PART I
IMPLEMENTATION OF
ABSORPTION-CORRECTED .
FLUORESCENCE MEASUREMENTS

PART II I
FLUOROMETRIC AND OTHER STUDIES
OF THE REACTION OF ALUMINUM (III)
AND FLAVONOL IN ABSOLUTE ETHANOL

Dissertation for the Degree of Ph D. *
MICHIGAN STATE UNIVERSITY
PATRICK MICHAEL KELLY
1‘976

 

This is to certify that the

t ‘ It'ld
PART I: IMPLEMENTATTON‘OE ABSORPTION-
CORRECTED FLUORESCENCE MEASUREMENTS

PART II: FLUOROMETRIC AND OTHER STUDIES
OF THE REACTION OF ALUMINUM (III)
AND FLAVONOL IN ABSOLUTE ETHANOL

presented by

Patrick Michael Kelly

has been accepted towards fulfillment
of the requirements for

_Eh_.D_.__ degree in My

Z a. .5

Major professor

Date Mill/27‘
V I ,

0-7 639

 

 

5: am a son:

 

 

 

 

. \

Income .7 ‘3
E BOOK BIIIUERY NC.

uamnv muons
_ WIQNI'Llfllw I,

FLUO
OF ALU

A sect
“'35 COnst:
include 1:]
ComPuter I
tance Sta
and the u
extend t}:

The p-
mately :0

Retric 3C

 

while the
The i
work On a

C

p
or the CC

 

6 {/0 / '/ /l’é

ABSTRACT

PART I
IMPLEMENTATION OF ABSORPTION-CORRECTED
FLUORESCENCE MEASUREMENTS
PART II

FLUOROMETRIC AND OTHER STUDIES OF THE REACTION
OF ALUMINUM (III) AND FLAVONOL IN ABSOLUTE ETHANOL

BY
Patrick Michael Kelly

A second computer-centered spectrophotofluorometer

was constructed. Improvements over the prototype instrument

include the capability of using several different sources,
computer controlled wavelength scans, the use of a reflec-
tance standard to calibrate the emission detection system
and the utilization of a methylene blue quantum counter to
extend this calibration to 7001mn.

The photometric accuracy of the instrument is approxi-
mately t0.l% T in the range from 1 to 100% T. The fluoro-
metric accuracy was estimated to be approximately i5%,
while the fluorometric precision was found to be 11%.

The instrument was used to extend some of the previous
work on absorption-corrected fluorescence measurements.
Some of the assumptions previously made in the derivation

of the correction equation were carefully examined. This

cfitical ex
emission op
the assumpt
Plementatic
Tents based
measurement
fumd to be
This in
electronic
HR band is
excitation
fluorescenc
scrption 135‘
although th
Based 0-
smctures
teen prOPOS‘
am only on
experimental
ing structu
iientical e
ethoxide tc
ifidge conf

The pre

 

C1
«avonol c}

in

detail.

 

Patrick Michael Kelly

critical examination resulted in the construction of an
emission optical system which met all of the criteria of
the assumptions. A scheme was then developed for the im-
plementation of absorption-corrected fluorescence measure-
ments based on this model. Subsequently, absorption-corrected
measurements were made up to an absorbance of 2.0 and were
found to be accurate to 13%.

This instrument was used in the discovery of a new
electronic absorption band in the Spectrum of flavonol.
The band is located in the spectral region near 410 nm, and
excitation at this wavelength results in the emission of
fluorescence at 484 nm. Tentatively, this electronic ab-
sorption band has been assigned to a (n,n*) transition
although the experimental result are somewhat inconclusive.

Based on the available experimental evidence, various
structures for aluminum (III) in absolute ethanol have
been proposed. All three are hexameric cyclic structures,
and only one of these could be eliminated on the basis of the
experimental evidence produced in this study. The two remain-
ing structures seem to fit the experimental criteria of
identical environments for all of the six aluminum ions, an
ethoxide to aluminum (III) ratio of 2.5 and an alternating
bridge configuration.

The previous work on the formation of aluminum (III)-
flavonol chelates in absolute ethanol has been re-examined

in detail. The formation of a 1:1 chelate in dilute base

and the l:
Tbuffered
Bladditio:
aluminum-U
solutions.
and a SCIII
for these I

The fll
hlsolutiox
dilute base
chelates 11
of 0.01, o.
of the 1:1
absolute 6t
reSpecuve]
been attrit
in unbuffez
intersySte:

The prc

Vhi
‘Ch Presx

 

appear to 1
Of .

ROI/ed Line-
3
me Ethan

f0 5

tallograp‘.

 

Patrick Michael Kelly

and the 1:1, 2:1 and 6:1 aluminum-to-flavonol chelates in
unbuffered absolute ethanol solutions has been confirmed.

In addition, this study has also shown that 1:1 and 2:1
aluminum-to-flavonol chelates also exist in acidic ethanol
solutions. Based upon the proposed aluminum (III) structures
and a series of potentiometric titrations, various structures
for these chelates in absolute ethanol have been proposed.

The fluorescence quantum efficiencies of the chelates
in solution have also been measured. The 1:1 chelate in
dilute base, and the 1:1, 2:1 and 6:1 aluminum-to-flavonol
chelates in unbuffered solutions, have quantum efficiencies
of 0.01, 0.01, 0.03 and 0.70, respectively. The efficiencies
of the 1:1 and 2:1 aluminum-to-flavonol chelates in acidic
absolute ethanol solutions were found to be 0.03 and 0.45,
respectively. These trends in quantum efficiencies have
been attributed to an increase in rigidity for the chelates
in unbuffered solutions and to a decrease in the rate of
intersystem crossing for the acidic chelates.

The prospects for obtaining the solids of the chelates,
which presumably exist in dilute absolute ethanol solutions,
appear to be quite small. Crystallization by the addition
of various solvents to the ethanolic chelate solutions has
proved unsuccessful. Evaporation of the solvent from these
same ethanolic solutions did yield crystalline powders. Un—
fortunately, the crystals were unsuitable for an X-ray crys-

tallographic structure analysis. In addition, Raman, infrared

and alumint
only one cl
the solid :

lieved to l

 

 

Patrick Michael Kelly

and aluminum-27 nmr spectroscopy have all indicated that
only one chelate is formed in concentrated solutions and in
the solid state. The stoichiometry of this chelate is be-

lieved to be three flavonate ions per aluminum ion.

FLUOROI
or ALUMI‘

 

in

 

PART I

IMPLEMENTATION OF ABSORPTION-CORRECTED
FLUORESCENCE MEASUREMENTS

PART II

FLUOROMETRIC AND OTHER STUDIES OF THE REACTION
OF ALUMINUM (III) AND FLAVONOL IN ABSOLUTE ETHANOL

BY
Patrick Michael Kelly

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1976

The au
guidance,
of this st

The au
for his ma

Specia
assistance
photofluon
aid in the
system, to
MI. Norman
Ronald Haas

 

ACKNOWLEDGMENTS

The author wishes to thank Dr. Andrew Timnick for his
guidance, encouragement and friendship throughout the course
of this study.

The author also wishes to thank Dr. Stanley R. Crouch
for his many helpful suggestions as second reader.

Special thanks are given to Dr. John Holland for his
assistance and advice during the construction of the spectro-
photofluorometer, to Dr. Thomas Edwards for his invaluable
aid in the design of part of the fluorometer emission optical
system, to Mr. Deak Watters for his machining expertise, to
Mr. Norman Young for his programming assistance and to Mr.
Ronald Haas whose abilities in electronic design and fabrica-

tion far exceed his abilities as a fisherman and hunter.

ii

LIST OF TAB

LIST OF FIG

Part I - IM

IV,

CO
ME

. Introd

. Theore

Evalua

. Instru

Optica
EleCtr
Comput
Instru

Instru
precis

Experi

 

Emissi

AbSor;

 

Automa

o Regult

Basic
Implep

Corr9r

Summa

 

LIST
LIST

Part

II.

III.

IV.

VI.

OF TABLES . .

OF FIGURES. .

TABLE OF CONTENTS

0 I O I O O O O O C O

I - IMPLEMENTATION OF ABSORPTION -

CORRECTED
MENTS . .

Introduction.
Theoretical .
Evaluation of
Instrumental.
Optical . . .
Electrical. .

Computer. . .

FLUORESCENCE MEASURE-

Instrumental Improvements . . . . .

Instrumental Accuracy and

Precision . .

Experimental.

Emission Optical System . . . . . .

Absorption-Corrected Fluorescence .

Automated Corrections . . . . . . .

Results and Discussion. . . . . . .

Basic Assumptions . . . . . . . . .

Implementation. . . . . . . . . . .

Correction Accuracy . . . . . . . .

Summary and Conclusions. . . . . .

iii

Page
vii

viii

ll
18
20
20
22
26
28

30
33
33
4O
45
47
47
49
SO
53

Chapter

Part II - F?
O:

A
I.Introd
IL Experi
Instru
Chemic

Purifi

Prepar

EXperi

 

Chapter

Part II - FLUOROMETRIC AND OTHER STUDIES

I.

II.

III.

OF THE REACTION OF ALUMINUM (III)
AND FLAVONOL IN ABSOLUTE ETHANOL

Introduction. . . . . . . . . . . .
Experimental. . . . . . . . . . . .
Instrumentation . . . . . . . . . .
Chemicals . . . . . . . . . . . . .
Purification of Flavonol. . . . . .
Preparation of Stock Solutions. . .
Experimental Procedures . . . . . .
Spectrofluorometer. . . . . . .
Photometric and Fluorometric

Titrations. . . . . . . . . . .
Apparent pH Measurements. . . .
Emission Measurements . . . . .
Concentrated Chelate Solutions.
Raman Spectra Measurements. . .
Infrared Spectra Measurements .
x-ray Powder Diffraction

Measurements. . . . . . . . . .
Aluminum-27 NMR Measurements. .

Results and Discussion. . . . . . .

Absorption and Fluorescence Spectra
of Flavonol . . . . . . . . . . . .

Emission Spectrum of Flavonol at
77 KO 0 O O O O O O O O O O O O O O

(n,n*) Excited State of Flavonol. .

Absorption and Fluorescence Spectra
of the Chelates . . . . . . . . . .

Emission Spectra of the Chelates
at 77 K O I O O I O O O O O O O O O

Chelate Stoichiometries . . . . . .

Basic Solutions . . . . . . . .

iv

Page

54
55
61
61
62
63
64
65
65
66
67
67
67
68
68

69
69

71

71

71
76

84

86
93
93

Chapter

‘

Raman
Infra
Alumi
Poten

Alumi
in Ab

 

 

 

Chapter
Neutral Solutions . . . . . . . . .
Acidic Solutions. . . . . . . . . .
Raman Spectra . . . . . . . . . . . . .
Infrared Spectra. . . . . . . . . . . .
Aluminum—27 NMR . . . . . . . . . . . .
Potentiometric Titrations . . . . . . .

Aluminum (III) Species Structures
in Absolute Ethanol . . . . . . . . . .

Basic Solutions . . . . . . . . . .
"Neutral" Solutions . . . . . . . .

Acidic Solutions. . . . . . . . . .
Chelate Structures. . . . . . . . . . .
Basic Solutions . . . . . . . . . .
"Neutral" Solutions . . . . . . . .

Acidic Solutions. . . . . . . . . .
Chelate Quantum Efficiencies. . . . . .
Preparation of Solid Chelates . . . . .

IV. Summary and Conclusions . . . . . . . .

APPENDICES

Appendix One - Mathematical Proof of
Assumption Four . . . . . . . . . . . .

Appendix Two - Buffered Digital I/O Pin
Configurations (DEC DRB-EA) . . . . . .

Appendix Three - Instructions for the
Construction and Implementation of an
Emission Detection System Correction
Table 0 I O O O I I O O O O O O O O O 0
Building the Correction Table . . .

Rhodamine B Section. . . . . . .
Methylene Blue Section . . . . .

Program PMT O O O O O O O O O O O O

Page
94
94
96
99

110

114

123
123
123
128
129
129
130
135
135
139

140

144

148

150
151

151
153

153

Chapter

Appen
Imple
Fluor

P
P

BIBLIOGRAP

 

Chapter

Data Processing . . .

Appendix Four - Instructions for the
Implementation of Absorption-Corrected

Fluorescence Measurements
Program LSSQ. . . . .

Section One. . . .
Section Two. . . .

Program RTFACT. . . .
Program ARTCAL. . . .

BIBLIOGRAPHY . . . . . . . . .

vi

Page

157

161
163

164
166

167
171
181

Table

II

III

Fl

De

ce

Wa

O)

DI

 

LI ST OF TABLES

Table Page
I Variables for 90°, Steady State
Fluorescence Measurements . . . . . . . . . 3
II Dependence of the Delayed Fluores-

cence of Flavonol on Excitation
Wavelength and the Presence of

Oxygen. . . . . . . . . . . . . . . . . . . 74

III Dependence of Delayed Fluores—
cence on Chelate Stoichiometry,
Excitation Wavelength and the

Presence of Oxygen. . . . . . . . . . . . . 88

IV Raman Spectra of Flavonol, 2:1

and 6:1 Metal-to-Ligand Chelates. . . . . . 97

V Correlation of Quantum Efficiency
with Chelate Stoichiometry and

Ionic Charge. . . . . . . . . . . . . . . . 137

vii

Figure

Op

Ci
Os

F1
til
Po

 

Bl
mu

Figure

10

11

12

LIST OF FIGURES

Optical Diagram for the Computer-
Centered Spectrophotofluorometer. .

Circuit for the Vibrating Bridge
Oscillator. . . . . . . . . . . . .

Flag Circuit for Computer Recogni-
tion of the Vibrating Bridge Mirror
Position. 0 O O O O I O O O O O O 0

Electronic Circuit for Photo-
multiplier Signal Amplification . .

Geometric and Optical Configuration
for the Collection of Fluorescence.

Limiting Conditions for the Fluores-
cence Observation Angles for Point
Sources Across the Sample Cell. . .

Variation of Optical Parameters as
a Function of Secondary Mask Dis-
tance from the Source of Fluores-
cence . . . . . . . . . . . . . . .

Fluorescence as a Function of
Absorbance for Quinine Sulfate in
1.0 N Sulfuric Acid . . . . . . . .

Fluorescence as a Function of
Absorbance for Increasing Amounts
of the Chromophore, 2,5—Dihydroxy-
benzoic Acid, in the Presence of

1 x 10"5 M Quinine Sulfate in 0.1
N Sulfuric Acid . . . . . . . . . .

Absorption and Fluorescence
Spectra of Flavonol in Absolute
Ethanol . . . . . . . . . . . . . .

Emission Spectrum at 77 K for
Flavonol in Absolute Ethanol. . . .

Excitation Spectrum of Flavonol
in Absolute Ethanol . . . . . . . .

viii

Page

21

23

25

27

36

37

38

42

43

72

73

77

Figure

13

14

15

16

l7

18

mng

 

M'U

 

 

 

:rao

390033

091?,

NH rrH "1H "TJH H'flJ’HIC/D

03H

Figure

13

14

15

16

17

18

19

20A

208

21A

213

22A

22B

Absorption and Fluorescence
Spectra of Flavonol in Absolute
EthanOI O O O O O O O O O O O O O O 0

Plot of Partial Quantum Ef-

ficiency (PQ) as a Function

of Wavelength for Flavonol in
Absolute Ethanol. . . . . . . . . . .

Absorption and Fluorescence
Spectra of Flavonol in Acidic
Ab801ute Ethan01 I O O O O I O O O O 0

Absorption and FluorescenCe
Spectra of the Aluminum-Flavonol
Chelates in Absolute Ethanol. . . . .

Emission Spectrum at 77 K for

the 2:1 and 6:1 Aluminum to

Flavonol Chelates in Absolute
Ethanol . . . . . . . . . . . . . . .

Energy Level Diagrams for the

2:1 and 6:1 Aluminum to Flavonol
Chelates in Absolute Ethanol at

77 K. O O O O O O O O I O O O O O O O

Spectrophotometric and Spectro-
fluorometric Titration Curves for
Aluminum (III) Titrated with
Flavonol with the Proton to Metal
Ion Ratio 5:1 . . . . . . . . . . . .

Infrared Absorption Spectrum of
Flavon01 O O C O I O O O I I O O O O 0

Infrared Absorption Spectrum of
Flavon01 O I O 0 I O O I O O O O O O 0

Infrared Absorption Spectrum of
the 2:1 Aluminum to Flavonol Chelate.

Infrared Absorption Spectrum of the
2:1 Aluminum to Flavonol Chelate. . .

Infrared Absorption Spectrum of the
6:1 Aluminum to Flavonol Chelate. . .

Infrared Absorption Spectrum of the
6:1 Aluminum to Flavonol Chelate. . .

ix

Page

78

79

81

85

87

9O

95

100

101

102

103

104

105

Figure

23A

238

MB

25

26

27

28

29

Ir
Rc-I

 

 

(It/IQ”?

IV

110

Figure

23A

23B

24A

24B

25

26

27

28

29

30

31

Infrared Absorption Spectrum of the
Residue Formed When the Solvent from
an Alcoholic Solution of Aluminum
Chloride is Evaporated. . . . . . . . .

Infrared Absorption Spectrum of the
Residue Formed When the Solvent from
an Alcoholic Solution of Aluminum
Chloride is Evaporated. . . . . . . . .

Infrared Absorption Spectrum of
Absolute Ethanol. . . . . . . . . . . .

Infrared Absorption Spectrum of
Absolute Ethanol. . . . . . . . . . . .

Typical Aluminum-27 NMR Spectrum
for the Aluminum-Flavonol Chelates
Formed in Absolute Ethanol. . . . . . .

Aluminum-27 NMR Shift as a Function
of Aluminum (III) Concentration for
Solutions with 2:1 and 6:1 Aluminum
to Flavonol Ratios in Absolute Ethanol.

Aluminum-27 NMR Shift as a Function
of the Aluminum to Flavonol Mole
Ratio in Absolute Ethanol . . . . . . .

Potentiometric Titration Curve
for Aluminum (III) Titrated with
Hydroxide Ions in Absolute Ethanol. . .

Potentiometric Titration Curves

for the Titrations of Various Aluminum
to Flavonol Ratios with Hydroxide

Ions in Absolute Ethanol. . . . . . . .

' Potentiometric Titration Curves

for the Titration of Various Aluminum
to Flavonol Ratios with Hydroxide
Ions in Absolute Ethanol. . . . . . . .

Potentiometric Titration Curves for
the Titration of Various Aluminum to
Flavonol Ratios with Hydroxide Ions

in Absolute Ethanol . . . . . . . . . .

Page

106

107

108

109

111

112

113

115

118

119

120

32

’igure

33

34A

35

36

37

38

Figure

32

33

34A

34B

35

36

37

38

39
40
41

Potentiometric Titration Curves for
the Titration of Various Aluminum to
Flavonol Ratios with Hydroxide Ions

in Absolute Ethanol . . . . . .

Potentiometric Titration Curves for
Fresh and Aged Solutions with an
Aluminum to Flavonol Ratio of 2:1

Titrated with Hydroxide Ions in
Absolute Ethanol. . . . . . . .

Various Proposed Structures for
Aluminum (III) in Absolute
Ethanol . . . . . . . . . . . .

Various Proposed Structures for
Aluminum (III) in Absolute
Ethanol . . . . . . . . . . . .

Various Proposed Structures for
the 6:1 Aluminum to Flavonol
Chelate in Absolute Ethanol . .

Various Proposed Structures for
the 2:1 Aluminum to Flavonol
Chelate in Absolute Ethanol . .

Various Proposed Structures for
the 1:1 Aluminum to Flavonol
Chelate in Absolute Ethanol . .

Proposed Structures for the

1:1 and 2:1 Aluminum to Flavonol

Chelates in Acidic Absolute
Ethanol . . . . . . . . . . . .

Program PMT Flowchart . . . . .
Program LSSQ Flowchart. . . . .

Program ARTCAL Flowchart. . . .

xi

Page

121

122

126

127

131

133

136

136
156
170
180

PART I

IMPLEMENTATION OF ABSORPTION-CORRECTED

FLUORESCENCE MEASUREMENTS

Fluore
important
fact, the
with the i

I
Consequentl
tion and t
that preciI
Only 1'
Obtain hifil
cause of tl
the meaSur
and Until
will remai
VOrk has (I
variables_
variables
tained in

For tt
ha‘le been

sourCe in S

 

I. INTRODUCTION

Fluorescence spectroscopy has become an increasingly
important analytical tool in the past few decades. In
fact, the volume of fluorescence literature has mushroomed
with the introduction of each new fluorescence probe.
Consequently, it is very important that the instrumenta-
tion and theory behind fluorescence be well understood so
that precise and accurate measurements may be made.

Only in the past few years has it become possible to
obtain high quality fluorescence spectra routinely. Be-
cause of the many instrumental and photophysical variables,
the measurement of fluorescence is a complicated process
and until these variables are eliminated or minimized, it
will remain that way. Consequently, in recent years much
work has gone into defining and eliminating many of these
variables. A list of the instrumental and photophysical
variables involved in fluorescence measurements is con-
tained in Table I.

For the most part, many of the instrumental variables
have been eliminated or minimized. The problems with
source instability were overcome by the use of a beam
splitter and second detector which monitored the source
(1-3). Unfortunately, the problem of source spectral

distribution was not remedied by this innovation. To solve

 

 

wUCOEOHDEOOZ OUCOUOOMOSHE Oumum kUQOUW .oom MOM EOHQMfiHM>

H manuE..H.

 

 

muoomwm owmouuomflca

unmflq poncuuoom
xoocHto>wuomummm

mommoooum coHDQMOmQ< muopcooom

mommoooum cofiumuomnd ahmfiflum

Eoumzm usocmom may mo hufiaflnmuw can mufiumocflq
muaaflamum new muaumocflq .muw>euemcom uouomuoo
Eoumhm Hoowumo cowmmHEm

muuofiooo HHOU

EcumMm Hwowumo cowumuwoxm

coflusnfiupmwo Houuoomw condom

sueawnmum oousom

 

Havananaouonm

HmucmfisuumsH

 

 

mucoamusmmoz mosoomouosam wuoum womoum .oom Mom moHnowum>

H OHDMB

this P
tor we
before
to a g1
wavelen
refinem
to the 1
cell. 7
intensit
tunately
had negl,

In a}
one seric
t0 monito
This fact
Spectral (
I"'aVEIEmgt}

WEIIS. TC

had to hav

  
   

3r to the
work has I:

I“fictions .

this problem partially, the beam splitter and second detec-
tor were placed after the wavelength isolation device, but
before the sample cell (4-6). These systems were refined
to a greater extent by correcting automatically for the
wavelength dependence of the beam splitter (7,8),and further
refinements led to the elimination of the beam splitter and
to the placement of the second detector behind the sample
cell. The fluorescence intensity was then ratioed to the
intensity measured by this second detector (9,10). Unfor-
tunately, this system could only be used with samples which
had negligible absorbances.

In all of the systems described previously, there is
one serious drawback. The second detector, which was used
to monitor the source, had a non-linear spectral response.
This fact precluded complete correction for the source
spectral distribution, the transfer coefficient of the
wavelength isolation device and any scattering at the cell
walls. To correct for these variables, the source detector
had to have a response which was proportional to the energy
or to the number of quanta in the incident beam. Much
work has been carried out in this area. For energy cor-
rections, bolometers and thermOpiles have been used (7,9,11),
while quantum counters have been used in quantum-corrected
instruments (8,12,13). Good results have been obtained by
the use of these devices so that excitation and source-
corrected fluorescence spectra are now routinely recorded.

Much progress has also been made in the elimination

of the emi

that almos

 

sion syste
devices wi
tum counte
fluorescen
are just n
choice is
linear res
Of the devI
rESponse.
Success (1
adIl<>cated
tiOns for
At any rat
quite ted:
Quite
Struct a
Corre-Ctiv
Push the‘
to «>11qu

‘1

 

 

of the emission instrumental variables. It seems obvious
that almost all of the instrumental variables in the emis-
sion system could be eliminated by the use of detection
devices with linear responses such as a bolometer or a quan-
tum counter. Unfortunately, due to the low levels of
fluorescence intensities normally measured, such detectors
are just not sensitive enough. The next most obvious
choice is to compare a more sensitive device with a non-
linear response, such as a photomultiplier tube, with one
of the devices mentioned above or with a detector of known
response. This has been done by several workers with good
success (14,15). In addition, several other workers have
advocated the use of a series of standard fluorophore solu—
tions for the calibration of the emission detector (16,17).
At any rate, most of the corrective procedures are either
quite tedious or complicated.

Quite recently, extensive work was undertaken to con-
struct a spectrofluorometer which incorporated all of the
corrective measures mentioned previously (18). To accom-
plish their goal, Holland §t_§l., utilized a minicomputer
to collect, manipulate and correct both absorption and
fluorescence data. The result was a fully automated and
instrumentally corrected spectrophotofluorometer. As of
this date, fully instrumentally corrected instruments are
becoming commercially available and as a result, good

quality fluorescence measurements, which are essentially

free from
routinely/II
about the I.
The pl
be divide<
cludes boi
while the
tive indi;
variables
is desira;
tion.
phOtOI
in fluore‘I
of the phI
System LISI
tensity i
fluoropho
selves as
in fluore
in mind,
been unde
The e
bathe.

purificat

 

free from instrumental variables, can now be made
routinely. On the other hand, the same cannot be said
about the corrections for the photophysical variables.

The photophysical variables, as listed in Table I, may
be divided into two subclasses. The intrinsic class in-
cludes both the primary and secondary absorption processes,
while the extrinsic class includes the effects of refrac—
tive indices and scattering. Since these photophysical
variables are not well understood by many researchers, it
is desirable at this point to formulate a general defini-
tion.

