Polymeric Fluids

GM05 Part 1
Dr. Helen J. Wilson
9.1.2006 19.1.2006


  

!#"$%&'%(*)+,-!.!/*'!"0,$1*&23#546%7'(*7-(*)89%)68;:*!-<8;"%2="#,$->,-!.!/*'!"@?A%
,(B,!"#C#DEF:*6G$#H89:*JIB$"#7-"F2K$%'ML*'$%"N8;7'.J,$<8;!7-$-"F$%(*1OL*P1*/*.68;"#D
QR/*7-(*)S89:*T$%(U/B23$%.689/*#7'(B)VL*#P.!!"#"#WX'7-YP/*7'1*"K.!(;8;$7-(*7'(*)ZL[%\,!"K$#"#/*I]46!.689!1V89Z^*`_aDTE:*
_R$`,89:*!"#b-7'YP/*7-1*"c!$%.68d7-"c1*689!#,7'(*!1eI;,89:*b"#:*$%L[!"#Wf%T.!%(*g*)/*#$<8;7'%(*"h8;:*$<8c89:*b,-!.!/*'!"
$%1*L89DjiH!.!$%/*"#%2M8;:*!7'J"#7'k!9W
89:*L[%\,!R,-!.!/*'!">.!$(I[!.!%"l89#689.m:*!1I;89:*^*`_aW
)%7\n7-(*)#7'"`H89,"#,J"#/*#L*7-"#7'(B)I[!:*$onf7-/*#D
pq_>7-'r"N8;$N8ZI;O'PsP7-(*)>7-(*2t#$%'O$<8b$>2t6_u"`$,L*'!"%2P89:*!"#V^*/*7-1*"$%(*1v:*`_+8;:*6OI[!:*$`nr!D
EF:*J"N8;/*.68;/*J%2r89:*J,#J2t#$%
L*$%N8>%2B8;:*J.!%/*"#J7-"#w
xTyPz%{9|h}~yP0KKRP1*!-'7-(*)8;L`1*o_>(*%wfLf!7-!(;8;"e/*"#!1&8;.m:*$%$%.689!#7'"`$@^*/*7'1*W>$(*1
,U1*!-"F/*"#!18;g88;:*!MI[!:*$`n7-/*F2r8;:*!"#JP1*!-"F7'("#:*!$#7'(*),^*`_a"#D
xTyPz%{9|h}~yP,%KKRP1*!-'7-(*)jI[8#89%d`/*L*%w/*(B1*!"N8;$(*1B7'(*)V:*`_7'(*1B7\n7'1*/*$%%'!.!/*-!"If!:*$`nB
$%(*1O!$%.688;^*o_>M68;:*P1*"F2tF1*!7n7'(*),$%.!%"#.!LB7'.J$`nr!$%)>I[!:*$`n7-/*#D

v J

EF:*J"#7-L*-!"N8a"N89!$%1G^*`__a:*7-.o:O7-"F6L[!#7',!(;89$%'#!$%'7'"`$I*-J7'"V$=9KyUB}vK]vw

EF:*!#J$#H8_dR8L[!"F%2K7'(*"N8;/B!(;8>/*"`!1O23%F"#:*!$#7'(*),^*`_a"#w
r
{!}~rANyU9{!}~*N#:*!,689!#"d7-(e_>:*7-.o:e89:*"N8;$%7'(?A"#:*!$%Cd$%(*1"l89#$7-(#$<8;?N"#:*!$d#$<8;!C
$%JL*#!"#.!7-I[!1*WM$(*18;:*J1*6n7'.!J,!$%"#/*!"89:*J#!YU/B7'#!1O23%.!!
r
{!}~rANyU9{!}~yU9@#:*!,689!#"#W]7-(Z_>:*7-.o:=$d.!(;8;%'-!12t%.!c7-"$L*L*-7'!1=$%(*1Z89:*c!.m:*$%(*7'.!$%
#!"#L[%(*"#J!$%"#/*#!1*D

EF:*HnX$%7-$I*-!"_a:*7'.m:O.!$%(OI[J7',L[%"#!1O$%(*1*%,!$"`/*!1O$%!w
8;:*J"N89#!"#"V2t%.!J$%L*L*'7-!1OL[!F/*(*7\8a$%!$%W
$%(*1
8;:*J"#:*!$F#$<8; "#Lf!!1O2r89:BH89%LOL*'$89J1*7n7'1*!1OI;e89:*J)%$LI[68_d!9(89:*JL*-$<8;!"#D
!. %,(g*"N8"N8;!L7'($(*$%\"#7-(*)G$(*6_$89!#7-$7'"R89$"#"`/*=8;:*$88;:*!O7'"$/*(*7-YU/*
#!'$897-(B"#:*7'LIf68_d!!(,89:*!"#!W#!)$%1*-!"#"V%2^*`_:*7'"N8;NDp+_a7'-.!,>I*$%.os8;R8;:*nX$%'7-1*7\8
%2r89:*7-"F$"`"#/*L8;7'%(O-$<8;!D
EF:*J!-$<8;7'%(*"#:*7'LOI[68_d!!(O"N89#!"#"F$(*1O"N8;$%7'(&?A%"N8;$%7'(O:*7'"l89%AC7-".!$%'-!189:*dz%[h9{9A{!|K{!Nry
}~yPA*{9ArD
EF:*R"#7-L*-!"N8>7'"8;:*R'$o_2t$=yU{![hN*c
A9zr|Ka^*/*7-1*W7-(=_a:*7'.m:=89:BR"N89#!"#"7-""#7-L*
L*#%L[N897-(*$%]89R89:*J"#:*!$%#$89!w
Z
2P_dZL*'8b"N89#!"#"T$)%$7-(*"N8b"#:*!$%$89!W[%2.!/*#"#8;:*7-"7-"M46/*"N8b$a"N8;$%7')%:;8b'7-(*V%2"#-L[=8;:*%/*):
8;:*J#7-)7-(*D
2M8;:*"N8;!"#">7'(*.!#!$"#!">%"#-`_a\8;:*$(-7'(*!$%W
89:B^*/*7'17-"J.!$-'!19KyP*}~{!KAKhNhDj2
787-(*.!!$%"#!"d2t$"N8;!8;:*$(-7'(*!$%W[8;:*=^*/*7'17'"d.!$%'-!109hyP*}~{!KAzryPKAKfDE:B!"#$%#"`:*`_>(
"`.o:*!,$<8;7-.!$-\GI[!-`_aw_dJ$%>L*'88;7'(*),"#:*!$%"N8;!"`"F$)%$7-(*"N8a"#:*!$F$89!D

