.xmwwué...é
.. . . ... .-

 

 

....fl. .
«m. L...
x
.

3:3 ..
3|)c.)

swan.

.. all: 14‘
3.; u

...-l

, , 90%.“. La}...
1.? irnzvm§r . . PREV”. fivfimwwulr
{3.3.583 . . . ‘ , hi“. n.i..u....h..»5
. ...un - . x .‘ . .919: .Mwufiruuflrt.
‘ ...!

.flv

. 693.. u
'3‘.-

‘uiv I.“

attifits
L"?

... .1 .‘p..uur.wn

‘0 .
. .SIIEIA.K.X£

. ......v9a‘..\.\n§|(

\! l G.)

£.-,
vii!"
.0. hot;
9
0..

:9 .
. , I , .r >3h¢x£h§ an.
..I £311.. .IBAI v. 1

if vb

.fihfiufinifl?

MICHIGAN STATE EIVUN

II III II III III IIII IIIIIIIIIIIIIIII

LIBRARY 301572 1917

Michigan State
University

 

 

"‘-«E-“si& m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

thesis entitled

THE USE OF RAMAN AND FLUORESCENCE
SPECTROSCOPY TO STUDY CRYSTALLIZATION
IN SITU

presented by
Robert John Richards

has been accepted towards fulfillment
of the requirements for

Master of Science degree in Chemical Engineering

%0W

Major professor

Date ”Lg/- 14’) 77/6

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

PLACE ll RETURN BOX to tuna-this chockwtfiun your record.
TO AVOID FINES return on or before date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

usu IoAnAffitmativo Action/Equal Opportunity Institution
WM‘

THE USE OF RAMAN AND FLUORESCENCE SPECTROSCOPY TO STUDY
CRYSTALLIZATION IN SIT U

By
Robert John Richards

A THESIS

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

MASTER OF SCIENCE

Department of Chemical Engineering

1996

ABSTRACT

THE USE OF RAMAN AND FLUORESCENCE SPECTROSCOPY TO STUDY
CRYSTALLIZATION IN SITU

By
Robert John Richards

Crystallization is the oldest separation and purification process known to man.
It can be used for a wide range of processes. Much of the basic research in
crystllization centers on understanding, factors which affect crystal size distribution
(CSD), bulk density, filterability, slurry viscosity, and dry solids flow properties.
These properties in turn depend on CSD and crystal morphology. The factors which
affect the fore-mentioned properties are also wide and varied, and often require a great
deal of trial and error to obtain parameter information needed to produce the desired
product.

To improve production, it would be beneficial to develop tecniques that measure
crystallization parameters quickly and efficiently, without trial and error. Ideally, these
methods could be done in situ, thus minimizing disruptions to the system.
Spectroscopic methods offer a means to determine crystallization conditions, with the
goal being to improve the process, and ultimately develop a system for on-line control.

In this work, fluorescence and Raman spectroscopy are investigated as means
to determine crystallization parameters. Fluorescence spectroscopy is used to detect
solution changes in aqueous sugar systems, with emphasis on sensitivity in the
supersaturated region. Raman spectroscopy is used to observe a solvent-mediated

phase transformation from both the solution and the solid sides of the system.

ACKNOWLEDGEMENTS

I give my professional acknowledgements to Kris Berglund, Tom Carter,
Beatrice Torgerson, Laurie Tanner, and Robert Buxbaum. I also wiSh to thank LEK
pharmaceuticals, Ljubljana, Slovenia for the cimetidine.

Funding was provided by the National Needs Fellowship Program and the
Cooperative State Research Service of the United States Department of Agriculture,
Grant Number 90-34189-5014. Additional Support was provided by the Crop and
Food Bioprocessing Center/Research Excellence Fund at Michigan State.

I thank everyone who accepts and respects me for who I am. Their

understanding has allowed me to make it this far.

iii

TABLE OF CONTENTS

CHAPTER PAGE
1. BACKGROUND ..................................... l
1.1. Introduction ..................................... l
1.2. Overview of Crystal Polymorphism . . , ................... 3

1.3. Common Analytical Techniques to Study Crystal Polymorphism. . 7

1.4. General Theory of Raman Spectroscopy .................. 8
1.5. Low Frequency Raman Spectra of Crystals ................ 10
1.6. Theory of Emission Spectroscopy ...................... 14

1.7. Instrumentation and Applications of Fluorescence Spectroscopy . . 17
REFERENCES ...................................... l9

2. THE USE OF RAMAN SPECTROSCOPY TO INVESTIGATE
SOLVENT MEDIATED PHASE TRANSFORMATIONS IN

CIMETIDINE ....................................... 21

2.1. Introduction ..................................... 21
2.2. Procedure ...................................... 23
2.3. Results and Discussion ............................. 25
2.4. In Situ Raman Spectra of Cimetidine Slurries .............. 34

2.5. Summary and Conclusions ........................... 36
REFERENCES ...................................... 38

iv

TABLE OF CONTENTS (Continued)

CHAPTER PAGE
3. THE USE OF CARMINIC ACID AS A TRACE FLUORESCENT
PROBE IN AQUEOUS SUGAR SOLUTIONS ................ 39
3.1. Introduction ..................................... 39
3.2. Carminic Acid as a Tailor-made impurity ................. 41
3.3. Procedure ...................................... 41
3.4. Results and Discussion ............................. 43
3.5. Summary and Conclusions ........................... 57
REFERENCES ...................................... 59
4. FUTURE WORK ..................................... 61
4.1. Assignment of the Fundamental Lattice Raman Bands for
Cimetidine Crystals ................................ 61
4.2. Raman Spectral Investigations of Other Cimetidine Forms ...... 62
4.3. Cimetidine Crystallization Experiments ................... 63
4.4. Improvements in Raman Spectroscopy Instrumentation ........ 63
4.5. Carminic acid .................................... 65
4.6. Crystallization of Carminic Acid-Doped Sugar Solutions ....... 66
REFERENCES ...................................... 68
APPENDIX: Original Data from Scatter Plots .................. 7O

LIST OF TABLES

TABLE PAGE

I. Conditions and Results for phase transformation experiments done in
Isopropanol ........................................... 29

11. Conditions and Results for phase transformation experiments done in
1.75/1 (v/v) isopropanol-water .............................. 29
III. Original Data from Figure 12 ................................ 70
IV. Original Data from Figure 13 ................................ 70
V. Original Data from Figure 15 (All PIR’s 675 cm- 1/494 cm-l) .......... 70
VI. Original Data from Figure 16 (All PIR’s 675 cm-1/494 cm-l) .......... 71
VII. Original Data from Figure 24 ................................ 71
VIII. Original Data‘from Figure 25 ................................ 71
IX. Original Data from Figure 26 ................................ 72
X. Original Data from Figure 27 ................................ 72
XI. Original Data from Figure 28 ................................ 72
XII. Original Data from Figure 31 ................................ 73
XIII. Original Data from Figure 32 ................................ 73
XIV. Original Data from Figure 33 ................................ 73
XV. Original Data from Figure 34 ................................ 74
XVI. Original Data from Figure 35 ................................ 74

vi

LIST OF FIGURES

FIGURE PAGE
1. Idealized curves of montropic and enatiotropic systems ............ 4
2. Schematic of a solvent-mediated phase transformation ............ 5
3. Idealized changesin supersaturation during a solvent-mediated phase

phase transformation .................................. 6
4. Schematic of Rayleigh, Stokes Raman, and anti-Stokes Raman

Scattering .......................................... 8
5. Contrast between resonance and non-resonance Raman scattering ..... 9
6. Schematic of light absorption and emission .................... 16
7. Block diagram of typical fluorescence spectroscopy instrumentation . . . l7
8. Chemical structure of cimetidine ........................... 22
9. Schematic of experimental setup for cimetidine solubility, phase

transformation, and in situ studies. The cuvette holder and dissolution
chamber were water jacketed for temperature control ............. 24

10. Raman spectra of crystalline cimetidine. Excitation'wavelength
647.12 nm, temperature 25 °C ............................ 26

11. Typical Raman spectra of cimetidine in isopropanol or
isopropanol-water solutions. Excitation 647. 12 nm, temperature 25 °C . 27

12. PIR’s of 675 cm-1 (cim) to 494 cm-1 (isoprop) as a function of cimetidine
concentration in isopropanol solutions. Solubilities are represented by
ovals and calibration curve points are represented by squares. Excitation
wavelength 647. 12 nm, temperature 50 °C ..................... 28

13. PIR’s of 675 cm-1 (cim) to 494 cm-1 (isoprop) as a function of cimetidine
concentration in 1.75/1 (v/v) isopropanol-water solutions. Solubilities
are represented by ovals and calibration curve points are represented by
squares. Excitation wavelength 647. 12 nm, temperature 25 °C. ...... 28

14. Solid state Raman spectral changes during phase transformation of
cimetidine. Excitation wavelength 647. 12 nm, temperature 25 °C ..... 3O

15. Change in PIR’s of 675 cm-1 (cimetidine) to 494 cm-1 (isopropanol)

during phase transformation experiments in isopropanol. Excitation
wavelength 647.12 nm, temperature 50 °C .................... 32

vii

LIST OF FIGURES (Continued)
FIGURE PAGE

16. Change 1n PIR’s of 675 cm- 1 (cimetidine) to 494 cm- I (isopropanol)
during phase transformation experiments in 1.75 (v/v) isopropanol-water.

Excitation wavelength 647. 12 nm, temperature 25 °C ............. 33
17. In situ Raman spectra of cimetidine form A in isopropanol.

Excitation wavelength 647. 12 nm, temperature 25 °C ..... g ........ 35
18. In situ Raman spectra of cimetidine form B in isopropanol.

Excitation wavelength 647.12 nrn, temperature 25 °C ............. 35
19. Chemical structure of carminic acid .................. V ....... 40

20. Absorption spectra of carminic acid in aqueous glucose solutions.
Carminic acid concentration 5 ppm, temperature 25 °C ............ 44

21. Absorption spectra of carminic acid in aqueous sucrose solutions.
Carminic acid concentration 5 ppm, temperature 25 °C ............ 45

22. Emission spectra of carminic acid 1n aqueous glucose solutions.
Carminic acid concentration 2. 5 ppm, excitation wavelength 500 nm,
temperature 25 °C .................................... 47

23. Emission spectra of carminic acid in aqueous glucose solutions.
Carminic acid concentration 2.5 ppm, excitation wavelength 310 nm,
temperature 25 °C .................................... 47

24. PIR’s (594 nm/ 610 nm) of carminic acid emission in aqueous glucose
solutions. Carminic acid concentration 2.5 ppm, excitation wavelength
500 nm, temperature 25 °C .............................. 48

25. PIR’s (594 nm/ 450 nm) of carminic acid emission in aqueous glucose
solutions. Carminic acid concentration 2.5 ppm, excitation wavelength
310 nm, temperature 25 °C .............................. 48

26. pH of glucose solutions as a function of glucose concentration.
Carminic acid concentration 2.5 ppm, temperature 25 °C ........... 50

27. Comparison of carminic acid PIR’s in aqueous glucose solutions and
buffer solutions. Carminic acid concentration 2.5 ppm, excitation
wavelength 500 nm, temperature 25 °C ...................... 51

28. Comparison of carminic acid PIR’s in aqueous glucose solutions and
buffer solutions. Carminic acid concentration 2.5 ppm, excitation
wavelength 310 nm, temperature 25 °C ...................... 51

29. Emission spectra of carminic acid in aqueous sucrose solutions.
Carminic acid concentration 2.5 ppm, excitation wavelength 500 nm,
temperature 25 °C .................................... 53

viii

LIST OF FIGURES (Continued)
FIGURE PAGE

30. Emission spectra of carminic acid in aqueous sucrose solutions.
Carminic acid concentration 2.5 ppm, excitation wavelength 310 nm,
temperature 25 °C .................................... 53

31. PIR’s (610 nm/ 580 nm) of carminic acid emission in aqueous sucrose
solutions. Carminic acid concentration 2.5 ppm, excitation wavelength
500 nm, temperature 25 °C .............................. 54

32. PIR’s (594 nm/ 450 nm) of carminic acid emission in aqueous sucrose
solutions. Carminic acid concentration 2.5 ppm, excitation wavelength
310 nm, temperature 25 °C .............................. 54

33. pH of glucose solutions as a function of sucrose concentration.
Carminic acid concentration 2.5 ppm, temperature 25 °C ........... 55

34. Comparison of carminic acid PIR’s in aqueous sucrose solutions and
buffer solutions. Carminic acid concentration 2.5 ppm, excitation
wavelength 500 nm, temperature 25 °C ...................... 56

35. Comparison of carminic acid PIR’s in aqueous sucrose solutions and

buffer solutions. Carminic acid concentration 2.5 ppm, excitation
wavelength 310 nm, temperature 25 °C ...................... 56

ix

LIST OF SYMBOLS

Ak = the preexponential constant

E = the electric field vector of incident light
e = the exponential function

K9 = the propagation direction

m = the dipole moment vector

Qk = the set of normal coordinates

t = time

a = the polarizability tensor

(1),, = the direction-cosine matrix

v9 = the monochromatic wave frequency

Chapter 1
Background
1.1: Introduction

Crystallization is the oldest unit operation in chemical engineering. The first known
application was the separation of sodium chloride from sea water during biblical times.
Almost every chemical industry employs crystallization as a separation or purification step
during production. Crystallization’s primary purpose is to separate a pure product from an
impure mixture, but often additional criteria such as crystal size distribution, bulk density,
and filterablility also need to be met. These properties are dependent upon path, thus
system changes directly affect the final product. System parameters have been, and are still
usually determined through trial and error, causing the optimization of a production process
to be laborious and time consuming. Crystallization has often been thought of as an art
form rather than a science due to these methods of improvement.

