chain orientation and crystallization · the prerequisite for crystallization is the retention of...

CHAIN ORIENTATION AND CRYSTALLIZATION

A. KELLER and M. R. MACKLEY

H.H. Wills Physics Laboratory, University of Bristol, Royal Fort,Tyndall Avenue, Bristol BS8 I TL

ABSTRACTThe article starts with the assertion that in order to obtain fibrous crystals withextended chain type characteristics the molecules have to be extended priorto crystallization otherwise chain folded lamellae will result. Further it assertsthat in the case of non-homogeneous chain extension those chains of the assemblywhich are inadequately extended will give rise to lamellar crystals nucleatedalong the fibrous ones already formed. The main body of the article providessubstantiation of this generalization and demonstrates its consequences in anumber of diverse examples. It is divided in two parts according to the twobroad categories of chain extension. (1) Static elongation which is applicableto crosslinked melts (elastomers) and (2) chain extension by flow achieved byelongational flow in solutions and melts. The part dealing with (1) is basically areview centred largely on past works by one of us while the part concerned with(2) is a preliminary account of work in progress. The latter in particular describeshow characterized elongational flow fields can be purposefully created and howthe resulting chain extension and the ensuing crystallization can be studiedin situ as a function of a number of variables. In the course of it reference ismade amongst others to criteria for complete chain extension, to the role ofentanglements and to the role of localized flow fields in the vicinity of thesolidified portions of the flowing system. Attention is drawn to the diversityof effects which all arise from the same few basic principles and to the numerous

fundamental and technological implications.

I. INTRODUCTION

The importance of crystallinity in macromolecules both technological andbiological needs no introduction, nor does the long-recognized fact that suchcrystallinity arises through parallel alignment of sufficiently regular chainswhere the repeating units in the adjacent chains are in appropriate crystallo-graphic register. From the very beginning therefore the chains were consideredto be running straight through the crystal as long as entanglements, con-sidered inevitable at some stage, permitted. Correspondingly, the basiccrystal unit was thought to be elongated along the chain direction andessentially of fibrous character. This conception was in accord with thewell-known fibre-forming tendency of long-chain molecules abundantlyutilized by nature and technology. In brief, it seemed all too evident thatfibre formation should be intimately tied up with a basic fibrous crystalmorphology.

195

J TQ

llJq MI(JJ fpC nuqGLJAiU cpnu CXIGU2IOU If ''lI qq MJfJJ fJJGJL 0LIIU uqbccu 1Lf!CJC Mill pc COUCGLuCq Mifp f pG J0LUJJ1J0U 0L1PLOf12 CLA2fIJ2

2fiLLcq 20JnfiOiJ'jioiu pA Ic jiicp 12 jWIJWL cxrmbjG O CLA2ffJJi1fiOU LLOW r

couboc 2fLf1CfflLG UTWGJA 1 jLOfl2 pqpouc M1f lWClJ&L OAGLL0MfJJcpnu-oJqcq cLA2Wllirfiou °L fIJG o. {JJG UJOJGCI1JG2 JJJG LC2(IJf 12 fpGU1IJ2f1UCG JAiU LI2G fO ppLIj2 JJJG2C UPLJJ2 MijJ fpcu bLoAJqc LJfICJG! LOL fG20 fpGIi OIJIA fJJO2G pJAiIJ 2fIjjiCiGUf GXfGU2IOU Mill CLA2fjJiG iIJ fJJG UL2fMJJCJJ iii 1AC1J 2A2fCIJJ UOf JJJ fG C!HU2 TLG cxtcuqcq OL jc UOf cdnjjAcxcuqcq Gj9LG CLA2WJJ11tJOU LOfl2 CLA21Jj CLJfIfIG2 MiJJ LG2fIJV b11LfPGLqo 20 pA C1IU 10N!U !t!U LI2G fO J7UJGJJ1G fJ0MCACL MGU fJJG CJJflU2 LGoiu u rIuqi2fnLpcq 2pfi2ficrJJA Lrnqou couunuwou !t iqj cucurjiA

10 fG CUCLfl couqrl2iou: iipcu cpuu woccnjc CL?2frJjic2

j qibjiA i cowboic UpLon2-IJcJc1 (p!p-jccpp) rLnCu1Lc uuiu2y Abic CL I!IiOIJ bLoqnc opmiq pA 2iwu2 rIbcLcoofcq oJnjiou o bojA

COGXI2f ill bLGAiOn2JA flU2fl2bGcfGq 1JJTUUCL (LoL LCAI©M2 2G LGj2 J 5)qcbduq2 OU fJJG coIJqifiou2 Ot cLA2f&JJI1fioU 1uq pGdncufJA Of IOLUJ2

MGLG G22GUf1qJA 0LLCCf MJJGfGL rpc CLA2fJj2 JLG IJWGI11L OL UPLOfl2qJJcwwT Jf pcrwc bbLduf fjJIf fJJGLG M12 HO LCJ C0UuiCf &uq fjJf 0fJJ

DGAGI0bWCUf2 qf1LiJJ fjJG l2f GM AGL2 20MG fJJG MA Ofif 0J fJJ12LLOW bLiwLA iUfflif!0UCJJfiu2 LfrJJu1u 2fUflf pLonp pcur i bJcfnLc MpJcp 2CGUJG to 0JJ0Thipjc LGJ11CfUCG fO paiqou fJJG G1LJ1GL COIJCGbf2: IJJG UPLOfI2 CLA2tJ2 M!fJJ1GfJTLG 01 CLA2 jiIU JJJfCL01JJ0JGCf1jG2 EACH 20 1CLC Lcwwcq 1 cou2iq©L-rrnqcLJAJiJ ofqcq-cjrnrn C0UJf1L1fI0U MJ2 200U LCC0UiCq 12 ¶1 GUGLfJtG UJ0JGC11JJL COIJCGbf 0 CJJIJTJ oJqiU 1P12 I1UJGIIL CLA2PTJ jJpi MJtp IIJCJ0U fJJC J0IJC21 pfrt 2JJOLfC2f CLA2WI qImcU2i0U ¶1 jjUqiU MJJ!CJJ f2 jcq fOGAGLJtfHIIJA 0ptJJuCq in i20Ifi0Ua bLoAcq fo C jTWCJjJL Mifp f JiG CHU2 UOf

If IL1G fPGLCIOLG J2 1 WJ10L 2HLbL12C fJJ1f tpc p&2ic CLA2fj (TUita MCIJ

\/ KEFFEI� /1.4D 1/f ic 1/IVCKFEA


structure. Properties, interesting as they are, will not be included exceptoccasionally mentioned in passing.

The material is most readily subdivided according to the method bywhich the chains are extended. Depending on the consistency of the systemchain extension can be achieved essentially in two ways:

(1) Static extension. In this case the material is stretched and kept in thisstate until crystallization sets in. The prerequisite for crystallization is theretention of orientation, normally associated with the absence of stressrelaxation on the time scale of the crystallization process. This is satisfied bythe networks of crosslinked systems where the ensuing crystallizationcorresponds to the familiar stress-induced crystallization effects of elastomers.

(2) Dynamic extension. In this case chain extension and ensuing crystal-lization is induced by flow. This clearly applies to true melts and solutions.As will be shown, chain extension sufficient to induce crystallization willusually occur under rather special flow conditions where the flow velocityhas a gradient along the flow direction (elongational flow).

Highly viscous melts form a bridge between (1) and (2) as here, owing toentanglements; the melt may sustain sufficient stress to display the charac-teristics of a crosslinked elastomer on the time scale required for localizedchain extensions and the ensuing crystallization even if the molten systemitself may be in a continued state of elongation (until of course flow ceasesdue to solidification).

The present article will give a broad survey of phenomena under (1) andwill briefly report some recent material under (2).

2. CRYSTALLIZATION OF CROSSLINKED MELTS ANDELASTOMERS

2.1. General scheme; columnar crystal texturesExtensive past work has led to a generalized scheme of crystal textures

formed under stress such as represented diagrammatically by Figure 2 whichwill now be briefly described and explained.

The basic observation, whether directly observed microscopically, ordeduced from x-ray diffraction patterns as it was done in the first instance6'7is as follows. The final crystal texture is columnar with the columns lyingall parallel along the stretch direction where the columns themselves havecross-striations on a finer scale. Depending on the column diameters, on thescale of the observation or on the size of the area under view, the overallimpression can be either (a) a cross-striated structure (fine striations per-pendicular to the stretch direction) or (b) a fibrous structure with theimpression of a fibrosity along the stretch direction. While (b) is at leastconsistent with the traditional expectations from fibrous crystals arisingalong the stretch direction, (a) appears to be at variance with it.