EF:*V.!%(*^*%/*
_RVL*$%"#"#!1v#/*(*1>_d$"T$(v6$%L*-V%2f$a"#:B!$Nl89:*7-.ms!(*7-(*)>^*/*7-1*wc$"B%/v"N8;7'#!1
7-"2t$"N8;!W
7\8>!"#7-"N8;!1j=/*.m:j:*$%1B!DEKP<8;:*L*$"l89R$(*1!=/*'"#7-(L*$%7'(;8J$%c8_R"#:*!$%Al8;:*7'(*(B7'(*)
^*/B7'1*"#wc23%8;U89:*L*$%"N8;!WB-`_"l89#!"#"#!"S'7-sF)%$`n781*(*8V$%sF78V#/*(%a89:*FI*#/*"#:*WBI*/8V:*7')%:*!
"l89#!"#"#!"H'7-sdI*#/*"#:*7'(*)=1*$%sd7\8R^*o_>D*7-7-'$%WfL*$%7'(;8R^*o_>"H:*$L*LB7'0$"Kr/jL*/8R78R(89:*
_R$-'WMI*/88;:*!(O1*P!"#(*8>^*`_/*.m:O/*(*1*!F)#$`n7\8G(*.!Hr/O-!$`nrJ7\8F89:*!#!D
T

qhdhBXhuXRRd

;( )#$L*:%2
"#:*!$%"N8;!"#"c$)%$7-(*"N8b"`:*!$T$89Z.!$%(I[Z#!L*9"#!(;89!1$%"T$a"`7'(*)%'Z2t/*(*.68;7'%(a?~ C#D
#I*$%I*\089:*g*#"N8;W$%(*1G.!!A8;$7-(*\08;:*"#7',L*'!"l89WP1*!2(*(oZ6_F89%(*7'$%(GI[!:*$`n7-/*R7'"c8;

,!$"`/*89:*7-"3nf7-"#.!%"#7\82t/*(*.6897-(*$%(*1>8;:*!(/*"#b7\8d$%"8;:*Sn7-"#.!"#787-(>8;:*Z/*"`/*$[b6_8;(*7-$(
^*/B7'1O!YP/*$<8;7'%(*"#D
b<8;2tc8;:*%"#d_>:*O:*$`nB!789:*!J^*/*7'1B"R%c8;!(*"#%"J6L[!#7'!(*.9!w7'($O^*`__a:*7'.m:7-"J(*<8
L*/*#"#:*!$%WK_de(*!!18;01*!g*(*e$0"#:B!$$89!D&*7-(*.!!WH7-($ Q"#:*!$^*`_aWK89:*e$89%2
"N8;$%7'(8;!(*"#%

? C`

_dJ1*9g*(*H89:*J"#:*!$%#$89J7-(O)!(*!#$]89,I[G
wr D
QR7'f!!(;8>^*/B7'1O,U1*!-"#w
b6_8;(B7'$%( ?~ CS

`_d!N`'$`_

?~ CSV B

$#!$%/

?~ CSr

iH7-(*):*$%
?NL*'$%"N897-.!C

2S
8;:*!(
rVO  #W 
rV a
#Wa  

?%TrbC9

b89 7-"Fb6_8;(*7-$(*W
& "#:B!$ 89:*7-.os%!(*7-(*)$%(*1
a "#:B!$Nl89:*7-(*(*7'(*)%D
?~ C93 B
r

%) $7-(*W] 7-"Fb6_8;(*7-$(*D
r% WM$ "G
W* Bb
$%(*1O$%"j
W* %9D
-"#sP(*`_a(O$%"89:*H7-!'1O^*/*7-1*D

p:*$<8a1*d8;:*!"#J"N8;!"`"#!"F'P%se'7-s7'#"N8;W
L*-<8#897-(*)"l89#!"#"F$%)$%7'(*"N8a"#:*!$%#$<8;!w

 


 


 
 


 
 




 

 
 


$%(*18;:*!(*W
L*-<8#897-(*)d8;:*J![!.68;7\nrHn7'"#.!%"#7\8j$)%$7-(*"N8a"#:*!$F$89!w


T

0T6]hd

7-"#!/*7-'-V^*`_ 7'"T^*`_q89:*#/B):v$a.!7'#.!/*-$L*7'LfV$F8_R<`1*7-!(*"#7-(B$r"#'789D
b!#_dS_>7-'r'P%s
$88;:* Qnr!"#7-(*D
b`_ 1BU!"V$)%!(*!$%'7-"#!1Ob6_8;(*7-$(O^B/*7'1OI[!:B$`nBJ7-(O$^*o_ -7's%H89:*7-" 

h689-"R"N8;$N8R_>789:j89:*23/*-)!(*!#$-7'"#!1b6_8;(*7-$(^*/*7'1*W$%(*1Gs%!!LG78,$"J)!(*!#$$%"RL[%"#"#7-I*'

