R h e ol o g i ca l Behav i o r and P o l ym er Pr op er t i es G. C. Berry Department of Chemistry Carnegie Mellon University Colloids, Polymers and Surfaces e-mail: gcberry@andrew.cmu.edu web site: http://www.chem.cmu.edu/be rry Carnegie Mellon 1 Int roduction 3 (12) Rheological methods 16 (19) Linear elastic parameters 26 (5) Linear visc oelastic functions 33 (12) Several viscoelastic experiments 44 (16) Relations among linear visc oelastic functions 62 (10) Examples of linear visc oelastic functions 73 (9) Time-temperature equivalence 83 (9) The glass transition temperature 93 (13) The visc osity 107 (26) Effects of polydispersity 134 (4) Network formation 139 (13) Isochronal B ehavior 153 (6) Examples from the literature 160 (45) Branched and li near metallocene polyolefins 161 (10) Colloidal d ispersions 172 (9) Wormlike Micelles 182 (4) Deformation of rigid materials 187 (4) Nonlinear shear behavior 192 (16) 209 (6) Carnegie Mellon Linear and nonlinear bulk properties 2 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 3 POLYMERS NATURAL PROTEINS POLYNUCLEOTIDES POLYSACCHARIDES Carnegie Mellon SYNTHETIC GUMS RESINS THERMOPLASTIC THERMOSETTING ELASTOMERS 4 Some Common Elastomers, Plastics and Fibers ELAS TOME RS PLAS TI CS Polyisoprene polyethylene polyisobutylene polytetrafluoroethylene poybutadiene polystyrene FIBERS poly(methyl methacrylate) Phenol-formaldehyde Urea-formaldehyde Melami ne-formaldehyde Poly(vinyl chloride) Polyurethanes Polysiloxanes Polyami de Polyester Polypropylene Carnegie Mellon 5 Fraction of Molecules With Molecular Weight M Mn Mw Mz Molecular Weight M A Schematic Illustrat ion of a Typical Distribution of Molecular We ights, showing Mn, Mw, and Mz Carnegie Mellon 6 A generalized Average of molecular weights: wµ is the weight f raction of polymer with molecular w eight Mµ: M() = wµM µ µ Special C ases: Number average: = M()/M = 1/ wµM µ µ Mw = M()/M() Mz = M/M() Mn Weight average: = wµMµ µ z-average: = wµMµ wµMµ µ µ G. C. Berry "Molec ular Weight Distribution" Encyclopedia of Materials Science and Engineering, ed. M. B. Bever, Pergamon Press, Oxford, 3759-68 (1986) Carnegie Mellon 7 Specific Volume Tm Temperature A schematic v-T diagram fo r a typical nonpolymeric material. Carnegie Mellon 8 Specific Volume Tg Tm Temperature A schematic v-T diagram for a typical semi-crystalline polymeric material. Carnegie Mellon 9 Specific Volume Tg Temperature A schematic v-T diagram fo r a typical noncrys talline polymeric material. Carnegie Mellon 10 Stress Rigid Plastic Flexible Plastic Elastomer Strain Typical Stress-Strain Behavior for Plastics and Elastomers Carnegie Mellon 11 F. W. Billme yer Jr. (1976): J. Po lym. Sc i.: Symp. (1976) 55: 1-10 "…characterization of polymers is inherently more difficult than that of other materials. Polymers are roughly equivalent in comple xity to, if not more compl ex than, other materials, at every physical level of organization from microscopic to macroscopic…" "We would wish, ideally, to characterize all aspects of a polymer structure in enough detail to predict its performance from first principles. I seriously doubt that this will ever be possible, and I am sure that even if it were, it would n ever be economically feasible." Carnegie Mellon 12 Carnegie Mellon 13 2-D projection of a random arrangement of a chain with 1000 non-overlapping bonds, each step otherwise randomly selected Carnegie Mellon 14 Mean chain dimensions: For a linear chain with contour length L (without excluded volume effects): Mean square-end-to-end dimension: RL2 = 2âL â is the persistence length (2â is the Kuhn length) for a flexible chain, â << L. Mean square-radius of gyration: RG2 = RL2 /6 = âL/3 Carnegie Mellon 15 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 16 Schematic of Rheometer System Computer System for Data Acquisition and Instrument Control Shear Stress vs Time (Frequency) Shear Strain vs Time (Frequency) Normal Force vs Time (Frequency) Temperature vs Time Carnegie Mellon Output Interfaces Torque Transducer Force Transducer Position Transducer Shape Transducer Temperature Transducer Rheometer 17 CONTROLLED STRESS IN TENSION "Frictionless" Bearing Position Transducer Sample Removable Weight Device Carnegie Mellon Tare Output Remov able Weigh t Input Controlled weight Positi on Transduc er Measure of sha ft positi on Volt age (current) Controll ed force 18 CONTROLLED DEFORMATION IN TENSION Drive Screws Crosshead Position Transducer Sample Device Carnegie Mellon Output Crosshead D rive Input Controlled Drive Positi on Transduc er Measure of sha ft positi on Volt age (current) Controll ed force 19 CONTROLLED STRESS RHEOMETER Controlled Torque Drive Angle Position Transducer Shaft "Frictionless mount" Sample Fixtures Fixed Shaft (Alternate: controlled rotation) Carnegie Mellon Device Input Output Controlle d To rque Drive Controlle d vol tage Controlle d torque Ang le Posit ion Transdu cer Measure of shaft angl e Volt age (current) 20 CONTROLLED DEFORMATION RHEOMETER Controlled Rotation Drive Angle Position T ransducer Shaft "Frictionless mount" Sample Fixtures T orque T ransducer (Force T ransducer) Carnegie Mellon Device Input Output Controll ed Deformation D rive Controll ed vo lt age Controll ed shaft rotation 21 Electromagnetic Coils I: A-F II: a-f d E F e , c D f G Iron Core C b g a • • Carnegie Mellon B A h H Aluminum Cylinder Attached to Rotor Phasing of the currents in Coils I and II can produce a timedependent torque: ³ Constant torque amplitude ³ Sinusoidal torque amplitude Torque amplitude may readily be varied over a factor of 10. 