ism-1414a?»

4 :30‘“ 7‘ «523':

.u..
A.

 

§fi

v '— -
{'1 ¥\Ir ~34-

 

 

. “‘
fil‘" (I...

Jig:

 

"1b"

.
:25;
'1

'.

‘
-

v‘I ,
‘7‘”.

‘ Li.“-

3 :-

v

o V't'

3‘.

1. my “Ear ,.,

‘..: Yf

;
735:3" 32‘“ V‘ur. “3

:1 I‘.'|",:\"_

"h‘h‘
(7.3». «,‘(Hhtnk

r
, .

3': .-

‘ <-
:33

.5; £2:

21“;
A

a: ., w
: 4 5‘7???

4. ‘é‘xfi’fag‘

56

.4;-
flfig‘a 1'.
.‘J 52:32.»-
$
- 1K) «.

1

‘xfpp

} .s

‘
LL‘E

~ I N
A, . 3

5‘ ‘ .4.
1-5:"

.1‘

Ԥ'v' 5'
m . :

T1}, Mfg}? ;‘ x
A ‘0 5"? u
“(I \_ :ifi'l’: ;:?:¢‘ ,

1193,15.

\J.K
:v‘ I

,3: 25,“. v
‘ ‘ Ir

‘ ii ‘» .'.\:-‘.' '12".
)‘3, a $5.]? .?.,_:.
3' - f‘..‘€;“.’.

If Y:

.
‘1 .J
m; 1" 1‘? "‘3'?
. ' ‘1. Cr ~
: 5 {Iv
. . » AV.
;’ ‘ '. §:"‘:‘2 'n 3

fix} ‘Ic VI‘ 4593’

fit 1W :tjwjm "H

.Mr‘z‘v“ 5‘97”?"
’5‘:

:11-

$5214“:-
“I 46‘, 5215
5“»:

.‘I' "
u’éfifiitjs I
3‘31): J

.3? .5 M ‘ -.
. .. In], ”x

V 5:,” -:.< 55: «,5

.‘HI‘- u

"-(-.. 4n

.

i
.1

iv-f-h‘SK ‘
14:1; 53': 5
' m.

an. .

III-u. 1.:
Hr“ i1'firf';...ta
ifs/5 ".5qu
y»- - '

 

 

1m 157-7

IIIIIIIII”

IIIIIIIIIIIIIIIII
I uamm .
Michigan State‘

University

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the
thesis entitled
AROMA ANALYSIS OF A FRUIT FLAVORED CEREAL PRODUCT:
INFLUENCE OF PACKAGE LINER SYSTEMS ON HEADSPACE D-LIMONENE

CONCENTRATION AS MEASURED BY GAS CHROMATOGRAPHY

AND CORRESPONDING SENSORY EVALUATION
presented by

Sherry Hsien-Wen Hsia

has been accepted towards fulfillment
of the requirements for

Packaging

 

 

M' 3 ° degree in

90K K {Jéé 4w

aor professor

KJack R. iaci n
Data May 27, 1988

 

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

 

 

MSU

 

 

RETURNING MATERIALS:
Place in book drop to

 

 

L13RAR155 remove this checkout from
—:-—. your record. FINES will

be charged if book is
returned after the date

“ stamped below.

93"

wfii

mi?

at

r55 0 5 £31

APR 0

9603

. paid: ,2 3
W 8.732.900
dé'b % ‘6 £004

3 £008

 

 

 

AROMA ANALYSIS OF A FRUIT FLAVORED CEREAL PRODUCT:
INFLUENCE OP PACKAGE LINER SYSTEMS ON HEADSPACE D-LIMONENE
CONCENTRATION AS MEASURED BY GAS CHROMATOGRAPHY

AND CORRESPONDING SENSORY EVALUATION

BY

Sherry Helen-wen naie

A THESIS

Sublitted to
nichigen State University
in pertiel fulfillment of the requirements
for the degree at

MASTER OP SCIENCE

school of Packaging

1989

ABSTRACT

AROMA ANALYSIS OP A FRUIT FLAVORED CEREAL PRODUCT:
INFLUENCE OF PACKAGE LINER SYSTEMS ON HEADSPACE D-LIMONENE
CONCENTRATION AS MEASURED EY GAS CHROMATOGRAPHY
AND CORRESPONDING SENSORY EVALUATION

by
Sherry Brien-wen neie

Two analytical procedures, namely package headspace
analysis for quantification of flavor volatiles and the
analysis for sorbed flavor constituents by the cereal liner
structures, as well as two sensory evaluation methods (i.e.
triangle test and consumer preference test) were applied to
assess the relative performance of two commercial package
liner structures on the flavor quality change of a fruit
flavored cereal product, where quality is dependent upon the
retention of aroma constituents. Storage stability studies
were carried out over a period of twelve months. For these
studies, d-limonene was selected as the probe compound due
to its dominant role in the cereal product's aroma profile
and its characteristic flavor.

The results of the objective and subjective tests
performed were evaluated to establish a relationship between
consumer preference to the product during storage and the

d-limonene headspace concentration for the respective

package liner structures.

Dedicated to my parents,
for their love, support, and strength
with sincere gratitude.

ii

ACRNOILEDGEMENTS

I would like to thank my major advisor, Dr. Jack Giacin,
for his guidance, encouragement, and support through my
academic endeavors.

Special thanks is extended to Dr. Mary Ann Filadelfi-
Keszi for her guidance in the sensory analysis and time
spent in evaluating my progress.

Appreciation is also expressed to Dr. Bruce Harte and
Dr. Dana Ott for their serving on the guidance committee.

I would like to acknowledge the efforts of Dr. Heidi
Hoojjat, members in the lab, and panels participated in the
sensory evaluation, for their patience, support and

technological advice through this research study.

iii

TABLE OF CONTENTS

Page
LIST OF TABLES.......................................... vi
LIST OF FIGURES........................................viii
INTRODUCTION............................................ 1
LITERATURE REVIEW....................................... 5

Barrier and Shelf Life............................... 5
Flavor Characteristics............................... 6
Flavor Analysis...................................... 10
The Use of Sensory Evaluation........................ 13
Environmental Effects on Stability of Flavors........ 15
Product/Package Interaction.......................... 16

MATERIAIS ANDMETHODSIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 20

Product and Materials................................ 20
Experimental Method.................................. 22
Storage Stability.................................... 22
Sampling Plan........................................ 23
Permeation Studies................................... 26
D-limonene Headspace Concentration Analysis.......... 31
Thermal Distillation/Desorption Test................. 33
Sensory Evaluation................................... 34
Triangle Test........................................ 34
Preference Test...................................... 36

RESULTS AND DISCUSSION. I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 38
Permeation Studies I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 38
D-Limonene Headspace Concentration................... 43
Sorption of D-Limonene by Liners..................... 50
Triangle TestI I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 62
Consumer Preference I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 64

CONCLUSIONIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 68

APPENDIXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 70

iv

Appendix A: Standard Calibration ........... ........ .
Appendix B: Sensory Evaluation Questionaire..........
Appendix C: Headspace Analysis Data..................
Appendix D: Sorption Analysis Data........... ....... .

BIBLIOGRAPHY. . . . . .

70
72
75
80

82

LIST OP TABLES

Table Page
1 Chemical Characteristics of D-Limonene............. 9

2 Contribution of Volatiles to Orange Juice
FlavorIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 11

3 Composition and Physical Properties of the
Two Liner systemSIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 21

4 Permeability Data for the HDPE Based Liner
StmctureIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 39

5 Permeability Data for the EVOH Based Liner
SthtureIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 41

6 D-Limonene Headspace Concentration Levels in
Two Cereal Package Liner Systems During 12
Months of Storage at 23 oC and 50% RH............... 44

7 Sorption Concentration Levels of D-Limonene
by Two Cereal Package Liners During 12 Months
of Storage at 23°C and 50% RH...................... 51

8 Triangle Test of Cereal Packaged in Two Liner
Systems During 12 Months of Storage at 23 oC
and 50% R110.IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 63

9 Consumer Preference for Cereal Packaged in
Two Package Liner Systems During 12 Months of
Storage at 23 oC and 50% RH......................... 65

10 Limonene/Ethyl Acetate Standard Calibration
Curve Data as Measured by Hewlett Packard
Model 5890A Gas Chromatograph...................... 70

11 Data of the D-Limonene Headspace Concentration
Analyses (Area Response)........................... 75

12 Data of the D-Limonene Headspace Concentration
Analyses (D-Limonene Headspace Concentration)...... 77

vi

13 ANOVA Table for Headspace Analysis on Storage
Time (Independent Variable) Versus D-Limonene
Headspace Concentration in EVOH Liner System
(Dependent Variable)............................... 79

14 Data of the Thermal Distillation Sorption Test..... 80

vii

LIST OP PIGURES

Figure Page
1 D-Limonene I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 1
2 Stacking Pattern of a Palletload.................. 25

3 Schematic of Quasi-Isostatic Permeation Test
ApparatuSIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 27

4 Permeation cell systemIIIIIIIIIIIIIIIIIIIIIIIIIIII 29

5 Permeation of D-Limonene Through the HDPE Based
Liner Structure at 23°C and 50% RH................ 40

6 Permeation of D-Limonene Through the EVOH Based
Liner Structure at 23° C and 50% RH................ 42

7 D-Limonene Headspace Concentration Levels in the
HDPE Based Cereal Package Liner Structure During
12 Months of Storage at 23°C and 50% RH........... 45

8 D-Limonene Headspace Concentration Levels in the
EVOH Based Cereal Package Liner Structure During
12 Months of Storage at 23° C and 50% RH........... 46

9 D~Limonene Headspace Concentration Levels in Two
Cereal Package Liner Structures During 12 Months
of Storage at 23° C and 50% RH..................... 48

10 Sorption Concentration Levels of D-Limonene by
the HDPE Based Cereal Package Liner Structure
During 12 Months of Storage at 23°C and 50% RH.... 52

11 Sorption Concentration Levels of D-Limonene by
the EVOH Based Cereal Package Liner Structure
During 12 Months of Storage at 23°C and 50% RH.... 53

12 Sorption Concentration Levels of D-Limonene by
Two Cereal Package Liner Structures During 12
Months of Storage at 23 oC and 50% RH.............. 54

13 Concentration Profile of a Penetrant in a
Polymer Film as a Function of Time................ 55

viii

14

15

16

Concentration Profiles of Methyl Salicylate
at 25° C in Two Hypothetical Multilayer Walls
at Steady-State................................... 58

Hypothetical Concentration Profile of D-limonene
at 25° C in Two Cereal Liner Structures at Steady
State of Transmission............................. 61

Limonene/Ethyl Acetate Standard Calibration

Curve as Measured by Hewlett Parkard Model
5890A Gas ChromatographIIIIIIIIIIIIIIIIIIIIIIIIIII 7].

ix

INTRODUCTION

For many years researchers have been studying consumer
behavior to foods. This information is not only of value
theoretically, but it gives a better understanding of those
factors which influence the acceptance of new food products
or product/package interrelationships. In the past, consumer
reactions were observed by the fluctuation of monthly sales
figures or by the sudden appearance of numerous letters from
consumers critical of the product (Horton, 1987).
Researchers agree that a much more productive method would
be to predict consumer preferences. Product development
scientists believe that by precise and diverse instrumental
measurements, consumer attitudes of food quality can be
accurately predicted, and consistent control of food
properties maintained (Rutenbeck, 1985). Unfortunately, the
buying public is not at all reticent about making its
sentiments known regarding its likes and dislikes. While
instrumental analyses are effective in assessing many food
properties, they cannot mimic human perception of food
quality precisely. Thus, sensory evaluation techniques have
gained more and more attention through the years. A new
approach for better understanding of consumer attitudes
towards foods is the correlation of human sensory responses

with the traditional chemical and physical measurements.