Attempts to characterize crystallization were first done in the early 1960’s when
Mullin1 published the first work that treated crystallization phenomena in a general sense,
rather than treating each system separately. Another important publication was Randolph
and Larson’s2 “Theory of Particulate Processes” which gave a mathematical interpretation
of the population balance, allowing for prediction of crystal size distributions in realistic
industrial crystallizers. These two books discussed the effects of solution characteristics
and phase equilibria on crystallization, pointing out the need for analysis techniques that
can accurately determine these properties. I

In order to achieve crystals with desired characteristics, system parameters that

affect the fore-mentioned criteria must be determined to allow for appropriate control. The

1

2
use of several analysis techniques to look at crystallization from both the solution and solid

sides of the system have been developed to aid in this. Randolph and Larson pioneered
two methods, one being the use of a Coulter particle counter to determine crystal size
distributions and the other the use of a mixed suspension, mixed product removal
(MSMPR) crystallizer. Myerson and coworkers3 have looked at concentrated solutions of
both electrolytes and non-electrolytes using diffusion and viscosity measurements.
Narayanan and Youngquist4 have attempted to use electrical conductivity to determine
solution properties in several inorganic salt systems.

The use of spectroscopic techniques in the field of crystallization has received much
attention in recent years due to their ability to observe systems in situ. When a system can
be monitored in situ, disturbances can be minimized, reducing experimental error and
allowing for better control. Richardson et. al.5 used NMR to determine water mobility in
aqueous sucrose solutions. Mathlouthi6 undertook x-ray diffraction studies to investigate
molecular association in aqueous sugar solutions.

Raman and fluorescence spectroscopy are two techniques which offer promise as
analytical tools to monitor crystallization in situ. Previous work has shown that both
methods have potential in this area. Cerreta and Berglund7 showed that the Raman
spectroscopy could be applied to the aqueous glucose system for observation of conditions
from both the solid and solution sides. of the system, with glucose solutions, anhydrous
glucose, and glucose monohydrate exhibiting unique spectra in the fingerprint region.
Chakraborty and Berglund8 have demonstrated that pyranine could be used as a trace
fluorescent probe to study the structure of aqueous sugar solutions.

This study further explores the potential of fluorescence and Raman spectroscopy to
f determine crystallization parameters. In specialty chemical production, it is advantageous
to use in situ analysis techniques for on line control since high purity standards are
required. In situ monitorization greatly reduces or eliminates system disturbances caused

by on-line sampling while providing more accurate real time data. In the making of

3

pharmaceuticals, four nines (99.99 %) purity and batch to batch consistency are necessary
criteria to meet FDA standards, thus good control of system parameters is needed to meet
these demands. Raman spectroscopy’s ability to observe conditions of both the solid and
solution sides of a system makes it ideal to study parameters in the crystallization of
pharmaceuticals. Along with this, crystal polymorphism can be observed since different
crystalline phases have different Raman spectra.

Trace fluorescent spectroscopy has proven to be a viable method to observe
conditions in supersaturated solutions. Pyranine is a valuable probe to study aqueous
sugar solutions, but it is not accepted by the FDA for use in foods. This incompatibility
with food doesn’t allow pyranine to be used as an on-line trace fluorescent probe, thus its
potential for industry is confined to non-food applications. Carminic acid, an RD and C
approved dye was chosen as a possible replacement since many of the same qualities of

pyranine, but it is safe for use in food and drugs.9

1.2: Overview of Crystal Polymorphism

Crystal Polymorphism is the ability of a chemical compound to exist in more than
one crystalline form. '0 Different structures are called polymorphs or polymorphic
. modifications. The ability of a solid to crystallize into more than one crystalline phase
dependson several conditions that exist during crystallization. Temperature, pressure, and
solvent are all parameters that can affect the polymorph obtained.

Polymorphs have similar chemical properties but dissimilar physical properties.
For example calcium carbonate can exist in two forms, calcite and aragonite. Both forms
have the same chemical properties, but exhibit different physical properties. Calcite is a
rhombohedral uniaxial crystal with a density of 2.71 gm/cc whereas aragonite exists as an
orthorhombic biaxial crystal with a density of 2.94 gm/cc.

The importance of studying polymorphic phenomena in pharmaceutical compounds

lies in the different physical properties between polymorphs.11 Differences in crystal form

4

can cause differences solubility, which in turn causes differences in bioavailability and drug
effectiveness. Tableting characteristics are also affected depending on the polymorph being
processed.”13 These types of discrepancies are simply unacceptable to the Food and
Drug Administration, thus it is very important to understand crystallization parameters that
affect polymorphism.

A particular area of polymorphism that causes problems in the pharmaceutical
industry is that of unwanted phase transitions. 14’15'16 Often times the desired phase
produced is the metastable phase which under certain conditions can transform to the stable
phase. A phase transition such as this will reduce the solubility and bioavailability of the
drug, making it less effective. In other cases the stable phase may be preferred, but it can
be difficult to make large amounts of it due to kinetic factors. For a situation such as this,
it is often advantageous to make the metastable phase and then set the conditions to force

a phase transformation to the stable phase.

 

 

 

 

 

Monotropic Enantiotropic
Phase II Phase I
2‘ 2‘
3 Phase I 3’ x
O O .
__/ ___/ Phase II
Temperature Temperature

Figure 1 : Idealized curves of monotropic and enantiotropic systems

There are two types of polymorphic systems, enantiotropic and monotropic. An
enantiotropic system has an interconversion temperature below the melting point of each
polymorph whereas a monotropic system doesn’t. A monotropic system has one

polymorph that is the stable form throughout the temperature range. Figure 1 shows

5

idealized monotropic and enantiotropic systems. For monotropic systems, the stable phase
exhibits lower solubility throughout the temperature range.
There are two ways that a phase transformation can take place. In the first the

metastable phase can undergo an internal rearrangement of molecules or atoms that occurs

 

 

 

 

 

 

Ptasei
Xi
:
.2
‘5
..E:
C
a)
o
c
o
0
2
2 (x29tx)
o
(D ———__/
Temperature

Figure 2: Schematic of a solventrnediated phase transition

in the solid state.17 The second route is aided by the presence of a solvent in contact with
the metastable phase which allows it to dissolve and allows the growth and nucleation of
the stable phase independently. This type of phase transformation is called a
solvent-mediated phase transformation.18

The changes in solute concentration as a solvent-mediated phase transformation
takes place can be explained in the following way. A solution at composition xi is
supersaturated with respect to both phase I and phase 11. Figure 2 gives a schematic of this.

As nucleation and growth occurs, the majority of the crystals will be the metastable phase,

with a few of the crystals being that of phase 11, following Ostwald’s law of stages.19

6
During the initial growth phase the solute concentration will fall from xi to x1. This being
the case, it is now possible for the phase transformation to take place via solvent-mediation
since there is a slurry of metastable and stable phase crystals in contact with a solution
saturated with respect to phase I and supersaturated with respect to phase II. As the phase
transformation takes place, phase II grows at the expense of the dissolution of phase I, thus
the solute concentration remains constant as this occurs. Once all of phase I is dissolved,
phase II will continue to grow until the solution reaches the solution composition at x2,
which is the saturation composition of phase 11.

Figure 3'shows the changes in supersaturation with respect to time as the phase
transformation takes place. The 512 point represents the initial supersaturation at the
beginning of the phase transformation where the solute concentration is x1. Region 1 of the
curve is the initial onset of the phase transformation where the growth of phase two is

faster than the dissolution of phase one.

012 ,
\ Region1
\

C
.9
g 1
(U
w 0
8 \ Region 3
D.
3
U) \
\ ‘
\

 

 

Tim e
Figure 3: Idealized Changes in Supersaturation During a Solvent-Mediated Phase Transformation.

As the transformation proceeds, region 2 represents a region of approximately

constant supersaturation, where the dissolution rate of phase I equals the growth rate of

7

phase 11. Region 3 represents the time domain where the phase transformation is complete,

and where only phase II is growing.
1.3: Common Analytical Techniques Used to Study Crystal Polymorphism

There are several methods by which to characterize polymorphic systems. Methods
include X-ray techniques, FT-IR, thermal techniques, solubility measurements and hot
stage rrricroscopy20‘23.

X-ray crystallography is the most common way to determine molecular information .
of a crystalline solid. The method is an excellent way to determine the lattice structure of
crystals and can also be used to obtain the packing relationship of molecules in a solid.
X-ray crystallography is not practical to use with a mixture of crystals or crystals that are
small in size, since the technique requires one large crystal for analysis.

FT-IR spectroscopy is extremely sensitive to both structure and conformation, and
is very good to investigate the solid state. There are several examples in literature that
show that FT-IR spectral differences can be used to distinguish different polymorphs. It is
difficult to use FT-IR in solvent-solute systems because the solvent spectra is usually much
stronger than the solute spectra, especially when the solubility is low. Another drawback is
that FT-IR cannot look at the far-IR region of the spectrum due to instrumentation
limitations, thus the lattice vibration region of the spectrum cannot be observed by FT-IR.

Thermal techniques such as Differential Scanning Calorimetry (DSC) and Thermal
Gravimetric Analysis (T GA) are used to determine melting points and to detect the presence
of solvates. These techniques lack resolution, which causes problems with polymorphs of
similar melting points. It is also difficult to differentiate between solid-solid
transformations and recrystallization processes. When done in conjunction with hot stage
microscopy, DSC endotherms and exotherrns can be correlated with either solid-solid

phase transitions or solid-liquid phase transitions.

8
Since polymorphs» exhibit different solubilities, dissolution and/or solubility
measurements are often used to distinguish between two or more crystal forms. These
studies are especially important when done on pharmaceutical products since solubility

differences result in bioavailability differences.
1.4: General Theory of Raman Spectroscopy

Light scattering can occur when light interacts with matter. When the scattering is
elastic and of the same frequency as that of the incident radiation, it is known as Rayleigh
scattering. The other type is inelastic and can be further broken up into two types of light
scattering: scattering of longer wavelength light which is known as Stokes Raman
scattering and scattering of shorter wavelength which is known as anti-Stokes Raman
scattering.24

Anti-Stokes
Raman Scattering

° vo+Av

v0

'"°“'°"‘ W v0 3......"

Light Scattering

  

vo—Av

Stokes Raman
Scattering

Figure 4: Schematic of Rayleigh, Stokes Raman, and Anti-Stokes Raman Scattering

Raman scattering is the result of light interacting with the electrons in a molecule.
There are two types of Raman scattering. Resonance Raman scattering occurs when the

incident frequency is near the electronic absorption band of the molecule, resulting in a

9
coupling of electronic and vibrational transitions. N on-resonance or spontaneous Raman
scattering occurs when the energy of the incident light is not sufficient to raise the
molecular electrons to a higher electronic state.

The frequency of Raman scattering is independent of the excitation wavelength, but
the intensity is directly proportional to v4. This means that regardless of the excitation
wavelength, the Raman spectrum will show the same shift in cm], but the spectrum will be
more intense as the excitation wavelength is decreased. Ideally it seems that the excitation
wavelength should be kept as low as possible, but it is quite cemmon to encounter
problems with fluorescence as excitation wavelength is decreased.25

Interference from fluorescence can pose a problem since it is much more intense
than Raman scattering. The intensity of Raman spectra compared to that of fluorescence
spectra is often 10'3 lower than that of fluorescence intensity. This can be a problem for
many compounds such as specialty chemicals and pharmaceuticals, which often will
fluoresce when excited with UV, blue, or green light. To avoid this difficulty, Raman
spectroscopy of pharmaceutical compounds is done using visible red or IR as the excitation

source.26
1.5: Low Frequency Raman Spectra of Crystals

Based on the theoretical work done by Turre1127, Raman spectral differences in the
low frequency region of the spectrum can depend on several factors. Crystal class,
molecular orientation, bond length, and force constants can all affect the spectra seen at low
frequencies. It is possible for a compound to exist in two crystalline phases which are of
the same crystal system, but the crystal class, molecular orientation, bond length and force
constants among molecules are different. This accounts for Raman spectral differences in
lattice vibrations. Since a polymorphic substance by definition can exist in more than one

crystalline state, different polymorphs should exhibit Raman spectral differences in the

10

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Excited
Electronic _______ _ -
State "" —/\ ----- Z - l- -
J v
v , V
Stokes Anti-Stokes Stokes Anti-Stokes

R nan m n
Non-Resonance Raman 950 C9 Ra a

Figure 5: Contrast Between Resonance and Non-Resonance Raman Scattering

1 1
lattice vibration region, even if they don’t show Raman spectral differences in the
fingerprint region.

Raman spectra of crystals in the lattice vibration region has been used for several
decades to observe polymorphic changes of crystal structure in the solid state.25 These
experiments involved the observation of Spectral changes in a large single crystal irradiated
at known orientations. Up until now, solvent-mediated phase transformations have not
been investigated with Raman spectra in the lattice vibration region.

When a molecule interacts with monochromatic light, there are two oppositely
directed forces that act on the electrons and the nuclei, respectively. Due to this, the nuclei
and electrons will be displaced relative to each other, causing an induced dipole moment.

This can be expressed as

m=ocE (1)

where m is the dipole moment vector, Otis the polarizability tensor, and E is the electric

field vector of the incident light. Equation (1) can also be used to describe the induced
dipole moment in a crystal when it interacts with light.

E represented as a function of time is

-2 K- -
E: Eoe 1t( of Vol) (2)

where ve is the monochromatic wave frequency, K6 is the propagation direction, and t is

time. The actual occurrence of Raman scattering depends on the polarizability tensor,

which in turn depends on the instantaneous position of atoms in the lattice. If 0: is a

function of the set of normal coordinates Qk, it can be expressed as

12

 

. n 80
or — ao+§,(an)oQk+ 2:2:(6anék. )QQk' +..

