a collaborative study of the dynamic mechanical and …

INTERNATIONAL UNION OFPURE AND APPLIED CHEMISTRY

MACROMOLECULAR DIVISION

A COLLABORATIVE STUDY OFTHE DYNAMIC MECHANICAL

AND IMPACT PROPERTIESOF PVC—I!

A Report of the JUPAC Working Party on 'Structure andProperties of Commercial Polymers'

Prepared Jbr publication byA. GONZE and J. C. CHAUFFOUREAUX

Solvay and Cie, Brussels, Belgium

LONDON

B UTTERWORTHS

315

PAC—35—3F

MACROMOLECULAR DIVISION

A COLLABORATIVE STUDY OF THE DYNAMICMECHANICAL AND IMPACT PROPERTIES OF

PVC

ABSTRACT

The paper summarizes the work done by a working party of IUPAC ondynamic mechanical, tensile and impact properties of a rigid and a toughenedPVC. Storage modulus and loss factor of both PVCs have been measured overa wide range of temperatures and frequencies in torsion pendulum andflexural vibration tests. A secondary transition of PVC and the glass transitionassociated with the CPE present in the toughened PVC have been locatedwith fairly good agreement between the five participants. Charpy, fallingweight, Izod, tensile and prestressed impact tests have been utilized over awide range of temperature to study impact properties. The brittle—toughtransition temperature range of both PVCs has been located for each testmethod and the results were discussed.

Tensile properties of both PVCs were specially studied in the linear rangeof deformation, at yield point and at rupture. Yield stress measurements arein good agreement with the generalized Eyring theory proposed by Bauwens—Crowet. Rupture strains show two transition zones; the first one, associatedwith a tough—brittle process, has the same activation energy as the secondarytransition found by dynamic mechanical measurements; the second one,associated with a tough—tough process, seems to be strongly correlated withthermal dissipation in test pieces.

The relaxation modulus of each PVC has been calculated over eight decadesof time at several temperatures from 20 to 64°C on the basis of dynamicmechanical, tensile modulus and stress relaxation measurements.

The identity of the activation energy measured for the transition zones bymeans of dynamic mechanical, tensile and impact tests respectively shows thatthe molecular relaxation processes measured in the linear range of deforma-

tion of PVC, have a strong effect on the impact or tensile properties.

INTRODUCTION

The present paper covers .the second part of a work done by an IUPACWorking Party, under the chairmanship of J. W. Barrett, on the dynamicmechanical and impact properties of a rigid (normal) and a toughenedpolyvinyl chloride (PVC). The objective of the work was to study the effectof molecular parameters of polymers on those properties.

The first part of the work on PVC was published in Pure and AppliedChemistry, the official Journal of IUPAC, in 19698. It was concerned withthe measurements, by means of a torsion pendulum, on rigid and toughenedPVCs supplied by Solvay and Cie. It also included some impact tests.

This first part was carried out by I. Franta, Technical University of317

A. GONZE AND J. C. CHAUFFOUREAUX

Prague; J. Heijboer, TNO, Deift; S. Baxter and T. T. Jones, MonsantoChemicals Ltd, England; G. Pezzin, Montedison, Italy and A. Gonze,Soivay and Cie, Belgium.

While reasonable agreement has been reached for the dynamic mechanicalproperties, the impact tests varied widely in the rate of elongation andspecimen shape and size. No conclusion about correlations could be drawn.The importance of working with samples prepared in one laboratory waspointed out.

The second investigation, using samples prepared by Solvay with rigidand toughened PVCs, was concerned with a more detailed study of thedynamic and standard mechanical properties.

For this further work additional collaborators joined the Working Party:J. Zelinger, Technical University of Prague; M. Chatain, CEMP, Paris; H.Oberst, Farbwerke Hoechst, Germany and W. Retting, BASF, Germany.

This paper summarizes the work done by the working party and the agree-ment reached in similar experiments, conducted in separate laboratories,on identical materials. A tentative attempt is made to correlate the resultsof the dynamic mechanical tests with those of the relaxation, tensile andimpact tests.

This paper is divided into five sections:1. Results of dynamic mechanical tests in a wide range of frequencies and

temperatures2. Results of impact tests3. Results of tensile tests4. Results of relaxation tests5. Correlation between the testsOn each diagram, the results of individual participants are identified by

means of a roman figure.I for BASF

Ii for CEMPIII for HoechstIV for MonsantoV for Montedison

VI for SolvayVII for Technical University of Prague

VIII for TNO

1. RESULTS OF DYNAMIC MECHANICAL MEASUREMENTS

Materials, equipment and conditions of testingAll the results obtained in the various laboratories for the dynamic

mechanical measurements, during the first and the second part of the workare given here.

The samples used are a rigid PVC (suspension PVC in a lead—cadmiumstabilized formula) and the same rigid PVC with 12 parts per hundred (pph)of chlorinated polyethylene (CPE) added.

During the first part of the work, the Specimens were prepared in eachlaboratory from granules furnished by Solvay. For the further work thespecimens were cut from sheets prepared in the Solvay laboratory.

318

DYNAMIC MECHANICAL AND IMPACT PROPERTIES OF PVC

Both PVCs, rigid and toughened, were studied at a number of frequenciesin the range 0.2 to 2 Hz on a torsion pendulum and in the range 25 to 3700 Hzon devices for the flexural vibration test. For this test four laboratories usedthe complex modulus apparatus, type 3930, Brüel and Kjaer, Copenhagen.Information about the apparatus and test conditions used by each laboratoryis given in Table 1.

The measurements on a torsional pendulum are well known. They aredescribed in the ISO recommendation 537. For the flexural vibrationexperiments on the Brüel and Kjaer device, the specimen is clamped at theupper end, in a vertical position, and is free at the lower end. The free lengthof the sample varies between 12 and 25 cm depending on the temperature andthe frequency. At the lower end, the bar is excited into stationary bendingvibrations by means of an electromagnetic transducer. The vibration ampli-tudes are measured by means of a capacitive or electromagnetic transducernear the clamp.

The resonance frequencies of different order, for a given length of thesample, are measured as well as the half-width of the resonance curve. Theloss factor d equals the quotient of this half-width and the resonance fre-quency. It is then possible to calculate the storage modulus E' from the reson-ance frequency, the free length, the thickness and the density of the bars.

Plots of E' and d versus frequency are obtained for each temperaturefrom which values at fixed frequencies are interpolated to give plots of E'and d versus temperature.

As for the torsional pendulum measurements, a transition temperature islocated by means of the maximum in the curve of d versus frequency or bymeans of a dispersion step of E' which is associated.

For the flexural vibrations the TNO Laboratory used an apparatus des-cribed by Dekking6. The sample is in an horizontal position and both endsare free.

The dynamic mechanical properties were studied in the various labora-tories over a wide range of temperatures from —160°C to + 125°C.

ResultsA typical plot of storage modulus E' and loss factor d or tan ö versus

frequency for the rigid PVC in bending vibration is given in Figures 1.1and 1.2. The same plot is given for the toughened PVC in Figures 1.3 and 1.4.The curves are given for five participants at about 20°C and 50C.

Figure 2 shows the variation of the storage modulus E' and G' versustemperature for rigid and toughened PVC at five resonance frequencies,between 1 and 2000 Hz.

Each participant has obtained similar plots for the loss factor and thestorage modulus, but the quantitative agreement is not very good. Thecurves show the lower flank of the so-called /3 peak of rigid PVC due to asecondary relaxation mechanism in the main molecular chain. At 20°Cthe maximum of the damping peak is situated at a frequency higher than 2000Hz. At about 1000 Hz the values obtained by the various participants forthe loss factor and the storage modulus vary from 2.8 x i0_2 to 4.0 x 1O_2and from 3 x io kgfcm2 to 3.8 x io kgfcm2 respectively.

