. .- “.l-
V0.1.

:- ; gm: :1 .-.,..~7

 

 

 

 
     
         
     
     
    

  

. .I'.
f“; If.{
e~'_'io.‘f:;’:E

31‘
xEAfi‘,I

,Al
.3.

'ég.
. m...

rte.

.,.

,3
33.51:- {m‘
“1.7..“ .

     

  
 

     

‘6‘.

  
 
 
 
 
    

  
    
 
   
 

I?“ I.
E git
I Ii: IIIIII‘IMH
.,. III-II" iii‘
DI IIIII iI‘II II‘
’ .i 'I' III

. In;
14.94%;
39’ .
M

1.1"”

SI ‘1: 'r-
ifIII'I
:E r'I

' ..ul‘:
flv'

“iii?“
'- 33?;
\_ “"3““ I"t‘o i‘

  

  
 
  
  
    

l m; LE.
“III II;. .‘I.
j?:‘ H” -(l' "I

     

 
 
   
     
   
 

I!" -H II
#433511.

,jy‘fig m

  
     
           
      
    
     
   
  
   
     
      
   
    
   
 
 
 

 
 

mg???"
H
4;?

"L
' ”my... V

   

. 61‘.-
.‘ III?“ II.
I .,I r-I .

I I '

    

  

I

‘IM- '1;

     
 

    

J; I lg, ‘
'13 IIIjIII5I":EI_ . ‘

‘..
“r

.‘
~v

    
 
 
   
  
 
   
  
    

  
 
   

 

.l‘ I.
~ >‘V A
.. . ._
. '. , ..J is} ‘_

. . ' 0': ‘2'.
5-153}... 1: '. -..""." '33; +1.
gang", 3;.- 7. ._ xii-Mn:

. .2 .9 r. L

. . j. _.

V ‘ ‘lr l‘l
. _.. . "rm :- .

 

    
  
 
 
 
 
  
 

     

       

        
 

 

  
     
 
   

  

  
      

"I
hi | TB." :35 .1 .’ :
SIIJIINIIE = . g E.
.-‘?'i. III-.1333" WI: we... . I
I iii) 1£7.55%*;:‘$é%;;:f§bi.al... .: “W‘s. ‘ ' "
x : 4321 II 6' . ’I""‘»i‘:I‘-“I' #190}? ‘ “*3 ’ '
a? 5 «III . ..,:-.:5;:*:-. "’“Ififi‘ifipiiigs. .n-E. I‘ ..Iiffist’
kin-E‘I‘I I; lit-LIN : .2‘ M .::;:':a:f _ .. “M ‘” I =
vi 1. M’Ii ' Wm»! . I 'I‘ - -. ‘ ‘
SII'IIJIIIIIIH HIE: ‘II I"!!! . ,‘ I '2‘ ‘-‘ I‘ ' . n “in
51“.”, 1 (“5‘1“ I‘O' ”It 5:0II5t - » :IF'fi‘ffixtgfi'.
I539”II-III?II'I‘iIiII-‘EII‘PI.5. . :I .' Q “‘-’ ’ {’35
1}”II‘I‘H'é'I’Iéib-‘V’I-‘I‘I'IIIII s ‘ = ,. ' ‘
III;.sé??£f.aigzfitugi.;§§EIIIEIIIIII . I _ I? .-
l; IIIIIJI :13 . ; I ' :I . ' I F,).{ , "f, .
WWWWI$%2» fififiI
IIIII’III' 3.5%.}?ij “WIT rII'IIII’ I I 1,85 3'":
IIIII‘II‘I‘III'III‘ [I’jIIIIIIIIE‘iIE‘ I {I I .
'j?"FIRI‘IIIFIIESE‘IIIIIII §5IZIIII5I ' i $35
i,;"'IItIIIIIII§fI§I Eiva ,IEEE
.I ”why“? I.“ III
‘WI‘II‘IIE5I5E IIIHIII 5;;5Ei5I11555II’! ' 3;! I‘
III III. IEIEIIII,1II.IIIIIIIII f
vitresIIIIIIIIIEIIII; ié ‘I'I E .
IIIII ”:1...“va '3 I” I {I
eia‘i‘II 5+ MINE I‘I‘II‘IIIH ‘ Iii
MI“ «J IIII' IIIIII'IIII‘I JII‘
II III w. I I -
III» :I W III:
5 5 I . ..

was;

                                                                 

                                           

willllll

31293 01571 1504

This is to certify that the

thesis entitled

DETERMINATION OF DIFFUSION”AND PREDICTION OF MIGRATION
OF XYLENE AND 2,4-DICHLOROPHENOL THROUGH HIGH
DENSITY POLYETHYLENE

presented by
RICHA SHARMA

has been accepted towards fulfillment
of the requirements for

MASTER degree in PACKAGING

 

 

Jim“ z 1%

Major professor

 

Date JULY 8, 1997

0.7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY

Michigan State
Universlty

 

 

 

PLACE II RETURN BOX to remov- thb chockout from your record.
TO AVOID FINES Mum on or him dd. duo.

DATE DUE DATE DUE DATE DUE

       

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU In An Affirmative ActionlEquII Opportunlty Inflation
Wm:

 

DETERMINATION OF DIFFUSION AND PREDICTION OF MIGRATION 0F
XYLENE AND 2,4-DICHLOROPHENOL THROUGH HIGH DENSITY
POLYETHYLENE

By

Richa Sharma

A THESIS

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

MASTER OF SCIENCE

School of Packaging

1997

ABSTRACT
DETERMINATION OF DIFFUSION AND PREDICTION OF MIGRATION OF
XYLENE AND 2,4-DICHLOROPHENOL THROUGH HIGH DENSITY
POLYETHYLENE
By

Richa Sharma

A study was performed to predict migration of two possible contaminants, xylene
and 2,4 dichlorophenol into food from a package structure incorporating recycled HDPE
and a food contact virgin barrier. The virgin barrier is expected to reduce or delay the
migration of the contaminants. The prediction of migration was done using two
mathematical models, proposed by Begley and Hollifield and by Laoubi and Vergnaud.
On the basis of the prediction, the dietary concentrations of the contaminants in the food
were also determined. For the prediction of migration, the models require the diffusion
coefficients of the compounds through the polymer. These were obtained experimentally
using isostatic and quasi isostatic test procedures.

The results of the models were analyzed and the predicted migration from the two
models were compared. The results of the predictions of the models indicated that the
virgin barrier is not very effective in controlling the migration into the food. However it
was also observed that the simplifying assumptions used by these models lead to a
probable over-estimation of the migration of the contaminants from the packaging into

the food.

To my family

iii

ACKNOWLEDGMENTS

I am very thankful to my major professor Dr. Susan Selke, for her guidance,
encouragement and assistance throughout my research.

I would also like to express my heartfelt thanks to Dr. Jack Giacin for his valuable advice.
and to Dr. Alec Scranton, Department of Chemical Engineering, for his suggestions while
serving on my committee.

I am grateful to the Center for Food and Pharmaceutical Packaging Research for
providing financial support for this project.

I thank my family at home for their love and encouragement throughout my study.

Last, but not the least, I thank all my friends in the school of packaging for their constant

help, kind words and moral support.

iv

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

1. INTRODUCTION
1.1.1. Background

1.1.2 Objectives of the Study

2. LITERATURE REVIEW

Part: I

2.1.1 The Threshold of Regulations

2.1.2 The need For Functional Barriers

2.1.3 Description of a Functional Barrier

2.1.4 Recent Developments in the Incorporation of Functional
Barriers for Recycled Polymers

Partzll

2.2.1 Efficiency of Functional Barriers

2.2.3 Empirical Determination of Functional Barriers

2.2.3 Theoretical Models

ix

11

13

16

17

18

2.2.4 Model 1 : Begley and Hollifield Model
2.2.5 Model 2 : S. Laoubi and J .M. Vergnaud
2.2.6 Model 3 : L. L. Katan

2.2.7 Prediction of Dietary Concentration
Part III

2.3.1 Permeability Theory

2.3.2 Measurement of Permeability

2.3.3 Absolute Pressure Method

2.3.4 Isostatic Method

2.3.5 Quasi Isostatic Procedure

3. MATERIALS AND METHODS

3.1.1 Film Sample

3.1.2 Model Test Contaminants

3.1.3 2,4 Dichlorophenol Physical and Chemical Characteristics
3.1.4 Xylene : Physical and Chemical Characteristics

3.1.5 Permeability Measurements on 2,4 Dichlorophenol

3.1.6 Permeability Experiments on Xylene

3.1.7 Determination of Vapor Pressure

vi

18

26

32

33

35

37

38

39

41

45

45

47

48

48

50

52

3.1.8 Calibration of the Gas Chromatograph

3.1.9 Theoretical Modeling

4. RESULTS AND DISCUSSION

4.1.1 Calibration Curves for Xylene and 2,4 Dichlorophenol

4.1.2 Results of Permeability Experiments of Xylene

4.1.3 Results of Permeability Experiments on 2,4 Dichlorophenol

4.1.4 Incorporation of the Diffusion Coefficients into the
Mathematical Models: Begley and Hollifield

4.1.5 The Laoubi and Vergnaud Model

4.1.6 Comparison of the Two Models

5. CONCLUSION
5.1.1. Summary

5.1.2 Recommendations and Future Work

6. APPENDICES
Appendix 1. Gas Chromatograph Conditions

Appendix 2. Calibration Curves

vii

53

54

55

55

60

73

73

8O

86

88

90

92

Appendix 3. Transmission Profile Curve

7. BIBLIOGRAPHY

viii

94

95

LIST OF TABLES

Table 1. Xylene: Results of Isostatic Permeability Experiments 56
Table 2. 2,4 Dichlorophenol: Summary of Permeability and Diffusion Coefficients 56

Table 3. Xylene: Prediction of Migration Using the Begley and Hollifield
Model for Monolayer Recycled Strucure (1 Mil) 63

Table 4. Xylene: Prediction of Migration Using the Begley and Hollifield
Model for Two Layer Barrier Structure (10 mils) 65

Table 5. 2,4 Dichlorophenol: Prediction of Migration Using the Begley and

Hollifield Model for a Monolayer Recycled Structure (1 mil) 70
Table 6. 2,4 Dichlorophenol: Prediction of Migration Using the Begley and

Hollifield Model for Two Layer Barrier Structure (10 mil) 72
Table 7. Xylene: Laoubi and Vergnuad Model: Prediction of Migration (1mil) 74

Table 8. 2,4 Dichlorophenol: Laoubi and Vergnuad Model: Prediction of
Migration (1 mil) 75

Table 9. Xylene: Laoubi and Vergnaud Model: Prediction of Migration (10mil) 76

Table 10. 2,4 Dichlorophenol: Laoubi and Vergnaud Model: Prediction of
Migration (10 mil), 77

ix

LIST OF FIGURES

Fig. 1. Schematic of Begley and Hollifield Model

Fig. 2. Scheme of the Two Layer Packaging

Fig. 3. Typical Transmission Profile Curve

Fig. 4. Plot of Diffusion vs Vapor Activity For Xylene
Fig. 5. Plot of Permeability vs Vapor Activity for Xylene
Fig. 6. 2,4 Dichlorophenol: Transmission Profile Curve

Fig. 7. Xylene: Comparison of Migration Through Single
Layer and Two Layer Structure

Fig. 8. 2,4 Dichlorophenol: Comparison of Migration Through
Single Layer and Two Layer Structure

Fig. 9 Xylene: Concentration of Contaminant in the Package as
a Function of Time.

Fig. 10 2,4 Dichlorophenol: Concentration of Contaminant in the Package as
a Function of Time

Fig. 11. Xylene: Comparison of Migration Between the Two Models

Fig. 12. 2,4 Dichlorophenol : Comparison of Migration Between the Two Models

23

28

43

58

59

61

68

69

81

82

83

84

1. INTRODUCTION

1.1.1 Background

The disposal of solid wastes is a major issue that has been faced by most
governments during the past few years. Since the existing landfills have already reached,
or are rapidly reaching, their capacity, there is a need to turn towards recycling as the
solution to the ever increasing problem of solid waste. Packaging, in various forms, is a
major component of the solid waste stream in the United States. This includes materials
such as paperboard, metal, plastics, glass, etc. as well as various types of composites of
these materials. Within these packaging materials, plastics by themselves represent a
significant share of the solid waste. It is for this reason that within the last few years,
pressures to utilize post consumer recycled content in plastics packaging have increased
greatly.

Plastics are used in all kinds of industries to package a wide range of
products because of their versatility. They combine properties of all the existing
packaging materials to give a product with features such as reasonable barrier properties,
ease in processing, light weight, economy, printability and many other advantages.
Particularly in the food industry, plastics are rapidly replacing conventional packaging

materials such as glass and metals. They are used in a wide variety of applications and

may be either in the form of rigid or flexible packages for food products. These packages
can be made up of single or multi-layered plastic structures and used either for primary or
secondary packaging.

