Hi

I run ddscat 7.3,and calculate nearfield ,but only get electrical field, I change ddpostprocess.f90 according instruction in the end code but in output I cant get polarization vector,. Should I change in ddpostprocess.par? or ddpostprocess.f90 ?

Thanks,

in discussion DDSCAT / General discussion » Longitudinal Polarization

Dear Hirak,

Insufficient information to be able to comment on this. Size, aspect ratio, wavelength, orientation…?. What version of DDSCAT are you using? Did DDSCAT claim to meet the error tolerance?

Also: DDSCAT does not work well when the dielectric function is very large, so you may not be able to calculate for gold in the

infrared.

Respected Sir,

DDA solution for polarization considering a gold nanorod shows an interesting pattern. The well-documented pattern (i.e. maximum @90°and minimum @180°) was obtained when the transverse wavelength was considered and beta, theta, and phi were set to zero. Using the same settings, the polarization patterns obtained for the longitudinal wavelength show patterns which deviate from their physical picture.

Please suggest the required modification that I need to make for obtaining the correct polarization pattern.

Thanks and regards,

Hirak

I am using the following version (copied from the wxxxryyykzzz.sca files)

DDSCAT —- DDSCAT 7.3.2 [17.01.25]

Hi.

I want to calculate the polarization of two spheres.

One sphere has 1.33+0.01i refractive index and another sphere has 1.9+0.45i.

So I set ddscat.par such as :

2 = NCOMP = number of dielectric materials

'../diel/m1.33_0.01'

'../diel/m1.90_0.45'

Then, m1.33+0.01i is number 1 and m1.90++0.45i is nuber 2. Is it right?

And I set BAM file such as :

x y z a IC1 IC2 IC3 theta_DF phi_DF beta_DF

0.000000 0.000000 0.000000 0.500000 1 1 1 0.00000 0.00000 0.00000

0.000000 1.000000 0.000000 0.500000 2 2 2 0.00000 0.00000 0.00000

Because a sphere at (0,0,0) has m=1.33+0.01i which is number 1, so I set IC1=IC2=IC3=1.

And also at a sphere at (0,1,0) , I set IC1=IC2=IC3=2 because m=1.90++0.45i is nuber 2.

Is it right?

I can't sure that means of number of IC1, IC2, IC3.

What version of DDSCAT are you running? The output files such as wxxxryyykzzz.sca will begin with a line like

DDSCAT —- DDSCAT 7.3.2 [17.01.25]

which will tell you the vintage of your software. Early versions of DDSCAT did have bugs in the MPI implementation, but the current release was thought to handle MPI correctly.

in discussion DDSCAT / General discussion » Running ddscat with MPI

What version of DDSCAT are you running? The output files such as wxxxryyykzzz.sca

will begin with a line like

DDSCAT —- DDSCAT 7.3.2 [17.01.25]

which will tell you the vintage of your software. Early versions of DDSCAT did have bugs in the MPI implementation, but the current release was thought to handle MPI correctly.

Hi,

I am a new user of ddscat.

I try to calculate scattering and absorption efficiencies of metal nanoparticles of arbitrary shapes and sizes.

I did a serial calculation with a sphere easily. but when I try a parallel calculation with MPI, I find very high values.

It is the same for all the algorithms

ZBCG2 IT= 1 f.err= NaN

…

cgcommon ckpt 20

iter= 8 frac.err= NaN

…

GPBICG IT= 10 f.err= 6.167E+00

…

…

sqrt(rnorm/bnorm)= 3 174106950333229.

sqrt(rnorm/bnorm)= 4 NaN

Other codes using MPI have been successfully implemented in our system, and the job submission procedures are correctly implemented.

The administrator followed all the instructions in the USer Guide. Did we miss an important information?

Regards

Hi,

I am a new user of ddscat.

I try to calculate scattering and absorption efficiencies of metal nanoparticles of arbitrary shapes and sizes.

I did a serial calculation with a sphere easily. but when I try a parallel calculation with MPI, I find very high values.

It is the same for all the algorithms

ZBCG2 IT= 1 f.err= NaN

…

cgcommon ckpt 20

iter= 8 frac.err= NaN

…

GPBICG IT= 10 f.err= 6.167E+00

…

…

sqrt(rnorm/bnorm)= 3 174106950333229.

sqrt(rnorm/bnorm)= 4 NaN

Other codes using MPI have been successfully implemented in our system, and the job submission procedures are correctly implemented.

The administrator followed all the instructions in the USer Guide. Did we miss an important information?

Regards

Were you able to compile the 7.3 and run it? Are you getting poln file? In which example? In general, you should look at nearfield option (User Guide) and DDPOSTPROCESS code if you want to get polarizations close to the object and inside. You can plot electric and magnetic field in this case. It is possible to output polarization vector just inside the shape. Let us know how far you got.

