during one
cycle of excitation and observation for pulsed
source time resolved phosphorimetry . . . 30
8. Experimental decay curve for a single
component sample (codeine, 0.4 mg/ml in EPA). 35
9. Semi-logarithmic plot of phosphorescence
intensity versus time for a single component
sample (codeine, 0.4 mg/ml in EPA). . . . 34
10. Experimental decay curve for a binary mixture
(morphine and codeine in ethanol) . . . 35
11. Semi-logarithmic plot of phosphorescence
intensity versus time for a binary mixture
(morphine and codeine in ethanol) . . . 37
12. Chemical structures of selected
phosphorescent drugs. . . . . . . 40
Figure Page
13. Variation in peal: to peak noise levels with
increasing amplification and load resistance 453
14. Comparison of (A) sourceless noise to (B)
phosphorescence decay of ethyl morphine
(0.0032 mg/nl in ethanol). . . . . . 44
15. Plot of the signal to noise ratio, S/N, versus
increasing sweep counts, N . . . . . 46
16. Plot of the signal to noise ratio versus
concentration of Amobarbital in chloroform . 48
17. Plot of the signal to noise ratio versus
concentration of Phenobarbital in chloroform 49
18. Phosphorescence emission spectra of drugs in
chloroform showing peak positions. . . . 53
19. Analytical curves for (A) Amobarbital, (B)
Quinine, (C) Cocaine, (D) Morphine, in
chloroform . . . . . . . . . 57
20. Analytical curves for (A) Codeine, (B) Ethyl
Morphine, and (C) Morphine, in ethanol . . 58
vii
Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
PULSED SOURCE TIME RESOLVED PHCSPHORIMETRY FOR QUA2NTITATIVE
AND/OR QUALITATIVE ANALYSIS OF DRUGS
By
Karen Frances Harbaugh
December, 1973
Chairman: James D. Winefordner
Major Department: Chemistry
A pulsed source time resolved phosphorimeter has been
described and its use for qualitative and quantitative
phosphorescence analysis of drugs in single and multicomponent
mixtures has been demonstrated. This system offers the
advantages that both qualitative identification and
quantitative determinations may be performed on certain
compounds without prior physical separation by use of
spectral and temporal resolution under various environmental
conditions.
A Xenon flashlamp, which provides high spectral
irradiance in the region most phosphors absorb, was
utilized as the excitation source in the experimental
system. A monochromator was used to spectrally isolate
viii
the components of interest, and a signal average was used
to monitor the decay of phosphorescence intensity as a
function of time. Based on differences in phosphorescence
lifetimes, the graphical separation of components was
accomplished by a semilogarithmic plot of phosphorescence
intensity versus time. Quantitation was achieved via
the proportional relationship between the initial phosphores-
cence intensity and the concentration of the species in
solution. Qualitative identification was accomplished by
the changes in the characteristic phosphorescence lifetimes
with changes in environmental conditions.
CHAPTER I
INTRODUCTION
Phosphorescence is a luminescence process whereby
light is emitted by a molecule following the absorption of
radiation of higher energy. It was first discussed by
Jablonski in 1935 [1]. The qualitative possibilities of
phosphorescence were suggested by Lewis and Kasha in 1944,
with considerable theoretical discussion [2], and Keics,
Britt and Wentworth evaluated phosphorescence as a quanti-
tative tool for chemical analysis in 1957 [3]. They showed
that a binary mixture of structurally similar compounds
could be resolved by judicious selection of excitation and
emission wavelengths and by choice of the delay before the
observation time after termination of the exciting radiation.
This was the first use of temporal resolution for chemical
analysis by phosphorimetry.
Time resolution in phosphorimetry utilizes the dif-
ference in molecular lifetimes just as spectral resolution
utilizes the differences in excitation and emission wave-
lengths. Phosphorescence lifetimes, for a given set of
environmental conditions, area unique characteristic of a
molecule, along with the excitation and emission spectra.
Should the spectral information of two (or more) molecules
be so similar that delineation of components is impossible
by spectral means, the molecules may be phosphorimetrically
resolved temporally, if their phosphorescence lifetimes
differ sufficiently.
Because most molecules generally absorb between 250
and 400 nm and emit between 350 and 500 nm, conventional
phosphorimetry, in which only excitation and emission
spectral resolution is utilized, is at best a rather
non-selective method of analysis. However, the phosphores-
cence lifetime- of molecules can differ greatly and the
influence of various solvents, particularly those containing
heavy atoms, such as iodide, upon the measured phosphores-
cence lifetimes can be even more striking. Therefore
time resolved phosphorimetry should be a powerful tool
possessing all of the sensitivity advantages of conventional
phosphorimetry as well as the selectivity advantages due to
time resolved. Fitzgerald and Winefordner [4] have shown by
information theory how time resolution plus spectral reso-
lution can provide up to 450 times the analytical information
provided by conventional phosphorimetry alone.
The initial time resolution studies made use of a
spectral continuum source with a mechanical shutter for
termination of exciting radiation. Keirs et al. [3]1 used a
mechanical phosphorescope which included a continuous light
3
source and a modified Becquerel type rotating disk.
St. John and Winefordner [5] later used an Aminco-Bowmnan
spectrophotofluorometer fitted with a manual guillotine
shutter and a dc readout coupled with a logarithmic con-
verter to time resolve mixtures of slow decaying phosphors
(of the order of seconds). Hollifield and Winefordner [6]
described a single disk mechanical phosphoroscope which
measured phosphorescence intensity at several intervals
along the emission decay curve.
The continuum source, mechanical shutter phosphoro-
scopic instruments, however, have certain limitations that
can be overcome by means of a pulsed source gated detector
system. The shutter transit time, and excitation and emission
viewing times are inflexibly linked in the former, and
limited by mechanical considerations (see [7]). Winefordner
[8] was first to suggest the advantages of pulsed source
gated detector instrumentation for time resolved phos-
phorimetry of fast phosphors (of the order of milliseconds)
in 1969, and O'Haver and Winefordner [7,9], had already put
continuous (CW) and pulsed source phosphorimetry on a sound
theoretical basis. Fisher and Winefordner [10] recently
compared conventional phosphorimeters to a unique pulsed
source system and demonstrated the advantages of the
instrument for the time resolution of selected phosphors.
Flash lamps have a greater spectral radiance in the region
where most phosphors absorb, which could lead to greater
sensitivities for certain phosphors, especially the fast
decaying species. Also, with portions of the excitation
and observation cycle independently variable, selectivity
is greatly increased.
O'Donnell, Harbaugh, Fisher and Winefordner [11]
used a modified version of this instrument to demonstrate
the utility of time resolved phosphorimetry for the analysis
of synthetic mixtures of halogenated biphenyls, Since this
time, however, no other applications of pulsed source time
resolved phospiorimetry have been shown.
The use of drugs, both legitimate and illicit, in
our time is extremely widespread and analysis of pure drugs
and of drug levels in human fluids is an important part of
the effective use and/or control of pharmaceuticals. Be-
cause conventional phosphorimetry has already proven to be
a sensitive analytical method for drugs which phosphoresce
[12-16], the application of time resolved phosphorimetry to
drug analysis should greatly extend the usefulness of
phosphorimetric drug assay.
The purpose of this research is to show the
versatility of a pulsed source, gated detector time resolved
phosphorimeter as applied to the analysis of selected drugs
in synthetic mixtures. The advantages and limitations are
5
illustrated as well as the quantitative and qualitative
possibilities. The feasibility of drug analysis in real
samples is demonstrated.
