 

A S‘

y .r-uqr 1;

 

2-.“ . u u .-

 

”I...

 

 

 

.
r
u

i

i.
l
I
..,
i.
i.
5'
I.
I
'f
v

 

hu'a

 

 

 

 

5“. m r em] - £31.53 '1 '
,

.

3.12;»: ‘1. ~*

$75.52;. .
"1‘L

1.13”“
.K'.‘ 4.:

:73) 3I:IL\:"
It». ”a”: :35 «7172'
“~.~2.£}1 law.”

“£41335;

.
, Viv
w”

‘3 13.37.?)
'i‘k‘é'fi
.223 “gang: .

 

:3*.;§;:‘3*.r ”If“? 'G.
”I... .AA-suéhigxrtx} a"

‘21."
w; m.” r .2 5:2“
.. 3:1}. .3525? $2». "" \
4U 'illwu. .»g¢.~.. .
w c '"k
137:3;
..
:M"

if
(us. “\3
h,- 2. ‘u »
...y.“7...6\.1-v.. '-¢fu.7~s..L-m.u
1

 

. - Erma-a1}

“ .
...~‘:..:“' g"\x.u.‘4'uv;«“ :73:
5323: . uni-1w“; an"
".1" 7-11“;- . a.”
s! .271: 3-: 33' 1. " xii
cakg._"~m-Jv $.71. 7 iii-7Q
25' . $35.2.

 

 

7 ‘ . . 1 2
42.75:“: gig. ”
. .

 

.1 2: ;%‘2 ‘u
my ' .1 .72 -
.‘igu. 1. .{V‘i—rﬁ- .Ehr‘
.9 u r ‘

 

1:. v u
.7 32.4.15... 7:22: a: 7"-
yea-24.12.33 M
ya“! “on." ‘ .IV
v mm. '
"n" 44...“. n.312l ‘x‘
u, ~ 5.12: W
”‘21.”...Iv.’ "“1.ij ‘6.
. {1:17
a... ’

1
>5.
1:.

   

 

haul

 

:< u

-32? uni-1.4.0.. ’13.»; 3';— v

meta-4A“ gauge.” .1.
”'35: 35%.

   

2M“
4.4-,»

» gs.

 

Lu. \
.11.-..

, “at
m r”:
”khtggﬁg—w .12;
L

 

“iiig’wa 1. n‘ 2"
=. .. .z. »
’"d‘li ‘5. km sizivrru"

 

 

 

llJl“Lillllglﬂlllllllllllllllllllllllllllllllllllll

1055 3315

This is to certify that the
thesis entitled

The Adhesion of Ion Beam Irradiated
Ultra-High Molecular Weight Fibers to

Epoxy Matrix
presented by

Richard Schalek

has been accepted towards fulfillment
of the requirements for

Master's Materials Science

 

degree in

 

 

 

Major professor

MS U is an Affirmative Action/Equal Opportunity Institution

 

 

msnw

7'3? ,— , ‘
mlthgw

 

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

DATE DUE DATE DUE DATE DUE

q n

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Afﬁrmative ActiorvKual Opportunity Institution
cMicthedmprnS—pt

 
 
   

.n. _.l¢-a-1da - o..- M‘L‘UN‘4“"“ H.

   

,..»JJ.¢ ¢

 

THE ADHESION OF ION BEAM IRRADIATED ULTRA-HIGH
MOLECULAR WEIGHT POLYETHYLENE FIBERS
TO EPOXY MATRICES

By

Richard Lee Schalek

A DISSERTATION

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

MASTER OF SCIENCE

Department of Materials Science
and Mechanics

1992

 

ABSTRACT

The Adhesion oflon Beam Irradiated Ultra-High
Molecular Weight Polyethylene Fibers
to Epoxy Matrices

By

RichardLeeSchalek

Ultra-high molecular weight polyethylene (UHMW—PE) ﬁbers have a highly
chain-extended and crystalline structure which is functionally inert and requires surface
modiﬁcation before it can operate as a reinforcement in polymer matrix composites.
Traditional fiber surface techniques provide adequate ﬁber-matrix adhesion, but also
degrade the ﬁber's mechanical properties. Low energy (100 to 1000 eV) ion irradiation
is an emerging material surface modiﬁcation technique. UHMW-PE Spectra-10m ﬁbers
were irradiated by high (>30 keV) and low energy ions to doses of 11th13 cm"2 to
1x10” cm*2. Interfacial shear strength (188) tests show a maximum increase of 900%,
with little loss in tensile strength for the low energy irradiations. The responsible
adhesion mechanisms were the improved wetting properties, due to surface oxidation,
and increased ﬁber-untrix mechanical interlocking. The achievable 188 is dependent on
thetailoredﬁbersmfacemicrostructmeandthecohesive strengthoftheﬁber.

 

 

Table of Contents

List of Tables ....................................................................................................................... v
List of Figures ..................................................................................................................... vi
Chapter One
Introduction ............................................................................................................. 1
Chapter Two
Review of the Literatme .......................................................................................... 5
The Composite Interphase .................................................................................... 5
Ultra-High Molecular Weight Fibers ................................................................... 9
Traditional Surface Treatments .......................................................................... 13
Plasma and Chemical Treatments ....................................................................... 13
Electron and Photon Irradiation .......................................................................... 15
Ion Implantation of Polymers ............................................................................. 15
General Considerations .................................................................................... 15
Examples of Irradiated Polymer Studies ............................................................... 19
Chapter Three
Experimental Methods ........................................................................................... 24
Sample Preparation ............................................................................................. 24
Mechanical Tests ................................................................................................ 27
Surface Chemistry Characterization ................................................................... 34
Physico-Chemical Characterization ................................................................... 35
Microscopy ......................................................................................................... 39
Chapter Four
Results and Discussion ......................................................................................... 41
High Energy Ion Implantation of Spectra-1000 ................................................. 41
Visual Observations ......................................................................................... 41
Surface Chemistry ........................................................................................... 43
SEM Observations ........................................................................................... 54
TEM Observations ........................................................................................... 57
Gel—Fraction of High Energy Irradiated Fibers ............................................... 61
Low Energy Ion Irradiation of Spectra-1000 ..................................................... 63
Visual Observations ......................................................................................... 63
Mechanical Properties ..................................................................................... 63
Surface Chemistry ........................................................................................... 67
Surface Energy of Low Energy Irradiated Fibers ............................................ 75
SEM Observations of Low Energy Irradiations .............................................. ‘78
TEM Observations of Low Energy Irradiations .............................................. 86
Chapter Five
Conclusions ........................................................................................................ 88

iv

Table 2-1:
Table 3-1:
Table 3-2:

Table 3-3:
Table 3-4:

LIST OF TABLES

 

 

Mechanical properties of Spectra- -1(X)0 and Kevlar-49 ﬁbers ..................... 12
High energy ion implantation parameters ---26
Low energy ion irradiation parameters, at a constant energy of 1 keV... ....29
Low energy ion irradiation parameters, at a dose of 1x1016 cm’2 ............... 30
Surface enagies of characterizing liquids for the micro-Wilhelmy

experiments ............................................................................................ 37

.. 7‘- ‘ 91-. ‘. to! "T‘NFI. 'FJA‘J

 

 

 

 

Figure 2- 1:
Figure 2-2:

Figure 2-3:

Figure 3-1:
Figure 3-2:
Figure 3-3:
Figure 3-4:
Figure 3-5:
Figure 3-6:
Figure 4-1:

Figure 4-2:
Figure 4-3:
Figure 4-4:
Figure 4-5:

Figure 4-6:
Figure 4-7:

Figure 4-8:

Figure 4-9:

Figure 4-10:
Figure 4-11:

Figure 4-12:
Figure 4-13:
Figure 4-14:
Figure 4-15:
Figure 4-16:
Figure 4-17:

Figure 4-18:

LIST OF FIGURES

Illustration of the three dimensional ﬁber-matrix interphase ........................... 6
Schematic representation of conventional polyethylene ﬁbers

 

 

 

 

 

 

 

and chain-extended polyethylene ﬁbers 10
Schematic illustration of chain scission and crosslinking of

polymeric ﬁbers 20
Fiber and fabric lay-up procedure 2‘5
Illustration of low energy ﬁber irradiation 28
Schematic diagram of the rnicrodroplet pull-off apparatus ........................... 32
Examples of pull-off load vs embedment length for 1 keV Ar+ ton ............. 33
Schc matic diagram of attenuated total internal reﬂectance technique ............ 36
Schematic diagram of the micro-Wilhelmy r 38
Photographs showing the discoloration of Spectra-1000 irradiated

as a function of (a) dose and (b) Inn 42
ATR-F'I‘IR spectra for various doses of N+ ion implantation ..................... 44
ATR-FI'IR spectra for N+. Ti+ and Xe+ ions at 1x1015 cm'2 ..................... 46
A typical survey scan of untreated Spectra-KKK) fabric 48
A typical survey scan of Spectra-1000 fabric irradiated by 30 keV Ni” ions

irradiated to a dose of1x1015 cm:2 49

 

A high resolution scan of the Cls region for untreated Spectra-1000 fabric5l
A high resolution scan of the Cls region for Spectra-1000 fabric
irradiated by 30 kev N+ ions to a dose of 1x1015 cm'2 ......................... 52
Oxygen:carbon ratios as a function of ion dose for UHMW-PE fabric
irradiated by 30 keV N+ Inns
SEM micrographs of (a) control and (b) ion irradiated (1x1015 Ni'lcmz)s
treated fibers
SEM micrographs of (c) plasma treated and (d) corona treated ﬁbers ...... 56
TEM micrograph of a control ﬁber (a) low magniﬁcation
and (b) high magniﬁcation showing ﬁbrils pulling away from ﬁber ...... 58
TEM micrograph of a ﬁber irradiated by 100 keV ions, 1x1015 N+/cm'2,
(a) showing irradiation zone and (b) showing ﬁbrillar decohesion ........ 60
Gel-fraction as a function of dose for 30 keV N+ ions .............................. 62
The tensile strength of 1 keV Ar'+ irradiated ﬁbers as a function of dose..64
The tensile strength of Ar“+ irradiated ﬁbers as a function of
ion energy for a constant dose of 1x1016 cm‘2 ....................................... 65

53

 

The normalized interfacial shear strength for 1 keV Ar+

irradiated ﬁbers as a function of dose 66
The normalized 188 as a function of Art ion energy at

constant dose of 1x1016 cm“2 68

 

The dose dependence of 188 for 100 and 250 eV Art irradiations ............ 69

 

Figure 4-19:

Figure 4-20:
Figure 4-21:

Figure 4-22:
Figure 4—23:
Figure 4-24:

Figure 4-25:

Figure 4-26:
Figure 4-27:

Figure 4-28:

The oxygen:carbon ratios of fabric irradiated to a dose of 1x1016 cm'2

as a function of M ion energy

 

Theoxygen:carbon ratios for the 100and250eV irradiations
asafunctionofdose

 

Inﬁ'aredspectraforfabricirradiatedwithlkeVArtionsand
The sm‘facefreeena'gies ofthecontrolandirradiated ﬁbers

ladc micrometer

 

Microdroplct on untreated ﬁber indicating the ﬁber’s
poor wetting properties

 

TEM rnicrographs of a ﬁber modiﬁed by 500 eV Ar'+ ions with

70
The oxygenzcarbou ratios of 1 keV Ar“ ions as a function of dose ........... 72
73
1 keV N+ ions to doses of 1x1016 cm'2 and 1x1015 cm‘2 ...................... 74
as a function of Art ion energy for irradiations of 1x1016 cm:2 ............. 76
(a) Debond of the microdroplet from the ﬁber (note impressions made
by shearing blades) and (b) droplet failure on the side opposite the 9
b 7
an
Illustration of cohesive failure occurring deep within a ﬁber ..................... 82
SEM micrograph showing surface morphology of Art irradiations of
(a) 1 keV of dose 3x10ls cm'2 and (b) 250 eV of dose 1x10l6 cm'2 ..... 83
27

a dose of 6.0x1015 cm‘2

 

 

Chapter One
INTRODUCTION

The development of high performance organic reinfacing ﬁbers has lead to signiﬁcant
advances in the area of structural composites. Organic ﬁbers typically display tensile
strengths and tensile modulii of 4 GPa and 200 GPa, respectively. This matches the tensile
performance of metal and glass ﬁbers when compared on an absolute scale, while
surpassing them when compared on a weight-to-strength basis.

The effectiveness of the load transfer from the brittle matrix to the ﬁber, via the
fiber-matrix interphase, often is the limiting factor in the success or failure of a strucnual
composite. A fundamental understanding of ﬁber-matrix adhesion in different composite
systems is therefore a ﬁrst step in realizing the potential of these high performance organic
reinforcing ﬁbers.

High performance organic ﬁbers can be divided into three major classes: graphite
ﬁbers, synthetic rigid polymer ﬁbers, and ﬁbers of chain-extended, ﬂexible polymers.
Only graphite and synthetic rigid polymer ﬁbers have found commercial success as
reinforcing agents in structural composites. The commercial use of chain-extended ﬂexible
polymers (comprised of only highly oriented polyethylene) in structural composites has
been limited by their relatively low adhesion level to epoxy matrices. The pom interfacial
bonding results from the chain extended crystalline morphology of the ﬁber combined with
the lack of chemically active sites on the polymer backbone. The poor adhesive properties
of polyethylene ﬁbers have limited their use to composites for ballistic applications where
energy intensive absorption by ﬁber-matrix delamination is the primary goal.

The adhesion of polyethylene ﬁbers to polymeric matrices is commonly improved
by some type of ﬁber surface treatment, such as plasma, corona or chemical processes,
with plasma treatments being the most successful, economical and the most widely studied
[1, 2, 3, 4]. These surface modiﬁcation processes are designed to control the chemical

 

n

 

 

 

 

 

2
sm'face functionality, alter stuface topography and/or to remove weak boundary layers
containing low molecular weight oligomers tr other additives, with the general objective of
improving adhesion and wettability of the ﬁber surface.

Many ﬁber surface treatments have met with success; however, the tee ' ues are
highly parameter-sensitive and often negatively affect bulk ﬁber properties. Improved
surface processing methods are therefore actively sought. Low energy broad-beam ion
processing is a rapidly developing materials modiﬁcation technology which complements
andhas severaldistinctadvantagesoverdremoreuaditionalﬁbersmfacealtering
techniques [5]. Precise control of experimental parameters, such as ion direction, energy,
ﬂux and background pressure are currently not available from other processes. This
increased experimental parameter control capability allows a more direct correlation
between the observed surface modiﬁcations and fundamental ion-solid interactions, than is
available from traditional surface modiﬁcation techniques.

As a separate issue from surface modiﬁcation of polymers for adhesion
enhancement, there has been increasing interest in studying ion beam intaactions with
polymeric and insulating materials. Compared to metals and semiconductors, polymers
have distinct radiation damage mechanisms arising prirmrily from the differences in
molecular structrue and chemistry. Bombardment by energetic ions causes the polymer
components within the depth of ion penetration to interact with the incident particle,
forming ionized stable and metastable radicals which subsequently thermalize and
recombine or leave the surface resulting in substantial surface modiﬁcation. The extent to
which these radiation effects alter the polymer properties depends on the chemistry and
structureofthepolymeraswellasdrenatrneoftlreionbeamparticularlytheiontype,
energy and dose.

