}§M.Wn:§: 6.“...
H.921 .3 urn»,
. F

tum

5 {1.4.1

I. A

2.... «if! 3:311“? t

v. «fitumfiwv hm‘
‘ In,

«N».

3:. ‘

u. h

. ,I‘U
Jfiunfifiw. . ‘ ,. ,r
. €37 a. . in. .
?.=ur..4uah. “will-v.5 L y

'9‘

(I.

L\£ 3:: out! 5.1
.a 1. Li... ifs"..-
i... . . ..
$455.? a...
4 .
. 123393.
ah 10.3.1.4. r
«fit.

 

A,

it

PM 2:} }4..I(¢J..Hn.£!l.
.. 1N1)“: find
3.: E “mlfiiitfnhu Illa
3.». f. "w“
flag} .1‘ int 3.. .5!
1.: Lun1l‘.§|.¢vhlnuytfl.lfl\¢rllh
I.

.

 

THESIS

(1 DEC

 

IIIIIIII IIIIIIIIIIIIIIIIIIIIIIIII IIIIIIIII

3 1293 02061 5880

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

thesis entitled

Structure and Energy Absorbing Properties
of a Polypropylene and Polyethylene
Terephthalate Fiber Composite Material

presented by

Tammy S. Cummings

has been accepted towards fulfillment
of the requirements for

MasteLLs—degree in Mechanics—

W7,

Major professor a

Date 9/34/93;

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

 

PLACE iN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11/00 chlRGDateDmpSS-p.“

STRUCTURE and ENERGY ABSORBING PROPERTIES of a
POLYPROPYLENE and POLYETHYLENE TEREPHTHALATE
FIBER COMPOSITE MATERIAL
By

Tammy S. Cummings

A THESIS

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

MASTER OF SCIENCE
Department of Materials Science and Mechanics

1 999

lh
ve
pr
an

rat

ac.

Sho

DTOc

ABSTRACT

 

STRUCTURE and ENERGY ABSORBING PROPERTIES Of a
POLYPROPYLENE and POLYETHYLENE TEREPHTHALATE
FIBER COMPOSITE MATERIAL

By

Tammy S. Cummings

Due to stricter safety standards enforced by the Federal Government,
such as FMVSS 201, standards of what is acceptable for automotive interiors are
raised to higher levels. As safety and cost become the number one goal of auto
manufacturers, materials must be fabricated to meet existing standards, as well
as future ones.

One such inhomogeneous material is a polypropylene (PP) and
polyethylene terephthalate (PET) energy absorbing material (EAM). This thesis
thoroughly investigates measurements of the amount of energy absorbed at
various strain rates and impacts, and the effects of processing on material
properties such as stiffness. In order to gain an understanding of the morphology
and functionality of the material, failure modes were examined at various strain
rates and impact speeds through microscopy.

The results of this project show that due to the inhomogeneity of the EAM,
a correlation between the way the material functions at various strain rates or
impact speeds and the amount of energy it absorbs is difficult to obtain. It also
shows that on can change the mechanical properties of the EAM by changing

process conditions.

l0\

de

W8
Se;

Rid

hOpI
WW

and

ACKNOWLEDGEMENTS

I would first like to thank two very special people: Brian Erickson and
Linda Chamberlain. \NithOUt the two of them, this would have never been
possible. i would like to thank Linda for giving me inspiration and support for
going back to school and getting my Master’s Degree. Brian, thank-you so much
for being the older brother that i never had. You supported me from beginning to
end. You were not only a good boss, but also what a TRUE MENTOR should
be, as well as my friend. i thank-you for believing in me when i didn’t believe in
myself. You will never know how much your support meant to me.

Mother and Father, I thank you both for your continued support and love. i
love you both very much. it is because of you that l was able to do this. i
dedicate this thesis to the two of you.

i would like to send out a heartfelt thanks to all the guys from my lab who
were there to help me out and lend me their support: Phil Cuicasi, Scott
Seymour, Shawn Corbin, Arvind Krishnaswamy, Ed Drown, and especially
Richard “Ricky” Schalek.

Dr. Drzal, I am honored to say that l was once a student of yours, and l -
hope that i lived up to what it means to be a student of LT. Drzal. I appreciate

your guidance and patience, and most of all for being a wonderful mentor, guide,

and friend.

iii

TABLE OF CONTENTS

 

1. introduction
2. Literature Review
General Information
FMVSS 201
Various Materials
3. Materials and Methods
Materials
Materials
Materials and Processing
Methods
Differential Scanning Calorimetry
Optical Microscopy
Dyna-tup impact Testing
Dynamic Mechanical Analysis
High Strain Tensile Machine
Impact/Penetration Machine
4. Results and Discussion
Thermal analysis of PP and PET EAM Fiber
Optical Microscopy of the Failure Modes of EAM
Processing and EAM
Dyna-tup Characterization of the EAM
DMA Analysis of the EAM Part l
Voids and EAM
DMA Analysis of the EAM Part II
E’, E” and Tan 8 Comparisons
Results of High Strain Tensile Tests
ESEM Microscopy of High Strain Tensile Samples
Results of Impact/Penetration Tests
ESEM Microscopy of impacted Samples
5. Conclusions
6. Suggestions for Future Work
7. References

iv

(DOOM

are

17
18
20
23
27
30

35
36
40
50
51
57
61....
79
83
92

104
115
120
121

 

 

ll

11

12
13.
14.
15.‘

16.?

17¢

LIST OF TABLES

 

1. Table 1: Density vs. Temperature parameters used for first group Of samples.
2. Table 2: Matrix of Time, Temperature, and Pressure Processing Conditions.
3. Table 3. Glass transition and melt temperatures of PP and PET [16].

4. Table 4: Matrix of Density vs. Temperature tests and 1200 glrn2 density vs.
Time.

5. Table 5: DMA parameters for various thickness tests.

6. Table 6: Summary of Initial Storage Modulus (E') and inflection Temperatures
(T g) for various thicknesses and frequencies.

7. Table 7: Summary of Storage Modulus Values for 1.2 mm thick samples
tested from -30-100°C at 1.5 Hz.

8. Table 8: Summary of initial Storage Modulus for various thicknesses run
intennittentiy at 150 Hz.

9. Table 9: Summary of initial Storage Modulus (E') for various thicknesses run
consecutively at 150 Hz.

10.Table 10: Void data of 1.5, 1.2, and .65 mm thickness samples taken from the
edges Of compressed sheets.

11.Table 11: Void data of 1.5, 1.2, and .65 mm thickness samples taken from the
center of compressed sheets.

12.Table 12: Processing parameters of the EAM.

13.Table 13: E' results of DMA machine direction samples at -100°C.
14.Table 14: E' DMA results in cross-machine direction at -100°C.
15.Table 15: E' DMA results of machine direction samples at 20°C.

16.Table 16: Comparison of E’ values from production and enlarged EAM
samples.

17.Table 17: E' DMA results of cross-machine direction samples at 20°C.

18.Table 18: Results of high strain data for all three speeds.

19.Table 19: Results of impact/penetration data for all three speeds.

vi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7. I
8. l
9. I
10.I
11.F
12.F
13.F
14.F

15.F

16F

17.F

18":

19F

1. Figure 1:
. Figure 2:
. Figure 3:

. Figure 4:

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

. Figure 9:

10. Figure 10:
11.Figure11:
12.Figure 12:
13.Figure 13:
14. Figure 14:

15. Figure 15:

16. Figure 16:

17.Figure 17:

18. Figure 18:

19. Figure 19:

LIST OF FIGURES

 

Chemical structure of PET [8].
Chemical structure of PP [8].

Carver Laboratory Press Model 2731.
Drawing of enlarged EAM mold.
Mounted sample specimen of the EAM.
Dyna-tup testing set-up.

TA instruments three-point bending set-up [12].
High strain tensile set-up.

High strain tensile specimen.

Sample high strain tensile curve.
Impact/penetration set-up.
impact/penetration sample set—up.
Sample impact/penetration curve.
EAM element examined by regions.

Non-impacted sample with little or no fluorescent epoxy in the
EAM element.

impacted EAM samples with damaged areas showing high
concentrations of fluorescent epoxy.

Second layer of EAM element showing consistent results from
layer to layer.

lnstron Data Young’s Modulus Values of 1000, 1200, and 1400
g/m2 Density samples vs. Temperature for machine direction.

lnstron Data Total Energy Values of 1000, 1200, and 1400 g/m2
Density Samples vs. Temperature for machine direction.

vii

 

20. Fig

21.Fig
22.Fig
23.Fig

24. Fig
25. Fig
26. Fig
27.Fig
28. Fig

29. Fig

30. Fig
31.Fig
32.Fig
33.Fig
34. Fig

35. Fig
36. Fig.

37. FigL

20. Figure 20: lnstron Data Young’s Modulus Values of1000, 1200, and 1400

21.Figure 21:

22. Figure 22:

23. Figure 23:

24. Figure 24:
25. Figure 25:
26. Figure 26:
27. Figure 27:
28. Figure 28:
29. Figure 29:

30. Figure 30:
31.Figure 31:
32.Figure 32:
33. Figure 33:
34. Figure 34:

35. Figure 35:

36. Figure 36:

37. Figure 37:

glm2 Density Samples vs. Temperature for cross-machine
direction.

lnstron Data Total Energy values of 1000, 1200, and 1400 g/mz
Density samples vs. Temperature for cross-machine direction.

Pseudo-Energy Values of 1000, 1200, and 1400 g/m2 Density
Samples vs. Temperature for machine direction.

Pseudo-Energy Values of 1000, 1200, and 1400 g/m’ Density
Samples vs. Temperature for cross-machine direction.

1400 g/m2 Density Average Weight vs. Temperature plot.
1200 glm2 Density Average Weight vs. Temperature plot.
1000 glrn2 Density Average Weight vs. Temperature plot.
1200 g/m’ Average Weight vs. Time plot.
1200 g/m2 Psuedo—Energy vs. Time plot.

ESEM micrographs of 1.5, 1.2, and .65 mm samples and their void
contents.

Summary of void data by thickness and location.

E' Graph for machine direction samples at -100°C.

E' Graph for machine direction samples at 20°C.

E' Graph for cross-machine direction samples at -100°C.
E' Graph for cross-machine direction samples at 20°C.

Initial modulus values for cross-machine and machine direction
EAM as a function of time in the heater.

initial modulus values for cross-machine and machine EAM as a
function of time in the mold.

Initial modulus values for cross-machine and machine EAM as a
function of temperature.

viii

 

 

38.F

39F
40F
41F
42F

43.F

45F
46F
47.F

48F
49F
50F
51F
52.1:

53;

55. F,-

56. Fj

38. Figure 38:

39. Figure 39:
40. Figure 40:
41. Figure 41:
42. Figure 42:
43. Figure 43:

44. Figure 44:

45. Figure 45:

46. Figure 46:

47. Figure 47:

48. Figure 48:
49. Figure 49:
50. Figure 50:
51.Figure 51:
52. Figure 52:
53. Figure 53:
54. Figure 54:
55. Figure 55:

56. Figure 56:

initial modulus values for cross-machine and machine EAM as a
function of pressure.

E” values for machine direction at -100°C.

E” values for machine direction at 20°C.

E” values for cross-machine direction at —100°C.
E” values for cross-machine direction at 20°C.
Tan 8 values for cross-machine direction at 20°C.

3-D graph of total energy absorbed during high strain rate testing
as a function of time in the heater.

3-D graph of total energy absorbed during high strain rate testing
as a function of time in the mold.

3-D graph of total energy absorbed during high strain rate testing
as a function of temperature.

3-D graph of total energy absorbed during high strain rate testing
as a function of pressure.

9-mph clean failure edge with little tearing.

15-mph torn and “chunky” failure edge.

3-mph with a mixture of nine and 15-mph failure edge.
PP seen in clumps throughout the EAM.

Fractured PET fiber surface.

PET fiber surrounded by PP (interfaciai failure).

PP fibrils attached to PET fibers.

3-D graph of total energy absorbed during impact testing as a
function of time in the heater.

3-D graph of total energy absorbed during impact testing as a
function of time in the mold.

ix

57. Figure 57:
58. Figure 58:

59. Figure 59:
60. Figure 60:
61. Figure 61:
62. Figure 62:
63. Figure 63:

64. Figure 64:

65. Figure 65:
66. Figure 66:
67. Figure 67:
68. Figure 68:

69. Figure 69:
70. Figure 70:

71 . Figure 71:

72. Figure 72:
73. Figure 73:

74. Figure 74:

3-D graph of total energy absorbed during impact testing as a
function of temperature.

3-D graph of total energy absorbed during impact testing as a
function of pressure.

3-mph impact clean impact edge.

15-mph impact torn and “chunky” failure edge.

9-mph mixed failure edge.

Brittle, clean surface failure of PET fibers at 3—mph impact.
Jagged and torn PET fiber failures at 15—mph impact.

Mixture of brittle, clean and torn and jagged PET failure at 9-mph
impact.

Stretched and broken PP throughout the EAM.
Brittle failure of PET fibers.
Breaking of PP among PET fibers.

Non-unifon'n dispersion of PP, along with evidence of PET fiber
pullout of PP.

Failure of EAM in areas of high PET fiber concentrations.
Remnant PP on PET fiber.

High temperature processed EAM with more uniform PP
distribution.

High pressure EAM sample with more compacted and denser PP
appearance.

Summary of effects of processing parameters (Time in the heater;
time in the mold, temperature and pressure) for high strain testing
as percent changes from baseline samples.

Summary of effects of processing parameters (Time in the heater;
time in the mold, temperature and pressure) for high strain testing
as percent changes from baseline samples.

INTRODUCTION

 

An energy absorbing material manufactured from polypropylene and
polyethylene terephthalate fibers was developed. This material is installed on the
A, B, C pillars along with the side rails of the headliner in order to help OEM’s
meet the requirements of Federal Motor Vehicle Safety Standard (F MVSS) 201 -
Occupant Protection in Interior Impact by 1998. In order to be an acceptable
energy absorbing material for this application, it must meet a HIC number of
1000 or below for various OEM’s. The material is relatively new and performs in
an acceptable manner in its current configuration but it has not been analyzed
for its optimum energy absorbing potential relative to its composition and
configuration. The goal of this project is to determine how this material absorbs
energy, and then to maximize the energy absorbing potential in order to meet a
HIC number Of1000.

In order to accomplish this, the EAM will be tested and analyzed on both
macroscopic and microscopic levels. impact testing will be performed in order to
gain an understanding of how the EAM responds to damage, looking both at
large scale damage and deformation, along with microscopic, fiber level, damage
and deformation. Thermal analysis of the material will be performed in order to
gain an understanding of the constituent materials themselves and their
interactions with one another under various processing conditions. Microscopy

will be necessary to evaluate and explain material behavior.

LITERATURE REVIEW

 

General Information

The use of composites is becoming more and more prevalent in our
society today. We see composites being used in a number of industries such as
recreational equipment, automotive, aircraft, and aerospace. Manufacturers are
using composites more to reduce weight in their products, but also to take
advantage of properties like high strength to weight and high stiffness to weight
ratios, and to dissipate energy. Car bumpers and interior car component
structures are examples of composites being using to provide a degree of
occupant safety. Composites are being used to dissipate energy now that safety
is becoming a higher priority of automakers due to the creation and modification
of stringent safety standards.

