|
Lovely article. > Creating a working device typically takes them dozens of tries. And even then, each device behaves differently, so specific experiments are almost impossible to repeat. This is frustrating. You can make two twisted bilayer graphene samples at 1.10 degrees precisely (to within 0.01 degrees), and they will show completely different phase diagrams. One will superconduct, but the other will not. Things like that. What I learned recently is that every transport paper's twist angle report is wrong. The two hypothetical samples are actually probably not both 1.10 degrees. The uncertainty in twist angle should be of order 10-20%, rather than <1%. I even made this same mistake in my own paper last year! When creating these TBG samples, we used to literally tear the graphene in half, to get accurate relative alignment of the two halves. It was very clever, but it imparts a huge amount of strain to the two layers, generally of order 0.1-0.3%. This seems like a small amount, but moire patterns are extremely sensitive to this (roughly strain amount divided by twist angle, but the twist angle is very small), so the unit cell area gets modified by anywhere from 5-30%. In transport measurements, we can only measure moire unit cell area, but not twist angle. The number 1.10 +\- 0.01 deg is calculated assuming no strain, and this is an incorrect assumption. An STM paper from 2019 first pointed this out, but it was just a couple sentences buried in the supplemental material, and I (and most others) completely missed it. Even four years after moire materials took over the condensed matter world, we still don't understand the basics of how the materials work. It's very exciting, hot stuff. |
> When creating these TBG samples, we used to literally tear the graphene in half, to get accurate relative alignment of the two halves. It was very clever, but it imparts a huge amount of strain to the two layers, generally of order 0.1-0.3%.
Does "it" mean the mechanical tearing of the crystal imparts the strain? or instead is it the newly introduced surface boundary (in 1D) that is imparting strain?
[I ask because long ago I was familiar with some of the crazy surface physics that would happen in IV-IV and III-V systems, and just wondering what effects the 1D termination of the 2D lattice might cause.]