Photophysical variables are those possible error sources
in fluorescence measurements which result from the nature
of the physical phenomenon observed or from the chemical
system used. As a result, the measured fluorescence in-
tensity is not proportional to the concentration of the
fluorophore. These discrepancies usually manifest them-
selves as an apparent change in absorptivity or a change
in fluorescence quantum efficiency. With this definition
in mind, it is now desirable to examine the work which has
been undertaken to eliminate these variables.

The extrinsic variables are usually of a controllable
nature. Many times quenching can be controlled by careful
purification of the solvent and reagents as well as de-
oxygenation. On the other hand, changes in refractive

indices and scattering are not easily controlled. In

systems
scatter
ing can
measure
(19) .
rathema‘
effects
if adeqt
not the
and they
ments.
Intr
are some
ObserVed
include 1
attermate
tiOn PrOc
cence fro
ES will b
into two
Occurs in
Spectrum

spectrum.

 

Called er.
PhoreS ar

CEIICQ VP- 4

A.._

 

systems of suspended particles or macromolecules, Rayleigh
scattering can be a considerable problem. Often scatter-
ing can be minimized by the use of filters or can be
measured and manually subtracted as part of a background
(19). In cases where refractive indices are a problem,
mathematical corrections can be made (20,21). Thus, the
effects of the extrinsic variables are usually not severe
if adequate precautions are taken. Unfortunately, this is
not the case with the intrinsic photophysical variables
and they have consistently plagued fluorescence measure-
ments.

Intrinsic variables or inner-filter effects, as they
are sometimes called, are concerned with the nature of the
observed phenomenon. In the case of fluorescence, these
include the primary absorption processes, which act to
attenuate the excitation beam, and the secondary absorp-
tion processes, which act to attenuate the emitted fluores-
cence from a fluorophore. The secondary absorption process-
es will be considered first. These processes can be divided
into two categories. The first is self-absorption and it
occurs in the spectral region where the fluorescence
spectrum of a fluorophore overlaps its own absorption
spectrum. The second category includes what might be
called environmental-absorption and it occurs when chromo-
phores and possibly other fluorophores absorb the fluores-

cence which is generated by the fluorophore of primary

concer
fluore
attent
To dat
mathem.
when 56
more wc
The
measure;
fluores<
phore cc
the cage
cell, it
the abso
absorbing
cence int
Sity, mOr
Cell than
ObSerVed
Phore Co“
than 0.06
EEtOrS (2
Cf workin'
E'Jer' thi
REEL-e the

tohcentra

concern. Both categories of processes lead to inaccurate
fluorescence measurements and unfortunately, very little
attention has been given to the elimination of their effects.
To date, only one known attempt has been made to predict
mathematically the expected fluorescence intensity measured
when secondary absorption occurs (21). It is certain, that
more work in this area will follow soon.

The other major intrinsic variable in fluorescence
measurements is primary absorption. Intuitively, the
fluorescence intensity should be proportional to the fluoro-
phore concentration. Unfortunately, this is not always
the case. As the excitation beam passes through the sample
cell, its intensity is decreased exponentially because of
the absorption by the fluorophore as well as any additional
absorbing chromophores which are present. Since the fluores-
cence intensity is proportional to the excitation beam inten-
sity, more fluorescence is generated in the front of the
cell than in the back of the cell. As a result, the total
observed fluorescence is not proportional to the fluoro-
phore concentration in solutions with absorbances greater
than 0.06. This effect has been noted by several investi-
gators (22-24). To minimize it, the general recommendation
of working with dilute solutions has been advocated. How—
ever, this is not a "universal" remedy. Consider the case
where the fluorophore is initially present at a very low

concentration level but the solution absorbance is high

due to
dilutil
in a d:
in cont
sociati
um abs
large u
hand, f.
error it
rents a:
T0 Circu
and Earn
Supposed
fluoresc,
Primary 6
tion Was
in the ca
Later, Pa;

ltsaPPlic

.C-zOre .

 

due to the presence of other absorbing species. Continued
dilution of such a system would not increase the accuracy

in a determination. In addition, dilution may cause changes
in conformation, bonding, solvation and the degree of as-
sociation as well as other chemical events which may alter
the absorption-fluorescence processes and thereby, introduce
large unknown errors into the measurements. On the other
hand, failure to dilute such a system will introduce serious
error into the determination if the fluorescence measure-
ments are not corrected for the primary absorption processes.
To circumvent problems such as the one described, Parker

and Barnes (25) proposed the use of a correction factor.
Supposedly, when this factor was applied to the measured
fluorescence, the fluorescence was then corrected for the
primary absorption processes. Unfortunately, no deriva-
tion was presented and the correction factor was used only
in the case of a chromophore and a fluorophore mixture.
Later, Parker and Rees (26) extended its use by advocating
its application to systems which contained only pure fluoro-
phore. Like the first publication, no supportive experi-
mental evidence was given for this extension. As a re-
sult, many investigators are reluctant to use a correc-

tion factor which has no apparent theoretical basis, no
criteria for its proper application and no evaluation of

the resultant corrections. In an effort to rectify this

situation, several investigators have turned their

attention
fects on
geometric
partially
same resu
the deriv
a good at
the corre
Only
out by Ho
for the p
rection f
attempt h
intensiti
absorPtio
The p
how the w
all of th
correctic
dition' t
“Emerita

CQnCe mea

0‘ .
Sb .
sorptlo
‘66
.911 eval

capabilit

 

10

attention to the problem. Ohnesorge (27) examined the ef-
fects on fluorescence measurements with changes in the
geometric model presented by Parker and Barnes. Gill (28)
partially derived a correction factor which gives the

same results as the Parker and Barnes factor. Although
the derivation was not quite satisfactory, Gill did make

a good attempt at elucidating the theoretical basis of

the correction factor.

Only recently has the theoretical basis been worked
out by Holland and others (29). In addition, the criteria
for the proper application of the Parker and Barnes cor-
rection factor have been partially defined and an initial
attempt has been made to correct measured fluorescence
intensities automatically for the effects of the primary
absorption processes.

The purpose of Part I of this thesis is to report on
how the work of Holland 33 31. has been extended so that
all of the criteria for the proper application of the
correction factor have now been clearly defined. In ad-
dition, this report describes the procedures for the im-
plementation of automated absorption-corrected fluores-
cence measurements which have been developed in the course
of this study. Furthermore, the accuracy of the resultant
absorption-corrected fluorescence measurements have also
been evaluated with respect to the present instrumental

capabilities and the availability of ideal chemical systems.

To dat
cal basis
Barnes. I
tions nece
delineated
from its a]

Quite 1
tion. Holj
correction
Their deriv
Structed ar
The details
detailed de

The dev.
Corifiguratil

II . THEORETICAL

To date, very little has been known about the theoreti—
cal basis of the correction factor presented by Parker and
Barnes. In addition, the chemical and experimental condi-
tions necessary for its application have never been fully
delineated,and the accuracy of the resultant corrections
from its application never determined.

Quite recently, work was begun to remedy this situa-
tion. Holland, Teets and Timnick were able to derive a
correction factor identical to the Parker and Barnes factor.
Their derivation was based on a theoretical model con-
structed around certain chemical and experimental restraints.
The details of their work are summarized below but a more
detailed development may be found elsewhere (29).

The development of the model began with a geometrical
configuration similar to the one chosen by Parker and Barnes.
Observation of the fluorescence was to be made at 90° to
the excitation beam and a mask was to be used to minimize
scattering from the cell walls. Unlike the Parker and
Barnes configuration, the edges of the mask were to be
placed as close to the edges of the cell as possible in
order that a maximum amount of the fluorescence would be
generated within the window defined by this mask. Thus,
the basic geometry of the system was established and the

next step was to define the instrumental and chemical

11

restraints 4
Upon c01
tions or re<

include the

(l) The
int
tie
anc‘
int

qua

(2) The

the

 

(3)
the

|
the

the

 

12

restraints of the model.
Upon consideration of the problem, seven basic assump-
tions or requirements were enumerated or implied and they

include the following:

(1) The incident intensity, Io, and the transmitted
intensity, I, represented by the measured quanti-
ties R (reference) and S (sample), respectively
and the measured quantity, F, the fluorescence
intensity, are proportional to the number of

quanta, Q, involved.

R = er
S = kQS
F = k'Qf

(2) The absorption processes within the cell follow

the Beer-Lambert Law.
_ -abc
I — Ice

(3) The quanta fluoresced by the fluorophore within
the observation window are linearly related to
the quanta absorbed by that same species within

the same window.

= k"¢(AI)

F . .
Window Window

(4) The

due

of

phc

(5) A1
ger

by

(6) OM

is

cen

(7) The
die

to

The 1a8t st
was the den
uere Used an
the ake of I
Based On

13

(4) The fraction of the total absorbance which is
due to the fluorophore is equal to the fraction
of the total quanta absorbed by that fluoro-

phore.

(5) A fixed fraction of the fluoresced radiation
generated within the observation window is viewed

by a detector with uniform sensitivity.

Fmeasured = k"'Fwindow
(6) Only the fluorescence of a single fluorophore
is measured and any absorption of this fluores-
cence is negligible.
(7) The effects of scattered light, refractive in-
dices and anisotropic characteristics are assumed

to be negligible.

The last step in the development of the theoretical model
was the derivation of a correction factor. Two approaches
were used and both arrived at the same relationship. For
the sake of brevity, only one of these will be outlined.
Based on assumptions 1 and 2, it was shown that the in-
tensity of the excitation beam at any fraction of the cell

length, w, from the point of entry could be calculated

from the fol!

where Iw 15
is the inten
T is the tra

expression f

where K is a
the absorbaq
of One or mq
Citation bed
positioned a
Combination
described th

quantities I

 

14
from the following equation,

Iw = R exp(w zn'r) = RTw (1)
where Iw is the intensity at a fractional distance w, R
is the intensity of the excitation beam before entry and
T is the transmittance. From assumptions 3 and 4, the

expression for the measured fluorescence was obtained,

Af
F = K KE—x—XZ-(le - IWZ) (2)
where K is an instrumental and geometric factor, Af is
the absorbance of the fluorophore, Ac is the absorbance
of one or more chromophores, and le and Iw2 are the ex-
citation beam intensities at the mask edges which are
positioned at the fractional distances of WI and wz. The
combination of equations 1 and 2, yielded an equation which
described the measured fluorescence in terms of measurable
quantities,
F = 2.303 123-3w” - Twl) (3)
in T
At this point, the absorption-corrected fluorescence
was defined as the measured fluorescence divided by the

average intensity of the excitation beam across the obser-

vation window from w1 to w2. This definition is quite

logical sin<
of fluoresce
tion of the

for this get

BY substitm

ing the int

The absorpt

This
reduce<

 

 

15

logical since absorption-corrected fluorescence is the amount
of fluorescence that would have been measured if no absorp-
tion of the excitation beam had taken place. The expression

for this general definition was written as,

_ F
Fco - ———-—~— (4)
W

I 2 dew

 

By substituting Equation 1 into the denominator and perform-

ing the integration, the general definition yielded,

(5)
TWZ - Twl

F
Fco _ E

 

The absorption correction factor, fa, was defined as

2n T(w -w )
f= 21 (6)
a Tw2 - Twl

 

This reduced Equation 5 to the following expression,
= E
F R x f (7)

where the ratio, g, is the source-corrected fluorescence.
Note that Equation 6 is identical, except in form,
with the correction factor presented by Parker and Barnes

(25):

where A is t
x1 and x2 ar
meters from
Holland, Tee
derivation f

have also en

 

met in orde
90ne one Ste
substituting

eSCencE iS F

960 ,
metry ls

 

preSSion

 

16

2.303 A (xl-xz)

 

io‘Axl - lo'AXZ

where A is the absorbance per centimeter pathlength and

x1 and x2 are the distances of the mask edges in centi-
meters from the entry point of the cell. Consequently,
Holland, Teets and Timnick have presented a theoretical
derivation for the Parker and Barnes correction factor and
have also enumerated the basic requirements which must be
met in order for it to be valid. In addition, they have
gone one step further. By solving Equation 3 for KAf and
substituting into Equation 5, the following relationship

was obtained,

This expression shows that the absorption-corrected fluor-
escence is proportional to the fluorophore concentration

if Beer's law is followed and the observation and detection
geometry is kept constant. Consequently, the overall ex-
pression can be written as,

in T (wz-wl)

x = 2.303 K (w -w ) abc
TW2 _ TWl 2 1 f

 

w
u
de

CO

where a is the absorptivity of the fluorophore, b is the
pathlength of the cell and cf is the concentration of the

fluorophore.

Since t
only remaini
an optical
quirements
corrections ‘
constructed ,\
which met th'
corrections

chosen for t

Previously 1;

liminary tes

 

17

Since the theoretical model had been developed, the
only remaining task was to construct a fluorometer with
an optical geometry which met all of the instrumental re-
quirements of the model and in addition, could make the
corrections automatically. Once this fluorometer had been
constructed, tests could then be made with chemical systems
which met the remaining requirements. The accuracy of the
corrections could then be determined. The fluorometer
chosen for these tests was one which was similar to that
previously built by Holland, Teets and Timnick (18). Pre-
liminary tests by these workers with various fluorophores
and chromophores indicated that the effects of primary
absorption on measured fluorescence was dependent on the
observation window size and independent of wavelength and
fluorophore which was used. Consequently, they showed
that a general correction scheme was possible. Further
tests showed that they were able to make automated absorp-
tion-corrected fluorescence measurements on solutions with
absorbances up to 1.3. Unfortunately, when a second instru-
ment was constructed, the same results could not be achieved.
Consequently, this study, which was partially involved with
the construction of this second instrument, was undertaken

for the following purposes:

(1) To re-evaluate the basic assumptions in the

theoretical model.

 

(2) To
si .
(3) To

CO

EVALUATION

Upon ir
sumPtions,
the theore
Category.
the first
tics of t}
90ry inc1\

the instrl

18

(2) To make any instrumental modifications neces-
sitated by this re-evaluation.

(3) To clearly define the method by which absorption-
corrected fluorescence measurements are made.

(4) To evaluate the accuracy of the resultant cor-

rected measurements.

EVALUATION OF THE BASIC ASSUMPTIONS

Upon inspection, it becomes obvious that the basic as-
sumptions, which formed the basis for the development of
the theoretical model, fall into two categories. The first
category, which embodies all of the assumptions except for
the first and the fifth, is concerned with the characteris-
tics of the chemical systems involved. The second cate-
gory includes the two remaining assumptions and deals with
the instrumental aspects of the measurements.

Since the chemical systems for this and the previous
study were carefully chosen to meet the criteria of the
theoretical model, it must be assumed that all of the as-
sumptions in the first category, except for number four,
are valid. On the other hand, the validity of this remain-
ing assumption was rigorously proved and the details are
contained in Appendix one. In addition, another approach,
which can lead to the same result, has recently appeared
in the literature (30). In any case, it is certain that

all of the assumptions in the first category are valid if

the correct

 

ConsiderI
clusion that
warrented b1,
in this stud
does indeed,

tional to ti

 

only the val
host certair

1uisunc'iersto:
its requirei
assumption

implementat
ments . Sin
OPtical SYS
ments wi 1 1

Chapter ' W17

optical 5y:

19

the correct chemical systems are utilized.

Consideration of the second category leads to the con-
clusion that assumption one is valid. This conclusion is
warrented by the fact that the instrument, which was used
in this study and described in the Instrumental chapter,
does indeed, make intensity measurements which are propor-
tional to the number of quanta involved. At this point,
only the validity of assumption five remains questionable.
Most certainly, it is the most complex and the most easily
misunderstood of all of the assumptions. In addition,
its requirements are difficult to fulfill and consequently,
assumption five presents the largest impediment to the
implementation of absorption-corrected fluorescence measure-
ments. Since this assumption deals with the emission
optical system of the spectrophotofluorometer, its require-
ments will be delineated and evaluated in the Experimental
chapter, which deals in part with the construction of the

optical system.

OPTICAL

Figure l
photofluoror
of this stuq

Holland 3 J

 

data acquisi
is either a]
a Hanovia 2
Powerpacs M
Wt is 0011
the slits C
This nmime}

in COnjunm

1180 line

III. INSTRUMENTAL

OPTICAL

Figure 1 presents the optical diagram for the spectro-
photofluorometer which was constructed during the course
of this study. It is similar to the one constructed by
Holland gt 31. (18) and likewise, uses a minicomputer for
data acquisition, manipulation and correction. The source
is either an Illumination Industries 150 watt Xenon arc or
a Hanovia 200 watt Xenon-Mercury arc powered by an Electro
Powerpacs Model 352 variable power supply. The source out-
put is collimated by a quartz lens and is passed through
the slits of a GCA McPherson Model EU-700 Monochromator.
This monochromator utilizes a Czerny-Turner configuration
in conjunction with a 48 x 48 mm grating which is ruled at
1180 lines per mm and is blazed for 2500 A. After leaving
the monochromator, the excitation beam is directed to a
Beckman Vibrating Bridge Assembly (parts #571868 and B-28128)
which is located in the sample compartment. This assembly
splits the excitation beam in time and directs it to a quan-
tum counter. The quantum counter consists of a solution
of rhodamine B in ethylene glycol which is contained in a
5 mm quartz cell, a 6100 A sharp cut-off filter, and a
selected RCA 1P28A photomultiplier tube which is powered

by a McPherson EU-42A high voltage supply. The useful

’20

21

.pomecpo:HMOponnouvommm nonopcoonuopsnsoo asp

 

 

you sapwafio Haoapno

.« shaman

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mosomama moa<zommoozos azmzem<msoo moe<somxoozos
onmmHsm onmmHsm game ona<aHoxm
moaaaom ,
mo<¢H\
\\\
. _ m :
_ _

 

 

 

 

 

\ l
x
\

 

 

 

_/l \\ mmazaoo
Saaz<ao

 

 

 

 

 

 

 

 

mombom

 

 

 

 

 

 

 

 

 

spectral r
i.

The fl
sample cel
a device c
rotates th
perpendicu
at the sli
lens is t
monochroma
the fluore
Ga-As phot.

high VOlta‘

ELECTRICAL

The so
drives the
Pi(Jilre 2.
Electric c
aSSGmbly.
any bridge

22

spectral range of the quantum counter is from 2500 to 6000
i.

The fluorescence radiation which is generated in the
sample cell is gathered at 90° to the excitation beam by
a device called an image rotator. This optical system
rotates the fluorescence image 90° in the plane which is
perpendicular to the optical axis and presents a 1:1 image
at the slit of another McPherson monochromator. A field
lens is then used to match the image size to the emission
monochromator optics. Upon emergence from the monochromator,
the fluorescence is then detected by a Hamamatsu Model R666
Ga-As photomultiplier tube which is powered by another McPherson

high voltage power supply.

ELECTRICAL

The schematic diagram of the oscillator circuit, which
drives the Beckman Vibrating Bridge Assembly, is shown in
Figure 2. The circuit utilizes only one of the two opto-
electric chopping devices which are a part of the bridge
assembly. This opto-electric chopper is used to convert
any bridge motion into an ac signal which is then passed
through an operational amplifier which is used in a follower
configuration. The signal is then ac coupled to the base
of a power transistor which is biased above ground by adjust-
ing the 10 k9 potentiometer. As the power transistor is

switched by the movement of the bridge assembly, current is

 

>hw+

sm-+

.popaflaaomo amawum manpaupfl> one son panache .N spawns

 

 

 

 

 

 

 

 

 

 

 

 

 

 

as :momzm
nHmo isssslnllfiw <mfia
chmdm an H ENN om /
wmam onm .
+ - H
ON I "
fol! IIIIIIIIIIII L
s mummomo
J 1 III! OHmHomAm
v5." loamo
ammzfi fin."
.3 o." A
xm.a

>mHH+ >md+ >ma+ >mafi+

alternate lY

 

the bridge I
is set up a1
amplitude 0
frequency,
at the ampl
Power if
made 115V (3

powered by

 

used in a It

 

In add:
Assembly, fil
Plies a Sicl
With regpeC
nal iS used
ure 3. Th:

the eXcita.

data “(1111.

Similar
each Set 1
to pOsitio
when the e
referenCe
set determ
to adjuSt
by the use

Of the dat

i

 

24

alternately supplied to the electromagnet which controls
the bridge movement. Consequently, a feedback situation

is set up and the bridge will continue to oscillate. The
amplitude of the bridge oscillation and consequently, the
frequency, is controlled by the 10 turn 25 k9 potentiometer
at the amplifier input.

Power for the oscillator circuit is supplied by a home-
made 115V dc supply while the operational amplifier is
powered by a Deltron Model OS-lS bipolar 15V power supply
used in a master—slave series tracking mode.

In addition to driving the Beckman Vibrating Bridge
Assembly, the oscillator circuit described above also sup-
plies a signal which indicates the position of the bridge
with respect to the sample and reference cells. This sig-
nal is used by the flagging circuit which is shown in Fig—
ure 3. This circuit is used to notify the minicomputer of
the excitation beam position and in addition, it also triggers
data acquisition. The flagging circuit consists of two
similar sets of monostables. The first monostable in
each set is used as a delay and this delay time is adjusted
to position the data flag for each channel to the time
when the excitation beam is in the center of either the
reference or the sample cell. The second monostable in each
set determines the duration of each data flag and is used
to adjust visually the delay times of the first monostables
by the use of an oscilloscope. The negative going edges

of the data flags initiate data acquisition by the

.J. cam
a W... M s N a“

5r“... r J 4‘

25

pounds mwudum mcdpmunfl> on» mo sowpwcmooom sopaosoo you a“:

made
oHAEMm

mafia
mucouomom

.:0Hpfimom

 

ohao wwam .m onswam

 

 

 

2 H .LV

 

 

 

>m+ .7

u mmmmz
ET

 

 

 

 

 

ET
xum ac.

Ho.
mom

 

_ “flofi
- mmmmz
w sou

 

 

 

Hmcwam
1. $3.3

 

_

 

 

 

i
T

 

 

IQ

 

 

 

 

 

 

>m+nl.

w H a H a
me nmmmz .wlT mans o
8. Sam 8.
uom W sow

minicomputer
The amp]
and fluores
one of the
The first 5
PET input 0
tage config
Devices Mod
with gain w
Model 118A

Conjul’lctiOn

Stage is of

 

analOSJ‘-t:o-c1
Channel has
nanoampeIe .

ampere and

The Cor
thronghOUt
Corporat i0

SpaCe Of

with the

26

minicomputer.

The amplification system for both the reference-sample
and fluorescence data channels is shown in Figure 4. Only
one of the channels is shown since they are both identical.
The first stage consists of an Analog Devices Model 142B
FET input operational amplifier used in a current to vol-
tage configuration. The second stage consists of an Analog
Devices Model 183J operational amplifier used as an inverter
with gain while the third stage utilizes an Analog Devices
Model 118A Operational amplifier as a follower with gain in
conjunction with a low pass filter. The input to the third
stage is offset so that the input limits of the computer
analog-to-digital converter can be fully utilized. Each
channel has a variable gain from zero to about one volt per
nanoampere. The maximum sensitivity is about one pico-

ampere and is limited by the stability of the offset.

COMPUTER

The computer utilized in conjunction with the fluorometer
throughout the course of this study was a Digital Equipment
Corporation (DEC) LAB 8/e minicomputer with a minimum core
space of 8K. The following LAB 8/e peripherals were used
with the spectrofluorometer. The fluorometer program FLUORO
was loaded into the computer through the use of a DEC Model
TU56 DECtape transport. A DEC Model AD8-EA analog-to-digital

converter (ADC) was used for data acquisition. This ADC

.006
'.L

 

 

 

27

.:ofi9m0fiame< Hmcmflm poflaofipHSEoponm noM Pflsopflo oficoppomam .: ohzmflm

>68 n.

V

H pmmto
no. :«mo xom
oo< oIH)>>>7I 4m 3 Rm a I_|l<w>m<r| mm: H

E: v:

 

 

 

 

 

0 92m

 

 

 

 

 

 

on

 

 

v5.“ 20H

 

 

.—
J—

000.

 

is a 10 bit
a DEC Model
and the con‘
used for th
for the flu
sists of a

with a lier
display is

output rang
Model DRB-E'
Various Opel
for the red
negatiVe gq
iflitiate da
USed for t
pen drop cl
DEC Buffen

Severa
which was
(t
“he abilit

as the Son

ESPQQiallyj

I
l

 

28

is a 10 bit successive approximation converter equipped with
a DEC Model AM8-EA multiplexer. The input range is 1 1V

and the conversion rate is 50 KHz maximum. Channel 0 was
used for the reference-sample signals and channel 1 was used
for the fluorescence. The fluorometer output display con-
sists of a DEC Model VC8-E display control in conjunction
with a Hewlett-Packard Model 7044A X-Y recorder. The DEC
display is a two axis digital-to-analog converter with an
output range of : 5.12V and a resolution of 10 mV. The DEC
Model DR8-EA 12 channel digital I/O buffer was used for
various operations. Input channels 10 and 11 were used

for the reference and sample data flags, respectively. A
negative going transition for either of these flags can
initiate data acquisition. Output channels 11 and 0 were
used for the monochromator wavelength drive and the recorder
pen drop control, respectively. The pin assignments for the
DEC Buffered Digital I/O are given in Appendix Two since

they could not be found in the DEC literature.

