Mooney–Rivlin solid

Continuum mechanics
Laws Conservation of mass Conservation of momentum Conservation of energy Entropy inequality
Solid mechanics Solids Stress · Deformation Compatibility Finite strain · Infinitesimal strain Elasticity (linear) · Plasticity Bending · Hooke's law Failure theory Fracture mechanics Frictionless/Frictional Contact mechanics
Fluid mechanics Fluids Fluid statics · Fluid dynamics Surface tension Navier–Stokes equations Viscosity: Newtonian, Non-Newtonian
Rheology Viscoelasticity Smart fluids: Magnetorheological Electrorheological Ferrofluids Rheometry · Rheometer
Scientists Bernoulli · Cauchy · Hooke Navier · Newton · Stokes · Jamal Nazrul Islam · Fazle Hussain
v · d · e

In continuum mechanics, a Mooney–Rivlin solid^[1][2] is a hyperelastic material model where the strain energy density function $W\,$ is a linear combination of two invariants of the left Cauchy–Green deformation tensor $\boldsymbol{B}$. The model was proposed by Melvin Mooney in 1940 and expressed in terms of invariants by Ronald Rivlin in 1948.

The strain energy density function for an incompressible Mooney–Rivlin material is^[3][4]

$W = C_{1} (\overline{I}_1-3) + C_{2} (\overline{I}_2-3), \,$

where C₁ and C₂ are empirically determined material constants, and $\overline{I}_1$ and $\overline{I}_2$ are the first and the second invariant of the unimodular component of the left Cauchy–Green deformation tensor:^[5]

$\begin{align} \bar{I}_1 & = J^{-2/3}~I_1 ~;~~ I_1 = \lambda_1^2 + \lambda_2 ^2+ \lambda_3 ^2 ~;~~ J = \det(\boldsymbol{F}) \\ \bar{I}_2 & = J^{-4/3}~I_2 ~;~~ I_2 = \lambda_1^2 \lambda_2^2 + \lambda_2^2 \lambda_3^2 + \lambda_3^2 \lambda_1^2 \end{align}$

where $\boldsymbol{F}$ is the deformation gradient. For an incompressible material, J = 1.

Derivation

The Mooney–Rivlin model is a special case of the generalized Rivlin model (also called polynomial hyperelastic model^[6]) which has the form

$W = \sum_{p,q = 0}^N C_{pq} (\bar{I}_1 - 3)^p~(\bar{I}_2 - 3)^q + \sum_{m = 1}^M D_m~(J-1)^{2m}$

with C₀₀ = 0 where C_pq are material constants related to the distortional response and D_m are material constants related to the volumetric response. For a compressible Mooney–Rivlin material N = 1,C₀₁ = C₂,C₁₁ = 0,C₁₀ = C₁,M = 1 and we have

$W = C_{01}~(\bar{I}_2 - 3) + C_{10}~(\bar{I}_1 - 3) + D_1~(J-1)^2$

If C₀₁ = 0 we obtain a neo-Hookean solid, a special case of a Mooney–Rivlin solid.

For consistency with linear elasticity in the limit of small strains, it is necessary that

$\kappa = 2D_1 ~;~~ \mu = 2~(C_{01} + C_{10})$

where κ is the bulk modulus and μ is the shear modulus.

Cauchy stress in terms of strain invariants and deformation tensors

The Cauchy stress in a compressible hyperelastic material with a stress free reference configuration is given by

$\boldsymbol{\sigma} = \cfrac{2}{J}\left[\cfrac{1}{J^{2/3}}\left(\cfrac{\partial{W}}{\partial \bar{I}_1} + \bar{I}_1~\cfrac{\partial{W}}{\partial \bar{I}_2}\right)\boldsymbol{B} - \cfrac{1}{J^{4/3}}~\cfrac{\partial{W}}{\partial \bar{I}_2}~\boldsymbol{B} \cdot\boldsymbol{B} \right] + \left[\cfrac{\partial{W}}{\partial J} - \cfrac{2}{3J}\left(\bar{I}_1~\cfrac{\partial{W}}{\partial \bar{I}_1} + 2~\bar{I}_2~\cfrac{\partial{W}}{\partial \bar{I}_2}\right)\right]~\boldsymbol{\mathit{1}}$

For a compressible Mooney–Rivlin material,

$\cfrac{\partial{W}}{\partial \bar{I}_1} = C_1 ~;~~ \cfrac{\partial{W}}{\partial \bar{I}_2} = C_2 ~;~~ \cfrac{\partial{W}}{\partial J} = 2D_1(J-1)$