Due to increasing pressures to reduce solid waste, most industries are being
pushed towards incorporating recycled plastics into their packaging. Specifically, the
food industry, which previously stayed away from using recycled packaging materials in
contact with food, is now beginning to utilize post-consumer recycled plastics in
packaging.

One of the main concerns in using recycled plastic as a packaging material for
food contact applications, is that after its use by the consumer, there is a possibility that
the package could come in contact with various substances for which it was not originally
intended. Some of these substances could be harmful for human consumption and could
even be potential carcinogens. An example of this contact with toxic materials is the
possible storage of pesticides, oils, or various other household chemicals in plastic bottles
afier the original product has been consumed. Because of the relatively poor barrier
properties of plastics compared to glass and metal, if the toxic material is stored for a
sufficiently long period of time, some of the contaminant may be sorbed into the plastic
package wall. When these contaminated packages are taken for reprocessing, there is a

strong possibility that the cleaning and processing operations are not efficient enough to

remove all of the contaminants from the polymer resin. Finally, after the reprocessed
resin is converted to a new packaging material for food contact purposes, there is a
danger of these residual contaminants migrating into the food product over its shelf life.
This may be may be harmful if the product is consumed by humans, with the extent of
damage caused depending on the nature of the contaminant and the amount of
contaminant that has migrated into the food product.

This kind of contamination of food products due to contact with post-
consumer recycled plastics is a major concern of the FDA for consumer safety. FDA’s
primary interest is to keep the consumer safe from the possible existence of these toxic
materials, and to keep a control on the maximum possible intake of any residual
contaminants from post-consumer recycled plastics used in food contact packaging. For
this purpose the FDA has recently adopted a policy known as the “threshold of
regulation,” which limits the amount of contaminant that can be present to a level which

it considers safe for repeated human consumption.

1.1.2 Objectives of the study

Among the various plastics used for packaging, high density polyethylene,

(HDPE) is a polymer that has very wide applications in the food and consumer packaging

industry. Large volumes of this polymer are used in food packaging, and due to
environmental issues there is an increasing need to incorporate the recycled HDPE back
into the food packaging system. It can be risky to have the recycled HDPE in direct
contact with the food because of the safety issues surrounding the possible migration of
contaminant into the food product. Hence it appears feasible to consider a “safer”
alternative of incorporating a food contact barrier made of virgin HDPE separating the
recycled HDPE and the food.

The use of “functional barriers,” a layer of virgin polymer between the
recycled layer and the food product, has been suggested to reduce the amount of
migration of contaminants from the recycled layer by acting as a barrier between the
recycled layer and the food. However, for the safety of the consumer, it is important to
ensure that these “functional barriers” actually prove to be effective in preventing or
reducing migration from the recycled layer into the food product.

This study attempts to determine the effectiveness of a virgin HDPE layer as a
“functional barrier” in reducing migration of contaminants from a recycled layer into a
food to a minimum level that is accepted by the FDA. As an alternative to performing
actual migration tests, which can be elaborate and time consuming, mathematical models
are used here which predict the migration of compounds through packages incorporating

functional barriers. This prediction will help determine if the virgin barrier proves to be

effective in reducing or preventing migration.

For the study, two representative compounds, xylene and 2,4
dichlorophenol were chosen from a range of recommended contaminants. The models
utilized for the prediction of migration are the Begley and Hollifield model and the model
proposed by S. Laoubi and J. Vergnuad. The study aims to determine the predicted
migration obtained from these selected models, compare the results between the two, and
to judge if either of the models portray what can be expected in the real-life situation of
the migration process. For the prediction of migration, these models require the diffusion
coefficient of the contaminant through the film and the initial concentration of the
contaminant in the recycled layer. The diffusion coefficients for both the selected
contaminants were determined through experimental methods using isostatic and quasi-
isostatic test procedures. The results of the predicted migration of the contaminant can
then be used to predict its dietary concentration in the food and hence can determine if
this concentration is expected to be within the limits specified by the FDA (i.e. below

0.5ppb).

The Specific objectives of this study are as follows:
- To determine the permeability and the difiusion coefficients of 2, 4, dichlorophenol and

xylene through high density polyethylene,

- To predict migration using the Begley and Hollifield model and the model suggested by
Laoubi and Vergnuad,
- To compare the predicted migration of the two models, both for monolayer packages,
and for two layer packages incorporating a functional barrier,
- To determine the dietary concentration based on these predictions and to determine if
this level of contamination is below the threshold of regulation and hence determine the
effectiveness of the functional barrier, based on the theoretical models.

A similar study is also being performed for HDPE with two other contaminants,

namely isopropanol and tetracosane, by W. Lertsiriyothin in the School of Packaging.

2. LITERATURE REVIEW

PART: I

FUNCTIONAL BARRIERS

2.1.1 The Threshold of Regulation

The law governing the use of food additives in the United States is the Federal
Food, Drug and Cosmetic Act. According to this law, a substance is deemed to be a food
additive if its “intended use results or may reasonably be expected to result either
directly or indirectly in its becoming a component, or otherwise affecting the
characteristics of the food unless the substance is generally recognized as safe (GRAS) or
is used in accordance with prior sanction of approval”. (FDA, Proposed Rules ,1993).
Food packaging materials are thus considered indirect food additives and are subject to
the regulations concerning indirect food additives.

Currently there do not exist any separate laws for recycled plastics used in direct
contact with food, and the FDA regulations apply to recycled plastics the same way as
they do for the virgin polymers. Hence this law is applicable to all kinds of packaging

materials whether they are made from 100% virgin polymer or have some component of

post-consumer recycled material.

Based on the study of the toxicological effects of a large number of potentially
hazardous chemicals, the FDA has concluded that the presence of a substance in the daily
diet at or below 0.5 ppb (parts per billion) is generally so negligible as to cause no health
problems to humans. As a result of this study, the FDA has established a dietary
concentration of 0.5 ppb as the threshold of regulation for substances used in food contact
articles. This means that the substance must not migrate, and not be expected to migrate
into food in amounts that result in dietary concentrations that are above 0.5 ppb. The
value of 0.5 ppb is equivalent to dietary exposure of 1.5 micrograms/person/day based on
a diet of 1500 grams of solid food and 1500 grams of liquid food/person/day in the
United States. (FDA, Proposed Rules,1993). This value gives the FDA 3 wide margin of
error even if the compound is later found out to be a human carcinogen. This is because a
0.5 ppb threshold is 2000 times lower than the dietary concentration at which the vast
majority of studied compounds are likely to cause non-carcinogenic toxic effects and 200
times lower than the chronic exposure level at which potent pesticides induce toxic
effects.

In recent times, the analytical methods to detect migration into food have become
capable of measuring and detecting small quantities of migrants into food products. This

level of 0.5 ppb corresponds to a migratory level that is above the detection limit for

many of the analytical methods used to quantify migrants from food contact materials.
For example if a consumption factor of 5% is used (minimum value recommended by the
FDA in absence of any Specific market volume data), it corresponds to a migration level
of 10 ppb. The analytical methods used to detect migration of food contact materials can

quantify well within this range.

2.1.2. The need for functional barriers

In order to reduce migration from the post-consumer recycled plastic into the
food, it is important that the packaging have low concentrations of contaminants to start
with. The greatest problem with post-consumer recycled plastics is that the reclamation
processes currently used are not very effective in removing contamination completely.
Some of these processes also have the added disadvantage of not being economically
feasible.

Among the currently used reclamation methods for recycled polymers in contact
with food products, the depolymerization process proves to be the most effective in the
elimination of possible contaminants. This is because depolymerization breaks down the
recycled (contaminated) polymer into basic monomeric and oligomeric units. Due to this

process, the resulting polymer has a purity that is quite comparable to the virgin polymer

and is generally considered safe for food contact materials. (Franz et. al., 1994).
However, a shortcoming of this process is that it is very expensive and is limited in its
suitability and application. Currently the only polymer that can be effectively recycled
using depolymerization is PET. Recycled PET from this process is being used for soda
beverage bottles, but as mentioned, it is not a commercially feasible process because of
high costs involved. (Overton, 1994)

Physical reprocessing is the most common recycling method for post-
consumer plastics and is generally an economically feasible method too. The high
processing temperatures involved help in removing bacterial and surface contamination
but the cleaning and processing operations of the recycled polymer may not always be
adequate to remove all the contaminants that may have absorbed into the polymer. Hence
physical reprocessing is most effective if the recycler has control over the life cycle of the
material so that the material does not come in contact with potentially dangerous
contaminants. As discussed earlier, these contaminants may migrate into the food
products when the polymer resin is processed into a food contact packaging material.
(Franz et al., 1994; Overton, 1994)

Reuse of post-consumer plastics (e.g. reuse of plastic bottles afier cleaning
and washing) is even more economically feasible than the above mentioned recycling

processes and is gaining popularity in the beverage industry. The process by itself may

10

not prove to be effective in removing contaminants from the plastic material in the event
of ‘misuse’ of the packages by the consumers. For this reason, a very tight source control
must be maintained in order to minimize contamination of the package to reduce risk of
migration. This process, however, has proved to be very suitable for certain polymers
such as PET which is generally resistant to absorbing extraneous chemicals. (Overton,
1994)

Because of the limitations of the existing reclamation processes, there is a need to
look for other options or systems in which recycled polymers may be put to use in food
packaging. One such concept, slowly gaining acceptance in the food industry, is the
combination of a recycled polymer layer and a virgin polymer layer in the structure of a
food contact package, with the virgin polymer in contact with the food to prevent

migration of residual contaminants from the recycled layer into the food.

2.1.3 Description of a functional barrier

The term “functional barrier” is generally the term used to describe a layer which,
when used in combination with recycled plastic in a packaging material, reduces
migration from any leftover contaminants in the recycled polymer into the food product

to a toxicologically acceptable level. This is so as not to endanger human health caused

11

by the possible consumption of such toxic substances in undesired levels.

Thus a functional barrier essentially incorporates the use of a virgin layer
of polymeric material which separates the food from the recycled plastic material. This
virgin layer then. acts as a barrier to the migration of possible contaminants from the
recycled plastic to the food contact layer. Migration of package components is less of a
problem with dry food products, but is more likely to occur from the packaging material
when the food product contains alcohol, acid or fat. Although the barrier layer may not
completely eliminate the migration of contaminants from the recycled plastic to the food
product, it can sufficiently slow down the process so that essentially there is negligible
migration during the life time of the product. It is important to determine whether this
level is below the threshold level specified by the FDA in their regulations.

For making a barrier structure, the post-consumer polymer is first cleaned
and re-processed by standard methods. It can then be coextruded or laminated with the
virgin resin in such a way that the virgin resin comes in contact with the food product and
the recycled layer is on the outside of the package. Most of the functional barriers
developed are made of the same polymer as the recycled resin. This is mainly because of
the ease of fabrication involved when both the layers are of the same polymer resin. It is
also important to use the same polymers in the structure in order to facilitate further

recycling of the plastic materials.

12

The package can also be fabricated as a three or more layered structure in which
the recycled polymer is sandwiched between two layers of virgin polymers. Development
of multilayer coextrusion techniques has enabled recycled polymer to be used in the core
layer. These are already being used in quite a widespread manner for detergents and other
non-food applications where migration is not really the major issue.

Recently some developments have been made and approvals have been given by
the FDA for the incorporation of barrier layers in food packaging containing post

consumer recycled plastics. These are discussed in brief below.

2.1.4 Recent developments in the incorporation of functional barriers for recycled

polymers.

According to FDA spokespersons, some materials may be safely used as a
functional barrier, even under the most severe conditions, without any migration from the
recycled layer to the food contact layer. Aluminum foil properly fabricated is one such
material that acts as a functional barrier preventing migration from the recycled layer into
the food product. Hence a recycled plastic film may be used without any concerns when
aluminum foil acts as the food contact layer separating the recycled layer from the food.

( PRTF Guidelines, 1994). However the same guarantee cannot be given for plastic

13

materials acting as functional barriers because of their relatively inferior barrier properties
compared to aluminum.

Some structures incorporating a barrier layer in contact with food products have
been prepared recently. The FDA has reviewed the use of coextruded laminated structures
where a food contact layer of virgin polymer separates the food from the recycled plastic
material. Many polyolefins, polyesters, and other polymeric materials of adequate
thicknesses have been found to effectively act as functional barriers preventing significant
migration from the recycled layers. Experimental analysis and theoretical modeling have
shown that in structures of virgin/recycled PET laminates and structures of
virgin/recycled PS laminates, a 1 mil layer of virgin PET and a 1 mil layer of virgin PS,
respectively, as the food contact layers, act as functional barriers and allow minimum
(less than 0.5ppb dietary concentration) migration of contaminants into the food when
stored for about 2 weeks at room temperature or lower. (Thorsheim and Armstrong,
1993.)

