Hdpe Butt Fusion Welding Theory

BUTT FUSION WELDING OF POLYETHYLENE
PIPES
A thesis submitted for the degree of

Doctor of Philosophy
by
Haroon Rashid
DEPARTMENT OF MATERIALS ENGINEERING
BRUNEL UNIVERSITY
JANUARY 1997
ýY
_
` `D
6j
-ý ý
ý
ýý
r'"
ABSTRACT
The butt fusion process is extensively used in the joining of
polyethylene (PE) pipes by the water and gas industries. This
welding process although deceptively simple, is rather poorly
understood, with much of the initial developments being of a
rather empirical nature. The Water Research centre (WRc) have
funded the present research in an attempt to optimise the
welding of high pressure pipeline (PE100) systems.
The main aims of this research were to investigate the effect
of different welding conditions on the physical and
mechanical properties of the joints produced and to
investigate these effects on the micro- and macro-structures
of the joints produced.
A series of welds were made using Eltex Tub 124 and Rigidex
002-50 pipes of 180mm diameter. The fusion pressure and heat-
soak times were varied. A milling machine witha twin cutter
arrangement was used to obtain the test specimens from around
the circumference of the pipes. Differential scanning
calorimetry was used to study the effect of sample
preparation methodology on the thermo-oxidative stability.
Polarised light microscopy and image analysis were used to
study the macro- and micro-structural developments in the
weld joint. Joint strength was evaluated via standard and
non-standard tensile test methods.
Milling the samples to produce the test specimens was found
to decrease significantly the thermo-oxidative resistance of
the polymer. Reasons for this behaviour have been proposed.
In order to achieve high quality thin films from microtomy,

custom-made blades were used. This programme also developed
the optimum polishing method for the microtomed blades. The
macro-structure of the bead: its shape and dimensions were
found to be a function of temperature and pressure.
Correlation was found between the bead geometry and the
position around the circumference of the pipe. The macro-
structures within the weld zone also showed this dependence
on the position along the circumference of the pipe. An
examination of the microstructures of each weld had shown the
presence of five different zones.
The feasibility of using microtomed thin sections in a
tensile test was demonstrated. The test method provides a
means to study failure initiation and propagation in the
tensile test specimen. Initial deformation was found to occur
in the centre of the melt-affected zone (MAZ) and the final
failure occurs at the junction of the weld bead and the bulk
polymer. Tests on films without the weld bead showed that
maximum deformation occurred at the centre of the sample
within the MAZ.
The presence of the bead and the asymmetry in the test

specimens caused by the welding process were found to have a
significant influence on the failure mode and the failure
strain. The strain rate was also found to play a significant
role in both beaded and debeaded samples. The failure was
initiated from the pseudo notches in the beaded samples. In
the debeaded sample the failure was within the MAZ.
ill
TITLE (i)
ACKNOWLEDGEMENT (ii)
ABSTRACT (iii)
CHAPTER 1 INTRODUCTION 1
1.1 Introduction 1
1.2 Objectives- 4
1.3 Outline of the thesis 5
CHAPTER 2 LITERATURE REVIEW 6

2.1 Materials 6
2.2 Effect of processing on polyethylene pipes 7
2.2.1 Thermal Oxidation 10
2.3 Morphology 11
2.4 Degradation and stabilisation of polyethylene 14
2.4.1 Additives 15
2.5 Evaluation of thermal degradation 16
2.5.1 Thermal analysis 17
2.5.2 Oxidation induction time (OIT) test 17

-
2.5.3 Correlation of OIT with antioxidant 19
concentration
2.6 Butt fusion welding 21
2.6.1 The butt fusion welding process 22
2.6.2 Theory of butt fusion welding 25
2.6.3 Microstructure of plastic welds 28
2.6.3.1 Specimen preparation 34
2.6.4 Factors influencing butt fusion welding 35
2.6.5 Temperature distribution within the weld 40
iv
2.6.6 Butt fusion welding of dissimilar grades 41
of polyethylene
2.6.7 Butt fusion weld quality and testing 42
2.6.8 Mechanical properties 43
2.6.8.1 Mechanical testing 44
CHAPTER 3 EXPERIMENTAL TECHNIQUES 50
3.0 Pipe materials so
3.1 Mater ials characterisation 50
3.1.1 Density gradient column 51
3.1.2 Gel permeation chromatography 52

3.1.2.1 GPC operating conditions 52
3.1.3 Differential scanning calorimetry 53
3.1.3.1 Melting point range and enthalpy 53
of fusion
3.1.3.2 Oxidation induction time 54
3.2 Butt fusion welding 55

3.2.1 Equipment 55
3.2.1.1 Thermal mapping of heater plate 56

3.2.2 Welding procedure 56
3.3 Micro structure 58

3.3.1 Microtomy 58
3.3.2 Polishing of blades 64

3.3.2.1 Disposable blades 64
3.3.2.2 D-profiled steel knives 64

3.3.3 Transmitted light microscopy 66
3.4 Analysis of welded specimen 66
3.4.1 Weld bead measurement 66
3.4.2 MAZ-Internal weld dimensions 68
3.5 Tensile testing 68
V
3.5.1 Microtensile testing-sample preparation 68
3.5.2 Macro-tensile specimen preparation 69
3.5.3 Micro tensile testing procedure 71
3.5.4 Macro tensile testing procedure 72
CHAPTER 4 MATERIAL CHARACTERISATION 73
4.1 Introduction 73
4.2 Density 73
4.3 Molecular weight determination 74
4.4 Thermal analysis 75
4.4.1 Differential scanning calorimetry 75
4.4.2 Oxidation induction time (OIT) 82
4.4.2.1 Sample mass 87
4.4.2.2 Volatility of antioxidants 91
4.4.2.3 Electron spin resonance (ESR) 92
4.4.2.4 X-ray photoelectron spectroscopy (XPS) 93
4.4.2.5 Surface damage 94
CHAPTER 5 WELD GEOMETRY AND DIMENSIONS 102
5.1 Introduction 102
5.2 Sample coding 102
5.3 Bead geometry visual assessment 104

-
5.4 Weld geometry - quantitative assessment 110
5.4.1 Bead widths 110
5.4.2 Bead areas 113
5.4.3 Melt affected zone (MAZ) quantitative 120

-
assessment
vi
5.4.4 Shear zone measurement 129
CHAPTER 6 MACRO AND MICROSTRUCTURAL EXAMINATION 133
6.2 Microtomy 133
6.2.1 Chemical polishing 134
6.2.2 Manual polishing 135
6.3 Macro-structure 136
6.4 Microstructure 147
6.4.1 Microstructure of the extruded pipe 147
6.4.2 Microstructure of the MAZ 152
CHAPTER 7 TENSILE TESTING OF MICROTOMED WELDED 162

SECTIONS
7.2 Image analysis 162
7.3 Microtensile testing of Eltex 164

7.3.1 Beaded sample 164
7.3.2 Debeaded sample 174
7.4 Microtensile testing of Rigidex 181
7.5 Effect of testing speed 183

7.6 Thin films - Rigidex 186
7.7 Final failure modes for welded films 191
7.8 Summary 194
vi'
CHAPTER 8 TENSILE PROPERTIES 195
8.2 Preliminary tensile tests 195
8.3 Effect of machining marks on tensile testing 198
8.3.1 Effect of defects on tensile testing 200
8.3.2 Effect of notching (skewing) on tensile 200

testing
8.4 Initial welding of Eltex and Rigidex 201
8.5 Eltex PE welds 205
8.6 Rigidex PE welds 214
8.7 Mixed welds between Eltex and Rigidex 215
8.8 Dual pressure welds 216
8.9 Failure modes in simple tensile testing of 223

butt welds
CHAPTER 9 CONCLUSIONS AND RECOMMENDATIONS FOR 225

FURTHER WORK
9.2 Conclusions 226
9.2.1 Material Characterisation 226
9.2.2 Weld Geometry and Dimensions 226
9.2.3 Macro and Micro-Structural Examination 228
9.2.4 Tensile Testing - Microtomed Welded 229

Sections
9.2.5 Tensile Properties 230
9.3 Recommendations for Further Work 231
REFERENCES 233
APPENDICES 254
viii
CHAPTER 1
INTRODUCTION
1.1 INTRODUCTION
The use of thermoplastic pipes, particularly for the
distribution of water and gas, has increased rapidly since

they were first introduced in the 1950's. In the UK, for
example, some 5000km of polyethylene (PE) pipe ranging in
diameter from 12mm to 1600mm is installed every year

[Lamond, 1995]. Thermoplastics are relatively cheap and
easily extruded into pipes. Furthermore, they are
corrosion resistant, lightweight and flexible [Bowman,
19951. For distribution of water and gas, PE is often the
preferred material mainly because of its flexibility which

makes laying easier. At present the use of PE pipe of
diameters more than 250mm is relatively low, but there is
a growing interest. in the water and gas industries to use
pipe diameters of more than 500mm [Watson', 1990).

Currently, these large sizes can only be joined by butt-
fusion welding. Butt-fusion welding is a reliable and
straight forward process, but it is slow when used on

large diameter, thick walled pipes. For example, the
recommended welding procedure for 400mm outer diameter PE
pipes with a wall thickness of 40mm, can take more than
one hour.
Page 1
In terms of pipe joining the alternative methods are
socket fusion and electrofusion joining [Marshall' et al.,
1995]. Socket fusion involves heating the inside of a
plastic socket fitting and the outside of the two pipe
ends to be joined with a specially shaped heating tool.
Once sufficiently heated, the pipe-ends are pushed into
the fitting and a joint is formed on cooling. In
electrofusion joining the pipes to be joined are pushed
into a coupler, which contains an integral electrical

heating coil embedded into its inner surface. When
energised from an external power supply the material
surrounding the coil at the outer pipe wall surface meets
and expands and fuses upon cooling to form the joint
[Marshall' et al., 19951
Since electofusion welding requires the use of an
additional plastic moulding for each joint, this
substantially increases the cost of each joint. On the
other hand this has the advantage of providing
reinforcement at the joint due to increased wall thickness
[Torsun et al., 1995]. In addition to this it is difficult
to produce large diameter fittings within reasonable

dimensional tolerances as the fittings are generally
injection moulded.
In recent years, there has been a rapid introduction of
new pipeline resins. In conjunction with this, there has
been a general trend in Western Europe to move over to
large diameters and/or high pressure plastic pipeline
systems (PE100). However, in general the welding
conditions have not been optimised for these new materials

[Wilson, 1995]. Very often the end user or fabricator is
Page 2
recommended to refer either to the manufacturer or
undertake their own weldability trials in order to

determine the optimum joining parameters [Land and
Erlenkamper, 1985].
Butt-fusion joining is a deceptively simple process,

complicated by the variety of parameters involved at each
stage of the joining process, all of which may change if
the different grades of materials are joined. The majority
of research in polymer welding has tended to concentrate

on the study of processing parameters [Stokes, 1988;
Schlarb and Ehrenstein, 1989] and on the testing of welded
joints [Nessel and John, 1984; Barber and Atkinson, 1974;
DeCourcy and Atkinson, 1977]. However, this has been
largely empirical; the more fundamental aspects of butt
fusion joining concerning the process, how material
variables effect it and the implications that these may
have on the joining conditions required for different
materials have not been considered [Folkes et al., 1991].

As such it is only in the water'and gas industries where
standards for the joining procedures, equipment and
inspection methods exist [Watson3,1990; Davitt, 1995].
With current trends towards the automation of welding

equipment in an effort to produce more consistent, high
quality joints, a thorough understanding of the joining
process is essential if this is to be achieved

successfully. The consequences of installing pipes without
a full understanding of their behaviour have been amply
demonstrated in the past, with early polyvinylchloride
(PVC) pipes which were prone to cracking when poorly laid;
and electrofusion joints experiencing failures when the
pipe surfaces were not properly prepared.
Page 3
Another reason for understanding the joining process in
detail is the lack of effective non-destructive testing
(NDT) techniques [Hale, 1990] to assess the quality of
joints. Although many of the traditional weld inspection
techniques have been tried with plastics welding, due to
the nature of the material itself they are largely
unsuccessful in identifying any defects, voids or other
problems associated with poor welds. Destructive tests are
used to predict optimum joining conditions but there is no
consensus of opinion on the best test method to assess the
quality of PE joints [Hale, 19901.
The current work is set against this background of
widespread usage of a joining process that is both more
complex than first assumed and which is relatively poorly

understood.
1.2 OBJECTIVES
(i) To study the effect of welding conditions on the macro
and microstructure of butt-fused PE pipes. The term macro
and microstructures are used here to describe the shapes
and geometry of the weld zone and the microstructural

features within the weld zone.
(ii) To investigate the deformation modes of the butt-
fused PE welds using macro and micro-tensile testing. The
macro-tensile tests refer to conventional tensile tests.

Micro-tensile test involved the tensile testing of
microtomed sections of the butt joint.
Page 4
(iii) to correlate the deformation modes to the macro and
microstructures of the butt-fused welds.
1.3 OUTLINE OF THE THESIS
This thesis comprises of nine chapters. Chapter two
surveys the current literature and covers aspects of
structure and properties of polyethylene, morphology and

the effect of degradation on the pipe. Issues relating to
the butt fusion welding, microstructures of welds and
mechanical properties are also covered in chapter two. In
chapter three, the materials and experimental procedures
start with a brief description of the experimental

techniques which were used during the course of this
research. The experimental results are presented and

discussed in chapters four to eight. The material
characterisation and the oxidation induction time results
are presented in chapter four. The weld bead geometry and

dimensions of the melt affected zone as a function of the
position along the circumference of the pipe are discussed

in chapter five. Chapter six deals with the macro and
microstructure of the weld zone. Chapter seven reports on'

the micro-tensile testing of welded sections. Chapter
eight deals with the macro-tensile properties of the pipes

and the welds. The conclusions and recommendations for
further work are presented in chapter nine.
Page 5
Chapter 2 Literature Review
CHAPTER 2
LITERATURE REVIEW
2.1 MATERIALS
There are three main types of commercial water-grade
polyethylene (PE) resins which are generally available.

Type (I) corresponds to low density polyethylene (LDPE)
with a density range of 910-925 kgm-3 and is produced by
the high pressure process [Brydson, 19821. Type (II) is
classified as medium density polyethylene (MDPE) with

densities in the range of 926-960 kgm-3. Type (III)
corresponds to high density polyethylene (HDPE) and has a

density range of 941-965 kgm-3. MDPE and HDPE are produced
by a low pressure process utilising either metal oxide
catalysts (Philips Process) or aluminum alkyl or similar
materials (Ziegler Process) [Brydson, 1982]. A simplified
schematic illustration of the Philip's HDPE production
route is presented in Figure 2.1 [BP, 19861. In the UK,
LDPE is used for low pressure applications whereas MDPE
and HDPE are used in high pressure pipe systems. 8001 of
all mains and 90% of all services are made of HDPE

[Brinken, 1982]. The usage of PE for pipe related
applications accounts for about 3-4% of the total PE
consumption worldwide. The total quantity of PE resin
consumed globally for pipes has been estimated at
Page 6
1. UNREACTED
SOLUTION ARE
MONOMERS
PROCESS; FLASHED OFF.
POLYMER IS 2. THE SUSPENDED
SOLUBLE IN CAT. IS REMOVED
THE DILUENT SOLUTION BY CENTRIFUGING.
PROCESS 3. SOLUTION IS
COOLED TO
120-160°C
30-30 ATMS. PRECIPATE THE
CHROMIUM POLYMER.
4. THE POLYMER I
ETHYLENE s OXIDE CAT.
SEPARATED BY
CO-MONOMERS DILUENT
CENTRIFUGING.
HYDROCARBON
SLURRY
PROCESS
90-100°C
SLURRY PROCESS;
POLYMER IS NOT
SOLUBLE IN THE
DILUTENT
1. POLYMER GRANULES
FORMED AROUND
CATALYST PARTICLES.
2. UNREACTED
MONOMERS ARE
FLASHED OFF.
3. THE POLYMER IS
SEPARATED BY
CENTRIFUGING.
4. THE POLYMER IS
CONTAMINATED WITH
SMALL AMOUNTS OF
CATALYST
Figure 2.1 Schematic illustration of the" production

processes in the manufacture of HDPE via
the Philip's process (BP, 1986).
1,060,000 tones for 1995 with European consumption
accounting for half of global total [Lamond, 1995].
2.2 EFFECT OF PROCESSING ON POLYETHYLENE PIPES
Pipes are manufactured by an. extrusion process. In
principle the extrusion process consists of metering
polymer usually in the granular form into a heated barrel
with a rotating screw. The rotation of the screw causes
Page 7
shear mixing of the molten polymer and also moves the
polymer up the barrel where it is forced under pressure
through the breaker plate and into an annular die. The
resulting extruded pipe is calibrated by means of water
cooled sizing die or vacuum sizing [Reitemeyer, 1981].
Monitoring and control of dimensions of the finished
product is achieved by the use of various ultrasonic
devices [Cist and Smith, 1978; Limpar, 1988] Pittman et

.
al. [Pittman et al., 1992] have carried out a
comprehensive study of the cooling phase of large diameter
pipe including in-situ measurements of temperature within
the pipe wall. The computer simulation was shown to
predict accurately the pipe profiles.
The nature of pipe extrusion is such that it can introduce
a range of defects such as gas bubbles, foreign particles
and spiderlines. These defects can reduce the performance
of the pipe depending on the size and nature of the

defect. Extrusion, combined with the cooling phase also
introduces changes in microstructure and residual stress
distribution through the pipe wall. This is due to the
differential cooling between the surface and the bore.
This factor can also influence the mechanical properties
of the pipe [Doshi, 1989; Beech et al., 1988]. The outer
wall of the pipe solidifies first, whereas inside, the
material is still in a molten state. As cooling
progresses, the material in the inside of the pipe
attempts to contract, but is prevented from doing so by
the material in the outer wall that has already
solidified. This results in a compressive stress build up

in the outer wall due to constrained elongation [Gebler
and Racke, 1982]. In addition, the extrusion process tends
Page 8
to orient molecular chains thus imparting a degree of
anisotropy. Bhatnagar and Broutman [Bhatnagar and

Broutman, 1985) have investigated the effect of annealing
on residual stresses in polyethylene pipes. They showed
that annealing time and temperature altered the residual
stress distribution in pipes. These stresses were

completely removed by annealing at temperatures between
100°C and 200°C for one hour. Pittman and Farah [Pittman
and Farah, 1995) carried out computer simulations of the
cooling process in pipes, including sag, thermal stresses
and morphology. These simulations provided a valuable tool

for relating product properties to processing conditions,
thus helping to optimise process operating conditions and
design. Isaac et al. [Isaac et al., 19951 carried out a
study on the mechanical properties and residual stress

distribution through the wall of PE pipes. Significant
differences in these properties have been recorded for
samples taken from different regions of the pipe wall. As
well as having the best tensile properties, samples from
the inner regions of the pipe wall (with the slowest
cooling rate) exhibited the lowest creep rates.
Density profiles through the pipe wall thickness were

determined by Gebler, Mallinson and Gwynn [Gebler, 1983;
Mallinson and Gwynn, 1988]. They observed that the
external surface had a relatively greater amorphous
content compared to the bore region. This was due to the
variation in the cooling rates for the surface and bore.
'Gebler [Gebler, 1983] reported that for a 190mm outer

diameter HDPE pipe, the external surface layer had a
density 952 kgm 3 the bore layer had density
of whereas a
of 959 kgm'3.
Page 9
Both the short term and long term mechanical properties of
polyethylene are dependent on the molecular weight and

degree of branching of the polymer. As with other
polymers, the short term properties are also dependent on
the rate of testing, the temperature of the test, the
method of the specimen preparation, the size and shape of

the specimen and the conditioning of the samples before
testing. In general, increasing the density or testing
rate or decreasing temperature causes an increase in
modulus and lowers the ductility [Brydson, 1982).
2.2.1 THERMAL OXIDATION
The rate of extrusion is slow for larger diameter pipes
and this can result in thermal oxidation of the surface
and bore. Muller and Gaube [Muller and Gaube, 1982] found
that processing under unfavourable conditions lead to
considerable oxidative damage on the bore and surface of
an HDPE pipe. The effect of this was a shorter lifetime.
This was found to be true even in the case where the
damaged layer on the inside of the pipe was of the order
of 100 micrometres. Gedde et al. and Terselius et al.
[Gedde et al., 1981; Terselius et al., 1982] report the
thickness of the oxidised layer in 200-1000 mm outside

diameter HDPE pipes to be typically about 0.1-0.2 mm. The
strength of these oxidised pipes in most cases was found
to be significantly lower than the strength of the non-
oxidised pipes.
Page 10
2.3 MORPHOLOGY
The general morphology of crystalline polymers has been
extensively studied. Polymers are considered to be either
crystalline or amorphous, although they may not be
completely one or the other.
Crystalline polymers are more rightly termed semi-
crystalline as the measured densities for perfect
materials differ from those obtained for such polymers

[Sawyer and Grubb, 19871. The dominant and most widespread
morphological entity formed when polymers crystallise from
the melt under normal conditions is the spherulite

[Keller, 1984] Spherulites consist principally of chain
.
folded lamallae radiating from a central point. Khoury and
Passaglia [Khoury and Passaglia, 19761 were among the
first to show the existence of the lamallar structure in
melt crystallised polymers. In polyethylene, spherulites
may vary in size from a fraction of a micrometre to
several millimetres in diameter, depending on the cooling
rate from the melt [Edwards et al., 1982].
It has been suggested that semi-crystalline polymers are
made up of spherulites which are composed of lamallae.
However, if the spherulites are to be considered to be
discrete entities, the mechanical properties of such
polymers would be extremely poor due to fracture along

inter spherulite boundaries. It has therefore been
suggested that lamallae are joined together by inter
crystalline links or tie-molecules [Lustiger, 1986).
Evidence for these links has been provided by Keith et al.

[Keith et al., 1980; Keith and Padden, 1984; Keith and
Page 11
Padden, 19861 who crystallised polyethylene in solution on
mica substrate.
Lustiger and Markham [Lustiger and Markham., 1983]
considered the influence of the tie-molecules in
preventing brittle failures in polyethylene under long
term loading conditions - slow crack growth. Tie-molecules
have been used to explain both ductile and brittle modes
of failure. During ductile deformation, the tie-molecules
stretch when a tensile load is applied to the lamallae, as

illustrated in Figure 2.2a [Lustiger and Markham., 1983].
At a critical point a limit may be reached as shown in
Figure 2.2b, where the lamallae break up into smaller
units, as shown in Figure 2.3. These so called `mosaic
blocks' and are directly incorporated into a new fibre
morphology according to a model advanced by Peterlin
[Peterlin, 1973]. In general, the tie-molecules act to
produce ductile deformation. In brittle failure, the
analysis follows the above where the failure occurs over a
much longer period and at lower stress levels. However,
the stress needed to obtain fibre pull-out is not attained

due to lower stress levels. Therefore, it has been
suggested that the loading situation remains as in Figure
2.2b. Tie-molecules begin to untangle and relax over long
periods. After a critical period of time, most of the tie-
molecules are presumed to have untangled, leaving behind
only a few tie-molecules which cannot maintain the applied
stress where the material consequently fails in a brittle
manner. Lustiger and Markham [Lustiger and Markham, 1983]
have also concluded that materials containing few tie-
molecules are more vulnerable to brittle type failures. Other
Page 12
T (b)
(a)
Figure 2.2 Tensile deformation applied to lamallae

[Lustiger and Markham, 1983].
t
"ý.
ý ..
ý(' 1_
f ýý
i
Figure 2.3 Incorporation of so called `mosaic blocks'

into new fibre morphology [Lustiger and
Markham, 19831.
Page 13
structural parameters affecting tie-molecules are as

follows: -
(i) molecular weight - the longer the polymer chain, the
greater the number of tie-molecules and therefore the
greater the number of tie-molecule entanglements;

(ii) co-monomer content - small amounts of monomer tend to
inhibit crystallinity due to the formation of short
branches, therefore producing improved brittle fracture
resistance;
(iii) degree of crystallinity - the higher the
crystallinity content in a material, the fewer the
amorphous intercrystalline tie-molecules and hence
increased brittle behaviour;
(iv) lamallae orientation - orientation of lamallae
perpendicular to the tensile stress direction results in a

greater tendency for the interlamellar failure compared to
the situation where the lamallae are orientated parallel

to the applied stress direction.
2.4 DEGRADATION AND STABILISATION OF POLYETHYLENE
The oxidation of PE is a well-researched topic where both
photo and thermal oxidation can occur. On exposure to
sunlight most polymers can be degraded to some extent. The
process of degradation by light, natural or artificial is
called photo-oxidation [Grassie and Scott, 1985; Allen,

1983; Amin et al., 1975; Scott, 19811. For the purpose of
characterisation of butt fused welds, an understanding of

the thermal oxidation process is important. The mechanism
for oxidation of simple hydrocarbons was established many
years ago and it is now generally accepted that polyolefin
Page 14
oxidation follows a quite similar pattern. The mechanism

involves an initiation step in which a hydrogen atom is
removed from a polymer molecule creating a free radical

R'. Propagation then follows in a series of reactions. The
first of these is a rapid reaction of R' with oxygen to

form a peroxy radical ROO'. This is then followed by the
rate controlling reaction, the abstraction of H from the
same or another polymer molecule by the ROO' radical.

Evidently each propagation step produces another polymer
radical R' which is essentially the same as the initiation
reaction. A comprehensive review of polyolefin oxidation

was given by Carlsson and Wiles [Carlsson and Wiles,
1986].
2.4.1 ADDITIVES
The additives normally present in water distribution pipes

include thermo-oxidative stabilisers, carbon black,
pigment and a UV stabiliser. Understandably, detailed

information of this nature is very seldom available to the
general public. However, the general information on the
nature of typical additives used in polyethylene

formations can be obtained from trade literature [BP,
19881 and standard text books [Brydson, 1982; Lutz, 19891.

Published trade literature for BP's Rigidex PE pipe
compounds and Solvay Eltex Tub 124 pipe extrusion
compounds have indicated that the following additives were
present [BP, 19881:
(i) ethylene homopolymer and copolymer;

(ii) other olefins and copolymers of ethyl acrylate and
vinyl acetate;
Page 15
(iii) antioxidants;
(iv) lubricants;
(v) antistatic agents;

(vi) pigments;
(vii) W stablisers; and
(viii) carbon black.
Antioxidants such as 4 aminophenol N-stearate, 4,4-thiodi-
(6-terbutyl-m-cresol) and 1,1,3-tri-(5-terbutyl-4-hydroxy-
2-methylphenyl) butane of the order of 0.3% by mass are
added to water grade polyethylene pipe to provide
protection against degradation during processing. Ideally,
the antioxidant used should not show any tendency to
bloom, bleed and discolour [Ritchie, 1972].
Pipe materials in the UK are colour coded for the pressure
pipe application using the following codes: yellow for
gas, blue for potable water and black for sewers and
chemicals. It is widely reported that titanium dioxide and
ultramarine blue are the two main inorganic pigments used

in water grade polyethylene pipes.
2.5 EVALUATION OF THERMAL DEGRADATION
Many analytical techniques can be used to monitor the
thermal degradation of polymers. As a result of

irreversible chemical reactions, changes in the molecular
composition during degradation provide the means for
following this deterioration of the polymer. The
analytical techniques for monitoring oxidation can be
divided, into two main areas, namely spectroscopic and
thermal analysis.
Page 16
The measurement of the carbonyl index using infra-red
spectroscopy is well known and widely used to study
thermo-oxidative degradation of polymers [Grassie and
Scott, 1985; Allen, 1983]. Many excellent treatments of
infra-red absorption theory exist and the text by Tadoharo
[Tadoharo, 1979] provides further details.
2.5.1 THERMAL ANALYSIS
Currently, thermal analysis is being used to measure the
stability of the polymer under processing conditions as
well as to evaluate the long-term degradation behaviour.
Differential scanning calorimetry (DSC) is a widely used
analytical tool in the study of polymers. The theoretical
aspects of the DSC are well chronicled and, its
quantitative capabilities have been established. The basic
technique involves heating an aluminium pan containing the
sample and an empty reference pan. The two pans are
maintained at the same temperature. By monitoring the
difference in energy input required to do this, it is
possible to observe both endothermic and exothermic

thermal events.
2.5.2 OXIDATION INDUCTION TIME - (OIT) TEST
The Water Industry Specification (WIS 4-32-03) has
specified an oxidation induction time test in order to
evaluate the oxidative thermal stability of water grade
polyethylene. This involves heating the sample to the

ti
desired temperature under a nitrogen atmosphere so that
oxidative degradation cannot occur. When the desired
temperature has been reached, oxygen is introduced to the
Page 17
chamber. An exotherm is observed when the sample degrades.
The time taken for this exotherm to occur is called the
OIT. The specification requires a minimum OIT of 20
minutes for MDPE pipes [WRc standard]. Variations of this
test have been used as part of evaluation of both long and
short term stability in polyethylene and other polyolefins

[Allen et al., 1988; Kramer and Koppelmann, 1986]. In butt
fusion welding of pipes, the material in the heat affected
zone is subjected to elevated temperatures. It is possible
that this region will show a reduction in antioxidant
concentration. This issue has not received any significant
attention in the literature.
Kramer and Koppelmann [Kramer and Koppelmann, 19861

investigated the use of OIT to predict service lifetimes.
The predictions were based on extrapolating the OIT values

which were generated at high isothermal temperatures to
lower temperatures using an Arhenius relationship. This
can only be used as a guide since the work has been
contradicted by Howard and Gilroy [Howard and Gilroy,
1975] who found `knees' in the Arhenius plots. This
implied that there were differences in the apparent
activation energy for thermal decomposition at different
temperatures. Howard [Howard, 1972] discussed the issue of
predicting service lifetimes of a material from elevated

temperature tests. He concluded that the nearer the
temperature was to the actual service temperature, the

better the agreement to service lifetime predictions.
However, if the test is to be used to compare the relative
performance of materials, then an extrapolation to low
temperatures is not needed and the selection of a higher

temperature is valid.
Page 18
The method of sample preparation, sample size and shape
are important parameters in conducting OIT experiments

[Marshall et al, 1973] Broutman et al. [Broutman et al.,
.
1986] carried out an OIT study on various grades of
polyethylene pipes. However, no details were given of
sample'size or shape.
2.5.3 CORRELATION OF OIT WITH ANTIOXIDANT CONCENTRATION
The measured OIT for a material is a representation of its
effective resistance to oxidation under the conditions of

the test. Allen et al. [Allen et al., 1989; 1990] carried
out the thermal oxidation of blue-water grade PE pipe via

DSC. They found that the oxidation of the pipe occurred
predominantly in the outer surface and to a lesser extent

in the bore, with little or no change in the middle
layers. Moore et al. [Moore et al., 1989] found that the
sample preparation technique strongly affected the OIT of

MDPE pipe. Samples obtained by drilling gave the lowest
and the most variable results, whereas, samples which were
obtained by sectioning gave the highest and most
consistent results. The OIT profile across a 45mm pipe
wall thickness was presented by Gebler [Gebler, 19891 and

is shown in Figure 2.4. From Figure 2.4, it can be seen
that the OIT value is dependent on the position of the
specimen in relation to the pipe wall. In addition to
this, the processing conditions were shown to affect the

OIT. The top curve represents a sample which was tested
immediately after extrusion and the bottom curve
represents the pipe which had been conditioned at 80°C for
3000 hours.
Page 19
40
0 hr straight after extrusion

v
30
H
0
3000 hr, 80 °C
20
0
".-+ 10
4J
b
r4
0
Bore Surface
Figure 2.4 Illustration of the dependence of the OIT

as a function of pipe wall for a 45mm
thick pipe [Gebler, 19891.
OIT experiments were conducted by Bjork and Stenberg
(Bjork and Stenberg, 19851 on samples taken at different
depths from thick walled vulcanised natural rubber
cylinders. Materials with the antioxidant were compared
with materials with no antioxidants added. The samples
were aged for up to 1,150 hours in air at 100°C. The
antioxidant was found to delay the on-set of oxidation in
the centre of the material by up to 900 hours of ageing.
Oxidative induction times have been used to study the
Page 20
stability of polyethylene and polypropylene in raw
material form and on a wire by Marshall et al. (Marshall
et al., 1973). Two stabilising additives were mixed with
low density polyethylene over the range from 0 to 0.20%.
The oxidative stability of the LDPE matrix was increased
by 8-fold at the 0.10% level. They also demonstrated that
the OIT was a linear function of the antioxidant
concentration.
2.6 BUTT FUSION WELDING
Butt fusion welding has been developed over the past fifty
years [Becker and Streese, 1969] with a great deal of the
early work being carried out in Germany during the Second
World War on polyvinylchloride (PVC) fabrication [Neumann
and Bockhoff, 1959]. It is one of the most common plastic
welding methods and is also known as a mirror, hot-plate
or butt fusion welding when used for joining pipes

[Lamond, 1995]. Polyethylene pipeline systems are being
extensively used in Europe by the Gas, Water and Chemical
Industries. Plastic pipeline systems are also utilised
extensively in North America, Japan and in a number of

developing countries in the Far East. Therefore there is a
need to understand factors which can influence the

integrity of butt fused welds.
Butt fusion welding has a number of advantages over other

joining methods and requires relatively simple equipment.
It is the most suitable process for joining large diameter
pipes where the welds usually have strengths approaching

that of the parent material [Grimm, 1990; Atkinson and
Ward, 19891. The major disadvantage of the process is the
Page 21
time required for each cycle. Weld times range from 10 to
20 seconds for small items and as much as one hour for 600
mm diameter pipe with wall thickness of 40 mm [Tirumalai
and Lee, 1990]. The length of time at this stage can be
minimised by automation of the process which also aids the
elimination of human error and gives greater consistency

in the quality of joints produced. Optimisation of the
parameters used are difficult as the process has been
shown to be highly tolerant of abuse [Maine and Stafford,
1988; Marshall et al., 1988; Wolters and Venema., 1982].
Butt fusion welding processes for pipe-to-pipe joining, '
procedures and equipment have evolved through many years
of experience supported by proof testing to establish that
a given process results in good joint strength and

durability. However, these developments were empirical in
nature and meant that a detailed understanding of the
microstructure of butt fusion welds has not been addressed

in any great detail [Benkreia et al., 1991]. With the
current trend towards mechanisation of the process and the
use of automatic equipment, greater control of the welding
process is possible, placing less dependence on skilled
operators and allowing more consistent results to be
achieved [Dib et al., 1992].
2.6.1 THE BUTT FUSION WELDING PROCESS
In essence, butt fusion welding is comparatively simple.