1

Although information on the importance of various
quality attributes for food purchase and consumption is
still fragmentary, the relative importance of sensory
attributes such as texture, flavor, and appearance have been
accentuated (Peryam, 1963; Szczesniak and Kleyn, 1963:
Schutz and Wahl, 1981). Moskowitz and Chandler (1978)
investigated consumer "trade off" propensities for flavor,
nutritional quality, and cost, and found that flavor usually
prevailed over the other sensory attributes. Schutz et a1.
(1986) studied the relative importance of nutrition, brand,
cost, and sensory attributes to food purchase and
consumption. It was found that flavor was consistently the
most important attribute in consumer response.

Loss of flavor from foodstuffs during storage has become
a major stumbling block to higher consumer acceptance.
Flavor scalping, a phenomenon resulting in the loss of
flavor compounds from products due to their absorption or
adsorption by the packaging material, is a potential concern
in those foods whose quality is associated with the
retention of volatile aroma constituents. It can shorten
product shelf life dramatically. Mohney (1986) indicated
that consumer preference was based, in part, on the
intensity of the aroma moieties (i.e. d-limonene) found
within the headspace of a fruit-flavored cereal package
liner. The use of appropriate packaging can maintain a
desirable intensity level of volatile aroma constituents in

foodstuffs. The retention of a foodstuff’s aroma will, in

part, depend upon: (1) the vapor pressure of the individual
components of the total aroma; (ii) the interaction of these
volatile organic moieties with other food components; and
(iii) the mass transfer characteristics of the package.
(Mohney, 1986).

The latter point, which includes the permeability and
solubility properties of the aroma moieties/package system,
should be of major concern in the selection and use of
plastic packaging materials for foods. The transport and
sorption of organic vapors through polymeric packaging
materials has been, and continues to be, the subject of
numerous investigations (Gilbert et al., 1983; Murray and
Dorschner, 1983: Zobel, 1982, 1984, 1985; Hernandez, 1984:
Baner et al., 1985: DeLassus, 1985; Murray, 1985; Hernandez
et al., 1986: and Mohney, 1986).

This study represents an extension of the studies
reported by Mohney (1986) and considers the effect of two
cereal package liner structures, namely a high density
polyethylene/ethylene vinyl acetate (HDPE/EVA) coextrusion
and a high density polyethylene/ethylene vinyl alcohol/
Suryln (HDPE/EVOH/Surlyn) coextrusion, on the quality of a
fruit-flavored cereal product, during twelve months of
storage. D-limonene was selected as the probe compound due
to its dominant role in the cereal product’s aroma profile

and its characteristic flavor.

The major objectives of this study are summarized below:

(1)

(2)

(3)

(4)

Evaluation of the relative performance of two
commercial cereal package liner structures with
respect to flavor retention (i.e. d-limonene
headspace concentration) and flavor sorption
(i.e. the amount of d-limonene being absorbed)
during twelve months of storage.

Sensory evaluation of cereal product packaged in
the two commercial cereal package liners
structures, with respect to flavor quality
change, during twelve months of storage.
Determination of consumer acceptance/preference
for the respective cereal package liner
structures with respect to product quality.
Evaluation of quantitative relationships between
the respective package liner systems with regard
to the headspace concentration of d-limonene and
the sensory responses to product quality, as a

function of storage time.

LITERATURE REVIE'

BARRIER AND SHELF LIFE

There has been an increase in recent years in the use of
plastics as food packaging materials. This has also prompted
researchers into investigating the potential interactions
that might arise between foods and packages and the effects
that these might have on food quality (Karel and
Heidelburgh, 1975; Gilbert and Mannheim, 1982).

In most cases, the environmentally omnipresent gaseous
reactants, water vapor and oxygen, can seriously effect
product stability (i.e. quality) under the usual food
storage and distribution conditions. The rate of transport
of such reactants across the partial barrier of the package
wall (intact or through breaks) can become the limiting
factor in shelf life. The corresponding transport rate of
low molecular weight compounds affecting flavor and aroma
may also be involved in quality changes during storage
(Gilbert, 1985).

Factors invloved in food/package compatibility which
influence the acceptability of foods include migration from
the packaging material to the product, absorption and
adsorption of flavor compounds inherent to product by the
packaging material (commonly referred to as scalping), and

5

wicking (Harte and Gray, 1987). Scalping of flavor compounds
is a concern for many products currently being packaged. For
example, citrus flavored products contain volatile, highly
aromatic compounds. When these compounds are selectively
sorbed by the packaging material, they no longer function as
flavor compounds and thus, the perceived quality of the
product is diminished.

These problems can be eliminated or reduced by employing
the appropriate barrier materials. Thus, selection of the
appropriate barrier to maintain the expected shelf-life of
foods is critical for successful marketing. Shelf-life of
foods is usually defined as the length of time that a
container or the material in a container will remain in a
saleable or acceptable condition, under specified conditions
of storage.

Product shelf-life is controlled by four factors: (1)
product characteristics: (2) the environment to which the
product is exposed during distribution: (3) the properties
of the package: and (4) the interaction of product and

package (Harte and Gray, 1987).

FLAVOR CHARACTERISTICS

Flavor represents a vast array of chemical classes.
Often, a flavor is composed of a number of related compounds

in precise proportions. Flavor of foods is the subtle

balance of various volatile and nonvolatile flavor compounds
(Gianturco and Biggers, 1974). It is not uncommon to
encounter flavors of 200 constituents or more (Heath and
Reineccios, 1986). Maarse (1984) reported that by the mid-
1980's, 4,300 different flavor compounds had been identified
in foods. Rijkens and Boelens (1975) have estimated that
probably 5,000-10,000 flavor compounds actually exist.

While flavors are most often considered as the summation
of a large number of compounds, there are studies which show
that the quantitative relationship of the individual flavor
components present is critical to full fruit flavor (Ahmed
et al., 1978b: Shaw, 1979; Schreier, 1981; Moshonas and
Shaw, 1986). Hence, 2—isobutylthiazole is critical to tomato
flavor; methyl- and ethyl- cinnamates to strawberry flavor:
methyl anthranilate to grape flavor: and benzaldehyde to
cherry flavor (Whitaker and Evans, 1987).

Cold-pressed peel oil, essence oil, aqueous phase
essence (industrially termed "natural citrus aroma"), and
folded cold-pressed oil are usually added to foodstuffs to
enhance citrus flavor. Over 90% of the essential oil of
orange is composed of the monoterpene hydrocarbon, d-
limonene (Figure 1). The chemical characteristics of d-
limonene are listed in Table 1.

D-limonene is an unsaturated hydrocarbon. This makes it
highly susceptible to oxidative and photooxidative
degradation. The hydroperoxides from oxidation of d-limonene

can decompose very rapidly to yield stable products

Table 1. Chemical Characteristics of D-Limonene.(a)

 

Molecular Structure
Molecular Weight
Density

Boiling Point
Melting Point

Molar Density
Soluble in

Vapor Mole Fraction in
Equilibrium with Pure
Liquid

C10H16
136.24
0.842 gm/mL
178°C
-74°c

6.17 x 10'3 gmole/mL
at 25°C

Ethyl Alcohol,
Diethyl Ether

1.9 x 10"3 at 25°C

 

(a) Weast et al., (1985).

10

including, limonene-1,2-epoxide and carvone (Farmer and

Alvapillai, 1942: Anandaraman, 1986). Such deteriorative
reactions are responsible for the off-flavors in orange

juice, which are often described as "turpentine-like" or
"painty" (Anandaraman, 1986).

Other terpenes in citrus oils include (5, p pinene and I3-
myrcene. Although these compounds have little flavor impact,
they can provide a solvent for the more potent flavorings
(Marshall, 1985). The contribution of volatiles to orange

juice flavor is shown in Table 2 (Durr, 1981).

FLAVOR ANALYSIS

Advances in chromatographic separation and instrumental
identification techniques (Shaw, 1977a, 1977b, 1979) have
helped to expand our knowledge of citrus flavors during the
past 25 years. Much information has been obtained on
specific flavor-contributing components of the major citrus
cultivars - orange, grapefruit, tangerine, lemon, and lime.
Extensive research conducted by Kirchner and Miller, 1953:
Wolford et al., 1961, 1964: Kesterson et al., 1971: and
Shaw, 1977a, 1977b has tried to identify, quantitate, and
organoleptically evaluate these fruit volatiles. In
addition, studies have also attempted to assess the
synergistic properties of various flavor blends (Ahmed et

al., 1978a: Shaw and Wilson, 1980).

11

Table 2. Contribution of volatiles to orange juice flavor. (a)

 

 

Contribution to Contribution to
Typical Flavor Off-Flavor
Important Desirable Precursors Detrimental
ethylbutyrate linalool linalool d-terpineol
neral limonene limonene carvone
geranial d-pinene valencene t-carveol
valencene nootkatone
acetaldehyde hexanal
octanal t-Z-hexenal
nonanal hexanol
O-sinensal 4-vinyl-
guaiacol
fi-sinensal 2,5-dimethyl-
4-hydroxy-3-
(2H)
furanone

 

(a) Durr, 1981.

12

Moshonas and Shaw (1984) reported that the use of fused
silica capillary column gas chromatography (GC) permitted
accurate quantification of volatile components in flavor
fractions, such as aqueous essence prepared from fresh
orange juice. This method led to studies of various indices,
component mixtures, and formulations in model systems, as
aides for determining profiles of flavor quality in specific
types of citrus flavored products (Johnson and Vora, 1983).

Nagy and Klim (1986) suggested that the use of infrared
spectroscopy, gas chromatography, and mass spectrometry
could greatly reduce the amount of time necessary to
identify and to quantify a flavor compound. They added that
the mose versatile instrument in flavor analysis is the gas
chromatography-mass spectrometer (GC-MS). Resolution of a
complex flavor mixture (i.e. fruit-flavored cereal) can be
attained by gas chromatography, utilizing highly efficient
capillary columns. After separation by GC, the eluted
components are identified by the mass spectrometer, through
their fragmentation patterns.

Although identification techniques have improved and,
many components, both volatile and nonvolatile, have been
identified, their subtle influence on quality and flavor
characteristics, and their relationships among various
citrus components are not fully known. The contribution of
essential oils to specific types of citrus flavor, their
stability and utilization are also not fully known. Thus,

more research on the relationship between flavor chemistry

13

and sensory evaluation of the individual flavor fractions in

foods will be required in the future.

THE USE OF SENSORY EVALUATION

In sensory evaluation, human observers or panelists act
as the analytical tool. Depending on the type of
investigation, the panelists can be grouped into three
types, namely: (i) highly trained experts: (ii) laboratory
panels; and (iii) large consumer panels. Highly trained
experts evaluate quality, and large consumer panels are used
to determine consumer reaction to a product. Sensory tests
performed by fairly large panels are valuable in predicting
consumer reactions. Evaluations by experts and trained
laboratory panels can be useful for control purposes, for
guiding product development and improvement, and for
evaluating quality. The trained panel can be particularly
useful in the assessment of product changes, for which there
is no adequate instrumentation (Larmond, 1977). Amerine et
al. (1965) and Larmond (1977) have provided general
guidelines for the selection and training of panelists.

The use of sensory evaluation to monitor flavor change
or deterioration during a shelf-life study requires careful
planning and a thorough understanding of the sensory
evaluation: physical and chemical composition of the

product: formulation: processing and packaging treatments;

l4

storage conditions: projected storage life; anticipated
sensory changes: experimental design: and sensory evaluation
procedures (Dethmers, 1979).

Both analytical and affective sensory methods may be
used to determine the shelf-life of a food/beverage product
(Prell, 1976). The method selected will be determined by the
purpose of the test. This may include the extension of
shelf-life by a change in processing, formulation, or
packaging: establishment of the shelf-life of a new product:
or determination of the shelf-life of an existing product.

Preference/acceptance tests are affective tests based on
a measure of preference or acceptance. In these tests,
samples are considered unacceptable when a sample obtains an
average panel score of 3.5 based on a seven point scale
(Gacula and Kubala, 1975), or when various deteriorative
changes in a freshly packed product accumulate, such that
the market quality would be judged lower than excellent or
that a taste panel would consider the quality to be at the
limit of acceptability.