There have not been many developments on the use of a barrier structure of virgin
HDPE/recycled HDPE, which will be used in our study. Only very recently the FDA has
given approval to Union Carbide for use of post-consumer recycled HDPE for packaging
dry foods with no surface fats. It must be separated from the food by a food-grade HDPE

layer at least 4 mils thick. The resin is being offered for use in producing multilayer rigid

l4

containers, and is priced higher than virgin HDPE and recycled HDPE for non-food
packaging use.

Approval has already been given to Continental PET Technologies, ( Florence,
KY), manufacturers of containers for food and beverages, for the use of a multi-layered
PET bottle with recycled PET sandwiched between layers of virgin PET. The food
contact layer is one mil of virgin PET and separates the recycled PET from the beverage.
Migration tests conducted with this package, using the actual product, proved that the
virgin layer was an effective barrier, i.e. it reduced the migration of contaminants into the
product to a level that was much below the acceptable daily intake. The results of these
tests were submitted to the FDA and the bottle was approved for food contact purposes.
(Miller, 1993)

Extensive tests on the effectiveness of a functional barrier have been performed
on a 3 layered polypropylene cup with recycled PP as the sandwiched layer and virgin
resin as the food contact layer. The thickness of this cup was 0.8m with 50% mass
fraction of recycled polymer. The results of these tests showed that the virgin layer of the
polymer (in this case PP) functioned as an effective barrier to the migration of

contaminants from the recycled surface to the food product. (Franz et al., 1994)

15

PART: II

METHODS FOR DETERMINING THE ADEQUACY OF A FUNCTIONAL

BARRIER

2.2.1 Effectiveness of a Functional Barrier

Whenever a package incorporating a recycled layer and a functional barrier is
introduced, it is very important to determine its effectiveness in preventing migration.
The most reliable way to determine if a functional barrier is effective, is by performing
migration experiments using actual conditions. But this method may not always be
practical because of the long times involved in migration testing. Recently some
investigators have developed mathematical models to predict the migration of the
contaminants from the recycled layer into the food product without having to perform
actual tests. These models require the diffusion coefficient values of the
contaminant/polymer system and initial concentrations of the contaminant in the recycled
layers in order to make the prediction.

The effectiveness of a functional barrier can be determined by three different
methods: 1) empirically, using representative model systems; 2) theoretically, using

dimensionless parameters to define general idealized systems and 3) also theoretically,

16

but using quantitatively defined systems relevant to real life.

Various factors contribute to the adequacy of a barrier in preventing migration
from the recycled layer to the food. These include the type of the polymer, the initial
concentration of the contaminant in the recycled plastic, the nature of the food product,
the thickness of the barrier layer, conditions of use such as time, temperature, humidity,
etc. Standard testing protocols to evaluate the adequacy of a functional barrier in given

situations are followed by the FDA and industrial scientists.

2.2.2 Empirical Determination of Functional Barrier

Empirical work has been done on PET by deliberately contaminating resin with a
number of substances considered to represent those which might be found in post
consumer recycled (PCR) plastics. An experiment was done using a three layer
coextruded PET bottle with the contaminated layer sandwiched between plies of virgin
PET. Migration tests were performed using appropriate test procedures. The migration
levels were compared with the migration from single layered bottles made from recycled
PET. From the experimental data it was found that migration levels from the 3-layered
PET bottles with a functional barrier were reduced by more than two orders of magnitude

compared to the migration from single-layered bottles. (Franz et.a1., 1996)

17

2.2.3 Theoretical Models

Theoretical models have been developed recently for the prediction of the
migration of a contaminant from a recycled package system into the food product. The
models under consideration in this study can predict migration of a contaminant over a
period of time. For this prediction, it is necessary to have knowledge of the diffusion
coefficients of the contaminant/package system, thickness of the material, and the initial
concentration of the contaminant in the polymer. There are also some models that require
the determination of the partition coefficient to predict migration. The models that are

used in this study to predict migration are explained in brief below.

2.2.4 Model 1: Begley and Hollifield Model

A mathematical model was proposed by Begley and Hollifield (1993) to predict
migration into the food from a package structure. The model predicts migration through a
monolayer recycled package structure (Case A) as well as migration through a two-
layered package structure with a virgin layer separating the recycled polymer from the
food. (Case B). These models include many assumptions that simplify the real-life

situation of migration through the package structure into the food.

18

CASE A: Migration Through a Monolayer Package

In the first case involving the prediction of migration through the monolayer
package, the package is assumed to be made of 100% recycled polymer and is in direct I
contact with the food product. This situation represents a worst case scenario. The
migration from this type of package into the food is assumed to follow Fickian behaviour

and is mathematically expressed by the following equation (Crank, 1975) ;

Mt=2C .|2—‘
° W (1)

Where:

M, is the amount of substance migrating from a unit surface area of package,
expressed in ug/cmz,
C0 is the concentration of the contaminant in the polymer at time 0, expressed as ug/cc;
t is the time, in seconds;
D is the diffusion coefficient of the contaminant through the polymer film, expressed in

cmz/sec

19

The initial concentration of both the contaminants in the recycled film, Co, was
assumed, based on the Plastics Recycling Task Force document (PRTF) , “Guidelines for
the Safe Use of Recycled Plastics for Food Packaging Applications.” The guidelines state
the seven compounds that best represent the possible contamination in post consumer
recycled plastic. This is discussed in more detail in Chapter 3. The guidelines also give
an indication of the maximum level at which each of the seven compounds are expected
to remain in the post-consumer HDPE before it undergoes the cleaning and recycling
operations. According to the guidelines, each of these seven compounds will have an
effective contamination rate of a maximum of 1% in the polymer before it has been
cleaned and recycled into new packaging material. This implies a contamination rate of a
maximum of 1% each, for both xylene and 2, 4 dichlorophenol in HDPE, which reduces
after recycling and cleaning operations. In order to arrive at this contamination in the
recycled plastic, the PRTF arranged for the cleaning and reprocessing of the HDPE
spiked with 1% of each of the contaminants. The contaminant remaining in the plastic
after cleaning and processing gave the maximum concentration likely to remain in the
post consumer recycled plastic. This concentration represented the initial concentration
C0 , and was determined as 120 ug/cc for xylene and 9.6 ug/cc for 2,4 dichlorophenol.

The prediction by this model is based on the assumption that the contaminant

concentration in the polymer does not change during storage, and the food is an infinite

20

sink for the contaminant, providing no resistance to mass transfer. Also the model does
not consider the effects of partition equilibrium on the migration of contaminants from
food packaging into the product.

Due to over-simplifying assumptions, the values of migration
predicted from this expression are expected to be much higher than those obtained from
actual migration testing. For example, in real life conditions, the concentration of the
contaminant in the recycled layer will reduce with time as the contaminant migrates into
the food. In this model, however, the concentration in the recycled layer is assumed to be
constant with time. So when applied in the expression it will depict more migration than
would be generallyexpected, especially for longer periods of time. Another assumption
of the model is that it does not consider migration of the contaminant out through the
outer layer of the recycled polymer which will further reduce the amount of contaminant

that will migrate into the food.

CASE B: Migration Through a Barrier Structure

The second case considers migration through a two-layered barrier package

structure. In this case the package is a laminated package structure made of a layer of

21

recycled polymer and a layer of virgin polymer where the virgin polymer layer is the food
contact layer and separates the food product from the recycled layer. Fig. (1 B) is a
schematic illustration of this situation.

In this case, the contaminant will first enter the layer of virgin polymer before it
enters the food. The initial concentration of the contaminant in the virgin layer is zero and
there is a finite concentration in the recycled layer. For simplification purposes, this type
of migration can be treated as a pseudo membrane problem where there is a fixed
concentration on one side of the film and a zero concentration on the other side.

The solution to this type of problem is expressed by the equation below;

 

°° _ n
Mt=[22£-l-i£ ] exp(-Dn21t2—%)]Col
I l I

5 n2 n2

(2)

Where;
I is the thickness of the virgin layer in cm.

Co is the initial contaminant concentration in the recycled layer.

22

Fig. (1)

 

 

 

 

 

 

 

Food

1 C0 C0
’1
. — 1—
l

Recycled Recycled Virgin

HDPE HDPE HDPE

(A) (B)

Schematic of the Begley and Hollifield model with single layer and a two layer structure
incorporating a functional barrier.

23

This equation is also simplified by various assumptions. As in the previous
expression for the prediction of migration through a single layer, this expression also
assumes that the concentration of the contaminant in the recycled layer is constant which
will tend to give a higher prediction of migration than will be norrnaly expected in real
life situations. Also it is likely that the contaminant will not move freely into the virgin
layer (as is assumed in the model), but will first diffuse out of the recycled layer before it
enters the virgin layer. Because of these assumptions, the prediction of migration is
expected to be much more than what is expected in the real-life situation, as is the case in
the monolayer package.

The model also describes a variable 1:, given by:

,2 (3)

which gives a general picture of the rate of mass transfer. The larger the value of 1: , the
more contaminant will have crossed the virgin layer and contaminated the food. Also, at
values of 1: >06, the migration from the two layer package is expected to approach that

from a single layer package. The model is not valid for the prediction of migration for a

two layer structure for values of 1.’ >06. This is due to the fact that at higher values of 1:,

24

either the diffusion coefficient is so high or the thickness of the virgin layer/recycled
layer is so low that there will be no effect of a barrier in preventing migration. Also a
high value of 1: could be due to larger time periods which may give the contaminant an
opportunity to eventually make its way through the functional barrier into the food. In
either of these cases there is practically no use of employing a functional barrier.

These equations have many visible flaws because of the over-simplification of the
diffusion process with various assumptions, and consequently, an over-estimation of
migration. However, they can still be used as a starting point estimate to predict
migration through a single layer and a two layer package system. The results from these
models can also be compared with other models or with actual migration tests. The
predictions can also be used to determine if the models can be used for a wide range of
contaminants and if they are applicable to what is normally expected in real life migration

situations.

25

2.2.5 MODEL 2 : S. Laoubi and J.M. Vergnaud

Another mathematical expression was developed by S. Laoubi and J.M.Vergnaud
(1995), to predict the migration of a contaminant through a two-layer package structure
incorporating a functional barrier.

Like the model developed by Begley and Hollifield, this model too makes
certain assumptions in order to simplify the complex situation of mass transfer through
this system. The assumptions made by this model are as follows :

1) The recycled and the virgin layer are made of the same polymer so that the diffusivity
of the contaminant is the same through both the layers.

2) The contaminant is initially present in the recycled polymer in a uniform
concentration.

3) The contaminant migrates through the two films and through the surface of the food-
contact virgin polymer but there is assumed to be no migration from the external surface
of the recycled polymer.

4) The mass transfer follows Fickian diffusion.

5) The food does not migrate into the polymer.

6) The thickness of the film changes negligibly because of the very small concentrations

of the contaminant.

26

7) The concentration of the contaminant in the food is negligible because of the large
volume of the food in contact with the functional barrier.

In order to arrive at the prediction of migration through such a structure,
the investigators started by expressing diffusion with the help of F ick’s equation with a

constant diffusivity , D

6C 2C
1,! :D * é x.t
5‘ Ox 2
(4)

where C,” is the concentration of the contaminant in the package structure at position x
and time t.

A scheme of the barrier structure is shown in Fig. (2) .The initial conditions state that at
time =0, at any position x in the recycled layer, the concentration of the contaminant C is
equal to the initial concentration Co or Cm.

Hence the initial boundary conditions can be given by,

at t=0, for 0<x<H, C=Cin;

and for H<x<L, C=0

27

Fig. (2)

 

 

Cin
I
Food
_ I —
O H L

 

 

 

Scheme of the two layer packaging, with a recycled film as functional barrier

suggested by Laoubi and Vergnaud.- Initial boundary conditions

28

Where L is the total thickness of the package and H is the thickness of the
recycled layer.
At the position x =0 there is no mass transfer. However at position x=L, according to the
mathematical prediction, there will be mass transfer with a convective coefficient of mass
transfer h. Using these boundary conditions,

when t > 0 , =0 6C/6x=0 - there is no transfer

when x=L , -D OC/Ox = h(CL-Cm)

when h is very large , h-m, and the above equation simplifies to

at =L CL=Cm

In both the cases, the concentration in the food is assumed to be negligible

and hence Cext =0

The concentration of the contaminant is 0 at the food contact surface at the start of the

migration process. For h=oo, the concentration at position x and time t, CM is expressed as

a fraction of the uniform initial concentration of the contaminant in the recycled polymer.

