In Summer thru Fall 2017, I used UV-Vis spectroscopy to investigate the bonding behavior of telechelic metal-coordinating polymers in hydrogels under various chemical stimuli such as pH and oxidants. I did this by fitting the spectral data to a gaussian peak superposition model. This repository contains the MATLAB scripts I've used to do this.

A hydrogel is a gelled network of polymers with a dispersed substantial amount of water. They have impressive properties and potential for applications -- for example, hydrogels are similar to many of the viscoelastic tissues in the body required to frequently store and dissipate energy (cyclic load), suggesting biomedical applications like replacing damaged cartilage.

In the intertidal zone of beaches, mussels, like the one shown here, remain fixed to inorganic substrates in changing environments. This includes wet and dry environments and the cyclic load from bombardment by turbulent waves. The mussel secretes a thread called the "byssus"; the cuticles on these byssal threads can remarkably withstand the harshness of these conditions.

Inspired by the mussel's cuticles, the Laboratory for Bio-Inspired Interfaces led by Niels Holten-Andersen studies synthetic metal-coordinating polymers that exhibit promising mechanical properties such as self-healing and tunable viscoelastic relaxations.

To go deeper into the details: by combining various metal salts with catechol- or nitrocatechol-modified polyethylene-glycol (PEG) polymers, we are able to form strong, hydrophilic coordination polymer networks that impressively mimic the properties of the mussel's cuticles. Due to the unique, metal-dopa bonds that form between the metal ion and the ligand that spontaneously reform when broken and are only slightly weaker than a covalent bond, the formed hydrogel is soft, yet can be subjected to large amounts of strain and withstand underwater conditions, just as the mussel's cuticles can.

As for the functions, most of the files are well-documented in the sense that you may type, for example, help gauss, to see the documentation in MATLAB's command window. Additionally, you may find this presentation helpful. Nevertheless, if something is unclear, feel free to reach out to me.

Both CSV2MATRIX and UVVIS convert a .csv file from the Denovix DS-11 FX+ spectrophotometer into a matrix. The former is an earlier and bulkier version of the latter, added for completeness. The matrix returned is used in several of the functions here including SPEC, GAUSS, and DECOMP.

SPEC simply plots spectrograms from the data matrix returned by UVVIS into a figure like the one below:

Given an output from GAUSS, SPEC can also make a nice .gif slideshow.

If you have a list of values that correspond to each spectrograph (for my research, this would be pH values), you can use REDBLUE and COP to make color matrices ready for use by SPEC.

Using a Monte Carlo method, suggests a guess matrix Gi to be used with GAUSS.

The MATLAB function I called GAUSS takes in a matrix M of experimental data whose size is (the number of measurements in M)+1 rows by (the number of wavelengths in M) columns. The +1 is for the top row, which lists all the wavelengths, usually 220 thru 750 (nm). GAUSS also takes in a matrix Gi, the guess matrix, which is (the number of peaks used in your model) by 2. The first column of Gi represents positions and the second represents the widths (specifically FWHMs) for each peak in your model.

GAUSS takes M and Gi, and iterates to find a better match of the model to the data. The iteration is from a gradient descent strategy for trying to fit the peak positions. And at each iteration, it uses a non-negative least squares algorithm to calculate the best-fit peak heights and widths. After it's finished, GAUSS returns a structure W with several useful fields, such as an updated guess matrix G and a matrix H containing the best-fit heights of the peaks at each measurements. This H is used for HPH below.

An interactive version of GAUSS. DECOMP uses human input via keyboard to guide the fitting. In the first mode, you can change the current measurement you're looking at and update the guess matrix by one iteration. In the second mode, you can move around the peaks individually.

The sensitivity affects by how much a peak is adjusted in the second mode and by how many solutions are skipped when browsing in the first mode.

Given W from GAUSS, plots the height of each peak versus the pH.

COG is a function created to help me with parsing arguments.

Asks a yes-or-no question and returns a logical.

Prints a list of choices and waits for your input. Returns your choice.

Makes creating .gifs easier than by using GIMP or IMWRITE or RBG2IND. It takes the filename and the frame handle, then adds the frame to the .gif.

In order from most abstract to most specific