Discriminatory tests are used to determine whether a
difference exists between two samples. This type of testing
requires trained or experienced panelists (Larmond, 1977).
It is considered limited at the point when a difference
between test and control samples can be detected by trained
panelists (Van Arsdel, 1969; Jul, 1969) or when the percent
of correct judgments in a triangle test comparing stored and

control samples ranges between 70-80 percent (Guadagni,

15

1957). The use of this procedure is limited to experiments
where a control sample can be maintained with a negligible
change over time and to studies where the number of
treatments is small (Gacula and Kubala, 1975).

Descriptive tests require trained panelists and are used
to determine the nature and intensity of the difference. If
a comparison of results from Profile Descriptive Analysis,
before and after storage, indicates changes in character or
intensity of the flavor, the samples are generally deemed

unacceptable (Dethmer, 1979).

ENVIRONMENTAL EFFECTS ON STABILITY OF FLAVORS

Few researchers have been working in this area. Durr et
al. (1981) investigated the influence of the behavior of d-
limonene and other aroma components on the sensory quality
of orange juice during filling and storage. Significant
losses of d-limonene, neral, geranial, ortanal, and decanal
from orange juice in carton packages were found. However,
these investigators concluded that the main quality
parameter for shelf-life was the storage temperature.

Hatzidimitriu et al. (1987) reported that some
multilayer packaging films lost their barrier properties
with respect to the permeation of organic vapors at higher

relative humidities.

l6

Wartenberg (1982) compared the quality of orange juice
when packaged in aseptic packages and glass bottles (with
and without headspace) during 6 months of storage. For the
first 5 months of storage, juice in glass bottles and carton
packages not having a headspace showed little difference in
the decomposition of ascorbic acid. The loss of ascorbic
acid was higher in the carton packages with headspace.
During the first 5 months of storage, sensory changes were
greater in the glass bottles than in either of the carton
packages. There was no difference bewteen the carton package
with headspace and that without headspace.

Granzer (1982) also reported on the quality of orange
juice in two carton packages designed to have a high barrier
against oxygen ingression. The material composition of the
packages was the same as used in Wartenberg’s (1982) study.
Deterioration of color and loss of ascorbic acid occured
during the first year of storage. These changes were said to

be due to the influence of oxygen.

PRODUCT/PACKAGE INTERACTION

Most of the potent flavorings in citrus flavored
products are highly lipophilic and can potentially interact
with the hydrophobic polymers used in the packages.

Much of the research in this area has focused primarily

on the permeability of plastics to gases and water vapor

17

(Lasoski, 1960: Karel et al., 1963; Masi and Paul, 1982).
Standard methods are readily available in the literature for
measuring the permeability of plastics (ASTM D 1434-82: ASTM
B 96-80). The gas and moisture protection requirements of
various foods are also well documented (Quast and Karel,
1972: Labuza, 1981). Another interaction that has received
considerable attention has been the migration of package
components into food (Crompton, 1979: Kashotock et al.,
1980: Crosby, 1981).

The loss of flavor components due to sorption by
plastics is also important and must be considered. Salame
and Temple (1974) reported that a 1% loss of an aroma
component to packaging results in a quality change that is
detectable by human olfactory senses. The emphasis in aroma
sorption has so far been on the sorption of citrus aroma
components by plastics (Durr et al., 1981: Mannheim et al.,
1987: Shimoda et al., 1984).

The absorption of d-limonene and specific flavorings
into the polyethylene lining of packages has been reported
by Hatzidimitriu et al., 1987: Ikegami et al., 1987; Kwapong
and Hotchkiss, 1987; Mannheim et al., 1987; and Shimoda et
al., 1987.

In 1981 Durr et al. studied the storage of orange juice
packaged in polyethylene lined cardboard and glass. It was
found that a loss of approximately 40% of d-limonene by
absorption onto the polyethylene layer of the soft packages

occurred within 6 days of storage. The sensory differences

18

of the juices from the two packages was not detectable until
after 27 days of storage at 20°C. After 62 days, the glass-
bottled juice was described as stale and musty. The soft-
packed juice was similarly described as stale and musty
after 90 days. Linalool, octanal, decanal and neral showed a
similar decrease in both packages.

Marshall (1985) investigated the absorption of d-
limonene by various polymers. He indicated that loss of d-
limonene into sealant layers between the juice and the
polyvinylidene chloride (PVDC) barriers of various
permeabilities (P = 0, 1, 4, and 8 cc oxygen @STP/m2 atm
day) was directly related to the thickness of the
polyethylene or polypropylene sealant layer rather than the
oxygen permeability of the film.

Hatzidimitriu et al. (1987) reported the permeation
rates (at 23°C) of several organic compounds found in
composite films at 0% and 75% relative humidity (RH).
Ethylene/vinyl alcohol copolymer and nylon combinations
exhibited superior barrier properties compared to
polyethylene terephthalate-glycol (PET-G) and polyvinylidene
chloride (PVDC) laminations even at elevated RH, provided
that moisture barrier films were also present in the
laminate construction.

Shimoda et al. (1987) revealed that the absorption of
flavor compounds into the inner layer of films depended on
the number of carbon atoms and molecular structure of the

flavor compounds, flexibility of the structure, and the

19

position of functional groups on the ring compounds. The
study showed that higher levels of sorption were observed in
crystallized polypropylene pouches than in high density
polyethylene ones, particularly for d-limonene and mycrene.
The authors proposed that the higher levels of sorption were
due to the difference in molecular structure between plastic
materials. In addition, density, geometrical isomerism,
polarity, crystallinity and steric hindrance of the
structure of the polymer films should also be taken into
consideration for the mechanism of sorption (Brown et al.,
1973: Peterlin, 1975: Kreituss, 1981: Ng et al., 1985).

Mohney et al. (1986) investigated the solubility and
permeability of limonene vapor in two cereal package liners.
They concluded that this was a necessary consideration when
estimating the storage quality of a packaged fruit-flavored
cereal product, since the quality was related to the
retention of aroma constituents within the package
headspace.

Although previous researchers had demonstrated flavor
sorption by packaging materials, to thebest of our
knowledge, no studies have been reported describing the
relationship between the sensory quality of the product and
the absorption of flavor compounds by packaging materials.
Even less is known about consumer acceptance for a product

after a determined period of storage.

MATERIALS AND METHODS

PRODUCT AND MATERIALS

A fruit flavored cereal product, whose quality depends
on the retention of volatile aromatic constituents, was
chosen for this study. The product, supplied by the cereal
manufacturer, was packaged in two commercial package liner
systems. Product evaluated thorough this study came from the
same production lot, which was shipped (August 7, 1987) to
the School of Packaging, Michigan State University, two days
after the production run.

The two package liner systems used in this study were a
high density polyethylene/ethylene vinyl acetate (HDPE/EVA)
coextrusion, and a high density polyethylene/ethylene vinyl
alcohol/Surlyn (HDPE/EVOH/Surlyn) coextrusion. Hereafter,
these films will be referred to as the HDPE and EVOH
structures respectively, thorough this thesis. The
composition and physical properties of the two liners are
listed in Table 3.

In addition to the packaged cereal, samples of the liner
materials were also obtained from the product supplier and
were stored in a desiccator over CaSO4 (0% relative
humidity) at ambient temperature (23°C). The roll stock

liner material samples were used as controls in the thermal

20

21

Table 3. Composition and physical properties of the two
liner systems. (

 

Physical Property EVOH HDPE

 

Description HDPE/EVOH/Surlyn HDPE/EVA
HDPE 1.70 mil
Tie 0.15 mil
EVOH 0.20 mil
Tie 0.15 mil

Surlyn Blend

0.40 mil
Caliper (P) 0.0026 0.0020
Yield (C) 11,200 14,663
WVTR (d) 0.25 0.23
02 Transmission (e) 0.15 N/A

 

(a) data obtained from product supplier.
(b) unit, inches.
(c) unit, inz/lb.

(d) WVTR, the wager vapor transmission rate, units in
grams/100 in /24 hrs.

(e) unit, cc/100 in2/24 hrs.

22

distillation analysis for sorbed volatiles, for comparison
with the liner materials which had been contacted with the
product for up to twelve months storage. Roll stock liner
samples were also used in permeation studies, to determine
the permeability constants for d-limonene in the respective

liner structures.

EXPERIMENTAL METHOD

STORAGE STABILITY

Eight pallet loads, each pallet load contained sixty six
shipping containers with twelve 11 oz. cereal boxes per
shipping container and a total of 6,336 boxes of cereal,
were received from the supplier. Four pallet loads of
product were packaged in the high density polyethylene
(HDPE) based liner and the other four pallet loads of
product were packaged in the ethylene vinyl alcohol (EVOH)
based liner. The individual pallet loads were identified
with respect to the liner structure and stored in the School
of Packaging storage room at temperatures ranging from 21°C
(70°F) to 26°C (80°F) and 50% relative humidity, thoroughout
the study. The storage environment maintained good
ventilation and no foreign flavors were detected within 350

square feet of the storage room.

23

Each pallet load was shrink-wrapped and loaded on wood
pallets to minimize the influence of ground temperature and
moisture. There was about one foot between each pallet load.
The cereal packages were set in the storage room for at
least three days, to allow the volatile aromas within the
liner structures to equilibrate, before the initiation of

the headspace concentration analysis.

SAMPLING PLAN

A random sampling plan was developed by Mr. Dennis Young
of the School of Packaging, Michigan State University, and
was employed in this study. This program was designed to use
individual random numbers to represent the location of boxes
of product within cases and tiers, under the condition that
the orientation and stacking pattern of the cases was well
defined. In this sampling plan, random numbers were
generated by a computer program. The random number is
comprised of a three digit number. The hundredth number
represented the tier number, the tenth number the case
number, and the last number the box number, which is counted
clockwise from the manufacturer’s joint for the container
from which the sample was withdrawn. For example, the code
number indicates 231 means that the sample was pulled from
the first box in the third container from the second tier of

the pallet load. The random number was then affixed to the

24

box chosen from the pallet load, according to the sampling
plan. The storage location of any package can be traced by
this random number in order to examine factors relating to
aroma loss during storage.

The product was shipped in eight pallet loads, with each
pallet load having six tiers. Every tier was composed of
eleven cases, each containing twelve individual cereal boxes
arranged in four by three order. The stacking pattern of the
pallet load is illustrated in Figure 2.

Sampling was done monthly over a period of nine months,
followed by one additional sampling after twelve months
storage. Sampling monthly was designed to sort products in
both liner structures, by using a horizontal split of the
three tiers of product starting from the top of a pallet
load. Each month, three hundred and thirty (330) individual
cereal packages were sampled from each liner structure. Five
(5) packages were used for the quantitative analysis of d-
limonene headspace concentration and the thermal
distillation/desorption of d-limonene from the liners.
Seventy five (75) packages were prepared for the triangle
testing, while the remaining two hundred and twenty five
boxes of cereal (250) were sorted for consumer preference

testing.

26

PERMEATION STUDIES

The permeation studies were carried out by employing a
Quasi-Isostatic method to determine the transmission rate of
d-limonene through the two liner structures. The
transmission rate is defined as that quantity of vapor
passing through a unit area of the parallel surfaces of a
plastic film per unit time, under specified conditions of
test. This procedure is referred to as "Quasi-Isostatic"
because the test compartments are maintained at an
essentially constant total pressure of one atmosphere. It
is, however, an accumulation procedure where permeant
collects, as a function of time.

A schematic diagram of the permeation test apparatus is
presented in Figure 3. A constant concentration of permeant
vapor is produced by bubbling nitrogen gas through the
liquid permeant. This is carried out by assembling a vapor
generator consisting of a gas washing bottle, with a fritted
dispersion tube, containing the organic liquid. Vapor
concentration (ppm) in nitrogen is expressed throughout on a
weight per volume basis. The permeability studies were
carried out at temperatures ranging from 21.1°C (70°F) to
26.7°C (80°F) and approximately 0% relative humidity.