Therefore, the Cauchy stress in a compressible Mooney–Rivlin material is given by

$\boldsymbol{\sigma} = \cfrac{2}{J}\left[\cfrac{1}{J^{2/3}}\left(C_1 + \bar{I}_1~C_2\right)\boldsymbol{B} - \cfrac{1}{J^{4/3}}~C_2~\boldsymbol{B} \cdot\boldsymbol{B} \right] + \left[2D_1(J-1)- \cfrac{2}{3J}\left(C_1\bar{I}_1 + 2C_2\bar{I}_2~\right)\right]\boldsymbol{\mathit{1}}$

The quantity

$p := - 2D_1~J(J-1) + \cfrac{2}{3}\left(C_1~\bar{I}_1 + 2~C_2~\bar{I}_2~\right)$

has the form of a pressure. The stress can then be expressed in the form

$\boldsymbol{\sigma} = \cfrac{1}{J}\left[-p~\boldsymbol{\mathit{1}} + \cfrac{2}{J^{2/3}}\left(C_1 + \bar{I}_1~C_2\right)\boldsymbol{B} - \cfrac{2}{J^{4/3}}~C_2~\boldsymbol{B}\cdot\boldsymbol{B} \right]$

For an incompressible Mooney–Rivlin material with J = 1

$\boldsymbol{\sigma} = -p~\boldsymbol{\mathit{1}} + 2\left(C_1 + I_1~C_2\right)\boldsymbol{B} - 2C_2~\boldsymbol{B}\cdot\boldsymbol{B}$

Note that if J = 1 then

$\det(\boldsymbol{B}) = \det(\boldsymbol{F})\det(\boldsymbol{F}^T) = 1$

Then, from the Cayley-Hamilton theorem,

$\boldsymbol{B}^{-1} = \boldsymbol{B}\cdot\boldsymbol{B} - I_1~\boldsymbol{B} + I_2~\boldsymbol{\mathit{1}}$

Hence, the Cauchy stress can be expressed as

$\boldsymbol{\sigma} = -p^{*}~\boldsymbol{\mathit{1}} + 2 C_1~\boldsymbol{B} - 2C_2~\boldsymbol{B}^{-1}$

where $p^{*} := \tfrac{2}{3}(C_1~I_1 - C_2~I_2). \,$

Cauchy stress in terms of principal stretches

In terms of the principal stretches, the Cauchy stress differences for an incompressible hyperelastic material are given by

$\sigma_{11} - \sigma_{33} = \lambda_1~\cfrac{\partial{W}}{\partial \lambda_1} - \lambda_3~\cfrac{\partial{W}}{\partial \lambda_3} ~;~~ \sigma_{22} - \sigma_{33} = \lambda_2~\cfrac{\partial{W}}{\partial \lambda_2} - \lambda_3~\cfrac{\partial{W}}{\partial \lambda_3}$

For an incompressible Mooney-Rivlin material,

$W = C_1(\lambda_1^2 + \lambda_2 ^2+ \lambda_3 ^2 -3) + C_2(\lambda_1^2 \lambda_2^2 + \lambda_2^2 \lambda_3^2 + \lambda_3^2 \lambda_1^2 -3) ~;~~ \lambda_1\lambda_2\lambda_3 = 1$

Therefore,

$\lambda_1\cfrac{\partial{W}}{\partial \lambda_1} = 2C_1\lambda_1^2 + 2C_2\lambda_1^2(\lambda_2^2+\lambda_3^2) ~;~~ \lambda_2\cfrac{\partial{W}}{\partial \lambda_2} = 2C_1\lambda_2^2 + 2C_2\lambda_2^2(\lambda_1^2+\lambda_3^2) ~;~~ \lambda_3\cfrac{\partial{W}}{\partial \lambda_3} = 2C_1\lambda_3^2 + 2C_2\lambda_3^2(\lambda_1^2+\lambda_2^2)$

Since λ₁λ₂λ₃ = 1. we can write

$\begin{align} \lambda_1\cfrac{\partial{W}}{\partial \lambda_1} & = 2C_1\lambda_1^2 + 2C_2\left(\cfrac{1}{\lambda_3^2}+\cfrac{1}{\lambda_2^2}\right) ~;~~ \lambda_2\cfrac{\partial{W}}{\partial \lambda_2} = 2C_1\lambda_2^2 + 2C_2\left(\cfrac{1}{\lambda_3^2}+\cfrac{1}{\lambda_1^2}\right) \\ \lambda_3\cfrac{\partial{W}}{\partial \lambda_3} & = 2C_1\lambda_3^2 + 2C_2\left(\cfrac{1}{\lambda_2^2}+\cfrac{1}{\lambda_1^2}\right) \end{align}$

