Thermoanalytical study of an aged XLPE-pipe

115

Thermochimica Acta, 114 (1987) 115-124 Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

THERMOANALYTICAL

STUDY OF AN AGED XLPE-PIPE

E. FIGUEROA K. Studsvik Energiteknik

AB, Materials

Technology,

S-611 82 Nykoping

(Sweden)

SUMMARY A high quality pipe made of XLPE (Engel method) for industrial applications was used for the study of the material deterioration process upon accelerated aging. A sample aged at llO°C while submitted to an hydrostatic pressure of 0.53 MPa, giving a hoop stress of 2.66 MPa, was studied by means of Thermal Analysis, Density, and Gel Content measurements. The effects of aging are compared to a reference unexposed sample and the experimental results are discussed in terms of annealing effects, free volume and structural relaxation processes.

INTRODUCTION High quality materials

crosslinked

that, despite

during manufacturing,

of advanced

present

the high density

process,

with a peroxide

techniques

polyethylene

(Engel method).

Peroxide

in the polymeric

radicals

carries

an unpaired

tional bridges forward

chemical

conditions materials

in turn crosslink

in a recombination reaction

during

the manufacturing

and the pipe forming

system, will also affect processing

conditions

random distribution throughout During

structural

the material

molecules

changes

process,

0040-6031/87/$03.50

crosslinks

straight-

to the thermodynamic the mixing

of the raw

present

and homogeneity.

in the

Thus,

that there is a perfect homogeneously

the XLPE-pipe

This thermal

throughout

at the molecular

spherulitic

of the chain.

distributed

(ref. 2).

themselves

embrios.

is formed and then rapidly

quench

results

in not only a

the pipe wall, but also results

level. From the structural

in randomly

lamellae with no or low macroscopic incipient

quality

al-

macro-

of tetrafunc-

as well as additives

the final material

activity

rearrange

the building

Besides,

process.

of intermolecular

throughout

(ref. 1). This seemingly

very sensitive

is treated

polymeric

on an inner member

affect the basic assumption

the manufacturing

chemical

forming

conditions

cooled from the melt with water. different

dissociation

throughout

control

5261-Z)

homolytic

reaction

is obviously

(Lupolen

are

In the production

environment,

electron

applications

process

inhomogenties.

raw material

reacts,

These macroradicals

and rigorous

some structural

koxy radicals, which

(XLPE) for industrial

polyethylene

oriented

orientation

The stacking

stacks of parallel

which

of several

0 1987 Elsevier Science Publishers B.V.

in

point of view,

can be regarded lamellae,

as

parallely

116 ordered,

produces

local orientations

talline microstructure fore it can be affected pipe, leading,

tallization material,

evolves.

and the cooling

they can result

activated

behaviour

crosslinking

as a paracrys-

surface

of the

of the peroxide

leads to a reduction

crosslinking,

in local fluctuations

These kinds of inhomogeneities

should,

crys-

of the HDPE raw

process,

are such that

crystallinity,

of material

as the

of further

conditions

rate after the pipe forming

treated

of density

the development

XLPE. The processing

in an inhomogeneous

This should be reflected XLPE-pipes.

at the external

This in turn obstructs

of the resulting

regarded

from the melt, and there-

to a higher degree of amorphicity.

the chemically

raw HDPE (density of about 0.95 gcms3) crosslinking

evolves

by the rapid cooling

for example,

Additionally

that are normally

(ref. 3). This process

density,

properties

in turn, affect

etc.

in the

the aging

of the material.

The aim of the present experimental

parameters

order to gain insight

research was to study the distribution

throughout

the pipe wall,

into the understanding

of some

in a comparative

of aging processes

study in

of XLPE.

EXPERIMENTAL High quality thickness

XLPE-pipe

with an overall

diameter

of 110 mm and a nominal

of 10.0 mm, were aged in an oven with circulating

pipes were submitted

to 0.53 MPa hydrostatic

pressure

air at llO+Z°C.