CHAPTER II
PHOSPHORESCES:CE THEORY AND IEASURI-ENT
Phosphorescence Theory
Absorption of radiation by an unsaturated organic
molecule promotes an outer electron from the lowest vi-
brational level of the ground state, where it normally
exists at room temperatures or below, to one of a number
of higher energy levels, depending on the energy of the
absorbed light. The molecule will eventually regain its
ground .state configuration by giving up this energy to
the environment via a combination of various deactivation
steps, which are illustrated by a modified Jablonski
diagram in Figure 1. The excited electron will first re-
turn to the lowest excited singlet level via vibrational
relaxation, internal conversion, and tunneling (for a de-
tailed discussion of these processes, see [17]). From the
excited singlet state, the molecule may return to the ground
state by internal conversion, whereby the molecule loses
its energy to the environment by nonradiative means. In
molecules with a large number of vibrational modes in the
ground state, as with aliphatic compounds, this process is
favored. Aromatic molecules, having rigid ring structures,
V
S2
S- S
V
T1
P /C
A / 01
Fig. 1.-=odified Jablonski diagram showing the possible
Transtio Tie of tr siio, seconds
A = Absorption. 1015
V = Vibrational relaxation1of
C = Internal conversion 10-13
F = Fluorescence 10-8
S = Intersystem crossing 10- 10-
P = Phosphorescence 10 10
generally have an energy gap that precludes internal con-
version and tunneling, and so luminescence becomes possible.
The molecule will fluoresce if a quantum of light is emitted
as the molecule drops to the ground state energy level.
This process is easily observed at room temperature.
Alternatively, the electron in the excited singlet
state may undergo a change in spin and occupy an excited
triplet state. This transition, called intersystem crossing,
is forbidden classically but allowable quantum mechanically
via spin-orbit coupling, which results in mixing of the
spin and orbital moments, so that the process has a finite,
though small probability. Spin orbit coupling is enhanced
by the presence of heavy atoms and paramagnetic ions, al-
though these are not a prerequisite for the phenomenon.
Once in the triplet state, the molecule may again
relax to the ground state configuration by internal con-
version, intersystem crossing back to the excited singlet
state with subsequent fluorescence (delayed fluorescence),
or it may phosphoresce by emitting a quantum of light of
lower energy than would be emitted via fluorescence.
The absorption of light energy and the return of the
electron to the first excited singlet state via vibrational
relaxation and internal conversion is much faster than the
subsequent transitions. Intersystem crossing and fluores-
cence have comparable rate constants and their transition
times are therefore about the same. Once the triplet state
is populated, though, reverse intersystem crossing is
favored over phosphorescence at room temperatures in low
viscosity liquids, and so luminescence from the triplet
state is rarely observed under these conditions. At low
temperatures, however, and in rigid media, the molecule can
maintain the triplet state configuration long enough for
phosphorescence to occur because collisional and vibrational
processes, which contribute heavily to the conversion back
to either the excited or ground singlet states, are at a
minimum.
Phosphorescence Lifetimes
As noted above, phosphorescence is in competition
with internal conversion and reverse intersystem crossing
for depopulation of the triplet state. Letting T represent
the number of molecules in the triplet state configuration,
then the rate of loss of this species with time is
dT= T+ k + ktT = k T + kT (1)
dt p ic ts p
where t is time and k k ic, and kts are the rate constants
for phosphorescence, internal conversion between the triplet
and ground states, and intersystem crossing from the triplet
state back to the first excited state, respectively. The
rate constants for all Dossible deactivation steps determine
the mean observable phosphorescence lifetime in the following
way,
i = -1 (2)
kp + k (2)
In the absence of competing nonradiative processes for de-
activation of the triplet state, Equation 2 reduces to
0 1 (3)
p
which is the inherent (natural) mean radiative lifetime, a
theoretical quantity independent of environmental para-
meters. Unlike the inherent mean lifetime, the mean
observable lifecime.is dependent upon temperature, matrix,
and impurities that may be present in the sample or solu-
tion. These effects can be at least partially controlled
to optimize analytical conditions.
The mean observable phosphorescence lifetime (here-
after referred to as the lifetime or phosphorescence life-
time) may vary with environmental temperatures above 2000 K,
but will do so negligibly at lower temperatures (for example,
see [18]). A change in lifetime with temperature is not
easily utilized analytically because it is difficult to
regulate temperatures in the low range, and the amount of
change may not be significant. The liquid nitrogen used to
maintain a temperature of 770 K for phosphorimetric analysis
eliminates the variation in lifetime with temperature as
well as maintaining a rigid matrix to reduce collisional de-
activation of the triplet state.
11
Phosphorescence spectra tend to exhibit much less depen-
dence on the dipole moment and hydrogen bonding properties of
the solvent than do fluorescence spectra [19] The observable
lifetimes, however, can and do change considerably with a
change in solvent characteristics [11, 20-2] due to the
effects which the solvent may have on the nonradiative rate
constants.
Heavy atoms will increase the spin-orbit coupling
coefficient of the wave function, thereby increasing the
probability of singlet-triplet intersystem crossing. Thus,
solvents containing a heavy atom will cause an increase in
. ki, and, accordingly, a decrease in qp. (These pertur-
i p
bations often remove the exponentiality of the decay, however,
so care must be taken that the nonexponential character
will not interfere [25,26]).
An increase in the polarity of the solvent can
decrease the energy separation between the excited triplet
or singlet state and the ground state when the molecule is
more polar in the excited state than in the ground state.
If this occurs, internal conversion becomes more probable,
I k. increases, and p again decreases. In general, then,
i 1 p
longer phosphorescence lifetimes are observed in nonpolar
solvents. Changes in solvent polarity will have a larger
effect on n->T1* transitions carbonyll type) than on->TTf*
transitions (aromatic ring type), because the carbonyl group
is more polar itself, and thus more susceptible to polar
interactions.
The ring structure and molecular substituents will
also affect the lifetime observed, but the degree and di-
rection of change depends to a great extent on the primary
molecular system itself. Generally, an increase in the size
of the fT system will result in a red shift in the spectrum
and a decrease in 1 Electron donating substituents also
p
tend to cause a red shift, but the direction of the change
in phosphorescence lifetime depends on the position of the
substituent [23,27].
The above statements apply generally to the factors
influencing lifetimes. In addition to the above general
considerations, other environmental conditions may also
affect the phosphorescence lifetimes which are observed.
Pulsed Source Phosohorescence 1,'easurement
The observed phosphorescence intensity for a phosphor,
as a function of time after termination of the exciting
radiation, is an exponential decay of the form
It = exp (- ) (4)
where It is the phosphorescence intensity at any time t,
I is the initial phosphorescence intensity (at t = 0),
.and T is the phosphorescence lifetime. In a mixture of
phosphors, the total phosphorescence intensity, I tT, is
given by
tT = ItA + ItB (5)
where the subscripts A and 3 refer to the different
components. For a multicomponent mixture of independently
absorbing and emitting phosphorescent species,
ItT = Iti exp (- t-) (6)
Time resolved phosphorimetry is based on Equation 6, and
is a means of separating the total intensity into its
individual components.
In order to accomplish quantitative analysis by time
resolution (measurement of phosphorescence signal as a
function of time), a strict relationship must exist between
the radiant flux observed, P and the concentration of a
given phosphor in a single or multicomponent sample. The
radiant phosphorescence flux emitted, P is related to the
radiant flux absorbed, Pabs, by
P = P 0abs (7)
p abs p
where 0 is the ratio of radiant flux emitted to radiant
flux absorbed. It is assumed in Equation 7 that there is
no pre-filter or post-filter effects and that only the
analyte absorbs and emits.
The radiant flux absorbed is given by
Pabs = o Pt = P (1 exp (-2.3 abc)) (8)
where P is the radiant flux (either CV or pulsed source;
however, for the latter case, P0 is the average radiant
flux) incident upon the absorbing sample, Pt is the radiant
flux transmitted, a is the molar absorption (extinction)
coefficient of the absorber, b is the absorption path length
(assuming a rectangular cell), and c is the concentration
of the analyte, in moles per liter. If(2.3 abc) is less
than 0.01, as is the case for dilute solutions of the
analyte, then
Pabs = P (2.3 abc) (9)
and,
P = P (2.5 abc)0 (10)
p o p
Thus a linear relationship does exist between the concentra-
tion of the phosphorescent species and the radiant flux
emitted as phosphorescence as long as the above assumptions
are true and as long as the analyte is dilute. Also, it is
necessary to assume a steady state condition of excitation
and emission.