The explanation will be clear from the generalities laid out in the Intro-duction. Applying these to the present case we may state the following.The extension introduces true stress-induced crystallization along isolatedlocalities. At this stage we abstain from specifying the nature of these stress-induced crystals beyond stating that they represent only a very small fraction

197

A. KELLER AND M. R. MACKLEY

>cz __>z>zThin fiLm

(a) (C)

200200

020( ) 020 ( ( ) ) 020(

(b) (e) Id)

Figure 2. Schematic representation of the crystal texture originating during the crystallization oforiented polymeric melts shown for thc particular case of polyethylene.(a) Low stress. (b) The main features of the x-ray diffraction pattern corresponding to (a). (c) Highstress. (d) The main features of the x-ray diffraction pattern corresponding to (c). (e) The mainfeatures of the x-ray diffraction pattern for stresses intermediate between those in (b) and (d). Thecorresponding texture (not drawn) is envisaged as in (a) but with lamellae which are only partiallytwisted.Only two reflections, 200 and 020, are represented in the x-ray diffraction patterns and themorphologies are drawn both as expected in the bulk and in a thin film. It is the latter case whichis more directly comparable with electron microscopic evidence. (Keller tnd Machin4 based on

ref. 5 by the same authors.)

of the total amount of material, comprising those chain elements whichhave been most highly stressed in the network and that by inference (ratherthan by direct observation so far) have an overall fibrous character. Theseentities then provide lines of nuclei for essentially chain folded crystallizationof the rest of the material which is in a lower state of chain extension. Thelamellae themselves will grow transversely outwards from the nucleatingcentral line producing a transversely striated entity of cylindrical symmetry.In the melt, the resulting entity will be a compact column [as opposed to thelooser structure resulting from solutions (Figure 1) referred to in the Intro-duction]. For low extensions the nucleating lines will be few, hence thecolumns will be wide, and the overall impression may become that of abasically cross-striated texture [(a) above]. High extensions producenucleating crystals in greater abundance, hence the columnar texture willbe on a finer scale, the overall impression becoming that of a fibrous texture[(b) above].

2.2. 11w transverse lamellar crystalsThe degree of stretch, however, can affect not only the number of nucleating

198

Thin fiLm


threads, hence the scale of the columnar texture, but also the configurationof the transversely growing lamellae.

In an unstressed melt the crystallization usually starts from point dis-continuities due to nuclei of some kind, fanning outwards resulting inspherulites. The crystalline units are lamellae which normally twist in aperiodic airscrew-like manner and are responsible for the familiar bandedstructure of most spherulites. The corresponding texture in the case thatnucleation of the lamellae occurs at closely situated localities along a linediscontinuity is illustrated by Figure 2a. This is the case expected if the stressis low enough not to interfere with the twisted lamellar development, but issufficient to produce the line nuclei. As the representative part of the materialwill be in the form of lamellae the overall crystal orientation will be such aswill result from this particular twisted lamellar arrangement. The cor-responding x-ray diffraction pattern will be neither a random pattern, norsuch as from a fibre with chain alignment along the draw direction, howeverpoor such an alignment may be, but will have characteristic features of itsown such as sketched for the case of polyethylene for two principal re-flections in Figure 2b. Indeed it was the consistent appearance of this peculiardiffraction pattern in standard, melt-extruded polyethylene films whichprompted one of us to postulate a texture such as in Figure 2a for melt-extruded polyethylenes (termed 'row structure') as far back as 19546.7.

Higher stresses produce not only more nucleating lines, hence a finer scalecolumnar structure, but will also influence the lamellar development. As tobe expected, crystallization under high stress will align all the crystals withthe chain axes along the stress direction. Thus, the lamellae will be increasinglyprevented from twisting until they are all parallel within the columns assketched by Figure 2c. As now all the chains are along the stretch direction,the corresponding diffraction pattern will be as in a drawn fibre (sketched inFigure 2d for two reflections in polyethylene). For intermediate stresses thelamellar alignment will only be partial and in the case of polyethylene theresulting x-ray pattern will have features as in Figure 2e where the lamellaralignment is most readily characterized by the magnitude of the splittingof the 200 reflection*.

As already stated solidification during flow, such as occurs during themanufacture usually [even if not always (e.g. ref. 8 and comments in ref. 5)]produces textures such as in Figure 2a presumably because the moltensystem can only support low stresses. However, the full sequence in Figure 2is readily realized if the system is permanently crosslinked. The correspondingseries of diffraction patterns on polyethylene were first obtained by Steinand Judge9 and the sequence was interpreted in the manner just summarized

Ln the case of polyethylene a diffraction pattern which is similar to that obtained in Figure 2bis often reported which nevertheless does not correspond to complete randomization around 1,but shows a preference of a along the stretch or flow direction (the frequently mentioned 'a axisorientation'). This is sometimes quoted as evidence against the scheme in Figure 2 (e.g. ref. 8). Aspointed out in ref. 5 there need be no contradiction. Such an orientation can be accommodatedwithin the scheme of Figure 2 by considering the lamellar planes in Figure 2a to show somedegree of alignment along the direction of the orienting influence instead of complete randomi-zation. The lamellar, as opposed to molecular, alignment such as would lead to Figure 2c, mightoccur at the very low stresses where the a axis orientation' is usually observed.

199

YHJADAM .51 .M It4A SIHJJEDI A

alili b Qi1idias31 3111 b3111t3v aiorllus ialisl odT .anid3sM bits 13113)1 sd b3si11siasl3-aa3lia no 'qooaoIoirn no1i3313 noiaarniansii d noiisi3lql3tni

.amlil null aa3lia no 3on3bn3qob 31s3a all bits owixoi isnmuloo b3isiita-aaoi3 3clT

ln3fftqolov3b all .lsal3vinu 3d ol aisqqs noitJtSiIIslal(13 b33ubni aa3lta 111

e215ub315m 3n3motasl3 us lo Is3iqi taom th bowollol (13n13nwfIO3 3d bliroD

3(11 bovl3ado tail! "'°'aw3ibnA lull 313(1 asw II .l3ddul Isiuisn 1(13mM!

b3133qx3 3(11 ni ion bits 1s1u3ibn3qi3q 5 th W313 alsiasçrc odi 15(11 losI gnbliila

.aoftliJl3clqa 'b33n313u0 woi niwoc1s hjdo ôô nolyM bsbluom noii3SifIi fin 10 notiio'l I (e hi a wlbA) abioisloq b3aaon 4qnonimoiod

3(11 asw ii tsili bits 1131311210 noilD3lib 3(11 01 133qa31 1111w efiOiIO3llb 1311515q

its aa3i1a 3111 01 1311slsq 313W rbirlw a3nil no1s b3lluaDo rhirlw noils3lDun 3111 olni all! jboth2oq ln3a3lq mo moil wonA 3W as 4oiilw nOilsVl3adO ni oals as l3ddui lo 3253 3111 ni e13V3WOH .b3ln3aolq won 3m3f13a 15131133

('3isnodmsos(loq bits c '3n3ms(ias(loq 3itDsioai .3.3) alsimaism 13(110 302 3(11 313H .b3vi3ado n33d Ion asil & nsk ni as laiwi lsll3msl lo 33fl3U32 3(11

aa3lia 3111 ol l3Ilslsq anisrla 01 b3bnoqa3mmoo a1(swls noitsin3ilo msll3msl lo 1(1110 3ilan3loslsflo ai 33it3U32 laiwl 3111 13111311W .zS siuvik ni as noii33lib

-aaom 10 33133b oili as rlona e2101351 131110 3ffl02 13(11311w 10 alsn3ism nislm3o ln3a3iq 3111 lo 33b3IWOIIA 3111 oT .In3a3lq Is '(52 Ionnso 3W '(slq 183113 e3itAfl1

1311113 lo a3111u1311qa lii 351131(151 lo 3nulaiwt no hoq3l on ai 313111 aioillus Ion ob 35113ms1 odl 31311 IsdI 3101313111 3ldiaaoq ai II .3n31'{Ia'(loq 10 l3ddU1

1111w b3ishoaas ai gnilaiwi 15111 1(13111 all!) .oithsi & nriol b3taiwt s iii wom Ias3l Is 3n3v(la'(loq lo 3aso 3111 ni rloidw lids11 msll3msl 3)hl-13ms311a WOllSfl

(.'3aso 3111 ton '(1n151133 21

noilsmraulli as 3V132 bluorla ?T-t nss½1 111 arlqsigoiorlq lo 33fl3Up3 ,IIT .boin3a3iq tauL noiissilsm3rl33 3111 lo 23253 lsio3qa 3ffl02 lo

oo

VIO!TA1JJATY5LD UI4A 14OITATMIJMO I4IAHD

rro-b3 tiwod (nii1CL) (3rihthsm'xo)1oq lo isd (xbhiom noiio,jrii as lo á1iu . rio biid 'iivsU) .&li!q31 dqs1o13im no13I nol tib nO1)Irir oi i ranst nuI3msI