Of the nearly 5800 cases of severe head injuries involving passenger cars
and light-duty trucks and vans, 37% are fatalities and 67% are serious injuries
with an abbreviated injury scale (AIS) rating of three or greater. These injuries
involve primarily the front-seated occupants, with most of them involving the
driver [2]. impact with the upper interior components of automotive vehicles is
the leading cause of head injury for non-ejected occupants killed in an
automobile crash. Accident data shows that occupant head injuries result
primarily from head contact with a vehicle’s pillars, side rails, headers and other

components during a crash [1].

Various standards created by the National Highway Traffic Safety
Administration (NHTSA) and the Department of Transportation (DOT) are in
effect to address this issue:

Std. NO. 208 — Occupant Crash Protection, limits the forces and
accelerations that are imposed on the head of a crash dummy in a frontal, 30-
mph crash test. It has been effective at reducing actual fatality risks and together
with the use of safety belts, has significantly reduced fatality risk.

Std. No. 201 - Occupant protection in an interior impact too effect on
January 1, 1968. The standard set requirements for instrument panels (lP),
interior compartment doors, seat backs, sun visors, and armrests (AIR) to lessen
injuries to persons thrown against them in crashes. A large number of occupant
injuries and fatalities result from head impacts with upper interior components not
covered by Std. No. 201 [1].

FMVSS 201

 

To perform the tests of FMVSS 201, a modified headforrn is used. This
headforrn lacks the nose of the Hybrid ill head to eliminate interference from the
nose during testing. The current headforrn is instrumented with tri-axial
accelerometers, positioned to measure the acceleration at the headfonn’s center
of gravity. These measurements are used to calculate the magnitude of the
potential for injury resulting from the impact.

The 15-mph test speed was chosen because it is the current test speed
used in Std. No. 201. It is the average speed at which the onsets of serious

injuries occur. it also represents the velocity at which the headfon'n contacts the

 

UPP'
isin

isag

adva

finpa

com;
4.5 kg
m/s.

relatiI

free n
II TEpr

actual

Std. No

schOol t

upper interior component and is lower than the actual speed at which the vehicle
is impacted. NHTSA has concluded that the proposed 15-mph F MH impact test
is appropriate for all components regardless of their locations.

The flight of the headforrn should be “free-motion” rather than guided. The
advantage is that the F MH can simulate the glancing and non-perpendicular
impacts experienced in real world crashes.

A new revised test procedure mandates that all upper-vehicle interior
components should be tested by {impacting a featureless Hybrid lll headforrn of
4.5 kg that travels in free-flight mode for not less than 25 mm at a velocity of 6.70
m/s. The tests are to be conducted at a temperature of 225°C 1 35°C with a
relative humidity between 10-70%.

HIC is calculated using the acceleration readings from an instrumented
free motion headforrn (FMH) and transforming it to a dummy equivalent HlC(d).
It represents the HIC that would normally be experienced by a full dummy or
actual vehicle occupant.

HlC(d) = .75446(FMH HIC) + 166.4

FMH HIC = [1/(t2 - t1) [a an"? (12 - t1)

a = resultant acceleration expressed as a multiple of g

(t; — t1) are any two points in time during the impact which is not greater

than 36 ms.

The HIC is an appropriate injury criterion to measure injuries by NHTSA.
Std. No. 208 occupant crash protection, No. 213 child restraint systems, No. 222

school bus passenger seating and crash protection each uses a HIC limit of

 

1C

58

39F
hea
the
heat
cont
head
cont;

conua

sheet
impaci

depenl

cOmbin

load/Clel

redesig,
pIaSIIC 5

material

1000. Research has shown that using an upper limit of 1000 HIC prevents
serious injuries in actual crashes.

For each impact zone, the proposed test procedures defined a range of
approach angles at which the FMH would strike any point in that zone. The
headforrn could be launched from any location inside the vehicle, provided that
the specified approach angles and the following restrictions were met. The
headfonn had to travel through the air for a distance of at least 25-mm before
contacting the vehicle interior surface. At the time of initial contact between the
headfonn and the vehicle, a specified portion of the headforrn's forehead must
contact some portion of the target circle and on a portion of the headforrn may
contact any part of the vehicle outside of the specified impact zone.

impacting points usually involve areas of the car structure where several
sheet— metal stampings or reinforcements come together. The distance of the
impact targets from these joints is usually around 125 mm. HlC(d) values
depend on the location evaluated and the direction of impact or approach angles.

The acceptance criterion is now based on a functional relationship that
combines the acceleration and time rather than an absolute headforrn
acceleration value. The ideal energy dissipater is represented by a square-wave
load/deflection response [1].

According to GE plastic engineers, the revised FMVSS 201 will force the
redesign of the upper material trim, with pillars being the greatest challenge. GE
plastic engineers also think that the HlC(d) is driven mainly by design rather than

materials [2].

 

 

pl;
me

be

sen
load
red u
load

more

mUItIp
They I"
are cur
bOISIerg
meet F)
shafierir.
and 5m

Dans elin

 

Energy Absorbing Materials

Various materials have been investigated in order to meet energy
absorbing requirements in F MVSS standards such as steel, thermoplastics,
polymer and polyurethane foams. Steel "crush-cans," are manufactured to yield
plastically under load. This usually involves axial crushing against the rigid
member. Crush-cans, however, are limited since they have a potential for
bending rather than crushing axially when non-axial forces are encountered.
Therrnoplastics are a bit more flexible than steel. They, however, are also
sensitive to non-axial loading like steel crush-cans. They offer greater off-axis
load-bearing capability than steel crush-cans, but the load-bearing capability is
reduced, and may cause plastic behavior. Polymer foams can perform well in all
load directions and can have a large range of crush characteristics. They are
more versatile since they can be tailor made [3].

Polyurethane (PU) foam is used in various light duty trucks and
multipurpose passenger vehicles (MPV) to improve vehicle crashworthiness.
They have a very good history of effective energy-absorption properties. They
are currently used in vehicle interiors and exteriors for applications such as knee
bolsters to help OEM's meet FMVSS 208 and hip and shoulder bolsters to help
meet FMVSS 214. Energy is absorbed through elastic deformation or the
shattering of cell walls and struts in PU foams. PU foams can be easily altered
and since it originates as a liquid, it can be foamed in place behind interior trip

parts eliminating the need for fasteners and adhesives. PUs are also

 

ar
pn
kn.
Wa
Cor

the

advantageous since they exhibit isotropic energy-absorption behavior and are
not sensitive to impact direction [3].

PU foams can absorb energy through different mechanisms. Recoverable
foams store a great deal of energy through elastic deformation of semi-flexible
cell walls and struts. After impact, the stored energy is released as rebound
energy. Some energy is also absorbed by the defamation of the PU cells and
struts. Since they are very flexible, recoverable foams can usually recover
almost all of their original shape after impact and can even withstand multiple
impacts. Rigid foams, however, dissipate energy through shattering of cell walls
and struts. They are usually non-recoverable. They will also absorb more
efficiently with less rebound than recoverable foams. They are more useful in
limited space areas where impact energy must be absorbed [4].

Honeycomb structures are another method of absorbing energy. They are
continually used in impacting faces to simulate actual crash conditions without
damaging test fixtures. Man-made honeycomb is manufactured by fusion
bonding, adhesive bonding, or welding thin sheets of paper, plastic, or metal
together. Aluminum and polypropylene and polyethylene honeycomb structures
are some examples. Honeycomb absorbs energy by crushing under loads. it
provides efficient “9” limit protection in applications such steering columns and
knee bolsters. Under compression, it carries load through the stabilized cell
walls, as failure occurs, the walls locally fail along the edge. This failure
continues along the cell walls until a solid block of material is achieved through

the continual collapsing of the cell walls [5].

Energy absorbing thermoplastics such as general-purpose low glass
acrylonitrile butadiene styrene (ABS), polycarbonate (PC), low gloss PC/ABS,
and high crystalline polypropylene (HCPP) are helping to meet and exceed
various FMVSS requirements. These materials are mainly used in pillar trims in
the interior of a vehicle. Testing of these four materials in FMH testing by Locke g
and Clark showed the ways in which these materials managed energy varied
significantly. This leads to the conclusion that HiC(d) is driven mainly by design
than material. Pillars molded from ABS, PC/ABS, and PC crush during impact,
which is the desired behavior since it provides continual energy management
during impact. HCPP ribs fractured at and around the impact site. Due to this
type of failure, the force of the impact is transferred to the underlying steel
surface since energy absorption by the plastic pillar ceases [6].

New materials are being developed constantly to meet the continually
changing and more stringent FMVSS standards. One of these materials is an
energy absorbing material composed of PP and PET fibers. The material is
confined in a 75/25 ratio has been fabricated for its energy absorbing potential.
The material is processed at set conditions, which melt the PP bUt not the PET
resulting in a stiff fibrous material. The material has a unique material geometry
and polymer combination. This project focuses on figuring out how the material

functions and ways to improve the energy absorbing potential.

 

MATERIALS and METHODS

Materials

A new material composed of recycled polypropylene and polyethylene
terephthalate post-industrial fibers was created. The patented unique structure
formed from this material is being used as an energy absorbing material in' the
interior of vehicles, in areas such as the A, B, and C pillars, along with the side
areas along the headliner.

The process of manufacturing the EAM starts with post-industrial PP and
PET fibers mixed in a 75I25 ratio, followed by processing in a special “weaving”
machine, which weaves the fibers into a “vertical” fabric-like material at a density
of 1200 g/m’. The weaved material is then cut into shape and then put in a
convection heater at 216°C for one minute while being compressed to a
thickness of five millimeters. The hot material is then taken out of the oven and
placed in the EAM mold. The material is compressed over a mold, which also
stretches the material in a water-cooled mold (16°C) for approximately 30
seconds.

Polyethylene terephthalate is a high-performance thermoplastic. its

chemical structure is shown in Figure 1 below.

‘ {IF—ma.

Figure 1: Chemical structure of PET [8].

in its crystalline form, the PET has a very high flex modulus and a high
heat deflection temperature, but very low impact strength. Impact strength can
be improved by adding elastomers or other fillers [7]. Polypropylene (PP) has
good chemical resistance along with satisfactory mechanical properties. Its

chemical structure is shown in Figure 2 below.

CH3

:ECHZ—cH—]» n

Figure 2: Chemical structure of PP [8].

It is relatively inexpensive and is easy to process. PET has good chemical and
solvent resistance and better mechanical properties than PP. It is difficult to
process though. Both materials are used frequently for packaging applications
[7].

PP and PET are incompatible due to differences in their chemical nature

and polarity. Blending them will create a clear two-phase morphology.

I0

Generally, the strength and stiffness of their blends increases with increasing
PET content, but they still exhibit very poor impact strength [7].

Studies analyzing the compatibility of the two materials by Morye et al,
showed that the viscosity of a PP-PET blends was lower than that for PP even
though PET has a higher viscosity. This seemed to indicate incompatibility
between PP and PET. In addition, adding PET to PP resulted in a lower tensile
strength of PP showing again that the two are incompatible [7].

Studies by Cheung and Chan showed that increasing the weight
percentage of PET to PP increased the modulus. “Good interfaciai bonding is
required if the blend is to have good mechanical properties. The mixing of PP
and PET results in clumps of one phase dispersed in the other" [9]. Due to the
incompatibility between the two polymers, the interfacial adhesion is very poor.
Thermodynamics does not favor the mixing of the two polymers. “The
interpenetration of chains from the two polymers at the interface is poor” [9].
This poor interpenetration leads to poor interfaciai adhesion.

An energy absorbing material consisting of PP and PET fibers confined in
a 75/25 ratio has been fabricated for its energy absorbing potential. The fibers
are processed at certain conditions, which melt the PP but not the PET resulting
in a stiff fibrous material. The unique material geometry and polymer
combination provides energy absorption capacity in the structure.
Material and Processing

The first set of samples was fabricated using the parameters shown in

Table 1. Mat material at three different densities were evaluated at five different

I]

 

densities were evaluated at five different temperature conditions to see what
effect the amount of material and the temperature had on the EAM. Mat material
consisting of 70% recycled PP and 30% recycled PET was first cut into various
sections from three different density material rolls. After the mat material was
gathered, the material was heated to the desired temperature for two minutes in
a convection oven. While the sample is being heated the oven, the sample is
compressed to a five-mm thickness. The heated and compressed mat material
is then placed in the corrugated mold, which is completely automated with a pre-
programmed open and close cycle time.

During the fabrication of these samples, large amounts of shrinkage of the
material were seen when the material was removed from the heater to the mold
at both the 193°C and 204°C temperatures. As a result, the samples fabricated
at these two temperatures were thicker than the rest of the samples fabricated at

216°, 227°, and 238°C.

 

 

 

 

 

 

 

 

 

1 93°C 204°C 21 6°C 227°C 238°C
1 000 glm2 1 2 3 4 5
1200 glma 6 7 8 9 1 0
1400 glm2 11 12 13 14 15

 

 

 

 

 

 

 

 

Table 1: Density vs. Temperature parameters used for first group of samples.

The second set of samples was fabricated at Michigan State University

(MSU) to evaluate stiffness as a function of thickness within the material. Sample

 

 

sheets of various thicknesses had to be fabricated. The thicknesses were 1.5,
1.2, and .65-mm. These measurements corresponded to different thickness
regions (A, B, C, D) of the EAM structure determine from previous micrographs
where regions A and C were approximately 1.5-mm thick, region B
approximately 1.2-mm thick, and the apex region D was approximately .65-mm
thick. Flat samples were processed in a four platen Carver Laboratory Press
dual daylight Model 2731 hydraulic press (Figure 3). The first set of platens were
heated to 216°C, the second set of platens were set at approximately 16°C.
Samples were made with virgin PP and PET at a 75/25 PP and PET mixture
ratio. Unconsolidated mat samples were placed between Teflon coated glass
fabric release film and then compressed to approximately five-mm using steel
shims at a maximum pressure of 24 KSI on the press. After heating, the hot
mats were then compressed between two steel plates to the various thicknesses
that were desired. While this was being done, the platens were cooled with
running water to approximately 16°C for approximately three minutes until the
bottom platen reached 66°C. After three minutes, the compressed material and
steel plates at various thicknesses were then inserted in the upper platen for two
minutes with the appropriate shims and then removed. The fabricated sheet
samples were then removed from the steel plates and then measured using a

micrometer along the edges of the sheets to verify the thickness of the samples.

I3

 

Figure 3: Carver laboratory press Model 2731.

In addition, a third set of samples was fabricated using the processing
conditions shown in Table 2 to evaluate the effects of different processing
conditions on the material. Time of the mat material in both the convection
heating oven and the time in the mold, the temperature of the heater, and
displacement of the mold were of concern. Samples were fabricated on a large,

scaled up mold and were processed (Figure 4) in order to get sample sections

large enough for testing. Samples needed to be representative of production run
samples since the samples fabricated at MSU were not representative. Each
region experiences different stresses when being formed in the mold, but due to
the lack of constraints experienced by the samples fabricated at MSU, these

samples experience none of the same stresses.

{— 30mm

 

Figure 4: Drawing of enlarged EAM mold.

 

 

Samples were first heated in a convection oven and formed in the

enlarged mold for the temperatures and times shown in Table 2. Samples were

approximately three times the original size of the EAM element.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TIME 1 TIME 2 TEMP PRESSURE
(HEATER sec) (MOLD sec) (°C) QISP mm)
60
TIME 1 90 60 60 60
(HEATER sec)
120
TIME 2 45 30 45 45
(MOLD sec)
60
216
TEMP (°C) 216 216 227 216
238
PRESSURE 1.5 1.5 1.5 1.5
(DISP mm)
1.25

 

 

 

Table 2: Matrix of Time, Temperature, and Pressure Processing Conditions.

