4. .uul- .-

 

‘_II IA'A.‘ nL

m‘um‘yg. 4 . wnxtg smszxg viva;
4' w 4 ., ' :I’ w 4. 2'44: fig." .2:
‘ '44.}. I“ 3.35% 2;: saggy: 9;.- imm
, u .
WNW

't-wlifdi ,
"i

" w"

4 t.
r“! .1

73.1 _
aw

4:44-32”,

"I Wu!
(1‘.

‘
I 1;; ”m/irrifl- '
- 4

if
.flbh.?

A"? J I
m‘flfrff :44:
7 ""f m figfil.”

1
. H ,
Jpn-u In"

,
. 3 . .
1455,41 :3“
4
. ,I 4,4

I

2"?
'JII'IJ. Bfiflg” C ‘
. l

'J‘II

J
W." a: *'
$4221; J .
' "W2.

., -¢’;1-1.
‘ ';l-' 04‘
Iii; I v ‘r"fr/-y,""1.""5‘afi

guru-r"

mi
4:324.
rw J:
J {5%, '
-‘?‘ ' I'é‘zf.

"'- U

. .I',

- «fit-Irv»?
W: - ' ""‘ ‘ Jarw I’Z‘I-T'

H, (’1’: L n I @jjfl’fl "if:

, I as? ' .

1);: 'I}
' "L‘Vr’E‘J‘

2

(TH?
\ 4:4

I' .3 .61
.414?“
’/ 3,55
"Mf;
‘!

‘h

a: 3-

a.
.‘

:\ 4
«1* u
11%;.

-- Ma)

23‘ '
34$: 94%”.
4" ‘ 4";

\-

I
r {-3- -
I 'f‘ni

r. g‘
3!:

ant:

flay:

'vuy

.. .,

 

 

 

."- rim“:
«1%:

940‘19 79679

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

THE FORMATION OF DIFRUCTOSE DIANHYDRIDES AND THEIR
A IMPACT ON FRUCTOSE CRYSTALLIZATION

presented by

Yi-ding Chu

has been accepted towards fulfillment

of the requirements for
Agricultural

Ph . D 0 degree in Engineering

 

 

fli/W

Major professor

Date

 

fife/V
’ /

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

 

MSU A RETURNING MA'. ERIALS:
Place in book drop to
LJBRARJES remove this checkout from

n your record. FINES will
, be charged if book is

 

returned after the date
stamped below.

 

 

are)
”timl

-~ 1"” *ri‘} «d
JAN % 34:? C

C3

 

 

 

 

 

THE FORMATION OF DIFRUCTOSE DIANHYDRIDES AND THEIR
IMPACT ON FRUCTOSE CRYSTALLIZATION

BY
Yi-ding Chu

A DISSERTATION

Submitted to
Michigan State University
in partial fullfilment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY
in

Agricultural Engineering

Department of Agricultural Engineering

1988

ABSTRACT

THE FORMATION OF DIFRUCTOSE DIANHYDRISES AND THEIR
IMPACT ON FRUCTOSE CRYSTALLIZATION

BY
Yi-ding Chu

Fructose undergoes dehydration reactions and forms
difructose dianhydrides in gitg during crystallization.
Difructose dianhydrides have been proposed as a cause in the
decrease of overall yields of fructose crystallization. In
this study, the kinetics of difructose dianhydrides
formation under industrial crystallizations and the effect
of difructose dianhydrides on the crystal growth of
anhydrous fructose from aqueous solution were investigated.

The kinetics of difructose dianhydrides formation were
determined by HPLC analysis. Four difructose dianhydrides
were detected. A second-order irreversible kinetic model
was proposed for the reaction. The extent of reaction
increased with increasing temperature and decreasing pH
value of the solution. The amounts of two of the difructose
dianhydrides stopped increasing when the fructose
concentration was 70% or less. On the other hand, the
formation of all difructose dianhydrides were retarded when
the initial fructose concentration was higher than 80%.

The kinetics of fructose crystal growth in the presence
of difructose dianhydrides were studied using

photomicroscopic contact nucleation techniques. Difructose

Yi-ding Chu
dianhydrides were found to cause inhibition of crystal
growth. Since the molecular structures of difructose
dianhydrides are similar to that of fructose, the decrease
of crystal growth is probably caused by the incorporation of
these impurities or adsorption to the crystal face which
would accept fructose molecules in the orientation that

existed in the difructose dianhydride.

ACKNOWLE DEGMENTS

The author is deeply indebted to her major professor,
Dr. Kris A. Berglund for his guidance, understanding and
encouragement throughout this work.

Appreciations are also expressed to Drs. Eric Grulke,
Pericles Markakis, and James Steffe for their helpful

suggestions and as the members of the guidance commitee.

1v

 

TABLE OF CONTENTS

LIST OF TABLES ....................................... Vi
LIST OF FIGURES ...................................... viii
INTRODUCTION ......................................... 1
CHAPTER 1 LITERATURE REVIEW ......................... 3

1.1 Crystallization ............................. 3
1.2 Fructose .................................... 17
1.3 Difructose Dianhydrides ..................... 34
1.4 References .................................. 40

CHAPTER 2 KINETICS OF DIFRUCTOSE DIANHYDRIDES
FORMATION UNDER FRUCTOSE CRYSTALLIZATION
CONDITIONS OOOOOOOOOCOOOOOCOOOOOCOOOOCOOOOO 44

Abstract .................................... 44
Introduction ................................ 45
Experimental ................................ 51
Results and Discussion ...................... 52
Conclusions ................................. 74
References .................................. 76

NNNNNN
00....
GUI-FUND”

CHAPTER 3 EFFECT OF GLUCOSE AND DIFRUCTOSE
DIANHYDRIDES ON CRYSTAL GROWTH IN FRUCTOSE
CRYSTALLIZATION .OOOOOOOOOCCOOOCO0.00.00... 78

Abstract .................................... 78
Introduction ................................ 79
Experimental ................................ 83
Results and Discussion ...................... 88
Conclusions ................................. 103
References .................................. 106

GUI-FuNH

UUUUUU

SMARY AND CONCLUS IONS O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O 1 O 7
RECOWENDATIONS O O O O I O O O O O O O O O O O 0 O O O O O O O 0 O O O O O O O O O O O O O 109
APPENDICES

A Methods and Materials ........................ 110
B original Data OOOOOIOOOOOOOOOOOOO0.0...0...... 114

CHAPTER 1

Table 1.1

Table

Table
Table
Table

Table

Table

Table

Table

1.2

1.8

1.9

CHAPTER 2

Table

Table

Table

Table

Table

2.1

2.2

LIST OF TABLES

Average fructose content in common foods....

Relative sweetness of some sugars and
noncaloric sweeteners OOOOOOOOOOIOOOOOOOOOOO

Solubilities of some sugars in water .......
Density of aqueous fructose solution .......
Viscosity of aqueous fructose solution .....

Refractive index of aqueous fructose
salution 000......0.0...OOOOOOOCOOOOOOOOOOOO

Tautomeric equilibria of D-fructose
SOIutions OOOOOOOOOOOOOOOO~OOOOOO0.0.0.000...

1

Percentage of water absorbed by some sugars
from moist air .............................

Known difructose dianhydrides ..............

Tautomeric equilibria of D-fructose
salutions 0.0.0....OOOCOIOOCOOOOOOOOOOOOOOOO

Known difructose dianhydrides ..............

Second-order rate constants of difructose
dianhydrides formation reaction ............

Second-order rate constant of difructose
dianhydrides formation reaction at 60 °C
ande 2.65 0.0.0...OOOOOOOOOOOOOOOOOOOOOOOO

Molar ratio of water to fructose in reaction
solutions at 50 oC .........................

vi

18

20

23

24

25

28

31

32

36

47

50

60

65

67

Table 2.6

Table 2.7

CHAPTER 3

Table 3.1

Table

Table

Table

Table

Table

 

3.2

The effect of fructose concentration on the
reaction rate of difructose dianhydrides
formation at 60 oC and pH 5.90 ............. 70

Viscosity of fructose solutions at 60 °C ... 73

Tautomeric equilibria of D-fructose
selutions 0.0.0.0.........OIOOOOOOOOOOOOOOOO 82

Conditions of crystal growth experiments for
fructose-water-glucose system at 40 oC ..... 86

Conditions of crystal growth experiments for
fructose-water-difructose dianhydrides
system ...-0......0.7.0.........OOOOOIOOOOOOO 87

Summary of parameter estimation for growth
rate data 00......OOOOOOOO......OOOOOOOOOOOO 92

Summary of statistics of regression
equations (4) and (6) ...................... 94

Known_difructose dianhydrides ........ ...... 97

vii

CHAPTER 1
Figure 1.1
Figure 1.2
Figure 1.3

Figure 1.4

Figure 1.5

Figure 1.6

Figure 1.7

Figure 1.8

Figure 1.9

CHAPTER 2

Figure 2.1

Figure 2.2
Figure 2.3
Figure 2.4

Figure 2.5

Figure 2.6

LIST OF FIGURES

Typical solubility plot. ..................
Surface of a growing crystal. .............
Two-dimensional nucleation models. ........

Development of a growth spiral starting
from a screw dislocation. ............ .....

D-Fructose-water phase diagram. ...........

Possible tautomeric forms of D-fructose in
salution. 0.000.000.........OOOOOOOOOOOOOOO

Fructose crystal showing different
development of faces. .....................

Structures of some difructose dianhydrides.

The effect of difructose dianhydrides on
yield of fructose crystallization. ........

Possible tautomers of D-fructose in
salution. ......OOOOOOOOO......OOOOOIOOC...

Diheterolevulosan I. ......................
HPLC analysis of difructose dianhydrides. .
GC analysis of difructose dianhydrides. ...
The formation of difructose dianhydrides
at so °c and pH 2.65 with a initial

fructose concentration of 80.3%. ..........

Second-order kinetic plot of difructose
dianhydrides formation at 60 °C. ..........

viii

 

10

12

27

30

33

37

39

46

48

53

54

57

58

Figure

Figure

Figure.

2.8

The Arrhenius plot of the first reaction
section of difructose dianhydrides
fomation. ................ ...... ... ....... 61

Second-order kinetic plot of odifructose

. dianhydrides formation at 60° C and pH

CHAPTER 3

Figure

Figure
Figure

Figure

Figure

Figure
Figure
Figure

Figure

3.3

2. 65 with different initial fructose
concentration. ............................ 64

The formation of difructose dianhydrides
at 60 °C and pH 2. 65 with different
initial fructose concentration. ........... 66

Second-order kinetic plot of odifructose
dianhydrides formation at 60° C and pH

5. 90 with different initial fructose
concentration. ...........................0 69
The effect of initial fructose

concentration on the formation of (a) D4 '
(b) 03 (c) 01+02 at 60°C and pH 5.90. ..... 71

Possible tautomers of D-fructose in
SOlution. ................................. 81

Schematic diagram of nucleation cell. ..... 85:

Size versus time plot of fructose contact
nuclei grown in the presence of glucose. .. 89

The influence of glucose on the
relationship between mean growth rate and
supersaturation at 40 oC. ................. 90

The influence of impurities on the variance
Of growth rate. ........................... 93

DiheterOIevulosan I. ...................... 96
HPLC analysis of difructose dianhydrides. . 98

The effect of difructose dianhydrides on
normalized mean growth rate. .............. 100

HPLC analysis of difructose dianhydrides
incorporation. ............................ 102

ix

INTRODUCTION

Fructose, also called levulose or fruit sugar, has
great potential use as a natural sweetener due to its wide
abundance and high sweetness. Pure crystalline fructose has
until recently remained a noncommercial product because of
the expense involved in its production. Earlier methods of
fructose isolation from either hydrolyzed sucrose or
hydrolyzed fructose polymers were not economically
competitive with the very low price for sucrose processed
from sugar cane or sugar beets. The large scale industrial
production of fructose became possible when ion-exchange
techniques were developed to separate the glucose_and
fructose components in aqueous solutions of inverted sucrose
or a mixture obtained by glucose isomerase conversion of
glucose to fructose. In 1988, the wholesale price of
commercial crystalline fructose in the United States is
about $2.2/Kg. At the retail level, crystalline fructose
still costs about 10 times as much as sucrose. This is 3
mainly because the crystallization step in its production is
very time-consuming and requires careful control of process
conditions.

The physico-chemical properties of fructose are
unfavorable to crystallization. It is highly soluble in

1

 

‘water resulting in a very viscous saturated solution. It
tends to crystallize as hemihydrate and/or dihydrate phases
rather than the anhydrous form under centain conditions.
These hydrate compounds melt near room temperature which
makes them unacceptable as products. Fructose also
undergoes a complex mutarotation reaction in aqueous
solution with three tautomeric forms existing in significant
amOunts.

In addition to physico-chemical properties, the
presence of impurities in fructose syrup may retard crystal
growth. These imputities include glucose and difructose
dianhydrides which may exist in the fructose syrup in
substantial amounts.- Residual glucose is often found after
the ion-exchange enrichment of fructose syrup. Difructose
dianhydrides are formed in sign under industrial ;
crystallization conditions and have been reported to
decrease the overall yield in fructose crystallization.

The objectives of this research are:

(l) to study the kinetics of difructose dianhydides
formation under fructose crystallization conditions;

(2) to determine the effect of difructose dianhydrides on
the growth of fructose crystals.

The results of this study will not only provide
information necessary to the control of fructose
crystallization, but also contribute to the basic
understanding of fructose chemistry as well as the effect of

imputities on the growth of crystals.

CHAPTER I

LITERATURE REVIEW

1.1. Crystallization

1.1.1. Introduction .

Crystallization is one of the oldest unit operation
used for production of solid material in the chemical
industry. It has the advantage of yielding an end product
that has good flow, handling, and packaging characteristics,
and also attractive appearance. Major concerns in
production of crystalline product include control of
uniformity, or crystal size distribution, and purity.