2t%$R_>:*7-'!D
EF:*!YP/*$<8;7'%(*"b)%`nr!(*7-(*)>8;:*,<8;7'%(?_d!'-K$"#"#/BV8;:*^*/*7'17'"b7-(*.!%L*#!"#"#7-I*'!Cb$#,$"#"
.!%(*"#!NnX$<8;7-(*w






 
 


8;:*#!J"N89#!"#"F1*!g*(*7897-(*"`w
u

?~ C 
  

+?~ C

$%(*18_R,,!(;89/* !YU/*$8;7'%(*"#w



 






 


 





 

  

?~ C   





 

 


 














EF:*^*/*7-11*!(*"#78j7'"  W,7'"
89:*FL*!"#"`/*!WB$%(*1
$%(*1
$#89:*.9,L[(*!(;8;"%2%8;:*nr!'P.!78D






pqh_a7'-X$%'"#d:*$onBHI[%/*(*1*$N.!%(*1*7\8;7'%(*"f89:*$8I[89:.!,L[%(*!(98;"28;:*H^*/*7'1bnB!-U.97\8e$#ck!!#
$8>$,"#%'7-1_d$%'-w


$8


  


?p+J:B$`nBJL*/8 %(89:BR.9!(98;!-7'(BJ2r89:*J.m:*$%(*(*!'D-C
EF:*JYP/*7'.ms%!_d$oe89R_>#7\8;J1*`_a(89:*!"`J!YU/*$897-(*"V7'2rB%/*-J:*$%L*L;_a7\8;:8;!(*"#F(*<8;$<8;7'%(O7-"

"! 
!
!
$#  ?~ C
? C



pqF$%V'PsP7-(*)>$<8Z$>^*`_+_>:*7-.m:,7'"$-B7'(J89:* `1*7-!.68;7'%(*Wr$(*1v(*=nX$#7-!"S7-(R89:B ` 1*7-!.68;7'%(*w


%v? C
&

1'%)(1 




*

b8989:*$8c7'(V89:*7-"K^*`_>W~8;:*,+.-/.012+43516/ +.+[_dT:*$`nrIf!!(F8;$-sU7-(*)Z$If/8c$%(*1a.!$%'-7'(*)bv7'"K$.68;/*$%'
D
EF:*7'"V^*`_ $%/8;,$<8;7'.!$%'G"#$<8;7'"`g*!"F$%"#"F.!%(*"#!An]$<8;7'%(*W
$%(*18;:*J"N89#!"#"#!"F$%#!w


u
+?~ C 
  ?~ CX
_a:*7'.m:O7-"8;:*J"#:*!$F"N8;!"#"V$"_RJ6fLf!.689!1*DFE:*8_d,%!(;8;/* !YP/*$897-(*"FIf!.!,!w













$%(*1

hU%sP7'(*)R$<8V W_dH.!$%("#!K8;:*$878F (*e1*!L[!(*1*"%( w$%(*1"#,?N23#  8;:*H

 "#!.!%(*1,,!(98;/*

!YP/* $897-(*C2
 .!$(%(*\&1*!Lf!(*1( D7V6823#8;:*g*"N8!YP/*$<8;7'%(*W $
 (
.!$(%(*\&1*!L[!(*1

%(

wF"#,7\8a/*"N8aI[J$,.!(*"N8;$(;8;D8a2t-'o_>"8;:*$<8

v8
%T:
9
_a:*!#)7-"Z$($#I*7\8;$%AI*$%"#6`-7'(*aLB!"#"#/B!Wh$%(*;
1 97-"S8;:*aL*#!"#"#/*#a)#$1*7-!(;8Z_>:*7-.o:1*7nB!"
8;:*J^*`_aDb`_ 8;:*J"#!.!(*1O,,!(;89/* !YP/*$<8;7'%(O7'"V"#$<8;7'"#gB!1O$(*18;:*Jg*"l8>I[!.!%,!"




1 %
'

 ??~ CX CS

9
9
1  
EF:*7'"7'"a$%"a2t$>$%"F_dv.!$%(&),_>789:*%/8"`L[!.!7'2 7'(B)_a :*7'.m:&.!(*"l89789/8;7\nrv'$`__dv$%v/*"#7-(*)DG8
7-"$01*7'f!!(;8;7'$%!YP/*$<8;7'%(&2t% $(*1:*!(*.!2t<%v? C#Wh_>78;:&If/*(*1*$%A+.!(*1*7897-(*=
" %v>? Ca
%v?l
CS D
0yU{![hN*h|KA
?b"#7-(*)$,b6_8;(*7-$(^*/*7'1*W*7'"F.!%(*"N8;$(;8>$%(*1O%/*!YP/*$897-(O7-"


1%

 u9  

1
EF:*J"#%'/8;7'%(89R89:B7'"#WM"#$<8;7'"`2 7-(*)@%v?>CS&%v?lA0CS WM7'"


NAKrKPO

C

?>

CB

EF:*7'"V7-"F$DCE0160
F.G
H IKJnB!-U.!78L*%g*'!]8;:*J,$<7-/* "#Lf!!1O7-"F$8F89:*>.!!(;89#!'7-(*

%&9LvM( fD
%v?

W]_a:*!

