Back in 1999 when I started my postdoctoral work, the "hot" and sexy topic to study on high-Tc superconductors was the underdoped regime of the phase diagram. This is where a bunch of exotic properties that are still mysterious appear, such as the pseudogap state, the unusually large energy gap in the superconducting state even though Tc is dropping, etc. I, on the other hand, decided to go the other way, into a territory that most people in the community thought was boring and not many exotica - the overdoped regime. This is where the cuprate compounds were highly overdoped with holes that its Tc also starts to drop. Most people thought that in this regime, the cuprates will now start to resemble the conventional superconductivity - the Fermi Liquid picture will start appearing. That's why they didn't find it interesting since this is well-known.
In any case, I jumped head first into it. The main reason is that I obtained the most overdoped samples of the Bi-Sr-Ca-Cu-O family ever created and has never been studied. So that in itself is new. Using angle-resolved photoemission spectroscopy, I studied the spectral function along 2 high-symmetry directions of the crystal structure. We managed to obtain two very important results:
1. The spectral peak (or the quasiparticle peak) in the antinodal direction persists well into the normal state. This is different than the optimally-doped and the underdoped cuprates where the peak disappears above Tc.
2. The behavior of the quasiparticle spectra now tends to approach that of the Fermi liquid predictions with some deviation. Such behavior was only predicted, but never measured before, for the overdoped regime.
We thought we had a good enough result to submit to PRL (and we did). However, during the writing of this paper, I attended an electron spectroscopy conference in Berkeley and presented a poster of the result there. My postdoc supervisor alerted several of his contacts at Stanford (he graduated from there) regarding the results we had obtained and the fact that I will be presenting them there. So as expected, a bunch of Stanford people came over to the poster, especially one person who showed a keen interest in it. I've met him before at previous physics conferences, so we were on friendly terms. He was interested in it because the existence of the peak way above the normal state slightly contradicted a paper that he published in Science about a year earlier.
We had a lively and interesting conversation. He paid extra attention to the data that were on the poster and we exchange several info regarding it.
About 8 months later, a preprint came via e-mail from him to me and my boss. In it, he had made the same experiment on an identically-doped material. He reproduced and confirmed all that we had discovered, but he found something else, something that WE MISSED!
The Bi-Sr-Ca-Cu-O that we were looking at has 2 CuO layers per unit cell. This compound is a popular material used in photoemission because one can obtain a large single crystal, but more importantly, it is easily cleaved parallel to the CuO planes. Thus, you can obtain a clean, freshly cleaved surface in vacuum. Early photoemission spectroscopy done on this material in the optimally-doped and underdoped regime have revealed a wealth of information, but many reported that they did not detect the bands coming from the 2-layer CuO planes - the so-called bilayer bands. All that have been observed so far is just a single band.
Now this is puzzling from the band-structure point of view because when you have two "active" layers, you should get an interaction between them. In fact, there have been theoretical calculations showing that the coupling between the two layers will produce a "split" bilayer bands resulting in the bonding band, and the antibonding band. Angle-resolved photoemission should be able to detect this, and none has been reported.
... till that manuscript. The Stanford guy, after looking at my data, went back and obtained some highly overdoped samples. He and his group looked at the same symmetry directions as I did, but in a different orientation of the detector. When they did that, they CLEARLY noticed something that appear to be the long sought-after bilayer split bands! They analyzed the data further and the degree of splitting agrees with theoretical predictions. They also showed that the split band doesn't appear when they used the compound with just a single CuO plane.
The manuscript was sent to us for us to comment on. They acknowledged the fact that we gave them the idea. In any case, both my paper and the Stanford paper were published in PRL. I went back and looked at the raw data that I obtained, and a head-smacking moment occured several times when I realized that I actually DID see the bilayer splitting in the data. Although the configuration wasn't idential to the Stanford measurement, it was also very clear in the data I collected. Had I realize what they were, we could have easily made appropriate further studies on it and would have gotten the same thing.
The bilayer splitting since then had received a lot of attention. It is another puzzle that has been solved, and it gave theorists the energy scale of coupling that occurs between intra-CuO planes. My subsequent meetings with the Stanford guy has been very pleasent - in fact we have become very good friends and continued our professional contact. He has publically admitted that the impetus to look for the bilayer bands came directly from the data he observed on my poster. While I was happy with what we observed and published, there is always in the back of my mind, the regret of not looking any closer at what I observed. I think I was thrilled at getting the initial data that clearly was new, without realizing that there's another even more spectacular discovery hiding in the same set of data.
No matter how much we know, or how hard we work, and element of chance, luck, and pure accident/coincidence can play a huge role. Someone once said that the more prepared you are, the luckier you get. I wish I was more prepared to have not let this big one get away...
Zz.
No comments:
Post a Comment