The use of crystallization as a purification and
separation process has extended far outside the traditional
chemical industry. For example, the recent advances in
microelectronics have been made possible in part because of
the ability to grow single crystals of precisely controlled
composition and structural perfection. Thus, basic
understanding of the crystallization process is necessary
not only for the design and control of industrial
crystallizer, but also for advances in materials science
[1].

Any crystallization operation can be considered to be

comprised of three basic steps: achievement of

3

supersaturation, formation of crystal nuclei, and growth of
the crystals. All three processes may occur simutaneously
in different regions of the crystallization unit and control

the quality and uniformity of the end product [2].

1.1.2. Supersaturation

Supersaturation is defined as the presence of more
solute in solution than would exist at equilibrium. The
condition of equilibrium is called the solubility of a
particular solute. A typical solubility diagram is shown in
Figure 1.1. Three regions exist in the solubility diagram
[2]:

(1) stable zone (undersaturated including solubility
curve) where no nucleation or growth occurs:

(2) metastable zone (supersaturated) where growth may
occur but nucleation does not;

(3) labile zone (supersaturated) where both nucleation and
growth occur.
The boundary between labile and metastable zones is not well
defined.

Supersaturation can be achieved by cooling
(corresponding to line AC' in Figure 1.1), evaporation (line
AC) or a combination of cooling and evaporation (curve AC").
The choice of method depends on the compounds to be
crystallized. Cooling is employed to create supersaturation
for compounds having a steep solubility curve, while

evaporation is used to achieve supersaturation for compounds

CONCENTRATION

 

 

LABILE

 

STABLE

 

TEMPERATURE

Figure 1.1 Typical solubility plot.
(from Mullin [2])

   
 
  

 

having a flat solubility curve [3].
The most common expressions of supersaturation are the
concentration driving force AC, the supersaturation ratio S

and relative supersaturation s defined as follows.

Ac = c-c (1)
s = we" (2)
s = AC/C* = 5-1 (3)

where C is solution concentration and C* is saturation
concentration. Another often used expression is
supercooling AT defined as

AT = 'r*-'r (4)

[2].

1.1.3. Nucleation
Nucleation is the formation of new solid phase. It is
generally classfied as primary and secondary based on the '
presence of growing crystals in the solution [4].
a. Primary nucleation
Primary nucleation occurs in systems that do not
contain crystals. There are two types of primary
nucleation: homogeneous and heterogeneous. In homogeneous
systems there is no foreign substance and the formation of
nuclei is due to supersaturation only [2]. The mechanism is
generally believed to be a series of bimolecular reactions:
A +A=A2
A2 +A

A 1+ A=A

n-

The overall free energy of these aggregates goes through a
maximum at some critical size which is inversely
proportional to the logarithm of the solution
supersaturation [5]. Aggregates larger than the critical
size will decrease their free energy by further growth.

Thus stable nuclei are formed and grow to macroscopic

crystals.
The rate of nucleation3is given by the equation
'C Ys
B = Cl expt-5-3---5-1 _
T (lnS) (5)

where B is the number of nuclei formed per unit time per
unit volume, C1 and C2 are constants, TS is interfacial
tension, T is temperature, and S is supersaturation.
Homogeneous nucleation is therefore dominant only at high
supersaturation with the rate negligible at low
supersaturations [1]. Most primary nucleation that occurs
in practice is likely to be heterogeneous, i.e. induced by
surfaces of foreign substances. The mechanism is unclear,
yet it is thought to be the result of a local ordering
process brought about by interactions across the interface
[6].
b. Secondary nucleation

Secondary nucleation requires the presence of growing
crystals in solution. It occurs at much lower
supersaturation as compared to that in primary nucleation.
In general most industrial crystallization processes operate

under conditions favoring secondary nucleation [4].

Secondary nuclei can originate from different sources
as discussed by Botsaris [4] and Jancic and Grootscholten
[3]. The most important types of secondary nucleation
are summarized in the following:

(1) initial breeding - when a crystal whose surface
contains tiny crystallites is introduced into the solution,
the crystallites fall off and form nuclei:

(2) needle breeding - when dendrites are formed on the
surface of a crystal growing at high supersaturation, the
dendrites may break off to form nuclei:

(3) fluid shear - when shearing action occurs at the
crystal surface where certain ordered aggregated of solute
molecule are attached, the aggregates may be removed and
develop into new crystals: 1

(4) contact nucleation - when the surface of a growing
crystal contacts the wall of the crystallizer, impeller or
another crystal, nucleation occurs.

Contact nucleation has been more widely studied than
other categories and is suggested to be the most significant
secondary nucleation in most industrial crystallization [1].
There are two explanations for this phenomenon. The first
possibility is that the contact results in microattrition of
the crystal surface [7]. The second explanation involves
the removal of an absorbed layer of solute molecules at the
crystal surface [8].

The rate of secondary nucleation depends on three

factors: supersaturation, degree of agitation ,and

suspension density. Empirical expressions which employ
power law relations are generally used to model the kinetics

of secondary nucleation [1].

1.1.4. Crystal growth

The growth of crystals from solution involves two
steps. The solute is first transported from the bulk of
solution to the vicinity of the crystal surface, then it is
adsorbed and incorporated into the crystal lattice. The
latter stage is often referred to as the surface integration
process [3].

a. Mass transfer limited growth model

In some instances, the crystal growth is limited by the
mass transfer of solute. The growth model usually used for
this case is

G = kcA Ac (6)
where G is crystal growth rate, kc is mass transfer
coefficient, A is surface area, and AC is the concentration
difference between the solid surface and the bulk of
solution [2].
b. Surface integration limited growth models

The mechanism of surface integration is depicted in
Figure 1.2. Instead of being flat, a growing crystal surface
has steps and the step may be incomplete, having one or more
kinks. A growth unit which is adsorbed on the surface will
diffuse across the surface looking for a favorable site to

incorporate into the lattice. The most favorable site is a

10

 

 

 

 

 

Figure 1.2 Surface of a growing crystal showing (A) flat
surfaces (8) steps (C) kinks (D) surface-
adsorbed growth unit (8) edge vacancies and
(F) surface vacancies. (from Mullin [2])

11

kink followed by a step. Growth occurs when the units fill
out the steps causing them to move across the surface [9].
Growth models are differentiated by the source of steps as
follows.
(1) Two-dimensional growth models

A new step could be created on a flat crystal surface
by surface nucleation provided that the supersaturation is
large enough. Ohara and Reid [10] discussed three models
of this type as shown in Figure 1.3. They differ from each
other in the relative rates of formation and growth (i.e.
spreading) of the two dimensional nuclei. The mononuclear
model applies for the case of infinite fast spreading: the
polynuclear model corresponds to the case of near—zero
spreading velocity. The birth and spread model describes
cases between the two extremes, i.e. finite rate of both
nuclei formation and spreading, and nuclei can be formed on
unfinished layers. The general form of two-dimensional
growth models is given by

G a Alspexp(A2/s) (7)
where G is growth rate, A1 and A2 are constants, and s is
supersaturation. The exponent p a 1/2, 3/2 and 6/5 for
mononuclear, polynuclear and birth and spread model
respectively.
(2) Burton-Cabrera-Frank (BCF) model

Burton et al. [11] developed a model explaining growth

at low supersaturations which do not allow the formation of

nuclei as required in the two-dimensional growth model.

12

 

 

 

(c) Birth and Spread Mode]

Figure 1.3 Two-dimensional nucleation models.
(from Chara and Reid [10])

13

Here a self-perpetuating step is provided by the screw
dislocation on the surface. Figure 1.4 shows the development
of a growth spiral. The mathematical form of this model is

c - c (sz/sc) tanh(sc/s) (a)

where G is growth rate, C is constant, s is supersaturation,

and 3c is a critical supersaturation. For s << sc, i.e. at
low supersaturation, this reduces to

G = c (sz/sc) (9)
while for 3 >> sc at high supersaturation,

G a C s (10)

Since the constants in those models are not known, the
simple power law equation
G = a 3b (11)
is generally used to correlate growth rate data [1]. This
equation represents the two limiting cases of the BCF model
(equations (9) and (10)) and is a good approximation in the
intermediate regime when 1 < b < 2.
c. Growth rate dispersion
When exposed to constant conditions of supersaturation,
temperature, and hydrodynamics different crystals of the
same material usually grow at different rates. This
phenomenon is known as growth rate dispersion [12]. Two
possible mechanisms have been proposed: ‘
(1) random fluctuation (RF) model - the growth of crystal
fluctuates during the course of time [13];

(2) constant crystal growth (CCG) model - the growth rate

of each crystal is constant, yet the rate varies from

14

 

(o) I (o) (c)

Figure 1.4 Development of a growth spiral starting from
a screw dislocation: (a) screw dislocation
(b) growth spiral, top view (c) growth spiral,
side view (from Mullin [2]).

15

crystal to crystal [14].

Observations on single crystals for shorter periods (at
least a few hours) generally support the CCG model. Since
the fast-growing crystals will increase in size more rapidly
than slow-growing, the larger crystals are more likely to
have faster growth rates, thus size-dependent growth has
been incorrectly employed to interprete experimental growth

data and model the crystal size distribution [1].

1.1.5. ,Impurities in crystal growth

' Most industrial crystallization processes take place in
the presence of some impurities. An impurity can exert a .
profound effect not only on the nucleation and growth
kinetics, but also on the purity and morphology of the
product.

The influence of impurities is generally classified

into four groups as dicussed by Mullin [2]:

(1) impurities which exert a dilution effect, retard
diffusion and thus the growth rate;

(2) impurities which alter the properties of solution or
the solubility of solute, by either promoting or retarding
the formation of the more ordered species necessary for
nucleation and growth:

(3) impurities which is adsorbed to the crystal surface
and interferes with the growth by presenting steric or
energetic barriers to the subsquent addition of growth
units. The Langmuir adsorption isotherm has been employed

in modelling this type of impurity effect and fairly good

16

correlation have been reported by Blitznakov and Kirkova
[15];

(4) impurities which are built into the crystals.
Impurities can be incorporated into crystal lattice if there
is some degree of structural similarity or they can enter a
growing crystal through the inclusion of liquid. Inclusions
may be formed when dendritic (needle) growth is followed by
a filling in process that emtraps fluid in pockets. Growth
instabilities which result in depressions in the crystal
surface may also cause liquid inclusion. In general large
crystals and/or fast growth are more likely to form'
inclusions [3].

The effect of impurities on crystal habit modification
has been the focus of recent studies in crystallization.
The habit of a crystal is defined as the type of external
shape which results from the different rates of growth of
the various faces. The growth of a given face is governed
by the crystal sturcture as well as the environmental
conditions including solvent, impurities, supersaturation
and temperature [2]. An impurity which affects the growth
of specific faces can be used to modify the habit and
produce crystals of desired properties for washing,
centrifugation/filtration, handling and subsequent use.
This has been referred to as "tailor-made impurities" in the
literature. Davey et al. [16] studied the growth of urea
crystals in the presence of biuret, which is the dimer of

urea formed during its synthesis reaction. When grown from

17

pure aqueous solutions urea crystallizes as long needles
'with a length to breadth ratio exceeding 50:1. The presence
of biuret results in crystals with a much smaller length to
breadth ratio. Such crystals are easier to separate and
handling. Black et al. [17] and Weissbuch et al. [18] both
reported the crystallization of racemic mixtures of amino
acids in the presence of chiral additives, e.g. a different
amino acid. The habit of crystals were modified because the
additives were adsorbed only onto surfaces of similar

chirality.

1.2. Fructose

1.2.1. Occurrence

Fructose, also called levulose or fruit sugar, is one
of the most common natural sugars. It is found in the free
form in almost all sweet fruits and vegetables. About 50%
of the dry matter of honey is fructose. The fructose
contents of some common foods is presented in Table 1.1 [19].

Fructose is also found in the polymer form, as is the
case with inulin, the storage polysaccharide in dahlia and
artichoke tubers. Levan is another fructosan which occurs
in some bacteria. In addition, fructose as well as glucose

constitutes the abundant disaccharide, sucrose [20].

1.2.2. Usage of crystalline fructose
The rising interest in using sugars other than the

conventional sucrose results from:

18

Table 1.1 Average fructose content in common foods

 

 

Fruit Fructose % total % total
content carbohy- solid as
(g/100g drate fructose
of edible
portion
Honey 40.5 82.3 48.9
Banana 5.9 22.2 39.0
Apple 5.9 14.5 38.0
Grape 6.5 15.7 35.5
Pear 5.6 15.3 33.3
Cherry 5.4 17.4 27.3
Strawberry 2.1 8.4 21.1
Blueberry 3.3 15.3 19.5
Grapefruit 2.3 10.6 19.5
Orange 2.6 12.2 18.3

Blackberry

12.9

17.7

 

19

(1) diabetics must avoid sucrose in their diet:

(2) sucrose-containing snacks have been proven harmful to
the teeth:

(3) the synthetic sweeteners, e.g. saccharin, have a
bitter aftertaste, and the safety of using synthetic
sweeteners has been questioned [21].

Fructose meets many of the requirements when a
substitute for sucrose is called for.

(1) It is a natural sugar. It has not shown any toxic
properties. Reasonable intake of fructose has not been
found to cause side-effects, such as diarrhea, which is
typical of the sugar alcohols [22].

(2) Fructose metabolism is insulin-independent and it is
tolerated better than sucrose or glucose by diabetics [23].

(3) Fructose causes less formation of dental plaque than
does sucrose [21].

(4) It is the sweetest natural sugar [19]. Table 1.2 shows
the relative sweetness of some sweetners [24]. The caloric
saving provided by its greater sweetness (i.e. less intake)

make fructose a desirable sweetener for dietetic foods.