INSTRUMENTAL IMPROVEMENTS

Several improvements have been made in the instrument
which was used in this study, when compared to the proto-
type built by Holland gt 31. (18). The first of these is
the ability to use a xenon, mercury or a xenon-mercury lamp
as the source. This improvement is a significant advantage

especially when sensitivity is needed in a fluorometric

determinat:
a mercury I
the speed (
in the new
length dri‘
computer 5
speed whil
has the ad
acquisitio
tion, data
by deactiv
the most i
multiplier
Since the
“P to abon
SYStem eff
are much e
tube has e
the flUOr
system had
of a meth}
of the Con
from that
a reflect
been used

frOm 2500

 

29

determination when the analyte absorbs in the region near

a mercury line. The second improvement is concerned with
the speed of data acquisition. Since the monochromators

in the new instrument utilize stepping motors in the wave-
length drive circuits, it is quite convenient to have the
computer step the monochromators at their maximum scanning
speed while pausing only for data acquisition. This process
has the advantage of nearly zero dead time as well as data
acquisition while the monochromators are stopped. In addi-
tion, data at one wavelength may be taken and averaged merely
by deactivating the wavelength drives. The last and perhaps
the most important improvement is the use of a Ga—As photo-
multiplier tube in place of a 1P28 for the emission detector.
Since the spectral response of these types of tubes is good
up to about 9000 A, corrections for the emission optical
system efficiency and the photomultiplier spectral response
are much easier to make. In addition, the use of the Ga-As
tube has extended the emission wavelength capabilities of
the fluorometer, and corrections for the emission detection
system have now been extended up to 7000 A through the use
of a methylene blue quantum counter (31). The construction
of the computer correction table has been somewhat modified
from that used by Holland. Instead of using smoked M90 as

a reflector, Kodak's White Reflectance Standard (32) has
been used with excellent results. The correction table

from 2500 to 6000 A was obtained by scanning the

30

monochromators in tandem with the reflectance plate in the
sample cell position. Five scans of this wavelength region
were run, and the correction factors were output on paper
tape via the teletype. Next, the rhodamine B quantum counter
was replaced with a methylene blue counter and five more
scans were made from 5500 to 7000 A. The correction factors
for this wavelength range were also output on paper tape.

The paper tapes for each quantum counter were then input

into a FORTRAN program named PMT where each set of cor-
rection factors was averaged and scaled. Mergence of the

two resulting correction tables in the region between 5500
and 6000 A, yielded the final overall correction table which
resides as part of the fluoroescence program. Besides
extending the photomultiplier correction table to 7000 A,

the foregoing procedure eliminated much of the noise en-
countered by Holland in the region around 4600 A, where xenon
has some strong emission lines. Details on the use of pro-
gram PMT and directions for implementing the photomultiplier

correction table are contained in Appendix Three.

INSTRUMENTAL ACCURACY AND PRECISION

The accuracy and precision of the instrument used in
this study is nearly the same as that which was quoted for
the prototype instrument (18). The optical resolution is
better than 2 A for high resolution work and tests with a

didymium filter showed better resolution than many

 

 

 

31

commercially available instruments. On the other hand, the
instrumental sensitivity is low due to the slow optical speeds
of the excitation and emission monochromators.

The photometric accuracy of the spectrophotometer was
found to be t 0.1% transmittance over the range from 1 to
100% T and it seems to be limited by the accuracy of the
ADC. In terms of absorbance, this error in transmittance
leads to an accuracy of1:0.001 at an absorbance of 0.10 and
i 0.04 at an absorbance of 2.0. Due to the stray light
characteristics of the monochromators used in this instru-
ment, filters may have to be used above an absorbance of
1.6 in order to maintain photometric accuracy. In addi-
tion, when fluorophore solution absorbances are near zero
or close to 2.0, fluorescence emitted along the optical path
of the excitation beam can sometimes present a serious prob-
lem although these effects on the photometric accuracy can
be eliminated through the use of cut-off filters.

Tests of this instrument as a fluorometer indicate a
relative standard deviation of less than 11% can be obtained
routinely for the measurement of fluorescence. The error
in the emission detection system correction table relative
to the quantum counters used in this instrument is about
i 1% in the wavelength region from 2500 to 6800 A. The
precision in quantum efficiency measurements is better than
t 1% and the accuracy over a two order of magnitude concen-
tration range of quinine sulfate has been found to be about

i 5%. This last figure is also probably a good estimate

32

of the fluorometric accuracy of the instrument although
this specification is difficult to determine. Tests have
also shown that absorption-corrected fluorescence measure-

ments are accurate to i 3% up to a solution absorbance of

2.0.

 

 

 

('1

'4.

IV. EXPERIMENTAL

As noted in the theoretical chapter of this thesis, the
fifth assumption for the theoretical model, which was de-
veloped for absorption-corrected fluorescence measurements,
is concerned with the emission detection system and it em-
bodies some of the most stringent requirements. In addition,
the fulfillment of these requirements is of utmost impor-
tance if absorption-corrected measurements are to be made
accurately. Therefore, each requirement of assumption five
was carefully delineated, and the emission detection system
constructed in such a way that this assumption was valid
for the spectrophotofluorometer used in this study. Once
this goal had been achieved, a correction scheme was de-
veloped and the accuracy of the resultant corrections

determined.

EMISSION OPTICAL SYSTEM

Briefly stated, the model assumes that a fixed fraction
of the fluorescence generated within the observation window
is viewed by a detector with uniform sensitivity. In rela-
tion to the emission optical system, this assumption breaks

down into the four following requirements:

(1) The excitation beam must be homogeneous.

(2) The fluorescence generated in a horizontal slice

33

 

 

34

within the observation window must reach the
detector.

(3) The observation angles for all point sources
along that slice must be equal.

(4) The detector must have a uniform sensitivity over
its sensing area which is normally illuminated by

the fluorescence.

The first requirement is satisfied if the excitation
beam is well collimated. Since Beer's law also has this
requirement, the excitation system of the fluorometer used
in this study was constructed so that this requirement was
met. The second requirement of the model involved a modi-
fication of the optical systems normally found in fluorom-
eters. Since the slit height of the emission monochromator
was larger than the horizontal dimension of the fluores-
cence observation window, either of two courses of action
could have been followed. The first involves the 90° rota-
tion of the monochromator so that its slits are in a hori-
zontal position while the second involves the 90° rotation
of the cell fluorescence image through the use of a dove
prism or a set of front surface mirrors. Based on the.in-
convenience of monochromator rotation and the high cost of
a quartz dove prism, front surface mirrors were used in
this study. Furthermore, the utilization of mirrors
allowed the use of conventional optical configurations.

The configuration, which was chosen for this study and

35

is shown in Figure 5, consisted of a focusing lens positioned
to give a 1:1 image ratio at the entrance slit and a field
lens which was compatible with the optical speed of the
emission monochromator. Note that with this optical system,
the mask edges at x1 and x2 in conjunction with the mono-
chromator slit width define a horizontal slice within the
fluorescence observation window which is viewed by the
detector. Consequently, the second requirement has been

met.

The fulfillment of the third requirement, which is con-
cerned with the observation angles, can never be totally
realized; only closely approached. To help meet this
requirement, a second mask with the same window dimensions
as the primary mask was placed at a distance, d, from the
source of fluorescence as shown in Figure 6. Preliminary
investigation showed that for a constant distance, d, the
observation angles ranged from a maximum at the center of
the cell to a minimum at either edge of the observation
vaindow. Consequently, these two limiting angles, 8 and
h, respectively, were used as a measure of uniformity in
the observation angles across the fluorescence window.
Their differences for various secondary mask distances were
calculated and expressed as relative percent error. The
results of these calculations are presented in Figure 7A.
Curve B shows the variation in the observation angle at
Cell center, 8, with d. Points for.these curves were cal-

culated for a window width of 0.80 cm..

36

.ooCoououosa mo
coavooaaoo on» you :oavmuswfimc°o Huowpno cam awhpoWOou .m unawam

 

 

 

moEeom
mzmj 3m; meals.
tom 2 \
”10228100202 _7 ’

 

 

 

 

 

__\ I

m2”: @2588 AV

 

 

 

 

 

 

 

, ,

v5.52 xmds.
>m<DZOUmm >m<2Ea

37

 

 

 

 

 

 

 

 

PRIMARY SECONDARY
MASK MASK
A.
—<: 6? VV
W l d >
EXCITATION
BEAM
PRIMARY SECONDARY
MASK [MASK
B.

 

 

 

 

 

 

 

 

/\ I d
EXCITATION
BEAM

\/

 

A. Maximum Observation Angle
8. Minimum Observation Angle
Figure 6. Limiting Conditions for the Fluorescence 0b-

servation Angles for Point Sources Across the
Sample Cell.

38

 

 

 

 

30 i I I I r
~e .
E); w O.8cm
._ 25... -.
X
on
\
ls 20~ J
S
a“
t I5r "
0
E
Q
2
3, lO- 4
Q B
.2
0 — -4
;§ 5
.A
O 1 l l J 1
O I 2 3 4 5 6

Location of the secondary mosk(d). cm

90’

75°

6C?

45°

30'9

|5°

00

A. Relative Percent Error in the Observation Angles

B. Observation Angle at Cell Center

Figure 7. Variation of Optical Parameters as a Func-
tion of Secondary Mask Distance from the

Source of Fluorescence.

 

 

 

t}

id

tht

del

Wit

M

urea

and

uh

39

Note that as the error between 8 and o approaches zero,
the observation angles go to zero. In other words, the
ideal situation would involve the collection of only the
radiation which is parallel to the Optical axis. Unfor-
tunately, this would require a large secondary mask dis-
tance and would drastically reduce the radiation collection
efficiency. Consequently, a mask distance of 4.0 cm was
chosen as a compromise. At this distance, the error in
the observation angles across the window is less than one
percent while the angles are still large enough to permit
the passage of a reasonable amount of fluorescence to the
detector.

The last requirement is concerned with the uniformity
of the detector sensitivity. The fulfillment of this
requirement by the photomultiplier tube used in this study
was tested in the following way. The fluorescence from a
series of quinine sulfate solutions was measured with and
without a diffuser in front of the photomultiplier. Ob-
viously, the fluorescence intensities of these two sets of
measurements were not the same because of the absorption
and scatter introduced by the frosted quartz but when the
data were normalized, the shapes of the fluorescence-concen-
tration curves were identical. From this result, it was
concluded that the sensitivity of the photomultiplier was
uniform throughout the cathode area normally illuminated by
fluorescence radiation. Consequently, the last requirement

has been met and the model developed for absorption-corrected

 

 

 

Th:

aCit

0‘15 1

40
fluorescence measurements could then be tested.

ABSORPTION-CORRECTED FLUORESCENCE

The masks for the fluorescence collection system were
constructed so that the observation windows were square and
had a dimension, w, of 0.80 cm. The masks were then posi-
tioned in front of the cell so that x1 and x2 were 0.10 and
0.90 cm, respectively. Obviously, w1 and w2 should have
the same values except that they are unitless. A 13" quartz
biconvex lens with a focal length of 2" was used as the
focusing lens while a l" quartz plano-convex lens with a 3"
focal length was used as the field stop. The relationship
between the focal lengths of the lenses and the monochromator

is given by the general lens formula,

1 + 1 l
ZII'I')focusing TI'I')monochrom. (f'l')field

 

The front surface mirrors were obtained from Edmund Scien-
tific Company (#591) and were resurfaced with aluminum and
a soft magnesium fluoride overcoat.

Once the optical system had been constructed and posi-
thoned, experimental fluorescence curves were obtained
for two series of solutions. The first series was composed
of a number of quinine sulfate solutions in 1.0 N sulfuric
aCid. These solutions ranged in absorbance from 0.007 to

over 2.0. The second was a series of solutions with a

 

 

cons
acid
the
tion
Fig:
flu<
shai
and
for

val

COT

ti:

pa

P'fl

In

41

constant quinine sulfate concentration in 0.1 N sulfuric
acid but, in addition, had an increasing concentration of
the chromophore, 2,5-dihydroxybenzoic acid. The first solu-
tion in this series contains only the pure fluorophore.
Figures 8A and 9A show the experimental results for the pure
fluorophore and the mixture, respectively. Since the

shapes of these experimental curves are dependent on W1

and w2, these curves were used to calculate the best values
for these window parameters. In addition, the calculated
values of w1 and w2 should be in close agreement with the
measured values if the experimental emission optical system
conforms to the model which was developed for making absorp-
tion-corrected fluorescence measurements.

In order to determine the best values for the window
parameters, w1 and wz, the absorption-corrected fluores-
cence for each experimental point for the pure fluorophore
solution series was calculated by the following process.
Since the absorption-correction factors, fa (Equation 6,
Page 15), for the fluorophore solutions which have absor-
bances below 0.06, are close to unity and almost independent
Of small variations in wl and wz, the absorption-corrected
fluorescence intensities for all of the pure fluorophore
Solutions under this absorbance limit were calculated by
the use of Equation 5 and the measured observation window
Parameters of 0.10 and 0.90. A least-squares fitting
lToutine was then used to determine the equation of the

best straight line through these theoretical points. Once

FLUORESCENCE

42

 

 

 

 

'01leL 1 1. L1 Lilli 1 Pl llLJll L

i r

3 A. Source-Corrected Fluorescence :

1 ,.

- B. Absorption-Corrected Fluorescence .

1 .

LO: :

q t

. r

i L.

. .

i e

/’(

OJ: :

I i

- t

. L

i _

1 r-
.O' 111' I T l IIIIUI I I Y I IIITU f
.0! OJ If) 2!)

ABSORBANCE
Figure 8. Fluorescence as a Function of Absorbance

for Quinine Sulfate in 1.0 N Sulfuric Acid.

43

.oao< oaksmasm
z “.0 c“ opmmasm oCHCHza s mica x a mo oesomohm on» ca
.owo< oaoucophxonohnuotm.m ouonooeouno on» no @925084

 

wcwmmonocH you oocmpuomp< no magpocsm o m< cocoononosam .m seawam
moz<mmomm<
ON 0.. md Nd _.o
h p P p p _ F _ . F _.O
44
cocoomonosam concounooncofipouomn< .m Mm
. -Nd
oocoomouosam concoupoonoousom .< mw
n:
l . S
«a
3
L i N
«J
1 -mxu n:
l v

 

 

 

 

 

a -Ifiq[~ q u |q| _ J1

de1

SO:

in

 

hi

0!

al

NH;

 

44

determined, this equation was used to calculate the ab-
sorption-corrected fluorescence intensities for the remain-
ing solutions. All of the calculational processes mentioned
above were performed through the use of a FORTRAN program
named LSSQ. The use of this program is outlined in Appen-
dix Four.

The absorption-corrected fluorescence intensities at
each experimental point for the mixture solution series
were calculated from the first experimental point by the
use of Equation 5 and the measured window parameters.
Obviously, all of absorption-corrected fluorescence inten-
sities are identical since the concentration of the fluoro-
phore was kept constant throughout the solution series.

This intensity value is also utilized by program LSSQ.

The paper tape outputs from LSSQ are presented in a
format which is compatible with another FORTRAN program
named RTFACT. Program RTFACT is a program which is based
on a modified Simplex optimization method and it is similar
to the ones outlined by Deming and Morgan (33). The program
accepts data from LSSQ in the form of transmittance, measured
fluorescence and absorption-corrected fluorescence for each
experimental point. It then uses a modified form of Equa-
tion 5 to fit the measured fluorescence to the calculated
corrected fluorescence. This is accomplished by varying w1
and w. The observation window width, w, was used because w1
and w are independent parameters whereas w1 and w2 are not.

The window parameters determined from the experimental

 

 

and
0.9

mea

AU'I

abs
as

110

pa:
Th:

fl‘

45

and calculated fluorescence data were found to be 0.11 and
0.91. These values are in excellent agreement with the

measured values of 0.10 and 0.90.

AUTOMATED CORRECTIONS

In order to eliminate the problem of tedious manual
absorption corrections, the spectrophotofluorometer program,
as initially outlined in (18), has been modified by Teets
to include a subroutine which automatically calculates the
absorption-correction factor, fa' from the observation window
parameters and the transmittance at each experimental point.
The program then applies these factors to the source-corrected
fluorescence, F/R, to yield the desired absorption-corrected
fluorescence. The subroutine uses a close approximation of
Equation 6 to calculate the correction factors. The de-
nominator of this equation, (Twz-Twl), was replaced with
the first seven terms of its Taylor expansion. The expansion

yielded the following expression:

2 2 3 3

w - w w - w

I _ -
2.(w2 w 3!(w2 wl)

TWZ-Twl = 1 + in T + ...

1)
Since the quantities in brackets depend only on wl and w2,

these quantities are constant and are calculated by the use
of another FORTRAN program named ARTCAL. This program uses
wl_and w2 to calculate the second through seventh bracketed

terms in the Taylor expansion and then converts them into

46

a binary floating point format which can be directly sub-
stituted in the subroutine which is contained in the PAL8
spectrophotofluorometer program.

The procedure outlined above was followed. The calcu-
lated window parameters of 0.11 and 0.91 were used and the
output from ARTCAL was substituted into the subroutine.

The automated application of the correction factors to the
experimental data yielded absorption-corrected fluorescence
curves similar to the ones shown in Figures 8B and 9B.

Details for the implementation of automated absorption-
corrected fluorescence measurements as well as the listings
and descriptions of the various FORTRAN programs mentioned

above are contained in Appendix Four.

V. RESULTS AND DISCUSSION

BASIC ASSUMPTIONS

Of the seven basic assumptions which pertain to the
theoretical model for absorption-corrected fluorescence
measurements, only two needed a closer examination. The
first of these was concerned with the relationship between
the fractional absorbance of the fluorophore and its frac-
tional absorption of quanta. This relationship, through

rigorous mathematical treatment, was shown to be,

The importance of this relationship is paramount since it
was used to establish that absorption-corrected fluores-
cence measurements can be made on solutions which contain
a fluorophore and several absorbing chromophores as well as
solutions which contain only pure fluorophore. Consequently,
Gill's question (28) about the application of the Parker
and Barnes correction factor to both the case of a pure
fluorophore and the case of a fluorophore and chromophore
mixture has been answered.

The second and last assumption examined in this study
was concerned with the observation of the fluorescence

generated within the sample cell. It seems almost obvious

47

at

t8!

ab'

 

Si

0!

de

To .

...

 

48

at this point that unless the measured fluorescence in-
tensity is predictable, there would be no way to make
absorption-corrected measurements. In other words, the
measured fluorescence intensity must be predicted by Equation
3, which is written as

KAfR

= 2.303 —— (Twz - Twl)

Fmeasured
in T

Since adherence to this equation depends almost entirely
on the geometric and optical configuration of the emission
detection system, assumption five as listed in the Theo-
retical chapter was examined with respect to the detection
system. As a result, this assumption was broken down into

the four following requirements:

(1) The excitation beam must be homogeneous.

(2) The fluorescence generated in a horizontal slice
within the observation window must reach the
detector.

(3) The observation angles for all of the point
sources along that slice must be equal.

(4) The detector must have a uniform sensitivity over
its sensing area which is normally illuminated

by the fluorescence.

In essence, the fulfillment of these requirements not only

predicts Equation 3 but it also outlines the instrumental

49

conditions for the implementation of absorption-corrected

fluorescence measurements.

IMPLEMENTAT ION

As noted in the preceeding chapter, the emission
optical system was designed and constructed so that the
requirements listed above were met. Whether it actually
does meet them is a point which has to be shown.

Equation 3 indicates that if all of the assumptions
are met, the measured fluorescence intensity is dependent
on the Optical system parameters, w1 and w2. Consequently,
if these parameters are calculated from the measured source-
corrected fluorescence intensities and corresponding trans-
mittances, their values should agree with the values of wl
and w2 which were obtained by the actual measurement of
these parameters when the optical system was constructed.
If good agreement exists, it may be assumed that the
optical system meets the requirements. Since excellent
agreement has been shown to exist between the actual and
calculated window parameters, it must be assumed that the
emission optical system, as outlined in the Experimental
chapter, meets the criteria of assumption five and sub-
sequently, corrections have been made on solutions with

total absorbances as high as 2.0.

CO

 

in

I‘d

ma

 

50

CORRECTION ACCURACY

In theory, Equation 5 is an exact expression. However,
in practice, the accuracy of the results obtained from its
application is limited by certain instrumental factors.
Since absorption-corrected fluorescence measurements are
dependent on R, S and F, these measurements are limited by
the accuracy of these three measured quantities. Evalua-
tion of the measurement capabilities of the instrument used
in this study should lead to a good estimate of the accu-
racy to which absorption-corrected measurements can be
made.

As noted in the Instrumental chapter, the photometric
accuracy of the instrument used in this study is r 0.1%

T. Because this error is most prominent near zero trans-
mittance, it is expected to have the greatest effect on the
absorption-corrected measurements in this region. Indeed
this is the case, the maximum predicted error in the cor-
rection factor, fa, at an absorbance of 0.50 is .14% where-
as it is 2.8% for an absorbance of 2.0. This error increases
to 7.1% for an absorbance of 2.5. Obviously, the greatest
error occurs in the transmittance region where the correc-
tion factors are increasing at a rapid rate. Consequently,
based on the photometric accuracy of the instrument, an
absorbance of 2.0 was chosen as a limiting value to ensure
a reasonable accuracy in the absorption-corrected measure-

ments.

51

Based on the foregoing discussion, it is expected that
the correction factors should have a relative error of 2.8%
or less if only the photometric accuracy of the instrument
is considered. However, the accuracy of the absorption-
corrected fluorescence measurements should be somewhat less
since these measurements are also dependent on the accuracy
of the source-corrected fluorescence, F/R. The fluorometric
accuracy of a particular instrument is a difficult quantity
to evaluate. Usually, the source-corrected fluorescence is
ratioed to the quanta absorbed within the fluorescence
observation window. This is done for a series of solutions
which contain a fluorophore with a constant quantum efficiency
throughout the concentration range chosen. Quinine sulfate
in sulfuric acid is usually used as a standard since its
quantum efficiency is believed to be constant to within
l-2%. Because the calculated ratios are directly propor-
tional to the instrumental constant, the agreement among
them over the whole series of solutions is considered to
be a measure of the instrumental fluorometric accuracy.
Unfortunately, this method must utilize the transmittance
of each solution to calculate the number of quanta absorbed
within the observation window (Equation 1). Therefore, the
ratios are somewhat dependent on the photometric accuracy
of the instrument, and the values obtained by this method
for the fluorometric accuracy should be used only as esti-
mates. For the instrument used in this study, the fluoro-

metric accuracy was found to be about i5%. Based on

ex;

at

in

E18

it

 

VF].
.ou

fo

si

li

WE

52

experimental data, as expected, the error becomes greater
at low fluorescence levels. Fortunately, the worst errors
in fluorometric accuracy occur in a region where the photo-
metric accuracy is the best and vice versa. Consequently,
it is expected that the absorption-corrected fluorescence
measurements should have a maximum relative error of less
than 15%. Of course, this figure does not take into account
for the errors incurred because of cell positioning, but
since this source of error can be minimized by the use of

a tight fitting cell chamber, it need not be considered

for the purposes of this discussion.

Based on the experimental results, the error from
linearity for the absorption-corrected fluorescence curves
was found to be a little greater than 13%. This figure is
in good agreement with the predicted maximum value of 15%
and it indicates that accurate absorption-corrected mea-

surements are possible.

 

0E
Tc
0‘
0n
he

 

VI. SUMMARY AND CONCLUSIONS

A second computer-centered spectrophotofluorometer
has been constructed. Improvements over the prototype
instrument include the capability of using several dif-
ferent sources, computer controlled wavelength scans, the
use of a reflectance standard to calibrate the emission
detection system and the utilization of a methylene blue
quantum counter to extend this calibration to 7000 A.

The instrument was found to have a photometric accuracy
of about i0.1% T in the range from 1 to 100 % T. The
fluorometric accuracy was estimated to be about 15%, while
the fluorometric precision was found to be il%.

The instrument was used to extend some of the previous
work on absorption-corrected fluorescence measurements.
Some of the assumptions previously used in the formula-
tion of a theoretical model were carefully examined. This
examination resulted in the construction of an emission
Optical system which met the criteria of the theoretical
model. A scheme was then developed for the implementation
of absorption-corrected fluorescence measurements based
on this model. Subsequently, absorption-corrected measure-
ments were made up to an absorbance of 2.0 and these
measurements were found to be accurate to about i3%. With-
out corrections, the error due to non-linearity would be

about -8S% for solutions with absorbances of 2.0.

53

PART II
FLUOROMETRIC AND OTHER STUDIES OF THE
REACTION OF ALUMINUM (III) AND

FLAVONOL IN ABSOLUTE ETHANOL

54

 

Ch,
bec

fla

I. INTRODUCTION

For years flavones have been known to exist in nature.
Generally, they are found in plants as the polyhydroxy or
polymethoxy derivatives of the parent flavone molecule.

In recent years, however, many other flavones have been
synthesized so that today well over fifty derivatives exist.
The structure of the parent molecule, flavone, reveals the

reason for such a myriad of derivatives. Of all the

 

various derivatives, those which contain hydroxy groups
at the 3 and 5 positions are the most interesting. It is
these derivatives which have the possibility of forming
chelates with metal ions and many of these chelates have
become analytically useful. As a result, the volume of

flavone literature has grown immensely in the last decade.

55

56

Fortunately, there are several excellent reviews available
in the literature and in all,they list well over two hundred
references (34-36). In addition, they also provide an ex-
tensive list of the chemical elements which have been
determined through the use of flavone derivatives.

Of all the hydroxyflavones , quercetin (3 , 5 , 7 , 3 ' , 4 ' -
pentahydroxyflavone), quercetrin (3-g1ucoside of quercetin),
rutin (5,7,3',4'-tetrahydroxy-3-rutinoside flavone) and
morin (3,5,7,2',4'-pentahydroxyf1avone) are probably the
most important. With these four flavones, over thirty
chemical elements may be determined and many of the deter—
minations have detection limits which are well below the
part per million level.