Generally, only first order (single photon) scattering is considering, thus the polarizability

tensor can be approximated as

oc=oro+§n‘,(aa)oQk (4)

k k

 

Since the lattice waves can be represented by

Qk = Akeflnl (kk0r-th) (5)

Equation (1) can be expressed as: (6)

m=aoE°e-21ri(Keor—vg)+z(%)oAkE092fli[(Kafi‘kl'r-(Velmltl
k k

The first term in equation (5) represents Rayleigh scattering.If in equation (6)

 

at 0,then Raman scattering occurs. For a free molecule— or or' can be represented

601
6Q.

in tensor form as a matrix:

’aa—I—Q—k

1k 1k 1k
axx() axy() axz()

80' 1k 1k 1k 1k 7
_5a zayxuayyuayzu ()

rk‘ 1k "rk
012x" azyn ant)

The orientations of each molecule in a crystal lattice can also be represented in tensor form:

13

ax ay 32 (8)
<1), = b, by b,
cx cy cZ

To account for more than one molecule in a given crystal lattice, the respective polarizability

tensors must be transformed into crystal fixed coordinates. This can be done as follows.

T
an“) = (1),, or“) (1),, (9)

If additivity of polarizabilities is assumed, the unit cell polarizability written as a function of

molecular coordinates becomes.

or = 212% '(k)Qk,nl (‘0)
’7

Using factor group analysis, the molecular normal coordinates can be combined by using

the unit cell symmetry. The general case for this gives:

(i)
Qk = 20km (11)
n

where i represents Ag, Bg, Au, Bu, etc., and n represents the number of molecules in the

unit cell. The signs of Q,“ depend on the orientation of the molecules in the unit cell. By

1 4
substituting the inverse of (10) into (9) , an expression for the polarizability of the unit cell

can be obtained.

(i) T (i)
01k =§l(2¢n°¢k¢n )Qk (12)
n

The derivative of equation (12) will give the tensor necessary to determine the fundamentals

of crystal lattice vibrations for a given system.

(i )

- T

1(l ) __ aak _
(1k =—"'. -2[(2¢nakq)n ) (13)
ac " ’ k n

k
Depending on the crystal system, the form of equation (12) will be different. The
mathematics outlined above can be carried out rigorously for the system under study, or
tabulated results of the tensor forms can be found in literature28 for each of the seven
crystal systems. This gives a general method by which to predict the number of

fundamental lattice vibrations, but it doesn't predict the frequencies of these vibrations, nor

does it account for combinations and overtones.

1.6. Theory of Emission Spectroscopy

Fluorescence and phosphorescence are processes in which radiation is emitted by
molecules that have been excited by the absorption of radiation. Fluorescence and
phosphorescence can be distinguished by the multiplicities from which each respective
molecule emits radiation. If the states from which the emission originates and terminates
have the same multiplicity, then fluorescence occurs. Fluorescence usually occurs between
the lowest S1 vibrational level and the S0 states, although molecules excited to a higher
singlet state can exhibit fluorescence. If the states are different in spin, I then

phosphorescence occurs. The most common phosphorescence occurs between the T1 state

1 5
and the So state. A practical way to distinguish between the two processes would be to
determine excited state lifetimes, since the fluorescence lifetimes are several orders of
magnitude shorter than phosphorescence lifetimes”.
Radiationless processes are caused by the conversion of electronic energy to
vibrational energy. When this occurs between the lowest singlet state (31) and the lowest
triplet state (T1) it is referred to as inter-system crossing. All other cases are known as

internal conversion, which results in no light being emitted.

Light can be represented in terms of energy by Planck’s equation, E=hv, thus

absorption and emission can be thought of as an energy absorption and emission process.
When light impinges on matter, two things can happen: it can pass through the matter with
no absorption or it can be absorbed by the molecules causing them to go from a lower
energy state to a higher energy state. In a sense, the absorption and emission process
could be looked at as an energy cycle with the same beginning and end points, where
energy is neither created or destroyed, but is converted to either radiative or non-radiative
processes. Radiative processes are those in which the excited molecule falls from a higher
energy level to a lower energy level and emits light, whereas non-radiative processes are
those in which the excited molecule falls to from a higher energy level to a lower energy or
vibrational level without emitting light. For gases and crystals, it is possible that the
absorption and emission energies can be the same, thus the excitation wavelength and
emission wavelengths are also the same. This process is known as resonance radiation.
For solutions, the emission energy is lower due to solvent interactions. Lower emission
energy'results in a spectrum of longer wavelength. This process is known as Stokes
fluorescence.30

Each fluorescent molecule has two characteristic spectra: the excitation spectrum
and the emission spectrum. The shape of the excitation spectrum is identical to the
absorption spectrum, and it is independent of the emission wavelength at which it is

measured.”32 The emission spectrum is usually a mirror image of the excitation

1.6

 

d
d

 

 

 

 

 

 

 

 

   
 
       
 
   

 

 

 

 

 

    
   

3 __—:='_ 10
2 ___———____ 9
1 _-:=-_——__
$2 v=0 ——————.— '8’
6 S1
5
4
3 .
2 lntersystem
.1 Crossing
V=°sMa 5
4
3 T
2 1
1 l
. Internal . 0=v
Absorpt'on Conversion Fluorescence
Phosphorescence
s
4
S 3
0 2
1
v=o

Figure 6: Schematic of light absorption and emission

1 7
spectrum, and is independent of excitation wavelength. This wavelength independence
doesn’t mean that the emission intensity will also be independent of excitation wavelength,
since the intensity depends on factors such as the relative strength of absorption and

non-radiative contributions.
1.7 Instrumentation and Applications of Emission Spectroscopy

The measurement of fluorescence requires a light source, excitation and emission
filters, detector, amplifier, and a readout device (usually a computer). The light source is
preferentially filtered 'at the excitation stage so that the sample can be excited by a
monochromatic beam of light. The emission signal wavelength is measured by the placing
the emission filter between the sample and the detector usually located at a 90° angle from
the incident excitation beam. The detector is amplified and the final signal is sent to a
readout device, where the signal intensity can be plotted as a function of emission

wavelength.

 

 

Emission
Filter

 

 

Figure 7: Block Diagram of Typical Emission Spectroscopy Instrumentation

The use of emission spectroscopy as an analytical tool offers several advantages,

1 8

namely extreme sensitivity, excellent specificity and dependence on environmental factors
such as temperature, viscosity, and pH. Quantitative methods of fluorimetry can be broken
up into two branches: The intrinsic fluorescence of a material can be measured as a
function system conditions or a fluorescent probe can be added as an extrinsic monitor of
system conditions in an amount that will have little or no effect on the system. Since the
majority of compounds do not fluoresce, it is often advantageous. to use the method of a
trace fluorescent probe. Since fluorescent techniques have high quantum efficiencies, it is
possible to use small quantities (< 103) that have little or no effect on system conditions.
Trace fluorescent probes can be used in a qualitative fashion to detect the presence or
absence of a reagent. Staining techniques make use of a fluorescent additive to determine
various cell components. Fluorochromes can be used in vivo to define the structures
involved in the production of hydrolytic enzymes and to determine enzymatic activity.
Fluorescent-labelled specific antibodies detect visually specific proteins in foods.
Nanogram quantities of metal and metalloid ions in solution will complex to various
additives allowing for quantitative concentration measurements by correlating the
fluorescence intensity of the complex to ion concentration.

Fluorescence spectroscopy can be applied to crystallization by adding trace amounts
of fluorophore to a system to look at the solution conditions. Both steady state and
time-resolved fluorescence spectroscopy have recently been used. Chakraborty and
Berglund10 used the steady-state fluorescence spectra of Pan33 er al. showed that
fluorescence lifetimes and rotational correlation times of pyranine and pyrene butyric acid in
aqueous sucrose solutions indicated the formation of a network of hydrogen bonding
interactions between sucrose molecules. He further used pyranine in supersaturated protein
solutions to measure anisotropy and correlate these measurements with optimal

crystallization conditions.

19

References

l. J. W. Mullin, W, (Butterwlo and Company, London 1961), preface.
2. A. D. Randolph and M. A. Larson, 1111ng gf Pagigglgte Prmesses, (Academic Press,

\OWQQUI-h

ll.

12

13.

14.
15.
l6.

17.
18.
19.

20.

Inc. 1971) chapter 3.

A. S. Meyerson and K. Toyokura, WWW (American
Chemical Society, 1990) chapter 1.

. H. Narayanan and G. R. Youngquist, AICHE Symposium Series, 253, (1987) 1.
. S. J. Richardson and M. P. Steinberg, Journal of Food Science, 52, (1987) 806.
. M. Mathlouthi et. al., Carbohydrate Research 81, (1980) 213.

. M. K. Cerreta and K. A. Berglund, Journal of Crystal Growth, 102, (1990) 869.
. R. Chakraborty and K. A. Berglund, Journal of Crystal Growth, 125, (1992) 81.

. mm, (Merck and Company, Inc. 1983) pp. 256.
10.

1 t , A.R. Verrna and P. Krishna,
(Russian press, 1967) pp. 8-9.

1 ' t , S. R. Bym, (Academic Press, New York, 1982) pp.
29-58, 79-102.

M. D. Tudalhar, J. E. Carless, and M. P. Summer, Journal of Pharmaceuticals
and Pharmacology , 35, 269-274 (1983).

G. A. Ghan, and J. K. Lalla, Journal of Pharmaceuticals and Pharmacology , 44,
(1992) 678.

R. J. Davey and J .Richards, Journal of Crystal Growth , 71, (1985) 597.
K. Matsuda et. al. Journal of Pharmaceutical Science , 73, (1984) 173.

R. J. Davey, P. D. Guy, and A. J. Ruddick, Journal of Colloid and Interface
Science, 108, (1985) 189.

A. I. Kitaigorodsky, et. al. Journal of Physical Chemistry of Solids , 26,
(1965) 463.

P. T. Cardrew, and R. J. Davey, Proceedings of the Royal Society of London A,
398, (1985) 415.

We, W. Ostwald, (W. Engelmann, Leipzig,1899).

I . Haleblian, and W. McCrone, Journal of Pharmaceutical Science, 58,
(1969) 91 1.

21.
22.

23.

24.

25.

26.
27.

28.

29.

30.

31.

32

33

20

M. Stolz, et. al., Journal of Pharmaceutical Science, 78, (1989) 758.

F. A. Chrzanowski et. al., Journal of Pharmaceutical Science, 73 (1984)
1448.

H. Kamiya, M. Kazuhiro, and K. Morisaka, International Journal of Pharmaceutics,
26, (1985) 197.

n i Pri i l 8 1i t' , A. T. Tu, (John
Wiley and Sons, New York, 1982) pp. 6-14.

r l l s, M.M. Sushchinskii,(Israel Program for
Scientific Translations Ltd., New York, 1972) pp. 1-15.

mm, D. A. Long, (McGraw-Hill, New York, 1977) pp. 1-20.

S r f , G. Turrell,(Academic Press, New York,
1972) Chapter 5 and 6. .

MELWMLMWE flirtatious W. G.
Fateley, et. al.(Wiley Interscience, New York, 1972) Appendix I and II.

s 'n s, L. Munck, (Longman Scientific and Technical UK

W
Limited, 1989) Chapter 1 and 2.

i - e t s ' , G. G. Guibault, (Marcel
Dekker, New York, 1973) Chapter 1.

t r i n c n ,R. S. Bekker, (John Wiley and Sons,
New York, 1969) pp. 3.

W. E. S. Bowen, (Longmans, Green and Company,London,
1953) pp. 5.

B. Pan et. al., Journal of Crystal Growth, 130, (1993) 587.

Chapter 2
The Use of Raman Spectroscopy to Investigate Solvent Mediated Phase
Transformations in Cimetidine

2.1 Introduction

The study of crystal polymorphism is an important aspect of insuring proper
consistency in pharmaceutical manufacturing. It is known that different polymorphs
exhibit different physical and chemical properties, which in turn cause variances in
bioavailability, stability, and tableting characteristics.'°2

In the production of pharmaceuticals, crystallization is the method most commonly
used to achieve high purity and batch to batch consistency. For a substance that can exist
in several different polymorphs, changes in crystallization conditions can change the
amount and type of polymorph upon manufacture. Since manufacturing processes are
usually operated under non-equilibrium conditions, metastable phases frequently occur.
This thermodynamic instability can lead to phase transformations in solution, thus it is
important to determine the conditions under which such transformations occur.

Several techniques have been used to characterize polymorphic systems, including,
DSC, FT-IR, dissolution studies, and hot-stage microscopy.‘ All of these techniques can
be used to distinguish crystalline polymorphs after the fact, but they lack the ability to
measure changes in situ.

The application of Raman spectroscopy to the analysis of pharmaceutical
compounds has become a growing field in the last several years.3'4 Much of the molecular
information that can be obtained from Raman spectroscopy is very similar to that found
with FT-IR, but Raman has several key advantages. Raman spectroscopy can be used on
samples in the liquid, solid, and gaseous state.5 In low-solubility systems with water and

water-alcohol solvents, the solute spectrum can be seen since these solvents have much

21

22

sharper bands that don't interfere with the solute spectrum. [The capability to watch system
changes in situ is another advantage that Raman spectroscopy has over FT-IR since the
fore-mentioned solvent transparency allows for detection of solute. Raman spectroscopy
also allows for a practical means to obtain low-frequency crystal lattice vibration spectra,
which is often important in distinguishing one polymorph from another and can also help
determine crystal structure.6 It is difficult to obtain lattice vibration spectra using
mid-infrared instruments since they lack the capability to acquire low frequency or far-IR
spectra.

Cimetidine (Figure 1) is a commonly used Gastrointesinal-tract acid secretion
inhibitor that exhibits several crystalline polymorphs (A, B, C, and D), all of which have
different bioavailabilities.7 Several crystallization studies have been undertaken to
determine conditions which produce each respective polymo’rph.8'lo Recently, the
solvent—mediated phase transformation of the metastable form A to the stable form B has

been studied in isopropanol solutions.“l

Figure 8: Chemical Structure of Cimetidine.