Then the expressions for the Cauchy stress differences become

$\sigma_{11}-\sigma_{33} = 2C_1(\lambda_1^2-\lambda_3^2) - 2C_2\left(\cfrac{1}{\lambda_1^2}-\cfrac{1}{\lambda_3^2}\right)~;~~ \sigma_{22}-\sigma_{33} = 2C_1(\lambda_2^2-\lambda_3^2) - 2C_2\left(\cfrac{1}{\lambda_2^2}-\cfrac{1}{\lambda_3^2}\right)$

Uniaxial extension

For the case of an incompressible Mooney–Rivlin material under uniaxial elongation, $\lambda_1 = \lambda\,$ and $\lambda_2 = \lambda_3 = 1/\sqrt{\lambda}$. Then the true stress (Cauchy stress) differences can be calculated as:

$\begin{align} \sigma_{11}-\sigma_{33} & = 2C_1\left(\lambda^2-\cfrac{1}{\lambda}\right) -2C_2\left(\cfrac{1}{\lambda^2} - \lambda\right)\\ \sigma_{22}-\sigma_{33} & = 0 \end{align}$

Simple tension

Comparison of experimental results (dots) and predictions for Hooke's law(1, blue line), neo-Hookean solid(2, red line) and Mooney–Rivlin solid models(3, green line)

In the case of simple tension, σ₂₂ = σ₃₃ = 0. Then we can write

$\sigma_{11} = \left(2C_1 + \cfrac {2C_2} {\lambda} \right) \left( \lambda^2 - \cfrac{1}{\lambda} \right)$

In alternative notation, where the Cauchy stress is written as $\boldsymbol{T}$ and the stretch as α, we can write

$T_{11} = \left(2C_1 + \frac {2C_2} {\alpha} \right) \left( \alpha^2 - \alpha^{-1} \right)$

and the engineering stress (force per unit reference area) for an incompressible Mooney–Rivlin material under simple tension can be calculated using $T_{11}^{\mathrm{eng}} = T_{11}\alpha_2\alpha_3 = \cfrac{T_{11}}{\alpha}$. Hence

$T_{11}^{\mathrm{eng}}= \left(2C_1 + \frac {2C_2} {\alpha} \right) \left( \alpha - \alpha^{-2} \right)$

If we define

$T^{*}_{11} := \cfrac{T_{11}^{\mathrm{eng}}}{\alpha - \alpha^{-2}} ~;~~ \beta := \cfrac{1}{\alpha}$

then

$T^{*}_{11} = 2C_1 + 2C_2\beta ~.$

The slope of the $T^{*}_{11}$ versus β line gives the value of C₂ while the intercept with the $T^{*}_{11}$ axis gives the value of C₁. The Mooney–Rivlin solid model usually fits experimental data better than Neo-Hookean solid does, but requires an additional empirical constant.

Equibiaxial tension

In the case of equibiaxial tension, the principal stretches are λ₁ = λ₂ = λ. If, in addition, the material is incompressible then λ₃ = 1 / λ². The Cauchy stress differences may therefore be expressed as

$\sigma_{11}-\sigma_{33} = \sigma_{22}-\sigma_{33} = 2C_1\left(\lambda^2-\cfrac{1}{\lambda^4}\right) - 2C_2\left(\cfrac{1}{\lambda^2} - \lambda^4\right)$

The equations for equibiaxial tension are equivalent to those governing uniaxial compression.