Under the same conditions the measurements on the toughened PVC319

Tab

le 1

. Dyn

amic

mec

hani

cal

mea

sure

men

ts:

test

con

ditio

ns an

d eq

uipm

ent

Pezz

in,

Tor

sion

M

onte

diso

n vi

brat

ions

Fl

exur

al

vibr

atio

ns,

Brü

el an

d K

jaer

Tra

nsve

rse

vibr

atio

n m

easu

rem

ents

(c

antil

ever

be

am)

Shee

ts a

nd s

trip

s fr

om S

olva

y

0

Con

trib

utor

A

ppar

atus

T

est m

etho

d ac

cord

ing

to

ISO

No.

533

M

ater

ial u

sed*

Fr

eque

ncy

rang

e (H

z)

Tem

pera

ture

ra

nge

(°C

)

Prep

arat

ion

of te

st

spec

imen

s

Dim

ensi

ons o

f tes

t sp

ecim

ens

(len

gth

= fre

e le

ngth

)

I R

ettin

g,

BA

SF

Flex

ural

vi

brat

ions

, B

riie

l an

d K

jaer

Tra

nsve

rse

vibr

atio

n m

easu

rem

ents

(t

wo

ends

cl

ampe

d)

1 an

d 3

100

to 1

000

23, 25 a

nd 5

3 Sh

eets

, pla

tes

and

stri

ps fr

om S

olva

y 15

3 x

10 x

15

mm

111

Obe

rst,

Hoe

chst

Fl

exur

al

vibr

atio

ns,

Bri

iel a

nd

Kja

er

Tra

nsve

rse

vibr

atio

n m

easu

rem

ents

(c

antil

ever

be

am)

1 an

d 3

30 to

200

0 —

40

to 6

0 Sh

eets

, pl

ates

and

st

rips

from

Sol

vay

150

x 10

x 1

5 m

m

V IV

B

arre

tt an

d T

orsi

on

B m

etho

d 1,

2 an

d 3

0.2

to 2

—

100

to 5

0 G

ranu

les m

illed

5.

72

x 0.63 x

Baxter,

vibr

atio

n, o

wn

at 1

70—

180°

C

0.05

1 cm

and

M

onsa

nto

cons

truc

tion

acco

rdin

g to

Nie

lsen

(thi

ckne

ss 2

mm

):

shee

ts pr

esse

d at

16

0°C

und

er 7

0 kg

cm

5.72

x 0

.63

0.16

5 gr

n x

p

—

I and 3

1

— 1

80 to

12

5

lOO

tol0

00

—70

to75

100

x 10 x 1

mm

170

x 8

x 1.

5 m

m

VI

Gon

ze, S

olva

y T

orsi

on

vibr

atio

ns,

Zw

ick

Flex

ural

vi

brat

ions

, B

rUel

and

K

jaer

Tra

nsve

rse

vibr

atio

n m

easu

rem

ents

(c

antil

ever

be

am)

— 6

5 to

20

Gra

nule

s pre

ssed

at

190

°C u

nder

45

kg

cm2

— 5

0 to

23

Plat

es p

ress

ed a

t 17

5°C

und

er

60 k

g cm

2 or

ex

trud

ed

50 <

10

x 0.

28

mm

215 and 130 x 1

0

x 4 mm

* M

ater

ials

: 1,

rig

id P

VC

; 2, P

VC

+ 6

pph

CPE

: 3, PV

C +

12

pph

CPE

(-3

—1 C

ru

tn 0

-11

(-3

A m

etho

d I a

nd 3

0.75 to 2

25 to

120

0

VII

Fr

anta

, T

orsi

on

B m

etho

d 1,

2 a

nd 3

ito

1.7

— 6

0 to

66

Gra

nule

s pre

ssed

80

x 1

0 x

2 m

m

Tec

hnic

al

vibr

atio

ns,

own

See

K.

H.

Ille

rs

at 1

80°C

und

er

Uni

vers

ity of

co

nstr

uctio

n an

d H

. Bre

nar,

12

0 kg

cm

2 an

d of

Pra

gue

acco

rdin

g to

Nie

lsen

N

oniu

s tor

sion

vi

brat

ions

Kol

loid

-Z.,

176,

11

0 (1

961)

re

laxe

d 4

h at

90

°C

z (-3 ru

C) z

VII

I H

eijb

oer,

TN

O

Non

ius

tors

ion

B m

etho

d 1,

2 a

nd 3

0.

2 to

3.2

—

16

0 to

70

Gra

nule

s pre

ssed

15

0 x

7.02

x 3

.52

vibr

atio

ns

at 1

85°C

und

er

76kg

cm2

(mec

hani

cal

mm

; 50

x 6.

5 x

0.51

or

1.65

mm

Flex

ural

T

rans

vers

e 1

and

3 14

0 to

370

0 —

80

to 4

0 ad

just

men

t of

184.

2 x

7.02

x

vibr

atio

ns,

vibr

atio

n th

ickn

ess)

3.

52 m

m

own

mea

sure

men

ts

cons

truc

tion

(fre

e-fr

ee b

eam

)


+20°C

p

Ill PVVI 0

vur +8x10..6x10 -

+20°C +23°C

20 40 60 80 102 2x102 1x102 2x103 3x1038x102

Frequency (Hz)

Figure 1.1. Rigid PVC 20—23 C.

E 4xi04 +40°CU +50°C

2x10

8x1026x10

+50°C4x1O+

2x102 .——-—--—-.:TT -... + +40°C

20 40 60 80 2 2x102Frequency (Hz)

Figure 1.2. Rigid PVC; 40—53C.

322


E4X104 +20°C

2xi0 A

IinnnI+

8x102 +23°C6x10 .I_-----D ._+

+20°C

+20°CI I I I I

20 0 60 80 102 2x102 4x102 6x102 8x02103 2x103 3x103

Frequency (Hz)

Figure 1.3. Toughened PVC; 2024 C.

4x104+53°C +40 C

+ + +

2x104+50°C

rIII DlVIII 1j

8x1026x102'° 4x102C * +

0 ° —. .

3U , .. --1' +40°C2x10 2 -

20 0 60 80 102 2x102 4x102 6x102 io 2x103 3x103

Frequency (Hz)

Figure 1.4. Toughened PVC; 40—53C.

323

PAC—35—3 .0

1< 1O

ir50E

301u20

E

30 -

2OEU

.3o20 -

50E

30

iu2020

E10

-100 -90


103

-3020

50E

— 50E

—50Hz 50

EU

30lu 20

20

E10

(0-100 -80

show a very similar scattering, the damping at about 1000 Hz is found to varybetween 4.1 x 102 and 5.5 x 102 and the storage modulus between2.4 x io and3 x i0 kgfcm2.

Figure 3 represents a plot of the storage moduli E', G' and the loss factorversus temperature obtained on rigid and toughened PVCs by meansof torsional and bending vibrations. The mean shape of the curves arerepresented at four frequencies 1, 24, 200 and 2000 Hz.

Rigid PVC2000Hz

Toughened PVC

1000 HzVIII

Ill

100Hz)—,

1000HzVIII -.

—VIII 200 HzVIVII100Hz)V50Hz

-60 -40 -20 0 20 40 60Temperature (°C)

Hz IV VIII

Figure 2.

-60 -40 -20 0 20 40 60Temperature (°C)

Toughened PVC

x104

E

2x 10

5 x102

C

2x102

-210

-180

Toughened

PVC

Rigid PVC

-100 -50Temperoture (°C)

Figure3.