23

Work done by Tudor et. al.12, demonstrated that FT -Raman using IR excitation could be
used to distinguish cimetidine polymorphs A, B, and C in the solid state while maintaining
effective fluorescence rejection. Their instrumentation, however, did not permit '
determination of the lattice vibration spectrum. Futhermore, high powers at the sample are
required to obtain spectra with acceptable signal to noise ratios due to Raman
spectroscopy's inverse wavelength to the fourth power dependence on spectral intensity.
In our study, we have used conventional Raman spectroscopy with a Krypton ion laser and
a dispersion grating spectrometer to demonstrate the feasibility of the technique and also to
answer some new questions about phase transitions in cimetidine-water and
cimetidine-water-isopropanol systems, and have investigated the possibility of following

these phase transformations in situ.

2.2. Procedure

Polymorphs A and B were supplied by LEK Pharmaceuticals in powdered form
and were used without further purification. Reagent grade isopropanol (Baker) and HPLC
grade water (Sigma) were used for solubility and phase transformation experiments. A
schematic of the experimental setup is shown in Figure 4. The processes are outlined in
detail below

A Coherent Radiation Innova 90-10 watt constant wave Kr+ ion laser at 647. 12 nm
was used as a monochromatic light source for two reasons. First, fluorescence problems
were minimized, as the excitation lies away from the absorption of most fluorescing
substances. Second, the higher disperSion of red light allowed for easier access to the
Raman lattice vibration region without interference from Rayleigh scattering.

The detection system consisted of an EG & G Model 1530-C/Cuv 10248 256 x

1024 CCD red-enhanced multichannel detector coupled to a Spex 1877 Triplemate triple

24
monochrometer. Solid samples were analyzed by directing the laser beam through a Zeiss
standard microscope that was attached to the Triplemate, whereas liquid samples were
analyzed by directing the laser beam through a quartz cuvette mounted before the filter
stage. In situ spectra of agitated slurries were obtained by directing the laser through a

quartz cuvette in a back-scattering sample configuration.

 

 

Cuvette Holder
or Microscope

 

Raman
Scatter

 

Spectrometer

 

 

 

Computer

      

Dissolution
Chamber

 

Magnetic
Stir Plate

 

 

Figure 9: Schematic of experimental setup for cimetidine solubility, phase transformation, and in situ
studies. The cuvette holder and dissolution chamber were water jacketed for temperature
control.

Solutions saturated with respect to phase A were prepared by adding excess A
and allowing the mixture to equilibrate in an agitated vessel for 5 hours. Forty-five ml of
this saturated solution were filtered and put into a temperature controlled batch crystallizer.
Exactly 0.30 g of a known mixture of seed crystals was added. Slurry samples were then
drawn off at known time intervals, with the solid and liquid phases being separated through

filtration. The spectrum of each phase was recorded in the range of Raman shift 64 to 685

25

cm". Power at the sample for liquid samples was 500 mW and for solid samples was 100
mW. Each spectrum consisted of 30, 10 second scans that were accumulated to improve
the signal to noise ratio.

Calibration curves were made by taking spectra of cimetidine/solvent solutions of
known composition and plotting the spectral changes versus solute weight percent. The
Peak Intensity Ratio (PIR) of a solvent band (494 cm") to a solute band (675 cm") was
used to act as an internal standard, to account for fluctuations in the spectral baselines.
Once the calibration curve was made, spectra of the saturated solution of forms A and B
were taken and the PIR was compared to that of the calibration curve, thus giving the

saturation concentration in weight percent.

2.3. Results and Discussion

The solid state spectra of phase A and phase B in the lattice vibration region are
shown in Figure 5. The spectra show distinct differences from 100 to 200 cm", 300 to
350 cm“, 500 to 525 cm“, and 600 to 650 cm".

Typical Raman spectra of cimetidine in isopropanol and/or isopropanol-water
solutions are shown in Figure 6. Three distinct isopropanol lines are seen at 494 cm", 425
cm“, and 375 cm" along with cimetidine spectral lines at 675 and 620 cm“. Raman
scattering from water was insignificant in this region of the spectrum, but a strong Rayleigh
tail, typical of Raman spectra of solutions is seen.16

The solubility curves (Figure 12 and Figure 13) show that cimetidine can be
detected down to levels of 0.5 weight percent in both isopropanol and water. Results
similar to the work of Sudo et. al8 show that Raman spectroscopy is comparable to

common methods in determining solubility curves.

26

 

 

 

 

 

 

2105--
sl .
1.8 10-_ : —FormB
g; . :'. ----- FormA
5 1.6105" ""3
2‘ ,_ I
E .
$ 1.410 _
g I-
(I) 5"
5 1210--
§ .
1105--
8104.11.:.1..1F....j14411:1.11:1LL1:.1.1:

 

 

 

100 200 300 400 500 600 700
Raman Shift (cm-1)

0

Figure 10: Raman spectra of crystalline cimetidine. Excitation wavelength 647. 12 nm, Temperature 25 °C

27

 

 

 

3105:
33.2.5105:
.E I.
3 .
S 2105-
: n
15 I
L-
33 5‘
1510 "
a u-
'17. I
c I
.93 5-
.E 110 _
5104bleLlllljllllllllllllllllllllllllll
0 100 200 300 400 500 600 700

Raman Shift (cm-1)

Figure 11: Typical Raman spectra of cimetidine in isopropanol. Excitation wavelength 647.12 nm,
Temperature 25 °C

28

0.5

0.4
a Solubility '3
9 03 (Form A)
8 ' '
2 .
.§ ,
3 °2 _ Solubility
a: (Form B)
a” 0.1

O I
1 2 3 4 5 6 7

Weight Percent Gmet idine

Figure 12: PIR’s of 675 cm-1 (cim) to 494 cm-1 (isoprop) as a function of cimetidine concentration in
isopropanol solutions. Solubilities are represented by ovals and calibration curve points are
represented by squares. Excitation wavelength 647.12 nm, Temperature 50 °C.

   

 

2.5
2 .. o
Solublllty (Form A)
’6
1’ 1.5
E Solubility (Form b)
9,
g 1
D.
0.5
0' n: lnlnanjlnnlnnalanal
0 2 4 6 8 10 12 14 16

Weight Percent Cimetidine

Figure 13: PIR’s of 675 cm-1 (cim) to 494 cm-1 (isoprop) as a function of cimetidine concentration in
1.75/1 (v/v) isopropanol-water solutions. Solubilities are represented by ovals and calibration
curve points are represented by squares. Excitation wavelength 647.12 nm, Temperature 25 °C.

29

Table 1: Conditions and results for phase transformation experiments done in isopropanol.

Temperature: 50 °C

Solvent: Isopropanol

Total Weight of Seeds: 0.30g

Total Volume of Solution Saturated w.r.t Form A: 45 ml

 

 

Complete Approximate Approximate
Transformation to Transformation Time to Reach
Weight % A Weight % B B? Time B Solubility
100 0 no ------------
75 25 no ------------
50 50 no ------------
25 75 yes 10 hours 24 hours
0 100 ------------- 15 hours

 

Table 2: Conditions and results for phase transformation experiments done in 1.75/1 (v/v)
‘ isopropanol-water.

Temperature: 25 °C

Solvent: 1.75/1 (v/v) Isopropanol/Water

Total Weight of Seeds: 0.30g

Total Volume of Solution Saturated w.r.t Form A: 45 ml

 

 

Complete Approximate Approximate
Transformation Transformation Time to Reach
Weight % A Weight % B to B? Time B Solubility

100 0 . no ------------
99 1 yes 16 hours 24 hours
75 25 yes 1.5 hours 10 hours
50 50 yes . 1.0 hour 10 hours
25 ‘ 75 yes 30 minutes 10 hours

0 100 ------------- 10 hours

30

 

 

 

 

1.2 105
, --------- 0 hours

I”: i: : —----28 hours 1
'E . Cr: . . -\ . =
3 1 10 "\J'"! "-....'J-.{h..«. ." ll
a “‘"~°-”-' \~..... 1' “I
N 9 10‘ “MM” \..\ _ l
h 0.0,:
4—0
25
h 4
s 8 10 l
U) 7 10 g E
c .-
£3 1
5 6 10‘ ' ' .

4 1 1 1 l 1 n L 1 l L .1....i.”1".l .1"...1"°°1'°"l"'.l "1.53""? l l
5 10 .

300 450 550 600 700
Raman Shift (cm-1)

Figure 14: Solid state Raman spectral changes during phase transformation of cimetidine. Excitation
wavelength 647.12 nm, Temperature 25 °C.

Two sets of experiments were run, one with isopropanol (T able 1) as the solvent,
and the other with 1.75/ 1.0 (v/v) isopropanol-water (Table 11). Each set consisted of trials
where the BIA ratio (w/w) varied from O to 100 percent, at 25% intervals. During the phase
transformation, system changes were observed in the solid and solution phases. The
solution PIR’s for both experimental sets are shown as a function of time in Figures 10
and 11, while typical spectral changes in the solid state are shown in Figure 9. The
solution PIR’s were used to quantitatively measure cimetidine concentration while the solid
state spectra were used to estimate the time of total phase transformation.

With isopropanol as the solvent, detectable phase transitions were only seen in the
slurry starting with 75%B/25% A (Figure 10). This particular experiment resulting in a
total transformation also had a change in solute concentration after the phase

transformation was complete. For the conditions and results given in Table I, there must

31

be a large percentage of form B present to achieve total phase transformation. During
phase transformation, the solute concentration remained at the saturation value of phase A
until all of phase A had transformed to B. These results are consistent with those of Sudo
et. a1,11 thus showing that the transformation is growth limited with respect to phase B.

The experiments with less than 75% B present in the initial slurry resulted in little
or no detectable phase transformation, along with little or no changes in solute
concentration. This is not to say that a phase transformation could not take place, but under
the given conditions a phase transformation was not detectable in the 36 hour period of
observation. Several reasons can be offered for this. The low solubility of cimetidine in
isopropanol could have allowed the kinetically favored form A to exist for long periods of
time in a slurry, despite the fact that the stable form B was present. The small difference in
solubility of form A and form B gave a relatively small driving force for phase
transformation. The dependence of total phase transformation upon the amount of form B
present in the slurry also suggests that a secondary nucleation mechanism may play a larger
role for a phase transformation in this system, since secondary nucleation has been Shown
to depend on solid phase content in a slurry.17

The constant or near constant solute concentration during phase transformation and
the very slow approach to equilibrium in the trial runs with 75% B or more give support for
a growth limited mechanism, which was discussed earlier and outlined in detail by Cardrew
and Davey,13 and also concluded by the Sudo et. a1.“

When a mixed solvent of 1.75/1 (isopropanol/water) was used, the results were
considerably different. Phase transformations occurred in all the experiments except the
one run without phase B present in the initial slurry (Figure l 1). The phase
transformations were much faster than in isopropanol. A full transformation occurred in

less than one hour for each experiment except for the slurry seeded with a 99% A/ 1% B

32

 

 

 

 

 

0.321:
' 0 Form A Solubility
0.31% 5 ' . o—
: o t
I- A
0.3";- g ‘ .
I o o
A 0.29-E . 00A, B
6 I A 25% B
3 0.28-E o 50% 3
‘2') l- . O
E O 27 I: '3 75°/o B
'9: ' E . 100% B
9; 0.26-E 0
°- 5 Form B Solubility o . o
0.25 _ .
0.24_WW+W

. 0 5 10 15 20 25 30 35 40
Time (hours)

Figure 15: Change in PIR’s of of 675 cm-1 (cim) to 494 cm-1 (isopropanol) during phase
transformation experiments in isopropanol. Excitation wavelength 647. 12 nm, temperature

50 °C.

PIR (Cim./lso)

33

 

 

 

 

 

 

2
1.9 Form A
2 solubility A
s .er a
e A o
1.8 0 0
8 x O 100% B Initial
D 75% B Initial
1'7 3 0 50% B initial
0 a o X 25% B Initial
1,5 . 0 1% B Initial
8 A 0% B Initial
Form B
1-5 solubiltiy Q a
. i—
1_4 1 1 1 1 1 L 1 1 1 IL 1 1 . 1 1 1 L 1 . 1 J 1 1 1 1
0 5 , 10 15 20 25

Time (Hours)

Figure 16: Change in PIR’s of of 675 cm-1 (cim) to 494 cm-1 (isopropanol) during phase

transformation experiments in 1.75 (v/v) isopropanol-water. Excitation wavelength
647.12 nm, temperature 25 °C.

34

mixture, where a full phase transformation occurred after 16 hours. The solute
concentration curves as a function of time show that this transformation appears to be
dissolution limited with respect to phase A.13 The higher overall solubility and larger
solubility differences between phases probably created a stronger driving force for phase
transformation, thus creating conditions for a much quicker phase transformation than that
seen with isopropanol as the solvent. Based on the shape of the solute concentration
profiles, transformation kinetics appear to be dissolution limited, as opposed to the growth

limited kinetics with isopropanol as the solvent.
2.4. In Situ Raman Spectra of Cimetidine Slurries

The largest advantage that Raman spectroscopy offers over more traditional
analytical methods is that it can be used to observe system changes in situ. This eliminates
the need to draw off sample aliquots, thus reducing the possibility of experimental error.
When applied to a crystal slurry, in situ measurements offer the ability to determine
conditions in the solid and the liquid phase simultaneously, while leaving the system intact.