In time resolved phosphorimetry, phosphorescence
intensity is measured at various delay times. Because P0
decreases with time after termination of the exciting pulse,
P will also decrease with time, and so the slope of an
P
analytical curve will decrease with increasing delay times.
This fact is used to advantage in the multiple analytical
curve method of time resolved phosphorimetry [35].
O'Haver and Winefordner [7] derived equations which
relate the operating parameters for a rotating can
phosphoroscope to the observed signal; these authors later
extended their considerations to the pulsed source gated
detector system [9]. They reported that the integrated
luminescence intensity, Pp, measured with a gated detector
for a single cycle is given by
t d t
[exp(- )][1-exp(- -)][1-exp(- -)]
P = I[ p (11)
[l-exp(- -)]
where I is the initial phosphorescence intensity, td is
the delay time between excitation and observation, t is
the observation (measurement) time, and q is the phosphores-
cence lifetime. Because the duration of the flash is small
compared to the lifetime of the phosphor (tf is microseconds
compared to in milliseconds or seconds), Equation 11
reduces to
t t
[exp(- -)][l-exp(- --)]
Pp = Io tf (12)
[1l-exp(- c)]
-I-r
This equation can be simplified further if the observation
time, t and therefore the total cycle time (tf + td +
t + t ) is much larger than '. Therefore,
p a
td
exp(- --)
P = Io t tt- -, (13)
which describes the phosphorescent species measured by the
instrumental system described in this work.
Some of the advantages of the pulsed source system
are apparent from an examination of these equations. It is
possible to achieve the enhancement of the signal of one
phosphor over that of another with a different lifetime by
the adjustment of the parameters td, t., and tc [17].
This exemplifies a major advantage of a pulsed source
system over a mechanical rotating disk or can arrangement,
because these parameters are not individually adjustable in
the latter. Figure 2 is a plot of the relative response
(P /1 ) versus lifetimes calculated with Equation 13 and
with typical parameter values, as stated in the legend.
Curve A illustrates the increased sensitivity pulsed source
phosphorimetry offers for fast phosphors in the presence of
long-lived species. On the other hand, curve B of Figure 2
shows how the influence of a short-lived species can be
eliminated by proper choice of a delay time. This is one
of the techniques that adds to the versatility of time re-
solved phosphorimetry.
17
io-2 (A)
10-3_
10 -
l-5
S10 (B'
10-6
10-4 10- 10-2 10-1 1 10
Phosphorescence lifetime (seconds)
Fig. 2.-Plot of relative response (P /1 ) versus
phosphorescence lifetime,, o
according to Equation 15.
The parameters used to calculate the curves are:
Curve A: tf = 1.2x10-6 s, t = 10 s, td = 5.2xl0-3 s.
Curve B: t = 1.2x10-6 s, t = 10 s, td = l.OxlO-2 s.
CHAPTER III
EXPERIMENTAL SYSTEM AITD PROCEDURE
Instrumentation
Pulsed source spectroscopy is not a recent develop-
ment. Physical chemists have been using this technique to
study fluorescence and phosphorescence decay characteristics
since Backstrom and Sandros first used such a system in
1958 [28]. Many workers since that time have improved upon
the technique (for example [29-31]); most workers have
utilized gated photomultipliers. Winefordner [8] suggested
the advantages of using this technique for phosphorimetric
analysis, and Fisher and Winefordner [10] built a time re-
solved pulsed source gated detector phosphorimeter and
compared it to conventional phosphorimetry. A time resolved
phosphorimeter utilizing a pulsed source has also been
described by Hamilton and Nagvi [532]. Hattori and Kato [335]
have reported time resolution studies using a continuum
source with a rotating disk to effect time variation of
pulsed irradiation of the sample. Mathiasch [34] also used
a continuum source but with a time averaging computer to
measure lifetimes of weak phosphors.
The system developed by Fisher and Winefordner [10]
has been modified for this work. For completeness, a full
description of the experimental set-up will be given. A
block diagram of the entire system is shown in Figure 3.
The components represented by dashed lines are options
available but were not used for this study. These compo-
nents could be used for a routine analysis to reduce costs,
for example, the use of emission filters in place of a
monochromator, or to facilitate data handling, as with the
use of a log converter or computer.
A Xenon Novatron 599 flashtube (Xenon Corporation) is
used as the source, and is more intense than the flashtube
previously used [535]. The specifications and recommended
operating conditions are listed in Table 1. (Flashtube
theory is briefly outlined in Appendix I and discussed more
fully in [36].) The voltage necessary for the triggering
pulse, as well as for the discharge in the lamp, is supplied
by a Model 457 Hicropulser power supply (Xenon Corporation).
For this work, 8 kV was applied across the lamp. This is
enough above the lamp threshold voltage to provide repro-
ducibility and good spectral radiance, but is below the
maximum value allowable in order to prolong the lamp life.
The flashtube is mounted in a plexiglass housing,
which is illustrated in Figure 4. The housing in turn
rests on a steel mount designed so that its distance from
Component
Designation
A
B
C
D
KEY TO FIGURE 3
Component
Xenon Nodel 457 Nicropulser (Xenon Corp.,
Medford, Mass.)
Xenon Novatron 599 Flashtube (Xenon Corp.)
Modified Aminco SPF sample compartment
(American Instrument Co., Silver Spring,
Md.)
Heath Company EU-700 scanning monochromator
(Heath Co., Benton Harbor, Michigan)
RCA IP28A photomultiplier tube in a Heath
EU-701-30 Photomultiplier Module
Keithley 244 High Voltage Po-.:er Supply
(Keithley Instruments, Cleveland, Ohio)
Load box (see text)
Tektronix 26A2 Differential Amplifier with
2601 Mainframe Power Supply (Tektronix,
Inc., Beaverton, Oregon)
Biomac 1000 Signal Averager (Data Labora-
tories, Ltd., Mitcham, Surrey, England)
Sargent Model SR strip chart recorder
(E.H. Sargent and Co., Birmingham, Alabama)
Impedance matching device (see text)
Log converter
Paper tape punch
Computer
Filter
Multiplier phototube with power supply
21
F
D J
FL T---
0
I
I --
P' .I
I
L-------
A
S!
i
Fig. 5.-Block diagram of the instrumental components of the
pulsed source phosphorimete'r.
TABLE 1
SPECIFICATIONS A2TD OPERATING CONDITIONS FOR THE FLASHTUBE
EXCITATION SOURCE
Xenon Novatron 599 Flashlamp
(Xenon Corp., Medford, Mass.)
Arc length
Maximum energy per pulse
Repetition rate
Rise time
Pulse width
Average maximum watts
Peak power
Life
3.8 cm
5 J
0-60 Hz
500 ns
1.2 us
300 W (at 60 Hz)
5 x 106 W
5 x 104 pulses
9 i
A= side view
B= back view
C= top view
D= flashlamp
E= " plexiglass
F= stainless steel base
G= banana lead socket
H= clip
B 2
G@
--u---j
1 1- 3%"----
C -T
3-3/4"
Fig. 4. Diagram of the Xenon flashlamp plexiglass
housing.
the sample compartment can be adjusted. A quartz lens 2 in
in diameter and with a 4 cm focal length is placed between
the lamp and sample cell to focus the lamp discharge image
on the sample to obtain maximum irradiance.
The flash repetition rate is controlled by a signal
from the signal average (discussed later) and was in no
case greater than 0.3 Hz. This allowed sufficient recovery
time for the lamp between flashes to insure greater repro-
ducibility, avoid spurious discharges, and prolong lamp life.