The parts to be joined are held against a hot plate to
provide the molten material, the plate is withdrawn and
the two faces are then pressed together under pressure to
Page 22
ro
a
0.2
0.1
Q) 0.0
U 200
0
150
Q3
100
W
50
H
0
::
II
'I
3
I
c2
C)
E Ti
C) Heating
U
T2 Bead-up
-1 T3
a Plate Removal
N
-4 T4 Weld time
O
0 4 10
T2 T3 T4
I'IME
Figure 2.5 Various welding parameters plotted against

time [Wolters and Venema, 1982]
Page 23
form a weld [Potente et al., 1988; Barton and Cherry,
1980; Watson2,1990]
.
The WIS 4-32-03 [WIS standard] specifies a "dummy" joint
to be made first. This is to remove contamination from the
heater plate. Maine and Stafford [Maine and Stafford,
1988] also recommend the making of a dummy joint. The butt
fusion process for joining pipes is shown in Figure 3.3
(section 3.2.2, pages 59-60). The effect of temperature,
pressure and displacement with time during welding of
components is shown schematically in Figure 2.5. This
figure shows the relationship between pressure and
temperature with the associated displacement as a function
of time. With reference to the time axis, the overall
heating time T1 + T2 corresponds to the heating and bead
up time. The plate removal time is indicated as T3 and the
joining and cooling time is coded T4. Critical aspects of
the butt fusion process are the melt temperature and the
melt volume. These parameters, in turn, depend on the
heater plate temperature and the heat soak time [Bucknall
et al., 1980; Harris, 1966; Barber and Atkinson, 1972;
1974; Atkinson and DeCourcy, 1981; Pimputkar, 1989;
DeCourcy and Atkinson, 1977]. Welding pressure [Barton and
Cherry, 1980; Barber and Atkinson, 1972; 1974; Atkinson
and DeCourcy, 1981; Potente and Tappe, 1988] and proper
alignment of the components also have an influence on the
joint strength. Weaknesses in the weld may arise due to
any one or combinations of the following:
(i) incomplete adhesion may be due to insufficient
heating, the welding temperature being too low due to
cooling, the heating time being too short or insufficient
Page 24
pressure [Bucknall et al, 1980; Atkinson and DeCourcy,
1981; Pimputkar, 1989];
(ii) void formation may be due to entrapment of air,
evolved degradation products as a result of excessive
heating, heater plate and pipe-face contamination by dust
and grease, too high a heater plate temperature and
thermal contraction of the material on cooling [Atkinson
and DeCourcy, 1981; Ziatsev, 1972; Ziatsev and Bukin,
1973] ;
(iii) narrow or thin welds may be due to excessive welding
pressure [Atkinson and DeCourcy, 1981; ];
(iv) axial misalignment of the weld [Parrnar, 19861; and

(v) pipe movement during the cooling phase of the butt
fusion process.
2.6.2 THEORY OF BUTT FUSION WELDING
There are a number of theories that attempt to explain

adhesion or autoadhesion of polymers:
(i) diffusion theory - assumes thatýthe molecules or chain

segments diffuse into each other. The diffusion theory is
applicable if the following conditions are fulfilled:

(a) the polymers must be mutually soluble and
(b) the macromolecules must have sufficient mobility.
(ii) viscoelastic contact theory is based on the fact

-
that the surfaces of the parts to be joined are brought so
close to each other that Van der Waals forces, such as
dipolar, induction, dispersion or hydrogen bridge bonds
come into effect.
Page 25
(iii) adhesion theory - an energy input is needed to
increase the surface of a weld isothermally which can be
related to surface tension. When two surfaces are brought
into contact with one another, a contact surface energy is
produced. Maximum adhesion is obtained when the specific
contact surface energy (surface tension) is zero, that is
when the materials are identical.
(iv) flow process theory - assumes that during the joining
process the welding temperature creates flow fields, i. e.

flow movements [Potente and de Zeeuw, 1979] As a result,
.
the thermal displacement processes are superposed by
mechanical displacements which can become so intense that
mixing occurs in the joining zone. These processes are

said to be faster than diffusing or viscoelastic contact
processes.
Potente [Potente, 1977] considered the diffusion theory

[Voyutskii, 1963; Kinloch, 1987] and viscoelastic contact
theory [Anandu and Karam, 1969; Anandu and Dipzinski,
1970] to explain the adhesion of polymers. He suggested
that both of these mechanisms occur concurrently during
the butt fusion welding. Brinken, (Brinken, 19823 like
Potente, [Potente, 1977] reviewed the diffusion theory

[Voyutskii, 1963; Kinloch, 1987] and viscoelastic contact
theory [Anandu and Karam, 1969; Anandu and Dipzinski,
1970] and in addition considered two other theories:
adhesion (Kaelble, 1971] and flow process theory [Potente,
1977]. He established a set of criteria for the
compatibility of various polyethylene resins. Brinken

[Brinken, 1982] stated that for the butt fusion joining of
pipes, it was necessary for the melt flow indices (MFI) of
Page 26
the polymers concerned to be similar. Materials with a
difference in MFI value could be joined by using two
independent heated hot-plates. Both the plate temperatures
and contact times can be controlled to match the degree of
melting of both materials [Grimm, 1990; Potente and

Gabler, 1986]. Malguarnera and Earles [Malguarnera and
Earles, 1982] also favoured the diffusion theory. They
believed that the strength of the joint was due to
diffusion of molecular segments from one pipe to the
other. Wool et al. [Wool et al., 1989] agreed that both
viscoelastic flow and diffusion occur to some extent

during welding. They stated that when two molten polymer
surfaces are brought into contact, wetting or molecular
contact resulted in Van der Waals attraction followed by
interdiffusion of chain segments across the interface
[Yuan and Wool, 1990].
Dodin [Dodin, 1981] in considering autohesion, suggested

that molecular contact was required and that it was
necessary to remove the contact surfaces to achieve good

bonding. This may be achieved by heating the polymer to a
viscous state and then physically causing chain

interdiffusion by applying pressure [Grimm, 1990].
Grandclement [Grandclement, 1989] whilst considering
electrofusion joining suggested that the extent of
molecular diffusion increases with an increase in joining
temperature or heating time. Maine and Stafford [Maine and

Stafford, 1985] also stated that polymer diffusion and
molecular mixing occur readily in the molten state.
Page 27
The dependence of molecular mobility on joining
temperature can also be influenced by the molecular
weight. Stokes and Hobbs [Stokes and Hobbs, 1989] noted

that the molecular weight affected the diffusion and hence
the joint strength. They found that higher molecular

weight materials required longer heating times and higher
temperatures for diffusion to occur. Watson [Watson',
1988] and Dodin [Dodin, 1981] considered the molecular
weight and the temperature effects and stated that as the
viscosity of the polymer decreases the molecular diffusion
could occur more readily. Dodin [Dodin, 1981] suggested

that autohesion occurred at the joint interface to a depth
of between 0.1 and 1µm.
Similar effects were observed by Potente and de 'Zeeuw

[Potente and de Zeeuw, 19791 and Neubert and Mack [Neubert
and Mack, 1973]. They found the weld quality to depend

largely on the relative flow of molten material displaced
from the weld as the two parts are pushed together. This
shearing motion generates intermixing and also helps to
push contamination away from the centre of the weld. Much
of the research was concerned with producing a perfect

joint interface indistinguishable from the parent
material. However, Maine and Stafford [Maine and Stafford,
1985] have observed that even in a good welded joint, an
interface between the various components can be
distinguished.
2.6.3 MICROSTRUCTURE OF PLASTIC WELDS.
Microstructural studies on butt fusion welds have been
carried out by a number of researchers [Bucknall et al.,
Page 28
1980; Barton and Cherry, 1980; Barber and Atkinson, 1972;
1974; Atkinson and DeCourcy, 1981; DeCourcy and Atkinson,
1977; Potente and Tappe, 1985; Menges and Zohren, 1967].
Menges and Zohren [Menges and Zohren, 1967] used
transmitted polarised light microscopy and identified four
zones in polyethylene welds, see Figure 2.6a. They stated
that the different zones were due to temperature effects.
Barber and Atkinson [Barber and Atkinson, 1972; 1974; ],
Atkinson and DeCourcy and DeCourcy and Atkinson [Atkinson
and DeCourcy, 1981; DeCourcy and Atkinson, 1977] observed
the microstructural features in microtomed sections of
polyethylene welds using transmission optical microscopy
and scanning electron microscopy. The microstructures were
explained on the basis of temperature gradients in the
region of the welds and the flow of the molten material

during welding. The etched surfaces of the weld exhibited
five different zones and are illustrated in Figure 2.6b.
The five weld zones were identified because of

differential attack by the chromic acid. The differential
attack was explained in terms of the amorphous or the low
molecular weight fraction being attacked first leaving the
crystalline material behind [Barber and Atkinson, 1972;
1974]. They suggested that there were columnar structures

in zone 3. They also proposed that long term weakness
could result because of the columnar structure which
caused separation at the boundaries. However, the later
work of DeCourcy and Atkinson [DeCourcy and Atkinson,
1977] suggested that zone 3 represented an area of
molecular orientation in the welds and it was this fact
that determines the extent of the etching in chromic acid
rather than any differences in crystal structure. The
Page 29
3 Hottest Material
4 Boundary
2 Cooler Material Nucleation
i1
li Reminant of Skin 2 Spherulitic,
Slightly Elongated
4 Boundary Layer 1 Reminant
of Skin
3 Columnar
1
Surface Cooling
5 Spherulitic
Figure 2.6 A schematic illustration of the five zones

in PE welds (Atkinson and DeCourcy, 1981)
latest view according to Atkinson and DeCourcy [Atkinson
and DeCourcy, 19811 is that a butt fusion weld with its

inevitable molecular orientation would be more susceptible
to cracking due to external loading or environment effects
than the rest of the pipe. Barton and Cherry [Barton and
Cherry, 1980] applied the same techniques as Barber and
Atkinson [Barber and Atkinson, 1972; 1974] to study the
morphological features of butt welds in polyethylene
pipes. They also considered alternative etching methods
and found etching in toluene vapour for 15 seconds to
support the observation made under the chromic acid etch.
Page 30
In their opinion, the degree of attack by the etchant
depended on the angle the flow lines made with the
boundary. The degree of the flow of the melt was detected
in the microtomed section of the weld taken between
crossed polarisers. They suggested that the contrast

between the weld and the melt affected zone (MAZ) was due
to the difference in orientation of the material. Folkes
et al. [Folkes et al., 1991] also found the presence of

flow lines in polyethylene welds. They found the molecular
orientation close to the weld was not parallel to the weld

line but at ± 450 suggesting that shearing flow had
occurred. Stevens [Stevens, 1993] found the crystalline
phase in polypropylene to be more easily orientated by the
welding process than the amorphous phase. The orientation

direction 'changed from along the axis of the pipe to the
flow direction of the welds made under the manufacturer's
recommended conditions. With the non-optimum conditions,
less orientation occurred and at the weld centreline the
orientation was the same as in the pipe. DSC was used to

determine the crystal structure in the welds to provide
information on the spherulite size/maximum lamellar
thickness, spherulitic size distribution and the
distribution of lamellar thickness [Stevens, 1996].
Bukin [Bukin, 1972] also observed orientation in the weld

zone by viewing microtomed sections using polarised light
microscopy. He found that the changes in direction of
orientation of material at the boundary of the MAZ between
the parent material and that within the weld was between
600_700. This was also said to give rise to
microstructural changes in the weld region. The
spherulites in the weld were elongated and orientated
Page 31
perpendicular to the parent material. Bukin also observed
a region of partially molten spherulites at the MAZ
boundary. These differences were attributed to variations

in the cooling rate across the weld as it was formed. Low
et al. [Low et al., 1978] investigated the microstructure
and mechanical properties of compression moulded HDPE
plaques and found that the difference in microstructure

from edge to centre was related to the difference in the
cooling rate. The slow cooled specimens gave the
characteristic fracture surface associated with brittle
failure. A similar type of behaviour was also observed in
injection moulded specimens of HDPE when the mould

temperature was varied from 35°C to 80°C. Tensile tests
showed the elongation to failure of the specimens prepared

with mould temperatures of 35°C and 80°C to be 184% and
79% respectively. Both types of specimen necked and the
fracture surfaces were similar except for the fact that
the specimens moulded at 80°C always started to fail in
the central region. When the microstructure was studied,

it was found that the central region did not show banes
structure [Low et al., 1978]. The differences in
mechanical properties were attributed to the effect of the
moulding temperatures on the resulting microstructures.

Gehde et al. [Gehde et al., 1992] have also identified a
recrystallised MAZ separated from the parent material by a
region of deformed spherulites for polypropylene samples.
They investigated the deformation modes of the weld using
a micro-tensile test and found a number of preferential
deformation sites in and at the vicinity of the joint.
Egen [Egen, 1987] has also identified a number of

morphologies as a function of the welding conditions used.
In an investigation on polypropylene, Kamenova et al.
Page 32
[Kamenova et al., 19881 used micro-hardness measurements

to establish the existence of a MAZ in a parent material
next to the weld. This region was found to have a low
strength due to a localised overheating and also exhibited

low hardness values. Stevens [Stevens, 1995] studied the
morphological features of butt welds in polypropylene

pipes and found deformed spherulites separated the weld
from the adjacent parent material. However, in vibration
welding, the weld was featureless and there was a sharp
transition between the weld and the parent material.
Parmar and Bowman [Parmar and Bowman, 1989] found that the
microstructure had little influence on the crack
propagation path within the weld. During mechanical

testing, the samples failed in this region. Nieh and Lee
[Nieh and Lee, 1993] found three zones around the weldline
in polypropylene which they define as stressless
recrystallisation area, columnar area and the slightly
deformed area. The size of the MAZ under different heating
conditions was established. The presence of voids in the

joint interface was attributed to a high heating
temperature and prolonged heating time.
Stevens [Stevens, 1988] and Watson [Watson', 1990] used

Fourier Transform Infra-Red spectroscopy to analyse and
quantify the variations in weld crystallinity due to the
thermal history of the welding cycle. They found'a change
in crystallinity across the pipe walls. The crystallinity
within the weld zone was fairly constant but was found to
be lower at the centre of the weld.
It is difficult to conclude from the review the

of
literature, the ideal microstructure necessary for a good
UNIVERSITY
BRUNEL LIBRARY 33
Page
weld. This is because a number of contradicting factors
have been stated by different authors as the parameter(s)
that would yield a good butt joint. The shear zone has
been identified by a majority of researchers as the
microstructural feature. that has a significant influence
on the mechanical properties of butt-fused joints. It is
also apparent from the literature that attention needs to

be given to the test method and test geometries that are
used to evaluate the mechanical properties of butt-fused

joints.
2.6.3.1 SPECIMEN PREPARATION
There is little published information on techniques for
microtoming MDPE and HDPE pipe sections. Most of the

literature in the public domain deals with techniques for
obtaining small thin sections of biological and polymeric

samples. Good quality microtome sections are necessary if
microstructural information on the pipe and weld zones are

required. The quality of the microtome sections depend to
a large extent on the quality of the microtome knife. Day
and Gitos [Day and Gitos, 19871 recommend samples of 8-10
µm and the cutting of these may be aided by cutting at low
temperatures, for example -30°C.. The presence of

scratches, chatter, compression and holes are some of the
problems that occur with microtoming. Scratches are the
result of an uneven knife edge whereas chatter arises from

fine periodic variations in thickness within a section
resulting from friction between the knife and the specimen
block.
Page 34
2.6.4 FACTORS INFLUENCING BUTT FUSION WELDING
The weld is usually the weakest point in any fabricated
system and is often the site of brittle fracture in
stressed applications (Bucknall et al., 19801.
Consequently a knowledge of factors which have an

influence on the joint strength is useful in trying to
obtain the optimum conditions for joining. With these aims

in mind a number of researchers have carried out work on
the analysis of the joining process.
DeCourcy and Atkinson [DeCourcy and Atkinson, 1977]
carried out work on polyethylene pipes and found that
insufficient heating produced weld fractures which were

brittle in nature. Welds obtained with excess heating had
entrapped gas and foreign bodies in the weld zone, thus
reducing weld integrity and strength. A maximum weld
strength was obtained with intermediate heating time.
Watson and Murch [Watson and Murch, 1989] studied the
effect of welding temperature on polypropylene, high
impact polystyrene and polyphenylene oxide. They found the
strength to be dependent on the welding temperature.
The importance of both heater plate temperature and the

heat-soak time was reported by Haworth et al. [Haworth et
al., 1988]. Combinations of low temperatures and short

heating times gave weak welds whilst higher temperatures
and long heating times gave rise to degradation and

interfacial voids. Pimputkar [Pimputkar, 19891 outlined a
joining parameter by linking bead volume, wall thickness
and the volume of melt. This highlighted the fact that low
temperatures gave rise to welds with lower strengths.
Page 35
In order to demonstrate the effects of welding conditions

on the melt flow zone, Bucknall et al. (Bucknall et al.,
1980] used specimens of different colours to illustrate
the flow path of the material. They discovered that at a

low displacement, the amount of flow was small, trapped
air remained in the weld and the weld bead was small. At
the highest displacement the maximum flow had occurred and

the weld bead was large. The maximum strengths were
achieved with intermediate pressures or displacements
which were sufficient to expel trapped air yet left a band

of molten material to form the joint. Watson [Watson,
19881 and 2] found that limited

only relatively amounts of
flow actually occurred within the weld in the range of
pressures examined. This was found to be dependent upon

temperature rather than on pressure or displacement. This
seems somewhat odd since the flow of the molten material

during the welding operation depends on the pressure.
Watson [Watson2,1988] found that higher temperatures
produced larger weld beads. In most joints, the material

at the edges was found to be pushed out to form the weld
bead with the original interface being stretched in the
process. In most joints the material at the centre of the
weld was said to remain undisturbed with the material at

the edges being pushed to form the weld bead. The
temperature dependence of this process showed that melt
viscosity played a large part in determining the geometry
of the weld bead. According to Watson (Watson 3,1990] the
temperature at which the two halves of the joint are
brought together must be above the crystallisation
temperature (120°C for PE) for sound welds with good
mechanical properties to be obtained. As long as this
Page 36
requirement is met then variations in the welding
parameter were said to have little effect.
Weld pressure also has a marked effect on the strength of
welds [Barber and Atkinson, 1972; 1974; Potente and Tappe,
1985]. Voyutskii [Voyutskii, 1963] considered the effects
of the welding pressure and welding time on the joint
strength. He found that increasing pressure increased the
joint strength up to a maximum value. Above this value no
gain in strength was apparent. It was also found that at

lower temperatures, the pressure had little influence on
the weld strength, but at higher temperatures, increases
in strength occurred more readily with increased pressure.

As the weld time was increased, a rise in joint strength
to an upper limit was observed until at excessive hold
times the material began to degrade. This was attributed

to molecular diffusion between the two components being
time-dependent. DeCourcy and Atkinson [DeCourcy and

Atkinson, 1977] also reported a sharp drop in joint
strength at low welding pressures. This was viewed as an

indication that insufficient contact of the joint faces
had occurred during welding. The effects of welding
parameters on the butt welds were investigated by Lin and

Benator [Lin and Benator, 1993]. Increasing the heating
time increased the weld strength and the heating
displacement. This was attributed to an increase in the
thickness of the molten layer. Increasing the bead-up
pressure decreased the weld strength due to squeezing of

the molten polymer thereby decreasing the thickness of the
molten layer which resulted in weaker welds. Increasing
the plate removal time decreased both the weld strength
and welding displacement. This was explained on the basis
Page 37
that during the plate removal time, the molten polymer was
cooled due to conduction to cooler portions of the sample
and convective losses to the air. Increasing the heating
time increased weld strength and increasing the welding
pressure decreased the weld strength. Longer joining times
enabled the weld to cool and resolidfy under pressure

thereby increasing the strength. Higher joint pressure
resulted in the molten polymer being squeezed out as well
as adverse molecular orientation.
On a more microscopic level, Davey [Davey, 1988]

investigated the effect of weld-line defects on the
resistance to fracture. Chalk was used on the heater plate

to provide a controlled degree of contamination. It was
found that even relatively small amounts of contamination
caused a rapid deterioration in the fracture performance
of the joint. The predominant failure mode under impact
loading was brittle fracture across the weld. Watson
[Watson2,1988] proposed that because only relatively
little flow of molten polymer actually occurred in the
weld zone, any contaminants in the centre of the weld, may
not be expelled from the weld zone as the bead is formed
thus leaving an inherent weakness in the joint.
Barker et al. [Barker et al., 1983] have shown the

presence of large inclusions in the pipe wall to have a
detrimental effect on the pipe performance. A number of
likely sources for these inclusions were speculated upon,
including catalyst residue, materials handling and
contaminants from processing. The study of the influence
of defects on the mechanical. properties of butt fusion
welds in MDPE was carried out by Girardi [Girardi, 1992]

.
Page 38
Various test geometries with known defects, were evaluated
and compared to the British Gas Standard [BG standard) for

butt fusion welds. It was concluded that small defects
were more significant as the pipe wall thickness
increased. At pipe wall thicknesses corresponding to
greater than 25mm, defects less than 1mm in diameter were

seen to result in brittle behaviour. Lu et al. [Lu et al.,
19921 investigated the effect of welding pressure and
surface contaminants on butt fused joints via slow crack
growth resistance tests. Contaminants such as Vaseline,
graphite and Teflon were coated on the surface and weld

pressures of 1.61 and 0.31 MPa employed. It was found that
welds produced with the high pressure resulted in better
slow crack growth resistance. At high weld pressure, the

samples which were contaminated with Vaseline and graphite
showed a negligible effect on weld quality. Teflon was
found to reduce the lifetime significantly especially at
low weld pressure. This was attributed to insufficient
melt movement. Sufficient weld pressure was said to be

required in order to squeeze out the thin "skin layer"
that occurred at the weld interface, along with any
contaminants.
DeCourcy and Atkinson [DeCourcy and Atkinson, 1977] noted

that melt flow rate (MFR) itself had an influence on the
welding characteristics. As the MFR decreased, the optimum
welding temperature and the lower limit of the welding
temperature were said to be raised. This has the effect of
decreasing the range of temperatures over which the
material can be welded. In effect this means that
materials with high MFR require less rigorous control of
the welding parameters as they are more tolerant to the
Page 39
welding conditions used. McGreevy [McGreevy, 1995]
investigated the effect of melt index and density on the
weld strength of HDPE. He found the weld strength to be
less for higher melt flow index materials. The density of
the polymer was also found to affect the weld strength.

For constant melt index, increasing the polymer density
resulted in a higher percentage of tensile failures. For
polymers with melt flow index values in the range and
densities of 935,952 and 963 kgm-3, the percentage
tensile failures rose with increasing density ranging from
20% to 40% to 100%. This was explained on the basis that
higher density polymers had a greater proportion of
crystalline region in their structure. Furthermore, in the
highly crystalline materials the melting temperature
region is narrower than for polymers of lower
crystallinity. In welding, the MAZ was found to be thinner
for HDPE materials than those of lower density. This led
to more brittle welds and as a result the weld strength
was lower for high density' materials.
2.6.5 TEMPERATURE DISTRIBUTION WITHIN THE WELD
Due to the difficulties involved in measuring temperatures

in or close to the weld region, relatively little work has
been reported on the temperature distribution within the
weld zone. Shillitoe and Wood [Shillitoe, 1991; Wood,
'1993] carried out work in this area to validate

theoretical models. Wolters and Venema [Wolters and
Venema, 1982] used small thermocouples placed around the
pipe wall to record temperature variations throughout the
welding cycle. They found that during the heating phase,
the temperature did not reflect the heater plate
Page 40
temperature. Furthermore during the change-over, as the
heater plate was removed, the temperature in the melt
dropped by about 109116. When the pipe ends were butted
together and pressure applied, the initial temperature
drop was found to be approximately 120°C within 20
seconds. After this, the cooling was seen to be much
slower throughout the remainder of the welding period. The
temperature within the weld was 60°C after 13 minutes.
This rapid initial cooling was also observed by Watson
[Watson2,1988] who notes that the initial rate of cooling
as material is pushed out to form the weld bead is much
faster than if the melted pipe ends were allowed to cool
in air, hence the welding process is even reasonably
tolerent of extended heater plate removal times.
2.6.6 BUTT FUSION %1ELDIlIGOF DISSIMILAR GRADESOF POLYETHYLENE
DeCourcy and Atkinson [DeCourcy and Atkinson, 1977], and

El Barbari et al. [El Barbari et al., 1986] looked into
the welding of dissimilar grades of PE due to the

increasing usage of polyethylene in pipe systems. The main
emphasis was for repairing and joining pipes of different
resin grades. Four polyethylene resins were used, each
requiring similar welding pressures. Only the temperature
and time needed to make the weld were the variables.

Welding was carried out by using conditions which were
appropriate to the material with the lower MFI. This meant

that-the higher MFI pipe was in effect, being joined at a
higher temperature than normal. Weld quality was assessed
using tensile tests and failure was observed to occur in
the higher MFI component and not in the weld zone. This
indicated that by balancing the parameters of pressure,
Page 41
temperature and time, welds could be produced between
dissimilar grades of polyethylene. They concluded that
joints could be made where the weld was at least as strong
as the parent material [DeCourcy and Atkinson, 1977].
Stokes [Stokes, 19931 used a dual platen machine in which
different hot-plate temperatures were used for obtaining
high strength in welds between dissimilar materials. In
addition to the hot-plate temperature, the weld strength
was found to depend on the heat-soak time.
Gabler and Potente [Gabler and Potente, 1980] used a
selection of six materials to make welds between amorphous
and semi-crystalline materials. They used two
independently controlled heated plates to control the
temperature and heat-soak time. Welds were produced and

tensile tests were used to compare the welding factors
(this is discussed in section 2.6.8.1) in comparison to
the weaker partner. Some successful welds were produced

but many fractured on cooling indicating that the material
properties had a large influence on weldability. Various
factor were considered which could have an influence on
the weldability. However, a more positive correlation was

found with thermal expansion. They concluded that, ideally
the coefficient of thermal expansion of the two materials
should be similar to obtain the strongest joints.
2.6.7 BUTT FUSION WELD QUALITY AND TESTING
The most frequently used method for checking a weld is a
visual inspection. Although this is very simple and
subjective, the shape and size of the weld bead can give a
good indication of some parameters relating to joint
Page 42
quality [Maine and Stafford, 1988; Barber and Atkinson,
1972; Potente, 1977; Barton and Cherry, 1983; Yeh, 19961.
With reference to visual inspection, the weld bead should
be relatively uniform and symmetrical around the
circumference of the pipe. Unevenness in the bead
dimensions is usually caused by poor joint preparation.
Visual inspection is limited to detecting welds which show
surface defects, insufficient melting, discoloured
material due to overheating, surface cracks and flaws due
to contamination or mechanical damage.
Two of the most important non-destructive testing methods
for polyethylene butt fusion welds are x-ray radiography
and ultrasonic testing. Both of these techniques have been
used with a varying degree of success (Maine and Stafford,
1988; Venema and Wolters, 1991; Yeh, 1966]. Radiography
only shows major contamination and voids while ultrasonic

inspection proves very difficult to interpret. Ultrasonic
tests are also difficult to apply in a practical manner

due to the attenuation characteristic of the polymer
involved.
2.6.8 MECHANICAL PROPERTIES
In order to assess the weld quality, routine quality
control and/or the suitability of the joining method, a

test method is required which gives good differentiation
between good and poor welds. Due to the long servihe life
of a pipe and the pipelines, lifetime testing of a
completed weld is a very lengthy process [Menges et al.,
1987]. Short term tests are therefore advantageous for
establishing weld quality. These tests should, first,
Page 43
relate to long term lifetime tests and secondly, be able
to differentiate between welds of high and low quality
[Diedrich and Kempe, 1980; Hessel and John, 1984].
2.6.8.1 TENSILE TESTING
The tensile test is a popular test applied to the routine
evaluation of the welds. In the case of the butt welded
pipe, longitudinal samples are taken from the welded
region such that the weld-line lies in the middle of the
test specimens. The usual procedure is to carry out a
tensile test with the weld bead removed. A welding factor,
(f) defined by the equation is then reported [Barber and

Atkinson, 1972; 1974]:
f= yield strength of welded material (2.1)

yield strength of parent material
DeCourcy and Atkinson [DeCourcy and Atkinson 19771] used

tensile bars machined from the pipe to assess the weld
quality. The results, reported in terms of welding

factors, showed little changes in (f) within the welding
condition ranges examined. The highest variation in the
welding factor was found to be between 0.89-0.95 for the
welding temperature variations between 150°C to 250°C. For
the other grades of polyethylene tested, (f) was found to
be between 0.98-1.00 and 0.96-0.98 when the welding
temperature was varied in the similar range. With debeaded
specimens, they found that the good quality welds gave
elongations of over 500%. Malguarnera and Earles

[Malguarnera and Earles, 1982] reported the results in
terms of yield stress. A significant reduction in the
Page 44
yield stress for the butt welded specimen was apparent
under the following conditions: (i) very low welding
temperature (204°C) and high welding temperature (370°C);
(ii) low heating time (5 seconds); (iii) low heating
(3.5MNm 2) and (iv) low fusion (14MNm 2)

pressure pressure .
Hale [Hale, 1990] recommended the use of a dumb-bell test-
piece with the weld being located at the centre of the
wasted section. The Water Research centre (WRc) and
British Gas use a test specimen that has been adopted for
welded joints, see Figure 2.7. As shown in Figure 2.7,
small circular ended side notches ensure that the
deformation starts in the weld. However, as deformation
continues, parent material tends to be pulled out from one
side of the weld and the weld itself does not deform
[Hale, 19901. Furthermore since a gauge length can not be
defined, it is difficult to measure elongation accurately.
Hale also recommended that the beads should be removed
prior to testing. This was said to eliminate effects
arising from changes in the size of the weld reinforcement
and in removing the stress concentration in the vicinity
of the bead and the parent material.
Watson [Watson2,1988] recorded high elongation values for
WRc type specimens having circular side notches with the
weld bead retained. These value were only half of that
achieved by the parent material due to the fact that the
specimen yielded and failed only on one side of the weld.
The weld strengths were found to be similar to that of the
parent material and failures were ductile with fracture
being initiated at the weld bead. Wolters and Venema
[Wolters and Venema, 1982] used tensile tests to
differentiate between various welding conditions. They
Page 45
Weld line
10/
16 215 16 75mm
ý
10
160mm
Figure 2.7 WRc's butt fused test piece for tensile

testing [Hale, 19901.
found in all tests, yielding started away from the weld.
The weld strength was greater than that of the parent
material with failure occurring away from the weld.
However, some researchers [Wolters and Venema, 1982;
Greig, 1977] prefer not to use the tensile test for
assessing the optimum welding conditions for butt welds.
This is because the results obtained using the test method
vary only slightly as exemplified by the work of Barber
and Atkinson, and DeCourcy and Atkinson [Barber and
Atkinson, 1972; 1974; DeCourcy and Atkinson, 1977] and
Malguarnera and Earles [Malguarnera and Earles, 1982]. The
welding factor method overestimates the weld strength (f
of 0.95 1.00) and it only detects a very bad weld (butt

-
welds made under poor/extreme conditions). Table 2.1 gives
a summary of preferred tensile test geometry for testing
welds by various researchers.
Page 46
Menges et al. [Menges et al., 1987] found the degree of

differentiation in tensile tests can be increased by
lowering the test temperature or by increasing the speed
of the test. Potente and Tappe [Potente and Tappe, 1988]
temperature °C to the
used a of -30 evaluate strength of
the fused joint.
Narisawa and Nishimura [Narisawa and Nishimura, 1993]
conducted tensile and impact tests on butt fusion pipes to
evaluate the short-term performance of the joints. The

joint strength of high molecular weight MDPE pipe was
greatly reduced. The joint strength was found to be
related to the heating conditions. The impact strength

decreased when the heating temperature was below 170°C
and/or the heating time was less than 30 seconds, and/or

the applied pressure exceeded a certain value.
According to Hale [Hale, 19901 the notch location is of

prime consideration when performing impact tests on
plastic welds. With large weld beads, it becomes
increasingly difficult to locate the central line of the
weld. However, when the weld regions are very small, the
notch necessary for impact testing extends outside the
weld area and thus the measured energy will not be
representative of the weld properties. It is therefore
recommended [Hale, 1990] that the weld bead be removed. It
was also recommended that the notch root tip radius be
sharper than 0.25mm and that the fractured specimens be
viewed under a microscope to obtain the maximum amount of

information on the failure.
Page 47
According to Pimputkar [Pimputkar, 1989], the impact tests
relate to joining conditions more closely than tensile
test methods. Instrumented falling weight impact testing
was carried out by McGrath [McGrath, 1991] to assess the
weld quality. He found that the weld centre-line was

inherently weak compared with the surrounding parent
material.
Tensile Test References
Geometry
Dumb-bell Barber and Atkinson, 1972; Atkinson and DeCourcy,
specimen 1981; DeCourcy and Atkinson, 1977; Low et al.,

1978; Hawarth et al., 1988; Hale, 1990; Nariswa
and Nishimura, 1993
WRc specimen DeCourcy and Atkinson, 1977; Potente and Tappe,
1985; Watsonl and 2j 1988; Girardi, 1992; El

Barbari, 1986; Hale, 1990
Table 2.1 List of preferred tensile test geometry

for testing welds by various researchers.
A new method of analysis was developed independently by

Williams (Williams, 1984] and Ward (Ward, 1983] using a
linear elastic fracture mechanics (LEFM) approach.
Bucknall et al. [Bucknall et al., 1980] combined LEFM and
tensile tests by using a razor-notched tensile specimen
for welded samples. They were able to distinguish between
low weld strengths caused by weld defects and those due to
weaknesses of the material associated with orientation.
Page 48
Marshall [Marshall, 1991] has tested a series of notched

impact specimens made according to WRc 4-32-03 [WIS
standard]. Samples were cut from a 12mm compression
moulded sheet of polyethylene Rigidex 002-50 PE resin
welded under standard conditions. The critical strain

energy release rate-Gc values showed very little change in
behaviour as compared with the parent material. However,
some test samples gave much lower G, values and these were
attributed to the fact that they were the only samples
notched at the interface. Folkes et al. [Folkes et al.,
1991] used charpy impact and instrumented falling weight
impact to test welded joints. Notches were also introduced
by placing lengths of PTFE tape into the joint after

trimming. They found that the two test methods gave
similar results.
In summary, the butt fusion process for pipe-to-pipe

joining, procedure and equipment has evolved through many
years of experience supported by proof testing to
establish that a given process resulted in good joint
strength and durability.
There is a need to understand fully the development of

microstructure under different welding conditions. The
correlation of this microstructure with tensile testing is
of significant importance.
Page 49
EXPERIMENTAL TECHNIQUES
3.0 PIPE MATERIALS
Two different commercial grades of pipes were used in this
study, namely, Eltex Tub 124 and BP Rigidex 002-50. These
materials were used as they represent the majority of pipe
resins in the UK. The pipes had an outside diameter of

180mm and a standard dimension ratio (SDR) of 11. The SDR
is defined as the ratio of the outside diameter to wall
thickness of the pipe. The technical data sheet for these
materials are presented in Appendix 1.
3.1 MATERIALS CHARACTERISATION
Various techniques were used to establish basic material
properties of the two resins used in the welding triäls.
Granulated pipe samples were used in the measurement of
the molecular weight distribution by gel permeation
chromatography. The crystalline melting temperature was
measured by differential scanning calorimetry using
samples taken from the outer, mid and bore of the pipe.
The density was determined using a density gradient
column.
Page 50
3.1.1 DENSITY GRADIENT COLtTh
The density gradient technique is frequently employed as a
rapid means of determining the densities of plastics
accurately. The basic principle of the technique depends
upon the hydrostatic equilibrium between a solid and a

liquid of identical density. The apparatus consists of a
column which contains a mixture of two liquids whose

densities differ. The density of the liquid in the column
,
varies in a linear scale from the top to bottom of the
column. Glass markers (beads) of known density provided

points of reference at intervals throughout the height of
the column. The floats were used to calibrate the column
against a vertical scale. The entire column was maintained

at a constant temperature of 23°C by a water bath.
MAZ
i II
IIli
III
20mm
Figure 3.1 Location within the pipe and welded regions

where the samples for the density
measurements were obtained.
Page 51
In this experiment, the two liquids used were isopropanol,
having a density of 0.784 gml-1 and diethylene glycol, of

density 1.118 gml-1, to give a density range of 0.93 to
0.98 l
gml-.
The density of the samples were obtained from the
specified regions in the pipe and welded regions shown in
Figure 3.1. The specimens were approximate squares
corresponding to 25mm2. The specimens were introduced into
the column by lowering them very slowly in a basket and

left to equilibrate for a period of 24 hours. Once the
equilibrium position had been reached, the relative
position of each specimen was recorded and their density
evaluated by comparing these positions to those of the
glass floats of known density. Three individual
determinations were made for each of the samples.
3.1.2 GEL PERMEATION CHROMATOGRAPHY
The gel permeation chromatography was carried out by RAPRA

(the Rubber and Plastics Research Association). Granulated
samples of each pipe resin were used for the analysis.
3.1.2.1 GPC OPERATING CONDITIONS
Sample solutions were-prepared by adding 50m1 of solvent

to 50m1 of sample and boiling under reflux for 20 minutes
to dissolve the samples. A part of each solution was then

filtered through the fibre pad at 180°C and the filtered
solutions were re-heated to near boiling immediately prior

to the chromatography. It was assumed that the non-
polymeric material present was negligible and that the
Page 52
concentration was not affected by filtration: both of
these assumptions are open to question.
Two determinations were carried out on each sample to
check for consistency and the operating conditions were as

follows:
Column; P1 gel (2 x mixed gel) 10 .tm
packing and 30cm column.