(kinD d b3bivotq noiqitni thiw r1qnioorIq 1nniito t)11Ei::

i9I3un to niI bubn -9d nIt 10 3luitrn 34T £S ion i nmu!o r!i rlif!iiw rnI niiiokun !inin3o rIi lo b!worrA iuO di o thom nimIdwivo ccii i1i i ir1i iol noi r1T .viiiriikb

on !.si1 innoqmo3 aniikusn th briE rinol iB11mEI r1i ni i Ishiim r1qE1o13im di ni o 1IiEq noiiatiThb 3c11 ni 13thi3 qu niworf 10 rnrk

bluow iuixi iii1iutriq B 1rio ii iuodiiw ai Jsiiv i 3n1ix ii Ii nv .1

3mo3d aiu! iv3wor! jBniEm gnuis!3un i3rrii2ib lo 3r13iix r1T b3sillEie1-noiflJIo no how in11uflIo3 ni I!EMqOoo13im in5LEqqB

b11E3 o r(i) jim 31uixi in3Isviup 3rh iii rni! uBa3d 'IthEm emi iooI iibum i iui3Intg IrIIrn13I IVrIEII th ( iti1 th dsdI-HirIg b3daiLdEi aw Ii iH 2nOil 1i1trrn3 ,rli 10 noii3b IEueiv niflirrriq

cfi cf bmio1 ioIor1qiom 2urndil 3n!Ufi i asri \ff13 riiisburi rli iBril IEmr1i tthiuI .noii3iib 3ldfl dli noli3 aniEr!3 r1i diiw mIoq rna bbnix lo inuomB thrti s 1as! 1i3 biE3ibni aihqoiq IB3irtEd3rn bnE

.idfl 3arIi nirliiw if1tfto3 niBfI3 !Jnjn3 3111 bisi as nibnu1 3di IqqE 01 nuiqm3i 001 lIE b3m3 ii

31u13tn12 ulnmuIo3 auogolEns 3111 01 m3iava nwo1 rioiinloa 3111 Iii bB3lth lo iIzJ3i11th 3111 10 W31v iii 223lOf1i13v3V1 . .1i th as iI3rrr 3111 th

.iqa31 airli mu IfI3m331B bii,n on i 3111i n3biv3 3viu13r1o3 ninjE1do I O

_______*

Y3DflAM $1 .M UfrM R3JJ1A .A

nurbEM I ,M) bnu bilIEJi (lliir1Eq ,3n1iloq 3i1nJoi io mlii nidT .. soulU iii bdiIduq

ai b,simmmu 3d iim noiiEuii2 di noiiii1i!qmiivo 3ffl02 lo 1ii di iA wo11o1

3di uboiq rbidw i3bun 3di bivoiq 31i2 isdi bis 'HJ13fl3 i ii niinoi 1idW .nmu1o3 niiIui di bitE IbmE1 rTiqobv3b \I1wI3jEI niio bisnoI 3ash1u2 Isnioini aB ibu ii3no13j,d 2uo3nJ1ix3 iirfi

2nmuIo3 rli oi ii 3Vi bru3 i3bun ai tai oIs ni in31q Ii ,ibod liE s1q IE13n3 iii ob 11ii 13m1oq 3di tEdi n3biv3 in3i3ffiue i idi

£ tEch a1E3v31 oIE 3z)n3biv3 ma dT .1n3v3 noiiibun 3di lii tiEq trIEtloqfrn B nivEd 12111 ciriol ibiriw 21Ei-I3 3d1 ,b,vlovni 2! noiiEsi!1Ei2'13 3Ei2-owi

111$lOvo ,fli lo 3vi133q2311i nOii33lib 2oIt 3th 01 13!IElsq noiiIin3i1o thirb 2idT .aEi2 l3iEI c1i cit nimio1 (3ELI3mEI) 21Bi2\13 3d1 lii floiiBifl3ilO fliBfi3

fioiiEs.i11Ei2(13 3di t3dw 3n3tditoq no ain3mil3qx3 d b3bivolq ak n3biv3 iq3A ni3d aiw 3IqmE2 3di 31111w noiizElIlib vi-x d b3wolloI 11Eui3E 213W

fn3iiBq tEll £ oi bo1 01 as d3u2 wot asw nodo 1i2 3111 iH . 1b3rbbli2 biqx 3111 113111 oviic3ldo r1i rliiw o ?i% ni 213 bnbt di lo 31dBr12iunui2ib 3d bluo.da rn3231q tuit Ii oasdq fIiiE3bun 3di lo fIoiiEin3flo

r conarw joiq pc 2nubic JLa cxcuqa uq jct CO!ULC2 jjic UL2MCI}U OL Ojjc1 qicLcu cxbJrnfl!ou2 jjjna in pc COflL2C O cLA2Wjll!OUflLflJ cLAa11qps11oU cicu Jpc qqjcciu fflfOL2 iAG ! q!GLC1UOp2CLMtIOU2 o ai.caa LG1flCiOU—OL JGUp CJJiWC2 9 couanJ joq—

jpc u1O-211JG cLAa11JJ1iaf!oiJ bpcuomcuou 12 & jca coua!2(ciu 'wia 2nwWL!Cq pcLc ¶LC MCjJ C2WPJ!21JCq

accouq 4LUflLC accina o qcAcjob coiu!urlonaJA pow ipc uL2) JJc caaClUWjo pcac bpcuomcirn iLG bLC2CIU!A flUCL LcucMcq 2cLIIpUA1 (tpna q.Jciuqnccq cLA2J1I!ou o LflCL n LOOin CWbC1 flLC 2OUJC i12bCC2

fcmbcuflnLce ijncp ia pc cjcct O2CLAC qnLw pc IIJJI1!L LCACL2jC 2LC22qi2bbc1Lcq ou LCWOAJ o pc 21LC22 Mp1c 211Jj U JJC OLU1J CLA2jj!OUbccur i cLAaf1IIJc11ou jma uo Ac bLoLcaacq t1La IJC Iu1i11 bccwcuq 'c MCLC LcboLcq) cILcnuJalucc conjq 2MiWb pC iu1t1) C JX12 f AbcCACIJ ij pA uo UJCInJ2 jj (acc Lcj jg MJJCLC comboajic bucLna OMpicpii ojjocq pA i bmci.u ancp ca u % %LC' OL ipicp nuqc woaçjc nJJJcq juqccq llL2 C LX12 ouciicq bcLu ia in bbcacqow oJ pc anpacdncu jrmcjjrn OACLLOMW lpcac cbccwioua MCLC in

o bJLrn1 ll1i22 JLflJUJ!20U C1C1L0U wicio2ubp 211G22 qJLGCP0IJ ACLPCJ (iccJIcl wqJJ0MflJ2 uppøu fL&UAGLC fO IJJC LM qLGCI 100 JJCJJ OLG CIIJJCL JIj1A IM11G 01 OLG !J ALAIIJ2

1trn jprn m o bo1A1p)rJc ct?aopq nuq j0 01 j0-iDlLWq&1G 21LG22c

CHYIkI OJflEWIVIIOI4 VMD cKAaLvrrIw1IoLI

101W

rh 1

(-4

lEnm

uIo3

ni

woc

k ii

did

ibnu

b ill'r

3nIr

!boq

lo

m

Iib

nifi

T

()

. uk

no

-b

oi

bnoq

11o3

11

3111

W

anoi

tni

iniu

i th

Icvr

(f

i)

10

1iii3

b bm

1n3

(d)

.utu

i .ll

,rnI

b

wiv

bn

rI

I3A

) I

iht

ioiii

ib

eri

.dqi

o13i

m

noiI

zi

oiai

mas

iiiT

302

1.OL JJILfJ.JGL MOLJC aponjq IJOf O2CflLG IJJG OACLIIJJ ASJIqIIA 01. IJiG 2CJiGtUGJJiG SOAG COU1LOACL2WJ !2211C2 JJOMCAGL' MJJIJC bLoiqqu Th0M1p boru:a

qiLccf UJOLbJJOJOICt1J OpCLA&1JOU2 ! LflGLqfluLJ IJJG UL2I 2WG 0 CLA2WIJ!StiI!OU &2 WGIJflOUG FlpOAc7 &uq pA owcboaaipijiiA 12 0LUG orq pA tpc LGACL21pJC UtJUILG 01. fiG qLffC1OU bffffGLU10 fiG qcAGJOBIUE suq &sqntiA !ufGLJOcY!u JffUJGJJffL 2fLflcfflLG JJJ!22{LG22 OLUG pA fiGUJ LGJff1G 2ffp2cdrlGufA MJJGU fib JOff GCffUJG {LSU21.GLLGCJCOIIL2G IJJ&I 2fICJJ IJJLG&CJ2 conJq JJ&AG L0LWJ UL2Ie Pill CJRSBBGaLGq ti? IJJG°1. ff 1iioLbpOJOJcffJJA qiafuci uncjctijw lpLcffq1 . a boaaipjc 01.jpcq jo bL0AIqG qILccj 2nbbou 1.OL 2ncp cJffIuJ2 tiuq in tcc qcuA fib GXJ2{GUCGUioLbJioJoIcaJ qtiitt iucnqiu qtrqc-pcjq GJCCf LOU micLoLsbp2a Mpicp 20 1.ffLJG22 ! L'J CAffJrnIfioU fl!2 Jiff2 10 po ptijtiuccq ffffIu2 B bLcbouqcLsucG 01.suq ti GGLf ff111 tiWOJIUf 01. fiGLIJ.IffJ qw W 1.SAOIIL 01. fiG JffUGL2. . WGACLIPO-MOflJ LGCIIIILG JLPGLG tiLt 2OIiiG wOLbpoJolcffJ Iuqicffulou2 1.OL rue [OLUJGLJ .Abc CLA2IffJ2 ff2 fiG 1.1J ff1JffJOA MIfJ lP 2011111011 &OJAJi e2pi2p-jcGpffp2WOLbJiOJO!cBJJA q2r1ucr CCUILffJ pLcffq UOL fib Gx!2rcucG 01. c1lGuqGq-cptiubpff2G ff2 G1GCfG pA qiffLffcflolr qoca uoi Aci bLoAt G!fiGL fit GZI2fGUcG 01.

