Menu

Discussion Forum

New Discussion
Filtra

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.

Electromagnetic waves, transient. Complex dielectric constant.

Posted 20 set 2010, 22:14 GMT-4 9 Replies

Alejandro Rodriguez Perez
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi,

I am calculating the electric field as a function of time for an STM junction. The dielectric constant for silver at 800nm is -27+.4*j. If I use scattered light time harmonic, it solves no problem but if I use transient, Comsol 3.5a gives me:
Error:
Failed to find consistent initial values.
Undefined_value_found_Linsolv
NaN_or_Inf_found_when_solving_linear_system_using_SOR

If I use transient, with a dielectric of 5+5*j solves no problem. with dielectric of -27, it gives me the same error.

Any ideas how to fix this will be greatly appreciated.

Thanks in advance.
Alejandro

9 Replies Last Post 31 ago 2016, 17:32 GMT-4
Yevgeniy Davletshin
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 set 2010, 16:46 GMT-4
Check boundary and initial conditions, it's not about dielectric function that you are using.
Check boundary and initial conditions, it's not about dielectric function that you are using.
Alejandro Rodriguez Perez
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 set 2010, 18:37 GMT-4
Yevgeniy,

thank you for your reply.
Without changing anything, if I use dielectric=5+.4*j I don't get the error. If I change it to -27+.4*j, then I get the error.

Alej
Yevgeniy, thank you for your reply. Without changing anything, if I use dielectric=5+.4*j I don't get the error. If I change it to -27+.4*j, then I get the error. Alej
Robert Koslover Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30 set 2010, 19:29 GMT-4
Any chance that the problem is simply that you have a *negative* value for the real part of the dielectric constant? That is not the behavior of any ordinary material. Are you trying to represent some kind of "artificial" dielectric, such as a meta-material?
Any chance that the problem is simply that you have a *negative* value for the real part of the dielectric constant? That is not the behavior of any ordinary material. Are you trying to represent some kind of "artificial" dielectric, such as a meta-material?
Alejandro Rodriguez Perez
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30 set 2010, 20:30 GMT-4
Hi Robert,

thank you for your response. Good point. I believe metamaterials have negative index of refraction. A negative real part of dielectric constant is representative of a metal (Ag, Au, Cu) at 800 nm. I took these values from a paper published on Physical Review b , volume 6 number 12, Dec 15th 1972 "Optical constants of noble metals". But I agree with you that Comsol is not happy with that big negative number in transient calculations.

Best,
Alejandro
Hi Robert, thank you for your response. Good point. I believe metamaterials have negative index of refraction. A negative real part of dielectric constant is representative of a metal (Ag, Au, Cu) at 800 nm. I took these values from a paper published on Physical Review b , volume 6 number 12, Dec 15th 1972 "Optical constants of noble metals". But I agree with you that Comsol is not happy with that big negative number in transient calculations. Best, Alejandro
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 1 ott 2010, 02:14 GMT-4
Hi

read carefully the chapter on the index notation int eh RF doc, COMSOL uses (for internal coherence) a different sign convention than what is usual in optics, so if you load in n&k pls check carefuly the signs, as in the publication worls there are two conventions and not always explicitely mentionned

--
Good luck
Ivar
Hi read carefully the chapter on the index notation int eh RF doc, COMSOL uses (for internal coherence) a different sign convention than what is usual in optics, so if you load in n&k pls check carefuly the signs, as in the publication worls there are two conventions and not always explicitely mentionned -- Good luck Ivar
Shakeeb Bin Hasan
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 1 ott 2010, 09:15 GMT-4
Ivar has a valid point there, in Comsol's formulation, positive (negative) imaginary part of dielectric constant denotes gain (losses). Make sure you are really doing what you want.

On the other hand, even with correct sign of dielectric constant, you will still have this problem. You cannot use negative real part of permittivity, in time simulations, without introducing dispersive character of your material. There are a few forum posts on this subject, some from me actually, that might give you some direction to proceed.

All the best
Ivar has a valid point there, in Comsol's formulation, positive (negative) imaginary part of dielectric constant denotes gain (losses). Make sure you are really doing what you want. On the other hand, even with correct sign of dielectric constant, you will still have this problem. You cannot use negative real part of permittivity, in time simulations, without introducing dispersive character of your material. There are a few forum posts on this subject, some from me actually, that might give you some direction to proceed. All the best
Alejandro Rodriguez Perez
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 6 ott 2010, 16:52 GMT-4
Shakeeb and Ivar,

thank you for the heads up. You are both right. E and M books like Jackson and Greiner have nc = nr + ni * j and ec = er + ei*j whereas Optics Hecht has nc =nr - ni * j which is the convention that Comsol follows.

Alejandro
Shakeeb and Ivar, thank you for the heads up. You are both right. E and M books like Jackson and Greiner have nc = nr + ni * j and ec = er + ei*j whereas Optics Hecht has nc =nr - ni * j which is the convention that Comsol follows. Alejandro
Alejandro Rodriguez Perez
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 8 ott 2010, 17:53 GMT-4
Hi everyone,

somehow is working now. In the help files, it says that if you want represent absorption in a material you can either use an imaginary dielectric constant an set conductivity to zero or use the real part of the dielectric and in put the conductivity as frequency of light times the complex part of the dielectric. I think my problem was that I was using both a complex dielectric constant and a conductivity in the transient calculation.

Thanks for your help
Alej
Hi everyone, somehow is working now. In the help files, it says that if you want represent absorption in a material you can either use an imaginary dielectric constant an set conductivity to zero or use the real part of the dielectric and in put the conductivity as frequency of light times the complex part of the dielectric. I think my problem was that I was using both a complex dielectric constant and a conductivity in the transient calculation. Thanks for your help Alej
Rishad Arfin
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 31 ago 2016, 17:32 GMT-4
Hello Perez,

I am going through the same issue right now. I have a frequency dependent permittivity of a material which is -400+85.57*i. If I compute the model using this in the permittivity then It exactly gives me the same error as you encountered.

What is your suggestion in this case ?

PS: I have already set the conductivity to 0 of that material but what should I do about the negative real part and Imaginary part of the permittivity of the material?

Please kindly anyone let me know.

Thanks,
Rishad
Hello Perez, I am going through the same issue right now. I have a frequency dependent permittivity of a material which is -400+85.57*i. If I compute the model using this in the permittivity then It exactly gives me the same error as you encountered. What is your suggestion in this case ? PS: I have already set the conductivity to 0 of that material but what should I do about the negative real part and Imaginary part of the permittivity of the material? Please kindly anyone let me know. Thanks, Rishad

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content