WittﬁnthepastZOyearslowenergy broad-beamionsourceshavebecome
important surface modiﬁcation devices. The technological development of these low energy
ion sources originated in the program for electric space propulsion [6, 7]. Signiﬁcant use

 

 

 

3
0f Fidded, broad beam ion sources started about 1970and increased thereafter. At

present, these ion som'ces are available in diameters of 1 cm to 38 cm, with ion-beam
currents ranging from a few milliarnperes to several Amperes.

The large current capability of a broad-beam ion source results in substantial
processing capacities and, hence, is useful in industrial applications. In a gridded source,
the large currents are made possible with optics that allow the beam proﬁle to be focussed
or defocussed (diverging). Thus, the ions are accelerated into a beam with a well-deﬁned
and controlled direction, density and energy. These experimental characteristics simplify
the control and process deﬁnition for ion beam sources over that of the competitive plasma
processes.

The present study is a continuation and expansion of a previous investigation
initiated by Ozzcllo, et a1. [8] concerning the effects of high energy (30-135 keV) ion
implantation of Spectra 1000 UHMW-PE ﬁbers. Ozzcllo, et a1. [8] found that high energy
ion irradiation resulted in a 300% increase of the ﬁber-manix interfacial shear strength
(158), with little, if any, degradation of the ﬁber tensile strength. These results provided
the impetus for the cm'rent study that investigates the mechanisms responsible for the
observed increases in 188 in both high energy and low energy (100 - 1000 eV) ion
irnplantations.

Tensile strength and 188 experiments determined the effects of low energy ion
irradiation on the mechanical properties of the ﬁbers. Surface chemistry characterization
(x-ray photoelectron spectroscopy (XPS), Fourier transform infrared spectroscopy (FI'IR))
of both the high and low energy ion irradiated ﬁbers have been performed, followed by
physico-chernical characterization using surface energy and gel-fraction wchniques.

Chapter Two presents a review of the literature including discussions of the general
concepts of adhesion, a description of the method of producing UHMW-PE ﬁbers along
with a summary of physical and mechanical property data of Spectra -1000 ﬁbers. A
summary of traditional ﬁber surface modiﬁcation techniques currently used to enhance

 

4
ﬁber-matrix adhesion is presented. The general effects of high and low energy ion

implantation of polymers and their relevance to ﬁber-matrix adhesion are discussed.
Chapter Three discusses sample preparation, surface chemistry characterization techniques
and experimental methods for determining mechanical properties. Chapter 4 presents and
discusses the experimental results . Chapter 5 draws conclusions regarding the probable

adhesion enhancement mechanisms and their contributions.

Chapter Two
REVIEW OF THE LITERATURE

The Composite Interphase
Theconceptoftheinterphasehasbeendescribedastheregionexistingbetweenthebulk
ﬁber and the bulk matrix [9]. The interphase region, shown in Figure 2-1, presumably
consists of unreacted species, impurities, voids and the ﬁber sm'face morphology resulting
mmmlmthcalandchenucdmvhonnenstdiﬁaentﬁomdmdmebulk
matrix and reinforcing phase. The effective bonding between the ﬁber and matrix is thus
complicated, requiring both chemical and mechanical characterization to completely
understand the underlying adhesion mechanisms.

Thephysical andmechanicalproperties ofacompositedependtoagreatextenton
the structure and properties of the ﬁber-matrix interphase; therefore, the adhesive bond
betweendreﬁberandmauixisofpnmeimpmmncemdrepcrformanceofcomposite
materials. This interphase bond is not only signiﬁcant to the strength and stiffness of the
composite (particularly off-axis), but also governs mechanisms of damage accumulation
and propagation.

Theadhesionofdreﬁba’mdremauixcanbeatuibutedtoadsorpdmandwetﬁng,
interdiffusion, electrostatic attraction, and mechanical interlocking mechanisms [10, 11].
These mechanisms are complex combinations of chemical and physical interactions.
Adsapﬁonismemolecuhramacﬁonmdconecﬁonofmemolectdarspedesonmthe
surface of another and results either through chemical interactions or physisorption
interactions. In general, chemical bonding interactions are two orders of magnitude greater
in strength than physisorption interactions and, hence, more desirable. Physisorption
mechanisms include nonpolar (Van derWaals) and polar (dipole and acid/base)

interactions, which, though weaker, act over longer distances.

 

 

 

 

 

  

Polymer of Differe t
Pro erties

Impurities

Unpolymerized Resin

Voids

 

 

 

 

Figure 2-1: Illustration of the three dimensional interphase.

7

The production of ﬁber reinforced polymeric composites requires that a liquid resin
make close contact with the swim of the ﬁber. An intimate ﬁber-resin interface is formed
when the liquid resin complewa spreads onto, or wets, the solid surface. Poor wetting of
the ﬁbers results in voids between the ﬁber and the solidiﬁed matrix, decreasing the
ultimate strength of the composite. Maximizing the interfacial molecular adhesion between
the ﬁber and the liquid resin, by matching the surface energy components of the two
adhering phases, maximizes the amount of ﬁber-matrix wetting.

Though wetting does not directly affect bonding mechanisms, it is a prerequisite to
achieving maximum levels of ﬁber-matrix adhesion. For a general ﬁber-matrix system, a
criterion for adequate wetting is difficult to deﬁne. Operational deﬁnitions, based on
thermodynamic considerations, provide a convenient method of expressing the relative
degree of ﬁber-matrix wetting. For instance, good wetting can be deﬁned as the spreading
ofa liquid on the surface of a solid with the ﬁber-resin contact angle less than 90 degrees.
Also,ifthe surface energy ofthe ﬁberis greaterthanthatoftheresin adequate wettingwill
occur.

Molecular interdiffusion is another possible ﬁber-matrix adhesion mechanism. The
bond strength depends on the amount of molecular entanglement, the number of molecules
involved and the strength of the bonding between the molecules. For most polymers,
interdiffusion between the epoxy and ﬁber is limited, but in the case of UHMW-PE, where
the composite consolidation temperature is approximately 85% of the melting temperatrne,
interdiffusion may occtn‘, resulting in enhanced adhesion.

Differences in the electrostatic charge between the ﬁber and the resin matrix,
resulting from the transfer of electrons across the interface, may also contribute to attractive
forces which bond the ﬁber to the resin. For polymers, these electrostatic attractive forces
are usually small when compared to other interactions and can probably be neglected for

most ﬁber-matrix systems.

 

8
Lastly, mechanical interlocking between the ﬁber and matrix may substantially

increaseadhesion bymaximizingtheﬁber-matrixcontactareaandprovidingmore
possibilities for chemical interactions and additional ﬁictional resistance to shear stresses.
Increased ﬁber surface rougbening, induced by ﬁber surface modiﬁcation, increases the
amount ofmechanical interlocking and adhesion. Adequate wetting, combined with the
minimization of voids, must be achieved or maximization of ﬁber-matrix adhesion will be
compromised. Furthermore, surface pits associated with surface roughening act as stress
concentratorsanddegradethemechanicalpropertiesoftheﬁber.

In addition to these mechanisms, the ﬁber—matrix coefﬁcient of thermal expansion
mismatch can create residual stresses and cracks drning cool-down after composite
processing degrading ﬁber-matrix interfacial properties. Weak boundary layers (WBLs)
consist of imprnities and low molecular weight oligomers on the ﬁber surface. WBLs
have lower cohesive or adhesive properties than their bulk substrates and serve as weak
interfacial links. Removal or strengthening of WBLs is beneﬁcial to achieving high levels
of adhesion.

The interfacial shear strength ([88) characterizes the adhesion of the ﬁber-matrix
interphase. Thethreemostcormnon methods fordetermining theISS are the single ﬁber
fragmentation technique [12,13], the microdroplet pull-off technique [14,15] and the in—sr’tu
microindentation technique [16]. For the case of UHMW-PE ﬁbers, the strain to failure of
theﬁberis toohigh tobetestedinthesingleﬁberﬁagmentationtest. Furthermore, the
inherent polishing difﬁculties of polyethylene ﬁbers, combined with the ﬁber’s viscoelastic
nature, discourages the use of the microindentation technique. The microdroplet pull-off
technique is therefore the appropriate test for UHMW-PE ﬁbers.

Studies by Herrera-Franco et al [17] and Drzal and Herrera-Franco [18] analyzed
and compared these techniques. They show that loading conditions of the microdroplet
may affect the interfacial stress distribution in the case ofthe microdroplet pull-off
technique. Furthermore, differences in the shape of the droplet meniscus and the length of

 

 

 

 

 

9
theﬁbersupportingthedropletontheinterfacialsmesseshelpexplain thefrequently
observed scatter of experimental results. Also, Gaur and Miller have investigated the
microdroplet pull-off experimental parameter effects on the interfacial shear strength of
Kevlar, glass and carbon ﬁbers [19]. Their results indicate that the shear strength
dwmaseswiminaeasingﬁber-manixmtafaciﬂmealmgmmdmmdmismsigniﬁcmt

load rate effect on shear strength.

Ultra-High Molecular Weight Polyethylene Fibers

A ultra-high molecular weight polymer combined with a high degree of crystalline chain
orientation results in the achievement of high strength and high modulus ﬁbers. For
polyethylene ﬁbers, producing both ulna-high molecular weight and a high degree of chain
orientation is difﬁcult since the extremely long, flexible, chain polymer tends to tangle in
the melt.

To reduce the number of entanglements and increase the chain extension in
UHMW-PE ﬁbers, a unique drawing method developed by Cappaccio and Ward [20] is
currently used as a general production technique. This gel spinning process uses ultra-
high molecular weight polymers dissolved in a solvent which disentangles the polymer
chains. This network is then drawn in an oven to high draw ratios (1>>20) where
molecular chains are transformed M a folded-chain conformation into a nearly extended-
chain conformation as shown in Figure 2-2. The quenching of the extrudate leads to the
formation of a gel ﬁber ﬁom which the solvent is either evaporated or extracted by a
second, low melting point solvent. General reviews of this process and a solid state
extrusion method are given by references [21,22]. The cm'rent trade names and producers
of polyethylene ﬁbers are: Spectra (Allied-Signal Corp., Petersburg, VA), Celanese
(Celanese Research, New Jersey), Snia (Snia Fibre, Milan, Italy) and Tekmilon (Mitsui
Petrochemicals Ltd., Tokyo, Japan). Table 2-1 compares some mechanical properties of
Spectra-HID ﬁbers with those of Kevlar 49.

10

 

Conventional Fiber

a; /'
gig:

. Relatively low molecular weight

 

 

 

. Moderate oreintation

. Crystalline regions chain folded

Extended-chain fiber

 

 

 

 

 

 

 

 

Q Very high molecular weight
0 Very high degree of orientation
. Minimum chain folding

 

 

Figure 2-2: Schematic representation of conventional polyethylene
ﬁbers and chain-extended polyethylene ﬁbers.

 

 

1 l
The supermolecular structure of UHMW-PE ﬁbers consists of either a shish-kebab
morphology or smooth microﬁbrillar morphology. Recently, Hofmann, et a1. [23]
investigamd, via transmission electron microscopy, the ﬁber surface morphology of

Spectra 900 and Tekmilon using a two—stage replica sample preparation technique. The
TEMmicmgraphsrevealthatﬂreSpecnaﬁbersweremainlycomposedofsmooth
microﬁbrils with negligible lamellar overgrowth. In contrast, the Tekmilon ﬁber had a
distinct shish-kebab structure. The smooth microﬁbrils observed in the Spectra were
believed to be a result of the high degree of drawing resulting in the unfolding of lamellar
crystallized molecular ends.

The segregation of impurities and low molecular weight species to the ﬁber surface
during ﬁber manufacturing produce WBLs. Several investigators have studied the effects
of molecular processes and impurity effects on the mechanical properties of the UHMW-PE
ﬁbers. For instance, Kalb and Pennings [24, 25] investigated the inﬂuence of the
microporosity and maximum strength of high molecular weight polyethylene ﬁbers on the
drawing process and concluded that drawability of polymer ﬁbers is inﬂuenced by the
mobility of the polymer molecules. Inhomogeneities inherent in the ﬁbers included
variations in the ﬁber diameter, polymer concentration and orientation, and the inclusion of
segregated impurities acquired during the spinning process.

Barham and Keller [26] systematically studied the crystallization conditions and
drawing conditions of polyethylene ﬁbers. They used single crystal mats of PE and found
that the most stringent condition for achieving high moduli was the minimization of low
molecular weight species inherent in the starting material. They also found that impurities,
in segregated form, were required to maximize the draw ratio. Thus, to produce ﬁbers
withoptimalmechanicalproperties, strictcontrol overthenattneandamountoflow
molecular weight species is essential.

'at 11..
r. .

 

 

 

 

12

Table 2-1: Mechanical properties of Spectra-10m and Kevlar-49 ﬁbers.

 

 

 

Kevlar 49 125 3.4 1.47 0.39-0.48 . _

 

 

 

 

 

 

 

 

FiberDiameter=30rnn
M" = 10‘

 

 

 

   

 

 

 

 

13

Traditional Surface Treatments

Plasma and Chemical Treatments Traditional methods for altering the smface and near-
surface properties of polymers and, in particular, UHMW-PE ﬁbers include chemical,
corona and plasma treatments. All have met with varying degrees of success, but the
overwhelming method of choice, with economical feasibility and adhesion improvement, is
plasma treatment. These methods modify the polymer swim in similar ways. Since the
physical processes occtn'ring during plasma treannent are closely related to those of ion
implantation, most of the literature review focuses on plasma treatments.

Liston [27] recently published a review article on the inﬂuence of low-pressure
glow-discharge plasmas on the surface properties of polymers, and a description of the
necessary equipment for plasma production. Though he does address the issue of the
bonding characteristics of the treated polymers to other materials, he does not speciﬁcally
treat the issue of adhesion between polymer ﬁbers and epoxy matrices. The two major
physical effects of plasma surface treatment of polymers are microablation and surface
cleaning. Crosslinking, or chain scission, and srn'face activation are the result of chemical
processes occurring in the polymer during plasma processing. The predominance of one
process over the other is manipulated by varying the process conditions.

Microablation denotes removal of the host material, whereas cleaning is the removal
of an unbound layer of contamination. Microablation and sm'face cleaning both remove
organic material by volatilizing low molecular weight hydrocarbons formed ﬁom either
oxidative or other free radical induced degradation reactions. For crystalline polymers the
ablation rate differs between the amorphous and the crystalline regions, resulting in a
preferentially etched surface.