The

giving a hoop stress of

2.66 MPa. A sample that failed after 17.136 hours was chosen for the purpose of the present

comparative

study with respect

to a reference,

The aged sample shows a rupture with a multiple

unexposed

was initiated

at the internal wall side. That type of failure

to be typical

of final chemical

rupture mechanism

long term testin of plastic The present (crossliking

deterioration

has been classified

pipes for industrial

study comprises

of Thermal

degree) measurements

Density,

THERMAL

of the measured

and Gel Content

the pipe wall. The measurements

were done on a number of samples that was considered tative distribution

study of

use (ref. 4).

Analysis,

throughout

This typical

in the experimental

III

that

has been proven

of aged polymers.

as Stage

pipe.

brittle microfracture

paremeters

enough to have a represen-

throughout

the pipe wall.

ANALYSIS

A Mettler

thermoanalytical

instrument

was used to study the glass transition both reference

and aged XLPE-pipes.

in the present work, determined

with a cryogenic

temperature

The glass transition

is the temperature

at the enthalpic

in XLPE at about 153 K (usually

DSC measurement

of samples

selected

temperature

cell

from

of XLPE

change that can be

refere as -12O'C transition).

Samples were cut from the inside, the middle

and the outside

of the pipe

117

wall

in the form of a disk with a diameter

0.5 mm. DSC analysis temperature

was also performed

on the same set of samples

up to 523 K. The DSC results,

T n, and the glass transition a The measurements

of about 5 mm and a thickness

temperature,

the change

the sample to liquid nitrogen

temperature

were performed

thermogram

at the glass transition

the shift in the base line, the change temperature

known relationship

changes

were registered

in the heat capacity

them at in the

and, based on

and the transition

for the aged sample can be discussed

in terms of the well

(ref. 5).

- KIM,

gm

In eqn. weight,

1.

were calculated.

The Tg results

Tg = T

The entalpic

temperature

heat at

by first cooling

(77 K), and then heating

rate of 10.0 K/min, up to the room temperature.

from room

in the specific

are given in Table

at the glass transition

of

(l), K is a constant,

and T

age molecular deffined,

is the limiting

gm weight

Mm is the average

for XLPE, and consequently

T

=T

-T

for the aged sami?e to':he respective

molar masses,

weight.

providing

that the ratio of the T

is inversely

that the constant

rela-

of the polymer molar

, it can be demonstrated sample

an aver-

case as a rather empiric

as representative

uzexposed

Although

Tgm, can not be unambiguosly

eqn. (1) would be used in the present

tion. The Mm value may be considered mass. Defining

value for the molecular

Tg at high molecular

proportional

K is unaffected

to the

ge

upon

aging.

(2)

Tge(aged)/Tge(ref.) = Mm(ref)/Mm(awd) With this relationship, can be estimated.

The calculation

pipe, the Mm is reduced the reduction

the relative

indicates

that at the inside of the aged value, while at the middle

is only to 99.4%.

first scan, however, This experimental

at the outside

of the aged pipe was not found on the

a second scan showed a transition

result

is affected

be used in the same manner

by the thermal

for the discussion.

is known to introduce

cation of a molar mass change.

that could be evaluated.

cycling,

Nevertheless,

a shift in Tg towards

can still be used in eqn. (2) to calculate

shows,

in the molar mass upon aging

to 97.7% of the original

The glass transition

cycling

change

higher

and thus cannot since the thermal

values,

a ratio that provide with an indi-

The calculated

ratio for the outside

indeed, a rather high value for the molar mass reduction,

93.7% of the original

value, while

On the other side, the expected part of XLPE on aging,

this result

a reasonable molecular

of the pipe

a reduction

value should be around 99%.

reacommodation

can also be a reason for the observed

of the amorphous change

in Tg

to

118

(ref. 6). This entanglement fore in Tg changes.

will result

In addition,

changes

also affect the values of Tg throughout lated relative

changes

in a free volume

reduction

and there-

in the degree of crystallinity the pipe wall. Consequently,

should

the calcu-

in the molar mass may be overestimated.

Fig. 1. (left) The glass transition temperature at the inside, the middle, and the outside of the pipe for the unexposed reference sample. The results above and below the T values represent the temperature width of the transition. Fig. 2. (right)gThe glass transition temperature at the inside , the middle and the outside of the aged sample after 17.136 hours at llO°C and a hoop stress of 2.66 MPa. The T at the outside of the pipe were obtained at the second scan and therefore thisgresult is affected by thermal relaxation. DSC results,

at the same sample positions,

scans were obtained and evaluated

with the same equipment

by the computerized

for the onset of exothermal geneous

distribution

unit.

entalpic

throughout

are presented

at a heating

change

at oxidation,

the unexposed

The endothermal

melting

phase transition, while

similarly sample

distributed

crease of crystallinity crease

change

for both reference

shows an increase

in the relative experienced

and the inside pipe wall sides.