The output of a used (but not exhausted) flashtube
is shown in Figure 5. To obtain the spectrum, the lamp was
fired repeatedly while the monochromator was scanned across
the wavelength range. The source spectrum is not corrected
for photomultiplier response, which falls off below 300 nm.
Also, a Corning 754 UV filter is in place between the flash-
tube and the sample compartment to reduce stray light above
400 nm.
The quartz sample tube (25 cm x 5 mm i.d.) is placed
in a modified Aminco SPF sample compartment (American
Instrument Co.) which is fitted to a Heath Company EU-700
scanning monochromator by means of an adapter plate. The
compartment holds a quartz dewar flask (American Instrument
Co.), and the sample tube is held in position by the modi-
fied Varian spinner assembly described by Lukasiewicz,
Rozynes, Sanders and Winefordner [37]. The sample, how-
ever, was not rotated.
S0 00
4)
4-
-r4
400 350 300
Wavelength
Fig. 5.-Spectral irradiance of Xenon Novatron 599 Flashtube as a function of
wavelength.
The scanning monochromator is a single-pass f/6.8
Czerny-Turner unit, with reciprocal linear dispersion of
O
about 20 A/mn which will cover the region from 190-700 nm
with an IP28A multiplier phototube. Resolving power is
about 0.1 nm. Light through the monochromator is detected
by an RCA IP28A multiplier phototube housed in a suitable
light-tight compartment (EU-701-30 Photomultiplier Module,
Heath Co.). A high voltage power supply (Hodel 244,
Keithley Instruments) provides 800 volts to the phototube.
Because the differential amplifier (Hodel 26A2,
Tektronix, Inc.) required a voltage input, a load box with
variable load resistors was used to develop a voltage from
the anode current of the multiplier phototube. This voltage
was then amplified by the differential amplifier (which is
powered by the Tektronix 2601 mainframe power supply); the
resulting voltage is detected by a signal average. Care
must be taken that the load resistor used is not so large
that it will increase the instrumental time constant enough
to distort the phosphorescence decay signal. Because the
measuring system contains stray capacitance, C, to ground
as well as the load resistance, R, the measured voltage
will decay exponentially with a time constant of
t = R C (14)
where t is in s, R is inn_, and C in F. This time constant
should be an order of magnitude smaller than the shortest
phosphorescence lifetime measured in order to minimize dis-
tortion of the measured decaying signal. (The measured
lifetime would be longer than the phosphorescence lifetime
if the effect of the time constant of the measurement device
is significant.)
The signal average (Biomac 1000, Data Laboratories,
Ltd.) has many modes of operation [38]; however, only
signal averaging was necessary in this work. The delay
time between termination of the excitation flash and obser-
vation of the phosphorescence decay, the time of observation
(the sweep time, and the delay time between observation
of the phosphorescence and the initiation of the next flash
are variable: 0.32 ms to 5.12 s for the delay before; 5 ms
to 81.92 s for the sweep time, and 40 ms to 655.4 s for the
delay after, all in increments of two. (The delay after
observation may also be zero.) In addition, the number of
sweeps to be averaged is variable from 1 to 16,384, in
increments of two; a "counts per address" control allowed
scale adjustment for the oscilloscope display (and thus
for the output).
Output from the signal average is recorded by a
potentiometric recorder (Model SR, 3. H. Sargent and Co.);
the observed phosphorescence signal is plotted versus time.
The recorder is calibrated with a time mark generator for
an accurate time base. A log converter could have been
used between the signal average and the recorder to produce
a straight line, with the slope inversely proportional to
the lifetime, rather than an exponential decay; however,
this refinement was not deemed necessary for the present
studies.
The flashlamp power supply is triggered by a signal
from the signal average at the beginning of each observa-
tion cycle, and the two are interfaced with a small
impedance matching device. The electronic diagram of this
system is given in Figure 6.
Instrumental Procedure
The events occurring during one excitation and
observation cycle are illustrated in Figure 7. The signal
average initiates the cycle with a trigger pulse to the
flashlamp power supply. During the time width of the flash,
tf,phosphors in the sample are elevated to the triplet
state, as evidenced by their phosphorescence and represented
by the dashed line. When the flash terminates, the
phosphorescence intensity decays exponentially. After a
preset delay time, t (initiated at the beginning of the
flash, unlike td in Equations 11 through 15) the signal is
monitored for some chosen time, t and then a delay after,
ta, is used to lengthen the cycle if needed. This cycle is
repeated for the necessary number of sweeps, and the
Fig. 6.-Circuit diagram of the impedance matching device
between the signal average and the flashlamp
power supply.
T =D40D3 transistor
R =10 K
C = .001 F
V =+ 9V (dry cell)
I \ ___-- t --*
I
<-------------^ -------------- 4
Fig. 7.-Diagram of the events that occur during one cycle
of excitation and observation for pulsed source
time resolved phosphorimetry.
tf = Flash duration
t = Delay between excitation flash and observation
td = Delay between initiation of flashlamnp dis-
charge and observation time
t = Observation time (sweep time)
ta = Delay after observation
t = Total cycle time
c
resulting "averaged" decay curve is plotted by the recorder
for subsequent data reduction.
The delay before the observation time is necessary
to eliminate initial scatter associated with the flash or
to minimize luminescence signals due to fluorescence or
fast phosphorescence when measuring a phosphor with a long
lifetime; this delay is at least 0.32 ms. The sweep time
used for proper lifetime determination should be at least
5 xT'. This allows the phosphorescence intensity to decay
to within 2 per cent of zero luminescence, which is neces-
sary to establish an accurate baseline for the lifetime
calculations. It is possible to adjust the sweep time to
measure fast phosphors in the presence of longer decaying
species. The observation time in this case should be at
least five times the lifetime of the short phosphor, but
need only be long enough to establish a straight line portion
for the long-lived phosphor in order to allow subtraction
of the phosphorescence of long-lived species from the total
phosphorescence signal at the various delay times.
The delay after is used to maintain the flash rate
below 0.3 Hz and to allow the phosphorescence signal to
decay to near zero so that the initial intensity does not
steadily increase with successive excitation flashes. For
accurate quantitative work, the total cycle time should
always be at least five times the lifetime of the longest
lived phosphor.
Data Reduction
Various methods of data reduction for time resolution
of single and multicomponent mixtures have been described by
Fisher [35]. Only the methods employed in this work will be
described here.
A typical recorder tracing of the phosphorescence
signal as a function of time is illustrated in Figure 8.
By drawing a smooth curve through the tracing, average values
at various delay times can be plotted on semi-logarithmic
graph paper, as shown in Figure 9. The initial intensity,
I is measured by extrapolating the straight line of the
semi-logarithmic curve back to t = 0. The lifetime, 'T0,
is equal to the time it takes for the phosphorescence signal
to change from I to 36.8 per cent of I The points of
decay of a single component solution may also be fitted to
an exponential curve of the form y = y exb (compare to
Equation 4) by a calculator (Wang 600 Series, Wang Labora-
tories, Inc., Tewksbury, Mass.) to determine Io (yo) and
Tp(- -); this method was used to determine the phosphores-
cence lifetimes of the compounds of interest, and to
calculate the initial phosphorescence intensities of stan-
dard solutions for use in analytical curves.
Multicomponent curves were analyzed graphically. A
decay curve of a binary mixture is shown in Figure 10; the
4-
4,
H
O
CO
I I I I
26 58
tie Ins
ig 8.1orina dea urefra igecopnn
sapl (cden,0. g/ni.EA)
T = 49
i40- o
: = 44 ms
0
0 20 0
0
0
I I I
20 40 60 80
time, ms
Fig. 9.-Semi-logarithmic plot of phosphorescence intensity
versus time for a single component sample
(codeine, 0.4 mg/ml in EPA).