A study by Holmes and Schwartz [28] investigates the plasma treatment of ultra-
high strength polyethylene (Spectra 900) fabric using an ammonia gas to attach surface
amine groups. Watercontact anglesdecreased to zero afteronlyS minutes in theplasma,

>J’A

 

 

 

14
rendering the fabric surface completely wettable. The addition of primary and secondary
aminegroupstotheﬁbersurfaceincreasedthewetting. Theyalsomeasmeda 10%
reduction in ﬁber tensile strength.
The tensile strength and the modulus ofUHMW-PE ﬁbers increases as the draw
ratio increases. Ladizesky and Ward [29, 30] have extensively studied the eﬁ‘ects of
chemical acid and plasma treatments on the adhesion of UHMW—PE ﬁbers to epoxy

matrices on different draw ratio ﬁbers. They found the acid treatment is most effective for
theintanediatedrawratiomaterial, whereastheplasmatreatmentismosteﬁ‘ectiveforthe
higher draw ratio ﬁbers. The maximum pull-out adhesion was found to be 4.9 MPa (a
100% increase over the untreated ﬁbers), but at the expense of a 42% decrease in tensile
strength. Scanning electron micrographs indicated a cellular structure on the ﬁber surface,
with the pit size increasing with increasing draw ratio, accompanied by a decrease and even
disappearance of the ﬁbrillated sm'face. They assert that the pull-out failure occurred by
peeling-offofthe ﬁlaments and took place in two stages, namely initiation and propagation
of the fracture.

A microdroplet technique was used to determine the interfacial shear strength [31]
of plasma surface modiﬁed UHMW—PE Spectra 1000 ﬁbers. An SEM investigation of the
ﬁber surface indicated that uniform cellular microcavitation of the Spectra ﬁbers occurred.
The interfacial shear strengths ofthe Spectra ﬁbers increased for all the plasma treatments
irrespectiveofthecherniealnatureoftleionizedgas. They hypotlesizedthattexturingof
the ﬁber surface improved the adhesive bond strength.

A recent study by Tissington, et a1. [32] investigated the effects of oxygen plasma
treatments on the mechanical properties of different types of ultra-high strength
polyethylene ﬁber composites. They found that the interlaminar shear strength increased
(79%) by preheating the reinforcement with an oxygen plasma. Composite mechanical
propertiessuchastensilesn'engthandmodulusandtleultimateﬂexmal strengthand
flexural modulus were shown to be largely independent of the plasma treatment.

 

15
one: investigators [33] studied the effects of surface ﬂuorination on the adhesion
ofSpectra 1000 ﬁberstoepoxymatrices. Contactanglesdeterminedthewettabilityoftle
ﬁbers, while XPS and infrared spectroscopy elucidated the surface chemistry. The
increase in ﬁber surface wettability, along with surface roughening, contributed to the
400% increase in interfacial shear strength. The role of the oxygen incorporation during
ﬂuorination on chemical bonding needs further study.

Electron and Photon Irradiation Other treatments, such as electron and photon
irradiation do not directly improve ﬁber—manix adhesion, but improve the mechanical
properties of UHMW-PE ﬁbers. A series of studies by Chen and Ranby [34, 35, 36]
details the effects of photocrosslinking on the mechanical properties, high temperature
resistance and creep behavior of Spectra 900. No appreciable changes in tensile strength or
modulus appeared for ﬁbers with a gel content ofup to 91% at room temperature.

Klein, et a1. [37] studied the structure and properties of electron irradiated UHMW-
PEﬁlamentsbeforedrawingasafunctionofirradiationdoseinanattempttodecreasethe
creep rate of the drawn ﬁbers. Gel content, differential scanning calorimetry, elasticity and
x-ray measurements, along with creep measurements, indicated that extensive chain
branching and formation of a permanent network occurred as a result of the irradiation.
Thetensilecreeprateisdecreasedbyforn'ordersofmagnitudeandisdescribedwellin
terms of two thermally activated processes, one crystalline and one amorphous, acting in
parallel.

Ion Implantation of Polymers

General Considerations Ion implantation is the introduction of dopants (impurities or

alloying elements) into the surface layers of solids by energetic ion beams. Stopping of

these energetic ions involves the production of a tremendous amount of damage (vacancies,

 

16
interstitials and amorphization) by collisional processes and bond breakage. The induced
damage alterstlrechernical, electricalandmechanicalpropertiesofthe hostmaterialsin
beneﬁcial ways. To date, the eﬁ‘ects of ion implantation on all classes of materials (i.e.,
metals, semiconductors, and insulators) have been studied [38, 39]. The ion implantation
of polymeric materials, a subset of the insulator class, has only recently (within the past 15
years) been aggressively investigated, due primarily to their importance in the electronics
industry. Several review articles concerning ion implantation of polymers have recently
been published [40,41].

Theenergydepositedﬁomtheincorningionstothe target accounts formostofthe
phenomena observed during and after ion bombardment of polymers. The energy loss is
phenomenologically partitioned into electronic interactions (ionization and excitation of the
host material atoms) and nuclear (hard sphere) interactions. The process by which a
particle loses energy when it moves through a medium consists of a random sequence of
independent events between the moving projectile and an electron a nucleus attached to an
atomoftlre solid. Foratargetthatcontainsmorethanoneelement, theenergylossistlre
sum of the losses of the constituent elements weighted by the abundance of elements. The
probability of interaction between the incident ion and the atom of the solid is described by
the stopping energy cross section, 8 .

The energy loss of an incident ion can then be characterind by Bragg's rule
(Equation 2— l) for the stepping power cross section[40].

e(polymer) = ne(I-l) + me(C) + ..... 2-1
wherenandmareﬂenumberofHandCatomsinthemonornerandeGDisthestopping
power cross section for hydrogen atoms, and e(C) is the stopping power cross section for
carbon atoms. Bragg's rule allows the energy loss, dE/dx (eV/A), of the accelerated ion to
be written in terms ofthe energy deposition per unit length along the ion track,

dEde = pNae/PM 2-2

S"?—

 

 

17

where N. is Avogadros number, p is the mass density (gm/cm3) and PM is the molecular
weight of the monomer . The depth distribution of energy loss along the ion path is
variable, withthemaximumvalueoftheenergylossparameter,dFde. appearingattheend
oftheionpatlr wheretheionvelocityisquitelow.

Thelarge amountofenergylostattheendoftheion travel resultsinasigniﬁcant
quantity of atomic displacements and radiation damage. Electronic interactions produce
intense chemical effects within the irradiated polymer, while nuclear interactions produce
point and linear defects.

Other useful parameters used in the analysis of implanted polymers are the total ion
range, deﬁned as the length of the ion track, and the projected ion range, which is the
length ofthe ion track perpendicular to the target stn'face. In the majority ofiorr
implantation studies, the projected ion range is the parameter of interest, since it determines
the lateral region of surface modiﬁcation. The number of ions striking the target per unit
area is deﬁned to be the ion dose.

The energy lost by the ions during the process of slowing down results in the
formation of excited states (free radicals), metastable molecular states and an ensemble of
other smaller molecules, many of which may be volatile. This process is fast (<10'13 s)
when compared to chemical reactions (~10'6 s). Strong gas evolution (molecular emission
orsputtering)duringionbombardmentlmsbeendetectedinalmostalloftleanalyzed
polymeric ﬁlms. Most of the ejected particles are small neutral molecules such as hydrogen
oracetylene with otherdetectable ionized species about twoordersofmagnitude lowerin
concentration [40]. This irreversible radiation damage cannot be annealed out since most
polymers are unstable at high temperatmes.

Puglisi [42] classiﬁes the modiﬁcations induced by high energy ion bombardment
into reactions giving products which do not contain the projectile and reactions between

projectile and target molecules, giving products which incorporate the projectile. The

 

18

end-of-track reactions occur when the projectile energy is about 20 eV, while the reactions
producing products not containing the projectile are present all along the ion track.

Puglisi also divided the ion dose into high dose ( >1014 cm-z) and low dose ( <
1013 cm-z) regimes. In the high dose regime, the bombardment of polymeric materials
pmducesamerphwscarbonacwusmataialalmosthrdependentofthenaumeofde starting
material. The progressive loss of volatile species and the continuous bond rearrangement
under ion bombardment induces the formation of stable carbon structures (amorphous
hydrogenated carbon). As a ﬁrst approximation, at high doses, the target loses memory of
its original structure and stoichiorrretry.

In the low dose regime, the overlap between ion tracks can be neglected and the
correlation between the energy input parameters, the molecular parameters of the target and
the observed phenomenology becomes clearer, allowing basic ion-solid research to be
performed.

Recent studies indicate that increasing the ion energy to a few keV generally
enhances radiation damage effects without signiﬁcantly changing the products [43, 44].
These studies also suggest that ion energies in the range of 500 to a few keV, produce near-
surface chemical products similar to those produced by 100 keV ions. Furthermore, since
ion energies in typical reactive plasmas range from a few eV's to several hundred eV's,
similar ﬁber surface modiﬁcation effects by high and low energy irradiations are expected.

As a group, the physical and chemical changes present in ion implaned polymers
inﬂuence and/or modify the polymer stoichiometry, molecular weight and polymer
solubility. Molecular emission occurs for all irradiated polymers and many times has a
profound effect on the polymer stoichiometry [40]. For instance, most ion irradiated
polymers show large reductions in hydrogen concentration, with reductions up to 80%
detected for ion doses in the range of 1013 to 1015 cm'2, for heavy and light ions,
respectively. Other non-carbon elemental constituents of the polymers, such as oxygen,
ﬂuorine and chlorine behave similarly. No large amounts of sputtered carbon have been

Hm
- - WWW" .,'

z

“9"“ U4 '

5.3.: ... .

 

19

observed. For the particular case of polyethylene, the molecular emission that occurs
during ion irradiation contains the species H2, CH4, C2H2 and C3H5 [40].

Bond breakage and reformation drning ion energy deposition induces chemical
changes, resulting in crosslinking or chain scission of the polymer backbone. Though
chain scission and crosslinking may both take place dtning ion bombardment, one process
usually dominates. The crosslinking process results in molecular weight increases and the
formation of a three dimensional network between the main chains. Chain scission, on the
other hand, results in molecular weight decrease with increasing ion dose as shown in
Figure 2-3. Polymers with a low heat of polymerization and high monomer yield during
pyrolysis tend toward chain scission processes during irradiation, while crosslinking
occurs in polymers with high heat of polymerization and low monomer yield [40]. Since
the heat of polymerization for polyethylene is 22 Kcal/mol and the monomer yield on
pyrolysis is 0.025 [40], crosslinking reactions dominate during the irradiation of
polyethylene.

Since polymers are formed from many thousands or millions of monomer units,
small numbers of bond changes, produced at relatively low doses (1013 to 1015 cm'2), are
enough to considerably modify the polymer viscosity or solubility. In general,
crosslinking decreases polymer solubility, while chain scission increases the polymer
solubility. For UHMW-PE, which has low solubility in most solvents, the irradiation
induced crosslinking increases the solubility making the polymer more environmentally
stable and heat resistant.

Examples of Irradiated Polymer Studies The study of ion-polymer interactions
provides insight into the physical and chemical irradiation induced effects which inﬂuence
fiber-matrix adhesion. Recent reviews [45, 46] examine the ion-beam degradative
processes occurring during irradiation. The cause of these processes are not well

understood [47, 48].

transverse: ?' a ' 3w: .f‘x‘ng', .

 

 

 

 

 

 

 

 

 

Figure 2-3: Schematic illustration of chain scission and crosslinking of

polymeric ﬁbers.

"mm r‘m rm“? “‘1‘“ ‘5'3‘ ‘5 w" V t J .l”

 

 

 

 

 

 

21
Koul, et a1. [49] studied the energy transfer mechanisms and the radiation chemistry
of noble gas ion irradiated polymer ﬁlms (Kapton, Teﬂon, mylar, cellulose nitrate, and
CR-39). Studies as a function of ion energy and dose elucidated the chemical and physical

effects of the irradiation. SEM micrographs suggested that, in these polymers, the gas
atoms form bubbles by coalescence, leading to stress and cracking after prolonged
irradiations.

For He ions in Teﬂon, cracking developed at doses near 3x10l4 cmi In contrast,
for Ar ions in Teﬂon, the critical dose is two orders of magnitude higher. Also, they
observed that the irradiation induced surface topography depends primarily on the radiation
sensitivity of the polymer. They concluded that, for Teﬂon, basic electronic energy transfer
mechanisms are responsible for the development of the surface topography, quite in
contrast to Kapton where nuclear mechanisms seem to dominate.

The modiﬁcation of polymers by ion irradiation occurs predominantly in the region
of the ion's implant depth. Studies of ion range proﬁles in irradiated polymers indicates
discrepancies between the theoretically predicted ion range and the experimentally observed
results. In one study by Fink, et al. [50] the depth proﬁles of light ions (Li, B and F1)
implanted into organic polymer ﬁlms were investigated. They found that, at a certain
threshold implantation energy, there is a transition from a regular ion proﬁle (as predicted
by TRIM [51]) to an irregular one. This transition occrn's whenever the electronic stopping
power starts to dominate the energy transfer process. They note that the variation in the
implanted dose did not affect either the range distribution parameters or the proﬁle shapes.
This occurs even though, at doses above 1x1015 cm‘z, polymers exhibit drastic changes in
their compositions.

Systematic measurements were carried out on the optical absorption depth proﬁle,
and the particle mass and dose dependence of the optical absorption. They show that the
blackening proﬁle resembles the vacancy proﬁle and not the proﬁle of the electronic energy

 

22
transfer (ionization proﬁle). This suggests that the discoloration results from nuclear

interactions.

Ion irradiation of polymers also affects the surface mechanical properties either
through crosslinking or chain scission reactions. Lee, ct al. [52,53] measured the
irradiation induced hardness changes of several different polymers (Kapton, Teﬂon, and
Mylar). Residual gas analysis dining irradiation identiﬁed the liberated gases as mostly
hydrogen, carbon monoxide, carbon dioxide, acetylene and hydro-ﬂuorocarbons. Teﬂon
suffered the most decomposition, as evidenced by the largest pressure increase during
irradiation. Substantial increases in the nanoindentation hardness of all the irradiated
polymers were found, with Kapton having a hardness 3 times that Fe-C‘r-N i stainless steel.
They assert that the irradiation induced three dimensional crosslinking state increased the
surface hardness of the polymer. The implanted materials also had lower wear rates than
unirradiated polymers. In general, higher outgassing rates resulted in lower hardness
values and higher wear rates.

Ion implantation of polyethylene has been investigated by several groups. For
instance, Ishitani, et a1. [54] used various analytical techniques (SIMS, FI‘IR, Raman and
ESR) to characterize oxygen implanted polyethylene. Depth proﬁling by SIMS shows a
monotonically decreasing oxygen content for an ion energy of 80 keV at a dose of 5x1014
cm"2 that, they suggest, results from the diffusion of the implanted oxygen out of the
polymer. In contrast, 5x10“S cm‘2 irradiated samples had the distribution predicted by
projected range theory. They identiﬁed hydroxyl, carbonyl, acid anhydride and
unsaturated hydrocarbons, including aromatic rings, on the polymer surface. The relative
amount of aromatic structure increased with dose, whereas the other functional groups
decreased with dose after reaching a critical dose. This correlates well with the monotonic
increase of amorphous carbon as detected by Raman spectroscopy. The depth proﬁle of
the generated functional groups indicated bound oxygen in the modiﬁed layer as oxygen

containing organic compounds.

