DEGRADATION OF ASCORBIC ACID IN A DEHYDRATED

MODEL FOOD SYSTEM DURING STORAGE

BY

Daniel B. Dennison

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Food Science and Human Nutrition

1978

ABSTRACT

DEGRADATION OF ASCORBIC ACID IN A DEHYDRATED
MODEL FOOD SYSTEM DURING STORAGE

BY

Daniel B. Dennison

The storage stability of reduced and total
ascorbic acid in a dehydrated model food system designed
to simulate a ready-to-eat breakfast cereal was studied
as a function of water activity, moisture content, oxygen
and temperature. Sorption isotherm data obtained at 10,
20, 30 and 37°C for the model system were used to calcu-
late the Brunauer, Emmet, Teller (BET) monomolecular
moisture content at these respective temperatures. The
model system, which contained 11.25 mg of ascorbic acid
per 100 g dry weight basis (25% Recommended Dietary
Allowance, RDA) was equilibrated at water activities
below, at and above the water activity corresponding
to the calculated BET monomolecular moisture content for
the adsorption isotherm. The samples were sealed in
thermal death time (TDT) and 303 cans and stored under
isothermal conditions to prevent any shift in water

activity. The TDT cans were filled with the model system

Daniel B. Dennison

so that no headspace remained and limited gaseous oxygen
was available due to the inter and intrastitial spaces.
The 303 cans were filled with an equal mass of model sys-
tem as the TDT, thus providing a large excess of gaseous
oxygen in the headspace. The effect of air and moisture
vapor transmission on the storage stability of ascorbic
acid in the model system was studied by packaging the
model system in l-oz. paperboard boxes with waxed liners.
The samples were stored at 30°C and 10, 40 and 85% rela-
tive humidities.

Reduced, dehydro and total ascorbic acid concen-
trations were determined as a function of storage time by
an automated o-phenylenediamine fluorometric assay pro—
cedure. Under all storage conditions, reduced and total
ascorbic acid losses could be satisfactorily described
by first-order kinetics.

Ascorbic acid destruction in the model system
stored in cans or boxes was linearly dependent on the
water activity, exhibiting maximum stability below the
BET monomolecular moisture content. These results con-
tradict the hypothesis that the BET monomolecular moisture
content should represent the water activity of greatest
stability.

Comparison of the rate constants for total and
reduced ascorbic acid degradation in TDT and 303 cans at

identical conditions of water activity and storage

Daniel B. Dennison

temperature showed a dependence on the availability of
gaseous oxygen. The results were interpreted in terms

of the consumption of dissolved oxygen in the degradation
of ascorbic acid and the transfer of gaseous oxygen into
the product moisture which would be governed by the equil-
ibrium constant K = (02)d/(02)g, where (02)d and (02)g
represent the concentration of dissolved and gaseous
oxygen present in the system, respectively.

The influence of riboflavin and vitamin A on the
degradation of ascorbic acid in the dehydrated model
system was studied as a function of water activity and
storage temperature. Ascorbic acid degradation was found
to be unaffected by the presence of either vitamin.

The catalytic influence of trace minerals (Fe,
Cu, Zn, Ca) on the rate of ascorbic acid degradation was
studied as a function of water activity. No catalysis
by the added metals was observed as 0.10 and 0.40 aw,
except for copper at 0.40 aw. This is interpreted as a
lack of metal ion mobility and/or insolubility at low
water activities. At 0.65 aw, which was in the capillary
region of the adsorption isotherm, a 2-3 fold increase
in degradation rate over the nonfortified system was
observed for each form of added trace mineral with the
exception of zinc oxide, which did not exhibit catalysis.

These results are explained of the basis of Fe and Cu

Daniel B. Dennison

ion mobility in the capillary region of the isotherm,
whereas Ca and Zn catalysis are interpreted as altering
the activity coefficient of oxygen.

Activation parameters were calculated for the
degradation of ascorbic acid in the model system stored
in metal cans. The free energy of activation remained
constant at all aw, thus it was concluded that the degra—

dation of ascorbic acid followed only one mechanism.

ACKNOWLEDGMENTS

The author wishes to express special gratitude
to Dr. James R. Kirk. His advice, personal direction
and expert counsel provided invaluable guidance through-
out this investigation and during the writing of this
dissertation.
The author also expresses grateful appreciation
to the members of his graduate committee, Drs. J. R.
Brunner, D. R. Heldman, L. L. Bieber and R. W. Luecke
for their review of the manuscript. A special note of
appreciation is extended to the members of the Department
of Food Science and Human Nutrition, Michigan State Uni-
versity, for their continual interest and encouragement.
The author expresses his most sincere appreciation
to his wife, Judy, for constant inspiration and encourage-

ment throughout his graduate studies.

ii .

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . .
LIST OF FIGURES . . . . . . . . . .
INTRODUCTION . . . . . . . .
LITERATURE REVIEW . . . . . . . . .
Sorption Phenomena . . . . . . .

Water Activity and the Sorption Isotherm
Theoretical Treatments of the Sorption

Isotherm . . . . . . . . . .
Physical State of Water . . . .
Thermodynamic Activation Parameters. . .

Reactions in Dehydrated Foods. . . . .

Microbiological. . . . . . . . .
Enzymatic. . . . . . . . . . .
Chemical . . . . . . . . . . .

Ascorbic Acid . . . . . . . . . .

Structure and PrOperties. . . . . .
Reactivity in Solution . . . . . .
Stability in Food Products . . . . .

Predictions of Product Storage Stability .
EXPERIMENTAL PROCEDURES . . . . . . .

Model Food System Composition. . . .
Model Food System Preparation. . . .
Addition of Other Nutrients to the Model
System . . . . . . . . . . .
Model Food System Equilibration . . . .
Model System Packaging and Storage . . .
Moisture Content Measurement . . . . .
Ascorbic Acid Determination . . . . .
Data Analysis . . . . . . . . . .
Experimental Design . . . . . . . .

iii

Page

vii

17
20

22
23
24

25

26
29
35

38
42

42
42

43
44
48
49
49
50
51

Page

Ascorbic Acid Stability in TDT Cans . . . . 52
Ascorbic Acid-—Vitamin Interactions in
TDT Cans . . . . 53
Ascorbic Acid Stability in Paperboard
Boxes . . . . . . 53
Ascorbic Acid Stability in 303 Cans . . . . 53
Ascorbic Acid--Mineral Interaction in
303 Cans . . . . . . . . . . . . 54
RESULTS . . . . . . . . . . . . . . 55
Ascorbic Acid Stability in TDT Cans. . . . . 55
Ascorbic Acid--Vitamin Interactions in TDT
Cans . . . . . . 62
Ascorbic Acid Stability in Paperboard Boxes . . 65
Ascorbic Acid Stability in 303 Cans. . . . . 69
Ascorbic Acid--Minera1 Interaction in 303
Cans . . . . . . . . . . 72
Thermodynamic Activation Parameters. . . . . 76
DISCUSSION. . . . . . . . . . . . . . 80
Ascorbic Acid Stability in TDT Cans. . . . . 80
Ascorbic Acid--Vitamin Interaction in TDT
Cans . . . . . . 84
Ascorbic Acid Stability in Paperboard Boxes . . 86
Ascorbic Acid Stability in 303 Cans. . . . . 91
Ascorbic Acid-~Mineral Interaction in 303
Cans . . . . . . . . . 103
Thermodynamic Activation Parameters. . . . . 107
SUMMARY AND CONCLUSIONS . . . . . . . . . 118
REFERENCES. . . . . . . . . . . . . . 122

iv

LIST OF TABLES
Table Page
1. Composition of model food system . . . . . 43

2. Mineral supplement, source and %RDA employed
in mineral fortification study . . . . . 45

3. Rate constants and half-lives for TAA and
RAA loss as a function of water activity
and storage temperature in ascorbic acid
fortified dehydrated model food system
packaged in thermal death time cans . . . 58

4. Activation energies for TAA and RAA loss in
ascorbic acid fortified dehydrated model
food systems packaged in TDT cans . . . . 62

5. Rate constants and half-lives for TAA and
RAA loss as a function of water activity
and storage temperature in multivitamin
dehydrated model food system packaged in
thermal death time cans . . . . . . . 64

6. Activation energies for TAA and RAA loss in
multivitamin fortified dehydrated model
system packaged in TDT cans . . . . . . 65

7. Rate constants and half-lives for TAA and
RAA loss at 30°C as a function of water
activity in ascorbic acid and multivitamin
fortified dehydrated model food systems
packaged in paperboard boxes. . . . . . 68

8. Rate constants and half-lives for TAA and RAA
loss as a function of water activity and
storage temperatures in a dehydrated model
food system packaged in 303 cans . . . . 71

9. Activation energies for TAA and RAA loss in
ascorbic acid fortified dehydrated model
food system packaged in 303 cans . . . . 72

Table

10.

ll.

12.

13.

14.

15.

16.

Page

Rate constants and half-lives for TAA loss as
a function of mineral fortification and
water activity in a dehydrated model food
system stored at 30°C in 303 cans . . . . 74

Rate constants and half-lives for RAA loss as
a function of mineral fortification and
water activity in a dehydrated model food
system stored at 30°C in 303 cans . . . . 75

Activation parameters for TAA and RAA degra-
dation in ascorbic acid fortified model
system stored in TDT cans . . . . . . 77

Activation parameters for TAA and RAA degra-
dation in multivitamin fortified model
system stored in TDT cans. . . . . . . 78

Activation parameters for TAA and RAA degra-
dation in ascorbic acid fortified model
system stored in 303 cans. . . . . . . 79

Linear regression of rate constant vs. water
activity for TAA and RAA degradation in
model system stored in 303 and TDT con-
tainers . . . . . . . . . . . . 94

Activation parameters for reduced ascorbic

acid degradation in an intermediate
moisture model system . . . . . . . . 115

vi

Figure

1.

10.

LIST OF FIGURES

General moisture sorption isotherm for foods
(A) and effect of hystersis on moisture
sorption isotherm (B) . . . . . . .

Stability profile of dehydrated foods (from
Labuza, 1976). . . . . . . . . .

Molecular structure of biologically active
forms of ascorbic acid. . . . . . .

Fraction of ascorbic acid species in solution
as a function of pH. . . . . . . .

Plots loglo k/kO vs. /I for ionic reactions
of various types. The lines are drawn
with slopes equal to ZAZB (Laidler, 1965)
Adsorption isotherm for the dehydrated model
food system at 20°C (from Bach, 1974). . .

Block diagram for equilibration of dehydrated
model food system (inner and outer chamber
represent Aminco-Aire unit) . . . . . .

Fraction of TAA remaining vs. time for
selected aw at 30°C in the ascorbic
acid only fortified model system stored
in TDT cans . . . . . . . . . . .

Fraction of RAA remaining vs. time for
selected aw at 30°C in the ascorbic acid
only fortified model system stored in
TDT cans . . . . . . . . . . . .

Fraction of DAA remaining vs. time for 0.10 aw
at selected temperatures in the ascorbic
acid only fortified model system stored
in TDT cans . . . . . . . . .

vii

Page

21

27

28

34

46

47

56

57

60

Figure Page

11. Arrhenius plot for the activation energy of
TAA loss in the ascorbic acid only forti-
fied model system stored in TDT cans as
a function of aw . . . . . . . . . 61

12. Fraction of TAA remaining vs. time at
selected relative humidities at 30°C in
the ascorbic acid only fortified model
system packaged in paperboard boxes. . . 67

13. Fraction of TAA remaining vs. time as a
function of (a) water activity at 20°C
and (b) temperature at 0.40 aw for the
ascorbic acid only model system stored
in 303 cans . . . . . . . . . . 70

14. Moisture content equilibration during
storage at selected relative humidities
in dehydrated model systems packaged
unequilibrated in paperboard boxes
(from Purwadaria, 1977). . . . . . . 87

15. Relationship of the rate of TAA degradation
vs. water activity at constant storage
temperature in 303 (A) and TDT (B) cans
(O 10, o 20, A30, 0 37°C) . . . . 93

16. Ratio of (02)/ (AA) as a function of water
activity . . . . . . . . . . . 99

17. Relationship of the rate of ascorbic acid
degradation vs. water activity at con-
stant temperature for an intermediate
moisture model food system (data from
Lee and Labuza, 1975) . . . . . . . 101

18. Dependence of entropy of activation (A) and
enthalpy of activation (B) on water
activity for TAA loss in the dehydrated
model system stored in TDT cans ( £5)
and 303 cans ( O ) . . . . . . . . 109

19. Dependence of entropy of activation (A) and
enthalpy of activation (B) on water
activity for RAA loss in the dehydrated
model system stored in TDT cans ( A)
and 303 cans ( O ) . . . . . . . . 110

viii

Figure Page

20. Isokinetic relationship of TAA degradation in
dehydrated model system stored in TDT cans
(A)and303cans(‘). . . . . . . 111

21. Isokinetic relationship of RAA degradation in
dehydrated model system stored in TDT cans
(A)and303cans(‘). . . . . . . 112

22. Isokinetic relationship for ascorbic acid
degradation in an intermediate moisture
model system (data from Lee and Labuza,
1975) . . . . . . . . . . . . . 116

ix

INTRODUCTION

Fortification of food products is not a recent
idea. As noted by Bauernfeind (1971), Boussingault, in
1833, first proposed the fortification of table salt with
iodine in South America to prevent goiter. The initial
impetus for nutrient fortification in the United States
- came principally from an executive World War II order in
1941 requiring the addition of thiamin, riboflavin, niacin
and iron to white bread and flour to combat beriberi,
ariboflavinosis, pellagra and iron deficiency anemia.

From the initial addition of vitamins to flour, the
practice of nutrient fortification of most foods has
been commonplace and, in many instances, expected by the
consumer.

The principle consideration that should govern
any fortification program is public health. The reso-
lution of nutrient deficiencies in humans must involve a
multi-phase approach. Aylward and Morton (1971) outlined
six areas for the improvement of the nutritional status
of the general population: (1) improvement in the initial

quality of plant or animal product, (2) changes in dietary

habits and household practices, (3) improved methods of
processing, preservation and storage, (4) biological
enoblements, (5) physical treatment and (6) addition of
supplements. The changing food patterns and the increased
use of manufactured foods have led to the acceptance of
food enrichment policies in more than thirty countries.
Bauernfeind (1971) has presented an extensive compilation
of fortified foods utilized in various countries.

Extensive fortification of cereals, flours and
other low moisture products has been employed for a
number of years in the United States. Addition of
nutrients above the label claim is generally required in
order to compensate for incomplete distribution, nutrient
degradation during processing and storage and analytical
error (Borenstein, 1971).

Labuza (1968) and Borenstein (1971) have noted
that little kinetic data are available describing the
stability of vitamins during processing and storage of
foods. The influence of factors such as pH, moisture
content, oxygen, product composition, trace metals on
vitamin stability are unknown or difficult to extrapolate
from one product to another.

Ascorbic acid is frequently not found in fortified
products such as flour and ready-to-eat breakfast cereals
because of its instability. Ascorbic acid destruction is

known to be affected by pH, moisture content, oxygen and

trace metals, but a detailed knowledge of the physio-
chemical properties governing the loss and mechanism of
ascorbic acid deterioration during storage of dehydrated
foods is negligible.

The purpose of this research was to investigate
the decomposition of ascorbic acid in dehydrated foods as
a function of selected product and storage variables. A
model food system was chosen to simulate the composition
of a ready-to-eat breakfast cereal. The goals were to
obtain kinetic data describing the degradation of ascorbic
acid as influenced by water activity, temperature, moisture
content, oxygen availability and interaction with other
nutrients, including trace metals. It was anticipated
that thermodynamic activation parameters could be calcu-
lated and thus provide information regarding the degrada-

tive process of ascorbic acid.

LITERATURE REVIEW

Sorption Phenomena

 

The common goal of food preservation is the
extension of the product shelf-life to permit convenient
storage and distribution. The most important consider-
ation in food preservation is controlling the growth of
microorganisms in the product. One method of preservation
is reduction of the availability of water to the micro-
organisms by dehydration. However, other physical and
chemical reactions may still occur in dehydrated foods
which adversely affect food quality. An understanding
of the relationship between the state of water and these
processes will contribute to an improved quality of the
product following storage.

Water Activity and the
Sorption Isotherm

 

 

Water is prevalent throughout nature, as would be
anticipated, and is very important to the structure,
texture and chemical interactions of foods. The thermo—
dynamic state of water in a food system is given by the

chemical potential (Labuza, 1976)

u = no + RT ln aw (1)

where no = standard state chemical potential for water,
R = gas constant, T = absolute temperature and aw = thermo-
dynamic activity of water. The activity of water may be

defined as

a z; =—___. (2)

where: P0 = vapor pressure of pure water, P = vapor
pressure of water in the food product and % ERH = per
cent relative humidity at which the system neither gains
nor loses water.

The removal of water from a food system and its
relationship to the water activity are depicted in
Figure 1. During the initial stages of water removal,
the water activity of the product remains close to unity
(Figure 1A). On continued removal of water, the product
undergoes a relatively rapid decrease in water activity
per unit loss of moisture content (9 HZO/g solid). This
phenomenon is typical of food systems and the curve is
referred to as the moisture sorption isotherm.

Equation 2 would predict the ratio of the
moisture content to water activity to be independent of
the determination of the sorption isotherm. That is,
whether the equilibrium moisture is obtained via the
removal of water (desorption) or the addition of water

(adsorption). Close examination of the sorption isotherm

.Amv EHOLDOmw :oflumuom wusumAOE co
mfimumum>n mo vomumw pom A<V mpoow How Sumsuomfl :ofluQHOm musumAOE Hmnwcwo .H musmam

34 34
m6 ¢.O md N6 _.0 0.. m6 QC 10 N6

 

  

<
wasuoo am;sgow
wawoo aJmsgow

determined by these two methods reveals a phenomenon
referred to as hystersis (Figure 1B). This effect shows
that dehydrated systems which reach their equilibrium
moisture content by desorption of moisture contain a
greater moisture content at the same water activity than
systems which reach moisture equilibrium via the adsorp-
tion of moisture. Brunauer (1945) identified five general
types of adsorption isotherms, designated by Roman
numerals. The sorption isotherm shown in Figure l is
referred to as type II. The type II isotherm is common
in the case of physical adsorption and is believed to
correspond to multilayer formation. The type II isotherm
is typical of most dehydrated food products (Labuza, 1968).
In an effort to make the moisture sorption iso-
therm more manageable, Labuza (1968) divided it into three
regions to define the state of water in a dehydrated food
system (Figure 18). Region A corresponds to the adsorption
of a monomolecular layer of water which is tightly bound
and unavailable for reaction (Labuza, 1976). The water
molecules in this region are considered as bound to car-
boxyl and amino groups (Rockland, 1969) and reactions
which depend on solvation by water are not measured
(Fennema, 1976). Region B is designated the multilayer
region and corresponds to the adsorption of additional
layers of water molecules in which the water is more

loosely bound, probably via hydrogen bonding to hydroxyl

and amide groups. Region C is referred to as the cap—
illary condensation area and describes water which exists
in the capillaries of the dehydrated food. This water is
relatively free to act as a solvent. The reason it has

a value of less than 1.0 is that capillary forces and
soluble constituents cause a lowering of the vapor pres-
sure of this water in accord with Raoult's law. These
regions of the moisture sorption isotherm are general

and are not defined at precise water activities. Con-
siderable overlap between regions exists for various
products.