In situ spectra of each phase in an agitated slurry of cimetidine saturated with
respect to each polymorph are shown in Figure 12 and 13. Spectral characteristics of both
the solid and liquid states can be seen in these graphs. The liquid state spectra of
isopropanol show bands at 494 cm", 425 cm", and 375 cm". The solid state spectra
show several cimetidine bands in the 600 to 700 cm'1 region, the 300 to 350 cm'1 region,
and the 100 to 200 cm‘1 region. These experiments are preliminary and demonstrate the
potential to use Raman spectroscopy as a tool for in situ analysis. Instrumentation
improvements need to be implemented for Raman spectroscopy to be used for on-line

analysis. Pallister et a1. has coupled holographic notch filters to a single monochrometer,

35

190000.0 ‘

180000.0 ‘

170000.01:

Cim Cim
160000.0 ‘

150000.01

Intensity (Arbitrary units)

 

 

140000.0 -
100 200 300 400 500 500 700 800

Raman Shift (cm-1)

Figure 17: In situ Raman spectra of cimetidine form A in isopropanol. Excitation wavelength 647.12
nm, Temperature 25 °C.

 

 

 

 

2.6105-
p-
5; ISO Cim
A 2.510E I50
(I)
3g r
2‘ 2.4105-
g - Cim
:5 I'
h " '
3 2.3105- Cim c'"
.e :
(D l.
5 2.2105-
E Z
1-
2.1105 lLllllllllllllllllllllleLlllAlAllll

O

100 200 300 400 500 600 700
Fiaman Shift (cm-1)

Figure 18: In situ Raman spectra of cimetidine form B in isopropanol. Excitation wavelength 647.12
nm, Temperature 25 °C.

36

increasing throughput by an order of magnitude over a Spex Triplemate while still
maintaining Rayleigh light rejectionmrl9 This configuration gives better signal to noise

ratios, thus reducing spectral accumulation times.
2.4. Summary and Conclusions

Raman spectroscopy can be used to investigate solvent-mediated phase
transformations, and also shows promise as a tool to study crystalline slurries in situ. A
Kr+ ion laser using an excitation of 647. 12 nm offers adequate signal to noise ratio while
eliminating the problems of fluorescence associated with excitation of lower wavelength.
The solid state spectra of cimetidine forms A and B show differences in the lattice vibration
regions. These spectral differences suggest differences in crystal structure, but assignment
of the lattice vibration bands for each polymorph must be done to confirm this. Spectra of
cimetidine in solution show a linear dependence of the PIR between cimetidine (675 cm")
and isopropanol (494 cm“) on concentration of cimetidine. These spectral characteristics
were used to observe the conditions during a solvent-mediated phase transformation for
cimetidine in both the solid and solution phases.

Experiments with isopropanol as the solvent showed that the transformation from
form A to form B is growth limited, and appears also to be dependent upon the density of
form B in the crystal slurry, suggesting that secondary nucleation is important in
isopropanol. Experiments with 1.75/1.0 (v/v) of isopropanol-water showed a
solvent-mediated phase transformation from form A to form B that was dissolution limited,
and occurred much faster than the transformations in isopropanol.

In situ Raman spectra of cimetidine slurries in isopropanol showed that spectral
lines of forms A and B in the solid state can be distinguished from spectral lines of

isopropanol in the liquid state. The instrumentation can be greatly improved by using

37
holograpic notch filters coupled to a single monochrometer. Once done, in situ Raman
spectroscopy should be excellent for further phase transformation and crystallization

studies, since it allows the observation of conditions without disruption of the system.

38

References

1. Solid State Chemists: 91 Drugs, 8. R. Bym (Academic Press, New York, 1982)
pp. 29-58, 79-102.

2. J. K. Haleblian, Journal of Pharmaceutical Sciences 64 (1975) 1269.

3. M. A. Heathman, Trends in Analytical Chemistry 8 (1989) 19.

4. P. Van Houng, Journal of Pharmaceutical and Biomedical Analysis 4( 1986) 811.

5.

CON

\0

n in i , A. T. Tu (John Wiley and Sons, Inc., New York,

1982) p. 7.

n t M. M. Sushchinski, (Israel Program for

. WWW;
Scientific Translations Ltd., New York, 1972) p. 83.
. M. Shibata, et. all, Journal of Pharmaceutical Sciences 72 (1983) 1436.

. S. Sudo, IK. Soto, and Y. Harano, Journal of Chemical Engineering of Japan 24
(1991) 237. ' ‘

. B. Hegediis and S. Gbrdg, Journal of Pharmaceutical and Biomedical Analysis 3
(1985) 303.

10. B. Prodic-Kojic et. al., Gazetta Chimica Italiana 109 (1979) 535.

11.

12.
13.

14.
15.

l6.
17.

18.
19.

S. Sudo, K. Soto, and Y. Harano, Journal of Chemical Engineering of Japan 24
(1991) 628.

A. M. Tudor, et. a1. SpectroChimica Acta 47A(1991) 1389.

P. T. Cardrew and R. J. Davey, Proceedings of the Royal Society of London
A398(1985) 415.

U. Bismayer, Phase Transitions 27 211 (1990) 211.

W. G. Turrell. (Academic Press. London.

1972) pp. 159-167.
E. F. Gross and M. F. Vuks, Nature 135 (1935) 100, 431.

P i r e , A. D. Randolph and M. A. Larson, (Academic
Press, New York, 1971) pp 104-107.

D. M. Pallister, et. al., Applied Spectroscopy, 46, (1992) 1469.

Kaiser Optical Product Literature, January 1995.

Chapter 3
The Use of Carminic Acid as a Trace Fluorescent Probe in

Aqueous Sugar Solutions

3.1 Introduction

The molecular interactions that occur in supersaturated solutions is an essential part
of understanding crystal nucleation and growth. Much work on understanding this
phenomena has been done in this area in recent years. Previous investigators have studied
the nature of molecular association and solute organization in supersaturated solutions
using a variety of techniques. Myerson"2 undertook diffusion studies in a variety of
electrolytes (NaCl, KCl) and non-electrolytes (urea, glycine, glycine-valine) indicating the
formation of aggregates in solution. Garside and Larson3'4 observed concentration
gradients in isothermal columns of supersaturated citric acid, potassium sulfate, urea and
sodium nitrate, which suggested solute structuring.

Attempts to characterize solute-solvent interactions in aqueous sugar solutions have
been carried out using several methods. Allen5 has done studies with sucrose similar to
Garside and Larson which showed that isothermal columns of supersaturated solutions
contained concentration gradients. Sugget6 has used NMR and dielectric relaxation
techniques to identify the interaction of water in sugar solutions, namely the presence of
associated water and bulk water. Richardson, et al.7 used 2H and 170 high field NMR to
show water mobility in sucrose solutions as a function of concentration. Similarly, I-Iills8
investigated the hydrogen exchange kinetics and water reorientation dynamics in aqueous
glucose solutions.

The use of extrinsic fluorescent probes gives another method to determine solution
structure. A recent investigation by Chakraborty and Berglund9 using trace amounts of
pyranine in aqueous sugar solutions showed that relative amounts of associated water and

bulk water could be detected by monitoring the relative amount of protonated and

39

4 0
deprotonated forms in solution. By relating the emission intensities of the protonated and
deprotonated forms, they showed that relative amounts of associated and bulk water can be
determined in a given solution. The use of pyranine has been demonstrated to work well
for the studies of water activity in aqueous sugar solutions but it is only a D and C
approved dye, thus it is not allowed for direct application in foods.

In this study we investigated the use of carminic acid, a naturally occurring, F, D
and C approved dye ‘0 as a substitute for pyranine in food applications. Its ground state
electronic spectrum is known to be influenced by pH “ and we found similar behavior in
the excited state. As can be seen from Figure 19, ‘0 carminic acid has a glucose group in its
structure, which may allow it to interact more easily with sugar molecules in an aqueous
solution. The glucose structure may also allow it to function as a tailor-made impurity,12
which will discussed in more detail below. These features suggest that carminic acid is a

suitable indicator of solution properties.

 

0 OH

Figure 19: Chemical structure of carminic acid

Previously, carminic acid has been used to determine trace quantities of Boron in
solution13r'4, and it has also been used in the spectrofluorimetric determination of several.
transition metal ions.15'16*‘7 Carminic acid has also been used as a staining agent in
histography.18 It has also received considerable attention in the food industry as a red
coloring agent.19 When complexed with aluminum, it forms carmine, which can give color

in the range from pale strawberry red to a dark blackcurrant red.20 Carminic acid

4 1
undergoes ester hydrolysis in alkali solutions,21 which accounts for its instability above

pH 7.5. Allevi and coworkers completed22 the first total synthesis of carminic acid.

3.2 Carminic acid as a tailor-made impurity

Carminic acid was also chosen because it is a glycoside, thus it may be able to
function as a tailor-made impurity. A tailor-made impurity as defined in literature”:23 is a
special class of additives designed to interact with selected crystalline faces in very specific
ways. These impurities have structural characteristics similar to the host molecule which
allow it to be more easily incorporated into the lattice during crystallization. The fact that
carminic acid has a glucose group makes it a candidate to function as a tailor-made
impurity12 in sucrose and glucose solutions. Generally, a tailor-made impurity is added to
inhibit crystal growth and/or change the crystal morphology. This is due to dissimilarities
between host and impurity that disrupt the normal crystallization process. In our study, we
were able to use a very small amount of acid and still obtain intense spectra. With low
probe concentration, (2.5 ppm) the change in crystal interface conditions would be kept to a
minimum, but the high emission intensity would still allow monitorization of the
crystallization process conditions at the crystal interface and the bulk solutions.
Crystallization of aqueous sugar solutions are currently being carried out by Racemis and

Blanchard.21

3.3 Materials and Methods

Absorption spectra were obtained using a Perkin Elmer Lambda 3A UV/V is
Spectrophotometer in conjunction with a Perkin Elmer R100A recorder. Emission Spectra

were obtained with a Spex 1681 FLUOROLOG 2 0.22m Spectrometer. Four sided

Suprasil quartz cuvettes with a 1 cm pathlength were used to contain the samples in both

42
absorption and emission. To aid in sugar dissolution, a Bransonic 220 sonic bath was
used. All solution pH measurements were taken with an Orion Research digital analyzer/
501 and a combination electrode.

Ultra pure grade RNase-free sucrose (Baker Chemicals), reagent grade anhydrous
glucose powder (Baker), carminic acid (Aldrich), and HPLC grade water (Fisher) were
used. A stock solution of 4 x 10'4 M carminic acid was prepared with spectroscopic grade
methanol (Malinckrodt). For emission fluorescence studies, solutions of 2.5 ppm
carminic acid were prepared. For absorption studies, solutions of 5 ppm were made.
Before each solution was prepared, the amount of carminic acid necessary for a given
concentration was calculated. The proper amount of stock solution was then calculated and
placed into a 20 ml vial with a precision pipette. The stock was allowed to evaporate, thus
giving the proper amount of probe. Each vial containing dry carminic acid was set aside
and kept away from light.

The sugar solutions were prepared to weigh exactly 10 grams. (i0.01%) The
sucrose samples varied in weight percent from 20% to 75%, with a saturation point of
67% at 25 C, whereas the glucose solutions varied from 30% to 75%, with a saturation
point of 48% at 25 C. The samples were prepared in 10% increments up to 60%. Above
60%, they were incremented by 5% in order follow changes in the supersaturated regions
more closely. In preparation it was absolutely necessary to ensure complete dissolution of
both sucrose and glucose solutions by heating and agitation.

After each solution was dissolved, it was quickly transferred into a vial with
pre-evaporated carminic acid stock solution, and allowed to cool to room temperature. To
insure homogeneous mixing of carminic acid, each cooled solution was reheated in the
water bath for one minute, and allowed to cool again to room temperature. Solutions that
were not reheated had visible and unwanted concentration gradients. Once the solutions

reached room temperature, spectrophotometric readings were taken.

4 3

Fluorescence emission of carminic acid in aqueous glucose and sucrose solutions
was measured using excitation wavelengths of 500 nm and 310 nm. The emission was
monitored from 530 nm to 700 nm for samples excited at 500 run, while the emission was
monitored from 400 nm to 700 nm for samples excited at 310 nm. Excitation and emission
slit widths of 2 mm were used for all spectra. For each sample, a background spectrum
was taken of the corresponding sugar-water solution. Samples were background corrected
by subtracting the fluorescence emission spectrum of each aqueous sugar solution .

Absorption spectra of carminic acid-doped sugar solutions was followed from 250
nm to 600 nm. For both sugars, a scanning speed of 120 urn/min. was used. The pH of

each solution was measured.
3.4 Results and Discussion

The absorption spectra of carminic acid in aqueous sugar solutions was determined
first to ascertain excitation wavelengths and also to determine the species that existed in the
ground state. The spectra of carminic acid in aqueous glucose solutions are shown in
Figure 20. Based on the work done by Schwing-Weil,ll the visible absorption spectra
indicate that carminic acid can exist in the LH3, LH2‘, and LH’2 states. In samples
containing less than 60% glucose, carminic acid exists as a mixture of LH'2 and LHZ'
states, whereas in solutions with concentrations higher than 60% glucose, carminic acid
exists in the LH3 species. Comparing these spectra to that of buffer solutions with similar
pH, it appears that visible absorption spectral changes in the ground state are a function of
glucose solution pH. The UV region of the absorption spectrum shows an equilibrium
between a species absorbing at 325 nm and a species absorbing at 280 nm that is a function
of glucose concentration. The assignment of these bands is as yet undetermined, but it is

being investigated.21

44

 

   
  
  

 

 

 

  

 

 

3.5 f r.f ' I I f I I I I I v I l I I 1’ v I u u u u I t v 1 t j rj r 1 4
3 20 °/o -‘
—' - - 40 % I
----- 70 % ‘
f: 2.5 ‘
D I
s 2 _:
a I
.9. .
a d
g' 1.5 1
m 1
g 1
‘ 1
0.5 .u

0

250 300 350 400 450 500 550 600

Wavelength (nm)

Figure 20: Absorption spectra of carminic acid in aqueous glucose solutions. Carminic acid concenuation
5 ppm, temperature 25 °C.