In order to evaluate this EAM and its properties, a variety of different

testing and characterization instruments were used, including differential

scanning calorimetry, optical microscopy, Dyna-tup impact testing, dynamic

mechanical analysis, high strain tensile testing, and impact/penetration testing.

The following sections discuss the instruments and methods in detail.

2,-4.-

 

 

Methods
Differential Scanning Calorimetfl

Differential Scanning Calorimetry (D80) is a device used to determine
temperature and heat flow for a given material as a function of time and
temperature. It can also determine information such as glass transition
temperature, crystallization, melting transitions, and heat capacity of a material.

The principal of operation of DSC involves monitoring the difference in
temperature between a reference and sample while both are being subjected to
a specified temperature profile. The energy difference between the two is
recorded as a function of temperature. The thermal events, endothermic (heat
absorption) and exothermic (heat evolution) reactions, appear as deviations from
the baseline given by the DSC. Exothennic responses .are positive which
corresponds to the evolution of heat in the sample compared to the reference,
while endothermic responses are negative which is indicative of absorption of
heat. These responses can be caused by melting, oxidation, and possible phase
changes occurring in the material. It is this information that helps one identify the
changes which a sample material undergoes with temperature as well as identify
both the material and its processing constraints.

The glass transition temperature (T9) is characterized by the onset of
motion in short segments of the polymer chain. This temperature is indicative of
the amorphous regions of the polymer. Below the T9, the polymer is hard and
behaves like a glass. Above the T9, the polymer becomes more flexible and

rubbery.

l7

The melting temperature (T m) is the temperature at which the material
goes from a solid phase to a liquid phase.

Both of these parameters are important and can be used to characterize
and verify the constituent materials of the EAM and its properties.

For this project, a TA Instrument DSC 2920 was used to identify glass
transition and melting temperatures of PP and PET. By obtaining these two
characteristic temperatures, verification of the composition of the two polymers
was achieved. The scans could also provide information about constituent
purity.
thical Microscogy

Optical microscopy is useful for evaluating a specimen at the microscopic
level using instruments such as optical microscopes or scanning electron
microscopes (SEM). For this research both an Olympus BH2-UMP optical
microscope with a barrier and band pass filter and an environmental scanning
electron microscope (ESEM) were used to examine specimens.

Samples were mounted in Buehler sampl-kwick fast cure acrylic resin and
polished on an Abramin sander with 240 to 4000 grit paper for three and a half
minutes with approximately 50 N of force. Optical micrographs were recorded
using an Olympus BH2 optical microscope. A typical mounted sample

containing a cross-section of the EAM is shown in Figure 5.

 

 

 

Figure 5: Mounted sample specimen of the EAM.

Samples were also impregnated with Struers Epofix epoxy resin
containing a fluorescent dye by applying a vacuum to the sample at room
temperature for 10 minutes. lmpregnated samples were then cured overnight
and then mounted in Buehler sampl-kwick fast cure acrylic and polished on an
Abramin sander with 240 to 4000 grit paper for 3.5 minutes and approximately
50 N of force as before. They were then examined using an Olympus BH2-UMP
optical microscope with a barrier and band pass filter. These samples were also
examined with an ElectroScan environmental scanning electron microscope

(ESEM).

Dyna-tup Impact Testing

impact testing is quite different from compression or tensile testing. in
impact testing, samples experience high loads in a shorter time over a small
surface area. The loading rate is very high so that there is no relaxation and
heat is generated. A totally different behavior of the material can be detected
under impact conditions than under slow strain rate testing.

The velocity and force are very critical in impact testing. The velocity is
measured immediately before contact. The sample is rigidly fixed to eliminate
vibration. The sample vibration after impact is dissipated energy. Heat is also
dissipated after the impact. Strain gauges are used to measure the deformation
experienced by the sample. A piezoelectric force transducer records the force
which is converted to a load and then to a stress. A force versus time or force
versus displacement graph can be generated from the impact data. The
following step by step process is used to calculate the energy.

The force is measured by knowing the acceleration, which is F(t)/m, where
F (t) is the force as a function of time, and m is the mass. This is the acceleration
at the contact point. By integrating the acceleration [Ia(t) = ldv/dt], one can find
the velocity. Since the initial conditions are known, the velocity history is known
at any instant of time. By integrating the velocity [iv(t) = idx/dt], one can obtain
the displacement. Plotting the force [F(t)] versus the displacement [x(t)] yields

the energy curve. Energy is equal to the force multiplied by the distance.

20

There are some factors to consider when doing Dyna-tup testing. One is
the assumption that the motion of the tup (impactor) impact is equal to the
movement of the specimen, assuming rigid bodies. The linear variable
displacement transducer (LVDT) only measures the displacement of the
backside of the specimen, so any deformation occurring in the other direction is
ignored. The energy from the tup is not necessarily equal to the energy
absorbed by the material. The exact velocity of the tup at the impact point must
be calculated.

Since it is known that the force on the specimen (the EAM sample) equals
zero right before impact and after the tup rebounds off the material, the period
between these two events is the force vs. time history. The following step by

process is used to calculate the impact velocity [11].

VELOCITY OF A FREE FALLING BODY
V1 = Vo + g(t1 - to) (initial velocity before impact)
V2 = V,, + g(t2 - to) (final velocity after impact)

Equate Vo yields:

V2 = V1 + g(t2 - t1) 1 (second velocity in terms of first velocity)
(1)

POSITION OF A FREE FALLING BODY

x1 = x0 + V,,(t1 - to) + g/2 (t1 - to)2 (first position)

x2 = x0 + V,,(t2 - to) + g/2 (t2 - t,,)2 (second position)

Equate xo yields:

21

l—

Va = (X2 ’ X1) 7' 9,2 [(ti ' to)2 " (t2 " toIZI/ t2 ' t1
Let t,,, x0, Vo = t,, x,, V1 yielding:

V, = (x2 - x,)/(t2 - t,) - g/2(t2 - t,) (initial velocity in terms of displacement
and time) (2)

Substitute eq'n (2) into (1) yields:
V2 = (x2 - x,)/(t2 - t,) + g/2(t2 - t,)

The Dyna-tup machine itself has a maximum drop height of approximately
five feet. Subtracting out the height of the impactor yields a total height of
approximately three feet. This translates to an impact velocity of approximately
15.2 km.lhr.

Preliminary tests using a Dyna-tup machine were conducted using a new
tup fixture, which was designed and fabricated especially for the testing, shown

in Figure 6.

22

 

Figure 6: Dyna-tup testing set-up.

The original set up was modified in order to accommodate the EAM and
its requirements for testing. The most crucial part of the set up was aligning the
tup and the bottom plate. It was crucial that the tup was directly over the point of
impact so that the transducer, which is located underneath the aluminum plate,
would not be damaged. The load cell could only handle a maximum load of

approximately 909 kg without damage.

23

Dynamic Mechanical Anafl/zer

A dynamic mechanical analyzer (DMA) is an instrument, which is
designed to measure visco-elastic properties, like modulus and energy
dissipation for both soft and stiff solid materials. Various shapes and sizes along
with various clamping arrangements can be used.

A DMA works by applying an oscillatory or sinusoidal strain or stress to a
material and then measuring the resulting stress or strain developed in the
material. Mechanical defamation is applied to samples by the DMA drive motor
sinusoidaliy. Deformations can be applied in a step fashion or at a fixed rate.
For an ideal solid obeying Hooke’s Law, the resulting stress will be proportional
to the amplitude of the applied strain. The phase angle between the stress and
strain will be 0°. For an ideal fluid obeying Newton’s Law, the resulting stress will
be proportional to the rate of strain. The phase angle of the stress will lead the
phase angle of the strain by 90°. For visco-eiastic materials, a complex modulus
is introduced, E*. This consists of a storage modulus E’ and a loss modulus E",
thus giving E* = E’ + E”. The storage modulus is a measure of stiffness, where
the loss modulus is a measure of the amount of energy retained after testing [12-
13]

For this project, both a TA Instruments 2980 and Du Pont 983 DMA were

used.

24

The TA Instruments 2980 DMA operates by choosing a particular set-up
and then programming the respective conditions. For this project, the three-point
bending clamping mechanism was installed and calibrated (Figure 7). Samples
were tested in a horizontal set-up. The set-up involved using the liquid nitrogen-
cooling accessory, along with compressed air. Samples were run using a
multifrequency mode, which tests the sample at a constant amplitude, and at a
constant frequency. The autostrain was also used. The autostrain puts a static

force on the sample and adjusts this value during the experiment [14].

 

Figure 7: TA Instruments three-point bending set-up [12].

The Du Pont 983 DMA operates under four modes of operation. For this

project, the fixed frequency oscillation mode was used. Samples were tested in

a vertical clamping set-up. Samples were clamped between two parallel arms,
which are mounted on low-force flex pivots which only allow motion in the
horizontal plane. The distance between the arms can be adjusted. An
electromagnetic motor attached to one of the arms drives the sample to the
desired amplitude. A linear variable differential transformer (LVDT) mounted on
the driven arm measures the sample response to the applied stress. An
adjustable thermocouple is mounted close to the sample to give accurate
feedback information to the temperature controller.

The DMA was used in this project to determine how the EAM was affected
by processing parameters. Obtaining these profiles of a material is a very
accurate way of predicting how a material will respond to certain conditions. By
plotting the E' as a function of temperature, the stiffness profile of a sample can
be seen as the temperature goes from law to high temperatures or vice versa.
The DMA was also used to see at what point a material will degrade when
subjected to various strain rates. This was done by subjecting samples at room
temperature to consecutive strain sweeps. (A strain sweep test is when the
sample is subjected to various levels of strain repeatedly.) The storage modulus
was then plotted versus various percent strain. When repeated several times,
the point at which the storage modulus decreases at a particular strain indicates
the point of material degradation. The DMA data also revealed the true material
machine direction of the material. This was done by testing a sample in machine

direction and cross-machine direction. The resulting storage moduli were plotted

26

against temperature. The difference in moduli was indicative of anisotropy
between machine and cross-machine direction.
High Strain Tensile Machine

In order to determine the ability of the material to absorb energy under
impact conditions, high strain tensile tests were performed using a hydraulic
tensile testing machine. The tests at various strain speeds, given in meters per
second (mps), were performed at the IKV at RWI'H — Aachen University using
the machine and instrument set-up shown in Figure 8. The auxiliary equipment
used in the experiments included a water cooling system, a ww700 transient

recorder, a Kistler 5001 charge amplifier, and Zwick REL control system.

Kistler
5001
charge
_ amplifier

Zwick REL .
- .. Control
ww70 _ ; g ' ‘ System
Transient ‘ ..
Recorder

 

Figure 8: High Strain Tensile Set-Up.

27

The high strain tensile testing system operates in the following manner.
The control system must be on and warmed up for a period of time. The system
parameters for warming up the machine are as follows: the machine is first reset,
the pump is turned on, a sinusoidal wave function is selected, and the negative
test direction is chosen. The Level and Amplitude control the height of the lower
testing arm and grip. The Level is raised to approximately 400 and Amplitude is
raised to approximately 180. The system should run for approximately 10
minutes before changing the parameters for testing.

For testing specimens, the following parameters were used: a ramping
wave function, and a single and positive (tensile) testing direction. The “active“
button was also set, which activates the machine to be able to test at the higher
speeds. A Kistler 9331A force gauge was used, which had a maximum force
potential of 20 kN was used for these experiments. The data was recorded
using a ww700 transient recorder.

The hydraulic tensile machine differs from a normal tensile machine in
that there is a pre-displacement distance, which is set on the lower grip (Figure
9) before the system is activated. This pre—displacement distance enables the
machine to accelerate to the constant velocity chosen for testing. For example, if
you want to test the sample at a strain of 6.7 mps, the system cannot
automatically attain the ample speed of 6.7 mps without allowing an initial

displacement to achieve this velocity.

28

 

Figure 9: High strain tensile specimen.

The samples were cutout using a steel dog-bone stamp and hand press.
The dog-bone specimens were eight cm long and 1.5 cm wide, and had a gauge
length of 3.0 cm. Samples were tested at approximately 6.7, 4.1, and 1.4 mps.
Five samples each of the B regions of all eight processing groups were tested.
Since the material was very inhomogeneous it was very important to test the

samples in a way that was completely consistent, measuring only the material

29

behavior and reducing the interference from other effects. Figure 10 shows a

sample plot of the high strain tensile data.

 

 

 

 

 

I .__ L -
‘ 150 T“ — __.___. _~__
I 120
I .
90 I l _ _ _ _ ___.__-_
2 .
E Maxrmum
0 Force
II.
60 E- ___- ____ l _ _
30
o v T r T f f 7* T r—‘—"“1
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50
' Displacemenflnln)
L

 

 

 

 

 

 

 

Figure 10: Sample high strain tensile curve.

Using the time-temperature superposition principle, testing samples at

very high strains is representative to testing samples at very cold temperatures.

Testing samples at lower strains is representative of testing at very hot

temperatures. This is an important principal because this data can indicate

representative material behavior at different temperatures and speeds.

Impact/Penetration Machine

In order to determine the ability of the material to absorb energy under

impact conditions, impact-penetration tests were performed using the same

30

hydraulic testing machine used for the high strain tensile tests. The tests were
performed at various speeds, given in meters per second (mps) at the iKV at
RWTH - Aachen University using the machine and instrument set-up shown in
Figure 16. The auxiliary equipment used in the experiments included a water
cooling system, a ww700 transient recorder, a Kistler 5001 charge amplifier, and

Zwick REL control system.

 

Figure 11: Impact/Penetration Set-up.

31

The impact-penetration machine operates in the same manner described
for the high strain tests. The only difference were the set-up shown in Figure 16,
which instead of tensile grips included an impact mandrel and sample platform.
The impact mandrel was one cm in diameter and 10 cm long. The sample
platform had a diameter of 12.4 cm. The testing parameters also varied slightly
using the following parameters: a ramping wave function, and a negative
(compression) testing direction. The "active" button was also set, which
activates the machine to be able to test at the higher speeds. A Kistler 9331A
force gauge was used, which had a maximum force potential Of 20 kN was used
for these experiments. The data was recorded using a ww700 transient
recorder.

Similar to the high strain tensile tests, the impact-penetration tests also
had a pre-displacement distance, which was set on the lower sample platform
(Figure 12) before the system was activated. This pre-displacement distance

enabled the machine to accelerate to the constant velocity chosen for testing.

32

Lower
Sample .
Platform i

 

 

 

Figure 12: Impact/penetration sample set-up.

The B regions were used for testing samples. They were trimmed to a
dimension of 10 cm by two cm in order to fit into the sample platform. Samples
were tested at approximately 6.7, 4.1, and 1.3 mps. Five samples each of the B
regions of all eight processing groups were tested. Figure 13 shows a sample

impact/penetration curve.

 

 

 

 

 

 

 

 

 

 

 

 

 

Maximum
Force A
400 . -
A m _
.2.
O
2
I2
200 , _—~ __ _
100 .L__
01 7' i " “ ' r
o 1 2 3 4 5
Displacemnum

 

 

 

 

Figure 13: Sample Impact/penetration curve.

34

RESULTS and DISCUSSION

Thermal analysis of PP and PET EAM Fiber
DSC scans of the recycled EAM showed Tm of PP and PET to
be around 167°C and 254°C respectively. These values correlated with T...

values Obtained from the literature shown in Table 3.

 

 

 

1g Tm
PP -10 -18°C 176°C
PET 69°C 265°C

 

 

 

 

 

Table 3. Glass transition and melt temperatures of PP and PET [16].