Theoretical Treatments of
the Sorption Isotherm

 

 

Various theories for the adsorption of nonpolar
gases on homogeneous and heterogeneous surfaces have
recently been reviewed by Adamson (1976). There are
four theoretical approaches which describe the sorption
isotherm to varying degrees of success: (1) kinetic,

(2) potential, (3) equation of state and (4) capillary
condensation. In addition, several empirical equations
have been proposed.

The kinetic approach proposed by Langmuir (1918)
for monolayer adsorption of gases was extended by Brunauer,
Emmett and Teller (1938) to apply to multilayer adsorption.

Their derivation led to the BET equation:

a (c-l)aw

w 1
-"_—— =——+———— (3)
(1 aw)m mOc moc

where aw = water activity, m = moisture content at aw
I

m0 = monomolecular moisture content and c = constant
related to the bonding energy. Three basic assumptions

were required for the derivation of the BET equation:
(1) Sorption occurs only on specific sites.

(2) Heat of sorption is constant for the first layer,
and equals the total heat of vaporization plus a

constant.

(3) Heat of sorption equals the heat of vaporization

for layers above the monolayer.

The major limitation of the BET equation is the
restricted water activity range (0.0 — 0.5) over which
it is applicable. The popularity of the BET equation is
derived from its ability to predict the monomolecular
moisture content and a hypothesis by Salwin (1959, 1962,
1963) that the monomolecular moisture content should cor-
respond to the water activity for maximum storage sta-
bility of a product.

The BET monomolecular moisture content may also

be used to calculate the water surface area, SO:

_N_

O 0 H20 - A (4)

10

where N = Avagadro's number, H20 = molecular weight of
water and A = area of a water molecule, 10.6 X 10-20m2
(Labuza, 1968). The water surface area value is generally
several orders of magnitude greater than that determined
by nitrogen surface areas (Berlin at al., 1966) or gas
permeability measurements (Fox et al., 1963). This dif-
ference in areas is rationalized by the ability of water
to plasticize long chain polymers exposing interior sites
for adsorption and that the water molecule is smaller

than nitrogen and thus is capable of entering smaller
pores.

The second approach to describing the sorption
isotherm is due to Polanyi (1928) and based on the force
field potential caused by the surface of the solid
material. The underlying assumption was that the total
work necessary to adsorb a gas molecule must equal the
work necessary to overcome the field strength as the
molecule approaches the surface of the solid and the
work of condensation. Frenkel (1946) extended the theory
by making the reasonable assumption that the interaction
between the solid and the adsorbed gas molecule is prin-
cipally of the dispersion type and the potential should
decrease with the inverse cube of the distance. This

led to an equation for the moisture content, m:

m = -a + 8 (1n aw) (5)

11

where a and B are constants. Although this method suc-
cessfully predicts an isotherm once an initial curve has
been determined, the primary disadvantage is its inability
to predict the monomolecular moisture content.

An alternative method to the formulation of the
force field potential for determining the sorption iso-
therm is that proposed by de Boer and Zwikker (1929).

The adsorption of gaseous molecules was explained by
assuming the polar adsorbent induces dipoles in the

first layer of adsorbed gas molecules which in turn
induces dipoles in the second layer, etc. This treatment
suggests the presence of strong orientational effects in
the multilayers; however, it suffers from the same short—
comings as the dispersion theory.

The third approach is that of Harkins and Jura
(1944) which utilizes a two-dimensional equation of state
for the film pressure. The Harkins-Jura equation takes

the form:

1n aw = B - A/v2 (6)

where A and B are constants and v = amount of adsorbed
gas.

The capillary condensation of the vapor, utiliz-
ing the Kelvin equation (Zsigmondy, 1911), is the only

approach that predicts the effect of hystersis. The

12

adsorption curve is dependent on the radius of curvature
of the capillary r, thus the capillary filling would be

described by

_ -21
aw - exp ( 1: cos BVO/RT) (7)

whereas on desorption, the retreat of the meniscus of

curvature was Zr and

exp (- % cos GVO/RT) (8)

D)
II

In equation (7) and (8), y = liquid surface tension,

9 = the contact angle of liquid in pores and V = molar
volume. Although an appealing hypothesis, Lubuza (1968)
has reviewed the problems associated with applying the

Kelvin equation to food systems:

(1) The equation is not applicable to situations

where the pores may be of molecular dimensions.

(2) The assumption of zero contact angle may lead

to errors .

(3) Surface tension of liquid in pores may vary

depending on the food matrix.

(4) Difficult to measure effects of nonideality

of vapor pressure lowering due to solutes.

Adamson (1976) and Labuza (1968) have noted that

the agreement of a theoretical equation with the

13

experimental data is necessary but insufficient proof of
its premises. The theoretical equation may well satisfy
the objectives of the user, or an empirical approach may
be equally suitable. Three empirical equations should
be mentioned.

Henderson (1952) described the sorption isotherm

empirically by:
l - a = e (9)

where c and n are constants and v = the amount of adsorbed
gas. Rockland (1957) has reported some success with food
systems for equation 9, but found other cases where a
composite isotherm of three straight lines described the
experimental data better. This led Rockland to the pro-
posal of "local isotherms," which correspond very closely
with the three regions of the sorption isotherm identified
by Labuza (1968).

More recently, Caurie (1970) prOposed that the
"natural" water sorption isotherm could be represented

mathematically by:
1n C = 1n Co - aw 1n r (10)

100 - %H20

%H20

r = magnitude of the gradient. This expression was

 

where C = at any aw, Co = C at zero aw and

further developed to yield

14

In C = -—————- - a 1n r (11)
' o

where Mo is the optimum safe moisture content for food
stability. Subsequently, Caurie suggested M to be
numerically equal to the water binding energy, Q, in

kcal and that MO is equivalent to the sorptive energy,

QC (Caurie, 1971a). The value obtained for M0 is some-
what greater than the BET monomolecular moisture content.
This has led Caurie (1971b) to postulate that the sorption
of water vapor in dehydrated foods is a penetration of

the sorbate throughout the sorbent rather than the surface

phenomenon assumed in the BET theory.

Physical State of Water

 

The description of the physical state of water in
dehydrated food systems is still incomplete. Much of the
problem arises from an inadequate understanding of pure
water and solute-water interactions. The physiochemical
factors of water important to the consideration of dehy-

drated foods have been summarized by Labuza (1976):

(1) Solvent properties--the ability of water to
dissolve molecules, to participate as a
reactant, act as a diffusion medium for other
molecules, and to concentrate molecules in the

aqueous phase.

15

(2) Solute effects-—the depression of water activity
by increasing solute concentration in accordance

with Raoult's law.

(3) Structural effects--the three-dimensional struc-
tural effects (size and shape of the capillaries)
in dehydrated food systems on lowering the water

activity.

(4) Solid surface interactions-~the strong binding
of water molecules to specific sites on the pro-

tein or other macromolecules of the food systems.

The state of water in dehydrated systems has been
studied by a variety of methods, such as differential
thermal analysis (DTA), differential scanning calorimetry
(DSC), nuclear magnetic resonance (NMR), electron spin
resonance (ESR) and phosphorescence. Results of each
technique suggest that water exists in at least two
states in dehydrated food systems. This has led to the
concept of bound water, which shows physical properties
different from free water (Kuprianoff, 1958).

One criterion for bound water has been to measure
the nonfreezable water which remains on cooling a sample
to sub-zero temperatures. The proton resonance signal
from flour samples adjusted to different moisture con-
tents was monitored as the temperature decreased by

Toledo et a1. (1968). A dramatic decrease in signal

16

intensity occurred between 30 and 20°F which they attri—
buted to the freezing of water. Below this temperature,
the signal for different samples was similar and the
intensity decreased slowly with continued cooling. This
was interpreted as reflecting the presence of unfrozen

or bound water. However, Toledo et a1. (1968) were
unable to determine the temperature at which the freezing
process was complete.

Differential thermal analysis revealed the temper-
ature (within narrow limits) at which a phase change from
ice to liquid occurs (Duckworth, 1972). Using samples of
carefully adjusted moisture contents, they determined the
nonfreezable water to correspond to the maximum moisture
content at which there is no evidence of a phase tran-
sition signal. Furthermore, the question of the complete-
ness of the freezing process could be easily determined
and in some cases erroneously high values for nonfreezable
water determined by other methods have been reported.

Of more practical importance is the binding of
water at ambient temperature since water binding may be
a temperature dependent phenomenon. Shanbhag et a1.

(1970) have reported an NMR method for determination

of "bound water capacity" (BWC). Careful adjustment and
determination of moisture content were critical, and as
later reported by Mousseri et a1. (1974), up to five days
for samples at the BWC moisture content were required to

achieve equilibrium.

17

Rockland (1969) has presented qualitative NMR,
ESR and phosphorescence results for water behavior in
gelatin. The NMR evidence showed a decreased band width
for water protons with increasing moisture content. ESR
signals from irradiation induced free radicals were found
to decay more quickly with increasing moisture content.
Phosphorescent decay times in the irradiated moisture-
adjusted gelatin samples were also found to decrease with
increasing moisture content, reflecting the stability of
the highly reactive triplet state. Rockland (1969) further
found that the physical behavior of the gelatin sample as
measured by these techniques correlated very well with
his proposed "local isotherm" concept (Rockland, 1957).

The ability of water in dehydrated foods to
solubilize chemical species and to allow diffusion of
these species was reported by Duckworth and Smith (1963).
They showed that migration of glucose, calcium and sulfate
occurred at moisture contents corresponding to monolayer
coverage. Thus, the solvent properties of water cannot
be neglected in dehydrated systems.

Thermodynamic Activation
Parameters

 

 

One main task of chemistry is to interpret and
predict reactivity. No general agreement prevails as to
the physiochemical quantity which should serve as a

measure of reactivity. Frequently, reactivity is

18

expressed by either the equilibrium constant, K, or the
rate constant, k, both of which are strongly temperature
dependent. An alternative approach is to separate the
reactivity into temperature dependent (enthalpic) and
temperature independent (entropic) parts. In kinetics,
classical Arrhenius theory or the absolute rate theory
yield an activation enthalpy (AHI) and an activation
entropy (ASI). The activation parameters are far less
sensitive to temperature, and within a narrow temperature
range may be taken as invariant (Schaleger and Long, 1963).
These parameters, entropy of activation, enthalpy
of activation and free energy of activation are calculated
from an equation based on the theory of absolute reaction

rates (Eyring, 1935):

= ka

eAS+/R e-AHI/RT
h

(12)

where k is the "best" value for the reaction rate constant

at a given temperature (T) and:

kb = 1.38 x 10.16 erg OK.l (Boltzman's constant)
h = 6.63 x 10.27 erg sec (Plank's constant)
R = 1.987 cal/OK-mole (gas constant)

AH+ = Ea -RT (for solutions).

The free energy of activation may then be calculated by

AG+ = AH+ - TAS+ (13)

l9

Enthalpy-entropy relationships are of two types,
(1) structural relationships for similar compounds under-
going the same reaction and (2) effects of moderate sol-
vent changes on a specific reaction. Frequently, for
both cases, AHI and AS+ tend to change in a compensating

manner and

AH+ = BAS+ + AH: (14)

where B is in OK and AH: is the intercept at AS+ = O.

This function is termed the compensation law or isokinetic
relationship (Leffler, 1955) and if obeyed results in

only small changes in AGI, the free energy of activation.

The significance of the isokinetic temperature
may be viewed as a constant, characteristic of the
reaction series and dependent on the experimental tem-
perature interval. Further, the value for the isokinetic
temperature may be either positive or negative and occur
either above or below the experimental temperature range.
Thus the significance is not the value of the isokinetic
temperature, but whether or not the isokinetic relation-
ship holds (Exner, 1973).

Blackadder and Hinchelwood (l958a,b) distinguished
three types of reaction series based on the activation
parameters: (1) series with constant entropy, reactivity
enthalpy controlled, interpretation based on electronic

effects; (2) series with constant enthalpy, reactivity

20

entropy controlled, interpretation based on steric and/or
solvent effects; (3) changes in AHI are paralleled by
changes in AS* such that the resulting effect on reactivity
is less than if controlled by one parameter, interpretation
based on steric and solvent effects. More recently, an
anti-compensation type series has been detected (Exner,
1949, 1972).

The existence of an enthalpy-entropy relationship
possesses important mechanistic implications. The iso-
kinetic relationship is expected to be followed only by a
reaction series in which structural or solvent changes do
not result in a change in mechanism. Large changes in
enthalpy or entropy of activation have often been con-
sidered indicative of changes in mechanism, but a more
reliable criterion of a change in mechanism is the con-

stancy of the free energy of activation (Leffler, 1955).

Reactions in Dehydrated Foods

 

Labuza (1968) summarized the general behavior of
the major classes of reactions in dehydrated foods as a
function of water activity as shown by the stability map
(Figure 2). This provides a guide to the types of
reactions that are dominant at a particular water
activity and some discussion as to the role of water is

warranted.

 

 

21

.LNBlNOO BHOISIOW

‘5;
O
a.

 

BLVH NOILOVEB BAIIV'IBH

 

0.2

0

Stability profile of dehydrated foods (from Labuza, 1976).

Figure 2.

22

 

Microbiological

. The initial concern in dehydration is the elimi-
nation of microbial growth. Scott (1957), in a compre-
hensive review, suggested that water activity, rather
than moisture content, determined the availability of
water for microbial growth. He reported water activity
limits for bacteria, molds and xerophilic fungi to be
0.91, 0.80 and 0.65, respectively. Furthermore, environ—
mental factors such as pH, oxygen, pressure, temperature
and nutritional adequacy may necessitate a higher minimum
water activity for microbial growth.

Recent research has identified additional effects

of water activity on microorganisms (Karel, 1973: Labuza,

1972):

(l) Microorganisms are most sensitive to heat, light,
and chemicals at high water activities. However,

an insensitivity develops upon moisture removal.

(2) Toxin production often occurs at water activities
lower than the minimum required for microbial

growth.

(3) Microbial growth is sensitive to the hystersis
effect. The minimum water activity for micro-
bial growth may be a higher water activity for

adsorption than desorption.

23

In a review of xerophilic fungi, Pitt (1976) has
noted the importance of the predominant solute present
in the medium at low aw, and the different responses of

the organism with different solutes.

Enzymatic

 

Enzymatic reactions that can occur in low moisture
foods which have not been heat treated are well known.
Acker (1962, 1969a,b) has concluded that enzymatic
reactions usually occur when fluid water is available,
that is, above the monomolecular moisture content. The
liquid water is required for solubilization and diffusion
or transport of the substrate to the enzyme. Acker and
Weise (1972a,b) have noted lipid splitting enzymatic
activity at water activities below the BET monomolecular
moisture content (as low as 10% RH) if the triglycerides
are in a fluid state and can be mobilized. Duden (1971)
reported hydrolytic activity in freeze-dried vegetables
at relative humidities of 1% concluding that the trans-
port medium for the substrate was the vapor phase.

In a recent study of the thermal inactivation of
wheat ribonuclease at varying low moisture levels, Multon
and Guilbot (1976) reported a two-stage process. The
first stage, which was more rapid than the second, was
interpreted as a water catalyzed step in which a reorgani-
zation of bond energies between water and polar sites

occurred. The second stage reflected the thermal effect.

24

Chemical

Nonenzymatic browning and lipid oxidation are two
classes of chemical reactions in dehydrated foods which
illustrate two distinctly different roles for water.

Comprehensive reviews of nonenzymatic browning
and a general discussion of carbonyl-amine reactions can
be found in the literature (Reynolds, 1962, 1965, 1969
and Feeney et al., 1976). The rate of browning has been
found to reach a maximum at a water activity between 0.3
to 0.7 depending on the food matrix, the most common
range being 0.5 - 0.7. Two explanations have been pro-
posed for this maximum in browning rate. Loncin et a1.
(1968) have suggested that the water produced in the
reaction increases the moisture content resulting in
product inhibition. However, Labuza et a1. (1970) believe
the maximum browning rate occurs at the point of greatest
reactant concentration and that the decreased browning
rate on increasing water activity is simply due to a
dilution of the reactants.

The kinetics and mechanisms for lipid oxidation
have been reviewed by Labuza (1971) and Schultz et a1.
(1962). Early studies showed improved stability toward
oxidation with increasing moisture content (Stevens et
al., 1948; Matz et al., 1955; Martin, 1958), however,
the interaction of water in lipid oxidation was unclear.

Investigations by Maloney et a1. (1966), Labuza et a1.

25

(1966) and Karel et a1. (1967) have suggested several

functions of water in retarding lipid oxidation:

(1) The water molecules may hydrogen bond to hydro-
peroxides at the lipid-water interface, thus,
slowing initiation. This antioxidant effect
increases until the interface becomes saturated

with water.

(2) Water may hydrate with trace metals reducing

their catalytic activity.

(3) Water may react to form insoluble metal hydroxides,
thus removing the metal catalysts from the reaction

phase.

(4) The lifetime of free radicals decreases as the

water content is increased.

The reaction profile for lipid oxidation is found
to increase as the water activity approaches that of the
intermediate moisture range. Labuza et a1. (1970) have
attributed this to the greater solubility and mobility
of the trace metals, thus overcoming the antioxidant

properties of water.

Ascorbic Acid

 

Zilva and co-workers (1923a,b; 1924a,b; 1925;
1927) isolated and reported the general properties of a

compound with antiscorbutic activity in the mid-19203.

26

The substance was found to be labile and exhibit strong
oxidizing power. Freshly oxidized solutions of this
antiscorbutic compound also possessed physiological
activity. Tillmans et a1. (l932a,b,c) resolved the
anomaly when they concluded that both the oxidized and
reduced forms possessed antiscorbutic activity, and that
the forms were reversible. Szent-Gyorgi (1928) isolated
an optically active, acidic reducing sugar with the
empirical formula C6H806. The subsequent demonstration
that this isolated sugar with the antiscorbutic factor of

Zilva was L-ascorbic acid was confirmed with its synthesis.

Structure and PrOperties

 

L-Ascorbic acid (Figure 3) is a white crystalline
solid melting at 192°C. It is an optically active ([a]o =
+23° in water) lactone ring system and shows selective
absorption of ultraviolet light at 245 nm to 265 nm as
the pH is increased. The hydroxyl groups on carbons 2 and
3 are enolizable due to the conjugated system. L-Ascorbic
acid possesses two acidic hydrogens with pKa's in water
of 4.25 and 11.79 (Figure 4). The solubility of L-ascorbic
acid is 32g/100 ml water (Bauernfeind and Pinkert, 1972)-

Dehydroascorbic acid (Figure 3), formed by the
removal of two equivalents of hydrogen from L-ascorbic
acid, retains the basic ring structure although it no
longer contains a conjugated system. The acidity associ-

ated with L-ascorbic acid is lost on oxidation to

27

.pwom oflnuoomw mo mEu0m m>wuom waamowmoHofln mo wusuosuum umasomaoz .m muzmflm

pwom oflnuoommoupzswplq Odom oflnuoomMIA

N

o
_
oflo\ No-0: old/\o 0-0-0:
/ \ : \/z

o 10 o:

 

 

 

28

 

.mm mo GOHDOCDM 6 mm coflusaom :fl mmflommm pflom canuoomm mo cofluomum

In
N. o. m o c N

 

.v musmflh

SEIOBdS :IO NOILOVHJ

 

29

dehydroascorbic acid. The planar structure of L-ascorbic
acid was confirmed by x-ray crystallographic analysis

(Cox et al., 1932).