22 Parallel Plates Sample Fixtures Height h 2R Cone & Plate Sample Fixtures Angle 2R Concentric Cylinders Sample Fixtures h 2R Carnegie Mellon R 23 Geometric Factors in Rheometry a Geometry Measured Translational geome tries Parall el Plate Force: F Stress: = F /wb Displa cement: Force: D F Strain: = D/h Displa cement: D Stress: Strain: = F /2š Rh = D/Rln(1 + /R) Rotational geometries Parall el Plate outer radius R; separation h Torque : M Stress: = (2r/R)M /R Rotation: Strain: Cone & Plate outer radius R; cone ang le š - Torque : Rotation: M Stress: Strain: = (3/2)M /R = (1/) Concentric Cyli nde rs inne r radius R; gap ; heigh t h Torque : Rotation: M Stress: Strain: (R/2h)M /R (r) (R/R) f(R,r) width,w; breadth b; separation h Concentric Cyli nde rs inne r radius R; gap ; heigh t h Calculated (r) = (r/h) 2 1 + R f(R,r) = (R/r) 1 + /2R a and are the shea r stress and s train, respectively Carnegie Mellon 24 Functions and Parameters Used Function/Para meter Symbol Units Time t T Frequency Strain Co mponen t ij --- Elong ation al s train --- Shea r strain --- Rate of shea r Ý, Ý T-1 Stress Componen t Sij Shea r stress ML-1T-2 Modulus G, K, E ML-1T-2 Compli anc e J, B, D M-1LT2 Viscosit y Carnegie Mellon T-1 ML-1T-2 ML-1T-1 25 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 26 Linear elastic phenomenology Shear stress Shear strain = J = (1/G) Elongational s tress Elongational s train = D = (1/E) Pressure ² P Volume change ² V ² V/V = B² P = (1/K)² P Carnegie Mellon 27 Linear Elastic Functions Shear Compliance J Shear Modulus G Bulk Compliance B Bulk Modulus K Tensile Compliance D = J/3 + B/9 Tensile Modulus 1/E = 1/3G + 1/9K Carnegie Mellon 28 Linear elastic phenomenology ij uj 1 ui = 2 x + x ; j i u is the displacement vector 2ij = J [Sij – 1 3 Sij = 2G [ij – ij S] + (2/9)ij B S 1 3 ij ] + ij K ij = 1 if i = j, and ij = 1 if i ° j In this notation, Shear stress = S 12 Shear strain = 212 Carnegie Mellon 29 Relations Among Linear Elastic Constants K, G E, G K, E K, E, G, K K EG 33G – E K K E 31 – 2 2G1 + 31 – 2 E 9KG 3K + G E E 3K(1 – 2) E 2G(1 + ) G G G 3KE 9K – E 3K1 – 2 21 + E 21 + G 3K – 2G 6K + 2G E 2G – 1 3K – E 6K Carnegie Mellon J = 1/G, B = 1/ K, D = 1/E 30 1 Pa = 1.45·10-4 psi Graphite whisker 12 Carbon fiber KevlarTM fiber PE chain direction Cellulose chain direction Log E/Pa 11 PVOH Avg textile fiber 10 PE Amorphous Glass 9 Nonpolar Polar Interchain stretch forces Carnegie Mellon Rotation Bending Stretch Intrachain forces Deformation Modes 31 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 32 Linear Elastic Functions Shear Compliance J Shear Modulus G Bulk Compliance B Bulk Modulus K Tensile Compliance D = J/3 + B/9 Tensile Modulus 1/E = 1/3G + 1/9K Carnegie Mellon 33 Linear Viscoelastic Functions Shear Compliance J(t) Shear Modulus G(t) Bulk Compliance B(t) Bulk Modulus K(t) Tensile Compliance D(t) = J(t)/3 + B(t)/9 Tensile Modulusa 1/Ê(s) = 1/3Ĝ(s) + 1/9K̂(s) a. The superscript "ˆ" denotes a Laplace transform. Carnegie Mellon 34 Linear viscoelastic phenomenology— Stress Controlled (t) J(t – ti) i = = i= t (t) •0 d(u) J(t – u) (u) u (t) = (t) J(u) • = Jo(t) + • du (t – u) u 0 (t) • duJ(t – u) -• 2 1 (t) (t) Carnegie Mellon t1 t2 t 35 Linear viscoelastic phenomenology— Strain Controlled (t) = G(t – ti) i i= = (t) •0 (t) = (u) • duG(t – u) u -• (t) = Go(t) + d(u)G(t – u) t • •0 du (t – u) G(u) u (t) (t) (t) Carnegie Mellon t1 t2 t 36 Linear elastic phenomenology ij uj 1 ui = 2 x + x ; j i u is the displacement vector 2ij = J [Sij – 1 3 Sij = 2G [ij – ij S] + (2/9)ij B S 1 3 ij ] + ij K ij = 1 if i = j, and ij = 1 if i ° j In this notation, Shear stress = S 12 Shear strain = 212 Carnegie Mellon 37 Linear viscoelastic phenomenology uj 1 ui ij = 2 x + x ; j i u is the disp lacement vector Sij(s) 2ij(t) = • ds{J(t – s) s – -• t 1 3 S(s) ij s S(s) + (2/9)ij B(t – s) s } ij(s) Sij(t) = • ds{2G(t – s) s – -• t (s) s 1 3 ij (s) + ijK(t – s) s } In this notation, Shear stress (t) = S12 (t) Shear strain (t) = 212(t) Carnegie Mellon 38 Relation between G(t) and J(t) 1 t t 0du • G(t – u) J(u) = 1 s2Ĝ(s)Ĵ(s) = 1 with Laplace transform: Carnegie Mellon 39 Shear Compliance J(t) and Recoverable Shear Compliance R(t) R(t) = J(t) – t/ = J• – J• – Jo(t) (t): Retardation Function Shear Modulus G(t) G(t) = Ge + Go – Ge(t) (t): Relaxation Function the (linear) viscosity, w ith 1/ = 0 for a solid, Ge the equilibrium modulus, with Ge = 0 for a fluid, Go the "instantaneous" modulus, with JoGo = 1, and J• the limit of R(t) for large t: Solid: J • = Je = 1/Ge; equilibrium compliance Fluid: J • = Js; steady-state recoverable compliance Carnegie Mellon 40 Creep Shear Compliance J(t) R(t) = J(t) – t/ = J• – J• – Jo(t) Shear Modulus G(t) G(t) = Ge + Go – Ge(t) Linear elastic solid: 1/ = 0, J• = Je = 1/Ge, (t) = (t) = Linear viscous fluid: 1/ > 0, Go = 0, (t) = (t) = (t) Linear viscoelastic soli d: 1/ = 0, J• = Je = 1/Ge, 0 < (t) < (t) Š 1 Linear viscoelastic fluid: 1/ > 0, J• = Js (= J oe), 0 < (t) < (t) Š 1 Bulk Compli ance B(t) = Be – Be – Bo(t) Bulk Modulus K(t) = Ke + Ko – Ke(t) 41 Carnegie Mellon 42 Simple example of the relation between G(t) and J(t) Maxwell fluid: G(t) = Goexp(- t/); J(t) = Js + t/; = /Go Js = Jo = 1/Go R(t) = Js Note: (t) = exp(-t/) and (t) = 0 for this model. Carnegie Mellon 43 Often used relations for (t) and (t) A weight set of exponentials with N relaxation times: N-1 (t) = exp(–t/ ) = J i i m N (t) = exp(–t/ ) = G i 1 i 1 • • dln L()exp(–t/) -• – J • o 1 • • dln H()exp(–t/) o – Ge -• Notes: i = i = 1, and m is equal to 0 o r 1 for a solid and flu id, resp. m0 > 1 > 1 > … > i > i > i+1 > … > N-1 > N (The contribution 0 is absent for a flu id) Carnegie Mellon 44 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 45 (b) Stress Relaxation (t) = o (t) 0 (• ) 0 (t) = a + bt R() Strain Strain o Stress Stress (a) Creep & Recovery (t) = o (t) t = Te 0 0 = t - Te t t Tim e Tim e (c) Ramp Deformation & Recovery o Stress Stress (t = Te ) (d) Sinusoid Deformation (t) 0 (t) P = 1/2 ( ) R o Strain Strain 0 (t) 0 . (t) = t t = Te 0 = t - Te t Tim e Carnegie Mellon Tim e 46 Creep and recovery w ith a step shear stress Stress Stress history: (t) = 0 t <0 (t) = o 0 Š t Š Te (t) = 0 t > Te (t) = o 0 Strain (t) = a + bt R() = (t = T e) - (t) (t) t = Te 0 t Carnegie Mellon q = t - Te Time 47 The strain in creep for t Š T e: (t) t = o• du J(t – u) (u - 0 = oJ(t) = oR(t) + t/ Carnegie