324

+ 50


For rigid PVC the plots show well-known features. For a constant fre-quency the modulus falls with increasing temperature and steps in the curvescan be observed. Associated with these steps we observe, for each frequency,two peaks in the mechanical damping versus temperature. The first one is abroad peak which appears at about —60°C at 1 Hz, —35°C at 24 Hz and0C at 2000 Hz; the second one is the well-known narrow peak which cor-responds to the glass transition of the PVC. The heights of the broad dampingpeaks are within the range 2.9 x 10-2 at 1 Hz to 3.4 x 10_2 at 2000 Hz.

For toughened PVC, the mechanical damping exhibits a third peak. Thisnarrower peak appears at about —10°C for 1 Hz. It has been shown in thefirst part of the work that the height of this peak varies with the amountof CPE present in the PVC. This peak, which corresponds to a very markeddispersion step of E', is associated with the relaxation process of the glass—rubber transition of the CPE.

The temperatures at which the damping due to the CPE occurs are signi-ficantly dependent on frequency. The temperature of the peak goes from—10°C at 1 Hz to + 10°C at 2000 Hz. The total height of the rubber dampingpeak is increased with the frequency but this is probably due to the progres-sive merging of the broad peak of the PVC into the narrow peak of the CPEwhen the frequency goes up.

DiscussionIt is known7 that the temperature dependence of the relaxation time t

of each motion of a group of the polymer chain or segments of the chaincan be expressed approximately in terms of an apparent energy of activa-tion Q:

exp(Q/RT)At any one temperature there will be a whole spectrum of relaxation

times, but the large majority will be clustered around a fairly well-definedtime and therefore the mean value of the spectrum is considered.

It may be shown that the relaxation time for the material investigatedin dynamic mechanical experiments may he equated to the inverse of thecircular frequency of the damping peak. Hence for temperatures of maximumdamping the frequencies are correlated to the activation energy by the rela-tion

w = a exp( — Q/RT)Therefore a plot of ln w versus reciprocal of absolute temperature should

be linear, the slope being ( — Q/R).Figure 4 shows the variation of log v with the inverse of the absolute tem-

perature for the maximum of the secondary relaxation process of PVC and forthe rubber damping peak of CPE.

It can be seen that the results obtained by five participants are situatedon straight lines for the PVC secondary transition as well as for the transitionassociated with CPE. The agreement between the experimental values isfairly good.

For the CPE transition temperature the linear relationship obtained is:

in v = — 30.070(TG)' + 113.92 with n = 0.961

325


where v is the frequency in Hz, T0 is the CPE transition temperature in Kand n is the correlation factor.

The energy of activation, Q, estimated from all the measurements in therange 0.2 Hz to 2000 Hz, is 60 kcal mol . This value is equal to that foundin the first part of the work on the basis of measurements at low frequencies.This order of magnitude of activation energy is characteristic of a main transi-tion.

For the PVC secondary peak, the relationship found is:

In v — 6.917 (Ta) 1 + 32.5 with n = 0.993

where T is the fi peak temperature in K.The points obtained at low frequencies may be divided into two groups

which give a higher and a lower value for the activation energy of the f3transition. The mean value is 13.8 kcal mol 1, extremes being 15 and 13kcal mol'. This value agrees very well with results published by differentworkers. It is slightly higher than those found previously (10 to 13 kcalmol 1)8

It is possible to make a similar examination of the modulus values at anyone temperature for the various frequencies. The discrepancies are moreimportant with regard to the loss-factor determination. It was concludedin the first part of the work that the specimen thickness had some effect on theshear modulus measured in torsional vibrations.

It seems that such an effect is observed in the bending vibrations for whichthe laboratories have used different thicknesses. The measurements of TNO

326

5

2

S

2

102

5N

ir0,0— 52

0s

0.2

Transitionassociated to

C FE

IrI

0IV IVVi OAV11 •VIII +

0÷

3.2

.1 I

3.4 3.6 3.8 4.0 4.2

K1(x 10)

Figure 4.

4.4 4.6 4.8 5.0


and Solvay were carried out with samples of thicknesses of 3.5 and 4 mm.They are in good agreement. Hoechst, BASF and Montedison used thinnersamples of l5 mm.

The results of the first group are about 10 to 15 per cent higher than thoseof the second one.

2. RESULTS OF iMPACT MEASUREMENTS

Materials, equipment and conditions of testingThe impact properties of the two PVCs whose dynamic mechanical

properties were studied above have been examined. Various kinds of impacttests were applied. The impact strength of the two PVCs was studied over alarge range of temperatures and at various rates of deformation.

The test methods and the experimental conditions used by each participantare given in Table 2.

Results and discussionThe results are given in graphical form in Figures 5, 6 and 7.It can be seen from the plots on Figure 5 that a well-marked brittle—tough

transition is observed for all the tests and that the results differ greatly fromthe rigid PVC to the toughened PVC and from one test to another.

The approximative range of the brittle—tough transition temperatures forrigid and toughened PVC is given in the following table.

Laboratory

Hoechst

Test Rigid PVC Toughened PVC

Charpy impact > 40C 520CMonsanto Falling weight 0—30C — 5—bCMontedison Izod impact >40'C — 10-0 CTNO Charpy impact 40—45C I5—20CSolvay Tensile impact

Prestressed impactb0—15C

+20CC0—5C0—b0C

As far as we consider only the mean values of the temperatures of thebrittle—tough transition for the two materials in the various tests, we noticethat the transition temperature of toughened PVC is always 10 to 15°Clower than that for rigid PVC.

For rigid PVC, impact tests carried out with notched pieces exhibit anincrease in impact strength at a higher temperature than unnotched pieces,namely, falling weight and tensile impact tests. This is probably due to thefact that the strain conditions are more severe at the root of the notch.

For toughened PVC the result obtained by Montedison in an Izod impacttest is not in agreement with this conclusion. The brittle—tough transition islower in temperature. These differences may be due to the quality of thenotch as shown by Oberst10.

In attempting to correlate the impact test transition temperatures fornotched impact tests, one is faced with the problem of the test pieces beingdifferent in size and shape. Some estimate of the strain rate of the test mustbe arrived at, as has already been discussed in the first paper of the workingparty on polystyrene7.

327

Tab

le 2

. Im

pact

test

s: te

st c

ondi

tions

and

equi

pmen

t

w

00

Con

trib

utor

A

ppar

atus

T

est m

etho

d

Ger

man

sta

ndar

d D

IN 53

453

Mat

eria

l us

ed*

1 an

d 3

Tem

pera

ture

ra

nge

(CC

)

— 4

0 to

60

Prep

arat

ion

of te

st

spec

imen

s

Shee

ts a

nd s

trip

s fr

om S

oiva

y

Dim

ensi

ons

and

num

bers

of

test

spe

cim

ens

Ill

Rac

ké,

Hoe

chst

C

harp

y no

tche

d im

pact

IV

B

arre

tt an

d Fa

lling

wei

ght

Bri

tish

stan

dard

1,

2 an

d 3

—77

to 6

0 G

ranu

les m

ilied

t D

iscs

2.

25 in

; B

axte

r, M

onsa

nto

test

27

82. M

etho

d 30

6 B

; he

ight

of f

all

cons

tant

: 60

and

96 c

m: t

he w

eigh

t of

the

stri

ker i

s va

ried

170C

; she

ets

pres

sed

at 1

80C

un

der 2

0 kg

cm

2 (t

hick

ness

2 m

m)

thic

knes

s 0.

06 in

20

spe

cim

ens

per

tem

per

V

Pezz

in,

Mon

tedi

son

Zw

ick

pend

ulum

T

est

Izod

M

etho

d: A

STM

D

-256

. Sp

eed:

3.4

m

s1. E

nerg

y: 13

.8

kg c

m

1 an

d 3

— 75

to

68

Shee

ts a

nd s

trip

s fr

om S

olva

y 63

x

12.7

x 1

.5 m

m

12 s

peci

men

s pe

r te

mpe

r

VI

VII

I

Gon

ze, S

olva

y

Hei

jboe

r, T

NO

Fran

k pe

ndul

um

Pres

tres

sed

impa

ct

resi

stan

ce

Ten

sile

im

pact

te

st

Cha

rpy

notc

hed

impa

ct

Met

hod:

D 5

3448

m

odif

ied.