Scans of PP fibers showed two melt temperatures of 167°C and 250°C
respectively. The cause of the minor second melt temperature shown in the
recycled PP scan is unknown and was not investigated further since the
constituent materials changed from one brand of recycled material to another
brand of recycled material. Scans of PET fibers showed a melt temperature of
254°C. All of the scans showed that the melt temperatures were consistent with
the individual materials and the finished product.

DSC results showed that the post-industrial individual PP and PET fibers
had melt temperatures around 163°C and 252°C respectively. The exothermic
peaks shown in the cooling curve of the graphs are indicative of crystallite
formation. Since both PP and PET are semi-crystalline materials both samples

will have some indication of crystallization.

35

in these scans, the sample T... is shown to decrease slightly after being tested a
second time. This lowering of the T... is possibly due to stress relaxation of the
PP and PET material. The DSC did not show the T9 for the PP material, which is
supposed to be visible around -20°C. Since the test was performed from
-30° to 300°C, this suggests that the Tg may be present in a range lower than
the range that was tested.
Optical Microscopy of the Failure Modes of EAM

To determine what type of failure each EAM element was doing while it
was impacted, impacted specimens were examined microscopically. It was
speculated that there was collapse in region A, buckling of the wall in region B,

and crushing in areas C and D (Figure 14).

. I
i, .. ,4. 0 .
.- , - I ,,
_ t a, i . 3
.. g. .. .
.‘ .
n4 ’ I', "’
.- .
_ a . _ .A a '
’1 .1 . , .. a , '- . [v
u. - I Avr-
b I ’ ‘
-' fit ‘ ' ' . u
., \_ .. . ‘ 4 . .
.. .. .‘ _. .
h 1 v I c

 

A

Figure 14: EAM element examined by regions.

Characterization of what the different regions of the element was very vital
to understand how the material performed at both a microscopic and

macroscopic level.

36

Failure modes of the EAM element could not be determined by optical
microscopy alOne. When recycled material EAM elements that were impacted
were examined under a microscope, "dice" looking objects were found scattered
throughout the material. it was speculated that they were cross-sectioned multi-
iobed fiber bundles. This was consistently seen throughout EAM samples that
were impacted or non-impacted.

in order to determine failure modes of the EAM element, non-impacted
elements had to be examined in order to see what the nominal amount of
damage was in an EAM element. Samples that were not impacted were
impregnated with fluorescent epoxy and mounted in the methods previously
described, and then examined under an Olympus BH2-UMP microscope. Non-
impacted samples showed little or no fluorescent epoxy anywhere within the
element shown in Figure 24. impacted samples impregnated and mounted in the
same way, however, showed that penetration of the fluorescent epoxy in high
concentrated areas correlated to damage areas within the EAM element shown
in Figure 25. By looking at the walls of the element, large concentrations of
fluorescent epoxy can be seen, leading to the conclusion that most of the force is
being absorbed by the walls of the element. By looking at micrographs, it was
evident that the tip of the EAM element remained intact with almost no
fluorescent epoxy showing, while the sides of the element displayed more break-
up during impact. Failure modes of the various EAM sections could not

determined at this time.

37

 

 

In addition to determining the order of failure within the EAM element,
consistency of the failure observations had to be verified. Each impacted and
non-impacted sample was cut in half and then examined to verify that the same
failure process occurred throughout the sample. The results were consistent
through the thickness with slight variations in the amount of breakup in the walls

of the element.

 

NON-iMPACTED' EAM ELEMENT
WITH FLUORESCENT EPOXY RESIN-DYE
(Magnification = 8.3X)

 

 

 

Figure 15: Non-impacted sample with little or no fluorescent epoxy in the
EAM element.

38

 

iMPACTED FIRST LAYER OF EAM ELEMENT
WITH FLUORESCENT EPOXY RESIN-DYE
(Magnification = 8.30

 

 

 

 

Figure 16: impacted EAM samples with damaged areas showing high
concentrations of fluorescent epoxy.

 

IMPACTED SECOND LAYER or EAM (ELEMENT
WITH FLUORESCENT EPOXY RESIN-DYE
(Magnification = 8.3X)

 

 

 

 

Figure 17: Second layer of EAM element showing consistent results from layer to
layer.

39

Processing and EAM

in order to determine how processing parameters affected the energy
absorbing properties of the EAM, samples were made at different processing
temperatures and densities. Unconsolidated mat samples having densities of
1000, 1200, and 1400 g/m2 were fabricated into consolidated panels at
processing temperatures of 193 -238°C in increments of 11°C. Density was used
as a variable instead of pressure since the press displacement was controlled to
a 1.5-mm gap. The temperature range used to evaluate EAM properties was
between the melting temperatures of PP and PET (T m of PP = 162°C, PET =
254°C). A total of 210 tensile tests were performed at 23°C and 50% relative
humidity. Tensile tests were done to measure material properties as functions of
modulus and adhesion between PP and PET. Samples were tested in both the
machine and cross-machine directions. Each testing group consisted of seven
samples (Table 4) with tests performed with an lnstron 4481 testing machine,
with a 2727 kg load cell. Samples were tested using ASTM Standard D638,
using a nominal strain rate of 50 mm/mm.

Modulus and total energy values extracted from the lnstron tensile test
data were plotted vs. temperature for both machine and cross-machine
directions. Figures 18 and 19 show the lnstron tensile data young's modulus and
total energy values of 1000, 1200, and 1400 glm2 density samples versus
temperature for machine directions respectively. Figures 20 and 21 show the

lnstron tensile data young's modulus and total energy values of 1000, 1200, and

40

1400 g/m2 density samples versus temperature for cross-machine directions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

respectively.
193°C 204°C 216° C 227° C 538°C
1000 glm’ 1 2 3 4 5
1200 glm’ 6 7 8 9 10
1400 glm2 11 12 13 14 15
193°C 204°C 216°C 227°C 238°C
1 200 glrn2 16 17 18
A = 30 A = 30 A = 30
sec. sec. sec.
B = 45 B = 45 B = 45
sec. sec. sec.
C = 60 C = 60 C = 60
sec. sec. sec.

 

 

 

 

 

 

 

Table 4: Matrix of Density vs.

tests.

Temperature tests and 1200 g/m2 density vs. time

41

 

 

 

YOUNG'S MODULUS vs. TEMPERATURE
for VARIOUS DENSITIES
(machine direction)

 

 

 

 

 

 

 

YOUNG'S MODULUS
(MP8)

 

 

 

.1000 g/m’
9 P P .0 P g > .12009/m'
('0 V (D B Q [U
22 a z; a: a 2.5 g Il4°°9’""
E '5
<
TEMPERATURE

 

 

Figure 18: lnstron Data Young’s Modulus Values of 1000, 1200, and 1400 glm2
Density samples vs. Temperature for machine direction.

 

TOTAL ENERGY vs. TEMPERATURE
for VARIOUS DENSITIES
(machine direction)

 

 

 

 

 

 

2
5
n: I
“é l [.100091m’
o o o o 0 Lu > f.12009/m’
:53 0% 2° 7‘ :9, 0 3 ’ 1400 Im1
.. N a a N g D g 9
'—
5 a)
TEMPERATURE

 

 

 

Figure 19 lnstron Data of Total Energy Values of1000, 1200, and 1400 glm2
Density Samples vs. Temperature for machine direction.

42

 

YOUNG'S MODULUS vs. TEMPERATURE

for VARIOUS DENSITIES
(c ross-direction)

 

 

 

 

U)
3
3 1000 -___- u... ___
8 A 800 .1 ,, —»
a g 600
E9 " 200 . 571
z 0 l
3 P P .0 .0 P w >
>- on g (D N co (D m
0) v- N (‘0 g Q
- N N N N o
g is
<
TEMPERATURE

 

 

.1000 glm’
.1200 g/m’
.1400 glm’

 

 

 

 

Figure 20: lnstron Data of Young’s Modulus Values of 1000, 1200, and 1400
glm2 Density Samples vs. Temperature for cross-machine direction.

 

TOTAL ENERGY vs. TEMPERATURE
for VARIOUS DENSITIES
(cross-direction)

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

S. l T 2:;- .1000 glm’
g I if — 1“"? e .1200 g/m”
g as; a It 1.1400 em
I“ .
o .0 P .0 .0 w '

... P.

> a)

<

TEMPERATURE

 

Figure 21: lnstron Data of Total Energy values of 1000, 1200, and 1400 g/m2
Density samples vs. Temperature for cross-machine direction.

43

 

To try and gain an understanding of the information the results were
giving, a pseudo energy was calculated using the formula W = O’yZIZE in order to

and normalize the data. This calculation was done by making the assumption
that plastic deformation after yield is a function of the sample preparation,
defects, etc . The sample behavior prior to the yield point, is the work of
deformation area under the curve, would provide a method to normalize the data
and show a dependence on the morphology of the part and not the quality of the
part itself. Looking at these pseudo energies improved the scatter, but no

correlation could be identified in Figures 23 and 24.

 

PSEUDO ENERGY (MPa) vs. TEMPERATURE FOR
ALL 3 DENSITIES
(machine direction)

 

 

 

 

 

 

 

 

 

 

o 0.800
t:

g 0'6“) l.1000 g/m’
8 0.400 .1200 glm1
a 0.200 .1400 glm’
U)

‘L 0.000

 

 

193°C 204°C 216°C 227°C 238°C
TEMPERATURE

 

 

 

Figure 23: Pseudo-Energy Values of 1000, 1200, and 1400 g/m2 Density
Samples vs. Temperature for machine direction.

44

 

PSEUDO ENERGY (MPa) vs. TEMPERATURE FOR ALL
THREE DENSITIES
(cross-direction)

 

 

 

>- 0.800
g .1000 g/m2
I.” 0.600 2
z .1200 glm
‘” 0400 2
O - .1400 glm
8 0.200

D

[2 0.000

 

193°C 204°C 216°C 227°C 238°C
TEM PERATURE

 

 

 

Figure 23: Pseudo-Energy Values of 1000, 1200, and 1400 g/m2 Density
Samples vs. Temperature for cross-machine direction.

Since little information was gained from examining the pseudo energies,
there was a possibility that there was a material or process uncontrolled variable.
The weights of each of the samples were measured and then examined for
uniformity. Average weight vs. temperature plots for both machine and cross-

machine directions were made for all three densities shown in Figures 24-26.

45

 

1400 glm2 AVERAGE WEIGHT vs. TEMPERATURE

 

     

is?
2 8.000
9? 6.000 machinfltsctjv ., °r9,5.sdirssti°_n_ _ 1
E ' .sro. DEV.
”00 ~" _ —3 AVERAGE

“o‘ . 1;.—
g 2.000 . ,- -
Ill
3 0.000 , _ r g

0 O U U 0 O O O O O

:4 ‘3 28 IL to so a :8 IL 3:.

0') ‘— N ('3 O) O 1- N ('3

v— N N N N N N N N

TEMPERATURE

 

 

 

Figure 24: 1400 g/m2 Density Average Weight vs. Temperature plot.

 

1200 glma AVERAGE WEIGHT vs.
TEMPERATURE

 

 

g 6.000
machinedireclion C'OSSdI'BCIIOI’i

g 5.000 ,. 7., 7- -
. 4.000
E 3000 .STD.DEV.‘
‘0’: 200° .AVERAGE'
§ 1.000
E 0.000

0 O 0 U 0 U 0 O O O

to 1: 20 is So in 1: in I" 30

O) O 1— N (‘0 0') O 1— N ('7

P N N N N F N N N N

TEMPERATURE

 

 

 

Figure 25: 1200 g/m2 Density Average Weight vs. Temperature plot.

46

 

1000 glm' AVERAGE WEIGHT vs. TEM PERATURE

6000 ,_ w. _.

machine direction croudiroclion

 

 

5000
4000
3000
2000
1000
0000

 

 

 

 

.STD. DEV.
.AVERAGE

 

 

 

AVERAGE WT. (grams

 

 

F

TEMPERATURE

 

 

 

Figure 26: 1000 glm2 Density Average Weight vs. Temperature plot.

It should be noted here that samples fabricated at 193°C and 204°C
showed large amounts of shrinkage. As the samples were heated to these
temperatures and then removed from the heater, considerable shrinkage of the
material occurred, thus increasing the density of the samples. Samples that
were supposed to be at approximately 1.5-mm thick ended up being as thick as
3.8 mm. Samples fabricated at 216°C and higher showed little or no shrinkage
and maintained sample thicknesses around 1.5-mm. Thicker sample weighed
more.

From looking at these graphs, parallel results were seen for both the 1000
and 1400 glm2 samples. However, the 1200 glm2 results had a large amount of

scatter. The 1200 glm2 samples made at 193°, 216°, and 238°C were made by

47

two different operators, and led to the conclusion that properties of samples are
operator dependent.

After examining these results of the 1200 glm2 samples, another set of
experiments using only the 1200 glm2 density mat was done. This time the
sample weights were measured at temperatures, 193° - 216°C in increments of
11°C, for different periods of time in the tool, 30—60 seconds in increments of 15
seconds. Average weight of samples fabricated at 193°, 204°, and 216°C vs.
time and pseudo energies of samples fabricated at 193°, 204°, and 216°C vs.
time were plotted as Figures 27 and 28. The average weight vs. time plots
shows that samples fabricated at 193°C showed the most amount of scatter with
time in the mold. The average weight of the samples also increased as the time
in the mold is also increased. Samples made at 204°C and 216°C were more
consistent with one another having smaller variability than the samples fabricated
at 193°C. This is probably due to decreased shrinkage of the material as the
temperature reaches the melting temperature of the PP. The average weight of
samples fabricated at 204°C showed a slight increased with time in the mold,
while the average weights of samples at 216°C showed a slight increase and
then decrease with time in the mold. These changes for the 204°C and 216°C
samples, however, were within the standard deviation. The pseudo-energy vs.
time plot in Figure 28 showed an increase in pseudo energy for the 204°C
samples as the time increased, while the samples as 216°C showed a decrease

in pseudo energy as time increased. The samples at 193°C showed the most

48

variability with increased time in the mold. The standard deviations were large,

however, making comparisons difficult.

 

AVERAGE WEIGHT vs. TIME for 1200 glm2

'- 8.00
5 7.00
m ... 6.00
a g 5.00
m a 4.00
o 5, 3.00
g ... 2.00
g 1.00
< 0.00

 

TIME (sec.), AVERAGE, STD. DEV.

 

 

 

Figure 27: 1200 g/m2 Average Weight vs. Time plot.

 

PSEUDO ENERGIES FOR VARIOUS TEMPERATURES
vs. TIME, AVG., STD. DEV. for 1200 glm’

 

 

PSUEDO ENERGY

30 45 60 AVERAGE STD. DEV
TIME (sec.), AVERAGE, STD. DEV.

 

 

 

Figure 28: 1200 g/m2 Pseudo-Energy vs. Time plot.

49

Dyna-tap Characterization of the EAM
Dyna-tup experiments were conducted on a sample of foam and a
sample of EAM with backing and without backing. The foam sample was
rectangular in shape and approximately four by six inches. The foam sample
was impacted at a velocity of 10.4 km/hr. The impact energy was 20.4J and the
total energy was around 16.3 J.

A sample EAM part that was approximately six by eight inches was
impacted with backing facing the impactor. This was done to simulate HIC
testing in a vehicle. The flat backing facing the impactor was ruptured during
testing. The peaks, however, were deformed but not ruptured. Data showed at
an impact velocity of 4.6 kahr produced an impact energy of 20.4 J and total
energy of 21.1 J. The maximum load was recorded to be around 2.7 kN.

