This blog post is intended to document my processes regarding the DDSCAT Multilayer Project. The processes become tedious and some are small, so I will try to provide step by step descriptions of what to do.

Project Set Up

In joining with the Freed-Hardeman University Computational Chemistry research group, my collegues and I are investigating the interaction of a multilayered Gold (Au) nanoparticle with light of varying wavelengths. Specifically, we are examining a nanoparticle with a Au core and a Au shell separated by a layer of glass (SiO2).

Chapter 1: Outer Sphere

Section 1: Job Submission

Step 1: Preparing the Shape File

We begin by preparing the shape file to describe the structure to the DDSCAT program. This preparation is complicated for a multi-layered, multi-materialed sphere, so we use Dr. Elise Chaffin’s modified shape program (that originally produced nanostar shape files). Through this step process, we will gather a list of all files required for this process.

> ./nanostar3

Note: inputs ask for diameter, but you should put in radii in $10\, \times$ nanometers (ex. an outer sphere with a $25nm$ radius is entered as $250$).

Lattice spacing is typically $1$. Enter three lines of $1$ to give value to the x, y, and z lattice spacing.

This script will produce a shape.dat file.

Step 2: Dielectric Files

Before we set up the ddscat.par file, we need to import any dielectric text files we will be using for our layers. In this project, we need two: Au_diel.tab and SiO2_diel.tab. These are tab separated files and can be found in many directories. The dielectric files for this project are sourced from Johnson and Christy 1972 and I. H. Malitson 1965 respectively.

Step 3: DDSCAT Set Up

When we run the ddscat job, we give it a file containing instructions on how to run our job. The ddscat.par file contains a lot of lines, but we are only interested in a few. You should copy someone’s ddscat.par file into your directory and run the following command to begin editing (or use another editor if you prefer):

> emacs ddscat.par

We are interested in the following sections:

This portion contains the reference to the dielectric files. Make sure you have these files in the same directory as the ddscat.par file you are editing now and that they are spelled correctly here.

3         = NCOMP = number of dielectric materials
'Au_diel.tab' = file with refractive index 1
'SiO2_diel.tab' = file with refractive index 2
'Au_diel.tab' = file with refractive index 3

This effective radii section must contain the correct value. To calculate this value, you find the total radius of the sphere and convert it to microns (micrometers). In this snippet, the radius is $25nm$ = $0.0250 \mu m$

'**** Effective Radii (micron) **** '
0.0250 0.0250 1 'LIN' = eff. radii (first, last, how many, how=LIN,INV,LOG)

In order to exit the emacs editor, hold CTRL and type x then c (C-x C-c). This exits and askes you to save as well.

Step 4: SBATCH Preparation and Execution

Now, we are ready almost ready for our first job! However, we need to tell HPC how we want our job ran. We use the submit_varylamdaddscatnew_array.sh script. The following changes need to be made to this script:

We need to change the array values to whatever our desired wavelength range is. Here, we are designating our job to run from wavelength $400nm$ to $900nm$ with steps of $10nm$.

> emacs submit_varylamdaddscatnew_array.sh

#!/usr/bin/csh
#SBATCH --cpus-per-task=8
#SBATCH --nodes=1
#SBATCH --mem=50000
#SBATCH --job-name=ddscat
#SBATCH --partition=computeq
#SBATCH --out=dd_scat_%A_%a.out
#SBATCH --array=400-900:10
...

Finally, we are ready to run the first job with the following command:

> sbatch submit_varylamdaddscatnew_array.sh

If you are curious about your job’s progress, type:

> squeue | grep <your_username>

Section 2: Post Processing

Step 5: Post Processing Scripts

While your files are running, you can prepare the post processing scripts.

ddpostprocess_radial.par

The first of which is the ddpostprocess_radial.par that sits in your files directory.

Here, we need just need to edit the radius at which the Near Field is measured. We standardly measure NF at $1nm$ off the surface. For a sphere of radius $25nm$ we measure NF at $0.025\mu m + 0.001\mu m = 0.0260\mu m$

0.0  0.0  0.0  0.0260      = xorig(1), yorig(1), zorig(1), Rsurf(1) in physical units

submit_ddpost_radial2_EXTNF_multi.sh

In order to submit the post processing file, we edit the submit_ddpost_radial2_EXTNF_multi.sh file. Make sure to have the same array here as you did for Step 4.

> emacs submit_ddpost_radial2_EXTNF_multi.sh 

#!/usr/bin/sh
#SBATCH --cpus-per-task=8
#SBATCH --nodes=1
#SBATCH --mem=32000
#SBATCH --job-name=ddscat_double
#SBATCH --partition=computeq
#SBATCH --out=dd_scat_%A_%a.out
#SBATCH --array=400-900:10
...

Once the ddscat.par job is finished, you can execute this file using the following command:

> sbatch submit_ddpost_radial2_EXTNF_multi.sh

average_Esphere_python3.py

Lastly, once the submit_ddpost_radial2_EXTNF_multi.sh jobs are done, we then need to extract the data points from the individual lam_ folders. We extract the NF data using average_Esphere_python3.py making sure that the start, stop, and step are the same as our job’s specifications.

