What phenomena occur in a low voltage arc between copper and graphite electrodes, and why is the result dependent on electrode polarity? I was playing around with a laboratory power supply, drawing arcs between electrodes of various materials. I noticed phenomena that I found interesting, and couldn't really explain myself:

The circular electrode is a 5 euro cent coin, which is composed of steel with a rather thick copper plating. The long, thin electrode is a 0.7 mm diameter mechanical pencil lead (mainly composed of graphite) which has been previously slowly heated until red hot in order to drive off any volatile constituents that would otherwise rapidly vapourize and split it apart.
The power supply is a 30 V, 10 A switching mode laboratory power supply with configurable voltage and current limits. Both limits are set to their maximum values.
Graphite anode, copper cathode
When the positive lead is connected to the graphite electrode, it gets quickly consumed in a steady arc after contact is made. A black, brittle, hard and flaky residue is left on the copper surface, presumably graphite which has either melted or has undergone plastic deformation.

CH1 is the arc current, 1 A = 23 mV. CH2 is the voltage drop, measured at a less than ideal location, the power supply terminals. A zoomed in portion of the complete waveform is pictured on the right.


I find this surprising considering the high melting point of graphite and the presence of oxygen in the atmosphere. The copper plating suffers surprisingly little damage.
Graphite cathode, copper anode
When the negative lead is connected to the graphite electrode, an arc is difficult to ignite. An ohmic contact forms instead, and the electrode heats up extremely rapidly to incandescent temperatures. When an arc is finally struck by gently pulling away the cathode, the graphite is hardly consumed at all, but the copper experiences heavy pitting.

CH1 is the arc current, 1 A = 23 mV. CH2 is the voltage drop, measured at a less than ideal location, the power supply terminals. A zoomed in portion of the complete waveform is pictured on the right.

Why does the system behave so differently when the polarity is inverted? What exactly is the black residue composed of, and how is it deposited?
 A: I'll put this out here for now - it's not a complete answer yet, but it's longer than a comment will hold. Nice experiment!
The physics of carbon arcs is interesting for many reasons - one is the production of nanoparticles and nanomaterials. Fullerenes such as "Buckeyballs" and Carbon nanotubes (CNTs) are often produced using Carbon arcs, and the physics of the process is currently an active field of research.
While your experiment is in air, it's possible that a small amount of carbon combines with (uses up) the oxygen in the arc region, such that other processes still take place. You might try the experiment with some relatively inert gas like nitrogen or helium, or even the standard trick of using a match or candle to remove most of the oxygen first. But be careful.
You shouldn't breath the air around your experiment either - please do it in an exhaust hood!
For example, an abstract from a google search:

Abstract: The atmospheric pressure carbon arc in inert gases such as helium is an important method for the production of nanomaterials. It has recently been shown that the formation of the carbondeposit on the cathode from gaseous carbon plays a crucial role in the operation of the arc, reaching the high temperatures necessary for thermionic emission to take place even with low melting point cathodes. Based on observed ablation and deposition rates, we explore the implications of deposit formation on the energy balance at the cathode surface and show how the operation of the arc is self-organised process. Our results suggest that the arc can operate in two different ablation-deposition regimes, one of which has an important contribution from latent heat to the cathodeenergy balance. This regime is characterised by the enhanced ablation rate, which may be favourable for high yield synthesis of nanomaterials. The second regime has a small and approximately constant ablation rate with a negligible contribution from latent heat.

From: Self-organisation processes in the carbon arc for nanosynthesis, J. Ng and Y. Raitses, J. Appl. Phys. 117, 063303 (2015); http://dx.doi.org/10.1063/1.4906784