Reactivity in Solution

 

Ascorbic acid is stable in the dry crystalline
state. In solution, it reacts readily giving a variety
of products depending upon the reaction conditions.
Oxygen, pH, metal ion catalysis and temperature are
reported as the principle factors determining the decom-
position of ascorbic acid. Although extensively studied
by numerous investigators, uncertainty still exists as
to the mechanism of degradation. A discussion of the
importance of these factors is most easily presented in
terms of the presence or absence of oxygen.

The anaerobic decomposition of ascorbic acid in
aqueous solutions has been reported to be pH dependent.
Finholt et a1. (1963) found that the maximum rate of
anaerobic degradation of ascorbic acid occurred at pH 4.0
at 96°C. Huelin et a1. (1971) reported an increase in
destruction rate as a function of pH at 50 and 100°C,
reaching a maximum at pH 2.3, decreasing to a minimum
at pH 4 and increasing again to pH 6, the highest pH of
the study. The discrepancy between these two studies is
difficult to rationalize; however, Finholt et a1. (1963)
did find a catalytic effect due to the buffer type and

concentration used to maintain the pH. No primary salt

30

effect on the anaerobic degradation rate was found. Fin-
holt et a1. (1963) suggested that the rate maximum at

pH 4 was due to the complexation of ascorbic acid and

the monoionic species of ascorbic acid as a reactive
intermediate.

Various reaction products from ascorbic acid
degradation have been reported in the literature. Huelin
et a1. (1971) measured carbon dioxide and furfural for-
mation during the anaerobic degradation of ascorbic acid.
One mole of CO2 was formed per mole of ascorbic acid over
-the pH range 2.2—6.0, but the yield of furfural decreased
with increasing pH. Investigations of Coggiola (1963)
identified, 2,5-dihydro-2-furoic acid as an end product
of the anaerobic degradation of ascorbic acid and Kurata
and Sakurai (1967) have reported 3-deoxy-L-pentosone as
a probable intermediate breakdown product.

The anaerobic catalysis of ascorbic acid by
cupric ion at pH 2.1 - 3.5 has been reported by Shtamm
et al.(1974). The reaction was second-order with respect
to c0pper (II) concentration, first-order in ascorbic
acid and inversely proportional to hydrogen-ion concen-
tration. Additionally, Huelin (1953) and Huelin et a1.
(1971) have reported the catalytic properties of fructose,
sucrose, fructose-l-phosphate and fructose-1,6-diphosphate
in the anaerobic decomposition of ascorbic acid. The

fructofuranose form of fructose was found to be the

31

catalytic agent. Sucrose catalysis was attributed to its
partial hydrolysis in the reaction medium.

The aerobic oxidation of ascorbic acid has
received intensive investigation. The reaction rate is
known to depend on oxygen, pH, metal catalyst, buffer and
temperature. The rate of the spontaneous oxidation of
ascorbic acid increases with pH, showing maxima at pH 5.0
and 11.5 (Csuros and Petro, 1955). This pH dependence
has been confirmed by Miller and Joslyn (1949) and Khan
and Martell (1967) for the pH range 2 - 6. Weissberger
et a1. (1943) found no rate maxima, although the same pH
dependence of the reaction from pH 4.7 - 9.2 was observed.
The monoionic and diionic species of ascorbic acid have
been interpreted as being more reactive in the presence
of oxygen than the neutral species. The reported order
of ascorbic acid oxidation with respect to hydrogen-ion
varies, -0.5, -l.0 and -2.0, Miller and Joslyn (1949),
Khan and Martell (1967) and Dekker and Dickinson (1940),
respectively.

The oxygen dependence of ascorbic acid oxidation
has been reported to be first-order above 0.2 atm oxygen
(Khan and Martell, 1967). Recently, Shtamm and Skurlator
(1974a) have reported a half-order dependence on oxygen,
which has been supported by the results of Jameson and

Blackburn (1975; 1976a,b). Jameson and Blackburn also

32

suggest that the oxygen dependence data of Khan and Mar-
tell (1967) more adequately fits half-order kinetics.

Reports of the catalytic properties of transition
metals, notably copper and iron, on the oxidation of
ascorbic acid abound in the literature. In many cases
the rates are found to increase linearly with electrolyte
concentration. Existing theories do not account for the
rate enhancement exhibited by ion-molecule reactions upon
the addition of neutral electrolytes.

It is generally assumed that all rate changes
resulting from the presence of neutral electrolytes may

be described by the Bronsted equation (Bronsted, 1922):

Y Y
k = k __.__A B (15)
O Y+

where YA' and y* are the activity coefficients for the

YB
reactants and the transition state, respectively. Once
the applicability of this equation was accepted, the
major problem in developing a theory of salt effects was
determining the relationship of the activity coefficients
of the reactants to the added electrolyte concentration.

The Debye-Huckel theory provided the first relationship

of this type in the form of the Debye-Huckel limiting law,
_ 2
Log Y — -QZ /I-, (16)

where Q is a constant for a given solvent, Z is the

valency of the ion and I is the ionic strength. The

33

limiting law is valid only below concentrations of 0.01M.
Following the procedure of Scatchard (1922), equation (16)

may be substituted into the rate equation (15) giving

Log k = Log k0 + 2QZAZB /I— . (17)

For aqueous solutions at 25°C, the value of Q is approxi-
mately 0.51 (Laidler, 1965). This equation has been

shown to describe salt effects for a large number of ion-
ion reactions (Figure 5), even though the applicability

of this equation is exceeded. According to equation (17)
a plot of log k vs. /I-wi11 give a straight line of

slope 1.022AAB. For the case where one reactant is a
neutral molecule, the product ZAZB in equation (17) is
zero and no salt effect would be predicted. The fact that
salt effects do exist for ion-molecule reactions has often
been explained by assuming that the solutions were not
dilute enough to allow application of the Debye-Huckel
limiting law. To surmount this shortcoming several
relationships using higher powers of I and purely emperi-
cal parameters have been developed.

Weissberger and LuValle (1944) reported that only
the monoionic ascorbic acid moiety was susceptible to
Cu(II) catalysis. More recently, studies by Khan and
Martell (1967) have shown the oxidation of ascorbic acid

to be linearly dependent on the concentration of Cu(II)

and Fe(III) ions. Ogata et a1. (1968) found that the

34

 

 

A
O
0.6 ’ 'o
. O
O B
0.4 ' f>°
o o: C
o 0 .
O , O .. . o
S 0.2 . ...o .
Q
g; " . . ., . , , D
.9 '. -: .
o . o o
’0.2 ’ O 0 0 O
.o o E
-O.4 P F
O
O _ ‘ .
0.1 0.2 0.3
/I

Figure 5. Plots loglo k/ko vs. /I for ionic reactions
of various types. The lines are drawn with
slopes equal to Z Z (Laidler, 1965).

A B
2+ 2+ _

A Co(NH3)5Br + Hg zAzB - 4
B s o 2' + I' z z = 2
2 8 A B
c Co(OC2H5)N:N02 + OH zAzB = 1
D (c (UREA) )3+ + H o z z - o

r 6 2 A B ‘
CH3COOC2H5 + OH zAzB = o
+ _
E H + Br + H202 ZAZB - -1
F Co(NH ) Br2+ + 0H" 2 z — -2
3 5 A B ’

2+ 3-

35

catalytic ability of Cu(II) ions was dependent on the
anion of the salt in the order CuC12>Cu(NO3)~CuSO4.
Shtamm and co-workers (1974a,c) and Jameson and Black-

burn (1976a,b) have also reported the catalytic proper-
ties of Cu(II) ions in ascorbic acid degradation. Pek-
karinan (1974) has reported evidence supporting Fe(III)
catalysis of ascorbic acid. Other catalytic agents which
have been reported include metal chelates, vanadyl and
uranyl ions, and iron (III) chelates of aminopolycarboxlyic
acids (Khan and Martell, 1968a,b, 1969). The mechanistic
interpretation of metal catalyzed ascorbic acid degra-
dation is varied, ranging from an ascorbate-metal-oxygen
complex involving a one electron transfer to oxygen (Khan
and Martell, 1967) to the formation of a metal-metal

dinuclear-ascorbate-oxygen complex with a two electron

transfer to oxygen (Jameson and Blackburn, 1975).

Stability in Food Products

 

The availability of kinetic data pertaining to
the destruction of ascorbic acid in food products is
much more limited than for destruction in model solutions.
Singh et a1. (1976) studied the disappearance of ascorbic
acid in a liquid infant formula as a function of initial
oxygen concentration and light intensity at 7.2°C. The
rate of ascorbic acid loss was found to be dependent on
the presence of oxygen and obeyed overall second-order

kinetics. If the oxygen level was maintained, pseudo

36

first-order loss of ascorbic acid was observed. The rate
of ascorbic acid loss also increased linearly with light
intensity up to 1600 lux. A companion study by Mack

et a1. (1976) showed a first-order relationship for the
uptake of oxygen in the same infant formula.

Stability of ascorbic acid in canned tomato juice
was studied by Lee et a1. (1977) as a function of pH,
Cu(II) and temperature. First-order reaction kinetics
were observed in the anaerobic system. The reported
activation energy was 3.3 kcal/mole, which was less than
that found by Pope (1972) for a similar tomato juice
system. Lee et al. (1977) reported that the degradation
rate of ascorbic acid reached a maximum at pH 4.06, and
was linearly dependent on the concentration of copper
at each pH.

Karel and Nickerson (1964) reported the destruc-
tion rates of ascorbic acid in dehydrated orange juice
crystals increased with increasing aw, but no dependence
on the storage atmosphere (air or vacuum) was observed.
The authors concluded that there was no value of aw below
which destruction of ascorbic acid ceased.

Vojnovich and Pfeifer (1970) studied the sta-
bility of reduced ascorbic acid in a corn-soy-milk (CSM)
blend at three moisture contents. Lee and Labuza (1975)
also studied the rate of ascorbic acid destruction in an

intermediate moisture (0.32-0.84 aw) glycerol, corn oil

37

and microcrystalline cellulose model system. Both of
these studies showed increased ascorbic acid degradation
rates with increasing aw. Lee and Labuza (1975) also
examined the effects of moisture sorption hystersis on
ascorbic acid stability and found greater degradation

of ascorbic acid on the desorption leg of the isotherm.
However, activation energies calculated for ascorbic
acid degradation were the same (~20 kcal/mole) for both
legs of the isotherm.

Recently, Waletzko and Labuza (1976) studied
ascorbic acid stability in an intermediate moisture food
system (aw=0.85) in an accelerated shelf-life test. They
reported that ascorbic acid degraded rapidly by first-
order kinetics whether packed in an air or Nz/H2 atmos-
phere. Ascorbic acid was more stable, however, in the
product packaged in the NZ/HZ atmosphere.

The retention of ascorbic acid in tomato juice
crystals as a function of aw was reported by Riemer (1977).
First-order kinetics were obeyed for the loss of reduced
ascorbic acid and dehydroascorbic acid. Activation
energies ranging from 16.2 - 24.6 kcal/mole were calcu-
lated. As with the study of Karel and Nickerson (1964)
no effect of oxygen was observed. Ascorbic acid stability
in products with high acidity may reflect a pH stabilizing

effect.

38

The stability of ascorbic acid in foods is gen—
erally improved at storage temperatures below 0°C. How-
ever, Grant and Alburn (1965) reported the rate of ascorbic
acid oxidation to be greater at —11°C than at 1°C in 0.02M
acetate buffer of pH 5.0 and 5.5. Freeze concentration
or the increase in concentration of the solutes in the
frozen state over the unfrozen state has been presented
as the most probable rationale for this phenomenon (Pin-
cock and Kiovsy, 1966). The detailed study by Thompson
and Fennema (1971) on the effect of freezing on ascorbic
acid stability has shown the discrepancy to be a function
of the solute concentration which governs the solubility
of oxygen. Thus, the higher the solute concentration,
which would be the normal pattern on freezing of foods,
the less dissolved oxygen and the slower the oxidation
rate of ascorbic acid at subfreezing temperatures.

Predictions of Product
Storage Stability

 

 

The prediction of the storage stability of food
products has received increased attention recently in
view of nutritional labelling, cost and time requirements
for conventional stability studies. Numerous mathematical
techniques have been proposed for the prediction of pro-
duct quality and deterioration rates of specific reaction

types (e.g., nonenzymatic browning, lipid oxidation).

39

The most commonly used model is that which depends
on a detailed understanding of the kinetics of the reac-
tion being modeled. This requirement necessitates the
experimental determination of those parameters affecting
the reaction kinetics.

Kwolek and Bookwalter (1971) developed a mathe—
matical model for storage stability based on time-temper-
ature data. The Arrhenius equation was satisfactory in
predicting flavor and peroxide values.

Mathematical models proposed for predicting the
stability of space foods in semi-permeable packages,
where lipid oxidation was the only deteriorative mechanism
of concern, were sufficiently accurate to aid in package
design (Karel and Labuza, 1969; Simon et al., 1971).

Mizrahi et a1. (l970a,b) and Karel et al. (1971)
developed a computer model for the prediction of the
storage life of dehydrated cabbage as a function of non-
enzymatic browning. The feasibility of using accelerated
storage tests for determining the necessary kinetic param-
eters have been evaluated. Conditions of high storage
temperature and high moisture content provided experi-
mental data in a short time that could be used to predict
the degree of browning at ambient storage conditions.
Correlation analysis between accelerated shelf life test
data and long-term storage studies indicated that accel-
erated storage tests could be utilized in the development

of prediction equations.

40

The storage stability of potato chips that undergo
loss of quality resulting from moisture adsorption and
oxidative rancidity was studied by Quast et a1. (1972)
and Quast and Karel (1973). A mixed model based on
reaction kinetics and empirical data fitting was used to
describe the deterioration process. The agreement between
the actual storage tests and the computer prediction was
good, although the procedure was less accurate for accel—
erated test studies.

Wanniger (1972) proposed a mathematical model for
prediction of ascorbic acid storage stability using the
equation ln k = -Ea/RT + a 1n H20 + b, where a and b are
constants and the other symbols possess their normal
meaning. Utilizing the data of Vojnovich and Pfeifer
(1970), he reported excellent correlation between the
experimental and predicted results.

A basic computer model for simulation of nutrient
stability in semi-permeable packages was postulated by
Heldman (1974). The loss of the nutrient was considered
to be a function of oxygen and water activity; storage
temperature was not incorporated into the model.

Lee et a1. (1976) developed a computer simulation
for determining ascorbic acid stability in canned tomato
juice. Copper‘ion concentration, pH and storage temper—
ature were the parameters incorporated into the model.

The experimental and predicted results were in excellent

41

agreement (i3%). The simulation also contained the addi-
tional capacity to calculate the effect of seasonal
variations as a result of storage temperature fluctuations
using the Fourier transform series.

Recently, Mizrahi and Karel (1977) proposed a
"no model" accelerated test method for the determination
of the deterioration rate of dehydrated products due to
moisture sensitive reactions. This technique is based on
theoretical considerations that the deterioration rate is
inversely proportional to the rate of moisture gain at
any given moisture content. This procedure was success-
fully applied to the prediction of ascorbic acid loss in
tomato power and browning in dehydrated cabbage.

From this overview of methods for predicting
product stability, it is obvious that presently used
models generally rely on one index of quality, e.g.
browning, lipid oxidation, nutrient loss, organoleptic
properties. To date, modeling proposals have lacked
general acceptance and utility. The reason for this
appears to be their inability to consider simultaneous

deteriorative reactions.

EXPERIMENTAL PROCEDURES

Model Food System Composition

 

The composition of the dehydrated model food sys-
tem is given in Table 1. The model system was designed

to simulate a ready-to-eat breakfast cereal.

Model Food System Preparation

 

The ingredients, except coconut oil, were mixed
dry in a ribbon blender, water was then added to give a
slurry of approximately 40% total solids. The slurry was
heated to 60°C, the coconut oil was then added and the
system was homogenized in a Manton-Gaulin homogenizer at
2000 psig (lst stage, 1500 psig: 2nd stage, 500 psig).
The pH of the homogenized slurry was 6.8.

The model system for all studies was fortified
with USP reduced ascorbic acid at a level of 25% NAS/NRC
RDA (11.25 mg ascorbic acid) per 1009 model system (dry
weight basis). The ascorbic acid was thoroughly mixed
into the homogenized model system slurry after it had
cooled to ambient temperature.

Following fortification, the model system slurry

was layered onto stainless steel freeze-drying trays,

42

43

placed in a Virtis Model FFD 42 WS Freeze-Dryer and
frozen at a platen temperature of -40°C. The model sys-
tem was then dried to 5p absolute pressure at a platen

temperature of 110°C.

Table 1. Composition of model food system

 

 

Component %d
Proteina 10.2
Fat 1.0
Carbohydrateb 76.6
Reducing SugarC 5.1
Sucrose 5.1
Salt 2.0

 

aSoya protein--Promine D. Central Soya.

bFood Grade Powdered Starch--A.E. Staley, Inc.,
and Corn Syrup Solids 15 D.E., American Maize.

cSupplied by the corn syrup solids, % dry weight
basis.

dCalculated on dry weight basis.

Addition of Other Nutrients to
the Model System

 

 

Fortification of the dehydrated model food system
with other nutrients for the interaction studies was
accomplished in a manner similar to the addition of
ascorbic acid.

In the case of the multivitamin system, vitamin A,

as retinyl acetate, was added to the model system slurry

44

prior to homogenization using the coconut oil as the
carrier. Riboflavin and ascorbic acid were then added

as aqueous solutions to the homogenized model system
slurry after it had cooled to ambient temperature. The
model system was then freeze—dried as described above.
Both vitamin A (retinyl acetate, Sigma) and riboflavin
(Sigma) were added at a level of 25% NAS/NRC RDA (0.287mg
and 0.450mg, respectively) per 100g model system (dry
weight basis).

For the mineral fortification study, the indi-
vidual mineral was added at the levels indicated in
Table 2 to the homogenized model system slurry. All
salts were of reagent grade quality. The model system

was then freeze-dried as described above.

Model Food System Equilibration

Water activities for the model system were
adjusted using equilibrium moisture content isotherm
data (Figure 6) determined by Bach (1974) for the freeze-
dried model system at 10, 20, 30 and 37°C using the
method of Palnitkar and Heldman (1971). All experiments
in this study were performed on the adsorption leg of
the sorption hystersis loop. Equilibration was accom—
plished by placing thin slabs of the freeze-dried model
system into an equilibration chamber and forcing con-
ditioned air of the desired aw and temperature from an

Aminco-Aire unit through the closed system (Figure 7).