Q
NyUAK|hN

b!#89:*O"#789/*$897-(7-",$'788;'O,#O.!,L*'6DSRZ!.!$-8;:*$<8,89:*OiH7'(B):*$%5^*/*7-1:*$"R8;:*!
1*7-[!#!(;8OL[%"#"#7-I*'$%.6897-(*"`w2S
R89:*!( R23%,
`WZ  = d$(B123%

a WMua

= DpqJ.!$(_a789J$%(O6L*!"`"#7'%(O2tj 7'(89!#"F%2?A% F C#w

?N
`CB(<

 

?N
`CB(< a

E :*J"#:*!$%#$<8;J7'"VL["`7\8;7\nrJ2tV $ (*1O(*!)%$<8;7\nrJ2tV a D
F
pqJ"N8;$N8>I;G"#n7'(*),2t89:*J"l89#!"#"#w




 9
U
V
&W@:9

2t%"#%>/*(*sU(*o_>(O.!%(*"N8;$(;8WZD
Zo__a7'-X_RJg*(*1XWY
EF:*!#J7'"V$,"N,68;AG7-(8;:*JLB%I*'!,wF7-2r_d>"N_d$%L8;:*H8;L $(*1I[88; _d$%'-"#W]_d_d%/*'1*(*8
6L[!.88;:*Z^*`_89.o:*$%(*)!DE:*7-"T,!$(*"8;:*$<8d7'2X_db!L*-$.! _>789:e 6nB!NU_a:*!#!W89:BZ(*
.m:*$%(*)!"Z_Ra_d%/*'1@6L[!.68=89"#!a_d%/*'1I[e.m:*$(*)%!"=2F"#7-)(*D (;+L*:;"#7-.!$cYU/*$%(;8978q.!$(
!789:B!bI[=/*(*.m:*$%(*)!1%b6$%.6896nr!"#!1[bI*/8(*8$e,7\P89/*#=%28;:*Z8_RZ
D hU%sP7'(*)j$<8d89:*
2t% 2 W8;:*[
 9 8;! _>7-'I[6nr!"#!1I*/8>8;:*[
 Wu899 7'"a/B(*.o:*$%(*)%!1*w"#;
 W /*"l8I[
k!!#D


pqJ$#J(*`_ _d%sP7'(B)d_a7\8;: 
9 `nr!8;:*J#$(*)%>
]\ \D
^2 9_  #W]8;:*!(
89:B%/*):*%/88;:*J^*`_aW
"# $%(*189:*!#J7'"V(*^*`_aD
^2 9_  #W]8;:*!H_a7'-
IfH89:*#!J!)%7'%(*"8;.!%(*"#7'1B!w



_>:*!# 
  a
.(`9
$(B1 ?N
9 q #CB(<
 _>:*!# a .(`9 
.(`9 $(B1

a
_>:*!#
B(9 

$(B1 ?N h:9 BC (<


pqJ.!$((*`_ 7-(;89!)%$89J/*L23# u
W]_>:*!#A% w2tV a
.(`9_RJ:*$`nr

9 


9 
1 %
'


u

%
1

a
$%(*1O/*"#7-(*)d8;:*J.!(*1B7\8;7'%(89:*$
8 %v?l
CS W

9


%
?
C
c?  C


b`_I[68_d!!(a
.(9  B(9Wd "#d8;:*Hnr!'P.!7\8G7'"F.!%(*"N8;$(;89w
9


%

9


$%(*1Og*(*$%'W
I[68_d!!( B(9 0WX_dJ:*$onB

1 %
'
9 


9 


u
%
1

b
a

$%(*1O,$<8;.o:*7-(*)dnr!'P.!7897-!"F$8 (9 g!"#w

9


c?  C
? b
0 C 
%

*/B,$#7'"#7-(*)d8;:*HnB9'P.!7\8LB%g*'!w



9>? vT 9C 
~?
C 
  a
.(`9

%
9>? u?l
.(`9CC 
a B(9   B(9

9?> C

~? :C B(9 

_a:*7'.m:O-PsP"F-7's%H89:*7-"#w
dfeMgKh`gjiMeMkliMmnoiqpVrMgKs>rMgjtVklscuveMgjwMiMkls>kxnoiqeMgKh`gjiMeMgjeynoiqscr4g^zM{}|o~klu


E :*.!!(;8;$%]#!)%7'%(7'"S/*(;7'!-1*!1$(B1^*`_a"7-($J"#-7'1ZKA|Kfw8;:*7'"S^*`_L*%g*'F7'"S.!$-'!1,$yKA|K
F
K]vD
EF:*7'!-1v^*/*7'1v7'"$a)PP1vP1*!r23%S^*/*7-1*"'7-s8;P<8;:*L*$"N8;!
D Ps$<8B%/*
8;P<8;:*L*$"N8;V$"Mr/
"`YU/*!!k!>78a%/8a%2r8;:*89/BI[!wV78a.!%!"V%/8a,#>V'!"#"V$%"V$,"`-7'1.6'7-(*1*!#WMI*/87'(2t$.687\87'"
-/*I*#7'.!$89!1OI;G$d8;:*7'(O-$`r!(*!$%8;:*H_d$%'-"#D
Uy }~zA9yUw $-.!/*'$89S8;:*Z%7'"#!/*7-'-Z^*`_ L*%g*'Z2tT$aL[o_d!-$`_ ^*/*7'1*DKp:*$<8d1*F8;:*V^*o_
L*#%g*'!"F-U%se'7-sJ2t%1*7-[!#!(;8nX$%'/*!"V2[