4
21 37 53
time, ms
Fig. 10.-Kx-erimental decay curve for a binary mixture
(morphine and codeine in ethanol).
corresponding semi-logarithmic plot is shown in Figure 11.
The decay curve at large delay times results in a straight
line which can be extrapolated back to t = 0 to find the
initial phosphorescence signal for the long-lived species.
When the contribution from this component is subtracted
from the original curve, the result is a straight line
representation of the fast phosphor. Again, extrapola-
tion back to t = 0 provides the initial phosphorescence
signal for the short-lived component. The respective life-
times are determined from the proper straight line sections.
When the lifetimes of the two components are quite similar,
as with morphine and codeine in this example, then the
measured lifetimes may be in error. The effect of this
error on the calculated initial phosphorescence intensity
and therefore on the quantitative results is best determined
by calculation of relative percent error of a determination;
this is discussed in a later section. For a ternary
mixture, this deletion procedure is repeated.
Because the curve fit method is used for lifetime
calculations as well as for the determination of initial
intensity of standard solutions, while the graphical method
is used for the calculation of the initial intensity of
components in mixtures for comparison with these standards,
it is important to know how the two methods correlate. In
Table 2 the initial intensities and lifetimes obtained by
I = 88
OOB
r-i 0
STB = 135 ms
-P \
o 0
S\ (B)
SO0
g 100-
0
0 0
S 50- O IoA 58
A O
O TA = 43 ms
50- 0
I IO
(A) O
20 40 60
time, ms
Fig. ll.-Semi-logarithmic plot of phosphorescence intensity
versus time for a binary mixture (morphine and
codeine in ethanol).
TABLE 2
COMPARISON OF CURVE FIT ANTD GRAPHICAL METHODS FOR DATA
REDUCTION
Percent
Error
Percent
CV Error
Curve fit 55.4 6.6 11.0 .81 .03 3.5
Graphical 53.9 6.7 12.4 .83 .054 4.1
The number of sweeps was 8 for each case. The component
measured was quinine (0.092 mg/ml in chloroform).
I = -nitial phosphorescence intensity
C6 = Relative standard deviation
T = Phosphorescence lifetime
the curve fit and graphical methods for quinine (.092 mg/ml
in chloroform) are comparable by both methods. The
graphical method offers the advantage of visibly indicating
any nonexponential character of a decay.
Quantitative analysis is accomplished by comparing
the initial intensity of the unknown to a standard plot of
initial intensities versus concentration (measured in the
same manner as the unknown), as determined for single
component standards. In this procedure, it is assumed that
the compounds in a mixture act independently.
Drugs and Reacents
The drugs used in this study were reagent grade
(Applied Science Labs, Inc., State College, Pennsylvania);
all drugs were ordered in accordance with BITDD regulations
and were used as received. Of the drugs ordered, those
found to be phosphorescent, and thus suitable for this
work, are illustrated in Figure 12.
Phosphorimetric quality anhydrous ethanol was pre-
pared in the laboratory according to the procedure outlined
in Appendix II. The chloroform is Certified A.C.S. Spectra-
analyzed (Fisher Scientific Company, Fair Lawn, New Jersey),
and was distilled once on a glass bead packed 6 ft column
to eliminate some of the phosphorescence background. The
ethyl ether, anhydrous (Nallinckrodt Chemical Works,
OH
Morphine
Codeine
W O 'OC2 H5
OH
Ethyl Morphine
C2H5 N
CH2 0
HC
"(C H3)2
Amobarbital
(C2H5)2
NH2
Procaine
Phenobarbital
CH3
H| I
H C-C-N H2
0H
Amphetamine
CH3
N
C6 H5 C-O-C2H5
0
Meperidine
CH3
H I
H -C-N-C H3
e th he t aine
Methamphetamine
OH 0
I H H H H I1
H CC --N--C H2 HC-C C-C-O-C H3
CH30C (CH2.2 N C3 C-o-c-C6H5
H H H -C CH
CH H H .H
CH2
Quinine Cocaine
Fig. 12.-Chemical structures of selected phosphores-
cent drugs.
St. Louis, Missouri), and cyclohexane (Hatheson Coleman and
Bell, Los Angeles, Calif.) were used as received. The EPA
used was a commercially prepared mixture (American Instru-
ment Co.).
Stock solutions were prepared by weighing a 1 mg
portion of the appropriate druS, and filling with solvent
to the mark in a 5 or 10 ml volumetric flask. Though these
small flasks could contribute to the error of analysis, they
were used because a very limited amount of each drug was
available. Serial dilutions for standard determinations
were made for each drug.
CI~t=IER IV
RESULTS AND DISCUSSION
Noise and Precision
The detection limit and accuracy of an analytical
method depends largely on the noise encountered in the
measurement, and the precision with which measurements can
be accomplished. Noise levels of this experimental system
were measured in the absence of sample or source irradiation,
and were found to be appreciable. In Figure 15 the
measured peak-to-peak noise level is shown at various
values of the differential amplifier and load resistor;
512 sweeps were performed by the signal average. As ex-
pected, the noise level increases with increasing amplifi-
cation, but it is not linear, as the increase in signal is
assumed to be. Noise is the main limitation in achieving
low detection limits, but Figure 13 shows that the relative
signal to noise ratio is slightly better at higher amplifi-
cation.
The major portion of the noise is due to dark current
shot noise from the photo detector and amplifier drift
and/or pickup noise. In Figure 14, the significance of the
instrumental noise level to an analysis is shown. An
(B) Load Resistance, 43
10 53 100
40-
SO-0
0O'
0 B
I 20-
0
0
10-
(A) Amplification, V
Fig. 13.-Variation in peaI: to peak noise levels with in-
creasing amplification and load resistance.
Curve A: 53 load resistance, varying amplification.
Curve B: 500V amplification, varying load resistance.
4Z,
(B)
-p
0
time
Fig. 14.-Comparison of (A) sourceless noise to (B)
phosphorescence decay of ethyl morphine (0.0052
mg/ml in ethanol).
analytical decay curve of ethyl morphine (0.0032 mg/ml in
ethanol) is compared to the sourceless noise level. There
is no significant contribution to the noise by the source
or sample variations.
In determining the lowest concentration level
detectable by an instrumental system, the important con-
sideration is the signal to noise ratio, S/N. The limit
of detection is commonly defined as the concentration of
species which provides a signal equal to twice the
peak-to-peak noise level observed. For a given solution,
the theoretical improvement in S/N with increasing sweep
counts on the signal averaSer (N) is shown in Figure 15,
curve A. Also shown in Figure 15, curve B, is the S/N
improvement for ethyl morphine (0.056 mg/ml in ethanol) with
increasing sweep counts. The improvement for the sample is
thus not ideal.
An increase in the number of sweeps does reduce the
noise level but by less than 47, where N is the number of
sweeps. There are disadvantages to extending the analysis
time by increasing sweep counts. The time of analysis could
become unreasonably long, especially in the case of slow
phosphors. The flash lamp has a limited lifetime, and is
rather expensive. Also, over a large number of sweeps, the
lamp intensity would change (as the lamp ages) making suc-
cessive analyses imprecise,
15-
-sH
CO
0 (A)
0-
o
40 80 120
Number of sweeps, N
Fig. 15.-Plot of the signal to noise ratio, S/N, versus
Curve A: Theoretical ("N) improvement in S/N.
Curve B: S/N improvement for ethyl morphine
(0.056 mg/ml in ethanol).
(0.055 mg/mi in ethanol).
In Figures 16 and 17 the increase in S/N with in-
creasing concentration is illustrated for amobarbital in
chloroform, and for phenobarbital in chloroform. If the
number of sweeps were the same for all concentration levels,
the curves would be linear, and a limit of detection for
each species could be calculated by an extrapolation of the
line to S/IN = 2. However, in this study, N was increased
for the lower concentrations in order to improve the pre-
cision and so a linear extrapolation is not valid because
of the uncertain improvement in S/N. For meaningful de-
tection limit v .lues a more detailed evaluation of the
signal to noise ratio at given sweep counts is necessary.