Tm, is also homogeneous

in aged samples

sides of the pipe wall. The relative

To,, shows a homo-

pipe wall, while the aged sample

drop of Tax at both the outside

material,

1. DSC

It can be noted that the temperature

shows a drastic

the reference

in Table

rate of 10.0 K/min,

throughout

it shows higher values

in (gravimetric)

and aged samples,

at both

crystallinity however,

is

the aged

percent change by about 23%. The in-

by the aged samples, may also cause an in-

in Tg (ref. 7). These results are consistent

since they correlate

well

119

with the deterioration nal surface. during

shown by the aged pipe, at both the internal

It is evident

that several

aging. The possibility

standing

competing

of uncoupling

of the aging process

effects

these effects

are taking

and exterplace

for a better

under-

(ref. 8), is far beyond the scope of the present

work. TABLE

1

Summary of thermoanalytical results for both the reference material and the aged sample. Results at the internal surface, in the middle, and at the external side of the pipe wall are presented.

Parameter

Position

Reference

Temperature T ox

for the Onset of Oxidation Inside Middle Outside Melting Temperature Inside Tm Middle Outside Relative Cristallynity K Inside Middle Outside Glass Transition Temperature T Inside g Middle Outside Change in the Specific Heat Inside AC P Middle Outside

*

These values were obtained thermal relaxation.

sample

Aged sample

217.8 224.4 223.0

'C 'C 'C

183.0 218.0 198.5

'C 'C 'C

129.8 130.2 130.0

'C 'C Oc

135.8 131.2 133.8

'C 'C 'C

49.5 48.8 47.0

z %

59.8 59.8 59.5

% % %

- 118.6 - 119.5 - 120.6

Oc Oc 'C

- 115.0 - 118.6 (- 110.4

0.175 0.127 0.064

J/N J/N J/N

at a second run, therefore

Oc Oc Oc)*

0.196 J/gK 0.149 J/gK (0.266 J/gK)* they are affected

by

DENSITY MEASUREMENTS The density, by the column 2782-70,

Method

instrument, the density

of the samples was determined

gradient

density

at a temperature

(ASTM D1505-85

509B, and B.S. 3715-64).

The column,

of the final sample positioning

are believed

to accurate

to +0.00005

were cut in form of disks with an average weight and aged samples

cal samples giving

a Davenport

Standard

B.S.

Ltd. (London)

of ethanol and distilled water, was calibrated for -3 range 0.92 to 0.98 gem . Due to careful degassing of the mixture

measurements

reference

of 23.00f0.05'C

or British

using a mixture

and the long term stability density

method

cut through

respectively.

in the column, the -3 . The samples

gem

of 8.4 mg and 7.9 mg for the

The disks were obtained

from cylindri-

the pipe wall. The samples were about 0.5 mm thick,

thus a good representation

of the density

distribution

throughout

the

120

pipe wall thickness good reproducibility

(10.0 mm nominal

thickness).

The experimental

results

shows

upon verification.

Fig. 3. (left) Density distribution throughout the pipe wall for both the reference and the aged sample. The results were obtained by the density gradien column method at 23.00+0.05°C. Fig. 4. (right) Differential reduced density change in percentage, throughout the pipe wall thickness.

The results of density for one set of samples differential

change

measurements

in the density

of the degree of compactation the pipe to a high hydrostatic The reference

unexposed

First, the density sample.

