v. u . .... u-I‘cfiuufln

HESIS

Mimiminimum:

01707 4141

This is to certify that the

thesis entitled

ANALYSIS AND EVALUATION OF BURST TEST METHODS USING
RESTRAINING FIXTURES

presented by

ROSAMARI FELIU-BAEZ

has been accepted towards fulfillment
of the requirements for

MASTER degree in PACKAGING

Major professor

DateMU/Ml/ /Cl /9?§/

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

 

 

 

‘ LIBRARY

, Michigan State
Unlverslty

 

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

DATE DUE MTE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M www.mna

ANALYSIS AND EVALUATION OF BURST TEST METHODS USING
RESTRAINING FIXTURES

By

Rosamari Feliu-Btia

A THESIS

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

MASTER OF SCIENCE

School of Packaging

1998

ABSTRACT

ANALYSIS AND EVALUATION OF BURST TEST METHODS USING
RESTRAINING FIXTURES

By

Rosamari Feliu-Baia

Three restrained burst tests were performed: two for blisters and one for pouches.
For both, blisters and pouches, four basic behaviors were found. First, burst pressure
varies inversely with package size. Second, unrestrained burst pressures are lower than
the restrained burst pressures. Third, burst pressure is inversely proportional to plate
separation. Fourth, restraining fixtures do not necessarily reduce variability or improve
repeatability. In fact, different patterns were found for raw variance and coefficients of

variation for unrestrained and restrained burst tests results for blisters and pouches.

Another experiment was performed with the purpose of correlating burst peel
strength and tensile peel strength for Tyvek/Plastic chevron seal pouches. Correlation
between burst peel strength and tensile peel strength could not be confirmed even though

burst peeling times and tensile peeling times were controlled to be the same.

To my husband, Ruben

iii

ACKNOWLEDGNIEN TS

I want to express my gratitude to my major professor, Dr. Hugh Lockhatt for all
his advice, support and help in the development of this research project. I would like also
to thank Dr. Gary Burgess and Dr. Dennis Gilliland for being part of my committee and
for all their contribution to this project. Their assistance was of great value.

I also want to thank my friends and the packaging graduate students for their
support. Special thanks to my parents, grandparents, my two brothers, my sister, and
other family members for all their love and encouraging words during the hard days. I
want to express my sincere gratitude to my husband Ruben for his love, support,
patience, and encouragement. Finally, I want to thank God for giving me the strength to

finish this project.

iv

TABLE OF CONTENTS

LIST OF TABLES ...............................................................
LIST OF FIGURES ...............................................................
CHAPTER 1. INTRODUCTION ..............................................

CHAPTER 2. LITERATURE REVIEW ......................................
Restrained Burst Test Methods for Blisters ..........................
Restrained Burst Test Methods for Pouches ............... . .........
Restrained Burst Test Studies Performed at MSU ..................

CHAPTER 3. THEORY ..........................................................
Relationship between Peel Test and Burst Test ......................
Burst Test Theory — Unrestrained and Restrained cases ............

CHAPTER 4. MATERIALS & METHODS ..................................
Part A. BLISTERS:

I. Unrestrained Vs Restrained Burst Test Results ..........................
II. Package Size and Gap Size Effect .......................................

Part B. POUCHES:
I. Unrestrained Vs Restrained Burst Test Results ..........................
II. Package Size and Gap Size Effect ...........................................

Part C. Correlation Between Burst Test and Peel Test ...........................

CHAPTER 5. RESULTS & DISCUSSION ......................................

RESULTS

Part A. BLISTERS:

I. Unrestrained Vs Restrained Burst Test Results .......................
Unrestrained Results ..........................................
Restrained Results ..........................................
Comparison between Unrestrained and Restrained Results ............
Comparison between Variances .................................
Comparison between Coefficients of Variation ..........................

vii

21

22
27

3O
3O

37

42

45
45
47
48
53
53

II. Package Size and Gap Size Effect ........................................... 55

General Results ...................................................... 55
Statistical Results ...................................................... 56
Comparison between Variances .............................................. 64
Comparison between Coefficients of Variation .......................... 66

111. Summary of Results for Blisters ............................................. 68

Part B. POUCHES:

I. Unrestrained Vs Restrained Burst Test Results ....................... 69
Unrestrained Results .......................................... 69
Restrained Results .......................................... 75
Comparison between Unrestrained and Restrained Results ............ 77
Comparison between Variances ........................................... 91
Comparison between Coefficients of Variation .......................... 95

II. Package Size and Gap Size Effect ........................................... 99
General Results .............................................. 99
Statistical Results .............................................. 100
Analysis of Failure Pattern for Different Package Sizes ................... 105

III. Summary of Results for Pouches ............................................. 108

Part C. Correlation Between Burst Test and Peel Test ........................... 109
General Results ................................. 109
Summary of Results ................................. 120

DISCUSSION ................................. 122

CHAPTER 6. CONCLUSIONS AND RECOMMEDATIONS ................ 125

1. Conclusions ....................................................................... 126

H. Recommendations ............................................................... 129

APPENDICES
Appendix A — Blister; Raw Data ....................................... 132
Appendix B - Pouches; Raw Data ....................................... 138
Appendix C — Correlation between Burst Test and Peel Test; Raw Data .. 147
Appendix D - Reasons Why Restraining Plates Should be Used ............ 165

REFERENCES ......................................................................... 169

Table 1.

Table 2.

Table 3.
Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

Table 11.

Table 12.

Table 13.

Table 14.

LIST OF TABLES

Sampling Procedure for Blisters — Unrestrained Vs. Restrained
Burst Test Results

Sampling Procedure for Blisters - Package Size and
Gap Size Effects

Sampling Procedure for Pouches; Peel Test
Sampling Procedure for Pouches; Burst Test

Sampling Procedure for Pouches — Correlation between
Burst Test and Peel Test

Unrestrained Results for Blisters

Unrestrained Burst Test for Blisters

AN OVA one-way Analysis — Package Size Effect
Restrained Results for Blisters

Restrained Vs. Unrestrained Results for Blisters

Overall Package Size and Test Method Effect on Burst
Test Results for Blisters — ANOVA Two-way Analysis

Test for equality of two variances for Blisters
(F -test for Unrestrained Vs. Restrained)

Comparison between Coefficients of Variation for Blisters
(Unrestrained Vs. Restrained)

Restrained Results for Blisters
Package Size and Gap Size Effects

Overall Package Size and Gap Size Effect on Burst Test
Results for Blisters — AN OVA Two-way Analysis

aaaaaa

......

oooooo

......

Ta

Ta.

Tat
Tab

Tab]

Table 15.

Table 16.

Table 17.

Table 18.

Table 19.

Table 20.

Table 21.

Table 22.

Table 23.

Table 24.

Table 25.
Table 26.
Table 27.

Table 28.

Table 29.
Table 30.

Table 31.

Comparison between the Variances for Blisters
Package Size and Gap Size Effects

(Bartlett’s Test for Homogeneity of Variances) ...... 64
Pairwise Comparisons of Variances for Blisters

(Package Size and Gap Size Effects) ....... 65
Pairwise Comparisons of Coefficients of Variation for Blisters

(Package Size and Gap Size Effects) ...... 66
Unrestrained Results for Pouches ...... 69
Overall Package Size and Chevron Effect in Unrestrained

Burst Test Results for Pouches — ANOVA Two-way Analysis ...... 70
Restrained Results for Pouches — per Gap Size ...... 75
Overall Unrestrained Vs. Restrained Burst Test Results

for Pouches — ANOVA One-way Analysis ...... 77
Overall Unrestrained Vs. Gap = 1.0” Burst Test Results

for Pouches — AN OVA One-way Analysis ...... 78
Comparison between Variances for Pouches

(Bartlett’s Test for Homogeneity of Variances) ...... 91
Comparison between Variances for pouches

(Bartlett’s Test for Homogeneity of Variances w/o gap = .025”) ...... 92
Pairwise Comparisons of Variances for Pouches ...... 93
Pairwise Comparisons of Coefficients of Variation for Pouches ...... 95
Restrained Results for Pouches — per Package Size ...... 99
Overall Package Size and Gap Size Effects on Burst

Test Results for Pouches — ANOVA Two-way Analysis 100
Location of Failures for Pouches ...... 106
Summary — Location of Failures for Pouches ...... 107

Burst Test Results for Pouches
Correlation between Burst and Peel Tests

viii

Table 32.

Table 33.

Table 34.
Table 35.
Table 36.
Table 37.
Table 38.
Table 39.
Table 40.
Table 41.
Table 42.
Table 43.
Table 44.

Table 45.

Peel Test Results for Pouches using Savg
Correlation between Burst and Peel Tests

Peel Test Results for Pouches using 8min
Correlation between Burst and Peel Tests

Relationship between Seal Strength and Burst Pressure
Unrestrained and Restrained Results - Blisters Raw Data
Package and Gap Size Effects — Blisters Raw Data
Unrestrained Burst Test Results - Pouches Raw Data
Restrained Burst Test Results — Pouches Raw Data

Peel Strength Results — Pouches Raw Data

Restrained Burst Test - Results for Correlation — Gap = 0.25”
Restrained Burst Test - Results for Correlation - Gap = 0.50”
Restrained Burst Test - Results for Correlation — Gap = 1.0”
Peel Test Results for Correlation, Gap = 0.25” - Raw Data
Peel Test Results for Correlation, Gap = 0.50” — Raw Data

Peel Test Results for Correlation, Gap = ID” - Raw Data

ix

uuuuuu

Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.
Figure 9.

Figure 10.

Figure 11.

Figure 12.
Figure 13.

Figure 14.

Figure 15.

Figure 16.

LIST OF FIGURES

Force Diagram In Seal Area ......
Pouch in its Flat Configuration ......
Pressurized Pouch — Unrestrained Case ......
Force Diagram of Half-Center Section Strip — Unrestrained Case ....
Diagram of Half-Center Section Strip — Inflated Pouch ......
Pressurized Pouch — Restrained Case ......
Force Diagram of Half-Center Section Strip — Restrained Case ......
Force Diagram of Half-Center Section Strip _ Restrained Case ......

Relationship between Critical Burst Pressure and Plate Separation ...

Burst Test Restraining Fixture for Blisters (14.5” x 11.5” x 3/8”)
at Plate Separation = 1.0” — Medtronic’s Design

Burst Test Aluminum Restraining Fixture for Pouches

(12’ x 12” x 3A”) at gap = 1.0” ......
Pouch Seal Locations for Pouches ......

Pouch Seal Locations for Pouches -— Correlation ......

Average Unrestrained and Restrained Burst Pressure Vs. Package

Seal Perimeter for Blisters ......

Box Plot of Unrestrained Burst Pressure Vs. Package Seal

Perimeter for Blisters ......

Box Plot of Restrained Burst Pressure Vs. Package Seal

Perimeter for Blisters ......

15

l6

l7

17

18

19

23

32

33

39

50

51

52

Figure 17.

Figure 18.

Figure 19.

Figure 20.

Figure 21.

Figure 22.

Figure 23.

Figure 24.

Figure 25.

Figure 26.

Figure 27.

Figure 28.

Figure 29.

Figure 30.

Figure 31.

Average Burst Pressure Vs. Gap Size for Blisters ...... 58
Box Plot of Burst Pressure Vs. Gap Size for Blisters

Package #1 ...... 59
Box Plot of Burst Pressure Vs. Gap Size for Blisters

Package #2 ...... 60
Box Plot of Burst Pressure Vs. Package Seal Perimeter

for Blisters - Gap = 0.20” ...... 61
Box Plot of Burst Pressure Vs. Package Seal Perimeter

for Blisters - Gap = 0.10” ...... 62
Box Plot of Burst Pressure Vs. Package Seal Perimeter

for Blisters - Gap = 0.01” ...... 63
Average Unrestrained Burst Pressure Vs. Package Seal

Perimeter for Pouches 72
Box Plot of Unrestrained Burst Pressure Vs. Package

Seal Perimeter for Pouches, Chevron Up ...... 73
Box Plot of Unrestrained Burst Pressure Vs. Package

Seal Perimeter for Pouches, Chevron Down ...... 74
Average Burst Pressure Vs. Package Seal Perimeter for Pouches
Unrestrained and Restrained (Gap = 1.0”) ...... 80
Average Burst Pressure Vs. Package Seal Perimeter for Pouches
Unrestrained and Restrained (Gap = 0.75”) ...... 81
Average Burst Pressure Vs. Package Seal Perimeter for Pouches
Unrestrained and Restrained (Gap = 0.625”) ...... 82
Average Burst Pressure Vs. Package Seal Perimeter for Pouches
Unrestrained and Restrained (Gap = 0.50”) ...... 83
Average Burst Pressure Vs. Package Seal Perimeter for Pouches
Unrestrained and Restrained (Gap = 0.25”) ...... 84
Average Burst Pressure Vs. Package Seal Perimeter for Pouches
Unrestrained and Restrained Results ...... 85

Figure 32.

Figure 33.

Figure 34.

Figure 35.

Figure 36.

Figure 37.

Figure 38.

Figure 39.

Figure 40.

Figure 41.

Figure 42.

Figure 43.

Figure 44.

Box Plot of Burst Pressure Vs. Package Seal Perimeter

for Pouches - Gap = 1.0” ...... 86
Box Plot of Burst Pressure Vs. Package Seal Perimeter

for Pouches - Gap = 0.75” ...... 87
Box Plot of Burst Pressure Vs. Package Seal Perimeter

for Pouches Gap = 0.625” ...... 88
Box Plot of Burst Pressure Vs. Package Seal Perimeter

for Pouches - Gap = 0.50” ...... 89
Box Plot of Burst Pressure Vs. Package Seal Perimeter

for Pouches - Gap = 0.25” ...... 90
Average Restrained Burst Pressure Vs. Gap Size for Pouches ...... 101
Box Plot of Restrained Burst Pressure Vs. Gap Size

for Pouches — Package #1 ...... 102
Box Plot of Restrained Burst Pressure Vs. Gap Size

for Pouches - Package #2 ...... 103
Box Plot of Restrained Burst Pressure Vs. Gap Size

for Pouches - Package #3 ...... 104
Average Burst Pressure Vs. Plate Separation

With “Average Seal Strength” Values ...... 116
Average Burst Pressure Vs. Plate Separation

With “Minimum Seal Strength” Values ...... 117
1n [Average Burst Pressure] Vs. ln [Plate Separation]

With “Average Seal Strength” Values ...... 118
In [Average Burst Pressure] Vs. ln [Plate Separation]

With “Minimum Seal Strength” Values ...... 119

xii

CHAPTER 1. INTRODUCTION

The assurance of the seal integrity of any package, but especially of packages
for medical and food products, is one of the most critical steps of any quality control
program. There are different ways to measure the seal strength of a package. The most
commonly used tests for this purpose are the tensile or peel test, and the burst or gas
pressurization test. Although the tensile test has been used through the years in
industry, it has some inconveniences: it is a time consuming test because many strips
have to be cut from a package in order to get a “true measure” of the seal strength. In
cases in which only some sample strips are tested, there is a possibility that some of the
weak areas may be overlooked. The burst test, on the other hand, has gained acceptance
in industry because it does not require as much time and it is easier to perform. Also, it
provides an evaluation of the entire package system not only of the seal. Burst testing
of packages subjects the entire sterile package system to some of the stresses that
packages encounter in the manufacturing, distribution and use environment [5].

This research project focused its investigation on the burst test. Burst testing
consists of increasingly pressurizing a package until it breaks. The pressure required to
break the package is recorded as a measure of the seal strength. Since most packages
for medical purposes are made with at least one flexible side the internal pressure tends
to deform the package during the test. In the case of pouches, it deforms both sides,
while in the case of blisters or trays, which are packages formed by a preformed plastic
sheet with flexible, semi-rigid, or rigid cover, it tends to deform only the lid or the lid
material. This deformation of the package may direct the force of the pressure to

specific areas depending on the package geometry and on the type of seal. In doing so,

it may influence the resulting burst values. It is also known that the package size, seal
peel strength, and material thickness, among other factors, affect the burst values.

In recent years, the idea of using restraining fixtures in the burst test has been
developed by engineers, researchers, and people from industry in general. A standard
method of restrained burst test has been proposed to the American Society for Testing
and Materials (ASTM). Members of leader companies like Carleton Technologies,
Medtronic, Rexam Medical Packaging, and TM Electronics have been providing
reasons for using restraining fixtures in the burst test.

One of these reasons is that with restraining plates there is a greater chance to
find the weakest point. As said before, when the package is inflated in an unrestrained
burst test it tends to deform. This deformation creates stress concentrators in some
areas of the package causing them to break at the point of stress concentration. The
main concern is that the package does not deform in the same way each time and that it
does not necessarily break on the weakest point but as a result of the stress
concentration caused by the deformation. Restraining plates are thought to limit the
extent and variation of deformation. Because all seal surfaces are exposed to the same
forces with restraining plates then there is a greater chance of finding the weakest point.
A second reason these people believe, is that restraining fixtures will provide more
consistent test results, and that the use of the restraining plates will help reduce or
eliminate the effect of other variables like package size and geometry. The possibility
that the repeatability of the burst test can be improved by using restraining plates is the

main reason why people in industry nowadays are proposing a new standard.

The purpose of this thesis project is to provide an analysis of the burst test
method using restraining plates, to study its advantages and disadvantages, and to
evaluate the applicability of this type of test in different situations. This analysis will
include the package size and package geometry effects on restrained burst tests results.
Since the restrained burst pressure is known to vary with the restraining plate
separation, an analysis of plate separation (gap) effect will be provided and the
theoretical relationship between the restrained burst pressure and plate separation will
be analyzed. Also an overall comparison between the unrestrained burst test results and

the restrained test results will be performed.

CHAPTER 2. LITERATURE REVIEW

RESTRAIN ED BURST TEST METHODS FOR BLISTERS

In 1994, David Bohn, fiom Medtronic, Inc., wrote in his article “Using Burst
Testing to Evaluate Sterile Blister Packaging ” about research being planned for totally
restricted burst testing [5]. It was not until 1996, that John Spitzley, fiom Medtronic,
he, wrote and designed a test plan (PTP9609121) in order to define the tests and
procedures required to determine the effect of a restraining fixture on the burst values of
packages of widely varying sizes and geometries [18]. They decided to do this because
blisters have one flexible side and they think that one of the effects of the internal
pressure in an unrestrained burst test is a “doming” of the lid which can alter the shape
of the package. This may direct the force of the pressure to specific areas depending on
package geometry thus influencing the resulting burst values. Medtronic’s theory was
that if the lid of the package were prevented from “doming” by a restraining fixture, the
result may be to minimize the effects of package size and geometry on the resultant
burst values. This test plan was put in practice in the School of Packaging. The results

will be shown later on in this report.

 

rel.
fin
obi;
0ft}
equa
of [h 1'

Ihebi

RESTRAINED BURST TEST METHODS FOR POUCHES

In January 1992, the use of restraining plates for food flexible pouches was
mentioned in section 4.3.4 of the MIL-P-44073 D - Military Specification Packaging
and Thermoprocessing of Foods in Flexible Pouches [13]. Its revised edition, from
February 1996, MIL-PRF-44073E — Performance Specification for Packaging of Food
in Flexible Pouches, mentions in section 4.3 .7 the use of restraining fixtures [14]. It
says that the internal pressure resistance shall be determined by pressurizing the
pouches while they are restrained between two rigid plates spaced 1/2 inch +/- 1/16 inch
apart. It mentions the use of the plates for open package (three-seal tester) and for
closed package (four-seal tester). Also, it specifies the parameters to be used in the test,
how the pouches should be examined and the criteria that should be used to consider a
test failure.

Professor Kit Yam, fiom Rutgers University, published in 1993 an article
“Relationship between Seal Strength and Burst Pressure for Pouches ”, in which he
mentioned the use of restraining plates [22]. The purpose of his study was to find the
relationship between peel and burst tests. The burst test was performed using restraining
fixtures. The article explained, based on force analysis, that the seal strength (S)
obtained from the peel test is equivalent to the product of the burst pressure (P) and half
of the plate separation (D) used for the burst test (S = P*(D/2)). Yam wrote the
equation as (S = (P*R)), where R = D/2. He emphasized in his article that the validity
of this equation is based on the assumption that the peeling times for the peel test and

the burst test should be the same.

Thomas Wachala, from Carleton Technologies, in 1994, published a study
“Restrained Vs Unrestrained Pressure T esting”, in which he compared both burst test
methods [20]. He explained in his article how the package is not the only factor
affecting the burst test results. He thinks that the method of holding the package during
the test can also have a big effect. Wachala also says that some of the advantages of the
restrained burst test method are: that it helps to test the packages more uniformly by
exposing all surfaces to the same forces and that this test provides a greater chance to
find the weakest point of the package. Also he mentioned as an advantage that the
restraining fixture would minimize the effects of package geometry and that the plate
separation could be standardized for use for specific packages at multiple locations. He
thinks that the unrestrained test, if done at multiple locations, has greater potential for
large differences in burst values. The disadvantages for using restraining fixtures, he
says, include higher burst values, and the need for a variety of plates to accommodate
various package sizes.

On January 21 1997, Committee F 2.6 of the American Society for Testing and
Materials (ASTM), presented a draft proposal for a Standard Test Method for Burst Test
Seal Strength Testing of Flexible Packages using Internal Air Pressurization within
Restraining Plates [4]. This standard method in particular is to be applicable to
packages with seals that are intended to have a peelable seal feature. In this proposed
standard method, the restraining plate burst method is described as a rapid means of
evaluating minimum seal strength and tendencies for package seal failure when the

package is exposed to an internal pressure. The use of the restraining fixtures is

 

recommended in order to maintain dimensional stability while the package is
pressurized.

Also, Neil Lorimer, from Rexam Medical Packaging made a presentation
“Understanding Restrained Burst Testing”, on April, 1997 to the ASTM F02
Subcommittee on Medical Packaging, in which he explained reasons for performing
restrained burst testing [12]. One of the reasons he provided was that restrained burst
testing provides a rapid means of evaluating minimum seal strength (burst strength).
The other reason is that this test is more efficient and economical to perform than force
gage testing of peel strength. He also mentioned in his presentation that restrained burst
testing can reliably detect the weakest area of a package seal placed around the
perimeter of a flexible package and that this is very important when developing
correlation between peel and burst test. It is important to recognize that tests values for
burst strength are correlated only to the weakest areas of the pouch seal and not to the
entire distribution of seal strength values. He thinks that in order to find correlation
between burst and force gage peel tests it is better to use restrained burst testing results
than the unrestrained burst test results. Pouches, when tested in an unrestrained mode,
tend to burst in the middle of the bag in spite of where the weakest point is really
located. This appears to be because a crease appears there, which concentrates stress.

Appendix D, at the end of this report, provides a detailed list of reasons why
members of leader companies are suggesting the use of restraining fixture in the burst

test.

10

RESTRAINED BURST TEST STUDIES PERFORMED AT

MICHIGAN STATE UNIVERSITY - SCHOOL OF PACKAGING

During fall 1996 we started working with restrained burst test methods for
blisters and pouches. Dr. Hugh E. Lockhart, Professor at Michigan State University
School of Packaging is in charge of this project. We have worked together on the design
of all the tests, the design of test fixtures, the experimental designs, and in the analysis
and interpretation of the results. We learned from our experiments three basic
behaviors that hold for both pouches and blisters. The first one is that in burst testing,
the burst pressure required to break a package decreases as the package size increases.
The second one is that unrestrained burst pressures are lower than restrained burst
pressures. The third one is that as the plate separation decrease, the burst pressure
required to break a package increase [9, 10, and 11]. These three behaviors were
observed while testing pouches and blisters and the results will be discussed in this
report.

The literature review presented above demonstrates that there is some work that
has been done in order to explain burst test methods using restraining fixtures. There is
certainly an effect of package size and of plate separation distance on the burst test
results. There is also a difference between the unrestrained burst test method and the
restrained method. The intent of this thesis project is to study and analyze these effects
on the burst test results, and to understand the main differences between restrained and

unrestrained burst test methods.

