Getting Started

Overview

Lightpile is a tool to simulate dipole emitters in thin film stacks. This is a physics problem involving eletromagnetism and a bit of quantum mechanics. You might be interested in using Lightpile if working in the field of (organic) light emitting devices, known as (O)LEDs, or quantum dots.

At the moment Lightpile calculates the emission behaviour of a monochromatic dipole emitter of arbitrary orientation within a stack (pile) of thin films.

It is written by Richard Pfeifer and licensed under the GNU AGPL license.

Installation

On Linux

You can build Lightpile from source with the following major dependencies (for a full list see requirements.txt)

  • Python 2.7
  • numpy (1.9.2), matplotlib (1.4.3), scipy (0.15.1)

Direct install with scipy available

If you have a Python interpreter with scipy installed and want to install Lightpile into this environment it is as easy as

$ pip install git+https://github.com/ri-p/lightpile

I do recommend to follow the second way of using a virtual environment in order not to clutter your python installation with numerous packages.

Install Lightpile into a virtual environment

It is recommended to install Lightpile into a virtual environment. Follow these steps, assuming you start with a blank installation of Ubuntu. They have been tested with Ubuntu 14.04 64bit.

  1. Get a virtualenv with pip
$ sudo apt-get install python-pip
$ sudo pip install virtualenvwrapper

Then add the following lines to your ~/.bashrc (for details see here)

export WORKON_HOME=$HOME/.virtualenvs
source /usr/local/bin/virtualenvwrapper.sh

and create your virtual environment

$ source ~/.bashrc
$ mkvirtualenv env-lp
  1. Install numpy, scipy, matplotlib
(env-lp)$ sudo apt-get install python-dev g++ libpng-dev libfreetype6-dev libjpeg8-dev liblapack-dev gfortran
(env-lp)$ pip install numpy==1.9.2
(env-lp)$ pip install matplotlib==1.4.3
(env-lp)$ pip install scipy==0.15.1
  1. Install Lightpile
(env-lp)$ sudo apt-get install git
(env-lp)$ pip install git+https://github.com/ri-p/lightpile

On Windows

Lightpile depends on the numpy-package. Due to the difficulties of compiling numpy from source on Windows we suggest using a Python distribution with precompiled scientific packages, i.e. Anaconda or Canopy. The following instruction uses Anaconda.

  1. Install Anaconda and prepare a new conda-environment ‘env-lp’

Download Python 2.7 Anaconda for Windows and install.

$ conda create --name env-lp python==2.7 numpy==1.9.2 scipy==0.15.1 matplotlib==1.4.3

  1. Install Lightpile

    $ activate env-lp
(env-lp)$ pip install https://github.com/ri-p/lightpile/archive/master.zip

Minimal Example

As an introduction we reproduce the situation shown in Fig. 10 of the paper of Ford and Weber

In your favourite text editor create a file with the following content and name it ‘ford_fig10.lps’ (lightpile scene):

# Vertical dipole above dielectric layer on silver mirror.
#
[materials]
# name      n       k
air         1.      0.
dielectric  1.4142  0.
ag          0.0767  4.366

[emitters]
# name      orientation     iqe
red_emitter vert            1.0

[stack]
# describe stack using 'materials' and 'emitters' names
air         0       # materialname  thickness in nm
red_emitter         # emittername
air         0.2
dielectric  800
ag          0

[dipolestudy]
# we start a task 'dipolestudy'. It needs to arguments:
# 1) the wavelength of the emitter is 633nm
spectralpoint wavelength nm 633
# 2) the intervall of directions or modes is: u = 0.1 ... 1.6 (no unit)
#    with an automatic adaptive sampling point spacing
angularrange u None 0.1 1.6 adaptive

[graph_p_a]
yscale log
ylim 1e-3 1e3

Let’s start the calculation

(env-lp)$ lightpile ford_fig10.lps

Lightpile creates for output files

  • dipolestudy_graph_f.png
  • dipolestudy_graph_p.png
  • dipolestudy_data_f.txt
  • dipolestudy_data_p.txt

Compare dipolestudy_graph_p.png with figure 10 of the paper.

For explanation concerning the physical meaning of p(u) and f(u) please read the detailed Lightpile examples.

Once you are done with your calculations, leave the environment with

(env-lp)$ deactivate