Mellon 48 The strain in creep for t Š T e: (t) t = o• du J(t – u) (u - 0 = oJ(t) = oR(t) + t/ The strain for = t – T e > 0 in recovery: Te t (t) = o•0 du J(t – u) (u - 0) – o• du J(t – u) (u - Te) T e () = oJ( + Te) – J() = oR( + Te) – R() + Te/ Carnegie Mellon 49 The strain in creep for t Š T e: (t) t = o• du J(t – u) (u - 0 = oJ(t) = oR(t) + t/ The strain for = t – T e > 0 in recovery: T t (t) = o•0edu J(t – u) (u - 0) – o• du J(t – u) (u - Te) T e () = oJ( + Te) – J() = oR( + Te) – R() + Te/ The recoverable strain R() = (Te) – (t) for > 0: R() = o{J(Te) – J( + Te) – J()} Carnegie Mellon = o{R() + R(Te) – R( + Te)} 50 Stress relaxation after a step shear strain Strain h istory: (t) = 0 t<0 (t) = o t •0 Stress o (t) ) (• Strain 0 (t) = o 0 t Carnegie Mellon Time 51 The stress response for t > 0: t (t) = o• du G(t – u) (u - 0) = oG(t) 0 = o{Ge + (Go – Ge)(t)} (•) = oGe Carnegie Mellon 52 Recovery after a ramp shear strain Strain history: Stress history: (t) = 0 t <0 (t) = Ýt 0 Š t Š Te (t) = 0 t > Te Stress (t = T e ) (t) 0 Strain R(q) = (t = T e) - (t) . (t) = t t = Te 0 =t-T t e Time Carnegie Mellon 53 The stress response for t Š Te : t t (t) = Ý• du G(t – u) = ÝGet + (Go – Ge)• ds (s) 0 0 For a fluid in steady-state deformation, = Ý , or • = (•)/Ý = Go • ds (s) 0 The strain for t > Te: Te (t) = 0 = Ý•0 (u) du G(t – u) + • du G(t – u) u Te t For large Te and t, (full recoil after steady flow) it can be shown that for a fluid this give s: c Js = Carnegie Mellon • • •0 ds s(s)/•0 ds (s) 54 Carnegie Mellon 55 The strain response for t > 0: t (t) = o• du J(t – u)cos(u) 0 In the steady-state limit with large t : (t) = o{J'()sin(t) – J''()cos(t)} In-ph ase (or real or storage) dyna mic compliance: • J'() = J• – J• – J o• ds(s)sin(s) 0 Out-of-phase (or imaginary or loss) dyn amic compliance • J"() = (1/) + J• – J o• ds(s)cos(s) 0 Carnegie Mellon 56 Alternatively (t) = o |J*()|sin t – "Dynam ic compliance": 2 2 2 |J*()| = J'() + J"() Phase angle (): tan () = J"()/J'() For small : J'() - J•, Carnegie Mellon J"() - 1/, and J"() – 1/ - 57 Oscillation with a si nusoid s hear strain Strain his tory: (t) = 0 t<0 (t) = osin(t) t •0 The stress response for t > 0 is giv en by t (t) = o • du G(t – u)cos(u) 0 In the steady- state limit with large t, (t) = o{G'()sin(t) + G''()cos(t)} In-ph ase (or real or storage) dyna mic compliance: • G'() = Ge + Go – Ge•0 ds(s)sins) Out-of-phase (or imaginary or loss) dyn amic compliance • Carnegie Mellon G''() = Go – Ge•0 ds(s)cos(s) 58 Alternatively (t) = o |G*()|sin t + "Dynam ic compliance": |G*()| = G'() + G"() 2 2 2 Phase angle (): tan () = G"()/G'() For small : • ds 0 G'() - Ge + Go – G e• s(s) fluid ( Js • flu id G''() = Go – Ge• ds(s) 0 Carnegie Mellon 59 Exact relations among the dynam ic modul i and comp liances: |G*()||J*()| = 1 2 J'() = G'()/|G*()| 2 J"() = G"()/|G*()| 2 G'() = J'()/|J*()| 2 G"() = J"()/|J*()| tan () = J"()/J'() = G"()/G'() Carnegie Mellon 60 The dynamic viscosity: In-pha se with the s train rate: '() = G"()/ Out-of-pha se with the s train rate: "() = G'()/ For small : • '() = Go – Ge• ds(s) 0 flu id • fluid ''() - Ge/ + Go – G e• ds s(s) Js 0 Carnegie Mellon 61 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 62 Linear Viscoelastic Functions Shear Compliance J(t) Shear Modulus G(t) Bulk Compliance B(t) Bulk Modulus K(t) Tensile Compliance D(t) = J(t)/3 + B(t)/9 Tensile Modulusa 1/Ê(s) = 1/3Ĝ(s) + 1/9K̂(s) a. The superscript "ˆ" denotes a Laplace transform. Carnegie Mellon 63 Relation between G(t) and J(t) 1 t t 0du • G(t – u) J(u) = 1 s2Ĝ(s)Ĵ(s) = 1 with Laplace transform: Carnegie Mellon 64 J(t) 8 Log(Compliance/cgs) R(t) -6 6 1/G(t) G(t) -8 0 2 4 6 8 10 4 12 Log(Modulus/cgs) -4 14 Log (Time/sec) Carnegie Mellon 65 J(t) 8 Log(Compliance/cgs) R(t) -6 6 1/G(t) G(t) -8 2 0 4 6 8 10 4 12 Log(Modulus/cgs) -4 14 Log (Time/sec) -4 Log(Compliance/cgs) -6 J”(w) 6 -8 G”(w) 4 G’(w) Log(Modulus/cgs) 8 J’(w) -10 -14 -12 -10 -8 -6 -4 -2 0 Log(Frequency/sec -1) Carnegie Mellon 66 -4 Log(Compliance/cgs) G(t) J’(w) -6 6 -8 4 G’(w) 0 2 4 6 8 10 Log (Time/sec) & –Log(Frequency/sec 12 Log(Modulus/cgs) 8 R(t) 14 -1) Figure 14 Carnegie Mellon 67 An often used relation between G(t) and J(t) A weight set of exponentials with N relaxation times: N-1 (t) = 1 • i exp(–t/i) = • dln L()exp(–t/) J• – Jo -• m N (t) = i exp(–t/i) = Go 1– Ge •-•• dln H()exp(–t/) 1 Notes: i = i = 1, and m is equal to 0 or 1 for a solid and fluid 0 > 1 > 1 > … > i > i > i+1 > … > N-1 > N (0 absent for a fluid) Carnegie Mellon 68 Determination of L() (or the i-i set) from J(t) (Similar considerations apply to the determination of H() (or the i-i set) from G(t)) Derivative methods for L(): 1st Approx.: L() - M(m) [R(t)/ln t]t = M(m) = 2nd Approx.: Carnegie Mellon L() - lnL()/ln (interative) [J(t)/ln t – J(t)/ln t)]t = 2 69 Carnegie Mellon 70 Determination of L() (or the i-i set) from J(t) (Similar considerations apply to the determination of H() (or the i-i set) from G(t)) Inverse transform methods for i-i: The inverse transform is "ill-posed", and a stable solutions requires constraints (e.g., i • 0) In an often used strategy, a set of logarithmically spaced i are chosen such that the span in 1/I does not exceed the time span in the experimental data. A constrained nonlinear least squares analysis is then used to determine the i. Commercial packages are available for this transform. Carnegie Mellon 71 Carnegie Mellon 72 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 73 Carnegie Mellon 74 Carnegie Mellon 75 H( ) Go '() = G''() t/ G(t) tan () L( ) J(t) Jo slope = 1/3 slope = 1 logor log log t log Low Molecular Weight Glass Former Go G(t) t/ H( ) '() = G''() slope = -1/2 tan () slope = 1 J(t) Jo slope = 1/3 L( ) logor log log t log Polymeric Fluid with M < Me Go G(t) t/ slope = -1/2 Jo JN '() = G''() H( ) GN tan () slope = 1 L( ) J(t) slope = 1/3 log t Carnegie Mellon logor log Polymeric Fluid with M >> Me log 76 log (L( )/Pa) log (R(t)/Pa) -4 Js -6 Slope = 1/3 Narrow MWD -8 -4 Slope = 1/3 -6 3 2 1 -8 -4 -2 0 Narrow MWD 2 4 6 8 10 log (t/s) or log (/s) Carnegie Mellon 77 Peak I with L() linear in 1/3 before the peak decreases sharply to zero. The behavior ascribed to peak I, first reported by Andrade, is seen in a variety of materials, such as metals, ceramics, crystalline and glassy polym ers and small organic mole cules; the decrease of L() to zero being eviden t in examples of the latter. The area under peak I provides the contribution JA – Jo to the total recoverable compliance J s. It seems likely that th e mechan ism giving rise to peak I may be distinctly different from the largely entropic origins of peaks II and III described in the following. Carnegie Mellon 78 log (L( )/Pa) log (R(t)/Pa) -4 Js -6 Slope = 1/3 Narrow MWD -8 -4 Slope = 1/3 -6 3 2 1 -8 -4 -2 0 Narrow MWD 2 4 6 8 10 log (t/s) or log (/s) Carnegie Mellon 79 Peak II that inc reases in peak area with increasing M until reaching a certain level, beyond which the peak is inva riant with increasing M, both in area and position in Peak II is ascribed to Rouse-like modes of motion, ei ther fluidlike for low molecular weight in the range for which the area increases with M, or pseudo-solid like (on the relevant time scale) in the range o f M after peak III develops. For low molecular weight, the Rou se model give s the area of peak II as Js – (JA + Jo) = (2M/5RT). For the pseudo-solid like behav ior, obtaining when peak III has developed, reflecting the e ffects of intermolecular entangl ement, the a rea of peak II becomes inva riant with M and given by JN – (JA + Jo) = (Me/RT). Carnegie Mellon 80 log (L( )/Pa) log (R(t)/Pa) -4 Js -6 Slope = 1/3 Narrow MWD -8 -4 Slope = 1/3 -6 3 2 1 -8 -4 -2 0 Narrow MWD 2 4 6 8 10 log (t/s) or log (/s) Carnegie Mellon 81 Peak III that deve lops as peak II area ceases to increase with increasing M, with peak III developing an a rea invariant with M, and a maximum at MAX that mov es to larger as MAX (M/Mc)3.4 for M > Mc The area under peak III, a lso invariant with M, ascribed to the effects of chain entang lements is given by 2+s Js – (JN + JA + Jo) = (kMe/ RT), where k is in the range 6-8 in most cases, and s - 2( – 1)/(3 – 2) - 0 to 1/4 with = ln RG2 /ln M Overall, = (2M/5RT)1 + ( 1+s M/kMc) 2.2 2.0 2 1.8 1.6 S Log (J ) + Cst. Js – (JA + Jo) 1.4 1.2 1.0 Carnegie Mellon 1.0 1.5 2.0• ~ Log (X) 2.5 3.0 3.5 82 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 83 Consider the following r educed exp ressions: [J(t/c) – J o]/Js = [R(t/c) – J o]/ Js + t/ Js [J(t/c) – J o]/ Js = [R(t/c) – J o]/ Js + t/c c = Js'(0) (= Js) The "time–temperature equivalence" approximation: [J(t/c) – J o]/Js is a singl e-valued function of t/c over a range o f temperature. Although rarely, i f ever, truly accurate for all temperature, it is never-the-less a useful and widely us ed app roximation for use with materials exhibit ing no phase transition over the temperature range of interest. Carnegie Mellon 84 Since c may not be kno wn over the range o f temperature of interest, it is often useful to "reduce" data to a common reference temperature TREF . Formally, this may be accomplished with [J(t) – Jo]/bTJs(TREF) = [R(t) – Jo]/bTJs(TREF) + t/bTJs(TREF) [J(t) – Jo]/bTJs(TREF) = [R(t) – Jo]/bTJs(TREF) + t/hTbTc(TREF) [J(t/aT) – Jo]/bT = [R(t/aT) – Jo]/bT + t/hTbT(TREF) [J(t/aT) – Jo]/bT = [R(t/aT) – Jo]/bT + t/aT(TREF) bT = b(T, TREF ) = Js(T)/Js(TREF ) hT = h(T, TREF ) = '(0)[T]/'(0)[TREF ] {=(T)/(TREF )} aT = bT hT Carnegie Mellon 85 Carnegie Mellon 86 bT = b(T, TREF ) = Js(T)/Js(TREF ) hT = h(T, TREF ) = '(0)[T]/'(0)[TREF ] (= (T)/(TREF )) aT = bT hT log R(t)/bT 0 T3 < T2 < T1 -2 -4 T3 T2 T1 -6 0 2 4 6 8 log t H. Markovitz J. Polym. Sci. Symp. No. 50: 431-56 (1975) Carnegie Mellon 87 Carnegie Mellon 88 4 T1 (Highest) 2 Experimental Time Range T2 log (J(t)/Pa) 0 T3 -2 T4 -4 T5 -6 T6 -8 T7 (Lowest) -10 -2 0 2 4 6 log (t/s) Carnegie Mellon 89 4 T1 (Highest) 2 Experimental Time Range T2 log (J(t)/Pa) 0 T3 -2 T4 -4 T5 -6 T6 Hypothetical example of time-temperature superposition (bT = 1). -8 T7 (Lowest) -10 -4 -2 0 2 4 6 8 10 log (t/s) or log (aT-1t/s) Carnegie Mellon 90 Carnegie Mellon 91 Carnegie Mellon 92 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 93 Specific Volume Tg Temperature A schematic v-T diagram fo r a typical noncrys talline polymeric material. Carnegie Mellon 94 A Free Volume Model: (vf)i = (v – v o)i at a certain position r i, v = (specific) volume vf = free volume vo = occupied volume Carnegie Mellon 95 Carnegie Mellon 96 The glass transition temperature Tg Tg depends on both intramolecular conformation and intermolecular interactions. Various Mod els/Treatments: Iso Free Volume: f(Tg) = constant Iso Viscous: (Tg) = constant Iso Entropic: ² S(Tg) = constant None of these are fully s atisfactory are free of arbitrary assumption s, and all contain pa rameters that can no t be independ ently evaluated. The free volume and entropic mod els provide similar expectations re the dependen ce of Tg on chain l ength and dilu ent. Carnegie Mellon 97 120 PMMA T g (°C) 100 80 60 40 0 0.2 0.4 0.6 0.8 1 Syndiotactic fraction Carnegie Mellon 98 Estimation of Tg and Tm via Group Contributions Tg - M-1Yg,i Tm - M-1Ym,i The Y x,i represent molar group contributions to the relevant property Higher order approximations a re available for both cases D. W. van Krevelen, Properties of polymers : their correlation with chemical structure, their numerical estimation and prediction from additive group contributions, 3rd Ed., Elsevier; Amsterdam ; New York, 1990. Carnegie Mellon 99 Carnegie Mellon 100 Carnegie Mellon 101 Carnegie Mellon 102 2.4 2.2 boyer kre ve len avg kre ve len cal c Tm /Tg 2.0 1.8 1.6 1.4 1.2 200 300 400 500 600 700 Tm /K D.W. Van Krevelen, op cit R. F. Boyer, Rubber Reviews 36:1303-421 Carnegie Mellon 103 Both free volume and entropic models give results that may be cast in the forms: Tg(M) - Tg (• ) {1 + kM/Mn} w 1 1 - w w + R1(1 - w) = + R Tg(w) Tg;DIL Tg(Mn) Both KM and R are model