effect of annealing and temperature on the morphological structure of

19 Pages·2007·0.37 MB·English

EFFECT OF ANNEALING AND TEMPERATURE
ON THE MORPHOLOGICAL STRUCTURE OF
POLYMERS
E. W. FISCHER
Inst liii! für Phys/kalische Chemie der Universitdt Mainz,
II. Ordinariat. 65 Mainz, Germany
ABSTRACT
During heat treatment of a semicrystalline polymer. various alterations of the
morphological structure can occur. Two special cases are treated in more detail:
theirreversible thickening of the polymer crystals during annealing and the
reversible premelting effects occurring during heating of a well annealed sample.
With regard to the thickening process. two mechanisms have to be taken
into account: melting and recrystallization on the one hand and refolding in the
solid state due to the high mobility of the chain molecules on the other. Which
type of process predominates depends on crystallization conditions, heating
rate and molecular weight.
The changes in the morphological structure due to premelting can be analysed
by x-ray small angle scattering studies. It is shown that the decrease of crystal-
linity within the so-called melting range is due to an increase of the thickness
of the disordered surface layer.
I. INTRODUCTION
It is well known that semicrystalline polymers generally change their
physical properties when they are heated to elevated temperatures far below
the melting point. These changes are related to alterations of the morpho-
logical structure and they indicate that in most cases the polymeric sample
is not in a state of thermodynamic equilibrium. Therefore the heat treatment
results in a reorganization of the structure leading to a state of order with a
lower free energy.
It is well established that two types of structure alterations can occur:
irreversible and quasi-reversible changes. The first group of phenomena
embraces numerous kinds of 'annealing effects' and generally the accompany-
ing changes of morphology depend not only on temperature but also on
annealing time. in the other case the measured quantities show a reversible
behaviour with regard to temperature if the time of passing through the
temperature cycle is sufficiently short. Very often a combination of both types
of processes is observed.
In this paper we will deal Lirstly with an important eliect of the first kind,
113
F 1K LI2CHFK
(JiG ILLGAGL21pJC !UCLG9BC Oj (JiG IP!CJ(UG2? 01 (JiG bOJAWCL CLABI9J: anq BCCOUqJA
MC MIJJ {LG9( (JiG df}921-LGACL?IpJG 2(LOC{I1LG CJJ9G? OCCflLiLJ qnLin bLC-
WGJ11U
II' IHICKEJ.4IWC Oh CKA21VF? DflHIJAC (k4JAE(FI14C
OUG O1(PC W02J !mb0LIUUt 9IJIJGUJi1J CJjGG(2 1? (JiG JUCLGUdG O (JiG {JJICJ<UG22
OjJJJC CLA2WJ2 !11 (PG cpain qILCC( ton MJJGIJ ?CWICLA?(aJJUJG bOJAWCL? 9LG patcq
apoc 9 CCL(9IL) (CWbGL9{OLG [JJG2G cpauca can pc WG9211LG pA CJCCILOU
wicLo2cobA auq 2W9JJ 9UJG Y-L9A ?C9{fGLiU nuq p9AC JJCGLJ 0JJ2CLAG 10L
IJFIWGLOI12 bOJAWCL2 auq 10L A9LiOfl2 wqa 0J ?9WbJC bLGb9L9{ iona' JJJC
TpiCJCC1JiIJ bLOCC22 M92 JIL?1 0JJ2GLAC 011 2011111011 L0MU CLA?(9J2 01 bojA-
GIpAJCnG1.5 f1{ 9Jao OGCIIL2 in pnjje WGJ( cLA2(9JJ1XGq aiJq in p1pjA C{L9MU
2aWbJGay
jJJC CUGL9J IC9WLG2 01 J9JJ9i. 1P!CrCU1U wcaanLcq pA (JiG J9LC IUCLG92G
111 ions 2bacin 9LC MCII G2{apJi?pGq JM0 qJflCLCLJI bow{? 01 A1CM JJ9AG pGGIJ
92?f1WC (0 cxbjaiu IP!? bpcnowcuow U IJJC OUC panq (JiG (piCJCGUiU
bLocC22 12 JJCJICACq (0 OCCfIL 111 IJJC ?Opq ?(9(G pA WC9IJ2 01 2bCCiaJ qwou
wCcpaui?uJ2 M{J1CJJ wajcc bo2?ipJc (pG (L91J2J91 iOU 01 WOJCCIIJC 2CdIICIJCC2
1pL0f1p ipe 19111CC 0" {JJC OIJJCL pauq i{ JJ92 JJCCIJ bLobo?Cq ipar (PC !UCLC92C