The spectra of carminic acid in aqueous sucrose solutions is shown in Figure 21.
There is little or no variance in the visible absorption spectra as a function of sucrose
concentration. In aqueous sucrose solutions, carminic acid exists in the L'3 and LH'2
states. When these spectra are compared to those of buffer solutions with comparable
pH’s, they appear to have spectra that is not dependent on pH. The fact that the sucrose
solution pH changes by two units. as a function of sucrose concentration with no
corresponding change in visible absorption spectra further exemplifies this. In a larger
sense, it implies that the carminic acid molecule in the ground state exists in one kind of
environment that is independent of sucrose concentration. Similar to carminic acid in
aqueous glucose solutions the UV absorption in aqueous sucrose solutions spectra show
an equilibrium between the species absorbing at 325 nm and the species at 280 run that is a
function of sucrose concentration.

The emission spectra of carminic acid in both sugars was done using excitation

wavelengths of 500 nm and 310 nm. These excitation wavelengths were chosen based on

45

 

 

 

 

 

 

1 .2

1

”‘- 0.3
:3
:3,

g 0.6
'5‘
3“

5 0.4

0.2

o

250 300 350 400 450 500 550 600

Wavelength (nm)

Figure 2]: Absorption Spectra of carminic acid in aqueous sucrose solutions. Carminic acid concentration
5 ppm, temperature 25 °C.

the UV-visible spectra of the sugars which showed isosbestic points at 310 nm and 500 nm
in the ground state.

Tentative band assignments of carminic acid in the excited state were made based on
the visible emission spectra in buffer solutions. In 1M HCl, the fully protonated state
(LH3*) emits at 594 nm. In 1M NaOH, the fully deprotonated state (L'3*) emits at 645
nm. It also appears that the LH2'* species emits at 610 nm and 450 nm, and the LH'2*
species emits at 580 nm. A full photochemical characterization of carminic acid is presently
being undertaken, but for the purposes of this study we used the fore-mentioned band '
assignments. Similar to pyranine in sugar solutions9, carminic acid indicates an equilibria
between different microenvironments which can be used to determine sugar concentration,
especially in the supersaturated region.

Emission spectra of carminic acid in solutions of varying glucose concentration

using an excitation of 500 nm are shown in Figure 22. The results show an exponential

46

increase in the peak intensity ratio (PIR) between 594 nm and 610 nm as glucose
concentration is increased and carminic acid concentration is held constant.

In Figure 23, the emission spectra of carminic acid in solutions of varying glucose
concentration using an excitation of 310 nm are shown. The PIR between 594 nm and 450
nm show a much stronger exponential dependence than that seen between 594 nm and 610
nm. There also appears to be a clear isosbestic point at 525 nm, indicating an equilibria
between two species of carminic acid as a function of glucose content.

According to Hill,8 there are two types of water that can exist in an aqueous
glucose solution. These are bulk water and solvation water. Bulk water refers to water
that is not hydrogen bonded to a glucose molecule. Solvation water refers to water that is
bound to a glucose molecule. The bulk water is capable of proton exchange. whereas the
solvation water is not. Using a context similar to Chakraborty and Berglund,9 we can
refer to a micro environment where there is either a presence or absence of exchangeable
protons in the vicinity of carminic acid. The presence of exchangeable protons would
cause carminic acid to exist as the LH'2* species, thus emitting at 610 or 450 nm, whereas
the absence of exchangeable protons would cause carminic acid to exist as the LH3*, thus
emitting at 594 nm.

In undersaturated aqueous glucose solutions at room temperature, Hill8 calculated
that there are 10 water molecules which surround a solvated glucose molecule. This model
assumes that each glucose hydroxyl group is hydrogen bonded by two water molecules in
solution. Since the solvation water is incapable of proton exchange, carminic acid residing
near the glucose molecule would exist in the LH3* state and thus emit at 594 nm. Carminic
acid residing in the bulk water would be in an area suitable for proton exchange, thus it
would emit in the LH’2* state of 610 nm or 450 nm.

In Figure 24, the PIR of 594 nm to 610 nm is shown. From this, it can be seen
that the PIR is a function of glucose Concentration, and thus a function of solvation and

bulk water content, where emission at 610 nm would represent solvation water and

47

 

1 l L
UtilullifITIIfi—Ijfili

‘

.l

d
--

 

 

 
 
  
  

 

75% Glucose
- - - 70% Glucose
......... O
1.5 106 - 60 A; Glucose _
- ' “-50 % glucose

 

 

110°-

............
......
I

Intensity (cps)

5105-

VIIIjITvaululliti

 

 

 

 

.,/-'
O l % 1 n 1 : n l n : n I l = 1 L n = n I u If: n n J. n l L : l L 1
540 560 580 600 620 640 660 680 ' 700

Wavelength (nm)

Figure 22: Emission spectra of carminic acid in aqueous glucose solutions. Carminic acid concentration
2.5 ppm, excitation wavelength 500 nm, temperature 25 °C.

 

 

 

 

 

   

 

 

 

1.210“ 7.....-“,....,.-.,-,._.,-fi..,..--‘
: — - 50 % Glucose I. \-. 1
1 10‘ :- ----- 60 % Glucose _: '--°~\ '1
. ——70%Glucose I . .
a 8105 '_ —----75%Glucose .- \1 L
g .
.33 6105 j
E 1
:50 4105 -
2105 -
0 {..L I l l A I l l l l l l l l 1 l kl I I I l I l I l A A}
400 450 500 550 600 650 700

Wavelength (nm)

Figure 23: Emission spectra of carminic acid in aqueous glucose solutions. Carminic acid concentration
2.5 ppm, excitation wavelength 310 nm, temperature 25 °C.

48

 

 

 

 

1.2PIIII[1'liriirr[1fifvtifillftfitjirrlq

:- 1

I- o d

1.15 f j

: j

. 9 «4

A 1.1 '- -l

o - .

8 E 1

F - o .

g 1 .05 f e j

a: Z Z
D—O

n. 1 j

1 _ o ‘

I o .

0.95 f 0 o I

: l

0.9 a L n J l 1 A A A l l A n 1 I n A n A l 1 n n A l n 1 4 n j n 1 1 J
20 30 40 50 60 7O 80 90
Weight Percent Glucose
Figure 24: PIR’s (594 nm/ 610 nm) of carminic acid emission in aqueous glucose solutions. Carminic

acid concentration 2.5 ppm, excitation wavelength 500 nm, temperature 25 °C.

 

 

16b-..-.....,#ff.,....,....,....,...
14'
E i
C: 12- O
o .
‘0 L
V!
E 10? o
3; b
19 F
a 8"
O4 - O
r
6L” 0
I o
4,WA—nnnlnnaJlnnnnlnkmmlnnn11.1.:
10 20 3O 4O 50 60 70
Weight Percent Glucose

v

ll'lllllllllllllllijfiL

 

 

on
C

Figure 25: PIR’s (594 nm/ 610 nm) of carminic acid emission in aqueous glucose solutions. Carminic
acid concentration 2.5 ppm, excitation wavelength 310 nm, temperature 25 °C.

49
emission at 594 nm would represent bulk water. At around 50% concentration, there is a
change in slope of the PIR. This occurs at about the saturation point of glucose in water at
25 °C. At this point, there is no longer enough water in solution for each respective
glucose molecule to be hydrogen bonded by 10 water molecules.

Figure 25 shows the PIR of 594 nm to 450 nm. This show a much stronger
dependence of PIR on glucose concentration. Similar to the above analysis, the emission
seen at 594 nm represents solvation water and emission seen at 450 nm represents bulk
water. There is also a change in slope of the PIR near the saturation point of glucose in
water.

Since carminic acid is a known pH indicator‘o'“, the pH's of each glucose solution
was taken. The pH readings were the same for samples with and without carminic acid.
The results given in Figure 6 show that the pH decreases linearly from 6.4 (30% glucose)
to 4.4. (75% glucose) Thus, it could be possible to assume that the change in emission
spectra with respect to glucose concentration could be a function of pH and not glucose
concentration. By comparing the emission spectra of the glucose solutions to emission
spectra of buffer solutions with similar pH's, spectral differences are seen between glucose
solutions and buffer solutions. Figure 27 shows the PIR’s (594 nm to 610 nm) of
carminic acid-doped glucose solutions along with the intensity ratios of buffer solutions
with similar pH's. From this, the PIR's are similar to the buffer solutions in the under
saturated regions, (up to 50% glucose) but once the solutions become supersaturated, their
PIR show a positive, exponential deviation fromthat of the buffer solutions with the same
pH. This deviation is probably due to the glucose content of the aqueous solutions, which
would decrease carminic acid's ability to exchange protons and exist in a deprotonated
state. The possibility of a pH effect was also investigated for solutions that were excited at
310 nm. Figure 28 shows the PIR’s (594 nm to 450 nm) of carminic acid-doped glucose

solutions along with the PIR’s of buffered carminic acid solutions of similar pH. From

50

 

I l
7fitIIIIfiIIjIIrIIIIII

9’
01
l

f

O}
1

pH 5.5 -

I

4.5 -

""IUTU‘IUi'tlr1U'rIIIIIU'UI
Innllnlnnlnlnnllnnllnlnll11L;
I

 

 

 

4 nannlnnhnlnnnAIIALnlnnnnllnnn
I I I

30 40 50 60 70

N
O
on
O

% Glucose (wlw)

Figure 26: pH of glucose solutions as a function of glucose concentration. Carminic acid concentration
2.5 ppm, temperature 25 °C.

this it is clear that in the supersaturated region there is no effect of pH, and that the spectral
changes seen are due to glucose content.

Previous work with sucrose has shown that there are two types of exchangeable
protons in aqueous sucrose solutions.6 This again would imply that there are two types of
micro environments in which the carminic acid can exist, one as a protonated species and
the other as a deprotonated species. Richardson et 01.7 showed through NMR water
mobility studies in aqueous sucrose solutions that samples up to 40 weight percent sucrose
contained water that was primarily free water, which is capable of proton exchange. At
about 40 weight percent, the water mobility drops sharply, thus causing the proton
exchange ability to decrease. In the first region, water would have the ability to exchange
protons, thus carminic acid would exist in the deprotonated state. In the second region,
there would be less likelihood for water to be able to exchange protons, thus, carminic acid

would be more likely to exist in the protonated state. Based upon this and the emission

51

 

 

 

 

 

 

1 .2 - I I I I I I I I I I I I I I I I I I I I I I —I I r I I I I I I I I I d
C 0 pm (594/610) of Glucose Solutions 0 I
1,15 : D PIR of Buffer solutions I
I o :
A 1.1 .b 2
O _ .
r4 _ .
§ - <> -
3 1.05 I o 2
s = =
1 : 0 1
' ° 0 pH=3.90 4
0.95 _' 0 0 pH=5.75 D :
. pH=4.35 .
o 9 I. I I I I I I I I I l I I J I I I I I I l I I I I l I I I I l I I I I d

20 30 4O 50 60 7O 80 90

Weight Percent Glucose

Figure 27: Comparison of carminic acid PIR’s in aqueous glucose solutions and buffer solutions.
Carminic acid concentration 2.5 ppm, excitation 500 nm, temperature 25 °C.

 

15 I I Ir'jTI I I I I I I I I II I I II I II III I I I I

I
0
d
1
14‘11113

..I

o

1
O

I

PIR (594 nm/ 450 nm)
01
, - .
o
o

pH=6.50 . pH=5.67 pH=4.15

 

 

IYTIIIII
LJAA+AALA

D

h

 

O I I I A I A I I I I L I I A I . ADA +1 4 . I I I I I I A I A A A
10 20 30 4O 50 60 70
Weight Percent Glucose

on
0

Figure 28: Comparison of carminic acid PIR’s in aqueous glucose solutions and buffer solutions.
Carminic acid concentration. 2.5 ppm, excitation 310 nm, temperature 25 °C.

52

spectra of carminic acid in aqueous glucose solutions results discussed earlier, it appeared
that carminic acid-doped sucrose solutions would give similar results to that of glucose.

The emission spectra of carminic acid-doped aqueous sucrose solutions are
different than those seen with carminic acid in aqueous glucose solutions. When excited at
500 nm, carminic acid emits in the LH2'* state or the LH‘2* state, whereas solution excited
at 310 nm emit from the LH3* state or the LH‘2* state. The sucrose emission spectra are
given in Figure 29 and 30. Figure 29 shows spectra of carminic acid in aqueous 'sucrose
solutions excited at 500 nm whereas Figure 30 shows spectra of carminic acid in aqueous
sucrose solutions excited at 310 nm. The solutions excited at 500 nm show species that
emit at 580 nm and 610 nm while the solutions excited at 310 nm show species that emit at
600 nm and 450 nm. Samples excited at 500 nm gave an increase in intensity with respect
to increasing sucrose concentration whereas samples excited at 310 nm showed large
changes in spectral characteristics as a function of sucrose content.