Spe

ed:

3.86

ms'

. Ene

rgy:

75

kg

em.

Ref

.: Pl

astiq

ues

Mod

. E

last

oniè

res

(Par

is).

20

(7) (

1968

) M

etho

d: T

NO

, V

eloc

ity o

f st

rike

r:

3.35

m s

AST

M D

256

M

etho

d B

Spee

d: 3.

35 m

s'

Free

leng

th: 4

in

1 an

d 3

1 an

d 3

1, 2

and

3

1.2

and

3

— 20

to 2

5

— 2

0 to

20

from

—

190

to 4

0

from

—

190

to 6

0 (f

luid

con

ditio

ning

)

Ext

rusi

on

of s

heet

s

Ext

rusi

on o

f she

ets

——

-—..—

...__

____

____

____

____

____

G

ranu

les

mill

ed a

t 17

0C, s

heet

s pr

esse

d at

185

C

unde

r 76

kg

cm2

Dum

b-be

lls: n

arro

w

sect

ion

12 x

5 x

1.

5mm

20

spe

ci-

men

s pe

r tem

per

AST

M D

638

—58

T

20 s

peci

men

s pe

r te

mp

Dum

b-be

lls: s

ectio

n 3.

2 x

1.3m

m

4 sp

ecim

ens

per

tem

per

127

(101

.6)

x 12

.7

x 2.

3 m

m

2 to

6 sp

ecim

ens

per t

empe

r

* M

ater

ials

1,

rig

id P

VC

: 2, P

VC

+ 6

pph

CPE

: 3,

PV

C +

12

pph

CPE

DYNAMIC MECHANICAL AND iMPACT PROPERTIES OF PVC

Charpy imPact strength

• Izod impact strength

: Tensile impact strenth

- Impact strength on•

prestressed sample VI

I60-100-80 -60 -40 -20 0 20

Temperature(°C)

8 x102

6x102

bxlO2

Toughened PVC

Falling

Charpy impactnhVW

-

Izotrengt>./\1- Tensile impact strength VI

- Impact strength on VI-

- prestressed sample

-100-80 -60 -40 -20 0 20 0 60Temperature 1°C)

Figure 6. Rigid PVC.

329

Rigid PVC

Follingmpo,,,/4V1.0

E0.5

20

10

2010

800

400

400200

1.0 -

E 0.5

20

g 10

2010

800

/.00

400

200

EUEU

Figure 5.

+53°c V

A

•365°C4 AifA _/i_I::/'

72C1.16ocz-. c I

rLyl •A

2.102

i0280

60

40

203 2 1 0

log-1 -2 -3 z -5


Tensile tests performed at well-defined elongation rates are better suitedfor a basic study of the ultimate behaviour of solid polymers. Results of ten-sile tests will be discussed in detail in Section 3 of this report.

in order to allow a direct comparison with the impact tests, the ruptureenergy values calculated from the tensile measurements are given in Figures6 and 7.

+23°C--'8x102

6x102

4x102

2x102

102 __________80

60

40

-'3 2 1 Ô 1

tog

Figure 7. Toughened PVC.

Investigations were carried out in a range of strain rates from iO to102 s'1 and in a range of temperatures from —20°C to 53°C.

For rigid and toughened PVC at about 20 C, a good agreement existsbetween BASF, Montedison and Solvay for the position of a first step whichis found at about 5 x iO s for the rigid PVC and l0_2s for thetoughened PVC. This transition is shifted to lower strain rates for lowertemperatures.

Solvay has carried out a great number of experiments at the higheststrain rates. In the range from 3 x 102 to 10 s 1 the rupture energy in-creases with the strain rate, because of the increase of the rupture stress. Inthe range above 10 s it is seen that a well-marked fall in the rupture energyoccurs at —16°C and + 22°C for the rigid PVC and only at —16°C forthe toughened PVC. It will be seen in Section 3 that these transitions corres-pond to a sharp decrease in the rupture strain associated with brittle fracture.

in contrast, the first step in the range of low strain rates corresponds to adecrease of the rupture strain in the plastic deformation range (see Section 3).

The CPE rubber is active in both transitions. Its effect is to shift the transi-tions to the higher strain rates.

330

00

VI I

—1

T •IVlviLVIT

-2 -3 -4 -5

Tab

le 3

. Ten

sile

test

s: te

st c

ondi

tions

and e

quip

men

t

0.05

to 2

00 m

m m

in'

1 an

d 3

23

Shee

ts a

nd

stri

ps fr

om

Solv

av

()

(2 2 2 H 0 rr

H

Con

trib

utor

A

ppar

atus

T

est m

etho

d Sp

eed

rang

e

6.7 x iO to

10-2

1 and 3

23 and 50

Mat

eria

l us

ed*

Tem

pera

ture

ra

nge

(C)

Prep

arat

ion

of te

st

spec

imen

s

Dim

ensi

ons a

nd

num

bers

of t

est

spec

imen

s

I R

ettin

g,

Wol

pert

Sh

eets

and

11

0(60

) x 1

5 x

1.5

BA

SF

Plas

tech

on

Rel

axat

ion

cm s

10 to

102

cm

s

1 an

d 3

23 a

nd 5

0

stri

ps fr

om

Solv

ay

Shee

ts a

nd

stri

ps fr

om

Solv

ay

mm

: gau

ge le

ngth

40

mm

24

0(20

0) x

15

x 1.

5

mm

: gau

ge le

ngth

20

0 m

m

11

Cha

tain

, O

wn

iO' to

3 x

iO cm

s-I

3

— 4

0 to

73

Stri

ps fr

om

CE

MP

spec

imen

, C

EM

P co

nstr

uctio

n Z

wic

k D

Y 1

0

3 x iO

to I c

ms

01 to

5 m

s

3 3 —

4O

to 7

3 —

40

to 7

3 So

lvay

St

rips

from

So

lvay

dum

b-be

lls

75(5

0 x

4 x

1.5

mm

R

elax

atio

n 20, 50 and 60

23, 5

0, 5

5, 6

0,

III

Obe

rst,

Rel

axat

ion

1

Hoe

chst

62

and

70

Gri

mm

inge

r,

3 23

and

50

Hoe

chst

R

elax

atio

n 1.

67 x

iO

" to

8.33

1

and

3 22

V

Pezz

in,

Inst

ron

type

A

STM

D 6

38

Mon

tedi

son

TT

-CM

(+

cre

ep t

est)

cm s

VI

Gon

ze,

Solv

ay

Shee

ts a

nd

stri

ps from

Solvay

Shee

ts a

nd

stri

ps fr

om

Solv

ay

Inst

ron

type

T

T-C

M

Ow

n co

nstr

uctio

n Fr

ank

AST

M D

638

--58

T

0.05

to 5

00 m

m m

in

1 1

and

3

AST

M D

638

-58

T

0.17

to 5

8 m

min

A

STM

D 6

38—

58 T

0.

5 to

50 m

s'

(220) x 8 x

1.5

mm

3

spec

imen

s pe

r sp

eed

and

tem

pera

ture

Gau

ge le

ngth

50

mm

Ref

eren

ce l

engt

h 75

mm

R

efer

ence

len

gth

75 m

m

Ref

eren

ce l

engt

h 75

mm

— 4

2 to

22

Ext

rusi

on o

f sh

eets

3

— 1

6 to

38

Ext

rusi

on o

f sh

eets

1 an

d 3

— 1

6 to

36.