To obtain a low vapor concentration, the permeant vapor
stream is mixed with another stream of pure carrier gas
(nitrogen). Before being directed to the permeation cell,

the vapor stream was passed through a glass reservoir as a

27

 

 

 

- Water Beth

= Gleee Mlxlng Device

= Gleee Vepor Generetor

- Reguletor

- Needle Velve

= Rotemeter

-= Cells (Double Chamber)

- To Waste end Gee Flow Bubble Meter
- Three Way Velve

IIGIU
end

‘4'HCTEP<:I

Figure 3. Schematic of Quasi-Isostatic Permeation Test
Apparatus

28

system was mounted in a constant temperature water bath,
maintained at 1°C above ambient temperature so as to avoid
condensation after the permeant vapor passed through the
glass reservoir. Flow meters were used to provide a
continuous indication that a constant rate of flow was
maintained.

Prior to the initiation of each run the permeation cell
is cleaned to be sure that it is free of limonene vapor.
This is achieved by mounting a sheet of foil in place of a
film sample and allowing the system to set for a period of 2
to 3 days, under closed conditions. After which time, the
headspace of each chamber is measured for traced amounts of
residual limonene which may have leached off from the side
walls of the cell. If any limonene is detected, the cell is
disassembled and heated in a 43°C (110°F) oven for 3 to 4
days and then re-evaluated. This procedure is performed
following each permeation test and repeated until the system
is clean.

The permeability of the respective films was determined
under identical conditions, so as to compare their relative
barrier properties. Duplicate runs on the same film type are
carried out simultaneously in specially designed
permeability cells. Figure 4 provides a detailed view of the
permeant cell system. Each permeability cell, constructed of
stainless steel or aluminum, is comprised of two cell

chambers and a hollow center ring. Both cell chambers and

29

 

 

PV - Permeent Vepor
H - Barometer
V - Needle Velve
o- Sempllng Forte
LC - Len Cell
CC - Center Cell
RC - ngnt Cell
F - To Weete end (he Flew Bubble Meter

Figure 4. Permeant Cell System

30

the center ring are equipped with an inlet and outlet valve
and a sampling port. The cell volume is approximately 50 cc.

In operation, test films are mounted in the permeability
cell so that the center ring effectively isolates the right
and left cell chambers. Hermetic isolation of the chambers
from each other and from the atmosphere is achieved by
compression of overlapping Viton "0" rings (from Detroit
Ball Bearing Company) on the film sample. Viton is a
fluorocarbon elastomer which is resistant to attack and
swelling by most organic vapors. For the permeability cell
with the lower center cavity from the atmosphere was
achieved through compression of the film against a smooth
metal face which resulted in a metal/film/metal seal. A
constant concentration of permeant vapor is then flowed
continuously through the high concentration (center) chamber
of the permeability cell at a flow rate in the range of 25
to 35 cc/min.

The increase in penetrant level in the low concentration
cell chambers is determined by gas chromatographic analysis.
At predetermined time intervals, a 0.5 mL aliquot of
headspace is removed from the low concentration cell
chambers with a gas tight syringe (Hamilton no. 1750, side
port type) and injected directly into the gas chromatograph
for quantitation. A constant total pressure of one
atmosphere is maintained in both the upper and lower cell
chambers by replacing the Smaple volume removed with an

equal volume of pure nitrogen. Samples are removed a number

31

of times over the period of test and an array of time vs.
area response values recorded.

To determine the permeability coefficient, the increase
in penetrant quantity in the lower concentration cell
chambers was plotted as a function of time and the resultant
transmission profile related to the permeability of the film
sample. The time interval during which the permeability data
was evaluated to obtain a steady state rate of transmission
was determined by graphical analysis of the time versus area
response values. In all cases, the data was evaluated
statistically by linear regression analysis to obtain the

best straight line fit.

D-LIMONENE HEADSPACE CONCENTRATION ANALYSIS

Quantitation of d-limonene concentration in the
headspace of a package was carried out by a gas
chromatographic analysis. Prior to headspace analysis, the
cereal boxes were opened and a square shaped gum sheet (1.27
cm. x 1.27 cm. x 0.1 cm.), which had an adhesive surface on
one side, was affixed to the liner structure by using the
adhesive side. This provided a sampling septum for headspace
gas chromatographic analysis. A 0.5 mL aliquot of headspace
was removed from the package headspace with a gas tight
syringe (Hamilton no. 1750, side port type) and injected

directly into a gas chromatograph (GC) for quantitation.

32

syringe (Hamilton no. 1750, side port type) and injected
directly into a gas chromatograph (GC) for quantitation.

A Hewlett Packard Model 5890A gas chromatograph,
equipped with a dual-flame ionization detector (FID), was
used for this analysis. The GC conditions were as follow:

Column: 2-4081 Fused Silica Capillary SUPELCOWAX 10

60 Meters, 0.25 mm. ID, 0.25 um. df (Supelco,
Inc., Bellefonte, PA).
Carrier Gas: nitrogen at 40 psi.
Temperature (°C): injector, 200.
oven, 75.
detector, 275.

Retention Time for D-Limonene (min.): 13.65.

A standard curve of response vs. limonene concentration
was constructed prior to analysis from standard solutions of
known limonene concentration. This allowed the determination
of the linearity and sensitivity of the method. Solutions
used for calibration were prepared by dissolving known
quantities of d-limonene in ethyl acetate. Peak responses
were obtained by injecting the various concentration of d-
1imonene into the gas chromatography. Calibration data are
detailed in Appendix A. D-limonene headspace concentration
levels in the respective cereal box liner structures were
determined for five individual randomly sampled boxes, and

the values averaged.

33

THERMAL DISTILLATION/DESORPTION TEST

The amount of d-limonene sorbed by the respective
package liner structures was determined by the thermal
distillation/desorption method. Following d-limonene
headspace analysis, liners were removed from the package and
cleaned throughly to make sure no cereal residue remained.
Liners were cut into small pieces (2.54 cm. x 2.54 cm.) and
weighed on an analytical balance. The liner pieces were then
added to 65 mL septa seal vials. The filled vials were gas
flushed with nitrogen before sealing with a teflon-lined
silicon septum and aluminum crimp cap. Vials were then
heated for one hour at 90°C. After one hour, the vial was
removed from the oven, a 0.5 mL aliquot of headspace was
taken from the vial with a heated syringe (Hamilton no.
1750, side port type), and injected directly into gas
chromatograph for quantitation.

The amount of d-limonene sorbed by the liners was

determined by reconstitution into the following equation:

Q = s x A x v x 1/vt x 1/wt (1)

where, S = the calibration factor from standard curve

(gm/A.U.).
A = d-limonene area response (A.U.).
V = volume of the vial (mL).

Vt = volume of the injection (mL).

34

Wt = weight of the liner (gm).
Q = the quantity of d-limonene absorbed

(gm of d-limonene/gm of liner).

SENSORY EVALUATION

Sensory evaluation was carried out using two tests, the

triangle test and the preference test.
TRIANGLE TEST

The purpose of using the triangle test was to determine
if a difference existed between the sensory quality of the
product in the two liner structures, as a function of
storage time.

In this test, twenty five (25) voluntary panelists (12
male and 13 female), ranging in age from 24 to 60 years,
were selected on the basis of their previous experience with
the test product and their sensitivity to the changing
concentration of d-limonene in the package. The latter was
established in preliminary screening studies. Panelists were
scheduled every month to participate in the triangle test,
shortly after analytical experiments were completed. The
testing laboratory was located in the School of Packaging.
It contained all the requirements needed for conducting

sensory evaluation tests (Larmond, 1977). Panelists were

35

asked to refrain from smoking, wearing perfume, drinking,
chewing gum, and eating for at least 30 minutes prior to
testing. If a panelist was ill at his/her scheduled time,
the person was allowed to skip the test.

On each day of testing, panelists were individually
shown the testing room, explained the nature and procedure
of the test, and asked to sign a consent form stating their
willingness to participate in the sensory test (Appendix B).
Panelistst were blind-folded to avoid making any decision by
observing a difference in the liners' opacity. A tape
recording of instructions was prepared to standardize the
testing condition. Three coded samples, two identical and
one different, were presented simultaneously to a panelist
with the help of an operator. Control and experimental
treatments were systematically varied so that each was
presented in odd and identical sample positions an equal
number of times. Panelists were asked to identify which of
the three samples presented differed from the other two. A
copy of the triangle test questionnaire developed by Larmond
(1977) and employed in this test is presented in Appendix B.
Analysis of the results was based on the probability that if
there was no detectable difference, the odd sample would be
selected by chance one-third of the time. Tables for rapid
analysis of triangle test data developed by Roessler et al.

(1948) were used in this study to determine whether a

36

significant difference existed between test liners, over

twelve months of storage.

PREFERENCE TEST

A preference test was used to give an indication of
consumer’s preference for the cereal products packaged in
the two liner structures. The preference test was carried
out by two hundred and fifty (250) untrained college
students, ranging in age from 18 to 25 years. Participation
was strictly voluntary.

The test took place in the standard sensory evaluation
laboratory in the Department of Food Science and Human
Nutrition at Michigan State University. The testing area was
equipped with eight (8) individual booths constructed along
a wall which divided the preparation area from the panel
room. Partitions had been built between the booths to
eliminate distraction and prevent communication among the
panelists.

During the test, two samples each containing a different
liner were simultaneously presented to the panelists.
Panelists were asked to evaluate each sample and marked
their preference accordingly on a nine-point hedonic scale
(Appendix B). Descriptive terms ranged from "like extremely"

to ”dislike extremely".

37

Hedonic scale ratings were later converted to numerical
scores ranging from "like extremely" (9) to "dislike
extremely" (1). The mean scores from each treatment were
then analyzed by using the Student's t-test. Results were
analyzed to determine panelists’ degree of liking between

the two package liners.

RESULTS AND DISCUSSION

PERMEATION STUDIES

The results of the studies on the permeation of d-
limonene vapor through the HDPE and the EVOH based liner
structures, are summarized in Tables 4 and 5. Representative
transmission rate profile curves for the respective film
structures are presented graphically in Figures 5 and 6,
where the total quantity permeated is plotted as a function
of time.

Permeation can be described in terms of its component
parts, the diffusion coefficient (D) and the solubility

coefficient (8) by Equation (2).

P=DxS (2)

The diffusion coefficient is a measure of how fast
molecules move in a film, and the solubility coefficient is
a measure of how many molecules can dissolve in a film. As
shown, the permeation behavior of the respective films had,
as predicted by theory, an initial induction period followed
by a non-steady state rate at diffusion, which preceded a
steady state transmission rate. For the low penetrant

concentrations used in this study, it was considered

38

39

Table 4. Permeability Data for the HDPE Based Liner

 

 

Structure
Total Quantity
Run Time of Limonene Pegmeated (a)

(min-) (gm x 10 )

0 0.00

15 0.60

30 1.25
45 3.59

60 28.71

 

Limonene Vapor Concentration (gm/mL, ppm) = 6.6 i 0.2
Lag Time (min.) = 35 i 2

Permeability Coefficient (gm/m2. day. 100 ppm) = 5.86 (a)

(a) Value is the average of duplicate runs.

40

 

onene
)

-8

Total Quantity of D-Li
Permeated (gm x 10

 

Figure 5.

 

 

 

70

Run Time (min.)

Permeation of d-limonene through the HDPE based
liner structure at 23°C and 0% RH.

41

Table 5. Permeability Data for EVOH Based Liner Structure

 

Total Quantity

 

Run Time of Limonene Pgrmeated (a)
(min.) (gm x 10' )
0 0.00
10 0.11
20 0.22
35 0.57
60 2.21
70 2.93
80 3.41
90 3.96
100 4.73
110 5.22

 

Limonene Vapor Concentration (gm/mL, ppm) = 6.6 i 0.2
Lag Time (min.) = 95 i 5

Permeability Coefficient (gm/m2. day. 100 ppm) = 0.34 (a)

(a) Value is the average of duplicate runs.

42

 

)

Total Quantity of D-ligonene
Permeated (gm x 10'

 

Figure 6.

 

 

 

T

l l I I r r r
17 34 51 68 85 102 119 136 153 170

Run Time (min.)

Permeation of d-limonene through the EVOH based
liner structure at 23°C and 0% RH.

43

appropriate to assume that the diffusion process was
Fickian. From the transmission data, the permeability
constant can be calculated by using the standard methods
(Barrer, 1939: Crank and Park, 1968), with a normalization
of the units of the permeability constant to 100 ppm (gm of
permeant/mL of nitrogen) permeant concentration difference
across the film.