Figure 31 shows that there is very little change in fluorescence PIR between 580
nm and 610 nm (excitation 500 nm). Based on this, the amount of LH2'* and LH‘2*
species in solution remains the same throughout the concentration range studied when an
excitation wavelength of 500 nm is used. Figure 32 gives the FIR between 594 nm and
450 nm when an excitation of 310 nm is used. Assuming that the emission at 594 nm
represents the LH3* form of carminic acid and the emission at 450 nm represents LH2'*
form of carminic acid, then Figure 32 illustrates how the lack of bulk water forces carminic
acid to exist in the fully protonated form as the sucrose concentration is increased. As with
glucose, the possibility of a pH effect on the emission spectra of sucrose solutions was
investigated. The pH's of each respective sucrose solution were taken, and are shown in
Figure 33. They range from 8.6 (20% sucrose) to 6.5. (75% sucrose) Figure 34
compares the PIR's of sucrose solutions to that of buffer solutions with similar pH's when

an excitation of 500 nm was used. From this it can be seen that the sucrose solutions have

 

 

l l l I l
IIIIIII'IIIIIIIIIIJIIIL

 

 

 

—-75°/o Sucrose
5 - - -70% Sucrose
4 10 ......... 60% Sucrose
- - - - -50% Sucrose
I“;
3 3 105
>1 _, — I" - ~—
\
3 2 105 / ‘ -
.5 ’ e
/ \ \
/ .. \ x ‘
1 105 I..-'..’.’ ............... ~ ..... ~.- 'o...... .......
/..‘ .’ '~.__‘.~
xv“; ’ " "
0I.’IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
I I I I
540 560 580 600 620 640 660 680 700

Wavelength (nm)

Figure 29: Emission spectra of carminic acid in aqueous sucrose solutions. Carminic acid concentration
2.5 ppm, excitation wavelength 500 nm, temperature 25 °C.

 

 

 

4105-....,...................,....

35105 75%sucrose _:
3105 , ‘\ ‘1

,5 I 70%sucrose :
5 ' s _

g. 2.510 ‘ :
>. Z
:5 2105 \ 1
g /¢.\ 60%sucrose ‘ Z
r: 1.5105 ,'/-~ ‘ .-"'~-. \ 1

H II ‘ I ‘.

’2 \“ ’/ —‘ \.§ ‘ :1

1105 . ~ ~ .' \-

, V‘ / 50%sucrose \ x. -

4 ~\ I \-.:

510‘ V" ‘4 '3

O I j__l L l l l l l l l l l l I l l l 1 l j l j J I l A l l d
400 450 500 550 600 650 700

Wavelength(nm)

Figure 30: Emission spectra of carminic acid in aqueous sucrose solutions. Carminic acid
concentration 2.5 ppm, excitation wavelength 310 nm, temperature 25 °C.

54

 

 

 

 

1.3 I. I I I f: I I I I = I I I I : I W I I : I I I I % I I I I : I I I I ‘
1.25 ._. ..-
p I
a 1-2 ‘r it
3 t :
E 1 15 3 1-
{O ' I- II
V + 0 II
a: r o o o 1
H F' O O O "
9.. 1.1 -g :-
1.05 -E 1..
1 - L I I I l I I I I = I I I I = I I I I : I j_1 I_l Ll #J : I L. L.
10 20 30 4O 50 60 7O 80
Wt% Sucrose

Figure 31: PIR’s (610 nm/ 580 nm) of carminic acid emission in aqueous sucrose solutions. Carminic
acid concentration 2.5 ppm, excitation wavelength 500 nm, temperature 25 °C.

 

 

 

 

8 b ' U ' I I' V V l I t i I t I l t t V I r T 1 l V ‘

b d

: o 1

7 r 1

’5‘ 6 f 1

E I

O I- -I

In 5 P 1

V r- l

I O I

4 f 1

1- n

3 * ‘

P .1

B 3 I' 1

a I I

Q: 2 - -‘

. O I

1 .' o o '2

o o I

o A L L A l A A A A l A A A A I A A A A L A A A A J 4 A A A l A A A A ‘

10 20 30 40 50 60 7O 80
Weight Percent Sucrose

Figure 32: PIR’s (594 nm/ 450 nm) of carminic acid emission in aqueous sucrose solutions. Carminic
acid concentration 2.5 ppm, excitation wavelength 310 nm, temperature 25 °C.

55

 

 

 

 

 

9 I I I I {—I I II I I I I I I I I I I%I I rfi I I I Ill I I I I
8.5 ...: :-
8 III- cu—
p «1
pH I .
p 1
7.5 -: 1*-
. J
i- q
7 “r 1"
6.5 I4 I I :4 I I4'LI I I I : I I I I = I I IITI I I I I : I tel
10 20 30 40 50 60 70 80

Weight Percent Sucrose

Figure 33: pH of glucose solutions as a function of glucose concentration. Carminic acid concentration
2.5 ppm, temperature 25 °C.

56

 

 

 

 

1.3 I. I I I T % I I I I : I I I I : I Ifi r: I I I I i r I Ii : I rI W i
1.25 -: :-
8 1.2 "if '2"
w " .1
8 1.15 _ o .
. o o 1
E - ° 0 o o 1
1.1 'l" I-
I a I
. _ a ..
_;_ pH‘7'75 pH=7.50 a ;_
1.05 . H-8 75 pH=6.96 .
t 1? " ' pH=8.14 :
1 I I I I l I I I I : I I I I ILI I I I : I I I I = I IJ L: I I I I -
10 20 30 ‘ 40 50 60 7O 80
Wt%Sucrose

Figure 34: Comparison of carminic acid PIR’s in aqueous sucrose solutions and buffer solutions.
Carminic acid concentration 2.5 ppm, excitation wavelength 500 nm, temperature 25 °C.

 

 

 

 

8 b T r I I I I I I I I I I I I I I I fit I W T T I I I I I I I I I I I j
E o 1
7 f 1
I 1
E 6 Z‘ 1
i 1
' 1
no: 5 f 1
V‘ L .
a = . r
4 :- '1
s E 1
53 3 , 1
o. E, j
2 : ~ 0 I
" J
, ph 3.30 pH 7.74 pH 6.50.
1 T o D D 0 Cl 1
E o O 1

O I I I I l I I I I l I I 4 I L I I I . I I I I I l I I I . l I I I I

10 20 3O 40 50 60 7O 80
Weight Percent Sucrose

Figure 35: Comparison of carminic acid PIR’s in aqueous sucrose solutions and buffer solutions.
Carminic acid concentration 2.5 ppm, excitation wavelength 310 nm, temperature 25 °C

57
different PIR's than that of buffer solutions with similar pH's, thus showing that spectra
seen with sucrose solutions is a result of sucrose content and not solution pH. The same
analysis done on solutions excited at 310 nm (Figure 35) shows that there is no dependence
of PIR on pH, thus it appears that in both cases, the spectral changes are dependent on

sucrose concentration
3.5 Summary and Conclusions

Trace amounts of carminic acid (2.5 ppm) can be used as a fluorescent probe to
measure solute concentration in aqueous glucose and sucrose solutions in a manner similar
to that done with pyranine. The possibility of a pH effect was considered, with spectral
differences being attributed to solution sugar content. An excitation of 500 nm and 310 nm
was used for each set of solutions.

In aqueous glucose solutions, an excitation of either 500 nm or 310 nm can be
used. For an excitation of 310 nm, the PIR of carminic acid between 594 nm (LH3*) and
610 nm (LH2'* ) shows an exponential dependence upon glucose concentration. Along
with this, the fluorescence intensity increases as a function of glucose concentration. For
an excitation of 310 nm, the PIR between 594 nm (LH3*) and 450 nm (LH2’* ) shows a
very strong exponential dependence upon glucose concentration. In aqueous sucrose
solutions an excitation of 500 nm shows little or no change in PIR between 610 nm
(LH2'*) and 580 nm (LH'2*) as a function of sucrose concentration, whereas an
excitation of 310 nm shows a strong dependence of the PIR between 594 nm (LH3*) to
450 nm (LH2'* ) on sucrose concentration.

The visible absorption spectra of carminic acid in aqueous glucose solution
indicates that carminic acid exists in the LHZ- and LH'2 states in undersaturated solutions
and in the LH3 state in supersaturated solutions. The visible absorption spectra of carminic

acid in aqueous sucrose solutions indicates that carminic acid exists in the LH'2 and 1:3

58
with no change as a function of concentration. Ultraviolet absorption spectra showed and
equilibrium between species at 325 nm and 280 nm for aqueous solutions of both sugars,
but these species are yet to be identified.

Carminic acid appears to be a good candidate to determine glucose and sucrose
concentration as a trace fluorescent probe. It has a high fluorescence intensity at low
concentrations (2.5 ppm) making it easy to use, and a benign effect upon ingestion, making
it acceptable for use in food and drugs. More work, both theoretical and experimental

needs to be done to verify exactly how and why carminic acid acts in sugar solutions.

59

References

1. A. s. Myerson and 1..s. Sorrel, AIChe Journal, 28, (1982) 778.
2. A. S. Myerson and Y.C. Chang in: Industrial Crystallization, ‘84, Ed. S. J. Jancic

and E. J. Jong (Elsevier, Amsterdam, 1984) p. 27.

3. M. A. Larson and J. Garside, Chemical Engineering Science , 41, (1986) 1285.

4. M. A. Larson and J. Garside, Journal of Crystal Growth, 76, (1986) 88.

5. A. T. Allen et. Al, Nature, 235, (1972) 236.

6. A. Suggett and A. H. Clark, Journal of Solution Chemistry, 5, (1976) 1.

7. S. J. Richardson, I. C. Baianu, and M. P. Steinberg, Journal of Food Science,

52, (1987) 806.

8. B. P. Hills, Molecular Physics, 72, (1991) 1099.

9. R. Chakraborty and K. A. Berglund, Journal of Crystal Growth, 130, (1993) 587.

10. mm, Eds. S. Budavari, M. J. O’Neil, A. Smith and P. E.

11.
12.

13.
14.
15.
16.

17.

18.

19.
20.
21.

22.

Heckelman (Merck & Co. Inc., 1989) p 280.
M. J. Schwing-Weill, Analusis, 14, (1986) 290.

S. N. Black, R. J. Davey, and M. Halcrow, Journal of Crystal Growth, 79,
(1986) 765.

R. C. Calkins and V. A. Stenger, Analytical Chemistry, 28, (1956) 399.
J. Anarez, A. Ferrer, J. M. Rabadan and L. Marco, Talanta, 32, (1985) 1156.
G. F. Kirkbright, T. S. West, and C. Woodward, Talanta, 13, (1966) 1637.

Poluektov, R. S. Lauer, and M. A. Sandau, Zhumal Analiticheskoi Khimii,
25, (1970) 2118.

A. T. Pilipenko and T. L. Shevchenko, Zhumal Analiticheskoi Khimii, 7, (1989)
1262.

J. C. Stockert, A. R. Llorente, P. Del Castillo, and G. Gomez, Stain Technology, 65,
(1990) 299.

Riboh, M., Food Engineering, (1977) 66.
A. G. Lloyd, Food Chemistry, 5,(l980) 91.
J. Racemis and G. Blanchard, personal communication.

Allevi, P. et al., Journal of the Chemical Society Communications, (1991)
1319.

6O

23. D. L. Klug, in: Handbook of Indnstn'al Crystallization, Ed. A. S. Myerson
(Butterworth-Heinemann, 1993) p. 82.

Chapter Four
Future Work
4.1 Assignment of the fundamental lattice Raman bands for cimetidine
crystals

Tudor et. al 1 has made tentative assignments for cimetidine in the fingerprint region
of the Raman spectrum, but there have been no assignments made of cimetidine in the
lattice vibration region. Turrell2 gives a detailed description, showing how various
fundamental lattice vibrations can be attributed to librational motions around the x, y, and z
axes.

The first step in determining fundamental lattice vibration assignments would be to
obtain the crystalline structure of each respective phase through x-ray crystallography. The
crystalline structure of ciimetidine form A has been determined to be P21 ,0 by several
authors3'4’5, but the crystalline structure of form B has not been published. In order to
determine the crystalline structure of form B, it would be preferable to grow a form B
crystal of suitable size6 and then determine the lattice parameters. An x-ray powder
diffraction could also be done, but x-ray crystallography would be preferred.

Large crystals of form B have proven difficult to grow”, but it appears that the
best solvent mix would be approximately 1.75/1 (v/v) isopropanol saturated with respect to
form B and undersaturated with respect to form A, where the overall solubility of
cimetidine is quite high relative to that in other solvent systems. The high solubility of
cimetidine in these solutions provides a better driving force for crystallization, especially
‘ for the crystallization of the thermodynamically stable form B. It would be best to seed the
fore-mentioned solution with a form B crystals in an unagitated solution, thus reducing
secondary nucleation of the stable phase. The optimal temperature of crystallization still
needs to be determined, but it appears the higher, the better since solubility increases with
increasing temperature. Transformation at higher crystallization temperatures should not be

a problem since cimetidine is monotropic with respect to forms A and B.9

61

62

. Once large crystals are grown and the x-ray crystal structure is determined, the
fundamental Raman band assignments can be made. For all crystalline structures,‘there are
tables of Raman polarizability tensors calculatedlo which can mathematically predict the
polarization of incident and scattered light giving rise to the respective Raman scattering
bands. By obtaining Raman spectra of cimetidine at various incident and scattering
orientations, the bands can be assigned with respect to unit cell and molecular symmetry.
The possibility of Raman spectra arising from multiphonon processes should also be I
considered, but it would be difficult to actually assign bands since these phenomena have

proven to be quite complicated for all but the simplest of molecules.11
4.2 Raman spectral investigations of other cimetidine forms

There have been seven different crystalline forms of cimetidinel2 discovered since
1976, with each exhibiting different infrared spectra. From the work of Tudor et al.1, it
has been shown that cimetidine exhibits different Raman spectral characteristics in the
fingerprint regions between forms A, B, and C, but the spectra of these forms have not
been investigated in the lattice vibration region. A more comprehensive understanding of
the Raman spectra of the other known forms of cimetidine should be done to complete the
basic research done up to this point. This would provide background information which
could be used to investigate the possibility of other phase transformations and/or

crystallization conditions.