specific parameters, best evaluated experimentally. For example, in the free volume model, KM and R arise from the extra free volume provided by chain ends and diluent, respectively: typ ically, R is in the range 0.5 to 1.5. Note, that if Tg;DIL > Tg(Mn), then Tg(w, Mn) is increased by the diluent. [G. C. Berry J. Phys. Chem. 70:1194-8 (1966) ] Carnegie Mellon 104 120 100 T g(°C) Free Radical p(Syndio) ~ 0.76 80 p(Syndio) ~ 0.50 60 0 1 2 3 4 5 10 4/Mn Carnegie Mellon 105 Carnegie Mellon 106 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 107 (T) - LOC(T) F (large scale structure, T) - LOC(T) F (large scale structure) "Arrheniu s" form: LOC(T) expW/T if T > (1.5-2)Tg For melts of crystalline polym ers, Tm > (1.5-2)Tg, permitting use of this simple form. "Vogel-Fulcher" form: For amorphous polymers with 0 Š (T – Tg)/K < - 200: LOC(T) expC/(T – To) Carnegie Mellon if T < (1.5-2)Tg 108 Carnegie Mellon 109 The temperature dependence of the viscosity: (T) - LOC(T) F (large scale structure, T) - LOC(T) F (large scale structure) For amorphous polymers with 0 Š (T – Tg)/K < - 200: LOC(T) expC/(T – To) if T < (1.5-2)Tg "WLF form": LOC(T)/LOC(TREF ) = expC/(T – To) – C/(TREF – To) C(T – TREF ) = exp – REF (T – TREF + REF ) with C and To being con stants, and ² Carnegie Mellon REF = TREF – To. 110 If TREF = Tg then K (T – T g) LOC(T)/LOC(Tg) = exp– T – T + g where = Tg – To and K = C/. For many polym ers: K = 2300 K and = 57.5 K These parameters may be in terpreted in t erms of the "free-volume" model Carnegie Mellon 111 Carnegie Mellon 112 Viscosity of Polymers and Their Solutions M, c, T - LOC(T) F (M, c, T) Dilute solutions LOC(T) - Solvent(T) F (M, c, T) - 1 + []c + … [] = šN AKRG2 RH/M G. C. Berry J. Rheology 40:1129-54 (1996) Carnegie Mellon 113 F (M, c, T) - 1 + []c + … [] = šN AKRG2 RH/M Spherical Particles R = RH = (5/3)1/2RG; K = 50/9 []c = (5/2) Carnegie Mellon 114 F (M, c, T) - 1 + []c + … [] = š NAKRG2 RH/M Flexible Chain L inear Polyme rs RG2 = (âL/3)2; the cha in expansion factor â the pe rsistence length Carnegie Mellon 115 F (M, c, T) - 1 + []c + … [] = šN AKRG2 RH/M Flexible Chain Linear Polyme rs RG2 = (âL/3)2; the chain exp ansion factor â the persistence length High M: 3RH/2 - RG L1/2; K - 10/3 ML[] = šN A(20/9)(â/3)3/23L1/2 = '(â/3)3/23L1/2 Carnegie Mellon 116 F (M, c, T) - 1 + []c + … [] = šN AKRG2 RH/M Flexible Chain Linear Polyme rs RG2 = (âL/3)2; the chain exp ansion factor â the persistence length High M: 3RH/2 - RG L1/2; K - 10/3 ML[] = šN A(20/9)(â/3)3/23L1/2 = '(â/3)3/23L1/2 Low M: RH - L; ML[] = šN A(â/3)L Carnegie Mellon K - 1 (Debye) 117 Flexible Chain B ranched Polymers ML[] = š NAKRG2 RH/L g = RG2 /(RG2 )LIN; calculated = 1 High M: h = RH/(RH)LIN; h - g1/2 K - KLINf(g, shape) [] = f(g, shape)g3/2[]LIN Carnegie Mellon 118 Flexible Chain B ranched Polymers ML[] = š NAKRG2 RH/L g = RG2 /(RG2 )LIN; calculated = 1 High M: h = RH/(RH)LIN; h - g1/2 K - KLINf(g, shape) [] = f(g, shape)g3/2[]LIN Star: [] = g1/2[]LIN Comb: [] = g3/2[]LIN Rando m: [] = g[]LIN Carnegie Mellon 119 Flexible Chain Branched Polymers ML[] = šN AKRG2 RH/L g = RG2 /(RG2 )LIN; calculated = 1 High M: h = RH/(RH)LIN; h - g1/2 K - KLINf(g, shape) [] = f(g, shape)g3/2[]LIN Star: [] = g1/2[]LIN Comb: [] = g3/2[]LIN Random: [] = g[]LIN Low M: [] = šN AKRG2 RH/LML RH - L; Carnegie Mellon K - 1 120 [] = g[]LIN (c) 2 1 Entanglement Interactions log([](c)/[]) 1 0.8 Scaled screening of Intramolecular Interactions 0.6 0.4 0.2 Virial Expansion 0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 log(RG/L) Carnegie Mellon 121 Viscosity of Polymers and Their Solutions M, c, T - LOC(T) F(M, c, T) Concentrated solutions and undiluted linear flexible chain polymers LOC(T) - LOC(Tg)exp{–K(T – Tg)/(T – Tg +² )} F(M, c, T) - 1 + [](c)c Low M (Rouse behavior; = 1): ~ ~ F(M, c, T) - 1 + X - X ~ X = [](c)c; ML[](c) = a modified Fox parameter š NA(â/3)L; ([](c) ind ependen t of c in this range ) Carnegie Mellon 122 Carnegie Mellon 123 High M (Entang lement regime) ~ ~ ~ ~ ~ ~ F (M, c, T) - 1 + XE (X/Xc ) - XE (X/Xc ) ~ ~ ~ ~ E (X/Xc ) = {1 + (X/Xc )4.8}1/2 ~ Xc = šN A(â/3)Mc - 100 for many polym ers ~ Mc = Xc/šN A(â/3) - 100/šN A(â/3) Carnegie Mellon 124 Carnegie Mellon 125 The dependenc e of Tg on the diluent conc entration must be considered for polymer solutions: K (T – Tg) LOC(T)/LOC(Tg) = exp – T – Tg + where = Tg – To and K = C/². For many polyme rs: K = 2300 K and = 57.5 K ² is approximately independen t of the polym er concent ration Carnegie Mellon 126 400 Polystyrene/Dibenzyl ether Temperature/K 300 Tg 200 To 100 Tg – To 0 0 0.2 0.4 0.6 0.8 1 Volume Fraction Polymer G. C. Berry and T. G Fox Adv. Polym. Sci. 5:261-357 (1968) Carnegie Mellon 127 1.0 5 0.75 4 0.50 log( /Pa·s) 3 2 0.25 1 0.125 0 -1 -2 3 4 5 6 log( M w) Carnegie Mellon 128 Viscosity of Polymers and Their Solutions M, c, T - LOC(T) F(M, c, T) Branched Chain Polymers (Concentrated or undiluted) LOC(T) - [LOC(T)]LIN; Rare excep tions to this kno wn F(M, c, T) - 1 + [](c)c ML[](c) = š NA(â/3) gL ~ ~ ~ F(M, c, T) - 1 + XE(X/Xc ); ~ X = [](c)c ~ ~ ~ ~ E(X/Xc ) = {1 + B(g, MBR /Mc)(X/Xc )4.8}1/2 B(g, MBR /Mc) - 1 unless the branch molecular MBR > Mc ~ Xc = š NA(â/3)Mc - 100 for many polym ers Carnegie Mellon 129 6 Slope >3.4 Linear Logc ) ~~ 4 Slope 3.4 2 Branched 0 Slope 1 -2 -2 0 2 Log (w) Carnegie Mellon 130 Carnegie Mellon 131 Moderately Concentrated Solutions (c) 2 1 Entanglement Interactions log([](c)/[]) 1 0.8 Scaled screening of Intramolecular Interactions 0.6 0.4 0.2 Virial Expansion 0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 log(RG/L) Carnegie Mellon 132 Viscosity of Polymers and Their Solutions M, c, T - LOC(T) F(M, c, T) Moderately Concentrated Solutions LOC(T) - [LOC(T)]1c -=µ0LOC(T)]µc = ; µ - = c/ F(M, c, T) - 1 + [](c)c ML[](c) = š NA(â/3)(c)2(RH(c)/L)L ~ ~ ~ F(M, c, T) - 1 + H(c)XE(X/Xc ); ~ X = [](c)c ~ ~ ~ ~ E(X/Xc ) = {1 + (X/Xc )4.8}1/2 ~ Xc = š NA(â/3)Mc - 100 for many polym ers [G. C. Berry J. Rheology 40:1129-54 (1996)] Carnegie Mellon 133 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 134 Molecular Weight Polydispersity LOC(T) scales with Mn through Tg LOC(T) scales with Mw, except pe rhaps for unusual distributions Peak I in L() is essentially unaffected by molecular weight dispersion Peak II in L() may comp rise two pieces: i) an area proportional to LMzMz+1 /Mw, with the averages calculated for chains with M < Me at volume fraction L, and ii) an area proportional to (1 – L)Me for chains with M > Mc at volume fraction 1 – L Peak III in L() has an a rea proportional to (1 – L)-2(Mz/Mw)2.5 The maxima for peak s II and III separate in as (1 – L)Mw Carnegie Mellon 135 Theoretical treatments are usually ca st in terms of G(t), often in the form: G(t) = { wiGi(t) } i Gi(t) = shear modulus for chains with Mi at weight fraction wi For example: = 1 in the "reptation mode l = 1/2 in the "double-reptation" model Carnegie Mellon 136 Carnegie Mellon 137 Theoretical treatments are usually ca st in terms of G(t), often in the form: G(t) = { wiGi(t) } i Gi(t) = shear modulus for chains with Mi at weight fraction wi For example: = 1 in the "reptation mode l = 1/2 in the "double-reptation" model The effects of increased dispersity of molecular species is usually most prominent in Peak III in L(), followed by effects in P eak II in L(). This is seen in L() for a polym er undergoing crosslinking to form a branched polymer, leading to a network polymer Carnegie Mellon 138 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 139 log (L()/Pa) -2 Cross link prior to gelation -4 -6 Initial -8 -4 -2 0 2 4 6 8 10 log (t/s) or log (/s) Carnegie Mellon 140 Incipient gelation log (L()/Pa) -2 ? Cross link prior to gelation -4 -6 Initial -8 -4 -2 0 2 4 6 8 10 log (t/s) or log (/s) Carnegie Mellon 141 Incipient gelation log (L()/Pa) -2 ? Cross link prior to gelation -4 -6 "Weak" gel "Gel” Initial -8 -4 -2 0 2 4 6 8 10 log (t/s) or log (/s) Carnegie Mellon 142 Carnegie Mellon 143 Carnegie Mellon 144 Carnegie Mellon 145 Power-law behavior G(t) = [Go – Ge](t) + Ge J(t) = Jo + (t) + t/ (t) = (Js – Jo)[1 – (t)] Suppose that for all t (note, this involves permissible, but peculiar behavior for large t): (t) = (t/) With this expression, and 1/ = 0: [J'() – Jo]/Jo = µ(µ)cos(µš /2) ( )-µ J"()/Jo = µ(µ)sin(µ š /2) ( )-µ Use of the convolution integral relating J(t) and G(t) gives (t) = Eµ(-kµt/)µ) with Ge = 0 and 1/ = 0, where kµ = µ(µ) and • Carnegie Mellon Eµ(x) = (nµx + 1) : n=0 n The Mittag-Leffler function 146 For small µ, G(t) - Go{1 + t/)µ} For any µ, for large t/ G(t) Gosin(µ š) /µš t/)µ G(t)J(t) sin(µ š) /µš < 1 (G'() – Ge)/(Go – Ge) (-) sin[(-)/2] ( ) G"()/(Go – Ge) (-) cos[(-)/2] ( ) [J'() – Jo]/Jo = µ(µ)cos(µš /2) ( )-µ J"()/Jo = µ(µ)sin(µ š /2) ( )-µ Carnegie Mellon 147 Bounded power-law behavior for (t) migh t be obtained in th e form (t) = 1; for t Š = (/t); for < t Š , with 0 < µ < 1 = (q/t)m; for t > , with m > 1 where q = (/)m.Then, G'() – Ge and G"() for << 1/; G'() = Go and G"() = 0 for >> 1/; (G'() – Ge)/(Go – Ge) (-) sin[(-)/2] () G"()/(Go – Ge) (-) cos[(-)/2] () for the interval 1/ < < 1/. Carnegie Mellon 148 An alternativ e relation th at also exhibits partial power-law behavior is given by: (t) n/m n/m = (/i) exp(–t/i)/ (/i) i = i = where i = /im; m = 2 and n = 0 in the Rouse model. For the intermediate int erval 1/ < < 1/, (G'() – Ge)/(Go – Ge) {/2m sin[(-)/2]} () G"()/(Go – Ge) {/2m sin[(-)/2]} () CarnegieµMellon where = (1 + n)/m (µ =, for the Rouse model). 149 0 0 -2 -2 N = 1000 µ = n/(1 + m) = 1/2 N = 300 µ = n/(1 + m) -4 µ = 1/3 m = 6; n = 1 -4 -6 0 log G'() or log G''() -6 m = 4; n = 1 0 -2 -2 -4 -4 -6 0 µ = 1/2 m = 4; n = 1 m = 3; n = 0.5 0 -2 -2 -4 -4 m= 2; n = 0 -6 0 m = 1; n = 0.5 -2 0 -2 µ = 2/3 m =3; n = 1 -4 -4 0 2 4 6 log( c ) Carnegie Mellon 8 µ =1 m = 2; n = 1 -6 -8 0 2 4 6 log( c ) 8 150 Carnegie Mellon 151 Carnegie Mellon 152 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 153 ISOCHRONAL BEHAVIOR • In some cases, the temperature is scanned while the dynamic properties are determined at fixed frequency; such experiments might typically be reported as G'(;T) and tan (; T) or '(;T) versus T, depending on the application. • Insofar as G'(c(T)) and G"(c(T)) as functions of c(T) are independent of T, the isochronal plots are seen to be mappings in which c(T) increases with decreasing temperature with: K c(T) expT - (Tg - ) • For a reference temperature equal to the glass temperature Tg, so that a = c(T)/c(Tg): K T - Tg ln a = ln – 1 + (T - Tg) k + k(T - Tg) + … with the linear approximation valid for (T - Tg) << ; k = ln and Carnegie Mellon k= K /. 154 1 Log G' /Go T - Tg = 0 0 Log tan Log G' /G o and Log tan -1 -2 -3 -4 -2 0 2 4 log (a T ) 1 =1s 0 -1 Log tan -1 Log G' /Go -2 -3 Carnegie Mellon -10 10 0 T - Tg 20 155 Carnegie Mellon 156 "Iso-chronal" behavior for Poly(v inyl C hloride); (No rotational is omers for the side group) 10 Poly(vinyl chloride) log G'/Pa (Fixed ) 8 tan = E"/E' Tg Main-chain rotation 0.1 0.01 -200 -100 0 100 Temperature (°C) Carnegie Mellon 157 Carnegie Mellon 158 Carnegie Mellon 159 Carnegie Mellon Int roduction Rheological methods Linear elastic parameters Linear visc oelastic functions Several viscoelastic experiments Relations among linear visc oelastic functions Examples of linear visc oelastic functions Time-temperature equivalence (Thermo-rheological simplicity) The glass transition temperature The visc osity Effects of polydispersity Network formation Isochronal B ehavior Examples from the literature 160 Examples from the literature Carnegie Mellon Branched and li near metallocene polyolefins Colloidal d ispersions Wormlike Micelles Deformation of rigid materials Nonlinear shear behavior Linear and nonlinear bulk properties 161 5 4 log G''() 3 2 log G'() 1 Unmodified Linear 0 -1 0 1 2 log Metallocene polyethylenes Claus Gabriel and Helmut Münstedt Rheol. Acta 38: 393-403 (1999) Carnegie Mellon 162 Carnegie Mellon Metallocene polyethylenes Claus Gabriel and Helmut