Besides elemental determinations, hydroxyflavone-metal
chelates have also been used for structural determinations
of the flavones themselves. Mixtures of boric acid and
citric or oxalic acid have been used to identify hydroxy-
flavones which have hydroxy groups in the 3 or 5 position (37-
41). Rutin and quercetin have been determined from their
complex formations with ZrOC12, AlCl3 and TiCl4 (42). In
addition, many other methods for the determination of
flavones have been developed, and these may be found in the
reviews previously referenced.

One of the simplest flavone derivatives is 3-hydroxy-
flavone or flavonol. Unlike many of the other flavones
which are used in the formation of chelates, flavonol has

only one chelation site. Consequently, it is expected

 

fc

Ch

th

thi

 

57

that the metallic chelates of flavonol should have the
structure shown below. Chelates of this type have been

used to determine quantitatively iron (43), tungsten (44,45),

tin (46), uranium (47,48), vanadium (49,50) and zirconium
(51). Flavonol also forms chelates with many other metal
ions, but they have not generally been used for quantita-
tive determinations.

As a whole, flavonol chelate structures in solution
are not well understood. Perhaps the most interesting
of these structures are those of the chelates which are
formed with aluminum (III). The first report of these
chelates was made by Jurd and Geissman (52) who measured
their absorption spectrum and proposed the structure shown
above. Later work by Urbach and Timnick (53,54) established

that a 1:1 chelate was formed in basic solutions while in

58

neutral solutions of absolute ethanol, three different
chelates were formed. Two of the chelates were found by
photometric means to have stoichiometries of 1:1 and 2:1
metal ions per ligand. The third chelate, which could only
be detected fluorometrically, had a metal to ligand ratio
of 6:1. In addition, from the data which were obtained
from the potentiometric titrations of the chelates and from
the structure of aluminum (III) in absolute ethanol, which
was proposed by Ohnesorge (55), these workers were able

to write an overall reaction equation for the formation of

the 2:1 chelate:

[A12(0Et)5(stoh)n]+ + 5H+ + FlOH +

[A12(OEt)3(F10)(EtOH)]2+ + 4H+ + 2EtOH

where FlOH represents flavonol. Unfortunately, no attempt
was made to explain why the spectra for the chelates were
all identical and how a bidentate ligand forms a chelate
with six aluminum ions.

In addition to work by Urbach and Timnick, Porter and
Markham (56) have attempted to elucidate the structures
and stoichiometries of the flavonol-aluminum (III) chelates
which are formed in absolute methanol. They have found
that a 2:1 metal-to-ligand chelate was formed in a 0.1 M
methanolic perchloric acid solution. They also noted that

the chelate fluoresced at 450 nm. However, when no acid

59

or base was added to the flavonol-aluminum (III) solutions,
the 1:2 and 1:1 chelates were formed. Upon the addition

of 0.1 M potassium acetate, two chelates were also formed.
These were the 4:1 and 1:2 metal-to-ligand chelates. As
noted by Urbach and Timnick and again by Porter and Markham,
the presence of excess base severely inhibits chelate forma-
tion. Consequently, when Porter and Markham added 0.1 M
sodium methoxide to the flavonol-aluminum (III) solutions,
they found that no chelation had taken place. However, when
equimolar amounts of methoxide or hydroxide and flavonol
were used, the 1:2 metal-to-ligand chelate was formed. In
addition, when the ratio of hydroxide to flavonol was 2.5,
they found that a 1:1 chelate was the predominant species.
Finally, Porter and Markham also noted that at slightly
higher hydroxide or methoxide ratios, the 2:1 and 3:1 metal-
to-ligand chelates were formed. In all cases, where either
hydroxide or methoxide was present, the solutions did not
fluoresce.

At this point, it should be fairly obvious that the
systems of chelates formed between flavonol and aluminum
(III) are quite complex. As a result, the data which were
obtained in the two studies outlined above have raised many

questions:

1. Why are the absorption and fluorescence spectra
identical for all of the chelates?

2. Are the chelates structurally similar?

60

3. Does a 6:1 chelate actually exist in solution and
if so, why?

4. What are the reasons for the vast differences in
quantum efficiencies among the chelates?

5. Why do the chelates formed in methanol differ in

stoichiometry from those formed in ethanol?

In an effort to answer at least some of the foregoing ques—
tions, a study was undertaken in order to re-examine the
formation of flavonol-aluminum (III) chelates in absolute
ethanol, and as a result, some new data have been added.
This part of this thesis is a report of that re-examination,
and it should serve as a stepping stone to a fuller under-
standing of the structure of aluminum (III) in solution as
well as the chelates which are formed upon the addition

of flavonol.

II. EXPERIMENTAL

INSTRUMENTATION

The following instruments were employed to make the

appropriate measurements:

Beckman Expanded Scale pH meter equipped with a Sargent
combination glass—saturated calomel electrode (#8-
30070-10) was used to measure the apparent pH of the
ethanolic chelate solutions.

Perkin Elmer Model 237B Grating Infrared Spectropho-
tometer was used to obtain the infrared absorption
spectra.

Spex Ramalog 4 Raman Spectrometer equipped with an
argon ion laser was used to obtain the Raman spectra.
A double monochromator system was used for this study.

A highly modified NMRS-MP-lOOO Nuclear Magnetic Reson-
ance Spectrometer, equipped with a General Radio Model
1164-A Frequency Synthesizer, a Fabri-Tek Model 1074
Multichannel Analyzer and a super conducting magnet,
was used to obtain the a1uminum~27 NMR spectra.

Aminco-Bowman Spectrophotofluorometer (SPF) equipped
with a rotating cam accessory and a liquid nitrogen
cooling system was used to obtain the delayed fluores-
cence and phosphorescence spectra.

Computer-centered spectrofluorometer, as outlined in

the Instrumental chapter in Part I of this thesis, was
used to measure the absorption and fluorescence spectra
as well as the quantum efficiencies of the chelate solu-
tions.

Picker Copper X-ray Source and Powder Camera were used
to obtain the X-ray powder diffraction patterns.

61

62

CHEMICALS

The following chemicals were used without further puri-
fication as standards or in the preparation of reagent solu-
tions described in this study:

Aluminum Chloride Anhydrous, Reagent Grade

Matheson, Coleman and Bell

Aluminum Chloride Reagent Grade
Allied Chemical Company

Ammonium Chloride Baker's Analyzed Reagent
J. T. Baker Chemical Co.

Ammonium Hydroxide Reagent Grade
Fisher Scientific Company

Calcium Oxide Baker's Analyzed Reagent
J. T. Baker Chemical Co.

2,5-Dihydroxybenzoic acid 99% Purity
Aldrich Chemical Company
Ethanol, absolute Commercial Solvents Corp.
Ethylenediaminetetracetic Baker's Analyzed Reagent
acid, disodium salt, J. T. Baker Chemical Co.
dihydrate
Hydrochloric Acid Analytical Reagent
Mallinckrodt, Inc.
Nitric Acid Analytical Reagent
Mallinckrodt, Inc.
Potassium Hydrogen Reagent Grade (ACS)
Phthalate Matheson, Coleman & Bell
Potassium Permanganate Baker's Analyzed Reagent

J. T. Baker Chemical Co.

Quinine Sulfate, Baker Grade
dihydrate J. T. Baker Chemical Co.
Silver Nitrate Baker's Analyzed Reagent

J. T. Baker Chemical Co.

63

Sodium Hydroxide Analytical Reagent
Matheson, Coleman & Bell

Sulfuric Acid Baker's Analyzed Reagent
J. T. Baker Chemical Co.

2-Thiouracil 97% Purity
Aldrich Chemical Company

Zinc Sulfate Reagent Grade
Allied Chemical Company

PURIFICATION OF FLAVONOL

Flavonol obtained from the Eastman Kodak Company was
purified by the following process. The impure flavonol
(yellow) was dissolved in a minimum amount of hot absolute
ethanol. About ten milliliters of dilute sulfuric acid
was added to this solution. The solution immediately turned
from a dark yellow to a very pale yellow. Water was added
to the hot ethanolic solution until obvious crystal forma-
tion had started. The solution was left to stand until it
had cooled to room temperature. The solution and crystals
were then cooled in an ice bath until complete crystalliza-
tion had taken place. The crystals were filtered and then
washed with cold water. The melting point range of the
fine pale yellow needle-like crystals was 167.8 - 169.7 C.
The literature melting point range is given as 168 - 170
C. The ultraviolet absorption spectrum of the product

matched that of flavonol.

64

PREPARATION OF STOCK SOLUTIONS

A standard aqueous solution of ethylenediaminetetra-
acetic acid was prepared by the direct weighing of the di-
sodium salt, dihydrate.

This standard EDTA solution was then used to standardize
an aqueous solution of zinc sulfate. Erichrome Black T was
used as the indicator.

Aluminum solutions were prepared by dissolving anhydrous
aluminum chloride in absolute ethanol. The dissolutions
were carried out in a dry ice and acetone bath to prevent
the formation of impurities due to the reactive nature of
the anhydrous salt. The solutions were then made up to be
about 0.01 M in aluminum chloride. Aliquots of these solu-
tions were taken to dryness on a hot plate and the solid
was dissolved in a small amount of concentrated hydrochloric
acid. Upon dilution, a small excess of the standard EDTA
solution was added, and the pH of each solution was adjusted
to between 7 and 8. A few drOps of Eriochrome Black T
were added, and the excess EDTA was immediately titrated
with the standardized zinc sulfate solution. The endpoint
was reached when the solutions turned from a pure blue to
a wine red color. The concentrations of the aluminum
chloride solutions were calculated, and the 0.01 M solutions
were diluted to make 1.00 x 10"3 M stock solutions. Various
concentrated aluminum chloride solutions were standardized

in a similar manner, and these solutions were used when

65

high concentrations of aluminum were required.

Standard ethanolic stock solutions of flavonol were
prepared by the direct weighing of the solid.

Stock solutions of sodium hydroxide in absolute ethanol
were prepared by dissolving the pellets in water to yield a
50% solution. This solution was allowed to stand for three
daysixJallow any carbonate to settle out. Three milli-
liters of this solution were then diluted to 500 ml with
boiled absolute ethanol. This ethanolic solution was then
standardized against potassium hydrogen phthalate. This
solution was then diluted further to obtain 1.00 x lo'3
M stock sodium hydroxide solutions.

Stock solutions of hydrochloric acid were prepared by
diluting five milliliters of 37% hydrochloric acid to 500
ml with absolute ethanol. This solution was standardized
against a standard solution of sodium hydroxide and diluted

further to obtain the 1.00 x 10'3 M stock solutions.

EXPERIMENTAL PROCEDURES

Spectrofluorometer

 

All of the absorption and fluorescence spectra were
measured on the computer-centered spectrofluorometer. These
measurements were made with a 150 watt Xenon arc lamp as
the source and a 4.0 nm bandpass for both the excitation

and emission monochromators. All spectra were background

66

corrected, and in addition, the emission spectra were cor-
rected for all instrumental variables.

The quantum efficiency measurements were made with the
same monochromator bandpasses and were relative to quinine
sulfate. The instrument was calibrated by the use of either

4 or 5.00 x 10"5

a 1.00 x 10- M quinine sulfate solution in
1.0 N sulfuric acid. The calibration was carried out at

an excitation wavelength of 365.3 nm, and the emission
monochromator was scanned from 370 to 620 nm. The generally
accepted quantum efficiency of 0.546 for quinine sulfate in
sulfuric acid was used. All quantum efficiency measure-
ments on the unknown chelate solutions were made at an
excitation wavelength of 404.6 nm. The emission wavelength

range was changed slightly from the range used for the

standard in order to eliminate any Rayleigh scattering.

Photometric and Fluorometric Titrations

 

Aliquots of the titrant and the reagent to be titrated
were placed in a series of volumetric flasks. In all cases,
flavonol was added to the aluminum aliquot. Acid or base
was then added, if necessary, and the contents of the flasks
were diluted to volume with ethanol. After a waiting period
of at least three hours, the appropriate measurements were
made. No corrections for dilution were necessary with this

procedure.

67

Apparent pH Measurements

 

The apparent pH measurements were made by calibrating
the meter at a pH of 7.0 with the use of an aqueous buffer.
The electrode was then allowed to equilibrate for thirty
minutes in absolute ethanol before any test measurements
were made. The test solutions were stirred between read-

ings and no temperature control was employed.

Emission Measurements

 

All delayed fluorescence and phosphorescence measure-
ments were made on samples contained in 5 mm quartz tubes.
The samples were deoxygenated by the bubbling of nitrogen
through the solutions for at least 15 minutes before being
frozen to 77 K. All the spectra were scanned on the Aminco-
Bowman Spectrophotofluorometer which used a 150 watt Xenon
arc lamp as the source and had a bandpass of 20 nm for both
monochromators. The period of the rotating cam was not
determined since the delayed fluorescence and phosphores-
cence lifetimes were quite long. All spectra were recorded

at a medium scan speed.

Concentrated Chelate Solutions

 

Concentrated chelate solutions could not be prepared
in the normal manner since flavonol is soluble only to the
extent of about 0.05 M in absolute ethanol. However, the

chelates themselves were found to be extremely soluble

68

in ethanol, and this characteristic was utilized in the
preparation of concentrated chelate solutions.

An appropriate amount of a concentrated aluminum chlor-
ide solution was placed in a volumetric flask, and the
required weighed amount of solid flavonol was dissolved
in it. The solution was then diluted to volume with
absolute ethanol. Solutions prepared in this fashion
were used in the aluminum-27 NMR and Raman studies. In
addition, these concentrated chelate solutions were also
used in the preparation of the solid samples which were

used in the infrared and X-ray powder diffraction studies.

Raman Spectra Measurements

 

All liquid and solid samples were sealed in 1.8 mm
glass capillary tubes. The spectrum of flavonol was ob-
tained on a solid sample because of its limited solubility
in ethanol. The spectra were run by the use of the argon
ion line at 5145 A as the source. The laser was operated
at 30.5 amperes and the spectra were measured at a scan

rate of 20 cm-1

per minute. The sensitivity, filtering
and slits were adjusted to meet the requirements of the

samples.

Infrared Spectra Measurements

 

Solid chelate samples were obtained by the evapora-
tion of the solvent from solutions which contained the

appropriate stoichiometric ratios of aluminum and flavonol.

69

The evaporations were carried out on sodium chloride plates
and yielded thin solid films. All of the spectra were
obtained at a slit setting of normal, a slow scan speed

and with air as the reference.
X-ray Powder Diffraction Measurements

Again, solid samples were prepared by the evaporation
of the solvent from liquid samples. Small amounts of the
solids were then placed on cellophane tape to hold them in
place. Finely powdered platinum metal was added to each
sample as a reference. The tape was then placed on a metal
disc so that the sample was positioned over a small hole
in the center. The sample was then placed in the target
chamber along with a camera and irradiated with the copper

K line at 1.541 A. The source was run at 35 kilovolts

c
1

and 18 milliamperes and the film was exposed for a period

of six hours at these settings. Subsequent development of

the film yielded the desired diffraction patterns.

Aluminum-27 NMR Measurements

All measurements were made on concentrated chelate
solutions which were contained in 5 mm sample tubes.
Aqueous aluminum chloride was used as a reference for all
the measurements. The resonance frequency for the reference
was at 59,133,015 Hz for a magnetic field strength of 53

kiloguass. All shifts for the chelates were to the low

70

field or high frequency side of the reference signal. The
number of scans per spectrum was varied and depended on
the concentrations of the samples but in all cases, the
dwell time was four milliseconds and the sweep width was
1684 Hz. In addition, all measurements were performed on

spinning samples and at room temperature.

III. RESULTS AND DISCUSSION

ABSORPTION AND FLUORESCENCE SPECTRA OF FLAVONOL

The absorption and fluorescence spectra of flavonol in
absolute ethanol are presented in Figure 10. The maxima
of the major bands in the absorption spectrum occur at 306
and 344 nm, and they can be attributed to carbonyl (w,n*)
transitions, since the molar absorptivities are quite large
(8306 = 10,600 and £344 = 16,500). The fluorescence spectrum
also consists of two major bands with the maxima located at
410 and 535 nm. These bands almost certainly result from
the deactivation of a (n,n*) excited state. However, the
fluorescence spectrum is a little unusual from the stand-
point that it is not the true mirror image of the absorption
spectrum as would be expected for an ideal fluorophore.
In fact, the intensity of the high energy band is quite
low compared to that of the band at 535 nm. In addition,
the wavelength region between the two maxima is nearly
void of fluorescence. These characteristics seem to suggest

that flavonol may have some unusual absorption and emission

features.

EMISSION SPECTRUM OF FLAVONOL AT 77 K

The emission spectrum of flavonol at liquid nitrogen

temperatures is presented in Figure 11. The spectrum

71

72

Fluorescence

.Hocnrpm

 

 

 

opsHomo< :H Hoco>mdm mo mnvoomm concomouosam one cowpouomn< .OH ohsmHm
a: .npwsoao>m3

one cow Omm com on: co: omm oom 0mm

0 p r h p p _ b p P b b . P . L LP b _ Intr . . h r b P L p . n L b h n p L L p c
0H4 vHoo
om; .N.o
Chi IMOO
cs. ea.o
owl Tmoo
co. .w.o

a: ::m u newcoflo>m3 soapmvaoxm

on. .5.0
cm. rw.o
col Hcccenaa s muofi w 6.: .o.c
OOH _ J J14 A a . q . 1 a 3 a a a a 1 . 14 1 3 «‘4 AAA 4 a a 1 q 0H

 

aoueqaosqv

73

one

.Hocnrpm opaflcnpa ca Hcccenfla nee x ea pa asnpcoam

a: .spwcoao>o3

cos omo coo omm com

b

coammwsm .HH onswdm

on: cos

D
P

 

 

‘ d

an sen one won I mnmeoHo>m3 cofipopdoxm

own

I

J

A

A

.0

.od

.ON

it on

.c:

.om

.ow

.On

.00

 

00a

 

Kitsueiux uorsstmg

74

consists of a single band which is attributed to delayed
fluorescence, since it is located on the high energy side
of the main prompt fluorescence band and has a fairly long
lifetime. The location and intensity of this band is de-
pendent upon the wavelength of excitation and somewhat on
the presence of oxygen. Table II summarizes these depen-

dencies. The data contained therein have not been corrected

Table II
Dependence of the Delayed Fluorescence of Flavonol

on Excitation Wavelength and the Presence of Oxygen

 

 

 

Excitation Emission Emission Dissolved

Wavelength Maximum Intensity Oxygen
306 435 1000 Present
344 462 55 Present
306 440 714 Absent
344 456 59 Absent

 

 

for inner-filter effects or for variations in the signals
due to sample positioning. However, it is quite apparent
that the production of delayed fluorescence is much more
efficient when flavonol is excited into its second elec-
tronic state. This phenomenon may be rationalized through
the following explanation.

Excitation at 306 nm would produce an excited singlet

75

state with sufficient energy so that intersystem crossing
results in a triplet state with a large excess of vibra-
tional energy. The second intersystem crossing to an
excited state then occurs before the lowest vibrational
level of the triplet state is reached. Since in this case
no additional energy is required for the second inter-
system crossing to occur, the production of delayed fluores-
cence would be fairly efficient. On the other hand, excita—
tion at 344 nm would produce a less energetic singlet state.
Intersystem crossing would then result in a triplet state
which was much lower in vibrational energy than that pro-
duced by excitation at 304 nm. Presumably, in this case,
the dissipation of this excess vibrational energy to the
lowest vibrational level would occur before a second inter-
system crossing to an excited singlet state would be achieved.
Since after this dissipation, the energy of the triplet
state would now be lower than that of the lowest vibrational
level of the first excited singlet state, a second inter-
system crossing could only occur with an increase in vibra-
tional energy. Consequently, the production of delayed
fluorescence by excitation at 344 nm would be relatively
inefficient.

At this point, it is interesting to note that the data
contained in Table II show that the delayed fluorescence
occurs in a spectral region where very little prompt

fluorescence is observed. This seems to indicate that

76

the delayed fluorescence originates from an excited state

other than the ones which give rise to prompt fluorescence.

(n,n*) EXCITED STATE OF FLAVONOL

In the ultraviolet absorption spectrum of flavonol,
shown in Figure 10, the absorption tails off in an unusual
manner in the region between 390 and 450 nm. The excitation
spectrum of flavonol, which is presented in Figure 12, was
obtained by monitoring the fluorescence generated at 535
nm. It is apparent that the tailing in the 390 to 450 nm
region is missing. Consequently, it must be concluded that
absorption in this region does not result in the generation
of fluorescence at 535 nm. However, as is shown in Figure
13, excitation at 410 nm does generate a single fluores-
cence band with a maximum located at 484 nm. The partial
quantum efficiency plot presented in Figure 14, shows
definitively that a previously undiscovered electronic
excited state is formed when flavonol is excited by radia-
tion within the tailing region.

At this time, it is believed that this new absorption
band arises from a (n,n*) transition, although the experi-
mental evidence is somewhat ambiguous. The molar absorp-
tivity of this band based on the concentration of flavonol
is small (e410 2 250) and is indicative of a (n,n*) transi-
tion. On the other hand, the spectral studies of flavonol

in different solvents by Urbach (53), show no correlation

Fluorescence

77

 

 

 

 

100 . . ....... . . . l l L J 4gL , 1 A 1 L
90* t
801 Emission Wavelength - 535 mm P
70‘ '
60 a _ I y-
504 _-
40- -
30‘ p
20: r
10* . t

O 1 1 Tea ... , . r. . . . . a I . . . r .. . 1
250 300 350 400 450 500

Wavelength, nm

Figure 12. Excitation Spectrum of Flavonol in Ab-
solute Ethanol.

78

Fluorescence

.Hocncem

 

 

opsaomp< Cw Hoco>mam mo unpooom cocoomososam one soapouoon< .ma onsmam
a: .nvmsono>m3
comet . .cmo. _ + .crmm. . . .cmmt r. .omsr Coma. .omm. . . roent . .+cmw
o”; (I: 1H.o
owl newcoao>m3 :uwvwwwomm .N.O
oml em.o
oer .s.o
cml .m.o
om. een n newcoaoeoz coavovwoxm .O.o
our .u.o
ow. .m.o
om. em.o
ooH . . ..ri . .ilwrii». i .... i . . . w. i it. .i i a. . aixax1... .0.“

 

 

aoueqaosqv

Relative Intensity

79

 

 

 

 

 

 

 

100 l l l 1 _l l_ l l L l 1 l 1 1 l 1 1 l l J l 1
Emission Wavelength - 484 nm
901 -
80‘ _
PQ
701 .
Absorbance

60. p -

50- .

no. r

301 .

20" r-

10- _
O T I T I Y T I I l l I r I I I f I T '—T I 1 T I

250 300 350 400 4 0 500

Wavelength, nm

Figure 14. Plot of Partial Quantum Efficiency (PQ)
as a Function of Wavelength for Flavonol
in Absolute Ethanol.

80

between band position and solvent polarity. In his studies,
the new absorption band was absent when flavonol was dis-
solved in chloroform, carbon tetrachloride or acetone and
present when dissolved in isooctane, methanol or ethanol.

In cases where the band was observed, no apparent shift in
wavelength was noticed. If indeed the transition in question
were a (n,n*) transition, it should undergo a hypsochromic
shift as the solvent polarity was increased. Since this is
not the case, a transition assignment cannot be made. The
absence of the absorption band in non-polar solvents could

be explained by the fact that (n,n*) transitions undergo

a bathochromic shift as solvent polarity is increased. In
this case, the absence of the band indicates that it is
located under the other major absorption bands when non-
polar solvents are used. However, this argument is extremely
weak since the band is present for flavonol solutions in
isooctane and absent for acetone solutions. Consequently,

it would appear that the solvent studies on flavonol are
quite inconclusive and assignment, as to the type of transi-
tion involved, is not possible.

Further spectral studies of flavonol in absolute ethanol
have indicated that the presence of acid affects the posi-
tion of the absorption band and its associated fluorescence.
This effect is shown in Figure 15. Upon the addition of
ethanolic hydrochloric acid, the absorption band undergoes

a slight blue shift, which may be indicative of a (n,n*)

81
Fluorescence

. .Hocmnvm ovsaomp<
oaoao< ca Hoco>mam mo unpooom oocoomouosam one coavnuomn< .ma onswwm

a: .nvwcoaoeus

 

 

..h. _ _ ..e. _ . ...h; .3“. it? _ ......L. e... L 12.: _ aw
OH. .H.O
ON. emu/03 I .N.o

gemsoao>o3 coaaovuoxm
on- K6
oer . is.o
on. w .3
col ac sen . rewroaoeas ccsnnpdoxm .o.c
o5. .e.o
om. . 66
cm. 6.0
cod . .1.. l .. . .4 a... 4 i?.. ilira. ... .. a i .. . ..r. .Ji... O.H

 

 

 

aoueqaosqv

82

transition. In addition, the fluorescence band also under-
goes a hypsochromic shift and its maximum is moved from 484
to 455 nm. The magnitude of this shift is larger than that
which would normally be expected for a change from a non-
polar to a polar solvent, but in this case, the change in
polarity is quite drastic.

At this point, it would seem that the evidence favors
the likelihood that the tail in the spectrum of flavonol
in absolute ethanol is actually a (n,w*) transition. How-
ever, one more possibility should be considered. Since
flavonol is an a-hydroxy ketone, there exists a possibility
for the formation of a keto tautometer. If a small amount
of the keto form does exist, it is then possible that the
newly discovered band is actually a (n,n*) transition for
the isomer. If this were the case, the previously evaluated
molar absorptivity for this band would be in error since
the concentration of the keto form would be much smaller
than the concentration of flavonol. Reconsideration on
this basis would indicate that the band is a (n,n*) transi-
tion.

Unfortunately, the results of the solvent studies by
Urbach are once again inconclusive. Since the enol form
is normally favored in non-polar solvents for reasons of
solubility (57), it is expected that the intensity of the
band in question would be depressed for these types of

solvents. Obviously, the presence of the band for flavonol

83

in isooctane and its absence for acetone solutions does
not support the existence of the keto tautometer.