29

 

Dr:

3,! :12 1 sin (2n +1)‘ItHcos (2n;2)mexp(
It

_(2n+l)21t2

(5)

The amount of contaminant remaining in the packaging at time t, Mm , is obtained
by integrating the concentration Cat: with respect to the space between the boundaries 0
and L . It is expressed in terms of time as the fraction of the initial concentration of the

contaminant located in the recycled polymer, and is given by the expression;

 

M °° _. n _ 2 2
Azii ( 1) sin(2n+1)1tHexp( (2n+1)1t D.
M, 1:2 Hn=0 (2n+1)2 2L 4L2 (6)

Where:
The value (Mm /Mi) is a dimensionless ratio,
L is the thickness of the entire structure,
H is the thickness of the recycled polymer,
D is the diffusion coefficient of the contaminant through the film, and

t is the time in seconds

The initial concentration of the contaminant in the polymer is determined as

discussed in the previous article.With the knowledge of the initial concentration of the

30

contaminant in the polymer, we can determine the amount that will migrate into the food
product knowing the diffusion coefficient of the contaminant/polymer system. While the
initial concentration of the contaminant in the polymer is dependent on the original
contaminant level in the polymer and the efficiency of the recycling and the cleaning
process, the diffusion coefficient through the film depends upon the fundamental
characteristics of the polymer film and the contaminant involved, and can be determined
experimentally. As is evident from the equation, this model is not specific to any
particular polymer and can be applied to any structure where the recycled layer and the
virgin layer are both made up of the same polymer.

This model is capable of giving the following results which will help in
determining the adequacy of the functional barrier :
— profiles of the concentrations of the contaminant through the structures for different
times,
- rates of transfer of contaminant in the functional barrier,
- concentration of the contaminant at the external surface of packaging,

- and rate of transfer of contaminants into the food.

31

2.2.6 Third approach : L.L. Katan

A third approach to this type of problem has been suggested by L.L. Katan (1996)
in which the solution to the prediction of the migration through a functional barrier
structure is based on the assumption that the virgin layer acts like a “sponge” . According
to this theory, at a time t, which is also called the burst-out time, a substantial portion of
the contaminant diffuses into the barrier, and then migrates into the food at an
accelerating rate. It states that in this situation the thickness of the barrier layer has more
effect on the burst out time than the diffusion coefficient of the compound. That is, the
diffusion coefficient does not play as much a role as the thickness of the polymer in
reducing the migration. The mathematical treatment gives the concentration of the

contaminants in the food contact layer as :

b (7)
where x, and xb are the thickness of the recyclate and barrier layers respectively. From
this model it follows that if x/xb is small ( i.e. the barrier layer is small), the barrier effect
is insignificant. Co is the initial concentration of the contaminant, and C. is the

concentration of the contaminant in the food contact barrier layer at time t=oo.This model,

32

unlike the above two models, does not take into account the diffusion coefficients of the
compound. This system is not incorporated in the study since we are interested in
predicting migration for individual specific contaminant/polymer systems. For this type

of prediction we need the diffusion values for the contaminant! polymer system.

2.2.7 Prediction of the Dietary Concentration

From the predicted migration of the contaminant into the food, the dietary
concentration of the contaminant in the food can be calculated. This can be a guideline to
determine how long the virgin layer will act as a functional barrier preventing migration
into the food above acceptable levels. In order to predict the dietary concentration, apart
from the migration values, the consumption factors and the food type distribution factors
also need to be determined. These are discussed below.

The FDA has defined consumption factors as the portion of the diet likely to
contact specific packaging materials. It is defined as the ratio of the weight of food
contacting a specific packaging material to the weight of all food packaging and varies
for different packaging materials. It is obtained from factors such as the types of foods,
the ratio of the food to the weight of the package, and the number of food packaging units

in each food packaging category. The consumption factors are well defined for different

33

polymers depending on the extent of their use in the market. In the event that they are not
expected to have a widespread use in the market, the minimum consumption factor that
can be used for any polymer is 0.05. On the basis of these factors, the consumption factor
of polyethylene is 0.13, assuming a wide spread use of the polymer. This is the value that
is used in our work in calculating the dietary concentration of the contaminant in the
food. Also, since the migration of a contaminant into a food depends on the nature of the
food, (ie, aqueous, acidic, alcoholic or fatty) food type distribution factors are also
considered to account for the variable nature of the food contacting each packaging
material. It indicates the fraction of the food contacting each material that is aqueous,
acidic, alcoholic and fatty. Another factor that has to be taken into consideration while
calculating the concentration of the contaminant migrated into the food, is the amount of
food in direct contact with the packaging material. According to the FDA guidelines, on
an average 1.55 grams of food contacts one square cm of packaging material. The dietary
concentration in the food is calculated by the following expression:
Dietary concentration = <M>*CF*Ft

Where CF is the consumption factor, M is the migration of the contaminant in the
food, Ft is the food type distribution factor for the system. In our studies, the food type
distribution factor has been taken as 1 since the models assume that there is no restriction

to the transfer of the contaminant at the interface of the packaging and the food. Hence

34

the dietary concentration is calculated on the basis of the predicted migration and
consumption factor alone.

The mathematical models can thus be used to predict migration of
contaminants over a period of time after determining the diffirsion coefficients of the
contaminant/ polymer systems by actual experiments. The diffusion coefficients obtained
can be incorporated into the equations along with the film thicknesses and the initial
concentration of the contaminants. The diffusion experiments are performed using
isostatic and quasi-isostatic procedures and a brief description of the theory and the

experimental setup is described in the following pages.

PART III

MASS TRANSFER: THEORY

2.3.1 Permeability Theory

Since plastics, unlike glass and metals, are not complete barriers to vapors, the
study of permeation and mass transport behavior of vapors through plastic films is
necessary to judge its extent of barrier properties. Permeation of organic vapors, water

vapor, and gases such as oxygen and carbon dioxide, is experimentally determined using

35

various methods.

Basically permeation is the transport of a substance through a substrate, involving
the following stages: adsorption of the permeating species onto the surface of the
polymer, solution of the gas or vapor into the polymer matrix, diffusion through the wall
along a concentration gradient and desorption on the other surface.

For films, permeability of a gas or vapor through the film is directly proportional
to the diffusion coefficient, usually denoted as “D,” and to the solubility coefficient,
usually denoted as “S.”

Mathematically the permeability is expressed as:

P = DxS (3)

where the diffusion coefficient ‘D’ is related to the rate of the advancement of the
permeant molecules and the time required for the system to reach steady state. For
simplification it is assumed that the concentration/distance relationship through the
polymer is linear and that diffusion takes place in one direction only

(ie. through the film). For gases and vapors that Show no interaction with the polymer,
the permeability coefficient is independent of the concentration of the diffusing gas.

The transmission rate is the amount of penetrating substance transferred per unit

36

time through a package or a specimen of stated dimensions. Transmission rate is
generally inversely proportional to the thickness of the morphologically uniform
specimens and is also proportional to the driving force. The permeance is the
transmission rate normalized for the area and the driving force. This study involves the
experimental determination of permeability data such as diffusion, solubility and
permeability. Under low concentrations of vapor, the vapor pressure of the penetrant is

proportional to its concentration in the polymer. From Henrys Law it follows that,

C=S x dP (9)

where;
C=concentration of the penetrant in the polymer phase
S=solubility coefficient

dP=partial pressure of the penetrant in the gas phase

2.3.2 Measurement of Permeability

The following three methods are generally used to measure gas and vapor

permeability for packaging materials:

37

(I) the absolute pressure method,
(ii) the isostatic method, and

(iii) the quasi- isostatic method.

2.3.3 Absolute Pressure Method

The earlier studies of permeability of organic vapor through polymer fihns were
carried out using this method. In the absolute pressure method, only the permeant vapor
in question is used and no other carrier gas or vapor is employed. The total pressure
gradient between the two chambers provides the driving force for the mass transfer
process. (ASTM D3985). The permeated gas or vapor is determined by a manometric or
volumetric technique. Steady state is attained when a constant increase in pressure or
volume with respect to time is observed in the low pressure chamber. However the
problem with using this method for permeability determinations is that the study cannot

be done at different relative humidities and co-permeants cannot be used.

38

2.3.4. Isostatic method

In the isostatic test procedure, the total pressure in both the chambers of the
permeation cell is kept constant. A constant permeant partial pressure or concentration
gradient is maintained between the two cell chambers, by continuously sweeping the
desired permeant concentration in an inert carrier gas stream over one side of the film and
maintaining the carrier gas flow on the other side concurrently. The permeant diffuses
through the film into the lower concentration chamber and is then taken up by the carrier
gas to a detector where it is quantified. The steady state permeation rate is the product of
the steady state concentration of the permeant in the carrier gas stream and the flow rate
of carrier gas. Continuous monitoring of the transmission rate of organic vapors through
the polymer film can be obtained using the isostatic test procedure from initial time to the
steady state conditions. The permeability of this vapor at different relative humidities can
also be determined.There are a number of test systems and equipments used to measure
permeation by the isostatic method. Recently, two commercial instruments have been
introduced for measuring the organic vapor permeability. MOCON’S AROMATRAN
Permeation test system (marketed by Modern Controls Inc., MN) is used for measuring
the permeation rate of various organic compounds through polymer films. This system is

equipped with flame ionization detector (FID) with a high degree of sensitivity.

39

The other system, which is the one that is used in this study, is the MAS 2000
organic permeation detection system marketed by Testing Machines Inc., NY. This
system is also equipped with an FID that provides sensitivity levels in the region of low
parts per billion. The determination of permeability, diffusion coefficient and solubility
can be obtained rapidly using the MAS 2000 system, which is interfaced with an IBM
computer. The computer controls the conditions of the experiment such as temperature,
pressure, flow rates and cell parameters such as opening and closing of the cell. The
computer is equipped with Lotus 123TM software that can be used to analyze the data
obtained from the permeability experiments.

The amount permeated through the film is plotted against time and the time required to
reach a rate of transmission equal to half the steady state is determined. This is then used
to calculate the diffusion coefficient as follows:

12
7.199t05 (10)

 

Where 1 is the thickness of the polymer and to5 is the time required to reach a rate of
transmission equal to half the steady state value. The permeability coefficient P can be

determined from the isostatic procedure by substitution into the equation below

40

= a *G *f*l

A*b (11)

where:

a= calibration factor

G= response units fi'om detector output at steady state.

f= flow rate of sweep gas conveying penetrant to the detector.
A=area of the film exposed to the permeant in the permeation cell
I= film thickness

b= partial pressure

2.3.5 Quasi-Isostatic Procedure

The quasi-isostatic method is an accumulation test method for determining
the permeability of organic vapors through barrier films. In this test method, the polymer
film is mounted in the permeability cell above a reservoir of the contaminant. The
permeant vapor which has diffused through the barrier fihn accumulates in the low
concentration chamber of the permeability cell above the film. The quantity of the
permeant accumulated in the low concentration chamber increases as a function of time
The vapor is then generally quantified using gas chromatography.

A quasi-isostatic test system was developed for determining the

41

permeability of organic vapors through packaging films by Hilton and Nee (1978).
However this system had the drawback that it was limited to only saturated vapors at a
given temperature.

Franz (1993) developed a system that could permit permeability experiments at
low vapor pressure. It is a variation of the quasi isostatic procedure and involves an
adsorption trap and solvent desorption techniques incorporated with gas chromatographic
analysis.

As described above, in the quasi-isostatic method, the quantity of permeant
accumulated increases as a fimction of time. When the quantity accumulated per unit time
reaches a constant value, we can say that the steady state of diffusion has been reached.
The total quantity of the permeant transmitted is plotted as a function of time and this plot
is known as the transmission profile curve. A typical plot of the transmission profile
curve is shown in Fig (3). The transmission profile curve gives the lag time 6, which is
calculated by extrapolating the linear segment of the total quantity of penetrant to have
transmitted through the film to the time axis.

The diffusion coefficient for the specific set of experimental conditions is
determined from the lag time and the thickness of the film as follows:

12
D:—
69 (12)

42

Quantity permeated (ug)

 

Fig(3)

l
l
I
I
l

Slopc(flux)=Q/T=Quantity IT ime

\

 

Time

Fig (3) Typical Transmission Profile Curve

43

where : I is the thickness of the film

0 is the intersection of the projection of the steady state portion of the
transmission profile curve with the time axis and denotes the lag time in seconds.
Also, the permeability coefficient at the steady state can be calculated by the equation

below;

 

A *b (13)

Where;
b= slope of the straight line portion of the transmission profile curve.
l= thickness of the film

A=surface area of the film exposed to the permeant in the permeability cell

44

3. MATERIALS AND METHODS

3.1.1 Film Sample

High density polyethylene (HDPE), obtained from Tredegar Film Products
(Richmond, Virginia) was used for the permeability experiments. (Trade name: Haute
Monax, Lot/S.O.# 49611). It was obtained in the form of a roll 43"x 1 meter. The roll
was kept at room temperature and carefully covered to keep it free from any kind of
damage. The film samples had thickness ranging from approximately 1.12-to-1.19 mils.
The average thickness of the roll was determined as 1.17 mils. The test samples were cut
from the roll and it was ensured that the samples did not have any kind of cracks, holes
or other types of defects that might lead to erroneous results in the permeability tests. It
was also ensured that the samples taken were free of dust particles or any other foreign

bodies.