5 E

xtru

sion

of

shee

ts

0


3. RESULTS OF TENSILE TESTS IN A VERY LARGE RANGEOF STRAIN RATES AND TEMPERATURES

Materials, equipment and conditions of testingThe same PVCs have been used for tensile and impact tests. All the

specimens have been cut from extrusion moulded plates and extrusionmoulded strips having a thickness of about 1.5 mm.

The experiments were carried out by BASF, CEMP, Hoechst, Montedisonand Solvay. Experimental conditions are given in Table 3.

The stress relaxation experiments are discussed in Section 4 of this report.They were mainly conducted in the laboratories of BASE' and Hoechst.

The experiments of the five laboratories arc concerned with tensile testsin the linear region and the non-linear region until fracture.

A very large range of strain rates was covered by means of special devices.The stress—strain curves were determined in each case. From these curves

the following quantities have been measured:(a) the Young's modulus in the linear region, Ekgfcm2)(b) the yield stress, o (kgf cm2)(c) the yield strain, (per cent)(d) the rupture stress, o (kgf cm 2)(e) the rupture strain, r (per cent)(J) the rupture energy, T4' (kgf cm cm 3)The results of the rupture energy measurements, given in Figures 6 and 7,

have been discussed in Section 2 of this report.All the results have been plotted versus the strain rate which is the ratio of

the crosshead speed to the gauge length of the sample:= V/L

In some cases, especially for Hoechst and BASF measurements of therelaxation spectra, the results have been plotted versus the yield time andthe rupture time which have been calculated from the corresponding elonga-tions and the crosshead speeds.

ResultsA typical plot of tensile stress—strain curves is given in Figure 8 for the

toughened PVC at — 16°C for speed of testing from 0.01 cm min1 to5.71 x lO cm min '. It can be seen that the yield strain passes through amaximum.

Tensile modulusFigure 9 shows the results of the tensile modulus measurements in the

linear region. Contributors are BASF and Montedison. The measurementswere made up to about 0.5 per cent elongation. The time was calculated bythe aid of the ratio AL/V The agreement is fairly good between the BASEand Montedison results.

•The curves show the shift of the softening region to shorter times when thetemperature is increased. This is directly correlated to the main molecularchain mechanisms of PVC (effect of glass transition temperature).

332


Rigi and toughened PVC Rigid and toughened PVC

+64°C +53°C

1

.71 xl04crn miri1

2.03 x104

lO3cm mm1

(°/126

2.5 3 3.5 2+ .s 5 5.5 6 6.5 7.5(cm)


8

5

2

8

6

4

2

EU

-x

2 4 6 81O2 4 6 81O_22 4 6 8O0.2 0.40.61 2 4 6810 2 4 6 8102 2 4 5810Time (s)

Figure9.

333

r .1V xlVI t.AI

Rupture stressFigures 10 and 11 show the values of the rupture stress for rigid and tough-

ened PVC respectively, versus the logarithm of the strain rate. Contributorsare BASF, Montedison, Solvay for rigid PVC plus CEMP for toughenedPVC.

The curves are characterized by a number of features:(i) They show a slight minimum at about 10 2 s and give an indication of

amaximumat iO s1.

r :V xjVIVII

A. GONZE AND J. C. CHAIJFFOUREAUX

-3°C

-42°C

1400

1200

1000

800

600

400

200

0

V—-y-+53°C+23°C

3 2 1 0 -1 -2 -3 -4 -5log F


EU

C

1200

1000

800 -

600 -

400 -

200 -

C-

+23°C

I I I I3 2 1 0 -1

log e

Figure Ii. Toughened PVC.

334

-2 —3 —1. -5


(ii) They show a discontinuity for each temperature after a zone where therupture stress is increasing with increasing strain rate.

(iii) The strain rate corresponding to this discontinuity is higher, the higherthe temperature.

(iv) in the upper part of the figure, the rupture of the samples is brittlefor the higher strain rates; the discontinuity corresponds, therefore, to atough--brittle transition region.

For toughened PVC, the curves are different from those obtained for rigidPVC at the highest strain rates. The ruptures are not brittle for tempera-tures above 9°C.

Rupture stress of modified PVC is about 20 per cent lower than that forrigid PVC at all the strain rates investigated.

II 0

VI 0.*Jo


Rupture strainFigures 12 and 13 are plots of the rupture strain versus the logarithm of

the strain rate for rigid and modified PVC respectively. In several cases thescattering of the measurements has been indicated. Contributors are thesame as for the rupture stress measurements.

It is obvious from these plots that the rupture strain does not vary mono-tonously when the strain rate is increased. Except for very low temperatures(—42 and —31°C), each plot can be divided into three well-defined rangesseparated by two transition zones. The rupture strain for toughened PVCat all temperatures is higher than that for rigid PVC.

The agreement between the four collaborators is very good as far as thelocalization of the strong rupture strain fall is concerned. This fall of ultimateproperties situated in the low-speed range indicates that a first variation ofthe tensile properties in the non-linear range takes place. This zone ofvariation is slightly displaced, for toughened PVC, to higher strain rates, i.e.to smaller extension times.

.335

21.0

200

160

120

80

40

0

Tough- britfietransition VI

- +36.5

V

+22°C

log e


After this first variation zone, the rupture strain of rigid PVC (Figure 12)keeps constant, at a level dependent on the temperature, for more than threedecays of variation of the strain rate. This rupture strain which is muchlower than that obtained in the range of lowest strain rates, is still higher thanthe rupture strains corresponding to a brittle fracture, so that the first varia-tion corresponds to a toughtough transition.

320- 7

280 a

-240 / ,I, /

'200 /

80 -+36.5°C -°c +22°C +38°C /1 7//-i2so.L2oC

tog £


If the speed of testing is still increased it is noticed that the rupture strainfalls abruptly for a given speed which is different for each temperature, thehigher the temperature, the higher the speed. After this fall the rupturestrain is reduced in each case to about 6 to 8 per cent and the rupture becomesbrittle.

As for the rupture stress, it is possible to define a tough--brittle transitionzone in Figure 12.

The behaviour is slightly different for toughened PVC (Figure 13). Inthis case, for speeds in the intermediate range, the rupture strain increasesfor tests above 2OC. The fall at high strain rate is only observed for thetests at lowest temperatures. As for rupture stress the influence of rubberin PVC is visible in the high strain rate region.

Comparing Figures 10 and 11, on the one hand, and Figures 12 and 13,on the other hand it is obvious that the brittle—tough transition corresponds,for the same strain rates and temperatures, to a sudden fall in rupture strainand a sudden increase in rupture stress.

Yield stressFigures 14 and 15 are plots of the ratio of yield stress to absolute tempera-

336

0

DYNAMiC MECHANICAL AND IMPACT PROPERTIES OF PVC


ture against the logarithm of the strain rate. This representation has beenchosen in view of the theoretical interpretation of the data, discussed below.

Figure 14 is a comparison, for rigid PVC, of the results obtained byBASF, Montedison and Solvay. Agreement is good at 20—23°C betweenMontedison and Solvay. The slope for BASF at 23 and 53°C is different.Figure 15 is a similar comparison for toughened PVC.

5

4

EU

0)

b

0

5

6

3

2

tog C

I +23°C •+53°C 0

V +23°C 'VI —42to+36.5°C

3 2 1 0 -1 -2 -3 -4 —5

logFigure15. Toughened PVC.

337


5

4

EU

03 2 1 0 -1 -2 -3 -L -5

log


Figures 16 and 17 show only the results from the Solvay Laboratory in avery wide range of strain rates and temperatures. Each point representsone single measurement in the range above 10 s_I and the mean value ofthree tests below 10 s1. The ratio of yield stress to absolute temperatureis given against the logarithm of the strain rate. Following Roetling3 and

5

E 3U

L2>Transitionassociated toCPE

1—

3i0C

of PVC

03

oc

I I I I

2 1 0 -1 -2 -3 -4 -5tog è


338


Bauwens4' the modified or generalized Eyring equation which describesthe non-newtonian viscous flow of polymers in the case where two or moredeformation processes are involved, has been applied to these data in orderto determine whether and how the results would fit this theory over a widerange of experimental conditions.