Pure shear

A pure shear deformation can be achieved by applying stretches of the form ^[7]

$\lambda_1 = \lambda ~;~~ \lambda_2 = \cfrac{1}{\lambda} ~;~~ \lambda_3 = 1$

The Cauchy stress differences for pure shear may therefore be expressed as

$\sigma_{11} - \sigma_{33} = 2C_1(\lambda^2-1) - 2C_2\left(\cfrac{1}{\lambda^2}-1\right) ~;~~ \sigma_{22} - \sigma_{33} = 2C_1\left(\cfrac{1}{\lambda^2} -1\right) - 2C_2(\lambda^2 -1)$

Therefore

$\sigma_{11} - \sigma_{22} = 2(C_1+C_2)\left(\lambda^2 - \cfrac{1}{\lambda^2}\right)$

For a pure shear deformation

$I_1 = \lambda_1^2 + \lambda_2^2 + \lambda_3^2 = \lambda^2 + \cfrac{1}{\lambda^2} + 1 ~;~~ I_2 = \cfrac{1}{\lambda_1^2} + \cfrac{1}{\lambda_2^2} + \cfrac{1}{\lambda_3^2} = \cfrac{1}{\lambda^2} + \lambda^2 + 1$

Therefore I₁ = I₂.

Simple shear

The deformation gradient for a simple shear deformation has the form^[7]

$\boldsymbol{F} = \boldsymbol{1} + \gamma~\mathbf{e}_1\otimes\mathbf{e}_2$

where $\mathbf{e}_1,\mathbf{e}_2$ are reference orthonormal basis vectors in the plane of deformation and the shear deformation is given by

$\gamma = \lambda - \cfrac{1}{\lambda} ~;~~ \lambda_1 = \lambda ~;~~ \lambda_2 = \cfrac{1}{\lambda} ~;~~ \lambda_3 = 1$

In matrix form, the deformation gradient and the left Cauchy-Green deformation tensor may then be expressed as

$\boldsymbol{F} = \begin{bmatrix} 1 & \gamma & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 1 \end{bmatrix} ~;~~ \boldsymbol{B} = \boldsymbol{F}\cdot\boldsymbol{F}^T = \begin{bmatrix} 1+\gamma^2 & \gamma & 0 \\ \gamma & 1 & 0 \\ 0 & 0 & 1 \end{bmatrix}$

Therefore,

$\boldsymbol{B}^{-1} = \begin{bmatrix} 1 & -\gamma & 0 \\ -\gamma & 1+\gamma^2 & 0 \\ 0 & 0 & 1 \end{bmatrix}$

The Cauchy stress is given by

$\boldsymbol{\sigma} = \begin{bmatrix} -p^* +2(C_1-C_2)+2C_1\gamma^2 & 2(C_1+C_2)\gamma & 0 \\ 2(C_1+C_2)\gamma & -p^* + 2(C_1 -C_2) - 2C_2\gamma^2 & 0 \\ 0 & 0 & -p^* + 2(C_1 - C_2) \end{bmatrix}$

For consistency with linear elasticity, clearly μ = 2(C₁ + C₂) where μ is the shear modulus.

Rubber

Elastic response of rubber-like materials are often modeled based on the Mooney—Rivlin model. The constants C₁,C₂ are determined by the fitting predicted stress from the above equations to experimental data. The recommended tests are uniaxial tension, equibiaxial compression, equibiaxial tension, uniaxial compression, and for shear, planar tension and planar compression. The two parameter Mooney–Rivlin model is usually valid for strains less than 100%.

Notes and references

1. ^ Mooney, M., 1940, A theory of large elastic deformation, Journal of Applied Physics, 11(9), pp. 582-592.
2. ^ Rivlin, R. S., 1948, Large elastic deformations of isotropic materials. IV. Further developments of the general theory, Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, 241(835), pp. 379-397.
3. ^ Boulanger, P. and Hayes, M. A., 2001, Finite amplitude waves in Mooney–Rivlin and Hadamard materials, in Topics in Finite Elasticity, ed. M. A Hayes and G. Soccomandi, International Center for Mechanical Sciences.
4. ^ C. W. Macosko, 1994, Rheology: principles, measurement and applications, VCH Publishers, ISBN 1-56081-579-5.
5. ^ The characteristic polynomial of the linear operator corresponding to the second rank three-dimensional Finger tensor (also called the left Cauchy–Green deformation tensor) is usually written

   $p_B (\lambda) = \lambda^3 - a_1 \, \lambda^2 + a_2 \, \lambda - a_3 \,$

   In this article, the trace a₁ is written I₁, the next coefficient a₂ is written I₂, and the determinant a₃ would be written I₃.
6. ^ Bower, Allan (2009). Applied Mechanics of Solids. CRC Press. ISBN 1439802472. http://solidmechanics.org/. Retrieved January 2010. 
7. ^ ^{a b} Ogden, R. W., 1984, Nonlinear elastic deformations, Dover

See also

• Hyperelastic material
• Finite strain theory
• Continuum mechanics
• Strain energy density function