F
rdfizd.
.. as. A.
x
1‘5.

nfixnun? h
. m. a

f.
1......

a L
mt .v :7.

 

.2155“.
. iii r

. , .2 a: th. $8.49..”
unhmrnsthimrtiv i; ‘3»)qu
than}!!! vein.“

3.2. 5:15:25

0

{(3.9}? ‘ .,
N . 1;..qu «an,

 

Meals

IIIIIIIII IIllIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIII

3 1293 01566 0784

LIBRARY

Michigan State
University

 

 

 

 

This is to certify that the

dissertation entitled
A Study of Structure in Supersaturated Aqueous

Sucrose Solutions Using Fluorescence Spectroscopy

presented by

Reena Chakraborty

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Chemical Engineering

 

 

% 0 4%“?
Majorprofes :-

Date 614/"

MS U is an Affirmative Action/Equal Opportunity Institution 0-1277!

PLACE ll RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date duo.

DATE DUE DATE DUE DATE DUE

-Im

I
I
-C». I
.| . II
b3 '
|
I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU IOMWWWOWIW
Wt

 

A STUDY OF STRUCTURE IN SUPERSATURATED AQUEOUS SUCROSE
SOLUTIONS USING FLUORESCENCE SPECTROSCOPY

By
Reena Chakraborty

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemical Engineering

1996

ABSTRACT
A STUDY OF STRUCTURE IN SUPERSATURATED AQUEOUS SUCROSE
SOLUTIONS USING FLUORESCENCE SPECTROSCOPY
By
Reena Chakraborty

The fluorescence of 8-hydroxy-l,3,6-pyrenetrisulfonate (pyranine), a trace
extrinsic probe, was measured in aqueous sucrose, glucose, lactose and fructose solutions.
The peak intensity ratio (PIR), the emission intensity of pyranine at 440 nm to that at 515
nm, shows a strong concentration dependence in supersaturated solution. The PIR
represents the ratio of solvation to bulk water per molecule of solute, and determines
concentration and activity simultaneously. Non-solvation water in aqueous sucrose
solutions, designated bulk water, exhibits one proton exchange rate. In other sugar
solutions, it exhibits more exchange rates, depending upon the strength of the hydrogen
bonds formed by the solvation water with the various moieties of the solute molecules.
Steady state and time resolved fluorescence spectroscopy in supersaturated aqueous
sucrose solutions indicate differences in structure and mobility of both bulk and solvation
water compared to undersaturated solutions.

Solvent distribution near growing and dissolving sucrose crystals was investigated
using laser optical fluorescence microscopy. The ratio of integrated half -peak intensifies
(IIR) of pyranine in sucrose solutions from 370 nm to 445 nm and 485 nm to 507 nm
exhibits a strong concentration dependence. Solvent microphase distribution in regions of
solution, 0.4 mu x 0.4 gm, indicated an isotropic distribution of pyranine in both excited
states. Interfacial solution structure differs from bulk structure. A strong crystal-size
dependence was observed in the extent of the concentration boundary layer; supersaturation

dependence was not observed.

Solvatons, molecules of sugar and their solvation water, are convenient structural
descriptors in non-dilute solution: in undersaturated solution, bulk-water solvates
solvatons; in supersaturated solution, the solvatons solvate bulk-water. Saturation is akin to
phase-inversion. The self-diffusion coefficient, Dself, is a ratio of the surface area of a
sphere with radius, R (the distance between identical solvation environments for a
solvaton), and the time, t, required to traverse R; the viscosity coefficient, 11., as the ratio of
the force of resistance, Fr, experienced by the solvaton in traversing R, to Dself. Near
saturation, R decreases, “c increases, and Fr becomes very large, explaining precipitous

drops in Dsefl' and order of magnitude increases in rt, reported in literature.

 

To my parents,
Dr. Bijoy Kumar Chakraborty
Mrs. Radha R31 Chakraborty
who believe in me,

and supported and encouraged
my educational pursuits.

ACKNOWLEDGEMENTS

Dr. Kris A. Berglund is thanked for serving as my graduate advisor. Dr. Daina
M. Briedis, Dr. Carl T. Lira, and Dr. George E. Leroi are thanked for serving on my
graduate committee and providing encouragement.

Dr. Donald K Anderson is thanked for many terms of financial support Many
terms were spent working for and with Dr. Martin C. Hawley and Dr. Krishnamurthy
Jayaraman, who are thanked for helping me appreciate process engineering design and
optimization as much as they do. Thanks also to Dr. Wei E. Kuan, Dr. William T. Sledd,
Dr. Mary Winter, and Mrs. Elizabeth D. Phillips of the Math Department at Michigan State
University who allowed me to learn how to teach Intermediate Algebra to freshmen during
the 2 semesters it took to dot the i's and cross the t's.

I looked to many friends for perspective, encouragement, and support. Chief
among them Afalah Radford, Maria Tjani, Shobha Ramanand, Susan Jones, Margaret
Green, Elizabeth Davenport, Suvama Padhalkar, V. Subba Rao, Anil Menon, Joel
Dulebohn, my sister Hena, and brother-in-law Kalyan Pyada. Thanks to the Berglund
Group, and Mike Bly, Ramkumar Subhramanian and the Pictionary Gang for perspective
and good company. Special thanks go to Tricia Moore and the folks at I-5, Carl and
Barbera Bourdelais and the folks at TBT, and Paul Hoelderle and all my friends at FPC.
Thanks also to Dr. Robert E Buxbaum and Mrs. Gail C. Buxbaum. Laura G. Berry is
thanked for demonstrating the power of the ACAS, positive thinking, and friendship.
Hobbes, Leo, and Axel and the late Calvin, Maurice and Portia, are thanked for my
hours of fine, furry, four-legged companionship.

TABLE OF CONTENTS

Chapter One: Introduction ........................................................................ 1
1.1 Crystal Growth from Solution ..................................................... l
1.2 Supersaturated Solutions are Difficult to Investigate ............................ 2
1.3 Consequences of Limited Experimental Techniques: Limited
Information Regarding Solution Structure ............................................ 3
1.4 A Broad Overview of this Work .................................................. 4

1.4.1 The Subject of Investigation ............................................ 4
1.4.2 The Experimental Approach ............................................ 5
1.4.3 Objectives ................................................................. 6
1.4.4 An Overview of the Chapters .......................................... 7
1.5 References ............................................................................ 9

Chapter Two: Literature Review ................................................................. 10
2.1 Background .......................................................................... 10
2.2 Crystal Growth from Solution ..................................................... 13

2.2.1 The Requirements for Crystal Growth from Solution ............... 13
2.2.2 Crystal Growth Kinetics ................................................ 13
2.2.3 Nucleation ................................................................. 14
2.2.4 The BCF Theory of Crystal Growth ................................... 15
2.2.4.1 ............................................................................ 17
2.2.5 Evidence for Non-Cluster Dependent Growth Mechanisms ........ 18
2.3 Solution Structure .................................................................. . 19

iii

2.3.1 The relevance of structure to crystallization research ................ 19

2.3.2 Ideal Solution and the classical treatment of solution structure ..... 21
2.3.3 Other Concepts of Solution Structure ................................. . 26
2.3.4 Non-Continuum Descriptions of Solvation and Structure ........... 30
2.4 Techniques commonly used to measure concentration in supersaturated
solutions ................................................................................... 32
2.4.1 Spectroscopic Sensing Techniques .................................... 33
2.5 References ........................................................................... 36

Chapter Three: The use of pyranine as a trace fluorescent probe to study structure in

aqueous sucrose solutions ........................................................................ 40
3.1 Background .......................................................................... 40
3.2 Photochemical Techniques using Trace Fluorescent Probes ................... 44
3.3 Materials and Methods: ............................................................. 48
3.4 Results and Analysis ................................................................ 50
3.5 Conclusions .......................................................................... 57
3.6 Acknowledgements ................................................................. .57
3.7 Literature Cited ....................................................................... 57

Chapter Four: Steady State Fluorescence Spectroscopy of Pyranine as a Trace

Extrinsic Probe to Study Structure in Aqueous Sugar Solutions ............................ 62
4.1 Background .......................................................................... 62
4.2 NMR .................................................................................. 64
4.3 Raman Spectroscopy ................................................................ 65
4.4 X-ray Diffraction .................................................................... 66
4.5 Diffusion and Viscosity Measurements ........................................... 66
4.6 Correlation of Macroscopic and Microscopic transport properties ............. 67
4.7 Photochemical Techniques using Extrinsic Trace Fluorescent Probes ........ 69
4.8 Materials and Methods .............................................................. 71

iv

4.9 Results and Analysis ................................................................ 73

4.10 Conclusions ......................................................................... 79
4.11 Acknowledgements ................................................................ 80
4.12 References .......................................................................... 81
Chapter Five: The PIR as a measure of deviation from Ideal Solution Behavior .......... 98
5.1 Background .......................................................................... 98
5.2 Materials and Methods .............................................................. 100
5.3 Results ................................................................................ 101
5.4 Analysis ............................................................................... 101
5.5 Discussion ............................................................................ 104
5.6 Conclusions .......................................................................... 105
5.7 Notation ............................................................................... 106
5.8 References ............................................................................ 106

Chapter Six: The Study of Structure in Saturated and Supersaturated Sucrose

Solutions using Steady State and Time Resolved Fluorescence Spectroscopy ............ 113
6.1 Background .......................................................................... 113
6.2 Experimental Methods .............................................................. 117

6.2.1 Fluorescent probes: ...................................................... 117
6.2.2 Preparation of Solutions ................................................. 118
6.3 Results and Discussion ............................................................. 119
6.3.1 Steady State Fluorescence Spectra and UV Absorption ............. 119
6.3.2 Time resolved Fluorescent Lifetimes ................................. . 120
6.3.3 Time Resolved Fluorescence Anisotropy ............................. 121
6.4 Structural Characteristics of Sucrose Solutions from Transport
Properties .................................................................................. 128
6.5 Self -Diff usion and Structure ....................................................... 133
6.6 Conclusions .......................................................................... 135

6.7 References ............................................................................ 137

Chapter Seven: Solvent Imaging Studies of Supersaturated Sucrose Solutions in the
Absence of Crystals and of the Near Interfacial Region of Growing and Dissolving

Sucme Crystals ................................................................................... 150
7.1 Background .......................................................................... 150

7.2 Materials and Methods .............................................................. 151

7.3 Results ................................................................................ 153

7.4 Discussion ............................................................................ 156

7. 5 Crystal Growth ..................................................................... 158

7.6 Conclusion ........................................................................... 160

7.7 Acknowledgements ................................................................. . 160

7.8 References ............................................................................ 161
Chapter Eight; Summary, Conclusions, & Recommendations for Further Work ......... 171
8.1 Summary ............................................................................. 171

8.1.1 Issues Related To Solution Structure ................................. 171

8.1.2 SRO and LRO in The Aqueous Sucrose System ..................... 173

8.1.3 Bulk Water and Solvation Water as indicators of SRO ............. 178

8.1.4 Other effects on structure and LRO ................................... 181

8.1.5 Issues related to structure near growing sucrose crystals. .......... 184

8.2 Conclusions ......................................................................... 186

8.3 Recommendations for Further Work .............................................. 187

vi

LIST OF FIGURES

Chapter Three
Figure 3.1. Chemical structure of pyranine or 8-hydroxy-l, 3, 6-pyrenetrisulfonate ...... 45

Figure 3.2. Emission equilibrium of pyranine. The equilibrium between the protonated
form (absorbs at 350 nm and emits at 440 nm when excited at 342 nm) and the
deprotonated form (absorbs at 405 nm and emits at 511 nm when excited at 342 nm) is
determined by ability of the microenvironment of the hydroxyl group to participate in
proton transfer ....................................................................................... 46

Figure 3.3. The emission spectra of pure water and sugar solutions of 20, 40, 60, and 70

wt.% sucrose doped with 100 ppm (1.9 x 10‘4 M) pyranine at 20 degrees C. Emission
maxima occur at 440 nm. and 511 nm. An isosbestic point is observed, represented by the
pointer at 485 nm .................................................................................... 50

Figure 3.4. Graph of PIR (peak intensity ratio) vs. wt% sucrose in solution at 20 degrees

C using 100 ppm (1.9 x 10'4 M) pyranine. The PIR is defined as the ratio of peak
intensities at 440 nm and 511 nm. The arrow represents concentration at saturation ....... 53

Figure 3.5. Total, bulk and solvation water (number of molecules of water per molecule of

sucrose) vs. wt% sucrose in aqueous solution with 100 ppm (1.9 x 10'4 M) pyranine at 20
degrees C. The values on the graph are obtained from the analysis presented in Tables 1
and 2. The vertical line represents saturation ..................................................... 55

Figure 3.6. Graph of total water, bulk water and solvation water per molecule of sucrose in

aqueous solutions doped with 100 ppm (1.9 x 10‘4 M). pyranine at 20 degrees C versus
the degree of supersaturation, DS. The DS is given by the difference between the solution
concentration and the saturation concentration at 20 degrees C divided by the saturation
concentration at 20 degrees C ...................................................................... 56

vii

Chapter Four
Figure 4.1. Chemical structure of pyranine or 8-hydroxy-1, 3, 6—pyrenetrisulfonate ...... 84

Figure 4.2. Emission equilibrium of pyranine. The equilibrium between the protonated
form (absorbs at 350 nm and emits at 440 run when excited at 342 nm) and the
deprotonated form (absorbs at 405 nm and emits at 511 run when excited at 342 nm) is
determined by ability of the microenvironment of the hydroxyl group to participate in
proton transfer ....................................................................................... 85

Figure 4.3. The emission spectra of pure water and sugar solutions of 20, 40, 60, and 70

wt.% sucrose doped with 100 ppm (1.9 x 10‘4 M) pyranine at 20 degrees C. Emission
maxima occur at 440 um. and 511 nm. An isosbestic point is observed, represented by the
pointer at 485 nm .................................................................................... 86

Figure 4. 4. Graph of PIR (peak intensity ratio) vs. wt% sucrose in solution at

20 degrees C using 100 ppm (1. 9 x 10'4 M) pyranine. The PIR1s defined as the ratio of
peak intensities at 440 nm and 511 nm. The arrow represents concentration at
saturation ............................................................................................. 87

Frgure 4. 5. Graph of PIR (peak intensity ratio) vs. wt% glucose 1n solution at

20 degrees C using 100 ppm (1. 9 x 10'4 M) pyranine. The PIR 18 defined as the ratio of
peak intensities at 440 nm and 511 nm. The dashed line represents concentration at
saturation ............................................................................................. 88

Figure 4.6. Graph of PIR (peak intensity ratio) vs. wt% fructose in solution at

210 degrees C using 100 ppm (1.9 x 10‘4 M) pyranine. The PIR is defined as the ratio of
peak intensities at 440 nm and 511 nm. The arrow represents concentration at
saturation ............................................................................................ 89

Frgure 4.7. Graph of PIR (peak intensity ratio) vs. wt% lactose in solution at

20 degrees C using 100 ppm (1.9 x 10'4 M) pyranine. The PIR is defined as the ratio of
peak intensities at 440 nm and 511 nm. The dashed line represents concentration at
saturation ............................................................................................ 90

Figure 4. 8. Total, bulk and solvation water (number of molecules of water per molecule

of sucrose) vs. wt% sucrose in aqueous solution with 100 ppm (1.9 x 10'4 M) pyranine
at 20 degrees C. The values on the graph are obtained from an analysis using a two state
model for water. The vertical line represents saturation ....................................... 91

viii

Figure 4.9. Graph of bulk water and solvation water per molecule of sucrose in aqueous

solutions doped with 100 ppm ( 1.9 x 10‘4 M). pyranine at 20 degrees C versus the degree
of supersaturation, DS. The DS is given by the difference between the solution
concentration and the saturation concentration at 20 degrees C divided by the saturation
concentration at 20 degrees C ...................................................................... 92

Figure 4.10. Total, bulk and solvation water (number of molecules of water per molecule

of glucose) vs. wt% glucose in aqueous solution with 100 ppm (1.9 x 104 M) pyranine
at 20 degrees C. The values on the graph are obtained from an analysis using a two state
model for water. The vertical line represents saturation ....................................... 93

Figure 4.11. Total, bulk, and solvation water (per molecule of fructose) as a function of %
fructose in solution. Data obtained for aqueous fructose at 293 K doped with 100 ppm
pyranine .............................................................................................. 94

Figure 4.12. Total, bulk and solvation water (number of molecules of water per molecule

of lactose) vs. wt% lactose in aqueous solution with 100 ppm (1.9 x 10'4 M) pyranine at
20 degrees C. The values on the graph are obtained from an analysis using a two state
model for water. The vertical line represents saturation ........................................ 95

Figure 4.13. Bulk water, solvation water, and non-bulk-non-solvation water (number of
molecules of water per molecule of glucose) vs. wt% glucose in aqueous solution with 100

ppm (1.9 x 10‘4 M) pyranine at 20 degrees C. The values on the graph are obtained from
an analysis using a three state model for water. The vertical line represents
saturation ............................................................................................. 96

Figure 4.14. Bulk water, solvation water, and non-bulk-non-solvation water (number of
molecules of water per molecule of lactose) vs. wt% lactose in aqueous solution with 100
ppm (1.9 x 10'4 M) pyranine at 20 degrees C. The values on the graph are obtained from

an analysis using a three state model for water. The vertical line represents
saturation ............................................................................................. 97

Chapter Five

Figure 5.1. The emission spectra of pure water and sugar solutions of 20, 40, 60, and 70

wt.% sucrose doped with 100 ppm (1.9 x 10'4 M) pyranine at 20 degrees C. Emission
maxima occur at 440 nm and 511 run An isosbestic point is observed, represented by the
pointer at 485 nm .................................................................................. 108

Figure 5.2. Chemical structure of pyranine or 8-hydroxy-l, 3, 6-pyrenetrisulfonate ..... 109

Figure 5.3. Emission equilibrium of pyranine. The equilibrium between the protonated
form (absorbs at 350 nm and emits at 440 nm when excited at 342 nm) and the
deprotonated form (absorbs at 405 nm and emits at 511 run when excited at 342 nm) is
determined by ability of the nricroenvironment of the hydroxyl group to participate in
proton transfer ...................................................................................... 110

Figure 5.4. Graph of xwlny“, versus ln PIR for aqueous sucrose solutions at 293 K ..... 111

Figure 5.5 Graph of xwlnYW versus ln Bulk water/Solvation Water (CA or Chemically

Associated water) for aqueous glucose solutions at 293 K ................................... 112
Chapter Six
Figure 6.1. Structure of PBA (left) and pyranine (right) ...................................... 138

Figure 6.2. UV-vis spectra of 10 ppm pyranine in aqueous sucrose solutions of various
concentrations at 293 K ........................................................................... 139

Figure 6.3. Fluorescent lifetimes of PBA in sucrose solutions observed at 380 nm and 293
K as a function of sucrose concentration. Squares ( El ) represent the long lifetime

component and triangles ( A ) represent the short lifetime component ..................... 140

Figure 6.4. Ratio of the intensities of the long lifetime component to the short lifetime
component of PBA at 380 nm .................................................................... 141

Figure 6.5. Fluorescent lifetimes of PBA in sucrose solutions observed at 400 nm and 293
K as a function of sucrose concentration. Squares ( El ) represent the long lifetime
component and triangles (A ) represent the short lifetime component ..................... 142

Figure 6.6 Ratio of the intensities of the long lifetime component to the short lifetime
component of PBA at 400 nm .................................................................... 143

Figure 6.7 Fluorescent lifetimes of pyranine in sucrose solutions at 293 K. Squares ( Cl )
represent the emission at 440 nm and triangles ( A ) represent the the emission at 511 nm.
Error bars indicate the standard deviation of the measurements .............................. 144

Figure 6.8 Kinetic diagram for the two-state excited state proton transfer of pyranine... 145

Figure 6.9 Dependence of the rotational correlation time of PBA on sucrose concentration.
Squares ( CI) represent the rotational correlation time for the long lifetime component and
triangles ( A) represent the rotational correlation time for the short lifetime component. 146

Fi ure 6.10 Dependence of the rotational correlation time of PBA on viscosity. Squares
represent the rotational correlation time for the long lifetime component and triangles
(A) represent the rotational correlation time for the short lifetime component ............. 147

Figure 6.11 Dependence of the rotational correlation time of pyranine on sucrose
concentration. Squares ( Cl) represent the rotational correlation time observed at 440 nm
and triangles ( A) represent the rotational correlation time observed at 511 nm ........... 148

Figure 6.12 Dependence of the rotational correlation time of pyranine on viscosity. Squares
( Cl) represent the rotational correlation time observed at 440 nm and triangles ( A )
represent the rotational correlation time observed at 511 nm. Error bars indicate the
standard deviation of the measurements ........................................................ 149

Chapter Seven

Figure 7.1. Dual detector images for aqueous sucrose solutions of 60, 74 and 76 wt.% at
293 K Detector 1 images the half peak intensity of pyranine from 485 nm to 507 nm.
Detector 2 images the half peak intensity of pyranine from 445 nm to 370 nm. One pixel
represents an area of solution 0.4 pm x 0.4 ram ............................................... 162

Figure 7.2. A plot of IIR (integrated emission intensity) of pyranine vs sucrose
concentration (wt.% sucrose) at 293 K Color values represent concentration in weight
percent sucrose ..................................................................................... 163

Figure 7.3 Dual detector (a) and concentration ratio mode (b) images of a sucrose cystal
growing in 68 wt.% solution ..................................................................... 164

Figure 7.4 Dual detector (a and c) and concentration ratio mode (b and d) images of several
different sucrose crystals growing in 71 wt.% solution ....................................... 165

Figure 7.5 Concentration ratio mode images of several different sucrose crystals growing
in 71 wt.% solution ....................................... 165

Frgure 7.6 Dual detector (a) and concentration ratio mode (b) images of a sucrose cystal
growing in 76 wt.% solution ..................................................................... 166

Figure 7.7. Samples of line scans in concentration ratio mode taken in the interfacial region
of growing sucrose crystals at 293 K. The big dip in concentration indicates the interface of
the crystal ........................................................................................... 167

Figure 7.8 A sample of a line scan taken near the interface of a dissolving sucrose crystal in
60 wt% sucrose solution at 293 K. The big dip in concentration indicates the interface of
the crystal ........................................................................................... 168

Figure 7.9 Dual detector images of a sucrose crystal in 74 wt. % solution at 293 K taken
using confocal microscopy in 0.4 gm vertical steps, and 0.2 pm x 0.2. gm horizontal
resolution. Instrument settings for this data differ from those of the rest of the data in this
work ................................................................................................. 169

Figure 7.10. Concentration ratio images of the sucrose crystal imaged in Figure 7.8... .170
Chapter Eight
Figure 8.1 Conceptual diagram of structure in dilute aqueous sucrose solution ........... 187

Figure 8.2. A conceptual representation of structure in concentrated aqueous sucrose
solutions ............................................................................................ 188

Figure 8.3. A conceptual representation of structure in saturated aqueous sucrose
solutions ............................................................................................ 189

Figure 8.4. A conceptual representation of structure in supersaturated aqueous sucrose

solutions ............................................................................................ 190
Figure 8.5 Structure in the aqueous glucose system ........................................... 191
Figure 8.6 Structure in the aqueous fructose system .......................................... 192
Figure 8.7 Summary of structure in aqueous citric acid systems ............................ 193
Figure 8.8 Regimes of solution structure in aqueous sucrose solutions .................... 194

xii

Chapter One

Introduction

Crystallization is one of the most widely used unit operations in the chemical
process industry (1). Many processes utilize crystallization from solution as a means to
purify, produce, or recover solid material. Cost-effective control of industrial processes
related to the production of electronic and optical materials, coatings, and the large scale
production of various chemicals including food and food additives, dyes, fertilizers,
pharmaceuticals, and fine chemicals is one of the primary objectives of engineering
research in separations. This is particularly true of research related to crystallization from
solutions. Control schemes for crystallization processes developed and described in
literature are derived from the population balance analysis of Randolf and Larson (2-4).
The crystal size distribution for a crystallizer can be determined from population balance
analysis given the mode of operation, crystallization kinetics, and operating variables of
temperature and supersaturation. The relation between the major variables of interest is
described using process control analogy. The main challenges to applying such a scheme in
a universal manner to provide on-line control of crystallization from solutions, are
continuous, rapid, and accurate determination of local supersaturation, and in situ

determination of particle size and morphology.

1.1 Crystal Growth from Solution

Crystal growth from solution differs inherently from vapor or melt
crystallization due to the presence of significant quantities of at least one non-crystallizing

species: the solvent The effect of solvent composition on the

2

morphology, habit, growth rate, and crystal size distribution in crystallization from the
solution phase has been well documented. Researchers in the field generally agree that there
is a cause and effect relationship involved regarding the solvent and its effects on the
crystallizing system.

The consensus ends in agreement that a causal relationship between solvent and
crystal properties in solution phase crystallization exists. The nature of such a relationship
is a subject of speculation. Even so, experimental evidence suggests the solvent affects the
crystallization at a microscopic level (involving one or several units of solvated solute
molecules and several molecules of solvent) regardless of how it affects the crystallization

on larger scales of order.

1.2 Supersaturated Solutions are Difficult to Investigate

Experimental investigation of supersaturated solutions on any scale of order is
difficult Supersaturated solutions are not in a state of equilibrium, rather in what Ostwald
first termed as a "metastable state" (5). As a result, localized changes in temperature,
concentration, pressure, fluid velocity, viscosity, the presence of foreign particles, or the
application of a sudden impact may each individually, or in combination, cause a loeal or
global instability to the system. The end result is often localized or catastrophic (global)
nucleation (i.e., not limited to a small region of solution). Reproducing experimental
conditions on a reliable basis is challenging. Rapid and accurate determination of local
and/or global solution concentration in saturated and supersaturated solutions is difficult.
These factors have hindered the experimental study of structure of and phenomena in
supersaturated solutions. Consequently, there are few experimental studies related to
structure and property determination in supersaturated solutions, and fewer still in

nucleating or crystallizing systems. Most structure and property determination related

3

studies have limited their focus to investigating the role of the solute in structure and

property relations.

1.3 Consequences of Limited Experimental Techniques: Limited
Information Regarding Solution Structure

The lack of experimental techniques for accurate determination of local
supersaturation and changes in local supersaturation has severely limited the ability to
measure and quantify interfacial properties during nucleation and crystal growth.
Consequently a clear and unambiguous determination of of the role of each of the solute
and solvent in crystallization from the solution phase has not been determined and remains
a subject of a great deal of speculation. This lack of information also affects the extent to
which control can be exerted over a crystallization operation in which a solute crystallizes
from a mixture of solute and solvent.

The role of solvent structure and properties and the relationship between solute and
solvent species in supersaturated solutions has not been a subject of research interest. As a
result, the spatial distribution of solvent and solute molecules in supersaturated solutions is
poorly understood; additionally, the functional roles each plays in the process of
supersaturation, nucleation, and crystal growth is unknown.

Another consequence of this lack of information is that there is no reliable way to
account for the solvent or its role in nucleation and crystal growth models. All theories of
crystal grth stem, historically, from the study and modelling of vapor or melt phase
separations. Such processes do not involve a solvent that has a different molecular structure
from that of the solute. Models of such processes thus neglect the presence of a solvent
which is chemically distinct from the crystallizing species and are unable to describe how it
affects the crystal growth process. These models describe the behavior of vapor and melt

phase systems well. Where a solvent differing in molecular structure from the solute is

4

present, as in crystallization from the solution phase, such models break down because the
parameters (which may be insufficient) lose their original physical significance and fail to
completely describe the behavior of the system

An understanding of solvent and the role it plays in determination of structure and
transport in supersaturated solution is necessary to explain several phenomena related to
crystal growth and nucleation from the solution phase. Unexplained issues related to such
phenomena include the origin of secondary nuclei , the growth of crystals from nuclei , the
tendency of certain substances, e.g., citric acid, not to nucleate from solutions even at very
high supersaturations, and epitaxial growth and polymorphism of nuclei as observed in

systems e.g., aqueous anhydrous glucose/ glucose monohydrate.

1.4 A Broad Overview of this Work
1.4.1 The Subject of Investigation

As a first step towards extending the understanding of the role of solvent and solute
in the structure and properties of supersaturated solutions and their roles in nucleation and
crystal growth it is useful to focus on how to accurately measure the one most critical
property in such phenomena; concentration in supersaturated solution and small changes in
concentration in supersaturated solution. Once it is possible to determine solution
concentration in supersaturated solution it is useful to analyze the experimental data to
obtain information related to the spatial distribution of the solute and solvent in
supersaturated solution. This information can then be related to other information relating to
the distribution of solute and solvent in solution and corroborated using another
experimental method. Finally it is useful to extend this technique to the determination of
local solution concentrations in solution on a time averaged basis both with and without the

presence of a crystal interface to determine whether such an approach has the ability to

5

provide any information regarding the loeal distribution and role of solvent and solute on a
time and space averaged basis.

This study is limited to the development of an experimental technique to determine
concentrations in supersaturated solutions of sucrose in water, to determine if such a
technique can provide information related to the spatial relationship between solvent and
solute in supersaturated solutions, and to relate such information to the body of information
available regarding the spatial relationship of solvent and solute in supersaturated solutions
of sucrose in water. Supporting experimental evidence in supersaturated solutions of
aqueous glucose, fructose, and lactose are presented when appropriate. Sucrose, and each
of the other sugars in this study, is a commercial sugar. Solution properties in dilute
solution, and crystallization data for each are readily available. Water is a well studied and

modelled solvent; its structure and behavior are fairly well known and documented.

1.4.2 The Experimental Approach

The experirrrental approach to this study has been to find a technique that allows the
detection of small changes in the local (microscopic) solvent concentration (in the vicinity
of a solute or other probe molecule) in supersaturated solutions where there is relatively
little solvent as compared to dilute solutions. Such a technique can then be used to
determine a standard calibration curve of sorts for concentration vs some solution property
in concentrated saturated and supersaturated solutions. In this way a bulk averaged value of
local supersaturation readings can be determined. The second phase of this study relates to
trying to determine truly local concentrations based on the findings of the first phase.

The focus of this work is to examine the basis for developing means to monitor
supersaturation in situ in a manner sensitive to small changes in solute and/or solvent
concentration, even in supersaturated solution. Once such means are identified the object is

to to test some of their limitations. The last step is to attempt to extend such techniques to

6

monitoring concentrations in the near interface region of growing crystals. At each step an
effort is made to correlate the information to understanding the spatial and functional
distribution of solute and solvent molecules in solution in order to shed further light on the
structure of supersaturated solutions. The determination of how and if the structure
observed affects the nucleation or growth processes is not attempted. Instead the object is
just to develop the means to observe such structure.

The usefulness of steady state fluorescence spectroscopy of extrinsic fluorescent
probe molecules as an experimental technique to determine supersaturation is first
demonstrated. The technique is then used in conjunction with fluorescence imaging to
investigate local concentrations in the absence of growing crystals. Finally, fluorescence
imaging is used to investigate concentration fields near the interface of growing sucrose

crystals.

1.4.3 Objectives

The primary objectives of this study are

1. to develop an experimental procedure that allows one to determine the concentration of a
saturated or supersaturated sucrose solution either directly i.e. , by measurement of a
solution property and correlating changes in this solution property to solution concentration
or indirectly i.e, by correlating changes in the quantifiable behavior of an extrinsic probe to
corresponding changes in solution concentration in saturated and supersaturated sucrose
solutions. The procedure or technique must be sensitive to, and capable of differentiating,
small differences in solution concentration in supersaturated solutions.

2. to demonstrate that such a technique may be applied, or modified so as to apply, to
measurement of concentration in supersaturated solutions of other sugars such as glucose

and/or fructose,

7

3. to briefly exarrrine the ability of this technique to provide information related to the
microscopic structure of supersaturated sucrose solutions by relating the analysis of
experimental data to related experimental data published in literature, and to published
studies of molecular dynamics of aqueous sugar solutions as well,

4. to investigate the extent to which this information is corroborated by time resolved
fluorescence spectroscopy which provides somewhat different information on different
space and time scales, and

5. to extend and modify this technique via two-dimensional microscopic imaging so as to
be able to determine time averaged local concentrations along a two dimensional cross—

sectional plane of supersaturated sucrose solution.

1.4.4 An Overview of the Chapters

A discussion of research in the field of solution phase crystallization is presented.
literature related to structure and property determination in solutions is also reviewed in
Chapter Two. Concepts are defined in terms of their use throughout this work.

The next few chapters cover the research as it was done and published in
chronological order, including details of materials and methods, data acquisition and
analysis, and results. The first paper, Chapter Three, deals with detection of small changes
in supersaturation rapidly, accurately, and in a manner suitable for use in on-Iine analysis
and control using any means possible. For this purpose spectroscopic techniques like
Raman, FTIR, and NMR were initially considered. NMR studies reported in literature
indicated the possibility of microheterogeneities in the system (6), but these techniques are
slow and inherently not suitable for meeting the objectives of this investigation.
Fluorescence techniques are commonly used in the study of microstructure in
microheterogcneous systems (7). There are several treatises delineating the procedure and

techniques available and the various kinds of information that can be extracted about the

8

molecule and its environment (e.g., 8—11). Similar techniques had already been used to
analyze water content in a variety of systems such as gels, sols, and sol-gels (12-18). It
was decided to use the extrinsic fluorescent probe, pyranine or 8-hydroxy-1,3,6-
pyrenetrisulfonate (18,19). The object of performing emission experiments on sucrose
solutions doped with pyranine was to correlate changes in the emission behavior of
pyranine to changes in solution concentration and to find an experimentally observed or
derived quantity that is very sensitive to small changes in solute and/or solvent
concentration in saturated or supersaturated solutions.

Fluorescent molecules in excited states are extremely sensitive to their solvation
environment It is possible to relate the changes in observed emission behavior to structural
changes in the microscopic solvation environment of the molecular probe (7, 18). The goal
was to use this information to correlate changes in the solvation environment of the
molecular probe to corresponding changes in the microphase distribution of the solute and
solvent molecules in solution. This work is introduced in Chapter Three and extended in
Chapter Four.

The next step involved investigating other sugar solutions with pyranine to
determine if the behavior observed was limited to sucrose or a more general trend among
sugars. The goal was to modify the experimental procedure, or data analysis, or both, to
enable supersaturation sensing and structural analysis. This is detailed in a second paper,
presented in Chapter Four. Chapter Five deals with the subject of simultaneous activity and
concentration measurement in a third paper.

Once information about the functional microphase distribution of the solute and
solvent was available it was then a goal to experimentally verify the validity of the
assumptions used to arrive at this conclusion. These assumptions pertained to the
assumption of a homogeneous distribution of the extrinsic probe molecule in solution and

the isotropy and homogeneity of the distribution suggested by analysis of the steady state

9

data This was accomplished using time resolved fluorescence spectroscopy. This is the
subject of the fourth paper, Chapter 6.

Scanning fluorescence imaging was used to investigate the distribution of solvent
molecules near the interface of growing sucrose crystals. This work is presented in the fifth
and final paper, Chapter 7.

The final chapter includes recommendations for future study. A summary of this

work and conclusions are also presented in Chapter 8.

1.5 References

1. Jancic SJ. and El de Jong eds., Industrial Cgstallization '84, Elsevier, Amsterdam,
(1984)

2. Randolf A. D. and M. A. Larson, Theory of Particulate Processes, Academic Press,
New York, NY (1971)

3. Randolf A. D. , Can. J. Chem. Engg., 42, 280, (1964).

4. Larson MA. and A. D. Randolf , Chem. Engg. Progress Symp. Ser., 65(95), 1,
(1969).

5. Ostwald W., Z. Phys. Chem, 22, 289-330, (1897)

6. Richardson S. J. and M. P. Steinberg, in Water Activity: Theoq and Applications to
Food, L B. Rockland and L. R. Beauchat eds, Marcel Dekker Inc., New York, 235,
(1987).

7. Kalyanasundaram K., Photochemim in MicrolLeterogeneous Systems, Academic Press
Inc., (1987).

8. Turro N. J ., Modern Molecular Photochemis_try, Benjamin Press, New York, NY,
(1978).

9. Lackowicz J. R., Principles of Fluorescence SpectrLoscopy, Plenum Press, New York,

NY, (1984).
10. O'Connor D. V. and D. Philips, Time Correlated Single Photon Counting, Academic
Press, Orlando, FL, (1984)

11. Demas J. N ., Excited State Lifetime Measurements, Academic Press, New York, NY,
(1983)

12. Clement, NR. and M. Gould, Biochemistry, 20, 1534 (1981).

13. Kano, K. and J.H. Fendler, Biochim. Biophys. Acta, 509, 289 (1978).

14. Kondo, H., I. Miwa and .I. Sunamoto, J. Phys. Chem, 86, 4826 (1982).

15. Bardez, E., B.T. Gouguillon, E. Keh and B. Valeur, J. Phys. Chem, 88, 1909,
(1984).

16.Pines, E and D. Happen. J. Phys. Chem, 87, 4471 (1984).