01 J9WCJJ9L IJJJCJCUC22 1? qncfO a wcjini Oj CJIJJCL MJJOJC CLA2I9J2 OL baLf? 01.
jpcCLA2{aJ? jojjomcq pA a LCcLA?faJJixaIiou ILOW IJJC WCJI Hi 011L Obinion IJJC
drIc2fiou !2 pA 110 WCIJ? 2cfijcq auq IJJG WCCJJaUJ2IJJ 01 1P!C1(CU!U !2 uoi
nuqcjooq bLobcq7c JJ {JJC jOJJOMi1J Ipi? M1JJ PC CWOU2JLS1G pA ouic
cxbcuwcujaj OpBCLMJJIOIJ2
IIJC anwbjiou 0{ 1 2oJiq 2{aJC qip?iou nJccpaniaw LO !?0fPCLWal
1P!CYCU!U !2 WTiIJjA pacq 011 f MO Op2cLA&11002 (1) II Jill? pCCIJ Of)2CI.ACq
LcbcarcqJA p? CJCCILOIJ WICLO2COb7\ 01 2!uIC janicjjac f[J91 1C OLICU{af 1011 01
fpc CJJa!1J2 LCWaJIJ2 con2fan. ' yucxawbjc Oj rpc 21 LIICfRLC oj an anncajcq
bojAcrpAjcuc 2iuJC cLAaraJ !2 1ACU !'-' ij\ui.c \ y4aLUJaIIJaUCC Oj WOJCCIIj&L
V EI'{" WiCL01.UbjJ 0J1!J ¶f1J1JG9G bo1AcipAjcnc !UIG L{L LPG CIJJ!U OL!CIJUI{IOIJ
! 'lI bcibcuqicnp;i io tpc bjuc O ipc
J1'1
ANNEALING AND POLYMER MORPHOLOGY
orientation during thickening has also been observed macroscopically in
oriented samples3 and suggests that a completely molten phase is not formed
during annealing. (ii) In many cases it was found that the tong spacing after
annealing depends not only on temperature but also on annealing time2 as
demonstrated in Figure 2. If the long spacing Lis supposed to represent the
130°C
400
128° C
126.5°C
300
25°C
.<
-J
0
120°C
200
115°C
100
1 10 100 1000 10000
t,min
Figure 2. Dependence of long spacings of polyethylene single crystals on annealing time for
various annealing temperatures7 8 The original spacing was 116 A.
average thickness of the crystals, then the assumption of continuously
occuring melting and recrystallization seems to be unreasonable.
The experimental results plotted in Figure 2 reveal some interesting
features of the thickening process. Within a certain temperature range the
time dependence can be described in good approximation by
L= L0+ B(T)log(- + I) (1)
where B(T) is a proportionality constant. Remarkably the very first jump
in long spacing after the shortest experimentally accessible times fits the
logarithmic line well, at least for lower temperatures. Further it is noticeable
that above 125°C some deviations from relationship (1) are observed.