A sample EAM part that was approximately six by eight inches was
impacted without backing, thus the peaks were facing the impactor. Data in
Figure 41 showed at an impact velocity of 10.4 km/hr produced an impact energy
of 19.8 J and total energy of 19.9 J. The maximum load was recorded to be
around 9.1 kN.

When the results of the-foam impact test were compared with the sample
EAM part that was impacted with backing facing the impactor, it was seen that '
the curves did not resemble one another since the EAM sample was damaged.

When comparing the load deflection curves of the EAM sample impacted

with and without backing, the resulting graphs looked different, again due to the

50

sample with backing being damaged. However, the maximum load was much
higher for the sample impacted without backing (9.1 kN compared to 2.7 kN).
DMA Analysis of the EAM Part I

It was also speculated that material properties varied as a function of
thickness. Measurements of various locations in the EAM element cross-section
were taken from previous micrographs of parts processed. In areas A and C, the
thickness was 1.5-mm, area B was 1.2-mm, and the apex was .65-mm thick.

The DMA was chosen to examine the EAM, because it would be able to
test relatively small samples (10 x 25 x 7 mm) as functions of both temperature
and frequency. Both stiffness values and levels of energy absorption could be
examined versus temperature and frequency.

Table 5 lists the parameters that were used when testing the various
thicknesses in the DMA. The thicknesses correlated to the corresponding areas
within the EAM element as a function of frequency. The frequencies were
chosen to see how the stiffness and energy absorption varied with frequencies
equivalent to velocities at impact. All samples were tested from -30° to 100°C in
order to see how the material behaved in both a cold and hot environment. The
samples were tested only up to 100°C to prevent melting of the sample. Samples

of each thickness were tested from highest frequency to lowest frequency.

SI

 

 

 

 

 

 

 

 

 

 

 

 

VARIOUS AUTO-STRAIN VALUES (%)
Frequency (Hz) 1 .5-mm 1 .2-mm .65-mm
150 170 170 160
15 160 140 140
1 .5 150 140 140

 

 

 

 

Table 5: DMA parameters for various thickness tests.

Table 6 summarizes the initial storage modulus values and inflection

temperatures. These parameters were examined to determine the sensitivity of

the stiffness and T9 (material properties) to temperature and frequency. The

same sample was used for all three different frequency trials.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

THICKNESS (mm)

FREQUENCY (Hz) 1.5 1.2 0.65
1.5

Initial E' (MPa) 3000 3100 4450

lnflection Temp. (°C) 5.66 3.74 1.22
15

Initial E ' (MPa) 2750 3400 4500

lnflection Temp. (°C) 6.31 8.04 6.14
150

Initial E ' (MPa) 2250 3900 4750

lnflection Temp. (°C) 14.22 11.53 7.81

 

 

Table 6: Summary of Initial Storage Modulus (E') and lnflection Temperatures
(T9) for various thicknesses and frequencies.

After this testing, four more samples at 1.2-mm thickness at 1.5 Hz, and

140% auto-strain, 2 N static load, and 450m frequency amplitude were tested for

52

repeatability. Tests were run with different samples to check for repeatability.
The initial storage modulus values were plotted and then compared. Table 7
gives a summary of these values. The results showed that the tests and results

were repeatable having a variability around 10%.

 

Initial E' LMPa) for 1.2-mm Samples
2802
3320
2966
2658
3254

 

 

 

 

 

 

 

 

Table 7: Summary of Storage Modulus Values for 1.2-mm thick samples tested
from -30 to 100°C at 1.5 Hz.

Results from the DMA run at three different frequencies showed that as
the thickness decreased, the storage modulus values increased. This result
indicated that the energy absorption characteristics would change depending on
the location (and thickness) of the EAM in the cross-section. The inflection
temperatures also showed some scatter of a few degrees. Since the
temperatures are usually associated with intrinsic material properties, the
differences detected with thickness may be due to the physical entanglements
between fibers in the EAM. Sample DMA graphs for a 1.5-mm thick sample
tested from -30° to 100°C at 150, 15, and 1.5 Hz are shown. The sample run at
1.5 Hz had the highest initial E' while the sample run at 150 Hz had the lowest E'
value. The inflection temperatures also decreased as the frequency decreased.

The data taken at 15 and 1.5 Hz showed the clearest results.

53

After examining the data at various frequencies, it was decided that testing
should be done at 150 Hz from -30° to 100°C in order to simulate results similar
to HIC testing. The total cycle time for an impact of HlC is around six
milliseconds, which is equivalent to approximately 167 Hz. Samples of each
thickness were tested three times. After each sample was run once in the DMA,
the sample was cooled to room temperature and the DMA experiment repeated.
The first group of samples that were tested at 150 Hz showed that the initial E'
values and inflection temperatures changed from run to run even though the
same sample was tested three times consecutively.

The initial storage modulus values are listed in Table 8 for all three

thicknesses.

1“t Run 2"d Run 3RD Run

 

Table 8: Summary of Initial Storage Modulus for various thicknesses run
intermittently at 150 Hz.

This data shows that the initial storage modulus changes as a function of
thickness. Since the material was allowed to cool to room temperature, the

material could be experiencing residual stresses or different degrees of

54

crystallinity. However, the storage modulus for each sample increases as a
result of the first thermal cycle and then remains the same after the second
thermal cycle.

Recalling the fracture behavior of the EAM element (Figure 22), region B
(1 .2-mm) fractures, but regions A and C (1 .5-mm) and region D (.65-mm)
remained completely intact. By comparing the E' data from Table 8, the1.5-mm
and 1.2-mm values correspond very closely to one another during the first run
with the .65-mm values being the lowest. The 1.2-mm data has the highest E'
values during the second and third runs with the 1.5-mm and .65-mm values
close to one another. These trends do not reflect what we would have expected -
the .65-mm samples having the highest E' values with the 1.5-mm and then 1.2-
mm samples having lower values respectively. Sinoe the .65-mm have
experienced the greatest degree of compaction, we would have expected them
to have the highest values with the 1.5-mm having the lowest since they are the
thickest. This led us to speculate that there may be some other factors within the
material affecting its performance such as voids.

A second group of samples were tested at 150 Hz for all three
thicknesses, but this time the samples were run consecutively three times in a
row without being cooled to room temperature for an extended period of time and

then restarted. A summary of these values is shown in Table 9.

55

1't Run 2'“l Run 3"D Run

 

Table 9: Summary of Initial Storage Modulus (E') for various thicknesses run
consecutively at 150 Hz.

The second group of samples showed that the storage modulus values
were very consistent from run to run with only the first run being lower than the
other two. This was consistent for all three thicknesses, which led to the
possibility that there is not a difference in the material properties (storage
modulus and inflection temperature) with thickness. Since runs two and three
were cooled at the same cooling rate they end up having the same crystalline
structure. The first DMA run would therefore be influenced by residual and
crystallization stresses, and perhaps physical entanglements between fibers.

Again, recalling the fracture behavior of the EAM element (Figure 14) -
region B (1 .2-mm) fractures, with regions A and C (1.5-mm) and then region D
(.65-mm) intact. By comparing the E' data from Table 9, the 1.5-mm data has the

highest E' values during all three runs, with the 1.2-mm and .65-mm being lower.

56

Therefore the way in which the samples are tested does affect the results.
Samples that were allowed to completely cool between thermal runs showed
different E' values in all three runs for all three thicknesses. Samples not allowed
to completely cool between runs showed that only the first E' for the first run was '
lower than the second and third runs for all three thicknesses. it should be noted
that due to these results, this “annealing” process (heating and cooling the
sample) may improve energy absorption of the EAM to achieve higher E’ results.
Voids and EAM

The DMA data showed that as the thickness of the samples decreased,
the storage, loss and tan delta values increased, leading one to believe that the
.65-mm thickness samples were stiffer due to having less voids.

To test this hypothesis, void content was measured using the ESEM.
Samples were polished and mounted as described earlier and then mounted in
the ESEM. By looking at representation ESEM micrographs of 1.5, 1.2, and
.65-mm compressed samples (Figure 29), it appears that the 1.5-mm specimens
show a larger and higher percentages of voids than the 1.2 and .65-mm samples.
The 1.2-mm samples showed that voids were evident but not as large or
prevalent as those seen in the 1.5-mm samples. The .65-mm samples showed
very few, and very tiny voids. The fluorescent impregnated samples examined
on the Olympus BH2-UMP microscope showed similar results and verified that

that void content does decrease with thickness.

57

 

Figure 29: ESEM micrographs of 1.5, 1.2, and .65-mm samples and their void
contents.

In order to try and quantify the amount of voids as a function of thickness,
Archimede's Principle was used. The first group of samples was taken from the
edges of the compressed sheets for all three thicknesses. This was done in
order to consistently obtain samples that were 1.5, 1.2, and .65-mm in thickness,
since metal shims of these thicknesses were used at the edges of the sheets.
First all samples were measured in both air and hexane using a scale. The

density of the sample (p3) was then calculated by using the formula:

P: = (W: Pl ‘ WI*Pa)/(Wa ' WI)

58

Where W3 is the weight of the sample in air, p. is the density of hexane, w.
is the weight of the sample in hexane, pa is the density of air. The void
percentage was then calculated by using the formula:

Void percentage = pa-ps/pa

The results are summarized in Table 10.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Thickness .65-mm 1.2-mm 1.5-mm

Voids Voids V_oiglg

0.07 0.08 0.07

0.07 0.11 0.11

0.08 0.16 0.08

0.06 0.09 0.09

0.06 0.15 0.10

0.08 0.07 0.09

0.09 0.08 0.10

0.08 0.10 0.10

0.07 0.07 0.07

0.08 0.07 0.09

0.11 0.10 0.10

0.11 0.11 0.09

Average 7.9% 9.8% 9.0%
Std. Dev. 1.7% 2.9% 1.4%

 

 

Table 10: Void data of 1.5, 1.2, and .65-mm thickness samples taken from the
edges of compressed sheets.

In a second experiment, the same procedure was done but samples were
taken from the center of the sheets rather than the edge of the sheets. It was
found that the center samples were thinner than the prescribed 1.5, 1.2, and .65-

mm thicknesses. The results of this experiment are shown in Table 11.

59

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mg .65-mm 1.2-mm 1.5-mm
Voids Voids Voids
0.14 0.13 0.13
0.13 0.13 0.13
0.14 0.13 0.13
0.12 0.13 0.13
0.13 0.15 0.13
0.12 0.13 0.12
0.13 0.13 0.13
0.14 0.13 0.14
0.13 0.16 0.14
0.14 0.13 0.12
0.14 0.13 0.13
0.10 0.13 0.12
Average 13.0% 13.5% 13.0%
Std. Dev. 1.2% 1.1% 0.7%

 

 

 

 

 

 

 

Table 11: Void data of 1.5, 1.2, and .65-mm thickness samples taken from the
center of compressed sheets.

 

 

 

 

 

 

 

 

 

 

 

 

 

1 .5-mm 1 .2-mm .65-mm
7.9% 9.8% 9.0%
Edges Voids Voids Voids
13.0% 13.5% 13.0%
Voids Voids Voids
Center

 

 

 

 

 

 

 

 

 

Figure 30: Summary of void data by thickness and location.

60

By examining both sets of void data, the results were as follows:

1. Samples taken at the edges of the sheets were more exact in the thickness
dimensions due to the shims used when fabricating the sheets.

2. Samples taken from the center were much thinner in the center sections,
seen mostly in the 1.2 and 1.5-mm samples due to not having shims near the
center of the sheets when they were fabricated.

3. All of the void content mean and standard deviation values overlapped for all
three thicknesses and contained a great deal of scatter.

4. The .65-mm samples had the most consistent void content values in both
studies done: mean thickness values were pretty close compared to the 1.2
and 1.5-mm samples.

DMA Analysis of the EAM Part II

The DMA was used once again to examine fabricated EAM parts at
various processing conditions. Samples needed to be tested in both the machine
and cross-machine directions to see if there were differences in stiffness
between the machine and cross-machine directions. The DMA was also used to
determine what samples of what sections of the various processed EAM were to
be tested for the high strain tensile and impact/penetration tests. Samples were
fabricated on a large, scaled up mold and were processed (Figure 4) in order to

get sample sections large enough for testing.

6]

 

Figure 4: Drawing of enlarged EAM mold.

Parameters were selected to see how the effects of time in the heater,
time in the mold, temperature, and pressure (taken here as displacement of the
mold) affected the material when processed. The baseline samples processed at
the plant are fabricated at 216°C for approximately 60 seconds in the heater and

45 seconds in the mold to a set displacement of 1.5-mm. The various

62

processing parameters dealing with times, temperatures, and pressures are
shown in Table 12. They were chosen both above and below production
processing parameters. Increasing the time in the heater should create more
flow of the PP and therefore increase mechanical adhesion. Increasing the time
in the mold would create a higher degree of consolidated material and therefore
increase adhesion between PP and PET. Increasing the processing temperature
should again create better flow of the PP to surround the PET fibers, thus
increasing mechanical adhesion. Increasing pressure should increase the level
of consolidation and reduce the presence of voids, creating a stiffer material. It
should be noted that 1A, 3B, and 4A are the same sample since the processing

parameters are identical.

 

 

 

 

 

 

 

 

 

1 2 3 4
TIME 1 TIME 2 TEMP PRESSURE
(HEATER sec.) (MOLD sec.) (°C) (DISP mm)
A 60*
TIME 1 B 90 60 60 60
(HEATER sec.)
C 120
TIME 2 45 A 30 45 45
(MOLD sec.)
B 60
B 216*
TEMP (°C) 216 216 C 227 216
A 238
PRESSURE 1.5 1.5 1.5 A 1.5*
(DISP mm)
B 1.25

 

 

 

 

 

 

 

 

Table 12: Processing parameters of the EAM

63

Three samples in both machine and cross-machine direction were tested from
each region (A, B, C). Samples were approximately 12 x 11.5 x 1.5-mm and
were tested at a frequency of one Hz from —100°C to 140°C. 5’ results of the
machine and cross-machine data at -100°C are shown in Tables 13-14

respectively.

64

 

Table 13: E' results of DMA machine direction samples at -100°C.

65

 

Table 14: E' DMA results in cross-machine direction at -100°C.

66

Since results were not very conclusive at -100°C, E' data at room

temperature (approximately 20°C) were extracted from the DMA thermal graphs.

The results are shown in Tables 15-16.

 

Table 15: E' DMA results of machine direction samples at 20°C.

67

 

Table 16: E' DMA results of cross-machine direction samples at 20°C.

68

The results are listed according to processing parameter - time in the heater,
time in the mold, temperature and pressure.

The results were then plotted for each temperature for both machine and
cross-machine directions for regions A, B, and C. The results show the average
and standard deviations for each of the processing groups in all three regions

(Figures 31-34).

 

E' Values @-100°C for Machine Direction

 

 

 

 

 

 

 

 

 

 

 

 

 

200000 -_ . .z~________”224.,4._, - _ ____-
A B V c 1r
1750.00 -___________. _ _ . -,
1r 1350 I 1'1“: 0
1500.00 .______-_-. { “so, 1513 g I 151:
l ; I 01:73
.. 1253
a , -______._______
“- 125000 1027 I l 130:
‘9 '53 1007 -r ‘I
~— E_§l'2°fi u: T I 01007
§ 1000.00 - . i f E
0 031 047
‘3 990 $ 0003
> 740 I -- fi 000 ..
in ..

 

 

 

 

 

750 .00 A

Im in? ._ T'"

500.00 ‘

 

 

 

 

250.00

 

 

 

 

 

 

fl- -.