This was not ascertained and so detection levels for the
drugs used in this study are not given in this work.
Precision depends on the same noise sources as are
present in determining detection limits. Confidence in the
results of any analytical technique requires that the pre-
cision be good, although good precision does not necessarily
insure good accuracy. The reproducibility of measurement
by the experimental system used in this work is demonstrated
in Table 3. In part A, high and low concentrations of
quinine in chloroform (a snowed matrix) and in EPA (a 5:5:2
volume ratio of ether, isopentane and ethanol which forms a
clear glass) were measured several times, with the results
48
15-
OO
0
10-
o
0
+I I
10-3 0-2 10-1
Concentration of Amobarbital, mg/ml
Fig. 16.-Plot of the signal to noise ratio versus
concentration of Amobarbital in chloroform.
.6 .12
Concentration of Phenobarbital, mg/ml
Fig. 17.-Plot of the signal to noise ratio versus
concentration of Phenobarbital in chloroform.
TABLE 3
ILLUSTRATION OF THE PRECISION OBTAINABLE WITH THE PULSED SOURCE PHOSPHORIMETER
Concentration T Percent Percent
mg/ml N Solvent ~ o Error I c- Error SC
0.28 10 C 52.4 2.8 5.4 0.85 0.02 2.1 32
A 0.003 10 C 63.7 6.5 10.1 0.94 0.06 6.4 52
0.25 8 EPA 66.5 6.3 9.4 0.98 0.04 4.1 32
0.02 10 EPA 48.9 6.5 12.9 1.11 0.05 4.6 32
0.092 8 C 55.4 6.6 11.0 0.81 0.03 3.8 32
0.092 8 C 57.7 8.8 15.2 0.82 0.04 4.9 16
B 0.092 8 C 57.9 8.9 15.4 0.80 0.05 6.2 8
0.28 10 C 45.0 1.1 2.4 0.88 0.02 2.3 32
0.51 11 EA 58.8 4.2 7.1 1.11 0.05 3.0 32
C 0.003 10 C 66.3 2.7 4.1 1.94 0.06 6.0 52
Sections A and C: the sample is renewed between determinations.
Section B: the sample is not renewed between determinations.
Quinine is the phosphor used in all cases.
N = number of determinations.
I = mean initial phosphorescence intensity.
G= relative standard deviation.
7= mean phosphorescence lifetime.
SC= sweep counts.
Solvents used are:
C = chloroform.
EA = ethanol.
EPA = 5:5:2 volume ratio of ether, isopentane and ethanol.
as shown. As expected, the lower concentrations show
poorer precision, because the signal to noise ratios are
lower.
Because a signal average is used, it is expected
that an increase in the number of sweeps would improve
precision. This is indicated in part B by the improvement
of precision with an increase from 16 sweeps to 52, re-
sulting in about the same relative standard deviation as
in part A. Although it is reasonable to expect further
increase in precision with an increase in the number of
sweeps, this is not always desirable because time of
analysis becomes significant at long observation times when
a large number of sweeps are made, and, as already indicated,
noise levels in measurement may still be significant at high
sweep counts.
Quantitative Analysis of Hixtures via Time
Resolved Pho:.-horimetry
The purpose of this study is to show how time re-
solved phosphorimetry can be applied to the measurement of
drug mixtures. The drugs chosen to demonstrate quantitative
analysis of binary mixtures are not necessarily the best
possible ones to demonstrate time resolution nor are the
mixtures chosen comprehensive in terms of type of specific
mixtures which can be measured. These drugs do, however,
illustrate that time resolution is applicable to drug
analysis and can overcome difficulties of temporally
separating spectrally similar molecules. With pulsed
source, time resolved phosphorimetry, a mixture of drugs,
whose phosphorescence spectra overlap, can be quantitatively
determined without a prior physical separation step neces-
sary for conventional phosphorimetry, which would be time
consuming if not difficult or impossible to accomplish in
many cases. The fact that a physical separation step is
not needed prior to analysis is the major advantage of this
technique.
In time resolution phosphorimetry it is first neces-
sary to measure the spectra and lifetime of the phosphores-
cence of each species present in the sample. From this
data, it is apparent which drugs can be spectrally resolved,
which can be temporally resolved, and which can be resolved
by a combination of these techniques. The phosphorescence
emission spectra are illustrated in Figure 18. Because
phosphorescence spectra do not change appreciably with
changes in solvent, the phosphorescence peaks obtained with
chloroform are representative of the peaks obtained in the
other solvents. The lifetimes, however, vary significantly
with a change in the matrix, as shown in Table 4.
A third important consideration in time resolution
studies is the relative phosphorescence signals of the
components being analyzed. For the best results, the
Pheno-
barbital
(385)
300
Cocaine (415)
Procaine (445)
S\C
400
500
Wavelength,
Fig. 18.-Phosphorescence emission spectra of drugs in
chloroform showing peas: positions.
600
TABLE 4
PHOSPHORESCE3TCE LIFETILMES (s) OF DRUGS
IN VARIOUS SOLVENTS
Drug
Meperidine
Amphetamine
He thamohe t amine
Amobarbital
Cocaine
Phenobarbital
Procaine
Quinine
Ethyl morphine
Morphine
Codeine
Emission
peak, nm
590
390
395
400
415
385
445
515
510
510
515
EPA.
5.9
ins
.0015
2.2
1.8
ins
1.0
.045
.028
.040
DEE
7.5
9.6
.0016
2. 4
1.65
2.9
1.1
.040
.020
.059
.55
.52
.55
.0015
.82
.77
1.45
.85
.033
.014
.028
i2
i.
i:
1:
i.
ns ins
ns ins
ns .0016
- 2.2
ns 1.7
ns ins
ns ins
ns ins
ns ins
.041 .041
= 5:5:2 volume ratio of ether, isopentane, ethanol
= anhydrous ethanol
= chloroform
= cyclohexane
= diethyl ether
= Drug was at least partially insoluble in the
solvent, and so was not run.
EPA
EA
C
CH
DEE
ins
phosphorescence signal of the components being measured
should be about the same, because the species with small
phosphorescence signals compared to the other will be lost
in the noise of the observed phosphorescence signal. The
relative initial phosphorescence intensities of the drugs
in the various solvents are sho.rwn in Table 5.
Quinine, morphine, ethyl morphine, and codeine
exhibit severe overlap of emission spectra, and all
combinations of these four drugs are temporally resolvable
except codeine and ethyl morphine. The best results would
probably be atained in ethanol solvent, because the life-
time differences are comparable to those in chloroform and
greater than in EPA, but the relative intensity for
morphine is higher in ethanol. Analytical curves for
several drugs in chloroform and ethanol are shown in Figures
19 and 20. The lower sensitivity of morphine in chloroform
results in a lower slope of the analytical curve; this could
be due to some quenching effect. Time resolution of
morphine from a mixture of codeine, morphine and ethyl
morphine is also possible, as shown in Table 6.
Time resolution of morphine in a mixture of morphine
and ethyl morphine, and morphine in a mixture of morphine
and codeine illustrates the usefulness of the technique
with phosphors having similar lifetimes. These results
also illustrate the usefulness of time resolution
TABLE 5
A COMPARISONT OF THE RELATIVE INITIAL PHOSPHORESCENCE
INTENSITIES OF SELECTED DRUGS IN VARIOUS SOLVEN'-TS
Log I arbitrary units: EA C DEE EPA
Amphetamine 3 4 3.5
Methamphetamine 2.8 4 -
Amobarbital 6 6 5.6 6
Cocaine 4.6 5.3 4 4
Phenobarbital 5 4.6 5 4.6
Procaine 4 4 --
Quinine 3.95 4.5 4
Ethyl morphine 3.7 3.5 3.3
Morphine 3.7 2 4
Codeine 3.8 3.7 3.5
EA = Ethanol
C = Chloroform
DEE= Diethyl ether
EPA= 5:5:2 volume ratio of ether, isopentane and
ethanol
10_
0
) C
S102-
0 D
1-1 10
10-9 10-1
Concentration, mg/mi
Fig. 19.-Analytical curves for (A) Amobarbital, (B)
Quinine, (C) Cocaine, (D) Morphine, in chloroform.