Secondly,

ward the outside

are shown in Fig. 3, and were obtained

for each material

category.

upon ageing.

of the material pressure

material

values

is a measurement

due to the long term exposure

under isothermal

shows two major

values show a somewhat the density

Fig. 4 shows the reduced

This change

aging.

interesting

higher spread compare

features. to the aged

show a profile with lower densities

of the pipe. This scattered

density

distribution

to-

throughout

of

the reference

material

of the PEX material

is believed

during

to be related

the processing

to an inhomogeneous

The aged sample shows a lower spread of density wall and also a much lower density and the bulk of the material. appears

a large increase

tion has been shifted

However,

towards

The shift of the density discussed

difference

in density.

towards

density

change,

that result during manufacturing,

volume fraction

f = (V-V(T,t))/V,

activation

approximation,

distribu-

where

higher values can be

Fig. 2. Voids and intermoiecan be described

V is the specific

and time. Molecular

are related.

and is given

the pipe

of the pipe

the total density

of the aged sample towards

of temperature

throughout

the outside

the inside of the pipe, there

In addition,

cular cavities,

and thermal

values

between

higher values upon aging.

in terms of the relative

is a function

formation

of the pipe.

The relationship

in the following

volume,

mobility,

by the freeand V(T,t)

free-volume

fraction

is known as the Doolittle

equation

(ref. 9).

In M = A - B/f

(3)

This expression the increasing

shows that the molecular

degree of packing,

eqn. (3) A and B are constants. glass transition

temperature,

mobility

or the reduction

The free-volume

T g, through

M clearly

decreases

of the free-volume

fraction

the following

f is related relationship

with

f. In to the (ref. 9).

f(T) = fg + Aa(T - Tg) Where f

denotes

tween the 4 cubical)

(4)

the free-volume thermal

fraction

expansivities

Using eqn. (4), a first estimation at the internal

at T

and Aa the difference 9' above and below T .

of the relative

that the thermal

expansivities

for the Tg values at the middle

change of 0.6%. These

do not change

results,

although

gree of cristallynity.

compared

somewhat

trated

density

change

the outside

upon aging. The

of the pipe gives a relative

to the relative

of a unaffected

density

change

(Fig. 4)

thermal

expansivity

of

investigations.

after 17.136 hours aging at llO°C, as illus-

in Fig. 4, shows a densification

Towards

significantly

high. This can be due to the raise in the de-

The assumption

XLPE with aging call for further The relative

in the free volume

pipe wall side, gives a result of 2.5%, base on the assumption

same calculation

are consistent,

changz

be-

of about 0.4% on average

for the bulk.

of the pipe, values of about 0.6% for the relative

densi-

fication

are most likely due to the high values for the initial free volume

fraction

of the material

due to the rapid cooling

ment. The value of the relative

density

change at the internal

about 1.8%, is not only due to the free volume structural

changes.

Electron

microscopy

after the pipe forming

fraction

studies,

equip-

pipe wall side,

changes,

but also to

(ref. lo), reveal

the onset of

122

a thin layer of degraded, pipes. The texture bril structure,

indicating

close packed structure process

oxidized

a different

is develop,

onsets upon ageing,

cracks and failure

polymer

as a common feature

of this layer, at fracture

surface,

molecular

where oxidized

with a resulting

configuration. molecules

weakness

of aged polymer

shows little or any fiA high density,

and segmentation

that allows

for micro-

initiation.

GEL CONTENT The Gel Content crosslinking,

of the samples,

was determined

used decahydronaphthalyn

an indirect measurement

by the ASTM D2765-68

(Dekalin)

standard

of the degree of method.

as the solvent with 1% addition

thylen-bis-(4-methyl-6-tert-buthylphenol)),

The method of (2.2-Me-

an antioxidant.

Fig. 5. (left) Gel Content of the unexposed reference XLPE material obtained by the method described in ASTM D2765-68. The shaded area indicates the variance around the mean value. Fig. 6. (right) Gel Content of the aged sample of XLPE, after 17 136 h, exposed to an hydrostatic pressure resulting in a hoop stress of 2.62 MPa, and a temperature of 11O'C.

123 Samples sentative formed

were selected distribution

and the average

ence sample,

the pipe wall

of the gel content.