4.3 Cimetidine crystallization experiments

The work done heretofore has shoWn that phase transformations can be observed
using Raman spectroscopy, but crystallization from a seeded, supersaturated solution has
yet to be investigated. Sudo et al.8 has used more traditional methods to observe the

crystallization of cimetidine under different conditions of seeding, supersaturation, and

6 3
temperature, but they have not investigated the effect of different solvents. As mentioned
previously, the capability Raman spectroscopy to obtain measurements in situ presents a
great potential to determine crystallization parameters with little or no disruption of the
system under study. In situ Raman spectra 0f cimetidine slurries can be greatly improved
through better instrumentation, which will be discussed below. Once the in situ Raman
spectroscopy of cimetidine slurries is perfected, the experiments of Sudo et. a1 should be
repeated using Raman spectroscopy in situ and the results compared to verify the use of
this technique as a plausible substitute for more traditional methods. After this,
crystallization experiments should be carried out in at least two solvents other than
isopropanol, preferably one with lower cimetidine solubility and one with higher cimetidine
solubility. Two logical solvents for these experiments would be water (lower solubility)
and water-isopropanol (higher solubility). Several parameters should be investigated. The
effect of supersaturation, seed crystal form, agitation rate, and temperature are all variables
that can be considered. The possibility of solvent-mediated phase transformations during
crystallization experiments should also be examined, especially if a mixture of seed crystal

forms is used.

4.4 Improvements in Instrumentation of Raman Spectroscopy

All the Raman spectra have been obtained using a Spex Triplemate 1877 coupled to
an EG & G CCD, red-enhanced detector. In the past this instrumental configuration has
proven to be quite useful in obtaining high resolution Raman spectra with good Rayleigh
line rejection, but the majority of these experiments have been done using blue excitation,
where the Raman efficiency is much better than experiments done with visible red or
infrared excitation. The triplemate configuration consists of three monochrometers and
three dispersion gratings. For each grating, a loss of 50% throughput oCcurs, thus the

signal that actually gets to the detector is reduced significantly by just the gratings alone,

6 4
not to mention many of the other optics that are downstream from the sample position.

When using visible red excitation, this can greatly diminish the signal to noise ratio since

the loss from the gratings is coupled to the v4 dependence of Raman scattering on excitation

frequency.

Considerable improvements in instrumentation for Raman spectroscopy have
potential to greatly improve the signal to noise ratio, thus allowing for higher quality
spectra and reduced power needs. The use of holographic notch filters coupled to a single
dispersion grating and CCD detector gives throughput up to an order of magnitude higher
than that of a triple monochrometer while still attenuating Rayleigh scatter enough to obtain
Raman spectra as low as 64 cm“. Pallister et al.15 has demonstrated the use of this new
technology by coupling holographic notch filters and beam splitters to a Raman
rnicroprobe, but this equipment could be modified for a conventional macroscopic sample
chamber.l4 By using this technology, the beam could be directed to a batch reactor with a
crystallizing slurry in a backscattering configuration, allowing for spectra to be detected
from both phases during the process. The high throughput would allow for much shorter
accumulation times, thus giving data points which would be more accurate as a function of
time.

Another possibility is the use of a fiber optic coupled to holographic notch filters
and a single dispersion grating. The fiber optic probe could be used as a remote sensing
device that could be dipped directly into a crystalline slurry. The use of a configuration
similar to this has been demonstrated for the emulsion polymerization of polystyrene.15
The authors obtained Raman spectra of both the solid phase polystyrene and the liquid
phase styrene monomer. They were able to follow the disappearance of the monomer
phase and the appearance of the solid polymer as a function of time. These spectra were
obtained in the fingerprint region with an excitation of 514.5 nm, but it appears feasible that
the experiment could be applied to crystalline slurries using visible red excitation to obtain

Raman spectra in the lattice vibration region.

65
4.5 Carminic Acid

Studies should be undertaken to better understand the potential use of carminic acid
as a trace fluorescent probe in aqueous food systems. The work done so far has been
limited to studies in sucrose and glucose solutions, thus work needs to be done in other
sugars and possibly other food systems to give a more complete picture of canninic acid’s
use as a trace fluorescent probe.

Experiments are necessary to assign fluorescence bands in both the UV and visible
regions to functional groups on the molecule. Similar analysis has been done for other
benzyl-anthraquinones,“5'l7 but carminic acid has not been characterized. This would
involve starting with a molecule of known fluorescence band assignments, such as
anthraquinone, and adding functional groups until the carminic acid molecule is completely
synthesized. After each functional group addition, the fluorescence spectra could be taken,
and the emission bands assigned. In this way, each successive fluorescence band that
arises could be accounted for by equating the spectral changes to the newly added
functional group. A study such as this would be very challenging since it would involve
expertise in both fluorescence spectroscopy and organic chemistry.

Several different protonated and deprotonated species that exhibit different spectra,
but a detailed analysis of the fluorescence characteristics should be done to better
understand what is happening in the solute-solvent system being probed. Determination of
excited state pKa’s is necessary to better understand the effects of pH on fluorescence
spectra. By using the Forster cycle”, the excited state pKa’s can be calculated with the
shift between. the absorption and fluorescence peaks. Similar studies have been done on
mono and bi-substituted phenols, but these compounds have well resolved fluorescence

spectra. The challenge lies in the deconvolution and assignment of the fluorescence bands.

66 _

Other FDA approved substances that are chemically similar to carminic acid offer
potential as trace fluorescent probes in sugar solutions”. Kermesic acid20 has the same
chemical structure as carminic acid except it has a hydrogen group in place of the glucose
group. Its violet-red color suggests that visible excitation may result in fluorescence
spectra similar to carminic acid. Some of the carmine lakes also show promise as trace
fluorescent probes provided that the system under study is free of transition metal ion

impurities.
4.6 Crystallization Studies of Carminic Acid Doped sugar solutions

The value of a trace fluorescent probe lies in its ability to monitor system parameters
with minimal effects on the system, thus it is important to understand these potential
effects. Carminic acid can be used to study the crystallization of sugars in situ, but the
effect of this probe on crystallization kinetics and the potential for carminic acid to function
as a tailor-made impurity should be studied first. The glucose group in carminic acid’s
structure makes it a tailor-made impurity,21 which could slow the crystallization kinetics,
change the crystal morphology, and increase or decrease the solubility of the sugars in
aqueous solution.

By varying the amount of carminic acid in aqueous sugar solutions during
crystallization, kinetics, crystal morphology, andinclusion can be studied as a function of
probe concentration. These studies should be done in both unagitated and agitated
solutions to factor in the effectof secondary nucleation. Seeded and unseeded solutions
also need to be looked at. Once carminic acid’s effect on aqueous sugar solution
crystallization are known, it can be used to study aqueous sugar solution conditions during
crystallization by using a probe concentration that minimally affects the system while still
allowing for monitorization of the system. Experiments with carminic acid-doped sugar

solutions would allow for observation of solution side conditions during crystallization.

67

An excitation wavelength of 310 nm would give the best results since spectral changes are

more pronounced when this excitation is used.

68

References

1.

l9

>1

oo

0

10. W. G. Fateley, et. al.

11.

12.
l3.
14.
15.

16.

17.

18.

19.

20.

A. M. Tudor, et. al. SpectroChimica Acta, 47A, (1991) 1389.

al , G. Turrell, (Academic Press, London,
1972), pp. 159-167.

B. Prodic-Kojic et. al., Gazetta Chimica Italiana, 109, (1979) 535.
M. Shibata, et. al, Journal of Pharmaceutical Sciences, 72, (1983) 1436.
E. Haddicke, F. Frickel, and A. Franke, Chem. Ber., 111, (1978) 3222.

J. W. Jeffery WW (London, New York, Academic Press,
1971) pp. 1-3.

B. Hegediis and S. Gorog, Journal of Pharmaceutical and Biomedical Analysis, 3,
(1985) 303..

. S. Sudo, K. Soto, and Y. Harano, Journal of Chemical Engineering of Japan, 24,
(1991) 237.

..S Sudo, et. al., Memoirs of the Faculty of Engineering, Osaka City University, 31,
(1990) 61-67.

 

M118, (Wiley Interscience, New11 York, 1972) Appendix I and II -- --

G. Turrell, Waging, (Academic Press, New York,
1972) Chapter 5 and 6.

British Patent Specification, No. 1543238 (1976)
D. M. Pallister, et. al., Applied Spectroscopy, 46, 1469 (1992).
Kaiser Optical Product Literature, January 1995.

C. Wang, T. J. Vickers, J. B. Schlenoff, and C. K. Mann, Applied Spectroscopy,
46,(1992) 1729.

N. V. Platonova, K. R. Popov, and L. V. Smimov, Optical Spectroscopy, 34,
(1969) 197.

D. K. Palitm, T. Mukherjee, and J. P. Mittal, Journal of the Indian Chemical Society,
63, (1986) 35.

E. L. Wehry, and L. B. Rogers, Journal of the American Chemical Society, 87,
(1965) 4234.

D. D. Gadgil, A. V. Rama Rao, and K. Venkataraman, Tetrahedron Letters, 18,
(1968) 2223.

O. Dimroth, Berichte der Deutschen Chemichen Gesellschaft, 43, (1910) 1387.

69

21. S. N. Black, R. J. Davey, and M. Halcrow, Journal of Crystal Growth, 79,
(1986) 765.

APPENDIX

Original Data from Scatter Plots

APPENDIX

Original Data from Scatter Plots

Table III: Original Data from Figure 12

 

0.00 0.0013 solubility A: Weight %=4.96, PIR=0.289
1.00 0.0630 solubility B: Weight %=3.77, PIR=0.217
2.00 0.128
3.00 0.175
4.00 0.242
5.00 0.294
6.00 0.378

 

Table IV: Original Data from Figure 13

 

WWW-4-

0.00 0.00 solubility A: Weight %=l4.2, PIR=0.1.69
2.50 0.329 solubility B: Weight %=10.8, PIR=0.1.39
5.00 0.600

7.50 0.955

10.0 1.24

12.5 1.59

15.0 2.02

 

 

Table V: Original Data from Figure 15 (All PIR’s 675 cm-l/494 cm-l)
PIR . PIR PIR PIR PIR

I. [I ] 0% B . °|i I 25% B . .l. I 59% B . .l. I 25% B i .I. I 11107 B
0.00 0.306 0.314 0.314 0.304 0.314
4.00 0.315 0.300 0.300 0.298 0.294
10.0 0.306 0.304 0.294 0.310 0.275
20.0 0.310 0.298 0.304 0.276 0.254
28.0 0.304 0.306 0.308 0.258 0.252
36.0 0.298 0.302 0.310 0.254 0.248

 

70

71

Table VI: Original Data from Figure 16 (All PIR’s 675 cm-ll494 cm-l)

 

Time PIR PIR ' PIR PIR PIR PIR
1.1.1111 llfiflnl'l 173.11 ZEfiB’TISQqB'TIZSVEHI'IIW
0.00 1.90 1.87 1.87 1.83 1.89 1.86
0.500 1.85 1.84 1.85 1.84 1.84 1.83
1.00 1.87 1.89 1.84 1.78 1.81 1.79
2.00 1.83 1.83 1.75 1.71 1.69 1.65
3.00 1.81 1.81 1.63 1.65 1.63 1.60
6.00 1.83 1.85 ' 1.54 1.56 1.57 1.55
10.0 1.85 1.83 1.49 1.51 1.50 1.50
16.0 1.84 1.63 1.49 1.48 1.49 1.45

24.0 1.89 1.49 1.47 1.46 1.48 1.45

 

Table VII: Original Data from Figure 24

 

Wm mm
30.0 0.949
40.0 0.948
50.0 0.973
60.0 0.994
65.0 1.036
, 70.0 1.062
75.0 1.1 16
80.0 ' 1.167

 

Table VIII: Original Data from Figure 25

 

Wilma: W
20.0 4.10
30.0 5.08
40.0 5.98
50.0 ‘ 7.16
60.0 9.75
70.0 12.1

80.0 14.9

 

72

Table IX: Original Data from Figure 26

 

W 211
30.0 6.35
40.0 5.95
50.0 5.45
60.0 5.05
65.0 4.90
70.0 4.65
75.0 4.38

 

Table X: Original Data from Figure 27

 

W88 Wm) w W
30.0 0.949 6.20 0.964

40.0 0.948 5.75 0.961

50.0 0.973 4.90 0.970

60.0 0.994 4.35 0.950

65.0 1.036

70.0 1 .062

80.0 1. 167

 

Table XI: Original Data from Figure 28

 

M12928 Wm) 111111213111 Wm.)
20.0 4.10 6.50 0.711

30.0 5.08 5.67 0.547

40.0 5.98 4.15 0.542

50.0 7.16

60.0 9.75

70.0 12.1

80.0 14.9

 

73

Table XII: Original Data from Figure 31

 

W W
20.0 1.12
30.0 1.11
40.0 1.11
50.0 1.12
60.0 1.14
70.0 1.12
75.0 1.11

 

Table XIII: Original Data from Figure 32

 

_1_2_S_u£r_Qs.eW 9’ W
20.0 0.750

30.0 0.390

40.0 0.438

50.0 0.868

60.0 1.70

70.0 4.28

75.0 7.20

 

Table XIV: Original Data from Figure 33

 

Mantras: 2H

20.0 8.65
30.0 8.35
40.0 8.03
50.0 7.75
60.0 7.35
70.0 6.95

75.0 6.51

 

74
Table XV: Original Data from Figure 34

 

 

mm mm fiufleufl Wm.)
20.0 1 . 12 8.75 1.03

30.0 1.11 8.14 1.05

40.0 1.1 l 7.75 1.09

50.0 1.12 7.50 1.07

60.0 1.14 6.96 1.06

70.0 1.12

75.0 1.1 1

 

Table XVI: Original Data from Figure 35

 

 

Menus: mm 1111M Wm)
20.0 0.750 8.30 0.671

30.0 0.390 7.74 0.847

40.0 0.438 6.50 0.875

50.0 0.868

60.0 1 .70

70.0 4.28

75.0 7.20

 

MICHIGAN srnr: UNIV. LrsRnRIEs
1111111111 1111111111111111111111111|
31293015721917