17. Kaufman, V.R., D. Avnir, E. Pines-Rojanski and D. Huppert, J. Non-Crystalline
Solids, 99, 397 (1988).

18. Pouxviel, J.C., B. Dunn and M. Zink, J. Phys. Chem, 93, 2134 (1989).

19. Kondo, H., I. Miwa and J. Sunamoto, J. Phys. Chem, 86, 4826 (1982).

Chapter Two

Literature Review

2.1 Background

Crystallization from solution is a unit operation of considerable importance in the
food, pharmaceuticals, and fine chemical industries (1). It is a separation technique that
yields high purity materials. Crystallization involves the transfer of mass from a vapor,
melt, or solution containing the material to be crystallized, onto a solid phase containing
nuclei, seeds or crystals of the same substance. This study is limited to aqueous sucrose
solutions; therefore, the discussion that follows will be limited to solution phase
crystallization. Mass transfer takes place against an interfacial concentration gradient. When
the solute loses its water (or solvent) of solvation, the interfacial region becomes more
solvent rich than the bulk solution; this gives rise to a concentration gradient. Mass transfer
of solute to crystal, and solvent to bulk solution, continues until the concentration of the
bulk solution approaches the saturation concentration.

A solution is defined as a mixture of two or more substances. The average particle
size in a solution is smaller than 10"7 m (2). It can be considered homogeneous on
macroscopic scales of volume e. g., a cubic centimeter. Inhomogeneities cannot he usually
be detected by such methods as light scattering. That substance which is present in the
largest molar quantity is the solvent Any substance present in a smaller molar quantity is a
solute.

A solution of a solute at a given temperature and pressure is said to be saturated if
solute added to the system at that temperature and pressure, after being allowed to
equilibriate, exists as a separate pure phase in equilibrium with the solution. The solubility
of a substance in a particular solvent at a given temperature and pressure, or saturation

concentration at the same temperature and pressure, is determined by the chemical nature of

10

11

the solute and solvent and their interactions. Only those substances which exist in the f orm
of a crystalline solid at a given temperature and pressure may be crystallized from their
solutions at those temperatures and pressures.

A solution of a solute which contains more solute than a saturated solution at the
same temperature and pressure is said to be supersaturated. If crystalline solute is added to
its supersaturated solution at constant temperature and pressure, and allowed to equilibriate,
the solution will tend to eventually attain saturation concentration with respect to the solute,
and the solute present in solution in excess of its saturation concentration will tend to
deposit on the crystalline solute. Thus the crystals will grow.

Increase in mass is an absolute measure of crystal growth (3). Crystal growth is
often described by an increase in characteristic length, for purposes of analysis. Different
measures of charactersistic length are in use, such as Volume of the crystal/Surface area of
the crystal. Characteristic length is not a very important parameter in this study, and thus it
will not be discussed in further detail. Charactersistic length is a parameter of major
importance while describing the growth kinetics, and crystal size distribution for a
crystallization process of interest. Randolf (4) together with Larson (5) developed the
population balance technique to predict the crystal size distribution in a system of given
operating system configuration, process parameters and growth kinetics. Such techniques
provide the analytical basis for on-line control of industrial crystallizations i.e., control of
the crystal size distribution or CSD. This work will not be discussed in any detail because it
is not relevant to this study. It is mentioned because there are significant practical
difficulties in applying population balance based analyses to control real crystallization
processes. Presently it is not possible to monitor solute and solvent concentration in
supersaturated solution both globally and locally on a continuous on-line basis.
Consequently, it is not possible to monitor CSD and the morphological quality of crystals
during growth and nucleation in a batch process, in a continuous yet non-intrusive manner.

Accurate and reliable information related to two key variables:

12

a a characteristic length "1" and

b. the supersaturation ”s"
is difficult to obtain, particularly on a local basis, e.g., in the neighborhood of the crystal
solution interface of a particular site on a growing crystal. The issue of monitoring
supersaturation both in the presence and absence of growing crystals is complicated by a
limited understanding of the phenomena of supersaturation on a structural and energetic
basis, and conflicting ideas and data relating to solution structure and transport. The
concept of supersaturation is easy enough to define in the sense of procedure or
phenomena Supersaturation is fairly easy to detect on a macroscopic scale; but, it is
difficult to quantitatively relate the saturation temperature, pressure, concentration, and
solution structure to chemical, physical, or transport properties of the solute and solvent
(6).

This work investigates the issue of measuring supersaturation in solution using
spectroscopy, both on a global and local basis for the aqueous sucrose system under
common industrial operating conditions. In this work global refers to the entire system, and
local refers to less than 10 microns in all directions . The term structure is used in this work
to refer to the spatial distribution of molecules in solution. Microstructure refers to structure
on the scale of 200 nanometers or less.

Sucrose was chosen as a substance suitable for investigation because its properties
and behavior are well known and also because it is a commodity of industrial significance.
It's properties have been tabulated and discussed in many articles. The reader is referred to
Kirk and Othmer's Encyclopedia of Chemical Technology (7) where a summary of
properties of sucrose and several excellent references may be found. Those studies, in
which properties which bear a direct relation to this work are determined, will be discussed
where appropriate.

Various spectroscopic methods were initially considered for measuring

concentration in supersaturated solution because they are capable of detecting phenomena

13

on a truly local basis. These methods record events on a molecular or sub—molecular space
scales, or on the time scales which correspond to such events. Further, interpreting data
obtained from such methods is fairly straightforward. There exists a vast and rich literature
that deals with various molecular spectroscopies, and spectroscopic analysis. Such
literature will be referred to as needed.

A limited amount of experimental data related to phenomena on such space and time
scales exists in literature. Of the systems for which some data exists, the aqueous sucrose
system is one of the best studied. The choice of the aqueous sucrose system was made in
the hope that any data obtained from this study could be compared with reference to
published results and add to the total understanding of the crystallization of sucrose from

aqueous sucrose solutions. A review of the literature pertinent to this study follows.

2.2 Crystal Growth from Solution
2.2.1 The Requirements for Crystal Growth from Solution

It is generally accepted that for crystal grth from solution to occur, the solution
must be supersaturated and there must exist a nucleus of greater than critical size. The
nature of the critical size of a nucleus is presently unknown. For growth at the interface of a
pre-existing nucleus to occur, the solute must get to the interface, release all solvent not
associated with the crystalline structure, assume an orientation favorable to the lattice

structure, and become incorporated into the lattice, not necessarily in that order (6, 8, 9).
2.2.2 Crystal Growth Kinetics
The rate of crystal growth is a function of the supersaturation of a solution.

Growth rate dispersion has been observed for a population of crystals in many

systems. Berglund showed that the growth rate of sucrose crystals is more or less

14

constant for every crystal, but that a dispersion in grth rates is often observed within
a population i.e., some crystals grow faster than others (10). Shiau showed for the
fructose water system that growth rate dispersion itself depends on supersaturation
(l l). Closely related to growth kinetics is nucleation kinetics which is also a function of

supersaturation.

2.2.3 Nucleation

Nucleation is a phenomenon by means of which a new crystal is formed.
Primary nucleation occurs in the absence of any pre-existing crystals. It may be
homogeneous or due to the presence of ordered solute molecules, or heterogeneous
i.e., due to the presence of dust or other particulate matter foreign to the system.
Secondary nucleation occurs due to the presence of growing crystals (13). It may occur
due to the shearing off of crystal dust on the crystal surface of uncured crystals (initial
breeding (13)), or due to collisions of crystals with each other, the impeller, or vessel
walls (collision breeding (14)), or due to sliding contact on solid interfaces (contact
nucleation). Most crystals produced in industrial crystallization from solutions are a
result of secondary nucleation (15). Secondary nucleation has an important effect on the
crystal size distribution, growth kinetics, and product crystal quality.

The origin and growth mechanism of secondary nuclei are areas of current
research effort There are two explanations for origin of secondary nuclei that are
prevalent in literature. One theory, developed by Mullin and Leci (16) and McCabe and
Clontz (17), is that the nuclei are produced as a result of coalescence of solvent-
exclusive pro-crystalline clusters of sub-critical size which are thought to exist in the
boundary layer at the crystal-solution interface. The other theory is that nuclei of critical
size are produced due to breakage of micro-dendritic outgrowths on the crystal surface

15

are sheared off by the fluid surrounding it (18), or that pre-crystalline clusters of critical
size are sheared away from the layer of solution adjacent to the crystal face (19).

The conventional analysis of growth incorporates two types of resistances. One
arises from volume diffusion or the resistance that a solvaton (a molecule of solute and
the solvent most associated with it (20)) experiences when it is transported from the
bulk solution to the interface. This resistance is probably very important in systems that
contain significant amounts of non-solvation and non-bulk solvent in supersaturated
solution. This volume or bulk diffusion resistance was first proposed in the crystal
growth theories of Volmer, Brandes, and Stranski. A detailed description of these
theories may be found in Crystal Growth by HE. Buckley (21).

The other resistance is thought to be experienced by a partially solvated solute
growth unit as it traverses from a point on the crystal surface to a "kink-site" so that
growth can proceed via propagation of a dislocation. Burton, Frank and Cabrera (22,
23) first proposed this mechanism to explain crystal growth from vapors onto real
crystals with surface defects and other imperfections, and then later extended the
analysis to growth of real (imperfect) crystals from solutions (hence referred to as the
BCF theory).

Nuclei being real rather than perfect crystals, are thought to grow by
propagation of growth defects like growth spirals. Some sites on the crystal surface are
thought to be more likely to result in growth than others, in particular an edge of a

surface dislocation or other defect These sites are called kink sites.
2.2.4 The BCF Theory of Crystal Growth
There are many theories of crystal growth from solution. The BCF theory is

perhaps most widely accepted The BCF theory proposes that growth units are present
in the solution or boundary layer surrounding a crystal. Such a growth unit is thought

16

to diffuse from the bulk solution to the crystal surface, be adsorbed on the surface so
that it is only mobile in two dimensions, and then diffuse to the kink site.

The BCF theory proposes that the dependence of growth rate on supersaturation
follows an Arrhenius type functionality with a clearly defined exponential coefficient.
The exponential exponent represents the primary mechanism of growth resistance, i.e. ,
limit 1 for volume diffusion and 2 for surface integration. Surface integration, is the
resistance experienced by a growth unit trying to incorporate itself on the surface of the
crystal. Such a growth unit is thought to be limited to two dimensional mobility, that is,
capable of diffusion in two directions. Volume diffusion, is the resistance experienced
by a growth unit in diffusing from bulk solution to the surface of the crystal.

The BCF analysis was originally intended to explain crystal growth from vapor
phase or melts. In the case of growth from the vapor phase the growth unit was clearly
defined as a single molecule of solute; in the case of melts it was convenient to define
the growth unit as one or more molecules of crystallizing substance. The BCF theory,
extended to crystallization from solutions, is understandably vague regarding several
key issues, among them the structure and composition of the growth unit, the role of
the solvent in crystal growth, at what point desolvation occurs, and what happens to the
desolvated solvent Over the years, particularly with the work of Strickland-Constable
(12, 14, 18) and also that of Clontz and Mcabe (17) , there has been a tendency to think
of a growth unit as a solvent exclusive pre-crystalline cluster of anywhere between 10's
and hundreds of thousands of solute molecules comprising organized structures smaller
than crtitical nuclei. It has been suggested that such clusters exist in saturated and
supersaturated solutions(24, 25, 26). It has also been suggested that perhaps the
presence of a nucleus or a growing crystal stabilizes such aggregates and that these may
be the source of secondary nuclei. There is no conclusive experimental evidence which

proves that this is the case.

17

2.2.4.1 Assumptions related to Solution Structure Implied By the BCF

theory

Although it is not clearly evident, the BCF theory pro-supposes that the
supersaturated solution is dilute in terms of absolute molar ratio of solute to solvent,
and hence the resistances proposed explain those instances where growth is transport
limited. In fact, transport limitations are proposed to be the only mechanisms of growth
resistance. What happens when an abundant supply of solvated solute is immediately
available at the crystal-solution interface, for example, is not considered.

The mechanism of surface integration assumes that there is very poor chemical
affinity between the solute and the solvent which allows the growth unit to remain
desolvated even though the solute is exposed to solvent many times during that time
interval. It further presupposes that the chemical affinity between the solute molecules
isn't very great either, because in order for a defect to propagate, and crystal grth to
occur, the growth unit must diffuse over an existing solute lattice to a kink site
(favorable for growth). Surface integration requires a partially solvated partially bonded
solute molecule to ”travel" across the crystal surface making and breaking bonds in one
dimension. The BCF theory thus implies that solute once (even partially) de-solvated,
is likely to stay that way for lengths of time on the scale of a diffusion process. Such
time-scales are at least an order of magnitude larger than lengths of time associated with
solvation-desolvation processes. Surface integration has appeared in various forms in
the growth models proposed by Gilmer (27), Chemov (28), Bennema (29) and others.
There is no conclusive and direct experiemental evidence available in support of surface
integration.

The BCF theory is limited by the fact that only two mechanisms of resistance to
crystal growth are proposed. In fact, there is experimental evidence that suggests

resistance to crystal growth may be due to other mechanisms in certain systems; for

18

example, in the secondary nucleation kinetics studies of fructose (30), lactose (31) and

ammonium dihydrogen phosphate (32).

2.2.5 Evidence for Non-Cluster Dependent Growth Mechanisms

There is evidence in literature which suggests a molecular grth mechanism
for the crystallization of certain organic compounds, among them sucrose. The reader is
referred to the work of Addadi and Lahav with amino acids (34-43) and Dunning and
Albon (44) and Smythe (45, 46) with sucrose. In the case of amino acids systems
studied were such that the absolute molar ratio of solute to solvent is low, even in a
supersaturated solution where the solute concentration is double that at saturation. The
solubility behavior of the chosen impurities was similar. Addadi and Lahav required
concentrations of several hundred grams per liter of the impurity to notice clear
morphological changes in crystal habit (growth inhibition without morphological
changes may be possible at much smaller impurity concentrations). Impurity
concentration in the crystals obtained was normally on the order of several thousandths
of a mole percent Crystal growth was always accomplished via seeding, suggestive of
the possibility that homogeneous nucleation is unlikely even at such high degrees of
supersaturation possibly because the likelihood of two (or more) solute molecules
desolvating and bonding to start the sequence of events leading to nucleation is small.
Addadi and Lahav have at all times suggested a molecular mechanism of growth and
inhibition and have not once alluded to clustering as an explanation of growth. Thus it
is possible, and even likely, that tailor made impurities act on a per molecule basis,
suggesting that grth is a sequence of molecular events. It is therefore possible that
systems which are affected by tailor made impurities experience crystal growth via a
molecule by molecule accretion process rather than by accretion of solvent exclusive

clusters.

19

2.3 Solution Structure

2.3.1 The relevance of structure to crystallization research

The study of structure in supersaturated solution derives its justification as a
direct result of three groups of experimental studies: studies of the inability of certain
substances to nucleate even at very high supersaturations from their solutions, studies
in epitaxial growth where the parent crystal is observed to support the grth of a non-
identical crystalline phase, and studies of polymorphism as a result of secondary
nucleation wherein contact of parent crystals resulted in the observation of more than
one crystalline form of nuclei in solution. Epitaxy refers to packing of the atoms or
molecues of a crystal and relative distances between atoms or molecules; morphology,
and habit both refer to external differences. Morphology refers to the basic
classification of the crystal eg; mono-clinic, or orthorhombic, habit refers to relative
extents of different faces given a particular morphology.

Tamman (47) and later Mullin and Leci (16 a.) demonstrated that certain
substances, for instance citric acid in aqueous solutions, formed glass-like amorphous
solids at high supersaturations which did not nucleate. Mullin and Leci (16 b.)
concluded that the ability or inability of a solution to nucleate must somehow depend
upon the structure of the supersaturated solution. They reasoned that the molecular
structure of citric acid favors intermolecular hydrogen bonding and association. Chains
or clusters of solute molecules in supersaturated (and even concentrated but
undersaturated) solutions were proposed as precursors to nuclei. This proposition will
henceforth be referred to as the cluster hypothesis. Subsequently, Mullin and Leci
studied the desupersaturation of citric acid solutions, and applied the cluster hypothesis

to explain how and why nucleation did or did not occur from supersaturated solutions.

20

It is important to note that Mullin and Leci, did not take a conclusive view as to the size
of the clusters nor did they rule out the possibility of solvent inclusive clusters. In
subsequent studies (none of which were conducted by Mullin or Leci), however, the
term "precrystalline clusters” has come to imply solventless solute clusters of 10 to
10,000 molecules.

In the study of contact nucleation of saccharides, Elankovan and Berglund (48,
49) and later Cerreta and Berglund (50, 51) showed that under certain conditions the
contact of a parent crystal of a particular form and composition yielded some nuclei of a
structure and composition different from that of the parent crystal in addition to nuclei
that were of the same structure and composition of the parent crystal. Cerreta observed
that under certain conditions the parent crystal supported growth of the non-identical
nuclei while under other conditions it would not. As a result, he concluded that neither
the attrition theory (small pieces of crystal which break off the surface of the parent
crystal by attrition serve as nuclei or precursors to nuclei) nor the cluster theory of
origin of secondary nuclei provided satisfactory answers as to the origin and
mechanism of this phenomena, and that there was probably a third mechanism
responsible.

In each instance the researchers conducted careful global or macroscopic
measurements of temperature, pressure and concentration, and were able to determine
the supersaturation of the solution. They had enough information to determine
macroscopic transport properties from correlations. Yet in each instance, the
researchers were of the opinion that a direct cause—effect relationship exists between the
local or global structure of the solution (not the macroscopic or global transport
properties) and the phenomena they observed. They were unable to define this
relationship because adequate information related to solution structure was not available
to them. They did not, in general, have any comments regarding what scale or scales of
structure were of significance to the phenomena they were studying.

21

2.3.2 Ideal Solution and the classical treatment of solution structure

The classical or continuum view of a solution is stated by Smith and Van Ness
in the third edition of Introduction to Chemical Engineering Thermodynamics (52) thus:
"Actually, the components of a solution are intimately intermixed and cannot have
private properties of their own. Nevertheless they may be assigned properties on some
arbitrary but universal basis, and (the pag’al molar propegty), defines how a solution
property is apportioned among the components of a solution. " The partial molar volume
of the i “1 component of a mixture wherein the number of moles of all other
components are fixed, at a specified temperature and pressure, is defined as the partial
deriviative of solution volume with respect to the number of moles of i.. A solution in
which the partial molar volumes of the components are equal to their pure component
volumes is said to be an ideal solution. The f ugacity of a component is a property of
that component with the units of pressure, such that in the limit of zero pressure the
ratio of fugacity to pressure has the value of unity, and the following equation defines
fugacity
dGi =Rlenfi,

Where the subscript i denotes the i m component and:

G = Gibbs Free Energy, Joules

R= Gas Constant, 8.314 Joules / mole 0K

T: Absolute Temperature, OK

The fugacity of a component in solution is defined by a similar relationship with the
partial molar Gibbs Free Energy, subject to the same condition with respect to pressure.
The fugacity is thus a property analogous to pressure, and its magnitude is a measure of
deviation from ideal gas behavior. The ratio of the fugacity of the i “l component in

solution to the fugacity of that component in some reference state is the activity, a ,2 .

22

The activity is an exponential function of the difference in the partial molar Gibbs Free
Energy and the pure component Gibbs Free Energy of a reference state, at a given
temperature and pressure. The activity is the most frequent choice as a measure of
deviation from ideal solution behavior because it is directly related to the difference in
the partial molar Gibbs Free Energy and the pure component Gibbs Free Energy of
some reference state. In the particular case of sugar water systems this is of
considerable practical significance because the correlation between shelf life, texture,

biological stability of foods, and water activity, aw, is well documented. Several

comprehensive reviews of the influence of water on chemical reactions in food are

available(53 -57). The measurement of aw, in concentrated and metastable aqueous

solutions of sugars is also an area of active interest (58). Official Methods of Analysis
of the AOAC (1984) 32004-32009 details analytical methods for determination of aw.
Hair hygrometers, electronic hygrometers, and isopiestic methods are commonly used.
Electronic hygrometers work on the principle that that changes in conductance,
resistivity, impedance, resistance, or capacitance of a thin film or an aqueous solution
of a hygroscopic substance can be correlated to corresponding changes in aw(59).
Rugged, stable, rapid and precise techniques for evaluation of aw are required. There is
a need for techniques that can rapidly, accurately, and simultaneously measure solute
concentration and water activity.

An excess property is the difference between an actual property and a property
that would be calculated at the same temperature, pressure and molar composition for
an ideal solution.The molar excess Gibbs Free Energy, GE , is related to the activity
thus:

GE/RT =2 ( x; In (54 / x,- ) ). The quantity in the inner brackets is the activity
coefficient for component i, y; , described by:
Y i = (a,- / x; ). The activity coefficient is also a widely used measure of deviation from

ideal solution behavior. The work of Kozak et al. succeeded in relating structural

23

effects of interactions between pairs of molecules to the activity coefficient; however,
there is no protocol to differentiate and account for site specific differences in
interactions between parts of molecules. Understanding such differences in interactions
may be a key element in understanding the relation between structure and other solution
phase phenomena.

It follows from this discussion that fugacity, activity, and activity coefficients
are all measures of the extent to which a solution deviates from the behavior of an ideal
solution. Each of these are in fact arbitrary quantities, none of these may be related to
an actual spatial or temporal quantity, and all of these are in fact based on the kinetic
theory of ideal gases as a model, and are measures of how much a real solution deviates
from a solution of ideal gases. None of these can be directly measured and each of
these assume that the solution is a continuous and microscopically homogeneous phase.
Further the kinetic theory of gases, assumes that the molecules of gas are hard spheres,
with mass but no intermolecular forces of attraction.

Models of ”real gases” deal with the consequence of the molecules having a
finite condensation volume and intermolecular forces of attraction. Yet molecules are
still treated as spheres and differences in relaxation and exchange rates among the
different moeities within a molecule are not taken into account, nor are differences in
function among molecules of the same species within a mixture. Such models do not
account for the fact that some fraction of molecules of a given species, particularly the
solvent in a solution, may be engaged in solvation of solute at any given time while
others may be engaged in pure component like behavior at the same time. Even the
most advanced treatments of solution thermodynamics such as those of Ben-Naim and
co-workers(20) which treat solute molecules and the solvent associated with it for the
longest time as a ”solvaton" in a sea of solvent, do not account for differences in
exchange rates at different moieties on the same solute molecule. These issues become

significant when time and length scales associated with solvation and desolvation

24

becorrre significant to the description of a process, as is the case for crystallization from
solution.

The relationship between supersaturation and deviation from ideal solution
behavior has often been used to define or describe structure in supersaturated solutions.
The traditional and accepted analysis is based on a modified version of Komk et al's
(60) approach of relating molecular neighbor pair potential functions to the excess
Gibb's free energy and thus the activity coefficient for each molecular species in
solution. Barone and co-workers (61—64) have demonstrated this technique. They
analysed solution structure by examining the concentration dependence of the
magnitude and nature of molecular pair potential f unctons calculated using a numerical
approach.

Myerson and co-workers (Sorrel, (65) and Chang, (66-69)) have attempted to
directly relate activity to solution structure through analytical correlations between
activity and other solution properties. They painstakingly and meticulously measured
diffusion coefficients of both non-electrolytes (urea, glycine, and glycine-valine) and
electrolytes (sodium chloride and potassium chloride) in solute diffusion studies in
concentrated aqueous solutions. Precipitous drops in diffusion coefficients were
observed near the saturation point in each case. The results indicate formation of solute-
solvent aggregates of an unknown composition. Whereas the measurement of the
diffusion coefficient was careful and direct, in each case, the calculation of cluster size
or aggregate size was not a very straight forward process and involved the implicit
assumption that the solution may be treated as a continuum. Transport phenomena in

solution are detemrined by non-equilibrium solvation dynamics which are governed by
short range interactions. These take place on a variety of time and length scales that are
shorter than the time and length scales associated with diffusion. This evidence will be

reviewed in another section. Myerson and coworkers attempted to relate the diffusion

25

coefficients to the structure of the solution resulting from short range interactions
without taking solvation dynamics into account.

Larson and Garside (24,23 performed an experiment with isothermal columns
of concentrated, saturated, and supersaturated solutions of citric acid, urea, sodium
nitrate, and potassium sulfate during which they observed concentration gradients along
the length of the column for only the supersaturated solutions. Larson and Garside
concluded that there must be solute structuring in supersaturated solutions. By making
the classical assumption that all solute in solution must be associated with other solute
molecules because of an absence of solvent to satisfy the solvation requirements, the
solvent exclusive clusters were proposed to be comprised of several thousand
molecules each. Myerson63 in recent work has argued that if the constraint of solvent
exclusion is relaxed, then the probable size of clusters would be two or three solvated
molecules of solute. Cerreta and Berglund (51,71) used Raman spectroscopy of
ammonium dihydrogen phosphate (ADP) to show that while no great differences could
be observed in the spectra of very concentrated, saturated and supersaturated solutions,
evidence did exist for dimer formation. Larson and Garside(24) cited Cerreta and
Berglund (51) as well as the Raman studies of Hussman et al.(7l) and MacMahon et
al.(73) with alkali nitrates as evidence of pro-crystalline clusters in concentrated
solutions.

Implicit to each approach is the assumption that any one molecule of a given
chemical species is considered to be identical in function to any other molecule of that
chemical species. Pair potential functions are time and space averaged and do not
account for dynamic structure and/or interactions on submolecular scales in a manner
that allows direct correlation with physical variables that can be directly measured;
neither do excess energies and activity coefficients derived therefrom The fact that
some solvent is involved in solvation and other solvent behaves as a pure component is

not accounted for, nor the dynamic nature of this distribution. The possibility of any

26

relation between deviation from ideal solution behavior and such a difference in the
function of the solvent is a-priori excluded from such analyses. This limitation arises
because of the implicit assumption that matter in solution is in a homogeneously

continuous phase at all times.

2.3.3 Other Concepts of Solution Structure

The fact that all water molecules in solution at any given instant do not have
identical solvation environments, even in pure solvent, has been accepted for a long
time now, e. g. Bousfield and Lowry (74). From the earliest post-quantum discourses
on solution structure and properties it is clear that a relationship between change in
solution structure and deviations from ideal solution behavior was thought to exist
although the specifics of cause-effect were undetermined. The focus on solution-
structure has been primarily limited to the study of the immediate solvation environment
of the solute molecules(75,76). In many cases the structure of this environment has
been observed to change little, if at all, with solution concentration up to saturation,
even though deviations from ideal solution behavior over the same range of solution
concentrations are observed to be significant. As a result, strong correlations between
structure and deviations from ideal solution behavior were not apparent. There has been
some speculation as to changes in the structure of the solvent that does not participate in
the solvation of solute with changes in concentration (77-81); however, until recently
there were few reliable estimates regarding the nature of such changes(83).

Physical chemists at the turn of the century were well aware that solution
structure (at that time attributed only to aqueous solutions) is a consequence of the
functional distribution of water molecules into "bulk" and ”solvated" rnicrophases. This
is to say that some of the water in solution behaves as if it is not associated (bulk or

free), and some behaves as if it is (solvated or bound). Even in pure water not all the

27

water molecules (or even parts of water molecules) are associated to the same extent
Bousfield and Lowry thought that water molecules existed as ”dimers and trimers"(74).
Pauling championed the idea of ”tetrahedral clusters” (75). Frank and Wen (84)
modified the tetrahedral cluster concept to reflect the dynamic natutre of distribution of
water molecules between clusters of different sizes , or ”dynamic or flickering
clusters”.

Most ideas regarding solution structure that are accepted today are in some way
related to Hildebrand's work (85). He proposed that in order for a molecule of solute to
enter solution it must first break the bonds with its nearest neighbors in the solid lattice,
then displace a molecule of solvent from the solvent lattice, and finally develop bonds
with its nearest neighbors in the solvent lattice.

Thus the processes of solvation and desolvation were proposed to be intimately
connected with the concept of solution. Frank and Wen and Némathy and Scheraga
(86) furthered the idea of structure in water by proposing the concept of a distribution
of clusters and flickering clusters respectively, wherein the chemically identical water
molecules are functionally distributed in solution according to how they participate in
nearest neighbor interactions defined in a large part by their ability to participate in
proton exchange.

A water molecule may be unable to participate in proton exchange in a solution
if it is hydrogen bonded to either another water molecule or a solute molecule. Water
that behaves in such a manner has been designated "associated", ”bound", or "water of
solvation (specifically applied to water hydrogen bonded in the proximity of a solute
molecule)". Water molecules may also be able to participate in solvent exchange and
exhibit behavior that is indicative of the absence of any solute molecules, such water
ins been designated "free”, "unbound”, or 'unassociated". In this work we will refer to
water which does not participate in solvent exchange on the time scale of the experiment

as "solvation water”, and water which behaves as if there were no solute molecules

28

present as “bulk water”. Water which behaves in a manner which is not consistent with
either solvation water or bulk water will be designated non-bulk-non-solvation water.
From this discussion it is apparent that it is important to account for not only what
points in space are occupied by solvent and solute molecules at given moments in time,
but also what role each molecule of solvent takes on at that time. For this purpose we
will use the terms spatial distribution, functional distribution, and functional
microphase distribution. The term ”spatial distribution" refers to how solute and water
or other solvent molecules are distributed and arranged in space in a solution. The term
”functional distribution" of solvent refers to the moles (or molecules) of water in
solution which can be assigned to each category. The term ”functional microphase
distribution” refers to the number of water molecules in the neighborhood of a solute or
probe molecule or moeity (when specifically designated) or some microscopic region in
space less than 200 nm in extent (when no specific molecule is designated) which can
be assigned to each category. It is also apparent from this discussion that the dynamics
of the molecules in solution is intimately related to the solution structure. In an absolute
sense, the discussion of one can only be incomplete without discussing the other. In the
following discussion we therefore review information related to both structure and
dynamics of aqueous sucrose solutions.

Franks (87-89), and Sugget (90-93) and coworkers in a series of studies on
various sugar solutions demonstrated that water in both dilute and non-dilute solutions
is also functionally distributed. Birch et. al. (94) demonstrated the functional
distribution of water on the basis of proton exchange rates in pulsed NMR experiments,
as did Richardson and Steinberg (83) in an NMR study of sucrose solutions of various
concentrations (including nearly saturated, saturated, and supersaturated solutions) in
the presence and absence of growing crystals. Each of these studies provides an
estimate of the amount of water that immediately solvates the sugar molecule in

solution, and thereby sheds light on the nature of short range and short time solute-

29

solvent interactions. None specifically addresses the issue of the structure or behavior
of the non-solvation water, and if the microphase distribution changes with
concentration.

Previous investigations have studied the nature of various molecular
associations by utilizing a variety of means. In the study of supersaturated solution
structure many studies have been conducted on the scale of molecular diffusion. Some
of these include the column studies of Larson and Garside (24-26), the studies of
Narayan and Youngquist (95), and the studies of Myerson and coworkers (65-69).
Dielectric relaxation methods (Suggett and Clark (92)) have identified two distinct
relaxation processes for water in sucrose solutions. These are designated associated
water, hydrating the sucrose, and bulk water comprising the remaining water in
solution, not associated with solute. Associated water has a longer relaxation time than
the bulk water. The 2H and 170 NMR studies of Richardson and Steinberg (83), on
the mobility of water in sucrose solutions, indicate three regions of decreasing
rotational rates of water molecules with increase of sucrose concentration through
saturation and continuing into the supersaturated regions. These effects were attributed
to the development of an extensive but dynamic network of hydrogen bonds of water to
sucrose, water bridging of the sucrose and direct sucrose to sucrose interactions
(although no direct evidence of sucrose-sucrose interactions were observed). The bulk
viscosities of these solutions also increase with increasing sucrose concentrations, in a
manner consistent with the rest of the data, suggesting that the same short range
interactions define both structure and transport

The studies of Myerson and coworkers (65-69) are particularly interesting
because they utilized Guoy Interferometry to measure diffusion coefficients. The
interesting aspect is that Myerson used continuum arguments to reach his conclusions.

This would indicate that for the systems under consideration the macroscopic transport

3O

coefficients in supersaturated solution must be very similar to the microscopic transport

coefficients in supersaturated solution.

2.3.4 Non-Continuum Descriptions of Solvation and Structure

A more recent trend in the description of solvation and solution structure is the
effort to develop non-continuum models to describe order, dynamics and interactions.
This work has been pioneered by the pico- and femto—second spectroscopy of Marcus
and coworkers. Structure, transport properties and thermodynamics of phase equilibria
are manifestations of interactions on submolecular length scales, e.g. hydrogen
bonding, hydrophobic interactions, which require anywhere from a picosecond to a
nanosecond to take place. The fact that solution structure and transport phenomena are
related in a way that structure governs transport properties and transport phenomena are
a directly measurable manifestation of structure, enables us to extract important
information about the structure from the study of transport processes and vice versa
Transport properties in solution are governed by non-equilibrium solvation dynamics.
There is evidence to suggest that transport phenomena are governed by short range
interaction processes that take place on a variety of time and length scales, which are
less than those associated with molecular diffusion (96,97). The work of Maroncelli et.
al.(98), Kenney-Wallace, Jonah, and co-workers (99-102), Wolynes ( 103, 104),
Cross and Simon (105,106), Migus et.al.(107), and Brady (108) demonstrates that the
time scales involved in the solvation (and desolvation) process, as well as the molecular
re-orientation process, are much faster than that of molecular diffusion. They suggest
that the length scales involved in the interactions most significantly associated with
desolvation and molecular reorientation, with respect to incorporation of solute
molecules in the crystalline lattice, are small compared to those associated with

molecular diffusion. Continuum models are still prevalent for solvation dynamics.

31

These models do justice to the long range interactions in solution, but fail to account for
the details of structure and interaction of inner solvation shells which are often
dominant factors (1(8).

There are now several non-continuum approaches to understanding solvation
dynamics. They differ primarily in the choice of a basis set of significant length/time
scales. One such widely discussed approach, the mean spherical approximation or
MSA, indicates that solvation of an ion or molecule occurs on a range of time scales. A
molecule can be thought to be made up of component dipoles. The limiting length scale
of each such component is on the order of length of the hydrogen bond. Local solvent
structure and dynamics near each of these components has been observed to differ
significantly. Even the solvation of a simple alcohol must be described by at least two
different time scales. One corresponds to the continuum longitudinal relaxation time,
while the other is associated with the time required by solvent molecules to rearrange
themselves near the dipole. For small dipoles or ions, in dilute solution, this time is of
the order of a picosecond (103,104). Furthermore, the molecular dynamics simulations
of Cross and Simon (105), suggests that if a dipole is strong enough (>17 D) then the
two ends may start behaving like point dipoles in the sense that the localized solvent
structure and dynamics may be different at each of the ends of the dipole. This is
summed up by Maroncelli et. al.(98) who conclude that the solvation response in a
polar solvent involves coupled reorientationall dynamics that occur on time scales
shorter than single particle diffusion. There have been studies in the solvation of an
electron in water and alcohols, using picosecond and femtosecond time correlated
spectroscopies, which indicate that solvent reorientational times are on the order of <
500 is (107). Brady's study (108) of the molecular dynamics of solvated glucose in a
15 wt% solution also indicates that the local structure and dynamics near each of the O

or CH moieties of the glucose molecule differ, as do the solvent relaxation times.

32

The significance of this information, simply put, is this: although solvation and
desolvation must necessarily play critical roles in the crystallization process, there are
no parameters for either length or time scales involved in these phenomena, in the
description of crystal growth or nucleation mechanisms, kinetics, transport, or
subsequent analyses of crystallization processes. Furthermore most models of solvation
in use are continuum models, it is not possible to use a non-continuum approach like
MSA for a sucrose-water system because the relevant length scales or time scales have
not yet been identified, besides which the large number of sites at which solvent
exchange occurs on a sucrose molecules may make such an appraoch impractical.
Therefore on the one hand non-continuum appraoches are still in their developmental
stages, and on the other the usefulness of any model of crystal growth based on an
approach which does not account for solvation and or desolvation will necessarily be
limited. The exent to which such limitations affect the ability to control crystallizations

is not an easy issue to address, and is beyond the scope of this work.