Münstedt Rheol. Acta 38: 393-403 (1999) 163 Carnegie Mellon Metallocene polyethylenes Claus Gabriel and Helmut Münstedt Rheol. Acta 38: 393-403 (1999) 164 Carnegie Mellon Metallocene polyethylenes Claus Gabriel and Helmut Münstedt Rheol. Acta 38: 393-403 (1999) 165 Carnegie Mellon Metallocene polyethylenes Claus Gabriel and Helmut Münstedt Rheol. Acta 38: 393-403 (1999) 166 0 log '( '( -1 -4 log J'()/b Unmodified Linear -5 Modified Branched 1 Modified Branched 2 2 3 4 5 6 log '(0) b U Carnegie Mellon M1 M2 log '(0) 3.28 3.68 4.00 log b -0.7 0 0 167 log J(t) or log R(t). Pa-1 0 mLDPE-Linear J(t) mLDPE-Branched LDPE-Branched -2 R(t) -4 -6 -2 0 2 4 log t/s C. Gabriel and H. Münstedt Rheo. Acta, 38:393-403 (1 999) Carnegie Mellon 168 From creep/recovery Carnegie Mellon 169 log '()/'(0) 0 mLDPE-Linear mLDPE-Branched -1 LDPE-Branched 0 2 4 6 log '(0) C. Gabriel and H. Münstedt Rheo. Acta, 38:393-403 (1 999) Carnegie Mellon 170 Carnegie Mellon 171 Examples from the literature Carnegie Mellon Branched and li near metallocene polyolefins Colloidal d ispersions Wormlike Micelles Deformation of rigid materials Nonlinear shear behavior Linear and nonlinear bulk properties 172 Colloidal dispersions: Linear and nonlinear viscoelastic behavior. Dilute dispersion of spheres interacting via a hard-core potential: LOC{1 + (5/2) + k'(5/2) + …} 2 Carnegie Mellon 2 = volume fraction = c/ (5/2) = c LOC - solv. k' - 1.0 173 Concentrated dispersion of hard-core spheres: Empirical relations: - LOC{1 – /n –5n /2 - LOC{1 – (5/2)1 – /n –5n k'/2 1 1 2 2 designed to force agreement with the v iri al expans ion at least to order and , respectively, n1 = 5/8 to give k' - 1.0 n1 = max - 0.64 Theoretical relations: = LOC{1 + (5/2) + k'1() + 2()(5/2)22} 1(): hyd rodyna mi cs 2(): thermodyna mi cs 1() + 2() =1 U: Carnegie Mellon 1() - (4/5)(1 – /max) 2() - (1/5)(1 – /max) (semi- empircial) 2 174 Carnegie Mellon 175 Concentrated dispersion of hard-core spheres: Linear Viscoelastic Response: '() = '(0) for small , as exp ected, but als o show a plateau '() - '(L) for a regim e at an in termediate range of - L, before decreasing to zero wit h increasing . '(L) is estima ted wit h () = 0, reflecting the suppression o f thermodyna mi c interactions at high G'(L) - G1; G1R3/kT2 - 0() for spheres of radius R 0() - 0.78('(L)/solv)g(2 , ) g(2 , ) is the radia l distribution at the contact cond iti on r/R = 2 Theory : g(2, ) = (1 – /2)2/(1 – )3 for < 0.5 and g(2, ) = (6/5)(1 – /max) for •0.5 Carnegie Mellon 176 Carnegie Mellon 177 Concentrated dispersion of hard-core spheres: Linear Viscoelastic Response: Theory: '() = LOC{1 + (5/2) + k'1() + 2()(5/2)22} '(L) = LOC{1 + (5/2) + k'1()(5/2)22} J'EFF() -1/2 for a rang e of < L J'EFF(L) - 1/G'(L) - 1/G1 - R3/kT20() Carnegie Mellon 178 Carnegie Mellon 179 Concentrated dispersion of interacting spheres: Van der Waals interactions Electrostatic interactions among charged spheres Interactions among spheres and a dissolved polymer True or apparent yield behavior may obtain Carnegie Mellon 180 Carnegie Mellon 170 nm beads (0.05 to 0.2 volume fraction), in 15% polystyrene solution D. Meitz, L. Yen, G. C. Berry and H. Markovitz J. Rheol. 32:309-51 (1988) 181 Examples from the literature Carnegie Mellon Branched and li near metallocene polyolefins Colloidal d ispersions Wormlike Micelles Deformation of rigid materials Nonlinear shear behavior Linear and nonlinear bulk properties 182 Wormlike micelles Certain amphillic molecules organize to form curvilinear cylinders, or wormlike micelles. For example, in an aqueous medium, the amphiphile might organize with its hydrophobic parts aggregated in the interior of the cylinder, and its hydrophopic pieces arranged on the "surface" of the cylinder The micelle structure will exhibit a lifetime ruptu re for rupture of its components If ruptu re is less than a longest rheological time constant rheol the intact wormlike micelle would exhibit, then the rupture dynamics may dominate the observed rheological behavior, The chain may respond to a deformation by micellar dynamics similar to those for a structure without rupture, abetted by the rupture process. With one model, this approximates Maxwell behavior with a time constant eff ective - ruptu reruptu re Carnegie Mellon 183 Cetyl triethylammonium tosylate -T CTA+ hydrophobic – + hydrophilic + + + + + ++ - - + + + + + + + + + + + + + + micelles grow 10 nm micellar network Schematic courtesy Dr. Lynn M. Walker Carnegie Mellon 184 In an extreme case, the system might approximate behavior for the Maxwell model, with a single relaxation time eff ectiveso that J(t) = Js + t/; with Js = eff ective G(t) = (1/Js)exp(-t/eff ective) With this simple model, J'() = Js 2 '() = (1/Js)/[1 + (eff ective) ] Carnegie Mellon 185 0 10% -1 '/p J'()/Jp T (°C) 30 35 -2 ◊ The rate of decrease of '() with increasing for larger , to the extent of an increase in '() with increasing for the data on the less concentrated Sample 40 p -3 3 Calculated 2 ◊ The increase of J'() above the imputed Js for smaller for the data on the more concentrated sample p 1 -1 -2 20% s log J /Pa 0 -3 solvent 4. 5 -2 log log '/ or log J'()/J These data reveal several deviations from simple Maxwell behavior, including: 4. 0 10% 3. 