45

Table 2. Mineral supplement, source and %RDA employed
in mineral fortification study.

 

 

Mineral mg per
Supplement source %RDA 1009 d.b.
Control -- —- --
FeSO4 ' 7H20 Mallinckrodt 10 9.0
25 22.4
FeCl2 ' 4H20 Mallinckrodt 10 6 4
25 16 0
Fe Fisher 10 1.8
25 4.5
ZnCl2 Mallinckrodt 10 3.1
25 7.8
ZnSO4 ' 7H20 Mallinckrodt 10 6.6
25 16.5
ZnO Fisher 10 1.9
25 4.7
CaCO3 Mallinckrodt 10 300.3
25 750.5
CuCl ° 2H 0 Baker 2.7
2 2

6.7

CuSO ° 5H 0 Mallinckrodt

 

24

BO
(3

MOISTURE_CONTENT
an ID

 

46

 

 

Figure 6.

 

0.8 LO

Adsorption isotherm for the dehydrated model
food system at 20°C (from Bach, 1974).

.Auflcs wufldnoocfl8¢ usmmwumwu Hwnfimno Hmupo paw umccflv Emumhm
pOOM Hmpofi pmumuphzmp mo coflumunflaflsqm MOM Emummflc xOOHm

 

 

 

.h muomflm

 

 

 

 

 

 

 

 

 

 

 

 

mmmZfimU WV mmm2<m0
mmBDO mMZZH
7
4
>
mmmfimo 1‘
rlll 20351qu3on I mmHmHnHzommo

 

 

 

 

 

 

 

 

48

A dehumidifier and cooling coil were placed in the closed
system when relative humidities were required which were
lower than could be provided by the Aminco-Aire unit
alone. Samples were equilibrated in approximately

twenty-four hours.

Model System Packaging and Storage

 

Following equilibration of the model system to
the desired equilibrium moisture content, the product was
immediately packaged in either enameled metal containers
(TDT-208x006 or 303- 303x406) or 1-oz. paperboard boxes.
The metal containers prevented the transfer of air and
moisture vapor into or out of the package. All containers
were filled with the same mass of model system, approxi-
mately 15g. This amount of the model system filled the
TDT cans leaving no headspace in the container, whereas
the 303 cans permitted a large headspace in the container.

Paperboard l-oz. cereal boxes (3cm x 7cm x 10.3 cm)
containing waxed liners (thickness 0.009cm) were packaged
with unequilibrated freeze-dried model system. Moisture
transfer coefficient for the liner and paperboard box

5 gHZO—cm/mz-h-mmHg) as

plus liner were equal (7.25 x 10-

reported by Purwadaria (1976).
The packaged model system was then stored in

constant temperature cubicles of the appropriate temper-

atures. The cubicle temperatures were estimated to be

il°C of the set temperatures of 10, 20, 30 and 37°C.

49

Relative humidities in the cubicles were estimated to
vary no more than 12% of the 10, 40 and 85% RH storage

conditions.

Moisture Content Measurement

 

Moisture content (dry weight basis) of equili—
brated model system was determined by drying the samples
in a vacuum oven at a vacuum of 28 inches Hg at the same
temperature at which the product was equilibrated and
stored. An ethanol-dry ice cold trap was inserted in
the line between the vacuum oven and the vacuum pump to
aid in the transfer of moisture from the product. Dry
air was admitted into the vacuum oven at a rate of
15-20 ml/min to aid in the displacement of water vapor
from the drying chamber. All samples were dried until
they reached a constant weight. Using this method, the
water activity of each sample could be determined to
ensure that the aw had not changed during storage. The
validity of this method for water activity determination
of a large number of samples was confirmed by the vapor

pressure manometric method of Sood and Heldman (1974).

Ascorbic Acid Determination

 

Ascorbic acid was measured by the continuous flow
o-phenylenediamine micro-fluorometric procedure described
by Kirk and Ting (1975). Ascorbic acid was extracted

from the sample matrix with a 3% m-phosphoric acid--3%

50

glacial acetic acid solution. Samples were filtered and
the extract introduced to a Technicon autoanalyzer.
Dehydroascorbic acid was permitted to condense with
o-phenylenediamine forming a fluorophor which was
detected fluorometrically with an excitation wavelength
of 360nm and emmission wavelength of 436nm. Total ascor-
bic acid was measured as dehydroascorbic acid following
oxidation of reduced ascorbic acid with 2,6-dichloro-
indOphenol. Boric acid was utilized as the blanking
reagent. Reduced ascorbic acid was determined from the

difference of total minus dehyroascorbic acid.

Data Analysis

 

The loss of ascorbic acid was analyzed by the

first-order kinetic equation for all storage conditions

studied:
_ d (C) _
—_dt_— - kC (18)
where (C) = molar concentration of ascorbic acid; t = time

(days); and k = first-order rate constant (days-l).
Ascorbic acid levels for zero time storage were deter-
mined after each aliquot of the dehydrated model system
was equilibrated to the desired water activity. Ascorbic
acid determinations were performed at preset intervals
for a minimum of two half-lives, except for samples
stored at 10°C in TDT cans which exhibited an extremely

long half-life.

51

The temperature dependence for ascorbic acid
degradation was analyzed according to the Arrhenius

equation:
k = A exp (-Ea/RT) (19)

where k = first-order rate constant; A = Arrhenius pre-
exponential; Ea = activation energy (cal/mole); R = gas
constant (1.987 cal/OK-mole) and T = absolute temperature
(OK). The reaction rate constants and the activation
energies were calculated by linear regression analysis
and by a computer program, KINFIT.

The KINFIT program differs from the usual least
squares techniques in that numerical integration pro-
cedures are used to provide a fit to the desired dif-
ferential equation (Dye and Nicely, 1971; Singh et al.,
1975). 'This method of calculation assists in accounting
for errors in vitamin assays and small variations in
storage times. This program is specifically written for
chemical reactions.

Thermodynamic activation parameters for the
destruction of ascorbic acid were calculated based on

the theory of absolute reaction rates. The appropriate

equations are outlined in the literature review.

Experimental Design

 

Various factors such as temperature, pH, oxygen,

metal ion concentration and water aCtivity have been

52

reported as influencing the degradation rate of ascorbic
acid in solution or dehydrated systems. As noted by
Labuza (1968) and Borenstein (1971), a dearth of kinetic
data exists on the storage stability of ascorbic acid

and other nutrients in food systems. It was, therefore,
the purpose of this study to examine some of these factors
and obtain satisfactory kinetic data for ascorbic acid
degradation in a dehydrated model food system.

Ascorbic Acid Stability
in TDT Cans

 

 

The model food system was prepared, fortified with
ascorbic acid only and equilibrated as previously described.
Equilibration conditions were 0.10, 0.24, 0.40, 0.50 and
0.65 aw at 10, 20, 30 and 37°C. Following equilibration,
the model system was then sealed in TDT cans and stored at
the respective equilibration temperatures. This provided
a total of twenty conditions with which to examine the
effects of water activity and temperature on the storage
stability of ascorbic acid. The 0.10 aw represents a
condition below the BET monomolecular moisture content,
0.24 aw is the experimentally determined BET monomolecular
moisture content, 0.40 and 0.50 aw are conditions in the

monolayer region and 0.65 a approaches the capillary

W

region of the adsorption isotherm.

53

Ascorbic Acid--Vitamin Inter—
actions in TDT Cans

 

 

The model system was prepared, fortified with
ascorbic acid and vitamins A and B2 and equilibrated as
previously described. Equilibration conditions were 0.10,
0.24, 0.40, 0.50 and 0.65 aw at 10, 20, 30 and 37°C.
Following equilibration, the model system was then sealed
in TDT cans and stored at the respective equilibration
temperatures. This provided twenty conditions with which
to examine the effects of vitamin interactions, water
activity and temperature on the storage stability of
ascorbic acid.

Ascorbic Acid Stability
in Paperboard Boxes

 

 

The model food system was prepared, fortified
with ascorbic acid only and with ascorbic acid, vitamins A
and B2 and packaged unequilibrated in 1—02. paperboard
boxes with waxed liners. The packages were stored in
cubicles at 10, 40 and 85% RH at 30°C. This permitted an
examination of the effects of moisture vapor and air
transmission on the storage characteristics of ascorbic
acid.

Ascorbic Acid Stability
in 303 Cans

 

 

The model food system was prepared, fortified
with ascorbic acid only and equilibrated as previously

described. Equilibration conditions were 0.10, 0.40 and

54

0.65 aw at 10, 20, 30 and 37°C. After equilibration,

approximately 15g of the model system was then sealed

in 303 cans and stored at the respective equilibration
temperatures. This provided twelve conditions for the
examination of the influence of a large gaseous oxygen
reservoir on the storage stability of ascorbic acid.

Ascorbic Acid--Mineral Inter-
action in 303 Cans

 

 

The model system was prepared, fortified with
the minerals in Table 2 and equilibrated as previously
described. Equilibration conditions were 0.10, 0.40 and
0.65 aw at 30°C. After equilibration, approximately 15g
of the model system was then sealed in 303 cans and
stored at 30°C. Zero, 10 and 25% RDA fortification
levels were employed to evaluate the effects of mineral
concentration at selected aw on the storage character-

istics of ascorbic acid. In addition, the effect of the

anion of the mineral employed could also be evaluated.

RESULTS

Ascorbic Acid Stability in
TDT Cans

 

The effects of water activity, moisture content
and storage temperature on the stability of ascorbic
acid were determined by studying the destruction of
reduced (RAA), dehydro (DAA) and total (TAA) ascorbic
acid in a dehydrated model food system. The model system
was equilibrated to 0.10, 0.24, 0.40, 0.50 and 0.65 aw at
10, 20, 30 and 37°C, packaged in TDT cans, sealed and
then stored isothermally at their respective equilibra-
tion temperatures. The dependence of TAA and RAA
destruction on aw at 30°C are shown in the first-order
kinetic plots of Figures 8 and 9. At all aw and storage
temperatures, the loss of TAA and RAA could be satisfac-
torily described by first-order kinetics. Correlation
coefficients associated with these plots were > 0.95.

The first-order rate constants and half-lives
for TAA and RAA degradation are presented in Table 3.
The experimentally determined TAA and RAA concentrations
at zero time storage were in good agreement with calcu-

lated ascorbic acid levels at t = 0 computed by the

55

 

56

 

A
1.0 w
.40
.8 014
o . ‘LJ‘O
.6 - . 0.50
. . . 055
.4»
C)
.2- o
' A
.' 1 4 1 n
20 40 60 80

Figure 8.

DAYS

Fraction of TAA remaining vs. time for
selected at 30°C in the ascorbic acid

only fortified model system stored in TDT
cans.

57

0J0
0.24
AAO

0.50
0.65

 

 

 

 

.IL I J J

1
20 40 60 80
DAYS

Figure 9. Fraction of RAA remaining vs. time for
selected aw at 30°C in the ascorbic acid
only fortified model system stored in TDT
cans.

58

 

 

 

 

 

Table 3. Rate constants and half-lives for TAA and
RAA loss as a function of water activity
and storage temperature in ascorbic acid
fortified dehydrated model food system
packaged in thermal death time cans.
Temp TAA RAA
°C 3w b b
a c a c
k 0 t% k 0 t%
10 0.10 0.31 0.03 224 0.43 0.07 161
0.24 0.37 0.04 187 0.45 0.05 154
0.40 0.42 0.05 165 0.47 0.09 147
0.50 0.49 0.03 141 0.58 0.05 119
0.65 0.50 0.02 139 0.55 0.05 126
20 0.10 0.45 0.06 154 0.65 0.01 107
0.24 0.95 0.12 73 1.34 0.21 52
0.40 1.28 0.11 54 1.69 0.22 41
0.50 1.12 0.08 62 1.30 0.13 53
0.65 1.44 0.28 48 1.93 0.54 36
30 0.10 0.91 0.08 76 1.11 0.15 63
0.24 1.78 0.23 39 2.30 0.31 30
0.40 3.13 0.26 22 3.84 0.42 18
0.50 3.99 0.34 17 4.63 0.50 15
0.65 4.77 0.42 15 5.29 0.51 13
37 0.10 0.98 0.11 71 1.23 0.20 56
0.24 5.01 0.36 14 4.44 0.16 16
0.40 7.03 0.24 10 7.87 0.28 9
0.50 9.24 0.59 8 8.90 0.57 8
0.65 15.74 0.66 4 16.85 1.11 4
aFirst Order Rate Constant, x 10".2 days_1 (KINFIT

analysis).

b

cHalf-Life, days.

Standard Deviation, x 10-

2

59

KINFIT program. The standard deviation of the rate con-
stants calculated by the KINFIT program was, in general,
less than 10% of the respective rate constants.

Data in Table 3 show that the stability of RAA
and TAA in the model system stored in TDT cans decreased
with increasing aw and storage temperature. Rate con-
stants for RAA destruction are slightly greater than
those describing TAA losses at the same conditions.
Little significance is placed on this rate difference
because of the standard deviations associated with the
rate constants describing RAA losses are in part a
function of the RAA determination (TAA-DAA=RAA).

The stability of DAA showed a strikingly dif-
ferent pattern than that observed for RAA and TAA. Dur-
ing the initial portion of the storage study, the con-
centration of DAA increased at 10°C, 0.10 and 0.24 aw
and 20°C, 0.10 aw followed by a decrease after approxi-
mately twenty-five days of storage (Figure 10). No
increase in DAA levels were observed at the other storage
conditions and DAA loss at the other storage conditions
appeared to fit a first—order equation. The rate con-
stants for the loss of DAA, however, were not calculated
for reasons which will be presented in the discussion.

The effect of temperature on the rate constants
describing the loss of TAA as a function of aw at 10, 20,

30 and 37°C is shown in Figure 11. A similar temperature

6O

.mcwo Boa cfl pmuoum Ewumwm Hmpoe pwflmauuow waco pflom cannoomm on» :w moms»

umummswu mmmuoum pwuomawm um

03
.

 

o... at A A. ... 1B

983%
can.

 

“Vb

3m oa.o How mafia .m> mcflcwmfimu <<o mo cofiuomum .OH mupmwm

mXSQ

 

 

 

20.0

I 0.0
8 0

6.0

4.0

3.2

Figure 11.

 

 

61

0.65
0.50

O
A
D
A
O

I I I n 1

3.3 3.4 _ 3.5 3.6
1xm3
7.

Arrhenius plot for the activation energy of TAA

loss in the ascorbic acid only fortified model
system stored in TDT cans as a function of aw.

62

dependence of the rate constant for RAA loss was also
found. Calculated activation energies (Ea) for TAA and
RAA destruction as a function of aw are reported in
Table 4. The activation energies for aw equal to or
above the monomolecular moisture content (aw = 0.24)
averaged l8.0:1.6 kcal/mole for TAA loss and 17.412.2
kcal/mole for RAA loss. The activation energy for both
TAA and RAA destruction at aw of 0.10 were confirmed by
the t-test to be significantly different from the acti—
vation energies for the other storage conditions at the
95% confidence level.

Table 4. Activation energies for TAA and RAA loss

in ascorbic acid fortified dehydrated
model food systems packaged in TDT cans.

 

Activation Energy

 

 

aw (kcal/mole)
TAA RAA
0.10 8.1 7.1
0.24 15.9 14.2
0.40 17.6 17.8
0.50 19.2 18.1
0.65 19.2 19.3

 

Ascorbic Acid--Vitamin Interactions
in TDT Cans

 

 

The model food system for this study was forti-
fied with ascorbic acid, vitamins A and B2, equilibrated

at 0.10, 0.24, 0.40, 0.50 and 0.65 aw at 10, 20, 30 and

63

37°C, sealed in TDT cans and stored isothermally at their
respective equilibration temperatures. The possible
effects of these added nutrients on the stability of
ascorbic acid were evaluated as a function of aw and
storage temperature.

The loss of RAA and TAA in the multivitamin model
system could be adequately described by first-order
kinetics. The degradation constants and half-lives are
presented in Table 5. Correlation coefficients were > 0.95
and the standard deviations of the rate constants as calcu-
lated by the KINFIT computer program were less than 10%.
The degradation rate of TAA and RAA in the multivitamin
fortified model system increased with increasing aw,
similar to that found for the model system containing
only ascorbic acid. There was no significant difference
in rate constants for TAA and RAA loss. Model system
containing vitamins A, B2 and C, which was equilibrated
at aw of 0.10 to 0.65 and stored at 10°C in TDT cans
showed apparent increases in RAA and TAA stability under
these conditions (Table 5) when compared to the model
system containing only ascorbic acid and stored under
similar conditions (Table 3).

The pattern of small initial accumulation of DAA
at low aw and storage temperatures found in the ascorbic

acid model system (Figure 10) was also observed for the

64

 

 

 

 

 

Table 5. Rate constants and half-lives for TAA and
RAA loss as a function of water activity
and storage temperature in multivitamin
dehydrated model food system packaged in
thermal death time cans.
Temp a TAA RAA
0C w ka 0b tC ka 0b tc
8 15
10 0.10 0.18 0.05 385 0.14 0.04 495
0.24 0.31 0.03 224 0.33 0.12 210
0.40 0.43 0.03 161 0.41 0.11 169
0.50 0.46 0.06 151 0.52 0.22 133
0.65 0.45 0.04 154 0.54 0.07 128
20 0.10 0.49 0.10 141 0.84 0.14 83
0.24 1.03 0.13 67 1.75 0.34 40
0.40 1.46 0.21 47 2.46 0.78 28
0.50 1.18 0.11 59 2.43 0.37 28
0.65 1.31 0.24 53 2.88 0.22 24
30 0.10 1.03 0.69 67 1.70 0.38 40
0.24 3.63 0.28 19 3.43 0.24 20
0.40 3.46 0.30 20 3.81 0.52 18
0.50 3.28 0.30 21 4.40 0.54 16
0.65 5.96 0.77 12 7.38 0.85 9
37 0.10 0.98 0.19 71 1.21 0.09 62
0.24 3.14 0.19 22 -- -- --
0.40 5.20 0.55 13 6.01 0.73 12
0.50 5.08 0.49 14 6.64 1.52 10
0.65 8.46 0.78 8 10.64 0.95 7
aFirst Order Rate Constant, x 10-2 days-1 (KINFIT
analysis).

CHalf-Life, days.

bStandard Deviation, x 10-

2

65

multivitamin fortified model system. At higher temper-
atures, the concentration of DAA decreased steadily with
time as noted previously.

Activation energies calculated from the kinetics
data describing the destruction of TAA and RAA in the
multivitamin model stored in TDT cans are presented in
Table 6. The values are similar to those calculated for
the ascorbic acid fortified system with the exception of
the 0.10 a condition (Table 4). Activation energies for

W

aw equal to or above the monomolecular moisture content

averaged 16.5i2.8 kcal/mole for TAA loss and 16.0:3.l
kcal/mole for RAA loss.
Table 6. Activation energies for TAA and RAA loss

in multivitamin fortified dehydrated
model system packaged in TDT cans.

