Minggu, 16 September 2012

[bs] Simulasi perambatan gelombang elektromagnetik dalam medium berindeks bias negatif

Pada material jenis tertentu, respon ketika dikenai gelombang elektromagnetik (GEM) dapat dirumuskan oleh model Drude sebagai berikut :



Untuk mendesain material agar memiliki indeks bias negatif, kita perlu menentukan frekuensi GEM target (f0), frekuensi tubrukan antar elektron dalam material atau yang disebut juga frekuensi plasma (Omega_pe dan Omega_pm) dan frekuensi redaman (gamma_e dan gamma_m). Dari Pers (16a) dan (16b), kita pilih secara acak frekuensi target, sebagai misal f0 = 30 GHz (Omega = 2*pi*f0). Kemudian tentukan juga frekuensi plasma. Untuk sederhanaya Omega_pe = Omega_pm = Omega_p. Kemudian gamma_e gamma_m = gamma = 10^8 rad/s. Selanjutnya untuk melihat respon bahan yang diungkapkan oleh Pers (16a) dan (16b) di atas kita lihat dulu bagaimana bila Omega_p bervariasi dari 100 GHz, 266.5 GHz, 500 GHz dan 1000 GHz. Berikut ini plot kurva antara frekuensi target (Omega) vs Omega_p.


Gambar 1. Drude model. (atas) Harga real dari permettivitas dan permeabilitas relatif bahan. (bawah) Harga imajiner atau faktor redamannya. Faktor redaman berharga cukup kecil seperti yang kita harapkan.


Dari Gb 1, ketika Omega_p/Omega (w/w0) = 1, kita ketahui bahwa pada frekuesi plasma (Omega_p) 100 GHz belum tampak respon negatif, sedangkan pada Omega_p = 266.5 GHz, 500 GHz dan 1000 GHz medium memberikan respon negatif, yakni n = -1, n = -6 dan n = -27 dan memiliki faktor redaman yang kecil, yakni kurang dari -0.02.


GEM yang merambat searah sumbu-z, dengan medan listrik E serah sumbu-x dan medan magnet H searah sumbu-y, diberikan oleh Persamaan Maxwell sbb :


Kemudian untuk tujuan simulasi, Pers (22) diterjemahkan dalam finite difference menjadi :


Metode komputasi yang kita pakai di sini, adalah pengembangan dari algoritma sebelumnya. Selain itu dipakai juga pulsa singgle cycle karena lebih stabil, yakni:

dengan Tp = 1/f0 adalah periode pulsa. Hasil komputasi tampak dalam gambar sbb:


Hasil 1. Uji algoritma. Media uji berupa vacuum space yakni material dengan Omega_p = 0 GHz.

Gambar 2. Slab berwarna hijau berisi udara atau vacuum space. 

Tampak dalam Gb 2, GEM tidak mengalami gangguan, hasil ini cocok dengan teori dasar. Bila mengalami gangguan, algortima tidak dapat dipakai untuk perhitungan lebih lanjut. Selanjutnya, kita substitusi slab hijau itu dengan medium berindeks bias 1 (n = + 1), namun bukan vacuum space, karena ia memiliki frekuensi plasma (Omega_p = 100 GHz).


Hasil 2. Medium dengan n = +1 dan Omega_p = 100 GHz.

Gambar 3. Tampak gelombang mengalami gangguan dalam slab hijau akibat frekuensi plasma dan faktor redaman.

Hasil 3. Kini, kita akan mencoba memodelkan slab yang berindeks bias negatif. Dimulai dari n = -1, yakni ketika frekuensi plasma (Omega_p) = 266.5 GHz dan faktor redaman (gamma) = 10^8 rad/s. 


Gambar 4. GEM yang merambat dalam medium berindeks bias n = -1.


Hasil 4. GEM melewati slab yang berindeks bias (n) = -6, yakni ketika frekuensi plasma (Omega_p) = 500 GHz dan faktor redaman (gamma) = 10^8 rad/s.

Gambar 5. GEM yang merambat dalam medium berindeks bias n = -6.


Hasil 5. GEM melewati slab yang berindeks bias (n) = - 27, yakni ketika frekuensi plasma (Omega_p) = 1000 GHz dan redaman (gamma) = 10^8 rad/s.

Gambar 5. GEM yang merambat dalam medium berindeks bias n = -27. Dalam interval waktu yang sama, GEM di dalam slab memerlukan waktu lebih panjang untuk merambat.