Two sets of measurements

value are presented

in Fig. 5 for the unexposed

XLPE material

deviation

sample was 77.6%,

shows an average

of 3.3. The average and the standard

percent

reduction

measured

values

It is worth

throughout

the pipe wall after aging

duced thermal

of the reference

sample

stability

and mechanical

surface

face of several Obviously,

is in agreement

XLPE qualities

of the material,

the changes

However,

can be discusen

exhaustive

order to have a more detailed

structure

at sur-

(ref. 10).

and consequently

in terms of the glass tempera-

termoanalytical

experimental

(ref. ll), and

of the rupture

of ageing

upon aging in the gel constent,

in the degree of crosslinking, ture changes.

degrees

in a re-

applications,

of fibril

in SEM studies

to a

that a

will result

at technical

with the absence

at various

values at

value for

(about 10.8% compared

strength

layer, as demonstrated

average

This is an indication

of the long term with standability

of

(about 18.2% better).

of the XLPE due to chain scissions,

(ref. 4). This result

result from

in the ditribution

to the resulting

of only 2.8% in the average).

molar mass reduction

of the aged

that the aged sample shows lower Gel Content

the total cross section

the internal

refer-

of 79.8% with a

2.9. The relevant

homogenization

pipe wall side, with respect

total reduction

consequently

were per-

degree of the aged sample shows a relative

of 2.8%, and a relative

noticing

gel Content

value for the Gel Content

deviation

this study is that the crosslinking

the internal

in order to obtain a repre-

and in Fig 6 for the aged sample.

The reference standard

throughout

studies

are in progress

in

support.

DISCUSSION Thermal

Analysis

single kinetic

pipes. The density increased

results

show that in general,

deterioration

process during

increase

by two different

shown by the sample upon aging can apparently routes.

Either by means of an increase

linity at a slow crystallization

rate during

a free volume

to a partially

relaxation

leading

both cases the density will the microhardness, The Gel Content

increase

reduction

itself

Although

aging or, by

amorphous

phase.

the elastic modulus

In and

to argue

at different

of aging

is assumed

to be due to

a higher degree of changes were expect-

of the pipe wall,

is not accurate

It seems to be possible processes

compressed

(and consequenlty

as a consequence

process.

ed in boths the inside and outside

ioration

the initial physical

be

of crystal-

(ref. 12).

a slow chain scission

that the method

it seems to be more than a

the long term aging test of XLPE

it should be remembered

enough for a more quantitative

in favour

chemical

of multiple

kinetics,

where

cooperative

analysis. deter-

the posibility

of un-

124

coupling

the various mechanism

acceleration

ageing

The present meters,

long and careful

and Gel Content,

as useful1

between

parameters

at molecular

the thermoanalytical

parameters

of the effect of aging upon XLPE. Attempts

ation, made clear the need for further ration processes

studies at several

rates.

study shows the relationship

density

scription

would demand

in the qualitative

at quantitative

insigth on the consequences

level, so that correlations

between

parade-

evaluof deterio-

different

may be calculated.

ACKNOWLEDGEMENT Thanks

are due to my colleagues

discussions I Cohen,

during the preparation

Elektrolux,

Stockholm,

M. Ifwarsson

and P. Eriksson,

of the manuscript.

For providing

Thanks

assistance

for fruitful

are also due to

with the low tempera-

ture measurements.

REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14

R. Rado, Int. Pol. SC & Tech., 3(8) (1976) T/107. A. Charlesby, J. Polym. Sci, 11 (1953) 513. J. Martinez Salazar and F.J. Balta Calleja, J. Cryst. Growth, 48 (1980) 283. M. Ifwarson and P. Eriksson, Kunststoffe, 76(3) (1986) 245. T.G. Fox and P.J. Flory, J. Appl. Phys., 21(1985) 581. E.A. Turi, Thermal Characterization of polymeric Materials, Academic Press, New York, 1981, pp. 248-251. E.A. Turi, Ibid, pp. 254-255. E.A. Turi, Ibid, pp. 290-291. L.C.E. Struik, Physical Aging in Amorphous Polymers and Other Materials, Elsevier, Amsterdam, 1978. F.J. Balta Calleja, J. Martinez Salazar, H. Cackovic and J. LobodaCackovic, J. Mater. Sci., 16(1981) 739. T.G. Fox and S. Loshaek, J. Polymer Sci, 15(1955) 371. P. Eriksson and M. Ifwarson, Kunststoffe, 76(6) (1986) 512. Ch. Hall, Polymer Materials, Macmillan Press Ltd, London, 1981, pp. 128-131. F.J. Balta-Calleja, D.R. Rueda, J. Garcia-Pena, F.P. Wolf and V.H. Karl, J Mat. Sci, 21(1986) 1139.