3.1.2 Model Test Contaminants

The guidelines given by the Plastic Recycling Task Force has suggested a list of

compounds which were selected to best represent the entire range of likely contamination

45

in the recycled polymers in the post-consumer recycled stream. These contaminants have
been chosen so that they cover the range of polar, volatile and non-volatile; non polar,
volatile and non-volatile; and heavy metal. Polar and non polar compounds are chosen
since the extent of polarity will affect the solubility of a substance in another. Volatile
and non-volatile compounds are chosen because the volatility of a compound is a
function of its vapor pressure and will affect migration. The model compounds and their
representation are as follows:

Isopropanol - polar-volatile - representative of paint strippers, household
cleaners etc.

2,4, dichlorophenol - polar non- volatile - representative of pesticides

Xylene - non polar-volatile - aromatic fraction of gasolene, paint thinners etc.

Tetracosane - non- polar, non- volatile

Heavy Metal - Copper-Z-ethylhexanoate - crab grass killer

Methyl Stearate- Fats and oils, Motor Oils, Brake fluids.

Ethylene Glycol- Antifreeze

As discussed, the two contaminants selected to study in these initial permeability
experiments were:

(1) 2,4 dichlorophenol and

46

(ii) Xylene
The physical and chemical properties of the compounds are discussed in brief

below.

3.1.3 2,4, Dichlorophenol : Physical and Chemical Characteristics:

This compound is a solid at room temperature. It was obtained from Aldrich
(Milwaukee, USA) and had a purity of 99%. It is a polar non -volatile compound
possessing the following chemical characteristics. (Aldrich Catalog)

Chemical Formula - C12 C6 H, OH
Molecular Weight - 163.00

Boiling range - 209 to 210 deg C

The saturation vapor pressure of 2, 4 dichlorophenol is very low at room
temperatures. Due to the difficulties of generating vapor at low pressures, and because of
the safety concerns surrounding the use of the compound, the permeability tests had to be
limited to the quasi-isostatic method. The determination of the diffusion coefficients was

done at saturation vapor pressure of the compound.

47

3.1.4 Xylene : Physical and Chemical Characteristics:

Xylene is a liquid at room temperature. The chemical was obtained from Aldrich
(Milwaukee, USA) and was 98 % pure HPLC grade. It is non polar, volatile with the
following chemical characteristics.(Aldrich Catalog, 1996)

Chemical Formula - C6 H4 (CH3)2
Molecular weight - 106.17

Boiling range - 143 tol45 deg C

3.1.5 Experimental Methods

1. Permeability Measurements on 2,4, Dichlorophenol Using Quasi-isostatic

Methods:

As mentioned above, permeability studies for 2,4 dichlorophenol were carried out
using the quasi-isostatic method because of the limitations in operating with this chemical
due to its physical properties and health safety concerns.

For the studies, a small quantity of, 2,4 dichlorophenol was placed in a

48

glass petri dish and placed in the lower chamber of the permeability cell. The sample test
film of high density polyethylene was placed between the lower chamber and the upper
chamber of the cell. Thus a partial pressure gradient was created between the upper
chamber and the lower chamber causing the vapor to permeate from the higher
concentration chamber to the lower concentration chamber through the HDPE film.

This kind of experimental set up limited the study to the determination of
diffusion for only the saturated vapor pressure of the contaminant.The amount of
contaminant permeated through the film to the upper chamber of the cell was determined
over a period of time. This was done by taking a fixed quantity (50 microliter) of the
headspace in a gas tight syringe and injecting it into the gas chromatograph which was
coupled with an FID detector. The gas chromatograph gives the corresponding area
response to the injected sample from which the quantity permeated can be calculated
using the calibration factor for the compound. The calibration of the gas chromatograph is
discussed in detail in section 3.1.7. The sampling of the headspace in the upper chamber
was done at regular intervals of time until steady state was reached and there was no
increase in the quantity permeated. The quantity permeated with respect to time was
plotted, which gave the transmission profile curve of the contaminant through the film. A
plot of a transmission profile curve obtained by the experiment is shown in Fig (1). The

experiment was repeated 3 times in order to obtain consistency of the results

49

3.1.6 Permeability Experiments for Xylene Using Isostatic Method:

The determination of the diffusion and the permeability coefficients of xylene was
carried out by the isostatic method using the Mas 2000 vapor permeability tester as
described earlier. The tests were conducted at four different vapor activities, 0.1, 0.2, 0.3,
and 0.4 corresponding to partial vapor pressures of 0.5, 1.0, 1.5, and 2.0 mmHg of the
contaminant. The tests were conducted in triplicate in order to ensure validity of the
results. At the start of the test, the test parameters and measurements and other relevant
information about the test sample and the conditions of the test were entered into the
computer interfaced with the Mas 2000.

For a test run, the vapor activity was calculated from the following equation;

 

VaporActivity = P
Psat (14)

Where ;
Psat is the saturation vapor pressure of the contaminant and
P is the partial pressure of the permeant.

The vapor was generated by bubbling nitrogen through the liquid permeant in a

50

glass vapor generator. The permeant vapor stream was then diluted with another stream
of nitrogen in order to obtain a low vapor pressure, and hence desired low vapor activity.
The area response of the gas chromatograph corresponding to the required vapor activity
was calculated using the gas law and the calibration factor. The determination of the
calibration factor is discussed in detail in section 3.1.7. The settings of the gas
chromatograph are shown in Appendix 1. The flow meters were adjusted correspondingly
to get the required area response when a sample of the permeant flow was taken from the
sampling port and injected into the column of the gas chromatograph.

After the vapor pressure, and consequently the vapor activity, was adjusted and
the signal base line reached a steady state, the Mass 2000 was calibrated for the permeant
by injecting the vapor from the sampling port into the calibration port of the Mass 2000.
This was done in the non-test mode. The detector response is indicated by pico amps. The
Mass 2000 calculates the calibration factor automatically. The consistency of the vapor
activity was verified periodically by injecting a sample of the vapor from the sampling
port into the gas chromatograph and ensuring a constant area response. The machine was
then run on the test mode and the quantity permeated was calculated over a period of
time. A sample of the raw data obtained for a specific vapor activity, from the Mas 2000,
is shown in Appendix 4. This gives the increase in the signal as a function of time. The

quantity permeated was calculated from the signal and the calibration factor that was

51

obtained from the Mas 2000. The diffusion coefficient for the vapor activity was
calculated by determining the t0, from the raw data and using it in Equation (10) as
discussed in Chapter 2. The permeability coefficient was calculated from Equation (11)

discussed in Chapter 2.

3.1.7 Calibration of the Gas Chromatograph

The gas chromatograph was calibrated using standard solutions of the
contaminants in appropriate solvents. The conditions of the gas chromatograph are shown
in Appendix 1. A standard solution of 2,4, dichlorophenol was prepared using
dichloromethanol and a standard solution of xylene was prepared using methanol. The
solutions were prepared by varying the concentrations of the contaminants in their
solvents, quantified in parts per million. A master batch of 1000ppm of the contaminant
in the solution was prepared which was further diluted to the required concentrations for
the compounds. The concentrations used for the calibrations can be seen in Appendix 2.
A 1 ul sample was removed from each of the standard solutions and injected into the gas
chromatograph. The response in area units was recorded for each concentration. A
standard curve of the area response vs the permeant quantity was constructed and the

calibration factor was determined from the slope of this plot. The standard calibration

52

curve for 2,4, dichlorophenol and xylene and the conditions used are shown in the

Appendix 2.

3.1.8 Determination of Saturation Vapor Pressure

The saturated vapor pressure was determined for both the compounds using the
headspace technique. Using this method, a small quantity of each of the compounds was
placed in a glass vial and sealed. The vials were allowed to stand for 2 days at room
temperature to ensure that the headspace was saturated with the vapor of the compound.
After two days, a 100ml sample of the head space was taken with the help of an air tight
syringe and injected into the column of the gas chromatograph. The area response
corresponding to the injected volume was recorded and the saturation vapor pressure was
calculated using the standard gas law equation and the calibration factor determined for
the compound. The saturated vapor pressure for 2,4 dichlophenol was determined as 0.24

mmHg and the saturation vapor pressure of xylene was 4.5 mmHg.

53

3.1.9 Theoretical Modeling:

The diffusion coefficients obtained for both the contaminants in high density j
polyethylene were incorporated into the models for the prediction of migration. The
results obtained from these models are described in detail in Chapter 4. Prediction of the
dietary concentration was done by using the predicted migration values and the

consumption factors and these results are also discussed in the following Chapter.

54

 

4. RESULTS AND DISCUSSION

4.1.1 Calibration Curves for Xylene and 2,4 Dichlorophenol

The calibration factors were determined for xylene and 2,4 dichlorophenol as
discussed in Chapter 3. The calibration curves, a plot of the quantity injected vs the area
response, are shown in Appendix 2. The calibration factor for xylene was determined as
5.6E-13(g/Au) and the calibration factor for 2,4 dichlorophenol was determined as
2.3E-13 (g/Au).The saturation vapor pressures for both the compounds at room
temperature were calculated from the calibration factor and the gas law, and were

determined as 0.24mmHg and 4.5 mmHg for 2,4 dichlorophenol and xylene respectively.

4.1.2 Results of Permeability Experiments on Xylene

The permeability results obtained from isostatic experiments for xylene are
illustrated in Table (1). The table shows the experimentally determined diffusion
coefficients for various vapor activities ranging from 0.1 to 0.4. It can be seen that the
relationship between the diffusion coefficient and the vapor activity is quite linear and

that the diffusion coefficients remain almost constant over the range of vapor

55

 

 

Table (1)

Xylene

Results of the Isostatic Permeability Experiments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Vapor Vapor D Permeability Solubility
Activity Pressure m2lsec coefficient Coefficien
mmHg gmillcm2.sec glcc.mmH
mmHg
0 0 0 0 0
0.4 2 1.1E-14 4.1 E-10 0.009
0.3 1.5 9.2545 3.8E-10 0.019
0.2 1 8.3E-15 3.4E-10 016
0.1 0.5 8.1 E-15 3.2E-10 0.01
Average Diffusion = 9.2E-15 m2lsec
Table (2)
2,4 Dichlorophenol
Results of Permeability and Diffusion Coefficients Using the
Quasi-isostatic procedure
Test Transmissio Permeability Diffusion Lag time
No. rate (g.millsqcm. Coefficient (secs)
(gliec) sec.mmH§) sqmlsec
1 4.9E-10 3.2E-11 i 2.0E-14 7000
i 2; 4.9E-10 3.2511; 2.0E-14 7100
% 3i 4.9E-10 285-11 1 2.1E-14 7200

 

 

 

56

activities indicating that the diffusion follows Fick’s law. A plot of the diffusion
coefficients as a function of the vapor activity is shown in Fig. (4). The diffusion
coefficient is measured in mz/sec.

The permeability coefficients were also calculated from the steady state data of
the isostatic experiment and the plot of the permeability coefficient and the vapor activity
is shown in Fig. (5). The plots indicate that the permeability coefficients remain constant
over a range of vapor activities.

Table (1) also gives the values of the solubility coefficients of the compound at
these vapor activities. The solubility coefficients were calculated using the relation
P=D*S described in Chapter 2.

Data mentioning the diffusion coefficients of these two compounds was not found
in literature and hence an accurate confirmation of the results could not be made.
However, a comparison was made with the diffusion coefficients obtained for compounds
with almost similar structures. The diffusion coefficient obtained for BHT ( which is
similar in structure to 2,4 Dichlorophenol ) was 4.7x 10'” m2lsec (Hoojat et. a1. 1987)
This is close to the diffusion coefficient obtained for dichlorophenol in our study (2.0 E-
14 m2lsec). Similarly the diffusion coefficient of toluene ( which is similar in structure to
xylene) was 1.6 E-13 mz/sec. ( Shu Jung Huang, 1996). This is somewhat close to the

diffusion coefficient of xylene in our study (0.92E-14m2/sec).

57

Fig (4)

Xylene

Plot of Diffusion Vs Vapor Activity for Xylene

 

 

Diffusion (sqmlsec)

Times 1E-12

 

 

Diffusion Vs Vapor activity]

 

 

 

 

 

 

 

 

0.02
Diffusion Coefficien
0.015 - h 1
I
0.01 —
I
I I
0.005 ~
0 1 1 1 L
0 0.1 0.2 0.3 0.4 0.5
Vapor Activity

 

58

 

Fig. (5)
Xylene

Plot of Permeabilty Vs Vapor Activity for Xylene

 

 

 

Permeability Coefficient vs Vapor Activity

 

 

 

 

 

 

 

 

 

 

F

E 0.6

E.