With the aid of this theory, according to a procedure outlined in the dis-cussion part of this section, the experimental points have been joined inthe diagram aT versus log by means of segments of straight lines.

It is obvious from Figures 16 and 17 that the experimental data may berepresented, in a first approximation, by two families of parallel straightlines indicating the presence of two deformation processes: an processdominating at high temperatures and low strain rates and a fiprocess whoseinfluence becomes apparent at lower temperatures and for higher strainrates.

Although the number of experimental results is reduced in the zone ofhigh temperature and low speed, it seems that for toughened PVC anadditional transition, namely the glass transition of CPE rubber, passesacross the two deformation processes of the rigid PVC. The statisticalanalysis of the experimental data seems to corroborate the existence of thisthird process, as will be further discussed below.

DiscussionRupture strain and stress

Th rupture behaviour of PVC gives interesting information on the modesof relaxation in the non-linear range of deformation. As the rupture strainis more sensitive than rupture stress to any modification of test conditions,the variations of the former as a function of temperature and strain rate(Figures 12 and 13) were only discussed.

For rigid as well as for toughened PVC two zones of fast variation ofrupture strain against strain rate were found.

The first one appears at high strain rates and is temperature dependent.As we wrote above, it is the transition zone between brittle (high speed)and tough (low speed) ruptures of PVC. It disappears at high temperature(22, 27 and 36.5C) in toughened PVC. This is probably due to an effect ofthe CPE which prevents brittle ruptures.

An Arrhenius relation can be used to correlate rupture time and tem-perature in the range of temperatures used, yielding an activation energyof 14 ± 2 kcal mol -'. As this value is the same as that found in the dynamicmechanical measurements for the fi secondary transit.ion of PVC, we canrightly assume that the brittle—tough transition of PVC in tensile-impacttests is due to the liberation of local motion of the macrornolecular chainwhich allows plastic deformation in the non-linear zone of deformation.The second transition observed in the rupture strain curves appears at

lower strain rates and is slightly temperature dependent (apparent activationenergy less than 2 kcal mol '). It is characterized by an important variationin the rupture strain in the tough zone of rupture. Since the deformationprocess of PVC does not change when the strain rate is increased fromvalues situated, for a given temperature, below the transition zone to valuesabove this transition, and as the temperature dependence of rupture time

339

XUA3$1UOF1UAF) .D .1. UkIA 3S1400 .A

3d1 lrfiio ol Idioq Ion i ft wof l3IIJffl 1TI f1OiIfiflI1t H1 113

oo1q noWxE11 i 10 fit grIis311 3th riliw vw biwio1ir1iiiI 13 fIt ffOIIiflI31i .nul3rk is1uo1orn rtijrn oth lo

b13cIrIoa tth3do1q i briodo nth1t ruJIqLn io nOiII$in3'1 gI31 dT (ot131 wol) 1Ernii-1ioj bnE (oti ffif1) aji13dthbE nxwixI r1oitinEfl 3rD iiiiw

2333O1 lint

noiismiobb iiac1q o no no birnbni I £ x lii nisfl (01 x) •1 .noiij1io no

i btn vq in I13qcu1I i lo ono 3lIJlqUl 3th woth bn13 \. sTh\ IA .Y E 1i 3flO fIoi1iflE1l r[uol—r1uol 3r11 gr1ioE3n3 ft31 rii13i1 Owl

b3tE13n iirI 3rD Th\°i) noiiinE11 3di lo oth n131li 13f!irj 3I131 nu13fl lonni noitErrrlol3b b 3f1O lEilifli 3f11 lii 131113ff! 3ff! lo cioilom uoiv odi d

3313iq 3)1J3I DVq th io rii!ir1iovo ni brul r1fJOrI3 '1biqii boiisqiib 3d '1noift 3cwb31 ol ffnon rigirl 3d 'Bffl 3luiJll3qlTt3flo noilthijw 1&lol idT

'[3t31qrno3 i noi!j3rnlol3b 3f11 313f1W 3rflUIOv E1isrni 13 ill DVT 3f1110 ri11it 3r11

.333iq !iI 3rD lo oiisiqui 3rli 01 ii 3vig bn bi1o1 noillsmlol3b 3ii13[q lEilini idgi1 3d! gnut133th3,o Eio1 3r11 10 rEui

3Ior!w oth rn 3IEgEqolq tonnED noiIEmlol3blo 3no lthtini orli bnflolE b3mlol .333iq i3I

-i123 n33d Ef1 Ii fIOt113ffi1013b 3rnulov bn13 'g13r13 31u1qu1 lo i13d 3r11 nO .D°OOI n&1l -Ifigid 3d n133 3wft313qrn3t lo 3i3313f1i 113301 3d! 113111 b3tErrI

rJ .( m\) ruiftisi1 r1I 10 qi1 3ff! lo '3n3uq2dtE1I orli rii131qx3 13131 ir1T 3vodE Ii 32i131 01 ffglJon3 3g113! f133d r1 311J11313qm3! lo a&n3ni 3111 no ir1l

DVq IO 1IT23o5N TDAM1 014A JADIfrTAHDaM DIMAfrTYU

n1n3u1oa 1J$3o1 3111 nth1qx3 fl3iffW TVS lo 3wiJn3qm3i noiiirnit 131 3111

.1i13i1$m 3111 lo 3th ni b3 3n3g tirf 3th noiiinfl 3r11 wobd i131 niifl IA i tni31q oi f1uon3 tbiqEi b3liqiib 3d n froiIim1o13b lo 3no lEitifli

3tJ1qo1q nf3 noiIkm1oi3b 3iti11q 3rIT .31u1E-I3qm3i lo 3ini inii1ini .nifni 3Ilnqul 3u11 i crnboiq bnl$ 333q I3i 3IofIw 3111 ni

01 x 1.1 JB1 niEii2 (01 x ) .- \M

Ii t33TIo3 i noifinffrt bv13cfo d11 lo noilEl3lqT3tni b3oqo1q 3111 11 lo aB1 3111 no in3bn3q3b '1gnofl 3d 111w 13th11 3r11 i&IJ bi33qx3 3d 'm

1o13J1 1&)ht3mo3g d b313311E 3d bluow bnE n3mi33q 3111 ni b3ttq11ib t3d .3331q 131 3i11 lo 3fD13if11 3111 .g.3 iu

31uIE13qfr13i 1n3idm bn 1oL ni 33n31311ib 3g1131 3111 Ia 3IJi333d o1A .\t3f13 noi1nvita 3111 Ia noithniml3l3b j iol b3u d .jonfu3c) it1is1 3111

itnioq lEIuOitII3q 3311i bui2 of nDhoW iuo Ia noifn3tni 3111 i 11

-3d 1iinLr133r. 3111 no 3mmI1o1q 3111 lo 1 3111 ni IuLt3b 3lOffl flu •\fq lo iuoiv&I

DvT b3n3r1uoi bnE biii no 3b&n 21n3m31u3rn 2if 31in31 3rfT .(fnioq bf,i) 1uoivd3d nib13i £ of bnoq3rro3 113111W mucnixIm i

tzn1 idT .niE313fli 2i niElf2 fit 3Iiflw 0 = tb\Db inioq ifli IA t DVq bnduof 3111 bn bigii rit lo woi113d noiismlol3b 3111 113th

14E

4


point is essentially viscous. This behaviour is observed in the glassy range.The yield stress is varying with the temperature and the rate of deformation.