Solvent; 1,2 dichlorobenzene with
antioxidant.
Flow-Rate; 1.0 mlmin-1.
Temperature; 140°C
Detector; Infra-red.
Calibration; polystyrene standards.
3.1.3 DIFFERENTIAL SCANNING CALORIMETRY
3.1.3.1 MELTING POINT RANGE AND ENTHALPY OF FUSION
A Perkin Elmer DSC-2 with a data station was used in this
programme to measure the melting point range and enthalpy

of fusion of the virgin polymer pellets and pipes.
In the DSC instrument the specimen under test and the
reference (empty sample pan) are placed in similar

environments and maintained at the same programmed
temperature by balancing the power fed to each specimen.
The amount of energy which is required to maintain the
specimen at the temperature of the reference even during
the thermal event is recorded. If the energy differential

between the amounts of energy absorbed or emitted by the
specimen and the reference is plotted against the
programmed temperature, the area beneath the exothermic or
Page 53
endothermic peaks is proportional to the total enthalpy
change in the specimen. The rate at which the specimen
absorbs or gives off heat is proportional to its specific

heat, as this is the amount of energy needed to change the
specimen temperature by a given amount.
Five specimens of approximately 5mg were cut from the
pellet and the pipe wall. The test conditions were as

follows:
Scan rate: 10 °C per

minute;
Atmosphere: nitrogen gas: 20m1 per minute;

Cooling rate: 40 °C per minute.
The following parameters were recorded for each sample:

(i) onset melting temperature;
(ii) the enthalpy of fusion and;
(iii) the melting temperature attained in the melt
endotherm.
3.1.3.2 OXIDATION INDUCTION TIME (OIT)
The OIT test was undertaken using DSC in compliance with

WAA specification 4-32-04. As outlined in Chapter 2,
sample preparation is of utmost importance. To obtain a

constant sample surface area, strips of approximately 1mm
in thickness were obtained from the surface, middle and
bore of the pipe. The strips were produced on a milling
machine. A hole-puncher of 5mm in diameter was used to
obtain specimens from the milled strips. However, in the
subsequent experiments the strips were annealed at 125°C

for 80 minutes under vacuum and cooled overnight.
Page 54
Samples weighing between 14.5mg to 15.5mg were cut from
the milled strips and heated from 40°C to 200°C at 20 °C
min-1 in a nitrogen atmosphere with a flow, rate of 20 ml
min-'. This was done to prevent any degradation during the

heating up period. At 200°C the DSC was switched over to
an isothermal mode and pure oxygen supplied at a rate of

20 ml min-'. The onset of the main exotherm was measured
by the baseline extrapolation method. OIT tests were also
carried out at 205°C, 215°C and 230°C.
3.2 BUTT FUSION WELDING
3.2.1 EQUIPMENT
All the welding was carried out on a commercial,

automatically controlled powered butt-fusion machine
(Fusion Equipment Ltd model ABF-3). The machine used a
microprocessor controller and a linear displacement
transducer on the frame of the machine to sequence and to
control the welding cycle. In particular, the displacement

in the bead-up stage in which at the start the pipes are
pushed against the hot-plate to ensure even contact was
monitored. The hot plate was ejected automatically from
the machine so that consistent short plate removal times
are achieved.
The hot-plate temperature once set, was checked

periodically during the course of this study via a digital
thermometer with a surface contact probe.
Page 55
3.2.1.1 THERMAL MAPPING OF HEATER PLATE
Thermal mapping of the butt fusion heater plate was
carried out using an Agema Thermopoint 8000 infra-red
camera. These experiments were carried out at 205°C and

230°C. The heater plate was divided into grids as
illustrated in Figure 3.2. The first reading taken 1cm
along the row and 2cm along the column. All the other
successive readings were taken at 5cm intervals, see
Figure 3.2.
5cm
Plate
Figure 3.2 Schematic illustration of the thermal

mapping of the heater plate.
3.2.2 WELDING PROCEDURE
The welding cycle in all cases consisted of bead-up

a
stage followed by the heat soak period and after rapid
removal of the heater plate, the joining or welding phase.
Before welding, the pipes were cut to a length of 25cm on
a power-saw. The pipe sections were wiped with a dry cloth
to remove the dirt and any surface contamination. Figure
3.3 gives a schematic illustration of the welding process.
Page 56
The pipes were clamped in the machine with the
manufacturers pipe marking aligned in order to reduce
mismatch in the diameter, Figure 3.3i. The pipe ends were
then trimmed until a continuous strip of polyethylene was
cut from both pipe ends. The trimmer was then removed and
the pipe ends were inspected to ensure that they were flat
and even, Figure 3.3iv. A check cycle was activated to
allow for inspection of misalignment, Figure 3.3v. The
pipe ends were separated and the hot-plate assembly was

fastened to the frame. The hot-plate was then locked into
position between the ends of the pipes. An initial-welding
pressure (bead up pressure) was used to melt any
asperities so that subsequent heating occurs equally

around the faces. This resulted in the formation of a thin
bead around the circumference where the pipe was in
contact with the heater plate Figures 3.3 vii-viii. The
pressure was released and the heating continued for a
certain time (heat soak time). After heating, the pipes

were separated and the heater plate ejected automatically,
(dwell time) ensuring that the cooling of the pipe faces
was not excessive. The pipes were then quickly joined
together and held under pressure (joining pressure). This
produced a bead of molten material around the joint. The

joint was then allowed to cool under pressure (cooling
time), Figure 3.3 ix. Finally the welded pipes were

numbered and top of the pipes were marked prior to removal
from the clamps. An initial set of welds were produced
with the Rigidex and Eltex pipes using the conditions
shown in Table 3.1. As a result of these trials a more
comprehensive study was undertaken, the welding conditions

for these welds are given in Tables 3.2-3.3. A number of
combined welds were also produced between Rigidex and
Page 57
Eltex. Table 3.4 gives the joining conditions employed in
each case. A limited number of dual pressure welds for
Eltex and Rigidex were also made. Table 3.5 gives the
joining conditions used for these welds. Welds were
allowed to cool for 1000 seconds.
CONDS TEMP BEAD UP BEAD UP SOAK JOINT COOLING
(°C) PRESSURE DISTANCE TIME PRESSURE TIME
(MPa) (m) (s) (MPa) (s)
S220 220 0.15 187.5 180 0.15 600
S230 230 0.15 187.5 180 0.15 600
S230A 230 0.15 12.5 323 0.03 600
S230B 230 0.15 187.5 323 0.15 600
S240 240 0.15 187.5 180 0.15 600
R220 220 0.03 125 140 0.23 600
R230 230 0.03 125 140 0.23 600
R230A 230 0.03 12.5 323 0.03 600
R230B 230 0.03 125 323 0.23 600
R240 240 0.03 125 140 0.23 600
Table 3.1 Summary of welding conditions which were

employed for making the Eltex and Rigidex
welds.
3.3 MICROSTRUCTURE
3.3.1 MICROTOMY
The welded pipes were cut into strips (25mm) on the band
saw. The strips were then cut into 40mm long sections.
These blocks were then ground and polished on 90,120,180
Page 58
(i)
(ii)
LLL
(iii)
(iv)
ýý
(v)
Page 59
(vi)
rf i
(vii)
T (viii)
(ix)
Figure 3.3 The butt-fusion process for joining pipe.
Page 60
and 320 grit silicon carbide paper to ensure that the
'samples were flat and to remove the milling marks.
The samples were then held in the vice mounted on the
microtome and thin (2 to 7µm) and semi-thick (7 to 15µm)
sections were cut on a Reichert-Jung 2050 Supercut
microtome. The microtome was fitted with a motor to ensure
SOLVAY TEMP BEAD UP BEAD UP SOAK JOINT COOLING
CONDS (°C) PRESSURE DISTANCE TIME PRESSURE TIME
(MPa) (m) (s) (MPa) (s)
SS1 230 0.15 187.5 180 0.15 600
SS2 230 0.15 187.5 180 0.02 600
SS3 230 0.15 187.5 180 0.21 600
SS4 230 0.15 187.5 180 0.31 600
SS5 230 0.15 187.5 180 0,42 600
SS6 230 0.15 187.5 180 0.63 600
SS7 230 0.02 12.5 180 0.02 600
SS8 230 0.02 12.5 360 0.02 600
SS9* 230 0.15 187.5 180 0.15 600
SS10 230 0.15 187.5 360 0.15 600

SS11 230 0.15 187.5 420 0.15 600
SS12 230 0.02 12.5 420 0.02 600
SS13 230 0.15 187.5 180 0.03 600
SS14 230 0.15 187.5 180 0.09 600
* crash cool

used for making the welds with Eltex pipes.
Page 61
RIGIDEX TEMP BEAD UP BEAD UP SOAK JOINT COOLING
CONDS (°E) PRESSURE DISTANCE TIME PRESSURE TIME
(MPa) (pm) (s) (MPa) (s)
RR1 230 0.15 187.5 180 0.15 600
RR2 230 0.15 187.5 180 0.02 600
RR3 230 0.15 187.5 180 0.21 600
RR4 230 0.15 187.5 180 0.31 600
RR5 230 0.15 187.5 180 0.42 600
RR6 230 0.15 187.5 180 0.63 600

RR7 230 0.02 12.5 180 0.02 600
RR8 230 0.02 12.5 360 0.02 600

RR10 230 0.15 187.5 360 0.15 600
RR12 230 0.02 12.5 420 0.02 600
RR13 230 0.15 187.5 180 0.03 600
RR14 230 0.15 187.5 180 0.09 600

used for making the welds with Rigidex
pipes.
CONDS TEMP BEAD UP BEAD UP SOAK JOINT COOLING
(°C) PRESSURE DISTANCE TIME PRESSURE TIME
(MPa) () (s) (MPa) (s)
SR1 230 0.15 187.5 180 0.15 600
SR2 230 0.15 187.5 180 0.02 600
SR3 230 0.15 187.5 180 0.21 600

SR4 230 0.15 187.5 180 0.31 600

used for making mixed welds between Rigidex
and Eltex.
Page 62
CONDS TEMP BEAD UP SOAK PRIMARY SECONDARY BEAD
(°C) PRESSURE TIME JOINT JOINT ROLL
(MPa) (s) PRESSUE PRESSURE OVER
(MPa) (MPa) TIME
(s)
SBRO10 230 0.15 225 0.15 0.02 10
SBR05 230 0.15 225 0.15 0.02 5
SHS450 230 0.15 450 0.15 0.02 10
RBRO10 230 0.15 225 0.15 0.02 10
RBR05 230 0.15 225 0.15 0.02 5
RHS450 230 0.02 450 0.15 0.02 10

used for making the dual pressure welds for
Eltex and Rigidex.
a uniform cutting action. The vertical guiding system
provided a much more rigid arrangement for cutting hard
and large samples without backlash from the sample arm.
As the weld was brought into contact with the cutting edge
of the blade, a thin section of the weld was detached from
the sample. The thickness of section depended on the
overlap between the sample and the cutting blade on the
microtome. Friction between the polymer and the cutting

edge did not present any serious problem. However, with
thin sections, the friction was relatively high in
relation to the mechanical strength of the sections and

hence some form of lubrication was necessary. Cedarwood
oil (supplied by BDH) was used as the lubricant. The
Page 63
Chapter 3 Experimental Techniques
lubricant was applied on the knife and the sample with the
aid of a fine brush. Whilst cutting, the sample was gently

brushed to prevent curling. The microtomed section was
mounted on the glass slide and entrapped air was expelled

by gently pressing the brush onto the section. After
pressing the air out, a drop of oil and a cover slip was
placed on top of the microtomed section. Once again the
remaining air bubbles were pressed out.
3.3.2 POLISHING OF BLADES
Two types of blade were used, namely disposable blades and

D-profiled steel blades.
3.3.2.1 DISPOSABLE BLADES
In order to obtain a consistent cutting edge, the
stainless steel blades (LKB Histoline) were polished using

standard metallographic techniques using a custom built
jig as shown in Figure 3.4. The polishing was achieved
using a 6µm diamond paste.
3.3.2.2 D-PROFILED STEEL KNIVES
The stainless steel blade (Reichert-Jung) used in this
study had effectively four surfaces on each side. The back
of the blade had an enclosed angle which is less than that
of the cutting angle. The secondary edge had an angle of

30 0 whereas the relief edge which was adjacent to the
cutting edge has an angle of 38°. Finally the cutting edge

had an angle of 40° as illustrated in Figure 3.5. The
following blade pre-polishing sequence was carried out
using 8mm diamond paste with a facet angle of 38° with
maximum load:
Page 64
(i) 30 minutes polishing with a turnover time of 25
seconds;
(ii) 20 minutes polishing with a turnover time of 10
seconds;
(iii) 10 minutes polishing with a turnover time of 5
seconds.
The blade was then thoroughly cleaned and a second
polishing wheel with a 3µm diamond paste was loaded on the
polishing machine. The blade was then returned to the
machine and the above procedure repeated. After completing
this polishing sequence, ' the blade was'polished at a facet
angle of 400 with 3µm paste for a period of 5 minutes at
minimum load, to give the blade the final optical edge.

The blade was periodically checked under an optical
microscope to evaluate the intergrity of the cutting edge.
blade stop pins
fastening dowel holes

screws
2 0mm
18°
zýw
Figure 3.4 Schematic illustration of the jig used for

polishing the blades.
Page 65
3.3.3 TRANSMITTED LIGHT MICROSCOPY
Microtomed sections of the welds were examined in a

Reichart-Jung transmitted light optical microscope. The
microscope was equipped with a rotating stage allowing the
specimen to be moved relative to the polarisation plane.

By observing the angle at which extinction-of the sample
occurred it was possible to get an indication of the
direction of the molecular orientation within the weld
zone.
Cutting edge
imary edge
Relief angle Secondary edge
Figure 3.5 Schematic illustration of the knife faces.
3.4 ANALYSIS OF WELDED SPECIMENS
3.4.1 WELD BEAD MEASUREMENT
Once each weld was produced the dimensions of the inner
and outer weld bead were measured with a digital vernier

calliper. These dimensions were taken for all strips (18)
Page 66
which were cut from the welded pipe. An average of 5
measurements were taken for each of the strips. The width

is the full width of the rolled over flash or weld bead,
see Figure 3.6.
The limitations of the manual measurements mentioned above
were recognised. In order to obtain an accurate measure of

the bead dimensions, the beads were viewed using a video
camera linked to a Cambridge Quantimet 520 image analysis
system, this enabled the bead dimensions to be identified,
the bead width and bead area were measured automatically

(see Figure 3.6).
ow
Surf i
BorE
Iw
0,,,: outer bead width OM: MAZ length outer

I,: inner bead width MM: MAZ length centre
OA: outer bead area IM: MAZ length inner
IA: inner bead area
Figure 3.6 Definition of the weld bead dimensions.
Page 67
3.4.2 MAZ - INTERNAL WELD DIMENSIONS
Microtomed sections of the welds were viewed between cross
polars in a Reichart-Jung transmitted light optical
microscope. The sections were photographed. The outlines
of the welds were traced onto graph paper. This was then
enlarged, cut out and weighed. A calibration was carried
out using a graph paper of known area and weight. In
addition to this, a selection of photographs were viewed
using a video camera linked to a Cambridge Quantimet 520

image analysis system. This enabled the outline of the
weld to be identified and a number of other dimensions
shown schematically in Figure 3.6 were measured. These
measurements were correlated to the width of the melt

zones in the outer, inner and centre sections of the pipe.
The angle made by the outer (Os) and inner beads (91) to
the pipe outer and inner surfaces as well as the area of

the MAZ were also measured.
3.5 TENSILE TESTING
3.5.1 MICROTENSILE TESTING-SAMPLE PREPARATION
The dimensions of the block samples used were 40 x 17.5 x

25mm (length x width x thickness). Each block was mounted
in a vice on the Reichert-Jung 2050 Supercut microtomer to
give a flat face. Custom-made grids with a line thickness
of 200µm were carefully transferred on to the previously
microtomed sample block by rubbing the back of HB pencil
on the grids. A thin film (401im) with the grid was then
obtained via microtomy. The microtomed sections were

placed between glass plates for a period of thirty minutes
prior to tensile testing in order to prevent them from
curling.
Page 68
3.5.2 MACRO-TENSILE-SPECIMEN PREPARATION
Seven strips were taken from each of the weld for tensile
testing. Initial set of welds were produced with both the
Rigidex and Eltex using the conditions shown in Table 3.1.
The weld bead from each of the welds was removed by. a
mechanical debeader. The samples were cut (25mm) from each
weld parallel to pipe axis on a band saw. The samples were
faced parallel on a milling machine to give a width of
17.5mm (the thickness of the pipe wall). After producing
the parallel strips, profiles were introduced into the
sample by using a milling cutter. The final width of the
gauge piece was approximately 10mm, see Figure 3.7.
17.5mm
Figure 3.7 Schematic illustration of the tensile test

profile used in this study.
Page 69
lt
15
14
11 10
Figure 3.8 Location of the strips numbered 1 to 18.
As a result of the preliminary trials, a more
comprehensive study was undertaken using the welding
conditions given in Tables 3.2-3.4. A special jig was

designed and built to enable rectangular strips to be
obtained from around the circumference of the pipe. Up to
18 strips were cut around the circumference of the pipe.

Position 1 corresponds to the top position when the weld
was made and position 10 corresponds to the bottom of the
pipe, the lowest position in the weld. This is
schematically shown in Figure 3.8. The dimensions of the
strips were 17.5mm in thickness, 17.5mm in width and a

length of 200mm. The cutting arrangement with the jig is
presented in Figure 3.9 It is worth noting that this
cutting arrangement was faster than the above method for
producing samples. The method also gave specimens with a
consistent dimension. However, it was not possible to cut

the Rigidex pipe welds into strips by this method. The
Page 70
Chapter 3 Experimental Techniques
Rigidex welds were produced using the first method. The
final width of the test piece was approximately 10mm. A
milling cutter of diameter of 20mm was used to machine the
beads off. In this way, samples without the beads were
also tested.
Figure 3.9 The double cutter arrangement which was

used for cutting the pipes into strips.
3.5.3 MICRO-TENSILE TESTING PROCEDURE
The tensile testing was carried out at ambient temperature
on an Instron 1026 machine using a crosshead speeds of 5,
50 and 500mm/min. A 50N load cell was used along a

with
LVDT which offered an extension range of 0 to 100mm.
Page 71
Due care was taken to ensure a consistent gripping
procedure for all the test specimens. The gauge length was
maintained at 20mm with the weldline being located at the
centre. The weldline was used to ensure vertical alignment
of the test specimen. All the samples were mounted with

the outer surface of the pipe being located on the right
hand side. This was done by marking a black spot on the
top right hand corner. This allowed the identification of
the top and bottom of the sample once failure had taken
place. Crossed polarised light was used along with a Sony

Hi-8 handycam video camera to record the test.
3.5.4 MACRO-TENSILE TESTING PROCEDURE
Tensile testing was carried out in a controlled room (20

°C) on an Instron 1056 testing machine head
using cross
speeds of 10 mm min-1 and 50 mm min-'. Stress at peak and
elongation at failure were recorded. Tests were also
carried out on the unwelded samples with a gauge width of
5 and 15 mm.
Page 72
MATERIAL CHARACTERISATION
4.1 INTRODUCTION
This chapter is concerned with the characterisation of the
Eltex Tub 124 and Rigidex 002-50 pipe materials. The
results from the density, GPC, DSC and OIT studies are
presented. The results are discussed at each stage to put

them in context with the rest of the work and provide
clarity in the following chapters.
4.2 DENSITY
The densities of the specimens taken from the locations in
the MAZ and 20mm away from it through the pipe wall
thickness are shown in Table 4.1. The number of specimens
in each case was three. The size of the sample from the
MAZ - mid position was small since the MAZ length is
relatively small. The mean values have been tabulated and
since the variation in these values was very small,

standard deviations have been omitted from the table.
However, the observed minimum and maximum variations are
stated. The following points emerge from the density
measurements:
Page 73
(i) As expected the density of the samples taken 20mm away
from the butt weld increases from the surface through the
bore of the pipe wall. This is because the external
surface of the pipe was cooled rapidly and the internal
surface was cooled slowly during the extrusion of the
pipe.
(ii) The density at the butt weld was found to be constant
through the pipe wall thickness but greater compared to
the specimens taken 20mm away from it.
Density, kgm-
Sample Standard Weld 20mm away

Location MAZ
Eltex Tub 124
Surface 955.3 943.7
Mid 955.5 953.2
Bore 955.8 955.3
Rigidex 002-50
Surface 941.5 934.4
Mid 941.2 938.8

Bore 941.3 940.2
Minimum variations = 0.1; Maximum variations = 0.4
Table 4.1 Density of samples taken from Eltex Tub

124 and Rigidex 002-50 butt welded pipes
under standard welding conditions (SSl and
RR1) from the MAZ and 20mm away.
4.3 MOLECULAR WEIGHT DETERMINATION
The molecular mass distribution were determined using GPC.
Page 74
Figures 4.1-4.2 illustrate the molecular weight data for
the two resins and the pipes. From Figure 4.1 and 4.2, the
difference in the molecular weight for the two resin
systems are readily apparent. The results show that there
is no significant change in the number average weight or

the weight average molecular weight for Rigidex 002-50.
However, for Eltex Tub 124 the results show differences
for the pipe. The surface exhibits the lowest molecular

weight value. This may be due to degradation as a result
of storage.
4.4 THERMAL ANALYSIS
4.4.1 DIFFERENTIAL SCANNING CALORIMETRY
The enthalpy of melting was calculated for Eltex and

Rigidex virgin pellets and pipes. The test conditions were
as follows:
(i) Scan rate 10°C per minute;
(ii) Atmosphere = nitrogen gas: 20m1 per minute;

(iii) Cooling rate 40°C per minute.
The following parameters were recorded for each sample:

(i) onset melting temperature; (ii) the enthalpy of fusion
and (iii) the melting temperature attained in the melt

endotherm.
The percentage crystallinity was calculated from the

following formula:
Hf
% Crystallinity =A 100 (4.1)
ex
Page 75
IC
® Nd
® ni ý1ý
r' 228 cw
$u
u 2XIO5
lxl0s
o' FZMN SUWACE MID BE

o
Figure 4.1 Number and weight average molecular weight

data for the Eltex Tub 124 resin and pipe.
Laos
® mw
2x105
pos
RESIN SURFAUE MID r-
Figure 4.2 Number and weight average molecular weight

data for the Rigidex 002-50 resin and pipe.
Page 76
where MHf is the enthalpy of fusion of the sample as
obtained from the DSC thermogram; OHf0 is the
corresponding value for the enthalpy of fusion for a 100%
crystalline linear polyethylene. The value of OHfe was
taken as 293Jg-1 [Gedde and Jansson, 1980].
The following series of experiments were carried out to
establish the reproducibility of the DSC thermograms. With
reference to Table 4.2, Specimens 1,2 and 3 correspond to
Eltex specimens with increasing mass. It is readily
apparent that the sample mass can have an influence on the
parameters measured. However, an important uncertainty in
these tests is the homogeneity of the specimen, both in
terms of degree of crystallinity and the concentration of

additives present. Specimens 4,5,6 and 7 again of Eltex
correspond to successive scans of specimen 4. In other
words, specimen 4 was first scanned at 10K per minute from
320K to 450K and then cooled (naturally) back to 320K.
This was repeated three times without removing the
specimen. A number of points can be concluded with
reference to Table 4.2:
(i) The DSC technique can be used with a high degree of
confidence provided the homogenity of the test specimen
can be assured.
(ii) The enthalpy of fusion AHf can be altered by the
cooling regime imposed on the specimen from the melt.
Tables 4.3-4.4 illustrate the measured properties of the
Eltex and Rigidex resin and pipes. The terms surface, mid
and bore refer to the relative positions in the pipe wall
Page 77
thickness direction. The general observed trends in Tables
4.3-4.4 are as follows:
(i) The Eltex virgin pellets exhibit slightly higher
crystallinity values as compared to the Rigidex pellets.

This is reasonable as the two polymer types are different:
Rigidex is a medium density material and Eltex is a high

density material.
(ii) The blue, yellow and black Eltex polymer grades show
similar crystallinity values. However, the Rigidex black
polymer shows a higher crystallinity value. On inspecting
the DSC thermogram, the presence of a secondary endotherm

is observed in the range of 356K-376K with a maximum at
370K. The presence of the secondary endotherm may be
attributed to the presence of a dual or multi-component

polymer system or ageing of the polymer below the melting
point.
On comparing the data for the virgin Eltex resin

(granulated pellets) and the data for the 180mm Eltex
pipe, some key difference are observed. The melt onset is
slightly higher for the extruded pipes. This is reasonable

as there is usually a pronounced dependence on the
previous thermal histories to which the polymer was
subjected. On comparing the data for the 180mm pipe, it is
clear that a gradient in property exists within the pipe

wall. The residual stresses resulting from the temperature
gradient which exists while the extrudate pipe solidifies
are complex. The outer wall of the pipe solidifies first,
on the inside the material is still in a molten state. As
the material below the surface starts to cool, it is
Page 78
Sample Weight Onset Enthalpy Crystallinity Melting
(mg) Temp (Jg-1) (%) Temp
(K) (K)
1 0.91 394.1 174.9 59.7 400.1
2 4.81 394.2 176.3 60.2 401.2
3 8.81 394.2 179.7 61.3 402.0
4 9.15 394.1 178.5 60.9 403.3
5. specimen 4 393.1 188.1 64.2 405.2
rescanned
6. specimen 5 393.3 187.7 64.1 405.1
rescanned
7. specimen 6 393.2 187.7 64.1 405.4
rescanned
8. specimen 7 393.3 187.5 64.0 405.2
rescanned
'4
Table 4.2 Summary of the results for the virgin blue

Eltex pellets. Specimen 1-3 illustrate the
influence of sample mass and specimens 4-7
indicate the reproducibility for a single
specimen where the thermal history was
identical.
restricted from contracting by the frozen surface layers.

This results in compressive stresses in the outer wall and
tensile stresses on the inside due to constrained
elongation. A similar trend is observed for the Rigidex
pipe in terms of melting characteristics as a function of

pipe wall thickness, ie., the observed values increase in
going from the outer surface of the pipe to the bore.
However, in Table 4.4 the on-set temperature of the outer
surface is slightly higher compared to the mid and bore of

the pipe. This is due to the presence of the secondary
Page 79
Sample Onset Enthalpy Crystallinity Melting
Temp (Jg-1) (%) Temp
(K) (K)
Virgin Eltex Resin
Blue 394.3 173.7 59.3 (0.6) 401.4
Aged blue 600 hrs 398.1 198.5 67.7 (0.8) 404.2
at 92°C vacnsn
No secondary endotherm was observed

Aged blue 700 hrs 409.6 213.1 72.7 (2.6) 414.0
at 125°C in vac un
Pronounced secondary endotherm between 376K-402K
Aged blue 700 hrs 403.7 183.1 62.5 (1.4) 410.6

at 125°C in air
Black 394.4 172.8 56.0 (1.4) 401.7
Orange 394.5 174.8 59.7 (0.9) 402.1
180mm OD SDR 11 Eltex Pipe
Surface 394.1 159.7 54.5 (0.9) 402.1
Mid 397.6 189.4 64.6 (0.. 7) 407.7
Bore 398.1 188.7 64.4 (0.5) 409.6
Surface 394.7 157.4 53.7 (1.4) 401.3
Mid 395.8 181.4 61.9 (0.5) 405.2
Bore 396.8 185.4 63.3 (0.8) 408.2
Surface 394.7 159.1 54.3 (1.1) 400.2
Mid 396.0 175.2 59.8 (0.6) 404.6
Bore 399.4 181.5 62.0 (0.8) 407.6
crystallinity: minimum varation u. 5; maximum variation 2.6.

melting temp.: minimum varation 0.2; maximum variation 2.9.
Table 4.3 Summary of the DSC results for the virgin

Eltex resin and three different diameter
Eltex pipes as a function of pipe wall
thickness.
Page 80
endotherm. This secondary endotherm was not observed for
specimens taken from the mid and bore of the pipe wall.
Table 4.3 illustrates the results from ageing a Eltex blue
virgin pellet. The pellet which was aged at 90°C for 600
hours under vacuum did not show the presence of any
secondary endotherm. However, it is clear from the table,
that ageing this material at 92°C results in a significant
change in the percentage crystallinity. Increasing the
ageing temperature to 125°C results in a further increase

in percentage crystallinity in the specimen along with
increases in other measured properties. However, after
ageing the specimen at 125°C in air, the secondary
endotherm was less pronounced compared with the vacuum-
aged specimens. The location of this secondary endotherm
was between 369K-396K. Rigidex blue pellet aged at 92°C in
a vacuum oven also showed the presence of the secondary

endotherm in the temperature range 362K-378K.
The increase in the crystallinity at 92°C for both Eltex
and Rigidex pellets seems to suggest a tendency to form a
more ordered and stable crystal structure. It is probable

that the secondary endotherm is due to increased order in
the amorphous fraction of the polyethylene as a

consequence of ageing the polymer below its melting point.
However, this may never be fully realised since at 92°C
the thermal motion of the polymer chains are insufficient
to bring about complete order. The secondary endotherm

probably has its origins in the limited motion of the
chains in the disordered region attempting to rearrange

into something approaching three-dimensional order.
However, in the case of Eltex pellets at 125°C, the
Page 81
melting temperature has been raised suggesting that a
significant ordered structure was formed. The values in
parenthesis correspond to the standard deviations.
Sample Onset Enthalpy Crystallinity Melting

Temp (Jgý'') (%) Temp
(K) (K)
Virgin Rigidex Resin
Blue 392.2 136.8 46.7 (0.7) 399.9
Aged Blue 600 hrs. 391.0 165.0 56.3 (2.2) 401.1
at 92°C vacznan
Secondary endotherm between 378K-387K
Black 392.4 145.3 49.6 (1.3) 399.5
Yellow 391.0 132.7 45.3 (0.7) 399.9
180mm OD SDR 11 Rigidex Pipe
Surface 391.9 136.5 46.6 (1.8) 398.1
Mid 389.9 141.1 48.2 (1.1) 400.0
Bore 391.0 152.5 52.0 (0.6) 401.4
crystallinity: minimum varacion U. b; maximum variation 2.2.

melting temp.: minimum varation 0.2; maximum variation 1.2.
Table 4.4 Summary of the DSC results for the virgin

Rigidex resin and pipe as a function of
pipe wall thickness.
4.4.2 OXIDATION INDUCTION TIME (OIT)
The oxidation induction test (OIT) is an industry-
specified test to assess the thereto-oxidative stability of

the polyethylene. An OIT value of 20 minutes at 200°C is
required. This is important for the following reasons:
Page 82
(i) Degradation of the polymer during processing and also
possibly during storage if volatile additives are used
would present serious problems but would be detected by
the OIT 20 minutes at 200°C requirement. The pipe may be
stored for up to one year outside before underground use.
Allen et al. [Allen et al., 1988] found that the OIT value
of the surface of a pipe decreased from 65 minutes to 1.8
minutes for a9 month exposure in the UK environment.
(ii) In the UK up to 25% regrind is allowed for pipes up
to 250mm OD. The antioxidant concentration is not
increased in the pipes with the use of regrind. This can

have undesirable effects in terms of thermal and
photostability of the pipe as well as taste problems

[Allen at al., 19881.
(iii) The current welding temperature used in the UK is

230°C ± 10°C. Thus the pipe can experience a temperature
of 240°C during the welding operation. The thermal
stability of pipe is particularly important, especially
with regard to further thermal treatment, e. g. at the weld
joints.
A summary of the OIT results for the Eltex and Rigidex
virgin resin and pipe samples as a function of test
temperature are presented in Tables 4.5-4.6. The OIT for
the different virgin pellets of Eltex and Rigidex at 200°C
show a very large spread in the results. On increasing the
temperature to 230°C this spread is dramatically reduced.
The information available suggests that the resins are
essentially the same, the only difference being the
pigment. Therefore, it is likely that the pigments may
Page 83
indeed influence the thermo-oxidative response of these
materials. It is readily apparent that the OIT values fall
from 177 minutes at 200°C to 26 minutes on comparing the
virgin resin and the extruded pipe surface. For the
Rigidex, the value falls from 89 minutes for the virgin

pellet to 77 minutes for the surface of the pipe at 200°C.
Furthermore, there are large differences in the surface
and bore OIT values. On comparing the values for the
Rigidex and Eltex surface at 200°C, a reduction in the OIT
of 33% was seen for the Eltex. The Eltex resin in the form
of the extruded pipe has a lower relative OIT value. In
comparison, the Rigidex pipe samples gave results typical
of that reported by other researchers.
As a consequence of the very large spread and low values
of OIT in the Eltex, an investigation was carried out to
gain an insight into the problem. It is worth pointing out

that the main concern was the welding temperature of 230°C
±10°C which is specified by the water and gas industries.
At these welding temperatures it is possible that the weld

interface is devoid of antioxidant. This in turn can lead
to serious loss of mechanical properties and embrittlement

of the weld zone.
Both the Eltex and Rigidex samples were prepared in a
similar manner. The virgin pellets were wrapped in tissue
paper and then cut using a sharp blade. For the pipe
samples, a stamping-die was used to produce 4mm diameter
disks for the OIT experiments. The OIT tests were carried
out in accordance to WIS 4-32-03 [Water Standard]. With
reference to Table 4.6, annealing the sample prior to
carrying out the OIT experiment was considered. This was
Page 84
OIT (minutes)
Sample 20C C 205"C 215 C 230"C
ELTEX RESIN
Blue 177 (4.2) 97 (3.5) 39 (2.5) 10 (1.1)
Black 70 (3.2) 44 (3.2) 17 (1.8) 9 (1.7)
Orange 218 (2.3) 135 (4.1) 47 (3.5) ii (2.1)
RIGIDEX RESIN
Blue 89 (2.7) 56 (3.1) 22 (2.5) 4 (2.2)
Black 57 (3.4) 36 (1.8) 19 (2.5) 5 (2.4)
Yellow 81 (1.8) 46 (2.5) 22 (1.5) 6 (1.8)
Standard deviation in bracxets.
Table 4.5 Summary of OIT results for virgin Eltex

and Rigidex pellets.
OIT (minutes)
Sample 200 C 205 C 215'C 2300C
ELTEX Tub 124 180mm OD SDR11 pipe
Surface 26 (22.9) 49 (28.7) 10 (12.9) 1 (0.3)
Annealed - 31 (40.2) - 7 (2.4)
surface "ýo k/c

Mid 116 (8.6) 50'Y(6.2) 25 (5.6) 7 (2.6)
Bore 39 (3.9) 27 (4.8) 7 (3.8) 4 (2.1)
RIGIDEX 002-50 180mm OD SDR1 pipe
Surface 77 (14.6) 38 (4.3) 21 (3.2) 2 (1.4)
Mid 77 (12.2) 43 (3.9) 15 (2.2) 2 (0.7)
Bore 72 (8.1) 35 (1.1) 9 (1.8) 2 (1.1)
Standard deviation in brackets.
Table 4.6 Summary of OIT results for Eltex Tub 124

and Rigidex 002-50 180mm OD SDR 11 pipe.
Page 85
because on inspecting the OIT specimen after each test, it
was observed that the extent of "flow" in the virgin resin
was significantly greater than that observed for the
extruded pipes. Furthermore, the extruded pipes exhibited
extensive curling, i. e. the sample had moved in the sample

pan during the test. This is likely to be because the
specimen stress relieves as the temperature is increased
from ambient to near melting. The residual stress in the
specimen is a direct consequence of the manufacturing

process. The extent of stress-relieving in the specimen
during the OIT experiment can be influenced by the
following factors: (i) specimen production method; (ii)
specimen dimensions; (iii) initial contact area between
the specimen and the sample holder; (iv) the temperature
gradient across the sample thickness and (v) contamination

of the specimen and/or the sample holder.
The obvious way to stop the distortion of the specimen

during the OIT experiment was to stress-relieve the
samples prior to testing. A series of experiments were
carried out where the previously-mentioned strips obtained

from the extruded pipe were annealed at 125°C for 80
minutes under vacuum and then left to cool to room

temperature.
The stress relieving solved the sample distortion
phenomenon but another problem was created. With reference

to Table 4.6, considerable scatter was obtained for the
various OIT test temperatures investigated. For the Eltex
surface, the 215°C series for the un-annealed pipes gave

OIT values between 2.0-24.8 minutes. The annealed OIT pipe
surface in the 205°C, exhibited values between 2.5-86
Page 86
minutes. Obviously, this scatter was unacceptable both
from an experimental view point and for the purpose of
quality control.
4.4.2.1 SAMPLE MASS
A series of experiments were carried out on the virgin

blue Eltex pellets to establish (i) the best way to
analyse the results and the investigate the influence of
sample mass. Figure 4.3 illustrates a typical DSC OIT
experiment. Point (A) represents the change-over from the
nitrogen gas to pure oxygen. Point (B) corresponds to OIT
(Ti) in Figure 4.4 and is representative of the point
where the instrument is starting to detect a measurable
exotherm. Slope (C) represents the primary exothermic
reaction where it is assumed the antioxidants- are
consumed. Points (C) and (D) correspond to OIT (T2) and

(T3) in Figure 4.4.
With reference to F: Lgure 4.4, OIT (Ti), (T2) and (T3)
correspond to three ways of interpreting the data in
Figure 4.3. The OIT values, regardless of the mode of

analysis indicate a very small dependence between the
sample mass and the resultant OIT value. However, the
surface area in these experiments was not controlled.
The effect of sample thickness on the OIT for the Eltex
pellets were investigated at 205°C. Three samples of

similar size but varying thickness were -Used. The results
from this investigation showed that the thickest sample
(1.43mm, 18.76mg) exhibited an OIT of 120 minutes. This
was followed by 111 minutes for the sample thickness of
Page 87
IL U
68 76 85
0 8CD
U
saz
N
F-
H
1L12 au 3.1 £&2 S1 B7 6302 Jß02 BJ T7 91Pa ii; c MM
TIME (MIN) DSC
Figure 4.3 Illustration of a typical DSC OIT

experiment.
150
V OIT (T3)
O OIT (T2)
0 OIT (T1)
125
100
vvv
00
0 00
75
000
ö 0
50
25
n
0 369 12 15 18
Sample Weight (mg)
Figure 4.4 Influence of sample mass on OIT for blue

Eltex virgin pellets.
Page 88
0.78mm and a mass of 9.37mg. An OIT of 85 minutes was
obtained for the thinnest sample having a thickness of
0.30mm and a mass of 2.24mg.
Figure 4.5 represents the OIT data from the knife-cut
samples corresponding to the outer surface of the Eltex
Tub 124 pipe as"a function of sample mass at 205°C. Figure
4.6 illustrates the dramatic effect on the measured OIT
values of Eltex prepared by machining and by using a sharp

blade. A thermocouple was used to measure the temperature
at the cutting tip of the milling cutter. Temperatures in
the region of 75°C were recorded. Moore et al. [Moore et
al., 1989] have found that the outer surface of the pipe
was very sensitive to the method of sample preparation.
Drilled samples gave the lowest and most variable results;
cut samples gave the highest and most consistent results.

A limited number of experiments were carried out on the
Rigidex pipes to investigate a similar detrimental effect
of machining on the OIT. The results from this study are
presented in Table 4.7.
OIT at 205 C
Sample (minutes)
Rigidex knife-cut 38 (3.2)
Rigidex milled 44 (7.1)
Eltex knife-cut 56 (4.2)
Eltex milled 49 (28.7)
Table 4.7 Summary of the OIT results for outer

surface of Rigidex and Eltex 180mm OD SDR
11 pipe.
Page 89
100
Eltex Tub 124 - )urige tut
80
00
0
0
60
00
40
20
A
V
0 10
,2468
Sanple Weight (mg)
Figure 4.5 Illustration of the observed relationship

between the OIT values for the outer surface
of Eltex pipe as a function of sample mass.
80
Knife Cit
® Nladzined
60
40
20
0 01234567
Sauple Number
Figure 4.6 Illustration on the effect on the measured

OIT values for test specimens prepared by
machining and by using a sharp blade.
Page 90
4.4.2.2 VOLATILITY OF ANTIOXIDANTS
Repeat experiments were carried out using a fresh supply
of Eltex virgin pellets directly from the 25kg polymer
container after five months storage. The reason for
carrying out the above mentioned test was to check that
the Eltex specimens were not being depleted of antioxidant
as a consequence of storage in the laboratory. Figure 4.7

illustrates the observed variation in the OIT values for
the Eltex pellets using "old" and a set of "fresh"
pellets. It was assumed that the stabliser was dispersed
uniformly in the pipe material and the oxidation induction
results were considered to be comparative and in no way
absolute. However, the true significant of the above
phenomenon may simply be due to material inhomogenity.
150
O New Sanples
O Old Sanples
O
13
O
13
100
41 d
O
00
00
50
A
V
0369 12 15 18
Sapt1e Weight (mg)
Figure 4.7 Observed variation in the OIT values for

the virgin pellets using "old" specimens
and a set of "fresh" pellets.
Page 91
Table 4.8 illustrates the OIT at 205°C from the ageing of
Eltex and Rigidex pellets at 92°C. The table shows a
reduction in the polymer resistance to thereto-oxidation
for the Eltex pellets aged for 650 hours. However, when
the ageing time was increased to 1500 hours, the OIT did
not show any reduction. For the Rigidex system, there was
no reduction in the OIT even after 1500 hours.
OIT at 205 C
Sample (minutes)
Aged for 650 hours at 92 C in vacuum
Eltex pellet 70 (8.21)
Rigidex Pellet 54 (3.80)
Aged for 1500 hours at 922C in vacuum
Eltex pellet 91 (6.43)
Rigidex Pellet 55 (4.21)
Table 4.8 Summary of the OIT results for 'the aged

Eltex and Rigidex pellets.
4.4.2.3 ELECTRON SPIN RESONANCE (ESR)
ESR studies were carried out on the machined, knife-cut
and machined-annealed samples for the outer surfaces of

Eltex Tub 124 and Rigidex 002-50 pipes. The samples were
freshly machined to a thickness of 1mm and were
subsequently cut into thin strips approximately 1.5-2mm in
width and 5mm in length. The test was carried out at room
temperature at x-band frequency using a Varian E-3 ESR
spectrometer.
Page 92
The general survey scan ESR spectra obtained for the Eltex
samples were similar in shape and form. The Rigidex
material did not give a trace. The electron free spin
factor (g) was calculated for the three Eltex samples. A
value of 1.99 was obtained for the three samples. The
presence of a an unpaired electron which may arise from
the additives present could catalyse the oxidation
reaction and hence may explain why the Eltex material
gives OIT values which are inconsistent.
4.4.2.4 X-RAY PHOTOELECTRON SPECTROSCOPY (XPS)
The previously prepared samples for ESR were cut to the
required size (10xl5mm) and mounted on the end of the
sample probe on a movable tip using double sided tape. The

investigations were made in a vacuum of 10-9 torr with a
Fisons Surface Science ESCALAB 210 XPS spectrometer.
The general survey scan XPS spectra obtained for the Eltex
samples were somewhat similar in shape. However, the
virgin pellet spectrum showed additional peaks related to
sulphur and nitrogen which were only present in minute
quantities. These elements could be from the additives

that are present in the polymer to prevent oxidation and
to aid processing as well as contaminants such as the
polymerisation catalyst, greases and dust. These additives
are intended to be well dispersed throughout the bulk but
in certain circumstances, surface segregation may take
place. The spectra from knife-cut and pellet samples of

Rigidex were also very similar in form. The spectra from
machined Rigidex samples does not exhibit a silicon peak.
Page 93
Table 4.9 lists the atomic% of elements present in the
Eltex and Rigidex samples.
Atomic
Sample C 0 Si S N
Eltex machined 87 10 3 - -
Eltex machined/annealed 86 11 3 - -
Eltex Knife-cut 87 10 3 - -
Eltex pellet 82 13 1 1 3
Rigidex machined 94 6 - - -
Rigidex knife-cut 91 6 3 - -
Rigidex pellet 90 8 2 0.27 -
Table 4.9 At% of elements present in the Eltex and

Rigidex samples.
It may be concluded that the information obtained from the

XPS investigations does not aid in the interpretation of
the DSC data. The information depth of XPS is around 5nm
where the first few atomic layer contribute the most to
the output signal, thereafter these contributions are
reduced.
4.4.2.5 SURFACE DAMAGE
Figure 4.8 is a low magnification view of the machined

Eltex 180mm OD SDR 11 pipe outer surface. The observed
features can be described as series of valleys and hills.
The valleys correspond to the zones where the cutter
scoops out the material and the hills correspond to the
Page 94
Chapter 4 Material Characterisation
Figure 4.8 Low magnification view of the outer

surface of the Eltex pipe.
Figure 4.9 High magnification view of the yielded

zones in the machined surface.
Page 95
Figure 4.10 High magnification view of the outer surface

of the Eltex pipe showing the crazes.
Figure 4.11 High magnification view of the machined

Eltex surface.
Page 96
material deformation at the edges of the cutter. The hills
are areas where the polymer has undergone gross yielding.
Figure 4.9 is a high magnification view of the this hilly
or yielded zone. Close inspection of this zone indicated
evidence for extensive crazing at the base of the hills
and the valleys. A magnified view of these regions are

illustrated in Figure 4.10. Clearer images of the crazes
observed in the machined Eltex pipes are presented in
Figures 4.11 and 4.12. With reference to Figures 4.11 and

4.12, the occurrence of craze zones in the machined
surfaces is clearly visible. The crazes form at 900 to the
machining direction. The inspection of these crazed
surfaces seems to indicate a high concentration of crazes

in the valleys. Within the valleys, the highest
concentration of crazes seems to occur in the vicinity of

the areas which exhibit extensive yielding. It is also
clear from Figure 4.13 that the observed crazes are not
due to beam damage. A craze once formed will extend and
grow if stressed. Figure 4.14 illustrates the observed
crazing in the machined Rigidex specimen. The results from
this limited study suggest that the extent of crazing
observed for the machined Eltex pipe sample is
significantly greater than that observed for the machined

Rigidex specimen.
The influence of the cutter speed was investigated by
carrying out machining operations at three different
speeds (234,958 and 1500 r. p. m). The OIT for all the
three speeds was approximately (0-12 minutes) when tested
at 205°C. The extent of yielding in the slow speed cutter
was significantly more than that observed for the fast
cutter speed.
Page 97
Figure 4.12 Illustration of the crazes for the

machined outer surface of the Eltex pipe.
Figure 4.13 High magnification view of the crazes region

in the machined outer surface of Eltex.
Page 98
Figure 4.14 High magnification view of the outer

surface of the Rigidex pipe.
Figure 4.15 High magnification view of the previously

machined Eltex pipe sample which was then
annealed between two glass slides 125°C
at
for 80 minutes in a vacuum.
Page 99
With reference to the above mentioned micrographs, the
observed crazing may help to account for the drastic
reduction in the OIT for machined specimens. However, the
following points need to be considered:

(i) The crazing significantly increases the surface area
of the test specimen;

(ii) The crazing present in the test specimens may also
provide an easier path for any migration of additives out
of the polymer. The presence of these crazes would also
make it easier for "external species" to diffuse into the
polymer and possible enhance the rate of thermo-oxidative
degradation during the OIT experiment.
Figure 4.15 illustrates the surface of Eltex pipe which

was previously machined and then sandwiched between two
glass slides and annealed at 125°C for 80 minutes in a

vacuum. This test was carried out to remove the crazes
introduced into the specimen as a consequence of the
machining operation. The morphology of the annealed
polymer is clearly visible. The crazes have been removed
as intended. OIT tests on the specimen from which Figure
4.15 was taken also gave very low values. Therefore, the
increase in surface area alone can not explain the
observed reduction in OIT for the machined specimens.

However, it is possible that the increased surface area
could make it easier for the antioxidant system to be less
effective - either as a consequence of migration of
species to the surface of the polymer and subsequent loss
from the polymer or because the system is made inactive as
a consequence of some chemical reaction with the
environment.
Page 100
4.4.2.6 SUMMARY
Sample preparation technique can be an important factor in
determining the thermal stability of PE pipe materials as
measured by DSC. Knife-cut samples gave consistent
results.
The OIT for the Eltex resin is significantly affected by
the action of machining. It is suggested that the decrease
in OIT may be due to volatility of the antioxidant or

surface damage caused by the action of machining which in
turn is creating extensive crazes on the polymer surface.

It is unlikely the crazes thenselves could lower the OIT
since these features will not be present above the melting

point (129°C). However, it is possible that air could be
absorbed in to the crazes leading to a premature decrease

in the OIT. It is also likely that the presence of a "weak
boundary layer" in terms of a surface layer which has been
depleted of antioxidant could accelerate the on-set of
degradation of the polymer. This is a reasonable
hypothesis since the proposed mode of degradation is via a
free radical reaction. ESR studies have shown that the
Eltex polymer has a higher concentration of free radicals.
Page 101
Chapter 5 Weld Geometry and Dimensions
WELD GEOMETRY and DIMENSIONS
5.1 INTRODUCTION
The most commonly used method for assessing the weld
quality in the field is the inspection of the weld bead.
The visual appearance and the width of the beads are used
as a qualitative indicator of the intregrity of the weld .

For example, the water industry specification WIS 4-32-08
[WIS, 19881 specifies a range of bead widths for the
various pipe diameters for butt fused welds. The visual

inspection of both the debeaded internal and external
beads are also specified. In this study, a detailed
characterisation was carried out on both the bead geometry
and-dimensions.
5.2 SAMPLE CODING
A schematic illustration of the coding system used in the
study is presented in Figure 5.1. With reference to
Figures 5.1 and 5.2, the top position corresponds to the
top section of the pipe in relation to the clamping
position when the weld was made. Likewise the bottom
section corresponds to the bottom section of the pipe.

Although eighteen individual specimens were cut from
around the circumference of the pipe, only the macro and
Page 102
Top Position
Outer bead
Inner bead
---ý' (surface)
(Bore) 'ý
_
N%ý
Lateral left
t
% Lateral Right
position position
,
"*
----ý"0
Bottom position
Figure 5.1 Schematic illustration of the coding system

used for the bead dimensions measurement.
Weld line
Top
position
Clamp I Pipe I Pipe I Clamp
Bottom
position
Figure 5.2 Schematic illustration of the pipe during

welding to highlight the top and bottom
positions.
Page 103
microscopic data for the samples for specimens from the

four positions indicated in Figure 5.1 are presented.
5.3 BEAD GEOMETRY - VISUAL ASSESSMENT
Figure 5.3 illustrates the visual appearance of the outer

and inner bead as a function of the position along the
circumference of the pipe for the standard Eltex weld

(SS1). With reference to Figure 5.3, a number of features
are readily apparent:

(i) the shape of the outer bead in Figure 5.3a, b, c and
d, are different. Fox example in Figure 5.3a, the extent
of sagging is greater than that observed in Figure 5.3c
(top and bottom). This has been attributed to the
influence of gravity on the beads. This effect is also
apparent on comparing the inner beads as shown in Figure
5.3a and 5.3c. In Figure 5.3a, the inner bead is seen to
sag away from the bore of the pipe whereas in Figure 5.3c,
it is seen to curl towards the bore of the pipe. A
schematic illustration of the influence of gravity on the
extent of sagging is shown in Figure 5.4.
(ii) on comparing the outer and inner beads, for example

in Figure 5.3a and 5.3c the difference in the bead
geometry may be attributed to the microstructure of the

bore and surface regions of the pipe. Figure 6.13,
(section 6.4.1, page 151) shows the presence of an
elongated spherulitic structure in the bore region of the
pipe. However, this structure was not observed on the
surface of the pipe. The elongated spherulitic structure
was observed on a significant proportion of the inner bead
as shown schematically in Figure 5.5. The retention this

of
Page 104
top (a)
lateral
lateral bore right
left (b)
(d)
... AM&, ýi
3f
bý
bottom (c)
10mm
Figure 5.3 Macrostructure of the SS1 Eltex weld around
the circumference of the pipe.
flow flow
of polymer at bead of molten polymer
bead
of the inner bore of the at of the outer
pipe without the surface of the pipe
influence of gravity
flow of molten polymer

at bead of the outer
surface of the pipe
without the influence
flow of molten of gravity
polymer at bead
of the inner bore
of the pipe V
Figure 5.4 Schematic illustration of the influence of

the gravity on the molten (extruded)
polymer around the circumference of the
pipe.
Page 105
elongated spherulitic structure in the inner bead is
because after the initial bead is formed, the outer layers
corresponding to the bore and surface regions of the pipe
are bent backwards as the bead-up process continues. This
process is shown schematically in Figure 5.6a, b, c and d.
As a consequence of the presence of the elongated
spherulitic structure, and the fact that the melting
points and yield strength of the material in the bore
region are greater, some structural support, is offered to
the inner bead. This support is not available for the
outer beads.
Outer bead
Surface
Bore
Inner bead
Figure 5.5 Schematic illustration of the weld bead

formed during the welding operation.
Page 106
Pipe Pipe (a)
Pipe Pipe (b)
Stagnant
region
Pipe I Pipe (c)
The skin is
bent back'
Outer bead
Molten
Outer shear
material
zone
Weldline Surface
(d)
Bore Inner shear

zone
Inner bead
Figure 5.6 Schematic illustration of the welding

process to. highlight the formation of the
beads and the different zones produced.
Page 107
An experiment was carried out to investigate if, the
melting and recrystallisation of the bore region would
give rise, to the elongated spherulitic structure as
observed in the as-received pipe. This experiment showed

that the original structure in the bore region could not
be reproduced by using similar conditions to that used in
the welding process.
(iii) on comparing the Figure 5.3a to d, there is a

general trend irrespective of the position along the
circumference of the pipe, for the outer surface near the
weld zone to be inclined towards the bore. This feature
was observed for all the welding conditions used in this
study. An investigation was made to identify the extent

and nature of the observed deformation. A section of pipe
was planned square on the welding machine. The inner and
outer diameters of the pipe were measured across the
circumference of the pipe at twenty points, at a distance
of 5mm and 35mm from the end face of the pipe. This
procedure was carried out at specified locations around

the circumference of the pipe. The measurements were
carried out on the Mitotoyo co-ordinate measuring machine,
model FN503.40cm of the pipe was then immersed in an oil

bath which was maintained at 125°C for a duration of 180
and 300 seconds. The inner and outer diameters were re-
measured at identical locations to that carried out
previously. The results from this study are presented in
Table 5.1.
With reference to Table 5.1, the largest deviation in
diameters occurs for the pipe with 300 second soaking

time. However in both cases the deviation for the inner
Page 108
surface at 5mm and 35mm is greater for the bore. The
observed deformation as a function of heating has been
attributed to the state of residual stress in the pipe.

These residual stresses arise due to the differential
cooling between the surface and the bore regions of the pipe.
OUTER SURFACE INNER SURFACE
SOAK TIME DISTANCE FROM END FACE (mm)
(SECONDS)
5 35 5 35
180 1.41 0.60 2.37 1.99
300 0.85 1.34 3.34 2.80
Table 5.1 Deviation in diameters for outer and inner

pipes at 5mm and 35mm away from the end
face of pipe.
The outside machining layer method using bending
measurements appears to be the common method for measuring

the residual stresses in the pipes. typical values for
63mm diameter MDPE pipe in the hoop direction are on

-3MPa
the outside and 1.5MPa of i the inside [Williams et al..
1981; Marshall et al., 1982]. However, the profiles of
residual stress through the pipe wall will depend on the
thermal history, pipe diameter and pipe wall thickness'.

Reinke and Potente [Reinke and Potente, 1982] have
reported that the magnitude of these stresses in 110mm
diameter SDR 17 HDPE pipe to be in the region of 12Nmm-2
and 5Nmm-2 for the outer and bore regions respectively.
Page 109
The proposed explanation for the observed bending of the
outer surface towards the bore during the welding process

has being attributed to the shrinkage that occurs due to
the relaxation of the residual stresses. This has also

being observed by Parmar [Parmar, 1986]. The action of
cutting a strip from the pipe will also result in some of

the residual stresses to be relieved [Doshi, 1989].
Typical examples of the bead geometeries for Eltex welds
made with a high and low pressure (0.42MPa and 0.02MPa)
are shown in Figures 5.7 and 5.8. A similar trend to that

discussed previously was observed for these welds, with
the exception of the dimensions of the beads.
Figures 5.9 to 5.10 show the observed bead geometry for
the Rigidex welds made under the standard and low pressure
(0.15MPa and 0.02MPa). Similar trends to that discussed
previously were observed for these welds with the
exception of bead shape and bead volume.
5.4 WELD GEOMETRY - QUANTITATIVE ASSESSMENT
5.4.1 BEAD WIDTHS
The coding for the various dimensions within the weld zone
have being presented in chapter 3, Figure 3.6 (page 67).
Figures 5.11 and 5.12 summarise the widths of the beads as
a function of the circumference along the pipe for
standard weld (SS1) and a low pressure weld (SS12).
Figures 5.13 and 5.14 show measured bead width for a
standard Rigidex weld (RR1) and a low pressure Rigidex
Page 110
top (a)
lateral
lateral right
left (b)
(3)
bottom (c)
10mm
Figure 5.7 Macrostructure of the SS5 Eltex weld around
the circumference of the pipe.
top (a)
lateral
lateral right
left (b)
(d)
bottom (c)
10mm
Figure 5.8 Macrostructure of the SS12 Eltex weld
around the circumference of the pipe.
Page 111
top (a)
lateral
lateral right
left (b)
(d)
bottom (c)
10mm
Figure 5.9 Macrostructure of the RR1 Rigidex weld
lateral
lateral bore
right
left (b)
(d)
bottom (c)
10mm
Figure 5.10 Macrostructure of the RR12 Rigidex weld

Page 112
weld (RR12) as a function of their respective positions
along the circumference of the pipe. It is readily
apparent that the width of the beads are a function of
their relative positions along the circumference of the
pipe. The outer-width for the top position was always
smaller than the outer bead width for the bottom position.
Whereas, for the inner bead, the bead width for the top
position was always greater than the inner bead at the
bottom position. These trends were observed for all the
welding conditions in this study. The data for the other

bead widths for the other welds investigated in this study
are presented in Appendix 2.
5.4.2 BEAD AREAS
The bead area gives an indication of the volume of

material extruded during the bead-up and heat-soak phases
of the welding process. The bead areas as measured by the

image analysis for the Eltex and Rigidex at the four
positions, top, bottom, lateral left and lateral right are
also represented in Figures 5.11 to 5.14. With reference

to Figure 5.11, it is readily apparent that the inner bead
always exhibits a larger bead area than the outer bead
area for the top position. The trends were similar to that
observed for the bead-width. An explanation for the

dependence of the bead-widths and bead areas in relation
to their position along the circumference of the pipe is
proposed with reference to Figures 5.15 (SS1, top
position) and 5.16 (SS1, bottom position). In comparing

Figure 5.15 and 5.16 the following conclusions have been
made:
Page 113
18 90
0 0°°0°0°0
16 00 80
v°ov
°ovv0vo
E-°
vv vvvvvv
E ry
14 70
'd
3
'd
12 X Aý 60
cý
X
X Outer Area
® Inner Area
10 0 Cuter Width 50
v Inner Width
A A^
-V
05 10 15 20
fbsitiam axcu pipe
Figure 5.11 Graph of bead width and bead area versus

position around the pipe for SS1 Eltex
weld.
18 80
70
16
O0OOO
O
E----0 O 60
114
0 O
b
vv0
0 50
a
pOg
12
oovo oý
40
0v
-v v
vv
10 vvvvvv
30
X
A LV
05 10 15 20
Position around pipe

position around the pipe for SS12 Eltex
weld.
Page 114
18 80
Q
16
° >v
vv °
60
v
14 <v
v
fi 0° Oo
ýE v 000 Ov
0 vo0
0o o ° °° 0
°v v° 0
12
40
10 0 res area
A outer area
0 Outer Width
v tracer Width
G 20
05 10 15 20

position around the pipe for RR1 Rigidex
weld.
18 60
o0o
o0
16 00 50
00
E----o 0
14 0v 40
°
'd 11
00
0 b
12 30
w
v l>
xvv
vv
°vv®
10 Inner Area 20
lIE Outer Area
v Irmer Width
0 Outer Width
0 in
05 10 15 20

position around the pipe for RR12 Rigidex
weld.
Page 115
surface
bore
5mm
Figure 5.15 Microtomed section of the SS1 Eltex weld,

from the top position of the pipe.
bore
surface
5mm
Figure 5.16 Microtomed section of the SS1 Eltex weld,

from the bottom position of the pipe.
Page 116
(i) The bottom outer bead was always greater than the top
outer bead and,

(ii) the top inner bead was always greater than the bottom
inner bead.
These conclusion are also supported by the evidence
presented in Figures 5.11 to 5.14 and the data presented

in Appendix 3.
Bead area was found to increase up to 0.21MPa, after which

it leveled off with further increases in pressure. This
non-linear relation between the applied pressure and the
total bead area is due to the fact the butt-fusion machine
used in this study was pressure controlled rather than

displacement controlled. In other words, for a given bead-
up pressure, heat-soak time and welding temperature, the
amount of material in the molten state will be constant.

Since the machine operates under pressure control, most of
the molten material will be extruded if the pressure is
increased.
The effect of the changes in the welding pressure on the
outer and inner bead areas for the Eltex welds are
presented in Figures 5.17-5.20. Figure 5.17 and 5.18
illustrate the effect of changes in the welding pressure
on the outer and inner bead areas for the four positions,
top, lateral right, bottom and lateral left. The outer

bead at the bottom of the pipe has a larger area than the
top, with the other two positions lateral right and lateral
left having areas somewhat in between the two. The reverse
occurs for the inner weld bead with the bead area being
larger at the top of the pipe. For clarity the lateral
Page 117
100
80
ý_
A ýýY
60
40
LateralLeftOuterBead
v LateralI ght OuterBead
Q
Q Q Top OuterBead
o -o BottomOrter Bead
0 Q2 0.4 0.6 0.8
V1kldngPressure(NPa)
Figure 5.17 Graph of outer bead area for the. top,

lateral right, bottom and lateral left
positions of Eltex pipe versus welding
pressure.
. LateralRgIt AmerBead
v lateral Left InnerBead
o-o Bdtorn Irrer Bead
Q-Q Top Irrer Bead
o a2 as 0.6 o.8
v angPressue
(M)
Figure 5.18 Graph of inner bead area for the top,

lateral right, bottom and lateral left of
Eltex pipe versus welding pressure.
Page 118
ý_
0 0.2 0.4 0.6 0.8
Wddng Pressire (M)
Figure 5.19 Graph of bead area for the top of pipe for
the Eltex versus welding pressure.
1W
80
¶ 60
4Q
Q BOftcm triff Bead

LORI Otter Bead
LO
20
welting Pres&re (NPa)
Figure 5.20 Graph of bead area for the bottom of pipe

for Eltex versus welding pressure.
Page 119
right and lateral left position data are omitted from
further plots. Figures 5.19 and 5.20 illustrate the effect
of welding pressure on the outer and inner bead areas for
the top and bottom of the pipe. With reference to Figure
5.19 and 5.20, it is readily apparent that the inner bead
at the top position always exhibits a larger bead area

than the outer bead, whereas the outer bead at the bottom
of the pipe has a larger area than the inner beads. Figure
5.21 to 5.24 illustrates the effect of changes in the
welding pressure on the outer and inner bead areas for the
top and bottom positions of the pipe for Rigidex welds.
The curves for the Rigidex are similar to the Eltex in
terms of their form. Bearing in mind that the heating
times and pressures are the same for both Eltex and
Rigidex, this means that the Rigidex material melts faster
and therefore produces more melt. However, the influence
of molecular weight is again apparent as lower molecular

weight Rigidex produces larger beads at a given pressure
than those of a higher molecular weight Eltex.
The data presented in Figure 5.25 5.26 indicates that as

the heat soak time is increased, a corresponding increase
in the bead area is obtained. This is obvious because as
the heat soak time is increased, the amount of molten
material at the pipe face will increase. Furthermore, the
trends observed for the series of experiments were similar
to those discussed previously.
5.4.3 MELT AFFECTED ZONE (MAZ) - QUANTITATIVE ASSESSMENT

Once the welds had been microtomed, they were
photographed. A selection of the photographs were viewed
Page 120
5 Weld Geometry and Dimensions
Chapter
80
60 Y
z
zY
40
o Top OuterBead
ý0 Q Top Irrer Bead
20
0 Q2 0.4 0.6 0.8
Weidrg Pressi.re (NPa)
Rigidex versus welding pressure.
1 -UI--
80
60
t1
9
40 00Bead
o BottomIma Bead
2n
°0 Q2 0.4 0.6 0.8
Weläng Pressire (NPa)

for Rigidex versus welding pressure.
Page 121
80
ý_
Q
40 0
Q
00 BottomOder Bead
QQ TapOuterBead
201
weläng Pressire (MPa)
Figure 5.23 Graph of outer bead area for the top and
bottom of Rigidex pipe versus welding
pressure.
IW
80
ý_
60
a
0
40 CP
o Betan OuterBead]
o TopImerBead
I
201
welärg Pressre(Wa)
Figure 5.24 Graph of inner bead area for the top and
pressure.
Page 122
using a video camera linked to a Cambridge Quantimet 520
image analysis system. This enabled the outline of the
weld to be identified and a number of other dimensions
shown schematically in Figure 3.6 were measured. With
reference to Figure 3.6:
OM: MAZ length outer;

MM: MAZ length centre;
IM: MAZ length inner.
These measurements were correlated to the lengths of the
melt zones in the outer, inner and centre sections of the
pipe. The angle made by the outer (60) and inner beads
(91) to the pipe outer and inner surfaces as well as the
area of the MAZ were also measured for the top, lateral
right, bottom and lateral left position around the pipe

circumference. In Figures 5.27-5.28, the outer and inner
MAZ lengths for the top and bottom of pipes for Eltex weld
are shown plotted against welding pressure. As the welding
pressure is increased, the outer and inner MAZ lengths
showed a gradual decrease in the MAZ lengths for pressures
up to 0.32MPa, followed by leveling off of the MAZ lengths
with further increases in pressure. The reasons for this
were discussed previously. The top and bottom positions
always follow the following trends:
(i) top outer was less than the bottom outer and
(ii) top inner was greater than the bottom inner.
This observation may also help to explain the shapes of

MAZ and the profiles of the shear zones; these parameter
have been defined schematically in Figure 5.6.
Page 123
150
140
120
100
T)
L
80
60
p Top Inner Bead
0 Top Outer Bead
ao
100 200 300 400 500
Heat Soak Time (Sec)
Eltex versus heat soak.
100 200 300 400 500

HestSoakTime(Sec)

for Eltex versus heat soak.
Page 124
1
Q Imer MAZ Length
0 Outer MAzLength
3
E
9 O
O
Q O Q
a o 13
0
0
a 0
0 0.2 0.4 0.6 0.8
V1Relcing
Pressure(NFa)
Figure 5.27 Graph of outer and inner MAZ length for

the top of Eltex pipe versus welding
pressure.
10
Q 1rra MAZ Length
o Outer MAZLength
I 6
78
0
0
0 0
4
Q
Q Q Q
13
0
13
0 0.2 0.4 0.6 0.8
Waking Pressare(MPa)

the bottom of Eltex pipe versus welding
pressure.
Page 125
In general the length of MAZ was found to be a function of
the position along the circumference of the pipe and the
welding pressure. The inner MAZ length was greater than
the outer MAZ length for the top of pipes, whereas the
trend was reversed for the bottom of the pipe. The length
of the MAZ was always found to be the greatest when the
bead area was the greatest. Micrographs as seen in Figures
5.3,5.7 ans 5.8 (SS1, SS5 and SS12) clearly illustrate
this point.
Figure 5.29-5.30 illustrates the outer and inner MAZ
lengths for the top and bottom of pipes for Rigidex welds.
The data for the Rigidex pipes showed a similar trend to
that observed for the Eltex welds.
Unlike these outer and inner MAZ lengths, when the centre
MAZ length is plotted against welding pressure, (Figures
5.31-5.32), we see the emergence of a curve common for the
top and bottom of the pipe for both Eltex and Rigidex
welds. With reference to these Figures, an increase in
pressure, decreases the length of this central MAZ region
and this is also true for the inner and outer MAZ lengths.
From the data presented in Appendix 3, it is possible to
see that although the physical dimensions of the MAZ
change with changes with pressure, the MAZ area is also
altered. This means that the melt is being redistributed
rather than having its volume changed and at higher
welding pressures more melt is forced out between the pipe
ends resulting in a wider and slightly higher bead and a
narrower central melt layer separating the solid bulk
material of the pipe. Appendix 3 lists the catalogue of
Page 126
Q InnerMAZLength
0 OuterMAz Length
Cb
8
6 Q Q
Q
o
13
o o
4
0
0
0
V dragPresssire(MPa)

the top of Rigidex pipe versus welding
pressure.
Q InnerMAZIeVh
0 Outer MAZlength
6
aý
1
4 1 0
0
0
,L
0 0.2 0.4
WeldngPressire(Na)
0.6 0.8

the bottom of Rigidex pipe versus welding
pressure.
Page 127
O Bottan of Pipe
0 Top of Pipe
'4
a a a
0
A
v00.2
0.4 0.6 0.8
Welding Pressure (MPa)
Figure 5.31 Graph of centre MAZ length, for the top and
bottom of Eltex pipe versus welding
pressure.
O Bottom of Pipe
0 Top of Pipe
14
41 2 a
D
0
a a 0
a
I,
V
0 0.2 0.4 0.6 0.8
Figure 5.32 Graph of centre MAZ length for

the top and
pressure.
Page 128
weld dimensions that had been produced for the Eltex and
Rigidex welds. The data in this Appendix coupled with the
deformation studied can facilitate a comprehensive data
base which can be used to assist on-site visual joint
inspection.
Figures 5.33-5.34 illustrate the MAZ area plotted against

the welding pressure for Eltex and Rigidex welds. The MAZ
area gives an indication of the volume of material melted

and depth of melt produced during the bead-up and heat-
soak phases of the welding process. As the welding
pressure is increased so there is a decline in the MAZ
area. The curves for the top and bottom for Eltex and
Rigidex are similar to one another in terms of their form
and gradient.
5.4.4 SHEAR ZONE MEASUREMENT
All the Eltex welds were microtomed and viewed in the
microscope to estimate the shear zones via an eyepiece
graticule for quantification at 3 different positions,
namely the outer, centre and inner sections of the pipe.

The shear zones in the welded joints are of significance
because other researchers have found that during a tensile
test the failure is initiated at the intersection of the
shear zone and the pseudo-notches. Furthermore, on
propagation the crack path was observed to go across from
one end of the shear zone pseudo-notch intersection to the
opposite diagonal. The shear zone size has also been
identified by a majority of researchers as the
microstructural feature that has a significant influence
on the mechanical properties of butt fused joints.
Page 129
5 Weld Geometry and Dimensions
Chapter
120
D Bottan of Pipe
0 Top of Pipe
90
60
Q
® 0
30
0
oý
0
welding Pressure (MPa)
Figure 5.33 Graph of MAZ area for the top and bottom
of Eltex pipe versus welding pressure.
120
Q Bottom of Pipe
O Top of Pipe
90
I
Nl
60 a
30
8 o
8 g
0
0
0 0.2 0.4 0.6 0.8
Figure 5.34 Graph of MAZ area for the top and bottom
of Rigidex pipe versus welding pressure.
Page 130
100
ä 80
60
40C
CD a Outer Pipe
o Inner Pipe
20
0 0.2 0.4 0.6 0.8
Welding Pressure(MPa)
Figure 5.35 Graph of shear zone width for the top

position, for the outer and inner Eltex
welds versus welding pressure.
i
Fm
ýý §
S.
iT. 3F
ao
o Outer Pipe
o Inner Pipe
'n
Welding Pressure(MPa)
Figure 5.36 Graph of shear zone width for the bottom

of pipe, for the outer and inner Eltex
welds versus welding pressure.
Page 131
Appendix 4 gives a summary of the results of shear zone
width measurements obtained for the top, lateral right,

bottom and lateral left positions around the pipe
circumference. The values listed are an average of 5
readings. From Appendix 4, it is apparent that the shear
zone width increases with an increase in the welding
pressure for both the inner and outer positions. The

difference in the shear zone width between the outer
surfaces and the mid-section can be attributed to the
faster cooling rate in the former. However, the width at
the inner position for both the top and bottom of the pipe
gives the highest values. Figure 5.35 and 5.36 illustrate
the outer and inner shear zone width for the top and
bottom of pipe versus welding pressure. As the welding
pressure is increased the shear zone widths increase up to
a welding pressure of 0.15MPa. Upon further increase of
the welding pressure, no significant change in the shear
zone is seen. This is attributed to the fact that the butt
fusion machine used during the course of this study was
pressure controlled.
Page 132
Chapter 6 Macro and Microstructural Examination
MACRO AND MICROSTRUCTURAL EXAMINATION
6.1 INTRODUCTION
In the welding process heat is applied to melt the
surfaces to be joined and these surfaces are then pushed

together and allowed to cool. It is inevitable, therefore,
that the different cooling rates in the welded region
cause a different number of zones to form. The number of

zones formed will depend on the welding temperature, the
melting point of the polymer, the thermal diffusivity of

the material at the relevant temperature, shear rate and
the welding pressure. In order to obtain information about
the weld structure, a selection of welds were investigated
by viewing thin sections in a transmitted light microscope
with the specimen being rotated between crossed polars to
allow changes in birefringence to be observed. In this
chapter the results of that investigation are recorded and

discussed and reasons for the observed structures are
suggested.
6.2 MICROTOMY
In order to carry out a detailed and reliable analysis of

the microstructure via optical microscopy, it is important
to obtain microtomed sections without scratches. However,
Page 133
this is very difficult in practice as it is dependent on
the quality of the cutting edge and the nature of the
polymer in question.
A detailed study was undertaken in the course of this
research programme to identify techniques for microtoming
"large sections" of butt fused polyethylene pipes. The
term large sections refers to sections of approximately 8-
10 µm thick, 40-50 mm long and 30-35mm wide. There is very
little published information on techniques for microtoming
MDPE and HDPE pipe sections. Most of the literature in the
public domain deals with techniques for obtaining thin
sections of biological and polymeric samples.
As expected, the quality of the cutting edge of the blade,
determines to a large extent, the quality of the
microtomed surface with respect to scratches. Disposable
blades in their às-received' state were not suitable for
microtoming PE. It was necessary to polish these blades
first to obtain a good quality cutting edge. Chemical
polishing and manual polishing were attempted in order to
improve the cutting edge.
6.2.1 CHEMICAL POLISHING
A 50% aqueous solution of hydrochloric acid, nitric acid
and phosphoric acid was prepared and heated to 60°C. The
blade was lowered into the solution until the cutting edge
was immersed in the solution. The blade was removed from
the solution every 20 minutes, ' washed and inspected under
the microscope for evidence for an improvement in the
cutting edge. The blade was then returned to the solution.
Page 134
The process was then repeated until the blade had been in
the solution for a period of 2 hours.
The chemical polishing resulted in the blade edge
corroding very rapidly on exposure to the atmosphere.
6.2.2 MANUAL POLISHING
Manual polishing was carried out using standard

metallographic techniques using a specially built jig as
shown in Figure 3.4. An improvement in the microtomed
samples was observed, however it was not possible to

obtain scratch-free microtomed sections.
As a consequence of the above mentioned difficulties with

microtoming, a conventional D-profiled steel knive was
modified as shown in Figure 3.5. The polishing sequence
was described in Chapter 3 (section 3.3.2.2, page 64).
Previous researchers working with filled polymers have
proposed diamond-tipped knives to be the best for ultra-

microtomy. Glass knives have been reported as a cheaper
second choice. The major disadvantage with glass knives is
that they do not maintain their cutting-edge.
It is readily apparent from Figure 6.1 that the polished

steel knife experiences progressive damage as a function
of the number of microtomed sections. With reference to

Figure 6. la, the micrograph shows the presence of a flaw
caused by the blade. This flaw becomes more apparent after

taking 12 successive microtome sections, Figure 6.1b.
Additional flaws also become more visible as more sections
are taken. Figure 6.1d represents a micrograph of the
Page 135
microtomed section after 25 sections. The extent of
microtomy-induced scratching was greater for the MDPE pipe
compared with the HDPE pipe. Possible reasons for this
observation could be:
(a) the different levels of filler loading in the two
polymers;
(b) the differences in the mechanical properties of the
two resins and;

(c) the nature of the fillers.
The success of microtoming may be attributed to the
modified knife cutting edge. Re-polishing after use should

be limited to compensate the state of wear of the blade.
The 38° polished surface becomes wider after each re-
polishing and therefore will take longer to refurbish. The
width of the polished surface on the cutting edge may be
allowed to increase up to a maximum 0.75mm, thus allowing

the operator to polish the 40° angle for the first few re-
polishes. With time the width of the relief angle will
increase to the point where it becomes uneconomical to
polish the blade further. It is thus recommended that the

blade be re-ground at 30° to restore the original relief
angle. This study has shown that the judgment and
competence of the operator is an important factor in
maintaining an efficient cutting edge on the knife.
7.3 MACRO-STRUCTURE
Figure 6.2 illustrates the macro and microstructural key

features which were observed in all the welded samples.
The term macro-structure is used to describe the general
features of the weld zone. The macroscopic appearance of
Page 136
(a)
4th cut
(b)
12th cut
\ý /
ý:0 cut
(d)
25th cut
2mm
Figure 6.1 Microtomed sections of the pipe to

illustrate the developments of scratches
from the knife section thickness 15 µm.
Page 137
the weld zone for the standard, low and high pressure
Eltex welds (SS1, SS12 and SS5), viewed under crossed
polarised light are presented in Figures 6.3-6.8. The
Rigidex welds made under standard and low pressure are
also presented in Figures 6.9-6.10. With reference to
these Figures the different shades are due to the degree
of retardation of the polarised light. The contrast

between the weld and the MAZ is due to the difference in
orientation of the material. For example, the sharp

weld/MAZ boundary indicates that a sharp change in
orientation had occurred. This is clearly indicated in
Figures 6.3-6.10. All the micrographs illustrating the
macro-structure in Figures 6.3-6.10 were photographed with

a favourable angle of the polariser to show as clearly as
possible both the weld and the MAZ regions. The material
adjacent to the weld interface had molecular orientation
which was not parallel to the weldline. Instead, the
material appeared to be aligned at about 450 to the weld

interface. This suggests that shear rather than
elongational flow occurred near the interface. In order

for shear flow to occur within the MAZ, there must be some
variation in the flow rates of the melt across the MAZ.
This may arise due to the viscosity gradient across the
MAZ between the hot molten polymer and the cooler parent
material. The situation may be exacerbated further the
as
heater plate is removed and the surface of the molten
material is allowed to cool slightly. This means that the
hottest material is within the centre of the MAZ between
the parent material and the cooler surfaces. As the pipe
ends are brought together to close the joint, this hotter
melt is squeezed out into the weld beads, resulting in
Page 138
We1dline
Oute
SL
ones
Inner
Weldline
5mm
Figure 6.2 Micrograph of the Eltex weld to illustrate

the key macro and microstructures features
in weld.
Page 139
Chapter Macro and Microstructural Examination
bore :>urf ace
5mm
Figure 6.3 Microtomed section of SS1 Eltex weld from

the top of the pipe.
bore :surface
5mm

the lateral right position of the pipe.
Page 140
bore surface
'.,.,, wº
,,. rr..
5mm

the bottom of the pipe.
5mm

the lateral left position of the pipe.
Page 141
shearing flow at both the weld interface and the melt-
parent material interface.
The general polymer flow pattern which occurs in the weld
zone during the welding operation can be seen on

inspecting Figure 6.7 (SS12, low pressure welds). The flow
direction appeared to be between 45° and 500 relative to
the weld interface. A schematic illustration of the
movement of the molten material has been shown previously

in Figure 5.6. Barton and Cherry [Barton and Cherry, 1980]
have also shown the presence of these flow lines in the
weld. The lines followed the direction of the melt flow in
the weld, away from the centre out towards the surface and
bore regions into the weld beads rolling around to form
the bead. To a first approximation, the flow field during
the process of welding appears similar to that established

when two fluids impinge from mutually opposed jets as
observed by Macley and Keller [Mackley and Keller, 1975].
At the centre of the weldline there is a stagnation point
where the lines diverge and turn through 90° before
running parallel to the weld interface indicating how the
melt had been squeezed between the pipe ends to form the
weld.
A good correlation was observed between the lengths of the
MAZ as shown by Figures 6.3-6.10 and the extent of
curvatures observed between the parent material and the

MAZ. The length of MAZ corresponding to the surface, bore
and mid-regions are summarised in Appendix 2 and were

discussed in the previous chapter.
Page 142
top
_ lateral left
_ý:
bottom
p
lateral right
5mm

four different position around the
circumference of the pipe.
Page 143
i
top
- f.
lateral left
g`
r
-ý
bottom
lateral right
ýt ýý
5mm

Page 144
top
lateral left
bottom
lateral right
04,
5mm
Figure 6.9 Microtomed section of RR1 Rigidex weld from

Page 145
top
lateral left
Sul
i: _i
bottom
5mm
Figure 6.10 Microtomed section of RR12 Rigidex weld
from four different position around the
Page 146
Macro-structures in the full range for the Eltex welds

investigated in this study are presented in Figures 6.11-
6.12 as a function of joining pressure and heat-soak time.
The macro-structural features observed in these welds

follow the same trend as that discussed previously for the
standard, low and high pressure welds.
6.4 MICROSTRUCTURE
6.4.1 MICROSTRUCTURE OF THE EXTRUDED PIPE
Three distinct microstructural features were observed for
the surface, bore and mid section of the extruded pipe.