W

11

l2

RELATIONSHIP BETWEEN PEEL TEST AND BURST TEST

As mentioned in the literature review section, Professor Kit Yarn, from Rutgers
University has worked on correlating peel test with restrained burst test results. In his
article [22], he explained that the seal strength (S) obtained from the peel test is
equivalent to the product of the burst pressure (P) and half of the plate separation (D)
used in the burst test (S = P*(D/2)). He derived this equation based on the assumption
that the walls of the pouch take approximately a circular shape when the air pressure
exerts a tensile force on the seal to peel it apart. The Y component of forces (tensile

peel) around the seal area can be represented by:

 

 

 

 

1 P Seal Plane
x

 

VFy

Figure 1. Force Diagram in Seal Area

l3

dF y = P R sin 9 d9
Fy = force peeling a one inch width of the seal
P = internal pressure
R = half plate separation
Fy=lo"/2 PR sine d0
Fy = P R ; Fy can be substituted by S (lb./in) at rupture
S = P R
He emphasized in his article that the validity of this equation is based on the assumption
that the peeling times for the peel test and the burst test should be the same. The tensile
peeling time is a function of gauge length, crosshead speed and strain-stress properties
of the pouch material and the seal. The burst peeling time is a firnction of plate
separation, rate of pressurization and stress-strain properties of the pouch material and

the seal. Professor Yam presented data in his article to support his theory.

l4

BURST TEST THEORY - UNRESTRAINED & RESTRAINED CASES

During summer 1997, Dr. Gary Burgess, Professor at Michigan State University,
got involved in the development of this project. He developed theoretical equations,
based on force diagrams, in order to explain the pouch behavior during an unrestrained
and a restrained burst test. The following equations and diagrams were provided by Dr.

Burgess.

Pouch Burst Testing:

 

 

L, W are the internal dimensions of the flat pouch

L before pressurization

 

 

Figure 2. Pouch in its Flat Configuration
When the package is pressurized the center section of the pouch tries to become
circular. The pouch “shrinks”.

l

The internal dimensions are now LI and W and are
smaller than LI and W,I respectively.

(L| < L) and (w'< W)

 

\
W Figure 3. Pressurized Pouch - Unrestrained Case

15

Unrestrained Case:

The force diagram, in Figure 4, represents a half center section strip of width h in an
unrestrained burst test at failure. The work needed to peel the seals apart can be
described by the vector component of the force in Y direction, perpendicular to the

plane of the seal.

 

 

 

 

~—-"‘"
_)
_)

’—

 

ls . st
t i

Figure 4. Force Diagram of Half-Center Section Strip - Unrestrained Case

I
the vertical component = ZFy = P W h = 2 S h ;

P= [(2 S)/w‘] (1)

where P — pressure

h = width ofthe strip

{/3
ll

seal strength (lb/inch) in a 180 ° degree peel test
(force required to peel seal apart / the width of the strip)

W'= diameter of the pouch (See Figure 5, next page)

In

In

L.)

(I)

16

 

Still

 

 

 

 

 

 

 

 

Figure 5. Diagram of Half-Center Section Strip - Inflated Pouch

Ifthe center section is approximated as a circle, assuming that the strip does not stretch

much along its perimeter, then;

n W‘ = 2 W (rt * diameter = circumference)

w'= [(2 W) / 1c]; w'= .636 w (it shrinks about 1/3) (2)
Substitute equation (2) in (1);

pm = [(1t S) / W)]; puma: = Burst Pressure (3)

So, for the unrestrained case the burst pressure is a function of seal strength and

pouch size. With this equation, some predictions can be made:

1. The burst pressure increases as the seal strength increases

2. The burst pressure decreases as the width of the package increases. Therefore,
bigger pouches are weaker in burst, even when seal strength is the same.

3. Dimension “L” has no effect on Panic... The burst pressure depends only on the
smaller dimension (W); so lengthening of the pouch while keeping the width the

same should not affect the burst strength.

17

| l———x—l

I/( Y

Figure 6. Pressurized Pouch - Restrained Case

Restrained case:

 

 

 

 

 

The restraining plates apply force over the contact length x. This force is equal and
opposite to the air pressure P inside and so these forces cancel and do not enter the force
balance. The vertical components of the pressure along the curved parts are balanced

by the seal tension, assuming the material does not stretch

'0

P

”H
it

 

4 \
t

T \
T on
T

 

 

  

 

 

b!

 

Figure 7. Force Diagram of Half-Center Section Strip - Restrained Case

SC

SC]

.w

18

 

  

P(D12)h D/2

 
 

 

 

 

 

 

Figure 8. Force Diagram of Half-Center Section Strip - Restrained Case

From the figure above;
x + [rt/2 * (D/2)] + [rt/2 * (D/2)] = W x = W - [1t* D/2] (4)
ZFy=ZSh=2*(pDh/2); p=[(2S)/D] (5)

SO Pei-meat = l(2 S) / D]

So, for the restrained case the burst pressure is a function of seal strength and plate
separation. With this equation, some predictions can be made:
1. The burst pressure increases as the seal strength increases.

2. The burst pressure increases as the distance between the plates decreases.

It
tht-

the

19

It is important to notice that the unrestrained case is a special case of the restrained case.
If D gets bigger and bigger, then eventually it will be unrestrained. This happens when
x = O; which from equation (4) above happens when D = [(2 W) / it]. When this D
value is substituted into equation (5), the following can be obtained;

Pcritical = {(2 S)/ [(2 W)/7t]} ; Pcm= [(1t S)/Wl

which is the burst pressure obtained for the unrestrained case. See equation (3)

Both cases can be put in a single graph.

Restrained;

PM“ Pcritlcal = (ZS/D)

Un restrained:

(1tS/W) __ Pctitlcal =(1tS/W)

 

 

 

I
(2W/rt)

D = Plate Separation (Gap Size)

Figure 9. Relationship between Critical Burst Pressure and Plate Separation

It can be seen from the graph that the restrained results could be represented
theoretically as a hyperbolic function. Beyond a certain D (plate separation) value,

there is no contact between the package and the plates and the test is similar to an

20

unrestrained one. In this case Puma] is independent of D and the data cannot be
represented with a hyperbolic function anymore. The results provided by Dr. Burgess
agree with the results provided by Dr. Kit Yam in his article “Relationship between

Seal Strength and Burst Pressure for Pouches ”[22].

CHAPTER 4. MATERIALS AND METHODS

21

22

PART A - BLISTERS:
I. UNRESTRAINED VS RESTRAINED BURST TEST RESULTS
Materials Tested:
1. Medtronic Accessories Package (P/N 119401-001)
Seal Perimeter (13.5”)
2. Medtronic Thera Small Outer IPG Package (P/N 119679-001)
Seal Perimeter (20.0”)
3. Medtronic Standard Leads Outer Package (P/N 119421-001)
Seal Perimeter (27.5”)
4. Medtronic Myocardial Leads Outer Package (P/N 119553-001)

Seal Perimeter (33.0”)

Test Methods Used:

1. ARC 2600 Burst and Creep Tester - Medtronic’s Operating Procedure PE026

2. Medtronic’s Test Plan PTP9609121

“Eflect of a Restraining Fixture on the Burst Values of Sterile Packages ”

Equipment:
1. Test-A-Pack 2600 Burst Tester - Carleton Technologies with closed fixture
2. Burst Test Restraining Fixture (14.5” x 11.5” x 3/8”) — Medtronic’s design and

construction. See Figure 10, next page.

23

=w_uon 93:95.62 I .5.— " aerauqaom 82m «a Asa? u :0: u ..mé: 2035 ...... 9:55.— wEEahmom «non. 3:5 .3 9...»:—

 

24

Procedure:

One half (50) of the packages of each kind were burst tested in an unrestrained
mode, and the other half (50) were burst tested using Medtronic’s restraining fixture.
Each group of (50) packages was tested on two days, 25 each day. This allowed
evaluation of day effect as a result of starting and stopping the test sequence.

The blisters were placed with or without the restraining plates in the closed package
fixture; depending if restrained or unrestrained burst test, respectively, was being
performed. A needle punctured the lid of the blister. The package was pressurized until

it broke. The pressure required to break the package was recorded.

Experimental Design:
See Table 1. Sampling Procedure for Blisters - Unrestrained Vs Restrained Burst

Test Results, next page.

Data Analysis:

Test results were analyzed statistically for significant difference between the means
using one-way and two-way analysis of variance, and t-tests as appropriate. The results
were further analyzed to determine if there were differences in variation between the

two test methods.

25

 

 

 

 

 

 

 

 

 

 

 

.86 a6 a u cesium a. .9:
m n 33:23 m u 8am 32m 50-2.3: ZR
m N 8nd 32m e u macaw 23me ..\o me H Ed .8d comm owed—cam 350
on 8383 mm sass. mm o . S u e 5&2 :26 82$
some u 95 a u §>Eaum Na .3:
m n 53:33 m n cam 26: $0.23: Ea
m n 89m 32m N. n aaom 335m a\o me H mm .Ed cmb owe—own c830
ow moans... 8 tags 3 o . a u H 82S: .35 SE.
.3 .o u so a u cesium
m u §>Em5m m n 3am 30E 3. .Uv:
m n BB 26; 2 n 3am 8:ch .x. a. .... mm :2 o3 839a: 23
0m 3383 mm sass mm o . om n e 85>: umfiga c8325.
..3 .o u 96 a u crease
m u §>Emcom m N 8nd 32m ta .Uv:
m n 39m 26E 2 n asaom 2:ch X. me H mm .Ed omnm 25-—ova: ZR
cm 8388 8 3383 mm o . S u H 332 uwfigm panacea
$885. 8.35am EouoEauam anoEauam 5:55.!
.3 Baa—=2 Eon. Boa. 9523— oEE. team
.53. 359.33— 195.289.: D Q 9.. 322—53. .25 09C. owed—cam

 

3.5.3— :oh 6...:— uofiabmom m> 35232:: . 9.8%:— ..e.. 9:539..— weznfiam ._ 03am.
luv—mm:— Aflhzmémmxm

 

u r ...}..d- a: :..—.—.’....~

26

 

 

 

 

 

 

..8.o n 96 m u asécum .a .3..—
m n beamsm m u 3Q Boa 80-38: 2E
m u use 32m a u madam 8:ch s a. n mm as one $35 .538»:
8 case 3 Enema 0 . a n e 53: 2355 5:6
as... n ewe m u 3322B 3.. .3..—
m u 53523 m u use 32m 25.33: 23
m N Sex 32m o .I. 9.5m 835m a\° me H mm .Ee omuw owed—own 363032
8 case 3 sass. mm o . om u H SEE escasm 550
am; u 96 m u seesaw nu .33
m u b32255 m u BB >6: 839m: 2E
m n 33— 265 _ u macaw 82me .x. me H mm .Ed 0on owaxoam 384
8 $383 mm 3388 mm o . om u a 85%: caeaam .830
..m _ .o u 96 m n 332.8% mu .3:
m u 53523 n n 22 32a 25.8%: 23
m N 8nd 30E _ .I. madam 83me .x. me u 5 .Ed 062 owed—omn— 384
8 case n sass. mm o o a u a 532 esteem ago
coach. moan—am 280593.— anoEacam LEE—5.:
..e .825: Z Bob 29,—. 9550“ 25% team
:33. 133.38”— 105.239.: D Q 9:522.th than 2.3. owed—ea.—

 

 

 

 

 

 

mzaom .88 3.5: 605.23.3— a> 35.2295: . 9.8%.: 3.. 9:539..— wEEEam A 033—.

ZO—H<DZ~HZOU .. ZU—mmn A<EZH§HQXH

 

27

II. PACKAGE SIZE AND GAP SIZE EFFECT

Materials Tested:
1. Medtronic Accessories Package (P/N 119401-001)
Seal perimeter = 13.5”
2. Medtronic Thera Small Outer Package (P/N 119679-001)
Seal perimeter = 20.0”
Test Methods Used:
1. ARC 2600 Burst and Creep Tester - Medtronic’s Operating Procedure PE026
2. Medtronic’s Test Plan PTP9609121
“Eflect of a Restraining Fixture on the Burst Values of Sterile Packages
Equipment:
1. Test-A-Pack 2600 Burst Tester - Carleton Technologies with close package fixture.
2. Medtronic’s Burst Test Restraining Fixtures (See Figure 10)

Box #1, #2, #3, #4, & #5 - the difference in boxes is the plate separation
Procedure:

All the packages were burst tested in the restraining fixture. Each group of 24
packages, at 3 different gaps, was tested on 2 days, 12 each day. This allowed
evaluation of day effects .

The blisters were placed within the restraining fixture. A needle punctured the lid
of the blister. The package was pressurized until it broke. The pressure required to

break the package was recorded. The same procedure was repeated for the three gaps.

23.3.3: _ <92..- nears—kl}..—

28

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9, 6V:
Bosses: 2a
..5. u 96 ..2. u a6 .8. u 96 m u assuaom .x. e. H 5 .ea 84.. 093.03 .36
S 2 S as V5m Q vem i xom m n 3am 32m 0 0 S u 9 39$. :95 82;
5. 6E
.. S. u 96 ..2. u 96 .8. n 96 m n 33388 x 8 n mm .8; 02 80-83: 23
S S 2 3a xom a 55 Q. xom m n 22 Boa o o 8 u 9 39.9». omega 98889.
«a dc:
Sosso: 2E
.. S. u 96 ..2. u AH6 ..om. n Q6 m u bsfimcom as as n mm .93 can 936$ 595
2 S 2 Na vam Q V5m i vam m u and so: u o a n a. 83S. :85 82F
a .9:
..S. u 96 ..2. u 96 .8. u 96 m n 93222 .x. a. n mm .as on: 898$: 23
.9 S 2 ya cam 3. £5 3. 5m m u 82 32m 0 o a u H RES. 639$ 90:82...
£295
.538 8.953 35 55 £825.89 9523— 2.5. Sam
.3 non—E:— 133. ..anaz nem— amoh Q 9.5.22.th .25 «9C. mind—cam

 

38cm.— onmm 95 can 3% $5.25 .. £035— ..8 2:38.; 3:955 .N 035—.

ZU—mm—G A<PZH§HAXH

 

29

Data Analysis:
The results were analyzed statistically for significant differences between means
using one-way and two-way analysis of variance. The results were further analyzed to

test the effect of gap size on variance.

30

PART B — POUCHES:
I. UNRESTRAINED VS RESTRAINED BURST TEST RESULTS
11. PACKAGE SIZE AND GAP SIZE EFFECT
A single test was conducted to compare unrestrained vs. restrained burst test results and
also to analyze the effects of package and gap size on the burst test results.
Materials Tested:
1. Package #1 - Oliver Products Company
(5” x 10”) Chevron Seal Pouch
Pouch Raw 1073 B/I‘yvek Pouch W/48 Gage PET
Plastic (2.6 mil) / Tyvek (7.0 mil)
Basis Weight (76.83 gym2 or 2.26 oz/ydz)
Average Peel Strength (2.] lb./in.)
Seal Perimeter (16.5”)

Seal Width (3/8”)

2. Package #2 - Oliver Products Company
(7” x 11”) Chevron Seal Pouch
Pouch Raw 1073 B/Tyvek Pouch W/48 Gage PET
Plastic (2.6 mil) / Tyvek (7.2 mil)
Basis Weight (75.10 g/m2 or 2.21 oz/ydz)
Average Peel Strength (2.2 lb./in.)
Seal Perimeter (19.0”)

Seal Width (3/8”)

w

E.

m

Lo)

31

3. Package #3 - Oliver Products Company
(9” x 12”) Chevron Seal Pouch
Pouch Raw 1073 B/I‘yvek Pouch W/48 Gage PET
Plastic (2.6 mil) / Tyvek (7.1 mil)
Basis Weight (74.85 g/m2 or 2.21 oz/ydz)
Average Peel Strength (2.0 lb./in.)
Seal Perimeter (21.5”)

Seal Width (3/8”)

Test Methods Used:
1. ARO 2600 Burst and Creep Tester Operating Procedure

2. INSTRON Tensile Tester (Model 4201) Operating Procedure

Equipment:
1. Test-A-Pack 2600 Burst Tester - Carleton Technologies with open package fixture.
2. Burst Test Aluminum Restraining Fixture (12” x 12” x 3A”).

See Figure 11, next page.

3. INSTRON Tensile Tester Machine — Model 4201

32

ac.“ u as» “a v.3 x ..N— u ..NC nose—Sm ...... chain win-3.58% BEE—52 «8h 5::— .2 Bani

 

33

Procedure:
A group of 30 packages were peel tested. There were 3 package sizes, 5 pouches
each, in 2 days. Four (4) locations (A, B, D, and E) were tested from each pouch. See

figure below.

 

 

Figure 12. Pouch Seal Locations
The peel strength of the sides (A and E) was compared with the peel strength of the
chevron area (B and D) to see if they were statistically different.

A group of 420 packages were burst tested. There were 3 package sizes, at 7
different modes, 10 samples each, in 2 days. The testing modes consisted of two (2)
types of unrestrained modes: chevron up or chevron down and five (5) types of
restrained modes: gap heights of 0.25”, 0.50”, 0.625”, 0.75”, and 1.0”, were used.
Chevron up means that the pouch was placed with the plastic side looking up and
chevron down means that it was placed with the Tyvek side looking up. So, in general,
there were 3 sizes * 7 modes * 10 samples * 2 days = 420 packages. The size of the
three different packages was determined by measuring the seal perimeter. The seal
perimeter values really represent the inner border of the sealed area.

The pouches were placed with or without the restraining plates in the open package

fixture; depending if restrained or unrestrained burst test, respectively, was being

 

d.

Sl.‘

34

performed. The pouches were slid over the test fixture’s inflation port until its top
portion touched the metal stops. The actuator was pressed down to clamp the pouch in
place. The package was pressurized until it broke. The pressure required to break the

package was recorded. The same procedure was repeated for the seven different modes.

Experimental Design:
See Tables 3 and 4, next page. Sampling Procedure for Pouches -— Peel Test and

Burst Test, respectively.

Data Analysis:

The results were analyzed statistically for significant differences between the means
using one-way and two-way analysis of variance. The results were further analyzed to
determine if there were differences in variation between the three different package

sizes and the seven different modes.

35

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m m m m 56% S n 80% .ea 8” : :5 a ..e
m a m < .2 n 5mg: swam 532 2 .9:
m m m m £55 2 u 80% .ea 8” : 5.2 a ..t
m a m < ..oa n 5mg: swam 58:: a do...
n m m m 58% Q n 80% Se 8“ : 9.2 n :9
m a m < ..o.~ u 533 swam 832: E .9:
m m m m 58E S u 30% ea 8a ofi a ..a
m o m < .3 n 5mg: swam 58>: ms. .93
m m. m m 58% S n 80% :3 8a is u E
m a m < .3 n 5mg: $80 532 g .3:
m m m m 58E 2 u 8% .ea 8a 2: u :9
m a m < ..oN n 5mg: swam 5&2 E .33
8.956 .3 53:52 EouoEEam 25,—. team
.3533 ZOEmZ— is 2.3. 993an

 

Emma. AH”: - nose-Sm .8.— og=uououm mils—am .m 03:.

ZU—mm—A— A<HZH§HQXH

 

36

35.58.— " .— MEEE 3.30:0 35.2.8.5: .I. 90.... ..5 .5525 ......a...8.:5 u DUE. "8:7.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

z u .....2.
... ... .. ... .. .. ... .u b.2360. :x: ... u .... .6... 86 ...N. a .6.
:33. :33. :33. :33. :3... 5.... 3.... m u 8.... :6... u . ... u h 8.3... Q. .9...
z u .....2.
... ... ... ... .. .. ... . u ...:Eaam .x. ... u .... .6... 86 ...N. n :5
:33. ......3. :33. :33. :3... 8.... 8.... m u 8.... :6... o . ... u .. 3.3... a... .9...
z u .....2.
.. ... .. .. .. ... ... . n 5.3.3. .x: ... n .... .6... 86 ...... a :m.
:33. :33. :39... :33. :3... 8.... 8.... m n 3.... :6... u . ... u .. 3.3.: .. .3...
z u .....o...
... ... ... .. .. ... ... . n @635. .x. ... m. .... .6... 86. a... . ....
.....3. :33. 38.... :33. :3... 8.... 8.... m u 3.... :6... o . .n u 6 3.3... a. .9...
z u .....2.
... ... ... .. ... .. .. . u £355. :x: N. u .... .6... 85. ...N. u :5
:33. :33. :33. :m...3. :3... 8.... 3.... m u a... :6... o . .N u a 3.3... a. .3...
z u .....2.
.. ... ... .. .. .. ... . u £336... ... ... u .... .6... 85. ...... u :3
:33. ......3. :33. :33. :3... 9...... 20.... m n 3.... :6... o . .~ u .. 53... ... .3...
£26....
683... 8.35am .... LEE—.2 rouofiauam 9520.— 052. ......m on»...
one: 30,—. 80,—. a. 95:22.59... :95 $8.2...

 

Emma. Hwy—Dm— - 8.3.8.— ..8 9:639... wEEEam .9 2......
luv—mm:— A<HZH§HA~XH

 

37

Part C - CORRELATION BETWEEN BURST TEST AND PEEL TEST

FOR POUCHES
Materials Tested:
1. Package #4 - Ethicon Endo Surgery
(6” x 10”) Plastic/Tyvek and Chevron Seal Pouch
Plastic (2.5 mil) / Tyvek (6.8 mil)
Seal Widths (A= 3.5/8”, B and D = V2”, E=3/8”)
Test Methods:

1. ARO 2600 Burst and Creep Tester Operating Procedure
3. INSTRON Tensile Tester (Model 4201) Operating Procedure

4. ASTM F-88 — 94 Standard T est Method for Sea] Strength of Flexible Barrier

Materials

Equipment:
1. INSTRON Tensile Tester Machine - Model 4201

2. Test-A-Pack 2600 Burst Tester - Carleton Technologies - with the open package

fixture.

3. Burst Test Aluminum Restraining Fixture (12” x 12” x 3A”).
(See Figure 11)

4. Stopwatch CASIO with sensitivity of .01 minute.

:1.

P8

110
lthe
Valu

Press

38

Procedure:

According to Yam [22], the validity of the equation relating burst pressure and
seal strength is based on the assumption that the peeling times for the peel test and the
burst test are the same. He explained that the tensile peeling time is proportional to the
elongation and inversely proportional to the crosshead speed; t9 = (60 * AL) / v.
tp is the tensile peeling time, AL (in) is the elongation, v is the crosshead speed (in/min),
and 60 is the factor used to convert from minutes to seconds. The tensile peeling time
(tp) changes as the crosshead speed in the INSTRON machine changes.

In order to find a crosshead speed to make the tensile peeling time equal to burst
peeling time (tb = tp ) the following equation was used: t1. = t1, = (60 * AL) / v in the form
v = (60 * AL) / tb . The burst peeling time (tb) was obtained from the burst test and AL
is taken as 2w, twice the seal width, so it was possible to solve for v.

The assumption that AL = 2w was made because the elongation of the materials
(Tyvek and plastic) was negligible when compared to the elongation of the seal at peak
load. AL, is the elongation of the specimen during the peel test and w is the seal width.
In order to verify that assumption, a tensile test for the Tyvek and plastic materials of
the pouch and a peel test for the seal were performed and compared.

As described by Professor Kit Yam in his article [22], a restrained burst test was
performed for all pouches. Three different gaps were used: 0.25”, 0.50”, and 1.0”. The
flow rate, which is the speed at which the air enters the package, was set at 1, 5, and 9
when testing the packages at each gap. The other two parameters were set at a specific
value and kept constant (sensitivity = 1 and prefill = N). The time between initial

pressurization and pouch bursting (tb = burst peeling time), changed as the flow rate was

 

rec

39

changed. The burst peeling time (tb), burst pressure values, and the location of failure
were recorded for each sample. The burst peeling time was measured using a
stopwatch.

One-inch wide specimens were cut, according to ASTM F88 - 94, from four

different pouch locations. See figure below.

 

 

Figure 13. Pouch Seal Locations

The gauge length in the INSTRON machine was set to RR (R = plate separation
divided by two) so that the area of the specimen acted upon by the tensile peel test was
the same as the area acted upon by the burst test [22]. Since the gaps tested in this
experiment were 0.25”, 0.50”, and 1.0”, the gauge length used were 0.40”, 0.80”, and
1.60”, respectively. The tensile peeling time (tp), and the seal strength at peak (S) were

recorded for each sample. The tensile peeling time was measured using a stopwatch.

40

The predicted burst pressure was calculated using the equation P = (S/R) or P =
(ZS/D); where P is the predicted burst pressure, S is the seal strength at peak obtained
from the peel test, D is the plate separation, and R is half of the plate separation. The

predicted burst pressure was compared with the experimental burst pressure.

Experimental Design:
See Table 5. Sampling Procedure for Pouches - Correlation Between Burst Test

and Peel Test, next page.

Data Analysis:
The predicted values were calculated with the minimum and total seal strength.

The predicted and observed results were plotted to see their agreement.