Tf

DMR]6]XFhBKhFPSvB]dX

(!$%K6L[!#7',!(;89"c_>789:Lff,!d^*/*7'1*"#W7\87-"d$-%"N8$_d$`"d%I*"#!NnB!18;:*$<8d8;:*=!"#Lf(*"#
8;Z89:*d"#:*!$%c$89d7'"H(*8R!L*#U1B/*.!7'I*-!wH7'2*B%/e:*%'1j(*d"#:*!$%c$89d2tc$=-(*)Z8;7',!WUr/j$o
$%\_R$`"#!$.m:=8;:*R"#$%J"N89#!"#"#W
I*/8>%2 8;!(=8;:*7'"F"N8;!"#"F7'"F(*<8>I8;$%7'(*!1YP/*7-.msUD88;$s%!""#%
8;7',v2t%8;:*v^*/B7'1O8;023/B'\+$146/*"N8a8;8;:*v(B6_ ^*o_ :*$L*L[!(B7'(*),890789we7-(&89:*!F_R#1*"#Wh89:*
^*/B7'1O:*$%"F=
$ O&y O&rM} Q&2K789"VL*6n7'%/*"F^*`_:*7-"N89%AD
EF:*,"N86P89#!Vnr!#"#7'%(28;:*7'"d7'"b$%(!-$"N8;7'.=b/*I*If!Z"#%'7-1*Wr_a:*7'.m:$%\_R$`"b#!!If!"
8;:*,"`:*$Lf,78>_d$"F8;"N89$%A8>_a7\8;:*Dvp :*!(O89:*v"N8;!"#"a$%L*L*'7-!18;7\87-"a!-!$"#!1*WT78>_>7-'}~yPA
I*$%.msa89,7\8;"F#7')%7'(*$%M"#:*$Lf!D
(89!#"F%2K"#:*!$%^*`_a"#WM$-7'(*!$%G!'$%"N897-.J"#%'7-1O.!$(OIfJP1*!-'!1O/*"#7-(*)


&F ? CF

>? ;CB'
1 M

38R7'"H(*8R1*7'#!.6890#!'$8;!1a8;Z8;:*b"#:*!$%c#$<8;!wH78R1*P!"#(*\8J.!$#b:*o_2t$"N8R789-"Honf7-(*)W%(*\:*`_
2t$%78>:*$%"F,`nB!1*D
"N89$%(*1*$%1L[!#7'.>^*/*7'1_d%(*\8a#!!=I[!F789"V"#:*$%L[J2t%6nr!I*/8aI[!.!$%/*"#J789"V,-!.!/*'!"
.!$%(eI[b"N8;68;.m:*!1I;0$^*`_aWf7\8c_>7-'f!,!I[9_>:*$8d:*$%"c:*$LBL[!(*!1a89=7\8d/B(98;7'U89:*b,-!.!/*'!"
:*$onB:*$%1j897-d8;#!'$fD8,7-"J$.!#"#">I[68_d!9(G$n7'"#.!%/*"J'7-YP/*7'1+?_>:*7-.m:23%)%689">7'(*"N8;$(;8;\C
$%(*1$(!'$%"N897-."#-7'1? _a:*7'.m:(*6nB!2t%)%689"#C#wT$b
A9z
yUN*9{!NzAc
|hNDcEKR,P1*!]$cn7'"#.!P!-$"N8;7'.
-7'YP/*7-1*W
_d,!$%'(*!!1"#%,689:*7-(*)89:*$8,6:*7-I*7\8;"J$O1*!.!$`7-(*
) O&y O&rM} QDO2_dR4/*"N8v$%'-`_$
-7'(*!$%R1*!Lf!(*1*!(*.!(j89:B^*o_ :*7'"l89%AW
_d)%68R8;:*NAKyU*}
A9z
yUN*9{!NzK|KAW_a:*7-.o:j_d
_a7'-M"#!J(*6P88;7',!D

Linear viscoelasticity

A linear viscoelastic fluid is a fluid which has a linear relationship between its strain
history and its current value of stress:
(t) =

G(t t0 )(t
0 ) dt0

The function G(t) is the relaxation modulus of the fluid. Because a fluid can never
remember times in the future, G(t) = 0 if t < 0.
Physically, you would also expect that more recent strains would be more important than
those from longer ago, so in t > 0, G(t) should be a decreasing function. There arent
really any other constraints on G.
A few often-used forms for G(t) where t 0 are:
Single exponential
Multi-mode exponential
Viscous fluid
Linearly elastic solid

G0 exp [t/ ]
G1 exp [t/1 ] + G2 exp [t/2 ] +
(t)
G0

Lets just check the last two. For the viscous form we have
(t) =

t
0

G(t t )(t
) dt =

(t t0 )(t
0 ) dt0 = (t)

as we would expect for a Newtonian viscous fluid. For the elastic solid we have
(t) =

t
0

G(t t )(t
) dt =

t
0

G0 (t
) dt = G0

(t
0 ) dt0 .

The integral on the right is the total strain (or shear) the material has undergone: so
this, too, gives the form we expect.

3.1

Creep

Given a sample of material, how would you go about modelling it? Even if you start by
assuming it is a linear material (and they all are for small enough strains), how would
you calculate G(t)?
One way would be to carry out a step strain experiment: for shear flow this means
you would set up your material between two plates and leave it to settle, so it loses the
memory of the flow that put it there. Once it has had time to relax, you shear it through
one shear unit (i.e. until the top plate has moved a distance the same as the distance
between the plates). Then stop dead. Measure the force required to maintain the plates
in position as time passes. The results will look something like this:
11

6Shear rate

Time
-

6Stress

Time
-

What is the real relationship between the stress function (t) and the relaxation modulus
G(t) in this case?
Suppose the length of time the material is sheared for is T . Then the shear rate during
that time must be 1/T (to get a total shear of 1):

t0 < T
0
0
1/T T t0 0
(t
)=

0
t0 > 0

Then the stress function becomes

Z
Z t
1 0
0
0
0
(t) =
G(t t0 ) dt0 .
G(t t )(t
) dt =
T
T

Now suppose that we do the shear very fast so that T is very small. Then
Z 0
G(t t0 ) dt0 T G(t)
T

and so the stress is

(t) G(t).
Thus the relaxation modulus is actually the response of the system to an instantaneous
unit shear.