0.00 . . - - _-.- . . , , ,_.__,
18 1c 2A 23 33 3c 3A 43 13 1c 2A 23 an ac 3A 43 13 to 2A 28 33 3c 3A 43

 

Sample ID.

 

 

 

Figure 31: E' Graph for machine direction samples at -100°C.

69

 

E' Values 9 20°C for Machine Direction

 

 

2000.0 v—— - —

17500 <»——-—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

15000 .»_._--
.3. 1250.0
3 002
£10000+— --___. -_ l
'5
> I
I“ 750 0 L I 000 0:14
‘ r 507
023 57. ‘ 500
500.0 7 --.-_._._- - H »_ .- f
‘33 417 400
373 £370 5:10
250.0 , - .

 

 

 

0.0 4— -

18 1C 2A 28 3B 30 3A 48

 

.7...

13 1c 2A 23 33 3c
Sample ID.

7r, __

 

 

 

 

 

 

 

707
004 "3
750 500 q
A «I
500 f 750 ”II
510
'I L
500 5" I
3A 43 13 1c 2A 23 :13 0c 0A 43

 

Figure 32: E' Graph for machine direction samples at 20°C.

 

E' Values @-1 00°C for Cross-Machine Direction

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. '—

 

18 1C 2A 28 3B 36 3A 43

.T_.

 

'1'

250 7

20000 .F_-____.__ -_ z - -

1750.0 4 - _

1500.0 _.
A 1250.0 <Ir v____
E. 0

1000.0 _ .
3 950.74}
C
> A.

10700
I” 7500 .- _ ._
500.3 501 7
§ b 570.0
500.0 - -- - , -W
' ' 430.7
3” 3 f {500.0 335.3
250.0 -.___---- - -______. _ --
41 —}
. 213.3
0.01—_-, ..-,,___,_ __

 

 

 

firm?“
i .3007 3

324.0

 

 

 

 

 

-r
., 1023.3
(>040 3 1r
755° 030.0
4} 551 7 .. § I
§ 401.7 570.0

 

 

 

w—v

T

jifi

18 1C 2A 23 38 3C 3A 40

Sample ID.

18 16 2A 28 38 SC 3A 48

 

Figure 33: E' Graph for cross-machine direction samples at -100°C.

70

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E' Values Q 20°C for Cross-Machine Direction
2000.0T .fl
1750.0 -- A ._ C
1500.0 HF_.__- “—3" _
S. 12500
g 1000.0 «-
§ 563
in 7500 -=
T T 540
500.0 1 .L 3$457
1‘ 317 222 231 ‘272 i Q
250.0 . i . i .
310 2:14 173 133 331
+133 225 .1 “3 Q 173 107 § 255 350
00...-.fi... ...f--.fi..- W
13 1c 2A 23 33 ac 3A 43 13 1c 2A 23 33 ac 3A 43 13 1c 2A 23 33 3c 3A 43
Samplel.D.

 

Figure 34: E' Graph for cross-machine direction samples at 20°C.

Examination of the data in Figures 54-57, show trends between the -

100°C and 20°C data. The -100°C data has higher E' values than the 20°C data.

The machine direction data also has higher E' values than the cross—machine

data. On average, the B regions of the cross-machine data have lower E' values

than the A and C regions. This correlation is not seen in the machine direction

data.

To determine the effect of processing conditions, morphology, and fiber to

fiber adhesion, the results were then graphed by processing condition: time in the

heater, time in the mold, temperature, and pressure. Mean and sample

71

 

deviations of all three processing parameters (i.e. time in the heater - 60, 90, and
120 seconds) for cross-machine at -100°C and 20°C and machine direction at -

100°C and 20°C for all three regions A, B, and C were plotted (Figures 35-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

38).
SmmaryofProcesParanaer-Tine1 (l-hater)
zmo ‘ e nag-1m T “— ‘ '
13QO
eugawc
17533 ~ # _ _
1530.0 _. _ - C
E 12510 .. —.
i 1337 an
n I I” 1*
I" «no.0 1120 ft..—
3 I «in
i I $330 r 588 an
;... .. ..
323 333 T am I
w Hf . .2. m . n
5000 A I 370 p ‘37 417 t Y ’45, ‘ j
4 as ; 579
2500* 342 ; {310 9m f T T £ -- 3m 5'6
' 242 . ‘ m 81 1
213 133
9133 .115 29 9
0.0 vvvvvvvvvvvvvvvv “MW—ff j,,,,,rYVT
33131033131c13131c23131c 33131cze131c33131caa131c £131C33131C33131ca131c
WELD.

 

 

 

Figure 35: Initial modulus values for cross-machine and machine direction EAM
as a function of time in the heater.

Figure 35 shows the results of time in the heater - 60, 90, and 120
seconds for cross-machine direction at 100°C ‘and then 20°C and machine
direction at 100°C. This is done for all three regions A, B, and C. It was
assumed that as the time in the heater increased, the E' values would also

increase since the PP material is allowed to flow more around the PET fibers,

72

increasing mechanical adhesion. This, however, was not always the case. As
the time increased to 120 seconds in the heater, the E’ values decreased for both
cross-machine and machine directions at both -100°C and 20° for regions A, B,
and C except for the cross-machine directions at both -100°C and 20°C for
regions B, and C, and machine direction at -100°C in region C. By looking at
Figure 58, one can also see that the E' values at -100°C have higher values than
at 20°C for all three regions (A, B, C). Region C E' values at -100°C have the
highest values, while E' values for region B at 20°C have the lowest values. This
is due to the material being colder, and therefore having a more brittle, stiffer
modulus. The standard deviations for the E' data at -100°C is also much higher
than those at 20°C. By grouping all of the data in each region by temperature
and direction, we see that the data at 20°C for both machine and cross-machine
directions appear the same with only deviations in size due to standard

deviations from the data.

73

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SumaryomeoessPaarHer-Tinazmbld)
mo- snag-1am
2mm
magma-c
17510
Bum B
15110
A I“? C 01373
Ema
a I
am pm - w" b "m
3 TN a? 0947 u
i In 3
> 7510 T . T
" 573 a I 1
553 9533 f I emf ant
f ‘3‘ Q l 5m 0 090 m
5110 a). tan 417 l
Ema—r f f .m " ‘ l
. {m as 9... .. w
0.0 y Y . . . , . r 7 T . Y , . v 1 . . . , 7 . . ............
2A$B2A$ZBZRIBEZAIBB 2A382A$282A$82A$8 fiSBflEEI‘SBZASB
mm

 

 

 

Figure 36: Initial modulus values for cross-machine and machine EAM as a
function of time in the mold.

Figure 36 show the results of time in the mold — 30, 45, and 60 seconds
for cross-machine at —100°C and 20°C and machine direction at —100°C and
20°C . It was assumed that as the time in the mold increased, the E' values
would also increase since the longer the material was kept in the mold, the
molten PP would flow to a greater degree. In region A, all E’ values increase
with time in the mold for both cross-machine'and machine directions at —100°C
and 20°C. For section B, only the cross-machine data at -100°C and 20°C
showed an increase in E’ values with increased time. The machine direction

.samples showed E’ values decreasing at 60 seconds. For region C, all E' values

74

at -100°C have higher values than at 20°C for all three regions (A, B, C). Region
C E' values at -100°C have the highest values, while E' values for region B at
20°C have the lowest values. This again is due to the material being more brittle
since it is colder. The standard deviations for the E' data at -100°C is also much
higher than those at 20°C, except for region B at 20°C. By grouping all of the
data in each region by temperature and direction, we see that the data at 20°C
for both machine and cross-machine directions, again, appear the same with only

deviations in size due to standard deviations from the data.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

zoooo «—
Burg-1m
sugar-c
17500 +4.11%
eugzrc . T
15am ~— —-————— A [
A B 1253 C 01373
512500 - - 4p—-—~~— ~— —
E l ‘
a 10000 T i‘m 1cm * 01m?
3 W I ~-—— * 9.7 a
753
5 960 I a” J .. §
in ma — — —~-— ' a 0
t” a?” 169’
(l ‘ T I l 5570 s79
5:110 - ..H L
m a 458 .. . a
{ .L {373 1 f4” 1
I i 5307 1272 55° 6 '3‘"
2500‘ — .213 o i r L
242 9‘33 0115 ‘73 292 f1:33
00 V Y 1 1 Y Y T V Y T .V f Y Y r—T—iv fl 1' W 1 V f Y 1% T r V T Y 1' f—‘l
wxaaxawxmwxa wxmwxuwxawxa sxuaxmwxasxm
WELD.

 

 

 

Figure 37: Initial modulus values for cross-machine and machine EAM as a
function of temperature.

75

Figure 37 shows the results of temperature - 216°C, 227°C, and 238°C for
cross-machine direction at —100°C and 20°C and machine direction at —100°C
and 20°C . It was assumed that as the temperature was increased, the E' values
would also increase since the PP material flows more and increases mechanical
adhesion. In region A, all E’ values increase with time in the mold for both cross-
machine and machine directions at —100°C and 20°C. For section B, only the
cross-machine data at -100°C and 20° showed an increase in E’ values with
increased temperature. The machine direction samples showed E’ values
decreasing at 238°C. For region C, all E’ values decreased at 238°C. By looking
at Figure 60, one can see that the E' values at -100°C have higher values than at
20°C for all three regions (A, B, C). Region C E' values at -100°C have the
highest values, while E' values for region B at 20°C have the lowest values.
Again this is seen due to the material being cold and thus being brittle. The
standard deviations for the E' data at -100°C is also much higher than those at
20°C, except for region B at 20°C. By grouping all of the data in each region by
temperature and direction, we see that the data at 20°C for both machine and
cross-machine directions, again, appear the same with only deviations in size

due to standard deviations from the data.

76

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ 'IIXW“

fl

 

 

 

 

SimmdnooasPaam-Prwe
mo
snag-«Ire .,
emgrc
175nm Bum
"mm
B T c
13:10 — - --
.. (M3100 "
é 12m
. m3
mo 4)
a mo 0 ., }m:n FL
2' I m 322*
i): 730 ' - fl
m Tm
mo :25 $ t slul T 4) J, (501 ‘53:.
I 1 .. *
. 13730 $350 fun ‘31]
31°F “ 9 ' 1
241.7 . ma 2‘33 mi W0 {1&7
oo.,.......,fi-....3..fifi..
3334333343 33333333 333493333
SHIELD.

 

 

 

Figure 38: Initial modulus values for cross-machine and machine EAM as a
function of pressure.

Figure 38 shows the effects of pressure (displacement of the mold) - 1.5-
mm and 1.25 mm of displacement for cross-machine direction at -100°C and
20°C and machine direction at —100°C and 20°C. It was assumed that as the
pressure increased, the E' values would also increase by diminishing the
presence of voids and therefore creating a stronger material. In all cases, the E’
values increased with increased pressure for all directions, temperatures, and
regions. By looking at Figure 38, one can see that the E' values at -100°C have

higher values than at 20°C for all three regions (A, B, C). Region C E' values at -

77

 

100°C have the highest values, while E’ values for region B at 20°C have the
lowest values. This due to the material being cold and thus brittle. The standard
deviations for the E' data at -100°C is also much higher than those at 20°C. By
grouping all of the data in each region by temperature and direction, we see that
the data at 20°C for both machine and cross-machine directions, again, appear
the same with only deviations in size due to standard deviations from the data.

By looking at the graphs, one can see that the data looks similar from
region to region. All E' values at -100°C are higher than at 20° for all 3 regions A,
B, and C. E' values at -100°C for region C seems to have the highest values
overall, while region B E' values at 20°C have the lowest values. Data at -100°C
have higher standard deviations than 20°C data. Regions A and C do not differ
greatly in appearance to region B at 20°, except the amount of standard
deviation.

The effect of increasing the time in the heater increased the scatter in the
B and C regions. Increasing the time in the mold and the temperature seemed to
have similar effects on the A, B, and C regions for cross-machine and machine
directions at both -100°C and 20°C. It could also be due to the material being
able to cool faster near the bottom of the mold. The effect of pressure increases
the E’ values for all directions, temperatures, and regions.
Comparability Between Small and Enlarged Samples

The E’ from the B sections at 20°C were extracted and compared to E’
values taken from B sections of actual EAM production samples tested at room

temperature. It was speculated that the two different sizes of samples would

78

 

possibly have different material properties since they underwent different
stresses while being formed in the mold. The results are shown in Table 16.
Table 16 shows that E’ values cover a wide range of values for both the large
and small molded materials, but they are comparable, leading one to believe that
the material properties are similar in both B regions for both the small production

samples and the enlarged EAM samples. ll

 

 

 

 

 

 

 

 

 

 

 

E’ MPa E’ MPa

(production EAM) (enlarged EAM) ‘
650 417 f
680 599 '3,
600 498 '

950 597

525 902

1250 758

569

580

Avg. = 775.8 Avg. = 615.0

 

 

 

 

Table 16: Comparison of E’ values from production and enlarged EAM samples.

Ejgm Ta_n 6 Comgan'sons

In addition to comparing the E’ values, E" values were extracted from the
DMA thermal graphs from the enlarged samples and plotted to see if any trends
were seen compared to the high strain and impact lpenetration energy results.
E” represents the loss modulus or the amount of energy absorption. Looking at
the E” values may be an indicator of energy absorption. The results for the E” for
machine and cross-machine directions at —100°C and 20°C are shown in

Figures 39-42.

79

 

 

m.m~r——-—+—--- __'- —— B C
60.00 «P *7 7 W ._ __ __ __ _
1?
5
~— I” h“ 1! I
g noon--
- T .7 a.
g 33
I'll 31{32 $31 27 31
, f3 2” i 93 a
20m l ‘21 ‘21 0
' l J.
l
0'm1 7‘- V fi' f v , . v v .33., *fi
13 1C 2A 28 38 x: 3A 4B

 

EWanflHNHHUthMHfiuflm

 

 

 

 

 

 

 

 

 

 

 

 

181C2A2838$3A48

iaiczAzeaaacaAaa
PfiKNCESEflIKBCfliOlINB

 

 

Figure 39: E” values for machine direction at -100°C.

E" Value. (UPI)

 

 

E'ValuooQZO‘CforMachimDincflon

2 as

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

74 —‘

°°‘°° A B c
..,._____E z _ -_.—— 1 l. .
l“ 'l .. {j
«w T I“ I §_ 1’ a A T31 lfliagw—

«winfm in}“+ul fyiu y- ”I {a {at I

13 1c 2A 28 38 ac 3A 4B
PFKJCEfiiquGlGfl00llfls

Figure 40: E” values for machine direction at 20°C.

80

181C2AZB3B3C3A48

 

 

 

 

 

 

 

 

0000| __ A— _~ _ M—e [ - — B C ._
8000 L — 444»a_#____._ —_._fic___ , ,43____~#___
l
E
g 40.00 .. . _--__ mwv ”th __
a ”in 29 30
”in? I in {a}. lagfll

 

zomlt 1 . 1. 1
{w (u 5“ w "fur“;«t” éwl

 

 

 

 

- 3,1.._ .._, .,- 3

1B 1C 2A 28 38 33 3A 48

_T_

 

18102A2838363A4818102AZBanSAJB
PROCESSIIBGROUPS

000

 

 

 

 

Figure 41: E” values for cross-machine direction at -100°C.

