Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Conductivity and relative permittivity of a liquid
Posted Feb 9, 2011, 3:06 p.m. EST 6 Replies
Please login with a confirmed email address before reporting spam
Hello,
I am modelling a capacitive electrode that is used in a liquid to stimulate a neuron sitting on top of it. I have attached the file, and I reference several things in it below.
The idea is that I apply a voltage to an electrode (the ring) which acts as the bottom plate of a capacitor. The electric field should be contained in the capacitor, and the top plate should then also receive a voltage. This voltage creates a current in the surrounding conductive solution, which will stimulate a brain cell (not in the simulation). The middle electrode is a recording electrode, which works in the opposite way as the stimulating electrode. However, it doesn't need a top plate.
I am having trouble figuring out which physics model I should use. I have used the electric currents model, but it asks for the relative permittivity of the solution, which is a conductor (known as PDS) and thus I don't know the value of the permittivity.
If anyone has any suggestions as to what I could use in this case, as well as any suggestions to make my model any better I would greatly appreciate it.
Thanks so much.
Michael
I am modelling a capacitive electrode that is used in a liquid to stimulate a neuron sitting on top of it. I have attached the file, and I reference several things in it below.
The idea is that I apply a voltage to an electrode (the ring) which acts as the bottom plate of a capacitor. The electric field should be contained in the capacitor, and the top plate should then also receive a voltage. This voltage creates a current in the surrounding conductive solution, which will stimulate a brain cell (not in the simulation). The middle electrode is a recording electrode, which works in the opposite way as the stimulating electrode. However, it doesn't need a top plate.
I am having trouble figuring out which physics model I should use. I have used the electric currents model, but it asks for the relative permittivity of the solution, which is a conductor (known as PDS) and thus I don't know the value of the permittivity.
If anyone has any suggestions as to what I could use in this case, as well as any suggestions to make my model any better I would greatly appreciate it.
Thanks so much.
Michael
Attachments:
6 Replies Last Post Feb 11, 2011, 1:55 p.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I believe that EC should do, but indeed you must fill in physical properties. At least use close to realistic values, +/-20% of given parameter will not change drastically your results, and you can always run a sensitivity analysis on your case.
One thing though: be careful with your geometry, you have small objects embedded in larger ones, I see that your electrodes are protrudung by some 32.5 um from the large volume, these small features make the meshing go mad and you end up with million of elements, while a few thousands should do. Often it's useful to slice your model into layers to ease the meshing. As well as you should consider some "air all around your electrodes on the top part too as you have also an E field therearound and you get physical (fringe) effects from both sides of the electrodes, but probably a few mm around your electrodes are sufficient, especially if you surround it by an "infinite elements layer",for the latter I prefere spherical physical domains for small MEMS devices, as the fields tend to drop off in a spherical way, seen from far away
--
Good luck
Ivar
I believe that EC should do, but indeed you must fill in physical properties. At least use close to realistic values, +/-20% of given parameter will not change drastically your results, and you can always run a sensitivity analysis on your case.
One thing though: be careful with your geometry, you have small objects embedded in larger ones, I see that your electrodes are protrudung by some 32.5 um from the large volume, these small features make the meshing go mad and you end up with million of elements, while a few thousands should do. Often it's useful to slice your model into layers to ease the meshing. As well as you should consider some "air all around your electrodes on the top part too as you have also an E field therearound and you get physical (fringe) effects from both sides of the electrodes, but probably a few mm around your electrodes are sufficient, especially if you surround it by an "infinite elements layer",for the latter I prefere spherical physical domains for small MEMS devices, as the fields tend to drop off in a spherical way, seen from far away
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Great, thanks for all your help Ivar.
For the properties associated with PBS, do you have a suggestion as to what to use for relative permittivity? The permittivity is not really a property that is associated with conducting materials, unless imaginary values are used.
Also, how do you suggest slicing my model to ease the meshing? Do I simply put a line through the model?
It seems that the version I uploaded didn't have the surrounding solution which should be above the electrodes, as you suggested to implement with air.
Thanks again. I really appreciate it.
Michael
For the properties associated with PBS, do you have a suggestion as to what to use for relative permittivity? The permittivity is not really a property that is associated with conducting materials, unless imaginary values are used.
Also, how do you suggest slicing my model to ease the meshing? Do I simply put a line through the model?
It seems that the version I uploaded didn't have the surrounding solution which should be above the electrodes, as you suggested to implement with air.
Thanks again. I really appreciate it.
Michael
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
if you have very different materials you can add a new principal "physics" node in the model and apply it to only one (or a few) materials, and change the way you define the material properties for these. I.e. complex permettivity or n,k are all possible (note the specific sign convention used in COMSOL for n-i*k material properties, as there are many "valid convention" around. COMSOL has choosen one, may not be trhe one you are used to.
In 2D you can slice up your volume with a line to introduce interiour boundaries and help the meshing. In defult union mode these geometrical objects are cut and identified as FEM elements (= domains and boundaries via the 3.5 geomanalyze() operator function, or via the "Finish" line in V4).
In 3D you need to cut your volumes with an extruded plane, or "box" with a volumic shape
--
Good luck
Ivar
if you have very different materials you can add a new principal "physics" node in the model and apply it to only one (or a few) materials, and change the way you define the material properties for these. I.e. complex permettivity or n,k are all possible (note the specific sign convention used in COMSOL for n-i*k material properties, as there are many "valid convention" around. COMSOL has choosen one, may not be trhe one you are used to.
In 2D you can slice up your volume with a line to introduce interiour boundaries and help the meshing. In defult union mode these geometrical objects are cut and identified as FEM elements (= domains and boundaries via the 3.5 geomanalyze() operator function, or via the "Finish" line in V4).
In 3D you need to cut your volumes with an extruded plane, or "box" with a volumic shape
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
As you predicted I have had some troubles meshing my design. I have split it with planes into four quarters, but I still run out of memory before the design is meshed. Do you have any other suggestions as to how I can simulate my design? I don't know how else to simplify it unless I only use 2D electrodes. I could try this but it would be nice to have the more accurate design.
Thank you.
Michael
As you predicted I have had some troubles meshing my design. I have split it with planes into four quarters, but I still run out of memory before the design is meshed. Do you have any other suggestions as to how I can simulate my design? I don't know how else to simplify it unless I only use 2D electrodes. I could try this but it would be nice to have the more accurate design.
Thank you.
Michael
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
check also in the "height" direction (thickness) you know (perhaps) you can turn off the scaling so that each axis has different scales, this is handy for MEMS or thin film systems. in v3.5 its in thelower border of your graphics view, double click on the "scale" name. In v4 its in the Definition view sub-node
It seemed for me that your electrodes crossed accroos the main boundary, try to curt your volume also that way.
Then you have the meshing options, that allow you to scale one direction (ie.e your thin thickness direction) then mesh it and scale it back down, this allows the mesher to work better, but you end up with mesh elements that have a poorer quality (but always better than no mesh ;)
--
Good luck
Ivar
check also in the "height" direction (thickness) you know (perhaps) you can turn off the scaling so that each axis has different scales, this is handy for MEMS or thin film systems. in v3.5 its in thelower border of your graphics view, double click on the "scale" name. In v4 its in the Definition view sub-node
It seemed for me that your electrodes crossed accroos the main boundary, try to curt your volume also that way.
Then you have the meshing options, that allow you to scale one direction (ie.e your thin thickness direction) then mesh it and scale it back down, this allows the mesher to work better, but you end up with mesh elements that have a poorer quality (but always better than no mesh ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I will try all of those suggestions. Thanks again!
Michael
Michael