 

23
Oxygen-implanted polyethylene had a high carbon radical concentration, (on the
order of 1x1015 spins/cmz) which saturated at a dose of approximately 1x10“s cm‘2.
Aliphatic ﬂee radicals formed in the beginning of the irradiation were unstable, but the
conjugated structure stabilized over a period of several months.
Schaible, et a1. [55] also studied the induced chemical changes and dielectric

properties of polyethylene irradiated by high energy ion implantation. At doses above
3x10ls cm'2, no appreciable increase in the number of ﬂee radicals occurs, even though
ions are continually being implanted. The stabilization of the ﬂee radical concentration at a
certain dose level indicated that either no new free radicals were being formed or new free
radicals react with the ﬂee radicals already present resulting in an equilibrium free radical
concentration. The dose rate had no effect on the ﬂee radical concentration.

They also found that higher doses of 0+ ions irradiations were required to generate
the same concentration of ﬂee radicals as Ar“ ions. This is consistent with the fact that
heavier ions generate more damage in the polymer than lighter ions. Increasing the
temperature of the sample dming irradiation results in enhanced radical formation. This can
be conceptually understood since at lower temperatures the polymer chains are relatively
rigid and the ﬂee radicals formed by bond cleavages exist in a relatively well deﬁned cage
where recombinations or rearrangements are more likely. At higher temperatures, the
polymer chains are more mobile and bond cleavages may result in chain displacement as
well. The free radicals formed may be moved to an environment where further reaction is

not energetically favorable.

 

 

 

 

Chapter Three
Experimental Methods

Sample Preparation
Prior to sample irradiation, Spectra-1000 ﬁbers and fatxic (Allied Signal Corporation,
Petersberg, VA) were mounted in a rectangular aluminum ﬂame and cleaned by soxhlet
extraction in absolute ethanol (99.99%) for 24 hours. Though this procedure cleaned any
physisorbed impurities or dirt loosely bound to the ﬁber surface, remnants of the
commercially applied UV stabilizers and antioxidants were found by x-ray photoelectron
spectroscopy. Soxhlet extraction in cyclohexane for 6 hours removed these additions. The
cleaned ﬁbers were then wrapped in acetone scrubbed aluminum foil and place in a
dessicator at room temperature. Oxygen plasma- and corona- treated ﬁbers processed by
Allied Signal Corporation, Petersberg, VA, were cleaned and stored in the same manner.

Both ﬁbers and fabric were prepared for ion irradiation by individually mounting
approximately 100 ﬁbers or several 100 mm x 0.5 mun fabric strips, with the use of
double-sided sticky tape, on a 75 mm square aluminum ﬂame. The ﬂame was attached to a
specially constructed aluminum ﬁxture that kept the ﬁbers and/or fabric in contact with a
chilled curved aluminum heat sink as shown in Figure 3-1. The ﬁxture also allowed the
ﬂamestobeinvertedsoboth sidesoftheﬁberscouldbeirradiatedduringseparateruns.

High energy ion irnplantations were performed in a medium current implanter, with
beam currents of approximately 20-30 [LA at the Spire Corporation, Danvers MA. Ion
species, accelerating potentials, total ion doses and computed implant depths of the high
energy ion implantations are shown in Table 3-1. Tie irradiation parameters were chosen
so that a wide variety of ion masses would be represened. The ion energies were selected
to correspond to approximately the same projected range. After irradiation the samples
were kept in a dessicator.

Low energy Ar+ ion irradiations were carried out with a 3-cm diameter Kaufman-
type ion gun (Ion-Tech, Denver, CO), equipped with diverging beam optical grids,

24

 

Fabric

 

 

 

 

Figure 3—1: Fiber and fabric lay-up procedure.

 

26

 

Table 3-1: High energy ion implantation parameters.

 

 

 

 

High Energy
Ion Implantation Parameters
Ion Energy Dose Implant
Species keV ionalcm Depth, run
Fibers:
N+ 30 5x1014 110
N” 30 1x1015 110
N* 30 5x10 15 no
N+ 100 1x1015 380
1
' A? 75 1x10 ‘ 130
A? - 75 5x10 1‘ 130
A? 75 1x1015 130
Ti+ 100 1x1015 135
4.
Na 40 1x10 15 ‘ 110
+ 15
Cl 80 1x10 145
Fabric: .
+ 15
N 30 1x10 110
N” 30 1x10 1‘ 110
N+ 30 1x10” 110
'ri+ 100 mo15 135
+ 15
Xe 135 1x10 90

 

 

 

 

 

 

 

27
mounted inside an in-house cylindrical vacuum chamber (35 cm high and 20 cm diameter)
pumped to a base pressure of 6.6x10'4 Pa as shown in Figure 3-2. The ﬁbers were

irradiated to doses ranging ﬂom 1x10ls to 1x1017 cm“2 at various accelerating potentials
between 100 and 1000 eV. Ion current densities ranged ﬂom 0.130 to 0.341 mA/cmz, as
measuredbyabiasedFaradayprobelocatedonashutterusedtoshieldthefiberarrayprior
to irradiation. The ion source to specimen distance was approximately 16.5 cm. Under
these conditions, processing times ranged ﬂom 3.5 to 220 seconds. Processing parameters
are listed in Tables 3-2 and 3-3.

Mechanical Tests
The tensile strength of both irradiated, plasma treated, corona treated and control Spectra
1000 ﬁbers was measured using a United Testing Systems Model SFM-20 screw driven
machine. The ends of individual ﬁbers were attached to kevlar/epoxy composite grips
(approximately 12 mm x 12 mm) with DER 331 epoxy cured with diethylenetriamine
(DETA) using a 10/1 ratio. The adhesive was allowed to cure at room temperature for
approximately 1 hour (until tacky). Binder clips were used to press the grips together for
the remaining 24 hours. The 25 mm gage length coupons were tested at a crosshead speed
of 2.54 mm/min. Prior to testing, an average ﬁber diameter was calculated from four
diameter measurements made with an Olympus Micro Cue-300 video caliper. The tensile
strength of the ﬁber was then calculated by using
s = F/A 3—1

where F is the load on the ﬁber at the failure and A is the ﬁber cross-sectional area.

Both treated and untreated ﬁbers were composited with Epon 828 microdroplets so
the ﬁber/matrix interfacial shear strength could be determined. Stoichiometric Epon 828
(diglycidyl ether of bisphenol-A, Shell Chemical Co.) with meta-phenelynediamire curing
agent (m-PDA, 14.5 phr) epoxy microdroplets were deposited on individual ﬁbers
suspended across an aluminum ﬂame by sliding a resin ﬁlled pipette along the length of the

 

28

 

3chan£manGrm

féfrtihf-C-‘m

 

Fibers and Fabric
Manned on Aluminum Frame

Water Cooled Sample Holder

 

 

 

Figure 3-2: Illusnation of low energy ﬁber irradiation arrangement.

 

29

Table 3-2: Low energy ion irradiation parameters, at a constant energy of 1 keV.

 

Table 3-3: Low energy ion irradiation parameters, at a dose of 1x1016 cma.

 

 

cram tame ocean.“ mum rounoracreo
DENSITYW “Mite-d («'5 (eV) mortar-t

 

 

 

 

 

 

 

 

 

0.124 22 1 I” 1.4
0.217 10 1 250 2.5
0.316 . 3 . 1 5w 3.5
0.267 ' 7.5 1 750 43
0341 5 1 1000 5.5

 

 

 

31
ﬁber. The microdroplets were cured with the ﬁbers in a horizontal position at room
temperaturefor 18 hoursandthenpostcmedforZhoursat348 K. Thiscuring schedule
ensured that no curing agent diffused out of the droplet during processing.
The microdroplet pull-out tests were performed using a blade micrometer attached

toamicroscope stageasshowninFigure 3-3. Acompositedﬁberwasgluedtoan
aluminum plate using a cyanoacrylate adhesive which was attached to either a 50 or 250
gram load cell. The length of the droplet along the ﬁber axis, called the embedment length,
wasmeasmedwiththevideocaliperandrecorded. Adropletwasgraspedbythe
micrometerblades and the micrometermoved away fromthe load cell at arate of20
microns/min. A 2 gram preload was applied to the dropkt to facilitate ﬁber/tension axis
alignment. The load vs displacement data was recorded either electronically or with a strip
chart recorder. An average ﬁber/matrix interfacial shear strength, 1, was then calculated

from

”pf/(now) 3-2

where PfiS the load at failure, (1 is the ﬁber diameter and l is the embedment length The
value of Pf , for each ﬁber treatment, was determined from the slope of a linear regression
line constrained to pass through the origin of a plot of failure load versus embedment length
for a number of 12 droplets as shown in Figure 3-4. Scatter in the data (for a given
treatment) is believed to result from ﬁber surface heterogeneity [19], inherent weaknesses
of the test itself (i.e., loading conditions and ﬁber ﬂee length [17]) and/or difﬁculty in the
precise determination of the embedment area due to variable ﬁber radii. The ISS errors
were determined by assuming a 95% conﬁdence level. A minimum of 12 droplets were
tested for each surface treatment.

The inherent irregular cross-sectional area of Spectra 1(X)0 ﬁbers makes an absolute
determination ofthe ﬁberdiameterextremely difﬁcult Tominimizetheuncertaintyintle

 

32

 

MiICRODHOPLET PULL-OFF APPARATUS

Stage

 

Load Cell

Movable Stage
Epoxy Droplets

Blade Mlcrometer

 

 

 

Figure 3-3: Schematic diagram of the microdroplet pull-off apparatus.

 

 

 

Figure 34: Examples of pull-off load vs embedment length for 1 keV Ar’ ions.

 

34
188 valuesduetotheﬁberdiameter,ISSvalueswerenormalizedwithrespecttotheconuol
value. ISSvaluescitedweredeterminedwithanominalﬁberdiameterofwum.

Surface Chemistry Characterization

X-ray photoelectron spectroscopy (XPS) spectrawereobtainedon aPerkin ElmerPhi
SMESCA systemwithamonochromaticAlKasomee(l486 eV) usedtoobtainboth
quantitative (relative surface concentrations) and qualitative (functional grorrp) information
within approximately theﬁrst6nmofthepolyethyleneﬁbers. Theuseofthe
monochromatic source precludes sample radiation damage. All samples used in this study
were analyzed at ambient temperature and exhibited no evidence of x-ray damage during
measurements. The pressure in the spectrometer was approximately 6.6x10’7 Pa. Five
minute survey scans were taken for all treatments. A 15 minute multiplex scan on carbon
(2-3 cycles) and oxygen (6-8 cycles) was performed for quantitative analysis. The
implanted ion spectral peak was not observed.

Stu-face functional group chemistry characterization of the ﬁber surface was canied
out on a Perkin Elmer 1850 Fourier Transform Inﬂated Spectrophotometer using an
attenuated total reﬂectance (ATR-FITR) technique. A KRS-S (n = 2.38) internal reﬂecting
elementCIRE) withanangleof45°wasusedtoobtaintheinﬂaredspecmaaccordingto
AST'MI Annual Book 42, E573, 1984. Both treated and untreated ﬁber mats were purged
of any adsorbed water by continuously ﬂushing the ﬁbers with dry air in a glove bag for a
minimum of 1 hour. While in the glove bag, the mats (approximately 12mm x 40mm)
were mounted on both sides of the crystal so that the beam underwent approximately 25
reﬂections, with a penetration depth of 2 to 3 pm before striking a mercury cadmium
telluride detector (Figure 3-5). The sample holder was then placed in the sample chamber
which was subsequently purged with dry air for either 15 or 30 minutes.

Spectra were scanned 1000 times (approx. 6 min.) and stored in memory.
Background spectra ofthe IRE only were taken with the same exact purging and scanning

 

 

 

 

 

 

35
conditionspriorandsubsequentmspectraofﬁbermats. Thebackgroundspectrumwas
then subuactedﬂomtheﬁbermatspectrumresultingintheﬁnalspectra.

Physico-Chmrical Characterization

Surface energy was measured via static advancing contact angles for single ﬁbers using a
micro-Wilhelmy technique. Water, methylene iodide and ethylene glycol were used as
characterizing liquids, as shown in Table 3-4. Fibers, 15 mm in length, were attached with
cyanoacrylate adhesive to a nickel hook which was hung on a Cahn Rg electrobalance. Tie
electrobalance rncasm'ed the force on the ﬁber as it was raised and lowered in the
characterizing liquid (Figure 3-5). Approximately 3 mm of the ﬁber was immersed before
beingmovedinandoutoftheliquidatarateonS tun/s. Theﬁberswereallowedto
stabilize 20 seconds before the static force measurement was taken. The entire apparatus
was enclosed in an environmentally controlled chamber to eliminate wind currents, reduce
contamination of the characterizing liquid, insulate the apparatus ﬂom vibrations and
maintain constant temperature.

Contact angles were determined (neglecting buoyant force) according to
m0g=7u0 Pcoso 3-3

where m is the mass measured by the electrobalance, g is the acceleration due to gravity, P
is tie perimeter of the ﬁber and 71,, is the total surface energy of the probe liquid. The
average ﬁber diameter, 31.2 +/- 2.8 pm, was determined by assuming complete wetting (4)
= 0) using hexadecane as a characterizing liquid. A minimum of 45 measurements were
taken for each ﬁber treatment.

Surface energetic analyses of the ﬁber were determined from the equilibrium
advancing contact angles in the manner described by Kaeble, et a1. [56]. This method
assumes, for low surface energy solids such as polymers, that the total surface free energy
across the interface, 7LT: is composed of a polar, Yip, and a dispersive part, Yip;

 

36

 

 

'urmw-PE Fabric

0
45 KRS-s Crystal

Approx. Penetration Depth I 31m

 

 

 

Figure 3-5: Schematic diagram of attenuated total internal reflectance technique.

.‘ _ 14L .-:: '.‘_‘.L._ 'A‘,,‘

1*”..4...

.. ~. v- 4!. Q1. .6

 

 

 

 

 

37

Table 34: Surface energies of characterizing liquids for the miero-Wilhelmy
experiments.

 

Table 3-4

 

Surface Energies of Characterizing Liquids, rrrJInr1

Liquids FY“ . Y“ Y1;

 

 

Water . 72.3 " 21.8 51.8
Ethylene

Glycol 43-3 29.3 19.0
Methylene

Iodide - 50.8 48" 2-4

 

 

 

 

 

 

38

 

 

   

    

VIII/[IIIIIIIJ

 

 

Liquid

 

Fiber

  

Contact Angle

 

Figure3-6: Schemmtic diagram of the micro-Wilhelmy apparatus.

 

 

 

39

Tu - 7i! + 7m 34
where i stands for either the ﬁber or liquid. The work of adhesion is given by:
W. = 2[ Yanm 75.01” + 71331” 71.91/21 3-5
orintermsofthecontactangleas
W.=’YL1-[l+cos¢]. 3-6
This leads to:
W. [2 ‘YLDIIZ = Yppln + anm [Yum/YLDm] 3-7

which is expressed in a slope-intercept form. Measuring the contact angle for a minimum
of two liquids and plotting W. /2 'YLDm vs YLpln/leﬂ yields a straight line with
slope and intercept equal to Yul/2 and ’Ypplﬂ respectively.