As shown in Tables 4 and 5, the averaged permeability
coefficient of the HDPE based liner structure is almost 22
times higher than the permeability coefficient of the EVOH
based liner structure. The lag time and the permeability
coefficient values for the HDPE based liner structure
obtained in this study were found to be in good agreement
with the values reported by Mohney et al. (1987). These
results support the hypothesis that the EVOH liner
structure possesses much better barrier properties with
regard to the retention of d-limonene within the cereal

package, than does the HDPE based liner structure.

D-LIMONENE HEADSPACE CONCENTRATION

The results of monitoring d-limonene headspace
concentration levels in the two cereal package liner
structures over 12 months of storage are summarized in Table
6 and presented graphically in Figures 7 and 8, where

headspace concentration is plotted as a function of storage

44

Table 6. D-limonene headspace concentration levels in two
cereal package liner systems during 12 months
of storage at 23°C and 50% RH.

 

Average d-limonene concentration(a)

(gm/EL. Ppb)
Storage Time

 

(months) HDPE liner EVOH liner
0 - -
2 0.24:0.21 6.04iO.37
3 0.07:0.15 4.2311.12
4 0.04:0.07 3.00:0.16
5 N.D 3.0810.17
6 N.D 4.09:0.10
9 N.D 3.54:0.64
12 N.D 1.45:0.33

 

(a) Values represent the means of five replicates.

(b) N.D means not detectable.

D-Limonene Headspace Concentration (gm/mL, ppb)

45

 

0.454
0.403
0.354
0.30-I
0.25-
0.20-
0.15-
0.10-

0.05‘

 

 

 

 

Figure 7.

 

2 3 4 5 5 7 8 9 10 11 12 13

Storage Time (months)

D-limonene headspace concentration levels in the
HDPE based cereal package liner structure during
12 months of storage at 23°C and 50% RH.

46

 

 

 

 

37

Q.

o.

7; I

86‘

554

«I

4H

m

H

‘5 I

84‘

c:

3 r

83- I}

‘0

a.

m

c

8

:82"

0

g, I

0-

a1

”-4

.4

I

o
llrlrlllrrfir
01 2345673910111213

Storage Time (months)

Figure 8. D-limonene headspace concentration levels in the
EVOH based cereal package liner structure during
12 months of storage at 23°C and 50% RH.

47

time. The headpace concentration levels for the respective
liner structures are superimposed in Figure 9 for
comparison.

At the time the cereal product was received from the
manufacturer, the gas chromatograph was inoperable and
therefore, the initial d-limonene levels were not recorded.
However, headspace analyses were completed from the second
month to the ninth month of storage, followed by one
analysis after the twelfth month of storage. Data for the d-
limonene headspace analyses performed are tabularized in
Appendix C.

As shown in Figure 7, the headspace concentration of d-
limonene in the EVOH liner structure decreased at a fairly
constant rate over the course of the study. However, a rapid
loss of d-limonene in the headspace of the HDPE liner
structure was observed over the first 4 months of storage,
with the concentration of d-limonene being below the level
of detectability for the remainder of the study.

As shown in Table 6, the standard deviation for the
third and fourth months' measure of the amount of d-limonene
remaining in the headspace was found to be higher than the
means for the HDPE based liner. A trace of the storage
location of the cereal boxes used in the third and fourth
months of testing showed that only boxes taken from cases
stored in the middle tier of a pallet, where one or two
sides of the case were exposed to air, had a detectable

amount of d-limonene remaining in the headspace of the

48

 

 

 

A 7
.0
8
. EVOH liner
I-l ..
a 6 I O HDPE liner
E
c 5—
o
w-l
U
n
5 I i
I: 4-
8
5 I
U I
o 3" ’ ii if
0
a
o.
m
“8
o 2‘
a:
2 i
2 1 -
o
a
.,..|
7‘
o 3 A
l I Y

 

 

01-4)

A -
r l l
6

Y

7 8 9 10111213
Storage Time (months)

Figure 9. D-limonene headspace concentration levels in two

cereal package liner structures during 12 months
of storage at 23°C and 50% RH.

49

package. Boxes withdrawn from cases located in the top or
bottom tier of a pallet, where two or more sides of the case
were exposed to air, showed a below detectable amount of d-
limonene in the headspace of the package (Appendix C).

A one-way analysis of variance (ANOVA) was performed to
test the time effect on the loss of d-limonene from the EVOH
based liner structure (Appendix C). It was found that the
storage time had a positive effect on the reduction of d-
limonene headspace concentration in the EVOH based liners
(P < 0.05).

An assessment of the two liners with regard to their
ability to retain d-limonene was made (Appendix C). It was
found that the EVOH liner was a better barrier to flavor
than the HDPE liner (P < 0.05).

D-limonene headspace concentration levels were observed
to be dependent on storage location within the pallet
stacking pattern. A linear regression method was employed to
analyze the relationship between the storage location of
cereal boxes, and the corresponding d-limonene headspace
concentration measured for the cereal packages. It was found
that samples pooled from cases which had two or three sides
exposed to air had lower d-limonene levels in the headspace
than samples from the middle cases, having only one side
exposed to air (Appendix C). However, the results of the
studies were found to be statistically insignificant
(P > 0.05) due to the limited number of paired sample

groups. The observation might still be valuable to explain,

50

in part, the broad range of standard deviations for the d-
limonene concentration levels obtained, particularly for the

HDPE liner.

SORPTION OF D-LIMONENE BY LINERS

The results of analyses to quantify the levels of d-
limonene sorbed by the two liner structures during storage
are presented in Table 7. As shown, the quantity of
d-limonene sorbed by the EVOH liner structure was
significantly higher than that sorbed by the HDPE liner
structure. For better illustration, the results of these
studies are presented graphically in Figures 10, 11, and 12,
where the levels of sorbed limonene are plotted as a
function of storage time. The high level of d-limonene
sorbed by the HDPE liner, shown in the fifth month of
storage, was found to be associated with the storage
location of the boxes in the pallet stacking pattern
(Appendix D).

Figure 13 illustrates an ideal transmission model of
penetrant across a monolayer barrier polymer film. In this
figure, the permeant concentration in the polymer film
progresses from the start of exposure until steady-state.
Shortly after exposure, the film has a high concentration of
permeant near the exposed surface. The permeant has not

however, penetrated very far into the film. With time, the

51

Table 7. Sorption concentration levels of d-limonene by two
cereal package liners during 12 months of storage
at 23°C and 50% RH.

 

Average quantity of d-limonene Sorbed (a)

. (gm/9m liner. ppb)
Storage T1me

 

(months) HDPE liner EVOH liner
0 _ _
2 — _
3 14.39i2.68 110.21i10.99
4 15.59:2.00 90.10il.88
5 22.10i2.53 69.05i9.84
6 9.64:0.81 47.3016.29
9 N.D 16.54i3.46
12 N.D 7.92:1.58

 

(a) Values represent the means of five replicates.

(b) N.D means not detectable.

52

 

 

 

 

 

 

 

A 221
.a

o.

D. 20-
5,7 -
t: 18
«4

r-‘l

g 16“
\.

g 14"
0 12‘
c

2

O 10~
a

or!

'7‘ 3-
O

‘H

O 6‘
5'

.3 4«
4.)

5

o 2 i
0'

o 1

Figure 10.

r %r r ’3
2 3 4 5 6 7 8 9 10 11 12 13

q
-
q
‘
u:
.

Storage Time (months)

Sorption concentration levels of d-limonene by
the HDPE based cereal package liner structure
during 12 months of storage at 23°C and 50% RH.

Quantity of D-Limonene (gm/gm Liner, ppb)

53

 

120
110-
100‘
90'
80‘
70-!
60"
50M
404
30 -‘
20-

10-1

 

 

 

Figure 11.

l l T’ l l l l l l l I
2 3 4 5 6 7 8 9 .10 11 12 13

Storage Time (months)

Sorption concentration levels of d-limonene by
the EVOH based cereal package liner structure
during 12 months of storage at 23°C and 50% RH.

120

54

 

110‘

100‘
90J
80‘
70"I
60q
50‘
40‘
30‘

20-

10‘-

Quantity of D-Limonene (gm/gm Liner, ppb)

 

Figure 12.

   
    

0 EVOH l iner

° HDPE liner

:1

2 3 4 5 6 7 8 9 10 11 12 13

 

Storage Time (months)

Sorption concentration levels of d-limonene by
two cereal package liner structures during 12
months of storage at 23°C and 50% RH.

55

 

Film

1n

steady state

 

 

Concentration of Penetrant
'6

 

 

 

High it' Low
Pressure (-——— ?OSF.i°n ——-—-) Pressure
Surface 1n 1 m Surface

Figure 13. Concentration profile of a penetrant in a polymer
film as a function of time.

DeLassus and Marcus (1986)

56

surface concentration does not change, but the penetration
depth increases. Finally, at steady-state, the concentration
profile is a straight line that drops uniformly across the
film. The area under the curve is proportional to the
quantity absorbed (DeLassus and Marcus, 1986).

Today, a monolayer barrier polymer film is less commonly
used, and a multilayer structure would typically be
fabricated and used in designing packages. In the multilayer
structure a thin barrier layer would be combined with
thicker layers of a less expensive, mechanically tough
polymer. For example, in this study, the 2.6 mil thick EVOH
based liner structure is a multilayer lamination comprised
of HDPE/EVOH/Surlyn, where the EVOH barrier layer (0.5 mil)
is placed between 1.70 mil thick HDPE and 0.4 mil thick
Surlyn blend.

The permeation process for large, organic molecules in
polymer films differs from the permeation process for small
gas molecules in polymer films (DeLassus and Marcus, 1986).
For small gas molecules, the diffusion coefficients are
large, and the solubility coefficients are small. This means
that few molecules are moving in a film: however, they are
moving quickly. For large, organic molecules, like d-
limonene, the diffusion coefficients are small, and the
solubility coefficients are large. This means that many
molecules are moving in a film: however, they are moving
slowly. These differences have an important consequence for

food package design. For good flavor retention, a multilayer

57

barrier container should have the barrier structure as the
inside layer (DeLassus and Marcus, 1986). This is
illustrated by considering the Saran coated oriented
polypropylene laminates described by DeLassus and Marcus
(1986). Such a structure has two 10 mil skin layers of
polypropylene and a 2 mil center layer of Saran copolymer.
If this structure was used to package a product which
contained methyl salicylate, the sorption in the interior
polypropylene layer would be great. This is depicted in
Figure 14 (A). The concentration profile in this layer would
be nearly constant because the activity gradient would tend
to be across the barrier layer in much the same way, as
described by DeLassus et al. (1986), that temperature drop
is across an insulator and voltage drop is across a
resistor. While the permeability coefficient was maintained
the same, switching the barrier layer position in the
laminate would result in the level of sorption to be quite
different. This is shown in Figure 14 (B). As illustrated in
this Figure, a lower permeant concentration gradient would
be expected to be seen in this structure. This means that if
the barrier were placed at the inner layer of the multilayer
structure, the barrier behavior would be enhanced, with
respect to respective losses.

DeLassus and Hilker (1987) further described a model to
explain the mass transfer of permeants such as flavor,
aroma, and solvent molecules across a multilayered

structure, as a function of the location of the barrier

3

Concentration of Methyl
Salicylate, kg/m

58

Polypropylene Saran

 

 

 

 

 

 

50
. '0 . ‘ ’0 J . .
0
Inside Outside
Edge 5’ Edge
(A)
Figure 14.

Saran Polypropylene

 

 

 

 

 

Inside

 

Edge

(3)

Concentration profiles of methyl salicylate at

25°C in two hypothetical multilayer walls at

steady-state.

DeLassus and Marcus (1986)

59

layer. The movement of low molecular weight molecules
depends on the concentration profile of the permeant across
the respective layers of the laminate structure, the
solubility of the permeant in the respective layers of the
lamination, the composition of the barrier layer, the
thickness of each layer of the lamination, and the location
of the barrier layer with respect to the penetrant phase.
The equation can be described as following:

AM (sorb) = 0.5 S x P x L (penetrated) x A (3)

where,lsM is the quantity of permeant that will cross a film
during the time to reach steady-state of permeation: S is
the solubility coefficient, which can be calculated from
Equation (2) by substitution of a known P and a known D:'P
is the permeability coefficient: L is the penetrating depth
of the film when permeation reaches steady-state: and A is
the film area.