E

8 04 I

S ' f I

E, .n'f I .. .

2’ g I . Permeability Coefficrent

c E

.g m

E 0.2 —

g

25

(D

g 0 r r r 1

g o 0.1 0.2 0.3 0.4 0.5
vapor activity

59

 

4.1.3 Results of Permeability Experiments on 2,4 Dichlorophenol

As discussed earlier, the permeability experiments for 2,4 dichlorophenol were
performed using the quasi-isostatic procedure at the saturation vapor pressure of the
compound. The experiments were performed in triplicate and the results of these
experiments are shown in Table (2). The diffusion coefficients were obtained
experimentally using equation (12) described in Chapter 2. An example of the
transmission profile curve obtained, which is a plot of the quantity permeated vs the time,
is shown in Fig (6). The transmission profile curves for the other two experiments are
shown in Appendix (3). The lag time 6, in seconds, was calculated from the x-intercept
of the straight line (steady state) portion of the curve and is shown in the table. The
permeability coefficient was also calculated for the steady state using Equation (13) and

can be seen in the table.

60

Fig. (6)

2,4 Dichlorophenol

Transmission Profile Curve Obtained from the Quasi-isostatic
Test Procedure

 

 

 

Transmission Profile Curve
2,4, Dichlorophenol

 

 

 

 

 

 

0.012
0.01 _
‘8' 0 008
m a . F
i a
a} g 0.006 _
z‘ 3
'a 2
g I- 0.004 _
3
o 0/
0 002 _ ,/
/¥
0 e er/ 1 =
0 5 10 15 20
Thousands
Time (seconds)
Regression Output:
Constant -4.8E-06
Std Err of Y Est 4.8E-07
R Squared 0.977815
No. of Observations 4
Degrees of Freedom 2
X Coefficient(s) 7.1E-10
Std Err of Coef. 7.5E-11
Lag Time 7100

61

 

4.1.4 Incorporation of the Diffusion Coefficients into the Mathematical Models :
Begley and Hollifield model
a) Xylene

The experimentally obtained diffusion coefficient for xylene using the isostatic
method was incorporated into the Begley and Hollifield model for the prediction of
migration. The results of the prediction of migration using this model are depicted in
Table (3). The column “Mtr” denotes the prediction of migration of xylene through unit
surface area of packaging (ug/cmz) through a single layer recycled HDPE film using
Equation (1). This is denoted as a function of time. The initial concentration of xylene
was assumed to be 120ug/cc which is the value obtained from the PRTF Guidelines,1993
as discussed in Chapter 2. The thickness of the recycled layer is assumed to be 1 mil. The
values of the variable “1:” are also depicted in the table. The model states that at values of
1' >06, the predicted migration from the two layer structure approaches that from a single
layer structure and exceeds that of the single layer at higher values. As discussed in
Chapter 2, the predicted migration results from this model for a barrier structure
incorporating a food contact virgin layer are not applicable at values of 1: >06. Since in
this situation, the value of 1: is greater than 0.6 on the first day itself, for these conditions,

the expression will not be valid for the two layered barrier structure.

62

Table (3)

Xylene

Prediction of Migration Using the
Begley and Hollifield Model

For a Monolayer Recycled Structure
Thickness of recycled Iayer=1 mil

 

 

 

Time Mtr Tau
(days) (ug/sgcm) (dt/l2)
0 0.00 0.00
* 0.38 1.23
2 0.54 2.46
3 0.66 3.70
4 0.76 4.93
5 0.85 6.16
10 1.21 12.32
15 1.48 18.48
20 ‘ 1.71 24.64
25 1.91 30.80
30 2.09 36.96
35 2.26 43.12
40 2.41 49.28
45 2.56 55.44
50 2.70 61.60

 

 

 

 

MtF-Migration through recycled layer
into the food.

Initial Concentration = 120ug/cc

* All the available contaminant migrates
after the first day. Prediction of
subsequent migration is invalid.

63

In order to reduce 1:, and to consequently make the model work for the prediction
of migration for a barrier structure, either the diffusion coefficient has to be decreased or
the thickness of the polymer has to be increased. D is characteristic of the contaminant!
polymer system and hence is a constant for this system, and we are interested in a
minimum shelf life of the food of at least 30 days for the prediction of migration.
Therefore in order to bring the value of 1: lower, so that the equation is valid for a two
layer structure, the thickness of the structure will have to be increased substantially. Table
(4) illustrates this, showing a reduced value of 1: corresponding to an increased thickness
of the virgin layer by 10 times the original thickness (i.e. 10 mils). This model does not
consider the thickness of the recycled layer in the prediction of migration. According to
this model, the migration of a compound in a barrier structure containing virgin layer is
assumed to be dependent only on the thickness of the virgin layer and is independent on
the thickness of the recycled layer. The results of the predicted migration from the two
layer structure and the monolayer structure is illustrated in Table (4). The predicted
migration from the two layer structure was calculated using Equation 2. described in
Chapter 2. Though the equation involves a summation of terms from 1 tom, the
calculations were done for the first 6 terms only, since the contribution from the
successive terms were almost negligible. If fewer terms are used, the predictions of

migration for the initial few days will not be accurate as per the equation. The predicted

64

Table (4)

Xylene

Prediction of Migration Using the
Begley and Hollifield Model
Two Layer Barrier Structure
Thickness of the Virgin Layer = 10 mils

 

 

 

Time Mtr Mtv Tau Dietary
(days) (ug/sqcm) (uglsqcm) Conc. (ppb)
0 0.00 0.00 0.00 0.00
1 0.38 0.00 0.01 0.00
2 0.54 0.00 0.02 0.00
3 0.66 9.5E-05 0.04 0.02
4 0.76 0.0007 0.05 0.15
*5 0.85 0.002 0.06 *0.55
10 1.21 0.05 0.12 9.88
15 1.48 0.15 0.18 30.97
20 1.71 0.30 0.25 59.47
25 1.91 0.46 0.31 92.08
30 2.09 0.63 0.37 126.94
35 2.26 0.82 0.43 163.03
40 2.41 1.00 0.49 199.78
45 2.56 1.18 0.55 236.90
50 2.70 1.37 0.62 274.22

 

 

 

 

 

Mtr = Migration through Food Contact Layer in mono-layer structure

Mtv = Migration through virgin layer in a two -layered structure
Initial Concentration =120ug/cc
Dietary Concentration in ppb

* Dietary Concentration

65

 

dietary concentration was also calculated from the expression in Chapter 2, which
involves multiplying the migration values by the consumption factors corresponding to
HDPE. It can be seen from the table that even though the thickness is increased to 10 mil,
the barrier will not prevent the migration of the contaminant to a value that is below the
threshold level as specified by the FDA. Part of the explanation for this could be due to
the assumption of the model that the concentration of the compound in the package is
constant even though it is obvious that in real life the concentration will decrease with
time as the compound migrates into the food. For example, the amount of contaminant
available to migrate from this barrier structure is lZOug/cc. This amount has already
migrated at approximately 100 days as can be derived from Table 4. Hence there must be
no more contaminant remaining in the polymer to migrate into the food. However,
because of the assumption that the concentration in the polymer is a constant, there is
continued predicted migration in this model which is contrary to what is possible in real
life situations. AS discussed in Chapter 2, this will give unrealistically high prediction of
migration, especially as the time increases. Also the predictions of migration in this case
after a period of 100 days are totally invalid. The same justification will also apply to the
single layer package structure in Table (3) where all the contaminant would have
migrated out of the structure at the first day itself. This situation will aggravate even

further with compounds with high diffusion coefficients such as xylene and 2,4

66

dichlorophenol.

On the other hand, if the assumptions of this model are utilized to predict the
dietary concentration, it indicates that the virgin layer will have to be very thick in order
to prevent substantial migration from the recycled layer into the food. This will not be
very useful from the economic, as well as environmental aspect because in such a
situation we are likely to end up using more polymer in order to make the barrier more
effective. This will not only increase the cost involved but will also defeat the purpose of
the environmental concerns of using more polymer.

But as discussed, we have to consider the fact that the prediction is not at all
accurate because of the assumption of constant concentration.

Fig. (7) shows a comparison of the predicted migration through a monolayer and

two layer package structure for a thickness of 10 mils.

b) 2,4, Dichlorophenol

For 2,4, dichlorophenol, which has a slightly higher diffusion coefficient, the
results of the prediction of the model are similar to that of the prediction of xylene
obtained above. Table (5) illustrates the results obtained for predicted migration from the

monolayer recycled HDPE structure. Here again, the initial concentration of the

67

Fig (7)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Xylene
Comparison of Migration Through a Single Layer and
Two layer Structure
Comparison Of Migration]
4
a 3 ~ .//'
g H
fit ‘Xr A
: //"
a: X
g” I Migration Through MonoLayer
. Migration Through Two layer
. . l
0.6 0.8 1
Tau

 

 

68

 

F19 (8)
2,4 Dichlorophenol

Comparison of Migration Through a Single Layer and
Two layer Structure

 

 

 

Comparison of Migrati0n|

 

 

 

 

 

 

 

 

 

0.3
A 0.25 L .-
g I/
0’ 0.2 — -
10
g: /I/
v 0.15 _ /
a //
E, 0-1 ” . Migration Through Monolayer
2 0 05 9 Migration Through Twolayer
//
O 1 L 1 1
0 0.2 0.4 0.6 0.8
Tau

 

 

69

Table (5)
2,4, Dichlorophenol

Prediction of Migration Using the
Begley and Hollifield Model

For a Monolayer Recycled Structure
Thickness of the Recycled Layer =1 mil

 

 

 

 

 

 

Time Mtr Tau
(days) (us/cc)

0.00 0.00 0.00
* 0.05 2.68
2.00 0.06 5.36
3.00 0.08 8.04
5.00 0.10 13.39
10.00 0.14 26.78
15.00 0.17 40.18
20.00 0.20 53.57
25.00 0.23 66.96
30.00 0.25 80.35
35.00 0.27 93.74
40.00 0.28 107.14
45.00 0.30 120.53
50.00 0.32 133.92

 

Mtr = Migration Through Food Contact Layer
Initial Concentration = 9.6ug/cc

* All the available contaminant migrates
after the first day. Prediction of

subsequent migration is invalid.

7O

contaminant was assumed from the PRTF guidelines as mentioned above. The initial
concentration of 2,4 dichlorophenol in the polymer was assumed as 9.6 ug/cc. Also for
this situation, the thickness of the recycled layer was assumed to be one mil. In this case,
again, assuming a thickness of one mil, the value of 1: is very high and the model is not
valid for the prediction of migration through a two layered structure at this thickness.
Also at a thickness of 1 mil, all the available contaminant would have migrated in the first
day, but because of the assumption of constant concentration, the model keeps showing
migration even after all the contaminant has left the polymer. When the thickness of the
virgin layer is increased to 10 mil, the value of 1 decreases, and the model can be used to
predict the migration from the two layer barrier structure incorporating a functional
barrier. Table (6) illustrates the prediction of migration through a two-layer structure with
a functional barrier of thickness of mil. Again, all the available contaminant would have
migrated out of the structure in 50 days making the model totally invalid after that period.
Though the equation involves a summation from O to co, only the first six terms are used
since the contribution from the remaining terms is negligible. Here again we can see that
the assumption of constant concentration is not very accurate, leading to an over
estimation in migration. If this model is followed then it demonstrates that increasing the
thickness of the virgin layer does not help reducing the dietary concentration of the

contaminant below the threshold value. Fig. (8) shows a comparison of the predicted

71

Prediction of Migration Using the

Table (6)

2,4, Dichlorophenol

Begley and Hollifield Model
Two Layer Barrier Structure
Ticknoss of the Virgin Layer = 10 mils

 

 

 

Time Mtr Mtv Tau Dietary
(days) (uglsqcm) (uglsqcm) Cone. mpb)
0.00 0.00 0.00 0.00 0.00
1.00 0.05 0.00 0.03 0.00
2.00 0.06 4.8505 0.05 0.01
3.00 0.08 0.00079 0.08 0.16
5.00 0.10 0.01 0.13 1.03
10.00 0.14 0.03 0.27 5.64
15.00 0.17 0.06 0.40 11.65
20.00 0.20 0.09 0.54 18.05
25.00 0.23 0.12 0.67 24.54
30.00 0.25 0.16 0.80 31.06
35.00 0.27 0.19 0.94 37.59
40.00 0.28 0.22 1.07 44.12
45.00 0.30 0.25 1.21 50.65
50.00 0.32 0.29 1.34 57.18

 

 

 

 

 

Mtr = Migration through Food Contact Layer in mono-layer structure

Mtv = Migration through virgin layer in a two -layered structure
Initial Concentration =9.6 ug/cc
Dietary Concentration in ppb

72

 

migration through a monolayer and two layer package structure with a 10 mil thick
recycled layer and a 10 mil thick virgin layer.