The microstructure of the bore in Eltex pipe is
illustrated in Figure 6.13 and the key feature here was
the presence of elongated spherulitic structures. This
microstructure has been attributed to the slower cooling

rate at the bore during the extrusion process.
The observed microstructure of the surface region of the
pipe is illustrated in Figure 6.14. However it was not

possible to resolve the microstructure of the surface
region using optical microscopy. During the extrusion
process, the pipe is sprayed with water at 'ambient
temperature to cool the pipe. This results in a finer
microstructure at the surface region. Figure 6.15

illustrates the observed microstructure of the mid section
of the extruded pipe. Although no quantification on the
relative size distribution of spherulites in the different
zones was undertaken, on the basis of visual examination,

the dimension of the spherulites in the bore and mid
section were similar but significantly larger than the
surface.
Page 147
Chapter_ 6 Macro and Microstructural Examination
top position bottom position
0.02MPR
-. ISNIL
',
0.21NIta
iP. 'i . 0)
Page 148
0.3lMPa
"q 1
ýx, .n
ý.
0.41MPa
1MPa
h
5mm
Figure 6.11 Macrostructural features of top and bottom

of Eltex welds as a function of joining
pressure.
Page 149
Heat-Soak (180 seconds)
r- "Ilt,
5mm
Figure 6.12 Macrostructural features of top and bottom

of Eltex welds as a function of heat soak
time. Joint pressure 0. lSMPa.
Page 150
20µm
Figure 6.13 Microtomed section of SS1 Eltex weld

showing elongated spherulites at the inner
bore of the pipe. Mag. 400.
20µm

the outer surface. Mag. 250.
Page 151
20µm

the central region. Mag. 400.
6.4.2 MICROSTRUCTURE OF THE MAZ
Figure 6.2 illustrates the coding system which was used to
describe the microstructure of the weld-zone. In addition
to this, an illustration indicating the weld in

region
question is also shown alongside the discussion.
With reference to Figure 6.3, two shear zones were
observed between the parent material and the MAZ. This
microstructure has been attributed to crystallisation
under shear. The dimensions of these shear zones for the
four positions along the circumference of the pipe (top,
lateral right, bottom and lateral left) were measured and
are presented in Appendix 3. The shear zones were measured
Page 152
at regions corresponding to the surface, bore and mid
section of the weld. The width of shear zones at the bore
regions were greater than the surface region with the mid
section showing the lowest value for all the Eltex welds
investigated in the study. The difference in the shear
zone width between the outer surfaces and mid section can
be attributed to the faster cooling rate in the former.
MAZ
shear zone
parent
20µm
Figure 6.16 Microtomed section of SS1 Eltex weld

showing MAZ, shear zone and the parent
material. Mag. 100.
Figure 6.16 illustrates the observed microstructure for
the parent material, MAZ and shear zone for the SS1 weld.
The shearing action of the molten MAZ causes immense
distortion to the crystal structure. The presence of a
pseudo-notch also coincides with the termination of the
shear zone. it is likely that a combination these two

of
factors will provide an initiation site for crack
Page 153
propagation if an appropriate tensile load was applied.
The shear zone size has been identified by a majority of
researchers as the microstructural feature that has a
significant influence on the mechanical properties of the
butt fused joints [Menges and Zohren, 1967; Atkinson and

DeCourcy, 1981; Gedhe et al., 1992; Nieh and Lee, 1993].
No significant d&fferences were detected on inspecting the
weldline from the surface to the bore region. However, the

dimensions of the spherulites were much finer than that
observed for the other regions in the weld. An estimation
of the extent of orientation may be inferred from Figures

6.3-6.10 by the presence of dark regions. These Figures
clearly show that an increase in the weld pressure results

in an increase in the extent of orientation in the weld.
High magnification micrographs of this region are
presented in Figures 6.17-6.18 (SS1 and SS12). The effect
of increasing the pressure resulted in a higher degree of
orientation. This can be seen on comparing the Figures
6.3-6.8 which correspond to standard, low and high
pressure Eltex welds. The presence of molecular

orientation in welds implies the presence of internal
stresses resulting from the desire of the material to
return to its normal, unorientated state. Bukin [Bukin,
1972] also observed orientation in the weld zone by
viewing microtomed section between crossed polars. He

found that the changes in direction of orientation the
at
boundary of MAZ between the parent material and within the
MAZ was 600-700. This was caused by the movement of the
melt during joining. The microstructure of the weldline is
determined by the welding pressure. At low welding
pressures (SS12) the length of the MAZ is relatively large
Page 154
10µm
Figure 6.17 Microtomed section of SS1 Eltex weld, from

the bottom of the pipe showing the weld
line with MAZ from the central region. Mag.
100.
10µm

the bottom of the pipe showing the weld
line with MAZ from the central region. Mag.
100.
Page 155
and the material adjacent to the weldline is relatively

hot; hence the weldline melts and subsequent slower
cooling produces ralatively large spherulites. At higher
welding pressure (SS1), the MAZ region is thin and the
cool material adjacent to the weldline is near the melting

temperature. Rapid cooling produces a fine grained
structure of small spherulites in the weldline. Stevens

(Stevens, 1988] used Fourier Transform Infra Red
spectroscopy (FTIR) to analyze and quantify variations in
the weld crystallinity due to the thermal history of the
welding cycle. A change in crystallinity was observed
across the pipe walls themselves and within the weld
crystallinity is fairly constant, but lower in the weld

centreline.
With reference to the schematic illustration of the weld

zone shown in Figure 6.2 and the micrographs presented in
Figures 6.3-6.10 (SS1, SS12, SS5, RR1 and RR12), evidence
of flow-induced orientation can be clearly seen. The
relative dimension of the spherulite within the weld zone

were found to be similar to that observed for the parent
material. However, the relative dimensions along the weld
line were smaller than the surrounding regions.
Figure 6.19 shows a micrograph from the bore of Eltex weld

made under standard conditions (SS1), highlighting the
different structures present in the weld. The parent
material is in the left corner and a whitened region at
the top of the micrograph is separated from the parent
material by a boundary layer - shear zone. The MAZ is on
the right hand side and the bead, on the top of the
micrograph.
Page 156
Figure 6.20 illustrates the micrograph from the bore of

the Eltex weld made using dual pressure welding. By
comparing Figure 6.20 with Figure 6.19, the weld structure

differs quite significantly from that of single pressure
weld. Firstly, due to the larger heat soak time and low
secondary fusion stress, the width of the MAZ is
significantly wider in the dual pressure weld. Secondly,
the structure at the junction of the bead and the pipe

surfaces is quite different, where the notch position no
longer coincides with the MAZ limit line. During the time
under the secondary joining stress, the position of the
MAZ boundary changes due to the melt zone advancing

further along the pipe wall. The additional heat in the
weld bead, due to larger soak time is sufficient to extend

the edge of the MAZ to a new position, thereby removing it
from the site of the sharp notch initially produced
between the bead and the pipe surfaces. Thirdly, there is
little evidence of orientation at the edge of the MAZ as
any orientation introduced under the primary joining
stress is able to relax during the secondary fusion stage.

The results presented by Marshall et al. (Marshal12 et al,
1995] are in agreement with the illustration presented in
Figure 6.20.
The skin region in the bore of the pipe was subjected to a

detailed microscopic investigation. The elongated
spherulitic structure which was observed on the bore of

the pipe was also observed in the inner bead skin region.
The implication here is that a portion of the inner skin
does not fully melt during the welding operation. This has
been discussed in chapter five, section 5.3. In Figure
6.21, the elongated spherulites in the skin region of the
Page 157
10µm

the bore at the bottom of the pipe. Mag.
100.
5 00µm

the dual pressure welding. Mag. 25.
Page 158
10µm

the bore showing the inner bead. Mag. 100.
10µm
Figure 6.22 Microtomed section of SS1 Eltex weld,

showing the inner bead. Mag. 100.
Page 159
inner bead are clearly visible. The structure in the bulk
of the inner bead is shown in Figure 6.22. In the beads
the material is hot where much of the flow has occurred.
Here, the heat is lost radially and solidification of the
surface occurs resulting in a conventional spherulitic
microstructure.
Figure 6.23 and 6.24 illustrate the microstructure for the
Eltex SS1 weld. The micrographs were taken from the mid
region of the shear/MAZ region highlighting that there was
no gross change in the size of spherulites. In other
words, going away from the MAZ, the microstructure does
not change.
The results of Figures 6.16 to 6.22 as above, show that

the microstructure of the welded region consists of five
zones. The five different zones were also observed for the
Rigidex welds but the relative size of the spherulites
were much finer. However, the elongated spherulitic
structure at the bore region was not observed. Similar
zones have been observed in PE welds by Menges and Zohren

[Menges and Zohren, 1967]. They offered no comment on the
nature of the zones, except to say that they were the
result of temperature effects. The zones were also

identified by a number of researchers [Barber and
Atkinson, 1972; 1976; DeCourcy and Atkinson, 1977; Barton
and Cherry, 1980], who explained that the chromic acid

attacked the material at different rates on the basis of
different crystalline structures in the different zones.
Page 160
20µm

the central region. The shear zone is on
the bottom of the micrograph. Mag 250
20µm

the central region which is 2mm away from
the MAZ. Mag. 250.
Page 161
TENSILE TESTING OF MICROTOMED WELDED
SECTIONS
7.1 INTRODUCTION
This chapter describes the tensile testing of microtomed
thin films for the Eltex Tub 124 and Rigidex 002-50 pipe
materials. The deformation of the weld during the tensile
testing were monitored using crossed polarised light and a
video camera. Image analysis was used to analyse the video

images as a function of displacement. The information on
the displacement was obtained by measuring the dimensions
of the surface mounted transfer grids.
7.2 IMAGE ANALYSIS
The image from the video was transferred to the computer
and image analysis was carried out using Global Lab image
software to measure the deformation. After trial and
error, 2 individual grids were chosen to indicate the
extent of deformation occurring in the sample in the Y-
direction. Deformation measurements were taken using the
following procedure:
Page 162
(i) the sample was mapped from the inner side to the outer
side at 7 equally placed points across the specimen for
the weldline;
(ii) above the MAZ;
(iii) below the MAZ and
(iv) away from the MAZ.
A schematic illustration of the position within the weld
zone where the dimension of the grids were taken is shown
in Figure 7.1. Five individual measurements were taken for
the each square, as indicated in Figure 7.1.
Clamp
Surface
,a7 ODOR Ell]]

(a
.A N
10 1 123 14
N9 yý
41
H 15 17 18 19 20 20
2]7Q] 24 26 27 28
Bore
Clamp
Figure 7.1 Schematic illustration of the position

within the weld zone where the dimension of
the grids were taken.
Page 163
The results from the manual measurement were more accurate
than the results of the automatic method, which either
produced negative percentage strain values or values which

fluctuated. This was mainly due to the problem of the
computer being unable to identify particles or complete
particles due to the difference in colour contrast of some

of the pixels. Removal of the problem particles still
yielded negative strain. The manual measurement of the
deformation did not produce either negative or fluctuating
results.
7.3 MICRO-TENSILE TESTING OF ELTEX
7.3.1 BEADED SAMPLE
A sequence of macroscopic deformation modes observed for

the Eltex Tub 124 weld made under standard welding
conditions (SS1) are illustrated in Figures 7.2 to 7.6.
With reference to Figure 7.2, the presence of orientation
at the weld line and shear zones are readily apparent. The
asymmetry in the weld is due to the residual stresses

which are developed in the pipe during the extrusion
process. In Figure 7.2, the right hand corresponds to the
outer surface. This sample was obtained from the top of

the pipe. The effect of 4mm extention during the tensile
test is shown in Figure 7.3. The stress concentration at
the pseudo-notches are also clear. The effect of extending
the sample by further 2mm (Figure 7.4) clearly showed that
the maximum deformation was within the MAZ. For extention
above 6mm, the stress concentration at the pseudo-notches
was seen to increase and yielding of the material within
this region took place as shown in Figures 7.4 and 7.5,
followed by the tearing of the material in this region. On
Page 164
Chapter 7 Tensile Testing of Microtomed Welded Sections
5mm
Figure 7.2 Eltex SS1 sample at the start of the test

showing no extention.
5mm
Figure 7.3 Eltex SS1 sample with 4mm extention. Tested

at 5mm/min.
Page 165
5mm

at 5mm/min.
5mm

at 5mm/min.
Page 166
the commencement of yielding at the pseudo-notches, the
material within the MAZ relaxes. This is simply because
the applied load is being carried by the reduced cross-
section in the region of these notches. This is
illustrated in Figures 7.5 and 7.6 respectively.
5mm
Figure 7.6 Eltex SS1 sample with 18mm extention.

Tested at 5mm/min.
Quantification of these deformation for the SS1 weld were
obtained using the surface mounted grids as described in
the experimental procedure, section 3.5.1. Selected
micrographs of the standard weld (SS1) with the surface
mounted grids are shown in Figure 7.7 as a function the

of
time. The deformation of these grids has been analysed
using image analysis and the results are presented in
Page 167
Figures 7.8-7.11. With reference to Figure 7.7, after
approximately 150 seconds during the tensile test a
portion of the grids within the weld zone became
significantly distorted. This made it difficult to obtain
accurate measurements in a reproducible fashion. This is
clearly seen in Figure 7.7.
With reference to the coding system used in Figure 7.1,
Figure 7.8 illustrates the effect of increasing the time

during a tensile test on the observed % strain at seven
different positions within the weld zone. The local strain
varies at the seven position within the MAZ, the highest
being at the centre. This clearly demonstrates that the

deformation is greatest in the centre region. However,
further away from the weldline, the deformation was fairly
constant within the seven points, (Figure 7.11). Figures
7.9 to 7.11 illustrate the % strain for region
corresponding to above, below and away from the MAZ (see
Fig. 7.1 for coding). On increasing the time from t=100
seconds to t=150 seconds during the tensile test, the
strain within the weld zone was found to decrease. This
was attributed to the extensive yielding within the
pseudo-notch regions. The deformation within this zone was

found to be relatively constant. The presence of the large
shear zones in these regions causes a plane of local
weakness. The notch in this region from the outer bead
also acts as a stress concentrator which is responsible

for the final failure. Further increase in the extention
resulted in extensive deformation primarily at the outer
surfaces. Out of 10 samples inspected 6001 showed maximum

deformation for the outer surface. A possible reason for
Page 168
t=0 seconds
t=40 seconds
t=80 seconds
t=150 seconds
5mm
Figure 7.7 Surface mounted grids on the SS1 beaded

sample, shown during a tensile test as a
function of time. Tested at 5mm/min.
Page 169
250
a--- dt- 150 Secs
-- -6 t-100Secs
t- --ft- 80 Secs
p- - -V t- 60 Secs
t- 40 Secs
200 O---fl
o- - -© t- 20 Secs
+-- +t-0Secs
150r
ý-ý
'r 100
50
Eil -
0
12 3' 4567
Position across the Sanple
Figure 7.8 The effect of increasing the % applied

strain at seven different regions for the
Eltex (SS1) weld with bead for the weldline.
250
+- " -+ t- 150 Secs
A--- t- 100 Secs
ýt - eosecs
f---"t-60Secs
200 p- - -c t- 40 Secs
O---f] t- 20 Secs
0---o t-0 secs
150
b \ /
w
4J
rA
x 100
50
n
ý1 234567
Position across the Sample
Figure 7.9 The effect of increasing the % applied

Eltex (SS1) weld with beads above the
weldline.
page 170
250
+- -+t- 150 Secs
ý- dt- 100 Secs
I- - 11 t -80 secs
0-- --at- 60 Secs
p- - -s7 t- 40 Secs
200 p--- fl t- 20 Secs
0- - -o t-0 Secs
r. 150
ro
J)
d'' 100
./
50
A
1234567
Figure 7.10 The effect of increasing the t applied

Eltex (SS1) weld with beads for below the
weldline.
250
+- - -+ t- 150 Secs
-6 t- 100 Secs
-- -6 t-80Secs
f-- --*t- 60 Secs
t- 40 Secs
200 p- - -v
t- 20 Secs
D---C
0---0 t-0 Secs
150
aJ
N
100
SO
0
23456?
Position across the Saaple
Figure 7.11 The effect of increasing the % applied strain

at seven different regions for the Eltex
(SS1) weld with beads away from the weldline.
Page 171
this observation was the microtomed sample used had
previously been deformed due to residual stresses. On the
application of load, the sample straighten out with the
outer surface being stretched (put in tension).
Gedhe et al. [Gedhe et al., 1992] have carried out

deformation studies on polypropylene butt fused welds and
found that the stress distribution in the weld zone was
not uniform under load. The microstructure in the weld
region, especially the zone of deformed spherulites (shear
zone), caused "stress concentration" in the joint area.

They were able to identify a number of preferential
deformation sites in and at the vicinity of the butt
joint. The two resin systems showed different failure
modes. Maximum stress was found at the junction of the

bead and the bulk polymer. Without the beads, they found
the maximum stress to be within the MAZ. The data
presented by Gehde et al. [Gehde et al., 1992] is in
agreement with the illustration presented in Figures 7.8-

7.11.
However, on comparing the % strain parallel to the stress

direction for the four regions at t=150 seconds for the
centre, the % strain is greatest in the centre region for

the weldline (Figure 7.12). The maximum deformation was
seen to be in the outer and inner regions. This is shown
in Figure 7.13 and 7.14 for the outer and inner surfaces.
This is due to the extensive yielding which occurs as a
result of stress concentration from the beads.
Page 172
250
Tv Q
Eltex without
Rigidex with
Beads
Beads
O Eltex with Beads
v
200
c i5o e
m
100
50 0
0'
MAZ Weldline Below MAZ Away MAZ
Above
Positon alcM the Sanple
Figure 7.12 Graph of % strain at the centre of weld at 4

different position-across the weld for Eltex
and Rigidex samples at t=150 seconds.
250
p Eltex without Beads
p Rigidex with Beads
p Eltex with Beads
200
V4 153
AD 100
m
50
8
0'
Above MAZ Weldline Below MAZ Away MAZ
Position alag the Sanple
Figure 7.13 Graph of % strain at the outer surface of

weld at 4 different position across the weld
for Eltex and Rigidex samples at t=150
seconds.
Page 173
Chanter 7 Tensile Testing of Microtomed Welded Sections
250
V Eltex without Beads
O Rigidex with Beads
0 Eltex with Beads
X00
50
100
ý.
7
Q
50
7
e
0
Weldline Below MAZ Away MAZ
Above MAZ
Positicn along the Satple
Figure 7.14 Graph of ä strain at the inner surface of

weld at 4 different position across the
weld for Eltex and Rigidex samples at
t=150 seconds.
7.3.2"DEBEADED SAMPLE
To investigate the, effect of beads on the deformation
modes, the outer and inner beads were removed. Figures
7.15 to 7.19 illustrate the sequence of macroscopic

deformation modes observed for the Eltex tub 124 weld made
under standard welding conditions (SS1). The effect of 9mm
extension during the tensile test is shown in Figure 7.16.
The stress concentrations that were present from the beads
are absent. The effect of extending the sample by a
further 4mm is shown in Figure 8.17. This clearly showed
that the maximum deformation was within the MAZ. Figure
7.18 to 7.19 illustrates the failure of the microtomed
sample within the MAZ, initiating from the scratch which
was present within the sample.
Page 174
5mm
Figure 7.15 Eltex SS1 debeaded sample at the start of

the test.
5mm
Figure 7.16 Eltex debeaded sample with 9mm extention.

tested at 5mm/min.
Page 175
5mm

tested at 5mm/min.
5mm

tested at 5mm/min.
Page 176
Figure 7.20 shows the debeaded SS1 sample with the surface
mounted grids as a function of time during the tensile
test. Quantification of the observed deformation that
occurred at the four different regions across seven
positions for the debeaded Eltex tub 124 weld made under
standard welding conditions (SS1) was also obtained.
5mm
tested at 5mm/min.
Figures 7.21 to 7.24 shows the extent of deformation for
the four regions at seven positions for the weldline,
above the MAZ, below the MAZ and away from the MAZ. On
comparing Figure 7.21 with 7.8, it can be seen that the
extent of deformation in the MAZ for the debeaded sample

is greater. In the weldline the beads offer structural
support (reinforce) to these positions and hence lower

a
deformation (% strain) is achieved.
Page 177
t=0 seconds
t=40 seconds
t=80 seconds
t=150 seconds
5mm
Figure 7.20 Surface mounted grids on the SS1 debeaded

weld, shown during a tensile test as a
function of time. Tested at 5mm/min.
Page 178
250
200
A--6 t-100Secs
{--
--
+t-150sß
it-80secs
l---ft-60Secs
150 Q- - -4 t- 40 Secs
13---0t-20Secs
.f
O- - -O t-0 Secs
-N--
100
50
A
1234567
Position across the sairple

at seven different regions for the Eltex (SS1)
weld without the beads for the weldline.
250
cý---6 t-100Secs
-- -a t- 80 Secs
ý---+ t= 60 Secs
p- - -i7 t- 40 Secs
200 Q --- fl t =20Secs
o---0t=0Secs
150
100
50
'- - _a - _"ý
-a- _
0"
1 35 7
Position across the sanple

at seven different regions for the Eltex
(SS1) weld without the beads above the
weldline.
Page 179
Chapter 7 Toneile Teetinq of Microtomed welded gectiona
250
r- "+t- 150 Secs

L4- fi t- 100 Secs
-- Am t- 80 Secs
i
ý- --t- 60 Secs
i
200 p- - -v t- 40 Secs
D---£1 t- 20 Secs
o---ot-0 Secs
z 150
.11
ro
I' 100
+-- .-
50 -+-- _- _
-- - -+ ---,
_-- -+ - __--a-- --ems
- -a---=-- ---- -- --- -- ---
-- -U-
A
1234 5' 67
Position across the Sarple

weld without the beads below the weldline.
250
+-- +t-150sß
e--- t- 100 secs
-- "t-e0 secs
f----t t- 60 Secs
200 40 Secs
p- - -V t-
O---at-20 Seca
o- - -o t-0 secs
150
0
100
50
Z-
1 234567

weld without the beads away from the weldline.
Page 180
The extent of deformation for above and below the MAZ in
the debeaded sample was found to be relatively constant
but significantly lower than the sample with the bead.
This is due to the absence of beads which act as stress
concentrators. The stress concentrations from the beads
are responsible for the final failures in the beaded
samples
7.4 MICRO-TENSILE TESTING OF RIGIDEX
The macroscopic deformation for the Rigidex 005-20 weld
made under standard conditions (RR1) is illustrated in
Figure 7.25. The quantitative deformation that occurs in
the weld at four different regions across 7 points are
presented in Figure 7.26 to 7.29. The extent of

deformation in the centre region is clearly seen to be the
greatest. However, on comparing Figure 7.26 with Figure
7.8, the extent of deformation in the Rigidex sample is
greater than that observed for Eltex. This is obvious as

the two polymers are different, Eltex being a HDPE whereas
Rigidex is MDPE. The yield stress of the MDPE is lower
than the HDPE. On comparing Figures 7.27 and 7.28 with 7.9
and 7.10, the extent of deformation at the sides for above
and below the MAZ is somewhat similar to the Eltex weld.

This is the result of extensive yielding that occurs from
the pseudo-notches. Figure 7.29 illustrates the extent of

deformation for the position away from the MAZ. The
deformation in this region is considerably lower than that
observed for other regions. This is because when the
centre region of the weld elongates, the stresses on the
outer and inner regions increase, so that these regions
also elongate. Shear zones were found not to creep much.
Page 181
t=0 seconds
t=40 seconds
t=80 seconds
t=150 seconds
5mm
Figure 7.25 Surface mounted grids on the RR1 sample,

shown during a tensile test as a function
of time. Tested at 5mm/min.
Page 182
The level of orientation in the shear zone was found to
have little influence in the failure behaviour.
The deformation modes for the Rigidex debeaded film tested
at 5mm/min behaved in a similar manner to the Eltex film,
which were shown previously in Figures 7.15-7.19. The
extent of deformation in the centre of MAZ was greatest.
The final failure occurred in this region which indicates
that the material in the centre of MAZ deforms more
readily.
7.5 EFFECT OF TESTING SPEED
Increasing the testing speed from 5mm/min to 50mm/min and

500mm/min for the debeaded Eltex and Rigidex welds
resulted in the failures taking place predominantly within
the centre of the weldline. The final failure modes were
believed to be influenced by a number of factors including
the following.
(i) scratch of test specimen;

(ii) Poission contraction and
(iii) geometry of sample.
Typical examples of failure mode for Eltex debeaded sample

tested at 5,50 and 500mm/min as a function of time are
presented in Figures 7.30 to 7.32.
Figure 7.30 illustrates a micrograph of the Eltex thin

film sample without the bead and tested at 5mm/min. The
fracture occurs near the inner surface within the MAZ.
Figures 7.31-32 are micrographs of the Eltex thin film
Page 183
250
A -- 't- 150 secs
__ 11t-100Secs
80 Secs
*---o t- {
p- - -p t- 40 Secs
O --- Ct- 20 Secs
1100
o---of-0Secs
r. 150
ro
Cl)
- 100
/ --- U,
__ ---i---
-.
50
n
1234567

at seven different regions for the Rigidex
(RR1) weld with beads for the weldline.
250
p- t1 t- 150 Seca
o-- --"t- 80 Secs
p- - -q t- 40 Secs
ß--- F] t- 20 Secs
200 o- - -0 t-0 Secs
150
gd
14
N
x 100
-
-'
50
3-"-"----9--------9"_
A
1234567
Pasiticn across the Sanple

(RR1) weld with beads above the weldline.
Page 184
250
1100
a 150
.,,
it
41
rn
'~ 100
50
1357

(RR1) weld with beads below the weldline.
250
6- dt- 150 Secs

--it- 100 Secs
h--f t- 80 Secs
O --- fl t- 20 Secs
200
o----0t-0Secs
150
u
r
-
100
50
- --" -i --- -mai ý-
ý=='-ý
-- _
0
135
Position across the Sale

(RR1) weld with beads away from the weldline.
Page 185
samples tested at 50 and 500 mm/min. The deformation is
seen to be greater in the centre region (MAZ) which
results in the fracture for films tested at 50 mm/min. For
films tested at 500 mm/min, fracture initially occurs
within the centre region of MAZ. Gehde et al. [Gehde et
al., 1992] found the maximum stress to be within the MAZ.
The data presented by Gehde et al. is in agreement with
the illustration presented in Figures 7.30-7.32.
7.6 THIN FILMS - RIGIDEX
A limited study of the effects of joining temperature,
heat soak and welding pressure was carried out on the
Rigidex 002-50 microtomed welds by varying the standard
conditions as given in Table 3.3. The difference in the
condition were as follows:
RR1: standard condition;
R2: temperature 220°C;
R3: temperature 240°C;
R4: heat soak 210 seconds,
R5: heat soak 70 seconds,
R6: welding pressure 0.11MPa,
R7: welding pressure 0.41MPa.
Appendix 5 gives the extent of deformation resulting in
the seven positions across the four regions for the
Rigidex welds for each of the six samples tested. The
extent of deformation in the welded samples was obtained

by studying the video evidence as the final failures modes
were very similar to those discussed previously. However,
Page 186
t=0 seconds
t=100 seconds
ý.
5
ý..
t=150 seconds
t=165 seconds
5mm
Figure 7.30 Illustration of the fracture in Eltex

debeaded film tested at 5mm/min.
Page 187
Tensile Testing o Microtomed Welded Sections
Chapter 7
t=0 seconds
t=9 seconds
t=16 seconds
t=2 se cc: 1
5mm
Illustration of the fracture in Eltex

Figure 7.31
Page 188
t=0 seconds
t=0.5 seconds
t=0.7 seconds
t=1 seconds
5mm
Figure 7.32 Illustration of the fracture in Eltex

Page 189
Table 7.1 lists the summary of the results obtained from
Appendix 5, which illustrates the generalised area or
deformation in the 4 regions across the thin film and thus
the generalised place of deformation for the weld, MAZ and
for the set of points away from the MAZ.
The effect of changes in the joining temperature, heat-
soak and joining pressure on the deformation modes is
given in Table 7.1. Within the limits of the temperature
and pressure variations used in this study, an increase in
the joining temperature and pressure had relatively little
effect on the deformation modes. Even the change in the
heat-soak time had little influence in the failure

behaviour.
DEFORMATION
POINTS 1-7 POINTS 8-14 POINTS 15-21 POINTS 22-28
RR1 centre centre centre sides
R2 centre centre centre sides/across
R3 centre centre centre sides
R4 centre centre centre sides/across
R5 centre/side centre centre sides
centre: deformation in the middle;

centre/sides: deformation across in the middle or at sides;
side: deformation occurs at the side;
side/across: deformation occurs at the sides or across the
set of points
Table 7.1 General area of deformation across the set

of data points.
Page 190
7.7 FINAL FAILURE MODES FOR WELDED FILMS
From the video evidence, it can be seen that three types
of fracture paths occur for the beaded samples of Eltex
and Rigidex in relation to the MAZ. This is schematically
shown in Figure 7.33. In the type I, the fracture
initiated from the region between the outer weld bead and
the inner weld bead. The fracture then propagates outside
the MAZ up to half or greater than half the wall thickness
after which it continues to propagate through the MAZ
until final fracture through the weld beads occurs. In
type II, the fracture initiates from the same location as

in type I, but propagates completely outside the MAZ to
the point of final fracture underneath the outside of the

inner weld bead which is on the same side as the outer
bead. In the type III, the fracture initiates again from
the same location as type I but propagates across the MAZ
and the failure is underneath the outside/inside weld bead
which is on the opposite side of the inside/outside weld

bead where initiation took place.
Type I and II fractures were found to be the most common
modes by which the overwhelming number of Eltex and
Rigidex samples failed while testing at 50 and 500mm/min.
Figures 7.2-7.6 illustrate type II fracture in SS1 films
tested at 5mm/min. Type III failures were associated with
SS12 samples (low welding pressure). SS1 and RR1 samples
tested at the slowest speed failed by the type I and type
II manners. Figures 7.34-36 illustrate the type III
failure in SS12 films tested at 5mm/min.
Page 191
Bore Surface
Type I
Type II
Type III
Figure 7.33 Schematic illustration of the three

different types of crack propagation path
by which films failed.
Page 192
Cater 7 Tensile Testing of Microtomed Welded Sections
5mm
Figure 7.34 Eltex SS12 sample at the start of the test

showing no extention.
5mm
Figure 7.35 Eltex SS12 sample with 9mm extention,

showing type III fracture. Tested at
5mm/min.
Page 193
5mm
Figure 7.36 Eltex SS12 sample with 16mm extention,

showing type III failure. Tested at
5mm/min.
7.8 SUMMARY
Micro-testing of butt fusion welds enable a study of the
deformation and failure behaviour of different regions of
welds. The extension in microtomed weld specimens can be
measured to determine the stress-strain behaviour of the
weld. In this way, it is now possible to quantify the
effects of the welding parameters on weld performance. The
stress-strain behaviour was found to same as that for the
macro-tensile specimens.
Page 194
Chapter 8 Tensile Properties
TENSILE PROPERTIES
8.1 INTRODUCTION
This chapter is concerned with the macroscopic tensile
test results and failure modes which were observed for the
butt fusion PE welds. As in the previous chapter, the
influence of welding conditions were investigated as a

function of position around the circumference of the pipe,
heat soak and testing speed. The influence of the bead on
the failure mode was also investigated.
8.2 PRELIMINARY TENSILE TESTS
A preliminary set of experiments were undertaken to select
an appropriate test specimen geometry for the welded and
unwelded pipe sections. This was necessary because the WRc
test specimen resulted in yielding at the pins and
therefore it was difficult to measure elongation
accurately. Table 8.1 shows the average values of seven
test specimens for the peak stress and elongation to
failure of Eltex pipe. The test specimen had gauge widths
of 15,10,8.5 and 5mm with a radius of curvature for the
profiles was 37mm. These results are presented in Figures
8.1 and 8.2. The peak stress values for specimens with a
gauge width of 15mm was observed to be lower than that for
Page 195
GUAGE PEAR STRESS ELONGATION AT FAILURE

(MPa) (%)
WIDTH
(mm)
AVERAGE RANGE AVERAGE RANGE

15 24.6 (0.9) 23.8-24.4 198.3 (104.2) 98.9-290.3
10 26.4 (0.4) 25.7-26.8 77.2 (14.4) 57.4-102.0
8.5 26.5 (0.2) 26.2-26.7 81.8 (23.3) 60.4-106.56
5.0 27.6 (1.1) 25.9-28.8 48.0 (11.6) 33.0-65.4
Table 8.1 Summary of the tensile results for Eltex

unwelded profiled pipe strips having 4
different gauge widths. The samples were
tested at 10 mm/min.
30 -
2B
26
24
22
2
p5 --
10 15 20
cage Wdth (mm)
Figure 8.1 Graph of peak stress versus sample gauge

width for Eltex unwelded pipe strips.
Tested at 10mm/mm.
Page 196
a
S
05 10 15 20
Gage Width (mm)
Figure 8.2 Graph of elongation to failure versus

gauge width for Eltex unwelded pipe
strips. Tested at 10mm/min
the 10,8.5 and 5mm specimens. This can be attributed to
the necking that occurred. All the specimens with gauge
widths of 15mm had elongation at failure in the region of
200%. The observed failure mode was the formation of a
neck which then propagates through the weld, drawing
material from the edge of the shoulder of the parent
material. Only 2 of the samples actually fractured. It was
therefore concluded that profiled specimens having a gauge
width of 15mm were not suitable for tensile test. Table
8.2 lists the peak stress and percentage elongation to
failure values for the Eltex welds made at 230 °C using a
gauge width of 10 and 5mm respectively. The extension to
failure for samples with a gauge width of 10 and 5mm were
around 40 and 50 percent of the values obtained for the
unwelded pipe.
Page 197
GUAGE PEAK SRESS ELONGATION AT FAILURE