The solubility of a permeant in a polymer layer depends
on the chemical affinity of sorbed molecules to the
structure of the layer (Ikegami et al., 1987), the molecular
weight of the permeant (Schimoda et al., 1987), and the
density of the barrier layer (Schimoda et al., 1984).

In this study, where the EVOH barrier layer was placed
as the middle layer of the multilayer structure, a modified

transmission profile model, based on the concept of DeLassus

and Marcus (1986) can be utilized to explain the

60

transmission and sorption of d-limonene through both liner
structures.

D-limonene is well known to have a high chemical
affinity to polyolefins. According to DeLassus and Hilker’s
(1987) model, the flavor compounds in cereal would be sorbed
rapidly by the interior HDPE layer in both liner structures
evaluated in this study. DeLassus (1985) pointed out that
the HDPE reaches a steady-state transmission rate quickly
with the permeant, d-limonene. It takes only 2.6 hours for a
2.0 mil HDPE film to reach steady-state when d-limonene was
selected as the permeant. Even if the thickness was
increased to 20 mil, steady-state would be reached in only
10 days. The same results were observed by Mohney et al.
(1987). The results of permeability studies have shown that
the HDPE based liner is a poor barrier to d-limonene (Tables
4 and 5). Therefore, for the HDPE liner structure, it is
anticipated that a high percentage of the d-limonene within
the cereal package would be lost from the system within
several months of storage, due to the permeation mechanism.
This is demonstrated by the low concentration levels of d-
limonene remaining in the headspace of the packages, and the
low quantity of d-limonene sorbed by the HDPE based liner
structure, over the storage time of the study. This model is
shown in Figure 15 (B). In contrast, the high level of d-
limonene concentration detected in the headspace of the EVOH
based cereal packages, and the high amount of d-limonene

sorbed by the EVOH based liner structure indicated that the

 

 

 

 

 

 

 

0 HDPE EVOH Surlyn
c
0
c
o
E
0d
A
l . ' L
a 1..
u :2
o -
c ‘-
o .
.H \o
u .‘ ~
6 . ,
u D.
u _.
c .
0 - .
o s o e
c ' ' 0:.
o .1
0
Inside Outside
Edge Edge
(A)
Figure 15.

61

HDPE

 

EVA

 

 

 

 

Inside Outside
Edge Edge
(B)

Hypothetical concentration profile of d-limonene

at 25°C in two cereal liner structures at steady
state of transmission.

62

transmission rate of d-limonene through the EVOH based
structure was lowered due to the EVOH barrier layer. The
model proposed to account for the results of headspace
ananlyses and sorption studies involving the EVOH liner

structure is illustrated in Figure 15 (A).

TRIANGLE TEST

The triangle test was used to determine if panelists
could sensorially identify differences in aroma intensity
between the two package liner systems as a function of
storage time. These results are summarized in Table 8.

It was found that panelists could identify differences
in d-limonene concentration intensity between the two cereal
package liner structures at the confidence level of P <
0.05.

Months 5 and 6 in the triangle design seemed to be less
significant (P < 0.05) than months 2, 3, and 4 (P < 0.001).
This was probably due in part to panelists' fatigue during
the last two testing months. Verification of this could be
observed by the gradual shortening of testing time taken by
panelists, by panelists' complaints, and by the truancy of

panelists from their arranged testing time.

63

 

 

Table 8. Triangle test of cereal packaged in two liner
systems during 12 months of storage at 23°C and
50% RH.
Storage Number of Panelists
Time Identifying difference Degree of dgfference
(months) (NC/Ht) a (D)
0 - .—
2 19/25 ** moderate (2.316)
3 17/25 ** much (2.706)
4 11/14 ** moderate (2.545)
5 ll/18 * moderate (2.545)
6 12/20 * much (2.750)
9 11/20 * much (2.681)
12 13/25 * much (2.750)

 

(a)
sample correctly and Nt
(b) D is the mean which was
the number of judges in
a numerical value given
2=moderate, 3=much, and
the summation by Nc‘

5% significant level.

** 0.1% significant level.

Nc is the number of judges who identified the odd

is the total number of judges.

obtained by multiplying
each degree of difference with
to that level (l=slight,
4=extremely) and then dividing

64

CONSUMER PREFERENCE

Consumer preference was quantitatively measured and the
results are presented in Table 9. Data showed that cereal
packaged in the EVOH liner system was rated more acceptable
than the same product packaged in the HDPE liner system
(P < 0.001), even after 6 months of storage. This was due to
the consumers' perception of a much higher aroma intensity
in the EVOH liner packages than in the HDPE liner packages
(P < 0.001). Cereal in an EVOH liner package was commonly
rated by consumers in the "like moderately" category
compared to the "like slightly" category for the product in
the HDPE liner package.

It was found in the quantitation of consumer preference,
that the degree of liking had a positive increasing
relationship with the storage time (P < 0.01). For example,
consumers felt that a longer storage time produced a more
acceptable product. This information gives an important
indication of the flavor quality of the product over the
storage stability test.

Further analyses were conducted to locate the source of
difference in the sensory preference. It was found that only
the EVOH liner structure showed a negative relationship with
storage time. Differences shown in the acceptability of
consumer preference for products in the EVOH based liner
structure were mainly due to consumers’ changing attitudes

over the storage time. The degree of acceptability of cereal

65

 

 

an

hm.m H u

 

 

Lev m4.m u moo: omaxsm u use: omfl\m~ u mean
101 30.6 n zo>m o~H\mm u mo>m Am~.mv zoos oma\mm u mo>m NH
4 mo.m u 0
Adv om.m u mean msa\4o I wear msa\>¢ u use:
Loo as.» u zo>m msa\moa n mo>m Ama.mc zoos msa\oma u =o>m m
4 4m.m n 0
Ace 4s.m u man: 4o\m~ u mean 4m\oH u man:
161 so.» u mo>m am\ao n mo>m AmH.mc zoos sexes u mo>m o
4 mo.4 u o
141 mm.o I won: maflxmm u use: maa\4a u use:
16v mH.s u ro>m mHH\om u mo>m Adm.mc zoos maa\moa u mo>m m
4 mo.m n 0
Adv 4m.m I wear moa\mm n meow moa\s~ n wear
16V m6.o n mo>m moaxoaa n mo>m 1mm.mc Boos moa\onfl u mo>m m
4 as.m n u
Lac ow.m u mean moa\mm u mac: ~ma\m~ u meow
16V s4.o n mo>m mma\mma u mo>m Ava.mc zoos moa\¢oa n mo>m m
w w Age 101 Roz\ozc M w Ace 165 Auz\0zc Amrocoec
o m Douche w dummmwo >aumcmucH mafia
mo monmmo >uflaflncumooo< uo omumwo a“ Homconum omcnoum
ocnseaocooH
MHmfldwcmm
uo Honezz
.mm won one 00mm pm omcmoum mo mSHCOE NH dawnso
msoumxm Hosea owmxomm 03» :fl pommxocm Hmmuoo Lou mocmumwoum hmesmcou .m dance

66

.H0>0H ucoowmwcmfim waoo.o

.:>Hu:mwau 0xflH= muc0m0um0u u0nasz
.:>H0ucu0008 0xflH= muc0m0ummn n0nezz

.>H0>w#o0mm0u >H0E0nux0

0xfiamfloua can >H0E0uux0 0xflaum .0Hcom oficoo0: ucflomlm 0:» :H .H0>0H
06:» on c0>flm 09Hm> Hmofiu0asc c :ufiz 00:0H0u0um no H0>0H :000 Cw m0mosm
Co M0285: 0:» mcflwaafluHaa h: U0cflcuno 003 :0a:3 c008 0:» muc0m0um0n q

.um0u :000 :a p0ummflofluucm 0:3 mumwa0ccm no H0nasc Hmuou 0:9 ms #2
.Ho0u0o 0:» 000:0H3Q Uazo3 >0:u #0:» uH0m 0:3 m0mpSn mo HOQED: 0:u 0H 02

.A>H0E0Mux0u¢ 6:0 .:0:Eum .0udu0ooe Hm .u:mflamuav H0>0H
06:» on 20>flm 09Hm> Hoowu0asc m :uflz 00:0H0wuflp mo H0>0H :000 :w m0qU3n
mo b0nss: 0:u mcfl>amwuasa x: o0:wcu:o 003 :ofl:3 :008 0:: muc0m0na0u o

.um0u :000 :w U0ucmfi0fiunom 0:3 mumHH0CMQ mo HOQES: chou 0:u
mH bz .>uflmc0u:« 080nm u0mcouum p0fimwuc0ofl 0:3 0006:“ ac 90:55: 0:» ma 02

om
Adv
16v

161

on

Any

Any

16.06001 .6 dance

67

packaged in the HDPE liner structure was seen to remain the
same. This can be interpreted as peOple preferred the
product packaged in the EVOH liner structure much more than

that packaged in the HDPE liner structure.

CONCLUSION

The relative performance of two typical liner structures

of a fruit-flavored cereal product, whose quality is

associated with the retention of volatile aroma

constituents, had been assessed by both qualitative (i.e.

triangle test and consumer preference survey) and

quantitative methods (i.e. d-limonene headspace

concentration analysis and thermal distillation sorption

test).

The results are summarized as follows:

(1)

(2)

(3)

The d-limonene headspace concentration was found to
be strongly time dependent. The EVOH liner

structure had better barrier properties with

respect to retaining aroma components within the

packages, than the HDPE liner structure over the
storage time considered.

The amounts of d-limonene sorbed by the EVOH liner
structure was significantly higher than by the HDPE
liner structure over time. This was felt to be due
to the differences in transmission rate of
d-limonene in these two laminate structures.

The sensorial difference between product packaged
in the EVOH liner structure and the HDPE liner

structure increased over time. However, it is felt

68

69

that, panelists' fatigue may be regarded in part
for the differences to be less significant for the
last two months' measurements.

(4) Consumers viewed the product packaged in the EVOH
liner structure to be more acceptable than the same
product packaged in the HDPE liner structure.
Sensory responses showed that the degree of
preference for only the EVOH lined packages
increased over time. This indicated that consumers’
preference was based on the intensity of the aroma
constituents within the package.

The results from both the analytical measurements and
the sensory evaluation support the hypothesis that the EVOH
package liner structure is a better barrier than the HDPE
liner structure in retaining the flavor quality of the fruit

flavored cereal product.

APPENDIX

APPENDIX A

Standard Calibration

70

Table 10. Limonene/Ethyl Acetate Standard Calibration Curve
Data as Measured by HP Gas Chromatograph Model
#5890.

Initial Calibration: 8/10/1987

 

 

Area Response Absolute Quagtity
(A.U.) (gm x 10' )
0 0.00
75094 16.83
154500 33.65
264050 50.48
320620 67.30

 

Calibration Factor = 2.02 x 10'13 gm/A.U.

Recalibration: 10/17/1987

 

 

Area Response Absolute Quagtity

(A.U.) (gm x 10 )
0 0.00
49929 16.83
86731 33.65
161300 50.48
326020 67.30
435210 84.13

 

Calibration Factor = 1.98 x 10'13 gm/A.U.

71

 

 

 

 

 

90
A Initial Calibration A
To 757 ‘ Recalibration
H
x
g, 60-
>4
4)
'3 J
c 45
c
5
<3
3 30—
a
H
o
0
fl .5.
l l l T T
0 8 16 24 32 40 48
Area Response (A.U. x 10")

Figure 16. Limonene /Ethyl Acetate Standard Calibration-

Curve as Measured by HP Gas Chromatograph
Model #5890.

APPENDIX B

consent torn For Sensory Panel nenbere
Triangle Test Questioneire

Consumer Preference Queetioneire

72

CONSENT PORN POE SENSORY PANEL MEMBERS

School of Packaging
Michigan State University

Fruit Flavored Cereal Ingredients:

Rice, sugar, hydrogenated coconut and/or palm oil, corn
syrup, salt, artifical color (includes yellow #5),
natural and artifical fruit flavors.