Yeay, its cool isn't ? :D

Jumat, 07 September 2012

[bs] Relative Permittivity of various metals originally coded by Collin Meierbachtol


Relative Permittivity of various metals originally coded by Collin 
Meierbachtol (C)2009 based on the Brendel-Bormann model described by  
A. D. Rakic, et. al., App. Opt., vol. 37, no. 22, 1998.



Optical Properties of Metallic Films for Vertical-Cavity Optoelectronic Devices

Aleksandar D. Rakic, Aleksandra B. DjuriĊĦic, Jovan M. Elazar, and Marian L. Majewski  

Applied Optics, Vol. 37, Issue 22, pp. 5271-5283 (1998)
http://dx.doi.org/10.1364/AO.37.005271, D


Abstract
We present models for the optical functions of 11 metals used as mirrors and contacts in optoelectronic and optical devices: noble metals (Ag, Au, Cu), aluminum, beryllium, and transition metals (Cr, Ni, Pd, Pt, Ti, W). We used two simple phenomenological models, the Lorentz–Drude (LD) and the Brendel–Bormann (BB), to interpret both the free-electron and the interband parts of the dielectric response of metals in a wide spectral range from 0.1 to 6 eV. Our results show that the BB model was needed to describe appropriately the interband absorption in noble metals, while for Al, Be, and the transition metals both models exhibit good agreement with the experimental data. A comparison with measurements on surface normal structures confirmed that the reflectance and the phase change on reflection from semiconductor–metal interfaces (including the case of metallic multilayers) can be accurately described by use of the proposed models for the optical functions of metallic films and the matrix method for multilayer calculations.

© 1998 Optical Society of America

[Optical Society of America ]

Simulasi Matlab.












Sabtu, 01 September 2012

[bs] Finite Difference Time Domain by Professor Sina Ataollah Khorasani


Picture of  Professor  Sina Ataollah Khorasani  (taken from google+)


From: Andri Husein <andri84@yahoo.com>
To: Sina Ataollah Khorasani
Sent: Thursday, 30 August 2012, 5:06
Subject: Finite Difference Time Domain in Matlab


Dear Professor S. Khorasani

I do hope Professor S. Khorasani is fine.

Allow me to introduce myself. I am Andri Husein, graduate student at University of Sebelas Maret, Surakarta, Indonesia.

I am currently studying for the FDTD simulation of metamaterials. I get your FDTD code from the site, but have not managed to run it. Here is the output from the matlab command history. 

%%%%%%%%%%%%%% Matlab Command History %%%%%%%%%%%
(C) Copyright 2005

Sharif University of Technology
School of Electrical Engineering
All Rights Reserved

Computing the bandstructure via FDTD
Direction GX ??? Error using ==> fprintf
Invalid file identifier. Use fopen to generate a valid file identifier.
Error in ==> SaveComponent at 17
fprintf(Output,'T: %.10f dt: %.10f dT: %.10f Total Steps: %d\n',T,dt,dT,nT);
Error in ==> SaveField at 23
SaveComponent(fname,Data,nT,dTT,dt,T)
Error in ==> RunFDTD at 46
SaveField(USave)
Error in ==> ScanPath at 24
RunFDTD
Error in ==> BandStructure at 54
ScanPath(kxrng,kyrng,fname);
Error in ==> Go at 22
BandStructure
Error in ==> emfdtd at 23
Go
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

I really hope Professor S. Khorasani pleased to inform our fault location? Thank you very much.

Regards,

Andri Sofyan Husein

*****


Re: Finite Difference Time Domain in Matlab
FROM: Sina Ataollah Khorasani
TO: Andri Husein
Message flagged Thursday, 30 August 2012, 12:33

1) Please download the latest version from here: EmFDTD

2) Study README.TXT before running. I guess you have not.

3) You must make an 'Output' folder first inside the directory where your MATLAB code has been unzipped. Otherwise, the MATLAB code will generate the error message you are asking. For instance, let's say your code is unzipped here:

C:\EmFDTD\

Then you have to create the following folder before running the program:

C:\EmFDTD\Output\

Good luck

****

Re: Finite Difference Time Domain in Matlab 
FROM: Andri Husein
TO: Sina Ataollah Khorasani


Dear Professor Sina Ataollah Khorasani,

I finally have managed to run the fdtd-program from Professor Sina Ataollah Khorasani and get some impressive picture in attach. Thanks a lot :).

I am also interested to understand the physical meaning of fdtd program that Professor  Sina Ataollah Khorasani made. If Professor Sina Ataollah Khorasani permit, may I ask journals using this computational method ?  

Thanks and Regards,

Andri Sofyan Husein