From these results we can conclude that the functional barrier for HDPE
would be effective in substantially controlling migration only if the thickness of the
barrier layer is very high. Since the prediction of migration from the model is a gross over
estimation of the actual migration due to the assumption of constant concentration, it is
obvious that in real life, the migration would not be as high as is predicted and using the
results from this model may be very erroneous. Actual migration studies have to be done
in order to confirm this. However it is very likely that the results obtained from this over-
simplified model will not be applicable to real life migration studies and it would either

have to be modified further or other models would have to be considered.

4.1.5 The Laoubi and Vergnaud Model

The predicted migration for both the compounds were also calculated using the
model pr0posed by Laoubi and Vergnuard. This model takes into account only a two
layered structure incorporating the virgin barrier. The thickness of the recycled layer and
the barrier layer is taken as one mil respectively. The predicted migration results for

xylene are shown in Table (7), and the predicted migration results for 2,4, dichlorophenol

73

Table (7)

Xylene

Laoubi and Vergnuad Model

Prediction of Migration Through a Two Layer Structure Incorporating a

Functional Barrier

Thickness of Recycled Layer=1mil, Thickness of Virgin Barrier=1 mil

 

 

 

Time Sum Mt Mtr Mtv Mtf
Qays) (Mt/Mi) (uglcm3) (gglcm3) (uglcm3L (u Icm3L
0 1.00 120.20 113.81 6.39 0.00
1 0.54 64.56 45.64 18.93 55.44
5 0.03 3.13 2.21 0.92 116.87
10 0.00 0.07 0.05 0.02 119.93
15 0.00 0.00 0.00 0.00 120.00

 

 

 

 

 

 

MFConcentration of the contaminant in the entire packaging
Mtr=Concentration of the contaminant in the recycled layer.
Mtv=Concentration of the contaminant in the virgin layer.
Mtf= Concentration of the contaminant leaving the virgin layer into the to
Initial concentration =120 uglcc

74

 

Table (8)

2,4 Dichlorophenol

Laoubi and Vergnuad Model

Prediction of Migration Through a Two Layer Structure Incorporating a
Functional Barrier

Thickness of Recycled Layer=1 mil, Thickness of Virgin Barrier=1 mil

 

 

 

 

Time Sum Mt Mtr Mtf

(Days) (Mthi) (rig/emit) (uglcm3) (uglcm3) (uflm3)
0 0.97 9.35 9.75 0.00 0.25
1 0.22 2.11 1.63 0.48 7.49
2 0.04 0.41 0.31 0.09 9.19
3 0.01 0.08 0.06 0.02 9.52
4 0.00 0.01 0.01 0.00 9.59
5 0.00 0.00 0.00 0.00 9.60

 

 

 

 

 

 

Mt=Concentration of the contaminant in the entire packaging.
Mt=Concentration of the contaminant in the recycled layer.

Mtv=Concentration of the contaminant in the virgin layer.

Mtf= Concentration of the contaminant migrating from the polymer into th
Initial concentration =9.6 uglcc

75

 

Table (9)

Xylene

Laoubi and Vergnaud Model

Prediction of Migration Through a Two Layer Structure Incorporating a
Functional Barrier (10 mil recycled layer; 10 mil virgin layer)

 

 

 

 

Time Mt/Mi Mt Mtr Mtv Mtf Dietary
(days) uglcc ufic uglcc uglcc Conc.(ppb)
1.00 120.48 113.81 6.39 0.00 0.00
1 1.00 120.09 111.10 9.05 0.00 0.00
5 1.00 120.01 103.10 16.91 0.01 0.07
10 0.99 119.10 96.24 22.86 0.90 8.97
15 0.98 117.05 90.92 26.12 2.95 29.53
20 0.95 114.26 86.43 27.83 5.74 57.36
25 0.93 111.08 82.48 28.60 8.92 89.22
30 0.90 107.68 78.90 28.78 12.32 123.17
35 0.87 104.20 75.61 28.58 15.80 158.05
40 0.84 100.69 72.55 28.13 19.31 193.14
50 0.78 93.77 66.96 26.81 26.23 262.27
60 0.73 87.15 61.93 25.22 32.85 328.47
100 0.54 64.56 45.64 18.93 55.44 554.37
150 0.37 44.24 31.25 12.98 75.76 757.65
200 0.25 30.30 21.41 8.89 89.70 897.00
365 0.07 _ 8.69 6.14 2.55 111.31 1113.07

 

 

 

 

 

 

MFConcentration of the contaminant in the entire packaging.
Mt=Concentration of the contaminant in the recycled layer.
Mtv=Concentration of the contaminant in the virgin layer.
Mtf= Concentration of the contaminant in the polymer migrating into the food.
Initial concentration =120 uglcc

76

 

Table (10)

2,4 dichlorophenol

Laoubi and Vergnaud Model

Prediction of Migration Through a'l'wo Layer Structure Incorporating a
Functional Barrier (10 mil recycled layer; 10 mil virgin layer)

 

 

 

 

 

Time Mt/Mi Mt Mtr Mtv Mtf Dietary
(days) uglcc uglcc uglcc glee Conc.(ppb)
0.00 1.00 9.60 9.10 0.50 0.00 0.00
1.00 1.00 9.60 8.67 0.94 0.00 0.00
2.00 1.00 9.60 8.33 1.27 0.00 0.03
3.00 1.00 9.59 8.06 1.53 0.01 0.11
4.00 1.00 9.56 7.82 1.73 0.04 0.43
5.00 0.99 9.50 7.62 1.89 0.10 0.96
10.00 0.94 9.05 6.80 2.26 0.55 5.48
11.00 0.93 8.94 6.66 2.28 0.66 6.61
15.00 0.88 8.46 6.17 2.30 1.14 11.36
20.00 0.82 7.85 5.64 2.22 1.75 17.47
25.00 0.76 7.26 5.17 2.09 2.34 23.41
30.00 0.70 6.70 4.75 1.95 2.90 29.03
35.00 0.64 6.17 4.37 1.80 3.43 34.27
40.00 0.59 5.69 4.02 1.66 3.91 39.13
50.00 0.50 4.82 3.41 1.41 4.78 47.76
. 60.00 0.43 4.09 2.89 1.20 5.51 55.10
I 70.00 0.36 3.47 2.45 1.02 6.13 61.32
I 100.00 0.22 2.11 1.49 0.62 7.49 74.86
i 150.00 0.10, 0.93 0.65 0.27 8.67 86.74

 

 

 

 

 

 

Mt=Concentration of the contaminant in the entire packaging.
MFConcentration of the contaminant in the recycled layer.
Mtv=Concentration of the contaminant in the virgin layer.
Mtf= Concentration of the contaminant in the polymer migrating into the food.
Initial concentration =9.6 uglcc

77

 

are shown in Table (8). For the prediction of migration, Equation (6), described in
Chapter 2, as used. As seen from the equation, a summation of terms is used. We used
only the first 6 terms in the expression since the contribution made by the consecutive
terms is almost negligible.

Here again we can see that the virgin polymer does not act as a functional barrier
at a thickness of one mil of recycled polymer and one mil of virgin polymer. For both the
compounds, almost the entire contaminant migrates out of the polymer in less than 10
days. Unlike the Begley and Hollifield model, this model does take into account the
decreasing concentration in the recycled layer and does not predict any further migration
once all the contaminant has left the polymer. To observe the effect of the migration with
an increase in the thickness of the barrier layer, the thickness of the barrier layer was
increased by 10 times (i.e.10 mils ) as done in the previous model. Tables (9) and (10) are
predictions of migration when the thickness of the recycled layer and the functional
barrier is increased to 10 mils each.

From the tables we can see that though the migration into the food reduces
substantially with an increase in the thickness, the dietary concentration is still much
above the threshold of regulation specified by the FDA even at 10 days. Hence even
according to this model, the functional barrier does not prove to be effective even at a 10

mil thick virgin layer in spite of the fact that the predicted migration from this model

78

is much lower compared to the Begley and Hollifield model. This is mainly because it
does not assume constant concentration of the contaminant throughout, but takes into
account that the concentration in the package will decrease as a function of time. Hence it
is more likely to give results that will be closer to real life situations. Still, it does not act
as a functional barrier in keeping the dietary concentration at levels below the threshold
value even at higher thicknesses. The Laoubi and Vergnaud model also gives the
concentration of the contaminant through the entire package structure. It gives the
concentration of the contaminant in the recycled layer, virgin layer, and food as a
function of time. Tables (9) and (10) show the concentration of the contaminant in the
different layers as a function of time. Mtr, Mtv, and Mtf are the concentration of the
contaminant in the recycled layer, functional barrier and in the food. Fig (9) and Fig (10)
are plots of the profiles of the concentration of the contaminant with respect to time,
through the two layers in the structure, and into the food. As is expected , it is seen that
for both the compounds, the concentration of the contaminant reduces gradually over a
period of time in the recycled layer as the contaminant diffuses into the virgin layer. In
this case, as against the Begley and Hollifield model, the concentration in the recycled
layer is not constant and decreases over a period of time. The concentration in the virgin
layer first increases and then decreases as the contaminant migrates from the virgin layer

into the food. The concentration of the contaminant in the food increases gradually with

79

time as the contaminant migrates from the functional barrier into the food. These profiles
of the contaminant transfer in the system are similar to the profiles obtained by the
researchers, Laoubi and Vergnaud, by inserting theoretical values of diffusion

coefficients into the mathematical model.

4.1.6 Comparison of the Two Models

The Begley and Hollifield model assumes that the concentration does not change
with time, whereas the Laoubi and Vergnuad model takes into account that the
concentration in the recycled layer will reduce with time. Hence, the predicted migration
from the Begley and Hollifield model is expected to be higher than that of the former as
discussed earlier. This is also the kind of observation that has been made in this study and
a comparison of the predicted migration between the two models can be seen in F ig. (9)
and Fig. (10) for both xylene and 2,4 dichlorophenol respectively. Also, the Begley and
Hollifield model does not take into account the thickness of the recycled layer for the
prediction of migration and assumes that the migration depends only on the thickness of
the barrier layer. In order to predict the amount of the contaminant that will migrate
through the polymer, it is also important to take into consideration the thickness of the

recycled layer as is done in the Laoubi and Vergnaud model. For the above two reasons

80

Fig (9)

Xylene
Concentration of the Contaminant in the Package System as Function of Time

Thickness of the Recycled Layer =10 mil, Thickness of Virgin Layer=10 mil

 

 

 

 

 

 

 

 

 

 

14° Concentration of Contaminant r‘
120 '
l
100 '
g . recycled
E 60 __ .virgin
g A in fOOd
40 ~—
20 \
1
o - 1 l
O 100 200 300 400
Time (days)

 

 

 

81

Fig (10)

2,4 Dichlorophenol

Concentration of the Contaminant in the Package System as Function of Time

 

 

uglcm2

 

 

Concentration of Contaminant |

 

 

   

I recycled
. virgin

 

 

 

 

tin fOOd

 

82

 

Fig (1 1)
Xylene

Comparison Of Predicted Migration Between the Two Models for a two layer
Barrier Structure with a Virgin Layer of 10 mil

 

 

l
l
I

Comparison Between the Models]

 

 

 

 

 

   

Migration (uglsqcm)

. Begley and Hollifield
. Laoubi nad Vergnaud

 

 

 

0 50 100 150 200
Time (days)

 

 

 

83

Fig (12)
2,4 Dichlorophenol

Comparison Of Predicted Migration Between the Two Models for a two layer
Barrier Structure with a Virgin Layer of 10 mil

 

 

 

Comparison Between The Models]

 

1

 

 

 

 

 

 

 

 

 

A 0.8 _
E
8
m 0.6
8’ I . Begley and Hollifield
c Laoubi and Ve naud
.2 0.4 .— ° '9
E
..C.”
2 0.2 A.

0 g 1 l

0 50 100 150 200

 

Time (days)

 

 

84

the Begley and Hollifield model will predict more migration than the Laoubi and

Vergnuad model.

85

5. CONCLUSION

5.1.1 Summary

The study involved determining the permeability experiments of the selected
compounds, xylene and 2,4 dichlorophenol using the isostatic and quasi-isotatic test
methods. The diffusion coefficients were obtained from these tests using the steady state
permeation data.

Mathematical models were used to predict migration of both the compounds
through a post consumer recycled HDPE film using the experimental values of the
diffusion coefficients. The initial concentration of the compounds were taken from the
PRTF Guidelines, on the basis of the contamination most likely to occur in post
consumer recycled plastics. The mathematical models were employed for single layer
recycled HDPE structures, and two layer package structures involving a recycled HDPE
layer and a food contact layer of virgin HDPE acting as a ‘funetional barrier.’ The models
used follow simplified assumptions which predict more migration than is likely to occur
in real life situations.