Upon the addition of acid to flavonol, it is expected
that the fluorescence, which is emitted from the newly
discovered excited state, would decrease in intensity,
since the enol form is now favored because of the forma-
tion of the enolic oxonium cation (57). In fact, the
fluorescence intensities before and after the addition
of acid are just about equal. This result also seems to
indicate that the new absorption band does not originate
from the presence of a keto tautometer. In addition, since
it is believed that the delayed fluorescence occurs via a
radiational deactivation of this excited state, it is highly
unlikely that excitation of the enol form at 344 nm would
result in the emission of delayed fluorescence from an
excited state which is supposedly associated with the keto
form. Based on the foregoing arguments, there seems to be
little doubt that the newly discovered absorption band
is not caused by the existence of a keto form. However,
the existence of a (n,n*) transition in the wavelength
region of 400 nm is quite unusual.

Normally, the (n,n*) transitions for aromatic carbonyl
compounds occur between 300 and 350 nm and are usually the
lowest energy electronic transitions. The efficiency of
intersystem crossing to the triplet state is ordinarily
quite high in these types of molecules and generally, they

will not exhibit much fluorescence. However, flavonol

84

is unusual in that the molecule contains a very highly con-
jugated n system. In fact, the whole molecule is involved
in this system. As a result, it is expected that the low-
est energy transitions for flavonol would be of a (n,n*)
nature. Indeed, this statement would probably be true if
it were not for the fact that flavonol is also an oxygen
heterocyclic molecule. In this case, conjugation across
the pyrone ring causes the energy of the (n,n*) state to

be lowered to a point where it is actually the lowest energy
transition. It has been found that conjugation in unsatu-
rated ketones has lowered the energy of the (n,n*) state to
such an extent that in molecules such as p-benzoquinone,
the (n,n*) transition occurs at 435 nm. Presumably, this
same type of conjugation is responsible for the low energy

(n,n*) transition in flavonol.

ABSORPTION AND FLUORESCENCE SPECTRA OF THE CHELATES

The absorption and fluorescence spectra of the flavonol-

aluminum (III) chelates are presented in Figure 16. All

of the chelates, no matter what the stoichiometry, have

the same spectra. The absorption spectrum consists of

two major transitions above 250 nm. The maxima of these
absorption bands are located at 326 and 405 nm and are
identical to the ones described by Urbach (53). It is

also apparent from the molar absorptivities of these bands

that (n,n*) transitions are involved. The fluorescence

85

Fluorescence

.Hcearem opaflcnna ca

 

 

 

 

movoaoso Hocoemamnsssfissa< on» no unpooom oocoomohosam use soavosomp< .OH onswwm
a: .svwcoao>m3
%ow can con . one cos . can con _ Tom m

Oar .«.o
ON. .~.o
on 1 v n o o

. "v

ca. .a on“

O

.m. “w

on. five

u

0

cm 1 I w 0 o a
on. .5.0
owl a: no: - remcofloeas ecseaeeoxm .m o
co. _ .m.o
cod 1. w . i . . . ., . . 1 . . . .,. a . w, i . . li4r . . . . o.”

 

86

spectrum for the chelates consists of a single band with
its maximum located at 450 nm. It also appears that the

fluorescence originates from a (n,n*) transition.

EMISSION SPECTRA OF THE CHELATES AT 77 K

The emission spectra for the 2:1 and 6:1 metal-to-
ligand chelates are presented in Figure 17. The spectra f
consist of two major bands. The band at 672 nm has been i
attributed to the emission of phosphorescence, while the i
band near 460 nm has been attributed to delayed fluores-
cence. The location and intensity of the delayed fluores-
cence band is dependent on the stoichiometry of the chelate,
the wavelength of excitation and the presence of oxygen.

Table III summarizes these dependencies. The data contained
in this table have not been corrected for inner-filter ef-
fects or for variations in the emission signals due to sample
positioning.

As was the case for flavonol, the excitation of the
chelates to the second excited state (326 nm) is more ef—
ficient in the production of delayed fluorescence in most
instances. The only exception to this phenomenom is the 6:1
chelate in the presence of oxygen. In addition, it is also
apparent that there are some major differences in the
emission characteristics of the two chelates. The presence
of oxygen with the 2:1 chelate seems to have little effect

on the intensity of the emitted delayed fluorescence. On

87

.Hocogpm opsaomn< CH movmaoco Hoco
u>mHm op secessa< H.O ocm H.~ on» you a an em sspvoonm scamofism

s: .gpwsoao>m3

one con 0mm cow 0mm com on:

p p b n p
I A 1

a: mom can cmn . nrpmnoao>ns rescuesoxm

 

.ea enemas

L

cos own
.0

 

 

OOH

Kitsueiul uotsstug

88

z IOH x ~.H I sowuosucoocoo Hoco>mamn

v

z IOH x N.m I cowuouucoocoo Hoco>mamo

w

 

 

 

usomnd «OH mmv mmm
m~.~
second man mmo omm naum
pcomoum oom omv mam
oo.H
usomoum com omv omm oauo
oceans chm one . mam
mm.~
ucomnfi mmn owe omm maum
anemone mmm owe mam
mm.m
ucomoum coca owe omm oHuN
commxo a: m H auwmcousH safiwxmz numsoao>mz Hoco>oamuad
OO>HOmmwo Ass ommvH scammHEm cowmmwsm cowumufioxm oumaono

 

 

commxo mo mocomowm on» one numcoao>o3 cowuouwoxm
.muuoEOAnoHoum ouoaonu so oosoomOHOSHm oomoaoo mo mocopcomoo

HHH manna

 

89

the other hand, the presence of oxygen has a marked effect
on the 6:1 chelate. In this case, the production of delayed
fluorescence from excitation at 330 nm is greatly decreased.
In fact, the fluorescence intensity is the same for both
excitation wavelengths. Upon the removal of oxygen, the
production of delayed fluorescence from excitation at 330

nm increases although the efficiency is not as high as for
the 2:1 chelate. This phenomenon may be explained by the
following argument.

If the mechanism for the production of delayed fluores-
cence in the chelates is similar to the one proposed for
flavonol, then the emission could occur in the following
fashion. Figure 18 shows the energy level diagrams for
the proposed mechanism. Figure 18A shows excitation to
the second excited singlet state (326 nm). A slight amount
of energy is lost to internal conversion before the inter-
system crossing occurs. The triplet state is then formed,
and some dissipation of the excess vibrational energy occurs
before a second intersystem crossing takes place. Note that
the loss of the vibrational energy does not result in a
triplet state of lowest vibrational energy. During the second
intersystem crossing, energy is again lost which results in
a singlet state from which the delayed fluorescence originates.

Figure 188 shows excitation to the first excited singlet
state (405 nm). Again, a slight amount of energy is lost

before the intersystem crossing can occur. Upon crossing,

 

90

Am
:a . a chow
‘ omp< cu m
Hocmnvmawwww ammo

as nmofin

.x 2. use .

now

one H.N 0:»

.0 . .

< «

cassa

ow so .

Hm

o:o>m

0:0 H

moved

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4 cm
m cm A: «W
. an. Ill”
lflm.) . 1 Hill mm
«a mMmm “Hum.
.3: W I
Hm I mm H
.. on mm
mm mm
nun” mmmu

 

 

 

 

 

 

 

 

||llflflmH

91

the resultant triplet state is formed with a relatively
little excess of vibrational energy. Consequently, before

a second crossing can occur, the molecule loses almost all

of this excess energy, and the second crossing must occur
with an increase in energy. Note that this was not the case
when the molecule was excited to the second electronic state.
As a result, the production of delayed fluorescence in

this second case is much less efficient. Since this mechanism
was used for flavonol to explain this phenomenon, it seems
likely that the emission mechanism for the 2:1 chelate would
be similar, since the data are nearly the same. On the other
hand, it would not explain the data which were Obtained

for the 6:1 chelate.

Apparently, the presence of oxygen has a serious effect
on the production of delayed fluorescence from this chelate.
Note that the production is fairly inefficient even though
the fluorescence is excited at the 326 nm band. This phen-
omenon may be explained by the energy level diagram in
Figure 18C. This figure shows that the amount of vibrational
energy which is lost to internal conversion in the triplet
state is increased by the presence of oxygen, and a low
energy level is reached before the second intersystem
crossing can occur. Consequently, it is expected that the
production efficiency of delayed fluorescence for this
mechanism would be the same as for the mechanism shown in

Figure 183 where the molecule is excited only to the first

 

 

92

electronic state. On the other hand, upon the removal of
oxygen from the system, excitation at 326 nm now results in
a more efficient production of delayed fluorescence, since
the mechanism shown in Figure 18A is now in operation.

At this point, an important question arises. Why does
the presence of oxygen affect the 6:1 and not the 2:1
chelate? Perhaps the answer to this question lies in the
structural differences between the two chelates. Since
it is believed that the 6:1 chelate is much larger than
the 2:1 chelate, it is likely that the quenching of the
larger molecule by oxygen is much more efficient. Other
evidence also seems to support this argument. Note that
even in the absence of oxygen, the intensity ratio (1330/
I395) for the delayed fluorescence from the 6:1 chelate is
not as large as it is for the 2:1 chelate. Apparently,
internal conversion of the triplet state for the 6:1 chelate
is more efficient even without the presence of molecular
oxygen. Consequently, it is logical to assume that oxygen
would have a greater effect on the internal conversion pro-
cesses of the 6:1 chelate, and as a result, the production
of delayed fluorescence is decreased.

Unlike flavonol, the delayed fluorescence for the two
chelates occurs at nearly the same wavelength as the prompt
fluorescence. Since it is fairly certain that the prompt
fluorescence originates from a (n,n*) excited state, it
then follows that the delayed fluorescence eminates from

that same state. However, for flavonol, the delayed

93

fluorescence seems to originate from a (n,n*) excited state.
Consequently, it appears that upon chelation, the lowest
lying (n,n*) transition of flavonol (344 nm) undergoes a
bathochromic shift to 405 nm. On the other hand, the low
lying (n,n*) transition of flavonol (410 nm) would be ex-
pected to undergo a hypsochromic shift and as a result, it
is lost under the (n,n*) manifolds of the chelate absorption
spectrum. Therefore, the (n,n*) transition at 405 nm is

now the lowest energy singlet transition, and the delayed
fluorescence is expected to originate from this excited

state.

CHELATE STOICHIOMETRIES

Basic Solutions

As noted by Urbach (53), two chelates are formed under
mildly basic conditions. When the hydroxide to aluminum
(III) ratio was one, a chelate with an aluminum to flavonol
ratio of 2:1 was formed. Upon standing, further chelation
occurs, and a 1:1 chelate is formed. When the hydroxide
to aluminum (III) ratio was raised to 2.5, only the 1:1
chelate was formed. In solutions where the hydroxide to
aluminum (III) ratio exceeded 2.5, no chelate formation

was indicated.

94

Neutral Solutions

 

The term "neutral" in this case indicates that no acid
or base has been added to the test solutions. As noted by
Urbach, the existence of two chelates is indicated by
photometric studies. In freshly prepared solutions, the
existence of a chelate with a metal-to-ligand ratio of 2:1
was indicated. Upon standing for two weeks, further chela-
tion resulted in the formation of a 1:1 chelate in addition
to the 2:1 species. A 6:1 metal-to-ligand chelate has
also been found to exist, but this chelate could only be

detected with fluorometric methods.

Acidic Solutions

Present studies indicate that two chelates are formed
in the presence of acid when the acid to aluminum (III)
ratio was 5:1. A 1:1 chelate can be detected by either
a photometric or fluorometric titration of aluminum (III)
with flavonol as shown in Figure 19. The other chelate
which has a metal-to-ligand ratio of 2:1 can only be detected
by a fluorometric titration. No significant stoichiometric
changes occur with time except that the break for the 2:1

chelate becomes better defined.

95

.Hum.oHpmm noH Haves op noponm one ans Hono>mHm 29H: oovmans aHHHv san
IESH< now mo>nno nOHpoane OHHPosononHmoneooom one Oanosoponoonwooom .oH onsmHm

AHHHV assassa< Op Hono>mam mo prmm oaoz

 

 

 

 

 

 

%.a . a: . on. . a”. . a: . 0.; . so . so . set no . as
I I I All
OHL. ..H.o
ONl. ..N.o
oml. :m.o
m oi:r an on: em monoomonosHm .N N :&.o
n
e an a cone non .
u a: me: a. 9 pa H ..mé
m mnoH snnHE=H< nH z mIOH x o.N
mowjr avooo
w.
os.. H $5.0
own: o nuwoO
co: . - - - ..m.c
OOH O.H

aousqzosqv

96

RAMAN SPECTRA

Of all the chelates, perhaps the most interesting are
the 2:1 and 6:1 chelates which are formed in "neutral" solu-
tions of absolute ethanol. Urbach (53) has noted that the
6:1 species is much more fluorescent than the 2:1 chelate.
Since an increase in structural rigidity can be responsible
for an increase in quantum efficiency, it was suspected that
the 6:1 chelate might have an unusually rigid structure,
which might account for the large increase in fluorescence
efficiency. If this were the case, Raman and infrared spec-
troscopy should give an indication of the nature of this
structure. Consequently, samples of flavonol and the two
chelates were subjected to analysis by these techniques.

The Raman spectra for these sample have been tabulated
in Table IV. The Raman spectrum of flavonol was obtained
with a solid sample since the solubility of flavonol in
absolute ethanol was not large enough to obtain reasonably
intense scattering peaks. The 2:1 and 6:1 chelate spectra
were obtained with ethanolic solutions which were 0.245 and
0.048 molar in flavonol, respectively.

It is apparent from the data contained in Table IV
that there is structurally little difference in the chelates
as far as flavonol is concerned. Since the reading error

1, apparently there are no significant shifts

is about :3 cm-
in the spectral features in either spectrum. There is only

one scattering peak which is not common to both of the

Raman Spectra of Flavonol, 2:1 and 6:1

Metal-to-Ligand Chelates

97

Table IV

 

 

 

 

Flavonol 2:1 Chelate 6:1 Chelate
cm-l Intensity cm"1 Intensity cm.l Intensity
198 3 275 l 278 1
263 3 402 l 402 l
297 3 432* l 430* S
332 l 490 2 492 5
378 1 520 1 520 1
433 4 578 15 578 46
438 3 617 1 617 l
510 7 667 3 667 8
579 9 683 6 683 17
618 3 844 6 843 14
624 3 881* 2 881* 36
672 9 997 12 997 32
777 3 1044* l 1052* 7
836 8 1097* 1 1097* 9
988 18 1152 7 1153 22
997 13 1168 3 1169 3
1032 3 1187 6 1188 17
1148 4 1235 14 1235 46
1187 11 1278* -- 1278* 5
1210 1 1327 9 1328 29
1225 3 1357 2 1358 9
1245 3 1427 10 1428 39
1277 3 1443 12 1444 55
1307 23 1459 7 1458 30
1318 9 1492 2 1492 3
1350 8 1497 1 1498 l
1410 7 1529 12 1530 47
1443 11 1557 9 1558 37
1468 14 1573 l 1573 2
1480 1 1595 24 1597 94
1489 4 1614 15 1615 57
1565 98 ‘ 2875* 54 2876* 58
1594 79 2925* 100 2927* 100
1618 100 2969* 40 2971* 43
3070 3 3073 l 3073 --
3077 2

 

 

,__
Ethanol Bands

98

Ichelate spectra and it appears at 3073 cm”1 for the 2:1
Iihelate. Since a vibrational peak in this spectral region
is normally associated with the -OH stretch of alcohols,
ii: may indicate that some free flavonol is present in solu-
tion. Unfortunately, the concentration of flavonol in the
6:21 chelate solution could not be increased because of the
lxirge scattering background which was encountered. Conse-
quently, this peak in the spectrum of the 6:1 chelate may
rust be intense enough to be detected, and as a result, no
conclusions can be drawn.

The only other interesting spectral features in the
cflielate spectra are the scattering peaks which are located
iJi the region near 280 cm-1. These peaks are due to aluminum
CIII) in solution. When no flavonol is present, this
scattering peak occurs at 282 cm-1. There is a shift to
278 cm-1 for the 6:1 species and to 275 cm"1 for the 2:1
Chelate. These shifts may indicate that the aluminum (III)
Structure in the 6:1 chelate is fairly similar to the struc-
tnire of normally solvated aluminum ions even though chela-
tion has occurred. Unfortunately, these scattering peaks
are extremely weak, even under conditions of high spectrometer

sensitivity. As a result, no additional conclusions can

be drawn.

99

INFRARED SPECTRA

The infrared spectra for flavonol, the 2:1 and 6:1 che-
1ates are presented in Figures 20, 21 and 22, respectively.
The flavonol spectrum was obtained from a potassium bromide
pellet of the sample. The chelate samples were prepared
as indicated in the Experimental chapter.

As was the case for the Raman spectra of the chelates,
no apparent shifts were found between the chelate spectra.
Again, there seems to be little structural difference in
the two chelates as far as flavonol is concerned. However,
one interesting piece of information was obtained from this
study. Figure 23 presents the spectrum of the dry residue
formed by the evaporation of ethanol from a solution of
aluminum chloride. The sample for this spectrum was prepared
in the same way as the chelates. Note that this spectrum
is markedly different from that of absolute ethanol which
is presented in Figure 24. This difference is explained
by the fact that the residue actually contains very little
free ethanol. Since this sample was evaporated to dryness,
the Spectrum is really that of the ethoxide ions and ethanol
associated with aluminum (III). Upon re-inspection of the
chelate spectra, it is evident that the spectral features
found in the residue spectrum are quite prominent in the
6:1 chelate spectrum, but not in the 2:1 spectrum. In
fact, parts of the 6:1 spectrum are almost obliterated

by the presence of these features. Based on these

100

.Hono>mHm mo ssnpoonm nOHPonomp< oosonmnH

HIEO .honosoonm

roman. . .cc.cm. . .

.ocMm r 8

. <8 3&2

d.

 

 

 

 

 

 

 

 

 

 

P b
1 J l. i q A 11 d

:00

 

 

 

rOOH

acusiirmsusam %

 

101

com

.Hono>mHm mo ssspoomm noHpmnomn< oonmnmnH .mON onsmHm

oooH

H

Inc .honosoonm

ommH

lP

omaH

1P

oooH

«I-

comH

4e

OOON
o

 

 

 

 

 

 

 

db

 

 

db

 

 

 

 

1‘.

 

 

 

 

 

 

 

 

 

 

..OH
Ion
:om

:ca

 

 

aousiirwsueam %

102

.oemHonO
Hono>on ow sanssH< Ham one Mo ssnpoonm noHeonomn< oonmnmnH .<HN onstm

HIEO .zononoonm
oomN coon .

b
‘ 1“ d d J R ‘I i I

oomH.

h h
d A d d 1

 

OOON

E b p
q '3 1 a

.TON

.om

1

L

.00

Ice

 

 

 

 

 

00H

aousiirmsusam %

103

.mvmamno
Hoco>mHm op assassH< H.N ms» mo ssnpowmm cofipmuomn< cmMMhmcH .mam mHzMwm

«lac .zoCmswmnm

00w 000a OONH

- p h
1 d J d * «I

003H

1

oowa coma OOON

J .4 d

d.

d-
O

 

d-

 

.. 3

 

 

 

 

 

 

 

 

 

 

 

00H

 

eouegzgmsueam %

104

.mpaamso
Hoco>wah op Eszwsza< «.o 0:9 mo sauvomgm COfiPonmn< cogwpth

.<~N musmflm

Hugo .hozmavmhm

ommm

d # d

ooom

¢

4&-

P L! b
d 4‘ d1

 

.
h-
dr-

 

 

 

 

 

 

 

aouezz;msuea¢ %

.opafloso
Hoco>mHm on E::HE:H< a.w 0:9 mo asugomgm coavguomn< omumuwcH

H...Eo .zocmsumhm

oom oood omaa

b
w

.mNN mhsmwm

coma OOON

fil-

 

coma coma

db
1}-

d

‘-

d

«u-
T
J.

105

 

 

 

 

 

 

 

d

 

 

aouegztmsuedm %

106

.covauoau>m
ma ocfiuoaso sacaaaa< mo soapsaom ouaonooa< cw scum vcm>Hom
¢sv 2mg: coeuom osvammm on» no asupooam :oavnnom94 vonauwcH .<m~ wuzmwm

disc .aonmsuoum

“SF * +0.3? . o 3

h h I
la 4‘ 4 I4 4 q a ‘ o

P b
q .

J. 8?#

 

:om

é>
~o

 

 

 

 

 

ooH

aouezarmsuaam g

107

.cmkuonw>m
ma «vauoaso assassa< mo :oapnaom oaaozoOH< ca scum vno>aom

 

   

as» can: cwauom oficumom 0:» mo sapwoomm coapnuomn< cohmnmcH .mnm onsmam
«lac .hocozwoum

V» own 090“ fi, coma amid vw owed omwu 09mm

:oH

lbw

.6m

.0:

 

 

 

 

 

Am

100

cu

 

 

aouezztmsuaam 5

.Hozmnpm mPSHomn< ho Esnpommm :oapmpomp< cmumnmcH .<¢N muzmfim

«use .moCozwmnm

ronN. . .opmm_ L‘ .

 

108

F P
d J ‘ 4

b , .oom~+ .opom; . 68%

...OH

.. ON

C
In

.00

L

 

 

00H

 

aoueg¢ymsueam %

1P

.Hocmspm mpsaomn< mo sappomgm :oflpmuomn< cmpmumcH

Hugo .zocosdwnm

oom OfioH coma ooafi oowH

. mau 98mg

oowH OOON

{b
O

 

109

 

1-
db
q»
d-

d W 1

d

tom

L

A

.3

.om

J

O
\O

JTON;

 

00H

 

aouegzgwsueam %

110

observations, it would appear that the 6:1 chelate is some-
what similar in structure to aluminum (III) in absolute

ethanol.

ALUMINUM-27 NMR

Since very little information about the chelates was
obtained from the Raman and infrared studies, aluminum-27
nmr spectroscopy was employed. A typical nmr spectrum for
the chelates is shown in Figure 25. The chelate resonance
peak lies to the high frequency side of the reference
signal. The widths at half height for the reference sig-
nals were about 45 Hz, while for the chelate peaks,they
ranged from about 350 Hz to a little greater than 400 Hz.

Figure 26 presents the shifts for the two chelates as
a function of aluminum concentration, and it indicates that
a concentration dependency exists. Since the concentrations
used in this study are about one thousand times larger than
those normally used in fluorescence studies, there is a
good possibility that the chelate structures change with
increasing concentration. It is interesting to note that
as the concentration of aluminum (III) increases, the dif-
ference in the shifts decreases. This may indicate that at
high concentrations, only one chelate exists in solution.
This phenomenon may explain why the Raman and infrared
spectra were identical for the two chelates.

Figure 27 shows the shift of the aluminum (III)

Relative Intensity

111

 

100*L

to
O

20 v

 

1 PPM = 59 H2

 

 

 

 

Figure 25.

J26 2 u 6 8 10121u161820 22
Shift, ppm

Typical Aluminum-27 NMR Spectrum for the Al-
uminum-Flavonol Chelates Formed in Absolute
Ethanol.

 

 

 

   

 

     

 

 

 

112
12
Aluminum to Flavonol
Ratio - 6:1
11~_
Aluminum to Flavonol
Ratio - 2:1
10..
E
D.
+; 9..
r...
..4
.c
U)
81-
7%
01‘ a: : : : LA
0 0.1 0.2 0.3 O.h 0.5 0.6

Aluminum (III) Concentration, I

Figure 26. Aluminum-27 NMR Shift as a Function of Aluminum
(III) Concentration for Solutions with 2:1 and

6:1 Aluminum to Flavonol Ratios in Absolute
Ethanol.

113

 

13

0.103 M in Aluminum Ions

12.4)-

117-

10 ir-

Shift. ppm

 

 

O I l 1 L 1 1 I 1 L
I

b i '2 3' LI 5 5 7 8 9

Mole Ratio of Aluminum (III) to Flavonol

Figure 27. Aluminum-27 NMR Shift as a Function of the Alum-
inum to Flavonol Mole Ratio in Absolute Ethanol.

114

resonance signal as a function of the mole ratio of aluminum
to flavonol for a constant aluminum (III) concentration of
0.103 M. The shape of this plot was quite unexpected and

is difficult to explain. Apparently, some unknown signifi-
cant structural change occurs at a mole ratio of five alu-
minum ions to one flavonol. The results of this aluminum-
27 nmr study as well as the Raman and infrared studies in-
dicate that the chelate(s) formed at high aluminum (III)-
flavonol concentrations may be different from those formed
at the concentration levels normally used in fluorescence

studies.

POTENTIOMETRIC TITRATIONS

When aluminum salts are dissolved in protic solvents,
the resultant solutions are acidic. Based on this charac-
teristic, much useful information has been gained from
the titration of these solutions with sodium hydroxide.
Ohnesorge (55) determined from potentiometric titrations
that five protons are released for every two aluminum ions
which are solvated in absolute ethanol. In addition, he
also found that with the presence of small amounts of water,
the normal titration curve for aluminum chloride in absolute
ethanol shows a shift toward higher proton to aluminum
ratios. Figure 28 shows the potentiometric titration curve
of anhydrous aluminum chloride in absolute ethanol. The

inflection point of this curve occurs at a hydroxide to

115

11

 

3.2 x 10’“ M in Aluminum Ions

Apparent pH

 

1JL

 

l 1 A
I I Y

2 3 u
Mole Ratio of Hydroxide Ions to Aluminum Ions

 

o
naqt

Figure 28. Potentiometric Titration Curve for Aluminum (III)

Titrated with Hydroxide Ions in Absolute Ethanol.

