Dispersion
The Dispersion subnode to Charge Conservation and Charge Conservation, Piezoelectric allows you to model possible dielectric losses in the material. The effect can be accounted for in Eigenfrequency, Frequency Domain and Time Dependent study.
Under the Material model, you can select one of the following option:
Debye (default)
The Dispersion subnode is contributing in a sense that the polarization contributions from two or more such subnodes added under the same parent node are additive.
Thus, you can add several Dispersion subnodes and even set them to use different dispersion models to build up an effective multipole model of any degree of complexity. For example, the Multipole Cole-Cole model can be set up this way.
In COMSOL versions 6.2 and earlier, the Dispersion node was exclusive in a sense so that only one subnode under the same parent was effective on the selected domain.
Debye and Multipole Debye model
For a single pole Debye dispersion model, you can enter the Relaxation time and Relative permittivity contribution. These parameters can be either scalar, or diagonal matrices to account for possible anisotropy in the material.
For the Multipole Debye model, you can enter any number of Relaxation time and the corresponding Relative permittivity contribution values using an input table with dynamically changed size. However, only scalar values can be entered.
Both models are directly suitable for Eigenfrequency, Frequency Domain and Time Dependent study types. And for these study types, you can specify how the relative permittivity input, εrS, on the parent node should be interpreted by selecting the Static response (the default value is Low frequency limit).
In the Thermal Effects section, you can select the Shift function type, which will be used to compute the effective relaxation time accounting for the temperature effects.
Constant Loss Tangent
In this case, you enter the Loss Tangent η(fc) together with the Center frequency fc. You also specify the model Bandwidth (decades) that defines a frequency interval centered at fc, in which the loss tangent will be approximately constant and equal to η(fc). You can also select the Accuracy for the approximation to be either Normal (default) or High. The software will use an equivalent Multipole Debye model and automatically deduce the necessary number of poles together with the values of the corresponding relaxation times and relative permittivity contributions, which will be used in computations to maintain the requested bandwidth and accuracy.
Cole-Cole and Havriliak–Negami models
These fractional models are extension of the Debye model. In some cases, they can provide a better representation of the relaxation spectrum.
For the Cole-Cole model, you enter a fractional order parameter, the Complementary broadening parameter αCC that controls a symmetric (in log-space) broadening of the spectrum. Note that its acceptable value is limited to the range from 0 to 1.
For the Havriliak-Negami model, you can enter two fractional order parameters, the Symmetric broadening parameter βHN with the value limited to the range from 0 to 1, and the Asymmetric broadening parameter γHN with the value limited to the range from 0 to 1/βHN.
At the Time Domain and Eigenfrequency section, you can compute an approximation for the complex valued relative permittivity contribution.
By default, the approximation is not computed. Without it, the Cole-Cole and Havriliak-Negami models are limited to Frequency Domain study type only.
The approximation computation needs to be done as a preprocessing step. Use the Compute approximation () action button that is available at the Time Domain and Eigenfrequency section. Once the computation has been performed, you can preview it using the Preview plot () action button that will become active at the section. The approximation corresponds to the Multipole Debye model. You can check the Show approximation data checkbox to inspect the computed parameter values.
Thus computed approximation will be used in Eigenfrequency and Time Dependent study steps. For the eigenfrequency computations, the eigenvalues problem will be always linear.
For all study types, you can specify how the relative permittivity input, εrS, on the parent node should be interpreted by selecting the Static response (the default value is Low frequency limit).
User Defined model
Enter expressions for the real and imaginary parts of the relative permittivity contribution, Δε and Δε''. Also enter the value of the real part of the contribution in the low frequency limit Δε'(0). Note that the imaginary part of the contribution is always assumed to be zero in the low frequency limit.
The expressions can be entered as functions taken directly from interpolated data, or can be analytical expressions of the frequency variable.
The internal variables for the frequency f and angular frequency ω are named phys.freq and phys.omega, respectively. Here, phys is the tag of the parent physics (for instance, solid). You can also use the variable freq defined by the solver.
At the Time Domain and Eigenfrequency section, you can compute an approximation for the complex valued relative permittivity contribution.
By default, the approximation is not computed. Without it, the User defined model is limited to Frequency Domain study type only.
The approximation computation needs to be done as a preprocessing step. Use the Compute approximation () action button that is available at the Time Domain and Eigenfrequency section. Once the computation has been performed, you can preview it using the Preview plot () action button that will become active at the section. The approximation corresponds to the Multipole Debye model. You can check the Show approximation data checkbox to inspect the computed parameter values.
Thus computed approximation will be used in Eigenfrequency and Time Dependent study steps. For the eigenfrequency computations, the eigenvalues problem will be always linear.
Note that for the User defined dispersion model, the relative permittivity input, εrS, on the parent node will be interpreted should be interpreted as the low frequency limit.
Thermal Effects
Dispersion properties depend on the temperature. For many materials, a change in the temperature can be transformed directly into a change in the time scale. Thus, the relaxation time is modified to aT(Tm, where aT(T) is a shift function.
Select a Shift function — None, Vogel–Fulcher, Arrhenius, Williams–Landel–Ferry, Tool–Narayanaswamy–Moynihan, or User defined.
When the default, None, is kept, the shift function aT(T) is set to unity and the relaxation time is not modified.
For Vogel–Fulcher enter values or expressions for these properties:
-
Reference temperature T0. The default is 230 K.
-
Activation energy Q. The default is 8000 J/mol.
For Arrhenius enter values or expressions for these properties:
-
Reference temperature T0. The default is 293.15 K.
-
For Williams–Landel–Ferry enter values or expressions for these properties:
-
Reference temperature TWLF The default is 293.15 K.
-
WLF constant 1 C1WLF. The default is 17.44.
-
WLF constant 2 C2WLF. The default is 51.6 K.
For Tool–Narayanaswamy–Moynihan enter values or expressions for these properties:
-
Reference temperature T0. The default is 293.15 K.
-
-
Activation energy fraction χ (0 < χ < 1).
-
-
For User defined enter a value or expression for the shift function aT.
Discretization
To display this section, click the Show More Options button () and select Discretization from the Show More Options dialog. Select the element order from the list box for the auxiliary electric field vector variables en.
You can change the solution algorithm by using the checkbox Use local time integration (checked by default). When unchecked, the solution of the equations for the auxiliary dependent variables in Time Dependent study will be governed by the time stepper algorithm used in the solver.
Location in User Interface
Context Menus
Ribbon
Physics tab with either Charge Conservation or Charge Conservation, Piezoelectric node selected in the model tree: