In 2014, Dr. Alfred Hübler designed an experiment with which he explored Complex Systems: "Stability and Conductivity of self assembled wires in a transverse electrical field." He explored the minutia of this complex system, and said later in an article, titled "Inhibitory nodes in self-assembling networks mimic the human brain," that this system puts us one step closer to designing "engineered brains." In this experiment, since we were unable to find any proven explanation for this phenomenon, we proposed a theory of our own. We then tested an aspect of this theory using conductance of various metals, and then found results which supported, but did not prove our proposition. But before talking about our results, we must discuss the terminology used in the original study.

From research and theorizing, we came to the agreement that the movement of these ball bearings was dictated by the passage of a positive or negative electrical charge through them. As is described in basic electrostatics, like charges repel, and opposite charges attract. Due to the like charges of all the ball bearings and the opposing charge between the ball bearings and the aluminum rim of the petri dish, the ball bearings would logically travel away from each other and towards the outer edge of the petri dish. Furthermore, as the ball bearings contact one another, their charge passes further from the center until it eventually escapes through the castor oil, so the bridges formed by the bearings between the cathode and anode of the power supply are similarly easy to explain. However, the trait of this phenomenon we struggled to grasp was the observed Lichtenburg-esque forks in the bearing trees. After contemplating this trait for some time and seeking the advice of Mr. Copeland, we realized that these forks are likely caused by the path the electrons take after reaching the ends of the bearing chains.

As the induced electric charge runs through chains of ball bearings, it is passed from one bearing to the next when two bearings come in contact. Electricity always takes the path of least resistance, and since both stainless steel and copper possess significantly higher conductivity properties than castor oil, the electric charge logically passes through the chain of bearings for as long a distance as possible. However, once the charge reaches the ball bearing furthest from the center of the pile, the charge begins to travel through the dialectic castor oil and falls into line with the high-entropy Lichtenburg travel patterns associated with unguided electric charge transfer. In simpler terms, once reaching the end of the ball bearing chain, the charge randomly travels from the bearings to the rim of the petri dish. It’s similar to how the spark on a grill or stove never travels between the lighter and burner in a straight line. Rather it moves in a random, arc-like shape. Of course, these arcs are slowed down by the presence of the castor oil, and the “trail” they leave in the oil by polarizing it as they pass draws the ball bearings along the path the arc took. Therefore, the chains created by the bearings form into Lichtenburg-esque branches which straighten out once they connect the two terminals of the power supply.

We struggled with the idea of confirming or denying our conceptual understanding of how these self-assembling wires functioned. Therefore, we decided that modeling the process with a computer would allow us to better grasp the concepts at play and determine if the ideal scenario we crafted played out in real life.

Simulating this entire process was a massive process, however, so we elected to start from the ground up with our real-life emulations. The first step we took was designing a physics engine capable of calculating electrostatic forces between two objects and adjusting the accelerations, velocities, and positions of the objects accordingly. Thus, we created two particle objects which each held a charge, mass, position, velocity, and acceleration, and used the below equation to calculate the electric force between the two particles.

In the video, the gray lines represent the path electrons have taken, the blue dots represent neutral oxygen molecules, and the boxes on the sides of the screen were our power supply terminal connections. In this simulation, black is negative and red is positive.

With additional time to pursue this project, we would continue to develop spark-making features (the physics of which we learned from this Walter Lewin Lecture) and eventually attempt building a proper simulation for the entire process of self-assembling wires.

Stainless Steel was found to have a slower circuit completion time on average. The average deviation of Stainless Steel was 4.08 with a percent AAD of 24.58%. The average deviation of Copper was 2.14 with a percent AAD of 16.96%.

The photos above highlight a phenotypic phenomena observed in self building wires of different metals. Stainless steel wires form fewer, longer branches from the origin while copper wires are more numerous initially. Branching in the stainless steel system tends to be contained to twofold forks, with the occasional threefold tree. The copper system contains trees with numerous branches.

https://en.wikipedia.org/wiki/Lichtenberg_figure

Lichtenberg Figures

https://www.nature.com/articles/srep15044.pdf

Self assembling wires original article

http://www.cs.unc.edu/~geom/LIGHTNING/lightning.pdf

Lightning simulation paper

https://www.frostphysics.org/electric-charge.html

Charges explanation

https://www.theworldmaterial.com/density-of-metals/

Density of Copper and Stainless Steel

https://www.tibtech.com/conductivite.php?lang=en_US

Conductivity/Resistance of Copper and Stainless Steel

https://www.youtube.com/watch?v=PeHWqr9dz3c&t=176s

Original Video on Self-Assembling wires

https://www.youtube.com/watch?v=2NKJ3x-iOAU

Other Video on Self-Assembling wires