1.2.3. Manufacture

Pure crystalline fructose has until recently remained a
noncommercial product because of the expense involved in its
production. Earlier methods of fructose isolation from
either hydrolyzed sucrose or hydrolyzed fructose polymers,

e.g. inulin, were not economically competitive with the very

20

Table 1.2 Relative sweetness of some sugars and
noncaloric sweeteners

 

 

Sugar Relative sweetness
Sucrose 1.0

Glucose 0.5
Fructose 1.7a
Lactose 0.2
Saccharin 400

Sodium cyclamate 30

Aspartame 180

Monellin 2000

 

a1.8 in Doty and Vanninen [19].

21

low price for sucrose processed from sugar cane or sugar
beets [25]. The large scale industrial production of
fructose became possible when ion-exchange techniques were
developed to separate the glucose and fructose components in
aqueous solution of inverted sucrose. The calcium form of a
sulfonated-polystyrene exchange resin is employed in the
separation step [22]. More recent technologies also involve
ion-exchange separation of fructose from glucose in a
mixture obtained by the isomerization of glucose by means of
immobilized glucose isomerase [26]. Another procedure uses
glucose-2-oxidase to oxidize glucose. The glucosone is then
selectively hydrogenated to fructose [25].

In 1988, the wholesale price of commercial crystalline
fructose in the United States is ca $2.2/Kg. At the retail
level, crystalline fructose still cost about 10 times as
much as sucrose. This is mainly because the crystallization
step in its production is very time-consuming and requires
careful control of process conditions [28].

Crystalline fructose can be obtained by crystallizing
from aqueous alcohol solutions. Lauer et al. [29] proposed
a process for crystallization of fructose from methanol.

The employment of solvents is undesirable from the economic
point of view for both solvent addition and the later
solvent removal. Kusch et a1. [30] reported a process
wherein a fructose syrup of 95% at a pH of 3.5 - 8.0 is
concentrated by vacuum evaporation to 2 - 5% water and

cooled to 60 - 85 °C. Pure crystalline fructose is added as

22

seeds and the whole mass is solidified. The solid is then
ground to achieve a crystalline product. In a subsequent
design, Forsberg et al. [31] presented another process for
crystallization from seeded aqueous solution.. The key to
this process is pH adjustment. A pH of 4.5 - 5.5 is found
to be the most favorable range. Another method presented by
Yamauchi [32] involves seeding and then concentrating or
cooling the solution while maintaining the sugar
concentration and temperature in the liquid phase within a
carefully defined range. Dwivedi and Raniwala [28]
presented a seeded process wherein large pellets are formed

which must be ground to the proper size.

1.2.4. Properties
a. Solubility
Fructose is the most water-soluble of all sugars [21].:
The solubilites of fructose, glucose and sucrose in water
Iare shown in Table 1.3. Fructose is also soluble in ethanol
up to 6.71 g/100 ml at 18 °C [19].
b. Density 1
The density of fructose-water solutions at different
temperatures and concentrations is given in Table 1.4 [33].
c. Viscosity
The viscosity of saturated and supersaturated aqueous
_fructose solutions is much higher than that of other sugars
due to its high solubility. Table 1.5 lists the viscosity

data reported by Watanabe [34].

23

Table 1.3 Solubilities of some sugars in water

 

 

Temperature Fructose Sucrose Glucose
(°c> (was) (Wt) (Wt)
20 79.0 66.6 47.1
30 81.5 68.2 54.6
40 84.3 70.0 61.8
50 87.2 72.0 70.9
60 90.3 74.2 -

 

(fructose & sucrose : Vanninen and Dotty [22]:
glucose : Dotty and Vanninen [19])

24

Table 1.4 Density of aqueous fructose solution

 

 

 

Temperature Concentration (wt %)

50% 60% 70% 80%
10 1.2357 1.2927 - -
20 .2301 .2860 1.3452 1.4068
30 .2242 .2792 .3378 .3990
40 .2176 .2721 .3303 .3915
50 .2106 .2647 .3227 “3832
60 .2036 .2574 .3145 .3744
70 .1962 .2498 .3062 .3661
80 .1890 .2415 .2979 .3584
90 .1824 .2343 .2907 .3502

 

25

Table 1.5 Viscosity of aqueous fructose solution

 

 

 

Concentration Temperature (°C)

(Wt H

25 40 60
30 2.6 1.8 1.1
50 3.5 5.2 2.8
60 24 13 5.2
70 170 39 18
80 ‘ - 92
90 - - 2100

 

(unit: c.p.)

26

d. Diffusion coefficient

The diffusion coefficient of fructose in concentrated
aqueous solutions is not available in the literature.
Uedaira and Uedaira [35] reported an equation for dilute
solutions at 25 OC:

0 - (7.002 - 0.813 o)x10"6 (cmZ/s) (12)
The applicable concentration range is C < 3%.
e. Phase diagram

Fructose is most often crystallized from aqueous
solution as an anhydrous phase. However, fructose
hemihydrates and dihydrates form rapidly in the vicinity of
room temperature or lower. The phase diagram of D-Fructose-
water system reported by Young et al. [36] is shown in
Figure 1.5. '

Fructose dihydrate is the stable phase and crystallizes
at conditions of below 19.9 °C and 79.4%. Although the
hemihydrate form is the stable form only in the range from
19.9 to 21.4 0C, it frequently has formed spontaneously at
24 to 26 oC and its transformation to anhydrous form is
quite slow. Below 19.9 °C it usually changes to dihydrate
form if it is crushed or vigorously stirred in a saturated
solution. Once the transformation is initiated, it proceeds
rapidly [36]. The melting points of anhydrous, hemihydrate
and dihydrate crystal are 102-104, 68 and 21.3 0C
respectively [36,37,38].

f. Refractive index
Refractive index of fructose solution at two

temperatures is given in Table 1.6 [33].

27

 

44o -

430 -

TEMPERA TURE ('6.)
l | t «L
'3 B o B 8
T l 5 l r”

I

(a

Q
l

 

 

T T T I I I l I [[1
Hal.) -‘
l
I
“I
F“ 25 .—
...,
‘1" /' H
(Vflxf /4’K
°\“ /‘9£ ., "J
#T /' Q“
./ [c
f g.» ’ .—
' f9 , .
p4 “Tr ,0. “g
’05 5 .0 I
; a ‘ / I§
a .L ./ 11¢} ..
I
1.0/q?!“
43/9 I”
/ («,8
I /\ § .-
‘ilg
° :4" ':
\\IV
is ‘—
1 l I - I l l 1 1 1

 

Figure 1.5

IO 20 30 40 50 60 70 80 .90 I00

COMPOSITION (Percent o-Fruc/ose by Weight)

D-Fructose-water phase diagram: ? ice, 0
anhydrous D-fructose, o- D-fructose hemihydrate,
‘b D-fructose dihydrate, q~metastable phase,
e.gel. (from Young et al. [36])

Table 1.6 Refractive Index of aqueous fructose solution

28

 

 

 

Concentration Temperature (°C)
(Wt %)
20 25

0 1.33300 1.33252
11.855 .34043 .34985
19.098 .36185 .35115
30.912 .38197 .38104
40.906 .40035 .39932
50.212 .41856 .41758
60.488 .44013 .43894
59.156 .45964 .45854
75.242 .47524 .47399
82.321 .49062 .48945

 

29

g. Tautomeric equilibrium
Crystalline anhydrous fructose has the
configuration of B-D-pyranose. In solution, fructose
establishes an equilibrium state of its various tautomers as
shown in Figure 1.6. The influences of temperature and
concentration on the tautomeric equilibrium has been
investigated by Hyvonen et al. [39]. It is found that the
proportion of furanose forms increases with temperature,
while concentration does not change the equilibrium
significantly (Table 1.7).
h. Optical rotation
The specific rotation for B-D-fructopyranose is -131°.
Because no crystalline form of the other tautomers has been
obtained, specific rotation values for them are calculated
from tautomeric equilibrium data and inconsistency is found
among different sources. For fructose solution, the
generally accepted specific rotation is -91° at 23 °C [40].
i. Hygroscopicity
Fructose is a very hygroscopic sugar. The rate of
water absorption by fructose is higher than that of glucose
and sucrose as shown in Table 1.8 [41].
j. Crystal morphology
a-D-fructopyranose crystallizes in the rhombic
bisphenoidal class of the orthorhombic system with a:b:c =
0.801:1:0.907: space group P212121. Figure 1.7 shows the
crystal and different development of crystallographic faces

[42,43].

30

(3F! (3}.
CHZOH Hom
CH OH
OH OH '\\I‘ (.ZH20H(//‘ 2
HO czo HO OH
8 -D- f ructopyranose H (I: 23H \ a -D- f ructopyranose
‘//‘ HQOH \
HOCHz OH CHZOH HOCHz ° cnzon
“0 CHZOH “0 OH
HO 7, HO
B-D-fructofuranose a-D-fructofuranose

Figure 1.5 Possible tautomeric forms of D-tructosc in
solution.

31

Table 1.7 Tautomeric equilibria of D-fructose solutions.

 

 

Tempera- Concentra- o-furanose B-furanose B-pyranose
ture tion (%) (%) (%)
(°C) (was)

23 20 6 21 73
23 50 4 21 75
23 80 5 21 74
0 20 4 18 78
22 20 6 21 73
67 20 8 ' 28 64

77 20 12 31 57

 

32

Table 1.8 Percentage of water absorbed by some sugars from
moist air

 

 

 

Sugar Relative humidity at 20 °C

60%, 1 hr 60%, 9days 100%, 25days
Sucrose 0.04 0.03 18.4
Glucose 0.07 0.07 14.5
Fructose 0.28 0.63 73.4

Maltose 5.05 5.1 -

 

33

OIO

 

 

 

 

 

 

 

 

 

 

 

 

 

"- s
0 MI. s
A I m -
I... s
s
--l-- ..................... m
.
m .
.m
.0
m .
0 0 .
m m P . m
s
s
s
m -r9'--li-fio..-‘-loh
’ '
nl .
4’" \\ .
'. oil! \‘a w
0 I: \\
... .............. . .. 1
0 mam and .m .

 

Figure 1.7 Fructose crystal showing different development of
A faces. (from Bates [42])

34

k. Crystal growth kinetics

The growth kinetics of anhydrous fructose crystals from.
aqueous solution have been studied by Shiau and Berglund
[44]. Crystals were formed by contact nucleation and the
growth was described by the Constant Crystal Growth model,
i.e. each crystal grows at constant rate while different
crystals have different growht rates.

The kinetic model is given by

‘G a 1.47x107exp(-6120/RT) s1°25 (13)

where G is growth rate in um/hr, R is gas constant in
cal/g-mole °K, T is temperature in °K, and s is
relative supersaturation. The variance of the growth rate
distribution was approximated by

062 = 0.165 01°35 (14)

1.3. Difructose Dianhydrides

Fructose dehydrates in acid solution and/or on
prolonged heating and forms a series of non-reducing,
dimeric dianhydrides. The dehydration-dimerization produces

an extremely stable central 1,4-dioxane ring [45].

1.3.1. Structure and properties
Difructose dianhydrides differ in that
(1) the acetal rings of fructose moieties may be both
pyranoid, mixed, or both furanoid:

(2) the dioxane ring may be formed by dehydration at

35

different carbon atoms of each fructose unit:

(3) additional anomerism is produced by the asymmetry of
the second carbon atoms of each fructose unit.

The first difructose dianhydride was isolated by Pictet

and Chavan [46]. It was later shown to be di-D-
fructopyranose-1,2':2,1'-dianhydride and known as
diheterolevulosan I. The other three compounds in these
series, diheterolevulosan II - IV, were found by Wolfrom and'
coworkers [47,48]. These compounds have at least one moiety
of fructose in pyranoid structure. The name "levulosan" is
generally used for this type of difructose dianhydride. A
second series, known as difructose anhydride I - III, was
reported by Jackson and coworkers [49,50]. In this series
both fructose moieties are in furanoid form. In a more
recent study Defaye et al. [51] added a new difructose
dianhydride to the list. Details about the structure and
properties of these eight difructose dianhydrides are
summarized in Table 1.9. The structures of some of these

compounds are also shown in Figure 1.8.

2. Analysis

The separation of different difructose dianhydrides is
generally achieved by chromatographic techniques. McDonald
[52] developed paper chromatograph of diheterolevulosan I,
II and difructose anhydride I - III. Hilton [45] used a
carbon-celite column to prepare diheterolevulosan I - IV.
The TLC and HPLC analyses of diheterolevulosan I and II were

reported by Velasco and Dowing [53] and Ramada et a1. [54]

36

Table 1.9 Known difructose dianhydrides

 

fructose moiety

 

 

linkage trivial name [a]D m.p.

ring anomeric C
pyr-pyr 3!. a 1,2:2,l diheterolevulosan I ' -44° 266
fur 3, a 1,2:2,1 11 -40° 261
fur a, 3 1,2:2,1 III -179° 257
pyr 3, 3 1,2:2,1 IV -309° 240
pyr 3, B 1‘,2:2,3 - -58.5° 206
fur-fur a, B 1,2:2,1 difructose anhydride I 26.9° 164
fur, _ 1,2:2,4 II 13.9° 198

fur a, 3 1,2:2,3 III 135.6° 162

 

37

 

“OCH, 0
O o§>////
5' H

 

H

  

Structures of some difructose dianhydrides:

l. diheterolevulosan I, 2. diheterOIevulosnn It,
3. diheterolevulosan III, 4. diheterolevulonnn
IV, 5. difructose nnhydridn l, 6. difructnnn
anhydride III.

38

respectively. Defaye et al. [51] analyzed the methylated
derivatives of six dianhydrides on a GC system.

The structure of these dianhydrides were first examined
by periodate oxidation to determine the form of fructose
rings [47]. Later studies have applied NMR and x-ray to
analyze the configuration and the ring conformation.

The conformation of diheterolevulosan I - IV and the
new dianhydride have been determined by 1H-NMR analysis of
their hexa-acetate derivatives: their anomeric configuration
was assigned by 13C-NMR studies [55,56,51]. Structure of
difructose anhydride I was elucidated from 1H-NMR spectra of ~
its 6,6'-dideoxytetra-O-acetyl derivative [57].
Configuration of difructose anhydride III was determined by
13C-NMR [58] and the conformation was established from X-ray

analysis of its crystal [59].