1O
A
/ 0O
10-2 i0-1
Concentration, mg/ml
Fig. 20.-Analytical curves for (A) Codeine, (3) Ethyl
Horphine, and (C) Horphine, in ethanol.
P-)
-P
H
10-'
0
O
02
P)
0
C 0)-
0
0
P-I
TABLE 6
RESULTS OF QUANTITATIVE ANALYSIS CF DRUGS BY PULSED
TIHE RESOLVED PHOSPHORIHrTRY
Mixture
Solvent
Amount
Present
(mg/ml)
Amount
Found
(mg/ml)
Relative
Error
SOURCE,
Percent
Morphine EA 0.067 0.068 1.6
Codeine EA 0.023 0.022 4.3
Morphine EA 0.11 0.12 9.0
Ethyl
morphine EA 0.052 0.042 20.0
Morphine EA 0.11 0.16 45
Codeine +
Ethyl
morDhine EA -
Morphine EA 0.16 0.16 0
Cuisine EA 0.062 0.058 7.0
phosphorimetry for the analysis of drugs and drug metabo-
lites in the presence of each other (codeine and ethyl
morphine are both metabolites of morphine).
Time resolution of a short and long-lived species
is illustrated by the analysis of morphine and quinine.
This particular combination also has a potential practical
application because quinine is used as a diluent for
heroin; morphine,being the immediate metabolic product of
that drug, in vivo, if adequately separated from urine,
could be determined in the presence of the quinine which
is most likely extracted along with the morphine.
Amobarbital and phenobarbital overlap enough to
prevent spectral isolation; however, time resolution can be
used to determine both barbituates. Conversely, it would
be desirable to time resolve amphetamine and methamphetamine,
but their lifetimes do not differ enough in any of the
solvents used to allow it. Cocaine and phenobarbital
similarly do not show a great enough difference in lifetimes
to allow a temporal resolution. If they did, it could be
done at the point of maximum overlap between the two peak
emission wavelengths, or at each emission peak to maximize
the signal of interest while at the same time eliminating
a portion of the "interferent."
A combination of spectral and temporal resolution
analysis is shown in Table 7. Although it is possible to
TABLE 7
RESULT OF UAtCTTITATIVE ANALYSIS OF DRUGS BY PULSED SOURCE,
TIHE RESOLVED PHOSPHO2I-.ETRY, WITH SPECTRAL ISOLATION
Solvent
Amount
Present
( g/ml)
Amount
Found.
(ms/ml)
Relative Percent
Error
Mixture
Morphine C 0.038 0.042 10
Cocaine C 0.021 0.018 14
perform a quantitative analysis by spectral isolation alone,
the measurement of the phosphorescence lifetime provides
additional qualitative information if the nature of the
drug is unknown. Measurement of a nonexponential decay at
a given wavelength could also show that more than one drug
species is present in the measured sample.
Qualitative Analysis via Time Resolved Phosphorimetry
Because excitation and emission spectra and phosphores-
cence lifetimes of a molecule in a given environment are
characteristic of that molecule, and because these parameters
may change to a different degree for different molecules
with a change in environmental conditions, time resolved
phosphorimetry could be a useful technique for molecular
identification. From an examination of Table 4, it is
obvious that, of these drugs, any single component sample
could be identified qualitatively with spectra and lifetime
data except for meperidine and the amphetamines, and ethyl
morphine and codeine. (It could well be, however, that
with an extension of solvent data, these combinations would
be resolvable and therefore identifiable under some other
environmental conditions.) Furthermore, qualitative
analysis could be accomplished on multicomponent samples
as long as the species are spectrally and/or temporally
resolvable.
As an example of the feasibility of drug identifica-
tion by time resolved phosphorimetry, a qualitative unknown
(prepared by a colleague) was analyzed and found to contain
procaine and quinine. The emission spectrum shows peaks at
445 nm and 510 nm. The lifetime of the species at 510 nm
was found to be 0.85 s and was 1.45 s at 445 nm. In addi-
tion to procaine and quinine, the sample also contained
morphine, but at a lower concentration than the quinine.
Because the relative intensity of morphine is much lower
than quinine in chloroform, the morphine signal is not
visible and its presence went undetected. This points out
a limitation of time resolved instrumentation for qualita-
tive and quantitative work. Hinor components will not be
easily seen unless they happen to be well isolated spectrally.
A second unkno:.n mixture was found to contain
quinine, procaine, and a component that did not fit any of
the data in Table 4 (e = 430, 1V= 1.6 s). This third
component turned out to be caffeine, and would have been
identifiable had the data for it been catalogued. It is
likely that for qualitative work (such as on drug tablets
in a state laboratory), the information on all (or most) of
the possible components will be at hand, so that this would
not be a major problem. This sample also contained
morphine, which was not detected for the same reasons given
above.
64
An important consideration for both qualitative and
quantitative analysis is that, if a component does not
phosphoresce, or does so weakly, it will not interfere,
unless its concentration is large enough to cause pre-
filter effects in absorption.
CHAPTER V
SUMMARY A:D FUTURE WORK
A pulsed source phosphorimeter has been described and
its use for qualitative and quantitative analysis of drugs
in single and multicomponent mixtures has been demonstrated.
This system offers the advantages that both qualitative
identification and quantitative determinations may be per-
formed on certain compounds without prior physical separation
by use of spectral and temporal resolution under various
environmental conditions.
A flashlamp was utilized as the excitation source in
the experimental system. Spectral isolation of the
components of interest was accomplished by a monochromator.
A signal average was employed to monitor the phosphores-
cence intensity as a function of time. The initial
phosphorescence intensity and the phosphorescence lifetime
were determined from a plot of the log of the measured
intensity versus time. Temporal resolution was performed
by a graphical delineation of components and subsequent
determination of individual intensities and lifetimes.
Quantitation is achieved via the proportional relationship
between the initial phosphorescence intensity and the concen-
tration of the species in solution.
There is much yet to be done with this system. The
noise level of the time resolved phosphorimeter is unusually
high. An optimization of this system with respect to noise
levels would result in much improved limits of detection.
An investigation into the application of this technique to
real samples, whether they be the qualitative identification
or pure or mixed drug forms (as tablets and powders) or the
quantitative analysis of drug levels in human fluids, is a
logical extension of this study. Also a study on variations
of the pulsed source technique with instrumentation improve-
ments (greater irradiance of the sample by means of mirrors
behind the source, or specific cycle programming to opti-
mize the procedure for a particular component) would improve
sensitivity and selectivity of the method. The use of a
computer for data reduction would facilitate numerical
computations.
Dedication of this technique to a specific analytical
problem could allow the simplification of instrumentation,
with an appropriate reduction in cost. This method would
then be more commercially feasible, although it is a
powerful research tool as it is.
APPENTDT CES
APPE11DIX I
General Theory of Flashlamp Operation
A flashlamp converts electrical energy into radiant
energy by discharging a high voltage capacitor through a
tube of inert gas. Xenon is used most often as the fill gas
because it has a high cross-sectional area for efficient
collisional excitation by high speed electrons, and low
excitation and ionization potentials.
Flashlamps emit high spectral irradiance in the
visible and ultraviolet regions. Thus they have considerable
analytical utility in luminescence spectrometry because most
luminescent species absorb in the ultraviolet region.