 

 

00.00 ——x - — .7 .—- _-__- f B _ c

 

 

 

 

so 00 I»___u_- . - _
‘6‘
3. I os 41

40.00 L ,____ _. - . ___ ,1}
g u 3. 37 1
3 {as
W :0 3°

.4 i. W

 

 

 

 

 

20.00 L_ $22 I mac. { 22 _- _ %20 T .t
l 1‘ 10 10 ‘7

l

000 ;__..z. r—fi - .ar»- 4 -—.—».~—-~r— —_~-._ 1

l 18 1C 2A 28 38 SC 3A ‘8 18 10 2A 28 38 3C 3A 48 18 10 2A 28 38 3C 3A 48

 

 

 

PROCESSING GROUPS

Figure 42: E” values for cross-machine direction at 20°C.

81

The results for the E” data showed no obvious trends or correlations
between the total energy absorbed values of the high strain and impact data.
However, process groups 2B and 4B in the machine direction at 20°C had the
highest values. This was consistent for all three regions, indicating that energy
absorption can be affected by processing.

The tan 8 values for cross-machine direction at 20°C (Figure 43) were

 

a
plotted and compared to the energy values for the high strain and impact data. .,
Tan 5 is the ratio of E”/E’. This ratio measures the possible energy absorption of l
the material normalizing the data from sample to sample. The results also did
not show any firm trends or correlations, however some process parameters did i-

have significant effects on the tan 6, and hence energy absorption.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Tan 5 Values @20’Cfor Cross-Machine Direction
0.16 B C
0.14 __
‘F
0.12
0.11
0.10 11
0 .,. 0 0.11 am 0.10 0,10 0 0.
g 010 g
0.10 0.08 L I 000
-- 0.00 . ., J
'3 00.01 0.093 “009' 00.09 I °~°° i”
> 0” vk - v A A
IO
5 00.07 J. 0007 0.00 0.00 °'°°-- (l I;
"' 0.03 0,07 .
“0C6
0.04
0.02 .
om f v V r v . f v I ‘F V 7 v f“ 7 "'T' 7" v 1? v v I
18102AZBSB$C§3A4B 181C2A2838303A4818102A2838303A48
PROCESSINGGROU’S

 

 

 

.Figure 43: Tan 8 values for cross-machine direction at 20°C.

82

Results of High Strain Tensile Tests
Five samples of each processing group (eight total) were tested at 3, 9,
and 15—mph. The average values of all five samples for total energy (J), total

displacement (mm), and initial force (N) are shown in Table 17.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample # Tot. Energy Tot. Disp. Initial Force.
Hi (my 1!)
3-mph 1 B7 0.0037 2.39 67.76
9-mph 1B12 0.0013 0.79 59.53
15-mph 1B8 0.0016 0.92 67.60
3-mph 1011 0.0038 2.55 65.26
9-mph 1019 0.0016 0.74 76.30
15-mph 1C12 0.0016 0.99 82.04
3-mph 2A7 0.0051 2.63 81.46
9-mph 2A13 0.0014 0.90 68.41
15-mph 2A8 0.0013 0.93 79.05
3-mph 2B11 0.0032 3.37 53.41
9-mph 2B19 0.0014 0.71 65.88
15-mph 2B12 0.0011 0.96 66.61
3-mph 388 0.0042 2.45 68.10
9-mph 3B13 0.0010 0.80 49.29
15-mph 3B9 0.0013 1.27 74.40
3-mph 307 0.0058 1.78 102.05
9-mph 3C13 0.0012 0.80 58.58
15-mph 3C8 0.0015 1.16 86.83
3-mph 3A7 0.0056 3.02 92.01
9-mph 3A13 0.0014 0.74 69.85
15-mph 3A8 0.0014 1.16 83.00
3-mph 437 0.0040 2.29 65.30
9-mph 4813 0.0011 0.81 55.60
15-mph 438 0.0012 1.17 70.90

 

83

 

 

Table 18: Results of high strain data for all 3 speeds.

Examination of the data in Table 18 shows the 3-mph samples had a
higher level of energy absorption, approximately three times higher than the 9
and 15-mph samples. To evaluate how the energy varied according to

parameter, the energy data was graphed according to speed and process

parameter (Figures 44-47).

 

3DGRAPHNHIGHSTRNN-TIE1mElnHEATER)

 

I” ”c (35)
.90”: (18)
.120 uc (16)

011310 ~ 1Nm(1C)

   

$500 (18)
WPLEID

9 mph
spam (mph) ‘5 '"P"

 

 

 

Figure 44: 3-D graph of total energy absorbed during high strain rate testing as a
function of time in the heater.

Figure 44 shows the total energy values are approximately three to four

times higher for the 3-mph data than the 9 and 15-mph data for all three times in

84

the heater. It also shows that the material tested at 3-mph are processed at 60
seconds has absorbed more total energy than at 90 and 120 seconds at 9 and
15-mph. The total energy is largest for the material processed at 120 seconds.
This could be due to 60 seconds being the “cut-off” time in the heater to obtain
“perfect" flow of the PP. . Longer times cause the PP to have a higher
temperature and therefore a lower viscosity, causing the PP is spread too thin

throughout the material and therefore reduces the level of adhesion.

 

3-D GRAPH HGH STRAIN -'I'IIE 2 (“ME In W)

mace»
I45uc(38)
lwsocQB)

 

 

SPEED (mph)

 

 

Figure 45: 3-D graph of total energy absorbed during high strain rate testing as a
function of time in the mold.

Figure 45 shows the total energy at 3-mph is approximately four times

higher than at 9 and 15-mph. It also shows the material processed at 30

8S

 

seconds in the mold to have the highest levels of total energy at 3-mph but
processing at 45 and 60 seconds results in large total energy at 9 and 15-mph.
This could be due to the material being constrained too long in the mold, which

then reduces the level of adhesion.

 

$0 GRAPH HIGH STRNN - TEMPERATURE

.216'1: (420$) (30)
.227-c (440*) (3c)
Ina-014003) (:w

  
  
 
  

WC (460'F) (3»
227°C (440‘F) (3C)

216‘C (420$) (33)
9 m

ph l
snail-mm '5 "‘P"

 

 

 

Figure 46: 3-D graph of total energy absorbed during high strain rate testing as a
function of temperature.

Figure 46 shows the total energy of 3-mph data is approximately four
times higher than the 9 and 15-mph data. It also shows that the material
processed at 227°C and 238°C have higher similar levels of energy

approximately 28% higher than at 216°C. Total energy at 9 and 15-mph is

86

 

greater at 227°C and 238°C than at 216°C. This could be due to the material
reaching a sufficiently high melting temperature for more even and consistent

flow of the PP over and through the PET fibers, thus leading to better mechanical

 

3-0 GRAPH HIGH STRAIN - PRESSURE

TOTAL EMU) -

l1.5ntndep(aa)
I1.25m'ndsp.(48) _

  
  
 

1.25 m dip. (4B)

1.5mmm SAMPLEID

15mph

SPEED (mph)

 

 

 

adhesion.

Figure 47: 3-D graph of total energy absorbed during high strain rate testing as a
function of pressure.

Figure 47 shows that the 3-mph data is approximately four times higher
than the 9 and 15-mph data. It also shows that both displacements have around

the same level of energy absorption.

87

By looking at Figures 44-47, all energy values are considerably higher at
3-mph than at 9 and 15-mph. Materials processed at 60 seconds in the heater,
30 seconds in the mold, at 227°C or above, and at 1.5-mm of pressure all have
the highest total energy values. All processing conditions tested at 3-mph
showed higher total energy values than the 9 and 15-mph data. There is a
change at 9-mph in which the data drops and then increases again at 15-mph. T‘
This could be due to the way in which the material behaves at slow speeds and
fast speeds. Since the material is being pulled in tension, it experiences

stretching and pulling. As it is stretched at the faster rate, fibers will fail in more :

 

of a brittle manner with less energy. Also, at slower speeds, sliding of the PET
fibers around each other and through the PP is more likely. Perhaps 9-mph is a
crossover point in which the material responds with a mixture of both kinds of
speeds. This could also be due to the method of failure of the material at various
speeds.

To test this hypothesis, the method of failure was then examined to try and
explain the gross difference in energy absorption. Samples were evaluated
visually for the amount of fiber pullout, or appearance of the samples. After
looking at representative samples of all eight processing groups at all three
speeds, it was seen that the 9-mph samples had very clean failure edges, with
little or no tearing (Figure 48). The 15-mph samples showed very torn and
“chunky” failure edges (Figure 49). The 3-mph samples exhibited a mixture of
the 9 and 15-mph failure edges (Figure 50). To examine and verify this further,

microscopy using an ESEM was done.

88

 

Figure 48: 9—mph clean failure edge with little tearing.

 

 

failure edge.

Figure 49: 15-mph torn and “chunky”

90

 

Figure 50: 3-mph with a mixture of 9 and 15-mph failure edge.

 

9|

ESEM Microscopy of High Strain Tensile Samples

High strain tensile samples at 3, 9, and 15-mph were examined using an
ESEM to see whether differences in the fracture surfaces would be evident from
one speed to another, and also to explain why the 3-mph samples absorbed
approximately three times more energy than the 9 and 15-mph samples. Gross
differences from sample to sample according to different strain rates were not
seen. The samples, however, were very consistent in that:

1. Clumps of PP were seen throughout all the samples (Figure 51)

 

Figure 51: PP seen in clumps throughout the EAM.

92

2. The PET fibers failed in a brittle manner (Figure 52)

 

Figure 52: Fractured PET fiber surface.

3. There is both cohesive and interfacial failure of the PP, though more

interfacial failure is seen (Figure 76). One can see PP does not completely

adhere to or surround the PET fiber.

93

 

 

Figure 53: PET fiber surrounded by PP (interfacial failure).

4. Baseline samples (3B) appeared to have "stringy" PP fibrils attached to the

PET fibers (Figure 54). One can see the PP fibrils attached to the PET. This

is likely due to a PET fiber being pulled out of the PP.

94

 

 

Figure 54: PP fibrils attached to PET fibers.

5. Samples at the higher temperature (3A) seem to have more uniform PP
through out the sample, with poor adhesion. Since the samples are more
compressed, the PP is spread throughout more of the sample and makes the
samples appear as though they have more PP than the others. The resulting
adhesion, however, is still poor with PET fiber pull out.

After evaluating the samples at both low and high magnifications, a more
substantive correlation between failure modes at various speeds and the amount

of energy absorption could not be found.

95

Results of Impact/Penetration Tests

Five samples of each processing group (eight total) were impacted at 3, 9,

and 15-mph. The average values of all five samples for total energy (J), total

displacement (mm), and initial force (N) are shown in Table 19.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample # Tot. Energy Max. Disg Tot. Disp. Max. Force
“1 (mm) (mm) 1!)
3-mph 189 0.029 3.89 5.99 358.78
9-mph 1B10 0.056 4.24 5.52 442.25
15-mph 1B11 0.013 3.26 4.00 334.30
3-rnph 1013 0.029 4.28 6.48 407.95
9-mph 1014 0.050 3.09 3.59 444.75
15-mph 1015 0.014 3.67 5.21 333.90
3-mph 2A9 0.018 3.36 5.83 392.65
9-mph 2A10 0.059 3.32 4.11 489.56
15-mph 2A11 0.017 4.07 5.35 401.30
3-mph 2B17 0.021 5.51 9.30 545.10
9-mph 2B15 0.045 3.29 4.02 426.13
15-mph 2B16 0.011 4.17 5.21 304.81
3-mph 3B10 0.019 4.55 6.95 386.00
9-mph 3B11 0.058 4.84 6.24 404.38
15-mph 3B12 0.018 4.24 5.55 469.00
3-mph 309 0.018 4.33 5.96 336.90
9-mph 3010 0.039 4.58 6.20 267.56
15-mph 3011 0.015 3.11 4.18 256.19
3-mph 3A9 0.021 4.28 5.78 420.05
9-mph 3A12 0.036 3.40 6.01 421.25
15-mph 3A11 0.016 3.70 4.58 291.55
3-mph 4B9 0.017 4.03 6.09 312.20
9-mph 4B10 0.064 4.52 5.51 462.60
15-mph 4B11 0.012 3.70 4.77 289.15

 

 

 

 

 

 

 

Table 19: Results of impact/penetration data for all three speeds.

96

 

 

By looking at Table 19, the 9-mph samples had a higher level of energy
absorption, approximately two to three times higher than the 3 and 15-mph
samples. The 3-mph had the second highest and the 15-mph samples had the
lowest. To evaluate how the energy varied according to parameter, the energy

data was graphed according to speed and process parameter (Figures 55-58).

 

3-D GRAPH "PACT - TIE 1 (TIDE In HEATER)

 

0.04 9.
TOTAL EIERBY (J) lgflcfi:
I sec
I120..c(1C)

     

IZOschC)

gaseous)
mu)
amph
9mm
SPEED (mph) "m9"

 

 

 

Figure 55: 3-D graph of total energy absorbed during impact testing as a
function of time in the heater.

Figure 55, a 3-D graph of the amount of energy absorbed at 3, 9 and 15-
mph as a function of time in the heater, shows total energy absorbed at 9—mph is

approximately three to four times higher than 3 and 15-mph. It also shows

97

material processed at 60 seconds in the heater has the highest total energy.
This again could be due to 60 seconds being the “cut-off” time in the heater to
obtain “perfect" flow of the PP. Anything longer than 30 seconds causes the PP

flow too much and spread in areas too thin and therefore reduces the level of

adhesion.

 

3-D GRAPH IMPACT URI 2 (TIME In MOLD)

 

I30uc(2A)
I45|oc(38)
loo-acne)

TOTAL EIERGY (J)

 

15mph

SPEED (mph)

 

 

 

Figure 56: 3-D graph of total energy absorbed during impact testing as a
function of time in the mold.

Figure 56, a 3-D graph of the amount of energy absorbed at 3, 9 and 15-

mph as a function of time in the mold, shows data at 9-mph is approximately four

to five times higher than at 3 and 15-mph. Again, It also shows the material

98

processed at 30 seconds in the mold to have the highest levels of total energy at
3—mph. This could be due to 30 seconds being the “cutoff” time in which

optimum levels of molten PP adheres to PET. Anything greater than 30 seconds

reduces levels of adhesion.

 

3-D GRAPH "PACT - TEMPERATURE

I 216‘0 (420$) (33)
I 221'c (440?) (so)
. 23:-c (nan-F) (3A)

   
 
 
      

230'6 (WT) (3A)

221-0 (440'?) (30)

. sun: In
3 mph 216‘0 (420'F) (as)
9 mph
SP- (mph) ‘5 '"P"

 

 

 

Figure 57: 3-D graph of total energy absorbed during impact testing as a
function of temperature.

Figure 57, a 3-D graph of the amount of energy absorbed at 3, 9 and 15-
mph as a function of temperature, shows the total energy absorbed at 9-mph

approximately four times higher than at 3 and 15-mph. The total energy is

99

approximately 33% higher for material processed at 216°C. This could be due to
216°C being the “ideal" processing temperature for the material to achieve
optimum flow and adhesion. Higher temperatures without added pressure would
allow surface tension forces to contract the melt thereby reducing the effective

mechanical adhesion.

 

01mm
0125mm

 

 

 

 

Figure 58: 3—D graph of total energy absorbed during impact testing as a
function of pressure.

Figure 58, a 3-D graph of the amount of energy absorbed at 3, 9 and 15-
mph as a function of pressure, shows 9-mph total energy absorbed to be

approximately four to five times higher than at 3 and 15-mph. It also shows

100

material processed at 1.25-mm of displacement has a higher total energy at 9-
mph. This could be due to more flow of the PP and thus better mechanical
adhesion and fewer voids.