 

116

aluminum (III) ratio of 2.7 and is identical to the curve
which was obtained by Urbach (53). Apparently, even minute
traces of water in the ethanol result in a shift from the
2.5 ratio which was found by Ohnesorge. This shift was not
expected since anhydrous aluminum chloride was used in this
study whereas Urbach used the nonanhydrate. In any case,
Ohnesorge's results have been fairly well confirmed.

In a similar manner, a number of solutions, with vary-
ing mole ratios of aluminum to flavonol, have also been
titrated with sodium hydroxide and some interesting results
have been obtained. Urbach demonstrated that with freshly
prepared aluminum-flavonol solutions, the titration curves
show two breaks which occur at hydroxide to aluminum ratios
of 2.2 and 2.7. Since it is assumed that water should also
affect the positions of these breaks, these ratios actually
indicate a 2.0 and 2.5 hydroxide to aluminum relationship.
Urbach also noted that as the ratio of aluminum to flavonol
was increased, the first break became less well defined.
Unfortunately, he did not go higher than a ratio of four
metal ions per flavonol. In addition, all of the titra-
tions which he performed were on freshly prepared solutions.
Consequently, there is no information available on the 1:1
chelate which is formed upon aging.

To obtain this desired information, similar potentio-
metric titrations have been performed on a series of solu-
tions with varying aluminum to flavonol ratios which were

aged for two weeks. The results of this study are shown

 

 

117

in Figures 29 through 32. It is apparent from these curves
that except for the solution with a 2:1 metal-to-chelate
ratio, two breaks appear. As noted by Urbach, the first
break becomes ill defined as the mole ratio increases. This
indicates that the predominent species, which should be the
6:1 chelate, is responsible for the liberation of five
protons for every two aluminum ions upon its formation.

On the other hand, the formation of the 2:1 chelate, as
proposed by Urbach (S4), liberates two different sets of
protons. The first set consists of four protons and accounts
for the first break in the curve. The second set consists

of one additional proton which is titrated for every two
aluminum ions and accounts for the second break. Suppos-
edly, this one additional proton is the result of an abstrac-
tion from one of the ethanol molecules within the coordination
spheres of the aluminum ions.

Upon aging the 2:1 chelate, it is known that it is
transformed into a chelate with a 1:1 stoichiometry. The
potentiometric titration of this 1:1 chelate is shown in
Figure 33. Curve A is the titration curve for the 2:1
chelate while curve B is for the 1:1 species. Note that
the curve for the 1:1 chelate essentially exhibits one
break at a hydroxide to aluminum ratio of two. This
indicates that this chelate is formed with the liberation
of four protons for every two aluminum ions and that the
structure is significantly different from that of its

predecessor, the 2:1 chelate.

118

.Hocmzpm opsaomn< :H mcoH mofixopoz: Spas moapmm Hoso>mam

 

 

    

 

   

 

ow answesa< msowpm> mo muowpmupfis on» pom mo>pso :oflpmupwa owhpoEowpcoPom .mN opswfim
AHHHV assassa< op mcflxopomm mo afivmm macs

s m m A o s m m A o
A A .1 A «VA A A A o
...-H
.N
A
.3
A
110

i.

HAN I owpmm H.H n oflpmm
Hono>wam ow assassa< Hono>wam op escHESH< :m
..m
mCoH escwesa< :H E duofi x N.m

 

 

0H

Hd iueaeddv

119

.Hocmnpm mpsaomn< CH mcoH oofixopozx cw“; mowvmm Hoco>mam

op sscfl53H< msowsm> mo sowpmppfia on» pop mo>uso coapmupfle owupmsofivcopom .on mpzmwm
AHHHV assassa< op meanest»: mo canon oaoz
H o
A A A A ..L A A. A o

 

 

  

A.s u owpmm
Hoco>mHm ov assassa<

mcoH escAesH< cs 2 squ x N.m

 

A.m . capmm
Hoco>mam op assassa<

 

 

OH

pd iueaeddv

120

.Hocmcpm mpsaomp< :H chH mowxouomz new; mofivmm Hoco>mam

op Escflesa< msowum> mo coapwppwe one new me>uso cowpmnpwe ownpmsowvcopom .Hm opswfim

AHHHV escsssa< op mefixoneaz no oApmm mac:

n
N
dPH

 

 

1

A . A A

an _
Ni

qL-H

 

A.m . ofipam A.m . aflpmm
Hoso>mam o» assassa< Hoco>mHm op answesa<

m:0H assassa< 2H 2 anoa x N.m

 

A

 

 

0H

Hd iuareddv

121

m

.Hocmnpm opsaomp< :fi mcoH mowxopozz an“: mowpmm Hoco>wam

op seawesa< mzownm> mo coflpmuvwe one you mo>usu noHPMpvfie owupmsoHpCepom

AHHHV escsssa< op meflxosezx no.0Apmm «do:

"02

A m A a

n
d

m

b

"N
'1'H

.Nm shaman

 

 

‘

 

H.m I oflpmm
Hoco>wam op sacHESH<

mcoH Escfissa< 2w 2

3|

4‘

 

OH x N.m

A.A . ofipmm
Hoco>mam ow assassac

co

 

 

0H

yd iueJeddv

122

 

 

 

 

11
3.2 x 10’“ M in Aluminum Ions
10<P
9...
8-r (Aluminum to Flavonol
Ratio - 2:1
(Aged)
74.
Ii 6«- B A
‘2
a 54»
m
e
<: 4.» Aluminum to Flavonol
Ratio - 2:1
(Fresh)
3*-
2.
1 4)-
0 t t : :
o 1 2 3 1+ 5
Mole Ratio of Hydroxide to Aluminum (III)
Figure 33. Potentiometric Titration Curves for Fresh and Aged

Solutions with an Aluminum to Flavonol Ratio of 2:1
Titrated with Hydroxide Ions in Absolute Ethanol.

123

In the following sections of this chapter, all of the
information, which was obtained from this and several other
studies, will be used to arrive at a proposal for the
structures of the various aluminum-flavonol chelates in

absolute ethanol.

ALUMINUM (III) SPECIES STRUCTURES IN ABSOLUTE ETHANOL

Since the structure of aluminum (III) species in solu-
tion is certainly responsible for the large variety of
chelate stoichiometries, it is appropriate at this point

to begin with a discussion of these structural properties.

Basic Solutions

 

In basic aqueous solutions, aluminum salts dissolve
to give aluminate ions. The same reaction almost certainly
occurs in basic absolute ethanol solutions. In addition,

A13+ + 40H" + 2EtOH + [Al(OH)4(EtOH)2]-

these aluminate ions are believed to be monomeric in nature
and consequently, any chelates which are formed in basic

solutions should also be of a monomeric nature.

"Neutral" Solutions

 

In "neutral" aqueous solutions, the structure of

solvated aluminum ions appears to be quite complex. Since

124

aluminum salts undergo extensive hydrolysis upon dissolu-
tion, the solutions which result are quite acidic. In
addition, the aluminum ions in solution are considered to

3+

A1 + 6H 0 + [A1(H20)6]3+ + [A1(H20)6_n(os)n](3’n)+

2
be octohedrally coordinated and polymeric in nature. Un-
fortunately, very little is known about the extent of polym-
erization. Studies by Aveston (58) suggest that the pre-
dominent ionic species in aqueous solutions may be
[A12(0H)2]4+ and [A113(OH)32]7+. 0n the other hand, Brosset,
gt 31., concluded that the predominent species may be
[A16(0H)15]3+ and they have suggested that a ring struc-

ture may be involved (59). In addition, cryoscopic measure-
ments by Kohlschfitter and Hantelmann (60) tend to support

a hexameric species.

In "neutral" solutions of absolute ethanol, the stoi-
chiometry of the ionic aluminum species is believed to be
similar to the one proposed by Brosset and others. Since
Ohnesorge has shown that the ratio of ethoxide to aluminum
ions is 2.5, the simplest polymeric species would have to
be at least a dimer. However, since a 6:1 chelate has been
shown to exist, it would appear that aluminum ions in
ethanol have a hexameric structure. Therefore, it would
appear that the ionic species formed in absolute ethanol
is [A16(0Et)15(EtOH)n]3+.

From aluminum-27 nmr studies, it has been found that

 

125

anhydrous aluminum chloride in ethanol gives only one
resonance signal. This indicates that the aluminum ions,
which are involved in the structure of the ionic species
in solution, have similar or identical environments. To
meet this criterion, it is quite obvious that the aluminum

ions must be a part of a ring or cyclic structure. 0f

- - + .+

\TAi/

A1

V - ./’]1\./1\

 

 

 

 

d

h

.B.

0 = Ethanol 0 = Ethoxide

the two simple structures shown above, which have been used
by Ohnesorge (55,61) in the explanation of some of his work
with aluminum-8-hydroxyquinolates, neither A or B are
adequate building blocks for the construction of a six mem-
bered aluminum ring. However, slight modifications of
these structures do result in the three possible struc-
tures which are shown in Figure 34. In structure A, the
aluminum ions are not quite equivalent since the coordina-
tion sites within the ring are occupied by alternating
ethoxide and ethanol molecules. On the other hand, they
may appear equivalent on the nmr time scale since proton

exchange should be quite rapid in this ionic molecule.

 

126

 

Q

O= Aluminum $8 Ethanol .' Ethoxide

Figure 3AA. Various Proposed Structures for Aluminum (III)
in Absolute Ethanol.

 

 

127

 

C): Aluminum é?” Ethanol [Du Ethoxide

Figure 3&8. Various Proposed Structures for Aluminum (III)
in Absolute Ethanol.

128

Low temperature aluminum-27 nmr might reveal the inequality
among the aluminum ions. Unfortunately, no known studies
have revealed a splitting of the single resonance signal
which is observed at room temperature. This would indicate
that structure A may be incorrect. In addition, another
factor makes this structure improbable. All the bridges
between the aluminum ions are identical. If this were the
actual structure, it would be expected that chelates other
than the 2:1 and 6:1 species would be formed, i;gL, the 3:1,
4:1 and 5:1 chelates. Since this is not the case, it would
appear that all the bridges are not identical and structure
A would have to be disregarded. However, structures B and
C seem to fit all of the criteria. All of the aluminum ions
are environmentally identical, and the bridges alternate
between two different types. These structures seem to be
intuitively sound, since the attack by flavonol to form

the chelates should favor one of the bridge types over the
other. Therefore, either structure B or C could represent

the structure of aluminum ions in absolute ethanol.

Acidic Solutions

 

In acidic solutions, aluminum ions are generally believed
to have all six coordination sites occupied by solvent mol-
ecules because of a shift in the equilibrium given below:

6+ + + +
A12(Et0H)n + [A12(0Et)5(EtOH)n_5] + 5H

129

The extent of polymerization is not known although it is
believed that at least a dimer exists since a 2:1 metal-to-
ligand chelate has been found to exist in acidic absolute

ethanol solutions.

CHELATE STRUCTURES

Basic Solutions

 

As prOposed by Porter and Markham (56) for aluminum-
flavonol chelates in methanol, the structure of the 1:1
chelate in basic absolute ethanol is probably a dihydroxy

complex.

/OH
A1

‘\\\\\OH

This structure is reasonable because the chelate should be

formed from the reaction of flavonate, FlO', and aluminate.
A1(oa)4” + F10- I [A1(OH)2F10] + 20H-

Apparently, flavonol is able to compete effectively with
the hydroxide ions for the coordination sites on the aluminum
ion if the concentration of hydroxide is relatively low.

However, at high concentrations, the equilibrium shown

130

above is shifted to the left and little or no chelate forma-

tion is indicated.

"Neutral" Solutions

 

Since it is logical to assume that the formation of
the 2:1 and 6:1 chelates occurs in a similar manner, the
simplest case, which should be the formation of the 6:1
species, will be considered first. As indicated in the
previous section of this chapter, the point of attack by
flavonol should favor one of the bridge types in the two
proposed aluminum ion structures (Figure 34). From the
potentiometric titration studies, it has been shown that
the formation of the 6:1 chelate occurs with no net change
in the number of free protons. This indicates that there
must be a loss of one ethoxide ion which is converted into

an ethanol molecule. This ethoxide ion is then replaced

[A16(0Et)15(EtOH)n]3+ + 15H+ + FlOH +

3+ +
[A16(0Et)14(F10)(EtOH)n+1] + 15H

by a flavonate ion. Figure 35 shows the possible structures
for the 6:1 chelate based on the proposed aluminum struc-
tures in absolute ethanol. The chelate structures have

been linearized for clarity but there is a possibility

that the 6:1 chelate may still have a cyclic structure.

For aluminum structure B (Figure 34), attack at the bridge

 

 

 

 

131

 

 

type C type d

rAi/krArAjAj/B‘

WCVMv—h v v v \

 

TAT/\T/RTATAT/

Clix/“V“v‘1—x'z‘1v1x

2

 

J

0 = Ethanol 0 s Ethoxide (: I Flavonate

Figure 35. Various Proposed Structures for the 6:1 Aluminum
to Flavonol Chelate in Absolute Ethanol.

132

type a should lead to the structure which is shown in
Figure 35A while attack at bridge type b produces the
chelate structure shown in Figure 35B. Obviously, an at-
tack at either bridge type produces a chelate which fits
the potentiometric titration data. For aluminum structure
C (Figure 34), attack at bridge type 3 should lead to the
chelate structure shown in Figure 35C, while attack at bridge
type d produces the structure in Figure 35D. Again, it
appears that the attack at either bridge type leads to a
chelate which fits the titration data. Consequently,
neither of the two proposed aluminum ion structures can
be eliminated from consideration on the basis of the 6:1
chelate formation.

For the 2:1 chelate, the potentiometric titration
curve had two breaks. The first break occurred at a proton
to aluminum ratio of two and indicates that there are four
free protons in solution after the formation of one chelate
molecule is completed. The second break, which occurs
at a proton to aluminum ratio of 2.5, indicates that one
additional proton is abstracted from the chelate at higher
pH's. Based on these results, the formation of the 2:1
chelate must occur with the production of four free protons,
and in addition, it must have one ethanol molecule within
its coordination sphere which is easily ionized. Figure
36 shows the 2:1 chelate structures which could possibly
be formed from the two proposed aluminum ion structures.

For aluminum structure B (Figure 34), attack by flavonol

133

, 2+

 

 

12+

 

 

 

 

In:
N
...
la.
...
O p E:
<H
I: \.

 

 

12+

 

 

0 = Ethanol o s Ethoxide (Z a Flavonate

Figure 36. Various Proposed Structures for the 2:1 Aluminum
to Flavonol Chelate in Absolute Ethanol.

 

134

at bridge types a and 2 yields the structures which are
shown in Figures 36A and 36B, respectively. For aluminum
structure C (Figure 34), attack by flavonol at the bridge
types 9 and d produces the structures shown in Figures 36C
and 36D, respectively. Note that structures A, C and D

are identical, and it is expected that the proton abstrac-
tion which occurred during the potentiometric titration in-
volves one of the ethanol molecules on the aluminum ion
which is not bound directly to flavonol. The reason behind
this expectation is that this aluminum ion is relatively
more electropositive than the one bound directly to the
flavonate ion with its large resonance system. On the other
hand, structure B in Figure 36 is quite different from the
others. In this case, proton abstraction would probably
occur at the bridging ethanol molecule. Note that this
abstraction results in a molecule which is identical to
structures A, C and D after proton abstraction. In any
case, with the experimental evidence presently available,
it is impossible to predict which bridging type undergoes
attack by flavonol, and in addition, neither of the proposed
aluminum ion structures can be eliminated.

For the 1:1 chelate which is produced from the 2:1
chelate upon aging, the potentiometric titration indicates
that four free protons are present in solution for every
two aluminum ions, and no proton abstraction was observed.
Although no mechanism can be prOposed at this time, the

length of time needed for the conversion indicates that a

135

gross structural change takes place. The proposed struc-
tures for the 1:1 chelate are presented in Figure 37.
Structure A results from the replacement of one ethoxide

ion and one ethanol molecule in structures A, C and D in
Figure 36. Structure B results from the same process for
structure B in the same figure. It is quite reasonable to
expect that structure B for the 1:1 chelate would undergo

a proton exchange and rearrangement for reasons of symmetry.
The resultant structure would then be identical to structure

A.

Acidic Solutions

 

The structures of the chelates in acidic solutions
should be identical to those which are found in "neutral"
absolute ethanol solutions although the compositions change
due to the protonation of the associated ethoxide ions. The
proposed structures for the 1:1 and 2:1 chelates are shown

in Figure 38.

CHELATE QUANTUM EFFICIENCIES

During the course of this study, an interesting trend
was noticed. Apparently, the amount of fluorescence, which
is produced from the excitation of a chelate, is dependent
upon the chelate stoichiometry and its ionic charge. ‘This
correlation has been tabulated in Table V. For the 1:1

chelate, the change in ionic charge seems to have only

 

 

0 = Ethanol

Figure 3?.

 

It”

0 = Ethoxide

 

 

2+

2+

(2 x Flavonate

Various Proposed Structures for the 1:1 Aluminum

to Flavonol Chelate in Absolute Ethanol.

F-

 

-

 

.

Figure 38.

TAT

11v 11

1:1 Chelate

TAM

2:1 Chelate

 

 

h+

- 5+

Proposed Structures for the 1:1 and 2:1 Aluminum

to Flavonol Chelates in Acidic Absolute Ethanol.

137

Table V
Correlation of Quantum Efficiency with

Chelate Stoichiometry and Ionic Charge

 

 

 

Solution . Chelate Ionic Quantum
pH Stoichiometry Charge Efficiency
Basic 1:1 0 0.01
Neutral 1:1 +2 0.01
Neutral 2:1 +2 0.03
Neutral 6:1 +3 0.70
Acidic 1:1 +4 0.03
Acidic 2:1 +5 0.45

 

 

a slight effect on the chelate quantum efficiency. On the
other hand, the quantum efficiency of the 2:1 aluminum-to-
flavonol chelate undergoes a large increase as its ionic
charge changes from +2 to +5.

At this point, it is important to note that even though
there seems to be a correlation between ionic charge and
quantum efficiency, it is actually the number of aluminum
ions in each of the chelates which causes the change in
quantum yield. Consider the chelates which are formed in
”neutral" absolute ethanol. As the ratio of aluminum to
flavonol increases, the quantum efficiencies also undergo

a drastic increase. The same is true for the two chelates

138

which are formed in acidic solutions. This phenomenon may be
explained by considering the structures of the chelates them-
selves.

For the chelates formed in "neutral" ethanol solutions, the
large increase in quantum efficiency between the 2:1 and 6:1
chelates could be caused by an increase in structural rigidity.
This may indicate that the 6:1 chelate has a cyclic structure
similar to that of aluminum (III) in absolute ethanol. On the
other hand, the quantum efficiency difference for the two
chelates formed in acidic solutions is not so easily explained.
Upon the addition of acid, it is expected that the ethoxide
ions involved in the chelate structures should be neutralized.
Consequently, the electropositivity of the aluminum ions would
increase, and this should result in a depletion of electron
density on the attached flavonate ion(s). It would appear
that this change in electron density might account for the
difference in quantum efficiency for the acidic chelates.

Generally, increases in quantum efficiency are attributed
to increases in rigidity, planarity and/or to the raising of
a (n,w*) transition energy relative to a (n,n*) transition.
However, an increase in rigidity or planarity does not seem
to explain the large difference in quantum efficiency for
the acidic chelates. On the other hand, it is possible that
the apparent change in electron density on flavonol does re-
sult in raising the (n,n*) transition energy since some non-
bonding electrons are utilized in the formation of the chelates.
This should decrease the rate of intersystem crossing and

increase the fluorescence efficiency.

139

PREPARATION OF SOLID CHELATES

Throughout this investigation, repeated attempts were
made to obtain the solids of the chelates which presumably
existed in solution. Unfortunately, it seems that all
attempts have ended in failure. In light of the results
which were obtained from the Raman, infrared and aluminum-
27 NMR studies, prospects for success in this venture are
dim since it appears that only one chelate is formed at
high aluminum-flavonol concentrations. However, a brief
outline of the recovery attempts is given below.

Evaporation to dryness of solutions which contained the
appropriate stoichiometric ratios led to the formation of
crystalline powders. The powders obtained from the solu-
tions of the 2:1 and 6:1 chelates gave identical X-ray
powder diffraction patterns. The infrared spectra of these
same powders in KBr pellets also indicated that they were
structurally identical.

Crystallization by the addition of various solvents such
as carbon tetrachloride, chloroform, benzene and cyclo-
hexane to ethanolic solutions of the chelates proved un-
successful. However, crystals were obtained over a period
of four days by allowing ethyl ether to diffuse into ethan-
olic chelate solutions. Unfortunately, the crystals were
too small and irregular for a X-ray crystallographic struc-
ture analysis. In addition, it is believed that the chelate,
which was formed by the above procedure, was identical to that

of the powders which were obtained through evaporation.

IV. SUMMARY AND CONCLUSIONS

From photometric and fluorometric spectral studies,
a new electronic absorption band in the spectrum of flavonol
has been discovered. The band is located in the spectral
region near 410 nm, and excitation at this wavelength re-
sults in the emission of fluorescence at 484 nm. Tentatively,
this electronic absorption band has been assigned to a
(n,n*) transition although the experimental results are
somewhat inconclusive. Further, support for this assign-
ment might be obtained by the conversion of flavonol'to
the corresponding oxime (62). If the absorption band dis-
appears from the absorption spectrum of the oxime, there
is a strong indication that the band is caused by a (n,n*)
transition. Unfortunately, this method of assignment has
only been shown to be reliable for nitrogen heterocyclic
compounds and consequently, further tests with oxygen
heterocycles would have to be carried out before a definite
assignment could be made. In addition, the conversion of
flavonol to its oxime in concentrated alkaline medium would
have to be modified since flavonol is unstable in concen-
trated base.

Based on the work of various investigators (55,59-61),
numerous aluminum-27 nmr studies, and the results of this
study, various structures for aluminum (III) in absolute

ethanol have been proposed. Of the three presented, all

140

141

are hexameric cyclic structures, and only one of these could
be eliminated on the basis of the available experimental
evidence. The two remaining structures (Figure 34, B and

C) seem to fit the experimental criteria of identical en-
vironments for all of the aluminum ions, an ethoxide to
aluminum (III) ratio of 2.5, and an alternating bridge con-
figuration.

The work of Urbach and Timnick (53,54) has been re-
examined in detail. The formation of a 1:1 chelate in
dilute base and the 1:1, 2:1 and 6:1 aluminum-to-flavonol
chelates in "neutral" absolute ethanol solutions has been
confirmed. In addition, this study has also shown that 1:1
and 2:1 aluminum-to-flavonol chelates also exist in acidic
ethanol solutions. Based upon the proposed aluminum (III)
structures in absolute ethanol and a series of potentiometric
titrations, various structures for these chelates have been
proposed. Although the 6:1 aluminum-to-flavonol chelate
in "neutral" absolute ethanol has been pictured in a
linear form, there is no experimental evidence to warrant
this representation. Indeed, there exists a good possibility
that this chelate actually has a cyclic structure which is
similar to that of aluminum (III) in absolute ethanol. As
far as the structures for the other chelates are concerned,
some are similar or identical to the ones proposed by Porter
and Markham (56) for the aluminum-flavonol chelates in

absolute methanol.

142

Fluorescence quantum efficiency measurements on the
chelates in solution have revealed an interesting trend.
It appears that as the number of aluminum ions in the chelate
structures increases, the quantum efficiencies of the chelates
also increase. For example, in "neutral" ethanolic solu-
tions, the existence of three chelates is indicated. These
are the 1:1, 2:1 and 6:1 aluminum-to—flavonol chelates. The
quantum efficiency of the 2:1 chelate is about three times
larger than that of the 1:1 chelate, and the 6:1 species
exhibits an efficiency that is about seventy times as large.
If the 6:1 chelate has a cyclic structure, the increase in
rigidity over the 1:1 and 2:1 chelates should account for
the large increase in quantum yield. An increase in quantum
efficiency can also be observed in the 1:1 and 2:1 chelates
which are formed in acidic ethanol solutions. There is ap-
proximately a fifteen fold increase in quantum efficiency
‘for the 2:1 chelate over the 1:1 species. This effect can-
not be explained by an increase in rigidity or planarity.
However, it might be explained by a decrease in the rate of
intersystem crossing caused by raising the energy of the
(n,n*) transition relative to the (n,w*) transition of the
chelates. This rise in energy could be caused by a depletion
of non-bonding electron density on the chelate flavonate
ion(s), which is caused by an increase in the electroposi-
tivity of the chelated aluminum ions. Since this effect
should be much larger in the 2:1 chelate, a large increase

in quantum efficiency over the 1:1 chelate is expected.

143

The prospects for obtaining the solids of the chelates,
which presumably exist in dilute absolute ethanol solutions,
appear to be quite small. Crystallization by the addition
of various solvents to the ethanolic chelate solutions has
proved unsuccessful. However, evaporation of the solvent
from the same ethanolic solutions did yield crystalline
powders. Unfortunately, the crystals were unsuitable for
an X-ray crystallographic structure analysis. Studies by
using Raman, infrared and aluminum-27 nmr spectroscopy have
all indicated that only one chelate is formed in concentrated
solutions and in the solid state. Comparison of the solid
state infrared spectra, which were obtained during the course
of this study, with the infrared spectrum of the solid che—
late obtained by Urbach, indicates that the solids in both
studies might be the same. Consequently, it is assumed that
the solid chelate obtained in this study probably has an

aluminum-to-flavonol ratio of 1:3.