3. Effect on fructose crystallization

Difructose dianhydrides have been reported to form in
[gitg under industrial crystallization conditions. It is
suggested that they are inhibitors for the growth of
fructose crystals because their presence decreases the

overall yields as illustrated in FigureL9 [31].

 

 

. ' I
Ylfild o |
7- ..
o . 8"
40r :
- I .-
30L :
l ..
zo- ' .
I
IOF , ' . I
l L L l I J
2 I 4 6 8 I0

9;, Disacchorides

Figure 1,9 The effect of difructose dianhydrides on yield of
fructose crystallization.

(from Forsberg et al. [31])

1.‘.

1.

2.

8.
9.

10.

11.

12.

13.

14.

15.

16.

17.

no

References

J. Garside, Chem. Engng. Sci. 40, 3 (1985).

J.W. Mullin, Crystallization (Butterworth & Co Ltd,
London, 1972).

S.J. Jancic and P.A.M. Grootscholten, Industrial
Crystallization (Delft University Press, Delft,
Holland, 1984).

G.D. Botsaris, in Industrial Crystallization, J.W.
Mullin Ed. (Plenum Press, New York, 1976) p. 3.

A.G. Walton, in Nucleation, A.G. Zettlemoyer Ed.
(Marcel Dekker, Inc., New York, 1969) p. 225.

A.G. Walton, The Formatin and Properties of
Precipitates (Interscience Publishers, New York, 1967).

J. Garside and M.A. Larson, J. Crystal Growth 43, 694
(1978). '

W. Drost-Hansen, Ind. Eng. Chem. 61, 10 (1969).

W. Kossel, Annln Phys. 21, 457 (1934).

M. Ohara and R.C. Reid, Modeling Crystal Growth Rates
from Solution (Prentice Hall, Inc., Englewood Cliffs,
NJ, 1973).

W.K. Burton, N. Cabrera and F.C. Frank, Phil. Trans.
Roy. Soc. 243, 299 (1951).

E.T. White and P.G. Wright, Chem. Engng. Prog. Sym.
Ser. 67, 81 (1971).

A.D. Randolph and E.T. White, Chem. Engng. Sci. 32,
1067 (1977).

M.A. Larson, E.T. White, K.A. Ramanarayanan and K.A.
Berglund, AIChE J. 31, 90 (1985).

G. Blitznakov and E. Kirkova, E., Krist. Tech. 4, 331
(1969).

R. Davey, W. Fila and J. Garside, J. Crystal Growth 79,
607 (1986).

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

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

N1

I. Weissbuch, L.J.W. Shimon, E.M. Landau, R. Popovitz-
Biro, z. Berkovitch-Yellin, L. Addadi, M. Lahav and

L. Leiserowitz, Pure App. chem. 58, 947 (1986).

T.E. Dotty and E. Vanninen, Food Technol. 11 (1975) 34.

R.S. Shallenberger and G.G. Birch, Sugar Chemistry
(The AVI Publishing Co., Inc. Wesport, CN, 1975).

L. Hyvonen and P. Koivistoinen, in Nutritive Sweeteners
G.G. Birch and K.J. Parker Ed. (Applied Science
Publishers, New york, 1982) p. 133.

E. Vanninen and T. Dotty, Sugar: Science and Technology
(Applied Science Publishers, New York, 1979) p. 311.

L. Sestoft, Nutrition Update 1, 39 (1983).

A.L. Lehninger, Principles of Biochemistry (Worth
Publishers, Inc., New York, 1982).

G.N. Bollenback, in Kirk-Othmer Encyclopedia of Chemical
Technology (John Wiley & Sons, New York, 1983) p. 944.

W.P. Chen, Process Biochem. 15, 36 (1980).
Chemical Marketing Reporter, May 9, 1988.

B.K. Dwivedi and S.K. Raniwala, U.S. Patent no.
4,199,373 (1980).

K. Lauer, P. Stephan and G. Stoeck, U.S. Patent no.
3,607,392 (1971).

T. Kusch, W. Gosewinkel and G. Stoeck, U.S. Patent No.
3,513,023 (1970).

R.H. Forsberg, L. Hamalainen, A.J. Melaja and J.J.
Virtanen, U.S. Patent no. 4,199,373 (1975).

T. Yamauchi, U.S. Patent No. 3,928,062 (1975).

J. Timmermans, The Physico-chemical Contants of Binary
Systems in Concentrated Solutions (Interscience
Publishers, New York, 1960).

T. Watanabe, Seito Gijutsu Kenkyu Kaishi 28, 70 (1978).

H. Uedaria and H. Uedaria, J. Phys. Chem. 74, 2211
(1970).

F.E. Young, F.T. Jones and H.J. Lewis, J. Phys. Chem.
56, 1093 (1952).

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

h2

F.T. Jones, F.E. Young and D.R. Balck, Anal. Chem. 25,
649 (1953).

F.E. Young, F.T. Jones and H.J. Lewis, J. Phys. Chem.
56, 738 (1952).

L. Hyvonen, P. Varo and P. Koivistoninen, J. Food Sci.
42, 652 (1977).

R.S. Shallenberger, Pure Appl. Chem. 50, 1409 (1978).

J.E. Hodge and E.M. Osman, in Principles of Food
Science Part I: Food Chemistry, O.R. Fennema Ed.
(Marcel Dekker, Inc. New York, 1976) p. 88.

F.J. Bates, Polarimetry, Saccharimetry and the Sugars
(United States Government Printing Office, Washington,
1942).

A.N. Winchell, The Optical Properties of Organic
Compounds, 2nd ed. (Academic Press, Inc., New York,
1954).

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

H.W. Hilton, in Methods in Carbohydrate Chemistry, Vol
II, R.L. Whistler and M.L. Wolfrom Ed. (Academic Press,
Inc, New York, 1963) p. 199. -

1

A. Pictet and J. Chavan, Helv. Chim. Acta. 9, 809
(1926).

W.L. Wolform and M.G. Blair, J. Am. Chem. Soc. 70,
2406 (1948).

W.L. Wolfrom, H.W. Hilton and W.W. Binkley, J. Am.
Chem. Soc. 74, 2867 (1952).

R.F. Jackson and S.M. Goergen, Bur. Stand. J. Res. 3,
27 (1929).

R.F. Jackson and E.J. McDonald, Bur. Stand. J. Res. 6,
709(1931)

J. Defaye, A. Gadelle and C. Pedersen, Carbohydr. Res.
136, 53 (1985).

E.J. McDonald and B.K. Goss, Anal. Chem. 24, 422
(1952).

V.S. Velasco and J.F. Dowling, U.S. Agri. Res. Serv.
South. Reg., [Rep.], ARS-S-88, 138 (1975).

54.

55.

56.

57.

58.

59.

43

K. Ramada, H. Yoshihara, G. Suzakamo and O. Hiroake,
Bull. Chem. Soc. Jpn. 57, 307 (1984).

R.W.
Res.

R.H.
Res.

R.U.

Binkley, W.W. Binkley and A.A. Grey, Carbohyd.
28, 365 (1973).

Binkley, W.W. Binkley and B. Wickberg, Carbohyd.
36, 196 (1974).

Lemieux and R. Nagarajan, Can. J. Chem. 42, 1270

(1964).

T. Uchiyama, Carbohyd. Res. 101, 138 (1982).

T. Taniguchi and T. Uchiyama, Carbohyd. Res. 107, 255
(1982).

CHAPTER 2*

KINETICS OF DIFRUCTOSE DIANHYDRIDES FORMATION UNDER
FRUCTOSE CRYSTALLIZATION CONDITIONS

2.1. Abstract

Fructose undergoes dehydration reactions and forms
difructose dianhydrides in gitg during crystallization.
These difructose dianhydrides can be incorporated into
fructose crystals and inhibit the crystal growth. In this
study, the kinetics of difructose dianhydrides formation
under industrial crystallization conditions were
investigated.

Four difructose dianhydrides were detected by HPLC

analysis. A second-order irreversible kinetic model was

proposed for the reaction. The extent of reaction increased

with increasing temperature and decreasing pH value of the

solution. The amount of two of the difructose dianhydrides

stopped increasing when the fructose concentration was about

70% or less. On the other hand, the formation of all dimers

were retarded when the initial fructose concentration was

higher than 80%.

 

*This chapter contains a paper submitted to "Biotechnology &

Bioengineering".

uh

45

2.2. Introduction

Fructose is one of the most abundant monosaccharides
and is known to be the sweetest natural sugar [1].
Crystalline fructose has not been widely available due to
its high cost resulting from a major processing difficulty
in the crystallization step of fructose from aqueous
solutions [2].

Fructose undergoes a complex mutarotation reaction in
aqueous solution with three tautomeric forms, B-pYranose, B-
furanose and a-furanose, existing in measurable amounts at
equilibrium. The reaction is shown in Figure 2.1 and the
equilibrium composition of aqueous solutions at different
temperatures and concentrations are given in Table2.1[3].
Fructose undergoes dehydration reactions in acid solution:
and/or on protracted heating. Unlike aldoses which
generally form monomolecular anhydrides, fructose forms
difructose dianhydrides possessing a central 1,4-dioxane
ring as shown in Figure 2.2.

Difructose dianhydrides have been suggested to be
inhibitors of fructose crystal growth. Forsberg et al. [4]
have reported that difructose dianhydrides were formed in
51:3 under industrial crystallization conditions and caused
a decrease in the overall yields of fructose
crystallization. Chu et al. [5] found that difructose
dianhydrides inhibit fructose growth. The inhibition is

probably caused by the incorporation of these compounds in

46

OH OH
CHZOH +10%
CH OH
OH OH Q4 (.ZHZOH.//‘ 2
HO $.10 HO 0H
8 -D-fructopyranose H3 g3 H ‘1 ‘D' f r uctopyranose
‘//‘ H¢0H ’\\1
' H
HOCH‘z OH CH20 HOCHz 0 011,014
H0 CHZOH ”0 OH
B-D-fructofuranose oeD-fructofuranose

Figure 2.1 Possible tautomers of D-fructose in solution.

47

Table 2.1 Tautomeric equilibria of D-fructose solutions.

 

 

Temperature Concentrat fin a-furanose B-furanose B-pyranose
(°c) (wt %) (%) (%) (%)
23 20 6 21 73
23 50 4 21 75
23 80 5 21 74
0 20 4 18 78
22 20 6 21 73
67 20 8 28 64

77 20 12 31 57

 

48

 

Figure. 2.2 Diheterolevulosan I.

1:9

the surface of crystal due to the structural similarity of
difructose dianhydrides to fructose.

The formation of difructose dianhydrides was first
reported by Pictet and Chavan [6] in a study of the action
of concentrated hydrochloric acid on fructose: the compound
studied was later known as diheterolevulosan I. Wolfrom and
Blair [7] isolated diheterolevulosan II on heating
concentrated aqueous fructose solution. Diheterolevulosan
III and IV were later found by Wolfrom and coworkers [8] and
Wickberg [9]. A second set of difructose dianhydrides has
been isolated mainly from syrup of hydrolyzed inulin, which
is a polyfructofuranose. Difructose anhydride I was
reported by Jackson and Goergen [10] and Irvine and
Stevenson [11] simultaneously, and difructose anhydride II
and III were prepared by Jackson and McDonald [12]. The
term "levulosan" was applied to those compounds with at
least one pyranoid structure, while the set named
”anhydride" possess furanoid forms only. Defaye et a1. [13]
have recently reported a new difructose dianhydride, thus
there are at least eight difructose dianhydrides in the
literature. Their sturctures differ in the configuration of
each fructose moiety and the linkage between them as
summarized in Table 2.2.

Most reports regarding difructose dianhydrides deal
with the isolation and structure elucidation of these
compounds. The kinetic information about their formation is

not available in the literature. The objective of this work

50

Table 2.2 Known difructose dianhydrides.

 

fructose moiety

 

 

::, linkage trivial name

Ping anomeric C

pyr-pyr B, o 1,2:2,1 diheterolevulosan I
fur B, a 1,2:2,1 II
fur B, B 1,2:2,1 III
pyr B. B 1,2:2,1 1v
PYr B. B 1,2:2,3 -

fur-fur a, B 1,2:2,1 difructose anhydride I
fur - 1,2:2,4 II
fur a, B l,2:2,3 III

 

51

is to study the kinetics of difructose dianhydrides

formation under fructose crystallization conditions.
2.3. Experimental

Fructose solutions were prepared by adding fructose to
hydrochloric acid solutions of desired pH. Reagent grade
fructose and double-distilled water were used. The reaction
was carried out in a 250 ml flask filled with 250 g
solution. Temperature was controlled by a water bath with a
Haake constant temperature circulator model E3 and a Curtin
Matheson Scientific 244-793 magnetic stirrer was used to
agitate the solution.

Samples were taken at intervals and further reaction
was minimized by immediate dilution and refrigeration.

About 2 g aliquot was taken and diluted with water in a 50
ml volumetric flask. The concentrations of difructose
dianhydrides and residual fructose were analyzed on a Biorad
Carbohydrate Analysis HPLC system, which was equipped with a
model 1770 refractive index detector and a HP 3392A
integrator. The column used was a 4.0x250 mm Bio-Sil Amino-
SS. The mobile solvent was acetonitrile/water with a volume
ratio of 70/30 and a flow rate of 1.0 ml/min. Samples of 5
ul were injected and eluted at room temperature. An
external standard composed of fructose and sucrose was used
to calibrate the analyses.

Some of the final reaction solutions were analyzed by

GC/MS spectrometry after silylation by TRI-SIL '2' purchased

52

from Pierce (trimethylsilylimidazole in dry pyridine, 1.5
meq/ml). The trimethylsilyl derivatives were analyzed on a
HP 5985 GC/MS system with a electron multiplier detector
with the ionization potential at 70 eV. The column was a 6
ft 3% OV-225 and the temperature program was 140-200 °C (3

°/min).