5 20% -3 30 -4 -3 35 T emperature (°C) -2 40 0 -1 1 2 ◊ It may be likely that these samples exhibit solid-like behavior with a Je at smaller than the experimental range, and that Jp is truly Js ◊ The relatively constant J'() is expected with the Maxwell model, but this may be fortuitous log p J p Carnegie Mellon J. F. A. Soltero and J. E. Puig Langmuir 12: 141-8 (1996) 186 Examples from the literature Carnegie Mellon Branched and li near metallocene polyolefins Colloidal d ispersions Wormlike Micelles Deformation of rigid materials Nonlinear shear behavior Linear and nonlinear bulk properties 187 Deformation of Rigid Materials Creep and Recovery in Tension Creep for 0 Š t Š Te (t) = oD(t) = o[DR(t) + DNR(t)] Recovery for = t – T e > 0 (, Te) = o[DR( + Te) – DR() + DNR(Te)] R(, Te) = (Te) – (, Te) = o{DR(Te) – DR( + Te) + DR()} Carnegie Mellon 188 G. C. Berry J. Polym. Sci.: Polym. Phys. Ed. 14:451-78 (1976) Carnegie Mellon 189 Andrade Creep (with DNR(t) = 0) A frequently ob served nonlin ear behavio r DR(t, o) = DA{1 + R(o)t1/3} oR 106 (sec 1/3) sinh(o/A) R(o) - R() ; A a constant o/A 30 299°C 20 231 50 10 34.5 0 0 20 40 60 80 100 /Mdyn/cm 2 Carnegie Mellon 190 Andrade Creep (with DNR(t) ° 0) A nonrecoverable logarithmic creep is frequently obs erved under larger stress: DNR(t) - DL ln(1 + µt/DL) (a) µt/DL <<1 µt (b) D(t)/MPa -1 3 2 1 0 0 5 10 (t/sec)1/3 Carnegie Mellon 15 20 0 5 10 15 20 25 (/sec)1/3 or [ + T )/sec] 1/3– ( /sec)1/3 191 Examples from the literature Carnegie Mellon Branched and li near metallocene polyolefins Colloidal d ispersions Wormlike Micelles Deformation of rigid materials Nonlinear shear behavior Linear and nonlinear bulk properties 192 An "Incompressible" Isotropic Elastic Material Suppose K >> G, then for infinitisimal strains Sij = 2 G {ij – 3ij } – ij P More generally, for finite strains: -1 Sij = W1 Bij + W2Bij – ij P Wi = Wi (I B;1, IB;2) – ŽW ŽIB;i For simple extension: f/A - 2(2 – -1)(W1 + W2/) For simple shear: S12 = 2(W1 + W2) G Carnegie Mellon S11 – S33 = 2W1 2 ; S22 – S33 = – 2W2 2 193 An expansion of the strain energy function gives the Mooney–Rivlin Equation for small deformations: W - C1 (IB;1 – 3) + C2 (IB;2 – 3) W1 = C1 and W2 = C2 For the original Kinetic Theory of Rubber Elasticity the contributions to C1 are entropic in origin, and.: 2C1 = EkT = RT/MXL 2C2 = 0 stress chains E = Number of chains under MXL = Molecular weight of between crosslinks The preceding estimates for C1 and C2 are not accurate, and have been modified in more modern Carnegie Mellon theories, e.g., these give C 2 > 0. 194 An "Incompressible" Viscoelastic Material Suppose K(t) >> G(t), then for infinitisimal strains t Sij(t) = 2 G(t – -• ij(s) s) s – ij (s) s ds – ijP Several relations are proposed for finite strains, including that due to Bernstein, Kearsley and Zapas:: t Sij(t) = U 2 I B;1 -• Carnegie Mellon U -1 B(t)ij(s) – I B(t)ij(s) ds – ijP B;2 195 An "Incompressible" Viscoelastic Material Suppose K(t) >> G(t), then for infinitisimal strains t Sij(t) = 2 G(t – -• ij(s) s) s – ij (s) s ds – ijP Several relations are proposed for finite strains, including that due to Bernstein, Kearsley and Zapas:: t Sij(t) = U 2 I B;1 -• Carnegie Mellon U -1 B(t)ij(s) – I B(t)ij(s) ds – ijP B;2 196 Nonlinear Response in Simple Shear for a Fluid (In the appro ximation with t >> R) Shear Stress (t) = S12(t): • G(u) (t) = – [(t,u)] F1[(t,u)] u du t (u) (t) = G(t – u) u M1[(t,u)] du -• (t,u) = (t) – (u) Carnegie Mellon F1( M1[(t,u)] = n F 1() = F1()1 + n 197 Nonlinear Response in Simple Shear for a Fluid (In the appro ximation with t >> R) First–Normal Stress Difference (1)(t) = 11(t) – 22(t) : • G(u) (t) = – [(t,u)] F1[(t,u)] u du t (u) (t) = G(t – u) u M2[(t,u)] du -• F1( M2[(t,u)] = Carnegie Mellon n F 1() = F1()2 + n 198 A Theoretical Expression for the Strain Function: The theory du e to Doi and Edwards F1() = [1 + (||/'')]; '' - 2.13 An Approximate form of the Strain Function: F1 () = 1 for || Š ' F1 () = exp[ – (||– ')/''] for || > ' log F1 ( ) 0 '/ '' [1 + (||/'')] -1 -1 exp(-| - '|/") -2 1 Carnegie Mellon 2 | 3 | / '' 4 199 Response to a Step Shear Strain Strain Jump: (t) t = 0+ ° = t ° G(t-s)(0) (t) = F1()1 °G(t) F1(°) = S12 (t, )/ ° n F1() + n ds (t) 0 -1 0.01 2.5 5 R -2 c -3 Carnegie Mellon -4 -2 0 log t/sec 2 200 Response to a Ramp Deformation (t) = · t t> Stress Growth: (t) = (t) = · G(s) n F1(· s) ds F1(· s)1 + n · s t ·2 sG(s) t F1(· s)2 + Carnegie Mellon · n F1( s) ds n · s 201 Steady-State Flow Viscosity · lim (t) = SS( t >> c · = SS(/ · · ( · = (0) = lim ( = · = (0) H · /'' ( c • G(u)M [ u]du 1 0 · Hc· /'' = Carnegie Mellon • G(u)du 0 202 Steady-State Flow First-Normal Stress Difference · lim (t) = () t >> c SS · = ()/2{ · · N() ()} SS SS · = Js lim N() = · = Js S /'' · N() c N • uG(u)M [ u]du 2 0 · · S c/'' = N Carnegie Mellon • uG(u)du 0 • -2 · 0 G(u)M1[ u]du • G(u)du 0 203 Steady-State Flow Steady-State Recoverable Compliance · lim (t,) = () R t; >>c R · = ()/ · () · R () SS R SS · = J lim R () s SS = · = J S /'' · R () s R c SS · SRc/'' = Result of an iterative calculation involving G(t) and F1() Carnegie Mellon 204 Suppose G(t) = Go•iexp(–t/i); •i = 1 Then, with the app roximate F1() given above (· ) = Go•i i H(·i/'') H(·i/'') - 1 ; · [1 + (i/ '')] - 6/5, - 1 By comparison, 1 '() =Go•i i [1 + (i)] In bo th cases, the factors i i in the terms in the su mmation are weighted by functions that decrease term–by–t erm with increasing ·or . Consequently, th ese expressions exhib it the Cox-Merz approximation: Carnegie Mellon (·) - '· 205 Narrow MWD Broad MWD -1 . . log H(); log SN() 0 -2 -2 -1 0 log Carnegie Mellon . 1 2 3 c 206 -1 s · 0 s log[J )/J ] · · s -1 0 (1) log[ )/ )] log[S )/J ] 0 -1 -3 -2 -1 0 1 2 3 · log( ) c Polyethylene K. Nakamura, C.-P. Wong and G. C. Berry J. Poly m. Sci: Polym. Phys. Ed . 22:1119 -48 (1984) Carnegie Mellon 207 -1 -1 s · 0 (1) 0 -2 -1 -2 log[S )/J ] s log[J'( )/J ] -2 0 · log[ ()/ )] log[ '( )/ )] 0 -1 -1 0 1 2 3 · log( ) c Linear and nonlinear behavior for a polymer with a relatively narrow MWD Carnegie Mellon 208 Examples from the literature Carnegie Mellon Branched and li near metallocene polyolefins Colloidal d ispersions Wormlike Micelles Deformation of rigid materials Nonlinear shear behavior Linear and nonlinear bulk properties 209 Carnegie Mellon 210 Carnegie Mellon 211 An Inherent Nonlinearity in Response B(t) = B() + B(t) ^ (t) = (t/) But = (V,T) An attempt to account for this effect makes use of an material time constant averaged over the time interval of interest: 1 (t ,t) = (t - t ) t2 t1 (u) du V(t) – V() P(s) t = B[(t – s) (t ,s) -• V() s ds Frequently, Carnegie Mellon B(t) = BA{1 + (t/A)1/3}; t < 212 Carnegie Mellon 213 Carnegie Mellon 214 Carnegie Mellon 215