JJJG tipOAG jjU1U2 01. ff q2fuc1 C ff112 OLlGUlG 21LG22-iUqnccq UfIcJGff1!UcpffiU-oJqcq cLA2fffJJJsffpou miqeL cCLffw 2bGcqJGq c!LcnmafffUccwffJ!UWGU1 MJIJG fib JffUCL tif JGff2f conjq p€ GUAi2ffGq 92 11 cOU2GdfIGUcc 01.pcpffA!O11L 12 cxbtcjcq JLOU.1 (JiG CLA2( ffJJ!sB1 iou pA couAcUliouffJ cpff!U

GJGC4LOU WicLo2LbJr ounnw açrnucq (uqLc/n1 )i' w JJJW o LI1CI CLA2rVJJJSGq nuqt LG22 2iLC2 qlLcc$Jou fJOIJ2 &LLO/A LUflJ2W!a2!Ou

CHVII4 OffIEI4JLVIIO14 VI4D cIcAaLvvrIsvnol4


presented here. This scheme has the merit to provide unity to a range ofotherwise disconnected phenomena and to bring together historicallyseparate aspects of polymer crystallization. The technological relevance ofthe foregoings should be self evident. As regards the disputed particularsrelating to the nucleating phase it will be apparent that the only firm point ofreference originates from flow-induced crystallization and this from dilutesolution (Figure 1). Consequently, it is to this aspect we shall now turn forfurther knowledge on this problem. This is along the established tradition ofwork on polymer crystal morphology, where the study of crystallizationfrom dilute solution has always led the way with applications to crystalliza-tion from the condensed phase to follow.

3. CHAIN ELONGATION AND CRYSTALLIZATIONDURING FLOW

3.1. Shishkebabs; their origin and structureThe decisive experiments in this field are due to Pennings and his

collaborators3'27 who precipitated polyethylene and polyolefins fromsolution by stirring. As well known the result was a precipitate of overallfibrous character. Microscopic inspection revealed the composite fibre-platelet shish-kebab texture already referred to by Figure 1. From the verybeginning this was interpreted in the manner already stated: fibrous, chainextended crystallization of those chains which were stretched sufficiently byflow followed by the lateral growth of chain folded lamellae due to un-stretched or less stretched molecules nucleating crystals along the fibre.

The degree of chain extension in a given flow field is primarily determinedby the molecular weight: the longer chains will be progressively moreextended. This was recognized first empirically; the physical reasons will bepresented later. Thus in the broad molecular weight distribution of a usualpolymer a spectrum of chain extensions is expected. It will be apparent belowthat the transition from slightly to highly extended chains is fairly abrupt.At this stage it suffices to state that the molecular weight spread can play apart in the formation of the basic two-component structure of the shish-kebabs. If the crystallization temperature is sufficiently high only the back-bone forms, the lamellar overgrowth then develops on subsequent coolingof the solution. Clearly this overgrowth can be prevented by filtration at thecrystallization temperature in which way the truly flow-induced componentcan be obtained in isolation. Even when formation of the overgrowth ispermitted the overgrowth can be dissolved by subsequent washing at hightemperatures leaving the bare backbone in isolation. In what follows weshall only be concerned with this backbone which is the basic flow-inducedcomponent of the system. (It is this isolation of the nucleating thread whichcould not be readily achieved in melt crystallized systems.)

As already stated, the chain orientation was along the shish-kebab axis.As this was true both for the central fibre and the overgrowth, the shish-kebabcan be regarded as a looser version of Figure 2c. Further, as already referredto, thermal properties (high melting point. superheatability) and mechanicalbehaviour (high modulus and lack of extensibility and associated clear

206


fracture across the fibre direction) suggested an extended-chain characterwithin the backbone.

For considerable periods it was held that this basic backbone is a compactfibre even if it was argued whether this fibre is structureless or striated asrevealed by certain electron micrograpbs28 In view of the fact that the chainslie along the fibre direction, and further that the striations were due tomaterial density fluctuations within otherwise smooth outlines, it wasinferred that a certain amount of back-folding is present within the fibre28' 29Subsequent works, however, revealed increasingly that even the centralbackbone was not always smooth, even less structureless, but could reveal aplatelet—fibre composite character27, i.e. they can be shish-kekabs them-selves but on a much smaller scale than the originally defined entity (see refs.27, 29). However, here the platelets are not removable by dissolution, hencethey must be molecularly connected to the central core as representedschematically by Figure 927, (Earlier such structures were obtained by nitricacid etching of smooth fibres28; however, in the later works referred to here

Figure 9. Schematic representation of the basic, flow induced crystal entity. Note it has a 'shish-kebab' structure where the central, essentially chain extended fibre is molecularly connectedwith the chain folded platelets, the two thus forming an inseparable unit (Pennings et a!? 7),

207

'I

Y3J)flAM .51 .M 014A 511JJ3A A

woll )laBd di : bvIovni trknE1if1 owi 3f1J niiiiitil1i 1LJiMJ1J d8dA-dir1 •)1 io½ iriupdu molt nij1ui dBd.A-rkid I iii1 thom i1i bOB (bworiB) nod±Ed bubni noiimniciT .ni!oo no biioqb fbirlw 1BniEm bfko$flInu 1euoiviq cJ diwoiivo 1BI1fflE1

.(13Ib){ bilE v1AoM) dqE1o1im f1o1i313

di nilq iiil di iii (.nir13i iuorliiw ..i brnio1 bniEido iw 13 iio iuoo lo idfl imiiq th lo 1uiEn d13dA-rIid rIi lo noiJifIoo1 niiino dBdA-fkid bvido t!sniiio rli rkiunui2ib oi wodlo m1doiq

.nodA3sd uu di moil diwoiivo niun bn 3ldfl b3ubni-wofl lo orli iii 'iiuidms fIiEil3o S 1o11f1i i 1f!i inioqwiv Jniq iuo moiI

imn3 di lo 1uiJ3n dr 1E!uoiiisq nI .iuisiJi1 iiIis ni iinu lo noiiinihb biIiinbi !isft d l3wLs ionns 1qJnoTMm-no1io3b ni niiuis1

-siiw!li ini owl diiw noitsirii di qu mu w iohidT .'bviixqo1i1 lii i iJ$di niJ3i d n& w ndw b1wonA intq iuo ni bniido noil ndil 3nod±iid uii i bus ( w%) 1uiu1l2 innoqmo owl o1dsmq di d iliw didw iiis1 di i ii bisl bJ111$ A .b3v!ovni i ii1) .titni bunilnoD iuolo itn f1i lo cn3Idolq odi i iuIuii bsx3 rli usd1 1sincnsbnu1 iom nv3

fioinix nisd3 isdi nwonA i II .wofl niiub noiriix n!13f13 lo thiio saomondq ngniibiid wofl iol 1dinoqi wofl -is1Iiqs d boouboiq

sw wns odT .noiiup iii snmon3dq ,di 101 inuo33s 01 !!sm 001 i bus hsM ib nsv lo inmnqx insiioqmi di d b3bivolq i 1si\çI3 uoidfl lo noiisrnlol di isril bvido iodius odT .OFiiooH

3i11ov 1o1sT b3!!s3 o n13iisq wofl 'çisbno3 ihidw is bq2 imii is lii x3ilov idi ni 1nmI3 !siin di bino31 dT .(.i i'X) boisqqs

3ft13vib bus onuflno31o 2nOr.31 315 3di : oquq in331q di iol m3iv 8O

4OITASIJJAThYSID 014A MOJTATfr1ELIRO VIEJAUD

borlnflsnu HA woil LsnoxisgnoI3 cf boouboiq as dscbA-daida ønodAasd iu1 .U siusi1 noflo3I3 rioiaaimannT .aiulswqmoi noflssn!siaçw 3rli is nidasw d bavomsi nood and Isnaism

.(ha11aA bus yaltsM) dqsionim

.s ax ororli anoior oaorb nI .aooiirov oil I lo noiionu[ orli is aonil rrisoria lo wofl Isool odT .rioiiooIib wofl 1530! out of !ollsisq ai rloirlw inoibsig ii3o!ov

isthrnsl orli ni boinoacto wofi isorla Isnan of boaoqqo as eIBnOiisno!oa ai oroil .noiioonb wofl oil of oaiovansii ai inoibsia ixoo1ov orb ororlw wofl çis!1iqso