2.4 Techniques commonly used to measure concentration in supersaturated

solutions

For industrial sugar solutions temperature measurement at a known pressure is the

most common technique of measuring supersaturation. Industrial sugar solutions are

normally crystallized by means of evaporating the water by means of boiling the solution.

On the basis of years of experience the temperature of a sugar solution can be accurately

correlated to its concentration or supersaturation.

Refractive index, viscosity, and density are other common means of determining

the concentration of a sugar solution. Additionally water of solvation data obtained via a

variety of techniques like NMR and dielectric relaxation can be related to solution

concentration.

33

2.4.1 Spectroscopic Sensing Techniques

All spectroscopic sensing techniques are based on extracting information about
the structure of a molecule or a part of a molecule, its environment, or the interactions
between them The literature which describes these techniques is vast and the how and
why of these techniques are largely not of significance to this work; however, the
results of studies of solution structure using these techniques affect this work a great
deal. A review of the results of various studies of aqueous sucrose solutions using

these techniques follows.

2.4.1.1 Raman SMU‘OSCOQ!

Recent experimental attempts to study interactions in supersaturated aqueous
sugar solutions have involved the use of Raman, IR, and NMR spectroscopies and X-
ray diffraction. X-ray diffraction, Raman and IR spectroscopy are sensitive to changes
in vibrational frequencies and space lattice ordering that occur as the concentration of
the solution increases. The effects of molecular association, and hydrogen bonding are
observed in Raman and IR spectra. Mathlouthi et al.(109-111) performed vibrational
spectroscopy of sugar solutions of various concentrations and observed that the
characteristics of the spectra were concentration sensitive. Bands assigned to the
assymetric vibrations of the CH and OH groups are observed to shift in frequency. In
an attempt to correlate the change in solution structure to change in concentration, the
ratio of the integrated band intensity at 1460 cm'l, assigned to the HCH bending mode,
to that at 1640 cm'l, assigned to the HOH bending mode was plotted as a function of
concentration. The shift in these frequencies were plotted against sugar concentration.
A ten wavenumber shift (approximately 1%) , from 1450 cm-1 to 1460 cm-1, was

observed in the band associated with the HCH bend as the concentration of sucrose in

34

solution increased from 10 to 66 wt%, and that of glucose increased from 5 to 54 wt%.
This indicates that the HCH bend can be accomplished with less energy in dilute
solution than in-vacuo or in crystalline form, the two standards that are commonly used
for calculation of band positions and contributions (112, 113).

While this indicates that this particular vibrational mode is about as hindered in
concentrated solution as it is in vacuo or in crystalline state, all one can safely conclude

from this is that the CH2 moiety must be hydrophobically hydrogen bonded to the same

extent in solution as it is in a crystal or otherwise limited in its ability to bend. It is not
clear from this observation per-se that the CH2 group is associated to another sugar
molecule: it might be associated with water. For similar reasons it cannot be concluded
that the shift in the HOH bend frequency is evidence of solvent exclusive solute
clustering; it can be concluded that some of the water molecules behave in an ice-like
manner (are tetrahedrally coordinated) and the bending motion is hindered concentrated
solutions. It cannot be said conclusively whether this hindering is due to association
with the solvent or with the solute. In addition to this uncertainty of interpretation is the
fact that Raman spectra are solute sensitive. As the concentration increases beyond
saturation the changes in the spectra are less pronounced; Mathlouthi's observations do

not extend into the supersaturated region.

2.4.1.2 Fourier Transform Infrared Smtroscopy

Evidence of solvent exclusive-solute clustering, if it occurs, could be observed
in IR spectra (114). The unequivocal interpretation of IR and Raman spectra is severely
hindered by the absence of accepted protocol for quantifying the effects of hydrogen
bonding on the spectra, by the inability to account for the effects of hydrogen bonding
in LCAO-MO or other calculations by means of which band assignments are

accomplished (since these calculations are for molecules in-vacuo), by the general lack

35

of resolution observed in all regions of the spectrum associated with water and its
vibrations, and by the lack of a clear understanding of the spectrum of liquid water.
This is more so true for IR. Suggested methods for obtaining good quality spectra
include the arbitrary subtraction of weighted solvent spectra; while this may be good
enough for qualitative analysis it is unsatisfactory for interpreting structure and
interactions. Although the there is evidence that suggests solute-solute interactions in
supersaturated solutions, there is also overwhelming evidence that all bands
corresponding to vibrational modes sensitive to hydrogen bonding are intensified. Thus

the possibility of solvent inclusive association cannot be ruled out.

2.4.1.3 X-rgyDiffraction

Mathlouthi's X-ray diffraction studies (81) of sugar solutions indicate formation
of intrarnolecular bonds with the increase in solute concentration. He was able to
convincingly demonstrate that some sucrose molecules might be associating with some
other sucrose molecules, a result consistent with the observations of band shifting and
intensification in IR and Raman spectra. He was, however, unable to provide any
convincing data or argument that suggests that such clusters are larger than two solute

molecules or that they were exclusive of associated solvent.

2.4.1.4 NMR

Franks, Suggett and co-workers (87-93) used dielectric relaxation and NMR to
determine solvation properties of sugars. Franks et al.(87, 88) confirmed the earlier
results of Jones (75) and Phillips (76) for dilute solutions (an average of 6 water
molecules of solvation per molecule of sucrose) and also confirmed that the sugar water
interaction was about as strong as the water-water interaction. Harvey and Symons (82)

used NMR to determine 11 water molecules of solvation per molecule of sucrose,

36

interpreted as 6 molecules of water strongly associated with the molecule of sucrose
and the remaining molecules more dynamically associated. Birch et al.(94), in a pulsed
NMR study of concentrated sucrose solutions, determined that there are four kinds of
protons in solution: bulk water, solvated water, exchangeable sugar hydroxyl, and non-
exchangeable 1ing(C-H) protons. Franks, Suggett and co-workers as well as Barone
and co-workers (who used numerical simulations based on Kozak et al.'s (60)
modification of McMillan Meyer (96) theory) initially hypothesized that solute-solute
interactions in dilute solutions were unlikely. More recent work by Barone et al.(6l)
indicates that the nature of interactions in solution may be more complex than
previously accepted without offering insight into the exact nature of the complexity.
The 170 NMR work of Richardson and Steinberg (83) indicates that sucrose solutions
can be modelled using an isotropic two states of water exchange model, with no
discontinuity in exchange rates related to saturation. The NMR work is sensitive to the
time scales on which the exchange of proton takes place between different proton
donor-acceptors. All of this work indicates that the solute molecules are at least partially
solvated even in supersaturated solutions and that there is water, even in supersaturated

solutions, that behaves like bulk water.

2.5 References

1. Jancic SJ. and E.J. de Jong eds, Industrial Crystallization '84, Elsevier, Amsterdam,
(1984)

2. Levine I. N., Physical Chemist_ry, 2nd edition, McGraw Hill Inc., New York, NY,
(1983)

3. Randolf A. D. and M. A. Larson, Theogy of Particulate Processes, Academic Press,
New York, NY (1971)

4. Randolf A. D. , Can. J. Chem. Engg., 42, 280, (1964).

5. Larson M.A. and A. D. Randolf , Chem. Engg. Progress Symp. Ser., 65(95), 1,
(1969).

6. Mullin, J.W., CM’zation, Butterworths, London, pp. 35-52, (1971).

7. Kirk , and Othmer eds, Encyclopedia of Chemical Technology, 3rd ed., (1981).

8. Brice J. C., J. Crystal Growth, 1, 218, (1967)

9. Bunn, C. W., Disc. Faraday Soc, 5, 32, (1949)

10. Berglund K A., PhD Dissertation, Iowa State University, 1981.

37

11. Shiau L. D., PhD. Dissertation, Michigan State University, 1988.

12. Mason R. E. A. and R. F. Strickland-Constable, Trans. Faraday Soc., 62, 455,
(1966)

13. Botsaris G. D., in J.W. Mullin ed. Industrial Crystallization, Plenum Press, New
York, NY, 3-32, (1976)

14. Lal D. P., R. E A. Mason, and R. F. Strickland Constable, J. Crystal Growth, 5, 1,
(1969)

15. Garside J. and R. J. Davey, Chem. Engg. Com, 4, 393-424, (1980)

16. a Mullin, J.W. and C. Leci, Philos. Mag, 1 9, 1075 (1969).
b.. Mullin, J.W. and C. Leci, AIChE Symposium Series 68, J. Estrin ed., 8, (1972).

17. Clontz N. A. and W. L. Mc Cabe, Chem. Engg. Prog. Symp. Ser., 67(110), 6, 1971.

18. Strickland-Constable R.F., Kinetics and Mechanism of Crystallization, Academic

Press, London, (1968)

19. Powers HE. C., Nature, 17 8, 139-140, (1956)

20. Ben-Naim, A. , Solvat_ion Thermodynamics, Plenum Press, New York (1987).

21. Buckley H. E, Crystal Growth, Chapman and Hall, London, (1952)

22. Burton W.K, N. Cabrera, and F. C. Frank, Phil. Trans, A243, 299, (1951)

23. Frank F. C., Disc. Faraday Soc., 5, 49, (1949)

24. Larson, M.A. and J. Garside, Chem. Eng. Science, 4 1, 1285 (1986).

25. Larson, M.A. and J. Garside, J. Crystal Growth, 7 6, 88 (1986).

26. Larson, M.A., AICHE Symposium Series no.240, 80, 39—44 (1984).

27.Gilmer G.H. , R. Ghez, and N. Caberera, J. Crystal Growth, 8, 79, (1971)

28. Chemov A. A., Soviet Phys. Usp., 4, 116, (1951)

29. Bennema P., in J.W. Mullin ed. Industrial ngstallization, Plenum Press, New York,
NY, (1976)

30. Shiau L. D. and K. A. Berglund, AIChE. J., 33(6), 1028, (1987).

31. Shri Y. , RW. Hartel, and B. Liang, J. Dairy Science, 72(11), 2906-2915, (1989)

32. Cerreta, M.K and K. A. Berglund, J. Crystal Growth, 102, 869, (1990)

33. Addadi L., Z. Berkovitch-Yellin, N. Domb, E. Gati, M. Lahav, and L. Leiserowitz, Nature,
296. 21, (1982).

34. Addadi L., Z. Berkovitch-Yellin, I. Weissbuch, M. Lahav, and L. Leiserowitz, Mol. Cryst.
Liq. Cryst, 96, 1, (1983).

35.Addadi L., and S. Weiner, Proc. Natl. Acad. Sci, USA, 82, 4110, (1985).

36. Addadi L., Z. Berkovitch-Yellin, I. Weissbuch, J. van Mil, L. J.W. Shimon, M. Lahav, and
L. Leiserowitz, Angew. Chem. Intl. Ed. Engl., 24, 466, (1985).

37. Addadi L., and S. Weiner, Mol. Cryst. Liq. Cryst, 134, 305, (1986).

38. Berkovitch-Yellin, Z., L. Addadi, M. Idelson, L. Leiserowitz, and M. Lahav, Nature, 296,
27, (1982).

39. Berkovitch-Yellin, Z., L. Addadi, M. Idelson, M. Lahav, and L. Leiserowitz, Angew. Chem.
Intl. Ed. Engl., 2 1(8), 631, (1982).

40. Landau, E M., M. Levanon, L. Leiserowitz, M. Lahav, and J. Sagiv, Nature, 3 18, 353,
(1985).

41. Landau, E.M., R. Popovitz-Biro, M. Levanon, L. Leiserowitz, M. Lahav, and J. Sagiv, Mol.
Cryst. Liq. Cryst, 134, 323, (1986).

42. Weissbuch I., L. Addadi, Z. Berkovitch-Yellin, E. Gati, M. Lahav, and L. Leiserowitz,
Nature, 310, 161, (1984).

43. Weissbuch 1., L. J. W. Shimon, E.M. Landau, R. Popvitz-Biro, Z. Berkovitch-Yellin, L.
Addadi, M. Lahav, and L. Leiserowitz, Pure and Appl. Chem, 58(6), 947, (1986).

44. Dunning W.J. and N. Albon, in Growth and Perfection of Crystals, ed. R.H. Doremus,
RW. Roberts, and D. Tumbull, Wiley, London, 416, (1958).

45. Smythe B. M., Aust. J. Chem, 20, 1115, (1967).

46. Smythe B. M., Sugar Technol. Rev., 1, 191, (1971).

479g5mman G., States of Agggeggtion, trans]. R.F. Mehl, van-Nostrand, New York,

1 .

 

48.
49.
50.

51.
52.

53.

$633943

84.

38

Eankovan, P. and K.A. Berglund, AIChE J., 33 .1844, (1987).

Elankovan, P., and KA. Berglund, Appl. Spectroscopy, 40 , 712, (1986).

Cerreta, MK. and KA. Berglund, in Industrial ngstallization '84, S.J.Jancic and E.J.
DeJong, (Eds), Elsevier, Amsterdam, 21 (1984).

Cerreta, M. K. and K A. Berglund, Journal of Crystal Growth , 84, 577 (1987).
Smith, J. M. and H. C. Van Ness, Introduction to Chemical Engineering Thermodynamics,
3rd ed., McGraw Hill Inc., Singapore, (1975)

Labuza T. P. 1980. The effect of water activity on rection kinetics of food
deterioration. Food T echnol. 3 4(4):36

Rockland L. B. and Nishi, S. K. 1980. Influence of water activity on food product
quality and stability. Food Technol. 3 4(4):42

. Duckworth, R B. 1975. Water relations ofFoods Academic Press, New York
. Rockland, L. B. and Stewart, G. F. 1981. Water Activity: Influences on Food Quality.

Academic Press, New York

Simato D. and Multon, J. L. 1985. Properties of Water ip Foods. Martinus Nijhoff
Publishers, Dordecht, The Netherlands

Le Maguer M. in EV_aterAc1_iviW: Theory and Applications to Foods, L. B. Rockland
and L. R. Beauchat eds, Marcel Dekker, Inc. , New York and Basel, pp. 1-25, (1987)

. Johnston MR. and RC. Lin, in Water Activigy : Theogy and Applications to Foods, L.

B. Rockland and L. R. Beauchat eds, Marcel Dekker, Inc. , New York and Base], pp.
287-294, (1987)

. Kozak, J.J., W.S. Knight and W. Kauzmann J. Chem. Phys, 48, 2, 675 (1968).
. Barone, G., G. Castronuovo, P. Del Vecchio and V. Elia, J. Solution Chem, 1 9, 1, 41

(1990).

. Barone, G., G. Castronuovo, P. Del Vecchio, V. Elia and M.T. Tosta, J. Solution Chem,

1 7, 10, 925 (1988)

. Barone, G., G. Castronuovo, V. Elia and M.T. Tosta, J. Solution Chem, 1 5, 2, 199 (1986).
. Barone, G., G. Castronuovo, V. Elia and V. Savino, J. Solution Chem, 1 3, 3, 209 (1984).

. Myerson, AS. and LS. Sorrel, AICHE Journal, 28,778 (1982).

. Myerson, AS. and Y.C. Chang, in Industrial ngstallization '84, SJ. Jancic and E.J. DeJong

(Eds), Elsevier, Amsterdam, 27 (1984).

. Chang, Y.C. and AS. Myerson, AIChE J., 3 l, 980 (1985).

. Chang, Y.C. and AS. Myerson, AIChE J., 32, 1567 (1986).

. Chang, Y.C. and AS. Myerson, AIChE J., 32, l747,(1986).

. Myerson A.S.,

. Cerreta, M.K, Ph.D thesis, Michigan State University (1988).

. Hussman, G.A., M.A. Larson and K.A. Berglund, in Industrfil Crystallization '84 SJ.

 

Jancic and E.J. DeJong (Eds), Elsevier, Amsterdam, 21 (1984).

. MacMahon, P.M., K.A. Berglund and M.A. Larson in Industrial ngstallization '84, SJ.

Jancic and E.J. DeJong (Eds), Elsevier, Amsterdam, 229 (1984).

. Bousfield, W.R and TM. Lowry, Trans. Faraday Soc., 6, 85 (1910).

. Jones, H.C., Amer. Chem. Journal, 32,319 (1904).

. Phillips, J.C., Trans. Faraday Soc., 3, 136 (1907).

. Bemal, JD. and RH. Fowler, J. Chem. Phys, 1, 515 (1933).

. Pauling, L. and O. Brockway, Proc. Natl. Acad. Sci, 20, 336 (1934).

. Franks, F. in Hydrogen Bonded Solvept System_s_, A.K. Covington and P.Jones (Eds),

Taylor and Francis Ltd, London, 31-47 (1968).

. Suggett, A., J. Solution Chem, 5, 1,33 (1976).

. Mathlouthi M., Carbohydrate Res, 9 1, 113, (1981).

. Harvey, J.M. and M.C.R. Symons, J. Solution Chem, 7(8), 571 (1978).

. Richardson SJ. and MP. Steinberg, in Eater Activity: Theory and Applications to

Food, L. B. Rockland and LR Beauchat eds, Marcel Dekker Inc., New York, 235,
(1987).
Frank, HS. and W.Y. Wen, Disc. Faraday Soc., 24, 1333 (1957).

39

85. Hildebrand, J H in Solubility, Wiley and Sons, N.Y., 2nd. ed., (1965).

86. Nemethy, G. and HA. Scheraga, J. Chem. Phys, 36, 3401 (1962).

87. Franks, F., J.R. Ravenhill and DS. Reid J. Solution Chem,l,1, 3 (1972 ).

88. Franks, F., M. Pedley and DS. Reid, J. Chem. Soc. Faraday Trans. 1, 7 2, 359 (1976).

89. Franks, F., Pure and Appl. Chem, 59, 9,1189 (1987).

90. Tait, M.J., A.Suggett, F. Franks, S. Ablett and RA. Quickenden J. Solution Chem, 1 , 2,

131 (1972 ).

91. Suggett, A. and AH. Clark, J. Solution Chem, 5, 1, 1 (1976).

92. Suggett, A., J. Solution Chem, 5, 1,33 (1976).

93. Suggett, A., S. Ablett and P.J. Lillford, J. Solution Chem, 5(1), 17 (1976 ).

94. Birch, G.G., J. Grigor and W. Derbyshire, J. Solution Chem, 1 8, 8, 795 (1989).

95. Narayanan, H. and GR. Youngquist, AICHE Symposium Series no. 253, 83,1-7 (1987).

96. McMillan, WC. and 1.5. Mayer, J. Chem. Phys, 1 3(7). 276 (1945).

97. Migus A., Y. Gauduel, J.L. Martin, and A. Antonetti, Phys. Rev. Lett. , 58(17),

1559, (1987).

98. aMaroncelli M., J. MacInnis, and GR. Fleming, Science, 243, 1674, (1989).
b.Chapman C. F., R. S. Fee, and M. Maroncelli, J. Phys. Chem, 94, 4929, (1990).

99. Kenny Wallace GA. and CD. Jonah, Chem. Phys. Lett., 3 9(3), 596, (1976).

100. Kenny-Wallace G.A., Can. J. Chem, 55, 2009, (1977).

101. Kenny-Wallace G.A., GE. Hall, L.A. Hunt, and K. Sarantidis, J.Phys Chem, 84,

1145, (1980).

102. Kenney-Wallaoe GA. and CD. Jonah, J. Phys. Chem, 86, 2572, (1982).

103. Wolynes PG, Annu. Rev. Phys. Chem, 3 1, 345, (1980).

104. Wolynes PG, J. Chem. Phys, 86(9), 5133, (1987).

105. Cross, A. J. and JD. Simon, J. Chem. Phys, 86, 7079, (1987).

106. Simon, J.D., Acc. Chem. Res, 2 l, 128, (1988).

107. Migus A., Y. Gauduel, J.L. Martin, and A. Antonetti, Phys. Rev. Lett. , 58(17), 1559,
(1987).

108. Brady J.W., J. Am Chem. Soc., 1 1, 5155, (1989).

109. Mathlouthi M and D.V. Luu, Carbohydrate Res, 7 8, 225, (1980).

110. Mathlouthi M. and D.V. Luu, Carbohydrate Res, 8 l, 203, (1980).

111. Mathlouthi, M., C. Luu, M. Meffrov-Biget and D.V. Luu, Carbohydrate Res, 8 l, 213

(1980).

112. Cael J.J., J.L. Koenig and J. Blackwell, Carbohydrate Res, 32, 79, (1974).

113. Szarek W.A., S. Korppi-Tommola, H.F. Shurvell, V. H. Smith Jr., and OR Martin, Can.
J. Chem, 62, 1512, (1984).

114. Back D.M., D.F. Michalska, and P. L. Polavarapu, Appl. Spectroscopy, 38(2), 173,

(1984).

Chapter Three

The use of pyranine as a trace fluorescent probe to study structure in

aqueous sucrose solutions"

3.1 Background

The effect of supersaturation on crystal growth from solution was established by
Ostwald in 1897, and the need to establish a relation between solubility and solution
structure was recognized in 1910 by Bousfreld and Lowry (10); but, the need to relate the
properties of saturated and supersaturated solutions to structure and to thereby characterize
the effect of solution structure upon crystal growth is fairly recent. In 1910 Bousfreld and
Lowry (10) accepted that water consisted mainly of dimers and trimers in the liquid state
and that structure in aqueous solution was a primary consequence of ordering of water
molecules. Jones (29) in 1904 showed that a molecule of sucrose is associated with 6
molecules of water on the average (values varied between 3 and 10 ), by measuring
freezing point depression in aqueous sucrose solutions of up to 47 wt.%. Phillips (49) in
1907 obtained similar results (with a similar range of values) using hydrogen gas
absorption in sucrose solutions. The mechanism of solvation was investigated by Eley
(19), who proposed a 2 step solvation model which correctly predicted the relationship
between temperature and heat capacity of solvation, but could not account for observed
water of solvation of various compounds. Pauling (48) proposed a tetrahedral model for

liquid water

 

' This material was published in part as R Chakraborty & K A. Berglund, "The Use of Pyranine as a Trace
Fluorescent Probe to Study Structure in Aqueous Sucrose Solutions,“ AIChE Symposium Series, no. 284,
p 113-124, (1991).

40

41

which could explain the solvation behavior of some ions, but could not account for
the conductivity of water or rapid rates of proton transfer.

Meanwhile Bemal and Fowler (8) first used Raman spectra of water and Hg
fluorescence of sugar solutions to speculate the relative strength of sugar-water
interactions, concluding that they were of the same order of magnitude as the water-
water interaction. Hildebrand (27) successfully explained solvation using a lattice
model of solution in which he considered the interaction of a molecule with its six
nearest neighbors. Thus, it was recognized that solution structure is the result of
different molecular interactions ie. solute-solvent, solvent-solvent, and solute-solute.
Solution structure was hence no longer understood to be an exclusive property of
aqueous solutions.

The structure of liquid water nevertheless continued to be a subject of interest
because a correct determination of the structure in liquid water defines the nature of
solvent-solvent interactions in aqueous solution. Frank and Wen (21) proposed a
flickering cluster model for water which allows dynamic interchange of single water
molecules between clusters of 2 to 7 water molecules. Nemethy and Scheraga (45)
authored another theory which assumes formation of clusters of a similar size with
the absence of dynamic interchange at the molecular level. Both these views are
widely accepted today. Once the nature of the solvent- solvent interaction was widely
accepted the attention focused on determining the nature of solvent-solute and solute-
solute interactions. The nature of these interactions is particularly complex in
concentrated solutions; therefore, most of the work reported involves techniques
developed for the study of undersaturated solutions.

Recent experimental attempts to verify concepts of solution structure have
involved the use of Raman, IR, and NMR spectroscopies. The possibility that
molecular conformation in solution could affect solution structure was recognized as

an important issue in the early 70's. Franks, Suggett and co-workers

ml

in

is r
51K
All

in:

Ida

SUR

42

(22,23 ,24,25,53 ,54,55,56) used dielectric relaxation and NMR to determine
solvation properties of sugars. Franks et al. confirmed the earlier results of Jones and
Phillips for dilute solutions and also confirmed that the sugar water interaction was
about as strong as the water-water interaction. Harvey and Symons (26) used N MR
to determine 11 water molecules of solvation per molecule of sucrose, interpreted as 6
molecules of water strongly associated with the molecule of sucrose and the
remaining molecules more dynamically associated. Birch et al. (9), in the only NMR
study of concentrated sucrose solutions, determined that there are four kinds of
protons in solution: bulk water, solvated water, exchangeable sugar hydroxyl, and
non-exchangeable ring (C-H) protons, using pulsed NMR.

Franks, Suggett and co—workers (22-2553-56) as well as Barone and co-
workers (4-6) (who used numerical simulations based on Kozak et al.'s (33)
modification of McMillan Meyer (39) theory) initially hypothesized that solute-solute
interactions in dilute solutions were unlikely. More recent work by Barone et al. (3)
indicates that the nature of interactions in solution may be more complex than
previously accepted without offering insight into the exact nature of the complexity.

Mathlouthi et al. (38) attempted to use vibrational spectroscopy of
concentrated solutions to correlate the change in solution structure to change in
concentration. It was demonstrated that vibrational spectroscopy (ie. Ranmn and IR)
is effective in solutions of low solute concentrations. Back et al. (1) used FI'IR to
successfully study the mutarotation behavior of glucose in dilute aqueous solutions.
Although vibrational spectroscopy yields significant information about solute solvent
interactions, in practice this information is very difficult to extract.

Larson (34) explicitly stated the need to understand and characterize the
relationship between the properties of supersaturated solutions (degree of
supersaturation) and concentrated solution structure.Work on cluster diffusion in

liquids (Cussler, 18) was cited as evidence that properties of non-ideal solutions

43

could deviate greatly from those of ideal and dilute solutions. Subsequently Myerson
and co-workers (l4-16,42,43) performed diffusion studies in concentrated aqueous
solutions of both non-electrolytes (urea, glycine, and glycine-valine) and electrolytes
(sodium chloride and potassium chloride) that indicated formation of solute-solvent
aggregates of an unknown composition as evidenced by precipitous drops in
diffusion coefficients. Cerreta and Berglund (12,13) used Raman spectroscopy of
ammonium dihydrogen phosphate (ADP) to show that while no great differences
could be observed in the spectra of very concentrated, saturated and supersaturated
solutions, evidence did exist for dimer formation. Larson and Garside (35) cited
Cerreta and Berglund (13) as well as the Raman studies of Hussman et al. (28) and
MacMahon et al. (37) with alkali nitrates as evidence of pre—crystalline clusters in
concentrated solutions. Larson and Garside (35,36) performed an experiment with
isothermal columns of concentrated, saturated, and supersaturated solutions of citric
acid, urea, sodium nitrate, and potassium sulfate during which they observed
concentration gradients along the length of the column for only the supersaturated
solutions. Larson and Garside concluded that there must be solute structuring in
supersaturated solutions. In other work, Narayanan and Youngquist (44) attempted to
relate the onset of saturation with a marked change in measurable physical properties
of bulk solution in solutions of sodium nitrate, potassium nitrate, magnesium sulfate,
potassium ammonium sulfate, and ADP. No unusual changes in density, viscosity,
or electrical conductivity with saturation and subsequent supersaturation were
observed.

In each of the previous studies (with the exception of the NMR study of
concentrated solutions) the approach has been to relate the behavior of the solute or
the state of the solute to a marked change in an observable bulk solution property at or
near saturation, then to correlate the changes in this property to degree of

supersaturation. The results of these experiments have been less than definitive;

44

therefore, we elected to develop a complementary approach which probes the solvent
rather than the solute.

3.2 Photochemical Techniques using Trace Fluorescent Probes

The unirnolecular photochemical reaction of electronically excited molecules in
homogeneous media can be used to study relaxation processes on a variety of time
scales. Fluorescence is the consequence of fast radiative transition to the ground state
in a molecule that has been excited by radiation to an allowed electronic state of higher
energy. Fluorescence techniques have high quantum efficiencies (0.7-0.9) but are
restricted in application to fast kinetic processes.

It is important to distinguish between the photochemical investigation of a
fluorescent compound and the experiment utilizing a trace fluorescent probe. The
experiment involving a trace fluorescent probe differs from the photochemical
investigation of a fluorescent compound in several important respects. The former
makes use of a compound for which the photochemical behavior is well known; the
photochemistry of the compound is the subject of the latter. Probe experiments
require very low molarities of the fluorescent compound; photochemical
investigations require high molarities. Changes in the emission of the molecule can be
correlated to corresponding changes in solvent properties in experiments using
fluorescent probes. In a photochemical investigation changes in emission of the
fluorescing species give information about its electronic state.

Often photochemical investigations are experimentally difficult because the
high molarity of the fluorescent species gives rise to a host of associated phenomena
which may also affect its excited state behavior. Fluorescent probes on the other hand
are usually present in sufficiently low molarities to preclude any effect other than that

due to the solvent environment of the probe (all other experimental renditions being

45

constant). In a homogeneous solvent or neat solution there is little possibility of non—
uniform probe distribution. Additionally, judicious choice of a probe that has an
affinity for a certain type of site on a molecule causes the nature of the
microenvironment of the probe to be reflected in its emission properties. Often the
sensitivity of a probe to solvent environment is enhanced in its excited state.
Fluorescent compounds at very low molarities are thereby sensitive molecular probes
of their solvent environment. Where the number of molecules that interact with the
probe (hosts) is finite, a statistical analysis of the hosts is a useful approach to
distinguish between various possible structural models. Information about the change
in emission properties of the fluorescent probe can therefore be correlated with
corresponding changes in solution properties to provide information about structure
in solution provided the number of hosts is finite.

For this study, we required a probe that is known to have a highly sensitive,
polarity dependent equilibrium between two excited states (both of which fluoresce)
with the same chromophore differing only slightly in structure. We chose pyranine or
8-hydroxy-1,3,6-pyrenetrisulfonate, Fig. 3.1. Its photochemistry has been well
characterized (Kondo et al. , 32).

 

Figure 3.1. Chemical structure of pyranine or 8-hydroxy-1, 3, 6-pyrenetrisulfonate.

In solution pyranine exists with the sulfonate groups entirely dissociated. It
has two excited fluorescent states: one where the hydroxyl proton is associated with
the molecule which emits at about 440 nm and the other where the hydroxyl proton is
dissociated from the molecule which emits at 511 nm, Fig. 3.2. Thus, the two excited
states differ only in the hydroxyl proton. The pyrene ring acts as the chromophore in
both cases. Pyranine in its excited state is sensitive to the presence or absence of
exchangeable protons in the vicinity of the hydroxyl proton (i.e. polarity or pKa).
Incident exciting radiation causes a Jt-Jt* singlet state transition in the probe molecule.
This energy is then released mostly through a fast kinetic radiative path on the order

of a 100 picoseconds with a quantum efficiency of nearly 80%.

440-450 nm 511-514 nmitem
work 1120‘ Pyo‘ ' + 1130“

 

hex:
342nm

Pyon+rr20 —" Py0+1130+
350nm 405nm Mb

Figure 3.2. Emission equilibrium of pyranine. The equilibrium between the
protonated form (absorbs at 350 nm and emits at 440 nm when excited at 342 nm)
and the deprotonated form (absorbs at 405 nm and emits at 511 nm when excited at
342 nm) is determined by ability of the microenvironment of the hydroxyl group to

participate in proton transfer.

The deprotonated state requires less energy of excitation and thus emits to the red of

the protonated state. Whether the observed emission is from the excited deprotonated

 

47

or protonated form is determined by the molecular environment in its immediate
vicinity. We can correlate changes in the fluorescence equilibrium between the two
emitting forms of the probe to changes in the nature of the rrricroenvironment of the
hydroxyl moiety of the probe. Once we establish the nature of the rrricroenvironment
in terms of the species in solution we can then relate changes in the fluorescence
equilibrium to changes in solution structure.

Pyranine has been used to probe water content of micellar dispersions
(Clement and Gould, (17); Kano and Fendler, (30); Kondo et al., (32); Bardez et al.,
(2); and, Pines and Huppert, (47) ), sols (Kaufman et al., (31) ) and sol-gels
(Pouxviel et.al., (50) ) among other systems. It has been used to relate water content
of these systems to the structure of various phases. It is photochemically active at
concentrations of 1 ppm (1.9 x 10‘6 M) or less. In addition, it is non-toxic, non-
carcinogenic, and suitable for use in food grade applications (licensed for use as a
food color additive, USDA regulations 1988).

The approach in this study is unique in two respects. First of all we decided to
use a solvent sensitive technique to study concentrated and supersaturated solutions.
We used pyranine as a trace fluorescent probe (very sensitive to solvent environment)
and related changes in its state to changes in solvent environment. Secondly, we
related changes in the state of the solvent to changes in solution structure. Our choice
of bulk solution properties also differs from those chosen previously, by correlating

changes in solvated and non-solvated (bulk) water with solute concentration, and in

particular the degree of supersaturation (DS).

 

3.3 Materials and Methods:

Emission spectra were obtained with either a 200 W Xe/Hg or a 150 W Xe
source in quartz cuvettes using an in-house constructed spectrometer described in
detail elsewhere (Mussel and Nocera, (41)). The lamps were operated at 80% rated
power. A Harnamatsu R1104 photomultiplier tube (maintained typically at 50%
maximum voltage) was used in conjunction with a preamplifier (maintained at a
sensitivity of 10 '6) and a Princeton Electronics lock-in amplifier (maintained at a
sensitivity of 3V). The photomultiplier tube was maintained at -60 degrees C to
improve the signal to noise ratio. A 375 nm KV or a 370 nm GG cutoff filter was
used to filter out source excitation at 342 nm. A 3 mm slit width was used and
shutters were kept fully open.

Reagent grade sucrose from Sigma Chemicals, 8-hydroxy 1,3,6,
pyrenetrisulfonate (pyranine) from Eastman Kodak, and 18 Mohm reverse osmosis
water were used. Stock solutions of 100 ppm (1.9 x 10‘4 M) pyranine were prepared
using 18 Mohm reverse osmosis water. Solutions of the 10 ppm (1.9 x 10'5 M) and
1 ppm (1.9 x 10‘6 M) were prepared by dilution. The amount of sugar and water
required for 10 grams (.01 kg) of each solution was calculated beforehand. Reagent
grade sucrose samples were carefully measured out to +/- 1% of the required amount
Samples were prepared in 20 ml glass vials by adding the predetermined amount of
stock solution to achieve the desired concentration. Samples varied in concentration
between pure water and 80 wt.% sucrose. In undersaturated solutions samples were
prepared in 10 wt.% intervals. Samples near saturation were taken at 2 wt.% intervals
between 60 and 70 wt%, and 76 wt.% and 80 wt.% solutions were also prepared.
Supersaturated solutions were prepared by heating the vials in a water bath until the

sugar dissolved. Spectra of the solutions were taken within 6 to 24 hours of

 

49

preparation to ensure equilibration of the sugar in solution and prevent biological
contamination of the sugar solutions.

Pyranine doped sugar solutions - with pyranine concentrations of 1 ppm (1.9
x 10‘6 M), 10 ppm (1.9 x 10'5 M), and 100 ppm (1.9 x 10‘4 M)- showed little
variation in peak intensity ratio (PIR). The PIR is defined as the ratio of the peak
intensity at 440 nm to that at 511 nm. This is the same as the F ratio described in
Pouxviel et.al. (50). It was decided to work at 100 ppm ( 1.9 x 10'4 M) levels of
pyranine to maximize the signal to noise ratio. Small changes in instrument settings
did not affect the values obtained for the PIR.

There was less than 3% variation in PIR of pyranine in undersaturated
sucrose solutions over a 30 degree temperature span from 20 to 50 degrees C . This
indicates that the equilibrium constant between the two excited states is not a strong
function of temperature in this range of operation. Therefore temperature corrections
were not applied.

Small variations in excitation wavelength (+/- 5 nm) did not change the PIR
(less than 3%); however, large changes in the excitation wavelength (335-365 nm)
indicated that the PIR were up to 5% less at longer wavelengths than at shorter ones.
For a solution of given sucrose concentration doped with a fixed amount of pyranine
the change in intensity of the peak at 511 nm with increase in wavelength is grcater
than that associated with the peak at 440 nm. This accounts for the difference in PIR
due to large changes in excitation wavelength. In order to get consistent results an
excitation wavelength of 342 nm was used for all experiments.