The apparent viscosity of the deformation (ratio of the stress to the strainrate) is usually not constant so that the behaviour is not linear. Applicationof the Eyring viscosity model to the yield behaviour of glassy polymers hasbeen proposed by various authors" 2 and more recently by Roetling3 andBauwens4' .

In the following we will apply the same model to the experimental dataof the present report.

The original Eyring equation, applied to the yield stress, may be writtenin the approximate form:

rjT ' = A ln (C/J) (1)

at sufficiently high stress levels, and sufficiently far outside of the transitions,where o- is the yield stress, T is the absolute temperature, A and C are con-stants, is the strain rate and J is the jump frequency of the rheological unit.

The fundamental process consists of the iump of segments of macromole-cules from one equilibrium position to another.

The jump frequency is proportional to the vibration frequency J0 of therheological unit and to the probability of a jump when no stress is acting onthe material. The probability of a jump is proportional to exp( — Q/RT) asrequired by the Boltzmann distribution law so that

J J0 x exp( — Q/RT)

and that equation (1) becomes

= A {ln (C'/J0) + QIR T] (2)

where Q is the apparent activation energy of t.he process and R the universalgas constant.

In an extended range of strain rates, the curves = f(ln c) do notobey equation (2) but may be represented by at least two segments of straightlines, which according to Roetling and Bauwens correspond to differentmodes of deformation. The intersection between two segments wouldcorrespond to a transition comparable to those revealed by dynamic mech-anical measurements. In this case the variation of the yield stress with strainrate and temperature can be described by the generalized theory of non-newtonian viscosity proposed by Ree and Eyring to represent the viscosityof polymer solutions and polymer melts. Following Roetling and Bauwenswe suppose that the stresses due to the various processes are additive andthat the values of A, C and Q are constant for a given process. ParametersA and C have a structural meaning.

The generalization of equation (2) gives

= A(ln C/JØ, + Q/RT) (3)

At low strain rates the terms containing the variables relating to the other

342


processes can be neglected. When the strain rate is going up we have to takeinto account the successive processes.

If one assumes that two processes are involved, equation (3) becomes

= (A + Afi) In + (A2Q + APQ) + constant (4)

At constant temperature:

= Aln + D1 (5)

in the range between the glass and the secondary transition and

iT= (A + A)ln + D2 (6)

in the range of strain rates above the secondary transition.At constant strain rate:

(7)

at the highest temperatures and

= AQ1 x - + D4 (8)

for temperatures below the secondary transition.D1, D2, D3 and D4 are constants.These equations show that a plot of o-T' versus log would give two

families of parallel straight lines, the first one with a slope A and the secondwith a slope ,4 + A. From the mean displacement of these lines versustemperature, for a given strain rate, it is possible to calculate the activationenergies Q and Q.

A plot of iT 1 versus T would also give two families of parallel straightlines with a slope of A2Q'R and (A2Q + AQe)/R respectively.

A similar treatment may be applied to the data on toughened PVC. Butin this case a new change of slope (unobserved in rigid PVC) might beinvolved, corresponding to the glass transition of the chlorinated poly-ethylene. This process is denoted '.

The coefficients of equation (4) have been calculated for the cx and /3processes of rigid PVC and for the ',x and /3 processes of toughened PVCusing a least squares analysis of the experimental data outside the transitionzones given by the graphical analysis. Indeed, as may be expected, additionof the experimental values, found in the transition zone, to the statisticalanalysis, increases the dispersion of the calculated factors.

The results obtained are:

for rigid PVC:

process = O.1O6ln + 26834 — 6.57kgfcm2 K1

fi process T1 = O.l97ln + 3l68 — 8.11

343

for toughened PVC


' process cT' = 0.08 1 in + 2003 4.79

process aT' = 0.0971n + 2869' 7.84

fi process 0T' = 0.2031n + 3456 — 9.77

The activation energies, associated with the various processes, within the95 per cent confidence interval are:

for rigid PVC:

= 50.1 keal mol' ± 9.5 kcal moi'Q = 10.6kcal mo1 ± 4.5 kcai mo1'

for toughened PVC:

Q = 49.3 kcal mo11 ± 12.4 kcal mo1'= 58.2 kcal mo11 ± 11.5 kcal mo11

Q 1 1.1 kcal mo11 ± 4 kcal mol'

The differences found between activation energies and parameters forrigid and toughened PVC are of the same order of magnitude as the experi-mental errors.

-36.1 -23.1 -6.8 +12.6

4,3x103 4.2 4.1 4.0x1033.9 3.8 3.7 3.6 3.5 3.4

T1(K1)Figure 18. Rigid PVC.

344

.13.3 3.2 3.1 3.0x103

5

I'

3

L.

102 .1

10 -1

1 s-i

+ 34.5 +60.2 T(°C)


For each temperature investigated we have plotted, on Figures 16 and 17,the variations of iT 1 versus log calculated with the results of the leastsquares analysis.

The experimental data are generally well represented by the calculatedlines. Figures 18 and 19 show, for a few values of testing speed, the calculatedvariation ofaT1 versus T1.

s1

s-i

s-iTransition

associated toCPE

- 36.1 - 23.1 -6.8 12.6 34.5 6O.2 T (°C)

4.3x1O 4.2 4.1 4.0x103 3.9 3.8 3.7 36 3.5 3.4 3.3 3.2 3.1 3 0x103T1(K)


Zitek and Zelinger'1 have applied the same treatment to these experi-mental data and obtained similar results.

We may conclude, therefore, that the analysis of the experimental data ina broad range of strain rates and temperatures, presented in this report,confirm Bauwens' contention5 that the /3 transition in PVC can be revealed,not only by dynamic mechanical or dielectric measurements but also by adetailed study of the yielding behaviour in the glassy state.

The molecular relaxation processes observed in the range of linearviscoelasticity appear, thus, also to influence the non-linear behaviour atlarge deformations, the technological importance of which is obvious.

4. RESULTS OF RELAXATION EXPERIMENTS

The relaxation spectrum of PVC was determined from the followingquantities:

345


(i) The relaxation modulus E(t) as a function of the measuring time t instress relaxation measurements.

(ii) The complex modulus of elasticity as a function of the circular frequencyw 2nv in flexural vibration experiments.

(iii) The tangential modulus doide as a function of the measuring time intensile tests.

The measuring methods and the theoretical basis to the experiments arewell known. They are described in a paper by Herman Oberst from Hoechstand Wolfgang Retting from BASF'2.

Figure 9 shows the results of the tensile modulus measurements in thelinear region, for BASF and Montedison, versus the measuring time calcu-lated by the aid of the relation t AL/V.

Figures 20 and 21 show the results of the relaxation modulus versusrelaxation time or measuring time. Contributors are Solvay, 23°C; BASF,23—42—53—64°C; Hoechst, 23—50—55—60°C. The results of BASE differ fromthose obtained by Hoechst. These differences may be due to differences inthe conditioning of the specimens.

Some results of flexural vibration experiments were presented in Figures1 and 2 (see Section 1).

The results of the three measuring methods, i.e. the stress relaxation, thetensile and the vibration tests have been summarized in Figures 22 and 23for the rigid and the toughened PVC respectively. Such a representationallows us to determine the time-dependent modulus .E(t) of viscoelasticsubstances in a wide range of the time t.

Considering the fact that the vibration experiments were made on samplesdiffering in shape, size and conditioning, it appears that at 20—23°C the

8

6

4 11 V1(o)

:--- --------- i-j—-.--o---o 23°C..J —2 — -

I — ..-:- — .. 5O°C- —..----E — -. ---- —.. +42°C

55°C- —-.. N—..—. 8 .