-riisrlo on asH isuli wofl Isnoiisgno!o oil ai ii isrli rorlirth bosingoooi iodT ni nonomonoulq orb lo noflsnudqxo out rol boithpoi aoiiroqorq griibnoixo

inoooi mo to inioq ruiiisia on arniol rioiiirioooi insiioqrni airiT .noiiaoup psmmna A lCnoiisox!dnq msnimi!oiq 5 iii boorwonns tbsomls anoiisiiaovnx

11kw rioiisoilctuq ill orb ;botmoqoi oct won Iliw tow orb 1 otsia inoaomq orb lo .o1or!woa!o isoqqs

to 193119 sdl rebnu lIoitRsiIlkltçJo bna uoitngnole nM,Io no alnsmiioqx3 £.E wofl 1Bnoilkn0I9

wok\nnoimno\i\o sifln sM nO .VU. orb of lo!lsmsq ai inoibsig yiioolov orli Ii Isnoiisgno!o bomioi ai wofl oriT

.wofl gnxisio!ooob bns gniismolooos ors aolqmsxo iao!qrnia .wofl lo noiiooiib -moo !sixsinu of rolls! orli enoianoixo lsixsinu of abnoqaoiioo iorrno oriT

orb gno!s moiasa gniwofl orb lo inornolo omu!ov insvoloi orli lo noiaaoiq bovoiulos e'(11U1 ion md 1!rnttsq oct bluow aoaso oaor!T .noiioorib wofl ismhq

ooslmua bi!oa s no gnigniqmi bns norionianoo B rignoirli gniaasq biupi! orli c! orcill enoialnixo 03 anoiisuiia oaor!i to oorisvo!oi lsoinr!ooi ouT) .'!oviiooqaoi

ni eC1lw noasom oriT (.inobivo IIoa oct bluoula gnibluom noitoolni bns gninniqa orb ni aoil !snoiisgnolo '(llsiirsq oil '(mo b!uow wofl ouli !20253 ovods orb

.anoitibnoo wofl oaorb lo noiissi!som !snoiinovnoo on ru a!Isw bi!oa lo oonoaorq oarovansrt ii oil oa!s taurn oroub oros '(lilsaa030n al al!sw is '(iioo!ov orb aA

Qo


Figure 12. Diagrammatic representation of 'Taylor vortices' developing between the cylindersof a rotating couette apparatus (after Pennings ci aL30).

velocity gradient (as in the usual capillary flow) in addition to the longitudinalgradient. Transverse gradients correspond to simple shear which will makethe fluid element rotate in addition to extending it. This rotation will limit theextension which can be achieved as the direction corresponding to extensionwill alternately become that corresponding to compression. The purpose ofthe experiments to be described was to eliminate this rotation, i.e. thetransverse component of the velocity gradient causing it. In formal language,to produce a flow where the antisymmetric components of the velocitygradient tensor, which represents rotation, are eliminated as far as possible

First, however, the effect of elongational flow on a macromolecule will besummarized. In a stationary solution the macromolecule will be in its usualrandom configuration. When in the course of elongational flow the cor-responding element of liquid is stretched the molecule will stretch out with itto an extent to which the stretching force is transmitted from liquid tomolecules and to an extent permitted by the entropic resistance of the chainto being extended. Thus at the resulting chain extension the stretching force,

210

:1


defined by the frictional drag, will balance the entropic retracting force of thechain, the determining quantity for chain extension being f/k where f is thefrictional force constant and k the force constant for elastic retraction. As canbe shown to a first approximation f/k = r where t is the characteristic exten-sional relaxation time of the chain molecule in a given system. Both f and kdepend on the molecular weight (M). k x 1/M and the correspondingrelation for f varies between f x M and f x M depending on whether thesystem is free draining or not. In any event r, hence the molecular extension,will be an acceleratingly increasing function of M.

For a given system the chain extension itself will depend on the strain rateidefined by the velocity gradient in the flow system. There are several theoriesrelating chain extension to strain rate and to molecular parameters33. For ourpurposes they are all of similar form, namely that the strain is expressed as

r cx B/(1 — B)where B = t. It follows that chain extension will be infinite as B approachesunity. Thus catastrophic extension is expected close to a given value of tfor aconstant t, or for a given value of r for constant c. i is defined by the flowconditions. Accordingly, for a given solution a high degree of chain extensionshould set in rather abruptly as the acceleration of the flow reaches a certaincritical value. Or under given flow conditions this abrupt change shouldoccur as r approaches a critical value which in view oft '-. M2,with l < < 2,is achieved by increasing the molecular weight. Thus we see that the criticaldependence of chain extension on molecular weight, already indicated above,follows from basic considerations. In plain words for substantial chainextension we need long molecules: beyond a certain M the chains will bepractically extended below only slightly affected. Thus for a broad distributiononly chains exceeding a certain length will be extended for a given strain rate.

0

tog MFigure /3. Molecular extension for uniaxial extension plotted as a function of molecular weight

for three selected strain rates (Mackley and Keller32).

211

6

h

2

0

-2

-4

-8


Conditions for a dilute solution of polyethylene are given in Figure 13.The abrupt change in c with M is clearly apparent and also the abovementioned effect of the strain rate. To quote a few approximate figures;for a molecular weight of 105—the usual average for commercial material—strain rates of io—iO s' are required. This is to be contrasted with strainrates in usual spinning or extrusion operations which are in the range of

N2 N2

Figure 14. Diagram showing essential features of impinginget a1.'),

jet' apparatus (based on Frank

100_102 s'. For these rates only chains above lo M will extend, henceonly a negligible fraction of the commercial material. (These figures apply tosolution, in the melt t will be larger and conditions will not be quite asunfavourable.)

3.2.2. Experimental realizationFor a systematic investigation on the effect of elongational flow a well-

characterizable flow field had to be created capable of achieving high strainrates. The Taylor vortices as obtained by stirring, were too erratic andproduced a flow field both too complicated and inhomogeneous for obtainingthe phenomenon of interest in its clearest form. For this reason the simplejet principle was adopted. Nevertheless, for reasons laid out above, trans-verse gradients had to be eliminated as far as possible. This was achieved bycreating a symmetrical system with the avoidance of solid walls in the regionof interest. The system consisted of an accurately aligned double jet im-mersed in a solution of polyethylene in xylene as shown in Figure 14. Therequired flow could then be created in two ways: (a) by sucking solution intoboth jets simultaneously; or (b) by blowing solutions from the jet reservoir

212

1

Glass reservoirtubes

Polmaintained at

elevated tempera

Jets


into the main vessel simultaneously through both jets. Method (a) corres-ponds to uniaxial extension with pure elongational flow along the symmetryaxis, and (b) to uniaxial compression with elongational flow resulting alongthe symmetry plane (the perpendicular plane halfway between the jets).The system was fitted with a suitable microscope system equipped withpolarizers enabling direct observation of the regions between the jets.

3.2.3. Basic observationsPronounced effects were observed both for suck and blow case. These fall

in two categories according to the temperature of the experiment. (1) Atsufficiently high temperatures the effects observed persisted only during theduration of the flow; they ceased as the flow stopped. This will be referredto as reversible effect which corresponds to alignment of molecules withoutthe creation of persistent crystals. (2) Below a certain critical temperature(TI max) the features observed persisted after the cessation of flow. These werethe effects of permanent crystallization.

Figure 15 shows some representative examples of observations abovemax' i.e. for the reversible case (1 above) for both the suck (b, d) and blow

(a, c) jets. As seen there is a line of birefringence along the symmetry axis andplane respectively apparent also with unpolarized light. In the suck case thedirection of polarizability was along the symmetry axis, thus indicating chainalignment according to expectations. In the blow case the birefringenceproved to be confined to a ring along the symmetry plane with tangentialorientation of the mblecules within it. Obviously the latter is a more com-plicated situation. It is accounted for in the comprehensive work32. In whatfollows we shall confine ourselves to the simpler suck case for the purposesof the present article.

Figure 16 shows the effect of sucking below max (case 2 above). It is seenthat again birefringence develops during suction which here corresponds topermanent fibrils.

These observations provided the basis for further study of both chainorientation preceding crystallization (above 7, max) and of the formationand structure of crystals resulting from a pre-defined state of chain alignment(below T max)

For a quatitative interpretation the flow field has to be first characterizedso as to define the strain rates (ê) at the desired locality. This was achieved bymeans of tracer particles in a solution of identical viscosity photographed indark field. Figure 17 shows an example. The high degree of symmetry of theflow field is immediately apparent. It is instructive to note the single stationaryparticle at the exact centre which indicates that the velocity there is zeroand that from here the flow will have to accelerate towards both jets. Quanti-tative characterization of this acceleration, in fact the velocities over thewhole field, is provided by photographs taken with faster shutter speedswhen continuous lines traced by the particles shrink to finite tracks of lengthsproportional to their velocity. In this case the velocity gradient could bedetermined. For the suck it case was found to be practically constant alongthe central symmetry axis which greatly simplifies the ensuing analysis.In fact it could be directly defined from the jet exit velocities and jet separationin the knowledge that the velocity is zero in the centre.