Emission spectra were obtained for each of the sample solutions prepared.
The PMT response was aboutthesameat440nmand 511 nmsothe spectrawere not

corrected for PMT response.

3.4 Results and Analysis

The emission spectra for pyranine d0ped sucrose solutions of 0, 20.40.60.
and 70 wt.% are shown in Fig. 3.3. It is evident that the intensity of both the blue
(440 nm) and the green (511 nm) peaks change with the amount of sugar and water
present in solution. The ratio of the peak intensity at 440 nm to that at 511 nm was
defined previously as the peak intensity ratio (PIR).The PIR increases as the
concentration of sugar increases and the concentration of water decreases. A clear
isosbestic point is observed at 485 nm which also indicates the presence of two

emitting species in equilibrium.

 

lntenslty
5 § 5 r (a 5

§

 

 

 

Wavelength

Figure 3.3. The emission spectra of pure water and sugar solutions of 20. 40, 60.

and 70 wt.% sucrose doped with 100 ppm (1.9 x 10"4 M) pyranine at 20 degrees C.
Emission maxima occur at 440 nm and 511 nm An isosbestic point is observed.
represented by the pointer at 485 nm.

   

51

There are two possible emitting forms of pyranine. The form that contains the
hydroxyl proton can exist only in an environment that does not facilitate the transfer
of protons. The other form exists where it can transfer protons. The fact that emission
from both forms is observed in solution leads us to conclude that microenvironments
suitable to each species must co-exist in solution. This argument is along the same
lines as that put forward by Zinsli (57) which led to a biphasic model for micelles.

In the current context microenvironments refer to the presence or absence of
exchangeable protons in the immediate proximity of the pyranine hydroxyl proton.
Emission observed at 440 nm indicates that there are protonated pyranine molecules
in solution. These can exist because there are no protons available for exchange in the
immediate vicinity of the pyranine hydroxyl proton. Similarly. emission observed at
511 nm indicates that there are deprotonated pyranine molecules in solution. So the
PIR, according to this model. is a measure of the number of pyranine molecules in
the protonated state versus the number in the deprotonated state. It is reasonable to
assume that the accessibility of a molecule of pyranine to either environment is the
same, because pyranine is present in very low molarity (of the order 10‘7 or less) in
solution. We now focus on what species or combination of species in solution could
give rise to the observed behavior of the PIR

Birch etal. (9) observed three kinds of exchangeable protons in solutions of
sucrose: bulk water. salvation water, and sucrose hydroxyl. Of these, the exchange
rate of the sugar hydroxyls is somewhat slower than that of the solvation water. For
reasons of mobility, it is unlikely that the sucrose molecules would be able to compete
with water molecules in interacting with pyranine molecules to a significant extent.
This leads us to propose that the pyranine is likely to interact only with water
molecules.

A water molecule in contact with other water molecules that behave like pure

water could provide an environment that is capable of proton exchange. If the

 

52

pyranine hydroxyl is in the proximity of such a water molecule it would emit at 511
nm. Thus, we propose that one of the two microenvironments is likely to be a bulk-
water-like microphase. A pyranine hydroxyl group existing near solvation water
already associated with a molecule of sucrose could not exchange protons and would
emit at 440 nm Thus the other likely rrricroenvironment could be due to the presence
of an envelope of solvation.

A graph of the PIR versus concentration of sucrose (wt.%) is shown in Fig.
3.4. Since pyranine has only one proton available for exchange. it is reasonable to
conclude that only one molecule of water at a time can participate in the exchange
process. It follows that the PIR (a ratio of pyranine molecules incapable of
exchanging protons to that of pyranine molecules capable of exchanging protons) is
also the ratio of water molecules that do not participate in proton exchange with the
pyranine to those that do. In addition we can scale this per molecule of sucrose. This
is purely for our convenience. because it then becomes possible to look at the solution
in terms of water of solvation per molecule of sucrose (unavailable for proton
exchange) and bulk water (available for proton exchange). These are terms that can be
compared with existing data, at least for undersaturated solutions.

Assuming that the quantum yield of the two emitting species is the same
(same chromophore). we can plot a graph of the number of molecules of water in
each of these nricroenvironments per molecule of sucrose. We calculate 6 molecules
of solvation water per molecule of sucrose in undersaturated solutions. dropping
(insignificantly perhaps) to 5.8 molecules at the saturation point. 5.5 at 70 wt. %, 5.2
at 76 wt.%. and 4.0 at 80 wt%; see Tables 3.1 and 3.2. It would seem reasonable
that solvation water per molecule would be a function of the molecular structure.
more specifically the number of sites (and conformation of these sites) on the
molecule that could hydrogen bond with nearby water. Perhaps more significantly the

number of molecules of water in the bulk water like environment drops from 165 for

 

53

a 10% sucrose solution to about 3 for a highly supersaturated solution of 70%
sucrose. It drops further to a value slightly less than 1.0 (0.&) for the 76% solution

and stays about constant for the 80% solution.

 

 

 

 

 

 

 

Wt % sucrose

Figure 3.4. Graph of PIR (peak intensity ratio) vs. wt% sucrose in solution at 20

degrees C using 100 ppm (1.9 x 10'4 M) pyranine. The PIR is defined as the ratio of
peak intensities at 440 nm and 511 nm. The arrow represents concentration at
saturation.

Although the determination of water of solvation is in agreement with the
results of Jones (29). Phillips (49). and Franks, Suggett and co-workers (22-25. 53-
56 ), we were not entirely comfortable with the assumption of equal quantum yields
for the two emitting species. In an attempt to isolate the two species and estimate
relative quantum yield. spectra of equal amounts of carefully measured pyranine were
obtained in methanol and water (under the same conditions as the rest of the

experiments). The only peak observed in methanol was that at 430 nm. In water, a

 

54

weak shoulder was observed at 440 nm. but the emission was primarily due to the
peak at 511 nm. The peak in methanol was approximately 1.8 times as intense as the
peak in water for any given concentration of pyranine (between 1-80 ppm (1.9x 10‘8
and 1.6 x 10’5 M) ). This would indicate 11 molecules of solvation water per
molecule of sucrose. in agreement with the results of Harvey and Symons (26) and
the earlier calculations of Stokes and Robinson (52). It would also indicate negative
numbers for the bulk water like phase in the supersaturated region. which is
unphysical. It is worth noting that the spectra in methanol were blue-shifted about 5-
10 nm from those in water. This may be due to a difference in polarity of
rrricroenvironment or due to a green bleed through in the spectra in water. At this
point we feel it is not appropriate to conclusively state that both emitting species have
the same quantum efficiencies although there is some experimental evidence to
support this and this is what our arguments would lead us to believe.

In either case (6 or 11 molecules of solvation water per molecule of sucrose)
the surprising observation was that there should be any bulk phase water at all in
saturated and supersaturated solutions. The number of water molecules in the bulk
phase per molecule of sucrose decays exponentially with increase of sucrose
concentration. Near saturation the number of water molecules per molecule of sugar
in the bulk phase is roughly equal to the the number of molecules of water of
solvation. For a highly supersaturated solution of 70% sucrose there are roughly
three molecules of bulk phase water and six molecules of solvation water per
molecule of sugar. It is probably the water in the bulk phase that is responsible for
structurally ”stable" supersaturated solutions. The ease with which sucrose can be

dissolved in solution appears to be directly related to the amount of bulk phase water.

55

 

200

—9— Total Water
—°'— Bulk Water
W" Solvation Water

  
  
 
 

100‘

Moleculeparmolaculodm

 

UUUUUUU

 

 

. ' 1’ a . .
O 20 4O 60 80
Wt. % Sucrose

 

Figure 3.5. Total, bulk and solvation water (number of molecules of water per

molecule of sucrose) vs. wt% sucrose in aqueous solution with 100 ppm (1.9 x 10‘4
M) pyranine at 20 degrees C. The values on the graph are obtained from the analysis
presented in Tables 1 and 2. The vertical line represents saturation.

An interesting feature of this analysis is that the water of solvation per
molecule of sucrose becomes approximately equal to the bulk water per molecule of
sucrose at saturation. In supersaturated solutions there is always less bulk water
available per molecule of sucrose than solvation water. While this feature is apparent
from Fig. 3.5, it becomes more so in Fig. 3.6. The crossing point of the two curves
indicates equilibrium At this point it becomes energetically unfavorable to add more

solute to the solution without supplying external energy.

 

/_

     

 

—B-— Bulk Water
—O— Solvation Water
-—*-—- Total Water

 

 

 

 

T

-O.8 43.6 ' -o'.4 -o'.2 ' -o.o ' 0.2
Degree of Supersaturation

Figure 3.6. Graph of total water. bulk water and solvation water per molecule of

sucrose in aqueous solutions doped with 100 ppm (1.9 x 10‘4 M). pyranine at 20
degrees C versus the degree of supersaturation. DS. The DS is given by the
difference between the solution concentration and the saturation concentration at 20
degrees C divided by the saturation concentration at 20 degrees C.

3.5 Conclusions

It appears that the PIR obtained from fluorescence spectra of aqueous sucrose
solutions doped with trace amounts of pyranine show a marked change at saturation
and are very sensitive to water content of supersaturated solutions. This is a sensitive
technique for measuring degree of supersaturation.

Structurally. the solvent appears to distribute itself in two kinds of
microenvironments. One is associated with the salvation envelope of the solute

molecule. designated solvation water. and the other is non-solvated solvent.

 

57

designated bulk phase water. Both microenvironments exist in solutions of all
concentrations examined The amounts and ratio of water in each phase change with
concentration of sucrose in solution. The water of solvation per molecule of sucrose
seems to be a constant. The amount of water present in the bulk phase decreases
exponentially with increase in sucrose concentration and correlates with the ease of
solvation of sucrose. At the saturation point. the amount of water in the bulk phase
per molecule of sucrose is roughly the same as the number of water molecules of

solvation per molecule of sucrose.

3.6 Acknowledgements

The authors wish to thank Dr. D.G. Nocera, of the Department of Cherrristry,
Michigan State University for kindly permitting the use of the emission spectrometer
in his laboratory. Financial assistance from the Department of Chemical Engineering
and the Crop Bioprocessing Center/ Research Excellence Fund at Michigan State
University is greatly appreciated.

3.7 Literature Cited

1. Back, D.M., D.F. Michalska and PL. Polavarapu, Applied Spectroscopy. 38.
2. 173 (1984).
2. Bardez. E, B.T. Gouguillon, E. Keh and B. Valeur, J. Phys. Chem, 88.
1909. (1984).
Barone. G., G. Castronuovo, P. Del Vecchio and V. Bia, J. Solution Chem,
19. 1, 41 (1990).
Barone, G., G. Castronuovo, P. De] Vecchio. V. Elia and M.T. Tosto. J.
Solution Chem. 1 7. 10. 925 (1988)
Barone. G., G. Castronuovo, V. Elia and M.T. Tosta, J. Solution Chem, 1 5,
2. 199 (1986).
Barone, G.. G. Castronuovo, V. Elia and V. Savino, J. Solution Chem. 1 3.
3, 209 (1984).
Ben-Naim, A. . Solvation Thermodypamics, Plenum Press. New York (1987).
Bemal, 1D. and RH. Fowler, J. Chem. Phys, 1. 515 (1933).
Birch, G.G., J. Grigor and W. Derbyshire, J. Solution Chem, 1 8. 8. 795
(1989).
10. Bousfield. W.R and TM. Lowry, Trans. Faraday Soc., 6. 85 (1910).
ll. Cerreta. M.K. Ph.D thesis, Michigan State University (1988).
12. Cerreta. MK. and K.A. Berglund, Journal of Crystal Growth , 84. 577
(1987).
13. Cerreta. MK. and KA. Berglund, in Industrial Cgstallization '84,
S.J.Jancic and E.J. DeJong, (Eds). Elsevier, Amsterdam. 21 (1984).
14. Chang. Y.C. and AS. Myerson, AIChE J., 3 1. 980 (1985).
15. Chang. Y.C. and AS Myerson. AIChE J., 32. 1567 (1986).

sesame?)

29.
. Kano. K and J.H. Fendler, Biochim. Biophys. Acta, 509. 289 (1978)
31.

32.
33.

. Larson. M.A., AICHE Symposium Series no.240, 80. 39-44 (1984).
35.
. Larson. M.A. and J. Garside. J. Crystal Growth. 76. 88 (1986).
37.

58

. Chang. Y.C. and AS Myerson. AIChE J., 32, l747,(1986).

Clement, NR. and M. Gould. Biochemistry, 20. 1534 (1981).
Cussler, E.L.. AICHEJournal. 26, 1, 43-51 (1980).

Eley. D.D.. Trans. Faraday Soc., 35,1281 (1939).

Frank. HS. and M.W. Evans. J. Chem. Phys, 13.11, 507 (1945).

. Frank, HS. and W.Y. Wen. Disc. Faraday Soc., 24, 1333 (1957).
. Franks. F. in Hydrogen Bonded Solvent System_s_, A.K. Covington and

P.Jones (Eds). Taylor and Francis Ltd.. London, 31-47 (1968).

Franks. F.. J.R Ravenhill and DS. Reid J. Solution Chem.,l.1. 3 (1972 ).

Franks. F.. M. Pedley and DS. Reid, J. Chem. Soc. Faraday Trans. 1. 7 2.
359 (1976).

. Franks. F.. Pure and Appl. Chem. 59, 9.1189 (1987).

. Harvey. J.M. and M.C.R. Symons. J. Solution Chem, 7 (8). 571 (1978).
. Hildebrand. J.H., in Solubility. Wiley and Sons. N.Y.. 2nd. ed.

. Hussman. G.A., M.A. Larson and K.A. Berglund. in Industrial

ngstallization '84. SJ. Jancic and E.J. DeJong (Eds). Elsevier, Amsterdam,
21 (1984).
Jones. H.C.. Amer. Chem. Journal, 32.319 (1904).

Kaufman, V.R., D. Avnir, E. Pines-Rojanski and D. Huppert, J. Non-

Crystalline Solids. 99, 397 (1988).

Kondo. H.. I. Miwa and J. Sunarnoto. J. Phys. Chem, 86, 4826 (1982).

Kozak, J.J., W.S. Knight and W. Kauzmann J. Chem. Phys. 48. 2. 675
(1968).

Larson, M.A. and J. Garside, Chem. Eng. Science, 4 1. 1285 (1986).
MacMahon, P.M., K.A. Berglund and M.A. Larson in Industrial

ngstallization '84. SJ. Jancic and E.J. DeJong (Eds). Elsevier, Amsterdam.
229 (1984).

. Mathlouthi, M., C. Luu, M. Meffrov-Biget and D.V. Luu, Carbohydrate Res,

8 1. 213 (1980).
McMillan. WC. and IE. Mayer, J. Chem. Phys. 1 3(7). 276 (1945).

. Mullin. J.W. and C. Leci, Philos. Mag, 1 9. 1075 (1969).

. Mussel. RD. and D.G. Nocera, J. Am Chem. Soc., 1 10, 2764 (1988).

. Myerson, AS. and LS. Sorrel. AICHEJournal, 28,778 (1982).

. Myerson, AS. and Y.C. Chang, in Industrial Crystallization '84. SJ. Jancic

and E.J. DeJong (Eds), Elsevier, Amsterdam. 27 (1984).

. Narayanan, H. and GR. Youngquist, AICHE Symposium Series no. 253.

83.1-7 (1987).

. Nemethy, G. and HA. Scheraga. J. Chem. Phys. 36. 3401 (1962).

. Ostwald. W., Z. Phys. Chem, 22. 289-330 (1897).

. Pines, E. and D. Huppert, J. Phys. Chem. 87. 4471 (1984).

. Pauling. L. and O. Brockway, Proc. Natl. Acad. Sci.. 20, 336 (1934).

. Phillips, J.C., Trans. Faraday Soc., 3, 136 (1907).

. Pouxviel. J.C., B. Dunn and 1.1. Zink. J. Phys. Chem, 93. 2134 (1989).

. Sorrel. LS. and AS. Myerson, AIChE J., 28, 772 (1982).

. Stokes, RH. and RA. Robinson, J. Phys. Chem, 7 6, 2126 (1966).

. Suggett. A. and AH. Clark. J. Solution Chem. 5. 1. 1 (1976)

. Suggett. A., S. Ablett and P.J. Lillford, J. Solution Chem. 5(1). 17 (1976 ).

Suggett. A., J. Solution Chem. 5. 1,33 (1976).

Tait. M.J., A.Suggett, F. Franks. S. Ablett and PA. Quickenden J. Solution
Chem. 1, 2. 131 (1972 ).

Zinsli, P.E., J. Phys. Chem, 83. 3223 (1979).

 

Table 3.1. Molar ratio of total water to sucrose at various concentrations.

 

 

 

__Sucrose Water Sucrose Water NW:
gm. rrroles moles N IN

(103111;) (lo-3 kg) Nsu Nwat wat 8"
HT 90 0.029 5.00 171.0
20 80 0.059 4.44 76.0
30 70 0.088 3.89 44.3
40 60 0.1 17 3.33 28.5
50 50 0.146 2.78 19.0
60 40 0.175 2.22 12.6
62 38 0.181 2.11 11.6
64 36 0.187 2.00 10.7
66 34 0.193 1.89 9.8
68 32 0.199 1.78 8.9
70 30 0.205 1.67 8.1
76 24 0.222 1.33 6.0
80 20 0.234 1. 1 1 4.87

 

 

 

 

 

 

 

Table 3.2. Material Balances. Basis: 1 molecule sucrose. The entries in the second
column are the same as those in the last column of Table 1. The entries in the third
column are the PIR's (defined as the ratio of peak intensities at 440 nm and 511 nm)
obtained experimentally. The PIR in each case is set. for the purposes of analysis, to
the ratio of solvated water to bulk water. This is based on assignment of the peak at
440 nm to the microenvironment associated with solvated water and that at 511 nm to
the bulk water. Total molecules of water are set equal to the sum of the water
molecules in each microenvironment. thus yielding the values in the last two

 

 

 

columns.
Total Solvation Bulk water Solvation
Wt % molecules of water/ Bulk water

Sucrose water water

10 171.0 0.040 164.5 6.5
20 76.0 0.076 70.7 5.3
30 44.3 0.094 40.5 3.8
40 28.5 0.192 23.9 4.6
50 19.0 0.358 14.0 5.0
60 12.6 0.849 6.8 5.8
62 11.6 1.011 5.8 5.8
64 10.7 1.450 4.9 5.8
66 9.8 1.470 4.0 5.8
68 8.9 1.875 3.1 5.8
70 8.1 2.120 2.6 5.5
76 6.0 4.000 0.83 5.2
80 4.87 4.740 0.82 4.0

 

 

 

 

 

 

Chapter Four

Steady State Fluorescence Spectroscopy of Pyranine as a Trace Extrinsic
Probe to Study Structure in Aqueous Sugar Solutions

4.1 Background

The structure of supersaturated solutions and its relationship with crystal nucleation
and growth is an area of significant current activity. Myerson and co-workers19'21’ 65. 66'
75 performed diffusion studies in concentrated aqueous solutions of both non-electrolytes
(urea. glycine. and glycine-valine) and electrolytes (sodium chloride and potassium
chloride) that indicated formation of aggregates of an unknown composition as evidenced
by precipitous drops in diffusion coefficients. Tamman83 and later Mullin and Leci62
demonstrated that certain substances, for instance citric acid in aqueous solutions. formed
glass-like amorphous substances at high supersaturations which did not nucleate. Mullin
and Leci concluded that the ability or inability of a solution to nucleate must somehow
depend upon which molecules are contained in a supersaturated solution and how they are
spatially distributed, which in turn depends upon the various molecular interactions. They
reasoned that the molecular structure of citric acid favors intermolecular hydrogen bonding
and association. Chains or clusters of solute molecules in supersaturated (and even
concentrated but undersaturated) solutions were proposed as precursors to nuclei.
Subsequently. Mullin and Leci60 studied the desupersaturation of citric acid solutions, and
applied the cluster hypothesis to explain how and why nucleation did or did not occur from
supersaturated solutions.

In the study of contact nucleation of saccharides, Elankovan and BerglundZ733 and
later Cerreta and Berglund15 showed that under certain conditions the contact of a parent
crystal of a particular form and composition yielded some nuclei of a structure and

composition different from that of the parent crystal, in addition to nuclei that were of the

61

62
same structure and composition of the parent crystal. They also observed that under certain

conditions it is possible to grow crystals of one form on parent crystals of another form
Cerretal'7 concluded that neither the attrition nor the cluster theory of origin of secondary
nuclei provided satisfactory answers as to the origin and mechanism of this phenomena,
and that there was probably a third mechanism responsible. As a result of this work it
becomes clear that theories which are solvent exclusive in their descriptions of growth and
nucleation phenomena, such as the attrition and cluster theory. are unsatisfactory because
they cannot explain epitaxy and polymorphism. There is a growing amount of experimental
evidence relating to growth and nucleation phenomena which cannot be explained on the
basis of solute behavior alone. This leads us to believe that we must attempt to explain
these phenomena in terms of both solute and solvent behavior. Solvent behavior in
crystallization phenomena is not well understood. In order that the role of the solvent in
growth and nucleation be understood. the structure and properties of supersaturated
solutions must be determined.

Several attempts have been made to observe changes in macroscopic solution
properties and to correlate these to changes in solution structure. Larson and Garside-51-52
performed an experiment with isothermal columns of concentrated. saturated, and
supersaturated solutions of citric acid, urea. sodium nitrate. and potassium sulfate during
which they observed concentration gradients along the length of the column only for
supersaturated solutions. They concluded that there must be solute structuring in
supersaturated solutions. Narayanan and Youngquist67 attempted to relate the onset of
saturation with a marked change in measurable physical properties of bulk solution in
solutions of sodium nitrate. potassium nitrate. magnesium sulfate, potassium ammonium
sulfate. and ADP. No unusual changes in density, viscosity. or electrical conductivity with
saturation and subsequent supersaturation were observed. The results of the experiments of
Larson and Garside-5L52 indicate changes which can be observed on a macroscopic scale

63
with the onset of saturation; however, it is not possible to accurately correlate these

observations to changes in bulk solution structure at a molecular level.

Recent experimental attempts to study interactions in concentrated aqueous sugar
solutions have involved the use of Raman. IR, and NMR spectroscopies and X-ray
diffraction. These experiments will be summarized in turn as sugars are the object of the

present study.

4.2 NMR
Franks, Suggett and co-workers32'36v73'30v82 used dielectric relaxation and NMR to

determine solvation properties of sugars. Franks et al. confirmed the earlier results of
Jones41 and Phillips'71 for dilute solutions (an average of 6 water molecules of solvation
per molecule of sucrose) and also confirmed that the sugar water interaction was about as
strong as the water-water interaction. Harvey and Symons38 used NMR to determine 11
water molecules of solvation per molecule of sucrose. interpreted as 6 molecules of water
strongly associated with the molecule of sucrose and the remaining molecules more
dynamically associated. Using a pulsed NMR study of concentrated sucrose solutions,
Birch et al.10 determined that there are four kinds of protons in solution: bulk water.
solvated water. exchangeable sugar hydroxyl, and non-exchangeable ring (C-H) protons.
Franks. Suggett and co-workers32‘3ét73'80v32 as well as Barone and co-workers5'7 (who
used numerical simulations based on Kozak et al.'s50 modification of McMillan Meyer 53
theory) initially hypothesized that solute-solute interactions in dilute solutions were
unlikely. More recent work by Barone et al.4 indicates that the nature of interactions in
solution may be more complex than previously accepted without offering insight into the
exact nature of the complexity.

Richardson and Steinberg74 used 170 and 2H high field NMR to investigate
sucrose-water or D20 water systems over a wide range of concentrations and states. The

64
170 NMR transverse relaxation rate (R2) in D20 increased markedly with sucrose

concentration in supersaturated solutions of up to 70 wt.% sucrose. An isotropic two-state
model of water with fast exchange was used to interpret results. Deviations in linearity in
the results were hypothesimd to be due to formation of intermolecular hydrogen bonds
between water and sucrose. hydrogen bond bridging between water and sucrose molecules

and sucrose-sucrose hydrogen bonding. No discontinuity in R2 was observed with the

onset of saturation.

4.3 Raman Spectroscopy

X-ray diffraction. Raman spectroscopy and IR spectroscopy are sensitive to
changes in vibrational frequencies and space lattice ordering that occur as the concentration
of the solution increases. The effects of molecular association. and hydrogen bonding are
observed in Raman and IR spectra Mathlouthi et al.56 performed vibrational spectroscopy
of sugar solutions of various concentrations and observed that the characteristics of the
spectra were concentration sensitive. Bands assigned to the assymetric vibrations of the CH
and OH groups are observed to shift in frequency. In an attempt to correlate the change in
solution structure to change in concentration. the ratio of the integrated band intensity at
1460 cm'l, assigned to the HCH bending mode. to that at 1640 cm‘l. assigned to the HOH
bending mode was plotted as a function of concentration. As the concentration of sucrose
in solution increased from 10 to 66 wt%. and that of glucose increased from 5 to 54 wt%, a
10 cm'1 shift (approximately 1%). from 1450 cm-1 to 1460 cm—1. was observed in the
band associated with the HCH bend. This indicates that the HCH bend can be
accomplished with less energy in dilute solution than in-vacuo or in crystalline form, the
two standards that are commonly used for calculation of band positions and contributions
(Cael et al.13. Szarek et. al.32). As a result one is tempted to conlude that the particular

vibrational mode is about as hindered in concentrated solution as it is in vacuo or in

65
crystalline state; however. all one can safely conclude is that the CH2 moiety must be

hydrogen bonded to the same extent in solution as it is in a crystal. It is not clear from this
observation per-se that the CH2 group is associated to another sugar molecule; it might be
associated with water. For similar reasons it cannot be concluded that the shift in the HOH
bend frequency is evidence of solvent exclusive solute clustering; it can be concluded that
some of the water molecules behave in an ice-like manner (are tetrahedrally coordinated)
and the bending motion is hindered in concentrated solutions. It cannot be said conclusively
whether this hindering is due to association with the solvent or with the solute. In addition
to this uncertainty of interpretation is the fact that Raman spectra are solute sensitive. As the
concentration increases beyond saturation the changes in the spectra are less pronounced;

Mathlouthi's observations do not extend into the supersaturated region.

4.4 X-ray Diffraction

Mathlouthi's X-ray diffraction studies-57 of sugar solutions indicate formation of
intramolecular bonds with increase in solute concentration. He was able to convincingly
demonstrate that some sucrose molecules might be associating with some other sucrose
molecules. a result consistent with the observations of band shifting and intensification in
IR and Raman spectra. He was, however, unable to provide any convincing data or
argument that suggests that such clusters are larger than two solute molecules or that they

were exclusive of associated solvent.

4.5 Diffusion and Viscosity Measurements

Myerson and co-workersl9'21r55-66 performed diffusion studies in concentrated
aqueous solutions of both non-electrolytes (urea. glycine. and glycine-valine) and
electrolytes (sodium chloride and potassium chloride) that indicated formation of an

unknown composition as evidenced by precipitous drops in diffusion coefficients. Larson

66
and Garside 51-52 performed an experiment with isothermal columns of concentrated,

saturated, and supersaturated solutions of citric acid, urea. sodium nitrate, and potassium
sulfate during which they observed concentration gradients along the length of the column
for only the supersaturated solutions. Larson and Garside concluded that there must be
solute structuring in supersaturated solutions. By making the classical assumption that all
solute in solution must be associated with other solute molecules because of an absence of
solvent to satisfy the solvation requirements. the solvent exclusive clusters were proposed

to be comprised of several thousand molecules each.

4.6 Correlation of Macroscopic and Microscopic transport properties

Myerson64 in recent work has argued that if the constraint of solvent exclusion is

relaxed, then the probable size of clusters would be two or three solvated molecules of
solute. The studies of Myerson and coworkersl9'21v65’66’76 are particularly interesting
because they utilized measurements on the time scale of diffusion. Transport properties in
solution are governed by non-equilibrium solvation dynamics. There is evidence to suggest
that such phenomena are governed by short range interaction processes that take place on a
variety of time and length scales all of which are less than those associated with molecular
diffusion (W olyne836’37). Continuum models are still prevalent for such dynamics. Such
models do justice to the long range interactions in solution, but fail to account for the details
of structure and interaction of inner solvation shells which are often dominant factors
(Ween-936,87).

In the particular case of supersaturated solutions, where the molar ratio of solvation
water to bulk water per mole solute is such that there is only enough water available on the
average for one solvation shell or incomplete solvation shells. one would expect
interactions to be dominated by the non-equilibrium behavior. There are now several non-

continuum heuristic approaches to understanding such phenomena. One such widely

67
discussed approach. the mean spherical approximation or MSA. indicates that solvation of

an ion or molecule occurs on a range of time scales. A molecule can be thought to be made
up of component dipoles. The limiting scale of each such component is that of the order of
the hydrogen bond. The local solvent structure and dynamics near each such component
has been observed to differ significantly: even the solvation of a simple alcohol must be
described by at least two different time scales. One corresponds to the continuum
longitudinal relaxation time, the other is associated with the time required by solvent
molecules to rearrange themselves near the dipole. For small dipoles or ions this time is of
the order of a picosecond (W olynes8’7 ). Furthermore. the molecular dynamics simulations
of Cross and Simon24. suggests that if a dipole is strong enough (>17 D) then the two ends
may start behaving like point dipoles in the sense that the localized solvent structure and
dynamics may be different at the two ends of the dipole. This is summarized by Maroncelli
et. al.55 who conclude that the solvation response in a polar solvent involves coupled
reorientational dynamics that occur on time scales shorter than single particle diffusion.
There have been studies in the solvation of an electron in water and alcohols. using
picosecond and femtosecond time correlated spectroscopies. which indicate that solvent
reorientational times are on the order of < 500 fs.

Interestingly. Myerson used continuum arguments to reach his conclusions. This
would indicate that for the systems under consideration the macroscopic transport
coefficients in supersaturated solution must be very similar to the microscopic transport
coefficients in supersaturated solution.

With the exception of NMR the focus of previous work has been primarily on
solute sensitive techniques. Steady state fluorescence spectroscopy using extrinsic trace
fluorescent probes is an experimental technique that is solvent sensitive like NMR, rich in
information content, experimentally simple, yet not currently utilized for studying structure
in solutions. This approach has been exploited and developed to a great degree of

sophistication for studying microhetreogeneous systems; Kalyansundaram42 has presented

68
a comprehensive treatise on principles involved and techniques available. The development

and analysis of a steady state fluorescence technique based on monitoring the solvent
sensitive emission behavior of an extrinsic fluorescent probe is discussed in the current

work.

4.7 Photochemical Techniques using Extrinsic Trace Fluorescent Probes

Fluorescence spectroscopy of extrinsic probe molecules in solution offers a
complementary approach to measuring many variables of interest. Fluorescence is the
consequence of fast radiative transition to the ground state in a molecule that has been
excited by radiation to an allowed electronic state of higher energy. Fluorescence
techniques have high quantum efficiencies (07-09). but are restricted in application to fast
kinetic processes. The unimolecular photochemical reaction of electronically excited
molecules in homogeneous media can be used to study relaxation processes on a variety of
time scales. Simple steady state measurements of intensity and wavelength of the observed
fluorescence are rich in information about to the immediate molecular environment of the
probe molecule (or moiety). The nature of the probe environment. concentrations,
distances. accessibility, possible complexing. electric field. polarity, lateral diffusion.
functional correlations. reactivities and pH are all reflected in the emission behavior of the
probe“.

It is important to distinguish between the photochemical investigation of a
fluorescent compound and the experiment utilizing an extrinsic trace fluorescent probe. The
experiment involving a trace fluorescent probe differs from the photochemical investigation
of a fluorescent compound in several crucial aspects. The former makes use of a compound
for which the photochemical behavior is well known; the photochemistry of the compound

is the subject of the latter. Probe experiments require very low molarities of the fluorescent

69
compound; photochemical investigations may require high molarities. In a photochemical

investigation changes in emission of the fluorescing species gives information about its
electronic state. Changes in the emission of the extrinsic probe molecule (or moiety) can be
correlated to corresponding changes in solvent properties. Often photochemical
investigations are experimentally difficult because the high molarity of the fluorescent
species gives rise to a host of associated phenomena which may also affect its excited state
behavior. Extrinsic fluorescent probes. on the other hand, can be designed or judiciously
chosen to be present at molarities low enough to preclude any effect other than that due to
the solvent environment of the probe (all other experimental conditions being constant).

The way the microenvironment of the probe responds to the perturbation can yield
valuable information about its structure. A probe that is chosen because it exhibits an
affinity for a certain type of site on a molecule, or a certain kind of interaction. will reflect
the nature of its microenvironment in its emission properties. Often the sensitivity of a
probe to solvent environment is enhanced in its excited state. Fluorescent compounds at
very low molarities are thereby sensitive molecular probes of their solvent environment

For this study, we chose a water soluble probe that is known to have a highly
sensitive. polarity dependent equilibrium between two excited states that fluoresce. We
chose pyranine (or 8-hydroxy-1,3,6-pyrenetrisulfonate), Figure 4.1. Its photochemistry
has been well characterized (Kondo et al.”). In solution pyranine exists with the sulfonate
groups entirely dissociated. It has two excited fluorescent states: one in which the hydroxyl
proton is associated with the molecule and one in which it is dissociated. Figure 4.2. The
two excited states differ only in the hydroxyl proton. Emission is observed from the
molecule in the former state at 440 nm and from that in the latter at 511 nm. when excited at

342 nm corresponding to an absorption isosbestic point. Pyranine in its excited state is
sensitive to the concentration of molecular moieties which can participate in exchange
processes in the vicinity of the hydroxyl proton (i.e. polarity or pKa). Incident exciting

radiation causes a rt-rt" singlet state transition in the probe molecule. This energy is then

70
released mostly through a fast kinetic radiative path on the order of a 100 picoseconds with

a quantum efficiency of nearly 80%. The deprotonated state requires less energy of
excitation and thus emission from the molecule in this state is observed to the red of that
emitting from the protonated state. Whether the observed emission is from the excited
deprotonated or protonated form is determined by the molecular environment in its
immediate vicinity. We can correlate changes in the fluorescence equilibrium between the
two emitting forms of the probe to changes in the nature of the microenvironment of the
hydroxyl moiety of the probe. Once we establish the nature of the nricroenvironment in
terms of the species in solution we can then relate changes in the fluorescence equilibrium
to changes in solution structure.

Pyranine has been used to probe water content of nricellar dispersions3. 23- 43. 49’
72, sols“4 and sol-gels73 among other systems. It has been used to relate water content of
these systems to the structure of various phases. It is photochemically active, even at
concentrations of 1 ppm (1.9 x 10‘6 M) or less. In addition. at these concentrations it is
non-toxic. non-carcinogenic. and suitable for use in food grade applications (approved for

use as a food color additive. USDA regulations 1988).

4.8 Materials and Methods

Emission spectra were obtained with either a 200 W Xe/Hg or a 150 W Xe source
in quartz cuvettes using an in-house constructed spectrometer described in detail elsewhere
(Mussel and Nocera53). The lamps were operated at 80% rated power. A Hamamatsu
R1104 photomultiplier tube (maintained typically at 50% maximum voltage) was used in
conjunction with a preamplifier (maintained at a sensitivity of 10 '6) and a Princeton
Electronics lock-in amplifier (maintained at a sensitivity of 3V). The photomultiplier tube

was maintained at -60 degrees C to provide a large signal to noise ratio. A 375 nm KV or a

71
370 nm GG cutoff filter was used to filter out source excitation at 342 nm. A 3 mm slit

width was used and shutters were kept fully open.

Reagent grade sucrose. glucose. and lactose monohydrate from Sigma Cherrricals.
fructose from A.E. Staley, 8-hydroxy 1,3,6. pyrenetrisulfonate (pyranine) from Eastman
Kodak. and 18 Mohm reverse osmosis water were used. Stock solutions of 100 ppm (1.9
x 10'4 M) pyranine were prepared using 18 Mohm reverse osmosis water. Solutions of the
10 ppm (1.9 x 10‘5 M) pyranine and 1 ppm (1.9 x 10‘6 M) pyranine were prepared by
dilution.