3.2

Storage and Loss Moduli

An step shear is very difficult to achieve in practice. Real rheologists, working in industry,
are far more likely to carry out an oscillatory shear experiment.
The material sample is placed in a Couette device, which is essentially a pair of concentric cylinders, one of which is fixed. The material is put in the narrow gap between
the cylinders, and the free cylinder is rotated. The flow set up between the two cylinders
12

is almost simple shear flow (as long as the gap is narrow relative to the radius of the
cylinders, and no instabilities occur).
The rotation is set up to be a simple harmonic motion, with shear displacement:
(t) = sin (t)
where is the shear, the frequency and the amplitude. In reality must be kept
small to ensure the material is kept within its linear regime; but if we are dealing with
an ideal linear viscoelastic material, can be any size. Note that the shear rate of the
fluid, (t),

will be
(t)

= cos (t).
Suppose this motion started a long time ago (t ). Then at time t, the stress profile
for a general linear viscoelastic material becomes
(t) =

t
0

G(t t )(t
) dt =

G(t t0 ) cos (t0 ) dt0

We can transform this integral by changing variables using s = t t0 :

(t) =

t
0

G(t t ) cos (t ) dt =

G(s) cos ((t s)) ds

0

and now if we write cos ((t s)) = <[exp [i(t s)]] we get:
Z

Z
(t) =
G(s)<[exp [i(t s)]] ds = <
G(s) exp [i(t s)] ds
0
0

Z
G(s) exp [is] ds
= < exp [it]
0

and the integral on the right is now a one-sided Fourier transform. Since it has no
dependence on t, it is just a complex number. By convention we define the complex
shear modulus, G , as:
Z

G = i
G(s) exp [is] ds,
0

and its real and imaginary parts:

G = G0 + iG00
are called the storage modulus G0 and the loss modulus G00 .
This gives
(t) = <[exp [it](iG )] = <[(cos (t) + i sin (t))(G00 iG0 )]
G00
(t).

= [G0 sin (t) + G00 cos (t)] = G0 (t) +

13

Now a purely viscous fluid would give a response

(t) = (t)

= cos (t)
and a purely elastic solid would give
(t) = G0 (t) = G0 sin (t).
We can see that if G00 = 0 then G0 takes the place of the ordinary elastic shear modulus
G0 : hence it is called the storage modulus, because it measures the materials ability to
store elastic energy. Similarly, the modulus G00 is related to the viscosity or dissipation
of energy: in other words, the energy which is lost.
Since the role of the usual Newtonian viscosity is taken by G00 /, it is also common to
define
G00
0 =

as the effective viscosity; however, the storage and loss moduli G0 and G00 are the most
common measures of linear rheology.
Let us check a few values:
G(t) = (t)
G(t) = G0
G(t) = G0 exp [t/ ]

G () = i
G () = G0 NB: By observation not integration
iG0
G0 ( 2 2 + i )
G () =
=
(1 + i )
1 + 2 2

Let us look more at the single-relaxation-time model. At slow speeds 1 we have

G iG0
so the material looks like a viscous fluid with viscosity G0 . At high speeds 1, we
have instead
G G 0
and the material looks like an elastic solid with modulus G0 .
If we plot the two moduli, G0 and G00 against the graph looks like this:

G0

G00

Frequency,
14

The point where the two graphs cross is given by:

G0 = G00
G0
G0
=
1 + 2 2
1 + 2 2
= 1.
2 2

A real fluid has more complex dynamics:

Typical dynamic data for an elastic liquid
6

Frequency,
but a first engineering approximation often used is to take the value of c at which the
two curves cross, and use = c1 as the relaxation time of the fluid. Alternatively, a
series of relaxation modes is used to fit the curves as well as possible.

15

16

Microscopic dynamics

In this section we will look at the first model that people came up with when they started
to model polymers from the microscopic level. Its called the Oldroyd B model. We
will spend one double lecture to derive it and in the last lecture, look at its macroscopic
(top-down) properties:
Linear rheology
Stress in shear
Stress in extensional flow

4.1

The stress tensor

Remember we had the three Navier-Stokes equations (in incompressible form):

u=0
This is mass conservation: integrated over a volume of space, it tells us that fluid flow in
matches fluid flow out, so fluid does not accumulate in one place.

u
+ u u =

t
= pI + (u + (u)> )
In the last two equations, is the stress tensor. The physical meaning of this is
calculated from the force the force exerted across a small surface element with area S
and unit normal n:
n

The force acting on the top surface of the tiny pillbox is n S.

4.2

Beads in a viscous fluid

Probably the simplest object you could put into a flow is a solid spherical bead. Lets
look briefly at the motion of a microscopic solid sphere in a Newtonian fluid.
There is really only one parameter in Newtonian fluid flow (once youve made everything
dimensionless): the Reynolds number, Re, which gives the ratio of inertial to viscous
forces. If the scale of the flow is very small, then the Reynolds number is very small and
we can assume it is zero. Then the governing equations are the Stokes equations:
2 u = p.

u=0
17

There isnt time in this course to go into the details of these flows: well just accept one
key fact. In a fluid which is otherwise at rest, a sphere of radius a, moving with constant
velocity U , experiences a drag force of
F = 6aU .
In the same way, a sphere moving with velocity U relative to the fluid around it experiences the same drag force.