213

cc () 0pjidnc AIGM O L11 LG1q (o1J1I?( ()pGLACq LOL (TC,

LOL Ch MCL rnq li' flCJ(Gq IIJIO ICV (C) 0pjidiic MGM °L pL!pi cq CO1JL OCLAG LOL(OL I JiG jO! CJG ((i) RILGLL1UCIJCG OGLAJ M!IP bojnoiq cLocq JI 10 IPC AWWC1LA

(LJUJ 4 ) () B!LCLL!UU 0pCLAGJ MIIp boJiLoiq CL02Gq 1I 10 LC aAWWCIL) xi2I bpo1oLbp i'" LOL ji bcL cu1 boAcp?cuc oInr!ou

(C) (q)

(a) (p)

'I KEITEE VL'ID I 1� V'LVCKFEA

I4OITAMJJATY5lD U4A MOITATMflMO 4IAHD

(di (Di

(b) (3) in q uIi B 1o noiu niiub bvido noit i1I313 o mmqoIvb dT (ii) .ö' wi1 ( I iufl A () ti boi etIoq ' OI x = J° 00! ii noUuIo qT xbigi$I

ioq rIi lo no bo!3oId ii! c1itIw noiu nhxb bniitdo tioqb iEi1D () .wofl 13fls biu !A3M) olrii bnix Ii niwod btwn3q3 diiw tud () A (b) .°

cii& MUT flX 1\) O11 iU b3ib11o3 3d b1uo noi1i11ii briE 110i21131x3 fllEr!3 rhod 2wo!101 i (1w nI

bIlE b3iiiiv s1l33lib 3d bluoo i ljii fnrnl2 3(1T .r brn 231d131nw 3d1 (11kw

ii lii iuo boinji w 2iflT .noiflni$q32 l bru noilu lo 31131 3rD d b3lIollnoo 3v12113(l3lqmo3 1 ii ni b3loIqx3 sw i' lo 1T13 3rlT .l3nnl3m 311Effl3122

32oqluq 2irll io'I .13m1oq 3111 lo lr1i3w 1s!u33Iom 3rD niiw d nrism briE x3bii51) bomqrno3 313W 2lrIgi3w mIu33Iom 1n313Thb lo 213ffl10q owl brni O1 x 10 21f1i3w w1u331om 1n3vE lr1iw riliw MUD ri3IEl2oH

1f1i3W 113liJ)3I01TI i23f1ifI 3111 10 133T13 orli noilibbs ni .(13vil33q1 O1 32113 1311131 3rD III .b3nimEx3 sw x3bti51 rn noiiudiilib bEold 3111 1 bn3

1EIiJ3310ffl l23(1if1 13V 3111 313(1W 3(10 1111W b5IIlqmo3 243W 31qffl112 1snino 3111

3111 (Ii 33ff313Thb rbu) noi1J3bE13b 311313d!13b d b3v0m31 aw bn3 R1i3w -1132 3(11 lo 211133ffi d borithb '13vi1i2n32 3d nso lisi lf1i3w 1131113310m r1if1

33ubolq 01 231(13310(11 123rIo1 3111 lo 1ilids 3(11 (10 r1uI31 3uprnrk)31 nib332 .( 2f1ofiU1o2 '(lEnoiisl2 11! II3V3 i3IDIJfl 11ia'13

313W 231d11111W 323(11 lo (10113(1(1113 213 b31u2133m 2311!insup 311T

-111113110 1131u331om 3rD fIi11il3b 24X13 1l3fflff12 3111 no1E 331I3(11it31id 3(11 (1) ;( Y 3v0d13) flE1 31di213v31 3(11 lo 3253 flu 11011

V .3.1 fl0iiJ3S1115i213 lo 31u11313q11131 fflumixEffl 3111 ()

31Q

IJJC pijiA °L fJJC 2OJ11IOU fO CLA2WJJJ5CMCJJU cuq °L fC qj2fLipnpou JGqf1C10U 111 fC W1J qCbLC22Cq 31 WL ic(ii) 31 " M2 2IIJJCUqA jCCfC pA MfLifi01J2 'II 1}JC JJ1JJ UJ0JCCflLU) B!LGLL!UCUCC (pcAoUq 31 UJ) M12 nUrijcCcq pA MLifi0U2 111 ft2 WI!

(ci) EIICCf OL 1JJOJGCfljL MCiJU (P!J UJ0JCCf1JL /CiJJf WI!)UJfCLWJ( i) OCCIIL2 f 1 J0NCL C0UCCUfLf IOU tO'- fJJG pJpcL IJJOJCCJT!81 !Cipf.CCpCq jOMCL 21LflU L12 (cr i) rnq f JIG C111 Oj MJfJJ C0UCGUfUIOUMcip LOL JIC CdrnAjCUf 21L1IIJ LTC2' J20 IJJC jUj COU2WU MfJflG 1(i) IJJC pJLctLiUCUc (rpoAC 31 WX) 1M12 JIfLCL LOL IJJC pJJJCL WOJCCFIJIL

(c) ELLCC OJ UJ0JCC1TJL McJpf (IACLcC)(ii) 31 W9X PM2 LLCPC1JJA 11UIIJCCfC pA CJJ91JC2 111 C0UCCU1UJ10U

CLJJC9J C0UCCUfLI0U f JIG MJJOJG GJjCC CCf2CqfOML? JOMCL C0UCC1JfLf 10112 ¶ nqqcu Cf1i-0j OCCnLLCq: pCjOM 1JJ12CCU1LPOU pcAouq J CCL Will C0UCCUtL1JOU HOMGACL MJICU oIuU) IIJC p1LCLJUGUCC (pcAouq 31 ) M12 qILGCqA boboornrj 0 Coil-

() EaCCI 0j C0UCC1JLp0UOfiK JU XAJCUC'

JJIJJC MiJnc LCICJIC i i c LOL pc bojAcpAjcuc H04&IGU(n) 31 '' iUCLcr2Cq /Jp 2L1U LC cUqJU fOML2 rn Awbof!C MJUGAWbOICi1JJA tO jJU1J AiJf1C t 0L poiit TO -U) BIG P!LCLLUGUCC (pCAouq 31 WX) IUCLC2Cq MIflJ 21Lfl1J LTtC tGUqJU

(ci) JJJC O 2fL9iU LtGIJIC 2r1WUJJLA O JJG LC2IIJt2 12 12 L0110M2:

cjqc? uq K1IcL)i UOM 1Jfq OL fp cc riiu n.&cci bnpcjca rnq r bJuL pi pcm

V ME! FEL! V1'1D I If k"!VCKFEA


(e) Effect of temperature(i) Birefringence (beyond TI; max) was essentially unaffected.(ii) The amount of crystalline material formed below TI; max was greatlyincreased by the lowering of the temperature.

Most of the above results are as expected by qualitative considerations.As a consequence of the present experiment we now have direct verificationin a quantitative manner. Even so there are certain features which have beenunexpected and are of special significance. These will be commented on inbrief.

The asymptotic limit of the birefringence with strain rate (a. i) implies thatsome physically significant limiting state of the system has been reached.The only feasible state of this kind is that of the fully aligned chain. Indeed,when allowing for the concentration dependence (by the proportionalrelation found under b. i) the observed limiting birefringence closely cor-responds to that expected from the fully extended chain from the polariz-ability figures of l3unn and Daubeny36 (in fact providing support for thesepolarizability figures as far as they are in dispute, see e.g. ref. 37).

The sudden cut-off value in the birefringence with the lowering of con-centration (b. i) implies that below a certain concentration chain alignmentdoes not decrease in proportion to solute material present but ceasesaltogether. This points to the fact that the birefringence, when observed, isdue to a cooperative effect. According to our interpretation this cooperativeeffect is provided by entanglements when a certain concentration is reached.Thus the cut-off value should give information on the onset of entanglements.

The finding that the high molecular weight tail affects TI; max implies thatthe longest molecules, even if present in very small proportions, play asignificant part in the onset of crystallization, but they do not affect theoverall chain alignment while in the solution. The latter is determined by therepresentative amount of material in the solution beyond a certain molecularweight.

3.2.5. Structure of the crystalsThe crystals formed (below TI; max) can now be removed from between the

jets and examined by electron-microscopy. Figure 11 shows an example.We see that they are the familiar shish-kebabs with a central thread clearlyäefined, containing transversely growing lamellae and a certain amount ofconnecting veil material. While this in itself is not quite novel our morpho-logical findings have the following important message to convey.

Firstly, crystals such as in Figure 11 grew at temperatures where the usuallamellar crystals were unable to form and grow, and thus are entirely stressinduced. They are indeed the central threads of shish-kebabs as in Figure 10.This proves that the primary product of crystallization is in itself a shish-kebab where the lamellae are interconnected with the stems.

The last conclusion is a welcome support to the model in Figure 9. What isnew, however, is that we have additional knowledge about the chain extensionpreceding the crystallization from birefringence measurements performedabove TI; max under the same flow conditions, an information which in viewof the independence of this birefringence of temperature (established undere. i) can be referred to the lowered temperature of the crystallization itself.