Especially at short times a large scattering of the values occurs which we
could not overcome by variation of the experimental conditions. We have
already explained this observation by the assumption of small unavoidable
115
F IA LI2CHEI{
qIUCLGUcG2 !' pG9filJ L9fG8 JSajGL pGafiu L9fG2 jcaq fo WGJIIU anq LG-
CLAIajjiS5fiOU MGLGS Mif p 2I!PIIA UJSjjGL pGSfiU L9fG tP!crGn!n OCCfILB
Mp!Cp 2f9piJiXG fiG CLAafSJ2 SSIUaj WGjfiIJ 20 OUG amA fL9M fiG cOUcJfl2iOn
1p91 fiG JIJCLG9BG 04 jOU abacin !2 U01 qne tO 9 FIUI4OLW bLocea pnf ipar pop
wccpaniwa C91i OccnL: f JJJCJCCWU pA a jiqru qqjn2!ou anq wcjtiu
jOjOMGq pA LGcLAaI9jJ!Saf IOU 2OW elect LOU WicLOcObicaj OBGLA9I loin
fJOM iuqccq 1JJ91 IJJG &nueajiuh G5M0flL 01 fiG bojActpAjGuG JSWGJjS !B 01
a U0UflW(OLW auq cowbpcacq U9{ccLG LOL cxawbjc uk\iwe LGAG9J C]GSLJA
I FI{wJ iJJic1OLLJbp ot o boActpAjcuc iujc cQatj &uucjcq u isc LOL Q win
uq c1pGdnduJJrlncncpcq8
ILGG qIjjGLGLJI LGtJOJi2 01 OUG anq fiG 5WG CLAaIa a CGIJILSJ nnajecteq SLGS
9 wiqqjc 9LGS Gxpipitni jjpuJJaL ifLfICfflLG ?fIbb0GqjA uAtajji.'ccq pow fiG
weji qnun dneucpin 01 {JJGcLA2f5j aIiq Sn OfIfGL LG!0U fjJicjCGIJGq qnLiLI
9IJJJG&J!U
JJJGLG ! J\SLICJ A 01 aqqi{iOUSJ GtJqGncc 11 I jiG L0jG 01 UJGJfJIJ qnLJu
SIJUGSJJIJ {LGSfWGIJ{ 9G[OLG MG qcaj MJ{JJ tjioae GXbGLIWGIJf aowc woqeja to
pc tpicrGuw bLocc waA pG conaiqcLGq- pG G5IJG 01 fiG OAGL&JJ afIMC{ITLG
can pe qcacLipcq aqednajejA pA tpc acpcwc 01 jtrn.e btobocq pA tattou;
Vt f pc IJJOIGCIIJ&L JGAGJ qiucLelif Wccpaiiiawa 101 C}JSJU LGjoJqw JJSAG pGGIJ
bLobo2cq pA KGUGrGLJ bGJGLJ!U1 Jq H0l1W1J1 MJJGLG fiG J92f f MO
bLoboaJ iarc info accornif IJJG fiwc auq fGWbGL9f IILG qcbcnqcncc 01 ipe
1P!crGU!U bLocGa qcacupcq pA cdnaiou j AJOLG LGcGUfJA GJJGL c civ -J
2flFGfGq 9 LGJOJU WGCj.JSIJI2W accotqrn fO ik\nis f5JUIJ pn0 acconuf
2OWG GXbGLJWGIJISJ LG211JJ Wi bojAawiqc
jpc qiEcfJ2aion Oj {pGG woqcja ! 01 LG5( !nrW !f amA 1JWflJ51G
IJGM cxbcuwcuj JJfU L?J i04 I1 5AG fO fLA fO qcciqc T[JG pSBJC dncaion
MG
iutojcq in rpcc woqcj2: ipc fpicrcn!u bLocca abbLobuafcjA qcacupcq
pA JiG S2aflWb{JOLJ 0j a opq 2fSfG ILSIJBJIJOJJS OPAIOIThIA 1}JC SIJ2MGL {0 1P!
dnc2Jiou i cjocjA aaaociatcq MUP rpG btopjcw 04 {G WGJfJ1J fGWbCLSfRLG 04
f}JG OLIthIJSJ CL?\2{5J2 2!IJCG pA qGjniipoir ncp ¶1 fLFJUa!f lOLl jayca bjacc Mi{pon{
IW
ANNEALING AND POLYMER MORPHOLOGY
(a)
(b)
Fiqure 4. The arrangements of chains in sections of a lamellar aggregate. (a) Before annealing
(b) After annealing. According to Stratton9.