4.3

Microscopic model

4.3.1

Bead-spring dumbbell

The dynamics of a single sphere in a viscous fluid are not very interesting. Interactions
between them can be. . . but thats a whole other research area.
In order to model a polymer, we need something which can in some way remember the
flows it has been through. We do this by creating something which can change its physical
configuration as it responds to flow: then when the flow stops, its configuration is the
way it carries the information about the flow history. This is exactly what real polymers
do: but our model will be much simpler than real polymers.
We take two spherical beads, and connect them together with a linear spring. This is
the simplest possible deformable object. The spring is intended to represent the entropic
tendency of the polymer coil to end up in a random, relaxed state. So if you stretch out
the coil, it will tend to recoil: the spring tends to bring the beads back together.
Now there are three effects acting on each bead:
The spring force
A drag force if it is moving relative to the fluid around it
Brownian motion
We will build up the behaviour of the whole dumbbell bit by bit.
4.3.2

Separate beads in deterministic motion

Let us just look at one bead of radius a. Well choose the radius to represent the typical
size of a relaxed polymer coil. Now say its instantaneously at a position x; and suppose
the other bead is instantaneously at x + r. Well look at what it does under the action
of a fluid flow and the spring force and nothing else.
Because it is very small, it has no inertia, and Newtons second law becomes a force
balance: total force on our bead is zero.
The spring force pulling it towards the other bead is r. is just a spring constant.
Because it comes from thermal forces, we know its size:
=

3kT
a2

18

which also depends on the typical size of the relaxed polymer coil.
If the bead at x is moving with velocity U 1 , and the fluid around it has velocity u(x),
the drag force on the bead will be
6a[u(x) U 1 ].
The total force on the bead (which must be zero) is
3kT
r + 6a[u(x) U 1 ] = 0
a2
so we can work out the velocity of the bead:
U1 =

kT
r + u(x).
2a3

The other bead is positioned at x + r and the spring force on it is r so its velocity is
U2 =

kT
r + u(x + r).
2a3

In the absence of Brownian motion, these two velocities tell us how the bead positions
evolve:
kT
dx
=
r + u(x)
dt
2a3
kT
d(x + r)
=
r + u(x + r)
dt
2a3
Finally we subtract the two to get the evolution of r:
dr
kT
=
r + u(x + r) u(x)
dt
a3
Since r is microscopically small relative to the scale over which u changes, we can use a
Taylor series for u(x + r):
dr
kT
=
r + r u(x)
dt
a3
which is correct up to terms of order r.
The constant 2kT /a3 is called 1/ ; it represents the inverse relaxation time of the
dumbbell following distortion caused by the flow.
dr
1
= r + r u.
dt
2
19

4.3.3

Stress in the fluid

Now suppose we have a suspension containing many of these dumbbells. What will the
extra stress they contribute be? Of course, the fluid they are suspended in, which has
its own viscosity , will continue to contribute a Newtonian stress; were just looking for
the polymer extra stress, p .
Again, we look at the tiny surface element with area S and unit normal n. We are
interested in dumbbells that cross the surface.
n

The force associated with a single dumbbell that crosses the surface is r; how many will
cross the surface?
Clearly a long dumbbell is more likely to cross the surface than a short one; and a
dumbbell which is lined up with n is more likely to cross the surface than one which
is perpendicular to it. If there are m dumbbells per unit volume, then we expect the
number crossing our surface element to be
mr n S.
Then the extra stress exerted by the dumbbells is
p n S = r(mr n S)

so

p = mrr =

3am
rr.
2

Again, we introduce a name for the new constant: this time, G = 3am/2 so that
p = Grr.
4.3.4

Adding Brownian motion

We have worked out how a dumbbell will evolve if its motion is deterministic, and the
stress it will cause. But to complete the model, we need to add Brownian motion of the
two beads.
We simply add a standard 3D Brownian motion (i.e. each of the x, y, z components is a
Brownian motion) to the evolution of the vector r:

1
dr = r + r u(x) dt + 1/2 dB t
2
We have chosen 1/2 as the coefficient for two reasons: it has the correct dimension; but
also because (as we will see later) this leads to a dumbbell with no flow having a natural
length of 1.
20

Because we eventually want to work out the extra stress, we then define the conformation tensor
A(x, t) = IE [r(x, t)r(x, t)]
which will give us
p = GA.
Now we use the first-step method (see GM01 part 1). As we take a time-step dt, the
position x of our dumbbell will move to
x + dx = x + u(x)dt.
So when we take our time-step, we get
A(x + u(x)dt, t + dt) = IE [(r + dr)(r + dr)]
= IE [r r + r dr + dr r + dr dr]
A(x + u(x)dt, t + dt) = IE [r r] + IE [r dr] + IE [dr r] + IE [dr dr]
Now lets look at these four terms separately. As usual, we only keep terms up to order
dt.
IE [r r] = A(x, t)

1
1/2
IE [r dr] = IE r
r + r u(x) dt +
dB t
2


1
= IE r
r + r u(x) dt
2
1
= Adt + A udt
2

1
1/2
IE [dr r] = IE
r + r u(x) dt +
dB t r
2

1
= IE
r + r u(x) dt r
2
1
= Adt + (u)> Adt
2
h r

i
IE [dr dr] = IE
+ r u dt + 1/2 dB t
+ r u dt + 1/2 dB t
2
1/22 1/2

= IE
dB t
dB t
1
= Idt
We also need to look at the left-hand side term.
A(x + u(x)dt, t + dt) = A(x + u(x)dt, t + dt) A(x, t + dt) + A(x, t + dt)
= (udt )A + A(x, t + dt)
= (udt )A + A(x, t + dt) A(x, t) + A(x, t)
A
= (u )Adt +
dt + A(x, t)
t
21

Putting these all together gives:

A
1
1
1
+ (u )A dt = Adt + A udt Adt + (u)> Adt + Idt
t
2
2

A
1
+ (u )A A u (u)> A = A I
t

and the polymer extra stress is

p = GA.
The terms on the left-hand side of the A equation are called the co-deformational time
derivative because they represent what happens to a line element that is deforming with
the flow.
We have now derived the Oldroyd-B equations:
u=0

+ u u =
t

= pI + u + (u)> + GA

A
1
+ (u )A A u (u)> A = A I
t

In the final lecture we will see how the fluid corresponding to these equations behaves.

