Material behavior is represented in physical theories by constitutive relations. A large class of physical behaviors can be represented by linear material models that take the form of a second-order tensor. The material tensor provides a relation between two vectors and can be written as
where d , f {\displaystyle \mathbf {d} ,\mathbf {f} } are two vectors representing physical quantities and K {\displaystyle {\boldsymbol {K}}} is the second-order material tensor. If we express the above equation in terms of components with respect to an orthonormal coordinate system, we can write
Summation over repeated indices has been assumed in the above relation. In matrix form we have
Examples of physical problems that fit the above template are listed in the table below.2
The material matrix K _ _ {\displaystyle {\underline {\underline {\boldsymbol {K}}}}} has a symmetry with respect to a given orthogonal transformation ( A {\displaystyle {\boldsymbol {A}}} ) if it does not change when subjected to that transformation. For invariance of the material properties under such a transformation we require
Hence the condition for material symmetry is (using the definition of an orthogonal transformation)
Orthogonal transformations can be represented in Cartesian coordinates by a 3 × 3 {\displaystyle 3\times 3} matrix A _ _ {\displaystyle {\underline {\underline {\boldsymbol {A}}}}} given by
Therefore, the symmetry condition can be written in matrix form as
An orthotropic material has three orthogonal symmetry planes. If we choose an orthonormal coordinate system such that the axes coincide with the normals to the three symmetry planes, the transformation matrices are
It can be shown that if the matrix K _ _ {\displaystyle {\underline {\underline {\boldsymbol {K}}}}} for a material is invariant under reflection about two orthogonal planes then it is also invariant under reflection about the third orthogonal plane.
Consider the reflection A 3 _ _ {\displaystyle {\underline {\underline {{\boldsymbol {A}}_{3}}}}} about the 1 − 2 {\displaystyle 1-2\,} plane. Then we have
The above relation implies that K 13 = K 23 = K 31 = K 32 = 0 {\displaystyle K_{13}=K_{23}=K_{31}=K_{32}=0} . Next consider a reflection A 2 _ _ {\displaystyle {\underline {\underline {{\boldsymbol {A}}_{2}}}}} about the 1 − 3 {\displaystyle 1-3\,} plane. We then have
That implies that K 12 = K 21 = 0 {\displaystyle K_{12}=K_{21}=0} . Therefore, the material properties of an orthotropic material are described by the matrix
K _ _ = [ K 11 0 0 0 K 22 0 0 0 K 33 ] {\displaystyle {\underline {\underline {\boldsymbol {K}}}}={\begin{bmatrix}K_{11}&0&0\\0&K_{22}&0\\0&0&K_{33}\end{bmatrix}}}
In linear elasticity, the relation between stress and strain depend on the type of material under consideration. This relation is known as Hooke's law. For anisotropic materials Hooke's law can be written as3
where σ {\displaystyle {\boldsymbol {\sigma }}} is the stress tensor, ε {\displaystyle {\boldsymbol {\varepsilon }}} is the strain tensor, and c {\displaystyle {\mathsf {c}}} is the elastic stiffness tensor. If the tensors in the above expression are described in terms of components with respect to an orthonormal coordinate system we can write
where summation has been assumed over repeated indices. Since the stress and strain tensors are symmetric, and since the stress-strain relation in linear elasticity can be derived from a strain energy density function, the following symmetries hold for linear elastic materials
Because of the above symmetries, the stress-strain relation for linear elastic materials can be expressed in matrix form as
An alternative representation in Voigt notation is
or
The stiffness matrix C _ _ {\displaystyle {\underline {\underline {\mathsf {C}}}}} in the above relation satisfies point symmetry.4
The stiffness matrix C _ _ {\displaystyle {\underline {\underline {\mathsf {C}}}}} satisfies a given symmetry condition if it does not change when subjected to the corresponding orthogonal transformation. The orthogonal transformation may represent symmetry with respect to a point, an axis, or a plane. Orthogonal transformations in linear elasticity include rotations and reflections, but not shape changing transformations and can be represented, in orthonormal coordinates, by a 3 × 3 {\displaystyle 3\times 3} matrix A _ _ {\displaystyle {\underline {\underline {\mathbf {A} }}}} given by
In Voigt notation, the transformation matrix for the stress tensor can be expressed as a 6 × 6 {\displaystyle 6\times 6} matrix A σ _ _ {\displaystyle {\underline {\underline {{\mathsf {A}}_{\sigma }}}}} given by5
The transformation for the strain tensor has a slightly different form because of the choice of notation. This transformation matrix is
It can be shown that A ε _ _ T = A σ _ _ − 1 {\displaystyle {\underline {\underline {{\mathsf {A}}_{\varepsilon }}}}^{T}={\underline {\underline {{\mathsf {A}}_{\sigma }}}}^{-1}} .