 

Activation Energy

 

 

a (kcal/mole)
w

TAA RAA
0.10 11.5 13.1
0.24 13.5 11.8
0.40 16.8 16.6
0.50 15.9 16.1
0.65 20.0 19.3

 

Ascorbic Acid Stability in
Paperboard Boxes

 

 

Two model systems were used in this study, one

fortified with ascorbic acid only and the other with

66

vitamins A, B2 and ascorbic acid. Both model systems
were packaged, unequilibrated, in paperboard boxes con—
taining waxed liners and stored in cubicles at 10, 40 and
85% relative humidity at 30°C. This study was designed
to evaluate the influence of air and moisture vapor trans-
mission on the storage stability of ascorbic acid.
Destruction of TAA in the ascorbic acid only
model system stored in boxes at 10, 40 and 85% RH are
described by first-order kinetics (Figure 12). Similar
conformity to first-order kinetics were found for the
loss of RAA in the ascorbic acid only fortified model
system and the loss of both TAA and RAA in the multi-
vitamin fortified model systems. It is recognized that
the moisture content of the boxed model systems changed
during storage and the effect of this change on the
destruction rate of ascorbic acid is discussed later.
The calculated first-order rate constants and half—lives
for the degradation of TAA and RAA in the ascorbic acid
and multivitamin fortified dehydrated model systems
packaged in paperboard boxes are presented in Table 7.
Higher relative humidities were associated with greater
destruction rates of TAA and RAA. Little difference is
noted in the destruction rates for TAA versus RAA in the
ascorbic acid only system. There is a slightly greater

rate of loss of RAA in the multivitamin fortified system.

67

 

LO
6
.2
a
o
< .l a
\
<
.06
%RH a
010
A40
.02 085
l2 24 36

DAYS

Figure 12. Fraction of TAA remaining vs. time at
selected relative humidities at 30°C
in the ascorbic acid only fortified
model system packaged in paperboard
boxes.

68

 

 

 

 

 

Table 7. Rate constants and half-lives for TAA and
RAA loss at 30°C as a function of water
activity in ascorbic acid and multivitamin
fortified dehydrated model food systems
packaged in paperboard boxes.
Model TAA RAA
System % RH b b
a c a c
k 0 t15 k 0 t8
Ascorbic 10 2.45 0.19 28 2.66 0.21 26
Acid
Only 40 3.63 0.16 19 3.77 0.25 18
85 7.01 0.45 10 7.06 0.46 10
Multi- 10 2.01 0.34 34 2.93 0.24 24
vitamin
40 3.00 0.21 23 3.48 0.33 0 20
85 10.88 0.19 6 13.44 1.91 5
aFirst Order Rate Constant, x 10-2 days-1 (KINFIT

analysis).

b

cHalf-Life, days.

Standard Deviation, x 10-

2

69

Ascorbic Acid Stability in
303 Cans

 

The degradation of total, reduced and dehydro
ascorbic acid in dehydrated model food system packaged
in 303 cans was studied as a function of water activity
and storage temperature. The model system was equili-
brated to 0.10, 0.40 and 0.65 aW at 10, 20, 30 and 37°C,
sealed in 303 cans and stored isothermally at their
respective equilibration temperatures. The experimental
results for the loss of TAA stored in 303 cans at 0.10,
0.40 and 0.65 aw at 20°C and at 10, 20, 30 and 37°C at
0.40 aw are presented in Figure 13. These data conformed
to the first—order kinetic function as did the destruction
data at all aw and storage temperatures. Figure 13 shows
the dependence of TAA stability on aw as well as temper-
ature. A similar first-order dependence was found for
the experimental data for RAA destruction in the model
system stored in 303 cans.

The first-order rate constants and half—lives for
TAA and RAA degradation in the model system stored in 303
cans are presented in Table 8. These data show an increase
in the rate of TAA and RAA loss with increasing water
activity at a constant temperature. No significant dif-
ference is apparent in the rate constants for TAA and RAA
destruction at the same storage conditions of temperature

and water activity.

70

.mcmo mom cw pmuoum
Emumwm H0605 Mano pwom UHnHoomm on» now So ov.o um mupumhwaEwu Any can Doom
um mufi>wuom kum3 “my mo :ofluocpw 6 mm mfifiu .m> mcflcflmfimu dds «0 cowuomum .MH musmflm

 

 

 

 

 

 

m><o
00 0? ON 00 Ct ON
W 1 '1 J a q
bno
0nd
ONn N.
o. o . 36.. 1
00 . O¢.Ou
O..Oo
. .1 3< .c.
.l
0
fig
0 I u A”.
n . . u 4 N
n 4 . o n. 4
n n . Ho.
. .4
m . . . 7.
I, < O.-

71

 

 

 

 

 

Table 8. Rate constants and half—lives for TAA and
RAA loss as a function of water activity
and storage temperatures in a dehydrated
model food system packaged in 303 cans.
TAA RAA
Temp a
0C w ka 0b tc ka 0b tC
1: *5
10 0.10 0.33 0.03 210 0.34 0.51 204
0.40 0.60 0.04 116 0.63 0.61 110
0.65 0.69 0.09 100 0.81 1.21 86
20 0.10 0.86 0.10 81 0.84 1.27 83
0.24 1.10 0.07 63 1.03 0.90 67
0.40 1.46 0.08 47 1.47 0.97 47
0.65 2.03 0.15 34 2.04 2.00 34
30 0.10 1.36 0.76 51 1.29 1.09 54
0.40 3.66 0.25 19 3.12 2.04 22
0.65 5.55 0.29 12 5.11 4.98 14
37 0.10 1.77 2.33 39 1.86 2.18 37
0.40 7.25 0.74 10 7.59 9.10 9
0.65 12.07 4.31 6 11.67 26.20 6
a . -2 -l
F1rst—order rate constant, x 10 days .

bStandard deviation, x 10-3.

cHalf—life, days.

72

The temperature dependence of TAA and RAA loss
could be described by the Arrhenius equation. The acti-
vation energies for TAA and RAA degradation in the ascor-
bic acid fortified model system stored in 303 cans is
presented in Table 9. The activation energies for 0.40
and 0.65 aw are within the range of experimental error
of those found for TAA and RAA degradation in the TDT
can studies. However, at 0.10 aw the activation energies
were 10.7 kcal/mole for TAA and RAA degradation in the
model system stored in 303 metal cans versus 7-8 kcal/mole
for the TDT containers.

Table 9. Activation energies for TAA and RAA
loss in ascorbic acid fortified

dehydrated model food system
packaged in 303 cans.

 

Activation Energy

 

 

a (kcal/mole)
w

TAA RAA
0.10 10.7 10.7
0.40 16.0 15.6
0.65 18.3 17.0

 

Ascorbic Acid--Mineral Interaction
in 303 Cans

 

Selected trace minerals were added individually
to the model system prior to freeze-drying to evaluate

their effect on the storage stability of ascorbic acid.

73

The minerals were added at two different concentration
levels (Table 2). Three forms of iron (FeSO4, FeClZ, Fe)
and zinc (ZnSO4, ZnClZ, ZnO) as well as two forms of
copper (CuSO4, CuClz) and CaCO3 were selected for the
study. The freeze—dried model systems were then equil-
ibrated to 0.10, 0.40 and 0.65 aw at 30°C and sealed in
303 cans and stored isothermally at 30°C.

TAA and RAA losses could be described by first—
order kinetics at all storage conditions. Correlation
coefficients were > 0.93 and the standard deviation of
the rate constants were approximately 10% of the rate
constants. The experimentally determined rate constants
and half-lives are presented in Tables 10 and 11 for TAA
and RAA degradation, respectively. No difference in TAA
and RAA destruction rates was observed, and a similar
dependence on water activity was found as for the other
studies.

No rate enhancement in ascorbic acid destruction
over the nonfortified system due to any of the minerals
studied was observed for the 0.10 and 0.40 aw storage
conditions. The only exception to this is the slight
catalysis in the CuCl2 and CuSO4 fortified systems at
0.40 aw.

At 0.65 aw, which is in the capillary region of
the adsorption isotherm, a 2-3 fold increase in the

degradation rate of TAA and RAA over the nonfortified

74

.msmn .mAAHumHm:

 

 

 

 

 

 

n
.H|m>mp~ ca x .ucmumcoo muwu Hmpuolumufimm
m oo.a AH hm.m ow mm.H v
ca oo.m ma m~.m am hH.H Omsu
o hm.HH ma om.v om mm.H N
ma Hm.m ma mm.m Hm mm.H HOSO
v ¢m.>a um m>.m he wv.H mm m
h hm.oa mm hm.~ we me.a ca oomu
ma Hm.m mm om.m me mw.H mm
mm mH.m Hm m~.~ mm vm.H oa ocw
v mm.ha om om.~ on mh.H mm v
v ow.mH em Hm.~ mv qv.a OH omcu
m mo.m~ mm he.~ me mo.H mm N
e mm.ha om mm.m om ov.a CH HUCN
v m~.>a Hm m~.~ mv mm.H mm
m mw.m om mm.~ mm Hm.H OH on
A H¢.m em va.~ we mm.H mm N
b mv.oa hm hm.~ mm mm.H OH Howm
o mm.HH mm HH.~ mm Hm.H mm v
w ~N.HH Hm o~.~ mm Hm.H oa omwm
«a mo.m om om.m om Nm.H s: 0:02
x x x
ucmewamasm
<0”;
36 mw.o m ov.o m oa.o Hmuwcfiz
.mcmo mom cw Doom um pmuoum
Ewumwm poom H0005 pwumupxnmp a Ca >uw>fiuom uwumz 0cm noduwowmwuuow
Hmumcwfi mo cofiuocsm 6 mm mmoH dds new mm>wanmamn paw mucmumcoo wumm .OH manna

75

.msmo .mmfifiumammn

. imamleoH x .ucmumcoo mum“ nopuolumuHmm

 

 

 

 

 

H

m mH.m NH Na.m mm om.H v
mH Nh.v pH oo.v om ov.H Omso
m Ho.NH NH mm.m ow m>.H N
NH em.m MN Ho.m em mN.H Hugo
v Nn.>H NN HH.m mm Hm.H mN m
n Hv.oH HN om.m mv Ho.H 0H ovmo
SH om.m mN mv.N mv mm.H mN
HN hm.m mN Nm.N oe eh.H CH OCN
v mm.mH mN No.m mm >5.H mN v
w MH.mH vN mm.N hm hm.H OH omCN
m mH.mN mN hv.N mm MH.N mN N
v oo.>H 5N mm.N mm mm.H CH Hucu
v Hm.>H Hm ON.N Nv mm.H mN
b mm.m wN vw.N Nm Nm.H OH on
b hm.a wN aw.N ov Nn.H mN N
b no.OH mN hm.N He on.H 0H Humm
o mn.HH wN vv.N Nm wH.N mN v
m m¢.HH mN bv.N ov vh.H OH omwh
VH Ho.m 0N mv.m mm vo.N In 0:02
x x x
an ax nu mx 2» mx

<omw ucwEmHmmsm

3 3 3 HmumcHz

m mo.o m ov.o m OH.o

 

.mcmo mom CH Ooom um pmuoum Emumam poem HwCOE
pwumupxcop m cH qu>Huom Hmum3 can coHumoHMHuHOM HwnmcHE
mo coHuocsu 0 mm mmoH (dz HON mm>HHImHms can mucmumcoo mumm .HH mHnt

76

system is noted for each trace mineral added, regardless
of form. The only exception is zinc oxide which did not

exhibit catalysis.

Thermodynamic Activation Parameters

 

The thermodynamic activation parameters associated
with the degradation reaction for ascorbic acid were
calculated from the first-order rate constants for TAA
and RAA destruction. The calculated entropy of activation
(AS+), enthalpy of activation (AH+) and free energy of
activation (AGI) for TAA and RAA degradation for the
three temperature dependent studies are presented in
Tables 12, 13 and 14. The free energy of activation for
TAA and RAA loss in each study remains constant at the
various water activities, whereas an increase in both
the entropy and enthalpy of activation is noted with

increasing water activity.

77

Table 12. Activation parameters for TAA and RAA
degradation in ascorbic acid forti-
fied model system stored in TDT cans.

 

 

 

 

 

Ea AH+ AS+ AG+
aw ______
(kcal/mole) (kcal/mole) (e.u.) (kcal/mole)
TAA
0.10 8.1 7.5 -43 20.2
0.24 15.9 15.3 -15 19.8
0.40 17.6 17.0 - 8 19.5
0.50 19.2 18.6 - 3 19.4
0.65 19.2 18.6 - 2 19.2
RAA
0.10 7.1 6.5 -46 20.1
0.24 14.2 13.6 -20 19.7
0.40 17.8 17.2 - 7 19.4
0.50 18.1 17.5 - 6 19.3
0.65 19.3 18.7 - 2 19.2

 

78

Table 13. Activation parameters for TAA and RAA
degradation in multivitamin fortified
model system stored in TDT cans.

 

 

 

 

 

Ea AH+ AS+ AG+
aw ______
(kcal/mole) (kcal/mole) (e.u.) (kcal/mole)
TAA
0.10 11.5 10.9 -31 20.2
0.24 13.5 12.9 -22 19.4
0.40 16.8 17.2 - 7 19.4
0.50 15.9 15.3 -14 19.5
0.65 20.0 19.4 1 19.1
RAA
0.10 13.1 12.5 -25 19.9
0.24 11.8 11.2 -28 19.4
0.40 16.6 16.0 -11 19.4
0.50 16.1 15.5 -13 19.3
0.65 19.3 18.7 - 1 19.0

 

79

Table 14. Activation parameters for TAA and RAA
degradation in ascorbic acid forti-
fied model system stored in 303 cans.

 

 

 

 

 

Ea AHI AS+ AG+
(kcal/mole) (kcal/mole) (e.u.) (kcal/mole)
TAA
0.10 10.7 10.1 -33 20.0
0.40 16.0 15.4 -13 19.4
0.65 18.3 17.7 - 5 19.2
RAA
0.10 10.7 10.1 -33 20.0
0.40 15.6 15.0 -15 19.5
0.65 17.0 16.4 - 9 19.2

 

DISCUSSION

The storage stability of ascorbic acid in low
moisture dehydrated food products was studied utilizing
a model food system designed to simulate a ready-to-eat
breakfast cereal. The model system provided a product
of known composition with which to conduct the study.
Model system preparation, as previously described, was
relatively simple and ensured a homogeneous distribution
of vitamins and minerals.

Ascorbic Acid Stability in
TDT Cans

 

The degradation of total (TAA) and reduced (RAA)
ascorbic acid in the low moisture dehydrated model food
system are adequately described by first-order kinetics
with correlation coefficients 2 0.95 (Table 3). This
first-order dependence was observed at all storage con—
ditions of water activity and temperature. Data in
Table 3 show that the storage stability of TAA and RAA
in the model system stored in TDT cans decreased as the
water activity increased from 0.10 to 0.65 at each

storage temperature. These data contradict the hypothesis

80

81

of Salwin (1959, 1962) that the BET monomolecular moisture
content should represent the equilibrium moisture content
for maximum storage stability of the product. The BET
monomolecular moisture content has been determined to be
equal to an aw of 0.24 in the dehydrated model system
(Bach, 1974), yet ascorbic acid degradation was signifi—
cantly reduced at 0.10 aw. Ascorbic acid degradation at
a measurable rate at 0.10 aw also conflicts with the view
of Fennema (1976), that reactions which depend on solva-
tion would not be measurable at water activities below
the monomolecular moisture content. The dependence of
RAA loss on water activity has been observed by other
investigators (Karel and Nickerson, 1964; Vojnovich and
Pfeifer, 1970; Lee and Labuza, 1975) but their studies
did not include aw below the BET monomolecular moisture
content. The dependence of TAA loss on water activity
and its conformity to first-order kinetics has not been
reported.

The rate constants for RAA loss are, in general,
slightly greater than the rate constants for TAA loss at
corresponding storage conditions. It is important to
note that little significance can be attached to the
difference in rate constants for TAA and RAA losses.

Bauernfeind and Pinkert (1970) have summarized
the possible degradation pathways for reduced ascorbic

acid. A major route for RAA degradation at neutral pH

82

is through dehydro (DAA) ascorbic acid. Following the
scheme shown in equation (20) it is evident that if DAA

is more stable than

k k

RAA ef=ié DAA ——39 products (20)
-1

RAA (k2 < kl)’ one would expect the rate constant describ—
ing RAA loss to be significantly greater than that for
TAA loss. In fact it would be fortuitous if TAA loss
could be described by first-order kinetics. In this
study, the rate of loss of TAA and RAA were essentially
equal at corresponding conditions, suggesting that either
RAA does not degrad via DAA, or that the degradation
rate for DAA is greater than for RAA (k2 > k1). This
latter view of ascorbic acid degradation is supported
by the investigations of Khan and Martell (1967) in
which they utilized the 2, 4-dinitrophenylhydrazone
method of Roe (1943) to measure the formation of dehy-
droascorbic acid with time.

Examination of the data in Figure 10 shows that
the concentration of DAA during the initial stages of
the storage study increased followed by a subsequent
decrease in DAA concentration at 10°C, 0.10 and 0.24 aw
and 20°C, 0.10 aw. At all other conditions a constant

decrease in DAA levels with time was noted. This is

interpreted as further evidence that RAA degrades

83

principally via DAA (equation 20). Kinetic treatment

of the data describing the destruction of DAA with time
was not attempted because it had not been independently
confirmed that RAA degraded solely via DAA in the dehy-
drated model system. The data for DAA loss does suggest,
however, that the rate of DAA degradation was faster than
the rate of RAA degradation.

The temperature dependence of TAA and RAA degra-
dation in dehydrated model system stored in TDT cans was
described by the Arrhenius equation (Table 4). The
activation energies calculated for aw equal to and above
the monomolecular moisture content (aw = 0.24) were in
the range of 15-19 kcal/mole, and showed a slight depen-
dence on the water activity. These values were not sig-
nificantly different from the activation energies reported
by Lee and Labuza (1975) for the destruction of RAA in an
intermediate moisture model food system at aw 0.32-0.84
on both the adsorption and desorption legs of the sorption
isotherm. The activation energy for TAA and RAA destruc-
tion in the model system stored at 0.10 aw in TDT cans
was 7-8 kcal/mole. This value was determined by the
t-test to be significantly different from the activation
energies at the other water activities. No stability
studies of ascorbic acid in systems at water activities
below the BET monomolecular moisture content have been

reported. Lee et a1. (1977), however, reported the

84

activation energy for the anaerobic destruction of
ascorbic acid in canned tomato juice (pH = 4.06) to be
3.3 kcal/mole. The observed change in activation energy
for ascorbic acid destruction in the model system at

the 0.10 aw storage condition could be interpreted as

a change in degradative mechanism. Based on these
limited data such a conclusion may be tenuous.

Ascorbic Acid--Vitamin Interaction
in TDT Cans

 

 

Kinetics data were obtained for TAA and RAA loss
in the same model system which was fortified with
vitamins A and B2 and then stored in TDT cans. The
calculated first-order rate constants and half-lives
for TAA and RAA loss are presented in Table 5. Cor-
relation coefficients for the determination of the rate
constants were 2 0.95. In general, the stability of TAA
and RAA in this multivitamin study showed a similar
dependence on water activity as was found for the system
containing only ascorbic acid. The degradation rate
for TAA and RAA loss in the multivitamin fortified system
increased with aw. As was found in the preceding study,
the BET monomolecular moisture content did not prove to
be the aw offering the greatest stability to ascorbic
acid.