The amount of sugar and water required for 10 grams (.01 kg) of each solution was
calculated. Reagent grade sucrose samples were carefully measured out to +/- 1% of the
required amount Samples were prepared in 20 ml glass vials by adding the predetermined
amount of stock pyranine solution to achieve the desired concentration. Samples varied in
concentration between pure water and 80 wt.% sucrose. In undersaturated solutions
samples were prepared in 10 wt.% intervals. Samples near saturation were taken at 2 wt.%
intervals, eg. for sucrose between 60 and 74 wt%, and solutions at high concentrations.
eg. 76 wt.% and 80 wt.% solutions of sucrose, were also prepared. Supersaturated
solutions were prepared by heating the vials in a water bath until the sugar dissolved.
Spectra of the solutions were taken within 6 to 24 hours of preparation to ensure
equilibration of the sucrose in solution [with respect to temperature] and prevent biological
contamination of the sugar solutions. Solutions of f ructose. glucose, and lactose were
prepared in a similar manner. by weighing out predetermined amounts of sugar and stock
solution. by agitating,and heating to dissolve all the sugar. and followed by cooling to 20
0C in order to obtain supersaturated solutions. Fructose samples were prepared at 10 wt.%
intervals up to 70 wt.% and then at 2 wt.% intevals up to 84 wt.%. Glucose solutions were
prepared at 5 wt.% intervals upto 40 wt.% and then at 2 wt.% intervals from 43 wt% to

55 wt%. lactose mono hydrate solutions were prepared at 5 wt.% intervals to 45 wt%.

 

72
Pyranine doped sugar solutions - with pyranine concentrations of 1 ppm (1.9 x 10'

6 M), 10 ppm (1.9 x 10-5 M). and 100 ppm (1.9 x 10-4 M)- exhibited little variation in
peak intensity ratio (PIR). The PIR is defined as the ratio of the peak intensity at 440 nm to
that at 511 nm This is the same as the F ratio described in Pouxviel et.al. (5_0). It was
decided to work at 100 ppm (1.9 x 10'4 M) levels of pyranine to maximize the signal to
noise ratio. Small changes in instrument settings did not affect the values obtained for the
PIR.

There was less than 3% variation in the PIR of pyranine in undersaturated sucrose
solutions over a 30 degree temperature span from 20 to 50 degrees C . This indicates that
the equilibrium constant between the two excited states is not a strong function of
temperature in this range of operation. Therefore temperature corrections were not applied.

Small variations in excitation wavelength (+/- 5 nm) did not change the PIR (less
than 3%); however, large changes in the excitation wavelength (335-365 nm) indicated that
the PIR were up to 5% less at longer wavelengths than at shorter ones. This behavior can
be explained on the basis of the difference in the amount of light absorbed by the two states
of pyranine. For a solution of given sucrose concentration doped with a fixed amount of
pyranine the change in intensity of the peak at 511 nm with increase in wavelength is
greater than that associated with the peak at 440 nm. This accounts for the difference in PIR
due to large changes in excitation wavelength. In order to get consistent results an
excitation wavelength of 342 nm was used for all experiments.

Emission spectra were obtained for each of the sample solutions prepared. The
PMTresponsewasaboutthesameatMOnmandSll nmsothespectrawere not corrected
for PMT response.

4.9 Results and Analysis

73
The emission spectra for pyranine doped sucrose solutions of 0, 20.40.60, and 70

wt% are shown in Figure 4.3. The intensity of both the blue (440 nm) and the green (511
nm) peaks change with the amount of sugar and water present in solution. The ratio of the
peak intensity at 440 nm to that at 511 nm was defined previously as the peak intensity ratio
(PIR).The PIR increases as the concentration of sucrose increases and the concentration of
water decreases. A clear isosbestic paint is observed at 485 nm which also indicates the
presence of two emitting species in equilibrium.

There are two possible excited states from which pyranine molecules may emit. The
farm that contains the hydroxyl proton can exist only in an environment that does not
facilitate the transfer of protons. The other farm exists where it can transfer protons. The
fact that emission from both forms is observed in solution leads us to conclude that
microenvironments suitable to each species must ca-exist in solution. This argument is
along the same lines as that put forward by Zinsli88 which led to a biphasic model for
micelles.

In the current context microenvironments refer to the presence or absence of
exchangeable protons in the immediate proximity of the pyranine hydroxyl proton.
Emission observed at 440 nm indicates that there are protonated pyranine molecules in
solution. Protonated pyranine molecules exist where there are no protons available for
exchange in the immediate vicinity of the pyranine hydroxyl proton. Sinrilarly. emission
observed at 511 nm indicates that there are deprotonated pyranine molecules in solution. So
the PIR is a measure of the ratio of protonated to deprotonated pyranine in solution; and
therefore (as explained in the following discussion) a measure of the ratio of water
associated with the salute to that not associated with the solute.

According to the pulsed NMR work of Birch et al. there are 3 kinds of exchangable
protons in concentrated sucrose solutions: bulk water. salvation water. and sucrose
hydroxyl. Of these. the exchange rate of the sugar hydroxyls is somewhat slower than that

of the salvation water. For reasons of mobility and polarity. it is unlikely that the sucrose

74
molecules would be able to compete with water molecules in interacting with pyranine

molecules to a significant extent.

A water molecule in contact with other water molecules that behave like pure water
could provide an environment that is capable of proton exchange. If the pyranine hydroxyl
is in the proxinrity of such a water molecule it would emit at 511 nm. Thus. we propose
that one of the two microenvironments is likely to be a bulk-water-like nricrophase. A
pyranine hydroxyl group existing near salvation water already associated with a molecule
of sucrose could not exchange protons and would emit at 440 nm. Thus. the other likely
microenviromnent could be due to the presence of an envelope of salvation in general and
more specifically due to the presence of water that chemically interacts with the molecule in
the sense of the three-state model for solutions. as in NMR.

A graph of the PIR versus concentration of sucrose (wt %) is shown in Figure 4.4.
Since pyranine has only one proton available for exchange, it is reasonable to conclude that
only one molecule of water at a time can participate in the exchange process. It follows that
the PIR (a ratio of pyranine molecules incapable of exchanging protons to that of pyranine
molecules capable of exchanging protons) is also the ratio of water molecules that do not
participate in proton exchange with the pyranine to those that do. Furthermore. we can
scale this per molecule of sucrose. This is purely forconvenience. It then becomes possible
to look at the solution in terms of water of salvation per molecule of sucrose (unavailable
for proton exchange) and bulk water (available for proton exchange); terms that can be
compared with existing data, at least f or undersaturated solutions. Similar results were
obtained from the analysis of glucose, fructose. and lactose. Figures 4.5, 4.6, and 4.7
respectively. It is apparent that the change in the PIR with saturation is not so marked in
the case of glucose and lactose as it is for sucrose and fructose.

Assuming that the quantum yield of the two emitting species is the same (same
chromophore). we can plot a graph of the number of molecules of water in each of these

microenvironments per molecule of sucrose. We calculate 6 molecules of salvation water

75
per molecule of sucrose in undersaturated solutions. dropping to 5.8 molecules at the

saturation point. 5.5 at 70 wt%, 5.2 at 76 wt.%, and 4.0 at 80 wt%; see Tables 1 and 2. It
would seem reasonable that salvation water per molecule would be a function of the
molecular structure. more specifically the number of sites (and conformation of these sites)
on the molecule that could hydrogen bond with nearby water. Of equal or greater
significance is the behavior of the bulk water. The number of molecules of water in the
bulk-water-like environment drops from 165 for a 10% sucrose solution to about 3 for a
highly supersaturated solution of 70% sucrose. It drops further to a value slightly less than
1.0 (0.83) for the 76% solution and stays about constant for the 80% solution.

Although the determination of water of salvation is in agreement with the results of
Jones“. Phillips”. and Franks, Suggett and co-workers32‘35» 7830' 32. we were not
entirely comfortable with the assumption of equal quantum yields for the two emitting
species. In an attempt to isolate the two species and estimate relative quantum yield. spectra
of equal amounts of carefully measured pyranine were obtained in methanol and water
(under the same conditions as the rest of the experiments). The only peak observed in
methanol was that at 430 nm. In water, a weak shoulder was observed at 440 nm. but the
emission was primarily at 511 nm. The peak in methanol was approximately 1.8 times as
intense as the peak in water for any given concentration of pyranine (between 1-80 ppm
(1.9x 10’8 and 1.6 x 10’5 M) ). This would indicate ll molecules of salvation water per
molecule of sucrose. in agreement with the results of Harvey and Symons38 and the earlier
calculations of Stokes and Robinson”. It would also indicate smaller numbers for both the
bulk-water-like phase and the solvated water in the supersaturated region. The smaller
amount of salvation water might indicate that the more dynamically associated water
(according to Harvey and Symons) might be depleted in order to solvate solute in excess of
the amount in a saturated solution. The basic concept that saturation occurs when it is no
longer energetically favorable to solvate more solute molecules without supplying external

energy. as evidenced by an equilibrium between the bulk and salvation water. remains

76
unchanged. It is worth noting that the spectra in methanol were blue-shifted about 510 nm

from those in water. This may be due to a difference in polarity of microenvironment or
due to a green bleed through in the spectra in water. At this point it is not appropriate to
conclusively state that both emitting species have the same quantum efficiencies although
this is what our arguments and observations would lead us to believe.

In either case (6 or II molecules of salvation water per molecule of sucrose) the
surprising observation was that there should be any bulk phase water at all in saturated and
supersaturated solutions. The number of water molecules in the bulk phase per molecule of
sucrose decays exponentially with increase of sucrose concentration. Near saturation the
number of water molecules per molecule of sugar in the bulk phase is roughly equal to the
the number of molecules of water of salvation. For a highly supersaturated solution of 70%
sucrose there are roughly three molecules of bulk phase water and six molecules of
salvation water per molecule of sugar. It is probably the water in the bulk phase that is
responsible for structurally ”stable" supersaturated solutions. The ease with which sucrose
can be dissolved in solution appears to be directly related to the amount of bulk phase
water.

An interesting feature of this analysis is that the water of salvation per molecule of
sucrose becomes approximately equal to the bulk water per molecule of sucrose at
saturation. In supersaturated solutions there is always less bulk water available per
molecule of sucrose than salvation water. While this feature is apparent from Figure 4. 8, it
becomes more so in Figure 4.9. The crossing point of the two curves indicates equilibrium.
At this point it becomes energetically unfavorable to add more salute to the solution without
supplying external energy.

When this analysis is applied to glucose (Figure 4.10), fructose (Figure 4.11). and
lactose (Figure 4.12). essentially the same trends are observed. except that the bulk water
and salvation water curves no longer cross at saturation in the case of glucose and lactose.

If we draw upon NMR analogy. sucrose solutions can be represented well by a two-state

77
isotropic model for water as in the work of Richardson and Steinberg74. Thus a two-state

analysis as presented f or sucrose is expected to work well. as it indeed does. Fructose
which has several hydroxyl groups that exchange at a rate similar to bulk water is observed
to have two molecules of water of salvation per molecule of fructose.

Pyranine can interact with water in one of two ways in the steady state emission
experiment. Thus, it is possible to measure only two "kinds” of water. that which will
exchange with it and that which will not. It is possible that glucose and lactose bath
associate with water in a third manner. so that pyranine interprets the water in the salvation
shell as "free” for purposes of exchange, whereas in reality it exchanges at an intermediate
rate, i.e., slower than bulk water but faster than salvation water.

In the case of glucose we may apply the following analysis. We observe from an
initial analysis along the lines of that for sucrose. that there appear to be four waters of
salvation per molecule of glucose. At saturation there now appear to be 8 molecules of bulk
water per molecule of glucose. We ask ourselves if this water is capable of solvating more
glucose then why is the solution saturated? Perhaps glucose molecules associate water at a
rate that is faster than what our probe percieves to be associated water but slower than bulk
water.

Brady's study of molecular dynamics of solvated glucose in a 15 wt% solution
suggests that the residence times of water molecules near each of the sugar oxygen atoms is
different He observed that it takes a small (on the scale of several fs) but finite amount of
time for solvent exchange to occur near each of the OH or O moieties in the glucose
molecule. Thus on some very small but finite time scale each of these groups remains
relatively ”unattached” to any specific water molecule. Four of these times are relatively
slow; two are of the same order of magnitude as the residence time of a water molecule near
another water molecule: roughly twice as fast as the other four. All exchange processes and
rotational correlation times can be classified into three groups. This is interesting in the light

of our steady state emission studies with pyranine as an extrinsic trace fluorescent probe in

78
solution. because it supports the idea that at least three significant exchange time scales are

involved.

Based on this information we assume that glucose associates 2 molecules of water
per remaining sugar oyxgen group (four of which are already accounted for as being
solvated). This adds up to 4 more water molecules (each of the other groups associate with
roughly double the water in the same period of time). When we take these extra molecules
into account and apply it to our analysis we observe that the bulk water does indeed becorrre
equal to the salvation water at or near saturation (for DS=0). Thus, a three state model for
water allows us to explain the behavior of aqueous glucose solutions well. Figure 4.13.

Lactose shows on the average of 3 waters of salvation per molecule, Figure 4.12.
If the three state model is postulated for lactose and a correction of 2 water molecules is
applied for each oxygen unaccounted for ie., 11-3=8. the curves still do not cross at

saturation, but a much better fit of data is obtained. Figure 4.14.

4.10 Conclusions

It is possible to calculate a quantity which we call the peak intensity ratio (PIR)
from steady state emission spectra of trace quantities of pyranine in solutions of sucrose,
fructose. glucose and lactose in polar solvents. It appears that the PIR obtained from
fluorescence spectra of pyranine in aqueous sucrose solutions shows a marked change at
saturation and is very sensitive to water content of supersaturated solutions. This is a
sensitive technique for measuring degree of supersaturation. The PIR obtained from the
fluorescence emission of pyranine in aqueous solutions of fructose also exhibits similar
behavior. Such a marked change in PIR was not observed with the onset of saturation in

the fluorescence emission of pyranine in aqueous solutions of glucose and lactose .

79
Structurally. the solvent appears to distribute itself in at least two. and in some

cases three, kinds of microenvironments. One corresponds to the water strongly associated
with the solute molecule. designated salvation water, and the other is solvent that
associates primarily with other solvent molecules, designated bulk phase water. The third
state corresponds to water in an envelope of salvation associated with the molecule in a
more dynamic fashion and cannot be measured by the probe directly. This water is viewed
by the probe to be bulk water. The amount of water present in the bulk phase decreases
exponentially with increase in solute concentration and correlates with the ease of salvation
of the solute. At the saturation point the amount of water in the bulk phase per molecule of
sucrose is roughly the same as the number of water molecules of salvation per molecule of
sucrose. This was also observed to be true for fructose. Lactose and glucose were
observed to behave in roughly the same way if it was assumed that some water was
available for exchange in a "third state", ie., at a third or intermediate rate.

Most of the previous studies of structure in concentrated and supersaturated
solutions have focused on solute-solute interactions and solute structuring or aggregation.
We have developed a means of extracting information about the solvent distribution in
sugar solutions using steady state fluorescence spectroscopy of an extrinsic trace
fluorescent probe. We have demonstrated that this technique works well for analysis of
aqueous solutions of polar organics and that the best data are obtained for systems that can
be described by an isotropic two—state exchange model for water. Reasonably good results
can be obtained for systems described by a three state model of exchange for water. Our
results do not indicate the presence of solvent-exclusive precrystalline clusters; instead we
observe the sugar molecules to be at least partially solvated in supersaturated solutions.
even at moderate degrees of supersaturation. It is suggested that the results of the various
studies in solution structure are not inconsistent with each other. We suggest they may be

consistent with each other if the supersaturated solution contains some moieties, solvent

80
and/or solute. that participate in bulk-water—like fast exchange and at least some molecules

of solute that are partially solvated at some moieties some of the time.

4.1 1 Acknowledgements

The authors wish to thank Dr. D.G. Nocera, of the Department of Chenristry.
Michigan State University for kindly permitting the use of the emission spectrometer in his
laboratory. Financial assistance from the Department of Chemical Engineering and the Crop
and Food Bioprocessing Center! Research Excellence Fund at Michigan State University is
greatly appreciated. Support from the Cooperative State Research Service (CSRS) of the
United States Department of Agriculture (Grant No. 90-24189-5014) is also
acknowledged. RC. wishes to thank Dr. R. I. Cukier. of the Department of Chenristry.

Michigan State University for several very useful discussions and comments.

4.12 References

1. Back D.M., D.F. Michalska, and P. L. Polavarapu, Appl. Spectroscopy, 38(2), 173.
(1984).

2. Barbara RF. and W. Jarzeba. Acc. Chem. Res. 2 1. 195. (1988).

3. Bardez. E., B.T. Gouguillon, E. Keh and B. Valeur, J. Phys. Chem, 88, 1909.
(1984).

4. Barone. G., G. Castronuovo, P. Del Vecchio and V. Elia. J. Solution Chem. 19. l.
41 ( 1990).

5. Barone. G.. G. Castronuovo, P. Del Vecchio. V. Elia and M.T. Tosto. J. Solution
Chem. 17. 10. 925 (1988)

6. Barone. G., G. Castronuovo, V. Elia and M.T. Tosta, J. Solution Chem. 1 5, 2, 199
(1986).

7. Barone, G.. G. Castronuovo, V. Elia and V. Savino, J. Solution Chem. 13, 3. 209
(1984).

8. Ben-Naim. A. . Salvation Themrodypm’cs, Plenum Press, New York (1987).

81

9. Bemal, JD. and RH. Fowler. J. Chem. Phys. 1, 515 (1933).

10. Birch. G.G., J. Grigor and W. Derbyshire, J. Solution Chem. 1 8, 8. 795 (1989).

11. Bousfield. W.R and TM. Lowry, Trans. Faraday Soc., 6. 85 (1910).

12. Brady J.W., J. Am. Chem. Soc., 1 1, 5155. (1989).

13. Cael J.J., J.L. Koenig and J. Blackwell. Carbohydrate Res, 32, 79, (1974).

14. Cerreta. MK. and K.A. Berglund. in Industrial prstallization '84, S.J.Jancic and
EJ. DeJong, (Eds). Elsevier, Amsterdam, 21 (1984).

15. Cerreta, M.K and KA. Berglund. Journal of Cryst. Growth, 102, 869, (1990).

16. Cerreta, M.K and K.A. Berglund, Journal ofCrystal Growth , 84. 577 (1987).

17. Cerreta. M.K.. Ph.D dissertation, Michigan State University (1988).

18. Chakraborty R. and K.A. Berglund, AIChE Symposium Series no. 284. 113.
(1991).

19. Chang, Y.C. and AS. Myerson, AIChE J., 3 1. 980 (1985).

20. Chang. Y.C. and AS. Myerson, AIChE J., 32. 1567 (1986).

21. Chang, Y.C. and AS. Myerson. AIChE J., 32, l747,(1986).

22. Chapman C. F., R. S. Fee, and M. Maroncelli, J. Phys. Chem. 94. 4929. (1990).

23. Clement, NR. and M. Gould. Biochemistry, 20, 1534 (1981).

24. Cross. A. J. and JD. Simon. J. Chem. Phys, 86, 7079, (1987).

25. Cussler, E.L., AICHEJournal, 26. 1. 43-51 (1980).

26. Delhaye M. and P. Dhamelincourt, J. of Raman Spectroscopy, 3. 33. (1975).

27. Elankovan, P. and K.A. Berglund, AIChE J., 33 .1844. (1987).

28. Elankovan, P., and K.A. Berglund. Appl. Spectroscopy. 40 . 712. (1986).

29. Eley. D.D.. Trans. Faraday Soc., 35,1281 (1939).

30. Frank. HS. and M.W. Evans, J. Chem. Phys, 1 3.11. 507 (1945).

31. Frank, HS. and W.Y. Wen. Disc. Faraday Soc., 24. 1333 (1957).

32. Franks. F. in Hydrogen Bonded Solvent Systems, A.K. Covington and P.Jones
(Eds), Taylor and Francis Ltd.. London, 31-47 (1968).

33. Franks, F.. J.R Ravenhill and DS. Reid J. Solution Chem..1.1. 3 (1972 ).

34. Franks, F.. M. Pedley and DS. Reid, J. Chem. Soc. Faraday Trans. 1. 72. 359
(1976).

35. Franks. F.. Pure and Appl. Chem. 59. 9.1189 (1987).

36. Giermanska J. and M.M. Szostak. Journal ofRaman Spectroscopy, 2 2. 107. (1991).

37. Hameka H.F.. G.W. Robinson. and OJ. Marsden, J. Phys. Chem, 9 1.
3150,(1987).

38. Harvey. J.M. and M.C.R. Symons. J. Solution Chem, 7 (8). 571 (1978).

82

39. Hildebrand. J.H., in Solubility. Wiley and Sons. N.Y., 2nd. ed.. (1965).

40. Hussman. G.A., M.A. Larson and K.A. Berglund, in Industrial ngstallization '84,
SJ. Jancic and EJ. DeJong (Eds). Elsevier, Amsterdam. 21 (1984).

41. Jones, H.C.. Amer. Chem. Journal. 32.319 (1904).

42. Kalyansundaram . Photochermjggy in MicroheterogpneousLSystemp.

43. Kano. K and J .H. Fendler, Biochim. Biophys. Acta, 509. 289 (1978).

44. Kaufman. V.R. D. Avnir, E. Pines-Rajanski and D. Huppert, J. Non-Crystalline
Solids, 99, 397 (1988).

45. Kenney-Wallace GA. and CD. Jonah. J. Phys. Chem. 86, 2572. (1982).

46. Kenny Wallace GA. and CD. Jonah, Chem. Phys. Lett., 39(3), 596, ( 1976).

47. Kenny-Wallace G.A., Can. J. Chem, 55, 2009. (1977).

48. Kenny-Wallace G.A., G.E. Hall. L.A. Hunt. and K. Sarantidis. J.Phys. Chem. 84.
1145. (1980).

49. Kondo. H., I. Miwa and J. Sunamoto, J. Phys. Chem, 86. 4826 (1982).

50. Kozak, J.J., W.S. Knight and W. Kauzmann J. Chem. Phys. 48, 2, 675 (1968).

51. Larson, M.A. and J. Garside. Chem. Eng. Science. 4 1. 1285 (1986).

52. Larson. M.A. and J. Garside. J. Crystal Growth. 7 6. 88 (1986).

53. Larson. M.A., AICHE Symposium Series no.240. 80. 39-44 (1984).

54. MacMahon, P.M., KA. Berglund and M.A. Larson in Industrial ngstallization '84.
SJ. Jancic and E.J. DeJong (Eds), Elsevier, Amsterdam. 229 (1984).

55. Maroncelli M., J. Maclnnis. and GR. Fleming. Science. 243, 1674, (1989).

56a. Mathlouthi M and D.V. Luu, Carbohydrate Res. 7 8, 225. (1980).

b. Mathlouthi M. and D.V. Luu, Carbohydrate Res, 8 1. 203, (1980).
c. Mathlouthi, M., C. Luu, M. Meffrov-Biget and D.V. Luu, Carbohydrate Res. 8 1.

213 (1980).

57. Mathlouthi M., Carbohydrate Res, 9 l. 113. (1981).

58. McMillan. W.G. and J .E. Mayer, J. Chem. Phys. 1 3(7), 276 (1945).

59. Migus A., Y. Gauduel, J.L. Martin, and A. Antonetti, Phys. Rev. Lett. . 58(17),
1559, (1987).

60. Mullin, J.W. and C. Leci, AIChE Symposium Series no 68. J. Estrin ed.. 8. (1972).

61. Mullin, J.W. and C. Leci. a. Philos. Mag., 1 9, 1075 (1969).

b. J. Crystal Growth. 5. 75, (1969).

62. Mullin J.W. and C. Leci, Chemy Ind., 1517, (1968).

63. Mussel. RD. and D.G. Nocera, J. Am. Chem. Soc., 1 10. 2764 (1988).

64. Myerson A.S.. private communications to K A. B.

65.
66.

67.

68.
69.
70.
71.
72.
73.
74.

75.
76.

83

Myerson, AS. and LS. Sorrel. AICHEJournal. 28.778 (1982).

Myerson, AS. and Y.C. Chang. in Industrial ngstallization '84. SJ. Jancic and E.J.
DeJong (Eds). Elsevier, Amsterdam. 27 (1984).

Narayanan, H. and GR Youngquist. AICHE Symposium Series no. 253, pp. 1-7
(1987).

Nemethy. G. and HA. Scheraga, J. Chem. Phys. 36. 3401 (1962).

Ostwald, W., Z. Phys. Chem, 22. 289-330 (1897).

Pauling. L. and O. Brockway, Proc. Natl. Acad. Sci. 20. 336 (1934).

Phillips. J.C., Trans Faraday Soc., 3. 136 (1907).

Pines. E and D. Huppert, J. Phys. Chem. 87. 4471 (1984).

Pouxviel. J.C., B. Dunn and J.I. Zink. J. Phys. Chem. 93. 2134 (1989).
Richardson SJ. and MP. Steinberg. in Met Acfitivity: Theory and Applications to
goo—,d L. B. Rockland and L. R. Beauchat eds. Marcel Dekker Inc., New York. 235.
(1987).

Simon. J.D.. Acc. Chem. Res, 2 1. 128. (1988).

Sorrel. LS. and AS. Myerson, AIChE J., 28. 772 (1982).

77. Stokes. RH. and RA. Robinson. J. Phys. Chem, 7 6. 2126 (1966).
78.
79
80

Suggett, A. and AH. Clark, J. Solution Chem, 5. l. 1 (1976).

. Suggett. A., J. Solution Chem. 5, 1,33 (1976).
. Suggett. A., S. Ablett and P.J. Lillford, J. Solution Chem, 5(1), 17 (1976 ).
81.

Szarek W.A., S. Korppi-Tommola. H.F. Shurvell, V. H. Smith Jr., and OR Martin,
Can. J. Chem. 62. 1512. (1984).

82. Tait, M.J., A.Suggett, F. Franks. S. Ablett and PA. Quickenden J. Solution Chem.

83.

84.

85.

$38

1, 2.131(1972).

Tamman G.. States of Aggregation, trans]. RF. Mehl, van-Nostrand. New York.
(1925).

Teale. F.W.J . and D. Constable,”lntramolecular Thiol-specific Probes” in Fluorescent
PM G.S. Beddard and M.A. West (Eds). Academic Press. New York, 1-19.
(1981)

Walafren G.E., in Water a Comprehensive Treaiige, Vol. 1, F. Franks ed.,Plenum
Press, New York. 151. (1972).

. Wolynes PG, Annu. Rev. Phys. Chem. 3 1. 345. (1980).
. Wolynes PG. J. Chem. Phys. 86(9), 5133, (1987).
. Zinsli, P.E., J. Phys. Chem. 83, 3223 (1979).

NaO 3 : SO3N a

Figure 4.1. Chemical structure of pyranine or 8-hydroxy-l . 3. 6-pyrenetrisulfonate

 

 

 

 

 

 

 

 

440nm 511 nm
km a to Abl“: t0
450nm 514nm
Pyorsr" ‘ e Pyo'*+ Igo“
A A
A-ex =342nm bex=342nm
W W
V V
PyOH . ‘ Pyo— +1130“
"ab' 350nm half 405nm

Figure 4.2. Emission equilibrium of pyranine. The equilibrium between the protonated
form (absorbs at 350 nm and emits at 440 nm when excited at 342 nm) and the
deprotonated form (absorbs at 405 nm and emits at 511 nm when excited at 342 nm) is

determined by ability of the nricroenvironment of the hydroxyl group to participate in
proton transfer.

 

a a

Intensity
8
O

b)
O
0

§

 

 

 

ryrrrrllr’

500 550 i 600
Wavelength

Figure 4.3 . The emission spectra of pure water and sugar solutions of 20. 40, 60, and
70 wt.% sucrose doped with 100 ppm ( 1.9 x 10‘4 M) pyranine at 20 degrees C. Emission
maxima occur at 440 nm and 511 nm. An isosbestic point is observed, represented by the
painter at 485 nm.

 

PIR

 

 

 

 

 

o 20 4o 60 ' 80

Figure 4.4. Graph of PIR (peak intensity ratio) vs. wt% sucrose in solution at 20
degrees C using 1(1) ppm (1.9 x 10-4 M) pyranine. The PIR is defined as the ratio
of peak intensities at 440 nm and 51 1 nm. The arrow represents concentration at
saturation

 

PIR

 

 

 

 

 

10'1'5'26'23'3b'3'5'4b'4'5 5'0'55
Wt% Glucose

Figure 4.5. Graph of PIR (peak intensity ratio) vs. wt% glucose in solution at 20
degrees C using 100 ppm (1.9 x 10-4 M) pyranine. The PIR is defined as the ratio
of peak intensities at 440 nm and 51 1 nm. The dashed line represents concentration
at saturation.

 

PIR
‘t’

 

 

 

 

Wt.% Fructose

Figure 4.6. Graph of PIR (peak intensity ratio) vs. wt% fructose rn solution at 20
degrees C using 100 ppm(1. 9x 104M) pyranine. The PIRis definedasthe ratio

of peak intensities at 440 nm and 51 1 nm. The arrow represents concentration

at saturation

 

 

 

 

 

0.3

0.2«
a:
_
m satn

0.1-

0.0 .........

o 10 20 30 4o 50

Wt% Lactose

Figure 4.7. Graph of PIR (peak intensity ratio) vs. wt% lactose in solution at 20
degrees C using 1(1) ppm (1.9 x 10-4 M) pyranine. The PIR is defined as the ratio
of peak intensities at 440 nm and 51 1 nm. The dashed line represents concentration
at saturation.

91

 

 

 

 

200

8
2 -——II— TotalWater
O ——-o— BulkWater
a . —-l—-—- Salvation Water
In
0
O
'3
8
3 100 -
E
In
8.
0
'3
a
O
S

0

0

 

Wt. % Sucrose

Figure 4.8. Total. bulk and salvation water (number of molecules of water per molecule
of sucrose) vs. wt% sucrose in aqueous solution with 100 ppm (1.9 x 10-4 M) pyranine
at 20 degrees C. The values on the graph are obtained from an analysis using a two state
model for water. The vertical line represents saturation.

 

 

 

 

 

g 2..

O

5

m

‘5‘

a " --.-— BulkWater

8 —fi—— SalvatianWater

'5

S

r. 100-

8.

in

ii

3 J

“5‘

.9.

5

rug 0 ‘ l ' r_' l ' I ' ' b
g -r.o -0.8 -0.6 43.4 -o.2 0.0 0.2 0.4

Degree of Supersaturation

Figure 4.9. Graph of bulk water and salvation water per molecule of sucrose
in aqueous solutions doped with 100 ppm (1.9 x 10-4 M). pyranine at 20

de C versus the degree of supersaturation, DS. The DS is given by the

di erence between the solution concentration and the saturation concentration
at 20 degrees C divided by the saturation concentration at 20 degrees C.

 

 

 

 

 

100

8

G

8

1580‘

‘54

0

iso-

a

0

a40

h d

8.

0

gm.

‘5

2
o 'T'T'I'I"l'
o 10 20 so 40 so so

Wt. % Glucose

4.10. Total. bulk and salvation water (number of molecules of water per
molecule of glucose) vs. wt% glucose in ueous solution with 100 ppm
(1.9x 10-4M)pyranineat20degrees C. T e valuesonthe grc'gfhareobtained
from an analysrs using a two state model for water. The vertr line represents
saturation.

 

..s
.4
.r
‘

Molecule per molecule of Fructose

 

 

 

o 20 4o 60 so

Wt.% Fructose

Figure 4.11. Total. bulk and salvation water (number of molecules of water per
molecule of fructose) vs. wt% fructose in aqueous solution with 100 ppm

(1. 9x 104M) pyranineat20degreesC. Thevaluesonthegrczphareobtained
from an analysrs using a two state model for water The vertr line represents
saturation.

 

§

Total Water

 

—O— BulkWater
————a-——- Salvated Water

Molecule per Molecule of Lactose
5.3

 

 

 

 

Wt. % Lactose

Figure 4.12. Total. bulk and salvation water (number of molecules of water per
molecule of lactose) vs. wt% lactose in aqueous solution with 100 ppm

(1.9 x 10-4 M) pyranine at 20 degrees C. The values on the grc'aarh are obtained
from an analysis using a two state model for water. The vertr line represents
saturation.

 

 

 

1 00

8

8 —I— Salvation Water

2 80 . —.—- Bulk Water
(3 .
0- ---I--- N on-bulk-non-oolvahon water
0

'3

0 so -

.2

§

h 40 "

8.

0

g 20.

"e'

E

0 r

 

 

   

o '10'26'30'43‘50 60
Wt.% Glucose

Figure 4.13. Bulk water. salvation water, and non-bulk—non—solvation water
(number of molecules of water per molecule of glucose) vs. wt% glucose

in aqueous solution with 100 ppm (1.9 x 10-4 M) pyranine at 20 degrees C.
The values on the graph are obtained from an analysis using a three state
model f or water. The vertical line represents saturation.

 

 

 

 

 

200
g —5— Salvation Water
8 —-O— Bulk Water
:4 ——G—-—- Non-bulk-non-aolvation water
fi-I
O
0
‘3
‘8' 100i
O
S
I-r
8.
0
‘5' r
U
2
O
S

0 fi 1 - r ‘ r T ' T ‘
0 IO 20 30 40 50

Wt. % Lactose

Figure 4.14. Bulk water. salvation water, and non-bulk-non-solvatian water
(number of molecules of water per molecule of lactose) vs. wt% lactose

in aqueous solution with 100 ppm (1.9 x 10-4 M) pyranine at 20 degrees C.
The values on the graph are obtained from an analysis using a three state
model for water. The vertical line represents saturation.

Chapter Five

The PIR as a measure of deviation from Ideal Solution Behavior

5.1 Background

The focus on solution-structure has been primarily limited to the study of the
immediate salvation environment of the solute molecules. In many cases the structure
of this environment has been observed to change little. if at all. with solution
concentration up to saturation. even though deviations from ideal solution behavior
over the same range of solution concentrations are observed to be significant. As a
result. strong correlations between structure and deviations from ideal solution behavior
were not apparent A relationship between change in solution structure and deviations
from ideal solution behavior has been surmised although the specifics of cause-effect
are yet undetermined. There has been some speculation as to changes in the structure of
the solvent that does not participate in the salvation of solute with changes in
concentration; however. until recently there were no reliable estimates regarding the

nature of such changes.

In a study of structure in aqueous sucrose solutions we used the salvation

environment related changes in the excited state equilibrium of trace quantities of pyranine

(a fluorescent probe molecule extrinsic to the system) to determine the peak intensity ratio.

or PIR (the ratio of water involved in solvating sugar molecules to water to water that is

not). We found to our surprise that while the immediate salvation environment of the solute

molecule remained fairly constant even in supersaturated solutions; the amount of water per

solute molecule not involved in its salvation (designated bulk water in the case of sucrose)

changed almost exponentially as a function of solution concentration. In an extension of

this work, with Pan's measurements using time resolved fluorescence spectroscopy of

pyrene butyric acid

98

99
and pyranine. we observed that the structure of the bulk water changed with solution

concentration as well. The most marked changes were observed in supersaturated solution.
Consequently. we were encouraged to search for a possible relationship between deviation
from ideal solution behavior for a solution of a given concentration and either the ratio of
salvation to bulk water per molecule of solute, or the amount of bulk water per molecule of

solute.

The obvious choice as a measure of deviation from ideal solution behavior is the
activity coefficient or the activity. In the particular case of sugar water systems this is of
considerable practical significance because the correlation between shelf life, texture.
biological stability of foods. and water activity, aw, is well documented. Several
comprehensive reviews of the influence of water on chemical reactions in food are
available1'5. The measurement of aw, in concentrated and metastable aqueous solutions
of sugars is also an area of active interest6.

Rugged, stable. rapid and precise techniques for evaluation of aW are required;
as are techniques that can rapidly. accurately, and simultaneously measure solute
concentration and water activity. Official Methods of Analysis of the AOAC (1984)
32004-32009 details analytical methods for determination of aw. Hair hygrometers.
electronic hygrometers. and isopiestic methods are commonly used. Electronic
hygrometers work on the principle that that changes in conductance, resistivity.
impedance, resistance. or capacitance of a thin film or an aqueous solution of a
hygroscopic substance can be correlated to corresponding changes in aw7. In this work
We will demonstrate for aqueous sucrose and glucose solutions. that a single
measurement of the emission spectrum of a trace fluorescent probe in a solution sample
can be used in practice to obtain both concentration and aw. We will suggest that this is
POSSible because the same structure-interaction effects are responsible for changes in

both the water activity as well as the ratio of salvation to bulk water.