APPENDICES

APPENDIX ONE

MATHEMATICAL PROOF OF ASSUMPTION FOUR

Assumption four states that the ratio of the fluoro-
phore absorbance, Af, to the total absorbance, At, in the
cell is equal to the ratio of the quanta, of, absorbed by

the fluorophore to the total quanta absorbed, Qt’

The mathematical proof of this statement is presented by
the use of the following approach. The cell is assumed to
contain a fluorophore and a constant number of chromo-
phores. The cell is then divided in a fashion such that
the fluorophore and the chromophores are segregated into

alternating equal volume slices as shown below:

A 0....

 

 

 

 

 

 

 

 

 

 

 

.A constant fraction, l/n, of the fluorophore is contained

in the first slice. The second slice contains the same

144

145

fraction of the chromophores while the third slice contains
l/n of the fluorophore and so on.

At this point, the following definitions are made:

(1) Af is the total absorbance due to the fluorophore.
(2) AC is the total absorbance due to the chromophores.

(3) At is the total absorbance, Af + Ac.

(4) The total number of cell slices is equal to Zn.

Consequently, the absorbance of the fluorophore in each of
the odd numbered slices is Af/n, while the absorbance for
the chromophores in each of the even numbered slices is
AC/n. From the general expression for the exponential
attenuation of the excitation beam, the following expression

may be written:
I = I e = Ice (1)

where a, b, c and A have the usual Beer's Law meanings.
From Equation 1, it follows that the intensity of the ex—

citation beam after passing through the first slice is

_ -(A /n)
I - Ioe f (2)

and I0 - 11' is the number of quanta absorbed by the fluoro-
phore in the first slice. The total number of quanta
absorbed by the fluorophore, Qf, across the whole cell is

given by the following expression

146

n
of = Io Z [e-(Q‘1)(At/n) _ e((Q’1)(At/n) + Af)/n] (3)

q 1

By factoring Equation 3, the following expression is obtained

n
of = 10 (1-e“Af/"’) (can‘t/’1’) f e‘At/“qu (4)
1

The total quanta absorbed within the cell is given by an

expression which is similar to Equation 3

n
Qt _ I X [e-(q-1)(At/n) _ e-q(At/n)] (5)

o q=l

By factoring, Equation 5 yields

<2t = 10(e(At/n) - 1) } e‘q(At/n’dq (6)
1

Consequently,

:2}.- - (1 _ e-(Af/n)) (amt/n),
Q elAt/n) _ 1

 

(7)
t

It should be noted at this point that as n approaches in-

finity, the cell model used in this proof approaches reality

and since

lim e(At/n) = 1

“+00

then,

147

Q _ -(Af/n)
—£ = 1 e (8)
t e (At/n) _ l

 

0

Unfortunately, Equation 8 cannot be evaluated in this form
because the denominator goes to zero as n approaches in-
finity. To circumvent this difficulty, Equation 8 can be
expressed as the series expansion of ex. Expansion leads

to the following expressions

2
(A /n)
e-(Af/m . 1 — (Af/n) + -£-—- -
2!
and
(A /n) (At/“)2
et =1+(At/n)+————+...

2!

Consequently, Equation 8 may now be written as

Qf (Af/n)2
_. = [(Af/n) + —— - ...]/[(At/n) +
Q 2!
t
(At/n)2
_______.+ ...] (9)
2!

As n approaches infinity and for reasonable values of At,it is
easily recognized that all of the terms in each series go to

zero infinitely faster than the first term. Therefore,

and assumption four has been shown to be valid.

APPENDIX TWO
BUFFERED DIGITAL I/O PIN CONFIGURATIONS

(DEC DR8-EA)

INPUT

(1) A l + 0 transition at any input bit leaves a l
in the corresponding accumulator bit.

(2) After clearing the input register, reading the
register into the accumulator leaves the accumulator

a logical 0.

OUTPUT

(1) Logical 1's in the accumulator give logical 0's

at the corresponding output bits.

ASSIGNMENTS
B_I_'£ am
0 ’ V2
l 81
2 T2
3 P1
4 82
5 Ml
6 P2
7 L1

148

149

a 2.13.
8 M2
9 J1

10 K2

11 H1

The assignments given above are for both the input and out-
put registers. The electrical ground for both registers

has been assigned to pin F2.

APPENDIX THREE
INSTRUCTIONS FOR THE CONSTRUCTION AND IMPLEMENTATION

OF AN EMISSION DETECTION SYSTEM CORRECTION TABLE

The emission system correction table is normally used
in conjunction with the fluorimeter computer program, FLUORO,
to correct the measured fluorescence for changes in sensitivity,
as a function of wavelength, of the emission optical system
and photomultiplier tube. The correction table resides in
field one of core memory, starts at address 40008 and can
extend up to address 47008. Consequently, there is room for
512 correction factors. The table consists of integer factors
which can range from 0 to 2047 and are stored as a function
of wavelength. Location 40008 contains the correction factor
for a wavelength of 200 nm, location 40018 contains the factor
for 201 nm, etc. Stored in this fashion, there is enough
room for all the correction factors from 200 to 700 nm.

The general process for the construction of the correc-
tion table is fairly simple. The process entails the compari—
son of the emission system response with that of a linear
detector. In this case, the linear detector is a quantum
counter. The comparison is made by placing a reflectance
plate in the sample position so that the excitation beam is
directed through the emission detection system when the
vibrating bridge is in the sample position. The true intensity

of the excitation beam is then measured by the quantum counter

150

151

when the vibrating bridge is in the reference position.

The ratio of the reference signal to the emission signal,
R/F, is then the correction factor for a particular wave-
length. The factors for the entire wavelength range are
obtained by scanning the emission and excitation monochrom-
ators in tandem and by measuring and outputting the ratio,
R/F, at one nanometer intervals.

Unfortunately, the use of just one quantum counter over
the wavelength range of 200 to 700 nm is not possible. Since
the rhodamine B quantum counter is only useful from 250 to
about 600 nm, it must be used in conjunction with a methylene
blue counter in order to cover the entire wavelength region.
Fortunately, there is a considerable overlap in the useful
wavelength ranges of these two counters. Since with the
proper filter, the methylene blue counter is useful in the
range from 550 to 700 nm, the overlap between 550 and 600
nm can be used to normalize the correction factors obtained

from the two counters.

BUILDING THE CORRECTION TABLE

Rhodamine B Section

(1) Load the main fluorometer program into the com-
puter and set up the instrument for a normal excita-
tion scan. A one nanometer data interval should

be used.

 

 

(2)

(3)

(4)

(5)

(6)

(7)

(8)

152

Wire the excitation and emission monochromators in
parallel so that they scan in tandem.

Place the reflectance plate, which has been re-
cently coated with Kodak's White Reflectance Stan-
dard, in the sample cell position.

Manually scan both monochromators to assure that
the reference and emission signals do not saturate
the analog-to-digital converter.

Reset the monochromators and scan from 200 to 610
nm. Note that the fluorometer program requires that
the scan be started at 200 nm even though meaningful
correction factors are obtained only from 250 to
610 nm.

At the end of the scan, make sure that the absorp-
tion and photomultiplier corrections are inactive.
To output the correction factors, the user types
PUNCH: R/F and the teletype responds with TURN ON
THE PUNCH. Once the paper tape punch is ready,
the user types a SPACE and the R/F values as a
function of wavelength are punched out.

Steps 5 through 7 are then repeated four addi-
tional times so that five individual paper tapes
are generated. These tapes will be processed at

a later time.

153

Methylene Blue Section

(1) The rhodamine B quantum counter is removed from
the instrument, and the rhodamine B solution and
the 610 nm sharp cut-off filter are replaced with
the methylene blue solution and a 700 nm sharp
cut-off filter. The quantum counter is then placed
back in the instrument.

(2) Steps 4 through 8 in the rhodamine B section are
then repeated except that the wavelength scans

are now from 550 to 700 nm.

PROGRAM PMT

The paper tapes which were obtained from the fluorometer
scans are processed by a FORTRAN program named PMT. This
program serves several functions. First, it enables the
user to average the multiple scans for each quantum counter
in order to minimize the random errors in the correction
table. Second, the correction factors may be scaled so that
the full integer range may be used and so that the tables
from the two quantum counters can be normalized and finally
merged. Finally, program PMT allows the user to punch out
the factors in a form which is compatible with PAL8 assembly.
A program listing and flowchart appear on the following

pages.

(0

Wk

28

38

27

13

ll
12

154

PROGRAM PMT

DIMENSION ARRAY(SOO)J N(!25)a LCIO)

R 3 I-

WRITE(I:2)

FOXMATC/I: 'TYPE IN THE BEGINNING AND ENDING WAVELENGTH IN NM')
READCIJS) ISTART

3EAD(133) IEND

FORMATCIIO)

LENGTH = IEND ' ISTART + 1

NUMBER 3 LENGTH/4

LOGOS 3 NUMBER * 4

LOG = LOGOS * I

DO 4 I = I: LENGTH

ARRAY(I) 3 @-

CONTINUE

J = I

NRITECIJQOJ

FORMAT(//J'PLACE THE TAPE IN THE READER AND TURN THE READER ON')
READCI:33) Z

FORMAT(IAG)

READ(I:6) A: B) C: D: E: F: G

FORMATC7A6)

READ(I:7) H: O: P: Q

FORMATCQAO)

DO 9 I 3 I) LOGOS: 4

A! A3RAY(I)

A2 ARRAYCI+I)

A3 ARRAY(I+2)

94 = ARRAY(I*3)

READCIae) NCJ): ARRAYCI): ARRAY<I+I>1 ARRAY‘I+2)' ARRAY(I*3)
FORMAT<II4¢4E1607)

ARRAYCI) = ARRAYCI) + A!

ARRAY(I+I) = ARRAY(I*!) + A2

ARRAY(I+2) = ARRAYCI+2> + A3

ARRAY(I*3) 3 ARRAY(I+3) + A4

J = J*!

CONTINUE

IF (LENGTH ‘ LOGOS) 11:11:27

AI 3 ARRAY<LOG)

82 = ARRAY(LOG+I)

A3 3 ARRAYCLOG+2)

REAO(IJIO) N(J)o (ARRAYCI): I 3 L03: LENGTH)
FORMAT‘IIA: 3EI607)

ARRAYCLOG) 3 ARRAY‘LOO) * AI

ARRAY(LOG*I) ARRAYCLOG+I> + AZ

ARRAY<LOG+2) ARRAY‘LOG+2) * A3

WRITECIJIQ)

FORMAT(//: 'TYPE IN THE RUN MODE! 'o/: 'I 3 DATA INPUT? /o
1'2 3 PRINT OUT AVERAGED SCALED DATA': /: '3 3 PUNCH OUT
ZDATA': /: '4 = RESTART THE PROGRAM': /)

13

14

15

16

17

18

19
28

21
22

39

23
34

35

41

36

24

37

25

26
38

155

READ(1313) IANS

FORMAT(111)

GO TO (14329329338) IANS
R=R*1o

GO TO 5

DO 15 1 3 13 LENGTH

ARRAY(1) 3 ARRAY(1)/R
CONTINUE

WRITEC1316)

FORMAT(//3 'TYPE IN THE SCALE FACTOR'3 /)
READ<1317) SCALE
FORMATC1F1405)

DO 18 I 3 13 LENGTH

ARRAY(I) 3 ARRAYCI) * SCALE
CONTINUE

RSCALE 3 SCALE

GO TO (21321319338) IANS
”3173(1323)

FORMAT(//3 'PRESS CONTINUE ON THE CONSOLE WHEN YOU HAVE THE
1PUNCH READY'3 /)

PAUSE

CALL OPEN

WRITEC1322)

FORMATC/3 '*4339'3 ’3 'DECIMAL')
NOW 3 (LENGTH/IO) * 1O

NUMI 3 LENGTH ' NUM

DO 34 1 3 13 NUM3 13

M 3 I

DO 39 J1 3 13 10

LCJ1) 3 ARRAY(M)

M 3 M + 1

CONTINUE

WRITE(1323) (L(N1)3 N1 3 13 10)
FORMAT(9(1I43'3'31X)3 114)
CONTINUE

1F (NUMl) 36336335

M 3 NON * 1

DO 41 1 3 13 NUM1

LCI) 3 ARRAY(M)

M 3 M + 1

CONTINUE

WRITEC1323) (LCN1)3 N1 3 13 NUM1)
W31TE<1324)

FORMATC/3 '$')

DO 37 1 313 LENGTH

ARRAY(I) 3 ARRAYCI) * R/SCALE
CONTINUE

GO TO (26326325338) IANS
PAUSE

CALL OPEN

GO TO 11

GO TO 1

CONTINUE

END

156

PROGRAM PMT FLOWCHART

( START )

INPUT
AVELENGTH
RANGE

 

 

 

 

INPUT
ORRECTION
FACTORS

1
AVERAGE
INPUT

 

 

 

 

 

 

 

 

 

 

 

 

 

  

INPUT
SCALE
FACTOR

      

 

Figure 39. Program PMT Flowchart.

157

DATA PROCESSING

The paper tapes from the methylene blue quantum counter

are processed first. This is because the methylene blue

table will contain the largest correction factor.

(1)
(2)

(3)

(4)

(5)

Load program PMT into the computer.

The teletype will respond with TYPE IN THE BEGINNING
AND ENDING WAVELENGTH IN NM. The user then types
in the appropriate information in integer format.
The teletype will then respond with PLACE THE

TAPE IN THE READER AND TURN THE READER 0N.
Subsequently, one of the five paper tapes for the
particular counter which is being processed, is
read into the program. Although the program reads
the heading at the beginning of each paper tape,
this information is not used.

The teletype then responds with:

TYPE IN THE RUN MODE:

1 = DATA INPUT

2 = PRINT OUT AVERAGED SCALED DATA
3 = PUNCH OUT DATA

4 = RESTART THE PROGRAM

If more tapes are to be read into the program, the
user types a 1. Steps 3 through 5 are then re-
peated until all the tapes for a particular quantum

counter have been read.

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

158

At this point, the user should type a 2 in response
to the alternatives given above.

The teletype will then respond with TYPE IN THE
SCALE FACTOR. The user should now locate the
largest correction factor and calculate the scale
factor which, when multiplied by this largest
factor, gives a number of about 2000.

The user then enters this scale factor into the
program and the scaled correction table is printed
out.

When the teletype responds with the list of
alternatives again, the user types a 3.

Again, the teletype responds as in step 7 and the
user enters the calculated scale factor.

The teletype will then type out PRESS CONTINUE

ON THE CONSOLE WHEN YOU HAVE THE PUNCH READY.

Once the CONTINUE switch is pressed, the PAL8
compatible correction table will be punched out

on paper tape. This tape is saved for final
editing and assembly.

Once the punching process is completed, the CONTINUE
switch is pressed again, and the teletype will
print out the alternatives. The user then chooses
alternative 4, and the program is restarted and
ready for more input.

The rhodamine B correction table is processed in

a similar manner except for steps 7 and 8. For

(15)

159

step 7, the user should type in the same scale
factor which was used for the final output of the
methylene blue correction factors. For the methylene
blue counter, all of the correction factors between
570 and 590 nm are averaged. The same is done for
the rhodamine B correction factors which were
obtained from step 7. The final scaling factor
for the rhodamine B table is merely the ratio of
the methylene blue average to the rhodamine B
average, multiplied by the scale factor used in
step 7.
The two PAL8 compatible paper tapes are then read
into the 08/8 EDITOR. The rhodamine B table is
read in first, then the other table. The two tables
are then merged at a point in the 570 to 590 nm
interval where two of the correction factors are
close to being identical. The overall correction
table should be edited until the following form is
obtained:

*4000

DECIMAL

table of correction factors

(16)

(17)

(18)

160

The file produced from EDITOR is then assembled

by the PAL8 assembler into a file named PMT.BN.
This file is loaded into field one when the main
FLUORO program is processed into an OS/8 SAVE file.
The effectiveness of the completed correction

table can be tested by running a tandem scan for
each quantum counter. If the table has been con-
structed properly, the ratio of the reference signal
to the emission signal, R/F, when plotted as a
function of wavelength, should be a straight hori-
zontal line. Obviously, the photomultiplier cor-
rection routine should be operative for these tests.
Once the tests are completed, the rhodamine B quan-

tum counter should be installed for routine use.

 

APPENDIX FOUR
INSTRUCTIONS FOR THE IMPLEMENTATION OF

ABSORPTION-CORRECTED FLUORESCENCE MEASUREMENTS

As indicated in the Theoretical chapter of this thesis,
the absorption corrections on source-corrected fluorescence
measurements are dependent upon the size of the fluores-
cence observation window and the geometry of the emission
Optical system. Consequently, absorption-corrected fluores-
cence measurements are only possible if the geometry of
the optical system is not altered and the size of the
observation window can be determined. Since the geometry
of the optical system will not be changed once it is
installed in the instrument, the only impediment to the
implementation of routine absorption-corrected fluores-
cence measurements is the determination of the window size.
Unfortunately, this process is somewhat tedious, but happily
it is also a one time venture.

Since absorption-corrected fluorescence deals with
pure fluorophore solutions as well as mixtures of fluoro-
phores and chromophores, these types of chemical systems
are used in the determination of the observation window
dimensions. The system of choice for the pure fluorophore
solutions is quinine sulfate in either 1.0 or 0.1 N sulfuric
acid. This system was chosen because quinine sulfate is

stable in solution over long periods of time; it can be

161

162

obtained in high purity; it has a constant quantum effic-
iency and there is little overlap between the emission and
absorption spectra.

For the mixtures of fluorophores and chromophores,
several systems have been found to be satisfactory. Of
these, the best are quinine sulfate and either 2-thiouraci1
or 2,5-dihydroxybenzoic acid in 0.1 N sulfuric acid. The
2-thiouraci1 is the best chromophore from the standpoint of
stability in solution but decomposition in the solid state
is a problem. On the other hand, 2,5-dihydroxybenzoic acid
is somewhat unstable in solution, and lifetimes of about
one week can be expected. The acid also has one more minor
drawback in that it tends to dimerize in concentrated solu-
tions, but the problem is not severe enough to warrant its
dismissal as a test chromophore.

In any case, at least two series of solutions should
be prepared for test purposes. The pure fluorophore series
should range in concentration so that the solution absor-
bances go from about 0.005 up to about 2.0. In the case of
quinine sulfate where the fluorescence at 450 nm is excited
at 312 nm, the solutions should range from about 1 x 10-6
to about 3.1 x 10"4 M. In the case of the fluorophore-
chromophore series, the first solution, which is pure
fluorophore, should have an absorbance of at least 0.1 so
that a reasonable amount of fluorescence is generated.

Obviously, this fluorescence is going to be severely

163

attenuated by the presence of large amounts of chrom0phore.
Consequently, there must be enough fluorophore present in
order to generate a quantity of fluorescence which can be
accurately measured even under conditions of severe attenua-
tion. The concentration for quinine sulfate chosen for this
study was 1 x 10'5 M but more concentrated solutions could
be used. Whatever the concentration chosen, it should be
the same for the whole series. The chromophore concentra-
tions should range so that the solution absorbances, which
do not include the fluorophore absorbance, run from about
about one-fourth that of the fluorophore all the way to
about 2.0. The chromophore absorbances in this study
ranged from 0.013 to about 1.85.

Once all of the solutions have been prepared, the source-
corrected fluorescence and transmittance of each solution
are measured. These data are then processed through a series

of FORTRAN programs.

PROGRAM LSSQ

Program LSSQ is a utility program which is divided into
two sections. The first section, which is connected with
run mode alternatives 1 and 2, is used to prepare the
source—corrected fluorescence and transmittance data for
input to the fitting program, RTFACT. The second section,
which is associated with alternatives 3 and 4, accepts the

window dimensions obtained from RTFACT as input. These

164

dimensions are then used in conjunction with the experimental

data in order to calculate the absorption-corrected fluores-

cence for each solution in the series. All the information

is then output in tabular form. This table may then be

used to determine whether the absorption-corrected fluores—

cence is linear with the absorbance.

Program LSSQ is used in the following manner:

Section One

 

(1)

(2)

(3)

The program is loaded into the computer and the
teletype immediately responds with

TYPE IN THE RUN MODE:

1 = LEAST SQUARES FIT FOR A PURE FLUOROPHORE

2 = FIT FOR A FLUOROPHORE AND CHROMOPHORE MIXTURE

3 = CALCULATION ROUTINE FOR A PURE FLUOROPHORE

4 = CALCULATION ROUTINE FOR A MIXTURE

The user then enters a 1 if the data are for a
pure fluorophore or a 2 if they are for a mixture.
In both cases, the teletype prints TYPE IN THE
NUMBER OF DATA POINTS AND YOUR DATA. The user;
then enters the appropriate information. An integer
format is used for the number of data points where-
as the transmittance and fluorescence data are
entered as real numbers. The data are arranged

in the following order:

(4)

(5)

(6)

165

transmittance /first solution
source-corrected fluorescence
transmittance /second solution

source-corrected fluorescence

etc.
The data for the least absorbant solution should
always appear first and the rest of the data are
ordered sequentially in increasing absorbance.
If the user chose the first option, the teletype
responds with TYPE IN THE NUMBER OF DATA POINTS TO
BE USED IN THE LEAST SQUARES FIT. The user then
enters in the number of pure fluorophore solutions
which have an absorbance under 0.05.
The teletype then prints out TYPE IN THE ESTIMATED
W1 AND W2, RESPECTIVELY. The required information
is entered and the program then calculates the ab-
sorption-corrected fluorescence for the solutions
to be used in the fit. The equation of the best
straight line is then calculated, and the program
prints out the values for the slope and intercept.
This equation is then used to calculate the absorp-
tion-corrected fluorescence for the remaining solu-
tions in the series.
If Option 2 was chosen, the teletype prints out
TYPE IN THE ABSORPTION-CORRECTED FLUORESCENCE FOR
THE PURE FLUOROPHORE. This value can be calculated

from the window dimensions which were obtained

(7)

166

from RTFACT for the series of pure fluorophore solu-
tions. The user then types in this fluorescence
value.

At this point, the teletype responds to options 1
and 2 with PRESS CONTINUE ON THE CONSOLE WHEN YOU
HAVE THE PUNCH READY. The user takes the ap-
propriate action and presses the switch. The program
then punches out the data in a form which is com-
patible with the input format of RTFACT. When the
punching process is completed, the program pauses
and the tape can be removed from the punch. The
user then presses the CONTINUE switch to restart

the program.

Section Two

 

(8)

(9)

(10)

(11)

Upon loading or restarting the program, the alter-
natives are typed out as in step one.

The user then enters a 3 if the calculations are

to be performed for a pure fluorophore series or

a 4 if they are for a mixture series.

The teletype then responds as in step three and
again, the user supplies the needed input.

If alternative 3 was chosen, the program follows

the same dialogue as outlined in steps three through
five. If alternative 4 was chosen, the dialogue

is the same as in step six.

167

(12) At this point, the teletype responds to options 3
and 4 with TYPE IN W1 and W2, RESPECTIVELY. The
user then enters the observation window dimensions
which were obtained from RTFACT.

(13) The program will next type out a table which con-
sists of four pieces of information for each solu-
tion in a particular series: the absorbance, the
source-corrected fluorescence, the theoretical
absorption-corrected fluorescence used in the
RTFACT fit, and the calculated absorption-corrected
fluorescence which was obtained by using the window
dimensions output by RTFACT.

(14) The user may now plot the calculated absorption-
corrected fluorescences against the solution
absorbances to determine whether the window dimen-

sions are satisfactory.

A program listing as well as a flowchart for LSSQ appear

on the following pages.

PROGRAM RTFACT

Program RTFACT is a program based on a modified Simplex
Optimization technique. This program fits the experimental
source-corrected fluorescence measurements to the LSSQ
calculated absorption-corrected fluorescences by varying

the window parameters, w1 and w. When a fit has been

32

31
I?

13

35

168

PROGRAM LSSQ

DIMENSION XAPPAY(IEO)1 YARP-AYCIOO): ZAPPAY(IOO): I-‘AEPA‘.’(IO¢)
DIMENSION OAPPAV(39)

SUHY‘.’ 3 7.

SUN)( 3 fl.

SUM" 3 90

SUMVZ 3 0‘.

WPITEC I: I)

FOPMATCI/a 'TVPE IN THE RUN MODE:':/a 'I 3 LEAST SOUAPES FIT
IFOP. A FUPE FLUOPOPHOPE'./.'2 3 FIT FOF‘A FLUOPOPHOPE AND
ZCHPOMOPHOPE MIYTURE'3/p '3 3 CALCULATION ROUTINE PO” A PUPE
3FLUOPOT’HOPE'J/a'4 3' CALCULATION ROUTINE FOP A MIXT‘UPE'a/)
PEAD(I:2) INPUT

YOPMAT(III)

VPITE(!.3)

FOPMAT(//a 'TVPE IN THE NUMBEP OF DATA POINTS AND YOUI5 DATA.) /)
PEADCIal‘) INTEP

EORMATTIIOJ

DO 6 I 3 I: INTER

PEAD(I:5) YAPPAVCI)

PEAD(I:5) ZAEPAVCI)

FOPMAT(1FIG.A>

CONTINUE .

GO TO (III7JIII7) INPUT

‘I'PITEII:8)

EOPMATT/l: 'TVPE IN THE ABSOPEANCE COPPECTED FLUOPESCENCE F09
ITI‘IE PURE FLUOPOPHOPE':/)

PEADCIag) PUPE

FOP"AT(IFIO¢4)

DO I? I 3 I: INTER

YAPPAYCI) 3 PUDE

GO TO (.IEII91I9131)'INPUT

WAPRAY(I) 3 ‘(ALOGCXAPPAV(I)))/2039258

CONTINUE

GO TO 16

VPITECI1I2)

FORMA‘N/la ""YPE IN THE NUMBER OF DATA POINTS TO BE USED IN
ITHE LEAST SQUARES FIT'al)

READC I: 13) NHIEST

FOPMATCIIa)

VT’ITEC I: 35)

FORMAT((:'TVT"E IN THE ESTIMATED VI AND ‘Izl PESPECTIVELVIJI)
PEAD<I1 27) 1‘33

PEAD(I:27) I34

DO 14 I 3 I: NHIEST

QAPPAVCI) 3 KAPPAYCI)*(ALOG(XARPAY(I3))3(V4-V3)/((XAEPAVCI)**
IVA) ' (YAHPAVCI)**V3))

14

IS
16

I7
18

I9
2?