Samples of irradiated fabric for gel-fraction determination were placed in acetone
washed 200 mesh stainless steel gauze. Extraction of the sol-fraction was performed by
suspending the gauze in boiling p-xylene, containing 0.5 wt.% of anti-oxidant (2,6—di-t-
butyl-4-methylcresol) for 24 h. The samples were then repeatedly dipped in fresh boiling
xylene until no polymer remants remained. The samples were deswollen in acetone and
dried under vacuum at 223 K for 6 h. The gel-fractions were determined by the weight
difference ofthe sample before and after extraction.

Microscopy

A Hitachi s-2soo scanning electron microscope (SEM), equipped with a L335 ﬁlament
operating at 10 or 15 kV, was used to obtain micrographs of the ﬁber and fracture surfaces.
For surface investigation, ﬁbers 10 mm long were mounted on a piece of double-sided
stickytape,placedon ascrewandsputtercoatedwith approximately20nmofAuto
provide a conductive link to the SEM. The screw was installed on a high resolution stub
whichcmﬂdbeﬁgidlyhckedmtheobjwﬁwpdepiwemmuumizeimagedegmdaﬁmdue
tovihration. Undertheseconditions, beamdamage began todegradethe ﬁbers afteronly

 

, we -‘-—‘ .v. ma)..." :.-..-‘;- 5' .7" v , ~ "

 

 

 

40
30—608econdsofexposureathighmagniﬁcation. Thus,afterfocusingandstigmation
waecmnpleedmespwimenwasmslaedshghdymrﬂmenﬁcmgmphremrdedfromm
undamaged section of the ﬁber surface. Resolution under these circumstances was
approximately 10 nm point to point.

Specimens oftested and untested microdroplets were attached to copper tape and
sputtercoatedwith approximatelyZOnmofAu. Thespecimens weretumedoverandthe
othersidecoatedwithAualso. Thetapedspecimenswerethenplacedonagraphiteholder
andmountedin the SEMforviewing.

Transmission electron microscopy (TEM) of the ﬁber-matrix interface was
accomplished by imaging room temperature ultra-thin microtomed sections of single ﬁbers
embedded in Epon 828 epoxy matrix as described in reference 57. The specimens were
imaged in a JEOL 1(1) CX TEM at an accelerating voltage of 100 kV.

Chapter Four
RESULTS and DISCUSSION

Thischapterisdividedinmtwomninsecﬁonscmrespondingwdetwodiﬁ‘aent
implantationenergyregimesstudied: highenergy(>30keV)irradiationsperfor-medina
medium current ion implanter, andlowenergyionirradiations performed bya3cm
diameterionKaufmantypeiongun. Thehighenergyimplantationexperiments,whichare
acontinuationofanearlierstudy,willbediscussedﬁrst

High Energy Ion Implantation of Spectra -1000

Visual Observations The most obvious effect of high energy ion implantation on Spectra
is the surface discoloration as shown in Figures 4—1a and 4-1b. The color change ofthe
fabric irradiated with 30 keV N+ ions as a function ofdose is illustrated in Figure 4-1a. No
color change occurs for either the 1x1013 or 1x1014 cm'2 irradiations (low dose regime),
while the 1x1015 and 1x1015 cm'2 (high dose regime) samples appear gold and dark
brown, respectively. In the high dose regime the darkening occurs because preferential
sputtering depletes the number of ﬁber surface hydrogen atoms leaving large amounts of
saturated carbon bonds in a graphite-like structure [58].

Figrue 4-1b illustrates the color differences for irradiations with different ion
species (Xe+, Ti+, N+, N‘) and different ion energies (135 keV, 100 keV, 30 keV and 1
keV respectively) at a constant dose of 1x1015 cm'z. Comparison of the 1 keV and 30 keV
N+, 1x1015 cm"2 implanted samples indicates a difference in color intensity. T‘le
diffaenceoccmsbecausedehighaenagyionproducesmoredamage(molecular
excitations and ionizations, combined with atomic displacements) altering the optical
propertiesofthematerialmoreextensively. Thedifferentionmassesandenergiesforthe
other irradiated samples make direct comparisons invalid, but the more massive and

41

42

mm

ONTROL

 

(b)

Figure 4-1: Photographs showing the discoloration of Spectra-1M irradiated.
asafunctionof(a)doseand(b)ion

 

43

energetic Xe+ and T’i+ implantations produce more radiation damage, resulting in a higher
level of carbonization.

Surface Chemistry An ATR-FI'IR investigation on irradiated UHMW-PE (in fabric
form) revealed a oxidized ﬁber sumface . The evolution of tle surface functional groups for
30keV N+irradiatedfabric,asafuretionofdose,isillustratedby thelepectrainFigure
4-2. TheprobedepthofATR-FI‘IR isZto3trm,muchdeeperthanthepredicteddepthof
irradiation induced surface modiﬁcation (approximately 135 nm). The data thus represent
both irradiated and unimadiatcd material characteristics. The control spectra shows
characteristic absorbances at 2925 cm'1 (CH stretch) and 1420 cm'1 (CI-12 scissors and in-
plane bending), seen also in the imadiated spectra. All specua exhibit a slight non-linear
baseline due to the large number of scattering centers in tle fabric. The absorbance peaks
below 14(1) cm“1 are due to the rubber pressure pads making intimate contact with the
reflecting crystal through the holes in the fabric.

Comparison of the control (unirradiated) and 1x10l3 N+ cm"2 dose spectra in
Figure 4-2 shows the appearance of absorbance bands at wavenumbers identiﬁed by the
letters A, B and C. The presence of these peaks at such low doses indicates the sensitivity
ofpolymers to ion beam irradiation when compared to ceramics or metals (where few
effects are seen at doses less than 1x10ls cm'z). The assignment of the functional groups
to these absorbance bands are made according to Gedde and Ifwarson [59] and Schaible, et
al. [55 ]. In Figure 42, the peak indicated by A, at 1644 ch, has not been conclusively
identiﬁed, but may be bicyclic quinone systems, which absorb in the region between 1635
and 1655 cm]. The peak at B (1704 cm'l) is due to carbonyl groups (C=O). At point C
is a broad band representing both hydroxyl (OH) and hydroperoxide (OOH) groups.

As the ion dose increases, small absorption structures develop into strong, broad
composite absorbance bands, indicated by letters C and D. These bands represent
extensive surface oxidation at the highest dose of 1x1016 N+lcm2. The carbonyl stretching

‘2 '53. “TIL-Toe"- z". '.‘f',.".‘..'"v' “.WMT-V-‘QM1YFW‘ ‘.- '0 i'Q-f - '

 

 

 

 

 

I

l

 

I

 

 

 

WWW

 

3:5 53 83:53:

I

l

1000

Iavenumber (cm")

Figrue 4-2: ATR-FI'IR specua for various doses of N+ ion implantation. .

 

 

 

r.--—_—-———““—“-—W ‘2 .gr ‘yr

 

 

 

 

45

hand, A, consists of three overlapping components which may be assigned to saturated
aldehydes (1733 cm'l), satmated ketones (1721 cm'l), saturated acids (1712 cm'l) and
possibly unsaturated hydrocarbons [60]. The poor speciﬁcity of these carbonyl groups is
inescapable because the separations between the peak positions are comparable to, or less
than, the half-wkidrsofdrethmepeaksandappreciably greaterthanthespectrometer
resolution (4 cm"). The permanent dipole moments of these surface groups produces a
polar molecules. At a dose of 1x1016 N'tlcm2 the large absorbance band at D changes
shape possibly due structural modiﬁcations such as formation of a crosslinking network,
severe hydrogen depletion and increases in aromatic ring su'uctmes, brought about by the
overall degradation of the polymer.

These results agree qualitatively with those of Ishitani, et al. [54] who studied 60
keV oxygen implanted polyethylene and reported a maximum surface oxidation level
(measured by ATR-FI'IR) at a dose of 5x1014 cm'2. They also observed that, as the dose
was increased above 5x1014 cma, most absorbance peaks decreased, except for an
increase in the concentration of unsaturated hydrocarbons.

Figure 4—3 shows the IR spectra for three different ion species (nitrogen, titanium
and xenon) with differing energies at a constant dose of 1x1015 cm’z. In general, the
structure of the bands at points A and B in these spectra are similar. The absorbance band
at point A in the Xe+ and Ti+ irnplantations appear closer in structure to 1x1016 Nr‘lcm2
irradiations than the 1x1015 Nr‘lcm2 samples. Since the Xe+ and Ti+ masses and ion
energy are greater than N+ ions, less radiation damage occurs during the NP irradiations
resulting in slight surface oxidation differences. With this qualiﬁcation, the degree and
type of surface oxidation is relatively independent of ion species. A plausible explanation
for this insensitivity can be made by considering the fundamental aspects of the ion-
polymer interactions. Upon irradiation, the ions lose energy by ionizing and exciting
molecules, creating both unstable free radicals (which recombine rapidly), and more stable
radical species. (The term free radical refers to any non-saturated fragment and does not

 

 

Transmittance

 

 

 

 

L I I l

 

soon 3000 3000 _tooeeoe

Figure 4-3:

Iavenurnber (curd)
ATR-FI'IR spectra for N". 11* and Xe+ ions at 1x10ls cm'z.

 

A“

‘ I ~ 9:. in".

- «‘1. m.

 

 

 

 

‘. .l [lllll ill}.

47
necessarily refer to spatial freedom ofthe ionized specie.) Species insensitivity has been
documented by Chipara and Wu [61], who determined that the nature of the radicals
muradiaedpdymersdependsmaiMyonthemagydeposiedbymeincidentparﬁcle
ratherthanonthetypeofincidentparticle. Thus,thesurfaceoxidationofionirradiated
Specma-ltxnmliesmmeﬁeeradicdgenaadmandmtmeexactdemﬂsofwhmimuawd
the production of these free radicals.

The source of oxygen has not been conclusively determined, though at processing
pressures of 5x10'6 Torr, low but ﬁnite oxygen partial pressures existed in the vacuum
chamber during implantation, making oxygen addition surface reactions possible. A
calculation of the impingement rate, R, of ()2 molecules on tie ﬁber surface during
irradiation follows from

R = p(21tMWkT)"°-s 4-1
where p is the oxygen partial pressure, MW the molecular weight of the gas, k is
Boltzmann’s constant and T the temperature [62]. For an oxygen partial pressure of 1x10“7
Pa at a temperature of 300 K the impingement rate is 2.7x10ll atoms cm'2 S'l. Since the
areal density of most solids is 5x10l4 to 3x1015 atoms cm'2 complete oxidation of the ﬁber
surface is not likely to occur before the irradiation chamber is bled to atmosphere.
Additional oxidation reactions then proceed from post-irradiation exposure to atmospheric
oxygen.

Further surface chemical analysis was perfumed using x-ray photoelectron
spectroscopy (XPS). Survey scans of UHMW-PE control fabric and 30 keV, 1x10ls
N+/cm2, irradiated fabric are shown in Figure 4-4 and Figure 4-5. The spectra show an
intense peak centered at 284.6 eV, due to tie carbon ls core level electrons (Cls) and one
centered at 532.0 eV for the oxygen ls core level electrons (018). The oxygen KLL Auger
peaks at 977.0 and 998.0 eV are also observed in these spectrum. Presumably the oxygen
present in tie control spectra is due to processing or commercially applied antioxidants and

 

die °8_.§8._m 38...... no 89. 3:8 3.5 < a... Bee

>0 .>ummzu uzaazan ,
new gum cam ocx oaw can ace“

h
a

«u-
q-
d

-
dl

1-
q-

 

 

 

 

 

+-
‘-
db
di-
--
«u-
ur-
dr-
"P
-D
-r
Th
“F
..

3/(3)N

«49

«as as: no as. a 9 3335 82 .2 >2. 8 .3 Bi 2.3 83.58%... as at... 3.5 < a... 2.5m

.6. .Ewm .22..

 

 

 

 

a emﬁ now can om? mam emu @QN cam new cam"

I n h .b h b b P b P b P n n b p h b D b

u u u q q u 1% 1 d d 1d d J H d
1- ..
LI ll
LI 1'
'I -I
ll .-
II II
II I
.. 1.
II 1'
L7 1...
I;- ll
II [I
ll ll
Irv ICI
JI if
II 1!
IT. it
It ......
1.! II

;p _ . _ . . . P p p p p p . u . p p .

u q u d d a 1 d u q d 1 J. d d u q I -

 

 

 

 

a“

3/(3)N

 

50
UV stabilizers not removed by the ethanol wash. No presence ofthe implant species was
found in the survey scan of the irradiated samples.
High resolution scans of the Cls region for control and fabric implanted by 30 keV
N+ to a dose of 1x1015 cm'2, shown in Figure 4-6 and Figure 4.7, were undertaken to

ftmther investigate the type chemical bonding occurring on the ﬁber surface. The control
Cls peak is relatively narrow (~1.75 eV FWI-IM) centered at 285.0 eV, but slightly broader
than that expected for pristine polyethylene, possibly due to small amounts of oxygen-
carbon bonding. The Cls peak for the N+ implanted specimen is asymmetric and broader
toward the higher binding energies, providing evidence for higher oxidation states of
carbon such as alcohol/ether (286.1 eV), peroxide (286.6 eV), carbonyl (287.6 eV) and
ester/carboxyl groups (290.0 eV). Line-shape analysis, using a deconvolution routine,
was performed on the high resolution scans to more exactly determine the chemical state of
the oxidation, but no reasonable ﬁts were obtained.

The relatively narrow width of the Cls peak indicates that the irradiation altered the
surface chemistry, but not sufﬁciently to produce signiﬁcant amounts of C-0 type bonds at
the higher binding energies. These results correlate well with those results reported by
Lanauze and Myers [63] for corona treated polyethylene. Other corona discharge treatment
studies [64] indicate large numbers of carbon-oxygen functionalities. Furthermore, our
results are also quite different from those obtained by an XPS study for argon, nitrogen
and oxygen plasma treated polyethylene, where extremely broad Cls peaks revealed the
highly oxidized nature of the surface [65]. Thus the degree and type of surface oxidation
varies widely for different processing techniques and processing parameters.

Figure 4-8 shows the oxygen:carbon composition ratio of UHMW-PE fabric
irradiated by 30 keV N+ ions as a function of ion dose, determined by XPS. The surface
oxygen concentration increases signiﬁcantly from the control value of 3.5 at.% to a
maximum value of 16.5 at.% for an ion dose of 1x1015 on2 then decreasing 11.2 at.% for

the highest dose of 1x1015 cm'2. Schaible, et al. [55] has reported that an equilibrium free

  

6.53 gen—~88 3825 8.. 838 :0 o... he 58 cog—82 .32 < 64. am

- - as'ausu-J- .n up.-¢:W\'v.~d.udouodow.- . ‘ -

>0 .>uumzm uzmazmn

 

up
up

n
1

51

 

ill

 

 

 

un-
-l-
A.
cu-
-u-
qr-

 

3/(3)N

 

 

52

«-80 22.: we once a 9 “:2 +2 >8. on 3 as“? 0.53 82-98% .8.“ SEE 20 05 we 53. eons—88 .3: < 8.? ”Suﬁ

. >0 .535 95!:
RN New v8 8w «8 .. 8N NR v8 8m 8w 8m

- — —. . . h p — L! b - b p _ _ — - -
u - - - - - u q - - < - u q - - - ‘ @

 

 

 

 

3

 

3/(3)N

 

....v, tr. 3"“. ;

 

 

 

 

 

53

 

O/C %

 

LOG” Dose, cm4

 

 

 

Figure 4-8: Oxygen:carbon ratios as a function of ion dose for UHMW-PE fabric
irradiated by 30 keV N+ ions.

