Discussion Forum

You have not defined the units. I added them and guessed that i01 = 2.7e-11 A/m^2 instead of 2.7e11. The parameters K1 and K2 remained me a bit unclear, what are they? See figure.

Your definition of the conductivity, k, had a small error. In the conductivity the charge numbers are raised to the second power, hence there are no negative terms in the summation.

In the node Transport of Diluted Species you have to couple the voltage variable V from the electrostatics node, see figure.

The solution converged, but I do not understad what you are simulating. Everything is constant in the solution plots. Why cannot you use Nernst-Planck equation instead of diffusion and electrostatics separately?

Wish this helps
Lasse

Hello Lasse ,

I actually tried with units too. It doesn't work Then I retried by removing units to simulate exactly the one I have in converted 3.5a version.It is very weird when same thing doesnt work in newer version.

I have a 6 species corrosionproblem(Fe,FeOH,H,OH,Na,Cl).It has 3 heterogenous electrochemical equations(using bulter volmer relationship) to define flux of H, OH, Fe at crevice tip and a fixed concentration of all species on other end of the crevice(crevice mouth) to simulate bulk.
I am trying to define a charge density by electrostatics module that will effect the electric field(governing electrostatic potential) down the crevice.
The three flux equations represent Fe oxidation , water reduction producing OH species and hydrogen reduction.
To include homogenous equillibrium reactions Fe+H2O-->FeOH+H and water dissociation with equillibrium constants K1 and K2.I intend to work the mass transport problem that is constrained by equillibrium constant such that concentrations of all species satisfies the constraint at all times.
I am guessing it is do with mesh.I have tried running each module while defining a good initial value of 0.2 V but cracks up for second constraint of metal hydrolysis recation mentioned above.
I have attached my model with units too. I will play with the value of io1 as you suggested.I have also attached a paper that I am trying to reproduce.
Thank you Lasse for your time. I appreciate it !

Please use Nernst-Planck equation instead of electrostatics node. But you have one small mistake there: surface charge density is not k*Vx but eps*Vx where eps is the permittivity of the medium, for water 78.4*8.854e-12[F/m].

Ok Lasse. You mean use nernst Planck equation interface only with pdes instead of transport of diluted species and electrostatics. I'll try that and update.
Thank you.

Neither I see from your model how weak form nodes are connected to the transport problem. And disable convection.

Lasse

Lasse ,
As you suggested I tried with Nernst planck equation. To approach the problem I went module by module. The model converges upto the 1st weak form pde and the constraint is also handled well keeping ratio of H and OH to 10^-14 mol/dm^3 . But fails to converge for second constraint such that K2=FeOH*H / Fe .
The idea of using the constraint is being derived from the model i sent you privately.I believe the way pde talks with NPE module is the condition given in the constraint and the extra pde node we add using -ut*test(u) ressembling the dc/dt in the transport equation we are solving. Please correct my understanding if I am wrong.
I tried makin mesh finer I get some undefined value for c_Na which I fail to see why is it occurring.With coarser fails to converge with 1000 iterations.This is exasperating! I have spent almost a semester to figure this thing out.Please find the attached model
My aim to is to basically replicate the shape of concentration profile in Sharlands paper even if figures are some what off.
I tried using eps in original model i attached with transport and diluted species and electrostatics, doesn't really help.
I appreciate your time and any suggestions !
Regards
Hetal M

There are some problems with the variables definitions, inconsistent units etc. What physical interpretation do the variables "u" and "u2" have in the weak form PDEs?

I believe u and u2 represent concentration if I am assuming ut in ut*test(u) represent dc/dt. The inconsistency in units for default weak form pde is I believe being over ridden by second weakform pde node we defined. I have tried to play with making units consistent for default pde node.It doesnt seem to help. As I mentioned before, I have taken this ide from the model my advisor gave me and it works in that version and even converted version. I am almost a beginner with Comsol .
Thank you for you time Lasse
Regards
HM

Hi Lasse,

We wanted to see if you can assist us with a simulation we are having. Attached is the file of out chip design from AutoCAD imported to COMSOL. We are having trouble setting up the domains and extruding the design to assign materials. The simulation is a DEP chip what should be surrounded by water. The electrodes are made of gold with a permittivity of 2. Any assistance you can offer would be greatly appreciated.

Hi Lasse ,
Do you think I am stuck in dead end probem? do you have any other suggestions or tips that I can try to implement !
Thank you
H M

You must change the name of the variable from u to the real one. Ho could Comsol guess that u is concentration?

What would you like to extrude?