2.4. Results and Discussion

The crystallization conditions of fructose
manufacture usually involve a temperature range of 30 to 60
°C and the pH value of solution is controlled between 4 and
6. The growth rate of fructose crystals is so low that the
crystallization step usually occurs in times on the order of
days [2,4,14,15]. Thus the reaction temperatures of 30, 40,
50, and 60 °c and pH values of 2.65, 4.35 and 5.90 were.
chosen to study the effect of temperature and pH on the
formation of difructose dianhydrides. The fructose
concentrations were 10% less than the saturated values and
the reaction was monitored for two weeks.

since difructose dianhydrides are not available
commercially and sucrose has been reported to have a
retention time very close to difructose dianhydrides in a
TLC and a GC systems [16], it was used to calibrate the
relative amounts of difructose dianhydrides. Four peaks
were found in both HPLC and GC analysis. The typical
chromatograms are shown in Figures 2.3 and 2.4 respectively.

Since difructose dianhydrides are non-reducing dimers, the

53

 

 

 

 

‘I’I'F'I'I
0 2 4 6 81012

 

 

Retention Time (min)

Figure 2.3 HPLC analysis of difructose dianhydrides.
l. Fructose 2. diheterolevulosan II 3.
diheterolevulosan I 4. diheterolevulosan III
5. diheterolevulosan IV.

54

 

 

 

D V...

v ‘ ' T I

I T I ‘ I I ' I
4 8 8 1O 12 14 16

Retention Time (min)

"

 

0'2

‘ Figure 2.4 GC analysis of difructose dianhydrides.

55

chance of having anomeric isomers overlapped in the same
peak does not exist and it is reasonable to assume that
there were only four compounds formed in the reaction.
Hereafter they are designated as DI, D2, D3 ,and D4 in order
of increasing retention time.

Hamada et al. [17] reported a method of synthesizing
diheterolevulosan I and II with a ratio of 1 to 2.5. Their
procedure was followed and analyzed by the HPLC system for
the retention times of diheterolevulosan I and II. It was
found that compound D1 has the retention time of
diheterolevulosan II and D2 has the retention time of
diheterolevulosan I. The other two compounds are believed
to be diheterolevulosan III and IV because they possess at
least one a-pyranose moiety, which is the most abundant
species in aqueous solution, and might be formed in
significant amounts. Diheterolevulosan IV might appear
faster and/or in larger amount than diheterolevulosan III
since it is composed of two B-pyranoses, thus the pudative
assignments for D3 and D4 are diheterolevulosan III and IV.
The set of difructose anhydrides is less likely to be
produced from fructose solutions because they consist of the
less abundant furanoid structures only. This assumption is
supported by the report of Hilton [18], in which difructose
dianhydrides were prepared from acid treatment of 100 g

fructose and the amount of difructose anhydrides formed was

less than 5 g while that of diheterolevulosans was about 35 g.

The formation of difructose dianhydrides at 60 oC and a

56

pH of 2.65 is shown in Figure 2.5. It was found that D3 and
D4 were formed first yet their concentration did not
increase after a certain period of time. On the other hand,
D1 and D2 were formed later and the extent of formation
increased throughout the two weeks reaction duration. For
an initial fructose concentration of 70 to 80%, the total
amount of difructose dianhydrides formed in two weeks was
10% or less with D3 and D4 being 1.5 to 3.5%.

Since the difructose dianhydrides formed are stable in
acidic solutions [19], an irreversible reaction is assumed

and a second order kinetic model is proposed as following

-§1-> D1

k
2 Fructose ”'2'> DZ
' -53-> 03

where ki is reaction rate constant, or

2 Fructose --§-> Difructose dianhydrides (2)

where K - Zki. Thus the rate of reaction is given by

-dC/dt - x02 (3)
where C is fructose concentration and t is time, which on
integration yields

1/c - 1/co - Rt (4)
where CO is initial fructose concentration. The model was
tested graphically by plotting the reciprocal of fructose
concentration versus time. Figure 2.6 shows the plot at 60 °C

as an example. It was found that a two section broken line

57

5.0 1
0 D1
‘ x 02
A [B
O D‘ a
4.0 J
a
a e
‘ e
I
II 9 a
3.0:-

 

Concentration (wt %)

 

r’ - r 7

r '— r
' 48 96 144 192

Time (hr)

Figure 2.5 The formation of difructose dianhydrides at 60

oC and pH 2.65 with a initial fructose
concentration of 80.3%.

1/[7] x 100 (wt 8")

58

 

 

 

1J4
--- 1 pH 5.90
4 —-x pH 4.35
1‘) -——-o pH1185
. 4 f v '— T T I 1— 1
0 18 9'6 114 152 240 288 336

Time (hr)

Figure 2.6 Second-order kinetic plot of difructose
dianhydrides formation at 60 oC.

59

fitted the data best for all pH values at 50 and 60 oC.
Data of 30 and 40 °C were fitted by single straight line.
The slope, which is the reaction rate constant, was
evaluated for each straight line section with the results

summarized in Table 2.3.

2.4.1. Effect of pH

The tendency that D3 and D4 formed first yet leveled
off after a certain period of time was observed for all pH
values. The amount of difructose dianhydrides formed was
increased by lowering the pH values of the solutions. However,
this acid-accelerating effect was demonstrated only in the
first reaction section. The reaction rate constants of the
second section appeared to be pH-independent as were shown

1 in Table 2.3 previously.

2.4.2. Effect of temperature

The formation of difructose dianhydrides showed no
qualitative difference at temperatures in the range of 30 to
60 oC. The extent of reaction increased with increasing
temperature. Figure 2.7 shows the Arrhenius plot for the
first reaction section. It was found that the data could be
fitted by straight lines, thus the Arrhenius equation was
applied to describe the relationship between temperature and
the rate constant. Since the slopes were similar for all pH
values, it was assumed that the mechanism of reaction does
not change within the pH range studied. Rate constant data

of all pH values were pooled to evaluate the activation

60

Table 2.3 Second-order rate constants of difructose dianhydrides
formation reaction

 

 

 

Rate Temperature Initial pH
constant (0C) concentration
(wt %) 2.65 4.35 5.90

K1 30 71.6 .00026 - .00004
40 74.4 .00022 .00024 .00026
50 76.9 .00221 .00100 .00107
60 80.3 .00353 .00239 .00185

K2 30 71.6 * - *
40 74.4 * * *
50 76.9 .00024 .00025 .00038
60 86.3 .00059 .00045 .00048

 

(K in units of wt%'1hr’1)
Data were fitted by single straight line with the slope
being K1. =

61

0.01000 __ a pH _ 2.35

—— x pH - 4.35
...... A pH - 5.90

   
 
 

0.00100

L. L L4 LL11]

0.000101

Rato Constant (wt%’1hr’1)

PLLJ LA
.

 

0.0000110 . 3:1 r 3:2 3:3

1/‘1' x 103 (°x"1)

Figure 2.7 The Arrhenius plot of the first reaction section
of difructose dianhydrides formation.

62

energy with the effect of pH being included in the
preexponential factor. The resulting equation is

K1 - Kl'o exp(-23300/RT) (5)
where R is ideal gas constant in units of cal/g-mole OK, T
is temperature in units of oK, and K1,o = 8.34x1012,
4.69x1012 and 3.87x1012 (wt%)'1hr'1 for pH = 2.65, 4.35 and
5.90 respectively.

Similar analysis were applied to rate constant data of
the second reaction section, although there were only two
data points for each pH values due to the limits of fructose
solubility and the concentration dependence of the reaction.
The resulting equation is

K2 = Kz’o exp(-12400/RT) (6)
where Kz'o a 8.58x104, 7.69x104 and 6.75x104.(wt%)'1hr‘1 for
pH = 2.65, 4.35 and 5.90 respectively. ’

The activation energy of a reverse reaction, acid
hydrolysis of some disaccharides, has been reported ranging
from 25 to 40 Kcal/g-mole [20]. The activation energy for
this dehydration reaction is 23.3 Kcal for the first
reaction section and 12.4 Kcal for the second reaction
section, which are in the same order of magnitude but less

than that of the hydrolysis reaction.

2.4.3. Effect of concentration
By comparing plots of dimer formation and fructose
consumption (reciprocal of fructose concentration versus

time), it was noticed that the point at which the

63

concentration of D3 and D4 levels off coincided with the
break point of the two section reaction. This phenomenon
appears for all temperature and pH values, thus it suggests
that only the formation of 01 and D2 may account for the
second section of reaction, while all four dimers are
produced in the first section.

It was also noticed that the fructose concentration at
the break point was about 70% or less independent of
temperature and pH values. Further experiments were
performed with initial fructose concentaration being 75, 70
and 65% at 60 °c and a pH of 2.65. The result of fructose
consumption is shown in Figure 2.8 along with the data of
80.3% from a previous experiment. In all three cases the
fructose concentration fell to 70% or less within one day
and thereafter a linear relationship was found for each set
of data. The slopes of these three lines are similar and
close to that of the second reaction section in the case of
80.3%. Table 2.4 lists the slope, or second order rate
constant, of data with different initial fructose
concentration. The formation of fructose dimers is
illustrated in Figure 2.9 and it is confirmed that the
concentrations of DB and D4 will reach a certain value and
stop increasing when the fructose concentration is about 70%
or less.

The molar ratio of water to fructose in the reaction
solution was calculated and the result of reactions at 50 °C

is shown in Table 2.5 as an example. It was found that the

64

 

i/[r] x 100 (wt x'1)

 

 

,/
1.8-1 ./'(
J n I /.(°/ /A
/"/o a ‘//
1.7-4 2' // .
1.6~
1.5~
a"
1.4~
J
mi
J
1.2.. ----—o 85%
.....A 70%
. ----x 733
11 —--:0 503 A_
' ' l ' I ' ' A r v r T
0 4a as 134 162 2b 250 330
Time (hr)

Figure 2.8 Second-order kinetic plot of difructose

dianhydrides formation at 60°C and pH 2.65 with

different initial fructose concentration.

65

Table 2.4 Second-order rate constant of difructose
dianhydrides formation reaction at 60 °C

 

 

and pH 2.65
Initial Rate *
concentration (wt %) constant (K2)
80.3 .00059
75.0 .00062
70.0 .00066
65.0 .00056

 

K in units of wt%'lhr'1)
Data were fitted by single straight line.

66

 

 

 

 

 

 

0 D1
x 02
(a) 75 % ‘ 03
a 04
3.0-!
a
I I I U /
2.0- " ' - "
d O
......... -1----------
1.0-
0.0 rfii‘j“rr'*I
.. (b) 70 %
J
a
3.
fl
0
w
u
d
H
u
a
O
0
a
O
U
(c) 65%
3.01
2.0--1

 

 

0.0

 

_fr

0 70 9 T1l4715727'210'2501350
Time (hr)

Figure 2.9 The formation of difructose dianhydrides at 60

°C and pH 2.65 with different initial fructose
concentration.

5.90

67
pH
4.35

2.65

 

solutions at 50 °C

time
(hr)

 

Table 2.5 Molar ratio of water to fructose in reaction

Reaction

700499142683236

233333444444444

025601 45754151

333344 44444444

392234543457345

334444444444444

 

 

68

ratio was four or higher when the concentration of D3 and D4
stop increasing. This suggests that the solvation of
fructose molecules may affect the formation of difructose
dianhydrides. Hyvonen and coworkers [3] have studied the
effect of concentration on fructose mutarotation in aqueous
solution and reported that the proportion of fructose
tautomers did not change much within the concentration range
of 20 to 80% as shown previously in Table 2.1. Thus the
equilibrium composition of solution could not account for
the concentration effect of this reaction.

The reactions at 60 °C, a pH value of 5.90 and initial
concentrations of 84, 88 and 92% were also studied to
determine whether concentration in this range has any
influence on the formation of difructose dianhydrides. The
results for fructose consumption are shown in Figure 2.10
which also includes data of 80.3% from a previous
experiment. Since the fructose concentrations were higher
than 70% at the end of two weeks, the slopes from linear
fits were compared with that of the first reaction section
in data of 80.3%. It was found that the slope decreased as
the initial fructose concentration increased with the values
listed in Table 2.6. The formation of different dimers is
shown in Figure 2.11. The amounts of D1 and D2 cannot be
determined separately for some data points, thus the sum of
these two compounds was used in Figure 2.11c. It appears that
in this concentration range the formation of dimers was

retarded by high fructose concentration. The effect of

69

 

i/(rj x 100 (wt t")

 

--“’ll 80:!

1'0“ _— . 04::
J —~—x “X
---- 0 823

 

 

T f t v . r 1 1
"-90 r ,3 9'5 114 152 230 20 3:56
Time (hr)

Figure 2.10 Second-order kinetic plot of difructose

dianhydrides formation at 60°C and pH 5.90 with
different initial fructose concentration.

70

Table 2.6 The effect of fructose concentration on the
reaction rate of difructose dianhydrides
formation at 60 °C and pH 5.90

 

 

Initial Rate *
concentration (wt %) constant (K1)

80.3 .00185

84.0 .00073

88.0 .00078

92.0 .00039

 

‘K in units of wt%'1hr'1)
Data were fitted by single straight line.

Concentration (wt %)

 

71

(a) D4

 

 

 

 

 

 

(c) D1+Dz

:107

J
2.0-J

J
‘LOJ

.J
0.0-1 . ,

 

‘1 ‘I ' l V 1* ' I j r ' ‘f
0 48 96 144 192 240 288 336

Time (hr)

Figure 2.11 The effect of initial fructose concentration on

the formation of (a) D4 (b) 03 (c) 01+02 at 60°C
and pH 5.90.