 

54
radical concentration (the value of which depends on the speciﬁc irradiation conditions)

plateaus at doses above approximately 3x1015 cm'z, independent of ion type and energy.
Theincreasedconcenmationofunsattmatedaromatic strucurres, atleastpartially, is
responsible for the decrease in oxygen containing groups.

SEM Observations The degree of mechanical interlocking between the ﬁber and matrix
is another important factor in the development of good interfacial bonding. Interlocking
tendstoincreaseas the stmfacemorphologyot‘theﬁberbecomesrougler. Figmes 4~9and
Figure 4-10 show a rather striking comparison between the surface morphology of Spectra-
10(X) ﬁbers in the untreated, ion beam modiﬁed, plasma and corona treated conditions (the
two latter representing commercially processed ﬁbers supplied by Allied-Signal Corp.,
Petersburg, VA).

Untreated ﬁbers, shown in Figure 4-9a, have a relatively smooth surface in the
axial direction of the ﬁber, with numerous loosely bound tendrils curling up from the
surface. After a 30 keV N+ irradiation to a dose of 1x1015 cm4, as shown in Figure 4-9b,
the tendrils have been removed, or else welded down to the ﬁber surface, but the surface
remains smooth. The welding effect probably results from thermal effects of the beam,
which may cause greater local heating in the tendrils due to their larger thermal separation
from the heat sink. The effect would be magniﬁed by the low thermal conductivity (0.15 to
0.67 W m1 K4) and low melting point of polyethylene, and the lack of a convective heat
transfer route in the vacuum of the implantation chamber.

The untreated ﬁber and the ion beam modiﬁed ﬁbers have surface morphologies
whicharein sharpcontrasttotheplasmaandcoronaueatedﬁbers showninFigures4-
10a and 4-10b, respectively. Here the ﬁber surfaces are extremely rough, with
numerous slit-like cavities in the transverse direction; however, the slit-like cavities are
not uniformly distributed. This type of surface morphology should provide enhanced
mechanical interlocking (assuming that good wetting can be achieved), but the cavities

 

 

 

 

 

 

 

(b)

Figure 4-9: SEM micrographs of (a) control and (b) ion irradiated (1x1015 N+/cm2)
treated ﬁbers.

PLASMA ‘, lf-_\rxls.\\ru)€5. , K
. ~ will???" '
”a“
\» :- ‘

1""
li‘.l\ .3

h r

‘\

 

 

 

 

 

(b)
Figure 440: SEM micrographs of (a) plasma treated and (b) corona treated ﬁbers.

 

.' '. "y'w. ‘

 

 

 

3". mew-«es:

T. J- “'1‘?

 

57
can act as stress concentrating flaws, weakening the plasma ueaed ﬁber in tension. A

reduction of 12% of the control tensile strength was measured. (The tensile strength of the
corona treated ﬁber was not tested.) Microdroplct pull-off tests indicated 200% and 300%
increasesintheintafacialslearsuengthofplasmaandcormameatedﬁbersoverﬂratof
untreated ﬁbers.

Our results indicate that high energy irradiations only slightly altered the ﬁber
surface morphology, while the plasma and corona treatments radically modiﬁed the surface
morphology. The pitted ﬁber surface produced by tle plasma and corona treatments will
increase the amount of mechanical interlocking resulting in better adhesion. Changes
produced by high energy irradiations may not actually increase mechanical interlocking;
however, by removing and welding loosely bound tendrils and on the ﬁber surface and
inducing crosslinking in the implanted region, the cohesive strength of the ﬁber surface will
increase, resulting in increased 188 as reported by Ozzcllo et al [8]. Moreover, the
sputtering that occurred during irradiation would tend to remove any low molecular weight
species and weak boundary layers present on the ﬁber Sta-face, improving the chances for
increased adhesion.

TEM Observations Ultra-thin sections of control and ion irradiated Spectra-1000
ﬁbers, microtomed transverse to the ﬁber axial direction for imaging in a transmission
electron microscope, help document the locus and mode of ﬁber-matrix interfacial
failure. A large number of transverse unheated UHMW-PE ﬁber-matrix sections were
examined, with ﬁber dropout from the matrix (the most commonly observed feature)
suggesting extremely poor ﬁber-matrix adhesion. A typical example where the ﬁber
remained in the matrix is shown in Figure 4-1 la. A higher magniﬁcation view of the
ﬁber-matrix interface in Figure 4-11b shows that decohesion results in ﬁbrillated
structures (the tendrils seen in Figure 4-9a) pulling away from the ﬁber surface,
suggesting that the locus of failure occurs cohesively within the ﬁber.

 

 

 

 

 

(b)

Figure 44 1: TEM micrograph of a control ﬁber (a) low magniﬁcation
and (b) high magniﬁcation showing ﬁbrils pulling away ﬁom ﬁber.

 

59
Thepleated structureandoccasionalstriationsseenonthe section surface

indicate that microtoming conditions may have been less than optimal. Since our
interest lies primarily in the ﬁber-matrix interface pm, the pleated structures will not
affect the observed results. Figrme 4-1 la also indicates the highly noncylindrical nature
of these ﬁbers which varies widely from ﬁber to ﬁber and along the length of a given
ﬁber. This varying morphology is a direct consequence of the ﬁber production
technique as indicated by similar observations of Dyneema Sk60 polyethylene ﬁbers.

Figure 4~12a shows a bright-ﬁeld TEM micrograph of a Spectra-1W ﬁber
irradiated by 100 keV N+ ions to a dose of 1x1015 cm'2. The overall ﬁber matrix
adhesion has improved considerably over that of the untreated ﬁber as evidenced by the
ﬁber-matrix interface remaining relatively intact after the microtoming process. A
higher magniﬁcation view of the ﬁber-matrix interface, shown in Figure 4-12b, shows
the epoxy as a dark region beginning at the ﬁber surface and the interior of the ﬁber.
The dark region is believed to delineate the ﬁber region affecwd by the ion irradiation
and roughly corresponds to the calculated ion implantation depth of 383 nm. A
smooth, sharp boundary exists between the ﬁber surface and the epoxy matrix region
indicating a high level of ﬁber-matrix adhesion. Cracks produced during the
microtoming are visible in the implanm region and indicate that this modiﬁed region is
more brittle than the sumounding matrix and ﬁber. The higher contrast between the
implanted region and the ﬁber is the result of a change in either mass or electron density
from the virgin ﬁber. Since UI-IMW-PE crosslinks upon ion irradiation, both the
cracking and darkness of the implanted region is evidence of the formation of a rigid 3-
dimensional network and a disruption of crystallinity.

Figure 4-12b also indicates that the locus of failure shifted from the ﬁber-matrix
interface to a point inside the virgin ﬁber. Separation of the ﬁber interior from the
implanted region, with only small ﬁbrillated structrues maintaining contact, is similar to
the mode of failure seen in the untreated ﬁbers. The cohesive failure suggests that the

ﬂ- ..E—“ ‘

:J..&’"'I.:. _I-a

 

 

 

 

 

 

Figure 4-12: TEM micrograph of a ﬁber irradiated by 100 keV ions, 1x1015 N+/cm-2,
(a) showing irradiation zone and (b) showing ﬁbrillar decohesion.

 

61
upperlimittotleﬁber-mauixinterfacialshearsuengthmaybeﬂeinuinsic shear
strength of the ﬁber itself.

Gel-Fraction of High Energy Irradiated Fibers The results of 501- gel experiments
performedon30keVN+irradiatedfabricasafrmctionofdosearepresentedinFigure4—
13. The gel-fraction (undissovable crosslinked mnterial) increases from a value of 0% for
conuol to 1.24% for a dose of 1x1015 cm4. The small amount ofcrosslinked material is
concentrated in the surface layers of the ﬁber and increases the cohesive strength of the
ﬁber surface. The degree of crosslinking achieved probably will not affect the ﬁber's creep

or high temperature resistance properties.

 

62

 

 

Gel-fraction, wt.%

 

 

 

 

 

 

 

 

 

 

 

1.5
1
q
1.25 . /
1.0
1
0.75 /
0-5 q //
0.25
0? I I IIIIIT I I'll!!! r TIIIIII
13 re 15 re
10 10 10 10
Dose, N’Ian’

 

Figure 4-13: Gel-ﬁaction as a function of dose for 30 keV N+ ions.

 

 

63
Low Energy Ion Irradiation of Spectra-1000

Visual Observations Low energy ion irradiations produced a slight surface discoloration
of the fabric. The most severe discoloration occurred for the lkeV, 1x1016 cm-2
irradiation’s which resulted in a light frown fabric. The other processing conditions
produced perceptible, but less intense discoloration. Color change is related to the amount
of damage incurred by the material. These results, along with the high energy implantation
results, indicate that a threshold dose (high dose regime) for discoloration exists. Since the
higher energy irradiations produced more intense color changes, the ion energy is a

contributing factor to irradiation induced discoloration.

Mechanical Properties The tensile strength of 1 keV Art irradiated ﬁbers as a
function of dose is shown in Figure 4—14. No statistically signiﬁcant decrease in strength
is observed since all data points fall within the scatter band, but a slight decrease (approx.
9%) in strength from the control is evident for all doses. Figure 445 shows the tensile
strength of Ar" irradiated ﬁbers as a function of ion energy for a constant dose of 1x1016
cm-2. A slight increase in strength is noted for irradiation at 100 eV, followed by a
decrease to a minimum at 750 eV, and a retum to near the control specimen suength (2.54
GPa) at 1 keV. The observed trends must be regarded with caution since the highly
irregular cross-sectional area of Spectra 1000 ﬁbers makes an absolute determination of the
tensile strength difﬁcult.

The normalized interfacial shear strength “mm/1mm) between the ﬁber
and epoxy matrix for 1 keV Art irradiated ﬁbers as a function of dose is shown in
Figure 4-16. A clear trend of increasing ISS with increasing dose is observed, with the
largest value of ISS being 7.75 MPa for the highest dose of 1x1016 cm'z. This ISS
value corresponds to over a 300% increase when compared to the control value of 2.41

MPa. Also indicated in Figme 445 are the normalized ISS values for commercial

 

 

 

N

Tenslle Strength, GPa

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a ’ T ﬁr
j .. ..
.1
: Argonlon
Irnplantatlone,1lrev
:
III In lIIiIII II III III III
0 4 6 8 10
I5 .2
Doee,x10crn

 

ﬁgure 4.14: The tensile strength of 1 keV Ar" irradiated ﬁbers as a function of dose.

 

 

 

 

Tensile Strength, GPa

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4
q l
‘ Ar’lmelantatlona
It 1
3 1 1x10 lone/em: ...
2 .
1
q
o I I '7 I I I I I I' I t r I I I T
0 250 750 1W0
‘ lonEnergy,eV '

 

 

Figure 4-15: The tensile strength ofAr+ irradiated ﬁbers as a function of
ion energy for a constant dose of 1x1016 cm'z.

 

 

 

Normalized rss, 1: “Jr“...

 

 

 

 

 

 

 

 

 

4 , r r
q
.‘ ArticuIn-adiatioue
. Ioanergy-lkev
3 _
U r I I r

 

 

 

 

 

 

 

 

 

Figme 4-16: The na'malized interfacial shear strength for 1 keV Ar"
irradiated fibers as a function ofdose.

 

67
plasma and corona treated ﬁbers. The corona value (4.6 MPa) is less than the plasma
value (7.4 MPa) which, in turn, is slightly less than the highest ISS value. This shows
that low energy ion implantation is capable of producing comparable levels of ﬁber-
matrix adhesion comparable to those ofcommercial treatments.
Figure4—l7 showsthenormalizedISS asafunctionofArtionenergyata

constant dose of 1x1015 cm'2. Irradiation has in all cases increased ISS levels with a
maximum increase of approximately 680% occurring at an ion energy of 100 eV.
Thereisalsoaslighttendencyforsmallina‘easesinthelss athigherimplantation
energies. The dose dependence of irradiation on ISS for 100 and 250 eV irradiations
shown in Figure 4-18 suggests an optimum dose of 5x1016 cma. At this dose, 250 eV
argon ions produced a ninefold increase in interfacial shear suength level compared to
the control specimens, somewhat exceeding the values for plasma and corona
treatments and the values obtained by Ozzcllo, et al.[8] for high energy irnplantations
(7.41 MPa. 300% increase).

Surface Chemistry Figure 4-19 shows the oxygenzcarbon ratio of UI-iMW-PE
fabric irradiated to a dose of 1x1016 cm"2 as a function of ion energy determined by
XPS. The surface oxygen concentration increases from the control value of 3.5 at.% to
15.2 at.% for the lowest ion irradiation energy of 100 eV. The surface oxygen
concentration appears to be roughly constant with increasing energy, with a maximum
increase of 18.2 at.% at an ion energy of 1000 eV. As in the high energy case, the
source of oxygen has not been conclusively determined and could be from oxygen
partial pressures in the irradiation chamber or as a result of post-irradiation exposure
to atmospheric oxygen. Though N2 was used to bleed up the vacuum chamber to
atmospheric pressure, no nitrogen signal was demoted. No implanted argon was
observed.

 

 

 

 

1
IV E

i
I §\ 1.1L

 

 

 

 

Normalized rss, 1: “It...“
a

 

 

 

 

 

 

4' ‘ \O
2:

g«@@
0 II I II :I I I I I I Ii‘I I I I I I'

o 250 500 750 1000

Ion Energy, eV

 

Figure 4—17: The normalized ISS as a function ofAr“ ion energy at
constant dose of 1x1016 cm'z.

 

 

10
t: 8
P8
\
a 6
P—
of
9 4
i
2
i
0
0 2 4 6 8
16 2
Dose,x10 ions/cm

Figure 4-18: The dose dependence of ISS for 1m and 250 eV Ar+ irradiations.

 

10

7O

 

 

.15

.05

 

 

111111114

 

 

 

 

1

j

 

 

 

 

Oxygen-Carbon
Ar’ Irradiations

 

Dose of 11:101 ‘cm' 3

 

 

 

 

 

 

 

 

 

 

 

Figure 4-19: The oxygenzcarbon ratios of falxic irradiated to a dose of lilth16 cm'2

as a function ofAr+ ion energy.

 

 

a. . ‘\&’-“-!,‘e-‘.~ ’m:;* x” yvmes .