There was no difference in degradation rates

between TAA and RAA loss in the multivitamin study. DAA

85

levels were measurable and followed a pattern similar

to that observed in the model system containing only
ascorbic acid. The measurement of DAA concentration
with time during the multivitamin study presents further
supporting evidence that the principal degradative path-
way for RAA in the model system was via DAA. As in the
preceding study, the loss of DAA in the multivitamin
fortified system could be inferred to be faster than the
loss of RAA.

Comparison of TAA and RAA degradation constants
between the multivitamin model system and the model sys-
tem containing only ascorbic acid showed no difference,
except at 10°C and low aw where an apparent increase in
ascorbic acid stability was observed in the multivitamin
system. This difference may be due to the relatively
long storage period which did not permit assays beyond
one-half life at these conditions. From these data, it
is concluded that neither vitamin A nor B2 significantly
interacted with ascorbic acid to either catalyze or
inhibit the degradation of ascorbic acid. The excep-
tional stability of riboflavin in dehydrated products
reported by Borenstein (1971) and Dennison et a1. (1977)
support this conclusion.

The temperature dependence of TAA and RAA degra-
dation in the multivitamin fortified model system stored

in TDT cans was described by the Arrhenius equation

86

(Table 6). The activation energies calculated for water
activities equal to and above the BET monomolecular
moisture content were 15-19 kcal/mole. These Ea values
were not significantly different from the activation
energies found for TAA and RAA degradation in the model
system containing only ascorbic acid and stored in TDT
cans. The higher Ea value at 0.10 aw for the multivitamin
system over the model system with ascorbic acid only
probably reflects the experimental imprecision associated
with the long storage period.

Ascorbic Acid Stability in
Paperboard Boxes

 

 

The experimental data for the degradation of
ascorbic acid in model systems, which were fortified
with ascorbic acid only and ascorbic acid, riboflavin
and vitamin A, and stored in paperboard boxes, was
treated by first-order kinetics (Table 7). The standard
deviations associated with the rate constants were rela-
tively large and the correlation coefficients were approx-
imately 0.90. The model system was packaged unequili-
brated and therefore the moisture content and water
activity changed during storage. Experimental data in
Figure 14 reported by Purwadaria (1976) using the same
model system and boxes demonstrated the time dependence
required for the unequilibrated model system to reach

moisture content equilibrium with the storage atmosphere

24

5;
I

MOISTURE CONTENT
E
t

 

Figure 14.

87

 

A
{A ‘afdrf”
A
A
A
A
% RH
. ‘0 10
u 40
A 85
. ’ ° ' o ' '
4 8 l2 l6 2O
WEEKS

Moisture content equilibration during storage
at selected relative humidities in dehydrated
model systems packaged unequilibrated in
paperboard boxes (from Purwadaria, 1977).

88

at 30°C. The rate of ascorbic acid degradation in the
previous studies was found to be dependent on aw, there-
fore, the rate of TAA and RAA loss in the model system
packaged in paperboard boxes would be expected to increase
with time until the moisture in the model system achieves
equilibrium with the storage atmosphere.

Treatment of the experimental data for TAA and
RAA loss in the model system stored in paperboard boxes
at 10% RH by first—order kinetics is believed to result
in little error in the calculated destruction rates
because of the short storage period required for moisture
equilibration at this condition (Figure 14). Determi-
nation of the rate constants for ascorbic acid destruc—
tion in the boxed model system stored at 40 and 85% RH
present a more complex situation. For simplicity, the
experimental data for TAA and RAA destruction in the
paperboard boxes at 40 and 85% RH were analyzed by the
first-order rate function in order to compare these data
with similar rate data obtained for the model system
stored in TDT cans at constant water activity.

The experimental first-order rate constants and
half-lives for TAA and RAA loss in both model systems
stored at 10, 40 and 85% RH at 30°C in paperboard boxes
are presented in Table 7. The stability of ascorbic
acid was dependent on the relative humidity of the

storage atmosphere and there was little difference in

89

degradation constants for TAA and RAA loss either in the
ascorbic acid only fortified model system or the multi-
vitamin fortified model system. These results are similar
to those found in the previous studies for ascorbic acid
stability in TDT cans.

There was a marked difference in the half-lives
for TAA and RAA loss in the model system packaged in
paperboard boxes at 10% RH and 30°C versus the TDT cans
at 0.10 aw and 30°C (28 vs. 76 days, respectively). This
observed difference in half-lives decreases as the rela-
tive humidity of the storage atmosphere increases. The
use of the paperboard boxes permitted relatively free
transmission of atmospheric gases and moisture vapor
across the waxed liner as shown by the moisture vapor

5g H20 - cm/m2 - h mmHg)

transfer coefficient (7.25 x 10-
reported by Purwadaria (1976). Since the product was
packaged unequilibrated at a very low moisture content,
the driving force for moisture equilibration would be
into the product resulting in a gradual increase in the
moisture content. Under these conditions, the observed
rate of TAA and RAA loss at 40 and 85% RH should be

less if the system had been equilibrated to these

aw conditions prior to storage. The increasing moisture
content and aw (Figure 14) with storage time are inter-

preted to account for the differences in the observed

destruction rates of TAA and RAA in the model system

9O

stored at 40 and 85% RH and 30°C in the boxes and the
model system packaged in TDT cans at 0.40 and 0.65 aw
and stored at 30°C.

The dramatic differences observed for TAA and
RAA stability in the model food system equilibrated to
0.10 aw and stored in TDT cans at 30°C versus the model
system packaged in boxes and stored at 10% RH at 30°C
cannot be accounted for on the basis of moisture content
or aw differentials. Equilibration of the boxed model
system at 10% RH is achieved in about two weeks (Figure 14),
yet the rate of loss of TAA and RAA in the model system
is 2-3 times faster than in the TDT containers. Since the
model systems were identical in composition, essentially of
equal aw, and the transfer of atmospheric gases could
occur only in the boxed model system, these data suggest
the involvement of oxygen in the stability of TAA and RAA
in the dehydrated food system at neutral pH. The influence
of oxygen on the stability of ascorbic acid has been
reported by Waletzko and Labuza (1976) in an intermediate
moisture model system at 0.85 aw. Karel and Nickerson
(1964), however, reported no significant difference in
the rate of ascorbic acid destruction in dehydrated
orange juice crystals stored in air or vacuum. Data from
this latter study probably reflects the stabilizing

influence of low pH and/or the chelation of metal ions

91

by the acids present in the soluble solids of the dehy-

drated orange juice crystals.

Ascorbic Acid Stability in
303 Cans

 

The comparison of the destruction rate at 30°C
for TAA and RAA at 10% RH in the boxed model system ver—
sus the 0.10 aw condition in the TDT cans suggested the
involvement of oxygen in the degradative process. This
led to a storage study in which a large reservoir of
gaseous oxygen could be provided with a constant water
activity. The 303 can was selected as the storage con-
tainer and was filled with the model food system forti-
fied with ascorbic acid equal to the mass (”15g) utilized
in the TDT can study. The model system stored in 303
and TDT cans at all storage conditions was identical in
composition and source of ingredients. The water activi-
ties and moisture contents of the two were identical
within experimental error; therefore, there should be
no effect of viscosity, reactant mobility or new catalytic
sites on the stability of ascorbic acid. The only vari-
able in the two systems is the relative amount of gaseous
oxygen at any condition of temperature and water activity.

The first-order degradation constants and half-
lives for TAA and RAA loss in the dehydrated model system
packaged in the 303 metal cans are presented in Table 8.

The dependence of the rate constants for TAA and RAA

92

degradation in the 303 cans on water activity and storage
temperature was similar to that found for ascorbic acid
destruction in the TDT cans. The experimental rate con-
stants for TAA and RAA loss in the 303 cans were found
to be greater than the corresponding rate constants in
the TDT cans.

The relationship between the rate constant and aw
was shown to be linear between 0.10-0.65 aw and is valid
for water activities below the capillary region of the
adsorption isotherm for the model system at all temper-

atures studied (Figure 15). Mathematically, this

function takes the form

kobs = k0 + (slope) aw (21)

where k0 is the intercept at zero water activity. This
relationship was also applied to the rate constants for
ascorbic acid degradation in the TDT can study. Although
the rate constants for TAA and RAA loss in the TDT can
are less than the corresponding rate constants for the
same product in 303 cans, they exhibit the same linear

vs. a (Figure 15).

relationship of kobs w

Calculation of the slopes obtained for the
observed rate constant vs. aw function (equation 21) at
constant temperature are presented in Table 15 for TAA

and RAA destruction in the model system stored in 303

93

.Auohm D
.om 4 .ON 0 .OH 0v memo Amy Hoe pcm A5 mom :H musumuwmswu wmmuoum
ucmuwcoo um >UH>Huom kumz .w> :oHumpmuqmp «NB NC mumu mcu mo QHnmcoHuMme

 

 

34
mo To Nd mo To Nd

11X? b\\\|q|\|\\#
. .N

. .¢

1 .m

. .m

A .0.

mA 4 .m.

 

 

.mH wusmHm

le>1

2.

( sfiop)

94

 

 

 

 

. «BAH. 55.4- . mmmH. m~.mu nm\aoa
Hm o mmsH. hm.H mm o HASH. mv.H- nm\mom
. mane. sm.¢ . mmso. mm.H om\eoe
mm 0 mass. -.m me H «who. mo.m om\mom
mH.H HmHo. hm.o mm.H moHo. sm.v om\aoe
mmmo. vo.m mHmo. mH.m o~\mcm

. omoo. om.m . smoo. om.~ OH\aoa
Hm m mmoo. Hm.~ mm H mwoo. mm.~ OH\mom

oHumu onHm mIOH x x oHumu onHm mIOH x ox
Uo\chHmucoo
a $9

 

.mumchucoo 909 can mom :H
pmuoum Emumwm HmGOE CH :oHumpmHmmp <<m paw mde How
>0H>Huom nwuw3 .m> ucmumcoo mama mo conmmHmmu HmwcHH .mH mHnme

95

and TDT cans. The ratio of the slopes for the 303 to TDT
containers decreased with increasing temperature, and
at 37°C the difference in slopes are within experimental
error. This relationship parallels the decrease of
oxygen solubility in water with increasing temperature.
The meaning of k0, which is obtained from equation 21
would be the degradation constant at 0.0 aw at the
respective temperatures. The relatively large standard
deviations associated with the ko values are a function
of the limited number of data points and the experimental
precision and are believed to be the reason for the dis-
parity in these values at the various temperatures.

The 303-TDT storage containers represent a
closed system, mass may neither enter nor leave the
container; however, mass may transfer across phase
boundaries within the container. The transfer of
gaseous oxygen into the product moisture at a given
water activity and temperature would be governed by the

equilibrium constant

 

K = (22)

where (02)d and (02)g represent the concentration of dis-
solved and gaseous oxygen, respectively. As the dis-
solved oxygen in the moisture of the dehydrated model

system is consumed via the destruction of ascorbic acid,

96

the concentration of dissolved oxygen would be maintained
in accordance with this equilibrium constant.

The model system used to study the stability of
TAA and RAA was prepared as described in the methods
section, equilibrated to the appropriate aw and packaged
in 303 cans and TDT cans, each with an equal mass of
product (15g). This left no headspace in the TDT con-
tainer, although a limited amount of gaseous oxygen was
present in the intra and intersitial spaces of the pro-
duct. The maximum concentration of dissolved oxygen
that could be present in any of the equilibrated model
system used (303 or TDT can studies) was estimated
assuming a moisture content of 10%, which corresponds
to an aw of 0.65 at 37°C. Using the solubility of
oxygen in water at 0°C (4.89 cm3/100cm3), even though
this temperature will yield an artificially high value,
the maximum oxygen content from air which could be
present in the moisture of the model system would be

4.36 x 10'7

moles of oxygen/cm3 of water. Assuming
complete solubility, calculation of the moles of
ascorbic acid in the fortified model system yielded

6 moles of ascorbic acid/cm3 of water in the

6.10 x 10'
model system at 0.65 aw and 37°C (based on 11.25 mg
ascorbic acid/100 9 model system, dry weight basis).
These data show that the maximum dissolved oxygen content

was approximately a factor of 10 less than the

97

theoretical stoichiometric ratio of one mole of oxygen/
mole of ascorbic acid reported by Hand and Greisan (1942)
for the oxidation of ascorbic acid in a system with

pH = 7.0. Thus, all equilibrated samples used in this
study, whether packaged in cans or boxes contained dis-
solved oxygen at a concentration less than that required
for a l to 1 mole ratio of oxygen to ascorbic acid.

No headspace was left in the TDT can after filling
it with model system. Thus, the moles of gaseous oxygen
which could be present in the TDT cans would vary depend-
ing upon the inter and intrastitial spaces in the packaged
model system. Using estimates that 20 to 40% of the
total volume of the TDT can could be occupied by air,
the total moles of gaseous oxygen were calculated and

5 to 5.4 x 10-5. This would repre-

ranged from 2.8 x 10-
sent approximately a lO-fold excess of moles of gaseous
oxygen/mole of ascorbic acid in the 159 of model system
in the TDT can. The moles of gaseous oxygen present

in the 303 can were estimated to be 5.1 x 10-3, provid-
ing an approximately 1000 fold excess of gaseous oxygen/
mole of ascorbic acid in the 303 can. Thus, for an equal
mass of product, the dissolved oxygen concentration
would be more nearly maintained at its initial level in
the container with the larger headspace.

Oxygen solubility in pure water is known to

decrease with an increase in temperature. Joslyn and

98

Supplee (1949) have reported that the solubility of

oxygen decreases as a function of soluble solids in

pure sugar solutions at constant temperature. There-
fore, as a function of temperature and soluble solids

the dissolved oxygen content in the moisture of the

model food system would be expected to be significantly
lower than that calculated for pure water. Assuming

that the ascorbic acid content (moles/ml) in the moisture
of the food product can only decrease in concentration
with increasing water activity and that the concentration
of dissolved oxygen (moles/ml) would be constant regard-
less of the water activity, the initial ratio of dissolved
oxygen to ascorbic acid would increase as a function of
water activity. This relationship was approximated by
assuming oxygen solubility in the model system to be
comparable to that in pure water, and plotting the
(Oz)/(RAA) ratio against aw (Figure 16). A sharp increase
in the slope of the curve is noted in the region cor-
responding.to capillarity of the adsorption isotherm

and suggests an increase in the rate of ascorbic acid
degradation not in accord with equation 21. This hypothe—
sis is supported by the ascorbic acid stability data
reported by Lee and Labuza (1975) for an intermediate
moisture food system equilibrated on either the adsorption
or desorption leg of the equilibrium moisture sorption

isotherm. Lee and Labuza (1975) found an increase in

99

.HAH>Huom umumz mo coHuocsu 6 mm H<<V \lmo. mo oHumm .GH musmHm
:4
m .0 N. .0 0.0 0.0 ¢.0 ”.0 N0 ..0

 

 

.N
4 1m.
kl
/
é 1m.
V
rv...
10 X
m
6.5. r
0.0». .o
6.8.
.o.

100

the rate of ascorbic acid destruction as a function of
the water activity (0.32-0.84). A sharp increase in
the slope at 0.65—0.70 aw is noted when their data are
treated according to equation 21 (Figure 17). This
break point corresponds to the capillary region of the
sorption isotherm.

The activation energies calculated for TAA and
RAA loss in the 303 cans at 0.40 and 0.65 aw (15-19
kcal/mole) in this study are within the range of those
found in the TDT can study. Similar activation energies
(16-20 kcal/mole) have been reported by Lee and Labuza
(1975) for the degradation of RAA in an intermediate
moisture model food system. Activation energies for
TAA and RAA degradation in the model system equilibrated
to 0.10 aw and packaged in 303 cans were found to be
10.7 kcal/mole versus 7-8 kcal/mole for the same model
system stored in the TDT containers. The discontinuity
in activation energies and the preceding discussion on
oxygen might suggest an anaerobic mechanism for aScorbic
acid destruction at 0.10 aw. Lee et a1. (1977) suggested
that above pH 4.1, the Ea for the degradation of ascorbic
acid under anaerobic conditions was a function of pH

according to

Ea = 1.840 (pH) - 4.178. (23)

2.8

2.4

2.0

|.6

k(days")

l.2

0.8

0.4

Figure

 

17.

101

 

l J J l

0.2 0.4 0.6 0.8 1.0
AW

Relationship of the rate of ascorbic acid
degradation vs. water activity at constant
temperature for an intermediate moisture
model food system (data from Lee and Labuza,
1975).

102

Applying this equation to the model food system at pH 6.8,
a predicted Ea of 8.5 kcal/mole is obtained, which cor-
relates favorably with experimental results for the model
food system stored in the TDT can. These data might
suggest the possible involvement of an anaerobic mecha-
nism for the degradation of ascorbic acid in the 0.10 aw
model system stored in the TDT can, but as emphasized
earlier, this is a tenuous conclusion.

Lee and Labuza (1975) have postulated that the
observed rate increase with water activity for ascorbic
acid degradation could not be rationalized by assuming
a dilution of the reaction species because the measured
half-lives did not remain constant. They attributed
the increased rate of destruction with increasing water
activity to the viscosity of their system. Using NMR
relaxation studies which showed that the viscosity was
inversely proportional to the moisture content of the
aqueous phase, Lee and Labuza (1975) suggested that as
the moisture content increased, the mobility of the
reaction species and/or catalysts would increase,
resulting in higher degradation rates of ascorbic acid.
Since viscosity was a linear function and the destruc-
tion rates a hyperbolic function of moisture content,
they concluded that several mechanisms must be expected
by which water in controlling the reaction, most likely
through a dilution effect and a control of the diffusion

rate.

103

As previously discussed, the increasing moisture
content of the system results in a dilution of the
ascorbic acid concentration, but the oxygen concentration
would remain constant. Assuming ascorbic acid degradation
to follow overall second-order kinetics and to be first-
order in both ascorbic acid and oxygen as reported by
Khan and Martell (1967), the ratio of oxygen to ascorbic
acid would increase as the water activity increases,
hence the reaction rate should increase. Thus, the
availability of oxygen in a low moisture food system
is interpreted as the major factor governing the degra-
dation of ascorbic acid. This rationale would also
apply to the effect of the desorption system on the sta-
bility of ascorbic acid.

Ascorbic Acid--Mineral Interaction
in 303 Cans

 

 

The effects of trace minerals on the stability
of ascorbic acid in dehydrated foods have not been pre-
viously studied. Extensive investigations concerning
metal ion catalyzed oxidation of ascorbic acid in
solution have been reported in the literature. Khan
and Martell (1967) have postulated an ascorbate-metal-
oxygen complex as the intermediate for copper and iron
catalysis of ascorbic acid. More recently, Blackburn
and Jameson (1975) have suggested a metal-metal dinuclear

ascorbate-oxygen complex to be the intermediate.