100
5.2 Materials and Methods

Reagent grade sucrose and reagent grade glucose were obtained from Sigma
Chenricals. and pyranine from Eastman Kodak. Water used was 18 Mohm reverse
osmosis water. Stock solutions of 100 ppm (1.9 x 10-4 M) pyranine were prepared by
measuring out 0.1 gms of pyranine to an accuracy of 1: 0.0002 gms and diluting with
0.001 m3 water. Sugar solutions of known wt % were nrade in 20 ml scintillation vials
by measuring out the required amount of sugar to 1 0.0002 gms. and adding the
required amount of stock solution. Sucrose solutions were obtained in 10 wt% intervals
upto 60 wt% and in 2 wt % intervals upto 74%. Glucose solutions were obtained in 5
wt% intervals upto 40% and at 2 or 3 wt % intervals upto 55 wt %. All concentrated
solutions were made by heating the mixture of sugar and pyranine doped water until the
sugar completely dissolved. Solutions were then slowly cooled back to 293 K.
Solutions were used within 6 to 24 hours to ensure equilibration. as well as avoid
biological contamination.

Emission spectra were obtained with either a 200 W Xe/I-Ig or a 150 W Xe
source in quartz cuvettes using an in-house constructed spectrometer described in detail
elsewhere (Mussel and Nocera9). The lamps were operated at 80% rated power. A
Hamamatsu R1104 photomultiplier tube (maintained typically at 50% maximum
Voltage) was used in conjunction with a preamplifier (maintained at a sensitivity of 10 '

6) and a Princeton Electronics lock-in amplifier (maintained at a sensitivity of 3V). The
Photomultiplier tube was maintained at 213 K to improve the signal to noise ratio. A
375 nm KV or a 370 nm 06 cutoff filter was used to filter out source excitation at 342
run. A 3 mm slit width was used and shutters were kept fully open. Upon obtaining the
enfission spectra of sucrose solutions doped with pyranine at 100 ppm (1.9 x 10-4 M)

val'ying in concentration from 0.5 wt% sucrose to 80 wt% sucrose. clearly a

101
contribution from molecules in each emitting form of pyranine was observed f or each

solution examined. Results are shown in Figure 5.1. Similar results were observed for

glucose solutions examined.

5.3 Results

The structure of pyranine is presented in Figure 5.2. The equilibrium behavior
of pyranine is described in Figure 5.3. In other work we have shown that the peak
intensity ratio (PIR). the ratio of peak intensity at 440 nm to that at 511 nm for a
solution excited at 342 nm, can be correlated to salute concentrations in concentrated
and supersaturated solutions of sugars3.

The water activity for aqueous sucrose was determined from the data of Stokes
and Robinson“). The xw'ln 7,, were plotted against experimentally obtained In PIR in
Figure 5.4. In each case the data fit well to a straight line. The activity coefficients for
aqueous glucose were obtained using UNIFAC 5. A reasonable linear fit is obtained for
a plot similar to that presented in Figure 5.4. It is interesting to note that the fit
improves considerably when a three state model for water is invoked and the xw*ln 7.,
is plotted vs the natural log of the ratio of chemically associated water to bulk water as
Opposed to PIR. Figure 5.5. The chemically associated water is obtained as a derived

quantity from experiment. by introducing a correction which is discussed in detail

elsewheres.

5“ Analysis

We will attempt to relate the dissociation equilbrium of the probe to the

sol vation of solute and dissociation equilibrium of water via the coupled species

102
equilibria. The various relations for a system of solute, Su, and water at a known

concentration, volume and temperature are presented briefly.
For a given unit volume V, and temperature T:
let there be
nw°=total molecules of water in pure solvent
nb= molecules in ”bound” state
nr= molecules in ”free” state
11 bulk = molecules in bulk water phase
n solvated: molecules in solvated phase
nw=molecules of water in solution
ns=molecules of solute in solution
now nw°=nb+nf ...... although nb and 11f need not be constant from moment to moment
or from one point in solution to another.
11 solvated: 6 ns 4» nb
1‘ bulk= nf
(although for most experiments we take n solvated= 6 as
n bulk: nr + no)
nw= n solvated + n bulk
PIR= n solvated/n bulk

Let us assume all the solute is solvated at any given moment.

Therefore the equilibria which affect the analysis are:

PyOH <=> PyO' + w ........................ ( 1)

and 2HzO + 2 (n'H20) <=> H+ + I-I(2n'+1)O'(nv+1) + OH' + H(2nv+1)O+(nr-1) ......... (2)
The solute molecules are entirely solvated

80 Su + 6 H20 => Su.6H20 ................ (3)

103
Let us recapitulate for the sake of clarity:

water involved in salvation of solute = 6 n3 ...............
water not involved in salvation of solute = nw - 6 n3 .............
So in a sense:
ns Su+ nw H20 <=> ns Su.6 H20 + (nw—6ns)*(2n'+ 2)'1[H+ + H(2nv+1)0(nv+1) +

OH' + H(2n'+1)0*(n'-1)]
and PIR = PyOI-I/ PyO' ............
Our argument has been that since there are on the order of 106 molecules of water for
every probe molecule. it is the behavior of this water that affects the local and global
instantaneous and time averaged equilibrium behavior of the probe. For this reason.
[PyO'l 0‘ [14*] + [H(2n'+1)0+(n'-1)] + [OH] + [14(2n'+1)0(n'+1)l
and [PyOI-I] or [Su.6 H20] + [I-IgO] + [n'HzO]
We find that when ns=0. in pure solvent, PIR is approximately zero. thus. for all
practical purposes we can assume that from the point of view of the probe molecule the
equilibrium of the water lies far to the right. Consequently:
[Pin or nw- 6ns and [PyOI-I] 0: 6ns,
so that PIRz ¢ 6ns/(nw-6ns) where o = a proportionality constant.
In PIR = In 615 + In ns - In ”available where ”available = nw - 6ns

0.0 d In PIR] d 11w = (nw + as), 115(1)“, ' 6113) .

Now xw=nw/(nw+ us)
so d xw = (nw—r- n5) dnw - nw d(ns+nw)/(nw + n,;,)2
= 1/(nw-r- ns) dnw also dxw=-dxs=-l/(nw+ us) (his
but, for the sucrose-water system In 1w = C1 xwxs2 and ln 1, = D1 xsxw2
so (I In Yo, / dxw = C1x32-2C1xwxs and d In y, / dxs = D1xw2-2D1xw
z -C1xs(2xw-xs)

so d In y../ d In PIR = -Crxs(2xw-xs)[xs(xw-6xs)l

104
= -C l [ZXWZXSZ' 13was3 6x84]

z ~C1[2xsz-17xs3+l 1x34]

as fairly constant because both xs and C1 are small
numbers and x32 does not change very much over the range of concentrations
considered.
and d xwln NJ (I In PIR = xw d In vw/d In PIR +ln 7w dde In PIR

= also fairly constant for the same reason.

5.5 Discussion

The PIR, and mol fraction. are both intensive state variables of the system.
They are directly obtained from measurements of the concentration of different species
in solution. The activity coefficient. gw. is not a directly measured variable; instead it is
obtained by estimating how much the contribution of the mol fraction of each species in
solution would have to change in order to correctly estimate a concentration dependent
system variable which can be directly measured, using a thermodynamic relationship
valid for an ideal solution. The deviation of the value of the activity coefficient from
unity is a measure of the deviation of the properties of the solution from those of an
ideal solution. To the extent that the actual value of the activity coefficient is related to
the values of system variables which are obtained from direct measurements, the
activity coefficient is also a measured quantity. The failure to reproduce any
experimentally observed correlation between such measured quantities via modelling or
argument must therefore be due to inherent flaws in the analysis or modelling and not in
the experimental observation (provided the experimental data are fairly accurate).

It is clear that we observe a linear correlation between PIR and xwlnyW (on

logarithmic coordinates). The analysis suggests this result. however. does not lead to it

105
in a direct and conclusive manner. A large part of this has to do with accuracy of the

model for the activity coefficient which is used to relate it to mol fractions. Structural
effects on solution properties are expressed via this model. The current practice in order
to form such a model is to start with an equation of state (of sorts) for the solution
where each of the coefficients represents the contribution of nearest neighbour
interactions. next nearest neighbor interactions, and so forth; where. ”nearest neighbor”
invariably refers to a molecule and ”interactions" refer to interactions between
molecules (on the scale of a molecule). Current practice is to consider each molecule of
a particular species exactly identical to every other molecule of that species. Functional
distribution of solvent cannot thus be accounted for, although strong correlations
between the deviations from ideal solution behavior and the change in microphase
distribution are experimentally observed. It is intuitively clear that both the change in
the PIR and the deviation from ideal solution behavior are both due to similar physica-
chemical reasons.

One of the ways to include the effects of structure would be to provide an
alternate model for the activity coefficient that takes into account the localized nan-
continuum nature of the interactions between the solvent and the solute. The basis for
such an approach lies in extending the principles of salvation thermodynamics as
developed by Ben Naim and colleauges. with the help of experiments and molecular
dynamics. This should allow the development of models which account for the
structural reasons for deviations from ideal solution behavior without having to account

for the exchange rates at each and every single atom of each molecular species

separately.

106

5.6 Conclusions

It has been shown that a single emission spectrum of a sample of aqueous
sucrose or glucose solution previously doped with pyranine can be analysed to yield aw
and concentration simultaneously because the salvation equilibria for all species in
solution are coupled via a mass balance on the solvent i.e. water. This analysis is
derived specifically for sucrose in water with pyranine as a probe. however, it may be
equally valid for other systems. The good correlation between derivation and

experiment point the way to the development of an optical sensor for water activity.

5.7 Notation

[A] I concentration of A, moles/ volume of solution (molality) [ML'3]
Kp I equilibrium constant for dissociation of pyranine. [dimensionless]
Ks I equilibrium constant for salvation of solute. [dimensionless]

KW I equilibrium constant for dissociation of water. [dimensionless]
aw I activity of water. [dimensionless]

xi I mole fraction of i,[dimensionless]

Yi I activity coefficient of i,[dimensionless]

R I gas constant

T I absolute temperature,K, [T]

M I mass. kg, or moles.[M]

L I length. meters. [L]

n(A) I males of A per unit volume of solution

A G8 I Gibbs' Free Energy of Salvation. such that A Gs/RT I [dimensionless]

107
5.8 References
l. Labuza T. P. 1980. The effect of water activity on rection kinetics of food
deterioration. Food Technol. 3 4(4):36

2. Rockland L. B. and Nishi, S. K. 1980. Influence of water activity on food product
quality and stability. Food Technol. 3 4(4):42

3. Duckworth, R. B. 1975. Water relations ofFoods. Academic Press, New York

4. Rockland, L. B. and Stewart. G. F. 1981. Water Activity: Influences on Food Quality.
Academic Press. New York

5. Simato D. and Multan. J. L. 1985. Properties of Water in Foods. Martinus Nijhoff
Publishers, Dordecht, The Netherlands

6. Le Maguer M. 1987. Mechanics and Influence of Water Binding on Water Activity. in
Water Activity : Theory and Applications to Foods, L. B. Rockland and L. R. Beauchat
eds. Marcel Dekker. Inc. , New York and Basel. pp. 1-25

7. Johnston MR. and RC. Lin. 1987. FDA Views on the Importance of aw in Good
Manufacturing Practice. in Water Activity .° Theory and Applications to Foods,

L. B. Rockland and L. R. Beauchat eds, Marcel Dekker, Inc. , New York and Basel, pp.
287-294

8. Chakraborty R and K. A. Berglund 1992. J. of Crystal Growth. submitted
9. Mussel. RD. and D.G. Nocera, J. Am. Chem. Soc., 1 10. 2764 (1988).

10. Robinson RA. and Stokes. R. H. 1965. Electrolyte Solutions, 2nd Edition (revised),
Butterworths, London.

108

 

 

 

 

 
  

 

'560‘ ' ' 'séo' 600
Wavelength

Figure 5.1. The emission spectra of pure water and sugar solutions of 20. 40. 60. and
70 wt.% sucrose doped with 100 ppm (1.9 x 10'4 M) pyranine at 293 K Emission

maximaoccurat440 nm. and 511 nm. An isobestic paint is observed. represented by the
pointerat485nm.

109

NaO 3 : SO3Na

Figure 5.1. Chemical structure of pyranine or 8-hydroxy-l , 3. 6-pyrenetrisulfonate

 

 

440nm
i.e,, .. to
450nm
PyOH"
A
hex =342nm
W
V
PyOH
A-a|,'350nm

110

511 nm
A'em " to
514 nm

 

 

— Pyo'*+ go"
A

Xex=342 um
W

 

 

V

 

 

* PyO- +119"
lab- 405nm

Figure 5.3. Emission equilibrium of pyranine. The equilibrium between the protonated
form (absorbs at 350 nm and emits at 440 run when excited at 342 nm) and the
deprotonated form (absorbs at 405 nm and emits at 511 nm when excited at 342 nm) is
determined by ability of the microenvironment of the hydroxyl group to participate in

proton transfer.

111

 

 

 

 

 

 

0.0
,E
5 -01 -
.?
-O.2 ' r r I
-4 -3 -2 -1 0 l
InPIR

Figure 5.4. Graph of ’Vv 1n Yw versus In PIR for aqueous sucrose
solutions at 293 K

112

 

 

 

 

 

-0.2 . - . . .
-1 o 1 2 3 4
In bulk water

 

salvation water

Figure 5.5 Graph of x” In 7w versus 1n Bulk water/Salvation Water (excluding
non-bulk-non-solvation water) for aqueous glucose solutions at 293 K

Chapter Six

The Study of Structure in Saturated and Supersaturated Sucrose Solutions
using Steady State and Time Resolved Fluorescence Spectroscopy

6.1 Background

Structure. transport properties and thermodynamics of phase equilibria are
manifestations of interactions on submolecular length scales. e. g. hydrogen
bonding, hydrophobic interactions, which require anywhere from a picosecond to a
nanosecond to take place. The fact that solution structure and transport phenomena
are related in a way that structure governs transport properties and transport
phenomena are a directly measurable manifestation of structure. enables us to extract
important information about the structure from the study of transport processes and
vice versa

Transport properties in solution are governed by non-equilibrium salvation
dynamics. They are governed by short range interaction processes that take place on a
variety of time and length scales, which are less than those associated with molecular
diffusion . Continuum models are still prevalent for such dynamics. These models do
justice to the long range interactions in solution. but fail to account for the details of
structure and interaction of inner salvation shells which are often dominant factors (1,2). In
the particular case of supersaturated solutions. a previous study of the steady state
fluorescence spectroscopy of pyranine in sucrose solutions of various concentrations

indicates the molar ratio of salvation water to bulk water per

 

‘Someofthismaterialwaspubliahedasapartasz.Pan,RChakraborty.&KA.Berglund.
"Steady State and Time Resolved Fluorescence Spectroscopy of Sucrose Solutions." Journal of Crystal
Growth. 1 30, pp. 587-599. (1993). Time Resolved Fluorescence Spectroscopy Data were collected and
analysed by B. Pan Steady State data were collected and analysed by this author. That part of the results and
discussion which this author contributed significantly to is included here. Some of Pan's data are mentioned
here in support of analysis presented. The reader is referred to the paper for a complete discussion of
materials. methods, and analysis.

113

114

male solute is such that there is only enough water available on the average for one
salvation shell or an incomplete salvation shell per molecule of solute (3). For this
reason. one would expect transport governing interactions in supersaturated solution
to be dominated by short range non-equilibrium solvent-solute interactions. The
absolute times required for these interactions to take place in supersaturated solutions
may be tens, hundreds, or even thousands of times longer than in dilute solution;
yet, in terms of distances involved, they are still short range.

There are now several non-continuum heuristic approaches to understanding
salvation dynamics. Even the salvation of a simple alcohol must be described by at
least two different time scales. For snurll dipoles or ions, in dilute solution. the time
required for solvent response is of the order of a picosecond (1,2). Furthermore, the
molecular dynamics simulations of Cross and Simon (4). suggests that if a dipole is
strong enough (>17 D) then the two ends may start behaving like point dipoles in the
sense that the localized solvent structure and dynamics may be different at each of
the ends of the dipole. This is summed up by Maroncelli et. al.(5) who conclude that
the salvation response in a polar solvent involves coupled reorientational dynamics
that occur on time scales shorter than single particle diffusion. There have been
studies in the salvation of an electron in water and alcohols, using picosecond and
femtosecond time correlated spectroscopies, which indicate that solvent
reorientational times are on the order of < 500 fs (6). Brady's study (7) of the
molecular dynamics of solvated glucose in a 15 wt.% solution also indicates that the
local structure and dynamics near each of the O or OH moieties of the glucose
molecule differ, as do the solvent relaxation times. In each of these studies in dilute
solution there is a large amount of non-salvation solvent Transport coefficients in
sucrose solutions of a wide range of concentrations have been reported in literature
by Stokes and Robinson (concentrated solutions(8)), and more recently by Sorrel

and Myerson (supersaturated(9)). In each case the transport coefficients in

115

concentrated. saturated. and supersaturated solutions are observed to be much
lower. One would expect that response and reorientation times would be much
greater as the amount of bulk water decreases in the system, or the solvaton to
solvaton distance decreases. however there are no experimental data in literature
which directly measure such parameters.

Previous investigations have studied the nature of these molecular
associations by utilizing a variety of means. In the study of supersaturated solution
structure many studies have been conducted on the scale of molecular diffusion.
Some of these include the column studies of Larson and Garside (10—12). the studies
of Narayan and Youngqrrist (13). and the studies of Myerson and co-workers (9.14—
19). Dielectric relaxation methods (20) have identified two distinct relaxation
processes for water in sucrose solutions. These were designated as associated water,
hydrating the sucrose. and bulk water comprising the remaining water in solution.
not associated with solute. Associated water has a longer relaxation time than the
bulk water. The 2H and 17O NMR studies of Richardson and Steinberg(21), on the
mobility of water in sucrose solutions, indicate three regions of decreasing rotational
rates of water molecules with increase of sucrose concentration through saturation
and continuing into the supersaturated regions. These effects were attributed to the
development of an extensive but dynamic network of hydrogen bonds of water to
sucrose. water bridging of the sucrose and direct sucrose to sucrose interactions
(although no direct evidence of sucrose-sucrose interactions were observed). The
bulk viscosities of these solutions also increase with increasing sucrose
concentrations. in a manner consistent with the rest of the data, suggesting that the
same short range interactions define both structure and transport

Techniques like x-ray diffraction (22) probe order on the right length scales
but may still be very noisy because many separate phenomena happen on similar

length/time scales and no single process can be distinctly identified. Vibrational

116

spectroscopy (non-time correlated), like IR or Rarnan spectroscopy (23. 2426)
probes faster time scales and smaller length scales than those involved, subsequently
important information is often impossible to extract or resolve because of poor
"signal to noise” ratios.

Both steady state and time resolved fluorescence spectroscopy of trace
fluorescent probes, like PBA and pyranine, yield data that directly measure transport
coefficients as well as data that can be indirectly related by means of correlations to
activity coefficients. Steady state fluorescence spectroscopy of pyranine is sensitive
to the time scale where the exchange of the hydroxyl proton of the pyranine occurs.
As the solution grows more concentrated the time required f or such an exchange
increases but the time required for exchange of the hydroxyl proton increases in the
same way. and the distance that is effectively probed remains the same.

Time resolved fluorescence techniques are able to resolve information on the
organization and dynamics of the intermolecular interactions in the
rrricroenvironment of a probe molecule (27). The advantage of time resolved
fluorescence spectroscopy is its ability to probe the salvation environment of the
entire probe molecule via measurement of lifetime and anisotropy. It is sensitive to
the longer times scales of molecular diffusion in the context of the microenvironnrent
of a probe molecule. In the nanosecond time resolved fluorescence experiment. we
use a nanosecond pulse of light to excite a sample of solution doped with a
fluorescent probe and detect the output emission. This signal, a waveform in the
form of an exponential decay. is fed to an oscilloscope (the triggering of which is
described later). The output is then analyzed to yield the emission lifetime and the
anisotropic rotational decay. Rotational correlation times can be subsequently
calculatedThe emission lifetime of the fluorescent probe is highly sensitive to its
salvation microenvironment and can provide information on the polarity of the

constituents of the solvent environment and the diffusive behavior of collisional

qrenche
about llr
ol the p

inrsllg

6.2 Er

6.2

PBA (
Which
highh
1l10-
long 1

of lor

“’ai'el

Min

117

quenchers. Measurement of the anisotropic rotational decay can provide information
about the microscopic transport properties and packing of the salvation environment
of the probe molecule. This work concerns the use of fluorescence techniques to

investigate microscopic transport in supersanrrated solutions.

6.2 Experimental Methods
6.2.1 Fluorescent probes:

For this study we chose two fluorescent probes. In addition to pyranine.
PBA (l-pyrene butyric acid. Figure 6.1) is a well characterized fluorescent probe
which has been extensively used as an oxygen sensor(27). The PBA molecule is
highly hydrophobic and readily forms excimers in water at concentrations above
1x10‘3 M. It has a blue emission with peaks at 380 nm and 398 nm and an unusually
long lifetime (~100nsec). The long lifetime makes this probe useful in measurements
of long rotational correlation times.

Pyranine (Figure 6.1). with its dissociable proton. is highly sensitive to
proton transfer processes and has been applied to the study of water content of
nricellar dispersions (Clement and Gould. (28); Kano and Fendler, (29); Kondo et
al.. (30); Bardez et al.. (31); and, Pines and Huppert, (32) ). sols (Kaufman eta1..
(33) ) and sol-gels (Pouxviel et.al., (34) ) among other systems. Because the
sodium ions of pyranine are believed to be completely dissociated in ionic solutions.
it is a highly soluble probe in water. The phatochenristry of pyranine has been
studied and is described elsewhere (Kondo, 35). The protonated forrrr emits with a
wavelength maxima at 440 run. while the dissociated form emits with a wavelength

maxima at 511 nm. Competition between these two forms determines the relative

118

intensities of the fluorescence maxima and influences the apparent fluorescence
lifetimes of each species. PBA and pyranine were obtained from Eastman Kodak

and were used without further purification.

6.2.2 Preparation of Solutions

Samples for the time resolved experiments were prepared in the following
manner: Ultrapure sucrose was obtained from Boehringer Mannhein. Sucrose
solutions were prepared at ~ 80°C by dilution of a 80 wt.% solution with either 10
mM sodium borate buffer at pH 8.5. for sufficient solubility of PBA, or 18 Mohm
reverse osmosis water. f or pyranine. to make 10 g samples with concentrations
ranging from 0 wt.% to 80 wt.%. The concentration of PBA used was 2x10“6
mole/Kg of solution. while the concentration of pyranine was 1x10'5 mole/Kg of
solution. The fluorescent probes were added to both the 80 wt. % solution and water
prior to dilution. Samples were then allowed to cool to 22 0C at least 3 hours and
were not deoxygenated prior to measurements. Suprasil Quartz cuvettes were used
to contain the samples during the measurements.

For the steady state fluorescence, and UV absorption experiments, reagent
grade sucrose from Sigma Chenricals. and 18 Mohm reverse osmosis water were
used. Stock solutions of 100 ppm (1.9 x 10'4 M) pyranine were prepared using 18
Mohm reverse osmosis water. Solutions of the 10 ppm (1.9 x 10'5 M) and 1 ppm
(1.9 x 10'6 M) were prepared by dilution. The amount of sugar and water required
for 10 grams (.01 kg) of each solution was calculated beforehand. Reagent grade
sucrose samples were weighed to :t 1 % of the required amount. Samples were
prepared in 20 ml glass vials by adding the predetermined amount of stock solution
to achieve the desired concentration. Samples varied in concentration between pure

water and 80 wt.% sucrose. In undersaturated solutions, samples were prepared in

119

10 wt.% intervals. Samples near saturation were taken at 2 wt.% intervals, eg. for
sucrose between 60 and 74 wt%. and solutions at high concentrations. eg. 76 wt%
and 80 wt.% solutions of sucrose. were also prepared. Supersaturated solutions
were prepared by heating the vials in a water bath until the sugar dissolved. Spectra
of the solutions were taken within 6 to 24 hours of preparation to ensure
equilibration of the sugar in solution and prevent biological contamination of the

sugar solutions.

6.3 Results and Discussion

6.3.1 Steady State Fluorescence Spectra and UV Absorption

Steady state fluorescence spectra of pyranine in sucrose solutions of various
concentrations have been presented elsewhere. UV absorption spectra of pyranine in
sucrose solutions are presented in Figure 6.2. These spectra were obtained in quartz
cuvettes of 10 mm path length at 298 K using a Perkin Elmer Lambda Array UV-Vis
spectrophotometer. These follow the same general trends as the steady state
fluorescence emission spectra (presented elsewhere) and exhibit an absorption
isosbestic point at 342 nm. Peaks at 350 nm and 400 nm correspond to absorption
of the protonated and deprotonated forms of pyranine. Note that the UV absorption
spectrum of pyranine in water follows the same general trends in profile as the
spectra of the pyranine doped solutions of sugar but exhibits an isobestic point at
325 nm instead of 342 nm. Two less concentration sensitive bands are observed at
280 nm and 220 nm respectively. The spectra in saturated sucrose solution and the
76 wt% solution show a peak at 400 nm and 525 nm respectively. The intensity of

the UV absorption peaks at 350 nm and 342 nm are a function of concentration

120

though not nearly as well behaved and sensitive to concentration as the fluorescence
spectra. It is worth noting that the UV absorption and steady state fluorescence
spectra are related in such a nranner as to suggest that fluorescence is observed from

single probe molecules and that no excimer formation is observed.

6.3.2 Time resolved Fluorescent Lifetimes

The total fluorescence lifetime measures the average time for a population of
probe molecules to return to the ground state from the photaexcited state. Total
fluorescence lifetimes of PBA. vary as a function of concentration, Figure 6.3. At
sucrose concentrations below 50%. the fluorescence decay at an emission
wavelength of 380 nm was adequately fit by a single exponential term. while two
exponential terms were required at higher concentrations. The value of the longer
lifetime is 100 ns in the buffer and increases to about 125 ns at 55 wt.%. Above this
concentration, the longer lifetime remains constant up to 80 wt.%. Values for the
short lifetime ranged from O to 40 us. The ratio of the pre—exponential terms for the
fast component to the long component is shown in Figure 6.4. Remaining at a
constant value of 0.1 from 50 wt.% to about 70 wt.% the ratio increases
dramatically near the saturation concentration. At 400 nm. as seen in Figures 6.5 and
6.6, the short component is present at all concentrations measured and shows a
higher intensity relative to the long component than at 380 nm. The same trend is
seen at both 380 nm and 400 nm as the sucrose concentration is increased.

The fluorescence lifetimes of pyranine in sucrose solutions displays greater
variation than the lifetimes of PBA as the concentration increases as depicted in
Figure 6.7. Three regions are again present for both wavelengths as the
concentration increases to 80 wt. % For the emission at 440 nm, two exponential

components are required to fit the total fluorescence decay for concentrations up to

121

40 wt.%. A low intensity long component (not shown). of about 10 ns. is present
along with the dominant short component. The pre-exponential factor for the low
intensity decay is less than approximately 2 % of the dominant decay. Because of the
relatively large intensity of the 511 nm band at these concentrations, the long lifetime
component is likely due to overlap from this emission. The fluorescence decay at
511 nm, attributed to the deprotonated form. can be fit by a single exponential term
throughout the entire range of concentrations. In the dilute solutions. the apparent
lifetimes of both species are relatively constant. with the protonated form exhibiting
a value of 0.5 ns and the deprotonated form exhibiting a value of 5.5 ns. At 40 wt.%
the lifetime at 440 nm begins to increase, while the lifetime at 511 nm shows a slight
decrease to 5.4 ns and remains relatively constant up to a concentration of 66 wt.%
In saturated solutions, the lifetime at 511 nm decreases sharply to a value of 4.4 ns
at 80 wt% The lifetime at 440 nm continues to increase less sharply and plateaus to
3.6 ns near 80%.

The increase in the lifetime of PBA is most likely due to the combined effects
of the decreased collisional quenching by oxygen and the decreased polarity of the
solvent Because of the long lifetime of PBA. its diffusional radius in the excited
state is also large. Interactions with collisional quenchers are possible during this
period which lead to a decrease in the fluorescence lifetime of PBA. At dilute
concentrations of sucrose, the lifetime may be shortened by collisions with dissolved
oxygen. As the solutions becomes more concentrated. the diffusion of both pyrene
and oxygen is hindered. causing an increase in the lifetime of PBA. Additionally.
PBA is known to have a longer lifetime in more hydrophobic solvents. due to an
increased rate of radiationless transitions from the excited state to the ground state
for rrrore polar solvents. The increase in the lifetime of PBA is most likely due to the
decreased diffusional radius of oxygen in the excited state of PBA and the increased
hydrophobicity of the solvent

122

The component that fluoresces with a lifetime of 0-40 ns in PBA doped
supersaturated solutions observed at 380 nm and 400 nm and pyranine doped
solutions observed at 440 nm could be due to some impurity in sucrose. This
impurity is probably natural or could occur as the result of the way the solutions
were prepared. It probably competes with the sucrose for water of salvation and
becomes more concentrated as the availability of bulk water decreases with increase
in concentration, contributing significantly to the fluorescence only in very
concentrated solutions. The fact that the contribution due to this component in dilute
solutions is low compared to the probe emission indicates that either the impurity has
a law quantum yield and/or the concentration of impurity in dilute solutions is very
low. Because of the high percentage of the total fluorescence that is attributed to this
impurity at high sucrose concentrations. it is likely that in these solutions it is present
in concentrations higher than that of the probe (which probably has a higher
quantum efficiency) which suggests that it may be a natural impurity or the result of
some reaction that takes place at 80°C involving the breakdown of sugar.

The changes in the fluorescent lifetimes of pyranine are probably caused by
changes in the rates of protonation and deprotonation. Pyranine is highly sensitive to
the proton accepting abilities of the solution. The various forms of the pyranine
probe as a proton donor in a two-state excited-state reaction are depicted in Figure
6.8. In bulk water. the protonated and the deprotonated ground state f arms of
pyranine exist at equilibrium with a pKa of 7.5. Upon excitation, the pKa for the
molecules in the excited state decreases to 0.5, causing a shift to the deprotonated
form The rates for this process in water have been reported to be 2.1x109 and about
1015 8'1 for the forward and reverse rates, respectively (Kondo. 35). Thus. the
excited pyranine probe exists almost completely in the deprotonated form in bulk
water. The observed fluorescence lifetime of each form of pyranine is dependent on

the intrinsic rate of excited state decay. ki, the rate of radiationless decay. kt. and the

123

rates of exchange between the protonated and deprotonated forms. The lifetime of
fluorescence errrission. excluding the effects of proton exchange. is related to the
intrinsic and radiationless decay rates by.

t=l=_1__

where kf= k5 + kr. A value of 1.6x108 8'1 for k; has been found and corresponds to
a lifetime of about 6 ns. This value agrees with the observed lifetime of the
deprotonated form and indicates that the reverse reaction is negligible at bulk water
conditions.

It has been argued that the relative increases of the steady state fluorescence
intensity of the protonated f arm as compared to the deprotonated f arm, due to
solvent effects, can be attributed to decreases in k1 (Kondo. 35). Otherwise, the
unlikely possibility that the intrinsic fluorescence lifetime must decrease. in order to
account for this effect, must occur. Thus, the observed increase in the apparent
lifetinre at440 nm can be attributed to a decreased rate of deprotonation. Similarly.
the decrease in the apparent lifetime at 511 nm indicates an increase in the rate of
protonation. Together. these rates determine the equilibrium observed in the steady
state spectra of pyranine. The rate effects are related to the availability of salvation
water and its distribution in sucrose solutions. Physically, the proton transfer
processes can be related to the associations of the water molecules in solution. A
water molecule not associated with solute is able to accept protons from donor
molecules owing to its ability to transfer the protons to other free water molecules in
proximity. However. if all of the tetrahedral hydrogen bonding sites of water are
occupied, proton exchange cannot occur. The time scales of the molecular motions

and associations are the determining factors governing the specific behavior of these

124

exchange effects. When water is involved in the salvation of sucrose. interacting
indirectly or directly. its mobility may be hindered by the sucrose molecule and
cause the subsequent decrease in the rate of proton transfer. This leads to the view
that the protonated farm of pyranine interacts with the water involved with the
salvation of sucrose. while the deprotonated form interacts with the bulk water.
From PIR ratios of the steady state fluorescence spectra, the ratio of the
salvation and bulk water molecules decreases as the sucrose concentration increases
and at saturation. there is one mole of salvation water for every male of bulk water.
Furthermore, the behavior of the fluorescent lifetime of pyranine can provide
dynamic information on the structure of sucrose solutions. In the dilute region the
relatively constant lifetimes of both the protonated and deprotonated species indicate
that the environment of pyranine does not change appreciably. As the concentration
is increased in the 40 wt% to 66 wt.% range. the behavior of the lifetimes suggests
that the water associated with sucrose becomes increasingly less mobile. while the
bulk water experiences a decrease in its mobility at 40 wt.% but remains relatively
mobile up to 66 wt%. At saturation, the bulk water begins to experience additional

effects of the sucrose by becoming less mobile.

6.3.3 Time Resolved Fluorescence Anisotropy

The rotational correlation time of the probe in solution is a measure of the
time required for reorientation of its excited state dipole moment and is related to its
rotational diffusion. Another indication of the nature of short range interactions
between the probe and its salvation environment is found in the nature of the
relationship between the rotational correlation time and viscosity. Rotational
correlation times of PBA in sucrose solutions are shown in Figure 6.9. It is apparent

thatthere are three regions as the concentration varies from 0 wt.% to 80 wt.%. In

125

the first region, with sucrose concentrations between 0 wt.% to about 40 wt.% the
measured rotational correlation times are nearly instantaneous relative to values at
higher concentrations. The accurate determination of these values are beyond the
limits of the instrumentation used. A more apparent increase occurs above about 40-
45 wt% up to about the saturation concentration (67 wt.%), in the second region. In
the third region, above about 66 wt.%, two components in the anisotropic decay
become present The fast component increases linearly as the concentration increases
from a value near zero up to about 10 ns. In contrast. a dramatic increase of almost
three orders of magnitude is seen in the long component. Figure 6.10 shows the
dependence of the rotational correlation time on viscosity up to 74 wt. % sucrose.
The short rotational correlation time increases only slightly as the viscosity
increases. A linear relation between the rotational correlation time and viscosity is
seen for the long component The slope of the linear fit is 6.2x10'2 ns/cP with an
intercept at 0.94 ns.

Another depiction of the three regions of concentration dependent behavior is
illustrated in Figure 6.11 with the rotational correlation time of pyranine. The
rotational correlation times of the protonated and deprotonated form of pyranine are,
within experimental error. the same in the entire concentration range. In the low
concentration range. below 40 wt.%. the rotational correlation time is nearly zero.
Above this concentration. the rotational correlation time increases as the
concentration increases. However, the increase above the saturation concentration is
less than that with PBA. This is seen more clearly in Figure 6.12. where the
rotational correlation time is plotted against viscosity. Below a concentration of
about 66 wt.%, the rotational correlation time increases linearly in relation to the
viscosity with a slope of 1.5x10'l ns/cP for both the protonated form and the

deprotonated form The intercepts occur at 1.7 ns and 0.6 ns. for the protonated

126

form and deprotonated form, respectively. These values clearly level of f . however.
as the solution becomes supersaturated up to 74 wt.% sucrose.

The bulk viscosity shows three characteristic regions of structural
organization. The relationship between rotational correlation time and viscosity is
linear for PBA. In the case of pyranine. this relationship is nonlinear above 66 wt.%
sucrose. This behavior may be interpreted in terms of the difference in molecular
structure and geometry. and the difference in the way each probe interacts with its
solvent environment. Both nonsphericity of the probe molecule and solvent
interactions contribute to the rotational behavior of the probe.