 

 

 

 

do

fur
in

en

6)

lo

7"

 

l‘

71
Figure 420 shows the oxygenzcarbon ratio of 1 keV Art ions as a function of
dose. The small variability (deviations less than 4 at.%) in oxygen concentration as a

function of dose, suggests that a somewhat constant value is achieved. To provide
furtherevidence forthis trend, higherdose irradiations were performed at low ion
energies of 1m eV and 250 eV. The oxygenzcarbon ratios for the 100 and 250 eV
irradiationsasafuretionofdoseareshowninfigme4—21. Foranion energyof250
eV, tle surface oxygen concentration appears to be a maximum (17.5 at.%) at the
lowest dose of rrtlol6 cut-2 and then tends toward a constant value (14.0 at.%). The
peakcouldbeanindicaﬁondetaﬁeeradicalequﬂibriumconcenuadonhadbeen
reached at a dose of 1x1016 cm'2. For higher doses, the onset of carbonization may
resultintheobserveddecreaseinsurfaceoxidation. Thedataforthe lOOeVionenergy
appears to follow the opposite trend of increasing oxygen concentration with increasing
dose. This mend suggests that 100 eV ions generate fewer free radicals, requiring
higherdoses toreach apeakvalueofoxidation.

Tle low energy results and the high energy results indicate that, for doses
above 1x1015 cm‘z, the oxygen surface concentration is somewhat constant.
Comparable levels of surface oxidation were found on the plasma (10.4 at.%) and
corona (15.6 at.%) treated ﬁbers. The ion dose, rather than energy, is therefore a more
important processing parameter for the production of surface oxidation. Gerenser [65],
studying oxygen plasma modiﬁed polyethylene, found a surface oxygen saturation
value of approximately 20 at.%. Our study combined with Gerenser's [65] results
suggests that tlere my be an upper limit to the amount of surface oxygen incorporation
possible under certain processing conditions.

Additional information on irradiated ﬁber surface chemistry was obtained from
ATR-FI‘IR measurements. Spectra for unirradiated fabric, and material irradiated with
1 keV Ar" ions and 1 keV N+ ions to doses of 1x1016 cm'2 and 1x1015 cmz,
respectively,areshownin Figure4-22. Thespectraare similartothehighenergy

72

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

It I-
Cw- T : d /
a -b -
.. l
1 I W“ l
Ar’rmdlauoru
: IlnalmsthI
1
cl
1
“II IIII IIII TIII IIII ITIU IIII IIII III!
1 2 4 s s 7 a 9 10
rs
Iondoee,x.10 cm'z

 

Figure 4-20: The oxygenzcarbon ratios of 1 keV Ar+ ions as a function of dose.

 

 

73

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

31
.25
//
.2 z
4.15
'1"
g .15;
. l
.1 q 3‘
1 Ar'rmdrraou
. loanergy-loodl,0
95‘ f 8250eV,D
1 l l I l
o “I IIII'IIII'IIII'IIII'IIII IIII “II IIII
1 2 3 4 5 6 7 8 9 10
1‘ -2
louduqucn

 

 

 

, Figme 4-21: Theoxygenzcarbonratios forthe lmandZSOeVirradiations
asafunctionofdose.

Transmittance (Arb. Units)

 

74

 

 

 

 

 

 

 

 

 

N" — 10“ cm"
A B
Control .
l I l l
coco soon soon 1000 one

Iavenumber (and)

Figure 4—22: Infrared spectra for fabric irradiated with 1 keV Ar" ions and
1 keV N+ ions to doses of 1x1015 cm'2 and 1x10” cm4.

wa- ;_--

 

 

 

 

hydl

llIlSi

larg

5w.
hyc
Tht
Th1
ftb

8111
dis
Sill

wt.

is

ﬁt

 

75
spectraofFigure4~2. Inthehradiatedmatuialatwobroadcompositeabsorption
bandshavedevelopedlabeledbyAandB. TheApeakisidentiﬁedashydroxyland
hydroperoxide groups and the B peak contains carbmyl. carboxylic acid and
unsaturatedhydrocarbon [55, 60] functionalities. Thefunctional groups generatedare
independentofionspecies. Mghmcarrirradinnondoscsarcanorncrormagninrdc
largerthantheNtnodistinctdegradativeprocesseﬂasobservedinthehighenergy
irradiations)areevident. Similarspectroscopicfeattueswerealsoobservedfor
specimensirradiatedatlowerenergies.

Surface Energy of Low Energy Irradiated Fibers UHMW-PE ﬁbers are
hydrophobic with low surface energies which decrease with increasing draw ratio [67].
The draw ratio for Spectra-10m ﬁbers is greater than 30 (considered a high ratio) [68].
T‘le total surface energy of 32 mJ/m2 for untreated ﬁbers causes poor wetting of the
ﬁber by the matrix resin. Figure 4-23 plots the surface free energies of the control and
irradiated ﬁbers as a function of Ar+ ion energy for irradiations of 1x1016 cm‘z. A
slightincreaseinthetotal smfaceenergyisobservedforallionpotentials, thoughthe
dispersivecornponentofthetotalenergyhasgenerallydecreased. Theincreaseintotal
surface energyis thus seen tobedominatedbyasharprisein thepolarcomponent,
which reaches a maximum of 12.2 mJ/m2 for the 750 eV irradiation.

For the case of Epon 828 epoxy resin the surface energy is 47.50 mJ/m2 (polar
= 10.8 mJ/mz, dispersive = 36.7 mJ/mz). Comparing the values found for the
irradiated ﬁbers and those of the epoxy, a closer match between the polar components
is achieved, while a larger gap in the dispersive component occurs. The plasma treated
ﬁber was found to have a total surface energy of 33.4 mJ/m2 with polar and dispersive
components of 9.4 rnJ/rn2 and 24.2 mJ/m2, respectively.

Theincreaseindepolm'componentofdresmfaceenergyisaresultofsmface
oxidation consisting of polar functional groups as identiﬁed by the IR analysis. The

 

76

 

 

 

ADlepereive
- Total

1111

Surface Energy, mllm 3
'8 8
th—q
. l—

 

 

 

 

 

 

 

 

 

 

 

 

 

. L r
q
10-:~-- ,« g ‘ ~
« é
o i
IIIIl'ﬁIIT TUII IIII—I
0 250 500 750 1000
IonEnergy,eV

 

 

 

Figure4—23: Thesurfacefreeenergiesofthecontmlandirradiatedﬁbers
asafunctionofAri’ionenergyforirradiationsoflxlomcma.

 

4"
“.WF
#—

 

 

 

0X

51]
01

fu

re
is

511

Si]

 

77
oxygen-containing groups are capable of participating in strong hydrogen-bonding
interactions, leading to improved wetting. The almost constant value of the polar

surface energy component for the energy range examined and the somewhat constant
oxidation level indicated in Figure 4—18, suggests that a constant surface density of
functional groups is achieved at a dose of 1x1015 cm'2. This also implies that the
Willelmymedrodprobesonlytheﬁrstfewmonolayersofdrematerialsincede
projected range of 100 eV ions is 1.4 nm (~2 unit cells).

The change in thepolarcomponentofthe surfaceenergy producedapronounced
effect on the wetting properties of the ﬁbers. Microdroplcts of liquid resin readily formed
andadheredtodreirradiatedﬁberswhﬂedropletsondeunueatedﬁbersbeadedupandfell
off. Thus, an oxygen surface concenuation of approximately 14 at.% (10 at.% if the
plasma treated sample is included) provides sufﬁcient chemical interaction for wetting to
occur.

Fiber surface modiﬁcation techniques tend to drive tie surface far from equilibrium.
All materials tend toward minimization ofthe surface free energy, so, ion induced surface
modiﬁed properties are, in principle, time dependent. A surface modiﬁcation techniqte
must impart a stable ﬁber surface chemistry to be useful. Hydrophobic recovery is the
reversion of a material from a hydrophilic state back to a hydrophobic state. This process
is common in many polymers, including polyethylene. Though no systematic aging
studies were performed, reproducible ISS results were obtained over a period of 3 months.
The individual chains in UI-IMW-PE are rigid and nuke hydrophobic recovery difﬁcult. In
addition, irradiation induced crosslinking increases the rigidity of the structme, furtler
discouraging hydrophobic recovery.

Although good wettability helps form a strong adhesive bond, it is not always
sufﬁcient for realization of increased ISS. Most commercial polyethylenes contain a low
molecular weight componart which, despite being perfectly miscible in the molten
polymer, is ejected to the surface dining processing. The high concentration of such

 

 

 

 

78
material on the ﬁber surface forms aweakboundary layerwhich must beremovedor

strengthened to achieve maximum adhesion levels. Ion irradiation may perform both
functions by sputtering the low molecular weight species and crosslinking the ﬁber surface.

SEM Observations of Low Energy Irradiations As discussed previously, ﬁber-matrix
adhesion occurs through a variety of mechanisms, all of which contribute to the overall
adhesion level and require investigation and characterimtion by an assortment of
techniques. A SEM investigation of the microbond pull-out results provide
phenomenological information concerning the mode(s) of interfacial failure. Figure 4-24a
shows a typical debond of the microdroplet from the ﬁber. Note that the impressions on
the droplet (B) are produced by the blades of the micrometer. The freshly exposed ﬁber
(A) appears smooth, with longitudinal striations which indicate the presence of ﬁbrils, and
the microdroplet failure is quite symmetric about the ﬁber's long axis. Also, the lack of
resin fragments on the ﬁber surface after the pull-out demonstrates that debonding is not
due to cohesive failure of the matrix, but rather to sliding of the ﬁber along the ﬁber-matrix
interface.

Figure 4-24b shows a droplet failure on the side opposite the blade micrometer.
The matrix resin has buckled in the thin region toward the end of the droplet, again
producing a more-or-less radially symmetric failure locus. Also shown, on the right hand
side of the image, is an undeformed microdroplet. Here, the high degree of ﬁber wetting
by the resin is suggested by the smooth, low angle, meniscus at the edge of the droplet
which tapers out to form a thin resin ﬁlm covering tie entire length of the ﬁber. This may
be compared to the meniscus of the untreated ﬁber, in which the contact angle is high and
the thin resin ﬁlm is absent, as shown in Figure 4-25. Furthermore, the embedment
lengths of the microdroplets on the control ﬁbers were, on the average, substantially
shorter than those on the irradiated ﬁbers, and axisymmetric conformation was achieved for
fewer than 60% of the microdroplets applied to the unheated ﬁbers. Axisymmetric

 

Figure 424: (a) Debond of tie microdroplet from the ﬁber (note impressions made
by shearing blades) and (b) droplet failure on the side opposite the
blade micrometer.

 

Figlne 4-25: Microdroplct on untreated ﬁber indicating the ﬁber’s
poor wetting properties.

 

81
conformation is known to always occurifthe contact angle is low and the ratio ofdrop

volume to ﬁber radius is above a critical valte [69]. Failure to achieve axisymmetry is thus
aﬁmﬁahdicaﬁonthﬂwetﬁngofﬂreunueaedﬁbasismfaiormthmdumdiawdones.

Though most failures developed symmetrically about the ﬁber's long axis,
occasionally cohesive failure occurs deep within the ﬁber, as shown in Figure 4-26.
Invariably, for a given length ﬁber, cohesive failures (referred to as peel-off) occur for
every microdroplet. Moreover, the peel-OE occurs at the same ﬁber-matrix location for
eachdroplet. Thissuggestsdiatsomeﬁbershaveaninuinsicweaknessdratmayberelated
to the lamellar structure of the ﬁber and possibly to the spinning and drawing conditions.
Similar fracture behavior is seen in melt spun polypropylene ﬁbers [70].

Figures 4-27 a and 4-27b show a comparison between the ﬁber morphology of
UHMW—PE ﬁbers in two irradiated conditions. A number of morphological features are
presentontheﬁbersmfaceafterirradiating with lkeVand250eVAr+ionsirradiatedto
doses of 3x1015 cm“2 and 1x1015 cm'2 respectively. First, the tendrils present before
irradiation have disappeared. Secondly, blister type structures are occasionally seen on the
1 keV ﬁber surface. Tlese two types of morphological structures result from thermal
effects produced by tle energetic ion beam and/or heat- generated by the neutralizing
element of the ion somee. These effects are magniﬁed by the fact that polyethylene has a
low value of thermal conductivity (0.15 to 0.67 W m'1 16‘) and low melting point, making
heat transfer from the ﬁber surface difﬁcult. Finally, small pits uansverse to the
longinrdinal axis of the ﬁber have developed on both ﬁber surfaces. These pits are similar
tothecrack-likepits foundin tleplasmaandcoronatreatedﬁbersinFigm-e4-10, butare
much smaller and more uniform.

Blister type structures were observed only in the 1 keV irradiations. All lower
energy irradiations had surface structures similar to the 250 eV irradiations shown in Figure
4-28b. The observed differences could be a result of irradiations at different dose rates
(beam current density) and different processing times. Though the 1 keV irradiation time

 

 

Figure 4-26: Illustration of cohesive failure occurring deep within a ﬁber.

 

..., ' 1121"“
250 9V Ar+ ..w“ ""‘vU'IWZV'utWJ. .mVth

_ .-. ...... , .-.-
.,,e..... .. ' ..‘t .~-.~ mu“.

.-4”

 

 

(b)

Figure 4-27: SEM micrograph showing surface morphology of Ar+ irradiations of
(a) 1 keV of dose 3x10” cm'2 and (b) 250 eV of dose 1x10l6 cma.

 

34
(5 s) islessthantheotherinadiationtimes, the 1 keV currentdensity (0.341 mA/cmz) was
signiﬁcantly higher than the other irradiations (less than 0.267 mA/cmz) except for the 750
eV irradiation which had a processing time of3 sec and a cut-tent density of0.316 mA/cmZ.
Theneuualizerﬁlamentcmrentofdreiongunwaskeptconsmntduringanofthe
irradiations andcannotberesponsible forthethernnl eﬂ'ects.

The observed morphological changes of the irradiated ﬁber are mostly responsible
fortheobservedincreasesininterfacialshearstrength. Forinstance,theremovalof
tendrils leaves tightly bound higher molecular weight species to the sin-face and may
remove a weak boundary layer on the ﬁber surface. Also, the pitting encourages
mechanical interlocking not previously available with the untreated ﬁber. Though the pits
also act as stress concentrators, their effect on the tensile suength seems to be minimal.
The blister type structures, however, may not be beneﬁcial to increasing the interfacial
shear strength, since they form weak surface layers on the bulk ﬁber and stress
concentration points, which may cause premature failine. Under proper processing
conditions, low energy ion implantation affects the ﬁber surface morphology in which
ways which tend to maximize fiber-matrix adhesion.

The drawing ofthe ﬁber and high processing temperatures produce ﬁber surface
tensile stresses upon cooldown. The ﬁber pitting is likely the result of the relaxation of
these tensile stresses. Plasma and corona surface modiﬁcation techniques many times lead
to similar ﬁber surface morphologies, but may vary signiﬁcantly for different processing
parameters. The surface morphologies presented hue are a representative example and are
not unique. The similarities can be attributed to the similar ion dynamics occtnring at the
ﬁber surface for plasma and corona processing. The ion beam dynamics of low energy ion
inadiationsarewellconholledardwitlrinacertainwelldeﬁnedarea, theionbeamis
extremely uniform. The relatively high temperatm'es of plasma and corona processing and
low energy irradiations increases the polymer chain mobility, effectively reducing the bond

 

85
breakingenergyandresultingineasierstressrelaxation. Thistypeofprocesscouldeasily
lead to the uniform pitting observed.