41

.53 ... 6......“ an ... . a. .6.
.8853... :2... ...... on.» ....» .38 ...... .58 2.2.8... A... ...... ... :38. 0.8:: 8:25.. 3. ...»... 8.97.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.3.... NM 8.3.8 NM 83...... NM ...... comm :3... u ....0

ooN. o. mmdu> o... 8 MM n> S. 8 mm. u> :ov. ...QMNB.
8.9:... NM 8.an. NM 8.9:... NM ...... co... :cm... H ....0

mod. 0. E... n > mNN 0. SM n > M... o. no. u > :3. 33.5.
83...... NM 8.3.... NM 83...... NM .5... ooM ...... u ....0

.9: 0.2.2 n> MM... o.vv.on> 3.8 v5n> :oo. SENS.

a n 32... m u .52... . u .52...
e. w=_....e..8....e0 3 w.....:e..8....e0 3 w.....=o..8....e0 £25 598..
8.8... 8.....an .... ...—=57. ....o... 330 2......
6.5:... .62.. 22.220 29.57.. 6.. on... .5
hm“... Am”... 1

m. m. m. Z M .50... .x. Nv H EM. ...... oMHw :mNd H 9.0

. u 3:35. o .. .8. n e 5.8....
m. m. m. 21E... 6\6 ...”..M E... oMmo. ..om.cn....0

. u 53.26. o . om u H 5......
m. m. m. Z n .....e... 6\: Mv n .... ...... 0me ...... n ....0

. u 3.235. u o .N u H 8.5.:

m n :5... m n :3... . u .52... 888E....... .3533.
3.8... 8.95% .e 8....5 Z .52. .8... 83...... .....h ......m
38.80 Q 25......th ...... Sam ....0
1 5...... 5...... 1

.8... .0... ...... .8... 3...... .8952. 3.3.9.80 - 85...... ...... 9.3.89... w..=.......m .m 2......
7.0.9.9 A<HZH§H~XH

 

CHAPTER 5. RESULTS & DISCUSSION

42

43

Even though blisters and pouches are different types of packages and required
different fixtures to perform the tests, the obtained results were similar. As mentioned,
in the literature review chapter, three behaviors were observed while testing blisters and
pouches. Part A and Part B of this chapter are intended to discuss these three behaviors

for the blisters and pouches, respectively.

Part A (for blisters) and Part B (for pouches) are both divided in three sections.
In the first section the results show that unrestrained burst pressures are lower than
restrained burst pressures. The second section will show results that demonstrate the
package size effects and the plate separation or gap effects on the burst values. It was
observed that the burst pressure required to break a package is inversely proportional to
package size and plate separation (gap size). The third section of both parts provides a

summary of the results obtained for the blisters and pouches, respectively.

Part C, show the results of an experiment that was performed with a different
pouch than the one used in Part B. The purpose of this experiment was to correlate burst
and peel test. The formulas used and the procedure followed were based on Professor
Kit Yam’s research, published in 1993 [22]. A summary of the results obtained in this

experiment is provided at the end of Part C.

The results were analyzed statistically for significant differences between the
means using one-way and two-way analysis of variance (ANOVA). Generally a log

transformation of the data helps to stabilize the variances. The ANOVAs were run for

both, the raw data and for the logged data. The analysis of residuals for each ANOVA
showed somewhat improved conformance to normality for the logged data. The
ANOVA results were essentially the same in regard to significance and we report only
the results of AN OVAs performed on the raw data. A comparison between the
variances and the coefficients of variation of the raw data was performed for each
experiment. These results are reported following the ANOVA results on each section.

Statistical significance is determined at the a = 0.05 level unless otherwise noted.

45

PART A - BLISTERS:

I. UNRESTRAINED VS RESTRAINED BURST TEST RESULTS
The main purpose of this section is to compare unrestrained with restrained burst
test results and to show how unrestrained burst pressures are lower than the burst
pressures obtained when using a restraining fixture. Even though the package size
effects will be discussed in section II, the results in this section also point out how
the burst values within the unrestrained mode and within the restrained mode vary

inversely proportional with package size.

UNRESTRAIN ED RESULTS:

Table 6. Unrestrained Results for Blisters

 

 

 

 

 

 

PKG n Avg. Std C. of Min Max Range Burst/Perimeter
# (in. H20) Dev Var (Burst/inch)
W»)
1 50 121.4 10.21 8.41 95.0 147.4 52.4 9.0
2 50 73.1 3.63 4.97 64.1 82.1 18.0 3.7
3 50 39.0 2.68 6.87 31.2 44.3 13.1 1.4
4 50 38.4 1.91 4.97 33.9 42.4 8.5 1.2

 

 

 

 

 

 

 

 

 

 

One (1) variable: Package Configuration
Seal Perimeter Package #1 = 13.5”
Seal Perimeter Package #2 = 20.0”
Seal Perimeter Package #3 = 27.5”
Seal Perimeter Package #4 = 33.0”

It can be seen from the table above that the average burst values vary inversely
with the package size. The smaller package has a higher burst value than the bigger

package.

 

Table 7. Unrestrained Burst Test for Blisters
AN OVA One-way analysis - Package Size Effect

 

 

 

 

 

 

Source Degrees Sum
Of of of Mean F Fem p ...... Conclusion
Variation Freedom Squares Square “14".“
PKG PKG Size
Size 3 229432 76477 23 86.46 2.6507 0.000 Effect
Error 196 6281 . 1 32 --- --- --- «-
Total 199 235713 --- --- -- --- ---

 

 

 

 

 

 

 

 

The one-way analysis of variance in the table above shows statistically significant

differences between package size. So, package size affects the unrestrained average

burst test values.

The t-test comparison of PKG #3 and PKG #4 resulted in a p value = 0.20.

Therefore, packages 3 and 4 are not statistically different from each other in an

unrestrained burst test. All other t-test results showed significant differences, so

packages 1 and 2 are different from each other and from packages 3 and 4. Differences

in shape and angles in the four types of blisters or trays could be responsible for these

results.

 

 

RESTRAINED RESULTS:

Table 8. Restrained Results for Blisters

47

 

 

 

 

 

 

PKG n Avg. Std C. of Min Max Range Burst / Gap
# (in. H20) Dev Var Perimeter (in)
(%) (Burst/inle
1 50 153.7 10.72 6.97 128.4 190.7 62.3 11.4 .14
2 50 113.2 8.37 7.39 84.1 135.0 50.9 5.7 .20
3 50 114.5 7.40 6.46 98.6 131.9 33.3 4.2 .15
4 50 86.8 3.56 4.10 76.5 94.0 17.5 2.6 .01

 

 

 

 

 

 

 

 

 

 

 

Two variables: Package Configuration and Gap Height. The variables cannot be
separated for statistical analysis because the experiment design does not allow it.
Seal Perimeter Package #1 = 13.5”
Seal Perimeter Package #2 = 20.0”
Seal Perimeter Package #3 = 27.5”
Seal Perimeter Package #4 = 33.0”

In this case it is not possible to estimate the package effect and the gap effect
individually because the two variables are changing at the same time.

The t-test comparison of PKG #2 and PKG #3 resulted in a p value = 0.41.
Therefore, packages 2 and 3 are not statistically different from each other in a restrained
burst test. All other t-test results showed significant differences. Because the gap and
package size were confounded, valid analysis for gap and size was impossible.
Therefore, another experiment was designed, and the analysis of this one is reported on

page 55.

48

COMPARISON BETWEEN UNRESTRAINED & RESTRAINED RESULTS:
A two-way analysis of variance was performed for each package configuration to
see the effect of test method (unrestrained Vs restrained) and day on the results. We
found statistical evidence of differences between test methods and no statistical
evidence that days affected the burst test values. For all reported analyses we pool the

samples fi'om the separate days for each package.

Table 9. Restrained Vs Unrestrained Results for Blisters

 

 

 

 

 

 

PKG Configuration n Unrestrained Restrained Ratio
Average Average
(in. H20) (in. H10)
PKG #1
Accessories 50 121.4 153.7 1.3
PKG #2
Thera Small Outer IPG 50 73.1 113.2 1.6
PKG #3
Standard Leads Outer 50 39.0 114.5 2.9
PKG #4
Myocardial Leads Outer 50 38.4 86.8 2.3

 

 

 

 

 

 

Table 9, above, shows that, at each package configuration, restrained average

burst pressures are higher than unrestrained average burst pressures.

Table 10. Overall Package Size and Test Method Effect on Burst Test Results for
Blisters - AN OVA Two-way Anal sis

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source Degrees Sum
of of of Mean F F p Conclusion
Variation Freedom Squares Square W cm value
PKG PKG Size
Size 3 317044 105681 2229.6 2.63 0.000 Effect
Test . Test Method
Method 1 240850 240850 5081.2 3.87 0.000 Effect
Interaction Interaction
3 26384 8794.6 185.54 2.63 0.000 Effect
Error 392 18576 47.4 --- --- --- «-
Total 399 602854 --- --- --- --- ---

 

 

 

49

The two-way analysis of variance in the table above shows strong statistical
evidence of differences between packages and test methods. So, in general, test
methods and package size, both affect the average burst test results. This analysis
also demonstrated an interaction between package and test method, but interaction
accounted for a relatively small percentage of the variation.

Figure 14, next page, shows the relationship between average unrestrained and
restrained burst test values Vs package seal perimeter (package size). It can be seen
from this figure that within the restrained testing mode, the burst pressure decreases as
the seal perimeter increases. The same behavior was observed when the package was
tested in an unrestrained mode. This figure also shows that restrained burst pressure
was higher than the unrestrained burst pressure for all package sizes.

Figures 15 and 16, in the following pages, present box plots of the burst pressure for
four different package sizes for unrestrained and restrained burst test, respectively.
These plots indicate that an increase in package seal perimeter produce a lower
unrestrained burst pressure. These box plot also shows the variability of the
unrestrained burst pressure within each package size (seal perimeter) as well as the
variability between different package sizes. It can be seen in both figures that the
distribution of unrestrained burst pressure at a particular package size is reasonably
symmetrical, and the variability in unrestrained burst pressure appears to be higher for
the smaller packages than for the bigger ones. Also, variability seems to be greater for
restrained than for unrestrained. Tables 11 and 12 summarize the analysis for variances

and coefficients of variation.

50

mm

 

r commabmom l 6955.58.59 x _

Amazofivaouofitom 13m anew—oak

 

 

 

 

 

     

OM MN ON m~ Ow
. 33.: u an
3.93 + 53.3 - "~88... u .m
ems; u an ,
M32 + nave - «548... u a _. -

 

 

38%:— a8 pogo—5.3m 13m away—cam .m>
c.5295 «35m configured can 6055.53.33 93.33.. .3 259%

o
m
cm em
9
ow ..aua
on m
ow W
o2 m
u
o2 )
m.
E: H
o2 m
cm:

 

. lliCavr-r: ~ VN ®K-UWHN

51

$225 “Beacon Rom owmxomm

 

 

 

 

 

 

 

0.? Ohm cdm W?
h b _ _
_ om
fl log
.. 102

 

 

 

magmzm 8m SBEEA 36m owed—0mm .m> 8385 35m uofiabmofib .2 Semi

(OZH ur) omssord 15mg poutensoru f1

52

A8325 coaofitom Eom owed—och

ogmm mfim odm Wmfi
_ _ _ _

*

#

 

 

_ ; loofi

 

 

 

 

 

 

 

 

 

 

lo:

 

 

 

 

 

 

 

a ll com

28m=m com aoHoEtom Rom owed—own .m> 23.85 55m cofimbmom .3 Eswfi

(OZH u!) omssord 15mg poutensag

 

 

53

COMPARISON BETWEEN VARIANCES:

Table 11. Test for equality of two variances for Blisters
(F— test for Unrestrained Vs Restrained)

 

 

 

 

 

Package Unrestrained Restrained F-Ratio p-value
(Variance) (Variance)

1 104.24 114.92 1.10 7.34E-01

2 13.18 70.06 5.32 3.28E-08

3 7.18 54.76 7.62 4.76E-11

4 3.65 12.67 3.47 2.58E-05

 

 

 

 

 

 

 

The F-ratio was calculated dividing the higher variance over the lower variance.
The calculated two-sided p values were compared against 0.05. It can be seen in the
table above that the variances for packages #2, #3, and #4 are statistically different. For

all cases, the restrained variances are higher than the unrestrained variances.

COMPARISON BETWEEN COEFFICIENTS OF VARIATION:

Table 12. Comparison between Coefficients of Variation for Blisters
(Unrestrained Vs. Restrained)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Package #1
Test Mode Standard Average Coeff. Of Standard
Deviation (in. HzO) Variation Error (CV)
Unrestrained 10.21 121.40 8.41 0.8469
Restrained 10. 72 153.70 6.97 0.7008
Comparing Difference Std Error Z-ratio p—value
in CVs (Difference)
UR&R 1.44 1.10 1.31 0.19
Packa e #2
Test Mode Standard Average Coeff. Of Standard
Deviation (in. HzO) Variation Error (CV)
Unrestrained 3.63 73.10 4.97 0.4978
Restrained 8.37 113.20 7.39 0.7434
Comparing Difference Std Error Ratio p-value
in CVs (Difference)
UR & R 2.43 0.89 2.71 0.01

 

 

 

 

 

hi

$12

for

 

54

Table 12. Comparison between Coefficients of Variation for Blisters
(U nrestrained Vs. Restrained) Continuation

Package #3
Test Mode Standard Average Coeff. Of Standard

Deviation (in. HzO) Variation Error (CV)

 

 

 

 

 

 

Unrestrained 2.68 39.00 6.87 0.6904
Restrained 7.40 1 14.50 6.46 0.6490
Comparing Difference Std Error Ratio p-value
in CVs (Difference)
UR & R 0.41 0.95 0.43 0.67
Package #4

 

Test Mode Standard Average Coeff. Of Standard
Deviation (in. H20) Variation Error (CV)

 

 

 

 

Unrestrained 1.91 38.40 4.97 0.4986
Restrained 3 . 56 86.80 4.10 0.4108
Comparing Difference Std Error Ratio p-value

in CVs (Difference)
UR& R 0.87 0.65 1.35 0.18

 

 

 

 

 

 

 

In order to compare coefficients of variation we used the standard errors of each
coefficient of variation [16] and the root mean square formula to determine the standard
error of the difference. The difference in coefficients of variation was calculated as the
higher coefficient of variation minus the lower coefficient of variation. The statistical
significance was determined using the standardized difference called the Z-ratio and
standard normal distribution. The obtained two-sided p values were compared against
0.05. It can be seen from the table above that package #2 was the only one that shows
statistical difference between the coefficients of variation. The coefficient of variation
for package #2 for the restrained case is higher than the coefficient of variation for the

unrestrained case.

 

55

II. PACKAGE SIZE AND GAP SIZE EFFECT

 

The results that will be shown in this section will demonstrate the effects of

changing the package (blister) and gap size on the burst pressure. It will be seen

that both, package size and gap size, vary inversely proportionally with burst

pressure.

GENERAL RESULTS:

Table 13. Restrained Results for Blisters — Package Size and Gap Size Effects

‘ Package #1 Accessories Packa e (P/N 119401-001)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gap n Avg. Std Coeff. Min Max Range Burst per
Size (in.HzO) Dev of Var. Perimeter
(i9 ("/o) (Burst/Inch)
0.20 24 124.10 8.41 6.78 109.30 139.90 30.60 9.19
0.10 24 151.93 11.16 7.34 125.00 170.70 45.70 11.25
0.01 24 235.70 13.56 5.75 214.20 265.20 51.00 17.46
Seal Perimeter = 13.5 inch

‘ Packa§#2 Thera Small Outer Packa e P/N 119679-0fl
Gap n Avg. Std Coeff. Min Max 1 Range Burst per
Size (in.HzO) Dev Of Var Perimeter
(in.) (%) (Burst/Inch)
0.20 24 104.07 8.00 7.65 85.50 116.90 31.40 5.24
0.10 24 130.63 5.19 3.97 120.20 139.70 19.50 6.53
0.01 24 157.77 4.55 2.88 150.60 164.70 14.10 7.89

 

Seal Perimeter = 20.0 inch

It can be seen from Table 13 that average burst values vary inversely with the

gap size. Smaller gaps produce higher burst values. Also, it can be noticed that the

average burst value at any gap size is different for different package geometries. The

smaller package (seal perimeter) has a higher average burst value than the bigger

package.

 

 

 

 

56

STATISTICAL RESULTS:

A two-way analysis of variance was performed for each package configuration
to see the effect of gap and day on the results. We found statistical evidence of
differences between the gap sizes. On the other hand we found no statistical evidence
that days affected the burst test values. For all reported analyses we pool the samples

fiom the separate days for each package.

Overall Package and Size Effects

Table 14. Overall Package Size and Gap size Effect on Burst Test Results
For Blisters - AN OVA Two way Analysis

 

 

 

 

 

 

 

 

 

Source Degrees Sum l
of of of Mean F F p Conclusion
Variation Freedom Squares Square W ...-me... value
PKG PKG Size
Size 1 56894 56894.2 695.53 3.91 0.000 Effect
Gap . Gap Size
Size 2 170359 85179.3 1041.3 3.06 0.000 Effect
Interaction Interaction
2 26236 13118.0 160.37 3.06 0.000 Effect
Error 138 1 1292 81.8 --- --- ---
Total 143 264781 --- --- --- ---

 

 

 

 

 

 

 

 

This two-way analysis of variance shows strong statistical evidence of
difference in average burst value between packages and among gap sizes. This

analysis also shows an interaction between package and gap sizes.

 

57

Figure 17, next page, shows the relationship between burst pressure and gap
size. It can be seen fi'om this figure that for both packages the burst pressure decreased
as the gap size increased. This figure also shows that package #1 with a seal perimeter
of 13.5 inches required higher burst pressure to break than package #2, which has a seal
perimeter of 20.0 inches. So, burst pressure varies inversely proportionally with
package and gap size.

Figures 18 and 19, on the following pages, presents box plots of the burst
pressure at three different gaps for packages #1 and #2, respectively. These plots
indicate than an increase in gap size produces lower burst pressures. These box plots
also show information about the variability within and between gap sizes. The
distribution of burst pressure at a particular gap, for package #1, was reasonably
symmetrical for gaps 0.10” and 0.20”. Also, the variability that was found at each gap
was very similar. For package #2, the distribution of burst pressure at a particular gap
was reasonably symmetrical for gaps 0.01” and 0.10”. The variability that was found at
gap = 0.20” was greater than for the other two gaps. Also see Tables 15, 16 and 17.

Figures 20, 21, and 22 are also box plots but presented in a different way: burst
pressure Vs. package seal perimeter at gaps 0.20”, 0.10”, and 0.10”, respectively.
These plots provide information about variability within and between different package
sizes. The three figures show that, for the three gaps, the distribution of burst pressure
within a certain package size was reasonably symmetrical. The variability between

package sizes differed more at gaps 0.10” and 0.01”. Also see Tables 15, 16 and 17.

 

58

mad

 

13H x :3 a a .-
$2.25 ofim new

Nd 2d fio mod

_ _ _ _

 

 

a u ”m
923 + $32 - "was”: u »

 

 

a n R
a}: + $.82 - ”2.3% u a

1 1
O
'13

l

l

 

 

Enema—m hem 3mm .30 .m> 9:58am «PEM— owaao>< .5 «Sara

O

03

QC

0
O
(\l

omm

(011.1 or) amssord rung afieraAV

 

59

90:25 8% new

cmd 3.0 5.0
L _ _ 1 oh:

 

 

 

 

 

 

 

To?

 

 

 

 

1 com

 

 

 

1 GR

 

 

 

 

 

Em owed—own - macaw—m Sm 05m 95 .m> 0.5%on ampsmdfi oSwE

(OZH u!) amssord 1s1ng

 

 

 

60

Amazons 35 new

omav cage zoae

L _ _ flow
I.om
1.oo_
iloaz
i.o~_
i.om~
_ l.ov_
.iom_
i.oe_
[lon_

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ma owed—own - Evasm how 3% new .m> 2:3on 5:5 .2 oSwE

(OZH u!) ornssord 13mg

 

 

61

Amazon: beaten 18m owmxomm

odm m4:

_ a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

..omd H 30 - meek—m Sm 338th Row owed—0mm .m> 2335 3:5 .

1% mu
18 m.
3
p
12: m
Io: m
m
roe m
9
1o: m.
H
Z
.19: m
82sz

 

 

62

EC Mahatma Rom owed—own

AVON mwmfi
_

 

iuom_

 

{IOMH

 

 

 

 

_ i.ov~

 

 

[Iomfi

 

 

 

iloofi

 

 

ilohfi

 

 

:36 H 95 - doaofitom Rom owed—own .m> 838.5 “EH—m AN 235

(OZH u!) ornssord 1s1ng pourensou

 

 

63

Amazon: Mahatma Rom owed—8m

 

 

 

 

 

 

 

 

 

odN mg: H

_ _ w

_ l Om~ n

a m.
H _ 4 w
p

8

m

loom pa

9

L m

9

m.

1 OR m

(Ox

 

 

 

.. 36 n 95 - moi—m 8m “Beacon Eom owed—0mm .m> 2385 “wasm .Nm 2&3

 

64

COMPARISON BETWEEN VARIANCES:

Table 15. Comparison between Variances for Blisters

Package Size and Gap Size Effects
(Bartlett’s Test for Homogeneity of Variances)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Package #1
Gap Standard Variance In Chi-Square p-value
(inches) Deviation (Variance)
0.20 8.41 70.73 4.26 5.06 7.95E—02
0.10 11.16 124.55 4.82
0.01 13.56 183.87 5.21
Average 126.38 4.77
Package #2
Gap Standard Variance In Chi-Square p-value
(inches) Deviation (Variance)
0.20 8.00 64.00 4.16 8.45 1.46E-02
0.10 5.19 26.94 3.29
0.01 4.55 20.70 3.03
Average 37.21 3.49

 

The Bartlett’s test for homogeneity of variances [16] is suggested when there are

more than two groups of variances to be compared. With this test an overall

comparison between the variances can be done. When comparing the obtained

p value with 0.05 it can be seen that there is no statistical difference between the

variances coming from different gaps for package #1. On the other hand, there are

statistical differences between variances for package #2.

 

 

 

 

Table 16. Pairwise Comparisons of Variances for Blisters
(Package Size and Gap Size Effects)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Package #1
Variance Variance at Variance at
at Gap = 0.20" Gap = 0.10" Gap = 0.1"
70.73 124.55 183.87
Pairwise F-Ratio p-value
Comparisons
0.20&0.10 1.76 0.18
0.20 & 0.01 2.60 0.03
0.10 & 0.01 1.48 0.36
Package #2
Variance at Variance at Variance at
Gap = 0.20" Gap = 0.10" Gap = 0.1"
64.00 26.94 20.70
Pairwise F-Ratio p-value
Comparisons
0.20 & 0.10 2.38 0.04
0.20 & 0.01 3.09 0.01
0.10& 0.01 1.30 0.53

 

The F-ratio was calculated dividing the higher variance over the lower variance.
For multiple pairwise comparisons of k treatments, p values less than [0.05/(k*(k-1)/2)]
were regarded as significant, the Bonferroni approach to multiple comparisons. For our
application k = 3 so the critical p value is .05/3 = 0.017. The obtained p value was
compared against 0.017. It can be seen in the table above that in package #2 the

comparison between gap 0.20” and gap 0.01” show a p value lower than the critical

value (0.017).

 

COMPARISON BETWEEN COEFFICIENTS OF VARIATION:

 

66

Table 17. Pairwise Comparisons of Coefficients of Variation for Blisters
(Package Size and Gap Size Effects)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Package #1
Gap Standard Average Coeff. Of Standard
(inches) Deviation (in. H2O) Variation Error (CV)
0.20 8.41 124.10 6.78 0.9826
0.10 11.16 151.93 7.35 1.0659
0.01 13.56 235.70 5.75 0.8331
Comparing Difference Std Error Z-ratio p-value
Pairs in CVs (Difference)
0.20 & 0.10 0.57 1.45 0.39 0.6949
0.20 & 0.01 1.02 1.29 0.79 0.4268
0.10&0.01 1.59 1.35 1.18 0.2392
Package #2
Gap Standard Average Coeff. Of Standard
(inches) Deviation (in. H20) Variation Error (CV)
0.20 8.00 104.07 7.69 1.1161
0.10 5.19 130.63 3.97 0.5744
0.01 4.55 157.77 2.88 0.4166
Comparing Difference Std Error Z-ratio p-value
in CVs (Difference)
0.20 & 0.10 3.71 1.26 2.96 0.0031
0.20 & 0.01 4.80 1.19 4.03 0.0001
0.10 & 0.01 1.09 0.71 1.53 0.1248

 

In order to compare coefficients of variation we used the standard errors of each

coefficient of variation [16] and the root mean square formula to determine the standard

error of the difference. The differences in coefficients of variation were calculated as

the higher coefficient of variation minus the lower coefficient of variation. The

statistical significance was determined using the standardized difference called the

Z-ratio and standard normal distribution.

