Quest for an organic solar cell.

[Nordland]

Quest for an organic solar cell.
I thought I wanted to share some work that I have been puzzling with a bit. It is a search for theoretical organic solar cell.
I have choosen to share it both because it would be fun to hear any input on this matter, but this might not be worthy of a paper, but still I find it personally quite interesting.

The outset
Being able to build a organic solar cell, would be an extremely promising source of clean energy, and the master of getting the most out of the solar energy, is nature itself, so clearly there is some potential for finding a good candidate.
The type of systems that I will consider, would be build from three components. An optical receptor, that will absorb photons to create excited electrons, an organic wire that will transfer the excited electron away from the receptor, and finally an interface between the organic wire and metallic contact (drain).
I got inspired by beta-Carotene because it has in nature an active optical group in each end of a isoprene wire, so my goal is to model an organic solar cell based on a beta-carotene hybrid.

Limitations
The only computational resource I have, is my own laptop, so therefore there is a limitation of the convergence studies that I have done on some of the parameters.

Method
Initially, I started out using LCAO-DFT, but I found out that I could get the exactly the same results with SlaterKoster, so I decided to be more elaborate in terms of the calculated derived properties, as I could do it much faster with SlaterKoster.

The organic wire
So in order for the organic wire to be well suited for the optical solar cell, it must be able to transfer a electron from the receptor to the drain (conducting). Secondly it would be preferable if it was not optical active in the range of the sun-light.

So the system I modeled is Isoprene, which I built in the builder in a minute or so starting from nothing. Then I optimized the system using DFT taking both stress and forces into consideration.

The bandstructure of Isoprene shows that it is semi-conductor with a small band gap around 0.2 eV, meaning that it would NOT be conducting at T=0 Kelvin, but possibly at finite temperature. If not, this could potential mean that isoprene is unsatisfying as the wire. I will return to this in a second.

I also calculated the absorption in the wire, and found it to not absorb any significant light in the visible range of the sun light. I have plotted the optical absorption together with the spectrum of the sun-light at earth level being reflected from a mirror at 37 degrees. (This is the typical solar cell setup where you have mirrors collecting the sunlight onto your solar cell) Therefore in terms of not being an optical active, it is a promising choice for an organic wire.

MD-NEGF
Since the organic wire has to be conducting and being so for a wide range of temperatures it is important to investigate whatever it is conducting or not.
The first naive approach is to calculate the conduction for a series of different electron temperatures and see the behavior of that. (This is the blue curve in the graph below) But it is an open question how good this relates to experiments. The band gap is very small and any vibrations in the chain would properly affect this directly, which this approach would not show.

Therefore I decided to try something else ( and perhaps new? ). I performed a MD calculation of 10000 time steps for a series of different temperatures, and for each step and each temperature I calculate the conductance using NEGF, and using the Born-Oppenheimer approximation, I justified that that the average of the ensemble of the configurations could be a valid estimated for a measurable conductance.

As you can see this method (Green dots) predicted that the conductance of isoprene is almost 6-7 times as good at room temperatures as the normal approach predicts, and more important that isoprene is conducting in a much broader temperature range. It is therefore likely that isoprene is an ideal wire for connecting the receptor to the drain.


Comming up
* Connection with the drain.

[Nordland]

One of my co-worker ask me if I had looked into what kind of mechanics that were behind this dramatic increase in conductance as a function of temperature, and therefore I decided to investigate it a bit. The isoprene has a small band gap and the increase in conductance is likely to be due to the band gap becoming comparable with the width of the electrons Fermi temperature.

I found out the vibration responsible for this decreasing the band gap was the separation of the methyl group from the backbone.


The higher temperature would give rise to a stronger oscillation in the methyl-backbone bond, and therefore a higher conductance averaged over time.
In order to fully understand the mechanics of this I decided to calculate the conductance as a function of the perturbation of the bond distance between the methyl group
and the backbone, and as a function of the electron temperature.


As shown it can be seen that the stretching of the bond distance leads to an increase in conductance, and the effect is increased with the electron temperature, but the mechanics is present at all temperatures.

[Nordland]

Optical properties of beta-Carotene
beta-Carotene consist a chain of isoprene molecules and have two beta-rings in the each end. It is the beta-rings that are the optical active, as the isoprene
segments are optical inactive in the spectrum of sunlight.


Despite it being well-established that beta-Carotene is very optical active, I still wanted to calculate the optical spectrum for it, such that I could compare to experimental data.
But I could not find really good experimental data where I could deduce the concentration of the beta-Carotene solution on which it was obtained, so it is hard to compare directly, hence I decided to scale the experimental data to have the same magnitude as the calculated.


The calculated curve and the experimental curve has some agreement. From an experimental point of it is very hard to measure the absorption in the UV area, since most lenses are made of quartz and this makes it hard to verify if the peak 3.3 eV exist or not. The top peak around 2.5 eV, and it is left shoulder peak at 2.3 eV, is present in both the simulated and the experimental curve. The calculated curve predict absorption in the red and infrared area. This absorption could properly be removed by including the right solvent conditions, but for now I will let it be.

[Nordland]

Reserved for the drain.