PBA can form hydrogen bonds only at the carboxyl group of is side chain
and has minimal interactions with solvent. However, pyranine has up to three
negatively charged sulfonate groups directly on the pyrene which can exert a greater
effect on its salvation shell. A molecule of PBA in aqueous solution can be
approximated by an oblate spheroid within its salvation shell. Because of the
additional polar groups on pyranine. however, this molecule may exist in a more
spherical cavity and thus rotate at a greater rate. From the slopes of the rotational
correlation time versus the viscosity the diameter of the spherical rotational cavity for
PBA is found to be 8 A. This corresponds well to the actual 7A value for the largest
diameter of a pyrene group. At concentrations below 60 wt.% sucrose. the diameter
for the spherical rotation of pyranine is found to be slightly larger at 11 A. This may
reflect the slightly larger size or greater charge interactions of the pyranine molecule
in solution. However, neither molecule is strongly solvated by water. because of the
similarity of the hydrodynamic diameters to actual molecular dimensions. Above
saturation. the rotational correlation rate f or pyranine does not vary linearly with
viscosity and partial slip boundary conditions are observed. This may be interpreted.
in view of the shape and free space arguments, as an increased ordering of the

solvent shell around the pyranine molecule. The ordered shell provides a cavity

127

within the solution in which pyranine may rotate with a decreased degree of solvent
displacement. In this ordered shell, the electropositive hydrogen atoms of the solvent
would interact with the negatively charged groups allowing additional rotational
movement with minimal disruption of the hydrogen bonding organization in the
surrounding fluid. As the pyranine probe rotates, charges on it may be relayed
among the hydrogen atoms, in a nranner similar to the transfer of protons in bulk
solution. This notion is supported by NMR studies. showing that the transverse
relaxation rates for 2H are an order of magnitude faster than the relaxation rate than
that of 170. Likewise, the linear relationship observed with PBA may be explained
by the nature of its rotational cavity. A possibility may be that the salvation
environment of the PBA molecule is less polar and more homogeneously distributed
about the pyrene molecule. and does not provide for the formation of a spherical
cavity. Because the rotational cavity of PBA may be considered more oblate than that
of the pyranine molecule. out of plane rotations would necessitate a greater degree of
solvent reorientation in the immediate salvation nricroenvironment of the probe.

This analysis is particularly interesting because it calls upon a macroscopic
hypothesis to explain a microscopic situation. An important issue is why this kind of
reasoning should apply at all. One reason could be that. at these concentrations, the
fluid environment of the probe is homogeneous. So the crux of the issue is the
structure of the solvent environment of each of these probes. At concentrations of
sucrose below saturation, the rotation of small probe molecules are essentially
similar. However. a marked divergence is seen at saturation indicating a fundamental
structural change in the structural organization of the solution. Regardless of the
specific explanations for increased rotational mobility. it is clear that solute mobility

in sucrose solutions decreases as the concentration is increased.

128

6.4 Structural Characteristics of Sucrose Solutions from Transport

Properties

The solution structure of dilute solutions up to a concentration of about 35-
45 wt.% has been described in terms of an isotropic two state model (e. g. (21)) in
which the rotational mobility of water is a weighted average of the rotational
mobilities of salvation water and bulk water. Anomalous behavior has been
observed in solution between 30 and 40 wt.% in FTIR and Raman spectra and X-
ray diffusion as well. The increase in the fluorescence lifetime of PBA at around 40
wt.%, is likely explained by the same underlying cause. The PIR calculated from
steady state fluorescence spectra of pyranine. the fluorescence lifetimes of the
protonated and deprotonated farms of pyranine and the rotational correlation times of
PBA and pyranine exhibit relatively small changes in this region. Thus, although the
solution becomes more concentrated and more water is bound to the sucrose. the
behavior of the rotational correlation times indicates that the structure of the bulk
phase water does not change appreciably from that of pure water.

As the concentration increases from 40 % to about 66 % sucrose, the
behavior of the fluorescence properties of the probes indicates a nonlinear structural
transformation where the solution becomes less fluid. This structural transformation
has been characterized in NMR and X-ray diffraction studies as the formation of a
network type structure of hydrogen bonding throughout the solution. These
hydrogen bonds have been postulated to be comprised of the water-water
interactions. hydrogen bond bridging between sucrose molecules and sucrose-
sucrose interaction. However, it is not possible to correlate these microscopic
interactions to specific locations relative to the sucrose molecule. Because of the

almost identical rotational correlation times measured f or the protonated and

129

deprotonated forms of pyranine. the mobilities of the salvation water and bulk water
appear to behave similarly at the nanosecond time scales of these measurements.
From this. it can be inferred that these solutions are isotropic at this time scale so that
sucrose-sucrose interactions if they occur at all are minimal. and the associated water
is not locally "bound" to sucrose. When in addition the steady state fluorescence data
are taken into account. it is clear that sucrose-sucrose interactions in these solutions
are unlikely to occur because on the average the sucrose molecules appear to be
completely solvated and there is still enough bulk water in solution to achieve the
complete salvation of all sucrose molecules. The average salvation
microenvironments of both protonated and deprotonated f arms of the pyranine
molecule is indicated to be roughly the same although the local salvation
environment around the probe hydroxyl moiety differs.

This region corresponds to the point where the number of molecules of bulk
water per molecule of sucrose decreases from about 40 to about 6. This is also the
region in which the sharpest changes in IR and Raman spectra are observed. A
molecule of solute in 35 - 45 wt.% solution of sucrose is effectively still in a dilute
solution because there are enough water molecules around it so that this water is still
not influenced by the effect of nearby sucrose molecules. As the number of solute
molecules increases. first, the effect of the water molecules dynamically associated
to salute and then, that of water more strongly associated. exerts itself on water
molecules not participating in the salvation of sucrose. Over a relatively small
concentration range. the number of molecules of bulk water drops exponentially.
This should be reflected by a sharp decrease in the rotational correlation time of the
probe however no such sharp decreases are observed. Most of the molecules of
water solvating the probe in solutions at these concentrations are already partially
associated, either strongly or dynamically, to the solute molecules. Transport in this

concentration range changes over from that governed essentially by the interaction of

130

probe with bulk water, to that resulting from the interaction of the probe with at least
partially solute associated water. It is likely that in this region the changes observed
are due to changes in structure associated with the salvation water.

Near the saturation point. another structural transformation is evident from
the sharp decrease in the lifetime of the deprotonated pyranine, the rapid increase in
the rotational correlation time of PBA and the relative leveling of rotational
correlation time of pyranine. Additionally. the PIR indicates that the molar ratio of
bulk to salvation water per molecule of sucrose is approximately unity. At this point.
it is likely that the structure of the bulk water phase is considerably different from
that of pure water. There are not enough water molecules per molecule of salute to
support the variety of flickering clusters that are found in pure liquid water under
identical conditions. The attractive forces between water molecules may take
precedence over the attractive forces between the deprotonated pyranine and water
for reasons of structural stability. It could be that there is enough bulk water to
dissociate a few pyranine molecules but not for as long a time as in dilute solutions.
because it is more important for the water to form clusters with other bulk water
molecules. As the bulk water continues to decrease in quantity and change in
structure and average coordination, the effects of molecular structure and salvation
environment on rotational correlation time and viscosity become more pronounced.

The most surprising and interesting observation, however. concerns the
implication of the sharp drop in the lifetime of the deprotonated pyranine on the
dynamics of the bulk water. The UV-vis absorption and the steady state fluorescence
spectra indicate that the absorption or emission is due to photon transfer between the
same excited states: i.e.; the the decrease in lifetime can be attributed to a decrease in
the rate of radiative transfer as opposed to a decrease in the rate of non-radiative
transfer. The behavior of the fluorescent lifetime. remaining more or less constant up

to the saturation point, indicates that the ”lifetime" (the interaction time) of the bulk

131

water up to the saturation point remains more or less the same despite the change in
structural organization. At saturation. the sudden drop in the lifetime of deprotonated
pyranine must also reflect an equivalent change in the "lifetime" of the bulk water.
The bulk water is not only reduced in quantity and changed in structure. it is also
available to pyranine for less time which means it must spend time trying to find
other water molecules to coordinate with and/or moving faster to satisfy all its needs
i.e.. solvating deprotonated pyranine. finding other water molecules to coordinate
with. coordinating with other water molecules, maintaining structural integrity of the
solution etc.

One might object to the idea that bulk water moves faster in supersaturated
solutions than in undersaturated ones because it seems contrary to intuition.
However. when the energetics of the formation of a supersaturated solution are
considered. this notion begins to make sense. It is impossible to make a phase
transition from undersaturated to supersaturated solution without the addition of
large amounts of external energy. This energy serves to increase the disorder of the
solution and also increase the probability that enough bulk water molecules will be
present in the vicinity of all the unsolvated solute molecules (one at a time) to solvate
them (one at a time). The most common way of accomplishing this is to heat the
solution until all the solute dissolves. Then as the solution cools. the solvated solute
molecules transfer their kT energy to the surroundings, and become less mobile.
(accounting for most of the cooling effect since these contain most of the mass) and
the bulk water loses sorrre kT energy as well. Under identical conditions of pressure
and temperature. a supersaturated solution contains more energy per unit mass and
volume than a saturated solution. In addition. the energy in a supersaturated solution
is probably partitioned differently. If we compare a supersaturated solution with a
saturated solution, the solvated solute in the supersaturated solution is much less

mobile. It contains most of the potential energy of the solution but very little of the

132

translational energy. The difference in energy between the two solutions cannot be
accounted for in terms of potential effects alone because there is an additional large
energy banier that must be overcome in order to reach supersaturation. What then is
responsible for keeping the supersaturated solution at a higher energy? One factor
might be that the bulk water moves faster relative to solvated water in supersaturated
solution than in a saturated solution; even though an an absolute basis. both bulk
and salvation water are less mobile.

The transport of solute and solvent in supersaturated solutions are not
necessarily coupled in the same manner as in dilute solution. The solute is solvated
and the time correlated fluorescence anisotropy measurements in supersaturated
solutions indicate that the salvation microenvironment of the probe molecule behaves
as if it is a homogeneous mixture of solvatons (solvated solute molecules). We were
thus, at the time this work was done, led to believe that in supersaturated solution
the diffusional motion of the solute is closely coupled to that of the salvation shell
surrounding it. That is to say the solute molecule is not likely to diffuse
independently of a salvation shell on the time scale required for diffusion. The
solvent. in addition to the option of being able to diffuse through the solution as bulk
solvent. or with the solute as a part of the salvation shell. also has the option of
being transported through exchange processes. These exchanges may be slower in
supersaturated solution than in dilute solution. but are still faster on a relative basis
than diffusion processes. Therefore it may be possible for water not to build up
locally due to release of solvent or to associate locally over short periods of time if
transported through solvent exchange. The solvent exchanged at one point in
solution is not the same molecule that participates in an exchange process elsewhere,

but it doesn't have to be so long as local requirements f or order are met.

133

6.5 Self-Diffusion and Structure

In retrospect the increase in viscosity. and decrease in diffusion coefficients
take on a different significance. Pan's data indicate that both the salvation water and
the non-salvation water exhibit decreased mobility and increased correlation times in
supersaturated solutions. In dilute solution where the salvation water is not strongly
associated with the solute is somewhat more appropriate to describe diffusion in
terms of solute and solvent. than in non-dilute solutions. When the salvation water
mobility is such that it becomes strongly associated with the solute it is necessary to
consider a solvaton as the appropriate unit to describe diffusion. As the solute
concentration increases in solution the average distance between solvatons
decreases. and the amount of total non-salvation water exponentially decreases.
Once the concentration is such that the solvatons are likely to solvate the non-
solvation water. the solvatons are the primary source of transport resistance.

It is useful to consider the resistance to mass transfer during self -diffusion of
a species in solution in terms of the time required for that species to encounter a
salvation environment identical to its initial salvation environment. The surface area
of the sphere with a radius equal to the distance a solvaton must travel from its initial
position to encounter an identical salvation environment divided by the time required
for it to travel that distance represents the coefficient of self—diffusion for the
solvaton. The viscosity can then be thought of as the ratio of the force of resistance
experienced by the solvaton in traveling this distance to the diffusion coefficient. As
the radius of the sphere approaches zero, so does the coefficient of self diffusion for
the solvaton. while the viscosity of the solution approaches some large number. In
dilute solution where the distance between solvatons (which are themselves more

loosely associated) is large and the mobility is high. diffusion coefficients are fairly

134

high, and viscosity is fairly low. As the solution becomes more concentrated the
radius of the sphere decreases as does the mobility of both the solvatons and the
bulk water. and the time required to diffuse increases, consequently the diffusion
coefficient decreases and the viscosity increases. While this is true in general of non
electrolytes, in the case of electrolytes the mobility of the hydrated ions will
generally remain high as the time required to travel between identical salvation
environments decreases. Therefore the self-diffusion coefficients in electrolyte
solutions may appear to increase up to the saturation point. after which the radius of
the sphere decreases so much as to cause the diffusion coefficient to drop
precipitously. In supersaturated solutions of non-electrolytes where the salvation
environment of the solvatons are fairly identical from point to point. the distance the
solvaton must travel to experience an identical salvation environment approaches
zero. and so does the coefficient of self-diffusion. The measurements of Sorrel and
Myerson (9,14). and Myerson and Chang (15-18) indicate that this is indeed the

case for many systems. Pan's observations also lend support to this interpretation.

135

6.6 Conclusions

We have demonstrated that fluorescence techniques provide a means by
which transport and microenvironmental properties can be measured and
consequently. the molecular structure of solutions can be inf erred. As a result of the
time resolved measurements with PBA and pyranine in sucrose solutions. further
evidence is available suggesting that the structure and composition of bulk water
changes with an increase in sucrose concentration from a nearly pure water like
structure to a very different structure in supersaturated solutions. Information on the
localized structure can be extracted from this data and incorporated with information
obtained from other means. As the solution becomes more concentrated, the
structure of the water transforms from a relatively unperturbed form to one in which
a hydrogen bonding network between dispersed sucrose molecules are bridged by
the salvation water with minimal effect on the bulk water. In supersaturated
solutions, however. the remaining bulk water also becomes associated to the sucrose
forming a third type of solvent interactions. The water in supersaturated solutions is
likely to exist in tetrahedrally coordinated farm for much shorter times than those
observed bulk water at saturation.

Analysis of the difference in the nature of the correlation between rotational
correlation time and viscosity for the two different probe molecules confirms the
importance of understanding the short range interactions that govern transport In
concentrated sucrose solutions. it may prove useful to describe the solution as a
homogeneous network of solvated solute molecules connected by more mobile and
dynamic bulk solvent molecules. In the case of water as a solvent. it may be more
important for structural integrity f or the bulk water to f arm stable coordinated

structures than to make itself available for salvation of probe molecules.

136

We saw evidence of both bulk and salvation water in solutions of all
concentrations examined. No indication of microheterogeneities in solution
microstructure was observed. The solutions appear to isotropic and thus. no
apparent microscopic clustering of the sucrose molecule are present in
noncrystallizing solutions. On the basis of our investigation of sucrose solutions
using steady state and time resolved fluorescence spectroscopy of fluorescent probes
we can conclude that it is unlikely that there are solvent exclusive clusters of solute

molecules in the metastable supersaturated sucrose solutions.

6.7 References

1. Wolynes P.G., Annu. Rev. Phys. Chem. 3 1, 345. (1980).

2. Wolynes P.G., J. Chem. Phys. 86(9), 5133, (1987).

3. Chakraborty R. and K. A. Berglund, AIChE Symposium Series no. 284. 114. (1991).

4. Cross. A. J. and JD. Simon. J. Chem. Phys. 86, 7079, (1987).

5. Maroncelli M., J. Maclnnis. and GR Fleming. Science. 243, 1674. (1989).

6. Migus A., Y. Gauduel, J.L. Martin. and A. Antonetti, Phys. Rev. Lett. . 58(17), 1559.
(1987).

7. Brady J.W.. J. Am. Chem. Soc., 1 1, 5155. (1989).

8. Stokes, RH. and RA. Robinson, J. Phys. Chem. 7 6. 2126 (1966).

9. Sorrel, LS. and AS. Myerson. AIChE J., 28. 772 (1982).

10. Larson. M.A. and J. Garside. Chem. Eng. Science. 4 l. 1285 (1986).

11. Larson, M.A. and J. Garside. J. Crystal Growth, 7 6. 88 (1986).

12. Larson. M.A., AICHE Symposium Series no.240. 80. 39—44 (1984).

13. Narayanan, H. and GR Youngquist, AICHE Symposium Series no. 253. 83.1-7 (1987).

14. Myerson, AS. and LS. Sorrel. AICHEJournal, 28,778 (1982).

137

15. Myerson. AS. and Y.C. Chang, in Industrial Crystallization '84. SJ. Jancic and E.J. DeJong
(Eds). Elsevier, Amsterdam. 27 (1984).

19. Ginde R and A. S. Myerson. AIChE Symposium Series no. 284, 124, (1991).

16. Chang, Y.C. and AS. Myerson, AIChE J., 3 1, 980 (1985).

17. Chang, Y.C. and AS. Myerson, AIChE J., 32, 1567 (1986).

18. Chang. Y.C. and AS. Myerson. AIChE J., 32. 1747,( 1986).

20. Suggett, A. and AH. Clark. J. Solution Chem. 5, 1, 1 (1976).

21.Richardson SJ. and MP. Steinberg. in Water Activity: Theory a_rrd Applications toFood, L.
B. Rockland and LR Beauchat eds, Marcel Dekker Inc., New York. 235. (1987).

22. Mathlouthi M., Carbohydrate Res. 9 1. 113. (1981).

23. Back D.M., D.F. Michalska, and P. L. Polavarapu, Appl. Spectroscopy, 38(2), 173.
(1984).

24. Mathlouthi M and D.V. Luu, Carbohydrate Res, 7 8. 225. (1980).

25. Mathlouthi M. and D.V. Luu, Carbohydrate Res, 8 1, 203, (1980).

26. Mathlouthi. M., C. Luu, M. Meffrov-Biget and D.V. Luu, Carbohydrate Res. 8 1. 213
(1980).

27. Kalyanasundaram K. Photochemist_ry in Microheterogeneous Systems
Academic Press Inc., (1987).

28. Clement, NR and M. Gould, Biochemistry. 20, 1534 (1981).

29. Kano. K and J.H. Fendler, Biochim. Biophys. Acta, 509, 289 (1978).

30. Kondo. H.. I. Miwa and J. Sunamoto, J. Phys. Chem, 86, 4826 (1982).

31. Bardez, E.. B.T. Gouguillon, E. Keh and B. Valeur, J. Phys. Chem. 88, 1909. (1984).

32.Pines. E and D. Huppert, J. Phys. Chem. 87. 4471 (1984).

33. Kaufman. V.R. D. Avnir, E. Pines-Rojanski and D. Huppert, J. Non-Crystalline Solids. 99.
397 (1988).

34. Pouxviel, J.C., B. Dunn and J.I. Zink. J. Phys. Chem, 93, 2134 (1989).

35. Kondo, H., I. Miwa and J. Sunamoto, J. Phys. Chem, 86. 4826 (1982).

Figure 6.1. Structures of PBA (top) and pyranine (bottom).

139

 

Satd Su

 

 

Absorbance
(units vary with concentration)

 

 

 

20% Su
0.20“
0.10 . u . . . r . . .
225 250 275 300 325 350
Wavelength nm

Figure 6.2 UV-vis spectra of 10 ppm pyranine in aqueous sucrose solutions
of various concentrations at 293 K.

Fluorescent Lifetime (ns)

 

140

 

 

 

 

 

60..
J
40-
4x A A
A
204 A MA%
. A A A
o s-A
0 10 20 3O 4O 50 60 7O 80
Wt.%Sucrose

6.3. Fluorescent lifetimes of PBA in sucrose solutions observed at
380 nm and 293K as a function of sucrose concentration. Squares ( E] )
represent the long lifetime component and triangles (A ) represent the
short lifetime component.

.0
ca

141

 

.0 .0 .0
U" 05 V
1 . 1 . 1 .

Intensrty Ratio of Long
Component to Short Component
0
A

 

 

 

.50....60....70 ...80

Wt.% Sucrose

Figure 6.4. Ratio of the intensities of the long lifetime component to the short
lifetime component of PBA at 380 nm

142

 

 

 

Fluorescent Lifetime (ns)
8

 

 

Wt.% Sucrose

6.5. Fluorescent lifetimes of PBA in sucrose solutions observed at
400 nm and 293K as a function of sucrose concentration. Squares ( El )
represent the long lifetime component and triangles (A ) represent the
short lifetime component.

143

 

Intensrty Ratio of Long
Component to Short Component

 

 

 

 

0.0 ‘7' ‘V 'II 'U'TYY rv It. 'I111 '1'71" '[V'UT'TT‘ '

010 20 304050607080
Wt.% Sucrose

Figure 6.6. Ratio of the intensities of the long lifetime component to the short
lifetime component of PBA at 400 nm.

144

 

 

 

 

 

6;
A " A A A AAAA
a 5‘
2;
4-
g I.
.2
Ofl I
I-l 37 II-
; .
0
a 2- -
3 1- '
.—
OITVV 1" riT'j firrvr l l I

 

 

Wt. % Sucrose

Figure 6.7. Fluorescent lifetimes of pyranine in sucrose solutions at
293 K. Squares (Cl ) represent the emission at 440 nm and triangles

( A) represent the emission at 511 nm. Error bars indicate the standard
deviation of the measurements.

145

pKa = 0.5

 

 

 

 

k -
Py0H* ‘ 1 - PyO *+ H
k-r
kf 440 nm 511 mn k'f
355 nm 355 nm
V _ V
pKa — 7.5 _
PyOH ‘ ‘ PyO + H

 

 

Figure 6.8. Kinetic Diagram for the two-state excited-state proton transfer of pyranine.

Rotational Correlation Time (us)

146

600

 

500 -

4oo -

 

 

 

0 l 0 20 30 40 50 60 70 80
Wt.% Sucrose

Figure 6.9. Dependence of the rotational correlation time of PBA on sucrose
concentration. Squares (El) represent the rotational correlation time for the
long lifetime component and triangles (A ) represent the rotational correlation
time for the short lifetime component.

Rotational Correlation Time (ns)

147

 

 

 

 

 

100
80‘
60'
40-
20"
i A A
0 “MW...“ --
0 500 1000 l 500 2000
Viscosity (cP)

Figure 6.10. Dependence of the rotational correlation time of PBA on viscosity.
Squares (Cl ) represent the rotational correlation times for the long lifetime
component and triangles (A ) represent the rotational correlation times for the
short lifetime component.

148

120

 

é

Rotational Correlation Time (ns)

 

 

 

 

o 10 20 30 4o ”s'o'meo 7o 80
Wt.% Sucrose

Figure 6.1 1. Dependence of the rotational correlation time of pyranine on sucrose

concentration. Squares (El )represent the rotational correlation time observed at

:40 nm and triangles ( A) represent the rotational correlation time observed at
11 nm.

149

80

 

601

 

    

 

Rotational Correlation Time (ns)
N a.
o o

 

 

 

o ' 1560' I ' ‘10'00' ' ' '15'00' ' ' 'zooo
Viscosity (cP)

Figure 6.12. Dependence of the rotational correlation time of pyranine on viscosity.
Squares ( El) represent the rotational correlation times observed at 440 nm and
triangles (A ) represent the rotational correlation times observed at 511 nm Error
bars indicate the standard deviation of the measurements.

Chapter Seven

Solvent Imaging Studies of Supersaturated Sucrose Solutions in the
Absence of Crystals and of the Near Interfacial Region of Growing and

Dissolving Sucrose Crystals'

7.1 Background

The spatial distribution and transport characteristics of solvent and solute at or near
the crystal-solution interface is a subject of much speculation. There is little experimental
data related to the study of this region. Current speculations regarding structure at the
crystal-solution interface propose the presence of solvent-exclusive solute aggregates, or
solvated species that differ in degree of solvation from those found in bulk solution. Direct
experimental verification of such structures has not been possible because of inherent
difficulties associated with simultaneous measurement of concentration and interactions
between each of the species in solution at various points along and near the interface. It is
also difficult to correctly identify the moving interfacial boundary. Consequently, the few
attempts to conduct measurements under such conditions have thus far yielded
inconclusive, and sometimes contradictory, results.

The object of this work has been to develop new solvent sensitive experimental
procedures that enable the study of solution structure. In previously reported work the
steady state fluorescent behavior of pyranine was used for determination of solution
supersaturation in various sugar solutions. This measurement was accomplished by relating
the ratio of peak intensities due to the fluorescent states of pyranine to the amount of water

present in the immediate vicinity

 

' Part of this work appeared as R. Chakraborty, M. H. Wade, & K. A. Berglund, ”Imaging
Concentration Gradients Near Growing and Dissolving Sucrose Crystals,” pre-print, in Separations
Conference Symposia, AIChE annual meeting, November 1992., Florida Confocal images were provided
by Dr. M. H Wade of Meridian Instruments, Okemos.

150

151

of the hydroxyl moiety of the probe molecule and the exchange behavior of this water. It
was convenient to describe water in the solution as solvation water, bulk water and non-
bulk-non-solvation water (12). It became clear that the functional nricrophase distribution
of the various species is an important indicator of solution structure. In subsequent work,
time resolved fluorescence spectroscopy of pyrene butyric acid and pyranine indicated the
absence of solvent exclusive aggregates in metastable solutions and also that the structure
changed with increasing supersaturation. The issue of whether or not the probe is

homogeneously distributed was unresolved (g).

7.2 Materials and Methods

Reagent grade sucrose from Sigma Chemicals, 8-hydroxy 1,3,6, pyrenetrisulfonate
(pyranine) from Eastman Kodak, and 18 Mohm reverse osmosis water were used. Stock
solutions of 100 ppm (1.9 x 10‘4 M) pyranine were prepared using 18 Mohm reverse
osmosis water. Solutions of the 10 ppm (1.9 x 10'5 M) pyranine and 1 ppm (1.9 x 10‘6
M) pyranine were prepared by dilution.

The amount of sucrose and water required for 10 grams (0.01 kilogram) of each
solution was calculated. Reagent grade sucrose samples were carefully measured out to +/-
1% of the required amount. Samples were prepared in 20 ml glass vials by adding the
predetermined amount of stock pyranine solution to achieve the desired concentration.
Samples varied in concentration between 60 and 80 wt. % sucrose. In undersaturated
solutions samples were prepared at 5 wt.% intervals. Supersaturated solutions were
prepared at 2 wt.% intervals for sucrose between 68 and 80 wt.%.. Supersaturated
solutions were prepared by heating the vials in a water bath until the sugar dissolved.
Spectra of the solutions were taken within 6 to 24 hours of preparation to ensure
equilibration of the sucrose in solution [with respect to temperature] and prevent biological

contamination of the sugar solutions.

152

Pyranine doped sugar solutions - with pyranine concentrations of 1 ppm (1.9 x 10‘
6 M), 10 ppm (1.9 x 10'5 M), and 100 ppm (1.9 x 10'4 M)- exhibited little variation in
integrated half peak intensity ratio (IIR). The IIR is defined as the ratio of the integrated
intensity ofthe peakobservedat440nmfrom370 nmto445nmto that of the peak at 511
nm from 485 nm to 507 nm when excited at 351 nm. This is not the same as the Peak
Intensity Ratio or PIR defined as the ratio of the intensity of the peak at 440 nm to that of
the peak at 511 nm (excited at 343 nm +/- 5 nm). The IIR may vary depending upon the set
up of the ACAS and is obtained from a working curve obtained at the time the experiments
are conducted. It was decided to work at 100 ppm (1.9 x 104 M) levels of pyranine to
maximize the signal to noise ratio. Small changes in instrument settings did not affect the
values obtained for the IIR. All experiments discussed in this paper were conducted at 20
0C.

The steady state technique can be easily adapted to an imaging system to study
supersaturation in the presence of crystals. The ACAS-570 confocal system consists of a
high quality optical light microscope with a confocal imaging system, a travelling stage,
laser excitation using a 12 W Krypton laser and multi-line UV, and two 2-dimensiona1
detectors which detect integrated fluorescence intensities at each "point” in solution, one
from 400-445 nm, and one from 485-510 nm. The ACAS 570 interactive laser cytometer
was used with an excitation of 351-360 nm, a laser power of 60 mW, 50% scan strength,
60% PMT on each detector, scan speed of 2 mm/s or less, and steps of 2 microns or .4
microns. A 10% neutral density filter, a 445 nm dichroic, and a 485 nm (22 nm band pass)
filter were used. All spectra were obtained using an oil objective. The solution sample is
placed in a sample holder which is then secured on the stage. When crystals are used, hot
undersaturated solution is gently dropped on top of the crystal to dissolve crystal dust and
also to ensure that the crystal adheres to the bottom of the sample holder. The scan area and

resolution as well as the scan speed are specified. The half micron layer of solution nearest

153

the bottom of the sample holder is scanned. When high resolution ( 0.5 microns or less) is
used it is necessary to focus on secondary nuclei or crystals formed therefrom

Sucrose crystals were placed in a compartment of the sample holder. Five milliliters
of solution were then gently introduced on top of the crystal. This, in effect, glued the
crystal down. The solution was initially hot and slightly undersaturated. The sample holder
was then placed in the slot on the moving stage and secured in place. An edge of an ink
spot that had been placed on the bottom of the holder for the purpose of locating the plane
corresponding to the bottom was then located and focused upon. The solutions were open
to the atmosphere and nuclei were normally observed. Most of the data were obtained from
observing well developed nuclei because it was most convenient and expedient to do so.
The stage was moved at speeds of 2 mm/s or less to minimize motion in the solution. There
was probably some agitation; however, the crystals appeared to remain stationary during
the course of an experiment. The ratio analysis and dual detector data analysis routines on
the Data Amlysis System (DASY) were used to analyze the data

7 .3 Results

The data obtained from the detectors consists of two dimensional plots of integrated
half-peak intensities. First a standard set of solutions of concentrations ranging from 60
wt.% sucrose to 76 wt.% sucrose were imaged in dual detector mode. Images for
solutions of 60, 74 arm 76 wt.% sucrose are presented in Figure 7.1. Data obtained from
solutions in the absence of crystals indicate an isotropic distribution of pyranine in solution
in both excited states. This is consistent with the time resolved experiments which also
indicate an isotropic distribution of probe molecules. This data was used to generate a plot
of [IR vs sucrose concentration (wt.% sucrose), Figure 7.2.

In the presence of small growing crystals, gradients in concentration and intensity

were observed in the distributions of both the protonated and the deprotonated species of

154

the pyranine molecule. In each case the emission intensity of probe nearest the interface
was lower than that observed in bulk solution. Our experiments with pyranine thus far
indicate that the probe does not tend to incorporate itself into the sucrose crystal; this is
supported by the fact that pyranine and sucrose are not likely to be attracted to each other.
These observations together indicate the presence of less probe immediately near the
interface than in bulk solution. Samples of crystals imaged in dual detector mode and the
same crystals imaged in concentration ratio modes are presented in Figures 7 .4, 7.5, and
7.6. A crystal in 68 wt.% sucrose solution is seen in Figure 7 .3. Several different crystals
observed growing in 71 wt% sucrose solution are presented in Figure 7.4. A crystal
observed growing in 76 wt.% sucrose solution is presented in Figure 7.5. The
concentration ratio images indicate a 2-10 micron region immediately adjacent to the crystal
surface that is more depleted of solvent than the rest of the boundary layer. Under the
optical microscope as well as in the fluorescence image both the crystal and the interface
appear to remain significantly stationary during the experiment. This data appears to lend
credence to the existence of an "intermediate phase” as proposed by Powers (4), and
supported by Clontz and McCabe (é), Mullin and Leci (6), and Berglund (Z) among others.
This data reinforces the idea that the 1-5 micron region near the crystal interface is solvent
depleted, however, it does not suggest the existence of solvent exclusive clusters. This
does not necessarily rule out the existence of solvent exclusive clusters becuase the
structure in the less than one micron region is entirely unknown.

The extent of the boundary layer referred to here is determined by that distance from
the crystal solution interface at which solution concentration is equal to the bulk solution
concentration The extent of the boundary layer was observed to be dependent on the size
of the crystal rather than on the supersaturation. The smallest crystals observed were about
a micron in size and had boundary layers that extended up to 10 microns into solution.
Consequently the distance between data points was specified as 0.4 microns and smaller

areas were scanned. Larger crystals that were 10-100 microns in length had larger

155

boundary layers extending several hundreds of microns into solution. As a result larger
distances were specified between data points. In all cases the solution concentration reaches
bulk solution value within about 10,000 microns from the crystal solution interface in a
relatively quiescent solution. On the time scale of the experiment, prior experience led us to
believe that the bulk concentration should not decrease by more than 0.5 wt.% due to
crystal growth. This was observed to be true in each case.

Line scans taken in dual detector and concentration mode through the interfacial
region indicate a fairly sharp transition in concentration in the 4 to 10 micron region,
followed by a more gradual change in concentration out to the bulk solution, Figure 7.6. A
great deal of noise can be observed in the data, particularly that in the blue (445 nm). This
is probably because the ACAS-570 has detectors which are green and red biased. The ratio
of bulk water to solvation water in the boundary layer differs significantly from that in bulk
solution: proportionally more bulk water and less solvation water was observed than in
bulk solution.

As expected, data taken in the presence of dissolving crystals indicate that the solute
concentration nearthe interface is greater than that in the bulk, Figure 7.7. For dissolving
crystals only concentration gradients were observed. We did not observe any gradients in
probe intensity. Transport of solute seems to proceed down a concentration gradient from
the interface out to the bulk

Dual detector and concentration images can be obtained along vertical sections of
the solutions to obtain three dimensional information related to concentration and solvent
distribution. Figures 7.8 and 7.9 respectively present the dual detector and ratio images
obtained by using confocal spectroscopy to observe a sucrose crystal in a 74 wt.% sucrose
solution. Vertical sections are taken at 0.4 micron intervals and, the horizontal resolution is

0.2 micron x 0.2 micron.

7 .4 Discussion

156

The data clearly indicates a marked difference between solvent distribution and,
consequently, the structure of bulk solution, and solvent distribution and solution structure
near the crystal-solution interface of growing crystals. Not only does the ratio of bulk water
to free water differ on the average; it differs spatially across regions that differ in their size
and extent. The total amount of water per molecule of solute appears to vary spatially in the
presence of a growing crystal; there appear to be discrete spatial regions with water
concentrations that indicate supersaturated solution separated by other discrete spatial
regions with water concentrations indicative of undersaturated solution, so that on the
average the solution concentration within the boundary layer is richer in solvent and poorer
in solute than in bulk solution.

Initially we thought differences in the emission intensity of the probe from point to
point in solution indicated less water”. both solvation water and bulk water. Our observation
of the behavior of water and how it distributes in supersaturated solutions in the absence of
growing crystals indicates that the solvent distributes into the bulk phase only after all the
solute has been solvated, even if this is accompanied by changes in the structure,
composition, and dynamics of the bulk water phase. Such observations were not real time
observations. While we can not rule out real time related differences we can say with
certainty that lower emission density is indicative of fewer molecules of probe per unit
volume of solution surrounding it The local concentration is still accurately represented by
the PIR (or the IIR in this case) because it is that quantity which is a measure of the nature
of the solvation environment of the probe. Therefore a decrease in the emission intensity
accompanied by an increase in the PIR or IIR indicates the presence of a greater number of
solvated solute molecules. A decrease in the emission intensity accompanied by a
corresponding decrease in the PIR or IIR indicates the presence of more bulk water in the

solvation environment of the probe than in bulk solution.

157

On this basis our data suggests a "relatively” solute depleted region, extending 2-10
microns from the interface in which the solute is completely solvated. This region is
adjacent to a ”concentration boundary layer, the extent of which appears to depend on the
size of the crystal, where there more bulk water is observed than in bulk solution. Bennema
(8), Burton, Cabrera, and Frank (2), and Chemov (19) in their work on modelling crystal
growth and the solution surrounding the growing crystal, each consider a boundary layer
of unstirred solution between the supersaturated bulk solution and the solid-solution
interface. Mullin (11) estimated the diffusion boundary layer to be about 10 microns away
from the interface into solution. Since we define the extent of the boundary layer by the
distance at which bulk solution concentration is observed, we measured larger layers. We
should also point out that our layers are not completely unstirred. The ”concentration
boundary layer” is likely to occur as a direct result of crystal growth, and the release of
solvation water. A supersaturation dependence on the extent of the boundary layer was not
observed, this may be because we were unable to observe the full extent of the boundary
layer in some cases. The diffusivity of the bulk supersaturated solution is low, and the
viscosity is high; also, it is probable that water will self-associate and form bulk water like
structures if it builds up because of transport resistance. Such build ups may lead to
localized density differences and convection involving pockets of solution of slightly
different concentrations. This would be consistent with the irregular "periodicity” observed
in the near interface x-ray diffraction data of Mathlouthi (1_2) for sugars. It may also
account for some of the patchiness observed in our images in concentration ratio mode
which can't be explained entirely by system noise caused by the red and green bias of the
detectors. At the same time we are unable to categorically rule out the possibility of
existence of pre-crystalline clusters or other solute rich structures in the near interface
region because we have thus far been able to investigate only a limited region. The structure

of the less than 1 micron region near the interface is still completely unknown.