 

 

67

For multiple comparisons of k = 3 treatments p-values less than 0.017 were
regarded as significant, the Bonferroni approach to multiple comparisons. For package
#2, the pair comparison between gap 0.20” and 0.10”, and between 0.20” and 0.01”

show a p value lower than 0.017.

 

68

HI. SUMMARY OF THE RESULTS FOR BLISTERS:

1. Restrained burst test pressures are higher than unrestrained burst test pressures.

2. The burst value varies inversely with the gap size. Smaller gaps produce higher
burst values.

3. In general, the package with smaller seal perimeter produces higher burst pressures
than the bigger packages. This behavior is also true for the restrained burst test
method. The burst value at any gap size is higher for smaller packages than for
bigger packages.

4. There was no pattern in the difference in variation between restrained and
unrestrained burst tests:

a. There is statistical difference in raw variances between restrained and
unrestrained burst test for packages #2, #3, and #4. For all cases the restrained
variances are higher than the unrestrained variances.

b. There is no statistical difference in coefficients of variation between restrained
and unrestrained burst test for packages #1, #3, and #4. Package #2 was the
only one that shows statistically significant difference between coefficients of
variation. For package #2 the restrained coefficient of variation is higher than
the coefficient of variation for the unrestrained case.

5. There was no pattern in the difference in variation between gaps:

3. There is statistically significant difference in raw variance and coefficients of
variation between gaps for package #2. No statistical difference, in raw variance
and coefficients of variation, between gaps was found for package #1.

Differences in shape and geometry can explain this gap effect on variation.

 

 

69

PART B — POUCHES:

As mentioned before, the first section of Part B is intended to compare unrestrained and
restrained burst test results. In section H package size and gap size effects will be
studied. The last section of this chapter consists of a summary of the results obtained for

pouches.

I. UNRESTRAINED VS RESTRAINED BURST TEST RESULTS

UNRESTRAINED RESULTS:

Table 18. Unrestrained Results for Pouches

Unrestrained Chevron Up

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PKG n Average Std C. of Min Max Range Burst per
# (in. H20) Dev Var Perimeter
(%) (Burst/inch)
1 20 59.69 9.73 16.30 43.00 77.30 34.30 3.62
2 20 45.65 11.51 25.21 31.40 65.60 34.20 2.40
3 20 41.55 6.85 16.49 26.20 49.90 23.70 1.93
Unrestrained Chevron Down
PKG n Average Std C. of Min Max Range Burst per
# (in. H20) Dev Var Perimeter
(%L (Burst/inch)
1 20 48.74 10.49 21.52 38.30 69.80 31.50 2.95
2 20 45.77 14.15 30.92 29.80 65.70 35.90 2.41
3 20 43.32 5.52 12.74 33.80 50.80 17.00 2.01

 

Seal Perimeter Package #1 = 16.5”
Seal Perimeter Package #2 = 19.0”
Seal Perimeter Package #3 = 21.5”

It can be seen from the table above that the average burst values vary inversely

with the package size. The smaller package has a higher burst value than the bigger

package.

 

70

Table 19. Overall Package Size and Chevron Effect in Unrestrained Burst Test
Results for Pouches - ANOVA Two-way Analysis

 

 

 

 

 

 

 

 

Source Degrees Sum
of of of Mean F F p Conclusion
Variation Freedom Squares Square we...“ cm value
PKG PKG Size
Size 2 2957 1479 14.50 3.08 0.000 Effect
Chevron No Chevron
Effect 1 274 274 2.69 3.92 0.104 Effect
Interaction Interaction
2 958 479 4.70 3.08 0.011 Effect
Error 114 11681 102 --- --- --- ---
Total 1 19 1587 1 --- --- --- --- «-

 

 

 

 

 

 

 

 

 

The two-way analysis of variance in the table above shows strong statistical
evidence of differences between packages. 0n the other hand, no statistical evidence of
difference between chevron up and chevron down was found. So, in general, package
size affects the average burst test results and the way the chevron is positioned
does not. This analysis also demonstrated a weak interaction between package and
chevron position but interaction accounted for a relatively small percentage of the
variation.

Figure 23, shows the relationship between unrestrained burst test values Vs package
seal perimeter. This figure shows that as the package seal perimeter increases the
unrestrained burst pressure decreases. It also shows that for package #1, with seal
perimeter 16.5”, there was a difference in burst pressure between chevron up and
chevron down. On the other hand, no difference was found between chevron up and
chevron down for packages #2 and #3. A possible reason for the difference in results
could be because packages, when tested unrestrained, do not deform in the same way

each time. It is possible that during this test package #1 varied more in deformation

 

 

71

than the other two packages. This could be the source of the weak interaction effect
found in the analysis of variance

Figures 24 and 25, on the following pages, are box plots which present unrestrained
burst pressure Vs. package seal perimeter for chevron up and chevron down,
respectively. In both cases it can be seen that as the package seal perimeter increased
the unrestrained burst pressure decreased. Both plots show that package #2 (seal
perimeter = 19.0”) was the one with higher variability and package #3 (seal perimeter
21.5”) the one with lower variability. When comparing both figures it can be seen that
there is not much difference between the two, meaning that there is not much difference
in variability due to chevron effects. Tables 23 to 26 summarize the analyses for the

variances and coefficients of variation.

 

72

 

:Ben— :e...>o._U ab 59525
353.5353 x 355.5353 I

  

Ame—~25 hogan—tom 13m owns—cam

NN 5N om 3 ma 2 3

_ _ _ LI‘

2

 

fl u "m
vwméw + xwvcwfi n Nxcfivcd H %
fill/I l.

  

5 n R
8.8.. + 36%.? - mama; u »

 

.3

ON

-om

.ov

.8

on

 

 

muu—Oflcam hOH

on

hogan—tom Rom owS—onm .m> 0558.5 35m 105330.:— D 09293.. .MN «...—ME

(01H ur) arnssard rsmg

paurensarufl ofiaraAV

 

 

73

A8525 “825:5 Rom ommV—omm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

WAN od~ m.©~ fl
_ b F m
w

_ 1 cm m.
9

_ E M.
m

- .. .

‘ m

I 1 8 m

9

l on m.

‘ H

m.

 

 

 

n5 cot/25 - monosom 8m 888:3 Rom ammo—0mm .m> 8385 35m MD .3. 835m

74

 

Genoa: “808:3 Rom owe—8m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2m 0.2 m2 0
_ _ _ w
W {cm m

n

B

m.

low P

8

g % _ m
_ 1% mu
m

_ m.

H

z

{E m

EsoD 5325 - monosom ..om bBEEm 13m owmxomm .m> 2:395 BSm MD .3 Eswfi

RESTRAINED RESULTS:

Table 20. Restrained Results for Pouches - per Gap Size

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gap = 1.0”
PKG n Average Std Coeff. Min Max Range Burst per
# (in. H20) Dev Var Perimeter
(%) (Burst/inch)
1 20 84.18 12.94 15.37 66.90 113.20 46.30 5.10
2 20 73.24 6.81 9.30 62.80 86.70 23.90 3.85
3 20 64.85 9.12 14.06 48.10 81.00 32.90 3.02
Gap = 0.75”
PKG n Average Std Coeff. Min Max Range Burst per
# (in. H20) Dev Var Perimeter
(%) (Burst/inch)
1 20 102.22 13.63 13.33 73.20 128.70 55.50 6.20
2 20 91.40 13.82 15.12 65.90 114.90 49.00 4.81
3 20 76.72 9.48 12.36 58.90 90.70 31.80 3.57
Gap = 0.625”
PKG n Average Std Coeff. Min Max Range Burst per
# (in. H20) Dev Var Perimeter
(%) (Burst/inch)
1 20 111.05 11.90 10.72 87.60 128.10 40.50 6.73
2 20 107.97 9.89 9.16 84.70 125.70 41.00 5.68
3 20 89.38 8.94 10.00 69.40 104.90 35.50 4.16
Gap = 0.50”
PKG n Average Std Coeff. Min Max Range Burst per
# (in. H20) Dev Var Perimeter
(%) (Burst/inch)
1 20 124.37 10.17 8.18 105.7 143.7 38.00 7.54
2 20 122.33 6.59 5.39 105.5 133.6 28.10 6.44
3 20 102.54 7.65 7.46 85.30 116.3 31.00 4.77
Gap = 0.25”
PKG n Average Std Coeff. Min Max Range Burst per
# (in. H20) Dev Var Perimeter
(%) (Burst/inch)
1 20 196.42 15.72 8.00 166.9 224.2 57.30 11.90
2 20 195.89 18.38 9.38 147.1 224.7 77.60 10.31
3 20 175.25 22.11 12.62 144.7 221.0 76.30 8.15

 

 

 

 

 

 

 

 

 

Seal Perimeter Package #1 = 16.5”
Seal Perimeter Package #2 = 19.0”
Seal Perimeter Package #3 = 21.5”

 

76

It can be seen from the table above that the average burst values vary inversely
with the gap size. Smaller gaps produce higher burst values. The average burst value is
also different, at any gap size, for different package sizes. The smaller package has a

higher average burst test value than the bigger package.

 

77

COMPARISON BETWEEN UNRESTRAINED & RESTRAINED RESULTS:

A two-way analysis of variance was performed for each package size to see the

effect of test method (unrestrained Vs restrained) and day on the results. We found

statistical evidence of differences between test methods and no statistical evidence that

days affected the burst test values. For all reported analyses we pool the samples from

the separate days for each package.

Table 21. Overall Unrestrained Vs Restrained Burst Test Results for Pouches

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Package #1 (5” x 10”) - ANOVA One-way analysis
Source Degrees Sum
Of of Of Mean F F p ...... Conclusion
Variation Freedom Squares Square m _os,1,133
Test Test Method
Method 1 137747 137747 1 1 1.43 3 .9097 0.000 Effect
Error 138 170596 1236 --- --- --- ---
Total 139 308343 --- --- --- --- ---
Package #2 (7” x 11”) - AN OVA One-way analysis
Source Degrees Sum
Of of Of Mean F F p ...... Conclusion
Variation Freedom Squares Square «IA-Led 054,138
Test Test Method
Method 1 149984 149984 104.60 3.9097 0.000 Effect
Error 138 197871 1434 --- --- --- «-
Total 139 347855 --- --- --- --- ---
Packa e #3 (9” x 12”) - AN OVA One-way analysis
Source Degrees Sum
of of Of Mean F F p ...... Conclusion
Variation Freedom Squares Square «IE—M 0541133
Test Test Method
Method 1 100510 100510 82.76 3.9097 0.000 Effect
Error 138 167600 1214 --- --- --- ---
Total 139 268110 --- --- --- --- «-

 

 

 

78

The one-way analysis of variance in the table above shows evidence of statistical

difference between unrestrained and restrained test methods on each package

configuration. So, the test method does affect the average burst values.

Another one-way analysis of variance was performed in order to compare the

unrestrained method with the biggest gap. See the table below.

Table 22. Overall Unrestrained Vs Gap = 1.0” Burst Test Results for Pouches

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ANOVA One-way analysis
‘ Package #1 (5” x 10”)
Source Degrees Sum
Of of Of Mean F F p ...... Conclusion
Variation Freedom Squares Stutare “Jelly... 0541453
Test Test Method
Method 1 l 1 970 1 1 970 83 .96 4.0069 0. 000 Effect
Error 58 8269 143 --- --- --- ---
Total 59 20239 --- --- --- --- ---
. Package #2 (7” x 11”)
Source Degrees Sum
Of of Of Mean F F p "In. Conclusion
Variation Freedom Squares Square deflated 053,53
Test Test Method
Method 1 10102 10102 81.32 4.0069 0.000 Effect
Error 58 7205 124 --- --- --- «-
. Package #3 (9” x 12”)
Source Degrees Sum
Of of Of Mean F F p ...... Conclusion
Variation Freedom Squares Square we...“ 05,1,53
Test Test Method
Method 1 6697.6 6697.6 126.14 4.0069 0.000 Effect
Error 58 3079.7 53.1 --- --- --- ---
Total 59 9777.3 --- --- --- --- ---

 

 

79

The results in the table above show that the average unrestrained burst values, even
when compared with the highest gap, are statistically different from the average
restrained test results. In general, average restrained burst test results are higher than
average unrestrained burst test results.

Figures 26 to 30, on the following pages, show the relationship between
unrestrained and restrained burst pressure Vs package seal perimeters, for gaps = 1.0”,
0.75”, 0.625”, 0.50”, 0.25”, respectively. These five figures show that the unrestrained
and restrained burst pressures vary inversely proportional with package seal perimeter.
It is also shown that at each package size, restrained burst pressure was always higher
than the unrestrained burst pressure. Figure 31, shows all testing modes together in one
graph. The line on the top represents the restrained burst pressure for the smallest gap
(0.25”) for three different packages. The lines on the bottom represent the unrestrained
burst pressure (chevron up and down) for the three packages. This figure shows that
gap 0.25” is a lot higher than the other gaps. This is evidence that this gap represents a
special situation.

Figures 32 to 36, present box plots of burst pressure vs. package seal perimeter for
gaps 1.0”, 0.75”, 0.625”, 0.50”, and 0.25”, respectively. These plots show that the
variability at each package size tend to be similar in the first four gap but increases at
gap 0.25”. The variability between package sizes and within gaps did not follow a
specific pattern. Tables 23 to 26 summarize the analyses for the variances and

coefficients of variation.

 

80

NN

 

:9.— " 9:0 5:5 :e..>2—U a: :e..>o__U
eon—mahmom 4 3395033 0 nonmahumouub +

 

 

 

Ange—5 33::an 13m owed—cam

2

 

 

 

 

 

 

 

8 cm 2 M: t 8
a n «m -
£3 + £33 - "gong... u a -
H N.“ I

T
ilk! 8.8.. + Egan - mama; u a -
vnénm + “63.: - «wvcmd Hf ,

 

 

AL: H mawv 35.23.3— 6...“ 6035.583:
mono—5m .Su .63.:th .aom owe—cam .m> c.5295 «95m cup—95‘ 6N charm

o
2 W
3
J
om a
AU 3
cm“ H
H m
7» S
.6 ....
u
8 w
n
....
8 a
2
cm
as

 

81

 

:mbé n 95 :Ben— 3.52.0 AD :33.—U
3:35mom x 35953.53 9 35:58.5: +

 

 

 

Anon—9:5 33.5.?” 30m nus—cam

 

 

 

 

 

 

 

.8 s cm 2 2 t 2 2
_ _ _ _ _ _ o
. on
_ u N:
33.; + £48.” - "3:5... u a
”H - O?
I”; D
/ 1 co
_ u N:
/ 3.8.. + 33.3 - "£32 1 a - 8
u H mm lT
name» + 386 + Nuance..- u a - 2:
In
em:

 

 

 

Air... H :53 35.23% 6:“ 3:35.333

33.—em .8: 33:53am 3% anew—cam .m> 9.33.5 3.5m 99295.. EN ensure

(ozH u!)

amssard isrng afiquV

82

NN

 

is... u :5
339.30% 0

 

525 :33.—U
35958.5: 9

:D 3.32.0
35.58.59 +

 

Ame—.25 33.53am 73m owed—cam

3.. om

_ _

a w:

r pl

2 2 2

_ _

 

 

 

 

$.me ..

: u N:
33.; + £33 - "3:5... u a

’
J

g u N:
8.8.. + Sewn». - "mums; n »

 

l

 

 

has... u 88 852.8: 23 523.82:

mono—Sm :8 33.:29— .«om owéoum .m> 9:53.:— umusm owa3>< .3 953%

cm

ow

o0

ow

2:

cm:

(om u!)

amssard wing afieJaAV

 

UN! Gh-In-EAHo-Ah fiUh-nflh nihifihn‘xifis -AVN mvvh-hwmrw~

83

NN

 

:3... u 550 5.53 55350 a 3 5.32—U
355.53% x 35953.53 0 355.535 3 +

 

A3555 33.533 13m mung—3m
5 on 2 3 S 3 m5

5 _ _ 3 3 _

 

 

3 H «M— .
vwméw + maveefi . «5332.6 H %
”'lfuln“ »

ll.
_ u «m
3.3.. + 33.2 - mama; u 5

l

l

5 n ”5
- a 33. + "£5- .... 5

V.

‘4 L

 

 

 

Azcmd .I. 55% 355.53% 6:: 355.5353
335.5 .5.— 335533 .aom 352$ .m> 0.53.5 3.53 muu3>< .mN 35w:—

om

ow

ow

cm

2:

cg

o:

(ozH n!)

amssard rsmg aBBJaAV

UQF‘Q-uca LCM LH‘fiQFfl-mkcflh —F\.QV. fluofluv—oH-MAH -V> coin-UUQo-Ah iUL-INH Q=E.e-Q\i< n‘N. QL-uhUmryh

84

 

:3..." 95 :35: 5.32.0 .5 5.525
35.53% I 35933.5: 9 35953.5: +

 

33:25 .335th 13m away—cam

 

 

 

 

 

 

mm 3.. cm 3 wfi : E 2
_ _ _ _ _ fl " N“ _ o
l > me.fiw + N®V©©oN I NNWMVGAv nu %
- fl Jq - om
g H «M
mcécv + nevwdm - uxumabé n h - o3
- cm 5
_ u R ._ I: - 8N
325m - 36m + «uwwocé- H .m
i. omm

 

 

Fm”: u QQUV 359.3% 6:“ 6053335:
352:5 .8“ hogan—tom 13m owaxuam .m> ouammuhm 3.5m uwauo>< .3 «...—am..—

(ozH u!)

aJnSSOJJ wing afieJaAV

85

mu

:3..." 55 :8... u 55 :33. u 56 :3... u 55 ...... u 55 ES: 5:35

35.53.. I 852.85 x 352.35 0 35.5.3. x 352.35 < 85.5.2.5 9 35.585: +

 

.5 28.—>25

 

A3595v 33.5.3.— _3m nus—cam

 

 

 

 

 

 

 

 

 

 

5N an 3 3 S 3 m—
l >
4 J‘"\’ Ill»
L, J.‘
T
T u ‘4...
T
LI! u I

 

 

 

553% 355.53% 5:. 35.53.55
3528.— ..8 3.25.3..— _aom ens—2E .m> 9533...— «3—5 02.33. .3 0.53%

c
-..m m
u
do
a
m.
.5: m
.d
.J
m
,3. m
m.
-55 m
:3

86

A3525 33.555 13m 03525

mém 0.3 WE
F _ . I m?

 

_ L.

 

5 [mo

 

 

 

 

 

 

 

 

 

 

“ A 11m:

 

1mm

 

 

s5l1- ..mo
{mo.

 

 

 

Tm:

 

:05 H 9.0 - 3:255 85 33555 73m 386.5 .m> 2335 355 .Nm 2sz

(OZH u!) amssaxd 13mg

87

A3525 335:5 53m 3835

mAN QE WE
_

_ _

 

I 8
1 E
I ow
5 I 8
.18.
. ¥
18.
51.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

..mnd H 9.0 .. 3:285 .55 33555 Rom owmxomm .m> 2335 355 .mm 2:me

(OZH u!) amssexd 13mg

88

A3525 522555 How owmxomm

mém 9% WE
_ 5 5

 

rem

{ow

 

 

 

$0

 

 

 

-oo 5

 

 

 

 

 

 

b:

 

{as

 

 

 

 

 

Tam _
..mmwd H 956 - 352:5 85 522555 Eom 3.3.25 .m> 2335 3:5 6m 2&5

 

(QZH u!) emssexd 15mg

89

@383 388:3 Rum mug—8m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EN 92 m9
w _ F
{a
H
m
13 .6.
Na
{we m
l m
In: W
fl # u
f _ LS m
w 13 (
In:

 

 

 

:36 H 95 - 8:0:an Sm Baofitom Rom owmxomm .m> 2385 Esm .mm 233m

90

Amazons bfifitom Rom «mg—0mm

mém 92
_ %

WE

 

*

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o¢_
I.Om_
1.oo_
1.05_
v.ow~
l.om_
..oom
r}o_~
[.omm
1.0mm

 

 

(OZH u!) amssexd 15mg

..mmd H 95 .. monosom Sm 588:3 Rom owmxomm .m> 2:3on 3:5 .3. oSwE

Table 23. Comparison between Variances for Pouches

91

COMPARISON BETWEEN THE VARIANCES:

(Bartlett’s Test for Homogeneity of Variances)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

, Package #1
Test Standard Variance In Chi-Square p—value
Mode Deviation (Variance)
URCU 9.73 94.67 4.55 7.16 3.06E-01
URCD 10.49 110.04 4.70
1.0" 12.94 167.44 5.12
0.75" 13.63 185.78 5.22
0.625" 11.90 141.61 4.95
0.50" 10.17 103.43 4.64
0.25" 15.72 247.12 5.51
Average 150.01 4.96
Package #2
Test Standard Variance In Chi-Square p-value
Mode Deviation (Variance)
URCU 11.51 132.48 4.89 30.73 2.86E-05
URCD 14.15 200.22 5.30
1.0" 6.81 46.38 3.84
0.75" 13.82 190.99 5.25
0.625" 9.89 97.81 4.58
0.50" 6.59 43.43 3.77
0.25" 18.38 337.82 5.82
Average 149.88 4.78
Package #3
Test Standard Variance In Chi-Square p-valne
Mode Deviation (Variance)
URCU 6.85 46.92 3.85 56.87 1.94E-10
URCD 5.52 30.47 3.42
1.0" 9.12 83.17 4.42
0.75" 9.48 89.87 4.50
0.625" 8.94 79.92 4.38
0.50" 7.65 58.52 4.07
0.25" 22.11 488.85 6.19
Average 125.39 4.40

 

 

 

92

The Bartlett’s test for homogeneity of variances [16] was used to make an

overall comparison between the variances of different testing modes. This test was

performed for each package. When comparing the obtained p values with the critical p

value (0.05) it can be seen that package #2 and package #3 show statistical difference

between the variances. When looking at the data closely we realized that the variance

obtained for gap = 0.25” was a lot higher than the variance obtained for the other gaps.

Since we thought that this gap was being responsible for the difference between the

variances, we ran a Bartlett’s test excluding gap 0.25” data.

Table 24. Comparison between Variances for Pouches

(Bartlett’s Test for Homogeneity of Variances w/o gap = 0.25”)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Package #1
Test Standard Variance Ln Chi-Square p—value
Mode Deviation (Variance)
URCU 9.73 94.67 4.55 3.63 0.60
URCD 10.49 110.04 4.70
1.0" 12.94 167.44 5.12
0.75" 13.63 185.78 5.22
0.625" 11.90 141 .61 4.95
0.50" 10.17 103.43 4.64
Average 133.83 4.86
. Package #2
Test Standard Variance Ln Chi-Square p-value
Mode Deviation (Variance)
URCU 11.51 132.48 4.89 19.44 0.0016
URCD 14.15 200.22 5.30
1.0" 6.81 46.38 3.84
0.75" 13.82 190.99 5.25
0.625" 9.89 97.81 4.58
0.50" 6.59 43.43 3.77
Average 118.55 4.60

 

 

93

Table 24. Comparison between Variances for Pouches - Continuation
(Bartlett’s Test for Homogeneity of Variances w/o gap = 0.25”)

 

 

 

 

 

 

 

 

Package #3
Test Standard Variance Ln Chi-Square p-value
Mode Deviation (Variance)
URCU 6.85 46.92 3.85 7.49 0.1868
URCD 5.52 30.47 3.42
1.0" 9.12 83.17 4.42
0.75" 9.48 89.87 4.50
0.625" 8.94 79.92 4.38
0.50" 7.65 58.52 4.07
Average 64.81 4.1 1

 

 

 

 

 

 

 

 

The results on the table above show that gap 0.25” was being responsible for the
differences in variance for package #3. This was not the case for package #2. We found
statistical difference between the variances for package #2. In order to follow up our
Bartlett’s test results for package #2 we ran an F -test for individual comparisons

between pairs of variances.