The nature of the proposed study involves panelists
comparing the sensory preference of a fruit flavored cereal
product packaged in two commercial cereal liner systems as a
function of storage time.

The product is a fruit flavored cereal product with
predominant citrus flavoring which includes d-limonene as a
major componene. D-limonene is a naturally occuring
constituent of cold pressed orange and lemon oil which are
used to provide the desired product aroma. Panelists will be
asked to identify the presence of fruity flavor which
includes d-limonene in the headspace of the cereal package
as a function of storage time.

I have read the above list of
ingredients and find none that I know I am allergic to. I
have also been informed of the nature of the study and the
procedures which will be used during the sensory analysis.

I agree to serve. on 'this survey’ panel, which, is being
conducted on this day of 1988.

I understand that my anonymity will be protected during the
reporting of results and that I am free to withdraw my
consent and to discontinue participation in the panel at any
time without panelity.

I understand that if I am injured as a result of my
participation in this research project, Michigan State
University will provide emergency medical care if necessary,
but these and other medical expenses must be paid by my own
health insurance program.

 

Signature

 

Date

73

 

 

 

NAME DATE

PRODUCT

W:

1. Two of these three cereal samples are identical, the

third is different. In the order indicated, open one box
of cereal at a time by tearing the liner. Then smell the
sample. Identify the odd sample based on perceived aroma.

999$ §h§£L_QQQ_§§mPl§

 

 

 

 

 

 

Indicate the degree of difference between the duplicate
samples and the odd sample.

Slight
Moderate
Much
Extreme

 

Acceptability:

Odd sample more acceptable
Duplicates more acceptable

 

Check the degree of acceptability for the sample you have
chosen in question 3.

like extremely

like very much

like moderately

like slightly

neither like nor dislike

 

 

 

 

 

Comments:

Please describe the flavor and aroma you perceived and
other factors which helped you to make your decision. If
you have any suggestions for better evaluating the
sensory qualities of the product, we welcome your
comments. Thank you very much and enjoy your cereal!

74

 

 

 

NAME DATE
PRODUCT
1. Evaluating the aroma in the cereal packages by opening

the box of cereal on the left first, tearing the liner
and smelling the sample. Check below which sample has the
stronger aroma. (Do not taste samples!)

9.991s W

 

 

 

 

. Check the degree of difference in odor between these

samples.

Slight
Moderate
Much
Extreme

. Which of these samples do you feel more acceptable?

 

 

 

 

. Check how much you like or dislike each one.

 

 

like extremely

like very much

like moderately

like slightly

neither like nor dislike
dislike slightly

 

 

 

 

 

 

 

 

 

 

 

 

dislike moderately
dislike very much

 

 

 

 

dislike extremely

 

 

Comments:

Please describe the flavor you perceived and other
factors which helped you to make your decision. If you
have any suggestions for better evaluating the sensory
qualities of the product, we welcome your comments.
Thank you very much and enjoy your cereal!

APPENDIX C

Headspace Analyses Data

75

Table 11. Data points of d-limonene headspace concentration
analyses (area response obtained from SC reading).

Storage Time(b) D-limonene Area Response/Sample Code(a)

 

(months) (A.U./#)
HDPE liner EVOH liner
2 618 (#132) 14627 (#359)
1160 (#367) 14372 (#232)
1182 (#167) 14146 (#128)
0 (#007) 14875 (#058)
O (#047) 16738 (#192)
Mean= 592 14952
Standard Deviation= 524 926
3 0 (#370) 14416 (#395)
0 (#365) 12320 (#214)
907 (#251) 10881 (#048)
0 (#326) 7932 (#036)
0 (#084) 6845 (#003)
Mean= 182 10479
Standard Deviation= 363 2784
4 0 (#314) 7298 (#308)
446 (#300) 7809 (#302)
0 (#358) 7806 (#396)
0 (#212) 7451 (#304)
0 (#381) 6766 (#321)
Mean= 89 7426
Standard Deviation= 183 386
5 0 (#220) 8031 (#062)
0 (#233) 8003 (#144)
0 (#057) 7074 (#242)
0 (#018) 7190 (#337)
0 (#251) 7868 (#223)
Mean= 0 7633
Standard Deviation= 0 415
6 0 (#102) 10463 (#346)
0 (#260) 9730 (#113)
0 (#041) 10048 (#290)
0 (#063) 10077 (#280)
0 (#123) 10304 (#321)
Mean= 0 10124
Standard Deviation= 0 249

Table 11. (continued)

76

 

Storage Time D-limonene Area Response/Sample Code
(months) (A.U./#)

HDPE liner IIEVOH line;

9 0 (#211) 11536 (#285)

0 (#006) 7317 (#341)

0 (#355) 8499 (#105)

0 (#249) 9284 (#117)

0 (#186) 7186 (#269)
Mean= O 8764
Standard Deviation= 0 1588

12 0 (#057) 5185 (#097)

0 (#012) 3312 (#257)

0 (#370) 3328 (#286)

0 (#224) 3103 (#275)

0 (#232) 2992 (#145)
Mean= 0 3584
Standard Deviation= 0 811

(a) The code was generated from the random sampling plan.
The hundredth number represents the tier number where
the sample was withdrawn. For example, 231 means that
the sample was pooled from second tier of the pallet-

load.

(b) Samples were pooled from the top three tiers of the
palletload at 2, 4, 6 months and from bottom three tiers

at 3, 5, 9 months.

77

Table 12. Data points of d-limonene headspace concentration
analyses (d-limonene headspace concentration a ).

Storage Time D-limonene Headspace Concentration

(months) (gm/mL, PPb)
HDPE line; EVOH line;

 

2 0.250 5.909

0.469 5.806

0.478 5.715

0 6.010

0 6.762

Mean= 0.239 6.040
Standard Deviation= 0.212 0.374
3 0 5.824

0 4.977

0.366 4.396

0 3.205

0 2.765

Mean= 0.073 4.233
Standard Deviation= 0.146 1.124
4 0 2.948

0.180 3.155

0 3.154

0 3.010

0 2.733

Mean= 0.036 3.000
Standard Deviation= 0.072 0.156
5 0 3.245

0 3.233

0 2.858

0 2.905

0 3.179

Mean= 0 3.084
Standard Deviation= 0 0.167
6 0 4.227

0 3.931

0 4.059

0 4.071

0 4.163

Mean= 0 4.090
Standard Deviation= 0 0.101

78

Table 12 (continued)

Storage Time D-limonene Headspace Concentration
(months) (gm/mL, ppb)
HDPE liner—___mn_liner

 

4.661
2.956
3.434
3.751
2.903
3.541
0.642

Mean=
Standard Deviation=

0000000

12 2.095
1.338
1.345
1.254
1.209
1.448

0.327

Mean=
Standard Deviation=

0000000

(a) Data obtained from multiplying the calibration factor
(0.404 x 10-12 gm/a.u./mL) of standard solutions
(d-limonene/ethyl acetate) by the area response from
each GC injection.

79

Table 13. ANOVA Table for Headspace Analysis on Storage
Time (Independent Variable) Versus D-Limonene
Headspace Concentration in EVOH Liner System
(Dependent Variable).

 

 

Source of Variation SS df MS F value
Treatments 59.21 5 11.84 35.88 *
Error 7.85 24 0.33

 

*
F1-.01(5,24)= 3°90-

APPENDIX D

sorption Analyses Data

Table 14.
test.

Storage Time

80

Data points of the thermal distillation sorption

Quantity of D-limonene Sorbed

 

(months) (gm/gm liner, ppb)
HDPE liner EVOH liner

3 13.86 101.51

12.65 98.82

10.98 103.66

15.66 124.25

18.80 122.79

Mean= 14.39 110.21
Standard Deviation= 2.68 10.99
4 18.21 92.10

13.06 88.40

14.09 87.34

14.98 91.09

17.63 91.57

Mean= 15.59 90.10
Standard Deviation= 2.00 1.88
5 25.21 55.92

17.63 81.07

21.54 75.35

22.83 74.01

23.27 58.91

Mean= 22.10 69.05
Standard Deviation= 2.53 9.84
6 8.78 42.50

10.50 52.11

9.22 57.15

8.97 40.68

10.71 44.06

Mean= 9.64 47.30
Standard Deviation= 0.81 6.29
9 N.D. (a) 18.71

N.D. 20.35

N.D. 14.15

N.D. 10.98

N.D. 18.51

Mean= N.D. 16.54
Standard Deviation= N.D. 3.46

81

Table 14. (continued)

Storage Time Quantity of D-limonene Sorbed

 

(months) (gm/gm liner, ppb)
HDPE liner EVOH liner

12 N.D. 10.11

N.D. 8.68

N.D. 5.37

N.D. 7.22

N.D. 8.22

Mean= N.D. 7.92
Standard Deviation= N.D. 1.58

(a) N.D. means not detectable.

BIBLIOGRAPHY

BIBLIOGRAPHY

Ahmed, E.M., Dennison, R.A., and Shaw, P.E. 1978a. Effect
of Selected Oil and Essence Volatile Components on
Flavor Quality of Pumpout Orange Juice. £1.89I121_E29§
thme, 26:368.

Ahmed, E.M., Dennison, R.A., Dougherty, R.H., and Shaw, P.E.
1978b. Flavor and Ordor Thresholds in Water of Selected

Orange Juice Components. 11.59Ii£1_£QQ§_Qh§mi. 26:187.
Amerine, M.A., Pangborne, R.M., and Rosseler, E.B. 1965.

”WW" . Academic
Press, New York, NY.
Anandaraman, S. and Reineccius, G.A. 1986. Stability of
Encapsulated Orange Peel Oil. £298.12280211. 40(11):88.
ASTM. 1984. D1434-82. Determining Gas Permeability
Characteristics of Plastic Film and Sheeting. Anngel

fieek_ef_5§1n_§tenge:g§, Section 8: Plastics, 08.01:632.
ASTM. 1984. E96-80. Water Vapor Transmission of Materials.

Anngel_§eek_ej_5§zu_§tengezge, Section 8: Plastics,
08.03:599.

Baner, A.L., Hernandez, R.J., Jayaraman, K., and Giacin,
J.R. 1986. Isostatic and Quasi-Isostatic Methods for
Determining the Permeability of Organic Vapors Through
Barrier Membrances. To be Published in Current

5 ’ S S .

82

83

Brown, W.R., Jenkins, R.B., and Park, G.S. 1973. J, £12m,
W. 41:45-

Crompton, T.R. 1979. "Additive Migration from Plastics into
Foods", Pergammon Press, Oxford.

Crosby, N.T. 1981. "Food Packaging Materials Aspects of
Analysis and Migration of Contaminants", Applied
Science Publishers Ltd., London.

DeLassus, P.T. 1985. Transport of Unusual Molecules in
Polymer Film. I c n
W: 445.

DeLassus, P.T. and Marcus, S.A. 1986. Diffusion and
Sorption of Organic Molecules by Polymer Films. In
”Activities Report of the R a D Associates",
Proceedings of the Fall 1986 Meeting, Research and
Development Associates for Military Food and Packaging
Systems, Inc., San Antonio, TX.

DeLassus, P.T. and Hilker, B.L. 1987. Interaction of High
Barrier Plastics with Food: Permeation and Sorption.
In "WWW”.
Gray, J.I., Harte, B.R., and Miltz, J. (Eds.),
Technomic Publ. Co., Lancaster, PA.

Dethmers, A.E. 1979. Utilizing Sensory Evaluation to
Determine Product Shelf Life. Eeeg_1eehnele, 33:40.

Durr, P., Schobinger, U., and Waldvogel, R. 1981. Aroma
Quality of Orange Juice After Filling and Storage in
Soft Packages and Glass Bottles. Lebenemittel;

W, 20:91. 773/7 -2 c/

\0

A » ‘ . ' I)
_/ S j J LI, ‘4/ .‘J

.5 ‘J

I
/

v

84

Farmer, E.H. and Alvapillai, S. 1942. The Course of
Autoxidation Reactions in Polyisoprenes and Allied
Compounds. I. The Structure and Reactive Tendencies of
the Peroxides of Simple Olefins. J, Chem, 809,, 121.