You can think about equal volume sphere as if you were to melt your nanorod and create a sphere. Good way to understand DDSCAT is to begin with a simple sphere for size parameter x = 1 to 10 and compare the results with Bohren and Huffman BHMIE code. Next, begin with more complicated shapes. For scattering geometry (choice of polarization, etc) I would examine first a copy of Bohren and Huffman book. Piotr

in discussion DDSCAT / General discussion » How to simulate a spherical Target at the interface air/dielectric (epsilon dielectric is 2)

If you look at equation (9) of Draine and Flatau, 1994 (J. Opt. Soc. Am. vol 11, no 4, April 1994) absorption cross section C_abs is given by summation over all dipoles:

C_abs= 4pik/E0^2 SUM_j=1^N [Im(P_j*(alpja_j) conj(Pj) - 2/3 k^2 P_j^2] (eq. 9)

and efficiency Q_abs is normalized C_abs (see DDSCAT User Guide https://arxiv.org/abs/1305.6497, section Absorption and Scattering By Finite targets). DDSCAT is calculating polarizations P_j in EVERY point of the shape/object. Thus, you can split the sum given in eq. 9 into contributions from subset of shape locations - for example upper part of disk and lower part of a disk. So if this is what you are interested in, the answer is yes. I think that the practical way of doing it is to store P_j and recalculate C_abs yourself. For that you can use nearfield option in DDSCAT. But, you can also get to the DDSCAT code and change the summation there (you will need to know FORTRAN and recompile the code).

in discussion DDSCAT / General discussion » How to simulate a spherical Target at the interface air/dielectric (epsilon dielectric is 2)

Hello, is there a way to obtain scattering and absorption efficiencies of two materials separately when your example is concerned?

More specifically, suppose that I am trying to model a surface as a rectangular disk (as you describe above) and there is a spherical particle of different material on top of it. Can I have absorption efficiency of the particle on top separately? When I do the calculations what I see is the absorption efficiency of the whole system.

If yes, I will try to model core-shell particle on top of a surface (there will be at least 3 different materials! surface, core and shell). Can I calculate absorption efficiencies of each material? Is it possible?

Thank you for your help.

Have a nice day.

Thanks alot for attention

"Wxxxryyykzzz.poln" file is unformatted(binary file), and I cant read this file.I want to read data of file.

in discussion DDSCAT / General discussion » Error in code when running for CONELLIPS target.

As per the UserGuide, the first three shape parameters are the lengths/d of the *outer* ellipsoid,and shape parameters 4,5,6 are the lengths/d of the *inner* ellipsoid.

Shape parameter 1 should be *larger* than parameter 4, parameter 2 larger than parameter 5, and parameter 3 larger than parameter 6.

Hence if you wish to do concentric spheres, you could set shape parameters 1-6 to

60 60 60 40 40 40

in discussion DDSCAT / General discussion » Error in code when running for CONELLIPS target.

Hello, Sir

I am using DDA to calculate the optical properties such as absorption, scattering and extinction for "CONELLIPS" shape (Two -layers). I have no idea about CONELLIPS shape, this type of target geometry is in-built or not. I have only changed the target geometry-(ELLIPSOID to CONELLIPS) and the composition-(40 40 40 60 60 60 = shape parameters 1 - 6) and run the code for two different materials. After that, i got the error as given below:

Error: FATAL ERROR IN PROCEDURE: TARCEL

AX < BX

EXECUTION ABORTED

1. The current release of DDSCAT is 7.3.2. DDSCAT 7.2 is no longer supported.

2. It is unclear what you mean by "cannot compile 2xxxryyykzzz.poln". Please clarify.

in discussion DDSCAT / General discussion » How to simulate a spherical Target at the interface air/dielectric (epsilon dielectric is 2)

Unfortunately, in its present form DDSCAT is not set up to treat the problem of a finite target near a semi-infinite slab.

For DDSCAT, the closest approximation to the posed problem would be to calculate the finite target placed at the surface of a rectangular block with the desired dielectric function representing the semi-infinite substrate. Use a block that is as large as you can afford, and repeat the calculation for different block sizes to see whether the results seem to be sensitive to the size of the block. This method is inelegant and unfortunately not very efficient; memory limitations will limit the size of the slab, given that the interdipole spacing d needs to be small enough to resolve the target geometry.

If your target happens to be a disk, you can use target option DSKRCTNGL. If your target happens to be a rectangular block, you can use target option RCTGLBLK2. If your target has another geometry, you can use target option FROM_FILE, and use your own software to prepare a file with the list of dipoles locations and compositions.

DDSCAT can work with complex structures. DDSCAT will calculate Muller matrix elements for the far-field, and can also do near-field calculations if desired. However, if the desired target geometry isn't one of the built-in target options, you will need to either (1) modify DDSCAT or (2) generate the list of "occupied" lattice sites (and compositions) using your own software tools, and use the FROM_FILE target option in DDSCAT to read in the list of occcupied sites. Option (2) is generally simplest since is doesn't require andy changes to DDSCAT.

The UserGuide describes how the input file for target option FROM_FILE should be formatted. There is a sample calculation in the examples/FROM_FILE directory.

Keep in mind that if one has a multilayer target, the interdipole spacing d should be small enough to resolve the thickness of the thinnest layer. One should always do a "resolution study" to verify that your results are not sensitive to your choice of interdipole spacing d.

Dear Sir,

I am working with ellipsoid Nanomatryushka particles having two or more layers. I need to simulate values for the Muller Matrices for the kind of particles. Can I do it with DDSCAT? Can I model Fano Resonance with it?

If I need to modify the CONELLIPSE section for my use, how can I do it?

Regards,

Hirak