Table 25. Pairwise Comparisons of Variances for Pouches - Package #2

 

 

 

 

 

 

 

 

 

Test Mode Standard Variances
Deviation

URCU 11.51 132.48

URCD 14.15 200.22
1.0" 6.81 46.38

0.75" 13.82 190.99
0.625" 9.89 97.81
0.50" 6.59 43 .43

0.25" 18.38 337.82

 

 

 

 

94

Table 25. Pairwise Comparisons of Variances for Pouches — Package #2
Continuation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Comparisons F-ratio p-value
URCU & URCD 1.51 0.3760
URCU & 1.0" 2.86 0.0271
URCU & 0.75" 1.44 0.4326
URCU & 0.625" 1.35 0.5148
URCU & 0.50" 3.05 0.0192
URCU & 0.25" 2.55 0.0478
URCD & 1.0" 4.32 0.0025
URCD & 0.75" 1.05 0.9191
URCD & 0.625" 2.05 0.1272
URCD & 0.50" 4.61 0.0016
URCD & 0.25" 1.69 0.2632
1.0" & 0.75" 4.12 0.0034
1.0" & 0.625" 2.11 0.1124
1.0" & 0.50" 1.07 0.8877
1.0" & 0.25" 7.28 0.0001
0.75" & 0.625" 1.95 0.1537

[0.75" & 0.50" 4.40 0.0022
|o.75" & 0.25" 1.77 0.2230
F1625" & 0.50" 2.25 0.0848
[0.625" & 0.25" 3.45 0.0096
|0.50" & 0.25" 7.78 0.0000

 

 

 

 

The F -ratio was calculated by dividing the higher variance over the lower
variance. For multiple comparisons of k treatments, p values less than [0.05/(k*(k-1)/2)]
were regarded as significant, the Bonferroni approach to multiple comparisons. For our
application k=7 so the critical p value is .05/21= 0.0024. When comparing the obtained
p values with the critical p value (0.0024) it can be seen that what is causing the
difference in variance in package #2, besides gap 0.25”, is the difference in variation
between unrestrained chevron down mode and gap 0.50”. The reason for that is not
explained. It may be an artifact of the experiment. Further experimentation will be

required to determine the cause.

95

COMPARISON BETWEEN COEFFICIENTS OF VARIATION:

Table 26. Pairwise Comparisons of Coefficients of Variation for Pouches

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Package #1
Test Standard Average Coefficient Std Error
Mode Deviation (in. H20) of Variation (CVar)
URCU 9.73 59.69 16.30 2.64
URCD 10.49 48.74 21.52 3.56
1.0" 12.94 84.18 15.37 2.49
0.75" 13.63 102.22 13.33 2.15
0.625" 11.90 111.05 10.72 1.71
0.50" 10.17 124.37 8.18 1.30
0.25" 15.72 196.42 8.00 1.27
Package #l-Pairwise Comparisons
Pairwise Difference Std Error Z-ratio p-value
Comparisons in CVar (Diff)
URCU & URCD 5.22 4.43 1.18 0.2388
URCU & 1.0" 0.93 3.63 0.26 0.7980
URCU & 0.75" 2.97 3.41 0.87 0.3837
URCU & 0.625" 5.58 3.15 1.77 0.0764
URCU & 0.50" 8.12 2.95 2.76 0.0059
URCU & 0.25" 8.30 2.94 2.83 0.0047
URCD & 1.0" 6.15 4.34 1.42 0.1565
URCD & 0.75" 8.19 4.15 1.97 0.0487
URCD & 0.625" 10.81 3.95 2.74 0.0062
URCD & 0.50" 13.35 3.79 3.52 0.0004
URCD & 0.25" 13.52 3.78 3.58 0.0003
1.0" & 0.75" 2.04 3.28 0.62 0.5350
1.0" & 0.625" 4.66 3.02 1.54 0.1232
1.0" & 0.50" 7.19 2.81 2.56 0.0104
1.0" & 0.25" 7.37 2.79 2.64 0.0084
0.75" & 0.625" 2.62 2.75 0.95 0.3403
0.75" & 0.50" 5.16 2.51 2.06 0.0399
0.75" & 0.25" 5.33 2.49 2.14 0.0326
0.625" & 0.50" 2.54 2.15 1.18 0.2381
0.625" & 0.25" 2.71 2.14 1.27 0.2039
0.50" & 0.25" 0.17 1.82 0.10 0.9239

 

 

 

96

Table 26. Pairwise Comparisons of Coefficients of Variation for Pouches

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Continuation
Package #2
Test Standard Average Coefficient Std Error
Mode Deviation (in. H20) of Variation (CVar)
URCU 11.51 45.65 25.21 4.23
URCD 14.15 45.77 30.92 5.33
1.0" 6.81 73.24 9.30 1.48
0.75" 13.82 91.40 15.12 2.44
0.625" 9.89 107.97 9.16 1.46
0.50" 6.59 122.33 5.39 0.85
0.25" 18.38 195.89 9.38 1.50
Package #2-Pairwise Comparisons
Pairwise Difference Std Error Z-ratio p-value
Comparisons in CVar (Diff)

URCU & URCD 5.70 6.81 0.84 0.4024
URCU & 1.0" 15.92 4.48 3.55 0.0004
URCU & 0.75" 10.09 4.89 2.06 0.0389

URCU & 0.625" 16.05 4.48 3.59 0.0003
URCU & 0.50" 19.83 4.32 4.59 0.0000
URCU & 0.25" 15.83 4.49 3.53 0.0004
URCD & 1.0" 21.62 5.54 3.90 0.0001
URCD & 0.75" 15.80 5.87 2.69 0.0071

URCD & 0.625" 21.76 5.53 3.93 0.0001
URCD & 0.50" 25.53 5.40 4.72 0.0000
URCD & 0.25" 21.53 5.54 3.89 0.0001

1.0" & 0.75" 5.82 2.86 2.04 0.0417
1.0" & 0.625" 0.14 2.08 0.07 0.9470
1.0" & 0.50" 3.91 1.71 2.29 0.0223
1.0" & 0.25" 0.08 2.11 0.04 0.9680
0.75" & 0.625" 5.96 2.85 2.09 0.0363
0.75" & 0.50" 9.73 2.59 3.76 0.0002
0.75" & 0.25" 5.74 2.87 2.00 0.0453
0.625" & 0.50" 3.77 1.69 2.23 0.0257
0.625" & 0.25" 0.22 2.09 0.11 0.9151
0.50" & 0.25" 4.00 1.72 2.32 0.0204

 

 

 

 

 

 

 

97

Table 26. Pairwise Comparisons of Coefficients of Variation for Pouches

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Continuation
Package #3
Test Standard Average Coefficient Std Error
Mode Deviation (in. H20) of Variation (CVar)
URCU 6.85 41.55 16.49 2.68
URCD 5.52 43.32 12.74 2.05
1.0" 9.12 64.85 14.06 2.27
0.75" 9.48 76.72 12.36 1.98
0.625" 8.94 89.38 10.00 1.60
0.50" 7.65 102.54 7.46 1.19
0.25" 22.11 175.25 12.62 2.03
Package #3-Pairwise Comparisons
Pairwise Difference Std Error Z-ratio p-value
Comparisons in CVar (Diff)

URCU & URCD 3.74 3.37 1.11 0.2666
URCU & 1.0" 2.42 3.51 0.69 0.4897
URCU & 0.75" 4.13 3.33 1.24 0.2151

URCU & 0.625" 6.48 3.12 2.08 0.0375

URCU & 0.50" 9.03 2.93 3.08 0.0021
URCU & 0.25" 3.87 3.36 1.15 0.2490
URCD & 1.0" 1.32 3.05 0.43 0.6654
URCD & 0.75" 0.39 2.85 0.14 0.8923

URCD & 0.625" 2.74 2.60 1.06 0.2913
URCD & 0.50" 5.28 2.37 2.23 0.0256
URCD & 0.25" 0.13 2.88 0.04 0.9651

1.0" & 0.75" 1.71 3.01 0.57 0.5710
1.0" & 0.625" 4.06 2.77 1.46 0.1431
1.0" & 0.50" 6.60 2.56 2.58 0.0099
1.0" & 0.25" 1.45 3.04 0.48 0.6342
0.75" & 0.625" 2.35 2.55 0.92 0.3552
0.75" & 0.50" 4.90 2.31 2.12 0.0341
0.75" & 0.25" 0.26 2.84 0.09 0.9270
0.625" & 0.50" 2.54 1.99 1.28 0.2014
0.625" & 0.25" 2.61 2.58 1.01 0.3110
0.50" & 0.25" 5.16 2.35 2.20 0.0281

 

 

 

98

In order to compare coefficients of variation we used the standard errors of each
coefficient of variation [16] and the root mean square formula to determine the standard
error of the difference. The differences in coefficients of variation were calculated as
the higher coefficient of variation minus the lower coefficient of variation. The
statistical significance was determined using the standardized difference called the Z-
ratio and standard normal distribution.

For multiple comparisons of k=7 treatments p-values less than 0.0024 were
regarded as significant, the Bonferroni approach to multiple comparisons. For package
#1, two pairs were different. For packages #2 and #3, nine pairs and one pair were
statistically different, respectively. Results of previous tests (not reported here) have

yielded a different response in pattern for variation.

99

II. PACKAGE SIZE AND GAP SIZE EFFECT

The results shown in this section are the same restrained burst test results that were
shown in the previous section. These results are arranged in a way that it is easier to
see, just by looking at Table 27, the effects of changing the gap size and package size
on the restrained burst pressure. The statistical results show evidence of difference
between the burst values coming from different package and gap sizes. It will be seen

that both, package size and gap size, vary inversely proportional with burst pressure.

GENERAL RESULTS:

Table 27. Restrained Results for Pouches - per Package
. Packa e #1 (5” x 10” Pouch) - Seal Perimeter = 16.50”

 

 

 

 

 

 

 

Gap 11 Average Std Coeff. Min Max Range Burst per
Size (in.H20) Dev Var Perimeter
(in.) (%) (Burst/Inch)
1.0 20 84.18 12.94 15.37 66.90 113.20 46.30 5.10
0.75 20 102.22 13.63 13.33 73.20 128.70 55.50 6.20
0.625 20 111.05 11.90 10.72 87.60 128.10 40.50 6.73
0.50 20 124.37 10.17 8.18 105.70 143.70 38.00 7.54
0.25 20 196.42 15.72 8.00 166.90 224.20 57 .30 11.90

 

 

 

 

 

 

 

 

 

Packa e #2 (7” x 11” Pouch) - Seal Perimeter = 19.00”

 

 

Gap 11 Average. Std Coeff. Min Max Range Burst per

 

 

 

 

 

Size (in.HzO) Dev Var Perimeter
(in.) (%) (Burst/Inch)
1.0 20 73.24 6.81 9.30 62.80 86.70 23.90 3.85
0.75 20 91.40 13.82 15.12 65.90 114.90 49.00 4.81
0.625 20 107.97 9.89 9.16 84.70 125.70 41 .00 5.68
0.50 20 122.33 6.59 5.39 105.50 133.60 28.10 6.44
0.25 20 195.89 18.38 9.38 147.10 224.70 77.60 10.31

 

. Packa e #3 9” x 12” Pouclfl - Seal Perimeter = 21.50”

 

Gap 11 Average Std Coeff. Min Max Range Burst per

 

 

 

 

Size (in.H20) Dev Var Perimeter

(in.) (E (Burst/Inch)

1.0 20 64.85 9.12 14.06 48.10 81.00 32.90 3.02
0.75 20 76.72 9.48 12.36 58.90 90.70 31.80 3.57
0.625 20 89.38 8.94 10.00 69.40 104.90 35.50 4.16
0.50 20 102.54 7.65 7.46 85.30 116.30 31.00 4.77

 

 

0.25 20 175.25 22.11 12.62 144.70 221.00 76.30 8.15

 

 

 

 

 

 

 

 

 

 

100

Table 27, shows that the average burst values vary inversely with the gap size

and package size. Smaller gaps and smaller packages produce higher burst values.

STATISTICAL RESULTS:

Table 28. Overall Package Size and Gap size Effect on Burst Test Results for
Pouches - ANOVA Two way Analysis

 

 

 

 

 

 

 

Source Degrees Sum
of of of Mean F F p Conclusion
Variation Freedom Squares Square W cm value
PKG PKG Size
Size 2 25980 12990 82.74 3.03 0.000 Effect
Gap Gap Size
Size 4 476795 1 19199 759.23 2.40 0.000 Effect
No
Interaction
Interaction 8 1479 185 1. 18 1.97 0.31 1 Effect
Error 285 44880 1 57 --- --- ---
Total 299 5491 34 --- --- --- ---

 

 

 

 

 

 

 

 

 

This two-way analysis of variance shows strong statistical evidence of differences
in average burst value between packages and among gap sizes.

Figure 37, next page, shows the relationship between burst pressure and gap
size. It can be seen fi'om that figure that as the gap size increases the burst pressure
decreases. Also it is shown that within a certain gap size the burst pressure varies
inversely with the package size.

Figures 38 to 40, which are box plots, besides showing that burst pressure varies
inversely with gap size, show the variability within and between gaps for each package.
These plots show that the variability at gap 0.25” was higher than the variability

obtained from other gaps for the three packages. Also see Tables 23 to 26.

101

N;

 

p.23 pad: 1 :25

 

 

 

       
   

 

 

an 003 a Na 95 x a. 35 a
@325 35 950
2 we no I. do o
_ _ . . . o
2%.: u an
3.3 + .543. - "338 u a - 8
me 85.25
- 2:
3%... u am
8:... n ”m as: + 42.2.. - «amass u »
2.9a + 54.8.. - mag: u a Na 88.63 - o:
E 88.8»
- com
8N

mafia—em .8.“ SE 95 .m> 2:295 «9:5 nonmabmom ouauo>< .bm charm

(OZH u!) amssoad Jung pouruisag oBBJoAv

102

$225 cum new

coo; owed mmod oomd ommd
_ L 2 _ 2

_

 

 

 

 

 

 

 

 

 

 

_ _ _ [oi

 

 

 

 

 

 

 

 

 

 

 

[of

 

.- 1 com

 

 

 

 

 

 

E owe—own - monosom How 05m 30 .m> 838.5 35m uofimbmom .wm 833m

(OZH u!) amssexd 15mg paurensaa

coo;
_

Genoa: cam 95

owed
_

mmod ocmd
T _

ommd
_

 

_

 

 

 

 

 

L

103

 

 

 

 

 

 

*

4. 0

 

 

 

 

 

 

me owe—own - monozom Sm oumm new .m> 2335 65m nonmebmom .em 2:me

1.0m

1.05

Flea

rlofifi
1.om~
11om~
Ironfi
Itoefi
[lofim
1|QMN

 

(OZH u!) emssexd 18mg paureusau

104

96:25 3% new

coo; cmhd mmod oomd ommd
_ u _ _ 2

 

é.
...

_
I L

L
L flw 12:

h 102

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

18m

 

 

 

mu owee—omm - monosom Sm 65m 96 .m> 8:mean “mesm wofiebmom .ov 8:me

(OZH u!) emsseld 15mg peurensau

105

ANALYSIS OF FAILURE PATTERN FOR DIFFERENT PACKAGE SIZES:
As part of this experiment a peel test was performed for the three different
packages. The results show no difference in sea] strength between the sides and the
chevron. However, the pattern of burst failures showed more failures on the chevron
than on the sides when the pouches were tested in a restrained mode. A possible
explanation for that behavior is that the side seals of the pouch experience more
pressure than the chevron seal when it is tested in an unrestrained test mode. In a
restrained burst test, the restraining fixture makes the force to be more uniformly
distributed around the seal perimeter. This will make the chevron seal receive more
stresses than it would receive in an unrestrained burst test. This fact plus the fact that
the chevron has corners which will act as stress concentrators make the chevron more
prone to have a failure, when testing with a restraining fixture. Table 29, next page
provides information about the location of failures obtained when the three pouches

were tested in a restrained and unrestrained test mode.

106

Table 29. Location of Failures for Pouches

. Package #1 ( ” x 10”)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Location of Failures
Mode A B BC C BCD CD D E
URCU 5 4 0 0 0 1 6 4
URCD 1 5 0 0 2 2 9 l
Gap = 1.0” 1 5 2 1 2 3 5 1
Gap = 0.75” 1 4 3 0 5 1 5 1
Gap = 0.625” 0 8 6 0 3 0 2 1
Gap = 0.50” 0 6 6 0 5 1 2 0
Gap = 0.25” 1 14 1 0 0 0 4 0

. Package #2 (7” x 11”)

Location of Failures
Mode A B BC C BCD CD D E
URCU 6 6 2 0 0 0 6 0
URCD 5 6 1 0 2 2 4 0
Gap = 1.0” 0 8 1 0 0 1 10 0
Gap = 0.75” 0 11 0 0 1 3 5 0
Gap = 0.625” 1 9 0 0 1 0 9 0
Gap = 0.50” 0 12 0 0 0 0 8 0
Gap = 0.25” 0 10 0 0 0 0 9 1

Package #3 (9” x 12”)

Location of Failures
Mode A B BC C BCD CD D E
URCU 2 7 1 0 0 2 8 0
URCD 5 6 1 0 1 0 6 1
Gap = 1.0” 0 11 1 0 3 0 5 0
Gap = 0.75” 0 11 0 0 0 0 9 0
Gap = 0.625” 0 12 0 0 0 0 8 0
Gap = 0.50” 0 13 0 0 0 0 7 0
Gap = 0.25” 0 9 0 0 0 0 11 0

 

 

NOTE: URCU = Unrestrained Chevron Up
URCD = Unrestrained Chevron Down.

 

107

Table 30. Summary Location of Failures for Pouches

. Package #1 (5” x 10”)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Number of Failures Number of Failures
Mode on the Sides on the Chevron
(Out of 20 samples) (Out of 20 samplesL
URCU 9 11
URCD 2 18
Gap = 1.0” 2 18
Gap = 0.75” 2 18
Gap = 0.625” 1 19
Gap = 0.50” 0 20
Gap = 0.25” 1 19
. Package #2 (7” x 11”)
Number of Failures Number of Failures
Mode on the Sides on the Chevron
(Out of 20 samples) (Out of 20 samples)
URCU 6 14
URCD 5 15
Gap = 1.0” 0 20
Gap = 0.75” 0 20
Gap = 0.625” 1 19
Gap = 0.50” 0 20
Gap = 0.25” 1 19
Package #3 (9” x 12”)
Number of Failures Number of Failures
Mode on the Sides on the Chevron
(Out of 20 samples) (Out of 20 samples)
URCU 2 18
URCD 6 14
Gap = 1.0” 0 20
Gap = 0.75” 0 20
Gap = 0.625” 0 20
Gap = 0.50” 0 20
Gap = 0.25” 0 20

 

As it can be seen from Tables 29 and 30, the pattern of failures showed in general

more failures on the chevron than on the sides when tested using the restraining fixture.

 

 

 

108

IH. SUMMARY OF RESULTS FOR POUCHES:

1. Restrained burst test pressures are higher than unrestrained burst test pressures.

2. No statistical difference was found for burst value between chevron up and chevron
down in the unrestrained test.

3. The burst values vary inversely with the gap size. Smaller gaps produce higher
burst values.

4. In general burst values decrease with an increase in package size. This behavior is
true for both restrained and unrestrained burst test methods. The burst value at any
gap size is lower for bigger packages than for smaller packages.

5. Since there was no pattern in the overall difference in raw variances and coefficients
of variation between restrained and unrestrained burst tests, there is no evidence that
restraining fixtures reduce variation.

a. There is statistical difference in raw variances between restrained and
unrestrained burst test for packages #2. No statistical difference in raw
variances was found for packages #1 and #3.

b. There is no pattern in the difference in coefficients of variation between
restrained and unrestrained burst test for the three packages.

6. A small gap (0.25” in this experiment) contributes to an increase in variation.

7. Even though the results showed no difference in seal strength between the sides and
the chevron, the pattern of failures showed more failures on the chevron seal than on

the side seal when the pouches were tested in a restrained mode.

109

PART C - CORRELATION BETWEEN BURST TEST AND PEEL TEST -

POUCHES

The main purpose of this section is to study and analyze the theoretical formulas

developed by Professor Kit Yam [22] that correlate burst test with peel test. It was

found that for the package under study, which is a Tyvek/plastic chevron seal pouch, the

results obtained from burst and peel tests did not correlate.

GENERAL RESULTS:

Table 31. Burst Test Results for Pouches
Correlation between Burst and Peel Tests

Gap = 0.25 inches

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Flow Average Burst Peeling Average Burst Burst
Burst Peeling Time Standard Pressure Pressure
Time — tb Deviation (in. H20) Standard
(sec) Deviation
1 78.37 3.43 197.03 11.88
5 8.78 0.23 200.42 8.44
9 4.86 0.27 223.21 21.32
Gap = 0.50 inches
Flow Average Burst Peeling Average Burst Burst
Burst Peeling Time Standard Pressure Pressure
Time — tb Deviation (in. HzO) Standard
(sec) Deviation
1 68.53 2.75 132.22 10.94
5 8.23 0.19 147.45 8.67
9 3.83 0.25 148.38 7.35
Gap = 1.0 inches
Flow Average Burst Peeling Average Burst Burst
Burst Peeling Time Standard Pressure Pressure
Time — tb Deviation (in. HzO) Standard
(sec) Deviation
1 60.89 2.38 76.65 4.76
5 6.99 0.23 84.79 6.06
9 3.43 0.12 86.61 5.33

 

 

 

 

 

Note: Setting for other parameters - Sensitivity = 1 and Prefill = N
The burst peeling time (tb) was measured using a stopwatch

 

 

110

Table 32. Peel Test Results using “Average Seal Strength” Value (8.")
Correlation between Burst and Peel Tests

. Gauge Length = 0.40” (Corres ondi

 

n to a ap = 0.25”)

4—4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Average Standard Average Standard
Velocity (8.") Deviation tp Deviation
(in/min) (lb./in) (Sm) (sec) tp
(Ib./in) (sec)
0.58 —0.78 1.515 0.24 82.76 3.38
5.50—7.00 1.813 0.24 9.15 0.29
9.55 — 12.60 1.779 0.29 5.15 0.18
. Gau e Len h = 0.80” (Corres onding to a gap = 0.50”)
Average Standard Average Standard
Velocity (8.“) Deviation tp Deviation
(in/min) (lb./in) (Sm) (sec) tp
(Ib./in) (sec)
0.66 — 0.88 1.776 0.24 69.02 6.63
5.47 — 7.29 1.891 0.28 8.63 0.27
11.74—15.75 1.987 0.24 4.15 0.18
. (hm—M3111 = 1.60” (Corres ondin to a ap = 1.0”)
Average Standard Average Standard
Velocity (Sm) Deviation tp Deviation
(in/min) (lb./in) (Sm) (sec) tp
(lb./in) (sec)
0.74 —0.98 1.776 0.21 62.58 1.50
6.44 - 8.58 1.928 0.30 7.34 0.18
13.13-17.51 2.000 0.27 3.61 0.09

 

 

Note: The tensile peeling time (tp) shown in this table was measured using a stopwatch
and it represent the average time it took for all samples to peel.

A peel test was performed on eight (8) pouches in four (4) locations each, so
there are 32 samples in total. The “average seal strength” value (Savg) is the average of
the peel strength values obtained from all 32 specimens. This “average seal strength”
value was substituted as the S, in the (P = 28 / D) formula, to get the predicted burst
pressure. In the formula, S is the seal strength, D is the plate separation or gap, and P is

the predicted burst pressure.

111

Table 33. Peel Test Results using “Minimum Seal Strength" Value (Sn...)
Correlation between Burst and Peel Tests

, Gauge Length = 0.40” (Corres ondin to a ap = 0.25”)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Average Standard Average Standard
Velocity (Sm...) Deviation tp Deviation
(in/min) (lb./in) (Sm) (sec) tp
(lb./in) (sec)
0.58 — 0.78 1.282 0.16 82.50 3.18
5.50 — 7.00 1.601 0.19 9.17 0.28
9.55 —12.60 1.462 0.16 5.03 0.14
Gau e Len h = 0.80” (Corres ondin to a ap = 0.50”)
Average Standard Average Standard
Velocity (8......) Deviation tp Deviation
(in/min) (Ib./in) (Sm...) (sec) tp
(Ib./in) (sec)
0.66 —- 0.88 1.486 0.20 66.4 1.03
5.47 — 7.29 1.617 0.23 8.69 0.42
11.74—15.75 1.732 0.22 4.15 0.19
. Gauge Length = 1.60” (Corres ondin to a ap = 1.0”)
Average Standard Average Standard
Velocity (8......) Deviation tp Deviation
(in/min) (lb./in) (Sm...) (sec) tp
(Ib./in) (sec)
0.74 —0.98 1.566 0.17 62.15 1.76
6.44 — 8.58 1.553 0.29 7.26 0.21
13.13-17.51 1.709 0.23 3.64 0.09

 

Note: The tensile peeling time (tp) shown in this table was monitored using a stopwatch

and it represent the average time it took for the samples with minimum peel strength to

break.