Thoughincreasedstnfaceoxidadonandinaeasedwetﬁngoftheﬁbabydtemsin
(asvisuanyobsavedbyeaseofdropletfamaﬁmwfbommephsmaandcmomueawd
ﬁbers was achieved, the ISS values found were lower than the values reported for low
energy irradiations. Theﬁbersm'facemtrphologyoftheccronaandplasmatreatedﬁbers
is nonuniform and may lead to complicated stress states along the ﬁber-matrix interface
whichencouragesfailure. Thelowenergyirradiationploducedauniformsurfacetextme
which tends to distribute the stresses at the interface more evenly, resulting in larger ISS
values. If complete wetting did not occm', voids would be produced, resulting in stress
concentration sites which promote failure.

The pitting present in the previously mentioned processing methods was absent in
the high energy irradiations shown in Figure 4-7. Since the high energy implantations
wereperformedatextremelylowbeamcurrents (2mm. comparedtoZOmAforlow energy
irradiations), thermal damage to the ﬁber surface is minimized, inhibiting the production of
stress relaxation mechanisms. Furthermore, the high amount of energy deposition
produced a substantial amount of crosslinking at the ﬁber surface, severely hindering stress
relaxation mechanisms.

Though TRIM-91 range calculations are not strictly applicable to ion energies
belowafewkeV,anestimatepredicts ionpenetrationsof 5.5 nm(~7 unitcells) and
1.4 nm (~2 unit cells) at 1000 eV and 100 eV, respectively. This indicates that little
wbsmfacemodiﬁcaﬁmﬁomdwkineﬁceﬁ‘ectsofdteaccelerawdionisandcipawd A
simple calculation estimates, with an assumed sputter yield of 3 atoms per ion, a
conservative upper bound on the sputtered layer depth of approximately 3.5 nm. The
SEM micrographs of the low energy irradiated ﬁbers indicated morphological changes
well below this depth. Thus ion irradiation not only generates free radicals, but also
acts as a catalyst for ﬁber structural modiﬁcation. The slight amount of material

 

86
removal, however, may play a signiﬁcant role in ISS improvement through the
elimination, or partial removal, ofsurface impurities and weak boundary layers.

TEM Observations of Low Energy Irradiations To more thoroughly investigate
the ﬁber-matrix interphase. TEM micrographs of a ﬁber modiﬁed by 500 eV ions with
a dose of 6.0x1015 cm-2 are shown in Figures 427a and 4-27b. The ﬁbers and the
surrourxlingmauixarelabeledmsisthernicrotomingdirection. Thecontrol
micrograph was previously shown in Figure 4—11 and clearly showed a low degree of
adhesion betweentheﬁberandmatrix. Themicrographoftheirradiatedﬁberindicates
a well bonded ﬁber-matrix interphase. with decohesion occurring near the ﬁber's
surface. A higher magniﬁcation view, shown in Figure 4-28b, of the irradiated ﬁber
indicates failure ensues when ﬁbrils pull away from the interface region. The same
behavior was seen in ﬁbers implanted at lower energies.

The ﬁber-matrix interfacial behavior for low energy ion irradiated ﬁbers is
slightly different from that of 100 keV N+ irradiated ﬁbers. In the latter, an irradiation
modiﬁed region was clearly discernible and cohesive failure occmred within the ﬁber
and not at the ﬁber-matrix interface proper. For the low energy irradiations, no ﬁber
modiﬁcation is detectable, but the same ﬁbrillar breakdown at the interface is observed.
Thus regardless of the type of surface modification technique, the interfacial shear
occrn'sascohesivefailureintheﬁber. Theﬁbersm'facemorphologyandnotthe
intrinsic cohesive strength limits the achievable ISS. Maximum levels of ISS are
therefore accomplished through adequate wetting and tailoring of the ﬁber srn'face

texture for an even distribution of interfacial stresses.

 

87

 

Figure 4-28: TEM micrographs of a ﬁber modiﬁed by 500 eV Ar" ions with

a dose of 6.0x1015 cma.

 

88
Chapter Five
CONCLUSIONS

1) High-energy ion irradiations of ultra-high molecular weight polyethylene (UHMW-PE)
ﬁbers produced surface carbonization which increased with ion dose.

2) Smalldecreasesintensile strength forlowenergy (100-1000eV)Ar" ionirradiated
UHMW-PE ﬁbers were found.

3) A 900% increase in the interfacial shearstrength between UHMW—PE ﬁbers and an
epoxy matrixoccurred atanAr’ion energy of250eV and adoseof5x1015 cm4.

4) Low and high energy ion irradiation of UHMW-PE ﬁbers induced sin-face oxidation
consisting of hydroperoxide, hydroxyl and carbonyl polar functional groups.

5) The oxygen:carbon composition ratio of UHMW-PE fabric irradiated by 30 keV N+

ions indicates a peak oxygen concentration of 16.5 at.% for a dose of 1x1015 cm‘z.

6) The gel-fraction content of 30 keV N+ irradiated fabric increased with increasing dose.

7) Increasesin thepolarcomponentofthe surfacefreeenergyoflowenergyAr’ion
irradiated UHMW-PE ﬁbers were observed at doses of 1x1016 cm‘z.

8) The ﬁber surface morphology of low energy Ar+ ion irradiated UTMW-PE ﬁbers
consisted of small, uniform pits which increase the mechanical interlocking between ﬁber

and matrix.

 

89

9) UHMW-PB ﬁbers modiﬁed by 5(1) eV ions toadose of6x1015 cm'2 indicated a well
bonded ﬁber-matrix interphase. with ﬁbrillar decohesion occurring near the ﬁber’s surface.

 

 

 

LIST OF REFERENCES

 

LIST 01" REFERENCES

[1] L.T. glsggyen, N-H. Sung and N.P. Suh, J. Polymer Sci.: Polym. Lett. Ed. 18, 541
(1 ).

[2] C. Y. Kim and D.A.I. Goring, J. Applied Polymer Sci. 15, 1357 (1971).
[3] C. Y. Kim, J. Evans and D.A.I. Goring, J. Appl. Poly. Sci. 15, 1365 (1971).

[4] S.L. Kaplgn, P.W. Rose, H.X. Nguyen and H.W. Chang, SAMPE Quarterly, July,
55 (19 8).

[5] JJ. Cuomo, S.M.RossnagelandH.R. Kaufmaninwm
(U r ' 'll' '14 GINO”... . 'an' "i "U“ l:.!.l.:.;. “1"}

Park Ridge, N. J. (1989).

[6] H. R. Kaufmam'Technology of Electron— Bombardment Thrusters", in '
36, L. Marton Ed., pp. 265-373, Academic
Press, New York (1974).

[7] HR Kaufman, J.J. Coumo and J.M.E. Harper, J. Vac. Sci. Technol., 21, 725
(1982).

[8] A. Ozzcllo, D. S. Grumman, L. T. Drzal, J. Kalantar, I-H. Low and R. A. Moody,
MRS Proc. 153, 217 (1989).

[9] L.T. Drzal, M.J. Rich and RF. Lloyd, J. Adhesion, 16, 1(1982).

[10] D Hull. WWW Cambridge University Press.
Cambridge, England (1981).

[11] K. W. Allen, J. Adhesion , 21, 261 (1987).

 

91
[12] A. Kelly and W.R. Tyson, J. Mech. Phys. Solids. 13. 329 (1965).

[13] L.T. Drzal, M.J. Rich, J.D. Camping and WJ. Park, 359' Annual Technical
Conference, Reinforced Plastics Composites Institute, The Society of Plastics

[14] LJ. Broutman, "Measurement of the Fiber-Polymer Interfacial Strength”, in
27%:69’) in Composites, ST? 452, American Society for Testing and Materials,

[15] Baal? P. Mini and L. Rebenfeld, Composite Sci. and Technology, 28, 17

[16] LP. Mandell, J.-H. Chen and FJ. McGarry, Research Report R80-1, Department of
llvgastgrials Science and Engineering, Massachusetts Institute of Technology, Feb.

[17] P. J. Herrera-Franco, V. Rao, M.Y.M. Chaing and L.T. Drzal, "Bond Strength
Mealntnl'ecgent in Composites -Analysis of Experimental Techniques— ",to be
pub 's

[18] L. T. Drzal and P.J. Herrera-Franco, "Composite Fiber-Matrix Bond Tests",
Engineered Materials Handbook, Vol. 3, Adhesives and Sealants, 1991.

[19] U. Gaur and B. Miller, Composites Sci. and Technology, 34, 35 (1989).
[20] G. Capaccio and I.M. Ward, Polymer Eng. Sci., 15, 219 (1975).
[21] P.J. Lemstra and R. Kirschbaum, Polymer, 26, 1372 (1985).

[22] P.J. Lemstra, R. Kirschbaum, T. Ohta and H. Yasuda, in '
Rom Ed. I.M. Ward, Elsevier Applied Science, London (1987).

[23] D. Hofmann and E. Schulz, Polymer, Vol. 30, Nov.(1989).
[24] B. Kalb and A.J. Pennings, J. of Materials Science 15, 2584 (1980).
[25] B. Kalb and A.J. Pennings, Polymer Communications, 21, January (1980).

[26] P.J. Barham and A. Keller, Journal of Materials Science 11, 27 (1976) .

92
[27] EM. Liston, J. Adhesion, 30, 199 (1989).

[28] S. Holmes and P. Schwartz, Composites Science and Technology, 38, 1 (1990).
[29] N.H. Ladizesky and I.M. Ward, J. Mat. Sci., 18 , 533 (1983).

[30] N.H. Ladizesky and I.M. Ward, J. Mat. Sci., 24, 3763 (1989).

[31] U. Gaur and T. Davidson, Mat. Res. Soc. Symp. Proc., 170, 309 (1990).

[32] B. Tissington, G. Pollard and I.M. Ward, J. Materials Science, 20, 82 (1991).

[33] H.X. Nguyen, R.G. Merrill, A.K. Schriver and E. A. Leone, Structural Composites:
Design and Processing Technologies, Proceedings of the 6th Annual ASM/ESD
Advanced Composites Conference, Detroit, MI, Oct. 1990.

[34] Y.L. Chen and B. Ranby, J. Polymer Sci.: Polymer Chemistry, 27, 4051 (1989).
[35] Y.L. Chen and B. Ranby, J. Polymer Sci.: Polymer Chemistry, 27, 4077 (1989).
[36] Y.L. Chen and B. Ranby, Polymers for Advanced Technologies, 1, 103 (1990).

[37] PG. Klein, N.H. Ladize and I.M. Ward, J. of Polymer Science: Part B: Polymer
Physics, 24, 1093 (198 ).

[381 WWW by G. Carter and WA. Grant. John Wiley and
Sons, New York, 1976.

139} WWW 2"“ Edition. Edited by 31‘ Ziegler.
Academic Press, Inc. San Diego, CA, 1988.

[40] T. Venkatesan, L. Calcagno, B. S. Elman and G. Foti, "Ion Beam Effects 1n Organic
MdecularSOHdsmdeYM” inImBeamMailiﬂrationgﬂmmulmEds P
Mazzoldi and G. W. Arnold, Elsevier, pp. 301 379, 1987.

[41] W.L. Brown, Radiation Effects, 98, 115 (1986)

[42] O. Puglisi, Materials Science and Engineering , B2, 167 (1989).

 

93

[43] D. Briggs and M.J. Hearn, Vacuum, 36, IWS (1986).
[44] D. Briggs and M.J. Hearn, Int. J. Mass. Spect. Ion Proc., 67, 47 (1985).
[45] J. Davenas, X.L. Ku, G. Boiteux and D. Sage, Nucl. Inst. Meth., B32, 754 (1989).

[46] G. Marletta, Nucl. Inst. Meth., B46, 295 (1990).
[47] D. Fink, M. Muller, L.T. Chadderton, RH. Cannington, R.G. Elliman and D.C.
McDonald, Nucl. Inst. Meth., B32, 125 (1988).

[48] J. Davenas, G. Boiteux and M. Fallavier, Nucl. Inst. Meth., B32, 796 (1989) .

[49] S.L. Koul, I.D. Campbell, D.C. McDonald, L.T. Chadderton, D. Fink, J .P.
Biersack and M. Muller, Nuclear Instruments and Methods in Physics Research,
B32, 186 (1988).

[50] D. Fink, M. Muller, L.T. Chadderton, P.H. Cannington, R.G. Elliman and D.C.
McDonald, Nuclear Instruments and Methods in Physics Research, B32, 125
(1988).

[51] 1E Ziegler. 1P. Biemck and U. Littmark. W
m, Pergamon, 1985.

[52] RH. Lee, M.B. Lewis, P.J. Blau and L.K. Mansur, J. Mater. Res., 6, No. 3,
Mar 1991.

[53] MB. Lewis and EH. Lee, Nuclear Instruments and Methods in Physics Research,
B61, 467 (1991).

[54] A. Ishitani, K. Shoda, H. Ishida, T. Watanabe, K. Yoshida and M. Iwaki, Nuclear
Instruments and Methods in Physics Research, B39, 783 (1989).

[55] M. Schaible, H. Hayden, and J. Tanaka, IEEE Transactions on Electrical Insulation
Vol. El-22 No. 6, Dec. 1987.

[56] DH. Kaeble, P.J. Dynes and EH. Cirlin, J. Adhesion, 6, 23 (1974).

 

94

[57] J. Kalantar, L.T. Drzal and KJ. Book, J. Vac. Sci. Technol., A8(5), Sep/Oct,
3878 (1990).

[58] A.M. Guzuman, J.C. Carlson, J.E. Bares and PP. Pronko, Nucl. Inst. Meth.,
B7/8, 468 (1985).

[59] U.W. Gedde and M. Ifwarson, Poly. Eng. Sci., 30, 4, 202 (1990).

[60] W.F. Maddams and SF. Parker, J. Polymer Sci.: Part B: Polymer Physics, 27,
1691 (1989).

[61] Ml. Chirara and L. Georgescu, Nucl. Inst. Meth., 209/210, 401 (1983).

[62] FA. Smidt, International Materials Reviews, 35, No 2, 61 (1990).

[63] J.A. Lanauze and D.L. Myers, J. Polymer Sci., 40, 595 (1990).

[64] H. Iwata, A. Kishida, M. Suzuki, Y. Hata and Y. Ikada, J. Polymer Sci.: Part A:
Polymer Chemistry, 26, 3309 (1988).

[65] L.J. Gerenser, J. Adhesion Sci. Tech., 1(4), 303 (1987).

[66] CH. Oudet and HH. Kausch, J. Mat. Sci. Lett.,10, 1061 (1991).
[67] M. N. and I.M. Ward, Mater. Sci. Technol., 3, 814 (1987).

[68] Personal communication, G. Weedon, Allied Signal, Petersburg, VA.
[69] BJ. Carrol, Langmuir, 2, 248 (1986).

[70] H. Bodaghi, J.E. Spuirell and 1.1. White, Intern. Poly. Proc., 3 (2), 100 (1988).

 

 

...-.. -- . -. . . .. .a .....- ..-“...aaaa-.- .....-caa-a—Iau—n-nym‘w‘ vandal-04¢».-

 

nrcnran STnTE UNIV. LIBRARIES
illWWWI”W||||WIWIIWWIHIH1Hll
31293010553315