Gacula, M.C. and Kubala, J.J. 1975. Statistical Models for

Shelf-Life Failures. J4_£Qed_§eienee, 40:404.
Gianturco, M.A., Biggers, R.E., and Ridling, B.H. 1974.

£1_Agrisi_rggd_§hem1, 22:758.

Gilbert, S.G. and Mannheim, C.H. 1982. "Shelf Life Testing
of Foods", P. A1. Center for Professional Advancement,
E . Brunswick, NJ .

Gilbert, S.G., Hatzidimitriu, E., Lai, C., and Passy, N.

1983. Studies on Barrier Properties of Polymeric Films

to Various Organic Vapors. 1ne§rumente1_bnelx§1§_efi
2999, 1:405.

Gilbert, S.G. 1985. Food/Package Compatibility. Feed
Techngli. 39(12):54.

Granzer, R. 1982. Yerpeekgng;ggng§hen, 33:35.

Guadagni, D.G. 1957. Time-Temperature Tolerance of Frozen
Foods: Organoleptic Evaluation of Frozen Strawberries,
Raspberries and Peaches. zeeg_1eennele, 11:471.

Harte, B.R. and Gray, J.I. 1987. The Influence of Packaging

on Product Quality- In ӣ229.212822t2232kage

QQuestibilitx_£rgseeding§”. Technomic. Lancaster. PA-
Hatzidimitriu, E., Gilbert, S.G., and Loukakis, G. 1987.

Odor Barrier Properties of Multi-Layer Packaging Films
at Different Relative Humidities. JL_Eeed_§eienee,

85

52(2):472.

Heath, H.B. and Reineccios, G. 1986. "Elexe;_gnemi§tzy_eng
Teehnelegy", AVI Publishing Co., Westport, CT.

Hernandez, R.J. 1984. Permeation of Toluene Vapor Through
Glassy Poly(ethylene) Terephalate Films. M.S. Thesis,
Michigan State University, East Lansing, MI 48823.

Hernandez, R.J., Giacin, J.R., and Baner, A.L. 1986.
Measuring the Aroma Barrier Properties of Polymeric
Packaging Materials. geekeging_1eehnele, l6(4):12.

Horton, H.W. 1987. Consumer Reaction to Flavor Innovations.
Eood_Tschn911. 41(6):80.

Ikegami, T., Shimoda, M., Koyama, M., and Osajima, Y. 1987.
Sorption of Volatile Compounds by Plastic Polymers for
Food Packaging- Eipn9n_Snoknnin_zogxo_§akkaisni.
34(5):267.

Johnson, J.D. and Vora, J.D. 1983. Natural Citrus Essences.
EQQQ_IQQBBQLI. 37(12)=92-

Jul, M. 1969. Quality and Stability of Frozen Meats. In
" t S s", Van Arsdel,
W.B., Copley, M.J., and Olson, R.L. (Eds.), Wiley
Interscience, New York, NY. .

Karel, M., Issenberg, P., Ronsivall, L., and Urin, V. 1963.
Application of Gas Chromatography to the Measurement of
Gas Permeability of Packaging Materials. Eeeg_1eennele,
17(3):91.

Karel, M. and Heldelburgh, N.D. 1975. Effect of Packaging
on Nutrients. In "Nntritional_Exaluation_of_£oo§

86

Exeeeeeing", Harris, R.S. and Karmas, E. (Eds.), 2nd
ed., p.412, AVI Publishing Co., Westport, CT.
Kashtock, M.E., Giacin, J.R., and Gilbert, S.G. 1980.
Migration of Indirect Food Additives: A Physical
Chemical Approach. g;_£eeg_§eienee, 45:1008.
Kesterson, J.W., Hendrickson, R., and Braddock, R.J. 1971.
Florida Citrus Oils. v o .

2311;, NO. 748.
Kirchner, J.G. and Miller, J.M. 1953. Volatile Oil

Constituents of Grapefruit Juice. l1_Ag£iQL_£QQQ_Qh§m1.
1:512.
Kreituss, A. 1981. W. 19:889-

Kwapong, O.Y. and Hotchkiss, J.H. 1987. Comparative
Sorption of Aroma Compounds by Polyethylene and Ionomer

Food-Contact Plastics. J&_£eeg_§eienee, 52(3):761.
Labuza, T.P. 1981. Prediction of Moisture Protection

Requirements for Foods. Qe;eel_£eede_flerlg, 26(7):335.

Larmond, E. 1977. ”Laboratory Methods for Sensory
Evaluation of Food", rev. ed. Pub. 1284, Food Res.
Inst. Canada Dept. of Agriculture, Ottawa, Canada.

Lasoski, S.W. 1960. Moisture Permeability of Polymers.
Effect of Symmetry of Polymer Structure. 1L_Apple
£212mer.§212n29. 4:118-

Maarse, H. 1984. Volatile Compounds in Foods. tat
nete, Vols. 1-3, Division for Nutrition and Food

Research, TNO, Zeist Netherlands.

87

Mannheim, C.H., Miltz, J., and Letzter, A. 1987.
Interaction Between Polyethylene Laminated Cartons and
Aseptically Packed Citrus Juices. g, Eeod Selence,
52(3):737.

Marshall, M.R., Adams, J.P., and Williams, J.W. 1985.
"Flavor Absorption by Aseptic Packaging Materials",

Proseeding§_A§entinak_L§§. Princeton. NJ.
Masi, P. and Paul, D.R. 1982. Modeling Gas Transport in

Packaging Applications. Je_nemb;enee_§eienee, 12:137.
k/Mohney, S.M. 1986. The Aroma Permeability and Solubility
of Two Cereal Liner Materials and Their Relationship
to Product Quality. M.S. Thesis, Michigan State
University, East Lansing, MI 48824.
Moshonas, M.G. and Shaw, P.E. 1984. Direct Gas Chromato-
graphic Analysis of Aqueous Citrus and Other Fruit

Essences. 11_Agrisi_zood_9hemi. 32:526-

Moshonas, M.G. and Shaw, P.E. 1986. Orange Juice Flavor and
Aroma Analyses. 21.59I191_£QQQ_§DQEI. in press.

Moskowitz, H.R. and Chandler, J.W. 1978. Consumer
Perceptions, Attitudes, and Trade-Offs Regarding Flavor
and Other Product Characteristics. Eeeg_meehnelegy,
32(11):34.

Murray, L.J. and Dorschner, R.W. 1983. Permeation Speeds
Tests, Aids Choice of Exact Material. £eekege_£ngineer,
28(3):76.

Murray, L.J. 1985. An Organic Vapor Permeation Rate

Determination for Flexible Packaging Materials. J;

88

Iunguarajfilntand_§heeting. 1:104-
Nagy, S. and Klim, M. 1986. Gas Chromatography—Mass

Spectral Techniques in Citrus Essence Research. [egg
Technology, 40(11):95.

Ng, H.C., Leubung, w.p., and Choy, C.L. 1985. 11.2212mer
£91, 20123. BREE: zg., 23:973.

Peryam, D.R. 1963. The Acceptance of Novel Foods. Feed
Ieehnelegy, 17:711.

Peterlin. A. 1975. J1_Hasronoli_§oii_£hxsi. 811:5?-
Prell, P.A. 1976. Preparation of Reports and Manuscripts

Which Include Sensory Evaluation Data. Eeeg_1eennelegy,
30(11):40.

Quast, D.G. and Karel, M. 1972. Effects of Environmental
Factors on the Oxidation of Potato Chips. J;_£eeg
figignge, 37:584.

Rijkens, F. and Boelens, H. 1975. The Future of Aroma
Research- In ”2r221_Insti_§¥mni_hroma_ze§earsh". ed.
Maarse, H. and Groenen, P.J., p. 203, Production
Wagenengen, Netherlands.

Roesslor, E.B., Warren, J., and Guymon, J.F. 1948.
Significance in Trianglar Tests. £ee§_3e§&, 13:503.

Salame, M. and Temple, E.J. 1974. High Nitrile Copolymers
for Food and Beverage Packaging. Ch. 6, In "Qhemietzy

ood n ", ed. Swalm, C.H., ACS Adv. Chem.
Series, 135, American Chemical Society.

Scheffe, H.A. 1959. The_Ane1ye1e_efi_yerienee. John Wiley &

Sons, New York, NY.

89

Schimoda, M., Nitanad, T., Kadota, N., Ohta, H., Suetsana,
K., and Osajima, Y. 1984. Adsorption of Satsuma
Mandarin Orange Juice Aroma on Plastic Films. 1 on
§DQEEQLB_EQQYQ_§§E£§1§hi7 31:697-

Schimoda, M., Matsui, T., and Osajima, Y. 1987. Effects of
the Number of Carbon Atoms of Flavor Compounds on
Diffusion, Permeation, and Sorption with Polyethylene
FilmS- Ni222n_Shokgnin_xogxo_§akkai§hi. 34(8).535-

Schreier, P. 1981. Changes of Flavor Compounds During the
Processing of Fruit Juices. Ereee_Leeg_Aenten_§ymee,
7:355. -

Schutz. M.G. and Wahl, O.L. 1981. Consumer Perception of
the Relative Importance of Appearance, Flavor, and
Texture to Food Acceptance. In ”QIiLiQQl_QI_EQQQ
Aceeptance", ed. Solms, J. and Hall, R.L., 97, Forster-
Verlag, Zurich.

Schutz, M.G., Judge, D.S., and Gentry, J. 1986. The
Importance of Nutrition, Brand, Cost, and Sensory
Attributes to Food Purchase and Consumption. Feed
Teehnelegy, 40(11):79.

Shaw, P.E. 1977a. Essential Oils. In "Qit;ee_§eieeee_ene
Ieehnelegy", eds. Nagy, 3., Shaw, P.E., and Veldhuis,
M.K., V01. 1, Ch. 11, p. 427, AVI publishing Co.,
Westport, CT.

Shaw, P.E. 1977b. Aqueous Essence. In "Qitre§_§eienee_ene
Teennelegy", eds. Nagy, 8., Shaw, P.E., and Veldhuis,

M.K., Vol. 1, Ch. 12, p. 463, AVI Publishing Co.,

90

Westport, CT.

Shaw, P.E. 1979. Review of Quantitative Analysis of Citrus
Essential Oils. l1_Ag:11_EQQQ_Qh§m1, 27:246.

Szczesniok, A.S. and Kleyn, D.H. 1963. Consumer Awareness
of Texture and Other Food Attributes. o chn ,
17:74.

Van Arsdel, W.B. 1969. Estimating Quality Change From A
Known Temperature History. In "Qeelity_ene_§;eeili§1_efi
£zeeen_£eee", eds. Van Arsedel, W.B., Copley, M.J., and
Olson, R.L., Wiley Interscience, New York, NY.

Wartenberg, E.W. 1982. yerpeekeeg;geeeenee, 33:58.

Weast, R.C., Astle, M.J., and Beyer, W.H. 1985. CRC
Handbook of Chemistry and Physics. 66th Edition, CRC
Press, Inc., Boca Ration, FL.

Whitaker, R.J. and Evans, D.A. 1987. Plant Biotechnology
and the Production of Flavor Compounds. Eood_Te§hn211.
41(9):86.

Zobel, M.G.R. 1982. Measurement of Odour Permeability of
Polypropylene Packaging Films at Low Odourant Levels.
221mm. 3:133-

Zobel, M.G.R. 1984. Aroma Barrier Properties of Coextruded
Films. MW

o;.- -1 - 1 . 11°93 5-1 0! --. V s'-: 4. .- ; .,.
. Technolegy, September 19-21, Princeton, NJ.

Zobel, M.G.R. 1985. The Odour Permeability of Polypropylene
Packaging Film. £e1yme;_1eeting, 5(2):153.

infighrfiulln w

 

nICHIGaN STATE UNIV. LIBRQRIES
ll11|“WW[Ill”IOWl1“IWI‘IWIHIIWIWHI
31293005510817