72

concentration on the formation of D4 was greatest, wherein
the amount of D4 formed decreased as the initial fructose
concentration increased from 80 to 92%. The formation of D3
and D1+D2 were less affected with the retardation effect
demonstrated only at initial stage of reaction when the
fructose concentration was still high.

Watanabe [21] has reported the viscosity of aqueous
fructose solution increases an order of magnitude as the
concentration goes from 70 to 90% at 60 0C. Table 2.7 lists
the viscosity values of concentration ranging from 30 to
90%. Although agitation was provided for these highly
viscous solutions, the extent of micromixing might not be
significant and the reactant molecules probably encountered
each other through molecular diffusion. The conditions of
diffusion control for reaction in solution have been
discussed by Moore [22]. For a bimolecular reaction, the
diffusion-limiting second-order rate constant is
proportional to the diffusivity of reactant molecules. With
typical diffusivity value of 10"5 cmzs'l, diffusion control
will take over from reaction (collision) control at about K
- 109 (mol/l)'ls'1. The diffusivity of fructose in dilute
aqueous solution has been reported by Uedaira and Uedaira
[23] at 25 °C and for concentration lower than 3%; the
diffusivity is given as

0 = (7.002-.813C)x10'6 (cmzs'l) (7)
where C is concentration in units of wt%. The diffusivity

data for highly concentrated solutions are not available in

73

Table 2.7 Viscosity of fructose solutions at 60 °C

 

 

ConcentratiOn (wt %) Viscosity (c.p.)
30 1.1
50 2.8
60 5.2
70 18
80 92

90 2100

 

74

the literature. The second-order rate constant found for
the dimer formation reaction is about 10"3 (wt%)‘1hr'1, or
0.5 in terms of (mol/1)'1s'1 approximately. Although the
diffusion-limiting rate constant under crystallization
conditions will be smaller than 109 due to the low
diffusivity value in highly viscous solution, the reaction
rate still seems to be too slow to be limited by diffusion.
However, the rate of mutarotation is relatively fast (in
times on the order of minites [1]) and is more probably
subjected to diffusion control as compared to the rate of
dimer formation reaction. Thus the high concentration may
retard fructose murarotation, which will affect the
composition of solution and then the formation of difructose
dianhydrides. Measurements of fructose diffusivity and
mutarotation rate in saturated and supersaturated aqueous

solution will be needed to prove this assumption.

2.5. Conclusions

1. Fructose underwent dehydration reactions and formed
four difructose dianhydrides under industrial
crystallization conditions.

2. The kinetics of difructose dianhydrides formation
reaction was modelled employing a second-order irreversible
rate equation. The extent of reaction increased with
increasing temperature and decreasing pH value of the

solution.

75

3. The amounts of two of the four difructose
dianhydrides stOpped increasing when the fructose
concentration was about 70% or less. This is probably
caused by the effect of solvation of fructose molecules.

4. The formation of all four dimers was retarded when
the initial fructose concentration was higher than 80%. The
rate of fructose mutarotation may be limited by diffusion in
highly concentrated solution, thus affects the formation of
difructose dianhydrides.

10.

11.

12.

13.

14.

15.

16.

17.

76

References

R.S. Shallenberger, Pure Appl. Chem. 50, 1409 (1978).

D.R. Dwivedi and 3.x. Raniwala, U.S. Patent no.
4,199,373 (1980).

L. Hyvonen, P. Varo and P. Koivistoninen, J. Food Sci.
42, 652 (1977).

R.H. Forsberg, L. Hamalainen, A.J. Melaja and J.J.
Virtanen, U.S. Patent no. 4,199,373 (1975).

Y.D. Chu, L.D. Shiau and K.A. Berglund, "Effect of
impurities on crystal growth in fructose
crystallization", submitted to J. Crystal Growth
(1988).

A. Pictet and J. Chavan, Helv. Chim. Acta. 9, 809
(1926).

W.L. Wolform and M.G. Blair, J. Am. Chem. Soc.
70, 2406 (1948).

W.L. Wolfrom, H.W. Hilton and W.W. Binkley, J. Am.
Chem. Soc. 74, 2867 (1952).

B. Wickberg, Acta Chem. Scand. 8, 436 (1954).

R.F. Jackson and S.M. Goergen, Bur. Stand. J. Res.
3, 27 (1929).

J.C. Irvine and J.W. Steveson, J. Am. Chem. Soc. 51,
2197 (1929). -

R.F. Jackson and E.J. McDonald, Bur. Stand. J. Res.
6, 709 (1931).

J. Defaye, A. Gadelle and C. Pedersen, Carbohydr. Res.
136, 53 (1985).

T. Kusch, W. Gosewinkel and G. Stoeck, U.S. Patent No.
3,513,023 (1970).

T. Yamauchi, U.S. Patent No. 3,928,062 (1975).

V.S. Velasco and J.F. Dowling, U.S. Agri. Res. Serv.
South. Reg., [Rep.], ARS-S-88, 138 (1975).

K. Hamada, H. Yoshihara, G. Suzukamo and 0. Hiroaki,
Bull. chem. Soc. Jpn., 57, 307 (1984).

18.

19.

20.

21.

22.

23.

77

H.W. Hilton, in Methods in Carbohydrate Chemistry, Vol
II, R.L. Whistler and M.L. Wolfrom Ed. (Academic Press,
New York, 1963) p. 199.

R.S. Shallenberger, Advanced Sugar Chemistry (The AVI
Publishing Company, Inc. Westport, CN, 1982) p. 216.

J.H. Pazur, in The Carbohydrates, 2nd Edition, Vol

IIA, W. Pigman and D. Horton Ed. (Academic Press, New
York, 1970) p. 76.

T. Watanabe, Seito Gijutsu Kenkyu Kaishi 28, 70 (1978).

W.J. Moore, Physical Chemistry (Prentice-Hall, Inc.,
Englewood Cliffs, NJ, 1972) p. 405.

H. Uedaria and H. Uedaria, J. Phys. Chem. 74, 2211
(1970).

CHAPTER 3*

EFFECTS OF GLUCOSE AND DIFRUCTOSE DIANHYDRIDES ON
CRYSTAL GROWTH IN FRUCTOSE CRYSTALLIZATION

3.1. Abstract

The influence of impurities on the crystallization of
anhydrous fructose from aqueous solution was studied. The
growth kinetics of fructose crystals in the fructose-water-
glucose and fructose-water-difructose dianhydrides systems
were investigated using photomicroscopic contact nucleation
techniques. Glucose is the major impurity likely to be
present in fructose syrup formed during corn wet milling,
while several difructose dianhydrides are formed in git;
under crystallization conditions and have been proposed as a
cause in the decrease of overall yields.

Both sets of impurities were found to cause inhibition
of crystal growth, but the mechanisms responsible in each
case are different. It was found that the presence of
glucose increases the solubility of fructose in water and

thus lowers the supersaturation of the solution. This is

 

*This chapter contains a paper submitted to "Journal of
Crystal Growth". In this paper, the effects of glucose and
difructose dianhydrides on fructose crystal growth were
studied. The author is responsible for the part related to
difructose dianhydrides.

78

79

probably the main effect responsible for the decrease of
crystal growth. Since the molecular structures of
difructose dianhydrides are similar to that of fructose,
they are probably "tailor-made" impurities. The decrease of
crystal growth is probably caused by the incorporation of
these impurities into or adsorption to the crystal surface
which would accept fructose molecules in the orientation

that existed in the difructose dianhydride.

3.2. Introduction

The monosaccharide D-fructose has great potential use
as a natural sweetener due to its wide abundance and high
sweetness (1.8 times that of sucrose). Unlike other
sweeteners such as sucrose, the crystallization of fructose
is very difficult due to unfavorable physico-chemical
properties [1]. Fructose is much more soluble in water as
compared to sucrose resulting in a highly viscous saturated
solution. For example, 100 g of water can dissolve 662 g of
fructose at 50 °C and the viscosity of this saturated
solution is about 4000 c.p. [2]. On the other hand, only
260 g of sucrose can be dissolved in 100 g of water with a
resulting solution viscosity of 102 c.p. [3]. Fructose
tends to crystallize as hemihydrate and/or dihydrate forms
rather than the anhydrous form under certain conditions.
These hydrate compounds melt near room temperature which
makes them unacceptable as products [4,5]. As a reducing

sugar, fructose undergoes mutarotation in solution as

80

depicted in Figure 3.1,but the only configuration of fructose
in the crystalline state is B-pyranose [6]. B-Furanose and
Q-furanose are also found in significant amounts in aqueous
solution with the equilibrium compositions of fructose
solutions at different temperatures and concentrations given
in Table3.1[7].

In addition to physico-chemical properties, the
presence of impurities in fructose syrup may retard crystal
growth. These impurities include glucose and difructose
dianhydrides which may exist in the fructose syrup in
substantial amounts. Residual glucose is often found after
glucose isomerase conversion of glucose to fructose and
subsequent ion exchange enrichment [8]. Difructose
dianhydrides are formed in sitg under industrial
crystallization conditions and have been reported to
decrease the overall yield in fructose crystallization [9].

The growth kinetics of fructose crystals from the pure
fructose-water system have been studied by Shiau and
Berglund [10]. It was found that the growth of fructose
crystals was size-independent with growth rate dispersion
among different crystals. The Constant Crystal Growth model
was employed and linear growth rates were evaluated from
plots of size versus time. The dependence of growth rate on
supersaturation and temperature was described by a power law
model and Arrhenius relation. The mean growth rate was
expressed as

c = A exp(-E/R'r)sn ' (1)

81

(3?! C)?!
CH on
0” \ CH OH/ 2
HO OH «\ ¢=8 / HO OH

B-D-fructopyranose "3&3” c-D-fructoPYraflose
‘//‘ H¢0H '\\1
HOW: on CHZOH HOCHz 0 011,014
@0120?! @011 7
HO HO
B-D-fructofuranose ceD-fructofuranose

Figure 3.1 Possible tautomers of D-fructose in solution.

82

Table 3.1 Tautomeric equilibria of D-fructose solutions.

 

 

Temperature Concentration c-furanose B-Turanose B-pyranose
(°C) (wt %) (%) (%) (%)
23 20 6 21 73
23 50 4 21 75
23 80 5 21 74
0 20 4 18 78
22 20 6 21 73
67 20 8 28 64

77 20 12 31 57

 

83

where A = 1.5x107 um/hr, E s 6100 cal/g-mole, and n = 1.3.
The variance of the growth rate distribution was
approximated by the following power law model:
062 = aGb (2)

where a = 0.17 (um/hr)°'65 and b = 1.4.

The objective of this study is to determine the effects
of glucose and difructose dianhydrides on the growth of
anhydrous fructose crystals under conditions likely to exist

in commercial crystallization.

3.3. Experimental

Solutions containing glucose were prepared by adding
glucose to pure fructose solutions. The concentration of
impurity was expressed as ratio of impurity to water (I/W).
Solutions containing difructose dianhydrides were prepared
initially by stirring an 80% pure fructose solution at 70 °C
for a certain period of time to generate difructose
dianhydrides. Extra fructose was then added to obtain
supersaturated solutions at the desired temperature.
Solubility data published by the National Bureau of
Standards [11] were used to calculate supersaturation in the
pure system. Since the influence of impurities on fructose
solubility is not clear, the same solubility as in the pure
system was used to calculate supersaturation in the impure
systems.

The concentration of difructose dianhydrides was

84

determined by a Biorad Carbohydrate Analysis HPLC system,
which was equipped with a Bio-Sil Amino SS column and a
refractive index detector. The mobile solvent was
acetonitrile/water with a volume ratio of 70/30 and a flow
rate of 1.0 ml/min. Samples were eluted at room
temperature.

Contact nucleation experiments were conducted to study
crystal growth rates. The experimental apparatus and
technique described by Shanks and Berglund [12] for the
sucrose-water system was used in this work with the growth
cell shown in Figure 3.2. Here the growth rates of crystals
in a stagnant cell were studied; since the aqueous fructose
solutions were highly viscous, stirring is unlikely to have
an effect on mass transfer. Contact nuclei were created by
sliding the parent crystal along a glass plate in the growth
cell with the resulting growth of nuclei monitored
photomicroscopically. Photographs were subsequently
analyzed by an image analyzer to determine the area of each
crystal and the equivalent circular diameter was taken as
the characteristic size. The experimental conditions for
the fructose-glucose and fructose-difructose dianhydrides
systems are listed in Tables 3.2 and 3.3.

Fructose crystals growing in the presence of glucose or
difructose dianhydrides were collected after completion of
the growth experiments. The adhering mother liquor was
washed from the surface by saturated alcohol solution. The

crystals were dissolved and the solution was analyzed for

85

 

(S)
10? VIEW

 

SIDE VIEW

Figure 3.2 Schematic diagram of nucleation cell with the
features:
(1) chamber containing solution: (2) parent
crystal: (3) glass cover slip where parent
crystal is slid: (4) support rods for glass cover
slip; (5) thermistor; (6) movable rod holding
parent crystal: (7) chamber containing constant
temperature water; and (8) water inlet and outlet.

86

Tablei3yz Conditions of crystal growth experiments for fructose-
water-glucose system at 40 °C.

 

 

Impurity/Water Supercooling No. of nuclei
Exp ratio (I/W) (0C) analyzed
1 0.05 1 12
2 0.05 3 l3
3 0.05 5 15
4 0.05 7 ll
5 0.05 9 l4
6 0.3 l 13
7 0.3 3 12
8 0.3 5 ll
9 0.3 7 15
10 0.3 9 12
11 0.3 11 13
12 0.6 5 14
13 0.6 7 13
14 0.6 9 13.
15 0.6 11 16
16 0.9 5 l8
17 0.9 7 16
18 0.9 9 12
19 0.9 11 15

 

Table:3.3 Conditions of crystal growth experiments for fructose-
water-difructose dianhydrides system.