217

A, KELLER AND M. R. MAKCLEY

It was found that composite platelet—fibre structures are obtained even when,by the above evidence, chain extension before crystallization is complete.Figure 11 is one example.

Our explanation of this surprising fact is as follows. When a system ofcompletely aligned chains crystallizes, first a truly extended-chain typeentity forms presumably from the longest molecules. This will be the back-bone in Figure 11. Once, however, a solid interface has arisen a transversevelocity gradient will necessarily arise thus reducing the purely longitudinalcharacter, hence the chain-aligning tendency of the flow. Molecules whichbecome less extended will then deposit in an extended form only alongparts of their length on the existing fibre. The rest of the same partiallyattached chain will then produce folded crystals either at the crystallizationtemperature itself or during subsequent cooling which incidentally wouldalso account for the molecular connectedness of the platelets and fibres.Accordingly. the purely extended-chain crystallization together with theunderlying chain alignment in a flow field would be self limiting: thecrystals would reduce and terminate the same chain alignment whichbrought them into being. However, at the tip of the fibre a pure extensionalgradient will always be present during flow. Hence growth would terminatelaterally while it could continue longitudinally.

Whatever the true explanation of the phenomenon the important factremains: a composite, molecularly interconnected chain-folded cumchain-extended crystal structure is obtained even when the chains are fullyaligned by flow before crystallization. Or looking at it in another way,completely extended structures cannot be obtained by flow-induced align-ment alone. The proportion of the extended-chain fibrous componentshould be enhanced with an increase in the molecular weight, which indeed isdemonstrated by our own morphological work to be reported in the fullpaper32.

3.3. Application to the meltThe principles used in the preceding double-jet experiments in solution

were also applied to the melt. In the first stage an apparatus was designed inwhich the behaviour of flowing melts could be observed in situ along thelines of the solution experiments. This having been achieved, flow conditionsand geometries are being varied and the resulting products examined asregards structure and properties. Preliminary results have already beenpublished38 and further work is in progress. Here a brief report will be madeon the situation as it now stands.

An Instron rheometer was used for the experiment in conjuction with aspecially constructed chamber which in its first application is as shown byFigure 18. The piston of the rheometer forces the molten polymer simul-taneously through the two jet orifices. This is the flow system as in Figure 14with the flow induced by pressure on the melt outside, as opposed to thesuction applied through the capillaries themselves. The region of the meltsituated between the capillaries could again be directly viewed by an opticalset-up similar to the one used for the solution case, care having been takenin the construction that the glass cell windows should sustain the pressuresapplied. The pressures at any instance could be measured with the con-

218


ventional facilities of the rheometer. Under steady flow they were in theregion of 50 bars and when blockage occurred due to crystallization (see later)they rose to 500 bars. The full information contained by the observations isrecorded in motion pictures; here a representative selection of results will bequoted with a very few illustrations.

Polymsr

Block

Mutually opposed orificuc

NWindows

Figure 18. Diagram of melt flow apparatus (Mackley and Keller38).

At temperatures which are sufficiently above the conventional solidificationtemperature, effects could only be observed during the flow itself. Thismanifested itself as birefringence with cylindrical symmetry around the jetaxis Figure 19a). The contour lines in Figure 19a are loci of equal retardation.As the temperature is lowered, cusping of these contour lines occurs along thesymmetry axis indicating highly localized birefringence along the axis. Thiseffect was still reversible and confined to a temperature range of a very fewdegrees above 140 °C. The effect corresponds to birefringence exactly alongthe jet axis direction which is additional to that produced by the rest of theflowing melt. On further lowering of the temperature a distinct line-discontinuity appears along the jet axis at around 140 °C for the pressures

219

—--- Barrsl

=p--

330

iciq . = i:ic (,cJqGA uq KcJIcL)= itoc ) j jj c'() bOJ&L CLOQ lu boiioiY = (U

in . boion = )?Oc () RLp cjq i. = JOC (c) bOJiJL CLO2C in bo2iiou\ö bpooLbp &Jcu qnuu o LCiOU päCu obbocq OU(CG (c.)bO1L cLocq

(5)

(C) (q)

(a) (p)

'f KEFUhI� VD Y'1 ic /'1vcKI LA

*I

-

MOITAIJJATY5D OfrlA 1'IOITATI43IMO 1IAH i%) idi1 ixiii2ib i lo niqqi r!i ii1 niI ifIT ii1) b3ilqqE

-wofi lo trnboiq di Iuoivdo i ii .wofl lo noliso no iq bnn (Q1 &1 nmi13qx noiu1o rli rliiw oIsnJ3 di ur1T .noissi!!Ei1 b3ubni

wofI !inoirinoI oi ub 1woivdo i noiiJ3i11Ei2çI3 dT .bv3ifI ad noiiuIo molt bonthido onbiv r1i id idw ii1oI iux rh Ji 1uo

on i ivwor! oIEnE 2irIT ioIv lEnibiflignol Iuq s i idi T vodi iil ciojnii1iid Idiivi th irIi noiiu1o aT .iIqmo rkifIw oi riuiinqmi irIirI rIi V wodE D° OQ ii bvido w JE1 ru

ii ii 10 e1U33O 0ff ob idi iI3m rli ni 3IiuIW .bbnix aiw 1!2h how r1i ir1i ni fbidw bnuolgA3l3d di molt br1iunui2ib 11Euu d ionn ii aob rIJ Ii1ik d m noitEni1o ith) lloJi bfn3i1o ,mo3d E3

n1Bviup iffl 1no 311T .(n1oifli mI rli oi niwo 1dsi i noiiEblsi3l Jw ni1 ih3mo1rkoi rI lo niqw 1di1vi c1i i t33 noiiu1o 3fIi OJ

.iuisiqmi nOiiESiIIBi1 r!i vod

ni ui niiiuq d bubrii noiusillsa'13 auoidfl 1Enoiiibbi niwotk r1qioiod .o ii .(i11A bnB v1AEM) Y O41 = V qqu tl3rn iorn1oq

wofl innTlib l3bnu o1E bv1do EW fIOth3ffIldI idi1 in3nsmi&1 uo iflut o I3ncd rLt flu bM3!q 813W SJJI3 11W 13 3fl138fl! no aT .2i1Jmo

orb noowJod bolEoqq13 aGri b13oluIi ofIiI!13i2'(13 orb oioH .aoiiiwqmi ii riiAo1d 11i3uJrnov ouu&g di molt nUJ3rn3mo bo1sJ2 idiI thnoiiibbE

I

1wwI

O uJuJ) jp coLLc2bouqiu x-i.A qiLac1ou trntctu i b1LticnIcLJ?cowbo2ic pjrmcui (tAbicj qifwGfCL2 UI1 0 1-3 UJUJ CCU1LJJ C0L0LiiU11 IIPLC conjq pG 2GGJJ UJ!CLO2COb!CaJIA JU CL022-2GCtIOIJ2 O 2flCcouAcupouqA cLA2ticq UJGL1fF jpc qi2tiuci ecUtLrJ COLG qric fO pcJJJC J1UCL 20pqic2 i jOMCL tGWbGUUULC iuq tPfl2 GUACJob2 tPC L!P 1tppGCrn2G 'P!P !2 2fiJj UJO{GU Mill GXfLI1C 2!mfl1t1UG0fl2l M'tP !UbLc2cuf wcpoq o bLcbwrfiou POMGAGLe 11 12 uoi: opnucq 0U OMIJ

0(1 GXIJ1q1IJ 01ff °L fp CJJUJJGL fP' II0M!U if fO pG cojjcctq 1 0(11

fflLG2 (\i.c \j] 12 COuliunon2JA PC1U CLGffG pcccu ipc lCf2 niq 1ccb2lpc UP'2 !U 1k\1.E \t1 [OL mnjfibjc pptc MGU 0LIJJG 1f JOMGL fcmbcu-

iuififCq lu UffCL CLUIf IC UJIU1JGLfC IIPLG2 MGLG 2CGU fO LOLW OAGL fpC MOJC COUC 0j COUjJIICUf 21LGUJI1UC22 1(1 fG C32G O fG qofipIc lCf LCCUJCUf M!fP fP i'tf' cxbccfit!ou GLC21IJ1C 1Cf2 fC COUcfiOUTJ JM 12 UCIfpGL 2 btouonuccq UOL jocipcqi fO 1 JG22CL CXfCUf fjJU fJJG qonp 1Gf2 (2 fO pG GxbGCfGq JU fJJG CJ2C