A peel test was performed on eight (8) pouches in four (4) locations each, so there

are 32 samples in total. The “minimum seal strength” value (Sm-n) is the average of what

could be the “weakest point” or the lowest peel strength value from each pouch. Sm is

the average of the minimum value of the 8 pouches. This “minimum” value was

substituted as the S, in the (P = ZS / D) formula, to get the predicted burst pressure.

Table 34, next pages, shows the predicted Vs observed burst values obtained at each

gap, using the “average” and “minimum” seal strength values.

 

 

112

 

 

 

 

 

 

 

 

 

 

 

8.x. NM. a -\+ 8.8m 5mm -1 8.2». .o: - . Rd -\+ 2:. so: 2 3.3 E .o .1 8w
8.3 a; -\+ $.08 :.:. 4+ 64.4mm m. 8.0 -\+ w; 83 8 one mac .1 t a
8.3. 3.: -\+ 8.8: 33 -\+ «33 :1 . mama... mom.» 1 a: 9 mm: m; -\+ 8%
5.: ...: 6.: ...: 6.: 3.5.5 6.:
E; 25...: 352...: E: e > 3
uEEEougo 2:895 3.5m 9589.5 2.5m 2:5. wE—oox— .5623» 2:5. mi??—
UOZOMQO UOuvmtuhh Hahn—m 0—_m=0.H
am: 2.3» ......wfiem ..am EEEE: 1:35 new
9.2 mm. a -\+ 38“ 2.8 .\+ 8.8». 5 Rd -\+ 9:. 86.5 3 3.3 3.0 -\+ m3
£8: a; -\+ $8.6. Sam .1 2.8.. m mg -\+ ”3 so: 2 one 63 -\+ 2.0
5.2 mm. : -\+ 8.3: 8.2 4+ comma _ a; -\+ 2:: a: 8 ”3 won -\+ 66%
5.: ...: 6a ...: 68 225.: As:
3.; 253.: 352...: 36E e > &
ouaEtnohgo 9:53...— aflam 2.58.:— uflsm 2:5. wE—oom 5629.4» «EC. mi??—
coEoSO 38:32.— 355 235,—.

 

 

 

 

 

 

Hré 2.3, ......»55 ..am $9.22.. 1 ..mg 9.6

9580.5 3.5: “E: 5w:o.=m 3% 5950mm nigezfiom .3. 033,—.

 

113

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

owdm mm.» -\+ mme 3...” .\+ 2.43 o 310 -\+ 85 30.2 9 fig: Ed -\+ 2%
3.3 3.x -\+ mg: Emma .\+ 3.”: m 2.0 -\+ mmw EN» 8 3.3 95 -\+ mow
mmém v0.2 -\+ SN»: “3.: -\+ 5.3: _ whm -\+ 3.8 Aww. 8 89 8: -\+ 3.8
SE ...: 6N: ...: A8: ESE: so:
ax: 25%.: 3.3.... to: e > 3
85559.30 9:58.5 5:5 9:585 5:5 258 3:85 5823» «EC. wE—oum
cutomno B8895 5:5 258,—.
Malamv 25$ ......mfibm .35 52:22.1 Em... mm:
5.? 2.5 -\+ wfiwi vmén 4+ mad—N a 3.0 -\+ QM 30.2 8 EL: 26 -\+ 2+
3. 3 3.x -\+ 3.5: 3.3” 4+ cmdom m 2.0 -\+ mmw Ems 8 3.3 Rd -\+ $5
$.wv v0.2 -\+ NNam: emén -\+ $63 _ Sum -\+ 3.3 gm. 9 89 8.0 -\+ Node
5": ...: 5N: ...: A83 AEEE: A8:
3: 353...: 3.3:...— E: e > 5
88559.25 9:585 5:5 9:585 5:5 2:; ”£85 .3823» 2:5. 9:85
automao 88395 5.55 258,—.

 

 

 

 

 

 

:sz 2.3» ......mfibm ..sm

4

.

 

$925 I :55 .5

5:82.380 I 9:585 5::— was saw—8:5 ism 8938m— Efigzatm .3 03.5.

 

 

114

 

 

 

 

 

 

 

 

 

 

 

a; 2.3+ .93 3.2.1.3.. a Q. -\+ a; :3. 9 m2: 8. -\+ 3m
3.. 8.0 -\+ on: 8.3 -\+ 3.3 m 8. 1 3o 33 9 3%. .3 -\+ o:
8.: o: -\+ SE m: 4+ 8.8 . m: -\+ 3.8 so. 2 VS 3.. -\+ 2a..
6”: ...: 5:. ...: 0:. 2.5.5. .8:
3... 253.... 23.2.... E... e > ...
35:58.30 9:58.... 3:5 9.38.5 3.5m 9E; witch 362.5 2:: w:=oom
60202—9 touom—00Lh «mug-mm 0=m:0.F
Elem. 2.3» .....msza ..am 5.552.. 1...... mac ‘ \
3: 2.. -\+ $8 3.: -\+ 8.2. a N3 -\+ a; an: 8 2.2V 8.0 -\+ .3
3.2 8.0 .1 ES and. -\+ 3.8. m 8.0 \+ 93 an. 2 $9 w; -\+ as
8.9. o: -\+ 33 E... -1 «33 . mom .1 3.8 :3. 9 v: on. -\+ .20
6:. ...: 5N: ...: .83 .553. .83
3... 253.... 3.3.... 32m ... > ..
8a..._umo..o>0 9.38.5 8.5m 0.580...— .m..=m «SF 9E3; @620.» 0...; mam—ooh
U0>L0mao 3002—00...‘ Huh—um 0:2..0H~

 

 

 

 

 

 

 

mum. 2...; ......mmohm :3 was}? 1...... 9:.

Eta—.5050 I 9.38...— uflam ...... 595.5 ..3m 5053— 3.3.8.029: .3” 03:.

 

115

It can be seen from Table 34 that when Savg is used the predicted burst pressure
deviates more form the observed burst pressure than when Sam. is used. Sm; and Smgn,
when substituted in the formula P=2 S/D, overestimate the burst pressure. The last
column of Table 34 shows the percentage by which Pprediaed is higher than Pobserved,
when using Savg and Sm...”

Figures 41 and 42, on the following pages, show the relationship between burst
pressure and plate separation (gap) and it plots both, the predicted and the observed
burst pressures, using Savg and 8m values, respectively. Both figures show how the
formula P = ZS/D tends to overestimate the burst pressure.

Another analysis was performed with this data. A linear fit was tried, using the
natural logarithm of burst pressures and burst separation. The equation P = ZS/D, can
be expressed as In P = ln ZS — In D. A plot was made using the natural logarithm of the
gaps and the predicted and observed burst pressures. These plots affirm that the
equation P = ZS/D overestimates the burst pressure at every gap size. Figures 43 and 44

show these results.

116

N.

 

7 .5525 o 3.3.5:. x L

 

€2.25 neuauanom 8a..—

 

 

 

SS... u "y.
mwfiau + www.mam u "Manua— H an

 

     
  

:3... u NM.
038 + in... - ":28 u »

 

 

95m u 82.; ...—awnohm .3m 09293.: 5.3
53.23% 02.3 .m> 9.38.5 025m 0929?. .3. «...—”E

On

co.

cm.

ooN

omN

com

omm

cow

cmv

(OZH u!) amssard rsmg afiuaAv

117

NA

 

—‘ COZOMA—o C “won—Ommuounm X L

€2.25 nowauanom 82m

 

 

    
   

. a... S. to N... o
5.8... u N... -
3.2" + ”8.8m - ”:2... n h
#3... n m.
55% + ”2%.. - "3.2% u s -
x -

 

 

saw - 2.3, ....»55 ..am 5.53.)... 5.3
nouauaaom 32m .m> 9.529..— 355 09295. .5. «...—ME

cm

cc.

of

com

03

com

omm

ccv

(02H u!) amssud rung afiuaAv

118

 

1 62.330...— 0 6009:0950:— x L

 

$2.25 Feuflaaom 82.: ...

o..-

 

 

 

 

 

 

o 3. .3- 3.- .3- .- N. .- v..-
- .
, N
- m
83... u "m
. a. 5...... + $.85- u .
x m INT
1T I.
. o 32... u a”. 1m
35. + "2.2...- u »

 

 

"am - 255 ....»55 .55 swab}... 5.3
FeuEaaom 82m. :- .m> 7:38...— uflam guano}; ... .9. 0.53m

(o‘H u!)
[amssard rung aficJaAv] ul

119

 

 

o..-

 

 

 

 

 

 

1 Acetomncvs o Gouomuoéfi x L
€2.25 753......aom own—E :—

o Nd- v o- 0.? ad- 7 N..- v..-
- .

- m

- m

33.: n «m
m v 5...... + 3.8...- n g
- m
MM
- 0 3:... u «m m
88... + .58...- u .

 

 

...am - 2.3, .....wsem .3m 58.5.)... ...?

783.23% 32.: ... .m> .9330...“ «9:5 owns}.— E .3 0.53m

(OZH u!)
[amssud rung ofimaAv] u]

1.

120

SUMNIARY OF RESULTS

CORRELATION BETWEEN BURST TEST AND PEEL TEST - POUCHES:

The formula P = ZS/D tends to overestimate the burst pressure. The overestimation

of burst pressure increases at smaller gaps. See Table 34 and Figures 41 & 42.

The correlation between peel and burst test does not seem to work well, at least for

the pouches that were tested, which are Tyvek/plastic chevron seal pouches. There

are some reasons that can explain why this correlation does not work:

a. The burst test and the peel test are two different types of test. In the burst test,
the force applied results in deformation of the package. On the other hand, in
the tensile peel test, deformation is being applied and it results in some force.

In the peel test, deformation is applied to a one-inch wide strip of material or
seal. In the burst test when the package is pressurized the seal perimeter does
not take the same load at all points because it deforms differently around the seal
perimeter. For example, the areas that have wrinkles are loose and slack, so
they do not take any force. It is only the perimeter of the stressed part that takes
the force. Also, corners, curves and straight line seals take force differently.

b. In the burst test, not only the seal perimeter affects the burst pressure. There are
other factors like sharp edges, corners, and angles that can act as stress
concentrators and therefore affect the results.

0. The variation of the seal strength around the seal perimeter of the pouch makes
the correlation harder. The area from which a strip is cut to be peel tested may
not necessarily represent the point in which the pouch would break in a burst

test.

121

d. The time that it takes for a sample to peel or burst can affect the results. If the
strip or sea] is being stressed slowly it will elongate and then break at a lower
value than if it is stressed quickly due to the nature of the material [7].
According to Professor Kit Yam [22], if the tensile peeling time and the burst
peeling times are controlled to be the same, a correlation between burst test and
peel test is possible. The percent difference between burst and peeling times
obtained in this experiment was approximately 5%, not much different from the
percent difference obtained by Yam. The obtained results on this thesis project
show that the correlation between the two tests was not possible to achieve for
Tyvek/plastic chevron seal pouches, even when the tensile and burst peeling
time were controlled to be the same. Professor Yam used a different type of
package. He used PET/aluminum/PP MRE pouches. Yam’s research results
show higher tensile and peeling times and higher burst values than the ones
obtained in this experiment. The difference in type of package, seal and

material could explain the differences in the obtained results.

122

DISCUSSION

Wachala, Bohn, Spitzley, Franks, Lorimer, and ASTM Committee F2.6, all give
reasons for using restraining fixtures in the burst test (see Appendix D). When these
reasons are compared with the results of this experiment, some agree and some do not.
It is true that when using the restraining fixture, all seal surfaces are exposed to the
same forces. The forces are distributed more unifome around the seal than when the
package is unrestrained. This gives the entire seal area a more equal opportunity to
burst. It was also observed that when the package was pressurized within restraining
plates, the dimensional stability of the package was maintained and the tendencies to
deform were minimized. Also it is true that the restrained burst test results will provide
a better way to correlate burst and peel test than unrestrained burst test results. But even
when using restrained burst testing there are still some difficulties in trying to find that
correlation for Tyvek/plastic chevron seal pouches. The predicted burst pressure
obtained with the formula described by Kit Yam [22] overestimated the observed burst
values.

Some of the reasons provided do not agree with the obtained results. For example,
the use of restraining plates to get more consistent results. It was seen that in the case of
the pouches a gap of 0.25” contributed to an increase in variation. In general, the results
for blisters and pouches showed no specific pattern in the variation.

Another common reason why the use of restraining fixtures is being proposed is that
it may reduce or eliminate the influence of package geometry and package size effects
on the burst values. The results of this project show that burst pressure varies inversely

with package size for both, unrestrained and restrained burst test. For both, blisters and

123

pouches, the package size effect was not reduced or eliminated with the use of
restraining plates, but stayed almost the same or even increased. See Tables 6 & 8, for
blisters and Tables 18 & 20, for pouches.

It is believed by some of the researchers mentioned above, that there is a greater
chance to find the weakest area when using restraining fixtures. The design of this
experiment does not allow to prove or disprove this because the pouches that were
tested have the same seal strength around the seal perimeter (sides and chevron). It is
true that when the pouches were tested in a restrained mode, they showed more failures
on the chevron than on the sides. It is not clear why this happened. There are some
factors that could explain this behavior. They are the following: first, the chevron is
receiving more stress when the pouch is restrained than when it is unrestrained. This is
because when tested unrestrained the pressure entering the package make the flexible
membranes of the pouch to form a “pillow” shape which pulls more on the side seal
than on the chevron seal. When using the restraining plates, they prevent these flexible
membranes from forming that “pillow” shape and allow testing the seal more
uniformly. Second, the peak and corners on the chevron act as stress concentrators,
which would make it more likely to break in a restrained test than the sides. It was
observed that the pouch still deforms a little bit on the chevron area while tested
between the restraining plates; this can create stress concentrators along the seal
depending on how it deforms. Another test should be conducted in the future using
pouches with a chevron seal made weaker than the sides. In that way, it will be possible

to see if the restraining fixture helps to find the weakest area or not.

124

Another reason why the use of restraining fixtures is being suggested is that it
provides a way of measuring the minimum seal strength of the package. I think this
really depends on the way “minimum seal strength” is interpreted and what is the
intended purpose of the burst test being conducted. The obtained results showed that
for blisters and pouches the restrained burst pressures are higher than the ones obtained
in the unrestrained test. When the pouches are tested in an unrestrained mode they
break at a lower value. There was a lot more deformation of the package when it was
tested unrestrained than when it was tested restrained. This deformation can fold the
seal area and create wrinkles. This will make the package easier to break in some
places than in others. This burst value can be seen as the “minimum seal strength” that a
package will show. This will represent the worst case even though it does not
necessarily represent the real strength of the seal. When the package is tested in a
restrained mode, because there is less deformation and fewer stress concentrators, the
package breaks at higher values. In this case the deformations are smaller, therefore the
stress concentrators must be lower, so the package seal remains secure until the air
pressure is higher. The “minimum seal strength” in this case represents more the

strength of the seal.

CHAPTER 6. CONCLUSIONS & RECOMNIEDATION S

125

126

CONCLUSIONS

The use of restraining plates in the burst test is currently being proposed in
industry. Members of leader companies have provided reasons for using restraining
fixtures in the burst test. Some of these reasons are not in accordance with the results
obtained in this experiment. Other experiments that have been performed previous to
this one have shown similar results. Therefore, some of these reasons need to be

reconsidered.

The use of a restraining fixture in the burst test has its advantages and
disadvantages. Some of the advantages are:
1. The tendency for package deformation (pouches) and donring of the lid
(blisters) is minimized
2. The dimensional stability of the package (pouches) and lid (blisters) is
maintained

3. Package seals are tested more uniformly.

Some of the disadvantages are:
1. There is no conclusive evidence that restraining fixtures reduce variation.
The data analysis shows that the raw variance and coefficients of variation
give different answers with respect to variation.
2. It requires higher pressure to break the package and longer time to complete

the test.

127

3. Restraining fixtures of different sizes are needed to test packages of different
sizes. Pouches that are extremely large or wide will probably need a bigger
restraining fixture than the rest of the pouches.

4. Gap heights needs to be specified depending of the package size and type. A
gap that is too big for a certain pouch will not minimize the effects of folding
and creasing of the flexible membranes. For blisters, choosing a gap height is
more complicated because the height of the tray needs to be taken into
consideration. Also, it is hard to determine which gap size will provide the best
results for a certain package and to determine if testing more than one gap will
be useful or not.

5. A restrained burst test does not necessarily represent what the most of the
packages will encounter in reality. Ifthe package in reality will behave like in
an unrestrained situation, and the restrained burst test results are being used as
an indicator of what is the pressure that the package can withstand, then the
results will indicate that the package is stronger than what it is in reality. That is

because restrained burst pressure is higher than the unrestrained burst pressure.

The burst test provides a means for measuring the overall strength of the
package, and determines the weakest point of the entire seal [15]. The internal burst
pressure is considered to be a good overall measure of the ability of a pouch to
withstand transport and handling [22]. It subjects the entire package to some of the
stresses that packages encounter in the manufacturing, distribution and use

environments [5].

128

If the goal of the burst test being conducted is to provide a measure of the
package integrity in the use environments then the unrestrained test should be used
since most of the packages behave like that in reality. It does not matter if deformation
lowers the burst pressure if in reality that is what is happening to the package. All
packages, during their useful life, can experience a series of different situations that are
difficult to predict. It is hard to predict how the package will deform or behave in those
situations. The unrestrained burst test will provide information about the lowest burst
pressure that a package can withstand. I think it is important to reconsider the reasons
why restraining plates should be used; if it is going to be used instead of the

unrestrained burst test or if it is going to be used in addition to it.

129

RECOMMENDATIONS

After completing the experiments and the data analysis, there are some things that

were learned and should be considered before repeating a fixture test.

1. If a restraining fixture will be used;

a. Define and specify the purpose of using it because:

1. There is no conclusive evidence that restraining fixtures reduce variation.
2. It was demonstrated that the use of restraining plates does not reduce or
eliminate the package geometry and size effects.

b. Find a way of estimating the best gap height to use for a specific pouch. For
example, we know that for a future test we should avoid using narrow gaps, for
example 0.25” (W’) or lower, for packages of the size range used in this research
project. The variation for gap 0.25” was actually higher than for the other gaps.

0. Since the raw variance (02) and coefficients of variation (CV) give different
patterns of response, it is essential that reports of repeatability or variation and
discussions make clear whether raw variance or coefficients of variation is
being reported.

2. When testing pouches in an unrestrained mode, it could be done either with the
chevron facing up or the chevron facing down. It was demonstrated that there is no
statistical difference between the unrestrained burst test results performed with the

pouch at this two different chevron orientations.

130

Even though we have accomplished some of our goals in terms of explaining the

basic behavior of the packages when tested in a restrained mode there is still more

research that needs to be done. Some of the recommendations for future research are:

1.

Determine why the restrained burst test is used

a. To predict real life performance?

b. As a quality control test to detect changes in sea] integrity and quality during
production?

c. Other?

Study, analyze, reconsider, and formulate reasons why restraining plates should

be used.

Conduct two restrained and unrestrained burst tests for Tyvek/plastic pouches.

In the first test, pouches that have a chevron seal weaker than the sides will be

tested. In the second, the same test will be performed using a package with the

same size and material, but with the side seals weaker than the chevron. The

main purpose of these tests is to see if the restraining fixture really help to find

the weakest point. If the pouches break mostly on the chevron in the first test

and on the side seals in the second test, then that will be a good indication that

the restraining plates help to find the weakest point. On the other hand, if the

pouches break mostly on the chevron during both tests, then it means that there

is something else that is causing breakage in that specific area. It could be the

way the pouch is held in the fixture, the deformation experienced while testing,

and others.

131

Use pouches made of different types of material (for example aluminum

pouches) to conduct:

a. An unrestrained burst test to study the chevron orientation effect

b. A restrained test to study the gap size effect

c. A peel and restrained burst test to see if a correlation between the two will
work or not

Conduct a restrained and unrestrained burst test with different types of blister

configurations.

a. Additional gap sizes should be included in the experiment.

b. Differences in tray shape and geometry and differences in angles and radii at
the comers should also be considered

Investigate if a correlation between peel test and burst test will work for blisters.

APPENDICES

APPENDIX A

APPENDIX A
BLISTERS -RAW DATA

1. UNRESTRAIN ED VS RESTRAIN ED BURST TEST RESULTS

Table 35. Unrestrained and Restrained Results - Blisters Raw Data

. Packa e #1 and Pachg;#2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PKG #1 Accessory Package PKG #2 Thera IPG outer Package
P/N 119401-001 P/N 119679-001
Sample UR Location R Location UR Location R Location
1 117.5 C 154.2 A 68.2 A 116.6 AC
2 132.7 A 157.0 B 73.4 C 100.2 A
3 125.3 D 146.9 A 75.6 A 94.1 A
4 119.6 D 152.4 AB 80.7 C 118.9 A
5 123.0 D 150.3 B 75.1 A 102.1 A
6 118.0 D 163.7 CD 75.6 B 112.6 A
7 126.9 A 147.6 B 72.6 B 102.4 AB
8 126.3 A 149.8 D 73.5 B 114.9 mf
9 127.9 C 156.5 B 70.9 C 121.5 A
10 125.1 A 147.4 A 72.7 C 116.1 C
11 125.3 D 152.6 CD 75.4 B 119.1 B
12 130.1 C 162.2 B 71.8 A 111.5 A
13 109.7 A 154.2 B 72.0 A 115.7 B
14 117.7 A 154.4 D 72.8 A 111.4 A
15 123.0 C 146.9 CD 73.6 C 129.9 B
16 129.6 D 151.8 B 64.1 A 112.8 B
17 99.8 D 152.2 B 72.9 A 113.6 B
18 118.2 C 141.4 CD 74.9 C 104.7 A
19 116.3 C 137.3 CD 73.3 B 113.4 A
20 128.2 C 144.5 CD 76.9 C 115.3 AC
21 115.9 C 154.4 A 69.0 A 107.6 AB
22 95.0 C 152.1 CD 80.2 C 118.1 C
23 115.0 D 150.5 D 76.9 A 119.0 A
24 100.2 C 151.2 AB 68.0 A 115.5 A
25 118.2 D 149.8 A 71.7 A 121.0 C

 

 

 

 

 

 

 

 

 

Note: mf = material failure; nb = non break

132

 

133

Table 35. Unrestrained and Restrained Results - Blisters Raw Data
Continuation

Package #1 and Packagfl

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PKG #1 Accessory Package PKG #2 Thera IPG outer Package
P/N119401-001 P/N 119679-001
Sample UR Location R Location UR Location R Location
26 120.3 C 145.2 C 69.3 C 105.2 A
27 124.2 A 149.2 C 69.9 A 115.3 A
28 116.6 C 128.4 D 74.6 B 106.7 A
29 121.3 C 162.0 C 69.4 A 116.0 A
30 121.0 D 151.1 C 73.5 A 100.7 A
31 114.6 A 180.3 A 74.5 A 105.0 B
32 122.5 C 144.8 C 75.7 A 119.5 CD
33 108.8 D 166.6 D 69.6 A 127.6 AC
34 112.3 D 153.4 C 76.0 B 110.9 AB
35 117.0 A 135.8 Nb 72.5 A 105.8 A
36 124.9 A 162.9 D 71.9 A 120.5 A
37 136.7 B 155.6 C 80.1 A 114.2 C
38 120.7 C 153.6 D 72.5 B 125.3 A
39 126.9 D 164.8 D 67.5 mf 113.3 A
40 141.7 C 190.7 C 66.9 A 107.2 A
41 125.7 D 163.6 C 70.0 A 101.2 CD
42 141.7 C 165.5 AC 82.1 B 125.3 D
43 147.4 C 153.0 D 72.2 B 135.0 C
44 135.2 C 160.7 A 71.3 A 112.2 A
45 124.9 D 162.2 D 73.5 A 109.2 A
46 112.9 D 168.7 Nb 70.0 C 101.8 A
47 112.9 C 143.4 AB 73.7 A 122.3 C
48 108.5 C 159.8 B 77.1 B 111.9 A
49 128.8 A 138.2 AB 76.4 B 104.4 B
50 116.4 D 145.7 D 74.7 C 116.8 B
Average
kin. H30) 121.4 N/A 153.7 N/A 73.1 N/A 113.2 N/A
[Std Dev 10.2 N/A 10.7 N/A 3.6 N/A 8.4 N/A

 

 

 

 

 

 

 

 

 

Note: mf = material failure; nb = non break

 