87

 

 

 

Exp Temperature Supersaturation Difructose dianhydrides (wtfii N0. of nuclei
(OC) (C-Ce)/Ce _EEEET____5?352 53 En analyzed
1 30 .033 1.6 .6 .6 .1'1 21
2 30 .055 1.2 .5 .4 .3 19
3 30 .068 1.2 .5 .5 .2 13
4 40 .039 3.0 .9 1.3 .8 13
5 40 .048 3.3 1.2 1.3 .8 10
6 40 .041 3.2 1.1 1.2 1.0 14
7 40 .021 4.5 .8 1.5 2.2 12
8 40 .045 4.6 .8 1.5 2.3 12
9 40 .045 5.5 1.3 1.9 2.3 25
10 50 .028 4.4 1.8 1.7 .9 14
11 50 .049 4.4 1.8 1.7 .9 12

 

lots: 01, 02, D3, Du represents diheterolevulosan II, I, III, IV
respectively.

88

the presence of impurities by the HPLC technique described

above.

3.4. Results and Discussion

Figure 3.3 shows examples of size versus time plots for
several crystals grown in the presence of impurity. It was
demonstrated that the growth rate of a single crystal was
constant while different crystals had different growth
rates. The phenomena of size-independent growth and growth
rate dispersion were observed for all impurity
concentrations in both the glucose and the difructose
dianhydrides systems. Thus, the Constant Crystal Growth
model used in the pure system was also employed in the

impure systems.

3.4.1. Effect of glucose on fructose crystal growth

From a plot of size versus time, the linear growth rate
was evaluated for each single crystal. The mean growth rate
and the variance of the growth rate distribution were then
calculated for each level of supersaturation at 40 oC.

The relationship between mean growth rate and
supersaturation at different values of impurity/water (I/W)
ratio is shown in Figure 3.4. The mean growth rate was
decreased to 25 to 50% of that in the pure system by the
presence of glucose at concentrations ranging from 0.05 to
0.9 in terms of I/W ratio. This can be described by the

following equation

 

89

 

:50-1

25~

/
O

’5 20— ‘ X
a.
S
.6; 15‘ A

10- A /0

5-1 a

00 5'411'5'8'1'0

Time (hr)

Figure 3.3 Size versus time plot of fructose contact nuclei

grown in the presence of glucose. Temperature =
40 C, supercooling = 5 °C and impurity/water
ratio = 0.3. Each line represents a single crystal.

Growth Rate (um/hr)

90

101

PURE
vw-ams

09x00

 

 

 

Supersaturation

Figure 3.4 The influence of glucose on the relationship

betgeen mean growth rate and supersaturation at
40 C.

91

c - A exp(-r/RT)(s-2.7x10‘4(I/W)°°39)n exp(-0.00981/W) (3)
Here A, E and n are assumed to be the same as in the pure
fructose system. The estimation of parameters is summarized
in Table 3.4.

It was observed that nuclei generated in a solution of
high supersaturation would dissolve at a temperature lower
than the supposed saturation temperature of the solution.
Nuclei were generated and grew only when supersaturation
reached a critical value which increased as impurity/water
ratio increased. This phenomenon suggests that solubility
of fructose increases with an increase of glucose
concentration. In the equation of mean growth rate, the
term 2.7x10'4(I/W)°°89 is subtracted from S to account for
this increase of solubility. Besides the solubility change,
the presence of glucose could also affect volume diffusion
in solution and/or surface integration of crystal molecule.
The need for the inclusion of exp(-0.0098I/W) in the model
may represent these effects, but the scatter of the data
precludes any definitive comment to be made.

Figure 3.5 shows the variance of the growth rate
distribution at 40 °C. A power law model was used to fit
the data resulting in following equation

as2 - 0.1661'4 (4)
The statistics pf regression is summarized in Table 3.5. No
significant difference of the variance of growth rate
distribution was found between pure and glucose impurity

systems.

92

Table 3.4. Summary of parameter estimation for growth rate data

 

 

 

Data Model parameters. Residual
sun of
a b 0 squares

(um/hr)2
Glucose 1. G:Aexp(-E/RT)S" - - - 383.58
2. G:Aexp(-E/RT)(S-a(I/W)b)" 4.58-4 0.85 - 11.90

3. G=Aexp(-E/RT)(S-a(I/W)b)n
exp(—cI/W) 2.78-4 0.89 9.8E-3 3.39
(equation (3))

Difructose 1. 0=Aexp(-E/Rr)s“ - - - 2274.35
dianhy-
drides 2. G:Aexp(-E/RT)S"exp(-aI/W) -23 - - 9.67

(equation (5))

 

'Nonlinear parameters were estimated by minimizing residual sum
of squares.

93

 

07 ”J —— o PURE
“ _— A GLUCOSE
fi ---- x DIFRUCTOSE DIANHYDRIDES
Ne 2.5-

:1

m

1; 2.0-«

m

.c

E:

o 1.5“

La

:9

H-l

o 1.0-4

m

u

c

m

0H 0.5-

H

m

>

0.0-1

 

 

"""""7‘.0'8'.0
0.0 1.0 210 3.0 4.0 5.0 6.0

Growth Rate (um/hr)

Figure 3.5 The influence of impurities on the variance of
growth rate.

94

Table 3.5. Summary of stat stigs*of regression equations
(4) and (6): 3G =aG

 

 

 

Equation Correlation parameters i standard error
coefficient
lna b
(4) 0.95 -l.86_~1_-_0.28 1.4_+_0.1
(6) 0.89 -2.64i0.18 1.710.l

 

*Linear parameters were estimated by ordinary least square
method after logarithmic transformation of the equation.

95

3.4.2. Effect of difructose dianhydrides on fructose
crystal growth

Fructose undergoes irreversible dehydration reactions
in acid solution or at high temperature and forms several
difructose dianhydrides. These dianhydrides possess a
central 1,4-dioxane ring as shown in Figure 3.6 with at least
eight difructose dianhydrides having been reported [13,14].
Their structures differ in the configuration of each
fructose moiety and the linkage between them, as listed in
Table 3.6.

Since the rates of crystal growth are so low the
crystallization step in fructose manufacture requires time
on the order of days. The temperature range involved is 30
to 60 oC and pH is controlled between 4 and 6 [5,9,15,16].
Under these conditions, four difructose dianhydrides,
diheterolevulosan I - IV, were found by HPLC analysis with a
typical chromatogram shown in Figure 3.7. The appearance of
these four dimers is expected since all of them have at
least one b-fructopyranose, which is the major form of
fructose in solution as shown previously in Table 3.1. When
kept at 60 0C for two weeks, an 80% fructose solution with a
pH value of 4.35 formed about 6% dianhydrides in total.

Since difructose dianhydrides are not commercially
available, the preparation of supersaturated solution for
studying the dianhydrides impurities was different from that
used for glucose. Thus, the supersaturation and impurity

concentration could not be similarly controlled for the two

96

 

Figure 3.6 Diheterolevulosan I.

97

Table 3.4 Known difructose dianhydrides.

 

fructose moiety

 

trivial name

 

 

' . _ linkage
ring anomeric C

pyr-pyr 8, c 1,2?2,1

fur 8, c 1,2:2,1

fur 8, 8 1,2:2,1

DY? B. B 1,2:2,1

DY? B. B 1,2:2,3

fur-fur c, 8 1,2:2,1

fur - 1,2:2,4

fur a, 8 1,2:2,3

diheterolevulosan I
II
III
IV
difructose anhydride I
II

III

 

98

 

 

 

—_—-‘ L-*Ji2 34 5

l I 1 ' '

111‘1'1
0 2 4.6 81012

 

 

 

4 Retention Time (min)

Figure1357 HPLC analysis of difructose dianhydrides.
1. Fructose 2. diheterolevulosan II 3.
diheterolevulosan I 4. diheterolevulosan III
diheterolevulosan IV.

99

sets of impurities.

Growth rate was normalized by division by the growth
rate of the pure system so that the effect of impurity could
be evaluated from data of different supersaturation and
temperature levels. Figure 3.8 shows the relationship between
normalized mean growth rate and impurity/water ratio (I/W).
The effect of difructose dianhydrides on crystal growth
inhibition was greater than that of glucose with a rapid
decrease in growth rate occurring over a narrow range of
impurity concentration. The growth rate was only one tenth
of that in the pure system at an impurity/water ratio of
0.1, or 0.006 in terms of mole fraction of total solid. The
following equation is applied to fit the data:

G - A exp(-E/RT)Sn exp(-9.91/W) (5)
Parameters A, E and n are the same as in the pure system.
The estimation of parameters is summarized in Table 3.4.

The impurity (I) used was the total amount of difructose
dianhydrides. Individual dianhydride levels were studied,
but no significant difference among the dianhydrides was
demonstrated by the experimental data.

The variance of the growth rate distribution is shown
in Figure 3.5 along with that of pure and glucose impurity
systems. The equation representing the variance data is

sG2 = 0.07161'7 (6)

The statistics of regression is summarized in Table 3.5.
Growth rate dispersion appears to be less in the difructose

dianhydrides impurity system, but the exponents in the

Normalized Growth Rate

 

100

 

 

Impurity/water ratio

Figure 3.8 The effect of difructose dianhydrides on
normalized mean growth rate.

101

variance equations of the two impure systems were compared
with that of the pure system by a t-test and no significant
difference was demonstrated in both cases (P = 0.05).
However, it should not be concluded by the comparison alone
because the range of growth rate data in dianhydrides system
is much smaller than those of the pure and glucose impurity
systems.

A small amount of difructose dianhydrides (< 1 %) was
found by HPLC analysis in fructose crystals grown from‘
solution containing these impurities (Figure 3.9). Crystals
grown in the presence of glucose were also analyzed;
however, glucose was not detected. The appearance of
impurities in the crystal may be caused by inclusion of
mother liquor as well as incorporation of impurity molecules
into lattice of the growing crystal. The amount of mother =
liquor inclusion should be negligible because fructose
crystals grow very slowly in the presence of impurities (0.5
to 2.0 um/hr in difructose dianhydrides impurity system) and
is further supported by the absence of glucose in crystals.
Thus, it is suggested that molecules of difructose
dianhydrides were incorporated into the fructose crystal.

Difructose dianhydides consist of two fructose moieties
and thus possess some of the structural characteristics of
fructose molecules. This suggests that the possible
mechanism of growth retardation occurring in the difructose
dianhydrides impurity system is the so-called "tailor-made

crystal growth inhibition", which means the structural

102

(3) solution

1 “ 1

 

234 5

 

 

 

.44

(b) crystal

F F 1

 

 

 

 

awn;

'l'l'l‘l'l'l
0 2 4 6 8101214

 

 

Retention Time (min)

Figure 3.9 HPLC analysis of difructose dianhydrides
incorporation. l. Fructose 2. diheterolevulosan
II 3. diheterolevulosan I 4. diheterolevulosan
III 5. diheterolevulosan IV.

103

similarity of the impurity and crystal molecules leads to
incorporation of impurity into the growing surface. This
will disrupt the bonding sequences in the crystal and
interfere with subsequent incorporation of crystal molecules
[17,18]. Based on HPLC analysis, the incorporation of
difructose dianhydrides had no preference among the four
dianhydrides. The effects of growth inhibition exerted by
different dianhydrides were also found to be similar. Since
all of these four dimers have at least one moiety of B-
fructopyranose, which is the form of crystalline fructose,
all of them are expected to be occluded into the crystal and
retard the growth of crystal similarly.

The morphologies of fructose crystals growing from the
pure and two impure systems were compared and no significant
difference was found. Since the sizes of crystals studied
were less than 50 mm, differences in crystal habit probably
could not be detected, and further observation on larger
crystals is necessary to show the effect of these impurities
on crystal habit modification.

3.5. Conclusions

1. The growth rate of fructose crystals in the
presence of impurities was size-independent, with growth
rate dispersion among different crystals.

2. Growth inhibition was observed in both impure

systems. At the same impurity level, the mean growth rate

104

in the presence of difructose dianhydrides was lower than
that in the presence of glucose.

3. The growth inhibition is probably caused by an
increase of solubility by glucose. Difructose dianhydrides,
on the other hand, appear to be incorporated into the
crystal, thus inhibiting subsequent surface integration of
fructose molecules. .

4. The effects of growth inhibition exerted by
different difructose dianhydrides were similar, probably
because they have the same structural characteristics of

crystalline fructose molecules.

105

Notation

:1
I

210-3012!
I

Greek

frequency factor, um/hr

correlation constants

concentration, g fructose/100 g solution
saturation concentration, g fructose/100 9 solution
activation energy, cal/g-mole

mean linear crystal growth rate, um/hr
impurity, g

growth rate order

ideal gas constant, 1.987 cal/g-mole °K
supersaturation, (C-Ce)/Ce

temperature, °K

water, 9

Symbol
variance of growth rate distribution, umz/hr2

3.6.

10.

11.

12.

13.

14.

15.

16.

17.

18.

106

References

G.N. Bollenback, Kirk-0thmer Encyclopedia of Chemical
Technology (John Wiley & Sons, New York, 1983).

T. Watanabe, Seito Gijutsu Kenkyu Kaishi 28 (1978) 70.
J. Timmermans, The Physico-chemical Contants of Binary
Systems in Concentrated Solutions (Interscience
Publishers, New York, 1960).

F.E. Young, F.T. Jones and H.J. Lewis, J. Phys. Chem.
56 (1952) 1093.

T. Yamauchi, U.S. Patent No. 3,928,062 (1975).
T.E. Dotty and E. Vanninen, Food Technol. 11 (1975) 34.

L. Hyvonen, P. Varo and P. Koivistoninen, J. Food Sci. 42
(1977) 652.

B. Vanninen and T. Dotty, Sugar: Science and
Technology (Applied Science Publishers, London, 1979).

R.H. Forsberg, L. Hamalainen, A.J. Melaja and J.J.
Virtanen, U.S. Patent no. 4,199,373 (1975).