104

The influence of mineral fortification on the
storage stability of TAA and RAA in the model food sys-
tem packaged in 303 cans was studied at 30°C and 0.10,
0.40 and 0.65 aw. The experimental data were treated
by first-order kinetics and the rate constants and half-
1ives for TAA and RAA loss are presented in Tables 10 and
11. It is apparent from the rate constants for TAA and
RAA degradation that there was no catalytic activity
associated with any of the metals studied at 0.10 aw.
Water activity of 0.10 is less than the BET monomolecular
moisture content and, therefore, extremely limited
mobility of the metal ions would be anticipated.

No rate enhancement, due to metal ion catalysis
for TAA and RAA degradation, was observed at 0.40 aw,
except for the slight catalysis exhibited by CuCl2 and

CuSO Apparently, the copper(II) ion possesses slightly

4.
greater mobility than the other minerals studied at
0.40 aw. Khan and Martell (1967) have suggested that
the rate determining step of metal catalyzed ascorbic
acid oxidation involves a one electron-transfer from
the ascorbate anion to the metal ion, thus requiring

a stable lower valence form of the metal ion for elec-
tron acceptance. This rationale may explain the lack
of observed catalysis by calcium and zinc ions.

At or below 0.40 a , the kinetic data from this study

W

for copper and iron catalysis are interpreted as

105

reflecting the reduced ability of the aqueous phase to
solubilize and/or mobilize the metal ions in the model
system.

At 0.65 aw, which is in the capillary region of
the adsorption isotherm, a 2-3 fold increase in the
degradation rate of TAA and RAA is noted for each of
the added trace minerals as compared to the model system
not fortified with minerals. The exception was zinc
oxide, which did not exhibit a catalytic effect. Cataly—
sis by iron and copper are anticipated in view of pub-
lished results in the literature. Iron is probably
oxidized to Fe(III) before catalyzing the ascorbic acid
reaction. It is difficult to discern the precise effect
of mineral concentration of the reaction, but the degra-
dation rate appears to increase with increasing copper
or iron levels. This dependence on concentration is in
accord with the results of Khan and Martell (1967) and
Ogata et a1. (1968). There appears to be a negligible,
if any, effect of the anion on the catalytic properties
of copper and iron. Mapson (1945) and Ogata et a1. (1968)
have suggested that the chloride ion possesses some
catalytic properties, but such an effect was not
observed in this study.

Reaction data indicate that free mobility of the
Cu(II) or Fe(II) may require the complete hydration of

the metal ion in its octahedral configuration, which may

106

only be possible in the free water of the capillary
region. Below this region, the metal ions may have

one or more charged species from the product matrix as
substituted liquids, effectively immobilizing or limit-
ing the migration of the ions.

The effect of various zinc salts on the catalysis
of ascorbic acid degradation was unexpected. The absence
of a catalytic effect associated with ZnO can be explained
in view of its characteristic insolubility in aqueous
solution. The catalysis of ascorbic acid degradation
by zinc chloride and zinc sulfate was unexpected because
Zn(II) is reportedly inactive as a catalyst of ascorbic
acid oxidation due to the lack of a stable lower valence
state (Khan and Martell, 1967). The same expectation and
rationale apply to calcium carbonate.

The effect of added electrolytes on the activity
of oxygen in aqueous solution has been reviewed by Long
and McDevit (1951). The activity coefficients of oxygen
increased linearly with the electrolyte concentration,
and was dependent on both the cation and anion present.
This phenomenon could explain the rate enhancement
observed for ascorbic acid degradation in the model
ZnCl

system with added CaCO and ZnSO4. Assuming

3’ 2’
the Bronsted equation discussed in the literature review

to be valid for ascorbic acid degradation,

107

k = k0 113.5%: (24)
any increase in the activity coefficient of oxygen should
result in an increased degradation rate. The importance
of this electrolyte effect may not be detectable in
aqueous solution where the ascorbic acid degradation
is studied in the presence of a large excess of oxygen.
In the model system, calculations have shown the ascorbic
acid concentration to be approximately lO-fold greater
than the dissolved oxygen content. Thus, an increase in
the activity of oxygen in the model system could have a
marked effect.

The rate data for ascorbic acid degradation
in the dehydrated food system at 0.65 aw is interpreted
as reflecting two types of behavior: (1) catalysis by
copper and iron ions as suggested by Khan and Martell
(1967) and (2) rate enhancement due to the increased
activity coefficient of oxygen on the addition of calcium
and zinc electrolytes. From a practical standpoint,
trace mineral fortification of a low moisture food would
have negligible effect on ascorbic acid stability at

moderately low water activities.

Thermodynamic Activation Parameters
The activation parameters for TAA and RAA degra—

dation in the dehydrated model food system are presented

108

in Tables 12, 13 and 14. As depicted graphically, AS+
and AH+ increased in value as a function of water
activity (Figures 18 and 19), while the free energy of
activation, AGI, remained constant. The effect of

water activity on the destruction of ascorbic acid in
the dehydrated model system could be interpreted in a
manner analogous to chemical reactions in various water-
alcohol mixtures, i.e., as a solvent effect. Solvent
effects, such as degree of solvation, dielectric con-
stant and hydrogen bonding, should be important to the

reactivity of the dehydrated system.

Activation parameters calculated for the degra-
dation of TAA and RAA in the dehydrated model system
indicate that the degradation of ascorbic acid obeys
the isokinetic relationship (equation 14), as illus-
trated by Figures 20 and 21. The results from the TDT
and 303 can studies appear to fall on the same isokinetic
line, yielding 8 values of 273°K and 276°K for TAA and
RAA destruction, respectively. These 8 values are less
than the temperature range of the study, which implies
that the degradation reaction of ascorbic acid is con-
trolled by the entropy of activation (Blackadder and
Hinchelwood, l958a,b). The validity of this isokinetic
relationship suggests that, although there is a large

change in AH+ (hence Ba) and AS+ with aw, the constant

109

440 memo mom can A0 0 memo
Boa :H pmuoum Emummm HmpoE pmumnpmnmp on» :H mmoH <49 HON huH>Huom umuw3
co Amy :oHum>Huom mo amHmcucm can Afiv coHum>Huom mo amonucm mo mocmpcwmoo .mH musmHm

..<

 

 

 

0.0 0.0 ¢.0 N.0 0 0.0 0
In J O—I
O
a J O. 4ON:
V V
H S
3- of
d 0 .im_ nAunmI
Q 0
16 0
low low-
0
m 1mm 4 Low.

 

 

110

.H 40 mCmo mom pCm HOV mCmo
Boa CH pmuoum Emumwm Hmpoa pmumupwnmp OCH CH mmoH «dz How >HH>Huom Hmum3
:0 Amy CoHum>Huom Ho aaHmcqu UCm A<v CoHum>Huom mo amouqu mo mocmpcwmwo .mH musmHm

:4.

 

 

 

 

 

Rb mYo vAV N0. 0 0A0 0H0 va _N0 0
O
a
O
40 40.-
c
.mv- gAUMwI
V V
H S
.8- .8.
.0. 40m:
4
EON 10v-
0
m .mm 4 .Om-

 

111

 

20 -
l6-
4‘
4»
I b
d
l2'
8 .
H—I 44L HI .j_l J
-40 -20 O
#
AS
Figure 20. Isokinetic relationship of TAA degradation in

dehydrated model system stored in TDT cans
(A) and 303 cans (A).

 

112

 

Figure 21.

i 1
-4O {-20
AS

Isokinetic relationship of RAA degradation in
dehydrated model system stored in TDT cans
(A) and 303 cans (fi).

113

free energy of activation indicates the degradation
reaction of ascorbic acid in the model system, in all

likelihood, follows one mechanism.

The data presented earlier for the involvement of
oxygen in the degradation of ascorbic acid suggest a
possible rationale for the observed dependence of AS+,
AH+ and ACT on aw. The entropy of activation is a
measure of the randomness of the activated complex rela-
tive to the reactants. A negative A81 represents a loss
in entropy or loss of freedom as the reactants proceed
through the transition state, whereas a positive AS+
suggests a loosely bound activated complex and a gain
in freedom. A decrease in AS* with increasing water
activity may be due to increased solvation of the reac-
tants and the activated complex or a decreased effective
charge on the transition state. At 0.10 aw in the model
system, the ASI of -43 e.u. obtained for TAA degradation
in the TDT can (Table 12) indicates a loss in entropy
and the formation of a tightly bound activated complex.
As the system approaches 0.65 aw, the entropy of acti-
vation approaches zero, suggesting little or no dif-
ference in the internal degrees of freedom of the tran-
sition state versus the reactants. The experimental
data for ascorbic acid degradation in the model system

show a dependence on the availability of oxygen. Thus,

114

a possible reaction intermediate would be the ascorbic
acid-oxygen complex postulated by Khan and Martell (1967).
Since the free energy of activation is constant, the
change in A81 could be interpreted as an increase in
degree of solvation of both the reactants and transition
state, as the water activity increases to 0.65. In a
like manner, the change in AHI as a function of aw
reflects increased solvation of the reactants and tran-
sition state with increased water activity, but the
degradation of ascorbic acid in the model system is
entropy controlled.

Little experimental data are available from the
literature which may be treated in a similar manner. The
activation parameters of Lee and Labuza (1975) for the
degradation of ascorbic acid in an intermediate moisture
system (aw = 0.32 - 0.84) have been calculated and are
presented in Table 16 and Figure 22. The free energies
of activation were about 18 kcal/mole compared to the
19 - 20 kcal/mole found in the present study. The
enthalpy and entropy of activation were also constant
at approximately 18 kcal/mole and zero e.u., respectively.
The activation parameters calculated for the adsorption
and desorption studies did not indicate a dependence on
the sorption hystersis phenomenon. These calculations

suggest that the degradation of ascorbic acid in the

115

.AmhmHv munnmq pCm 004 m0 mump Eoum pmpmeono muwumfimnmm CoHum>HuU¢m

 

 

 

 

 

 

e.uH o.>H ¢+ v: m.mH m.mH vm.o
m.nH m.>H MH+ m+ m.HN m.wH mh.o
m.mH H.mH ml H+ h.mH ¢.mH no.0
u.wH >.mH HH+ o o.NN m.mH .Hm.o
H.mH o.mH m+ H+ m.mH «.mH Nm.o
CoHHQHOmmp
m.>H v.5H N+ H+ w.hH w.hH vm.o
m.hH m.>H w: NI m.mH N.5H mb.o
v.mH «.mH 0 ml v.mH H.mH no.0
m.mH h.mH HH+ HI o.NN v.mH Hm.o
N.mH H.mH 5+ m+ m.HN m.om Nm.o
COHHQHOmpm
N Con H CCH N CCH H Con N CCH H can
3
m
HmHoa\Hmoxv +o< 1.6.mc +m< HmHoB\Hmoxv +m<
m.Emum>m Hobos musumHOE mumemEHmuCH Cm CH
CoHumpmHmmp pHom UHQHoomm pmoscmu How mumumfimumm COHum>Huo< .mH mHnt

116

20

 

no
-20 :5 1..”
AS

Figure 22. Isokinetic relationship for ascorbic acid
degradation in an intermediate moisture model
system (data from Lee and Labuza, 1975).

117

intermediate moisture model system of Lee and Labuza
(1975) followed a similar pathway as found in the
present study.

Limited data are available for the calculation
of activation parameters for the degradation of ascorbic
acid and other nutrients in low moisture food systems.
However, information of this type may lead to a better
understanding of the role of water in nutrient stability,

nonenzymatic browning and lipid oxidation.

SUMMARY AND CONCLUSIONS

The storage stability of total, reduced and
dehydroascorbic acid in a low moisture dehydrated model
food system was investigated as a function of storage
temperature, water activity, oxygen availability, vitamin
interactions and trace mineral catalysis.

The destruction of TAA and RAA was dependent on
the water activity and storage temperature, and its loss
could be adequately described by first-order kinetics.
Maximum stability of TAA and RAA was observed at low
storage temperatures and 0.10 aw, which was below the
BET monomolecular moisture content. A linear relation-
ship between the degradation constants for TAA and RAA
loss with water activity was found, which contradicts
the hypothesis of Salwin (1959) that the BET mono-
molecular moisture content should correspond to the
water activity of greatest stability. The data pre-
sented tends to support the suggestion by Karel and
Nickerson (1964) that there is no water activity below

which the degradation of ascorbic acid ceases.

118

119

Dehydroascorbic acid was measurable during the
storage study. DAA concentration decreased at storage
temperatures and water activities above 20°C and 0.24 aw,
respectively. An exception was noted at 10°C and low aw
in which there was a small initial increase in DAA con-
centration followed by a gradual decrease after 25 days.
It was concluded that the RAA degraded to DAA, which
underwent further destruction.

Oxidation reactions in dehydrated food systems have
been shown to be a function of reduced viscosity promot-
ing mobility of reactants, dissolution of precipitated
catalysts and swelling of solid matrices exposing new
catalytic sites (Labuza et al., 1970). These factors
certainly explain to some extent the increase in reaction
rates for TAA and RAA in dehydrated food systems. How-
ever, the identical model system was prepared and equil-
ibrated at the same aw, and stored in TDT and 303 cans
so that the only variable was the amount of gaseous
oxygen available. Factors other than oxygen would have
been internally compensated for in this study. The
increased destruction rates for ascorbic acid in the
303 cans was attributed to the presence of the large
gaseous oxygen reservoir in the 303 can and the main-
tenance of the initial dissolved oxygen concentration

according to the equilibrium constant.

120

The effects of added riboflavin and vitamin A
on the storage stability of ascorbic acid was studied
using the same model system. Comparison of degradation
rates for TAA and RAA loss in the multi-vitamin fortified
model system versus the ascorbic acid fortified model
system showed no significant effect.

The catalytic properties of added trace minerals
(Fe, Cu, Zn, Ca) on the degradation of ascorbic acid was
investigated at 30°C and three water activities. At

and below 0.40 a no rate enhancement of ascorbic acid

w'
degradation was observed over the nonmineral fortified
model system due to the presence of these minerals,
except for copper at 0.40 aw. At 0.65 aw, which is in
the capillary region of the adsorption isotherm, a 2-3
fold increase in degradation rate over the nonmineral
fortified system was observed for each form of the
added trace mineral. The lone exception was zinc oxide,
which did not exhibit catalysis due to its inherent
insolubility. The catalytic effect of Fe and Cu ions

in the capillary region of the adsorption isotherm were
explained on the basis of the mobility of these ions.
Catalysis by soluble zinc electrolytes and calcium

ions was attributed to the electrolyte effect on

increasing the activity coefficient of oxygen, thus

increasing the rate of ascorbic acid degradation.

121

Activation parameters were calculated for the
degradation of ascorbic acid in the model system stored
in TDT and 303 cans. Although a significant change in
the energy of activation was noted as a function of
water activity, the free energy of activation remained
constant. It is concluded that the degradation of
ascorbic acid in the dehydrated model system followed

the same mechanism at all water activities studied.

REFERENCES

REFERENCES

Acker, L. 1962. Enzymatic reactions in foods of low
moisture content. Adv. Food Res. 11:263.

Acker, L. 1969a. Enzyme activity at low water contents.
Recent Adv. Food Sci. 3:239.

Acker, L. 1969b. Water activity and enzyme activity.
Food Tech. 23:1257.

Acker, L. and R. Weise. 1972a. Lipase behavior in
systems of low water content. I. Influence
of the physical state of the substrate on
enzymic lipolysis. Lebensmittel - Wiss. u.
Technol. 5:181.

Acker, L. and R. Weise. 1972b. Behavior of lipase in
systems of low water content. II. Enzymic
lipolysis in the range of extreme low-water
activity. Z. Lebensmittelunters. u. - Forsch.
159:205.

Adamson, A. W. 1976. Physical Chemistry of Surfaces,
Wiley and Sons, New York.

Aylward, F. and I. D. Morton. 1971. Vitamin fortifi-
cation of foods. Proc. 3rd. Int. Congr.
Food Sci. Technol. 192.

Bauernfeind, J. C. and D. M. Pinkert. 1970. Food pro-
cessing with ascorbic acid. Adv. Food Res.
18:219.

Berlin, E., P. G. Kliman and M. J. Pallansch. 1966.
Surface areas and densities of freeze-dried
foods. Ag. Food Chem. 14:15.

Blackadder, D. A. and C. Hinchelwood. 1958a. Kinetics
of the rearrangement and oxidation of hydra-
zobenzene in solution. (1). Rearrangement
and spontaneous oxidation. 3. Chem. Soc. 2720.

122

123

Blackadder, D. A. and C. Hinchelwood. l958b. Kinetics
of the rearrangement and oxidation of hydrazo-
benzene in solution. (11). Catalyzed oxidation.
3. Chem. Soc. 2728.

Bronsted, J. N. 1922. Theory of the rate of chemical
reactions. Z. Physik. Chem. 102:169.

Brunauer, S. 1945. The Adsorption of Gases and Vapors.
Princeton University Press. Princeton, New
Jersey.

 

Brunauer, S., P. H. Emmet and E. Teller. 1938. Adsorp-
tion of gases in multimolecular layers. J. Am.
Chem. Soc. 66:309.

Caurie, M. 1970. A new model equation for predicting
safe storage moisture levels for optimum sta-
bility of dehydrated foods. J. Fd. Technol.
5:301.

Caurie, M. 1971a. A practical approach to water sorption
isotherms and the basis for the determination of
optimum moisture levels of dehydrated foods.

J. Fd. Technol. 6:85.

Caurie, M. 1971b. A single layer moisture absorption
theory as a basis for the stability and avail-
ability of moisture in dehydrated foods. J. Fd.
Technol. 6:193.

Coggiola, I. M. 1963. 2,5-dihydro—2-furoic acid: A
product of the anaerobic decomposition of
ascorbic acid. Nature, London. 200:954.

Cox, E. G. and T. H. Goodwin. 1936. Crystalline
structure of the sugars (III) ascorbic acid
and related compounds. J. Chem. Soc. 769.

Csuros, Z. and J. Petro. 1955. Autoxidation of ascorbic
acid as a function of temperature. Acta Chim.
Acad. Sci. Hung. 1:199.

de Boer, J. H. and C. Zwikker. 1929. Adsorption as a
result of polarization-adsorption isotherm.
Z. Phys. Chem. 133:407.

Dekker, A. D. and R. G. Dickinson. 1940. Oxidation
of ascorbic acid by oxygen with cupric ion as
catalyst. J. Am. Chem. Soc. 63:2165.

124

Dennison, D., J. Kirk, J. Bach, P. Kokoczka and D.
Heldman. 1977. J. Food Process. Preserv.
1:43.

Duden, R. 1971. Enzymic reactions in foods with very
low water activity. Lebensmitte1.-Wiss. u.
Technol. 6:205.

Dye, J. L. and V. A. Nicely. 1971. A general purpose
curve-fitting program for class and research
use. J. Chem. Edu. 66:443.

Duckworth, R. B. 1971. Differential thermal analysis
of frozen food systems. I. The determination
of unfreezable water. J. Fd. Technol. 6:317.

Duckworth, R. B. and G. M. Smith. 1963. Diffusion of
solutes at low moisture levels. Recent Adv.
Food Sci. 6:230.