134

Table 35. Unrestrained and Restrained Results — Blisters Raw Data
Continuation

Package #3 and Package #4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PKG #3 Standard Leads Outer PKG #4 Myocardial Leads Outer
PIN 119421-001 PIN 119553-001
Sample UR Location R Location UR Location R Location
1 40.3 C 123.2 D 38.1 mf 85.3 BC
2 41.2 C 103.4 A 42.0 B 89.5 BC
3 35.2 C 119.6 B 41.0 AB 89.5 BC
4 40.1 C 101.5 B 40.0 B 84.0 BC
5 41.4 A 104.4 B 36.6 mf 86.8 BC
6 33.5 C 120.9 C 38.7 B 84.9 BC
7 43.2 C 124.6 AD 34.9 mf 88.1 BC
8 42.1 C 115.4 B 36.8 AB 84.8 BC
9 43.3 C 109.5 B 39.9 B 83.9 BC
10 41.5 C 114.4 D 38.0 B 87.2 BC
11 39.8 C 122.2 B 40.0 B 85.8 BC
12 34.2 mf 122.7 D 37.3 AB 93.5 BC
13 39.5 C 119.7 B 36.7 B 83.3 BC
14 38.6 mf 112.9 A 40.0 B 82.0 BC
15 40.1 C 115.0 B 36.8 B 81.7 BC
16 38.9 C 119.2 C 40.1 AB 84.7 BC
17 41.2 C 131.9 D 35.9 mf 81.9 BC
18 37.0 mf 105.8 B 38.7 B 88.7 BC
19 42.9 C 118.3 D 36.9 A 86.2 BC
20 40.9 C 125.3 AD 38.9 AB 86.6 BC
21 40.5 C 108.2 B 38.3 A 88.6 BC
22 38.3 C 113.3 B 40.7 AB 92.2 BC
23 38.3 C 112.0 B 37.7 B 85.0 BC
24 44.3 C 101.9 B 38.1 mf 83.0 BC
25 39.2 mf 106.2 B 39.1 B 89.3 BC

 

 

 

 

 

 

 

 

 

 

 

Note: mf = material failure; nb = non break

135

Table 35. Unrestrained and Restrained Results — Blisters Raw Data
Continuation

Package #3 and Package #4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PKG #3 Standard Leads Outer PKG #4 Myocardial Leads Outer
P/N 119421-001 P/N 119553-001

Sample UR Location R Location UR Location R Location
26 36.8 C 106.5 A 38.0 B 89.0 BC
27 38.3 C 117.2 D 34.0 AB 82.9 BC
28 38.0 C 106.4 B 39.9 AB 92.6 AD
29 37.0 C 131.6 CD 42.4 A 85.8 BC
30 37.2 C 115.4 D 39.9 AB 89.3 BC
31 35.9 C 111.9 D 38.6 AB 85.0 BC
32 39.8 C 117.5 B 39.3 AB 83.9 BC
33 39.9 C 121.4 D 40.6 AB 81.1 BC
42.2 C 119.1 A 36.5 B 91.0 BC
35 36.7 C 115.4 D 39.4 AB 90.4 BC
38.2 C 98.6 B 38.8 AB 90.9 BC
37 36.5 C 116.3 B 33.9 AB 86.7 BC
38 43.3 C 111.2 AB 37.3 A 86.1 BC
39 38.2 C 109.5 AB 37.1 AB 90.9 BC
40 37.1 mf 111.0 A 38.2 AB 86.5 BC
41 38.9 C 114.4 A 39.6 AB 89.0 BC
42 41.6 C 120.9 D 35.4 AB 90.8 BC
43 31.2 mf 105.4 B 36.0 A 86.8 BC
44 37.8 C 111.4 B 39.3 AB 94.0 AD
45 36.9 C 114.7 CD 39.5 AB 84.9 BC
46 38.0 C 117.0 A 37.1 AB 76.5 BC
47 38.6 C 114.2 B 41.0 B 86.6 BC
48 39.1 C 118.6 D 38.1 AB 91.5 BC
49 37.0 mf 115.2 A 39.0 B 84.5 BC
50 42.2 C 111.4 B 40.6 A 88.3 BC

Average

Kin. H20) 39.0 N/A 114.5 N/A 38.4 N/A 86.8 N/A
| Std Dev 2.7 N/A 7.3 N/A 1.9 N/A 3.6 N/A

 

 

 

 

 

 

 

 

 

 

Note: mf = material failure; nb = non break

Table 36. Package and Gap Size Effects - Restrained Burst Test Results

11.

136

BLISTERS -RAW DATA

PACKAGE SIZE AND GAP SIZE EFFECT

Blisters Raw Data

Package #1 - Accessories Package -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gap = .20" Gap = .10" Gap = .01"
Sample Burst Value Location of Burst Value Location of Burst Value Location of
(in. H20) Failure (in. H20) Failure (in. H20) Failure
1 117.8 D 166.9 B 265.2 C
2 115.9 D 150.8 C-D 235.3 C
3 136.1 A 166.4 D 227.7 C
4 122.9 D 141.6 D 216.6 D
5 119.0 C 149.8 D 249.6 A - C
6 122.2 C 159.5 D 264.0
7 123.9 D 150.8 C - D 240.9 B - D
8 121.1 D 162.3 C-D 238.4 B-D
9 139.8 B 128.4 D 237.4 D
10 114.1 C 125.0 C - D 246.3 C
11 131.3 B 146.5 C-D 241.3 C
12 114.0 C 147.5 D 246.4 C
13 114.6 D 162.0 D 242.4 D
14 130.7 C 141.8 C 241.0 C
15 128.5 C 147.1 D 239.4 A - C
16 132.2 C 146.0 C 230.2 B - D
17 139.9 C 155.9 D 236.4 C
18 125.6 C 157.6 A 220.5 C
19 126.1 C 161.6 D 220.5 D
20 109.3 C 145.5 D 215.4 C
21 120.8 D 170.7 C 235.2 A - C
22 123.0 C 153.3 A - B 224.2 A - C
23 132.8 D 157.0 D 228.3 B - D
24 116.7 D 152.4 D 214.2 D
Average 124.10 N/A 151.93 N/A 235.70 N/A
Std Dev 8.4123 N/A 11.1619 N/A 13.5612 N/A

 

 

 

137

Table 36. Package and Gap Size Effects - Restrained Burst Test Results
Blisters Raw Data - Continuation

Package #2 - Thera Small Outer Package

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gap = .20" Gap = .10" Gap = .01"
Sample Burst Value Location of Burst Value Location of Burst Value Location of
(in. H20) Failure (in. H20) Failure (in. H20) Failure
1 116.9 B 135.3 B-D 151.3 A-C
2 98.8 A 139.7 B -D 157.9 A-C
3 100.6 B 130.9 B -D 164.7 B -D
4 96.8 A 129.5 A-C 156.6 B-D
5 112.5 A-C 129.4 A-C 155.0 A-C
6 99.6 A 134.1 C 159.1 A - C
7 116.0 A 132.2 A - C 160.2 D
8 100.4 B 137.4 C 162.6 C - D
9 112.4 D 128.4 A-C 161.1 A-C
10 108.7 A 137.1 B -D 164.3 A-C
11 108.6 A 121.8 A-C 150.6 A—C
12 99.8 B 136.0 C 159.7 B - D
13 101.8 B 133.7 B-D 150.6 B-D
14 90.3 A 134.4 A - C 156.4 B - D
15 107.8 A 129.4 A-C 151.3 A-C
16 105.4 B 130.4 B -D 153.7 B-D
17 103.1 B-D 132.5 D 159.3 B-D
18 114.1 D 127.0 A-C 151.9 A-C
19 100.6 A 120.2 B - D 159.9 A - C
20 85.5 B 127.8 B -D 163.7 B-D
21 113.2 A 133.5 A-C 163.3 A-C
22 103.5 B 128.7 B -D 160.9 A-C
23 95.0 A 121.1 A-C 154.8 A-C
24 106.2 B 124.5 A-C 157.7 A-C
Average 104.07 N/A 130.63 N/A 157.78 N/A
Std Dev 8.0029 N/A 5.1917 N/A 4.5503 N/A

 

 

APPENDIX B

APPENDIX B
POUCHES -RAW DATA
I. UN RESTRAINED VS RESTRAIN ED BURST TEST RESULTS

11. PACKAGE SIZE AND GAP SIZE EFFECTS

Table 37. Unrestrained Burst Test Results — Pouches Raw Data

Package #1 (5” X 10”)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Unrestrained Unrestrained
Chevron Up Chevron Down
Sample BP (in. H20) Location BP (in. 1120) Location
1 50.7 B 46.7 D
2 66.7 E 41.3 CD
3 43.0 CD 46.9 BCD
4 45.1 D 46.1 D
5 54.1 B 39.7 D
6 62.9 A 39.4 CD
7 56.7 D 41.4 D
8 77.3 D 60.4 D
9 50.9 E 38.3 D
10 52.9 B 49.7 BCD
11 64.4 D 69.8 B
12 76.0 A 57.1 E
13 59.4 B 68.9 B
14 55.5 E 64.0 A
15 69.3 A 41.3 B
16 71.5 D 41.1 D
17 54.0 A 38.8 B
18 57.6 D 40.5 D
19 69.9 A 58.8 B
20 55.9 E 44.5 D
Average 59.69 N/A 48.74 N/A
Std Dev 9.7289 N/A 10.4903 N/A
C Var (%) 16.2990 N/A 21.5252 N/A

 

 

 

 

 

 

138

139

Table 37. Unrestrained Burst Test Results — Pouches Raw Data
Continuation

Package #2 (7” X 11”)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Unrestrained Unrestrained
Chevron Up Chevron Down
Sample BP (in. H20) Location BP (in. H20) Location
1 40.2 B 31.5 CD
2 39.4 B 64.8 A
3 41.9 BC 30.2 B
4 32.7 BC 65.7 B
5 38.5 D 32.9 D
6 52.7 D 31.9 BCD
7 38.0 B 40.3 D
8 41.3 D 32.5 D
9 61.3 A 30.4 B
10 36.0 D 29.8 BCD
11 40.7 D 63.0 A
12 31.4 D 63.8 A
13 32.7 B 39.9 CD
14 60.4 A 54.0 B
15 40.6 B 63.4 A
16 59.5 A 50.4 BC
17 62.3 A 35.9 B
18 38.3 B 60.7 A
19 65.6 A 38.1 D
20 59.5 A 56.2 B
Average 45.65 N/A 45.77 N/A
Std Dev 11.5122 N/A 14.1515 N/A
C Var (%) 25.2183 N/A 30.9188 N/A

 

 

140

Table 37. Unrestrained Burst Test Results — Pouches Raw Data
Continuation

Package #3 (9” x 12”)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Unrestrained Unrestrained
Chevron Up Chevron Down
Sample BP (in. H20) Location BP (in. H20) Location
1 45.3 D 49.8 A
2 32.9 D 44.6 B
3 44.4 B 45.9 A
4 48.1 A 47.3 BC
5 36.2 D 33.8 D
6 41.9 B 46.1 BCD
7 47 .4 B 42.9 E
8 45.6 B 34.6 B
9 44.6 A 42.6 D
10 41.2 D 44.8 A
11 49.9 B 34.4 B
12 39.6 D 41.4 D
13 47.2 BC 50.8 B
14 46.0 CD 46.3 A
15 28.4 D 39.8 D
16 26.2 B 36.0 B
17 41.8 B 38.6 D
18 48.6 CD 50.1 D
19 32.8 D 47.0 B
20 42.8 D 49.6 A
Average 41.55 N/A 43.32 N/A
Std Dev 6.8456 N/A 5.5199 N/A
C Var (%) 16.4776 N/A 12.7416 N/A

 

 

 

141

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

<2 33.2 <2 e82: <2 mace: <2 3%.: <2 $8.2 3: Em
<2 «.32 <2 3...."— <2 35 <2 2.8. <2 2.2 6952
m one 8m 32 a .22 m 26 no 26 8
a 38 m 22 m 4N2 um 22 m on a
m we: mom 0. 2 2 m 22 o 3.6 8 an: 2
m 5.2 um 32 um 32 8m 32 o 6.8 S
m 0.8.6. 8m :3 m 32 m 32 < 4.8 S
m 38 m :2 m 3.2 a 22 um 0.8 m—
a is m 32 um 32 no n8 0 3w 3
m .6..va om on: m 4.3 m 33 mom New 2
m 32 m 32 a 6.6m om 6.02 m 92: S
o 38 um 3.2 m E: m 2: m wfi :
um 38 a 0.2: um 23 o 4.2: m 93 S
m 38 o f: m 92 a up m em a
a new m 2.2 8m 22 8m Em mom 6.8 w
m 49: m no: m 3: 8m ”.3 a 2: B
m we: um 32 um ea: om we: om N: e
m 32 no 3: 8m 32 mom n8 6 0.2 m
m 3: com 32 m 32 m . 22 a 2K 4
m 99: 8m 9»: on 0.5 o as o e; n
< we: um 32 um E: < 3: no 32 N
m S: om was mom ox: 8m 36 m w. E 2
SN: ...: 5N: ...: 5N: ...: 5N: ...: 5am ...:
5:236..— hm :32“ch mm :65:qu n:— =omuaucw~ mm :cmuauow— hm o-QEam
..mg n 95 ..cmd u 96 =de m 30 em; u 95 :3 u 30

 

 

 

 

 

 

 

ha: a :19 an o 5.2:— 1 5.5 >53 85.8% 1 3:53— amoh 3.5m away—“8% .mm 935.

 

142

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

<2 300.2 <2 £03 <2 9.00.0 <2 $8.2 <2 "20.0 an 3

<2 0002 <2 «.32 <2 3.08 <2 0:0 <2 3.? 00.26%
m 0:: m 32 a :2 a 0.5 a 0.0» 00
m 0.00m m SN. m 0.02 8 0.4.: no >9 2
o :2 o 0.02 a 0.2.02 m <9 m 0.: 0—
a 0.0:. m 32 a 0.3.2 o 0.: a 500 E
a 0:0 9 0.02 mom 2: m 0.8 a 0.00 0—
m 0.02 a 0.02 m 0.02 oo 0.02 a 3.0 m—
o 0.02 m S: m 0.02 8 0.02 a 0.0 E
m 5.8 m 22 o :2 m . 0.00 m ......» 2
o 4.0,: o 0.0: a is m 0.00 m 2.0 2
o 3.: m 0.03 a 3: m 2.0 m E :
m E: m 2: m 32 o 0.00 o 30 02
m 0.02 o S: m 2.0 m 2: m 0.8 0
m 0.000 a 0.02 m 4. _0_ m 0.3 Q 30 0
a 38 m 22 a 0.2.2 mom :0 m ~00 0
m 002 a 0.0.: o 2: a 0.00 m w; 0
o 0.08 a 32 < 0.00 m 20 a . 0.? m
m 0.3: m 0.02 m 22 a 20 um at 4
a 0.8m m 002 a 0.02 m 0.8 o :0 n
m 0.20. m 0.02 m 0.8 m 0.00 o A: N
m 2.8 m 32 m 0.0: m 002 m 0.0a _

5N: ..5 5E ...: 8N: ...: 5N: ...: 6N: ...:
=22“qu hm 5.3.361— mm :omwaucd Amm— =omuaucx~ mm Eomuauod n:— DEE-«m
:30 u 0.6 ..0m.0 u 96 28.0 1 new em; 1 00o ..0._ u 0.6

 

 

 

 

 

 

F: n :5 we owed—oun— - gum—.5300 I 5.5 Be“ 85.8.— 1 3.53— .8h 3.5m 355.23— .wn fiflh

 

143

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

<2 00:.00 <2 0000.0 <2 0000.0 <2 0000.0 <2 0020 >3 0.0

<2 00.0: <2 00.00. <2 00.00 <2 00.00 <2 00.00 00.222
a 0.0: m 0.00 m 0.00 m 0.00 m 000. 00
a 0.000 m 0.02 m 0.00 m 0.00 m 0.00 0_
a 0.00. a 0.000 m 0.00 a 0.00 8m 0.00 00
a 0.02 a 0.02 m 0.80 a 0.00 m 0.00 00
o 0.3: m 0.000 9 0.0.00 a 0.00 050 0. _0 00
a 0.002 m 0.02 m :0 a 0.00 a 0.00 0—
m 0.000 m 0.000 m 0.00 a 0.00 a 0.00 E
a 0.00 o 0.0: a 0.00 a :0 o 0.00 2
m 0.002 a 0.00 o 0.00 m 0.00 m 0.00 0—
m 0.00. a 0.00 m 0.00 m 0.00 m 0.00 :
m 0.002 m 0.0: m 0.00 a 0.00 m 0.0 00
a 30 m 0.02 o 0.00 m 0.00 o 0.00 0
m 0.0: m 0.00 m 0.00 m 0.00 m 0.00 0
m 0.000 a 0.000 m 000 m 0.00 O :0 0
m 0.000 a 0.00 m 0.00 m 0.00 mom 0.00 0
o 0.0.: m 0.00 o 0.00 a 0.00 m 000 0
m 0.02 m 002 Q 0.00 m 0.00 m 0.00 ..
a 0.0: m 0.02 a 0.00 m 0.00 om 0.00 0
9 0.0: m _.0: a 0.00 o 0.00 m 0.00 0
m 0.02 m 0.02 m 0.00 m 0.00 m 2.0 _

50: ...: 50m ...: 60: ...: 50m ...: 50: ...:
E539— mm .8539— mm :ezaoed mm .8339— mm 55.3.: mm 9.0—E05
..000 n 000 ..000 u 00o ..0000 u 00w ..000 n 0.5 :0; u 0.5

 

 

p.00 0 :3 mt «my—20m - 0850:5050 I 5.5 35— 02.2.: I 0:33— .00; 09:5 350.503— .wm 0340a.

 

 

144

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.00 00.00 00.00 00.00 0.00 ...e> D 00080.0
0000.0 0000.0 0000.0 0000.0 >000 000 .20
00.0 00.0 00.0 00.0 600.63 0000800

00.00 0000.0 00.0 _ 000.0 000.0 000.0 000.0 00 _

00.0 0000.0 00.0 000.0 000.0 000.0 000.0 0 _

00.00 0000.0 00.0 000.0 000.0 000.0 000.0 0 _

00.00 0000.0 00.0 000.0 000.0 000.0 000.0 0 _

00.00 0000.0 00.0 000.0 000.0 000.0 000.0 0 _

00.00 0000.0 00.0 000.0 000.0 000.0 000.0 0

00.00 0000.0 00.0 000.0 000.0 000.0 000.0 0

00.0 0000.0 00.0 000.0 000.0 000.0 000.0 0

00.0 0000.0 00.0 000.0 000.0 000.0 000.0 0

00.00 0000.0 00.0 000.0 000.0 000.0 000.0 0 _

0.0.0 00> o 600 000 600.630 0.0 00 00 < 6.0.000 _
0.90:5 .0000 00.0000“ Saracen 002:5

0:00 0 :00 00 00503.0 .50 00.3.30 £00200 .80 .00 0.000.

145

3&— 003.:
3.3 mean...
mméu mammé
3.: 02.0”...
3.3 25m...
”0.0 $3.:
90.00 0820
when mmmvé
3.: 0.03.:
0.0.3 3mm...
0.0.0 50> 0 >000 .00
0.2.20 .02. 00.0003—

aufi
NYN
m:.N
wcfi
3.0
m—d
mu."
afin
mud
3.N
090.0950

00x00 «00> 0 0030025

 

 

 

 

 

 

 

 

 

 

 

 

00.00 00.00 00.00 00.0
0000.0 0000.0 0000.0 0000.0 .600 0.00
00.0 00.0 00.0 100.0 00.20:
000.0 000.0 000.0 000.0 00
000.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
00 00 00 < 6.0.000

 

 

 

 

 

 

000300200— 0.2500

0000000000000 - 0:00 0 ..00 00 00303.0 00.0 0000800 000.0200 08.0 .00 6.00.0.

.02—
00.00030

146

«We— 33”.:
omd 20:5
3.3 cnmud
«flan vmmvé
3.3 ammné
3.: 0.50m...
bran mmwmé
3.3 ewwmd
0km: 5mg.
20.3 mam...
3.0 ...; u :00 0.00
0.2500 .020 .00—0003—

2:
3.0
Na;
”fin
me.—
«fin
fl:
am;
«A:
cad
mug—030

”Q:
wand
NN.N

mudm
waN...
K.—

 

 

 

00.00
0000.0 ,
00.0

in:
3.3.:
end

3.0 50> 0 50080.0

>05 3m
090.0950

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

000.0 000.0 00.0.0 000.0 00
000.0 000.0 000.0 000.0 0
T 000.0 000.0 000.0 000.0 0
000.0 000.0 00.0.0 000.0 0.
fl 000.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
000.0 000.0 0.00.0 000.0 0.
— 00.0.0 000.0 000.0 000.0 0
000.0 000.0 000.0 000.0 0
_ 000.0 000.0 00.0.0 00.0.0 0
_ 00 00 00 < 20:50

 

 

00080305 00280—

0800300005 - 0:00 .0 :00 00 «05.8.0 .00 00.030 500.200 .800 .00 20.0.0.