The elastic properties of a continuum are invariant under an orthogonal transformation A _ _ {\displaystyle {\underline {\underline {\mathbf {A} }}}} if and only if6 C _ _ = A ε _ _ T C _ _ A ε _ _ {\displaystyle {\underline {\underline {\mathsf {C}}}}={\underline {\underline {{\mathsf {A}}_{\varepsilon }}}}^{T}~{\underline {\underline {\mathsf {C}}}}~{\underline {\underline {{\mathsf {A}}_{\varepsilon }}}}}
The elastic properties of a continuum are invariant under an orthogonal transformation A _ _ {\displaystyle {\underline {\underline {\mathbf {A} }}}} if and only if6
An orthotropic elastic material has three orthogonal symmetry planes. If we choose an orthonormal coordinate system such that the axes coincide with the normals to the three symmetry planes, the transformation matrices are
We can show that if the matrix C _ _ {\displaystyle {\underline {\underline {\mathsf {C}}}}} for a linear elastic material is invariant under reflection about two orthogonal planes then it is also invariant under reflection about the third orthogonal plane.
If we consider the reflection A 3 _ _ {\displaystyle {\underline {\underline {\mathbf {A} _{3}}}}} about the 1 − 2 {\displaystyle 1-2\,} plane, then we have
Then the requirement C _ _ = A ε _ _ T C _ _ A ε _ _ {\displaystyle {\underline {\underline {\mathsf {C}}}}={\underline {\underline {{\mathsf {A}}_{\varepsilon }}}}^{T}~{\underline {\underline {\mathsf {C}}}}~{\underline {\underline {{\mathsf {A}}_{\varepsilon }}}}} implies that7
The above requirement can be satisfied only if
Let us next consider the reflection A 2 _ _ {\displaystyle {\underline {\underline {\mathbf {A} _{2}}}}} about the 1 − 3 {\displaystyle 1-3\,} plane. In that case
Using the invariance condition again, we get the additional requirement that
No further information can be obtained because the reflection about third symmetry plane is not independent of reflections about the planes that we have already considered. Therefore, the stiffness matrix of an orthotropic linear elastic material can be written as
C _ _ = [ C 11 C 12 C 13 0 0 0 C 12 C 22 C 23 0 0 0 C 13 C 23 C 33 0 0 0 0 0 0 C 44 0 0 0 0 0 0 C 55 0 0 0 0 0 0 C 66 ] {\displaystyle {\underline {\underline {\mathsf {C}}}}={\begin{bmatrix}C_{11}&C_{12}&C_{13}&0&0&0\\C_{12}&C_{22}&C_{23}&0&0&0\\C_{13}&C_{23}&C_{33}&0&0&0\\0&0&0&C_{44}&0&0\\0&0&0&0&C_{55}&0\\0&0&0&0&0&C_{66}\end{bmatrix}}}
The inverse of this matrix is commonly written as8
where E i {\displaystyle {E}_{\rm {i}}\,} is the Young's modulus along axis i {\displaystyle i} , G i j {\displaystyle G_{\rm {ij}}\,} is the shear modulus in direction j {\displaystyle j} on the plane whose normal is in direction i {\displaystyle i} , and ν i j {\displaystyle \nu _{\rm {ij}}\,} is the Poisson's ratio that corresponds to a contraction in direction j {\displaystyle j} when an extension is applied in direction i {\displaystyle i} .
The strain-stress relation for orthotropic linear elastic materials can be written in Voigt notation as
where the compliance matrix S _ _ {\displaystyle {\underline {\underline {\mathsf {S}}}}} is given by
The compliance matrix is symmetric and must be positive definite for the strain energy density to be positive. This implies from Sylvester's criterion that all the principal minors of the matrix are positive,9 i.e.,
where S k _ _ {\displaystyle {\underline {\underline {{\mathsf {S}}_{k}}}}} is the k × k {\displaystyle k\times k} principal submatrix of S _ _ {\displaystyle {\underline {\underline {\mathsf {S}}}}} .
Then,
We can show that this set of conditions implies that10
However, no similar lower bounds can be placed on the values of the Poisson's ratios ν i j {\displaystyle \nu _{ij}} .11
Geraldes DM et al, 2014, A comparative study of orthotropic and isotropic bone adaptation in the femur, International Journal for Numerical Methods in Biomedical Engineering, Volume 30, Issue 9, pages 873–889, DOI: 10.1002/cnm.2633, http://onlinelibrary.wiley.com/wol1/doi/10.1002/cnm.2633/full /wiki/International_Journal_for_Numerical_Methods_in_Biomedical_Engineering ↩
Milton, G. W., 2002, The Theory of Composites, Cambridge University Press. ↩
Lekhnitskii, S. G., 1963, Theory of Elasticity of an Anisotropic Elastic Body, Holden-Day Inc. ↩
Slawinski, M. A., 2010, Waves and Rays in Elastic Continua: 2nd Ed., World Scientific. [1] https://web.archive.org/web/20090210192845/http://samizdat.mines.edu/wavesandrays/WavesAndRays.pdf ↩
Boresi, A. P, Schmidt, R. J. and Sidebottom, O. M., 1993, Advanced Mechanics of Materials, Wiley. ↩
Ting, T. C. T. and Chen, T., 2005, Poisson's ratio for anisotropic elastic materials can have no bounds,, Q. J. Mech. Appl. Math., 58(1), pp. 73-82. ↩
Ting, T. C. T. (1996), "Positive definiteness of anisotropic elastic constants", Mathematics & Mechanics of Solids, 1 (3): 301–314, doi:10.1177/108128659600100302, S2CID 122747373. /wiki/Mathematics_%26_Mechanics_of_Solids ↩