transforming the material to the molten state. Unfortunately the melting
behaviours of polymer crystals is very complicated as well. We will see later
especially that the effect of partial melting and the influence of molecular
weight distribution will give rise to serious difficulties in the definition of the
fusion point of' a semicrystalline polymer. Nevertheless, according to
Mandelkern et al.'4 there is strong evidence that in the case of fractions of
polyethylene the temperatureT*, at which the increase of' the long spacing
starts, can be identified with the melting temperature of' the crystals. The
authors studied this critical annealing temperature T* as a function of' the
(a) (b) (c) (d) (e)
Fiqure 5.Schematic representation of a refolding mechanism proposed by Dreyfus and Keller1 3
117
E. W. FISCHER
initial crystallite thickness and showed that T* obeyed the well known
equation of the melting point T of a crystal of the small size c in the chain
direction:
/
2a
T* T =
T,
(1 —
Reasonable values of the equilibrium melting point T, the interfacial free
energy 7( and the enthalpy of fusion AH have been obtained using the
identity T*
In the case of unfractionated polyethylene the situation is more difficult.
For example, in Figure 6 the change in long spacing L as a function of anneal-
100
100 110 120 130
7;
Figure 6. Dependence of long spacing ot solution grown polyethylene crystals on annealing
temperature alter annealing for I hour'5. (-----—-f PE crystallized at 80 C:(- 0
PE crystallized at 90 C - - )fractionated PE (Mq =40000). crystallized at 90 C
ing temperature is plotted for various samples of polyethylene grown from
dilute solution15. In the case of a fraction M 40000,Lremainsconstant
up to 124C in agreement with the results of Mandelkern et a!.14. The
unfractionated materiall, however, shows thickening at 110°Cand 115°C
respectively, depending on crystallization temperature.
The remarkable differences between fractionated and unfractionated
samples are also reflected in the shape of the difierential thermoanalysis
curves, see Figure 7a. The crystals of unfractionated polyethylene show only
one endothermal maximum at about 132°C, whereas the fractionated
sample yields two peaks as already found by other authors'6' Thefirst
peak can be interpreted as due to the melting of the original crystals, the
second peak is supposed to be caused by the recrystallized material. This
t The crystals were filtered at the crystallization temperature, so the very low molecular
components have been removed.
118
ANNEALING AND POLYMER MORPHOLOGY
05 2 8 32
120 140 120 140 120 140 120 1/.0 1°C
L____
120 140 120 140 120 140 120 140 120 140 1°C
Figure 7a. DSC curves of solution grown polyethylene crystals recorded with the indicated
heating rates in C/min'. Upper row: IJnfractionated Marlex 50; Bottom row: fractionated
PE. M =40000. crystallization temperature 90 C.
136
C)
0
132
128
0 60
Heating rate °C/min
Fiqure7b. The temperature position of the endothermal maxima of the DSC curves depending
on heating rate.• Marlex 50. A fractionated PE.
interpretation is supported by the dependence of the shape of' the melting
curves on the heating rate, see Figures 7a and b. In the case of the fractionated
sample, only one endotherm maximum is observed when high heating
rates are applied. Obviously recrystallization from the melt is avoided by fast
heating.
119
E. W. FISCHER
So far the interpretation of the increase ot long spacing as caused by
melting and recrystallization seems to be fully confirmed. One has to keep
in mind, however, two other observations. Firstly, the appearance of two
endotherm maxima in the melting curves of the fractionated samples also
depends on the crystallization temperature as shown in Figure 8. At low
30
C
> 20
10
stut850Cj)i\,
0
100 120 1i0
°C
T,
Figure 8.USC curves of a polyethylene fraction (Mq = 40000)crystallized from dilute solution
at the indicated temperatures's. Heating rate 1 C mm.
crystallizat ion temperature (80C) only one maximum appears. Secondly,
the temperature position of' the second peak depends not only on annealing
temperature but also on annealing time, as demonstrated in Figure 9.