EJ1LfPCL cxbcLiwcuf2 20MG flJf 2iU1G lGf 2 C1IJ fI20 bLoqncc UPLG2 CACIJo fpI2 uqiu LOL qj J1Uq2 01 mcp bLoCG22iu UCCq2 uo cIPoLf IOU

f !'- ¶f fC2G 1°Pf! PCLOLG (UAMpGLG CJ2G pc bofcuprJ !Wb0LfUCCMJJCLCAGL fGLC 12 JoJJitnqiuq urqJcuf IU T ll0M1J UJGJf 2opqicftioucpflu-CXfCIJqU c1cf o Gjoufiouc UOM rI2o Pnfpiu fpc uicjç JJJn2o jpc 2flG2fiOU uq bLoAiqJu coUAIUC1u bLool o pc couccbfiou 01I pLC LOLUJ2 pcpiuq fiG 02f(CjC 1J0U fpc CCUfLC jHJG2 bLoAw fG pq1ffG2f pA LcbjrCJ!J ouG o ipc 1cf2 pA oriuqq biu ' 2OM2 fp iuqccqpGpiUq fiG L2 f jiG 1j0CfA LJqiGuf UJfl2f pG CL0 JjJ12 2flG2f IOU MI2 bnf fO

cjq qo/Au2fLcuJ pcpiuq fJJG 2011q (L2 lu AICM o pc jci fPf01ff fO 112 pA C 0 CUU0IT fjJGLG porqq pG GAGIJ UJOLG CUCCf JAGGGCf GAGU '1 jG22GUG qnc fO fJJG 2opq Mjj2 OL fJJC pIL2 H0MGAGL 12 boiufGqlf!0U JpG COUU(ICIJI UOM 1 fiG LJ UJG2JJG2 M0f1j 0 COIIL2G bLoAJqc fP'W02f CGCf JAG fIG if2Gp conjq bLoqncG oLicuffJOu jGqiu to CLA2WI-

3O IP!2 2ji0M2 fjJf MpJIG fiG qoclpjc-]cf 2A2fGIJJ MT2 nuqonprcqjA ipc

JX1 OL fJC 1GL uq pijj YbGLiwG1Jf i. = io c bo1L cocq fl f; (jAfcJqGA iiiq KGIiGL,)V hPOOLbP 1J(GU qnLiu2 uo/ 2polliu ci iiiiiiou occnwu ou pc AmmcfiA

f KEFFEI� VL1D 1 K V1VCKFEA


interesting. It has two components: a very highly oriented one, as in Figure 2d,plus a more intense pattern such as in Figure 2e, the splitting of the 200 arcsdepending on solidification conditions. Thus we have made a full circle: weobtained our starting structures in Figure 2 but in this case in a purposefullydesigned elongational flow experiment. Obviously the central thread ofFigure 19d acted as a nucleus for the subsequently crystallized materialwhich was in a much lower state of orientation. Here the macroscopicfilament itself can be regarded as a single column of the kind sketched inFigure 2 where the central nucleating line is now clearly identified. It is ratherremarkable that the uniform orientation of the overgrowth phase should havepersisted from centre to periphery indicating the effectiveness of row nucleationfor the whole structure.

Many questions arise as regards the properties of the distinct fibres. Theseshould now be clearly assessable. Birefringence evidence indicates that weare within a factor of 3 of the fully aligned chain structure38. Highermechanical strength is indicated but definitive statement on this will dependon the accurate assessment of the relative proportions of core and over-growth material. Work is in progress.

In the above experiment the flow of the polymer melt was continuous.For jets of much greater length than used in Figures 19-21 however blockagesets in. This can be seen visually in our set-up and is also registered as asudden increase of pressure by the rheometer. It was observed with the aidof our optical arrangement that as the pressure rises the whole molten masssolidifies. This solidification is seen to occur by crystallization nucleated atthe initial threads. The effect is reversible: the solidified material melts againon release of pressure, the initial thread alone being left39. Clearly, here wehave a truly reversible pressure-induced crystallization produced by thepressure due to the blockage, which in itself is caused by flow-induced crystals,with the flow-induced crystals acting as nuclei in addition. We believe this isthe phenomenon underlying the blockage of capillaries by crystallizable meltsabove the practical solidification temperature, a phenomena observed inrecent years4042. This phenomenon has received much recent attention inview of the remarkable products it gives rise to41' 42 Thus we see that ourpresent in situ observations promise to unravel complex and potentiallyimportant solidification effects such as could hardly be accounted for bymacroscopic measurements of a flow or pressure parameter.

The article will be concluded on the last note. In brief, the requirement for arational understanding of solidification during flow is the understanding oflocalized flow phenomena and of the resulting chain elongation. Further, itmay be added that there is no substitute for direct sight of what is actuallytaking place.

We shall end up with a flash-back to the starting theme: are polymercrystals fibres or platelets, do the chains try to stay extended or do they wishto fold? We started with stating the answer, namely, that the truth is both.We hope the article itself has borne this out and in addition has demonstratedhow this basic truth can manifest itself under a truly astounding variety ofcircumstances. it is up to our technology to adapt this variety to its basicrequirements.

223


REFERENCESA. Keller, Reports on Progress in Physics, 32, Part 2. 623 (1972).

2 A.Keller. MTP InternazionalReviewofScience, Physical Chemistry. Series One, v.8 Macromol.Science. Ed. C. E. H. Bawn. Butterworth: London University Park Press, Baltimore, p. 105(1972).A. J. Pennings. Crystal Grol4th. Proc. Tnt. Conf. on Crystal Growth. Boston, Pergamon,Oxford. p. 389 (1966).A. Keller and M. J. Machin, Polymer Systems, Deformation and Fiow; Proc. 1966 AnnualConf. of Brit. Soc. of Rheology. eds. R. E. Whetton and R. H. Whorlow. Macmillan: London.Melbourne, Toronto. p. 97 (1968).A. Keller and M. J. Macbin, J. Macrornol. Sd. B, 1, 41(1967).

6 A. Keller. J. Polymer Sci. 15, 31(1955).A. Keller, Nature, 174, 926 (1954).D. R. Holmes and R. P. Palmer, J. Polymer Sci, 31. 345 (1958).R. S. Stein and J. T, Judge, J. App!. Phys. 32, 2357 (1961).

10 E. H. Andrews, Proc. Roy. Soc. A277. 562 (1964).E. H. Andrews, J. Polymer Sci. A-2. 4. 668 (1966).

2 Dlugosz. D. T. Grubb, A. Keller and M. B. Rhodes, J. Materials Sci. 7. 142 (1972).13 G. S. Yeh and S. L. Lambert. J. App!. Phys. 42, 4614 (1971).

J. J. Klement and P. H, Geil, J. Macromol. Sci. B6, 31(1972).' H. A. Davis. J. Polymer Sci. .4-2.4. 1009 (1966).16 E. S. Clark, SPE J., 23, 46 (1967).' M. J. Hill and A. Keller, J. Macromol. Sci. 83, 153 (1969).' M. J. Hill, A. Keller and M. Walton, IUPAC Symposium on Macromolecules, Leyden (1970),

Book of Abstracts, Vol. II, p. 849, preprint V7." H. Jenkins and A. Keller, to be published.20 M. Iguchi, H. Tonami and T. Kawai, Kolloid Z. u. L. Polymere. 221,28(1967).21 A. N. Gent, J. Polymer Sci. A-2, 4, 447 (1966).22 M. J. Hill and A. Keller, J. Macromol. Sci. B5, 776 (1971).23 C, A. Garber and E. S. Clark, J. Macromol. Sci. 84, 499 (1970).24 S. B. Clough, J. Macromol. Sci. B4, 199 (1970).25 D. Luch and G. S. Yeh. J. App!. Phys. 43. 4326 (1972).26 E. H. Andrews. Lecture given at the Conference of the Royal Microscopical Society, Oxford

1968 (unpublished).27 A. J. Pennings, J. M. A. A. van der Mark and A. M. Kiel, Kolloid Z. u. Z. Polymere, 37. 336

(1970).28 F. M. Willmouth. A. Keller, I. M. Ward and T. Williams, J. Polymer Sci. .4-2. 6, 1927 (1968).29 F. M. Willmouth and A. Keller, J. Macromol. Sci. 86, 493 (1972),° A. J. Pennings, J. M. A. A. van der Mark and H. C. Booij. Kolloid Z. u. Z. Polymere. 236, 99

(1970).F. C. Frank, A. Keller and M. R. Mackley, Polymer. 12. 7,467(1971).

32 M. R. Mackley and A. Keller, Phil. Trans. in press.A. Peterlin, J. Polymer Sci. 84, 287 (1966).D. J. Blundell, A. Keller and A. J. Kovacs ,J. Polymer Sci. 84,481(1966).D. J. Blundell and A. Keller, J. Macromol. Sci. 82, 301 (1968).

36 C. W. Bunn and R. de P. Daubeny, Trans. Faraday Soc. 50, 1173 (l9S4).37 M. V. Vokenstein, Configurational Statistics of Polymeric Chains, lnterscience: New York

(1963).38 M. R. Mackley and A. Keller, Polymer, 14, 16 (1973).

M. R. Mackley and A. Keller, to be published.40 A. K. Van der Vegt and P. P. A. Smit, Advances in Polymer Sd. Monograph 26. Soc. Chem.

Ind., London, p. 313 (1967).J. H. Southern and R. S. Porter, J. Macromol. Sci. 84, 541 (1970).

42 J. H. Southern, N. E. Weeks, R. S. Porter and R. G. Crystal. Makromol. Chem. 162 19(1972).

224

chain orientation and crystallization · the prerequisite for crystallization is the retention of...

Documents