WIDTH (MPa) (%)
(_m)
10 27.1 (0.4) 26.6-27.7 30.3 (1.6) 27.8-31.9
5.0 29.2 (0.4) 28.1-29.5 23.8 (2.8) 19.4-27.4

weld made in accordance with Table 3.1 at
230°C, using 2 different gauge widths.
Tested at 10 mm/min.
Both the Eltex and Rigidex materials were studied in the
same way. Tensile tests were first carried out on unwelded
specimens and typical values of peak stress and percentage
elongation at fracture were recorded for each test
specimen.
8.3 EFFECT OF MACHINING MARKS ON TENSILE TESTING
The test geometry and the quality of the test specimen can
also have an influence on the tensile test fracture
surface. This coupled with the residual fabrication
stresses on the surface and bore of the extruded pipes

have an effect on the tensile properties. Therefore, any
induced gouges or nicks on the surface or bore of the pipe
may constitute a potential site of weakness for the
pipeline.
The effects of machining marks on the Eltex and Rigidex
samples were removed by polishing and re-melting the notch
area. The polishing was carried out using 1200 grade emery
paper. A 35mm diameter bar was heated to a temperature of
Page 198
150°C and was used to melt the aspirates from the notching
operation. Table 8.3 illustrates the results obtained
which show no improvement in the elongation to failure
values by the polishing and re-melting operations.
However, when the outer and inner skins were machined off
a significant improvement in the elongation to failure
values was seen for both the Eltex and Rigidex materials.
This could be due to the removal of macroscopic defects
leading to yielding which occurs at lower stress. As there

is a gradient property across the pipe wall, the outer
skin has lower mechanical (tensile) properties (see Table
8.3) .
PEAK STRESS ELONGATION AT FAILURE

(MPa) ($)

ELTEX 27.1 (0.5) 26.2-27.8 128.8 (23.9) 71.5-157.2
UNWELDED
POLISHED 27.2 (0.1) 27.2-27.2 132.5 (2.3) 130.9-134.1
MELTED 26.1 (0.1) 25.9-26.2 139.4 (15.0) 118.9-154.8
SKIN-OFF 27.0 (0.1) 26.8-27.1 157.0 (22.8) 137.0-187.3
RIGIDEX 20.5(0.4) 19.9-21.0 197.7 (49.9) 130.5-256.2
UNWELDED
POLISHED 20.0 (0.5) 19.9-20.4 202.1 (98.0) 114.8-308.8
MELTED 20.0 (0.6) 19.6-20.7 201.0 (69.7) 151.7-250.3
SKIN-OFF 20.2 (0.4) 19.8-20.8 243.3 (33.2) 188.7-268.9

and Rigidex unwelded pipe strips of gauge
width 10mm. Tested at 50mm/min.
Page 199
8.3.1 EFFECT OF DEFECTS ON TENSILE TESTING

tN`ý
The presence of defects on the notched Eltex unwelded
samples were investigated by using a drill of diameter
0.5mm. The samples were drilled to a depth of 1,2,3 and
4mm from each end to give a total defect length of 2,4,
6 and 8mm. The results are presented in Table 8.4 which
shows a dramatic change in the elongation at failure eg.

for Eltex unwelded it reduces from 128.8% to 45.1% for a
defect of 2mm.
DEFECT PEAK STRESS ELONGATION AT FAILURE

SIZE (MPa) (%)
2mm 27.3 (0.3) 26.2-27.8 45.1 (5.2) 41.1-49.5
4mm 26.8 (0.2) 26.4-27.1 45.9 (4.1) 43.3-48.8
6mm 27.3 (0.2) 26.8-27.6 27,9 (7.1) 26.3-29.6
8mm 27.1 (0.1) 26.3-27.5 27.4 (3.2) 25.4-30.9

unwelded pipe strips having various length
of defects, defect diameter 0.5mm. Tested
at 50mm/min.
8.3.2 EFFECT OF NOTCHING (SKEWING) ON TENSILE TESTING
The difficulty of making consistent notches, particularly

with regards to the symmetry can also have an influence on
the tensile result. A preliminary set of experiments were
carried out to establish the effect of shifting the notch
by 2,4 and 8mm from one side. The results illustrated in
Table 8.5 shows that similar elongation to failure values
Page 200
were obtained to the original samples highlighting that
skewing had no major effect.
SKEWED PEAR STRESS ELONGATION AT FAILURE

(MPa) (%)

2mm 27.1 (0.4) 26.6-27.4 124.5 (5.3) 118.3-131.3
4mm 26.6 (0.3) 26.1-27.2 135.1 (7.2) 126.4-141.6
8mm 26.4 (0.1) 26.3-27.2 134.9 (6.7) 128.3-142.8

unwelded pipe strips. The notch was
displaced from one side by 2,4 and 8mm
respectively. Tested at 50mm/min.
8.4 INITIAL WELDING OF ELTEX AND RIGIDEX
Table 8.6 shows the average values of peak stress and

percentage extension to failure obtained for welds made
using different welding temperatures and pressures. The
results showed that the Eltex and Rigidex specimens had a

lower elongation to fracture value than the parent
materials. The percentage reductions for the Rigidex
specimens was 30% for welds corresponding to the standard

welding conditions. 60% reduction was observed for the
Eltex samples made under standard welding conditions. The
extrusion temperature of Eltex is higher then the
extrusion temperature of Rigidex. This coupled with the

lower thermal conductivity of Eltex as compared to the
Rigidex results in slower cooling. Examination of
microtomed sections has shown that Eltex has a larger
distribution of spherulite sizes which results from slower
Page 201
PEAR STRESS ELONGATION AT FAILURE

(MPa) (%)

ELTEX 26.4 (0.4) 25.7-26.9 77.2 (14.4) 57.4-101.9
UNWELDED
S220 28.0 (0.4) 27.0-29.1 26.0 (4.7) 19.5-31.9
S230 27.1 (0.4) 26.6-27.7 30.3 (1.6) 27.8-31.8
S230A 28.3 (0.4) 28.1-28.8 37.5 (8.5) 24.3-52.2

.
S230B 28.3 (0.3) 28.0-28.7 32.4 (4.1) 24.4-38.0
S240 27.4 (0.5) 26.9-27.9 28.1 (4.2) 22.2-34.1
RIGIDEX 21.2 (1.9) 19.4-23.2 76.4 (20.1) 62.8-105.8
UNWELDED
R220 20.7 (0.2) 20.3-21.0 49.4 (5.3) 43.2-56.9
R230 22.2 (2.0) 20.8-24.8 53.3 (8.1) 42.9-63.6
R230A 21.6 (1.5) 20.6-22.5 52.1 (10.8) 43.1-70.2

R230B 21.0 (1.2) 20.8-21.4 46.8 (6.3) 39.6-55.6
R240 -
20. 3) 20.5-21.2 49.6 (3.4)
-9(o . 40.3-55.8
S Eltex Solvay
R BP Rigidex

and Rigidex welds tested at 10 mm/min.
cooling (see Figure 6.13, section 6.4.1, page 151). The

fracture of Eltex follows a characteristic pattern, in
that a pin hole appears in the weld line, which to

grows a
circular hole until fracture occurs at the shoulder at one
end of the gauge length. This is pictorially illustrated
in Figure 8.12. The results are graphically presented in
Figures 8.3 and 8.4 and show that acceptable welds may be
Page 202
produced in the temperature range 220-240°C with the
optimum welding temperature being about 230°C for both
Eltex and Rigidex. The peak stress values obtained at the
various welding temperatures varied by less than 5% but
the elongation to failure varied by 20%. However, the

Eltex samples which were welded under the condition 230A
(Table 3.3) varied by 38% of the average values quoted.
Even so these elongation to failure values were
sufficiently reproducible to indicate trends by the
variation in the welding temperature. For welds made at

230°C using long soak times and high pressure and 240°C
under standard conditions, no adverse affects were
noticed. However, as expected the weld beads were

relatively larger.
Vanspeybroeck [Vanspeybroeck, 1995] studied more than 100

different fusion conditions using 7 different PE resins of
MFI 0.3 to 1.2 and pipe diameters from 110 to 315mm. He
found that welds with large beads had shorter failure
times than welds with small beads. He also found that the
large beads do not always fail in the fusion zone. He also
observed that welds with small beads generally failed in
the middle of the fusion zone. The difference between
small and large beads is of course a function the

of
joining pressure and the MFI of the PE material itself.
Vanspeybroeck [Vanspeybroeck, 1995] carried out finite
element calculations on different geometeries of the bead
and found that the bead geometry influences the stress

distribution in the bead area. The large beads were found
to have a higher stress concentration which failed
prematurely giving a low elongation to failure values.
Page 203
a,
T'
28,26
24
n- Q Rigfdex
0 Btex
221
210 220 230 240' 250
Welcin9 tarperahre (OC)
Graph of peak stress for Eltex and Rigidex

Figure 8.3
versus welding temperature; pressure
constant at 0.15MPa. Tested at 10mm/mm.
60
50
ö
aý
5
.ý
40
s
oF gide'c
o Btex
I 30
2D
210 --- --- --
14 --
VVddngTerPeratuern
Graph of elongation to failure for Eltex and

Figure 8.4
Rigidex versus welding temperature; welding
pressure constant at 0.15MPa. Tested at
10mm/min.
Page 204
8.5 ELTEX PE WELDS
A summary of the tensile results for the beaded and

debeaded Eltex specimens are presented in Tables 8.7 and
8.8. The results for the beaded specimens are also
presented graphically in Figures 8.5 and 8.6. Variations

in the welding pressure had a minimal effect on the
tensile strengths of the joints. Typical tensile traces
for the pipe strips and the welds (SS1, RR1 and SR1) with
and with out the beads are presented in Appendix G.

(MPa) (%)

SS1 28.3 (0.6) 27.7-29.0 32.1 (1.3) 31.0-33.3
SS2 28.3 (0.4) 27.8-28.8 31.6 (1.2) 30.2-32.7
SS3 29.3 (0.7) 28.7-30.3 34.7 (5.9) 27.8-41.7
SS4 29.3 (0.6) 29.5-38.9 41.8 (1.0) 40.7-43.1
SS5 29.3 (0.1) 29.3-29.4 36.0 (4.8) 31.5-42.1
SS6 29.7 (0.4) .29.4-30.3 29.0 (5.1) 24.9-36.3
SS7 27.9 (1.1) 26.9-29.5 16.1 (4.1) 12.5-20.7
SS8 30.8 (0.6) 29.9-31.3 36.2 (2.5) 33.3-39.2
SS9 29.6 (0.4) 29.0-29.9 29.2 (0.9) 28.2-29.8

SS10 28.6 (0.5) 28.4-29.4 29.2 (1.7) 27.0-31.0
SS11 30.2 (1.1) 29.2-31.3 35.8 (6.7) 30.9-45.6
SS12 27.4 (2.2) 25.4-30.5 19.7 (6.6) 9.9-24.3
SS13 30.2 (0.7) 29.6-31.2 30.4 (1.2) 29.1-31.7
SS14 29.9 (0.6) 29.3-30.6 28.2 (2.0)' 25.4-29.5
Table 8.7 Summary of the tensile results for the

Eltex beaded welds tested at 50 mm/min.
Page 205
The spread of results from the debeaded specimens was
again small, a slight decrease in the tensile strength was
seen as the joining pressure was increased. The average
maximum tensile strength achieved with these debeaded
samples was 28 MPa, which was actually smaller than the

beaded samples (30 MPa) but similar to that of the parent
material (26 MPa). Figure 8.6 shows that good results are
achieved over a pressure range of 0.02 to 0.21 MPa. Table
8.8 shows that elongation to failure values had reached a
maximum at 0.09 MPa joining pressure for the debeaded
specimens (SS14). With the beaded specimens, the
elongation to failure was found not to be influenced by
the welding pressure. A possible reason for the this could

be due to the fact the butt fusion machine which was used
was pressure controlled. In other words, for a specified
heat-soak time, the amount of material in the molten state
will be the same. Therefore, the amount of the extruded
material in the bead will first increase for low pressures
and then level off for the higher pressures. Figures 8.7
to 8.10 give the peak stress and elongation to failure
values for welds made under varying heat soak times at

pressures of 0.15 and 0.02 MPa and a temperature of 230
oC
Variations in heat soak time had a minimal the

effect on
tensile strength of the joints at different pressures
although the elongation to failure values were for
greater
the debeaded samples at 0.15 MPa joining pressure as
compared with the elongation to failure values at an
applied pressure of 0.02 MPa.
Page 206

(MPa) (%)

DSS1 27.6 (0.3) 27.4-27.7 35.5 (5.1) 31.90-39.1
DSS2 27.0 (0.3) 26.6-27.3 49.5 (25.6) 23.9-75.1
DSS3 27.4 (0.2) 27.2-27.5 28.7 (6.8) 21.8-35.5
DSS4 27.3 (0.1) 27.2-27.4 25.6 (0.8) 24.8-26.4
DSS5 27.3 (0.4) 26.9-29.4 23.6 (1.6) 22.5-26.7
DSS6 27.4 (0.5) 26.8-27.9 22.5 (2.3) 18.7-24.8
DSS7 27.0 (0.6) 26.4-27.9 24.2 (3.7) 18.7-30.3

DSS8 27.5 (0.2) 27.1-27.6 31.4 (4.8) 26.3-38.4
DSS9 27.3 (1.0) 25.4-28.1 38.9 (25.2) 26.1-90.1
DSS10 27.7 (0.2) 27.3-27.9 90.2 (56.1) 29.9-160.5
DSS11 27.2 (0.3) 26.7-27.6 54.9 (51.9) 28.7-160.6
DSS12 27.2 (0.5) 26.4-27.7 23.9 (4.5) 18.0-29.0
DSS13 27.3 (0.3) 26.8-27.7 31.1 (6.7) 24.5-43.8

DSS14 27.5 (0.3) 27.0-27.8 85.5 (52.3) 30.1-157.7

Eltex debeaded welds tested at 50 mm/min.
With low joining pressures, fracture occurred in the weld

line. The test results from the debeaded specimens
highlight the fact that specimen geometry plays a major
role when evaluating welded joints. All of the welds
tested with the bead retained showed elongations to
failure to be much lower than that for the parent material
whereas the debeaded tensile specimen with a pressure of
0.09 MPa gave elongation to failure values equivalent to
the parent material. Surprisingly, in some cases welds
with beads from "bad" fusion conditions (SS2, SS13, SS14,
Page 207
D
30
2B
,L
0
welting Pr re (MPa)
Figure 8.5 Graph of peak stress versus welding

pressure for Eltex; welding temperature
constant at 230 °C. Tested 50mm/mm.
at
60
Q Debeaded
o Beaded
50
LL
40
.9
I .
30
0
20
0 a2 as 0.6 0.8
WeldingPressire(Wa)

welding pressure for Eltex; welding
temperature constant at 230°C. Tested at
50mm/min.
Page 208
o DeUxKW
o Beaded
2,9
28
Z7
100 200 300 400 500
HeatSoak(S)
Figure 8.7 Graph of peak stress versus heat soak time

for Eltex at the welding pressure of
0.15MPa. Tested at 50mm/min.
100
O Deeded
O Beetled
80
ö
IJ.
60
O
0
40
20
100 200 300 400 500
HeatSoak(S)
Figure 8.8 Graph of elongation to failure versus heat

soak time for Eltex at the welding
pressure of 0.15MPa. Tested at 50mm/min.
Page 209
30
Q Debeaded
o Beaded
29
28
27
26
100 200 300 400 500
HeatSo* (S)

for Eltex at the welding pressure of
100 200 300 400 500
HeatSoak(S)
Figure 8.10 Graph of elongation to failure versus heat

soak time for Eltex at the welding
pressure of 0.02MPa. Tested at 50mm/min.
Page 210
RR2, RR13 and RR14 - Tables 8.7 and 8.9) performed better
than welds from "good" fusion conditions. This was due to
the fact that small filaments of the yielded polymer
continued to be drawn from the surface region after the
central section of the weld had failed in some of the
welds. This coupled with the fact that in the beaded
specimens with the weld bead retained, fracture initiated
from the notch below the bead,. indicates the importance of

the notching effect. In the beaded samples, the failure
was in all cases initiated by the notch formed the

at
junction of the weld bead and the pipe wall. Our recent
data suggests that the actual failure tended to run in one
of two directions, either straight across the sample
through the parent material to the notch below the
opposite weld bead or it ran diagonally through the weld,
through the heat affected zone to meet the notch the
of
bead on the other side of the weld. This is shown in
Figure 8.11. For the debeaded samples, in the majority of
the cases, a pin hole appeared in the weld line, which
grows to a circular hole until fracture occurred. This is
illustrated in Figure 8.12.
The test results from low interface fusion stress between

0.02 to 0.09 MPa (Tables 8.7 and 8.8, welds SS2, SS13 and
debeaded) [increased
SS14 beaded and confirm that heat soak
time coupled with reduced interface fusion stress are
beneficial in promoting the required ductility in the
tensile weld test. This is due to the fact that melt is
retained in the weld zone which has a uniform spherulitic

structure similar to that of adjacent, non joint material.
Similar observations have been made by Marshall et al.,
[Marshall2 et al., 19951.
Page 211
.ý
ý:
t=0 seconds
t=12 seconds
r=2
r=27 ýý TP. QS
5mm
Figure 8.11 Illustration of the failure of Eltex welded

sample with beads. Tested at 50mm/min.
Page 212
t=0 seconds
v3 r.
ý1
t=2 seconds
t=2.5 seconds
t=3 seconds
5mm
Figure 8.12 Illustration of the failure of Eltex welded

sample without beads. Tested at 50mm/min.
temperature constant at 2300C. Tested at
50mm/min.
Page 213
8.6 RIGIDEX PE WELDS
The results of welds made over a range of pressures and
without beads are presented in Tables 8.9 and 8.10. The
results are graphically presented in Figures 8.13 and 8.14
and show that variations in welding pressure have a

minimal effect on the tensile strength of the joint
produced. The spread of the results from the beaded
samples is around 2 MPa at any one pressure value whereas

the spread from the debeaded samples is again small to
within 1 MPa. Figure 8.14 shows that acceptable welds are

produced in the pressure range 0.09 to 0.63 MPa, with the
optimum welding pressure being 0.21 MPa. The debeaded
samples RR3 to RR6 gave elongation to failure values

similar to that obtained for the parent material. However,
a number of samples did not break and the test was stopped
for the RR3 to RR6 samples. For samples tested with the
welding bead left on, the elongation to failure values
were smaller. Necking occurred away from the weld, the
failure was initiated by the notch between the bead and
the gauge length. The final fracture occured in this
region. Figure 8.15 illustrates the pseudo-notch failure
of the Rigidex weld made under standard welding conditions

(RR1). Figure 8.16 and 8.17 illustrate the affect of
varying heat soak time with pressure of 0.02 MPa and a
welding temperature of 230°C. It can be seen that the
debeaded samples gave the highest elongation to failure
values at a heat soak time of 380 seconds. This is
somewhat surprising as only a low pressure was used.
Page 214

(MPa) (%)
RR1 22.1 (0.4) 21.5-22.4 52.7 (6.7) 44.9-60.5
RR2 22.5 (0.1) 22.5-22.7 37.1 (7.4) 28.1-45.9
RR3 20.9 (0.3) 20.6-21.3 48.8 (7.1) 41.3-61.1
RR4 20.8 (0.2) 20.5-21.0 55.0 (8.0) 44.4-67.4
RR5 20.5 (0.3) 20.2-20.9 51.1 (1.8) 39.6-54.0
RR6 20.6 (0.7) 20.2-22.0 60.9 (12.4) 48.4-78.4
RR7 22.5 (0.3) 22.3-22.9 41.5 (1.6) 39.8-43.0
RR8 23.2 (0.4) 22.7-23.5 41.3 (1.8) 39.6-43.8
RR10 22.8 (0.2) 22.6-23.0 43.5 (2.1) 41.4-46.0
RR12 19.4 (0.2) 18.3-21.0 15.8 (5.0) 06.7-16.6
RR13 22.6 (0.2) 22.4-22.8 38.7 (1.5) 37.6-40.8
RR14 22.3 (0.4) 21.7-22.6 39.8 (2.3) 37.8-42.2

Rigidex beaded welds at 50mm/min.
8.7 MIXED WELDS BETWWEN ELTEX and RIGIDEX
Welds by adopting the conditions from Table 3.4

were made
over a range of pressures. Tensile tests were then carried
out in the usual way. Table 8.11 gives the peak stress and
to failure for both the beaded and debeaded

elongation
samples. Figures 8.18 and 8.19 shows the effect of
increasing the welding pressure on the peak stress and the
elongation to failure. The elongation to failure was seen
to reach a maximum at a pressure of 0.31 MPa, but only

falls slightly even when the welding pressure was 0.15 MPa
for beaded samples. When the welding pressure was 0.02
MPa, contraction cavities were seen in the fracture
surfaces indicating poor adhesion. For debeaded samples
Page 215
the maximum elongation to failure value is achieved at a
pressure of 0.21 MPa. For beaded and debeaded samples,
yielding always occurs in the Rigidex half of the
specimen. This was expected as the yield stress for
Rigidex PE is 20 MPa compared with 27 MPa for the Eltex
material. Two types of fracture were observed for the
specimens tested with the weld bead removed. Low welding
pressure resulted in the fracture occurring at the weld.
For welds made under or near optimum conditions

(industrial conditions), displacement of more than 65 mm
were obtained and the neck again initiated in the Rigidex
samples and ran through the weld. Fracture finally
occurred either at the end of the gauge length in the

Rigidex test specimens or at the weld interface. These
fractures were similar to those shown in Figures 8.12 and
8.15 for debeaded and beaded samples.
8.8 DUAL PRESSURE WELDS
Using the standard joining profiles as shown in Figure

2.5, page 23, the HDPE pipe welds with wall thickness
greater than 20mm gave a brittle type of response when

tested using the sample shape specified in Figure 2.7,
page 46. To overcome the brittle response the "dual
pressure welding" profile has been developed. This extends

the welding time whereby high interface pressures have
been retained only for the bead-up and the short term melt
mixing phases of the process.
A limited amount of dual pressure welds for Eltex and

Rigidex were made. Table 8.12 shows the results for
welds produced by the dual pressure process. In all cases
Page 216

(MPa) (%)
DRR1 20.7 (0.8) 19.5-21.3 87.4 (40.8) 43.6-142.2
DRR2 21.0 (0.3) 20.7-21.3 46.4 (5.9) 39.2-53.6
DRR3 21.3 (0.7) 20.4-22.3 264.5 (19.4) 231.. 5-278.5
DRR4 21.6 (0.8) 20.7-22.9 232.4 (72.5) 130.5-291.7
DRR5 21.1 (0.8) 20.1-23.7 248.5 (33.4) 190.2-290.2
DRR6 21.2 (0.7) 20.6-22.6 240.0 (43.1) 166.2-281.0
DRR7 21.3 (0.3) 21.0-21.7 69.7 (46.3) 42.4-163.7
DRR8 21.4 (0.2) 21.2-21.6 140.8 (74.5) 60.8-227.3
DRR10 21.3 (0.4) 20.7-21.9 209.3 (47.6) 118.8-251.0
DRR12 20.8 (0.2) 20.5-21.1 89.9 (39.0) 57.2-160.9
DRR13 21.3 (0.4) 21.8-21.8 50.8 (4.1) 44.9-54.6
DRR14 21.0 (0.4) 20.5-21.6 154.8 (88.1) 43.0-246.1

Rigidex debeaded welds tested at 50mm/min
D
23
I I¢ýý
22
21
20
0 0.2 a4 0.6 0.8
Welcing Pressire (PvPa)

pressure for Rigidex; welding temperature
constant at 230°C. Tested at 50mm/min.
Page 217
v
ê
c
1i
2
0 Q2 Q4 0.6 0.8
waking Pressure
(N9'a)

welding pressure for Rigidex; welding

(MPa) ($)

SRl 23.6 (0.3) 23.2-24.0 39.0 (1.2) 37.2-40.3
SR2 22.1 (0.5) 21.2-22.8 30.2 (2.2) 26.7-33.3
SR3 23.1 (0.4) 22.6-23.6 47.1 (1.8) 44.6-49.5
SR4 22.7 (0.3) 22.4-23.1 51.7 (6.7) 47.0-64.7
DSR1 22.9 (0.3) 22.5-23.4 73.2 (28.5) 46.3-132.7
DSR2 22.0 (0.4) 21.5-22.5 65.0 (25.1) 35.4-109.2
DSR3 23.0 (0.3) 22.4-23.4 97.8 (31.5) 73.2-143.4
DSR4 22.7 (0.3) 22.3-23.1 73.6 (7.1) 66.1-82.6
SR MIXED WELDS
D DEBEADED
Table 8.11 Summary of the tensile results for mixed

welds of Eltex and Rigidex, beaded and
debeaded. Tested at 50mm/min.
Page 218
t=0 seconds
t=12 seconds
t=20 seconds
t=27 seconds
5mm
Figure 8.15 Illustration of the failure of Rigidex

welded sample with beads. Tested at
50mm/min.
Page 219
25
Beaded
FO 0 D2beaded
24
.,
22
20
18
100 200 sW 4W am
HeatSoak(S)

for Rigidex at the welding pressure of
160
O Beaded
12D
S 80
6
01
100 2W xU 4W 5W
HeatSods(S)
Figure 8.17 Graph of elongation to failureversus heat

soak for Rigidex at the welding pressure
of 0.02MPa. Tested at 50mm/min.
Page 220
a Dýebeaded
o Beacfad
24
23
22
21F
0.1 UZ UJ U.4
WeldingPreswm (Wa)

pressure for mixed welds; welding
50mm/min.
113
0.1 Q2 0.3 0.4

0
Waking Pressire (NPa)
Figure 8.19 Graph of elongation to failure

versus
welding pressure for mixed welds; welding
50mm/min.
Page 221
a hot-plate temperature of 230°C was used. The sample
geometry was identical to that used for the welded pipe

tests. For the dual pressure welds, the peak stress
results remain relatively constant but the elongation to

failure results were found to vary when compared with the
single pressure welds (Tables 8.7 and 8.9). The dual
pressure method showed an improvement in the displacement
to break values of welded specimens for both Eltex and

Rigidex by 20V and 10% respectively. The improved weld
performance is due to the recrystallisaton that takes
place with a 0.025 MPa interface pressure during cooling.
Eltex pipe materials are extremely crystalline and
examination of the microtomed sections has shown that
large spherulites are generated by the slow cooling during
pipe production. This is shown in Figure 6.13. In the

conventional welding process with relatively high applied
pressure during cooling, the shearing action of the molten
MAZ causes immense distortion to the crystal structure.
However, when there is little or no net pressure applied
during the cooling phase, then the structural distortion
of the pipe wall : MAZ boundary is minimised. Similar
observations have been reported by Marshall et al.

[Marshall2 et al., 1995] whilst working on Eltex Tub 124
series materials. They report that the sensitivity of the
weld toughness to the applied cooling pressure arises

solely as a consequence of the generation of a plane of
local weakness caused by melt shear. This can only be
minimised by reducing the cooling pressure since shear is.

inevitable and desirable for melt mixing during the
initial surface contact phase. With the Rigidex pipe welds
it was not possible to discern any significant spherulitic
structure using conventional light microscopy. Since the
Page 222
spherulites are so small, the distortion of the structure
at the MAZ boundary is of almost no significant.
The fracture in the beaded samples occurred in the usual
way, the fracture being initiated by the notch between the

bead and the gauge length.

(MPa) ($)

SBRO10 28.7 (0.6) 28.2-29.4 38.7 (3.2) 35.6-41.6
SBRO5 28.1 (0.4) 27.6-28.4 37.1 (5.2) 34.4-44.9
SHS450 28.7 (0.5) 28.1-29.4 38.7 (5.5) 32.9-43.5
RBRO10 21.8 (0.3) 21.6-22.1 58.3 (13.1) 45.2-71.8

RBRO5 21.9 (0.4) 21.4-22.2 57.2 (11.4) 46.2-67.2
RHS450 22.4 (0.3) 22.0-22.6 56.9 (9.0) 47.1-68.1
S ELTEX
R RIGIDEX
BRO BEAD ROLL OVER TIME
HS HEAT SOAK
Table 8.12 Summary of the tensile results for beaded

dual pressure welds of Eltex and Rigidex
°C.
made at 230 Tested at 50mm/min.
8.9 FAILURE MODE IN SIMPLE TENSILE TESTING OF BUTT-WELDS
Tensile testing of Eltex and Rigidex rarely produced a

failure which was anything other than ductile. The only
exception to this was the planar interface failure which
results from inadequate molecular mixing at the interface
Page 223
between the 2 pipe ends during the fusion process. These
were seen for the long heat-soak time and low fusion
stress welds made in Eltex and Rigidex samples (SS8,
SS12, RR8 and RR12).
The influence that even quite large variation in fusion
pressure and heat soak have on the actual weld ductility
is not clearly revealed by this tensile test. The final
failure modes were similar to that observed for the
microtomed section, which is presented in Figure 7.33,
section 7.6, page 192. In the case of dual pressure welds,

increased heat-soak time and a lower secondary joining
stress benefited the weld and also resulted in stress and
orientation free weld zones.
Page 224
CHAPTER 9
CONCLUSIONS AND RECOMMENDATIONS FOR
FURTHER WORK
9.1 INTRODUCTION
The butt fusion joining of polymers is a very common
The main application is in the production of

process.
where many thousand joints are made each year.
pipelines
Many studies have concentrated on optimising the joining
parameters for particular materials and applications
than the more fundamental aspects underlying the

rather
joining process. This meant that a detailed understanding
the microstructure of butt fusion welds has not been

of
in any great detail.
addressed
The study concerns the more fundamental aspects of

present
the process which are vital if it is to be fully

welding
This knowledge will assist in the optimisation
understood.
improvement of the joining process. The macro- and
and
of the butt fused PE pipes have been
micro-structure
studied in detail. An understanding of the microstructure
the deformation modes in the welds will enable this

and
technique to be applied to the joining of different
materials. Since most of the welding of PE is carried out
in potentially uncontrolled conditions, it is

on site
Page 225
imperative to have confidence that the welding process is
tolerant of variations in the welding practice.
9.2 CONCLUSIONS
9.2.1 MATERIAL CHARACTERISATION
The density of the pipe was found to increase from the
surface through the bore of the pipe wall. The density of

the butt weld was constant through the pipe wall thickness
but was greater than that of the pipe.
The molecular weight distribution showed that there are no
significant changes in the number average weight or the
weight average molecular weight for resin and Rigidex 002-

50 pipe in terms of surface, mid and bore positions in the
pipe wall thickness. However, for the Eltex Tub 124 the
surface exhibited the lowest molecular weight value.
The Eltex virgin pellets were found to exhibit higher
crystallinity values compared to the Rigidex pellets. The
melt-onset for the extruded pipe was found to be higher
than the pellets, which increases in going from the outer
surface of the pipe to the bore. Milling the samples to
produce the test specimens for the OIT study was found to
significantly decrease the thermo-oxidative resistance of

the Eltex resin.
9.2.2 WELD GEOMETRY and DIMENSIONS
The flow of the molten polymer during the butt fusion
welding process was found to influence the shape and

dimensions of the weld. The bead dimensions of all the
Page 226
welds tested increased as the welding pressure increased.
The bead dimensions were also found to be dependent on

their respective position around the circumference of the
pipe. It was readily apparent that the outer bead width

for the top position was always smaller than the outer
bead width for the bottom position. Whereas, for the inner
bead, the bead width for the top position was always
greater than that for the bottom position. Furthermore,
the overall shape was also dependent on whether the bead
corresponded to the inner or outer surface of the pipe. It
was proposed that the overall shape of the beads may be
partly attributed to the influence of gravity on the
molten bead material and the resultant microstructure of

the pipe. The molten material 'sags' under its own weight.
This implied that a greater volume of material was
forcibly extruded due to the influence of gravity. The
width measurements, which were recorded using image
analysis, correlated very closely with the manual
measurements which were taken using digital vernier
callipers.
The bottom outer bead area was always greater than the top
outer bead; and the top inner bead area was always greater
than the bottom inner bead area. These trends were similar
to that observed for the bead-widths. It was also clear
that for welding pressures in excess of 0.21 MPa, the
increase in the extruded bead area was relatively small.
It was found that the flow of the molten material was
"
influenced by its relative position in the weld.
Furthermore, it was shown that with increased heat-soak
time, a corresponding increase in the bead area was
obtained.
Page 227
The MAZ lengths for the outer and inner at the top and
bottom positions always showed that the top outer length
less than the bottom outer length and the top inner
was
length was greater than the bottom inner length. In
general, the length of MAZ was found to be a function of
the position along the circumference of the pipe and the
welding pressure. The length of the MAZ was always found
to be the greatest when the bead area was the greatest.
However, the centre MAZ length was found to decrease along
with the welding pressure.
The bead data in conjunction with the deformation studies
of the welds (using microtomed sections) can facilitate a
comprehensive database which can be used to assist on-site
visual inspection of joints.
9.2.3 MACRO- and MICRO-STRUCTURAL EXAMINATION
Extensive wear of the polished steel knife was noted as a

function of the number of microtomed sections taken. A
procedure and methodology for polishing the microtome
knife was developed and defined in this research
programme. This study has shown that the judgement and
competence of the operator is important in maintaining an
efficient cutting edge for the knife.
Examination of the welded specimens in the transmission
light microscope suggested that the material adjacent to
the weld interface had an enhanced level of molecular
orientation. This orientation indicated that shear flow
had occurred in the region across the MAZ between the
molten material in the MAZ and the cooler parent pipes.
Page 228
This effect was also observed at the melt-solid boundary
at the edge of the weld.
The extent to which the shearing of the molten material
occurred varied with pressure. In welds produced with high
joining pressures, the MAZ was considerably narrower.

Although there was evidence that shear flow had occurred
at the melt-solid boundary, the weld interface was much

less defined. In welds produced with low joining
pressures, the weld interface could be clearly discerned,
with a large MAZ. The macrostructural features of the
welds were found to be dependent on the position around
the circumference of the pipe. The elongated spherulitic
structure which was observed on the bore region of the
Eltex Tub 124 pipe was also observed in the inner bead
region. The inner skin apparently does not melt during the
welding operation. Examination of the microstructure of
each weld had shown the presence of five different zones.

These zones have been explained on the basis of the flow
of the material at the weld and the consequent temperature
variations.
9.2.4 TENSILE TESTING - MICROTOMED WELDED SECTIONS
Microtomed tensile testing of the butt fused welds enabled
a study of the deformation and failure behaviour of
different regions in the welds. The deformation in the
microtomed weld specimen was measured using surface-
mounted grids and image analysis. The initial deformation
was greatest in the centre region of the MAZ. The presence
of beads was found to act as a stress concentrator which
was responsible for the final failure in the pseudo-notch
Page 229
regions at the junction of the beads and the bulk polymer.
Three types of fracture paths were observed for the beaded
samples. For the debeaded samples, the failure was found
to occur within the centre of the MAZ, which was also the
area of maximum deformation.
The testing speed was found to have an effect on the final

failure in the Eltex Tub 124 and Rigidex 002-50 debeaded
microtomed samples. Increasing the testing speed from

5mm/min. to 500mm/min. resulted in the failures taking
place predominantly within the centre of the weldline.
The effect of changes in the joining temperature heat-

,
soak and joining pressure on the deformation modes of
Rigidex 002-50 had relatively little effect the
on
deformation modes. Even the change in the heat-soak time
had little influence in the observed failure behaviour.
9.2.5 TENSILE PROPERTIES
A series of tensile tests were undertaken on the welded
specimens using the WRc tensile profile. This resulted in
yielding at the pins and therefore it proved very

difficult to measure elongation to failure accurately of
the material within the weld zone. A study into tensile
testing resulted in the development of a new tensile test
geometry.
The stress-strain curves obtained for the butt joints
showed higher elongation to failure values for good fusion
conditions, though the yield strengths were the same. The
tensile test showed differences in the elongation to
Page 230
failure only when the welding conditions were
unfavourable. All of the welds with the beads retained
showed elongations to failure to be much lower than that
for the parent materials. With the beads left on, failure
initiated from the notch, formed at the junction of the
weld bead, and the pipe wall. The removal of the beads
enabled a clearer picture of the weld to be obtained.
A limited study into dual pressure welding resulted in an

improvement in the elongation to failure values for the
welded specimens. This was attributed to the
recrystallisation that takes place with minimum structural
distortion at the pipe wall and MAZ boundary.
9.3 RECOMMENDATIONS FOR FURTHER WORK
The present and future widespread use of the butt fusion
welding process in the construction of pipelines systems
world wide suggests that active research into all aspects
of the process should continue. Future work should focus
on three distinct but complimentary directions: (i)
materials and microstructural characterisation, (ii)
modelling the butt fusion process for large diameter pipes

by computer simulation and (iii) experimental data to
validate the computer simulation.
A better understanding of the influence of thermal and

rhelogical properties on the weldability of pipe materials
is needed. This would allow a direct link to be made
between the morphology of the pipe material and its
optimum welding conditions. Furthermore, it is also
Page 231
necessary to gain an understanding in to the influence of
the physical ageing on the integrity of the weldline.
To enhance our knowledge in the area of butt fusion
welding, the next significant work has to concentrate on
the mechanisms involved in the melt fusion process. This
could be studied by using radioactive labeling or dopants
in the interface region. For example, SIMS and NMR can be
used to characterise the interface.
The surface mounted grids method for micro-tensile testing
be refined to include higher resolution polarising

could
This will enable the deformation behaviour to
microscope.
be observed in the microscope. Finite element analysis of
the weld zones could be correlated with the weld zone
microstructures.
Impact testing could be used to assess the relative
toughness of the interface and the MAZ in the welds.
Notching the samples can be carried out by inserting

of
blades at the interface during the welding
surgical
operation.
Fatigue test using single edge notched specimens could be
used to assess the quality of the welded sections and
pipes. This mode of testing may prove to be a

extruded
effective and reliable means of evaluating long term

cost
of the butt fused joints.
performance
Page 232
1ý\
REFERENCES
Allen, N. S., Degradation and Stabilisation of Polyolefins,
Elsevier Applied Science Publishers, London, (1983). -
Allen, N. S., Marshall, G. P., Vasillion, C., Moore, L. M.
Kotecha, J. L., Luc-Gordette, J. and Valange, B., oxidation
processes in blue water pipe, Polymer Degradation and
Stability, 20(1988)315-324.
Allen, N. S., Vasiliou, C., Marshall, G. P. and Chen, W.,
Light stabliser, antioxidant and pigment interactions in
the thermal and photochemical oxidation of polyethylene

films, Polymer Degradation and Stability, 24(1989) Part 1,
17-31.
Allen, N. S., Moore, L. M., Marshall, G. P., Vasiliou, C.,
Kotecha, J. and Valange, B., Diffusion and extractability
characteristics of antioxidants in blue polyethylene water
pipe: a DSC and radiolabeling study, Polymer Degradation
and Stability, 27 (1990) 145-157.
Amin, M. U., Scott, G. M. and Tillerkertne, L. M., Mechanisms
Photo-Initiation Process in Polyethylene, Pargammon

of
Press, London, (1975).
Anandu, J. N. and Dipzinski, L., Interfacial contact and
bonding in autohesion: IV-experimental verification of
theory, J. of Adhesion, 2(1970)16-22.
Page 233
Anandu, J. N. and Karam, H. J., Interfacial contact and
bonding in autohesion: I-contact theory, J. of Adhesion,
1(1969)16-23.
Atkinson, J. R. and DeCourcy, D. R., The assessment of

fusion joint quality and some evidence for the presence of
molecular orientation in butt fusion welds in polyethylene
pipes, Plastics and Rubber Processing and Applications,.