L.D. Shiau and K.A. Berglund, AIChE J. 33 (1987) 1028.
F.J. Bates, Polarimetry, Saccharimetry and the Sugars
(United States Government Printing Office, Washington,
1942).

8.3. Shanks and K.A. Berglund, AIChE J. 31 (1985) 152.

R.S. Shallenberger, Advanced Sugar Chemistry (The AVI
Publishing Company, Inc. Westport, CN, 1982).

J. Defaye, A. Gadelle and C. Pedersen, Carbohydr. Res.
136 (1985) 53.

T. Kusch, W. Gosewinkel and G. Stoeck, U.S. Patent No.
3,513,023 (1970).

B.K. Dwivedi and S.K. Raniwala, U.S. Patent no.
4,199,373 (1980).

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

R. Davey, W. Fila and J. Garside, J. Crystal Growth 79
(1986) 607.

SUMMARY AND CONCLUSIONS

The kinetics of difructose dianhydrides formation under
industrial crystallization conditions and the effect of
difructose dianhydrides on the crystal growth of anhydrous
fructose from aqueous solution were investigated.

The kinetics of difructose dianhydrides formation were
determined by HPLC analysis. Conclusions are:

1. Fructose underwent dehydration reactions and formed
four difructose dianhydrides under industrial
crystallization conditions.

2. The kinetics of difructose dianhydrides formation
reaction was modelled employing a second-order irreversible
rate equation. The extent of reaction increased with
increasing temperature and decreasing pH value of the
solution.

3. The amounts of two of the four difructose dianhydrides
stopped increasing when the fructose concentration was about
70% or less. This is probably caused by the effect of
solvation of fructose molecules.

4. The formation of all four dimers was retarded when the
initial fructose concentration was higher than 80%. The
rate of fructose mutarotation may be limited by diffusion in
highly concentrated solution, thus affects the formation of

difructose dianhydrides.
107

108

The kinetics of fructose crystal growth in the presence
of difructose dianhydrides were studied using
photomicroscopic contact nucleation techniques. Conclusions
are:

1. The growth rate of fructose crystals in the presence of
difructose dianhydrides was size-independent, with growth
rate dispersion among different crystals.

2. The presence of difructose dianhydrides caused
inhibition of fructose crystal growth.

3. Difructose dianhydrides appeared to be incorporated
into the crystal, thus inhibiting subsquent surface
integration of fructose molecules.

4. The effects of growth inhibition exerted by different
difructose dianhydrides were similar, probably because they
have the same structural characteristics of crystalline

fructose molecules.

RECOMMENDATIONS

l. The diffusivity of fructose in aqueous solution
under crystallization conditions should be measured to
understand the transport of fructose molecules and the
kinetics of both chemical reaction and crystallization.

2. The kinetics and equilibrium of fructose
mutarotation should be investigated to determine the actual
composition of solution under crystallization conditions.

3. The solvation of fructose molecules at different
concentrations and its relationship to the formation of
difructose dianhydrides should be studied.

4. The packing arrangement of fructose molecules in the
crystal should be studied to understand the incorporation of
difructose dianhydrides in the surface of the growing
crystal.

5. Single crystal growth experiments should be
performed to determine the growth rates of different
crystallographic faces and the effect of difructose
dianhydrides on the growth rates.

6. A mathematical model for crystal growth under
industrial crystallization conditions should be developed
employing the kinetics of crystal growth in the presence of
difructose dianhydrides and the kinetics of the formation of

difructose dianhydrides.

109

APPENDICES

110

APPENDIX A

METHODS AND MATERIALS

Kinetics of Difructose Dianhydrides Formation Reaction

1. Reaction

Experiments were conducted at 30, 40, 50 and 60 °C.
Hydrochloric acid solutions of pH 2.65, 4.35 and 5.90 were
first prepared. Fructose was added to these solutions to
get the desired concentrations, which were 10% less than the
solubility of fructose at the reaction temperature.
Analytical grade reagents and distilled-deionized water were
used.

250 g fructose solution was added to a 250 ml flask.
The flasks were put into a water bath controlled by a Haake
constant temperature circulator model E3. The solutions
were agitated by a Curtin Matheson Scientific 244-793
magnetic stirrers for a period of two weeks.

About 2 g samples were taken on a daily basis. After
weighting they were diluted immediately with water in a 50
ml volumetric flask and then refrigerated to minimize

further reaction.

2. HPLC analysis

The sample solutions were filtered through a Millipore

111

Millex-HA disc filter of size 0.45um and then analyzed on a
Biorad Carbohydrate Analysis HPLC system. The system
consists of a Model 1330 pump, a Model 1770 refractive index
detetor and a Model 3392A integrator. The column was a
250x4.0 mm Bio-Sil Amino-SS. A 30x4.6 mm guard column which
had the same packing material was installed in front of the
analysis column. Samples of 5 01 were injected and eluted
at room temperature with 70% (v/v) acetonitrile. The eluant
was filtered and degassed with a Millipore All Glass Filter
Apparatus using a 0.5 um Fluoropore membrane.

The attenuation range of the detector was set at 16X
(16x4x10'6 RIU/FS). The peak controls of integrator were
set as width a 0.16 min, area rejection = 0 area count (1
area count = 0.125 mV-sec) and threshold = 9 (29x1.25x10"4
mV). The analysis was calibrated by external standard
method. The standard solution was a mixture of fructose and
sucrose with a concentration of 1.0 g/100 ml for each sugar.
The response factor of sucrose was used to calibrate the
amounts of difructose dianhydrides. The sample solutions
were further diluted with water for the determination of
residual fructose concentration if the initial analysis
indicated that it was out of the range of column capacity.

The concentrations of dianhydrides were based on peak
area. The concentrations of residual fructose was based on
peak height. The concentration data were transformed from
g/100 ml to wt% using the data of sampling weight and

dilution ratio.

112

3. GC analysis

A 2 g of sample was diluted with water in a 50 m1
flask. 0.5 ml of the resulting solution was lyophilized and
then dissolved in 1 ml TRI-SIL ’2' purchased from Pierce
(trimethylsilylimidazole in dry pyridine, 1.5 meg/m1).

The GC/MS analysis was performed by the Mass
Spectrometry Facility in Department of Biochemistry,
Michigan State University. The spectra were taken by a HP
5985 GC/MS system which was equipped with electron
multiplier detector with the ionization potential at 70 eV.
The column was a 6 ft 3% 0V-225 and the temperature program

was 140 - 200 °C (3 °/min).

Effect of Difructose Dianhydrides on Fructose Crystal Growth

1. Supersaturated solution

Crystal growth experiments were conducted at 30, 40 and
50 °C. An 80% fructose solution was first stirred at 70 °C
for 24 or 48 hr to generate difructose dianhydrides. A
large excess of fructose was added to the resulting solution
and stirred for at least 24 hr at different temperatures to
provide the desired degree of supersaturation. The
solutions were filtered through a nylon mesh of 15 um and
then kept at the previous temperature (used in dissolving
fructose) until they were free of crystals when checked by a
microscope with 200x magnification. The concentration of
fructose and difructose dianhydrides were analyzed on the

HPLC system described earlier in this section.

113

2. Contact nucleation

A parent crystal of size of approximately 2 mm was
glued to the movable rod with epoxy. Solution was added to
the cell and heated at a temperature of 10 °C higher than
the experimental temperature for 30 min to dissolve any
nuclei that may have been generated in the solution and to
remove surface irregularities of the parent crystal. The
solution was brought to the desired experimental temperature
and the parent crystal was slid across the glass plate to
generate contact nuclei. Photographs were taken at
intervals and the experiment was stopped after the crystal
sizes reached approximately 40 um. The magnification of the
microscope was 200x; a 50x2 um scale was photographed and
used to calibrate the analysis.

The raw data obtained from each experiment consisted of
a series of slides, which were analyzed on a Joyce-Loebl
Magiscan 2 image analyzer to determine the area of each
crystal. The equivalent circular diameter was taken as the
characteristic size and linear growth rate was evaluated by

ordinary least square estimation.

114

APPENDIX 8

ORIGINAL DATA

Kinetics of Difructose Dianhydrides Formation Reaction

Concentration (wt%) vs. Time (hr)
T: Temperature, IC: Initial concentration, F: Fructose

Table 81. T = 60 C, pH = 2.65, IC = 80.3%

115

Table 32. T = 60 C, pH = 4.35, IC = 80.3%

Table B3. T = 60 C, pH I 5.90, IC 3 80.3%

116

II
\I
O)
)0
09

Table 84. T - 50, pH = 2.65, IC

Table BS. T I 50 C, pH 8 4.35, IC = 76.9% ’

117

Table B6. T 3 50 C, pH = 5.90, IC = 76.9%

Table B7. T = 40 C, pH = 2.65, IC 8 74.4%

118

Table B8. T = 40 C, pH = 4.35, IC = 74.4%

Table 39. T = 40 c, pH = 5.90, IC = 74.4%

 

119

Table 310. T = 30 c, pH = 2.65, IC = 71.7%

Table 311. T a 30 c, pH - 5.90, IC = 71.7%.

120

II
\3
U
as

Table 812. T I 60 C, pH I 2.65, IC

Table B13. T I 60 C, pH = 2.65, IC = 70.0%

121

Table 814. T I 60 C, pH I 2.65, IC = 65.0%

Table BIS. T I 60 C, pH = 5.90, IC = 92.0%

122

Table 816. T = so c, pH = 5.90, IC = 88.0%

Table 817. T I 60 C, pH I 5.90, IC = 84.0%_

123

Effect of Difructose Dianhydrides on Fructose Crystal Growth

Crystal size (um) vs. Time (hr)
(note: Time is listed in descending order in the following
tables.)
T: Temperature, S: Supersaturation, D: Difructose dianhydrides

Table B18. T I 30 C, S I 0.053, D I 1.6%

1
2
3
4 16.53 16.79 15.28 12.33 14.43 10.34 8.06 8.40 6.88
5 24.25 23.04 21.22 24.16 14.73 13.60 10.28 9.45 7.99
6 29.52 27.39 26.76 17.40 18.06 17.62 13.56 11.35 9.96
7 19.77 20.43 17.02 15.67 13.36 11.59 9.56 8.72 7.27
8 41.16 38.46 35.91 29.81 25.81 22.19 19.61 15.92 12.94
9 37.55 33.07 31.22 26.07 23.47 19.41 17.37 15.32 13.80
10 21.25 20.46 20.16 16.96 14.80 12.19 9.21 8.06 6.12
11 24.09 22.82 19.80 16.96 22.06 21.84 8.33 6.88 6.47
12 19.86 16.92 17.75 14.16 12.24 9.09 6.73 7.42 5.25
13 32.95 33.00 29.18 26.45 13.07 '11.97 10.18 12.90 11.01
14 30.07 28.14 25.44 19.63 14.47 ‘12.42 9.45 6.72 7.99
15 23.28 22.11 20.43 18.89 15.03 12.46 11.45 7.11 7.27
16 25.92 22.36 22.09 17.28 12.90 11.78 9.39 7.27 5.55
17 23.51 23.28 19.49 15.50 12.19 11.59 9.15 9.09 6.30
18 21.22 19.27 17.44 13.15 11.69 10.39 9.96 6.72 5.45
19 36.17 34.36 32.35 27.83 21.56 18.60 16.12 13.36 10.81
20 23.49 21.17 19.72 17.99 13.88 11.87 9.62 7.78 7.42
21 20.08 18.83 15.78 13.36 9.68 9.45 7.12 6.03 4.06

124

Table 819. T I 30 C, S I 0.055, D I 1.2%

125

Table 821. T - 40 C, S - 0.039, D - 3.0%

....

No. 40. 36. 32. 28. 24. 21. 17. 13. 9
54 27.05 26.15 25.50 22.74 21.19 22 50 18.43 15 80 12.
55 26.36 24.23 22.16 20.69 18.86 16 33 12.97 9.66 10
56 17.33 16.04 13.60 14.67 14.50 13.73 14.27 10.77 9
57 36.58 33.19 30.99 29.36 27 40 26.42 22 83 21.29 18.
58 22.52 22.46 19.10 16.43 17 18 14.35 12 27 10.62 10
59 19.63 20.59 18.16 16.09 15 26 14.73 9 88 10.09 7
60 20.07 18.30 16 71 13.23 13.60 13.48 13.16 10.25 7.
61 17.88 15.35 13.29 12.77 11.49 12 10 10.09 8.15 6.
62 21.51 17.95 17.69 14.41 11.82 12 54 11 00 9.40 7.
63 26.99 26.32 23.91 21.33 19.06 13.54 11 12 10.21 9.
64 22.65 22.24 20.77 17.78 17.40 17.85 16 35 13.66 10.
65 27.63 23.33 22.93 20.53 19.73 19.02 16.83 14.70 11.
66 20.06 20.07 17 54 16.02 14.24 15.48 13.63 11.85 8.
Table 822. T I 40 C, S I 0.048, D I 3.3%

NO. 12. 10. 8.25 6. 5. 4. 3.

67 28.03 23.90 20.87 16.35 13.51 11.99 8.50
68 27.02 24.33 19.43 13.42 11.96 10.77 7.94
69 34.40 25.91 21.09 16.86 13.00 10.54 8.30
70 26.88 21.60 17.78 14.38 12.20 9.57 7.22
71 29.42 23.45 19.60 13.19 11.31 9.17 7.22
72 29.73 27.33 20.40 14.98 13.45 11.38 7.78
73 28.15 25.61 23.74 16.68 12.41 10.85 9.12
74 20.61 17.25 14.50 11.64 8.45 6.86 7.62
75 24.47 22.91 20.30 15.91 13.10 7.78 7.99
76 21.45 20.09 16.63 12.57 9.96 7.99 7.39

126

Table 823. T I 40 C, S I 0.041, D I 3.2%

 

127

Table 825. T - 40 C, S - 0.045, D - 4.6%

128

Table 827. T - 50 C, S - 0.028, D - 4.4%

   

n11111'11111111111111171