The electrical discharge is initiated by a high
voltage trigger pulse. The energy dissipated is given by
J = 1 C V2 (A-l)
where J is in joules, C is the capacitance in pF, and V is
the voltage in kV. The duration of the flash is given by
t = RC (A-2)
7 R C1
where tf is the duration time, C is again capacitance, and
Rf is the effective arc resistance, which depends on arc
length, tube diameter, type and pressure of the fill gas,
and the voltage.
69
Xenon flash tubes with a quartz envelope have an
effective color temperature of about 7,0000 K, as do DC
xenon arcs, but the emissivity is much greater, especially
at the peak of the flash.
APPENDIX II
Preparation of Phosphorimetrically Pure Ethanol
Phosphorimetrically pure anhydrous ethanol is necessary
in order to obtain good analytical results, free from an
interfering background. Commercially prepared absolute
ethanol is unsuitable because it contains interferents
either as a result of the preparation method (as with a
benzene azeotrope) or due to phosphorescing denaturants.
A modification of the Lund and Bjerrum method [39]
for the preparation of anhydrous alcohols was employed to
prepare phosphorimetrially clean anhydrous ethanol. Into
a 2 1 round bottom flask is placed 1.5 1 of 95 per cent
ethanol, 10 g or more of magnesium turnings, and several
g of iodine. The mixture is refluxed for at least 6 h
during which time the water is converted to Mg(OH)2, with
the evolution of hydrogen and the formation of the
hydroxide quite visible. The reaction is as follows,
Mg + 2 EtOH = H2 + Mg(OEt)2 ,
Mg(OEt)2 + 2H20 = Mg(OH)2 + 2EtOH .
The magnesium must be kept in excess. The iodine catalyzes
the reaction and is reduced during the refluxing.
After refluxing, the alcohol is distilled. The middle
cuts show an appreciable decrease in phosphorescence back-
ground.
LIST OF REFERE::CES
1. A. Jablonski, Z. Physzik, 94, 38 (1935).
2. G. N. Lewis and M. Kasha, J. Am. Chem. Soc., 66, 2100
(1944).
3. R. J. Keirs, R. D. Britt, Jr., and V. E. Ventworth,
Anal. Chem., 29, 202 (1957).
4. J. J. Fitzgerald and J. D. Winefordner, Rev. Anal.
Chem., in press.
5. P. A. St. John and J. D. Winefordner, Anal. Chem., 39,
500 (1967).
6. H. C. Hollifield and J. D. Winefordner, Chem. Instr.,
1, 31 (1969).
7. T. C. O'Haver and J. D. Winefordner, Anal. Chem., 538,
602 (1966).
8. J. D. Winefordner, Accts. of Chem. Res., 2, 361 (1969).
9. T. C. O'Haver and J. D. Winefordner, Anal. Chem., 38,
1258 (1966).
10. R. P. Fisher and J. D. Winefordner, Anal. Chem., 44,
948 (1972).
11. C. M. O'Donnell, K. F. Harbaugh, R. P. Fisher and
J. D. Winefordner, Anal. Chem., 45, 609 (1975).
12. J. D. Winefordner and H. W. Latz, Anal. Chem., 35,
1517 (1965).
13. J. D. Winefordner and H. Tin, Anal. Chem. Acta., 32,
64 (1965).
14. Ibid, 31, 259 (1964).
15. W. J. McCarthy, P. A. St. John and J. D. Winefordner,
"Phosphorimetry as a Heans of Chemical Analysis," in
S. Udenfriend, ed., Fluorescence Assay in Biology and
Medicine. Academic Press, NT.Y., 1970.
16. H. C, Hollifield and J. D. Winefordner, Talanta, 12,
860 (1965).
17. J. D. Winefordner, S. G. Schulman and T. C. O'Haver,
Luminescence Spectronetry in Analytical Chemistry.
Wiley-Interscience, New York, 1972.
18. D. M. Grebenshchikov, N. A. Kovrizhnykh and S. A.
Kozlov, Opt. Spektrosk., 31, 755 (1971).
19. S. G. Schulman, Fluorescence News, Vol. 6, Number 4,
Hay, 1972.
20. K. F. Harbaugh, C. H. O'Donnell and J. D. Winefordner,
Anal. Chem., 45, 381 (1973).
21. J. J. Aaron, J. J. Mousa and J. D. Winefordner, Talanta,
20, 279 (1973).
22. R. J. Lukasiewicz, J. J. Mousa and J. D. Winefordner,
Anal. Chem., 44, 963 (1972).
23. J. E. Kent, H. F. O'Dwyer and R. J. Stephens, Chem.
Phys. Let., 10, 201 (1971).
24. D. H. Haaland and G. C. Nieman, J. Chem. Phys., 59,
1013 (1973).
25. G. C. Giachino and D. R. Kearns, J. Chem. Phys., 52,
2964 (1970).
26. S. P. NcGlynn, H. J. Reynolds, G. W. Daigre and N. D.
Christodoyleas, J. Phys. Chem., 66, 2499 (1962).
27. N. C. Yang, D. S. McClure, S. L. Murov, J. J. Houser,
R. Dusenbery, J. Am. Chem. Soc., 89, 5466 (1967).
28. H. J. T. Backstrom and K. Sandros, Acta. Chem. Scand.,
12, 823 (1958).
29. R. G. Bennett, Rev. of Sci. Instr., 31, 1275 (1960).
30. H. L. Bhaumik, G. J. Clark, J. Snell and L. Ferder,
Rev. of Sci. Instr., 36, 37 (1965).
31. C. Lewis, W. R. Ware, L. J. Doemeny and T. L. Neuzek,
Rev. of Sci. Instr., 41, 107 (1975).
32. T. D. S. Hamilton and K. Naqvi, Anal. Chem., 45, 1581
(1973).
33. S. Hattori and Y. Kato, J. Molec. Spec., 59, 432
(1971).
34. B. Mathiasch, Anal. Let., 4, 519 (1971).
35. R. P. Fisher, Ph.D. Dissertation, University of
Florida, 1971.
36. Amglo Corporation, Flashtube Tngineering Manual,
Chicago, Ill.
37. R. J. Lukasiewicz, P. A. Rozynes, L. B. Sanders and
J. D. Winefordner, Anal. Chem., 44, 237 (1972).
38. Biomac 1000 Operating Manual, Data Laboratories, Ltd.,
Surrey, England.
39. H. Lund and J. Bjerrum, Ber., 64B, 210 (1931).
BIOGRAPHICAL SKETCH
Karen F. Harbaugh was born in Andalusia, Alabama,
on January 8, 1945. After attending public schools in
Washington state, California, Alabama, and Florida, she
graduated from Renton High School in Renton, Washington.
She obtained a Bachelor of Science degree from the
University of Florida in December, 1966. Graduate studies
were begun immediately thereafter, but were temporarily
set aside when she went to work for Shell Chemical Company
in Deer Park, Texas. She returned to graduate school at
the University of Florida in January, 1970, and has worked
toward a doctorate in chemistry since that time.
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
,Gerhard H. Schmid
Associate Professor of Chemistry
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Willis 3. Person
Professor of Chemistry
This dissertation was submitted to the Graduate Faculty of
the Department of Chemistry in the College of Arts and
Sciences and to the Graduate Council, and was accepted as
partial fulfillment of the requirements for the degree of
Doctor of Philosophy.
December, 1973
Dean, Graduate School
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
G'erhard M. Schmid
Associate Professor of Chemistry
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Willis B. Person
Professor of Chemistry
This dissertation was submitted to the Graduate Faculty of
the Department of Chemistry in the College of Arts and
Sciences and to the Graduate Council, and was accepted as
partial fulfillment of the requirements for the degree of
Doctor of Philosophy.
December, 1973
Dean, Graduate School
M 7 2 2 5 8 1 4
|
RSS
HIDE