158

The issue of solution microstructure near the interface is still an open one. We see
no evidence to suggest the presence of solvent exclusive clusters of a lifetime greater than
the timescale of the experiment, or of a size greater than the resolution of the experiment
However, we are unable to conclusively rule out the existence of solvent exclusive
structures at this juncture because we do not have reliable information regarding the spatial

distribution of the solute molecules and their interactions.

7. 5 Crystal Growth

Partly as a result of observation and partly as a result of conjecture it is this author's
suggestion that the growth of a crystal is best described in at least 2 stages: initial growth,
and subsequent growth. The initial grth period follows some impact type catastrophic
“local” event that affects the structure and organization of those solvatons which are in
immediate contact with the event Such an event may be due to a heterogeneous nucleus,
the introduction of a seed, or mechanical or thermal disruption. It is possible that the
mechanism of initial grth differs according to the origin of the nuclei or seed crystal. The
introduction of a seed or heterogeneous nucleus puts a demand on solvent. Solvent is
required for wetting or solvation of those moieties of the solid in contact with solution
along the interface. Because the self diffusion coefficients are precipitously low, this
solvent must necessarily be supplied from the solution in contact with the solid interface.
As a result some of the solute is likely to be at least partially desolvated at least part of the
time. Depending on the chemical nature of the substance some or all of this solute may
crystallize quickly. During the initial grth period an abundant supply of solute is
immediately available to the nucleus or seed and transport of solvatons from bulk solution
is not likely to be a grth limiting factor. During initial growth the availability of water for

wetting is a primary factor.

159

For some systems the primary source of resistance may be related to how fast the
solvatons can reach the interface or volume diffusion. For other systems like sucrose-water
in which the chemical affinity of the solute for itself is not particularly high, the main
resistance to growth may be related to how quickly the partially solvated solute leaves the
solution. Once the solute is anchored to the crystal (to any extent) it may still be partially
solvated at other points on the molecule but is no longer mobile, because unlike the water
the solute crystal is immobile and the exchange rate is zero. Partially solvated solute in
solution phase may be displaced due to contact or fluid shcar or otherwise yield nuclei.

Once the initial grth period has ended, the crystal continues to grow. The water
of solvation released from the crystallized solute at some point will be in excess of the
water required to wet all the solute at the crystal-solution interface locally. Such water
diffuses into the supersaturated solution and builds up a region of less concentrated
solution adjacent to the crystal. Eventually bulk concentration gradients prevail and
convection is likely to occur, bringing the surface of the crystal in contact with
supersaturated solution of bulk concentration once again. It seems reasonable that the
structure near the interface of the growing crystal changes with time, and the crystal grows

in spurts.

7.6 Conclusion

The single most important conclusion that can be drawn as a result of this and
related work regarding the study of solution structure in supersaturated sugar solutions is
that in order to explain crystal growth and nucleation related phenomena it may not be
sufficient to treat the solution as a continuum. How the solvent is spatially and functionally
distributed and structured plays an important role in determining the properties and
behavior of solution. A material balance treatment of crystal growth or dissolution is

incomplete without an account of the solvent microphase distribution. The imaging of

160

pyranine in a sense allows us to determine the localized rrricrophase distribution of water in
solution. The role played by bulk water and solvation water, how the relative amounts
change both globally and locally, and the significance of such phenomena in crystal growth

and dissolution is not yet understood.

7 .7 Acknowledgements

We would like to thank the Crop Bioprocessing Center/ Research Excellence Fund
at Michigan State University for use of facilities and funding for RC. We are grateful to
Dr. Troscoe, of the Department of Pediatrics for use of the DASY, and to Dr. Asmina H.
Jiwa for set-up and assistance on the confocal ACAS-570 at Meridian Instruments.

7 .8 References

1. Chakraborty R. and K.A. Berglund, AIChE Symposium Series, no 284, 113, (1991).

2. Chakraborty R and K. A. Berglund, Journal ofCrystal Growth, 125, pp. 81-96
(1992).

3. Pan B., R Chakraborty and K A. Berglund, Journal ofCrystal Growth, 130, pp.
587-599, (1993).

4. Powers HE. C., Nature, 17 8, pp 139-140, (1956)

5. Clontz NA. and W.L. McCabe, Chem. Engg. Progress Symp. Ser., 67(6), 6-17,
(1971)

6. Mullin J.W. and CL. Leci, Phil. Mag., 1 9. Pp 1076-1077, (1969)

7. Beglund KA., PhD. Dissertation, Iowa State University, (1981).

8. Bennema P., Theory and Experiment for Crystal Growth from Solutions, in J.W.
Mullin ed. Industrial Crystallization, Plenum Press, New York, NY, (1976)

9. Burgon, W.K., N. Cabrera, and EC. Frank, Phil Trans. Roy. Soc. London, 243, 299,
(19 l)

10. Chemov A.A., Soviet Phys. USP, 4, 116, (1951)

ll. Mullin J.W., Crystallization, Butterworths, London, (1972).

12. Mathlouthi M., Carbohydrate Res., 9 1, 113, (1981)

161

    
   

f
I???
cw“

ugg'em'riifi .

Figure 7.1. Dual detector images for aqueous sucrose solutions of 60, 74 and 76 wt% at
293 K. Detector 1 images the half peak intensity of pyranine from 485 nm to 507 nm.

Detector 2 images the half peak intensity of pyranine from 445 nm to 370 nm One pixel
represents an area of solution 0.4 m x 0.4 pm.

162

 

 

1.00W
3.95-
0.96-
8.85-
0.80-
6.754
0.70-

0.55-

Ratio vs Wt.% Sucrose

 

 

0.50 . I
as 52

I II

I ' l' I 'l ' II I ll

64 55 58 7B 72 74 7B 78 8%

Ratio vs.

amount (wt%. Su)

 

 

Figure 7.2. A plot of [IR (integrated emission intensity) of pyranine vs sucrose
concentration (wt.% sucrose) at 293 K Color values represent concentration in weight

percent sucrose.

Values

..-9a
.22: 7 78'
if-rz
i-rs
5-7a
cu-rr
-7a

..- 71

 

83
75
50
25
00
75
SB

3 25
¢—-7a
g-sa
{-51
- 65.
:I-sr
g~sa
f-BL
-sa
(mr.s

00
75
50
25
00
75
SB
25
00
U)

163

Color Values

        

   

T?‘ 4095
\Detector 1 Data 3339 Detector 2 Data
3583
3327
3071
2815
2559
2303
2047
1791
1535
1279
1023
757
Values

..-saoa

-7a75

775a

7&25

7sao

7175

7a50

7L25

7aoo

5&75

5750

5&25

sass

5175

sass

BLZS

-B&00

un%.8u)

Figure 7.3 Dual detector (a) and concentration ratio mode (b) images of a sucrose cystal
growing in 68 wt% solution.

164

Color Values
:3— 4095
3839
3583
3327
3071
2815
2559
2303
2047
1791
1535
1279
— 1023
- 7B7
- 511

Detector 1 Data Detector 2 Data

       

255
mu '- 4695
- 3339 ‘ Detector 2 Data
— 3533 -
- 3327 _
— 3071 -

- 2815 -
h ‘- 2559 "' 77.;
- 2303 —‘
2047 —
1731 —
1535 -
1279 -
1023 -

- 757 -

Detector 1 Data

 

Figure 7.4 Dual detector mode images of several different sucrose crystals growing in 71
wt.% solution.

165

Values

80.00
78.75
77.50
78.25
75.00
73. 75
72.50
71.25
70.00
88.75
87.50
88.25
85.00
83.75
82.50
81.25
80.00
(wt%. Su)

 

.- 00.00
— 70.75
L" - 77.50
- 75.25
- 75.00
- 73.75
- 72.50
-' - 71.25

   

(wt%. Su)

Figure 7.5 Concentration ratio mode images of several'different‘sucrose crystals growing
in 71 wt% solution.

166

Color Values
-.- 4095
3839
3583
3327
3071
2815
2559
2303
2047
- 1791
- 1535
— 1279
— 1023

- 787

l

Detector 1 Data Detector 2 Data

  
 
 
 
 
 
 
 
  

l

     

(wt%. Su)

Figure 7.6 Dual detector (a) and concentration ratio mode (b) images of a sucrose cystal
growing in 76 wt.% solution.

167

 

Values (wt%. Su)

 

0 40 80 120

 

 

 

 

Values (wt%. Su)

 

0 51015202530

 

 

Values wt%. Su
80- ( )

78-
75~
74~
72-
70-
BB'WTUWWTH
0 4 B 12 16 20 24

 

 

 

 

 

Figure 7.7. Samples of line scans in concentration ratio mode taken in the interfacial region
of growing sucrose crystals at 293 K. The big dip in concentration indicates the interface of

the crystal.

168

Color Values
- 4095 — -
3939 -
— 3593 —
3327 -
; 3071 —_
- 2915 -‘
2559 -‘j
2303 -_f
2047 —?L;
1791 - 4
1535 -‘.
1279
1023
757
511 —
255 -
0 -“

   

Detector 1 Data Detector 2 Data

       

 

Calcium (wt%. Su)

 

 

0 2000 4000 6000

 

 

 

Figure 7.8 A sample of a line scan taken near the interface of a dissolving sucrose crystal in
60 wt% sucrose solution at 293 K. The big dip in concentration indicates the interface of
the crystal.

 

169

 

Figure 7.9 Dual detector images of a sucrose crystal in 74 wt. % solution at 293 K taken
using confocal microscopy in 0.4 um vertical steps, and 0.2 um x 0.2. 14m horizontal
resolution. Instrument settings for this data differ from those of the rest of the data in this
work.

_A /——.

170

1.1.
r‘ 913

3.303
1. 7 '-

 

Figure 7.10. Concentration ratio images of the sucrose crystal imaged in Figure 7.9.

Chapter Eight

Summary, Conclusions, & Recommendations for Further Work

8.1 Summary

This work can be conveniently summarized in two parts for the purposes of
discussion. The first part deals with issues related to the relationship between the solvent
and solution structure in general and how this relationship changes with the solute and
solvent concentration in solution in particular. The remainder deals with the attempt to
extend these ideas and experimental techniques to the study of the solution structure in the

near interface region of growing sucrose crystals.

8.1.1 Issues Related To Solution Structure

Short Range Order and Long Range Order

Solution structure is often described in literature in terms of short range
order and long range order. Short range order, hence referred to as SRO, is the result of
submolecular interactions which take place on relatively small space scales. In dilute
solutions those interactions are in effect on time scales of the order of less than one
picosecond and length scales of a few nanometers. In supersaturated solutionstime scales
are on the order of 100's of nanoseconds although length scales are still a few nanometers.
SRO determines the dynamics of a system. System dynamics determine rates and
mechanisms of physical transport properties and rates and mechanisms of chemical
reactions.

Long range order, hence referred to as LRO, describes long-time and time-averaged
structure which results from the organization of groups of molecules on the scale of tens of

nanometers or more. Long range order is sometimes described to be a result of long range

171

172

or long term interactions between molecules. This is somewhat misleading because first of
all the molecules themselves do not interact, it is the submolecular moieties on the
molecules which are affected by forces of attraction and repulsion exerted by submolecular
moieties of other molecules. The forces between atoms that make up these moieties can be
expected to follow Heitler London behavior; all forces of interaction dying off within 2
atomic diameters. Such forces are inherently short range and short term. These short range
forces acting on different parts of a molecule simultaneously can result in the formation of
multi-molecular groups. In fact this is why the first multi-nrolecular groups thought to exist
for water were dimers and trimers. These multi-molecular groups appear to exist for long
periods of time when one views the time-averaged ”snapshot” of spatial distribution of the
components of the system. They may remain constant in composition in terms of the
number and relative spatial orientation of molecules of each of the molecular species
involved, even though different molecules of each species may occupy those same relative
positions at different moments in time. There are short time periods, determined by SRO,
when an exchange of molecule occurs at each of those positions, during which those
positions may not be occupied by any molecule, and during which some other group may
come into existence. In a mixture of two or more molecular species LRO describes all the
different types of time-averaged structures, their composition, and relative spatial
orientations. It is possible for a system to exhibit LRO on more than one spatial or temporal
scale of order.

It is clear that LRO depends intimately on SRO, and that the description of
the structure of a system is incomplete without an understanding of both. It is this author's
hypothesis that there is a definite relationship for a system given a particular temperature
and pressure and composition and number of macroscopic phases which defines the
number of scales of order on which a system under consideration exhibits LRO of
significance. The number of such scales is likely a function of the number of chemically

distinct molecular species in the system; the number of distinct relaxation times for each of

 

173

those species, as deternrined by the chemical structure of each of those species, system
temperature and pressure, and the SRO; the number of relaxation times of significance with
respect to the mechanism of the phenomena being observed in that system; and the number

of molecules of each species in the system.

8.1.2 SRO and LRO in The Aqueous Sucrose System

Let us examine the sucrose water system from this point of view. There are
two chemically distinct molecular species, sucrose and water. Water has two distinct
relaxation times of significance, hence two solvation states. There are three different types
of possible interactions: solute-solute, solute-solvent, and solvent-solvent. Also the
cherrrical affinity of water for water is greater than the chemical affinity of water for
sucrose, which in turn is greater than the chemical affinity of sucrose for sucrose. Unless
the solvent concentration is such that enough solvent is not available to solvate all solute
molecules to the same extent that they are solvated in dilute solution, solute-solute
interactions are unlikely. For the sucrose water system, data generated in this work
indicates that such a situation is not observed below solution concentrations of about 76 %

by weight sucrose.

LRO in Dilute Solutions

In dilute aqueous sucrose solutions there are four scales of LRO of
significance. The first is related to the structure in bulk solvent, far from the solvated solute
and corresponds to the length and time scales associated with flickering clusters. The
second is related to the solvated solute and corresponds to the length and time scales related
to the solvatorr a single solute molecule and the water of solvation most strongly associated
with it The third is related to the solvation environment of the solvaton and corresponds to

the length and time scales associated with the structure of the bulk solvent immediately

174

surrounding the solvaton. And the fourth, to the structure of the solution on the scale of
many solvated solvatons in bulk solvent and corresponds to the length and time scales

associated with diffusion. A conceptual diagram of the dilute aqueous sucrose solution is

presented in Figure 8.1.
Complexities that may arise in other systems due to other relaxation mters

Dilute Aqueous Glucose solutions

For the sucrose water system the solvation environment of the solvaton in
dilute solution is likely very similar to the structure of bulk water since the behavior of the
solvent in the system is adequately described by two exchange rates for water. However in
other systems, such as the glucose-water system, this environment is likely to be
significantly different, due to the presence of significant amounts of non-bulk-non-
solvation water per molecule of solute. The molecular dynamics simulations of Brady and
the data analysis performed in this study both suggest that some water in the glucose water
system is associated with the solute on time scales which are significantly faster than
solvation water but significantly slower than bulk water. It is reasonable to assume that this
water is associated with the solvation environment of the solvaton, and that it is more
structured than bulk water. One possible reason could be that the solvent clusters are more
tetrahedral and less flickering in nature.

Dilute Fructose and Lactose solutions

The description of relaxation states in dilute solution is more complex yet
for aqueous fructose and lactose solutions. Whereas, dilute aqueous solutions of sucrose
and glucose are fairly easy to characterize, in the absence of some working empirical

correlation, these systems are not Fructose, for example is known to exhibit rather

175

complex enatiomeric equilibrium in addition to the ability to adopt different conformations.
Such complexities may require additional parameters for description, and hence increase the
number of relaxation states needed to adequately characterize SRO. Subsequently the
system is likely to exhibit LRO of significance on a greater number of spatial and temporal
scales. Such a characterization is not possible using the experimental techniques or methods
of analysis developed in this work, since the experiments indicate how water is distributed
into one of two states with relaxation rates relative to the rate of proton exchange.
However, data obtained using these methods may provide a bare framework against which
to corroborate data obtained using other methods more inherently suited to the measurement
of relaxation times. In any case, these limitations do not affect the characterization of
structure or concentration measurements in supersaturated sucrose solutions: the focus of

this work.

SRO and LRO in Concentrated Aqueous Sucrose Solutions

As the solution becomes more concentrated in solute, the bulk water per
molecule of solute exponentially decreases. Many changes in properties are observed for
sucrose solutions as the sucrose concentration approaches 40 wt. %. If one assumes that
flickering clusters are on the average tetrahedral in structure and that a solvaton is on the
average solvated by six such clusters, it is interesting to note that 40 wt% corresponds to
the concentration where there is only enough water available to provide six tetrahedral
water clusters (which still behave as bulk water) for every solvaton. This seems to indicate
a change in SRO of the bulk water. At this point the solution is far from ideal, and no
longer dilute. So, in terms of LRO for non-dilute solutions there appears to be one less
scale of significance to consider. A conceptual representation of structure in concentrated

sucrose solutions is presented in Figure 8.2.

 

176
Saturation described in terms of SRO LRO and Phase Inversion Analogy

 

With the addition of more sucrose, the bulk water per molecule continues to
exponentially decrease in availability until a concentration of 66.7 wt.% sucrose or
saturation is reached. Near this point the number of bulk water per molecule of sucrose
becomes equal to the number of solvation water, and at saturation the number of solvatons
equals the number of tetrahedral clusters that could be formed by the bulk water available
per molecule of sucrose (or per solvaton). Proceeding from dilute solutions to saturation,
thus corresponds to a precipitous drop in bulk water, corresponding to a sharp change in
the SRO of the bulk water, and duly evidenced by marked changes in transport properties.
This is consistent with the fact that transport properties are governed by SRO.

In terms of LRO saturation is somewhat akin to a phase inversion point
because at non—dilute concentrations less than saturation the solution can be viewed on
some scale as solvatons being solvated by flickering clusters which are tetrahedral on the
average (PC or TC), and for concentrations greater than saturation the solution can be
viewed on the same scale as a solution of FC solvated by solvatons. A conceptual
representation of structure in saturated aqueous sucrose solutions is presented in Figure
8.3.

Such an analogy holds true for aqueous glucose solutions, but not for
aqueous fructose or lactose solutions. The PIR is still a steep function of solvent content,
and hence indicator of solution concentration, and solvation water per molecule becomes

equal to bulk water per molecule at or near the saturation point

SRO and LRO in Suw Solutions

The definitions of diffusion of solute through solvent, and other transport
properties as envisioned for dilute solutions are not useful or of the same physical

significance in supersaturated solutions. One view of transport which might be more

177

realistic is that transport and transport coefficients are governed by the SRO, structure, and
availability (thus LRO), of the bulk solvent It might be more useful to describe measures
of resistance to transport in terms of the transport resistance experienced by some small but
inert probe molecule through a solution of interest Solutions which contain a significant
amount of non-bulk-non-solvation water at saturation are likely to exhibit LRO on
additional scales of significance and additional solution structure transition points

The bulk water per molecule of sucrose continues to decrease until at a
concentration of about 76 wt.% sucrose the number of molecules of solvation water per
molecule of sucrose changes significantly. At this point solute-solute interaction becomes
more likely, even though in absolute terms there is still enough water available in the
system to completely solvate every molecule of sucrose, although there would be no bulk
water if this occurred. This may correspond to a glass transition point for the sucrose water
system. A conceptual representation of structure in supersaturated aqueous sucrose

solutions of concentrations less than 76 weight percent sucrose is presented in Figure 8.4.

below.

8.1.3 Bulk Water and Solvation Water as indicators of SRO; Solvatons
and Structure in Bulk Solvent as indicators of LRO

From a practical point of view it is unrealistic to deal with more than two or
three different parameters to describe solution properties and behavior. Consequently the
challenge is to identify a set of parameters that is small enough to be of practical use and at
the same time of physical significance to the system and the changes taking place within it
Rather than identifying the parameters a-priori, we chose to investigate the the system and
its behavior as crystallization proceeded, hoping that such a study would suggest scales of
significance, and, consequently, such a set of parameters. We were searching in particular
for scales significant to the crystal growth process, which led us to a related issue of

significance: what these scales are and how they are identified. This is not a new problem

 

 

178

nor is it appropriate to attempt the solution in a generic manner. The approach we decided
upon was to study several related systems in a comparative manner in the hope of
determining a functional set of physically realistic parameters for the family of systems
examined.

It appears from the analysis of the data obtained that the number of
molecules of bulk water per molecule of solute, and the number of molecules of solvation
water per molecule of solute are each indicators of the nature of the SRO associated with
the solution. They are measures of the nature, extent, and time scales of short range
interactions (relative to proton exchange) between solvent and solvent, and solvent and
solute respectively. The non-bulk non-solvation water is an indicator of SRO and an
indicator of LRO on additional scales of significance, but cannot be obtained by direct
measurement. LRO in the sucrose water system may be described in terms of solvatons,
consisting of a sucrose molecule and the number of water molecules of solvation measured
at the concentration of the solution, and PC, which are determined by the number of
tetrahedral clusters that can be formed by the bulk water per molecule of sucrose at the
same concentration. Regimes of solution structure in aqueous sucrose systems are
summarized in Figure 8.5.

In previous analyses the evidence indicating solvent solute interactions has
sometimes been ignored. Micro and macroscopic transport properties are of the same order
of magnitude in supersaturated sucrose solutions. In light of the previous analysis this
makes sense. In supersaturated solutions the solvatons outnumber the bulk water
structures, and in all likelihood it is the solvatons which solvate the bulk water structures or
FC or TC. As a result there is not likely to be long range macroscopic transport in a
supersaturated solution on the order of magnitude of diffusional time scales in dilute
solutions.

In the particular case of supersaturated solutions, a study of the steady state

fluorescence spectroscopy of pyranine (PYR) in sucrose solutions of various

 

179

concentrations indicates the molar ratio of solvation water to bulk water per mole solute is
such that there is only enough water available on the average for one solvation shell or an
incomplete solvation shell per molecule of solute (Chakraborty and Berglund 15 a). Also
the data of Pan et. al.., as well as viscosity data for supersaturated aqueous sucrose
systems suggest that rotational as well as translational mobility of solvated solute is
severely limited. For this reason, one would expect transport governing interactions in
supersaturated solution to be dominated by short range non-equilibrium solvent-solute
interactions. The absolute times required for these interactions to take place in
supersaturated solutions may be tens, hundreds, or even thousands of times longer than in
dilute solution (Pan and Berglund, 45); yet, in terms of distances involved, they are still
short range, and in terms of time scales relative to diffusional transport, they are still fast
One might object to the idea that bulk water moves faster (relative to
diffusional time scales) in supersaturated solutions than in undersaturated ones because it
seems contrary to intuition. However, when the energetics of the formation of a
supersaturated solution are considered, this notion begins to make sense. It is impossible to
make a phase transition from undersaturated to supersaturated solution without the addition
of large amounts of external energy. This energy serves to increase the disorder of the
solution and also increase the probability that enough bulk water molecules will be present
in the vicinity of all the unsolvated solute molecules (one at a time) to solvate them (one at a
time). The most common way of accomplishing this is to heat the solution until all the
solute dissolves. Then as the solution cools, the solvated solute molecules transfer their
kinetic energy to the surroundings in the form of heat, and become less mobile, (accounting
for most of the cooling effect since these contain most of the mass) and the bulk water loses
some kinetic energy as well. Under identical conditions of pressure and temperature, a
supersaturated solution contains more energy per unit mass and volume than a saturated
solution. In addition, the energy in a supersaturated solution is partitioned differently. In an

undersaturated solution the bulk water structures form the immediate solvation environment

 

180

of the solvatons, in supersaturated solutions the solvatons form the immediate solvation
environment of the bulk water structures. The saturated system retains energy of
reorganization which is required to accomplish the transition from bulk water solvating the
solvatons to the solvatons solvating the bulk water. The higher energy of a supersaturated
solution when compared to the energy of a saturated system in equilibrium with solid at the
same temperature may possibly be accounted for in terms of this energy of reorganization.
Such a structural transition can also explain the long time structural stability of apparently
metastable solutions. Such a restructuring may allow the system to achieve a new
equilibrium between minimizing its total energy and nraxinrizing the entropy of itself and its
surroundings. Observations such as the fact that bulk water moves faster relative to
solvated water in supersaturated solution than in a saturated solution (even though on an
absolute basis, both bulk and solvation water are less mobile) could make more sense if

this is true.
8.1.4 Other effects on structure and LRO

Temrature

Since SRO and LRO are both related to the Gibbs Free Energy of the
system, there is an e"(-(RT)) dependence on absolute temperature. At temperatures near
300 K, a ten percent variation in absolute temperature is not expected to significantly affect
the magnitude of this term. Therefore we would expect relative spatial orientations of
molecules to be a fairly weak function of temperature, and not expect to see major changes
in solvent functional microphase distribution over the temperature range of interest. This is

in fact, supported by our observations.

Absence of Si git rcant Solute-Solute Interactions

181

Large (>10 molecule) solvent exclusive solute clusters are unlikely to exist
in supersaturated solutions of sucrose and glucose in water. All spectroscopic evidence for
highly supersaturated sucrose solutions indicates some molecular interaction between solute
molecules. Such associated groups likely consist of no more than two or three partially
solvated solute molecules. Spectroscopic evidence for solute association in supersaturated
glucose-water system is more doubtful, particularly when the large amounts of non-bulk-

non-solvation water in a system are taken into account Structure in the aqueous glucose

system is summarized in Figure 8.6.

 

182
Likelihood of Si gn_11' rcant Stine-80193 Interaction

Data for other systems such as the citric acid water system and the fructose
water system indicate that solute-solute association is probable and likely in those
supersaturated solutions because there is simply not enough water to maintain the solvaton
structure indicated in under-saturated solutions. In the aqueous fructose system there is
clearly not enough water per molecule of fructose to completely solvate it. Also the
assumption that bulk water tends to associate in tetrahedral clusters when time and space
averaged may no longer be a good assumption. Perhaps the appropriate structural module
of bulk water is a single molecule under these conditions. If this is so, calculating the
number of solvatons per bulk water and number of bulk water per solvaton, shows that the
number of bulk water molecules equals the number of solvatons at 72 wt.% fructose.
While this does not give an exact value of 78 wt% fructose as reported by Bates;
enantiomeric and conformational effects were not considered, nor were the presence of
difructose dianhydrides any or all of which may significantly affect the LRO of the system.

. This is summarized in Figure 8.7.

Weakly Dissociated Smgies with Partial Ionic Character

Saturation in the citric acid water system occurs near the point where the
number of water of solvation per molecule of citric acid equals the number of bulk water
per molecule of citric acid. At this concentration the citric acid molecule is still weakly
dissociated. There are 2 FC or TC per solvaton. One way to explain this behavior might be
to say that some of the bulk water solvates the dissociated protons so that on the average
the presence of an extra four molecules of bulk water is required per molecule of citric acid.

Such water might be involved in solvation of protons but would exhibit the exchange rate

183

of bulk water (since bulk water is water able to participate in proton exchange). This is
summarized in Figure 8.8.

8.1.5 Issues related to structure near growing sucrose crystals.

With the introduction of a crystal or nucleus a new phase is introduced into
the system and therefore the number of scales on which the system exhibits LRO must
increase. It is interesting to interpret the "patches” observed in the ratioed and time averaged
fluorescence density images obtained in dual detector mode using laser induced optical
fluorescence microscopy in this light. Most of these patches vary in extent from a few
square microns to a few hundred square microns. They appear to represent areas of
solution with sucrose concentrations comparable to bulk concentration separated by areas
of solution with considerably lower sucrose concentrations. None of the regions appear to
be solvent exclusive (except for regions clearly corresponding to the crystal itself). The
very near interface region of the crystal has a structure which is completely unknown,
however the water of solvation released due to crystallization to collect in relatively large
patches. The sucrose concentration in these patches corresponds to undersaturated but
concentrated solution.

Solvatons likely diffuse into the solvent rich regions while bulk-solvent
counter-diffuses into the solvaton rich region. Bulk solvent structure is likely different in
each region, and transport is much slower in the solvaton rich region than in the solvent
rich region. When one considers the energetics of saturation it becomes evident that there is
not enough energy available in the system for the undersaturated patches to become
supersaturated and thus the ”phase inversion" from the solvatons being solvated by the
bulk water to the bulk water being solvated by the solvatons to occur. In a like manner
because the chemical affinity of sucrose for water is greater than the chemical affinity of
sucrose for sucrose it is unlikely that there is enough energy available for the supersaturated

184

regions of solution to ever attain undersaturation. This may be a way to conceptualize in
spatial terms or on the level of LRO, why crystallization stops when the concentration of
bulk solution reaches saturation. Even though one might be tempted to apply the term
clusters to these regions, it is important to note that these are not solvent exclusive
aggregates and they are no more "pm-crystalline” than the supersaturated solution itself
prior to nucleation or seeding.

The crystallization itself (at bulk solution concentrations of less than 76%) is
likely initially driven by interfacial competition for water of solvation between the sucrose
molecules on the solution side and the sucrose molecules on the solid side on the one hand
and the fact that the affinity of water molecules for other water molecules is greater than the
affinity of water molecules for sucrose on the other. Some water molecules are likely to
associate with other water molecules in the neighborhood instead of making themselves
available for solvating sucrose.

The boundary layer which extends to the point that bulk solution
concentration is observed, likely constitutes the next scale of LRO of significance. This
research suggests that the spatial descriptor of LRO in the boundary layer must be relative
to the size of the crystal but not to the degree of supersaturation of the bulk solution. This

work does not provide information related to the temporal descriptor.

8.2 Conclusions

Steady state fluorescence spectroscopy of pyranine as a solvent probe can
be used to determine concentration in supersaturated solutions of sugars such as sucrose,
glucose, fructose and lactose. The PIR determined by this method may also be related to

the measurement of activity. In this sense, pyranine can be used as a simultaneous probe of

activity and concentration in such systems.

 

 

185

Steady state emission spectroscopy of pyranine is not a good indicator of
solute concentration in dilute solution. Water appears to have more than one function in
solution, some of the molecules at any given time are unavailable for participation in
solvent exchange and constitute the solvation water nricrophase, and other water molecules
are available for participation in solvent exchange to differing degrees and constitute the
bulk water nricrophase. The solvation water microphase for most substances studied
remains fairly constant until high degrees of supersaturation are reached. Steady state
fluorescence spectroscopy of pyranine can only directly measure solvent exchange rates
relative to the rate of proton exchange on two time scales. Such information is not enough
to completely describe the behavior of systems such as glucose-water where large amounts
of non-bulk-non-solvation water are present, or f ructose-water where a some solute
association is likely, or citric acid water where partial dissociation occurs.

Combined with other techniques such as time-resolved-fluorescence—
spectroscopy or NMR, useful information related to structure, can be obtained on more
than one scale. An analysis of all information available in literature combined with the
information generated for the sucrose-water, glucose-water, f ructose-water, and citric acid
water systems indicates that at or near the saturation point structural changes occur in terms
of both short range and long range order. The structure of the bulk water phase may also
change although the steady state fluorescence method by itself does not provide evidence
on the correct scale of long range order. The transport resistances in a system appear to be
related to the structure and quantity of bulk water present.

The steady state fluorescence method can be modified for use in laser
excited optical fluorescence spectroscopy to yield information related to the solvent
distribution in a 2-dimensional area of solution. Such images collected in supersaturated
solution, in the presence and absence of growing crystals, and in concentrated solutions in
the presence of dissolving crystals, indicate that structure in each case is different The

extent of the boundary layer appears to be related to the size of the crystal and have no

186

supersaturation dependence. The structure of the near interface region is completely

unknown.

8.3 Recommendations for Further Work

Information related to the structure of the bulk water phase in supersaturated
solution is critical to the understanding of crystal growth. The NMR data of Richardson
and Lang should be a good starting point with which to corroborate data obtained by other
methods. This author is unable to suggest any existing method which might be of use.

Two dimensional laser induced optical fluorescence microscopy and
confocal techniques have the capability of providing three dimensional concentration
profiles of solvent in supersaturated solution. Studies of growing crystals in different
systems may provide useful information regarding crystal growth mechanisms and the role
of solution structure in crystal growth.

Investigations of crystallization from non-aqueous solvents may provide
further information regarding the general role of solution structure in crystal growth from
solution.

Two dimensional imaging techniques may prove to be the best candidates
for developing on-line particle imaging and control instrumentation. The effects of noise,
detector response, movement of particles, and appearance of new nuclei need to be

characterized first.

'11.

187

A solvaton=
su rose molecule with 6 water molecules of salvation

single water molecule

 

Flickering cluster) 91% c.3098, 3° erg an
of 4 water molecules ’09

Figure 8.1 Conceptual diagram of structure in dilute aqueous sucrose solution.

 

188

A solvaton = a sucrose molecule and 6 water molecules of salvation

7’ . S5“ 9° ‘3"

9
§ A single water molecule
0
0
I

‘30

A flickering cluster of 4 water
molecules on the average ‘8'

Figure 8.2. A conceptual representation of structure in concentrated aqueous sucrose
solutions.

 

189

A solvaton

‘HWCQZ a»
a. ‘ 2:3 A flickering cluster,
01f 03" tetrahedral on the average
’3' g: or

3" a»

Figure 8.3. A conceptual represntation of structure in saturated aqueous sucrose
solution.

190

 
 
  

‘ A flickering cluster,
tetrahedral on the average

Figure 8.4. A conceptual representation of structure in supersaturated aqueous sucrose
solutions of less than 76% by weight sucrose.

191

Solutlon Structure Raglmas for the Aqueous Glucose System

 

 

 

 

 

30
El Ntc/Nsolvaton, undersaturated solution
A N(tc-S)/Nsolvaton, dlute solution
s Nthsolvaton. concentrated solution
‘ O Nsolvaton/Ntc. supersaturated solution
5 I (n(nbns/4))/Nsolvaton
D
c 20 -
o e
is Ntc - Nsolvston
o
i a A Ntc>Nsolvston Nsolvston >Ntc
sstrr.
2% a i
6 - N
a I
c 3 i a e
g! 10 " V; ; he
i U N g «I
t 1 l
. i i 3
' O
s. ; x
1 I 0
I I £ I £ I ' + I I I
0 r I ' I ' fi‘ ' ' ' '
O 10 20 30 40 50 60

Wslghtfit Glucose

Figure 8. 5. Structure in the aqueous glucose system.

 

I, Is, IthIsslvsm, Ir Isolvstsallts

192

Aqueous Fructose System

 

 
       

 

 

 

 

 

1003 —Cl— Nsolvaton/Nb
I -—I— M)
1 g Nb-Nsolvston —0— Ns

103 --t— Nsolvaton/Ntc
1‘:

~19 D/ i
: 2

.01 ' I v I I V; I v
0 20 40 60 80 100

Wt% Fructose

Figure 8.6. Solution structure in aqueous fructose systems.

 

193

Aqueous Citric Acid Solutions
100 —‘-"— M)
—0— Ms
i 3 —O— Ntc-Nb/4
II —I— Nsolvaton/Ntr

 

l l

10

A 4 llllll

1 1 Ntc/Nsolvaton-Z

 

"it, Is, Itcflsolvston or Isslvstsollte

 

 

 

 

 

~01 ' I 1 I I I I
0 20 40 60 80

Wt.% Citric Acid

Figure 8.7. Summary of structure in aqueous citric acid systems.

 

194

Solution Structure Regimes In Aqueous Sucrose Solutions

 

 

 

 

 

g 40
3 I N(tc-6)/Nsoivaton. Dilute Solution
'6 I A Nthsolvaton, concentrated solution
i O Nsolvaton/Ntc. Supersaturated Solution
1 3O - Nsolvstons - Ntc
° 3 Sam.
~ I
l a s 3
.1: g E 2
3 20 - a 3 : g
Q ' 3
5 .. s 3 5 s
g '3 s i! a i
3 i 3.: 2
Q i 2
‘ H; 3 5 3 2 s
'1' I N i s 2 h
I ; A F 2 v
3 10 .. ~ I a E 3 H
8 S 2 z a E 3
I s §
§ 1 I : ‘ f O
0 : ““ .. g
i 0 1 . . if I ”ind—H—
0 20 40 60 80 100

Wt% Sucrose

Figure 8.8. Regimes of solution structure in aqueous sucrose solutions.

111011an 51an UNIV. LIBRARIES
llMillimilliilmilililWillillilillilllliliil
31293015660784