22

The Oldroyd-B fluid

Last time we started from a microscopic dumbbell with a linear entropic spring, and
derived the Oldroyd-B equations:
u=0

u
+ u u =

= pI + u + (u)> + GA

A
1
+ (u )A A u (u)> A = A I
t

(1)
(2)
(3)
(4)

Note that since A = IE[r r], it must be symmetric.

5.1

Shear flow

Suppose we make our fluid carry out an unsteady shear flow:

u = ((t)y,

0, 0).
If the forcing all depends on y and t only, we expect all the physical variables only to
depend on y and t. The mass conservation equation (1) is satisfied. The momentum
equation (2) becomes
u
= .

t
Now the tensor u is

0 0
u =
(t)

0
so (3) gives
=

p (t)

(t)

+ GA

and (4) becomes

1 Axx 1
Axx Axy
Axy
A
xy 0
A
xy A
yy

=
Axy
Ayy 1
A
yy 0
0
0
t Axy Ayy

5.1.1

Steady shear flow

Now set as a constant. This means that all the variables should be independent of t:
= 0.

p
+ GA
=
p
23

1 Axx 1
Axy

=
Axy
Ayy 1

Lets look at the last equation, for the components of A. It gives three scalar equations:

A
xy 0
A
yy 0

A
xy A
yy
0
0

1
A
xy A
xy = (Axx 1)

1
A
yy = Axy

0 = Ayy 1.
Solving from the bottom up gives
Ayy = 1

Axy =

Axx = 1 + 2 2 2 .

The total stress becomes

=

p + G + 2G 2 2 ( + G )
( + G )
p + G

The presence of the polymers has made two changes to the Newtonian stress:
The viscosity is increased from to + G
There is a difference between the two diagonal stress components.
This difference is called the first normal stress difference
N1 = xx yy = 2G 2 2 .
As a stress acting along the flow lines (xx-direction) it has the opposite sign to pressure
so it acts like a tension: the streamlines are like stretched rubber bands. It is the driving
force behind the rod-climbing experiment:

24

This picture is from Boger & Walters book Rheological Phenomena in Focus. The
rod in the middle is rotated, causing a shear flow round the outside. The streamlines are
circular, so their tension causes the fluid to move to the middle and the only place it
can go is up the rod.
5.1.2

Linear rheology

Lets return to the time-dependent shear flow:

u
= ;
t

u = ((t)y,

0, 0).

p (t)

+ GA
=
(t)

and

Axx Axy
Axy Ayy

A
xy 0
A
yy 0

A
xy A
yy
0
0

1
=

Axx 1
Axy
Axy
Ayy 1

For the standard linear rheology experiment, we set

(t)

= cos (t).
Well assume the motion starts at t = 0. Before that there is no flow so the fluid was
relaxed and A = I.
Lets look at the evolution of A. To get the rheology, we will only need Axy , so well only
use the Axy and Ayy equations.
Axy
1
A
yy = Axy
t

Ayy
1
= (Ayy 1).
t

The second of these is satsfied by Ayy = 1, so the first gives

Axy
1
cos (t) = Axy
t

Axy 1
+ Axy = cos (t)
t

Axy e(t/ ) = cos (t)e(t/ )
t

This is one of those tricky integrals that works iteratively by parts:

Z t
0
I =
cos (t0 )e(t / ) dt0
0
Z t
0
0 (t0 / ) t
= [ cos (t )e
]0 +
2 sin (t0 )e(t / ) dt0
0
Z t
0
0
0 (t / ) t
2 2
0 (t0 / ) t
= [ cos (t )e
]0 + [ sin (t )e
]0
3 2 cos (t0 )e(t / ) dt0
= cos (t)e

(t/ )

2 2

+ sin (t)e

(t/ )

25

0
2 2

Finally the expression for Axy is

Axy =

1
cos (t) + 2 2 sin (t) e(t/ )
2
2
1+

For long times, the e(t/ ) term becomes very small so we will neglect it. The total shear
stress xy then becomes

G 2 2
G
(t)

+
(t)
xy = +
1 + 2 2
1 + 2 2
Thus the linear rheology functions for the Oldroyd-B fluid are
G0 =

G 2 2
1 + 2 2

G00 = +

G
.
1 + 2 2

This is just like the single exponential relaxation fluid if we set = 0; that would give the
Upper Convected Maxwell model. With a nonzero viscosity, although the relaxation
time is still , the storage and loss modulus no longer cross at = 1 .

5.2

Extensional flow

Finally, we look at a steady 2D extensional flow:

u = (x,
y).

Again, this satisfies mass conservation. This time we have

0
.
u =
0
The stress is
=

p + 2
0
0
p 2

+ GA

and the evolution of A (independent of time and position) becomes

1
A
xx A
xy
A
xx
A
xy

= AI
A
xy A
yy
A
xy A
yy

so
Axx =

1
(1 2
)

Axy = 0

Ayy =

1
(1 + 2
)

The total stress is

p + 2 + G/(1 2
)
0
.
=
0
p 2 + G/(1 + 2
)
Since in a Newtonian fluid we have

p + 2
0
=
0
p 2
26

we can define an extensional viscosity as

ext =

xx yy
.
4

The Oldroyd-B fluid has extensional viscosity

ext = +

G
(1 42 2 )

This strain-hardens (gets thicker with increasing speed) for low strain rates but for higher
strain rates disaster strikes:

The viscosity diverges at a strain rate of = 1/2 and for strain rates slightly larger, the
viscosity value is negative!
The moral of this: a linear spring is fine for shear flows, where the stretch is fairly
moderate; for stretching flows, a linear spring can stretch indefinitely and give infinite
forces. The standard workaround at this point is to use a nonlinear spring law (FENE
model, finie extensibility nonlinear elasticity) which brings with it its own complications.
There is no single right answer to polymer modelling but hopefully you now have an
idea about how to start!

27