During annealing at temperatures above the melting point of the original
crystals, the melting peak due to the recrystallized material still shifts to
higher temperatures with increascd time. Accordingly small angle x-ray
measurements showed that the long spacing of crystals obtained from the
traction increases from 160 Ato240 A when the annealing time is increased
I'
I I
I —Annealing time lh
20
> Uncinnecited
10 Annealing time
305
0
100 120 140
T, °C
Figure9. I)SC curves ofa solution crystalliied PEIraction annealed at 128 Cior various times'
120
ANNEALING AND POLYMER MORPHOLOGY
from 1 mm to 1000 at a constant temperature of 128°C. These results
mm
indicate that in addition to melting and recrystallization another process is
going on, which results in a continuously changing long spacing.
The same conclusion could already be drawn from the results obtained
with unfractionated samples which are plotted in Figure 2. Even if the
solution grown crystals melt at annealing temperatures beyond 125°C, it
is observed that subsequently the long spacing of the recrystallized mat
increases continously with time.
In this connection it may be interesting to look at the annealing behaviour
of solution grown crystals of polymers different from polyethylene. It is
known that polyethylene oxide (PEO) crystals also undergo a drastic
alteration of morphology during annealing'8. Our own studies were per-
formed using PEO of Mq = 32000crystallized from 04% xylene solutions'9.
The dependence of long spacings on annealing temperature are plotted in
Figure 10. There is a clear break in the curve at 57°—58°C. On the other hand,
2
g'200
U
60
50 52 54 56 56 60
Annealing temperature,°C
Figure10. Dependence of long spacing of solution grown polyethyleneoxide crystals on annealing
temperature after annealing for 10 mm19.
in the DSC melting curves of the same crystals, the first indication of melting
is observed at about 58°—59°C with a temperature position of the endothermal
maximum of 62.5°C. Taking into account the fact that during the DSC run
some reorganisation has already taken place, it may be reasonable to assume
that the break in the curve in Figure lOis caused by the melting of the original
crystals. Then the increase of long spacing at annealing temperatures below
57°C is probably due to a thickening mechanism which is different from
melting and recrystallization. Remarkably, in the L(T) plot of Figure 10 a
plateau at about 170 A is observed. This value agrees rather well with the
twofold value of the original long spacing. This observation may be considered
as additional evidence for the model proposed by Keller et al.'3 which was
pictured in Figure 3.
The described results of small angle x-ray scattering studies give rise to the
121
E. W. FISCHER
intriguing question whether there is sufficient mobility of the chain molecules
within the crystal lattice. In order to allow for a complete reorganization of the
structure without melting. a coordinated translational motion has to be
assumed. The mobility of the chains in the lattice can be studied by means
of the motional narrowing of the broad component in the nmr spectra.
Figure 11 shows the temperature dependence of the second moment and of
16
24
12
U)
a)
-c E
0
16 E
-o
c C0
4 c ta-)
U)
0 3
-100 -60 -20 0 40 80
Figure Il. Linewidth and second moment of the nmr broad line spectra of PEO solution grown
crystals as a function of temperature19.
the line width in the case of PEO solution grown crystals'9. A drastic
decrease of these quantities is observed and rather low values are obtained
at temperatures beyond 50°C, where the thickening process starts. In this
connection we refer also to the extremely interesting results of Kovacs20,
who observed thickening during the crystallization of PEO single crystals
from the melt.
Summarizing the results about the effect of annealing on the change of
morphological structure one may conclude that two competitive processes
occur. On the one hand, whole crystallites or parts of them may melt and new
crystallites develop from the molten phase. On the other hand there is
evidence for a thickening process occuring continuously and involving parts
of the original crystals. Which mechanism predominates depends on mole-
cular weight distribution, crystallization temperature, heating rate and
annealing temperature.
The distinction between a melting—recrystallization process and continuous
thickening does not mean that in the latter case no molten material is
involved. As we will show in the next section, many experimental results
indicate that during heating a partial melting at the surface of the crystal
takes place which we call boundary premelting. This process seems to play
an important role in the thickening mechanism.
122