> emacs average_Esphere_python3.py

...
start=400
stop=900
step=10
...

Then execute the script:

> python average_Esphere_python3.py

average_EXTABS_sphere_python3.py

Similarly, we need to edit average_EXTABS_sphere_python3.py to add the correct start, stop, and step.

> emacs average_EXTABS_sphere_python3.py

...
start=400
stop=900
step=10
...

Then execute the script:

> python average_EXTABS_sphere_python3.py

Step 6: Rename Output Files

You will be running the same post processing scripts next chapter, and they will output files with the same name (which will overwrite the ones you have just created).

Therefore, we need to rename these files to give them a more descriptive name:

> mv average_Esphere.txt outer_sphere_average_Esphere.txt
> mv average_EXTABS_sphere.txt outer_sphere_average_EXTABS_sphere.txt

Step 7: Combine Output Files

This step is optional, but it is nice to consolidate the columns in the two outfiles that we actually use in analysis into one file. We use a simple script that accomplishes this task: combineEsphereEXTABS.py. This script does not need to be edited unless your directory is not of the format xcore/ynm-Au_znm-SiO2/. In that case, just comment out the lines indicated in the script.

Then execute the script:

> python combineEsphereEXTABS.py

Chapter 2: Inner Sphere

Section 1: Job Submission

Step 1: ddpostprocessing editing

In order to take measurements now at the inner sphere’s surface, we need to change the radius at which the data is collected. To do that, we need to edit ddpostprocess_radial.par and reduce the radius. Here, our core has a radius of $10nm$, so we measure NF at $0.011\mu m$.

> emacs ddpostprocess_radial.par

0.0  0.0  0.0  0.0110      = xorig(1), yorig(1), zorig(1), Rsurf(1) in physical units

The submission of this change is simple:

> sbatch submit_ddpost_radial2_EXTNF_multi.sh

Step 2: Create, Rename, and Combine Output Files

This step is the same as Steps 5-7. Make sure that your output files have already been renamed, and then we can run the following scripts:

> python average_Esphere_python3.py
> python average_EXTABS_sphere_python3.py
> mv average_Esphere.txt inner_sphere_average_Esphere.txt
> mv average_EXTABS_sphere.txt inner_sphere_average_EXTABS_sphere.txt
> python combineEsphereEXTABS.py

Chapter 3: Peak Refinement (Optional)

Section 1: Refinement Discernemnt

When a job is completely finished, sometimes the interval in which you ran the job is insufficient to describe a peak that rapidly changes:

I will usually run a second set of jobs that lower the wavelength interval and isolates the peak.

Section 2: Process Notes

Here is the list of steps you’ll need to accomplish to refine a peak in the data.

Step 1: Array Modification

In the following files, you will need to change the array to the desired refined peak:

After changing the arrays, follow the steps above to run the ddscat job and then the ddpostprocess job.

Step 2: Post Processing Modification

Post processing is almost the same as the two previous chapters, but you will need to be careful of the names of the text output files. Therefore, I recommend renaming them using the following commands:

> mv average_Esphere.txt outer_sphere_average_Esphere_refined.txt
> mv average_EXTABS_sphere.txt outer_sphere_average_EXTABS_sphere_refined.txt

Step 3: Combining the Refined and Original Data

Finally, use the the addRefinedPeaks.py to combine the two data sets together. It will ask you for the path of the two text files to be combined.

Step 4: Combine Outputs Again

Execute the script:

> python combineEsphereEXTABS.py

Chapter 4: Image Creation (Optional)

Section 1: Mayavi Images

Refer to this blog.

Section 2: Spectra Images

Currently, I am using a Jupyter Notebook on the HPC (Tutorial) to manually pull all the data into graphs of my choosing. However, I have plans to make this process simpler by using a database containing the data and the metadata about each job and a connection to a Jupyter Notebook coded for image processing. That way, you can command (via SQL) what data you want to compare and have the Notebook visualize the data quickly. Here is a blog I am keeping updated on my progress: Au-SiO2-Au Multilayered Nanoparticle Project.

Appendix 1: Au Shell Only (Optional)

In the process of writing the manuscript, we realized that we need some data on just a Au shell. I created five jobs, each having a shell of thickness 10nm at different distances (outer radius 22.5nm to 32.5nm in intervals of 2.5nm).

To create these jobs, the nanostar3.sh file was used. The core layer, however, needed to be SiO2 instead of Au, so in the ddscat.par file, I changed the core refractive index 1 to be SiO2. The radius of the core was then set to 10nm in the nanostar3.sh script. This number should be arbitrary because the program should set the core and middle layer to the same marterial, causing it to be continuous.

File Structure:

<
Previous Post
Au - SiO2 - Au GIF Creation Notes
>
Next Post
Manuscript Process