In Figures 55-58, all energy values are considerably higher at 9-mph than
at 3 and 15-mph. At 3 and 15-mph the total energies are similar. Material
processed at 60 seconds in the heater, 30 seconds in the mold, at 216°C at 1.25-
mm displacement also have the highest levels of total energy. All processing
conditions tested at 9-mph showed higher total energy values than the 3 and 15-

mph data. There is a change at 9-mph in which the data increases and then

 

decreases at 15-mph. This could be due to the way in which the material reacts ...
to impacts at slow speeds and fast speed. During impact, the material

experiences shear forces. When it is impacted at a faster speed, the material

has less time to react and behaves in a different manner than at slower speeds.

It will fail in more of a brittle manner with less energy absorption and higher force

of impact. Perhaps 9-mph is a crossover point in which the material responds

with a mixture of both kinds of failure modes.

To test this hypothesis, impacted material was examined to determine
failure modes to explain the differences in energy absorption. Samples were
evaluated visually for the amount of fiber pullout, or appearance of the samples.
After looking at representative samples of all eight processing groups at all 3
speeds, it was seen that the 3-mph samples had very clean failure edges, with
little or no tearing (Figure 59). The 15-mph samples showed very torn and

“chunky” failure edges (Figure 60). The 9-mph samples exhibited a mixture of

101

 

the 3 and 15-mph failure edges (Figure 61). The impact failures should be
different to the high strain samples since they are failing in compression and
shear instead of tension during the high strain. More stretching of the PET fibers
should be seen during tension testing. To examine and verify this further,

microscopy using an ESEM was done.

Failure edge
of sample

 

Figure 59: 3-mph impact clean impact edge.

|02

 

Failure edge
of sample

   

Figure 60: 15-mph impact torn and “chunky" failure edge.

Failure edge
of sample

 

Figure 61: 9-mph mixed failure edge.

103

ESEM Microscopy of Impacted Samples

Impacted samples at 3, 9, and 15-mph were examined with an ESEM to
detemtine whether gross differences in the failure mechanisms of the EAM at
various speeds would be seen. Samples did not show gross differences from
speed to speed, except in the way in which the PET fibers failed. Samples at 3-
mph seemed to fail in a very brittle manner, with very clean surfaces (Figure 62).
Samples at 15-mph showed a different type of failure with PET fiber surfaces
being very jagged and torn (Figure 63). Samples at 9-mph showed a mixture of
both types of failures shown in both the 3 and 15-mph samples (Figure 64).

D

Brittle,
clean
PET

fibers

 

Figure 62: Brittle, clean surface failure of PET fibers at 3-mph impact.

104

Jagged,
torn
PET
fibers

 

Figure 63: Jagged and torn PET fiber failures at 15-mph impact.

Brittle,
clean
fibers

Jagged,
torn
fibers

 

Figure 64: Mixture of brittle, clean and torn and jagged PET failure at 9-mph
impact.

105

All samples showed consistent results in that:

1. The PP stretches and breaks when it fails (Figure 65)

Broken PP

 

Figure 88: Stretched and broken PP throughout the EAM.

2. The PET fibers fail more in a brittle than ductile manner (Figure 66)

 

Figure 66: Brittle failure of PET fibers.

106

3. There is some adhesion between the PP and PET but it is very poor with the

PP breaking off the PET fibers (Figure 67).

Breaking of PP
among PET fibers

4»: 3‘
I r —
l'flflI‘ 1‘

 

Figure 67: Breaking of PP among PET fibers.

4. PP is again found in chunks throughout the samples with the PP not being

very uniform (Figure 68).

107

A Fiber

‘ pull-out

 

Figure 68: Non-uniform dispersion of PP, along with evidence of PET fiber
pull-out of PP
5. There are large amounts of fiber pull out of the PET fibers from the PP
(Figure 68).
6. The samples fail in regions of high PET concentrations and low PP

concentrations (Figure 69).

108

 

High concentration of
PET fibers in fracture
area of EAM

 

Figure 69: Failure of EAM in areas of high PET fiber concentrations.

7. There is evidence of some residual PP seen on PET fibers (Figure 70).

 

Figure 70: Remnant PP on PET fiber.

109

 

8. PP in samples processed at a higher temperature (3A) appears to be more

uniformly distributed throughout the sample (Figure 71).

   

Figure 71: High temperature processed EAM with more uniform PP distribution.

”0

 

9. High pressure samples (4B) showed the PP appears more compacted and

denser (Figure 72).

   

Figure 72: High pressure EAM sample with more compacted and denser PP
appearance.

The results of the effects of processing on the energy absorption of the
EAM for both high strain and impact testing are summarized in figures 96 and 97.
The results are displayed as percent changes from the baseline samples (33)
for: i) time in the heater, ii) time in the mold, iii) temperature, and iv) pressure.
The baseline samples processing conditions against which the other processed

samples were compared are l) 60 seconds in the heater, ii) 45 seconds in the

111

 

 

mold, iii) at a processing temperature of 216°C, and at iv) 1.5—mm of

displacement. The graphs show trends for both high strain and impact samples.

 

High Strain Summary

 

Change In Energy from Buoflno 80mph. (51.)

 

 

 

Turret M Tin-20m“) Term Pusan
goand120 seamen Pronto-1mm 227'Cand 1.25-m
would: seconds 238'0 disp.

 

Figure 73: Summary of effects of processing parameters (Time in the heater,
time in the mold, temperature and pressure) for high strain testing as
percent changes from baseline samples.

The high strain data (Figure 73) shows that as time in the heater is
increased, the energy absorption increases at 90 and 120 seconds for both at 9-
mph and 15-mph. It also shows that decreasing the time in the mold to 30
seconds increases the energy absorption capacity of the EAM. Increasing time
in the mold to 60 seconds, again decreases the energy absorption at 3 and 15-

mph but increases it at 9-mph. Increasing the processing temperature to 227°

112

and 238°C increases the energy absorption capacity of the EAM. Increasing the
pressure to 1.25-mm of displacement decreases the energy absorption at both 3

and 15-mph but increases it at 9-mph.

 

Impact Summary

   

Change In Energy from the Beeellne Semplee
(%)

 

Tlme 1 (ne‘er) Tine 2 (meld) Terrp Preeeue
90 and 120 30 and 60 Proceeelng Parameter arc and 1.25-mm
seconds securds 238'0 dlep.

 

 

Figure 74: Summary of effects of processing parameters (Time in the heater,
time in the mold, temperature and pressure) for high strain testing as
percent changes from baseline samples.

The impact data (Figure 74) shows that as time in the heater is increased,
the energy absorption decreases at 90 and 120 seconds for both at 9 and 15-
mph and increased at 3-mph. It also shows that decreasing the time in the mold
to 30 seconds decreases energy absorption for 3 and 15-mph. Increasing time in

the mold to 60 seconds again increases the energy absorption at 3-mph but

113

 

decreases it at 9 and 15-mph. Increasing the processing temperature to 227°C
decreases the energy absorption for all 3 speeds, but at 238°C it increases for
the 3-mph data. Increasing the pressure to 1.25-mm of displacement decreases

the energy absorption at both 3 and 1.5-mph but increases it at 9-mph.

 

114

 

1.

CONCLUSIONS

 

Fluorescent impregnated samples showed that there was a relationship
between the amount of fluorescent epoxy absorbed by the EAM and the level
of damage seen in an element. Non-impacted samples showed little or no
penetration by the fluorescent epoxy anywhere in the element, while impacted
samples showed that there were large areas penetrated by fluorescent epoxy
in the side wall area of the element.

. Samples showed that there were crushing in the walls of the EAM element

 

(region B) with the rest of the EAM element intact (regions A, C, and D).

. Examination of these micrographs showed that failure occurs in the walls
of the EAM element and not at the apex of the element which is left
completely intact.

By evaluating micrographs, it was found that the process of fabricating the

sample sheets at MSU is not representative of what is seen in the EAM when

processed at the plant.

. The process of fabricating the sheets at MSU is “too perfect,” meaning it
did not under go bi-axial Stretching affects seen at the plant, along with
shear stresses seen when the EAM is formed in the wall regions.

The first set of samples fabricated revealed that the mechanical properties of

the EAM can change according to processes conditions.

. This was verified by the large amount of scatter seen when examining the

modulus, total energy, and pseudo energy values extracted from tensile

115

 

test data on various samples processed at different times and

temperatures.

4. By comparing all three different results from the Dyna-tup testing, it was

concluded that the EAM absorbs energy similar to if not better than foam to

achieve the optimal energy absorption square curve.

Dyna-tup testing was promising, but limited since the highest impact

speed that could be reached is around 12.8 km/hr.

5. Flat sample sections of EAM of 1.5, 1.2, and .65-mm thicknesses tested in

the DMA at 150, 15, and 1.5 Hz showed that as the thickness decreased, the

storage modulus increased. Transition temperatures were sensitive to

process conditions.

The way in which the samples are tested does affect the results. Samples
that were allowed to completely cool between thermal runs showed
different E' values in all three runs for all three thicknesses. Samples not
allowed to completely cool between runs showed that only the first E' for
the first run was lower than the second and third runs for all three
thicknesses.

Micrographs showed that the samples at 1.5-mm in thickness had more
voids than the 1.2 and .65-mm samples. Samples were thicker near the
edge and thinner in the center.

Examination of cross-machine and machine data in the DMA showed
trends between the -100°C and 20°C data. The -100°C data has higher E'

values than the 20°C data. The machine direction data also has higher E’

116

values than the cross-machine data. On average, the B regions of the
cross-machine data have lower E' values than the A and 0 regions. This
correlation is not seen in the machine direction data.

6. By determining the E’ from DMA data of the different processed EAM groups,
it was found that there were slight differences between the A, B, and 0
regions of the material.

. Increasing time in the heater increased E’ and the scatter in the E’ data for
both the B and 0 regions.

. Increasing time in the mold and the processing temperature seemed to

 

 

TIT; ‘1‘ 4: u -rr-LlL'J-n-r._.-

have similar effects on the E’ data in all three regions of the EAM. There
seems to be a difference in material properties from the material at the top
region of the EAM element represented by areas A and half of B, and the
bottom portion of the element represented by the bottom region of B and
C.

. Increasing the pressure increases the E’ values for all directions,
temperatures, and regions. 1

7. Evaluation and examination of the tensile failure surface of high strain tested

EAM samples showed:

. The data at 3-mph had the highest level of energy absorption for all
samples. Samples processed at 60 seconds in the heater, 30 seconds in
the mold, at 227°C or above, and at 1.5-mm of pressure all have the

highest amount of energy absorbed.

117

 

At nine and 15-mph the energy absorption is three to four times lower than
at 3-mph.

When examining the amount of fiber pullout, or appearance of the strained
samples, the 15-mph samples showed the most fiber pullout with very
jagged and “chunky” looking failure edges. The 9-mph showed the least
amount of fiber pullout with very clean and little or no tears in the failure
edges of the samples. The 3-mph samples showed a mixture in the
amount of fiber pullout and the appearance of the failure edges of both the

nine and 15-mph samples.

8. Evaluation and examination of the shear failure surface of EAM samples

subjected to impact loading showed:

The 9-mph samples all had the highest level of energy absorption.
Samples processed at 60 seconds in the heater, 30 seconds in the mold,
at 216°C, and at 1.25 mm of pressure all have the highest amount of
energy absorbed.

At three and 15-mph the energy absorption is two to three times lower
than at 9-mph.

When examining the amount of fiber pullout, or appearance of the strained
samples, the 15-mph samples showed the most fiber pullout with very
jagged and “chunky” looking failure edges. The 3-mph showed the least
amount of fiber pullout with very clean and little or no tears in the failure

edges of the samples. The 9-mph samples showed a mixture in the

118

 

amount of fiber pullout and the appearance of the failure edges of both the
nine and 15-mph samples.

a The ESEM revealed that the PET fibers fail in a very clean and brittle
fashion at 3-mph, but the 15-mph fibers failed in a very jagged and torn
fashion. 9-mph samples had fibers that failed in both ways.

9. Micrographs revealed that there is poor fiber to fiber adhesion of the PP and

 

PET fibers in the EAM, primarily due to mechanical adhesion.

119

 

FUTURE RESEARCH

 

The following is a list of recommended experiments for the EAM:

1. Determine if there is a way to minimize the heterogeneity of the material.

2. Investigate the “annealing” process of the material — how does it affect the
material and what exactly is happening to the material when it undergoes this
process.

3. Determine a method to evaluate the adhesion between single PP and PET
fibers.

. Measure the adhesion between individual PP and PET fibers at different
time, temperatures, and pressures in order to analyze whether or not
increasing adhesion will improve energy absorption.

4. Identify why the 9-mph impact and the 3-mph high strain samples absorbed
energy to a much greater extent.

5. Investigate the effect of the PP/PET ratio on energy absorption.

6. Investigate creating a uniform thickness throughout the element, especially

thickening region B.

120

 

REFERENCES

 

1. Federal Motor Vehicle Safety Standards; Head Impact Protection (Docket 92-
28, Notice 4), Federal Register, Vol. 60 (160), pp. 43032-43061, August 18,
1995.

2. Lorenzo, Luis et al., “Improving Head Impact Protection,” Automotive
Engineering, Vol. 104, No. 5, pp. 49-53, May 1996.

3. 'Syrowik, Glenn F. et al., “Energy-Absorbing Polyurethane foam to Improve
Vehicle Crashworthiness,” SAE Transactions: Journal of Materials and
Manufacturing (USA), Vol. 104, pp. 411-420, 1995.

4. Sounik, David F. et al., “Head-Impact Testing of Polyurethane Energy-
Absorbing (EA) Foams,” SAE Transactions: Journal of Materials and
Manufacturing (USA), Vol. 106, No. 5, pp. 211-220, 1997.

5. Porter, J.H., “Utilizing the Crushing Under Load Properties of Polypropylene
and Polyethylene Honeycomb to Manage Crash Energy,” SAE Transactions:
Journal of Materials and Manufacturing (USA), Vol. 103, pp. 659-666, 1994.

6. Locke Deborah J. et aI., “Energy-Absorbing Therrnoplastics for Head-Impact
Applications,” SAE Transactions: Journal of Materials and Manufacturing
(USA), Vol. 105, No. 5. PP. 133-137, 1996.

7. Morye, 38. et al., “Studies in Blends of Polypropylene and Polyethylene
Terepthalate,” Journal of Polymer Materials, September 1996, pp. 217-221.

8. Hart, Harold, “Organic Chemistry: A Short Course,” Eighth Edition, Houghton
Mifflin Company, pp. 383, 397, 1991.

9. Cheung, Man Ken et al., “Mechanical and Rheological Properties of
Poly(ethylene terpthalate)/Propylene Blends,” Polymer International, 43,
1997, pp. 281-287.

10.Sperling, L.H., "Introduction to Physical Polymer Science," Second Edition,
Lehigh University, Bethlehem, Pennsylvania, pp.322.

11.Dynatup General Research Corporation, “Dynatup Model 830-1 Manual,” pp.
A3-1 to A3-3.

12.TA Instruments, “DMA 2980 Dynamic Mechanical Analyzer Operator's
Manual,” Revision 0, pp. 1-4, 5-10, 1997.

121

 

13.“Dynamic Mechanical Testing,” Adhesives and Sealants, Vol. 3, pp. 318-320,
1990.

14. Du Pont, DMA Handbook

15.ASM International, Engineering Materials Handbook, Volume 3 Adhesives
and Sealants, pp. 318-320, 1990.

16.ASM International, Engineering Materials Handbook, Volume 2 Engineering
Plastics, pp. 50-53, 1988.

 

122