Eichner, K. 1976. The influence of water content on non-
enzymic browning reactions in dehydrated foods
and model systems and the inhibition of fat oxi—
dation by browning intermediates, in Water
Relations of Foods: ed. R. B. Duckworth.
Academic Press, New York.

 

Exner, O. 1964. Enthalpy—entropy relation. Coll.
Czech. Chem. Commun. 12:1094.

Exner, O. 1973. The enthalpy-entropy relationship.
Prog. Phys. Org. Chem. 11:411.

Eyring, H. 1935. Activated complex and the absolute
rate of reactions. J. Chem. Phys. 6:107.

Feeney, R. E., G. Blankenhorn and H. B. F. Dixon. 1976.
Carbonyl-amine reactions in protein chemistry.
Adv. Prot. Chem. 16:135.

Fennema, O. R. 1976. "Water and ice" in Principles of
Food Science, pt. 1, Food Chemistry. ed. 0. R.
Fennema, Marcel Dekker, Inc., New York.

Finholt, P., R. B. Paulssen and T. Higucki. 1963.
Rate of anaerobic degradation of ascorbic acid
in aqueous solution. J. Pharm. Sci. 61:948.

Fox, K. K., V. H. Holsinger, M. K. Harper, N. Howard,
L. S. Pryor and M. J. Pallansch. 1963. Measure—
ment of the surface areas of milk powders by a
permeability procedure. Food Tech. 11:127.

125

Frenkel, Y. I. 1946. Kinetic Theory of Liquids. The
Clarendon Press, Oxford.

 

Grant, N. H. and H. E. Alburn. 1965. Fast reactions
of ascorbic acid and H202 in ice, a presumptive
early environment. Biochemistry 6:1913.

Hand, D. B. and E. C. Greisen. 1942. Oxidation and
reduction of vitamin C. J. Am. Chem. Soc.
66:358.

Harkins, W. D. and G. Jura. 1944. A vapor adsorption
method for determination of the area of a solid
without assumption of molecular area. J. Am.
Chem. Soc. 66:1366.

Heldman, D. R. 1974. Computer simulation of vitamin
stability in foods. Paper presented at seminar
on Stability of Vitamins in Foods for the Associ-
ation of Vitamin Chemists, Chicago, IL.

Henderson, 8. M. 1952. A basic conception of equilibrium
moisture. Agr. Eng. 66:24.

Huelin, F. E. 1953. Studies on the anaerobic decompo—
sition of ascorbic acid. Food Res. 16:633.

Huelin, F. E., I. M. Coggiola, G. S. Sidhu and B. H.
Kennet. 1971. The anaerobic decomposition of
ascorbic acid in the pH range of foods and in
more acid solutions. J. Sci. Fd. Agric. 66:540.

Jameson, R. F. and N. J. Blackburn. 1975. The role of
Cu-Cu dinuclear complexes in the oxidation of
ascorbic acid by 02. J. Inorg. Nucl. Chem.
37:809.

Jameson, R. F. and N. J. Blackburn. 1976a. Role of
copper dimers and the participation of copper
(III) in the copper catalyzed autoxidation of
ascorbic acid. Part II. Kinetics and mechanism
in 0.100 mol dm'3 potassium nitrate. J. C. S.
Dalton 534.

Jameson, R. F. and N. J. Blackburn. 1976b. Role of
copper dimers and the participation of copper
(III) in the copper catalyzed autoxidation of
ascorbic acid. Part III. Kinetics and mechanism
in 0.100 mol dm‘3 potassium chloride. J. C. S.
Dalton 1596.

126

Joslyn, M. A. and H. Supplee. 1949. Effect of sugars
on conversion of dehydroascorbic acid. Food
Res. 16:216.

Karel, M. 1973. Recent research and development in the
field of low-moisture and intermediate foods.
CRC Crit. Rev. Food Tech. 6:329.

Karel, M. and J. T. R. Nickerson. 1964. Effect of
relative humidity, air and vacuum on browning
of dehydrated orange juice. Food Technol.
18:104.

Karel, M., T. P. Labuza and J. F. Maloney. 1967.
Chemical changes in freeze-dried foods and
model systems. Cryobiology, 6:288.

Karel, M., S. Mizrahi and T. P. Labuza. 1971. Computer
prediction of food storage. Modern Pack. 66:54.

Khan, M. M. Taqui and A. E. Martell. 1967a. Metal ion
and chelate catalyzed oxidation of ascorbic
acid by molecular oxygen. I. Cupric and
ferric ion catalyzed oxidation. J. Am. Chem.
Soc. 66:4176.

Khan, M. M. Taqui and A. E. Martell. 1967b. Metal ion
and metal chelate catalyzed oxidation of ascorbic
acid by molecular oxygen. II. Cupric and ferric
chelate catalyzed oxidation. J. Am. Chem. Soc.
66:7104.

Khan, M. M. Taqui and A. E. Martell. 1968a. The kinetics
of the reaction of iron(III) chelates of amino-
polycarboxylic acids with ascorbic acid. J. Am.
Chem. Soc. 66:3386.

Khan, M. M. Taqui and A. E. Martell. 1968b. Kinetics
of metal ion and metal chelate catalyzed oxi—
dation of ascorbic acid. III. Vanadyl ion
catalyzed oxidation. J. Am. Chem. Soc. 66:6011.

Khan, M. M. Taqui and A. E. Martell. 1969. Kinetics of
metal ion and metal chelate catalyzed oxidation
of ascorbic acid. IV. Uranyl ion catalyzed
oxidation. J. Am. Chem. Soc. 61:4668.

Kirk, J. R. and N. Ting. 1975. Fluorometric assay for
total vitamin C using continuous flow analysis.
J. Food Sci. 66:463.

127

Kurata, T. and Y. Sakurai. 1967. Degradation of
L-ascorbic acid and mechanism of non-enzymatic
browning reaction. II. Nonoxidative degradation
of L-ascorbic acid including formation of 3-
deoxy-L-pentosene. Agric. & Biol. Chem. 61:170.

Kurpianoff, J. 1959. "Bound water in foods," Funda-
mental Aspects of the Dehydration of Foodstuffs,
the Macmillian Company, New York, NY.

 

Kwolek, W. F. and G. M. Bookwalter. 1971. Predicting
storage stability from time-temperature data.
Food Tech. 66:1025.

Labuza, T. P. 1968. Sorption phenomena in foods. Food
Tech. 66:263.

Labuza, T. P. 1971. Kinetics of lipid oxidation in
foods. CRC Crit. Rev. Food Tech. 6:355.

Labuza, T. P. 1976. "Interpretation of sorption data
in relation to the state of constituent water"
in Water Relations of Foods. ed. R. B. Duck-
worth. Academic Press, New York.

 

Labuza, T. P., J. F. Maloney and M. Karel. 1966. Autoxi-
dation of methyl linoleate in freeze-dried model
systems. II. Effect of water on cobalt catalyzed
oxidation. J. Food Sci. 61:885.

Labuza, T. P., S. R. Tannenbaum and M. Karel. 1970.
Water content and stability of low-moisture
and intermediate moisture foods. Food Tech.
24:543.

Labuza, T. P., S. Cassil and A. J. Sinsky. 1972. Sta-
bility of intermediate moisture foods. 2:
Microbiology. J. Food Sci. 66:160.

Laidler, K. J. 1965. Chemical Kinetics, second edition,
McGraw—Hill, New York.

 

Langmuir, I. 1918. The adsorption of gases on plane
surfaces of glass, mica and platinum. J. Am.
Chem. Soc. 66:1361.

Lee, S. H. and T. P. Labuza. 1975. Destruction of
ascorbic acid as a function of water activity.
J. Food Sci. 66:370.

128

Lee, Y. C., J. R. Kirk, C. L. Bedford and D. R. Heldman.
1977. Kinetics and computer simulation of
ascorbic acid stability of tomato juice as
functions of temperature, pH and metal catalyst.
J. Food Sci. 66:640.

Leffler, J. E. 1955. The interpretation of enthalpy
and entropy data. J. Org. Chem. 61:533.

Loncin, M., J. J. Bimbenet and J. Lenges. 1969.
Influence of the activity of water on the
spoilage of foodstuffs. J. Food Technol.
3:131.

Mack, T. E., D. R. Heldman and R. P. Singh. 1976.
Kinetics of oxygen uptake in liquid food.
J. Food Sci. 61:309.

Maloney, J. F., M. Karel, and T. P. Labuza. 1966.
Autoxidation of methyl linoleate in freeze-dried
model systems. I. Effect of water on the auto-
catlyzed oxidation. J. Food Sci. 61:878.

Mapson, L. W. 1945. Influence of halides on the gxi-
dation of ascorbic acid. 2. Action of Cl on
the cupric-cuprous system. Biochem. J. 66:228.

Martin, M. F. 1958. Factors in the development of oxi-
dative rancidity in ready to eat oat flakes.
J. Sci. Food Agric. 6:817.

Matz, S., C. S. McWilliams, R. A. Larsen, J. H. Mitchell,
Jr., M. McMullen and B. Layman. 1955. The
effect of variations in moisture content on the
storage deterioration rate of cake mixes. Food
Tech. 6:276.

Miller, J. and M. A. Joslyn. 1949. Effect of sugars
on the oxidation of ascorbic acid. III.
Influence of pH and type of buffer. Food Res.
14:354.

Mizrahi, S., T. P. Labuza and M. Karel. 1970a. Computer
aided predictions of extent of browning in
dehydrated cabbage. J. Food Sci. 66:799.

Mizrahi, J., T. P. Labuza, and M. Karel. 1970b. Feasi-
bility of accelerated tests for browning in
dehydrated cabbage. J. Food Sci. 66:804.

129

Mizrahi, S. and M. Karel. 1977. Accelerated stability
tests of moisture sensitive products in permeable
packages by programming rate of moisture con-
tents. J. Food Sci. 66:1575.

Mousseri, J., M. P. Steinberg, A. I. Nelson and L. S.
Wei. 1974. Bound water capacity of corn starch
and its derivatives by NMR. J. Food Sci. 66:114.

Multon, J. L. and A. Guillot. 1976. "Water activity in
relation to the thermal in activation of enzymic
proteins," in Water Relations of Foods. ed.

R. B. Duckworth, Academic Press. New York.

 

Ogata, Y., Y. Kosugi and T. Morimato. 1968. Kinetics of
cupric salt-catalyzed autoxidation of L-ascorbic
acid in aqueous solutions. Tetrahedron 66:4057.

Palnitkar, M. P. and D. R. Heldman. 1971. Equilibrium
moisture characteristics of freeze-dried beef
components. J. Food Sci. 66:799.

Pekkarinan, L. 1974. The mechanism of the autoxidation
of ascorbic acid catalyzed by iron salts in
citric acid solution. Finn. Chem. Lett. 233.

Pincock, R. E. and T. E. Kiovsy. 1966. Kinetics of
reactions in frozen solutions. J. Chem. Edu.
43:358.

Pitt, J. I. 1976. "Xerophilic fungi and the spoilage of
foods of plant origin," in Water Relations of
Foods. ed. R. B. Duckworth. Academic Press,
New York.

 

 

Polanyi, M. 1928. Anwendurg der langmuirschen theorie
auf die adsorption von gasen an holzkohle.
Z. Physik Chem. A138:459.

Pope, G. G. 1972. Effect of time, temperature and for-
tification level on the retention of ascorbic
acid in fortified tomato juice. Ph.D. thesis,
Ohio State University, Columbus, Ohio.

Purwadaria, H. K. 1977. Simulation of nutrient sta-
bility in dry foods during storage. M. S. Thesis,
Michigan State University, East Lansing, MI.

Quast, D. G., M. Karel and W. M. Rand. 1972. Development
of a mathematical model for oxidation of potato
chips as a function of oxygen pressure, extent of
oxidation and equilibrium relative humidity.

J. Food Sci. 66:673.

130

Quast, D. G. and M. Karel. 1973. Simulating shelf-life.
Modern Pack. 66:50.

Reimer, J. 1977. Prediction of vitamin C. retention
in a stored dehydrated food. Ph.D. thesis,
Massachusetts Institute of Technology, Boston,
MA.

Reynolds, T. M. 1963. Chemistry of nonenzymic browning.
I. The reaction between aldoses and amines.
Adv. Food Res. 16:1.

Reynolds, T. M. 1965. Chemistry of nonenzymic browning.
II. Adv. Food Res. 16:167.

Reynolds, T. M. 1969. Nonenzymic browning: sugar-amino
interactions, in Symposium on Foods: Carbohy-
drates and Their Roles, ed. H. W. Schultz, R. F.
Cain and R. W. Wrolstad. Avi. Publishing
Company, Inc. Westport, Conn.

 

Rockland, L. B. 1957. A new treatment of hygroscopic
equilibria. Food Res. 66:604.

Rockland, L. B. 1969. Water activity and storage sta-
bility. Food Tech. 66:1241.

Salwin, H. 1959. Defining minimum moisture contents for
dehydrated foods. Food Tech. 16:594.

Salwin, H. 1962. The role of moisture in deteriorative
reactions of dehydrated foods. Proceedings of
Conference on Freeze-Drying of Foods. National
Academy of Sciences National Research Council.

Salwin, H. 1963. Moisture levels required for stability
in dehydrated foods. Food Tech. 11:1114.

Scatchard, G. 1922. Statistical mechanics and reaction
rates in liquid solutions. Chem. Rev. 16:229.

Schaleger, L. L. and F. A. Long. 1963. Entropies of
activation and mechanism of reactions in
solutions. Prog. Phys. Org. Chem. 1:1.

Schultz, H. W., E. Day and R. Sinnhuber. 1962. Lipids
and their oxidation. Avi. Publishing Company,
Inc. Westport, Conn.

 

Scott, W. J. 1957. Water relations of food spoilage
organisms. Adv. Food Res. 1:83.

131

Shanbhag, S., M. P. Steinberg and A. I. Nelson. 1970.
Bound water defined and determined at constant
temperature by wide-line NMR. J. Food Sci.
35:612.

Shtamm, E. V. and Yu. I. Shurlator. 1974a. Catalysis
of the oxidation of ascorbic acid by copper(II)
ions. I. Kinetics of the oxidation of ascorbic

acid in the copper(II) - ascorbic acid -
molecular oxygen system. Zhur. Fiz. Khin.
66:1454.

Shtamm, E. V. and Yu. I. Skurlator. 1974b. Catalysis
of the oxidation of ascorbic acid by copper(II)
ions. IV. Kinetics of the oxidation of
ascorbic acid in the copper(II) - ascorbic
acid - hydrogen peroxide system. Zhur. Fiz.
Khin. 66:1857.

Shtamm, E. V., A. P. Purmal and Yu. I. Skurlator. 1974a.
Catalysis of the oxidation of ascorbic acid by
copper(II) ions. II. Anaerobic oxidation of
ascorbic acid by copper(II) ions. Zhur Fiz.
Khin. 66:2229.

Shtamm, E. V., A. P. Purmal and Yu. I. Skurlator. 1974b.
Catalysis of the oxidation of ascorbic acid by
copper(II) ions. III. Mechanism of oxidation
by molecular oxygen. The role of copper(I)
ions. Zhur. Fiz. Khin. 66:2233.

Simon, I. B., T. P. Labuza, and M. Karel. 1971. Computer
aided prediction of food storage stability:
oxidative deterioration of a shrimp product.

J. Food Sci. 66:280.

Singh, R. P., D. R. Heldman and J. R. Kirk. 1976.
Kinetics of quality degradation: ascorbic
acid oxidation in infant formula during
storage. J. Food Sci. 61:304.

Sood, V. C. and D. R. Heldman. 1974. Analysis of a
vapor pressure manometer for measurement of
water activity in nonfat dry milk. J. Food Sci.
66:1011.

Stevens, H. H. and J. B. Thompson. 1948. The effect of
shortening stability on commercially produced
army ration biscuits. II. Development of oxi-
dation during storage. J. Amer. Oil Chemists
Soc. 66:389.

132

Szent-Gyorgi, A. 1928. Observations on the function of
peroxidase systems and the chemistry of the
adrenal cortex. Description of a new carbohy-
drate derivative. Biochem. J. 66:1387.

Thompson, L. U. and O. Fennema. 1971. Effect of freezing
on oxidation of L-ascorbic acid. J. Agr. Food
Chem. 16:121.

Tillmans, J., P. Hirsch and W. Hirsch. 1932a. Reduction
capacity of plant foodstuffs and its relation to
vitamin C(I) reducing substances of lemon juice.
Z. Untersuch. Lebensm. 66:1.

Tillmans, J., P. Hirsch and F. Siebert. 1932b. Reduction
capacity of plant foodstuffs and its relation to
vitamin C(II) reducing substance of lemon juice
as a stabilizer for the true vitamin. Z. Unter-
such. Lebensm. 66:21.

Tillmans, J., P. Hirsch and J. Jackisch. 1932c.
Reduction capacity of plant foodstuffs and its
relation to vitamin C(III) quantities of reducing
substance in various fruits and vegetables.

Z. Untersuch. Lebensm. 66:241. '

Toledo, R., M. P. Steinberg and A. 1. Nelson. 1968.
Quantitative determination of bound water by
NMR. J. Food Sci. 66:315.

Vojnovich, C. and V. F. Pfeifer. 1970. Stability of
ascorbic acid in blends with wheat flour, CSM
and infant cereals. Cereal Sci. Today 16:317.

Waletzko, P. and T. P. Labuza. 1976. Accelerated shelf-
1ife testing of an intermediate moisture food

in air and in oxygen-free atmosphere. J. Food
Sci. 61:1338.

Wanninger, L. A., Jr. 1972. Mathematical model pre-
dicts stability of ascorbic acid in food pro-
ducts. Food Tech. 66:42.

Weissberger, A., J. E. LuValle and D. S. Thomas, Jr.
1943. Oxidation processes. XVI. The autoxi-
dation of ascorbic acid. J. Am. Chem. Soc.
66:1934.

Weissberger, A. and J. E. LuValle. 1944. Oxidation pro-
cesses. XVII. The autoxidation of ascorbic acid
in the presence of copper. J. Am. Chem. Soc.
66:700.

133

Zilva, S. S. 1923a. Influence of reaction on the oxi-
dation of the antiscorbutic factor in lemon
juice. Biochem. J. 11:410.

Zilva, S. S. 1923b. Note on the conservation of the
potency of concentrated antiscorbutic extracts.
Biochem. J. 11:416.

Zilva, S. S. 1924a. Note on the conservation of the
potency of concentrated antiscorbutic prepara-
tions. II. Biochem. J. 16:186.

Zilva, S. S. 1924b. Antiscorbutic fraction of lemon
juice. Biochem. J. 16:632.

Zilva, S. S. 1925. The antiscorbutic fraction of lemon
juice. III. Biochem. J. 16:589.

Zilva, S. S. 1927. Note on the precipitation of the
antiscorbutic factor from lemon juice. Biochem.
J. 61:354.

Zsigmondy, R. 1911. Uber die struktur des gels der
kieselsaure theorie der entwasserung. Z. Anorg.
Chem. 11:356.

MICHIGAN smTE UNIV. LIBRQRIES
llHIWW“WIW”WWWmlWIII‘IHHIWIHII
31293008337416