23
24

25
26

27

28

29
38

169

x a -(ALOG(YAPPAV(I))/2.32258)
sunxv a SUMYY + (OAPBAYCI) * x:

sumx - SUMY + x

SUM? . SUM? + CARRAVCI)

SUMXE 8 SUMV2 + Yttz.

CONTINUE

HIEST . NHIES?

SLOPS . ((HIEST*SUMXV) - (SUMYtvaV))/((HIEST*SUMN2) -
1(SUM¥**2.))

PINTER a ((SUMY*SUMX2)-(SUMN*SUMXY))/((HIEST*SUMY2)-(SUMX#*2.))
DO 16 I a 1. INTEP

VAPPAV(I) . SLOPE * (~(ALOG(XAPPAY(I)))/2.32258) + PINTSP
GO TO (16.16.15.15) INPUT

VARPAY<I> = ~(ALOG(XAPRAY(I)))/2.30258

CONTINUE

GO TO (17.19.17.25> INPUT

wa'rSu. 18) SLOPE. RIN‘Y‘ER

FOPMAT(//. ‘SLOPE . '. 1813.6. /. 'INTEPCEPT . '. 1313.6)
GO TO (19.19.25.25) INPUT

PPIT£(1.2O)

roanarcxx. 'PPESS CONTINUE ON THE CONSOLE VHEN.YOU HAVE THE
lPUNCH READV'./)

PAUSE

CALL OPEN

D0 24 I = 1. INTEF

vaTS<I.23) YAPPAY(I). YARPAYtI). ZAPPAY<I)'
FOPMAT<2<IF1€.4. I). xrne.a>

CONTINUE

GO TO 30

NEITE(1.26)

FORMAT(//. 'TYPE IN v1 AND v2. RESPECTIVELY'. I)
PEAO<1.27> v1

PEAD(1.27) re

rOPMAT<1r1P.a>

PPIT£<1.28)

IOPMAT(//.'ABSORBANCE'oGX.'FLUORESCENCE':6X:'THEOR-AB-FLUOP':
IOX.'CALC-AB-FLUOR')

DO 39 I 8 I: INTER

A 8 ZARPAY(I)*(ALOG(¥APPAV(I)))*(V2-VI)/((XARPAYCI)**V2)
l-(XARPAY( I )IHII'I ))

VPITE(I:29) VAPPAY(I): TARRAY(I)o VAPPAY(
FOPMATCI: IflfoaaTXaIF!€-4a9XalFIO-4:9X:I
CONTINUE

PAUSE

CALL OPEN

GO TO 82

END

I); A
FIE-4)

170

PROGRAM LSSQ FLOWCHART

 

 

 

 

 

 

   
 

MIXTURE

 

 

CALCULATE
ABSORPTION
CORRECTED
FLUORESCENCE

 

 

 

 

I

 

 

 

 

 

 

 

CALCULATE
ABSORPTION
CORRECTED
FLUORESCENCE

 

 

 

 

I

PRINT
OUT
TABLE

 

 

 

 

Figure no. Program LSSQ Flowchart.

171

obtained within the specified convergence limit, the program
outputs the window parameters as well as a value for the
residual which is a measure of the exactness of the fit.
Consequently, a low residual indicates that a good fit

has been obtained. The input format for program RTFACT

is given below:

1 /uneven data intervals
20 /number of data points
transmittance /output from LSSQ

absorption-corrected fluorescence
source-corrected fluorescence

2 /number of variables
.100 /estimate of w

.800 /estimate of w

.01 /convergence limit of 1%

Once convergence has been attained, the window parameters,
wl and w, are converted to w1 and w2 which are quantities
utilized by programs LSSQ and ARTCAL. The program listing
for RTFACT appears on the following pages. No flowchart

is supplied because the program is well commented.

PROGRAM ARTCAL

Once the user is satisfied with the values for the
observation window parameters, they must be placed in
the FLUORO program subroutine which calculates the absorption-
correction factor, fa. Since this subroutine calculates

fa by using the first seven terms of a Taylor expansion,

UIUCIO C10

4()C)O OCT

u-ITIOCD
I-Q

I2

172

PROGRAM RTFACT

DIMENSION XARRAY(I@O): YARPAY(IOO). GUESS(IO:9); RSDL<IO)
DIMENSION H(102o BADCIO): OLD(9): PRES(9): RNEUCSa9). RENEV(S)
DIMENSION FINDCIO): ZARRAYCIOO)

KL I l

ITT I 2

NE = O

VATCI'I 3 '0

TYPE IN THE KIND OF INPUT DATA: A 2 FOR DATA TAKEN AT
EQUAL INTERVAL53 A I FOR DATA TAKEN AT UNEOUAL INTERVALS
READCIaI) INPUT

FORMAT(III)

IF (INPUT - 0) 2.2.6

TYPE IN THE EQUAL INTERVAL DATA INPUT PARAMETERS: FIRST AND
LAST INDEPENDENT VARIABLES: AND DATA INTERVAL

READCla3) START. FINISH: XINTER

FORMAT(3(IFIOo3))

INTER I (FINISH - START + XINTER)/XINTEP

FILL I START

LOAD THE INDEPENDENT AND DEPENDmT VARIABLES IN XARRAY AND
YARRAY; RESPECTIVELY

DO A I I I: INTER

XARRAY(I) I FILL

FILL I FILL + XINTER

CONTINUE

READ(IoS) (YARRAY(I): I I I. INTER)

FOPMATSIFlflo3)

GO TO IO

TYPE IN THE NUMBER OF VARIABLES AND LOAD THE INDEPENDENT
AND DEPENDENT VARIABLES IN XARRAY AND YARRAY. RESPECTIVELY
READ‘Ia7) INTER

FORMATCIIIO)

DO 9 I I I: INTER

READ<Ia8) XARPAY(I)

READ(I:8) YARRAYCI)

PEAO<1.O).2APPAY<I)

FORMAT(IFIO.4)

CONTINUE

TYPE IN THE NUMBER OF UNKNOUNS

READCIoII) IUNKN

FORMAT<1II)

RIUNKN I IUNKN

IUNKI I IUNKN + I

TYPE IN THE INITIAL ESTINATES FOR THE UNKNOVNS'

READCIOIE) (GUESS(IJI)J I P I: IUNKN)

FORMAT¢IFIOo3)

93

13

14
15
16

17

18
19
20
21

22

23

24

000010

26'
27
28

173

TYPE IN THE CONVERGENCE LIMIT

READ(1:93) OED

FORMATIIFIOoQ)

CONSTRUCT THE INITIAL SIMPLEX

DO 16 I 3 I: IUNKN

DO I5 J 3 2: IUNHI

GUESSCJ:I) I GUESSCI:I)

IF (J " I '1) 15.14.15

GUESSIJ:I) 3 0I 3 GUESSCJ:I) '0 GUESSCJ:I)

CONTINUE

CONTINUE

CALCULATE THE THEORETICAL DEPENDENT VARIABLES

DO 22 J 3 I: IUNKI

RESID 3 00

DO I8 L 3 I: INTER

THIS IS THE THEORETICAL EQUATION _
A 3 ZARRAYCL) 3 (ALOGIXARRAYCLII) 3 (GUESSIJ:2))/C(XARRAY(L)
I33CGUESSIJ:2) 4' GUESSCJ:I))) " (XARRAYIL)3*GUESS(J:I)))
RESID 3 RESID + (AESIA ' YARRAYCL)’)##2.

CONTINUE

IF (KL "' I) 19:19:3I

SAVE THE RESIDUAL FOR EACH VERTEX

RSDLCJ) 8 RESID

CONTINUE

START SORTING THE RESIDUALSI WORST RESID IN EADCI)

DO 22 I 3 I: IUNKI

FINDCI) 3 RSDLII)

CONTINUE - .

DO 25 J ,3 I: IUNKI

BI 3 20

DO 24,I 3 I: IUNKI

IF (BI ' FINDIII) 23:24:24

81 3 RSDLCI)

K1 3 I

CONTINUE

FINDIKI) 3 Go

EADCJ) 3 BI

KCJ) 3 KI

CONTINUE

SORTING COMPLETED} START GENERATION OF THE NEW SINPLEX
THE NEXT DO LOOP IS USED TO AVOID HAVING THE SIMPLEX STRANDED
IT IMPLEMENTS THE USE OI: OTHER VERTICES (RATHER THM THE ‘
WORST VERTEX) FOR THE CONSTRUCTION OF A NEW SIMPLE”

DO 56 INDDC 3 I: IUNKN

I 3 KCINDEX)

GENERATION OF THE HYPERI’ACE

DO 28 N 3 I: IUNKN

PRESCN) 3 0.

DO 27 J = I: IUNKI

I? (J ' I) 26:27:26 -

PRESCN) 3 PRES(N) '0' GUESS(J:N)/RIUNKN

CONTINUE ‘

CONTINUE

0000000000

30

31
32

33
34

42

43
44

174

GBNERATE NEV SIMPLEX USING THE FOLLOVING ALGORITHM

REFLECT THE WORST VERTEX (V) ACROSS THE HPERFACE TO FORM

A NEW VERTEX (R): IF (R) IS NOV THE BEST VERTEX: EXPAND
THE SIMPLEX TO (5): IF (S) IS NOV THE BEST: KEEP (5)3 IF
NOT: KEEP (R): IF (R) IS NEITHER THE BEST NOR THE

WORST VERTEX: KEEP (R): IF (R) IS THE WORST VERTEX AND
WORSE THAN (U): THEN CONTRACT INTERNALLY: IF (R) IS BETTER
THAN (U): THEN CONTRACT EXTERNALLY: IF INTERNAL OR EXTERNAL
CONTRACTION RESULTS IN THE NEU VERTEX BEING THE WORST:

THE SIMPLEX IS FURTHER DIMINISHED

NV I 1

FACK 3 I.

DIRECT REFLECTION

DO 32 N I I: IUNKN

RNEVCNV:N) I PRES(N) + FACK I (PRES(N) - GUESSCI:N))
GUESS(I:N) I RNEV(NV:N)

CONTINUE

KL I 2

J I I ,

GO TO 17

RENEUCNV) 3 RESID

GO TO (32:37:42:42) NV

' IF (BADCIUNKI) - RESID) 33:36:36

IF (BAD(2) - RESID) 39:39:34

HATCH I Go .

DO 35 N I 1: IUNKI

GUESS(I:N) I RNEUCNVSN)

CONTINUE

KEEP DIRECTLY REFLECTED VERTEX
RSDLCI) I RENETAWNV)

GO TO 47

EXTERNALLY EXPAND SIMPLEX

NV I 2

FACK 3 20

GO TO 29 H .

KEEP EXPANSION} GO TO STATEMENT 34
IF (BADCIUNKI) - RESID) 38:34:34
KEEP THE DIRECT REFLECTION

NV I I

GO TO 34

IF (BAD(1) - RESID) 40:40:41
INTERNALLY CONTRACT THE SIMPLEX
NV I 3

PACK 3 '05

GO TO 29

EXTERNALLY CONTRACT THE SIMPLEX
NV I 4

EACH 3 05

GO TO 29

IF (BAD(2) - RESID) 58:58:34
RECONSTRUCT THE SIMPLEX USING THE BEST VERTEX
VRITE(1:44) .
FORMATC/I: 'THE PROGRAM VILL RECONSTRUCT THE SIMPLEX
1 AND CONTINUE')

45

46

S2
53

63

175

NE I NE + 1

THE PROGRAM IS ONLY ALLOWED TO RECONSTRUCT 25 TIMES

IF (NE - 25) 45:45:54

DO 46 N I 1: IUNKN

IBEST I K(IUNK1)

GUESS(1:N) I GUESS(IBEST:N)

CONTINUE

KL I 1

GO TO 13 ,

COUNT ITERATIONS AND SEE IF CONVERGENCE WAS ATTAINED
ITT I ITT + I

DO 49 N I 2: IUNK!

IF (ABS(RSDL(1) - RSDL(N)) - OED) 49:49:48

CONTINUE CONVERGING IF THE MAXIMUM NUMBER OF ITERATIONS
HAS NOT BEEN EXCEEDED

IF (ITT - 606) 21:54:54

CONTINUE

‘VRITE(1:SE)

FORMAT(//: 'CONVERGENCE ATTAINEDI')

PRINT OUT THE UNKNOVNS

N I K(IUNKI)

DO 53 I I 1: IUNKN

WRITE(1:52) I: GUESS(N:I)

FORMAT(//: 'w': III: I P ': IEI206)

CONTINUE .

wRITE(1:63) PSDL(1)

FORMAT(//: 'RESIDUAL I ': 1E12o6)

GO TO 62

FAILURE TO CONVERGE CAN BE DUE TO TOO MANY RECONSTRUCTIONS
OR EXCEEDING THE MAXIMUM NUMBER OF ITERATIONS
MRITE(1:SS)

FORMAT(//: 'CONVERGENCE WAS NOT ATTAINEDI')

LIST THE'UNKNOUNS

GO TO 51

A JUMP TO STATEMENT 56 MEANS THAT THE PROGRAMS IS USING
THE SECOND WORST VERTEX TO AVOID BEING STRANDED
CONTINUE

A JUMP TO STATEMENT 43 MEANS THAT THE SIMPLEX IS STANDED
AND CONSEQUENTLY: THE SIMPLEX VILL HAVE TO BE RECONSTRUCTED
GO TO 43

A JUMP TO STATEMENT 58 MEANS THAT THE PROGRAM IS DIMINISHING
THE SIMPLEX EITHER INTERNALLY OR EXTERNALLY TO AVOID BEING
STRANDED.

IF (WATCH - 9.) 59:59:56

GO TO (56:56:62:61) NV

PACK 3 FACK ‘I’ 006

GO TO 29 .

FACK I FACK - .26

GO TO 29

CONTINUE

END

176

the constant part of each term must be calculated and the
resultant numbers converted into binary floating point nota-
tion. Program ARTCAL is used for this task. Upon prompting
from the program, the user types in the values of w1 and w2
which were obtained from RTFACT. The program responds by
outputting the decimal value for each term constant as well
as the binary floating point equivalent. ARTCAL calculates
the term constants for the second through eleventh terms of
the Taylor series. This is done so that the user can be
assured that the eighth through eleventh terms are negligible
compared to the first seven terms. The term constants ARTl
through ARTS are then substituted into FLUORO and the user
is now ready to make automated absorption-corrected fluores-
cence measurements. A program listing and flowchart for

ARTCAL appear on the following pages.

177

PROGRAM ARTCAL

DIMENSION IHOLD(2¢): IEASE(60): IEX(4): NOUT(20): APT(1?)

28
29
3?

31

DIMENSION APTN(1G)
HJ I O
VRITE(1:28)

FORMAT(/:'TYPE IN U1 AND V2: RESPECTIVELY')

READ(1:29) V1
FORMAT(1E12:6)
READ(1:30) V2
FOPMAT(1E12:6)
DO 32 NN‘I: I0
FACT'I.

LINN+1

DO 31 JJI1:L
R I JJ
FACTIFACTIR
CONTINUE
EXINN+1

APT(NN)=((V2IIEX)-(V1IIEX))/(FACTI(W2-V1))

IZII

J-n

ISCOREII

INDEXI2

NEXI?
ARTN(NN)IART(NN)
DO 2 131:20
APT(NN)=APT(NN)I8.
NUMBEPIAPT(NN)
IHOLD(I)INUMBEP
PEALMINUMBER
APT<NN)IART(NN)-REALM
CONTINUE

DO 3 I'I:60
IBASE(I)=¢
CONTINUE

D0 4 I31:6U:3
MII+1

NII+2
IHOGIIHOLD<J>+1

GO TO (12:5:6:7:8:9:1€:11): IHOG
IBASE(N)=1

GO TO 12
IBASE(M)I1

GO TO 12
IBASE(M)I1~
IBASE(N)=1

12

11

12
13

14
15

16

17

18
19

29

21

22
23

24

4C

41

42

37

178

GO TO 12

IBAS£(I)=1

GO TO 12

IBASE<1>Il

IBASEtN)Il

GO TO 12

IBASE(I)=1

IBASE(M)I1

GO TO 12

IBASE<I>II

IBAS£(M)=1

IBASE<N>=1

IF (J-2G) 13:4:4

JIJ+1

CONTINUE

DO 24 1:1:69

GO TO (14:17:18): ISCORE
If (IBASECI)) 16:15:16
NEY=NEY+1 ‘

GO TO 24
ISCOPEIISCORE+1
NOUT(1)-IBASE(I)I2

GO TO 24
NOUT<1>=NOUT<1)+IBASE(I>
ISCOPEIISCOPE+1

GO TO 24

GO TO (19:22:21): 12
NOUT(INDEX)IIBASE(I)I4
IZII7+1

GO TO 22
NOUT<INDE¥)=IBASE(I)I2+NOUT(1NDEX)
IZIIZ+1

GO TO 22
NOUT(INDEY)IIBASE(I)+NOUT(INDE¥)
IZIIZ+1

If (IZ-4) 24:23:24
1281

INDEYIINDEX+1

CONTINUE

If (NE?) 42:42:42

DO 41 JR I 1:4

IEXtJR) I Z

CONTINUE

GO TO 43

PNEX I NE?
DECIMALI4O9S.-PNEY
DIVIDECIMAL/B.
IDIVIDIV

SAVOP-IDIV
PDIVISAVORtB.
IEX<1>IDECIMAL-RDIV+1.
GO TO 34

RDEC I IDIV

38

39
43

33
25
26
27
34
35

36
32

179

DIVIRDEC/B:

IDIVIDIV

PDIVIIDIV*8

IEX(2)IPDEC-PDIV

GO TO 34

RDEC .3 IDIV

DIVIRDEC/Uo

IDIVIDIV

RDIV'IDIV*8

IEX(3)3RDEC'RDIV

GO TO 34

IEX(4) I IDIV

NJ 3 0

URITE(1:33) NN: ARTNINN)

FORMATI/ll: 'ART':112:' ' ':IEI206)
WRITE(I:25) IEX(4): IEX(3): 187((2): IEX(I)
FORMAT(/:'THE EXPONBH‘ I ':QII)
WRITE(I:26) NOUTCI): NOUTIE): NOUTC3): NOUT(4)
FORMAT(/:'THE UPPER MANTISSA 3 ':4I I)
URITE(1:27) NOUTIS): NOUT(6): NOUT(7): NOUT(8)
FORMAT(/:'THE LOWER MANTISSA ' ':4II)

GO TO 32

KJ' 3 KJ I I

IF (IEX(KJ)'8) 36:35:35

IEX(KJ) ' O

IDIV I IDIV I I

GO TO (37: 38: 39) KJ

CONTINUE

END

180

PROGRAM ARTCAL FLOWCHART

< START >

INPUT
H AND

 

 

 

 

 

CALCULATE
TERM
CONSTANT

 

 

 

 

 

CONVERT
T0 DCTAL
NUMBER

 

 

 

 

 

t
DETERMINE
EXPONENT
AND
MANTISSAS

 

 

 

 

 

     
      

OUTPUT
CONSTANT
AND ITS
BINARY
Fi'

 

ND

 

Figure 41. Program ARTCAL Flowchart.

BIBLIOGRAPHY

6.
7.
8.

10.

11.

12.

13.

14.

15.

16.
17.

18.

19.

BIBLIOGRAPHY

Hitachi, Ltd., New Marunouchi Bldg., Tokyo, Japan,

Photovolt Corp., 1115 Broadway, New York, Lumetron 402
EP.

Klett Mfg. Corp., 179 E. 87th Street, New York, Model
2070.

Parker, C. A., Nature, 82, 1000 (1958).

Slavin, W., Mooney, R. W. and Palumbo, D. T., J. Opt.
Soc. Am., 51, 93 (1961).

Lipsett, R. P., J. Opt. Soc. Am., 12, 673 (1959).

American Inst. Co., Bull. 2392 D (1967).

Perkin Elmer Corp., Instrument News, 21 (2) 12 (1970).
Turner, G. K., Science, 116, 364, 183 (1964).

Cravitt, S. and Van Duuren, B., Chem. Inst., 1, 71 (1968).

Howerton, H. K., "Fluorescence", Dekker, New York,
Chapter 5 (1967).

Parker, C. A. and Rees, W. T., Analyst, 85, 587 (1960).

Witholt, B. and Brand, L., Rev. of So. Inst., 32, (9),
1271 (1968).

Melhuish, W. H., J. Opt. Soc. Am., 52, 1256 (1962).

Lee, J. and Saliger, H. H., Photochem. Photobiol., g,
1015 (1965).

Argaver, R. and White, C. E., Anal. Chem., 36, 2141 (1964).
Parker, C. A., Anal. Chem., 32, 502 (1962).

Holland, J. P., Teets, R. E. and Timnick, A., Anal.
Chem., 15, 145 (1973).

Price, J. M., Kaibara, M. and Howerton, H. K., App. Opt.,
1, 521 (1962).

181

20.
21.
22.
23.

24.
25.
26.
27.
28.

29.
30.

31.
32.

33.

34.

35.

36.
37.
38.

39.

40.

182

Hermans, J. J. and Levinson, S., J. Opt. Soc. Am., 41,
460 (1951).

Van Slageren, R., Den Boef, G. and Van der Linden, W. E.,
Talanta, 29, 501 (1973).

Hercules, D. M., "Fluorescence and Phosphorescence
Analysis", John Wiley and Sons, New York (1966).

Passwater, R. A. and Hewitt, J. W., Fluorescence News,

Longsworth, J. W., Photochem. Photobiol., 8, S89 (1968).
Parker, C. A. and Barnes, W. J., Analyst, 82, 606 (1957).
Parker, C. A. and Rees, W. T., Analyst, 81, 83 (1962).
Ohnesorge, W. E., Anal. Chim. Acta., 31, 484 (1964).
Gill, J. E., Appl. Spectrosc., 24, 588 (1970).

Holland, J. F., Teets, R. E., Kelly, P. M. and Timnick,
A., submitted to Anal. Chem. for publication.

Michaelson, R. C. and Loucks, L. P., J. Chem. Ed., 52,
652 (1975).

Seely, G. R., J. Phys. Chem., 13, 125 (1969).

Eastman Kodak Co., publications JJ-3l and JJ-32.

Deming, S. N. and Morgan, 8. L., Anal. Chem., 45, 278A
(1973).

Katyal, M., Talanta, 15, 95 (1968).

Nevskaya, E. M. and Nazarenk, V. A., Zh. Analit. Khim.,
‘21, 1699 (1972).

Korac, A., Wiadom. Chem., 23, 345 (1969).
Wilson, C. A., J. Am. Chem. Soc., 61, 2303 (1939).
Taubak, K., Naturwiss., 39, 439 (1942).

Hbrhammer, L. and Hansel, R., Arch. Pharm., 285, 286
(1952).

HBrhammer, L. and Hansel, R., Arch. Pharm., 28 , 447
(1953).

183

41. Hérhammer, L. and Hansel, R., Arch. Pharm., 88, 315
(1955).

42. Hagadorn, P. and Neis, R., Arch. Pharm., 286, 486 (1953).

43. Kohara, H. and Ishibashi, N., Japan Analyst, 16, 470
(1967).

44. Battei, R. and Trusk, B., Anal. Chem., 25, 1910 (1963).

45. Battei, R. and Trusk, 8., Anal. Chim. Acta., 31, 409
(1967).

46. Bozhevol'nov, E. A., "Luminescent Analysis of Inorganic
Materials", Khimiya, Moscow (1966). p. 86.

47. Kanno, T., Sci. Repts. Res. Inst. Tohoku Univ., A12,
532 (1960).

48. Kanno, T., Sci. Repts. Res. Inst. Tohoku Univ., A13,
91 (1961).

49. Kohara, H., gt 31., Japan Analyst, 16, 315 (1967).

50. Kohara, H., EE.El°v Japan Analyst, $2! 938 (1966).

51. Elinson, S. V. and Petrov, K. I., "The Analytical
Chemistry of Zirconium and Hafnium", Nauka, Moscow

(1965), p. 150.

52. Jurd, L. and Geissman, T. A., J. Org. Chem., 21, 1395
(1956).

53. Urbach, F. L., Ph.D. Thesis, Michigan State University,
East Lansing, Michigan (1964).

54. Urbach, F. L. and Timnick, A., Anal. Chem., 40, 1269
(1968). '_—

55. Ohnesorge, W. E. and Capotosto, Jr., A., J. Inorg.
Nucl. Chem., 34, 829 (1962). .

56. Porter, L. J. and Markham, K. R., J. Chem. Soc. (C),
1970, 344.

57. Wheland, G. W., "Advanced Organic Chemistry", 3rd
ed., John Wiley and Sons, New York (1960), pp. 683-
693.

58. Aveston, J., J. Chem. Soc., 1965, 4438.

59. Brosset, C., Biedermann, G. and Sillen, L. G., Acta
Chem. Scand., 8, 1917 (1954).

60.

61.
62.

184

Kohlschfitter, H. W. and Hantelmann, P., z. Anorg.
Allegem. Chem., 248, 319 (1941).

Ohnesorge, W. E., J. Inorg. Nucl. Chem., 32, 285 (1967).

Yoshihara, K. and Kearns, D. R., J. Chem. Phys., 45,
1991 (1966).

 

"TTT‘TTTTTTTTTTT