'53°C- N+62°C

2 '64°C

3+70°C

10 _______________________________________________- --- ______2 4 6 8 10 20 40 60 80 102 2 6 8 10 2 4 6 8

Time (s)


346


4 IlL I

2_\23°C

23°C+ 23°C

53°C 42°C8

N NN \ 53°CN

N

"+64°C

io.2 4 6 8 10 20 40 60 80 102 2 4 6 B 2 4 6 8

Time (s)

Figure21. Toughened PVC.

results of the three methods agree reasonably well. The values of the storagemodulus at the low-frequency end of the range are similar to those of thetangent modulus for the short time stress strain experiments.

8

6 VIII 22 °C v

8 °C60--— N \.62— — — —. — .._ .._'. —-

\ç.03 __________Vibration tests Time (s)1

2 27rv

Tensile tests (0.5°/a of elongation)I . . ..

3 Relaxation tests (0.5 of elongation)10iü 2 5 2 5 102 2 5 lOl 2 5 1 2 5 10 2 5 10 2 5 2 5

Time (s)

Figure22. Rigid PVC.

347

÷23 °c


The curves show, for the long times of experiment, the shift of the softeningregion, corresponding to the glass temperature, to shorter times withincreasing temperature. Moreover they show, for the short tipes of experi-ment, the lower flank of the so called fi peak of the rigid PVC and, for thelowest temperature, the step due to the glass transition temperature of thechlorinated polyethylene.

i05

61

v 453-AQVU)VI 2040,60 and 23°C 7

N.r i0 °C °C

: \\\\\Vibration tests

2 - Time(sl1— ______________________ \\Tensile tests l0.5/ of elongation) —,-

Relaxation tests )0.S/. of elongation)102 Ill,,,,

10' 2 4 6 102 4 6 102 2 4 6 10 2 4 6 1 2 4 6 101 2 4 6102 2 6 io 2 4 6 10lime (s)


The relaxation spectra H(t) of rigid and toughened PVCs at 23 and 50°Cwere determined by Oberst and Retting by means of their results of thetime-dependent modulus and of the loss modulus E"2.

A comparison between those curves and the results of the tensile tests inthe non-linear range has been presented by Oberst and Retting. For thispurpose they have plotted the yield stress, the rupture stress, the rupturestrain and the rupture energy versus the yield time and the rupture timerespectively and they have related these plots to the relaxation spectra.

Figure 24 is an example of the work done by Oberst and Retting. Therelaxation spectra show distinctly that two maxima may be present in amore extended time range, though only the flanks can be observed in themeasuring range used.

5. STUDY OF THE CORRELATIONS BETWEEN DYNAMICMECHANICAL, TENSILE, RELAXATION AND IMPACT

MEASUREMENTS

The underlying theme of all the studies so far is that the dynamic mech-anical, the tensile, the relaxation and the impact measurements are based

348

DYNAMIC MECHANICAL AND IMPACT PROPERTiES OF PVC

EL)

a.

I

Time (s)

1O4 1Q3 i_2 10_i 1 10 102 10'S

4L.x10 4.2 4.0xi03 3.8 3.6

T( K)Figure 25,

349

---——--.—.._.....-2 3°C

50°C

1o8c_sEl.2

1O2

2

861.

A2L101

86

i0)23Oc

10

E i0 1

i0s1

(s)


102

10

t .- ey

ioDynamical tests rrQ.T-1- \ '•'- \

,w \, \\ PVC)-'-,. cPe,'rversus T-"--., \'rversusTfor rigid FD1 \for gloss transition of CPE

-36.1

\-23.1 -6.8 +12.6

I - - J___ 1 — I I+34,6 +60,2 T(°C)

3.4 3.2 3.0x103 2.8


upon the same relaxation processes possessing different relaxation times.The dispersion regions are met when the frequencies or the strain rates areincreased. When the temperature is changed to higher values the relaxationtimes are all shifted to shorter times so that the dispersion regions in dynamicmechanical, tensile, relaxation and impact measurements are now found athigher frequencies and strain rates, or at shorter times.

The main results we have obtained are summarized in Figure 4 for dynamicmechanical tests; Figures 6 and 7 for tensile impact tests; Figures 10 to 17for tensile tests, and Figures 22 and 23 for relaxation tests.

(a) In the dynamic mechanical tests at least one secondary dispersion zonetakes place, which is named a /3 transition. The activation energy of thistransition is about l4kcalmol1. Figure 3 shows that the dispersionsecondary peak is very broad and covers 100°C at low frequency. Thetemperature of the maximum is shifted from —60 'C to + 5°C when thefrequency is increased from 1 Hz to 2000 Hz.

The dispersion zone due to the glass transition of the CPE is well markedin the toughened PVC. The corresponding temperatures and frequenciesare — 8°C for 1 Hz and + 10°C for 2000 Hz.

in Figure 25 the average relaxation times of the secondary transition andof the CPE glass transition respectively, have been plotted as a function ofT . It has already been shown in Section 1 that the apparent activationenergies calculated from the slope of these plots, are equal to 14 kcal molfor the /3 transition in PVC and 60 kcal mol' for the CPE glass transition.

(b) In impact and tensile tests, it has been shown that two transition zonesmight be involved. The first one, in the low strain-rate range, correspondsto a tough—tough mechanism and is caused or strongly influenced byadiabatic heating of the specimens. Oberst and Retting12 attributed it to theglass transition of PVC.

The second one, at much higher strain rates, gives rise to a tough-brittlemechanism. Because the time—temperature relation of this transition yieldsthe same value of the activation energy (14 kcal mol 1) as the /3 transitionmeasured in dynamic mechanical tests, the tough--brittle transition can beattributed to the same relaxation processes.

The comparison between the usual impact tests and the tensile impactmeasurements proves that the fracture, in usual impact tests, takes place inthe time—temperature range of the /3 transition.

A transition with the same activation energy has also been found in theyield stress measurements in agreement with the generalized Eyring theoryproposed by Bauwens-Crowet5.

Figure 25 allows the comparison of the time—temperature dependence ofthe transitions measured in the dynamic mechanical and the tensile tests,and illustrates that the 'apparent energies of activation' found for thedynamic mechanical properties, for the rupture behaviour in the brittle—tough transition zone and for the yield stress in tensile measurements aresimilar.

Although the molecular relaxation processes observed in the range ofsmall deformations also influence the behaviour in the non-linear range, itis clear that the mechanical damping in the linear range and the tensileproperties in the non-linear range cannot be connected directly. In particular

350


the exact location in time of the two transitions in tensile tests cannot bederived from the location of the damping peak. In other words, yield timeand rupture time scales cannot be identified with the relaxation time scalefound in the linear range of deformation. This conclusion is in agreementwith the results of the collaborative study on polystyrene published byJones7.

REFERENCESS. Glasstone, K. J. Laidler and H. Eyring, The Theory of Rate Processes, pp 48Q483,McGraw Hill, N.Y. (1941).

2 R. E. Robertson, J. App!. Polymer Sci., 7,443(1963).J. A. Roetling, Applied Polymer Symposia. No 5, 161—169, Interscience, N.Y. (1967).J. C. Bauwcns, J. Polymer Sc!., A-2, 5, 1145 (1967).J. C. Bauwens et al., J. Polymer Sci., A-2, 7, 735 (1969).

6 P. Dekking, PhD Thesis, Leyden (1961) R. F. S. Hearman, Brit. J. App!. Phys., 9, 381 (1958).T. T. Jones, .1. Polymer Sci., C, 16, 3845 (1968).A. '3onze, Pure and Appi. Chem., 18, 551 (1969).A. Gonze, Plastiques Moderne.s et Elastomeres, 20 (7), 134 (1968).'° H. Oherst, Kunststoffe, 52, 4 (1962).' P. Zitek and J. Zelinger, J. App!. Polymer Sci., 14, 1243(1970).

12 H. Oberst and W. Retting, J. Macromol. Sc Phys., B, 5 (3), 559 (1971).

351

a collaborative study of the dynamic mechanical and …

Documents