The prediction of migration obtained from the Begley and Hollifield model was

observed to be higher than that from the Laoubi and Vergnuad model. This result was

86

expected because of the difference in assumptions used by the models.

The results of the predicted migration were compared for the two layer and
single layer package structures. The study also used the results of the predicted migration
to determine the dietary concentration in the food based on the parameters set up by the
FDA. Based on the results of the models alone, and not considering the simplifications
made by them, we find that the functional barrier does not prove to be very effective in
keeping migration below the threshold level. In order for the barrier to be effective, its
thickness will have to be increased substantially. This will in turn increase both the cost,
as well as the consumption of the recycled polymer thus defeating the objective of
reducing recycling waste.

However, as discussed earlier, we have to also take into consideration the fact that
both these models tend to over- estimate migration. This is more evident in the Begley
and Hollifield model. Due to this over-estimation, there will also be a subsequent over-
estimation in the prediction of the dietary concentration of the contaminant. The increase
in the predicted dietary concentration of the contaminant is also compounded by the
incorporation of safety factors in determining the consumption factors which are used to

predict dietary concentration of the contaminant.

87

5.1.2 Recommendations and Future Work:

The likelyhood of the over estimation of predicted migration makes it important
to perform actual migration studies with these compounds. The next step in this project
would be to perform actual migration studies using doped films with the compounds at
the recommended initial concentrations used. The migration studies can be done using the
standard food simulating solvents. An attempt should also be made to try to modify the
models so that they represent a more real life situation.

The models also assume that there is no resistance to the transfer of the
contaminant into the food once it reaches the food contact surface. This generally
represents a worst case situation and may be true for fatty products but may not be true
for types of food such as dry foods. In such a situation there will once again be an over-
estimation of migration.

Also another point that has to be taken into account is that the compound will be
able to diffuse through the non food contact end of the package, thus reducing the
available concentration of the compound that can move into the food product, leading to
lower migration of the compound into the food. Another important factor to be
considered is the time intervals between the manufacture of the package and filling it with

food. These time intervals may be large, up to many months, or very short, up to a few

88

minutes. In the first case the migration may start out of the package even before it is filled
with food. Hence it would not be right to assume that migration of the compound starts
when the package is filled with the food as this might lead to an under-estimation of
migration. The temperature of storage or use is also going to play a large part in the
migration of the contaminant into the food and also has to be considered while predicting
migration.

Though the over estimation of migration of the models gives a “safety factor”, a
gross over -estimation will most definately lead to increased use of polymer and hence
increased cost. Hence it is important to determine the extent of overestimation with actual
migration studies. If the results of the migration studies are close to the predicted
migration by the models, then it will be possible to use these models as a rough estimate
in cases where there is not enough time to perform the actual tests. However an
agreement of the migration tests with the predictions also mean that for these particular
compounds, though the models are good for predictions, there is no advantage of
incorporating a functional barrier since it does not help in reducing the migration even at
high thicknesses of the polymer film. Using a higher thickness of the polymer film will
only lead to the use of more volumes of polymer which is against our aim of reducing

plastic waste.

89

APPENDIX (1)

(A)

l. Xylene
1.Gas Chromatograph Conditions

HP 5890A Gas Chromatograph
Column : Supelcowax

Carrier Gas : Helium

Detector : Flame Ionization
Detector Temperature : 280 deg C
Injector Temperature : 220 deg C
Injection Volume : lul

2. Temperature cycle

Initial temperature - 45deg C
Initial Time - 5 min

Final Temperature- 200 deg C
Final Time - 5 min

Rate- 12 deg C/ min

90

APPENDIX (1)

(B)

2,4 Dichlorophenol
1.Gas Chromatograph Conditions

HP 5890A Gas Chromatograph
Column : SPB-5

Carrier Gas : Helium

Detector : Flame Ionization
Detector Temperature : 280 deg C
Injector Temperature : 220 deg C
Injection Volume : lul

2. Temperature cycle

Initial temperature - 45deg C
Initial Time - 5 min

Final Temperature- 200 deg C
Final Time - 5 min

Rate- 12 deg 0 min

91

Appendix 2

Fig.(1)

Calibration Curve For Xylene

 

 

 

Calibration Curve - Xylene
Quantity Injected Vs Area Response

 

 

/.

,/’

 

 

. Quantity injected ]

 

 

 

Lg

 

014
(112 L
3’ 0.1_
'O
93m
'35
EEO'OGF
‘E
5" 004A
C!
(102 —
0 r
0 50

150 200 250 300
Thousands

Area Response

100

 

Calibration Factor = Slope = 0.000557 (g/AuTI

 

 

 

 

 

 

 

 

 

 

Cone Area Quantity
(ppm) Response Injected
(Au) (9)
10 36620 1.674E-08
20 65671 3.348E-08
40 123781 6.698E-08
80 246782 1.339E-07
Regression Output:
Constant -3.0E-09
Std Err of Y Est 9.4E-10
R Squared 0.999779
No. of Observations 4
Degrees of Freedom 2
X Coefficient(s) 5.6E-13
Std Err of Coef. 5.9E-15

92

 

Appendix 2

Calibration curve for 2,4, Dichlorophenol

 

 

 

Calibration Curve : 2,4 Dichlorophenol

Quantity Injected V/s Area Response

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4 . Quantity Injected ]
e 3 ~ /-
E .
8 5
E g 2 r-
a a
if:
5 1 ~ //
0 l I 1 1 l 1
0 2 4 6 a 10 12 14
Thousands
Area Units
alibration Factor = Slope = 2.3 E-13(gIAU]]
Table (x)
Cone. Area Quantity
(ppm) Response Injected
(AU) (9)
200 1160.2 3.8E-10
400 2409 7615-10
800 5769 1 .55-09
1600 12256 3.0E-09
Regression Output:
Constant 1.5E-10
Std Err of Y Est 4.5E-11
R Squared 0.998994
No. of Observations 4
Degrees of Freedom 2

X Coefficient(s)

Std Err of Coef.
CALIBRATION FACTOR = 2.3 E-13 (gIAu)

2.3E-13
5.2E-15

93

 

Appendix 3
2,4 dichlorophenol

Transmission Profile Curve Obtained From
Quasi Isostatic Test Procedure

 

 

 

Transmission Profile Curve]

 

 

 

 

 

 

 

 

 

 

 

 

0.008 ,
g 0.006 _
E ..
m .c
“e’ a
5 g 0.004 — ] 9 Quantity Permeated (g)
D- 3
IE :5
E
g 0.002 e
o
0 0 o .4/ r m
0 5 10 15 20
] Thousands
1 Time (secs)
Regression Output:
Constant -9.4E-06
Std Err of Y Est 8.8E-07
R Squared 0.942331
No. of Observations 4
Degrees of Freedom 2
X Coefficient(s) 1.0E-09
Std Err of Coef. 1.8E-10

94

BIBLIOGRAPHY

Begley, T. and Hollifield, H., “Recycled Polymers in Food Packaging: Migration
Considerations,” Emdjechnglggy, Nov. 1993, pp.109-112.

Crank, J and Park, G.S., W, Academic Press, New York.

Division of Food Chemistry and Technology, Center for Food Safety and Applied
Nutrition, US Food and Drug Administration, “Points to Consider For the Use of
Recycled Plastics in Food Packaging:Chemistry Conditions.” 1992.

Franz R., “Testing and Evaluation of Recycled Plastics for Food Packaging Use -

Possible Migration Through a Functional Barrier, MW, vol
11, no.4 pp. 479-496, 1994.

Franz R, “Permeation of Volatile Organic Compounds across Polymer Films - Part -1
:Development of a Sensitive Test Method Suitable For high Barrier Packaging Films at

Very Low Perrneant Vapor Pressures. WWW, Vol 6., pp91-
102,1993.

Hoojat, P, Harte, B. R. ;,Hemandez R. J.; Giacin, J. R and Miltz, J., “ Mass Transferof
BHT from HDPE and its influence on Product Stability. ”W2.

leehnglggymoll, no.,3 pp78- -,81 1987.

Hilton, B.W. and Nee, S. Y., “Permeability of Organic Vapors Through
Packaging Films,” We 17(1), pp 80, 1978

Huang, 8., “Evaluationg Isostatic and Quasi Isostatic Procedures for Determining
Organic Vapor Barrier Properties Through Polymer Membranes,” Michigan State
University, 1996.

Katan, L.L., “Do ‘Funetional Barriers’ Function 7,” WW
5.012119; Vol 9,1996 pp.289-296.

Laoubi, S., and Vergnaud, J. M, “Process of Conataminant Transfer Through a Food
Package Made of a Recycled Film and a Functional Barrier.” Eagkagmgleghnglowd
ngenge, vol 8, pp. 97- 110, 1995.

Laoubi, S., et al., “Effect of Functional Barrier Thickness for a Food Package with
Recycled Polymer,” Wm, vol 8, 1995, pp. 249- 259.

95

Miller ,Chaz, “ Recycled PET Approved by FDA for Use in More Food
Packages,” Recyclinglimes, June 29,1993.

Miltz, J., et a1. ,“Mass Transfer From and Through Packaging Materials,”
Baekaainaleclmolmanflcienee vol 5 1992 PP 49- 56

Overton, B. W. ,“Impact of Recycling and Environmental Legislation,” E991
AddinyeaandQQmMnants vol 11 1102 1994 PP 235-293

Pasternak, R. A, et al., “A Dynamic Approach to Diffusion and Permeation
Measurements, WEEK—59191199, Part A-2, vol. 8, 1970, pp. 467.

Wis: vol 14 n01 1996 p06

‘ I ., _ o—L'. 9"“... S E 1‘.'-

Plastics Recycling Task Force, National Food Processors Association and Society of the
Plastics Industry,” Guidelines for the Safe Use of Recycled Plastics For Food Contact
Applications,” 1994.

Thorsheim, H., and Amstrong, D., “Recycled Plastics for Food Packaging,” Aug., 1993,
pp 55-58.

Tice, P. A. “Pira Project on Migration of Monomers and Overall Migration, Food
Addrtmsandfmntammants 5(SUPP1 1): PP 373- 330 1933

96

ADDITIONAL SOURCES

Ashby, R., Migration from PET under all Conditions of Use,ӣogd_Additiyes_and
Contaminants, 5 (Suppl): 485-492, 1988.

Barlas S., “FDA Plans to Speed Approval for PCR Use,” Eagkagingflqfld, June 1995,
pp. 24,27

Baner, A. L.; Hernandez R. J. ,Jayaraman K. and Gaiein, J. R., “Isostatic and Quasi
Isostatic Methods for Determining the Permeability of Organic Vapors Through Barrier

Membranes.” Current Wing, M. L. Trodel (Ed) American
Society for Testing Materials, Philadelphia, PA. 1984. pp. 49-62.

Blakesley, C.N., “Organic vapor Permeation Through Packaging Films.” LEQQd
Ieghnglggy, vol. 10 , 1975, pp. 365-367.

Blakesley, C.N., “Kinetic Considerations of the Permeation of Organic Vapors

Through a Flexible Packaging Film, 1mm vol. 9, 1974, pp. 285 -
296.

Boone, J., et al., “Deficiencies of Polypropylene in its Use as a Food Packaging
Material- a Review ” RackemnaleshnoleuaniScience vol 6 1993 PP 277- 231

Bruck, C. G., “Possible Intake of Food Packaging Ingredients by Consumers,’ ’_F_o_od
Additiyasandfiontaminants 5 (SUPPl 1) 363- 372 1988

Crosby, N.T. WW, Applied Science Publishers Limited,
London. (1981).

Figge, K. and Rudolph, F., “Diffusion on The Plastics Packaging/ Contents
System,”Angmand_te_Mnkmmnleknlare_Chemie, vol.78, no.1169, 1979. pp.157- 180.

Foulke J ., “Recycled Plastics,” EDALQnsnmer November, 1991, pp.

97

Hotchkiss, J .H. “Overview on Chemical Interactions Between Food and

Packaging Materials,” WNW TheRoyal
Society of Cambridge. (1992).

Katan, L.L, “Threshold of Regulation - Comment and Report,” Packaging
W, vol' 8, 1995, pp.29-32.

Laoubi, S. ,“Safety of Recycled Plastics for Food Contact Materials. Testing to Define
a Functional Barrier,” WW, vol. 8 pp. 17-27, 1995.

Plastics Recycling Task Force, Technical Committee, “Protocol for Evaluationg
Effectiveness of Plastics Reclamation Process,”Sept. 15, 1994.

Risch, S.J., Migration of Toxicants, Flavors, and Odor-Active Substances from
Flexible Packaging Materials to food,” W, July 1988, pp. 95-102.

Schwartz, P., “Food Packaging Regulations in the United States,” EoodAddjfim and
Contaminants, vol. 5(Suppl. 1), 1988, pp.537-541.

98

 

"IIIIIIIIIIIIIII