1(1981)287-292.
Atkinson, J. R. and Ward, I. M., Joining of biaxially
oriented polyethylene pipes, Polymer Engineering and

Science, 29(1989)1638-1641.
Barber, P. and Atkinson, J. R., Some microstructural

features of the welds in butt welded polyethylene and
polybutene-1 pipes, J. Material Science, 7(1972)1131-1136.
Barber, P. and Atkinson, J. R., The use of tensile tests to

determine the optimum conditions for butt fusion welding
certain grades of polyethylene, polybutene-1 and
polypropylene pipes, J. Material Science, 9(1974)1456-1466.
Barker, M. B., Bowman, J. and Bevis, M., The performance

and cause of failure of polyethylene pipe subjected to
constant and fluctuating pressure loadings, J. materials

Science, 18(1983)1095-1118.
Barton, S. J. and Cherry, B. W., Butt-welding of

polyethylene pipe, Proceedings of the 28th National
Welding Convention, Weldinf in the Eighties, Melbourne,
Australia, (1980) 1-8.
Page 234
Barton, S. J. and Cherry, B. W., Butt fusion welding of high
density polyethylene pipe, J. of Adhesion, 7(1983)59-85.
Becker, U. and Streese, G., The welding of plastics,
Kunststoffe, 59(1969)679-684.
Beech, S. H., Burley, C. and Bunn, H. C., Residual stress in
large diameter MDPE pipe, PRI-Proceedings of the VII
International Conference on Plastic Pipes, University of
Bath, September, (1988) 15/1-15/9.
Benkreia, H., Day, A. J., Shillitoe, S. and Stafford, T. G.,
Butt fusion of polyethylene pipes -a theoretical
approach, Proceedings of the welding Institute Conference
on Advances in Joining Plastics and Composites, Bradford,
Yorks, (1991)188-201..
Bhatnagar, A. and Broutman, L. J., Effect of annealing and
heat fusion on residual stresses in polyethylene pipe,
ANTEC Conference-Society of Plastic Engineers, Washington,
(1985)545-549.
Bjork, F. and Stenberg, B., An oxygen induction time study
on the behaviour of phenyl-2-naphthylamine in thick walled
TMTD-vulcanised natural rubber cylinders, Plastics and
Rubber Processing Applications, 5(1985) Part 3,277-279.
Bowman, J., Enhancing the performance of plastic pipes,
fittings and joints, The Institute of Materials Plastic
pipes IX Conference, Heriot-Watt University, Scotland (1995)6-18.
BP Trade Literature, (1986).
Page 235
BP Rigidex Polyethylene Pipe Compounds, Technigram,
T400/1, (1988).
Brinken, F., A theoretical approach to the practicle
problems of joining different PE grades, Proceedings of

the 5th International Conference on Plastic Pipes,
University of York, Paper 25 (1982).
British Gas Engineering Standards PS/PS2; Part 1- Pipes;
Part 4- Electrofusion fittings; Part 6- Spigot fittings.
British Gas p1c, UK.
Broutman, L. J., Duvall, P. E. and Nylander, L. R.,
Evaluation of aged plastic pipe used in gas distribution,
ANTEC Conference-Society of Plastic Engineers, (1986)pp.
639-643.
Brydson, J. G., Plastics Materials, Butterworth Scientific,
London, (1982).
Bucknall, C. B., Drinkwater, J. C. and Smith, G. R., Hot
plate welding of plastics: factors affecting weld
strength, Polymer Engineering and Science, 20(1980)432-
440.
Bukin, V. E., Structure of butt fused joints in
polyethylene tubes, Automatic Welding, 25(1972)35-37.
Carlsson, D. J. and Wiles, D. M., Degradation, Encyclopedia
of Polymer Science and Engineering, Vol. 4, Ed.
Kroschwitz, J. I., 2nd Edition, Wiley-Interscience, New
York, (1986)630-696.
Page 236
Cist, J. D. and Smith, J. G., On line inspection of plastic
pipe by rotational ultrasonics, Plastics Engineering,
34(1978)41-44.
Davey, T. G., The Welding Institute Research Bulletin,
November (1988).
Davitt, J., Flexible solutions: plastic pipes for the
water industry, The Institute of Materials Plastic Pipes
IX Conference, Heriot-Watt University, Scotland (1995)19-
24.
Day, M. J. and Gittos, M. F., The light microscopy of

plastic welds, TWI Research Bulletin, 28(1987)41-44.
DeCourcy, D. R. and Atkinson, J. R., The use of tensile

tests to determine the optimum conditions for butt welding
polyethylene pipes of different melt flow index, J.
Material Science, 12(1977)1535-1551.
Dib, N., Day, A. J., Benkreira, H. and Wood, A. S., Computer
aided design of the butt fusion welding processes on

polymer pipes, Proceedings of the Plastic Pipes VIII
Conference by the Plastic and Rubber Institute,
Koningshof, Netherlands, (1992), B2/2-1-B2/2-13.
Diedrich, G. and Kempe, B., Welding of pipes and fittings
made with different grades of HDPE, Kunststoffe,
70(1980)87-89.
Dodin, M. G., Welding mechanisms of plastics: a review, J.
of Adhesion, 12(1981)99-111.
Page 237
Doshi, S. R., Prediction of residual stress distribution in
plastic pipe extrusion, Journal of Vinyl Tech,
11(1989)190-194.
Edwards, M. F. Ellis, D. I. Georghiades, S. and Smith, R.,
Extrudate cooling of medium density polyethylene pipe
sections, Proceedings of the 5th International Conference
on Plastic Pipes, University of York, September, (1982),
Paper 8.1-8.9.
Egen, U., Microstructure in tool-butt-welding joints in
polypropylene pipes, Conference Proceedings: ANTEC, (1987)
794-799.
El Barbari, N., Michel, J., and Menges, G., Weldability of

dissimilar HDPE pipe materials for heated tool butt-
welding, Kunststoffe, 76(1986)341-345.
Khoury, F. and Passaglia, E., Chapter 6- The morphology
of crystalline synthetic polymers, in Treaties on solid
state chemistry, Edited by Hannay, N. B., vol. 3., Plenum
Press, London, (1976).
Folkes, M. J., Haunton, J. and Bowman, J., Microstructure
development and its effect on the mechanical properties of

butt fusion joints, Proceedings of the welding Institute
Conference on Advances in Joining Plastics and Composites,
Bradford, Yorks, (1991)118-127.
Gabler, K. and Potente, H., Weldability of dissimilar
thermoplastics - experiments in heated tool welding, J. of
Adhesion, 11(1980)145-163.
Page 238
Gebler, H., Influence of processing on the properties of
HDPE pressure pipes, Kunststoffe, 73(1983)9-11.
Gebler, H., Long term behaviour and ageing of PE-HD pipe,
Kunststoffe: German Plastics, 79(1989) Part 9,34-36.
Gebler, H. and Racke, H. H., Determining internal stresses

in thick-walled plastic pipes, Kunststoffe, 72(1982)17-20.
Gedde, U. W. and Jannsson, J. F., Determination of thermal
oxidation of high density polyethylene pipes using

differential scanning calorimetry, Polymer Testing,
1(1980)303-312.
Gedde, U. W., Terselius, B. and Jansson, J. F., A survey of
methods for the detection of thermal oxidation in high
density polyethylene pipes, Polymer Testing, 2(1981)Part3,
211-222.
Gehde, M., Bevan, L. and Ehrenstein, G. W., Analysis of the
deformation of polypropylene hot-tool butt welds., polymer
Engineering and Science, 32(1992) 586-592.
Girardi, M. A., The evaluation of mechanical testing
techniques and the influence of defects in MDPE butt
fusion welds, Proceedings of the Plastic Pipes VIII
Conference by the Plastic and Rubber Institute,
Koningshof, Netherlands, (1992), B2/6-1-B2/6-8.
Grandclement, G., Proceedings of the 11th Plastic Fuel Gas
Pipe Symposium, AGA, San Francisco, USA, (1989)159-168.
Page 239
References
Grassie, N. and Scott, G., Polymer Degradation and.
Stabilisation, Cambridge University Press, cambridge,

(1985).
Greig, J. M., Pipes Pipeline International, 22(1977)16-24.
Grimm, R. A., Fusion welding techniques for plastics,

Welding Journal, March(1990)23-28.
Hale, G., Measuring properties of plastic welds, Welding
Institute Bulletin, March/April(1990)30-33.
Harris, W. D., Joining of Plastics, Welding Research
Council Bulletin, 12(1966)1-12.
Haworth, B., Law, T. C., Skyes, R. A. and Stephenson, R. C.,

Factors determining the weld strength of window corners
fabricated from impact modified UPVC profiles, Plastics
and Rubber Processing and Applications, 9(1988)81-89.
Hessel, J. and John, P., Long term strength of HDPE welded

joints low temperature flexural test used as evaluation
-
criteria for butt welded heating elements, Kunststoffe,
74(1984)12-14.
Howard, J. B., Stabilisation problems with low density
polyethylene insulation, Proceedings of the, 21st
International Wire and Cable Symposium, (1972) 329-361.
Howard, J. B. and Gilroy, H. M., Some observations on the
long-term behaviour of stablised polyethylene, Polymer
Engineering Science, 15(1975)268.
Page 240
Isaac, D. H., Eccott, A. R., Pittman, J. F. T. and Farah, I.,
Mechanical properties and residual stresses through the
wall thickness of polyethylene piping, The Institute of

Materials Plastic Pipes IX Conference, Heriot-Watt
University, Scotland (1995)542-550.
Kaelble, K., Physical Chemistry of Adhesion, John Wiley
and Sons, London, (1971).
Kamenova, L., Alexiev, K., Korab, G. and Esaulenko, G., A
study of relations between the welding parameters, the
morphology and mechanical characteristics of polypropylene
welded joints, Welding in the World, 26(1988)146-158.
Keith, H. D., Padden, F. J. and Vadimsky, R. G., Origin of

intercrystalline links, J. polymer Science, Physics
Edition, 18(1980)2307-2309.
Keith, H. D. and Padden, F. J., Twisting orientation and the
role of transient states in polymer crystallisation,

Polymer, 25(1984)28-42.
Keith, H. D. and Padden, F. J., A discussion of spherulitic
crystallisation and spherulitic morphology in high
polymers, Polymer, 27(1986)1463-1471.
Keller, A., Polyethylene as a paradigm of polymer crystal

morphology, Plastics and Rubber Processing and
Applications, 4(1984)85-92.
Kinloch, A. J., Adhesion and Adhesive Science and

Technology, Chapman and Hall Ltd., London, (1987).
Page 241
Kramer, E. and Koppelmann, J., Measurement of oxidation of
polyolefins by thermal analysis, Polymer Degradation and
Stability, 16(1986)261-275.
Lamond, R. S., The polyethylene pipe market, The Institute
of Materials Plastic Pipes IX Conference, Heriot-watt
Land, W., and Erlenkamper, E., Experiences gained in the
welding of thermoplastic mouldings, Kunststoffe,
75(1985)315-320.
Limpar, A., extrusion technology for large diameter
polyolefin pipes, PRI-Proceedings of the VII International
Conference on Plastic Pipes, University of Bath,
September, (1988)16/1-16/12.
Lin, T. T. and Benator, A., Effects of hot plate welding
parameters on the weld strength of polypropylene, ANTEC
Conference Proceedings, New Orleans, USA, (1993) 961-963.
Low, A., Vesely, D., Allan, P. and Bevis, M., An

investigation of microstructure and mechanical properties
of high density polyethylene spherulites, J. Material
Science, 13(1978)711-721.
Lu, X., Qain, R., Brown, N. and Buczala, G., The effect of
pressure and contaminants on slow crack growth in a butt

fusion in a polyethylene gas pipe, J. of Applied Polymer
Science, 46(1992)1417-1427.
Page 242
Lustiger, A. and Markham, R. L., Importance of tie
molecules in preventing polyethylene fracture under long
term loading conditions, Polymer, 26(1983)1647-1654.
Lustiger, A., The effect of an environmental stress

cracking agent on interlamellar links in polyethylene, J.
Polymer Science, Part B. Polymer Physics, 24(1986)1625-
1629.
Lutz, J. T., Thermoplastic Polymer Additives-Theory and

Practics, Marcel Dekker, New York, (1989).
Mackley, M. R. and and Keller, A., Flow induced plymer

chain extention and its relation to fibrous
crystallisation, Phil. Tran. R. Soc. Lond. A., 278(1975).
Maine, L. and Stafford, T. G., PE joining techniques, PRI-
Proceedings of the 6th International Conference on Plastic
Pipes, University of York, 26-28 March (1985) B1-B31.
Maine, L. and Stafford, T. G., In-field quality control of

fusion jointing methods, PRI-Proceedings of the VII

Bath, 20-22 September, (1988).
Malguarnera, S. C. and Earles, L. L., The quality assurance
of heat fused thermoplastic pipeline joints, Proceedings
od the 40th ANTEC Conference, San Francisco, USA,
(1982)279-282.
Page 243
Mallinson, J. N. and Gwynn, D. P., Characterisation of blue
MDPE water pipe, PRI-Proceedings of the VII International
September, (1988)9/1-9/11.
Marshall, G. P., The influence of welding parameters on the
toughness of butt fusion welds in MDPE, Proceedings of the
welding Institute Conference on Advances in Joining
Plastics and Composites, Bradford, Yorks, (1991)202-215.
Marshall, D. I., George, E. J. Turnispeed, J. M. and Glenn,
J. L., Measurement of oxidation stability of polyolefins by
thermal analysis, Polymer Engineering and Science,
13(1973) Part 6,415-421.
Marshall2 G. P., Hepburn, D. and Cosgrove, B., The

,
development of toughness in butt fusion welds for PE
pipes, RAPRA, 11 (1995) 193-210.

pressure
Marshalls, G. P., Hepburn, D., Pearson, D. and Mackellar,
S., Factors affecting the toughness of electrofusion
joints in PE pipe systems for the water industry, RAPRA,
11 (1995) 211-233.
Marshall, G. P., Taylor, M. D. and Dickinson, A. J.,
Assessment of the influence of processing on medium-high
density PE pressure pipes, Proceeding of the 5th
International Conference, Plastic Pipes, York University,
September, (1982) 7.1-7.9.
Page 244
References
Marshall, G. P., Youd, S. and Birch, M., Correlation of
fast fracture, PRI-Proceedings of the 7th International
September, (1988).
McGrath, G. C., The analysis of welded joints in plastic
components using instrumented falling weight impact
analysis and long term loading techniques, Proceedings of

the welding Institute Conference on Advances in Joining
Plastics and Composites, Bradford, Yorks, (1991)148-155.
McGreevy, R. J., Hot plate welding of high density
polyethylene: the effect of melt index and density on weld
strength, ANTEC Conference Proceedings, Boston, USA,
(1995) 1224-1228.
Menges, G., Michel, J. and El Barbari, N., Testing of
dissimilar material welds of various PE-HD pipe materials,
Welding in the World, 2S(1987)181-187.
Menges, G. and Zohren, J. Untersuchungen zum
heizelementscheeiben von polyolefinen, Plastverabeiter,
10(1967)165-171.
Moore, L. M., Marshall, G. P. and Allen, N. S., Thermal
analysis of MDPE pipe: A critical review Polymer
Degradation and Stability, 25(1989)161-180.
Muller, W. and Gaube, E., Material and processing

influences on the time dependent creep rupture strength of
HDPE pipe, Kunststoffe, 72(1982)24-25.
Page 245
Narisawa, I. and Nishimura, H,, Short-and-long properties
of fusion-joninted polyethylene, ANTEC Conference
Proceedings, New Orleans, USA, (1993) 381-387.
Neubert, W. and Mack, W. A., Getting a bead on welding of
plastic pipe and fittings, Plastics Engineering,
29(1973)40-43.
Neumann, J. A. and Bockhoff, F. J., Welding of Plastics,
Reinhold, New York (1959).
Nieh, J. Y. and Lee, L. J., Morphological characterisation
of the heat-affected zone (HAZ) in hot plate welding,

ANTEC Conference Proceedings, New Orleans, USA, (1993)
388-392.
Parmar, R. and Bowman, J., Crack initiation and
propogation paths for brittle failures in aligned and
misaligned pipe butt fusion joints, Polymer Engineering
and Science, 29(1989)1396-1405.
Parmar, R., The long-term behaviour of butt fusion welds

in polyethylene pipeline systems, PhD Thesis, Brunel
University, (1986).
Peterlin, A., Morphology and fracture of drawn crystalline
polymers, J. of Macromolecular Science: Physics,
8(1973)83-99.
Pimputkar, S. M., The dependence of butt fusion bond
strength on joining conditions for polyethylene pipe,
Polymer Engineering and Science, 29(1989)1387-1395.
Page 246
Pittman, J. F. T. and Farah, I. A., Computer simulation of
the cooling process in plastic pipe manufacture including
sag, thermal stresses and morphology, The Institute of
Materials Plastic Pipes IX Conference, Heriot-watt
University, Scotland (1995)364-371
Pittman. J. F. T., Whitman, G. P., Beech, S. and Gwynn, D.,
The cooling phase of large pipe manufacture, Proceedings
of the Plastic Pipes VIII Conference by the Plastic and
Rubber Institute, Koningshof, Netherlands, (1992), E2/1-1-
E2/1-10.
Potente, H., On the theory of heated-tool butt welding,

Kunststoffe, 67(1977)98-102.
Potente, H. and Gabler, K., Welding of dissimilar
thermoplastics, Plastverarbeiter, 31(1986)203-207.
Potente, H., Michel, P. and Tappe, P., The principle- of

hot plate welding of semi-crystalline thermoplastics,
Joining Plastics in Production, Watson, M. N. (Ed. ),
Welding Institute, Abingdon, Carobs, UK, (1988) 105-126.
Potente, H. and Tappe, P., Heated tool butt welding of
polyethylene pipes - Welding parameters and testing
techniques, Materials and Design, 5(1985)273-280.
Potente, H. and Tappe, P., Scale up laws in heated tool

butt welding of semicrystalline thermoplastics, Welding in
the World, 26(1988)18-25.
Page 247
Potente, H. and de Zeeuw, K., PRI-Proceedings of the 4th
Sussex, Paper 24, (1979).
Reinke, M. and Potente, H., Effects of the processing

history on the internal properties of HDPE pipe and their
butt-welding and how they are affected by annealing,
ANTEC, (1982)59-61.
Reitemeyer, P., Extrusion of pipes, Kunststoffe,

71(1981)24-25.
Ritchie, P. D., Plasticisers, Stablisers and Fillers,
Butterworths, London, (1972)230,248-249.
Sawyer, L. C. and Grubb, -D. T., Polymer Microscopy, Chapman
and Hall, London, (1987).
Schlarb, A. K. and Ehrenstein, G. W., The impact strength

of butt welded vibration welds related to microstructure
and welding history, Polymer Engineering and Science,
29(1989)1677-1682.
Scott, G., Developments in Polymer stabilisation 4,
Applied Science Publication, London (1981)1-22.
Shillitoe, S., A study of the butt fusion welding of

thermoplastic pipes, PhD Thesis, Bradford University,
(1991).
Solvay high density Polyethylene, Technigram, (1989).
Page 248
Stevens, S. M., Fourier transform infra-red spectrometry
for the analysis of polymers, TWI Research Bulletin,
29(1988)210-213.
Stevens, S. M., The evaluation of molecular orientation and
structural changes in polypropylene welds, ANTEC
Conference Proceedings, New Orleans, USA, (1993) 378-380.
Stevens, S. M., The effect of welding on the a- and f3-
crystalline forms of polypropylene, ANTEC Conference
Proceedings, Boston, USA, (1995) 1196-1200.
Stevens, S. M., A morphological examination of
polypropylene hot plate welds, ANTEC Conference
Proceedings, Indianapolis, USA, (1996) 1275-1279.
Stokes, V. K., Vibration welding of thermoplastics. Part 1:
Phenomenology of the welding process, Polymer Engineering
Science, 28 (1988) 718-727.
Stokes, V. K., Experiments on the hot-tool welding of

dissimilar thermoplastics, ANTEC Conference Proceedings,
New Orleans, USA, (1993) 953-960.
Stokes, V. K. and Hobbs, S. Y., Strength and bonding
mechanisms in vibration-welded polycarbonate to
polyetherimide joints, Polymer Engineering and Science, 29
(1989)1667-1676.
Tadoharo, H.,. Structure of Crystalline Polymers, Wiley,
London, (1979).
Page 249
Terselius, B. Gedde, U. W., and Jansson, J. F., Structure
and morphology of thermally oxidised high dendity
polyethylene pipes, Polymer Engineering and Science,
22(1982)Part7,422-431.
Tirumalai, V. C. and Lee, L. J., Characterisation of
polystyrene in the vicinity of hot plate welded joints,
Proceedings of the ANTEC 1990 International Conference,
Society of Plastic Engineers, Dallas, (1990)1751-1757.
Torsun, I. S., Day, A. J., Wood, A. S., Dib, N., Rosala,
G. F., Folkes, M. J. and Hinchliffe, F. A., An intelligent
knowledge based system for the elec :tro-fusion welding of
polyethylene pipes., The Institute of Materials Plastic
Pipes IX Conference, Heriot-watt University, Scotland

(1995) 378-388.
Vanspeybroeck, P., Analysis of results of the constant

tensile load test for fused PE gas pipe materials, The
Institute of Materials Plastic Pipes IX Conference,
Heriot-watt University, Scotland (1995)92-106.
Venema, B. and Wolters, M., Evaluation of possibilities

for non-destructive inspection of butt fusion joints in PE
piping systems, Proceedings of the welding institute

Conference on Advances in Joining Plastics and Composites,
Bradford, Yorks, (1991)128-136.
Voyutskii, S. S., Autobesion and Adhesion of High Polymers,

Interscience Publishers, London, (1963).
Page 250
Ward, I. M., Mechanical Properties of Solid Polymers, Wiley
and Sons, Chichester, (1983).
Water Industry Specification, ISN 0267-0305, WIS No. No. 4-
32-03, "Specification for blue polyethylene (PE) pressure
pipe for cold potable water nominal sizes 90-1000 for
underground or protected use)" WRc, July (1989), Issue 2.
Watson', M. N., TWI Members Report 382/1988.
Watson2, M. N., The butt fusion joining process-fundamental
aspects, PRI-Proceedings of the VII International
Conference on Plastic Pipes VII, University of Bath,

(1988)18/1-18/13.
Watson", M. N., Plastics joining at the welding institute,
The Welding Institute Bulletin, Jan. /Feb. (1990)4-7.
Watson2, M. N., The butt fusion (hot plate) welding process

for polyethylene and polypropylene pipe - improvements to
productivity, Members Report 419/1990.
Watson3, M. N., Improvement to productivity with the butt
fusion process for joining plastic pipes, Proceedings of
the ANTEC 1990 International Conference, Society of
Plastic Engineers, Dallas, (1990) 1804-1807.
Watson, M. N. and Murch, M., Recent developments in hot
plate welding of thermoplastics, Polymer Engineering and
Science, 29(1989)1382-1386.
Page 251
Williams, J. G., Gray, A. and Hodgkinson, J. M., The
determination of residual stresses in plastic pipe and
their role in fracture, Polymer Engineering and Science,
21 (1981) 822-828
Williams, J. G., Fracture Mechanics of Polymers, Ellis and
Harwood, Chichester, (1984).
Wilson, K, A., Verification of butt fusion weld quality in
large diameter PE100 water pipes, The Institute of
Materials Plastic Pipes IX Conference, Heriot-watt
Wolters, M. and Venema, B., Mechanical quality of welds in
polyethylene pipe systems, Proceedings of the 5th
international Conference on Plastic Pipes, University of
York, September, (1982), Paper 26.1-26.11.
Wood, A. S., The butt-fusion welding of polymers, Chemical
Engineering Science, 48(1993)3071-3082.
Wool, R. P., Yuan, B. L. and McGarel, O. J., Welding of
polymer interfaces, Polymer Engineering and Science,
29(1989)1340-1367.
Yeh, H. J., Infrared welding of polyethylene pipes using
radiant heater, ANTEC Conference Proceedings,
Indianapolis, USA, (1996) 1245-1249.
Yuan, B. L., and Wool, R. C., Strength development at
incompatible semicrystalline polymer interfaces, Polymer
Engineering and Science, 30(1990)1454-1464.
Page 252
Ziatsev, K. I., The cause of fracture of joints in plastic
pipelines, and calculation of optimum process parameters

for welding them, Welding Production, 19(1972)20-22.
Ziatsev, K. I. and Bukin, V. E., On the relationship between
the paths of fracture and the structure of butt welds in
polyethylene tubes, welding Production, 20(1973)18-20.
Page 253
Appendix 1
ELTEX TU B 124 PROPERTIES
Properties Test methods Units Values

Density, standard,at 23C ISO R 1183 kg/M3 951
ASTM D 1505
Melt index, load 2.16 kg (MI2.16) DIN 53735 g/10 min < 0.15
ISO 1133
Melt index, load 5 kg (MI5) ASTM D 1238 0.5
g/10 min
Apparent dynamic viscosity at 190°C,at 100 s", SOLVAY (I) Pa.s 22.102
2
Inherent viscosi in tetraleneat 160°C SOLVAY j /g
Yield stressat 23C
at 50mm/min DIN 53455 Test Spec. MPa 25
at 100 mm/min No. 4 (3) MPa 25
Tensile strengthat break, at 23°C
at50mm/min ISO R 527TestSpec. MPa 39
at 100mm/min No.2 MPa 37
Percentageelongation at break, at 23°C
at 50mm/min % >600
at 100mm/min % >600
Modulus of elasticity at 23°C ISO R 527 (°) MPa 1200
SHORE D hardnessat GPC 62
20 °C -
at DIN 53505 60
-
at 40°C ASTM D 2240 (5) 56
-
at60C 54
-
at 80°C 52
-
VICAT softening point, load 1 kg DIN 53460 C 127
VICAT softening point, load S kg ISO 306 °C 77
Thermal conductivity, at 23°C ASTM D 1525 W/m. K 0.38
Coefficient of linear expansion DIN 52612 K"' 1.3.10'4
ASTM D 696
Specific heat, at 23°C Calorimetric kJ/kg. K 1.9
at 100°C kJ/kg. K 3.2
Brittleness temperature ASTM D 746 °C <-100
Molecular weight distribution SOLVAY medium
() Apparent dynamic viscosity of the material extruded at 190 °C through a 1.00 mm diameter and 15 mm long
die, at apparentshearrate 100 s'.
(2) Concentrationof the solution: 1 g/l.
(3) Measuredon compression-mouldedplates.
(4) Measured on compression-moulded plates. Testing rate = lmm/min; thickness: 4mm.
s 15 s value, under constant load.
Table A Selected properties of Eltex TU B 124, reproduced from Eltex data

sheet [Solvay, 1989].
RIGIDEX PC 002-50 R 102 PROPERTIES
Property Test methods Units Values

Melt flow rate-2.16 kg load BS 3412: 1976 g/10 min 0.2
ISO 1133 - 1981 Procedure4
5 kg load BS 3412:1976 g/10 min 1.0
ISO 1133 - 1981 Procedure5
Density BS 3412: 1976; ISO 1872-1972 kg/M3 950
Tensile strength at yield BS 3412:1976; ISO/R527-1966: Type MPa 19
2 SpeedD
Flexural modulus BS 2782:1978:335A MPa 700
Elongation at break BS 3412:1976; ISO/R527-1966: Type % >800
2Speed D
Hardness ISO/R868-1968 (Type D) - 65
Vicat softening temperature BS 2782:1976:120A C 116
Brittleness temperature ASTM D746-79 C <-70
Linear thermal expansion ASTM D696-79 Average value over C 1.5 x 10
temperaturerange 20°C to 60°C
Thermal conductivity BS 874: 1973 W/m C 0.4
Thermal stability-Oxidation induction ASTM D3350-81 C >240
temperature
Oxidation induction time Isothermal at 20d C min >30
Stresscrack resistance F) ASTM D1693-70 Methods A, B&C hours >1000
Electric strength BS 2782:1972:201B, sheet 1mm thick kV/mm >20
Volume resistivity BS 2782:1982:230A 12m >101.1
Surfaceresistivity BS 2782:1982:23IA Q >10
Relative permittivity BS 2067:1953 (between 1 and 20 2.6
-
MHz) IEC 250
Loss tangent BS 2067:1953(between 1 and - 3x10
20MHz) IEC 250
Table B Selected properties of Rigidex PC 002-50, reproduced from Rigidex

data sheet [BP, 19881.
The data presented in Appendix 2 for the measurement of
bead width was obtained by using image analysis. The data
are an average of seven readings and is accurate to 1
decimal place. However, the data from the image analysis

correlated very closely with the manual measurements which
were obtained using digital vernier callipers.
Appendix 2
OO
a
00 NM
i h
n a
N N 0
M
00
O
v
Oý O
- g r
N M
M
o
b 00
Oý
0 vn
.+
00 M M M N1 00 M N1 M N M M M M M 00
V1 º- O, In In Os O of 0% b Os In 05 O5 g 't h N O
N 2 e v1 'O M ýO rý IO r1: Oý 00 rý r": lý v1 lO o
lq: ..
V V e e d' e
-* -f at -e -e -e -e -e -e e 'et vi
1 0 ý b M N In In 00 V N N 0
-9 O -1 N
Oý 00 O5 0% O5 0% O5 0% Ö 0% T 0% .O5 0% N
O5 0% 0% Ö
r ~
_M
y
y , h 50 M %D h Ooo
' h N T N N ein 5 M h O
2 O5 v1 - - v1 h O O O5 ao O' oo of O IO In O
Ö Ö Ö Ö Ö Ö Ö Ö Ö Ö Ö Ö
Oý O' O'
^
1 00 tN M b M O' O h N O \O pý 0 0 O O O
7 N _ O: V R M 00 Os O V1 N N M
N 2: 2 h
Ilt
h h h 00 h a0 0o Oy N N N
N N r .. .
y
r 0 N -V M M In 50 N O O, 05 N 00 0% O5 of 00
V) O O
O O 1ý ei ý". ý+ n N N N ý+ 61 O, In 00 IC!
.N N N N
2 O'. 0% LO h: OÖ 0 ei N 0 0 0, 0'
.. . -. .- . -ý N N N N N N N N N N N
N O M O5 00 V1 M O 0 V1 O N vi - In
Oý Mf 1D a0 h 7 'ýf ýf 7 ýO h oo v1 M O
b b b b b b b b P
00 h 1G ýC oý 00 oÖ tý
y
I. In f- 0 00 0 M O5 h - In
y O N O
_ v1 N 00 NN . -r VS 00 h N . 00
O N O - 00 - Co -
N N N N N N
ö $ N 0 '0 00 00 ID N N 0
N1 In h N '0' Z Z M M M M tT a h h Os
h Z
0'
N
H 4! f O h $ M 0 M O, b N N h Y1 N b O
rý ; M M M M `7 N N O IO M N
h vi h h b ýO ýO ýD 2 b b b b m V1 n Yy Vy
pý O O N
h O - b - b O 00 M N p
O N n h oo '0
0O vO v1 b b O 1 ei N O
le: Ö N
N O ON ON 00 00 00 00 00 00 00 Q%
O ,
.. 00
. ..
N
U) w M_ N_ M N b eý b 00 b V) M N_ 00 ýp N In
y S 00 Nh OO M M N N N 00 'i O IC
O f`! !V lV M N1 1+1 M N N Ö
O
:
k 0' 00 b N - OO h h - 0' v1 h N NO N 0
ZI 00 Oý e'1 M Qý le: 00 O 1ý OO T 1.O lD b 00
'L Q% oa a 0% e% 00h N N N 00 h N tP 00 0% o, 0%
ýý
y O N ýD O ýO h o0 N 0% 0 M 0 N 0 0 O
2 M N1 V1 Oh n P o -i Cý T M O 1ý h
d d o1 2 ° : : 2 2 = Z Z ° ° o,
.. .. .,
p O h $ N v1 M h O1 'D 0 O In h N e+1 00
ýO h tß'1 M N N ýO Oý V1 h oý 00
M 00 NI R1 M M M M
ýp ýp M V1 1p V1 V1
h in b N h ')f O- M <f M h h Vf
ein v vi oo a n O m v)
'C e vi N V1 V1 vi vi V1 V1 M, M
c3
M
vi i --i N
b a0 b N N v1 M v1 N h Oi M O. of b N VY h r-1
00 00 b OO M - N M N M SO V1 h - Cl! o0 - 4)
V M M M M M N1 M M 2
N N h In In 2
NM K
CA
v N pý ^ O N N O N N b M V1 s N g cg, g N Q)
y ýn v p o0 0o en h o N en v, a C v o v h
Z % % oo
Z M vi b b wii vi en M
-i 9 9 .e
pý N b ýO ^ N p. O 0 M p N T
0% V1 a N1 O O O In a of V1 o o O
. 0
Nn -e en m «% ei en ein m ein en m e vi vi w
y
N
p O 00 M N N O'. h O b O h N 0D m In
V N 00 00 b N N V1 h 00 G V1 IO h t} M Vl
; b 4I
M Z Z V m b b 2 1O b hr mn m M m
.
10
. .: r
9 M tS ^+'1 Z C? ^ O N 00 00 N 0 Co t+1 O5 00 00
O5 O V e v V o0 N O1 M h O 7 N O'
yý ý ý of M M M M M ý M ý Pi Vi 1I) ý ra
ro
a)
N ý+ O h ý g $ O N N eý1 O h N ýp M ýO h
h N 00 lý l+l tý1 V 'R ý+ h O N Oý ýD h
N h b h fý h
N1 ý ýO h ^ r ý ý ý ý ý
N M NM M P V1 N o0 N 00 In N M
M -1 V M vi V1 < e: N M N M el' M
00 Co 00 00 00 00 00 00 00 00 00 00 00 Co 00 OG od aÖ
N
vi
Y N N M n Ö
N
y ~. p N 00 O
ýJ O
. . ,
Oý ON 0' O, 05 O. Ö
O O O O O C: O, p' p, o' pZ k
b
u
v
a
ä
1.. (- %ß 4MMM a' On H CO N &n %D MONN
e-4
rA 01 MOm OD l0 t- %D w t- ON Ul 01 Ow
rl
rr tn uº uý ui uý uý in ui uý ýo ko
H rl rl r4 HH ri H H r-1 r-i f-1 eIH VA ri rl 9-1 H
$4 co l0 Co 10 tO l! 1 M CD 0 lO O% tO -W 0 co OD v
sP
d MI OD N f1 CO M 10 CO 10 l11 H OD a0 l! 1
4J
MM fn wvw U1 Ul U1 ul lA Ln Ln %- MMM
"1HHH rl HHH r-1 HHHHHHH C-4 H
1.ý lA U1 Mw co M k0 HM ri ON H %D -4 M ON O CO l0
4) rl N rl CADNO rý kD d! sN MM V' 10 Cý O rý Cýr!
2
H1 rl OOO 01 O% 01 01 01 a1 01 0 o1 OOOH
H HHHH ri HH rl H r-i
M
a ON OmH %D t- rn Ln U1 O In nOm
$1 a 00 In lo
mmv [- 01 A %D t0 rý l0 U1 v Oº a% co I- MO t-
41 ...................
rl e4 NNNMMMMMMMMNNNNN r4
p,y H rHl '-1 HHHH HHHHH
e-i r-1 r-1 r-1 HH '-i
$i Q1 0% OD GO N CD sM %D (n NMV 00 V 00 CO OD t-
wON co º(1 NO co 01 ON G1 Co c) sr 0) "r Co U1
CO C0 (1 Cý (ý 1H
E 1N rl 01 01 C1 01 GD CO 00
H ri
rl
NNN
H NNN rl rl rl rl rl rl r-I rl ri rl
0
IX $4 1NMN rl N co M (v1 CD ON OD NM CO M OD r-1

