NLOPredict

A data analysis and prediction tool for the study of nonlinear optics.

  1. Introduction
  2. Download and Installation
  3. Getting Started
  4. Explaination of Features
  5. Instruction on how to preform a Gaussian calculation for input tensor data
  6. Bugs
  7. Features to be implemented: higher priority first
  8. Favorite PDB files
  9. Creative hints for Chimera

1. Introduction

When you look at the things around you, colors emerge from the absorption of light by molecules and atoms in the materials you see. Absorption is a single photon process where the electrons in molecule are excited in the presence of this light. If a much more intense light source is used, such as a laser, multiple photons can interact with a molecule at the same time.

Nonlinear Optical Processes

The most common nonlinear optical process is second harmonic generation. If you have ever seen a green laser pointer then you have witnessed it in action. A green laser pointer has a crystal inside that contains molecules which simultaneously absorbs two photons of infrared light and emits one green photon. There are many other types of nonlinear optical process such as frequency summation and frequency difference generation. Using lasers, scientists can use these processes to study molecules. Nonlinear optical processes can tell us more about the molecules being studied than simple absorption of light can.

Orientation of molecules

Nonlinear optical processes that involve an odd number photons (3, 5, etc.) only occur where molecules are not randomly oriented. So they do not occur in liquids and gasses because of a cancellation effect. They do occur at surface and interfaces where molecules have some orientational order. This is a useful tool! Scientists can use nonlinear optical process to measure how molecules are oriented and because the measurement is sensitive to changes in molecular conformation the signal measured will change with orientation. It is conventional do describe the orientation of a molecule with respect to a surface coordinate system by the use of 3 Euler angles.

Tensor elements

The intensity measured in an experiment generally arises from the coherent interference of several nonzero elements within the surface tensor. The interaction of one photon of light with a molecule can be described by a simple vector called a transition dipole moment (Mu). Two photons interacting with a molecule is described by a 3 by 3 matrix termed a transition polarizability (Alpha). A three photon process is described by a 3 by 3 by 3 tensor called a transition hyperpolarizability (Beta). So, there is a tensor of varying size that discribes the nonlinear response of an individual molecule. To descibe a surface, simply sum up the tensors for all the molecules which results in a surface tensor (of the same size).

Putting it all together in NLOPredict and using Chimera

There are two pieces of information needed by NLOPredict to predict nonlinear optical properties. (1) The molecular tensor elements are needed for a particular nonlinear optical process and (2) the molecular coordinates or positions of the atoms are needed. Once this data is inputted the molecular tensor in the surface coordinate system is calculated as well as the surface tensor elements, which can be directly connected to experiments. When viewing the molecule in the chimera window we visualize the molecule sitting on a surface which is the plane of the screen with the Z-axis pointing out at you, the Y-axis pointing up and the X-axis pointing to the right. Upon rotating the molecule around (changing it's orientation) the predicted change in the nonlinear optical response is readily viewable in the NLOPredict window.

Short video introduction to NLOPredict

After reading the above introduction it is benificial to see these things in action. Here is a short (~3 min.) video describing these things in real time. It is an .mp4 file that can be viewed with the current version of QuickTime.

2. Download and Installation

3. Getting Started

  1. Start up Chimera
  2. Go to File > Fetch by ID
  3. In the Fetch Structure by ID window select PDB ID
  4. Type in 2BRD and click the Fetch button
  5. Go to Tools > Indiana/Purdue > NLOPredict
  6. Click the Structure pull down menu and select #0 2BRD
  7. NLOPredict has now performed several calculations and is ready to configure

4. Explaination of Features

5. Instruction on how to preform a Gaussian calculation for input tensor data

Electronic transitions

geometry optimization; zindo/cis - zindo=values bad but differences good

Vibrational transitions

geometry optimization; iop(7/33)=1

6. Bugs

7. Features to be implemented: higher priority first

  1. Work on Export Config file.
    1. Export Contour plot numbers.
    2. Filename of the .pdb file.
    3. Input values of the tensors.
  2. Fix single molecule mode.
  3. Update help file.
  4. Give a name to groups of tensor values entered. Add an option to save values for later recovery (editable dictionary structure).
  5. Import molecular coordinates and tensors from Gaussian quantum calculations.
  6. New plots (3 histograms) for distributions in Euler angles only for proteins.

8. Favorite PDB files

9. Creative hints for Chimera

Lets say you have a protein and you only want to consider the helix response. Typing in the following commands will allow you to delete everything from the protein except the helices and then calculate the NLO response from the remaining structure. The following will assume that you have a protein file open and it has helices to view.

  1. In the Chimera window go to Favorites > Command Line. This will open a command line at the bottom of the window.
  2. sel :/isHelix
  3. sel invert
  4. del sel

Line 2. can be replaced with any of the following for a different selection.