.02—
00.0009:

APPENDIX C

APPENDIX C

CORRELATION BETWEEN PEEL TEST AND BURST TEST — POUCHES
RAW DATA

Table 40. RESTRAINED BURST TEST — Results for Correlation

Gap = 0.25” ; Flow = 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Burst Peeling Time Burst Pmsure Location
(sec) (in. 1120)
1 74.1 191.2 D
2 83.1 194.0 D
3 79.0 211.4 B
4 85.7 213.2 E
5 80.7 190.3 E
6 80.7 187.0 A
7 77.8 204.5 D
8 75.3 190.5 D
9 78.8 183.8 A
10 76.3 210.2 B
11 80.7 221.5 A
12 77.9 186.1 D
13 77.5 186.3 D
14 73.7 193.8 B
15 74.3 191.6 A
Avg. 78.37 197.03 N/A
Std Dev 3.4288 11.8841 N/A

 

 

 

 

 

Other parameters: Sensitivity = 1; Prefill = N

147

148

Table 40. RESTRAINED BURST TEST - Results for Correlation

Continuation

Gap = 0.25”; Flow = 5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Burst Peeling Time Burst Pressure Location
(sec) (in. H20)
1 9.2 217.5 B
2 8.9 198.7 D
3 9.0 202.9 B
4 8.6 197.3 B
5 8.4 191.5 D
6 8.8 207.0 D
7 8.7 209.3 B
8 8.7 189.6 B
9 8.9 201.0 D
10 8.9 209.7 B
11 8.7 196.6 D
12 8.6 190.2 D
13 9.1 202.8 B
14 8.8 203.8 B
15 8.4 188.4 D
Avg. 8.78 200.42 N/A
Std Dev 0.2305 8.4364 N/A

 

Other parameters: Sensitivity = 1; Prefill = N

 

149

Table 40. RESTRAINED BURST TEST — Results for Correlation

Continuation

Gap = 0.25”; Flow = 9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Burst Peeling Time Burst Pressure Location
(sec) (in. H20)
1 4.5 197.8 B
2 5.1 213.2 D
3 5.3 238.8 B
4 4.5 196.3 D
5 4.8 216.0 D
6 4.8 228.3 D
7 4.9 222.5 E
8 5.0 236.5 D
9 5.1 244.8 D
10 5.1 261.4 D
11 4.8 213.5 B
12 4.8 224.8 D
13 5.1 249.9 B
14 4.8 221.9 B
15 4.3 182.5 D
Avg. 4.86 223.21 N/A
Std Dev 0.2720 21.3154 N/A

 

Other parameters: Sensitivity = 1; Prefill = N

 

150

Table 41. RESTRAINED BURST TEST - Results for Correlation

Gap = 0.50”; Flow = 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Burst Peeling Time Burst Pressure Location
(sec) (in. H20)
1 63.0 129.9 B
2 72.0 136.9 D
3 68.0 117.9 D
4 71.0 131.1 D
5 70.0 134.6 D
6 69.0 123.9 D
7 72.0 155.3 D
8 72.0 144.9 B
9 67.0 125.4 D
10 70.0 146.3 A
ll 67 .0 126.1 D
12 68.0 133.0 D
13 64.0 113.8 B
14 67.0 127.0 E
15 68.0 137.2 B
Avg. 68.53 132.22 N/A
Std Dev 2.7482 10.9418 N/A

 

Other parameters: Sensitivity = 1; Prefill = N

 

151

Table 41. RESTRAINED BURST TEST — Results for Correlation

Continuation

Gap = 0.50”; Flow = 5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Burst Peeling Time Burst Pressure Location
(sec) (in. H20)
1 8.3 150.1 B
2 8.4 148.4 B
3 8.2 153.0 D
4 8.0 139.9 D
5 8.3 151.2 D
6 8.0 138.4 B
7 8.3 141.8 B
8 8.5 156.5 D
9 8.4 143.6 B
10 8.4 158.0 B
11 8.3 161.8 D
12 7.8 127.9 D
13 8.1 142.4 D
14 8.2 149.7 B
15 8.2 149.0 D
Avg. 8.23 147.45 N/A
Std Dev 0.1870 8.6715 N/A

 

Other parameters: Sensitivity = 1; Prefill = N

 

152

Table 41. RESTRAINED BURST TEST — Results for Correlation

Continuation

Gap = 0.50”; Flow = 9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Burst Peeling Time Burst Pressure Location
(sec) (in. 1120)
1 3.8 146.8 D
2 3.6 137.4 D
3 3.9 139.7 D
4 3.8 145.0 D
5 3.6 142.3 D
6 4.2 151.6 BC
7 4.1 154.9 CD
8 3.9 147.6 A
9 3.6 141.6 D
10 4.3 158.4 B
11 3.9 157.1 D
12 3.4 157.7 D
13 3.7 139.4 B
14 4.0 156.4 B
15 3.7 149.8 D
Avg. 3.83 148.38 N/A
Std Dev 0.2469 7.3510 N/A

 

Other parameters: Sensitivity = 1; Prefill = N

 

153

Table 42. RESTRAINED BURST TEST — Results for Correlation

Gap = 1.0”; Flow = 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Burst Peeling Time Burst Pressure Location
(sec) (in. H20)
1 56.7 70.2 B
2 64.2 82.7 B
3 63.1 77.2 B
4 63.2 83.1 D
5 60.4 72.7 B
6 62.3 74.8 B
7 60.3 74.8 B
8 63.3 79.9 B
9 62.0 80.9 D
10 59.2 76.9 B
11 62.2 77.8 B
12 60.9 79.1 D
13 61.0 81.5 E
14 57.3 69.7 D
15 57.2 68.5 A
Avg. 60.89 76.65 N/A
Std Dev 2.3820 4.7571 N/A

 

Other parameters: Sensitivity = l; Prefill = N

 

154

Table 42. RESTRAINED BURST TEST - Results for Correlation

Continuation

Gap = 1.0”: Flow = 5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Burst Peeling Time Burst Pressure Location
(sec) (in. H20)
1 6.8 79.0 A
2 7.3 90.8 A
3 7.2 92.5 B
4 7.2 95.7 B
5 7.0 85.3 B
6 7.0 80.5 D
7 7.1 84.1 D
8 6.9 83.9 D
9 6.9 86.6 CD
10 6.9 76.8 A
11 6.8 88.1 CD
12 7.3 90.9 D
13 6.5 75.2 D
14 7.2 83.9 D
15 6.8 78.6 E
Avg. 6.99 84.79 N/A
Std Dev 0.2251 6.0624 N/A

 

Other parameters: Sensitivity = 1; Prefill = N

 

 

155

Table 42. RESTRAINED BURST TEST — Results for Correlation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Continuation
Gap = 1.0”; Flow == 9
Sample Burst Peeling Time Burst Pressure Location
(sec) (in. H20)

1 3.5 95.3 D

2 3.3 83.5 A

3 3.3 84.1 D

4 3.3 84.4 D

5 3.5 83.1 B

6 3.4 78.6 E

7 3.7 93.5 E

8 3.5 90.6 B

9 3.3 87.0 D

10 3.5 86.6 D
11 3.3 88.0 CD

12 3.4 79.4 E

13 3.4 83.4 A

14 3.5 96.4 B

15 3.5 85.3 D
Avg. 3.43 86.61 N/A
Std Dev 0.1163 5.3330 N/A

 

 

 

 

 

Other parameters: Sensitivity = 1; Prefill = N

 

.00—03080008 .m 00:00 .0 a. m .< 00.0 E055 who can £50 .56 M 030622, ...ovd H Ewan: owsmw 6000: 0800800000 08002

 

 

 

 

156

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>00—
evemd ahead coma; was. Numb... vmmné m—mnfi vweud Em
end» «5.0 8.; En.— mmAw wnmé 3.30 Sm; .w><
00.9w ~30 90.20 0000.0 8.00 maué oedm 03.0 m
omww 30.0 Saab 3N.— omen mmm0 063 03.0 r
00.3 3000 3.; N3”.— owww 050.0 00.3 mm»: m
00.00 0&0 oon @050 000.; 800.0 cmfiw 003.0 m
ovsw Sm.— cfiuw 3N; 000.00. 090.0 100.5 9&0 v
3.0% 00.0.0 cmdw 80.0 Oman Nmmg 3.5 80.0 m
3.; So.— ofiwn $0.0 owww mow0 2.0.30 0.3.0 N
8.3 m3; 00; cm: 00.3 000.0 owdw 03.0 0
0800 2:00 EB: 0800 2:: 0.5.0.0 0800 2:00 0.5.0.0 0800 2:00 00.00.00
3:03 0:050. m witch 00:20:. m 3:300 9:050. m 9580— 0:003... m 00.0—Sam
H a m <
0.20.084

 

 

— H >55 E000 amNd .I. 0:00 - 03:00—00..th .08 00.0—00¢ Emma. GHQ.— .mv 9.0—ah.

 

157

.00—02080008 .m 00:00 .9 0% m .< 00.0 E050: omw new .000 .36 M 020622, ...ovd u fimfi: owsaw 00000: 0022500000 08670

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ba
Sand mam—é cmwud 02.0.: cmwnd 0305 Send man—d Em
3.» m3.— 0Nd v8.0 36 2.0.0 cud gm.— .w><
me 800 0.0 30.0 —.a v3.0 020 08.0.. a
00 mow; md 20..— 0.0.0 SON we coma 0.
0.0 @060 :6 man; we owed ow ~20 9
mm 03.0 0.00 wen; v.0 0.2.0 06 05.0 m
mm 000... 00.0.. 3%.— m.@ 09.0 No 80.0 v
ed 000.. bd 03.0 0.0 0:50 Wm 03.0 n
me com; v.0 3m; 0.0 30.0 v.0 03.0 N
mm 000.0 0.00 20..— Zw M030 we mmoN 0
0800 2:; ES: 680 20:. EB: 6000 2:: EB: 0800 2:; 0.5.0.0
9580— 0:00.00. m w==oum 00:05.0. m 9580— 00:05.0. m M058.— o=0coh m 00.0—Sam
m a m <
0020.305

 

 

0 n 32.0 0.0.0 ..000 u 0.6 - 000000000000 1 0.003230 .00 00.230 00000 000000 .00. 2000

 

158

.00—03080090 .m 0000 .Q d. m .< 00.0 E050: 00.0 0000.0 .00.~0 .00 . 00 n 0206230 .2000 H 0&02 omzmw 0000: 0000080000000 0802

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>20
000~.0 03.0.0 003.0 000~.0 0000.0 0.0.3.0 0000.0 --.0 000m
0m.m 0.00.0 00.0 000.0 00.0 30.0 00.0 08.0 .m><
0.0 000.0 0.0 000.0 - 1- ~.m 000.0 0.0 000.0 a
0.0 00.0.0 0.0 000~ 0.0» 000.0 0.0 30.0 0.
~.0 0v0~ 0.0 000.0 0.0 000.0 0.0 000.0 0
0.0 0000.0 0.0 0010 0.0 0v0~ 0.0 000~ m
~.0 000.0 0.0 000.0 0.0 30.0 0.0 20.0 v
0.0 ~00~ 0.0 mam.0 0.00 000.0 0.0 vM- m
0.0 :00 0.0 000.0 0.00 50.0 0.0 ~m0~ ~
00 ~05. 0.0 000.0 0.0 000.0 0.0 000.0 0
0800 20F 05.00 080 2.00 0.0.0.0 0800 2000 0.0.00 080 2:00 05.00
90:90.0 0:00.90. m 90:09.0 00:05.0. m 90:80 9:00:00. m w:=oo.0 0:00.00. m 20.0an
H 00 00 <
0.20390

 

 

0 u 300.0 0:: ..000 n .000 - 000000000000 I 00032.00 000 00.0030 0.0000 0.0000 .00. 0.000

 

.00—03080002 .m 0000 d 0% m .< 5.0 £0000: 00.0 00000 .000 .050 u 020_oo_o> .2000 u fiwfi: owsaw 000000 008000000000 08070

 

 

 

 

159

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>900
09.0.0 nv-.0 33.00 v~0~.0 030.0 0~w~.0 030.0 0900 05
0.0.00 00.0.0 00.00 000.0 00.00 800 3.00. 000.0 .w><
00.00 000.0 000.0 00000 0v.~0 ~0\.0 0~.00 M000 0
000.00 0000 000000 ~000 00.00 000.0 00.00 0~0.~ 0.
00.00 0~0.~ 00.00 0.00.0 00.00 000.~ 0~.00 0.000 0
00.00 0000 00.00 0000 00.00 000.0 00.00 0000 m
00.00 0000 00.00 000.0 00.00 00000 000.00 ~000 .0
00.00 0000 00.00 ~b~0 00.00 050 00.0% .3000 M
000.00 ~000 00.00 05.0 00.00 0000 00.~w 0~w0 ~
00.00 000~ 000.00 0000 00.00 000.0 00.00 000.0 0

0800 2.0.0 000.00 080 20:. 000.00 080 20F 0000.00 0800 2:: 0000.00
3:800 0:00.00. m M05800 0:00.90. m M05300 0:050. w 3:800 9.00.5.0. m 0.0—000$
0.0 00 m <
000000600 - ..

 

 

0 u 32.00 00:00 ..0m.0 H 0.006 - 0.000.000.0000 .5.— 00000030 Emma. 10000000 .3. 03.0.0.

 

160

.boZSQmmE .m 98 ”Q a. m .< 5m SEEM Dim can “as £00 M 82023, .236 n 5mg. omsmmv ”wow: EBQEEE ”8oz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fin
35nd wmand cmmfio 33.: amend amaud wflfic mmvud Em
5.x «2: 3.x Sm; med :5; an.» $5.— .w><
5w :0; Yw ovoN 9w mm»..— od 8v; w
Md mnw.— 9w «SN m.w «S: M; wnod n
0.x Sam m.w 30; 0.0 we..— nw Co; e
m.m womN 9w 3.0; Tm 2%.— 5w 39m m
Wm coo; Wm mu!— od :3; M; omfim v
«aw vmc; 0w Em; Om SNN 5w EON m
Wm coma Wm So; ow owed Nd 3a.“ N
ow oofim Mm mm»..— mw mom; Ow mvwg —
A83 2:; EB: Gav 25h EB: A8: 2:; E3: 33 25. EB:
@525 2.2:; m w:=oom 05.5% m w==oom «=25? w 9509— 0:2.th m “...—Sam
M.— Q m <
.5389—

 

 

m u 5E ...... :35 u 96 - 5.3.525 I 532:5 2: £33— 3”; as: .3. case

 

161

23838. d Ea d a. m .< E 582 3.: c5 .32 .32 u 3:603 ...8.o n 5mg; 3:50 ”BS 3255 382

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>9:
2.5.: $3.: $36 cam—é «me—.9 33.: 33.: 336 Em
m3. mm?" 36 flea; 86 as." and 3a; .w><

2“ women Nd 8w.— Né mmod «av who; a

N6 mom; 3 3a; 3“ god me So; 5

0+ 03 .N v.v 3%— ad wnod new smog e

0% TENN ma mam; QM c8.— oé NSN m

ad of .N m6 vmoN Qm coma .. IV ava v

3» wood 0+ mood o.v v2 .N «av cam; n

0v can; QM 3nd Maw vow; me mm»; N

0v :0; QM mam; 3‘ wood 3“ Nmmd —

39$ «ES. A515 33; 25h. :55: A33 258 :55: A83 25h $5.3:
3:3.— 0.12—oh m maze?— o=m=ok m 9509— 0:238 m 9599.— 0:239. m 0355
H a m <
:33qu

 

 

a u 32:.— 15 acmd u 92.9 - :33.—53:0 I .3359th .5.— nzsnom Emma. Amman .vv 03.;

 

5.3838“ .m 95 .Q a. m n< 5m 58>: vud can “mod dwd M 82023, .200.— “ Ema»: omsmO ”com: 380883 "892

 

 

 

 

162

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>2—
334 anNd mama; mN_N.c can»; 33.: ES; mwNNd Em
che 3mg mva Nb.— m—dw Nab." voNe Rm; .w><
003 m8; 2N0 m3; 05% En.“ ovNo to; m
omdo an; 0&3 gm; ovNo am: 09% SNN b
ovMo $0; 003 NS; omNc 5mm; 03% mmoN 9
Save mow; ovho 02..— oowo owe.“ 008 vaA m
8.5 NS: code omega. 3N0 3K.— oZo 23.— v
05va :0; 00.3 mi..— oodo new; 095 03.. n
05$ 3m: 03% mom; 02% man; 09% emu; N
odmc Em; 004% 3a; 02% wow; ooNo v3; _
A8: 2:: EB: Ci 2.5 EB: A83 2:3 E3: 33 05F EB:
958m 2.2:; m witch 0:25P m wE—oom o=m=oh m wE—oom o=m=oh m 29.—am
H a m <
555.5

 

u H >5:— E—a ac; H haw - :33.“:th .5.— m==ue¢ Emu—H. Amman .mv 035—.

 

163

503638. nm :5 .Q a. m .< ho: ESE 3&0 was .wmm .00: u 8:628, .200; H fimfix 0930 ”wow: 280:8qu 3ch

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>2—
Nve—d 3N1: chNd mnhNd 53.: vaN.: emnfic“ $3.: Em
5.: a3.— cnfi 35.— wNK N3; $2. 3%.— .w><
vs EVN “N mNNN N6 gm; 2. v3; w
ms SON ms wooN ms $0N ms m3..— 5
ms :84 2. Sn; ms wnoN 2. mooN w
2. ova; ms BNN fin 3m; 5.: omNN m
:6 ME; fin 3%.— _.n 03; ms oooN v
we SON 2. wQN ms mac; ms NHNN m
N.\. QEN :8 :3..— N.\. MEN we now; N
MN :3: NB com; 2. Cum: 0.: nmm; _
so: 2:; EB: 33: 2:? EB: ca: 2:: EB: 68: 2.5 EB:
w:=oom 9:23,: m 9:3: 9:25.: m w==oom 0:25P m 3:25 0:25.: m 2.3—am
a a m <
:23qu

 

 

 

m u 32,,— ...: :3 n 96 - 8:25:50 I 53230 3.. £33. 53 dim .3 2.3.

 

igneoame nm EB d a. m “< 5: EEE 2 .2 can 402 .Nm.m_ n moEooB> .200; H Emce— owsz “tom: muouoESmm "8oz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>93
38.: NmnNd 33.: 33.: 33.: bENd £33.: 33.: Em
9mm 3:.“ med 3:.— med N3.“ mmd :ch .w><
Om :VMN 5m 3N; 9m EON Wm oSN m
5m :NQ— 5m NENN m.m E—N m6. SON b
on MS; m.m vmoN m.m :8.“ 9m 0mm; w
m.m :3.— 5w So; Om meN Wm mmm.— m
9m ovoN 9m 2:; 5m Sm; 5m oov‘N v
m.m :3.— o.m 33; EM MEN Wm 02..— m
0m SQN 9m mNmN Om vaoN 5m mmNN N
VM ome 9m SAN Om 3w; m.m ommN _
E: 95F EB: E: 2:; EB: 63: 2:; EB: 33: 2:: EB:
wag—ooh o=m=oh m 9525 0:25P m witch 265,—. m wag—com o=m=oh m 2956
H a m <
:23qu

 

 

a n 32,.— EE ac.— " aaU . :¢_ua==_u=c0 I 5559th .3: 3.59m Emmi. dam—m .mv visa.

 

APPENDIX D

 

APPENDIX D
REASONS WHY RESTRAINING FIXTURES SHOULD BE USED
As mentioned before in this report, members of leader companies like Carleton

Technologies, Medtronic, Rexam Medical Packaging, and TM Electronics have

provided reasons for using restraining fixtures in the burst test. Also, Committee F 2.6

on Medical Packaging of the American Society for Testing and Materials (ASTM),

pointed out in its proposed standard method [4], some of the reasons and situations in
which restraining fixtures should be used in a burst test. The following paragraphs
show these reasons.

Thomas Wachala - CARLETON TECHNOLOGIES — [19]

1. “In an unrestrained test there is no guarantee that under pressure, each package will
fold and deform the same way. The test can not be considered statistically fair if
one or more tests are biased because of the lack of consistency in fixturing and
holding the package under test. This lack of consistency may adjust the mean burst
value to a point where a degree of uncertainty may develop concerning the
acceptability of a perfectly acceptable lot from the production floor.”

2. “If testing (referring to the unrestrained test) is being done at multiple locations, or
if the end user or customer verifies product supplied by the manufacturer, the
potential exists for large differences in burst values.”

3. “The unrestrained testing does not necessarily identify the weakest part of the seal”.

4. “The use of properly sized restraining plates will minimize the deformation of the

package during a burst test”

165

166

5. “By maximizing the effect of restraining plates, the working surfaces around the
perimeter of the package can be equalized, but not entirely. The surfaces at and
near comers and curves will not respond equally, and will remain as a localized area
of incongruity”.

6. “Restraining plates give the operator the opportunity to test the package seals more
uniformly with a greater chance of finding weak areas.”

7. “The effects of package geometry and tendencies to deform during a test are
minimized.”

8. “The plate separation gap can be standardized, for use with a specific package, at
multiple locations.”

9. “This increases the likelihood that the test procedure and package geometry will
remain consistent, thus reducing the variables that prevent repeatable test
perfinnnance”.

10. “Restraining plates are yet another tool for quantifying and qualifying the
performance of package seals. If used safely and properly, the information they
provide can help control the manufacturing process and give the manufacturer the
ability to design and control the strength of specific seal areas.”

David Bohn/John Sgitzley — MEDTRONIC-|5|

1. “Additional research may yield to improved methods that will remove more
variables from the burst tests and increase its sensitivity to variation within the

process.”

2.

167

“Restricting the lid from ballooning may reduce or eliminate the influence of
package shape on the burst test and give the entire sea] area a more equal
opportunity to burst. This will make monitoring of package burst zones more

sensitive to material, tooling and other processing variation.”

Stephen Franks - TM ELECTRONICS-[8|

1.

2.

3.

4.

5.

“With restraining burst test it is possible to get more consistent results.”

“It helps to get a more accurate picture of where the weakest area of the seal is.”
“With restraining plates a more uniform loading is being applied.”

“It provides a way of measuring the minimum seal strength of the package.”

“Do not have to deal as much with the geometry problem.”

Neil Lorimer — REXAM MEDICAL PACKAGING-Ill]

1.

“Restrained burst testing provides a rapid means of evaluating minimum seal
strength (burst strength).”

“Restrained burst testing is more efficient and economical to perform than force
gage testing of peel strength.”

“By restraining the pouch to maintain dimensional stability the stress is more
uniformly applied to the sealed perimeter of the pouc .”

“Restrained burst testing can reliably detect the weakest area of a package seal
placed around the perimeter of a flexible package.”

“Studies using 1” gap on restraining plates have shown correlation coefficients of
0.92 or greater between peel test and burst test results. Burst test results compared to
peel test values for areas of pouches where the lowest seal strength were observed

(burst in that location).”

168

“Burst testing with restraining plates cannot test for general package integrity, it is a
test of seal strength, not package integrity.”

“Burst testing cannot be used to evaluate the strength characteristics of the entire
sealed area since it provides a measure only of the weakest area of the seal. It
should be combined with other methods of seal evaluation to determine uniformity

of seal or other potential peel defects (i.e. bearding, fracturing, etc)”

ATSM Standard- Draft proposal - |4|

1.

“This test provides a rapid means of evaluating tendencies for package seal failure
and minimum seal strength when the package is exposed to a pressure differential.”
“This test provides an indicator of the minimum strength of a seal area around the
perimeter of a package. An indicator of the minimum seal strength may be of
importance to the package manufacturer and end user in ensuring package integrity
and conversely that the seal strength is not so high as to limit opening (peelable

seals) by end users.”

. “This test cannot provide a measure of the uniformity of seal strength above that

minimum seal strength detected by this test method. This test methods of evaluating
uniformity of seal strength or opening functionality or opening functionality.”
“Restraining plates maintain dimensional stability while the package is being
pressurized and more uniformly direct the stresses of pressurization to the perimeter
of the package.”

“In particular this test is intended as applicable to packages with seals that are
intended to have a peelable seal feature (peeled open by end user to remove contents

of package)”

REFERENCES

10.

ll.

12.

13.

REFERENCES

. ASTM Committee D-20, ASTM D1898-68 — “Standard Practice for Sampling of

Plastics”, pp. 470 — 477, 1968 (Reapproved 1989).

ASTM Committee F-2, ASTM F1140-88 — “Standard Test Methods for Failure
Resistance of Unrestrained and Nonrigid Packages for Medical Applications”,
pp. 1075-1077, 1988.

ASTM Committee F-2, ASTM F88-94 — “Standard Test Method for Sea] Strength
of Flexible Barrier Materials”, pp. 893 — 895, 1995.

ASTM Committee F-2, Draft Proposal ASTM Standard — “Standard Test Method
for Burst Test Seal Strength Testing of Flexible Packages using Internal Air
Pressurization within Restraining Plates”, January 1997.

. Bohn, David, “Using Burst Testing to Evaluate Sterile Blister Packaging”, Medical

P_lastics and Biomateriis, pp. 14 —20, Summer 1994,

Bohn, David, “ARO 2600 Burst and Creep Test PE026 Operating Procedure”,
Medtronic, pp. 1 -10, May 1996.

Burgess, Gary, “PKG 310 Technical Principles and Dynamics for Packaging”,
Michigan State University, pp. 76, Revised 1994

Franks, Stephen, TM Electronics, Inc., telephone conversation, august 1997

Lockhart, Hugh, F eliu, Rosamari, “Effect of a Restraining Fixture on the Burst
Values of Sterile Packages”, Medtronic, pp. 1-9, fall 1996.

Lockhart, Hugh, Feliu, Rosamari, “Effect of Gap Size and Seal Geometry on the
Burst Values of Sterile Packages”, Medtronic, pp. 1-9, spring 1997.

Lockhart, Hugh, Feliu, Rosamari, “Effect of Gap and Package Size on the Burst
Values of Pouches for Sterile Products”, Carleton, pp. 1-18, summer 1997.

Lorimer, Neil, “Understanding Restrained Burst Testing”, Rexam Medical
Packaging, April 1997.

MIL-P-44073D, “Military Specification Packaging and Thermoforming of Foods in
Flexible Pouches”, paragraph 4.3.4, 1992.

169

14.

15.

16.

17.

18.

19.

20.

21.

22.

170

MIL-PRF-44073E, “Performance Specification Packaging of Food in Flexible
Pouches”, paragraph 4.3.7, 1996.

Nolan, Patrick J ., “Physical Test Methods for Validating Package Integrity”,
pp. 6 -13, 1997.

Sokal, Robert R, Rohlf, James F., “Biometry - The principles and practice of
statistics in biological research”, W.H. Freeman and Company, San Francisco,
California, pp. 137, 369-376, 1969.

Spitzley, John, Larsen, Curtis, Liebler, Bernard, “HIMA Reference on Sterile
Package”, HIMA Publication 93-7, pp. 1-2, and 5-8, September 1993.

Spitzley, John, PTP9609121 - “Effects of a Restraining Fixture on the Burst Values
of Sterile Packages”, Medtronic, pp. 1 — 4, September 1996.

Wachala, Thomas P., “Correlating Tensile and Burst Test in Pouches”, Medical
Device & Diagnostic Industry. February 1991.

Wachala, Thomas P., “Restrained Vs. Unrestrained Pressure Testing”, Carleton
Technologies Incorporated, pp. 1-15, 1994.

Wachala, Thomas P., Technical Report #95-008 - “Process Capability for the F100-
2600 Package Tester”, Carleton Technologies Incorporated, pp. 1 — 4, April 1995.

Yam, Kit L., Rossen, Jack, and Wu, Xuan-Fei., “Relationship between Seal
Strength and Burst Pressure for Pouches”, Packaging Technology and Science .
Vol. 6, pp. 239—244, 1993.

   

HICHIGAN srnTE UNIV. LIBRARIES
lIHI”Wlllllllllllllllllllllllllllllllllllllllllll
31293017074141