O U1 v CO w a% M G1 CD C0 ul V' ri Vý rý
y rý
U) tN NN tO %D to sM V4
V' lA U1 WO tD IN N CC) NN
r4 H '-1 HH C-1 HHH ri HHHH '-I HH '-1
H E
E
$4 co NOOOOOr lf1 O U1 [- OO U1 tO O
Cl a (71 ON N r- m OD tf1 0% l- Cl ON r4 %0 %0 %D
H 4MMMMMt}ä1)
U1 tf1 1f1 U1 sN lfl Ll lfl U)
3 H e-1 r"i ri ri HHHHH rl H r-1 HHHHH ro
r1
ä a)
$.i tN co lO NOOOOOOO U1 NN U) O t0
A NNw %D H CO N Co 01 Ln
4) U1 NMO eP H [N
Q Aj .................
%D %D [- CO CD a% m a% 01 (A 0) 01 co 00 I- l0 l0 ro
W p ri H r-+ HH 'i HHHH r-+ -r H .-i HHH
"ri
C51
oý
$4 U1 co H Ln Co Co l0 M %D O 01 tf1 lD lo a% OH r- a% "ri
\O co t0 N C1 l- wMN rn mNNN U1 O tr t! 1 Ul (Yr
y
H ý.r
rl HH H"1000OOOOOOOO r-1 eý 14
HiHHHH rH HHHH ri HHHH .-1 ri H O
H
w
$4 wo co O GO U1 . -1 MO -ýM tf1 M Ul OD OmO ul N

V I. OM (- l0 N Iý h [. %D %D U1 MI ON r. tM co r-
3-1
41 ro
NMMMMMMNN
p HHHH rl HH rH HHHHHH ýH rl H r-1 H "rl
$4 . in O1 -. H LÖ [- Cc Ln ON tO O\ QO OM I- M co "o
4 01 M CO %D N [, Vw N ri N ri OOH to r. H rý CO M
y
:
OOOO 01 (l a 0; Ql (1 11 C1 Ol Ol Cý OO
H e1H ri HH .-1 e-1 H
$4 CD N V. ONNm Vw Ln oHN Ln 0) Ln Ln NN Co CN
y Mm to 0% H U1 O% co co H ON co ON co Ul [N 01 Lf1 H
He1HHNNNNMNNNNNNHHH
p .ý . -ý H .ýH .ýHHHH r-1 H r-1 HHHH ri H
d
r OHNMy Ul %D [, co Ol
rl NM Vý in l0 lý CO Oý 04
H 1-1 rl e-i ri 14 ri It r-i rl
04
U)
The data presented in Appendix 3 was obtained by image
analysis and is accurate to 1 decimal place.
Appendix 3
.. o 0 0 o O o o O o O O o O O
_, o 0 0 0 0 0 0 o O o o O o 0
. . . . . . . . . . . . .
m o o O o O O ýr to O H N O o
co N Ol M N N to M
ri H
ý. O O O O O O O O O O O O O O
v O O O O O O O O O O O
O O O O O O N C1 O N N N l0 d'
ý°
CO H M N c-4 I! ) M
O O O O O O O O O O O 0 O O
O O O O O O O O O O O
.. lD N O O O O N O O Q1 N rl c O b
(D d' Cl H H lD to Ln r-1 0) to W
r-I
a)
O
O O O O O O O O O O O O O O
v O O O O O O O O O O O a)
" dI Co N
H cl N O O O O N O O N N N O O
d' m H r-1 ýl %D in H 0) to
44
0
N Ul H d' to H
Q' H
"" Ul
M
c4
M
00
O cl CO 111 N N
U)
M
01
00
N
de
u)
O
Co
[N
to
.--1
a CD
0
a H e T-1 O O O to to e-i r1 H Cl M N 4J
a)
1J
N [- Co M 0 U) U) In d' H L() M 00 00 00 zW
-"
H O N O to N r-I lO Ql 01 00 -KM Co e M z
0
o
W H
a -- ýD l0 w U, in LA cc a di in lD lo to
N w
LY
G Ei
0
N c+t CO O O d' to O O of O CO Ln -ri Z
w
ý" -1 H-1 N u
A Q g O O Ln c% dw CO t M N N M O N to 0
a di N CO %0
N
`-' U) to u) -i M CM Lfl c1 Lý Ln
Um
. rj
b
" fu - M 0) to m
N
N
de
CO
M
O
41
0)
CO
e
de
0
rý
N (3) H CO ro
r-i
(ý N "' Ol 111 CO e l0 e e--1
H a a i CO N M Ln H de N -M N CO N M M,
ýO dý N CO CO CO N Ul N O N d' de lo
w
E-4
2
H H
0
a)
w
" RS N M O H Ch Lfl M CO 0) M O M a1 cM
-
(Tý U ^ H tD M it N N M O N to
U) M N O rn
Q R(,' 1!1 O Ol Ol H N cv tN Ul <M O Ill N O
0 9
U) M Ul U1 tO to H N U) CO Ql N M U)
4J
N N U) N Co tO N to Ol ' C1 CO Ql
U
"" e-1 U)
"dý N ýD tO O N Co m l0 O in M (D d'
,A M 1. H cn * c% 1, to c0 N to N al
'd' C1 M M N N M to ' U) U) Co 0) Ln
H H M
r-1
x W W
rd El
N Ö
0HNmv H
ri Nm 'ä 111 to N 00 al A H
En En
U)
U)
U)
U)
U)
U) cn
Lo U2 Cl) Cl)
r-1
Cl) En
rl
Cl)
ý--I
a
U) En Q H
U) U2 Cl) En U)
0 0 o 0 o o O O o 0 o 0 0 0 b
0 O o O o O O O O C) O O O o o
;
& O O O O O O (") t11 e-- 01 N M O O
r1 al m 01 r
IR: eM N N N W
r-H H
X
Q)
O O Co O O 0 Co O O O O O O O 41
0 Cl O O 0 0 0 O 0 O O O O 0 r-I
" M
O O O O O O O O N i-1 N M O (V
M C Ol d+ Lf1 N N to 4
H 0
0 C4
O W
0 Cl 0 o O o 0 o O CD O O O O r
C) 0 o 0 0 0 o O O C) O O CD O 4J 0
r-4
N N O O O O k0 O rý N N N N N E
CD q;p co co CA ýr Ln Ln rn r- ko
04
O Co 0 o O o o O o o O 0 O O
O o O o o O o 0 o O O o O O b)
N N O O O O LN O N CO N t0 N N m
CD 0)
d' co d' m L[) U) U) 0) N lO
N H 00 LIl v lo U) a1 a)
. -" U) r-i e-I co Ol c-I tO
M Ol a Ql co U) N N Ll 00 O H LO H 41
(d
a M 0 0 0 %D lD H H H 01 M M H
r1 rl
H
N W
E Wa
t '.Ulo N 'O ' o Cl o N qr rn En CO O
H" ., r-1 N rM ' 0l N Cl [ý M l0 01 ý--I o W
,
a m V M M N N %D to N M M CO M M
44
- U
Ö m
O QO M N 00 to H N O -to
N cal N M N * N
O M CO V * M CO WO l0 N %D N M co
to %D Ll Ul Ll Lfl CO CO Lfl [N N H lD Ln
H
rd
ro
" fd U1 dl IN Ell Lf1 O co H N CO e-1 H to N
"-"
W 1 Ul T-4 co Ol N M N M N Ol in N C) 3
'x
H a' ý--ý M ko O co N Ln m o c) 00 O 00 GAD [-4
to rn l0 N l0 l0 r-I eN %D Ql 0 M .0 ö z 2
" fd "" Cr al Co l0 rH ,,I cM (n r-i N 00 0) Lfl 01 aý a aý

(Yý ^ w H CV O H CV w l0 CV O O H H H
4 0 a ýQ
Ol m M O 0) rl d' M Ql CO M Cl N H
in M N N r-I m to Q1 O N rfl in (d
e-I 41
(S
C-)
CO Q1 m N %D M N O l0 O H N M N
O N N N O N O W O N N O N d+
ý+ N O N O U) U) Ol I O M Ol 0; N CO N
M 00 M m N N m M .0 in m 0) [- En
W a
I'D
O N M d' 4) O H
c-i
ri r4 M ' LI) %D N co 01 ri . -I H H H 04
U) En U) N to Ul U) U) U) ul U) m En En P4
Cl) Cl) Cl) Cl) M U) U) Cl) U) Cl) Cl) Cl) M U) 4 O H
.. O O O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
0
O
O
O
O
O
O
O
O O O
4
N O O O O O Ll ý4 Ol L11 O O O
("1 ON m . --I e M m N Ql l0
r-I . --1 r-1
N O O O O O M O to al to O O O
m°
M Cl M r-i U (n M N 0) to
H .--I .--I
+. O O O O O O O O O O O O O O rd
° O 0 0 O O 0 0 0 0 O O 0 O O H
4 N O
M N N O O O cc c; r N lO to N
CD qe 00 CD r-4 Ol to in d' 01 NN to W
'--I X
0)
4)
H
0 O O O O O O O O O O O O O
ý.
v O O O O O O O O O O O O O O
s7ý O lo O O O cc O cc l0 N N M O w m
qT Co ' N 0) lD Ln dI 01 N l0 0
. "'1
E
0
[ý7 c% Co in 4 p to 1.g a1 N Ol l0 1J
rl ý-1 1)
V dý cl O 01 CO Ln N N M CO N H l! 1 r-i 0 a
a - ri M r-1 O O O tO tO H H H O M N A w
ýZ
a)
H
WW
N - N r-1 N
N
. "1 tD N
L11
N
O
M
N
N
O1
M
tO
[-
N
O M H aa
N N r1 M M if
0
m e M N N N N In N m M 00 M M W
U
0 Um
. r-{
tý7 "" 0 M in Ln O N tO N to (3) Co M Co En
O in i* N M -T M M 0 N r-1 Co l0 O CO
a; 01
N
A - ul tG In in In In Ul CO w c to U')
E
ro
" «1 "-" N d0 c% (n to M
CO
0)
tO
Ul N M O (D Co d' rd
(iý ^ Co to 'd' V' to M co Co N Lfl 01 00
. . . . . 4 . .
a)
H 4 -1 O
.
e %D %D 'd* tD o0 c0 N
.
rý
. .
M
.
x
y to %0 l0 l0 l0 H Ve Ln 00 (D M e in
fi w
0
a) w
Ln lo O H N M 0) d* M O CO 0)
iYi N 'r H Ol to M de e-1 O CO N .-i 0 N en cl
0 a P4
c; R"i ý--I f ý1 M H M M tM t11 N al Ol N a1 Lt)
.E to M tD N N N ri M lD 0) N en in
r-1
H (d
II
U
lo CO in Ul U') N [N .--1 00 Co l0 O H to
Ql [N %D to N Ifs e-i U'. e O Co N M O
Ol tO CO 01 M N N O l0 in m N in M
t- M N. N N de N -cM to Ks O [- In
H r"H N M
a
Wz
. rl
b w
N Z
N H
w-i N (n cp Ln tO [- co (1
0rNM
r-1 Hi
Nit
A
0
r-I
U) U) U) U) U) U) U) U) U) U] Cl) U) En U
U) U) U) En Cl) V] U) U) H
CI) Vl Cl) Cl) Cl) U) 0
ý. O 0 O 0 0 O O O O O 0 O O 0
O 0 O O 0 O 0 O 0 O 0 0 0 0 b
. . . . . . . . . . . . .
° N O O O O O tD O O O O O L( O
m
in cl M O Ve Ul -e c9 Co to
N
V-1 r-i V-4
v
..
O 0 O O O O 0 O O O O O O O
N O O O O O IN O O O O O O O
m cl c'7 O -M Ul -e N 00 in
H H w
r-i 0
O O O O O O O O O O O O
0 w
O O 0
O O O 0 O O O 0 O O O O O O . ri
O
.
O
.
O
.
O
.
O
.
O
.
in
.
N
.
tll
.
d'
. . . . 0
' M co O
m H
q* Co Ol in in d* 01 N lo
U) W
0
a
. -. O O O O O O O 0 O O O O O O
° 0 0 O O O O O
O 0 O O O O O w
a)
m r-1 O O O O O 111 O N N e O lo 0) r--I 0
CD
d' Co H Ol in hfl d* 0) N in
CO in de %D
" . -" C1
-T
ri
O1 O Ql
-4
CO in N
r-1
N
in
M
C1
Co
N
N
in
O
in
tO
to
H
a)
r4
a O O O l0 to H W
y r-1 M e-i 14 c-i (1 M H
a)
W
0 O O O O O O m Ql d' H O1
4J LL
" .. to O
e H 5 a1 M CO ID M N M e M N in Co N O E-1
to 0
.l -- cri 'c in in ui in to to rý is cn ca to w U
Cl)
0 U
lD lD d'
0
N " "" Co N C1 00 N M N Ul co Ol H
O M rl N ri O N O O ce N ýo
ý ui Uý in in rn oa in a)
,. -- in 'c CO N o Ln In E
r-1
ro
fxs CO CO 0 (f'. H O N r-i N Ln dp in Co N ro
H
(Yi Ol ^ 00 CO r-1 [N M Q1 M U H M in r- to
3 a)
H Q
...
N
l0
d' N
to
M
N
CO
to
(V
N
N M N 1-I in N r Ql xH
M r-1 c' tO O O 4M M ä
w 0
r-1 r-1 0 z
a) w Pý
" (Q in c4 H M rn l' Cl N M H to r-I lD lo a
tD in r4 CO N lO H Ol ß1 in N 1n lf1 in b)
01 in 0 N 0 Co cM to Ln ca
0 0 P4
M N ca
in m tD to N to e-I M in CO O N M in
H 41
. 4 O %0
U
N - N in m in in 01 in N l0 l0 e
N M N N In tN M rn 00 Ql m H Ol
M to ý-1 O ca -i in Co to to CM ca M to
-e l- M M N N m N M in -e CO Ql to
H V-1 H M
x G4
. ri W Wz
H
ro Z
O CN M -e 0 H
rl N
U)
f+1
N
dý
U)
u"
U)
t0
U)
tý
U)
0ý C1 r-1
ri
r-1 r-1 r-1 r-I
a)
Cl) U] ý (n cn Cl) to
N U] U) U] cn U) U) cn U) Cl) U) rn cn H
0
,. O O O O O O O O 0 O 0 O O
O O C) O O O O O O O O O O
j
O c; O O O r-4 oo N CO lO r-1 W
00 N CO N r-I m .0 rr)
H
o O
de O O O O CO CO 00 l0 N l0
0 O r-I
co N CO N v-I M w M
r-I
O O
O O
; 4 i
O CO O O O O N d4 LA di
N Ch LA -i %m Ol l0
O QN
1--1 r-1
W
O O O O
Ei
O O O O O O O O O
O O O b
O O O O O O O O O O 0
N O LA N W 0) 3
O 00 O O O O
d1
Q)
rd
l0 O O N N Ol al 01
dý N !n N ri "14
e tD N N CO N H
Q1 N [N W Lfl r-4
"
3 O dý LA H O N dý N a
O ' O O O
...
4-4
0
CO O N N m Q1 Ql 04
ýp r4
_ '.0 12 .I N N CO l- 1-1 O
V cl N N
i
O O -e LA '-I 0 N N
a O de O O
..
N H
J-ý W
Pý
0 to N CO i-I CO
N M di O O O LA Ul
Na
44
hfl CO Cl (n ý lO 00 tD U
a to CO tD Ln LA
N "N
r.
O Um
N . rI
0 0 CO 0 w %D (N M
CO CO p\ r-1 w lo w
N Co de H 4 01 to -1
N CO p
V-1 r-i
b
c> dý M N al Ln M e Co N Co b
H
9 N 3
CD %0 d' N -r O cn o Ln .-4 N w-I
L4
W.
H
...
N Co N CO
CO 01 LA O M W l0 Ln lD 44 H
0
W-4 W-1 z
w
c% N d' Ln N LA
a) a
O CO N en .--I ö
l0 N H 01 M d' . --I s C\ lo
N L Pý
H CO N d' H
a
CO tf 01 dý 00 N
O O M mot' cd
N LA LA lD l0 en %0 00 cn -M
In J-)
fd
U
O O CO e-1 O e M l0 m w QO w
N .ý 01
pp N Ill 1,0 Ln N l0 O to e to
N a1 LA
%D N lie N 00 O Lfl rl e--I l--1
V-4 0' LA LA CO Ln
(n 00 en n N H--H
V-4 r-1
a
b Ei
4% 0 H
0 N le M dý
N (fl d' LA %D N co H ri r-1
ý-i
ä ä ä ä ä ö ýa ä 0 H
ä
b
o O O C) O O O O O CD 0 O
G
o O o 0 o O O o 0 0 0 O
o o 0 0 O o 4
G N 00 M M 0ý
l0 N Ul N
CM p1 N N
H
-e
H Zj
O O O O O O O O O O O O "rl
G O O O O O O O (D O O O O G14
o N o 0 0 0 o
0 00 If) CO 4-4
m Q1 N l0 N fV in 0
H H
O
O O O 0 0 0 0 0 O O O O N
O O O O O O O O O
0 O O O
0
t0 O O O Lfl U)
O 0 t0
M O O
CO
O 0
ei CO lf1 r-1 N %D O w
H a
V-1
o O O O O O
43
O o O CD O O
a-
C
0 o o O O o O o o O O
O
o o
14
Co
O o r+
M
H
Co o
L[1 H
H
N ýo
ý--I
O
N to C% r-I M CO 01 M
N CO N H-1
N tD U) dI '.O N N l0 W 1
U Q1 M
Wz
d' O O O dý dý ý-I Ol dý N
O d'
av
41 WW
Ltl (4 CO C> r- C:) e 0 aa
N^ N CZ
CO w de A d'
V-A
M M in 0) E-4
44
M N ýO Ill Q1 N ý1
U) CO dý M M
N
u "M
0 Um
%.0
N Co Ln [- (Na I- r-1
-e
IN
Co
r-1
%D Ln M
O
dý M M ca co ca c N N
dý N dý
b
b
O lý N
t- l0 N LI) CO CO de CO Ill N
M
m N 111
'. O 'D N x
H4,,, O1 M
de
M
IN
O
CO
N
N \D
M
de
M
00
111
0 le 0 %0 4 H
t0
z
a) G4
N t0 Ol H O t11 Ul H (n N
-, Cl in -1 CV r1 O %D N in O 0) Ln
0 fA
"N H M O O l0 l0 r--'H
a
Oý
CO de N (n r-1 d
ti/
t0 [N OD 0) [N H d' M '.O
In m
U
Ql C13 W ON O co [- co r 0 %D Ql
., t0 r (V e-1 cal [- l0 H IN w Ln N
H ký 1D kD
O N O rn Cl C) C4 00 M
rn ON cn N . -1 r-I 01 H in r- 0) In
a a
. r4 W Wz
b H
Z
O fV n 0 H
H Na rn d' in to [, co H H LD4
r-1 r-I
m m 2-14, m H
Lý a tý4 tZ P4 P4 0
o o o o o f
O O N
co
Ql
co 00 t.riO
cr rA
r-1 e-I
4
O O O 0 O Co c O Lt O
O to
N Ol M r1 00
V-4 H
rf
a
w
O O O O O O O O O O O O N H
,.. O O O O 0 0 O O O
0
O O
O lD N
0
H O O O O Co i.n lo
O
Qs N N U) -i N l0 w
H-I rl b a
0
01 ýa
O O O O O O O O O O O .P4
rA
O
... O O O O O O O O O O
O O
w
.0
(D
O (-1 0 O 0 O
N
M
in
N
Co
d'
CO
N l0 0
m
(Yl
. -I r--I E
0
cr [N to N lC . -1 rn C% N ßr-1 de CO 41
ä
o, rl N lO U) le U) 12 N CO N1 rl 4J
0
A
c; N
O ' O O O O dý dý r- i cM
H
p,S w
N O 111 O to Co d' tD [N N
r. Co rl
in N Ul U tO CY% CO tD
0 Ei
H fh CO (fl
t
4-4 Z
w
T N e r1 M M t d' U) [Z L-; U
º1 --
to "M
0 U
. rA
H O r1 - O c+1 Co N Ul lD Co -i
t-A - I- [- M de N
9 E L11 01 Ln H O ýO
O (h
v
a in N -r qr "r a, o, CO N N E
-- r-f
ro
o o N O in N 0% o in rn In Un ro
( O de N O O to 'T tO N CO Co N
a)
O x
H Q LI Ol Ol U1 rn "da H to N () N
H
.. tD ' to N N 10 d' 00 0 rn ' N w0 z
ä
c% N O lM de to to 00
C) aý
Qf [N N Co . -1
.+ N N CO N O d* Co è Ul
ýi C) ^ bl ri M -'
O lfl 01 a0
0
r1
0m
<i lD ri [r rn e-i r-1
-j de N N d' M W
v t0 rn ýO N CO
rd
U
M e CO CO m CO O M U) N CO
t', 1 r. de 00 N 0) tD Ul to
tD ii) N to N
M 01 N N H CO U) 1.4 . -I t! 1 O cl r-
m Co M N N A Co N Ln N O) U)
r-i 14
k cr:
. rj H
ONM de U) QH
lo t`- 00
r-i Np m cp to r-i r-i r-1 rl a
4
[l, ß; Ri Lri [ýi PG fei ýi PG P4' 4ýi fy OH
r6
O O O O o O o O O O O O N
,.. CD O o 0 o O o O o o O O
o
m 1n
O)
o o O O Ln
ýo
O
N
rN o
N
Ln
w
Lf
d'
k
H a)
b
O O O O O O O O O O O O rn
O O O O O O O O O O 0 O
O U1 O O O O tl O tý O U1 LC
%D N N N l0 d+ 44
Q1 0
c-I
r. W
0 Ei
O
0
O O O O O O U1 O 00 111 M M U)
O CO N Co eM O h ýo 0
H T-4 CL
41
O 0 O O O O O O O O O O rn
O O O O O 0 Ll c; 00 111 M M 0
CD
CD N 00 de O N l0
H r-1
cd
N . -" O to tO H N O M O If O H -e a)
() e Lfl CO to LIl Lfl O W N de 00 Cl 41
cd
a ýº tD M M M M N l0 ýD O N l0
u)
Ww
N r-{ N U1 tN 1o H Cl O N O Co Cl aa
O Cl N %O Lfl 111 CO N N %D N
0 E-4
a .ý V-{ e O O O O le de H M e N w
U)
G Um
0
N CO e Cl N H N N Cl 00 l0 u'1 to
" ..
in %D M H Co Co U)
J i Ný c' 4 .4 Cl Ol 00 lD 0 N N a)
`-'
b
N Ol N Cl 01 01 to M Cl H
rd
" fQ . -. F al H
N l0 H O O 0) 0% d* 0)
cA 0 [N CO ri a)
H Q N CO t0 ý.-1 l- O M dý M U dý dý W4
H
to Cl N CO IN Co " Co N nit g:M IN
44 0
H
0 z
N
w Pý
to %D tO 0% 01 CO lD N d' in 00
a
(ia 'r N dý N H ri H t0 M N t- 'i' co tm
lD co O M l0 O 01 0)
0 0
tC V' II1 %D l0 l0 N O dI M in cd
H
cd
.. N [N N CO d' Co Cl to 01 N Lf1 r-
e 0 in in N to A lD -M O t0 M N N
M In c1 M ri N O S 0 0 %D i 00
M CD N N N r-1 CD H in N Ol in
H H M
X P4 PG
W Wz
.rA E-4
Z
ONM CM 0 H
NMe U1 l0 N 00 r-1 r4 1 R.
-1
4 H
WWWW fý W Fr. Wýý fir' fei 0
Appendix 4
N VM .w
Ln
In
M Ln lD
Ol
In
N
tD
-1
d*
lD
lD
0
t.! )
l-
t.[)
CO
t.I
Lf1
In
--1
LI)
CD
--If
as
o -.a1
4-3
rr4 -)
aý 44
W o o
aW^ co
NN
0% c-1
cß'1
0)
r-1
00
ri
Ql
r-1
OMN
MMNNNMMM
01 01 de lle 0 41
U
aaý O
W
N
WU El) 44
r-I
CO
r-i
'd'
[-
L-
f'M
[-
co
t-
co
CO
r-i
Lf)
co
If)
,Ov
lD (-
r-
co
CO
'D
V'
In
Ul
LU 09
0
"rl U
}4
O ".-1
MO R41 O 00 . -1 N d' CV l0 co di -;v N P Q)

U) .0 Ltl %D LC1 l0 d0 l0 Lfl If I! ) %D q r
qc; N .0
E-4
0 Ord
r.
Ol NO Ol O ro :J
CO Ol Nw Ql O d' de N
0O
OW.. NNMN r-I N CV MNNMMMM
M
0 ý0
pu
In 00 OMOO de d' 'O l0 O O. N d'
a "' ZO MNNmMI in lD LN Ci l0 in in
UN O
ýý ý'
H
a
4J
(VI de M le 0 r-I r-1 dI 14.0 l0 (- i-{ O N
U) M tf1 l0 l0 w d$ W It) Lll Lfl LI) de Ul
M "d
ö ...
h
V
H
a .. Co O Cl a1 N O tD Cl N N m de le d*
W
0Iw
E N N m . -i .-I N de M N N N r1 M M
3
wro
. -" N O Co N Co CO U1 CO N P OD lO L11 W
CO de [N [- Co CO t1) U) l0 [N Co l0 L!1 II)
o
41
41
0
roýN
Ul -M Ql 1,O l0 Ul O lD N 00 00 CO N M
Ul (f1 U) l0 l0 l0 -N w Ul tll LI) d+ -e l! ) .r, r
F -
D 44
0 ö r1 k
a% M O 0) lD O 0) lr r, 00 M d+ Nei
N v
Cl N cn e-4 . -1 Cl N M N N N M M M rd r--I
a E-1 I-i pW
(0 (D
(1) 4j ý4
V (is O
,0
fA r-I 44
d' O U1 N CO 01 l0 01 N al Co N d+ N
CO d' CO CO CO CO U) U) tD 00 00 l0 U) lD
Irl
X
.,b-4
U)
Cl m" tl wS co a% .-i ri r-4 r-1 rl a
Ui () U) U) CI) En Cl) Cl) Cl) U) (! ) Ul Gl Cl)
ýa
Appendix 5
STANDARD POSITION
1 2 3 4 5 6 7
sample 1
pts 1-7 x xx
pts 8-14 xx x xx
pts 15-21 x x x
pts 22-28 xx x xx xx x
sample 2
pts 1-7 x
pts 8-14 xx x x
pts 15-21 x x xx
pts 22-28 x x xx xx
sample 3
pts 1-7 x x xx
pts 8-14
x xx
pts 15-21 x xx x x
pts 22-28 x x xx
sample 4
pts 1-7 x xx xx x x
pts 8-14 x x
pts 15-21 x xx x
pts 22-28
x xx
sample 5
pts 1-7 x xx x x
pts 8-14 x x xx x
pts 15-21 x xx x
pts 22-28 x xx
sample 6
pts 1-7 x xx x
pts 8-14 x xx x x
pts 15-21 x x xx x
pts 22-28 xx x xx
x: deformation
xx: large deformation
xxx: very large deformation
Appendix 5.1 Points of deformation across four layers for

each of the six samples for Rigidex welds.
R2 POSITION
1 2 3 4 5 6 7
sample 1
pts 1-7 x xx x x
pts 8-14 x x xx
pts 15-21 x x xx x
pts 22-28 xx x x
sample 2
pts 1-7 x xx x x xx
pts 8-14 x x x xx xx
pts 15-21 xxx xxx xx xxx xx
pts 22-28 x x x x
sample 3
pts 1-7 x xx x xx x
pts 8-14 xx x x
pts 15-21 xx xx
pts 22-28 xx x x x x x
sample 4
pts 1-7 x xx x xx
pts 8-14 x xx xx
pts 15-21 x x x x xx
pts 22-28 x x xx x
sample 5
pts 1-7 x x xx xx
pts 8-14 xx x x x
pts 15-21 xx x x
pts 22-28 x xx x x
sample 6
pts 1-7 x x xx x xx
pts 8-14 xx x xx x
pts 15-21 x xx x x
pts 22-28 xx xx xx
x: deformation

R3 POSITION
1 2 3 4 5 6 7
sample 1
pts 1-7 x x
pts 8-14 xx x x x
pts 15-21 xx x x x xx x
pts 22-28 x x x
sample 2
pts 1-7 x x x x
pts 8-14 x x x x
pts 15-21 x xx x
pts 22-28 x xx x
sample 3
pts 1-7 xx x x x x
pts 8-14 x xx x xx
pts 15-21 x x x
pts 22-28 x3oc x
sample 4
pts 1-7 x x x
pts 8-14 x x xx x
pts 15-21 xx xx x
pts 22-28 x x x xx
sample 5
pts 1-7 x x
pts 8-14 xx
pts 15-21 x xx x xx
pts 22-28 x xx
sample 6
pts 1-7 x xx
xx
pts 8-14 xx x x
pts 15-21 xx x x x
pts 22-28 x xx x x xx
x: deformation

R4 POSITION
1 2 3 4 5 6 7
sample 1
pts 1-7 x x xx
pts 8-14 xx xx x
pts 15-21 xx xx xx
pts 22-28
xx
sample 2
pts 1-7 xx xx
pts 8-14 x xx x
pts 15-21 x xx x x
pts 22-28 x _
sample 3
pts 1-7 x xx xx
pts 8-14 xx x x x
pts 15-21 x xx x
pts 22-28 x x x
sample 4
pts 1-7 x xx xx
pts 8-14 x xx x
pts 15-21 x x x xx
pts 22-28 x xx xx
sample 5
pts 1-7 x xx x xx
pts 8-14 x x xx x
pts 15-21 x x xx x
pts 22-28 x xx x
sample 6
pts 1-7 xx x x x x
pts 8-14 x x xx x
pts 15-21 x x xx x
pts 22-28 xx
x: deformation

R5 POSITION
1 2 3 4 5 6 7
sample 1
pts 1-7 x x
pts 8-14 xx x xx x
pts 15-21 x xx xx x
pts 22-28 xx xx
sample 2
pts 1-7 x xx x x xx
pts 8-14 x xx x
pts 15-21 xx x
pts 22-28 x x x xx
sample 3
pts 1-7 x x xx
pts 8-14 x x xx x
pts 15-21 xx xx x
pts 22-28 x x xx
sample 4
pts 1-7 x xx x xx
pts 8-14 xx x x xx x
pts 15-21 xx x xx
pts 22-28 x x x xx
sample 5
pts 1-7 xx x xx x
pts 8-14 xx xx xx x
pts 15-21 x xx
pts 22-28 x xx x x xx
sample 6
pts 1-7 x xx xx xx
pts 8-14 xx x xx x x
pts 15-21
pts22-28 xxx x xx x
x: deformation

R6 POSITION
1 2 3 4 5 6 7
sample 1
pts 1-7 xx x xx
pts 8-14 x x x xx x
pts 15-21 x xx xx
pts 22-28 xx xx x xx
sample 2
pts 1-7 x x xx
pts 8-14 xx x x x
pts 15-21 xx x xx xx x
pts 22-28 x xx x
sample 3
pts 1-7 x x x x x
pts 8-14 xx x xx
pts 15-21 xx x
pts 22-28
x xx
sample 4
pts 1-7 x x xx
pts 8-14 x xx x x
pts 15-21 roc x x x
pts 22-28 x xx x xx
sample 5
pts 1-7 x xx xx xx x x
pts 8-14 xx x x
pts 15-21 x xx
pts 22-28 x x x xx
sample 6
pts 1-7 x xx xx x
pts 8-14 x x xx xx x
pts 15-21 x xx x
pts 22-28 x x xx
x: deformation

R7 POSITION
1 2 3 4 5 6 7
sample 1
pts 1-7 xx x x x
pts 8-14 x xx x xx x
pts 15-21 x xx xx xx x
pts 22-28 c x
sample 2
pts 1-7 x x x x
pts 8-14 x xx x x
pts 15-21 x x xx
pts 22-28 x x x xx
sample 3
pts 1-7 x x x xx x
pts 8-14 xx xx
pts 15-21 x xx xx
pts 22-28 x xx xx x
sample 4
pts 1-7 xx x xx x
pts 8-14 xx xx x x
pts 15-21 xx x x
pts 22-28 xx x x x x xx
sample 5
pts 1-7 xx x x
pts 8-14 xx x xx x
pts 15-21 x xx x
pts 22-28 xx x xx xx
sample 6
pts 1-7 xx x
pts 8-14 x x xx x
pts 15-21 x x xx
pts 22-28 xx x
x: deformation

Appendix 6
O
O
to
I-'
C
o-S
C3.
U)
C
U
L
N
CL
0
0
(V
0 0 0
0 0
N
0
(T)
°dW ssaJgS
ra I
In
N
a)
4.3!
Wi
0
o
,ý
IN# cI
(D'
J 4. )I
C_
r-i
d WI
L
"V t'1
min
0l
44'
4)
L
cr m
U .ýi
L
Ui m
z CL'
M'
-I
.UI 4J
oho
D U)
Dý
N
a)
N
N
n
v
l0
x
ri
b
0 0 0
0 N 0 a
(T) a
ego-S
DdW
u
.Q1
411
ö
C) "
C
W'
O r. r
CD cn:
v
K
C
4J
d W..
L
Cl) 0
441
04.3 11
to r- 04
G
U
L
N
0.. 43
UI
0x
N
D
d'
a).
:>I
N
N
1!
4.), ,
D
N
x
.b:
r.,
0 o 0
O N 04
4
9:
m
DdW ss :s
o
o
"FI
U.
a'
Cl.. >
In
N a
C "b'
ýi "r
L
+-)
N
= 441
U
L "ri
Nb
00 o- },.i
N
in
oxa
(A
ON
O
t0
x
.d
b
OO Ki
a)
N a
a
DdW 9SBJ-4S
V.,
rcl
RI
al
Cl
.
C . rf
°N 31
r
Il
O
Sri
LT
V--f
.r
L
+) 0l
N 441
4 rý
C
Co rl
LP
Co 4.).
CL , t11_
O
4 o\0
a)
N
N
a)
O 4J
Ln M
I)
tC
x
.br.,
a)
0 0 a
04
0:4
N
0
odW ssaj, 4s
Aý
4
0 t
f
. ii r
0
0
N b
a, 1
t
r-I
n x
K aý
I) i
C a
",,
L
w
0
LA
44
43
C
CD
U
L
N N
o\0
H
N
N
U)
a)
41
I.O
ýD
:4
.br.,
a
00
d0
I
'. 1
b
J
o
w
Ln
N
,a
co ý
ý' .
U)
ra
4J W
14
dl ,ý
Uw
VA 0
.rI
to
OR
N
0
0 N
T-
a)
N
N
a)
4J
tG
X
.r,
93
0 0
0 0
aý
N
Ddw ssa-j-4s
Id
Q)
O
OO
m4
3
. r. {
rd
r1
r
O
ý..ý
0t
rd
a)
cx
L
N
O
Ö0 4a
m0
L "ri
0u nj
Cp
0'0
U)
Q U)
NN
a)
N
N
4.3
D U]
O
r-t
CD
.0
x
.br.,
a 0 a)
O 0 a
N a
DdW 99DJIS

Hdpe Butt Fusion Welding Theory

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hdpe Butt Fusion Welding Theory

Uploaded by

Copyright:

Available Formats

BUTT FUSION WELDING OF POLYETHYLENE

A thesis submitted for the degree of

DEPARTMENT OF MATERIALS ENGINEERING

In order to achieve high quality thin films from microtomy,

The presence of the bead and the asymmetry in the test

1.3 Outline of the thesis 5

CHAPTER 2 LITERATURE REVIEW 6

2.2 Effect of processing on polyethylene pipes 7

2.2.1 Thermal Oxidation 10

2.4 Degradation and stabilisation of polyethylene 14

2.5 Evaluation of thermal degradation 16

2.5.1 Thermal analysis 17

2.5.2 Oxidation induction time (OIT) test 17

2.6.1 The butt fusion welding process 22

2.6.2 Theory of butt fusion welding 25

2.6.3 Microstructure of plastic welds 28

2.6.3.1 Specimen preparation 34

2.6.4 Factors influencing butt fusion welding 35

2.6.5 Temperature distribution within the weld 40

2.6.8 Mechanical properties 43

2.6.8.1 Mechanical testing 44

CHAPTER 3 EXPERIMENTAL TECHNIQUES 50

3.0 Pipe materials so

3.1 Mater ials characterisation 50

3.1.1 Density gradient column 51

3.1.2 Gel permeation chromatography 52

3.2 Butt fusion welding 55

3.2.1.1 Thermal mapping of heater plate 56

3.3 Micro structure 58

3.3.2 Polishing of blades 64

3.3.2.2 D-profiled steel knives 64

3.4.1 Weld bead measurement 66

3.4.2 MAZ-Internal weld dimensions 68

3.5 Tensile testing 68

3.5.2 Macro-tensile specimen preparation 69

3.5.3 Micro tensile testing procedure 71

3.5.4 Macro tensile testing procedure 72

CHAPTER 4 MATERIAL CHARACTERISATION 73

4.3 Molecular weight determination 74

4.4 Thermal analysis 75

4.4.1 Differential scanning calorimetry 75

4.4.2 Oxidation induction time (OIT) 82

4.4.2.1 Sample mass 87

4.4.2.2 Volatility of antioxidants 91

4.4.2.3 Electron spin resonance (ESR) 92

4.4.2.4 X-ray photoelectron spectroscopy (XPS) 93

4.4.2.5 Surface damage 94

CHAPTER 5 WELD GEOMETRY AND DIMENSIONS 102

5.1 Introduction 102

5.2 Sample coding 102

5.3 Bead geometry visual assessment 104

5.4.1 Bead widths 110

5.4.2 Bead areas 113

5.4.3 Melt affected zone (MAZ) quantitative 120

CHAPTER 6 MACRO AND MICROSTRUCTURAL EXAMINATION 133

6.1 Introduction 133

6.2 Microtomy 133

6.2.1 Chemical polishing 134

6.2.2 Manual polishing 135

6.3 Macro-structure 136

6.4 Microstructure 147

6.4.1 Microstructure of the extruded pipe 147

6.4.2 Microstructure of the MAZ 152